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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
98020190-7121-4dbd-a562-e64a923ad4b9 | identifying-recurring-patterns-with-deep | 1806.05229 | null | https://arxiv.org/abs/1806.05229v3 | https://arxiv.org/pdf/1806.05229v3.pdf | Identifying Recurring Patterns with Deep Neural Networks for Natural Image Denoising | Image denoising methods must effectively model, implicitly or explicitly, the vast diversity of patterns and textures that occur in natural images. This is challenging, even for modern methods that leverage deep neural networks trained to regress to clean images from noisy inputs. One recourse is to rely on "internal" image statistics, by searching for similar patterns within the input image itself. In this work, we propose a new method for natural image denoising that trains a deep neural network to determine whether patches in a noisy image input share common underlying patterns. Given a pair of noisy patches, our network predicts whether different sub-band coefficients of the original noise-free patches are similar. The denoising algorithm then aggregates matched coefficients to obtain an initial estimate of the clean image. Finally, this estimate is provided as input, along with the original noisy image, to a standard regression-based denoising network. Experiments show that our method achieves state-of-the-art color image denoising performance, including with a blind version that trains a common model for a range of noise levels, and does not require knowledge of level of noise in an input image. Our approach also has a distinct advantage when training with limited amounts of training data. | ['Zhihao Xia', 'Ayan Chakrabarti'] | 2018-06-13 | null | null | null | null | ['color-image-denoising'] | ['computer-vision'] | [ 5.56049883e-01 -4.36148733e-01 4.57439482e-01 -5.57688117e-01
-1.01255810e+00 -4.25119460e-01 3.55012864e-01 -1.52840972e-01
-3.46366078e-01 3.74699235e-01 3.02661788e-02 1.21407714e-02
7.76974410e-02 -8.62124920e-01 -9.21731353e-01 -1.17987418e+00
1.80295277e-02 -1.92238986e-01 -4.40548845e-02 -3.46709698e-01
1.79406047e-01 3.98182571e-01 -1.50697410e+00 5.73078752e-01
6.09531820e-01 1.33517671e+00 3.04620743e-01 9.42362666e-01
1.30097978e-02 7.97647774e-01 -5.59456527e-01 -2.50340581e-01
5.26854515e-01 -6.53222144e-01 -3.08808535e-01 3.10341984e-01
7.59666562e-01 -4.53801364e-01 -5.71047723e-01 1.47864413e+00
3.17170054e-01 2.14808986e-01 4.26976174e-01 -4.42666858e-01
-7.59057522e-01 1.63193554e-01 -5.53527296e-01 1.43253475e-01
-1.08555825e-02 3.56543511e-01 6.74127400e-01 -8.32757771e-01
3.51471007e-01 1.17731917e+00 1.00357234e+00 4.87202227e-01
-1.70168066e+00 -3.83728534e-01 9.06772688e-02 2.06633024e-02
-1.05257297e+00 -6.11748219e-01 8.29967678e-01 -6.76739514e-02
6.60717964e-01 1.49873540e-01 5.14478028e-01 1.03247392e+00
2.03513458e-01 3.38724464e-01 1.47753215e+00 -4.74056751e-01
2.50151813e-01 -2.22198889e-01 1.39363453e-01 4.55779403e-01
5.72893396e-02 7.05118701e-02 -5.23728430e-01 4.53454852e-02
6.51876688e-01 1.09872892e-01 -4.71775770e-01 -1.49223223e-01
-8.22744668e-01 6.28739357e-01 6.95735395e-01 4.03928071e-01
-6.93044364e-01 3.71304482e-01 1.08235277e-01 6.73159778e-01
6.37672365e-01 1.88882247e-01 -3.43668967e-01 2.43198663e-01
-1.09927988e+00 1.33764297e-01 8.83315921e-01 2.11510852e-01
1.17728353e+00 5.10306582e-02 2.63048410e-01 9.74182367e-01
-2.66578025e-03 6.73053443e-01 2.79452235e-01 -1.33790624e+00
2.04269022e-01 1.17737927e-01 -3.32815908e-02 -1.32807982e+00
6.42733788e-03 -4.42358375e-01 -1.47891164e+00 7.26158738e-01
4.22729105e-01 7.93401822e-02 -1.36187565e+00 1.68962383e+00
5.06926030e-02 2.78890759e-01 1.28439814e-01 1.02336836e+00
5.84083438e-01 8.16576242e-01 -3.66483539e-01 -1.32808700e-01
9.23953414e-01 -6.07290328e-01 -5.60070753e-01 -5.54325879e-01
-1.80428311e-01 -9.55064774e-01 8.23791385e-01 9.10819232e-01
-1.04055500e+00 -7.44201303e-01 -9.23540115e-01 -4.56877761e-02
-2.55655944e-01 -4.49316949e-03 1.46533564e-01 5.55440366e-01
-1.23560750e+00 1.05061901e+00 -8.28827024e-01 -2.67070979e-01
4.66448992e-01 2.90276557e-01 -5.26072383e-01 -5.39731562e-01
-6.69178486e-01 6.85267746e-01 2.05207057e-02 5.35359025e-01
-1.10875285e+00 -5.20448983e-01 -9.18741524e-01 1.37605324e-01
2.46433526e-01 -4.42088485e-01 1.00992477e+00 -1.56749725e+00
-1.22647059e+00 7.37648606e-01 -5.26238561e-01 -3.83550227e-01
3.51157993e-01 -2.57070184e-01 -3.39705616e-01 1.56200558e-01
1.58029124e-01 2.96577364e-01 1.46276879e+00 -1.90713823e+00
-4.28334922e-01 -4.72170979e-01 -2.59917498e-01 -1.31742567e-01
-2.45611761e-02 -2.46830150e-01 -6.38542891e-01 -8.62344861e-01
5.69104195e-01 -5.30322134e-01 -4.98624831e-01 1.73370346e-01
-3.63957167e-01 4.31625903e-01 7.32268274e-01 -8.56112838e-01
7.05373347e-01 -2.48691654e+00 1.62736326e-01 7.05257356e-01
2.26105481e-01 5.50611392e-02 -5.89787722e-01 2.71102160e-01
-1.40589848e-01 -2.20364574e-02 -6.48632586e-01 -4.68773901e-01
-3.35190475e-01 3.44534039e-01 -4.66832459e-01 5.47659159e-01
4.14465696e-01 3.93804520e-01 -7.33126998e-01 8.63655005e-03
1.71113640e-01 6.29414856e-01 -3.20710182e-01 3.19273561e-01
-7.96444640e-02 5.97801328e-01 -2.10882332e-02 4.67072278e-01
9.72332239e-01 3.07542831e-02 1.47502020e-01 -4.74148095e-01
5.19316532e-02 -1.00084646e-02 -1.40596807e+00 1.52198446e+00
-4.76164401e-01 7.47452080e-01 7.43283570e-01 -1.37958097e+00
1.03107429e+00 8.18015188e-02 2.44893566e-01 -8.45956326e-01
1.90426335e-02 3.63842994e-01 -7.33148605e-02 -5.37970364e-01
2.52300084e-01 -3.04229468e-01 2.34652206e-01 3.96003127e-01
2.60129392e-01 -2.23566443e-01 1.09874822e-01 -2.99300868e-02
1.37294769e+00 -7.41499960e-02 -1.35126725e-01 -1.12525634e-01
4.68171865e-01 -1.80381387e-01 5.65553367e-01 1.28952277e+00
5.94485365e-02 1.10278797e+00 4.34765577e-01 -6.81162357e-01
-1.18052995e+00 -1.13182092e+00 1.82800926e-02 8.34200561e-01
1.98959365e-01 -8.34041014e-02 -7.78245211e-01 -4.54896927e-01
-4.25617248e-02 1.08213991e-01 -8.51657391e-01 -1.61563326e-02
-7.61767387e-01 -7.33471870e-01 1.60621315e-01 1.53794542e-01
6.23068690e-01 -9.70278203e-01 -2.66605794e-01 2.15615034e-01
-2.06985027e-01 -8.79955351e-01 -1.59682959e-01 7.03728676e-01
-9.91653264e-01 -1.02819240e+00 -6.64115727e-01 -9.99857008e-01
8.88173997e-01 6.00416660e-01 1.28183019e+00 4.40385014e-01
-2.71462262e-01 1.83047295e-01 -1.66596711e-01 3.68455537e-02
-5.04617691e-01 -4.63716865e-01 -1.33119315e-01 4.01784539e-01
1.40293732e-01 -8.08130503e-01 -7.86109209e-01 1.49627596e-01
-1.13367474e+00 -3.65624994e-01 8.09126616e-01 1.27446437e+00
9.12335515e-01 4.34688479e-01 2.71529526e-01 -8.68354857e-01
5.94230354e-01 -3.74893069e-01 -4.83202279e-01 7.71351755e-02
-2.32076585e-01 1.46013007e-01 8.39160144e-01 -3.99263293e-01
-1.08653045e+00 2.82378525e-01 -2.87547737e-01 -4.89951253e-01
-4.78191972e-01 6.69078708e-01 -1.07984856e-01 -7.32270777e-02
7.45328903e-01 3.74628484e-01 2.66936988e-01 -8.17170739e-01
3.83730322e-01 4.63644028e-01 1.17527807e+00 -3.31421524e-01
9.27807808e-01 7.74633586e-01 -8.83163512e-02 -9.48799968e-01
-8.18396568e-01 -5.13181627e-01 -6.10380948e-01 -2.54849158e-02
6.43271327e-01 -9.76602614e-01 -3.98592234e-01 7.97578633e-01
-1.05355859e+00 -5.48736930e-01 -2.04234302e-01 1.47164509e-01
-2.16394112e-01 5.10275304e-01 -7.82517195e-01 -7.38219976e-01
-1.34465396e-01 -9.79452193e-01 8.68090570e-01 1.31470695e-01
9.83697623e-02 -7.69159555e-01 -1.22387595e-02 9.79694575e-02
5.78753293e-01 1.35042131e-01 8.41606200e-01 -1.61806241e-01
-5.71215391e-01 -2.89462864e-01 -2.88335651e-01 9.87492681e-01
4.46114689e-01 -1.97160065e-01 -1.11614251e+00 -3.22055459e-01
6.18430734e-01 -3.25725406e-01 1.47161937e+00 7.31216490e-01
1.33145785e+00 -2.95326740e-01 1.47385120e-01 8.45110416e-01
1.79797649e+00 -1.17175847e-01 8.14529479e-01 3.16052496e-01
5.63295901e-01 6.74315274e-01 2.90246662e-02 1.41224831e-01
-2.14970529e-01 2.44462624e-01 5.65556347e-01 -5.01545429e-01
-2.23951444e-01 8.50081891e-02 2.68342227e-01 7.27924705e-01
4.62943353e-02 -2.41906159e-02 -5.20030320e-01 6.95222735e-01
-1.54467022e+00 -1.00140703e+00 -5.61163109e-03 2.10118413e+00
9.03205752e-01 7.69909844e-02 -3.56519848e-01 2.08825499e-01
6.74274802e-01 2.53929883e-01 -6.89841568e-01 -2.97382236e-01
-5.65560341e-01 6.39997900e-01 3.90662998e-01 5.30467927e-01
-1.27763879e+00 6.45570993e-01 6.41480017e+00 6.58988774e-01
-1.20870626e+00 -2.16879427e-01 1.06586957e+00 3.55381258e-02
-1.11040257e-01 -1.82334960e-01 -3.24619450e-02 3.12261403e-01
6.63566053e-01 3.74699235e-01 9.40852404e-01 4.00227278e-01
4.33366627e-01 -3.72375816e-01 -1.00342500e+00 1.01440930e+00
1.50014073e-01 -1.23864055e+00 -8.47152993e-02 -1.45002991e-01
9.64517653e-01 9.80105996e-02 2.19187304e-01 -1.05243042e-01
4.53463376e-01 -1.28019905e+00 5.67149401e-01 8.53256643e-01
4.46391910e-01 -5.56196213e-01 9.57495928e-01 2.70721883e-01
-7.89887249e-01 -3.86143364e-02 -5.15163600e-01 -1.48765087e-01
-4.80642207e-02 9.80624735e-01 -5.28085791e-02 2.88981169e-01
1.28335178e+00 8.52979124e-01 -5.07802248e-01 8.87640655e-01
-3.28816772e-01 7.57331371e-01 -3.35763395e-01 5.39948285e-01
1.14075556e-01 -5.20249248e-01 3.36517066e-01 1.20548463e+00
3.68031472e-01 1.55976981e-01 -3.90694104e-03 9.12879348e-01
-3.04401219e-01 -3.88235122e-01 -6.53060317e-01 3.13859493e-01
1.36159614e-01 1.31084251e+00 -6.89897537e-01 -3.61701220e-01
-4.32385653e-01 1.24367630e+00 1.79200530e-01 7.47643650e-01
-8.67329538e-02 -2.71370500e-01 8.22278738e-01 -1.56086147e-01
6.59107149e-01 -2.38470927e-01 -7.14472711e-01 -1.05861640e+00
2.18643352e-01 -1.37295330e+00 5.98805808e-02 -8.28076720e-01
-1.81618237e+00 6.80601001e-01 -7.18832076e-01 -1.17303753e+00
1.29711060e-02 -6.41139209e-01 -9.90048468e-01 1.14652753e+00
-1.44418406e+00 -8.35873365e-01 -5.75439215e-01 7.31869817e-01
4.20610219e-01 2.56649014e-02 7.54944682e-01 1.16960101e-01
-3.91421646e-01 2.19050661e-01 6.70508206e-01 3.07826728e-01
7.61372328e-01 -1.33377552e+00 4.33374643e-01 1.27587974e+00
2.79598653e-01 6.31222546e-01 8.49765599e-01 -5.58935583e-01
-1.52776098e+00 -1.05948746e+00 3.25896204e-01 3.41879167e-02
4.68095720e-01 -2.68657297e-01 -1.41790438e+00 3.26991469e-01
3.26780707e-01 3.74254555e-01 2.39422172e-01 -7.06170350e-02
-5.33518851e-01 -5.57626724e-01 -1.12586629e+00 4.71887022e-01
5.60584486e-01 -6.55162334e-01 -3.65910769e-01 2.94491462e-02
3.34076256e-01 -2.56910354e-01 -5.80704033e-01 9.80149582e-02
5.25519252e-01 -1.18211412e+00 1.03759801e+00 -3.67568195e-01
7.64498949e-01 -4.09789264e-01 -3.90459895e-01 -1.51344490e+00
-3.45908582e-01 -4.42524940e-01 2.27344587e-01 1.01448119e+00
3.25003147e-01 -3.77375633e-01 7.72271037e-01 4.09506202e-01
3.06442101e-02 -4.28960472e-01 -9.09545124e-01 -5.52620947e-01
-3.86206321e-02 -4.83357102e-01 2.59945631e-01 6.35627151e-01
-6.91139460e-01 2.26517506e-02 -5.02677858e-01 4.87283111e-01
1.06259978e+00 1.04667053e-01 7.98685551e-01 -1.00150824e+00
-3.52386355e-01 -3.47111881e-01 -2.61750281e-01 -1.00988996e+00
1.01579269e-02 -4.98328954e-01 6.03455484e-01 -1.47128797e+00
3.73651795e-02 -2.77504563e-01 -4.81591254e-01 4.92039114e-01
-2.39745602e-01 8.48088324e-01 6.43192828e-02 3.37858200e-01
-3.45393807e-01 4.49548095e-01 8.90385926e-01 -4.20547873e-01
-2.20619813e-02 -6.07220642e-02 -6.97350800e-01 8.40794623e-01
7.39623249e-01 -5.65361559e-01 1.61461398e-01 -8.22880208e-01
1.60970725e-02 -2.18247145e-01 7.88327754e-01 -1.06829369e+00
1.76545084e-01 1.11330308e-01 8.68371665e-01 -2.62341768e-01
2.94201642e-01 -9.27367806e-01 9.14974287e-02 4.06397015e-01
-3.33817244e-01 -2.08151579e-01 1.56789288e-01 7.95518935e-01
-5.69844365e-01 -2.74178982e-01 1.00441790e+00 -2.98306614e-01
-7.26339221e-01 1.88154671e-02 -5.39044023e-01 -2.62133956e-01
2.72798181e-01 -3.31994921e-01 -1.80304185e-01 -6.91742718e-01
-8.77917230e-01 -1.46490812e-01 5.04992008e-01 3.66985612e-02
7.84349620e-01 -1.06735682e+00 -7.64491022e-01 4.90881383e-01
-2.62217849e-01 4.36624959e-02 3.50542843e-01 5.40977716e-01
-5.56296110e-01 -4.55317050e-01 -6.91009685e-02 -7.65994549e-01
-1.12593901e+00 3.72704357e-01 5.89624882e-01 -1.13117032e-01
-7.22055852e-01 8.42185557e-01 1.85857847e-01 -2.83853710e-01
1.76527143e-01 -3.98003370e-01 2.43408129e-01 -2.50980973e-01
6.54881179e-01 3.33989002e-02 3.12692195e-01 -6.19316041e-01
-9.35466774e-03 6.99737430e-01 -4.08482552e-03 -1.49073645e-01
1.75249124e+00 -2.10826367e-01 -6.30020201e-01 2.98832536e-01
1.43070948e+00 -3.22995819e-02 -1.62555301e+00 -5.63306928e-01
-1.97355837e-01 -7.21776128e-01 4.23705161e-01 -8.65455925e-01
-1.56511450e+00 7.87681162e-01 9.32053804e-01 3.03790957e-01
1.68581605e+00 -1.16128385e-01 7.00352252e-01 5.95779359e-01
-2.81067435e-02 -8.65914822e-01 1.97530568e-01 4.35765862e-01
9.02472973e-01 -1.49031579e+00 -4.59468625e-02 -7.65561312e-02
-1.65710285e-01 1.33106148e+00 3.02439213e-01 -6.35170162e-01
7.46698380e-01 3.98730993e-01 5.50584853e-01 -2.50311762e-01
-5.06772041e-01 -2.74274826e-01 1.76832348e-01 7.54703045e-01
2.11288452e-01 -2.91077673e-01 1.63894877e-01 3.50249201e-01
-2.71731652e-02 -1.67313844e-01 4.43416893e-01 8.43420565e-01
-5.89194417e-01 -1.03658748e+00 -7.66533494e-01 5.18572867e-01
-6.85040534e-01 -3.03006738e-01 -2.68072218e-01 2.51284301e-01
3.11926872e-01 1.15128875e+00 5.48582636e-02 -3.03784162e-01
3.56069922e-01 -2.56982744e-01 3.40623617e-01 -3.71576160e-01
-6.56595647e-01 3.34526211e-01 -3.61948222e-01 -6.55875623e-01
-4.60833520e-01 -4.34769064e-01 -6.87061250e-01 -2.30569109e-01
-2.68836338e-02 -9.08977911e-03 7.47469366e-01 8.82245779e-01
5.18385842e-02 3.77492607e-01 7.93336868e-01 -1.32995391e+00
-5.05968392e-01 -8.79685938e-01 -7.36518919e-01 8.36612821e-01
9.41313565e-01 4.35262583e-02 -6.28851652e-01 5.14018476e-01] | [11.50780200958252, -2.2942845821380615] |
bff8dda1-c963-4b51-8211-e9ac4ca42da2 | identity-sensitive-knowledge-propagation-for | 2208.12023 | null | https://arxiv.org/abs/2208.12023v1 | https://arxiv.org/pdf/2208.12023v1.pdf | Identity-Sensitive Knowledge Propagation for Cloth-Changing Person Re-identification | Cloth-changing person re-identification (CC-ReID), which aims to match person identities under clothing changes, is a new rising research topic in recent years. However, typical biometrics-based CC-ReID methods often require cumbersome pose or body part estimators to learn cloth-irrelevant features from human biometric traits, which comes with high computational costs. Besides, the performance is significantly limited due to the resolution degradation of surveillance images. To address the above limitations, we propose an effective Identity-Sensitive Knowledge Propagation framework (DeSKPro) for CC-ReID. Specifically, a Cloth-irrelevant Spatial Attention module is introduced to eliminate the distraction of clothing appearance by acquiring knowledge from the human parsing module. To mitigate the resolution degradation issue and mine identity-sensitive cues from human faces, we propose to restore the missing facial details using prior facial knowledge, which is then propagated to a smaller network. After training, the extra computations for human parsing or face restoration are no longer required. Extensive experiments show that our framework outperforms state-of-the-art methods by a large margin. Our code is available at https://github.com/KimbingNg/DeskPro. | ['Jingwen Guo', 'Hao Tang', 'Wei Shi', 'Hong Liu', 'Jianbing Wu'] | 2022-08-25 | null | null | null | null | ['human-parsing'] | ['computer-vision'] | [ 1.46621495e-01 -2.42489323e-01 1.19213142e-01 -6.77735388e-01
-5.13099849e-01 -3.84938717e-01 1.43858373e-01 -4.23446476e-01
-4.46436912e-01 6.05298281e-01 1.17392272e-01 2.65845358e-01
-1.91003655e-03 -4.43343371e-01 -6.76424146e-01 -7.89383888e-01
5.51389754e-01 -8.66035074e-02 -1.46680534e-01 -6.00238629e-02
-3.01987082e-02 5.53551853e-01 -1.47937405e+00 -1.04650356e-01
8.43212605e-01 1.09545875e+00 -1.12277173e-01 2.14683354e-01
1.10806398e-01 2.36017182e-01 -3.06661963e-01 -1.07080948e+00
5.38758576e-01 -2.50711292e-01 -3.49723965e-01 1.97440669e-01
9.36659336e-01 -6.86119854e-01 -4.55717653e-01 1.39502656e+00
7.65234590e-01 -6.59856200e-02 9.93081853e-02 -1.14822328e+00
-8.86231840e-01 1.92149624e-01 -1.02439952e+00 -1.72734842e-01
3.36637557e-01 1.20444790e-01 2.76138633e-01 -1.04207170e+00
3.91533792e-01 1.43540215e+00 1.01924098e+00 1.04362869e+00
-1.04890418e+00 -1.03119743e+00 3.15947354e-01 3.89378399e-01
-1.81088710e+00 -9.43141282e-01 1.06059349e+00 -5.88463955e-02
6.65131509e-02 3.78384858e-01 5.14693022e-01 1.09114671e+00
-2.69964576e-01 6.61385834e-01 1.06380272e+00 -2.14173988e-01
-2.81972617e-01 1.57505184e-01 3.71034518e-02 8.83931518e-01
4.93794560e-01 -5.47203459e-02 -4.11281854e-01 3.69502865e-02
9.99874532e-01 3.24023247e-01 -4.07992721e-01 -1.89946845e-01
-8.90557110e-01 2.54858851e-01 4.44435239e-01 -1.18930407e-01
-3.01235735e-01 -9.99938920e-02 3.21041197e-01 5.65666240e-03
5.08574128e-01 -1.60998166e-01 -3.39422077e-01 3.11910391e-01
-6.95023537e-01 1.53259978e-01 4.19279158e-01 1.18915880e+00
6.67868495e-01 -1.40591994e-01 -1.35454118e-01 9.68892455e-01
3.03650141e-01 7.67053366e-01 3.02008279e-02 -8.17124784e-01
3.80138934e-01 5.16591251e-01 1.68674991e-01 -1.22609627e+00
-3.31843644e-01 -3.60171229e-01 -1.14860022e+00 -1.69436961e-01
6.46180928e-01 -2.40634918e-01 -7.93722689e-01 1.85461152e+00
7.50912130e-01 1.45915717e-01 -2.87616640e-01 1.29845452e+00
8.65585327e-01 2.57799387e-01 2.22126320e-01 -3.47853065e-01
1.57194257e+00 -7.51891971e-01 -7.03197837e-01 -2.21358404e-01
-9.99193639e-02 -8.08478236e-01 7.17387974e-01 1.05894081e-01
-9.97612715e-01 -7.78108478e-01 -7.52818704e-01 -1.71865731e-01
-1.74912095e-01 5.11739254e-01 5.71299314e-01 9.37560201e-01
-9.61656332e-01 3.24602813e-01 -5.59836626e-01 -5.26704252e-01
5.83628297e-01 5.98208368e-01 -4.44562405e-01 -3.96773338e-01
-9.57150459e-01 5.75380623e-01 -1.59548134e-01 7.86982894e-01
-3.14121574e-01 -5.50371468e-01 -1.02517331e+00 -2.42485523e-01
5.07914901e-01 -7.35064685e-01 8.82077694e-01 -1.15718257e+00
-1.55436730e+00 9.17299509e-01 -4.38872874e-01 2.26031691e-01
7.80373454e-01 -2.09969759e-01 -5.39499998e-01 -4.08018343e-02
8.92596319e-03 5.17903328e-01 1.16239798e+00 -1.34469199e+00
-2.68964082e-01 -8.86174142e-01 -1.79144695e-01 2.67011136e-01
-4.39700484e-01 3.26100022e-01 -1.02971673e+00 -7.65142739e-01
1.56058177e-01 -9.95855033e-01 9.58057344e-02 3.50387067e-01
-4.62090701e-01 -4.95267622e-02 5.01970649e-01 -1.26591420e+00
8.36039841e-01 -2.12466192e+00 -1.64809469e-02 1.38160333e-01
1.77107230e-01 2.46461391e-01 -2.01096267e-01 -2.20597789e-01
-1.95534322e-02 -1.98641196e-01 -1.59732118e-01 -7.04362690e-01
-1.30415127e-01 -1.33382365e-01 1.17412440e-01 7.76751757e-01
1.48280822e-02 8.03678811e-01 -5.36431551e-01 -6.58880711e-01
1.78137213e-01 7.05165625e-01 -1.54483244e-01 7.25735873e-02
3.17528069e-01 6.86135173e-01 -3.82762343e-01 1.05986309e+00
1.43151534e+00 1.13358116e-02 5.79494722e-02 -6.05220556e-01
-1.67623609e-02 -4.53622997e-01 -1.14586616e+00 1.79107904e+00
-1.18018754e-01 9.72092822e-02 5.80655217e-01 -7.01144040e-01
8.07462454e-01 1.05297357e-01 3.78121376e-01 -6.96564615e-01
4.31618392e-01 -2.26566829e-02 -3.14589024e-01 -4.41863567e-01
2.32759029e-01 5.23465015e-02 7.69692436e-02 -3.05028930e-02
-2.86964744e-01 5.94975770e-01 -2.08574086e-01 -2.79888183e-01
5.55086374e-01 3.39635640e-01 7.95131898e-04 -1.36704788e-01
8.73543620e-01 -3.67305636e-01 1.17032790e+00 3.10216695e-01
-5.63259244e-01 6.30321681e-01 -1.04195058e-01 -5.97595096e-01
-8.88173759e-01 -9.79267776e-01 -1.47150621e-01 9.35256541e-01
5.68704128e-01 -7.35479444e-02 -1.14580166e+00 -6.31407559e-01
1.56552330e-01 -2.36152913e-02 -6.19505048e-01 -1.11602852e-03
-6.88382745e-01 -8.48232090e-01 5.24739921e-01 4.97433871e-01
8.29924643e-01 -6.47549391e-01 1.68228582e-01 -1.37224153e-01
-2.67906547e-01 -1.17363048e+00 -1.00129962e+00 -9.51645494e-01
-6.15786374e-01 -1.03640497e+00 -1.29597747e+00 -8.88113737e-01
1.26025271e+00 5.32625616e-01 5.30083656e-01 1.81328923e-01
-5.21846771e-01 3.68326098e-01 -5.68981394e-02 -3.34026575e-01
1.65468633e-01 -1.56431884e-01 4.49981123e-01 5.85487604e-01
4.92148519e-01 -4.35824722e-01 -1.01796651e+00 5.37528336e-01
-4.91625547e-01 1.79812923e-01 6.99517548e-01 7.99471140e-01
5.91519177e-01 -1.93866864e-02 5.75674713e-01 -5.93805552e-01
3.73844892e-01 -3.98522876e-02 -7.01490939e-01 4.67258781e-01
-3.85711253e-01 -3.87775391e-01 6.27443671e-01 -5.72341502e-01
-1.39683378e+00 3.27483565e-01 -3.84775526e-03 -5.83677888e-01
-2.68508971e-01 -1.44511610e-01 -8.14683855e-01 -4.76922065e-01
1.61568716e-01 3.57269675e-01 7.80703202e-02 -6.97847903e-01
1.27355874e-01 6.67397797e-01 8.50639641e-01 -5.63928246e-01
1.04825366e+00 4.95103508e-01 -5.88782802e-02 -7.16443837e-01
-7.80167103e-01 -2.60221064e-01 -6.40847981e-01 -3.29424143e-01
8.01080167e-01 -1.08487010e+00 -1.14963591e+00 9.68084574e-01
-1.29111445e+00 2.43367016e-01 2.24565536e-01 3.05618614e-01
4.45673093e-02 6.89918876e-01 -7.52666831e-01 -1.00121450e+00
-6.44714832e-01 -8.44218075e-01 9.65620935e-01 8.37340057e-01
2.90364504e-01 -3.98512065e-01 -4.48566735e-01 7.45842159e-01
3.34502816e-01 1.73180252e-01 4.26877767e-01 -4.54730913e-02
-6.72107518e-01 -3.31023723e-01 -7.04730153e-01 2.94993460e-01
4.32818830e-01 -4.02203262e-01 -1.11688495e+00 -4.54687148e-01
-1.02844551e-01 1.04692943e-01 5.58690488e-01 3.29142898e-01
1.40580225e+00 -4.42517370e-01 -2.44886965e-01 9.38265443e-01
1.17088127e+00 3.03474348e-03 4.37038600e-01 7.90163800e-02
1.16799152e+00 8.30294430e-01 5.04096270e-01 3.97257954e-01
8.33126068e-01 6.48048759e-01 1.23432837e-01 -3.01532924e-01
-2.53230006e-01 -2.94972807e-01 1.76034451e-01 7.34048367e-01
-3.60230565e-01 1.50148422e-01 -4.70063865e-01 3.99037033e-01
-1.82598066e+00 -6.95590079e-01 9.97176096e-02 2.33903480e+00
8.64620149e-01 -4.00187135e-01 1.71070784e-01 -2.19916373e-01
1.24988854e+00 -5.14999554e-02 -9.64537323e-01 1.86460316e-01
-1.32001325e-01 -1.03381671e-01 6.47340178e-01 2.20093086e-01
-1.17939377e+00 8.96163285e-01 4.73005438e+00 6.57862902e-01
-8.47313523e-01 1.82181820e-01 6.37563169e-01 -3.02142147e-02
1.20422579e-01 -2.98418283e-01 -8.70961547e-01 6.63888752e-01
2.73385972e-01 5.18392473e-02 6.74058020e-01 7.63803005e-01
1.13275833e-01 1.19503357e-01 -8.32141876e-01 1.54987931e+00
4.77681935e-01 -7.79761970e-01 -1.83472425e-01 -3.34592573e-02
4.77983326e-01 -6.71387374e-01 -9.13799182e-03 3.52353882e-03
-1.76478013e-01 -7.44308650e-01 3.98025960e-01 9.51240957e-01
1.08343863e+00 -8.78650844e-01 8.33045065e-01 -3.69608440e-02
-1.38737953e+00 5.79467788e-02 -5.46481431e-01 1.91252738e-01
1.18951108e-02 4.21856374e-01 -2.37865567e-01 6.87633872e-01
8.61454546e-01 4.81460810e-01 -5.82328439e-01 9.50817108e-01
-1.15701005e-01 1.99006230e-01 -2.26815864e-01 2.68314600e-01
-5.67991316e-01 -9.22772288e-02 3.84143502e-01 1.02985442e+00
2.84302801e-01 3.35885316e-01 1.56105474e-01 6.99271381e-01
-4.14168179e-01 5.25260642e-02 -1.22430831e-01 4.50399995e-01
4.99864936e-01 1.36669946e+00 -3.34614366e-01 -8.21047872e-02
-4.27019268e-01 1.58408546e+00 3.58146131e-02 3.65062505e-01
-8.95637393e-01 -3.71185988e-01 8.83199155e-01 1.72476679e-01
2.54611783e-02 -8.19608197e-03 -1.97680190e-01 -1.37979472e+00
4.83710617e-01 -7.40253806e-01 2.20076978e-01 -6.57657743e-01
-1.59221184e+00 3.39194715e-01 -2.65692294e-01 -1.10611022e+00
3.41897041e-01 -4.10932422e-01 -3.84594679e-01 1.10725617e+00
-1.58936656e+00 -1.66532254e+00 -8.66558671e-01 8.36753368e-01
3.43811512e-01 1.17985625e-02 6.58444583e-01 6.77709162e-01
-1.13910711e+00 1.38535213e+00 -1.90641209e-01 5.59102833e-01
1.04918265e+00 -7.25162625e-01 4.78447407e-01 1.01488006e+00
-6.00056231e-01 9.59994912e-01 2.93210417e-01 -7.97412932e-01
-1.75007737e+00 -1.15701735e+00 6.54280424e-01 -2.39940852e-01
1.03414878e-01 -4.69586194e-01 -8.86223614e-01 4.35476720e-01
-1.61200717e-01 2.93073744e-01 6.18226826e-01 -7.46692345e-02
-5.31041324e-01 -7.47133374e-01 -1.37572432e+00 6.58220828e-01
1.48411512e+00 -4.54653412e-01 -2.10651457e-01 3.63194570e-02
4.38377172e-01 -4.45274174e-01 -1.00701535e+00 4.84348744e-01
9.73159850e-01 -6.96545720e-01 1.03325188e+00 -1.50946915e-01
-1.42969355e-01 -5.20065725e-01 8.81025493e-02 -8.07244837e-01
-3.19101602e-01 -7.34549344e-01 -2.55715800e-03 1.71139407e+00
-8.69445279e-02 -7.94926941e-01 9.10015643e-01 1.18733287e+00
3.86056036e-01 -4.09526199e-01 -8.84266376e-01 -6.92021787e-01
-4.20423985e-01 1.70523390e-01 9.15229619e-01 1.03509104e+00
-4.39240187e-01 5.09597845e-02 -8.43699992e-01 5.79981565e-01
1.23440194e+00 1.30702808e-01 9.17850494e-01 -1.22646618e+00
5.55115603e-02 -6.81073219e-02 -3.26899171e-01 -9.20160711e-01
1.64878935e-01 -4.40213680e-01 6.30388483e-02 -1.19473326e+00
4.81830567e-01 -4.23598498e-01 -3.62030923e-01 5.07931292e-01
-4.00938988e-01 4.61694390e-01 2.67223954e-01 2.61523634e-01
-5.09798944e-01 5.65698683e-01 1.38435805e+00 -3.54855768e-02
-5.21870591e-02 4.68835011e-02 -9.31102097e-01 8.14535677e-01
7.60134876e-01 -2.27629155e-01 -1.11618675e-01 -5.81163466e-01
-1.99084073e-01 -4.61975075e-02 6.61071658e-01 -8.89566004e-01
5.31982005e-01 -7.49363154e-02 1.15340924e+00 -3.63690704e-01
4.99742121e-01 -8.07060719e-01 1.69398770e-01 1.85097620e-01
8.15228969e-02 3.65263037e-02 2.65682548e-01 4.90649462e-01
1.17745772e-01 2.02484637e-01 8.29212248e-01 -6.98020235e-02
-5.98887742e-01 7.76063263e-01 1.75507098e-01 -4.46371228e-01
7.45114326e-01 -3.57393831e-01 -3.65521699e-01 -2.05705538e-01
-4.74965364e-01 2.10115507e-01 7.06723690e-01 5.09327531e-01
6.42982066e-01 -1.31699860e+00 -7.53767848e-01 2.86139131e-01
7.45983198e-02 9.60536860e-03 8.70987773e-01 8.13863277e-01
-2.08340064e-01 7.00014969e-03 -2.65414476e-01 -2.87333012e-01
-1.59781504e+00 5.94355404e-01 3.18262398e-01 3.26418161e-01
-4.82183188e-01 8.43728781e-01 2.59803444e-01 -4.52590644e-01
2.68047541e-01 1.97763115e-01 -1.32110203e-03 -1.55743808e-01
7.53556013e-01 5.27627110e-01 -1.63707495e-01 -9.17283893e-01
-6.58533335e-01 1.04267919e+00 -2.86853522e-01 2.99063474e-01
1.05921113e+00 -5.09359777e-01 -2.21719921e-01 -2.99973637e-01
9.55303848e-01 1.23734199e-01 -1.46045721e+00 -3.92774284e-01
-3.68841946e-01 -8.33011746e-01 -1.28957391e-01 -7.83734202e-01
-1.36757731e+00 5.78484476e-01 9.69800413e-01 -4.70016360e-01
1.51890063e+00 -2.87326396e-01 1.02543187e+00 1.61852986e-01
5.13281465e-01 -1.08044183e+00 -4.79728490e-01 -1.78940203e-02
8.89604986e-01 -1.42503059e+00 2.33390197e-01 -7.27781296e-01
-4.68728662e-01 9.24368680e-01 8.28700006e-01 9.40124243e-02
4.99005258e-01 2.13153493e-02 1.38103992e-01 2.37972960e-01
2.08021238e-01 -1.03505880e-01 3.53874892e-01 6.84300840e-01
1.35160357e-01 7.15571940e-02 -3.27596098e-01 1.01709747e+00
-2.41779778e-02 -5.79648241e-02 6.25032140e-03 6.67346656e-01
8.18610191e-02 -1.09948456e+00 -6.75435722e-01 1.86947271e-01
-4.82704222e-01 3.11583914e-02 -4.12146032e-01 5.63515604e-01
4.20144945e-01 9.07699883e-01 -2.32698098e-01 -4.08546001e-01
3.39217126e-01 -7.89424181e-02 7.18664765e-01 -1.19937763e-01
-3.20559621e-01 1.61653563e-01 -6.04625493e-02 -6.67563260e-01
-4.97715682e-01 -7.57799566e-01 -8.06180954e-01 -7.84782588e-01
-2.64304250e-01 -3.04312795e-01 6.37232721e-01 7.07691848e-01
5.95761478e-01 1.50226742e-01 6.21861577e-01 -9.57636952e-01
-3.62122953e-01 -8.07505190e-01 -5.04638195e-01 4.42651898e-01
3.21478963e-01 -5.96029460e-01 -7.98705779e-03 1.54916793e-01] | [14.697482109069824, 0.9426100254058838] |
57a746a2-689e-48d3-9084-bf93a866b4e0 | rethinking-crowdsourcing-annotation-partial | 2109.02688 | null | https://arxiv.org/abs/2109.02688v1 | https://arxiv.org/pdf/2109.02688v1.pdf | Rethinking Crowdsourcing Annotation: Partial Annotation with Salient Labels for Multi-Label Image Classification | Annotated images are required for both supervised model training and evaluation in image classification. Manually annotating images is arduous and expensive, especially for multi-labeled images. A recent trend for conducting such laboursome annotation tasks is through crowdsourcing, where images are annotated by volunteers or paid workers online (e.g., workers of Amazon Mechanical Turk) from scratch. However, the quality of crowdsourcing image annotations cannot be guaranteed, and incompleteness and incorrectness are two major concerns for crowdsourcing annotations. To address such concerns, we have a rethinking of crowdsourcing annotations: Our simple hypothesis is that if the annotators only partially annotate multi-label images with salient labels they are confident in, there will be fewer annotation errors and annotators will spend less time on uncertain labels. As a pleasant surprise, with the same annotation budget, we show a multi-label image classifier supervised by images with salient annotations can outperform models supervised by fully annotated images. Our method contributions are 2-fold: An active learning way is proposed to acquire salient labels for multi-label images; and a novel Adaptive Temperature Associated Model (ATAM) specifically using partial annotations is proposed for multi-label image classification. We conduct experiments on practical crowdsourcing data, the Open Street Map (OSM) dataset and benchmark dataset COCO 2014. When compared with state-of-the-art classification methods trained on fully annotated images, the proposed ATAM can achieve higher accuracy. The proposed idea is promising for crowdsourcing data annotation. Our code will be publicly available. | ['Z. Jane Wang', 'Tianze Yu', 'Jianzhe Lin'] | 2021-09-06 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 2.37816721e-01 3.99659723e-01 -2.76804745e-01 -5.02903759e-01
-9.12622690e-01 -9.01223838e-01 1.48452967e-01 4.38625693e-01
-7.94630349e-01 7.50293970e-01 -4.64243323e-01 1.44174233e-01
4.15222734e-01 -4.09956485e-01 -8.90443802e-01 -8.17001641e-01
2.56020069e-01 6.79668367e-01 5.95257342e-01 -1.39879003e-01
2.89838687e-02 2.36189738e-01 -1.57142961e+00 1.93593279e-01
8.66948366e-01 1.05536401e+00 1.93742946e-01 5.47472894e-01
-7.35285133e-02 1.20660841e+00 -4.75690454e-01 -7.50056446e-01
5.36494434e-01 9.68504846e-02 -1.21872461e+00 6.00770772e-01
2.95164824e-01 -1.88046917e-01 2.91284621e-01 1.04980731e+00
3.43898237e-01 8.14264789e-02 3.42323124e-01 -1.73113322e+00
-7.11089015e-01 2.59911120e-01 -6.66846991e-01 -3.42502654e-01
1.49659067e-01 7.82469288e-02 6.87438607e-01 -1.00517058e+00
3.79967988e-01 8.16689372e-01 8.54619086e-01 6.69201374e-01
-1.10319459e+00 -3.59964460e-01 2.19508961e-01 1.05246872e-01
-1.65243816e+00 -2.01107025e-01 4.50016201e-01 -7.42137909e-01
3.35272819e-01 4.43549395e-01 3.97698224e-01 7.58564591e-01
-1.84747845e-01 8.94369662e-01 1.48376274e+00 -6.80457652e-01
5.19247353e-01 7.80437469e-01 -1.54762372e-01 6.69704795e-01
-2.82125808e-02 -3.34225804e-01 -6.44860983e-01 -4.95892704e-01
1.56845793e-01 1.79977521e-01 -8.61824825e-02 -2.81552404e-01
-1.06922424e+00 7.83358097e-01 3.67159396e-01 -7.77937472e-03
-1.71642855e-01 2.36272603e-01 4.81114030e-01 -1.38592422e-01
8.57772708e-01 2.30051339e-01 -5.59795916e-01 2.17357039e-01
-9.01398659e-01 4.84196059e-02 6.15341783e-01 1.30179322e+00
1.26215267e+00 -4.15693641e-01 -1.70634851e-01 9.03603792e-01
3.03379864e-01 6.67383790e-01 4.12223995e-01 -8.75279248e-01
1.30558893e-01 8.82633984e-01 6.33598089e-01 -8.13949287e-01
-2.56795526e-01 4.58756745e-01 -6.09054208e-01 2.13257179e-01
4.69454199e-01 -1.23966411e-01 -7.00765729e-01 1.05763125e+00
5.34775078e-01 -3.34630348e-02 -2.38314316e-01 1.12959361e+00
8.41814816e-01 2.26437569e-01 5.63337862e-01 -6.48872107e-02
1.45795047e+00 -1.46067810e+00 -7.31045008e-01 -2.90578961e-01
7.56952226e-01 -8.37631524e-01 9.91494060e-01 2.45204732e-01
-7.39153028e-01 -6.84733093e-01 -5.61417580e-01 -6.25348464e-02
-7.31690109e-01 5.52764893e-01 5.12152433e-01 9.15239453e-01
-9.56383109e-01 4.07322109e-01 -5.14364481e-01 -3.09972137e-01
5.50008714e-01 4.74078238e-01 -3.86230409e-01 6.88655823e-02
-1.10791993e+00 7.84629226e-01 3.99593890e-01 2.14355752e-01
-1.00158775e+00 -6.92126304e-02 -8.66440296e-01 -5.05890727e-01
5.24097741e-01 1.33843303e-01 1.32436597e+00 -1.51787674e+00
-1.00255966e+00 1.44093740e+00 -2.63785571e-01 -4.32038903e-01
7.78122067e-01 -3.21109295e-02 -1.93790585e-01 2.98362911e-01
4.16937917e-01 1.14352417e+00 9.06296432e-01 -1.72373569e+00
-6.89279437e-01 -1.80768937e-01 2.22068399e-01 8.56244266e-02
-3.72402132e-01 2.54919350e-01 -2.82808274e-01 -4.28466380e-01
-2.03079417e-01 -1.46996045e+00 -4.84821349e-01 2.55756557e-01
-3.42506021e-01 -2.66228586e-01 7.14322388e-01 -4.45166826e-01
6.38494909e-01 -2.02370667e+00 -4.46668506e-01 -9.08090025e-02
2.63101906e-01 4.15117413e-01 1.05542071e-01 2.02801615e-01
2.30435684e-01 3.82463813e-01 -3.55237335e-01 -6.73333645e-01
-2.56734528e-02 4.29664284e-01 -4.14087266e-01 6.13838553e-01
4.95112300e-01 1.15800655e+00 -1.07710135e+00 -1.00506175e+00
1.34872466e-01 1.41361207e-01 2.47953832e-01 2.96367615e-01
-2.28056371e-01 8.70448470e-01 -1.26947209e-01 9.02179658e-01
8.32666576e-01 -4.39045548e-01 -2.93827001e-02 2.09289312e-01
-6.79629296e-02 -6.70049012e-01 -1.21402836e+00 1.29696214e+00
-3.11790794e-01 4.84423071e-01 -1.17788568e-01 -7.41598189e-01
1.01308572e+00 5.63501477e-01 5.64161301e-01 -4.73282814e-01
7.02402145e-02 5.00251889e-01 -8.44212651e-01 -9.25166070e-01
8.43805075e-01 1.85153916e-01 -1.28376171e-01 4.69526947e-01
-1.26110762e-01 -3.41245264e-01 1.41021192e-01 9.09996182e-02
6.46213651e-01 3.57802838e-01 1.64313629e-01 -2.51661271e-01
5.43824852e-01 4.98666257e-01 6.15482092e-01 8.22239697e-01
-6.94184124e-01 7.56252646e-01 1.67247489e-01 -9.78408456e-01
-1.29839087e+00 -3.55705023e-01 -1.13787659e-01 1.52512753e+00
5.95576227e-01 -4.92941849e-02 -9.13921595e-01 -9.87514257e-01
-2.05533490e-01 -7.71442652e-02 -8.90459716e-01 5.28232038e-01
-2.42672324e-01 -7.83289194e-01 6.97447360e-01 5.43989003e-01
8.55835140e-01 -9.35565114e-01 -6.27947509e-01 5.16894506e-04
-4.02020365e-01 -1.47160399e+00 -4.77573037e-01 3.30368340e-01
-2.60768235e-01 -1.11262965e+00 -9.69023108e-01 -9.28853333e-01
1.06038904e+00 4.56931263e-01 1.03000093e+00 5.07724524e-01
-1.56128824e-01 3.79278809e-01 -8.02200437e-01 -9.78495836e-01
-2.57086277e-01 -8.35008547e-02 -1.73701197e-02 4.15740788e-01
6.51308358e-01 1.34177998e-01 -6.56799912e-01 1.02745771e+00
-1.13245547e+00 1.42226117e-02 2.19063386e-01 6.77872777e-01
7.33619571e-01 -2.91875005e-01 6.68714583e-01 -9.89413261e-01
1.03872217e-01 -3.42634082e-01 -7.46227205e-01 5.71820319e-01
-5.15201092e-01 -2.91295767e-01 5.83944201e-01 -6.41680062e-01
-8.75077128e-01 1.08610082e+00 1.54099867e-01 -1.58181861e-01
-5.34471452e-01 -3.23541135e-01 3.04920495e-01 -7.96920657e-01
9.87917066e-01 -3.48801017e-02 -3.25810611e-01 5.78589588e-02
2.13340089e-01 1.03162253e+00 1.73342213e-01 -5.29510498e-01
6.35557413e-01 8.79270554e-01 -1.94262221e-01 -5.04342556e-01
-1.09008002e+00 -1.09515822e+00 -9.46509123e-01 -6.26178741e-01
1.20159149e+00 -1.32354259e+00 -8.28065693e-01 5.60260534e-01
-1.29832447e+00 -5.06952167e-01 -2.72668779e-01 2.09117621e-01
-4.05272156e-01 6.44303024e-01 -4.72964942e-01 -1.27946520e+00
-2.80172825e-01 -1.45466292e+00 1.78412461e+00 3.84265304e-01
-1.33588493e-01 -9.53691006e-01 -9.96091589e-02 9.48190093e-01
3.26051474e-01 3.54065537e-01 7.34297335e-02 -8.39806437e-01
-4.05611336e-01 -4.88360703e-01 -2.66216010e-01 4.61801201e-01
-3.23503315e-01 -1.29565090e-01 -1.66681325e+00 -8.47875103e-02
-1.81757808e-01 -1.16806674e+00 5.18913388e-01 -1.03734180e-01
1.24210536e+00 -1.72981068e-01 -2.52561301e-01 -2.38892183e-01
1.55871642e+00 -1.05820306e-01 4.93099004e-01 3.92391354e-01
6.65161133e-01 8.87677848e-01 1.04553795e+00 4.58397716e-01
6.34454787e-01 6.39060915e-01 6.29927218e-01 -4.69142556e-01
4.29994226e-01 1.73319995e-01 3.26797478e-02 4.18222189e-01
-3.12641948e-01 -1.90705881e-01 -1.14334357e+00 7.17245162e-01
-2.30493236e+00 -4.68276292e-01 -7.82696307e-01 2.27325249e+00
9.64704812e-01 -2.76803106e-01 2.26696417e-01 2.37781182e-01
1.12089777e+00 -2.30969489e-01 -3.20517629e-01 -1.40513927e-01
-2.05491796e-01 -1.86963111e-01 9.61151481e-01 4.16554868e-01
-1.56627858e+00 8.65400791e-01 5.51862669e+00 9.66103435e-01
-7.81902552e-01 7.02732444e-01 1.05879760e+00 3.68481606e-01
1.94834277e-01 8.26349407e-02 -8.26445460e-01 6.10259712e-01
5.25744677e-01 3.74736369e-01 2.44329318e-01 1.20197165e+00
1.06743015e-01 -6.86097324e-01 -8.36901188e-01 9.52844560e-01
2.29881912e-01 -1.01477790e+00 -4.01865512e-01 1.86495353e-02
1.27283394e+00 -8.03790316e-02 -4.28824425e-01 6.23170622e-02
3.59782517e-01 -9.22783911e-01 1.12209725e+00 3.35731894e-01
6.65324152e-01 -3.05998176e-01 1.21163654e+00 5.98754644e-01
-1.28042996e+00 -8.16406980e-02 -5.80155492e-01 -1.52736977e-01
7.29552284e-02 7.08234668e-01 -7.55718410e-01 4.34001327e-01
8.29603076e-01 1.13195002e-01 -1.03014171e+00 1.02170992e+00
-4.33662474e-01 4.82054979e-01 -6.53606548e-04 -2.35033333e-01
2.17993200e-01 1.38012215e-01 -1.50229767e-01 1.15528595e+00
-2.55687773e-01 2.27776263e-03 7.02415645e-01 4.42870915e-01
-1.80655614e-01 3.03403795e-01 -5.71918249e-01 2.99076736e-01
4.11098063e-01 1.68193614e+00 -1.20154202e+00 -4.97031838e-01
-3.46812695e-01 1.32854366e+00 3.73050898e-01 2.94076771e-01
-9.99177158e-01 -2.03128606e-01 -2.83961743e-01 -8.67787823e-02
-1.98732987e-01 9.68512669e-02 -3.78230631e-01 -1.01820254e+00
3.19103032e-01 -4.37382162e-01 -4.44838684e-03 -9.81363714e-01
-1.29810429e+00 6.95737898e-01 -2.70752013e-01 -1.37663031e+00
1.89103559e-01 -6.45224988e-01 -1.34937987e-01 7.06093252e-01
-1.61198890e+00 -1.53639555e+00 -9.17883039e-01 4.63697463e-01
5.85932672e-01 1.60082981e-01 9.84773636e-01 3.80411029e-01
-1.77441821e-01 3.63071561e-01 -1.07133761e-01 4.94657278e-01
8.11492741e-01 -1.32146800e+00 2.58666992e-01 5.58704615e-01
1.34094998e-01 -4.27554175e-02 4.59071070e-01 -5.88403463e-01
-8.38813424e-01 -1.48454678e+00 8.90978813e-01 -9.39268112e-01
5.25134563e-01 -6.07524455e-01 -7.81832933e-01 5.51025867e-01
1.45877570e-01 6.64123178e-01 7.91940391e-01 -6.54996514e-01
-1.43421322e-01 1.93308800e-01 -1.35159075e+00 8.47954750e-02
5.35663247e-01 -5.71595073e-01 -2.54688989e-02 1.21518505e+00
7.10743010e-01 -4.27121520e-01 -7.44649947e-01 2.80408740e-01
2.12250531e-01 -6.57440066e-01 5.26507616e-01 -2.70983636e-01
2.84694254e-01 -6.07772231e-01 -7.20488802e-02 -9.78511870e-01
2.36819282e-01 -3.81373793e-01 5.18104494e-01 1.37051380e+00
4.65783447e-01 -3.19262415e-01 7.98988938e-01 1.15494239e+00
2.00769097e-01 -6.63630188e-01 -8.60897601e-01 -8.99962127e-01
-1.20198853e-01 -3.50921214e-01 5.02820730e-01 1.20797753e+00
-1.11390144e-01 -6.10177219e-02 -5.65573096e-01 1.80169255e-01
5.60859203e-01 -2.88599461e-01 8.38727355e-01 -1.28740215e+00
1.39403194e-01 2.73300856e-01 -4.55467373e-01 -5.37132621e-01
2.76113391e-01 -4.89854455e-01 5.32781899e-01 -1.28719246e+00
3.59612703e-01 -8.08181465e-01 4.71637547e-01 8.96484971e-01
-3.46304804e-01 1.17876303e+00 1.84717238e-01 6.12303019e-01
-1.31392932e+00 2.40324318e-01 9.46458340e-01 -4.21380192e-01
1.48547143e-01 -1.40193909e-01 -1.60919234e-01 8.29768121e-01
8.53232503e-01 -8.19106221e-01 -5.88443018e-02 -3.70525360e-01
5.81647456e-01 -2.52581388e-01 6.07813478e-01 -8.83050144e-01
4.91087765e-01 -1.89455003e-01 1.72286287e-01 -1.33785009e-01
2.68226951e-01 -1.08466184e+00 3.24560441e-02 1.81077182e-01
-5.10458231e-01 -1.10951908e-01 -7.87387341e-02 7.05509305e-01
-2.25203797e-01 -8.00086558e-01 8.18342984e-01 -5.12296557e-01
-8.08571875e-01 2.48306662e-01 -4.29518312e-01 2.82061417e-02
1.45853293e+00 -1.89735815e-01 -4.82454687e-01 -1.25870064e-01
-7.05888987e-01 3.53914559e-01 7.50185609e-01 2.20584676e-01
2.42371380e-01 -1.43265617e+00 -6.59220278e-01 -3.55418712e-01
7.20195830e-01 3.28552961e-01 2.06489012e-01 8.27377617e-01
-6.87872529e-01 4.20076726e-03 1.88945591e-01 -8.28735650e-01
-1.27331161e+00 7.59961367e-01 1.87989295e-01 -1.08384401e-01
7.66622424e-02 7.72914886e-01 -2.42514431e-01 -8.32925677e-01
2.16757238e-01 4.84857075e-02 -1.35044768e-01 1.99543461e-01
6.23354852e-01 3.74256045e-01 3.56386811e-01 -1.00130451e+00
-3.75466555e-01 6.99776530e-01 2.72834629e-01 9.26104411e-02
8.48987341e-01 -3.31250817e-01 -3.49177331e-01 5.51534891e-01
8.56231511e-01 -4.67400774e-02 -1.19579566e+00 -2.73049593e-01
1.35882810e-01 -4.58701849e-01 -4.75083083e-01 -5.57774663e-01
-8.32131803e-01 7.34821796e-01 8.11199546e-01 5.69350183e-01
8.39053750e-01 -4.99234423e-02 4.96674538e-01 6.48070574e-01
9.32052791e-01 -1.63422728e+00 3.34507555e-01 1.65508062e-01
5.13328373e-01 -2.20454693e+00 -2.74489503e-02 -6.12333298e-01
-1.26721609e+00 9.33134556e-01 6.89576805e-01 1.27290815e-01
3.55116010e-01 7.46219531e-02 4.06231999e-01 -1.99284792e-01
-1.60354555e-01 -4.70601052e-01 5.60007989e-02 6.39383018e-01
2.18079403e-01 1.62191764e-01 -1.61191911e-01 3.62500787e-01
4.76190448e-01 9.02511925e-02 7.65048206e-01 1.04488325e+00
-4.53903139e-01 -8.52457106e-01 -8.31457853e-01 7.97415450e-02
-5.84495962e-01 3.20728831e-02 -6.65805399e-01 3.57026964e-01
6.73913419e-01 1.61966193e+00 -2.66499162e-01 -2.82559425e-01
3.14597227e-02 5.58325686e-02 -9.91418213e-02 -6.07428908e-01
-8.92504811e-01 -2.77787685e-01 -5.91080561e-02 -3.16995800e-01
-1.17513466e+00 -2.22419545e-01 -1.16684043e+00 -1.44927669e-02
-8.36366832e-01 3.51548821e-01 8.01113605e-01 1.03957880e+00
1.23288423e-01 -1.20065436e-01 8.24863434e-01 -1.10279477e+00
-9.75291729e-02 -9.33895826e-01 -5.78871310e-01 7.26465523e-01
5.26345074e-02 -5.40706754e-01 -4.10051793e-01 7.10793972e-01] | [9.599787712097168, 4.274991989135742] |
833f2cbf-f36a-4567-a099-300c5dda750d | exploring-the-limits-of-indiscriminate-data | 2303.03592 | null | https://arxiv.org/abs/2303.03592v3 | https://arxiv.org/pdf/2303.03592v3.pdf | Exploring the Limits of Model-Targeted Indiscriminate Data Poisoning Attacks | Indiscriminate data poisoning attacks aim to decrease a model's test accuracy by injecting a small amount of corrupted training data. Despite significant interest, existing attacks remain relatively ineffective against modern machine learning (ML) architectures. In this work, we introduce the notion of model poisoning reachability as a technical tool to explore the intrinsic limits of data poisoning attacks towards target parameters (i.e., model-targeted attacks). We derive an easily computable threshold to establish and quantify a surprising phase transition phenomenon among popular ML models: data poisoning attacks can achieve certain target parameters only when the poisoning ratio exceeds our threshold. Building on existing parameter corruption attacks and refining the Gradient Canceling attack, we perform extensive experiments to confirm our theoretical findings, test the predictability of our transition threshold, and significantly improve existing indiscriminate data poisoning baselines over a range of datasets and models. Our work highlights the critical role played by the poisoning ratio, and sheds new insights on existing empirical results, attacks and mitigation strategies in data poisoning. | ['Gautam Kamath', 'YaoLiang Yu', 'Yiwei Lu'] | 2023-03-07 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 1.46580279e-01 -2.30550706e-01 -5.13776302e-01 1.97522208e-01
-8.69601309e-01 -1.11961520e+00 7.72773683e-01 4.90617454e-01
-5.40845692e-01 5.53455055e-01 -5.34891598e-02 -7.32169986e-01
3.31009924e-02 -5.75535476e-01 -9.03317869e-01 -9.45713341e-01
-3.05861861e-01 3.39630127e-01 2.19934642e-01 -7.84292072e-02
4.60167140e-01 8.36699843e-01 -1.07405639e+00 1.97294995e-01
5.39127946e-01 5.37101150e-01 -7.56401002e-01 7.41297007e-01
3.43971461e-01 8.96253109e-01 -1.05780458e+00 -5.82988203e-01
5.09018302e-01 -2.80696660e-01 -7.98750997e-01 -5.21966554e-02
5.64223349e-01 -5.18017054e-01 -5.41828811e-01 1.28226376e+00
5.08417845e-01 -5.94989896e-01 4.85743254e-01 -1.90665340e+00
-2.14750245e-01 9.67180490e-01 -5.20858526e-01 5.25447369e-01
-1.63910002e-01 7.82552719e-01 7.56481647e-01 -4.30353522e-01
3.22507054e-01 1.12955797e+00 7.27858663e-01 1.01206362e+00
-1.42687273e+00 -1.03039551e+00 9.29344147e-02 -2.08487168e-01
-1.30509722e+00 -4.34880733e-01 6.72622323e-01 -3.15073669e-01
7.14363337e-01 6.29355252e-01 2.05729604e-01 1.61346912e+00
1.09281138e-01 7.74856210e-01 1.21336102e+00 -1.48729458e-01
4.24490035e-01 9.91138071e-02 4.58562344e-01 6.05284512e-01
9.29716706e-01 3.06718081e-01 -6.64889157e-01 -9.98854458e-01
5.20855248e-01 -2.28725672e-01 -3.17261368e-01 -4.17998910e-01
-7.70272851e-01 9.97845590e-01 1.47879660e-01 2.64938846e-02
9.42212939e-02 6.06830716e-01 5.02774715e-01 4.21089590e-01
2.43591473e-01 7.03886449e-01 -7.93283463e-01 2.76603475e-02
-5.61131418e-01 4.31947351e-01 8.41737032e-01 5.59911013e-01
2.57276922e-01 2.26650015e-01 -7.94817209e-02 1.67606413e-01
2.33237103e-01 6.60217583e-01 2.16354743e-01 -6.07645631e-01
6.70153975e-01 5.14213502e-01 1.88535929e-01 -5.93812585e-01
-3.11631352e-01 -6.23224497e-01 -6.94614708e-01 3.85285132e-02
8.52317333e-01 -2.57568866e-01 -5.76374292e-01 2.01624250e+00
5.29159486e-01 2.64718413e-01 5.87936863e-02 4.05188888e-01
1.53593227e-01 1.45560518e-01 5.31174719e-01 -2.31796667e-01
1.10952461e+00 -3.13879579e-01 -2.15750232e-01 8.58864561e-02
1.26854336e+00 -1.99025542e-01 1.27879763e+00 5.75971127e-01
-9.10762072e-01 1.30102485e-01 -1.09988022e+00 3.37861061e-01
-4.02025431e-01 -6.47306800e-01 6.11088097e-01 1.23561382e+00
-6.39804721e-01 5.70145369e-01 -1.16854393e+00 -1.48244530e-01
8.34917963e-01 6.29954040e-01 -7.52917230e-02 2.29762822e-01
-1.09943807e+00 5.92679441e-01 5.98533787e-02 -2.53334314e-01
-1.49576688e+00 -1.14199483e+00 -3.36493045e-01 -3.07413042e-01
2.75462300e-01 -6.11828327e-01 1.12144661e+00 -1.56083539e-01
-7.84617901e-01 7.28248715e-01 2.66485155e-01 -1.10484874e+00
9.09146369e-01 -3.63059282e-01 -1.17134616e-01 1.46097243e-01
-4.48085874e-01 3.32657814e-01 7.26041913e-01 -1.62913501e+00
-4.04126048e-01 -3.72849703e-01 2.28443876e-01 -1.94384053e-01
-9.48271811e-01 2.71444976e-01 1.06390931e-01 -4.88941103e-01
-3.55376571e-01 -7.62896299e-01 -5.38749576e-01 -1.84823647e-01
-9.76653934e-01 -7.25113451e-02 1.02001536e+00 -7.89861381e-02
1.51177895e+00 -1.98556030e+00 -1.44046798e-01 2.84692794e-01
6.88502610e-01 4.84271824e-01 -6.86924383e-02 6.14355981e-01
3.49317640e-02 7.39181101e-01 -5.68788230e-01 -3.00593317e-01
2.57670451e-02 5.93962371e-02 -8.41858566e-01 9.35702741e-01
1.34842545e-01 8.43319356e-01 -6.93807185e-01 -1.93755701e-01
1.46091981e-02 2.66643405e-01 -8.94734919e-01 4.81792837e-02
-4.13048416e-01 3.11602235e-01 -5.89389443e-01 5.93531311e-01
5.91031492e-01 -1.88033447e-01 1.31464735e-01 -1.39989451e-01
1.37684852e-01 3.55672181e-01 -8.64909351e-01 7.52531350e-01
4.88772430e-02 1.66683227e-01 5.98268174e-02 -7.44294822e-01
5.59149146e-01 6.06339946e-02 5.31995356e-01 -2.58206040e-01
3.98366839e-01 2.39222571e-01 1.65030763e-01 -2.27558225e-01
7.97845945e-02 -1.41355187e-01 -3.67982209e-01 7.27079690e-01
-4.98528898e-01 1.54233605e-01 -1.69462830e-01 3.70179981e-01
1.75488877e+00 -5.07592142e-01 1.23224095e-01 -5.41876435e-01
3.12428445e-01 2.40591839e-01 1.77273408e-01 1.26725960e+00
-3.39891791e-01 2.50643373e-01 1.00377214e+00 -4.41689700e-01
-1.07244706e+00 -1.07514989e+00 -3.28063637e-01 7.93498576e-01
1.14364922e-01 -6.29565656e-01 -1.12438452e+00 -1.31522214e+00
3.03114802e-01 5.24151862e-01 -7.57486284e-01 -7.87405491e-01
-6.91577017e-01 -1.51273882e+00 1.48219907e+00 4.89637673e-01
6.22951388e-01 -6.12577200e-01 -5.48722923e-01 -1.37860969e-01
1.06163464e-01 -8.58625889e-01 -1.68161273e-01 4.47516829e-01
-9.80068028e-01 -1.45728028e+00 5.56987748e-02 -2.59429604e-01
6.35820806e-01 1.35159850e-01 1.06787884e+00 6.53641582e-01
-1.42056793e-01 2.34888986e-01 -3.51041965e-02 -5.20206928e-01
-9.44020927e-01 2.99860626e-01 3.22183222e-01 -3.45057815e-01
5.77504396e-01 -5.36449730e-01 -4.48640108e-01 2.47570068e-01
-1.36563122e+00 -4.57348406e-01 2.66698390e-01 4.32687908e-01
1.92622885e-01 1.64382532e-01 5.66492736e-01 -1.24596775e+00
7.41664529e-01 -7.35780239e-01 -7.85944343e-01 5.95137067e-02
-5.87790132e-01 2.37372920e-01 9.20285285e-01 -8.51311505e-01
-1.93655699e-01 -2.02927995e-03 -1.94167241e-01 -6.18426085e-01
-1.43081561e-01 1.03438653e-01 -5.40742815e-01 -1.75847799e-01
1.06886315e+00 1.38032153e-01 -1.91891253e-01 -6.58560872e-01
2.89810210e-01 3.84053290e-01 4.41653788e-01 -8.47502172e-01
1.35006118e+00 6.83005810e-01 2.09470242e-01 -6.70693994e-01
-7.21740603e-01 6.00089598e-03 -4.11761463e-01 3.80028188e-01
3.96242142e-01 -7.18900561e-01 -9.84702528e-01 8.30936432e-01
-9.51978147e-01 -7.13598609e-01 -1.93648696e-01 -1.51458219e-01
-2.54050672e-01 4.88684386e-01 -8.46746922e-01 -7.41635978e-01
-2.59983420e-01 -9.27579880e-01 7.19364524e-01 -2.71341383e-01
-1.54425725e-01 -1.05237794e+00 -1.38973142e-03 1.74455762e-01
2.11116582e-01 4.35490817e-01 1.23278737e+00 -1.32198358e+00
-4.56144631e-01 -2.45612279e-01 4.09362502e-02 2.70255417e-01
-2.70512432e-01 1.11719459e-01 -9.56612885e-01 -5.98875105e-01
3.41958106e-01 -5.29922843e-01 8.49832594e-01 -7.87553191e-02
1.50306928e+00 -8.85478258e-01 -3.85572046e-01 6.12011969e-01
1.51346314e+00 -1.25657126e-01 5.28617322e-01 3.78382415e-01
8.49274397e-01 1.48713455e-01 1.83475256e-01 6.67761087e-01
-2.08421163e-02 3.64644557e-01 7.35725760e-01 -9.30422246e-02
6.85279071e-02 -5.10439932e-01 3.54528785e-01 2.49998808e-01
5.06326377e-01 -3.58199894e-01 -1.10478675e+00 3.59419078e-01
-1.28136849e+00 -6.73677027e-01 -4.16488320e-01 2.50966144e+00
1.33183038e+00 5.45224905e-01 5.30951977e-01 5.14881253e-01
4.45280731e-01 1.31154433e-01 -7.68777668e-01 -1.71692401e-01
-1.42360702e-01 3.23364913e-01 1.07410288e+00 5.79754472e-01
-1.22757602e+00 1.06432366e+00 7.31965637e+00 9.35329258e-01
-1.10444009e+00 2.21587881e-01 7.30974078e-01 -4.00462598e-01
-3.11442971e-01 -9.32234973e-02 -1.08509457e+00 4.19142962e-01
1.30332065e+00 -2.09697798e-01 2.38295734e-01 7.65327215e-01
-1.19372711e-01 2.75736362e-01 -1.39118803e+00 4.55330968e-01
-1.24574021e-01 -1.27989042e+00 2.45830297e-01 6.10440373e-01
5.50151527e-01 1.73242897e-01 4.45918083e-01 1.36678606e-01
1.00031137e+00 -1.14027190e+00 4.47412938e-01 -1.83792993e-01
5.41099191e-01 -1.08647025e+00 2.53258318e-01 4.49653447e-01
-6.11776829e-01 -4.57488656e-01 -2.21129000e-01 -1.21654123e-02
-3.90022457e-01 3.42077613e-01 -6.31483793e-01 1.26550570e-02
2.80944318e-01 9.60122272e-02 -1.02538741e+00 8.45317721e-01
-1.38651326e-01 1.22979569e+00 -7.09715605e-01 6.93730414e-02
2.36735120e-01 5.08289337e-01 5.35961628e-01 1.04886281e+00
-3.49086851e-01 -8.87794197e-02 7.41917342e-02 6.88476205e-01
-4.34286386e-01 -4.10931885e-01 -7.13980198e-01 2.14640200e-02
8.59734654e-01 8.03532600e-01 -5.48066914e-01 -1.84286878e-01
1.53576769e-02 4.91851002e-01 3.22159320e-01 2.14930832e-01
-1.16924977e+00 1.34430071e-02 9.42163825e-01 2.78086096e-01
-1.09430276e-01 -3.72360975e-01 -7.40373611e-01 -1.14398921e+00
-1.17159091e-01 -1.58776319e+00 5.66247940e-01 1.18992239e-01
-1.32425451e+00 2.14213893e-01 3.21891785e-01 -1.00156701e+00
1.24956742e-01 -4.64491904e-01 -5.20982087e-01 3.59996229e-01
-1.05991697e+00 -9.55056608e-01 2.21198067e-01 6.81088567e-01
4.17220220e-02 2.66409516e-02 6.76340759e-01 1.29934832e-01
-9.24072802e-01 1.23366487e+00 1.80798247e-01 3.06790590e-01
2.66045839e-01 -9.90903020e-01 7.92751431e-01 1.14418232e+00
2.76249617e-01 9.45303440e-01 1.05113804e+00 -7.84504950e-01
-1.69093847e+00 -1.24111807e+00 3.75552326e-01 -1.12209654e+00
1.04443204e+00 -7.47183442e-01 -9.51640964e-01 6.89548969e-01
-3.96068066e-01 9.00116637e-02 6.66832030e-01 -9.71334204e-02
-1.02228034e+00 4.36668843e-02 -1.53705680e+00 9.35355008e-01
1.06013644e+00 -3.75790954e-01 -8.02653804e-02 4.71651107e-01
8.15763652e-01 -1.06237449e-01 -6.89115107e-01 5.96201837e-01
1.30188301e-01 -8.42935383e-01 1.01736450e+00 -1.21968281e+00
3.48339021e-01 4.27715033e-02 -2.07254961e-01 -7.98000574e-01
2.77023047e-01 -8.63122106e-01 -6.29249156e-01 1.14617765e+00
3.41040045e-01 -7.17820406e-01 9.90299344e-01 7.61841238e-01
3.54659319e-01 -7.70120442e-01 -1.02675855e+00 -9.57172275e-01
8.93887579e-01 -5.27011037e-01 6.97686195e-01 1.11083460e+00
-4.97297645e-02 -1.26045808e-01 -4.13654536e-01 6.58685625e-01
1.17253530e+00 -5.38955569e-01 9.83985722e-01 -9.78259921e-01
-3.39172870e-01 -4.88073349e-01 -2.26507232e-01 -8.74141335e-01
-1.64864212e-02 -6.74881220e-01 -3.96465927e-01 -6.35972500e-01
3.34417909e-01 -5.77984810e-01 -3.75816673e-01 7.91049540e-01
-2.71020532e-01 6.66664183e-01 2.39039928e-01 5.01857519e-01
-5.61704338e-01 1.52175114e-01 6.57334626e-01 1.18985295e-03
-6.43270612e-02 -3.57373133e-02 -1.06717241e+00 6.35316849e-01
1.13615084e+00 -1.04298568e+00 -3.42849284e-01 -2.04335377e-01
3.22820932e-01 -5.00106990e-01 7.27445900e-01 -9.78632927e-01
7.41661713e-02 -3.44474196e-01 1.21695958e-01 -3.97273451e-01
-4.80170660e-02 -7.37098217e-01 -2.04027578e-01 1.16226554e+00
-6.53410554e-01 -3.16103324e-02 2.84018010e-01 5.64037323e-01
5.58171213e-01 -6.61010593e-02 1.05548799e+00 1.19021736e-01
9.75009352e-02 4.03474391e-01 -4.49406385e-01 4.38973576e-01
1.16525698e+00 2.29045928e-01 -9.24393833e-01 2.42934674e-02
-3.44277143e-01 1.97276711e-01 6.72486365e-01 2.52423674e-01
2.81093597e-01 -1.03593957e+00 -7.63050675e-01 3.33487302e-01
1.09906152e-01 -2.16136575e-01 -3.28912318e-01 7.52185106e-01
-4.22437072e-01 1.56872228e-01 1.25915006e-01 -4.27606583e-01
-1.48070025e+00 8.62628520e-01 5.13584316e-01 -4.31754351e-01
-4.40725982e-01 7.63113201e-01 1.60582826e-01 -1.88369423e-01
4.95435983e-01 -2.38848731e-01 4.22514915e-01 -4.34473664e-01
6.28221631e-01 3.68286133e-01 6.08727243e-03 -2.89495468e-01
-3.34759444e-01 -5.22969812e-02 -3.28798771e-01 4.20245677e-02
1.01688576e+00 3.47446054e-01 1.23935178e-01 1.08030759e-01
1.13268137e+00 1.62577823e-01 -1.31665838e+00 -6.73283711e-02
1.01572700e-01 -3.49597543e-01 -2.59873301e-01 -8.52755904e-01
-9.62498069e-01 8.36851239e-01 3.56333852e-01 6.63866162e-01
8.63901258e-01 1.62018001e-01 7.69534826e-01 3.88144106e-01
4.29624170e-01 -5.36419928e-01 1.54091567e-01 1.53695270e-01
5.46531379e-01 -8.35717976e-01 1.47650763e-01 -3.09876561e-01
-2.71158576e-01 5.33967316e-01 7.44292974e-01 -3.32447857e-01
6.30116820e-01 8.42001915e-01 -4.45239358e-02 -2.12957770e-01
-9.96595979e-01 2.30611980e-01 -3.32518071e-01 7.33295739e-01
-1.15398005e-01 -2.55804896e-01 -2.79759765e-01 6.26630008e-01
-3.53917778e-01 -4.15331662e-01 6.24180198e-01 9.25608993e-01
-4.38275814e-01 -1.32989061e+00 -5.45556247e-01 3.14986259e-01
-9.92513657e-01 -2.32519284e-01 -9.63223934e-01 1.17853320e+00
-2.14709982e-01 1.09145069e+00 -2.00418949e-01 -5.35457730e-01
1.79010078e-01 8.28261599e-02 3.92403275e-01 -2.44491726e-01
-1.09232664e+00 -2.69866824e-01 -1.38977423e-01 -4.05439347e-01
1.42924979e-01 -4.64125991e-01 -1.03826463e+00 -8.39540422e-01
-3.40126783e-01 1.67929843e-01 4.79147851e-01 8.75929475e-01
2.83980429e-01 -9.68656503e-03 8.94183397e-01 -2.83906430e-01
-1.38384271e+00 -7.21388221e-01 -3.59248161e-01 6.02488279e-01
5.29813528e-01 -2.50234067e-01 -1.06694710e+00 -1.35803536e-01] | [5.835359573364258, 7.5794525146484375] |
f6f4cb92-c940-45ba-9f8e-db91c19cf8b5 | domain-shifts-in-dermoscopic-skin-cancer | 2304.06968 | null | https://arxiv.org/abs/2304.06968v3 | https://arxiv.org/pdf/2304.06968v3.pdf | Domain shifts in dermoscopic skin cancer datasets: Evaluation of essential limitations for clinical translation | The limited ability of Convolutional Neural Networks to generalize to images from previously unseen domains is a major limitation, in particular, for safety-critical clinical tasks such as dermoscopic skin cancer classification. In order to translate CNN-based applications into the clinic, it is essential that they are able to adapt to domain shifts. Such new conditions can arise through the use of different image acquisition systems or varying lighting conditions. In dermoscopy, shifts can also occur as a change in patient age or occurence of rare lesion localizations (e.g. palms). These are not prominently represented in most training datasets and can therefore lead to a decrease in performance. In order to verify the generalizability of classification models in real world clinical settings it is crucial to have access to data which mimics such domain shifts. To our knowledge no dermoscopic image dataset exists where such domain shifts are properly described and quantified. We therefore grouped publicly available images from ISIC archive based on their metadata (e.g. acquisition location, lesion localization, patient age) to generate meaningful domains. To verify that these domains are in fact distinct, we used multiple quantification measures to estimate the presence and intensity of domain shifts. Additionally, we analyzed the performance on these domains with and without an unsupervised domain adaptation technique. We observed that in most of our grouped domains, domain shifts in fact exist. Based on our results, we believe these datasets to be helpful for testing the generalization capabilities of dermoscopic skin cancer classifiers. | ['Titus J. Brinker', 'Julia Niebling', 'Roman C. Maron', 'Sireesha Chamarthi', 'Katharina Fogelberg'] | 2023-04-14 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [ 4.36905921e-01 -2.44785160e-01 -8.31687357e-03 -3.41377050e-01
-3.70781273e-01 -9.35287774e-01 4.33757544e-01 4.99214113e-01
-4.92551863e-01 7.59598494e-01 -1.91220090e-01 -4.56340551e-01
-2.52273530e-01 -7.26472080e-01 -8.31380546e-01 -5.74974954e-01
5.30322045e-02 3.36920768e-01 3.69487703e-01 -1.87870506e-02
1.07957534e-02 7.82467961e-01 -1.41131020e+00 4.69093770e-01
9.15270686e-01 7.03061223e-01 1.35148481e-01 5.80079257e-01
2.10964959e-02 1.59445107e-01 -8.25213492e-01 -2.89291948e-01
4.46294278e-01 -3.18587035e-01 -6.52590454e-01 1.90942809e-01
7.53431261e-01 -5.97436666e-01 -1.12352178e-01 9.08166170e-01
4.10721630e-01 -8.14914927e-02 8.05496633e-01 -7.78524518e-01
-4.05806839e-01 4.22324464e-02 -2.94785172e-01 2.38923863e-01
1.76706091e-01 3.79896283e-01 3.38929445e-01 -2.25526109e-01
1.06990635e+00 5.76815486e-01 7.01735914e-01 6.27036035e-01
-1.25539792e+00 -4.88913983e-01 5.31527326e-02 3.69320698e-02
-1.00242364e+00 -1.23851173e-01 3.93923968e-01 -5.56981921e-01
5.24312437e-01 2.14476258e-01 7.56912768e-01 1.37366807e+00
2.54971474e-01 3.58860157e-02 1.40252006e+00 -4.37946707e-01
2.37906098e-01 6.37632012e-01 -3.26191217e-01 3.96745890e-01
6.01312995e-01 1.28343523e-01 -1.07685372e-01 5.39736683e-03
8.73305142e-01 -8.99272487e-02 -2.58393049e-01 -3.09102207e-01
-1.00532281e+00 4.62655813e-01 5.73974013e-01 5.06767988e-01
-2.53295094e-01 -1.71784267e-01 3.31778854e-01 3.39030594e-01
2.43300349e-01 7.44019806e-01 -4.12324697e-01 -6.22872747e-02
-8.76790404e-01 -9.61646661e-02 7.36279666e-01 5.70818186e-01
3.84968132e-01 -3.80179495e-01 8.62112194e-02 7.43754327e-01
-4.56186742e-01 1.32367879e-01 6.01814210e-01 -5.81782997e-01
2.19890401e-02 7.16135859e-01 9.29986686e-02 -6.68981373e-01
-5.83650112e-01 -3.63329291e-01 -5.51935792e-01 2.88081318e-01
8.40809047e-01 -1.06981434e-01 -1.39396703e+00 1.55419815e+00
2.67547876e-01 -4.43039723e-02 -3.98823544e-02 9.28141236e-01
7.76824653e-01 -3.03561389e-01 3.26895028e-01 5.74149005e-02
1.21209204e+00 -1.01316668e-01 -4.76765633e-01 -8.56065080e-02
7.33258784e-01 -9.98263955e-01 9.89489794e-01 4.67259705e-01
-6.59693062e-01 -3.41681361e-01 -1.01969790e+00 2.79880792e-01
-6.34544015e-01 5.48253991e-02 6.04399681e-01 7.69011915e-01
-7.81744421e-01 8.06567967e-01 -8.99223506e-01 -1.03244638e+00
4.57507342e-01 3.62916738e-01 -6.79151058e-01 -1.02507293e-01
-1.11128759e+00 1.03718901e+00 2.34254539e-01 -1.83916867e-01
-5.35782218e-01 -7.93181479e-01 -5.36146820e-01 -3.52128625e-01
2.43103698e-01 -6.28869593e-01 1.15784347e+00 -1.32604122e+00
-1.01618826e+00 1.20265424e+00 4.57658246e-02 -4.18448389e-01
7.29072630e-01 3.29268068e-01 -7.55357444e-01 2.37038985e-01
-1.47361934e-01 4.91124511e-01 7.24346399e-01 -1.11146224e+00
-5.90027750e-01 -5.50846994e-01 1.90356374e-01 -8.90248269e-03
-4.03720051e-01 -1.08783111e-01 -8.05879384e-02 -4.68046725e-01
-2.55523890e-01 -1.25705040e+00 -1.73812151e-01 3.93021256e-01
-1.21679448e-01 3.20573956e-01 7.16168344e-01 -6.97416246e-01
8.59020233e-01 -2.14746022e+00 -4.29202229e-01 2.73764610e-01
3.27336341e-01 6.18577302e-01 -1.25257373e-01 4.17520255e-01
-2.83718079e-01 3.88211817e-01 2.42191479e-02 1.75374493e-01
-5.63209355e-01 9.13306102e-02 2.11263254e-01 5.96494138e-01
4.01538253e-01 5.79624712e-01 -7.46033013e-01 -4.43121880e-01
4.05143023e-01 4.62737828e-01 -2.72088230e-01 -1.21391721e-01
-2.38838583e-01 5.98203123e-01 -1.08917341e-01 6.37896121e-01
6.86021328e-01 -2.58072525e-01 2.23346487e-01 -4.02543992e-01
1.72420219e-02 9.99748334e-02 -8.03773582e-01 1.38851714e+00
-5.50720513e-01 6.70209289e-01 -7.35677183e-02 -7.12336063e-01
6.99555576e-01 3.24420094e-01 5.92715621e-01 -5.41425467e-01
5.66187613e-02 4.17857915e-01 5.66123426e-01 -6.68206930e-01
3.84644896e-01 -4.37027991e-01 3.80963385e-01 9.76666063e-03
-7.32991844e-03 -1.15797698e-01 1.27618879e-01 -1.06386796e-01
1.20542216e+00 -3.51819605e-01 1.84294850e-01 -1.46225974e-01
6.70026541e-02 5.11350632e-01 3.36357176e-01 5.11501789e-01
-4.70550954e-01 1.02502012e+00 5.28901398e-01 -3.02914500e-01
-1.14318621e+00 -1.18674302e+00 -6.45392776e-01 4.51859236e-01
4.60326225e-02 4.93620802e-03 -5.44064939e-01 -7.46752262e-01
1.99961126e-01 3.90385121e-01 -9.35145140e-01 -3.54850829e-01
-1.39021933e-01 -7.79428542e-01 5.36222160e-01 5.33465624e-01
2.76573837e-01 -7.93851435e-01 -6.87366128e-01 8.38990733e-02
2.25829616e-01 -1.12769842e+00 -9.77630317e-02 2.60470957e-01
-7.13092089e-01 -1.46736610e+00 -8.69723856e-01 -6.33393347e-01
9.06977415e-01 8.28796104e-02 9.18617010e-01 -5.51422732e-03
-8.73350024e-01 5.53224266e-01 -3.14516783e-01 -6.17173433e-01
-7.63477147e-01 2.18428239e-01 -3.39101031e-02 -2.17831507e-01
3.86199653e-01 -3.09266120e-01 -1.02031195e+00 3.32108736e-01
-1.27343106e+00 -4.33795184e-01 7.40275085e-01 8.20043862e-01
4.89835501e-01 1.33803576e-01 4.37374383e-01 -1.20242453e+00
8.81656647e-01 -4.67216343e-01 -4.38572943e-01 2.40412980e-01
-5.36440372e-01 -2.82706618e-01 6.24938428e-01 -7.41757333e-01
-8.31892073e-01 9.57078412e-02 1.08926162e-01 -3.70531172e-01
-7.04344571e-01 5.67354202e-01 3.62893850e-01 -2.69708216e-01
1.22438920e+00 -5.96072190e-02 4.75719362e-01 -2.50291169e-01
-1.44059822e-01 9.15088654e-01 2.95840293e-01 -2.65660733e-01
6.67608857e-01 6.25788450e-01 8.89396518e-02 -8.95210445e-01
-4.69256818e-01 -5.30321598e-01 -5.88531792e-01 -9.48729068e-02
7.84567177e-01 -7.44879723e-01 -3.30683768e-01 5.33096492e-01
-6.83457255e-01 -4.89739031e-01 -2.74148554e-01 4.31695312e-01
-2.50378281e-01 4.00998414e-01 -3.26179147e-01 -4.27248210e-01
8.92980024e-03 -1.15599823e+00 7.27364480e-01 4.59243149e-01
-4.94382203e-01 -1.39483154e+00 -1.58445865e-01 1.95468292e-01
5.55065274e-01 6.23884082e-01 9.07059550e-01 -6.87911987e-01
-1.13908522e-01 -4.24990088e-01 -4.02024053e-02 4.70559955e-01
7.86384284e-01 4.57550257e-01 -9.45843160e-01 -5.34833431e-01
-3.02691191e-01 -2.00935543e-01 7.48905659e-01 3.39404613e-01
1.35448599e+00 6.41731098e-02 -4.16397750e-01 5.87345302e-01
1.40887308e+00 3.91367227e-01 5.60603321e-01 4.00326431e-01
2.59122014e-01 8.37499917e-01 5.72461128e-01 2.94745918e-02
-1.18119583e-01 5.96159816e-01 3.65455627e-01 -4.97240931e-01
-3.75668705e-01 -8.53771269e-02 -9.33260843e-02 -2.96025537e-02
-1.56212896e-01 -9.57222059e-02 -9.78424609e-01 8.52780521e-01
-1.14840496e+00 -6.70259595e-01 8.71769711e-02 2.52058411e+00
7.15326786e-01 1.76272944e-01 2.86839336e-01 -1.69527531e-01
8.12228739e-01 -2.42801636e-01 -8.23507190e-01 -4.93574709e-01
3.04689854e-02 3.26799482e-01 9.18426514e-01 8.73414800e-02
-8.54578912e-01 5.62318206e-01 6.32625914e+00 3.90459210e-01
-1.85070395e+00 -2.71823138e-01 5.35173118e-01 -1.80630371e-01
-1.12273246e-01 -2.03346133e-01 -3.79039437e-01 5.43173850e-01
7.09584713e-01 -1.51467323e-03 3.49565536e-01 5.77714324e-01
1.34950932e-02 -3.10503662e-01 -1.11175644e+00 6.94898903e-01
-1.57202572e-01 -1.14296365e+00 -2.57925659e-01 1.98400974e-01
5.66258192e-01 -4.46499139e-02 5.01215279e-01 -1.22463837e-01
9.75638628e-02 -1.29913807e+00 -6.02261387e-02 3.02656800e-01
1.42834496e+00 -3.88608932e-01 8.29700470e-01 2.21407637e-02
-5.10291159e-01 2.40313843e-01 -2.02269778e-01 3.43979657e-01
-2.58584201e-01 5.98562539e-01 -1.59773040e+00 3.22333217e-01
7.04553723e-01 3.53358835e-01 -7.66496003e-01 1.19488406e+00
4.58944812e-02 6.18534982e-01 -4.93136168e-01 -2.63741296e-02
9.56368446e-02 2.25242227e-01 3.85025561e-01 8.77270579e-01
2.48292550e-01 -3.63740116e-01 -5.16442358e-01 6.90033197e-01
6.65466115e-02 -4.48557958e-02 -6.46452606e-01 -3.59760821e-01
4.88760412e-01 1.27072310e+00 -6.49406910e-01 1.62496328e-01
-5.51842690e-01 9.61757660e-01 -5.15212677e-02 4.52954024e-01
-6.14800215e-01 -2.56945848e-01 8.89771521e-01 7.78468609e-01
1.05610386e-01 -1.84301939e-02 -1.95985436e-01 -9.20817733e-01
1.07443057e-01 -9.81814384e-01 4.83091712e-01 -3.97459656e-01
-1.53874314e+00 4.96880352e-01 -9.90612060e-02 -1.36331975e+00
-2.04634309e-01 -9.21095908e-01 -4.94562745e-01 1.00670850e+00
-1.44609571e+00 -1.09431088e+00 -8.23199093e-01 5.71746409e-01
-3.82566005e-02 -7.95928240e-02 8.77493560e-01 1.57644138e-01
-3.11455399e-01 9.12625909e-01 1.55589089e-01 4.53922525e-02
1.40559912e+00 -1.36751401e+00 7.08016232e-02 4.38301295e-01
-1.11231372e-01 8.94132912e-01 6.74166262e-01 -6.55446529e-01
-7.81600773e-01 -1.14045107e+00 2.03909799e-01 -5.26740968e-01
5.52329838e-01 -1.44116297e-01 -1.04820466e+00 5.99982142e-01
-1.80790052e-01 8.64757784e-03 9.62409317e-01 2.36666963e-01
-2.91282356e-01 -2.78130561e-01 -1.54995763e+00 6.96981966e-01
8.13824952e-01 -3.77677619e-01 -1.84256509e-01 5.02367675e-01
2.55877167e-01 -7.53304660e-01 -1.18801022e+00 6.42928541e-01
6.20525718e-01 -1.01087999e+00 7.86282301e-01 -7.68936634e-01
6.43991172e-01 -6.51015714e-03 2.84440845e-01 -1.42188358e+00
-1.19000457e-01 -6.20701760e-02 4.61336315e-01 1.01309609e+00
5.26461244e-01 -9.87391353e-01 1.04155159e+00 8.20648193e-01
1.24820583e-01 -8.47602069e-01 -7.86970794e-01 -8.17027330e-01
5.60107112e-01 9.82585326e-02 5.86188197e-01 9.78486419e-01
-8.12044591e-02 -4.06292319e-01 2.31932908e-01 1.70731872e-01
1.35115355e-01 -5.75340465e-02 7.09074497e-01 -1.06036222e+00
-2.40517274e-01 -2.91280717e-01 -7.55477607e-01 -1.61473915e-01
-2.60620683e-01 -7.89028406e-01 -3.09330434e-01 -1.49290788e+00
3.73922959e-02 -7.65508950e-01 -4.11323965e-01 4.69870895e-01
-1.24270394e-01 2.04165801e-01 -2.93778069e-02 2.47394383e-01
-1.23617006e-02 -2.78339475e-01 1.38390768e+00 1.37894135e-02
-3.56063545e-01 1.07326224e-01 -7.88705707e-01 5.19905865e-01
1.21115518e+00 -2.49721199e-01 -3.45319301e-01 -1.75011605e-01
-4.93003801e-02 -2.29049116e-01 4.40326452e-01 -1.23097634e+00
1.04991339e-01 -3.16843122e-01 7.03769565e-01 -1.33207723e-01
2.27341741e-01 -9.32639718e-01 2.03769639e-01 6.49272382e-01
-4.52222005e-02 -3.76912981e-01 6.12567067e-01 3.67930919e-01
-2.86628723e-01 -1.86923489e-01 1.01379192e+00 -2.39978001e-01
-6.98294163e-01 1.51237264e-01 -4.61858600e-01 -8.20580721e-02
1.36809337e+00 -6.19533777e-01 -4.99820590e-01 -3.44269931e-01
-7.53561616e-01 -1.35734662e-01 1.10053253e+00 3.58716905e-01
2.63393849e-01 -8.53972256e-01 -5.08995473e-01 1.38368472e-01
5.43610036e-01 1.34917766e-01 5.39956331e-01 8.47563565e-01
-8.39547634e-01 3.24994445e-01 -5.65101862e-01 -5.90128362e-01
-1.41213095e+00 3.94159287e-01 7.08162248e-01 -1.24468297e-01
-2.07647383e-01 6.42031074e-01 4.60863188e-02 -3.77645046e-01
-2.39157006e-02 -3.15891415e-01 1.94719627e-01 -2.34185345e-02
1.64945558e-01 -7.29555190e-02 5.26443362e-01 -7.41395056e-02
-3.02869111e-01 3.07285845e-01 -5.10793567e-01 2.57048815e-01
1.05967426e+00 2.06687331e-01 2.92787462e-01 2.21241534e-01
9.18419957e-01 -1.13990409e-02 -1.17750418e+00 1.25999108e-01
-4.20917302e-01 -4.70704913e-01 -2.65696049e-01 -1.24332547e+00
-9.32563126e-01 8.17865252e-01 1.00297129e+00 8.59486014e-02
1.32220423e+00 1.62218162e-03 5.44936299e-01 1.86087228e-02
3.87759149e-01 -1.19684660e+00 -2.51196977e-02 -3.56234685e-02
7.26106346e-01 -1.27458334e+00 -3.36607732e-02 -6.26623988e-01
-5.40670276e-01 1.07548594e+00 7.80849874e-01 4.69692275e-02
4.19528276e-01 3.29416662e-01 5.00207484e-01 -1.01930141e-01
-4.38267231e-01 -1.93364471e-01 1.74147397e-01 1.13064289e+00
5.35278440e-01 1.51289910e-01 -5.94876826e-01 2.60242134e-01
-2.97923952e-01 2.33816832e-01 8.03877532e-01 9.01643276e-01
4.17824872e-02 -1.33500946e+00 -2.94018805e-01 1.01014614e+00
-5.91068864e-01 2.32444122e-01 -7.84530401e-01 1.27304482e+00
3.55651319e-01 6.49878800e-01 1.13057986e-01 -3.84024650e-01
3.88745546e-01 -7.10322261e-02 7.18275547e-01 -9.59160089e-01
-5.87260306e-01 -4.63428050e-01 3.25835161e-02 -2.46993616e-01
-2.16599256e-01 -6.80523634e-01 -1.01492965e+00 -2.35666499e-01
-3.13682616e-01 -4.41673726e-01 8.30162227e-01 5.16773820e-01
2.06622303e-01 6.20925784e-01 1.22712746e-01 -1.25677720e-01
-6.07617855e-01 -9.12739992e-01 -6.99196398e-01 8.55081916e-01
2.53574193e-01 -6.02953076e-01 -4.65767831e-01 1.68069452e-01] | [15.4923734664917, -2.9318130016326904] |
f7c4358b-f7f9-4fa1-b8ce-240e61094325 | overlapping-community-detection-using-dynamic | 2210.11174 | null | https://arxiv.org/abs/2210.11174v1 | https://arxiv.org/pdf/2210.11174v1.pdf | Overlapping Community Detection using Dynamic Dilated Aggregation in Deep Residual GCN | Overlapping community detection is a key problem in graph mining. Some research has considered applying graph convolutional networks (GCN) to tackle the problem. However, it is still challenging to incorporate deep graph convolutional networks in the case of general irregular graphs. In this study, we design a deep dynamic residual graph convolutional network (DynaResGCN) based on our novel dynamic dilated aggregation mechanisms and a unified end-to-end encoder-decoder-based framework to detect overlapping communities in networks. The deep DynaResGCN model is used as the encoder, whereas we incorporate the Bernoulli-Poisson (BP) model as the decoder. Consequently, we apply our overlapping community detection framework in a research topics dataset without having ground truth, a set of networks from Facebook having a reliable (hand-labeled) ground truth, and in a set of very large co-authorship networks having empirical (not hand-labeled) ground truth. Our experimentation on these datasets shows significantly superior performance over many state-of-the-art methods for the detection of overlapping communities in networks. | ['Md Saidur Rahman', 'Md Iqbal Hossain', 'Md Nurul Muttakin'] | 2022-10-20 | null | null | null | null | ['graph-mining', 'community-detection'] | ['graphs', 'graphs'] | [-3.0486293e-03 3.3583844e-01 2.2348334e-01 3.0376082e-02
-2.9274660e-01 -4.2394441e-01 3.6528021e-01 2.4616629e-01
-2.7022013e-01 6.0007858e-01 -1.7178611e-01 -5.7294512e-01
1.8020160e-02 -1.2640074e+00 -9.6929473e-01 -4.1723546e-01
-5.8716965e-01 7.4326313e-01 3.2109234e-01 1.2573995e-01
6.7692801e-02 2.9080370e-01 -9.0075582e-01 3.2197484e-01
9.2329973e-01 4.1201267e-01 1.6267602e-01 1.0061151e+00
-6.2150106e-02 8.6074495e-01 -3.7612945e-01 -9.5234114e-01
2.9379100e-01 -2.2028777e-01 -8.9898539e-01 2.0405526e-01
3.8041455e-01 -2.0339790e-01 -7.7615041e-01 1.1378870e+00
4.0954691e-01 -4.1929451e-01 4.7639924e-01 -1.4301987e+00
-8.1830090e-01 1.0458109e+00 -8.8701355e-01 2.7515566e-01
3.2414887e-02 2.1395455e-01 1.2900236e+00 -3.2487559e-01
9.5193952e-01 1.3572737e+00 8.1251001e-01 1.7796160e-01
-1.3004287e+00 -8.2821411e-01 1.0877720e-01 -1.3474914e-03
-1.4117146e+00 4.9689889e-02 5.0740433e-01 -5.9868222e-01
7.5867718e-01 -2.8823876e-01 6.8895257e-01 1.1583012e+00
2.0560829e-01 6.7119914e-01 6.1274529e-01 -2.8086033e-01
5.7620004e-02 -1.9165580e-01 2.7809983e-02 9.1691852e-01
7.3066109e-01 -1.7473291e-01 -9.7082347e-02 -2.6233023e-01
1.1114696e+00 1.9084691e-01 -1.5314655e-02 -3.1474528e-01
-1.0008366e+00 1.0181394e+00 8.4685284e-01 2.6910001e-01
-2.2859420e-01 6.3164836e-01 5.4498762e-01 4.6842939e-01
5.5440688e-01 1.6328157e-01 -7.6540343e-02 3.6559632e-01
-9.9784958e-01 1.5427446e-01 1.1823567e+00 1.1306080e+00
9.1136247e-01 -2.0612334e-01 -1.6106379e-01 3.5835353e-01
2.9472476e-01 -3.9336845e-02 -1.9848186e-01 -6.0518450e-01
4.5906398e-01 9.2509598e-01 -5.0439280e-01 -1.2531087e+00
-2.2602531e-01 -6.9948941e-01 -1.3633833e+00 -9.8776303e-02
4.6073523e-01 -4.1216683e-01 -8.4711665e-01 1.5382941e+00
1.3173053e-01 5.1774591e-01 -7.3018201e-02 8.4337753e-01
9.0941292e-01 2.9675993e-01 -2.6982865e-01 2.3800382e-01
1.2597096e+00 -9.7999996e-01 -3.5533637e-01 -2.6433667e-02
7.2203916e-01 -5.6006968e-01 4.8086867e-01 2.2097625e-01
-7.2729290e-01 -2.1372762e-01 -7.1443129e-01 -7.2953470e-02
-2.9272720e-01 1.6289981e-01 8.0783945e-01 6.1679816e-01
-1.5015516e+00 7.0487052e-01 -7.5728273e-01 -4.6560532e-01
8.7537462e-01 3.6564299e-01 -5.2355444e-01 -3.3787572e-01
-9.3765175e-01 1.4158471e-01 4.0514463e-01 -1.9986417e-02
-1.2029253e+00 -3.3835763e-01 -7.3173493e-01 3.6848393e-01
6.6173828e-01 -8.5369545e-01 6.3412440e-01 -9.8807174e-01
-8.0164844e-01 9.3113315e-01 1.8504640e-01 -8.0886394e-01
6.0808063e-01 2.7181304e-01 -1.8003611e-01 3.2381222e-01
1.9705057e-01 5.6322241e-01 5.0519478e-01 -1.0209244e+00
-2.0820761e-01 -1.8002608e-01 2.0592472e-01 -3.7179533e-01
-2.6700586e-01 2.6712054e-02 -5.0884050e-01 -6.5458328e-01
-1.3398263e-01 -1.0619670e+00 -4.7171307e-01 -6.1750021e-02
-9.4937247e-01 -2.7044415e-01 6.5625250e-01 -6.7335862e-01
1.1922716e+00 -1.9649518e+00 9.4236344e-02 5.4647416e-01
1.0050890e+00 1.7911719e-01 -4.5998928e-01 4.5091903e-01
-1.0157191e-02 4.6211034e-01 -2.6828200e-01 -4.3378046e-01
-2.5305778e-01 1.1624106e-01 2.3469372e-01 4.4946268e-01
5.1412070e-01 8.7203407e-01 -1.0354921e+00 -4.4177920e-01
-3.6428872e-01 3.3404696e-01 -7.6486421e-01 1.5043914e-01
-4.0096337e-01 2.1712148e-01 -2.3716915e-01 5.5974358e-01
8.8382179e-01 -9.9689680e-01 6.3576603e-01 3.3671942e-01
2.3699948e-01 1.1053585e-01 -1.3166414e+00 1.3778753e+00
5.4753896e-02 6.8580663e-01 3.3818218e-01 -9.8186541e-01
1.0137221e+00 2.3708580e-01 3.7354928e-01 -9.6082121e-02
8.4512614e-02 2.3720275e-01 4.8972556e-01 -1.0488587e-01
3.9154413e-01 2.7307367e-01 3.0068448e-01 6.8809819e-01
2.8615940e-01 5.1217747e-01 5.8847463e-01 7.7456909e-01
2.0683448e+00 -5.3743613e-01 -6.9105878e-02 -2.8845528e-01
2.6687023e-01 -1.4894426e-01 3.7338206e-01 9.3070161e-01
-6.1203115e-02 7.8041142e-01 1.3611734e+00 -4.3101838e-01
-1.2190784e+00 -7.8115255e-01 3.4141421e-01 5.4677486e-01
-1.4096427e-01 -5.7783955e-01 -6.8308997e-01 -9.3511868e-01
1.4714980e-01 -7.2610125e-02 -5.9071207e-01 1.8888369e-01
-4.2742589e-01 -7.5578117e-01 7.4122828e-01 3.3666459e-01
6.0775840e-01 -1.0265934e+00 1.6208585e-01 4.4627529e-01
-8.7986797e-02 -1.4903510e+00 -5.7618672e-01 -7.7915758e-02
-6.2914401e-01 -1.7502550e+00 -6.7837727e-01 -7.8622955e-01
6.1235511e-01 4.7775587e-01 1.3540679e+00 5.5616289e-01
-3.3975384e-01 1.7923042e-01 -3.0253765e-01 -1.1623383e-01
-6.4319086e-01 4.7314358e-01 -5.2130336e-01 7.4574031e-02
4.1807392e-01 -8.1774116e-01 -5.4300445e-01 1.5482386e-02
-8.5962659e-01 -5.3053439e-02 6.6292703e-01 6.2584418e-01
1.7448381e-01 2.2244148e-01 6.2945545e-01 -1.2339518e+00
7.5890046e-01 -8.3577144e-01 -7.6790857e-01 1.4714582e-01
-4.5007235e-01 -1.6495362e-02 5.3528529e-01 -2.4880156e-01
-2.7476135e-01 -5.3645860e-02 9.9814117e-02 -6.8934536e-01
2.0036420e-02 7.7086055e-01 -6.9412529e-02 -1.0990099e-01
4.6488160e-01 -2.1283604e-01 2.9120243e-01 -2.6924753e-01
1.9400880e-01 5.5968004e-01 1.9927886e-01 -2.8719342e-01
6.1283797e-01 5.4826111e-01 1.9266374e-01 -6.8781477e-01
-4.2227760e-01 -6.5369600e-01 -6.8677551e-01 -1.3369229e-01
7.4503320e-01 -1.2448165e+00 -7.9314685e-01 4.3341807e-01
-1.4382081e+00 -3.5672462e-01 1.5725458e-01 1.9850557e-01
-9.1895521e-02 7.7281743e-01 -9.9832428e-01 -6.6883385e-01
-5.8056974e-01 -1.1820830e+00 1.0608282e+00 -4.5575716e-02
1.7143355e-01 -1.1234907e+00 -7.8503720e-02 1.7304286e-01
2.5222477e-01 6.4625442e-01 7.6801181e-01 -8.2296175e-01
-1.0688797e+00 -2.0749776e-01 -8.3486056e-01 2.4067649e-01
-7.5060345e-02 2.6811460e-01 -6.7827386e-01 -4.8243415e-01
-8.6292022e-01 -9.1261491e-03 1.2462136e+00 3.1399930e-01
1.2695622e+00 -3.4376115e-01 -5.1658946e-01 4.8697940e-01
1.4546627e+00 -5.5291617e-01 8.8755912e-01 1.2129496e-01
1.0495877e+00 3.4303120e-01 -2.7638313e-01 2.9038367e-01
6.2475944e-01 1.8870060e-01 7.5907087e-01 -1.9075835e-01
-1.0419256e-01 -3.7196752e-02 2.8367054e-01 7.6458609e-01
-6.3666053e-02 -7.5134909e-01 -1.0951473e+00 8.1605899e-01
-1.9943554e+00 -1.0077083e+00 -8.6919558e-01 1.8445749e+00
8.1987989e-01 2.8532884e-01 2.8211281e-01 -1.5988596e-01
1.3703918e+00 -2.8302450e-02 -2.3350471e-01 -1.7672278e-01
-2.1414688e-01 2.0548691e-01 6.0403830e-01 2.3076619e-01
-1.2891680e+00 7.8706783e-01 5.5175004e+00 6.7277312e-01
-8.4286451e-01 1.1074624e-01 7.5808036e-01 4.5420584e-01
-3.2862216e-01 2.2894399e-01 -6.1306417e-01 4.9926090e-01
8.8476378e-01 -2.2062771e-02 5.0873822e-01 7.2750294e-01
-1.3419987e-01 7.2084971e-02 -9.6883667e-01 7.7353948e-01
-1.5280637e-01 -1.5603096e+00 3.6572009e-02 6.0681784e-01
1.0045668e+00 3.0799234e-01 -4.4492400e-01 2.3275223e-01
1.0099643e+00 -1.0697109e+00 3.5757869e-01 4.3687624e-01
9.0286762e-01 -6.1719996e-01 9.3801391e-01 3.0173746e-01
-1.2595427e+00 5.6015212e-02 -6.1677271e-01 8.8014841e-02
-1.3901880e-01 1.1952013e+00 -1.1236154e+00 8.1456161e-01
5.7393974e-01 1.0221044e+00 -6.7937028e-01 1.2321184e+00
-5.4371174e-02 7.8311288e-01 -4.2220747e-01 5.9636004e-02
4.0437099e-01 -2.5874642e-01 6.9320619e-01 1.3480716e+00
3.1740329e-01 -4.5886746e-01 2.9197270e-01 1.3966315e+00
-8.1610626e-01 -2.2949912e-01 -8.2614374e-01 -4.3095154e-01
2.8634930e-01 1.5133567e+00 -1.0339416e+00 -2.4791388e-01
-5.2486527e-01 8.7357134e-01 7.9926610e-01 2.2610173e-01
-6.5813577e-01 -2.9999310e-01 3.9427298e-01 5.0679779e-01
6.0692191e-01 -3.4479812e-01 -1.7517990e-02 -1.2213089e+00
-2.1436001e-01 -8.2351512e-01 5.6232536e-01 -4.7555482e-01
-1.6958253e+00 4.6921873e-01 -3.3688259e-01 -7.1942055e-01
2.8716481e-01 -5.4285151e-01 -1.0360055e+00 8.9077616e-01
-1.3726840e+00 -1.2574595e+00 -5.4636043e-01 3.7999329e-01
-1.6678201e-01 -3.1855831e-01 4.5533705e-01 5.1931387e-01
-6.4429587e-01 5.4627782e-01 2.1286441e-02 6.9089448e-01
5.1915276e-01 -1.5323282e+00 8.1736106e-01 1.0650568e+00
2.7506785e-02 6.2799084e-01 2.2624955e-01 -1.1507301e+00
-1.1473596e+00 -1.6163622e+00 7.6097518e-01 -1.2890717e-02
7.9393691e-01 -6.7354304e-01 -1.0429803e+00 8.1792539e-01
2.5273749e-01 2.6057661e-01 4.1933694e-01 2.1670453e-01
-3.2825258e-01 1.6811812e-01 -9.7447991e-01 3.9685008e-01
1.2821660e+00 -3.8951963e-01 2.3971823e-01 5.2210337e-01
8.5916954e-01 -2.0717424e-01 -7.7250057e-01 1.6260344e-01
2.2665104e-01 -9.2395931e-01 5.3309560e-01 -6.0029036e-01
8.3024418e-01 -2.6922888e-01 2.8238595e-01 -1.1747427e+00
-4.6625090e-01 -5.9184772e-01 -2.0979159e-01 1.1690408e+00
2.1249826e-01 -5.3705114e-01 1.0098828e+00 -6.7441285e-02
8.0101073e-02 -5.4838735e-01 -7.2913426e-01 -6.3855398e-01
1.3062911e-02 -1.0749882e-01 6.4734071e-01 1.0897456e+00
-3.3289003e-01 4.1259930e-01 -1.0344612e-01 2.5085843e-01
7.8126502e-01 -1.5945738e-02 1.0869994e+00 -1.5623560e+00
-2.1610418e-01 -4.2585424e-01 -7.4558938e-01 -7.3393297e-01
2.2623736e-01 -1.2781075e+00 -4.2437214e-01 -1.8129736e+00
6.4611948e-01 -4.3268374e-01 1.3302056e-01 3.3940870e-01
-1.2876184e-01 1.5267685e-01 -1.1812257e-02 1.1239685e-01
-9.1867107e-01 2.7789384e-01 1.3347933e+00 -2.7579743e-01
1.2060640e-01 -5.5928458e-02 -7.2935319e-01 2.8487974e-01
7.0162547e-01 -5.4452515e-01 -1.3302875e-02 -3.1123373e-01
6.0350996e-01 -6.9771454e-02 8.2500589e-01 -9.5209187e-01
4.2848849e-01 4.0925255e-01 1.4395472e-01 -6.4220732e-01
-4.4514588e-01 -6.6564584e-01 2.7973819e-01 4.8550829e-01
-7.7908367e-02 2.0267019e-01 -4.2520864e-03 1.1066140e+00
4.2069107e-02 -8.5927621e-03 6.2459821e-01 -3.8552004e-01
-1.9395711e-01 6.0012156e-01 -3.8345078e-01 1.6679950e-01
8.8924497e-01 -1.3007884e-01 -6.0994858e-01 -4.5158249e-01
-8.3177596e-01 5.0027567e-01 5.6750494e-01 1.4984897e-01
4.8718157e-01 -1.0929998e+00 -1.0719353e+00 7.6895662e-02
-6.6360772e-02 3.3825043e-01 5.9411626e-02 8.8783950e-01
-9.1608799e-01 5.3888224e-02 -1.5308203e-02 -6.1475670e-01
-1.1613239e+00 4.8430765e-01 4.9456099e-01 -8.2730550e-01
-5.5030453e-01 7.1220076e-01 6.7243576e-02 -6.8047309e-01
1.4665478e-01 -1.6243698e-01 -1.2038213e-01 -1.6733535e-01
1.2468967e-01 3.9465898e-01 4.4929408e-02 -1.5571138e-01
-1.7958689e-01 -2.1209663e-01 -2.5807050e-01 5.1541466e-01
1.4267687e+00 1.2443743e-01 -7.2236311e-01 -2.8311886e-02
1.1252328e+00 -1.8017220e-01 -8.4793150e-01 -4.6474391e-01
1.8078807e-01 -2.2350791e-01 -1.3446668e-01 -3.7225750e-01
-1.4318343e+00 8.5673338e-01 1.8295465e-01 5.7272011e-01
6.1914909e-01 4.9272198e-02 6.3650900e-01 1.9804235e-01
2.6386788e-01 -7.0494705e-01 3.3454818e-01 7.8176796e-01
6.7083150e-01 -1.3467995e+00 5.6327619e-02 -7.2028595e-01
-2.7615523e-01 1.3588810e+00 8.0067635e-01 -3.4424150e-01
9.5396024e-01 2.4440117e-01 -6.2360662e-01 -7.4682534e-01
-8.5104430e-01 -5.2417922e-01 -5.7208888e-02 5.4177403e-01
4.2992085e-01 2.0005038e-01 3.1326529e-02 5.3331995e-01
3.6991168e-02 -8.6686440e-02 9.4798690e-01 4.5945460e-01
-1.8969634e-01 -1.0663115e+00 -7.9260752e-02 8.7471503e-01
-6.1141682e-01 -3.2776144e-01 -7.0077932e-01 8.0549121e-01
-3.7787754e-02 8.0932391e-01 2.4759404e-01 -3.9693844e-01
-6.4126514e-02 -3.9992547e-01 2.3447181e-01 -8.8466567e-01
-8.2626766e-01 1.3989422e-02 2.2221510e-01 -2.0439266e-01
-3.2399744e-01 -4.2169815e-01 -1.1073011e+00 -9.2314577e-01
-6.0947371e-01 2.8166631e-02 1.2723097e-01 6.6781533e-01
6.4197129e-01 8.1798798e-01 4.6426672e-01 -4.3728396e-01
-1.2010595e-01 -1.1972651e+00 -8.0814868e-01 2.4185498e-01
2.2119084e-01 -2.8768235e-01 -3.5690644e-01 -2.7684301e-01] | [6.942113399505615, 6.085938453674316] |
99804930-fe8d-4843-95d7-943ef9662cbe | high-impedance-fault-detection-through-quasi | 2212.09989 | null | https://arxiv.org/abs/2212.09989v1 | https://arxiv.org/pdf/2212.09989v1.pdf | High Impedance Fault Detection Through Quasi-Static State Estimation: A Parameter Error Modeling Approach | This paper presents a model for detecting high-impedance faults (HIFs) using parameter error modeling and a two-step per-phase weighted least squares state estimation (SE) process. The proposed scheme leverages the use of phasor measurement units and synthetic measurements to identify per-phase power flow and injection measurements which indicate a parameter error through $\chi^2$ Hypothesis Testing applied to the composed measurement error (CME). Although current and voltage waveforms are commonly analyzed for high-impedance fault detection, wide-area power flow and injection measurements, which are already inherent to the SE process, also show promise for real-world high-impedance fault detection applications. The error distributions after detection share the measurement function error spread observed in proven parameter error diagnostics and can be applied to HIF identification. Further, this error spread related to the HIF will be clearly discerned from measurement error. Case studies are performed on the 33-Bus Distribution System in Simulink. | ['Newton G. Bretas', 'Sean Meyn', 'Arturo Bretas', 'Austin Cooper'] | 2022-12-20 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 8.65647718e-02 -1.97210997e-01 -9.12975818e-02 3.42156813e-02
-1.01461709e+00 -5.08598506e-01 2.42123023e-01 3.26895982e-01
4.69107032e-01 1.10897553e+00 -2.36713231e-01 -4.33587670e-01
-9.10987794e-01 -4.27006632e-01 -3.00920129e-01 -9.41092193e-01
-8.29882503e-01 4.55530196e-01 -1.21974573e-01 -2.90363915e-02
3.20333093e-01 7.71406233e-01 -9.45552707e-01 -9.68642235e-02
9.55191791e-01 8.52645814e-01 -4.36195642e-01 4.98585016e-01
7.57495642e-01 8.44903052e-01 -1.37220407e+00 6.84523821e-01
1.94986686e-01 -4.83886480e-01 -7.15802312e-01 1.66976396e-02
-2.62991041e-01 -5.92416465e-01 -3.05953741e-01 1.07646358e+00
6.79764330e-01 7.32248323e-03 7.73145914e-01 -1.89952350e+00
5.79805970e-01 4.83555615e-01 -5.48110902e-01 6.56487525e-01
9.44694102e-01 2.52330631e-01 5.99499106e-01 -6.36645794e-01
7.76220784e-02 7.44050324e-01 8.24927986e-01 -4.76803780e-01
-1.54485261e+00 -5.65975189e-01 -5.35618663e-01 2.46005043e-01
-1.53806281e+00 4.20268551e-02 9.87211466e-01 -6.56833172e-01
1.47799432e+00 4.87551957e-01 6.27822340e-01 4.31054831e-01
6.37149155e-01 5.86804330e-01 1.33165872e+00 -2.01854289e-01
4.77467895e-01 -5.43435216e-02 3.27870190e-01 3.29657584e-01
6.75572038e-01 3.37415695e-01 -2.90451199e-01 -3.64794075e-01
7.15307593e-01 -4.32689279e-01 -9.99116600e-01 -2.00370386e-01
-7.40490854e-01 5.71124077e-01 2.41853446e-01 4.29069638e-01
-3.75589013e-01 4.10557128e-02 3.91988993e-01 6.09931290e-01
2.04963252e-01 8.53345335e-01 -7.16864228e-01 -4.05698776e-01
-1.28587449e+00 -5.16255237e-02 1.04948151e+00 6.31598413e-01
6.18052363e-01 1.00866246e+00 5.20003922e-02 2.57156283e-01
5.69260180e-01 8.28269124e-01 3.41321290e-01 -5.60795426e-01
2.41445705e-01 6.28911138e-01 2.97164291e-01 -7.47042656e-01
-7.42489278e-01 -8.47411156e-01 -6.19846582e-01 5.41849732e-01
3.01747143e-01 -6.53577685e-01 -4.69274640e-01 1.07192385e+00
-8.57562199e-03 3.65746379e-01 -2.56741289e-02 6.92717850e-01
-2.51800358e-01 5.20106196e-01 -5.10107219e-01 -6.56175911e-01
1.23507893e+00 -2.02560022e-01 -9.55616593e-01 -1.12385906e-01
8.90599072e-01 -5.09634614e-01 2.17112422e-01 5.62694073e-01
-1.03642213e+00 -9.66819003e-02 -1.48644638e+00 1.03503716e+00
4.82659973e-02 7.76598603e-02 1.37694702e-01 8.56239438e-01
-4.54405665e-01 8.50896955e-01 -8.51754308e-01 -5.14663942e-02
2.09896579e-01 3.49499136e-02 -2.24317685e-01 2.30080113e-01
-1.48708558e+00 1.51880121e+00 2.69957393e-01 5.24315774e-01
-8.02399695e-01 -1.33035326e+00 -9.36428010e-01 8.69643092e-02
1.18132062e-01 -1.08508812e-02 9.95903552e-01 -4.08472329e-01
-1.25500476e+00 -2.28482395e-01 1.21407308e-01 -5.65370798e-01
2.75670499e-01 -2.46467248e-01 -9.22305942e-01 5.84337354e-01
-4.58785985e-03 -8.06881368e-01 8.38612318e-01 -1.04757559e+00
-5.02109289e-01 -1.56436507e-02 -7.17338741e-01 -2.19607100e-01
3.27790856e-01 -6.73689619e-02 9.11744773e-01 -4.28109169e-01
4.08959508e-01 -2.63544321e-01 -2.63866514e-01 -6.65790915e-01
-5.63736618e-01 3.28012824e-01 9.54690754e-01 -1.18359447e+00
1.58295178e+00 -1.60232937e+00 -2.49354154e-01 1.03380656e+00
-3.22645903e-01 3.09226848e-02 4.81341302e-01 1.19074416e+00
-6.14542365e-01 -3.37316573e-01 -3.24666560e-01 6.30584717e-01
4.10878479e-01 -6.15559109e-02 -2.15035707e-01 1.30323470e+00
4.71508682e-01 4.96546388e-01 -7.87479997e-01 5.30433476e-01
5.62413514e-01 8.49271566e-02 -1.10957390e-02 2.22640827e-01
4.91151631e-01 3.09637725e-01 -1.15561843e-01 4.46861625e-01
6.75993145e-01 -1.05405591e-01 -1.46323713e-02 -7.58265972e-01
1.42924398e-01 1.55202433e-01 -1.91997182e+00 1.02056217e+00
-2.96795070e-01 7.57009506e-01 1.46242782e-01 -1.22793090e+00
8.86660695e-01 5.76729476e-01 7.03807771e-01 -5.44940829e-01
3.08236722e-02 4.68866855e-01 2.91941673e-01 -5.62956154e-01
-1.28888249e-01 -7.62237981e-02 1.55288860e-01 6.20230913e-01
4.38214213e-01 -4.43138391e-01 4.03498054e-01 2.46991571e-02
1.28841436e+00 -1.97933570e-01 5.28566480e-01 -8.41625631e-01
6.32547617e-01 2.38856557e-03 9.35702801e-01 2.22386092e-01
-1.01496965e-01 2.41498753e-01 8.04227531e-01 3.34008247e-01
-7.51597881e-01 -1.09062207e+00 -6.84279442e-01 -5.01749992e-01
5.91322556e-02 -1.76802158e-01 -2.19712645e-01 -5.20028293e-01
4.07251120e-01 1.19485414e+00 -1.67277753e-01 -4.98006433e-01
-5.21103859e-01 -1.05541646e+00 4.62661296e-01 6.87833190e-01
1.41111106e-01 -3.27776909e-01 -9.20072258e-01 5.26319802e-01
5.12161434e-01 -7.69329309e-01 1.12487577e-01 8.28356981e-01
-6.72314048e-01 -1.69200730e+00 -4.79390323e-01 -2.31087446e-01
1.02411735e+00 -6.04072511e-01 9.20989335e-01 6.50221854e-02
-8.38965774e-01 5.66471994e-01 -1.41171515e-01 7.09774047e-02
-2.95178950e-01 -7.94499934e-01 3.03207070e-01 -2.31000081e-01
-5.51706962e-02 -4.14118320e-01 -4.49474424e-01 5.30310214e-01
-4.94177788e-01 -6.57279909e-01 5.33403158e-01 1.15916169e+00
-1.58386901e-01 9.35337782e-01 1.08180594e+00 -7.14871526e-01
7.04749525e-01 -7.21731722e-01 -9.91713285e-01 8.65026265e-02
-1.20908999e+00 -2.47020975e-01 6.02643609e-01 -1.07267782e-01
-1.03063905e+00 -4.24105287e-01 -4.40130048e-02 -1.24945685e-01
-3.19135249e-01 8.72520983e-01 -3.45415533e-01 -3.08180600e-01
3.89523745e-01 -2.26702675e-01 9.25651342e-02 -3.19239259e-01
-2.96747476e-01 5.35376072e-01 3.04930091e-01 -2.38792002e-01
1.25033474e+00 -3.75663191e-02 4.88578767e-01 -7.23609507e-01
-2.72918418e-02 -6.64944828e-01 -4.25090164e-01 -2.65264809e-01
8.98350626e-02 -1.03070080e+00 -8.05758119e-01 7.70291328e-01
-5.57427585e-01 -4.01076794e-01 -4.41959679e-01 8.01205814e-01
-4.27214205e-01 5.64088404e-01 -8.44476104e-01 -1.07651734e+00
-1.53753012e-01 -1.29375148e+00 6.89936161e-01 1.30208150e-01
-4.78092462e-01 -1.51478541e+00 -1.11509962e-02 -2.03654110e-01
3.42432052e-01 3.82030159e-01 9.69403863e-01 -8.19899082e-01
-2.15439454e-01 -6.28299713e-01 2.71691114e-01 6.60815001e-01
1.75908580e-01 6.56692609e-02 -7.52683580e-01 -1.08286285e+00
4.00989354e-01 9.07837749e-02 -4.88802493e-01 1.55467913e-01
2.06283510e-01 -2.61106461e-01 -3.46725792e-01 4.42971498e-01
2.02467918e+00 4.50300932e-01 6.29243314e-01 4.22230840e-01
1.69117391e-01 2.60520488e-01 7.08451688e-01 5.99676788e-01
-2.21078157e-01 3.98877949e-01 7.50029981e-02 -2.65190512e-01
2.66751140e-01 1.15216397e-01 5.03321767e-01 7.14233398e-01
5.52830696e-01 -1.47034854e-01 -1.13586271e+00 4.57160830e-01
-1.26508141e+00 -7.06959128e-01 -7.15477884e-01 2.19124174e+00
5.42488635e-01 2.35618711e-01 -4.30871010e-01 1.25034523e+00
4.04919326e-01 -2.81392515e-01 -3.40933204e-01 -3.22119921e-01
-3.05515468e-01 5.40263653e-01 8.60504627e-01 5.98331749e-01
-6.67069018e-01 -4.23025429e-01 6.66267776e+00 8.32553506e-01
-8.15736353e-01 -3.30859691e-01 2.07881749e-01 4.34352070e-01
-4.33266200e-02 1.92243010e-01 -5.47227323e-01 5.42424560e-01
1.48022687e+00 -6.36988640e-01 9.09035653e-02 2.65611440e-01
5.88419139e-01 -7.41684616e-01 -9.97696936e-01 7.42754042e-01
2.77059432e-02 -7.40788937e-01 -5.48560739e-01 -1.89987049e-01
8.65035772e-01 -4.66275036e-01 -5.74658751e-01 1.37620091e-01
-8.66346434e-02 -8.46484482e-01 2.13925257e-01 4.35655832e-01
5.54522872e-01 -8.06467295e-01 1.20286012e+00 2.92124093e-01
-8.74475598e-01 -5.68585157e-01 1.19829565e-01 -1.14776641e-01
9.44034874e-01 1.15248430e+00 -1.10539126e+00 1.15614915e+00
5.53409636e-01 6.05616331e-01 -3.39183420e-01 1.50884807e+00
-7.97353625e-01 1.21451962e+00 -3.98020953e-01 4.40091431e-01
-3.08425307e-01 -1.88440606e-01 6.82222068e-01 8.85782838e-01
3.59288514e-01 -1.60159722e-01 1.09878086e-01 6.91529930e-01
7.44416416e-01 -3.60906661e-01 -1.52747855e-01 2.73295432e-01
7.90773153e-01 1.25409889e+00 -5.78809857e-01 -2.14179859e-01
-3.41085076e-01 4.68055844e-01 -6.84741557e-01 8.18334699e-01
-6.39848769e-01 -9.34277654e-01 7.18825758e-01 -1.09084137e-01
-1.88982710e-01 1.22331709e-01 -3.68453383e-01 -8.04172814e-01
-7.22887665e-02 -9.36900973e-01 1.70256302e-01 -9.05915797e-01
-1.44187462e+00 1.44337192e-01 3.96122664e-01 -1.54422045e+00
-1.05324948e+00 -4.44581509e-01 -9.24972713e-01 1.51664424e+00
-1.52140832e+00 -3.94422740e-01 2.73875296e-02 2.03657910e-01
3.28100860e-01 -6.98612556e-02 7.06798971e-01 4.72477823e-01
-1.02239537e+00 4.81881261e-01 3.81213129e-01 2.22448006e-01
1.28838003e-01 -1.36746693e+00 -1.23396918e-01 1.29299307e+00
-4.55569923e-01 2.64461547e-01 1.15971291e+00 -8.48153472e-01
-1.73072839e+00 -6.08605027e-01 3.87139320e-01 1.81360260e-01
1.11591458e+00 -1.49617165e-01 -1.20949876e+00 6.46721900e-01
4.88069206e-01 -2.49024190e-04 3.49806428e-01 -5.99917710e-01
3.45343471e-01 -8.73148814e-02 -1.64443064e+00 1.32008061e-01
-2.73514483e-02 -7.62099504e-01 -8.89119864e-01 1.03450656e-01
-3.05167586e-01 -3.44602108e-01 -1.76625991e+00 9.33171630e-01
-3.91226299e-02 -3.60362947e-01 7.87844777e-01 -2.36503974e-01
-4.25464481e-01 -7.46313512e-01 1.03554919e-01 -2.09015489e+00
-1.70213982e-01 -9.94741261e-01 -4.22628134e-01 1.25602639e+00
4.02217776e-01 -1.24111831e+00 4.68008608e-01 1.28259033e-01
-3.05574741e-02 -4.48107302e-01 -1.20475030e+00 -9.11583841e-01
7.98011106e-03 -2.86846906e-01 5.30173481e-01 1.11739242e+00
8.19967687e-01 9.63528454e-02 1.84524000e-01 1.01854312e+00
9.58432555e-01 -2.42389724e-01 2.93295592e-01 -9.39980090e-01
-2.62055367e-01 -2.74812490e-01 -7.11150646e-01 -2.83077538e-01
-6.96571842e-02 -2.80405343e-01 2.27409407e-01 -1.51850879e+00
-6.20004654e-01 -2.19129726e-01 -1.77865535e-01 1.57341525e-01
-3.02523226e-01 -5.50625697e-02 -2.73758382e-01 -4.17075381e-02
3.36518437e-01 3.68537456e-01 3.08362067e-01 -4.37708944e-02
4.99152869e-01 1.74466401e-01 5.14564872e-01 4.68834549e-01
8.49692881e-01 -1.81608915e-01 -5.36887825e-01 2.40657121e-01
1.54768556e-01 7.98307121e-01 2.07432106e-01 -1.50560796e+00
3.68997842e-01 2.53533423e-01 5.16963780e-01 -8.24055076e-01
-1.84031218e-01 -1.43637168e+00 6.17223561e-01 9.00203466e-01
1.97264120e-01 4.36606377e-01 2.76823282e-01 3.43557000e-01
-4.73060578e-01 -5.42453289e-01 4.63010013e-01 4.84651804e-01
-6.83795691e-01 -3.44413310e-01 -8.56998682e-01 -2.10874096e-01
1.22789156e+00 -2.24301353e-01 -7.37329900e-01 -2.37337679e-01
-4.93610352e-01 5.37808239e-01 1.56161249e-01 -1.34913862e-01
5.39231122e-01 -9.98449326e-01 -7.58209527e-01 6.42663896e-01
-1.86207622e-01 -5.39891243e-01 3.23330849e-01 1.32882988e+00
-4.73806679e-01 3.80505264e-01 -6.02456927e-03 -8.53968501e-01
-6.45112634e-01 6.15265220e-02 7.55413651e-01 -6.61108494e-02
-4.36982125e-01 4.01265502e-01 -8.30206692e-01 5.36119714e-02
-1.38072431e-01 -1.40802458e-01 -1.34878827e-03 2.95490682e-01
3.95935774e-01 1.08227909e+00 6.02437139e-01 -3.55485499e-01
-3.29051822e-01 3.35301757e-01 6.53558969e-01 3.47439162e-02
1.16374373e+00 -2.17036858e-01 -6.28094096e-03 7.37281382e-01
1.31621051e+00 -2.53047377e-01 -1.36522532e+00 4.22731161e-01
4.11517680e-01 -3.39683414e-01 1.22927316e-01 -1.34588516e+00
-1.24689889e+00 4.56025422e-01 6.87082887e-01 5.78844428e-01
1.08117914e+00 -8.21553946e-01 1.85484260e-01 -6.10504411e-02
6.70031428e-01 -9.80866253e-01 -4.70284998e-01 7.35842586e-02
5.14047980e-01 -5.59788108e-01 3.09261113e-01 -3.64020556e-01
-5.62982112e-02 1.12337315e+00 1.92529425e-01 -5.13578415e-01
9.14773166e-01 1.14162481e+00 8.25780481e-02 -9.86878574e-02
-7.62208164e-01 3.61182958e-01 6.32110015e-02 6.15371346e-01
2.17421934e-01 -2.40379274e-01 -5.35959899e-01 3.06573510e-01
1.41473264e-01 -2.86031365e-01 8.84881556e-01 1.34330285e+00
-2.72145867e-01 -6.42185807e-01 -7.60448217e-01 7.46799529e-01
-3.64107519e-01 4.07645047e-01 6.23999000e-01 9.68115151e-01
-5.08881629e-01 1.33292341e+00 1.16519593e-01 2.04765983e-03
7.68619657e-01 1.84627131e-01 2.57540882e-01 -2.31524214e-01
-9.49406266e-01 2.37585958e-02 3.39190632e-01 -7.09472656e-01
3.10922805e-02 -8.21063101e-01 -1.42785752e+00 1.37984380e-01
-9.82622087e-01 8.61391544e-01 7.76329935e-01 1.06736875e+00
-2.38731310e-01 1.13383782e+00 1.03146064e+00 -5.85404038e-01
-1.00727594e+00 -1.21423161e+00 -1.26924968e+00 2.95455933e-01
3.50432754e-01 -8.02055538e-01 -1.31673694e+00 -5.15042126e-01] | [6.053831100463867, 2.5664336681365967] |
40972be2-cff9-4d27-bef7-816e363b23e0 | context-sensitive-temporal-feature-learning-1 | 2204.03270 | null | https://arxiv.org/abs/2204.03270v2 | https://arxiv.org/pdf/2204.03270v2.pdf | Context-Sensitive Temporal Feature Learning for Gait Recognition | Although gait recognition has drawn increasing research attention recently, it remains challenging to learn discriminative temporal representation, since the silhouette differences are quite subtle in spatial domain. Inspired by the observation that human can distinguish gaits of different subjects by adaptively focusing on temporal clips with different time scales, we propose a context-sensitive temporal feature learning (CSTL) network for gait recognition. CSTL produces temporal features in three scales, and adaptively aggregates them according to the contextual information from local and global perspectives. Specifically, CSTL contains an adaptive temporal aggregation module that subsequently performs local relation modeling and global relation modeling to fuse the multi-scale features. Besides, in order to remedy the spatial feature corruption caused by temporal operations, CSTL incorporates a salient spatial feature learning (SSFL) module to select groups of discriminative spatial features. Particularly, we utilize transformers to implement the global relation modeling and the SSFL module. To the best of our knowledge, this is the first work that adopts transformer in gait recognition. Extensive experiments conducted on three datasets demonstrate the state-of-the-art performance. Concretely, we achieve rank-1 accuracies of 98.7%, 96.2% and 88.7% under normal-walking, bag-carrying and coat-wearing conditions on CASIA-B, 97.5% on OU-MVLP and 50.6% on GREW. | ['Bin Feng', 'Wenyu Liu', 'Botao He', 'Bo Yang', 'Hao Wang', 'Xinggang Wang', 'Duowang Zhu', 'Xiaohu Huang'] | 2022-04-07 | context-sensitive-temporal-feature-learning | http://openaccess.thecvf.com//content/ICCV2021/html/Huang_Context-Sensitive_Temporal_Feature_Learning_for_Gait_Recognition_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Huang_Context-Sensitive_Temporal_Feature_Learning_for_Gait_Recognition_ICCV_2021_paper.pdf | iccv-2021-1 | ['multiview-gait-recognition'] | ['computer-vision'] | [-7.84118846e-02 -7.07422078e-01 -3.73233616e-01 -3.07912618e-01
-5.83578646e-01 -7.26759881e-02 1.75475046e-01 -8.16627219e-02
-2.47680962e-01 5.78730822e-01 2.51019210e-01 3.60600263e-01
-3.45568806e-01 -6.38762772e-01 -1.84992492e-01 -8.78794134e-01
-6.94448590e-01 -3.20571624e-02 5.40155292e-01 -2.34295532e-01
1.43407613e-01 4.18040961e-01 -1.67207217e+00 2.88935214e-01
9.69851255e-01 1.25049686e+00 8.22353177e-03 2.82890707e-01
1.88039929e-01 5.34671247e-01 -5.24055541e-01 1.88170895e-02
1.22086853e-02 -3.30329657e-01 -4.14337486e-01 2.00831920e-01
1.79834574e-01 -2.38174498e-01 -4.55921501e-01 6.75068974e-01
5.98496556e-01 2.92653561e-01 5.47587156e-01 -1.28023553e+00
-5.45397818e-01 9.98925418e-02 -8.73482108e-01 6.34235561e-01
5.72052896e-01 4.20172215e-01 8.63685966e-01 -7.94358134e-01
2.66413599e-01 1.11024272e+00 6.83621705e-01 2.14725390e-01
-9.32555616e-01 -7.22100198e-01 3.84257466e-01 6.76066577e-01
-1.33132863e+00 -2.25192502e-01 1.11922145e+00 -4.32242990e-01
7.14065731e-01 2.29679689e-01 7.56311178e-01 1.07767439e+00
5.00523627e-01 8.77531111e-01 1.02382708e+00 4.08173278e-02
4.82908227e-02 -6.65503025e-01 2.28210032e-01 7.08220005e-01
1.07671633e-01 9.97324809e-02 -8.17256451e-01 1.21497944e-01
7.63675511e-01 2.63767183e-01 -1.73789263e-01 -2.00454906e-01
-1.12496114e+00 3.44646245e-01 3.78254890e-01 4.68739986e-01
-4.37427074e-01 -5.20128477e-03 6.75617635e-01 2.13985234e-01
2.62758523e-01 -2.01983407e-01 -3.14392626e-01 -3.48361999e-01
-9.84163046e-01 5.24342917e-02 2.90060818e-01 6.54245853e-01
5.55353224e-01 2.51367062e-01 -2.97336608e-01 9.50508118e-01
1.60097182e-01 5.47317088e-01 8.80771041e-01 -6.34391069e-01
5.14577508e-01 5.44758737e-01 -1.04315512e-01 -1.46217585e+00
-5.15088499e-01 -5.89947045e-01 -1.05950153e+00 -1.95912749e-01
3.35332721e-01 2.87592649e-01 -8.99124682e-01 1.62576067e+00
2.03248426e-01 5.90916932e-01 -3.41039211e-01 1.13781035e+00
5.16259372e-01 4.92561162e-01 2.24719331e-01 -3.95100415e-01
1.43597138e+00 -6.30861998e-01 -7.13113070e-01 -1.88392162e-01
5.87963760e-01 -5.02505302e-01 1.07548952e+00 3.78067166e-01
-8.08148205e-01 -8.68478119e-01 -1.24292922e+00 1.63122013e-01
-2.30959058e-01 5.17446995e-01 5.50187290e-01 4.21733081e-01
-5.94828248e-01 6.60118580e-01 -1.19254827e+00 -5.21951079e-01
2.45014563e-01 2.43437424e-01 -3.91062051e-01 -1.59649048e-02
-1.35264242e+00 6.07565403e-01 2.76346922e-01 4.67521369e-01
-5.84725499e-01 -4.16753471e-01 -8.11229289e-01 -2.06006557e-01
3.75006318e-01 -4.28474188e-01 7.36269176e-01 -4.41841543e-01
-1.31340492e+00 5.89681566e-01 -2.49775082e-01 -4.56914276e-01
5.64341307e-01 -3.23787987e-01 -9.05003309e-01 3.81980091e-01
4.22493279e-01 3.42548639e-02 8.46320510e-01 -7.35167384e-01
-5.89344561e-01 -5.35129786e-01 -3.61855417e-01 1.86600834e-02
-4.41888779e-01 -1.73392534e-01 -5.55626631e-01 -1.02619350e+00
5.30638635e-01 -7.23454595e-01 -9.36599597e-02 4.44294289e-02
-1.27721250e-01 -1.93806663e-01 1.09281886e+00 -9.38982129e-01
1.63220501e+00 -2.27658343e+00 1.35379970e-01 2.76656091e-01
-1.69793200e-02 3.06091368e-01 -2.86727063e-02 3.17159981e-01
6.88098976e-03 -2.12421194e-01 -2.51806200e-01 -2.07231805e-01
-4.98503745e-02 3.24682206e-01 -1.57915413e-01 5.09523749e-01
1.74819529e-01 6.33418143e-01 -8.74415457e-01 -7.11944938e-01
4.85438049e-01 3.19060296e-01 -4.54778254e-01 2.29734299e-03
2.88231045e-01 4.39557225e-01 -7.63668597e-01 1.05809021e+00
5.97395122e-01 8.73208046e-03 2.23350450e-02 -5.55155337e-01
-1.43284932e-01 -5.29705137e-02 -1.29687166e+00 1.71573949e+00
-1.50144339e-01 1.98263720e-01 -2.91625082e-01 -1.24933124e+00
1.06481183e+00 3.09772551e-01 9.90476310e-01 -9.47802961e-01
8.42036381e-02 3.86879295e-01 -2.05172658e-01 -7.95333087e-01
3.35551232e-01 5.39545640e-02 -3.87449026e-01 8.66580606e-02
-2.99804984e-03 5.69241464e-01 4.01059598e-01 -1.29117146e-01
1.10607231e+00 9.14683416e-02 3.60894412e-01 -2.32905194e-01
7.87826180e-01 -4.14266229e-01 1.05974662e+00 3.09864402e-01
-7.57835686e-01 5.85703254e-01 2.52147585e-01 -6.80409253e-01
-5.44687510e-01 -1.32351184e+00 1.15884528e-01 9.93122101e-01
3.52232069e-01 -5.16227543e-01 -2.86385715e-01 -5.48517168e-01
5.57391420e-02 9.82645601e-02 -6.88610137e-01 -5.36915481e-01
-9.84995067e-01 -7.56127834e-01 5.75783372e-01 9.56297398e-01
1.05597162e+00 -8.73912692e-01 -6.97742462e-01 4.31815356e-01
-3.07045311e-01 -1.04650116e+00 -7.79143393e-01 -1.80876195e-01
-9.19632077e-01 -1.06317878e+00 -5.90749919e-01 -8.15266490e-01
2.21843094e-01 2.47511327e-01 6.47659361e-01 6.14875145e-02
-4.57615823e-01 1.80329889e-01 -4.50144142e-01 3.79960179e-01
5.03883839e-01 4.44992371e-02 2.51954675e-01 3.09253186e-01
2.67122120e-01 -1.05507946e+00 -8.22793424e-01 6.88189685e-01
-5.82854092e-01 -8.36567804e-02 5.88986874e-01 9.42086458e-01
6.22846365e-01 1.74144208e-01 5.15204787e-01 3.74486437e-03
3.86618137e-01 -3.74859571e-01 -7.06289262e-02 2.05495998e-01
-2.29533419e-01 -7.12586194e-02 6.56700850e-01 -5.57642221e-01
-8.52730513e-01 -8.72961730e-02 3.16551402e-02 -5.01746356e-01
-8.13237131e-02 5.92069864e-01 -2.81050205e-01 1.31695166e-01
3.58296692e-01 6.10585988e-01 -1.41405120e-01 -5.45854449e-01
-3.68983522e-02 3.55207711e-01 7.84887969e-01 -8.97906005e-01
8.21525574e-01 5.77085495e-01 4.37164754e-02 -7.75778711e-01
-5.46465158e-01 -4.86792326e-01 -7.78112948e-01 -5.18224418e-01
9.10030544e-01 -7.60371983e-01 -6.89085960e-01 7.73607492e-01
-5.80840647e-01 1.02708638e-02 -8.51971656e-02 4.96781707e-01
-4.66028869e-01 5.38698852e-01 -7.14272082e-01 -6.18891776e-01
-2.69737273e-01 -9.96347249e-01 1.09067953e+00 3.35513443e-01
-1.82602257e-01 -6.15369678e-01 -1.39281482e-01 2.23335177e-01
2.85511017e-01 4.79154557e-01 5.68655074e-01 -1.88824520e-01
-3.74764025e-01 -1.73793212e-01 -9.88515466e-02 1.04111589e-01
4.90721941e-01 3.27574275e-02 -6.30195856e-01 -3.82783204e-01
-1.33409828e-01 -1.44439742e-01 9.44005966e-01 3.98276359e-01
1.02064109e+00 -4.63812798e-02 -4.28702414e-01 5.33267081e-01
1.04809344e+00 5.23141205e-01 6.25279546e-01 3.74752045e-01
7.25699246e-01 4.43780184e-01 1.01993644e+00 7.08773077e-01
3.33477587e-01 9.26767647e-01 4.93429825e-02 2.24173784e-01
-9.85820219e-02 -1.61830515e-01 5.41942358e-01 9.62613106e-01
-3.19339514e-01 2.43426844e-01 -8.15159440e-01 5.65539837e-01
-2.03088355e+00 -1.23255706e+00 1.21147625e-01 2.00268459e+00
5.51786602e-01 3.54164809e-01 4.12980825e-01 4.50832009e-01
7.47134984e-01 6.08016610e-01 -5.77342093e-01 9.36096460e-02
-8.41464028e-02 7.10089058e-02 2.64053255e-01 8.78982991e-02
-1.47839022e+00 6.53123319e-01 5.05295229e+00 9.52060878e-01
-1.24581552e+00 -4.09148298e-02 4.74067539e-01 -9.96502340e-02
3.46879035e-01 -2.36536682e-01 -4.53952849e-01 8.67456973e-01
5.54235160e-01 -2.45886520e-01 -2.71466691e-02 6.91902578e-01
5.76446831e-01 4.78984825e-02 -6.80753589e-01 1.11540246e+00
-3.32800522e-02 -8.31473291e-01 8.05093069e-03 -1.49409190e-01
2.79684514e-01 -4.79619771e-01 2.12365136e-01 3.44165087e-01
-2.84898043e-01 -7.57668257e-01 6.11967027e-01 6.56723976e-01
5.14036298e-01 -7.31618702e-01 6.59301341e-01 2.15723347e-02
-2.29133868e+00 -2.22143129e-01 -2.05420610e-02 -1.49929859e-02
2.99595684e-01 6.30277514e-01 -4.63282876e-02 8.83894324e-01
8.99914742e-01 1.32856464e+00 -6.42295778e-01 1.19489610e+00
-1.12006284e-01 5.97807348e-01 -2.56625086e-01 2.18063772e-01
1.36736929e-01 -1.17081173e-01 5.06344378e-01 1.15739977e+00
4.51578051e-01 1.71381027e-01 6.06922507e-01 3.64273131e-01
4.61303443e-01 5.63738234e-02 -2.57936656e-01 1.22218326e-01
4.52230990e-01 8.37648451e-01 -6.01991832e-01 -1.64865568e-01
-2.15662852e-01 1.02666199e+00 -8.62088706e-03 3.16361189e-01
-9.56329226e-01 -3.95420879e-01 7.21276581e-01 4.90236171e-02
3.24410617e-01 -5.18513024e-01 -2.18955711e-01 -1.32104540e+00
5.32773256e-01 -6.33481264e-01 7.88805366e-01 -5.78274488e-01
-1.36612797e+00 3.52236211e-01 9.37636867e-02 -1.82158875e+00
-1.08217500e-01 -3.99801522e-01 -7.23675013e-01 5.73843241e-01
-1.13574994e+00 -1.21305549e+00 -3.70730788e-01 7.20857561e-01
6.75802231e-01 -5.44105582e-02 4.01001930e-01 5.91234386e-01
-8.29077303e-01 7.00773299e-01 -1.95039079e-01 2.77446926e-01
7.29636788e-01 -8.96345377e-01 1.57743126e-01 9.59383845e-01
-3.59597832e-01 7.12875485e-01 6.36427641e-01 -7.27458298e-01
-1.38036430e+00 -1.12014592e+00 6.97475612e-01 1.23300135e-01
7.17234373e-01 4.46708724e-02 -1.10918975e+00 3.73158962e-01
-3.39680225e-01 3.00861061e-01 4.37526017e-01 -4.71978784e-02
-4.25157279e-01 -4.97208565e-01 -1.00457811e+00 5.80452263e-01
1.31825984e+00 -5.48405051e-01 -6.67544723e-01 -1.07034341e-01
2.74214745e-01 -3.41324389e-01 -1.12604380e+00 7.17985928e-01
9.30979133e-01 -8.18771183e-01 1.07051599e+00 -2.12188274e-01
1.48974732e-01 -6.90793693e-01 -3.48717958e-01 -1.01717341e+00
-5.13875663e-01 -2.53797621e-01 -2.39551619e-01 1.16149819e+00
-2.60630816e-01 -6.77807510e-01 8.32201660e-01 1.13070115e-01
-3.16970319e-01 -7.72881567e-01 -1.24996829e+00 -1.12588644e+00
-3.23211014e-01 -5.09921670e-01 4.95227844e-01 8.46043468e-01
5.71018159e-02 -1.07078619e-01 -7.78746963e-01 2.29178458e-01
6.40088499e-01 2.90415794e-01 4.40361202e-01 -1.20736217e+00
-2.30394214e-01 -4.25488383e-01 -9.22712982e-01 -9.49732840e-01
-8.49222094e-02 -5.06247222e-01 -4.43125740e-02 -1.25290525e+00
-1.40823694e-02 -1.93550199e-01 -7.96318412e-01 6.27191305e-01
-1.96537331e-01 2.85880685e-01 4.26816717e-02 2.77929127e-01
-6.51603520e-01 8.20722222e-01 1.29569554e+00 -3.53367835e-01
-1.44716173e-01 -5.23776747e-02 -2.12997168e-01 5.08291543e-01
9.11593556e-01 -1.07041977e-01 -5.65088511e-01 -1.20633923e-01
-4.56202894e-01 6.82762861e-02 4.39894348e-01 -1.55249512e+00
1.90071329e-01 -3.30796450e-01 5.07032752e-01 -8.62872183e-01
4.54592943e-01 -5.96275449e-01 1.86114997e-01 7.02076733e-01
-3.92308123e-02 1.45507067e-01 1.87112451e-01 6.31663799e-01
-4.12141532e-01 5.48707724e-01 7.09175050e-01 8.34493935e-02
-1.23984945e+00 5.69012463e-01 -3.51109713e-01 2.18254700e-02
9.52047527e-01 -6.51179910e-01 -2.16506422e-01 -2.14153472e-02
-9.23260629e-01 3.25332135e-01 2.37691388e-01 5.82937777e-01
8.40297639e-01 -1.72119296e+00 -4.11200285e-01 3.87559265e-01
2.96647727e-01 -4.71023679e-01 7.84055054e-01 1.16526210e+00
-1.76201761e-01 2.43425772e-01 -4.78065014e-01 -7.66564965e-01
-1.26772106e+00 2.93622732e-01 3.69119853e-01 -4.54103202e-01
-7.68429160e-01 4.72325176e-01 2.46090218e-02 2.07181826e-01
3.97538720e-03 -4.13043678e-01 -3.00399989e-01 2.31033787e-01
4.96844471e-01 6.06832325e-01 1.11240521e-01 -8.87986422e-01
-7.30793774e-01 1.02723575e+00 -8.71283785e-02 6.46287426e-02
1.04126215e+00 -2.66258977e-02 1.43145457e-01 5.22955120e-01
1.26877022e+00 -2.57734865e-01 -1.31452584e+00 -1.76437780e-01
1.32549658e-01 -5.59319317e-01 -3.68922174e-01 -5.41132987e-01
-1.24120915e+00 7.95984387e-01 8.53678346e-01 -7.86763728e-02
1.55719995e+00 -3.72702479e-01 1.09606719e+00 1.60781648e-02
7.31083632e-01 -1.16348720e+00 5.00594258e-01 4.58822310e-01
7.93590128e-01 -8.66300046e-01 2.29373276e-02 -4.56924051e-01
-4.13888425e-01 9.17306721e-01 7.51983285e-01 -3.81851554e-01
7.01113582e-01 -5.21027576e-03 -2.53114998e-01 -5.88933751e-02
-6.56099498e-01 -1.62181795e-01 4.69749600e-01 5.04160285e-01
3.25584441e-01 2.17211425e-01 -6.83337688e-01 7.22336292e-01
-3.69868577e-02 1.23260565e-01 -8.33559409e-02 1.28243947e+00
-4.81056839e-01 -1.00979292e+00 -2.65494198e-01 4.40725833e-01
-1.64613307e-01 3.20063740e-01 1.53722599e-01 8.12933028e-01
2.28584230e-01 9.22083199e-01 -9.29770917e-02 -9.99153435e-01
4.97715712e-01 -6.47573322e-02 3.37678254e-01 -1.60381883e-01
-3.31220567e-01 2.08389804e-01 3.20814513e-02 -9.64040160e-01
-4.54116225e-01 -9.48143482e-01 -1.30391610e+00 -1.77141249e-01
1.61037594e-01 1.50523275e-01 -1.46478593e-01 9.56572473e-01
3.52947086e-01 6.31130636e-01 6.04352355e-01 -9.00442839e-01
-2.50201732e-01 -8.73148620e-01 -8.36927950e-01 5.83256483e-01
3.48524630e-01 -1.19409657e+00 -2.72607625e-01 7.12658539e-02] | [14.281294822692871, 1.4163490533828735] |
df68e75e-a546-459a-9fa9-36fad633cb3d | a-priority-map-for-vision-and-language | 2207.11717 | null | https://arxiv.org/abs/2207.11717v4 | https://arxiv.org/pdf/2207.11717v4.pdf | A Priority Map for Vision-and-Language Navigation with Trajectory Plans and Feature-Location Cues | In a busy city street, a pedestrian surrounded by distractions can pick out a single sign if it is relevant to their route. Artificial agents in outdoor Vision-and-Language Navigation (VLN) are also confronted with detecting supervisory signal on environment features and location in inputs. To boost the prominence of relevant features in transformer-based architectures without costly preprocessing and pretraining, we take inspiration from priority maps - a mechanism described in neuropsychological studies. We implement a novel priority map module and pretrain on auxiliary tasks using low-sample datasets with high-level representations of routes and environment-related references to urban features. A hierarchical process of trajectory planning - with subsequent parameterised visual boost filtering on visual inputs and prediction of corresponding textual spans - addresses the core challenges of cross-modal alignment and feature-level localisation. The priority map module is integrated into a feature-location framework that doubles the task completion rates of standalone transformers and attains state-of-the-art performance on the Touchdown benchmark for VLN. Code and data are referenced in Appendix C. | ['Rico Sennrich', 'Leonardo Impett', 'Jason Armitage'] | 2022-07-24 | null | null | null | null | ['trajectory-planning'] | ['robots'] | [ 3.20492685e-01 -4.50403988e-02 2.02377081e-01 -5.67556262e-01
-1.06077564e+00 -5.58085620e-01 9.83890474e-01 2.46509328e-01
-8.94682705e-01 3.08922768e-01 6.39807999e-01 -5.87695897e-01
-3.46000135e-01 -6.29783869e-01 -5.45458078e-01 -2.43823737e-01
-5.15254997e-02 5.87001026e-01 3.92704725e-01 -4.65397954e-01
3.79771113e-01 4.95916337e-01 -1.87299335e+00 6.05136275e-01
4.84169871e-01 8.83956969e-01 6.05532587e-01 7.22964287e-01
1.73516124e-01 6.32867277e-01 -3.08985472e-01 -2.25795135e-01
3.05510700e-01 -4.97054011e-02 -6.85420334e-01 -2.82763779e-01
9.26968992e-01 -1.02046408e-01 -6.64397895e-01 5.92611313e-01
8.33410382e-01 4.44404989e-01 7.67811120e-01 -1.49515855e+00
-4.30878639e-01 3.00001144e-01 -1.13410644e-01 8.24647784e-01
3.95862311e-01 8.86032343e-01 9.88809943e-01 -1.27657819e+00
5.63536823e-01 1.45343709e+00 6.86828315e-01 1.82779551e-01
-1.35899460e+00 -3.21097165e-01 5.70152044e-01 4.85247672e-01
-1.12974274e+00 -9.70676243e-01 2.47755721e-01 -5.93024135e-01
1.74681473e+00 2.90868670e-01 6.92842960e-01 1.36873889e+00
-5.10412417e-02 6.85451984e-01 9.70205605e-01 -1.22614168e-02
6.72431067e-02 5.94003759e-02 -1.57454297e-01 6.59672618e-01
-1.22956365e-01 6.16020560e-01 -9.38328028e-01 3.64455909e-01
5.53907156e-01 -2.65795112e-01 -1.12414114e-01 -3.83074254e-01
-1.75415397e+00 5.95611989e-01 8.66623461e-01 1.26532856e-02
-5.16290009e-01 1.37080908e-01 5.15029848e-01 1.92969233e-01
-8.01364183e-02 3.46540689e-01 -2.99173117e-01 -2.35531598e-01
-1.12265456e+00 3.12178582e-01 1.39199376e-01 1.13547623e+00
6.10445201e-01 7.56798163e-02 -7.48793483e-01 7.43927300e-01
4.09353763e-01 7.78160214e-01 3.51267636e-01 -8.61956060e-01
1.03503418e+00 5.06546259e-01 1.01851746e-01 -8.57487023e-01
-1.01852906e+00 -7.14456201e-01 -3.74927223e-01 4.80364859e-01
6.32302403e-01 1.76565617e-01 -1.03052962e+00 1.79594326e+00
2.59800404e-01 -3.79492909e-01 -2.31050223e-01 1.13068151e+00
7.73940682e-01 2.27484152e-01 4.09771770e-01 5.44512868e-01
1.56846881e+00 -1.03410685e+00 -2.20666379e-01 -9.00543571e-01
5.42323470e-01 -3.16851199e-01 1.41407669e+00 1.12647258e-01
-9.35406148e-01 -8.66274357e-01 -1.01448226e+00 -6.96082532e-01
-9.11508203e-01 8.56969506e-02 6.44779861e-01 3.81114900e-01
-1.52445185e+00 2.43202955e-01 -5.59256911e-01 -6.62518919e-01
6.05171680e-01 4.52613771e-01 -7.48893797e-01 1.07957833e-01
-1.11817777e+00 1.51361108e+00 2.70700961e-01 2.65883118e-01
-1.14396691e+00 -6.99082315e-01 -1.12668681e+00 -8.91190544e-02
1.54985664e-02 -9.93570030e-01 1.23935175e+00 -5.82614124e-01
-1.06606567e+00 1.07749379e+00 -2.79560387e-01 -4.16712523e-01
6.19272292e-01 -8.81904364e-02 -5.16929388e-01 1.89132430e-02
5.98156452e-01 1.32279742e+00 8.14780295e-01 -7.64177322e-01
-1.09841025e+00 -3.95507753e-01 -3.46699841e-02 5.13078213e-01
2.68930316e-01 -5.82308620e-02 -3.30639273e-01 -2.36464903e-01
1.03226207e-01 -5.79989910e-01 -3.51306558e-01 8.92997459e-02
-3.01659167e-01 -4.94319536e-02 4.19111252e-01 -7.20687091e-01
8.55337620e-01 -2.05161119e+00 5.29581755e-02 3.45402747e-01
4.34751920e-02 -8.85782316e-02 -5.63164651e-01 3.85366768e-01
-7.62743354e-02 -2.67942518e-01 -6.68782592e-02 -4.65133309e-01
4.98242378e-01 -1.17313080e-01 -4.07373786e-01 5.13241172e-01
3.72635454e-01 1.31414044e+00 -1.04046190e+00 -2.42871702e-01
5.41656554e-01 5.23235917e-01 -6.37617528e-01 -2.75618970e-01
9.35796183e-03 5.07565737e-01 1.54819682e-01 6.14052534e-01
3.20204586e-01 1.49148673e-01 -3.89749199e-01 -1.01175614e-01
-5.07867813e-01 8.30882013e-01 -1.02822125e+00 1.94901800e+00
-5.97461343e-01 1.11286581e+00 -6.03870414e-02 -6.67159677e-01
6.00882471e-01 -1.82144538e-01 -2.82302946e-01 -1.60692477e+00
-5.80686517e-02 8.93368497e-02 1.49720564e-01 -4.54020947e-01
7.62413621e-01 4.14705426e-01 -2.72863477e-01 8.11904892e-02
1.88616440e-01 1.22349903e-01 1.70308664e-01 7.44466931e-02
1.24026823e+00 3.77465099e-01 1.82025164e-01 -3.14853400e-01
4.21947420e-01 1.59123629e-01 -3.82645167e-02 1.00078762e+00
-3.03518146e-01 4.79404688e-01 3.90275091e-01 -4.73430276e-01
-1.13598585e+00 -1.17574501e+00 1.24055021e-01 1.59408581e+00
-2.56952494e-01 -5.01198173e-01 -4.78051305e-01 -3.75361323e-01
-2.43303813e-02 1.02231312e+00 -8.06821525e-01 -1.56634212e-01
-4.96417731e-01 -2.48968691e-01 5.62811255e-01 7.41311014e-01
3.42859298e-01 -1.31919599e+00 -1.35645139e+00 2.04176828e-01
-3.28194886e-01 -1.08987069e+00 -2.91338563e-01 5.57902634e-01
-2.47379705e-01 -9.40700054e-01 -4.10429388e-01 -7.30068147e-01
5.53560317e-01 3.24451327e-01 1.23736310e+00 -3.14389467e-01
-3.32665116e-01 6.09264851e-01 5.08024171e-02 -4.03692305e-01
7.92323425e-02 3.22081029e-01 -6.82742968e-02 -2.92279005e-01
3.85759652e-01 -5.05320847e-01 -8.00539672e-01 3.79376352e-01
-1.40115097e-01 3.00538003e-01 7.87659109e-01 7.22589612e-01
5.55647910e-01 -5.25613248e-01 3.35912824e-01 -1.75355420e-01
4.13045794e-01 -4.16658312e-01 -8.46275389e-01 -6.50579631e-02
-3.10159355e-01 1.12982161e-01 2.77623624e-01 -1.73134357e-01
-6.18929088e-01 2.42607459e-01 -1.51200175e-01 -3.20229977e-02
-4.04107571e-01 2.90397614e-01 -3.73507589e-01 1.18427925e-01
9.21176791e-01 3.63216013e-01 -2.42931351e-01 -5.89763336e-02
7.66554475e-01 2.83529192e-01 9.57676649e-01 -2.87367851e-01
7.30087817e-01 4.14341152e-01 1.80687845e-01 -6.17323637e-01
-6.43069208e-01 -5.50115585e-01 -9.01533842e-01 -3.36585224e-01
7.09368408e-01 -1.11637831e+00 -9.06382859e-01 1.75832808e-01
-1.12972796e+00 -8.34881902e-01 -3.30800235e-01 3.77650142e-01
-9.34339821e-01 -3.84107202e-01 1.71786472e-01 -5.14542758e-01
2.19322443e-02 -1.18844604e+00 1.21860886e+00 5.59441261e-02
-3.84378731e-01 -5.77476025e-01 -1.10619396e-01 3.51898419e-03
7.21911192e-01 -1.32878509e-03 6.80483162e-01 -6.99541211e-01
-6.85742795e-01 -5.46541512e-02 -5.93251407e-01 -5.55116117e-01
-4.98842061e-01 -7.23477006e-01 -1.27932954e+00 3.03965732e-02
-5.14389575e-01 -1.54504195e-01 1.18849254e+00 4.25204903e-01
7.02590466e-01 -1.95017502e-01 -5.59527159e-01 8.37288499e-01
1.09419513e+00 -1.39979571e-01 5.08030474e-01 9.44140732e-01
5.52619219e-01 9.64288116e-01 5.37134588e-01 3.14323872e-01
9.65558410e-01 9.04644430e-01 7.28219628e-01 -1.62724987e-01
-4.08246100e-01 -4.43015307e-01 3.76787066e-01 -3.06563288e-01
2.57610679e-01 -4.13471237e-02 -1.22483563e+00 9.52333927e-01
-1.89653647e+00 -1.35464740e+00 -1.96088880e-01 2.12823224e+00
5.04649937e-01 3.87269586e-01 2.76021123e-01 5.24085527e-03
3.75632226e-01 1.06460400e-01 -6.12152219e-01 -3.36728513e-01
-1.70821726e-01 -3.96083854e-02 7.73397565e-01 6.58789635e-01
-1.33472967e+00 1.18123627e+00 6.14559078e+00 4.81267154e-01
-8.89393985e-01 2.06489772e-01 3.68719220e-01 -4.97216344e-01
-8.25246200e-02 -2.60621309e-01 -1.08391881e+00 2.70956814e-01
1.28956330e+00 2.87918717e-01 7.24767387e-01 8.20462167e-01
4.73800838e-01 -3.48337471e-01 -1.58387458e+00 1.17705047e+00
2.25697365e-02 -1.20878291e+00 -1.09225795e-01 8.67751464e-02
4.90384810e-02 6.14578903e-01 6.12419069e-01 6.55558467e-01
3.01187932e-01 -1.47584176e+00 1.43442214e+00 8.16746652e-01
8.16623747e-01 -6.99507773e-01 2.10220531e-01 2.21710995e-01
-1.30949306e+00 -3.92614216e-01 -2.81859279e-01 -1.64225399e-01
2.96633989e-01 4.91658524e-02 -1.04632223e+00 1.24854743e-01
6.87727153e-01 5.55248618e-01 -1.05799353e+00 1.42140377e+00
-4.58696812e-01 7.22796023e-02 -3.34210664e-01 7.00730756e-02
5.33309579e-01 3.22007895e-01 5.50436378e-01 1.61297536e+00
3.19520295e-01 -2.30262130e-01 1.83443483e-02 8.35198402e-01
5.03171802e-01 -7.98491016e-02 -8.64567697e-01 5.75700164e-01
5.88999987e-01 1.24180210e+00 -7.14442432e-01 -1.02788426e-01
-3.68963093e-01 9.16646302e-01 4.22847807e-01 3.65012467e-01
-8.05388570e-01 -3.90567511e-01 8.32310915e-01 3.16382080e-01
6.65254235e-01 -3.67666304e-01 -4.47587758e-01 -3.67745042e-01
5.17922975e-02 -6.80309951e-01 1.41991660e-01 -1.10507607e+00
-7.76717365e-01 6.59970522e-01 -1.89541504e-01 -1.23336625e+00
-3.26471955e-01 -8.41975510e-01 -3.17528307e-01 1.27055097e+00
-1.75049186e+00 -1.58940959e+00 -5.07268310e-01 9.29940760e-01
6.31747603e-01 -1.43512115e-01 8.87141824e-01 2.49580875e-01
-1.52907386e-01 5.94347835e-01 -3.97052526e-01 1.62700027e-01
6.22780740e-01 -1.16781223e+00 1.00174999e+00 1.04375219e+00
1.21879749e-01 5.52080214e-01 6.07112110e-01 -5.28330743e-01
-1.08285391e+00 -1.22459733e+00 1.13980424e+00 -1.15101659e+00
5.78293085e-01 -7.33350694e-01 -2.93865740e-01 7.33682454e-01
1.73103675e-01 6.73724115e-02 3.68804455e-01 2.87231088e-01
-5.75049579e-01 -4.20405753e-02 -9.32860076e-01 8.05989742e-01
1.68913710e+00 -8.92477036e-01 -7.19863415e-01 3.88970822e-01
5.25352240e-01 -4.63137478e-01 -1.11484051e-01 -1.07236192e-01
6.23605967e-01 -1.08354497e+00 1.20307457e+00 -5.95549285e-01
5.35238199e-02 -5.59358656e-01 -3.81382763e-01 -1.47002244e+00
-8.36052895e-01 -4.61721480e-01 2.45252624e-01 9.25405562e-01
6.32309914e-01 -4.01513875e-01 2.38611951e-01 5.76077640e-01
-4.72972512e-01 -1.53557792e-01 -1.22496700e+00 -4.69450921e-01
-5.14106214e-01 -8.87459517e-01 5.07584989e-01 3.13026100e-01
2.13015936e-02 5.57993174e-01 1.17069758e-01 3.46728832e-01
4.37581450e-01 -4.31967229e-01 6.94535375e-01 -1.26296985e+00
1.94917396e-01 -9.73922491e-01 -6.18008852e-01 -8.04699361e-01
8.87038261e-02 -1.17448771e+00 2.03160331e-01 -1.96954286e+00
-3.46264273e-01 -1.76598534e-01 -3.46157819e-01 7.24093497e-01
1.82681978e-01 7.24614710e-02 2.54253477e-01 -2.04111010e-01
-7.82533407e-01 3.29057813e-01 8.39929998e-01 -2.91540027e-01
-8.65033790e-02 -2.04474226e-01 -7.98983991e-01 5.15303910e-01
6.05477571e-01 -1.93919495e-01 -6.52445078e-01 -7.04186440e-01
4.36392576e-01 -1.35869443e-01 9.85608518e-01 -1.49725699e+00
5.81201673e-01 8.31841305e-02 8.31041038e-01 -1.07531428e+00
5.88466287e-01 -7.80779183e-01 -1.80523008e-01 1.93459734e-01
-5.14818788e-01 5.04886150e-01 5.10638416e-01 3.97427320e-01
2.90752649e-01 2.74715632e-01 4.81875062e-01 -3.23467493e-01
-1.26975286e+00 1.65026292e-01 -5.29308558e-01 1.96162567e-01
6.31833613e-01 -4.78575557e-01 -6.71781898e-01 -1.65152907e-01
-7.28042126e-01 4.84835893e-01 2.46102989e-01 7.10571289e-01
7.81953692e-01 -1.39482510e+00 -6.59915805e-01 4.96028662e-01
2.03082651e-01 -2.08737537e-01 1.80107951e-01 1.21988726e+00
-1.17023990e-01 1.03354907e+00 -7.11293697e-01 -5.54187000e-01
-7.54320383e-01 7.28021324e-01 4.74294841e-01 1.89401023e-03
-7.00899839e-01 1.16990542e+00 2.62791127e-01 -5.03941476e-01
4.57006693e-01 -4.51513737e-01 -5.62725842e-01 5.35171628e-01
7.64267862e-01 3.25685471e-01 3.74368459e-01 -1.08592463e+00
-7.84616292e-01 3.68469000e-01 1.25317603e-01 -6.95897937e-01
1.17920637e+00 -2.98351556e-01 2.49899313e-01 2.89149523e-01
7.65376687e-01 -1.95210472e-01 -1.66326010e+00 -1.62968904e-01
1.91355184e-01 -2.97362655e-01 1.52229965e-01 -1.33902109e+00
-4.79977340e-01 1.05423856e+00 8.63788545e-01 -4.20929193e-01
7.88445890e-01 -1.32683009e-01 2.29621783e-01 5.07034004e-01
3.10510308e-01 -1.24434578e+00 -1.53146237e-01 8.74127090e-01
1.20141995e+00 -1.27014327e+00 -3.59582663e-01 2.54889131e-01
-7.00308025e-01 7.52525985e-01 7.76120007e-01 1.28764778e-01
2.28371367e-01 3.34943458e-02 -1.18286602e-01 -3.07704538e-01
-9.37177598e-01 -8.57283711e-01 5.65065026e-01 1.29940736e+00
1.26709659e-02 7.40320906e-02 4.43445414e-01 5.88992715e-01
-7.56141365e-01 -5.91639221e-01 -2.61576492e-02 5.44985771e-01
-6.46330655e-01 -4.43433613e-01 -3.13155562e-01 3.73726219e-01
2.24083886e-01 -6.04865253e-01 3.48030515e-02 7.30526209e-01
2.32898399e-01 1.01353800e+00 3.80236626e-01 -2.95172632e-01
7.95464993e-01 3.26941162e-01 3.93078417e-01 -4.29947227e-01
-7.47733057e-01 -2.30491519e-01 4.34584528e-01 -1.09269285e+00
-1.31457493e-01 -1.12979364e+00 -1.01709044e+00 8.42561051e-02
4.25291806e-01 -4.85001504e-01 7.88909912e-01 9.31596160e-01
5.31472027e-01 8.43926668e-01 -6.11715280e-02 -1.38083887e+00
-8.59345421e-02 -8.69426966e-01 1.56007260e-01 1.09228753e-02
6.82504594e-01 -8.90168011e-01 1.22877367e-01 -2.03213602e-01] | [6.222151279449463, 0.5793972611427307] |
a4b1d23b-3d67-4647-8936-a7a8307b1084 | a-multi-view-joint-learning-framework-for | 2301.11608 | null | https://arxiv.org/abs/2301.11608v1 | https://arxiv.org/pdf/2301.11608v1.pdf | A Multi-View Joint Learning Framework for Embedding Clinical Codes and Text Using Graph Neural Networks | Learning to represent free text is a core task in many clinical machine learning (ML) applications, as clinical text contains observations and plans not otherwise available for inference. State-of-the-art methods use large language models developed with immense computational resources and training data; however, applying these models is challenging because of the highly varying syntax and vocabulary in clinical free text. Structured information such as International Classification of Disease (ICD) codes often succinctly abstracts the most important facts of a clinical encounter and yields good performance, but is often not as available as clinical text in real-world scenarios. We propose a \textbf{multi-view learning framework} that jointly learns from codes and text to combine the availability and forward-looking nature of text and better performance of ICD codes. The learned text embeddings can be used as inputs to predictive algorithms independent of the ICD codes during inference. Our approach uses a Graph Neural Network (GNN) to process ICD codes, and Bi-LSTM to process text. We apply Deep Canonical Correlation Analysis (DCCA) to enforce the two views to learn a similar representation of each patient. In experiments using planned surgical procedure text, our model outperforms BERT models fine-tuned to clinical data, and in experiments using diverse text in MIMIC-III, our model is competitive to a fine-tuned BERT at a tiny fraction of its computational effort. | ['Yixin Chen', 'Bradley Fritz', 'Christopher King', 'Lecheng Kong'] | 2023-01-27 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [ 1.28734216e-01 3.23490620e-01 -5.78065336e-01 -3.79901350e-01
-8.42906237e-01 -3.80445540e-01 3.73551399e-01 8.18578899e-01
-3.60832989e-01 4.57854569e-01 8.06006253e-01 -7.42500663e-01
-3.41218293e-01 -6.98047042e-01 -5.11761129e-01 -3.78902018e-01
-3.26361567e-01 1.13993990e+00 -3.81733477e-01 -6.04028627e-02
-3.87860715e-01 4.11648840e-01 -9.28779304e-01 7.88563788e-01
4.87647951e-01 8.95819008e-01 1.92382127e-01 9.85570669e-01
-1.63629681e-01 1.38435388e+00 -2.00094655e-01 -5.05171537e-01
9.74463597e-02 -3.39217097e-01 -6.59809232e-01 -1.87797070e-01
1.17276005e-01 -3.02041411e-01 -8.58848393e-01 6.16463661e-01
7.00780213e-01 -2.45787799e-01 7.96834707e-01 -7.95321107e-01
-8.32219422e-01 7.70428896e-01 -1.21110089e-01 1.92348510e-01
2.67063439e-01 -5.30952681e-03 1.04115021e+00 -7.93570220e-01
9.01576281e-01 9.70455468e-01 7.25627005e-01 7.17147708e-01
-1.26987302e+00 -3.86773258e-01 4.80414741e-02 -7.95076489e-02
-1.20697236e+00 -2.01409221e-01 5.03252208e-01 -5.97970963e-01
1.09218764e+00 2.30800405e-01 7.17567205e-01 1.39690220e+00
1.05223525e+00 5.98545074e-01 7.46366560e-01 -1.32723689e-01
2.32660219e-01 4.72762622e-03 -2.02471483e-02 9.89043295e-01
1.50050372e-01 1.74136963e-02 -5.36870122e-01 -5.24568081e-01
5.50950944e-01 8.56553018e-01 -3.21457505e-01 -5.56393266e-01
-1.38977337e+00 1.05233002e+00 5.91268182e-01 2.41130218e-01
-3.60457540e-01 1.46147668e-01 7.51820624e-01 2.32160226e-01
4.30399060e-01 2.09964782e-01 -5.16216278e-01 -2.27829173e-01
-9.92489815e-01 -9.22037438e-02 1.06083596e+00 8.80481899e-01
1.85313508e-01 -2.38304988e-01 -1.16181023e-01 8.40437710e-01
1.58406243e-01 4.54912573e-01 8.01615894e-01 -3.90724719e-01
7.32162595e-01 7.42568076e-01 -5.87200761e-01 -9.49437976e-01
-7.99007058e-01 -5.31248510e-01 -1.40606737e+00 -2.57848352e-01
-1.07037343e-01 -1.74656108e-01 -1.26357436e+00 1.39841115e+00
-1.39261425e-01 -1.40503244e-02 2.38254637e-01 4.58449215e-01
1.06059563e+00 2.72462010e-01 9.84178334e-02 -1.34656027e-01
1.49935532e+00 -7.33706057e-01 -8.13993692e-01 -4.14436251e-01
1.16401827e+00 -5.45108974e-01 4.70386475e-01 2.11584494e-01
-9.98486876e-01 -1.34394690e-01 -6.13464594e-01 -2.72952974e-01
-3.81708741e-01 1.63904443e-01 5.69339395e-01 2.19591618e-01
-1.10557413e+00 6.03110433e-01 -1.20945215e+00 -3.49229008e-01
7.20479608e-01 4.35440272e-01 -6.54855549e-01 -4.91997987e-01
-1.05518377e+00 9.65868413e-01 2.73550689e-01 -7.81292915e-02
-8.37458849e-01 -9.49299753e-01 -1.24244571e+00 2.45129764e-01
4.05671030e-01 -1.20502400e+00 8.98422003e-01 -6.10288501e-01
-1.05463505e+00 1.02269650e+00 -1.55626953e-01 -5.73819697e-01
4.86164987e-01 3.11139058e-02 -6.13606215e-01 2.56603867e-01
-9.89518762e-02 3.52409631e-01 5.48219085e-01 -8.61290991e-01
-1.43620312e-01 -6.11591101e-01 -4.88350056e-02 6.50071427e-02
-2.78676480e-01 -3.66008729e-01 -5.03887355e-01 -7.41323113e-01
3.81026417e-02 -1.17937231e+00 -7.08013654e-01 3.29426497e-01
-4.81300861e-01 1.14315622e-01 3.20666045e-01 -7.33609438e-01
1.56121361e+00 -2.02129292e+00 3.22818488e-01 2.83853531e-01
7.58097470e-01 8.71674940e-02 1.11267135e-01 9.04946268e-01
-3.82684678e-01 9.15279463e-02 -2.65885770e-01 -4.88233328e-01
-4.48026359e-01 7.06840992e-01 -2.81267673e-01 5.81502438e-01
7.04507232e-02 1.06394219e+00 -1.02333033e+00 -7.74193525e-01
1.48728445e-01 4.50487524e-01 -1.02133930e+00 2.77294338e-01
-2.22571567e-03 5.15301377e-02 -4.08580601e-01 5.84595084e-01
1.51520461e-01 -9.71235454e-01 6.28754497e-01 -6.18431121e-02
5.03518641e-01 2.59170324e-01 -6.74042106e-01 2.06762528e+00
-8.13629210e-01 3.16870332e-01 -1.41124800e-01 -9.87780035e-01
4.46718752e-01 4.86045718e-01 8.55189741e-01 -2.56382227e-01
1.78953066e-01 1.55208195e-02 2.20979765e-01 -6.27283275e-01
2.33765021e-02 -3.08418393e-01 1.34145208e-02 3.58081400e-01
1.92133561e-01 -4.11226377e-02 -7.35867992e-02 6.44911051e-01
1.56287742e+00 -3.67492884e-01 6.47504449e-01 -1.93985682e-02
1.76882416e-01 -3.16528697e-03 4.84625936e-01 5.88676453e-01
7.67403096e-02 6.21002376e-01 7.45317638e-01 -7.61963964e-01
-8.89651775e-01 -1.17561829e+00 -3.09826642e-01 7.67464638e-01
-4.68089819e-01 -8.09785604e-01 -4.61004674e-02 -9.80793476e-01
8.57196376e-02 6.13503039e-01 -1.02469552e+00 -2.21903756e-01
-4.34547961e-01 -6.09689534e-01 3.78315002e-01 8.34112167e-01
-4.99247372e-01 -8.86575401e-01 -4.84494597e-01 3.80784214e-01
-9.76316980e-04 -1.18661261e+00 -5.58650255e-01 5.25668144e-01
-9.87703383e-01 -1.22792435e+00 -6.04175806e-01 -7.62430429e-01
7.89208710e-01 -1.70026898e-01 1.32263541e+00 1.82997450e-01
-6.53275311e-01 4.02888536e-01 -2.53992796e-01 -5.60798824e-01
-7.88866460e-01 -4.31743963e-03 -3.44995074e-02 -1.62351623e-01
2.53132910e-01 -4.71831620e-01 -5.84506631e-01 -2.98443347e-01
-1.14808202e+00 2.97543794e-01 7.38071799e-01 1.29375100e+00
6.10746980e-01 -5.28730750e-01 9.20061544e-02 -1.47380674e+00
6.12500489e-01 -7.23624706e-01 -1.30291760e-01 1.78592488e-01
-7.13159800e-01 3.03892493e-01 9.40243185e-01 -1.67713180e-01
-3.76390427e-01 3.91415246e-02 -2.15799928e-01 -9.65006411e-01
9.54612792e-02 1.06143022e+00 3.98472190e-01 4.02515769e-01
6.67919874e-01 1.93222135e-01 3.36797535e-01 -3.20984632e-01
3.51566911e-01 5.78987300e-01 1.07288614e-01 -1.65295601e-01
4.31713313e-01 6.07143164e-01 2.03407750e-01 -4.01880562e-01
-1.06848097e+00 -6.12302661e-01 -5.28185785e-01 1.77209973e-01
1.00321376e+00 -1.12281370e+00 -7.18711257e-01 -2.45493889e-01
-9.06785548e-01 -1.39652595e-01 -3.07291716e-01 6.38842404e-01
-6.75220430e-01 1.67503059e-01 -8.14107776e-01 -2.88519442e-01
-4.78239328e-01 -1.17719507e+00 1.27863097e+00 -4.82386410e-01
-3.58469963e-01 -1.48255324e+00 2.33957767e-01 1.50715977e-01
1.35700271e-01 4.43183362e-01 1.38321209e+00 -1.08591318e+00
-1.23009525e-01 -4.89494920e-01 -7.16626570e-02 1.17368065e-01
3.68501127e-01 -2.66828030e-01 -7.70401776e-01 -3.97896945e-01
-1.62093624e-01 -3.37344259e-01 9.28307593e-01 4.73872274e-01
1.65051222e+00 -4.26428825e-01 -7.65203416e-01 7.95648038e-01
1.47608304e+00 -2.71861106e-02 2.68570274e-01 -2.18776286e-01
9.94150281e-01 3.09673488e-01 2.05211397e-02 6.73504889e-01
6.29758537e-01 3.91062737e-01 4.50920314e-01 -2.00503275e-01
-5.72579466e-02 -3.00365508e-01 1.24905869e-01 1.24057245e+00
9.54568461e-02 -1.50267944e-01 -1.46878958e+00 5.32851279e-01
-1.82824683e+00 -1.01201570e+00 1.55753046e-01 1.91299880e+00
8.88388097e-01 1.42372906e-01 -3.24331433e-01 -2.39132226e-01
1.74022868e-01 7.25215375e-02 -6.49449408e-01 -6.02610171e-01
2.63872892e-01 3.81991565e-01 6.11457527e-01 1.52395368e-01
-9.13156748e-01 3.71175945e-01 6.31001186e+00 6.16810441e-01
-1.06281137e+00 1.24476634e-01 7.82982767e-01 -3.79100829e-01
-2.91391283e-01 -4.36898768e-01 -3.66990298e-01 3.57078940e-01
1.43646026e+00 -2.76168138e-01 2.55177766e-01 9.48083341e-01
8.52790549e-02 3.61830354e-01 -1.64749813e+00 1.29733896e+00
3.32025766e-01 -1.95624268e+00 2.97239542e-01 3.13282907e-01
5.37742674e-01 3.78314435e-01 5.30112572e-02 2.73134798e-01
7.42480159e-01 -1.20941472e+00 2.55767196e-01 4.86940891e-01
1.19598031e+00 -4.09072191e-01 9.06451106e-01 3.42399269e-01
-1.08505666e+00 -3.52110177e-01 -2.54247695e-01 4.77190971e-01
-2.03599990e-03 5.17542958e-01 -1.24298894e+00 9.00922000e-01
4.99852121e-01 1.17105246e+00 -5.48351705e-01 5.93319476e-01
1.47693336e-01 5.00539303e-01 -1.02695629e-01 1.66055575e-01
3.18925858e-01 1.90693244e-01 2.28398234e-01 1.23056102e+00
2.80748934e-01 2.36061618e-01 4.60760474e-01 5.88404179e-01
-3.31315786e-01 1.45740956e-01 -1.06700897e+00 -4.09608513e-01
4.60130088e-02 1.24128342e+00 -4.00865674e-01 -5.73787987e-01
-6.75459266e-01 7.62779772e-01 5.27170300e-01 2.11881459e-01
-6.84673905e-01 5.65326549e-02 6.58417165e-01 2.82063663e-01
1.62545815e-01 -1.18115032e-02 -3.25156629e-01 -1.34306324e+00
-1.00709304e-01 -1.00771916e+00 1.01678908e+00 -6.25280559e-01
-1.46693349e+00 8.85111392e-01 -3.45978945e-01 -1.74696863e+00
-6.13187194e-01 -8.78171325e-01 -1.52190775e-01 8.30448151e-01
-1.36367142e+00 -1.30530953e+00 -5.87588921e-02 6.97494149e-01
5.04666686e-01 -3.26309144e-01 1.42213118e+00 1.94443762e-01
-3.27492982e-01 5.34857333e-01 3.72628003e-01 6.38807237e-01
8.09695303e-01 -1.26111031e+00 1.48112997e-01 2.63810903e-01
1.33328751e-01 7.98527718e-01 2.60567695e-01 -6.47218764e-01
-1.77813697e+00 -1.41261101e+00 8.80878210e-01 -7.36638784e-01
9.29436982e-01 -4.17640626e-01 -7.71015346e-01 1.06099963e+00
-2.41147494e-03 5.04302621e-01 1.32640350e+00 1.55767351e-01
-3.54919374e-01 2.18204409e-02 -8.40749979e-01 5.02247989e-01
9.78878736e-01 -7.50671148e-01 -7.16253996e-01 6.51683450e-01
8.31162274e-01 -7.09876001e-01 -1.25477409e+00 2.10825235e-01
3.25357109e-01 -5.99323332e-01 9.20266211e-01 -1.20940244e+00
1.00561857e+00 3.28649014e-01 -1.49935797e-01 -1.54479218e+00
-2.88427383e-01 -2.09676445e-01 -1.95529670e-01 3.67042154e-01
5.92419028e-01 -5.94550133e-01 6.33823812e-01 6.39877021e-01
-2.81245112e-01 -1.32287705e+00 -9.12503123e-01 -4.07558620e-01
2.05101728e-01 -3.91734272e-01 2.19959900e-01 1.16725338e+00
3.71935725e-01 2.39313826e-01 -2.16266349e-01 1.45552456e-01
3.21626574e-01 2.75815874e-01 2.46568605e-01 -1.29667342e+00
-3.84970516e-01 -2.80624002e-01 -5.64812005e-01 -5.11591315e-01
1.36061311e-01 -1.54886067e+00 -2.95764893e-01 -1.99224854e+00
3.67781878e-01 -4.96079594e-01 -4.51630086e-01 6.19709015e-01
-1.55141667e-01 -1.90055028e-01 2.66673863e-01 1.55832797e-01
-3.86168122e-01 2.84326345e-01 1.22726357e+00 -3.77192318e-01
-2.49335133e-02 -1.64116532e-01 -5.71547449e-01 6.87597275e-01
3.28652680e-01 -7.12910712e-01 -3.57296914e-01 -4.46690708e-01
4.53888744e-01 7.53827453e-01 1.56240702e-01 -7.73946524e-01
3.89534324e-01 -5.68194650e-02 5.79555154e-01 -4.82143164e-01
3.64218980e-01 -1.18991220e+00 4.18805555e-02 7.68530786e-01
-5.97162962e-01 3.26401263e-01 2.82230139e-01 9.57625270e-01
-2.29848981e-01 1.45934284e-01 4.58766997e-01 -3.92779350e-01
-9.62224677e-02 8.71722877e-01 -3.86551172e-01 3.33349049e-01
8.63118351e-01 8.19737390e-02 -1.89260393e-01 -4.17971760e-01
-1.06909382e+00 3.29458654e-01 2.10705683e-01 5.53795993e-01
8.68368268e-01 -1.26169634e+00 -7.08539248e-01 2.07508191e-01
4.51211601e-01 2.07731426e-02 4.20426786e-01 9.58386123e-01
-6.77115023e-01 5.45236349e-01 9.22927782e-02 -6.21119559e-01
-1.27914453e+00 1.06516826e+00 1.65598124e-01 -9.13501263e-01
-9.53332245e-01 5.69299400e-01 3.06881905e-01 -4.16607440e-01
2.69128114e-01 -5.77784598e-01 -1.03592522e-01 9.73793417e-02
6.34815812e-01 -2.25897834e-01 2.02542007e-01 -2.92058408e-01
-4.39785540e-01 2.83777237e-01 -3.68783951e-01 2.62410879e-01
1.74445343e+00 1.95166662e-01 -6.63658902e-02 6.20144188e-01
1.40628910e+00 -2.17401043e-01 -5.75983703e-01 -4.64439332e-01
-1.54080227e-01 -9.20718238e-02 2.07725912e-01 -7.91108131e-01
-1.12663043e+00 1.18653309e+00 5.04105449e-01 -6.31331280e-02
9.77834821e-01 3.02503426e-02 7.06754088e-01 4.85012680e-01
2.05037862e-01 -7.53212810e-01 4.63157333e-02 2.91654617e-01
8.74944508e-01 -1.27917278e+00 5.85038103e-02 -1.73738509e-01
-9.91996050e-01 1.43744946e+00 2.52633095e-01 5.82587346e-02
1.14178991e+00 4.79165822e-01 2.49097452e-01 -5.20075321e-01
-1.32251394e+00 1.26541585e-01 4.10974920e-01 3.01109135e-01
4.86647695e-01 3.08097839e-01 2.01576486e-01 5.75025797e-01
-3.11791543e-02 8.33300874e-02 6.87689364e-01 1.05840588e+00
1.81928977e-01 -9.57135975e-01 -3.78468670e-02 1.21456122e+00
-7.36627519e-01 -6.00462139e-01 -1.41003132e-02 6.10472500e-01
-1.01837628e-01 6.23572052e-01 1.21767499e-01 -2.77327001e-01
3.40559989e-01 2.10820004e-01 9.49016139e-02 -1.05876565e+00
-6.87187791e-01 -4.37728502e-02 9.34962928e-03 -8.37984681e-01
-1.59333378e-01 -4.59222436e-01 -1.42542171e+00 -2.71277010e-01
8.25825855e-02 -9.23954248e-02 4.52678174e-01 6.86747193e-01
6.44647956e-01 9.83350933e-01 3.33254576e-01 -3.29985857e-01
-5.53371310e-01 -6.22385621e-01 -5.36258399e-01 6.07827961e-01
6.83926582e-01 -4.08376604e-01 5.33378907e-02 3.01245064e-01] | [7.948308944702148, 6.743780136108398] |
dd834058-51cf-4c5d-ba8d-544df564a54b | what-makes-sound-event-localization-and | 2107.10469 | null | https://arxiv.org/abs/2107.10469v2 | https://arxiv.org/pdf/2107.10469v2.pdf | What Makes Sound Event Localization and Detection Difficult? Insights from Error Analysis | Sound event localization and detection (SELD) is an emerging research topic that aims to unify the tasks of sound event detection and direction-of-arrival estimation. As a result, SELD inherits the challenges of both tasks, such as noise, reverberation, interference, polyphony, and non-stationarity of sound sources. Furthermore, SELD often faces an additional challenge of assigning correct correspondences between the detected sound classes and directions of arrival to multiple overlapping sound events. Previous studies have shown that unknown interferences in reverberant environments often cause major degradation in the performance of SELD systems. To further understand the challenges of the SELD task, we performed a detailed error analysis on two of our SELD systems, which both ranked second in the team category of DCASE SELD Challenge, one in 2020 and one in 2021. Experimental results indicate polyphony as the main challenge in SELD, due to the difficulty in detecting all sound events of interest. In addition, the SELD systems tend to make fewer errors for the polyphonic scenario that is dominant in the training set. | ['Woon Seng Gan', 'Douglas L. Jones', 'Ngoc Khanh Nguyen', 'Zhen Jian Lee', 'Karn N. Watcharasupat', 'Thi Ngoc Tho Nguyen'] | 2021-07-22 | null | null | null | null | ['direction-of-arrival-estimation', 'sound-event-localization-and-detection'] | ['audio', 'audio'] | [-3.11162502e-01 -7.14910686e-01 8.07230949e-01 9.00061056e-02
-1.25068605e+00 -5.45317769e-01 9.32440013e-02 1.36063725e-01
-2.26873815e-01 4.94291812e-01 3.02157968e-01 -6.72461838e-02
-2.55994368e-02 -3.05283993e-01 -5.48050761e-01 -4.70931947e-01
-2.33167484e-01 -5.41732349e-02 4.23867136e-01 -9.21855643e-02
-6.01593778e-02 2.40465388e-01 -1.77857327e+00 1.88523918e-01
4.18636709e-01 1.00456452e+00 4.45056885e-01 1.00700617e+00
1.96167111e-01 6.56563461e-01 -1.24964440e+00 2.37099845e-02
-1.52775757e-02 -6.31156266e-01 -1.48161411e-01 -4.64226753e-01
4.18006748e-01 -7.56843612e-02 -2.11918443e-01 7.02388525e-01
1.01279783e+00 2.10123762e-01 2.89461076e-01 -1.28230584e+00
4.17481959e-01 3.27593714e-01 -2.51430035e-01 8.41512978e-01
4.95856076e-01 -1.47328451e-01 9.05593753e-01 -1.32059932e+00
-6.88624308e-02 8.47114801e-01 9.42430913e-01 9.87286493e-03
-7.45415211e-01 -9.06985939e-01 -2.49852702e-01 3.84657830e-01
-1.52614772e+00 -7.01832771e-01 6.45599306e-01 -3.94379139e-01
7.53109574e-01 5.65047383e-01 3.28198075e-01 1.06224191e+00
5.12392856e-02 2.73029536e-01 6.01084173e-01 -5.41348100e-01
2.15389878e-01 1.29366860e-01 3.33562233e-02 -7.46269748e-02
1.61685169e-01 2.84555852e-01 -9.43645418e-01 -1.79801404e-01
2.10579708e-01 -2.93567210e-01 -4.95541692e-01 5.24333894e-01
-1.09108281e+00 2.91576773e-01 1.16317630e-01 3.79592329e-01
-2.74970025e-01 2.11220264e-01 2.70502120e-01 1.75000504e-01
4.44781244e-01 5.78242302e-01 -5.04217803e-01 -4.21747237e-01
-9.09441769e-01 1.52663171e-01 8.97855759e-01 7.86231756e-01
4.48299766e-01 3.10162127e-01 -6.82796612e-02 9.10339296e-01
3.01140845e-01 7.28987098e-01 1.90901831e-01 -5.28378785e-01
4.89197224e-01 -3.77672255e-01 2.81452894e-01 -1.19022310e+00
-6.05320334e-01 -1.25283539e+00 -4.84071314e-01 -2.71857351e-01
2.89003789e-01 -4.65421379e-01 -3.29529256e-01 1.67167318e+00
2.91905880e-01 5.25539875e-01 -2.47223467e-01 7.71457553e-01
5.30203164e-01 7.72622287e-01 -5.78539073e-02 -3.17415982e-01
1.25356889e+00 -8.12481284e-01 -8.43263388e-01 -4.14742440e-01
-9.79095921e-02 -1.25386596e+00 6.66710377e-01 5.12570143e-01
-7.69542754e-01 -6.91533267e-01 -8.32299054e-01 5.77331781e-01
3.22118476e-02 1.45931333e-01 1.83636695e-01 7.15739131e-01
-5.73725343e-01 1.60277992e-01 -6.28951788e-01 -1.16819046e-01
-2.41897345e-01 -2.95943856e-01 1.34621903e-01 1.78443447e-01
-1.26371980e+00 6.16969764e-01 -2.17685074e-01 7.35338554e-02
-1.01915276e+00 -1.11649764e+00 -5.74779272e-01 3.10443044e-01
1.78089857e-01 1.50028476e-02 1.68759775e+00 -3.85537833e-01
-9.51229811e-01 1.72827691e-01 -3.04519802e-01 -4.28148568e-01
4.29803610e-01 -6.52242959e-01 -1.20733953e+00 1.22770645e-01
3.38865310e-01 -3.40297580e-01 8.81060004e-01 -9.95742500e-01
-1.04207599e+00 1.19883336e-01 -5.80926359e-01 5.40918782e-02
-2.88389057e-01 4.09714699e-01 -2.66202569e-01 -7.57029116e-01
3.49728823e-01 -5.70180714e-01 1.35455523e-02 -4.18316811e-01
-2.30288655e-01 7.19850883e-03 6.60796583e-01 -6.08331680e-01
1.44510341e+00 -2.60264158e+00 -6.21435761e-01 4.12383042e-02
2.15629879e-02 -3.23430374e-02 3.06923568e-01 6.93050861e-01
-2.32514933e-01 -1.70554176e-01 3.67169291e-01 -3.00025672e-01
-4.21621948e-02 -1.37181833e-01 -6.99799418e-01 3.18522602e-01
1.75649375e-02 4.65630665e-02 -1.03769326e+00 -9.42070782e-02
6.46289214e-02 2.12947071e-01 -3.06262434e-01 2.81883448e-01
2.27601737e-01 5.02896905e-01 -3.26537117e-02 4.09994841e-01
6.67892098e-01 9.51550156e-02 -1.85363069e-01 -2.82480657e-01
-4.38467473e-01 5.65772295e-01 -1.49902999e+00 1.00950742e+00
-7.06988037e-01 1.02258945e+00 3.49066079e-01 -6.71094894e-01
7.02183902e-01 8.18041027e-01 4.87561136e-01 -6.93880439e-01
-2.05063671e-01 4.47035134e-01 3.23809534e-02 -5.66789210e-01
4.49078619e-01 -1.99423060e-01 7.72483507e-03 3.88777286e-01
-1.66325290e-02 -1.24801978e-01 1.72325969e-01 2.44701793e-03
1.32423747e+00 -4.59380865e-01 9.51175019e-02 -3.81657109e-02
1.29579112e-01 -4.89762843e-01 7.79227734e-01 1.05803561e+00
-2.61936069e-01 9.47581172e-01 7.33136088e-02 -6.69545447e-03
-3.97049129e-01 -1.27975857e+00 -1.35467008e-01 1.16003168e+00
-3.57507989e-02 -4.94905800e-01 -4.80662912e-01 -2.62352765e-01
-2.21541435e-01 9.14771557e-01 4.14743945e-02 -1.35655388e-01
-4.74066108e-01 -7.09587991e-01 7.42325485e-01 3.53736520e-01
2.39628538e-01 -7.87036419e-01 -6.59009039e-01 5.87389350e-01
-6.20516002e-01 -1.42696571e+00 -5.72964668e-01 4.20151234e-01
-1.17479011e-01 -9.20407474e-01 -4.43998426e-01 -5.79926491e-01
3.47535163e-01 6.52351141e-01 1.23745799e+00 -2.13442847e-01
-5.60853362e-01 5.88228822e-01 -5.94720006e-01 -1.09398758e+00
-3.13889802e-01 -4.13469017e-01 2.37940580e-01 1.19496182e-01
-4.73319665e-02 -7.48938203e-01 -6.93279743e-01 7.57391274e-01
-4.78878140e-01 -4.85877156e-01 2.73136824e-01 3.94421220e-01
6.33964092e-02 7.42932260e-01 9.38613534e-01 -2.34222203e-01
5.12424052e-01 -8.09593260e-01 -5.00067294e-01 -1.64660141e-01
-3.02154422e-01 -7.75072694e-01 5.94401240e-01 -3.24888051e-01
-1.06085336e+00 -1.17752656e-01 -2.97853649e-01 -7.75877386e-02
-3.39190304e-01 3.47356766e-01 -3.77953760e-02 1.96596965e-01
9.09250259e-01 1.75206706e-01 -6.25861824e-01 -6.67975307e-01
-3.42921764e-01 8.28890204e-01 6.12618268e-01 -3.11834365e-01
6.74069524e-01 2.60058880e-01 -2.13813260e-01 -1.11487472e+00
-9.34457481e-01 -8.53666842e-01 1.13390788e-01 -4.34584886e-01
5.05985439e-01 -1.23733592e+00 -2.28569120e-01 6.56005800e-01
-1.27882886e+00 2.32980892e-01 -1.17708996e-01 9.45434332e-01
-1.05098784e-02 1.41818568e-01 -3.50448549e-01 -1.20333588e+00
-1.73951074e-01 -8.27174842e-01 8.08661938e-01 2.28689253e-01
-5.29740751e-01 -3.65493208e-01 1.20205291e-01 1.83626011e-01
5.25151432e-01 -1.35584604e-02 6.35083735e-01 -4.27086234e-01
-1.94381028e-01 -4.88640666e-01 2.33918637e-01 4.64945942e-01
3.54031563e-01 -1.30802229e-01 -1.33544707e+00 -4.50315252e-02
3.02061945e-01 3.00019234e-01 4.51525360e-01 4.70754743e-01
7.35597968e-01 1.40681535e-01 -1.86259881e-01 1.67624593e-01
9.55383599e-01 4.76866603e-01 4.66513127e-01 -4.25843000e-02
2.82660484e-01 4.19793159e-01 7.54643738e-01 6.55616701e-01
1.46830782e-01 8.68959129e-01 3.72233987e-01 -1.82509586e-01
-4.17461008e-01 -1.56305224e-01 3.81398261e-01 1.01718473e+00
3.74891758e-01 -4.50606138e-01 -1.03691971e+00 7.60541260e-01
-1.37765276e+00 -9.91364658e-01 -6.08976901e-01 2.49458647e+00
7.42122948e-01 6.78540766e-02 -8.12209770e-03 5.58826625e-01
8.45811665e-01 -7.97686130e-02 -4.54378518e-04 5.53480862e-03
3.48835625e-02 3.89688760e-01 2.50938118e-01 2.45764747e-01
-1.04574013e+00 2.34374210e-01 6.99601984e+00 6.22241080e-01
-1.03495574e+00 2.60855317e-01 1.66963786e-01 -3.13045263e-01
2.25223854e-01 -2.83407062e-01 -7.56953657e-01 7.13946939e-01
1.17419255e+00 -1.03586912e-01 1.17950730e-01 6.17544591e-01
5.54557025e-01 -4.66555119e-01 -1.05497634e+00 1.00554323e+00
8.85973349e-02 -8.19089532e-01 -5.72233438e-01 -3.46664071e-01
5.13523161e-01 1.41577020e-01 4.76433476e-03 4.29679632e-01
-2.06865400e-01 -6.73140347e-01 1.19345760e+00 3.09889704e-01
5.18705845e-01 -7.49043584e-01 6.13478363e-01 4.28461045e-01
-1.30509043e+00 -9.75593328e-02 -6.42092898e-02 -4.08328652e-01
2.88482457e-01 1.27139902e+00 -1.06759131e+00 5.64156651e-01
1.05826843e+00 2.24826127e-01 -1.42521515e-01 1.66610134e+00
-3.07706505e-01 1.19231200e+00 -5.42795002e-01 1.10323645e-01
-3.59193921e-01 5.13891280e-01 9.82083797e-01 1.32216990e+00
6.89926445e-01 -7.98062235e-02 1.86370015e-01 4.19623345e-01
5.85253835e-02 -2.10126743e-01 -1.85201705e-01 1.94968805e-01
7.73648202e-01 1.03694201e+00 -6.56590879e-01 1.20018609e-01
-3.23640883e-01 6.15107000e-01 -4.87213969e-01 4.84103322e-01
-1.13207495e+00 -6.96172595e-01 6.75948858e-01 8.10386688e-02
4.23430860e-01 -2.25288570e-01 -3.34839374e-02 -6.33042336e-01
2.11589739e-01 -7.56418228e-01 1.09513655e-01 -8.82283330e-01
-9.79327559e-01 5.20305157e-01 -2.99378932e-01 -1.50889373e+00
-1.11554295e-01 -1.40352368e-01 -8.08792233e-01 7.99627841e-01
-1.25942600e+00 -3.38489771e-01 -3.51047158e-01 2.63188243e-01
7.69631624e-01 9.81492996e-02 9.53770638e-01 1.09573913e+00
-5.31948030e-01 5.22854030e-01 1.85177371e-01 -1.41187176e-01
1.00632405e+00 -1.02057934e+00 4.26341265e-01 1.10046017e+00
3.10428739e-01 1.73718542e-01 1.07087004e+00 -4.60699022e-01
-1.13220143e+00 -1.21367192e+00 1.12249303e+00 -4.47827607e-01
7.01398194e-01 -5.65916002e-01 -5.04026711e-01 1.67543948e-01
-3.36613566e-01 2.08014265e-01 7.14551985e-01 2.25834712e-01
-8.28282163e-02 -3.32613915e-01 -6.59308195e-01 3.17558438e-01
6.44544542e-01 -5.87797403e-01 -4.12124157e-01 4.34093267e-01
6.79101408e-01 -4.81457978e-01 -3.41939449e-01 1.22368246e-01
4.36223328e-01 -1.00406587e+00 1.01039791e+00 1.36753395e-02
9.80614945e-02 -4.46732402e-01 -2.59025216e-01 -1.36403680e+00
-2.35796973e-01 -8.54466498e-01 -9.97820199e-02 1.63764882e+00
3.96897405e-01 -5.38108051e-01 8.31887275e-02 -1.24400027e-01
-2.10688010e-01 -1.29014507e-01 -9.43462968e-01 -1.03725100e+00
-6.74843192e-01 -9.67918158e-01 4.13905829e-01 4.54482973e-01
-4.22384709e-01 4.30805206e-01 -5.19631922e-01 7.83647656e-01
5.49756527e-01 -8.19076300e-02 5.24272382e-01 -1.21645236e+00
-5.08150458e-01 -4.50642258e-02 -2.18565702e-01 -7.92912185e-01
-5.75592935e-01 -3.11657906e-01 8.87975991e-01 -1.17521584e+00
-3.72009963e-01 -3.99809867e-01 -6.59394026e-01 -7.19185770e-02
-3.87426823e-01 2.18463764e-01 3.77609357e-02 -1.67585500e-02
-6.52915657e-01 2.80190408e-01 4.80354160e-01 2.19153151e-01
-1.18067391e-01 7.31414676e-01 -5.73971510e-01 8.79268408e-01
6.69864118e-01 -9.83599365e-01 -2.93282747e-01 -6.26767337e-01
3.94484103e-01 2.06729397e-01 5.55459797e-01 -1.49198318e+00
4.64925855e-01 1.69333190e-01 1.72560900e-01 -7.80182123e-01
2.42824331e-01 -1.00158644e+00 3.62905741e-01 2.28572950e-01
7.07964525e-02 -9.34195891e-02 4.62573558e-01 6.00951612e-01
-3.93075377e-01 -5.99551238e-02 5.35969555e-01 1.22831866e-01
-5.20663321e-01 -2.20152438e-01 -8.82664263e-01 3.69987130e-01
7.47724414e-01 1.93033904e-01 -1.48373812e-01 -5.56364119e-01
-5.82079113e-01 -1.06441595e-01 -4.94250864e-01 4.53987002e-01
3.47197294e-01 -9.47252393e-01 -8.95871520e-01 1.05963491e-01
-7.76188150e-02 -2.26086766e-01 4.68435317e-01 1.07306504e+00
-2.50705957e-01 2.60388464e-01 2.11916938e-01 -5.91522217e-01
-1.14102983e+00 -8.27361569e-02 4.09946293e-01 1.48975998e-02
-4.97743130e-01 1.22765887e+00 1.94417521e-01 5.26121482e-02
6.37453139e-01 -2.38899320e-01 -1.78070292e-02 1.77097261e-01
7.58647859e-01 8.74274850e-01 6.38823390e-01 -3.04500341e-01
-5.89247108e-01 2.51153350e-01 2.02855036e-01 -2.66147166e-01
1.05922794e+00 -2.69387335e-01 -5.14693325e-03 7.12024093e-01
9.67056632e-01 6.37716711e-01 -8.05694222e-01 -8.48930627e-02
-4.41763364e-02 -2.77629882e-01 2.76478589e-01 -9.70446825e-01
-7.42876530e-01 7.33241498e-01 8.08582485e-01 5.24985135e-01
1.26468194e+00 -9.20557156e-02 9.27840292e-01 1.41558334e-01
4.16823864e-01 -9.64037120e-01 4.26139049e-02 7.10613728e-01
8.76889825e-01 -8.44279170e-01 -2.29996458e-01 -3.35329980e-01
-7.27185532e-02 8.37356150e-01 3.83673370e-01 2.50119627e-01
7.93512166e-01 5.72681963e-01 3.98845345e-01 -5.63989766e-02
-5.10268211e-01 -7.41761252e-02 1.48783609e-01 4.53099102e-01
2.74482578e-01 1.91298611e-02 9.42654442e-03 8.96540046e-01
-4.90283519e-01 -5.29105902e-01 3.29344511e-01 8.84084940e-01
-4.49630588e-01 -5.57445288e-01 -8.62920344e-01 2.00992450e-01
-7.15553284e-01 -2.75624335e-01 -7.51303583e-02 2.59965688e-01
1.88159421e-01 1.57010722e+00 8.49358588e-02 -3.83256674e-01
7.99686551e-01 -1.13980379e-02 -6.94407895e-02 -7.15685785e-01
-6.83167636e-01 3.83547306e-01 1.20641112e-01 -4.62319702e-01
1.97888538e-01 -7.95282960e-01 -1.17002118e+00 1.49092808e-01
-6.29355967e-01 5.76461911e-01 8.26125145e-01 8.43945444e-01
3.70105773e-01 1.33792782e+00 8.91757250e-01 -5.87700784e-01
-6.17462397e-01 -9.54960585e-01 -6.89471483e-01 1.89212516e-01
6.39331102e-01 -5.48062444e-01 -8.85773361e-01 3.22022066e-02] | [15.18349552154541, 5.319964408874512] |
a1eb4efa-3f50-49ef-9bf9-bb9809377065 | decision-making-in-non-stationary | 2202.13003 | null | https://arxiv.org/abs/2202.13003v1 | https://arxiv.org/pdf/2202.13003v1.pdf | Decision Making in Non-Stationary Environments with Policy-Augmented Monte Carlo Tree Search | Decision-making under uncertainty (DMU) is present in many important problems. An open challenge is DMU in non-stationary environments, where the dynamics of the environment can change over time. Reinforcement Learning (RL), a popular approach for DMU problems, learns a policy by interacting with a model of the environment offline. Unfortunately, if the environment changes the policy can become stale and take sub-optimal actions, and relearning the policy for the updated environment takes time and computational effort. An alternative is online planning approaches such as Monte Carlo Tree Search (MCTS), which perform their computation at decision time. Given the current environment, MCTS plans using high-fidelity models to determine promising action trajectories. These models can be updated as soon as environmental changes are detected to immediately incorporate them into decision making. However, MCTS's convergence can be slow for domains with large state-action spaces. In this paper, we present a novel hybrid decision-making approach that combines the strengths of RL and planning while mitigating their weaknesses. Our approach, called Policy Augmented MCTS (PA-MCTS), integrates a policy's actin-value estimates into MCTS, using the estimates to seed the action trajectories favored by the search. We hypothesize that PA-MCTS will converge more quickly than standard MCTS while making better decisions than the policy can make on its own when faced with nonstationary environments. We test our hypothesis by comparing PA-MCTS with pure MCTS and an RL agent applied to the classical CartPole environment. We find that PC-MCTS can achieve higher cumulative rewards than the policy in isolation under several environmental shifts while converging in significantly fewer iterations than pure MCTS. | ['Abhishek Dubey', 'Ayan Mukhopadhyay', 'Geoffrey Pettet'] | 2022-02-25 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 1.58389226e-01 4.19720709e-02 -3.09051991e-01 8.38179886e-02
-8.53053570e-01 -7.80232966e-01 6.56301022e-01 2.65079767e-01
-8.27124417e-01 1.23568034e+00 1.27607867e-01 -5.49295843e-01
-1.65168896e-01 -8.92972052e-01 -7.74932265e-01 -6.59806728e-01
-3.74302983e-01 7.75987267e-01 5.10029793e-01 -1.97187468e-01
3.34199876e-01 3.71900439e-01 -1.43946230e+00 -1.28616929e-01
9.09950972e-01 6.31803155e-01 5.86603284e-01 9.48817372e-01
9.78054330e-02 9.03102040e-01 -3.79272014e-01 2.56541848e-01
4.79984522e-01 -4.63974923e-01 -6.37986600e-01 -1.03566088e-01
-3.80348057e-01 -6.79960847e-01 -1.69251606e-01 9.82557595e-01
5.06214797e-01 7.39110649e-01 3.96711290e-01 -9.35744226e-01
1.82110861e-01 6.15808785e-01 -3.52405101e-01 1.47545338e-01
4.55856770e-01 8.20437133e-01 6.54793680e-01 -6.03962988e-02
5.57638407e-01 1.54513276e+00 3.66429180e-01 5.64396858e-01
-1.38312721e+00 -3.27614307e-01 7.03041971e-01 1.05575718e-01
-9.44084644e-01 -2.24313274e-01 3.47828418e-01 -1.92176774e-01
1.06024253e+00 -2.64470205e-02 9.92491722e-01 9.96973574e-01
5.75660765e-01 1.01480091e+00 1.50190401e+00 -2.73401022e-01
1.17239845e+00 -4.46598023e-01 -6.87889159e-01 4.09643441e-01
1.16633289e-01 8.57736170e-01 -4.22937185e-01 -3.80131006e-01
7.29912639e-01 -1.07236855e-01 3.78901660e-02 -3.26765507e-01
-1.14552438e+00 6.60225213e-01 1.33371845e-01 -6.89472556e-02
-8.46708417e-01 5.98296225e-01 1.42322227e-01 4.64679211e-01
1.01346202e-01 7.72544920e-01 -5.16172349e-01 -6.83655500e-01
-7.75530696e-01 9.48242009e-01 6.96857452e-01 5.47604084e-01
6.83632076e-01 1.61002293e-01 -3.70160639e-01 2.32009977e-01
3.00855666e-01 6.48755014e-01 4.21632439e-01 -1.38235092e+00
4.58068222e-01 -2.00747065e-02 9.53152359e-01 -3.96838546e-01
-3.17588210e-01 -1.97577029e-01 -1.32391006e-01 7.54574835e-01
4.55392241e-01 -7.37640142e-01 -1.05790710e+00 1.88169026e+00
7.04602659e-01 2.83876121e-01 3.09097409e-01 6.84334636e-01
-4.81127918e-01 7.58359730e-01 1.08262539e-01 -5.20779848e-01
6.35445714e-01 -6.63278937e-01 -7.19770372e-01 -6.08547866e-01
3.79274487e-01 -3.45047355e-01 7.99511492e-01 4.53430176e-01
-1.12924910e+00 -3.06870162e-01 -9.88371849e-01 8.77210438e-01
-2.16328483e-02 -6.67125046e-01 4.54372972e-01 2.93708414e-01
-1.05844402e+00 1.19637299e+00 -1.31543934e+00 -2.32062787e-01
2.42322236e-01 3.39195788e-01 3.42099160e-01 -5.62188849e-02
-1.18931174e+00 1.32780910e+00 5.88225782e-01 -3.47747914e-02
-1.58403206e+00 -1.87477350e-01 -6.97205126e-01 -2.30941370e-01
1.22444284e+00 -4.74864244e-01 2.04171610e+00 -7.43020952e-01
-2.13638854e+00 -2.82061070e-01 -1.83723062e-01 -7.14777470e-01
8.79937530e-01 -1.82225078e-01 -8.30206499e-02 -2.30833709e-01
1.53221548e-01 5.70524096e-01 8.58622551e-01 -1.11413956e+00
-9.59815145e-01 -2.22723439e-01 1.21718712e-01 7.82702208e-01
5.30699611e-01 -3.31429511e-01 6.55101612e-02 -1.28604904e-01
-5.49476082e-03 -1.15361834e+00 -1.08507276e+00 -4.34179425e-01
-1.39344975e-01 -9.19217244e-02 4.54408228e-01 -1.74656168e-01
1.15851057e+00 -1.59970140e+00 1.67881072e-01 2.87622392e-01
-2.88455635e-01 1.71379566e-01 -1.61184445e-01 8.10415983e-01
5.48560441e-01 -1.12077355e-01 -3.25797439e-01 -8.56153220e-02
1.36711568e-01 6.08926117e-01 -4.42455679e-01 1.81665182e-01
-1.11142471e-02 8.12119901e-01 -1.41027415e+00 -1.77435279e-01
1.60211354e-01 -3.50956917e-01 -5.68419933e-01 1.70175865e-01
-9.28723574e-01 7.71140575e-01 -9.23487246e-01 5.84279776e-01
2.06541479e-01 1.52323022e-01 6.09253287e-01 7.69963086e-01
-4.86913055e-01 2.90967792e-01 -1.50016856e+00 1.39111912e+00
-2.99839109e-01 2.26427063e-01 1.13774389e-01 -8.63750160e-01
6.15671515e-01 1.77555636e-01 5.03499568e-01 -7.25903928e-01
-1.47036118e-02 2.25409478e-01 1.82901010e-01 -4.59361851e-01
4.70474035e-01 -2.85471708e-01 -1.08262844e-01 7.08605528e-01
-4.23419207e-01 -5.51315725e-01 1.06660552e-01 -7.05128983e-02
1.47136390e+00 6.65480316e-01 6.94577456e-01 5.37437759e-02
-9.87180471e-02 3.75835836e-01 8.35165739e-01 1.26157510e+00
-4.88593996e-01 -2.04804763e-01 4.16077197e-01 -5.23197055e-01
-7.39322722e-01 -1.04156613e+00 1.65843576e-01 9.48653579e-01
3.17794681e-01 -1.87782153e-01 -5.00573933e-01 -6.50381982e-01
3.13286811e-01 1.14506316e+00 -4.32229429e-01 9.32900992e-04
-5.34057677e-01 -6.81241333e-01 -3.59918363e-02 3.12793553e-01
5.13240695e-01 -1.23127985e+00 -1.29790711e+00 9.14845765e-01
8.68634582e-02 -6.60685003e-01 -2.63041735e-01 3.44598114e-01
-9.75014925e-01 -9.15376246e-01 -4.01020586e-01 5.07975854e-02
5.02831221e-01 4.86512156e-03 7.39501238e-01 -3.01429629e-01
3.56227040e-01 5.90453029e-01 -1.65976688e-01 -6.11586273e-01
-6.70966506e-01 -2.67307550e-01 4.70298618e-01 -3.30323160e-01
-1.54292136e-01 -5.57753444e-01 -4.72463369e-01 3.47219944e-01
-6.68813586e-01 8.04204419e-02 6.38696253e-01 7.98415303e-01
7.81976998e-01 4.69050944e-01 5.11241913e-01 -7.34739602e-01
8.87519956e-01 -4.96194273e-01 -1.08986938e+00 1.01045832e-01
-7.91919172e-01 4.00810122e-01 5.33533514e-01 -7.46491611e-01
-1.18142486e+00 1.33929044e-01 1.87459588e-01 -4.05448109e-01
-1.05882891e-01 6.03291333e-01 1.38134196e-01 2.07077548e-01
5.63422143e-01 2.09715441e-01 -3.05115748e-02 -3.28654081e-01
2.11125135e-01 2.80741394e-01 2.68079579e-01 -1.04995155e+00
7.47852862e-01 2.05212817e-01 -5.93052134e-02 -2.89864957e-01
-6.21329248e-01 1.23518325e-01 -4.17929828e-01 -5.75624824e-01
5.64757049e-01 -6.49554014e-01 -7.02282667e-01 3.58967572e-01
-8.29396009e-01 -1.16178524e+00 -6.00712478e-01 5.69336176e-01
-9.30900455e-01 -4.29285765e-02 -2.27470160e-01 -1.42519557e+00
2.94384748e-01 -1.17884052e+00 6.69531584e-01 4.23179299e-01
-1.45538896e-01 -8.80482197e-01 5.95107675e-01 -7.04325587e-02
4.13046539e-01 4.30858523e-01 5.27443409e-01 -3.26333553e-01
-7.80725241e-01 1.33961901e-01 5.70370913e-01 -1.78633675e-01
-2.19500475e-02 -2.28283226e-01 -6.76010430e-01 -6.37891233e-01
-2.36034729e-02 -3.02997559e-01 6.26379490e-01 6.48119986e-01
7.62612641e-01 -5.84556162e-01 -4.24467057e-01 2.92815082e-02
1.37306547e+00 9.32822645e-01 2.85562843e-01 7.50737846e-01
-2.28097718e-02 4.23965544e-01 1.22769153e+00 7.42221355e-01
4.87808764e-01 6.19877160e-01 6.69407547e-01 3.44511718e-01
5.20295978e-01 -6.71437800e-01 8.92849326e-01 1.77307054e-01
-2.03022167e-01 -8.50540176e-02 -9.88878429e-01 3.81028026e-01
-2.32038188e+00 -1.18387341e+00 5.54492176e-01 2.52588320e+00
9.06814814e-01 4.67850566e-01 2.54754156e-01 -3.01389664e-01
4.77574140e-01 9.14849266e-02 -1.42885351e+00 -5.10612428e-01
3.77688438e-01 1.83251854e-02 7.52207100e-01 7.48432100e-01
-9.11872029e-01 1.09981573e+00 6.98072338e+00 4.74724591e-01
-8.60127449e-01 -2.21382882e-02 4.88882363e-01 -3.63780230e-01
-1.32600695e-01 2.99069881e-01 -7.11431563e-01 5.91337502e-01
1.18337727e+00 -3.48762363e-01 9.46456850e-01 8.25879812e-01
7.49216914e-01 -8.32092106e-01 -1.04601467e+00 3.63787115e-01
-6.92162871e-01 -1.12714469e+00 -5.31706274e-01 1.74115434e-01
8.93388093e-01 2.72025913e-01 -1.38088912e-01 6.87388301e-01
1.47380233e+00 -8.23790610e-01 9.44176555e-01 4.69252169e-01
3.37510109e-01 -8.64281535e-01 7.01427221e-01 9.57606852e-01
-1.03491569e+00 -5.16824305e-01 -3.30809355e-01 -4.83896911e-01
3.57380539e-01 1.49777830e-01 -1.04017961e+00 3.86771858e-01
4.55630690e-01 3.83286357e-01 -6.83385581e-02 1.00981700e+00
-4.91297215e-01 7.54896820e-01 -5.01305401e-01 -4.05908674e-01
6.93510354e-01 -3.41881633e-01 8.58851016e-01 5.38701117e-01
3.51767421e-01 1.91763505e-01 8.19405437e-01 8.81488085e-01
7.24042714e-01 -3.43424827e-01 -7.55002618e-01 -2.41479673e-03
7.83054829e-01 5.61105847e-01 -6.53442621e-01 -4.09049541e-01
1.66512817e-01 6.38320804e-01 2.94636369e-01 5.33850193e-01
-6.13848627e-01 9.92999151e-02 7.27288663e-01 -2.69222021e-01
2.71279097e-01 -4.09733713e-01 1.07551090e-01 -9.21241403e-01
-4.06195641e-01 -9.41484332e-01 2.26903334e-01 -5.46271205e-01
-7.34726548e-01 1.83105156e-01 3.29198390e-01 -1.29028296e+00
-8.94455612e-01 -6.42779022e-02 -6.18161023e-01 6.57538712e-01
-1.49657607e+00 -4.52263385e-01 4.42779988e-01 3.77203524e-01
8.16453636e-01 8.57412592e-02 6.25591099e-01 -5.93836784e-01
-6.05518699e-01 -2.12173909e-01 4.90456611e-01 -5.12586772e-01
3.46664518e-01 -1.53612959e+00 4.48637068e-01 8.83988857e-01
-1.89854339e-01 2.01380283e-01 1.05193782e+00 -1.08669209e+00
-1.45216095e+00 -9.64972973e-01 1.49665147e-01 -2.21200064e-01
7.42931783e-01 3.09383273e-02 -7.44586885e-01 6.72527015e-01
-1.95011795e-02 -3.42654347e-01 1.47866637e-01 -8.61345753e-02
3.40587199e-01 1.36136875e-01 -1.22971249e+00 8.65657747e-01
8.46696913e-01 -1.12478912e-01 -5.26597977e-01 2.12119594e-01
6.30697489e-01 -7.33390152e-01 -5.71800470e-01 1.29426450e-01
4.65675980e-01 -8.15925717e-01 5.66529393e-01 -7.45121777e-01
-7.15883225e-02 -4.75531399e-01 -1.26789659e-01 -1.89281344e+00
-1.78554296e-01 -1.25687361e+00 -2.59494931e-01 6.23335481e-01
3.39774787e-01 -8.50600123e-01 5.89486539e-01 6.73648000e-01
6.96412250e-02 -8.39968920e-01 -1.18771327e+00 -9.74870145e-01
6.33610561e-02 -6.44815564e-01 7.49740124e-01 4.43464547e-01
-6.93563148e-02 -7.98941851e-02 -3.35197955e-01 3.25644702e-01
7.50401020e-01 1.64638534e-01 7.76930392e-01 -7.91144967e-01
-6.63241863e-01 -1.85295999e-01 7.75439367e-02 -1.00514889e+00
5.71020059e-02 -2.97649682e-01 5.94400346e-01 -1.54464078e+00
-1.61706164e-01 -5.69903135e-01 -2.97041833e-01 6.15754962e-01
-2.53427058e-01 -8.42622757e-01 4.20048267e-01 2.57573694e-01
-6.80689156e-01 7.73563802e-01 1.52977955e+00 3.49584408e-02
-7.74883687e-01 4.80109274e-01 -3.07051986e-01 8.28880489e-01
8.96435857e-01 -5.35097539e-01 -6.40417278e-01 -2.03441963e-01
1.59351096e-01 5.88349164e-01 -1.44919097e-01 -1.15671146e+00
2.79435843e-01 -1.13399470e+00 1.01698399e-01 -6.57597780e-01
2.22918525e-01 -5.96141577e-01 2.73396313e-01 1.02423048e+00
-5.27269602e-01 2.15282038e-01 2.36669704e-01 1.14064407e+00
1.92045227e-01 -2.79556394e-01 6.89707339e-01 -4.48511124e-01
-8.53012204e-01 3.10683399e-01 -1.13977706e+00 -1.40336733e-02
1.24122691e+00 -1.24581374e-01 2.05987133e-02 -6.58062577e-01
-8.21424305e-01 8.61915946e-01 4.89018649e-01 1.97194815e-02
5.59140980e-01 -1.01238632e+00 -4.18183774e-01 -1.59136817e-01
-4.75488901e-01 2.68036246e-01 5.38598234e-03 5.82719445e-01
-7.91413561e-02 2.15088457e-01 -6.42023683e-02 -3.34925652e-01
-7.55890965e-01 4.71232563e-01 6.62560463e-01 -7.07928896e-01
-5.93945146e-01 3.32980305e-01 -2.08594948e-01 -4.95714515e-01
1.72423542e-01 -4.47224677e-01 -7.26341084e-02 -9.01795998e-02
6.66311860e-01 4.89330202e-01 -3.82210016e-01 8.53591859e-02
-3.99564207e-02 1.90273911e-01 1.55637145e-01 -8.67690921e-01
1.35790467e+00 -1.39462382e-01 3.46535951e-01 6.48584962e-01
1.58633202e-01 -2.95598000e-01 -2.10999060e+00 -3.66730362e-01
2.93334991e-01 -6.32429361e-01 3.44528645e-01 -1.04075146e+00
-3.80592227e-01 3.23962003e-01 5.32175839e-01 1.01392597e-01
7.88116217e-01 -4.15392488e-01 5.32415926e-01 7.92133808e-01
1.04107165e+00 -1.52282357e+00 6.70897737e-02 6.66989803e-01
7.01852083e-01 -1.08903897e+00 2.79175434e-02 4.16213602e-01
-9.94297266e-01 8.70449603e-01 6.75294399e-01 -1.20139606e-02
3.05752307e-01 3.19627881e-01 -1.42412916e-01 1.56204730e-01
-1.34338820e+00 -6.05992973e-01 -3.84528041e-01 7.24253237e-01
-5.68222940e-01 4.69782412e-01 1.41731143e-01 4.15816195e-02
-6.37333766e-02 1.33685991e-01 5.76526940e-01 1.51756442e+00
-8.98756683e-01 -1.13846874e+00 -6.81150973e-01 5.63535929e-01
-2.45394260e-02 2.95350254e-01 -1.28505930e-01 6.94413304e-01
2.38067582e-02 9.99018610e-01 6.87615722e-02 -2.47502521e-01
2.15693444e-01 -5.02644992e-03 4.49971497e-01 -6.07952118e-01
-3.34409952e-01 3.37406486e-01 2.50781089e-01 -1.05863380e+00
-2.34007031e-01 -1.32888210e+00 -1.41970408e+00 -2.22829446e-01
-5.40570356e-02 1.95655808e-01 6.07624292e-01 1.13256395e+00
2.18765542e-01 5.01010537e-01 8.99239600e-01 -1.12675023e+00
-1.26524067e+00 -7.03573942e-01 -4.30092245e-01 -3.20938379e-01
5.00143230e-01 -1.00685036e+00 -2.07453489e-01 -5.32962441e-01] | [4.261341571807861, 2.0276641845703125] |
44d93468-4231-4928-ba02-ac3fee91012e | hybrid-subspace-learning-for-high-dimensional | 1808.01687 | null | http://arxiv.org/abs/1808.01687v1 | http://arxiv.org/pdf/1808.01687v1.pdf | Hybrid Subspace Learning for High-Dimensional Data | The high-dimensional data setting, in which p >> n, is a challenging
statistical paradigm that appears in many real-world problems. In this setting,
learning a compact, low-dimensional representation of the data can
substantially help distinguish signal from noise. One way to achieve this goal
is to perform subspace learning to estimate a small set of latent features that
capture the majority of the variance in the original data. Most existing
subspace learning models, such as PCA, assume that the data can be fully
represented by its embedding in one or more latent subspaces. However, in this
work, we argue that this assumption is not suitable for many high-dimensional
datasets; often only some variables can easily be projected to a
low-dimensional space. We propose a hybrid dimensionality reduction technique
in which some features are mapped to a low-dimensional subspace while others
remain in the original space. Our model leads to more accurate estimation of
the latent space and lower reconstruction error. We present a simple
optimization procedure for the resulting biconvex problem and show synthetic
data results that demonstrate the advantages of our approach over existing
methods. Finally, we demonstrate the effectiveness of this method for
extracting meaningful features from both gene expression and video background
subtraction datasets. | ['Micol Marchetti-Bowick', 'Benjamin J. Lengerich', 'Ankur P. Parikh', 'Eric P. Xing'] | 2018-08-05 | null | null | null | null | ['video-background-subtraction'] | ['computer-vision'] | [ 1.35000214e-01 -3.42414260e-01 -1.01109460e-01 -1.92273736e-01
-7.23482311e-01 -7.94124007e-01 4.65430439e-01 -3.45089942e-01
-1.36521384e-01 5.70579708e-01 3.80572408e-01 4.64670062e-02
-3.54974806e-01 -3.21826696e-01 -5.22712171e-01 -1.17906857e+00
9.50279552e-03 3.23011220e-01 -1.18664637e-01 1.05573505e-01
-5.92615642e-02 7.97005594e-01 -1.41724968e+00 9.94223729e-03
5.55670738e-01 7.19279110e-01 -2.60163434e-02 5.30719221e-01
-2.24167779e-02 1.63580537e-01 -3.24987829e-01 1.46521524e-01
6.27115905e-01 -3.63620728e-01 -3.41765940e-01 3.33686680e-01
2.86041468e-01 -6.20925725e-02 -4.79086727e-01 1.21864533e+00
3.91658545e-01 1.73509978e-02 7.50173450e-01 -1.38652062e+00
-6.01360127e-02 8.42932835e-02 -6.81617022e-01 5.19241579e-02
3.28036360e-02 -1.93013906e-01 6.97017193e-01 -9.80462551e-01
5.75747907e-01 1.27468801e+00 4.86627787e-01 3.63320440e-01
-1.89519727e+00 -5.78345895e-01 -1.21798404e-01 -5.93996383e-02
-1.58435261e+00 -6.09649718e-01 1.08164704e+00 -7.15763807e-01
1.97370157e-01 5.96976101e-01 4.45245445e-01 1.06556642e+00
-4.89145294e-02 7.80997813e-01 1.20309889e+00 -4.14707661e-01
3.99389863e-01 1.82918012e-01 4.02304918e-01 2.89972126e-01
3.74036431e-01 -1.31127983e-01 -4.91806895e-01 -6.79759920e-01
7.09986031e-01 3.01348209e-01 -5.47475457e-01 -1.20795286e+00
-1.26755309e+00 9.39128339e-01 -2.11061120e-01 2.94388950e-01
-4.03136820e-01 -1.67418525e-01 1.97735325e-01 1.39953449e-01
4.36883599e-01 2.89616853e-01 -3.43241721e-01 -2.25047484e-01
-1.03262854e+00 3.07555776e-02 8.17875981e-01 7.20672846e-01
6.68947935e-01 -1.28728554e-01 -1.19690612e-01 8.24135780e-01
2.38647327e-01 4.21666116e-01 5.41103542e-01 -1.08651114e+00
1.49330676e-01 6.04038060e-01 2.86771536e-01 -1.12250519e+00
-4.06919628e-01 -3.44130546e-01 -1.08471727e+00 1.88766271e-01
5.87463737e-01 -8.44034404e-02 -6.16642952e-01 1.86330664e+00
4.35274035e-01 4.87792403e-01 2.55994499e-01 9.26157355e-01
1.87213421e-01 6.65718973e-01 -4.70141560e-01 -7.29514182e-01
1.06895792e+00 -4.78297681e-01 -7.25649595e-01 -7.87553098e-03
3.79026532e-01 -4.86893296e-01 9.91913795e-01 5.09236813e-01
-6.15938962e-01 -1.35236964e-01 -9.85768139e-01 2.21753255e-01
5.36165573e-02 4.14889604e-01 6.24328852e-01 6.22168601e-01
-6.74758911e-01 4.60596859e-01 -1.05132174e+00 -4.25254703e-01
2.72880673e-01 4.12151873e-01 -6.12266898e-01 -1.73135981e-01
-7.07232594e-01 1.24370769e-01 9.70334932e-02 4.05327566e-02
-6.34806335e-01 -3.13466460e-01 -5.63393891e-01 1.58655286e-01
5.44115663e-01 -3.96101713e-01 6.71028018e-01 -6.60430014e-01
-1.24013805e+00 5.29811800e-01 -4.88890082e-01 -1.09281391e-01
3.52359593e-01 -3.02566916e-01 -1.82921052e-01 1.69484869e-01
-7.23360181e-02 7.54054487e-02 1.17626357e+00 -1.21224988e+00
-1.64356738e-01 -7.57053554e-01 -3.46448958e-01 1.82918951e-01
-7.46196210e-01 2.58664191e-02 -7.69083917e-01 -6.63639247e-01
7.62410581e-01 -1.31673896e+00 -3.60267907e-01 2.25407660e-01
-3.71857941e-01 1.26902387e-01 9.85777080e-01 -4.03285652e-01
1.16552353e+00 -2.56543875e+00 6.41459644e-01 4.14850682e-01
4.49123502e-01 1.11582436e-01 -3.10044996e-02 3.55613738e-01
-3.17459166e-01 -1.46833166e-01 -1.85574040e-01 -2.40886897e-01
-1.38091922e-01 3.33827406e-01 -5.85807383e-01 9.27040696e-01
-4.18997109e-02 4.81907398e-01 -7.13629007e-01 -2.09731996e-01
7.46376887e-02 4.76851404e-01 -4.21176881e-01 1.67400435e-01
3.01912904e-01 3.85097802e-01 -5.41116714e-01 2.97657102e-01
7.69665778e-01 -1.97349712e-01 3.53458643e-01 -2.91371703e-01
-1.36814369e-02 -2.74260312e-01 -1.74467492e+00 1.61282396e+00
2.48108581e-02 6.78960145e-01 3.26259673e-01 -1.33942771e+00
8.04961920e-01 2.44788378e-01 9.46775794e-01 -1.12834208e-01
2.75905970e-02 6.35075197e-02 -6.88617527e-02 -4.11571234e-01
4.65964191e-02 -8.41838717e-02 1.10530585e-01 4.11852211e-01
-1.23330310e-01 2.05775529e-01 5.92654310e-02 1.50002569e-01
1.08039188e+00 -1.89104274e-01 4.53214347e-01 -3.50936949e-01
5.67628503e-01 -3.20790201e-01 9.23103094e-01 5.91945112e-01
-2.87665069e-01 6.94867432e-01 8.74582112e-01 -2.48053700e-01
-9.86664951e-01 -8.20328712e-01 -3.92034620e-01 5.49419165e-01
-3.51673216e-02 -4.90373880e-01 -7.30664968e-01 -6.00142896e-01
2.42891666e-02 3.98568690e-01 -4.46715027e-01 -2.50344098e-01
-3.84368479e-01 -9.74355817e-01 2.38326401e-01 3.28724325e-01
1.41126951e-02 -2.14246377e-01 -3.62316817e-01 -6.04417548e-02
-1.35325387e-01 -1.02145076e+00 -3.31100643e-01 3.33155841e-01
-1.16610348e+00 -9.81507719e-01 -6.15005672e-01 -4.14371133e-01
8.27456474e-01 6.92317784e-01 5.83725035e-01 -4.18324172e-01
-2.08611473e-01 5.02640188e-01 -9.20968130e-02 -2.29069814e-01
-2.14103639e-01 -3.95359933e-01 6.18433118e-01 4.19148564e-01
6.92706048e-01 -5.99772394e-01 -3.01586121e-01 4.16931391e-01
-9.14359152e-01 1.56032005e-02 4.81241554e-01 1.07514894e+00
8.18007290e-01 3.71091366e-01 1.65960684e-01 -7.76690841e-01
4.20445114e-01 -4.30531174e-01 -6.71729088e-01 3.86721641e-02
-4.20422405e-01 3.43134373e-01 7.10552812e-01 -5.94950378e-01
-4.93517846e-01 5.85652053e-01 2.80140281e-01 -9.88564193e-01
-1.33902431e-01 4.05327559e-01 -5.54368913e-01 4.59230319e-02
6.59382999e-01 5.17355800e-01 2.99970895e-01 -8.13802004e-01
3.18091691e-01 7.47900426e-01 2.70641565e-01 -4.71066386e-01
1.01953804e+00 7.35230744e-01 3.98041576e-01 -1.07524204e+00
-5.95086098e-01 -9.24072504e-01 -9.37012374e-01 1.53929517e-01
4.63104159e-01 -1.02954662e+00 -4.26209271e-01 2.11977109e-01
-6.56942189e-01 2.03573465e-01 -5.43375611e-01 7.64389873e-01
-6.69264376e-01 5.92449665e-01 -1.34525597e-01 -9.91517782e-01
1.67803571e-01 -1.19767201e+00 1.16285336e+00 -1.85249031e-01
-1.47220179e-01 -6.92782164e-01 3.18400621e-01 1.07041493e-01
4.53831442e-02 2.35410899e-01 9.44683790e-01 -6.30601346e-01
-2.91909039e-01 -3.78988206e-01 -9.98575706e-03 4.29682046e-01
2.55539089e-01 -2.40918472e-01 -1.03924787e+00 -6.21339560e-01
5.15155435e-01 -4.75153215e-02 9.45347428e-01 4.86186087e-01
1.26916468e+00 -2.15219036e-01 -5.69174469e-01 7.89181530e-01
1.31044543e+00 -1.41586829e-02 3.28904390e-01 2.24371441e-02
6.85165465e-01 5.67958832e-01 4.81723934e-01 4.71656531e-01
-1.95987463e-01 9.52187598e-01 1.67522267e-01 -4.73404415e-02
1.77092403e-01 -7.89325088e-02 3.83438736e-01 8.97151709e-01
2.50464976e-01 1.58243012e-02 -7.84436882e-01 3.47417414e-01
-2.03978610e+00 -8.78508091e-01 -4.28393632e-02 2.55362892e+00
5.78775048e-01 -1.76988989e-01 2.92304516e-01 4.23919648e-01
6.55229986e-01 -6.57553896e-02 -7.39834487e-01 2.51997888e-01
-4.13747162e-01 -3.44935745e-01 3.32048625e-01 1.89609945e-01
-1.27785861e+00 4.94184136e-01 6.63018894e+00 6.48649275e-01
-1.16078269e+00 -1.88006938e-01 4.36235309e-01 -2.34290138e-01
-2.31371343e-01 8.11600313e-02 -6.91375494e-01 3.44829410e-01
9.32679713e-01 -3.75095010e-01 4.27199751e-01 1.01028907e+00
4.46654290e-01 1.91222101e-01 -1.18562853e+00 1.40763342e+00
1.92878395e-01 -1.07481480e+00 5.86725324e-02 5.20128310e-01
6.33482754e-01 -3.02808136e-01 4.99184430e-03 9.77112874e-02
-1.24848105e-01 -8.08249950e-01 2.78379589e-01 5.52898765e-01
8.02588999e-01 -5.74952006e-01 3.31111073e-01 7.47814298e-01
-8.09765100e-01 -2.99972266e-01 -7.74448574e-01 3.63320820e-02
-2.01674417e-01 8.57934773e-01 -6.37936950e-01 2.99916565e-01
5.02065778e-01 8.44322562e-01 -5.30773282e-01 1.04348719e+00
1.50009528e-01 7.70353079e-01 -5.11647999e-01 2.51275241e-01
-1.25339285e-01 -7.74310708e-01 9.92132306e-01 8.86650205e-01
5.31893015e-01 2.83257484e-01 2.32799068e-01 6.39195800e-01
4.64618616e-02 3.00306946e-01 -8.69855106e-01 -3.16207826e-01
2.51874447e-01 1.33068454e+00 -7.35399961e-01 -1.11394502e-01
-6.64409637e-01 9.21441019e-01 2.79549241e-01 5.27783394e-01
-4.79388893e-01 -1.37986943e-01 9.67560053e-01 1.94853880e-02
2.93483227e-01 -5.17535150e-01 -8.46655667e-02 -1.66118526e+00
9.63324234e-02 -9.91661012e-01 3.16718847e-01 -4.47014451e-01
-1.10073316e+00 2.28332996e-01 -1.02989182e-01 -1.57436502e+00
-3.47201914e-01 -6.56261027e-01 -2.44112536e-01 8.56711864e-01
-9.86899197e-01 -7.72917688e-01 -1.56856462e-01 6.94443166e-01
3.79689366e-01 -4.15147752e-01 9.84546900e-01 1.34768143e-01
-7.34397829e-01 2.91399926e-01 1.05381548e+00 -1.92660186e-02
7.17676163e-01 -1.27627361e+00 -9.56867933e-02 9.96972620e-01
5.71718454e-01 9.00587857e-01 7.90790975e-01 -4.22282279e-01
-1.87496245e+00 -9.03157055e-01 3.71083170e-01 -4.97504264e-01
5.92058778e-01 -7.19263136e-01 -1.04201627e+00 6.52349770e-01
-4.84177142e-01 2.90160686e-01 1.10230887e+00 2.83469081e-01
-3.07614177e-01 -1.74774334e-01 -9.03850317e-01 6.34495914e-01
7.56628990e-01 -4.39083606e-01 -4.50542480e-01 3.52267683e-01
3.49790752e-01 -4.59311530e-02 -6.76034331e-01 1.68997154e-01
6.78450227e-01 -7.33458996e-01 9.69683468e-01 -9.73778188e-01
4.42589335e-02 -5.77221751e-01 -6.02180898e-01 -1.28178775e+00
-5.50796509e-01 -6.75242901e-01 -5.43411314e-01 1.07714665e+00
-1.52386934e-01 -2.89029896e-01 1.07408488e+00 6.50731266e-01
3.61322731e-01 -5.36850810e-01 -1.13098788e+00 -9.41769958e-01
-2.86584496e-02 -5.14841557e-01 3.55864167e-01 1.06988919e+00
5.27809970e-02 4.27064210e-01 -4.99907374e-01 2.80673444e-01
9.18863297e-01 3.76198977e-01 1.08845222e+00 -1.48180950e+00
-3.14028949e-01 -3.54352817e-02 -8.37503672e-01 -1.10667229e+00
3.26137960e-01 -7.78994560e-01 -1.61391079e-01 -1.02804911e+00
7.19464481e-01 -2.06636339e-01 -3.61613274e-01 2.92000651e-01
-1.15597337e-01 1.75006881e-01 8.20721611e-02 5.98670483e-01
-2.80980796e-01 6.45360708e-01 8.41764331e-01 -1.35566533e-01
-4.04842287e-01 7.54569173e-02 -8.68049502e-01 8.19417834e-01
4.12197858e-01 -4.67444390e-01 -4.02782559e-01 -1.07494093e-01
-9.39374194e-02 3.35395113e-02 1.62669837e-01 -8.72323394e-01
9.19606909e-02 -2.97778845e-01 6.23778403e-01 -4.36794341e-01
5.39475918e-01 -1.14940453e+00 3.48362356e-01 1.76659390e-01
-1.57830939e-01 -4.76329327e-01 1.11271448e-01 8.54541659e-01
-2.37131447e-01 -1.26319408e-01 7.52538502e-01 2.52961487e-01
-3.82915676e-01 1.33786649e-01 -2.33943477e-01 -2.79794127e-01
1.09960485e+00 -2.65480638e-01 2.41090767e-02 -5.13735592e-01
-8.79582107e-01 6.42203987e-02 5.96942306e-01 3.21751475e-01
5.35190046e-01 -1.59185445e+00 -6.55800462e-01 6.15449727e-01
2.08180845e-01 -3.39595079e-01 1.31318241e-01 1.00576103e+00
-3.07986513e-02 5.77431619e-01 -1.04384966e-01 -1.00566304e+00
-1.52316320e+00 9.35659587e-01 5.33645228e-02 7.76724666e-02
-7.17796504e-01 4.11516041e-01 5.96426427e-01 -1.31076172e-01
5.11140861e-02 3.63146216e-02 -3.02206933e-01 1.31070197e-01
6.37571156e-01 5.42975307e-01 -2.53449589e-01 -9.95347142e-01
-2.10312173e-01 6.07317150e-01 6.71367347e-02 -6.84258416e-02
1.51048720e+00 -2.89403737e-01 -2.21577018e-01 7.17708588e-01
1.52019715e+00 3.37362558e-01 -1.33937812e+00 -4.35534894e-01
-4.62274207e-03 -8.13566625e-01 2.52129257e-01 -7.86368176e-02
-8.46637011e-01 8.55590582e-01 8.39721918e-01 9.19140354e-02
1.31320333e+00 -1.84093025e-02 2.00511366e-01 6.11442149e-01
2.67000794e-01 -9.75260675e-01 -1.58089682e-01 1.80869833e-01
8.71109962e-01 -1.22667730e+00 1.70693353e-01 -5.72833300e-01
-4.11647052e-01 1.19854891e+00 2.14161798e-01 -5.02306968e-02
6.92679822e-01 1.36134118e-01 -1.18901066e-01 -1.53612867e-01
-6.86053872e-01 -1.36191363e-03 3.81509602e-01 4.11139518e-01
1.96090981e-01 3.58592384e-02 -2.44417816e-01 6.88839555e-01
-1.00355735e-02 -1.56326160e-01 6.59116387e-01 7.45784163e-01
-3.23416829e-01 -1.14940667e+00 -7.47518659e-01 5.82007349e-01
-3.81787747e-01 2.42799237e-01 -4.85675901e-01 4.40256178e-01
-2.71234155e-01 6.47856057e-01 -1.46561548e-01 -1.54565066e-01
3.43271643e-01 3.47333848e-01 2.44342625e-01 -6.03675246e-01
3.89891922e-01 8.40261161e-01 -4.14687306e-01 -7.87341893e-01
-3.42028618e-01 -1.12674785e+00 -8.17860305e-01 1.44597322e-01
-3.14550161e-01 3.08493942e-01 6.47777021e-01 6.73510551e-01
4.40815121e-01 1.26364425e-01 8.74555886e-01 -6.88754201e-01
-9.92232859e-01 -6.08972013e-01 -1.20135200e+00 6.22965813e-01
4.36605394e-01 -7.93700159e-01 -5.41054189e-01 2.18927860e-01] | [7.746199607849121, 4.294505596160889] |
451e5f92-6153-4131-a51d-abe260332662 | dropkey-for-vision-transformer | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_DropKey_for_Vision_Transformer_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_DropKey_for_Vision_Transformer_CVPR_2023_paper.pdf | DropKey for Vision Transformer | In this paper, we focus on analyzing and improving the dropout technique for self-attention layers of Vision Transformer, which is important while surprisingly ignored by prior works. In particular, we conduct researches on three core questions: First, what to drop in self-attention layers? Different from dropping attention weights in literature, we propose to move dropout operations forward ahead of attention matrix calculation and set the Key as the dropout unit, yielding a novel dropout-before-softmax scheme. We theoretically verify that this scheme helps keep both regularization and probability features of attention weights, alleviating the overfittings problem to specific patterns and enhancing the model to globally capture vital information; Second, how to schedule the drop ratio in consecutive layers? In contrast to exploit a constant drop ratio for all layers, we present a new decreasing schedule that gradually decreases the drop ratio along the stack of self-attention layers. We experimentally validate the proposed schedule can avoid overfittings in low-level features and missing in high-level semantics, thus improving the robustness and stableness of model training; Third, whether need to perform structured dropout operation as CNN? We attempt patch-based block-version of dropout operation and find that this useful trick for CNN is not essential for ViT. Given exploration on the above three questions, we present the novel DropKey method that regards Key as the drop unit and exploits decreasing schedule for drop ratio, improving ViTs in a general way. Comprehensive experiments demonstrate the effectiveness of DropKey for various ViT architectures, e.g. T2T, VOLO, CeiT and DeiT, as well as for various vision tasks, e.g., image classification, object detection, human-object interaction detection and human body shape recovery. | ['Luoqi Liu', 'Tiande Guo', 'Xiangjian Jiang', 'Congying Han', 'Xuecheng Nie', 'Yinhan Hu', 'Bonan Li'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['human-object-interaction-detection'] | ['computer-vision'] | [ 6.28599972e-02 5.70381880e-02 1.18607311e-02 -1.72825456e-01
-9.13106278e-02 -2.11593300e-01 -1.47638544e-02 -1.75322801e-01
-6.13932312e-01 5.53278506e-01 5.55361509e-02 -1.25841796e-02
-1.21533265e-02 -6.26086950e-01 -1.03008282e+00 -8.00671458e-01
3.20246458e-01 -6.80225715e-02 7.48026013e-01 -3.29262882e-01
4.40108106e-02 2.99392581e-01 -1.57326388e+00 1.05916746e-01
9.48728204e-01 1.05181336e+00 3.38432282e-01 2.60591358e-01
1.02902576e-01 9.94114816e-01 -5.66420972e-01 -6.06020749e-01
1.11044765e-01 -5.06803751e-01 -6.34576201e-01 1.48039818e-01
5.05546570e-01 -4.80020404e-01 -4.93652254e-01 1.13277066e+00
6.91116035e-01 2.77125202e-02 5.51800191e-01 -1.12047052e+00
-1.04040444e+00 5.49727976e-01 -7.95412183e-01 5.24751425e-01
-1.42377108e-01 4.47969466e-01 8.10117424e-01 -6.66350782e-01
3.19296271e-01 1.16443253e+00 7.16112256e-01 7.32983649e-01
-8.41235876e-01 -5.89435637e-01 5.23537040e-01 4.36417818e-01
-1.23078871e+00 -1.12504959e-01 8.41563106e-01 -2.51833498e-01
6.72758162e-01 1.59765139e-01 6.42883062e-01 1.14054823e+00
2.67092347e-01 1.18118024e+00 8.20318282e-01 -2.13550925e-01
-1.17020056e-01 1.33514598e-01 5.44330359e-01 9.69122946e-01
3.71204495e-01 -2.21362710e-01 -3.31327260e-01 2.82875985e-01
9.82233524e-01 8.83733556e-02 -4.97674912e-01 -2.58641466e-02
-9.38498318e-01 5.73331475e-01 6.53906107e-01 3.04014832e-01
-3.42010498e-01 2.62381494e-01 4.20909852e-01 3.25388819e-01
3.62193793e-01 1.05391644e-01 -5.95938861e-01 2.26004288e-01
-7.72825897e-01 2.40220148e-02 3.48756462e-01 1.09072268e+00
6.66571200e-01 4.65773761e-01 -6.05534196e-01 8.90745521e-01
-2.88072173e-02 3.70466173e-01 5.37655711e-01 -5.26025951e-01
3.18995386e-01 6.80823743e-01 -2.37113625e-01 -8.58530521e-01
-4.29280937e-01 -8.37156117e-01 -1.16099966e+00 1.19338483e-01
2.86492825e-01 -3.80613297e-01 -1.21930981e+00 1.94104815e+00
3.16427231e-01 3.49551439e-01 -3.48221660e-01 1.07080066e+00
1.14833486e+00 2.90333360e-01 8.94940123e-02 -6.18058629e-02
1.58586347e+00 -1.09472227e+00 -7.82526731e-01 -1.78679526e-01
4.42972988e-01 -4.59334880e-01 1.28038919e+00 2.68889934e-01
-1.28674006e+00 -9.53133047e-01 -1.10001361e+00 -3.91103148e-01
-1.16859294e-01 4.22571629e-01 6.32824719e-01 3.74781042e-01
-1.06263483e+00 8.52057099e-01 -8.71539056e-01 -2.48058647e-01
5.78647017e-01 5.51370203e-01 -2.56587356e-01 1.83333442e-01
-1.26436782e+00 8.68872285e-01 1.40779883e-01 4.62429076e-01
-7.50265479e-01 -6.93249464e-01 -8.25017452e-01 2.43755579e-01
4.29678380e-01 -9.42944348e-01 9.35787320e-01 -1.19295621e+00
-1.41818821e+00 6.58376098e-01 -2.67783403e-01 -5.59869349e-01
4.60114241e-01 -3.77874076e-01 -3.36453132e-02 3.27057354e-02
-1.76534653e-01 6.73768401e-01 1.14903140e+00 -1.03567243e+00
-5.82546115e-01 -3.89513522e-01 2.08088279e-01 1.12896204e-01
-7.08243549e-01 6.29543280e-03 -5.72822750e-01 -7.32754469e-01
1.26243159e-01 -6.54479086e-01 -1.24704011e-01 1.52741268e-01
-3.01680505e-01 -4.21578586e-01 9.66007352e-01 -7.34214604e-01
1.36943734e+00 -2.25604010e+00 2.21070170e-01 -1.49099916e-01
3.11938375e-01 5.31649351e-01 -7.09576756e-02 -1.89142689e-01
-1.07708767e-01 9.61501002e-02 -3.89781326e-01 -5.25712729e-01
-1.99943319e-01 4.05488998e-01 -9.37210396e-02 4.52100992e-01
7.51085997e-01 1.04368043e+00 -7.03584313e-01 -6.42609715e-01
1.48603916e-01 6.47987366e-01 -7.17222452e-01 1.27669469e-01
6.70584664e-03 3.64666075e-01 -4.85883087e-01 6.02021456e-01
8.07858348e-01 -2.15542063e-01 -4.13142115e-01 -5.34184456e-01
-1.20324768e-01 1.14743903e-01 -1.16902196e+00 1.31112778e+00
-2.32140601e-01 4.31773365e-01 2.90242106e-01 -1.11494982e+00
7.94336498e-01 1.15446702e-01 1.30275145e-01 -7.84351230e-01
4.28391814e-01 -5.70366867e-02 1.21254228e-01 -7.74255693e-01
3.20931941e-01 -4.03578579e-02 3.65432531e-01 -5.20166308e-02
3.48371148e-01 4.02661651e-01 3.24198343e-02 -3.61612886e-02
9.57452357e-01 2.46394366e-01 -1.71828002e-01 -3.57413113e-01
6.11856222e-01 -3.97985160e-01 6.69875681e-01 6.57036245e-01
-3.92804652e-01 1.00859189e+00 6.20893240e-01 -2.48963103e-01
-7.91057885e-01 -6.28615022e-01 -3.36547010e-03 1.07755005e+00
1.65029511e-01 -1.04757838e-01 -1.06406021e+00 -7.65082598e-01
-1.24096923e-01 2.52888829e-01 -9.52554643e-01 -5.00002265e-01
-8.04008901e-01 -9.70937610e-01 4.68994677e-01 8.49267006e-01
8.18748415e-01 -1.40289950e+00 -5.44195294e-01 1.51704133e-01
-9.16732773e-02 -9.86338496e-01 -7.40267694e-01 5.12488723e-01
-8.60738814e-01 -1.02349889e+00 -1.07593989e+00 -1.05442989e+00
7.36928046e-01 2.83669293e-01 8.95319819e-01 5.11969268e-01
-1.64643809e-01 1.16432428e-01 -2.97522098e-01 -4.49467748e-01
2.43776411e-01 3.83231848e-01 -2.60687500e-01 2.43533298e-01
1.00283206e-01 -6.69091821e-01 -8.61711323e-01 2.65448749e-01
-9.64889526e-01 -2.40204662e-01 7.51401901e-01 9.58976507e-01
4.16444391e-01 -4.77744676e-02 4.41281825e-01 -6.91743970e-01
3.85206670e-01 -3.04035336e-01 -3.34274739e-01 1.19562171e-01
-1.40907854e-01 1.92233935e-01 7.72365451e-01 -7.39472985e-01
-7.86626875e-01 -1.25005290e-01 -4.84465063e-01 -8.14680934e-01
7.50525445e-02 1.39423102e-01 -3.94841760e-01 -9.95801315e-02
3.35961312e-01 4.26875055e-01 -6.37227893e-02 -6.13682568e-01
1.69482753e-01 3.70213389e-01 5.08120716e-01 -3.88587028e-01
7.78493404e-01 5.09914637e-01 -1.24334946e-01 -8.28564167e-01
-9.42578375e-01 -2.86550224e-01 -5.55163741e-01 4.56016697e-02
9.85545456e-01 -7.67253816e-01 -1.09896863e+00 7.47108638e-01
-1.10651302e+00 -2.94848353e-01 -2.99252450e-01 3.18086654e-01
-1.59742251e-01 3.84202272e-01 -7.55790770e-01 -7.46898413e-01
-6.17246807e-01 -1.18840373e+00 1.00688612e+00 4.35024381e-01
1.02998309e-01 -8.45111728e-01 -4.19946879e-01 1.13870926e-01
4.58653599e-01 8.32451209e-02 8.60759676e-01 -2.77048081e-01
-4.32193875e-01 2.20730558e-01 -4.19055730e-01 7.92081475e-01
6.07522763e-02 -1.08978324e-01 -9.81434345e-01 -4.04636949e-01
4.31626588e-01 -2.46920228e-01 1.39239287e+00 5.98237157e-01
1.38258493e+00 -2.41455972e-01 -1.21912047e-01 9.82324004e-01
1.18820632e+00 -1.24849090e-02 9.56332743e-01 3.26749414e-01
9.59997773e-01 4.77778196e-01 1.34412378e-01 1.78105876e-01
3.27372879e-01 4.90337998e-01 5.50199032e-01 -5.82153201e-01
-5.32109678e-01 -7.89148733e-02 3.70939016e-01 8.40826988e-01
-2.66256303e-01 1.31656490e-02 -3.87886763e-01 6.17834687e-01
-1.70534718e+00 -7.52463400e-01 -2.77800322e-01 2.08866572e+00
8.30727577e-01 3.86784971e-01 3.44678275e-02 3.06938048e-02
7.28723824e-01 2.19182789e-01 -6.47965372e-01 -2.63686657e-01
-3.06882769e-01 5.59444070e-01 5.59353411e-01 2.80803561e-01
-1.07641220e+00 1.04284322e+00 5.38662004e+00 9.66006637e-01
-1.28928554e+00 3.14189792e-01 5.80170035e-01 -7.40534812e-02
-8.62378180e-02 -4.12597694e-02 -1.04667318e+00 6.36794269e-01
3.77322406e-01 3.52161437e-01 1.71153829e-01 7.12914944e-01
1.80284113e-01 9.60176066e-02 -9.87510920e-01 8.88958275e-01
1.38172239e-01 -9.65348184e-01 2.89945453e-01 -3.70338351e-01
4.43207830e-01 -1.69035614e-01 1.21790126e-01 4.40654755e-01
-9.91767868e-02 -7.94186473e-01 9.20001686e-01 5.16833246e-01
5.41286767e-01 -6.94844663e-01 8.23340058e-01 2.25675970e-01
-1.14593315e+00 -1.05195552e-01 -5.56980252e-01 -1.99388084e-03
-1.15619153e-01 5.35685956e-01 -1.38105959e-01 7.34854758e-01
8.69470716e-01 7.96567738e-01 -7.03558624e-01 1.15906549e+00
-4.01341438e-01 7.39071369e-01 -3.15679669e-01 -3.34922373e-02
3.73907864e-01 -8.58573020e-02 4.42418277e-01 1.13068306e+00
4.96901572e-02 1.37811884e-01 -1.52331516e-01 9.09429073e-01
-1.79645717e-01 -1.44928455e-01 -1.35453820e-01 5.09516656e-01
-6.33377861e-03 1.17567587e+00 -6.64045572e-01 -2.81685352e-01
-3.50122482e-01 1.22838509e+00 4.17192310e-01 5.25216222e-01
-1.17018974e+00 -4.98445630e-01 5.44437826e-01 3.93589407e-01
7.19119430e-01 -7.11712986e-02 -2.81869859e-01 -1.17805886e+00
3.91995132e-01 -8.01809371e-01 2.22280130e-01 -6.65060520e-01
-1.20299542e+00 7.09546626e-01 -1.26280099e-01 -1.06160045e+00
4.37377214e-01 -5.84879458e-01 -8.87128532e-01 8.43345046e-01
-1.83907163e+00 -1.26867747e+00 -3.55227202e-01 8.99572253e-01
5.90517581e-01 2.50704020e-01 1.05680071e-01 5.81520021e-01
-1.04966819e+00 1.01437759e+00 -2.25120485e-01 2.77001172e-01
6.15794241e-01 -1.03948164e+00 2.42170751e-01 8.95595670e-01
-1.50089845e-01 8.40177357e-01 5.19963384e-01 -5.61718702e-01
-1.31022465e+00 -1.09657872e+00 4.14426684e-01 -2.38676623e-01
4.59441602e-01 -4.38491374e-01 -1.22624171e+00 5.92730522e-01
2.81916946e-01 2.40858927e-01 -6.37954799e-03 -8.48771036e-02
-9.89471972e-02 -2.51377314e-01 -9.88392472e-01 6.36857390e-01
1.19755793e+00 -1.37499586e-01 -6.84533298e-01 1.44762456e-01
9.29341137e-01 -4.52758521e-01 -4.56394225e-01 4.98395056e-01
4.22259927e-01 -8.23803544e-01 8.70544791e-01 -6.14876926e-01
5.10639727e-01 -3.56252193e-01 2.01819822e-01 -9.78986204e-01
-5.66971600e-01 -6.87357783e-01 -2.12825730e-01 1.41690671e+00
2.51592577e-01 -6.19502783e-01 8.81765425e-01 3.15657258e-01
-5.13822079e-01 -1.00094450e+00 -9.21857834e-01 -7.26039052e-01
1.68163583e-01 -7.48813599e-02 2.82159090e-01 7.25326121e-01
-5.55301189e-01 3.18110228e-01 -5.67886293e-01 2.39523783e-01
6.55431986e-01 -3.02945912e-01 6.15074098e-01 -9.83213425e-01
-3.43422651e-01 -4.14330840e-01 -3.17448199e-01 -1.33009887e+00
-7.26645738e-02 -5.29513478e-01 -8.98506790e-02 -1.57385039e+00
2.93463260e-01 -1.71255618e-01 -5.43828189e-01 7.52851844e-01
-6.58201337e-01 3.48050088e-01 2.29209810e-01 1.18199982e-01
-4.44868237e-01 8.10767174e-01 1.58678794e+00 -2.29074284e-01
-1.41550705e-01 -6.09104112e-02 -8.27896297e-01 8.03969264e-01
4.73793745e-01 -4.30067748e-01 -4.22860503e-01 -7.02407598e-01
-1.58385664e-01 -4.27852511e-01 6.28363550e-01 -1.11020708e+00
3.62038642e-01 1.47719026e-01 4.29838717e-01 -5.19749522e-01
8.88568535e-02 -6.72344446e-01 -3.13560367e-01 5.18078864e-01
-2.51725707e-02 -2.87067750e-03 4.78726208e-01 3.74803931e-01
-2.09714428e-01 -3.45578820e-01 9.99451637e-01 -1.82872206e-01
-7.22215176e-01 4.30469066e-01 -7.84320459e-02 1.30531535e-01
6.67242408e-01 -4.50657725e-01 -1.41131133e-01 -7.49387965e-03
-8.06250989e-01 5.01105309e-01 8.82009491e-02 3.54257017e-01
6.01296067e-01 -1.10157728e+00 -7.52848685e-01 3.41071755e-01
-2.79523700e-01 8.26459005e-02 5.68084359e-01 1.12676883e+00
-1.93554565e-01 1.57141045e-01 -2.30635896e-01 -5.60290813e-01
-1.19030166e+00 6.25556886e-01 5.29478610e-01 -2.07848296e-01
-8.22685003e-01 1.14543378e+00 5.90031147e-01 8.49988759e-02
5.28585017e-01 -8.28473926e-01 -3.89416784e-01 1.84971169e-02
4.80681658e-01 6.91603571e-02 1.77272141e-01 -3.44230860e-01
-4.07990336e-01 7.39925802e-01 -3.25980455e-01 5.10889888e-01
1.39593327e+00 -1.83901116e-01 5.71638159e-03 2.65232086e-01
1.09815764e+00 -3.09289396e-01 -1.61692071e+00 -1.76900830e-02
-2.93257535e-01 -8.33936259e-02 1.24735340e-01 -4.35732096e-01
-1.57336140e+00 9.88001883e-01 7.61497557e-01 1.05885535e-01
1.32920194e+00 -7.34418258e-02 1.04110312e+00 8.00149515e-02
1.05090380e-01 -8.78778219e-01 1.62155092e-01 5.45361221e-01
9.35047925e-01 -1.10851383e+00 -1.38165787e-01 -3.22969586e-01
-5.66478491e-01 7.66330600e-01 1.00286829e+00 -3.97296816e-01
5.75197637e-01 3.33616436e-01 -2.12436348e-01 -1.84676290e-01
-5.80197394e-01 -5.07923722e-01 2.54040629e-01 5.60852528e-01
3.91499192e-01 -5.47845423e-01 -3.94223005e-01 8.29755783e-01
4.81033474e-02 5.43868504e-02 3.92085075e-01 8.19994509e-01
-5.16873479e-01 -8.40460598e-01 -2.47211024e-01 3.64357531e-01
-5.93822241e-01 -3.98855090e-01 -1.40706971e-01 8.96464705e-01
4.79868174e-01 5.87309480e-01 -1.43073853e-02 -2.77198672e-01
7.51750052e-01 -9.06753466e-02 4.36226726e-01 -4.41748619e-01
-1.01854539e+00 2.61552520e-02 -2.75177449e-01 -3.65381569e-01
-3.78835142e-01 -3.49670738e-01 -1.07717705e+00 -1.46452919e-01
-6.31835163e-01 -8.00906345e-02 1.57073274e-01 1.06994712e+00
2.70248234e-01 9.60196316e-01 3.24594945e-01 -7.09933102e-01
-7.15533435e-01 -1.10593462e+00 -5.21707296e-01 4.82095301e-01
6.12487316e-01 -9.19944346e-01 -4.43819433e-01 3.99751812e-02] | [9.855834007263184, 0.21868382394313812] |
c9a77000-7269-4f82-aece-d76b3a47ab17 | ca-pmg-channel-attention-and-progressive | null | null | https://www.researchgate.net/publication/351338170_CA-PMG_Channel_attention_and_progressive_multi-granularity_training_network_for_fine-grained_visual_classification | https://www.researchgate.net/publication/351338170_CA-PMG_Channel_attention_and_progressive_multi-granularity_training_network_for_fine-grained_visual_classification | CA-PMG: Channel attention and progressive multi-granularity training network for fine-grained visual classification | Fine-grained visual classification is challenging due to the inherently subtle intra-class object variations. To solve this issue, a novel framework named channel attention and progressive multi-granularity training network, is proposed. It first exploits meaningful feature maps through the channel attention module and captures multi-granularity features by the progressive multi-granularity training module. For each feature map, the channel attention module is proposed to explore channel-wise correlation. This allows themodel to re-weight the channels of the feature map according to the impact of their semantic information on performance. Furthermore, the progressivemulti-granularity trainingmodule is introduced to fuse features cross multi-granularity. And the fused features pay more attention to the subtle differences between images. The model can be trained efficiently in an end-to-end manner without bounding box or part annotations. Finally, comprehensive experiments are conducted to show that the method achieves state-of-the-art performances on the CUB-200-2011, Stanford Cars, and FGVC-Aircraft datasets. Ablation studies demonstrate the effectiveness of each part in our module. | ['Maoguo Gong', 'Ruyi Liu', 'Xiangzeng Liu', 'Hang Yao', 'Qiguang Miao', 'Peipei Zhao'] | 2021-04-21 | null | null | null | iet-image-processing-2021-4 | ['fine-grained-visual-recognition', 'fine-grained-image-classification'] | ['computer-vision', 'computer-vision'] | [-7.86607936e-02 -2.85393745e-01 -2.94199377e-01 -4.40737098e-01
-1.04778802e+00 -4.54070300e-01 4.76758510e-01 -3.73458229e-02
-2.24776760e-01 5.59121966e-01 1.47649646e-01 -1.08511224e-01
-8.55994746e-02 -5.71793437e-01 -8.78644288e-01 -5.04119873e-01
1.25500029e-02 3.22077721e-02 3.22182417e-01 1.55016007e-02
2.61809938e-02 4.41747099e-01 -1.53564703e+00 9.21028793e-01
7.72651672e-01 1.70554113e+00 3.24154407e-01 5.58380544e-01
-1.93928555e-01 6.47921383e-01 -4.19498801e-01 -3.72427136e-01
3.22712302e-01 -3.71102504e-02 -6.66100085e-01 3.22705060e-01
8.84290755e-01 -5.04560411e-01 -1.54451936e-01 9.96183336e-01
1.96768209e-01 -3.19087803e-02 3.83376807e-01 -1.24826586e+00
-6.70725346e-01 4.97102171e-01 -8.86641026e-01 3.73049498e-01
-1.50165349e-01 2.73476958e-01 1.23015344e+00 -1.20021200e+00
2.17630923e-01 1.35232484e+00 5.31559885e-01 1.91241860e-01
-1.15433931e+00 -9.10277069e-01 7.97163904e-01 3.30851734e-01
-1.23711956e+00 -8.96425322e-02 5.77499926e-01 -4.85646248e-01
9.35760617e-01 7.92228281e-02 4.86232638e-01 8.87682319e-01
2.28949964e-01 8.32041144e-01 9.61114347e-01 -1.75527394e-01
-1.99326158e-01 -4.52410839e-02 2.37702861e-01 9.57572043e-01
8.51606056e-02 3.21185648e-01 -4.85971242e-01 1.20975956e-01
7.06339538e-01 4.88670766e-01 -1.67820990e-01 -4.59168434e-01
-1.19111609e+00 8.32958519e-01 1.10752690e+00 1.65399134e-01
-3.26677173e-01 4.48199660e-01 4.96236205e-01 4.61147949e-02
4.60850179e-01 4.14735168e-01 -6.24428213e-01 9.29882899e-02
-7.18419313e-01 4.27336283e-02 8.61576572e-02 1.24992514e+00
9.11603272e-01 -9.23804343e-02 -8.08432102e-01 5.92938304e-01
3.86297971e-01 1.87227935e-01 3.68799627e-01 -6.22776866e-01
7.18508244e-01 6.77042305e-01 -2.22043842e-02 -7.61192322e-01
-3.78564417e-01 -7.03141510e-01 -6.84663296e-01 1.66369319e-01
1.56515017e-01 8.31940211e-03 -1.22932637e+00 1.42084706e+00
3.44231188e-01 3.48322183e-01 -4.58204687e-01 1.10840845e+00
9.44877625e-01 4.32838380e-01 3.87533337e-01 2.83579439e-01
1.50791168e+00 -1.55404353e+00 -3.51456434e-01 -2.24122122e-01
4.37628984e-01 -6.82126641e-01 1.20534229e+00 -1.53144533e-02
-8.17761302e-01 -1.04160583e+00 -1.19614410e+00 -2.11171120e-01
-4.90680128e-01 3.88434798e-01 7.36283779e-01 2.01046348e-01
-6.12783492e-01 5.83332956e-01 -6.82098091e-01 1.88747384e-02
8.32674801e-01 1.55239850e-01 -1.53038844e-01 -1.25277415e-01
-1.00866449e+00 4.96487826e-01 4.16235387e-01 1.20321698e-01
-1.08809018e+00 -9.58629847e-01 -7.91858435e-01 3.55302453e-01
2.46612668e-01 -6.55100286e-01 1.12509859e+00 -9.37574029e-01
-1.16768873e+00 5.24593353e-01 3.77564952e-02 -4.00391191e-01
3.78461629e-01 -3.49878639e-01 -1.94970220e-01 1.13043629e-01
2.61512935e-01 8.33267629e-01 1.09712958e+00 -1.13577366e+00
-1.13544130e+00 -3.58581692e-01 2.43320182e-01 9.22722965e-02
-1.55359104e-01 -1.30220443e-01 -8.26214254e-01 -8.70601773e-01
-2.52572060e-01 -7.64010727e-01 -2.09472626e-01 -8.68073106e-02
-5.24666905e-01 1.83871519e-02 8.04724574e-01 -4.01062846e-01
1.17454505e+00 -2.46621346e+00 4.84942943e-02 5.02268821e-02
3.77844930e-01 6.63477108e-02 -3.36129367e-01 -6.04226142e-02
-6.66111782e-02 1.57445610e-01 9.87588894e-03 -3.66894692e-01
2.01171469e-02 -1.50938511e-01 -2.61043429e-01 2.85555393e-01
5.31455874e-01 1.17434704e+00 -6.44672036e-01 -3.41406018e-01
4.22659427e-01 3.90950322e-01 -6.66782796e-01 3.53078216e-01
-8.92886445e-02 3.74946833e-01 -7.00797319e-01 9.47970152e-01
7.83203840e-01 -5.56280255e-01 -2.04265773e-01 -5.14405191e-01
-6.61037341e-02 -1.26750216e-01 -6.96835697e-01 1.79961061e+00
-7.21575022e-01 4.51228321e-01 -3.43156345e-02 -6.02132022e-01
4.95850205e-01 -3.21559273e-02 6.48374707e-02 -8.71381402e-01
2.59860784e-01 7.50096962e-02 -1.36494031e-02 -2.35773012e-01
4.73165721e-01 6.52867779e-02 -4.12063658e-01 -5.35922088e-02
1.21253476e-01 3.60825300e-01 4.95829992e-02 1.24174900e-01
8.02611411e-01 3.65004665e-03 2.12133199e-01 -1.57457009e-01
4.34293479e-01 -2.42229179e-01 3.27559143e-01 6.12526059e-01
-2.46529177e-01 6.41823828e-01 1.95401683e-01 -5.29474258e-01
-7.42613316e-01 -1.02904272e+00 -1.93405256e-01 1.44717634e+00
3.83926749e-01 -4.52298194e-01 -4.70360428e-01 -1.13049257e+00
3.78548026e-01 2.92190611e-01 -1.19936287e+00 -3.29123616e-01
-9.00610834e-02 -5.16370475e-01 1.94540590e-01 1.12216389e+00
7.52940834e-01 -8.88037622e-01 -8.41430724e-01 2.02146396e-01
-1.11740969e-01 -1.22627652e+00 -5.62176108e-01 2.69510239e-01
-6.56934679e-01 -1.02490938e+00 -6.45408213e-01 -5.33347368e-01
5.42984486e-01 4.34645236e-01 1.12133420e+00 8.33263248e-02
-5.81134677e-01 1.55252650e-01 -5.33613801e-01 -2.84480929e-01
3.47505420e-01 1.96855664e-01 -6.48044348e-01 2.90177077e-01
2.61707097e-01 -5.57882972e-02 -7.15986788e-01 3.53360206e-01
-4.79249001e-01 1.87691320e-02 9.48301792e-01 1.32434666e+00
8.98939192e-01 -2.49403417e-01 4.01943922e-01 -8.69476795e-01
3.43922555e-01 -3.48725289e-01 -5.70765853e-01 2.15982318e-01
-2.72380382e-01 5.55064641e-02 7.11457193e-01 -8.96823853e-02
-8.82713258e-01 8.96823555e-02 8.42989832e-02 -8.84871185e-01
-3.35338771e-01 2.90715426e-01 -3.91691059e-01 -1.23737007e-01
2.82630265e-01 -9.19391140e-02 -6.63770735e-01 -5.91886222e-01
6.26655698e-01 4.66406584e-01 4.00589406e-01 -4.72280055e-01
6.76667094e-01 4.25639182e-01 -3.57146680e-01 -4.10572678e-01
-1.11953592e+00 -5.49854040e-01 -6.45701647e-01 -1.59106012e-02
1.16463554e+00 -1.28901851e+00 -4.03768122e-01 2.66669303e-01
-9.90984321e-01 -2.31624305e-01 -2.71021336e-01 4.17559296e-01
-3.37468743e-01 -1.22965440e-01 -4.65521544e-01 -3.30294430e-01
-2.09125355e-01 -1.35123849e+00 1.47823358e+00 2.84240931e-01
1.85330898e-01 -6.82196438e-01 -5.25009155e-01 2.19134942e-01
3.94446611e-01 2.57803909e-02 8.74207973e-01 -3.51819098e-01
-1.01288760e+00 -1.81026593e-01 -9.07890618e-01 2.39322022e-01
-9.44545772e-03 -1.59727976e-01 -1.19023573e+00 -5.06145716e-01
-6.64528430e-01 -4.58103448e-01 1.42229629e+00 3.45497400e-01
1.86388540e+00 2.41611172e-02 -5.13571680e-01 9.30895329e-01
1.46874619e+00 -8.61999020e-03 3.79235536e-01 3.39173436e-01
1.07343960e+00 2.65885741e-01 8.07631314e-01 5.76002121e-01
4.13155049e-01 7.62766361e-01 6.87874138e-01 -3.62669349e-01
-3.20188403e-01 -1.87143072e-01 -1.16746211e-02 1.55299008e-01
1.34706154e-01 -3.34899873e-02 -5.52786946e-01 5.15057564e-01
-1.69067669e+00 -7.34687150e-01 1.62307426e-01 2.02631402e+00
2.55786061e-01 3.05102587e-01 4.08536978e-02 -1.58884466e-01
5.79947352e-01 3.29369098e-01 -5.07015288e-01 -2.69820720e-01
2.51093596e-01 -1.51844034e-02 5.69206119e-01 2.16964126e-01
-1.36924052e+00 1.12149441e+00 5.42164516e+00 1.04521418e+00
-1.22597015e+00 2.35673770e-01 8.93733680e-01 -1.18472435e-01
-1.27723426e-01 -2.58156568e-01 -9.41879690e-01 6.45868242e-01
3.95888150e-01 1.83637351e-01 3.37955534e-01 9.41998363e-01
-3.39546084e-01 1.23216935e-01 -9.31368947e-01 9.28963006e-01
-2.80782729e-02 -1.27214110e+00 1.01778418e-01 -1.37213439e-01
6.79376364e-01 1.08740367e-01 3.45992923e-01 6.12266421e-01
1.18077092e-01 -1.03940308e+00 8.59272838e-01 3.43270361e-01
1.12430131e+00 -1.06265736e+00 6.63038909e-01 -7.51896277e-02
-1.75985026e+00 -4.59382027e-01 -4.73402202e-01 1.57817900e-01
-2.71483436e-02 5.10598838e-01 -7.07106113e-01 6.49272680e-01
1.04919159e+00 7.27766812e-01 -9.45985377e-01 1.12621486e+00
-1.39215022e-01 3.45352709e-01 7.12463334e-02 1.69278532e-01
6.30034864e-01 4.00574714e-01 8.37505013e-02 1.40730536e+00
2.65982926e-01 -2.55844504e-01 4.77703601e-01 7.10123301e-01
-2.13366792e-01 -1.52097702e-01 -2.51268148e-01 9.88201052e-02
3.25043947e-01 1.58624923e+00 -6.55423462e-01 -4.26136017e-01
-7.41802096e-01 1.01815653e+00 5.99175990e-01 4.70451981e-01
-1.25111759e+00 -5.32353461e-01 7.73405910e-01 2.43144780e-02
1.09471083e+00 1.41812578e-01 -4.02891219e-01 -9.71156001e-01
-9.31284726e-02 -5.87541223e-01 5.48812747e-01 -6.71204984e-01
-1.33362639e+00 9.37782288e-01 -2.41172090e-01 -1.38041711e+00
1.55739725e-01 -6.29549265e-01 -4.45289850e-01 9.07881498e-01
-1.88799262e+00 -1.58372271e+00 -6.57779634e-01 6.60052121e-01
7.18404531e-01 -1.86101213e-01 6.34579718e-01 4.14065242e-01
-6.35208964e-01 1.11312151e+00 -9.33037698e-02 1.19802758e-01
6.65535271e-01 -1.21182084e+00 3.81413519e-01 8.04239035e-01
-7.53592551e-02 2.64078587e-01 2.33440712e-01 -3.82017285e-01
-8.02596927e-01 -1.65789819e+00 5.55350423e-01 -3.11633795e-01
5.12205899e-01 -5.03432155e-01 -7.61245847e-01 6.99336886e-01
1.49825692e-01 6.93887353e-01 6.09027028e-01 3.66811901e-01
-7.96124935e-01 -3.30080658e-01 -7.98029840e-01 1.46877393e-01
1.04821479e+00 -5.45241237e-01 -3.18850189e-01 3.47721055e-02
9.31327283e-01 -4.64165658e-01 -7.02237010e-01 5.58211863e-01
6.62022114e-01 -9.66307461e-01 9.07690704e-01 -7.93606937e-01
6.02424979e-01 -3.97990584e-01 -3.13434750e-01 -1.59431684e+00
-9.13940132e-01 -5.63422963e-02 -2.22134694e-01 1.11564887e+00
6.56811953e-01 -2.97497064e-01 4.21723932e-01 1.99811772e-01
-5.08814692e-01 -9.89577234e-01 -8.13284338e-01 -6.63859785e-01
-6.75153509e-02 -3.05530816e-01 9.26028490e-01 5.42177916e-01
-3.45580995e-01 4.37948793e-01 -4.32172269e-01 2.78096586e-01
4.23369944e-01 8.18597794e-01 7.05351114e-01 -1.00237584e+00
-5.09615362e-01 -4.84496683e-01 -4.90193486e-01 -1.21269143e+00
7.32027665e-02 -8.28600347e-01 1.28049508e-01 -1.29972363e+00
4.21488971e-01 -3.67329717e-01 -7.36695051e-01 5.31646967e-01
-5.68575919e-01 4.94392633e-01 5.38437366e-01 2.54707243e-02
-1.18828440e+00 6.93263292e-01 1.50359368e+00 -3.43582153e-01
1.51469335e-01 -2.37388730e-01 -9.84088242e-01 5.18961906e-01
4.87320304e-01 -2.97051460e-01 -2.78149277e-01 -5.13703644e-01
-4.11778122e-01 -2.55966932e-01 6.83315217e-01 -1.08197689e+00
-5.11861183e-02 6.84882328e-02 9.92815137e-01 -8.13103497e-01
2.08784774e-01 -1.02214956e+00 -4.08327371e-01 1.98080167e-01
-3.65736067e-01 8.69958401e-02 4.44183230e-01 8.14796329e-01
-4.23397005e-01 4.71838921e-01 1.00813246e+00 8.70635267e-03
-1.08581829e+00 6.99000299e-01 1.21757984e-02 9.01231244e-02
1.32105863e+00 5.97062446e-02 -3.06228220e-01 5.51445410e-02
-6.33094490e-01 4.92169440e-01 2.76512682e-01 6.94008172e-01
5.15787542e-01 -1.67882931e+00 -5.92754126e-01 4.86633122e-01
6.87628210e-01 -5.50949760e-02 8.26372921e-01 7.65204549e-01
-8.45826268e-02 6.16338074e-01 -3.90846699e-01 -8.16658437e-01
-1.17211163e+00 9.14081514e-01 3.59756231e-01 -4.95022506e-01
-3.86777520e-01 1.29210639e+00 1.00786793e+00 -8.02528933e-02
2.58284599e-01 -8.01701009e-01 -9.54053998e-02 1.10463180e-01
7.74905026e-01 1.34316936e-01 1.76312372e-01 -6.56955302e-01
-5.02862573e-01 7.11377203e-01 -4.36526358e-01 3.97914708e-01
9.23282564e-01 -2.36379594e-01 5.26042521e-01 2.06284523e-01
1.45937526e+00 -1.17126308e-01 -1.84329474e+00 -1.33603543e-01
-4.67195541e-01 -8.00698459e-01 1.57772958e-01 -1.04980743e+00
-1.34392750e+00 1.18259072e+00 8.17798913e-01 -3.40937302e-02
1.28210700e+00 1.00245759e-01 7.00555444e-01 -1.06548786e-01
2.36406118e-01 -9.28223848e-01 1.81978285e-01 5.62282801e-01
9.99649286e-01 -1.43556440e+00 -2.15430483e-01 -5.49129367e-01
-7.11537540e-01 8.49430382e-01 1.01043308e+00 -1.37620702e-01
7.80316055e-01 1.88715577e-01 -3.49289924e-02 -1.36838809e-01
-8.25962961e-01 -4.50192571e-01 5.77433288e-01 3.83149952e-01
4.10993636e-01 2.39919305e-01 8.60509276e-02 9.63770151e-01
1.03340149e-01 -1.61647633e-01 -2.98089057e-01 6.66230440e-01
-3.95255774e-01 -6.11552477e-01 -1.01582460e-01 5.35352290e-01
-3.39365721e-01 -2.41162196e-01 -3.24712306e-01 7.69274473e-01
5.20700514e-01 7.10389555e-01 3.09205383e-01 -7.47291446e-01
5.59271276e-01 -2.67925829e-01 3.73388052e-01 -4.61690903e-01
-6.90162957e-01 2.86924183e-01 -5.09392358e-02 -9.97229457e-01
-1.38019189e-01 -3.51331174e-01 -1.20371532e+00 -1.05162617e-02
-4.33253050e-01 4.47296649e-02 3.62302721e-01 8.41480792e-01
7.81954527e-01 1.19940805e+00 7.13198900e-01 -1.04772580e+00
-4.48186517e-01 -9.76725876e-01 -4.57828462e-01 4.48770583e-01
4.72724229e-01 -1.00978470e+00 -4.60246950e-02 -2.19501331e-01] | [9.588634490966797, 2.030710458755493] |
02729033-275c-430e-8cf5-7a032e21c104 | maked-multi-lingual-automatic-keyword | null | null | https://aclanthology.org/2022.lrec-1.664 | https://aclanthology.org/2022.lrec-1.664.pdf | MAKED: Multi-lingual Automatic Keyword Extraction Dataset | Keyword extraction is an integral task for many downstream problems like clustering, recommendation, search and classification. Development and evaluation of keyword extraction techniques require an exhaustive dataset; however, currently, the community lacks large-scale multi-lingual datasets. In this paper, we present MAKED, a large-scale multi-lingual keyword extraction dataset comprising of 540K+ news articles from British Broadcasting Corporation News (BBC News) spanning 20 languages. It is the first keyword extraction dataset for 11 of these 20 languages. The quality of the dataset is examined by experimentation with several baselines. We believe that the proposed dataset will help advance the field of automatic keyword extraction given its size, diversity in terms of languages used, topics covered and time periods as well as its focus on under-studied languages. | ['Dwaipayan Roy', 'Adam Jatowt', 'Sriparna Saha', 'Anubhav Jangra', 'Yash Verma'] | null | null | null | null | lrec-2022-6 | ['keyword-extraction'] | ['natural-language-processing'] | [-5.69095314e-01 -4.25864130e-01 -7.77628601e-01 -1.31417781e-01
-1.41439056e+00 -1.15819383e+00 8.56388867e-01 4.69330221e-01
-8.18236709e-01 9.09748197e-01 8.14876199e-01 -4.78631109e-01
-3.04042846e-01 -3.77251476e-01 -5.69006920e-01 -4.64754969e-01
8.09692033e-03 4.35230345e-01 3.90221268e-01 -1.48779646e-01
3.79001558e-01 2.72105157e-01 -1.19573045e+00 5.28336883e-01
5.76788545e-01 8.36071193e-01 3.32765996e-01 4.44304556e-01
-4.66736704e-01 5.87862015e-01 -4.83954906e-01 -3.23004782e-01
-9.75368693e-02 -6.57573119e-02 -1.00811887e+00 -8.44671354e-02
2.23384380e-01 3.81563365e-01 -1.59187078e-01 9.41699803e-01
4.64697748e-01 -1.68131828e-01 6.84691787e-01 -8.74824047e-01
-3.29740196e-01 1.36327231e+00 -5.20524144e-01 5.65761328e-01
4.34742540e-01 -7.20425844e-01 1.28503823e+00 -8.62803459e-01
9.76806760e-01 9.60326672e-01 6.68871582e-01 -9.69543681e-02
-6.77528441e-01 -7.26750016e-01 -4.99723945e-03 5.41464686e-02
-1.72967565e+00 -4.73163992e-01 6.49244130e-01 -4.10915792e-01
1.19277036e+00 2.92644590e-01 3.56596172e-01 9.42677796e-01
-8.21348876e-02 7.36340523e-01 1.40631509e+00 -8.50298882e-01
-1.47002246e-02 7.03686118e-01 2.97326595e-01 2.32675597e-01
2.36291796e-01 -6.73261940e-01 -6.82481289e-01 -4.66342151e-01
1.07276661e-03 -6.18899524e-01 -3.86348605e-01 2.22088903e-01
-1.26684022e+00 8.27191353e-01 -5.45762420e-01 6.62751198e-01
-4.20398623e-01 -1.03280708e-01 6.80931866e-01 4.31728065e-01
8.21885943e-01 6.30671680e-01 -1.24314082e+00 -3.93756658e-01
-1.24337125e+00 5.92444122e-01 1.14763963e+00 1.28883135e+00
4.90909576e-01 -2.81895459e-01 1.98433578e-01 1.00848746e+00
3.51681620e-01 5.28861523e-01 6.33999288e-01 -5.28952122e-01
7.08101451e-01 5.10970473e-01 -1.19162593e-02 -1.10334361e+00
-5.63464046e-01 -3.94897461e-01 -1.37030751e-01 -9.33362007e-01
1.68509230e-01 -3.34833711e-01 -8.21376801e-01 1.39715195e+00
3.62977922e-01 -3.73096317e-01 2.28744939e-01 4.55587536e-01
1.34338510e+00 9.11285579e-01 -1.81282610e-01 -7.05537379e-01
1.95247865e+00 -5.71835935e-01 -9.02830780e-01 2.00850606e-01
7.15912998e-01 -1.36999452e+00 7.41821110e-01 5.69393575e-01
-7.26597488e-01 1.03176072e-01 -5.21948218e-01 -1.02928139e-01
-1.03218579e+00 1.62251666e-01 7.87165999e-01 7.09043384e-01
-7.40369916e-01 -3.50195169e-01 -3.45066607e-01 -7.64518857e-01
-3.03094592e-02 -1.32334530e-02 -1.93547890e-01 -4.95682210e-02
-1.50634313e+00 6.25408947e-01 7.32992828e-01 -5.83606362e-01
-6.23904228e-01 -7.12453008e-01 -5.92573047e-01 -3.70582670e-01
7.34607220e-01 1.35767728e-01 1.16276574e+00 -3.30671966e-01
-8.86240184e-01 9.28384006e-01 -1.22426331e-01 -5.38076043e-01
-2.44693998e-02 -1.39227360e-02 -9.74403203e-01 5.01649007e-02
3.54106367e-01 1.77022845e-01 3.14234793e-01 -9.04835522e-01
-1.24690557e+00 -1.63541645e-01 -2.43732661e-01 2.34705776e-01
-5.13479233e-01 9.88598168e-01 -1.27598023e+00 -1.20075119e+00
2.85384450e-02 -8.05518448e-01 3.39297280e-02 -1.00533521e+00
-4.37517494e-01 -5.43927550e-01 7.80169904e-01 -9.69312310e-01
1.88127172e+00 -2.02473879e+00 -1.29452020e-01 3.20672452e-01
-1.25065476e-01 -3.96150678e-01 3.72930586e-01 8.40249062e-01
1.50788382e-01 4.19529349e-01 4.47275698e-01 2.14125682e-02
5.37258293e-03 1.83329374e-01 -5.27901709e-01 6.44737661e-01
-6.24950111e-01 7.98344612e-01 -6.77066624e-01 -9.25111890e-01
-4.52615350e-01 1.44196391e-01 -2.34413773e-01 -4.80365366e-01
-2.94099629e-01 -2.36639380e-01 -7.18912899e-01 1.14740384e+00
2.02292055e-01 1.42270997e-01 1.14336528e-01 -2.99763322e-01
-6.43656969e-01 7.86183119e-01 -1.17903519e+00 1.63875079e+00
-3.10722351e-01 7.59922683e-01 1.39470980e-01 -7.50936985e-01
3.75699371e-01 7.38465905e-01 6.64358258e-01 -5.55257797e-01
8.74170065e-02 6.99311852e-01 -1.24917194e-01 -3.37379038e-01
1.03403354e+00 2.71115720e-01 -8.33116829e-01 4.07834917e-01
1.29433185e-01 -1.87214136e-01 6.14821970e-01 4.24338967e-01
7.32374370e-01 -2.31738776e-01 5.93539596e-01 -8.44571590e-01
3.68168145e-01 4.71404254e-01 5.23212135e-01 6.52492642e-01
1.15338705e-01 2.09472388e-01 3.86770070e-01 -1.20357119e-01
-1.06772292e+00 -1.70769423e-01 -6.36639237e-01 1.13338602e+00
-4.49084610e-01 -8.40918541e-01 -5.53451121e-01 -5.59879184e-01
1.56958491e-01 5.88886857e-01 -3.98357213e-01 4.40356255e-01
-5.50955594e-01 -1.05124855e+00 9.23838973e-01 -1.85372025e-01
6.19688556e-02 -8.59753609e-01 1.00709246e-02 2.93470681e-01
-3.84397686e-01 -1.62521970e+00 -6.00850463e-01 3.85445118e-01
-3.36631984e-01 -7.83158481e-01 -7.90997744e-01 -9.69321609e-01
2.24462479e-01 2.01114506e-01 1.23519158e+00 -5.59009731e-01
1.01157047e-01 4.45926487e-01 -8.66472244e-01 -7.90323436e-01
-3.44674975e-01 6.36077225e-01 2.47247815e-01 -3.65521103e-01
8.34700763e-01 -1.61867082e-01 -4.18917611e-02 1.46595612e-01
-6.56174183e-01 -2.53674895e-01 3.35461795e-01 2.58875996e-01
7.54619002e-01 7.47683704e-01 6.28690839e-01 -9.85551834e-01
9.86596107e-01 -7.34742999e-01 -7.36661315e-01 3.17641497e-01
-7.61509299e-01 -2.40006015e-01 2.59014189e-01 -4.99173731e-01
-8.59309793e-01 -3.52875412e-01 2.49687463e-01 3.11807096e-01
-5.08657321e-02 1.30961108e+00 -2.02872232e-03 1.07115805e-02
3.86959165e-01 3.98452580e-01 -8.14661980e-01 -1.08392143e+00
3.75307262e-01 1.20695353e+00 4.40298885e-01 -6.56169355e-01
4.66167748e-01 1.69018894e-01 -7.55658507e-01 -1.13148785e+00
-8.90934825e-01 -1.14045405e+00 -3.25496495e-01 -7.01521188e-02
6.73108935e-01 -1.25851417e+00 -5.21228425e-02 4.40216839e-01
-8.24463487e-01 2.60008395e-01 8.13414976e-02 8.05360198e-01
1.11505605e-01 2.63613939e-01 -5.82906902e-01 -8.24135423e-01
-3.50332826e-01 -8.36151600e-01 9.06921744e-01 -1.03953853e-01
-3.42284799e-01 -9.90942955e-01 1.38584182e-01 3.63306940e-01
1.69639990e-01 5.61466590e-02 8.10460269e-01 -1.00970495e+00
-8.42664987e-02 -3.21798235e-01 1.74895853e-01 -5.87883964e-02
3.47187817e-01 3.82499536e-03 -4.39925045e-01 -2.55740225e-01
-2.39241689e-01 -4.21918392e-01 8.80200028e-01 5.44872940e-01
9.90555167e-01 -5.65177619e-01 -5.82493305e-01 2.18219563e-01
1.39271176e+00 2.94627249e-01 1.36193439e-01 7.75169730e-01
5.45570493e-01 5.17849505e-01 6.15298092e-01 4.49267775e-01
7.67835915e-01 7.97083616e-01 -3.91205311e-01 1.56552330e-01
3.86217274e-02 -9.08973534e-03 4.34967130e-01 1.54766750e+00
1.58933997e-01 -4.74755377e-01 -1.29944444e+00 1.11870253e+00
-1.29131758e+00 -7.46954322e-01 -1.00152269e-01 1.85305154e+00
1.51083374e+00 3.56808364e-01 3.40952694e-01 1.74014017e-01
2.78763473e-01 1.91486630e-04 1.08523175e-01 -3.62541050e-01
-4.79550898e-01 9.32111870e-03 1.03048670e+00 3.21068674e-01
-1.31277680e+00 1.33643842e+00 6.60111332e+00 1.43842626e+00
-8.25943351e-01 1.19386576e-01 2.77493000e-01 -7.42885983e-03
-4.96424377e-01 8.93900990e-02 -1.60495603e+00 6.29313648e-01
1.21381927e+00 -5.05042672e-01 1.98693499e-01 6.86747551e-01
2.89264768e-01 -2.52002209e-01 -4.77551937e-01 6.86391234e-01
3.48309338e-01 -1.47036839e+00 -1.61777571e-01 3.26161355e-01
8.59019876e-01 5.83122849e-01 -3.17362785e-01 4.32285875e-01
2.64856339e-01 -6.77797198e-01 9.47633147e-01 3.21434103e-02
7.43015349e-01 -1.13295186e+00 7.89103627e-01 5.36823452e-01
-1.13236666e+00 3.11460763e-01 -2.02094279e-02 3.95242959e-01
2.17285588e-01 8.22809100e-01 -7.91975498e-01 6.71848953e-01
1.14196694e+00 5.35635293e-01 -5.86094737e-01 8.15222621e-01
2.96196759e-01 1.22244096e+00 -5.95659435e-01 -4.36422646e-01
6.49708271e-01 -9.26841199e-02 5.39874196e-01 1.87181449e+00
2.39399701e-01 5.46293184e-02 5.07757902e-01 -7.08928257e-02
-2.43367389e-01 8.79243851e-01 -4.32227522e-01 -5.51118255e-01
7.60660529e-01 1.22558773e+00 -1.02667916e+00 -3.98520947e-01
-7.02718258e-01 2.71897435e-01 -1.23493239e-01 4.24704611e-01
-6.42548680e-01 -6.78028286e-01 3.44124854e-01 -2.82437988e-02
4.35417801e-01 -4.49267387e-01 -1.65297389e-02 -1.08838010e+00
-1.10492513e-01 -1.44751072e+00 6.77814305e-01 -2.67216504e-01
-1.05308151e+00 6.66106045e-01 3.98623228e-01 -9.75901961e-01
-8.32561776e-02 -3.80648702e-01 3.71837199e-01 8.14391613e-01
-1.59850061e+00 -1.20354831e+00 4.78773654e-01 4.55830276e-01
7.66361237e-01 -7.60238826e-01 6.80216432e-01 7.61518359e-01
-3.93223017e-01 6.92141712e-01 4.79853481e-01 2.12444633e-01
7.69051254e-01 -1.17568231e+00 2.24917099e-01 6.08888447e-01
5.16898990e-01 7.11912692e-01 8.41070533e-01 -8.24763000e-01
-1.35195851e+00 -8.47395301e-01 1.44839275e+00 -6.16888881e-01
1.21918809e+00 -5.91124833e-01 -6.04815304e-01 4.83673751e-01
6.94221199e-01 -4.68065262e-01 9.63531852e-01 3.05647880e-01
-2.04291105e-01 -6.20587077e-03 -8.03624511e-01 4.27197754e-01
3.71839523e-01 -7.23573387e-01 -6.40541971e-01 7.17238665e-01
9.56731915e-01 -3.34044367e-01 -1.20437741e+00 1.93241566e-01
4.56598282e-01 2.82652024e-02 7.12517738e-01 -5.57546616e-01
-7.78127909e-02 -2.19721481e-01 -5.25398374e-01 -1.00850463e+00
-2.79023852e-02 -6.34391606e-01 -1.62963659e-01 1.88766098e+00
1.07338309e+00 -3.17005098e-01 5.13015807e-01 1.30545199e-02
-4.69439924e-02 -7.65028417e-01 -9.69502449e-01 -6.75946593e-01
1.69740066e-01 -1.06687295e+00 3.99623096e-01 1.30541170e+00
8.39430466e-02 3.70706230e-01 -5.10081470e-01 3.32146287e-02
2.80696601e-01 2.24277109e-01 4.82077986e-01 -9.37850833e-01
1.59886591e-02 -2.64826387e-01 -8.07470456e-02 -8.11956286e-01
1.17832921e-01 -8.16158116e-01 -1.97150379e-01 -1.33113825e+00
2.83782423e-01 -5.95834553e-01 -6.42153397e-02 4.86806780e-01
2.75891751e-01 8.75021070e-02 -5.65063477e-01 3.47700685e-01
-6.65046692e-01 7.31256325e-03 7.17753768e-01 2.06479244e-02
-1.59121156e-01 -1.33596927e-01 -7.79500842e-01 7.68102050e-01
8.07913423e-01 -8.95165265e-01 -2.60988355e-01 -1.61207706e-01
6.67827725e-01 -5.31847596e-01 -5.43335557e-01 -3.64191711e-01
3.38350147e-01 -4.11521137e-01 -6.40244931e-02 -1.09374237e+00
-1.43684059e-01 -6.98898017e-01 5.81283569e-02 -7.95836188e-03
-2.39633605e-01 4.24325645e-01 2.65932411e-01 3.60533416e-01
-5.93495965e-01 -2.03703403e-01 2.77331501e-01 -4.51773256e-01
-1.02006054e+00 1.44643039e-01 -8.10420632e-01 6.04478836e-01
8.15686703e-01 1.01876341e-01 7.98639059e-02 -1.69701338e-01
-1.51837587e-01 3.55901957e-01 2.84587055e-01 5.85667253e-01
1.60541639e-01 -1.22195303e+00 -9.53335285e-01 -4.29534525e-01
3.96616489e-01 -4.35328096e-01 -4.28492874e-01 8.52408409e-01
-6.15916848e-01 9.70729470e-01 6.24484241e-01 -9.00197998e-02
-1.38624644e+00 5.02886772e-01 -4.03478779e-02 -4.03914362e-01
-5.40568471e-01 8.69564712e-01 -4.27030653e-01 -4.84074116e-01
6.11680865e-01 -4.67496097e-01 -7.85934567e-01 6.50645852e-01
3.84781271e-01 5.20724058e-02 2.06094652e-01 -8.95892799e-01
-6.44993603e-01 4.66432005e-01 -4.03283983e-01 -5.09178460e-01
1.33718204e+00 -4.42852408e-01 -4.86191273e-01 6.56190276e-01
1.21286082e+00 6.16965413e-01 7.73578435e-02 -5.58967888e-01
8.57765853e-01 -2.67903749e-02 5.53705394e-01 -7.76469350e-01
-7.26814747e-01 -1.17146395e-01 9.62466672e-02 4.20302808e-01
1.06902790e+00 3.43985021e-01 7.61017919e-01 4.45036739e-01
4.37733024e-01 -1.76312602e+00 -5.39420009e-01 6.46917939e-01
7.99281716e-01 -1.19507957e+00 5.40137887e-01 -2.90819138e-01
-4.97045100e-01 7.42286325e-01 2.36351881e-02 5.67778945e-01
1.14925694e+00 4.96384770e-01 3.06938171e-01 -6.44917488e-01
-8.39629769e-01 -2.08805993e-01 7.94270813e-01 -4.16797996e-02
7.67751992e-01 3.26512195e-02 -1.03876829e+00 7.31900692e-01
-5.39871395e-01 -3.37190717e-01 4.96196091e-01 1.00421929e+00
-3.99776191e-01 -1.13591707e+00 -2.72162527e-01 5.09757400e-01
-1.44419158e+00 -5.15513480e-01 -5.46401739e-01 9.41052079e-01
1.41129538e-01 8.76574159e-01 -4.52204883e-01 -3.65044475e-02
1.15979955e-01 -9.56856161e-02 1.73756573e-02 -6.62700057e-01
-5.63463271e-01 7.20693231e-01 7.51228571e-01 2.23034844e-02
-8.22090924e-01 -9.73681450e-01 -8.46196830e-01 -1.72375736e-03
-6.32314861e-01 9.39901829e-01 9.26915467e-01 8.98620725e-01
2.19953820e-01 1.79461360e-01 5.81464648e-01 -1.12552486e-01
-7.64802694e-02 -1.15066385e+00 -7.71936059e-01 -1.62384555e-01
1.27541751e-01 -4.25314039e-01 -3.82204354e-01 1.81599945e-01] | [10.19558048248291, 9.622690200805664] |
c6b18734-29b3-4380-953f-35688183a35d | a-general-method-for-amortizing-variational | 1811.05090 | null | http://arxiv.org/abs/1811.05090v1 | http://arxiv.org/pdf/1811.05090v1.pdf | A General Method for Amortizing Variational Filtering | We introduce the variational filtering EM algorithm, a simple,
general-purpose method for performing variational inference in dynamical latent
variable models using information from only past and present variables, i.e.
filtering. The algorithm is derived from the variational objective in the
filtering setting and consists of an optimization procedure at each time step.
By performing each inference optimization procedure with an iterative amortized
inference model, we obtain a computationally efficient implementation of the
algorithm, which we call amortized variational filtering. We present
experiments demonstrating that this general-purpose method improves performance
across several deep dynamical latent variable models. | ['Yisong Yue', 'Joseph Marino', 'Milan Cvitkovic'] | 2018-11-13 | a-general-method-for-amortizing-variational-1 | http://papers.nips.cc/paper/8011-a-general-method-for-amortizing-variational-filtering | http://papers.nips.cc/paper/8011-a-general-method-for-amortizing-variational-filtering.pdf | neurips-2018-12 | ['inference-optimization'] | ['audio'] | [-1.16842195e-01 -8.96443501e-02 -7.33900070e-02 -2.04232246e-01
-8.41434956e-01 -5.44003367e-01 1.15206885e+00 -4.80873734e-01
-4.06634629e-01 8.04034472e-01 1.63825721e-01 -2.85293549e-01
1.09749973e-01 -8.25047374e-01 -7.73944676e-01 -8.84399712e-01
5.92528060e-02 8.73039365e-01 4.25956585e-02 2.32264712e-01
5.71570247e-02 2.40241781e-01 -1.25596130e+00 -1.74961418e-01
5.05360842e-01 5.27673066e-01 1.03854388e-01 9.85511482e-01
-1.58508949e-04 5.20442247e-01 -3.86314452e-01 -4.50365841e-01
-2.32631397e-02 -7.03332007e-01 -7.51503229e-01 3.58002000e-02
1.36498287e-01 -6.03659213e-01 -3.89175147e-01 8.92152309e-01
2.38752306e-01 8.18736851e-01 9.05820370e-01 -1.03134716e+00
-5.02153277e-01 6.11283898e-01 -3.26284111e-01 4.91339087e-01
8.48742500e-02 1.99336410e-01 1.21622264e+00 -1.01059449e+00
1.00714433e+00 1.74411213e+00 5.89116633e-01 6.16148472e-01
-1.94837201e+00 -1.99920848e-01 3.13506722e-01 -2.93017346e-02
-9.90395606e-01 -4.98142302e-01 6.02870464e-01 -7.52551854e-01
1.19860423e+00 7.35783651e-02 8.97664905e-01 1.25343275e+00
4.30614769e-01 9.95052159e-01 8.48293006e-01 -1.56366050e-01
5.00427604e-01 -2.71627426e-01 4.45812643e-01 6.05594099e-01
-2.02666044e-01 2.93311357e-01 -4.92648572e-01 -7.53695190e-01
8.83272231e-01 1.56350464e-01 9.82413664e-02 -2.22900003e-01
-8.75819623e-01 1.13258350e+00 -2.92342365e-01 -2.49034733e-01
-5.28105736e-01 8.04965079e-01 3.00600588e-01 4.49593872e-01
9.05991077e-01 -1.62756279e-01 -5.10411501e-01 -4.97859120e-01
-1.30677068e+00 8.27889860e-01 7.53132582e-01 5.34760714e-01
8.10040772e-01 3.60838681e-01 -3.04005504e-01 5.44265211e-01
8.13458264e-01 8.70731115e-01 -1.39326617e-01 -1.49536991e+00
-2.60423064e-01 -2.17531666e-01 4.27083522e-01 -4.53469753e-01
-2.15807855e-01 -2.93615818e-01 -6.32540166e-01 7.80396760e-02
1.81389377e-01 -3.92637253e-01 -1.18367755e+00 2.06698704e+00
4.88821000e-01 5.94422698e-01 -2.25251123e-01 4.45450366e-01
4.24941927e-01 9.54037786e-01 3.61522101e-02 -7.76477754e-01
9.07699704e-01 -8.29822481e-01 -1.08582592e+00 1.22230157e-01
-7.04056174e-02 -5.76629460e-01 6.08212054e-01 2.88113922e-01
-1.53935695e+00 -3.47477853e-01 -6.95600927e-01 -1.54280916e-01
-1.13309689e-01 -3.14971864e-01 9.29503918e-01 5.51629603e-01
-1.26143992e+00 9.18072164e-01 -1.94173265e+00 -6.33626282e-02
-2.79335864e-03 1.88394517e-01 2.29459628e-01 5.79767466e-01
-1.00424242e+00 6.09414041e-01 -1.07686520e-01 8.61931816e-02
-1.66929424e+00 -9.57315803e-01 -7.66295373e-01 4.35613953e-02
1.47582099e-01 -1.02475369e+00 1.70471549e+00 -5.37980139e-01
-1.80642903e+00 5.80361128e-01 -1.21835685e+00 -7.70599961e-01
6.75389886e-01 -3.86538804e-01 -1.27347514e-01 7.03582261e-03
-3.25820446e-02 2.69563079e-01 1.29176414e+00 -1.06358969e+00
-2.35834122e-01 -1.89377084e-01 -2.13844627e-01 -3.43934298e-01
2.77199715e-01 -2.69280393e-02 -6.32763505e-01 -5.58814764e-01
2.09733039e-01 -1.02996099e+00 -3.25388193e-01 -2.10997760e-01
-2.59173125e-01 -4.24575537e-01 7.69727468e-01 -7.31607378e-01
1.42484677e+00 -1.72515225e+00 7.04495668e-01 1.49819031e-01
6.32229567e-01 -2.68559515e-01 3.28678936e-01 4.34354395e-01
2.23536298e-01 1.49741217e-01 -5.76135576e-01 -1.10707986e+00
1.32873222e-01 5.26435018e-01 -6.82659090e-01 6.13079250e-01
6.82122037e-02 1.14283419e+00 -1.02681673e+00 -4.06002283e-01
4.15921688e-01 5.03558874e-01 -1.00536895e+00 2.86569685e-01
-8.55672240e-01 6.87167406e-01 -4.20444548e-01 3.44624341e-01
7.62338519e-01 -4.42005813e-01 3.36412340e-01 2.22971722e-01
-2.90302157e-01 2.74594635e-01 -1.18582463e+00 1.66182888e+00
-3.14725608e-01 9.73304510e-01 2.42449775e-01 -8.45229208e-01
2.03305975e-01 4.83850449e-01 6.94833636e-01 -7.05397278e-02
-9.40131489e-03 -1.61901504e-01 -5.98311305e-01 -3.26912314e-01
4.34916198e-01 -3.95252615e-01 -3.47415134e-02 6.77736700e-01
3.94765943e-01 1.64105482e-02 2.77106166e-01 6.30969405e-01
9.29106772e-01 4.59229678e-01 -3.83997709e-02 -3.30565959e-01
9.44616646e-02 -1.29788771e-01 7.50573397e-01 1.38668823e+00
-1.64869532e-01 2.24029377e-01 3.57372582e-01 -4.28698838e-01
-1.16059053e+00 -1.93461442e+00 -3.75411928e-01 1.07887685e+00
-3.29669088e-01 -9.05038476e-01 -5.45897007e-01 -9.66498256e-02
2.66171992e-02 6.25944972e-01 -8.22069526e-01 1.43394724e-01
-6.53941154e-01 -9.94989038e-01 2.07730487e-01 3.79313916e-01
2.17514455e-01 -7.99001157e-01 -4.74990457e-01 5.17296791e-01
-2.34134927e-01 -4.78382319e-01 -1.97971791e-01 2.47096773e-02
-1.06990254e+00 -4.85772610e-01 -5.76651990e-01 -8.85885805e-02
2.08628476e-01 -1.21031471e-01 1.50078642e+00 -3.40528786e-02
-2.30446160e-02 4.77231920e-01 2.37384439e-01 -1.10414162e-01
-4.86244321e-01 -1.87075540e-01 2.75630087e-01 -1.35554085e-02
3.31437975e-01 -7.37835944e-01 -4.76788670e-01 -2.19675198e-01
-6.63278818e-01 -1.07285403e-01 -4.47059348e-02 1.00264132e+00
9.82969642e-01 -2.47900024e-01 -5.03279567e-02 -9.38259602e-01
5.84118843e-01 -6.84520245e-01 -1.03025794e+00 6.27434775e-02
-4.02188361e-01 4.01848823e-01 1.99170604e-01 -4.51599389e-01
-1.43545592e+00 -3.93697470e-02 -3.79564404e-01 -7.42396355e-01
2.68048853e-01 7.04933763e-01 2.51506865e-01 4.20690805e-01
2.48080492e-01 3.75508457e-01 7.89741427e-02 -8.54494333e-01
7.13576019e-01 1.15215927e-01 5.61481774e-01 -6.93054855e-01
8.60610425e-01 9.47444439e-01 1.03105657e-01 -9.73814547e-01
-7.15523481e-01 -2.20364362e-01 -6.17273688e-01 -2.70750791e-01
1.18349087e+00 -9.97719109e-01 -9.75891948e-01 6.38379931e-01
-1.12630093e+00 -6.27134204e-01 -1.64934084e-01 4.62149382e-01
-7.58440197e-01 2.25475833e-01 -8.21864486e-01 -1.32195067e+00
-1.38868108e-01 -9.78996515e-01 1.08074319e+00 -4.07051258e-02
-2.47097015e-01 -1.58323467e+00 7.43499637e-01 -2.80750543e-01
2.43029296e-01 2.19902784e-01 6.30678356e-01 2.27877703e-02
-8.43595743e-01 1.39197022e-01 3.54577124e-01 5.08259907e-02
-2.75062472e-01 5.95191836e-01 -8.19432497e-01 -3.63758087e-01
2.35664308e-01 -1.07911684e-01 1.46305120e+00 1.03758872e+00
8.84290278e-01 -2.19327152e-01 -5.43922365e-01 7.96881735e-01
1.44247377e+00 -1.74267203e-01 4.02803868e-01 -3.82818460e-01
5.56801260e-01 -1.79176554e-01 1.67256713e-01 7.40639329e-01
2.88518429e-01 6.28164887e-01 -5.13035022e-02 1.66937515e-01
2.19742373e-01 -1.67340517e-01 4.95526850e-01 9.30471480e-01
-2.85365015e-01 -1.61585376e-01 -7.15619385e-01 7.37317204e-01
-2.20486188e+00 -1.58890319e+00 -1.30197018e-01 1.82938445e+00
7.95771480e-01 4.84775268e-02 2.60647893e-01 -5.09807527e-01
4.45763618e-01 3.52810979e-01 -8.74739528e-01 -4.34138060e-01
4.78870906e-02 5.04876852e-01 2.13196814e-01 1.10406446e+00
-1.09508264e+00 1.18471622e+00 8.76522160e+00 7.26325393e-01
-5.80051124e-01 6.77798986e-01 7.27449805e-02 -6.24145269e-01
-6.59370661e-01 1.47682026e-01 -1.00051665e+00 4.55968142e-01
1.31471384e+00 -2.81800330e-01 8.08561385e-01 7.05425262e-01
6.32805288e-01 7.80578256e-02 -9.53932345e-01 6.31162286e-01
-4.94485945e-01 -1.66353881e+00 5.81406318e-02 2.42324188e-01
1.01263058e+00 4.08212990e-01 1.82051897e-01 3.63795072e-01
1.01022732e+00 -5.67320347e-01 7.48610556e-01 7.88995266e-01
2.42303312e-01 -6.41878664e-01 4.55962084e-02 1.53313443e-01
-1.22078598e+00 1.23183213e-01 -3.86808246e-01 -2.87394702e-01
8.54056776e-01 7.43152678e-01 2.63199937e-02 -8.15776587e-02
6.00890219e-01 9.29970562e-01 1.83185637e-01 6.48190141e-01
-3.42838824e-01 1.05929387e+00 -6.00054443e-01 2.12662905e-01
2.29816601e-01 -6.53599381e-01 1.06756639e+00 1.15995145e+00
-4.53795977e-02 7.39447549e-02 1.47077978e-01 1.58486080e+00
1.47829473e-01 -7.48387754e-01 -2.75689393e-01 -1.90241292e-01
5.90685725e-01 9.34068859e-01 -5.14181376e-01 -6.95596576e-01
-3.82743597e-01 8.98586333e-01 3.54435176e-01 6.99686110e-01
-1.07043374e+00 2.48693064e-01 1.30111694e+00 -3.53564829e-01
7.21791565e-01 -6.46856248e-01 -2.17389867e-01 -1.58417511e+00
-4.60792035e-01 -4.21966434e-01 5.48974574e-01 -6.00626528e-01
-1.15740407e+00 1.10990182e-01 4.04220432e-01 -4.81441885e-01
-7.57305086e-01 -2.65242040e-01 -8.02656889e-01 1.16126108e+00
-9.54768062e-01 -8.14802289e-01 1.79315537e-01 7.70222127e-01
5.97457230e-01 2.24459007e-01 6.42227769e-01 -3.11152544e-02
-7.47209430e-01 3.56821030e-01 7.09624767e-01 -1.86497182e-01
4.60451767e-02 -1.51387441e+00 9.01729822e-01 1.20486367e+00
1.27957284e-01 1.20185590e+00 1.33379734e+00 -8.83014858e-01
-1.59361327e+00 -7.44994462e-01 7.66486585e-01 -6.62720263e-01
8.27414036e-01 -7.14438498e-01 -8.22399974e-01 1.41310287e+00
2.35069886e-01 -6.16509497e-01 5.30592620e-01 6.14362299e-01
-9.04140819e-04 6.37247264e-01 -9.52465117e-01 7.58050799e-01
1.10010254e+00 -8.27087164e-01 -7.67750561e-01 4.44908738e-01
9.62430537e-01 -4.90691781e-01 -1.04807079e+00 5.73040023e-02
6.20700777e-01 -3.80524129e-01 1.23536158e+00 -7.92475462e-01
3.64134222e-01 -5.13911769e-02 -1.61632121e-01 -1.01963115e+00
-6.05820358e-01 -1.15780187e+00 -1.15688217e+00 8.71442378e-01
4.31272298e-01 -6.35728717e-01 6.08056784e-01 7.77597606e-01
1.66710839e-01 -4.09481257e-01 -1.04010463e+00 -6.29221797e-01
3.98594677e-01 -6.65669799e-01 3.41205716e-01 6.81219041e-01
-3.30801100e-01 2.71588236e-01 -6.69171929e-01 1.78478777e-01
1.22864985e+00 3.37591499e-01 7.98984826e-01 -1.28002512e+00
-8.24116766e-01 -2.54296899e-01 6.89320341e-02 -1.51484287e+00
4.83516037e-01 -6.95750892e-01 -7.58673549e-02 -1.32000721e+00
4.47012424e-01 1.90890849e-01 -2.04898536e-01 1.87570974e-02
-3.64988416e-01 4.15441431e-02 -5.28574847e-02 4.14729655e-01
-4.42633599e-01 7.92493284e-01 9.59298611e-01 -6.65826872e-02
-4.26060408e-01 1.23291694e-01 4.73300889e-02 6.07095003e-01
3.63487244e-01 -5.72601378e-01 -2.72640854e-01 -4.22314852e-01
2.72275925e-01 4.03150678e-01 6.00977123e-01 -4.70027089e-01
1.11007698e-01 -3.38026881e-01 3.18173110e-01 -1.09351897e+00
7.54825413e-01 -9.07245874e-02 6.45589650e-01 5.04405200e-01
-2.05689415e-01 1.30583704e-01 9.23578590e-02 7.92662919e-01
1.73779905e-01 3.47016342e-02 9.46347892e-01 -2.60349125e-01
-6.34513617e-01 4.41959620e-01 -9.54865813e-01 7.28430226e-02
4.56747711e-01 2.88436592e-01 3.21734957e-02 -5.25095940e-01
-1.62644446e+00 2.18857959e-01 3.99251312e-01 -2.84520108e-02
3.51971060e-01 -1.21905851e+00 -7.00340092e-01 8.85668918e-02
-5.60974896e-01 -4.39069390e-01 2.98756272e-01 8.38361979e-01
-2.14917004e-01 3.25486422e-01 2.37771392e-01 -8.86398554e-01
-1.07568204e+00 7.02295721e-01 5.89482963e-01 -3.88794154e-01
-8.66688609e-01 1.07278311e+00 9.65866521e-02 -3.08156550e-01
-1.16090000e-01 -8.08551386e-02 3.50029200e-01 7.94038102e-02
6.73113167e-01 6.63361013e-01 -4.63533938e-01 -4.36710030e-01
-2.30885312e-01 2.71684676e-01 2.36966070e-02 -9.53865588e-01
1.21139038e+00 -5.33331811e-01 -2.63365626e-01 1.04058874e+00
1.18239748e+00 -1.64810553e-01 -1.62280846e+00 -3.20235372e-01
-4.28761899e-01 -3.28815788e-01 5.05408168e-01 -3.72657508e-01
-9.35403585e-01 8.29924107e-01 3.77841949e-01 3.47391367e-01
6.96215987e-01 1.17369376e-01 7.08881199e-01 3.63107830e-01
2.04069391e-01 -1.11006212e+00 -3.54310066e-01 7.95003414e-01
4.14275914e-01 -9.03185368e-01 -6.64595934e-03 1.75909661e-02
2.15342008e-02 6.55829787e-01 -3.99872027e-02 -3.91238749e-01
1.28606164e+00 1.76903605e-01 -2.31174812e-01 -4.45363492e-01
-1.28138077e+00 -7.74335340e-02 6.71938583e-02 2.23527029e-01
2.29183555e-01 2.32633755e-01 -3.35336924e-01 2.99498230e-01
-1.51737660e-01 1.78974718e-02 1.94126323e-01 8.71981978e-01
-4.42304403e-01 -1.07642472e+00 -7.35593066e-02 4.66998726e-01
-5.46872139e-01 -1.69352800e-01 -2.00145721e-01 4.72051769e-01
-4.74330783e-01 1.12457907e+00 4.92244333e-01 -4.24161740e-02
-3.55690986e-01 3.01765770e-01 6.56971157e-01 -2.72174358e-01
-3.65861118e-01 3.43941897e-01 -6.61610290e-02 -7.71225989e-01
-4.65684295e-01 -1.16319561e+00 -1.07958305e+00 -9.19068992e-01
-2.96904087e-01 1.42400742e-01 4.90924567e-01 1.30114925e+00
2.51021534e-01 4.99502510e-01 3.70331913e-01 -1.05750680e+00
-5.64770699e-01 -9.33666468e-01 -4.42372590e-01 3.82619470e-01
6.18856132e-01 -1.01520932e+00 -5.94504535e-01 1.98323637e-01] | [6.906907081604004, 3.8281140327453613] |
b5266bac-6756-42a3-9cd6-fa6cb42b36fd | fs6d-few-shot-6d-pose-estimation-of-novel | 2203.14628 | null | https://arxiv.org/abs/2203.14628v1 | https://arxiv.org/pdf/2203.14628v1.pdf | FS6D: Few-Shot 6D Pose Estimation of Novel Objects | 6D object pose estimation networks are limited in their capability to scale to large numbers of object instances due to the close-set assumption and their reliance on high-fidelity object CAD models. In this work, we study a new open set problem; the few-shot 6D object poses estimation: estimating the 6D pose of an unknown object by a few support views without extra training. To tackle the problem, we point out the importance of fully exploring the appearance and geometric relationship between the given support views and query scene patches and propose a dense prototypes matching framework by extracting and matching dense RGBD prototypes with transformers. Moreover, we show that the priors from diverse appearances and shapes are crucial to the generalization capability under the problem setting and thus propose a large-scale RGBD photorealistic dataset (ShapeNet6D) for network pre-training. A simple and effective online texture blending approach is also introduced to eliminate the domain gap from the synthesis dataset, which enriches appearance diversity at a low cost. Finally, we discuss possible solutions to this problem and establish benchmarks on popular datasets to facilitate future research. The project page is at \url{https://fs6d.github.io/}. | ['Qifeng Chen', 'Jian Sun', 'Haoqiang Fan', 'Yao Wang', 'Yisheng He'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/He_FS6D_Few-Shot_6D_Pose_Estimation_of_Novel_Objects_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/He_FS6D_Few-Shot_6D_Pose_Estimation_of_Novel_Objects_CVPR_2022_paper.pdf | cvpr-2022-1 | ['6d-pose-estimation-1', '6d-pose-estimation'] | ['computer-vision', 'computer-vision'] | [ 2.80498117e-02 1.68120250e-01 -6.55674860e-02 -5.55288494e-01
-5.80387115e-01 -5.02755702e-01 3.34346741e-01 -4.39229161e-01
-5.89134283e-02 2.76327759e-01 -1.90016136e-01 1.89430356e-01
-2.34110251e-01 -6.06173217e-01 -9.88358676e-01 -5.11764109e-01
2.56509155e-01 8.08424771e-01 3.31202835e-01 -9.50350910e-02
2.06578411e-02 9.59422946e-01 -1.78231764e+00 1.64364800e-02
4.93428051e-01 1.50707841e+00 2.73449451e-01 1.42711952e-01
1.39244288e-01 3.92675579e-01 -3.30545664e-01 -5.26160717e-01
9.65931535e-01 1.79369971e-02 -4.50298846e-01 4.94919568e-01
1.00526226e+00 -6.20440423e-01 -4.33265865e-01 9.55319643e-01
6.05596125e-01 2.54302800e-01 5.46225071e-01 -1.57757127e+00
-5.16212106e-01 -7.46461973e-02 -6.64276719e-01 -2.40571886e-01
9.32763889e-02 1.83505088e-01 8.87185454e-01 -1.20247495e+00
8.64665985e-01 1.06280053e+00 8.09511125e-01 7.16583490e-01
-1.10537577e+00 -6.11135244e-01 2.58197367e-01 1.98685125e-01
-1.49510550e+00 -6.43958330e-01 1.21880710e+00 -2.54554629e-01
7.85820782e-01 1.86754227e-01 7.89155960e-01 1.25749946e+00
-4.88406569e-01 6.62622035e-01 9.18613911e-01 -2.84395963e-01
3.02796632e-01 3.91530134e-02 -1.94006041e-01 6.95290387e-01
2.52917469e-01 3.28813642e-02 -5.78524828e-01 -9.34798717e-02
1.28267908e+00 2.65886158e-01 -2.48124093e-01 -1.17325819e+00
-1.05676389e+00 5.81383228e-01 6.05490983e-01 2.39330474e-02
-2.90924400e-01 2.04273894e-01 1.58294514e-02 2.60995984e-01
4.98124123e-01 3.12442213e-01 -5.78342676e-01 8.57085884e-02
-5.09702206e-01 1.71934500e-01 7.96036899e-01 1.42941904e+00
8.09673190e-01 -4.55299839e-02 2.34279230e-01 9.38780010e-01
2.96243727e-01 6.72426283e-01 -1.71431199e-01 -1.28467977e+00
4.45022404e-01 7.02230990e-01 1.75617293e-01 -1.13880825e+00
-2.28038371e-01 -3.62854809e-01 -8.54108989e-01 3.07822615e-01
5.04747272e-01 4.20412749e-01 -9.02713895e-01 1.53562438e+00
8.89125764e-01 2.75835872e-01 -2.50571370e-01 1.17623973e+00
9.38011646e-01 2.85268635e-01 -5.52424192e-01 1.23131521e-01
1.14479947e+00 -8.34359765e-01 -1.23540275e-01 -2.53032565e-01
1.87399834e-01 -7.07387686e-01 1.10732555e+00 2.57696062e-01
-1.14847219e+00 -5.34406781e-01 -1.00868833e+00 -3.85405213e-01
-1.84097931e-01 2.40904957e-01 7.09681094e-01 4.50811207e-01
-8.80307496e-01 5.07888496e-01 -8.02460730e-01 -3.26431781e-01
6.41191900e-01 4.91453022e-01 -5.45683801e-01 -4.11838681e-01
-5.83991766e-01 8.73580098e-01 2.45691076e-01 3.21063012e-01
-8.24710667e-01 -7.41149783e-01 -8.02967072e-01 -2.60918766e-01
7.38898873e-01 -8.79896581e-01 1.15559018e+00 -9.64883029e-01
-1.54378748e+00 1.13755476e+00 2.55367577e-01 5.92131317e-02
7.08799183e-01 -6.52322844e-02 9.05898213e-02 2.99206764e-01
-1.11127190e-01 8.26292932e-01 9.93243694e-01 -1.38010585e+00
-2.76358753e-01 -7.79670596e-01 3.39172155e-01 3.77504438e-01
-3.12073559e-01 -3.13885301e-01 -8.65136802e-01 -5.05864680e-01
5.54748654e-01 -1.01442969e+00 -1.85337007e-01 7.44338334e-01
-1.72876343e-01 -1.64893821e-01 7.48864532e-01 -4.11102742e-01
2.38350347e-01 -2.33603668e+00 2.22732887e-01 2.27449328e-01
3.44547451e-01 2.88625304e-02 -3.33354026e-01 9.91671607e-02
9.53668952e-02 -3.29466641e-01 -6.30358979e-02 -6.73624873e-01
2.54849583e-01 3.49502206e-01 -2.03128085e-01 7.42169559e-01
1.73659295e-01 9.08525527e-01 -6.19070947e-01 -3.52350354e-01
2.95965552e-01 6.60087228e-01 -4.37044442e-01 3.15195858e-01
-4.79660720e-01 3.86892051e-01 -4.87400323e-01 9.92862463e-01
1.04040456e+00 -4.36147332e-01 -9.41410065e-02 -6.43667102e-01
2.64153749e-01 8.39832332e-03 -1.56040072e+00 2.14229059e+00
-4.02953744e-01 2.37467095e-01 2.17171669e-01 -9.07802701e-01
1.11873496e+00 8.85924846e-02 6.21916354e-01 -6.45911396e-01
3.02991927e-01 4.25891995e-01 -3.47828478e-01 -3.93850476e-01
3.08552235e-01 -6.28113821e-02 2.08639145e-01 3.85779589e-01
1.80485606e-01 -4.45698589e-01 -1.99358389e-01 -6.84209317e-02
6.40049398e-01 4.44969952e-01 1.16795875e-01 9.35829058e-02
1.39730543e-01 -7.13558272e-02 5.76604307e-01 5.12457132e-01
-4.53459173e-02 8.95725250e-01 9.14534107e-02 -6.45377636e-01
-1.31426346e+00 -1.09751797e+00 -2.06299424e-01 5.83378434e-01
4.93711978e-01 -9.77304950e-02 -3.22872669e-01 -6.47342265e-01
2.63230473e-01 3.09525400e-01 -6.06551886e-01 1.02110747e-02
-5.12210011e-01 -3.04089278e-01 1.37238532e-01 6.81824207e-01
4.77062225e-01 -7.45027721e-01 -6.31922424e-01 -2.61609763e-01
-1.10945284e-01 -1.29404426e+00 -3.32718402e-01 1.11772515e-01
-1.04673171e+00 -1.07984900e+00 -8.70131850e-01 -8.24291468e-01
7.74219871e-01 5.84069073e-01 1.04616570e+00 -2.52459422e-02
-4.41155195e-01 6.83909237e-01 -2.84642607e-01 -2.24761084e-01
-4.66396175e-02 -1.06354699e-01 1.40441805e-01 -3.51391197e-03
3.28443468e-01 -9.27316368e-01 -7.53199041e-01 5.54053664e-01
-6.43847942e-01 3.66839141e-01 5.83726227e-01 5.03762722e-01
8.57115984e-01 -2.35329613e-01 9.31424871e-02 -2.77084053e-01
-1.77529469e-01 -1.34509951e-01 -8.09284031e-01 2.44981885e-01
-2.77612627e-01 -2.35357493e-01 2.86195934e-01 -6.77495480e-01
-8.60357344e-01 4.09431636e-01 -5.84281841e-03 -1.17300856e+00
-2.71396130e-01 -9.17181224e-02 -3.45092475e-01 -4.24777299e-01
4.74109054e-01 7.07269758e-02 2.49971807e-01 -7.61973619e-01
4.45392728e-01 3.64427686e-01 4.66058433e-01 -7.14953244e-01
1.09423506e+00 8.32828641e-01 1.86903924e-01 -7.21339881e-01
-8.70688915e-01 -5.82716048e-01 -8.45777810e-01 -2.50618070e-01
5.02149105e-01 -1.05070198e+00 -7.55156577e-01 4.37634021e-01
-1.28138459e+00 -3.73170853e-01 -5.84518433e-01 4.67019111e-01
-8.14461887e-01 2.98331290e-01 -4.78339940e-01 -6.77455008e-01
-2.19413310e-01 -8.95204961e-01 1.38632882e+00 -2.00664066e-02
3.10458422e-01 -5.55502355e-01 -1.02403544e-01 3.80181760e-01
1.43072173e-01 4.37032640e-01 4.97395337e-01 -3.22508276e-01
-1.04223943e+00 -4.21373919e-02 -4.42887485e-01 3.95130783e-01
2.31517032e-02 -1.50562018e-01 -1.01524782e+00 -2.99068213e-01
3.17806810e-01 -4.89979416e-01 4.95581537e-01 3.38152289e-01
1.20680714e+00 3.87894660e-02 -1.07376121e-01 9.91916060e-01
1.60899913e+00 -1.33804470e-01 4.20077473e-01 1.47825852e-01
1.00964665e+00 8.34703445e-01 6.88523710e-01 4.75563467e-01
3.36442769e-01 8.89972925e-01 7.44275987e-01 -7.18405321e-02
-3.61809582e-01 -2.89129019e-01 -2.02122219e-02 7.91996479e-01
-3.36578459e-01 -1.38986155e-01 -7.55071402e-01 1.90580457e-01
-1.65649164e+00 -6.35019898e-01 -2.88620554e-02 2.24519849e+00
5.05799830e-01 -8.70327652e-02 1.16174370e-01 -1.45885989e-01
6.15921021e-01 5.02093695e-02 -8.41212928e-01 4.17317003e-01
-2.37422824e-01 4.57254834e-02 3.59200656e-01 2.03250244e-01
-8.21743011e-01 8.25135171e-01 5.01642656e+00 8.13776314e-01
-9.89965260e-01 1.33437410e-01 4.98916328e-01 -1.54865429e-01
-1.84466347e-01 -4.89431880e-02 -8.17615032e-01 9.90724415e-02
2.33735561e-01 3.15047920e-01 4.69852567e-01 1.01990402e+00
-2.74365276e-01 2.47553363e-02 -1.28809607e+00 1.41674030e+00
3.80201101e-01 -1.19586384e+00 -1.00096613e-01 1.15196280e-01
7.03718245e-01 1.86122492e-01 5.33424541e-02 -6.24767914e-02
2.93934103e-02 -6.67150795e-01 9.81707394e-01 5.08901596e-01
9.78852093e-01 -5.19183695e-01 3.51430893e-01 3.32763433e-01
-1.16448212e+00 -1.22215981e-02 -7.52812505e-01 -3.39965485e-02
7.80855715e-02 5.33932865e-01 -7.53375888e-01 6.07685328e-01
9.50199306e-01 8.51468623e-01 -6.57619357e-01 1.15515411e+00
-8.63301530e-02 -9.19123292e-02 -6.92084908e-01 4.47972119e-02
-1.05780356e-01 -2.91252822e-01 4.57989722e-01 4.61375386e-01
4.34774220e-01 1.44479543e-01 3.04415617e-02 1.03426898e+00
-9.81342345e-02 -7.16647506e-02 -6.17895901e-01 2.12155372e-01
3.87118906e-01 1.26331747e+00 -9.53666925e-01 -3.29866298e-02
-3.56263310e-01 1.05683625e+00 4.28340614e-01 3.30134034e-01
-7.92386830e-01 -3.10721584e-02 4.78973955e-01 1.54437959e-01
5.94263256e-01 -3.93107682e-01 -1.33695111e-01 -1.32557833e+00
4.81083721e-01 -8.65888119e-01 1.75967649e-01 -1.05456519e+00
-1.57575130e+00 4.57816392e-01 1.25727609e-01 -1.58312690e+00
2.06134945e-01 -8.27600121e-01 -7.28051737e-02 5.10881066e-01
-1.36156785e+00 -1.40822804e+00 -7.10216284e-01 7.10819840e-01
4.70236510e-01 3.14299017e-02 5.57304204e-01 3.62913191e-01
-2.49443710e-01 4.17405576e-01 -9.79675278e-02 -8.62759575e-02
6.38155222e-01 -8.98518443e-01 3.53650689e-01 3.93528610e-01
2.23292872e-01 4.24313635e-01 3.83524626e-01 -4.05839115e-01
-1.83296573e+00 -9.32511091e-01 4.14012611e-01 -6.44233465e-01
3.71339500e-01 -8.89476955e-01 -7.25363493e-01 6.04242861e-01
-3.11971605e-01 4.12270665e-01 2.71212637e-01 -8.56033638e-02
-5.74302733e-01 -3.45924556e-01 -1.14807296e+00 5.32240570e-01
1.62169206e+00 -5.48357189e-01 -4.35004056e-01 3.77867043e-01
8.16174924e-01 -7.76443660e-01 -9.42810059e-01 6.37715638e-01
6.93253219e-01 -9.90264297e-01 1.29247677e+00 -3.61812264e-01
3.19919586e-01 -3.88572633e-01 -5.40079296e-01 -8.71102393e-01
-9.13731605e-02 -3.91272455e-01 -3.56625408e-01 1.19347632e+00
-6.82717934e-02 -5.48600078e-01 1.07442641e+00 6.65213168e-01
-1.27627000e-01 -8.94918978e-01 -1.07439029e+00 -9.82127905e-01
-2.40611538e-01 -4.52021152e-01 7.09357858e-01 8.77712607e-01
-7.60928869e-01 1.94647670e-01 -4.12963063e-01 3.66214216e-01
7.96996295e-01 5.08870423e-01 1.17588294e+00 -1.30185950e+00
-4.69801754e-01 -2.26605833e-01 -6.01116538e-01 -1.34183562e+00
-1.81638021e-02 -8.21810961e-01 -3.73795107e-02 -1.16279328e+00
1.95905387e-01 -8.14751863e-01 3.35118547e-02 3.07635278e-01
2.76260316e-01 5.77390969e-01 2.09539860e-01 2.68992037e-01
-5.94287038e-01 8.61919284e-01 1.47597623e+00 -3.21475975e-02
-8.98984149e-02 1.28438592e-01 -3.74110252e-01 5.88424325e-01
6.31146967e-01 -4.40459222e-01 -5.26675940e-01 -6.59112990e-01
3.02830487e-01 -3.62451524e-02 8.73265564e-01 -8.67114544e-01
2.00950667e-01 -1.44958243e-01 4.55627799e-01 -7.80751050e-01
9.16538775e-01 -1.18522847e+00 4.73681360e-01 1.30675673e-01
-5.20443767e-02 -6.98128790e-02 2.51047034e-02 6.62395716e-01
1.30648032e-01 -2.52220154e-01 6.90528691e-01 -2.56768942e-01
-7.98163652e-01 9.14065599e-01 5.37700355e-01 1.99541897e-02
1.11347342e+00 -6.09798670e-01 -5.97271211e-02 -2.17844963e-01
-5.66323817e-01 4.22962680e-02 9.12001550e-01 4.97710407e-01
7.63482869e-01 -1.48251593e+00 -4.23579425e-01 3.51973385e-01
3.40930253e-01 5.60220540e-01 3.95788372e-01 8.62961173e-01
-5.16260087e-01 -8.03873613e-02 -2.74292409e-01 -9.52684879e-01
-1.19722986e+00 6.29863977e-01 5.42262673e-01 3.51310223e-01
-7.98076630e-01 1.00454938e+00 2.32929006e-01 -7.98502743e-01
6.40995681e-01 -3.24874580e-01 3.17412972e-01 -1.67412356e-01
2.52474487e-01 4.00980502e-01 1.16398372e-01 -6.65255964e-01
-2.54767716e-01 9.51199889e-01 1.95688695e-01 1.27628729e-01
1.75203538e+00 -1.38224661e-01 -8.39445367e-02 4.78496671e-01
1.37380373e+00 -2.85665661e-01 -1.70097172e+00 -4.28669542e-01
-2.97083825e-01 -8.52263153e-01 -7.51328468e-02 -4.69603747e-01
-1.30012679e+00 8.51084292e-01 7.02895999e-01 -2.13029027e-01
1.04916418e+00 3.56021971e-01 6.15549386e-01 5.95193982e-01
6.03432298e-01 -1.01852965e+00 5.15114546e-01 2.73659378e-01
1.12365305e+00 -1.32074332e+00 8.21670219e-02 -7.13428378e-01
-3.70888203e-01 1.22658813e+00 8.28722775e-01 -2.91263938e-01
5.85998654e-01 7.44032636e-02 -7.41489455e-02 -4.71632540e-01
-3.77838761e-01 -9.02290791e-02 4.34833080e-01 7.87579954e-01
-2.67439485e-01 -2.36367017e-01 3.80038768e-01 2.26810172e-01
-1.64627843e-02 -1.83273196e-01 1.38403237e-01 7.91159928e-01
-8.96637961e-02 -9.91760254e-01 -3.34738314e-01 2.67593980e-01
9.18443426e-02 2.96731025e-01 -4.37156379e-01 8.88131320e-01
2.27749690e-01 3.63071948e-01 1.13951176e-01 -2.19915107e-01
4.63243157e-01 -2.31197625e-01 1.09065711e+00 -5.37556529e-01
-7.02944025e-02 1.11923821e-01 -1.17143027e-01 -7.01308608e-01
-7.11825967e-01 -5.13452947e-01 -6.90355480e-01 -1.75052032e-01
-5.40613472e-01 -4.02910113e-01 7.12527514e-01 5.81445098e-01
4.47063059e-01 -4.39451123e-03 6.84930742e-01 -1.33590770e+00
-8.05656314e-01 -6.49911463e-01 -6.77928925e-01 4.20740545e-01
2.03582972e-01 -9.70733583e-01 -4.44456041e-01 -1.54023260e-01] | [7.676603317260742, -2.7748820781707764] |
f5257cc7-86d7-456b-891e-e2c96212e968 | mapping-process-for-the-task-wikidata | 2210.12659 | null | https://arxiv.org/abs/2210.12659v1 | https://arxiv.org/pdf/2210.12659v1.pdf | Mapping Process for the Task: Wikidata Statements to Text as Wikipedia Sentences | Acknowledged as one of the most successful online cooperative projects in human society, Wikipedia has obtained rapid growth in recent years and desires continuously to expand content and disseminate knowledge values for everyone globally. The shortage of volunteers brings to Wikipedia many issues, including developing content for over 300 languages at the present. Therefore, the benefit that machines can automatically generate content to reduce human efforts on Wikipedia language projects could be considerable. In this paper, we propose our mapping process for the task of converting Wikidata statements to natural language text (WS2T) for Wikipedia projects at the sentence level. The main step is to organize statements, represented as a group of quadruples and triples, and then to map them to corresponding sentences in English Wikipedia. We evaluate the output corpus in various aspects: sentence structure analysis, noise filtering, and relationships between sentence components based on word embedding models. The results are helpful not only for the data-to-text generation task but also for other relevant works in the field. | ['Grigori Sidorov', 'Alexander Gelbukha', 'Hoang Thang Ta'] | 2022-10-23 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [ 2.61745062e-02 5.71009278e-01 7.66253918e-02 -1.36208400e-01
-7.09149241e-01 -4.16077882e-01 6.46152496e-01 5.79017699e-01
-4.66528475e-01 1.05909085e+00 7.91373551e-01 -2.27093875e-01
1.56696245e-01 -1.13442886e+00 -2.78559148e-01 -1.08309865e-01
2.67615378e-01 2.53057629e-01 1.27695993e-01 -6.82344377e-01
5.28149188e-01 -1.49276420e-01 -1.52080929e+00 3.88723940e-01
1.51814520e+00 3.50511879e-01 6.07840359e-01 3.07293177e-01
-1.12904084e+00 1.05807674e+00 -8.09520006e-01 -1.00111997e+00
-1.27268329e-01 -5.17585337e-01 -9.16743457e-01 -1.81451306e-01
1.79485843e-01 5.65851256e-02 6.93509802e-02 1.25730586e+00
6.10224068e-01 -1.28376499e-01 4.78643298e-01 -1.05599403e+00
-9.89489555e-01 1.32115567e+00 -4.20477241e-01 -9.68608931e-02
6.08027279e-01 -2.95008570e-01 1.23181117e+00 -1.09915233e+00
1.04885912e+00 9.87065196e-01 5.01094580e-01 5.96633852e-01
-9.48088586e-01 -2.68685073e-01 -1.77599996e-01 1.77912325e-01
-1.05362976e+00 -3.46679240e-01 6.94989622e-01 -6.39427900e-01
1.14658940e+00 3.50474536e-01 5.62415957e-01 7.84879744e-01
7.75839761e-02 6.12281799e-01 5.87213516e-01 -8.11858773e-01
3.00775729e-02 5.00099778e-01 -3.06548569e-02 7.20523536e-01
5.78200877e-01 -9.58009064e-01 -6.09031081e-01 -3.22111882e-04
2.67408162e-01 -3.83106530e-01 -2.61550695e-01 -4.72302474e-02
-1.38850260e+00 9.43314612e-01 1.47654563e-01 5.63622177e-01
-4.64232594e-01 -1.92628145e-01 7.57861853e-01 2.86098927e-01
7.41903961e-01 7.79037178e-01 -2.15617672e-01 -5.05970299e-01
-6.43806636e-01 4.85282004e-01 9.96466696e-01 1.04745209e+00
6.26071274e-01 -1.12262383e-01 -3.02253127e-01 1.02925718e+00
2.10561126e-01 3.95478100e-01 7.06084311e-01 -6.25583231e-01
1.14060259e+00 9.71715271e-01 1.82997763e-01 -1.37487614e+00
-1.21453330e-01 -1.60188526e-01 -7.38328457e-01 -3.41831833e-01
2.01270524e-02 -5.27083099e-01 -1.47479802e-01 1.42206788e+00
1.72305778e-01 -6.55167878e-01 2.55429566e-01 5.21760702e-01
1.23668945e+00 9.44627285e-01 -6.45628059e-03 -1.89018592e-01
1.35143030e+00 -1.02246451e+00 -1.04841948e+00 -4.35727954e-01
9.27779019e-01 -8.57056916e-01 1.05486929e+00 -9.05173719e-02
-1.03089786e+00 -3.67039502e-01 -8.69861722e-01 -2.66674012e-01
-5.46227992e-01 4.18688178e-01 5.29715896e-01 5.05564094e-01
-1.05921197e+00 2.85139591e-01 -4.80775446e-01 -6.70715332e-01
2.02658802e-01 -3.20234239e-01 -3.71535450e-01 4.49580774e-02
-1.54724514e+00 1.19368267e+00 5.74877739e-01 -5.06367423e-02
8.22991505e-02 -5.66974461e-01 -1.11388075e+00 1.17972501e-01
2.70898342e-01 -7.50624776e-01 8.69626582e-01 -4.94971961e-01
-1.01581287e+00 8.45505536e-01 -2.02384755e-01 -3.58419597e-01
3.51125211e-01 -1.05581217e-01 -3.58196169e-01 -2.09986076e-01
6.37863696e-01 5.45474887e-01 1.28361776e-01 -7.45218813e-01
-7.51973689e-01 -1.58415228e-01 1.15697496e-01 4.59507167e-01
-1.02573454e+00 2.87215173e-01 -4.08835948e-01 -6.19806051e-01
-2.56414622e-01 -5.00871122e-01 -3.11302155e-01 -2.61227340e-01
-6.84014797e-01 -5.84454954e-01 2.39995211e-01 -1.13174057e+00
1.61285496e+00 -1.93847132e+00 2.35456228e-01 -1.01558551e-01
9.34675783e-02 4.95897532e-02 -3.20369869e-01 1.14880538e+00
1.15788698e-01 4.94860917e-01 -5.17965965e-02 -1.83488131e-01
2.27998242e-01 -2.02062860e-01 -5.22630751e-01 -3.13958436e-01
3.02762985e-01 8.31269562e-01 -1.11222494e+00 -7.09922493e-01
-1.41091451e-01 1.15613170e-01 -2.52816975e-01 1.63145170e-01
-1.71594784e-01 -2.47547209e-01 -6.13568187e-01 1.50569957e-02
3.35864842e-01 -3.59587148e-02 2.99724191e-01 -5.00575900e-02
-4.45334077e-01 6.29942477e-01 -8.59851599e-01 1.56916177e+00
-7.60158956e-01 8.35135758e-01 -1.94861561e-01 -7.91473210e-01
1.23994756e+00 3.06221157e-01 3.68203700e-01 -6.85393453e-01
-1.93098515e-01 1.99321568e-01 -1.94720954e-01 -8.62724900e-01
1.20891511e+00 3.52266543e-02 -3.43768597e-01 5.78316927e-01
-7.30611570e-03 -3.80718976e-01 1.02849019e+00 7.35215545e-01
9.40322101e-01 -1.30723894e-01 4.10847604e-01 -2.06899345e-01
5.38594663e-01 2.49300554e-01 2.64097661e-01 4.25339818e-01
4.39659685e-01 3.30468059e-01 8.00165594e-01 -1.83300331e-01
-1.28184223e+00 -7.33725131e-01 1.54952228e-01 6.93257689e-01
-1.34966105e-01 -7.72844493e-01 -8.12027454e-01 -5.31240642e-01
-7.55707026e-02 1.05720687e+00 -4.59959954e-01 1.22658052e-01
-3.63674611e-01 -4.54993635e-01 4.17837203e-01 2.78927356e-01
4.82785493e-01 -1.37010336e+00 -1.71194598e-01 3.87405485e-01
-8.57165635e-01 -9.74433482e-01 -4.05574024e-01 -2.33465105e-01
-4.88596231e-01 -8.06699038e-01 -7.63256609e-01 -1.18959653e+00
9.40614045e-01 3.05147529e-01 1.15319002e+00 4.26234156e-02
-3.79125923e-01 1.12753026e-01 -6.64515913e-01 -6.19598508e-01
-6.83819890e-01 2.47785360e-01 -9.10688713e-02 -3.93143624e-01
4.04406369e-01 -9.72260162e-02 3.68555672e-02 -2.33644128e-01
-9.46727514e-01 3.86749506e-01 3.13188076e-01 6.46570206e-01
9.39529836e-02 1.65796906e-01 8.95615160e-01 -1.09320879e+00
1.36963224e+00 -4.68827814e-01 -2.80478805e-01 6.32285893e-01
-4.62894738e-01 1.18504703e-01 6.68000519e-01 1.36968300e-01
-1.31767845e+00 -3.27107966e-01 -1.98200569e-01 7.36127675e-01
3.29941034e-01 1.20156431e+00 -7.11356774e-02 4.59046781e-01
8.58047843e-01 2.80255824e-01 7.79658854e-02 -2.17339173e-01
6.03957415e-01 1.05721092e+00 1.96569234e-01 -4.50492203e-01
8.21059167e-01 -4.15519960e-02 -5.22335112e-01 -1.06763196e+00
-7.85235703e-01 -2.91332722e-01 -3.32111686e-01 -3.94435585e-01
8.39175820e-01 -9.62163091e-01 -3.12145054e-01 1.09421819e-01
-1.55665302e+00 6.42977208e-02 -3.91128361e-01 8.93449038e-02
-9.73594114e-02 5.91965735e-01 -4.54609364e-01 -5.98588765e-01
-5.43748558e-01 -7.57726073e-01 6.22979164e-01 2.96664417e-01
-4.01044995e-01 -1.13578403e+00 2.16987729e-01 6.03053808e-01
5.70227325e-01 1.45268347e-02 1.06421602e+00 -5.16228080e-01
-2.60620207e-01 -1.72121540e-01 -2.95526177e-01 4.50236410e-01
3.17718863e-01 1.84467316e-01 -2.13474184e-01 5.02171330e-02
-1.95542723e-01 -3.64279121e-01 5.80774307e-01 -8.46111998e-02
8.49418342e-01 -5.89391589e-01 -8.46095905e-02 -2.28135571e-01
1.37785518e+00 -1.93561077e-01 8.77877235e-01 6.40751660e-01
5.05449116e-01 1.00896645e+00 6.00497961e-01 5.55575132e-01
7.31686413e-01 5.83145142e-01 1.58955038e-01 8.58175308e-02
-1.10722393e-01 -5.21622419e-01 5.00116706e-01 1.53239322e+00
4.58982661e-02 -3.23435634e-01 -1.00699735e+00 8.86096895e-01
-1.77819467e+00 -1.25441813e+00 -4.80699092e-01 1.92138708e+00
1.18344474e+00 -3.32440548e-02 -2.32785657e-01 -2.75680900e-01
8.37276161e-01 1.73780859e-01 1.23656921e-01 -4.16247576e-01
-1.80941522e-01 -1.21464141e-01 1.42753735e-01 2.40519732e-01
-7.07330763e-01 7.59421825e-01 5.29867744e+00 7.85988569e-01
-8.77312839e-01 -7.06144646e-02 2.27866709e-01 3.35629106e-01
-8.21245253e-01 1.75031670e-03 -7.47304082e-01 6.26253605e-01
6.21095657e-01 -1.15150940e+00 2.14942530e-01 9.27849710e-01
3.75328302e-01 -8.56196359e-02 -6.90154850e-01 7.53734827e-01
3.79110515e-01 -1.78258264e+00 2.54535556e-01 -2.69966394e-01
8.97125840e-01 -4.41710763e-02 -2.70642668e-01 4.71876323e-01
3.59162480e-01 -6.22619927e-01 7.14088619e-01 3.95700693e-01
5.34385383e-01 -6.22179270e-01 1.11426890e+00 5.93230367e-01
-1.11109364e+00 1.57129481e-01 -7.67680585e-01 -9.58352089e-02
3.93899381e-01 1.01000583e+00 -8.55286539e-01 7.92365968e-01
4.77763027e-01 6.14377797e-01 -6.54742599e-01 9.58412230e-01
-5.13881445e-01 3.18901330e-01 9.59291086e-02 -8.44788074e-01
-6.29344285e-02 -4.36498493e-01 4.04281765e-01 1.28671229e+00
7.13544846e-01 -2.62600183e-01 1.83670726e-02 8.29701006e-01
-4.23159659e-01 8.17951322e-01 -9.16612923e-01 -5.75639248e-01
7.81270802e-01 1.40422082e+00 -5.65720260e-01 -2.21954212e-01
-5.44188619e-01 8.80126059e-01 6.78779483e-01 1.15759388e-01
-4.81730849e-01 -1.26111925e+00 3.30900252e-01 7.68493116e-02
5.41465655e-02 -1.28824666e-01 -2.55934894e-01 -1.25306237e+00
5.80003083e-01 -6.95197940e-01 -8.42358470e-02 -8.80192280e-01
-1.23604822e+00 6.25237525e-01 -4.08213079e-01 -1.18294656e+00
-2.74801135e-01 -3.20560932e-01 -8.07388365e-01 1.00200307e+00
-1.19396293e+00 -8.68181229e-01 -2.02816769e-01 -1.65672883e-01
5.45550227e-01 -2.94593424e-01 9.26900148e-01 3.66772652e-01
-4.82434958e-01 3.06321293e-01 3.13260496e-01 4.63705450e-01
6.15494728e-01 -1.20453262e+00 6.82948649e-01 9.47060585e-01
2.09064424e-01 9.49640453e-01 6.67292774e-01 -8.25169861e-01
-1.13364792e+00 -1.16804624e+00 1.99777818e+00 -3.18112731e-01
1.12304902e+00 -4.89811182e-01 -5.74525237e-01 3.13864946e-01
6.82639599e-01 -6.99643135e-01 6.59468889e-01 2.72376120e-01
4.11627106e-02 3.94571945e-03 -7.69052148e-01 8.60616148e-01
6.37284756e-01 -5.30747354e-01 -7.15127409e-01 6.32688344e-01
8.55308354e-01 -3.19291174e-01 -8.65981340e-01 -2.86387712e-01
6.78388253e-02 -6.53854132e-01 5.80222130e-01 -4.86452430e-01
1.13852930e+00 -2.31659934e-01 2.09698707e-01 -1.57537568e+00
-1.50118724e-01 -5.11700094e-01 4.10298258e-01 1.76054835e+00
9.75293696e-01 -4.29333568e-01 7.19427228e-01 7.23187268e-01
-2.40672305e-01 -6.00332618e-01 -7.71971524e-01 -5.90184331e-01
1.17409527e-02 -1.97428674e-01 4.37344193e-01 9.40902352e-01
7.21158206e-01 7.16194987e-01 -3.34545076e-01 -4.46707875e-01
3.76268536e-01 -6.71399431e-03 7.81687915e-01 -1.12732482e+00
1.04159102e-01 -4.27383453e-01 -1.94763407e-01 -7.86269009e-01
3.55005227e-02 -1.20395005e+00 1.94919989e-01 -2.39707470e+00
5.27469814e-01 -2.29661688e-01 4.91420388e-01 4.38779503e-01
-4.52185631e-01 -9.21946615e-02 3.24202448e-01 -5.49287610e-02
-5.52449346e-01 5.96670687e-01 1.13896847e+00 -3.82968992e-01
-1.53400507e-02 -3.10453326e-01 -1.14786124e+00 3.17200422e-01
6.74367785e-01 -3.97479296e-01 -3.23537678e-01 -6.34013712e-01
1.03130817e+00 -3.98173779e-02 -1.00099921e-01 -7.04949915e-01
2.55120724e-01 -1.21247187e-01 -2.38788828e-01 -3.72868866e-01
-1.44064113e-01 -4.47346091e-01 2.53541991e-02 2.90815830e-01
-6.40434444e-01 9.89006832e-02 -6.80924766e-03 1.27658218e-01
-5.20665288e-01 -9.16879177e-01 1.38722092e-01 -9.56070274e-02
-5.77148974e-01 -2.07021475e-01 -5.23200929e-01 3.68438810e-01
9.46442544e-01 4.16456982e-02 -7.82614112e-01 -5.60054898e-01
-2.99473733e-01 4.13796812e-01 3.37939918e-01 7.08454549e-01
5.57251453e-01 -1.29800856e+00 -1.16598117e+00 -2.51978308e-01
3.58839184e-01 -1.44692317e-01 3.47868413e-01 5.78615665e-01
-7.88931668e-01 5.02246022e-01 -2.28941649e-01 1.81734934e-01
-1.17980790e+00 1.22006968e-01 -3.46721798e-01 -5.40122271e-01
-4.17759717e-01 7.63780057e-01 -3.72866213e-01 -6.81273520e-01
1.76230390e-02 4.94751008e-03 -7.30205238e-01 3.48290980e-01
7.34403968e-01 2.74778575e-01 2.08310690e-02 -4.68330503e-01
-9.82584525e-03 1.20758325e-01 -2.17506617e-01 7.25753326e-03
1.62391174e+00 -2.71756835e-02 -8.23323071e-01 3.23240042e-01
1.05999720e+00 4.02041167e-01 -2.40328059e-01 -5.45492694e-02
2.62301773e-01 -2.16332301e-01 -2.40417525e-01 -6.24775171e-01
-7.78164506e-01 7.42196083e-01 -1.09783329e-01 7.40014493e-01
5.19176722e-01 8.70891958e-02 7.59939373e-01 6.73071444e-01
4.28701729e-01 -1.34803009e+00 1.12616621e-01 7.25851059e-01
1.05931592e+00 -1.10071945e+00 1.13650501e-01 -5.43640077e-01
-9.31443870e-01 1.27614594e+00 6.11164808e-01 3.87051135e-01
2.38345742e-01 1.88020453e-01 1.50418356e-01 -5.55379428e-02
-6.86978579e-01 -4.27634232e-02 2.37171333e-02 7.11575329e-01
9.55112576e-01 -1.61725432e-01 -1.03589451e+00 7.03681529e-01
-4.10871178e-01 -1.19066283e-01 9.53112602e-01 7.18849540e-01
-7.53540218e-01 -1.47573781e+00 6.62204921e-02 8.71919513e-01
-3.50698054e-01 -4.35765266e-01 -4.91044819e-01 3.89184147e-01
-1.14203475e-01 9.87129867e-01 -3.87503998e-03 -2.10043669e-01
4.23670411e-01 -4.88698184e-02 2.72032823e-02 -1.03186131e+00
-5.87840855e-01 -6.80705488e-01 7.04480529e-01 5.85915260e-02
-3.29331964e-01 -6.98013902e-01 -1.41517186e+00 -3.77841890e-01
-3.51294041e-01 7.07095206e-01 9.86280560e-01 6.63784921e-01
4.30623591e-01 4.68830794e-01 5.74503779e-01 -4.55201149e-01
-5.70284903e-01 -1.17043114e+00 -3.87096435e-01 4.76166129e-01
-5.22458553e-01 -1.31265193e-01 -2.35199019e-01 2.18973324e-01] | [12.144311904907227, 9.470080375671387] |
c9e2095e-4614-447c-8367-064a492f01a3 | neighbortrack-improving-single-object | 2211.06663 | null | https://arxiv.org/abs/2211.06663v2 | https://arxiv.org/pdf/2211.06663v2.pdf | NeighborTrack: Improving Single Object Tracking by Bipartite Matching with Neighbor Tracklets | We propose a post-processor, called NeighborTrack, that leverages neighbor information of the tracking target to validate and improve single-object tracking (SOT) results. It requires no additional data or retraining. Instead, it uses the confidence score predicted by the backbone SOT network to automatically derive neighbor information and then uses this information to improve the tracking results. When tracking an occluded target, its appearance features are untrustworthy. However, a general siamese network often cannot tell whether the tracked object is occluded by reading the confidence score alone, because it could be misled by neighbors with high confidence scores. Our proposed NeighborTrack takes advantage of unoccluded neighbors' information to reconfirm the tracking target and reduces false tracking when the target is occluded. It not only reduces the impact caused by occlusion, but also fixes tracking problems caused by object appearance changes. NeighborTrack is agnostic to SOT networks and post-processing methods. For the VOT challenge dataset commonly used in short-term object tracking, we improve three famous SOT networks, Ocean, TransT, and OSTrack, by an average of ${1.92\%}$ EAO and ${2.11\%}$ robustness. For the mid- and long-term tracking experiments based on OSTrack, we achieve state-of-the-art ${72.25\%}$ AUC on LaSOT and ${75.7\%}$ AO on GOT-10K. Code duplication can be found in https://github.com/franktpmvu/NeighborTrack. | ['Hong-Yuan Mark Liao', 'Yung-Yu Chuang', 'Youn-Long Lin', 'Hung-Shuo Chang', 'Cheng-Yun Yang', 'Chien-Yao Wang', 'Yu-Hsi Chen'] | 2022-11-12 | null | null | null | null | ['visual-object-tracking'] | ['computer-vision'] | [-2.58356631e-01 -2.82083213e-01 -2.76582748e-01 -1.97623730e-01
-7.42915273e-01 -9.30638313e-01 1.64482638e-01 -1.85201854e-01
-4.51344818e-01 6.68860495e-01 -1.57360911e-01 -1.09744363e-01
-7.55584985e-02 -5.34708440e-01 -1.01047778e+00 -4.74622011e-01
-4.10078406e-01 4.69325602e-01 7.89600253e-01 2.55542696e-01
-2.26693630e-01 4.60403442e-01 -1.33741498e+00 4.48568314e-02
5.72729886e-01 1.25858688e+00 -1.55643895e-01 5.94373107e-01
1.09155312e-01 5.73032975e-01 -6.15906239e-01 -3.52291554e-01
6.77682519e-01 1.65611282e-02 -2.38660052e-01 -4.11635786e-01
1.27347982e+00 -3.67649227e-01 -5.46820879e-01 1.33340430e+00
3.94671798e-01 -1.19955316e-01 2.80146956e-01 -1.59695208e+00
-3.94036680e-01 6.34707510e-01 -8.34973633e-01 4.66067761e-01
-4.45333496e-02 6.30624592e-01 7.42270112e-01 -7.40916908e-01
6.93076074e-01 1.18656135e+00 1.18307817e+00 4.78083611e-01
-1.09234476e+00 -1.49831903e+00 4.92422640e-01 3.20135243e-02
-1.35783565e+00 -4.34603125e-01 3.10386449e-01 -4.31611121e-01
4.11194712e-01 2.65943468e-01 7.12661028e-01 1.03442407e+00
2.19514266e-01 6.78669631e-01 8.33066761e-01 2.71039158e-01
-7.87288919e-02 -1.77334085e-01 3.00665438e-01 7.95280635e-01
6.20879591e-01 7.79295564e-01 -4.23688322e-01 -5.50780520e-02
3.76860052e-01 8.88545737e-02 -2.44555846e-01 -3.72199655e-01
-1.37728024e+00 4.55460250e-01 9.13550496e-01 4.68755793e-03
-8.27921405e-02 5.29844999e-01 2.63415039e-01 1.96150467e-01
3.04759532e-01 1.64530933e-01 -5.65040112e-01 -4.75608781e-02
-1.02054226e+00 8.93047825e-02 3.65897179e-01 1.00099552e+00
6.24702036e-01 2.90436476e-01 -4.93434221e-01 3.33167970e-01
5.11287212e-01 1.12667716e+00 -6.49553984e-02 -1.14096379e+00
3.32838714e-01 4.95695740e-01 3.70027483e-01 -9.37227368e-01
-4.08305258e-01 -8.63694131e-01 -4.54042494e-01 5.24694443e-01
6.54080212e-01 -3.86654586e-01 -1.14912236e+00 1.92426717e+00
7.49133766e-01 6.01535320e-01 -1.82228729e-01 1.11138153e+00
8.83177996e-01 3.17188472e-01 4.24126573e-02 -1.43168643e-01
1.26023734e+00 -1.00184572e+00 -4.35717970e-01 -2.35298485e-01
5.40825069e-01 -8.95013094e-01 4.81398195e-01 2.11101711e-01
-5.72495401e-01 -7.39196181e-01 -8.88281405e-01 5.44493735e-01
-1.50565475e-01 1.87910050e-01 5.01984835e-01 4.63159859e-01
-9.33174014e-01 6.67992234e-01 -1.03963780e+00 -2.40661755e-01
7.42022872e-01 4.23390776e-01 -2.58223623e-01 -1.01079345e-01
-8.99861813e-01 6.46748066e-01 1.82241976e-01 2.15522140e-01
-1.14943111e+00 -1.10732973e+00 -5.97440720e-01 -2.36229673e-01
8.97702038e-01 -5.28441489e-01 1.17920113e+00 -6.09522045e-01
-9.79465961e-01 3.39734018e-01 -1.34142756e-01 -6.78169787e-01
8.59412134e-01 -4.23203260e-01 -6.44072831e-01 -1.69083372e-01
4.00706828e-01 8.79393399e-01 9.29196894e-01 -1.08907890e+00
-9.54243481e-01 -3.64828169e-01 -1.97545752e-01 -2.26724535e-01
5.96074238e-02 -2.25609899e-01 -8.48490059e-01 -8.13057423e-01
1.10558070e-01 -1.35142756e+00 -3.57608348e-02 7.44119227e-01
-5.35960019e-01 -1.31335989e-01 1.19895446e+00 -3.64905924e-01
8.72869730e-01 -2.24184680e+00 -5.43884158e-01 1.91410467e-01
5.28104186e-01 5.59779704e-01 -3.13771337e-01 -1.28904715e-01
7.27803856e-02 -7.64627978e-02 2.00402543e-01 -1.40413865e-01
5.43786809e-02 -4.09044549e-02 -1.67164728e-01 6.63181484e-01
-2.29985803e-01 1.06649506e+00 -8.67267728e-01 -4.69842851e-01
2.49208584e-01 5.74633896e-01 -3.47485632e-01 -1.75853148e-01
-4.31165129e-01 4.47309941e-01 -5.64137697e-01 7.58697987e-01
8.01771700e-01 -4.07869726e-01 -8.56382847e-02 -4.77048129e-01
-2.96511322e-01 -2.14830399e-01 -1.35721183e+00 1.47091138e+00
2.32330441e-01 7.42471218e-01 3.70216668e-02 -2.69930333e-01
6.77958190e-01 1.46608986e-02 6.84178114e-01 -5.44752598e-01
2.68817753e-01 1.03021964e-01 1.68261126e-01 -2.20416505e-02
3.09639722e-01 3.63729447e-01 5.77533320e-02 2.30615482e-01
-2.12649703e-01 6.01564288e-01 2.35627547e-01 3.38527113e-01
1.36476231e+00 3.56631070e-01 -3.15151691e-01 -2.01794267e-01
1.14860147e-01 3.97592664e-01 1.04568779e+00 9.67716098e-01
-5.40484548e-01 2.50235438e-01 -1.56738590e-02 -4.82272923e-01
-5.64544022e-01 -1.21444726e+00 -7.27504939e-02 9.32942569e-01
5.37358522e-01 -3.78899187e-01 -3.34620714e-01 -8.67763400e-01
5.73851645e-01 3.64139557e-01 -7.52975821e-01 -2.19942674e-01
-4.83680487e-01 -2.78804839e-01 6.94334984e-01 6.16797507e-01
5.69084346e-01 -7.94284344e-01 -4.34094995e-01 1.94187671e-01
2.45067347e-02 -1.15095031e+00 -9.15272653e-01 -1.94832683e-01
-7.06882238e-01 -1.25125003e+00 -4.39457715e-01 -1.43287942e-01
4.66175437e-01 6.22030854e-01 7.94705153e-01 3.15768540e-01
-2.43157670e-01 3.10224175e-01 -1.93838164e-01 -3.03916425e-01
-1.73994973e-01 -1.19563229e-01 4.24417973e-01 -9.24677253e-02
1.98819056e-01 -2.19245702e-01 -7.28698194e-01 8.93106639e-01
-2.34340504e-01 -4.16796297e-01 5.81518471e-01 3.05902064e-01
8.93000484e-01 1.22138813e-01 2.68117279e-01 -6.85320318e-01
-3.08006704e-01 -3.29703510e-01 -1.19521165e+00 1.73507646e-01
-5.54024160e-01 -5.83415441e-02 2.91497499e-01 -7.95230091e-01
-6.76636696e-01 1.32023424e-01 2.29412854e-01 -1.30746639e+00
5.88954166e-02 -2.88855992e-02 8.33198428e-02 -4.55207139e-01
5.74628651e-01 -1.39486626e-01 7.75090605e-03 -5.17137647e-01
1.87257051e-01 2.66377116e-03 7.24111259e-01 -3.73582780e-01
1.43271554e+00 6.35772884e-01 -7.97362253e-02 -2.39113182e-01
-1.04939198e+00 -3.15784454e-01 -1.89029828e-01 -5.47817945e-01
7.83053219e-01 -1.11427164e+00 -1.12879467e+00 3.33910048e-01
-8.59465837e-01 -3.87895018e-01 -3.58594686e-01 7.24670529e-01
2.24198312e-01 2.01689437e-01 -3.80461276e-01 -6.26332223e-01
-3.92071366e-01 -1.05921841e+00 8.07043493e-01 4.83345866e-01
-3.65232565e-02 -5.55728674e-01 -3.98388840e-02 2.70440608e-01
5.68846345e-01 3.34776521e-01 2.23377533e-03 -5.77582717e-01
-1.09236610e+00 -3.36331338e-01 -4.25166219e-01 -2.48623616e-03
2.96511855e-02 1.45824909e-01 -8.25666130e-01 -8.24233115e-01
-6.48225963e-01 -6.16523772e-02 9.80598748e-01 5.19530237e-01
8.36553097e-01 -1.84179470e-01 -8.71179402e-01 6.47940576e-01
1.14707172e+00 1.61115140e-01 1.37622252e-01 9.11646262e-02
8.68237615e-01 5.20708188e-02 9.62648213e-01 1.46023467e-01
1.85709447e-01 8.06258440e-01 7.87675440e-01 1.38103828e-01
-6.14366055e-01 -2.60276854e-01 5.96106589e-01 2.57036686e-01
1.32409751e-01 -1.98967919e-01 -7.33056486e-01 4.40836161e-01
-1.83271921e+00 -1.04085696e+00 -4.37855870e-01 2.16309261e+00
4.63791817e-01 5.17319858e-01 2.78728843e-01 -5.23310721e-01
1.03611302e+00 3.21497470e-02 -1.05924988e+00 6.23529911e-01
-5.78690926e-03 -1.18129671e-01 8.90170336e-01 4.30226535e-01
-1.15662444e+00 1.15622282e+00 4.93721008e+00 8.92110229e-01
-1.12879205e+00 1.66482329e-01 8.51049572e-02 -4.46415812e-01
3.01555693e-01 -4.55873013e-02 -1.32079136e+00 6.76727176e-01
7.03469574e-01 1.33518167e-02 7.36483485e-02 9.76902962e-01
5.22615090e-02 -1.21774711e-01 -8.97480965e-01 7.34806836e-01
-1.84393391e-01 -1.38527620e+00 -4.30866390e-01 1.30421653e-01
5.52233875e-01 6.62899017e-01 1.75167084e-01 4.82605308e-01
6.87544048e-01 -5.20451486e-01 9.61640418e-01 3.51572335e-01
7.29465127e-01 -3.33950788e-01 6.59219742e-01 5.15827052e-02
-1.84068060e+00 9.92981158e-03 -1.51323900e-01 2.58594006e-01
1.15954988e-01 5.59151828e-01 -7.72275090e-01 5.23700953e-01
1.31666219e+00 9.14871097e-01 -8.60601664e-01 1.55066931e+00
-2.27963477e-01 5.89801908e-01 -8.37892830e-01 2.89681405e-01
3.07383817e-02 3.48461896e-01 1.14284539e+00 8.60567212e-01
2.77812004e-01 -6.48562685e-02 5.67915618e-01 6.55144095e-01
-1.78305000e-01 -4.40771461e-01 -4.18226302e-01 2.10297048e-01
8.11486602e-01 1.32471800e+00 -8.75862300e-01 -4.53058690e-01
-5.26235402e-02 4.43043113e-01 8.25821087e-02 2.52105564e-01
-1.42080867e+00 -1.62723083e-02 9.15842175e-01 1.21923365e-01
8.62786710e-01 1.12326881e-02 2.63046592e-01 -1.13845158e+00
1.47139579e-01 -5.67889035e-01 6.00865066e-01 -7.39878714e-01
-1.26645100e+00 7.37784803e-01 1.33008696e-03 -1.65587163e+00
1.84951946e-01 -3.79703313e-01 -4.03966367e-01 4.36350077e-01
-1.27095330e+00 -1.19324350e+00 -4.87117589e-01 5.51449597e-01
2.10272208e-01 -5.23506571e-03 1.36708483e-01 5.83869338e-01
-6.16545200e-01 8.31511736e-01 -1.42239094e-01 4.28639144e-01
9.51001465e-01 -8.17245662e-01 2.75325865e-01 1.08404481e+00
1.10363744e-01 4.82936203e-01 7.34787166e-01 -1.11816609e+00
-1.48819506e+00 -1.68675494e+00 3.58584434e-01 -7.34825253e-01
8.32806051e-01 -1.58776507e-01 -8.90198469e-01 9.45053816e-01
-2.10229158e-01 7.74642110e-01 2.09283113e-01 6.20499589e-02
-6.77281916e-01 -3.30274522e-01 -1.03597105e+00 4.66235608e-01
1.45472646e+00 1.01280464e-02 -2.52691060e-01 1.71979532e-01
9.53587353e-01 -6.79589212e-01 -1.03616989e+00 6.67126000e-01
8.30921054e-01 -6.48493886e-01 1.07402313e+00 -4.67202395e-01
-4.73179519e-01 -1.02820957e+00 -1.00247070e-01 -8.31867456e-01
-3.90307873e-01 -5.41371107e-01 -2.36933932e-01 1.16654551e+00
5.83669364e-01 -6.39235318e-01 1.13160133e+00 4.65818852e-01
-2.99830645e-01 -3.75947118e-01 -1.18552065e+00 -1.26011562e+00
-2.32988819e-01 -6.49900794e-01 4.79375124e-01 9.05580640e-01
-8.70220423e-01 1.38550267e-01 -5.26281953e-01 6.03233874e-01
1.24966013e+00 3.52992058e-01 1.01188934e+00 -1.46052718e+00
-1.75829619e-01 -1.80948928e-01 -2.93899983e-01 -1.00880897e+00
-7.87521377e-02 -8.02874088e-01 2.05731928e-01 -1.04827547e+00
3.80565748e-02 -8.18366468e-01 -3.25930804e-01 9.59831774e-01
-8.21040869e-02 6.96087956e-01 6.41956389e-01 4.09732729e-01
-1.20633531e+00 4.44186926e-01 1.16560256e+00 -2.65762478e-01
-4.61584143e-02 2.08128408e-01 -4.35543925e-01 6.60030544e-01
7.82755077e-01 -1.13596225e+00 2.61627184e-03 -5.61287165e-01
1.12386812e-02 6.31206408e-02 8.35901082e-01 -1.42355442e+00
5.22862792e-01 1.02059402e-01 6.11048937e-01 -1.07259059e+00
3.27265501e-01 -1.17347980e+00 6.80772543e-01 7.60037422e-01
3.95361111e-02 1.15445703e-01 5.71949482e-01 8.91539991e-01
2.41872266e-01 1.71924725e-01 9.55111563e-01 2.05373541e-01
-8.03516090e-01 7.92062521e-01 8.00250098e-02 1.64776191e-01
1.11905253e+00 -3.07777256e-01 -7.04426885e-01 -8.22993740e-02
-6.77971840e-01 7.45336950e-01 5.29809237e-01 6.42547607e-01
2.54661202e-01 -1.42840099e+00 -5.33015430e-01 -6.23369664e-02
7.70083219e-02 -1.78023577e-01 3.68264049e-01 1.15881932e+00
-1.84705220e-02 1.95408791e-01 -3.78510356e-02 -1.10697997e+00
-1.49300933e+00 5.22817433e-01 4.27706718e-01 -1.49523854e-01
-7.96128869e-01 8.91827106e-01 5.48286214e-02 -2.62509346e-01
3.78805280e-01 -3.27589303e-01 2.25347266e-01 -2.05805078e-01
5.59450209e-01 3.51008594e-01 -1.48813471e-01 -7.19256163e-01
-7.09322274e-01 7.75070488e-01 -2.80157417e-01 1.59391627e-01
9.19831216e-01 6.89675882e-02 3.20321172e-01 -8.00983980e-03
9.45696712e-01 -2.45537125e-02 -1.84191859e+00 -4.70829666e-01
-1.68024659e-01 -6.20633125e-01 1.25082489e-02 -1.10470927e+00
-1.68975091e+00 3.67875606e-01 1.02225721e+00 -1.17555879e-01
6.57671392e-01 4.36100736e-02 8.47413599e-01 3.25890690e-01
5.81489444e-01 -5.98163903e-01 4.59466614e-02 4.65488940e-01
4.51330930e-01 -1.36625206e+00 8.95518512e-02 -2.57934153e-01
-2.51788914e-01 6.48270786e-01 1.05278575e+00 -6.01008534e-02
6.93569064e-01 4.17538881e-01 1.53854907e-01 -3.17251503e-01
-6.78330123e-01 -2.70718783e-01 4.46308821e-01 5.06308019e-01
-2.27518663e-01 -1.00970507e-01 3.32805783e-01 2.30903342e-01
-1.46212623e-01 4.52717207e-02 -6.53367564e-02 9.03248429e-01
-4.67746496e-01 -7.06279039e-01 -6.73418343e-01 4.93149728e-01
-3.57166767e-01 1.69880748e-01 -2.40416557e-01 1.04645085e+00
3.36749434e-01 7.98842251e-01 4.69232500e-02 -5.79214513e-01
4.59126681e-01 -3.81829828e-01 2.54655719e-01 -2.21374795e-01
-6.25704587e-01 2.36298323e-01 6.29311800e-02 -9.07192528e-01
-4.64098722e-01 -7.55887568e-01 -1.32459784e+00 -6.32717848e-01
-5.60816169e-01 3.01147997e-01 3.73162478e-01 6.72530472e-01
7.01420724e-01 5.76720178e-01 1.61618501e-01 -8.35106015e-01
-2.29884803e-01 -7.67184615e-01 -1.12363651e-01 1.63546801e-01
4.86330777e-01 -1.13688135e+00 -2.82532245e-01 -1.83583945e-01] | [6.353831768035889, -2.1097798347473145] |
1891a55c-98e9-445b-bb4e-f8a664c04b0b | l0learn-a-scalable-package-for-sparse | 2202.04820 | null | https://arxiv.org/abs/2202.04820v2 | https://arxiv.org/pdf/2202.04820v2.pdf | L0Learn: A Scalable Package for Sparse Learning using L0 Regularization | We present L0Learn: an open-source package for sparse linear regression and classification using $\ell_0$ regularization. L0Learn implements scalable, approximate algorithms, based on coordinate descent and local combinatorial optimization. The package is built using C++ and has user-friendly R and Python interfaces. L0Learn can address problems with millions of features, achieving competitive run times and statistical performance with state-of-the-art sparse learning packages. L0Learn is available on both CRAN and GitHub (https://cran.r-project.org/package=L0Learn and https://github.com/hazimehh/L0Learn). | ['Tim Nonet', 'Rahul Mazumder', 'Hussein Hazimeh'] | 2022-02-10 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [-8.84058297e-01 -2.82097340e-01 -3.64425272e-01 -5.07608950e-01
-1.38916433e+00 -4.31478262e-01 1.48436517e-01 2.35915333e-01
-1.68368459e-01 6.77131176e-01 4.43843603e-01 -1.83082357e-01
7.97568038e-02 -6.59906149e-01 -7.31663465e-01 -7.46331871e-01
-2.88587600e-01 5.09928048e-01 -3.04307908e-01 5.14300577e-02
3.47823091e-02 1.44395128e-01 -1.33405077e+00 4.36903507e-01
3.77702802e-01 9.83122885e-01 1.50476679e-01 8.85077417e-01
-1.31917283e-01 8.59824955e-01 1.51521489e-01 -2.03951567e-01
4.72760648e-01 -2.93994337e-01 -4.85454649e-01 -6.12619936e-01
4.37093675e-01 1.09490141e-01 -3.12246799e-01 9.17193592e-01
7.15528548e-01 1.05682351e-01 4.54190105e-01 -1.04982996e+00
-6.56834126e-01 4.44456905e-01 -6.53987169e-01 3.97209257e-01
4.27657574e-01 4.40981776e-01 1.30594504e+00 -1.71817052e+00
2.44067147e-01 1.28286564e+00 1.12595284e+00 1.50885507e-01
-1.49130344e+00 -9.69962478e-01 -4.27746437e-02 -1.27297521e-01
-1.73779714e+00 -6.76091135e-01 4.79283184e-01 -4.64757532e-01
1.01342368e+00 2.91696578e-01 5.38070977e-01 7.93048739e-01
5.54550774e-02 9.52841103e-01 1.14515805e+00 -1.25582159e-01
3.59513134e-01 2.81802956e-02 3.62008601e-01 1.05136013e+00
-3.50682177e-02 1.94380227e-02 -8.95031154e-01 -6.50865674e-01
9.66466010e-01 9.05178562e-02 4.82503548e-02 -2.26916090e-01
-1.16536570e+00 1.15190089e+00 4.48953539e-01 1.85197115e-01
-3.28950495e-01 4.71608728e-01 5.26152372e-01 2.54730463e-01
9.27886546e-01 2.02967763e-01 -7.06285655e-01 -3.87969732e-01
-9.88717854e-01 4.05558586e-01 1.00744951e+00 9.97156560e-01
8.25312316e-01 3.01456958e-01 2.31471717e-01 1.14700198e+00
2.48962045e-01 5.01517117e-01 5.65659106e-01 -1.25625563e+00
1.02662779e-01 1.75490141e-01 -4.76968102e-02 -1.12645411e+00
-5.01908422e-01 -6.66339755e-01 -9.13401365e-01 -5.01971468e-02
2.82346547e-01 -2.62576818e-01 -3.92529011e-01 1.28084898e+00
5.39947212e-01 4.75333363e-01 -2.37755015e-01 9.11362588e-01
1.31385863e+00 7.25339711e-01 -5.56152239e-02 -1.14157751e-01
9.04142380e-01 -1.40569866e+00 -3.51300955e-01 -3.49995852e-01
9.11799967e-01 -9.30339396e-01 8.75523031e-01 4.67961162e-01
-1.19587672e+00 -4.38680083e-01 -4.84025300e-01 -3.28406423e-01
-4.90409553e-01 1.53565168e-01 1.17940509e+00 2.32633561e-01
-1.22790635e+00 3.53628367e-01 -1.31045306e+00 -2.82487005e-01
5.59692979e-01 9.41630825e-02 -5.07775664e-01 -4.69091117e-01
-7.14763045e-01 6.84796989e-01 5.06738871e-02 -1.37098599e-03
-1.09309530e+00 -1.12435412e+00 -1.23843181e+00 -2.39158481e-01
2.40289658e-01 -6.47548735e-01 1.29750133e+00 -5.65384388e-01
-1.00725937e+00 8.56734216e-01 -6.57318592e-01 -4.66433838e-02
1.28523901e-01 -4.93252814e-01 -1.38814405e-01 -2.49431759e-01
2.31333479e-01 2.27058455e-01 6.65933728e-01 -7.76504815e-01
-8.70127231e-02 -4.36014831e-01 -8.28048110e-01 1.18897840e-01
-9.24377814e-02 3.34378153e-01 -5.96263170e-01 -6.78478122e-01
8.28379542e-02 -7.79215157e-01 -6.05959594e-01 1.67309806e-01
-1.66679427e-01 -2.19679549e-01 4.40886050e-01 -8.78682375e-01
1.17421377e+00 -2.05883217e+00 1.48726583e-01 2.90581763e-01
3.36737722e-01 7.04413280e-02 -3.97674292e-01 6.32718742e-01
-3.51573795e-01 -2.29721919e-01 -2.70454168e-01 -5.93767464e-01
-9.88392085e-02 -7.48131946e-02 1.34463266e-01 1.03773749e+00
-2.02060983e-01 9.29275692e-01 -9.53606248e-01 -3.46939236e-01
2.74037123e-01 7.58018494e-01 -8.09269369e-01 2.79993773e-01
-9.43921581e-02 5.03792942e-01 -3.99851263e-01 9.57037985e-01
6.25156105e-01 -4.84193087e-01 1.27736881e-01 2.22645566e-01
-3.65882874e-01 2.62767375e-01 -1.48174441e+00 2.01393247e+00
-4.88085181e-01 3.50143194e-01 7.33168602e-01 -1.06321990e+00
8.65232766e-01 5.66225760e-02 5.53065836e-01 -6.20348565e-02
1.41545266e-01 3.13446909e-01 -7.45024502e-01 -4.05432403e-01
1.13057114e-01 2.35340521e-01 -9.94380414e-02 4.81532156e-01
2.92559236e-01 -2.25762367e-01 1.54123262e-01 3.56247753e-01
1.34351623e+00 4.02472407e-01 6.66964769e-01 -4.97822553e-01
1.52756229e-01 1.41752347e-01 5.51478744e-01 7.81338573e-01
2.66011477e-01 4.19009537e-01 3.97838980e-01 -7.32786000e-01
-1.05320001e+00 -1.01875389e+00 -4.82187659e-01 1.69490695e+00
-5.77488244e-01 -8.95503402e-01 -2.85685688e-01 -1.55541301e-01
6.08019769e-01 6.80973053e-01 -4.90409344e-01 3.88326079e-01
-2.44652495e-01 -1.02624321e+00 5.12561142e-01 4.55019206e-01
-1.05659902e-01 -9.48335826e-01 2.34838203e-01 1.05720580e-01
1.45030364e-01 -6.53546691e-01 -4.49519008e-01 5.21911204e-01
-9.33314383e-01 -8.55588198e-01 -4.93323326e-01 -6.77383602e-01
7.64476299e-01 2.06011519e-01 1.37264192e+00 7.61741176e-02
-7.33630359e-01 5.61386406e-01 -3.39369059e-01 -2.34625161e-01
1.91677764e-01 3.00392568e-01 3.01730037e-01 -3.43889385e-01
5.60371578e-01 -7.70762920e-01 -6.41344607e-01 8.78037885e-02
-4.70294863e-01 -1.76051736e-01 3.00987095e-01 9.29050744e-01
1.02952945e+00 -2.56423265e-01 4.44436163e-01 -1.07652700e+00
5.51393211e-01 -1.19980621e+00 -9.04202521e-01 -2.11896092e-01
-6.36606693e-01 -1.28462598e-01 6.79442227e-01 5.52156866e-02
-4.46027189e-01 3.50994319e-01 -5.77189863e-01 -6.74272835e-01
-1.27020285e-01 7.82131553e-01 2.28981018e-01 -5.12778759e-02
6.77029431e-01 4.49732393e-01 2.44786859e-01 -8.12808394e-01
6.77463055e-01 6.38287067e-01 2.60265708e-01 -6.38953149e-01
6.92011893e-01 1.79363251e-01 -2.73901373e-01 -1.01995349e+00
-1.22165179e+00 -8.72885644e-01 -5.51982164e-01 4.04648148e-02
4.32354867e-01 -1.52820158e+00 -6.26934409e-01 3.58120143e-01
-6.08848453e-01 -6.69091463e-01 -3.25150251e-01 4.36559975e-01
-5.65908790e-01 1.00338846e-01 -5.82073271e-01 -5.45673132e-01
-4.30103868e-01 -8.29937518e-01 9.37430143e-01 -6.05463609e-02
-4.66514081e-01 -1.16134655e+00 5.80043554e-01 4.49702829e-01
6.69156849e-01 1.27036020e-01 2.93200940e-01 -8.64602149e-01
-3.01314920e-01 -4.16442811e-01 2.06517261e-02 3.90449136e-01
-1.59269929e-01 -8.73872265e-02 -8.69172573e-01 -4.74119127e-01
-1.94314480e-01 -8.78626585e-01 9.50494051e-01 6.65463209e-01
1.37414587e+00 -7.45424867e-01 -2.14422598e-01 1.22186375e+00
1.52418137e+00 -6.16665304e-01 4.35462058e-01 4.86131608e-02
6.75904095e-01 2.41631866e-01 4.91678685e-01 8.36562693e-01
5.37739098e-01 4.59635377e-01 1.05109081e-01 -1.84021518e-01
1.01104148e-01 -2.82006443e-01 3.72780770e-01 1.21188772e+00
1.52113393e-01 6.45910561e-01 -1.05965650e+00 5.42501867e-01
-2.19864869e+00 -1.08587086e+00 -2.83707231e-01 2.00771523e+00
1.13091791e+00 -3.74028027e-01 1.68581769e-01 -4.39704508e-01
3.24727774e-01 3.87569070e-01 -6.67177439e-01 -3.83666426e-01
-1.03441663e-01 5.18253326e-01 5.19089818e-01 1.07866454e+00
-9.43268955e-01 1.14972150e+00 5.91938972e+00 1.16499293e+00
-1.14027488e+00 6.11925423e-01 8.11289132e-01 -8.28021169e-01
-1.17474154e-01 -2.17036605e-02 -1.03109217e+00 3.99294257e-01
1.00128114e+00 -2.47045621e-01 8.95848751e-01 1.39629984e+00
3.83653522e-01 -1.46275491e-01 -7.70118833e-01 1.11746550e+00
-2.36784443e-02 -1.49424458e+00 -7.63540864e-01 -1.68466240e-01
6.53080702e-01 1.00131834e+00 8.89217257e-02 5.58385134e-01
8.04762065e-01 -1.36312282e+00 4.59631354e-01 5.62363327e-01
9.90220845e-01 -6.68307960e-01 5.48406482e-01 2.95570821e-01
-1.33781087e+00 1.31992906e-01 -5.68926156e-01 -2.41659939e-01
-1.73407033e-01 1.23938763e+00 -4.81911749e-01 2.31378809e-01
1.10808671e+00 1.29946017e+00 -7.21155286e-01 9.75828052e-01
-1.80127949e-01 9.43084002e-01 -5.50622761e-01 7.61077926e-02
-1.00076184e-01 -4.63544756e-01 4.42748517e-01 1.57661283e+00
1.98938727e-01 3.11624557e-01 5.19687474e-01 7.58770764e-01
-3.79438102e-02 4.74141479e-01 -7.22338200e-01 5.76607920e-02
5.52028835e-01 1.61583889e+00 -3.53148609e-01 -9.72232223e-02
-7.34046519e-01 5.09893537e-01 8.30614269e-01 3.07157785e-01
-6.49970353e-01 -3.24147612e-01 8.84411037e-01 2.67454624e-01
4.30805117e-01 -5.23539603e-01 -4.23437059e-01 -1.43282139e+00
-3.28670710e-01 -1.01946723e+00 6.60321891e-01 -8.48142147e-01
-1.60826147e+00 4.01026726e-01 -3.54341477e-01 -8.02388966e-01
-8.80300719e-03 -5.99853992e-01 -4.40968573e-01 8.68693709e-01
-1.11977518e+00 -9.68669116e-01 -2.74384230e-01 7.63832211e-01
4.40150559e-01 -2.32086256e-01 1.09110260e+00 8.22641850e-02
-7.44879246e-01 4.50045496e-01 6.43196821e-01 1.09081142e-01
8.17014039e-01 -1.06561446e+00 1.83083862e-01 3.16564202e-01
6.00870661e-02 8.57111454e-01 6.38410747e-01 -5.42227566e-01
-1.80186343e+00 -1.27702391e+00 8.97112012e-01 -5.98310769e-01
1.02204657e+00 -5.26705801e-01 -7.65051365e-01 1.07313025e+00
1.03830747e-01 5.66609204e-01 1.21302533e+00 6.12514019e-01
-6.76992595e-01 -2.42613748e-01 -1.06309044e+00 2.63381392e-01
7.51354635e-01 -4.99871314e-01 -1.81747288e-01 7.75198281e-01
3.35221529e-01 -4.17161703e-01 -1.25722909e+00 1.34770915e-01
6.92424715e-01 -8.72224510e-01 9.93214846e-01 -4.02182370e-01
3.84535730e-01 -1.38761476e-01 -3.51914316e-01 -1.25250626e+00
-6.12287343e-01 -6.73289239e-01 -3.44745070e-01 7.70502985e-01
5.88962615e-01 -9.06500041e-01 7.12328374e-01 3.83254170e-01
-1.25702336e-01 -1.16848099e+00 -1.06549871e+00 -5.19198775e-01
1.52303249e-01 -7.11421192e-01 2.00245798e-01 1.08159673e+00
2.79020876e-01 1.33394361e-01 -4.22203571e-01 -5.10916971e-02
7.83879995e-01 1.54005334e-01 1.03558731e+00 -7.17145562e-01
-5.21270514e-01 -2.19417170e-01 -1.99029803e-01 -8.11768055e-01
3.41385722e-01 -1.36237693e+00 -2.70848334e-01 -1.21561122e+00
2.84052610e-01 -7.16293693e-01 -3.73995990e-01 1.16995156e+00
-9.16092016e-04 4.70181704e-01 -8.91841650e-02 3.86308581e-01
-8.58673692e-01 5.86704612e-01 7.88420141e-01 -3.88061702e-02
-1.53768718e-01 -4.08689342e-02 -9.16031301e-01 7.54945934e-01
9.34679210e-01 -8.42865407e-01 2.94541031e-01 -3.81588727e-01
3.66692752e-01 1.38065025e-01 2.46253118e-01 -7.27088571e-01
2.68549502e-01 -1.56712726e-01 6.01294994e-01 -3.98618400e-01
4.62088645e-01 -5.37146389e-01 7.50387311e-02 3.31628889e-01
-4.25306022e-01 2.89269507e-01 3.51542979e-01 2.42610231e-01
-1.19560868e-01 -2.00590611e-01 8.72191966e-01 -5.13543010e-01
-3.25625211e-01 4.61529970e-01 -2.06865996e-01 2.69902915e-01
8.46577823e-01 4.78417307e-01 -3.60156447e-01 -5.77349663e-01
-8.76137793e-01 3.82505745e-01 4.91665840e-01 1.80176452e-01
6.10770047e-01 -1.21300912e+00 -1.02455842e+00 3.65646213e-01
4.77643833e-02 -3.41394544e-02 5.51217236e-02 1.21760142e+00
-6.54077709e-01 3.98278087e-01 2.40651086e-01 -4.31202233e-01
-1.15239620e+00 5.06828189e-01 2.95052171e-01 -1.75849870e-01
-5.32216609e-01 1.25467598e+00 -2.67388940e-01 -1.01282752e+00
3.02123964e-01 -7.41232857e-02 2.91034341e-01 -1.58822834e-01
6.99438572e-01 6.04220450e-01 -5.69534674e-02 -4.64745373e-01
-4.68559265e-01 3.81158888e-01 4.44972813e-02 3.25522482e-01
1.93644178e+00 2.58694082e-01 -7.57810295e-01 5.65468132e-01
1.33627498e+00 3.62224132e-01 -8.97093952e-01 -2.73782611e-01
-3.38506877e-01 -4.21426266e-01 2.06104383e-01 -4.79497910e-01
-1.21859443e+00 4.07231569e-01 2.56140947e-01 -2.07117185e-01
8.37536097e-01 3.08363408e-01 4.41611350e-01 4.38245684e-01
4.50034291e-01 -7.51973808e-01 -9.81030464e-02 7.67529964e-01
1.05701566e+00 -1.44565094e+00 5.89131176e-01 -2.56074131e-01
-4.92118776e-01 7.17746615e-01 3.55588734e-01 -6.03177726e-01
1.24422538e+00 6.41537488e-01 1.33487061e-01 -3.89833838e-01
-1.11133575e+00 1.78332552e-02 9.54267010e-02 1.95737198e-01
9.12575960e-01 1.28610209e-01 -1.27334267e-01 9.95015204e-01
-4.83760536e-01 1.35004669e-01 1.23364247e-01 8.37140024e-01
-2.89110333e-01 -1.02589273e+00 -6.55135632e-01 8.29404354e-01
-5.43946385e-01 -5.30319691e-01 1.48553047e-02 3.24601918e-01
-1.78146213e-01 5.98550618e-01 -1.39899030e-01 -1.04457282e-01
-1.65930212e-01 1.45330369e-01 1.80506259e-01 -7.71708012e-01
-7.94054925e-01 2.86356211e-01 1.48458704e-01 -1.13077843e+00
6.29974380e-02 -9.79000270e-01 -1.08857524e+00 -4.60543782e-01
-1.63917784e-02 4.12886769e-01 8.28882575e-01 3.75661314e-01
6.84601307e-01 3.23747210e-02 5.92128932e-01 -1.07308114e+00
-3.25036705e-01 -1.01541173e+00 -9.28523242e-01 -1.19687170e-01
1.87499225e-01 -2.53603637e-01 -7.23498166e-01 -1.36940226e-01] | [7.2239580154418945, 4.414794445037842] |
296d7d08-4377-4637-b8e4-75b72f7ee0b8 | risa-net-rotation-invariant-structure-aware | 2010.00973 | null | https://arxiv.org/abs/2010.00973v1 | https://arxiv.org/pdf/2010.00973v1.pdf | RISA-Net: Rotation-Invariant Structure-Aware Network for Fine-Grained 3D Shape Retrieval | Fine-grained 3D shape retrieval aims to retrieve 3D shapes similar to a query shape in a repository with models belonging to the same class, which requires shape descriptors to be capable of representing detailed geometric information to discriminate shapes with globally similar structures. Moreover, 3D objects can be placed with arbitrary position and orientation in real-world applications, which further requires shape descriptors to be robust to rigid transformations. The shape descriptions used in existing 3D shape retrieval systems fail to meet the above two criteria. In this paper, we introduce a novel deep architecture, RISA-Net, which learns rotation invariant 3D shape descriptors that are capable of encoding fine-grained geometric information and structural information, and thus achieve accurate results on the task of fine-grained 3D object retrieval. RISA-Net extracts a set of compact and detailed geometric features part-wisely and discriminatively estimates the contribution of each semantic part to shape representation. Furthermore, our method is able to learn the importance of geometric and structural information of all the parts when generating the final compact latent feature of a 3D shape for fine-grained retrieval. We also build and publish a new 3D shape dataset with sub-class labels for validating the performance of fine-grained 3D shape retrieval methods. Qualitative and quantitative experiments show that our RISA-Net outperforms state-of-the-art methods on the fine-grained object retrieval task, demonstrating its capability in geometric detail extraction. The code and dataset are available at: https://github.com/IGLICT/RisaNET. | ['Lin Gao', 'Yu-Kun Lai', 'Fang-Lue Zhang', 'Jiawei Sun', 'Jie Yang', 'Rao Fu'] | 2020-10-02 | null | null | null | null | ['3d-object-retrieval'] | ['computer-vision'] | [-4.03934598e-01 -4.35187668e-01 -1.06473304e-01 -4.51746881e-01
-9.97588992e-01 -1.00303102e+00 9.37097609e-01 4.02268410e-01
1.92779407e-01 -2.76723281e-02 3.15727502e-01 1.94729820e-01
-5.49831867e-01 -1.20971060e+00 -5.89372039e-01 -6.15635753e-01
-5.98567910e-03 1.15292156e+00 9.26403478e-02 -9.50270593e-02
3.24962735e-01 1.63387632e+00 -1.62691569e+00 2.89093286e-01
2.24142760e-01 1.42459643e+00 4.60738912e-02 1.34097233e-01
-3.04811716e-01 -2.96870530e-01 -4.40785885e-01 3.39706242e-02
4.70991760e-01 4.73931760e-01 -6.78891838e-01 -1.19101204e-01
9.36932921e-01 -5.33238828e-01 -4.30374205e-01 7.73514688e-01
7.67978668e-01 1.44465202e-02 1.32883620e+00 -9.56043184e-01
-1.04395723e+00 -1.98590964e-01 -2.32706070e-01 -1.36508465e-01
3.24955672e-01 -6.32309020e-02 1.17789733e+00 -1.48691380e+00
7.28064120e-01 1.62783933e+00 3.97097081e-01 3.12916517e-01
-9.35403526e-01 -5.76580644e-01 -6.02153167e-02 -3.84918392e-01
-1.78560412e+00 -3.13517272e-01 9.21598554e-01 -3.44718784e-01
1.26083994e+00 2.95932680e-01 6.20376706e-01 6.43676996e-01
4.81921174e-02 7.39333510e-01 5.91968060e-01 3.16540934e-02
9.07526761e-02 -4.02921647e-01 -8.61535221e-02 7.76976109e-01
4.97833401e-01 1.38854817e-01 -2.33805463e-01 -4.00133342e-01
1.15049648e+00 4.37050164e-01 1.49348453e-01 -8.30826998e-01
-9.98267770e-01 6.14309609e-01 7.28683352e-01 5.95481873e-01
-3.71212274e-01 2.08525643e-01 1.97370186e-01 2.80963808e-01
6.69828594e-01 3.48908246e-01 -4.07178938e-01 9.24473479e-02
-7.96880305e-01 5.92873394e-01 5.91796279e-01 1.26630270e+00
7.68870175e-01 -6.17036037e-02 -4.02524740e-01 9.70679104e-01
5.93702078e-01 1.14090264e+00 1.57307342e-01 -7.67524183e-01
1.86873794e-01 1.18420255e+00 -1.35907009e-01 -1.10435879e+00
-3.47598523e-01 -3.05766344e-01 -8.50556195e-01 1.29495621e-01
4.15018760e-02 1.00101924e+00 -1.10083091e+00 1.34495258e+00
3.46939206e-01 -4.24227864e-01 -3.97077680e-01 1.09085703e+00
1.43002069e+00 4.64984089e-01 -1.05134912e-01 5.70579946e-01
1.33876801e+00 -4.58273500e-01 1.14477649e-01 1.58525810e-01
4.65768248e-01 -9.67910886e-01 9.74307775e-01 -3.92826915e-01
-1.21404684e+00 -5.09913027e-01 -8.07839394e-01 -3.53231519e-01
-7.63461113e-01 1.41359061e-01 5.57666421e-01 3.66611868e-01
-9.89511967e-01 5.70705950e-01 -5.65918207e-01 -2.34496623e-01
7.88293302e-01 3.74360353e-01 -6.02018118e-01 -2.87805438e-01
-8.13011825e-01 6.93567038e-01 2.10394681e-01 -1.33552387e-01
-9.00488913e-01 -8.22865188e-01 -1.03925991e+00 3.77553642e-01
-1.25429228e-01 -8.91689718e-01 8.58857930e-01 7.24461675e-02
-1.01869416e+00 1.42466056e+00 -1.17247671e-01 1.80370182e-01
7.70260245e-02 1.13250874e-01 2.83857770e-02 3.39424312e-01
1.77532002e-01 8.41517031e-01 9.54779565e-01 -1.28486848e+00
-3.44154797e-02 -7.44784117e-01 1.52776688e-01 1.34922013e-01
-1.24875970e-01 -2.43013740e-01 -5.65910995e-01 -8.97649884e-01
4.36602712e-01 -8.66146207e-01 2.09059209e-01 3.31971884e-01
-1.02530792e-01 -7.90374100e-01 9.41008031e-01 -1.52185103e-02
7.93467402e-01 -2.14459252e+00 -2.81766560e-02 4.59773093e-01
3.99656892e-01 2.98250079e-01 -7.48331964e-01 5.54123521e-01
-1.92518830e-02 3.56244832e-01 -8.74622241e-02 -2.65557289e-01
5.51768005e-01 6.74077217e-03 -4.17431444e-01 3.89698178e-01
4.26829159e-01 1.57153356e+00 -7.40440845e-01 -2.95729488e-01
4.27864909e-01 8.16242039e-01 -3.58138174e-01 2.67221183e-01
-1.62598819e-01 -1.47653427e-02 -1.05361235e+00 1.19249451e+00
9.14686620e-01 -3.97187740e-01 -4.45323944e-01 -5.47519267e-01
1.19468994e-01 2.61224389e-01 -8.83829713e-01 1.62443745e+00
-4.14022356e-01 1.65594831e-01 -1.98054284e-01 -6.00441575e-01
1.43360472e+00 7.14486092e-02 6.85963511e-01 -1.03289044e+00
9.03330669e-02 5.96661627e-01 -6.27323091e-01 1.35131523e-01
4.73816395e-01 2.39381753e-02 -3.76867712e-01 7.43656814e-01
3.47993076e-02 -9.42778707e-01 -1.64432392e-01 4.81883846e-02
8.63951623e-01 -5.31651415e-02 1.80741519e-01 -3.53856623e-01
5.11554539e-01 -3.40801060e-01 -1.50161870e-02 7.59303987e-01
2.96678636e-02 8.89677703e-01 -2.41833851e-01 -7.85309851e-01
-1.27353501e+00 -1.33462512e+00 -2.98268735e-01 8.50006402e-01
2.86029309e-01 -3.70767832e-01 -2.44006038e-01 -5.79729974e-01
6.56466961e-01 1.10840775e-01 -5.08613884e-01 -1.80068284e-01
-5.34221530e-01 -5.25559857e-03 5.17991602e-01 5.62598407e-01
2.47634649e-01 -1.21454942e+00 -3.98530275e-01 -1.83383465e-01
1.67179704e-01 -8.06398749e-01 -7.97724962e-01 -1.00578636e-01
-9.58259881e-01 -1.06334877e+00 -9.57146704e-01 -9.41579700e-01
9.39185560e-01 6.79454625e-01 1.34292138e+00 3.84627551e-01
-5.16516030e-01 8.48774910e-01 -3.53525817e-01 -2.23058790e-01
-1.45687073e-01 3.40061516e-01 -1.15256175e-01 -2.88483471e-01
5.44914901e-01 -6.01178885e-01 -6.89330101e-01 3.81079555e-01
-1.17802799e+00 -3.38790268e-01 6.52153850e-01 6.66489065e-01
1.11834359e+00 -3.71952541e-02 3.51757824e-01 -2.61150539e-01
5.63502252e-01 -1.44995987e-01 -5.62638521e-01 2.35775650e-01
-3.11686456e-01 3.20581853e-01 2.75813580e-01 -1.34590060e-01
-5.05630910e-01 -1.01763271e-01 -1.87537670e-01 -8.49309504e-01
-5.42131841e-01 2.85658419e-01 -2.06542030e-01 -3.20646197e-01
3.20448756e-01 4.28569019e-01 -1.61131471e-01 -9.25054491e-01
3.86774212e-01 5.22827446e-01 -7.65067115e-02 -9.37997937e-01
1.17250776e+00 6.61534250e-01 3.45589727e-01 -7.99273491e-01
-8.10754180e-01 -5.97036660e-01 -8.69231522e-01 1.54558290e-02
4.12185043e-01 -9.04882669e-01 -8.25409889e-01 5.25488198e-01
-1.09765637e+00 -3.56767066e-02 -4.80189085e-01 -6.16539316e-03
-7.12326109e-01 2.29411542e-01 -4.56475973e-01 -4.08021063e-01
-8.19046795e-01 -9.04392660e-01 2.23619843e+00 -1.41842887e-02
-1.15501724e-01 -8.61176670e-01 1.57399699e-02 9.41794664e-02
6.34365916e-01 1.74253598e-01 1.27466750e+00 -5.75418949e-01
-9.55535769e-01 -4.75662947e-01 -6.93202198e-01 -8.15960318e-02
3.62311125e-01 -2.81140208e-01 -7.77538955e-01 -6.70523286e-01
-3.43618542e-01 -4.42484319e-01 8.97847295e-01 3.15475136e-01
1.21401358e+00 -1.52356893e-01 -2.61224598e-01 5.49671769e-01
1.36265445e+00 -1.38071269e-01 3.04856092e-01 1.34521313e-02
7.32349992e-01 4.15961921e-01 4.86192822e-01 4.16177601e-01
3.55712414e-01 8.13508987e-01 4.88275617e-01 6.30175397e-02
-4.81422931e-01 -4.30280566e-01 -2.07440451e-01 8.69918704e-01
-1.68977808e-02 -8.70605707e-02 -8.28793943e-01 6.76333725e-01
-1.47569966e+00 -7.63867378e-01 3.34602714e-01 2.08736706e+00
5.96307218e-01 -1.45630136e-01 -1.25158966e-01 -2.05772996e-01
3.75465035e-01 3.86460364e-01 -6.26027465e-01 -1.98532104e-01
-3.37574124e-01 4.69414860e-01 1.78723976e-01 1.90793023e-01
-1.10846519e+00 9.55555081e-01 5.41235256e+00 1.22167969e+00
-1.11160934e+00 -2.75110543e-01 3.37297648e-01 4.28545661e-02
-7.67363727e-01 -3.58120650e-01 -8.42420995e-01 -7.60640949e-02
2.79729992e-01 -1.29211277e-01 1.78707376e-01 7.67798364e-01
-1.49941221e-01 4.88402039e-01 -1.11732614e+00 1.19262218e+00
3.12452111e-02 -1.39981651e+00 8.44694078e-01 7.50783756e-02
7.57342994e-01 2.07685515e-01 8.92940238e-02 7.79269040e-02
-4.93350141e-02 -1.25000024e+00 7.72383869e-01 7.18694091e-01
1.18683565e+00 -6.10000074e-01 4.53183919e-01 4.32336293e-02
-1.72853863e+00 2.97937721e-01 -6.85407639e-01 3.97679567e-01
-1.59955472e-01 5.58020413e-01 -6.72942162e-01 5.65995455e-01
6.18643820e-01 8.42268407e-01 -6.21583283e-01 1.19453776e+00
1.22821040e-01 -5.97697459e-02 -5.74902594e-01 -7.61702657e-02
4.53336865e-01 4.72461283e-02 6.95449293e-01 1.11071658e+00
5.74361086e-01 1.74819946e-01 2.23926648e-01 1.19597852e+00
-2.69158334e-01 2.04352364e-01 -9.60139215e-01 -3.26632649e-01
7.65417159e-01 1.15986419e+00 -6.72222018e-01 -1.47177488e-01
-1.54857337e-01 6.15089178e-01 2.76244551e-01 2.17912227e-01
-1.41968250e-01 -2.92887539e-01 7.12918282e-01 3.46561193e-01
5.79376042e-01 -5.46019912e-01 -2.14860454e-01 -9.08653438e-01
1.70362443e-01 -5.20112395e-01 3.37411523e-01 -8.92589271e-01
-1.70826542e+00 5.63429236e-01 3.40287238e-02 -1.41907752e+00
-1.48172649e-02 -8.22821200e-01 -1.58235252e-01 1.00007784e+00
-1.63680375e+00 -1.71511924e+00 -4.85464573e-01 6.07816696e-01
3.55336010e-01 -2.80074745e-01 1.19814360e+00 3.43030572e-01
3.17729831e-01 5.02536654e-01 3.30163166e-02 2.15671882e-01
7.10957885e-01 -9.93730962e-01 7.33763099e-01 -8.44014436e-02
3.37365150e-01 7.79343784e-01 -4.04707566e-02 -7.00934172e-01
-1.78608108e+00 -1.26911926e+00 9.46501076e-01 -5.80470979e-01
2.81885773e-01 -4.08033907e-01 -8.31076145e-01 1.56665787e-01
-5.67161262e-01 3.02993625e-01 4.19401556e-01 -5.08150645e-02
-9.88244951e-01 -2.34124176e-02 -1.16985428e+00 2.46218234e-01
1.37479877e+00 -9.73294735e-01 -7.10323632e-01 2.36776248e-01
6.27364576e-01 -3.57457131e-01 -1.27907777e+00 7.65506923e-01
8.53974223e-01 -5.94714165e-01 1.52684367e+00 -5.70593178e-01
2.36100242e-01 -2.28544816e-01 -5.46408474e-01 -1.06058717e+00
-6.34794235e-01 2.35008411e-02 -1.93369344e-01 7.40265489e-01
-3.46995704e-02 -5.12329102e-01 7.06395805e-01 3.40884924e-01
-8.16481709e-02 -9.24747825e-01 -9.40317988e-01 -9.57121849e-01
4.12686467e-01 -2.71072775e-01 1.14842284e+00 6.03179276e-01
-8.39382887e-01 5.58546558e-02 3.21200252e-01 1.83209851e-02
7.33256280e-01 1.13559699e+00 5.64762950e-01 -1.75241339e+00
3.33980322e-01 -1.02671993e+00 -7.56799042e-01 -1.23109424e+00
3.72374982e-01 -1.37688220e+00 -3.38229001e-01 -1.64171684e+00
3.26913655e-01 -8.55140507e-01 -3.45711857e-01 7.23983884e-01
3.77059549e-01 7.14922369e-01 3.44942749e-01 4.45959896e-01
-5.90663791e-01 8.94453406e-01 1.60215580e+00 -6.72012568e-01
5.55133335e-02 -5.66070266e-02 -4.92816091e-01 3.68968099e-01
4.87569511e-01 -2.85407424e-01 -1.85987413e-01 -5.39459109e-01
2.38754749e-01 -1.88449383e-01 7.17089653e-01 -4.16190714e-01
5.59398718e-02 5.39719909e-02 6.30848706e-01 -1.12326479e+00
5.85488915e-01 -1.07115006e+00 6.59141093e-02 1.34665683e-01
-6.62030056e-02 -1.68847755e-01 4.04642522e-01 2.16505781e-01
-3.50410879e-01 -8.67850855e-02 6.26361549e-01 -2.52746969e-01
-5.07326484e-01 1.00326490e+00 2.01562598e-01 1.32441536e-01
5.20955622e-01 -2.85034865e-01 -2.17643484e-01 -2.16138721e-01
-5.43901622e-01 -7.20146224e-02 6.55474007e-01 8.56838942e-01
1.05149198e+00 -2.05251217e+00 -7.00723648e-01 5.22124350e-01
5.66007793e-01 1.48151204e-01 2.95530111e-01 1.37677982e-01
-3.79104465e-01 9.74420607e-01 -2.92356670e-01 -7.06409216e-01
-1.17279351e+00 3.41192186e-01 3.73698831e-01 -1.60184100e-01
-5.43771148e-01 5.82631707e-01 5.18600821e-01 -9.56569016e-01
9.48626846e-02 -4.01602209e-01 2.66546514e-02 9.54686776e-02
3.37495506e-01 8.06889236e-02 4.18947577e-01 -9.09035325e-01
-5.88351250e-01 1.40161753e+00 1.31423712e-01 4.99220043e-01
1.70704877e+00 1.04425587e-01 -2.99336284e-01 7.72877485e-02
1.66753864e+00 -7.70685747e-02 -8.32866490e-01 -5.45604050e-01
-2.02915058e-01 -6.85806513e-01 -4.57315929e-02 -8.06595623e-01
-1.07987499e+00 8.61365736e-01 5.10592163e-01 7.69882202e-02
9.11483884e-01 5.15396595e-01 8.01104307e-01 6.62392974e-01
6.82307005e-01 -5.87072194e-01 1.48219749e-01 9.60644245e-01
1.51746178e+00 -1.06207180e+00 6.82551041e-02 -3.15052360e-01
1.67432785e-01 1.23032796e+00 2.34066308e-01 -3.50953609e-01
1.03799963e+00 8.22498128e-02 -2.10758984e-01 -8.66644621e-01
-4.60493922e-01 -3.14659059e-01 1.20229626e+00 4.57051456e-01
2.42207974e-01 2.73035944e-01 1.99661911e-01 3.28819096e-01
-1.00436665e-01 -5.70633769e-01 -4.28590745e-01 7.18060374e-01
-4.54771399e-01 -1.16155434e+00 -2.47847468e-01 4.64926034e-01
1.08429499e-01 3.13836448e-02 -7.90240109e-01 5.82861423e-01
-3.31276417e-01 2.67774493e-01 3.78391892e-02 1.64352860e-02
6.17346346e-01 -5.17787226e-02 7.43928015e-01 -6.28412604e-01
-3.00507754e-01 4.36176844e-02 -2.98767179e-01 -7.99034774e-01
-2.66057312e-01 -4.09994781e-01 -1.22874153e+00 -2.68575579e-01
8.82792752e-03 -1.85594067e-01 5.73702037e-01 7.41811156e-01
7.84250855e-01 1.79369748e-02 6.20136440e-01 -1.40810096e+00
-5.41556478e-01 -6.34990573e-01 -6.54273152e-01 8.36607397e-01
2.25915566e-01 -9.62233424e-01 -1.77712530e-01 -5.50664783e-01] | [8.141122817993164, -3.850538730621338] |
ee7bab5f-1f1b-487b-b81d-adf6a79106c2 | convolutional-lstms-for-cloud-robust | 1811.02471 | null | http://arxiv.org/abs/1811.02471v2 | http://arxiv.org/pdf/1811.02471v2.pdf | Convolutional LSTMs for Cloud-Robust Segmentation of Remote Sensing Imagery | Clouds frequently cover the Earth's surface and pose an omnipresent challenge
to optical Earth observation methods. The vast majority of remote sensing
approaches either selectively choose single cloud-free observations or employ a
pre-classification strategy to identify and mask cloudy pixels. We follow a
different strategy and treat cloud coverage as noise that is inherent to the
observed satellite data. In prior work, we directly employed a straightforward
\emph{convolutional long short-term memory} network for vegetation
classification without explicit cloud filtering and achieved state-of-the-art
classification accuracies. In this work, we investigate this cloud-robustness
further by visualizing internal cell activations and performing an ablation
experiment on datasets of different cloud coverage. In the visualizations of
network states, we identified some cells in which modulation and input gates
closed on cloudy pixels. This indicates that the network has internalized a
cloud-filtering mechanism without being specifically trained on cloud labels.
Overall, our results question the necessity of sophisticated pre-processing
pipelines for multi-temporal deep learning approaches. | ['Marco Körner', 'Marc Rußwurm'] | 2018-10-28 | null | null | null | null | ['segmentation-of-remote-sensing-imagery'] | ['miscellaneous'] | [ 4.91368830e-01 -4.88074332e-01 3.44756573e-01 -3.82145762e-01
-2.81684995e-01 -1.09952283e+00 6.88546896e-01 2.08031222e-01
-3.53185236e-01 6.82371259e-01 -3.56726319e-01 -7.26408601e-01
1.41191289e-01 -8.36350620e-01 -6.58147395e-01 -8.72206509e-01
-3.63409132e-01 1.02456763e-01 4.98548299e-02 -2.22955763e-01
3.17545235e-01 7.70273685e-01 -1.87470901e+00 6.24680221e-01
7.42009401e-01 7.46817172e-01 1.58921927e-01 8.09149384e-01
-2.09497049e-01 5.46787560e-01 -6.53141677e-01 2.64311910e-01
2.90634006e-01 -2.50852853e-01 -6.63611948e-01 -9.44318697e-02
7.36351490e-01 -1.90513115e-02 -7.83635601e-02 1.12201965e+00
4.60686654e-01 -2.40310222e-01 4.31008458e-01 -8.32623303e-01
-3.07737440e-01 4.00749534e-01 -4.41032141e-01 6.69349551e-01
-2.63504595e-01 6.75559223e-01 8.47765923e-01 -9.11785185e-01
5.72484910e-01 7.08105922e-01 7.29840398e-01 2.65290052e-01
-1.62673163e+00 -6.71836793e-01 5.65956950e-01 -3.69576603e-01
-1.64934754e+00 -6.35396183e-01 1.59071922e-01 -9.06210721e-01
1.28391492e+00 7.34619439e-01 1.02293003e+00 6.05017006e-01
5.24923243e-02 2.46705800e-01 1.69835126e+00 -1.69160590e-01
2.46115923e-01 -5.71640916e-02 5.43839633e-02 2.29849845e-01
4.19952840e-01 3.68103623e-01 -1.97820961e-01 6.18553068e-03
7.33718634e-01 1.47473872e-01 -3.02251875e-01 1.26021385e-01
-1.00219059e+00 5.27327240e-01 7.60481775e-01 4.72527772e-01
-1.77149117e-01 4.73727018e-01 1.78523183e-01 3.89504403e-01
7.00412273e-01 7.58826375e-01 -5.36687970e-01 4.11746293e-01
-1.41996789e+00 2.89200634e-01 3.07427257e-01 5.02391994e-01
9.97062206e-01 3.49745452e-01 -2.18732417e-01 2.08838895e-01
1.12319328e-01 5.90656459e-01 -7.91307166e-02 -6.02130473e-01
-1.97410509e-02 5.76333165e-01 3.54469091e-01 -6.65923953e-01
-4.61335033e-01 -9.74447846e-01 -6.71270549e-01 7.14744508e-01
3.70234549e-01 -1.30826935e-01 -1.29930937e+00 1.44000304e+00
-5.97792193e-02 4.25337821e-01 8.93169791e-02 1.00999689e+00
6.62027836e-01 3.77318531e-01 2.73662776e-01 1.83500558e-01
1.24868941e+00 -5.01057923e-01 -4.98219758e-01 -4.16547507e-01
5.85211575e-01 -3.74941826e-01 1.20228362e+00 1.21710710e-01
-5.77261329e-01 -1.63013771e-01 -1.18620145e+00 2.16236129e-01
-1.01691341e+00 6.29709139e-02 1.00911283e+00 7.01934516e-01
-1.16860700e+00 7.39059091e-01 -1.06814575e+00 -3.50915462e-01
6.19257629e-01 3.28949332e-01 5.92901604e-03 3.82642820e-02
-1.01890838e+00 9.63576019e-01 1.66357651e-01 5.40916920e-01
-1.15091240e+00 -4.57910091e-01 -4.32550579e-01 3.89864802e-01
1.91239920e-03 -5.42789638e-01 8.40567529e-01 -1.41829371e+00
-1.02312684e+00 1.30194092e+00 -1.94249630e-01 -4.22341228e-01
1.46970823e-01 4.98582162e-02 -4.40899760e-01 -7.31014460e-02
-2.37387806e-01 6.64352775e-01 7.60657251e-01 -1.44120657e+00
-5.17662644e-01 -3.98676932e-01 6.52652010e-02 -5.36218733e-02
2.43557878e-02 1.91262681e-02 6.04408048e-03 -5.84241509e-01
3.55570525e-01 -1.04176688e+00 -3.97883177e-01 1.15628600e-01
-1.04742564e-01 4.52402711e-01 7.39075303e-01 -2.58698761e-01
1.21376622e+00 -1.99174702e+00 -3.26492935e-01 3.78300875e-01
2.46031687e-01 4.88435686e-01 -2.86952466e-01 3.00319850e-01
-2.65894711e-01 7.15919912e-01 -5.88156760e-01 -1.05949789e-01
-3.44532400e-01 2.28592291e-01 -5.92645586e-01 7.36612201e-01
6.21880591e-01 8.97595406e-01 -9.33326006e-01 -2.23270636e-02
1.44625276e-01 6.83620214e-01 -2.75175154e-01 -9.41858888e-02
-5.49429178e-01 7.11889148e-01 -5.83380647e-02 8.82951677e-01
8.97121429e-01 -3.96158278e-01 4.03775871e-01 2.53031224e-01
-8.07929635e-01 5.51727593e-01 -8.49909067e-01 1.58779585e+00
-2.93956369e-01 9.83384430e-01 2.90377676e-01 -5.88297486e-01
6.74894273e-01 1.96120754e-01 7.25549832e-02 -5.41763604e-01
2.54534706e-02 3.59861732e-01 2.75458068e-01 -3.13377351e-01
3.84503931e-01 -8.06177631e-02 4.12877172e-01 1.41498640e-01
-2.88622618e-01 -2.89756089e-01 -2.16126800e-01 -2.17874944e-01
8.13423336e-01 1.74949616e-01 -2.54165173e-01 -7.30329216e-01
1.09514594e-01 3.09516430e-01 2.16461524e-01 9.62439597e-01
-8.35003853e-02 7.76863217e-01 3.60110581e-01 -5.52226543e-01
-7.09042728e-01 -6.81548953e-01 -4.07336384e-01 9.66978967e-01
-2.24721432e-01 -2.53204226e-01 -4.48976576e-01 -1.79338336e-01
3.14527489e-02 5.88338196e-01 -7.66524732e-01 1.48757949e-01
-3.18521470e-01 -1.11200035e+00 9.58194256e-01 2.10751444e-01
3.26774806e-01 -8.85251820e-01 -9.70949769e-01 1.20696850e-01
3.81233804e-02 -8.81977737e-01 3.26493382e-01 6.73460126e-01
-1.06544471e+00 -6.90290987e-01 -3.86227310e-01 -3.68480772e-01
6.52318776e-01 3.91627878e-01 1.31785607e+00 6.17734373e-01
-3.10773015e-01 -2.70424515e-01 -3.03065404e-02 -4.66498911e-01
8.71733353e-02 3.74840856e-01 -3.57764274e-01 -1.01983041e-01
5.58528721e-01 -8.11820328e-01 -7.24479795e-01 -1.54426351e-01
-8.24565530e-01 -1.01190597e-01 3.77914935e-01 6.73742890e-01
6.06973886e-01 1.11736663e-01 6.71822205e-02 -8.41295183e-01
8.32431465e-02 -4.64811891e-01 -1.18035126e+00 -7.81176938e-03
-7.41006017e-01 -7.40686283e-02 1.50430650e-01 -8.53130892e-02
-5.31032860e-01 4.30784464e-01 1.46285235e-03 -5.06442308e-01
-2.94674098e-01 9.37742949e-01 5.58022177e-03 -3.07503819e-01
7.13055789e-01 3.04455876e-01 -3.23992670e-01 -2.23849520e-01
-6.54960983e-03 5.01723349e-01 4.17840332e-01 -2.23265976e-01
7.39701509e-01 9.76085424e-01 -1.32560134e-01 -9.14773464e-01
-6.95121527e-01 -3.52324963e-01 -6.32759273e-01 -2.84037471e-01
8.78498971e-01 -1.33045876e+00 -7.02156425e-01 7.13805854e-01
-1.00420892e+00 -7.49353111e-01 -3.08754854e-02 1.90582871e-01
2.77558148e-01 -1.01954915e-01 -2.31502846e-01 -1.11673939e+00
-2.46832862e-01 -8.57832730e-01 9.93665636e-01 8.11801702e-02
4.65738513e-02 -7.15827882e-01 3.26836035e-02 -2.68336236e-01
8.70532691e-01 4.59855080e-01 5.38718939e-01 -1.70094013e-01
-1.02852905e+00 -3.41019481e-02 -4.51334625e-01 1.89539101e-02
1.63438812e-01 4.72818047e-01 -1.91265428e+00 -3.80950570e-01
-2.43040070e-01 -3.18896957e-02 1.60635448e+00 4.37920898e-01
1.26418591e+00 -9.03518498e-02 -3.94596696e-01 9.80981946e-01
1.83728230e+00 -2.73461133e-01 9.72722530e-01 3.51153463e-01
6.57857537e-01 5.38529634e-01 1.33707926e-01 1.94097802e-01
6.81161508e-02 2.12380230e-01 9.84161794e-01 -6.14381671e-01
3.04308027e-01 1.24882981e-01 1.18408360e-01 7.12806582e-02
-2.22750500e-01 -3.50678712e-01 -1.15654457e+00 9.28278089e-01
-1.59904087e+00 -1.15424442e+00 -4.63970155e-01 2.27616620e+00
6.79227948e-01 3.12135607e-01 -4.37850595e-01 -5.26816957e-02
3.88873607e-01 3.89878064e-01 -5.15502214e-01 -2.06727341e-01
-6.76347971e-01 5.02714574e-01 1.06896675e+00 6.35175169e-01
-1.36488676e+00 1.32840538e+00 7.10361719e+00 1.89449176e-01
-1.97346425e+00 7.38291368e-02 5.41468143e-01 -3.19131970e-01
-5.34521699e-01 3.52033943e-01 -6.47142470e-01 2.80139595e-01
1.06478846e+00 7.13863075e-01 5.17716944e-01 2.46870637e-01
4.83537465e-01 -3.60629022e-01 -7.08758891e-01 6.72573566e-01
-6.53090060e-01 -1.60079122e+00 -1.58210352e-01 9.80817825e-02
6.88796341e-01 8.98951232e-01 1.39772832e-01 5.58787435e-02
4.25686896e-01 -1.56569302e+00 6.07365966e-01 5.88291466e-01
1.08728004e+00 -1.67114362e-01 4.50978488e-01 7.92534426e-02
-1.20158684e+00 8.44268948e-02 -2.92334080e-01 -7.14320540e-01
-3.41984868e-01 7.76234329e-01 -5.27711928e-01 1.35540217e-01
9.03897226e-01 5.99161625e-01 -8.35257053e-01 9.38810766e-01
-1.35939177e-02 8.85130167e-01 -6.44362688e-01 3.02996606e-01
4.15384531e-01 -1.36790007e-01 2.75775552e-01 1.55248606e+00
1.97762325e-01 6.87372014e-02 1.19534479e-02 1.15605211e+00
2.21281543e-01 -3.02539647e-01 -7.35444784e-01 -4.43391442e-01
4.49449241e-01 1.24897623e+00 -9.81304169e-01 -4.20835316e-01
-3.60461950e-01 7.76297092e-01 8.30404833e-02 4.84278440e-01
-6.25953555e-01 -1.19702011e-01 1.08415484e+00 2.28656903e-01
4.85331267e-01 -4.03289109e-01 -8.24381053e-01 -1.33708882e+00
-1.92525417e-01 -7.67281651e-01 4.92117181e-02 -8.95699263e-01
-8.34435344e-01 6.63797438e-01 -5.01612425e-01 -1.21436894e+00
2.63298631e-01 -5.61505675e-01 -7.24613011e-01 1.40957201e+00
-2.11928701e+00 -1.25862479e+00 -7.20236301e-01 3.46898407e-01
-1.71540096e-01 3.24957073e-01 1.12580132e+00 3.15494150e-01
-4.45023179e-01 6.67811856e-02 2.49999270e-01 1.46742119e-02
3.66701126e-01 -1.16878486e+00 6.68279171e-01 1.12267625e+00
-2.64159311e-02 7.71419406e-01 8.18907499e-01 -5.56667984e-01
-1.18740666e+00 -1.43299472e+00 8.85115981e-01 -4.56028163e-01
3.82627845e-01 -6.86212420e-01 -1.18920469e+00 8.01458299e-01
3.66249591e-01 4.49823081e-01 5.05004346e-01 4.75788526e-02
-4.82447296e-01 -1.20087102e-01 -8.85865927e-01 3.87225449e-01
9.04898524e-01 -8.66713822e-01 3.98558751e-02 4.27865505e-01
3.67303699e-01 -2.66484976e-01 -4.35890168e-01 4.32661355e-01
5.31191885e-01 -9.86588657e-01 5.96967280e-01 -7.36800551e-01
1.28341570e-01 -8.70343268e-01 -2.36704439e-01 -1.19456398e+00
-3.94666135e-01 -2.13873714e-01 3.93596053e-01 9.39462423e-01
6.25627220e-01 -7.63301909e-01 5.50045788e-01 2.63127774e-01
-2.67304033e-01 -2.57157981e-01 -7.78921843e-01 -6.56439006e-01
3.80419195e-01 -2.05552861e-01 6.57082260e-01 1.38774192e+00
-5.52023470e-01 1.74293257e-02 8.43331069e-02 8.12672317e-01
1.16251044e-01 4.02789801e-01 4.94455606e-01 -1.44033539e+00
3.17923678e-03 -5.77188790e-01 3.21100801e-02 -5.54238200e-01
-8.41166079e-02 -9.98383820e-01 1.05059683e-01 -1.39249468e+00
-8.27734545e-02 -6.41634583e-01 -4.03711438e-01 7.86908269e-01
-8.77079591e-02 5.07420659e-01 -4.88490090e-02 3.82220805e-01
2.70058271e-02 9.55756530e-02 8.82556856e-01 -2.96670288e-01
-2.17217714e-01 -2.39027753e-01 -4.70968813e-01 4.47904140e-01
9.38639343e-01 -6.89492345e-01 -2.09883898e-02 -1.03542912e+00
6.13417447e-01 -4.82615620e-01 8.73036206e-01 -1.07424641e+00
-8.16327855e-02 -2.66150177e-01 5.26147664e-01 -6.36185884e-01
2.57314116e-01 -8.45664501e-01 3.69192421e-01 5.96101582e-01
-3.52650061e-02 8.60779360e-02 6.81416512e-01 9.36526582e-02
-1.83932021e-01 1.93905920e-01 9.44083393e-01 -5.70556104e-01
-6.50695860e-01 1.00904964e-01 -9.96952236e-01 -3.90357792e-01
4.12101716e-01 -2.50243783e-01 -5.57481527e-01 6.64355680e-02
-8.56550455e-01 2.73178726e-01 7.89748788e-01 3.55512276e-02
2.58467168e-01 -5.82701027e-01 -8.39254379e-01 4.78160024e-01
2.29187936e-01 -1.23183373e-02 -1.52028883e-02 8.75959396e-01
-7.08891630e-01 4.19301510e-01 -2.10938275e-01 -9.74214733e-01
-1.03626323e+00 3.66683453e-01 1.05492306e+00 1.42329186e-02
-3.75786453e-01 8.73410225e-01 -1.33166211e-02 -5.44559658e-01
-1.46157295e-01 -7.49392092e-01 -9.37564522e-02 3.81786704e-01
1.57837167e-01 -1.85281098e-01 4.92615253e-01 -3.25288445e-01
-5.70910037e-01 2.26953521e-01 3.60200644e-01 -2.24745303e-01
1.35042584e+00 -1.21762104e-01 -2.96847552e-01 6.90231681e-01
6.84717417e-01 2.95658852e-03 -1.18015409e+00 4.37455103e-02
4.78729382e-02 -4.61784929e-01 3.79988790e-01 -1.08298087e+00
-1.33749378e+00 1.17290378e+00 1.21844459e+00 2.10369721e-01
1.05230653e+00 -4.30462480e-01 -3.31201293e-02 4.66578066e-01
-1.38503006e-02 -7.70636320e-01 -7.23112345e-01 7.06076503e-01
7.06536472e-01 -1.24468052e+00 -2.85196081e-02 -1.64440870e-01
-2.16356248e-01 8.56014609e-01 6.11616731e-01 -2.26284638e-01
8.15937698e-01 4.83292580e-01 5.28590083e-01 -6.12966001e-01
-9.28736448e-01 -6.91889286e-01 -2.01647714e-01 3.42186004e-01
7.31819212e-01 3.27520549e-01 3.11289425e-03 7.89554641e-02
3.76896597e-02 1.21436484e-01 5.43817878e-01 1.11710322e+00
-4.81800914e-01 -6.23011053e-01 -3.35200101e-01 5.64147949e-01
-4.37646240e-01 -8.13509285e-01 -6.43283546e-01 5.79736829e-01
2.42727086e-01 7.76539981e-01 5.04862547e-01 -2.99970299e-01
-1.27190083e-01 2.07556054e-01 2.96211660e-01 -8.25020492e-01
-1.16934657e+00 2.45637417e-01 7.97135085e-02 -2.86079735e-01
-5.74078321e-01 -8.34451675e-01 -1.10530674e+00 -2.20953152e-01
-6.56571686e-01 -1.69148833e-01 9.08231735e-01 7.67849386e-01
6.09203279e-01 7.04298019e-01 2.28228986e-01 -9.33512509e-01
-7.75808543e-02 -1.05280364e+00 -5.94848514e-01 -1.25234470e-01
8.11919570e-01 -4.38442886e-01 -4.28447008e-01 -4.33760285e-02] | [9.631141662597656, -1.6334415674209595] |
3bd07adf-4c90-4990-8ac2-0da1059026aa | counterfactual-explanation-with-missing | 2304.14606 | null | https://arxiv.org/abs/2304.14606v1 | https://arxiv.org/pdf/2304.14606v1.pdf | Counterfactual Explanation with Missing Values | Counterfactual Explanation (CE) is a post-hoc explanation method that provides a perturbation for altering the prediction result of a classifier. Users can interpret the perturbation as an "action" to obtain their desired decision results. Existing CE methods require complete information on the features of an input instance. However, we often encounter missing values in a given instance, and the previous methods do not work in such a practical situation. In this paper, we first empirically and theoretically show the risk that missing value imputation methods affect the validity of an action, as well as the features that the action suggests changing. Then, we propose a new framework of CE, named Counterfactual Explanation by Pairs of Imputation and Action (CEPIA), that enables users to obtain valid actions even with missing values and clarifies how actions are affected by imputation of the missing values. Specifically, our CEPIA provides a representative set of pairs of an imputation candidate for a given incomplete instance and its optimal action. We formulate the problem of finding such a set as a submodular maximization problem, which can be solved by a simple greedy algorithm with an approximation guarantee. Experimental results demonstrated the efficacy of our CEPIA in comparison with the baselines in the presence of missing values. | ['Yuichi Ike', 'Ken Kobayashi', 'Takuya Takagi', 'Kentaro Kanamori'] | 2023-04-28 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 4.33195978e-01 5.05570412e-01 -5.51316082e-01 -4.91873503e-01
-8.44609022e-01 -4.29261416e-01 3.80710453e-01 -2.67071091e-02
-9.24287811e-02 1.22475350e+00 4.29704905e-01 -3.10080141e-01
-3.86141479e-01 -6.82195544e-01 -1.21325111e+00 -8.09572041e-01
2.28910401e-01 3.88979435e-01 -6.23375118e-01 2.23159477e-01
4.44741368e-01 -5.67149231e-03 -1.45787835e+00 6.06472135e-01
1.21370780e+00 6.29035294e-01 -2.09743574e-01 8.65101162e-03
8.42280686e-02 6.56221628e-01 -5.07817626e-01 -5.20492971e-01
4.83709425e-01 -4.93355691e-01 -8.20839882e-01 3.29630733e-01
8.13767090e-02 -5.79602361e-01 -3.13011333e-02 1.06445360e+00
3.07116896e-01 1.05249777e-01 5.66781640e-01 -1.85223389e+00
-6.06056035e-01 1.02302718e+00 -6.89342082e-01 -3.39834690e-01
4.69231039e-01 3.32625687e-01 8.77248704e-01 -8.06492150e-01
5.50409913e-01 1.25766909e+00 4.77942973e-01 6.40395463e-01
-1.43937123e+00 -8.05740058e-01 4.99809504e-01 2.99963146e-01
-1.04625142e+00 -4.67181385e-01 8.17399859e-01 -2.25149602e-01
4.73512203e-01 6.20341480e-01 3.39804798e-01 1.10042930e+00
9.73641276e-02 9.11589622e-01 1.15583217e+00 -2.21429393e-01
5.80905557e-01 1.82341397e-01 1.07467473e-01 5.80080040e-02
4.45915371e-01 3.10389906e-01 -6.96256697e-01 -7.15473235e-01
2.00109541e-01 2.74879009e-01 -5.44256151e-01 -4.44084257e-01
-1.25703466e+00 7.83286333e-01 2.45440066e-01 -3.78403813e-01
-8.06041539e-01 1.03382774e-01 1.13155320e-01 3.55702668e-01
2.58227378e-01 6.65645599e-01 -8.05206060e-01 7.94487521e-02
-6.15307808e-01 6.59931302e-01 6.49160385e-01 7.40498960e-01
5.77939928e-01 -2.88545668e-01 -6.69072747e-01 3.82085353e-01
-3.82482894e-02 3.39581907e-01 2.53322214e-01 -1.32953835e+00
6.51147783e-01 6.80138052e-01 8.37790906e-01 -7.77934134e-01
-3.30150016e-02 -3.18341225e-01 -7.26174176e-01 1.20634258e-01
5.66240072e-01 -4.47611928e-01 -4.25137013e-01 2.14020300e+00
5.68421006e-01 5.24871349e-01 2.13174626e-01 8.91460598e-01
2.55947322e-01 4.22794938e-01 -1.66570857e-01 -8.10865521e-01
9.69676554e-01 -5.73510051e-01 -8.44389856e-01 -2.19721675e-01
7.19324231e-01 -3.54161233e-01 8.95133615e-01 4.00532901e-01
-9.16307449e-01 -7.86879193e-03 -6.57389462e-01 4.53042984e-01
1.47260189e-01 -1.30680382e-01 7.65778780e-01 4.95829940e-01
-4.44546610e-01 7.19299436e-01 -4.94148076e-01 2.64790773e-01
5.57681561e-01 5.30455291e-01 -3.39920312e-01 -2.01938555e-01
-1.10096395e+00 6.48841321e-01 2.69427896e-01 1.60994902e-01
-9.14482832e-01 -1.26732004e+00 -7.54421115e-01 3.13186646e-01
8.74561310e-01 -9.64093685e-01 1.28536415e+00 -1.13373756e+00
-1.04601502e+00 3.95784259e-01 -6.42293990e-01 -6.77463651e-01
9.65816736e-01 -1.29407272e-01 -1.17024086e-01 -5.20029843e-01
3.88979822e-01 4.37465608e-01 7.62864351e-01 -1.37819004e+00
-6.71938837e-01 -6.70664072e-01 2.38420516e-01 1.88886926e-01
8.30827877e-02 -3.69676530e-01 -3.48609523e-03 -4.88128603e-01
1.36014566e-01 -7.89617360e-01 -5.78492045e-01 -1.73034847e-01
-9.58595812e-01 -1.67607233e-01 5.78917682e-01 -4.34333831e-01
1.17866039e+00 -2.04348230e+00 -1.56679392e-01 3.67388010e-01
1.81735232e-02 -2.48052686e-01 -5.39581664e-02 3.24930817e-01
-2.96945840e-01 3.49362642e-01 -5.80620348e-01 -2.29498178e-01
9.11884531e-02 1.27321571e-01 -7.62706101e-01 5.40533781e-01
-1.25915617e-01 6.43840849e-01 -7.45508015e-01 1.68868192e-02
-6.18436523e-02 -2.82085873e-02 -8.04855287e-01 1.99250728e-01
-3.32330912e-01 5.92074990e-01 -3.76263857e-01 5.05626857e-01
1.07730651e+00 -4.95649539e-02 5.01446426e-01 1.65460020e-01
-1.24441953e-02 2.57984251e-01 -1.47803092e+00 1.15724492e+00
-2.20253170e-01 1.10228449e-01 -1.90904185e-01 -1.11067748e+00
6.50068939e-01 2.69376934e-01 5.25649011e-01 -2.59709626e-01
-9.47741419e-02 1.30694568e-01 -5.33808693e-02 -4.72743750e-01
5.84556088e-02 -2.03554660e-01 -1.12921752e-01 6.19739711e-01
-7.81741738e-01 4.21568543e-01 -1.62141725e-01 3.84724955e-03
1.06274021e+00 -1.31098092e-01 5.74591875e-01 1.02154724e-02
3.26445073e-01 2.17922055e-03 1.16647053e+00 1.21071887e+00
-6.95806695e-03 5.35797954e-01 7.85742044e-01 -4.94453639e-01
-6.35854185e-01 -8.23820531e-01 -1.75131634e-01 4.76562381e-01
2.13136107e-01 -1.63518876e-01 -5.60435116e-01 -1.06035244e+00
4.94908899e-01 1.34305358e+00 -8.99514258e-01 -3.21680456e-01
-2.65404522e-01 -9.69620466e-01 -5.85687533e-02 3.57421309e-01
6.14437819e-01 -1.10981452e+00 -3.94790173e-01 1.63217992e-01
-5.35223424e-01 -7.09325194e-01 -6.27329528e-01 -1.02008440e-01
-9.88214552e-01 -1.37060320e+00 -1.00934297e-01 -1.45720422e-01
9.82435882e-01 2.88152367e-01 9.01676774e-01 1.15624584e-01
2.57102847e-01 -3.43330503e-02 -6.40818179e-02 -5.72261095e-01
-2.51679152e-01 -5.07330358e-01 1.90760165e-01 3.90104115e-01
4.42055821e-01 -5.96613884e-01 -7.10386693e-01 2.41881102e-01
-7.23632216e-01 1.09168164e-01 5.69778144e-01 9.56385255e-01
7.45397031e-01 1.84636861e-01 9.48058844e-01 -1.35263860e+00
6.95515215e-01 -7.43834198e-01 -5.48473120e-01 4.26793963e-01
-8.98295045e-01 3.38937044e-01 7.45913148e-01 -3.17436010e-01
-1.16708994e+00 2.54141212e-01 4.15223360e-01 -2.75273800e-01
-2.31160149e-01 6.68540001e-01 -6.88863993e-01 5.07549763e-01
5.10119319e-01 2.65438408e-01 -8.86756228e-04 -4.89343435e-01
2.89321989e-01 6.65769994e-01 4.70905095e-01 -6.19959593e-01
6.32220089e-01 5.86601615e-01 7.01008812e-02 -1.14006273e-01
-1.07909465e+00 -2.23993242e-01 -3.00139099e-01 3.77287380e-02
2.48763144e-01 -7.05940545e-01 -1.12323487e+00 1.55934617e-01
-1.17616296e+00 6.41586781e-02 -4.29203153e-01 3.74033809e-01
-6.74600840e-01 2.61936635e-01 2.33901516e-01 -9.61047173e-01
-4.59753275e-02 -1.06443226e+00 7.04800129e-01 -6.01251684e-02
-2.56749690e-01 -5.40089190e-01 -2.39111066e-01 5.57474732e-01
-1.91858530e-01 6.21452034e-01 9.56943333e-01 -6.24092937e-01
-7.83519208e-01 -2.68120646e-01 1.36910439e-01 -6.70868754e-02
1.31188452e-01 -3.69200408e-01 -8.06630135e-01 -2.40383640e-01
-8.12804550e-02 1.28751159e-01 7.34995902e-01 7.93123305e-01
1.77772439e+00 -1.11681604e+00 -4.46712166e-01 5.46339810e-01
1.18069005e+00 3.09826463e-01 7.19566762e-01 4.64066535e-01
2.35639736e-01 5.81414402e-01 8.63959074e-01 9.36223805e-01
2.82417387e-01 7.94609129e-01 7.83537984e-01 8.30779821e-02
5.24736404e-01 -4.61743653e-01 3.50705147e-01 -3.71665210e-01
9.13359448e-02 -3.73919606e-02 -4.44925785e-01 7.27309287e-01
-2.34155416e+00 -1.18083167e+00 -1.99288532e-01 2.79307628e+00
9.02379811e-01 -2.14248523e-01 4.86611947e-02 1.47620752e-01
8.49450946e-01 -3.32098365e-01 -1.13847983e+00 -3.72037441e-01
3.43950801e-02 -3.46664935e-01 4.88773674e-01 3.60476881e-01
-8.37287068e-01 3.09353381e-01 6.11527157e+00 3.53585184e-01
-5.29663920e-01 -9.10743773e-02 8.07227850e-01 -3.59243125e-01
-8.28515410e-01 2.69927502e-01 -5.77858210e-01 7.47402191e-01
6.26791358e-01 -6.07132852e-01 6.46021962e-01 8.70684862e-01
7.33846605e-01 -6.24982752e-02 -1.52528441e+00 7.61901617e-01
-2.70236701e-01 -1.52439785e+00 3.18835646e-01 2.76274145e-01
1.02226090e+00 -5.80300868e-01 1.30047172e-01 2.15734392e-01
4.51957434e-01 -9.07347679e-01 5.72668254e-01 5.32234311e-01
5.28728545e-01 -1.01186287e+00 7.57478893e-01 7.32333958e-01
-3.34754050e-01 -5.57243645e-01 -3.25415075e-01 -4.23879981e-01
-1.16056442e-01 7.91712642e-01 -8.64401639e-01 5.99168658e-01
3.82909983e-01 5.30280888e-01 -9.07300562e-02 1.10459232e+00
-5.24082303e-01 7.83527792e-01 -6.99932426e-02 2.74269700e-01
-1.14857361e-01 -1.76300421e-01 7.58167565e-01 4.84455824e-01
4.31554645e-01 3.28281701e-01 9.79277268e-02 1.02710664e+00
-3.80591154e-01 -8.03850517e-02 -7.09261656e-01 3.74303043e-01
8.55658531e-01 5.92794895e-01 -9.54013839e-02 -1.87717244e-01
-2.05219463e-01 7.76961625e-01 1.83865443e-01 4.98799592e-01
-8.78788829e-01 5.01675457e-02 1.04506159e+00 1.54242575e-01
4.55580018e-02 7.20057130e-01 -9.22394633e-01 -1.31731713e+00
3.31367701e-01 -1.20516658e+00 7.41821527e-01 -7.54857957e-01
-1.22366059e+00 -2.10997313e-01 -4.23888676e-02 -1.31121433e+00
-2.98570961e-01 5.93084916e-02 -8.40735674e-01 7.89678514e-01
-1.19027710e+00 -6.31204247e-01 -1.12266457e-02 4.43748176e-01
4.23159361e-01 3.93882990e-02 7.40070045e-01 -3.54074270e-01
-6.91753566e-01 8.47055435e-01 3.86761755e-01 -2.35733002e-01
5.53673327e-01 -1.24504042e+00 -1.54546514e-01 1.03722906e+00
-1.22294120e-01 7.19731450e-01 1.03064573e+00 -6.96128666e-01
-1.24760973e+00 -1.17064118e+00 1.17944086e+00 -4.18309122e-01
9.79266167e-02 -4.86784382e-03 -9.13060904e-01 9.15275097e-01
-1.49592161e-01 -2.99060255e-01 7.39092350e-01 4.21620905e-01
-1.31606936e-01 -2.49405697e-01 -1.57993329e+00 8.84914041e-01
1.09496295e+00 5.11045158e-02 -7.15943694e-01 5.30710340e-01
6.47045970e-01 -3.20762604e-01 -3.78437817e-01 3.77003551e-01
4.29601729e-01 -9.63274360e-01 9.79697347e-01 -1.23695958e+00
7.83075273e-01 -3.73032182e-01 -1.03269823e-01 -1.48779249e+00
-1.89769506e-01 -7.24776685e-01 -1.73595622e-01 1.06976569e+00
6.52778506e-01 -7.95024931e-01 7.89618015e-01 1.37028027e+00
7.66009986e-02 -8.43942821e-01 -1.09321046e+00 -7.07023442e-01
-9.48672146e-02 -6.08693659e-01 1.43867660e+00 1.13549268e+00
2.18208835e-01 -1.70192853e-01 -8.01565647e-01 6.20455682e-01
9.14764047e-01 7.83702314e-01 9.95537996e-01 -1.08533835e+00
-3.27113986e-01 2.89461762e-02 -5.33986948e-02 -7.09433854e-01
4.25854474e-01 -8.80070567e-01 -1.72338501e-01 -1.33243322e+00
7.71161199e-01 -1.63415074e-01 -2.72762597e-01 7.81025827e-01
-5.76088428e-01 -4.16215777e-01 1.73962474e-01 2.38228038e-01
-3.12484920e-01 5.78070462e-01 1.16279578e+00 -1.76043674e-01
-5.07345974e-01 5.38671553e-01 -1.41754949e+00 7.74214268e-01
6.97246015e-01 -8.86209369e-01 -3.15963805e-01 -1.56087190e-01
1.99580699e-01 4.85110074e-01 6.73671007e-01 -2.37041175e-01
6.25777766e-02 -7.02839553e-01 3.67060661e-01 -5.38648546e-01
-6.54717684e-02 -9.69192326e-01 4.34848219e-01 5.42674661e-01
-7.69074976e-01 -2.84090340e-01 -1.01956934e-01 8.27284873e-01
3.48661728e-02 -2.82715887e-01 3.61042321e-01 1.09135702e-01
-4.22833979e-01 2.93479145e-01 -1.60795018e-01 -8.75849947e-02
1.15623796e+00 -1.65685013e-01 -3.79666090e-01 -6.32311344e-01
-8.79026234e-01 6.66026711e-01 3.38255346e-01 2.58012652e-01
6.65034771e-01 -1.44342160e+00 -8.93737555e-01 1.77658051e-01
5.76694533e-02 -2.92085751e-04 2.82077104e-01 8.39430332e-01
4.83991802e-01 2.38014877e-01 1.43509088e-02 -1.34751618e-01
-1.17286646e+00 7.16484785e-01 3.36275041e-01 -2.18313515e-01
-4.68982428e-01 3.27940643e-01 3.51380616e-01 -5.72433174e-01
1.34816617e-01 -9.01262239e-02 -5.12461029e-02 -1.74937502e-01
6.01938903e-01 5.10080159e-01 -8.35000649e-02 -1.31692262e-02
-3.80217373e-01 -2.26459578e-01 -1.41713312e-02 -3.61648947e-02
1.47373259e+00 -1.56035945e-01 -1.88887686e-01 1.48694873e-01
7.23085701e-01 -3.41810547e-02 -1.45573509e+00 -1.73858270e-01
-1.00188628e-01 -1.09371614e+00 -1.07688665e-01 -1.20777917e+00
-9.19596970e-01 2.18931854e-01 8.12487751e-02 -8.86330977e-02
1.26056075e+00 -2.07971200e-01 5.05284429e-01 3.89442146e-01
2.79779822e-01 -8.59614730e-01 -5.46282411e-01 -9.95061919e-02
1.17490244e+00 -1.29569519e+00 -6.64587542e-02 -5.43796599e-01
-7.73994505e-01 7.01101363e-01 6.54676676e-01 2.35552683e-01
4.32274699e-01 -2.89646666e-02 -1.98852539e-01 9.06000286e-02
-1.06257713e+00 1.51627347e-01 1.55381039e-01 6.15851343e-01
9.89127383e-02 4.11615908e-01 -6.33304417e-01 1.15143919e+00
-2.18332395e-01 2.89247811e-01 8.04377556e-01 6.55651927e-01
1.07991165e-02 -1.01975358e+00 -7.14758456e-01 8.12265873e-01
-4.32811618e-01 8.67680386e-02 -6.01324737e-01 5.49724758e-01
1.23927310e-01 1.22457600e+00 -1.23101652e-01 -2.41870478e-01
5.49614787e-01 1.28487527e-01 6.04014099e-02 -4.96102959e-01
-3.82512718e-01 -3.11546415e-01 8.32982082e-03 -7.26560712e-01
-1.62352026e-01 -1.09219825e+00 -1.16359627e+00 -5.07256627e-01
-2.55998105e-01 3.20737362e-01 2.16868356e-01 1.31723189e+00
5.86666346e-01 2.78345227e-01 1.11275911e+00 -2.99469978e-01
-1.19363081e+00 -6.46749020e-01 -6.09764576e-01 6.48312628e-01
3.16833317e-01 -6.19686007e-01 -6.59361660e-01 -5.44888526e-02] | [8.666192054748535, 5.58976411819458] |
8bfcb154-a056-45d3-9436-0ac1f34c8ed3 | image-pre-compensation-balancing-contrast-and | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Ji_Image_Pre-compensation_Balancing_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Ji_Image_Pre-compensation_Balancing_2014_CVPR_paper.pdf | Image Pre-compensation: Balancing Contrast and Ringing | The goal of image pre-compensation is to process an image such that after being convolved with a known kernel, will appear close to the sharp reference image. In a practical setting, the pre-compensated image has significantly higher dynamic range than the latent image. As a result, some form of tone mapping is needed. In this paper, we show how global tone mapping functions affect contrast and ringing in image pre-compensation. In particular, we show that linear tone mapping eliminates ringing but incurs severe contrast loss, while non-linear tone mapping functions such as Gamma curves slightly enhances contrast but introduces ringing. To enable quantitative analysis, we design new metrics to measure the contrast of an image with ringing. Specifically, we set out to find its "equivalent ringing-free" image that matches its intensity histogram and uses its contrast as the measure. We illustrate our approach on projector defocus compensation and visual acuity enhancement. Compared with the state-of-the-art, our approach significantly improves the contrast. We believe our technique is the first to analytically trade-off between contrast and ringing. | ['Jinwei Ye', 'Yu Ji', 'Sing Bing Kang', 'Jingyi Yu'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['tone-mapping'] | ['computer-vision'] | [ 6.01149559e-01 -1.35230169e-01 1.35666579e-01 -1.97893187e-01
-4.40525413e-01 -5.41896343e-01 3.15007716e-01 7.75851160e-02
-4.89193052e-01 6.36263371e-01 2.93289185e-01 -2.35732660e-01
-7.08203614e-02 -6.75926149e-01 -6.65796041e-01 -7.40930736e-01
2.28151619e-01 -3.57454926e-01 5.08243203e-01 -7.65597522e-02
5.62702119e-01 3.83788705e-01 -1.52668452e+00 7.61533678e-02
1.15586150e+00 8.38407099e-01 4.52262819e-01 7.30502784e-01
3.05203795e-01 7.55752385e-01 -6.43664896e-01 -1.30023926e-01
4.25991118e-01 -6.72816038e-01 -6.09084249e-01 8.03037435e-02
9.52978492e-01 -4.63697195e-01 -3.55119556e-01 1.45578027e+00
5.31821907e-01 4.72820513e-02 5.40479124e-01 -6.45072043e-01
-4.89304334e-01 2.58473039e-01 -7.39402592e-01 2.46644214e-01
1.13431998e-01 2.51280874e-01 3.95002186e-01 -5.88668525e-01
5.27259707e-01 8.33436370e-01 4.90745246e-01 3.70294631e-01
-1.55788863e+00 -3.44468534e-01 -3.80792707e-01 9.10352822e-03
-1.15349460e+00 -4.23716664e-01 7.06394434e-01 -4.04401839e-01
5.92331052e-01 3.71861100e-01 5.59087813e-01 2.22531229e-01
7.21417367e-01 1.88692987e-01 1.69384432e+00 -7.97188699e-01
9.97778326e-02 1.18588299e-01 -7.14231506e-02 5.16543210e-01
2.34866694e-01 4.30300474e-01 -2.39035413e-01 1.69910461e-01
1.06333292e+00 -1.77992120e-01 -1.03977180e+00 -5.03894567e-01
-1.10477912e+00 2.47479036e-01 5.44321120e-01 3.92939746e-01
-1.82645202e-01 1.99311286e-01 -2.35257506e-01 2.09991843e-01
1.46765500e-01 9.43758547e-01 9.57307443e-02 -3.68253067e-02
-1.08741450e+00 1.55399621e-01 3.05507630e-01 4.25659031e-01
7.17796683e-01 -2.44801819e-01 -3.97447288e-01 8.31105888e-01
-1.77858979e-01 4.76619005e-01 4.28062379e-01 -1.20406413e+00
1.40127778e-01 2.47022882e-01 3.59387100e-01 -5.64348936e-01
-2.44036853e-01 -6.84712648e-01 -5.30419886e-01 9.38897133e-01
6.87782466e-01 -3.74691710e-02 -1.00107098e+00 1.66358232e+00
-1.99368358e-01 1.43819004e-02 7.83664989e-04 1.13340878e+00
2.47879103e-01 4.64661628e-01 -4.52906996e-01 -4.94168788e-01
1.24520898e+00 -7.86424279e-01 -9.06862080e-01 -1.15387663e-01
3.72292921e-02 -1.26365244e+00 1.34237778e+00 5.02029359e-01
-1.55595732e+00 -4.95532811e-01 -1.22848916e+00 -2.32049879e-02
2.53625453e-01 -8.99308771e-02 1.42461300e-01 7.13707149e-01
-1.25908291e+00 6.51239991e-01 -3.52454334e-01 -2.22952515e-01
-6.45968691e-02 9.65740085e-02 -6.16785809e-02 -1.18861899e-01
-7.93273449e-01 1.14968443e+00 -1.18919471e-02 -4.99204099e-01
-4.12912011e-01 -1.06521606e+00 -5.95108986e-01 8.68046358e-02
2.09860310e-01 -7.17063427e-01 1.35043633e+00 -8.93432736e-01
-1.62999845e+00 8.90540481e-01 -2.31306806e-01 -5.19034743e-01
5.25588512e-01 -1.17222808e-01 -2.71290720e-01 3.15613955e-01
-1.75221905e-01 5.10000229e-01 1.12599695e+00 -1.33030438e+00
-4.58284616e-01 -1.33676052e-01 8.97987485e-02 3.83792102e-01
6.99421167e-02 4.75900546e-02 -5.02745271e-01 -5.23031414e-01
1.44748569e-01 -7.48992562e-01 3.48548032e-02 9.02199745e-03
-1.48721442e-01 6.00013018e-01 6.56925261e-01 -3.57393622e-01
1.32948923e+00 -2.33752918e+00 -3.32352608e-01 -5.38312038e-03
4.78888452e-01 1.83747724e-01 1.42209716e-02 8.45793784e-02
-1.56943902e-01 -3.37783813e-01 -3.35478574e-01 8.66394788e-02
-3.77598405e-01 -2.76632786e-01 -2.79822260e-01 6.50425136e-01
2.12219963e-03 6.75654292e-01 -8.35521281e-01 -2.88228005e-01
5.16286671e-01 6.32429361e-01 -7.43911207e-01 8.98835361e-02
1.08561374e-01 3.99988383e-01 3.65582764e-01 3.89309466e-01
1.18123174e+00 -1.55749589e-01 8.31484869e-02 -5.21366537e-01
-6.57516956e-01 -1.01339579e-01 -7.51393974e-01 1.21602559e+00
-4.72116232e-01 1.07522368e+00 -4.19638120e-02 -3.26036692e-01
7.94877172e-01 -1.35253385e-01 2.17083588e-01 -1.14860821e+00
1.45021945e-01 3.80799562e-01 1.98541522e-01 -2.81622857e-01
5.98938465e-01 -4.03739601e-01 3.34059328e-01 1.48946986e-01
-2.59225905e-01 -5.02282083e-01 -1.08445987e-01 -1.91063300e-01
8.90884280e-01 -1.96558222e-01 3.97311419e-01 -5.28762162e-01
4.46148217e-01 -4.72923964e-01 4.84605134e-02 7.65541375e-01
-3.92838269e-01 1.07066345e+00 4.88262832e-01 -5.48134074e-02
-1.16963208e+00 -1.23790205e+00 -5.22588730e-01 4.61556643e-01
5.10479927e-01 -8.29542950e-02 -1.21180689e+00 2.79859062e-02
-2.20079005e-01 6.41106725e-01 -3.74927968e-01 -2.07493957e-02
-4.47040856e-01 -5.04334331e-01 1.47514641e-01 6.59044683e-02
8.10323536e-01 -6.55618608e-01 -1.05067980e+00 4.52767871e-02
-2.47573584e-01 -8.88443530e-01 -8.73682082e-01 8.77421945e-02
-6.18795156e-01 -9.32287812e-01 -1.15367246e+00 -5.86002111e-01
8.12440634e-01 5.35238504e-01 9.87082601e-01 -2.78146088e-01
-4.52413768e-01 1.65735260e-01 -3.82342166e-03 -1.78227156e-01
-2.06081256e-01 -6.02041423e-01 -2.50190586e-01 2.77910270e-02
-4.82995026e-02 -5.39906681e-01 -1.18853867e+00 3.97882581e-01
-1.03893304e+00 9.63466093e-02 7.34316528e-01 7.25021482e-01
6.27328873e-01 3.89957339e-01 -2.02545270e-01 -4.94481921e-01
7.70533800e-01 3.24683934e-01 -9.93235290e-01 2.11960569e-01
-8.20106566e-01 2.26808503e-01 5.03343225e-01 -4.62278932e-01
-1.21268106e+00 -1.35330856e-01 4.04320359e-01 -4.46130604e-01
1.56874970e-01 7.96181485e-02 1.24824762e-01 -5.81603467e-01
1.09255886e+00 1.71544850e-01 1.47581279e-01 -2.70607233e-01
1.77288100e-01 5.08370519e-01 1.07906091e+00 -2.19161049e-01
5.59366047e-01 6.72421098e-01 2.41103962e-01 -8.18531275e-01
-5.94581783e-01 -3.15946519e-01 -7.04421699e-02 -3.58767629e-01
8.14872682e-01 -4.65108037e-01 -7.71520019e-01 6.57881796e-01
-9.71266031e-01 -4.88132179e-01 -4.33950454e-01 7.62932301e-01
-7.41034150e-01 3.61845106e-01 -4.44757491e-01 -7.37582862e-01
-3.70470667e-03 -1.05232108e+00 5.56714177e-01 5.64560950e-01
1.52553514e-01 -7.07298100e-01 1.58768341e-01 1.87548995e-02
8.82562637e-01 5.43768480e-02 7.45785236e-01 4.98342156e-01
-7.52364814e-01 2.63670623e-01 -4.65694100e-01 4.75001276e-01
3.05484235e-01 -2.33655572e-01 -1.25640976e+00 -3.09418947e-01
4.15564477e-01 2.18881994e-01 9.79228735e-01 1.10397911e+00
9.77064669e-01 -3.99874970e-02 1.10473745e-01 9.92482722e-01
1.74390352e+00 3.74454021e-01 1.25930965e+00 3.52698356e-01
1.27028733e-01 6.27426982e-01 6.54047728e-01 1.74771503e-01
-2.40740359e-01 9.11410689e-01 1.90073773e-01 -6.03208899e-01
-5.81830800e-01 -1.24032915e-01 7.24962801e-02 2.84682274e-01
-8.48173816e-03 -1.81466192e-01 -6.74090922e-01 2.81476140e-01
-1.45154786e+00 -9.79436219e-01 7.54715083e-03 2.76417804e+00
8.83540750e-01 1.42496035e-01 -2.04485476e-01 9.53901187e-02
7.41790891e-01 -7.85822794e-03 -3.10376316e-01 -2.50758767e-01
-3.55244994e-01 3.91209483e-01 8.93628955e-01 1.18160033e+00
-8.87950718e-01 6.22331321e-01 7.15287352e+00 6.19555056e-01
-1.52876759e+00 -1.76110223e-01 5.89385927e-01 -1.35499969e-01
-3.88501257e-01 1.70208246e-01 -3.92317653e-01 6.04424238e-01
6.13634229e-01 -2.48725563e-01 5.79617143e-01 3.25671375e-01
3.95438731e-01 -6.42754674e-01 -8.57736886e-01 1.20161343e+00
1.21211164e-01 -1.02204478e+00 -2.71595895e-01 1.87155515e-01
6.91733479e-01 -2.96537697e-01 5.26805043e-01 -3.46045166e-01
-3.42076004e-01 -1.12951946e+00 4.76899892e-01 8.87465954e-01
1.02982843e+00 -7.17453957e-01 5.49732685e-01 -7.82161504e-02
-7.42086112e-01 1.07969485e-01 -3.72472465e-01 5.07088788e-02
8.29235986e-02 8.81731451e-01 -7.22161353e-01 1.45480096e-01
4.28628832e-01 3.56908739e-02 -5.06470799e-01 1.53058171e+00
2.35422440e-02 1.56681895e-01 -9.14980471e-02 2.36464739e-01
-1.98729888e-01 -3.37441593e-01 5.97565293e-01 1.05270660e+00
4.15673643e-01 2.65935987e-01 -2.69753873e-01 9.39719975e-01
1.16707951e-01 -2.18205974e-02 -4.15973663e-01 2.89115608e-01
3.57282281e-01 8.63668323e-01 -5.55699885e-01 -1.73436984e-01
-4.49694425e-01 1.04548967e+00 -1.87007293e-01 6.30169272e-01
-5.83218992e-01 -8.05128813e-01 4.94053155e-01 6.83342159e-01
3.34049203e-02 -2.08916087e-02 -3.58224809e-01 -9.39312577e-01
3.97851244e-02 -6.88413203e-01 -1.52390271e-01 -1.27567387e+00
-7.31359303e-01 5.72861373e-01 -8.40403065e-02 -1.33146787e+00
-2.00175941e-02 -5.25798500e-01 -5.27332246e-01 1.20226681e+00
-1.74376309e+00 -5.49886644e-01 -5.55871725e-01 5.76229692e-01
1.46676555e-01 1.98958054e-01 3.29989821e-01 3.10228139e-01
2.17562556e-01 4.79693711e-01 8.38625208e-02 -3.39771986e-01
1.01697195e+00 -1.26673818e+00 9.67528764e-03 1.15357280e+00
-4.47325885e-01 6.26855254e-01 1.10427880e+00 -4.92265880e-01
-9.24408376e-01 -5.89317977e-01 7.38355935e-01 -1.67165741e-01
4.72030431e-01 6.02118708e-02 -9.69237208e-01 1.98556229e-01
3.85711253e-01 -7.47711062e-02 2.54844159e-01 -4.52449232e-01
-1.43673524e-01 -3.77653003e-01 -1.23199546e+00 7.44296432e-01
6.45197928e-01 -6.01668119e-01 -3.29803109e-01 -7.06821382e-02
6.04132771e-01 -5.96607447e-01 -5.42265296e-01 5.00681698e-01
5.88098824e-01 -1.65578175e+00 8.42279613e-01 3.00286323e-01
4.14482802e-01 -7.20096409e-01 -1.55318439e-01 -1.27926469e+00
-4.36765552e-01 -9.77864504e-01 3.70832980e-01 8.27532053e-01
2.33785138e-01 -8.76392603e-01 5.82311153e-01 3.68145972e-01
-7.00934157e-02 -3.04015428e-01 -5.98610401e-01 -7.80243456e-01
-1.48561567e-01 1.54643226e-02 1.52748093e-01 6.15374029e-01
1.01368442e-01 -1.86215877e-03 -4.14867759e-01 1.37125105e-01
1.01377249e+00 1.39769807e-01 5.39944828e-01 -7.38722026e-01
-5.14305949e-01 -4.74082738e-01 -5.03489792e-01 -1.30157411e+00
-5.53457499e-01 -3.56561571e-01 1.07556604e-01 -1.11491323e+00
3.26490968e-01 -2.61503667e-01 -3.35135192e-01 1.88651858e-04
-1.18903697e-01 4.30856317e-01 1.58474877e-01 1.66344061e-01
1.07216626e-01 1.42135844e-01 1.70643246e+00 1.00017726e-01
-5.93591928e-01 -7.53451660e-02 -7.05609620e-01 4.54621851e-01
7.36542583e-01 -6.00669906e-02 -6.43110812e-01 -2.38369256e-01
9.85288471e-02 2.61666298e-01 4.73244578e-01 -1.16455913e+00
3.77543092e-01 -1.23184413e-01 3.90440375e-01 -4.93114293e-01
2.96168327e-01 -5.95491409e-01 1.95744991e-01 4.63726133e-01
-3.22945327e-01 -2.92899728e-01 1.10141173e-01 1.15075357e-01
-4.02732104e-01 -4.13877517e-01 1.40370464e+00 8.71644095e-02
-2.93190718e-01 -1.85179830e-01 -2.33819306e-01 -2.05504148e-06
7.24879086e-01 -3.05296063e-01 -7.65930712e-01 -5.78435123e-01
-3.44779968e-01 -4.10179347e-01 1.15628648e+00 -1.38561264e-01
6.99619949e-01 -1.03576636e+00 -5.22659540e-01 3.91700923e-01
-1.41372196e-02 -4.98242140e-01 3.67602974e-01 1.10035610e+00
-8.15423906e-01 3.82675648e-01 -3.94326329e-01 -6.53168917e-01
-1.40666533e+00 5.69263160e-01 7.77817369e-01 1.47983521e-01
-5.53073704e-01 6.93852901e-01 6.58644021e-01 3.95174086e-01
8.00408795e-02 -5.63457429e-01 -2.79020462e-02 -4.67357874e-01
7.82481909e-01 2.07328573e-01 5.20729721e-02 -1.79358333e-01
-4.31114398e-02 1.04153681e+00 -3.21054161e-02 -6.31877780e-01
7.84243941e-01 -2.85632730e-01 -1.12109803e-01 2.65080243e-01
1.15804541e+00 6.27408683e-01 -1.55439389e+00 -7.59109929e-02
-5.11516511e-01 -1.10591722e+00 3.69667202e-01 -8.01670730e-01
-8.33077192e-01 8.84863079e-01 1.17459190e+00 3.62880230e-01
1.69963980e+00 -1.91173971e-01 3.64710897e-01 -1.11092962e-01
6.94481805e-02 -9.46574271e-01 2.59832978e-01 1.55591920e-01
7.69509017e-01 -9.08335030e-01 -4.31607245e-03 -6.00715280e-01
-4.38635498e-01 9.50829625e-01 3.46708864e-01 -5.66175096e-02
3.93750399e-01 5.08166194e-01 3.27083558e-01 -7.01985285e-02
-2.91382372e-01 -3.36734325e-01 4.45528239e-01 6.66666746e-01
4.15314078e-01 -5.36952019e-02 -3.73538524e-01 -1.67438075e-01
-4.42246437e-01 1.81430161e-01 8.45542312e-01 7.55908310e-01
-9.23557341e-01 -9.46731687e-01 -6.34092093e-01 1.40035972e-01
-5.61906874e-01 -3.44434708e-01 -8.93594846e-02 4.72739369e-01
1.29899886e-02 8.56961370e-01 2.93726355e-01 -3.34962122e-02
4.16728526e-01 -4.13236797e-01 1.06150043e+00 -1.20301820e-01
-2.91499853e-01 4.67571408e-01 -2.65206039e-01 -5.18886507e-01
-4.25151616e-01 -1.63007826e-01 -1.07070208e+00 -3.57310414e-01
-1.69063777e-01 -1.68118238e-01 6.19113684e-01 2.01942697e-01
1.16335742e-01 5.17195284e-01 5.94282568e-01 -5.70035934e-01
-1.83687896e-01 -5.98745286e-01 -8.19234550e-01 3.85653079e-01
8.08063745e-01 -4.21370029e-01 -6.11876905e-01 9.67002809e-02] | [10.875617980957031, -2.4571712017059326] |
e1088f9d-16fa-4251-b1b0-dac2b89d94f7 | use-of-a-taxonomy-of-empathetic-response-1 | 2305.10096 | null | https://arxiv.org/abs/2305.10096v1 | https://arxiv.org/pdf/2305.10096v1.pdf | Use of a Taxonomy of Empathetic Response Intents to Control and Interpret Empathy in Neural Chatbots | A recent trend in the domain of open-domain conversational agents is enabling them to converse empathetically to emotional prompts. Current approaches either follow an end-to-end approach or condition the responses on similar emotion labels to generate empathetic responses. But empathy is a broad concept that refers to the cognitive and emotional reactions of an individual to the observed experiences of another and it is more complex than mere mimicry of emotion. Hence, it requires identifying complex human conversational strategies and dynamics in addition to generic emotions to control and interpret empathetic responding capabilities of chatbots. In this work, we make use of a taxonomy of eight empathetic response intents in addition to generic emotion categories in building a dialogue response generation model capable of generating empathetic responses in a controllable and interpretable manner. It consists of two modules: 1) a response emotion/intent prediction module; and 2) a response generation module. We propose several rule-based and neural approaches to predict the next response's emotion/intent and generate responses conditioned on these predicted emotions/intents. Automatic and human evaluation results emphasize the importance of the use of the taxonomy of empathetic response intents in producing more diverse and empathetically more appropriate responses than end-to-end models. | ['Pearl Pu', 'Anuradha Welivita'] | 2023-05-17 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [-2.50264853e-01 4.68120754e-01 2.06575155e-01 -7.75179982e-01
-1.11703709e-01 -4.68895018e-01 8.05302918e-01 -1.07702993e-01
-1.67250603e-01 8.85462642e-01 6.78801298e-01 2.54392296e-01
3.14363912e-02 -6.22015774e-01 4.20053184e-01 -2.78544843e-01
3.85939479e-01 7.87070036e-01 -4.30281699e-01 -9.82406378e-01
5.27740479e-01 3.85240138e-01 -1.21144366e+00 8.18577945e-01
6.70711160e-01 6.90686584e-01 -1.87481657e-01 1.00677800e+00
-1.87434033e-01 1.82910061e+00 -8.22905660e-01 -5.07394433e-01
-1.93538100e-01 -1.06854010e+00 -1.63500619e+00 -2.28490457e-01
-7.49572754e-01 -2.71528721e-01 1.99778944e-01 4.64042515e-01
6.31824732e-01 7.67649770e-01 8.82990777e-01 -1.63349485e+00
-7.78709114e-01 6.23400450e-01 2.86554277e-01 -2.04739302e-01
1.31255198e+00 3.35408062e-01 7.09227204e-01 -3.88747841e-01
8.08058918e-01 1.36180353e+00 7.49203205e-01 1.39032149e+00
-9.76431847e-01 -3.95260483e-01 -3.65871698e-01 2.26372227e-01
-4.88014728e-01 -5.03677487e-01 8.95275712e-01 -5.86933076e-01
1.08852458e+00 2.96471566e-01 6.43305659e-01 1.64763534e+00
5.50329834e-02 4.60631400e-01 1.22778034e+00 -4.35014248e-01
3.44322652e-01 6.30604327e-01 1.32946417e-01 3.12772691e-02
-8.75952840e-01 2.93783039e-01 -5.93577981e-01 -4.59426135e-01
5.11471987e-01 -3.47502530e-01 -1.14559121e-01 2.67141581e-01
-9.16146159e-01 1.39506781e+00 4.85862494e-01 4.27387744e-01
-1.01429057e+00 -1.42878711e-01 8.78211975e-01 6.21660113e-01
4.33375925e-01 1.11912847e+00 -7.17625916e-02 -6.70831203e-01
-2.25996509e-01 4.24469352e-01 1.58449769e+00 6.50007010e-01
5.38080692e-01 -1.16213337e-01 -3.97394091e-01 1.27347279e+00
1.02761991e-01 -2.50716377e-02 5.14593899e-01 -1.37967944e+00
-1.94367051e-01 8.43831182e-01 5.76138616e-01 -1.00740099e+00
-7.54437208e-01 3.82134497e-01 -4.00242150e-01 3.50360394e-01
3.09724629e-01 -7.67292976e-01 1.69724137e-01 1.83335781e+00
5.07883370e-01 -4.97377545e-01 6.79596126e-01 1.15296340e+00
1.10760021e+00 9.36866283e-01 5.23561656e-01 -3.46916437e-01
1.24757290e+00 -1.07330823e+00 -8.64262700e-01 -1.36230782e-01
8.37875307e-01 -7.43694842e-01 1.13492942e+00 1.97687954e-01
-1.12942529e+00 -2.98275739e-01 -4.29558516e-01 -2.09877901e-02
-2.51267493e-01 -2.47714132e-01 8.46583366e-01 2.30633736e-01
-8.95380199e-01 3.93262058e-01 1.39561072e-01 -7.41670012e-01
-3.63475800e-01 2.62250364e-01 -3.15437198e-01 5.51083446e-01
-1.58253717e+00 1.59567440e+00 4.89486568e-02 -3.02504301e-01
-1.09373391e-01 -5.14528692e-01 -6.87227607e-01 -8.39083642e-02
-3.65445703e-01 -7.99838245e-01 1.74868512e+00 -1.56610370e+00
-2.46494794e+00 1.00667691e+00 2.27661416e-01 -2.04269439e-01
2.84913987e-01 1.22140720e-01 -3.42844009e-01 3.23009133e-01
-7.83406943e-02 9.71065879e-01 2.41200984e-01 -1.16827977e+00
-4.34497595e-01 -3.28831300e-02 1.40878990e-01 5.51943898e-01
-2.63210177e-01 9.13038492e-01 7.10377634e-01 -1.44552946e-01
-5.98992467e-01 -8.40341091e-01 -2.45590717e-01 -3.11664373e-01
1.00932434e-01 -6.96104169e-01 5.16879559e-01 -4.17334616e-01
8.06365430e-01 -1.77120137e+00 7.11427182e-02 -1.51381567e-01
1.20846570e-01 1.92250103e-01 -2.23002180e-01 1.22826886e+00
-1.59537137e-01 -3.21085393e-01 2.66731322e-01 -1.61096588e-01
4.04735982e-01 6.18825629e-02 -3.63018930e-01 -4.28278977e-03
1.47531435e-01 7.84668684e-01 -1.04587090e+00 -5.07638752e-01
2.64142126e-01 3.51508081e-01 -6.44103050e-01 1.05475008e+00
-1.38582557e-01 9.04678285e-01 -6.19125605e-01 -4.05934639e-02
4.44933660e-02 1.99144021e-01 4.87002507e-02 3.99720669e-01
-8.40560347e-02 4.46653128e-01 -4.25761044e-01 1.14024270e+00
-9.42951202e-01 4.75964665e-01 1.06846400e-01 -5.67381799e-01
1.64054465e+00 9.69723165e-01 4.06903118e-01 -5.00467122e-01
6.00248873e-01 -1.72713958e-02 1.31481647e-01 -9.42023695e-01
5.88295579e-01 -7.39376545e-01 -7.29550898e-01 1.24335408e+00
-1.06624197e-02 -4.77779806e-01 -1.48337170e-01 1.65116280e-01
1.01536906e+00 1.50804371e-01 5.98042965e-01 -5.33337519e-03
7.12631106e-01 2.74948001e-01 3.09188187e-01 3.89244020e-01
-5.71601272e-01 7.42672458e-02 6.11594915e-01 -7.64180064e-01
-7.40002334e-01 -4.14362133e-01 4.71953183e-01 1.50120437e+00
-5.22217378e-02 7.69209713e-02 -9.27780509e-01 -2.71413982e-01
-6.64575756e-01 1.32193065e+00 -5.93730211e-01 -4.71793681e-01
-3.64160478e-01 -1.97127134e-01 9.71975327e-01 2.78166354e-01
2.98595011e-01 -2.11408758e+00 -1.00071335e+00 4.54237252e-01
-7.89899707e-01 -9.51767087e-01 -1.78680614e-01 9.47596133e-02
-4.02814031e-01 -7.16127515e-01 -1.78349525e-01 -7.46834099e-01
3.96611691e-01 -1.10428013e-01 1.04526806e+00 3.04090083e-02
1.89784661e-01 4.95705366e-01 -1.01551890e+00 -4.17925119e-01
-1.33696342e+00 -2.95193285e-01 -1.10965278e-02 -7.74431042e-03
5.65717995e-01 -7.06688762e-01 -3.63413990e-01 4.57564116e-01
-4.16329563e-01 2.56977767e-01 1.53711159e-03 8.01871419e-01
-4.83406901e-01 -7.77214348e-01 1.21236300e+00 -7.53457487e-01
1.78753757e+00 -8.68448019e-01 4.85558480e-01 1.58276603e-01
-2.26260364e-01 -2.50227749e-01 1.15760350e+00 -7.96156168e-01
-1.73507011e+00 3.72199826e-02 -3.52474004e-01 -5.26530668e-02
-6.77576244e-01 2.04745471e-01 3.42904687e-01 -3.89265120e-02
1.18128836e+00 -1.76285475e-01 3.50142866e-01 4.87288646e-02
5.40780306e-01 1.25404334e+00 6.77891731e-01 -1.08936095e+00
5.51065281e-02 -1.22238934e-01 -4.73524451e-01 -4.32243794e-01
-5.01504540e-01 -4.78704274e-01 -2.24469170e-01 -9.66217995e-01
8.36838841e-01 -5.95126867e-01 -1.42796409e+00 3.59357715e-01
-1.56839275e+00 -7.06022859e-01 -3.39807898e-01 3.08795005e-01
-1.27283287e+00 1.76267922e-01 -1.10791278e+00 -1.05360806e+00
-7.93809593e-01 -7.21922338e-01 4.95086491e-01 5.94567418e-01
-1.61535609e+00 -1.15982914e+00 4.91787821e-01 6.29907191e-01
6.29146278e-01 3.31567854e-01 8.60033751e-01 -1.16311550e+00
7.73676157e-01 -3.86965811e-01 1.60200700e-01 1.29000485e-01
-8.16134140e-02 -5.79485260e-02 -8.18915367e-01 4.81285453e-01
4.64226127e-01 -1.18280816e+00 -4.07740951e-01 -2.60513961e-01
2.27975368e-01 -7.75357842e-01 2.75319755e-01 -2.29307562e-02
6.99340820e-01 4.61453021e-01 8.02650571e-01 8.51690918e-02
-1.33104622e-01 1.44798946e+00 9.53886390e-01 8.52692485e-01
6.20476723e-01 6.64565086e-01 1.75712645e-01 5.10332026e-02
3.30353945e-01 -3.15505147e-01 6.49886608e-01 4.49709296e-01
-1.79768026e-01 -1.35304958e-01 -6.66047037e-01 3.02517086e-01
-2.00004721e+00 -1.52706730e+00 -5.14174223e-01 1.53869569e+00
1.06832778e+00 -5.93660355e-01 4.03948665e-01 -1.95331767e-01
7.04698801e-01 -9.80761945e-02 -1.81696847e-01 -1.82556033e+00
2.28762522e-01 2.52880275e-01 -5.44421434e-01 7.42691219e-01
-2.96593338e-01 1.20387149e+00 5.99799156e+00 1.24095909e-01
-9.82984424e-01 1.08376935e-01 5.17197728e-01 1.41220480e-01
-8.38162154e-02 1.27485581e-03 -1.42597854e-01 1.62321687e-01
1.12381983e+00 -2.03932077e-01 5.60301542e-01 1.03845000e+00
6.97015941e-01 -6.02937788e-02 -1.26978517e+00 6.95871055e-01
5.99486753e-02 -9.31004524e-01 -3.27453554e-01 -4.57784861e-01
3.69433641e-01 -7.11005926e-01 -5.01456022e-01 6.50340497e-01
7.43554950e-01 -1.04308283e+00 5.11389971e-01 7.62358129e-01
2.72667229e-01 -6.24576867e-01 8.83105755e-01 7.50389814e-01
-4.60124195e-01 -1.20566688e-01 -4.85572070e-02 -9.53311682e-01
4.23885465e-01 -5.59571385e-01 -1.20092320e+00 3.24272178e-02
3.19290936e-01 2.99854755e-01 3.11324924e-01 3.83632302e-01
-5.52847981e-01 1.82678789e-01 7.07829073e-02 -7.42960930e-01
2.08202824e-01 -3.68132710e-01 4.88491088e-01 1.34267950e+00
9.40738060e-03 7.29632735e-01 -7.79666677e-02 1.11598778e+00
3.49486738e-01 4.41815406e-01 -6.58322573e-01 5.73840439e-02
8.27258527e-01 1.53982127e+00 -2.36991733e-01 -1.50036663e-01
4.40828055e-02 1.06440771e+00 3.85430396e-01 1.03652000e-01
-8.70016158e-01 -2.38961264e-01 6.59706414e-01 -4.38003689e-01
-5.73054075e-01 5.39095044e-01 -3.20446074e-01 -6.33100033e-01
-5.52109003e-01 -1.14206421e+00 3.10615182e-01 -1.43613029e+00
-1.64654088e+00 9.00976777e-01 -1.78425193e-01 -9.50271964e-01
-1.06887484e+00 -2.87841797e-01 -1.38750947e+00 1.07273769e+00
-6.07577324e-01 -1.16374099e+00 -4.67340827e-01 7.56576896e-01
4.87907708e-01 -1.76984910e-03 1.56318247e+00 -1.90665379e-01
-2.47947171e-01 3.84014636e-01 -7.20849752e-01 -1.03412345e-02
9.37094212e-01 -8.31459761e-01 -2.48803049e-01 7.06545562e-02
-6.86948776e-01 4.60939348e-01 1.11862206e+00 -1.65660530e-01
-7.73354650e-01 -4.72655833e-01 1.39592159e+00 -4.71544415e-01
6.63926959e-01 1.87404547e-02 -7.40016818e-01 5.53643465e-01
7.42350101e-01 -6.97933555e-01 1.11637509e+00 5.87344170e-02
-1.00613922e-01 4.39484656e-01 -1.54364073e+00 8.59971821e-01
6.57568932e-01 -3.41117382e-01 -1.12528372e+00 4.16462541e-01
4.33338284e-01 -1.71214342e-01 -1.07107759e+00 -1.14786088e-01
3.82888675e-01 -1.30842090e+00 6.80294991e-01 -1.08171034e+00
8.80004644e-01 2.36719549e-01 1.76304281e-01 -1.56343079e+00
-1.63111746e-01 -1.15557277e+00 3.16087037e-01 1.35263705e+00
1.37809157e-01 -7.19230652e-01 4.43376243e-01 1.27170670e+00
-2.68049836e-01 -5.77353239e-01 -6.34265184e-01 -5.65369539e-02
2.44747251e-01 -2.23995939e-01 5.18850863e-01 1.22890890e+00
1.32340109e+00 9.28229630e-01 -6.42798781e-01 -4.89948332e-01
-2.11101025e-01 5.61869219e-02 1.18181181e+00 -1.27629089e+00
-1.40733451e-01 -6.51036978e-01 -4.33480144e-02 -6.52686715e-01
7.13874161e-01 -6.64066911e-01 5.06691575e-01 -1.42906666e+00
1.73819717e-03 -3.58354151e-01 5.42370677e-01 6.13485456e-01
-1.89067982e-03 -1.38079405e-01 2.15053082e-01 2.27422744e-01
-4.69658375e-01 6.69184208e-01 1.08620000e+00 5.14354706e-01
-5.78671694e-01 2.90331662e-01 -8.99067819e-01 8.88126433e-01
1.07085562e+00 -5.48776746e-01 -3.88726503e-01 3.30618262e-01
3.07065904e-01 8.98198247e-01 4.00462776e-01 -6.35737240e-01
3.01404238e-01 -7.24699736e-01 -1.71075895e-01 1.82564348e-01
6.07841134e-01 -4.52606797e-01 1.12631038e-01 3.52932811e-01
-8.26684296e-01 1.46424677e-02 -1.50932312e-01 -4.44319993e-02
-3.39595467e-01 -5.47963738e-01 1.06422436e+00 -2.26723373e-01
-6.22219145e-01 -4.56261516e-01 -1.20636308e+00 -6.16129413e-02
1.29950917e+00 -5.11317611e-01 -3.72461021e-01 -1.35282183e+00
-6.91835403e-01 3.14904571e-01 4.55822766e-01 6.87116146e-01
5.93644679e-01 -1.14291739e+00 -7.96543419e-01 -3.62605602e-01
1.59123600e-01 -6.26263857e-01 4.34853047e-01 7.47147560e-01
-5.05728960e-01 1.85428053e-01 -8.66475582e-01 4.08238173e-02
-1.24059308e+00 2.77908981e-01 6.88404799e-01 -3.36895347e-01
-4.12939757e-01 7.79074550e-01 -3.48378979e-02 -1.00825632e+00
1.50135309e-02 5.98672152e-01 -8.31162751e-01 6.11383021e-02
5.24482608e-01 3.22793096e-01 -4.79798347e-01 -8.32899988e-01
-3.68924513e-02 -2.36250497e-02 3.45998138e-01 -4.67914850e-01
1.21954167e+00 -4.11181450e-02 -3.66088897e-01 2.96801627e-01
7.49050677e-01 -2.72645414e-01 -7.69522905e-01 1.32732958e-01
1.93451606e-02 -1.31389886e-01 -7.51795828e-01 -1.18814468e+00
-3.28501999e-01 6.64068103e-01 -1.97435424e-01 4.74767804e-01
1.00015986e+00 -9.00461227e-02 8.90231133e-01 3.87571901e-01
2.94140249e-01 -1.45603788e+00 5.34535289e-01 8.43021750e-01
1.29734373e+00 -9.74821806e-01 -4.98531967e-01 -1.71112537e-01
-1.61204529e+00 1.39223218e+00 1.06716490e+00 -7.83383176e-02
-4.18506861e-02 1.39362603e-01 8.04888129e-01 -2.68068284e-01
-1.32503712e+00 2.49104559e-01 -2.39131138e-01 9.11246896e-01
7.01664805e-01 1.24904752e-01 -7.16848910e-01 1.00273168e+00
-5.51245928e-01 4.05614227e-02 8.83883893e-01 5.90708375e-01
-3.43633413e-01 -1.14311004e+00 -4.24079299e-01 -1.12482347e-01
-4.86786664e-02 3.47208977e-01 -1.42260242e+00 5.82130253e-01
-2.67870307e-01 1.73976135e+00 -1.88815683e-01 -6.31342411e-01
4.18773443e-01 4.59915787e-01 5.36860488e-02 -7.33066916e-01
-1.63092351e+00 -7.81461656e-01 8.24716210e-01 -4.70203072e-01
-6.28504395e-01 -5.85772991e-01 -1.62694347e+00 -8.10432494e-01
-6.36588186e-02 4.81966376e-01 4.61186409e-01 1.03473878e+00
2.63407171e-01 -2.88037099e-02 1.12914824e+00 -9.78013277e-01
-9.25492823e-01 -1.38536489e+00 -1.38811022e-01 1.09906220e+00
-5.12552142e-01 -3.19617391e-01 -4.38305855e-01 -1.25562727e-01] | [13.121760368347168, 7.676385879516602] |
0509cf77-0b7c-40bf-839e-54913a3183e4 | amelioration-de-la-qualite-d-images-avec-un | 2303.07151 | null | https://arxiv.org/abs/2303.07151v1 | https://arxiv.org/pdf/2303.07151v1.pdf | Amélioration de la qualité d'images avec un algorithme d'optimisation inspirée par la nature | Reproducible images preprocessing is important in the field of computer vision, for efficient algorithms comparison or for new images corpus preparation. In this paper, we propose a method to obtain an explicit and ordered sequence of transformations that improves a given image: the computation is performed via a nature-inspired optimization algorithm based on quality assessment techniques. Preliminary tests show the impact of the approach on different state-of-the-art data sets. -- L'application de pr\'etraitements explicites et reproductibles est fondamentale dans le domaine de la vision par ordinateur, pour pouvoir comparer efficacement des algorithmes ou pour pr\'eparer un nouveau corpus d'images. Dans cet article, nous proposons une m\'ethode pour obtenir une s\'equence reproductible de transformations qui am\'eliore une image donn\'ee: le calcul est r\'ealis\'e via un algorithme d'optimisation inspir\'ee par la nature et bas\'e sur des techniques d'\'evaluation de la qualit\'e. Des tests montrent l'impact de l'approche sur diff\'erents ensembles d'images de l'\'etat de l'art. | ['Thomas Tamisier', 'Olivier Parisot'] | 2023-03-13 | null | null | null | null | ['nature-inspired-optimization-algorithm'] | ['computer-code'] | [ 3.28625828e-01 -1.66663870e-01 4.05935109e-01 -4.96758670e-01
-4.71301466e-01 -5.26845515e-01 8.57053101e-01 6.55449033e-01
-9.25105333e-01 3.85431617e-01 5.24339303e-02 7.66810551e-02
9.90714785e-03 -9.86547410e-01 -7.90820062e-01 -5.10696530e-01
9.20089558e-02 3.75222951e-01 5.67347407e-02 -3.53350639e-01
3.47154766e-01 5.97369730e-01 -1.64058208e+00 2.56014854e-01
9.37483072e-01 9.12019789e-01 4.83700842e-01 3.81357878e-01
-9.65825394e-02 4.10232902e-01 -4.50651139e-01 -3.74366581e-01
7.16982901e-01 -1.02634811e+00 -5.00942111e-01 2.06747577e-01
5.68668544e-01 5.56575477e-01 8.32702313e-03 1.26892745e+00
3.01007718e-01 6.83987960e-02 1.58514309e+00 -5.98692656e-01
-6.84290528e-01 4.20439363e-01 -2.51440912e-01 1.01736033e+00
3.92278135e-01 3.21917087e-01 1.07568312e+00 -4.34708834e-01
8.41315329e-01 6.74574852e-01 8.06647956e-01 3.67056638e-01
-1.31179774e+00 -1.86334237e-01 -1.72751263e-01 2.18519419e-01
-1.26180661e+00 -4.13736254e-01 2.21933782e-01 -6.68311059e-01
3.79521877e-01 3.43755007e-01 2.91195899e-01 6.21523857e-01
3.72014433e-01 7.23178208e-01 1.44796634e+00 -3.07382524e-01
2.27942422e-01 6.80475950e-01 -3.14794600e-01 7.80392408e-01
-8.57654586e-02 2.96691298e-01 -1.55354440e-01 3.30197066e-02
7.90408194e-01 -6.24454655e-02 -2.71816790e-01 4.95841354e-02
-9.50533509e-01 8.25167894e-01 9.94644761e-01 6.52853727e-01
-5.17903030e-01 -3.13632697e-01 3.87658566e-01 2.43020669e-01
2.94652969e-01 8.11328351e-01 2.49845404e-02 -1.61915183e-01
-5.76651216e-01 3.30671847e-01 8.15017819e-01 5.59294105e-01
5.68320036e-01 -1.23269871e-01 1.93347678e-01 6.55834317e-01
1.77194010e-02 2.72963911e-01 2.01695383e-01 -2.76869595e-01
4.31360304e-02 4.97783899e-01 -6.78876042e-02 -8.51109266e-01
-2.20194250e-01 -6.29978418e-01 -7.38089085e-01 5.00351727e-01
4.41295743e-01 2.88684428e-01 -3.42104197e-01 1.38007283e+00
-1.20504446e-01 -1.85509846e-01 1.20651528e-01 9.50497746e-01
1.34938240e-01 1.63382202e-01 3.05316001e-01 -4.37564015e-01
1.40212500e+00 -2.12031767e-01 -6.75974265e-02 1.16218127e-01
4.06250715e-01 -9.69786704e-01 1.09676862e+00 7.90449977e-01
-8.56703162e-01 -6.80078626e-01 -9.87684369e-01 3.68932970e-02
-1.81648731e-01 3.32883269e-01 -1.77690208e-01 3.72216076e-01
-7.61786759e-01 6.51902556e-01 -2.76719511e-01 -3.42246473e-01
2.25206524e-01 1.00907348e-01 -4.96109962e-01 3.65178436e-01
-3.34646463e-01 1.33030045e+00 6.81437075e-01 2.63708621e-01
-6.29912853e-01 -6.68061431e-03 -8.15555274e-01 2.26615161e-01
6.78547561e-01 -4.07481551e-01 1.38455677e+00 -1.42020392e+00
-1.06194520e+00 9.36375260e-01 -4.08849657e-01 -5.83547235e-01
8.51556361e-01 2.37943232e-01 -4.52303737e-01 2.24643975e-01
4.40643961e-03 5.96281588e-01 6.24772787e-01 -1.40053976e+00
-9.25402701e-01 4.30404283e-02 -5.18804695e-03 1.71177641e-01
-4.75908011e-01 6.63185492e-02 -1.91572636e-01 -1.12408447e+00
1.79739296e-01 -1.23558176e+00 1.06083415e-01 4.15052146e-01
-3.36659819e-01 9.03237313e-02 8.74407828e-01 -3.45656425e-01
9.33423579e-01 -1.95084536e+00 5.96904576e-01 1.37052804e-01
4.30363089e-01 7.62526751e-01 -6.43613219e-01 2.39549711e-01
-1.65732935e-01 9.64218676e-02 -7.86981821e-01 -1.55827031e-01
2.87143085e-02 -4.99847457e-02 4.41981584e-01 1.59947291e-01
-1.87505648e-01 5.43277383e-01 -6.80562437e-01 -5.50537765e-01
1.84140190e-01 3.51334602e-01 -5.49404442e-01 6.30627051e-02
-5.81808686e-01 4.56291497e-01 -4.48869616e-01 -1.83488414e-01
3.57166886e-01 -2.27515981e-01 3.15877907e-02 -4.19052750e-01
-5.04560590e-01 -1.96865007e-01 -1.28193879e+00 1.45702112e+00
-4.60645080e-01 4.11922157e-01 -3.87675175e-03 -9.62412655e-01
1.70603192e+00 2.84071356e-01 4.74615037e-01 -5.31004250e-01
5.87540090e-01 4.18037593e-01 -9.55290645e-02 -3.37861836e-01
1.80162653e-01 -3.90404791e-01 6.29464149e-01 9.31273878e-01
-3.25713158e-01 -1.41572207e-01 9.94704366e-01 -7.12132692e-01
1.00850224e+00 6.86418861e-02 1.30435854e-01 -6.58351064e-01
1.22421801e+00 -3.48685719e-02 2.03915924e-01 2.24742517e-01
1.81558684e-01 1.42169118e-01 4.35030222e-01 -4.82946336e-01
-1.52626824e+00 -9.78428364e-01 -1.69767052e-01 7.41240025e-01
-3.11054230e-01 -4.46883179e-02 -8.24702024e-01 -8.12484860e-01
-1.78620189e-01 8.54306459e-01 -6.21735215e-01 1.88283041e-01
-2.12906122e-01 -9.99731779e-01 2.28637338e-01 5.17719686e-01
9.84378994e-01 -9.67643857e-01 -1.06864917e+00 3.56889844e-01
-2.26738378e-01 -1.02803183e+00 -2.20138341e-01 -3.44416410e-01
-6.30696714e-01 -9.79027867e-01 -5.80214500e-01 -2.77946740e-01
8.52276802e-01 2.46327773e-01 1.15419030e+00 -4.38126713e-01
-7.66463399e-01 1.29010215e-01 -6.62695050e-01 -6.86650395e-01
-1.18104005e+00 1.34930685e-01 4.12936658e-02 2.57347941e-01
2.84307003e-01 -2.69778818e-01 -1.11922550e+00 8.53310406e-01
-1.26883948e+00 -2.90663928e-01 9.72307622e-01 3.73355240e-01
7.48247325e-01 1.31011769e-01 1.28247172e-01 -7.55007207e-01
9.53964531e-01 -6.38449490e-02 -6.19626403e-01 3.26682925e-02
-4.55514729e-01 4.27493632e-01 1.13972235e+00 1.35316942e-02
-1.12387395e+00 1.62591994e-01 -4.55635160e-01 -2.61724710e-01
-6.33344352e-01 3.51649910e-01 -1.61654219e-01 3.15312773e-01
1.24103606e+00 5.43715894e-01 -1.58520579e-01 -6.30255938e-01
2.93212265e-01 3.10650557e-01 7.10886776e-01 -4.87459540e-01
5.31865358e-01 3.94904912e-01 -2.55415410e-01 -7.10292161e-01
-2.20091179e-01 -1.12824038e-01 -7.59382188e-01 -2.09265918e-01
1.26808059e+00 -3.65651429e-01 -6.80479348e-01 8.54398087e-02
-1.10305011e+00 -2.62365013e-01 -2.84511805e-01 9.60653782e-01
-5.86886466e-01 -2.32222199e-01 -3.12375069e-01 -2.96647131e-01
7.63746798e-02 -1.02472293e+00 8.19028258e-01 8.98132622e-02
3.13688884e-03 -6.37678027e-01 5.35395682e-01 3.28709215e-01
1.07805580e-01 7.90513027e-03 7.01850414e-01 -1.04309358e-01
-1.01981366e+00 -1.91369608e-01 -1.05794497e-01 6.98862553e-01
3.80548596e-01 -2.74304468e-02 -5.94373524e-01 -2.10530624e-01
-2.73744345e-01 8.05333376e-01 7.48199522e-01 7.60749355e-02
8.15883398e-01 -3.09679270e-01 2.43958056e-01 1.92011327e-01
2.07861567e+00 4.43093330e-01 7.51044989e-01 1.78127155e-01
-1.33983761e-01 6.12521172e-01 8.24217975e-01 -2.22710013e-01
4.88219000e-02 9.39165890e-01 3.16670477e-01 2.73003966e-01
4.65015411e-01 5.59834912e-02 8.16597193e-02 8.86087596e-01
-8.17171097e-01 -2.17044801e-01 -8.45242977e-01 4.66722339e-01
-1.44269383e+00 -4.46011186e-01 -3.96423131e-01 2.12023997e+00
5.06671011e-01 3.81932259e-01 3.94233972e-01 6.04548343e-02
6.70172393e-01 1.91688716e-01 -2.58111835e-01 -9.15326178e-01
-7.63010532e-02 3.17593098e-01 1.94454417e-01 4.01899099e-01
-6.45704806e-01 4.79002774e-01 4.17788935e+00 5.46591699e-01
-1.49062848e+00 1.08881243e-01 6.33866251e-01 2.05391482e-01
-2.19153136e-01 -3.98013830e-01 -4.86745507e-01 5.89272499e-01
7.99255192e-01 -1.09860495e-01 1.54846981e-01 5.71810484e-01
2.81292349e-01 -4.71096396e-01 -1.05193365e+00 8.30099523e-01
4.83056158e-01 -1.36791980e+00 -5.70791721e-01 -2.11961418e-01
2.76675940e-01 5.85675165e-02 5.59803434e-02 1.26106873e-01
1.64482266e-01 -9.84704792e-01 6.42095983e-01 6.42730117e-01
8.00809860e-01 -7.07279146e-01 9.16500166e-02 2.18319550e-01
-9.61875916e-01 4.90247458e-02 -1.92201242e-01 3.30526799e-01
3.80765378e-01 1.94677889e-01 -5.42916179e-01 1.00848043e+00
4.00756001e-01 3.97283733e-01 -8.05114746e-01 1.23563468e+00
-2.52314359e-01 3.52699459e-01 -2.25956693e-01 -4.02280182e-01
2.84375161e-01 -7.19379961e-01 6.96282923e-01 1.30068982e+00
5.73072433e-01 5.37691653e-01 -6.02342665e-01 1.00710034e+00
1.63686633e-01 8.56283307e-01 -4.03504997e-01 2.57701986e-02
-2.99452364e-01 5.90223610e-01 -1.35985422e+00 5.98903783e-02
-1.57128155e-01 9.76164281e-01 1.41069852e-02 -2.14112476e-01
-5.51375270e-01 -6.13666415e-01 7.02949613e-02 4.86913845e-02
2.21204564e-01 -5.60025692e-01 -8.13350752e-02 -8.50718915e-01
-3.78531307e-01 -1.05929208e+00 4.25677299e-02 -7.45177448e-01
-8.44889045e-01 1.22986066e+00 1.75481543e-01 -9.35639143e-01
-5.90340197e-01 -3.67955685e-01 -4.32328582e-01 5.22736490e-01
-7.43427694e-01 -4.43256885e-01 -1.38579067e-02 5.04097879e-01
5.98253787e-01 -3.13906074e-01 1.00046468e+00 3.82995725e-01
-7.24965096e-01 -1.20672239e-02 2.06231833e-01 -2.55151331e-01
5.34726262e-01 -1.27427185e+00 1.83695719e-01 1.04562759e+00
7.79135108e-01 2.08810270e-01 1.07712233e+00 -4.95981604e-01
-9.11204994e-01 -8.60093474e-01 7.26659477e-01 -3.84183586e-01
2.07703725e-01 -1.10399224e-01 -9.85206127e-01 9.87266541e-01
1.79187283e-01 -3.25783879e-01 2.81630963e-01 -6.95456862e-02
-1.18928239e-01 -4.96221900e-01 -8.07890654e-01 6.70974195e-01
3.23847651e-01 -1.88757464e-01 -2.87296951e-01 -2.57193446e-02
-2.56811231e-01 -2.63389856e-01 -1.06511021e+00 5.07560313e-01
-9.09507647e-03 -1.69033098e+00 5.34111977e-01 2.34059598e-02
5.24036944e-01 -5.08982599e-01 4.29129571e-01 -9.76034522e-01
5.10441422e-01 -3.90125364e-01 4.89298195e-01 6.86311245e-01
6.35039657e-02 -2.96565503e-01 3.13448578e-01 -2.94524968e-01
1.43223792e-01 -2.15126351e-01 -1.06735301e+00 -3.44324440e-01
-7.67134055e-02 -9.42040384e-01 4.45811063e-01 5.24246395e-01
-7.32148826e-01 4.25772905e-01 4.75853570e-02 -4.61849943e-02
3.61793727e-01 4.59639877e-02 8.78847539e-01 -9.56574619e-01
-7.26487994e-01 -7.21673131e-01 -6.40698314e-01 -8.60329807e-01
5.89654176e-03 -8.81854534e-01 -3.44148815e-01 -1.18342102e+00
4.72142957e-02 -2.36424953e-01 -2.08835274e-01 1.55442029e-01
9.89123359e-02 5.12129724e-01 8.20816219e-01 7.37252608e-02
-5.73503792e-01 4.52889264e-01 1.37335742e+00 1.14654720e-01
3.44306529e-02 6.00569062e-02 -4.23888832e-01 8.12028766e-01
5.20817637e-01 -4.32714015e-01 -4.49729443e-01 -1.87724829e-01
3.38518947e-01 -2.27389932e-01 3.15060079e-01 -1.13252878e+00
4.95902412e-02 -3.19204144e-02 4.75061178e-01 -2.82615960e-01
7.04824567e-01 -6.53901160e-01 3.77639592e-01 4.25555438e-01
-3.62414300e-01 2.47299984e-01 2.57130653e-01 1.98173717e-01
-2.29547068e-01 -5.43964386e-01 1.08359873e+00 -1.88671529e-01
-1.13048768e+00 4.19637859e-01 -5.28604031e-01 9.90620553e-02
1.00410724e+00 -4.11461353e-01 2.59378910e-01 1.60278548e-02
-8.27991605e-01 -7.91522339e-02 5.43680429e-01 3.62881750e-01
5.46009183e-01 -8.24159265e-01 -1.39104450e+00 2.16507077e-01
4.06466164e-02 -4.08709198e-01 8.76194954e-01 3.22071642e-01
-1.23089480e+00 1.15891613e-01 -6.84197187e-01 -1.17195666e+00
-1.97045851e+00 6.54510319e-01 4.56316292e-01 1.75922200e-01
-7.90975571e-01 8.70744467e-01 -2.88369209e-01 8.50669295e-02
-5.77473819e-01 -2.33564854e-01 -8.01501051e-02 -1.46226853e-01
7.55970627e-02 5.71648002e-01 -2.16477007e-01 -9.71793711e-01
6.68087527e-02 7.23781884e-01 -4.84728068e-02 -5.52081347e-01
1.13663948e+00 1.61112711e-01 -3.45104188e-01 1.68860137e-01
1.34163558e+00 1.49498135e-01 -1.54421997e+00 5.65071642e-01
-3.61846119e-01 -4.58184391e-01 -3.03896606e-01 -8.43605697e-01
-1.12701130e+00 6.90066755e-01 1.00941062e+00 1.61449730e-01
1.09309626e+00 -2.97832012e-01 1.00804172e-01 5.19670963e-01
3.69463325e-01 -9.99972522e-01 -1.40775561e-01 1.88590273e-01
1.32758367e+00 -1.16346920e+00 1.12966578e-02 -2.14968860e-01
-6.34606004e-01 1.20947456e+00 -8.11486393e-02 5.75861335e-03
6.69380188e-01 -1.23399243e-01 4.46313262e-01 -1.73958838e-01
-1.98845163e-01 -6.42964125e-01 6.27625942e-01 1.21385537e-01
4.20146942e-01 -2.79849380e-01 -1.16697454e+00 -1.69539049e-01
-6.03418827e-01 4.61519271e-01 3.66523594e-01 7.64746666e-01
-4.84312177e-02 -1.22536278e+00 -5.50557911e-01 2.46130958e-01
-1.58659279e-01 1.94466278e-01 -3.91804576e-01 6.23814285e-01
6.28267825e-01 7.17367887e-01 1.44193713e-02 -2.74162650e-01
7.57322073e-01 -2.42165431e-01 7.08961487e-01 -4.58096772e-01
-8.12666297e-01 2.81646222e-01 -2.10488006e-01 -2.92247593e-01
-9.73289311e-01 -7.23296523e-01 -5.51506698e-01 -4.91245985e-01
-1.69142261e-01 3.69113177e-01 8.69556904e-01 1.11901712e+00
1.79556966e-01 6.24905884e-01 2.72856861e-01 -8.85038555e-01
-2.31785119e-01 -6.95295215e-01 -7.05285847e-01 5.90635955e-01
-1.23238392e-01 -5.88303268e-01 7.63532370e-02 1.10546574e-01] | [14.103784561157227, 13.31803035736084] |
35916c05-abae-4823-a8d0-655b493a6092 | a-multi-domain-framework-for-textual | null | null | https://aclanthology.org/L18-1432 | https://aclanthology.org/L18-1432.pdf | A Multi-Domain Framework for Textual Similarity. A Case Study on Question-to-Question and Question-Answering Similarity Tasks | null | ['Hern', 'Amir Hazem', 'Basma El Amal Boussaha', 'Nicolas ez'] | 2018-05-01 | a-multi-domain-framework-for-textual-1 | https://aclanthology.org/L18-1432 | https://aclanthology.org/L18-1432.pdf | lrec-2018-5 | ['question-similarity'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.185973167419434, 3.7379775047302246] |
2764d4b5-e911-44cc-9361-1957c3acb996 | a-simple-strategy-to-provable-invariance-via | 2209.11916 | null | https://arxiv.org/abs/2209.11916v1 | https://arxiv.org/pdf/2209.11916v1.pdf | A Simple Strategy to Provable Invariance via Orbit Mapping | Many applications require robustness, or ideally invariance, of neural networks to certain transformations of input data. Most commonly, this requirement is addressed by training data augmentation, using adversarial training, or defining network architectures that include the desired invariance by design. In this work, we propose a method to make network architectures provably invariant with respect to group actions by choosing one element from a (possibly continuous) orbit based on a fixed criterion. In a nutshell, we intend to 'undo' any possible transformation before feeding the data into the actual network. Further, we empirically analyze the properties of different approaches which incorporate invariance via training or architecture, and demonstrate the advantages of our method in terms of robustness and computational efficiency. In particular, we investigate the robustness with respect to rotations of images (which can hold up to discretization artifacts) as well as the provable orientation and scaling invariance of 3D point cloud classification. | ['Michael Moeller', 'Adam Czapliński', 'Zorah Lähner', 'Jonas Geiping', 'Kanchana Vaishnavi Gandikota'] | 2022-09-24 | null | null | null | null | ['3d-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision'] | [ 4.61527348e-01 3.17342103e-01 -2.30204463e-02 -3.44299287e-01
1.18828215e-01 -8.97453725e-01 8.04550529e-01 -1.49228290e-01
-4.00637001e-01 5.03504217e-01 -7.08879456e-02 -4.23447669e-01
-3.53108406e-01 -9.12586689e-01 -1.14030910e+00 -7.87363589e-01
-2.18505815e-01 2.92747498e-01 -4.37566824e-03 -4.10043061e-01
1.45837843e-01 1.48615873e+00 -1.50768876e+00 -2.02848554e-01
5.99333942e-01 8.22802365e-01 -6.53945804e-01 5.05575418e-01
4.44561630e-01 5.97357571e-01 -5.20997703e-01 -8.38366523e-02
7.40877748e-01 -8.19907114e-02 -8.74318063e-01 1.54962137e-01
5.52826107e-01 -1.66707054e-01 -3.10839415e-01 1.22432327e+00
1.70208558e-01 2.75671750e-01 6.88871861e-01 -1.59747720e+00
-8.02155197e-01 2.95638829e-01 1.03623085e-01 -5.32050692e-02
6.81022182e-02 8.93705245e-03 6.24931931e-01 -5.86193323e-01
7.39971399e-01 9.80864286e-01 7.41305649e-01 7.67676115e-01
-1.41747308e+00 -3.58693272e-01 1.03597865e-01 -1.64199486e-01
-1.38593245e+00 -5.66807568e-01 1.06240284e+00 -3.58487725e-01
7.01977015e-01 5.83115816e-01 3.78251970e-01 9.99680221e-01
1.80814937e-01 3.68283242e-01 7.73933172e-01 -4.97699827e-01
4.61152017e-01 1.05676673e-01 6.14403039e-02 4.95221436e-01
4.95023876e-01 1.55794904e-01 1.31150231e-01 -6.00551106e-02
9.23728526e-01 -1.29933551e-01 -4.02160972e-01 -8.96461666e-01
-1.07277334e+00 6.87545002e-01 8.19559634e-01 3.33440006e-01
-2.00141683e-01 1.23338990e-01 2.70054072e-01 7.13523865e-01
3.24432999e-01 9.38402891e-01 -5.10130584e-01 5.33481836e-01
-5.27082920e-01 1.31253123e-01 6.67080581e-01 1.09229934e+00
6.55049443e-01 4.32729959e-01 2.98779160e-01 2.54598767e-01
1.25045598e-01 2.05848470e-01 5.01060545e-01 -8.89819801e-01
2.38404855e-01 5.04522622e-01 -4.02881801e-02 -1.24836528e+00
-6.35564387e-01 -4.50406343e-01 -1.16414046e+00 6.02193952e-01
3.89740497e-01 -5.29551394e-02 -9.55390930e-01 2.20440865e+00
4.12128031e-01 1.89942494e-02 2.59874105e-01 6.06021404e-01
3.71347994e-01 3.03132564e-01 -3.18819553e-01 -1.88526258e-01
8.20951760e-01 -4.77387726e-01 -3.81441861e-01 -4.57777865e-02
5.11134803e-01 -3.92866254e-01 1.06035018e+00 1.10051572e-01
-1.01808298e+00 -6.86468422e-01 -1.44669390e+00 1.51934713e-01
-5.67679882e-01 -1.48439139e-01 4.15311694e-01 8.17602992e-01
-1.10904992e+00 9.92422163e-01 -9.50191438e-01 -3.17539662e-01
2.27191970e-01 8.75318766e-01 -6.97305679e-01 2.82075107e-01
-1.17624390e+00 9.31457996e-01 5.90836346e-01 8.28634109e-03
-4.25992429e-01 -6.42776072e-01 -8.79069746e-01 -4.98387776e-02
2.12720722e-01 -8.06501925e-01 9.79635894e-01 -1.13290310e+00
-1.42799294e+00 6.75697088e-01 2.43759781e-01 -6.72983348e-01
6.80635870e-01 -6.20354228e-02 -3.13207597e-01 1.12134956e-01
-2.92786092e-01 7.70101249e-01 1.24814022e+00 -1.15808260e+00
6.60384670e-02 -3.58979285e-01 5.02564132e-01 5.14889285e-02
-6.40723526e-01 -2.10446015e-01 3.58714908e-03 -8.29727411e-01
3.10642183e-01 -1.34033644e+00 -3.33448172e-01 2.43256748e-01
-5.21047473e-01 -1.16898224e-01 1.02362061e+00 -1.32911026e-01
5.68308592e-01 -2.27327061e+00 2.78988212e-01 6.63547218e-01
-1.38889104e-02 2.71337956e-01 -3.26818317e-01 1.54025480e-01
-6.82562709e-01 4.92139578e-01 -2.51965106e-01 -2.36458331e-01
4.72649746e-02 5.40065289e-01 -5.25820136e-01 9.13987100e-01
6.35423660e-01 7.45136559e-01 -5.10028780e-01 -4.45142277e-02
2.96440333e-01 4.74633008e-01 -7.36461997e-01 6.87002167e-02
-1.03493720e-01 6.27225578e-01 -4.53207552e-01 4.11640376e-01
5.22749066e-01 5.72148003e-02 -2.63230428e-02 -8.01059827e-02
1.14788212e-01 2.40778327e-01 -1.22022688e+00 1.32664847e+00
-3.93633068e-01 5.10505855e-01 -1.63867529e-02 -1.26401162e+00
9.02084827e-01 2.05926210e-01 4.79475826e-01 -1.27620384e-01
3.62603664e-01 9.13098380e-02 1.98056772e-01 1.98088661e-02
4.99282509e-01 7.92675018e-02 2.70251185e-02 4.77219999e-01
2.33795196e-02 -1.34864077e-01 -1.36742890e-01 -2.44025320e-01
8.46459866e-01 5.41484915e-02 1.96196377e-01 -4.70280856e-01
6.23212457e-01 -1.36377677e-01 3.17625791e-01 8.05124640e-01
1.43358402e-03 5.79359710e-01 4.54761505e-01 -5.70857882e-01
-1.35984552e+00 -9.12935853e-01 -3.39470416e-01 5.93375742e-01
-2.60179173e-02 1.53579235e-01 -8.63015890e-01 -7.03324974e-01
-3.28074098e-02 6.52761102e-01 -1.03890705e+00 -7.08682358e-01
-8.37218404e-01 -4.59111273e-01 8.19655955e-01 6.13296568e-01
5.43872774e-01 -1.00717676e+00 -5.68630457e-01 -8.29780549e-02
3.64607036e-01 -1.19340563e+00 -1.56551421e-01 3.87496620e-01
-1.10245621e+00 -9.11989152e-01 -4.39480275e-01 -7.15729952e-01
1.15807784e+00 8.62914473e-02 6.50453925e-01 1.72132894e-01
2.44139865e-01 3.96604598e-01 -1.69771895e-01 -3.49979520e-01
-6.12438023e-01 3.47721338e-01 7.59310722e-01 4.77970466e-02
-1.74737111e-01 -9.58972812e-01 -1.17620990e-01 4.52731133e-01
-1.41226029e+00 -1.51333913e-01 3.23476464e-01 6.43823802e-01
4.94702220e-01 2.79696077e-01 4.42232013e-01 -6.02964938e-01
5.93434036e-01 -5.27993143e-02 -5.33344567e-01 1.28841311e-01
-5.41784942e-01 3.79412472e-01 1.08664978e+00 -6.24254048e-01
-4.57342595e-01 1.68696165e-01 -2.53114074e-01 -8.70004177e-01
-3.69390160e-01 2.86121964e-01 -3.63226682e-01 -7.55608380e-01
1.02684808e+00 1.58048585e-01 3.04255426e-01 -1.72773123e-01
5.37712753e-01 1.25697315e-01 6.38082623e-01 -7.77260065e-01
1.46216738e+00 6.46056831e-01 5.50138295e-01 -8.44578028e-01
-5.35537362e-01 1.05861038e-01 -1.03547084e+00 -7.97670931e-02
5.64794779e-01 -3.66982192e-01 -5.14029860e-01 4.46208000e-01
-1.08632171e+00 -1.35110840e-01 -6.89568222e-01 3.35504562e-01
-7.80246973e-01 2.49898553e-01 -3.64249438e-01 -3.26548636e-01
-2.22240821e-01 -1.09447002e+00 3.93592596e-01 -7.62576461e-02
-1.85961902e-01 -1.02323794e+00 -9.74580124e-02 -3.84829581e-01
4.12595093e-01 5.39659858e-01 9.10514414e-01 -1.05358493e+00
-4.33131069e-01 -5.04465401e-01 9.44605172e-02 7.47182488e-01
2.05414861e-01 1.00490034e-01 -8.51474524e-01 -5.05195618e-01
3.42479557e-01 -1.17528871e-01 5.12844920e-01 1.40583470e-01
1.21186745e+00 -7.54988611e-01 1.25924228e-02 8.95450473e-01
1.16504192e+00 2.64263581e-02 6.16939068e-01 5.99464834e-01
7.01355398e-01 5.84008038e-01 1.91555575e-01 -1.39438763e-01
-4.12695289e-01 4.84603703e-01 7.10751712e-01 5.61098568e-02
2.59219319e-01 -1.28366813e-01 1.54236183e-01 6.17720902e-01
-4.72279817e-01 -7.68263787e-02 -7.36073196e-01 2.64891326e-01
-1.45576513e+00 -7.97723949e-01 2.11588532e-01 2.14332294e+00
4.56931651e-01 4.18796450e-01 -9.07578319e-02 5.53839624e-01
7.21720099e-01 2.07301930e-01 -5.19442797e-01 -5.15531957e-01
-1.58051118e-01 2.90608257e-01 7.39790618e-01 3.08585525e-01
-1.28065741e+00 7.73046553e-01 6.52756929e+00 3.71590257e-01
-1.42097020e+00 -2.34762922e-01 2.52908170e-01 2.08437860e-01
-2.00266242e-01 -1.22417137e-01 -4.25473362e-01 4.01334465e-03
6.44475520e-01 4.80967201e-02 5.07577896e-01 9.00057673e-01
-1.37619376e-01 7.99139023e-01 -1.31459367e+00 2.33853579e-01
1.31315753e-01 -1.44267547e+00 2.95881659e-01 2.34736606e-01
8.23039651e-01 -1.16569310e-01 2.15538293e-01 5.26802940e-03
1.32374778e-01 -9.80898857e-01 7.31057227e-01 4.67066258e-01
6.67972922e-01 -9.56405580e-01 4.96062011e-01 4.41907197e-01
-7.23587811e-01 9.28034931e-02 -3.12160611e-01 -1.51737601e-01
-2.89522797e-01 1.55079097e-01 -7.75312006e-01 6.53757036e-01
3.68506461e-01 5.82031965e-01 -5.32191634e-01 5.99172950e-01
-3.56765538e-01 2.91654140e-01 -6.42550290e-01 5.76593168e-02
3.28167558e-01 -2.10040018e-01 8.10954809e-01 7.50781059e-01
2.61848688e-01 -4.25003171e-02 -6.87933415e-02 8.45788836e-01
-2.70762295e-01 -1.56454310e-01 -1.01554751e+00 7.98584744e-02
3.95225763e-01 9.98089969e-01 -8.20416808e-01 -8.12070891e-02
-5.12745678e-02 7.24766314e-01 1.53109565e-01 3.82968932e-01
-6.32170618e-01 -4.06604886e-01 8.01859260e-01 -4.08574753e-02
2.87824750e-01 -4.11773860e-01 -4.19752717e-01 -1.09204340e+00
3.08360070e-01 -8.51489365e-01 -3.58287953e-02 -7.38082051e-01
-9.83733714e-01 7.75750279e-01 -2.18232814e-02 -1.46626318e+00
-2.95308739e-01 -9.29433525e-01 -6.75086796e-01 4.87573266e-01
-1.25626755e+00 -1.22313571e+00 -4.96876836e-02 7.80814230e-01
-5.14730401e-02 -2.18202949e-01 8.48124266e-01 1.23854324e-01
-4.51840669e-01 8.74055624e-01 -9.56985578e-02 3.05247664e-01
2.43777260e-01 -1.06197298e+00 6.38240933e-01 1.04592156e+00
2.79974520e-01 9.28070962e-01 9.34415638e-01 -3.05954933e-01
-1.25375688e+00 -1.40119064e+00 6.06948614e-01 -5.02908170e-01
7.02549756e-01 -2.28625000e-01 -1.08782017e+00 1.11494648e+00
-1.00260943e-01 1.81128874e-01 2.91845381e-01 3.04054171e-02
-4.23770159e-01 -6.65120333e-02 -1.29448831e+00 9.16604280e-01
1.06358087e+00 -6.07905149e-01 -8.59631240e-01 2.98680574e-01
9.38246667e-01 -6.35306239e-01 -1.08023036e+00 7.78824031e-01
4.57949787e-01 -6.65845871e-01 1.23915052e+00 -1.19163644e+00
2.54150540e-01 -4.08370167e-01 -2.20516369e-01 -1.30428910e+00
-3.34681809e-01 -6.84412539e-01 -1.04045030e-02 1.01811588e+00
3.66120428e-01 -8.62775028e-01 8.14390898e-01 7.00006068e-01
-3.45573008e-01 -5.32455802e-01 -1.09154296e+00 -1.06844389e+00
4.60490227e-01 -5.80736578e-01 9.58501518e-01 1.42450583e+00
-2.74413884e-01 2.71142088e-03 -2.89610773e-01 6.49478734e-01
1.19095370e-01 -1.23577602e-01 8.85013103e-01 -1.16968405e+00
-3.50673497e-02 -4.74591196e-01 -9.24536049e-01 -7.89903641e-01
3.54735076e-01 -9.53403831e-01 -1.74285024e-01 -9.42927957e-01
-5.89276075e-01 -5.26500940e-01 -3.54979336e-01 7.02309132e-01
4.55231041e-01 4.06231523e-01 2.42260441e-01 4.24432725e-01
-1.78635176e-02 4.30997878e-01 1.26394784e+00 -1.81156071e-03
-3.54668081e-01 2.82349735e-01 -5.50830185e-01 9.55612183e-01
1.22568488e+00 -4.18292552e-01 -4.50441241e-01 -4.60270941e-01
3.54460835e-01 -3.24576944e-01 6.69900775e-01 -1.15195107e+00
1.89266622e-01 -2.15891257e-01 3.72612417e-01 -2.22632587e-01
3.40420276e-01 -1.09821784e+00 2.46129647e-01 5.88571012e-01
-5.03677428e-01 2.27116406e-01 3.78482610e-01 3.09202641e-01
-5.85531117e-04 -4.15328562e-01 8.57312739e-01 2.11141139e-01
-3.41379374e-01 2.53722817e-01 -2.24102899e-01 -3.06843549e-01
1.00008106e+00 -4.07477498e-01 -1.85968861e-01 -1.91632032e-01
-8.94364595e-01 -1.49034768e-01 7.37583876e-01 4.69036013e-01
3.29602271e-01 -1.50873971e+00 -2.73717374e-01 6.44068778e-01
-6.76999381e-03 1.62793756e-01 -2.06795245e-01 6.77959800e-01
-5.53069413e-01 4.12595212e-01 -5.21972775e-01 -5.67969918e-01
-1.09897912e+00 8.42759609e-01 6.77899301e-01 3.22631747e-03
-5.91276348e-01 3.38969052e-01 -2.06312120e-01 -8.13407421e-01
1.79705441e-01 -5.78726232e-01 -1.31169572e-01 -3.47537816e-01
4.70631160e-02 2.16962341e-02 3.39098364e-01 -5.49364686e-01
-3.46599281e-01 4.63561922e-01 1.68362364e-01 5.47086895e-02
1.18093538e+00 3.87870163e-01 -5.17985597e-02 1.26405343e-01
1.31444359e+00 -3.37780602e-02 -1.04970849e+00 -1.81141376e-01
-2.13800877e-01 -6.87749386e-02 -2.05238700e-01 -1.79203659e-01
-1.20129418e+00 6.40536487e-01 3.89861375e-01 6.73077524e-01
1.07575548e+00 -2.67212629e-01 1.85731903e-01 9.51867759e-01
8.34164172e-02 -7.09273100e-01 -2.49366567e-01 6.52395129e-01
1.03699207e+00 -7.69009948e-01 -1.09164976e-02 -3.01946133e-01
-1.02952771e-01 1.23201406e+00 4.76225644e-01 -7.86058307e-01
8.07161748e-01 7.46355727e-02 -4.60969806e-02 1.27132788e-01
-3.77828836e-01 1.37600735e-01 5.81378639e-01 6.25315964e-01
-6.82027489e-02 -2.67452717e-01 5.72315557e-03 4.22589928e-02
-5.93111038e-01 -2.47476667e-01 5.30229032e-01 9.83974218e-01
-2.51965255e-01 -1.02866530e+00 -5.40091157e-01 1.32648692e-01
-3.64422798e-01 2.23833978e-01 -5.83872199e-01 1.34482646e+00
1.37914598e-01 4.30105567e-01 1.33468360e-01 -4.51157063e-01
4.01605785e-01 3.33082117e-02 4.92685348e-01 -2.80449450e-01
-3.96625549e-01 -3.64750713e-01 -9.48769078e-02 -2.64247060e-01
-6.97193205e-01 -6.90504611e-01 -9.86709774e-01 -3.08484852e-01
-2.61489242e-01 -5.84658124e-02 6.18100703e-01 1.06457746e+00
2.85753131e-01 6.58071697e-01 8.34472656e-01 -9.72589254e-01
-9.33850944e-01 -7.79046893e-01 -2.56688148e-01 4.46083248e-01
4.93879378e-01 -6.19317710e-01 -5.88120997e-01 3.97793390e-02] | [8.863134384155273, 2.5939319133758545] |
bcdb092a-9e45-40cd-b6c5-0c39d3417696 | multi-source-diffusion-models-for | 2302.02257 | null | https://arxiv.org/abs/2302.02257v3 | https://arxiv.org/pdf/2302.02257v3.pdf | Multi-Source Diffusion Models for Simultaneous Music Generation and Separation | In this work, we define a diffusion-based generative model capable of both music synthesis and source separation by learning the score of the joint probability density of sources sharing a context. Alongside the classic total inference tasks (i.e., generating a mixture, separating the sources), we also introduce and experiment on the partial generation task of source imputation, where we generate a subset of the sources given the others (e.g., play a piano track that goes well with the drums). Additionally, we introduce a novel inference method for the separation task based on Dirac likelihood functions. We train our model on Slakh2100, a standard dataset for musical source separation, provide qualitative results in the generation settings, and showcase competitive quantitative results in the source separation setting. Our method is the first example of a single model that can handle both generation and separation tasks, thus representing a step toward general audio models. | ['Emanuele Rodolà', 'Luca Cosmo', 'Michele Mancusi', 'Emilian Postolache', 'Irene Tallini', 'Giorgio Mariani'] | 2023-02-04 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 3.44906718e-01 1.28778443e-01 8.89262557e-03 1.53699249e-01
-1.34706414e+00 -8.38392317e-01 9.92976248e-01 -2.10337430e-01
4.94522676e-02 7.38666713e-01 6.63199484e-01 1.19945139e-01
-3.36661249e-01 -6.10770583e-01 -7.75327802e-01 -9.77284253e-01
-1.94904059e-01 7.82402337e-01 -6.56695217e-02 -5.92212565e-02
-1.13175824e-01 1.14157297e-01 -1.36278558e+00 3.05519342e-01
7.28283584e-01 6.80733204e-01 3.66358683e-02 1.01405144e+00
1.22703135e-01 9.29658830e-01 -9.59792376e-01 -3.77933532e-01
1.67477354e-01 -1.00483656e+00 -5.87045252e-01 4.29416001e-02
3.39809239e-01 -5.31246811e-02 -3.05654965e-02 8.11228395e-01
6.11798108e-01 1.18021086e-01 9.94404495e-01 -1.43894756e+00
-3.99501562e-01 1.34051883e+00 -4.53403711e-01 -1.36852592e-01
3.74332041e-01 -5.13564944e-02 1.17853141e+00 -7.32358634e-01
5.61608195e-01 1.06028128e+00 7.43176520e-01 3.25928152e-01
-1.82417583e+00 -5.87304890e-01 -2.24795446e-01 -1.78303465e-01
-1.35365415e+00 -9.19726789e-01 8.70669305e-01 -5.69855392e-01
4.41956967e-01 2.96116352e-01 5.70032716e-01 1.56693757e+00
-1.73374340e-01 9.71391976e-01 9.14852202e-01 -4.54174876e-01
3.05922866e-01 -1.22471884e-01 -1.26113519e-01 1.21085756e-01
-6.40958697e-02 1.41070783e-01 -1.11408591e+00 -5.35701156e-01
6.29742444e-01 -5.07501304e-01 -2.90623337e-01 -1.70813367e-01
-1.41098332e+00 6.04469121e-01 -6.23496510e-02 8.99870247e-02
-2.98767805e-01 4.62782800e-01 -2.37428434e-02 1.73069574e-02
4.40492570e-01 3.42008203e-01 -1.08546518e-01 -2.74872720e-01
-1.56488883e+00 8.04177523e-01 1.20704639e+00 1.13441086e+00
4.23971087e-01 3.38100195e-01 -5.15585244e-01 8.31730247e-01
3.45541924e-01 5.18149495e-01 2.24217072e-01 -1.28871393e+00
1.95576206e-01 -2.39343002e-01 3.15385938e-01 -2.88731903e-01
-1.94556993e-02 -7.00624168e-01 -7.12929547e-01 3.14872891e-01
8.67805183e-01 -4.65255737e-01 -8.08372676e-01 2.10727382e+00
-1.29387798e-02 9.51715469e-01 9.85591933e-02 6.59581006e-01
5.86437643e-01 5.50223053e-01 -2.52085030e-01 -3.11371028e-01
1.18107760e+00 -8.31984580e-01 -6.04016185e-01 -4.23958339e-03
-6.38557747e-02 -1.06242168e+00 7.07629204e-01 8.73684347e-01
-1.52534890e+00 -4.60942626e-01 -9.32822347e-01 8.45692828e-02
3.08400810e-01 1.86602503e-01 6.65608346e-01 4.54578161e-01
-9.19203639e-01 9.89178717e-01 -9.29933012e-01 2.25991666e-01
2.67048955e-01 -1.72769397e-01 5.64467646e-02 2.71788359e-01
-9.36030090e-01 3.07521850e-01 1.37619853e-01 -1.22662589e-01
-1.25104439e+00 -9.68113363e-01 -4.46307451e-01 6.32707104e-02
2.00604364e-01 -8.77206743e-01 1.40352988e+00 -8.89663041e-01
-1.60018671e+00 5.40048540e-01 -1.15907833e-01 -5.55467546e-01
5.06286025e-01 -3.40346277e-01 -4.15280014e-01 -1.65532470e-01
8.59641135e-02 4.65166986e-01 1.11958730e+00 -1.39875519e+00
-4.95391995e-01 7.42709450e-03 -3.69315296e-01 3.19903046e-02
3.22939515e-01 -1.75656155e-01 -3.20488751e-01 -1.09712255e+00
-2.01621000e-02 -1.02171457e+00 2.13074014e-02 -4.52983499e-01
-9.83616054e-01 5.06112911e-02 1.34080142e-01 -7.78035223e-01
1.10546649e+00 -2.20870781e+00 6.48474932e-01 2.21544638e-01
3.57545316e-02 -2.17686087e-01 -3.05039715e-02 5.83346605e-01
-1.50565729e-01 3.92146781e-02 -4.66325313e-01 -1.05233908e+00
4.45531011e-01 1.07606435e-02 -8.23241889e-01 1.97156042e-01
1.69608921e-01 7.40348458e-01 -8.79013777e-01 -1.69632643e-01
-3.14940661e-01 6.19214416e-01 -6.87638223e-01 2.45342986e-03
-5.08642972e-01 5.82967758e-01 4.14036438e-02 3.91162604e-01
4.61073935e-01 2.35886853e-02 6.88285381e-02 -1.72845647e-02
4.33241241e-02 5.86425483e-01 -1.63356042e+00 2.22711372e+00
-3.85355115e-01 5.16759634e-01 3.96705091e-01 -2.38689199e-01
6.21567726e-01 5.18046319e-01 3.75748903e-01 1.55925214e-01
-1.52602136e-01 3.34636718e-01 9.74725857e-02 3.66521478e-02
6.49801433e-01 -4.99141842e-01 -2.00984567e-01 7.46983826e-01
5.78334808e-01 -3.54492426e-01 1.57499880e-01 4.97395575e-01
1.00137520e+00 5.92723012e-01 6.61311150e-02 -1.16964653e-01
-2.46968806e-01 -3.02542925e-01 2.75061309e-01 1.10815465e+00
3.88475150e-01 1.14634049e+00 6.27835989e-01 4.82543588e-01
-9.88972902e-01 -1.67275500e+00 -2.64880676e-02 8.30156803e-01
-2.61075258e-01 -7.46011019e-01 -7.38257289e-01 -1.84629872e-01
-6.17839172e-02 1.04566967e+00 -2.65463322e-01 -1.15019567e-01
-3.80917400e-01 -8.71343851e-01 9.84782338e-01 3.46303970e-01
-7.24273697e-02 -9.33218420e-01 -1.45469904e-01 2.78752834e-01
-5.64085960e-01 -7.05381989e-01 -4.27776515e-01 3.88276488e-01
-5.52932620e-01 -7.25494206e-01 -6.68551803e-01 -4.97303814e-01
-2.82876968e-01 -2.72547275e-01 1.47254610e+00 -2.87757933e-01
-1.46532714e-01 2.99749017e-01 -7.49905705e-02 -5.51411748e-01
-6.71637774e-01 1.38760239e-01 1.20128668e-03 1.63654953e-01
-1.59774765e-01 -1.13664389e+00 -3.69291693e-01 -2.13593975e-01
-8.01947534e-01 9.50614735e-02 2.61764318e-01 6.01759076e-01
5.08127570e-01 1.70739621e-01 9.45489585e-01 -6.86066508e-01
8.39440465e-01 -8.87157023e-01 -1.90815359e-01 -9.97442156e-02
-2.73392737e-01 4.82375100e-02 3.77889425e-01 -4.73231524e-01
-1.19320786e+00 -7.09797349e-03 -1.65111139e-01 -3.67370278e-01
-3.17112595e-01 3.23000163e-01 -3.65114003e-01 6.52761519e-01
8.21119249e-01 2.19778106e-01 -2.35302135e-01 -8.69955778e-01
7.23068297e-01 4.29884315e-01 9.94481385e-01 -7.70301521e-01
8.69380593e-01 5.08989692e-01 5.04493862e-02 -7.03624070e-01
-8.39202821e-01 -2.13623717e-01 -5.31497240e-01 7.05853030e-02
8.02674651e-01 -1.15577555e+00 -4.32352155e-01 6.37599707e-01
-1.14165711e+00 -5.73600471e-01 -8.46481562e-01 5.18719435e-01
-9.68491971e-01 1.09898247e-01 -6.85517251e-01 -1.07707429e+00
1.87485918e-01 -7.04640627e-01 1.35105085e+00 6.70487341e-03
-5.11762500e-01 -1.03995240e+00 5.23149788e-01 -7.31191859e-02
1.41108483e-01 3.69286597e-01 7.62373805e-01 -6.44292355e-01
-7.60749459e-01 1.09773219e-01 4.09693986e-01 4.02511179e-01
1.14960499e-01 4.51792032e-02 -1.39823258e+00 1.45213738e-01
1.22427948e-01 -2.35119581e-01 1.31011581e+00 4.68032539e-01
6.45702720e-01 -2.86122769e-01 -2.49428973e-01 6.92202508e-01
1.01898289e+00 -1.55367553e-01 6.34213388e-01 -3.48941565e-01
6.52412891e-01 3.92417014e-01 6.29880577e-02 6.12952888e-01
1.64638057e-01 8.36471975e-01 7.66447037e-02 4.70908619e-02
-7.30486929e-01 -7.31164932e-01 4.82444644e-01 6.94009721e-01
-1.92331493e-01 -5.76784015e-01 -4.73802418e-01 7.13179410e-01
-1.80137980e+00 -1.35401416e+00 -2.67573714e-01 2.34160471e+00
1.28547609e+00 1.28838688e-01 5.35035491e-01 4.08240497e-01
4.48136032e-01 1.57841906e-01 -3.32800359e-01 1.77280381e-01
-2.58470237e-01 7.02859998e-01 3.84734869e-02 8.53251398e-01
-1.04152727e+00 6.55438602e-01 7.20294762e+00 1.22283196e+00
-6.26691878e-01 1.82058632e-01 2.01984078e-01 -6.45610869e-01
-6.76720142e-01 1.54573783e-01 -7.59977698e-01 6.28470540e-01
1.03965640e+00 -5.13578318e-02 8.63860130e-01 2.61387229e-01
-4.49321717e-02 -8.22528899e-02 -1.38902044e+00 7.49651253e-01
1.20229445e-01 -1.19420457e+00 -1.06828064e-01 5.03219776e-02
7.67730296e-01 -1.42353818e-01 9.15404931e-02 1.94185212e-01
7.53356278e-01 -1.16997623e+00 1.31036186e+00 9.46267664e-01
6.06047988e-01 -6.83600366e-01 1.31337985e-01 6.30654097e-01
-8.31391513e-01 2.29084119e-01 1.00961633e-01 -6.21085353e-02
6.75006390e-01 8.70368838e-01 -6.85042620e-01 7.38904893e-01
3.34890127e-01 6.41079307e-01 -2.56265134e-01 1.10584748e+00
-4.72927570e-01 1.23610485e+00 -5.21743059e-01 5.97368062e-01
-3.25366825e-01 -3.94841701e-01 1.21255481e+00 1.21101964e+00
6.05955005e-01 -3.63468677e-01 6.83428273e-02 1.51901376e+00
-9.05545354e-02 -3.86385411e-01 -4.78991777e-01 -8.47617984e-02
4.83374119e-01 1.24308097e+00 -7.11523712e-01 -2.29639962e-01
1.23371959e-01 1.00277865e+00 9.95842218e-02 5.28583348e-01
-9.73769546e-01 -3.49778086e-01 6.27851367e-01 1.18567891e-01
3.76837671e-01 -2.04717666e-01 -3.25516522e-01 -1.26650262e+00
-2.51741171e-01 -8.91084731e-01 4.14400697e-02 -9.53061461e-01
-1.40558481e+00 3.15542072e-01 1.83893606e-01 -1.03247321e+00
-8.32834601e-01 -7.18452260e-02 -7.94769287e-01 1.27882934e+00
-1.12264311e+00 -1.06066418e+00 8.40152428e-02 4.78667557e-01
2.73841619e-01 -8.59816745e-02 8.38548303e-01 1.82782367e-01
-1.39786139e-01 2.37808302e-01 1.11274451e-01 -2.02401802e-01
8.78096700e-01 -1.61801136e+00 5.57829320e-01 8.56445789e-01
9.95370030e-01 6.43246412e-01 8.77695501e-01 -6.68687940e-01
-9.08025920e-01 -8.81454468e-01 8.48679185e-01 -8.76164734e-01
6.96826637e-01 -8.15006852e-01 -6.27281547e-01 8.01372349e-01
3.65829706e-01 -4.56847340e-01 8.82536530e-01 2.74500847e-01
-2.93595374e-01 2.14135811e-01 -7.62722552e-01 6.61878705e-01
1.33267689e+00 -5.05395055e-01 -4.81715083e-01 2.72577047e-01
7.13420093e-01 -3.74302387e-01 -6.12822831e-01 -4.13444042e-02
5.45809150e-01 -1.26764762e+00 1.27931774e+00 -3.98604214e-01
6.21225953e-01 -4.65991706e-01 -1.94343448e-01 -1.76522505e+00
-3.41475517e-01 -1.18545783e+00 -3.93709362e-01 1.86649001e+00
4.43068564e-01 -9.89753976e-02 4.28152680e-01 2.03449335e-02
-2.32136562e-01 -1.21177286e-01 -8.93942058e-01 -7.87144065e-01
1.78963631e-01 -7.09814250e-01 6.36483848e-01 5.33896923e-01
-2.49207094e-01 6.26046479e-01 -6.77646458e-01 8.07231851e-03
9.84648526e-01 3.87472093e-01 6.23468220e-01 -1.24309337e+00
-1.22366154e+00 -6.71704292e-01 1.20918378e-01 -1.02815676e+00
3.06409597e-01 -1.19402730e+00 2.68632889e-01 -1.35156190e+00
1.60135135e-01 -5.43938279e-01 -9.89731476e-02 1.81154013e-01
-1.78869754e-01 4.36743349e-01 3.45501125e-01 3.45824420e-01
-3.28002840e-01 5.13933241e-01 9.24274802e-01 1.24073094e-02
-2.17599273e-01 4.25320596e-01 -1.02298594e+00 8.84490490e-01
5.31830013e-01 -6.93472445e-01 -5.81005454e-01 -1.84387982e-01
4.09129739e-01 3.76141310e-01 7.20932484e-01 -1.04944205e+00
1.66122973e-01 5.42314388e-02 3.44518065e-01 -3.63549739e-01
5.92299759e-01 -3.99020284e-01 6.43783271e-01 -1.41811846e-02
-6.06571913e-01 -5.76623738e-01 5.06860577e-02 7.68468380e-01
-1.35956436e-01 -3.96375656e-01 3.58632088e-01 2.92779189e-02
8.76752660e-02 -5.69169857e-02 -3.59163493e-01 5.00639975e-01
4.17470843e-01 1.89065889e-01 -2.05024838e-01 -7.76978135e-01
-1.18470836e+00 -3.59029174e-01 2.74504185e-01 2.22193941e-01
1.87529251e-01 -1.53924382e+00 -1.03769898e+00 2.72039235e-01
-2.47041181e-01 4.79183607e-02 3.02940588e-02 7.25012124e-01
1.84126467e-01 -4.40447703e-02 2.09279880e-01 -3.68322700e-01
-9.26659584e-01 3.13149691e-01 2.69514740e-01 -2.05710694e-01
-3.70488495e-01 9.19156134e-01 2.59372443e-01 -1.60914987e-01
2.18372241e-01 -2.87205040e-01 1.32099226e-01 2.84358948e-01
4.59586561e-01 5.54898739e-01 -1.09407008e-01 -4.45203006e-01
-2.21598372e-01 6.49867654e-02 6.23682439e-01 -1.02159309e+00
1.11843264e+00 5.80552369e-02 -1.52756453e-01 1.12600422e+00
5.82941175e-01 8.70772898e-01 -1.46130860e+00 -4.60995510e-02
-1.37571648e-01 -1.49674132e-01 -1.72093093e-01 -1.13542545e+00
-6.42199755e-01 8.25496972e-01 1.22279627e-02 4.26969886e-01
8.89094293e-01 2.07825035e-01 5.40037334e-01 -2.83162683e-01
3.01693588e-01 -6.95264935e-01 5.24462126e-02 3.89424533e-01
9.42967534e-01 -6.14062071e-01 -2.37689748e-01 -1.98333696e-01
-5.77414811e-01 7.31062591e-01 -1.16017491e-01 -3.82074207e-01
6.97791755e-01 8.66137922e-01 -2.48409778e-01 1.46026136e-02
-8.62302542e-01 -1.40971348e-01 4.22994435e-01 6.86210752e-01
6.13683641e-01 2.66705036e-01 2.64522254e-01 1.29945338e+00
-7.52957821e-01 -2.01305114e-02 6.28433466e-01 5.08900940e-01
-2.03169793e-01 -1.35370350e+00 -3.95636648e-01 2.13257283e-01
-5.31283915e-01 -5.24590254e-01 -5.89850605e-01 3.95223647e-01
3.52609843e-01 1.17731810e+00 9.82596427e-02 -1.60716102e-01
1.61733091e-01 5.57652652e-01 7.03676105e-01 -7.76086628e-01
-6.08212233e-01 7.82871187e-01 3.19199741e-01 -2.11572573e-01
-4.19116288e-01 -9.07815456e-01 -1.02830517e+00 -1.46499813e-01
-3.41907173e-01 2.39355013e-01 4.19790328e-01 6.28197968e-01
2.27814898e-01 8.23335588e-01 2.07011163e-01 -1.00110292e+00
-6.06567860e-01 -1.12009442e+00 -1.05885923e+00 3.96301031e-01
4.88911867e-01 -4.08966690e-01 -5.69108009e-01 3.91803443e-01] | [15.653407096862793, 5.67043924331665] |
0c769b16-0cf7-49f0-ae68-a3fb0dea6de6 | toward-leveraging-pre-trained-self-supervised | 2306.12714 | null | https://arxiv.org/abs/2306.12714v1 | https://arxiv.org/pdf/2306.12714v1.pdf | Toward Leveraging Pre-Trained Self-Supervised Frontends for Automatic Singing Voice Understanding Tasks: Three Case Studies | Automatic singing voice understanding tasks, such as singer identification, singing voice transcription, and singing technique classification, benefit from data-driven approaches that utilize deep learning techniques. These approaches work well even under the rich diversity of vocal and noisy samples owing to their representation ability. However, the limited availability of labeled data remains a significant obstacle to achieving satisfactory performance. In recent years, self-supervised learning models (SSL models) have been trained using large amounts of unlabeled data in the field of speech processing and music classification. By fine-tuning these models for the target tasks, comparable performance to conventional supervised learning can be achieved with limited training data. Therefore, in this paper, we investigate the effectiveness of SSL models for various singing voice recognition tasks. We report the results of experiments comparing SSL models for three different tasks (i.e., singer identification, singing voice transcription, and singing technique classification) as initial exploration and aim to discuss these findings. Experimental results show that each SSL model achieves comparable performance and sometimes outperforms compared to state-of-the-art methods on each task. We also conducted a layer-wise analysis to further understand the behavior of the SSL models. | ['Yuya Yamamoto'] | 2023-06-22 | null | null | null | null | ['self-supervised-learning', 'singer-identification', 'music-classification'] | ['computer-vision', 'music', 'music'] | [ 5.54560646e-02 -4.73280847e-01 -2.78491586e-01 -3.55300933e-01
-1.03924417e+00 -5.85069895e-01 2.64713645e-01 -4.23849136e-01
-1.81603253e-01 3.58192593e-01 2.74663389e-01 -1.24989748e-01
-7.36475512e-02 -6.54145554e-02 -3.76656443e-01 -6.55507624e-01
1.85572095e-02 3.34780037e-01 -9.87014025e-02 4.31849249e-02
7.10501075e-02 2.45525002e-01 -1.88378727e+00 4.62250352e-01
6.89105213e-01 9.22627985e-01 2.50003368e-01 7.53251851e-01
-2.16418579e-01 7.18402624e-01 -7.80366004e-01 2.19180718e-01
-5.88782737e-03 -7.11630046e-01 -5.79711080e-01 1.26846418e-01
6.12861276e-01 -1.24257058e-01 -2.46542484e-01 6.25156105e-01
9.87599254e-01 3.42543811e-01 4.59700286e-01 -1.01359093e+00
-3.95799488e-01 9.39305484e-01 -8.61939266e-02 2.31836364e-01
2.00685620e-01 3.16774070e-01 1.36103642e+00 -1.22857416e+00
2.03564435e-01 1.18486500e+00 7.94101954e-01 1.05254900e+00
-1.03107882e+00 -9.10850286e-01 -1.53466746e-01 3.23342353e-01
-1.28529155e+00 -9.95228052e-01 9.62425530e-01 -3.38282436e-01
7.63824344e-01 5.00611365e-01 3.53480250e-01 9.65848863e-01
-4.96683061e-01 1.18905473e+00 1.19233191e+00 -5.13882637e-01
1.20403558e-01 1.29840970e-01 1.46000445e-01 3.70938778e-01
-4.27820325e-01 2.25270122e-01 -1.22399068e+00 -4.39585038e-02
4.13554490e-01 -5.93667924e-01 -3.40020508e-01 7.19885677e-02
-1.16604018e+00 6.24970853e-01 1.03759959e-01 6.02100968e-01
-8.20728838e-02 -1.24645360e-01 6.57398820e-01 4.39448297e-01
5.71976781e-01 7.94836044e-01 -3.97101194e-01 -4.04416591e-01
-1.35133696e+00 2.03120947e-01 9.21057880e-01 5.33717215e-01
7.20990375e-02 6.60268784e-01 -2.20622241e-01 1.52369070e+00
6.35073259e-02 1.57954916e-01 8.19594681e-01 -8.96830261e-01
4.08778876e-01 1.34483844e-01 -3.17861885e-01 -3.53249609e-01
-3.40318263e-01 -7.75677741e-01 -6.33367598e-01 -1.41390488e-01
3.70085001e-01 -1.30063772e-01 -7.65107930e-01 1.78107965e+00
-9.48860273e-02 4.03013974e-01 -1.80167612e-02 9.76001501e-01
1.28026450e+00 6.82415247e-01 -1.23653695e-01 -5.34155786e-01
8.00223053e-01 -1.26965725e+00 -1.10799384e+00 -4.32180054e-02
3.09604138e-01 -9.16997969e-01 1.50238359e+00 5.82330823e-01
-1.08775735e+00 -1.01046658e+00 -9.26294506e-01 1.01500504e-01
1.13092111e-02 4.71498042e-01 5.13120770e-01 8.32221568e-01
-8.47936332e-01 6.73845232e-01 -5.88205755e-01 -1.76823258e-01
3.64255190e-01 5.21623671e-01 -1.41909942e-01 2.80439913e-01
-1.11279762e+00 4.53410000e-01 1.65543318e-01 1.14429668e-01
-1.26317441e+00 -5.77504575e-01 -4.71251816e-01 3.00811119e-02
3.26859266e-01 -3.37199926e-01 1.65888846e+00 -6.80668652e-01
-1.77125418e+00 9.71074581e-01 -2.99918294e-01 -3.96251857e-01
2.15407945e-02 -2.62185931e-01 -5.48108637e-01 -2.07800984e-01
-2.75886714e-01 4.26559031e-01 1.14016557e+00 -1.25415242e+00
-5.01023591e-01 -3.40568781e-01 -3.63579988e-01 3.10571015e-01
-5.89748383e-01 3.15542758e-01 -1.40752673e-01 -7.24990904e-01
1.80078104e-01 -9.19685543e-01 1.76422387e-01 -6.10116899e-01
-4.13682580e-01 -4.97118831e-01 8.51344764e-01 -4.94901031e-01
1.69303155e+00 -2.57321215e+00 2.48117775e-01 -2.86659718e-01
1.16664328e-01 7.36314058e-01 -2.14551866e-01 4.97965842e-01
8.66778940e-03 1.00502307e-02 -2.56959975e-01 -6.77211106e-01
-2.59050094e-02 1.61610812e-01 -4.12206471e-01 2.08747223e-01
-5.81811555e-02 7.26694167e-01 -7.21762955e-01 -3.30412805e-01
2.23999366e-01 1.44182429e-01 -4.94072378e-01 7.04832852e-01
-7.25982040e-02 7.97715306e-01 -9.24453884e-02 8.93665254e-01
1.36404455e-01 1.91230178e-01 5.81083223e-02 -1.16604380e-01
-1.96375668e-01 7.85807073e-01 -1.09930396e+00 1.59517276e+00
-8.74017715e-01 8.28137338e-01 3.54877084e-01 -9.48774934e-01
1.03345668e+00 7.32897758e-01 3.10223341e-01 -2.06468612e-01
1.04282506e-01 6.59050286e-01 5.83612621e-01 -5.61728835e-01
3.31325918e-01 -5.38306057e-01 1.34455115e-01 5.25512457e-01
4.79981214e-01 -2.98799366e-01 -2.27185544e-02 -3.74575585e-01
8.03602695e-01 -1.84290081e-01 -9.19340476e-02 -3.05463448e-02
6.27663255e-01 -2.08531067e-01 6.17962897e-01 6.63920760e-01
-2.10352927e-01 8.34878266e-01 -3.02321352e-02 -1.66758761e-01
-6.49637878e-01 -8.86760950e-01 -3.51212293e-01 1.45588851e+00
-3.46199125e-01 -5.08904457e-01 -9.02745187e-01 -3.22447330e-01
-7.65290707e-02 5.98826408e-01 -1.78770036e-01 -8.26181546e-02
-6.90598071e-01 -6.28566921e-01 9.11720872e-01 6.37436330e-01
4.11919922e-01 -1.52889359e+00 -5.86181693e-02 3.22334617e-01
-7.65348673e-02 -1.03772414e+00 -6.00853264e-01 3.68319333e-01
-9.09560144e-01 -8.08229446e-01 -7.25351393e-01 -1.12375987e+00
1.74041670e-02 2.12487608e-01 9.54250216e-01 -2.02093506e-03
-4.60066386e-02 3.05234313e-01 -2.57092327e-01 -4.36618567e-01
-7.27269888e-01 4.41845000e-01 6.28442883e-01 2.36568615e-01
2.17722625e-01 -5.72015524e-01 -3.04664373e-01 4.26484406e-01
-5.92172325e-01 -2.97090858e-01 2.58793116e-01 1.10235548e+00
5.99754870e-01 9.44997221e-02 1.17284334e+00 -7.72623539e-01
1.11964989e+00 -1.90326452e-01 -6.08419031e-02 7.83960894e-02
-4.60593075e-01 -3.06146652e-01 6.72138214e-01 -7.07836270e-01
-7.89506197e-01 7.35040661e-03 -4.47376311e-01 -7.09145248e-01
-3.48331481e-01 5.02090573e-01 -3.04898769e-01 7.78900310e-02
5.86833477e-01 2.89118022e-01 1.34001136e-01 -1.08720481e+00
2.25883335e-01 1.52644205e+00 6.35295391e-01 -3.70167494e-01
6.08142972e-01 2.44134124e-02 -5.37546217e-01 -1.19138587e+00
-1.22973144e+00 -7.36973345e-01 -5.62999845e-01 -3.51252347e-01
5.37924707e-01 -8.63432527e-01 -5.64958096e-01 7.81757236e-01
-8.11022639e-01 -3.14172626e-01 -4.74599153e-01 7.05927551e-01
-7.15600967e-01 1.20256342e-01 -7.27800429e-01 -1.33659780e+00
-4.47445303e-01 -1.10130894e+00 9.93095279e-01 -3.76405427e-03
-5.52824020e-01 -8.49104881e-01 1.30185813e-01 7.12730050e-01
4.39522803e-01 -4.44437832e-01 9.21329975e-01 -8.90536070e-01
-8.65332633e-02 -1.98651757e-02 3.07803929e-01 9.93809700e-01
4.08579350e-01 -2.89150745e-01 -1.62623215e+00 -3.84447843e-01
2.30924129e-01 -6.82908058e-01 1.02100074e+00 4.33027536e-01
1.49036527e+00 -1.25359371e-02 2.31833503e-01 5.93925893e-01
4.77971673e-01 3.04630011e-01 1.69881359e-01 -3.78540456e-02
8.18495512e-01 7.95693874e-01 6.30979478e-01 8.22874978e-02
-1.28061129e-02 7.64544904e-01 5.20205200e-02 -1.25996217e-01
-6.27865374e-01 -5.17142773e-01 4.32863504e-01 1.58849800e+00
-1.06515050e-01 -1.62071869e-01 -6.67201638e-01 5.97010016e-01
-1.63218153e+00 -8.39801371e-01 1.03260368e-01 2.29725575e+00
9.51149583e-01 -7.42252395e-02 5.24135649e-01 8.47824931e-01
6.34467006e-01 4.85549808e-01 -7.72510290e-01 -3.89263660e-01
-5.24960794e-02 3.62168163e-01 -1.44825116e-01 2.47181207e-01
-1.08241642e+00 1.24827313e+00 7.02657795e+00 1.12499058e+00
-1.39868355e+00 1.32668957e-01 2.38344118e-01 -2.26379633e-01
4.89749340e-03 -4.06575590e-01 -8.32596004e-01 3.08170855e-01
1.01116419e+00 1.10608928e-01 8.23272526e-01 6.81446791e-01
4.19492453e-01 4.19743687e-01 -1.19118500e+00 1.26768637e+00
2.35712454e-01 -1.11553228e+00 -4.17965688e-02 -1.73023939e-01
6.41236246e-01 3.00738271e-02 2.42810011e-01 4.78526831e-01
-4.25438017e-01 -1.18274319e+00 6.50158525e-01 1.93938538e-02
1.04558885e+00 -5.56869149e-01 5.88896632e-01 5.04410505e-01
-1.01181924e+00 -2.38684610e-01 -2.21774238e-03 -1.70186967e-01
1.30470663e-01 1.55882373e-01 -1.04999840e+00 1.35340244e-01
5.91699958e-01 7.06853628e-01 -1.71254277e-01 9.94257569e-01
-1.93938956e-01 1.54864609e+00 -1.01950169e-01 -2.93941330e-02
4.45908532e-02 3.46364342e-02 8.22083414e-01 1.13201857e+00
1.79034531e-01 -2.02146754e-01 -1.25576938e-02 8.13725591e-01
-2.30230838e-01 3.75517637e-01 -4.48067188e-01 -5.97872972e-01
4.99655306e-01 9.69281375e-01 -2.79161036e-01 -1.54422313e-01
-2.66999424e-01 4.74886656e-01 1.43193424e-01 2.49910310e-01
-3.03226829e-01 -3.28230369e-03 7.08498836e-01 4.24383849e-01
7.97295719e-02 -4.86517966e-01 -3.53365123e-01 -9.15373862e-01
-1.72161296e-01 -1.25613415e+00 2.26142824e-01 -4.53000307e-01
-1.33435607e+00 8.10275137e-01 -3.10568899e-01 -1.43369079e+00
-5.31570017e-01 -5.71130633e-01 -7.77927935e-01 6.85217679e-01
-1.39092910e+00 -9.36558843e-01 -1.98782124e-02 4.64464456e-01
1.17440534e+00 -9.23505127e-01 1.01764226e+00 4.41232413e-01
-6.79894388e-01 7.46855021e-01 1.75391242e-01 -7.45765446e-03
8.63881648e-01 -1.22685969e+00 4.00316685e-01 3.46632421e-01
9.65929151e-01 6.32513046e-01 3.89926255e-01 -5.81903309e-02
-1.45336246e+00 -7.12148130e-01 9.00493026e-01 -2.72253066e-01
5.59243858e-01 -4.58722293e-01 -1.12827826e+00 1.05788425e-01
3.30039561e-02 -9.14176926e-02 1.22344375e+00 5.83169281e-01
-1.06887005e-01 -2.27567166e-01 -6.51205659e-01 3.37915361e-01
1.13847554e+00 -9.73944008e-01 -6.85336053e-01 1.90588251e-01
4.33618098e-01 -3.60030323e-01 -5.74911952e-01 6.26830578e-01
6.68525279e-01 -9.36675668e-01 7.69794703e-01 -8.08901072e-01
1.29387125e-01 -2.67553210e-01 -1.28970429e-01 -1.44758809e+00
-3.46309990e-02 -6.49875343e-01 -2.64353842e-01 1.32445824e+00
4.51640427e-01 -1.19282037e-01 5.70177078e-01 -7.87195712e-02
-3.79573613e-01 -7.20345497e-01 -9.50847626e-01 -1.01208484e+00
-4.98262718e-02 -7.74379194e-01 4.41554725e-01 9.88483369e-01
-6.47493452e-02 7.22719193e-01 -6.21969044e-01 -2.03075826e-01
4.19216096e-01 3.36740494e-01 7.47998416e-01 -1.06899643e+00
-5.78020692e-01 -3.44475210e-01 -1.35522425e-01 -1.09600627e+00
5.96055984e-01 -1.13852096e+00 2.33868837e-01 -1.27238500e+00
2.82493606e-03 -4.48705226e-01 -5.76458454e-01 4.43297863e-01
-1.50905386e-01 5.36149979e-01 3.38191777e-01 4.66802239e-01
-4.38405007e-01 6.42395556e-01 1.36860895e+00 -1.27941698e-01
-6.25738144e-01 6.64625049e-01 -4.34800088e-01 9.41121638e-01
9.85221267e-01 -1.89895377e-01 -4.10107732e-01 -3.78228694e-01
-4.65636402e-01 2.15559229e-01 2.88357660e-02 -1.01151490e+00
1.25811696e-01 2.91830689e-01 -1.01776235e-01 -6.29558444e-01
6.05569661e-01 -2.77055502e-01 -2.96314389e-01 1.31171897e-01
-6.83244467e-01 -6.83495700e-01 1.96762651e-01 1.98665783e-01
-7.44266808e-01 -3.03057998e-01 7.58960962e-01 9.21027139e-02
-6.22954845e-01 7.57922605e-02 -4.49030191e-01 2.52017602e-02
3.24850529e-01 -1.60011411e-01 2.83760667e-01 -6.43020511e-01
-1.01083994e+00 -8.26012045e-02 -1.42865062e-01 9.01555419e-01
6.37679577e-01 -1.26010990e+00 -8.36935699e-01 4.42487001e-01
1.06004186e-01 -2.54826605e-01 1.92373887e-01 6.54677510e-01
6.26912192e-02 4.39840376e-01 2.32309345e-02 -6.68799877e-01
-1.66894495e+00 2.17567250e-01 3.66706312e-01 -5.44503704e-03
-3.53383034e-01 8.80494058e-01 1.96138248e-01 -7.67052114e-01
7.94254959e-01 -2.06502587e-01 -4.10657108e-01 1.21775426e-01
4.72833186e-01 4.81916577e-01 1.93997845e-01 -6.90249801e-01
-2.53741801e-01 4.93982166e-01 3.78869511e-02 -7.09419698e-02
1.09296179e+00 4.36675362e-02 2.15413421e-01 1.05781889e+00
9.80156958e-01 2.26256520e-01 -8.34023595e-01 -2.43282795e-01
1.44585371e-01 -4.06829149e-01 3.50843459e-01 -7.77632415e-01
-1.00125194e+00 1.30231977e+00 6.22699797e-01 2.45098457e-01
1.10539675e+00 1.26210734e-01 9.76924956e-01 2.45671436e-01
1.64117157e-01 -1.10025001e+00 -7.87595958e-02 5.81319869e-01
1.07507610e+00 -1.32915986e+00 -4.36224014e-01 -4.33276296e-01
-6.94357574e-01 8.77728462e-01 5.83459079e-01 7.29767755e-02
5.02222538e-01 3.26634318e-01 4.26802039e-01 7.79496282e-02
-7.67671943e-01 -4.63329762e-01 6.28043354e-01 3.78334016e-01
9.28065777e-01 2.50230759e-01 -4.20553476e-01 9.19296443e-01
-5.49815893e-01 -1.16814375e-01 -5.90962246e-02 4.20006335e-01
-3.79733860e-01 -1.36540151e+00 -4.07251924e-01 4.72861469e-01
-5.98368108e-01 -2.13552907e-01 -9.26859200e-01 4.66623247e-01
5.78762777e-02 1.28664124e+00 -1.21849015e-01 -7.47414410e-01
4.65289682e-01 3.35431725e-01 5.32183468e-01 -9.45864737e-01
-9.54132199e-01 3.22856784e-01 2.83296973e-01 -8.69408101e-02
-4.50159937e-01 -5.67969143e-01 -1.10169446e+00 2.24033654e-01
-5.63513935e-01 3.79898041e-01 5.88466108e-01 8.74917567e-01
2.13764030e-02 6.79895401e-01 1.06836212e+00 -8.18392515e-01
-9.44368362e-01 -1.28321040e+00 -9.96399522e-01 2.76714116e-01
4.15040046e-01 -5.83768845e-01 -6.95966780e-01 -4.58739847e-02] | [15.260927200317383, 6.09234619140625] |
ea8155cf-0ce1-4831-a8a1-ef99854ac48c | training-neural-networks-for-aspect | null | null | https://openreview.net/forum?id=HyxoAoxOwN | https://openreview.net/pdf?id=HyxoAoxOwN | Training Neural Networks for Aspect Extraction Using Descriptive Keywords Only | Aspect extraction in online product reviews is a key task in sentiment analysis and opinion mining. Training supervised neural networks for aspect extraction is not possible when ground truth aspect labels are not available, while the unsupervised neural topic models fail to capture the particular aspects of interest. In this work, we propose a weakly supervised approach for training neural networks for aspect extraction in cases where only a small set of seed words, i.e., keywords that describe an aspect, are available. Our main contributions are as follows. First, we show that current weakly supervised networks fail to leverage the predictive power of the available seed words by comparing them to a simple bag-of-words classifier. Second, we propose a distillation approach for aspect extraction where the seed words are considered by the bag-of-words classifier (teacher) and distilled to the parameters of a neural network (student). Third, we show that regularization encourages the student to consider non-seed words for classification and, as a result, the student outperforms the teacher, which only considers the seed words. Finally, we empirically show that our proposed distillation approach outperforms (by up to 34.4% in F1 score) previous weakly supervised approaches for aspect extraction in six domains of Amazon product reviews. | ['Luis Gravano', 'Daniel Hsu', 'Giannis Karamanolakis'] | 2019-03-14 | null | null | null | iclr-workshop-lld-2019 | ['aspect-extraction', 'topic-models'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.87480682e-01 4.44759995e-01 -6.37609661e-01 -5.61649024e-01
-6.83593035e-01 -7.45260656e-01 6.27759695e-01 7.18262196e-01
-4.79066819e-01 5.94208419e-01 1.02094404e-01 -4.51660037e-01
2.04879194e-01 -8.93997431e-01 -6.82254553e-01 -7.39433467e-01
2.14425430e-01 3.61570954e-01 -3.55667397e-02 -2.81579107e-01
5.65572642e-02 1.40167996e-01 -1.54508686e+00 2.91924715e-01
6.95643187e-01 1.16480434e+00 -3.24928045e-01 3.75622809e-01
-3.43483508e-01 6.32425964e-01 -6.78107440e-01 -5.08060634e-01
1.44815758e-01 -2.71562725e-01 -6.37696624e-01 2.87011981e-01
3.49245548e-01 -2.78239936e-01 3.45904499e-01 1.04263282e+00
-1.07148872e-03 1.24126226e-01 8.25536728e-01 -1.12092066e+00
-5.34430325e-01 1.02257550e+00 -5.24444640e-01 -1.22146621e-01
-4.73423041e-02 -2.23546371e-01 1.58297300e+00 -9.88007128e-01
4.57934797e-01 7.22580671e-01 5.49540341e-01 2.48741210e-01
-8.93016636e-01 -3.58880937e-01 7.39653111e-01 -2.62228549e-01
-9.24579263e-01 -2.16079503e-01 9.14613664e-01 -2.48252079e-01
1.29755127e+00 -2.82058865e-02 8.56783628e-01 7.57097065e-01
-6.91494495e-02 1.06475723e+00 9.35122967e-01 -5.43176055e-01
5.16558170e-01 7.96948314e-01 8.13372731e-01 5.07575333e-01
7.57363260e-01 -2.63596833e-01 -5.79623520e-01 -4.21856910e-01
1.42254815e-01 -1.21706285e-01 1.10354826e-01 -3.09730470e-01
-5.93918443e-01 1.20153654e+00 1.61836550e-01 3.98461461e-01
-6.63040400e-01 -1.04394384e-01 3.40617537e-01 1.26212761e-01
1.03539658e+00 7.38546312e-01 -9.14399683e-01 -2.78985202e-02
-1.05841672e+00 9.58670378e-02 1.11928082e+00 1.03246033e+00
9.36641276e-01 1.92481697e-01 6.90441951e-02 7.59825408e-01
2.52577960e-01 5.03891230e-01 4.78221923e-01 -3.66502851e-01
2.75061071e-01 9.87922490e-01 -8.63442943e-02 -8.43875289e-01
-1.76117867e-01 -7.18288600e-01 -5.33530533e-01 2.67299786e-02
1.81895703e-01 -4.48203415e-01 -1.02961314e+00 1.37370813e+00
4.48845923e-01 -2.09538862e-01 2.52715886e-01 4.26168054e-01
1.00807750e+00 4.69189137e-01 1.63611814e-01 -2.81740457e-01
1.63284755e+00 -1.19337308e+00 -7.19525933e-01 -6.46146894e-01
6.66927159e-01 -7.22344100e-01 8.06022823e-01 5.89195251e-01
-8.28399181e-01 -2.58043617e-01 -1.24236691e+00 2.63659298e-01
-9.21356738e-01 3.18290055e-01 1.03905475e+00 6.37431920e-01
-7.24723279e-01 3.96593392e-01 -5.94367862e-01 -1.94464281e-01
6.38099194e-01 5.05714595e-01 -2.83907682e-01 9.89630520e-02
-1.02480817e+00 6.24241590e-01 4.05608714e-01 2.26885192e-02
-5.75496912e-01 -4.15464342e-01 -9.58931744e-01 1.67925641e-01
7.42179453e-01 -4.17066127e-01 1.31181753e+00 -1.42161047e+00
-1.24364805e+00 6.51031315e-01 -3.41797709e-01 -4.97836709e-01
-2.25668445e-01 -5.23524940e-01 -1.52944416e-01 2.98761755e-01
5.91064394e-02 4.66442972e-01 1.03820705e+00 -1.39262974e+00
-9.38333571e-01 -1.91674799e-01 5.60011268e-01 1.63622275e-01
-7.32573032e-01 -5.28060645e-03 -3.44577610e-01 -5.06346583e-01
-1.79283097e-01 -9.78957415e-01 -4.12547082e-01 -3.29518169e-01
-6.83334768e-01 -4.09703583e-01 7.11647809e-01 -2.15701088e-01
1.05520785e+00 -2.01192307e+00 -3.10519189e-01 4.46407706e-01
3.96019369e-01 3.75211656e-01 -1.47509038e-01 1.07797682e-01
-2.66428769e-01 1.85191646e-01 -2.10853547e-01 -4.41956401e-01
-1.16948433e-01 2.26147369e-01 -3.72128367e-01 1.06729090e-01
4.16349798e-01 9.03279960e-01 -7.39023030e-01 -3.50003928e-01
-1.22907922e-01 4.32782918e-01 -4.90045488e-01 8.66445675e-02
-5.26042819e-01 -3.49242657e-01 -5.28397560e-01 6.77107275e-01
4.71746057e-01 -4.79485571e-01 6.00775098e-03 -3.04379076e-01
2.13700980e-01 7.51517892e-01 -1.07998037e+00 1.09367740e+00
-6.46963894e-01 4.73097146e-01 -1.72442004e-01 -9.36590254e-01
8.21523964e-01 3.56588751e-01 5.34512281e-01 -2.06231728e-01
3.40236932e-01 1.24541737e-01 -1.12762660e-01 -1.47964373e-01
6.59336627e-01 -4.30419117e-01 -1.07231185e-01 8.73226345e-01
4.55879867e-01 -2.28353411e-01 4.86965507e-01 2.73783803e-01
8.45575988e-01 -5.75859323e-02 6.33235812e-01 -1.18590459e-01
4.94297594e-01 1.82595611e-01 3.61748278e-01 6.50970757e-01
7.10173771e-02 5.11521339e-01 9.11027253e-01 -4.05381501e-01
-8.40655148e-01 -5.34992099e-01 3.16409990e-02 1.31102657e+00
-2.90860385e-01 -6.50693774e-01 -7.34597147e-01 -1.25584161e+00
-2.01896444e-01 8.39814067e-01 -8.70695233e-01 -1.56106889e-01
-3.31757337e-01 -8.44936788e-01 1.16676755e-01 5.77225685e-01
1.54140294e-01 -1.17442977e+00 -3.12299490e-01 7.74736628e-02
6.68113008e-02 -1.33767450e+00 -1.54784665e-01 8.04133892e-01
-9.39950526e-01 -1.06323934e+00 -4.20625478e-01 -7.04032958e-01
9.92123604e-01 2.05810860e-01 1.37119377e+00 1.30959302e-01
3.29536378e-01 2.39914402e-01 -5.03178895e-01 -7.94048488e-01
-2.22508445e-01 6.14374936e-01 -1.37504742e-01 -6.37560338e-02
1.13083875e+00 -4.90914255e-01 -3.41209054e-01 -4.31245305e-02
-1.07607019e+00 -2.77819991e-01 8.06026816e-01 7.34269679e-01
6.26754045e-01 1.02211684e-01 5.78227580e-01 -1.58791828e+00
8.30980301e-01 -4.04045820e-01 -4.58735704e-01 -6.33141994e-02
-9.45345581e-01 5.05240597e-02 5.93827605e-01 -7.72553325e-01
-9.08868134e-01 1.94686145e-01 -2.33402073e-01 2.02702984e-01
-3.90522301e-01 8.02353978e-01 -7.60506317e-02 3.00654441e-01
6.41235590e-01 1.23910800e-01 -2.06981644e-01 -2.42810652e-01
4.04503435e-01 7.34344125e-01 -1.70195609e-01 -2.51203865e-01
7.22391248e-01 3.99338901e-01 -2.90856421e-01 -9.16842461e-01
-1.53544009e+00 -7.48712420e-01 -7.15157747e-01 1.72153398e-01
5.62361240e-01 -9.59567308e-01 -2.65707582e-01 3.26402068e-01
-1.04556775e+00 8.72118324e-02 -7.15668082e-01 4.03330535e-01
-2.69475669e-01 4.81472537e-02 -5.35940468e-01 -8.38340461e-01
-6.76975131e-01 -9.72664893e-01 1.05115592e+00 3.65203142e-01
-4.95132267e-01 -1.09251893e+00 -1.21513493e-01 2.57517964e-01
3.92252147e-01 -7.46169686e-02 9.99530315e-01 -1.55166960e+00
-2.22788900e-01 -6.45071805e-01 7.14055896e-02 7.48555422e-01
4.64138299e-01 -1.09244846e-01 -1.23087084e+00 1.33329540e-01
3.18923652e-01 -2.88700789e-01 8.70481730e-01 3.61016184e-01
5.76070249e-01 -6.07802749e-01 -1.29598588e-01 1.66572914e-01
1.29920542e+00 -7.40383379e-03 1.55039653e-01 4.77117717e-01
6.90281689e-01 7.11708903e-01 6.41970456e-01 2.02837840e-01
2.83913434e-01 2.42971122e-01 1.91036120e-01 -2.32729837e-01
2.92334199e-01 -1.88018054e-01 3.86671722e-01 9.18255806e-01
6.49951398e-02 -1.79852083e-01 -5.50809264e-01 9.34062362e-01
-1.59426153e+00 -2.98482627e-01 -1.09984316e-01 2.03309083e+00
1.04428291e+00 6.57044768e-01 2.83753812e-01 1.75603658e-01
2.53755778e-01 2.88373590e-01 -4.40928966e-01 -5.63366294e-01
-8.14090371e-02 4.65746552e-01 4.14468348e-01 5.42336643e-01
-1.30621243e+00 1.05547571e+00 5.26245975e+00 6.95259869e-01
-8.76662672e-01 1.65733159e-01 6.71173930e-01 -9.59550962e-02
-4.47326362e-01 1.20141715e-01 -1.16729355e+00 -1.72512420e-02
1.01507723e+00 -3.10329795e-02 -2.19040945e-01 1.35281408e+00
-2.52677891e-02 -3.54420006e-01 -1.16440833e+00 3.31698120e-01
4.88505036e-01 -9.23827946e-01 2.52085924e-01 1.42888010e-01
8.78941357e-01 -5.97562268e-02 5.27726188e-02 3.73226553e-01
4.27277505e-01 -9.12467241e-01 2.84258574e-01 -2.66549010e-02
7.18766525e-02 -1.02187681e+00 1.20757258e+00 3.14144284e-01
-9.81054366e-01 1.76967084e-01 -4.48164999e-01 1.05773188e-01
1.10476829e-01 1.13363445e+00 -7.91494191e-01 2.08858132e-01
4.05300587e-01 6.32184505e-01 -4.98221487e-01 6.66079104e-01
-6.49930120e-01 1.02049649e+00 -3.95469725e-01 -3.82419586e-01
6.45463824e-01 -1.04674228e-01 3.87686849e-01 1.16625786e+00
-1.04027040e-01 -4.28413749e-02 1.29939556e-01 5.71485400e-01
-1.26196295e-01 3.73394102e-01 -6.53233349e-01 -3.72751087e-01
-1.62227601e-01 1.50216591e+00 -9.62234139e-01 -6.05287790e-01
-5.54231644e-01 5.19875586e-01 1.66490898e-01 4.63207543e-01
-2.95614392e-01 -4.38549906e-01 5.80154300e-01 5.89790903e-02
8.69533062e-01 -4.96600531e-02 -4.53836471e-01 -1.12376916e+00
2.48751208e-01 -1.04425871e+00 2.39569947e-01 -4.71119970e-01
-1.28938091e+00 1.04087746e+00 6.63475459e-03 -1.05610228e+00
-4.85208958e-01 -8.19649398e-01 -5.38083076e-01 7.53524125e-01
-1.87595713e+00 -1.31126261e+00 9.63796005e-02 1.48271322e-01
6.03691161e-01 -2.33318806e-01 9.40711081e-01 -2.87018530e-02
-2.26568714e-01 4.25215662e-01 -3.22732747e-01 2.55846560e-01
4.77239788e-01 -1.39458799e+00 4.06424552e-01 6.21595740e-01
5.01677930e-01 9.92611289e-01 8.41298461e-01 -7.55196154e-01
-1.00149596e+00 -8.87087643e-01 1.29323936e+00 -3.35051179e-01
8.04681599e-01 -4.51789290e-01 -8.99969041e-01 6.23803020e-01
5.19142687e-01 -3.57136242e-02 1.20190716e+00 5.80481887e-01
-5.15851915e-01 -6.81056082e-02 -9.60247159e-01 4.31930840e-01
1.72322512e-01 -4.66466755e-01 -7.33087540e-01 5.01043618e-01
8.45114529e-01 3.22126858e-02 -5.57776451e-01 1.99158266e-01
4.57063556e-01 -5.72224557e-01 6.84823215e-01 -9.05244708e-01
7.36778200e-01 -9.63427871e-02 4.60609198e-02 -1.38880765e+00
2.50760734e-01 -2.31464088e-01 -3.99921566e-01 1.37213218e+00
1.20898747e+00 -5.67636967e-01 1.11474574e+00 5.00872195e-01
4.09914970e-01 -1.18687379e+00 -5.04908264e-01 -4.16875303e-01
1.47333086e-01 -5.76660514e-01 4.11547273e-01 6.84652686e-01
7.26373643e-02 9.52151656e-01 -2.23437637e-01 -4.59171794e-02
3.38735133e-01 3.47668439e-01 3.93404812e-01 -1.18110263e+00
-4.42731559e-01 -3.13039601e-01 -1.03627071e-01 -1.02236891e+00
2.10595727e-01 -5.50043762e-01 2.82713413e-01 -1.56891274e+00
2.53277957e-01 -2.78925002e-01 -3.33678305e-01 7.40756392e-01
-3.46220315e-01 2.64762372e-01 -3.23489420e-02 -1.15368739e-01
-5.71263373e-01 4.95902210e-01 8.79700065e-01 -3.36375028e-01
-2.98296481e-01 6.33653581e-01 -1.24195254e+00 1.12215507e+00
9.43814993e-01 -8.40353191e-01 -4.80366945e-01 5.70752099e-03
7.12809920e-01 -5.78751683e-01 -1.94899172e-01 -3.91867876e-01
2.35298246e-01 8.72202739e-02 1.40617028e-01 -6.18554711e-01
2.67642826e-01 -1.01037979e+00 -8.19074094e-01 -1.83868203e-02
-4.41530824e-01 -8.77765268e-02 1.95656598e-01 4.73423630e-01
-4.57530975e-01 -7.64338493e-01 3.28743875e-01 -2.09238097e-01
-1.39665246e-01 1.38767272e-01 -5.40118575e-01 1.13971919e-01
6.60678983e-01 9.63510480e-03 -1.22330427e-01 -5.86124241e-01
-5.80190957e-01 3.43229645e-03 1.74396131e-02 3.51330012e-01
5.44412315e-01 -1.04505920e+00 -5.23725629e-01 7.49835968e-02
3.43343049e-01 1.76033020e-01 -2.34948263e-01 5.40570915e-01
1.28871985e-02 4.80117679e-01 3.63600612e-01 -3.88897985e-01
-1.39111400e+00 5.17541707e-01 -6.77945092e-03 -5.96718132e-01
-3.22648346e-01 8.48624766e-01 1.81752264e-01 -6.36142492e-01
2.67668903e-01 -5.39702058e-01 -6.95714653e-01 5.44026434e-01
4.55104083e-01 -3.13686818e-01 4.24005330e-01 -5.68435431e-01
-1.55426383e-01 4.08107132e-01 -5.57934761e-01 -9.72368270e-02
1.83329475e+00 1.72208294e-01 -1.41714349e-01 6.57956958e-01
1.12443876e+00 2.25173056e-01 -7.79839337e-01 -4.70188498e-01
1.20327353e-01 1.16618030e-01 1.72047928e-01 -7.52582431e-01
-1.27037084e+00 7.61499524e-01 1.51412874e-01 4.74481791e-01
1.11500394e+00 1.12431869e-01 6.94101989e-01 8.29014719e-01
-4.82740179e-02 -1.12389767e+00 -8.53126124e-02 5.91168523e-01
3.01463515e-01 -1.33617735e+00 2.88056314e-01 -6.18144155e-01
-8.61090064e-01 1.01495123e+00 6.03183806e-01 -1.95995271e-01
1.05879676e+00 5.38159549e-01 2.18355581e-01 -5.07118285e-01
-9.17793751e-01 -4.64260936e-01 5.66712022e-01 4.01271671e-01
6.19187474e-01 -6.80267885e-02 -2.21095487e-01 9.42020714e-01
-3.50556731e-01 -2.41609231e-01 5.92174232e-01 1.08836043e+00
-5.91642737e-01 -1.19042587e+00 5.84215038e-02 8.42505515e-01
-8.92670393e-01 -5.98392367e-01 -6.29485011e-01 5.67548990e-01
7.93955922e-02 1.07602632e+00 -2.96294630e-01 -7.47747421e-02
2.99056083e-01 1.92892686e-01 -1.33689359e-01 -1.05236351e+00
-9.46833789e-01 2.24448845e-01 3.28946888e-01 -1.65940687e-01
-6.26659095e-01 -4.09982592e-01 -8.47098589e-01 3.20386142e-01
-7.57790625e-01 5.39476633e-01 9.02664542e-01 1.39783847e+00
1.17905386e-01 4.84302312e-01 5.35622418e-01 -3.66938084e-01
-3.66533756e-01 -1.09796965e+00 -5.21615148e-01 1.05576120e-01
4.29794669e-01 -3.57985258e-01 -6.87806189e-01 1.98380113e-01] | [11.337011337280273, 6.722313404083252] |
e2ead0c3-8269-4b47-a8d4-df728ad92951 | dealing-with-typos-for-bert-based-passage | 2108.12139 | null | https://arxiv.org/abs/2108.12139v2 | https://arxiv.org/pdf/2108.12139v2.pdf | Dealing with Typos for BERT-based Passage Retrieval and Ranking | Passage retrieval and ranking is a key task in open-domain question answering and information retrieval. Current effective approaches mostly rely on pre-trained deep language model-based retrievers and rankers. These methods have been shown to effectively model the semantic matching between queries and passages, also in presence of keyword mismatch, i.e. passages that are relevant to a query but do not contain important query keywords. In this paper we consider the Dense Retriever (DR), a passage retrieval method, and the BERT re-ranker, a popular passage re-ranking method. In this context, we formally investigate how these models respond and adapt to a specific type of keyword mismatch -- that caused by keyword typos occurring in queries. Through empirical investigation, we find that typos can lead to a significant drop in retrieval and ranking effectiveness. We then propose a simple typos-aware training framework for DR and BERT re-ranker to address this issue. Our experimental results on the MS MARCO passage ranking dataset show that, with our proposed typos-aware training, DR and BERT re-ranker can become robust to typos in queries, resulting in significantly improved effectiveness compared to models trained without appropriately accounting for typos. | ['Guido Zuccon', 'Shengyao Zhuang'] | 2021-08-27 | null | https://aclanthology.org/2021.emnlp-main.225 | https://aclanthology.org/2021.emnlp-main.225.pdf | emnlp-2021-11 | ['passage-ranking', 'passage-re-ranking'] | ['natural-language-processing', 'natural-language-processing'] | [-1.48375764e-01 -4.77155268e-01 -1.01621680e-01 -7.86032621e-03
-1.41926825e+00 -7.39892423e-01 7.23517120e-01 6.30030096e-01
-6.19182229e-01 5.17686665e-01 5.06297946e-01 -1.36415482e-01
-5.91197670e-01 -7.18272507e-01 -9.20808733e-01 -1.12424660e-02
9.58423167e-02 8.23397458e-01 6.09026730e-01 -7.71998525e-01
5.79836965e-01 2.23595962e-01 -1.71890116e+00 6.98648751e-01
9.33475554e-01 8.04933846e-01 2.04076514e-01 8.55788052e-01
-2.67953128e-01 8.26993704e-01 -8.99208784e-01 -3.14978868e-01
1.61981627e-01 -2.77512640e-01 -1.19423401e+00 -6.90199077e-01
7.65884280e-01 -5.10437310e-01 -6.41821027e-01 5.73019207e-01
8.33647490e-01 4.87735778e-01 7.10885763e-01 -6.87352479e-01
-7.58104801e-01 8.11015844e-01 7.30780559e-03 7.29269564e-01
7.22474277e-01 -3.72704893e-01 1.36563921e+00 -9.83569801e-01
6.12205744e-01 1.22695601e+00 6.47364914e-01 3.83396506e-01
-8.77548099e-01 -2.13237405e-01 4.57590620e-04 4.01208580e-01
-1.46563542e+00 -3.59029830e-01 5.18826365e-01 -8.69907886e-02
1.16485775e+00 6.58241391e-01 1.31761044e-01 1.07233274e+00
-5.40242083e-02 8.42561185e-01 5.42862236e-01 -5.05935490e-01
6.98164571e-03 7.35545903e-02 4.46825564e-01 8.53620991e-02
1.01174138e-01 -4.28286120e-02 -5.04007161e-01 -2.98818946e-01
1.86761543e-01 1.01287410e-01 -3.31944734e-01 1.18035637e-01
-9.02011395e-01 6.82412505e-01 5.08441746e-01 5.03811121e-01
-2.65650362e-01 1.67199727e-02 3.93406659e-01 8.73262227e-01
4.88844723e-01 1.09556556e+00 -5.76918185e-01 -6.17242232e-02
-1.09142387e+00 7.57211208e-01 9.98520374e-01 9.49432254e-01
6.34475946e-01 -6.29320800e-01 -8.69069397e-01 1.30143642e+00
2.97022108e-02 4.67823267e-01 7.35868037e-01 -6.23143971e-01
4.00094956e-01 5.97877264e-01 4.49311823e-01 -1.03753865e+00
-3.93866092e-01 -6.99248135e-01 -2.63507545e-01 -8.83336663e-01
1.80904239e-01 3.30792397e-01 -7.87604809e-01 1.58400619e+00
-2.14840602e-02 4.71951291e-02 1.99166209e-01 7.57348359e-01
1.13931787e+00 6.64796829e-01 -2.04873793e-02 6.33070339e-03
1.30379021e+00 -1.09516191e+00 -5.39401352e-01 -1.99483350e-01
8.93315196e-01 -1.16482830e+00 1.44629705e+00 -3.97684462e-02
-1.03446841e+00 -4.52140987e-01 -8.48329782e-01 -4.03672069e-01
-6.45333111e-01 -1.63151428e-01 8.51525590e-02 1.27746433e-01
-1.27300513e+00 6.39529407e-01 -1.29436344e-01 -7.28386939e-01
-1.92595005e-01 7.44501054e-02 7.44302645e-02 -3.74211729e-01
-1.66733897e+00 9.71174896e-01 2.64332235e-01 -1.42965972e-01
-9.37270224e-01 -9.59805906e-01 -3.34669739e-01 3.60213488e-01
3.62973183e-01 -9.50930953e-01 1.61234176e+00 -6.42650127e-01
-1.02264488e+00 7.78122723e-01 -3.51984024e-01 -4.10768330e-01
3.12974304e-01 -5.38413942e-01 -6.04379833e-01 4.55501050e-01
2.41093695e-01 2.36681029e-01 6.29903674e-01 -1.25127435e+00
-6.10358953e-01 -1.11920558e-01 4.75999892e-01 4.42694068e-01
-4.42556024e-01 2.02372268e-01 -5.99702001e-01 -6.33557975e-01
-4.80791070e-02 -6.44947410e-01 -2.58337101e-03 -6.78550601e-01
-2.79412568e-01 -6.03167474e-01 3.40449363e-01 -6.55028820e-01
1.65216887e+00 -1.78059268e+00 -2.14062572e-01 1.79351762e-01
1.76606953e-01 4.65789765e-01 -6.59627378e-01 1.02543604e+00
4.89654243e-01 3.53692174e-01 1.76320687e-01 -1.35486811e-01
8.94875675e-02 1.64823774e-02 -6.38385773e-01 -1.83624834e-01
-1.34906426e-01 1.13937128e+00 -1.09910882e+00 -4.07524168e-01
-2.83234775e-01 2.19791427e-01 -5.73093355e-01 4.81122881e-01
-4.64690298e-01 7.33835623e-02 -6.55115068e-01 5.76033533e-01
2.95799196e-01 -3.79827231e-01 -3.62639219e-01 -9.31443926e-03
2.09997237e-01 8.67189109e-01 -5.67935109e-01 1.61098230e+00
-7.52261937e-01 5.34344196e-01 -3.38744819e-01 -6.77434325e-01
7.85129189e-01 2.67987490e-01 2.91743487e-01 -1.41345811e+00
-9.82166752e-02 6.82272851e-01 -3.24429452e-01 -6.61847949e-01
1.22987676e+00 1.03760041e-01 -9.16185006e-02 4.49873358e-01
-1.37974143e-01 -2.96758674e-02 6.07517600e-01 6.25449896e-01
1.44797838e+00 -3.41738939e-01 -1.49238005e-01 -2.51132518e-01
5.53919494e-01 1.50839970e-01 1.32450489e-02 1.59323192e+00
1.43972114e-01 8.56944442e-01 7.35218599e-02 -6.35522157e-02
-1.02497101e+00 -8.11334848e-01 -1.10553540e-01 1.58138537e+00
4.50303972e-01 -4.52705890e-01 -5.20753443e-01 -4.67854172e-01
1.43400252e-01 6.06627405e-01 -3.58999610e-01 -5.70597112e-01
-7.21882582e-01 -4.14203584e-01 6.30019248e-01 3.39418739e-01
1.59847677e-01 -1.06872976e+00 -2.24107996e-01 3.61422449e-01
-6.04688108e-01 -1.06285501e+00 -5.87042749e-01 -2.35788777e-01
-7.67259061e-01 -9.92303729e-01 -1.07835639e+00 -8.45272779e-01
2.29583710e-01 6.53402209e-01 1.74858558e+00 6.64133012e-01
7.34132715e-03 9.57605541e-01 -1.00327361e+00 -1.54749885e-01
-4.59195197e-01 5.41666687e-01 -2.19994932e-01 -2.89226890e-01
4.18111980e-01 -3.99529427e-01 -1.05039251e+00 5.84378004e-01
-1.36505771e+00 -5.05190194e-01 4.09959257e-01 7.60989785e-01
4.03133035e-01 -4.65112925e-01 9.69556391e-01 -7.15063930e-01
1.25897300e+00 -7.39605784e-01 -1.79688960e-01 7.20734000e-01
-7.50246525e-01 1.46412119e-01 6.68367088e-01 -5.37303030e-01
-6.05627835e-01 -9.19730604e-01 -3.29125494e-01 -1.71642408e-01
2.24459395e-01 9.42066491e-01 2.81355470e-01 -9.94115137e-03
9.56825376e-01 1.56213596e-01 -5.44443309e-01 -8.36987019e-01
2.30798587e-01 9.08999562e-01 2.03155413e-01 -6.02850199e-01
6.15532935e-01 2.79770017e-01 -4.41188574e-01 -6.92316294e-01
-1.13306701e+00 -1.19503367e+00 -1.54092848e-01 -7.70234093e-02
2.86804467e-01 -9.42771673e-01 -2.55878717e-01 2.66176224e-01
-1.12314880e+00 -2.62387604e-01 -2.95289397e-01 2.22946569e-01
-2.95177341e-01 4.17790681e-01 -6.64078116e-01 -3.44582587e-01
-5.18421888e-01 -7.70703435e-01 1.16556966e+00 2.99639732e-01
-1.60233811e-01 -1.06332469e+00 3.71587008e-01 4.40467983e-01
9.30118561e-01 -3.48307073e-01 1.09441149e+00 -1.33674884e+00
-4.24124688e-01 -5.67582011e-01 -1.17035940e-01 1.42529652e-01
-1.52257517e-01 -3.64901751e-01 -8.02687347e-01 -4.52047974e-01
-2.54327804e-01 -4.87009436e-01 1.12593579e+00 -6.02067858e-02
9.30750787e-01 -5.34242451e-01 -1.87788963e-01 2.11782634e-01
1.25790131e+00 3.10539035e-03 6.00693226e-01 5.52442372e-01
4.75950241e-01 4.45723444e-01 6.09046936e-01 3.25747341e-01
4.85830545e-01 7.57507324e-01 1.37828678e-01 9.95063782e-02
-2.73291826e-01 -6.38628662e-01 1.22287318e-01 1.00235164e+00
4.72091734e-01 -7.59613276e-01 -9.11087930e-01 8.53724182e-01
-1.79326653e+00 -7.82435775e-01 2.13529356e-02 2.41829896e+00
9.10522044e-01 -2.15811238e-01 -3.38479318e-02 -1.31951049e-01
4.71328378e-01 1.15211166e-01 -3.98973078e-01 -2.86607653e-01
-3.03640127e-01 3.71568620e-01 3.08780402e-01 6.37905598e-01
-7.19932675e-01 1.00079620e+00 6.48001385e+00 1.01765776e+00
-7.58029640e-01 3.37669104e-02 2.70798534e-01 -1.53298721e-01
-7.27753997e-01 6.55180439e-02 -7.59086847e-01 3.64906341e-01
1.04774702e+00 -4.73227113e-01 5.68571150e-01 4.55609292e-01
1.65754408e-01 5.11808926e-03 -1.30439854e+00 7.18869328e-01
3.95762295e-01 -1.21725023e+00 6.13231540e-01 -4.88472402e-01
6.48519695e-01 9.47675183e-02 -6.22570440e-02 9.21183467e-01
1.94476768e-02 -9.69516516e-01 5.72672069e-01 6.68006122e-01
3.11796874e-01 -3.97424638e-01 8.10675085e-01 3.59409928e-01
-8.18161368e-01 -1.06190421e-01 -6.06805146e-01 1.01899996e-01
-2.52304599e-02 3.37890923e-01 -7.50745714e-01 6.12752020e-01
7.75835216e-01 5.02694488e-01 -8.47743690e-01 1.42967272e+00
5.69390170e-02 5.89477241e-01 -3.66824597e-01 -3.24619651e-01
2.40900993e-01 3.21269602e-01 8.48316491e-01 1.21631348e+00
3.25075388e-01 7.86629319e-02 -3.06853056e-02 5.86616158e-01
-3.70990634e-01 4.20897871e-01 -3.21358681e-01 -1.18555881e-01
7.07229197e-01 6.69520736e-01 -3.56209219e-01 -5.53201139e-01
-2.70975471e-01 1.05229008e+00 4.56725866e-01 8.03441882e-01
-4.81731921e-01 -4.32929933e-01 3.75223070e-01 4.15959895e-01
3.08428794e-01 1.97551116e-01 2.68071949e-01 -1.26684988e+00
3.62555176e-01 -1.07626557e+00 6.79446757e-01 -7.40879297e-01
-1.47755933e+00 6.01209462e-01 1.31689385e-01 -1.13498342e+00
-2.82563537e-01 -1.29412204e-01 -4.26097423e-01 7.04427779e-01
-2.00456643e+00 -8.45010161e-01 -9.70006883e-02 5.51575541e-01
6.45869732e-01 6.20868877e-02 6.22701347e-01 7.65268505e-01
-1.22539498e-01 8.73960674e-01 6.33048475e-01 -6.03741519e-02
8.07667136e-01 -1.05157244e+00 1.87688068e-01 6.48397326e-01
3.35631341e-01 1.05280530e+00 7.71997750e-01 -3.77846688e-01
-1.31423640e+00 -9.55053329e-01 1.45670545e+00 -6.47443891e-01
5.70486844e-01 -1.34172365e-01 -1.19649601e+00 3.51572871e-01
4.94371690e-02 -2.94617414e-01 5.68306684e-01 2.66654223e-01
-5.96241117e-01 -2.13762313e-01 -8.33447754e-01 6.84346616e-01
1.05250823e+00 -8.83993924e-01 -1.00293100e+00 6.77967787e-01
1.09307349e+00 -3.02900195e-01 -6.35532081e-01 6.36171460e-01
4.48317200e-01 -8.35941792e-01 1.21539724e+00 -8.56097460e-01
2.84199506e-01 -7.89568722e-02 -1.15704730e-01 -1.18981957e+00
-1.28775597e-01 -5.42944372e-01 -2.51065910e-01 1.09271348e+00
5.41654766e-01 -4.37859178e-01 2.10242242e-01 4.74040896e-01
-2.32048810e-01 -5.73188305e-01 -9.63358581e-01 -9.16777849e-01
4.79395837e-01 -1.40609279e-01 7.34271705e-01 6.54115081e-01
-5.75491227e-02 4.09268409e-01 -1.11773238e-01 -8.63215607e-03
-6.39590099e-02 1.71629772e-01 6.67100966e-01 -1.09808397e+00
-1.63111433e-01 -6.62966967e-01 -4.40202542e-02 -1.55679643e+00
6.40114620e-02 -1.01899636e+00 1.81614563e-01 -1.87435555e+00
3.26383293e-01 -4.61834222e-01 -6.16562605e-01 4.49936911e-02
-5.98339915e-01 1.46451309e-01 6.51444495e-02 5.91579914e-01
-1.18791878e+00 5.36440074e-01 1.04113936e+00 -2.65035808e-01
-2.79283583e-01 1.63830355e-01 -8.10139239e-01 1.73430189e-01
6.21830583e-01 -6.96858823e-01 -3.91441166e-01 -7.35278606e-01
9.02100623e-01 8.06298554e-02 3.95205855e-01 -7.63362169e-01
2.67376512e-01 1.74499333e-01 -2.45651111e-01 -4.87799317e-01
2.59662252e-02 -6.41258597e-01 -4.18827474e-01 1.13009177e-01
-1.07435203e+00 4.11614954e-01 2.24664971e-01 6.10348880e-01
-4.60046947e-01 -4.35883909e-01 2.54523873e-01 -1.42039835e-01
-6.88236773e-01 2.10482255e-01 -4.66978639e-01 6.31656885e-01
2.57700831e-01 1.19139224e-01 -5.54386199e-01 -7.44659424e-01
-3.41751277e-01 5.07797658e-01 2.63785571e-01 8.53919983e-01
6.39006078e-01 -1.22572660e+00 -9.64331686e-01 -2.59856313e-01
5.11214793e-01 -2.39129186e-01 2.72929043e-01 6.25807583e-01
-3.88198972e-01 8.28312933e-01 5.92310369e-01 -2.96264410e-01
-1.00230038e+00 4.48185325e-01 4.03378159e-01 -7.10885584e-01
-3.89727235e-01 8.71489346e-01 -1.20626152e-01 -5.61657190e-01
4.64083105e-01 -4.46443200e-01 -4.64686692e-01 5.29876724e-02
5.74770868e-01 3.66136253e-01 5.57220697e-01 -5.11174738e-01
-1.71504855e-01 4.89598632e-01 -3.75872999e-01 -1.34581104e-01
7.69458115e-01 -4.10510719e-01 -1.13089398e-01 3.52419019e-01
1.50886583e+00 2.10789517e-02 -3.83875608e-01 -6.64182007e-01
3.80487263e-01 -3.72830003e-01 -4.07556854e-02 -1.09428990e+00
-5.54476976e-01 6.57669783e-01 4.80071485e-01 3.38548630e-01
1.04761541e+00 1.23612419e-01 1.23173296e+00 8.15623999e-01
3.35164487e-01 -1.10674727e+00 2.66015351e-01 9.65931356e-01
1.00600863e+00 -1.08126235e+00 -2.35975266e-01 1.50983587e-01
-2.32103229e-01 7.23499000e-01 4.63239521e-01 5.45748137e-02
4.55622882e-01 -5.67620456e-01 2.69485176e-01 -3.38954121e-01
-8.93785059e-01 -4.33177084e-01 6.94936812e-01 1.78853527e-01
4.05514121e-01 -2.79817611e-01 -6.38246775e-01 4.44300443e-01
-2.13046595e-01 -1.16018154e-01 3.07759464e-01 1.00093901e+00
-6.09076083e-01 -9.49340820e-01 -1.77947402e-01 6.49347544e-01
-5.25646210e-01 -6.17384374e-01 -6.22524500e-01 5.04767716e-01
-5.59400380e-01 1.31296241e+00 1.84304908e-01 -3.38407665e-01
5.54634571e-01 5.63696884e-02 1.75375476e-01 -5.99948049e-01
-1.03034449e+00 -2.64295667e-01 2.12400585e-01 -5.13753116e-01
-1.69052705e-01 -2.03926921e-01 -8.93996835e-01 -8.16082880e-02
-5.10104895e-01 6.92868114e-01 2.84374475e-01 1.01372516e+00
6.86262667e-01 3.60588491e-01 6.63240731e-01 -1.59271762e-01
-9.31334317e-01 -1.16871524e+00 -3.96881521e-01 8.16247404e-01
5.72514832e-01 -3.22220683e-01 -6.84688449e-01 -5.19044697e-01] | [11.53713607788086, 7.679036617279053] |
ec1cc8f4-8017-42d8-ba8f-bcc402c56018 | gode-integrating-biochemical-knowledge-graph | 2306.01631 | null | https://arxiv.org/abs/2306.01631v1 | https://arxiv.org/pdf/2306.01631v1.pdf | Gode -- Integrating Biochemical Knowledge Graph into Pre-training Molecule Graph Neural Network | The precise prediction of molecular properties holds paramount importance in facilitating the development of innovative treatments and comprehending the intricate interplay between chemicals and biological systems. In this study, we propose a novel approach that integrates graph representations of individual molecular structures with multi-domain information from biomedical knowledge graphs (KGs). Integrating information from both levels, we can pre-train a more extensive and robust representation for both molecule-level and KG-level prediction tasks with our novel self-supervision strategy. For performance evaluation, we fine-tune our pre-trained model on 11 challenging chemical property prediction tasks. Results from our framework demonstrate our fine-tuned models outperform existing state-of-the-art models. | ['Pengcheng Jiang'] | 2023-06-02 | null | null | null | null | ['knowledge-graphs', 'property-prediction'] | ['knowledge-base', 'medical'] | [ 6.30518258e-01 2.86544655e-02 -7.06981897e-01 -2.97407806e-01
-7.03821599e-01 -5.61463594e-01 5.31601727e-01 9.52856660e-01
-6.10905588e-02 1.15992677e+00 2.24034309e-01 -4.71556634e-01
-3.94898653e-01 -8.31612229e-01 -1.07145917e+00 -5.66872418e-01
-1.68459624e-01 2.76531696e-01 1.51373938e-01 -8.56344998e-02
1.06792510e-01 6.48274243e-01 -8.20697904e-01 4.50074017e-01
1.17475331e+00 8.41305852e-01 9.19643044e-02 3.25660646e-01
1.50742173e-01 8.38197410e-01 -1.93674877e-01 -5.57277262e-01
-1.44476414e-01 -2.24607050e-01 -8.75064075e-01 -4.43758816e-01
4.45733875e-01 3.33950758e-01 -3.48491162e-01 8.97198200e-01
4.54764843e-01 1.41317904e-01 9.83012617e-01 -7.09514320e-01
-8.73451054e-01 5.85600853e-01 -3.73757601e-01 8.22649002e-02
5.14909804e-01 2.53113329e-01 1.14742100e+00 -5.93000531e-01
7.94098318e-01 1.13176215e+00 6.76570535e-01 4.19520199e-01
-1.64543021e+00 -6.45122468e-01 3.31498057e-01 4.68528539e-01
-1.30025148e+00 -4.01908815e-01 6.06502652e-01 -5.23358345e-01
1.33110118e+00 -1.51879236e-01 3.20713103e-01 1.28767645e+00
6.55704081e-01 4.13095325e-01 7.61365235e-01 2.37713195e-02
2.27123290e-01 -1.44081607e-01 -6.89949316e-04 9.02230322e-01
5.90610445e-01 -1.28672987e-01 -6.15800977e-01 -4.88850087e-01
4.05836225e-01 1.10454611e-01 -4.69545633e-01 -4.03367609e-01
-1.00636744e+00 7.39756823e-01 7.89358139e-01 2.94908047e-01
-5.77312171e-01 2.06452206e-01 2.43109345e-01 8.27661529e-02
6.77039027e-01 9.09360051e-01 -9.53503549e-01 3.66995066e-01
-5.51654577e-01 -4.64063771e-02 8.20842981e-01 7.47108102e-01
6.62937164e-01 -1.98833272e-01 -1.06535554e-01 6.30398750e-01
3.41255844e-01 -3.26233581e-02 1.93193421e-01 -2.07505785e-02
4.66654271e-01 8.38395596e-01 -1.99128136e-01 -7.02853084e-01
-6.32228792e-01 -6.66404188e-01 -7.72248924e-01 -3.33799213e-01
1.71689481e-01 1.27061680e-01 -1.10511994e+00 1.64467144e+00
4.88224745e-01 5.56027353e-01 2.29990736e-01 3.34458649e-01
1.15284014e+00 6.12135530e-01 1.04119456e+00 -5.74503839e-02
1.18535292e+00 -8.84525836e-01 -6.10159576e-01 -2.48867143e-02
8.03144634e-01 -2.95777798e-01 5.27353942e-01 3.53724062e-01
-6.70539200e-01 -3.95634413e-01 -1.30439949e+00 -5.23575433e-02
-8.73001158e-01 -9.19793546e-03 1.01464295e+00 4.11760181e-01
-5.84693193e-01 1.28323543e+00 -6.77333713e-01 4.49755788e-02
8.88280451e-01 6.48990750e-01 -7.89337933e-01 -4.74803209e-01
-1.43216109e+00 1.04293931e+00 5.69381118e-01 -1.48666084e-01
-9.49163616e-01 -1.41903925e+00 -1.00713062e+00 5.83059862e-02
5.16515374e-01 -1.05784321e+00 6.84202969e-01 -9.78463739e-02
-1.37944102e+00 4.53810692e-01 -9.00875777e-02 -4.49664384e-01
8.94863345e-03 -2.04213802e-02 -4.95400518e-01 2.16649011e-01
-1.98787041e-02 3.24851185e-01 5.32132208e-01 -1.01992679e+00
-1.20452523e-01 -5.18939912e-01 -2.78130323e-02 4.91333492e-02
-2.53158778e-01 -4.01580095e-01 -1.28650203e-01 -4.38103855e-01
-4.50291663e-01 -6.66663170e-01 -5.61399102e-01 -3.60095978e-01
-6.90318525e-01 -3.08306754e-01 3.52887571e-01 -5.31977713e-01
1.15883088e+00 -1.74435115e+00 4.45879370e-01 2.90255278e-01
5.57365894e-01 3.42807204e-01 -3.95526409e-01 8.45369756e-01
-4.55496818e-01 3.23547065e-01 -1.88150167e-01 -7.03762397e-02
-2.70675510e-01 -1.57463148e-01 -5.88532053e-02 3.63714963e-01
5.45360148e-01 1.15512168e+00 -1.18686759e+00 -2.99671501e-01
2.90885456e-02 7.06446409e-01 -4.44854975e-01 1.56649128e-01
-7.38600612e-01 6.09420776e-01 -8.68808448e-01 9.55526710e-01
5.29182553e-01 -8.26103389e-01 7.12278068e-01 -6.64097190e-01
3.03602129e-01 6.11584663e-01 -6.78003132e-01 1.78449512e+00
-3.09502602e-01 -3.02166581e-01 -4.98691618e-01 -1.00227392e+00
5.53546906e-01 2.88346320e-01 5.43957829e-01 -5.76521337e-01
-5.76338433e-02 5.28990477e-02 1.50743211e-02 -9.76866707e-02
-5.00009768e-02 -1.37786567e-01 2.31366783e-01 4.71330322e-02
5.65532982e-01 7.74630681e-02 1.98315546e-01 3.10482830e-01
1.36656153e+00 3.53765815e-01 9.11412895e-01 -1.56072423e-01
8.10978234e-01 -1.19538404e-01 3.32872182e-01 3.01166773e-01
1.63237542e-01 -3.61867598e-03 5.51920772e-01 -3.82380456e-01
-4.50031400e-01 -7.07417309e-01 -2.53694266e-01 1.18999076e+00
-6.78571835e-02 -8.24610054e-01 -4.06737149e-01 -1.10206270e+00
4.16493922e-01 4.16850448e-01 -9.32136834e-01 -5.20251989e-01
-4.41183290e-03 -9.20778334e-01 3.76703262e-01 4.96030927e-01
-1.14162816e-02 -7.50774145e-01 5.44274390e-01 5.50878704e-01
3.35643381e-01 -1.27082384e+00 -2.07134619e-01 5.71104825e-01
-8.89350653e-01 -1.41439104e+00 -2.31662914e-01 -5.29729486e-01
5.27954340e-01 2.33230460e-02 1.21833050e+00 -9.66641027e-03
-4.44748372e-01 -9.95981097e-02 -1.86160892e-01 -5.73980689e-01
-4.87310529e-01 3.64261955e-01 -1.56267002e-01 -1.71833247e-01
1.39444843e-01 -7.61856079e-01 -7.36426413e-01 -1.32833809e-01
-8.29164147e-01 -1.80715501e-01 8.18014383e-01 6.55379295e-01
9.46963966e-01 -1.09528720e-01 9.76426601e-01 -1.43060911e+00
6.30078137e-01 -7.42625237e-01 -4.34683114e-01 5.56164443e-01
-6.87014580e-01 5.44905901e-01 9.25991595e-01 -2.18576774e-01
-7.46556640e-01 3.08573097e-01 -2.28804991e-01 -1.74921587e-01
-1.38511077e-01 1.01624107e+00 -4.86660600e-01 -5.40720820e-01
6.58341408e-01 2.58553643e-02 -3.25985134e-01 -5.84785402e-01
6.28935218e-01 2.56573915e-01 3.30827057e-01 -7.97493875e-01
6.05251372e-01 5.69703914e-02 6.11309707e-01 -5.27568042e-01
-1.18748963e+00 -6.01051569e-01 -6.94612086e-01 3.71443391e-01
7.60185599e-01 -1.10040367e+00 -8.32769752e-01 2.55229566e-02
-9.89549160e-01 -1.60937488e-01 1.61892846e-01 2.94521362e-01
-3.56520236e-01 4.95997727e-01 -5.56871474e-01 -2.00299293e-01
-5.15598893e-01 -1.00877821e+00 1.15701532e+00 -1.06375501e-01
-3.13417912e-02 -1.51491332e+00 4.81227010e-01 4.56995577e-01
1.25520006e-01 6.19330287e-01 1.31105745e+00 -9.15498137e-01
-4.87881243e-01 -1.12501562e-01 -2.54096985e-01 1.32418834e-02
8.07202876e-01 -1.51717290e-01 -1.04676664e+00 -1.37116373e-01
-7.92648137e-01 -5.38394034e-01 1.32954490e+00 2.00571492e-01
1.28282166e+00 -2.54835814e-01 -7.63572037e-01 7.37199187e-01
1.40847540e+00 9.79174860e-03 5.13969362e-01 7.35451281e-02
1.14306378e+00 3.33938658e-01 4.69946295e-01 3.37259501e-01
4.20702159e-01 4.32198435e-01 5.08495212e-01 -1.09611824e-01
-4.24323268e-02 -5.56022763e-01 8.91424417e-02 2.67196506e-01
-3.43296200e-01 -3.79820943e-01 -7.09549248e-01 2.21408024e-01
-1.81273949e+00 -8.81326675e-01 1.76978391e-02 1.88612890e+00
1.46373224e+00 -4.93331216e-02 -4.71174717e-02 -1.63298443e-01
3.74135107e-01 1.25529364e-01 -9.63087797e-01 -1.91571087e-01
9.19584930e-02 6.96411371e-01 6.20617509e-01 4.05812651e-01
-1.41282284e+00 1.23335993e+00 6.66134977e+00 1.06809580e+00
-1.01467299e+00 -3.85374665e-01 6.59325600e-01 4.36013371e-01
-3.07677895e-01 -1.88798502e-01 -9.15413797e-01 2.68893629e-01
1.42946446e+00 -4.13372785e-01 2.03650087e-01 7.22500622e-01
4.63956269e-03 2.78179973e-01 -1.63003695e+00 6.07687950e-01
-1.63156882e-01 -1.96372461e+00 5.14507294e-01 2.11731806e-01
7.39597082e-01 7.34936595e-02 -3.83956172e-02 7.85134286e-02
5.89149892e-01 -1.28095853e+00 -3.89584489e-02 6.03482723e-01
8.51788521e-01 -5.57752728e-01 5.63188255e-01 8.23151618e-02
-1.36196578e+00 2.85583526e-01 -3.23059261e-01 3.30589950e-01
-1.72227606e-01 7.97023118e-01 -1.27597106e+00 1.45696223e+00
-6.25419617e-02 1.47169340e+00 -6.87554479e-01 8.24566901e-01
-3.94464463e-01 4.84324634e-01 -9.58659872e-03 1.62326425e-01
9.07034352e-02 -2.25570485e-01 -1.02688789e-01 1.38738716e+00
-9.49470699e-02 2.34744295e-01 2.20771372e-01 7.48128593e-01
-5.36325753e-01 3.02394450e-01 -7.11718559e-01 -7.02865541e-01
3.01013887e-01 1.31238472e+00 -3.69075447e-01 -4.72245127e-01
-4.56109524e-01 6.59127414e-01 6.69176280e-01 2.84526199e-01
-7.41190672e-01 -9.87183303e-02 8.62294436e-01 -7.75048211e-02
3.83097649e-01 -1.11752547e-01 1.15144590e-03 -1.10189342e+00
-6.04084492e-01 -8.79267693e-01 6.83494389e-01 -3.27344507e-01
-1.68458855e+00 3.95076215e-01 -2.56882757e-01 -8.88898492e-01
4.28792415e-03 -1.03810537e+00 -2.75741726e-01 8.21165562e-01
-2.10429311e+00 -1.37805831e+00 6.31619692e-02 5.55181801e-01
-4.42761071e-02 1.76725537e-02 1.14446616e+00 2.72441059e-01
-7.83916295e-01 5.02541780e-01 1.94884866e-01 -3.10412973e-01
8.92517328e-01 -1.27603102e+00 4.26559001e-01 1.90541759e-01
5.89635707e-02 9.79514539e-01 3.26171577e-01 -9.77426350e-01
-1.64901447e+00 -1.50867987e+00 5.67684829e-01 -6.49613857e-01
9.45175588e-01 -2.96904266e-01 -1.03193796e+00 3.86802942e-01
-2.06197098e-01 2.35527858e-01 1.50034535e+00 3.51313949e-01
-7.75650918e-01 4.11738232e-02 -1.05871773e+00 1.64247438e-01
1.13637936e+00 -5.92213154e-01 -3.28157544e-01 6.62112772e-01
8.08836043e-01 -1.93262458e-01 -1.59954238e+00 6.87642932e-01
4.72544432e-01 -2.57432729e-01 1.21627223e+00 -1.22386575e+00
5.19239247e-01 -1.79259822e-01 5.74792735e-02 -1.45926487e+00
-6.22406423e-01 -4.28419560e-01 -4.32285041e-01 5.78868568e-01
8.39139640e-01 -5.64676762e-01 7.59921730e-01 4.06608433e-01
-2.16411486e-01 -1.06306005e+00 -4.73248303e-01 -6.65949225e-01
1.42672576e-03 8.62600002e-03 5.36163211e-01 1.11506176e+00
3.26435834e-01 7.98338294e-01 -2.57747173e-01 3.07120770e-01
5.78903913e-01 1.70284763e-01 4.20311719e-01 -1.50865829e+00
-5.84199071e-01 -3.05329740e-01 -6.78750396e-01 -6.56469166e-01
4.56675678e-01 -1.16703343e+00 -4.98601198e-01 -1.86144865e+00
5.11857212e-01 5.53536713e-02 -9.61337388e-01 9.04006958e-01
-5.33125639e-01 5.99879958e-02 -2.78082043e-01 -7.17606843e-02
-8.39895725e-01 6.55942857e-01 1.41222382e+00 -5.14577568e-01
-3.52156647e-02 -2.62675971e-01 -1.14877892e+00 3.08833748e-01
5.49146771e-01 -4.29000080e-01 -5.61657310e-01 8.53538513e-02
3.22764039e-01 -1.74624681e-01 3.61710601e-02 -7.13384867e-01
1.37695223e-01 -4.51202154e-01 7.07824647e-01 -2.89921075e-01
3.50329369e-01 -3.90070081e-01 1.32874638e-01 5.04355490e-01
-3.42475981e-01 -5.68941951e-01 4.43550915e-01 1.22707987e+00
-1.39197260e-01 2.49623224e-01 6.22141063e-01 -1.07417233e-01
-4.34258491e-01 8.37034702e-01 7.18556568e-02 -2.49007195e-01
9.32331860e-01 6.04570508e-02 -5.53895354e-01 -1.64444745e-02
-8.70544791e-01 2.81985641e-01 2.71462291e-01 1.86690599e-01
6.26228750e-01 -1.04243362e+00 -3.63381326e-01 -1.44782126e-01
4.22463566e-01 -2.04839095e-01 1.70910388e-01 5.27478278e-01
-2.59442836e-01 7.62359977e-01 -1.28410771e-01 -5.61780520e-02
-1.16240382e+00 7.76267469e-01 2.28487968e-01 -7.41910100e-01
-1.19914092e-01 8.00784409e-01 5.36976635e-01 -2.77287632e-01
-4.03314829e-03 -3.88328880e-01 -5.96223772e-01 -3.54962004e-03
3.64607573e-01 6.91538677e-02 3.93184870e-01 -2.16603041e-01
-6.93902612e-01 4.80654240e-01 -4.70019847e-01 8.15722346e-01
1.82877982e+00 6.53398395e-01 -1.77353427e-01 7.08480254e-02
1.18484616e+00 -1.02471434e-01 -9.72714424e-01 -2.46091411e-01
2.03583732e-01 -9.10272729e-03 8.04355517e-02 -1.21006811e+00
-6.18684888e-01 6.28170729e-01 -5.25453538e-02 -3.37266028e-01
8.93754005e-01 1.83055922e-01 4.50136930e-01 8.67133021e-01
2.68696815e-01 -7.03656316e-01 2.25118324e-01 2.42917880e-01
7.55746841e-01 -1.16524065e+00 6.42671406e-01 -9.83336747e-01
-4.41192985e-01 1.08492172e+00 4.40873653e-01 8.13622028e-02
6.46274626e-01 -1.04391508e-01 -5.01210868e-01 -5.38385212e-01
-1.20717752e+00 -2.89573610e-01 8.94299984e-01 6.33129060e-01
9.17264879e-01 1.46303819e-02 -1.58687040e-01 7.50130892e-01
3.42627853e-01 5.69157489e-02 4.94407602e-02 9.05238211e-01
-3.47343326e-01 -1.66490495e+00 2.70318449e-01 5.09092391e-01
-5.87464333e-01 -5.67096233e-01 -1.03241324e+00 5.41178882e-01
-8.87473673e-02 7.30826139e-01 -8.63150656e-01 -4.75482285e-01
2.86634594e-01 1.96155682e-01 8.52446735e-01 -9.27387536e-01
-4.68831092e-01 -9.64236483e-02 4.11507070e-01 -6.95471764e-01
-5.67380965e-01 -1.68811738e-01 -1.20147443e+00 -2.69129202e-02
-3.43246371e-01 3.15023184e-01 5.66365719e-01 8.50840211e-01
9.33810532e-01 8.29385161e-01 5.61263978e-01 -7.01327324e-01
-4.70749378e-01 -8.27451527e-01 -4.82328683e-01 3.06019157e-01
1.36468485e-01 -6.40783012e-01 5.66008948e-02 1.37230471e-01] | [5.103039741516113, 5.89339017868042] |
d5567898-8bad-4f37-887c-470a7a7d10f3 | learning-word-embeddings-from-speech | 1711.01515 | null | http://arxiv.org/abs/1711.01515v1 | http://arxiv.org/pdf/1711.01515v1.pdf | Learning Word Embeddings from Speech | In this paper, we propose a novel deep neural network architecture,
Sequence-to-Sequence Audio2Vec, for unsupervised learning of fixed-length
vector representations of audio segments excised from a speech corpus, where
the vectors contain semantic information pertaining to the segments, and are
close to other vectors in the embedding space if their corresponding segments
are semantically similar. The design of the proposed model is based on the RNN
Encoder-Decoder framework, and borrows the methodology of continuous skip-grams
for training. The learned vector representations are evaluated on 13 widely
used word similarity benchmarks, and achieved competitive results to that of
GloVe. The biggest advantage of the proposed model is its capability of
extracting semantic information of audio segments taken directly from raw
speech, without relying on any other modalities such as text or images, which
are challenging and expensive to collect and annotate. | ['Yu-An Chung', 'James Glass'] | 2017-11-05 | null | null | null | null | ['learning-word-embeddings'] | ['methodology'] | [ 4.52470660e-01 -6.16789237e-02 -1.19733579e-01 -3.86416286e-01
-8.81628394e-01 -4.23494756e-01 3.00626427e-01 2.17502266e-01
-5.98307192e-01 4.20851469e-01 7.23933697e-01 -1.71177402e-01
1.84505597e-01 -5.08167267e-01 -5.43605804e-01 -7.02573836e-01
-6.56224042e-02 4.24103327e-02 3.48331444e-02 -1.11851722e-01
9.54893827e-02 1.44620344e-01 -1.81550038e+00 2.28535995e-01
2.29647204e-01 1.30369830e+00 3.43303382e-01 7.97996640e-01
-3.28682482e-01 7.85635948e-01 -5.73945045e-01 -1.55530423e-01
2.36154608e-02 -4.62020069e-01 -7.54754305e-01 7.13447854e-02
1.08867511e-01 -3.44909757e-01 -7.19648242e-01 9.26207423e-01
6.81444526e-01 4.85922664e-01 7.64269352e-01 -9.17201340e-01
-7.33148277e-01 8.01294088e-01 2.73917429e-02 2.38796964e-01
3.63624007e-01 -1.76058769e-01 1.46578860e+00 -9.38336551e-01
3.57770562e-01 1.00226676e+00 4.83285725e-01 4.35281187e-01
-7.71524727e-01 -5.25393546e-01 2.09336355e-02 5.45115650e-01
-1.62745965e+00 -6.71988666e-01 8.74062896e-01 -5.76831043e-01
1.00243628e+00 2.03489333e-01 3.60438824e-01 1.39800227e+00
-8.35093483e-02 7.35581875e-01 3.33808810e-01 -5.21840155e-01
3.59218955e-01 1.99244067e-01 -1.44868745e-02 1.82556987e-01
-3.39914203e-01 -8.87566656e-02 -2.82792985e-01 -1.13947682e-01
5.48846304e-01 2.00345621e-01 -4.50015783e-01 -3.07383627e-01
-1.03334045e+00 9.30496633e-01 4.50426787e-01 5.86715221e-01
-4.58972633e-01 1.49235770e-01 9.20687854e-01 2.62835979e-01
3.28946114e-01 1.72389522e-01 -2.59053767e-01 -3.98479044e-01
-8.83706987e-01 -1.67498052e-01 3.18797857e-01 9.31294918e-01
4.19480443e-01 6.50011182e-01 -7.19460919e-02 1.08932102e+00
3.98347199e-01 4.45631325e-01 1.21414518e+00 -6.06212258e-01
5.36738455e-01 3.03218603e-01 -2.38261670e-01 -1.06824505e+00
-8.09002668e-02 -3.34632009e-01 -7.70034075e-01 -3.36410314e-01
-1.89250082e-01 -6.10155985e-02 -8.03239763e-01 1.58760726e+00
-1.57335121e-02 4.36110377e-01 3.01049680e-01 1.06265628e+00
1.30658853e+00 1.01923978e+00 -6.53273612e-03 -1.82694606e-02
1.20362401e+00 -1.01332605e+00 -8.44899476e-01 -1.99528739e-01
5.51226318e-01 -7.27603674e-01 1.10616946e+00 2.19773322e-01
-9.46764290e-01 -8.02109897e-01 -1.11203861e+00 -8.64580199e-02
-4.80132312e-01 3.37446295e-02 8.65429044e-02 2.94094890e-01
-1.01571774e+00 5.77396512e-01 -4.97985125e-01 -3.95049065e-01
7.85593837e-02 4.25342023e-02 -5.46175480e-01 1.60402492e-01
-1.40082121e+00 4.74037200e-01 7.35693872e-01 2.35243402e-02
-1.13809812e+00 -2.69582927e-01 -1.23675215e+00 4.58510280e-01
1.76043510e-01 -1.23242229e-01 1.37311661e+00 -1.11810267e+00
-1.45828474e+00 4.96037930e-01 -1.62486315e-01 -6.69891357e-01
-1.23710722e-01 -2.39953116e-01 -8.28566968e-01 3.73760909e-01
-2.13193923e-01 4.85982031e-01 1.22977579e+00 -9.32377160e-01
-3.60678613e-01 -6.42357320e-02 5.54983355e-02 1.04144759e-01
-9.69551504e-01 1.50746986e-01 -3.95842135e-01 -9.92153406e-01
-2.11109892e-01 -6.67846739e-01 -2.42591769e-01 -1.98048815e-01
-2.73180991e-01 -2.20407054e-01 7.90717244e-01 -9.11218226e-01
1.44168580e+00 -2.40774012e+00 2.06062376e-01 1.78834982e-02
5.47126085e-02 5.40776908e-01 -5.40062726e-01 6.76406264e-01
-3.05444628e-01 -2.07461670e-01 -2.34426692e-01 -1.98856503e-01
9.81231704e-02 2.56585985e-01 -7.58784056e-01 1.92112654e-01
8.78741220e-02 6.71804488e-01 -9.49537039e-01 -4.17060643e-01
1.71274006e-01 7.07657397e-01 -4.06389028e-01 5.21190584e-01
-6.86894581e-02 1.88877657e-01 -2.07422197e-01 2.68379867e-01
2.39692137e-01 8.67391080e-02 5.22406846e-02 -1.56965688e-01
9.14262235e-02 7.96784878e-01 -9.81706738e-01 1.98547649e+00
-6.37257397e-01 6.17948115e-01 -1.09741643e-01 -1.41146994e+00
1.12047434e+00 8.21820378e-01 3.13978672e-01 -5.71738720e-01
3.53715837e-01 5.02922870e-02 -1.41964689e-01 -7.63315320e-01
5.05791664e-01 -1.50877833e-01 -2.56634057e-01 4.93392140e-01
6.12609446e-01 -1.61521375e-01 -3.16660404e-02 1.19761350e-02
9.51825500e-01 -2.34888762e-01 3.53269130e-01 3.22383732e-01
8.74872923e-01 -6.44558907e-01 4.68955249e-01 3.42646152e-01
-2.35236045e-02 6.87010586e-01 1.08705133e-01 -2.01965049e-01
-1.10203636e+00 -1.14899313e+00 3.58217359e-02 1.26627350e+00
-2.22146332e-01 -7.33683169e-01 -6.47504508e-01 -5.57012618e-01
-2.01503560e-01 6.94287837e-01 -3.57113361e-01 -3.74396533e-01
-2.89185733e-01 2.24021852e-01 8.21700633e-01 9.00238752e-01
-1.34020224e-01 -1.16848743e+00 -2.98300683e-01 4.46062803e-01
-3.73965532e-01 -1.22556877e+00 -5.64855039e-01 1.94492891e-01
-6.70088887e-01 -5.89456499e-01 -9.30740535e-01 -1.08912408e+00
2.56266117e-01 2.87544072e-01 8.27805877e-01 -2.87912697e-01
-2.77715892e-01 4.39667284e-01 -8.53125513e-01 -4.08219278e-01
-2.27258116e-01 8.63953233e-02 2.54410982e-01 4.13217813e-01
3.87114763e-01 -8.56424153e-01 -4.41397876e-01 1.83331773e-01
-1.07443821e+00 -4.47411418e-01 4.29885417e-01 7.87262857e-01
7.99855113e-01 -4.52985056e-02 7.18026221e-01 -2.08769485e-01
6.30760312e-01 -4.83997554e-01 -1.18205465e-01 -1.49660751e-01
-1.94402143e-01 -1.93216875e-02 8.44002724e-01 -4.73269284e-01
-5.13162673e-01 -8.18623975e-02 -4.47168767e-01 -9.09897208e-01
-3.49516362e-01 5.62321603e-01 -2.28433356e-01 2.83116609e-01
2.29735926e-01 5.52495182e-01 -8.41909572e-02 -8.63028705e-01
5.63282192e-01 1.29183698e+00 7.66338944e-01 -2.38146290e-01
5.80139101e-01 1.99619383e-01 -7.17261314e-01 -1.03507769e+00
-6.40479088e-01 -7.77731001e-01 -5.42563856e-01 -1.19892009e-01
8.92978787e-01 -9.18944180e-01 -2.62561113e-01 4.35955860e-02
-1.25721478e+00 2.65335977e-01 -5.05533516e-01 9.10103500e-01
-5.55679321e-01 5.76954722e-01 -6.34291291e-01 -8.41194212e-01
-5.61201572e-01 -1.02210796e+00 8.97494316e-01 2.77766176e-02
-3.83066237e-01 -9.78918314e-01 1.44410625e-01 2.61867493e-01
4.87183183e-01 -2.05102324e-01 9.71415937e-01 -1.21908498e+00
-4.61660698e-02 -5.05877078e-01 5.32957725e-02 9.34789717e-01
1.84573576e-01 -2.47374937e-01 -1.42450714e+00 -3.87303412e-01
5.50133660e-02 -5.09647727e-01 8.28523874e-01 1.02861673e-01
1.65918076e+00 -5.63968420e-01 5.83967790e-02 4.93914783e-01
1.19697976e+00 4.03328300e-01 5.56006968e-01 5.35059758e-02
5.97547472e-01 6.02096975e-01 4.18355584e-01 5.88201761e-01
2.16497898e-01 7.24868715e-01 4.96575773e-01 1.09083124e-01
2.16943622e-02 -4.92509186e-01 5.37638009e-01 1.82637250e+00
4.92196269e-02 -2.07460612e-01 -5.80654502e-01 1.01715159e+00
-1.60622025e+00 -9.76843297e-01 2.99267381e-01 2.23529267e+00
7.39573002e-01 4.81733903e-02 2.91722342e-02 6.03680730e-01
7.81951964e-01 4.27824110e-01 -2.67240435e-01 -6.59816742e-01
1.18992724e-01 5.29012203e-01 -1.98317245e-02 2.34366149e-01
-1.14944077e+00 7.85872340e-01 6.09737587e+00 1.11381161e+00
-1.22093904e+00 2.50227302e-01 -3.10044177e-02 -2.84897476e-01
-3.61228585e-01 -2.99646348e-01 -3.55300903e-01 4.72269684e-01
1.41465926e+00 -6.22254424e-02 2.69302100e-01 9.60658729e-01
1.00427516e-01 6.19917810e-01 -1.19000292e+00 1.16634786e+00
2.24509820e-01 -1.03954673e+00 3.13807398e-01 -3.42548639e-01
3.79112363e-01 -3.09181903e-02 1.34294808e-01 4.04998660e-01
-3.92340273e-02 -1.15804911e+00 6.44043744e-01 3.45481515e-01
8.48089933e-01 -9.36640859e-01 9.71181393e-01 1.21557526e-01
-1.39540923e+00 -3.52153704e-02 -6.85021758e-01 -3.73137593e-02
3.05386186e-01 3.78548235e-01 -1.03107345e+00 6.31613851e-01
6.83960557e-01 8.50708902e-01 -2.08660081e-01 9.28009629e-01
-2.77109057e-01 9.30548847e-01 6.87630624e-02 -7.83844516e-02
5.06033123e-01 1.60725545e-02 4.72743154e-01 1.47368062e+00
5.01868069e-01 -3.07650238e-01 4.05881070e-02 5.69229007e-01
-2.91276664e-01 5.28050721e-01 -8.43253136e-01 -6.34946525e-01
5.78170598e-01 1.16672194e+00 -1.65610164e-01 -3.63242656e-01
-6.08210087e-01 8.83767486e-01 1.22773051e-01 3.83319974e-01
-8.19851160e-01 -1.18864155e+00 9.30659950e-01 -1.07694462e-01
8.29462588e-01 -8.61216933e-02 1.75114527e-01 -9.80075836e-01
1.64570719e-01 -8.11100304e-01 2.70125508e-01 -9.25640643e-01
-1.35222089e+00 9.67611909e-01 -3.03969890e-01 -1.71742189e+00
-6.47023380e-01 -3.50389063e-01 -6.38080060e-01 7.13250756e-01
-1.43888807e+00 -8.16894889e-01 1.78745890e-03 8.65691960e-01
6.41604900e-01 -5.78518033e-01 1.19688654e+00 5.05223572e-01
-1.79965153e-01 9.30498600e-01 3.85540783e-01 5.18297791e-01
5.75378180e-01 -7.37662435e-01 4.97880638e-01 5.38280845e-01
6.83296859e-01 5.37319303e-01 3.59097749e-01 -6.47154301e-02
-1.05166996e+00 -1.23236227e+00 1.14976108e+00 2.38196328e-02
9.19932604e-01 -5.70266962e-01 -1.08570063e+00 5.95149636e-01
2.91828215e-01 -1.23195462e-01 1.12580061e+00 -1.42819211e-01
-6.48636043e-01 -1.81182757e-01 -6.63639069e-01 2.74831414e-01
9.33197200e-01 -1.23904788e+00 -1.08714914e+00 1.73595220e-01
1.18104315e+00 -4.27705459e-02 -8.70206594e-01 2.02699631e-01
3.81819487e-01 -5.39894640e-01 1.18172240e+00 -8.69298100e-01
6.16609871e-01 -2.84093291e-01 -6.68643177e-01 -1.39061272e+00
-1.61008149e-01 -3.47009540e-01 -2.66147237e-02 1.45806837e+00
2.41215840e-01 -3.62668127e-01 3.76449317e-01 -2.16368198e-01
-3.58525217e-01 -5.72032034e-01 -1.01186275e+00 -9.91457582e-01
-2.71700770e-01 -8.37691844e-01 7.92035222e-01 9.11578715e-01
1.72920555e-01 4.88169968e-01 -5.04359305e-01 1.23999692e-01
1.82131276e-01 3.69852372e-02 5.36595941e-01 -1.14602971e+00
-4.33174759e-01 -1.59521401e-01 -1.04142356e+00 -1.10838270e+00
5.30466914e-01 -1.14825070e+00 1.24944873e-01 -1.26977503e+00
-2.53322661e-01 4.96347956e-02 -7.40876317e-01 4.01447684e-01
2.40342528e-01 3.67103875e-01 2.06401885e-01 6.89243823e-02
-5.42290866e-01 1.16885448e+00 7.57477164e-01 -4.78033245e-01
2.40197685e-02 -1.01369500e-01 -4.51452345e-01 8.49435925e-01
6.11056447e-01 -4.44809139e-01 -5.90699494e-01 -6.24373972e-01
-3.09481025e-01 6.77257329e-02 1.86695486e-01 -1.04918396e+00
3.96172553e-02 2.66935855e-01 5.42070754e-02 -5.88594496e-01
6.61057770e-01 -9.75369036e-01 -1.35157481e-01 1.40787229e-01
-7.60912240e-01 -1.17162801e-01 1.08221278e-01 7.09933460e-01
-8.08187842e-01 -5.93556464e-01 4.96907026e-01 1.73156187e-01
-8.24316025e-01 2.10550949e-01 -5.32653391e-01 3.69584337e-02
8.13458920e-01 -4.36227769e-02 2.92542070e-01 -8.03961158e-01
-7.43193626e-01 -2.08779275e-01 5.15499115e-02 7.45936453e-01
9.86883402e-01 -1.87212873e+00 -5.84091425e-01 2.98730105e-01
4.99880075e-01 -3.09616506e-01 1.83814988e-01 4.39276099e-01
-1.18517533e-01 6.71218693e-01 -1.17735699e-01 -4.95197982e-01
-1.14613652e+00 8.17175329e-01 -1.04293965e-01 1.03903450e-01
-6.79351687e-01 9.05631065e-01 2.40643412e-01 -4.10763830e-01
6.53011143e-01 -4.65374678e-01 -5.60664177e-01 2.65389383e-01
8.93346250e-01 1.60860389e-01 2.50082642e-01 -1.00606239e+00
-2.95796573e-01 5.45690596e-01 -1.12526886e-01 -1.80403918e-01
1.43255675e+00 -1.65477768e-01 8.77748877e-02 8.37288857e-01
1.62586486e+00 -1.68409392e-01 -7.19601512e-01 -6.39788151e-01
-1.37082696e-01 -2.84214973e-01 1.56538099e-01 -9.33561623e-02
-1.20817137e+00 1.36723602e+00 5.44754863e-01 6.78000450e-02
9.38947976e-01 2.85725985e-02 1.29421353e+00 3.92244071e-01
1.28725335e-01 -1.22593951e+00 9.96475890e-02 6.95200145e-01
8.61916363e-01 -7.66016543e-01 -5.27540147e-01 1.13384850e-01
-7.63309062e-01 1.21834671e+00 2.15706110e-01 -1.50973544e-01
6.19746149e-01 -2.07636915e-02 1.25994056e-01 1.30735680e-01
-7.25324273e-01 -3.41454625e-01 4.90657061e-01 8.08359802e-01
5.06720722e-01 -6.01976812e-02 -7.67070800e-02 1.01617503e+00
-3.34697962e-01 -2.85445333e-01 3.01640749e-01 7.60652304e-01
-4.46978211e-01 -8.58722150e-01 -2.53268033e-01 2.50811160e-01
-4.39896882e-01 -2.96128720e-01 -3.56384099e-01 1.43196926e-01
-1.49776697e-01 1.04053354e+00 3.39735627e-01 -7.82201231e-01
3.89124006e-01 3.66286725e-01 -3.50207351e-02 -5.49237728e-01
-4.49158251e-01 2.47518480e-01 3.90439108e-02 -4.30435359e-01
-2.82328904e-01 -2.96135247e-01 -1.23405814e+00 1.79067582e-01
-3.19049269e-01 4.93363321e-01 7.68456161e-01 8.97464335e-01
3.77652884e-01 6.87085092e-01 1.01128137e+00 -1.01015604e+00
-6.49038970e-01 -1.17296648e+00 -7.61881173e-01 5.31440079e-01
5.25024652e-01 -5.23976922e-01 -3.53866130e-01 3.42740446e-01] | [15.138367652893066, 5.090996742248535] |
b0ad9b2d-024f-4804-a0b4-090e438882d4 | a-decentralized-spike-based-learning | 2305.16748 | null | https://arxiv.org/abs/2305.16748v1 | https://arxiv.org/pdf/2305.16748v1.pdf | A Decentralized Spike-based Learning Framework for Sequential Capture in Discrete Perimeter Defense Problem | This paper proposes a novel Decentralized Spike-based Learning (DSL) framework for the discrete Perimeter Defense Problem (d-PDP). A team of defenders is operating on the perimeter to protect the circular territory from radially incoming intruders. At first, the d-PDP is formulated as a spatio-temporal multi-task assignment problem (STMTA). The problem of STMTA is then converted into a multi-label learning problem to obtain labels of segments that defenders have to visit in order to protect the perimeter. The DSL framework uses a Multi-Label Classifier using Synaptic Efficacy Function spiking neuRON (MLC-SEFRON) network for deterministic multi-label learning. Each defender contains a single MLC-SEFRON network. Each MLC-SEFRON network is trained independently using input from its own perspective for decentralized operations. The input spikes to the MLC-SEFRON network can be directly obtained from the spatio-temporal information of defenders and intruders without any extra pre-processing step. The output of MLC-SEFRON contains the labels of segments that a defender has to visit in order to protect the perimeter. Based on the multi-label output from the MLC-SEFRON a trajectory is generated for a defender using a Consensus-Based Bundle Algorithm (CBBA) in order to capture the intruders. The target multi-label output for training MLC-SEFRON is obtained from an expert policy. Also, the MLC-SEFRON trained for a defender can be directly used for obtaining labels of segments assigned to another defender without any retraining. The performance of MLC-SEFRON has been evaluated for full observation and partial observation scenarios of the defender. The overall performance of the DSL framework is then compared with expert policy along with other existing learning algorithms. The scalability of the DSL has been evaluated using an increasing number of defenders. | ['Shirin Dora', 'Suresh Sundaram', 'Shridhar Velhal', 'Mohammed Thousif'] | 2023-05-26 | null | null | null | null | ['multi-label-learning'] | ['methodology'] | [ 1.99827775e-01 -2.03069866e-01 2.72734553e-01 1.87455386e-01
-7.13165998e-01 -1.23526394e+00 2.55963832e-01 2.87679434e-01
-7.25551486e-01 8.03517878e-01 -7.90236235e-01 -1.34371355e-01
-4.67901021e-01 -7.48133838e-01 -6.18457496e-01 -1.43227446e+00
-1.97687685e-01 8.59305322e-01 5.19089162e-01 -1.30435020e-01
5.58102369e-01 9.40346360e-01 -1.18783987e+00 4.01061147e-01
1.64363772e-01 1.32582819e+00 -1.05084002e-01 5.47209978e-01
4.41660643e-01 1.51429623e-01 -9.05345559e-01 2.66803920e-01
9.24931049e-01 -1.14375047e-01 -3.25750887e-01 -7.95323104e-02
-1.42753869e-01 2.44026527e-01 2.00244591e-01 9.86553848e-01
2.83106714e-01 -5.82420081e-02 8.77408326e-01 -1.59656537e+00
3.31041336e-01 1.52059570e-01 -5.24270654e-01 3.50783139e-01
-2.93558687e-01 2.46656552e-01 7.11009383e-01 -1.47388875e-01
6.87585175e-01 9.01298523e-01 3.11993539e-01 5.38505018e-01
-1.61880684e+00 -8.65157187e-01 5.38918436e-01 1.14801265e-01
-1.31347239e+00 3.32806647e-01 7.04971790e-01 -3.77609372e-01
7.12287903e-01 1.97363973e-01 6.18858337e-01 1.28551698e+00
7.40488946e-01 3.04923981e-01 1.54410744e+00 -3.63583229e-02
8.78618777e-01 -6.76792786e-02 1.08362958e-01 3.22298408e-01
1.44399896e-01 5.45829892e-01 -4.22871441e-01 -3.88659686e-01
7.63532400e-01 -1.96995437e-01 -3.59643288e-02 -2.07470804e-01
-6.25333667e-01 8.61248851e-01 3.78643245e-01 4.06115621e-01
-3.80180597e-01 -1.43861607e-01 5.17215073e-01 6.14692867e-01
1.30562574e-01 4.79418486e-01 -3.16349655e-01 3.52077544e-01
-1.05904067e+00 2.65558898e-01 9.70967531e-01 2.50456363e-01
7.62707710e-01 1.87083840e-01 3.16835232e-02 2.05958709e-01
1.93632558e-01 7.51344502e-01 -1.60929516e-01 -1.00121641e+00
1.93808556e-01 7.04052269e-01 3.03238988e-01 -8.24278712e-01
-6.59706771e-01 -5.49481988e-01 -6.34427071e-01 1.38240230e+00
5.06095529e-01 -3.24328482e-01 -6.22884154e-01 1.80220902e+00
3.30157518e-01 3.87824535e-01 6.44021630e-02 8.50537956e-01
-3.21709841e-01 7.79862344e-01 -1.68505058e-01 -3.07480723e-01
9.16515648e-01 -5.29258430e-01 -1.80459693e-02 -1.09311245e-01
3.15301389e-01 -1.23451367e-01 5.10863885e-02 8.60985041e-01
-5.14147341e-01 -3.46009582e-02 -8.39761198e-01 1.07412791e+00
-6.02680087e-01 5.34772612e-02 -7.34066963e-02 2.76843667e-01
-8.51475775e-01 5.68143129e-01 -7.64134347e-01 -2.70124286e-01
2.97382236e-01 7.54609883e-01 -4.00570482e-01 5.89782596e-01
-9.28284168e-01 9.93369997e-01 3.53936344e-01 -1.24982465e-02
-1.93988490e+00 -1.14927702e-01 -4.28682297e-01 5.66140562e-02
3.28699231e-01 -1.37395233e-01 6.29237831e-01 -9.37408507e-01
-1.27023578e+00 7.34800518e-01 4.16720361e-01 -8.68076622e-01
2.51922727e-01 6.67178869e-01 -6.80014044e-02 4.20260936e-01
2.90776789e-01 4.23539072e-01 9.86928821e-01 -1.69137180e+00
-9.37482119e-01 -5.42531967e-01 7.97132254e-02 3.39926183e-01
1.53385505e-01 4.59621064e-02 4.15154815e-01 -7.54440904e-01
1.20955236e-01 -1.18175519e+00 4.47647050e-02 -2.18701318e-01
-2.99047559e-01 -1.25446290e-01 1.24645507e+00 -1.61788046e-01
7.75127769e-01 -1.82207417e+00 6.89496040e-01 5.00926316e-01
-1.77089237e-02 5.64724922e-01 -2.97791839e-01 7.93368638e-01
-2.84774438e-03 -1.53095946e-01 -5.81264138e-01 -4.11326468e-01
-3.63091052e-01 4.31485325e-01 -7.63436258e-01 7.37811148e-01
4.64344993e-02 1.31792799e-01 -5.16541123e-01 -8.73826370e-02
1.35791689e-01 5.90375997e-02 -8.59373435e-02 1.18251987e-01
-3.56817901e-01 8.70514333e-01 -5.90804100e-01 7.85188854e-01
5.78690648e-01 2.15117201e-01 -1.72788084e-01 4.16448802e-01
-5.24805307e-01 -3.56258512e-01 -1.22796226e+00 1.15870106e+00
-3.69605273e-01 1.18830524e-01 7.70443678e-01 -1.41640580e+00
1.07533193e+00 3.09471816e-01 7.58365691e-01 -2.86284000e-01
4.34942126e-01 2.86901414e-01 1.00437300e-02 1.42434523e-01
-5.47607124e-01 -2.68945277e-01 -5.85234344e-01 1.13264644e+00
1.70827538e-01 1.22857422e-01 9.62654352e-02 4.85709757e-02
1.54669642e+00 1.39700770e-01 -5.25460579e-02 -3.98890436e-01
7.97567487e-01 2.53694266e-01 9.70824599e-01 6.99199677e-01
-4.45113391e-01 -8.28067511e-02 4.58532244e-01 -7.18644261e-01
-6.35896027e-01 -1.51376593e+00 -1.62870705e-01 9.20626521e-01
5.27640283e-01 1.68458506e-01 -9.49690342e-01 -7.05046296e-01
4.46420237e-02 6.57589912e-01 -8.10371518e-01 -1.71823174e-01
-3.78516018e-01 -6.33875012e-01 8.18049788e-01 5.81950834e-03
4.26786244e-01 -1.53898501e+00 -1.33746028e+00 4.54689860e-01
2.46484175e-01 -6.37884378e-01 1.25325680e-01 1.09781981e+00
-5.40806592e-01 -1.20274353e+00 -1.30772680e-01 -5.84839284e-01
7.42265224e-01 -1.38248755e-02 9.43215340e-02 -1.69235155e-01
-2.59241998e-01 2.77067631e-01 -2.44509205e-01 -5.24657607e-01
-1.66062891e-01 -1.54274195e-01 5.31718373e-01 4.87629831e-01
7.24681169e-02 -1.02582741e+00 2.27276348e-02 6.68542445e-01
-9.35760677e-01 -2.66609758e-01 3.90098125e-01 7.97623098e-01
8.57200742e-01 2.32746884e-01 7.16629744e-01 -3.43186021e-01
4.67260808e-01 -4.33538884e-01 -1.26571667e+00 3.77085447e-01
-2.22707719e-01 2.25721732e-01 1.08598757e+00 -8.04043114e-01
-6.06234968e-01 1.71998814e-01 4.79263037e-01 -7.44559944e-01
-5.36571860e-01 -9.57767963e-02 6.22609444e-02 -5.42302191e-01
6.39037132e-01 3.53813380e-01 -9.25103500e-02 -2.11028621e-01
-2.01897487e-01 6.21183693e-01 4.84851897e-01 -7.50069737e-01
8.56502950e-01 6.76817477e-01 3.36986989e-01 -4.13157314e-01
-8.61427307e-01 -3.17727447e-01 -6.89739645e-01 -6.72872245e-01
8.56837749e-01 -4.67245698e-01 -1.09745526e+00 8.62356484e-01
-1.15329504e+00 -3.10592055e-01 -2.09043980e-01 6.63615689e-02
-7.12999582e-01 -2.83406377e-02 -5.78363180e-01 -1.15200949e+00
-4.11106944e-01 -1.26267564e+00 1.08559096e+00 3.82316142e-01
3.33734989e-01 -9.72954929e-01 5.05173981e-01 1.21621437e-01
1.74202979e-01 6.39056087e-01 6.79657757e-01 -1.14467192e+00
-4.89478976e-01 -3.44337285e-01 2.26160064e-01 2.63507456e-01
-3.02096695e-01 -3.47925931e-01 -7.39437819e-01 -7.66523421e-01
5.10330021e-01 -5.08645415e-01 7.11729348e-01 4.05924648e-01
5.32893658e-01 -1.33848473e-01 -4.74545687e-01 5.09339213e-01
1.63340807e+00 5.64810693e-01 -7.66161159e-02 8.50977302e-01
4.44402322e-02 6.74243748e-01 6.48501277e-01 4.37338412e-01
1.01342551e-01 6.08516991e-01 1.03441465e+00 3.08224678e-01
4.76390809e-01 3.52581084e-01 5.58811903e-01 1.16993850e-02
1.10203266e-01 -1.50263995e-01 -8.52321267e-01 1.63225815e-01
-1.96055126e+00 -1.13583446e+00 1.78442508e-01 2.38230443e+00
2.77616650e-01 3.13842297e-01 2.00336680e-01 1.30529508e-01
1.00899434e+00 -5.13333231e-02 -9.71785784e-01 -6.74045861e-01
-3.04820001e-01 2.48306155e-01 7.66152084e-01 4.09871548e-01
-1.28956163e+00 9.75161791e-01 4.81110859e+00 9.48927939e-01
-1.19931102e+00 -3.24116764e-03 2.69007355e-01 -3.61188978e-01
3.86413485e-01 3.62833500e-01 -1.24668658e+00 5.86141467e-01
1.06818724e+00 -9.35993940e-02 1.01989937e+00 5.21584332e-01
1.84663519e-01 -6.01128876e-01 -6.57359898e-01 8.11497211e-01
3.07068585e-05 -1.08067989e+00 -3.03510457e-01 5.21300435e-01
5.67240059e-01 2.52303958e-01 9.24890563e-02 1.32755071e-01
4.95562911e-01 -5.77696741e-01 6.61131263e-01 5.39158165e-01
3.14068943e-01 -8.51099789e-01 6.35205150e-01 1.09918833e+00
-1.29857147e+00 -7.22113132e-01 -3.08960617e-01 -3.88362110e-02
3.71454090e-01 1.26297683e-01 -3.04172456e-01 6.46180034e-01
6.86163723e-01 1.84407324e-01 -3.33682299e-01 9.10227716e-01
-4.36969012e-01 3.76992583e-01 -7.58443177e-01 2.72530079e-01
7.72342920e-01 -5.52494049e-01 1.02430928e+00 6.95737243e-01
9.09889936e-02 -3.70442681e-02 4.43657160e-01 7.80026495e-01
3.66942734e-01 -3.38552356e-01 -7.61884630e-01 3.19020718e-01
5.42248428e-01 1.72741246e+00 -1.00418818e+00 -1.18948668e-01
3.07749480e-01 8.03278804e-01 4.27974164e-01 3.60049874e-01
-9.59530413e-01 -2.14049637e-01 5.84311426e-01 -2.54299659e-02
2.88102716e-01 6.19131513e-03 -3.81678462e-01 -7.25300312e-01
-2.94435889e-01 -7.39349067e-01 6.30900681e-01 -3.47094089e-01
-1.17282534e+00 9.54567611e-01 8.74066651e-02 -1.45362508e+00
-5.99262305e-02 -6.13502264e-01 -9.42812562e-01 9.07438338e-01
-1.01498997e+00 -1.37834263e+00 1.90488413e-01 7.70329595e-01
3.61680508e-01 -6.89609051e-01 9.70021009e-01 -4.11911517e-01
-7.95632958e-01 -6.26193061e-02 7.57915005e-02 -1.00138970e-01
5.13102591e-01 -1.14725327e+00 -2.10906088e-01 1.00275421e+00
2.60108829e-01 5.01167886e-02 5.40439427e-01 -9.39291298e-01
-7.12979913e-01 -1.21741128e+00 2.71550179e-01 -3.80553424e-01
7.77739644e-01 -5.63681841e-01 -6.31903052e-01 5.59050798e-01
-1.88284796e-02 4.03992599e-03 7.57132471e-01 -6.64945126e-01
-4.07890409e-01 -1.63661644e-01 -1.50963748e+00 1.98844165e-01
3.28279674e-01 -4.32948649e-01 -6.20056808e-01 2.08716705e-01
-9.14050415e-02 2.65858248e-02 -5.78187108e-01 1.93777785e-01
2.20941439e-01 -8.14431369e-01 6.07165337e-01 -4.31696028e-01
-4.73048955e-01 -8.17096174e-01 -1.20816082e-01 -1.43090034e+00
-2.16331661e-01 -3.60378444e-01 4.12317991e-01 6.73037231e-01
1.94478139e-01 -7.23837793e-01 6.58061564e-01 2.22049709e-02
-1.27993494e-01 -7.76019216e-01 -1.87405002e+00 -9.76589382e-01
3.23934138e-01 -7.12480024e-02 -2.99154103e-01 4.96175647e-01
5.35107497e-03 3.78042385e-02 -4.42346871e-01 7.61008203e-01
1.33550608e+00 4.43397671e-01 2.46076390e-01 -1.49629128e+00
-4.61618781e-01 -2.47328237e-01 -3.12079340e-01 -1.94824263e-01
4.34184462e-01 -1.09096551e+00 -2.16282941e-02 -1.29486346e+00
-3.42260391e-01 -4.21634614e-01 -4.78065640e-01 1.04441965e+00
7.21881032e-01 -1.81617588e-02 2.21683651e-01 4.21156019e-01
-7.31236100e-01 7.32366517e-02 6.84687436e-01 -5.85638136e-02
-2.72005498e-01 2.87502021e-01 -9.49858427e-02 6.09982014e-01
1.01802421e+00 -1.13788176e+00 -8.67351443e-02 1.15229038e-03
-3.10175359e-01 3.05218309e-01 4.99988049e-01 -1.45490730e+00
6.30011499e-01 -1.23771735e-01 2.74905503e-01 -5.54813206e-01
5.14949262e-01 -1.07906258e+00 3.77493441e-01 9.25373733e-01
-1.64772123e-02 1.02968693e-01 3.90889823e-01 7.15413570e-01
-6.73393160e-02 -3.95847857e-01 1.19288099e+00 -2.11338848e-01
-1.96452603e-01 7.08854795e-02 -1.10113597e+00 -2.32228205e-01
2.02010942e+00 -1.37816176e-01 -3.54460359e-01 1.46285325e-01
-1.08326876e+00 5.71799338e-01 3.61069381e-01 5.91224693e-02
5.91459453e-01 -9.17553306e-01 -4.30538386e-01 4.65705156e-01
-1.58361971e-01 -1.96033433e-01 2.52820551e-01 8.94128144e-01
-2.45945901e-01 3.00166517e-01 -8.90145123e-01 -3.86856198e-01
-1.14950240e+00 7.86041200e-01 5.68866730e-01 -6.82452202e-01
-3.93330276e-01 7.08404362e-01 -9.07314010e-03 -3.07402760e-01
2.00381175e-01 2.33252794e-01 -2.82916009e-01 4.06202793e-01
4.08793896e-01 3.23639721e-01 -1.57998681e-01 -8.36846888e-01
-4.44011033e-01 6.66885972e-01 1.28470317e-01 -5.05617559e-01
1.52966249e+00 2.58319974e-01 -2.81387210e-01 2.92251319e-01
6.66643381e-01 -5.05005717e-01 -1.47242200e+00 3.89713287e-01
-1.34614967e-02 1.94166541e-01 -1.54493243e-01 -9.53352034e-01
-9.88605142e-01 9.09716487e-01 5.69201767e-01 2.04856575e-01
1.27606845e+00 -2.79919446e-01 3.71022582e-01 4.02278483e-01
1.32201469e+00 -1.21775079e+00 3.14125195e-02 4.40694958e-01
7.56234348e-01 -5.77790260e-01 -5.56954801e-01 2.66026765e-01
-7.29294717e-01 1.24090576e+00 5.76868117e-01 -8.94523978e-01
6.00884795e-01 7.02063799e-01 2.50645012e-01 -2.12518215e-01
-9.41589415e-01 4.18625548e-02 -3.43792617e-01 6.83592677e-01
-9.97949839e-01 -7.39143863e-02 -1.71732545e-01 5.03408313e-01
3.20354581e-01 -6.14869416e-01 4.09871966e-01 9.98939633e-01
-8.65038991e-01 -1.48301566e+00 -8.19561422e-01 2.75405467e-01
-6.39027134e-02 3.99359018e-01 -6.61408186e-01 4.60944653e-01
5.16361117e-01 9.21349049e-01 3.89273348e-03 -5.65636337e-01
9.76629481e-02 2.60921456e-02 1.53797179e-01 -7.57562160e-01
-1.30692613e+00 3.48301381e-01 -4.22426373e-01 -5.22829711e-01
-1.32084131e-01 -7.53789008e-01 -1.52695823e+00 2.07186475e-01
-2.02099949e-01 6.28527701e-01 7.31139362e-01 8.97944272e-01
6.54394925e-02 1.51472315e-01 9.20346320e-01 -9.60124254e-01
-6.42163336e-01 -6.30070448e-01 -9.62617278e-01 -2.60907989e-02
-4.20511439e-02 -1.13000917e+00 -1.03221560e+00 -4.17373240e-01] | [3.811410903930664, 2.307469367980957] |
83555a4a-bb33-4271-a9e5-77dab1e4dd7b | constructing-category-specific-models-for | 1802.09292 | null | http://arxiv.org/abs/1802.09292v1 | http://arxiv.org/pdf/1802.09292v1.pdf | Constructing Category-Specific Models for Monocular Object-SLAM | We present a new paradigm for real-time object-oriented SLAM with a monocular
camera. Contrary to previous approaches, that rely on object-level models, we
construct category-level models from CAD collections which are now widely
available. To alleviate the need for huge amounts of labeled data, we develop a
rendering pipeline that enables synthesis of large datasets from a limited
amount of manually labeled data. Using data thus synthesized, we learn
category-level models for object deformations in 3D, as well as discriminative
object features in 2D. These category models are instance-independent and aid
in the design of object landmark observations that can be incorporated into a
generic monocular SLAM framework. Where typical object-SLAM approaches usually
solve only for object and camera poses, we also estimate object shape
on-the-fly, allowing for a wide range of objects from the category to be
present in the scene. Moreover, since our 2D object features are learned
discriminatively, the proposed object-SLAM system succeeds in several scenarios
where sparse feature-based monocular SLAM fails due to insufficient features or
parallax. Also, the proposed category-models help in object instance retrieval,
useful for Augmented Reality (AR) applications. We evaluate the proposed
framework on multiple challenging real-world scenes and show --- to the best of
our knowledge --- first results of an instance-independent monocular
object-SLAM system and the benefits it enjoys over feature-based SLAM methods. | ['J. Krishna Murthy', 'Rishabh Khawad', 'Parv Parkhiya', 'Brojeshwar Bhowmick', 'K. Madhava Krishna'] | 2018-02-26 | null | null | null | null | ['object-slam'] | ['computer-vision'] | [-8.94985721e-02 -2.98926562e-01 1.56842638e-02 -5.31927228e-01
-8.22026849e-01 -6.09126806e-01 6.21719539e-01 1.15248457e-01
-2.02208161e-01 5.00537574e-01 -3.31733108e-01 1.20690405e-01
-1.86749637e-01 -6.44534767e-01 -7.91288435e-01 -4.12916839e-01
6.45038784e-02 1.17713392e+00 4.22732174e-01 -9.89811495e-02
3.75781536e-01 1.16993487e+00 -2.05581236e+00 -1.14763945e-01
6.21845245e-01 9.91163135e-01 5.98165929e-01 3.98557007e-01
-1.67472914e-01 4.43638265e-01 -2.09066778e-01 1.06969602e-01
5.95325947e-01 2.16650926e-02 -3.81010801e-01 4.81892556e-01
1.09237778e+00 -2.57823497e-01 -3.05722684e-01 8.48216057e-01
2.52980739e-01 2.14198589e-01 4.20362920e-01 -1.32947075e+00
-1.65622458e-01 -4.18124706e-01 -2.91868001e-01 -5.30928254e-01
6.85137391e-01 -3.34679673e-04 9.76203680e-01 -1.34746718e+00
1.06927812e+00 1.11855781e+00 5.04503608e-01 1.53295234e-01
-1.18052769e+00 -3.01569521e-01 8.51240382e-02 1.06423929e-01
-1.61821783e+00 -4.21949834e-01 9.69451785e-01 -5.43488264e-01
7.15145290e-01 3.57067287e-01 9.25302565e-01 4.15058225e-01
1.73835941e-02 5.60381889e-01 9.80784833e-01 -3.72720599e-01
3.56001198e-01 3.00479144e-01 -4.33337595e-03 7.40666747e-01
3.50467086e-01 4.38267253e-02 -7.87934959e-01 -2.58725613e-01
8.62246692e-01 4.44083005e-01 -2.39292040e-01 -1.43884218e+00
-1.30496967e+00 6.38301611e-01 8.07353556e-01 1.17944151e-01
-3.60289156e-01 2.46024027e-01 -8.20369646e-02 8.50762278e-02
3.53943378e-01 4.12706763e-01 -5.04838169e-01 1.78393826e-01
-9.43910599e-01 3.21064234e-01 6.02279544e-01 1.53772509e+00
1.47974765e+00 -3.63584831e-02 4.05368924e-01 5.20283520e-01
3.74571025e-01 7.81939149e-01 1.13708913e-01 -8.17935169e-01
1.58085153e-01 9.56942022e-01 3.95256758e-01 -8.87273014e-01
-5.43314695e-01 -4.27567333e-01 -3.55480671e-01 3.15448791e-01
1.41847311e-02 7.12322831e-01 -8.85685921e-01 1.29000998e+00
6.86372817e-01 4.27591056e-02 -1.62298769e-01 1.11579669e+00
6.40812814e-01 2.20078021e-01 -5.03003001e-01 -2.03154646e-02
1.19121730e+00 -7.31528282e-01 -3.86102885e-01 -4.29092705e-01
6.97286010e-01 -9.87775981e-01 8.82965684e-01 4.63454604e-01
-7.00964153e-01 -5.60913086e-01 -8.51011038e-01 -2.58139759e-01
-3.27347010e-01 2.73798674e-01 7.31468678e-01 3.72696429e-01
-8.54116678e-01 2.10183814e-01 -6.79731131e-01 -5.72683394e-01
1.93671554e-01 5.23109436e-01 -7.59443402e-01 -4.70870256e-01
-3.54693145e-01 1.08590233e+00 3.98096710e-01 1.46132603e-01
-8.72698843e-01 -6.20904148e-01 -1.04061592e+00 -3.10233861e-01
5.68246484e-01 -7.45094121e-01 9.41992342e-01 -5.19884050e-01
-1.16674089e+00 9.98699844e-01 -3.48295689e-01 -1.54247135e-01
4.74319667e-01 -5.18879294e-01 1.03372708e-01 8.64290521e-02
1.22836538e-01 7.53125906e-01 6.85806096e-01 -1.76759851e+00
-5.66249192e-01 -6.59713209e-01 1.64102182e-01 4.57059354e-01
2.86275655e-01 -3.95502061e-01 -5.92165291e-01 -5.80515340e-02
1.03909647e+00 -1.19855964e+00 -2.53073692e-01 5.21945834e-01
-2.24345982e-01 1.65385738e-01 1.05962551e+00 -3.02387029e-01
3.15708548e-01 -2.08289051e+00 1.99875623e-01 1.70657828e-01
-1.07275730e-04 -1.37051061e-01 -6.73825815e-02 4.68409061e-01
2.61789799e-01 -4.22097832e-01 -7.52332732e-02 -6.18266046e-01
-1.97881144e-02 4.94494766e-01 -2.28995487e-01 8.79565895e-01
2.58342866e-02 7.96462536e-01 -8.07271481e-01 -4.43935275e-01
8.68907392e-01 5.30613959e-01 -6.22829437e-01 2.86141634e-01
-4.18992758e-01 7.52464831e-01 -3.43685061e-01 9.49307263e-01
9.25977707e-01 1.22733966e-01 -1.38355093e-02 -3.00940514e-01
-2.34186515e-01 2.70405591e-01 -1.52591550e+00 2.28623343e+00
-8.18402827e-01 4.61843520e-01 -7.80845620e-03 -6.16078436e-01
1.12130845e+00 -3.56155820e-03 5.56322515e-01 -3.20459872e-01
-1.36486785e-02 6.17486715e-01 -3.29208463e-01 -3.90069112e-02
9.15286660e-01 -3.73249315e-02 -5.00149764e-02 2.00336031e-03
2.70591378e-01 -9.47929323e-01 -1.46808907e-01 1.52019784e-01
7.15130508e-01 7.10445046e-01 6.19474530e-01 -1.96018577e-01
4.55996722e-01 3.30757588e-01 4.41323191e-01 6.04185104e-01
3.68450791e-01 7.20278203e-01 -2.61500955e-01 -5.85950136e-01
-9.31348205e-01 -9.63477314e-01 -3.35752636e-01 5.47076046e-01
6.69233024e-01 -3.89642060e-01 -6.72597662e-02 -3.90878230e-01
2.77875185e-01 5.57801783e-01 -2.22122774e-01 2.94122607e-01
-4.54418063e-01 -1.15037136e-01 -3.39242905e-01 1.37471408e-01
1.46808803e-01 -8.00866067e-01 -1.02553427e+00 2.19445229e-01
1.44345954e-01 -1.28320611e+00 -5.00590540e-02 2.25261837e-01
-9.81794238e-01 -1.12426364e+00 -2.76900172e-01 -5.13785064e-01
8.21315885e-01 7.24223733e-01 1.04463494e+00 1.68543860e-01
-6.22865081e-01 7.32357502e-01 -2.46199757e-01 -2.45824352e-01
-1.19024262e-01 -1.90693364e-01 2.81742841e-01 1.04882374e-01
2.55758256e-01 -5.74008286e-01 -3.30824852e-01 5.57112277e-01
-6.11830473e-01 1.58310547e-01 4.25299138e-01 6.25266254e-01
9.71611202e-01 -2.89064407e-01 1.49769172e-01 -6.51674092e-01
-5.14881253e-01 -7.95525536e-02 -1.20528710e+00 6.93545714e-02
-3.51645321e-01 -1.68482289e-01 2.11372420e-01 -4.22088295e-01
-7.26719737e-01 7.49758363e-01 2.58740485e-01 -7.15101779e-01
-3.51696968e-01 3.50296438e-01 -3.94394487e-01 -5.40986240e-01
5.63974142e-01 2.74361312e-01 -2.23662779e-01 -7.42395818e-01
4.94064748e-01 4.09499198e-01 4.77220803e-01 -5.74892640e-01
1.16594708e+00 7.88633645e-01 4.18274611e-01 -9.07419801e-01
-7.77012885e-01 -1.02710128e+00 -1.09824741e+00 -1.66501045e-01
5.04693210e-01 -1.21589434e+00 -4.38294381e-01 1.57237470e-01
-1.30337965e+00 5.33565134e-02 -6.16908610e-01 7.49615192e-01
-9.12007391e-01 2.62876183e-01 3.40978354e-02 -1.01291120e+00
1.98807701e-01 -1.33268833e+00 1.63378811e+00 -1.24139458e-01
9.15423483e-02 -6.84462309e-01 -1.38856610e-02 1.41948402e-01
-7.61905536e-02 2.20653102e-01 3.70008558e-01 -4.44000959e-01
-1.42530274e+00 -5.63861787e-01 -8.84232968e-02 -1.11117221e-01
1.57359734e-01 -2.29377642e-01 -1.08580041e+00 -4.01221663e-01
-1.19726285e-02 -2.32479960e-01 3.76804262e-01 1.82717049e-04
6.48440480e-01 1.74547657e-01 -4.35989887e-01 7.89980650e-01
1.70376158e+00 3.82625498e-03 2.96485960e-01 2.86198974e-01
9.72417951e-01 6.60020888e-01 1.20154202e+00 5.70909083e-01
3.58558357e-01 1.26825380e+00 8.87783051e-01 7.68015161e-02
-4.55074740e-04 -4.03187960e-01 -4.30243202e-02 6.74336493e-01
1.91873491e-01 1.43207163e-01 -8.75318825e-01 6.10144436e-01
-1.85313690e+00 -3.77122194e-01 -3.58956963e-01 2.41766238e+00
2.86519349e-01 -2.46868715e-01 -1.78140953e-01 -1.46702111e-01
2.70094424e-01 3.02564353e-02 -5.11064768e-01 1.57608271e-01
-1.61739171e-01 -3.01271752e-02 5.16663492e-01 6.67622626e-01
-8.97507608e-01 1.10582399e+00 4.71010065e+00 4.07922417e-01
-1.04533315e+00 1.47475317e-01 -5.06636262e-01 -2.43598446e-02
-3.57335895e-01 7.17232704e-01 -1.05967534e+00 -1.18654922e-01
3.76704276e-01 3.98233421e-02 2.30152130e-01 1.25598884e+00
-2.05204874e-01 -4.12338912e-01 -1.44228923e+00 1.38717735e+00
2.94656128e-01 -1.29217339e+00 3.90560515e-02 4.70826268e-01
7.09417701e-01 1.53201789e-01 -3.55905652e-01 -4.93965186e-02
9.17112827e-02 -4.35224831e-01 1.04268384e+00 4.59444523e-01
6.94104135e-01 -4.34815973e-01 5.47194123e-01 7.43871212e-01
-1.30898094e+00 6.39511645e-02 -6.44282281e-01 -2.76868008e-02
3.14307362e-01 7.27816701e-01 -1.09859741e+00 9.48816061e-01
4.56543267e-01 7.40507483e-01 -6.25443757e-01 1.31176054e+00
-1.65038019e-01 -1.29901469e-01 -6.87977970e-01 3.33084911e-01
2.09653266e-02 -2.56488979e-01 6.09881878e-01 6.98882699e-01
5.01653135e-01 -1.64955657e-03 5.72080493e-01 6.50529027e-01
1.20903872e-01 2.05924690e-01 -9.00484622e-01 4.95334655e-01
4.18396801e-01 1.33992839e+00 -6.68689668e-01 -2.33088031e-01
-2.52743065e-01 8.80229414e-01 2.42451474e-01 4.13053408e-02
-3.90927106e-01 2.24797457e-01 4.65957791e-01 2.59769022e-01
2.10702941e-01 -7.44521618e-01 -8.67547691e-02 -1.44811177e+00
1.75278038e-01 -5.62079608e-01 -9.88324285e-02 -1.15192282e+00
-8.63460541e-01 5.96859872e-01 1.60655364e-01 -1.68484521e+00
-2.44358033e-01 -5.31369567e-01 1.30642399e-01 7.86893249e-01
-1.61669469e+00 -1.61750495e+00 -7.60576546e-01 5.05183578e-01
6.47474766e-01 -6.39901161e-02 8.30881596e-01 8.20214748e-02
3.02821696e-01 -1.51271194e-01 -4.62167971e-02 -4.66551632e-01
7.29548395e-01 -1.02646637e+00 1.97334394e-01 7.20923305e-01
8.65027547e-01 6.80248320e-01 5.58120549e-01 -7.94943631e-01
-1.90918899e+00 -9.25451517e-01 7.01557875e-01 -7.49231696e-01
3.46428007e-01 -7.35643983e-01 -7.86339164e-01 8.18065107e-01
-5.76987863e-01 5.50050974e-01 9.22724903e-02 3.74993272e-02
-3.15141886e-01 -1.16453536e-01 -1.15392792e+00 1.73717424e-01
1.20197380e+00 -6.22410178e-01 -4.25353199e-01 7.15694785e-01
4.75348860e-01 -8.02764058e-01 -6.44367278e-01 4.72178280e-01
6.41643703e-01 -9.84253943e-01 1.01879179e+00 -2.21012190e-01
-3.59201133e-01 -7.31341243e-01 -7.33133733e-01 -1.08968365e+00
8.26275907e-04 -3.68974954e-01 1.29134860e-02 8.90050411e-01
-3.08857441e-01 -5.80170393e-01 7.98452735e-01 3.76082361e-01
-3.26403528e-01 -2.78625607e-01 -1.19348383e+00 -9.90313411e-01
-6.93152964e-01 -4.82076466e-01 6.45300984e-01 7.01750398e-01
-7.66058385e-01 9.67881083e-02 -4.16963041e-01 6.91970527e-01
7.77180851e-01 6.90930426e-01 1.47957075e+00 -1.67178082e+00
-1.67054802e-01 2.12757811e-01 -9.45220947e-01 -1.22348750e+00
2.90151596e-01 -7.91271031e-01 2.64129937e-01 -1.43432319e+00
-4.79620583e-02 -9.65941906e-01 1.76464334e-01 4.17930454e-01
2.73447305e-01 4.91166770e-01 4.76105541e-01 5.65231144e-01
-5.03658175e-01 5.70010900e-01 9.76038635e-01 1.00887239e-01
-1.59754902e-01 -9.53132287e-02 1.14837056e-02 8.43869328e-01
2.74589360e-01 -6.20375395e-01 -2.50231296e-01 -4.32877332e-01
9.65423658e-02 1.89508855e-01 7.32617378e-01 -1.04433024e+00
2.18732521e-01 -3.43179822e-01 1.11250818e-01 -1.25527728e+00
1.10605049e+00 -1.44599521e+00 6.34877861e-01 2.65920013e-01
2.72584051e-01 -2.29298726e-01 -7.08211586e-02 5.07567108e-01
-1.72006935e-01 -3.69790077e-01 6.32243574e-01 -2.98496872e-01
-9.64781702e-01 4.91642177e-01 2.05774873e-01 -5.29619813e-01
9.30664480e-01 -4.59443539e-01 -1.54780019e-02 -3.54213148e-01
-6.52082860e-01 -6.67283386e-02 1.17932260e+00 3.86666358e-01
7.21902609e-01 -1.37747920e+00 -4.47763622e-01 5.34259796e-01
7.00909972e-01 6.39729798e-01 2.13553950e-01 6.66198134e-01
-7.04428971e-01 5.08401513e-01 -1.75748125e-01 -1.28047276e+00
-1.18588352e+00 6.88014150e-01 1.06327586e-01 2.17870519e-01
-7.09498525e-01 4.61073816e-01 6.13104463e-01 -8.67342114e-01
-5.20780496e-02 -3.00338835e-01 3.91336620e-01 -1.02055490e-01
1.34643599e-01 1.23318627e-01 3.67157578e-01 -1.11420703e+00
-6.15987718e-01 1.13871002e+00 3.63044798e-01 -1.46010891e-01
1.40342760e+00 -1.94787011e-01 -1.66028112e-01 7.76219726e-01
1.01816225e+00 2.22182751e-01 -1.21446586e+00 -4.82143819e-01
6.79729953e-02 -1.02717876e+00 7.23012239e-02 -4.05720353e-01
-5.94629943e-01 9.29020166e-01 5.68778396e-01 -3.17751914e-01
7.83298433e-01 1.51162148e-01 2.45143011e-01 6.38059795e-01
1.36295724e+00 -5.20078123e-01 -4.35052961e-02 3.27878743e-01
1.09912372e+00 -1.17475414e+00 4.41887140e-01 -6.13559783e-01
-2.94372410e-01 1.05608368e+00 4.73389179e-01 -2.52612501e-01
4.57996994e-01 9.12113041e-02 -1.53132686e-02 -2.53640622e-01
-2.69213349e-01 -3.05330366e-01 4.37417448e-01 7.16153502e-01
-1.26464590e-01 -1.40321453e-03 1.32559210e-01 -2.50953585e-01
-1.80924922e-01 -1.01711832e-01 4.50562239e-01 1.20604527e+00
-5.55142462e-01 -1.24302208e+00 -5.75170875e-01 9.84425992e-02
4.11209106e-01 1.53937757e-01 -1.82437569e-01 9.65954423e-01
1.41399100e-01 4.67103779e-01 7.78366402e-02 -7.18582124e-02
4.74597543e-01 3.88183072e-02 8.31101060e-01 -9.61273074e-01
-2.82873988e-01 3.61927807e-01 -4.89204340e-02 -7.71576345e-01
-6.51776373e-01 -8.03142369e-01 -1.10604107e+00 2.16896921e-01
-7.64123201e-01 -1.13780625e-01 1.29229367e+00 7.80988455e-01
3.67268652e-01 -5.42299561e-02 5.29172301e-01 -1.53238702e+00
-3.72269720e-01 -7.06951499e-01 -8.07964921e-01 3.02464932e-01
4.28767323e-01 -1.15529037e+00 -2.45977342e-01 -4.22966480e-02] | [7.3274102210998535, -2.3462979793548584] |
2b980dde-fdf9-4c30-8db9-2ddb2bd820ee | shaddr-real-time-example-based-geometry-and | 2306.04889 | null | https://arxiv.org/abs/2306.04889v1 | https://arxiv.org/pdf/2306.04889v1.pdf | ShaDDR: Real-Time Example-Based Geometry and Texture Generation via 3D Shape Detailization and Differentiable Rendering | We present ShaDDR, an example-based deep generative neural network which produces a high-resolution textured 3D shape through geometry detailization and conditional texture generation applied to an input coarse voxel shape. Trained on a small set of detailed and textured exemplar shapes, our method learns to detailize the geometry via multi-resolution voxel upsampling and generate textures on voxel surfaces via differentiable rendering against exemplar texture images from a few views. The generation is real-time, taking less than 1 second to produce a 3D model with voxel resolutions up to 512^3. The generated shape preserves the overall structure of the input coarse voxel model, while the style of the generated geometric details and textures can be manipulated through learned latent codes. In the experiments, we show that our method can generate higher-resolution shapes with plausible and improved geometric details and clean textures compared to prior works. Furthermore, we showcase the ability of our method to learn geometric details and textures from shapes reconstructed from real-world photos. In addition, we have developed an interactive modeling application to demonstrate the generalizability of our method to various user inputs and the controllability it offers, allowing users to interactively sculpt a coarse voxel shape to define the overall structure of the detailized 3D shape. | ['Hao Zhang', 'Hang Zhou', 'Zhiqin Chen', 'Qimin Chen'] | 2023-06-08 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 3.96616608e-01 5.49597979e-01 5.59352458e-01 -2.17824563e-01
-7.61614263e-01 -7.39106953e-01 8.27036321e-01 -4.24786121e-01
5.60645878e-01 6.56621635e-01 1.73777834e-01 -2.38085026e-03
1.88897122e-02 -1.41650510e+00 -1.13018751e+00 -5.96323669e-01
-1.48471281e-01 8.38305652e-01 -1.48716941e-01 -2.16889575e-01
1.90126270e-01 1.16369736e+00 -1.75235057e+00 5.31014979e-01
5.73661387e-01 9.50078130e-01 5.35608716e-02 6.74297869e-01
-1.52304592e-02 1.49172112e-01 -3.90027195e-01 -1.17359616e-01
5.51255286e-01 -1.90994069e-02 -5.11940718e-01 5.37613571e-01
9.09794390e-01 -4.99583095e-01 -7.84004629e-02 6.50593638e-01
2.46427864e-01 4.79172188e-04 1.03718746e+00 -6.99658394e-01
-1.09084547e+00 1.68905631e-01 -4.55643237e-01 -7.22712159e-01
2.71351188e-01 3.58191490e-01 6.31027818e-01 -1.16614676e+00
1.13947439e+00 1.60971510e+00 5.55201471e-01 5.66738248e-01
-1.99117732e+00 -5.12669921e-01 -1.28565609e-01 -8.40818524e-01
-1.41007400e+00 -2.33476341e-01 9.51632142e-01 -5.09072244e-01
8.42995226e-01 3.71784776e-01 7.95285165e-01 1.10422623e+00
5.13415992e-01 2.66250551e-01 1.35845935e+00 -1.65429443e-01
3.87803614e-01 -1.10678218e-01 -6.08177006e-01 9.84472871e-01
2.10194066e-02 4.39465225e-01 -1.83696404e-01 -3.02463323e-01
1.96138990e+00 4.81541343e-02 -1.34783372e-01 -6.61102295e-01
-1.31786978e+00 6.92860782e-01 6.09906137e-01 -5.27289882e-02
-3.71286541e-01 4.43778247e-01 -1.48210570e-01 3.32963169e-02
7.77083457e-01 5.64305067e-01 -1.02490015e-01 8.46670642e-02
-6.19682550e-01 5.28892636e-01 7.13418305e-01 1.18919694e+00
8.50982428e-01 5.29255092e-01 -6.15418032e-02 6.25106335e-01
-4.24307361e-02 9.12141800e-01 -1.76047981e-01 -1.29770863e+00
1.35495104e-02 4.79581058e-01 2.06973106e-01 -8.31859648e-01
8.20191577e-02 -1.97664648e-01 -1.08619356e+00 9.81334925e-01
1.30630016e-01 1.88928396e-01 -1.35699570e+00 1.47908390e+00
3.57476503e-01 -1.14562310e-01 -2.82351434e-01 7.92042673e-01
8.20252061e-01 6.46301985e-01 -2.77520537e-01 2.75987834e-01
9.65273380e-01 -3.43870401e-01 -2.46793643e-01 3.96629810e-01
-1.10093966e-01 -8.07022870e-01 1.43903375e+00 3.44088316e-01
-1.41629624e+00 -5.97499907e-01 -8.87148321e-01 -1.40107960e-01
-1.64333984e-01 -5.13962545e-02 7.02831388e-01 4.21853840e-01
-1.13562083e+00 8.69839191e-01 -8.37218046e-01 3.15674730e-02
7.16494799e-01 2.50629157e-01 -4.58943069e-01 1.23500079e-02
-5.11937380e-01 3.97591442e-01 -8.98487419e-02 -2.80617267e-01
-1.13071895e+00 -1.04232216e+00 -8.57192457e-01 1.69385299e-02
-2.46401448e-02 -1.10086477e+00 9.86338854e-01 -6.79776251e-01
-1.90594625e+00 9.59431052e-01 -9.75382105e-02 1.51921973e-01
5.42757571e-01 3.46344970e-02 4.76061665e-02 4.71436121e-02
9.13602710e-02 8.28886747e-01 1.03971088e+00 -1.93156612e+00
1.80154797e-02 -1.52374178e-01 6.23190477e-02 5.01192287e-02
5.53119123e-01 -7.36190379e-01 -1.13833830e-01 -1.03320312e+00
3.70254666e-01 -8.00560474e-01 -2.69515365e-01 3.11798930e-01
-5.94812870e-01 2.60155052e-01 9.52103198e-01 -2.55216181e-01
2.19064698e-01 -2.19885921e+00 2.51601160e-01 5.35967648e-01
2.26438612e-01 -4.49355602e-01 -2.41322875e-01 5.56615889e-01
-1.02219075e-01 4.27930981e-01 -3.65359247e-01 -4.17381912e-01
1.59109801e-01 1.82226986e-01 -6.98021293e-01 1.83751002e-01
3.84631127e-01 1.09442616e+00 -6.51710570e-01 4.86071371e-02
3.91469300e-01 1.04826736e+00 -7.30055869e-01 3.53843719e-01
-6.00990355e-01 9.21152055e-01 -4.80910838e-01 6.79290533e-01
6.76724076e-01 -2.59361118e-01 -1.17824316e-01 -2.57999122e-01
-9.05741230e-02 -9.13267676e-03 -1.14519382e+00 1.60215855e+00
-7.50096202e-01 1.57381862e-01 1.90715209e-01 -3.24050874e-01
1.22728693e+00 1.76226944e-01 2.10389048e-01 -5.68323314e-01
-2.85430811e-02 3.23221505e-01 -5.64733148e-01 -1.64685901e-02
4.53540325e-01 -5.46982706e-01 -1.91456631e-01 8.06142151e-01
-2.80768454e-01 -1.22203994e+00 -4.91983205e-01 3.94364521e-02
5.96468925e-01 6.37598276e-01 -5.58825098e-02 -4.01990116e-01
-7.21854577e-03 -7.53333420e-02 -1.58181012e-01 5.57205498e-01
8.47887754e-01 1.06549525e+00 2.52236158e-01 -8.02721918e-01
-1.73802602e+00 -1.79699349e+00 -4.09699112e-01 4.85282898e-01
-4.94908132e-02 3.52423638e-02 -7.67348707e-01 -1.72441870e-01
2.18557358e-01 7.20242858e-01 -9.55177605e-01 1.63814709e-01
-8.32367301e-01 -2.16008469e-01 1.52819172e-01 4.36637312e-01
3.66161853e-01 -1.22610176e+00 -4.99793559e-01 1.21450737e-01
3.13142478e-01 -7.74760187e-01 -4.49756444e-01 -1.74407333e-01
-1.10698068e+00 -8.00605059e-01 -6.17433310e-01 -8.11683357e-01
1.08409452e+00 -1.11387886e-01 1.36419749e+00 1.27198445e-02
-5.58739185e-01 4.74801153e-01 1.68622971e-01 -8.19120109e-02
-7.00337350e-01 -2.75807709e-01 -7.96575397e-02 -2.68787276e-02
-6.94090009e-01 -1.22939765e+00 -5.19486248e-01 1.17571242e-01
-1.14371610e+00 8.20425212e-01 3.13965976e-01 8.72376800e-01
1.33287513e+00 5.17846160e-02 2.77258247e-01 -9.73574042e-01
5.95351398e-01 -7.53413215e-02 -6.72107756e-01 -1.72759324e-01
-1.40087441e-01 3.66027057e-01 7.36898601e-01 -5.07413507e-01
-1.16864169e+00 5.80922849e-02 -4.22682315e-02 -6.15470111e-01
-4.98684108e-01 1.16404314e-02 -9.19357389e-02 -1.27360448e-01
6.38249159e-01 3.16429287e-01 1.79027636e-02 -5.55529654e-01
8.41609895e-01 3.79481316e-02 6.19946063e-01 -1.19511437e+00
1.11131716e+00 9.11537230e-01 3.19119513e-01 -8.00323606e-01
-3.73917043e-01 8.08964193e-01 -7.71416366e-01 -4.72258590e-02
7.86353409e-01 -6.76239014e-01 -7.87487566e-01 3.16457838e-01
-9.41605866e-01 -8.40311825e-01 -7.60802984e-01 -8.18099976e-02
-1.10273218e+00 -1.34140670e-01 -6.62923217e-01 -6.75252974e-01
-3.23562354e-01 -1.16866732e+00 1.76011574e+00 -1.89524576e-01
-3.35639387e-01 -7.38937199e-01 -1.62745222e-01 -9.78568569e-02
7.21911192e-01 1.02837372e+00 1.39206350e+00 3.90389830e-01
-1.28095508e+00 5.89422807e-02 -1.26960233e-01 -1.39421299e-01
4.75728661e-01 1.62501976e-01 -9.13132429e-01 -2.51230150e-01
-3.00235808e-01 -4.00438547e-01 4.79923069e-01 4.10515517e-01
1.43719327e+00 -4.28004801e-01 -1.74055189e-01 8.49447966e-01
1.47876871e+00 6.38966449e-03 6.71995640e-01 -1.87814072e-01
8.64092827e-01 3.72010112e-01 -1.82870273e-02 3.32882941e-01
-1.58785284e-02 5.76146364e-01 4.00897235e-01 -2.96977788e-01
-4.20752406e-01 -6.02810204e-01 -1.10541299e-01 4.68667537e-01
-5.49575746e-01 1.67794317e-01 -5.87555170e-01 2.25339696e-01
-1.21690154e+00 -8.03079188e-01 1.56235501e-01 2.30651450e+00
7.45855629e-01 7.51992166e-02 -1.90515682e-01 -2.94340551e-01
4.41530168e-01 3.35546061e-02 -6.37173414e-01 -5.70669711e-01
-3.42988856e-02 8.59866381e-01 1.05407342e-01 7.74859846e-01
-7.35515535e-01 1.07893026e+00 6.86754036e+00 7.76815832e-01
-1.46069825e+00 -2.64820576e-01 9.61564481e-01 -3.26058656e-01
-1.10845315e+00 -3.03043067e-01 -2.24216372e-01 -9.74971280e-02
3.79406273e-01 -1.25795081e-01 8.19813013e-01 8.21807981e-01
1.78184822e-01 2.63611555e-01 -1.24241602e+00 9.55408335e-01
-2.50541717e-01 -1.81410861e+00 7.20166445e-01 2.28458330e-01
1.09474432e+00 -3.11776280e-01 3.85325253e-01 -6.39742464e-02
7.04766989e-01 -1.59750688e+00 9.77880120e-01 8.09948087e-01
1.65503669e+00 -8.17527056e-01 -1.05704777e-01 2.21289083e-01
-1.19367695e+00 4.57792789e-01 -3.06169242e-01 2.37479173e-02
2.85782188e-01 5.69663107e-01 -7.36853600e-01 3.56031954e-01
6.21847153e-01 2.79059857e-01 -1.18950397e-01 3.39667350e-01
-9.31375623e-02 1.54403031e-01 -3.86589319e-01 1.57165170e-01
1.11922354e-01 -3.43600512e-01 5.96857548e-01 7.90474951e-01
6.11939251e-01 3.36404711e-01 1.14124343e-01 1.76126826e+00
-1.10857598e-01 -6.36675209e-02 -1.05111730e+00 6.36907071e-02
3.72037381e-01 1.03293240e+00 -6.51439190e-01 -4.17427808e-01
1.54143542e-01 9.64679301e-01 3.86360943e-01 5.99004626e-01
-5.93768299e-01 -2.67994683e-02 6.63517654e-01 5.73450923e-01
4.88303453e-01 -6.22930169e-01 -7.76572406e-01 -9.13749278e-01
-1.03453785e-01 -6.92460716e-01 -4.92816865e-01 -1.28925180e+00
-1.35240626e+00 9.03823435e-01 4.53205295e-02 -1.12980568e+00
-1.79317370e-01 -4.62750584e-01 -5.49999416e-01 1.25923836e+00
-8.86687577e-01 -1.37359357e+00 -2.74487555e-01 3.89772505e-01
3.67240459e-01 -5.34371100e-02 1.21957648e+00 -2.96688825e-01
1.71353877e-01 1.27903685e-01 -1.44795492e-01 -1.24500662e-01
1.90241963e-01 -1.27032995e+00 9.56536293e-01 2.16687888e-01
7.68661052e-02 6.04239106e-01 5.68534970e-01 -9.06584978e-01
-1.55380321e+00 -1.23015130e+00 1.79845318e-01 -5.54871559e-01
2.89982632e-02 -6.05850101e-01 -8.87819052e-01 7.64879107e-01
9.52271521e-02 2.01684772e-03 2.67896026e-01 -2.84808874e-01
-5.34215987e-01 2.86239415e-01 -1.50007010e+00 9.71548557e-01
1.43268180e+00 -4.91322428e-01 -4.40050125e-01 1.01018056e-01
6.60746932e-01 -8.84244859e-01 -1.13531113e+00 5.21964371e-01
6.97680056e-01 -1.00142658e+00 1.09340656e+00 -3.93628925e-01
8.00071120e-01 -3.40805084e-01 -2.54348367e-01 -1.50351191e+00
-6.73208177e-01 -5.51965117e-01 2.03001946e-01 6.99080110e-01
3.20680261e-01 -6.00086808e-01 7.80486107e-01 6.71886861e-01
-2.08787441e-01 -9.08335507e-01 -7.73902297e-01 -5.67232966e-01
3.91831517e-01 -3.39656085e-01 1.22098112e+00 7.97154546e-01
-5.28179705e-01 -8.27149972e-02 -9.05501246e-02 2.18243197e-01
8.80432725e-01 8.57763171e-01 9.88586605e-01 -1.13401639e+00
-3.39786142e-01 -4.48872358e-01 -4.99677472e-02 -1.03297210e+00
-8.78550783e-02 -9.54635739e-01 -1.56032324e-01 -1.41456854e+00
-2.55089737e-02 -6.47108734e-01 5.11960387e-01 3.36614907e-01
3.33427399e-01 6.14057004e-01 2.51099039e-02 9.39944088e-02
2.78429419e-01 7.94270813e-01 2.10757804e+00 -1.08772561e-01
-2.44741634e-01 -4.06698436e-01 -5.56823909e-01 6.57428443e-01
4.78183389e-01 -7.13480562e-02 -4.38884944e-01 -5.10570526e-01
5.84572144e-02 2.92735487e-01 6.61822021e-01 -7.56998479e-01
-5.69733799e-01 -3.13962191e-01 8.39181721e-01 -4.88911569e-01
6.56600296e-01 -6.42893851e-01 9.34261680e-01 1.24938950e-01
-3.45886856e-01 -9.99978334e-02 3.43187481e-01 3.38588804e-01
1.66995972e-01 5.59345424e-01 8.74509275e-01 -4.78618085e-01
-1.29034802e-01 6.96057260e-01 -1.30479217e-01 -1.45541295e-01
7.60163307e-01 -4.27680641e-01 -1.32393569e-01 -4.38891977e-01
-1.05993056e+00 -4.82737660e-01 1.09970760e+00 2.94634700e-01
9.53640044e-01 -1.86832297e+00 -6.93709075e-01 8.58553946e-01
-8.70943293e-02 5.09598017e-01 3.23935568e-01 -1.85741320e-01
-8.36453795e-01 1.02459975e-01 -3.78011733e-01 -7.88939118e-01
-7.27528155e-01 4.11022335e-01 5.00724554e-01 8.52961093e-02
-1.24038637e+00 4.92861420e-01 8.71332288e-01 -6.71769202e-01
-2.19565466e-01 -6.12002969e-01 4.66650188e-01 -8.67694557e-01
2.94229001e-01 -1.08455129e-01 -1.97424591e-01 -4.55641270e-01
1.35711044e-01 9.63537693e-01 4.35014188e-01 -3.42563063e-01
1.64440227e+00 2.61138976e-01 -5.97512163e-02 3.40904653e-01
1.06799722e+00 3.03312540e-01 -1.77640557e+00 6.32086545e-02
-8.79395604e-01 -6.44142568e-01 -2.26835519e-01 -7.84594297e-01
-1.11215270e+00 8.00461173e-01 2.51349658e-01 1.32060364e-01
7.15768695e-01 1.67310223e-01 7.49230206e-01 2.37704858e-01
8.29023540e-01 -4.84835565e-01 2.37652332e-01 4.14987832e-01
1.65880370e+00 -6.65696561e-01 -1.74411088e-01 -8.50777745e-01
-2.86554247e-01 1.24298060e+00 3.43240619e-01 -7.62983859e-01
8.10326219e-01 5.75338662e-01 -3.65978554e-02 -6.07377827e-01
-6.63165867e-01 3.51540744e-01 4.09821510e-01 6.53846741e-01
3.36241573e-01 4.37374145e-01 4.66460079e-01 1.17949590e-01
-6.94685757e-01 -1.33397475e-01 4.87529516e-01 5.55683494e-01
-1.89006224e-01 -1.00293112e+00 -5.17421782e-01 3.85415345e-01
-3.69812474e-02 -4.58850265e-02 -2.05097452e-01 8.04803848e-01
-5.03277639e-03 9.11826715e-02 3.80954772e-01 -2.20549345e-01
4.70209152e-01 -1.24369906e-02 8.94076228e-01 -8.89477372e-01
-1.43994883e-01 2.37550750e-01 -8.44084751e-03 -7.85529017e-01
7.87163153e-03 -3.75300050e-01 -1.18623149e+00 -4.41612095e-01
4.03139651e-01 -3.28043222e-01 4.11623776e-01 5.07213116e-01
6.50360942e-01 4.16681468e-01 6.25267148e-01 -1.67542613e+00
-9.20455381e-02 -6.24081194e-01 -8.12282443e-01 8.00427735e-01
2.65311033e-01 -8.08925390e-01 -2.51847297e-01 2.69960940e-01] | [8.996967315673828, -3.576079845428467] |
ab7daef9-78c8-48a2-8061-3c46fc464572 | graph-based-stock-recommendation-by-time | null | null | https://dl.acm.org/doi/10.1145/3451397 | https://dl.acm.org/doi/pdf/10.1145/3451397 | Graph-Based Stock Recommendation by Time-Aware Relational Attention Network | The stock market investors aim at maximizing their investment returns. Stock recommendation task is to recommend stocks with higher return ratios for the investors. Most stock prediction methods study the historical sequence patterns to predict stock trend or price in the near future. In fact, the future price of a stock is correlated not only with its historical price, but also with other stocks. In this article, we take into account the relationships between stocks (corporations) by stock relation graph. Furthermore, we propose a Time-aware Relational Attention Network (TRAN) for graph-based stock recommendation according to return ratio ranking. In TRAN, the time-aware relational attention mechanism is designed to capture time-varying correlation strengths between stocks by the interaction of historical sequences and stock description documents. With the dynamic strengths, the nodes of the stock relation graph aggregate the features of neighbor stock nodes by graph convolution operation. For a given group of stocks, the proposed TRAN model can output the ranking results of stocks according to their return ratios. The experimental results on several real-world datasets demonstrate the effectiveness of our TRAN for stock recommendation. | ['Zhao Li', 'Shichao Zhang', 'Jianxin Wang', 'Cong Xu', 'Xiaoting Ying', 'Jianliang Gao'] | 2022-02-01 | null | null | null | acm-transactions-on-knowledge-discovery-from-3 | ['stock-prediction'] | ['time-series'] | [-7.66406178e-01 -4.08732086e-01 -5.07734358e-01 -2.92912662e-01
3.82375687e-01 -5.93906641e-01 5.73322415e-01 2.50813048e-02
-1.71772867e-01 3.11702549e-01 8.27186644e-01 -3.63853693e-01
-4.72637266e-01 -1.54986167e+00 -5.45252323e-01 -3.37250799e-01
-3.44476998e-01 2.71922559e-01 5.28237462e-01 -6.77038670e-01
6.72225118e-01 5.40918946e-01 -1.15802264e+00 1.31721720e-01
4.40441221e-01 1.24451983e+00 2.62848854e-01 1.69736415e-01
-7.32106924e-01 1.41308999e+00 -5.47986090e-01 -8.17501605e-01
5.13310194e-01 -1.04317293e-01 -1.87433511e-01 -5.16337827e-02
4.80182767e-02 -6.83724344e-01 -9.62831438e-01 1.23367298e+00
5.81738427e-02 3.80146615e-02 3.80410880e-01 -1.13572705e+00
-1.33420038e+00 1.55784750e+00 -4.96405452e-01 9.46294069e-01
-4.39301342e-01 -4.13411260e-02 1.64578032e+00 -7.55679846e-01
3.93886387e-01 8.68359268e-01 2.85759896e-01 -1.18492141e-01
-3.83149296e-01 -8.74528527e-01 7.83726871e-01 2.38037497e-01
-7.78292835e-01 2.93072522e-01 1.05008161e+00 -5.31064153e-01
9.64613259e-01 1.36774853e-01 1.35816681e+00 1.81828290e-01
4.65027362e-01 5.26572347e-01 3.87304336e-01 5.11612236e-01
-2.00129598e-01 -8.48116875e-02 4.79720026e-01 4.46956277e-01
7.42719173e-01 2.94067897e-04 -4.78637427e-01 2.12101638e-01
7.89476812e-01 8.14229369e-01 -1.67260483e-01 1.28183126e-01
-1.35098839e+00 8.82849216e-01 1.11849535e+00 7.54857898e-01
-6.66695714e-01 4.11713988e-01 2.55992353e-01 5.57926714e-01
9.03156340e-01 5.14220953e-01 -3.13163519e-01 4.18835819e-01
-5.90408981e-01 -1.70700662e-02 6.77155435e-01 1.01497650e+00
5.53098619e-01 5.79846323e-01 -1.67610973e-01 9.99826267e-02
5.31152546e-01 3.06286871e-01 8.88705373e-01 -2.43876547e-01
7.19112694e-01 8.61849189e-01 6.78189918e-02 -1.55964160e+00
-4.02667522e-01 -1.10750079e+00 -6.73641860e-01 -2.25302026e-01
-1.17685467e-01 -3.59176807e-02 -5.67229271e-01 1.15781057e+00
-2.10677192e-01 7.14559019e-01 1.18724860e-01 8.83814037e-01
7.86169946e-01 1.20258188e+00 -2.07880750e-01 -3.85965854e-01
9.32061076e-01 -8.32022667e-01 -8.16674769e-01 -3.52780335e-02
2.59876370e-01 -2.91422069e-01 5.16853988e-01 -1.89337283e-01
-7.93282628e-01 -3.82963181e-01 -1.02743459e+00 4.03677374e-01
-5.23420393e-01 -2.10530296e-01 6.00113988e-01 7.00896457e-02
-1.07260144e+00 1.09273136e+00 -4.13616538e-01 2.20870316e-01
1.91220250e-02 3.57735842e-01 4.70000058e-01 4.92056251e-01
-1.63005161e+00 6.40172780e-01 4.73791867e-01 4.08840477e-01
-2.16193765e-01 -8.54003012e-01 -4.55942005e-01 4.74318326e-01
3.05534601e-01 -1.58846304e-01 8.37874472e-01 -7.64254928e-01
-9.94481266e-01 1.22734010e-01 5.66637337e-01 -8.63648474e-01
3.21611822e-01 7.35579357e-02 -7.78516471e-01 -2.43987337e-01
-1.13309652e-01 -1.95348382e-01 3.85676146e-01 -5.61871767e-01
-9.82401192e-01 -6.75935596e-02 3.95285368e-01 2.36504018e-01
-7.54199028e-01 -1.92364424e-01 -3.54048073e-01 -9.56807494e-01
-2.42351219e-02 -3.91891330e-01 -2.11627632e-01 -7.53737748e-01
-1.74080595e-01 -2.90928453e-01 4.25161332e-01 -5.27605057e-01
1.57036698e+00 -2.00483227e+00 -5.11740334e-02 5.05314291e-01
5.07006466e-01 -3.42805013e-02 -1.24505498e-01 5.96617341e-01
-1.35517552e-01 2.78570294e-01 3.62731576e-01 5.61698616e-01
9.74858403e-02 -1.11233830e-01 -8.77626419e-01 2.17207849e-01
6.22784570e-02 1.35600972e+00 -9.94643390e-01 -3.71308550e-02
-1.08651102e-01 9.58950520e-02 -1.77255034e-01 2.36723810e-01
-4.37931389e-01 -3.98521572e-01 -8.39830935e-01 3.24615121e-01
3.11083704e-01 -6.30067468e-01 -2.17399076e-02 -3.73549312e-01
-4.54127312e-01 5.43534398e-01 -9.28807676e-01 5.45991063e-01
-8.68052840e-02 5.28306127e-01 -7.44175851e-01 -4.13524717e-01
1.33580196e+00 -1.17512472e-01 4.93984580e-01 -7.93373942e-01
1.18578412e-01 3.61989737e-01 2.77887166e-01 -2.40385476e-02
9.34265792e-01 -5.02014905e-02 3.70942727e-02 8.13343525e-01
-2.65570641e-01 3.42367679e-01 3.77583772e-01 4.94059980e-01
8.54978263e-01 -3.57013613e-01 6.02564104e-02 -3.54837894e-01
3.07252973e-01 -2.52426714e-01 4.77496594e-01 2.60833740e-01
2.89218873e-01 -1.07069574e-02 7.44147956e-01 -8.11132967e-01
-8.29064250e-01 -6.37683153e-01 4.93044585e-01 7.98894405e-01
2.11978361e-01 -8.79761279e-02 3.42008732e-02 -5.28534532e-01
4.46862578e-01 6.54223025e-01 -7.16204941e-01 -2.17026383e-01
-4.28356349e-01 -7.15495050e-01 -2.44726464e-01 6.60035312e-01
5.24465680e-01 -1.27649367e+00 -2.28245884e-01 4.84314501e-01
4.51392114e-01 -5.78178108e-01 -1.12528515e+00 -2.19268411e-01
-7.86939502e-01 -8.31370950e-01 -1.05314016e+00 -5.73348701e-01
4.60657954e-01 4.82227296e-01 1.12226987e+00 5.42519510e-01
4.91283327e-01 3.71533513e-01 -4.36825752e-01 -4.80326861e-01
1.84192911e-01 7.43671283e-02 -1.17264450e-01 5.70825279e-01
1.88160107e-01 -5.98935604e-01 -7.64677763e-01 1.46690935e-01
-8.67000461e-01 -2.54314244e-01 6.71903014e-01 4.04284418e-01
4.72413659e-01 6.70424223e-01 7.66219258e-01 -1.16692328e+00
1.02483165e+00 -9.36815083e-01 -1.20377302e+00 3.92227024e-01
-1.18862975e+00 9.86077338e-02 3.58059585e-01 -5.54377973e-01
-7.70558298e-01 -5.84912658e-01 4.14646417e-01 -5.43811142e-01
1.10111344e+00 1.31746602e+00 1.88569859e-01 3.84067208e-01
-2.09939107e-01 6.20470226e-01 -2.84522623e-01 -2.54840076e-01
1.43028140e-01 2.04778954e-01 2.84801602e-01 2.17090473e-01
1.02142787e+00 -1.95225440e-02 4.97938767e-02 -6.54173791e-02
-9.70893681e-01 -2.80750990e-01 -3.82072479e-01 -5.77085555e-01
6.26036465e-01 -7.90378571e-01 -7.52559423e-01 5.54595232e-01
-8.07118773e-01 -7.95937479e-02 -2.04278931e-01 6.27906322e-01
8.25278908e-02 -1.16983719e-01 -9.46350396e-01 -9.06754315e-01
-6.36881590e-01 -7.41163254e-01 3.32565963e-01 2.48285711e-01
3.54908913e-01 -1.14746046e+00 -1.58667505e-01 -2.73633510e-01
5.82283914e-01 -1.26437214e-03 8.77324522e-01 -1.09973860e+00
-1.33597028e+00 -4.60893840e-01 -4.33940828e-01 4.47329693e-02
3.56844157e-01 2.42574960e-02 -2.26154879e-01 -6.53481996e-03
-1.24069430e-01 5.84934413e-01 1.05882943e+00 3.87014568e-01
5.93105197e-01 -7.12245226e-01 4.62154262e-02 3.81148785e-01
1.51402688e+00 6.76285982e-01 4.54477698e-01 4.75273460e-01
1.04579294e+00 7.06473768e-01 5.89110136e-01 4.28591251e-01
4.57408726e-01 4.90347147e-02 4.25609499e-01 4.11436856e-01
1.96481675e-01 -4.84912932e-01 3.35407168e-01 1.42717969e+00
-2.68428981e-01 -4.59123075e-01 -6.28871977e-01 4.21159208e-01
-1.82237911e+00 -1.33913040e+00 -2.60080904e-01 1.93387413e+00
2.63489366e-01 7.39328265e-01 2.85175502e-01 -2.20039085e-01
8.42308939e-01 6.57947719e-01 -8.39169502e-01 4.16637838e-01
-1.74030364e-01 -5.82782209e-01 1.07922125e+00 3.08408260e-01
-6.67430818e-01 6.80541456e-01 5.20739794e+00 3.70784134e-01
-1.24655867e+00 -4.64102358e-01 6.81663811e-01 -1.98322088e-01
-1.04446983e+00 -2.46550173e-01 -8.81497979e-01 9.45061326e-01
8.36206973e-01 -1.18252563e+00 5.30472875e-01 7.70958781e-01
-2.16507658e-01 5.62246859e-01 -6.22123778e-01 4.71347094e-01
1.20925745e-02 -1.99122977e+00 3.98496211e-01 1.23730458e-01
5.17789721e-01 1.32126182e-01 3.93365264e-01 1.51637927e-01
4.94258612e-01 -4.28316325e-01 1.12775779e+00 1.07822621e+00
-4.24230248e-02 -1.14392233e+00 1.05759597e+00 6.49315864e-02
-1.88726890e+00 -3.67817491e-01 -5.50135374e-01 1.06828921e-01
1.18552461e-01 6.61534727e-01 -5.53014219e-01 7.16188848e-01
7.38927364e-01 1.51166761e+00 -4.42489892e-01 8.29460740e-01
-4.89531234e-02 4.40838546e-01 1.86420351e-01 -9.20739055e-01
4.58381206e-01 -7.91409552e-01 2.37761319e-01 6.98739052e-01
7.39521801e-01 6.09829962e-01 -2.36983985e-01 7.28073359e-01
-3.64286780e-01 1.93946347e-01 -6.19298697e-01 -7.64402032e-01
3.72270942e-01 9.76456881e-01 -9.43676770e-01 -3.94612908e-01
-5.81051171e-01 3.08987319e-01 2.52419382e-01 2.64112085e-01
-5.01283824e-01 -4.29196835e-01 4.37434733e-01 3.56572300e-01
5.73882163e-01 -1.13569692e-01 8.81025859e-04 -9.56447363e-01
-1.76045150e-01 -1.76329583e-01 5.68859756e-01 -9.87527788e-01
-1.72432387e+00 9.11486387e-01 -2.13744491e-01 -1.49546540e+00
1.38214171e-01 -5.72358787e-01 -1.03016841e+00 8.80459666e-01
-1.77487195e+00 -7.20994532e-01 1.67790174e-01 3.50947022e-01
1.84336722e-01 -8.06934118e-01 -1.96940199e-01 2.77864188e-01
-5.01527905e-01 -3.47955637e-02 -4.51889485e-02 4.93629336e-01
-1.28912479e-01 -1.41484559e+00 8.31472874e-01 8.19688916e-01
3.59241039e-01 6.50750816e-01 4.61097062e-01 -1.31812203e+00
-1.41841817e+00 -1.31846011e+00 9.68244731e-01 -3.53139676e-02
1.60221338e+00 1.62744194e-01 -1.13971055e+00 9.23217714e-01
2.31461734e-01 -5.22047617e-02 3.13285649e-01 -1.88928291e-01
-4.06699747e-01 -6.28977478e-01 -5.34297407e-01 5.85864782e-01
1.00670362e+00 -2.01715022e-01 -6.68820739e-01 3.48962814e-01
1.48512650e+00 2.82601956e-02 -8.63657534e-01 1.14388049e-01
3.05277735e-01 -8.61822307e-01 8.34582508e-01 -5.75970888e-01
3.87066573e-01 -2.67753959e-01 8.27976242e-02 -1.49289715e+00
-1.02474773e+00 -3.17806393e-01 -1.15178190e-01 1.36404574e+00
5.75167298e-01 -1.02893794e+00 7.01288044e-01 5.65846622e-01
-8.61402750e-02 -5.56523800e-01 -3.83779258e-01 -6.53352916e-01
-2.74725884e-01 -1.19933942e-02 1.53222001e+00 8.53586733e-01
-3.63992117e-02 4.18369353e-01 -4.66037840e-01 2.45062590e-01
3.20672512e-01 8.61917257e-01 3.57628637e-03 -1.34232986e+00
-2.15824157e-01 -9.63182747e-01 -4.16913509e-01 -7.78818369e-01
1.56232670e-01 -1.15062785e+00 -4.15207118e-01 -1.76483655e+00
-5.91706261e-02 -2.55506486e-01 -1.05840552e+00 -1.92414597e-02
-2.39787728e-01 -2.28067070e-01 3.77861470e-01 5.14000416e-01
-4.32160467e-01 4.29846674e-01 1.42293525e+00 -4.60800737e-01
-1.31859735e-01 4.24785703e-01 -7.33494997e-01 3.29179883e-01
8.40942502e-01 -2.99778461e-01 -6.19445562e-01 -3.30096453e-01
1.01914394e+00 2.86789179e-01 -2.07276091e-01 -5.31663716e-01
4.93269950e-01 -4.14540797e-01 2.17826560e-01 -1.10489607e+00
-5.89389652e-02 -9.01184976e-01 4.92178440e-01 8.45736265e-01
-5.29191792e-01 7.19841063e-01 -2.76370138e-01 8.40245068e-01
-3.10635477e-01 -1.15622304e-01 1.65686399e-01 -2.40915343e-01
-8.92511666e-01 9.23924446e-01 -2.04776257e-01 -3.86010230e-01
7.98575222e-01 2.67975032e-01 -5.75721860e-01 -7.11018205e-01
-2.69913375e-01 5.37385046e-01 -1.62584975e-01 6.25125170e-01
7.76510656e-01 -1.53067088e+00 -7.51076043e-01 -1.20122954e-01
4.72748317e-02 -4.64000732e-01 7.92663395e-02 3.45731556e-01
-4.72731054e-01 4.52863485e-01 -1.16740823e-01 2.63674438e-01
-7.21865535e-01 1.10737717e+00 5.66075742e-01 -4.38164711e-01
-5.62738240e-01 8.85252595e-01 5.58001287e-02 1.79154754e-01
1.14751488e-01 -5.27763724e-01 -9.55591977e-01 6.20943129e-01
7.92530656e-01 3.18808585e-01 -1.68051198e-01 -4.43066686e-01
-3.14257108e-02 5.50097704e-01 -3.23727839e-02 4.39021997e-02
1.85887265e+00 5.77457948e-03 -4.78151232e-01 7.95761585e-01
8.79374325e-01 1.96830556e-01 -9.18272495e-01 -5.32913983e-01
5.29643238e-01 -5.26603758e-01 2.70520031e-01 -2.81403154e-01
-2.15477943e+00 3.32744688e-01 5.82986185e-03 9.85177755e-01
8.68679881e-01 7.61529803e-03 8.65527868e-01 4.40292805e-01
3.67029697e-01 -5.63814342e-01 1.26360863e-01 4.21037167e-01
1.14777970e+00 -6.44180059e-01 1.47457346e-01 -1.98617384e-01
-7.52173841e-01 1.13565135e+00 5.73347569e-01 -5.44134676e-01
1.29743671e+00 1.50443226e-01 9.72088799e-03 -5.34012079e-01
-8.22121203e-01 -2.90176392e-01 7.03957319e-01 -1.00954160e-01
1.85923740e-01 1.11190546e-02 -1.49219483e-01 7.29973435e-01
-4.14231956e-01 -2.84110695e-01 6.56169295e-01 5.76882780e-01
-7.04819739e-01 -6.93385839e-01 1.28332779e-01 1.10111213e+00
-4.30080891e-01 -3.92604649e-01 -3.27849686e-01 5.29128015e-01
-3.40659499e-01 5.78533530e-01 7.78353453e-01 -8.41658652e-01
6.33167744e-01 -3.46425235e-01 -2.40736574e-01 -6.44236982e-01
-8.36840630e-01 2.02953398e-01 -1.36887506e-01 -1.41111955e-01
-2.02141896e-01 -7.37543643e-01 -1.39200795e+00 -5.24681449e-01
-1.37768358e-01 3.11099380e-01 1.34659588e-01 6.27278626e-01
2.50821084e-01 9.56650019e-01 1.04967928e+00 -3.91361684e-01
-4.82468069e-01 -7.91171074e-01 -1.39699662e+00 9.48533416e-02
3.96839380e-01 -4.22188193e-01 -2.84944564e-01 -2.68432975e-01] | [4.312275409698486, 4.340070724487305] |
11c75bda-9e11-413f-a35d-789857cc841f | high-order-interaction-for-weakly-supervised | null | null | https://www.sciencedirect.com/science/article/abs/pii/S0925231221013060 | https://www.sciencedirect.com/science/article/abs/pii/S0925231221013060 | High-Order-Interaction for weakly supervised Fine-Grained Visual Categorization | Fine-Grained Visual Categorization (FGVC) is a challenging task due to the large intra-subcategory and small inter-subcategory variances. Recent studies tackle this task through a weakly supervised manner without using the part annotation from the experts. Of those, methods based on bilinear pooling are one of the main categories for computing the interaction between deep features and have shown high effectiveness. However, these methods mainly focus on the correlation within one specific layer but largely ignore the high interactions between multiple layers. In this study, we argue that considering the high interaction between the features from multiple layers can help to learn more distinguishing fine-grained features. To this end, we propose a High-Order-Interaction (HOI) method for FGVC. In our HOI, an efficient cross-layer trilinear pooling is introduced to calculate the third-order interaction between three different layers. Third-order interactions of different combinations are then fused to form the final representation. HOI can produce more discriminative representations and be readily integrated with the two popular techniques, attention mechanism and triplet loss, to obtain superposed improvement. Extensive experiments conducted on four FGVC datasets show the great superiority of our method over bilinear-based methods and demonstrate that the proposed method achieves the state of the art. | ['Shaozi Li', 'Zhun Zhong', 'Zhimin Luo', 'Junzhen Wang', 'Nanyu Li'] | 2021-11-13 | high-order-interaction-for-weakly-supervised-1 | https://github.com/puallee/HOI-Net | https://github.com/puallee/HOI-Net | neurocomputing-2021-11 | ['fine-grained-image-recognition', 'fine-grained-image-classification', 'fine-grained-visual-categorization'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.87471533e-01 -5.58952868e-01 8.93994793e-02 -5.44216514e-01
-6.90362692e-01 -4.05001909e-01 6.54307961e-01 -2.32205540e-03
-4.29160774e-01 4.49256718e-01 2.17910886e-01 1.04220480e-01
-3.03618908e-01 -6.79997087e-01 -6.41473591e-01 -8.87342513e-01
1.49734125e-01 -1.66180491e-01 5.47255039e-01 -3.50399539e-02
3.19364756e-01 3.67265105e-01 -1.60896170e+00 6.70186996e-01
1.10192585e+00 1.53604269e+00 1.69837117e-01 -3.15748267e-02
-4.22202796e-01 6.00817442e-01 -3.41982186e-01 -2.49270126e-01
1.85342550e-01 -1.51265636e-01 -5.17652333e-01 -6.01423569e-02
3.37181687e-01 -1.42254025e-01 -5.85120730e-02 1.15198362e+00
4.95776951e-01 -5.40980324e-03 8.84968340e-01 -1.26649582e+00
-7.49335170e-01 4.21517938e-01 -9.64706123e-01 1.48680866e-01
-1.11051813e-01 -4.91792522e-03 1.05596626e+00 -1.22160304e+00
6.05927929e-02 1.45648444e+00 6.20501161e-01 4.27304506e-02
-1.03117633e+00 -9.16470110e-01 6.41118586e-01 5.75379252e-01
-1.73960805e+00 -2.15522442e-02 8.98494005e-01 -5.97699821e-01
6.36098444e-01 2.16068268e-01 3.57712209e-01 6.28522635e-01
1.78005453e-02 7.76869118e-01 1.41303587e+00 -2.17298925e-01
-1.17870003e-01 2.92417526e-01 4.77808177e-01 8.21653247e-01
9.58217084e-02 -1.71543062e-01 -2.20684558e-01 -2.17191558e-02
5.71891308e-01 3.08821440e-01 -2.86342084e-01 -1.89142719e-01
-1.12863874e+00 7.89815903e-01 1.14107788e+00 5.37074685e-01
-3.51234645e-01 -5.72084375e-02 4.45990503e-01 -1.96173266e-02
4.70757902e-01 5.18156961e-02 -3.52089852e-01 3.34064424e-01
-6.59951806e-01 -4.54070792e-03 2.56179929e-01 6.97375417e-01
1.00995135e+00 -3.10906738e-01 -7.81897545e-01 1.11866796e+00
4.14651453e-01 -2.99100615e-02 6.47193789e-01 -5.28199315e-01
6.09276772e-01 1.07675803e+00 -1.53720766e-01 -1.29623973e+00
-2.88969606e-01 -6.91134691e-01 -1.16466391e+00 1.31423756e-01
2.91477829e-01 2.07992703e-01 -8.19933534e-01 1.57203841e+00
2.11083800e-01 9.86413471e-03 -4.14867729e-01 9.82776523e-01
1.02666950e+00 6.62021756e-01 3.44167829e-01 -8.83981287e-02
1.40895247e+00 -1.12506533e+00 -5.13288975e-01 -5.13839163e-02
2.03102320e-01 -7.74216413e-01 1.17635512e+00 2.24465936e-01
-6.57706082e-01 -9.34660017e-01 -1.08401895e+00 -1.66404277e-01
-5.77060342e-01 5.74282825e-01 6.92979097e-01 3.49837035e-01
-8.61615777e-01 5.38817286e-01 -4.95663166e-01 6.85943589e-02
6.70577168e-01 3.79872203e-01 -4.38424885e-01 -2.15491459e-01
-1.15951967e+00 6.12464786e-01 4.19867218e-01 4.27100956e-01
-4.24274921e-01 -6.22448325e-01 -6.15084946e-01 2.74375618e-01
1.82952523e-01 -3.47039938e-01 8.29588950e-01 -8.15959215e-01
-1.16646063e+00 6.21290505e-01 -2.24134207e-01 -3.92822400e-02
5.34854889e-01 -8.94107595e-02 -4.49520759e-02 5.75977042e-02
9.24634412e-02 6.45684600e-01 6.84131265e-01 -1.21508443e+00
-7.21776962e-01 -5.04371703e-01 1.25059262e-01 2.03488082e-01
-5.75837195e-01 -8.74139518e-02 -5.85718274e-01 -6.53104782e-01
2.92184651e-01 -7.29221582e-01 -5.38502298e-02 9.98717844e-02
-3.95031571e-01 -8.32139671e-01 7.93871403e-01 -6.00487709e-01
1.30441439e+00 -2.16808009e+00 2.42419858e-02 9.36680883e-02
2.44501233e-01 4.11723703e-01 6.97552562e-02 1.98032528e-01
1.44183651e-01 1.98655754e-01 3.92288342e-02 -1.53514192e-01
1.13540679e-01 -2.33265430e-01 -8.15779865e-02 3.20742786e-01
4.27039981e-01 9.01334643e-01 -5.02248287e-01 -7.33333230e-01
2.53159583e-01 6.02195263e-01 -4.89133716e-01 2.40250126e-01
1.98188275e-01 2.52508163e-01 -6.51675820e-01 5.32600999e-01
1.09128273e+00 -2.73270369e-01 -1.56989783e-01 -8.68118227e-01
-3.71690094e-01 -1.52055144e-01 -1.06720901e+00 1.26760864e+00
-4.18688953e-01 1.94467887e-01 -1.04237452e-01 -1.04879487e+00
1.02038229e+00 4.13723029e-02 1.40864894e-01 -6.59911513e-01
3.88845205e-01 2.92711288e-01 1.64556667e-01 -3.88420224e-01
7.13897049e-02 -1.02000318e-01 -1.19836427e-01 1.39306441e-01
1.97702259e-01 2.86582798e-01 1.00804560e-01 -2.33376697e-02
5.62871337e-01 -4.49847318e-02 3.83064657e-01 -3.17317247e-01
8.88758063e-01 -4.71296251e-01 6.93177044e-01 5.95785439e-01
-4.10587788e-01 6.73916936e-01 4.55314964e-01 -3.24929327e-01
-6.67963982e-01 -7.40067124e-01 -2.94358581e-01 1.03491187e+00
3.80345583e-01 -2.38309219e-01 -6.87458515e-01 -8.88989449e-01
1.22185916e-01 1.85026094e-01 -8.42692316e-01 -2.34070882e-01
-4.40732896e-01 -9.45789814e-01 2.47412384e-01 8.32327843e-01
1.08203816e+00 -8.72572780e-01 -9.65311229e-02 7.13292882e-02
-3.32348824e-01 -9.93117571e-01 -5.66977382e-01 6.72366917e-02
-5.58831096e-01 -8.69862795e-01 -9.94452953e-01 -9.59292114e-01
6.31872535e-01 5.28530300e-01 5.99963963e-01 7.56284148e-02
-3.44850659e-01 -1.74195111e-01 -4.40691888e-01 -2.72065103e-01
3.59164834e-01 -6.38558529e-03 -1.94247276e-01 5.44587672e-01
4.95809555e-01 -3.58422160e-01 -8.07005763e-01 4.38764542e-01
-6.71627283e-01 2.43666954e-03 9.46158528e-01 1.00495934e+00
7.34550476e-01 1.81336105e-01 4.98074025e-01 -6.68515742e-01
4.88681465e-01 -3.68831277e-01 -3.48661661e-01 4.29943800e-01
-3.15159202e-01 2.01500982e-01 9.54920769e-01 -4.50279146e-01
-1.26859593e+00 -1.04601592e-01 -9.75667164e-02 -4.71946687e-01
-3.73590477e-02 3.77675921e-01 -4.35745269e-01 -3.75668049e-01
2.89216667e-01 3.11119139e-01 -4.10552442e-01 -7.24797010e-01
2.81718314e-01 8.91927183e-01 1.57594249e-01 -5.26691318e-01
7.22710550e-01 4.42662120e-01 -1.59452900e-01 -4.57838356e-01
-9.36462581e-01 -5.59728205e-01 -8.10090125e-01 -9.16574150e-02
9.36653495e-01 -9.74436998e-01 -9.00177598e-01 6.73528969e-01
-1.08710301e+00 1.51217997e-01 1.75858498e-01 5.11425436e-01
-8.64668489e-02 3.97325039e-01 -4.77602214e-01 -7.91443527e-01
-3.98696065e-01 -1.29050744e+00 1.01725721e+00 5.91025352e-01
4.09218639e-01 -6.58523917e-01 -4.23455954e-01 2.45480448e-01
4.83360529e-01 -2.77833287e-02 1.02265954e+00 -4.92639840e-01
-5.45711279e-01 -1.20338954e-01 -1.03387761e+00 5.78302264e-01
2.01334506e-01 -7.80565143e-02 -1.09825146e+00 -1.76698133e-01
-2.54410416e-01 -3.94818604e-01 1.31200838e+00 1.79704413e-01
1.68294907e+00 -1.61029339e-01 -2.68475294e-01 5.88967443e-01
1.59781623e+00 1.17558539e-01 4.97136533e-01 2.92946637e-01
9.78849590e-01 7.32380450e-01 6.74157679e-01 3.58725607e-01
4.92922783e-01 8.08726609e-01 2.32879251e-01 -1.67174578e-01
-2.17749223e-01 -1.69872254e-01 6.36402518e-02 9.75791395e-01
-3.37318659e-01 1.24702491e-01 -5.55545032e-01 3.44801843e-01
-1.84232426e+00 -8.09220612e-01 -2.60106772e-01 2.01189709e+00
7.03297317e-01 5.50261736e-02 -1.01270758e-01 2.14845896e-01
8.82385373e-01 1.80150926e-01 -4.57277536e-01 -3.36908817e-01
-2.18319207e-01 2.27252692e-02 3.56975555e-01 7.72014186e-02
-1.44849157e+00 8.39745998e-01 4.95306683e+00 1.37702334e+00
-1.09033012e+00 1.06464036e-01 7.67554820e-01 2.67046869e-01
-3.84828113e-02 -2.00406641e-01 -8.69156241e-01 6.76504791e-01
2.54234463e-01 3.57334986e-02 2.41762087e-01 6.64387405e-01
-3.54498103e-02 7.26267602e-03 -7.76557207e-01 1.10695243e+00
5.35238869e-02 -9.08298016e-01 2.31587484e-01 -4.81510796e-02
7.19564795e-01 -3.59262794e-01 6.62098676e-02 5.00476062e-01
7.06197135e-03 -8.96045864e-01 7.24880934e-01 6.33528709e-01
6.76909089e-01 -7.81161427e-01 9.85820711e-01 2.89473414e-01
-1.76922083e+00 -3.72005314e-01 -7.06857324e-01 4.33398150e-02
-1.76837519e-01 6.06481254e-01 -2.66869247e-01 8.26821625e-01
1.01484132e+00 7.29475081e-01 -8.38096321e-01 1.31504548e+00
-2.07620651e-01 4.46093440e-01 -8.81013870e-02 -1.34107381e-01
3.63030434e-01 -2.13860035e-01 1.12139121e-01 1.21683025e+00
2.25829348e-01 1.65982321e-01 2.53128678e-01 7.74594486e-01
-6.64741099e-02 4.39124823e-01 -2.10587293e-01 1.71062529e-01
1.62963927e-01 1.48912287e+00 -6.62315428e-01 -3.68658274e-01
-5.96420765e-01 8.72045100e-01 7.02981234e-01 3.37512374e-01
-7.48560548e-01 -7.64985919e-01 4.28397119e-01 -1.43095016e-01
5.88326573e-01 3.53933573e-02 -4.35339361e-01 -1.20547724e+00
3.27207029e-01 -6.54463172e-01 5.06789505e-01 -4.98255521e-01
-1.71990609e+00 7.29218006e-01 -7.11802840e-02 -1.44119692e+00
2.34666780e-01 -6.36347830e-01 -6.08592153e-01 1.06145835e+00
-1.71918380e+00 -1.50187159e+00 -6.64584160e-01 6.65190697e-01
4.43769515e-01 4.72590923e-02 5.52350879e-01 4.69890743e-01
-7.98732400e-01 8.38380039e-01 5.09348065e-02 1.37331888e-01
6.40520513e-01 -1.11894643e+00 -3.12095255e-01 6.27563953e-01
-1.78476989e-01 7.26718545e-01 1.27724379e-01 -3.97451252e-01
-8.45476449e-01 -1.31139612e+00 7.92035103e-01 3.40312757e-02
4.50780779e-01 -4.23778564e-01 -1.17558432e+00 2.35456318e-01
4.36800048e-02 3.42002273e-01 6.02118373e-01 5.08011952e-02
-7.03882813e-01 -5.71858943e-01 -1.03177869e+00 2.81922877e-01
9.66291070e-01 -5.00536680e-01 -6.10003233e-01 2.03808714e-02
5.75850070e-01 1.34750620e-01 -6.58753455e-01 6.85077488e-01
7.19471097e-01 -1.08825159e+00 9.48726892e-01 -3.25296462e-01
4.45435703e-01 -5.56520283e-01 -1.94885090e-01 -1.20695424e+00
-7.33567774e-01 1.24825686e-01 2.19145402e-01 1.62687886e+00
1.12816691e-01 -7.36635208e-01 3.74127239e-01 4.14600044e-01
-6.64750636e-02 -1.01874506e+00 -7.37918258e-01 -7.65170813e-01
1.68988824e-01 -6.55279458e-02 5.50275803e-01 6.75132155e-01
-1.99051231e-01 2.37167314e-01 -3.05406094e-01 7.76089877e-02
6.25706673e-01 5.78361809e-01 4.67014134e-01 -1.31755853e+00
-6.60265312e-02 -6.05934978e-01 -6.28445208e-01 -8.95140052e-01
2.88623571e-03 -9.57365632e-01 1.40088588e-01 -1.72684360e+00
7.57034421e-01 -6.93051815e-01 -7.70145774e-01 5.29967427e-01
-5.75587511e-01 6.37590051e-01 3.01593661e-01 3.76959920e-01
-7.12909818e-01 8.05497408e-01 1.30972719e+00 -2.03704521e-01
4.30531204e-02 -1.94758356e-01 -1.01212883e+00 8.38445544e-01
6.46833718e-01 -2.28497788e-01 -9.06507969e-02 -3.22050095e-01
-4.13670599e-01 -5.78889310e-01 4.72936571e-01 -1.03833556e+00
3.23621333e-01 2.66357251e-02 8.40391576e-01 -6.34666920e-01
2.38510221e-01 -7.19361663e-01 -1.46336406e-01 2.38253340e-01
-2.37943262e-01 -3.46103311e-01 1.37381881e-01 5.13316929e-01
-6.66930318e-01 8.08480680e-02 9.03175652e-01 -1.40925467e-01
-7.98638105e-01 5.02093792e-01 1.36237368e-01 3.19223069e-02
1.01191545e+00 -1.32802278e-02 -2.99661458e-01 -1.32651748e-02
-4.22134995e-01 2.71608412e-01 -1.57382786e-02 4.46868181e-01
4.94415760e-01 -1.66223860e+00 -6.39032602e-01 1.84010550e-01
2.50280529e-01 -2.55661100e-01 7.25459576e-01 9.49822366e-01
-9.48410928e-02 6.20631635e-01 -3.91382813e-01 -5.84601998e-01
-1.15071154e+00 5.50855041e-01 2.36761078e-01 -4.41273212e-01
-2.81370133e-01 1.05811882e+00 1.06290424e+00 -2.70873398e-01
2.16117769e-01 -2.68383145e-01 -7.38792181e-01 3.61447453e-01
6.59447134e-01 3.08390319e-01 1.28272146e-01 -1.01002610e+00
-6.13377154e-01 1.08748770e+00 -2.81204343e-01 3.75517577e-01
1.17962790e+00 -1.01520397e-01 -1.38675690e-01 3.64689231e-01
1.66954422e+00 -1.08019993e-01 -1.37538612e+00 -4.48760360e-01
-2.41289064e-01 -6.84886932e-01 8.49351212e-02 -7.21488237e-01
-1.30424273e+00 1.38444519e+00 7.09571362e-01 9.43260193e-02
1.32081914e+00 1.08071893e-01 5.97696722e-01 5.10801263e-02
4.04586345e-01 -8.54634047e-01 2.67032385e-02 2.87893623e-01
1.11267114e+00 -1.31769872e+00 -1.38557209e-02 -7.69510627e-01
-5.86862743e-01 1.03706193e+00 8.61067951e-01 -2.71891087e-01
7.88114667e-01 -2.11676300e-01 -7.67356753e-02 8.17815736e-02
-4.62854505e-01 -4.83167708e-01 7.17828989e-01 3.14629465e-01
4.37025070e-01 6.44142106e-02 -6.79307818e-01 1.14148414e+00
1.65659681e-01 -1.35662584e-02 -1.52682453e-01 5.11497736e-01
-3.87791276e-01 -1.05368042e+00 -2.41410360e-01 7.30876982e-01
-6.02599263e-01 -1.30976364e-01 -2.51843005e-01 6.10850215e-01
5.73979735e-01 9.34256136e-01 -6.45944774e-02 -5.34421206e-01
4.88780022e-01 -1.44349664e-01 4.18654978e-01 -3.15585673e-01
-6.77709281e-01 1.08725101e-01 -3.48402292e-01 -5.71024299e-01
-2.95847416e-01 -4.50673401e-01 -1.08195448e+00 8.19165781e-02
-3.99453312e-01 1.55766204e-01 5.00854671e-01 9.15345788e-01
3.88347745e-01 6.41639411e-01 9.23738360e-01 -1.03676486e+00
-6.71321869e-01 -1.02078414e+00 -6.24694824e-01 6.31587505e-01
2.37374634e-01 -1.06766367e+00 -5.28054893e-01 -2.98959941e-01] | [9.672012329101562, 1.976745843887329] |
03f16acd-0d90-4ee3-a78f-27954f0dae7c | smc-uda-structure-modal-constraint-for | 2306.08213 | null | https://arxiv.org/abs/2306.08213v1 | https://arxiv.org/pdf/2306.08213v1.pdf | SMC-UDA: Structure-Modal Constraint for Unsupervised Cross-Domain Renal Segmentation | Medical image segmentation based on deep learning often fails when deployed on images from a different domain. The domain adaptation methods aim to solve domain-shift challenges, but still face some problems. The transfer learning methods require annotation on the target domain, and the generative unsupervised domain adaptation (UDA) models ignore domain-specific representations, whose generated quality highly restricts segmentation performance. In this study, we propose a novel Structure-Modal Constrained (SMC) UDA framework based on a discriminative paradigm and introduce edge structure as a bridge between domains. The proposed multi-modal learning backbone distills structure information from image texture to distinguish domain-invariant edge structure. With the structure-constrained self-learning and progressive ROI, our methods segment the kidney by locating the 3D spatial structure of the edge. We evaluated SMC-UDA on public renal segmentation datasets, adapting from the labeled source domain (CT) to the unlabeled target domain (CT/MRI). The experiments show that our proposed SMC-UDA has a strong generalization and outperforms generative UDA methods. | ['Zhicheng Jiao', 'Harrison Bai', 'Xinbo Gao', 'Rama Chellappa', 'Ihab Kamel', 'Li Yang', 'Lulu Bi', 'Jie Li', 'Zhusi Zhong'] | 2023-06-14 | null | null | null | null | ['unsupervised-domain-adaptation', 'self-learning'] | ['methodology', 'natural-language-processing'] | [ 4.48455215e-01 2.85203695e-01 -4.04592991e-01 -6.16719544e-01
-1.05399859e+00 -5.27163804e-01 3.46830398e-01 -1.28622964e-01
-1.55033335e-01 5.86791933e-01 2.90241331e-01 -5.38599305e-02
-2.36294195e-01 -8.51873994e-01 -6.52769446e-01 -1.10613251e+00
2.65612662e-01 9.77236807e-01 4.60989028e-01 -3.38490978e-02
6.20954111e-03 2.53719121e-01 -7.89940834e-01 3.84031594e-01
1.26469898e+00 7.65617073e-01 5.39105237e-01 1.74735844e-01
-4.07379478e-01 4.90552008e-01 -2.06152365e-01 9.84880105e-02
1.84816420e-01 -7.61692584e-01 -1.05683434e+00 3.77608955e-01
3.02434355e-01 -2.15156451e-01 -8.90370160e-02 1.15858364e+00
5.76534688e-01 -7.57658109e-02 1.07314777e+00 -8.81476104e-01
-9.28151608e-01 3.56179446e-01 -6.56624079e-01 3.10359567e-01
-2.30747327e-01 -5.18056601e-02 5.78626990e-01 -5.42413354e-01
1.13273668e+00 1.03226817e+00 7.15722203e-01 6.89950109e-01
-1.38270581e+00 -5.14275014e-01 2.61788636e-01 6.16750084e-02
-1.27028549e+00 2.77467631e-02 1.18754184e+00 -7.13972390e-01
2.65211433e-01 -1.81392804e-01 3.05493504e-01 1.15873015e+00
2.95642670e-02 1.01041710e+00 1.40632689e+00 -2.24293783e-01
3.77115101e-01 -9.12652686e-02 1.04163513e-01 6.54510558e-01
1.37254342e-01 4.05779891e-02 2.77897459e-03 -5.49495257e-02
1.09138930e+00 1.01870056e-02 -8.15499648e-02 -1.02486515e+00
-1.16434455e+00 8.45867693e-01 5.69431245e-01 4.13938671e-01
-3.13862562e-01 -4.47553933e-01 5.17420053e-01 8.20742324e-02
3.90976816e-01 -2.33055977e-03 -4.35977966e-01 3.63961041e-01
-8.59416425e-01 2.59683058e-02 5.31243980e-01 1.18926537e+00
9.48133945e-01 -9.06951055e-02 -4.11647946e-01 1.02654564e+00
4.04145390e-01 4.02533531e-01 7.80186951e-01 -6.96254551e-01
3.61270532e-02 7.53726482e-01 -4.39815372e-01 -6.24995053e-01
-5.07021248e-01 -5.05734682e-01 -1.03382421e+00 -3.38942893e-02
4.71825004e-01 -7.03193769e-02 -1.56327045e+00 1.67131245e+00
5.82903624e-01 2.93780982e-01 2.02198207e-01 1.05404449e+00
1.13380098e+00 3.32741439e-01 3.98517132e-01 -1.25591531e-01
1.31419158e+00 -8.90674114e-01 -5.36967456e-01 -3.24574828e-01
7.09708512e-01 -6.10558629e-01 9.27876472e-01 7.67650753e-02
-6.76656008e-01 -5.98520100e-01 -8.87352049e-01 -1.12286448e-01
-1.29354745e-01 8.35393965e-02 4.03890043e-01 5.74787736e-01
-7.55000591e-01 1.91379249e-01 -9.15047765e-01 -3.79103005e-01
8.33676875e-01 1.51740015e-01 -4.92412955e-01 -4.28328812e-01
-9.65269387e-01 6.09919131e-01 7.05439985e-01 -1.66724831e-01
-8.76428962e-01 -6.11050427e-01 -1.03448009e+00 -3.00229102e-01
3.06869566e-01 -8.05554926e-01 9.37475681e-01 -1.04431951e+00
-1.50028837e+00 1.44118583e+00 -8.22906010e-03 -3.20073605e-01
6.46656573e-01 1.86440244e-01 -1.84050173e-01 4.10484493e-01
5.26983500e-01 6.53350472e-01 8.13425422e-01 -1.48184752e+00
-4.32103872e-01 -5.97362757e-01 -4.42110002e-01 2.07330137e-01
3.43826897e-02 -5.58526516e-01 -3.49026680e-01 -8.58971298e-01
5.02688944e-01 -9.84282017e-01 -3.50481093e-01 -1.88273802e-01
-3.74759138e-01 -2.80996025e-01 9.86171007e-01 -6.57107472e-01
8.90857816e-01 -2.30499220e+00 2.58768260e-01 2.49913901e-01
1.68662697e-01 1.45676851e-01 5.26661910e-02 -1.57022238e-01
-1.23596191e-01 -1.98917091e-01 -9.34344649e-01 8.27035382e-02
-1.54307052e-01 7.09945560e-01 1.64113104e-01 3.94143850e-01
1.12800032e-01 9.34784651e-01 -9.62850213e-01 -1.16863501e+00
1.39516667e-01 1.59841880e-01 -6.75627410e-01 2.87844419e-01
-2.30530024e-01 1.03489482e+00 -8.49967062e-01 8.19784105e-01
9.75401103e-01 -5.35161436e-01 4.50281709e-01 -2.56991297e-01
2.33084142e-01 -1.48487568e-01 -1.00812161e+00 2.37112331e+00
-1.76286787e-01 5.54080419e-02 6.85842186e-02 -1.51651824e+00
1.13388300e+00 1.19605966e-01 8.11318815e-01 -9.08818722e-01
-1.65733676e-02 3.65011752e-01 8.96456465e-02 -6.94656491e-01
-1.79985955e-01 -4.35813308e-01 -1.83392510e-01 1.14138938e-01
4.25535500e-01 -1.47119723e-02 -2.27493308e-02 7.97349289e-02
8.46979499e-01 5.52961886e-01 2.94108897e-01 -5.91523290e-01
6.09612107e-01 2.59696364e-01 1.09555328e+00 5.86537659e-01
-4.63038236e-01 1.03783011e+00 3.70033979e-01 -4.14491892e-01
-9.47054207e-01 -1.48021650e+00 -6.16424441e-01 9.85930443e-01
5.59985697e-01 2.41486430e-01 -7.55915165e-01 -1.17806685e+00
1.23506330e-01 4.07347172e-01 -7.54756987e-01 -3.27395767e-01
-6.84228361e-01 -9.94983494e-01 3.35480511e-01 7.37936914e-01
9.05552208e-01 -8.80102038e-01 -1.42369851e-01 2.35126525e-01
-3.93234611e-01 -9.23308432e-01 -5.71912408e-01 2.34685883e-01
-1.17022860e+00 -9.58825529e-01 -1.26729500e+00 -1.45661986e+00
8.66336405e-01 -1.05958931e-01 1.05084968e+00 -4.85016972e-01
-2.48152375e-01 3.46625477e-01 -3.12319547e-01 -1.07773930e-01
-5.01384199e-01 2.44184479e-01 -3.97472054e-01 7.51164928e-02
4.47986424e-01 -6.50088847e-01 -8.52751732e-01 5.84541619e-01
-8.46259832e-01 3.30516025e-02 1.09098375e+00 1.02963364e+00
1.20216584e+00 -2.57045358e-01 8.08192968e-01 -1.48564219e+00
2.48562440e-01 -7.32084990e-01 -2.39434540e-01 3.34320456e-01
-7.82553434e-01 2.27317095e-01 1.56076446e-01 -4.63364512e-01
-1.47667968e+00 4.75625455e-01 -6.57489598e-02 -4.45675224e-01
-5.92080772e-01 5.43736041e-01 -4.07154351e-01 5.83387539e-02
9.47764218e-01 5.40859282e-01 5.04315607e-02 -6.72158599e-01
3.29752922e-01 4.20930952e-01 8.58557284e-01 -9.68761384e-01
5.08091390e-01 7.12959528e-01 3.63074476e-03 -6.07878387e-01
-7.14508176e-01 -6.64213181e-01 -9.21148658e-01 -3.29300389e-02
1.05388963e+00 -8.68760049e-01 7.03777745e-02 4.33166653e-01
-7.67997801e-01 -4.29452866e-01 -3.80319953e-01 4.86096561e-01
-7.66121984e-01 6.67907417e-01 -6.59489095e-01 -8.90339464e-02
-2.78608680e-01 -1.12676072e+00 1.15745544e+00 4.61949170e-01
1.15402006e-01 -1.11228681e+00 3.41865838e-01 3.67021620e-01
1.87025204e-01 5.27871072e-01 1.16134477e+00 -8.71328652e-01
-4.76318210e-01 1.25258833e-01 -3.90013278e-01 3.77198815e-01
3.99757773e-01 -6.37450278e-01 -7.20230699e-01 -2.49758795e-01
2.26499792e-02 -2.23991901e-01 9.50839102e-01 9.05630350e-01
1.10011840e+00 1.09626986e-01 -4.98053044e-01 8.68022680e-01
1.39664972e+00 2.93730736e-01 5.68177521e-01 3.22785974e-01
8.06149065e-01 3.67276430e-01 5.69357216e-01 1.57936573e-01
4.19901818e-01 2.33941257e-01 2.22901627e-01 -5.60193837e-01
-3.32521409e-01 -2.90602207e-01 -1.35905415e-01 6.56937063e-01
-1.16877295e-01 2.21711457e-01 -1.15254676e+00 8.65634739e-01
-1.75115299e+00 -5.44782102e-01 -7.16576949e-02 1.80629170e+00
1.01506221e+00 8.85582864e-02 2.07246304e-01 -5.37048697e-01
8.99037957e-01 -1.41163781e-01 -9.36673820e-01 8.12035128e-02
-9.29689929e-02 2.24807635e-01 5.63886106e-01 1.75988078e-01
-1.40136719e+00 7.47221649e-01 5.84178591e+00 8.63424182e-01
-9.96407926e-01 2.55094886e-01 5.97326636e-01 6.18470430e-01
-2.50480711e-01 -8.29676464e-02 -4.28215802e-01 5.05744696e-01
4.02251393e-01 -1.61876112e-01 -2.49834552e-01 1.20056796e+00
-2.40827322e-01 2.56579947e-02 -1.05402148e+00 9.91935194e-01
-9.36362669e-02 -1.28380561e+00 4.48303446e-02 2.16687813e-01
9.88849878e-01 6.98989406e-02 9.67992172e-02 5.19494176e-01
3.57657939e-01 -8.00876558e-01 1.44560412e-01 5.43604970e-01
7.56645322e-01 -4.10712928e-01 8.34119260e-01 2.99272299e-01
-1.13694549e+00 2.75191307e-01 -2.66156167e-01 5.42828381e-01
-6.02102354e-02 5.06346643e-01 -9.07693505e-01 8.31255019e-01
6.92932904e-01 9.27076459e-01 -4.22374547e-01 9.78951156e-01
-6.89910501e-02 6.26980007e-01 -9.80731025e-02 5.40411949e-01
2.62060165e-01 -4.86825198e-01 4.22502965e-01 1.33815396e+00
6.98829740e-02 1.41442895e-01 5.45128286e-01 1.06318140e+00
4.00324948e-02 2.20851392e-01 -5.09586394e-01 1.43318847e-01
1.57903761e-01 1.13026345e+00 -1.12707615e+00 -3.76628101e-01
-3.66038442e-01 1.02524507e+00 -1.34843932e-02 4.31824327e-01
-7.03589499e-01 -6.25224262e-02 1.78813159e-01 1.54407859e-01
3.39748263e-01 -6.35496974e-02 -4.46444005e-01 -1.08094203e+00
-1.77048400e-01 -6.14265084e-01 9.49490070e-01 -4.49259222e-01
-1.78717458e+00 4.00649041e-01 1.64410725e-01 -1.42825961e+00
-1.46638721e-01 -4.43179727e-01 -3.69316548e-01 7.89740145e-01
-1.68286252e+00 -1.52154434e+00 -4.11022395e-01 7.77754724e-01
5.05610049e-01 -1.13508806e-01 7.38432229e-01 5.35778165e-01
-2.49216646e-01 4.21590835e-01 3.47726077e-01 6.44097388e-01
1.03372812e+00 -1.52115846e+00 -1.28952295e-01 5.50876021e-01
-2.69263923e-01 4.62800860e-01 2.95154274e-01 -8.38204622e-01
-8.99552107e-01 -1.19120336e+00 2.65822470e-01 -2.17720762e-01
1.50682002e-01 -5.90108223e-02 -1.28170896e+00 8.17405641e-01
3.98221910e-02 4.02304888e-01 8.43594491e-01 1.12410247e-01
-2.10472181e-01 5.25864288e-02 -1.47642303e+00 1.16978325e-01
1.11683047e+00 -3.57174903e-01 -1.00855815e+00 2.59205550e-01
4.09818798e-01 -7.08462536e-01 -1.10628402e+00 5.80433428e-01
2.60805368e-01 -8.16510081e-01 1.03368425e+00 -6.50013506e-01
4.40252393e-01 -3.24267149e-01 -4.81004082e-02 -1.13008523e+00
-6.26290977e-01 -6.06648326e-02 1.36707291e-01 1.21600020e+00
1.03002667e-01 -6.23128593e-01 9.96927559e-01 2.55831122e-01
-5.59965312e-01 -4.27896976e-01 -1.00283408e+00 -8.94027710e-01
6.96281552e-01 3.81135382e-02 2.67652661e-01 1.38023496e+00
-3.86489689e-01 2.82592475e-01 8.88169333e-02 2.72294760e-01
8.35342109e-01 6.52363241e-01 4.44905311e-01 -1.53871512e+00
-1.66733369e-01 -3.86674970e-01 -4.99248981e-01 -1.02389812e+00
1.04782380e-01 -1.44703293e+00 1.01139970e-01 -1.80425918e+00
4.37585890e-01 -5.69811046e-01 -5.17366290e-01 5.00644565e-01
-7.85009265e-02 6.07781205e-03 -2.98812330e-01 4.17935073e-01
-6.07806981e-01 4.41530138e-01 1.62980580e+00 -3.97426546e-01
-2.48980463e-01 -7.76041672e-02 -5.87804496e-01 7.94791400e-01
7.21440077e-01 -4.70412046e-01 -5.88407993e-01 -2.89116234e-01
-4.44979995e-01 -2.44674087e-02 3.35503429e-01 -8.98225248e-01
9.47704241e-02 -1.73674807e-01 5.03850639e-01 -5.83602190e-01
-3.40095997e-01 -7.17752516e-01 -2.19333731e-03 4.96372908e-01
-2.36683577e-01 -6.13502264e-01 8.57976824e-02 8.20788741e-01
-2.80079514e-01 -7.96980187e-02 1.30629325e+00 -3.84496003e-01
-1.17149818e+00 4.29951876e-01 -1.02497302e-01 6.09486580e-01
1.01070154e+00 -2.88200736e-01 1.03555918e-02 2.17321128e-01
-1.30871439e+00 3.17631155e-01 3.50379229e-01 2.16183901e-01
5.58150589e-01 -1.40202653e+00 -7.86518395e-01 3.07254016e-01
4.06691045e-01 7.11089015e-01 7.63969779e-01 8.73306632e-01
-5.26710451e-01 7.63823465e-02 -5.58195829e-01 -1.31418693e+00
-8.51242840e-01 6.45606458e-01 5.51398039e-01 -2.27626368e-01
-1.02420187e+00 7.98153877e-01 8.87739360e-01 -6.82385266e-01
-1.12259604e-01 -3.02330017e-01 -6.66502640e-02 -9.72930808e-03
-7.44940192e-02 9.47456360e-02 1.53842336e-02 -6.24437153e-01
-3.84706348e-01 6.57361507e-01 -2.94449985e-01 2.40588665e-01
1.17294931e+00 -1.65575773e-01 1.45692348e-01 1.29212230e-01
1.17307377e+00 -4.49873269e-01 -1.36935067e+00 -7.58521259e-01
1.67818546e-01 -2.33700305e-01 7.70602077e-02 -8.94913018e-01
-1.18646634e+00 7.76547909e-01 1.06255937e+00 -1.99207917e-01
1.22410631e+00 3.91889662e-01 8.40964735e-01 -6.82857931e-02
3.01159233e-01 -1.34291792e+00 9.52390134e-02 2.51275659e-01
4.95184332e-01 -1.51612580e+00 -1.83070257e-01 -5.07820070e-01
-8.92708600e-01 9.05888319e-01 7.47875631e-01 -2.10723549e-01
7.64855802e-01 7.67407343e-02 2.03107581e-01 -2.79850215e-01
3.06157060e-02 -4.90374535e-01 2.57179737e-01 1.08908534e+00
2.83803076e-01 1.00774780e-01 -4.00016069e-01 8.11107397e-01
3.74382615e-01 1.61218256e-01 -3.89325470e-02 1.01605117e+00
-3.36808115e-01 -1.20715904e+00 -3.12643647e-01 2.56177694e-01
-1.59864277e-01 2.81776100e-01 -1.54895172e-01 9.48400021e-01
4.08008039e-01 3.05164367e-01 9.49025992e-03 1.27260208e-01
3.44661027e-01 2.47166187e-01 5.48780084e-01 -6.68881416e-01
-1.59905106e-02 5.39446175e-01 -4.83839601e-01 -2.61789858e-01
-6.17722094e-01 -8.21326137e-01 -1.69511688e+00 3.17253619e-01
2.06724361e-01 1.72578488e-02 2.25486353e-01 9.65571523e-01
4.10533845e-01 6.63107753e-01 4.91485149e-01 -4.64015603e-01
-2.06293121e-01 -8.21331859e-01 -7.94244111e-01 8.23472083e-01
2.36318648e-01 -8.66407573e-01 5.08174263e-02 4.95097488e-01] | [14.576336860656738, -1.9771854877471924] |
ebe5c9a4-424a-48b2-866b-8bb98536bfe4 | mediar-harmony-of-data-centric-and-model | 2212.03465 | null | https://arxiv.org/abs/2212.03465v1 | https://arxiv.org/pdf/2212.03465v1.pdf | MEDIAR: Harmony of Data-Centric and Model-Centric for Multi-Modality Microscopy | Cell segmentation is a fundamental task for computational biology analysis. Identifying the cell instances is often the first step in various downstream biomedical studies. However, many cell segmentation algorithms, including the recently emerging deep learning-based methods, still show limited generality under the multi-modality environment. Weakly Supervised Cell Segmentation in Multi-modality High-Resolution Microscopy Images was hosted at NeurIPS 2022 to tackle this problem. We propose MEDIAR, a holistic pipeline for cell instance segmentation under multi-modality in this challenge. MEDIAR harmonizes data-centric and model-centric approaches as the learning and inference strategies, achieving a 0.9067 F1-score at the validation phase while satisfying the time budget. To facilitate subsequent research, we provide the source code and trained model as open-source: https://github.com/Lee-Gihun/MEDIAR | ['Se-Young Yun', 'Joonkee Kim', 'Sangmook Kim', 'Gihun Lee'] | 2022-12-07 | null | null | null | null | ['electron-microscopy-image-segmentation'] | ['computer-vision'] | [ 9.44042951e-02 -2.96205610e-01 -1.61353230e-01 -2.05898672e-01
-1.25981247e+00 -5.82580745e-01 3.47802013e-01 4.34853464e-01
-7.10398376e-01 1.08480942e+00 -3.19615692e-01 -3.35451066e-01
7.18174577e-02 -5.41627347e-01 -6.33711040e-01 -1.07814538e+00
4.55907315e-01 6.71019197e-01 3.68098229e-01 1.99662864e-01
4.58505780e-01 6.43059731e-01 -1.06857145e+00 4.46620762e-01
6.03239417e-01 8.39211941e-01 3.97764832e-01 1.01044261e+00
-4.29892261e-03 2.76694417e-01 -8.81318003e-02 -1.48449227e-01
3.37739959e-02 -2.81767905e-01 -9.48788345e-01 -2.78169304e-01
7.07772896e-02 -5.97163960e-02 -4.86149192e-02 8.62986028e-01
8.90547156e-01 -2.54568160e-01 6.60949171e-01 -1.05030835e+00
-1.23626836e-01 1.57189891e-01 -6.21947765e-01 5.66857457e-01
-2.21637506e-02 4.19175565e-01 6.70681357e-01 -7.15053856e-01
1.06941664e+00 6.86504066e-01 3.62144172e-01 8.69803250e-01
-1.34688878e+00 -5.77104628e-01 -1.31334141e-01 -1.06022522e-01
-1.38757408e+00 -6.56341910e-01 3.51459146e-01 -5.89705527e-01
8.37385535e-01 1.79236755e-01 6.36721492e-01 1.05586660e+00
2.71856815e-01 9.04582918e-01 1.20230210e+00 8.67960975e-02
2.75759399e-01 -3.20429951e-01 1.84348509e-01 6.77895963e-01
2.26474211e-01 -3.61627460e-01 -4.19489235e-01 -1.24679804e-01
1.00008953e+00 2.24829003e-01 -2.17672095e-01 2.23908067e-01
-1.49780619e+00 4.65974957e-01 -3.12427245e-02 5.44245064e-01
-1.99886948e-01 1.45310462e-01 4.81207520e-01 -9.87015292e-02
4.96891826e-01 1.89238563e-01 -7.75638521e-01 -9.08016488e-02
-1.05461466e+00 3.68646085e-01 4.33750719e-01 7.28284657e-01
4.71828759e-01 -4.61272329e-01 -2.14088693e-01 6.33179545e-01
2.38203719e-01 -8.05391930e-03 4.56239432e-01 -1.02253246e+00
-2.27716535e-01 8.30523551e-01 -1.57969758e-01 -3.61051619e-01
-7.00811863e-01 -3.40781450e-01 -9.49060678e-01 -8.86075199e-02
9.69006121e-01 -2.33857617e-01 -9.04756010e-01 1.43956923e+00
6.96157753e-01 2.82122612e-01 -1.32829770e-01 9.45985973e-01
1.02463543e+00 3.90477210e-01 8.88255686e-02 -2.75534898e-01
1.55278683e+00 -6.45928144e-01 -6.23574257e-01 1.64159596e-01
9.20515180e-01 -6.64690137e-01 8.26113403e-01 3.96667212e-01
-9.73127544e-01 -4.80589978e-02 -6.16346955e-01 -4.96693134e-01
-6.72135234e-01 1.80562913e-01 6.48437440e-01 1.42927691e-02
-9.15851057e-01 3.30564231e-01 -9.74128306e-01 -6.20731533e-01
9.60905254e-01 4.06304449e-01 -5.75422049e-01 -8.99024084e-02
-6.06780469e-01 5.83756566e-01 2.18685448e-01 6.15955852e-02
-1.03063798e+00 -1.16753864e+00 -4.30328637e-01 -3.22703481e-01
6.09449707e-02 -7.23060012e-01 1.00409484e+00 -1.28896266e-01
-1.30241466e+00 1.66073132e+00 -5.62826276e-01 -2.44387552e-01
6.26746655e-01 1.63936689e-02 6.73546419e-02 3.87782156e-01
1.06533319e-01 8.77187669e-01 -7.44297728e-02 -1.14185011e+00
-6.74037457e-01 -6.03365600e-01 -3.12076360e-01 -1.64243087e-01
7.68652633e-02 1.89375074e-03 -6.97150111e-01 -3.45952123e-01
-1.40176117e-02 -7.35392392e-01 -3.51442844e-01 1.50795311e-01
-5.55760324e-01 -7.41946474e-02 6.50921106e-01 -5.09845495e-01
7.49088585e-01 -1.87298810e+00 2.75142968e-01 -2.04021111e-01
3.69125664e-01 1.56677485e-01 7.50474706e-02 1.55429170e-01
1.53514087e-01 2.94214129e-01 -3.16319525e-01 -5.05416870e-01
-3.45193595e-01 -3.34414601e-01 1.77517459e-01 7.87792563e-01
7.04214945e-02 1.07336414e+00 -7.79119074e-01 -7.03506172e-01
5.53272702e-02 5.98114252e-01 -1.95901439e-01 1.13604963e-01
-2.68815488e-01 1.14165235e+00 -4.28600520e-01 1.19485521e+00
5.28461456e-01 -6.21369004e-01 1.88963071e-01 -3.40493351e-01
-1.87027112e-01 -3.26533675e-01 -5.64513147e-01 1.78334606e+00
1.70925051e-01 5.05276084e-01 4.78588074e-01 -1.13916945e+00
5.96031427e-01 4.07572925e-01 9.05973613e-01 -4.70075369e-01
4.34118629e-01 5.06090462e-01 -9.95758995e-02 -3.97801012e-01
-1.16132079e-02 -2.58739114e-01 1.47378489e-01 3.87519822e-02
7.99153149e-02 -4.85790707e-02 4.43279743e-01 1.75066561e-01
1.19396269e+00 2.41828442e-01 1.31128326e-01 -2.80482471e-01
3.79195690e-01 2.94036925e-01 8.31415832e-01 3.65488201e-01
-6.63597047e-01 8.58951032e-01 6.28966153e-01 -2.27526337e-01
-1.07008100e+00 -7.72791326e-01 -5.39690018e-01 7.73421049e-01
3.79753821e-02 -3.83907417e-03 -7.72123635e-01 -4.23054785e-01
7.14016380e-03 -9.20640863e-03 -5.49236953e-01 3.10549974e-01
-3.14398348e-01 -1.16083038e+00 9.32921350e-01 1.42968997e-01
2.90058941e-01 -7.97933161e-01 -4.47891206e-01 1.51621237e-01
-2.27493435e-01 -1.28876781e+00 -1.34183556e-01 2.99426436e-01
-7.37580061e-01 -1.43566477e+00 -1.14612126e+00 -8.47733974e-01
8.00031960e-01 -6.32133633e-02 6.56860530e-01 2.36514226e-01
-8.27382803e-01 3.69252861e-02 -8.27215239e-02 -2.78987080e-01
-1.65220164e-02 1.47166625e-01 -1.75975263e-01 -2.55445778e-01
4.67199445e-01 -3.86200041e-01 -8.95641088e-01 4.80155610e-02
-8.99837196e-01 3.36593837e-01 4.50570881e-01 8.10612381e-01
1.47487700e+00 -3.14940840e-01 8.16193938e-01 -1.18141878e+00
2.36839816e-01 -4.82916325e-01 -7.02442586e-01 5.62530793e-02
-2.23951325e-01 -5.72396874e-01 5.68086326e-01 -2.58219928e-01
-6.53279960e-01 1.91044480e-01 -5.35414517e-01 -4.96655330e-02
-5.86967409e-01 5.12977839e-01 -1.22684479e-01 -1.46041423e-01
1.89388320e-01 3.17395002e-01 1.93158314e-01 -2.66852856e-01
-2.27561906e-01 5.02914548e-01 3.76410395e-01 -5.70628345e-01
1.57936916e-01 8.32121611e-01 2.82686949e-01 -7.30859816e-01
-9.55692053e-01 -7.81887114e-01 -6.07661843e-01 -3.34315389e-01
1.11614132e+00 -8.91749680e-01 -1.07470703e+00 8.50928843e-01
-9.08874333e-01 -8.40773165e-01 2.89652109e-01 2.92185664e-01
-6.88690901e-01 1.72711164e-01 -1.02386034e+00 -4.85397428e-01
-5.34104824e-01 -1.22263539e+00 1.34406543e+00 5.39460957e-01
-1.41345128e-01 -1.01333725e+00 1.68553218e-01 8.05546105e-01
1.80694580e-01 6.50100470e-01 1.00538766e+00 -9.01862741e-01
-6.81282580e-01 -1.52301818e-01 -2.74925709e-01 -4.22433197e-01
9.15842727e-02 4.25450951e-01 -1.03749835e+00 -2.57304579e-01
-6.35383666e-01 -4.09296870e-01 1.00341964e+00 9.16870952e-01
1.36095834e+00 2.77981222e-01 -7.48071551e-01 8.14965248e-01
1.48628497e+00 8.86299238e-02 5.85015953e-01 3.94248188e-01
7.24598408e-01 5.39249361e-01 5.83961785e-01 4.47485834e-01
2.59374291e-01 3.89800936e-01 3.03369671e-01 -4.15638477e-01
6.07559234e-02 2.09672228e-01 -2.47696117e-01 4.53942537e-01
6.42281771e-02 -3.81597847e-01 -1.29295313e+00 7.65579224e-01
-2.00911283e+00 -8.00474942e-01 -4.50414091e-01 1.86167169e+00
1.06202281e+00 1.27862334e-01 1.47797674e-01 -4.35146764e-02
7.55771399e-01 -2.84042358e-01 -8.72836351e-01 1.53838217e-01
-3.33186388e-01 -3.06092091e-02 2.84000129e-01 4.26971555e-01
-9.89908993e-01 1.12409425e+00 5.99076653e+00 1.06187451e+00
-1.14119470e+00 5.97532503e-02 1.31365550e+00 -2.54498482e-01
-7.98172578e-02 -9.93691236e-02 -1.06653798e+00 6.56613410e-01
7.70564139e-01 1.35078160e-02 1.16587460e-01 1.91688061e-01
5.80026627e-01 -3.77663523e-01 -1.05886829e+00 8.90245140e-01
-5.15571892e-01 -1.87051618e+00 -1.20407969e-01 3.44372183e-01
5.10864019e-01 3.22065473e-01 4.22401056e-02 5.98631762e-02
-7.52340406e-02 -1.18584037e+00 3.75017971e-02 7.35586166e-01
1.01569355e+00 -5.28642058e-01 8.32558095e-01 4.73972291e-01
-8.79517317e-01 3.65959197e-01 -2.54995435e-01 2.93984354e-01
3.43560338e-01 1.00726402e+00 -5.49512088e-01 2.82266110e-01
6.03030801e-01 7.75864005e-01 -3.36445540e-01 1.18140233e+00
5.04378915e-01 5.92209995e-01 -3.42852592e-01 2.76814699e-01
-1.26821801e-01 -2.12029487e-01 4.43792790e-01 1.32713616e+00
2.13859260e-01 2.59419709e-01 1.31069779e-01 8.69338810e-01
-1.84849381e-01 -2.43916698e-02 -3.81439298e-01 -3.00620824e-01
4.17669952e-01 1.86521530e+00 -1.33789790e+00 -2.25090817e-01
-1.74894005e-01 6.25975549e-01 3.31428170e-01 3.71429086e-01
-7.37572253e-01 -2.24212244e-01 6.54112041e-01 2.76899338e-01
-3.19717154e-02 -9.31039900e-02 -4.35492367e-01 -9.79790688e-01
-4.34479028e-01 -6.40218914e-01 5.05699933e-01 -5.09032369e-01
-1.20685899e+00 1.04814120e-01 -3.40085477e-01 -8.36064637e-01
3.80703270e-01 -9.29012954e-01 -4.94389683e-01 6.97881997e-01
-1.59783506e+00 -1.31429577e+00 -2.59257108e-01 2.50288159e-01
4.43374127e-01 -4.83483560e-02 8.28494906e-01 4.74886447e-01
-1.04969716e+00 3.44150960e-01 2.57961929e-01 1.50157243e-01
7.36202955e-01 -1.16125369e+00 -1.27267852e-01 5.09846866e-01
-4.32691902e-01 7.39401460e-01 6.56333208e-01 -6.32314205e-01
-1.37079453e+00 -1.08437765e+00 5.99331319e-01 -4.06209528e-01
5.95521629e-01 -2.49753714e-01 -9.76309776e-01 6.12792909e-01
-7.48039922e-03 3.54916275e-01 1.23675394e+00 -1.41914800e-01
1.92310229e-01 7.26579353e-02 -1.36271453e+00 6.26875758e-01
7.48128593e-01 -3.70052218e-01 2.49620795e-01 4.84921575e-01
4.48677480e-01 -7.68624663e-01 -1.43431437e+00 5.36433578e-01
6.06746972e-01 -7.97637582e-01 6.63582206e-01 -4.64654654e-01
5.92478812e-01 -5.25852859e-01 4.22237813e-02 -6.46659017e-01
-1.35586530e-01 -3.89245570e-01 -3.18571121e-01 1.27395380e+00
5.94651520e-01 -5.46405196e-01 1.06954396e+00 3.72281879e-01
-2.39309639e-01 -1.45836890e+00 -1.11505210e+00 -3.37738156e-01
4.34973180e-01 1.91551615e-02 1.47739246e-01 8.10185730e-01
6.47441223e-02 1.94257066e-01 1.53610826e-01 -8.49154368e-02
7.31473505e-01 1.61408752e-01 7.05167174e-01 -1.09970927e+00
-7.82525763e-02 -6.47392929e-01 -2.32746512e-01 -7.49823391e-01
2.27362037e-01 -9.44348395e-01 -7.81886205e-02 -1.79938686e+00
6.78042650e-01 -1.98680669e-01 -5.28751671e-01 4.92605746e-01
-1.78392410e-01 3.73296082e-01 -2.69047201e-01 3.28731030e-01
-8.89539421e-01 1.32743657e-01 1.53075433e+00 -3.36220376e-02
1.55442774e-01 -1.83079198e-01 -7.61802673e-01 5.07672071e-01
1.03679979e+00 -2.83122063e-01 2.53499467e-02 -3.16196501e-01
-3.24966758e-02 5.67000099e-02 4.35666710e-01 -9.58446383e-01
5.19242942e-01 -3.13871950e-01 6.57507956e-01 -6.79822624e-01
3.56753945e-01 -4.76698756e-01 3.22296768e-01 5.07681847e-01
-5.39615393e-01 -3.07213873e-01 3.17730308e-01 3.31385404e-01
-1.29030496e-01 1.55362755e-01 1.07155013e+00 -3.05038720e-01
-2.93567777e-01 7.45243013e-01 -5.37719131e-01 1.99828967e-01
1.12272716e+00 -3.51463526e-01 -8.94926548e-01 1.65414199e-01
-7.47055769e-01 5.71936190e-01 7.46494055e-01 -1.87300503e-01
4.98755127e-01 -8.57336521e-01 -6.76178515e-01 -2.02682093e-01
1.67071968e-01 4.84613240e-01 5.18317044e-01 1.48751390e+00
-7.02258408e-01 4.63355154e-01 -2.82501400e-01 -6.91412568e-01
-1.24784625e+00 1.01964846e-01 5.07228494e-01 -3.81958187e-01
-3.33031356e-01 9.53911960e-01 2.86857426e-01 -5.79626203e-01
-1.74774453e-01 3.33571695e-02 -2.99526811e-01 -6.87925443e-02
3.73053730e-01 3.02309692e-01 -6.75617009e-02 -7.21112311e-01
-5.15314877e-01 5.60265541e-01 -3.98242146e-01 1.13808878e-01
1.32912707e+00 -1.22560136e-01 -4.08285886e-01 6.56987906e-01
1.23696625e+00 -3.01459074e-01 -1.29928482e+00 9.85698700e-02
-2.28121839e-02 2.19725966e-02 1.23327874e-01 -7.85277367e-01
-1.11649382e+00 8.20610762e-01 4.11843896e-01 -1.12517215e-01
9.60868478e-01 5.81572689e-02 9.64255333e-01 1.65814728e-01
2.86256611e-01 -1.06401396e+00 -3.16264421e-01 5.37995398e-01
2.85412312e-01 -1.42865920e+00 4.11618464e-02 -4.83501464e-01
-3.63301665e-01 8.82645965e-01 9.33323205e-01 2.12596595e-01
6.45248532e-01 5.92782378e-01 2.09445521e-01 -3.64637792e-01
-1.07096326e+00 -2.53174126e-01 -2.53037661e-01 5.14476180e-01
9.60717618e-01 -1.85574561e-01 -3.96986574e-01 8.68533373e-01
3.52239221e-01 2.73555607e-01 2.27163211e-01 8.92217636e-01
-3.78423363e-01 -9.58669245e-01 9.99055728e-02 6.20647013e-01
-9.84896183e-01 6.32248744e-02 -3.90507519e-01 4.95349854e-01
3.49989757e-02 6.46095514e-01 -5.02642095e-02 3.34871709e-02
-1.19961217e-01 -1.18213847e-01 4.40954059e-01 -5.91022015e-01
-4.30533022e-01 4.80624110e-01 -3.72034311e-01 -5.37078977e-01
-5.09893954e-01 -6.47789776e-01 -1.99825978e+00 -3.20414066e-01
-3.66543569e-02 -1.19699180e-01 4.09449190e-01 1.03661442e+00
6.38994634e-01 5.39513648e-01 8.92104432e-02 -8.39856446e-01
4.07705188e-01 -6.96194172e-01 -5.51119626e-01 1.50401860e-01
2.75347710e-01 -3.36456567e-01 -2.65931457e-01 4.93054837e-01] | [14.549108505249023, -3.1437408924102783] |
b2ec9f12-f13b-4a40-8910-28ee27544fcf | leveraging-machine-learning-for-multichain | 2306.07972 | null | https://arxiv.org/abs/2306.07972v1 | https://arxiv.org/pdf/2306.07972v1.pdf | Leveraging Machine Learning for Multichain DeFi Fraud Detection | Since the inception of permissionless blockchains with Bitcoin in 2008, it became apparent that their most well-suited use case is related to making the financial system and its advantages available to everyone seamlessly without depending on any trusted intermediaries. Smart contracts across chains provide an ecosystem of decentralized finance (DeFi), where users can interact with lending pools, Automated Market Maker (AMM) exchanges, stablecoins, derivatives, etc. with a cumulative locked value which had exceeded 160B USD. While DeFi comes with high rewards, it also carries plenty of risks. Many financial crimes have occurred over the years making the early detection of malicious activity an issue of high priority. The proposed framework introduces an effective method for extracting a set of features from different chains, including the largest one, Ethereum and it is evaluated over an extensive dataset we gathered with the transactions of the most widely used DeFi protocols (23 in total, including Aave, Compound, Curve, Lido, and Yearn) based on a novel dataset in collaboration with Covalent. Different Machine Learning methods were employed, such as XGBoost and a Neural Network for identifying fraud accounts detection interacting with DeFi and we demonstrate that the introduction of novel DeFi-related features, significantly improves the evaluation results, where Accuracy, Precision, Recall, F1-score and F2-score where utilized. | ['Leandros Tassiulas', 'Iason Ofeidis', 'Sandro Scherrers', 'Georgios Palaiokrassas'] | 2023-05-17 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [-5.01691520e-01 -1.53328493e-01 -2.50513703e-01 -3.94058861e-02
-3.42161059e-01 -9.91970778e-01 9.27623272e-01 4.02821004e-01
-5.95544755e-01 1.07395136e+00 4.72326726e-02 -5.25456250e-01
-9.83412340e-02 -9.26649570e-01 -4.29283082e-01 -5.26740849e-01
-3.53303969e-01 7.30915189e-01 2.17235371e-01 -3.03388834e-01
7.32306898e-01 4.85243440e-01 -8.90329063e-01 2.01354727e-01
5.53098142e-01 1.25334942e+00 -6.07138157e-01 1.82318896e-01
-1.05804399e-01 1.21446967e+00 -5.31402171e-01 -1.01834106e+00
7.51482964e-01 1.05541917e-02 -6.75692558e-01 -8.75914991e-01
-4.10007536e-01 -6.34935200e-01 3.55201699e-02 9.89547849e-01
9.69986022e-02 -6.04642510e-01 5.07029355e-01 -1.47645164e+00
-1.97378963e-01 1.04857671e+00 -6.21353686e-01 2.07784742e-01
2.24537641e-01 4.14642572e-01 1.28056943e+00 -5.92030585e-01
9.06531274e-01 6.67494357e-01 8.01718295e-01 -6.94878250e-02
-9.17213082e-01 -9.07676995e-01 -5.89840770e-01 2.59298712e-01
-5.57044148e-01 -1.01343170e-01 6.65109873e-01 -4.05197650e-01
1.00704741e+00 3.24842781e-01 9.65027928e-01 8.96585464e-01
3.65904212e-01 6.96166754e-01 1.53298807e+00 -4.48095322e-01
1.87075987e-01 2.94689029e-01 4.44342732e-01 1.19144991e-01
8.95504713e-01 1.07482292e-01 -5.57425499e-01 -8.35415602e-01
3.59526366e-01 3.06821316e-01 -2.34798327e-01 -2.31969789e-01
-1.38081706e+00 1.05600965e+00 5.29075786e-02 8.30081880e-01
-6.85322702e-01 2.51636147e-01 8.73346150e-01 8.02731097e-01
1.56658248e-03 4.06861037e-01 -6.92059755e-01 -5.90010285e-01
-8.65959764e-01 5.46634793e-01 1.42034483e+00 5.98839283e-01
4.79313582e-01 -3.84280473e-01 1.59837589e-01 -2.19905470e-02
5.88980377e-01 2.42840946e-01 5.75655818e-01 -7.96837568e-01
7.91967630e-01 8.49079192e-01 5.38082659e-01 -9.56756592e-01
-8.60161483e-02 -2.05791205e-01 -7.07219839e-01 4.61555868e-01
8.16889584e-01 -1.31629884e-01 -1.80033997e-01 1.11545205e+00
1.14877127e-01 -5.31527638e-01 -4.84100021e-02 6.67066872e-01
3.83600965e-02 4.35011417e-01 -1.33226871e-01 -4.56678689e-01
1.33491683e+00 -6.56838477e-01 -8.16597044e-01 5.83545506e-01
3.11675102e-01 -8.20344746e-01 3.49616498e-01 9.96080935e-01
-1.02107406e+00 4.05914485e-01 -9.92716730e-01 3.16063106e-01
-5.23934662e-01 -6.57917500e-01 8.82445216e-01 8.34397674e-01
-5.67964196e-01 1.09086156e+00 -6.88259184e-01 -1.05990887e-01
6.61134362e-01 4.23475355e-01 -4.92162079e-01 5.42631686e-01
-1.26099956e+00 9.79228854e-01 4.87938374e-01 -5.50259240e-02
-7.99880922e-01 -3.38032216e-01 1.17973737e-01 1.51664138e-01
1.26265466e-01 -3.32582116e-01 9.93667483e-01 -9.62751746e-01
-1.31170118e+00 7.76649177e-01 8.53948116e-01 -9.83188868e-01
1.43013144e+00 -3.21743309e-01 -2.95748383e-01 8.44951496e-02
-1.16017185e-01 -2.27766842e-01 4.60912645e-01 -6.72727704e-01
-7.34910965e-01 -4.76996064e-01 7.95156136e-02 -2.99008071e-01
-2.47790828e-01 5.38869262e-01 5.05598605e-01 -7.60222495e-01
-3.27440873e-02 -7.05260038e-01 3.19964141e-01 -4.67930466e-01
-3.21421951e-01 -3.49256456e-01 8.89719903e-01 -9.77815270e-01
1.04441953e+00 -1.62135279e+00 -4.29548740e-01 6.90127790e-01
2.50237107e-01 3.59678954e-01 8.20519090e-01 1.09946358e+00
6.46279976e-02 4.10395443e-01 -1.95879146e-01 1.48793012e-01
2.72083044e-01 -1.38312712e-01 -2.45573133e-01 6.42818153e-01
-3.91511232e-01 7.46105850e-01 -8.25220227e-01 -3.06398809e-01
-8.65628719e-02 2.76783463e-02 -2.16994032e-01 1.51286080e-01
-1.16821729e-01 4.45782810e-01 -4.83915806e-01 1.00150573e+00
6.83655977e-01 -3.11834335e-01 4.86558884e-01 3.52453417e-03
-2.76027292e-01 1.17126040e-01 -1.16541469e+00 1.21677077e+00
1.53006852e-01 4.15454119e-01 -8.21410120e-02 -8.30930233e-01
8.46290231e-01 7.39023805e-01 7.60320127e-01 -5.00670373e-01
3.70554924e-01 8.04896176e-01 -1.10947356e-01 -4.12970454e-01
2.34149903e-01 -8.28397274e-02 1.02244057e-01 1.19834208e+00
-1.64886922e-01 5.17404497e-01 1.98025361e-01 1.43798813e-01
1.35713398e+00 1.24440752e-01 6.51849210e-01 -2.93414384e-01
5.01553237e-01 2.92942766e-03 5.92175066e-01 5.57600498e-01
-6.90325499e-01 3.75728346e-02 8.97089660e-01 -9.30360436e-01
-1.17431152e+00 -7.04811633e-01 -4.76999059e-02 4.69347686e-01
-1.03438474e-01 -9.27786082e-02 -5.35044193e-01 -1.02948964e+00
5.40424287e-01 4.15292650e-01 -1.69921830e-01 4.65862960e-01
-7.54879713e-01 -7.09578216e-01 7.10133791e-01 -6.84120953e-02
9.37768638e-01 -1.30567074e+00 -1.07346678e+00 4.80403483e-01
-1.01285525e-01 -7.01062679e-01 2.18514018e-02 2.50098646e-01
-1.01834452e+00 -1.57176220e+00 -6.04624748e-01 -1.32462859e-01
2.21641034e-01 -3.16137403e-01 1.06219935e+00 2.70412326e-01
-1.66104566e-02 -3.99762660e-01 -4.56207722e-01 -3.89523894e-01
-6.30136669e-01 4.10474837e-02 -4.38118875e-02 1.04397051e-01
5.78093827e-01 -8.29719424e-01 -8.40547085e-01 2.42590040e-01
-6.88714266e-01 -2.66744256e-01 5.22408485e-01 6.95709705e-01
-1.90489516e-01 -4.07620519e-01 8.99563789e-01 -1.22971177e+00
6.82903826e-01 -6.01245046e-01 -9.05881464e-01 2.03562424e-01
-1.27056384e+00 -6.36535883e-02 5.50431728e-01 7.91081116e-02
-7.45127976e-01 -6.72069073e-01 3.42953473e-01 -1.48376659e-01
2.15537071e-01 3.84805351e-01 9.45668817e-02 -1.58018500e-01
4.32291389e-01 -1.63732842e-01 5.73773459e-02 -6.64344549e-01
1.85208783e-01 9.03609335e-01 1.23709299e-01 -3.56837004e-01
7.27348685e-01 4.31353480e-01 -1.24052078e-01 1.12388708e-01
1.56775981e-01 -3.34711790e-01 -2.27905482e-01 -4.29020748e-02
4.09290105e-01 -5.39804697e-01 -1.39777493e+00 1.05908549e+00
-1.30368328e+00 3.22109401e-01 -1.18119933e-01 6.66470110e-01
-2.46092707e-01 5.49549222e-01 -9.59137380e-01 -9.05444086e-01
-8.60644758e-01 -7.81208277e-01 -1.30973190e-01 1.56853065e-01
-4.11174595e-01 -6.67752028e-01 3.65020543e-01 6.84174895e-01
7.83384264e-01 8.21814060e-01 7.62309194e-01 -1.30149436e+00
-9.23293829e-01 -3.58399481e-01 -2.00113773e-01 5.76717794e-01
1.48684770e-01 4.20516655e-02 -7.99914658e-01 -4.63235110e-01
2.50123352e-01 -2.70949394e-01 4.29120094e-01 -2.65228391e-01
3.13092679e-01 -8.46682131e-01 -1.11637376e-01 2.49713272e-01
1.62767494e+00 4.40748692e-01 5.75014770e-01 9.42704022e-01
1.41208231e-01 4.09100324e-01 3.96574527e-01 7.73260772e-01
1.90133378e-01 4.60776120e-01 6.39244378e-01 5.39509654e-01
6.32808566e-01 -5.51257469e-02 2.86076605e-01 7.09211826e-01
-8.46187592e-01 2.37980306e-01 -9.78297591e-01 3.95844489e-01
-1.94579995e+00 -1.11606586e+00 -2.49920413e-01 2.27543640e+00
8.13159287e-01 3.78627270e-01 4.96020317e-01 5.62459826e-01
6.74241006e-01 -2.21100077e-01 -3.03231776e-01 -5.43646932e-01
-3.15546393e-02 1.14733882e-01 1.06577671e+00 5.21205775e-02
-9.14085925e-01 3.94301653e-01 5.37844992e+00 3.85300934e-01
-9.47629511e-01 4.48101789e-01 7.81913817e-01 3.82225364e-02
-3.01106006e-01 5.52606046e-01 -2.66566753e-01 1.05384994e+00
9.27418232e-01 -3.31232131e-01 6.64369047e-01 7.25992024e-01
2.01342657e-01 -1.35974333e-01 -9.28816199e-01 7.80321538e-01
-4.41509753e-01 -1.50486684e+00 -4.21156496e-01 3.53431970e-01
7.60787725e-01 3.08572084e-01 -4.84638512e-01 -1.07781522e-01
5.66971779e-01 -7.63601482e-01 9.60894048e-01 7.81727552e-01
2.26154760e-01 -7.92153358e-01 1.24233770e+00 2.66941428e-01
-4.39509749e-01 -3.34561676e-01 2.72705913e-01 -1.54614881e-01
1.44563764e-01 6.45039320e-01 -7.23717868e-01 8.51216853e-01
8.12460065e-01 4.39085037e-01 -1.20433301e-01 9.86731231e-01
-2.95410991e-01 6.46881700e-01 -4.32457834e-01 -2.31828600e-01
2.61442125e-01 -5.95968127e-01 5.58083236e-01 8.68130386e-01
2.79748559e-01 -1.55541152e-01 -5.42831063e-01 8.62792611e-01
-3.12675685e-01 3.08341056e-01 -5.66846132e-01 -7.30823427e-02
5.97370028e-01 1.29070771e+00 -7.20544636e-01 -2.87015140e-01
-4.58367944e-01 4.50797051e-01 -1.04481310e-01 1.48326293e-01
-5.97978890e-01 -5.49899340e-01 1.29895672e-01 3.04637641e-01
2.83712238e-01 2.17273366e-02 -5.11436284e-01 -1.24316847e+00
3.34147513e-01 -1.10502243e+00 4.00514096e-01 -1.86234832e-01
-1.34673047e+00 1.60981685e-01 -2.21114680e-01 -1.22564960e+00
-2.43610188e-01 -3.48705411e-01 -6.35998547e-01 8.78357351e-01
-1.51380622e+00 -1.13384724e+00 1.92433279e-02 5.70952177e-01
-3.16584021e-01 -6.44908667e-01 4.85703379e-01 4.54948753e-01
-1.49139881e-01 2.92686909e-01 5.02833724e-01 3.83539438e-01
6.33490384e-01 -1.31154728e+00 2.11715728e-01 6.08846784e-01
-6.12800717e-02 8.97159517e-01 6.69477582e-01 -7.60081232e-01
-1.45909965e+00 -3.37252587e-01 1.22879291e+00 -4.41913098e-01
9.76781309e-01 -1.75120294e-01 -5.82566381e-01 7.52952516e-01
4.57522929e-01 -6.89268038e-02 2.67504364e-01 -2.40870655e-01
-2.76682943e-01 -5.39926112e-01 -1.59769440e+00 -1.53680399e-01
5.32910526e-01 -3.77208531e-01 -7.06716180e-01 2.42108345e-01
6.58269972e-02 1.64228559e-01 -9.62787449e-01 6.77926689e-02
1.00246644e+00 -1.64190114e+00 4.82535481e-01 -5.55077732e-01
1.84808120e-01 -1.15829155e-01 1.37373786e-02 -6.15593791e-01
2.39732757e-01 -1.31230521e+00 -4.96314824e-01 1.41799712e+00
2.42397502e-01 -1.28077269e+00 9.54594433e-01 7.02968597e-01
6.39384091e-01 -5.72626650e-01 -1.38987339e+00 -6.22477353e-01
7.85821751e-02 3.35521847e-02 9.37789679e-01 1.35001063e+00
3.52089524e-01 -3.70149016e-01 -3.73633355e-01 -5.60118735e-01
8.88345957e-01 3.84762466e-01 8.20600152e-01 -1.42129862e+00
-3.52738261e-01 -4.36475962e-01 -5.05927920e-01 -2.13151425e-01
-5.67688584e-01 -7.87840664e-01 -9.14991975e-01 -1.02457547e+00
1.78300485e-01 -7.44065404e-01 -7.55202174e-01 6.05188191e-01
3.86286408e-01 -6.15252890e-02 6.05515778e-01 1.04044914e+00
-1.30296201e-01 3.26073281e-02 8.36312950e-01 -2.11313814e-01
1.76751316e-01 2.51152158e-01 -5.18264234e-01 5.99979222e-01
5.83474815e-01 -7.92149186e-01 3.00910860e-01 7.83375502e-02
8.23563099e-01 1.82040066e-01 2.26524711e-01 -6.79138064e-01
2.65783906e-01 -1.08667240e-01 7.26619037e-03 -5.21681428e-01
-4.98241723e-01 -1.11933434e+00 6.69962943e-01 9.48820829e-01
1.66786537e-01 3.30470949e-01 -7.15974987e-01 3.72174412e-01
-3.71063471e-01 -5.58938444e-01 4.40347612e-01 -4.82561290e-01
-2.25716876e-03 -7.62626305e-02 -3.73322256e-02 -2.63971329e-01
1.36167455e+00 -2.23486558e-01 -5.40539622e-01 -2.19331831e-01
-3.65584105e-01 1.46634385e-01 7.05592096e-01 9.64427963e-02
-2.86499131e-02 -1.04223585e+00 -7.73815930e-01 -1.42952045e-02
-2.76868433e-01 -3.89533132e-01 -4.47172046e-01 1.13025296e+00
-1.18393290e+00 4.80247557e-01 -6.25439405e-01 -1.85951233e-01
-9.02935326e-01 2.40511581e-01 1.77993879e-01 -1.01492500e+00
-4.97767121e-01 2.47612432e-01 -7.24833310e-01 -8.60065073e-02
8.01570341e-02 -4.91472095e-01 -2.90241808e-01 4.64191020e-01
2.62477577e-01 8.53369176e-01 2.10125491e-01 -3.01628411e-01
-3.85617584e-01 8.25489014e-02 -1.69953704e-01 -2.91067481e-01
1.67883790e+00 2.76304394e-01 -6.83478475e-01 3.80057424e-01
7.28413165e-01 2.46024787e-01 -8.36782753e-01 8.37377086e-02
7.17111230e-01 -7.44654596e-01 -3.66892010e-01 -1.15352428e+00
-1.14282715e+00 5.40207505e-01 4.76343542e-01 7.82311499e-01
5.39943695e-01 -5.08665204e-01 9.83893514e-01 2.95925319e-01
1.05593610e+00 -1.13005257e+00 -3.89401019e-01 2.25867435e-01
5.97128510e-01 -1.18880463e+00 1.28063336e-01 3.05709630e-01
-3.66552353e-01 1.32334006e+00 -1.59789607e-01 -1.85132205e-01
6.11242473e-01 1.89381465e-02 2.12273598e-02 -2.22099692e-01
-4.85853553e-01 8.30729723e-01 -6.64115846e-01 2.52513647e-01
2.97199488e-01 2.25807741e-01 -1.17077768e+00 9.06705558e-01
-1.72138214e-04 6.28524184e-01 7.16238916e-01 1.19018340e+00
-1.90859288e-01 -1.62918401e+00 -4.11851197e-01 5.74419618e-01
-1.19832122e+00 -1.46158859e-02 -2.39550084e-01 9.77399886e-01
2.88457096e-01 6.46897554e-01 -3.73281032e-01 -3.43623525e-03
1.24871813e-01 1.80623382e-01 -6.18922338e-02 8.82203206e-02
-1.51550114e+00 -4.15747762e-01 4.15294677e-01 -5.13581812e-01
-4.58569914e-01 -9.34593797e-01 -1.05262482e+00 -9.33508337e-01
-5.51970124e-01 5.41750908e-01 8.95280540e-01 7.30275214e-01
8.82535800e-02 -3.68250608e-01 9.94765878e-01 -2.46535435e-01
-1.05121362e+00 -1.17830563e+00 -8.71457875e-01 6.88610733e-01
2.45998219e-01 -3.65173250e-01 -5.18742621e-01 2.49529500e-02] | [6.393923759460449, 6.40379524230957] |
69bac929-5704-4d07-8531-4e84f92c2c7a | from-disfluency-detection-to-intent-detection | 2209.08359 | null | https://arxiv.org/abs/2209.08359v1 | https://arxiv.org/pdf/2209.08359v1.pdf | From Disfluency Detection to Intent Detection and Slot Filling | We present the first empirical study investigating the influence of disfluency detection on downstream tasks of intent detection and slot filling. We perform this study for Vietnamese -- a low-resource language that has no previous study as well as no public dataset available for disfluency detection. First, we extend the fluent Vietnamese intent detection and slot filling dataset PhoATIS by manually adding contextual disfluencies and annotating them. Then, we conduct experiments using strong baselines for disfluency detection and joint intent detection and slot filling, which are based on pre-trained language models. We find that: (i) disfluencies produce negative effects on the performances of the downstream intent detection and slot filling tasks, and (ii) in the disfluency context, the pre-trained multilingual language model XLM-R helps produce better intent detection and slot filling performances than the pre-trained monolingual language model PhoBERT, and this is opposite to what generally found in the fluency context. | ['Dat Quoc Nguyen', 'Thinh Hung Truong', 'Mai Hoang Dao'] | 2022-09-17 | null | null | null | null | ['xlm-r', 'slot-filling'] | ['natural-language-processing', 'natural-language-processing'] | [-1.57009766e-01 5.06864116e-02 -3.45817477e-01 -1.28946543e-01
-6.80339754e-01 -7.30992019e-01 6.20793939e-01 1.19135313e-01
-8.33828509e-01 7.69628704e-01 9.24685180e-01 -8.75098884e-01
5.45611501e-01 -4.43492651e-01 -4.27577972e-01 -8.22887421e-02
-2.03216374e-02 5.18126965e-01 1.90033403e-03 -4.42389250e-01
7.38677010e-02 -2.45593756e-01 -9.38081205e-01 2.16684341e-01
1.23199916e+00 1.46628082e-01 7.71683097e-01 2.47456998e-01
3.22814658e-02 1.06900442e+00 -7.59557366e-01 -1.01121880e-01
2.35312909e-01 -4.75279599e-01 -1.09167075e+00 -2.97786921e-01
-4.83073257e-02 -3.46463591e-01 -3.32254261e-01 6.99864984e-01
1.83826566e-01 7.72775263e-02 4.31152105e-01 -6.56130314e-01
-7.74766684e-01 1.11833560e+00 3.16148289e-02 7.22445190e-01
6.42861009e-01 3.64182442e-01 1.13377905e+00 -9.41474617e-01
1.18598533e+00 1.27479434e+00 4.64451820e-01 4.24002171e-01
-1.20462251e+00 -5.95228851e-01 4.23910260e-01 -3.40081453e-02
-1.51289523e+00 -6.13186538e-01 3.90791059e-01 -6.06227040e-01
1.97014916e+00 5.52850701e-02 3.98161709e-01 1.12073791e+00
1.97473526e-01 7.01248467e-01 1.21773028e+00 -7.91622996e-01
-1.77887678e-01 1.86762542e-01 3.13197464e-01 4.12014306e-01
-3.52093093e-02 3.98721665e-01 -6.20773137e-01 2.34225184e-01
6.37298882e-01 -4.99299288e-01 -2.90584117e-01 6.90182507e-01
-1.54958248e+00 9.03536201e-01 1.46102160e-01 9.48652208e-01
-1.87303960e-01 -3.48880798e-01 7.65101612e-01 7.13861823e-01
8.24783981e-01 7.45750904e-01 -8.52439702e-01 -3.40998560e-01
-9.08322453e-01 1.59857094e-01 7.63337314e-01 7.37178802e-01
5.49796760e-01 1.99611615e-02 -5.22614062e-01 9.26195025e-01
1.95950419e-01 3.81617337e-01 6.65926158e-01 -4.82810676e-01
9.23834562e-01 3.69454116e-01 1.44993797e-01 -3.88847589e-01
-5.31147063e-01 -1.83384314e-01 -4.27918062e-02 -4.20135379e-01
5.65599680e-01 -3.48946065e-01 -1.10085857e+00 2.16153479e+00
-1.31742969e-01 -5.19427717e-01 2.25459248e-01 8.67411196e-01
5.37970603e-01 8.10604572e-01 8.79066467e-01 -6.24086738e-01
1.53236628e+00 -9.64310467e-01 -1.02408946e+00 -8.13835204e-01
1.36400652e+00 -1.03291738e+00 1.64132178e+00 -9.18385480e-03
-8.71625125e-01 -5.77924371e-01 -9.07734513e-01 -4.43554014e-01
-3.92443746e-01 1.41551554e-01 7.92823136e-01 6.45069778e-01
-1.17322063e+00 -4.81777173e-03 -9.25420702e-01 -4.07589406e-01
-3.58041942e-01 -6.41702190e-02 -3.17561477e-01 -2.16790140e-01
-1.82173622e+00 1.30877376e+00 7.57680118e-01 -3.09371084e-01
-6.34329081e-01 -6.63214803e-01 -1.20088553e+00 -2.65759706e-01
3.25527608e-01 -1.47089213e-01 1.28727627e+00 -9.00491774e-01
-1.00527942e+00 1.19822395e+00 -4.52231079e-01 -2.52712816e-01
2.19472453e-01 -4.92740035e-01 -6.17484868e-01 -4.49426532e-01
6.54676318e-01 6.94009066e-01 1.54148921e-01 -7.14469552e-01
-7.31323600e-01 -3.56662899e-01 -1.01443157e-01 7.12839961e-01
3.45491692e-02 4.85623747e-01 -3.16927344e-01 -9.62458193e-01
-4.59553339e-02 -8.41955006e-01 -5.86184375e-02 -1.07712972e+00
-2.69127995e-01 -5.26878834e-01 6.85660839e-01 -1.24266839e+00
1.83140111e+00 -2.19230866e+00 -1.48928419e-01 -1.71601161e-01
-1.99385330e-01 3.25910836e-01 -2.74883714e-02 5.37265003e-01
-6.79689869e-02 4.17229980e-01 5.09123020e-02 -3.77446532e-01
3.58885974e-02 3.74718130e-01 -3.90723407e-01 2.97653913e-01
1.62787050e-01 1.21119392e+00 -7.61393547e-01 -3.30768138e-01
2.16798589e-01 1.89121127e-01 -7.68268883e-01 2.92450581e-02
-4.96389806e-01 5.96639812e-01 3.31072479e-01 6.94415212e-01
3.25957775e-01 1.66458562e-01 6.36477172e-01 3.08099955e-01
-6.99840784e-01 1.29387617e+00 -3.24331611e-01 1.84139717e+00
-7.95983911e-01 5.90263367e-01 -5.98636791e-02 -6.61662400e-01
4.06956732e-01 6.37677073e-01 -2.47338954e-02 -1.28383029e+00
2.34980676e-02 5.54646492e-01 4.52171028e-01 -2.32423291e-01
7.23196149e-01 -4.22880054e-01 -6.03767216e-01 5.92214525e-01
1.48078635e-01 2.34688539e-02 4.30303574e-01 2.46428028e-01
1.03113270e+00 -1.04242645e-01 4.87303078e-01 -6.11090660e-01
3.49091321e-01 4.68413562e-01 6.99952662e-01 6.99086785e-01
-3.44300449e-01 5.95624298e-02 2.86998063e-01 -1.75575554e-01
-7.71685004e-01 -1.03195965e+00 -2.71341771e-01 1.60789013e+00
-7.05417842e-02 -7.65266955e-01 -6.62817419e-01 -7.13456094e-01
-2.64690787e-01 1.29268897e+00 -3.21149588e-01 7.20594227e-02
-1.02345383e+00 -6.29149914e-01 6.01306617e-01 5.43708026e-01
4.09028769e-01 -1.62636399e+00 -2.78113335e-01 3.62823576e-01
-8.29068720e-01 -1.35901833e+00 -7.78590024e-01 3.72575939e-01
-4.87860888e-01 -6.14474297e-01 -2.89513201e-01 -1.23647141e+00
9.03054252e-02 -1.81530789e-01 1.45691288e+00 4.43837106e-01
1.20584115e-01 -5.99637590e-02 -5.00208676e-01 -9.94129777e-02
-6.01634145e-01 2.72181749e-01 2.25236863e-02 -1.09672320e+00
7.45206118e-01 -1.99095964e-01 -1.70198858e-01 5.22022769e-02
-6.93829954e-01 2.09378302e-02 4.78040367e-01 6.61848009e-01
2.67315447e-01 -1.64056703e-01 4.88518357e-01 -9.57657814e-01
8.21770906e-01 -5.69562078e-01 -3.82685691e-01 1.58707313e-02
-6.92835331e-01 -7.81737715e-02 5.79579890e-01 -2.99149752e-01
-1.35154092e+00 -1.89995483e-01 -5.42228162e-01 2.51760602e-01
-2.60231256e-01 7.92680621e-01 -2.10143536e-01 4.24032837e-01
7.35651791e-01 1.03602216e-01 -3.72775793e-01 -4.81790066e-01
3.31454843e-01 7.04042614e-01 6.64374232e-01 -5.30636191e-01
3.60551059e-01 -8.98070782e-02 -8.66002858e-01 -9.72028553e-01
-7.43518770e-01 -5.60716689e-01 -6.73043668e-01 2.05286101e-01
1.06709754e+00 -1.32117808e+00 -1.55313700e-01 5.03675342e-01
-1.24213219e+00 -9.16624784e-01 4.64670807e-02 5.87764800e-01
-4.51136798e-01 1.63963243e-01 -1.21126938e+00 -5.18115580e-01
-2.88503557e-01 -1.19124746e+00 8.38708639e-01 -2.99968094e-01
-6.48514092e-01 -1.25918567e+00 2.67582685e-01 2.17754856e-01
5.08097172e-01 -4.71365601e-01 1.30271924e+00 -7.00411022e-01
-2.41743505e-01 4.12962824e-01 -1.32773682e-01 2.40399316e-02
1.57285631e-01 -6.56498313e-01 -5.69471896e-01 -6.22102991e-02
3.82049978e-02 -4.44712847e-01 9.36880350e-01 4.36548680e-01
-1.88453626e-02 -4.15565073e-01 -4.21284765e-01 3.93514901e-01
1.23613870e+00 4.42252070e-01 5.88622570e-01 3.83671284e-01
4.58557427e-01 4.46109205e-01 8.04174364e-01 1.01371422e-01
8.27333510e-01 7.34616816e-01 -3.28662306e-01 -1.38954386e-01
-2.51404226e-01 -6.16762936e-01 7.81069875e-01 7.97901213e-01
3.93167198e-01 -5.17912328e-01 -1.06066191e+00 8.85255396e-01
-1.69046652e+00 -4.91119772e-01 7.92994872e-02 2.10143399e+00
1.22100365e+00 9.60775092e-02 1.52033031e-01 4.01245765e-02
4.96874332e-01 3.46491635e-01 1.00305855e-01 -5.19368947e-01
-4.68475036e-02 3.13077301e-01 3.09180170e-01 9.64412093e-01
-1.23225558e+00 2.16036034e+00 6.90594244e+00 5.98668396e-01
-1.14526129e+00 5.77152073e-01 4.49519575e-01 2.10353687e-01
-4.73674238e-01 1.67446882e-01 -7.99961984e-01 3.62527460e-01
1.12357688e+00 7.81445801e-02 5.07597744e-01 5.01924813e-01
1.98168665e-01 -5.26210487e-01 -9.75509644e-01 4.87220675e-01
-1.69616744e-01 -8.21608067e-01 -4.22088593e-01 -1.80380136e-01
5.13348877e-01 2.00065911e-01 -1.79941848e-01 8.37678194e-01
5.02610207e-01 -1.09700072e+00 1.08850443e+00 -9.92808789e-02
1.03874910e+00 -6.28371239e-01 5.40480971e-01 4.26247686e-01
-1.09764576e+00 1.02720961e-01 -1.97142988e-01 -6.36982322e-01
7.51136541e-01 5.76651990e-01 -9.21607375e-01 7.48259798e-02
3.33249778e-01 4.85574096e-01 -4.11376715e-01 3.76162797e-01
-6.11282825e-01 1.15504742e+00 -2.54817337e-01 2.50284761e-01
1.97228789e-01 -3.97610292e-02 6.21223986e-01 1.57007897e+00
-9.25697107e-03 2.48961169e-02 5.94448268e-01 8.63647699e-01
2.48006523e-01 4.02011037e-01 -6.04761600e-01 -5.17632484e-01
4.79987413e-01 7.21565843e-01 -1.07273757e+00 -2.97865272e-01
-7.10787594e-01 1.38351810e+00 6.08025968e-01 3.62581134e-01
-6.12112403e-01 -7.20974132e-02 8.99637997e-01 -5.55683766e-03
8.59907120e-02 -6.96437061e-01 -5.21655619e-01 -1.31544387e+00
-2.03236818e-01 -7.76621997e-01 5.73415279e-01 -2.53822565e-01
-9.60203886e-01 7.02671885e-01 3.92318470e-03 -4.68031347e-01
-7.76161373e-01 -4.12319332e-01 -4.01206285e-01 1.25878787e+00
-1.39413464e+00 -1.01630509e+00 4.62271959e-01 5.81186056e-01
7.33956218e-01 2.18707129e-01 9.35009599e-01 4.13261861e-01
-5.13625443e-01 4.84677285e-01 -6.10740721e-01 3.23537230e-01
7.10038304e-01 -1.10638297e+00 8.80084693e-01 1.26692617e+00
2.58738995e-01 8.93745482e-01 6.59359634e-01 -1.34424829e+00
-6.43225074e-01 -1.00875819e+00 1.91628993e+00 -5.18997729e-01
5.84917545e-01 -6.98726952e-01 -7.77791202e-01 1.20689833e+00
3.66586357e-01 -3.60065311e-01 5.45824111e-01 5.79783022e-01
-2.32522964e-01 8.84157121e-01 -9.38417494e-01 6.02892578e-01
1.61742866e+00 -7.96320021e-01 -8.33939493e-01 5.17902136e-01
1.30234694e+00 -4.67155993e-01 -4.41030592e-01 3.43326032e-01
1.83176756e-01 -8.70035231e-01 6.68964148e-01 -4.52409029e-01
8.79802480e-02 -1.49220545e-02 -1.65963307e-01 -1.46865034e+00
-6.41527116e-01 -3.35800260e-01 3.39394450e-01 1.17429256e+00
7.66005695e-01 -5.79869330e-01 1.07532807e-01 3.69230539e-01
-4.65480208e-01 -1.49144188e-01 -8.03194761e-01 -9.25324678e-01
4.12617177e-01 -6.22307360e-01 1.93143725e-01 9.04435396e-01
4.64758843e-01 6.12394869e-01 -1.51201755e-01 -9.81979743e-02
-8.59748796e-02 -5.67949601e-02 1.74860079e-02 -9.13817108e-01
-1.69006988e-01 -1.48799703e-01 1.37517706e-01 -9.21575189e-01
7.45191216e-01 -1.09854662e+00 4.27477866e-01 -1.36992419e+00
-1.57733843e-01 -7.52719045e-01 1.00778423e-01 8.11953306e-01
-4.42589372e-01 1.21506408e-01 3.22433621e-01 4.46296126e-01
-4.29055005e-01 3.03186893e-01 9.22357380e-01 4.24167216e-01
-8.33372653e-01 -2.66977578e-01 -8.34111392e-01 4.36803043e-01
6.96820378e-01 -3.78929526e-01 -4.14415896e-01 -4.55103248e-01
7.48729929e-02 2.38093421e-01 -8.52663219e-02 -6.34618521e-01
-3.28912765e-01 -1.59894750e-01 1.55541331e-01 -2.09835768e-01
-1.17838994e-01 -2.14988589e-01 -2.08006948e-01 5.93520284e-01
-1.01417601e-01 4.49092716e-01 5.69115341e-01 -2.97311544e-01
-1.34190708e-01 2.96965353e-02 3.82247627e-01 -4.87538278e-01
-1.10720372e+00 -1.90475270e-01 -1.11570191e+00 7.08582282e-01
5.89960337e-01 1.32959917e-01 -4.77694839e-01 -2.66760498e-01
-5.94701648e-01 5.33940420e-02 5.30480027e-01 7.21938908e-01
-5.43892533e-02 -1.05466878e+00 -4.72381264e-01 6.67221904e-01
2.39721928e-02 -5.47293544e-01 -3.80243063e-02 8.55788887e-01
-3.01218450e-01 1.18748403e+00 2.93306559e-02 -9.51513797e-02
-6.81125581e-01 6.11879885e-01 1.21735141e-01 -6.53737247e-01
-4.48098660e-01 9.30777013e-01 2.66090840e-01 -6.99557304e-01
3.95513661e-02 -5.24326444e-01 -1.24766327e-01 -7.75703192e-02
3.27274919e-01 2.63566822e-02 2.07341775e-01 -1.11064065e+00
-7.83476889e-01 -1.63511232e-01 -1.95884094e-01 -6.31162167e-01
6.31998241e-01 -3.28310430e-01 -1.54665232e-01 3.30783457e-01
8.29768598e-01 4.27375644e-01 -6.47286236e-01 -2.51061916e-02
2.54003793e-01 -1.98392078e-01 1.31345779e-01 -1.19945478e+00
-4.74573374e-01 2.92411953e-01 1.17982231e-01 -4.94197533e-02
8.04783404e-01 1.99527398e-01 8.96604776e-01 -7.30211139e-02
3.40574086e-01 -1.42420554e+00 -3.04943711e-01 1.37459016e+00
7.32688427e-01 -1.11146712e+00 -7.59873211e-01 -5.18148482e-01
-8.45375657e-01 3.03890944e-01 9.28332508e-01 1.49050429e-01
6.60679519e-01 4.70221192e-01 4.78266180e-01 -2.49969140e-01
-7.63125479e-01 -5.18187821e-01 1.59835517e-01 4.84809875e-01
1.17377424e+00 4.14269477e-01 -1.32512379e+00 8.70256603e-01
-7.90740967e-01 -9.27500427e-02 7.47517042e-04 1.06458712e+00
-2.44969755e-01 -1.29970551e+00 -2.33852506e-01 3.36412430e-01
-6.77419007e-01 -8.41472328e-01 -4.60250586e-01 9.30191517e-01
2.37879753e-01 1.24974728e+00 4.71905142e-01 -4.06509787e-01
5.43054268e-02 5.63647747e-01 2.38815740e-01 -1.13547623e+00
-5.78535020e-01 1.40049860e-01 3.34920466e-01 -7.46245384e-01
1.05821833e-01 -7.75531113e-01 -1.47306550e+00 7.43675185e-03
-2.69006252e-01 1.76291078e-01 1.66717172e-01 1.48425627e+00
1.60007715e-01 1.10622600e-01 4.34923172e-02 -6.32396996e-01
1.63096131e-03 -1.35405385e+00 -5.91758192e-01 3.57613027e-01
2.05362916e-01 -6.36246204e-01 -5.23944259e-01 -2.69471526e-01] | [10.701452255249023, 9.88488483428955] |
309fb2ec-17a5-4782-aac1-426928717888 | serial-contrastive-knowledge-distillation-for | 2305.06616 | null | https://arxiv.org/abs/2305.06616v1 | https://arxiv.org/pdf/2305.06616v1.pdf | Serial Contrastive Knowledge Distillation for Continual Few-shot Relation Extraction | Continual few-shot relation extraction (RE) aims to continuously train a model for new relations with few labeled training data, of which the major challenges are the catastrophic forgetting of old relations and the overfitting caused by data sparsity. In this paper, we propose a new model, namely SCKD, to accomplish the continual few-shot RE task. Specifically, we design serial knowledge distillation to preserve the prior knowledge from previous models and conduct contrastive learning with pseudo samples to keep the representations of samples in different relations sufficiently distinguishable. Our experiments on two benchmark datasets validate the effectiveness of SCKD for continual few-shot RE and its superiority in knowledge transfer and memory utilization over state-of-the-art models. | ['Wei Hu', 'Zitao Wang', 'Xinyi Wang'] | 2023-05-11 | null | null | null | null | ['relation-extraction'] | ['natural-language-processing'] | [-2.89810617e-02 3.28612179e-01 -5.33652186e-01 -2.44093344e-01
-1.34656399e-01 1.70601681e-01 4.30296987e-01 2.09087551e-01
-4.18103427e-01 1.10186708e+00 5.60138002e-02 -1.04293628e-02
-2.78176069e-01 -8.73447895e-01 -4.38927978e-01 -3.93531948e-01
1.49270847e-01 7.42467940e-01 5.48697770e-01 -4.20474738e-01
-2.10396603e-01 4.94643539e-01 -1.16587973e+00 2.25655437e-02
9.48536396e-01 8.37371409e-01 7.90799186e-02 9.82847810e-02
-3.12366098e-01 1.21481216e+00 -5.19699097e-01 -6.73111975e-01
-2.73529232e-01 -4.15787041e-01 -1.01938057e+00 -1.05354294e-01
-2.24666998e-01 -2.02929333e-01 -1.00062144e+00 9.24258828e-01
4.39068645e-01 6.20499074e-01 5.33196390e-01 -1.10197008e+00
-1.32660496e+00 8.41282606e-01 -5.61705112e-01 6.64119005e-01
-3.31811309e-02 -1.78789690e-01 7.55645871e-01 -1.29103851e+00
7.00328827e-01 1.17127085e+00 5.95310926e-01 6.11953020e-01
-1.14170682e+00 -8.39417815e-01 3.70024055e-01 6.20819628e-01
-1.69534552e+00 -6.48238242e-01 6.12094641e-01 -1.73648283e-01
1.21929455e+00 -4.04162630e-02 5.09851635e-01 8.91085625e-01
-1.74346372e-01 8.78332973e-01 4.84300971e-01 -4.66980487e-01
1.62169382e-01 1.12738885e-01 5.61207831e-01 6.20594501e-01
5.08194685e-01 8.52335244e-04 -7.17866123e-01 1.14196576e-01
6.49018228e-01 2.96812207e-01 -2.05401734e-01 -3.48003685e-01
-8.67491364e-01 3.90768737e-01 5.34704924e-01 4.20472175e-01
-2.71640331e-01 -1.94194496e-01 3.93399090e-01 4.79493678e-01
6.50000155e-01 4.77188528e-01 -6.72644019e-01 6.44715205e-02
-4.03046310e-01 -1.26403823e-01 5.94317138e-01 1.36386228e+00
9.00197685e-01 -6.15150034e-02 -5.40626407e-01 1.03113663e+00
-3.12427849e-01 5.95735461e-02 7.82546043e-01 -3.86111498e-01
5.34395695e-01 8.70097041e-01 -2.09524378e-01 -8.76710713e-01
-2.89513409e-01 -5.08972466e-01 -1.14861143e+00 -5.25133371e-01
-2.04830617e-01 -2.39625767e-01 -1.07904935e+00 1.48562443e+00
3.60975146e-01 6.55796587e-01 4.66634959e-01 4.18262035e-01
1.15391648e+00 5.48130214e-01 3.23987722e-01 -6.92235827e-01
1.04988885e+00 -1.19836128e+00 -9.71145332e-01 -6.45645976e-01
7.25834489e-01 -2.01896816e-01 8.67592633e-01 -1.95004884e-02
-8.15424263e-01 -7.90899515e-01 -1.14554262e+00 -2.51563728e-01
-4.61666942e-01 6.94427872e-03 9.50883567e-01 1.56800449e-01
-2.72111118e-01 7.63715804e-01 -8.55516136e-01 -2.54025757e-01
8.57847095e-01 2.65734226e-01 -3.52893054e-01 -2.23732606e-01
-1.64786327e+00 1.15651691e+00 1.04981160e+00 1.02780119e-01
-6.91216052e-01 -9.36044693e-01 -9.20535207e-01 4.52452391e-01
1.03304935e+00 -3.87456030e-01 1.06966996e+00 -1.40460610e-01
-1.30417883e+00 5.99765480e-01 -1.95748135e-01 -4.25025582e-01
1.20179862e-01 -5.68619132e-01 -7.78561175e-01 -2.10648909e-01
-1.36580527e-01 2.13504553e-01 4.56059396e-01 -1.23647308e+00
-5.44231236e-01 -2.18677089e-01 -1.14951700e-01 1.47893950e-01
-7.47653306e-01 -2.93275982e-01 -5.41448534e-01 -6.65748119e-01
1.27640918e-01 -5.34593761e-01 -1.63829625e-02 -5.07106185e-01
-2.43159965e-01 -5.61335862e-01 7.48632431e-01 -4.46399957e-01
1.55639899e+00 -2.26766372e+00 2.50588685e-01 -2.13422373e-01
4.38107550e-01 8.78271759e-01 -2.31206805e-01 6.25738055e-02
-3.01122427e-01 -1.52216941e-01 1.23356752e-01 -2.14869499e-01
-5.58571637e-01 4.10748065e-01 -2.71151781e-01 -5.62232174e-02
3.73482734e-01 1.33119106e+00 -1.37604761e+00 -5.77944875e-01
-8.67995098e-02 2.20841303e-01 1.32897079e-01 3.26388061e-01
-9.66306403e-03 2.35719495e-02 -3.20642024e-01 5.54800093e-01
5.07446170e-01 -5.16719460e-01 3.35031688e-01 -2.16116011e-01
4.58177298e-01 1.96438134e-01 -8.27926099e-01 1.72301018e+00
-2.46159956e-01 2.77069896e-01 -7.63598979e-01 -9.44345593e-01
1.17486596e+00 2.98949927e-01 1.33955598e-01 -6.34979427e-01
2.20463023e-01 -8.62200037e-02 4.93850000e-02 -4.98915762e-01
4.23651516e-01 -4.45254803e-01 2.56434113e-01 3.04329723e-01
5.84157825e-01 1.70592397e-01 1.09466337e-01 4.10946459e-01
1.13354480e+00 -1.45416245e-01 6.76952779e-01 2.24270359e-01
2.80960947e-01 -1.92152604e-01 1.04397893e+00 4.49405640e-01
-1.03795514e-01 1.72397524e-01 4.17525679e-01 -4.70267385e-01
-6.00059032e-01 -9.41508055e-01 2.37305224e-01 1.01681507e+00
3.45739126e-01 -5.31020939e-01 5.21794334e-02 -1.05607259e+00
1.94799565e-02 9.91945088e-01 -7.02147603e-01 -1.10736632e+00
-6.07252061e-01 -9.00284708e-01 1.46107987e-01 7.92606413e-01
6.62850797e-01 -1.16802573e+00 -1.12601876e-01 3.73248726e-01
-3.01849991e-02 -1.14042044e+00 -3.19325954e-01 4.05047596e-01
-8.28975320e-01 -1.00216782e+00 -5.01817107e-01 -1.05940020e+00
7.41666257e-01 5.85637689e-01 1.05702400e+00 -1.55589152e-02
-1.94200903e-01 -3.18759769e-01 -6.66150630e-01 -2.57643878e-01
-1.36221200e-01 4.44781870e-01 1.43759906e-01 -2.12254256e-01
7.24247754e-01 -8.59675467e-01 -1.91454068e-01 1.07787602e-01
-5.61666369e-01 1.68017242e-02 7.78006971e-01 1.18025243e+00
6.69410765e-01 5.62403142e-01 8.96710873e-01 -1.41739345e+00
7.48527765e-01 -6.69456184e-01 -1.50528818e-01 8.89342725e-01
-1.04423094e+00 1.51532069e-01 5.46084940e-01 -8.57370257e-01
-1.38624644e+00 -1.95913374e-01 5.24235129e-01 -7.42357194e-01
3.08496147e-01 7.95392632e-01 -2.01827511e-01 9.38687623e-02
6.37066245e-01 2.14066803e-01 -2.70014852e-01 -5.73143065e-01
6.13895774e-01 4.07182187e-01 7.01994836e-01 -4.33416337e-01
6.73847973e-01 4.36491035e-02 -1.10498235e-01 -4.19851661e-01
-1.64070570e+00 -3.29368234e-01 -7.78911233e-01 2.00758561e-01
4.48668629e-01 -8.77766848e-01 -4.16998386e-01 5.20274639e-01
-1.09891319e+00 -2.13215187e-01 -8.35403025e-01 4.28695053e-01
-2.79846899e-02 8.48252699e-02 -7.81358242e-01 -7.30661094e-01
-4.17877287e-01 -3.77511591e-01 2.58899689e-01 6.89384818e-01
-1.63459748e-01 -8.01930785e-01 1.37093976e-01 -7.17534963e-03
8.75080824e-02 -2.25098014e-01 1.21412277e+00 -8.89914930e-01
-2.37715632e-01 -2.12164223e-01 -4.41658676e-01 3.12917948e-01
5.57281315e-01 -4.92479712e-01 -9.03015733e-01 -1.54010177e-01
-1.40970320e-01 -6.51406407e-01 1.24344695e+00 -2.50614941e-01
8.94408166e-01 -1.54178664e-01 -8.02308381e-01 3.37015599e-01
1.00832450e+00 4.13081855e-01 6.11273885e-01 -1.51459411e-01
1.03886962e+00 3.30478281e-01 9.83852804e-01 2.57785052e-01
4.68592644e-01 3.55590731e-01 -2.16377005e-01 1.30307376e-01
-4.06909615e-01 -5.22597253e-01 -2.15688184e-01 1.12669837e+00
-1.97017565e-01 2.97193527e-02 -7.89051533e-01 7.15212107e-01
-2.18566203e+00 -8.92476201e-01 4.98217374e-01 1.99861503e+00
1.27925944e+00 4.75215465e-01 -3.06803465e-01 6.14052005e-02
7.92506337e-01 4.23881365e-03 -8.17826748e-01 1.21874586e-01
-1.60492092e-01 4.88582343e-01 1.12988494e-01 3.24707657e-01
-1.03024209e+00 1.53197610e+00 6.02448463e+00 9.99001741e-01
-6.81509674e-01 3.16936404e-01 6.69860005e-01 -1.35832950e-01
4.17817496e-02 1.61342591e-01 -9.94815290e-01 6.75477013e-02
7.41275847e-01 -5.05347669e-01 2.55580574e-01 8.92916262e-01
-5.46814024e-01 -6.07912894e-03 -1.11749256e+00 9.75063682e-01
1.28849536e-01 -1.29164732e+00 7.13092908e-02 -4.74747777e-01
8.74624789e-01 -3.77130508e-01 -2.93438435e-01 9.84390795e-01
5.95556140e-01 -8.57665122e-01 3.54773249e-03 7.29894340e-01
7.78684497e-01 -8.69676948e-01 7.72529542e-01 2.91819423e-01
-1.20411730e+00 -2.94967052e-02 -6.11507773e-01 -1.77677125e-01
1.27399042e-01 7.89907277e-01 -8.65539432e-01 7.65043676e-01
3.29450041e-01 1.03414738e+00 -7.23300815e-01 6.73321903e-01
-6.02991343e-01 3.96013319e-01 6.96747452e-02 1.91419516e-02
-3.69360864e-01 3.11750591e-01 4.50380780e-02 8.67119908e-01
1.09448403e-01 6.97905242e-01 1.50002062e-01 5.72127461e-01
-3.41191888e-01 -4.53525782e-02 -2.15169072e-01 -1.59726590e-01
1.11363244e+00 1.10713005e+00 -4.90727276e-01 -5.94228983e-01
-4.09189224e-01 1.02460265e+00 1.16017509e+00 3.40134770e-01
-4.44516331e-01 -5.80094099e-01 2.85652727e-01 -3.14889133e-01
4.45832640e-01 -1.36899367e-01 -9.67407376e-02 -1.39613605e+00
-6.48942068e-02 -3.07756424e-01 6.61878824e-01 -5.86754739e-01
-1.58886051e+00 6.54131532e-01 4.28548455e-02 -7.32207716e-01
1.09531708e-01 -6.63094521e-02 -4.63598877e-01 6.83683872e-01
-1.44987249e+00 -1.10268652e+00 -3.95899028e-01 4.45973814e-01
4.65339303e-01 -4.18845862e-01 9.27951276e-01 4.09816861e-01
-9.40595269e-01 8.47957909e-01 -2.28428677e-01 3.13944966e-01
6.33986473e-01 -9.37658191e-01 5.47223806e-01 6.67220473e-01
2.24965915e-01 6.28079951e-01 3.67766470e-01 -1.05988121e+00
-1.05789018e+00 -1.21053052e+00 1.18113959e+00 -1.31542608e-01
6.70350015e-01 -1.43095151e-01 -1.34372342e+00 9.07666922e-01
-3.16356093e-01 4.56607699e-01 7.01818824e-01 6.97003841e-01
-4.96969342e-01 -4.12727267e-01 -8.84556830e-01 5.59576750e-01
1.35487139e+00 -3.87208164e-01 -1.04389203e+00 1.75095364e-01
1.16482091e+00 -2.30205998e-01 -8.61758232e-01 7.31972098e-01
2.98539370e-01 -2.88532734e-01 1.02385306e+00 -9.92828190e-01
2.58830160e-01 4.50074635e-02 2.65271425e-01 -1.46067226e+00
-7.14383841e-01 -4.38466936e-01 -1.08302486e+00 1.66727555e+00
3.20834219e-01 -4.97660577e-01 7.75833130e-01 5.91439009e-01
2.29856908e-01 -8.74841452e-01 -6.55747831e-01 -1.06268990e+00
-2.05877602e-01 1.30127087e-01 7.75177002e-01 1.26412833e+00
3.45593452e-01 1.24266601e+00 -6.21180773e-01 -4.19829711e-02
2.27209821e-01 1.32311791e-01 5.25303602e-01 -1.34262490e+00
-3.15302044e-01 3.08995191e-02 -4.44438666e-01 -8.71989489e-01
2.64458537e-01 -8.98365378e-01 -1.04206964e-01 -1.43022227e+00
5.12818992e-01 -5.77009857e-01 -8.92332971e-01 7.86668003e-01
-9.31860209e-01 -2.92663664e-01 -4.19385396e-02 2.98632205e-01
-9.59262490e-01 1.12555850e+00 1.39665878e+00 -1.47617087e-01
-5.09682238e-01 -1.68605596e-01 -8.90869558e-01 4.99621600e-01
5.68577468e-01 -6.46589398e-01 -9.13290560e-01 -2.07529902e-01
1.27358943e-01 -2.94590951e-03 -7.18738213e-02 -9.32378888e-01
4.63977903e-01 -3.04244190e-01 4.56448972e-01 -6.05844617e-01
4.74594831e-01 -6.40665174e-01 -8.48453119e-02 3.24747592e-01
-3.93686622e-01 -4.38519359e-01 1.35445148e-01 7.41813838e-01
-1.70598388e-01 -1.70016453e-01 6.13354445e-01 -1.15386300e-01
-1.12142587e+00 5.84791005e-01 3.93806309e-01 3.85116011e-01
1.14292288e+00 3.44651192e-01 -5.40089011e-01 8.44246522e-03
-1.10387862e+00 4.82276261e-01 -1.85037032e-01 5.51632881e-01
9.69810307e-01 -1.59281027e+00 -6.45208359e-01 6.28053024e-02
3.89226884e-01 3.60289961e-01 4.58687484e-01 4.29845721e-01
1.15955606e-01 9.60222855e-02 3.53330262e-02 2.38675982e-01
-1.14252210e+00 1.23425484e+00 4.69779409e-02 -5.59485435e-01
-8.50797176e-01 1.25930119e+00 -6.49295300e-02 -1.81510553e-01
2.10123628e-01 -1.19757624e-02 -3.44646901e-01 7.23138303e-02
7.77796626e-01 4.47087526e-01 2.01574657e-02 -1.22970212e-02
-3.76111925e-01 2.97460817e-02 -8.90203893e-01 3.26185554e-01
1.51137221e+00 -1.21966176e-01 -9.26790312e-02 7.46025324e-01
8.33565116e-01 -4.77531165e-01 -1.07582712e+00 -1.00645077e+00
2.56992787e-01 -3.62208217e-01 -1.03547074e-01 -8.32511246e-01
-1.08891571e+00 5.83958387e-01 1.99493363e-01 -9.56057683e-02
9.64783728e-01 3.53346050e-01 1.00786126e+00 8.00130427e-01
3.52451682e-01 -1.05343032e+00 1.95941418e-01 6.83752716e-01
5.64213037e-01 -1.12629437e+00 3.41321886e-01 -7.20427394e-01
-7.42234826e-01 7.27866232e-01 1.10339415e+00 1.20921163e-02
9.24522579e-01 1.44494221e-01 -2.94661283e-01 -2.99127042e-01
-9.57865238e-01 -3.91718715e-01 1.71981230e-01 6.24156117e-01
1.56513989e-01 -1.04643060e-02 -2.57214963e-01 1.04784834e+00
8.90400857e-02 5.13482034e-01 1.28939360e-01 1.18948805e+00
-3.63868982e-01 -1.04229403e+00 4.65452760e-01 6.13553286e-01
1.16503388e-01 -3.08821857e-01 -3.85988116e-01 5.63738644e-01
2.90402442e-01 9.15813446e-01 -1.59803525e-01 -8.36462319e-01
5.12900591e-01 2.93385416e-01 6.28578842e-01 -1.13572907e+00
-1.65154040e-01 -3.91386002e-01 2.54250944e-01 -5.35961725e-02
-1.96832776e-01 -3.82414758e-01 -1.44154179e+00 -2.14662164e-01
-6.53990626e-01 2.79072970e-01 -1.97542608e-01 1.25918758e+00
5.94092786e-01 1.03146660e+00 4.66573417e-01 -1.99666068e-01
-4.71199483e-01 -1.29600942e+00 -7.68826485e-01 3.35214198e-01
6.49465173e-02 -1.09641576e+00 3.64083238e-02 -1.56435251e-01] | [9.145071983337402, 8.522132873535156] |
001bfc3f-cfe0-4af1-8b94-9347e3c51a6c | benchmarking-robustness-of-ai-enabled-multi | 2306.03454 | null | https://arxiv.org/abs/2306.03454v1 | https://arxiv.org/pdf/2306.03454v1.pdf | Benchmarking Robustness of AI-enabled Multi-sensor Fusion Systems: Challenges and Opportunities | Multi-Sensor Fusion (MSF) based perception systems have been the foundation in supporting many industrial applications and domains, such as self-driving cars, robotic arms, and unmanned aerial vehicles. Over the past few years, the fast progress in data-driven artificial intelligence (AI) has brought a fast-increasing trend to empower MSF systems by deep learning techniques to further improve performance, especially on intelligent systems and their perception systems. Although quite a few AI-enabled MSF perception systems and techniques have been proposed, up to the present, limited benchmarks that focus on MSF perception are publicly available. Given that many intelligent systems such as self-driving cars are operated in safety-critical contexts where perception systems play an important role, there comes an urgent need for a more in-depth understanding of the performance and reliability of these MSF systems. To bridge this gap, we initiate an early step in this direction and construct a public benchmark of AI-enabled MSF-based perception systems including three commonly adopted tasks (i.e., object detection, object tracking, and depth completion). Based on this, to comprehensively understand MSF systems' robustness and reliability, we design 14 common and realistic corruption patterns to synthesize large-scale corrupted datasets. We further perform a systematic evaluation of these systems through our large-scale evaluation. Our results reveal the vulnerability of the current AI-enabled MSF perception systems, calling for researchers and practitioners to take robustness and reliability into account when designing AI-enabled MSF. | ['Baowen Xu', 'Zhenyu Chen', 'Lei Ma', 'Yang Feng', 'Zhijie Wang', 'Xinyu Gao'] | 2023-06-06 | null | null | null | null | ['self-driving-cars', 'object-tracking', 'depth-completion'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 5.28849423e-01 -1.16193667e-01 -4.94693108e-02 -4.50282902e-01
-4.70609784e-01 -4.79442686e-01 7.55350828e-01 8.84694234e-02
-1.82912692e-01 2.61522710e-01 -1.09192543e-01 -2.70949543e-01
-2.44334221e-01 -9.29460287e-01 -8.14858913e-01 -7.24240899e-01
-1.38641313e-01 7.35451058e-02 4.96596247e-01 -5.91909289e-01
1.47999540e-01 5.59876561e-01 -2.14883494e+00 2.29236171e-01
6.95877850e-01 1.48115468e+00 5.23570836e-01 6.89682841e-01
4.53552455e-01 6.69929862e-01 -5.16809762e-01 -2.17119440e-01
3.25752586e-01 1.39002293e-01 -3.62505674e-01 1.65396348e-01
3.92791301e-01 -1.28062680e-01 -4.57153320e-01 1.15084779e+00
3.71835351e-01 2.03774665e-02 4.21254992e-01 -1.79491675e+00
-6.29947245e-01 2.75311410e-01 -2.86542028e-01 2.98218485e-02
2.46913448e-01 7.55025864e-01 7.67066360e-01 -7.22851574e-01
1.23171963e-01 1.43957698e+00 5.55372775e-01 3.97879481e-01
-7.24244535e-01 -5.43239236e-01 2.70242125e-01 4.23216194e-01
-9.32199299e-01 -4.85069513e-01 8.09333146e-01 -4.39772576e-01
1.03212130e+00 4.68983464e-02 4.94943529e-01 1.17694032e+00
5.92904806e-01 8.05966854e-01 9.77528036e-01 -1.76195771e-01
4.72094059e-01 -2.04077363e-01 7.21578896e-02 4.45254445e-01
4.06988800e-01 5.00898719e-01 -4.08797175e-01 3.22214186e-01
4.46755379e-01 -5.53926341e-02 3.57417837e-02 -3.66996586e-01
-1.56404054e+00 6.98860705e-01 7.77070045e-01 1.40838578e-01
-5.15238881e-01 4.06549275e-01 4.30187106e-01 2.91081280e-01
-6.71796054e-02 3.92958343e-01 -3.49834025e-01 -1.09670743e-01
-4.04538184e-01 5.31615615e-01 4.70308393e-01 9.29976940e-01
7.30746686e-01 2.57627547e-01 2.22271234e-02 7.70753622e-01
3.64883482e-01 9.28986967e-01 2.59044260e-01 -1.24073553e+00
3.55504841e-01 7.15101302e-01 9.85399559e-02 -1.18158782e+00
-4.46299702e-01 -1.99587196e-01 -1.04264939e+00 6.85678244e-01
4.05568480e-02 1.68162242e-01 -9.00743127e-01 1.70718670e+00
1.68792561e-01 1.53339893e-01 4.00829077e-01 1.13835919e+00
6.73224628e-01 4.76960629e-01 -1.22349791e-01 4.30514067e-02
1.38808954e+00 -7.48303235e-01 -6.35046840e-01 -4.46328521e-01
2.08385184e-01 -7.19501078e-01 9.52538490e-01 7.10766733e-01
-8.09301794e-01 -1.11188877e+00 -1.44461024e+00 1.51832730e-01
-6.47052169e-01 -1.37020126e-01 8.09154391e-01 5.29902220e-01
-7.67039716e-01 4.15930718e-01 -7.96553075e-01 -3.42218220e-01
6.26682520e-01 3.18764061e-01 -4.10975754e-01 -3.64369839e-01
-1.32844472e+00 1.05257368e+00 4.06053990e-01 3.48425090e-01
-1.53724253e+00 -3.26879978e-01 -7.90834308e-01 -2.47229859e-01
6.33838415e-01 -7.68554866e-01 1.36925423e+00 -6.26472473e-01
-1.35002387e+00 5.40321469e-01 2.13351563e-01 -8.32553804e-01
2.50875801e-01 -3.37244868e-01 -5.97796738e-01 -1.64111227e-01
-3.63975093e-02 8.21446359e-01 8.84932995e-01 -1.54368484e+00
-9.93690312e-01 -3.99017870e-01 4.38388288e-01 -5.79698868e-02
-1.19593620e-01 -2.40730017e-01 -1.97470374e-02 -3.08336437e-01
-9.38316924e-04 -1.02388215e+00 -2.35174716e-01 8.52007493e-02
-3.46119285e-01 -3.89889985e-01 1.28221369e+00 9.12199020e-02
6.90701663e-01 -2.16697669e+00 4.54866849e-02 -1.54700667e-01
2.51624525e-01 4.71535802e-01 -2.75530219e-01 2.89193064e-01
1.89825520e-01 -3.23446512e-01 -2.58202910e-01 -3.70177120e-01
2.36387864e-01 5.15040517e-01 -4.77866709e-01 3.43101382e-01
3.82929593e-01 8.78945887e-01 -9.28672969e-01 -8.61095190e-02
8.29355597e-01 3.95192385e-01 -3.51267874e-01 2.95758367e-01
-5.05831063e-01 2.10321382e-01 -4.35910553e-01 9.24512506e-01
5.99566936e-01 2.09358394e-01 -4.73016888e-01 -3.56608748e-01
-1.36156306e-01 -1.01788923e-01 -1.01938772e+00 1.70381570e+00
-3.79046351e-01 5.91744840e-01 2.10304141e-01 -1.20289969e+00
1.08380353e+00 1.17935851e-01 5.58582008e-01 -9.15688396e-01
3.15664023e-01 1.63868681e-01 2.19264254e-01 -5.07204056e-01
5.98703623e-01 5.85833751e-02 -3.08311343e-01 -1.63815811e-01
-1.07238770e-01 -4.69912559e-01 2.45735064e-01 7.95450155e-03
1.23003745e+00 -1.30530708e-02 -3.50432517e-03 3.35155055e-02
6.08955085e-01 2.90429890e-01 5.37525952e-01 7.32260644e-01
-6.81853831e-01 4.60252345e-01 -1.57362387e-01 -5.64505041e-01
-7.12983787e-01 -1.22933447e+00 -5.16490638e-02 9.27141428e-01
5.27450144e-01 -1.29699409e-01 -5.52363575e-01 -4.13860619e-01
3.25956702e-01 5.96377015e-01 -4.58060950e-01 -5.70794344e-01
-3.21727216e-01 -6.33684278e-01 5.96109211e-01 6.68228269e-01
8.75273347e-01 -1.26868570e+00 -1.05904520e+00 3.23665231e-01
-8.79240315e-03 -1.56369364e+00 3.07730615e-01 1.67199552e-01
-4.34455156e-01 -1.19219530e+00 -2.58935004e-01 -5.80262542e-01
9.52240452e-02 8.37556422e-01 1.01514614e+00 -1.49259761e-01
-2.36691609e-01 5.35434723e-01 -3.74590218e-01 -1.03532183e+00
-6.50832772e-01 -2.44411483e-01 5.15331149e-01 -1.14570409e-01
2.00933248e-01 -4.92176771e-01 -6.03502572e-01 5.12166858e-01
-9.25960779e-01 -8.97375941e-02 7.93934107e-01 5.19095659e-01
4.84193981e-01 3.49149615e-01 8.40963364e-01 -2.23237321e-01
6.07092261e-01 -4.71247971e-01 -7.71962702e-01 3.86410132e-02
-5.46484768e-01 -4.42668796e-01 7.22722232e-01 -3.37086469e-01
-7.08858132e-01 1.46278981e-02 -2.12897807e-01 -6.84974372e-01
-5.68224072e-01 4.46894646e-01 -3.97720486e-01 -2.35000835e-03
6.98191106e-01 1.18794121e-01 1.20443761e-01 -5.43454774e-02
3.73002887e-01 8.37791026e-01 1.14554369e+00 -3.99684489e-01
8.40412736e-01 3.58812451e-01 2.74527222e-01 -8.13858211e-01
-6.65560663e-01 -2.98442125e-01 -2.67084539e-01 -4.87078100e-01
8.31384718e-01 -9.32486892e-01 -1.03563774e+00 6.77553058e-01
-1.12499702e+00 -1.97282091e-01 -1.64157599e-01 3.01198065e-01
-8.49024594e-01 2.67698258e-01 -2.11147994e-01 -8.89146566e-01
-2.13885829e-01 -1.56194556e+00 1.32564712e+00 1.40837163e-01
1.23267835e-02 -5.30556798e-01 -6.97290227e-02 6.85122907e-01
5.41980565e-01 3.73715252e-01 6.41359448e-01 -2.49156535e-01
-8.84763837e-01 -1.40518904e-01 -2.31479704e-01 5.19698679e-01
2.41937190e-01 1.16520241e-01 -1.16647506e+00 -3.85713667e-01
-1.69124484e-01 -5.14684796e-01 9.69202876e-01 3.29805315e-01
9.89530861e-01 2.33613357e-01 -3.19229245e-01 4.17954028e-01
1.15491998e+00 3.95568848e-01 5.55052280e-01 3.85160804e-01
6.66479826e-01 5.34346282e-01 1.10871363e+00 2.59980977e-01
6.02703750e-01 6.35514855e-01 1.26602185e+00 2.33348887e-02
-2.40888163e-01 1.71226473e-03 3.93205464e-01 4.09505427e-01
2.19665095e-03 -4.95439917e-01 -7.75833547e-01 3.62391979e-01
-1.90964556e+00 -8.91566575e-01 -8.09630007e-03 2.13079643e+00
1.44468099e-01 5.44375360e-01 -6.15807883e-02 6.00821435e-01
4.96021688e-01 8.05840790e-02 -9.14641917e-01 -2.86527067e-01
-2.30064765e-01 -3.62342983e-01 2.22462535e-01 1.38605177e-01
-1.40633154e+00 8.11671436e-01 5.72920990e+00 4.78914052e-01
-1.27421081e+00 -2.13836566e-01 3.66163582e-01 3.61763179e-01
2.99235471e-02 -2.79548675e-01 -6.06033146e-01 4.70220000e-01
8.23616862e-01 2.86400225e-02 3.71586561e-01 1.13263202e+00
2.77466565e-01 -2.80269474e-01 -1.02758753e+00 1.16654551e+00
-1.49611801e-01 -1.30619895e+00 -1.50288284e-01 9.45986286e-02
5.36556304e-01 2.58533984e-01 1.43821493e-01 4.41067785e-01
2.64146119e-01 -9.95057940e-01 7.43039787e-01 3.37177187e-01
4.46778029e-01 -6.66043341e-01 8.75778735e-01 4.05224383e-01
-1.23730791e+00 -5.45924604e-01 -5.31185091e-01 -3.40376407e-01
1.23611473e-01 8.98900628e-01 -6.98754787e-01 6.55031264e-01
8.99281144e-01 8.08486640e-01 -3.73718143e-01 8.47540259e-01
-7.12257102e-02 3.62561077e-01 -2.73219496e-01 4.81163785e-02
1.70090780e-01 2.32119128e-01 6.27521753e-01 8.91219318e-01
2.16170713e-01 -1.75845832e-01 3.50585550e-01 7.25195944e-01
1.50091425e-01 -6.64853156e-01 -8.64747167e-01 -4.98360991e-02
5.17941535e-01 1.34218085e+00 -6.02649927e-01 -2.19882876e-01
-4.85633165e-01 5.93474746e-01 -8.43679681e-02 7.91592151e-02
-1.00373137e+00 -1.43329158e-01 1.22895443e+00 -8.07635039e-02
1.26025617e-01 -5.24745286e-01 -3.67722362e-01 -7.27213264e-01
-7.88309947e-02 -9.87406254e-01 8.68222266e-02 -9.63077426e-01
-1.28686249e+00 7.10237086e-01 -2.87243910e-02 -1.34286714e+00
-2.47441351e-01 -7.75172949e-01 -4.23576891e-01 3.62900883e-01
-1.64714026e+00 -1.34802067e+00 -7.92048872e-01 5.31180084e-01
6.28066540e-01 -3.65504622e-01 7.18451381e-01 1.67323366e-01
-6.98122680e-01 1.70767203e-01 -2.16295302e-01 -1.17625482e-01
6.52029335e-01 -9.11691844e-01 5.59491813e-01 9.44753647e-01
-1.67601369e-02 3.38747561e-01 8.43762934e-01 -4.79236066e-01
-1.99417865e+00 -1.16757464e+00 1.15838833e-01 -5.86080372e-01
5.89018285e-01 -3.23821634e-01 -6.30034029e-01 3.55802685e-01
1.45594105e-01 2.16582507e-01 3.73402331e-03 -1.93433583e-01
-1.51439056e-01 -6.50990188e-01 -1.16781902e+00 3.60617876e-01
1.01058078e+00 -3.11022043e-01 -6.52968466e-01 1.62829712e-01
7.98934281e-01 -2.33227357e-01 -8.01677585e-01 1.06283438e+00
5.56730092e-01 -1.27539849e+00 1.21983516e+00 -8.53761844e-03
4.56702620e-01 -6.67844832e-01 -6.04233861e-01 -1.38822460e+00
-3.09340954e-01 -3.57732624e-01 -1.39642924e-01 1.00772548e+00
-2.13155448e-01 -7.75520504e-01 6.68384492e-01 1.91179797e-01
-7.25594521e-01 -8.53976488e-01 -9.93227601e-01 -7.13776588e-01
-1.31081596e-01 -9.71142590e-01 6.38835192e-01 4.25825298e-01
-4.03013706e-01 1.32874012e-01 -1.86976627e-01 4.26541597e-01
6.00294888e-01 2.04234883e-01 9.87947643e-01 -1.26535904e+00
8.70104358e-02 -3.95861536e-01 -7.87884653e-01 -9.56892788e-01
-8.99388343e-02 -3.83669466e-01 4.04216915e-01 -1.58860183e+00
-1.34225994e-01 -4.17743891e-01 -4.81180638e-01 5.68254173e-01
5.18430248e-02 3.45006227e-01 3.11353862e-01 9.99100059e-02
-7.72167504e-01 6.66965961e-01 1.37356937e+00 -4.25896764e-01
2.22445115e-01 1.05146825e-01 -8.02533567e-01 7.34185815e-01
7.26340652e-01 -1.34341633e-02 -6.36249065e-01 -3.69729161e-01
-3.22681852e-02 -1.36089370e-01 6.61783814e-01 -1.59218371e+00
4.03720111e-01 -2.64516532e-01 3.09360325e-01 -7.44872212e-01
6.23105347e-01 -1.11018074e+00 -1.00800820e-01 6.16527855e-01
6.86358437e-02 1.03139721e-01 2.35424384e-01 4.06619728e-01
-4.96294439e-01 3.47567290e-01 8.40546727e-01 -3.79354283e-02
-1.36499226e+00 2.57046193e-01 -5.15276492e-01 -4.05323386e-01
1.29016197e+00 -4.47508574e-01 -5.11759281e-01 -3.94431829e-01
-1.66087538e-01 3.25947106e-01 3.81036907e-01 9.42597747e-01
9.49567556e-01 -1.16200256e+00 -5.03299117e-01 5.95567465e-01
3.74775052e-01 3.51367205e-01 2.97158539e-01 4.38376307e-01
-7.22466856e-02 6.78491056e-01 -4.50156540e-01 -9.10453498e-01
-9.33853805e-01 8.43653202e-01 1.86927527e-01 1.70511171e-01
-2.61154234e-01 6.28451586e-01 2.26610914e-01 -4.37538832e-01
3.99184972e-01 -7.99454570e-01 7.39826262e-02 -3.01435441e-01
4.39291775e-01 5.29146612e-01 1.87830269e-01 -6.49866879e-01
-5.01451433e-01 5.44301748e-01 3.08062464e-01 4.11226600e-01
1.09335864e+00 -8.56176466e-02 3.76631439e-01 3.26765180e-01
7.50162721e-01 -4.19584990e-01 -1.45571506e+00 1.22203343e-01
-2.81585455e-01 -2.58171529e-01 2.49029055e-01 -8.53715420e-01
-1.06715381e+00 9.24490988e-01 7.85550117e-01 4.29836661e-01
1.30974531e+00 -3.08863055e-02 8.40497315e-01 6.76706016e-01
9.35044408e-01 -9.08360660e-01 3.18491846e-01 6.78055048e-01
1.05056024e+00 -1.68905234e+00 -1.37502030e-01 -4.88417327e-01
-4.79675829e-01 9.35517848e-01 8.87249649e-01 -4.21121642e-02
4.13519502e-01 5.41143537e-01 1.69978902e-01 -1.02614567e-01
-7.91235566e-01 -3.44934613e-01 1.95338622e-01 8.72637153e-01
-4.27585132e-02 1.40977046e-02 4.43477035e-01 5.18125772e-01
-2.42313623e-01 2.00574234e-01 2.12759316e-01 9.98733759e-01
-6.55729890e-01 -9.59563136e-01 -5.98529756e-01 2.24915624e-01
3.85567062e-02 4.13000405e-01 -8.21690559e-02 7.06030130e-01
5.51663339e-01 1.37667596e+00 -8.38944837e-02 -8.81074786e-01
7.26339996e-01 -4.44601536e-01 4.55825001e-01 -2.60146320e-01
-3.05528760e-01 -3.71097416e-01 3.52572612e-02 -8.05552900e-01
-5.31898618e-01 -4.66227025e-01 -1.25406313e+00 -3.18986714e-01
-1.44100770e-01 -4.82409060e-01 9.69662726e-01 9.69573081e-01
4.14705515e-01 7.76367068e-01 6.68697000e-01 -1.20978940e+00
-5.96792698e-01 -7.98820078e-01 -2.63842940e-01 4.87670660e-01
3.85594577e-01 -1.13526964e+00 -1.70995787e-01 -1.80541009e-01] | [7.857624530792236, -1.3868157863616943] |
7efd4258-5f7f-4701-9fe1-2868ec581830 | samossa-multivariate-singular-spectrum | 2305.16491 | null | https://arxiv.org/abs/2305.16491v1 | https://arxiv.org/pdf/2305.16491v1.pdf | SAMoSSA: Multivariate Singular Spectrum Analysis with Stochastic Autoregressive Noise | The well-established practice of time series analysis involves estimating deterministic, non-stationary trend and seasonality components followed by learning the residual stochastic, stationary components. Recently, it has been shown that one can learn the deterministic non-stationary components accurately using multivariate Singular Spectrum Analysis (mSSA) in the absence of a correlated stationary component; meanwhile, in the absence of deterministic non-stationary components, the Autoregressive (AR) stationary component can also be learnt readily, e.g. via Ordinary Least Squares (OLS). However, a theoretical underpinning of multi-stage learning algorithms involving both deterministic and stationary components has been absent in the literature despite its pervasiveness. We resolve this open question by establishing desirable theoretical guarantees for a natural two-stage algorithm, where mSSA is first applied to estimate the non-stationary components despite the presence of a correlated stationary AR component, which is subsequently learned from the residual time series. We provide a finite-sample forecasting consistency bound for the proposed algorithm, SAMoSSA, which is data-driven and thus requires minimal parameter tuning. To establish theoretical guarantees, we overcome three hurdles: (i) we characterize the spectra of Page matrices of stable AR processes, thus extending the analysis of mSSA; (ii) we extend the analysis of AR process identification in the presence of arbitrary bounded perturbations; (iii) we characterize the out-of-sample or forecasting error, as opposed to solely considering model identification. Through representative empirical studies, we validate the superior performance of SAMoSSA compared to existing baselines. Notably, SAMoSSA's ability to account for AR noise structure yields improvements ranging from 5% to 37% across various benchmark datasets. | ['Devavrat Shah', 'Sean Mann', 'Munther Dahleh', 'Abdullah Alomar'] | 2023-05-25 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 4.56222713e-01 -3.79598081e-01 1.28476411e-01 8.99277702e-02
-1.09550130e+00 -8.98332715e-01 5.90552866e-01 1.02878541e-01
4.95850714e-03 5.01343369e-01 3.97411287e-02 -5.57262421e-01
-5.68085790e-01 -2.68772542e-01 -9.06414270e-01 -1.04810977e+00
-3.87177944e-01 1.14211567e-01 -1.97816879e-01 -4.18224186e-02
-6.91687092e-02 4.88669425e-01 -1.34421146e+00 -5.67136467e-01
7.95635760e-01 1.32470787e+00 -1.32599428e-01 7.07355142e-01
2.16146797e-01 7.04708815e-01 -1.77210629e-01 -9.81869921e-02
5.22746027e-01 -3.43856066e-01 -1.88543975e-01 2.75087088e-01
7.50444643e-03 5.18468861e-03 -1.85621344e-02 9.04396832e-01
2.00019613e-01 1.58646300e-01 7.42891371e-01 -1.20221901e+00
-3.64227772e-01 2.80788034e-01 -7.80223131e-01 2.62159616e-01
8.05561692e-02 1.52762577e-01 1.08099806e+00 -8.77103508e-01
2.98652090e-02 1.20325315e+00 9.62516546e-01 1.78229474e-02
-1.59229088e+00 -4.70370024e-01 1.55363411e-01 -4.03870076e-01
-1.31780028e+00 -6.05349362e-01 9.63293493e-01 -7.77755916e-01
4.98336703e-01 2.53348202e-01 2.36416325e-01 1.23343790e+00
3.92884761e-01 6.83417678e-01 1.15502059e+00 -5.12867272e-01
4.20382500e-01 -2.83818305e-01 4.06247854e-01 2.71546513e-01
3.29683542e-01 2.25469887e-01 -2.87941158e-01 -5.36551118e-01
7.96928287e-01 8.61599818e-02 -1.08805068e-01 -4.33460176e-01
-1.07997334e+00 6.72948003e-01 -3.22932482e-01 1.65412039e-01
-9.38238025e-01 1.34921446e-01 2.48971194e-01 5.15221775e-01
6.27609909e-01 1.71802372e-01 -6.10144973e-01 -1.58831835e-01
-8.49649549e-01 4.12419066e-02 9.19733763e-01 7.42824912e-01
6.33433521e-01 6.21709347e-01 -6.83400705e-02 6.13723159e-01
3.50566894e-01 9.84075725e-01 4.45885897e-01 -8.36777508e-01
2.79378891e-01 1.34916544e-01 4.11952555e-01 -1.05774105e+00
-2.80844480e-01 -7.27634013e-01 -1.23539114e+00 -1.90739587e-01
5.92505991e-01 -3.91096413e-01 -6.21939838e-01 1.98347127e+00
2.56896228e-01 6.73558712e-01 2.98603892e-01 4.45103616e-01
-7.96705559e-02 6.37064099e-01 -2.12486997e-01 -9.05704618e-01
1.14960766e+00 -5.47883868e-01 -7.79559553e-01 -2.75812417e-01
1.83009297e-01 -7.08885193e-01 9.44121778e-01 4.44960386e-01
-8.74463916e-01 -3.74723881e-01 -8.13797176e-01 4.34333920e-01
1.42279536e-01 1.57099947e-01 4.52881128e-01 5.91721535e-01
-9.04807627e-01 5.03287435e-01 -1.22009826e+00 -1.83741182e-01
-5.33758737e-02 1.93761542e-01 -1.29325405e-01 2.81691670e-01
-1.11280417e+00 4.25854504e-01 -1.28506005e-01 4.85969186e-01
-9.31361079e-01 -9.73063767e-01 -6.60947323e-01 7.99670815e-02
5.90292573e-01 -7.18470871e-01 1.17355025e+00 -1.07723141e+00
-1.53964007e+00 1.89535484e-01 -6.26657307e-01 -5.98651171e-01
5.63494980e-01 -2.49760225e-01 -6.52073443e-01 7.94148892e-02
-5.24548255e-02 -4.99474406e-01 1.45701039e+00 -1.13267958e+00
-5.07357597e-01 -4.31317061e-01 -2.36406118e-01 -1.65509105e-01
-2.01043606e-01 -1.33834794e-01 -1.85255751e-01 -8.94656777e-01
3.35551113e-01 -1.30665016e+00 -3.22712988e-01 -4.53841627e-01
-3.97207260e-01 5.90076558e-02 4.26599652e-01 -7.15970516e-01
1.37740827e+00 -2.39325690e+00 -2.30435468e-02 4.43614572e-01
2.68451101e-03 -2.20489860e-01 -1.01400234e-01 6.63812757e-01
-3.36994737e-01 -5.46549410e-02 -3.71203184e-01 -5.05329490e-01
6.32937774e-02 1.29454136e-01 -1.04553545e+00 5.76038480e-01
2.51264840e-01 5.72608590e-01 -7.43044913e-01 1.26082465e-01
3.67824621e-02 4.85611767e-01 -9.66054350e-02 7.38912728e-03
1.57579124e-01 5.54297030e-01 -3.53136957e-01 5.47740221e-01
3.72689724e-01 -4.91731524e-01 1.06013715e-01 -8.71368200e-02
-1.55156881e-01 -7.52005354e-02 -1.66693044e+00 9.76880252e-01
-6.36623085e-01 5.01687050e-01 1.47119135e-01 -1.03250444e+00
8.05675149e-01 4.71073538e-01 7.32170045e-01 -1.57783896e-01
-5.75613342e-02 3.37491184e-01 -2.00372100e-01 -1.13778681e-01
1.69352710e-01 -2.56925493e-01 -1.42936117e-03 4.19645071e-01
-1.70018956e-01 3.31210822e-01 4.15497012e-02 -5.73954508e-02
1.11634636e+00 -5.82657084e-02 4.51568574e-01 -4.34274763e-01
5.71821451e-01 -2.75885880e-01 6.03117347e-01 8.82863700e-01
-5.77484444e-02 3.96022499e-01 5.06676495e-01 -1.33510411e-01
-8.80404294e-01 -1.24133003e+00 -1.12012476e-01 9.25300837e-01
-2.48782560e-01 -1.78497598e-01 -2.73882389e-01 -2.02990719e-03
1.24325514e-01 6.19015455e-01 -6.00451887e-01 -3.86820477e-03
-3.73159587e-01 -1.05398571e+00 2.00799167e-01 5.04338384e-01
8.17753281e-03 -3.09310377e-01 -3.19209218e-01 3.67048651e-01
6.29698411e-02 -1.16974497e+00 -5.09961784e-01 3.16279978e-01
-9.75888610e-01 -8.45871389e-01 -5.94169199e-01 -2.19909802e-01
4.50788438e-01 4.65463638e-01 8.34525645e-01 -4.43349570e-01
1.85140371e-01 1.00588095e+00 2.00335994e-01 -5.61881244e-01
-4.27942187e-01 -1.48726463e-01 4.33170527e-01 5.91476440e-01
2.19789166e-02 -8.48862648e-01 -4.16833252e-01 1.85166761e-01
-9.41938519e-01 -2.80290514e-01 7.28858531e-01 7.91122496e-01
7.54894614e-01 3.72360021e-01 7.41985142e-01 -5.89658260e-01
5.84004343e-01 -6.16093516e-01 -9.67525899e-01 2.69676805e-01
-8.44297349e-01 9.25496891e-02 7.09757864e-01 -7.21238434e-01
-8.67394865e-01 2.21139967e-01 3.98332387e-01 -7.56862879e-01
1.12460129e-01 8.33665013e-01 1.38945758e-01 1.59960002e-01
3.27300817e-01 5.47146976e-01 2.50803173e-01 -3.91199201e-01
7.65034407e-02 3.32238525e-01 6.56186223e-01 -6.56575143e-01
1.24854708e+00 6.10859275e-01 4.14738089e-01 -1.14745891e+00
-8.14730287e-01 -7.47044146e-01 -4.87490833e-01 2.70005893e-02
3.68362695e-01 -1.29616916e+00 -8.18644404e-01 6.24256432e-01
-7.55231738e-01 -1.89394638e-01 -2.68116862e-01 5.29916644e-01
-6.82213545e-01 4.39564079e-01 -7.07879663e-01 -1.48063469e+00
-3.28463435e-01 -7.72906721e-01 1.11698139e+00 -1.17945462e-01
-2.52204627e-01 -1.25593352e+00 1.80591673e-01 -3.57497185e-02
2.38144428e-01 3.89121741e-01 9.51460302e-01 -7.54497170e-01
-3.68643343e-01 -3.73625100e-01 2.38874406e-01 4.09006178e-01
4.09299612e-01 2.09348708e-01 -8.98834050e-01 -5.12140989e-01
4.48869675e-01 2.95574665e-01 4.78344619e-01 7.05837965e-01
4.53972250e-01 -4.50949699e-01 -6.49109110e-02 4.03198242e-01
1.45664096e+00 2.23198429e-01 2.17197970e-01 2.43442550e-01
5.95287979e-01 5.72295606e-01 2.46540353e-01 5.70768774e-01
3.38148504e-01 3.46329629e-01 4.99346107e-02 1.99158788e-01
5.44459105e-01 -2.21381143e-01 7.09981740e-01 1.16363192e+00
-9.64381695e-02 4.78445599e-03 -9.28916872e-01 5.77234328e-01
-1.78973818e+00 -8.93619299e-01 -2.68635303e-01 2.52263927e+00
6.97261512e-01 4.54776175e-02 3.78169894e-01 3.42011899e-01
5.56938469e-01 2.44354695e-01 -7.05608606e-01 6.93374127e-02
-3.36951107e-01 4.70679626e-03 6.40919089e-01 4.87616509e-01
-1.31024933e+00 3.21570128e-01 6.40444088e+00 6.70877516e-01
-1.16122413e+00 -1.13262556e-01 4.06007290e-01 1.21347085e-01
-1.95017412e-01 6.80264458e-02 -7.32226193e-01 4.81486440e-01
1.34426618e+00 -5.30731976e-01 6.38157070e-01 8.58506918e-01
5.66085517e-01 2.24027663e-01 -1.05042696e+00 9.40478921e-01
-2.83363551e-01 -7.50601351e-01 -2.32468724e-01 2.34500945e-01
8.75987470e-01 -1.31698295e-01 2.52338290e-01 3.86863083e-01
4.07502025e-01 -7.38740504e-01 6.43859327e-01 7.55804837e-01
3.21265012e-01 -6.77625537e-01 4.72476810e-01 3.91120523e-01
-1.26825726e+00 -4.00595337e-01 -8.00365731e-02 -9.52054709e-02
1.18539318e-01 9.65903819e-01 -2.79166460e-01 7.94920802e-01
3.85525078e-01 7.22825050e-01 -3.04781944e-01 6.53228879e-01
7.05831572e-02 1.19607747e+00 -5.88067591e-01 4.59927559e-01
1.89535707e-01 -7.03492701e-01 8.87592137e-01 8.75211656e-01
6.15104198e-01 -6.85729235e-02 1.25579283e-01 4.75450397e-01
3.74850899e-01 1.58395395e-01 -3.44737679e-01 -3.76991928e-01
4.37091976e-01 9.44814205e-01 -5.81750393e-01 -4.78392988e-02
-5.63699841e-01 7.30766952e-01 -2.80491948e-01 8.31102133e-01
-5.37068784e-01 1.06119342e-01 7.31629014e-01 6.97405711e-02
5.30763507e-01 -4.63049680e-01 -2.98080504e-01 -1.20204651e+00
2.98626602e-01 -1.15432274e+00 5.23837149e-01 -4.60798830e-01
-1.55904746e+00 2.93789804e-01 -1.25361696e-01 -1.34798694e+00
-7.96933651e-01 -2.82208174e-01 -4.68477219e-01 1.02141976e+00
-1.33897913e+00 -1.05208528e+00 1.33470476e-01 5.10925770e-01
4.95507270e-01 -1.50556505e-01 6.50105417e-01 4.97969752e-03
-8.78647566e-01 2.65562147e-01 8.17040503e-01 -2.59357661e-01
5.37183404e-01 -1.26943135e+00 3.17947865e-01 1.22289026e+00
6.72096089e-02 8.83283317e-01 1.06415248e+00 -5.84132731e-01
-1.89596665e+00 -1.00249457e+00 6.94059491e-01 -4.81016010e-01
1.44604683e+00 -2.74109960e-01 -1.15114164e+00 9.00846243e-01
-3.76793295e-01 1.29838744e-02 5.54019511e-01 1.74301043e-01
-3.47351015e-01 -3.80690902e-01 -5.76409042e-01 6.31472468e-01
5.54398239e-01 -6.57955170e-01 -3.99739951e-01 1.26997903e-01
5.26797891e-01 -1.02594845e-01 -9.16343272e-01 4.21538323e-01
5.86345732e-01 -4.44865257e-01 1.01462901e+00 -6.58836842e-01
2.75218099e-01 -4.14498746e-01 -4.45485353e-01 -1.16100538e+00
-2.76056498e-01 -1.27504802e+00 -5.23516238e-01 1.35797489e+00
2.73834735e-01 -9.10865843e-01 2.24717051e-01 7.24820614e-01
1.46554664e-01 -4.08048898e-01 -9.27731693e-01 -1.21130359e+00
-4.01156815e-03 -5.76913536e-01 2.74479598e-01 9.48682487e-01
-5.63535750e-01 3.55177373e-01 -6.97011650e-01 7.59035528e-01
7.81255186e-01 3.18781883e-01 8.22461963e-01 -1.49370360e+00
-4.91608828e-01 -2.71563917e-01 -4.12479490e-02 -1.03897417e+00
2.36485481e-01 -3.99607778e-01 1.94611579e-01 -8.35212827e-01
-1.23698428e-01 -9.02202576e-02 -4.74381298e-01 -7.99838081e-03
-3.15810770e-01 -1.85870469e-01 3.28590795e-02 5.37179053e-01
-2.03369290e-01 5.81721723e-01 6.71838641e-01 1.74650550e-01
-3.62341344e-01 5.23064256e-01 -7.59001613e-01 7.67262042e-01
5.22550762e-01 -3.63565713e-01 -7.05725670e-01 1.66326240e-01
1.80634722e-01 3.50034982e-01 4.34410572e-01 -8.12086225e-01
2.08080247e-01 -1.31052002e-01 1.41817957e-01 -4.75076735e-01
1.26616761e-01 -9.93948758e-01 5.67903817e-01 2.62929112e-01
-2.20885590e-01 2.33101577e-01 2.15543196e-01 1.25202143e+00
-5.93054295e-02 -1.03201354e-02 3.96369815e-01 4.25780684e-01
-4.15957004e-01 1.95622623e-01 -5.66032529e-01 -9.93080661e-02
7.36297607e-01 -4.62678168e-03 1.04419798e-01 -8.11939240e-01
-6.54832482e-01 5.17782345e-02 9.35073793e-02 1.26579747e-01
7.24349811e-04 -1.05137062e+00 -5.38069963e-01 1.85740530e-01
-2.00340003e-01 -2.37837210e-01 4.09755856e-01 1.21063101e+00
1.34657741e-01 4.81450766e-01 5.19507229e-01 -7.16806769e-01
-9.03498054e-01 8.41132820e-01 8.67383853e-02 -3.83484364e-01
-4.62951750e-01 3.46857607e-02 3.04563880e-01 -7.18527809e-02
7.77571276e-02 -3.75823259e-01 6.46599978e-02 2.55968630e-01
5.45531094e-01 4.84009683e-01 8.64516050e-02 -6.44185901e-01
-2.52727568e-01 5.49144506e-01 2.43607640e-01 -1.73415959e-01
1.34163392e+00 -6.56727493e-01 -8.99503082e-02 1.20631683e+00
1.05119669e+00 2.46030778e-01 -1.44431174e+00 -4.50637996e-01
4.95337427e-01 -1.96913704e-02 8.73086080e-02 -3.71195465e-01
-7.74818122e-01 5.66409171e-01 6.00997329e-01 5.14817417e-01
1.32670510e+00 -4.68823791e-01 5.97822785e-01 3.85779142e-01
2.26869956e-01 -6.71547949e-01 -2.02387586e-01 5.34356236e-01
8.40391219e-01 -1.17808330e+00 -1.86248153e-01 -3.71813655e-01
-5.78596234e-01 1.04707813e+00 -2.69689292e-01 -2.02265173e-01
8.28050137e-01 2.62240767e-01 8.14528540e-02 6.86617717e-02
-9.72048223e-01 -8.26749429e-02 4.67381358e-01 2.65404552e-01
1.50017262e-01 -2.49566138e-02 -3.68276983e-02 8.89041126e-01
-1.32094711e-01 -1.12731665e-01 4.72875237e-01 8.02846551e-01
6.73847739e-03 -5.72379112e-01 -5.90062797e-01 3.21858495e-01
-7.09876955e-01 -1.86294079e-01 6.37099072e-02 7.49010205e-01
-9.02328908e-01 1.12626171e+00 -4.86405566e-02 -4.21550609e-02
3.65280628e-01 4.12484944e-01 -1.04391105e-01 -1.44329205e-01
-2.25731298e-01 7.12393582e-01 -2.39126012e-01 -2.88036823e-01
-3.71540219e-01 -1.18244374e+00 -5.93350291e-01 -2.53027260e-01
-1.26670465e-01 3.25660706e-02 5.47009766e-01 1.12042427e+00
3.95613283e-01 4.00309592e-01 1.02394927e+00 -5.65245092e-01
-1.17322779e+00 -7.04790533e-01 -7.76362300e-01 1.43716872e-01
7.87991345e-01 -5.36132038e-01 -9.89080429e-01 3.35216075e-01] | [6.901791095733643, 3.76599383354187] |
a6acd7e3-eea0-435e-95da-82cb7834d6f0 | importance-aware-learning-for-neural-headline | 1912.01114 | null | https://arxiv.org/abs/1912.01114v1 | https://arxiv.org/pdf/1912.01114v1.pdf | Importance-Aware Learning for Neural Headline Editing | Many social media news writers are not professionally trained. Therefore, social media platforms have to hire professional editors to adjust amateur headlines to attract more readers. We propose to automate this headline editing process through neural network models to provide more immediate writing support for these social media news writers. To train such a neural headline editing model, we collected a dataset which contains articles with original headlines and professionally edited headlines. However, it is expensive to collect a large number of professionally edited headlines. To solve this low-resource problem, we design an encoder-decoder model which leverages large scale pre-trained language models. We further improve the pre-trained model's quality by introducing a headline generation task as an intermediate task before the headline editing task. Also, we propose Self Importance-Aware (SIA) loss to address the different levels of editing in the dataset by down-weighting the importance of easily classified tokens and sentences. With the help of Pre-training, Adaptation, and SIA, the model learns to generate headlines in the professional editor's style. Experimental results show that our method significantly improves the quality of headline editing comparing against previous methods. | ['Ying Zeng', 'Lei LI', 'Qingyang Wu', 'Hao Zhou', 'Zhou Yu'] | 2019-11-25 | null | null | null | null | ['headline-generation'] | ['natural-language-processing'] | [ 2.27545306e-01 3.36704671e-01 -2.09968552e-01 -6.61315918e-01
-8.88026893e-01 -3.23055476e-01 7.41494954e-01 5.04237786e-02
-6.84379339e-01 7.34352887e-01 6.72152817e-01 -1.45462081e-01
3.10167074e-01 -6.90484107e-01 -1.00173342e+00 1.45414129e-01
8.09193373e-01 3.80851716e-01 -8.51962566e-02 -4.42867249e-01
4.43719506e-01 2.33677533e-02 -1.06530452e+00 5.12375355e-01
1.38610911e+00 6.94184065e-01 4.92954820e-01 8.08108628e-01
-2.89492697e-01 1.08851182e+00 -9.36987340e-01 -1.01991498e+00
3.68260033e-02 -6.36776745e-01 -3.77669334e-01 2.86026984e-01
7.85632968e-01 -4.70486671e-01 -3.77083600e-01 1.09608197e+00
5.80965996e-01 1.67733446e-01 5.23311853e-01 -7.89940596e-01
-1.43125248e+00 1.21376395e+00 -5.11799455e-01 2.17891946e-01
3.12136441e-01 4.20530252e-02 1.16382623e+00 -1.11144471e+00
6.50933027e-01 1.19592190e+00 6.87931538e-01 5.50240815e-01
-9.32211936e-01 -6.10103130e-01 4.07681793e-01 -8.56222808e-02
-8.21131945e-01 -7.79715478e-01 8.00263524e-01 -4.74805504e-01
7.46317983e-01 1.02424771e-01 4.67094362e-01 1.29697561e+00
1.16327696e-01 9.81137812e-01 7.24407375e-01 -5.26661396e-01
-2.83641130e-01 4.67012644e-01 8.79665390e-02 6.28258944e-01
-3.22226547e-02 -3.80730629e-01 -7.70636797e-01 3.47555548e-01
6.27305388e-01 5.36966585e-02 -2.90369183e-01 4.45902556e-01
-1.07535076e+00 7.76146412e-01 2.90737510e-01 7.52408430e-02
-3.61772984e-01 -7.71518648e-02 6.03357017e-01 4.90810752e-01
8.42412055e-01 9.05805230e-01 -1.75834909e-01 -3.14187169e-01
-1.14552188e+00 1.79203853e-01 9.21144903e-01 1.32212555e+00
5.14518082e-01 -1.90593615e-01 -7.08654761e-01 1.29590976e+00
1.19808987e-01 5.29242635e-01 5.08665740e-01 -7.14087486e-01
8.96909773e-01 6.22793496e-01 1.85000040e-02 -1.14599371e+00
1.07437655e-01 -8.80612075e-01 -8.58955681e-01 -4.75098640e-01
1.79689452e-01 -3.22952151e-01 -5.05880296e-01 1.38759673e+00
-1.51905462e-01 -3.60832214e-01 -3.03417265e-01 6.83659315e-01
5.67238510e-01 9.92399871e-01 -3.76753837e-01 -3.12553287e-01
1.10550714e+00 -1.74066389e+00 -1.14737797e+00 -3.78974527e-01
4.38158780e-01 -1.07727873e+00 1.64234543e+00 3.01291406e-01
-1.44745708e+00 -7.29494512e-01 -1.04751146e+00 -5.26894152e-01
2.47093011e-02 6.40605330e-01 1.29741922e-01 1.73478331e-02
-7.70887494e-01 6.60410881e-01 -4.98227328e-01 6.73881322e-02
6.40120387e-01 -2.40077108e-01 -1.43805705e-02 5.38815297e-02
-1.27900279e+00 6.87048733e-01 -1.88297391e-01 5.40133519e-03
-6.00659847e-01 -1.03568363e+00 -8.89649749e-01 5.22386171e-02
3.78592193e-01 -5.79260409e-01 1.72619998e+00 -1.22082651e+00
-1.73813772e+00 7.40073025e-01 -1.47688404e-01 -2.72542387e-01
1.02541220e+00 -6.65447056e-01 -5.76816797e-01 -1.47900894e-01
4.94396567e-01 2.96348453e-01 9.23349857e-01 -8.91655207e-01
-8.02473545e-01 7.89717734e-02 2.26995535e-03 1.66933313e-01
-8.81795049e-01 1.23902589e-01 -8.64422381e-01 -1.11676347e+00
-5.91431022e-01 -7.69655943e-01 -1.15117252e-01 -8.21419433e-02
-8.56665015e-01 -3.59222770e-01 1.29793331e-01 -1.00599086e+00
1.70455396e+00 -2.22229958e+00 1.11524835e-02 -1.12734921e-03
4.92100418e-01 1.46117210e-01 -2.27476344e-01 3.49177361e-01
5.40788949e-01 4.24875915e-02 -8.90943855e-02 -7.80662477e-01
2.01970771e-01 -2.32131720e-01 -4.26696867e-01 2.46319231e-02
2.24862695e-01 9.20552969e-01 -9.60589051e-01 -3.77379477e-01
-4.26731795e-01 3.09235126e-01 -8.90218914e-01 4.99399900e-01
-2.57649153e-01 1.31087571e-01 -3.27015221e-01 2.01764494e-01
3.21304709e-01 -6.62568867e-01 -2.75653720e-01 -7.44883493e-02
-3.21359187e-01 6.03656769e-01 -5.66346824e-01 1.60591555e+00
-9.60595608e-01 1.00025821e+00 -1.19604804e-01 -3.94831955e-01
9.45381761e-01 -7.33651628e-04 -4.19427827e-02 -8.59252989e-01
1.14601679e-01 1.69241086e-01 -1.34507015e-01 -5.31429291e-01
1.01834786e+00 1.29727898e-02 -2.55404055e-01 8.04700494e-01
-9.64372829e-02 2.50367582e-01 5.20153284e-01 6.24853194e-01
9.77978587e-01 -2.42219567e-01 -1.52754813e-01 9.03655887e-02
4.39967781e-01 -2.43791804e-01 5.12890041e-01 9.91903245e-01
-1.41195804e-02 6.96213782e-01 5.52210748e-01 -1.22619085e-01
-1.35190737e+00 -5.55306792e-01 6.43268824e-02 1.54894793e+00
-1.67165935e-01 -5.75487435e-01 -9.45184171e-01 -7.18469679e-01
-2.41559222e-02 8.47982168e-01 -5.71370006e-01 -3.18397999e-01
-7.70023346e-01 -2.30291426e-01 4.36511338e-01 4.57847118e-01
5.25999010e-01 -1.07798862e+00 -1.27195544e-03 5.07655740e-01
-2.01498002e-01 -1.04346871e+00 -1.42137218e+00 -2.71421254e-01
-2.45473847e-01 -5.14174461e-01 -1.02272010e+00 -8.57723713e-01
1.04231012e+00 4.17004526e-02 1.07519996e+00 8.71019214e-02
1.80026934e-01 -9.32036340e-03 -4.68699396e-01 -5.01532197e-01
-7.90866673e-01 6.30695164e-01 1.81773286e-02 1.58027932e-01
2.08790332e-01 -2.62316018e-01 -3.16288143e-01 1.98385902e-02
-7.51339078e-01 5.30148268e-01 7.21495330e-01 9.08440590e-01
2.59010792e-01 -3.23751718e-01 9.98081863e-01 -1.63323474e+00
1.24409771e+00 -3.08024794e-01 -3.81195724e-01 3.49133760e-01
-8.09192717e-01 5.71961291e-02 1.22574258e+00 -5.47277153e-01
-1.31613648e+00 -4.10426795e-01 -2.62662828e-01 -1.13706887e-01
5.58492422e-01 7.99339652e-01 -5.12840711e-02 4.08161908e-01
6.88735425e-01 2.23266304e-01 1.69197377e-02 -6.48520529e-01
4.45089459e-01 1.05983126e+00 7.07734287e-01 -1.90244228e-01
8.91621113e-01 -5.57914376e-02 -8.95694375e-01 -4.86395329e-01
-1.56114781e+00 -1.57111898e-01 -3.41828614e-01 -3.59440118e-01
4.77283150e-01 -1.04452729e+00 -2.67797112e-01 5.58165789e-01
-1.44955754e+00 -3.26548338e-01 -3.28727454e-01 2.32420921e-01
-1.93115637e-01 2.30666861e-01 -8.99470031e-01 -3.02694649e-01
-5.13755202e-01 -9.68342006e-01 6.80830896e-01 3.72838944e-01
-3.28534096e-01 -9.49286819e-01 1.03370380e-02 5.44847488e-01
6.45962477e-01 -2.47259930e-01 6.00574374e-01 -6.12898648e-01
-2.46135116e-01 -5.71986914e-01 -4.16134238e-01 7.53161609e-01
8.28620493e-02 1.53199479e-01 -6.37234628e-01 -9.44367051e-02
-2.48176128e-01 -3.71526569e-01 8.83545339e-01 1.18995339e-01
1.26933002e+00 -1.04153502e+00 -5.23972213e-02 6.60367608e-01
6.57377958e-01 -2.44640231e-01 2.90966392e-01 3.56119394e-01
9.94806945e-01 5.25947809e-01 5.22022903e-01 7.80560672e-01
7.13147759e-01 3.95288467e-01 -4.25386071e-01 -1.55259043e-01
-3.10969770e-01 -8.65882695e-01 6.72323525e-01 1.65762556e+00
2.65410841e-01 -5.04311085e-01 -4.25522417e-01 4.07709569e-01
-1.68834817e+00 -9.08780694e-01 1.22193962e-01 1.75064278e+00
1.53875315e+00 4.86451119e-01 -2.65934616e-02 -2.29450136e-01
9.01013851e-01 2.50899702e-01 -6.41567826e-01 -4.73142147e-01
-1.32183611e-01 -1.84378013e-01 5.19340217e-01 7.16375411e-01
-8.91207576e-01 1.00773609e+00 5.48643017e+00 6.94837153e-01
-1.22433710e+00 1.67507157e-01 5.44139564e-01 -3.00201237e-01
-6.00661278e-01 -2.88540095e-01 -1.29050207e+00 1.00714660e+00
9.00586307e-01 -6.05434179e-01 4.53371972e-01 9.23341811e-01
4.01230097e-01 5.49861014e-01 -1.18712652e+00 7.57212460e-01
5.35412967e-01 -1.51739419e+00 2.13032275e-01 -2.49676049e-01
1.06480789e+00 -1.64008260e-01 1.20903984e-01 6.99750662e-01
3.13768774e-01 -7.00138867e-01 1.01738906e+00 8.21211457e-01
9.17949080e-01 -5.29479623e-01 4.98347789e-01 4.84074652e-01
-4.94688034e-01 2.63071135e-02 -3.47693354e-01 -1.90457761e-01
6.35061383e-01 1.00950027e+00 -6.75774336e-01 -1.74088448e-01
1.50564432e-01 1.14946818e+00 -7.42745876e-01 7.34720469e-01
-5.00466943e-01 6.16421282e-01 1.17794842e-01 -3.20138752e-01
2.09695891e-01 1.64135862e-02 3.69257599e-01 1.39173329e+00
2.55089372e-01 -2.86712646e-01 4.81666662e-02 9.60852563e-01
-8.94554079e-01 1.93204582e-01 -1.72324523e-01 -6.12431407e-01
7.87285924e-01 1.20715141e+00 -9.00221914e-02 -6.11999094e-01
-5.26824415e-01 1.40931451e+00 8.03796291e-01 2.40217030e-01
-7.73948967e-01 -9.91850436e-01 1.45871416e-01 3.94854635e-01
-7.68792853e-02 5.03853187e-02 -3.19480896e-01 -1.33633161e+00
2.67129391e-01 -1.10988557e+00 -9.23937261e-02 -5.70434570e-01
-1.64301503e+00 7.14881718e-01 -7.29604602e-01 -1.01463985e+00
-9.33233276e-02 -2.29525328e-01 -5.48681080e-01 7.18207836e-01
-1.49946976e+00 -1.04047120e+00 -2.35509470e-01 4.80522886e-02
1.03934491e+00 -4.71536726e-01 5.71226217e-02 7.45336235e-01
-8.76393437e-01 1.24736583e+00 1.76211134e-01 2.77755499e-01
1.26862895e+00 -1.24562585e+00 7.10245490e-01 8.50575626e-01
-2.02780589e-01 8.20336461e-01 6.05787396e-01 -8.00032794e-01
-1.02059734e+00 -1.40291309e+00 1.62405205e+00 -4.96478796e-01
7.60062754e-01 -4.35656190e-01 -8.30387235e-01 7.78730989e-01
3.56875509e-01 -4.69189465e-01 8.52178931e-01 2.07887664e-01
-1.72901303e-01 -1.17633894e-01 -4.69325989e-01 7.62282610e-01
1.03376520e+00 -6.96177304e-01 -6.54716074e-01 7.06055701e-01
1.12128651e+00 -4.44181561e-01 -8.17481875e-01 -1.28911227e-01
3.58515948e-01 -4.49584275e-01 5.03697872e-01 -5.20363986e-01
1.38720179e+00 8.08065310e-02 5.80741227e-01 -1.60720694e+00
-4.67363477e-01 -7.96597958e-01 -1.04981907e-01 1.57459879e+00
7.83627093e-01 -2.89422274e-01 3.86581808e-01 7.35499322e-01
-7.03298450e-01 -6.71854258e-01 -2.94952095e-01 -5.27392685e-01
-1.71043798e-01 -7.98011944e-02 4.65814471e-01 1.02448094e+00
9.95224714e-02 9.19471502e-01 -7.35651970e-01 -5.05126476e-01
3.23379695e-01 -1.50997639e-01 9.38236296e-01 -9.40413475e-01
-3.77375215e-01 -6.76062942e-01 4.99037206e-01 -1.60908735e+00
3.98389816e-01 -1.15180874e+00 2.53013372e-01 -1.47612023e+00
2.45175585e-01 -5.01416326e-01 -4.78122234e-02 2.65394479e-01
-4.89052415e-01 2.90024914e-02 8.73759314e-02 3.34900588e-01
-9.51122224e-01 6.70228660e-01 1.77267694e+00 -8.63747969e-02
-2.78488517e-01 4.68053669e-02 -1.22975850e+00 5.13846099e-01
3.52905273e-01 -3.21750522e-01 -2.10946143e-01 -8.67666125e-01
7.98552990e-01 -1.66418329e-01 -3.80275063e-02 -7.21072078e-01
5.99209905e-01 -4.79818583e-02 2.73649484e-01 -3.43292475e-01
-1.39536530e-01 -2.84886032e-01 -4.18492585e-01 1.97534561e-02
-1.35166812e+00 2.16112033e-01 -3.27716470e-01 5.03522873e-01
-2.42284253e-01 -2.72597164e-01 7.89403021e-01 -6.39928207e-02
-1.26873046e-01 3.39579403e-01 -4.11359847e-01 4.28201109e-01
5.53200245e-01 1.45268485e-01 -4.16060060e-01 -5.95464528e-01
-3.07296455e-01 3.72641265e-01 5.30247688e-01 6.77094042e-01
4.82038021e-01 -1.30640817e+00 -9.54185843e-01 2.72117764e-01
2.71991938e-01 1.14889123e-01 2.44054914e-01 6.24548495e-01
-6.31816566e-01 1.75803363e-01 -7.76847824e-02 5.65451421e-02
-7.17628062e-01 2.88931787e-01 -5.34667037e-02 -2.47487471e-01
-7.14979053e-01 1.24353814e+00 -5.16561186e-03 -4.87639844e-01
4.05058652e-01 -1.71398088e-01 -2.75874496e-01 2.29519367e-01
8.94477129e-01 3.89556140e-01 -8.72716308e-02 -3.93824279e-01
2.52764136e-01 -5.11760525e-02 -8.64208281e-01 -1.32544339e-01
1.36842334e+00 -4.22488570e-01 -1.09789073e-01 5.47500193e-01
1.34285295e+00 4.84981388e-01 -1.38513315e+00 -5.60682297e-01
-3.81824165e-03 -4.57733154e-01 1.01091348e-01 -8.90087187e-01
-9.64428663e-01 7.60519147e-01 -2.47009233e-01 3.19733843e-02
7.73639262e-01 -1.47916168e-01 1.48249912e+00 4.76861209e-01
-4.08936173e-01 -1.79485345e+00 4.23459440e-01 8.66940081e-01
1.14410758e+00 -1.26080012e+00 -1.29467681e-01 -1.14497878e-01
-9.59077775e-01 9.71414924e-01 8.83666515e-01 -9.36544687e-02
3.49103481e-01 2.64862180e-01 2.04795986e-01 3.40536819e-03
-8.65380049e-01 5.23086190e-01 4.07503366e-01 9.54183787e-02
7.37772882e-01 -2.20360775e-02 -4.38214600e-01 1.17956662e+00
-7.33684599e-01 1.22525886e-01 8.67938697e-01 5.80718279e-01
-4.45308417e-01 -8.92898262e-01 8.71076211e-02 9.50914145e-01
-4.83243197e-01 -3.86753052e-01 -4.67107207e-01 -8.90297350e-03
-3.10499221e-01 6.09435558e-01 1.60767645e-01 -3.65245253e-01
5.02028227e-01 -2.44687721e-01 2.94471309e-02 -8.01658988e-01
-8.12893033e-01 -1.77118368e-03 2.51090288e-01 -7.22194687e-02
1.55803144e-01 -3.52699935e-01 -9.21945691e-01 -5.47547638e-01
-3.30567777e-01 1.73563346e-01 5.47577143e-01 8.18145633e-01
5.23769140e-01 8.51603746e-01 9.05923069e-01 -4.87138510e-01
-8.93746614e-01 -1.20536494e+00 -4.60067838e-01 7.00178683e-01
4.26583052e-01 -1.73446938e-01 -3.58853042e-01 3.36477995e-01] | [12.131884574890137, 9.15505313873291] |
0568b137-a85a-4199-9d18-c143145327f0 | lsdnet-trainable-modification-of-lsd | 2209.04642 | null | https://arxiv.org/abs/2209.04642v1 | https://arxiv.org/pdf/2209.04642v1.pdf | LSDNet: Trainable Modification of LSD Algorithm for Real-Time Line Segment Detection | As of today, the best accuracy in line segment detection (LSD) is achieved by algorithms based on convolutional neural networks - CNNs. Unfortunately, these methods utilize deep, heavy networks and are slower than traditional model-based detectors. In this paper we build an accurate yet fast CNN- based detector, LSDNet, by incorporating a lightweight CNN into a classical LSD detector. Specifically, we replace the first step of the original LSD algorithm - construction of line segments heatmap and tangent field from raw image gradients - with a lightweight CNN, which is able to calculate more complex and rich features. The second part of the LSD algorithm is used with only minor modifications. Compared with several modern line segment detectors on standard Wireframe dataset, the proposed LSDNet provides the highest speed (among CNN-based detectors) of 214 FPS with a competitive accuracy of 78 Fh . Although the best-reported accuracy is 83 Fh at 33 FPS, we speculate that the observed accuracy gap is caused by errors in annotations and the actual gap is significantly lower. We point out systematic inconsistencies in the annotations of popular line detection benchmarks - Wireframe and York Urban, carefully reannotate a subset of images and show that (i) existing detectors have improved quality on updated annotations without retraining, suggesting that new annotations correlate better with the notion of correct line segment detection; (ii) the gap between accuracies of our detector and others diminishes to negligible 0.2 Fh , with our method being the fastest. | ['Evgeny Shvets', 'Leonid Erlygin', 'Lev Teplyakov'] | 2022-09-10 | null | null | null | null | ['line-segment-detection', 'line-detection'] | ['computer-vision', 'computer-vision'] | [-2.41228208e-01 1.41148925e-01 -2.39459246e-01 -7.61796348e-03
-6.39746785e-01 -5.71394145e-01 3.23718607e-01 3.56180102e-01
-6.10424340e-01 4.65025097e-01 -1.82623655e-01 -4.31109577e-01
3.48296016e-01 -6.52251601e-01 -9.36178029e-01 -1.64284557e-01
-9.96007249e-02 5.17004356e-02 1.00996053e+00 -4.46799099e-01
2.41366133e-01 8.91684651e-01 -1.31626236e+00 1.84614420e-01
5.33335805e-01 1.25295544e+00 -3.64854395e-01 8.44954073e-01
7.73694217e-02 7.45649219e-01 -5.81832886e-01 -4.74497676e-01
5.87834001e-01 -1.60615802e-01 -7.05613852e-01 -1.17404960e-01
9.18390632e-01 -7.05067217e-01 -5.64444244e-01 5.56776941e-01
5.24673462e-01 -2.37017766e-01 3.55897397e-01 -1.12749982e+00
-8.45404342e-02 5.76223850e-01 -7.54247904e-01 2.80224144e-01
3.65961343e-01 1.00618474e-01 7.52441108e-01 -9.22757924e-01
7.87550569e-01 6.92915559e-01 1.61174047e+00 1.91409662e-01
-9.89474952e-01 -4.45998222e-01 -2.82531798e-01 -1.38261507e-03
-1.76514888e+00 -2.24535793e-01 4.23248887e-01 -4.95204717e-01
1.04197717e+00 -3.27815637e-02 9.57136750e-01 7.33540893e-01
1.47274926e-01 7.84618676e-01 5.35937428e-01 -5.11387587e-01
-9.91172343e-02 7.72451162e-02 5.63970506e-02 9.07207370e-01
4.30046350e-01 3.72660868e-02 -2.81991243e-01 1.83618888e-01
1.08040881e+00 -3.25169802e-01 -2.19550952e-01 -4.44214523e-01
-1.32109487e+00 7.03693986e-01 6.37779832e-01 2.86895573e-01
-1.53152496e-01 3.72810781e-01 6.93504155e-01 7.61233792e-02
2.77211398e-01 2.81273693e-01 -4.86981064e-01 -3.26663047e-01
-1.44070911e+00 2.87516922e-01 7.46854961e-01 9.83002126e-01
8.16758633e-01 -2.53282934e-02 -1.40044838e-01 4.34178472e-01
-3.49181630e-02 1.76985502e-01 2.27013484e-01 -7.45670617e-01
3.85927409e-01 4.56486017e-01 1.46855593e-01 -1.13393986e+00
-8.67232144e-01 -7.94120908e-01 -4.08900023e-01 3.69493127e-01
8.81602526e-01 -3.34657788e-01 -6.92188084e-01 1.06756854e+00
1.84009880e-01 -5.17624505e-02 -3.77624243e-01 8.94800127e-01
6.29328370e-01 2.80885130e-01 -1.23571523e-01 3.87286842e-01
1.15980446e+00 -1.16185188e+00 -2.29046956e-01 -4.71654785e-04
1.22946143e+00 -9.60149944e-01 7.92420447e-01 4.59924787e-01
-9.57715750e-01 -7.47348845e-01 -1.56777525e+00 -2.32896909e-01
-2.15384305e-01 4.33692843e-01 7.61606634e-01 7.63236403e-01
-1.16980636e+00 7.82192409e-01 -6.74276948e-01 -6.42622888e-01
4.90072399e-01 3.84130388e-01 -4.88194346e-01 3.18075269e-01
-9.00191009e-01 6.88276052e-01 3.75502050e-01 8.85956585e-02
-2.52334893e-01 -7.92917371e-01 -7.34405637e-01 6.20256625e-02
3.54788750e-01 -5.10431349e-01 1.36530268e+00 -8.86199951e-01
-1.45019329e+00 8.48422825e-01 2.07389683e-01 -9.39724803e-01
1.10339296e+00 -5.97143233e-01 -2.72727013e-01 2.86282927e-01
-6.04991652e-02 8.52931499e-01 4.92368609e-01 -9.70261455e-01
-9.95019019e-01 2.66376346e-01 1.34971857e-01 -2.61710107e-01
-5.91918267e-02 -1.49785370e-01 -5.17528951e-01 -3.68658930e-01
8.76411870e-02 -8.81651163e-01 -7.16258585e-02 4.75167722e-01
-6.20553493e-01 -1.62957221e-01 9.11414266e-01 -4.10808235e-01
1.31562066e+00 -2.00961423e+00 -8.35005879e-01 2.51041114e-01
3.40124428e-01 5.64547360e-01 1.38278827e-01 5.48888505e-01
-2.48525947e-01 1.01665229e-01 1.20770663e-01 -3.81873995e-01
-2.72086620e-01 -2.42721066e-01 -8.64795521e-02 7.93314934e-01
2.22431332e-01 8.81545842e-01 -7.34353840e-01 -4.90134120e-01
4.53844219e-01 4.33461994e-01 -5.21383524e-01 -2.67753661e-01
2.67941982e-01 -8.93891265e-04 -7.21596554e-03 6.51436925e-01
7.81062424e-01 -1.68126777e-01 -3.00022185e-01 -6.40102804e-01
-8.09096277e-01 5.90748414e-02 -1.23845923e+00 1.88319314e+00
-6.51239604e-02 1.34250605e+00 -4.98095632e-01 -6.85616970e-01
1.08266199e+00 5.41483313e-02 7.02508986e-01 -5.61227441e-01
2.40173489e-01 3.29118669e-01 3.24789658e-02 -2.59020329e-01
7.49709129e-01 5.33387482e-01 1.33764446e-01 9.03959051e-02
2.12999284e-02 -1.07953764e-01 3.54817927e-01 2.24826857e-01
1.27782094e+00 3.72546464e-01 2.34418347e-01 -2.57715702e-01
2.44098619e-01 3.56369734e-01 3.71038705e-01 9.36815858e-01
-4.70178127e-01 1.03867221e+00 6.46919250e-01 -9.49217200e-01
-1.47644162e+00 -8.62482905e-01 -2.00392500e-01 7.00814545e-01
2.40240127e-01 -7.44021118e-01 -9.23994958e-01 -6.26399517e-01
-1.31757064e-02 1.14247635e-01 -6.27350807e-01 1.25281408e-01
-7.08134294e-01 -3.75458777e-01 1.15572703e+00 8.82613182e-01
8.48819554e-01 -4.44334090e-01 -9.20693278e-01 2.90644258e-01
1.71971112e-01 -1.42400432e+00 -3.12111259e-01 1.39017195e-01
-6.55329525e-01 -1.18877661e+00 -7.19523013e-01 -6.47652745e-01
4.33564842e-01 2.84114659e-01 1.05154717e+00 3.81590188e-01
-4.37701702e-01 6.97198585e-02 -4.42004293e-01 -3.73955190e-01
-8.92106369e-02 4.39687252e-01 -2.38695189e-01 -3.75288576e-01
1.26534775e-01 -3.25109176e-02 -9.54407573e-01 2.14503706e-01
-5.33279598e-01 2.15146139e-01 6.05197728e-01 5.77751994e-01
1.88795656e-01 -4.41024482e-01 5.89982755e-02 -8.03261817e-01
3.01432341e-01 -1.87268872e-02 -7.37801850e-01 -1.45523503e-01
-6.67710543e-01 -2.44697466e-01 4.44466054e-01 -2.04411507e-01
-6.39731169e-01 4.45681185e-01 -4.40655380e-01 -3.27237040e-01
-2.19318911e-01 1.70835346e-01 6.64116919e-01 -5.06664097e-01
1.03726530e+00 -2.35840932e-01 -2.37104565e-01 -2.05031425e-01
2.60905653e-01 1.98631331e-01 8.57388973e-01 -1.32680908e-01
7.96316922e-01 7.02865601e-01 3.88332307e-02 -7.87473440e-01
-8.11536849e-01 -6.63123190e-01 -1.01453161e+00 -3.68347526e-01
7.78674662e-01 -7.95489192e-01 -7.76803732e-01 5.73296905e-01
-1.26992941e+00 -4.40271854e-01 -4.23370987e-01 4.12867963e-01
-4.67235059e-01 5.46810269e-01 -9.16541636e-01 -5.37659824e-01
-2.56653547e-01 -9.82392490e-01 1.11493874e+00 3.78102839e-01
-2.45605782e-01 -8.30937684e-01 1.86737329e-01 -5.50097376e-02
4.49453294e-01 5.77556789e-01 3.21065605e-01 -5.47071993e-01
-3.63511741e-01 -6.53504372e-01 -6.74198031e-01 2.64674157e-01
-2.59377241e-01 6.67335093e-01 -9.78335381e-01 -5.69017045e-02
-7.15427697e-01 -1.34911388e-01 7.47859001e-01 4.29007292e-01
9.32505727e-01 2.47993901e-01 -4.36039597e-01 8.25650156e-01
1.59881818e+00 2.10415646e-02 9.15047824e-01 8.90827477e-01
6.97395384e-01 2.07535744e-01 4.81191635e-01 5.52762806e-01
3.98385704e-01 7.79048324e-01 2.73513079e-01 -8.12396586e-01
-4.11917239e-01 -3.01261038e-01 9.57995504e-02 4.30978715e-01
-3.70928675e-01 -2.78250277e-01 -1.04721582e+00 3.43223512e-01
-1.91241336e+00 -8.67405653e-01 -8.04576695e-01 2.18649602e+00
3.35512757e-01 6.82831585e-01 4.46425825e-01 3.09210479e-01
6.01647854e-01 -1.34820461e-01 -2.03763008e-01 -4.64331508e-01
-2.31251121e-01 1.63957104e-01 1.17174470e+00 4.70614359e-02
-1.37644076e+00 1.06767166e+00 7.32303524e+00 6.65686727e-01
-1.45106721e+00 -8.70827064e-02 4.64053184e-01 2.45506957e-01
4.44786489e-01 1.07639723e-01 -8.65742683e-01 1.87119916e-01
7.53827572e-01 1.38288036e-01 -3.80892843e-01 1.17855537e+00
3.14466149e-01 -3.12325478e-01 -1.00350702e+00 1.10264337e+00
3.04187536e-02 -1.71655273e+00 -4.11972880e-01 -7.11608157e-02
7.01440215e-01 2.63684869e-01 -2.84706950e-01 1.70271561e-01
-1.19320713e-01 -7.21393347e-01 1.07235670e+00 5.35444736e-01
7.35180378e-01 -8.54180455e-01 9.66160953e-01 -1.34661905e-02
-1.42600560e+00 1.91999048e-01 -3.68166387e-01 -8.87967646e-02
2.54388899e-01 5.25793314e-01 -1.11569977e+00 3.95918369e-01
7.82837868e-01 6.96900129e-01 -8.77700806e-01 1.56804514e+00
-6.29253732e-03 5.78284383e-01 -5.29145300e-01 1.75940797e-01
4.95023310e-01 2.39992395e-01 1.52675942e-01 1.77485538e+00
2.49628112e-01 -4.78983492e-01 1.81786314e-01 5.92070818e-01
7.44629130e-02 1.97719008e-01 -4.64331090e-01 3.98905009e-01
3.21384937e-01 1.45196342e+00 -1.16020477e+00 -4.14316714e-01
-6.53832316e-01 1.00778365e+00 3.25386435e-01 7.10103735e-02
-1.19045770e+00 -7.60648131e-01 3.65591705e-01 4.38661009e-01
4.04259592e-01 -5.69921017e-01 -3.62802416e-01 -8.65141392e-01
-1.27972469e-01 -2.74960220e-01 9.22036916e-02 -6.73634589e-01
-6.66583300e-01 6.19798481e-01 -2.34712601e-01 -1.51070642e+00
-2.92782396e-01 -6.06106222e-01 -5.76303124e-01 4.00553137e-01
-1.44846749e+00 -1.16962397e+00 -6.91613495e-01 4.00550365e-01
5.59716105e-01 2.56949872e-01 3.47756296e-01 4.85240787e-01
-5.54615080e-01 9.88411546e-01 -1.15761440e-02 7.01445818e-01
7.80501068e-01 -1.07145035e+00 8.08016300e-01 9.02523875e-01
-1.57474037e-02 2.34255359e-01 7.28469610e-01 -5.15699506e-01
-8.57336700e-01 -9.28739369e-01 6.39525235e-01 -2.65712649e-01
6.98443711e-01 -3.15943658e-01 -7.98714459e-01 6.76614165e-01
6.73953146e-02 3.40035915e-01 2.66668558e-01 -2.80241668e-01
-3.28744680e-01 3.27937230e-02 -9.35847819e-01 5.20294189e-01
1.08815849e+00 -2.71845073e-01 -8.17418769e-02 3.05223674e-01
6.14696801e-01 -6.98172510e-01 -8.22946191e-01 3.99405181e-01
9.69092011e-01 -1.57165444e+00 9.61548686e-01 -6.26817197e-02
2.07778305e-01 -3.89412642e-01 3.24574739e-01 -8.29021156e-01
-3.71723264e-01 -5.06511688e-01 1.24057949e-01 1.01461256e+00
4.09889370e-01 -5.48574507e-01 1.03632569e+00 3.57282698e-01
-5.18742740e-01 -8.99946272e-01 -7.46185362e-01 -8.27718794e-01
-2.03636467e-01 -6.74232781e-01 3.51047605e-01 8.23086917e-01
-1.88752815e-01 -1.50584638e-01 -2.69652247e-01 -7.76073411e-02
3.29392135e-01 -3.73878419e-01 1.14510572e+00 -1.04929018e+00
9.08067748e-02 -5.60204983e-01 -8.96872044e-01 -1.22171032e+00
-3.12996179e-01 -3.92782122e-01 2.88069192e-02 -1.30676532e+00
-2.06286848e-01 -5.77880383e-01 1.01882815e-01 5.43131053e-01
2.76705503e-01 7.31360793e-01 2.18231589e-01 2.49913812e-01
-6.85355544e-01 1.23222983e-02 8.38848472e-01 2.13344082e-01
-1.86510652e-01 -2.43123651e-01 -7.62384087e-02 1.01666927e+00
6.84340596e-01 -2.50498503e-01 1.21360704e-01 -4.07945335e-01
3.21955413e-01 -3.36120814e-01 4.96812463e-01 -1.67281675e+00
5.05957782e-01 4.37429845e-01 6.93426192e-01 -9.09406841e-01
7.54038021e-02 -6.22558892e-01 -8.98174047e-02 7.47142136e-01
-1.08661637e-01 3.40429574e-01 3.39937657e-01 1.17878847e-01
-7.30396584e-02 -2.39560157e-01 6.57239974e-01 1.13449454e-01
-1.14150012e+00 2.41267130e-01 -4.85004425e-01 -2.55707353e-01
1.05709088e+00 -8.59905839e-01 -4.91319239e-01 -3.19813997e-01
-4.34155434e-01 -2.75483876e-01 7.06867754e-01 3.35630715e-01
4.13907111e-01 -1.20673835e+00 -5.59014797e-01 4.68859971e-02
1.76828444e-01 -2.23552361e-02 5.06671257e-02 1.12877321e+00
-1.53459167e+00 5.77116668e-01 -1.85640529e-01 -7.89379954e-01
-8.50773990e-01 2.41810098e-01 5.29339135e-01 -7.72134811e-02
-9.55813646e-01 7.67608464e-01 -9.37686339e-02 2.10131168e-01
7.77717829e-02 -6.85329139e-01 1.72566599e-03 -5.84970340e-02
3.37983519e-01 5.95915794e-01 4.61533666e-01 -6.36282086e-01
-3.10896486e-01 8.26391101e-01 -4.84881289e-02 2.20931754e-01
1.03745842e+00 3.93138006e-02 4.12713170e-01 2.55188912e-01
1.13942969e+00 2.15293199e-01 -1.34698296e+00 1.63687333e-01
6.03877082e-02 -4.04923141e-01 9.09996312e-03 -3.09337914e-01
-1.00433469e+00 5.93065083e-01 8.17155898e-01 2.00889379e-01
8.01805675e-01 -1.32106036e-01 1.09665871e+00 2.19390199e-01
4.32230532e-01 -1.18630457e+00 -9.14224833e-02 5.06023765e-01
5.77415586e-01 -1.19292128e+00 1.14959739e-01 -5.93061149e-01
-1.21152364e-01 1.71252358e+00 7.42227554e-01 -5.49579322e-01
3.62172306e-01 5.47309756e-01 1.59654260e-01 -1.70272753e-01
-3.45767349e-01 -2.02866957e-01 3.11237812e-01 5.11472762e-01
5.88298321e-01 -2.44604826e-01 -3.80030304e-01 8.39606524e-02
-3.92186612e-01 7.41299689e-02 6.24355197e-01 1.01241434e+00
-4.54932690e-01 -7.91858137e-01 -3.82343829e-01 2.89550483e-01
-2.86756873e-01 2.35043224e-02 -1.37283176e-01 1.44617498e+00
3.10440719e-01 6.84360206e-01 3.47858906e-01 -5.81336975e-01
6.50544643e-01 -3.65098923e-01 2.61195928e-01 -2.44617492e-01
-7.73449659e-01 -7.44239315e-02 1.20611563e-01 -9.95495021e-01
-3.18863422e-01 -4.65936601e-01 -1.31210661e+00 -5.09466767e-01
-6.10859454e-01 -2.37437427e-01 7.89007664e-01 9.15141940e-01
2.44226143e-01 4.03660655e-01 3.14008445e-01 -1.10283017e+00
8.25589988e-03 -6.07750595e-01 -4.17806447e-01 2.88036346e-01
2.80146927e-01 -5.85497856e-01 -2.61093110e-01 5.07795997e-02] | [8.305620193481445, -1.6451045274734497] |
2c8d7ff2-5ba7-4669-b5f6-8ab49bef141e | transferring-deep-reinforcement-learning-with | 1809.00770 | null | http://arxiv.org/abs/1809.00770v1 | http://arxiv.org/pdf/1809.00770v1.pdf | Transferring Deep Reinforcement Learning with Adversarial Objective and Augmentation | In the past few years, deep reinforcement learning has been proven to solve
problems which have complex states like video games or board games. The next
step of intelligent agents would be able to generalize between tasks, and using
prior experience to pick up new skills more quickly. However, most
reinforcement learning algorithms for now are often suffering from catastrophic
forgetting even when facing a very similar target task. Our approach enables
the agents to generalize knowledge from a single source task, and boost the
learning progress with a semisupervised learning method when facing a new task.
We evaluate this approach on Atari games, which is a popular reinforcement
learning benchmark, and show that it outperforms common baselines based on
pre-training and fine-tuning. | ['Bing-Yu Chen', 'I-Chao Shen', 'Shu-Hsuan Hsu'] | 2018-09-04 | null | null | null | null | ['board-games'] | ['playing-games'] | [ 7.03694150e-02 -1.50168180e-01 -1.68573067e-01 3.02954484e-02
-5.05520225e-01 -7.26265728e-01 7.08389819e-01 2.67871097e-02
-9.26150203e-01 1.37745500e+00 -1.09962314e-01 -1.61103681e-01
-1.98094770e-01 -8.28389347e-01 -7.24130988e-01 -5.72521329e-01
-2.41950065e-01 8.68414044e-01 6.56209290e-01 -6.48311555e-01
1.76262751e-01 9.93739963e-02 -1.51131678e+00 2.21150648e-02
1.03119326e+00 7.10665882e-01 5.37756860e-01 5.70062637e-01
1.95727721e-01 9.56915736e-01 -6.30232394e-01 -2.52459824e-01
4.20535088e-01 -3.42646539e-01 -1.07373095e+00 -2.36385372e-02
3.26364487e-02 -4.36492026e-01 -4.82464433e-01 1.01754522e+00
2.77268440e-01 7.16398478e-01 4.52377200e-01 -1.13321280e+00
-3.52101535e-01 5.96515894e-01 -3.51411253e-01 5.34382701e-01
4.05552030e-01 6.48249567e-01 7.11943209e-01 -3.74224037e-01
5.70594490e-01 1.15127409e+00 4.94957805e-01 8.98933291e-01
-1.07835150e+00 -7.56123304e-01 2.64905930e-01 3.60553503e-01
-8.49425614e-01 6.47585746e-03 4.97698903e-01 -1.31654873e-01
1.14631414e+00 -4.20991808e-01 7.30017781e-01 1.33979940e+00
3.17179799e-01 9.97514486e-01 1.49496198e+00 -8.12304169e-02
7.11443663e-01 -2.63386369e-01 -3.24813128e-01 6.53189540e-01
1.66171834e-01 6.99709237e-01 -5.29304743e-01 3.38074267e-02
6.89390421e-01 6.61992207e-02 -1.49173997e-02 -6.61515772e-01
-9.91514444e-01 8.17714453e-01 3.86021286e-01 1.58537596e-01
-6.14059985e-01 2.57417291e-01 5.09856880e-01 9.02849734e-01
1.85114995e-01 1.00075483e+00 -7.17869520e-01 -5.15675962e-01
-6.53656423e-01 6.12762749e-01 7.47824550e-01 5.61328471e-01
9.22185063e-01 3.46790463e-01 -1.49907038e-01 6.09947026e-01
-2.29404002e-01 4.30605829e-01 1.00994766e+00 -1.11742735e+00
3.73600006e-01 3.30001533e-01 3.00775886e-01 -2.69581884e-01
-5.08903921e-01 -4.16475028e-01 -5.52869141e-01 8.17272782e-01
4.44036126e-01 -6.42149270e-01 -1.15816247e+00 1.94773996e+00
1.70225307e-01 3.74903679e-01 4.29981381e-01 8.48645151e-01
3.37833911e-01 5.94552696e-01 1.85262978e-01 -2.02288717e-01
8.37152183e-01 -1.11474597e+00 -3.95068973e-01 -7.12222219e-01
5.55145144e-01 -6.99077472e-02 1.16214204e+00 8.63563657e-01
-1.16772318e+00 -6.54753506e-01 -1.03758132e+00 4.26046371e-01
-2.98461437e-01 -7.27834463e-01 8.96935999e-01 4.69991803e-01
-1.06809592e+00 9.42492366e-01 -8.62583876e-01 -3.67223710e-01
5.14443696e-01 4.97551769e-01 -2.75994152e-01 -5.15111610e-02
-1.37456250e+00 1.20138085e+00 9.35764968e-01 -5.32180488e-01
-1.75405288e+00 -6.42968178e-01 -7.08952904e-01 1.64837286e-01
9.78552520e-01 -9.14931834e-01 1.72250938e+00 -1.16083920e+00
-2.12856865e+00 4.49138910e-01 4.36298251e-01 -7.66292274e-01
4.92816925e-01 -3.49752843e-01 -2.29398996e-01 -1.26134276e-01
8.39823782e-02 7.95017898e-01 1.14852631e+00 -9.26584244e-01
-1.08793390e+00 -2.63013661e-01 5.91533005e-01 7.09571302e-01
-1.87326834e-01 -3.63186210e-01 1.88344587e-02 -4.99857306e-01
-3.90430719e-01 -1.09951508e+00 -4.89733815e-01 -6.21741235e-01
3.06490034e-01 -4.93619323e-01 5.31852007e-01 -3.40800643e-01
6.49759352e-01 -1.90842390e+00 5.46605945e-01 -1.17514543e-01
1.74037695e-01 5.41811585e-01 -3.68791312e-01 2.59631217e-01
-6.00757189e-02 -4.25423592e-01 4.13830727e-02 6.26824424e-02
-1.58407316e-01 4.33688164e-01 -2.06057474e-01 -9.59909800e-03
9.32397321e-02 9.86370504e-01 -1.52296686e+00 1.38489574e-01
-4.17633131e-02 -3.01751971e-01 -6.62874997e-01 2.28357479e-01
-6.35196209e-01 5.81480563e-01 -2.51811445e-01 2.18327835e-01
1.77219585e-01 -2.11677849e-01 2.19549149e-01 6.74224555e-01
2.98625797e-01 3.37037295e-01 -1.12305510e+00 2.14880967e+00
-5.68318188e-01 3.18372726e-01 -3.10353249e-01 -1.38221300e+00
7.68770397e-01 2.15219826e-01 3.26546997e-01 -9.96688843e-01
1.46772072e-01 2.23034043e-02 3.55723023e-01 -3.92671108e-01
4.35308933e-01 -2.96857268e-01 -2.08398849e-01 7.05359578e-01
4.33827162e-01 -5.45812845e-01 6.69997096e-01 1.80966258e-01
1.42703867e+00 3.76331925e-01 3.14736247e-01 -1.08109243e-01
2.40092337e-01 3.31753492e-01 6.06237471e-01 1.21494806e+00
-2.68858939e-01 1.15804762e-01 2.95190752e-01 -8.76365602e-01
-8.76243711e-01 -1.41574204e+00 5.93851566e-01 1.40127838e+00
3.70392799e-02 -1.36961117e-01 -5.15127301e-01 -1.10958958e+00
9.10629183e-02 7.83048093e-01 -6.80284917e-01 -7.55933106e-01
-4.68060702e-01 -6.53735399e-01 1.65231958e-01 6.20583296e-01
7.75609255e-01 -1.59995651e+00 -9.66137648e-01 4.47329938e-01
1.28973514e-01 -7.92912066e-01 -6.52401447e-02 5.97067237e-01
-9.97115433e-01 -1.02181196e+00 -7.74816096e-01 -8.32009673e-01
1.98960170e-01 3.22997928e-01 1.35791874e+00 9.76164043e-02
-5.02220392e-02 4.65832978e-01 -3.28114957e-01 -3.51879328e-01
-5.50306141e-01 3.46289515e-01 6.69195116e-01 -6.10664368e-01
2.60484844e-01 -7.39436448e-01 -4.40853745e-01 6.15607649e-02
-8.64114761e-01 -1.86659768e-02 6.62732363e-01 1.26444376e+00
1.92248568e-01 4.25992370e-01 8.57170522e-01 -8.84947717e-01
1.07476640e+00 -5.97932279e-01 -7.24755466e-01 1.19027376e-01
-7.06690490e-01 4.35274720e-01 8.06214750e-01 -7.79226780e-01
-1.16833735e+00 -2.72413343e-02 1.10167228e-01 -2.26709291e-01
-2.60546535e-01 6.40395880e-01 2.73986191e-01 1.19220823e-01
1.03900695e+00 5.21845162e-01 1.08866453e-01 -8.59680027e-02
3.32642406e-01 8.08602720e-02 5.10074377e-01 -8.59031558e-01
9.13992047e-01 1.63673535e-01 -2.08721340e-01 -3.89731020e-01
-1.03724992e+00 -1.46646082e-01 -2.96945304e-01 -1.49340227e-01
4.87134725e-01 -9.45177794e-01 -8.81389022e-01 6.45867229e-01
-6.42869413e-01 -1.12819767e+00 -6.26704872e-01 4.50113714e-01
-9.65664983e-01 7.73612559e-02 -5.83475530e-01 -3.09263796e-01
3.13927568e-02 -9.81480718e-01 3.93792957e-01 7.80421257e-01
-9.88327637e-02 -1.07179022e+00 5.74021995e-01 9.94430408e-02
4.94085431e-01 -2.35574409e-01 6.85749471e-01 -6.23529553e-01
-1.91072166e-01 1.94669768e-01 1.06981851e-01 2.55875707e-01
2.45400906e-01 -6.66536093e-01 -6.52881205e-01 -8.11081886e-01
-3.85731943e-02 -1.15718234e+00 1.11160135e+00 1.27131179e-01
9.57112849e-01 -1.38761207e-01 -1.45798296e-01 3.34810406e-01
1.19619131e+00 3.88877034e-01 6.47640944e-01 7.21400440e-01
1.46007061e-01 3.36377397e-02 8.71323884e-01 4.50066507e-01
3.61573100e-01 4.25449401e-01 5.88355362e-01 1.60426989e-01
-3.75045128e-02 -3.41480851e-01 4.79047447e-01 3.53893548e-01
-2.77352631e-01 1.25389129e-01 -9.54275787e-01 5.06666183e-01
-1.99433267e+00 -1.08577681e+00 8.76437843e-01 2.04050326e+00
1.07900381e+00 6.20753527e-01 4.46728021e-01 -9.01186094e-02
2.20096558e-01 -8.03416967e-02 -1.18141305e+00 -3.32747102e-01
9.72354189e-02 7.43021131e-01 2.92913556e-01 3.77172202e-01
-1.03518534e+00 1.55922341e+00 7.02816391e+00 8.30540955e-01
-1.03615534e+00 1.04029357e-01 4.80885148e-01 -1.82029709e-01
2.29621857e-01 -1.34426534e-01 -5.50330043e-01 3.00386667e-01
7.24256814e-01 -4.09535557e-01 1.06825089e+00 9.32351828e-01
-1.29969209e-01 -5.18936038e-01 -1.02588713e+00 8.23905945e-01
-1.32212594e-01 -1.21411407e+00 -5.03611825e-02 -3.45447004e-01
1.18531775e+00 2.10142747e-01 3.86455536e-01 1.20477152e+00
1.51150405e+00 -1.12963033e+00 2.61610240e-01 2.99491346e-01
3.61665636e-01 -9.05666113e-01 5.88360786e-01 7.00655580e-01
-6.78914547e-01 -4.90916997e-01 -5.07268429e-01 -7.70644009e-01
-2.59113342e-01 -1.47346675e-01 -1.07360923e+00 3.25384945e-01
8.24652970e-01 5.78853846e-01 -5.54737747e-01 1.13864076e+00
-6.80029809e-01 4.59913373e-01 -1.29702359e-01 -3.43712181e-01
5.79542935e-01 -2.43575498e-02 3.01777214e-01 5.41775942e-01
2.65077710e-01 1.89594403e-01 5.00281870e-01 4.99377608e-01
-4.24555093e-02 -3.01741302e-01 -7.66792715e-01 2.87153795e-02
2.15028226e-01 1.03354192e+00 -7.59487391e-01 -3.74649882e-01
-2.49946490e-01 1.24823976e+00 8.93296242e-01 4.80853617e-01
-6.93835795e-01 -1.88446999e-01 7.12641060e-01 -2.49646932e-01
4.03003097e-01 -3.44163895e-01 1.17580570e-01 -1.19139755e+00
-3.48519295e-01 -1.25618756e+00 5.77594519e-01 -7.85856009e-01
-1.13400567e+00 3.99483323e-01 -2.20308393e-01 -1.02286017e+00
-7.01447725e-01 -5.85546732e-01 -6.31527960e-01 4.32193995e-01
-1.53800285e+00 -4.33181286e-01 -7.50678405e-02 9.33383822e-01
8.39782298e-01 -7.15001762e-01 7.78994620e-01 -2.11018026e-01
-3.18648130e-01 4.07129884e-01 2.98587263e-01 -8.94652903e-02
7.03102410e-01 -1.48716843e+00 4.08622861e-01 6.08511686e-01
3.53887618e-01 1.54922485e-01 8.34416151e-01 -6.03240252e-01
-1.27728891e+00 -7.04721749e-01 6.41342551e-02 -3.32545400e-01
7.85559237e-01 -2.33234882e-01 -8.04311037e-01 6.08219266e-01
3.75813067e-01 -2.53499985e-01 1.77628309e-01 4.17527586e-01
-2.72728413e-01 -1.23434409e-01 -9.26973760e-01 7.76892006e-01
1.11033070e+00 -3.24544013e-01 -1.04699719e+00 4.14564192e-01
5.43755770e-01 -6.49423242e-01 -3.44017386e-01 1.89937562e-01
1.76415235e-01 -9.11311448e-01 8.42657983e-01 -1.30897009e+00
4.82495666e-01 -1.63959891e-01 4.30302620e-01 -2.34674454e+00
-5.84195435e-01 -8.52175474e-01 -1.09639302e-01 4.99819428e-01
4.02032226e-01 -4.09822494e-01 1.14624679e+00 2.51699090e-01
9.81856417e-03 -3.33873153e-01 -7.80816734e-01 -1.20859969e+00
3.69889140e-01 -1.64844751e-01 3.98476809e-01 8.43585432e-01
4.41123217e-01 6.08749151e-01 -4.80347961e-01 -5.33352457e-02
4.29951221e-01 1.70975477e-01 9.82292593e-01 -1.31295478e+00
-6.55376434e-01 -5.06010532e-01 -2.41541922e-01 -1.14516187e+00
3.65670621e-01 -7.40097284e-01 2.76880264e-01 -1.31815541e+00
2.35788450e-01 -3.70371819e-01 -6.39313698e-01 7.10375071e-01
-4.32550758e-01 5.04048727e-02 1.63313732e-01 -2.19580874e-01
-1.23564231e+00 6.82069480e-01 1.52757871e+00 -3.17414343e-01
-4.67504442e-01 2.58354783e-01 -7.07075179e-01 7.82486081e-01
1.15915680e+00 -5.83837926e-01 -6.85172141e-01 -3.46612930e-01
6.00282013e-01 1.96235582e-01 5.92813306e-02 -1.39017129e+00
3.17229152e-01 -4.80294019e-01 4.88270909e-01 2.48207189e-02
2.68404156e-01 -6.29330754e-01 -2.36781865e-01 1.01885498e+00
-2.48796254e-01 1.41148552e-01 4.50540811e-01 6.33079112e-01
-5.88078052e-02 -4.10459220e-01 6.91110075e-01 -4.41618085e-01
-1.23952138e+00 3.41765195e-01 -7.66650915e-01 5.88008344e-01
1.31845605e+00 -5.68207428e-02 -2.29649231e-01 -6.61018133e-01
-1.01994133e+00 5.50775111e-01 3.65255922e-01 6.75820947e-01
5.66876531e-01 -1.14994228e+00 -6.79148555e-01 8.49796310e-02
-1.25207961e-01 -1.65210128e-01 2.34816656e-01 2.49560222e-01
-3.48713160e-01 6.54425919e-02 -9.29946065e-01 -3.42453569e-01
-8.26906323e-01 8.90507817e-01 3.07106435e-01 -7.68826246e-01
-4.87172335e-01 7.42319524e-01 1.13516770e-01 -3.91818404e-01
2.28718221e-01 -2.37395227e-01 -3.26355815e-01 -1.74000084e-01
5.93185663e-01 1.68476924e-01 3.92128117e-02 2.07260057e-01
5.44896908e-03 1.80425867e-02 -6.03580713e-01 -2.65994966e-01
1.46538627e+00 7.40438029e-02 5.74941337e-01 2.84363747e-01
3.14911276e-01 -7.12266505e-01 -1.87164605e+00 -5.78163445e-01
7.31790513e-02 -3.44922543e-01 -1.87360600e-01 -1.04332817e+00
-9.94354486e-01 6.27942801e-01 4.36434269e-01 1.68499231e-01
1.03304780e+00 -1.56633332e-01 5.26255071e-01 9.97503281e-01
9.02859151e-01 -1.43802834e+00 8.67945969e-01 1.19008958e+00
5.79055071e-01 -1.35143518e+00 -3.47006954e-02 3.57955873e-01
-9.76407349e-01 9.17116582e-01 1.00285637e+00 -5.35925150e-01
3.53010029e-01 1.68959409e-01 -1.25403449e-01 -5.42542385e-03
-1.04410076e+00 -3.53852212e-01 -2.07966864e-01 1.02403259e+00
-2.69923270e-01 -7.81620443e-02 9.87244546e-02 4.60388035e-01
-3.40045571e-01 2.43613452e-01 6.92691922e-01 1.05405712e+00
-6.63416624e-01 -1.17113245e+00 -2.86543705e-02 5.05271912e-01
-2.67060161e-01 7.00663626e-02 -6.84147999e-02 7.28494465e-01
-4.85284775e-02 6.81104422e-01 -2.91694868e-02 -3.66503119e-01
2.52417624e-01 8.60443264e-02 9.65250731e-01 -8.96586418e-01
-6.66432619e-01 -4.14543629e-01 -1.37126505e-01 -6.55920088e-01
-2.54979402e-01 -6.66997373e-01 -1.25930774e+00 -3.05077702e-01
1.22381181e-01 3.06522816e-01 6.66785538e-02 1.13667643e+00
7.32620806e-02 7.30856240e-01 5.11050463e-01 -8.11563253e-01
-9.77729380e-01 -7.85758138e-01 -6.00729942e-01 5.28553963e-01
1.95579290e-01 -1.11759233e+00 3.55896279e-02 -1.72637880e-01] | [4.046187400817871, 1.5248554944992065] |
515a7622-3bf2-4304-ad20-63bdef50e2e9 | empirical-study-incorporating-linguistic | 2210.07559 | null | https://arxiv.org/abs/2210.07559v1 | https://arxiv.org/pdf/2210.07559v1.pdf | Empirical Study Incorporating Linguistic Knowledge on Filled Pauses for Personalized Spontaneous Speech Synthesis | We present a comprehensive empirical study for personalized spontaneous speech synthesis on the basis of linguistic knowledge. With the advent of voice cloning for reading-style speech synthesis, a new voice cloning paradigm for human-like and spontaneous speech synthesis is required. We, therefore, focus on personalized spontaneous speech synthesis that can clone both the individual's voice timbre and speech disfluency. Specifically, we deal with filled pauses, a major source of speech disfluency, which is known to play an important role in speech generation and communication in psychology and linguistics. To comparatively evaluate personalized filled pause insertion and non-personalized filled pause prediction methods, we developed a speech synthesis method with a non-personalized external filled pause predictor trained with a multi-speaker corpus. The results clarify the position-word entanglement of filled pauses, i.e., the necessity of precisely predicting positions for naturalness and the necessity of precisely predicting words for individuality on the evaluation of synthesized speech. | ['Hiroshi Saruwatari', 'Shinnosuke Takamichi', 'Takaaki Saeki', 'Yuta Matsunaga'] | 2022-10-14 | null | null | null | null | ['voice-cloning'] | ['speech'] | [ 4.54322219e-01 5.30687928e-01 -2.10602283e-01 -3.27431887e-01
-6.86741233e-01 -5.20570815e-01 4.10816759e-01 -2.18571782e-01
-1.28914386e-01 5.19996107e-01 7.77982414e-01 -4.18314815e-01
1.58627793e-01 -3.78859460e-01 -3.91785055e-01 -3.83109957e-01
4.42472875e-01 1.97870553e-01 -1.69346228e-01 -2.73729026e-01
-5.89422658e-02 3.27734172e-01 -1.58589005e+00 2.15346470e-01
1.04649794e+00 4.19539630e-01 8.13783467e-01 9.12480295e-01
-1.68232650e-01 4.02016550e-01 -9.98756945e-01 -3.74334604e-01
-3.69056970e-01 -9.82123971e-01 -8.34066868e-01 2.60129064e-01
-2.54444387e-02 -2.14706182e-01 -5.64702898e-02 8.20255876e-01
7.78292596e-01 2.26931214e-01 4.25499231e-01 -8.40551913e-01
-7.48560011e-01 1.56672585e+00 5.43218255e-01 4.07488465e-01
6.26937449e-01 4.35470138e-03 1.20233715e+00 -6.51705623e-01
7.17104495e-01 1.23792696e+00 2.77566373e-01 1.03866231e+00
-1.19031751e+00 -2.50571609e-01 5.00403754e-02 -1.41887501e-01
-1.23466098e+00 -9.20445859e-01 1.02130163e+00 -4.19619054e-01
9.78110194e-01 7.45142996e-01 5.99559724e-01 1.47124994e+00
5.89166991e-02 7.60648310e-01 6.90490067e-01 -6.62856758e-01
1.52020618e-01 2.86913604e-01 -8.23091343e-02 3.10025215e-01
-4.74249929e-01 4.49934810e-01 -6.04743123e-01 2.98171967e-01
4.66809094e-01 -6.04264259e-01 -5.06008387e-01 5.20631909e-01
-1.36853445e+00 6.52426183e-01 -3.45429808e-01 8.06108892e-01
-2.23682925e-01 -2.94809669e-01 3.47758204e-01 2.23510250e-01
4.47962105e-01 7.56605268e-01 -4.51037437e-01 -6.11917615e-01
-8.26267838e-01 1.09060064e-01 6.83283269e-01 9.89201963e-01
2.35277504e-01 5.06706893e-01 -5.66513002e-01 1.13145053e+00
-1.04260370e-01 7.59323955e-01 9.17134285e-01 -8.48005772e-01
1.71464756e-01 1.20540679e-01 -3.26191597e-02 -7.32918918e-01
-4.10384417e-01 -4.05689031e-01 -3.72749060e-01 -7.12598622e-01
2.30327159e-01 -2.57377058e-01 -3.97535980e-01 2.15373659e+00
2.18094543e-01 -1.05162732e-01 2.59814620e-01 5.87549746e-01
9.41111505e-01 8.46622705e-01 -3.73136140e-02 -8.81873131e-01
1.23784065e+00 -9.32208478e-01 -1.20847166e+00 -9.88248810e-02
5.92186093e-01 -8.57112706e-01 1.86977506e+00 6.17528856e-02
-1.19917715e+00 -6.97247386e-01 -8.08221757e-01 -9.64637771e-02
-1.62282839e-01 2.61141062e-01 2.81457752e-01 1.13291788e+00
-6.59598589e-01 5.94780087e-01 -4.40446794e-01 -1.47800624e-01
-3.46019477e-01 1.74902827e-02 -1.96454465e-01 7.27196634e-01
-1.32427895e+00 6.96903944e-01 1.34470582e-01 -1.30344376e-01
-4.40204412e-01 -5.93070626e-01 -6.82629883e-01 9.98763517e-02
5.33666164e-02 -2.55605221e-01 1.44873512e+00 -8.04769814e-01
-2.08940864e+00 7.96125770e-01 -4.30349499e-01 -2.12496877e-01
1.58183038e-01 -1.25445962e-01 -8.91350091e-01 -9.62312520e-02
-2.10639667e-02 4.45607185e-01 7.62381196e-01 -1.00320017e+00
-4.84707534e-01 -1.62686676e-01 -6.62772834e-01 2.02595130e-01
-2.47345164e-01 2.44847894e-01 1.75841302e-01 -8.69774461e-01
-7.54268914e-02 -9.41411316e-01 1.70067906e-01 -8.75449479e-01
-7.12939620e-01 -5.31211674e-01 4.19838458e-01 -9.27620232e-01
1.57143927e+00 -2.31020546e+00 3.33899617e-01 -1.68124288e-01
-1.12161510e-01 3.16856474e-01 -1.90881446e-01 4.55430090e-01
4.47864346e-02 3.72135073e-01 -8.45207460e-03 -6.62353516e-01
1.73730895e-01 2.13203549e-01 -4.91067737e-01 6.53229579e-02
-4.42148633e-02 8.43728423e-01 -9.40503716e-01 -2.95626074e-01
1.47994444e-01 2.46039093e-01 -7.04104841e-01 4.85593289e-01
-4.67562199e-01 9.20464396e-01 -5.26412874e-02 4.44689244e-01
4.12365459e-02 4.73116606e-01 1.37861729e-01 2.24700972e-01
-5.77289939e-01 9.06084180e-01 -6.36076331e-01 1.46752679e+00
-6.24888957e-01 3.81596208e-01 -5.18721640e-02 -4.62191641e-01
1.01547170e+00 6.38916075e-01 9.43554044e-02 -7.77384341e-01
2.53651291e-01 3.61843795e-01 3.98260623e-01 -5.58137238e-01
8.99271786e-01 -3.23142856e-01 -1.10096604e-01 4.19270903e-01
-3.98312435e-02 -5.19000292e-01 -5.09836823e-02 -2.82410890e-01
8.30538452e-01 -1.82768211e-01 2.20820025e-01 -1.52534559e-01
3.84731501e-01 -4.40024793e-01 5.80230296e-01 5.12887597e-01
-1.56006768e-01 7.87878513e-01 3.99162263e-01 1.97916552e-01
-8.72426510e-01 -1.07665133e+00 4.91179638e-02 1.35039151e+00
-2.07582071e-01 -5.34994841e-01 -1.12148118e+00 -8.06612819e-02
-4.81164783e-01 1.63163376e+00 -1.28528327e-01 -3.74338329e-01
-7.67332196e-01 -5.96827865e-02 9.04756069e-01 1.93348184e-01
-1.13895319e-01 -1.58285630e+00 -2.54215114e-02 4.90256757e-01
-7.54599452e-01 -1.31837320e+00 -9.26376164e-01 -1.25633821e-01
-3.65686655e-01 -4.36274141e-01 -4.84061450e-01 -9.18849349e-01
-6.60721725e-03 -3.87469418e-02 7.11176336e-01 -5.69422245e-02
2.65941620e-02 3.10675859e-01 -6.00138843e-01 -4.87065554e-01
-1.22254407e+00 3.39702934e-01 3.23866695e-01 -1.46745786e-01
-1.23712882e-01 -7.14700580e-01 -1.29876602e-02 3.50185215e-01
-7.39014626e-01 1.75631210e-01 -4.73743821e-05 7.91101992e-01
4.14524108e-01 -1.73248738e-01 8.82091403e-01 -7.06526995e-01
1.11879098e+00 -1.38351157e-01 -2.70411193e-01 3.42750341e-01
-3.02793324e-01 -1.31968871e-01 8.33071589e-01 -6.46395981e-01
-1.28548229e+00 -2.75309116e-01 -7.95365214e-01 -6.95772991e-02
-2.61434942e-01 2.29718015e-01 -6.73707604e-01 5.50578415e-01
7.73991346e-01 5.35833299e-01 1.12033486e-02 -6.41953588e-01
6.64730608e-01 1.10918534e+00 5.61924636e-01 -5.16266823e-01
2.77371883e-01 -2.82260984e-01 -5.92278242e-01 -1.58315384e+00
-3.91725659e-01 -1.85522780e-01 -3.36995840e-01 -6.90592527e-02
9.04794335e-01 -3.24626327e-01 -5.80927551e-01 5.60032189e-01
-1.46075308e+00 -3.95867050e-01 -5.67623019e-01 5.79618633e-01
-8.75372469e-01 4.73720700e-01 -6.31368637e-01 -8.43831062e-01
-3.84213418e-01 -1.30642653e+00 9.61883724e-01 4.74537574e-02
-7.55515397e-01 -7.84651518e-01 1.49693027e-01 2.90386558e-01
4.22632158e-01 -3.50524992e-01 1.22790146e+00 -8.57839286e-01
-1.53613672e-01 2.93063432e-01 6.11329675e-01 6.15258694e-01
4.60780680e-01 3.33796698e-03 -9.28225815e-01 3.41915905e-01
2.29053661e-01 1.59795344e-01 1.36390015e-01 2.58856684e-01
7.73302972e-01 -6.95467293e-01 2.71375357e-05 5.10019541e-01
4.63508487e-01 5.39211869e-01 6.07817709e-01 -3.80546302e-01
5.11301160e-01 8.22492301e-01 3.71224046e-01 5.43696225e-01
3.29340160e-01 7.12554395e-01 -1.96352169e-01 3.35988998e-01
-4.07623440e-01 -7.78517187e-01 5.18670619e-01 1.45827258e+00
8.63711834e-02 -7.88598955e-01 -6.36397481e-01 6.34812236e-01
-1.22377253e+00 -9.19194758e-01 -7.61353895e-02 2.13248777e+00
1.19839311e+00 -1.75613016e-01 5.90438843e-02 2.63596803e-01
9.68124866e-01 3.06142747e-01 -4.88512181e-02 -7.78693557e-01
-3.81102651e-01 3.49945903e-01 -1.25232905e-01 7.86683083e-01
-4.40886974e-01 1.58548009e+00 6.45296764e+00 8.90019894e-01
-1.26198423e+00 8.45166519e-02 2.79442370e-01 1.48217365e-01
-8.74813557e-01 -2.21212849e-01 -8.96741986e-01 6.42589927e-01
1.25120068e+00 -3.88962835e-01 9.16299999e-01 7.21616685e-01
5.56627572e-01 2.00078547e-01 -1.19774306e+00 8.47743750e-01
-2.53011677e-02 -1.11133194e+00 -7.51253739e-02 -2.46472597e-01
4.34519321e-01 -3.71007681e-01 1.25336023e-02 1.22502275e-01
-3.68753761e-01 -9.16427553e-01 1.05126405e+00 4.87257242e-01
9.04986203e-01 -4.50222462e-01 1.53363481e-01 5.40434957e-01
-7.43865311e-01 2.20588610e-01 -1.35939568e-01 -5.48480311e-03
5.14906526e-01 5.33995211e-01 -1.09834671e+00 3.65390331e-02
-1.31585866e-01 1.22743264e-01 -1.71382695e-01 5.16570628e-01
-5.18342853e-01 1.01099730e+00 -1.97962046e-01 -4.31965351e-01
-3.20021093e-01 -4.73861061e-02 9.74379838e-01 1.10105407e+00
5.33992648e-01 8.93080011e-02 -1.31872728e-01 9.11706388e-01
1.59848690e-01 5.56348383e-01 -4.10784960e-01 -8.09997141e-01
8.21001112e-01 7.04442918e-01 -5.53749502e-01 8.86368006e-02
-3.86012793e-02 1.11550975e+00 7.33480528e-02 2.85160720e-01
-6.84101582e-01 -1.65030107e-01 8.00963283e-01 1.14582911e-01
3.88861783e-02 -3.04629862e-01 -2.44835258e-01 -8.67770791e-01
-6.49997517e-02 -9.82152939e-01 -3.94956201e-01 -6.75852478e-01
-1.07523501e+00 9.13254321e-01 -2.36371353e-01 -7.24471748e-01
-7.65792012e-01 -1.65336937e-01 -5.68781972e-01 8.71477664e-01
-8.38229001e-01 -1.01995933e+00 2.53054053e-01 4.06916767e-01
8.27926099e-01 -2.27773383e-01 1.22598469e+00 8.84408578e-02
-4.90416080e-01 8.33448648e-01 -2.10291609e-01 -1.03422895e-01
4.35411274e-01 -9.05014455e-01 8.10633361e-01 7.45648265e-01
1.39901891e-01 6.32387996e-01 9.15414035e-01 -7.38435268e-01
-9.96922910e-01 -7.75678456e-01 1.67291069e+00 -1.61153659e-01
4.75632995e-01 -5.32245398e-01 -6.79753482e-01 5.58354795e-01
9.37508866e-02 -5.04138649e-01 9.00936663e-01 2.65556067e-01
1.80737332e-01 8.96778703e-02 -9.08019245e-01 9.11337197e-01
1.39394557e+00 -8.17307055e-01 -8.77891839e-01 2.51885116e-01
1.49848306e+00 -2.91915894e-01 -5.41763842e-01 6.24899454e-02
3.52208376e-01 -1.05667531e+00 4.94265795e-01 -3.62354845e-01
2.04082593e-01 1.18216708e-01 -2.35745028e-01 -1.72457397e+00
-2.38624394e-01 -1.27248299e+00 3.13076228e-01 1.57647240e+00
4.74032462e-01 -5.51806331e-01 1.98983386e-01 4.55058396e-01
-7.61408329e-01 -1.36260092e-01 -1.09597445e+00 -8.82771134e-01
4.74746898e-02 -6.02434516e-01 8.57209504e-01 5.23465693e-01
4.43323612e-01 5.65576196e-01 -5.17172813e-01 -1.25421822e-01
-1.60632163e-01 -1.57488406e-01 4.55021799e-01 -8.72638106e-01
-4.03856903e-01 -4.29001123e-01 7.40424097e-02 -1.12414551e+00
4.97139394e-01 -7.81351566e-01 3.60638738e-01 -1.03912640e+00
-5.28674245e-01 -3.21648598e-01 3.82689953e-01 3.84163894e-02
-6.29716888e-02 -5.37518322e-01 3.23050767e-01 -1.04415618e-01
4.06799838e-02 1.00748217e+00 1.36608446e+00 3.27443480e-01
-6.51038051e-01 5.72426438e-01 -5.14545679e-01 5.74591100e-01
8.24855924e-01 -4.03656036e-01 -6.30898178e-01 -1.13624863e-01
-6.18073763e-03 6.69027925e-01 -1.12876542e-01 -7.47006595e-01
3.24671715e-02 -2.57920742e-01 -3.72883797e-01 -4.24169242e-01
3.68523538e-01 -2.86924362e-01 1.71129078e-01 2.00660869e-01
-6.59448385e-01 7.46212248e-03 2.97046024e-02 7.23075878e-04
-5.74751683e-02 -3.78000021e-01 5.30153871e-01 -2.02378750e-01
-3.07492286e-01 -9.85847116e-02 -8.35635006e-01 2.60753810e-01
6.34733438e-01 -2.92600960e-01 -2.07615227e-01 -4.40550506e-01
-9.45589066e-01 -4.26842779e-01 1.96024030e-01 8.57042968e-01
5.50059259e-01 -9.71188843e-01 -5.97765148e-01 5.39136529e-01
-7.45283812e-02 -5.80179870e-01 4.34196264e-01 5.20149589e-01
-3.70317101e-02 4.95469391e-01 1.45667538e-01 -2.07653925e-01
-1.19291151e+00 5.57338715e-01 3.79471689e-01 2.77005017e-01
-4.60578442e-01 1.03347909e+00 8.49840790e-02 -5.80786467e-01
3.20053756e-01 -5.72007120e-01 -6.75729439e-02 -4.53322232e-02
5.23093224e-01 5.19029856e-01 3.96342166e-02 -8.02058935e-01
-1.58510074e-01 -1.14229918e-01 2.46530145e-01 -4.54782724e-01
7.95686841e-01 -5.18242300e-01 -1.94744840e-01 8.98713827e-01
9.50342596e-01 8.59883845e-01 -7.13809490e-01 3.23535465e-02
-1.37693733e-01 -1.81560516e-01 -1.06970206e-01 -8.23647976e-01
-4.84369218e-01 6.56703055e-01 1.63139582e-01 3.97699356e-01
6.90654993e-01 1.79176882e-01 1.19897413e+00 4.80301976e-02
2.29574427e-01 -1.36494017e+00 -1.59594253e-01 8.18154752e-01
1.12751567e+00 -8.02491486e-01 -8.29866052e-01 -6.53688312e-01
-8.95786226e-01 8.80431831e-01 4.89284068e-01 4.75627363e-01
6.07019424e-01 9.94303748e-02 2.18770709e-02 3.56933117e-01
-4.62883741e-01 -1.69053853e-01 2.35172823e-01 6.94911301e-01
6.97886288e-01 6.87730551e-01 -8.21042895e-01 1.10200202e+00
-1.22267020e+00 -3.02131712e-01 4.21122432e-01 3.92424643e-01
-5.84355652e-01 -1.30912423e+00 -2.42405549e-01 -3.20524201e-02
-3.33414018e-01 -4.76460457e-01 -5.98748386e-01 3.36600751e-01
2.59257436e-01 1.37704074e+00 8.31834823e-02 -6.37999296e-01
3.58077526e-01 3.61738890e-01 3.49919349e-01 -8.87494624e-01
-6.09635949e-01 2.35209808e-01 3.82975698e-01 -3.02113861e-01
-7.25178421e-02 -7.40732193e-01 -1.43502045e+00 -2.42765844e-01
-3.56627762e-01 3.22031111e-01 8.26843500e-01 1.12755585e+00
2.18448430e-01 6.28621995e-01 6.02806509e-01 -5.74909389e-01
-5.67225814e-01 -1.18771398e+00 -9.41769481e-01 1.03779875e-01
3.13132495e-01 -2.03678891e-01 -6.24778688e-01 -6.37710169e-02] | [14.7843656539917, 6.742637634277344] |
d824afbf-f080-4821-b303-012265d1b254 | stride-scene-text-recognition-in-device | 2105.07795 | null | https://arxiv.org/abs/2105.07795v1 | https://arxiv.org/pdf/2105.07795v1.pdf | STRIDE : Scene Text Recognition In-Device | Optical Character Recognition (OCR) systems have been widely used in various applications for extracting semantic information from images. To give the user more control over their privacy, an on-device solution is needed. The current state-of-the-art models are too heavy and complex to be deployed on-device. We develop an efficient lightweight scene text recognition (STR) system, which has only 0.88M parameters and performs real-time text recognition. Attention modules tend to boost the accuracy of STR networks but are generally slow and not optimized for device inference. So, we propose the use of convolution attention modules to the text recognition networks, which aims to provide channel and spatial attention information to the LSTM module by adding very minimal computational cost. It boosts our word accuracy on ICDAR 13 dataset by almost 2\%. We also introduce a novel orientation classifier module, to support the simultaneous recognition of both horizontal and vertical text. The proposed model surpasses on-device metrics of inference time and memory footprint and achieves comparable accuracy when compared to the leading commercial and other open-source OCR engines. We deploy the system on-device with an inference speed of 2.44 ms per word on the Exynos 990 chipset device and achieve an accuracy of 88.4\% on ICDAR-13 dataset. | ['Gopi Ramena', 'Sukumar Moharana', 'Nikhil Arora', 'Arun D Prabhu', 'Rachit S Munjal'] | 2021-05-17 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 3.83738518e-01 -4.17771906e-01 -9.96553749e-02 -4.92773414e-01
-4.44512933e-01 -3.35088521e-01 3.22326958e-01 4.89242189e-02
-7.76325643e-01 1.74813852e-01 -3.73613179e-01 -6.85296178e-01
2.85478204e-01 -6.46858573e-01 -6.64532959e-01 -4.17783111e-01
7.54517615e-01 2.31793433e-01 3.12032998e-01 2.31690526e-01
5.44545889e-01 6.92770779e-01 -1.42039812e+00 3.77918631e-01
7.45004892e-01 1.55970955e+00 3.12996745e-01 8.84054661e-01
-3.07160616e-01 8.98232162e-01 -7.41181850e-01 -7.67184198e-01
-1.05472624e-01 2.52091676e-01 -3.67431343e-01 -1.39924452e-01
7.81714439e-01 -7.32465386e-01 -6.98450923e-01 9.15256023e-01
7.18090713e-01 -4.26720589e-01 5.78258872e-01 -8.15438807e-01
-8.07821870e-01 3.72017443e-01 -5.34323096e-01 3.45318578e-02
-6.04850948e-02 -6.73935637e-02 5.94584644e-01 -8.31845462e-01
5.75334616e-02 8.52227569e-01 5.43364882e-01 7.07681894e-01
-8.60784471e-01 -8.29488516e-01 -1.40782103e-01 3.34696293e-01
-1.57636511e+00 -7.49439061e-01 6.03569269e-01 -1.64618120e-01
1.53763890e+00 3.65085930e-01 1.31734520e-01 1.30707991e+00
5.61880410e-01 9.53534663e-01 9.01098669e-01 -3.24106425e-01
2.61889607e-01 5.13749838e-01 4.36031789e-01 7.04039276e-01
4.46929425e-01 -6.20115042e-01 -7.06827581e-01 3.04329365e-01
7.30001807e-01 3.29617053e-01 8.00175443e-02 2.05103323e-01
-7.95847476e-01 4.42192048e-01 2.27829412e-01 1.62704214e-01
2.32288893e-02 5.14658391e-01 5.41188419e-01 -9.15349051e-02
1.98926315e-01 4.04598899e-02 -1.52268022e-01 -2.16731295e-01
-9.00179565e-01 -2.55080521e-01 5.42295218e-01 1.22772920e+00
1.93842977e-01 1.85451314e-01 -1.43252209e-01 1.06313169e+00
4.14245307e-01 1.08192885e+00 5.26662290e-01 -1.46520957e-01
6.78791106e-01 6.16162181e-01 -1.22745469e-01 -8.62549841e-01
-3.03773105e-01 -3.16900760e-01 -9.50301647e-01 -2.90132165e-02
4.80532721e-02 6.79445490e-02 -1.29982507e+00 7.99127162e-01
-3.42509449e-01 -2.03450307e-01 -1.45149291e-01 6.52661622e-01
6.69713914e-01 6.09982848e-01 1.15339300e-02 2.78967261e-01
1.64449084e+00 -7.51434147e-01 -9.65328872e-01 -4.33316290e-01
5.70668161e-01 -9.52834368e-01 1.13093519e+00 6.48124278e-01
-8.42121005e-01 -6.45059407e-01 -1.50144172e+00 -6.04911804e-01
-7.09515870e-01 7.84294128e-01 4.91066128e-01 1.18420732e+00
-8.04488659e-01 3.44455302e-01 -8.33486497e-01 -3.43069196e-01
7.37660110e-01 7.73814678e-01 -1.48620633e-02 -1.02713048e-01
-5.87448061e-01 5.86035609e-01 1.64717898e-01 3.47584724e-01
-3.61852169e-01 -2.22558066e-01 -6.08759403e-01 3.92044336e-02
2.03648403e-01 -3.28680336e-01 9.67203915e-01 -6.87691927e-01
-2.03484035e+00 7.53918588e-01 -2.45433107e-01 -5.24224699e-01
2.71896869e-01 -5.43656468e-01 -8.09771657e-01 1.81161761e-01
-4.04334366e-01 5.70188344e-01 1.15110111e+00 -5.18610418e-01
-5.78949511e-01 -5.52494228e-01 -4.56698269e-01 -5.17803580e-02
-9.67415154e-01 1.73963860e-01 -8.38851213e-01 -5.56333661e-01
-9.96628851e-02 -7.60512769e-01 4.65235591e-01 1.83575630e-01
-6.58918440e-01 -3.55349854e-02 1.37430882e+00 -6.07006073e-01
1.29054177e+00 -2.22523427e+00 -4.82087195e-01 4.91549522e-02
8.96402523e-02 8.15195441e-01 2.42152244e-01 -1.31250128e-01
2.87024468e-01 -1.47292484e-02 2.26642147e-01 -6.85824394e-01
6.52764365e-03 -3.16605419e-02 -5.78420877e-01 5.44929981e-01
1.96190104e-01 9.06336904e-01 -7.05203936e-02 -4.71176594e-01
7.31644273e-01 6.44348443e-01 -4.77457047e-01 5.14911301e-02
-7.56924674e-02 -3.33297729e-01 -3.88443947e-01 8.91482651e-01
7.84916639e-01 -2.37430304e-01 1.79676697e-01 -2.91821629e-01
-1.39632970e-01 2.82651424e-01 -9.27517712e-01 1.53453910e+00
-4.38201100e-01 1.05443394e+00 -2.19161838e-01 -6.72451973e-01
1.20800495e+00 1.10550821e-01 -7.95696750e-02 -1.10595584e+00
6.29244864e-01 2.68712193e-01 -2.71671116e-01 -6.97290599e-01
8.52028251e-01 3.61201555e-01 -2.78417976e-03 3.86977270e-02
-1.13672562e-01 2.74676830e-01 -3.94225270e-01 -1.02022737e-01
9.97496545e-01 -1.29481286e-01 -2.57277638e-01 -2.82239050e-01
5.12706757e-01 -4.37146515e-01 2.99763065e-02 8.01758409e-01
2.18487345e-02 4.04555500e-01 1.19829990e-01 -4.36537266e-01
-1.07065201e+00 -8.97458017e-01 -4.32165653e-01 8.14022660e-01
1.20150469e-01 -4.49742585e-01 -9.14404750e-01 -4.55536842e-01
-1.63654029e-01 6.79168344e-01 -1.59731925e-01 1.22974582e-01
-4.31735992e-01 -7.31337965e-01 1.13318336e+00 8.49303126e-01
1.09830606e+00 -7.15619981e-01 -7.89739370e-01 1.33753922e-02
6.09175973e-02 -1.59768343e+00 -4.35464352e-01 1.71227008e-01
-8.39640081e-01 -5.41686118e-01 -5.75759888e-01 -7.12582707e-01
5.60591936e-01 1.85213815e-02 4.63081151e-01 -2.22816408e-01
-6.80781841e-01 2.20567316e-01 -1.69342235e-01 -6.09678984e-01
2.22979441e-01 3.24198902e-01 6.02723509e-02 3.84199053e-01
8.42205882e-01 -1.05821602e-01 -6.53426945e-01 3.71233761e-01
-7.66205847e-01 6.20489568e-02 7.24809408e-01 5.66723108e-01
4.08483595e-01 5.51965833e-02 2.18489796e-01 -7.32561886e-01
2.66683310e-01 1.82572931e-01 -8.18968475e-01 1.67428628e-01
-6.62516832e-01 1.52739689e-01 8.97621036e-01 -4.70001549e-01
-9.87235487e-01 2.27727711e-01 -3.42727214e-01 -4.48172331e-01
-1.14752926e-01 -2.36537643e-02 -2.41611615e-01 -2.60519981e-01
3.08090836e-01 3.26287806e-01 -2.51308173e-01 -5.52976429e-01
-1.49873778e-01 1.57561576e+00 5.21673977e-01 -1.94457605e-01
3.36812645e-01 5.63649714e-01 -1.31836534e-01 -1.42473674e+00
-3.33250880e-01 -3.84586036e-01 -3.37722152e-01 2.50641908e-03
1.02150643e+00 -1.01025677e+00 -1.31494379e+00 1.19554913e+00
-1.12064660e+00 -3.12551111e-01 3.55325282e-01 3.30703437e-01
-2.04274014e-01 3.22838485e-01 -6.42020583e-01 -1.06745267e+00
-8.05098891e-01 -1.16741145e+00 1.36929464e+00 1.83638558e-01
9.04248096e-04 -5.75449646e-01 -7.25601912e-01 4.24262106e-01
5.66089809e-01 -4.10711229e-01 7.87547171e-01 -4.78325248e-01
-8.51325870e-01 -5.49140930e-01 -7.24754453e-01 4.22717988e-01
-2.28859052e-01 -2.79487446e-02 -1.46671546e+00 -1.49375618e-01
-8.97050370e-03 -2.51715451e-01 8.61141860e-01 2.51777679e-01
1.80952358e+00 -5.83763048e-02 -4.79183942e-01 6.82819366e-01
1.62578154e+00 3.21650714e-01 1.32308567e+00 2.36671697e-02
1.11487842e+00 1.28407449e-01 2.58170724e-01 3.48223090e-01
1.37805611e-01 8.88007641e-01 8.28093067e-02 9.74373370e-02
5.86277107e-03 -2.25805804e-01 4.57878649e-01 8.16402078e-01
2.57996559e-01 -7.31200278e-01 -8.80267739e-01 1.93704799e-01
-1.68051839e+00 -5.74511349e-01 -2.46157795e-01 2.32673120e+00
3.83332998e-01 5.21973431e-01 -3.26135010e-01 3.09132636e-01
6.79478109e-01 4.16173041e-02 -5.36014795e-01 -8.31691027e-01
3.76156345e-02 5.97235560e-01 1.14427638e+00 2.45597273e-01
-1.13536656e+00 9.42222834e-01 5.98840284e+00 1.15686715e+00
-1.56948137e+00 -1.30922556e-01 6.72719777e-01 -3.18781644e-01
2.60583878e-01 -5.78815699e-01 -1.36812699e+00 5.98189354e-01
1.30356348e+00 6.04581773e-01 2.61641026e-01 1.02450109e+00
-6.31158873e-02 -5.99077344e-02 -1.18059599e+00 1.55268157e+00
4.24970388e-01 -1.28659225e+00 1.42659217e-01 2.51374006e-01
1.16145998e-01 3.92913483e-02 4.34578300e-01 7.78247863e-02
-3.53348523e-01 -1.23404050e+00 7.01929569e-01 3.67570072e-01
1.44416726e+00 -8.71394277e-01 8.35767746e-01 1.19349584e-01
-1.02905560e+00 -2.57760733e-01 -5.39098442e-01 5.28710671e-02
-1.62370324e-01 5.85735559e-01 -6.41428530e-01 -3.67120020e-02
8.31659019e-01 4.42654073e-01 -6.15811050e-01 4.88030881e-01
1.74196422e-01 4.15283948e-01 -4.72659558e-01 -7.07083941e-01
1.28957853e-01 1.69539645e-01 5.04590161e-02 1.37488019e+00
3.35841209e-01 -1.97668850e-01 -4.86902058e-01 6.05323076e-01
-3.76919508e-01 7.08394721e-02 -4.87422794e-01 -2.52290934e-01
3.54328126e-01 1.19619215e+00 -6.64942861e-01 -3.08782697e-01
-4.07706320e-01 1.64946222e+00 1.37038082e-01 4.62706089e-02
-1.09777629e+00 -1.08340347e+00 5.67167580e-01 1.29109159e-01
5.75809836e-01 -3.89008015e-01 -6.54519200e-01 -1.03272104e+00
4.11310017e-01 -6.63422406e-01 -8.38746969e-03 -8.98868501e-01
-8.00045669e-01 5.63134313e-01 -7.35030472e-01 -8.20377111e-01
2.06519887e-01 -1.49757445e+00 -2.63019919e-01 7.82729924e-01
-1.38899386e+00 -1.15569854e+00 -4.88748372e-01 5.92677653e-01
6.18823707e-01 -2.69209921e-01 9.06982303e-01 6.65183246e-01
-1.05779088e+00 1.26308918e+00 3.35656375e-01 5.21258295e-01
6.40810609e-01 -8.16943944e-01 6.88219666e-01 7.53437221e-01
8.31962898e-02 7.45307863e-01 1.87627599e-01 -5.69551706e-01
-2.06209779e+00 -1.17419946e+00 7.64971972e-01 -4.98081714e-01
3.61748129e-01 -9.59373474e-01 -6.30773306e-01 5.84642351e-01
3.80149744e-02 -4.72200736e-02 6.50386631e-01 -1.68823570e-01
-6.01049483e-01 -5.08400798e-01 -1.19668698e+00 8.26460123e-01
8.59112918e-01 -1.01614618e+00 -2.10684421e-03 9.66397375e-02
4.84266311e-01 -3.24567229e-01 -5.90161085e-01 1.35356948e-01
8.40855479e-01 -7.24696398e-01 7.29141057e-01 -6.26325160e-02
3.44913691e-01 -1.87649742e-01 -4.07831311e-01 -2.77704358e-01
-8.82289112e-02 -5.73785484e-01 -1.27092868e-01 1.25629616e+00
4.13936019e-01 -8.12565327e-01 1.04349363e+00 6.66151941e-01
4.94110510e-02 -5.42509198e-01 -9.25559998e-01 -6.79447234e-01
-3.38295937e-01 -8.02984059e-01 1.63284928e-01 3.47700179e-01
3.80962454e-02 4.92242068e-01 -4.87556666e-01 3.15524250e-01
6.73442304e-01 -3.21743816e-01 5.14355063e-01 -8.67114186e-01
-2.62809664e-01 -2.77211934e-01 -8.01259100e-01 -1.40863764e+00
-8.29475746e-02 -6.34422302e-01 -1.10269897e-01 -1.02900577e+00
3.78513685e-03 -2.78714716e-01 -1.36568204e-01 2.82762766e-01
3.91348124e-01 6.57882452e-01 -6.54143095e-02 -1.03805505e-01
-6.16115928e-01 3.50619674e-01 4.84590083e-01 -2.28065521e-01
8.23887736e-02 -3.70662689e-01 -5.08175015e-01 5.94434321e-01
8.23385179e-01 -1.07721999e-01 -2.63182700e-01 -8.10290813e-01
1.20586440e-01 -5.26831985e-01 3.79947662e-01 -1.33549583e+00
5.03719151e-01 5.16891718e-01 8.35550427e-01 -8.97752523e-01
5.34123182e-01 -1.08323967e+00 -7.07366318e-02 4.06418502e-01
-1.73367247e-01 -1.99487790e-01 5.46587467e-01 3.64938498e-01
2.22408250e-01 -2.82287419e-01 7.58787274e-01 4.05280620e-01
-7.22576201e-01 1.82409152e-01 -5.36494732e-01 -4.36158448e-01
8.32990348e-01 -4.80011433e-01 -7.20486164e-01 2.29954198e-02
-3.82837169e-02 -2.08550185e-01 2.28672355e-01 4.60869491e-01
8.58626902e-01 -1.08449948e+00 -1.69751003e-01 5.34631550e-01
2.54444331e-01 -3.06189775e-01 3.80211055e-01 5.07226229e-01
-9.35881913e-01 1.03013158e+00 -6.11972325e-02 -6.63568676e-01
-1.36653793e+00 5.44645727e-01 3.00045609e-01 1.03763781e-01
-6.59727633e-01 6.79007411e-01 -1.59938380e-01 2.15305284e-01
5.05007863e-01 -3.85679513e-01 1.14937127e-01 -2.62833446e-01
9.64928508e-01 4.75845724e-01 3.86133254e-01 -1.18560739e-01
-4.17825818e-01 9.38608944e-01 -5.36204755e-01 -1.67773142e-02
9.08549726e-01 -3.52217369e-02 9.12433937e-02 2.66467303e-01
1.42920065e+00 -2.64519826e-02 -9.18056905e-01 -6.06018677e-02
-3.07887077e-01 -4.87748712e-01 4.43332046e-01 -8.16309690e-01
-9.52034533e-01 1.23548186e+00 1.00389397e+00 -5.51153496e-02
1.09586990e+00 -4.51141417e-01 9.82216299e-01 6.66625381e-01
2.96215653e-01 -1.60094440e+00 8.36967826e-02 3.85534406e-01
4.58021969e-01 -1.06829154e+00 -1.09925911e-01 -4.17690188e-01
-4.45153326e-01 1.26926029e+00 5.81594467e-01 9.15099233e-02
4.48455572e-01 8.70465815e-01 -8.86123106e-02 -1.29437670e-01
-3.57200414e-01 2.23320350e-01 2.60692358e-01 4.86790776e-01
4.04198855e-01 9.80546251e-02 2.34410971e-01 3.51565599e-01
1.41909957e-01 1.81143329e-01 2.92649180e-01 7.20804989e-01
-2.93328404e-01 -8.73889387e-01 -2.30974078e-01 7.44460940e-01
-7.85962522e-01 -3.42264801e-01 -3.52522075e-01 2.69922197e-01
-1.48503378e-01 8.77015889e-01 5.84928930e-01 -7.26805031e-01
1.69369802e-01 1.39009625e-01 3.90950203e-01 -1.12416878e-01
-3.12232941e-01 -7.92079568e-02 1.14735872e-01 -5.75703800e-01
9.50939506e-02 -2.45459482e-01 -9.35015500e-01 -7.58620799e-01
-4.39942598e-01 -6.59892142e-01 1.32162309e+00 7.54340231e-01
7.56617904e-01 7.25273490e-01 3.99531692e-01 -4.37412858e-01
-2.83603787e-01 -7.56182790e-01 -5.47327518e-01 -3.29474866e-01
1.05538748e-01 -1.80838749e-01 1.29913241e-01 -2.89469436e-02] | [11.940177917480469, 2.2237861156463623] |
6cab085c-e52c-4581-9d58-7a49397e6528 | streaming-sparse-linear-regression | 2211.06039 | null | https://arxiv.org/abs/2211.06039v2 | https://arxiv.org/pdf/2211.06039v2.pdf | Online Linearized LASSO | Sparse regression has been a popular approach to perform variable selection and enhance the prediction accuracy and interpretability of the resulting statistical model. Existing approaches focus on offline regularized regression, while the online scenario has rarely been studied. In this paper, we propose a novel online sparse linear regression framework for analyzing streaming data when data points arrive sequentially. Our proposed method is memory efficient and requires less stringent restricted strong convexity assumptions. Theoretically, we show that with a properly chosen regularization parameter, the $\ell_2$-norm statistical error of our estimator diminishes to zero in the optimal order of $\tilde{O}({\sqrt{s/t}})$, where $s$ is the sparsity level, $t$ is the streaming sample size, and $\tilde{O}(\cdot)$ hides logarithmic terms. Numerical experiments demonstrate the practical efficiency of our algorithm. | ['Qiang Sun', 'Xiuneng Zhu', 'Yuhao Yan', 'Shuoguang Yang'] | 2022-11-11 | null | null | null | null | ['variable-selection'] | ['methodology'] | [ 2.76796073e-01 -1.01877317e-01 -3.93901050e-01 -2.97369570e-01
-1.08647811e+00 -2.98851520e-01 -3.37578177e-01 5.60601532e-01
-5.93626678e-01 9.02686417e-01 -3.85055542e-01 -4.59412932e-01
-3.87545526e-01 -6.14344478e-01 -8.55770648e-01 -9.07252133e-01
-4.53487188e-01 1.39906317e-01 -1.14795893e-01 8.78687650e-02
2.57053077e-01 1.47328407e-01 -1.20497608e+00 -5.39805889e-01
1.11009264e+00 1.53000414e+00 2.76303720e-02 5.66574574e-01
1.11448824e-01 6.67065680e-01 -4.74351048e-01 -3.09930772e-01
4.73335683e-01 -4.27795231e-01 -2.51760453e-01 2.07477987e-01
1.80235729e-01 -7.31994435e-02 -2.92405277e-01 1.14557600e+00
3.14008147e-01 4.01757926e-01 1.58586711e-01 -1.03080761e+00
-4.35491167e-02 5.61601222e-01 -1.18825889e+00 3.93532634e-01
4.90453169e-02 -1.73145577e-01 1.00249243e+00 -8.62161517e-01
3.72326583e-01 6.28003120e-01 5.87542474e-01 -9.20677632e-02
-1.24467957e+00 -9.85488892e-01 4.15323496e-01 -6.14838749e-02
-1.67871404e+00 -3.16165924e-01 6.32337034e-01 -2.45623454e-01
4.11353439e-01 4.24249828e-01 3.99953783e-01 2.52893925e-01
-1.64768755e-01 6.51829004e-01 9.69846427e-01 -4.03252751e-01
4.70298111e-01 3.10088433e-02 2.35068530e-01 8.39770913e-01
6.74933910e-01 -1.93751946e-01 -6.85526550e-01 -3.38022530e-01
7.76491880e-01 2.01406166e-01 -2.33587325e-01 -2.07286075e-01
-8.73438835e-01 1.05369985e+00 1.04296813e-02 -4.07896042e-02
-4.27192599e-01 3.59015524e-01 4.13590223e-01 3.33055913e-01
7.46836782e-01 -1.50056696e-02 -5.62138796e-01 -3.09591323e-01
-1.25990224e+00 9.53003466e-02 5.53670228e-01 1.13102329e+00
7.05518067e-01 3.47386032e-01 1.01574562e-01 8.61684322e-01
-1.26537710e-01 8.68052304e-01 1.12746824e-02 -1.02803898e+00
9.33919013e-01 3.10792923e-01 2.49764949e-01 -1.23508179e+00
-2.94468224e-01 -5.80903649e-01 -1.00690389e+00 -3.34156454e-01
6.39962792e-01 -6.34055912e-01 -2.42727101e-01 1.42748725e+00
3.72793585e-01 1.42602906e-01 -4.05361474e-01 7.72209644e-01
1.24404170e-01 6.61168754e-01 -2.05033332e-01 -8.69292855e-01
8.18764329e-01 -6.28275990e-01 -5.11582971e-01 -1.44962698e-01
5.88749290e-01 -5.18956602e-01 9.48588729e-01 6.06335521e-01
-1.31321645e+00 -9.01376754e-02 -8.27965200e-01 3.53040963e-01
3.72197181e-01 6.95159659e-02 7.83606470e-01 6.87219024e-01
-5.34196675e-01 3.03514838e-01 -8.89493704e-01 2.88253456e-01
5.75702488e-01 5.52411258e-01 -8.10350478e-02 -1.76435202e-01
-5.91258824e-01 -1.70995314e-02 -1.32251889e-01 2.21883178e-01
-4.66583282e-01 -7.78970540e-01 -6.39935970e-01 1.96656644e-01
8.79009604e-01 -2.33629048e-01 8.62171233e-01 -6.54281497e-01
-1.26378775e+00 4.60073650e-01 -7.10220516e-01 -6.15946770e-01
4.46334511e-01 -7.98682719e-02 -1.41867206e-01 1.43360287e-01
1.15743943e-01 -3.68435919e-01 9.96542573e-01 -8.45581234e-01
-7.77450621e-01 -7.90663362e-01 -1.83446407e-01 -1.32327378e-01
-3.72761786e-01 1.76078945e-01 -4.18450177e-01 -7.43541420e-01
3.92797619e-01 -8.24056804e-01 -8.31131637e-01 3.93589176e-02
1.74747016e-02 1.20242961e-01 1.91034421e-01 -4.97959584e-01
1.71634901e+00 -2.31806350e+00 -8.82973000e-02 7.83206463e-01
2.27920517e-01 -1.06541231e-01 3.69694203e-01 2.20567524e-01
1.21349931e-01 5.35223819e-02 -5.14920056e-01 -1.55787811e-01
-2.43015930e-01 -1.75640717e-01 -2.72637516e-01 8.63282323e-01
-3.76665473e-01 2.16512457e-01 -7.58051157e-01 -2.44817719e-01
-5.67080602e-02 2.46584579e-01 -7.22672522e-01 -3.14949714e-02
-4.48502563e-02 2.43800998e-01 -7.59070456e-01 5.77789426e-01
7.74581015e-01 -6.23921633e-01 3.98007125e-01 3.05605888e-01
-1.12796165e-01 5.37432767e-02 -1.81546009e+00 1.28049600e+00
-6.06540561e-01 4.94265616e-01 4.62363362e-01 -1.41901886e+00
9.86649811e-01 -2.54042838e-02 8.84078979e-01 -6.60182774e-01
1.83519140e-01 3.12600344e-01 -4.55063403e-01 -2.36359030e-01
2.40904987e-01 -1.91040739e-01 -8.99323970e-02 2.87508160e-01
-6.24523342e-01 2.96366096e-01 3.12455684e-01 -1.42523676e-01
1.19104552e+00 -3.00933599e-01 6.28336251e-01 -2.14991093e-01
4.47585434e-01 -2.97203884e-02 9.26125169e-01 7.88529575e-01
5.55319828e-04 3.35079163e-01 7.90985525e-01 -2.38682628e-01
-8.03881884e-01 -6.54512525e-01 -2.25367516e-01 1.30461979e+00
1.84256926e-01 -3.50269526e-01 -5.66743970e-01 -4.59594041e-01
1.97814569e-01 6.58431649e-01 -5.52112639e-01 3.43880057e-01
-8.18812549e-01 -9.05182362e-01 1.02789804e-01 4.92024124e-01
2.24717990e-01 -5.87947965e-01 -5.74409008e-01 2.47142076e-01
2.11118069e-02 -9.26973224e-01 -7.39630282e-01 3.62012058e-01
-1.18553758e+00 -8.56001019e-01 -6.02930367e-01 -5.43650568e-01
1.06367159e+00 6.52124107e-01 7.57867396e-01 -8.65729619e-03
-1.85896128e-01 2.59571429e-02 -3.86261404e-01 -4.28332776e-01
3.51735681e-01 -7.04921111e-02 -6.69814497e-02 9.17216986e-02
1.69581547e-01 -5.18923998e-01 -8.17005932e-01 2.91379690e-01
-8.19988966e-01 -4.63428169e-01 3.45336676e-01 9.30065036e-01
1.03597403e+00 1.91906750e-01 4.61251140e-01 -1.28092599e+00
2.34567523e-01 -6.76799595e-01 -1.21980679e+00 -2.41820831e-02
-7.86607325e-01 7.49965608e-02 1.00835562e+00 -3.71913552e-01
-6.80991769e-01 1.38498679e-01 2.68257886e-01 -4.98460501e-01
4.45338100e-01 8.99014652e-01 1.29435927e-01 -1.14157215e-01
3.51579756e-01 4.08564091e-01 -9.46134627e-02 -3.98337781e-01
1.26624286e-01 4.67659295e-01 2.63241172e-01 -5.31234503e-01
6.90487742e-01 6.29684985e-01 1.84328228e-01 -8.42142582e-01
-6.97794616e-01 -4.56803381e-01 -1.92095637e-01 1.50643483e-01
2.65820511e-02 -8.94433796e-01 -1.11339664e+00 -2.46670872e-01
-3.80051970e-01 -3.27175677e-01 -3.23102683e-01 5.53761959e-01
-5.00947893e-01 4.74427372e-01 -1.28547192e-01 -1.26768982e+00
-3.80535781e-01 -8.13412845e-01 6.86394751e-01 -5.35662398e-02
-5.55324778e-02 -6.58814132e-01 -3.45104575e-01 5.76352835e-01
2.40878403e-01 3.03277373e-01 7.22353697e-01 -3.65398586e-01
-5.06307125e-01 -5.24691284e-01 -3.22069019e-01 1.06570698e-01
9.76993423e-03 -2.09612280e-01 -1.76206321e-01 -4.73973721e-01
1.61406368e-01 2.61092912e-02 6.76190913e-01 6.69601083e-01
1.83901322e+00 -7.35177219e-01 -2.02087283e-01 7.69730747e-01
1.45175898e+00 9.61496308e-02 3.45024526e-01 -2.34176591e-02
3.73793542e-01 2.30193421e-01 9.93870556e-01 1.40044808e+00
1.17287427e-01 3.84441197e-01 2.42596552e-01 -9.64736380e-03
5.42325377e-01 3.67832445e-02 3.81681800e-01 8.12114120e-01
-3.77937704e-02 -1.48537949e-01 -6.68487310e-01 5.77599823e-01
-1.77278066e+00 -6.95355296e-01 -2.22382918e-01 2.88454437e+00
9.75224555e-01 1.60755679e-01 2.90248752e-01 2.43639871e-01
6.81846857e-01 1.82083741e-01 -6.43071830e-01 -4.05884981e-01
-3.70834433e-02 7.69160092e-01 1.06368291e+00 5.20153224e-01
-7.62701392e-01 6.39132977e-01 5.22339678e+00 1.04467738e+00
-1.13046789e+00 -4.56145294e-02 6.81358159e-01 -6.83101892e-01
-1.81436390e-01 7.80929625e-02 -7.25488722e-01 6.96837902e-01
9.92055237e-01 -4.20947462e-01 5.64043164e-01 1.00951111e+00
6.64241135e-01 -4.04040843e-01 -8.28014553e-01 1.13490129e+00
1.30994627e-02 -1.23143971e+00 -6.01246178e-01 2.09270850e-01
7.08648026e-01 -2.38803774e-01 2.65364379e-01 4.71488498e-02
1.53177530e-01 -1.09632361e+00 3.85276288e-01 1.38370499e-01
9.68483031e-01 -9.97009754e-01 5.07540286e-01 4.36659783e-01
-1.44911814e+00 -3.12656462e-01 -3.02110851e-01 -4.27225567e-02
7.85190091e-02 9.92665112e-01 -7.18532085e-01 3.37598145e-01
7.38095343e-01 3.86904061e-01 -1.25975430e-01 8.82609725e-01
1.39018938e-01 9.18554008e-01 -7.53016949e-01 -2.57543713e-01
1.21187538e-01 -4.33877707e-01 4.81168121e-01 1.03403735e+00
4.50271785e-01 5.39775252e-01 2.88880885e-01 1.84934616e-01
-2.37594321e-01 5.36326528e-01 -1.29928902e-01 1.07054353e-01
7.56054282e-01 8.10389042e-01 -7.09388316e-01 -8.31205398e-02
-5.34390330e-01 6.35477304e-01 2.53555477e-01 3.11625421e-01
-9.61300254e-01 -4.16460842e-01 4.29431915e-01 4.38295662e-01
7.89992034e-01 -4.63623792e-01 -7.53045857e-01 -1.04185653e+00
4.96406645e-01 -8.32655847e-01 6.38666034e-01 5.44061027e-02
-1.01331294e+00 1.08205408e-01 -2.30506897e-01 -1.10597527e+00
-5.98444901e-02 -2.18308330e-01 -1.80247679e-01 5.62707484e-01
-1.21770716e+00 -3.60004514e-01 -1.78026363e-01 6.05545938e-01
4.07665133e-01 -6.16675541e-02 5.13629556e-01 2.99718231e-01
-9.87457931e-01 9.44091797e-01 7.96772122e-01 -1.46151319e-01
4.41940874e-01 -7.85336733e-01 -2.78496087e-01 8.05411935e-01
-2.04193503e-01 6.51177168e-01 9.09412026e-01 -5.30994117e-01
-1.65529740e+00 -8.89990747e-01 7.67330527e-01 2.78612882e-01
7.04462826e-01 -1.81851864e-01 -7.84238756e-01 4.58948880e-01
-3.89862061e-01 4.66163397e-01 1.03826571e+00 1.86462983e-01
-2.67747283e-01 -6.50233090e-01 -1.24223578e+00 5.37527502e-01
7.85718381e-01 -1.03460312e-01 3.10901195e-01 3.22821856e-01
3.56415123e-01 -3.45499337e-01 -1.25452983e+00 3.54835421e-01
6.20122850e-01 -6.36294127e-01 7.19157159e-01 -4.43513066e-01
1.25141934e-01 -1.37378603e-01 -3.92596751e-01 -6.18900001e-01
4.96636778e-02 -9.66121376e-01 -4.33504820e-01 6.72878623e-01
6.37235165e-01 -6.79472208e-01 1.05775988e+00 7.48116910e-01
4.00458843e-01 -9.93822336e-01 -1.22039592e+00 -6.57448649e-01
-1.65852364e-02 -5.47788382e-01 2.51003891e-01 7.86828458e-01
4.14336771e-02 -1.37493342e-01 -4.95205611e-01 2.17738345e-01
7.34589517e-01 2.62960672e-01 8.05524588e-01 -1.06928670e+00
-6.76720560e-01 -2.01292619e-01 -4.45956141e-02 -1.25662601e+00
-1.68209717e-01 -5.08534610e-01 -1.68791816e-01 -8.17232609e-01
2.46108294e-01 -9.46977675e-01 -4.81634736e-01 2.57733226e-01
-3.68283987e-01 -9.17364135e-02 3.67528051e-02 4.78869826e-02
-6.80571496e-01 4.80170637e-01 7.98024178e-01 2.28019133e-01
-4.19973463e-01 3.07677090e-01 -7.49555349e-01 6.04347825e-01
6.33668125e-01 -7.00171947e-01 -2.20954463e-01 -3.99694145e-01
5.57226837e-01 8.07814538e-01 -5.74788526e-02 -4.35450494e-01
1.99556664e-01 -5.14846623e-01 6.71363249e-02 -4.65525240e-01
1.51828274e-01 -7.85254776e-01 -6.92762211e-02 3.43763053e-01
-3.78325731e-01 -7.11897612e-02 -3.21566276e-02 9.59294856e-01
-8.88233706e-02 -2.65464813e-01 7.91958034e-01 2.17043206e-01
4.78328951e-02 4.90442872e-01 -1.24293357e-01 1.79621577e-01
1.05943680e+00 -1.63321599e-01 2.30320886e-01 -5.53054154e-01
-4.04041111e-01 4.21740711e-01 1.50402024e-01 -2.00591862e-01
5.71753740e-01 -8.08224320e-01 -7.77730465e-01 1.15862675e-01
-1.75007507e-01 2.37668231e-01 3.12861741e-01 1.23871493e+00
-4.37752932e-01 4.40433860e-01 6.49599791e-01 -4.66688901e-01
-1.17862630e+00 4.27629203e-01 -1.82555571e-01 -3.62553984e-01
-5.00930846e-01 1.08821595e+00 -1.50301754e-01 1.67155191e-01
2.87119001e-01 -1.94965526e-01 3.06653827e-01 -3.88481878e-02
3.83661330e-01 9.45281506e-01 -8.85397941e-03 -1.94848537e-01
-3.37625325e-01 2.40878418e-01 -8.50959569e-02 -5.83606325e-02
1.48678756e+00 -3.15703869e-01 -1.52394414e-01 4.63319808e-01
1.37595189e+00 4.21310216e-01 -1.34903789e+00 -4.44805115e-01
-5.65106496e-02 -6.50409460e-01 7.68766366e-03 -6.53824508e-02
-1.25284564e+00 5.36432981e-01 3.74047875e-01 2.37050965e-01
1.32312500e+00 -2.09150910e-01 9.34951425e-01 2.93395311e-01
6.42048120e-01 -1.24449158e+00 -2.17957914e-01 4.13197547e-01
5.22134542e-01 -1.25067937e+00 3.90841424e-01 -7.12881029e-01
-4.72657681e-01 9.65251982e-01 2.65510798e-01 -4.83433217e-01
6.43472195e-01 9.39237922e-02 -6.18667126e-01 1.02975577e-01
-5.60357690e-01 3.10141165e-02 -8.38836655e-02 -5.03217045e-04
4.83365387e-01 2.21119583e-01 -6.82022929e-01 9.52352464e-01
-2.46296912e-01 7.79663473e-02 3.79548192e-01 9.53829885e-01
-4.67755258e-01 -8.69842172e-01 -3.51263970e-01 7.13803113e-01
-8.10023487e-01 -1.94610104e-01 2.58690119e-01 6.42428815e-01
-2.76337355e-01 1.02529407e+00 -2.49338355e-02 1.29167229e-01
1.13949262e-01 -1.80851519e-01 3.44034493e-01 -3.72009456e-01
-3.63197356e-01 3.79320055e-01 -5.01466468e-02 -5.68123937e-01
-1.24010161e-01 -9.10958052e-01 -1.38580203e+00 -4.36104387e-01
-3.08672965e-01 3.03803980e-01 5.55638850e-01 9.19155002e-01
2.88731635e-01 1.52886361e-01 1.30582571e+00 -1.47092015e-01
-6.00865662e-01 -5.57865441e-01 -7.75730729e-01 -7.36560374e-02
3.26027423e-01 -5.13052762e-01 -5.11062562e-01 2.73321127e-03] | [6.679098129272461, 4.514091968536377] |
efba3ee7-0fcb-4767-bd14-92afbe7e2ec6 | atom-search-optimization-with-simulated | 2005.08642 | null | https://arxiv.org/abs/2005.08642v1 | https://arxiv.org/pdf/2005.08642v1.pdf | Atom Search Optimization with Simulated Annealing -- a Hybrid Metaheuristic Approach for Feature Selection | 'Hybrid meta-heuristics' is one of the most interesting recent trends in the field of optimization and feature selection (FS). In this paper, we have proposed a binary variant of Atom Search Optimization (ASO) and its hybrid with Simulated Annealing called ASO-SA techniques for FS. In order to map the real values used by ASO to the binary domain of FS, we have used two different transfer functions: S-shaped and V-shaped. We have hybridized this technique with a local search technique called, SA We have applied the proposed feature selection methods on 25 datasets from 4 different categories: UCI, Handwritten digit recognition, Text, non-text separation, and Facial emotion recognition. We have used 3 different classifiers (K-Nearest Neighbor, Multi-Layer Perceptron and Random Forest) for evaluating the strength of the selected featured by the binary ASO, ASO-SA and compared the results with some recent wrapper-based algorithms. The experimental results confirm the superiority of the proposed method both in terms of classification accuracy and number of selected features. | ['Kushal Kanti Ghosh', 'Soulib Ghosh', 'Ram Sarkar', 'Suman Kumar Bera', 'Ritam Guha'] | 2020-05-10 | null | null | null | null | ['facial-emotion-recognition', 'handwritten-digit-recognition'] | ['computer-vision', 'computer-vision'] | [ 2.58728296e-01 -2.61420250e-01 -2.68095970e-01 -4.28835452e-01
-3.21146965e-01 -2.72975177e-01 4.36639607e-01 3.62008631e-01
-4.88228500e-01 1.09777331e+00 2.87697781e-02 2.59631108e-02
-6.37457490e-01 -8.43996644e-01 -2.30958849e-01 -8.03708792e-01
-1.30691960e-01 7.03569233e-01 3.18124920e-01 -4.18230087e-01
8.23369145e-01 7.31048465e-01 -2.11343908e+00 4.09822494e-01
9.34252203e-01 1.25346303e+00 5.96083328e-02 4.91905242e-01
-2.96108782e-01 2.22238272e-01 -6.29856884e-01 5.08727953e-02
4.07738000e-01 -4.58982021e-01 -8.31496716e-01 1.50290746e-02
-4.86076891e-01 6.26666427e-01 5.10520041e-01 9.16622400e-01
5.87363124e-01 4.98191386e-01 7.76780784e-01 -1.41032934e+00
-1.81335554e-01 5.62854171e-01 -6.21853828e-01 2.53316969e-01
6.00105762e-01 -2.33493194e-01 5.77055573e-01 -4.94723558e-01
8.27355981e-01 1.37308872e+00 4.78031516e-01 3.03272814e-01
-8.91005516e-01 -3.85349631e-01 -1.96834996e-01 6.55367494e-01
-1.43150735e+00 -2.95612693e-01 7.93218732e-01 -1.58077925e-01
1.35063243e+00 7.23262310e-01 7.00452685e-01 4.53291744e-01
4.29477513e-01 7.36374438e-01 1.61355603e+00 -1.11706269e+00
6.34135365e-01 4.72851276e-01 3.83261055e-01 8.92302394e-01
2.04886898e-01 1.01748168e-01 -6.49636805e-01 -6.60227716e-01
-1.74845174e-01 -2.76263505e-01 -7.06539862e-03 -1.34665176e-01
-7.65553236e-01 9.46327448e-01 1.12234563e-01 6.52497113e-01
-6.70479834e-01 -3.88108820e-01 3.44608128e-01 2.48292148e-01
1.35771558e-01 6.94203675e-01 -5.66837370e-01 6.61984086e-02
-6.72794104e-01 3.24592113e-01 8.12870681e-01 4.91938084e-01
5.34546971e-01 -4.28554676e-02 -1.69255823e-01 8.59585166e-01
4.20707464e-01 2.07814634e-01 1.12529063e+00 1.45742008e-02
1.80925190e-01 9.82526958e-01 2.23211020e-01 -9.31780696e-01
-4.94442314e-01 -4.15001780e-01 -3.26176882e-01 5.11872530e-01
-1.99508697e-01 -2.96100706e-01 -1.07726014e+00 1.14656651e+00
5.93808115e-01 -3.65040481e-01 3.29604328e-01 6.35072768e-01
6.84173465e-01 8.31645548e-01 -1.87455654e-01 -4.57276613e-01
1.02888298e+00 -1.22238410e+00 -6.89447820e-01 2.21535072e-01
1.96991205e-01 -8.66412103e-01 5.14372706e-01 7.82186866e-01
-8.87168229e-01 -4.71397728e-01 -1.27855241e+00 6.43668354e-01
-9.07785892e-01 1.16875738e-01 6.52384996e-01 9.08803999e-01
-7.83644736e-01 8.13257337e-01 -6.44740760e-01 -4.30771291e-01
2.48099834e-01 7.80870557e-01 -3.65606785e-01 1.78184792e-01
-9.64639485e-01 9.82333660e-01 6.49595022e-01 3.89539264e-02
-2.82916248e-01 2.66496778e-01 -4.41875786e-01 1.12862930e-01
2.27790669e-01 -3.60486060e-01 5.92962027e-01 -1.29869545e+00
-1.62503028e+00 5.78591526e-01 -2.57325768e-01 -3.36872667e-01
1.48083627e-01 2.24970102e-01 -6.23076677e-01 -4.45989221e-02
-1.73083931e-01 3.62203032e-01 8.34633350e-01 -9.94218409e-01
-7.67306626e-01 -6.66401505e-01 -4.24358487e-01 4.71624702e-01
-1.27610385e-01 2.93058932e-01 2.82176673e-01 -4.59965438e-01
2.40272105e-01 -9.50697303e-01 -1.80947855e-01 -6.17029309e-01
-5.00901103e-01 -3.36009622e-01 7.47271657e-01 -2.94337094e-01
1.34800494e+00 -1.91641223e+00 2.93329239e-01 9.24630344e-01
-7.79304147e-01 3.58947188e-01 1.28713667e-01 5.71232498e-01
-1.88081607e-01 9.79908258e-02 -2.92223603e-01 -1.39140198e-02
-1.01963237e-01 1.36507317e-01 2.62066841e-01 2.98953056e-01
-1.40624776e-01 3.15575838e-01 -4.37542766e-01 -6.37549281e-01
1.26786202e-01 2.16795713e-01 -1.55814797e-01 -8.75834078e-02
-1.32817030e-01 -1.30469367e-01 -7.36706853e-01 1.11372793e+00
5.57500720e-01 3.51702124e-01 4.07442115e-02 8.01239684e-02
-3.21721047e-01 -2.63845414e-01 -1.50360453e+00 1.08734286e+00
-2.35346064e-01 1.16043933e-01 -2.27402940e-01 -8.79150450e-01
1.32091761e+00 1.90610975e-01 3.81095290e-01 -6.30796254e-01
3.84517729e-01 4.18578237e-01 2.90151928e-02 -6.19859576e-01
2.47056529e-01 1.03598922e-01 3.20396453e-01 3.71067911e-01
4.27879207e-02 1.80261165e-01 2.32095316e-01 -7.57983088e-01
8.01122248e-01 1.98696271e-01 7.72300541e-01 -4.30133343e-01
1.02016973e+00 4.78980899e-01 4.84179914e-01 6.09394073e-01
-1.20472901e-01 2.89688796e-01 4.11488861e-01 -5.19678593e-01
-6.20166600e-01 -2.99852520e-01 -2.55510777e-01 9.36633527e-01
-1.38244182e-01 -1.41304918e-03 -7.41770864e-01 -7.20571816e-01
1.22052461e-01 8.64123106e-01 -5.99271476e-01 -8.14400539e-02
-3.69999319e-01 -1.15428758e+00 2.19026849e-01 -1.40377641e-01
6.12384975e-01 -1.36972785e+00 -1.00583804e+00 3.10525268e-01
3.57362956e-01 -3.70473146e-01 2.81653672e-01 7.69166827e-01
-8.26082647e-01 -7.59781122e-01 -3.39239866e-01 -6.24617636e-01
5.25735080e-01 -1.84543744e-01 7.32130766e-01 1.82518139e-01
-6.24724388e-01 -2.30207667e-02 -1.06259668e+00 -7.77460217e-01
-2.34114587e-01 2.13199243e-01 -7.42476135e-02 3.76611084e-01
2.58589894e-01 -3.86765808e-01 -2.94542134e-01 2.90424615e-01
-6.74428463e-01 -4.49306428e-01 7.59523869e-01 1.00074744e+00
6.03725255e-01 4.89313215e-01 3.96263927e-01 -8.79888833e-01
8.09049010e-01 -5.31478465e-01 -5.73501289e-01 5.44445872e-01
-9.59847391e-01 3.78034264e-01 5.77036083e-01 -1.17960788e-01
-8.45366061e-01 3.00131381e-01 -5.01699597e-02 -1.06319547e-01
-1.74173683e-01 5.24624705e-01 -1.16813362e-01 -5.38942039e-01
8.84134650e-01 2.69553781e-01 7.92713612e-02 -3.59787852e-01
-3.56685191e-01 1.01527524e+00 -4.74574789e-02 -4.91559088e-01
2.34649748e-01 2.98014730e-01 2.67980069e-01 -5.92896879e-01
7.41234943e-02 -3.87926847e-01 -2.59691358e-01 -1.15322620e-01
8.73429835e-01 4.09694575e-02 -6.38024867e-01 6.28566742e-01
-7.47178733e-01 2.44095162e-01 -2.51009949e-02 2.95478493e-01
-2.55461544e-01 1.96327299e-01 2.75076628e-01 -1.24559259e+00
-7.87717581e-01 -1.36555994e+00 7.19564855e-01 5.98568022e-01
-1.06778309e-01 -5.51611006e-01 -5.33351973e-02 5.25106713e-02
4.20357406e-01 7.44087875e-01 1.21054757e+00 -1.06832325e+00
7.25423619e-02 -3.85331631e-01 4.74991530e-01 1.54941574e-01
1.36271685e-01 1.66589394e-01 -7.21569180e-01 -6.59262687e-02
6.54445291e-02 -2.34185189e-01 6.49186134e-01 3.51083636e-01
9.40089762e-01 -2.05859393e-01 -4.99771059e-01 5.26452422e-01
1.67425430e+00 1.09492683e+00 4.88584071e-01 1.09518421e+00
-1.86693564e-01 3.20926130e-01 1.16538811e+00 8.34608674e-01
-3.11756730e-01 8.34369302e-01 3.46366823e-01 1.95883855e-01
3.39198083e-01 3.30264121e-01 1.87139645e-01 6.01372756e-02
-3.20382476e-01 -4.31358218e-01 -9.34701562e-01 2.37058342e-01
-1.80348611e+00 -7.84283757e-01 1.57149166e-01 2.15332294e+00
4.58595723e-01 3.13701332e-01 2.21018672e-01 7.47735262e-01
9.47085857e-01 -1.63429871e-01 -2.61820644e-01 -1.18950605e+00
-2.58465052e-01 6.60186052e-01 7.14672267e-01 4.74659413e-01
-1.09573948e+00 6.14008665e-01 5.31632280e+00 1.00267386e+00
-1.37954199e+00 -1.70116693e-01 5.73828042e-01 -5.89634441e-02
-5.84100969e-02 -5.37835211e-02 -7.86752820e-01 5.38932920e-01
7.89258718e-01 -1.71488188e-02 7.22720742e-01 1.02370489e+00
-2.63632778e-02 -5.74130952e-01 -5.56474984e-01 8.59913528e-01
1.54265285e-01 -1.00588977e+00 6.95913509e-02 -3.04588735e-01
8.01032901e-01 -3.85752290e-01 -2.74488200e-02 -9.11461785e-02
-1.47336364e-01 -1.17582726e+00 6.79415226e-01 5.56113839e-01
3.14927965e-01 -1.18537235e+00 1.07703149e+00 1.95068985e-01
-9.60772038e-01 -4.09194678e-01 -3.06032717e-01 1.84402332e-01
-2.24248976e-01 4.13298041e-01 -8.77362669e-01 8.27676833e-01
8.20573270e-01 -6.77852109e-02 -7.20436037e-01 1.19414222e+00
3.18980873e-01 1.89494655e-01 -6.31784678e-01 -8.22676301e-01
4.58405584e-01 -1.74162313e-01 7.34900296e-01 8.57553542e-01
2.01417238e-01 7.75742531e-02 -9.03990045e-02 3.09137136e-01
5.73276997e-01 5.15580058e-01 -2.92391270e-01 -4.79826368e-02
5.98943293e-01 1.04141533e+00 -1.11969662e+00 -1.80743545e-01
1.32171631e-01 6.54869974e-01 -1.35214508e-01 -4.83739153e-02
-6.54280543e-01 -1.04248798e+00 -4.14934829e-02 -3.96671236e-01
4.27719384e-01 2.19742134e-01 -3.57639551e-01 -4.33872938e-01
-1.23205766e-01 -1.10830700e+00 6.49757266e-01 -5.85215867e-01
-8.59587908e-01 1.20211112e+00 2.43371099e-01 -9.35474038e-01
-2.34805956e-01 -5.79872787e-01 -5.12228668e-01 9.00480866e-01
-9.67766047e-01 -8.77040863e-01 -1.69338614e-01 7.08081961e-01
5.90135455e-01 -8.34710240e-01 6.89006329e-01 -4.02110666e-02
-4.93652135e-01 5.76673031e-01 3.84537786e-01 -5.81005096e-01
4.20136303e-01 -9.98229682e-01 -1.40211776e-01 4.19560343e-01
-7.23587945e-02 3.00717533e-01 9.08460617e-01 -6.64185345e-01
-1.38674128e+00 -5.19088268e-01 8.73291433e-01 3.02435666e-01
-3.26760225e-02 9.03893411e-02 -4.03621644e-01 1.85567573e-01
1.87000573e-01 -1.53155252e-01 5.59732378e-01 -1.21762648e-01
4.01644915e-01 -2.39053220e-01 -1.89218128e+00 1.30513981e-01
6.50770366e-01 3.43858302e-01 -5.22345603e-01 4.94016051e-01
1.18188225e-02 -3.58953267e-01 -7.03647852e-01 4.85453784e-01
6.09297335e-01 -1.37339389e+00 7.55056560e-01 -6.03592575e-01
1.03937704e-02 -2.22776383e-01 -3.62026691e-01 -1.31004655e+00
-2.87576199e-01 -4.94215280e-01 3.53884190e-01 1.15890789e+00
6.04099035e-01 -1.01468837e+00 8.20589781e-01 3.45824003e-01
-9.49651971e-02 -1.48455811e+00 -1.01206601e+00 -4.58554894e-01
-5.41303515e-01 2.59779572e-01 1.03994417e+00 8.45266938e-01
-1.11222744e-01 -2.57247128e-02 3.35933790e-02 -1.85750842e-01
2.56945401e-01 3.33868086e-01 2.09822565e-01 -1.12065148e+00
-5.16002893e-01 -5.64690351e-01 -7.79630065e-01 2.11449176e-01
-5.13129532e-02 -7.15653658e-01 -2.87197232e-01 -1.19059300e+00
-3.56962457e-02 -7.07390606e-01 -4.99305844e-01 6.10625505e-01
8.08700845e-02 -1.18790016e-01 7.16722235e-02 -6.85585290e-02
-7.13102147e-02 3.36332947e-01 7.12276638e-01 1.26031086e-01
-6.51959419e-01 2.60330260e-01 -3.14877272e-01 5.54875016e-01
7.54173517e-01 -7.66692281e-01 -2.51059592e-01 1.31192371e-01
1.12649366e-01 1.35114655e-01 -3.00760686e-01 -1.13432491e+00
4.12443057e-02 -4.35284466e-01 5.65033317e-01 -5.88259220e-01
1.71994552e-01 -8.47130060e-01 3.34347278e-01 7.27762163e-01
-2.55666167e-01 5.45728385e-01 8.05163309e-02 2.67557770e-01
-3.68710786e-01 -9.35375631e-01 9.79894221e-01 -3.26683581e-01
-7.50498354e-01 -3.62430483e-01 -2.79763490e-01 -6.37651682e-01
1.44601572e+00 -6.61680937e-01 -1.36152342e-01 9.12042558e-02
-7.23708451e-01 3.47924680e-02 2.63563663e-01 3.15886773e-02
6.57844603e-01 -9.08377469e-01 -3.41614395e-01 1.23514779e-01
-1.65893525e-01 -4.15728390e-01 -4.66352068e-02 6.44460678e-01
-9.18812931e-01 6.31537855e-01 -8.53996396e-01 -2.69853085e-01
-1.70827591e+00 5.60105741e-01 3.00694227e-01 -3.03440601e-01
-1.00768536e-01 8.52214634e-01 -1.13732767e+00 -3.08784366e-01
2.04137027e-01 -4.32799011e-02 -7.94436216e-01 6.80383593e-02
4.08790521e-02 9.71901655e-01 4.77413207e-01 -4.99172062e-01
-7.06850708e-01 6.91151738e-01 1.59320280e-01 -3.60827893e-01
1.40582871e+00 4.54946220e-01 -4.98115718e-01 1.38411313e-01
1.01109469e+00 1.32389292e-01 -1.87805980e-01 3.51428062e-01
1.15758106e-01 -5.83029509e-01 1.22100130e-01 -1.21711910e+00
-7.84148932e-01 2.42040977e-01 1.09351635e+00 2.01639295e-01
1.53196359e+00 -6.48209572e-01 4.54815060e-01 6.41964018e-01
4.67847228e-01 -1.31083643e+00 -5.87539256e-01 2.41913468e-01
8.48235667e-01 -9.76027131e-01 3.01151514e-01 -3.14679742e-01
-8.42931390e-01 1.36554992e+00 4.47453439e-01 -1.99097171e-01
6.33659244e-01 1.49437398e-01 -2.58624047e-01 6.58917502e-02
-8.91512752e-01 -1.76020265e-01 1.82897002e-01 4.28912014e-01
2.20423818e-01 -1.13475427e-01 -1.00052869e+00 5.26333213e-01
-2.79253930e-01 1.36996016e-01 1.99052751e-01 1.46188903e+00
-6.64626241e-01 -1.33476412e+00 -8.53815496e-01 6.51615739e-01
-3.79497439e-01 2.18111396e-01 -7.09930003e-01 8.51016402e-01
5.32929599e-01 9.10465479e-01 -3.78219843e-01 -5.91591895e-01
1.08187594e-01 3.36944759e-01 6.10915780e-01 -2.31031448e-01
-1.23231113e+00 -2.36453071e-01 3.61087292e-01 -3.91697407e-01
-4.03273284e-01 -8.23094070e-01 -1.29999113e+00 1.06639259e-01
-5.53615808e-01 7.86873579e-01 1.25627124e+00 9.31517482e-01
3.08597803e-01 1.63324490e-01 8.50924671e-01 -7.22139537e-01
-6.26856565e-01 -8.56071293e-01 -4.22849894e-01 6.87732175e-02
7.86498263e-02 -1.09167016e+00 -2.63023794e-01 -5.20823240e-01] | [7.856736660003662, 3.726134777069092] |
ebb5f802-9cc6-4875-b14c-62c2a5538902 | learning-multi-scale-representations-for | 1408.2938 | null | http://arxiv.org/abs/1408.2938v1 | http://arxiv.org/pdf/1408.2938v1.pdf | Learning Multi-Scale Representations for Material Classification | The recent progress in sparse coding and deep learning has made unsupervised
feature learning methods a strong competitor to hand-crafted descriptors. In
computer vision, success stories of learned features have been predominantly
reported for object recognition tasks. In this paper, we investigate if and how
feature learning can be used for material recognition. We propose two
strategies to incorporate scale information into the learning procedure
resulting in a novel multi-scale coding procedure. Our results show that our
learned features for material recognition outperform hand-crafted descriptors
on the FMD and the KTH-TIPS2 material classification benchmarks. | ['Wenbin Li', 'Mario Fritz'] | 2014-08-13 | null | null | null | null | ['material-classification', 'material-recognition'] | ['computer-vision', 'computer-vision'] | [ 2.43525192e-01 -4.12216514e-01 -3.13499480e-01 -4.56436485e-01
-7.94903159e-01 -4.65099365e-01 7.81782568e-01 2.47294143e-01
-1.75172597e-01 6.10435188e-01 3.70530009e-01 1.44163311e-01
-3.79845679e-01 -6.74263060e-01 -7.28247046e-01 -7.58174002e-01
-1.80541654e-03 3.39634329e-01 1.79397717e-01 8.76836702e-02
5.68333328e-01 8.24209869e-01 -1.68282712e+00 5.78854442e-01
2.65642375e-01 1.45996225e+00 2.11703375e-01 6.77281439e-01
-3.65478456e-01 9.09008086e-01 -5.54596603e-01 -8.44425187e-02
4.64075923e-01 -1.78762108e-01 -7.36753702e-01 5.72423220e-01
7.82799542e-01 -1.45873800e-01 -7.25406706e-01 6.08671904e-01
3.86910021e-01 2.06512153e-01 1.21373856e+00 -9.37415242e-01
-7.22814977e-01 3.61843884e-01 -4.82413232e-01 2.22458124e-01
6.08235955e-01 -2.50099719e-01 1.20278049e+00 -1.10029864e+00
8.18170547e-01 1.03624547e+00 8.10886860e-01 1.46826535e-01
-1.08123553e+00 -2.15817764e-01 -1.31974578e-01 4.41728324e-01
-1.50558066e+00 -5.12932718e-01 1.11640346e+00 -6.66538835e-01
1.19629383e+00 1.82158247e-01 6.48288429e-01 1.01754797e+00
2.35762998e-01 1.11137617e+00 1.06686878e+00 -8.01335275e-01
2.49157995e-01 1.60390496e-01 2.40176078e-02 8.34777296e-01
4.46830750e-01 2.70375907e-01 -6.87998295e-01 -1.72328427e-02
9.42297935e-01 3.04420471e-01 1.43536165e-01 -6.30419552e-01
-1.12938881e+00 1.15809608e+00 5.12893856e-01 5.90375125e-01
-3.69840562e-01 1.63313389e-01 5.86939275e-01 4.69632655e-01
3.85968328e-01 6.26171231e-01 -4.31850463e-01 -3.11373770e-01
-1.22570682e+00 1.53155595e-01 8.76668572e-01 9.56296682e-01
6.65378034e-01 3.25565785e-01 -3.76017243e-01 1.07742989e+00
4.47924510e-02 6.31027758e-01 4.69141632e-01 -5.96807778e-01
5.19548059e-02 7.06488013e-01 -1.85958639e-01 -1.26767111e+00
-4.16403681e-01 -2.98034668e-01 -7.64151871e-01 4.72471006e-02
3.70809883e-01 3.08340847e-01 -9.84685898e-01 7.56041646e-01
-7.39352629e-02 -5.09158969e-02 -1.76358312e-01 7.42301762e-01
8.37289929e-01 4.18082237e-01 -3.75278473e-01 2.28819072e-01
8.99726689e-01 -1.02343023e+00 -3.62692147e-01 1.49807185e-01
5.43846488e-01 -9.96304333e-01 6.06221139e-01 6.68574154e-01
-8.06982696e-01 -6.45861268e-01 -1.02226150e+00 3.90045233e-02
-4.45507646e-01 3.50268364e-01 1.13810480e+00 5.24435043e-01
-7.49039233e-01 7.54667997e-01 -6.77746832e-01 -3.68961185e-01
9.01926160e-01 4.51709181e-01 -4.82180774e-01 -4.06075567e-01
-6.04158163e-01 7.14302242e-01 2.45378464e-01 -1.54551968e-01
-1.00832331e+00 -3.55509877e-01 -7.73223937e-01 -1.62956752e-02
2.28271186e-01 -4.16874290e-01 8.40547323e-01 -8.55707824e-01
-1.58173418e+00 8.87277961e-01 9.40380916e-02 -5.73736310e-01
8.26915428e-02 -2.05760464e-01 -1.92688927e-01 4.92928594e-01
-1.20012514e-01 5.69678366e-01 1.29999459e+00 -1.10222340e+00
-4.24933583e-01 1.14762083e-01 -4.97986078e-02 -1.57064408e-01
-5.63586831e-01 -5.22201248e-02 -5.98820895e-02 -8.48697603e-01
6.21701889e-02 -8.65991175e-01 -7.85348266e-02 2.51811385e-01
-3.55369151e-01 -3.86244714e-01 7.77330160e-01 -3.50050420e-01
7.87742913e-01 -2.20377612e+00 2.11733133e-01 3.41496557e-01
4.70054075e-02 1.92077845e-01 -2.11018100e-01 5.36231101e-01
1.83928788e-01 -3.85192990e-01 5.40309474e-02 -2.30620131e-01
1.10687368e-01 2.93087959e-01 -8.55599865e-02 6.78017259e-01
3.44731867e-01 9.16566193e-01 -7.39894211e-01 -5.82742870e-01
4.56419438e-01 6.18709385e-01 -7.17532992e-01 1.82430074e-01
6.28950596e-02 7.73654431e-02 -4.38556373e-01 1.01849830e+00
3.79869044e-01 -2.96274245e-01 -2.89483249e-01 -2.12354109e-01
-1.00126877e-01 3.25698733e-01 -9.83663917e-01 1.97102475e+00
-4.50396210e-01 9.66161311e-01 -2.82238424e-01 -1.27214515e+00
1.21608031e+00 1.77428022e-01 6.95404470e-01 -6.20364606e-01
1.18213311e-01 1.53018326e-01 -2.76036382e-01 -3.85150492e-01
3.97366494e-01 -1.75734565e-01 -7.76285604e-02 3.58695015e-02
7.50400007e-01 -3.50797474e-01 2.55986571e-01 1.61784977e-01
1.35980749e+00 -1.17002197e-01 3.16623718e-01 -3.67873132e-01
6.66820407e-01 -6.15960471e-02 1.51545316e-01 8.77288222e-01
8.88093784e-02 8.27645600e-01 1.40445203e-01 -6.40798092e-01
-1.02602339e+00 -1.00197446e+00 -3.42693567e-01 1.23337674e+00
-4.04405713e-01 -5.63409865e-01 -1.80007577e-01 -6.88747942e-01
4.03023481e-01 5.01480512e-02 -7.00013876e-01 -1.99369416e-01
-3.69288206e-01 6.71211332e-02 3.26578975e-01 7.97996044e-01
1.47465050e-01 -1.29151499e+00 -3.26030761e-01 1.85859844e-01
5.58882296e-01 -8.28301728e-01 -3.87805969e-01 7.52848923e-01
-8.82843792e-01 -9.66421783e-01 -1.01023698e+00 -1.15723646e+00
7.19064295e-01 3.87614250e-01 1.03934252e+00 -4.86582108e-02
-7.51151085e-01 6.05705559e-01 -7.76725292e-01 -3.43235254e-01
-1.37871623e-01 2.89035499e-01 2.36677006e-03 -7.97329918e-02
4.15727317e-01 -5.70010185e-01 -5.15305102e-01 1.92672312e-01
-8.78653049e-01 -3.43154639e-01 9.94003236e-01 1.10641325e+00
5.61375141e-01 -9.76892263e-02 4.93870467e-01 -8.02164197e-01
5.59383750e-01 -1.44129992e-01 -2.60894567e-01 7.52357617e-02
-3.93825054e-01 3.22041512e-01 5.68747163e-01 -4.20439571e-01
-6.16250455e-01 4.32315677e-01 -9.34975222e-02 -6.20648921e-01
-3.73068154e-01 6.21339679e-01 6.52755946e-02 -7.01216161e-01
7.26516187e-01 4.80637014e-01 -1.86623126e-01 -5.79466701e-01
3.59221637e-01 6.44451737e-01 3.02430630e-01 -8.15113544e-01
8.25252593e-01 3.71200711e-01 3.71406265e-02 -1.07632816e+00
-1.05004597e+00 -8.90460432e-01 -8.49129021e-01 -2.31516942e-01
6.11296296e-01 -7.63146102e-01 -1.58435330e-01 3.12699378e-01
-8.95320773e-01 -1.35988295e-01 -8.96902800e-01 4.30912167e-01
-9.29289818e-01 2.18600079e-01 -5.63165426e-01 -4.67942059e-01
-2.16513649e-01 -7.92816699e-01 1.34455931e+00 2.17494249e-01
-1.28635630e-01 -1.12062788e+00 3.26621085e-01 1.05705753e-01
7.93975890e-01 -6.62802625e-03 6.12088203e-01 -7.97129452e-01
-2.19893292e-01 -6.46156609e-01 -2.77103394e-01 5.23704231e-01
4.23983097e-01 -1.22467212e-01 -9.73312318e-01 -4.92608935e-01
-1.00533783e-01 -6.29539967e-01 1.19780195e+00 3.49800497e-01
1.30897975e+00 -1.15818903e-01 -1.38647899e-01 6.60659075e-01
1.36272025e+00 5.77947274e-02 5.10192990e-01 3.83129269e-01
6.13643825e-01 2.02426970e-01 1.97946787e-01 7.23655701e-01
-1.29114613e-01 5.72364509e-01 -8.22600443e-03 1.68918088e-01
-4.45245981e-01 -2.32287973e-01 1.99337035e-01 1.19553316e+00
-2.09477648e-01 1.04623184e-01 -7.57851779e-01 5.75414240e-01
-1.54908633e+00 -9.88520503e-01 4.56829488e-01 1.77283180e+00
5.92663586e-01 2.57463664e-01 2.05379143e-01 5.61916769e-01
3.56014043e-01 8.12000707e-02 -3.49869400e-01 -4.09892023e-01
-2.10073099e-01 6.80666983e-01 4.57895190e-01 -1.20723575e-01
-1.23996663e+00 7.39215016e-01 7.40221643e+00 9.17992830e-01
-1.23422658e+00 -1.14765115e-01 1.19755849e-01 -1.98234711e-02
-5.49299084e-02 -2.56931514e-01 -7.47182965e-01 9.15859044e-02
8.17096412e-01 1.06554389e-01 2.55395293e-01 1.14077032e+00
-4.15407658e-01 1.52661547e-01 -1.40939736e+00 1.35891569e+00
3.70893747e-01 -1.61558688e+00 1.45326946e-02 -1.03895357e-02
1.00348508e+00 1.03869237e-01 1.36084199e-01 4.59221095e-01
8.02390128e-02 -1.02841330e+00 6.32236540e-01 6.77016020e-01
6.16514862e-01 -6.14674568e-01 6.54802442e-01 -4.58178632e-02
-1.34741032e+00 -7.01970980e-02 -7.12816238e-01 -7.26755708e-02
-1.45112932e-01 5.79867125e-01 -8.52677345e-01 4.02521282e-01
3.05968523e-01 1.22385299e+00 -7.87872195e-01 1.59488988e+00
-4.52523269e-02 8.04651439e-01 -2.51477808e-01 -2.72002280e-01
4.41410810e-01 1.72512725e-01 4.13326681e-01 1.53844070e+00
1.89326227e-01 -5.45996726e-01 4.74911541e-01 5.34330904e-01
-6.70011714e-02 1.37466729e-01 -7.69295573e-01 -4.56161201e-01
9.37144756e-02 1.18808687e+00 -9.88446772e-01 -3.20539415e-01
-6.13656580e-01 9.81047928e-01 3.63127172e-01 -3.85476388e-02
-3.09809804e-01 -5.33278584e-01 2.67202675e-01 2.65949517e-01
1.06798661e+00 -4.84088928e-01 -1.02499291e-01 -1.19456601e+00
-2.38118306e-01 -7.74357736e-01 2.34114558e-01 -3.77567917e-01
-1.83489132e+00 4.42687392e-01 -1.73888192e-01 -1.32752216e+00
-2.47144364e-02 -1.02025843e+00 -5.19550622e-01 2.06666574e-01
-1.37928391e+00 -1.04624927e+00 -1.44198388e-01 8.76946092e-01
7.48110533e-01 -7.44461715e-01 1.07298946e+00 3.27200741e-01
-1.05901025e-01 7.16710389e-01 6.38705194e-01 4.45377290e-01
6.83417678e-01 -1.25024462e+00 2.46868413e-02 1.56583995e-01
8.55371714e-01 4.06746566e-01 4.99605358e-01 -2.95061290e-01
-1.75270879e+00 -6.93658173e-01 5.08737564e-01 -1.44381896e-01
8.02948236e-01 -4.47283626e-01 -6.29999101e-01 3.56447458e-01
2.32148748e-02 4.33196694e-01 9.09023941e-01 3.88830721e-01
-6.60711527e-01 -1.17786966e-01 -1.20007741e+00 -1.61637858e-01
7.73922801e-01 -8.74931514e-01 -7.83707380e-01 5.09032607e-01
3.01905304e-01 -3.28287967e-02 -9.95086908e-01 1.17960155e-01
7.02097178e-01 -4.32678819e-01 1.06018543e+00 -6.14689767e-01
3.67615283e-01 1.45174667e-01 -5.70047617e-01 -1.20168710e+00
-6.12957656e-01 -2.99328625e-01 -5.01545787e-01 8.81058037e-01
1.41931728e-01 -1.49444744e-01 1.12600577e+00 6.64123893e-02
-2.48627439e-01 -7.41348207e-01 -9.63149905e-01 -1.14754415e+00
2.31658444e-01 -2.06468984e-01 2.61989627e-02 6.68435454e-01
-1.20226340e-03 2.37493694e-01 -4.63801652e-01 -5.81318140e-01
5.15091002e-01 4.99702424e-01 4.88278776e-01 -1.47567189e+00
-3.93968076e-01 -4.43639457e-01 -1.23024368e+00 -1.05656385e+00
8.60151276e-02 -1.19316268e+00 1.12301126e-01 -1.48295069e+00
3.88700575e-01 -3.86703253e-01 -7.31451154e-01 4.28028792e-01
3.89408797e-01 2.37207711e-01 2.53406972e-01 1.66078344e-01
-1.09224725e+00 5.35768867e-01 1.18848753e+00 -5.27096808e-01
1.12362333e-01 -1.57532215e-01 -4.94188607e-01 4.05988842e-01
7.10234284e-01 -5.83011866e-01 -5.87256532e-03 -2.40527257e-01
1.80801228e-01 -3.02386105e-01 2.21902847e-01 -1.51361620e+00
1.73436359e-01 3.51886600e-02 7.25965321e-01 -5.70110559e-01
3.19237918e-01 -8.76735985e-01 -3.04304838e-01 3.66880238e-01
-5.01861036e-01 -3.84272128e-01 8.71156007e-02 7.63879299e-01
-4.82700765e-01 -4.56810981e-01 3.98338199e-01 -1.65628031e-01
-5.98843455e-01 2.80606359e-01 -5.95813394e-01 -1.02077842e-01
7.65906453e-01 -2.51345634e-01 1.20556466e-01 -2.35718578e-01
-7.89098382e-01 -3.79798174e-01 2.57445514e-01 4.22176927e-01
9.33385193e-01 -1.67641068e+00 -6.59590900e-01 3.52968067e-01
2.21938297e-01 -1.81943595e-01 -4.21567149e-02 6.40452981e-01
-5.00533223e-01 9.50637400e-01 -5.77186644e-01 -6.88678265e-01
-1.03583455e+00 6.96705520e-01 -2.23871037e-01 -3.00883085e-01
-7.10388839e-01 1.09077406e+00 -1.31888062e-01 -1.23513520e-01
4.48219746e-01 -3.27398777e-01 1.36435051e-02 1.08212888e-01
4.71634179e-01 2.55547255e-01 2.89519459e-01 -5.85584581e-01
-3.71571481e-01 6.68422103e-01 -4.54702586e-01 3.79839152e-01
1.96543658e+00 3.34730029e-01 2.17991948e-01 6.10654235e-01
1.66687369e+00 -7.52694905e-02 -1.27416778e+00 -5.28230667e-01
2.97837585e-01 -5.02170563e-01 2.55221128e-01 -6.41509354e-01
-1.21183324e+00 7.45910764e-01 6.14803076e-01 2.32294723e-01
8.75885427e-01 1.00955002e-01 7.43404806e-01 9.03620541e-01
6.91924810e-01 -1.16989815e+00 5.90440392e-01 6.40725255e-01
9.99477088e-01 -1.38883173e+00 3.18311065e-01 -1.69562176e-01
-4.37408239e-01 1.49081802e+00 1.46195769e-01 -9.56034839e-01
9.97575760e-01 4.04888541e-01 -4.78412509e-01 -4.26810056e-01
-6.33795440e-01 -2.50403583e-01 8.20229530e-01 6.52647853e-01
5.57794452e-01 3.46471220e-02 -1.75361019e-02 4.30288404e-01
-1.40871853e-01 9.74025875e-02 6.89617544e-02 1.41269445e+00
-8.24548960e-01 -1.16286719e+00 -3.09888184e-01 9.28895712e-01
-2.99085826e-01 -1.40202343e-02 -7.40154147e-01 4.29808468e-01
-7.50577748e-02 6.05926931e-01 -6.84921890e-02 -4.67151612e-01
1.94300085e-01 1.26816198e-01 1.02014613e+00 -7.59041667e-01
-6.82316184e-01 4.48165275e-03 -9.90017354e-02 -5.23998499e-01
-5.76865852e-01 -6.97629571e-01 -7.88763583e-01 -5.95738664e-02
-7.73054600e-01 1.98574886e-01 6.54503167e-01 9.53392208e-01
1.75823212e-01 4.87120658e-01 7.17140257e-01 -1.18296754e+00
-5.48787594e-01 -9.52241540e-01 -8.15548241e-01 4.61513758e-01
2.22213596e-01 -1.01274180e+00 -4.14702058e-01 1.24698952e-01] | [9.4940185546875, 2.26234769821167] |
b96cf205-21da-4930-9ba4-e583d0e4e1d9 | adjointnet-constraining-machine-learning | 2109.03956 | null | https://arxiv.org/abs/2109.03956v1 | https://arxiv.org/pdf/2109.03956v1.pdf | AdjointNet: Constraining machine learning models with physics-based codes | Physics-informed Machine Learning has recently become attractive for learning physical parameters and features from simulation and observation data. However, most existing methods do not ensure that the physics, such as balance laws (e.g., mass, momentum, energy conservation), are constrained. Some recent works (e.g., physics-informed neural networks) softly enforce physics constraints by including partial differential equation (PDE)-based loss functions but need re-discretization of the PDEs using auto-differentiation. Training these neural nets on observational data showed that one could solve forward and inverse problems in one shot. They evaluate the state variables and the parameters in a PDE. This re-discretization of PDEs is not necessarily an attractive option for domain scientists that work with physics-based codes that have been developed for decades with sophisticated discretization techniques to solve complex process models and advanced equations of state. This paper proposes a physics constrained machine learning framework, AdjointNet, allowing domain scientists to embed their physics code in neural network training workflows. This embedding ensures that physics is constrained everywhere in the domain. Additionally, the mathematical properties such as consistency, stability, and convergence vital to the numerical solution of a PDE are still satisfied. We show that the proposed AdjointNet framework can be used for parameter estimation (and uncertainty quantification by extension) and experimental design using active learning. The applicability of our framework is demonstrated for four flow cases. Results show that AdjointNet-based inversion can estimate process model parameters with reasonable accuracy. These examples demonstrate the applicability of using existing software with no changes in source code to perform accurate and reliable inversion of model parameters. | ['Maruti K. Mudunuru', 'Bulbul Ahmmed', 'Satish Karra'] | 2021-09-08 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-3.33601683e-01 1.47907129e-02 -2.44112834e-02 -1.63317800e-01
-1.31158248e-01 -5.92461288e-01 4.66073990e-01 2.00880975e-01
-3.80023420e-01 9.79557633e-01 -2.90845186e-01 -5.08501470e-01
-5.87154448e-01 -9.37466383e-01 -8.11287582e-01 -8.06907892e-01
-4.11956310e-01 6.49796546e-01 -2.52615921e-02 -1.36607781e-01
1.87716261e-01 1.08981478e+00 -1.21890867e+00 -6.73942208e-01
1.16914499e+00 9.94258583e-01 -1.84117988e-01 6.84080601e-01
-2.71998197e-01 8.36423337e-01 -4.99774888e-02 3.14365089e-01
4.89619434e-01 -3.59060675e-01 -5.43707550e-01 -3.54803175e-01
2.40708053e-01 -4.57118571e-01 -3.96057397e-01 8.40527534e-01
2.69681036e-01 9.52680647e-01 7.68175244e-01 -1.35580230e+00
-4.55847502e-01 1.32655755e-01 -8.76172781e-02 -5.93934916e-02
-3.08326840e-01 5.80496728e-01 6.62227511e-01 -6.48439705e-01
3.36936146e-01 9.59708631e-01 1.15549409e+00 2.61520535e-01
-1.32365274e+00 -5.40632486e-01 -2.48922780e-03 -1.92502439e-02
-1.34412730e+00 -1.43032253e-01 7.95518339e-01 -7.90193915e-01
9.69686568e-01 2.61876225e-01 9.47349548e-01 7.35788345e-01
4.78570551e-01 -2.43367031e-02 9.08717990e-01 -1.87512308e-01
8.86636615e-01 2.57586151e-01 -9.01086163e-03 9.27210808e-01
3.43140781e-01 7.32725143e-01 -4.82644618e-01 -2.96639621e-01
1.08168483e+00 -2.05984026e-01 -2.02901348e-01 -5.90720892e-01
-9.32766497e-01 1.02187145e+00 5.60577691e-01 -2.25181326e-01
-4.70787644e-01 2.64866263e-01 3.49975318e-01 1.28313312e-02
5.15066266e-01 8.44146252e-01 -6.16704404e-01 -1.28242999e-01
-1.03388298e+00 3.39530677e-01 1.28185701e+00 6.48214698e-01
8.59796405e-01 5.50857723e-01 5.28657399e-02 4.66409445e-01
6.64083242e-01 7.53398836e-01 -9.58273187e-02 -1.53272879e+00
-2.37111017e-01 4.26912457e-01 3.35643500e-01 -8.66066873e-01
-1.93460569e-01 -4.46275085e-01 -1.05279684e+00 9.48602974e-01
3.08622718e-01 -4.73280877e-01 -8.87033343e-01 1.50913572e+00
5.35121441e-01 6.28307045e-01 2.48230565e-02 1.12841332e+00
6.00495040e-01 8.54046047e-01 1.87554508e-01 -9.42311883e-02
8.57804418e-01 -6.98242962e-01 -6.30722940e-01 -9.42317247e-02
3.56481373e-01 -3.67256343e-01 9.75976944e-01 8.25872347e-02
-1.11135793e+00 -6.05514348e-01 -1.04299271e+00 1.42642170e-01
-5.00613630e-01 -2.96554416e-01 8.63688886e-01 4.48179126e-01
-7.43262529e-01 1.38279629e+00 -1.28552139e+00 -1.36274397e-01
2.89499611e-01 2.77034879e-01 -3.74539532e-02 7.39373684e-01
-1.35454738e+00 1.01687741e+00 3.62430245e-01 2.99975187e-01
-1.17881536e+00 -1.55938923e+00 -8.84564698e-01 2.97302067e-01
6.07659929e-02 -8.94825816e-01 1.10063040e+00 -6.20529413e-01
-2.01518631e+00 3.28888834e-01 2.60930002e-01 -4.46114868e-01
7.51349449e-01 -2.99291611e-01 -1.79137602e-01 1.09794594e-01
-2.54143476e-01 4.16035950e-01 5.73070467e-01 -1.13777602e+00
-1.94797710e-01 3.73101056e-01 1.94698513e-01 7.43035674e-02
-3.08365732e-01 -3.59261930e-01 1.20428115e-01 -1.67616919e-01
1.64267197e-01 -9.91782427e-01 -3.98306459e-01 8.28306675e-01
-3.09421331e-01 2.25115925e-01 8.96052897e-01 -6.43855989e-01
4.39273775e-01 -1.80976152e+00 1.53597981e-01 2.06967771e-01
1.26434162e-01 2.78637916e-01 3.96214545e-01 3.70073766e-01
-1.03614606e-01 2.66701698e-01 -7.16019690e-01 -2.74789959e-01
2.24331215e-01 4.46314037e-01 -5.54666817e-01 7.40305245e-01
4.44586664e-01 7.13995278e-01 -8.17009211e-01 -5.14603436e-01
7.57581711e-01 7.66953290e-01 -5.13316572e-01 4.40768719e-01
-3.80290836e-01 1.07510579e+00 -3.07659477e-01 1.64589971e-01
8.92003179e-01 -2.56005019e-01 -1.91066742e-01 -1.92572355e-01
-5.55144310e-01 -1.23436721e-02 -1.42275798e+00 1.52304745e+00
-9.02197242e-01 5.42697072e-01 5.56716383e-01 -1.20352519e+00
7.83499479e-01 2.30298996e-01 7.42446542e-01 -3.95835340e-01
8.83934051e-02 1.02727711e-01 -1.07189175e-03 -6.04090035e-01
2.67028421e-01 -5.18667161e-01 2.89953321e-01 2.87889302e-01
5.93717135e-02 -7.70395994e-01 -1.08173089e-02 6.86764345e-02
7.52259433e-01 6.80853665e-01 7.23591596e-02 -6.10889256e-01
7.09679723e-01 5.10823905e-01 7.11597979e-01 6.94564104e-01
-1.21775500e-01 2.25513801e-01 2.12713763e-01 -2.90451944e-01
-1.27787519e+00 -1.20114744e+00 -3.20511997e-01 5.89890838e-01
6.57659769e-02 1.45020738e-01 -3.78857374e-01 -1.08592929e-02
3.90151024e-01 7.49148250e-01 -4.31675285e-01 -3.23878139e-01
-6.65482342e-01 -4.78683859e-01 4.37533796e-01 4.94415849e-01
7.61168778e-01 -7.89811432e-01 -6.35389090e-01 2.19966203e-01
3.80564749e-01 -9.34736252e-01 2.32788950e-01 3.07420850e-01
-1.09745514e+00 -8.83849144e-01 -3.65292937e-01 -2.84660071e-01
4.60888118e-01 -5.96474171e-01 9.97014701e-01 1.23945735e-01
-3.44522953e-01 4.22367126e-01 1.22434080e-01 -3.81305456e-01
-5.48758566e-01 -6.21398203e-02 1.57741040e-01 -3.68688464e-01
-4.62228239e-01 -1.02625489e+00 -5.20780683e-01 1.39927939e-02
-6.39319241e-01 -1.18054144e-01 3.46840203e-01 5.93846083e-01
5.13164341e-01 2.46342346e-01 7.42829964e-02 -5.20250142e-01
5.11879325e-01 -4.47965264e-01 -1.09962499e+00 -1.24559905e-02
-6.00567758e-01 1.13320336e-01 9.84386146e-01 -4.26884770e-01
-1.49062538e+00 1.26901463e-01 -6.74935877e-02 -7.98241258e-01
-3.44707400e-01 6.78131521e-01 1.42787471e-01 -3.70922983e-01
7.17511475e-01 -1.70226581e-02 2.08803222e-01 -3.81098330e-01
1.15487576e-01 7.75908530e-02 4.30373549e-01 -1.02687240e+00
1.09556091e+00 6.49672031e-01 7.26768315e-01 -8.78491938e-01
-7.16387570e-01 -2.66971797e-01 -6.35142982e-01 -2.90151805e-01
6.69846952e-01 -7.30999947e-01 -1.18424594e+00 4.08445954e-01
-8.91659617e-01 -6.47054732e-01 -6.65369809e-01 8.89957726e-01
-4.97531444e-01 3.65646809e-01 -6.22884214e-01 -1.23009312e+00
-3.39196622e-01 -8.92539263e-01 5.69202900e-01 3.77412826e-01
-2.86818892e-01 -1.69296992e+00 2.92510599e-01 -2.67815590e-01
7.56936967e-01 5.23673356e-01 7.59451866e-01 -1.68981105e-01
-6.43834054e-01 1.11679837e-01 -2.57042870e-02 3.47723931e-01
5.52495643e-02 5.04647553e-01 -1.15316570e+00 -7.54371732e-02
2.36553535e-01 -1.07260436e-01 7.61028647e-01 6.77145362e-01
1.01148283e+00 -1.70947433e-01 -2.41829127e-01 9.88342941e-01
1.61371636e+00 -9.46287587e-02 2.62184411e-01 3.92711125e-02
7.13893414e-01 5.31810105e-01 2.53093243e-01 4.55751956e-01
-2.35856012e-01 1.59203321e-01 5.74925601e-01 -2.92834133e-01
1.09926559e-01 5.52039295e-02 1.64375514e-01 7.19698966e-01
-1.12139203e-01 3.63307819e-02 -1.14131558e+00 3.20908159e-01
-1.69853783e+00 -8.56554031e-01 -4.77119416e-01 2.08790970e+00
8.47139597e-01 1.59209788e-01 -4.78338897e-01 -1.10241793e-01
4.93686110e-01 -1.73469290e-01 -9.62179244e-01 -6.17154062e-01
2.31504098e-01 3.89776707e-01 7.91602671e-01 8.79374743e-01
-1.19212043e+00 4.92277592e-01 6.06741381e+00 3.20622563e-01
-1.55916631e+00 1.87224507e-01 1.94475532e-01 5.20924851e-02
4.09044661e-02 4.52044070e-01 -5.39446473e-01 3.53368402e-01
1.05833447e+00 -1.69051453e-01 4.51624095e-01 9.60198104e-01
8.66259754e-01 -1.71960562e-01 -1.02675271e+00 4.74437028e-01
-6.72303855e-01 -1.64489317e+00 -3.59883577e-01 1.35497943e-01
8.05346131e-01 1.88070200e-02 -2.71326393e-01 4.56751555e-01
5.32788575e-01 -1.03552294e+00 7.71837950e-01 1.02145636e+00
4.61260617e-01 -4.73932177e-01 6.00580096e-01 4.42679167e-01
-1.00926650e+00 4.21050489e-02 -3.05314392e-01 -5.23672223e-01
6.07457161e-01 9.98217404e-01 -4.63460356e-01 5.82351744e-01
6.01821005e-01 7.15279639e-01 4.65064868e-02 1.14617407e+00
-3.41808908e-02 8.74075949e-01 -7.16444910e-01 1.58830330e-01
1.66570544e-01 -6.90321386e-01 7.52447724e-01 9.79462683e-01
2.76253551e-01 -1.05156131e-01 2.83604443e-01 1.71050072e+00
3.79856110e-01 -1.54496625e-01 -3.75164151e-01 -1.41955107e-01
2.79061943e-01 1.38366628e+00 -7.22394705e-01 -2.47196287e-01
-4.00419198e-02 3.99604112e-01 3.85904238e-02 5.69067001e-01
-1.23314226e+00 -1.87901154e-01 9.58938181e-01 -2.77889427e-02
9.90931913e-02 -5.80949128e-01 -4.46165830e-01 -8.01176012e-01
-1.37189195e-01 -2.46871531e-01 9.57750380e-02 -7.68217564e-01
-1.51250505e+00 6.24127453e-03 3.75123143e-01 -1.10136771e+00
-1.43923238e-01 -9.06860054e-01 -9.32113886e-01 1.23569548e+00
-1.63056922e+00 -9.12622988e-01 -5.48342049e-01 2.28527829e-01
-2.77183019e-03 1.68978591e-02 8.11990976e-01 1.18748792e-01
-5.89023829e-01 -2.43798733e-01 4.34149653e-01 1.67723835e-01
3.37405384e-01 -1.31039715e+00 7.96139911e-02 6.72690451e-01
-6.20262742e-01 9.27552760e-01 1.04504955e+00 -9.43152070e-01
-1.57056153e+00 -1.06342530e+00 2.32870817e-01 7.45228007e-02
9.29810107e-01 -1.78438291e-01 -1.15961778e+00 4.47574943e-01
7.29437172e-02 5.01914322e-01 2.84355313e-01 -6.84894845e-02
2.07330376e-01 -1.27008632e-01 -1.22858059e+00 1.83862299e-01
8.03420663e-01 -4.63122368e-01 -2.73380965e-01 4.72821891e-01
4.58854705e-01 -6.02824152e-01 -1.01151109e+00 4.93976027e-01
4.39075559e-01 -7.57541716e-01 1.10620499e+00 -8.38439286e-01
4.95799690e-01 -3.49965811e-01 1.61657527e-01 -1.33166111e+00
-1.49272338e-01 -4.85020250e-01 -4.57450897e-01 1.11782241e+00
1.16963357e-01 -9.46158469e-01 6.10747993e-01 9.94847953e-01
-4.17750150e-01 -5.13837397e-01 -8.38257194e-01 -1.07963562e+00
5.57811439e-01 -4.75278944e-01 3.92519832e-01 1.24129748e+00
-4.04031783e-01 -1.27433583e-01 8.01150128e-02 6.98144317e-01
7.85019040e-01 7.76594058e-02 4.63623613e-01 -1.59350097e+00
-3.21774721e-01 -3.79491776e-01 -1.02466688e-01 -4.40194935e-01
5.95914662e-01 -6.81071818e-01 9.25751328e-02 -1.72256982e+00
-3.50700825e-01 -6.69209898e-01 -1.17387004e-01 2.65321761e-01
2.08867356e-01 -1.89494729e-01 -1.23359554e-01 2.67552853e-01
2.00322881e-01 7.93228209e-01 1.25887322e+00 -6.57724887e-02
-2.66191393e-01 -2.72959173e-01 6.19113669e-02 7.71192729e-01
1.08239865e+00 -5.36950827e-01 -5.37550986e-01 -6.72392547e-02
2.71689415e-01 -1.11042209e-01 1.10573566e+00 -1.30491745e+00
4.95932996e-01 -5.02946138e-01 4.26175624e-01 -2.99319446e-01
5.44494152e-01 -1.02863264e+00 4.31894064e-01 6.49685800e-01
-2.37242237e-01 -3.74273300e-01 7.22403109e-01 3.78219008e-01
-3.18729021e-02 -3.13207030e-01 9.53928113e-01 -4.17259336e-01
-7.10958004e-01 2.55181521e-01 -4.92755026e-01 1.40432432e-01
8.80811870e-01 -1.72778234e-01 -3.63910664e-03 -2.91130483e-01
-7.66117990e-01 1.35313526e-01 3.61194402e-01 -2.03237489e-01
2.17072845e-01 -1.00345492e+00 -5.24228513e-01 8.19880217e-02
-4.71546769e-01 3.66979390e-01 4.04976159e-01 8.73841166e-01
-1.15194547e+00 1.70979515e-01 -3.08318704e-01 -7.59964347e-01
-5.30382156e-01 1.86224148e-01 1.17174125e+00 -9.07388702e-02
-7.15041220e-01 6.70892239e-01 -1.31286576e-01 -9.76070821e-01
-7.35465214e-02 -4.54558939e-01 3.04181457e-01 -2.08838642e-01
-7.53633007e-02 5.64880192e-01 1.05130807e-01 -2.44554624e-01
-3.76657397e-01 4.92322147e-01 6.94172204e-01 -2.28405297e-01
1.51526892e+00 1.45229205e-01 -2.37359896e-01 5.91853142e-01
9.53039944e-01 -1.57320723e-01 -1.72700179e+00 1.63591281e-01
-4.37669486e-01 -8.41570422e-02 7.60086775e-01 -7.90988743e-01
-1.23385417e+00 1.06958604e+00 6.30154729e-01 1.69194922e-01
5.51582396e-01 -6.39965355e-01 4.96505141e-01 5.95760942e-01
-1.31697312e-01 -1.25493562e+00 -3.08563739e-01 7.22079158e-01
8.15492868e-01 -1.04797900e+00 3.86520326e-01 -3.52700412e-01
-8.68163928e-02 1.17668831e+00 6.76206350e-01 -3.30076903e-01
1.16350448e+00 6.05479419e-01 2.77907383e-02 -1.82846323e-01
-2.82196432e-01 1.94108561e-01 9.28475931e-02 3.50653172e-01
3.46697748e-01 -2.47168690e-01 1.95302248e-01 -7.89589882e-02
-1.21883802e-01 1.14077359e-01 2.88466990e-01 1.03210390e+00
-2.72183925e-01 -6.39673471e-01 -5.79423487e-01 2.14919150e-01
1.85943693e-01 2.49100253e-02 -2.36544721e-02 1.00980878e+00
2.08109081e-01 6.33424282e-01 3.47080797e-01 3.13976586e-01
2.35805660e-01 2.48747677e-01 2.37046555e-01 -3.92546892e-01
-4.99776691e-01 -3.76745284e-01 -1.50207192e-01 -3.92771482e-01
-3.22993159e-01 -5.69397271e-01 -1.70967889e+00 -7.05775082e-01
-1.39592275e-01 2.78178513e-01 9.00502741e-01 9.10060287e-01
3.71524483e-01 8.64704430e-01 1.46931112e-01 -1.07624662e+00
-7.11256921e-01 -9.04765725e-01 -6.87437296e-01 1.51898280e-01
3.26947540e-01 -1.09034097e+00 -8.71388316e-01 1.33816853e-01] | [6.435507297515869, 3.441756248474121] |
4865cec8-f730-4c6b-9704-d32932c973c3 | graph-fuzzy-system-concepts-models-and | 2210.16730 | null | https://arxiv.org/abs/2210.16730v1 | https://arxiv.org/pdf/2210.16730v1.pdf | Graph Fuzzy System: Concepts, Models and Algorithms | Fuzzy systems (FSs) have enjoyed wide applications in various fields, including pattern recognition, intelligent control, data mining and bioinformatics, which is attributed to the strong interpretation and learning ability. In traditional application scenarios, FSs are mainly applied to model Euclidean space data and cannot be used to handle graph data of non-Euclidean structure in nature, such as social networks and traffic route maps. Therefore, development of FS modeling method that is suitable for graph data and can retain the advantages of traditional FSs is an important research. To meet this challenge, a new type of FS for graph data modeling called Graph Fuzzy System (GFS) is proposed in this paper, where the concepts, modeling framework and construction algorithms are systematically developed. First, GFS related concepts, including graph fuzzy rule base, graph fuzzy sets and graph consequent processing unit (GCPU), are defined. A GFS modeling framework is then constructed and the antecedents and consequents of the GFS are presented and analyzed. Finally, a learning framework of GFS is proposed, in which a kernel K-prototype graph clustering (K2PGC) is proposed to develop the construction algorithm for the GFS antecedent generation, and then based on graph neural network (GNNs), consequent parameters learning algorithm is proposed for GFS. Specifically, three different versions of the GFS implementation algorithm are developed for comprehensive evaluations with experiments on various benchmark graph classification datasets. The results demonstrate that the proposed GFS inherits the advantages of both existing mainstream GNNs methods and conventional FSs methods while achieving better performance than the counterparts. | ['Shitong Wang', 'Kup-Sze Choi', 'Zhenping Xie', 'Zhaohong Deng', 'Fuping Hu'] | 2022-10-30 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 1.55801937e-01 -1.48688152e-01 -3.78507487e-02 -2.81736225e-01
4.79322106e-01 -1.43490642e-01 1.98161393e-01 3.81963313e-01
-7.18550682e-02 5.01412451e-01 -4.85048354e-01 -5.14313161e-01
-7.57681847e-01 -1.29710162e+00 -3.88326496e-01 -5.45500636e-01
-1.50466532e-01 2.53091156e-01 6.21059299e-01 -6.28839791e-01
3.48072141e-01 4.75220382e-01 -1.62226033e+00 -1.65757045e-01
1.43284500e+00 1.06321776e+00 2.12067366e-01 1.23630501e-01
-3.53486389e-01 4.28799391e-01 -4.33695138e-01 -1.64144099e-01
1.89824238e-01 -3.01034510e-01 -4.61499602e-01 1.81534942e-02
-3.44449520e-01 2.09696859e-01 -2.58918583e-01 1.28878748e+00
7.99900144e-02 6.95876718e-01 8.57306004e-01 -1.53869843e+00
-8.64534259e-01 2.25435928e-01 -3.81255448e-01 -4.72904555e-02
1.41995743e-01 -2.25716159e-01 4.17705953e-01 -4.13727999e-01
3.65807265e-01 1.35467565e+00 5.56442380e-01 1.75289631e-01
-7.50287294e-01 -6.89163864e-01 2.50660181e-01 5.80485702e-01
-1.76466584e+00 2.61929929e-01 7.91141748e-01 -3.02369744e-01
7.74446368e-01 2.85152346e-01 8.02660942e-01 2.54619241e-01
2.83489767e-02 5.36569774e-01 1.07749486e+00 -5.89106381e-01
2.81735003e-01 1.80339038e-01 5.30064940e-01 8.62558544e-01
5.18357575e-01 -7.66246244e-02 1.93740442e-01 1.39131516e-01
7.42349327e-01 2.72969361e-02 -2.55132675e-01 -1.56638920e-01
-6.39862716e-01 7.38306999e-01 7.20252156e-01 5.81769347e-01
-2.34112427e-01 -4.56321955e-01 4.14794922e-01 1.19310379e-01
-8.49923305e-03 -6.79532662e-02 -9.72749852e-03 4.06297207e-01
-5.96258581e-01 1.82599783e-01 6.59105420e-01 1.10302889e+00
8.93882275e-01 1.98253170e-01 2.76049925e-03 6.86576366e-01
4.50844854e-01 5.44974506e-01 5.80723464e-01 -3.58378887e-01
2.01275989e-01 1.38312173e+00 -2.85980374e-01 -1.80102420e+00
-6.14881635e-01 -1.43630520e-01 -1.27868271e+00 -4.79679704e-02
-1.71350986e-01 1.25793785e-01 -8.33147228e-01 1.35282552e+00
4.89151388e-01 6.12395644e-01 2.20522925e-01 6.34274125e-01
9.52983260e-01 8.07286441e-01 6.00952581e-02 -4.43108648e-01
1.14531851e+00 -4.90768611e-01 -7.30658412e-01 3.01993668e-01
3.03895086e-01 -3.01915824e-01 9.82879758e-01 3.94641370e-01
-5.82834601e-01 -8.81637394e-01 -1.23057652e+00 3.45486611e-01
-9.36648011e-01 5.82406810e-03 6.46872520e-01 7.04877496e-01
-1.01008916e+00 4.75599289e-01 -6.57384276e-01 -5.91462493e-01
2.20692396e-01 5.16039073e-01 -1.85016006e-01 -8.65083784e-02
-1.67889178e+00 8.49463105e-01 1.19041252e+00 3.55047405e-01
-1.76784657e-02 -2.15493515e-01 -9.47324753e-01 1.48734838e-01
4.55353796e-01 -5.31045556e-01 5.45394242e-01 -8.92092884e-01
-1.35014844e+00 1.71180606e-01 1.76545739e-01 -3.76238823e-01
6.11686222e-02 6.17024541e-01 -1.23084927e+00 2.38005683e-01
-1.66611657e-01 7.25920945e-02 7.91802764e-01 -1.00853515e+00
-6.37483001e-01 -3.70398343e-01 8.97300094e-02 1.84078932e-01
-5.99227548e-01 -3.14916044e-01 -4.66096878e-01 -4.55038071e-01
8.99023861e-02 -6.85639203e-01 -1.41212031e-01 -5.62722027e-01
-3.73748243e-01 -4.67015982e-01 9.30671871e-01 -3.65588009e-01
1.86200213e+00 -1.91960144e+00 3.40425149e-02 1.17411494e+00
4.07074168e-02 9.44576263e-01 1.85020298e-01 4.27460849e-01
1.26922559e-02 5.66701479e-02 -4.69485462e-01 6.24729514e-01
-5.07098064e-02 3.35279137e-01 2.15186313e-01 9.62149650e-02
1.63166955e-01 6.38859749e-01 -7.98935056e-01 -5.58483064e-01
7.22389579e-01 3.13275039e-01 -2.51460403e-01 -8.33394565e-03
-1.44818515e-01 5.59901893e-02 -7.25534916e-01 4.87137794e-01
9.08431053e-01 -3.18014234e-01 3.75960052e-01 -6.00254178e-01
-1.46730751e-01 -6.47608638e-01 -1.84671402e+00 1.04009092e+00
-1.34843156e-01 -2.32066348e-01 -2.56240629e-02 -1.42169261e+00
1.41681588e+00 -7.14952499e-03 3.20123017e-01 -7.27454841e-01
4.03937161e-01 2.74485260e-01 1.78235531e-01 -6.86214805e-01
1.85898736e-01 -1.22637391e-01 1.93582222e-01 3.23720500e-02
-8.02861676e-02 1.40066028e-01 8.22795093e-01 2.73749530e-01
6.36456192e-01 -5.71207181e-02 5.03907979e-01 -5.20690739e-01
1.45857501e+00 1.68183044e-01 5.59859097e-01 2.81110227e-01
8.79591480e-02 -1.15700766e-01 2.23805398e-01 -3.85938406e-01
-3.01973522e-01 -9.77824986e-01 -4.47969027e-02 3.21087122e-01
7.21166015e-01 -5.05440354e-01 -8.87370884e-01 -5.84998310e-01
1.32989092e-02 6.60554051e-01 -3.00298452e-01 -5.28290570e-01
-3.75982106e-01 -7.09910691e-01 5.35566032e-01 4.04980421e-01
9.12447751e-01 -1.09739780e+00 -1.97659463e-01 1.32802814e-01
9.85327512e-02 -8.26405346e-01 -1.04221188e-01 -4.28653717e-01
-6.64863586e-01 -1.58845675e+00 -8.85533243e-02 -1.11040485e+00
8.52774024e-01 4.89119351e-01 5.75664222e-01 5.95538259e-01
-8.48335624e-02 2.42371604e-01 -5.27635694e-01 -4.32450950e-01
-4.96864647e-01 -1.81939453e-04 3.02749634e-01 3.73674691e-01
4.83532131e-01 -5.62158883e-01 -3.07786793e-01 2.94183314e-01
-1.25806594e+00 1.30057484e-01 6.68793201e-01 6.63676262e-01
6.54003739e-01 9.79934394e-01 1.06046426e+00 -1.02245164e+00
8.22163224e-01 -5.61344087e-01 -1.01828885e+00 6.55657649e-01
-1.08432233e+00 -1.19998418e-01 1.26299036e+00 -2.31556837e-02
-1.14826512e+00 -3.52949500e-01 -3.69575433e-02 -4.19374228e-01
-1.76192373e-01 8.56537461e-01 -4.97812688e-01 -2.91197389e-01
5.25457501e-01 4.24491942e-01 4.06501472e-01 -1.98303834e-01
1.68999642e-01 8.15444946e-01 4.62485880e-01 -5.99655986e-01
8.53499234e-01 1.13194026e-01 4.25106913e-01 -9.37628150e-01
-4.00805138e-02 -3.49670887e-01 -5.42589486e-01 -4.26034987e-01
7.15513527e-01 -3.76630068e-01 -1.11316454e+00 4.69533741e-01
-8.24704945e-01 2.18520075e-01 1.61537677e-01 5.90478897e-01
-2.42853686e-01 4.61891651e-01 -3.91383708e-01 -8.67827296e-01
-2.39757627e-01 -1.09755957e+00 4.50667471e-01 7.11692333e-01
3.05724710e-01 -1.17697942e+00 -2.96356320e-01 1.04357362e-01
3.32486480e-01 5.86984634e-01 1.29949784e+00 -7.58843601e-01
-3.43078762e-01 -8.16114992e-02 -2.85985649e-01 6.33271754e-01
3.24391305e-01 1.82838291e-01 -2.00965211e-01 -1.50674865e-01
-1.57067463e-01 9.64683741e-02 2.97409952e-01 2.42968470e-01
1.08705270e+00 -3.28893185e-01 -3.90829891e-01 4.58369255e-01
1.85902691e+00 8.51046205e-01 6.66368365e-01 3.36207807e-01
7.54587352e-01 6.16422474e-01 7.70172000e-01 1.64185598e-01
5.07369637e-01 3.16590220e-01 2.92987764e-01 1.68545451e-02
2.64754593e-01 -1.11204803e-01 1.34056166e-01 1.16319311e+00
-3.31424862e-01 -2.25044772e-01 -7.11651027e-01 3.75126675e-02
-1.95631099e+00 -8.89132500e-01 -4.46529567e-01 1.97956240e+00
3.25056612e-01 1.71721920e-01 1.85618639e-01 5.91879725e-01
1.28722751e+00 -2.19536871e-01 -5.33222795e-01 -3.92703831e-01
-5.49877062e-02 3.32447201e-01 2.31247634e-01 3.39248061e-01
-9.55026269e-01 8.15400720e-01 4.58703947e+00 1.19105744e+00
-1.07254148e+00 -3.96505803e-01 1.90325230e-01 5.86627483e-01
-1.83754221e-01 1.03823543e-02 -5.27249694e-01 6.20077908e-01
7.35652030e-01 -3.86197001e-01 8.20716858e-01 6.38270259e-01
3.20054233e-01 1.66460052e-01 -3.42373669e-01 1.10027981e+00
-4.18539718e-02 -1.21964478e+00 4.04959023e-01 -3.31840098e-01
7.13263929e-01 -6.35019720e-01 -2.61323214e-01 3.23529780e-01
-2.21334537e-03 -9.61886227e-01 3.67846787e-01 6.04344666e-01
7.43350387e-01 -9.72515941e-01 7.51344025e-01 3.26614767e-01
-1.63833964e+00 -1.39236778e-01 -4.57268804e-01 -3.61401252e-02
6.56538317e-03 4.42554772e-01 -7.13926494e-01 1.69289625e+00
3.93837482e-01 8.82615864e-01 -7.57303476e-01 8.72283816e-01
3.64194997e-02 4.68413204e-01 -4.36804771e-01 -2.65521705e-01
2.95050859e-01 -8.93345654e-01 2.39760011e-01 9.19659257e-01
4.46793646e-01 4.20662701e-01 2.89250314e-01 8.55796039e-01
3.11848581e-01 6.91028714e-01 -5.22587299e-01 -8.16049725e-02
7.84798980e-01 1.39679515e+00 -1.07179964e+00 -4.94779050e-01
-4.68452513e-01 2.68590778e-01 1.79971009e-01 3.58920991e-01
-9.26879108e-01 -8.98622155e-01 -1.35251805e-02 2.09497496e-01
-2.87591368e-02 8.09132531e-02 2.84253750e-02 -9.71909344e-01
2.91112121e-02 -6.41084850e-01 7.77421117e-01 -4.41309392e-01
-1.28836942e+00 5.84346712e-01 3.44033301e-01 -1.33664072e+00
4.19909731e-02 -8.19083989e-01 -8.99458408e-01 7.48233795e-01
-1.45434952e+00 -1.26346219e+00 -6.85319960e-01 1.11951542e+00
-8.27587396e-02 -3.06484163e-01 4.62149709e-01 5.48364282e-01
-8.75211120e-01 4.52128023e-01 1.59778580e-01 -9.72879678e-03
1.30738199e-01 -9.86261249e-01 -1.55310988e-01 9.75216508e-01
-1.17453955e-01 9.18730915e-01 2.43695095e-01 -8.22290123e-01
-1.51236951e+00 -1.38620722e+00 3.78897905e-01 5.06419897e-01
5.20473301e-01 1.38995692e-01 -1.07219636e+00 4.62400109e-01
-3.26117367e-01 9.43662375e-02 3.17223191e-01 -2.92378366e-01
1.03351250e-01 -4.61873382e-01 -1.38041914e+00 5.91472030e-01
9.13829625e-01 -2.52216551e-02 -4.07410800e-01 2.01056808e-01
4.12252098e-01 -2.41680652e-01 -1.07989740e+00 6.59569263e-01
4.30711120e-01 -1.01639354e+00 9.36271071e-01 -4.22241628e-01
-1.83508992e-01 -1.08371472e+00 -5.53639838e-03 -1.06284833e+00
-5.69810092e-01 -2.00228363e-01 -8.21766406e-02 1.31280959e+00
-8.83951485e-02 -9.00634766e-01 5.43556035e-01 2.97984183e-01
-4.60947096e-01 -7.81980157e-01 -6.33619249e-01 -1.04895186e+00
-4.08676237e-01 -1.68200031e-01 9.91705716e-01 9.65047240e-01
9.04121473e-02 4.05606568e-01 1.46970406e-01 2.13754803e-01
4.35275495e-01 6.67185411e-02 4.50156868e-01 -1.48156023e+00
-1.19792148e-01 -5.88132679e-01 -8.15840423e-01 -3.42878044e-01
2.14268491e-01 -1.33951116e+00 -2.94235021e-01 -1.75049305e+00
-5.24338543e-01 -6.80721700e-01 -5.30743420e-01 2.13312432e-01
-2.16208845e-01 -1.59783866e-02 1.32858306e-01 -2.64218990e-02
-5.15936494e-01 5.31392217e-01 1.21999550e+00 -1.86444089e-01
-4.00184959e-01 -4.23198380e-02 -6.73678160e-01 6.87053859e-01
7.63120472e-01 1.78225234e-01 -1.11943698e+00 1.17432974e-01
-6.57296926e-02 3.05655114e-02 1.46444395e-01 -1.22091305e+00
4.02341753e-01 -3.89073133e-01 1.92056656e-01 -4.96685594e-01
-2.18429714e-01 -1.05696440e+00 5.86297214e-01 5.90388954e-01
2.04891518e-01 -3.29947770e-02 3.86013910e-02 6.93960309e-01
-3.29490364e-01 -2.15870544e-01 7.69666016e-01 2.80290507e-02
-1.12985432e+00 4.68957663e-01 -1.95394158e-01 -4.38311100e-02
1.44983864e+00 -5.03009975e-01 -1.98787943e-01 1.84391215e-01
-6.30092382e-01 3.91009450e-01 2.02706292e-01 3.61984015e-01
7.23735213e-01 -1.63348639e+00 -3.31252009e-01 4.71839368e-01
1.86802015e-01 3.34845692e-01 6.18550777e-01 7.45560169e-01
-7.56776631e-01 3.43224466e-01 -3.26327085e-01 -3.20719928e-01
-9.32140648e-01 9.22261059e-01 -1.03742359e-02 -6.77379370e-02
-4.08486933e-01 1.97848111e-01 -1.52643267e-02 -4.51292813e-01
-1.22030355e-01 -4.31436360e-01 -8.07596564e-01 -1.37535661e-01
4.05715764e-01 7.38429427e-01 3.36235344e-01 -9.25524354e-01
-4.68199104e-01 7.12581098e-01 2.51711160e-01 4.68716770e-01
9.41572547e-01 -7.70255476e-02 -3.61486644e-01 1.33389398e-01
9.99955475e-01 -3.81014496e-01 -5.57162583e-01 -5.53820357e-02
6.31664395e-02 -1.49499103e-01 -3.04232657e-01 -6.41678452e-01
-1.20200300e+00 5.53822041e-01 4.70191181e-01 4.89265859e-01
1.44866967e+00 -5.22324204e-01 9.59705472e-01 3.90324175e-01
4.96011645e-01 -1.29787409e+00 -3.87338370e-01 3.95327985e-01
4.51492161e-01 -7.94453323e-01 -1.35807961e-01 -6.36377811e-01
-5.26596427e-01 1.46850359e+00 8.78829360e-01 -3.88500392e-01
9.12575245e-01 -1.91284722e-04 -4.81322110e-01 -2.62909353e-01
-1.42534345e-01 -2.45270133e-01 4.51967418e-01 8.43434811e-01
-4.24309596e-02 8.24652687e-02 -6.90616548e-01 1.19859695e+00
-1.97066039e-01 3.47730696e-01 3.27816665e-01 7.59848237e-01
-5.72707474e-01 -9.62497532e-01 -3.82354707e-01 8.08365464e-01
1.77907154e-01 1.66519180e-01 -2.43782857e-03 9.51987088e-01
6.55886292e-01 1.07412398e+00 -6.39941245e-02 -8.95505726e-01
5.23102641e-01 -1.57219067e-01 3.96355063e-01 -3.52673382e-01
-2.77305931e-01 -4.33137119e-01 -2.31553271e-01 -2.85082042e-01
-6.69651628e-01 -2.56418049e-01 -1.91346741e+00 -4.09124076e-01
-6.21887147e-01 6.82527661e-01 2.79522270e-01 9.88505721e-01
2.54541695e-01 6.22124493e-01 6.20090067e-01 -2.93297052e-01
-5.24734035e-02 -6.40341759e-01 -9.22904372e-01 4.66135412e-01
-4.67163742e-01 -1.12981546e+00 -9.21230391e-02 -1.30636200e-01] | [7.561838150024414, 4.690603256225586] |
83e70d42-d326-4efc-ab8a-f8dff149ad1f | developing-personalized-models-of-blood | 2007.12802 | null | https://arxiv.org/abs/2007.12802v1 | https://arxiv.org/pdf/2007.12802v1.pdf | Developing Personalized Models of Blood Pressure Estimation from Wearable Sensors Data Using Minimally-trained Domain Adversarial Neural Networks | Blood pressure monitoring is an essential component of hypertension management and in the prediction of associated comorbidities. Blood pressure is a dynamic vital sign with frequent changes throughout a given day. Capturing blood pressure remotely and frequently (also known as ambulatory blood pressure monitoring) has traditionally been achieved by measuring blood pressure at discrete intervals using an inflatable cuff. However, there is growing interest in developing a cuffless ambulatory blood pressure monitoring system to measure blood pressure continuously. One such approach is by utilizing bioimpedance sensors to build regression models. A practical problem with this approach is that the amount of data required to confidently train such a regression model can be prohibitive. In this paper, we propose the application of the domain-adversarial training neural network (DANN) method on our multitask learning (MTL) blood pressure estimation model, allowing for knowledge transfer between subjects. Our proposed model obtains average root mean square error (RMSE) of $4.80 \pm 0.74$ mmHg for diastolic blood pressure and $7.34 \pm 1.88$ mmHg for systolic blood pressure when using three minutes of training data, $4.64 \pm 0.60$ mmHg and $7.10 \pm 1.79$ respectively when using four minutes of training data, and $4.48 \pm 0.57$ mmHg and $6.79 \pm 1.70$ respectively when using five minutes of training data. DANN improves training with minimal data in comparison to both directly training and to training with a pretrained model from another subject, decreasing RMSE by $0.19$ to $0.26$ mmHg (diastolic) and by $0.46$ to $0.67$ mmHg (systolic) in comparison to the best baseline models. We observe that four minutes of training data is the minimum requirement for our framework to exceed ISO standards within this cohort of patients. | ['Harlan M. Krumholz', 'Bassem Ibrahim', 'Nathan C. Hurley', 'Bobak J. Mortazavi', 'Roozbeh Jafari', 'Lida Zhang', 'Erica Spatz'] | 2020-07-24 | null | null | null | null | ['blood-pressure-estimation'] | ['medical'] | [ 1.94301739e-01 2.22829223e-01 3.91802862e-02 -4.58135098e-01
-7.17419267e-01 -2.81171978e-01 -1.41904771e-01 3.80765855e-01
-5.93767643e-01 1.31061804e+00 -2.99689174e-01 -5.08801460e-01
-1.95073903e-01 -1.03346050e+00 -8.17808390e-01 -4.82176453e-01
-5.73605895e-01 3.09026450e-01 -2.95709997e-01 1.43090054e-01
-1.50841057e-01 2.39801958e-01 -9.11157668e-01 -1.94855198e-01
1.32909405e+00 1.24988663e+00 -4.86056417e-01 7.52641439e-01
1.70393571e-01 4.06562775e-01 -8.13975990e-01 -9.89405885e-02
5.66611230e-01 -4.20553207e-01 -1.51454592e-02 -1.34816334e-01
8.09771061e-01 -4.23002869e-01 2.65557528e-01 8.93730104e-01
7.31950819e-01 9.23767015e-02 2.60843724e-01 -7.01318443e-01
-2.98682332e-01 4.60552961e-01 -4.29233670e-01 1.57508045e-01
-5.22043779e-02 1.52581796e-01 3.94019753e-01 -2.41168305e-01
-1.90782055e-01 6.50176585e-01 8.01328182e-01 5.61921000e-01
-1.73389566e+00 -6.01090372e-01 -2.06473470e-01 -4.42872941e-01
-1.43270910e+00 -5.37517309e-01 4.69594359e-01 -6.65209830e-01
4.40444678e-01 6.40873075e-01 7.42697477e-01 6.08820081e-01
5.39743364e-01 -7.13720322e-02 1.50549233e+00 -3.21364135e-01
1.77018240e-01 6.74332917e-01 1.61413044e-01 4.51811939e-01
4.34732109e-01 1.26096994e-01 2.08866745e-01 -2.01786518e-01
1.01718771e+00 2.68931538e-01 -1.87048450e-01 5.84077798e-02
-5.86187899e-01 4.84911263e-01 4.35407192e-01 8.66220221e-02
-5.23984313e-01 -1.22463696e-01 2.38752052e-01 4.50331241e-01
4.21800315e-01 5.18419027e-01 -4.77527022e-01 -1.15419000e-01
-1.06846797e+00 4.15906578e-01 9.32062566e-01 6.10931695e-01
5.20737052e-01 1.96768656e-01 -2.52142519e-01 7.29897022e-01
1.91601411e-01 8.82539451e-01 2.29538605e-01 -9.71417427e-01
3.72575641e-01 4.52133417e-01 5.71469188e-01 -1.06221688e+00
-5.70866585e-01 -7.28180230e-01 -1.27694368e+00 1.69883147e-01
7.59481251e-01 -6.01578534e-01 -7.50719070e-01 1.67060459e+00
1.09846912e-01 1.26975760e-01 -1.39694676e-01 7.91469634e-01
4.36495960e-01 2.17328042e-01 3.99705261e-01 -5.29463291e-01
1.32137704e+00 -3.66689295e-01 -4.46037084e-01 -1.98582247e-01
2.63026386e-01 -3.90590161e-01 1.00277221e+00 3.03891271e-01
-1.10185850e+00 -8.59609544e-01 -8.96475732e-01 3.02272975e-01
-2.26520404e-01 -8.83912146e-02 2.46875510e-01 1.16805410e+00
-6.73616529e-01 8.58697832e-01 -7.77453244e-01 -1.27587095e-02
3.31862748e-01 3.60560298e-01 1.72064453e-01 1.37570396e-01
-1.38240492e+00 9.11002278e-01 -1.83282644e-01 2.43345171e-01
-3.07114154e-01 -1.53648889e+00 -6.04688168e-01 4.04009270e-03
-8.18130076e-02 -7.77299702e-01 7.35546350e-01 -2.36556381e-01
-1.55207455e+00 6.40570462e-01 -1.48022413e-01 -8.56078684e-01
8.73832464e-01 -6.00235641e-01 -6.36523545e-01 1.15479566e-01
-8.71427543e-03 2.07280934e-01 4.92172360e-01 -7.20460892e-01
-2.61616230e-01 -6.93659246e-01 -3.59764665e-01 -2.65488714e-01
-1.09717064e-02 -1.82052732e-01 2.92539835e-01 -2.58305997e-01
-6.88620955e-02 -9.72224772e-01 -5.88176072e-01 8.07369873e-02
-2.62623489e-01 3.40730995e-01 2.74114460e-01 -1.12672126e+00
1.07841980e+00 -1.57216454e+00 -4.56350505e-01 5.38685799e-01
4.50376183e-01 5.06036162e-01 5.81672490e-01 -9.01172236e-02
2.29967777e-02 2.63040572e-01 -2.91475117e-01 -8.27971548e-02
-1.71879411e-01 -7.45260268e-02 -2.16726623e-02 3.90797287e-01
-7.56050274e-02 8.63468051e-01 -6.04984939e-01 -1.78225383e-01
5.53304613e-01 5.77199757e-01 -4.50798869e-01 2.12216780e-01
7.37136602e-02 9.41283047e-01 -2.37418622e-01 4.96516019e-01
7.64376044e-01 -1.33143142e-01 1.51103377e-01 -1.54691011e-01
-3.19515705e-01 -1.82403009e-02 -1.11521852e+00 1.41833758e+00
-4.33873534e-01 1.97226360e-01 -8.48764107e-02 -1.26028776e+00
1.57788563e+00 6.17998958e-01 8.56185019e-01 -8.91442001e-01
-4.79181446e-02 2.02152759e-01 3.92344326e-01 -5.22181153e-01
-3.22041273e-01 -6.54880285e-01 4.45105098e-02 2.09004313e-01
-5.76676309e-01 1.83681101e-01 -1.78092401e-02 -4.14837301e-01
9.78644371e-01 -1.77192628e-01 2.77719140e-01 -5.56216002e-01
6.82634652e-01 -2.55256504e-01 7.61238158e-01 8.89413357e-01
-6.50479853e-01 1.27566397e-01 1.13091700e-01 -7.46168375e-01
-9.39671934e-01 -1.17498255e+00 -8.78188670e-01 2.73982346e-01
-3.70813161e-01 2.62778141e-02 -3.82728636e-01 -3.41766775e-02
5.45301676e-01 4.28562671e-01 -3.85150731e-01 -2.76267789e-02
-7.11620808e-01 -9.13722396e-01 4.66926068e-01 5.05653381e-01
7.82147586e-01 -5.24446070e-01 -9.01599467e-01 5.71232021e-01
5.67026250e-02 -7.61057615e-01 1.36496155e-02 2.92721763e-02
-1.14908659e+00 -8.46627653e-01 -8.73064160e-01 -8.13862681e-02
5.09486973e-01 -5.05900919e-01 1.15646815e+00 -2.62267649e-01
-4.95212972e-01 1.17799632e-01 1.95979178e-01 -6.91156268e-01
-1.27622753e-01 -1.34610891e-01 3.20558608e-01 5.74502014e-02
7.86940098e-01 -1.00468814e+00 -1.20447683e+00 2.17258096e-01
-1.78704724e-01 -2.60399848e-01 3.88202250e-01 4.52903360e-01
7.32822180e-01 -3.65112633e-01 9.21882510e-01 -1.09930193e+00
4.68124419e-01 -3.60530913e-01 -7.83803046e-01 8.94761905e-02
-1.10151923e+00 -3.33995461e-01 9.03693020e-01 -2.76724339e-01
-5.00259340e-01 -1.47209987e-01 1.64923847e-01 -3.47360075e-01
-2.91483998e-01 4.21660334e-01 2.28163153e-01 -2.97190640e-02
9.62401688e-01 1.95744678e-01 3.10574383e-01 -5.66459060e-01
-6.53903782e-02 7.16378629e-01 5.13371408e-01 -7.31522024e-01
6.34611905e-01 1.26605898e-01 4.42519695e-01 -8.74834359e-01
-7.46194363e-01 -4.75919619e-02 -6.04706705e-01 -1.71000078e-01
6.47757292e-01 -9.12149131e-01 -1.17306817e+00 4.87968214e-02
-2.47212291e-01 -2.46985555e-01 -2.37540364e-01 6.06104076e-01
-4.03164476e-01 2.05328092e-01 -4.40324306e-01 -1.00233114e+00
-9.15870667e-01 -7.43731320e-01 3.64164501e-01 2.20457181e-01
-4.92857486e-01 -1.10560191e+00 -1.16449259e-01 4.17292714e-01
1.04390085e+00 1.06887710e+00 7.40196109e-01 -3.17950428e-01
-2.90189773e-01 -7.13140905e-01 -1.36623904e-01 6.17278159e-01
6.78261995e-01 -4.29419339e-01 -6.68375432e-01 -3.80308867e-01
3.31636071e-01 5.52245378e-02 3.46785903e-01 8.63491893e-01
1.12428463e+00 -3.86785388e-01 -2.32422352e-01 4.08387184e-01
1.55629063e+00 5.23419082e-01 7.01304674e-01 1.31962106e-01
3.50545794e-01 8.94425139e-02 1.92128435e-01 6.87793791e-01
1.60587564e-01 4.30706441e-01 -1.77174643e-01 -2.95651972e-01
2.61372149e-01 1.28484787e-02 -1.04257472e-01 1.79312438e-01
-3.44571561e-01 6.16107881e-01 -9.43517983e-01 1.62445933e-01
-1.42322099e+00 -8.12475801e-01 -1.41727388e-01 2.66407442e+00
1.22069418e+00 1.93343714e-01 3.79581630e-01 1.17247753e-01
6.17050767e-01 -2.68205971e-01 -7.62030482e-01 -4.75759864e-01
4.30845380e-01 7.70419061e-01 5.70673108e-01 5.84943295e-01
-1.08913445e+00 1.81220949e-01 5.32752466e+00 -4.58367407e-01
-1.32956612e+00 -2.85472900e-01 9.37487185e-01 -2.60775417e-01
4.73907024e-01 -1.72327116e-01 -7.03713536e-01 9.45230484e-01
1.53200960e+00 -6.01649463e-01 2.66218781e-01 8.74280930e-01
7.19639719e-01 -2.78019011e-01 -1.22514808e+00 9.77236629e-01
-2.74847120e-01 -1.04834545e+00 -5.26693761e-01 7.13891834e-02
4.22345340e-01 -2.84483016e-01 -7.10909516e-02 3.60710889e-01
-1.79904804e-01 -1.16317034e+00 -1.54725969e-01 9.15791512e-01
9.79746878e-01 -4.45313454e-01 5.49913049e-01 3.29599619e-01
-9.07793164e-01 1.01965524e-01 -4.03984010e-01 -4.80149001e-01
2.17892125e-01 1.29890764e+00 -7.09367871e-01 2.94684589e-01
6.57095492e-01 4.29539770e-01 -2.01780751e-01 1.17065716e+00
1.45673171e-01 6.00626767e-01 -4.65085298e-01 1.27065420e-01
-3.06853980e-01 -4.90938872e-01 3.85267675e-01 8.57264578e-01
2.91553617e-01 1.12296090e-01 4.87428874e-01 1.14932024e+00
1.47760510e-01 1.14596151e-01 -3.59728307e-01 5.26274204e-01
3.39231461e-01 7.28424489e-01 1.44417092e-01 -5.97915888e-01
-3.36421281e-01 5.36886394e-01 -1.15906581e-01 3.05477858e-01
-9.60306346e-01 -6.14234090e-01 6.16387725e-01 5.96577227e-01
-2.50007331e-01 -7.10875774e-03 -7.35203862e-01 -9.52765822e-01
4.53266442e-01 -4.37467694e-01 3.47746730e-01 1.57988146e-02
-1.15415728e+00 2.08728805e-01 -1.89493060e-01 -1.24573803e+00
-1.47650436e-01 -2.84024298e-01 -4.98151898e-01 1.74718726e+00
-1.57721031e+00 -4.18568760e-01 -5.59588432e-01 4.15495753e-01
-9.05906931e-02 1.04432784e-01 1.30215073e+00 6.29593194e-01
-7.42628276e-01 7.91440308e-01 -1.67824998e-01 2.47144967e-01
7.51915872e-01 -1.32271588e+00 -1.34244084e-01 5.80796540e-01
-7.52500892e-01 9.59008574e-01 7.11438119e-01 -5.02273679e-01
-1.16101217e+00 -1.36199617e+00 1.02607381e+00 -3.04378569e-01
2.10645661e-01 -6.81670085e-02 -9.49151158e-01 5.62714517e-01
-3.28467131e-01 4.07967597e-01 1.09043980e+00 2.25470573e-01
-3.57064009e-02 -7.88488686e-01 -1.66805899e+00 3.91933084e-01
3.58170390e-01 -2.34344706e-01 -3.59229863e-01 3.38951617e-01
2.67980456e-01 -6.71760798e-01 -1.92536223e+00 6.08723640e-01
9.01247203e-01 -7.53988445e-01 1.02008533e+00 -5.02034962e-01
2.34539762e-01 -1.61656365e-01 5.76547347e-02 -1.14086783e+00
-2.67612696e-01 -7.49255657e-01 -1.05690584e-01 9.66577590e-01
5.50293028e-01 -1.14275825e+00 8.63044500e-01 1.66861212e+00
-1.11955581e-02 -7.08024979e-01 -9.63770330e-01 -7.30936706e-01
5.19348085e-01 -2.28008069e-02 2.83927113e-01 7.87711143e-01
2.07378611e-01 1.06074125e-01 -5.93568802e-01 2.12953120e-01
1.06830573e+00 2.99663872e-01 8.67480576e-01 -1.45999169e+00
-4.36240524e-01 2.37819161e-02 -3.16203386e-01 -8.51858139e-01
-4.53817666e-01 -6.91164136e-01 -3.58313560e-01 -1.23885751e+00
-1.25103556e-02 -7.65015960e-01 -7.85298526e-01 4.06621814e-01
-2.14678451e-01 -7.95255452e-02 -4.07190211e-02 7.76222034e-04
3.09514970e-01 8.05156082e-02 1.12288344e+00 -1.64542943e-02
-7.86081970e-01 5.13069034e-01 -6.94813490e-01 2.11388096e-01
1.42842102e+00 -3.28013033e-01 -2.19928086e-01 -3.35495323e-02
-2.02837467e-01 5.00852585e-01 4.65516090e-01 -1.31423616e+00
-1.16926096e-01 -1.56664729e-01 9.73481238e-01 -1.30169570e-01
2.95039117e-01 -8.47095370e-01 3.41298252e-01 1.02105618e+00
-1.94280565e-01 -3.04866731e-01 3.36634487e-01 2.30163679e-01
1.19020335e-01 2.88354397e-01 1.01810944e+00 -6.32116795e-01
3.90858464e-02 2.12289721e-01 -1.03575908e-01 -5.94635345e-02
9.34327364e-01 -2.22979292e-01 -2.13115707e-01 -9.05757546e-02
-1.35385871e+00 2.84277260e-01 -1.09662816e-01 1.59640938e-01
3.41333210e-01 -1.08000398e+00 -8.13855767e-01 4.46847826e-01
-2.00963259e-01 -8.65516961e-02 3.23657125e-01 1.10625052e+00
-2.88085192e-01 5.94709992e-01 -1.46721572e-01 -6.76429331e-01
-1.06194007e+00 2.79184133e-01 8.16036701e-01 2.00237595e-02
-9.15051162e-01 2.45062679e-01 -5.46147943e-01 -8.05594251e-02
4.08993065e-01 -6.61922038e-01 1.34381786e-01 -4.67718303e-01
6.84412837e-01 5.87426782e-01 9.64380130e-02 2.53334552e-01
-2.04998463e-01 6.28518343e-01 1.01680970e-02 2.23980054e-01
1.18470263e+00 -4.44784947e-02 -1.11422859e-01 5.55214226e-01
8.69526148e-01 -1.68879151e-01 -9.15346324e-01 -2.88001388e-01
-1.65565491e-01 -4.95154679e-01 -2.28637248e-01 -1.22681463e+00
-8.26359153e-01 7.84058034e-01 1.16666687e+00 2.46241629e-01
9.78136122e-01 -4.42748219e-01 6.60935342e-01 1.94700733e-01
2.96878278e-01 -9.53891218e-01 -5.19201994e-01 -1.18310735e-01
7.34521568e-01 -1.07809293e+00 1.66957349e-01 -3.67201090e-01
-2.18839586e-01 8.61910760e-01 5.29729903e-01 -2.77083218e-01
1.08943152e+00 1.36662629e-02 4.70990211e-01 8.76193941e-02
-2.89229035e-01 2.91096270e-01 6.61639273e-02 4.73833352e-01
7.62802184e-01 5.64846277e-01 -5.59931219e-01 4.59203601e-01
-2.71177351e-01 4.78704244e-01 3.44437659e-01 9.24843550e-01
-5.36952198e-01 -1.04099178e+00 -1.62061617e-01 9.34138715e-01
-6.44703805e-01 2.07201019e-03 3.54008704e-01 4.24048871e-01
3.44264470e-02 1.10413873e+00 1.31495073e-01 1.06208101e-01
7.52426505e-01 6.45729005e-01 3.60402852e-01 -5.95474422e-01
-5.73830426e-01 4.53205854e-02 -1.85619801e-01 -3.72980505e-01
-4.66099590e-01 -4.89661217e-01 -1.07179523e+00 -3.24622989e-01
2.82906562e-01 1.31698444e-01 5.26034176e-01 5.59844017e-01
5.97968400e-01 4.68244135e-01 7.45374620e-01 -2.60304779e-01
-9.22067702e-01 -1.15978217e+00 -7.85976768e-01 4.95511115e-01
4.26097721e-01 -3.75833839e-01 -3.90868813e-01 2.61953156e-02] | [14.067906379699707, 2.9746549129486084] |
4b6b37d2-1b08-40f4-881f-36d85dc661f8 | set-the-scene-global-local-training-for | 2303.13450 | null | https://arxiv.org/abs/2303.13450v1 | https://arxiv.org/pdf/2303.13450v1.pdf | Set-the-Scene: Global-Local Training for Generating Controllable NeRF Scenes | Recent breakthroughs in text-guided image generation have led to remarkable progress in the field of 3D synthesis from text. By optimizing neural radiance fields (NeRF) directly from text, recent methods are able to produce remarkable results. Yet, these methods are limited in their control of each object's placement or appearance, as they represent the scene as a whole. This can be a major issue in scenarios that require refining or manipulating objects in the scene. To remedy this deficit, we propose a novel GlobalLocal training framework for synthesizing a 3D scene using object proxies. A proxy represents the object's placement in the generated scene and optionally defines its coarse geometry. The key to our approach is to represent each object as an independent NeRF. We alternate between optimizing each NeRF on its own and as part of the full scene. Thus, a complete representation of each object can be learned, while also creating a harmonious scene with style and lighting match. We show that using proxies allows a wide variety of editing options, such as adjusting the placement of each independent object, removing objects from a scene, or refining an object. Our results show that Set-the-Scene offers a powerful solution for scene synthesis and manipulation, filling a crucial gap in controllable text-to-3D synthesis. | ['Daniel Cohen-Or', 'Raja Giryes', 'Gal Metzer', 'Elad Richardson', 'Dana Cohen-Bar'] | 2023-03-23 | null | null | null | null | ['text-to-3d'] | ['computer-vision'] | [ 5.58342934e-01 -9.73653495e-02 2.27980003e-01 -2.83668935e-01
-3.56305391e-01 -8.44627023e-01 7.37091422e-01 4.95466553e-02
-5.62616512e-02 4.84925717e-01 1.39517263e-01 1.43342558e-02
7.66759664e-02 -1.03600240e+00 -7.94948995e-01 -8.07842135e-01
5.98764122e-01 5.68563342e-01 3.12579870e-01 -5.70556819e-01
3.19181323e-01 1.09173656e+00 -1.76419234e+00 1.18665569e-01
7.41993189e-01 7.96835005e-01 6.77415907e-01 8.25211048e-01
-5.15272975e-01 5.36370695e-01 -8.45828295e-01 -1.20119192e-01
5.40743828e-01 -5.32584846e-01 -4.23610210e-01 6.16851091e-01
6.35281861e-01 -2.47402251e-01 -1.37972921e-01 7.47174799e-01
5.10958135e-01 2.31031731e-01 6.40008688e-01 -9.98067677e-01
-5.75335383e-01 1.95979372e-01 -4.77129638e-01 -5.71384907e-01
1.91847757e-01 1.37132555e-01 8.29721630e-01 -6.48465812e-01
7.44635940e-01 1.00720704e+00 2.72601902e-01 5.38702190e-01
-1.56624186e+00 -1.40773579e-01 2.00308606e-01 -3.65566969e-01
-1.25510395e+00 -5.02280176e-01 1.03115594e+00 -5.40525615e-01
6.52153373e-01 6.48636997e-01 7.88329840e-01 6.51452124e-01
5.21476194e-02 5.55736542e-01 9.69037473e-01 -8.28190446e-01
2.82856405e-01 3.04486185e-01 -3.99821192e-01 6.33685052e-01
-1.65206045e-01 -1.97497085e-01 -5.99761009e-01 7.20995069e-02
1.28584194e+00 -1.85067907e-01 -4.23185050e-01 -8.15556824e-01
-1.32148314e+00 5.01786768e-01 3.98757100e-01 5.09467646e-02
-1.30908832e-01 3.51860732e-01 -5.08264974e-02 1.94115698e-01
3.19376707e-01 9.82616305e-01 -4.88360882e-01 2.68656433e-01
-7.46256411e-01 6.24033213e-01 6.06974840e-01 1.22781789e+00
1.04007471e+00 8.22990686e-02 -3.44014704e-01 9.38387156e-01
-7.79707506e-02 5.10562181e-01 9.94683430e-02 -1.18083978e+00
2.42660403e-01 6.54184163e-01 3.21690530e-01 -8.24487746e-01
-2.00640127e-01 -3.43892217e-01 -6.62660539e-01 7.70661652e-01
3.56535256e-01 -4.37869281e-02 -1.19930565e+00 1.68866062e+00
5.72936714e-01 -3.55174720e-01 -3.64277363e-01 6.71959281e-01
5.56747317e-01 8.57694805e-01 -3.15702826e-01 5.00892997e-02
1.17508745e+00 -8.82456899e-01 -5.01770735e-01 -1.18805200e-01
3.94371271e-01 -8.85612309e-01 1.22042108e+00 4.40662622e-01
-1.25782132e+00 -5.38754702e-01 -9.18071508e-01 -3.52711767e-01
-4.42478925e-01 1.68445155e-01 5.33984125e-01 4.32955414e-01
-1.11905944e+00 4.83900249e-01 -3.93907219e-01 -1.37751192e-01
6.82340637e-02 3.50374907e-01 -2.44368091e-01 1.40583172e-01
-7.15557158e-01 8.25678170e-01 2.46394709e-01 -2.84593608e-02
-6.48014426e-01 -8.02894354e-01 -8.16679537e-01 1.04862936e-01
5.57926834e-01 -9.87180769e-01 1.26920927e+00 -1.02177322e+00
-1.75976121e+00 9.72085297e-01 -8.49882215e-02 1.17633991e-01
5.20603597e-01 1.95943378e-02 1.30244434e-01 -7.06865713e-02
4.60714214e-02 9.06675160e-01 1.11375284e+00 -1.73492157e+00
-5.03651619e-01 -2.15830624e-01 2.25908443e-01 4.82400537e-01
-3.70093733e-02 -1.68279216e-01 -7.07722306e-01 -9.36835766e-01
3.30166727e-01 -7.68362343e-01 -4.80962455e-01 5.75814426e-01
-5.79810083e-01 1.17142320e-01 8.92714500e-01 -1.14218362e-01
8.34212661e-01 -2.15650129e+00 3.65903556e-01 1.36891246e-01
1.03684053e-01 5.57959527e-02 -1.70742333e-01 5.27586162e-01
-2.36781430e-03 1.85344517e-01 -4.02649164e-01 -6.91803575e-01
6.01407699e-02 1.77613199e-01 -3.30861330e-01 1.78881094e-01
2.26779029e-01 7.35131860e-01 -7.41426289e-01 -3.46185654e-01
6.56979918e-01 6.70586407e-01 -7.51754999e-01 2.66119301e-01
-7.99148798e-01 7.14437068e-01 -6.01703465e-01 2.87379861e-01
6.14859819e-01 -4.30031233e-02 -6.60915002e-02 -2.01420322e-01
-4.95055586e-01 1.07154474e-01 -1.45552862e+00 2.04858112e+00
-6.40504897e-01 6.20450556e-01 2.81436920e-01 -6.26779735e-01
1.10919726e+00 8.28793719e-02 5.18717766e-01 -5.62525749e-01
5.79737090e-02 -4.17772643e-02 -5.08166909e-01 -2.47499287e-01
8.08049083e-01 -1.72650978e-01 -4.09930348e-02 4.81482655e-01
-2.12916806e-01 -1.07738066e+00 1.60780594e-01 1.27231985e-01
8.09251368e-01 4.06669140e-01 9.74024534e-02 -2.32447341e-01
3.29596370e-01 3.52145098e-02 1.99671313e-01 8.42960954e-01
5.09090900e-01 1.14519310e+00 3.92630607e-01 -2.77285129e-01
-1.30467558e+00 -1.06920362e+00 -1.89464390e-01 9.17689979e-01
2.23902270e-01 -3.45379442e-01 -7.83969879e-01 -3.26796621e-01
-1.73633948e-01 7.44918942e-01 -5.46198845e-01 2.25201353e-01
-8.33625197e-01 -4.51294571e-01 -7.54871592e-02 1.29101261e-01
4.54065740e-01 -1.03191626e+00 -8.67896080e-01 2.13926032e-01
-2.30371486e-02 -1.04604685e+00 -6.22394323e-01 3.66823882e-01
-6.85083866e-01 -7.36002386e-01 -7.19806910e-01 -7.08715022e-01
9.85751748e-01 4.87532169e-01 1.13246250e+00 4.07645777e-02
-4.81872201e-01 3.02817822e-01 -2.40052864e-01 -2.80895144e-01
-5.54155648e-01 3.07979230e-02 -2.49812469e-01 2.32920334e-01
-6.31348670e-01 -7.20667541e-01 -5.04564285e-01 3.87443990e-01
-1.22495985e+00 5.44650853e-01 1.35522351e-01 3.44639987e-01
6.96718276e-01 1.79156184e-01 3.03064045e-02 -7.50616491e-01
2.77848572e-01 1.03813753e-01 -9.10274982e-01 1.85977206e-01
1.62026156e-02 3.70246470e-01 7.95204759e-01 -3.22556436e-01
-1.24675131e+00 4.35209394e-01 3.78034413e-02 -3.24978739e-01
-3.51238459e-01 7.21350014e-02 -4.41813946e-01 -7.20599815e-02
7.04985380e-01 1.20404303e-01 -2.78355509e-01 -5.92381775e-01
6.19753838e-01 2.72440076e-01 6.25167906e-01 -8.22529793e-01
9.44399595e-01 6.11434162e-01 3.09135109e-01 -1.02682304e+00
-7.21503973e-01 -1.62986159e-01 -9.02905762e-01 -1.67214081e-01
9.58991647e-01 -5.81653118e-01 -4.17668790e-01 5.62501371e-01
-1.33958697e+00 -8.00740302e-01 -6.86062574e-01 1.53266629e-02
-6.03389382e-01 -1.08237065e-01 -9.98679101e-02 -5.43811202e-01
1.88528284e-01 -1.37768686e+00 1.45562351e+00 8.37322846e-02
-1.91108674e-01 -8.71173263e-01 -5.70361391e-02 1.35523826e-02
3.98697466e-01 4.89402354e-01 1.08047652e+00 3.11622709e-01
-9.70081091e-01 -8.19108710e-02 -7.94949681e-02 1.84465572e-01
4.29569244e-01 3.61892760e-01 -8.77793074e-01 8.19396973e-03
2.22538449e-02 2.06274986e-02 5.83094835e-01 2.75963008e-01
1.43061948e+00 -9.87666659e-03 -2.06621945e-01 8.45328629e-01
1.56978285e+00 9.45383310e-02 6.04454398e-01 3.00048441e-01
9.27679420e-01 7.30805457e-01 1.87988222e-01 4.76689637e-01
6.42917454e-02 1.15879977e+00 5.65162539e-01 -2.01848269e-01
-5.80803216e-01 -3.49233568e-01 5.64081855e-02 2.39407539e-01
-5.42145781e-03 -6.38793409e-01 -6.38654649e-01 2.08082601e-01
-1.50289643e+00 -7.64777899e-01 -2.66519189e-01 2.31673837e+00
7.34738588e-01 -1.11466341e-01 -2.13509321e-01 1.36066943e-01
6.65456474e-01 3.24294746e-01 -4.05976772e-01 -2.87028879e-01
-3.61014038e-01 2.13922486e-01 3.91940504e-01 6.93037152e-01
-7.99127519e-01 1.02565110e+00 6.49664688e+00 7.40480244e-01
-1.37208998e+00 -3.33733290e-01 4.19051021e-01 -1.85460746e-01
-7.79159307e-01 5.30818701e-02 -8.23287725e-01 1.60659879e-01
9.94361639e-02 9.62517504e-03 5.88762403e-01 5.45812547e-01
5.08490920e-01 -3.23803931e-01 -1.09514809e+00 9.54907238e-01
1.57522202e-01 -1.39992499e+00 3.91036659e-01 7.44144851e-03
9.81514573e-01 -3.89500767e-01 8.84533394e-03 -2.43800417e-01
2.56617695e-01 -8.62636566e-01 1.07627749e+00 5.50241828e-01
9.98453915e-01 -6.16791725e-01 -2.46926710e-01 5.03474057e-01
-9.74309504e-01 2.35383242e-01 -1.75364584e-01 3.92279029e-02
2.76323050e-01 6.87889218e-01 -5.92129588e-01 3.27578217e-01
4.56035554e-01 2.42712423e-01 -4.34130222e-01 8.75927269e-01
-3.74597192e-01 4.30861525e-02 -4.39832211e-01 1.04534447e-01
8.17011297e-02 -3.32833827e-01 6.87970340e-01 9.45860445e-01
3.58860016e-01 2.03877643e-01 1.59516588e-01 1.22257352e+00
-2.69242018e-01 1.35752335e-01 -6.12008929e-01 3.41903985e-01
2.36964658e-01 1.20687211e+00 -8.82030487e-01 -2.85162687e-01
-6.61288053e-02 1.12990177e+00 2.61886179e-01 4.52468485e-01
-8.04248810e-01 -4.84312356e-01 5.13262153e-01 4.25160557e-01
3.14493567e-01 -5.46098650e-01 -5.58212757e-01 -1.09438515e+00
1.72188133e-01 -6.93046212e-01 -2.98882961e-01 -1.31986904e+00
-7.39129841e-01 4.33954656e-01 1.00012898e-01 -1.03231812e+00
1.21391125e-01 -5.61662614e-01 -4.22723413e-01 8.80620837e-01
-1.26869082e+00 -1.10880482e+00 -5.46877325e-01 6.00302696e-01
7.21237957e-01 1.93243340e-01 6.87119663e-01 1.07821375e-01
-2.44382158e-01 2.10177571e-01 4.02448885e-02 -2.20964268e-01
6.22064769e-01 -1.26255810e+00 3.72851104e-01 7.28354931e-01
2.08900422e-01 3.67235065e-01 6.98890090e-01 -4.14309919e-01
-1.44483089e+00 -9.97756958e-01 6.29392445e-01 -5.69543719e-01
2.74735540e-01 -7.85256684e-01 -7.00977087e-01 4.59999830e-01
4.68589850e-02 -2.23296821e-01 5.11577986e-02 -2.75133073e-01
-9.25846696e-02 -2.32042983e-01 -1.03515303e+00 1.05735898e+00
1.08619905e+00 -3.95564526e-01 -1.28411204e-01 5.09648025e-01
7.90742457e-01 -7.31807351e-01 -4.38643128e-01 3.06797355e-01
3.30740213e-01 -1.12091637e+00 9.75030363e-01 -1.50615513e-01
5.46632946e-01 -6.87655568e-01 -1.43808916e-01 -1.46877480e+00
-2.92190492e-01 -8.67906153e-01 2.86339283e-01 1.06115532e+00
2.67938077e-01 -3.69062304e-01 8.27166617e-01 6.83971405e-01
-5.37319064e-01 -4.60020214e-01 -4.51622248e-01 -3.72344285e-01
-3.51153165e-02 -4.59214658e-01 8.06949675e-01 7.93679655e-01
-6.20672584e-01 2.55258143e-01 -2.97174335e-01 1.92644477e-01
4.73503590e-01 2.53975421e-01 1.21675527e+00 -1.01884496e+00
-3.79885525e-01 -5.79438448e-01 -4.13316227e-02 -1.46836138e+00
-1.07122995e-01 -6.71443224e-01 2.71239549e-01 -1.66167247e+00
-1.68234542e-01 -8.65817070e-01 4.78714794e-01 4.47614372e-01
1.52635900e-02 3.50796402e-01 3.79068226e-01 8.64474848e-02
2.16712300e-02 5.08266687e-01 1.77648139e+00 -9.18021053e-02
-5.21287024e-01 -1.34079695e-01 -5.96729279e-01 7.31582999e-01
8.15154433e-01 -1.74482465e-01 -4.55489546e-01 -9.03002203e-01
3.35699439e-01 2.44798139e-02 4.56902355e-01 -8.57548714e-01
1.34961903e-01 -4.49356794e-01 4.09011215e-01 -6.12299860e-01
6.40510798e-01 -8.54049087e-01 4.06386673e-01 -8.57104212e-02
-3.68591815e-01 -2.47917324e-01 1.24606840e-01 1.49108544e-01
8.81905258e-02 -3.52428496e-01 9.38567579e-01 -4.90795255e-01
-5.15285432e-01 3.27827781e-01 -1.41487449e-01 -3.05596679e-01
8.50841105e-01 -4.52276707e-01 -6.91098720e-02 -3.08732182e-01
-7.74525821e-01 4.06279229e-02 9.11585510e-01 3.34372163e-01
3.46080035e-01 -1.18931484e+00 -5.78430653e-01 4.73765135e-01
3.28112692e-02 7.12565660e-01 1.36749089e-01 1.71209887e-01
-7.83024311e-01 1.44983202e-01 -2.73218378e-02 -7.52376378e-01
-1.12192023e+00 3.94786388e-01 5.95477104e-01 8.45052078e-02
-6.23235166e-01 6.81453347e-01 6.62417531e-01 -4.77120876e-01
1.54755279e-01 -3.31286222e-01 2.14254096e-01 -1.22573838e-01
2.89437175e-01 5.07573783e-02 8.25866759e-02 -4.25318867e-01
1.83583707e-01 1.05864227e+00 2.86477804e-01 -3.84397209e-01
1.33837688e+00 -1.35558471e-01 -2.68440127e-01 4.33464736e-01
9.42052245e-01 4.57849026e-01 -1.54934645e+00 -9.75317210e-02
-6.43933892e-01 -7.65703261e-01 1.52965397e-01 -6.71670616e-01
-1.02275622e+00 7.60967493e-01 2.90898923e-02 2.42505565e-01
1.13146949e+00 -3.44039761e-02 4.48869586e-01 3.37614924e-01
4.45389003e-01 -9.15846109e-01 2.55968302e-01 5.21774828e-01
1.18225491e+00 -8.76781046e-01 1.06867909e-01 -6.08277917e-01
-4.29488361e-01 1.18400455e+00 4.12030935e-01 -5.40336631e-02
3.83162707e-01 3.61901432e-01 1.21730253e-01 -1.89480349e-01
-2.90063232e-01 -9.35696959e-02 2.72597909e-01 5.01926184e-01
3.69763941e-01 -1.06144212e-01 2.19858021e-01 -3.37970734e-01
-4.71590370e-01 -3.66519332e-01 6.01086557e-01 9.62126076e-01
-5.41430771e-01 -1.39436018e+00 -5.98525763e-01 1.43253908e-01
-1.19708236e-02 -5.93782738e-02 -3.77181023e-01 7.23429263e-01
3.44217747e-01 5.93253255e-01 5.74248135e-02 9.18890610e-02
5.54439664e-01 -1.71399146e-01 8.36963415e-01 -9.18385088e-01
-3.90871793e-01 3.21251184e-01 -1.74955755e-01 -3.79865855e-01
-3.33117694e-01 -5.28014183e-01 -1.10870242e+00 -1.48849785e-01
-3.53785664e-01 -7.62948766e-02 9.64984059e-01 6.12881303e-01
1.92471474e-01 5.48046052e-01 8.64849806e-01 -1.31589699e+00
3.31212878e-02 -3.64122272e-01 -6.40129328e-01 2.77692944e-01
4.85586643e-01 -5.01876771e-01 -3.33588004e-01 3.44994992e-01] | [9.304643630981445, -3.0763027667999268] |
47246374-bb6f-402d-a36a-0904ea655b9a | nilut-conditional-neural-implicit-3d-lookup | 2306.11920 | null | https://arxiv.org/abs/2306.11920v1 | https://arxiv.org/pdf/2306.11920v1.pdf | NILUT: Conditional Neural Implicit 3D Lookup Tables for Image Enhancement | 3D lookup tables (3D LUTs) are a key component for image enhancement. Modern image signal processors (ISPs) have dedicated support for these as part of the camera rendering pipeline. Cameras typically provide multiple options for picture styles, where each style is usually obtained by applying a unique handcrafted 3D LUT. Current approaches for learning and applying 3D LUTs are notably fast, yet not so memory-efficient, as storing multiple 3D LUTs is required. For this reason and other implementation limitations, their use on mobile devices is less popular. In this work, we propose a Neural Implicit LUT (NILUT), an implicitly defined continuous 3D color transformation parameterized by a neural network. We show that NILUTs are capable of accurately emulating real 3D LUTs. Moreover, a NILUT can be extended to incorporate multiple styles into a single network with the ability to blend styles implicitly. Our novel approach is memory-efficient, controllable and can complement previous methods, including learned ISPs. Code, models and dataset available at: https://github.com/mv-lab/nilut | ['Radu Timofte', 'Michael S. Brown', 'Javier Vazquez-Corral', 'Marcos V. Conde'] | 2023-06-20 | null | null | null | null | ['color-manipulation', 'photo-retouching', 'image-enhancement', 'tone-mapping'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.13702583e-01 -4.34964985e-01 -1.75655082e-01 -4.77346301e-01
-6.04280412e-01 -7.08077312e-01 3.81928027e-01 -1.13859951e-01
-3.12506557e-01 3.46221119e-01 -1.85073957e-01 -7.20320404e-01
3.28016758e-01 -7.29368091e-01 -9.63470697e-01 -3.81942153e-01
1.26530245e-01 1.41426280e-01 3.06674361e-01 -7.70941675e-02
2.27356225e-01 1.02599692e+00 -1.36270332e+00 2.47649863e-01
4.20064628e-01 1.29468203e+00 6.44534379e-02 1.06153607e+00
-3.02236378e-01 7.09223390e-01 -5.83005726e-01 -5.00293076e-01
6.65370047e-01 -2.03091487e-01 -2.72282243e-01 -5.54893985e-02
9.41383719e-01 -7.77449191e-01 -3.84232700e-01 1.03700912e+00
5.63518941e-01 -1.58585399e-01 3.91914397e-01 -1.36295259e+00
-4.99543846e-01 3.92774612e-01 -5.64460397e-01 -9.69562083e-02
-3.42319831e-02 2.46460855e-01 6.37923837e-01 -8.23637426e-01
2.95713454e-01 1.30525422e+00 7.87339151e-01 6.37243032e-01
-1.46056449e+00 -7.85290599e-01 9.22322497e-02 4.80803242e-03
-1.35944104e+00 -4.96745676e-01 8.93861353e-01 -1.78775668e-01
9.86392140e-01 4.16767567e-01 9.26827133e-01 8.99343967e-01
1.93674028e-01 8.02041471e-01 1.36271226e+00 -4.59347188e-01
2.01865420e-01 2.01520339e-01 1.28932953e-01 8.10772955e-01
2.98005402e-01 -6.43149465e-02 -8.17653716e-01 -2.41043065e-02
1.45484149e+00 -2.70986836e-02 -2.04793543e-01 -5.17883301e-01
-9.66809154e-01 4.39730436e-01 5.05111814e-01 -1.90868512e-01
6.77464232e-02 8.14078510e-01 3.75360847e-01 3.31275761e-01
2.88543791e-01 2.21380726e-01 -5.40819824e-01 -2.86544174e-01
-9.98621643e-01 1.37650475e-01 8.23383033e-01 1.17675149e+00
1.17711520e+00 2.94957578e-01 -5.85506484e-02 6.60091221e-01
4.66434769e-02 5.84704161e-01 2.63936013e-01 -1.36618197e+00
1.01909161e-01 6.64585590e-01 -6.42829761e-02 -7.73310661e-01
-3.23928028e-01 -3.25650156e-01 -9.35690820e-01 7.63924897e-01
2.87394762e-01 1.68192595e-01 -1.04051983e+00 1.52275860e+00
1.35886505e-01 3.40434432e-01 -3.93797040e-01 7.94160783e-01
6.67929888e-01 6.65958881e-01 -1.53270379e-01 2.58399159e-01
1.29419959e+00 -9.03562307e-01 -4.55971211e-01 -1.78158000e-01
2.08026499e-01 -8.88696969e-01 1.55517578e+00 6.38319910e-01
-1.28941882e+00 -6.47211611e-01 -1.38545573e+00 -5.99525571e-01
-5.95818818e-01 3.54392409e-01 8.83085430e-01 1.12350368e+00
-1.59863520e+00 5.91235578e-01 -1.00407004e+00 1.35549000e-02
4.76538509e-01 7.57007957e-01 4.76494767e-02 1.58526093e-01
-7.11601436e-01 6.79281890e-01 1.53492451e-01 1.19623601e-01
-6.28183365e-01 -7.64317393e-01 -6.55964792e-01 -1.02624604e-02
1.40225604e-01 -8.10806751e-01 1.47812533e+00 -1.03486550e+00
-1.98190296e+00 9.50077295e-01 -1.38150021e-01 -5.59084833e-01
4.71734732e-01 -2.29101658e-01 -1.63241193e-01 6.47251457e-02
-3.74879122e-01 9.63075280e-01 1.15758348e+00 -1.16661239e+00
-4.00755197e-01 -7.69585511e-03 4.10319358e-01 2.43270434e-02
-5.63497722e-01 -7.75397345e-02 -9.99809742e-01 -6.81017458e-01
-2.72240527e-02 -8.75469089e-01 -3.18872601e-01 7.99499035e-01
-3.03779632e-01 4.04512882e-01 9.43020225e-01 -4.89724666e-01
1.07459152e+00 -2.11132574e+00 -1.89453915e-01 1.91696405e-01
3.31666440e-01 4.26919192e-01 5.56151010e-03 -9.28491652e-02
2.76364595e-01 1.27075076e-01 -2.17943698e-01 -6.94115758e-01
3.00356388e-01 3.13288718e-01 -2.97521919e-01 1.36428326e-01
3.46079290e-01 8.35910916e-01 -5.36078930e-01 -2.48973787e-01
5.33787310e-01 8.80296350e-01 -7.64750719e-01 4.82206084e-02
-2.02011406e-01 9.28639397e-02 -9.60280970e-02 7.65739024e-01
9.82207358e-01 -4.16712463e-01 2.21080527e-01 -6.49792314e-01
-2.41021231e-01 3.10378224e-01 -1.32355452e+00 1.75825119e+00
-6.55271053e-01 9.39063072e-01 7.32077062e-02 -4.74203110e-01
8.60757947e-01 -3.43470573e-02 6.67629540e-02 -6.00675106e-01
3.05127829e-01 3.84136677e-01 -4.92965341e-01 2.33088359e-01
8.45009625e-01 3.20122451e-01 -1.26707274e-02 3.96567196e-01
1.64832234e-01 -4.62368190e-01 8.57884288e-02 1.80244774e-01
9.25470233e-01 3.76058131e-01 1.58756241e-01 -9.57903937e-02
3.41095775e-01 -2.52034783e-01 4.26017791e-01 7.55633771e-01
-1.56054115e-02 6.86501741e-01 7.15342045e-01 -6.33307397e-01
-1.17893052e+00 -1.20703948e+00 -1.51143536e-01 6.87432170e-01
1.89960450e-01 -4.78743821e-01 -7.96894610e-01 -2.90896565e-01
-1.29581586e-01 4.80870068e-01 -3.66147161e-01 8.27780813e-02
-8.20766509e-01 -4.91745323e-01 7.03927279e-01 6.37703717e-01
6.97459400e-01 -6.12475514e-01 -9.37383115e-01 1.29951656e-01
4.37731445e-01 -1.28192806e+00 -6.65698946e-01 4.72057551e-01
-1.15006149e+00 -8.02290976e-01 -6.82823062e-01 -5.50976813e-01
8.26322854e-01 2.49993041e-01 1.23712862e+00 3.73683833e-02
-1.17597453e-01 4.73890573e-01 1.13643654e-01 -2.50828892e-01
-3.72661680e-01 -6.09289296e-03 1.19442018e-02 -5.68981171e-02
1.81689203e-01 -7.81077683e-01 -7.83792019e-01 1.56677976e-01
-9.19378340e-01 5.97170055e-01 5.78544140e-01 7.75122583e-01
9.02770042e-01 -1.87109813e-01 -9.48992670e-02 -9.50432301e-01
4.48012710e-01 2.86796927e-01 -1.05912936e+00 6.54356480e-02
-5.38958669e-01 1.53464809e-01 6.19775712e-01 -6.56626344e-01
-9.70424592e-01 3.16428572e-01 -2.23001510e-01 -6.43481374e-01
-9.43244100e-02 -1.17317311e-01 -3.08375061e-01 -7.06999063e-01
4.45166022e-01 -7.24602267e-02 -1.44938722e-01 -5.58576167e-01
4.97769445e-01 4.58062708e-01 5.97052932e-01 -7.23163426e-01
8.67381275e-01 4.14945453e-01 1.94162920e-01 -6.18121088e-01
-4.25840557e-01 1.68142049e-03 -4.90768433e-01 -2.92786598e-01
5.32570302e-01 -9.53703105e-01 -7.79109061e-01 7.56637096e-01
-1.23956561e+00 -8.27961862e-01 -3.75889122e-01 8.87721628e-02
-5.05594432e-01 1.44605726e-01 -8.37109685e-01 -5.24596930e-01
-3.59305292e-01 -1.62876832e+00 1.17930520e+00 5.14371812e-01
-4.76957858e-02 -8.85240555e-01 -4.41334635e-01 -6.81858137e-02
6.28155947e-01 1.71428636e-01 9.89210188e-01 1.48874223e-01
-1.10564494e+00 -1.38759357e-03 -6.17511213e-01 5.57072222e-01
1.40323624e-01 1.89726979e-01 -1.19257796e+00 -1.84915647e-01
-6.36612102e-02 -2.16514990e-01 6.41044140e-01 4.90875751e-01
1.40909195e+00 -3.17651153e-01 1.21366596e-02 1.20710802e+00
1.56512451e+00 1.56699196e-01 5.30949891e-01 4.09268171e-01
7.92769909e-01 -7.89592341e-02 -6.89529106e-02 4.56672221e-01
3.67316753e-01 7.32255220e-01 3.29973102e-01 -4.22880739e-01
-5.13619959e-01 -2.35502586e-01 4.02843118e-01 8.57461929e-01
-4.16079834e-02 -5.36707081e-02 -6.98611856e-01 4.68012467e-02
-1.38579154e+00 -4.27129835e-01 5.10716764e-03 2.23172092e+00
1.04201102e+00 4.47065800e-01 -1.50543913e-01 2.28014603e-01
3.83194089e-01 1.01733431e-01 -7.97397792e-01 -6.68697894e-01
-3.05720955e-01 7.40296483e-01 9.49508727e-01 4.45014089e-01
-9.05067861e-01 8.01629066e-01 6.19668579e+00 8.11017096e-01
-1.49757648e+00 2.60038255e-03 9.22807872e-01 -2.18078151e-01
-3.99520069e-01 -2.26420145e-02 -9.66189325e-01 2.92741865e-01
9.34502423e-01 1.04583979e-01 6.98092043e-01 6.45495236e-01
1.45726502e-01 -4.07416858e-02 -1.13646293e+00 1.34909379e+00
-2.19723973e-02 -1.53399801e+00 3.48620445e-01 -8.47426355e-02
6.43857718e-01 -2.69006163e-01 5.23895383e-01 6.48955852e-02
5.51129207e-02 -6.93342447e-01 9.34005857e-01 3.21399599e-01
1.37161672e+00 -7.71165609e-01 1.75123796e-01 -1.76455930e-01
-1.13725376e+00 1.72343656e-01 -2.75132954e-01 1.75017908e-01
1.72961801e-01 5.18036246e-01 -5.92145205e-01 7.63606280e-02
8.99486423e-01 6.21127427e-01 -8.72287810e-01 1.05845499e+00
-2.16199219e-01 4.81651694e-01 -3.90366942e-01 -2.27643941e-02
7.83981085e-02 -1.46201462e-01 2.18022466e-01 1.25059068e+00
5.25042713e-01 -1.54372737e-01 -2.67231017e-01 8.93171847e-01
-4.41142052e-01 -1.43859804e-01 -2.44023040e-01 2.49914471e-02
3.92430633e-01 1.20135379e+00 -9.94260967e-01 -3.82817239e-01
-5.18411636e-01 1.25855637e+00 1.05715290e-01 3.05276066e-01
-8.71906817e-01 -5.11456668e-01 9.27830458e-01 -7.86543731e-03
1.50961533e-01 -5.68529963e-01 -5.95762074e-01 -1.19128442e+00
-2.60633528e-02 -9.87696350e-01 -2.00171292e-01 -1.01355922e+00
-9.79102314e-01 5.59414268e-01 -1.29639223e-01 -1.42878759e+00
9.09243375e-02 -1.26920795e+00 -1.20329268e-01 8.09566557e-01
-1.74876285e+00 -1.16841292e+00 -5.10465801e-01 6.08757734e-01
5.16846478e-01 9.69784148e-03 8.49523962e-01 6.39366984e-01
-4.27251428e-01 8.12183142e-01 -1.91492170e-01 -1.71977431e-02
8.69784415e-01 -1.29476452e+00 9.02035534e-01 8.13679397e-01
2.63707072e-01 6.70009077e-01 3.84574860e-01 -2.45168805e-01
-1.94465768e+00 -9.02787149e-01 3.30987543e-01 -1.96534529e-01
5.09254873e-01 -6.85676396e-01 -7.51347303e-01 6.60751998e-01
2.85461962e-01 6.96758106e-02 6.32200301e-01 -3.16260457e-01
-5.01497269e-01 -4.47959393e-01 -1.01385331e+00 8.85188341e-01
9.88577902e-01 -6.91864371e-01 1.85453326e-01 -5.07260533e-03
7.89253533e-01 -1.05579340e+00 -5.93213856e-01 1.24903351e-01
6.94798768e-01 -1.22838414e+00 1.22420454e+00 6.55262023e-02
2.54332483e-01 -5.76855659e-01 -2.58909702e-01 -8.75053823e-01
3.23506556e-02 -8.49101067e-01 -5.67293763e-01 1.00005400e+00
2.64305085e-01 -7.15837657e-01 8.85950387e-01 7.42594779e-01
-1.64855957e-01 -7.39427507e-01 -7.48148620e-01 -5.62089801e-01
-2.48650312e-01 -7.12787092e-01 9.01637137e-01 5.28735518e-01
-7.11636245e-01 -3.41579020e-02 -5.49572289e-01 3.28332424e-01
7.80235469e-01 -6.65567964e-02 9.11692679e-01 -7.78074503e-01
-6.31626606e-01 -8.28469872e-01 -5.41135967e-01 -1.59651017e+00
-2.80412018e-01 -5.78796864e-01 -3.37207168e-01 -1.18686640e+00
-1.90383002e-01 -7.82861948e-01 -3.22309792e-01 4.55526859e-01
2.05966264e-01 7.52819121e-01 3.86876255e-01 -6.72653168e-02
-2.18330279e-01 2.45712221e-01 1.19536543e+00 -2.24804640e-01
-2.95796484e-01 -7.42808431e-02 -6.13266528e-01 8.11155856e-01
1.06097686e+00 -1.47100240e-01 -4.60661322e-01 -8.85298908e-01
3.45275015e-01 -1.40397921e-01 4.77761447e-01 -1.23515701e+00
3.46959889e-01 1.65107295e-01 6.64132059e-01 -6.62272871e-01
8.20714712e-01 -8.72622788e-01 3.09554785e-01 2.29426056e-01
-1.14691317e-01 2.02649117e-01 6.58015370e-01 3.05275973e-02
-4.99659888e-02 -1.39731213e-01 7.12226987e-01 -4.85865250e-02
-9.06555891e-01 4.26679462e-01 -2.04280511e-01 -2.40698308e-01
5.28429985e-01 -4.64638531e-01 -1.84317872e-01 -2.63644129e-01
-3.45302880e-01 -2.96351522e-01 6.77650750e-01 5.45261540e-02
7.76607454e-01 -1.42589891e+00 -1.01467133e-01 5.46889603e-01
-3.03062409e-01 1.95317180e-03 1.79162040e-01 3.73982459e-01
-1.02532196e+00 1.10482886e-01 -3.89609694e-01 -6.13876164e-01
-1.19296598e+00 1.93779901e-01 4.85494018e-01 -1.12014942e-01
-4.49156702e-01 9.02847469e-01 9.99535918e-02 -3.04416358e-01
3.75888407e-01 -6.99731946e-01 4.43662643e-01 -2.17776105e-01
6.27599657e-01 2.83284396e-01 1.36220604e-01 -2.31866434e-01
-7.67399231e-03 8.97919893e-01 1.36375427e-01 -1.17923938e-01
1.28347313e+00 6.47781417e-03 -1.74713865e-01 3.79453331e-01
1.18896580e+00 5.63827856e-03 -1.64017034e+00 -1.33186311e-01
-3.82603139e-01 -5.75394750e-01 2.31722504e-01 -6.70381248e-01
-1.34758818e+00 1.00752282e+00 7.92526603e-01 -1.30293608e-01
1.68134415e+00 -4.33876306e-01 9.22687113e-01 4.15888458e-01
5.55295944e-01 -7.92853832e-01 -5.56649342e-02 4.10047382e-01
5.93672931e-01 -8.76015127e-01 7.49144182e-02 -4.26175237e-01
-2.24050328e-01 1.40582407e+00 6.09359264e-01 -1.18099593e-01
7.10040867e-01 8.52470696e-01 2.99955666e-01 2.97278255e-01
-5.44945717e-01 1.36776939e-01 2.46486217e-01 5.25306463e-01
4.29615885e-01 -1.72350276e-02 8.12097564e-02 2.91096598e-01
-2.48857617e-01 8.95412713e-02 6.42835677e-01 9.77604508e-01
4.54888530e-02 -1.53818417e+00 -2.29760930e-01 4.30898517e-01
-3.72059047e-01 -2.92098045e-01 7.50304833e-02 6.06739044e-01
1.81645736e-01 3.47798496e-01 1.63640201e-01 -4.21414405e-01
3.03070039e-01 -2.30004236e-01 7.78074384e-01 -2.64255494e-01
-7.11197615e-01 5.70682548e-02 -1.30617499e-01 -7.90786505e-01
-4.28068995e-01 -4.21862394e-01 -9.77876067e-01 -4.85173166e-01
3.91715430e-02 -4.94837791e-01 9.62734461e-01 1.32541373e-01
3.94946128e-01 6.19955540e-01 4.86082166e-01 -1.34986496e+00
-1.83008641e-01 -1.73126161e-01 -4.04453933e-01 -9.63194221e-02
4.58392739e-01 -4.39409345e-01 -2.95033753e-01 3.88966858e-01] | [10.553651809692383, -2.0437729358673096] |
fe14a2af-c106-41a8-8f4e-2b0f715a4e45 | estimate-then-optimize-versus-integrated | 2304.06833 | null | https://arxiv.org/abs/2304.06833v1 | https://arxiv.org/pdf/2304.06833v1.pdf | Estimate-Then-Optimize Versus Integrated-Estimation-Optimization: A Stochastic Dominance Perspective | In data-driven stochastic optimization, model parameters of the underlying distribution need to be estimated from data in addition to the optimization task. Recent literature suggests the integration of the estimation and optimization processes, by selecting model parameters that lead to the best empirical objective performance. Such an integrated approach can be readily shown to outperform simple ``estimate then optimize" when the model is misspecified. In this paper, we argue that when the model class is rich enough to cover the ground truth, the performance ordering between the two approaches is reversed for nonlinear problems in a strong sense. Simple ``estimate then optimize" outperforms the integrated approach in terms of stochastic dominance of the asymptotic optimality gap, i,e, the mean, all other moments, and the entire asymptotic distribution of the optimality gap is always better. Analogous results also hold under constrained settings and when contextual features are available. We also provide experimental findings to support our theory. | ['Yunfan Zhao', 'Haofeng Zhang', 'Henry Lam', 'Adam N. Elmachtoub'] | 2023-04-13 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [ 4.64297645e-02 1.17746331e-01 -4.81048375e-01 -3.44940960e-01
-1.19841325e+00 -8.07645917e-01 3.36508870e-01 6.49635494e-02
-4.65379536e-01 9.13891435e-01 2.23027825e-01 -2.06833467e-01
-6.84362829e-01 -2.46187881e-01 -6.95683002e-01 -9.48745131e-01
1.45376891e-01 6.38280988e-01 -2.72202700e-01 7.06173554e-02
4.16478783e-01 1.76341787e-01 -1.27554917e+00 -2.88034916e-01
8.97440076e-01 1.18494415e+00 3.74013960e-01 5.69568276e-01
1.27524108e-01 2.20150754e-01 -3.38691652e-01 -2.96078414e-01
4.86855090e-01 -3.15230995e-01 -3.94954205e-01 2.71470755e-01
2.36561969e-01 -1.86292022e-01 1.22462198e-01 1.27533293e+00
5.92240155e-01 2.43620053e-01 6.03777766e-01 -1.31179285e+00
-4.03390169e-01 4.28544730e-01 -5.65950394e-01 2.18639597e-01
2.44630888e-01 7.94600695e-02 1.29802716e+00 -9.54904079e-01
4.54234570e-01 1.07303333e+00 8.08451414e-01 2.26776123e-01
-1.57357371e+00 -3.41543078e-01 4.93754536e-01 -2.70126134e-01
-1.47148156e+00 -6.85992002e-01 6.02375865e-01 -5.31375825e-01
6.22680068e-01 1.28910154e-01 2.63544410e-01 7.52673864e-01
1.12778060e-01 5.66436768e-01 9.72920358e-01 -4.70248669e-01
2.84295529e-01 5.51504791e-01 2.23294526e-01 5.55715919e-01
5.63836396e-01 4.03153121e-01 -6.27190709e-01 -3.64636034e-01
3.41884106e-01 -2.79808372e-01 -3.02768856e-01 -4.41377550e-01
-9.03802395e-01 8.53943348e-01 -2.13568613e-01 -4.35629673e-02
-4.90435958e-01 2.49624550e-01 1.08646549e-01 2.46022522e-01
6.84497416e-01 7.00734496e-01 -5.63517451e-01 -2.29136676e-01
-1.10252774e+00 4.41184700e-01 1.02851284e+00 1.13478184e+00
6.43580854e-01 -1.90772885e-03 -1.75453514e-01 7.30882406e-01
4.06017005e-01 6.01298749e-01 2.84291744e-01 -1.09311771e+00
7.53760934e-01 1.85552239e-01 7.24722683e-01 -8.61082971e-01
-4.01121080e-01 -9.81365085e-01 -5.22508681e-01 -2.57039815e-01
7.50895500e-01 -5.87718248e-01 -5.03095269e-01 1.96114600e+00
3.84917438e-01 -7.79847801e-02 6.75603151e-02 9.44327116e-01
1.88046142e-01 5.44590056e-01 -2.36913130e-01 -4.56225663e-01
9.86845791e-01 -5.77226520e-01 -8.29276085e-01 -3.63975555e-01
5.67710400e-01 -7.13964522e-01 1.14526677e+00 1.93671331e-01
-1.03020895e+00 -1.68404877e-01 -9.72763777e-01 3.08737814e-01
1.49536073e-01 1.19139694e-01 3.77982736e-01 5.77183485e-01
-8.56375694e-01 5.38298845e-01 -5.53527236e-01 1.98316760e-02
1.65310666e-01 4.49461401e-01 1.86282303e-02 1.42100602e-01
-8.70013952e-01 6.98321044e-01 3.58049005e-01 2.64797658e-01
-8.67738128e-01 -8.51873457e-01 -6.60169899e-01 1.90633059e-01
7.55223393e-01 -6.71598971e-01 1.38435888e+00 -1.02071631e+00
-1.35597265e+00 5.85086823e-01 -5.12907624e-01 -5.76892793e-01
8.72442484e-01 -4.33834583e-01 1.42992020e-01 -8.55136365e-02
2.17472345e-01 -1.06200241e-01 7.76575446e-01 -1.28174853e+00
-6.91481411e-01 -2.81744480e-01 8.23787451e-02 3.60761404e-01
-9.78929400e-02 -7.21219033e-02 -4.06516075e-01 -4.93598849e-01
-1.60483755e-02 -1.14117670e+00 -5.71104825e-01 -9.01797637e-02
-7.12722838e-01 -8.83364584e-03 3.02863300e-01 -5.10216296e-01
1.57529294e+00 -2.18175697e+00 -5.26492856e-02 5.91150343e-01
7.61000142e-02 -2.92580187e-01 7.24481046e-02 6.18067384e-01
-4.67396565e-02 2.84475267e-01 -3.13269526e-01 -4.85292226e-01
2.84437269e-01 9.31616947e-02 -3.74793470e-01 9.14175808e-01
-6.98702112e-02 6.46099210e-01 -6.21623099e-01 -3.84459198e-01
1.99769046e-02 -9.95370522e-02 -7.06421793e-01 1.44475818e-01
-1.97633922e-01 4.84745413e-01 -6.26846611e-01 3.48464012e-01
5.15136123e-01 -5.27820230e-01 1.72577485e-01 1.36254698e-01
-7.29535818e-02 3.78910750e-02 -1.55239856e+00 1.19738078e+00
-6.16807282e-01 6.72989845e-01 3.59757513e-01 -1.21283305e+00
6.86800182e-01 3.11153322e-01 7.06229806e-01 -9.99501422e-02
-1.16935782e-01 2.51260161e-01 5.70953488e-02 -4.20058578e-01
2.94122487e-01 -3.81255358e-01 -1.25820311e-02 3.41603369e-01
-2.23920524e-01 -3.84767205e-02 5.44008054e-02 -7.34247044e-02
5.65448880e-01 3.87166603e-03 5.52398562e-01 -7.89844751e-01
2.98373699e-01 4.94072624e-02 6.14872754e-01 1.15551615e+00
-5.00240438e-02 5.52245915e-01 7.11840630e-01 8.32417756e-02
-9.31110501e-01 -9.82688606e-01 -3.45991880e-01 9.39586222e-01
2.84761667e-01 -1.21781714e-01 -7.86814749e-01 -5.09463608e-01
2.49561355e-01 9.49968815e-01 -6.52606487e-01 5.00057340e-02
-4.59050775e-01 -9.48891461e-01 -7.63663501e-02 3.63830239e-01
1.46944940e-01 -3.13611299e-01 -2.82601565e-01 2.91430265e-01
-1.32562682e-01 -1.22032487e+00 -8.29732180e-01 7.38283843e-02
-8.32064331e-01 -9.85483289e-01 -6.38753295e-01 -2.48899952e-01
6.17946744e-01 6.76756026e-03 1.07713151e+00 -8.65482092e-02
2.17134714e-01 4.11349624e-01 1.05316509e-02 -5.20016134e-01
-1.29654154e-01 2.19612122e-02 1.58008814e-01 2.10829198e-01
-9.18613970e-02 -4.66992021e-01 -5.92421353e-01 4.11189765e-01
-6.42519057e-01 -2.46647939e-01 4.07111347e-01 9.18273509e-01
8.10819864e-01 1.56499371e-01 6.82273090e-01 -9.03073430e-01
7.39673376e-01 -6.79687083e-01 -9.24399078e-01 2.72475332e-01
-9.72160459e-01 3.44873607e-01 6.50395393e-01 -4.47032779e-01
-1.12329888e+00 -3.63038555e-02 9.67998207e-02 -4.04241651e-01
6.39529750e-02 8.41782928e-01 -3.08915347e-01 2.14976013e-01
3.91510189e-01 1.92056268e-01 -7.59220868e-02 -6.41444206e-01
2.16255069e-01 4.49639231e-01 3.47386539e-01 -9.28727627e-01
6.48225427e-01 5.28963923e-01 1.57902658e-01 -9.08889771e-01
-1.25618279e+00 -4.99107331e-01 -3.88685167e-01 -3.65179628e-02
3.61809283e-01 -6.97775424e-01 -5.51060259e-01 -5.34835719e-02
-8.94048095e-01 -1.28200665e-01 -6.86424732e-01 8.24914277e-01
-1.01943541e+00 1.74940020e-01 -3.11949234e-02 -1.31632149e+00
3.41953225e-02 -1.25487149e+00 1.08823442e+00 8.43050554e-02
-1.96688950e-01 -1.28656030e+00 -9.17893369e-03 -1.21744890e-02
1.89595565e-01 3.43790382e-01 8.77913415e-01 -1.05942535e+00
-3.92427474e-01 -4.99816060e-01 -1.66649789e-01 3.30177754e-01
4.62776534e-02 1.93148851e-02 -7.05548942e-01 -2.88165271e-01
2.66825348e-01 2.41633981e-01 6.74461722e-01 1.13396609e+00
1.06454563e+00 -7.00408280e-01 -3.00625414e-01 7.28477359e-01
1.66704404e+00 1.17921755e-01 5.95156699e-02 1.41434535e-01
3.87157083e-01 7.83077419e-01 6.71787918e-01 7.98184454e-01
3.40815067e-01 7.22973466e-01 1.99151576e-01 1.96203813e-01
2.80084431e-01 -3.07908684e-01 2.73905009e-01 5.05223811e-01
1.86095923e-01 -4.15417731e-01 -8.77261043e-01 6.45658076e-01
-1.95037615e+00 -9.09136534e-01 2.39201576e-01 2.51960540e+00
6.79669261e-01 6.26039878e-02 2.12026790e-01 -2.30872989e-01
7.93400943e-01 7.51143470e-02 -7.35444427e-01 -2.04317927e-01
-1.45703018e-01 -1.33551806e-01 8.06419253e-01 9.06428993e-01
-8.66355956e-01 5.92552900e-01 7.35512304e+00 9.21013176e-01
-7.75635898e-01 6.17878735e-02 6.40453994e-01 -3.72517139e-01
-3.97464901e-01 2.82581031e-01 -9.96338725e-01 6.73940241e-01
8.72130632e-01 -5.93369007e-01 4.28070068e-01 8.25787842e-01
6.25823617e-01 -3.01653653e-01 -1.35257971e+00 8.02253842e-01
-3.37965757e-01 -1.23843312e+00 -3.58147770e-01 4.69530761e-01
9.28551972e-01 -2.24764064e-01 2.51083463e-01 -1.51164621e-01
4.40688580e-01 -1.00253236e+00 8.89374912e-01 8.41342747e-01
6.41629457e-01 -7.96856523e-01 8.04411829e-01 6.84077919e-01
-9.87175345e-01 -2.70599008e-01 -1.36105612e-01 -3.65147889e-02
3.39534283e-01 9.12895679e-01 -6.60549641e-01 3.98645610e-01
4.32940274e-01 5.58460593e-01 -1.42936289e-01 1.17687023e+00
4.30386551e-02 7.10442960e-01 -4.61328566e-01 -9.78487059e-02
2.00242609e-01 -5.08090556e-01 1.06055403e+00 1.09114838e+00
2.50050396e-01 -1.59431890e-01 4.15552020e-01 6.27396464e-01
2.66620249e-01 2.60962665e-01 -6.44332230e-01 -1.24935560e-01
5.25427938e-01 6.10402584e-01 -5.05366743e-01 -2.32389867e-01
-3.84572536e-01 4.22462910e-01 1.79219529e-01 7.82499492e-01
-6.88953578e-01 -2.14815751e-01 7.71514237e-01 -5.91459125e-02
3.28322202e-01 -3.09633136e-01 -4.88379687e-01 -1.09103537e+00
2.96294123e-01 -7.21032381e-01 4.15257961e-01 -2.41895214e-01
-1.29378486e+00 6.78598732e-02 3.39909285e-01 -1.03777659e+00
-5.85869491e-01 -6.02759719e-01 -3.69685739e-01 9.10271883e-01
-1.23690045e+00 -4.49680060e-01 2.91237831e-01 2.70589501e-01
5.85939229e-01 1.92406110e-03 3.83526981e-01 7.01054111e-02
-7.90550292e-01 6.41672194e-01 7.52201736e-01 -1.90041780e-01
3.59906346e-01 -1.24263680e+00 -2.50782788e-01 8.39110851e-01
9.40367281e-02 6.45958424e-01 1.20581424e+00 -5.21309853e-01
-1.30882120e+00 -7.43528843e-01 6.59551561e-01 -4.21292484e-01
8.25483263e-01 -2.39662051e-01 -5.71698487e-01 5.29119372e-01
-1.07815459e-01 -1.95143372e-01 7.46131957e-01 3.59635055e-01
4.29390520e-02 -2.09365144e-01 -1.12082076e+00 4.83085632e-01
8.05979967e-01 -3.53362918e-01 -2.04024956e-01 5.97975969e-01
6.43379807e-01 -3.52205753e-01 -9.02311742e-01 3.38768333e-01
7.82948911e-01 -8.66419375e-01 8.07371557e-01 -9.75214362e-01
2.86609381e-01 -1.96256228e-02 -7.75595486e-01 -1.22431338e+00
3.19362469e-02 -9.46075618e-01 -3.77332330e-01 1.06816304e+00
7.74873137e-01 -9.55497026e-01 8.06182027e-01 7.96518624e-01
1.23959959e-01 -1.15975034e+00 -1.08672261e+00 -1.05795407e+00
7.31790364e-02 -7.40959764e-01 5.72149634e-01 5.02340674e-01
-4.75941420e-01 1.92240804e-01 -4.21050131e-01 4.23118055e-01
7.29901135e-01 4.42869067e-01 6.40912592e-01 -1.04333377e+00
-6.72475517e-01 -6.25341535e-01 -8.83004144e-02 -1.34167743e+00
2.69805998e-01 -6.64528251e-01 2.96834737e-01 -1.22515690e+00
2.04987049e-01 -5.73441803e-01 -2.45129690e-01 -2.37323597e-01
-2.75285989e-01 -4.27932322e-01 1.95942283e-01 3.71035099e-01
-3.53597254e-01 4.87880081e-01 1.14218497e+00 2.70889282e-01
-2.97743708e-01 5.77358007e-01 -1.08195901e+00 8.95170808e-01
7.08655536e-01 -5.55148542e-01 -5.59329271e-01 -2.07786411e-01
2.10657045e-01 3.35007757e-01 2.61089414e-01 -4.36974645e-01
6.84742779e-02 -6.07224822e-01 1.36322275e-01 -3.77184719e-01
3.56660008e-01 -8.88782978e-01 5.85573353e-02 2.76974857e-01
-7.13096023e-01 1.71676248e-01 1.85218211e-02 9.02636170e-01
1.91592835e-02 -5.18557549e-01 7.55734801e-01 1.48195624e-01
-9.40414295e-02 4.19318050e-01 -1.34782881e-01 5.33182681e-01
9.04339254e-01 -3.55177402e-01 6.17294870e-02 -9.12568152e-01
-8.20133984e-01 3.44436973e-01 3.32395643e-01 -3.88371274e-02
4.00209904e-01 -1.12269735e+00 -7.82323539e-01 -8.86794850e-02
-1.29277080e-01 -1.63944393e-01 4.76441521e-04 1.26684523e+00
3.79646160e-02 5.70595562e-01 7.08610058e-01 -5.68622053e-01
-8.88677657e-01 4.36508507e-01 5.66328764e-01 -4.68523890e-01
-2.57560760e-01 7.20462799e-01 4.71611738e-01 -1.84315085e-01
3.06111157e-01 -3.02807480e-01 3.18486869e-01 1.30402103e-01
2.32654303e-01 5.25520384e-01 -1.77442044e-01 -6.37665391e-01
-3.18733722e-01 5.73437989e-01 2.75622994e-01 -4.92141306e-01
1.11655998e+00 -6.51070237e-01 1.47621274e-01 6.13134444e-01
1.30409598e+00 3.49136651e-01 -1.58242881e+00 -5.25844336e-01
2.56192416e-01 -6.64346933e-01 5.02152145e-02 -6.34421349e-01
-9.12633300e-01 6.55046344e-01 4.08038259e-01 3.46481979e-01
1.04181135e+00 -5.65534122e-02 1.06648892e-01 3.73275578e-01
1.82650238e-01 -1.28650200e+00 -1.92971289e-01 3.02408159e-01
9.59258676e-01 -1.31586373e+00 2.47618228e-01 -3.95426661e-01
-7.72995710e-01 8.99056196e-01 2.25432530e-01 -2.81802148e-01
1.06575692e+00 9.90831926e-02 -3.17557007e-01 1.41286543e-02
-8.82153213e-01 -2.09684178e-01 5.45956850e-01 4.27417308e-01
1.62950829e-01 7.83839524e-02 -2.44603828e-01 9.07203496e-01
-4.67284381e-01 -4.90209371e-01 2.39312634e-01 5.77662468e-01
-4.07942832e-01 -6.98606849e-01 -3.91016126e-01 7.10507870e-01
-5.86849570e-01 -2.92434573e-01 -1.22537807e-01 1.03316045e+00
-2.76342392e-01 1.08822858e+00 -7.08307251e-02 4.20391075e-02
3.22489977e-01 -4.62190360e-02 3.39193732e-01 -5.25725842e-01
-9.49023142e-02 2.90977120e-01 2.60018528e-01 -6.01006985e-01
-1.56835094e-01 -9.41310883e-01 -7.53857911e-01 -1.90408289e-01
-7.72307217e-01 3.30507010e-01 6.02321804e-01 1.25337982e+00
2.89474458e-01 9.30114239e-02 9.18210924e-01 -4.83652562e-01
-1.20438528e+00 -6.39250100e-01 -6.96606636e-01 -4.99646366e-02
8.39039743e-01 -7.20094562e-01 -5.80629587e-01 -1.08202510e-01] | [6.578854560852051, 4.119256973266602] |
8c567b30-931d-4aef-8b5b-2e1aecaee78d | prosodic-enhanced-siamese-convolutional | 1808.01026 | null | http://arxiv.org/abs/1808.01026v1 | http://arxiv.org/pdf/1808.01026v1.pdf | Prosodic-Enhanced Siamese Convolutional Neural Networks for Cross-Device Text-Independent Speaker Verification | In this paper a novel cross-device text-independent speaker verification
architecture is proposed. Majority of the state-of-the-art deep architectures
that are used for speaker verification tasks consider Mel-frequency cepstral
coefficients. In contrast, our proposed Siamese convolutional neural network
architecture uses Mel-frequency spectrogram coefficients to benefit from the
dependency of the adjacent spectro-temporal features. Moreover, although
spectro-temporal features have proved to be highly reliable in speaker
verification models, they only represent some aspects of short-term acoustic
level traits of the speaker's voice. However, the human voice consists of
several linguistic levels such as acoustic, lexicon, prosody, and phonetics,
that can be utilized in speaker verification models. To compensate for these
inherited shortcomings in spectro-temporal features, we propose to enhance the
proposed Siamese convolutional neural network architecture by deploying a
multilayer perceptron network to incorporate the prosodic, jitter, and shimmer
features. The proposed end-to-end verification architecture performs feature
extraction and verification simultaneously. This proposed architecture displays
significant improvement over classical signal processing approaches and deep
algorithms for forensic cross-device speaker verification. | ['Nasser M. Nasrabadi', 'Sobhan Soleymani', 'Hadi Kazemi', 'Jeremy Dawson', 'Ali Dabouei', 'Seyed Mehdi Iranmanesh'] | 2018-07-31 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [-4.58060093e-02 -2.79983133e-01 7.34375045e-02 -5.97095728e-01
-8.82327914e-01 -3.90072137e-01 2.47657925e-01 2.16527686e-01
-4.26854312e-01 4.25884485e-01 2.58647442e-01 -2.69172192e-01
-4.86464910e-02 -1.58972025e-01 -2.84636170e-01 -7.41874218e-01
-1.69330820e-01 -2.16741800e-01 -2.85596669e-01 -2.75963396e-01
3.36251378e-01 6.73981369e-01 -1.69650292e+00 8.57786536e-02
6.92322552e-01 1.32057452e+00 -3.04811567e-01 8.13151300e-01
5.76537326e-02 3.62106115e-01 -8.94887090e-01 -3.63002300e-01
-1.09888807e-01 -4.01993573e-01 -4.29066837e-01 -1.61660343e-01
3.35513473e-01 -3.83881092e-01 -1.64033428e-01 1.02312756e+00
9.63757277e-01 5.35955764e-02 4.91761535e-01 -1.01579702e+00
-7.54092574e-01 7.50426769e-01 -2.19626114e-01 3.47952217e-01
3.87804866e-01 1.68378770e-01 8.43932569e-01 -7.36427546e-01
1.19506866e-01 9.34181333e-01 9.54109609e-01 4.27140146e-01
-7.57916033e-01 -8.42398942e-01 -3.21830027e-02 4.88778740e-01
-1.32585704e+00 -9.34186161e-01 1.34491968e+00 -3.66687417e-01
1.04856265e+00 1.03063986e-01 2.33058080e-01 1.08819544e+00
2.78138220e-01 5.75502038e-01 8.32511365e-01 -4.11364973e-01
1.70402929e-01 2.60504574e-01 3.60474557e-01 5.33708990e-01
-4.20381814e-01 1.85536936e-01 -8.16148996e-01 -1.53627068e-01
1.57717437e-01 -4.14144278e-01 -3.98349941e-01 2.04443082e-01
-9.29218650e-01 6.29196167e-01 1.59936816e-01 7.80165613e-01
-4.68787700e-01 -1.93187594e-01 1.03459513e+00 1.94432035e-01
2.70931542e-01 5.48045039e-02 -3.60440105e-01 -4.06068712e-01
-1.54216242e+00 -1.25870481e-01 5.65981090e-01 4.23528373e-01
2.81810760e-01 7.97641575e-01 -1.07421607e-01 7.91931570e-01
5.46370804e-01 4.53664780e-01 1.01869464e+00 -5.95508933e-01
3.18786561e-01 2.23723054e-01 -1.04932733e-01 -8.84672582e-01
-4.00234491e-01 -5.78197122e-01 -9.28606451e-01 3.23821694e-01
2.29960591e-01 -1.62299380e-01 -6.56958997e-01 1.75837612e+00
1.47495076e-01 1.00837424e-01 2.65855610e-01 9.56903756e-01
9.03147042e-01 3.32871258e-01 -9.65531617e-02 -1.80987790e-01
1.33236694e+00 -6.01820767e-01 -1.15813982e+00 2.78953463e-01
-4.83367182e-02 -9.22737598e-01 9.40659344e-01 3.46176028e-01
-1.13138425e+00 -1.07557964e+00 -1.41903293e+00 -7.01233745e-02
-4.56529528e-01 4.57939625e-01 -2.63643302e-02 1.23445904e+00
-1.04680288e+00 6.60797060e-01 -7.05526292e-01 7.66001046e-02
4.41471040e-02 3.86198610e-01 -2.65214264e-01 6.12247407e-01
-1.51240754e+00 7.81992733e-01 1.85271814e-01 5.85041106e-01
-7.32237220e-01 -4.57607806e-01 -9.85285938e-01 4.29764986e-01
-3.49120021e-01 -1.82711825e-01 1.28683281e+00 -7.98675120e-01
-2.14714003e+00 4.35680211e-01 -4.78317261e-01 -5.04949868e-01
2.90390342e-01 1.78528637e-01 -1.00447118e+00 2.57659674e-01
-3.72008532e-01 2.28776291e-01 1.15930820e+00 -6.84650779e-01
-3.00845444e-01 -4.05371845e-01 -3.29544812e-01 -1.26703843e-01
-5.83574772e-01 4.24285829e-01 1.20082334e-01 -4.68069702e-01
4.00914326e-02 -6.12107098e-01 3.12313586e-01 -2.44837105e-01
-5.35383403e-01 -2.08111316e-01 1.09282613e+00 -1.17053556e+00
1.23756361e+00 -2.53585863e+00 -3.56662214e-01 3.08582168e-02
-1.26781836e-01 5.18196225e-01 8.73425826e-02 1.85948268e-01
-1.24303736e-01 -6.84008077e-02 -1.73213124e-01 -8.25999320e-01
3.79295886e-01 -3.50720733e-01 -1.79363757e-01 7.21847892e-01
3.12422961e-01 4.68048841e-01 -1.79699272e-01 -4.61026907e-01
4.47895437e-01 1.09799933e+00 -1.28847778e-01 -1.84324965e-01
4.19536233e-01 4.73672301e-01 -1.60513129e-02 6.41592741e-01
8.61131907e-01 3.93095523e-01 -1.09810472e-01 -2.10241735e-01
-1.34535402e-01 7.44670212e-01 -1.02774835e+00 1.68125021e+00
-5.50503552e-01 8.87834311e-01 5.48678994e-01 -7.73212790e-01
1.24335229e+00 9.15917516e-01 3.69235724e-01 -3.65282029e-01
5.15967548e-01 3.37080359e-01 3.70472550e-01 -4.58893090e-01
7.04992414e-01 -2.83376396e-01 7.58486986e-02 2.44250119e-01
1.97290093e-01 1.64537698e-01 -1.50732309e-01 -2.21665069e-01
5.84747970e-01 -1.67012036e-01 1.10312179e-01 -3.12924832e-02
1.11289907e+00 -6.42948210e-01 6.13577247e-01 1.56979367e-01
-8.61255586e-01 6.56117380e-01 8.94624591e-02 6.02236763e-03
-8.77057254e-01 -8.41832161e-01 -3.21277827e-01 8.32420588e-01
-4.25582796e-01 -3.90238427e-02 -9.49010372e-01 -3.60598356e-01
-1.16284080e-01 5.69258392e-01 -4.73165721e-01 -2.62216311e-02
-6.16905868e-01 -2.48286620e-01 1.39417815e+00 5.51561177e-01
5.14974177e-01 -1.11336863e+00 -4.42992032e-01 4.27201867e-01
4.53910604e-03 -1.02669883e+00 -7.97128081e-01 2.03741789e-01
-5.27332723e-01 -7.74946630e-01 -6.15711927e-01 -7.49305189e-01
-2.16413856e-01 -2.68590331e-01 5.24659872e-01 -3.54076177e-01
1.54896232e-03 1.49774551e-03 -2.30616033e-01 -3.54124039e-01
-5.81206203e-01 1.32097945e-01 4.55757886e-01 3.69685501e-01
5.70376456e-01 -8.10562730e-01 -4.52274382e-01 1.19766451e-01
-6.52402461e-01 -7.96729982e-01 2.23327443e-01 8.17594588e-01
2.58471340e-01 1.29156172e-01 1.13183022e+00 -1.10811681e-01
9.88732576e-01 -1.59950957e-01 -3.27215374e-01 1.24895222e-01
-3.35708737e-01 -6.83634579e-02 1.02923226e+00 -4.54414606e-01
-1.10465157e+00 -2.52673924e-02 -6.57048881e-01 -5.16702056e-01
-4.97861892e-01 7.00038016e-01 -3.26416939e-01 -2.92610703e-03
4.34990466e-01 4.64798242e-01 1.63014635e-01 -5.24311960e-01
1.47656454e-02 1.29703367e+00 7.41361797e-01 -1.00621507e-01
3.78822982e-01 1.37588188e-01 -3.91726017e-01 -1.02382874e+00
-1.97251141e-01 -5.95973372e-01 -7.24319935e-01 -1.42601877e-01
8.35805953e-01 -7.39629090e-01 -1.29022002e+00 7.13300467e-01
-1.36789477e+00 4.42495793e-01 -1.69816568e-01 7.15965688e-01
-3.72707725e-01 6.81019843e-01 -8.55596185e-01 -1.27739894e+00
-9.17486548e-01 -1.40650916e+00 9.92088258e-01 2.54117876e-01
-3.13669950e-01 -8.05144727e-01 -5.12077883e-02 3.68631572e-01
8.86717319e-01 7.91436881e-02 5.94574511e-01 -8.30026329e-01
9.09885913e-02 -5.07918715e-01 1.50187626e-01 8.05324614e-01
2.50931025e-01 2.31895074e-01 -1.53749061e+00 -3.10058296e-01
4.98958051e-01 -7.97040015e-02 6.01966798e-01 5.49422741e-01
8.98304224e-01 -1.76779568e-01 2.32039005e-01 3.86786133e-01
1.01117420e+00 4.52393651e-01 3.83718491e-01 -1.20557085e-01
2.99718350e-01 5.15225887e-01 7.39937723e-02 4.65027779e-01
2.57004112e-01 6.27027750e-01 1.58927634e-01 1.42422318e-01
-1.63002744e-01 3.78804617e-02 6.96637630e-01 1.32674694e+00
2.56094068e-01 -1.13310106e-02 -7.94339120e-01 6.78651333e-01
-1.24270427e+00 -1.04151881e+00 -6.09762818e-02 1.94645262e+00
5.61038554e-01 1.65636435e-01 1.34838611e-01 9.23584104e-01
9.42750514e-01 2.14712143e-01 -6.35583520e-01 -8.43362927e-01
-2.91443408e-01 4.08330113e-01 1.29020095e-01 4.66732234e-01
-1.05128467e+00 3.02044451e-01 5.97704172e+00 7.72786736e-01
-1.72192836e+00 2.55083174e-01 1.37531266e-01 -3.03805359e-02
7.47995637e-03 -5.96490979e-01 -7.09070146e-01 5.90828717e-01
1.46137416e+00 1.26900166e-01 1.59847721e-01 7.64996707e-01
5.65788090e-01 2.23231718e-01 -1.22403812e+00 1.03519523e+00
3.24911743e-01 -8.74647439e-01 -4.67358649e-01 -1.06780596e-01
2.73245573e-01 -1.38976695e-02 5.35773098e-01 2.24393681e-01
-4.63311106e-01 -1.04824197e+00 9.61530328e-01 2.73742914e-01
7.56865263e-01 -8.58012021e-01 9.80596960e-01 1.22707419e-01
-1.32700312e+00 -5.07082455e-02 7.84420893e-02 1.47949055e-01
2.67209977e-01 4.66765881e-01 -1.01098943e+00 6.34218395e-01
5.32119095e-01 3.59465927e-01 -3.39955181e-01 9.58133757e-01
6.80589452e-02 7.72039294e-01 -2.65725046e-01 3.89092267e-02
1.52369961e-01 2.43051365e-01 5.37170768e-01 1.39682817e+00
4.35325623e-01 -4.53105807e-01 -3.23940814e-01 8.20761442e-01
3.94683145e-02 1.52616307e-01 -3.20983022e-01 -1.58376873e-01
5.21279335e-01 9.80896950e-01 -2.59718120e-01 -2.30418935e-01
-1.60867646e-01 8.19368303e-01 -2.31001973e-01 2.82988042e-01
-8.63755226e-01 -9.46051300e-01 7.86598325e-01 -2.70011187e-01
3.40345949e-01 -2.85829097e-01 -4.62998509e-01 -1.03275084e+00
3.09852511e-01 -9.21603501e-01 3.08022574e-02 -3.79839301e-01
-1.29151464e+00 9.73527551e-01 -5.90169430e-01 -1.27414525e+00
-6.01464868e-01 -4.65088278e-01 -8.59981179e-01 1.38631368e+00
-1.81466484e+00 -1.16848719e+00 5.86104132e-02 7.25689828e-01
6.15596831e-01 -4.92990881e-01 9.53506708e-01 6.32333457e-01
-6.05869293e-01 1.01171947e+00 2.25873426e-01 4.24970001e-01
6.58377111e-01 -9.39315856e-01 6.63647830e-01 9.12175000e-01
-9.47280526e-02 7.16938734e-01 6.05925739e-01 -1.63548887e-01
-1.03660452e+00 -6.55168951e-01 1.16917050e+00 1.54977411e-01
5.16986847e-01 -1.93106264e-01 -9.87820923e-01 2.25198984e-01
4.19459105e-01 -5.93384542e-02 9.51141119e-01 8.17147270e-02
-5.15807211e-01 -2.43427843e-01 -1.35869944e+00 5.68197370e-02
2.11658642e-01 -1.29689646e+00 -6.35353267e-01 -1.74725652e-01
4.38070208e-01 -1.43680125e-01 -9.42499816e-01 1.33755311e-01
8.23832333e-01 -1.14362240e+00 5.94341636e-01 -1.68915570e-01
9.44115296e-02 -4.32206064e-01 -2.61936545e-01 -1.18383384e+00
-3.52945477e-02 -6.34231031e-01 4.42497097e-02 1.62230396e+00
5.02825737e-01 -5.89379907e-01 7.37966120e-01 4.39393669e-01
-4.23428893e-01 -3.24180871e-01 -1.44151652e+00 -8.21888149e-01
1.11983173e-01 -6.14160359e-01 7.25219727e-01 8.72256577e-01
3.19533288e-01 2.11389989e-01 -4.09289926e-01 3.18479180e-01
4.22122449e-01 -8.64196047e-02 1.99946821e-01 -1.10485816e+00
-2.21605361e-01 -7.43744671e-01 -5.58754325e-01 -6.56781673e-01
3.95994484e-01 -7.28060007e-01 1.36536837e-01 -9.39009964e-01
-2.62731642e-01 -5.81658967e-02 -7.54674137e-01 1.58217788e-01
1.00254379e-01 1.28060028e-01 2.25192487e-01 -1.70399278e-01
-4.72535603e-02 7.47774482e-01 7.03731120e-01 -4.24544692e-01
-2.83117086e-01 1.54498115e-01 -1.55163944e-01 5.19403696e-01
9.42749202e-01 -9.62561071e-02 -6.97387829e-02 -1.55887276e-01
-5.88258982e-01 3.71969044e-01 1.37228638e-01 -1.17607760e+00
5.27555287e-01 4.82711792e-01 3.27119917e-01 -8.36935461e-01
8.08402300e-01 -8.67794752e-01 -3.91849466e-02 4.04009759e-01
-4.03549105e-01 5.76549061e-02 3.42305064e-01 1.97554097e-01
-8.52171540e-01 -2.47201458e-01 8.79110813e-01 2.87486941e-01
-1.23242177e-01 -1.07929565e-01 -7.91634321e-01 -6.21434152e-01
5.32133818e-01 -4.88797665e-01 -2.38471791e-01 -3.80094767e-01
-7.78986216e-01 -3.75488013e-01 3.35681178e-02 3.96598309e-01
6.60183370e-01 -1.19151604e+00 -7.49379277e-01 4.32237357e-01
-2.13106692e-01 -5.80784261e-01 6.96819484e-01 9.54321682e-01
-2.51798868e-01 7.86826849e-01 -4.00639236e-01 -5.11625111e-01
-1.42829096e+00 4.69035566e-01 5.46483934e-01 3.06304581e-02
-8.30362141e-02 7.53911734e-01 -5.28110564e-01 -4.52383041e-01
5.84285259e-01 -5.88146329e-01 -1.97779536e-01 2.33188003e-01
4.96088296e-01 3.95014882e-01 5.68813980e-01 -1.04723001e+00
-6.67148829e-01 4.76162970e-01 4.99613658e-02 -3.15004259e-01
1.12677372e+00 -2.98145503e-01 1.46832481e-01 5.54531157e-01
1.46064103e+00 2.63837487e-01 -8.07194471e-01 -1.54006675e-01
-5.62124029e-02 5.61874919e-02 3.70439172e-01 -8.19844961e-01
-1.02177584e+00 1.35313284e+00 8.63969505e-01 3.31777334e-01
1.21492302e+00 -7.46743321e-01 1.08747232e+00 1.29203379e-01
6.42328407e-04 -1.13963342e+00 -3.46670032e-01 4.74846870e-01
7.55257845e-01 -1.15237939e+00 -4.43015695e-01 7.35904276e-02
-5.53801477e-01 1.45218432e+00 1.75310507e-01 1.29899338e-01
8.93581629e-01 3.24258238e-01 3.34231049e-01 2.02749580e-01
-3.12358052e-01 -5.01933061e-02 3.56348395e-01 6.08010292e-01
7.95306027e-01 -2.70329732e-02 -2.21674010e-01 7.61743486e-01
-6.90152645e-01 -1.42572075e-01 3.73589724e-01 5.83308458e-01
-1.83651134e-01 -1.01913917e+00 -6.71768963e-01 -6.60361126e-02
-7.96995997e-01 -8.48350525e-02 -3.22886258e-01 4.04180288e-01
2.53360689e-01 1.43208957e+00 -6.13891371e-02 -6.19301081e-01
9.10677388e-02 6.28731012e-01 1.34826347e-01 -2.88182721e-02
-1.13798523e+00 1.14198864e-01 -2.09331140e-02 -1.60456672e-01
-4.69439834e-01 -6.97093785e-01 -1.17650318e+00 -3.80952209e-01
-4.77699161e-01 1.98353961e-01 1.45667386e+00 9.43608880e-01
3.34960550e-01 6.89770877e-01 7.70440340e-01 -9.05687809e-01
-7.04255641e-01 -1.46274912e+00 -6.60555124e-01 1.20964609e-01
1.05926764e+00 -3.27125460e-01 -3.17051917e-01 8.60852674e-02] | [14.423074722290039, 6.039051532745361] |
455009ad-e691-4aad-a32e-d18c8a8b90b5 | predicate-argument-alignment-using-a-global | null | null | https://aclanthology.info/papers/N15-1002/n15-1002 | https://www.aclweb.org/anthology/N15-1002 | Predicate Argument Alignment using a Global Coherence Model | null | ['Mark Dredze', 'Benjamin Van Durme', 'Travis Wolfe'] | 2015-05-01 | null | null | null | hlt-2015-5 | ['cross-document-coreference-resolution'] | ['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] | [-1.5391852855682373, 15.869174003601074] |
5088f3d6-4957-404a-a250-595381771e4b | grf-learning-a-general-radiance-field-for-3d-1 | 2010.04595 | null | https://arxiv.org/abs/2010.04595v3 | https://arxiv.org/pdf/2010.04595v3.pdf | GRF: Learning a General Radiance Field for 3D Representation and Rendering | We present a simple yet powerful neural network that implicitly represents and renders 3D objects and scenes only from 2D observations. The network models 3D geometries as a general radiance field, which takes a set of 2D images with camera poses and intrinsics as input, constructs an internal representation for each point of the 3D space, and then renders the corresponding appearance and geometry of that point viewed from an arbitrary position. The key to our approach is to learn local features for each pixel in 2D images and to then project these features to 3D points, thus yielding general and rich point representations. We additionally integrate an attention mechanism to aggregate pixel features from multiple 2D views, such that visual occlusions are implicitly taken into account. Extensive experiments demonstrate that our method can generate high-quality and realistic novel views for novel objects, unseen categories and challenging real-world scenes. | ['Bo Yang', 'Alex Trevithick'] | 2020-10-09 | grf-learning-a-general-radiance-field-for-3d | http://openaccess.thecvf.com//content/ICCV2021/html/Trevithick_GRF_Learning_a_General_Radiance_Field_for_3D_Representation_and_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Trevithick_GRF_Learning_a_General_Radiance_Field_for_3D_Representation_and_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 2.71682531e-01 -3.70574966e-02 2.04628125e-01 -4.58805174e-01
-3.36982936e-01 -7.52298355e-01 7.27700412e-01 -4.34837252e-01
9.16314572e-02 2.60330409e-01 -2.25682883e-03 -1.29766598e-01
2.34768853e-01 -1.03477848e+00 -1.10939109e+00 -4.53440577e-01
2.43492082e-01 4.66754138e-01 6.46392331e-02 2.89239623e-02
2.69171208e-01 1.43387341e+00 -1.76005471e+00 1.03445053e-01
3.21906567e-01 8.93441796e-01 3.88200909e-01 9.77577150e-01
4.55630012e-02 3.94905508e-01 -4.08000082e-01 -1.92989141e-01
7.53656089e-01 1.71844885e-01 -5.31146824e-01 8.48650813e-01
1.11854899e+00 -8.44553292e-01 -6.29762650e-01 8.23630571e-01
2.25003168e-01 2.73582637e-01 8.01748216e-01 -8.61100256e-01
-1.04255962e+00 -3.86866242e-01 -8.01809907e-01 -2.12478228e-02
7.00906157e-01 3.20255309e-01 7.72239149e-01 -1.01638055e+00
7.48531818e-01 1.49139547e+00 5.06344438e-01 3.83926570e-01
-1.54726112e+00 -1.76618576e-01 3.96950662e-01 -2.90874779e-01
-1.27435613e+00 -3.22055817e-01 1.19953132e+00 -4.07664835e-01
1.20597100e+00 1.18125468e-01 8.08525503e-01 1.12940967e+00
1.76959664e-01 5.33123672e-01 8.54970515e-01 -3.73701423e-01
1.67073831e-01 -1.08262394e-02 -1.97605222e-01 6.93635821e-01
-1.40472561e-01 2.82998204e-01 -2.62328982e-01 -8.85703564e-02
1.43102360e+00 5.69287360e-01 -3.18696260e-01 -9.69708204e-01
-1.33524704e+00 5.18766105e-01 8.60618353e-01 -2.41437033e-01
-5.31648636e-01 2.27465451e-01 -4.06866699e-01 -8.84920545e-03
4.80501324e-01 3.59078914e-01 -4.08592314e-01 3.31112355e-01
-2.31644496e-01 4.33263361e-01 3.84260446e-01 1.23862028e+00
1.15931475e+00 8.23023394e-02 3.58024985e-01 5.71296751e-01
1.54068008e-01 9.55034077e-01 -1.51961704e-03 -1.64835346e+00
2.96341717e-01 6.18403554e-01 2.81448931e-01 -1.20523632e+00
-2.73575991e-01 -3.96122068e-01 -5.86714864e-01 7.04593599e-01
-1.01161294e-01 2.08158299e-01 -1.21204519e+00 1.64245677e+00
4.71206486e-01 1.66285768e-01 -9.17223170e-02 1.02386367e+00
8.06897223e-01 8.24676752e-01 -5.96467793e-01 2.11047873e-01
9.67665255e-01 -5.82212567e-01 -7.78371692e-02 -3.67671698e-01
8.63827989e-02 -5.68000436e-01 7.49708712e-01 2.90184915e-01
-1.52597558e+00 -8.42292190e-01 -8.77888560e-01 -5.20800650e-01
-4.07385677e-01 1.89692471e-02 5.76440275e-01 2.71183670e-01
-1.15111399e+00 4.43909794e-01 -6.87105477e-01 -3.31957012e-01
3.48378479e-01 2.06325829e-01 -6.02405727e-01 -2.44585186e-01
-4.86879021e-01 7.90955603e-01 6.18718602e-02 1.26423970e-01
-1.01356220e+00 -6.20546520e-01 -1.16401565e+00 -2.39570588e-02
1.84828475e-01 -1.29059589e+00 1.04830146e+00 -9.22639370e-01
-1.38383818e+00 1.22680783e+00 -3.10119748e-01 6.99028671e-02
2.41740927e-01 -1.49924457e-01 -1.28656983e-01 3.05015922e-01
1.12529002e-01 9.58177865e-01 8.58100295e-01 -1.86729479e+00
-3.86175960e-01 -6.91822946e-01 5.76014757e-01 7.01800287e-01
4.05938655e-01 -4.20558035e-01 -4.60452467e-01 -3.51040304e-01
7.03136504e-01 -8.09059083e-01 -4.35138047e-01 5.58740914e-01
-4.12914842e-01 2.13152468e-01 8.40324342e-01 -3.57171416e-01
4.35407870e-02 -2.22353292e+00 3.14747095e-01 1.33309305e-01
3.63795191e-01 -2.04973906e-01 -2.80541271e-01 7.07219690e-02
-1.91247851e-01 -1.46108782e-02 8.50422680e-02 -4.50779945e-01
-1.99189395e-01 3.30462337e-01 -4.21356201e-01 4.19479460e-01
4.69971299e-01 1.06052852e+00 -9.81273592e-01 1.43668652e-01
9.69537795e-01 8.31779540e-01 -7.32637227e-01 3.82050961e-01
-4.53902781e-01 5.01357257e-01 -5.90271831e-01 4.75785941e-01
1.04492342e+00 -4.10522193e-01 -3.40617031e-01 -3.05611074e-01
-1.48943722e-01 2.68169403e-01 -1.18854356e+00 1.99148846e+00
-5.50602436e-01 4.92304355e-01 -2.13848710e-01 -7.44379282e-01
1.07772386e+00 2.64355261e-02 3.67382467e-01 -5.04925430e-01
3.49385701e-02 -1.89890489e-01 -7.21156597e-01 -4.90068972e-01
4.51563418e-01 4.80263084e-02 7.28366002e-02 3.23329508e-01
5.07066697e-02 -7.58873045e-01 -4.34328228e-01 1.65962964e-01
7.28446722e-01 4.44179803e-01 2.76962578e-01 1.80604398e-01
3.48218471e-01 -1.85409486e-01 3.10185701e-01 6.52842462e-01
3.16896796e-01 1.26741433e+00 9.10646766e-02 -9.79025245e-01
-1.34238338e+00 -1.69399118e+00 -1.45344734e-01 3.81496996e-01
4.42360014e-01 6.40258417e-02 -2.51520038e-01 -5.17394066e-01
1.24145187e-01 6.11237943e-01 -7.26913810e-01 7.62031740e-03
-4.35067326e-01 -9.74512175e-02 -3.74724686e-01 5.05548775e-01
4.04361337e-01 -8.23407114e-01 -8.89434099e-01 -3.48588675e-02
-1.74982175e-01 -1.13686991e+00 -2.74182022e-01 2.27410048e-01
-8.92606497e-01 -9.20021892e-01 -6.04891002e-01 -7.49467909e-01
9.40689087e-01 1.00033760e+00 1.34190691e+00 -1.60464168e-01
-5.54141641e-01 6.43111467e-01 -5.20115867e-02 -2.46736884e-01
-1.79309353e-01 -4.60137695e-01 4.00698632e-02 1.81422144e-01
3.53922755e-01 -9.06338096e-01 -6.79118574e-01 2.27056518e-01
-7.05007911e-01 4.30268884e-01 3.72545302e-01 5.88094234e-01
6.98688865e-01 -2.08214432e-01 -1.30309209e-01 -5.61805487e-01
-1.86753005e-01 -1.66872799e-01 -7.34277129e-01 8.85511860e-02
3.48901808e-01 -1.32680029e-01 5.27239382e-01 -1.78201988e-01
-1.20243907e+00 3.18559140e-01 1.40104666e-01 -8.78027558e-01
-7.93482304e-01 -1.01949714e-01 -3.86653781e-01 -4.24800552e-02
8.55057716e-01 2.51350552e-01 -1.11292616e-01 -5.53137720e-01
7.58408368e-01 4.14607108e-01 7.15117395e-01 -5.81371725e-01
1.02390194e+00 1.05059659e+00 2.04207495e-01 -9.11104500e-01
-9.81436253e-01 -2.71778375e-01 -1.28811538e+00 -1.13041945e-01
1.00808275e+00 -1.20772231e+00 -6.67890549e-01 3.46990019e-01
-1.43586826e+00 -1.88777596e-01 -4.87038434e-01 3.82519245e-01
-8.05795729e-01 2.75282472e-01 -2.77236074e-01 -7.54728496e-01
3.01565737e-01 -1.08093607e+00 1.71173024e+00 2.55287111e-01
6.63694926e-03 -9.63601768e-01 -5.68143427e-02 2.89707929e-01
-4.33229730e-02 5.42204559e-01 9.42215204e-01 3.31385076e-01
-1.36629784e+00 -7.86529481e-02 -4.30737048e-01 3.52900267e-01
2.68436521e-01 1.57889888e-01 -1.28899038e+00 -1.22424729e-01
1.60200924e-01 -2.84977078e-01 5.24789870e-01 6.12651885e-01
1.42862558e+00 5.64013384e-02 -3.48409623e-01 1.12722242e+00
1.57727158e+00 1.13373809e-01 4.31842506e-01 -2.14563403e-02
8.97802949e-01 5.97071052e-01 6.24728687e-02 3.32302690e-01
4.05461729e-01 6.88354611e-01 9.53683496e-01 -3.50446731e-01
-4.96092066e-02 -3.28264862e-01 2.64821649e-02 3.94064069e-01
-2.07810715e-01 -1.92226812e-01 -5.37709355e-01 3.81357819e-01
-1.41311836e+00 -9.52871621e-01 7.50324577e-02 2.21718240e+00
1.06395185e-01 -7.89530873e-02 -2.80731916e-01 -1.72596440e-01
4.74661142e-01 2.86812216e-01 -9.02719975e-01 -2.99736142e-01
-3.41757268e-01 3.43589075e-02 3.40893507e-01 5.01238108e-01
-8.01455557e-01 7.64739394e-01 6.96846533e+00 -1.57636642e-01
-1.07646835e+00 -3.66949946e-01 5.86598277e-01 -2.61788130e-01
-7.90202439e-01 3.22605646e-03 -4.66122478e-01 -1.35863870e-01
1.11624300e-01 9.59960222e-02 6.09972715e-01 7.75631666e-01
4.80983481e-02 1.80421144e-01 -1.35945392e+00 1.26587903e+00
3.88293445e-01 -1.43454301e+00 5.04628718e-01 3.95239502e-01
1.04066324e+00 3.19940299e-01 3.18604171e-01 -2.43701681e-01
5.59644282e-01 -8.65900457e-01 7.09677100e-01 7.19505072e-01
7.33127356e-01 -5.70254266e-01 8.26699063e-02 4.26788926e-01
-8.09663236e-01 -5.56247048e-02 -7.15804815e-01 -2.78786510e-01
3.06843221e-01 4.32805955e-01 -6.11663878e-01 4.81601059e-01
9.18962479e-01 1.14363837e+00 -4.82879549e-01 7.25614905e-01
-2.12254003e-01 -3.25257927e-01 -4.98897135e-01 3.69242221e-01
2.81163365e-01 -3.41139644e-01 6.37658596e-01 5.92655838e-01
4.56363648e-01 2.36207381e-01 1.28202200e-01 1.33552396e+00
-1.08679503e-01 -4.63194519e-01 -1.25009787e+00 6.25302970e-01
1.89893529e-01 1.22416067e+00 -3.29833269e-01 -3.14407438e-01
-5.41240335e-01 1.24489725e+00 4.80436832e-01 9.50483859e-01
-5.50391138e-01 -9.69159380e-02 9.39208448e-01 2.17982650e-01
5.03308237e-01 -4.82013792e-01 -5.48597276e-02 -1.47046316e+00
2.49703318e-01 -3.11892271e-01 -9.16583464e-02 -1.85131395e+00
-1.29841030e+00 4.98464078e-01 9.26403518e-05 -1.49664104e+00
-2.64572561e-01 -1.04939771e+00 -5.63873649e-01 1.15558553e+00
-1.55030227e+00 -1.07444763e+00 -6.77337468e-01 6.01616621e-01
4.67458546e-01 1.44201294e-01 7.58125722e-01 -2.29427621e-01
6.24203160e-02 -6.90990090e-02 8.77529308e-02 -3.18062156e-02
2.00711802e-01 -1.27936208e+00 1.06570315e+00 6.19236827e-01
4.33117390e-01 6.55894935e-01 2.88747370e-01 -1.31161630e-01
-1.45528972e+00 -1.05191422e+00 6.00871146e-01 -1.14645445e+00
7.16070607e-02 -6.62159801e-01 -7.91976094e-01 9.61072445e-01
-3.59647833e-02 3.60176444e-01 1.92353725e-01 -1.70906812e-01
-5.98499238e-01 1.07653218e-03 -1.14769804e+00 8.39262128e-01
1.53042531e+00 -7.52947152e-01 -7.58612454e-01 4.52776104e-01
8.55297923e-01 -7.19722450e-01 -6.46663308e-01 2.30344445e-01
4.79784340e-01 -1.33328938e+00 1.60467017e+00 -6.08807325e-01
4.12660927e-01 -4.01095629e-01 -4.26826477e-01 -1.51485693e+00
-5.52001476e-01 -3.24609816e-01 -4.70641442e-02 4.78383452e-01
8.64862725e-02 -5.53844094e-01 8.18311036e-01 6.16020441e-01
-1.47462547e-01 -4.61780131e-01 -6.48742616e-01 -4.38167959e-01
4.28035930e-02 -4.17395383e-01 8.51571083e-01 6.62460268e-01
-7.89000094e-01 3.42279494e-01 -2.00514033e-01 6.50250912e-01
9.31863666e-01 5.53852320e-01 1.23030138e+00 -1.38086772e+00
-2.47642651e-01 -1.99900359e-01 -7.53745079e-01 -1.79597747e+00
2.29510128e-01 -7.12793887e-01 -7.88895413e-02 -1.50659573e+00
6.08794615e-02 -2.85289764e-01 1.66295677e-01 2.86890734e-02
5.55177853e-02 3.96297485e-01 1.78729802e-01 5.00558093e-02
-2.03286827e-01 6.45315170e-01 1.49203289e+00 -1.03217117e-01
-9.62238386e-02 -7.50002041e-02 -6.09972239e-01 8.69949400e-01
3.66209447e-01 5.38357608e-02 -5.45082629e-01 -1.05115092e+00
1.07117109e-01 2.30429284e-02 9.30936337e-01 -8.48112464e-01
-2.14683190e-01 -4.30598140e-01 1.21499360e+00 -1.08539867e+00
9.11794007e-01 -1.05661714e+00 3.08497310e-01 -1.79356202e-01
-1.81782544e-01 5.02106063e-02 -1.61559898e-02 7.13096619e-01
1.33671731e-01 6.12158291e-02 5.57253063e-01 -6.79381967e-01
-6.50792658e-01 6.22892857e-01 5.00215031e-02 -2.49157131e-01
9.69395578e-01 -6.97538078e-01 -1.77581996e-01 -3.85915786e-01
-7.82411993e-01 -3.75195779e-02 1.18473268e+00 6.38024628e-01
9.58581209e-01 -1.52408159e+00 -5.94249189e-01 8.97633493e-01
3.64937782e-01 7.24099040e-01 5.77034235e-01 -1.26585439e-01
-7.69118726e-01 3.32863033e-01 -3.60709548e-01 -1.28317416e+00
-8.94037306e-01 8.66561353e-01 6.40799046e-01 5.01511872e-01
-1.05015576e+00 6.72349632e-01 9.55073774e-01 -9.26126480e-01
5.21906316e-02 -5.32539010e-01 1.60588101e-01 -6.90536857e-01
4.42804515e-01 -1.10469237e-01 -9.89855453e-02 -9.64916885e-01
-9.99283865e-02 1.29360855e+00 3.60687673e-02 -1.29896283e-01
1.37085998e+00 -3.93407464e-01 7.26131350e-02 5.55251539e-01
1.59407008e+00 -1.21988744e-01 -1.88972819e+00 -4.73995507e-01
-7.69469023e-01 -1.10118520e+00 1.21778920e-01 -4.00236934e-01
-1.07359755e+00 1.10103166e+00 3.49990934e-01 -9.44253504e-02
8.95477295e-01 1.45570368e-01 4.26595360e-01 5.78990400e-01
5.42560101e-01 -5.07095456e-01 2.23512977e-01 4.31516379e-01
9.87098753e-01 -1.19089603e+00 -3.66804451e-02 -4.04192239e-01
-2.75282145e-01 1.10622692e+00 7.01780915e-01 -4.40389395e-01
6.76768541e-01 -1.36485830e-01 1.99289471e-01 -3.31614792e-01
-6.52642310e-01 6.75004125e-02 3.71707618e-01 1.04091167e+00
-2.48045428e-03 -6.78146258e-02 9.13948655e-01 -2.82878816e-01
-1.67993054e-01 -2.82733202e-01 5.95568776e-01 7.30636179e-01
-2.90272623e-01 -6.04305089e-01 -3.87275875e-01 8.22857544e-02
1.48269385e-01 1.81023777e-01 -3.13058048e-01 5.50181329e-01
1.08206630e-01 3.93692791e-01 6.23210251e-01 -2.14898869e-01
5.06518841e-01 -1.71098739e-01 8.14127147e-01 -8.97730172e-01
2.04413068e-02 4.24957797e-02 -3.89449120e-01 -7.72450566e-01
-4.77648586e-01 -6.52905345e-01 -8.57353210e-01 -1.05947487e-01
-6.26533777e-02 -3.77854437e-01 8.47824931e-01 4.94112670e-01
6.11531556e-01 3.12105685e-01 1.16547799e+00 -1.59525990e+00
-3.18276256e-01 -3.70044291e-01 -5.85159659e-01 7.27859855e-01
8.59996319e-01 -7.60240734e-01 -6.83086038e-01 2.29865044e-01] | [8.839462280273438, -2.9981157779693604] |
47cc835b-f5a2-4213-ba4c-c9f5737b6226 | modelling-multi-issue-bargaining-dialogues | null | null | https://aclanthology.org/L16-1500 | https://aclanthology.org/L16-1500.pdf | Modelling Multi-issue Bargaining Dialogues: Data Collection, Annotation Design and Corpus | The paper describes experimental dialogue data collection activities, as well semantically annotated corpus creation undertaken within EU-funded METALOGUE project(www.metalogue.eu). The project aims to develop a dialogue system with flexible dialogue management to enable system{'}s adaptive, reactive, interactive and proactive dialogue behavior in setting goals, choosing appropriate strategies and monitoring numerous parallel interpretation and management processes. To achieve these goals negotiation (or more precisely multi-issue bargaining) scenario has been considered as the specific setting and application domain. The dialogue corpus forms the basis for the design of task and interaction models of participants negotiation behavior, and subsequently for dialogue system development which would be capable to replace one of the negotiators. The METALOGUE corpus will be released to the community for research purposes. | ['Volha Petukhova', 'Harmen de Weerd', 'Fokie Cnossen', 'Niels Taatgen', 'Andrei Malchanau', 'Christopher Stevens'] | 2016-05-01 | modelling-multi-issue-bargaining-dialogues-1 | https://aclanthology.org/L16-1500 | https://aclanthology.org/L16-1500.pdf | lrec-2016-5 | ['dialogue-management'] | ['natural-language-processing'] | [ 0.12178647 0.7899332 0.23188695 -0.6167947 -0.3578375 -0.8684339
1.2352629 0.26653305 -0.5558646 1.0011463 0.7017243 -0.16023932
-0.3153593 -0.7497026 0.47146255 -0.21530883 0.14333962 1.4049481
0.2415939 -0.95996463 0.62572336 0.42056257 -1.1677452 0.60586566
0.5127109 0.29336557 0.4344511 0.94213116 -0.43189946 0.7504068
-0.7597865 -0.3827892 0.08674663 -0.68319666 -1.5813937 0.25234544
-0.30495727 -0.33927986 0.216117 0.7431493 0.62382585 0.4114819
0.5952743 -1.1307739 0.10838504 1.0875849 0.27923858 0.1304956
0.9506188 0.05619524 0.89194906 -0.27806884 1.0762655 1.5606555
0.3301681 0.7371836 -1.1341449 -0.28895625 -0.34262434 -0.04749024
-0.8799249 -0.7013846 0.43467736 -0.34953648 1.0798544 0.3749525
0.51217866 1.1476818 -0.07749114 0.60510474 1.2814308 -0.9170627
0.12711063 0.6227395 0.37420878 0.24314082 -0.32355815 -0.1938585
-0.3105974 -0.5904244 0.86361784 -1.1138365 -0.08514795 -0.13141277
-1.0849572 0.9798623 -0.384182 0.90945804 -0.58407754 -0.4555738
1.0833625 0.71986735 0.0722596 0.6891076 -0.5535008 -0.96734494
-0.08980471 0.8988422 1.6655865 0.7363277 0.18176806 -0.33801708
-0.24830888 1.2372912 0.6209314 -0.19720715 0.30625987 -1.1828576
0.53204405 0.80778617 0.47969633 -0.5658524 -0.85177195 0.6007027
-0.2535813 0.06593885 0.8198653 -0.45544153 -0.16635397 1.4867038
0.54454947 -0.7137052 0.62363935 0.72661835 0.8105458 0.5470931
0.35446808 -0.5542409 1.7069845 -0.46399814 -1.016947 0.32151905
0.9157041 -1.0314415 0.8418179 0.39847642 -1.3198788 -0.32798344
-0.5428015 0.12058848 -0.16965728 -0.30003348 0.67624617 0.7127713
-1.0348804 0.22524576 -0.23206994 -0.8158904 -0.46002752 0.46909946
-0.34006467 0.4432426 -1.5931734 1.3611219 0.8487545 0.0087267
-0.14775555 0.13416426 -0.8434537 -0.28374946 0.24545592 -0.543938
1.8119304 -1.0019329 -1.9011679 1.1338351 0.47592568 -0.47240517
0.7481945 0.18907066 -0.16112262 -0.1546576 0.15006049 0.5254741
-0.11621353 -1.1781318 -0.8312244 -0.32645738 0.4145169 0.9164315
0.31918103 0.69938177 -0.07853099 -0.17129058 -0.13906723 -0.7175422
-0.27078 -0.84501475 -0.16638072 -0.6116672 0.3263637 -0.6135241
1.1287118 -1.613932 0.16434647 0.21710345 -0.12137327 0.39905027
0.10360455 1.1658095 0.2512766 -0.16255052 0.11502671 0.08035277
0.46770236 0.35344258 0.13975675 -0.01643544 -0.38275033 0.34789145
-0.76523405 -0.5883595 0.38242486 -0.13007446 -0.07287045 0.483066
-0.5189642 0.60889375 -0.728139 0.12961417 0.16732061 0.49903628
0.7122675 0.24508753 -0.51997983 0.5696815 -1.1048167 1.5954671
-0.58928925 0.31535015 0.7044763 -0.7038067 1.1518883 0.9480104
0.43799108 -0.73313713 0.44247127 0.23217668 0.50035197 -0.8112652
0.8819261 -0.06693145 -0.479185 0.79183745 -0.02137107 -0.16609502
0.29634786 0.07797632 0.69290227 0.14495657 0.7423082 -0.31048176
0.8249023 0.46888554 0.37862888 0.48593304 -0.4541175 -0.34658337
0.61213964 -0.65481174 -1.3299348 -0.26158682 -0.29162085 1.2106198
0.04179465 -0.26856738 -1.0275027 -0.54648757 -0.53486264 1.0503427
-0.09162395 0.5192687 -0.81960297 -0.42337435 0.76049805 -0.25916734
0.76212144 -1.6193238 -0.6925421 0.9145525 -0.4576378 -0.8800564
0.05011537 0.06619107 -0.2357917 -1.2509819 -0.06080425 -0.7255535
-0.01167082 -0.36499965 0.9911445 0.08112874 -0.11840922 0.60196203
-0.61053526 -0.48903367 -1.2915647 0.18440203 -0.21484394 -0.5804116
0.3843067 -0.26090443 -0.30639267 0.6675579 -0.80649376 0.36478698
-0.04856075 0.9138473 -0.3739469 -0.28913766 0.8855816 -1.1769383
1.6659936 -0.32191113 -0.44148877 0.33472523 -0.47935432 -0.2603767
0.2218181 0.00608702 -1.655465 -0.24478231 -0.59766746 0.812722
-0.73617613 0.23583105 -0.49188232 -0.01600375 0.7443597 -0.14545417
0.14761502 -0.3192931 0.26826254 1.1639242 0.25509104 -1.1195056
0.08743458 -0.40327725 -0.43463975 -0.9971859 0.07586782 -0.75631285
-0.932927 -0.52517015 0.77380157 -0.5020087 -0.96002823 0.26706022
-1.3680427 -0.65925205 0.05705588 0.12899391 -0.8448818 0.31065974
-0.8059552 -1.2518903 -0.4601199 -0.90787786 0.51091754 0.374084
-1.0631946 -1.2991081 0.3532005 0.76699084 0.26537862 0.20930399
0.78490585 -1.199423 0.04698212 0.12297729 0.02268453 -0.03331744
-0.07399389 -0.11593081 -0.6569322 0.14687593 0.08462071 -0.7217725
-0.18712321 -0.11336803 0.02715473 -0.38145578 -0.02896264 -0.38327235
0.9199744 0.77307516 0.65337944 0.6995357 -0.14414693 1.2411542
1.18481 0.6206973 0.6444897 1.2772598 -0.12290421 0.29183
0.29511142 0.23824781 0.34505355 0.48392713 -0.2160857 -0.28374603
-1.0741808 0.4631563 -2.122187 -1.0435095 -0.48568657 1.6770895
0.8631077 0.08473535 0.34502697 0.02467022 0.7145784 0.01872314
0.27426615 -1.2900977 0.29350427 0.13739142 0.02746907 0.8301111
-0.75657743 0.9990022 5.411288 0.3774492 -0.3395045 0.12435997
0.38316518 0.64829814 -0.0299509 0.04760181 -0.68994635 0.11897791
1.2006229 -0.5485864 0.44116148 0.3360637 0.79315794 -0.28877056
-0.81115663 0.34318653 -0.35371277 -1.1735451 -0.36728296 0.2087542
0.10563893 -0.40276337 -0.89029354 0.49357146 0.5803941 -0.66132647
0.40879387 0.33047947 0.39505935 -0.70653015 0.99015045 0.6679076
-0.87539953 0.05508719 0.04535426 -0.1880822 0.5640043 -0.47990385
-1.4830709 0.55765367 0.01336907 -0.37962806 0.2165014 0.658541
0.07955848 0.43721405 -0.20432036 -0.23861018 0.42911437 -0.5358623
0.85731775 1.4355215 -0.25729945 0.6635288 0.7354391 0.42462534
0.25273445 0.6722772 -0.52657974 0.1696922 0.77523744 1.1379557
-0.7370398 -0.12403056 -0.06509633 0.8769094 -0.04044857 -0.13583246
-0.23058546 -0.25900766 0.5592489 0.08574051 -0.49632013 -0.07388084
-0.24221826 -0.5792146 -0.4268974 -1.1370133 0.51928216 -0.3536276
-0.9109727 0.5305816 0.40417442 -0.36839622 -0.99227095 -0.4724474
-0.7980685 1.0478088 -0.821751 -1.2643698 -0.07904986 0.396326
0.9612595 -0.5066601 1.4634697 -0.03711263 -0.34346557 0.06201021
-0.36979294 0.10523658 0.64625293 -1.4168893 0.05111179 -0.09184935
-0.36409885 0.39969987 0.8602477 -0.6183646 -0.9716152 -0.25373116
1.2319931 -0.03418208 0.8063551 -0.08429481 -0.7439665 0.33267626
0.7286973 -1.0699594 0.8592251 0.12739843 0.46200687 0.47107977
-1.4280312 0.4625784 0.68650335 -0.2870253 -1.0798912 0.31560498
0.1438355 -0.7560856 -1.203029 0.09419198 0.44383553 -1.041509
0.5840051 -0.42481738 -0.00958791 0.13675308 0.08584259 -1.014228
0.09393436 -1.0849724 0.7521258 1.5575274 0.45613524 -0.57549053
0.5370449 1.28157 -0.06289565 -0.1724579 -1.0254695 0.08445455
-0.0070294 -0.33304432 0.4590515 1.0348228 0.70333457 0.60027665
-0.16177945 -0.17205007 0.1271052 -0.20559287 1.1091907 -1.2151246
-0.2127898 -0.66852015 -0.28648546 -0.5795181 0.11330764 -0.46501014
0.12506728 -1.6181529 -0.414144 -0.7638934 0.500538 0.23695996
0.19285047 -0.358336 0.26769575 0.2566149 -0.3461303 0.309768
1.1400944 0.36984304 -0.67182696 0.42739183 -0.5427186 0.5854364
1.0606699 0.04985891 -0.43282995 0.40052015 -0.0558458 0.77511203
-0.16443408 -0.30811924 0.2524102 -0.501892 -0.13091192 -0.17609558
0.20239457 -0.7023894 0.42182028 0.39901513 -0.75568455 0.03889739
0.14861724 0.02802039 -0.21149382 -0.8727174 0.59082025 -0.48595423
-0.8250823 -0.3058421 -0.95719486 -0.07758414 1.3950588 -0.35898688
-0.15758768 -0.48896736 -1.0607405 0.5959761 0.5028431 0.27696818
0.17407301 -0.85390306 -0.9229324 -0.29853362 0.00934262 -0.03156882
0.09561626 0.24530286 -0.8657916 0.4165974 -0.68287903 -0.08129684
-1.7668743 -0.40599853 0.51317 -0.75677645 -0.65241754 0.40951326
-0.37816066 -0.8489117 0.27208588 0.29347804 -1.0232763 0.3214209
0.41603968 0.23290205 -0.1858814 -0.8033342 0.07961359 -0.2854601
0.01055619 -0.63736814 1.2417227 -0.45000923 -0.34880775 0.56960994
0.38763747 -0.11926337 -0.80469793 -0.14824805 0.65950406 -0.20733008
-0.40286246 -1.2111664 0.0786263 0.14922449 0.27019048 0.96735364
0.62581944 -0.22599319 0.36906543 0.3952438 0.58513826 -1.4322112
-0.3953801 0.8586999 1.2495617 -1.0045111 -0.14683251 -0.50744593
-1.0908358 1.5665559 0.792871 0.39887285 0.1714151 0.3274103
0.39132884 -0.49991512 -1.0712105 -0.10604821 -0.168568 0.67272
0.7847195 0.22645839 -1.1325737 0.5496256 -0.32468662 -0.01659643
0.6727048 0.9350557 -0.49123088 -1.9197624 -0.56948215 0.33680585
-0.24906649 0.31755838 -1.0751481 1.2723157 0.1282248 1.0615447
-0.18564957 0.0820011 0.841078 0.36530644 0.3479442 -0.7404813
-1.592325 0.10621879 1.4034415 -0.1900705 -0.59170926 -0.8849458
-1.3677113 -0.415601 -0.3586492 0.71046156 0.83824146 1.0054146
-0.20672704 0.08268078 0.54417 -0.7366422 -0.6266597 -1.4989095
-0.45491892 0.42063943 -0.49819154 -0.43560877 0.24972245 -0.07954247] | [12.944530487060547, 7.963223457336426] |
1fedcb59-a066-4060-afe7-b382da6b30cb | director-field-model-of-the-primary-visual | 1310.1341 | null | http://arxiv.org/abs/1310.1341v2 | http://arxiv.org/pdf/1310.1341v2.pdf | Director Field Model of the Primary Visual Cortex for Contour Detection | We aim to build the simplest possible model capable of detecting long, noisy
contours in a cluttered visual scene. For this, we model the neural dynamics in
the primate primary visual cortex in terms of a continuous director field that
describes the average rate and the average orientational preference of active
neurons at a particular point in the cortex. We then use a linear-nonlinear
dynamical model with long range connectivity patterns to enforce long-range
statistical context present in the analyzed images. The resulting model has
substantially fewer degrees of freedom than traditional models, and yet it can
distinguish large contiguous objects from the background clutter by suppressing
the clutter and by filling-in occluded elements of object contours. This
results in high-precision, high-recall detection of large objects in cluttered
scenes. Parenthetically, our model has a direct correspondence with the Landau
- de Gennes theory of nematic liquid crystal in two dimensions. | ['Ilya Nemenman', 'Rebecca Butterfield', 'Vijay Singh', 'Martin Tchernookov'] | 2013-10-04 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 2.18063563e-01 -2.90146470e-01 1.18062392e-01 -2.19636276e-01
1.67314738e-01 -6.94250584e-01 8.44762921e-01 -3.43055695e-01
-7.90359318e-01 6.21107757e-01 4.28062379e-02 1.71868727e-01
-1.64400443e-01 -5.35023332e-01 -3.98955166e-01 -1.17584920e+00
-3.42129081e-01 4.50495422e-01 6.51075065e-01 1.05432644e-01
4.95611787e-01 8.05804908e-01 -1.70194173e+00 4.02629405e-01
5.10730863e-01 9.06237006e-01 4.02816534e-01 8.79367113e-01
-8.83408710e-02 4.34380978e-01 -5.42512357e-01 -8.41255486e-02
1.84628308e-01 -2.55045176e-01 -3.14722240e-01 1.64019257e-01
8.36502314e-01 -7.73172975e-02 -2.32749119e-01 1.39662421e+00
2.93440312e-01 -3.12179029e-01 1.16985786e+00 -6.24788105e-01
-6.72930121e-01 -5.11974003e-03 -7.22360909e-01 7.43463159e-01
-1.74248278e-01 8.53651091e-02 7.14899242e-01 -1.01168752e+00
1.05992985e+00 1.36747026e+00 2.04966798e-01 7.27386117e-01
-1.58510649e+00 -3.92123550e-01 5.00165880e-01 -3.43289286e-01
-1.31877220e+00 -4.19778138e-01 5.05132675e-01 -8.65877271e-01
8.83878231e-01 3.96227926e-01 1.12084258e+00 9.25867736e-01
9.86830592e-01 5.42687118e-01 1.35986459e+00 -2.44791955e-01
4.72820252e-01 -4.09513950e-01 6.04060769e-01 6.56071603e-01
7.61554360e-01 3.94993693e-01 -5.54777563e-01 -4.10281360e-01
1.25744665e+00 -1.72586828e-01 -1.93075150e-01 -5.31500041e-01
-1.15468109e+00 4.99541998e-01 4.42775071e-01 5.25472462e-01
-4.12393063e-01 8.93412754e-02 -3.81780863e-02 1.58514410e-01
4.47591841e-01 3.39066178e-01 -5.99806979e-02 5.35903335e-01
-9.84481215e-01 4.99049395e-01 8.78058136e-01 6.39230072e-01
1.97133228e-01 2.69345175e-02 -1.14284471e-01 5.82134008e-01
3.23675513e-01 8.72841835e-01 2.37393633e-01 -9.54516053e-01
-4.04278576e-01 8.75602737e-02 -9.48015526e-02 -8.96136701e-01
-4.13412482e-01 -5.89579165e-01 -1.21302927e+00 8.88773143e-01
6.33385599e-01 1.76827341e-01 -1.32133925e+00 1.72919464e+00
-9.04595666e-03 -2.19001144e-01 -1.46730915e-01 9.24991667e-01
4.22493458e-01 4.86021399e-01 4.84885015e-02 -6.86679900e-01
1.41105151e+00 -2.46235952e-01 -8.89943421e-01 -3.63643110e-01
-2.96213895e-01 -5.35293937e-01 5.57430029e-01 5.28038263e-01
-8.83458793e-01 -2.28985235e-01 -9.66789484e-01 2.33496144e-01
-5.37143230e-01 -2.98199743e-01 6.87249064e-01 3.01533073e-01
-1.12923169e+00 1.65932730e-01 -9.06268299e-01 -3.95778775e-01
4.07192469e-01 2.62281626e-01 -1.55575261e-01 1.29621997e-01
-4.57527190e-01 9.12075639e-01 1.55718938e-01 2.93772906e-01
-6.45223558e-01 -2.65540093e-01 -4.80620742e-01 -1.86820388e-01
1.35547653e-01 -6.14713311e-01 8.85117829e-01 -1.00925899e+00
-7.80008912e-01 1.22966707e+00 -5.03418326e-01 -1.81907356e-01
5.32331586e-01 -1.14032470e-01 -1.25259310e-01 2.12243691e-01
-3.39989625e-02 8.43958080e-01 9.01469886e-01 -1.59898651e+00
-9.07507837e-02 -6.31135225e-01 -5.54385841e-01 9.00599882e-02
3.33256304e-01 1.38683125e-01 -2.72698641e-01 -8.53075385e-01
6.83110595e-01 -7.82590389e-01 -4.76930827e-01 3.11221540e-01
-3.44548494e-01 9.68310535e-02 6.58871830e-01 -1.76759586e-01
8.79690826e-01 -2.17000651e+00 5.04400656e-02 4.80607688e-01
7.95654237e-01 2.03396112e-01 -2.44604707e-01 -9.76109356e-02
7.06579089e-02 1.02094956e-01 -3.32735240e-01 2.31405437e-01
-5.65989852e-01 2.51114726e-01 -2.72044182e-01 8.04367065e-01
4.41936016e-01 9.31596279e-01 -8.14882219e-01 -4.79090452e-01
-7.09736422e-02 4.72314358e-01 -5.28674126e-01 -1.14288896e-01
-1.75099239e-01 3.95082593e-01 -4.47900355e-01 6.88477218e-01
8.98339808e-01 -4.37826246e-01 1.30025819e-01 1.03415981e-01
-4.06369716e-01 -3.09197664e-01 -1.10311937e+00 1.04307258e+00
4.43902880e-01 9.76107240e-01 4.65110242e-01 -5.45404315e-01
7.21558988e-01 -5.70454262e-02 2.76623189e-01 -6.81675613e-01
1.25440806e-01 2.24661261e-01 5.29579520e-01 -2.85736769e-01
1.48549438e-01 6.86595142e-02 3.04618090e-01 2.29663357e-01
1.09066807e-01 -9.33944434e-02 2.27942616e-01 1.37035564e-01
9.46081519e-01 -3.92536283e-01 5.22851348e-01 -1.02913451e+00
6.37976220e-03 -9.83030945e-02 3.72112244e-01 1.26533484e+00
-2.78079957e-01 5.75970709e-01 4.45262164e-01 -6.60408258e-01
-1.00197458e+00 -1.48022330e+00 -6.54039681e-01 5.93435884e-01
1.67558223e-01 1.30924463e-01 -7.52473772e-01 1.83912456e-01
1.99414268e-02 2.66932026e-02 -8.28973413e-01 8.96847248e-02
-6.27187312e-01 -1.15377104e+00 2.46960029e-01 -5.51801138e-02
4.10552144e-01 -1.40160751e+00 -1.09804714e+00 1.05437651e-01
2.00350896e-01 -9.72865582e-01 -2.90785670e-01 7.03046978e-01
-8.71620119e-01 -1.02540612e+00 -9.62201893e-01 -9.31876421e-01
9.92852986e-01 1.32002249e-01 1.12462842e+00 5.09706214e-02
-7.78940260e-01 1.07076235e-01 1.52269885e-01 -3.12720388e-01
-1.42960310e-01 -5.12612641e-01 2.31551960e-01 -2.03495875e-01
7.07166195e-01 -4.12201822e-01 -4.94427502e-01 1.62436381e-01
-8.57677162e-01 -1.25312403e-01 6.07165098e-01 1.01991689e+00
7.59353220e-01 -2.45399624e-01 1.20946370e-01 -7.22668529e-01
6.57671988e-01 -1.09094247e-01 -9.96813357e-01 1.24566503e-01
-3.79490077e-01 1.39813676e-01 8.65001231e-02 -8.26666713e-01
-7.21578419e-01 6.25399500e-02 2.61194378e-01 -1.38345689e-01
-3.78568351e-01 1.78205013e-01 3.17414135e-01 -4.28516537e-01
7.34156072e-01 3.89864773e-01 -1.55186966e-01 -2.81551838e-01
1.27458259e-01 2.67684102e-01 7.93470860e-01 -4.24601436e-01
5.11326194e-01 7.44952440e-01 8.14678520e-02 -1.45852697e+00
-4.32582587e-01 -3.49155992e-01 -9.62094069e-01 -3.51725578e-01
9.24084485e-01 -5.66674769e-01 -7.99158275e-01 9.56552923e-01
-1.48418212e+00 -2.81969100e-01 -3.12651932e-01 4.74926025e-01
-5.45020580e-01 4.94961031e-02 -7.45212376e-01 -1.38005507e+00
6.31981269e-02 -7.90070772e-01 8.45517337e-01 2.85353512e-01
-4.88646403e-02 -5.63812971e-01 1.55413419e-01 -5.79719901e-01
4.63874936e-01 3.03075999e-01 1.02647007e+00 -4.08166498e-01
-1.02264571e+00 1.01477534e-01 -4.09841955e-01 6.56786114e-02
-1.79320827e-01 2.34965980e-01 -1.22746062e+00 -1.26822546e-01
2.84714788e-01 2.29783580e-02 1.43716919e+00 1.27868044e+00
6.58843458e-01 -1.69144526e-01 -4.88954037e-01 6.82151020e-01
1.46849978e+00 3.90363067e-01 4.84043270e-01 -1.10788733e-01
2.55654216e-01 8.50845695e-01 1.69474453e-01 2.89034843e-03
-4.78917003e-01 1.97650462e-01 2.32824504e-01 -1.60892293e-01
-1.49853855e-01 2.95664907e-01 -1.26027793e-01 6.79933608e-01
-2.99528241e-01 -2.49287561e-01 -9.40621912e-01 4.98385131e-01
-1.62294316e+00 -1.16989744e+00 -1.26546264e-01 1.99924755e+00
5.62201023e-01 4.86837178e-01 -2.23742321e-01 -3.36465985e-01
6.95926070e-01 2.41319984e-01 -7.29315639e-01 -4.86566648e-02
-6.86768055e-01 -1.36859221e-02 6.12591386e-01 7.24570274e-01
-1.00640237e+00 9.43152845e-01 8.73133755e+00 5.53987265e-01
-1.03303623e+00 -1.67249948e-01 7.50672162e-01 -9.95965898e-02
-5.38146049e-02 -3.14459264e-01 -9.23489749e-01 2.06648201e-01
5.50921381e-01 1.06540047e-01 4.04952824e-01 3.66267622e-01
1.34802163e-01 -7.00003028e-01 -7.82643378e-01 9.01684523e-01
3.99400257e-02 -1.45369112e+00 9.18754041e-02 5.03459752e-01
5.52778304e-01 2.58784562e-01 9.50744599e-02 -4.49968964e-01
2.93101668e-01 -1.03793526e+00 7.58639276e-01 8.62322330e-01
9.23813879e-01 -1.73097193e-01 3.47742796e-01 5.85725546e-01
-9.66954350e-01 8.92413855e-02 -5.58319747e-01 -6.94585219e-02
-1.17555805e-01 7.27471650e-01 -2.19950140e-01 -5.30802071e-01
6.92982793e-01 2.23280936e-01 -4.49975789e-01 1.33591115e+00
3.33584100e-01 4.34143282e-02 -4.46529537e-01 -3.34217697e-01
3.39646637e-02 -4.13708419e-01 9.34701145e-01 1.40217388e+00
-9.26562473e-02 4.84518319e-01 2.71901526e-02 1.12035370e+00
3.60488713e-01 -1.37483150e-01 -1.05495560e+00 2.53457632e-02
7.87012428e-02 9.92477834e-01 -1.32718801e+00 -2.92511731e-01
-1.85426161e-01 8.59488130e-01 2.66506314e-01 7.05984533e-01
-2.99483687e-02 -3.79110128e-02 6.76787317e-01 1.95986480e-01
3.17338228e-01 -5.09398222e-01 -2.92380810e-01 -1.33826947e+00
-6.00036196e-02 -5.38239360e-01 -2.86585763e-02 -7.96282887e-01
-1.52399075e+00 8.50993097e-01 1.66442990e-02 -6.70050263e-01
-5.62410131e-02 -1.17388463e+00 -3.99666727e-01 1.03572762e+00
-1.11614525e+00 -8.04404259e-01 -9.10224952e-03 6.53007388e-01
3.55275303e-01 -3.75602767e-02 8.30336452e-01 -2.06727862e-01
5.82456291e-02 -1.58657521e-01 1.84197098e-01 3.22082132e-01
3.99900496e-01 -1.31596351e+00 3.65883261e-01 9.81060445e-01
1.55741930e-01 1.00387108e+00 9.28810358e-01 -1.08292317e+00
-8.31748188e-01 -5.72987914e-01 8.39915574e-01 -2.30056167e-01
4.54557210e-01 -8.36125731e-01 -1.07311726e+00 2.47712672e-01
9.61663648e-02 3.98452252e-01 3.87324631e-01 -1.09015979e-01
-3.91054004e-01 3.02196234e-01 -9.99308407e-01 5.84687948e-01
1.09960151e+00 -3.95296216e-01 -5.97051561e-01 2.14982197e-01
2.03690931e-01 -1.57636955e-01 -2.01415300e-01 2.59357870e-01
1.01732779e+00 -1.05223989e+00 8.21633339e-01 -6.82427704e-01
-1.54396877e-01 -1.60044074e-01 -1.93085849e-01 -9.46737409e-01
-7.66803741e-01 -3.49193543e-01 1.48939982e-01 5.19716978e-01
1.67563483e-01 -5.74312806e-01 7.29470074e-01 1.66642010e-01
2.54674256e-01 -4.95900482e-01 -1.14611030e+00 -6.61143839e-01
1.84772134e-01 8.96255150e-02 -2.92425811e-01 6.52667284e-01
-2.92929024e-01 1.20012671e-01 3.89300436e-02 8.66659805e-02
1.04519749e+00 8.54946449e-02 6.27571866e-02 -1.79398036e+00
1.03587285e-01 -6.19937181e-01 -6.76969349e-01 -1.14361358e+00
-5.25736585e-02 -4.78704959e-01 2.93053061e-01 -1.05118704e+00
5.40666759e-01 -2.97275513e-01 -2.32443795e-01 -1.12756774e-01
2.27763817e-01 2.66765445e-01 -6.40792213e-03 5.70471942e-01
-3.61930937e-01 -2.94513386e-02 1.58057499e+00 5.30468859e-02
-7.35713169e-02 -8.44679698e-02 -4.38560963e-01 1.05632234e+00
3.73936653e-01 -6.52228177e-01 1.11295983e-01 -9.77847278e-02
9.86507684e-02 -1.26636267e-01 7.11338520e-01 -9.46746469e-01
3.88105184e-01 -3.70970279e-01 9.62904274e-01 -7.42461681e-01
4.11152452e-01 -7.74940431e-01 -8.68567675e-02 8.51158857e-01
-3.00750226e-01 9.60329995e-02 7.52790496e-02 7.74747372e-01
-1.81767628e-01 -1.15690202e-01 1.31683624e+00 -4.80675131e-01
-6.89066231e-01 1.89673498e-01 -1.02454066e+00 5.54502197e-02
8.77411664e-01 -3.02162349e-01 -7.15499043e-01 -7.69767165e-02
-9.94405985e-01 -2.18704060e-01 5.17969012e-01 6.78571612e-02
6.27647758e-01 -9.81129169e-01 -4.87459213e-01 7.50750661e-01
-1.01171121e-01 -2.48501375e-01 -7.62359425e-02 4.79864836e-01
-6.36675656e-01 5.98510027e-01 -7.12272942e-01 -9.17507470e-01
-1.24664426e+00 4.37750161e-01 6.59373641e-01 -8.19813609e-02
-4.62621629e-01 8.04134309e-01 6.60107493e-01 1.19516164e-01
1.93775728e-01 -5.35768092e-01 -3.96598935e-01 -1.08373119e-02
7.15031564e-01 -1.22830123e-01 -2.08063319e-01 -7.03348756e-01
-5.11129797e-01 8.34359109e-01 -8.72200057e-02 -3.59149873e-01
1.10295737e+00 2.38536857e-02 -3.66742074e-01 8.17489803e-01
7.00422883e-01 2.12899014e-01 -1.35821176e+00 -1.97948471e-01
7.81621113e-02 -4.31314588e-01 -3.97126861e-02 -7.30125129e-01
-7.36842036e-01 7.19977379e-01 7.90844500e-01 3.48425776e-01
9.59018350e-01 1.95820525e-01 -1.75187975e-01 5.49596429e-01
3.32084954e-01 -5.96603453e-01 1.00801170e-01 7.49969542e-01
9.47458029e-01 -1.02808785e+00 -9.84168947e-02 -4.48135406e-01
-3.40077907e-01 1.11120296e+00 5.84868431e-01 -7.84834385e-01
1.05586946e+00 9.19030488e-01 2.58971840e-01 -5.87764680e-01
-9.21009183e-01 -2.88442820e-01 5.21478653e-01 9.83955741e-01
4.12786484e-01 6.52669817e-02 -3.25624943e-01 3.29237171e-02
8.98989066e-02 -2.52901852e-01 4.50420648e-01 8.28171849e-01
-9.85460997e-01 -4.80105996e-01 -3.60729247e-01 6.31609261e-01
-5.31840563e-01 -3.01845759e-01 -4.60879892e-01 7.30511308e-01
7.78552517e-02 5.90098619e-01 5.97783744e-01 1.46008685e-01
4.33849506e-02 5.03292046e-02 7.10017860e-01 -5.43017626e-01
-2.61536330e-01 8.97234142e-01 -3.86416405e-01 -5.54831088e-01
-5.82664609e-01 -8.01308811e-01 -8.48280430e-01 -7.29386136e-02
-3.56427401e-01 -7.26884156e-02 4.32238579e-01 7.90301919e-01
-5.21884188e-02 3.80688190e-01 -1.93429604e-01 -1.04470074e+00
-3.44425619e-01 -8.19644630e-01 -1.16995680e+00 2.37073079e-02
9.54027236e-01 -7.88590074e-01 -5.09762824e-01 2.67040610e-01] | [9.486054420471191, 2.3166518211364746] |
c8c90ef7-29df-4915-b359-c98a0bf12ef0 | discovery-of-natural-language-concepts-in | 1902.07249 | null | http://arxiv.org/abs/1902.07249v2 | http://arxiv.org/pdf/1902.07249v2.pdf | Discovery of Natural Language Concepts in Individual Units of CNNs | Although deep convolutional networks have achieved improved performance in
many natural language tasks, they have been treated as black boxes because they
are difficult to interpret. Especially, little is known about how they
represent language in their intermediate layers. In an attempt to understand
the representations of deep convolutional networks trained on language tasks,
we show that individual units are selectively responsive to specific morphemes,
words, and phrases, rather than responding to arbitrary and uninterpretable
patterns. In order to quantitatively analyze such an intriguing phenomenon, we
propose a concept alignment method based on how units respond to the replicated
text. We conduct analyses with different architectures on multiple datasets for
classification and translation tasks and provide new insights into how deep
models understand natural language. | ['Dong-Hyun Lee', 'Yo Joong Choe', 'Seil Na', 'Gunhee Kim'] | 2019-02-18 | discovery-of-natural-language-concepts-in-1 | https://openreview.net/forum?id=S1EERs09YQ | https://openreview.net/pdf?id=S1EERs09YQ | iclr-2019-5 | ['concept-alignment'] | ['computer-vision'] | [ 3.94360781e-01 2.22932268e-02 -1.99826345e-01 -5.91509759e-01
-5.73602617e-02 -8.31136346e-01 7.88339913e-01 4.45998251e-01
-4.50848460e-01 4.24352139e-01 6.37179732e-01 -4.76843446e-01
3.40656847e-01 -8.46960306e-01 -7.21492827e-01 -1.45718738e-01
1.57204375e-01 6.03193402e-01 -7.08202273e-02 -4.89476234e-01
1.85614586e-01 4.11538303e-01 -1.11224949e+00 8.55977774e-01
6.91352248e-01 5.05540192e-01 7.36710057e-03 2.21392885e-01
-5.08091807e-01 7.79207647e-01 -6.95494115e-01 -3.32128942e-01
-1.06736068e-02 -7.14427829e-01 -1.04051411e+00 -1.67971209e-01
3.90263855e-01 -6.34315982e-02 -3.24497461e-01 1.07824135e+00
7.16608763e-02 -1.72390640e-02 9.13751006e-01 -7.51224637e-01
-1.40293169e+00 9.48610187e-01 -2.53110766e-01 4.09289181e-01
2.46550232e-01 1.06807813e-01 1.47676229e+00 -7.53836989e-01
5.80112278e-01 1.46077621e+00 3.55817616e-01 8.16041291e-01
-1.60697675e+00 -5.48251867e-01 3.49638641e-01 5.17677777e-02
-9.13921893e-01 -5.08387506e-01 5.64322650e-01 -3.67872119e-01
1.13018072e+00 -6.30072579e-02 3.13773900e-01 1.41809082e+00
4.85397816e-01 9.27804828e-01 9.88017499e-01 -4.41494703e-01
-2.77460013e-02 3.40036303e-02 4.56589222e-01 6.77714765e-01
4.44208115e-01 -1.96062654e-01 -4.42254990e-01 6.63103759e-02
6.25010312e-01 1.98865637e-01 -2.92550296e-01 1.63540989e-01
-1.49456298e+00 9.45068598e-01 7.76870966e-01 8.73662114e-01
-1.98737711e-01 3.20317507e-01 3.95448864e-01 5.07566750e-01
4.15325344e-01 9.50453341e-01 -6.44749820e-01 2.26215810e-01
-3.66993755e-01 6.07123077e-02 6.08509302e-01 6.13905430e-01
8.77580762e-01 1.90893054e-01 -1.70675099e-01 8.13829362e-01
1.70834288e-01 2.45874956e-01 8.52263391e-01 -4.73879516e-01
5.34226120e-01 8.45283210e-01 -2.80790895e-01 -1.05974483e+00
-5.32366216e-01 -5.56089044e-01 -9.96217310e-01 -1.99145585e-01
4.94393945e-01 1.47053460e-02 -9.65467989e-01 2.02863050e+00
-6.63485229e-01 -4.08961952e-01 2.35471949e-01 7.27025628e-01
7.96872020e-01 7.14026272e-01 3.28036636e-01 2.17443749e-01
1.46733105e+00 -7.39670336e-01 -6.35579705e-01 -7.99559474e-01
8.33594441e-01 -4.87571985e-01 1.30779505e+00 4.60936800e-02
-1.01136065e+00 -6.43800676e-01 -8.94721091e-01 -4.38951254e-01
-6.75518513e-01 -8.60702842e-02 6.50318623e-01 2.00220436e-01
-1.18164659e+00 5.02692461e-01 -6.53753936e-01 -6.59761369e-01
5.14085650e-01 1.66221708e-01 -3.61325771e-01 1.55178785e-01
-1.27574944e+00 9.55282807e-01 4.28637207e-01 2.62495037e-02
-7.49323249e-01 -3.85456592e-01 -8.75161827e-01 6.43949926e-01
-6.64516240e-02 -7.47180939e-01 1.35551035e+00 -1.36867309e+00
-1.15437114e+00 1.15340221e+00 -4.86183226e-01 -5.91248512e-01
-2.36081511e-01 2.80554872e-02 -3.86656106e-01 2.25397162e-02
-4.58818525e-02 6.57721460e-01 4.97329921e-01 -1.07830763e+00
-2.49141797e-01 -3.80359888e-01 1.88226417e-01 -9.79243591e-02
-4.70817983e-01 2.07151741e-01 9.67272744e-02 -9.28964376e-01
3.45491916e-01 -8.23670506e-01 -1.85453594e-01 -4.05757800e-02
-2.84787178e-01 -3.00130039e-01 3.80556762e-01 -4.07467455e-01
9.37548101e-01 -2.02281857e+00 2.92743415e-01 -1.21158883e-01
4.80995238e-01 1.28224283e-01 -3.92966956e-01 5.82460105e-01
-1.53836980e-01 7.31521189e-01 -4.70969290e-01 -1.78994641e-01
-4.70421137e-03 3.64513814e-01 -9.99573708e-01 7.33740851e-02
6.98189914e-01 1.29337370e+00 -1.13801730e+00 1.58276428e-02
-1.86639428e-01 2.91053087e-01 -4.99789476e-01 9.70678255e-02
-4.17215198e-01 2.88206518e-01 -5.84569037e-01 4.33480650e-01
2.19695970e-01 -5.61360419e-01 4.54833925e-01 -7.54761845e-02
1.54920012e-01 7.66581535e-01 -3.48348022e-01 1.44237602e+00
-4.40065026e-01 1.20110071e+00 -3.19600999e-01 -1.39096427e+00
9.64234054e-01 4.40241247e-01 -2.53997240e-02 -9.50475752e-01
3.00266653e-01 1.86901718e-01 4.50527012e-01 -2.46811360e-01
3.77544016e-01 -3.92369896e-01 -2.71746010e-01 8.44559669e-01
2.43128106e-01 -2.12998502e-02 1.39319763e-01 2.81268686e-01
1.09218597e+00 -3.36833715e-01 4.69552964e-01 -3.50791216e-01
1.77275658e-01 -1.67512894e-01 3.39512914e-01 6.42693460e-01
-1.40963301e-01 6.77502632e-01 7.77899444e-01 -8.94000828e-01
-8.53997171e-01 -1.16472602e+00 -3.96292061e-02 1.26823270e+00
-4.36092913e-02 -4.66319382e-01 -4.76692796e-01 -3.81379873e-01
-8.01378936e-02 7.76643872e-01 -9.77439404e-01 -5.73468089e-01
-5.37340701e-01 -7.60546863e-01 7.90439546e-01 7.64373779e-01
3.67601454e-01 -1.30756187e+00 -4.54389125e-01 2.23581761e-01
-9.28763822e-02 -1.07202399e+00 -1.79956555e-01 4.45754886e-01
-8.15167487e-01 -9.59375858e-01 -3.37033629e-01 -9.20401394e-01
9.69832182e-01 3.98975551e-01 1.51209366e+00 2.81671822e-01
1.32168695e-01 1.62516430e-01 -2.75850773e-01 -5.44749200e-01
-6.79929912e-01 3.72177154e-01 7.84888342e-02 -9.09747332e-02
8.71912539e-01 -6.07026756e-01 -3.55349779e-01 1.04567803e-01
-1.13475943e+00 1.26834676e-01 6.63398504e-01 9.08776581e-01
1.14508249e-01 -4.22008067e-01 4.73794848e-01 -1.08312380e+00
1.19819736e+00 -6.96743071e-01 -1.99597001e-01 3.67989123e-01
-4.46809292e-01 5.20439625e-01 1.06992555e+00 -3.19124728e-01
-7.06193805e-01 -2.98575938e-01 -1.61129553e-02 1.46912441e-01
-3.37415516e-01 7.51635551e-01 2.51185954e-01 3.35511923e-01
8.40191364e-01 3.60549003e-01 -2.96041369e-01 -4.97626394e-01
4.13302571e-01 4.89538968e-01 3.99722934e-01 -9.31155860e-01
7.97490001e-01 3.70943874e-01 -3.79811972e-01 -6.42904162e-01
-1.02220404e+00 -3.59380548e-03 -6.51622832e-01 3.36270481e-01
9.70135033e-01 -7.53082752e-01 -5.68531394e-01 1.48215219e-01
-1.56100774e+00 -3.41247976e-01 -2.86855940e-02 5.61292805e-02
-2.78546512e-01 3.36380713e-02 -6.77824378e-01 -4.39607054e-01
-1.66775659e-01 -1.14558363e+00 8.12592745e-01 1.67465866e-01
-6.48669660e-01 -1.22923672e+00 3.22035886e-02 7.42833270e-03
6.68773293e-01 -7.73750618e-02 1.46754611e+00 -1.04932654e+00
-4.32992727e-01 -5.65083101e-02 -4.05241907e-01 1.94451362e-01
4.37528759e-01 8.29645023e-02 -1.03633821e+00 -1.73986077e-01
-1.64788514e-01 -4.93681431e-01 1.24225330e+00 5.00126136e-03
1.60270047e+00 -4.23034340e-01 -3.11080396e-01 5.33702612e-01
1.06528234e+00 1.42346427e-01 4.35731053e-01 3.19271296e-01
4.15229440e-01 6.46082878e-01 -3.66802335e-01 -1.78719774e-01
2.17967704e-01 3.17874491e-01 1.04936883e-01 -1.60674036e-01
6.33573234e-02 -3.80305201e-01 4.08484548e-01 8.10347259e-01
5.81942461e-02 -6.13843620e-01 -1.15245903e+00 5.77533126e-01
-1.65747428e+00 -8.93087924e-01 2.08318159e-01 1.73464000e+00
9.58239913e-01 3.91589552e-01 -3.84547293e-01 -2.40772918e-01
4.36375439e-01 4.75389987e-01 -4.91819143e-01 -7.69010067e-01
-5.10114133e-01 5.23550630e-01 5.13717718e-02 3.26290399e-01
-6.73425198e-01 1.13960135e+00 7.27205372e+00 1.58391267e-01
-1.49360812e+00 -9.43613797e-02 8.79161298e-01 3.04161996e-01
-6.59663618e-01 -3.01421974e-02 -3.81530106e-01 2.14744553e-01
9.32650685e-01 -2.54539639e-01 4.49067563e-01 6.17389083e-01
6.34846985e-02 3.52585852e-01 -1.62439036e+00 6.10739708e-01
-9.78388563e-02 -1.58949184e+00 6.99932933e-01 -2.91265607e-01
7.26772130e-01 2.93122143e-01 2.18346685e-01 4.27407742e-01
6.24069631e-01 -1.46477199e+00 5.35572231e-01 4.01970536e-01
4.72704202e-01 -1.99091777e-01 3.87883544e-01 1.88380450e-01
-8.81012142e-01 -2.73799095e-02 -5.80177605e-01 -5.91276705e-01
-1.32871717e-01 1.37285456e-01 -8.10197890e-01 -9.19159278e-02
4.84601647e-01 6.37574434e-01 -6.49688959e-01 3.31668735e-01
-4.83406693e-01 7.75844872e-01 1.26734629e-01 -2.36076221e-01
3.75560403e-01 -3.33093554e-02 1.74693882e-01 1.28016102e+00
6.94038644e-02 3.10668908e-02 -9.88546684e-02 1.51602852e+00
-6.23225927e-01 -5.91401476e-04 -7.89738178e-01 -7.67206907e-01
2.10563585e-01 9.63928223e-01 -7.42424488e-01 -4.81759936e-01
-5.92224360e-01 9.67523098e-01 7.44146466e-01 8.53489816e-01
-2.81680942e-01 -2.59749413e-01 1.03450692e+00 -1.39492322e-02
-2.47400060e-01 -5.24896026e-01 -5.37469923e-01 -1.51514935e+00
-8.12773034e-02 -8.23620856e-01 1.23406179e-01 -9.34037089e-01
-1.47115028e+00 6.95971847e-01 -2.81002969e-01 -8.32329512e-01
-3.90389472e-01 -1.12678492e+00 -8.85713220e-01 8.71212065e-01
-1.30702055e+00 -7.72087276e-01 8.21106881e-02 2.25804046e-01
5.81135571e-01 -4.19021875e-01 1.16440618e+00 -8.32324103e-02
-5.70482135e-01 6.18466973e-01 2.46268585e-01 6.79669976e-01
6.55684471e-01 -1.09208632e+00 9.80066061e-01 6.57691360e-01
7.33275533e-01 1.35077310e+00 6.26075804e-01 -1.32575780e-01
-1.19172132e+00 -8.89461517e-01 1.11710274e+00 -5.34304619e-01
8.42985272e-01 -6.96472406e-01 -1.29841971e+00 1.04911137e+00
5.90911686e-01 -1.11232854e-01 7.76006281e-01 4.03852642e-01
-8.59986126e-01 1.22243203e-02 -6.60718679e-01 9.12198782e-01
9.62940574e-01 -8.94223571e-01 -1.10635984e+00 3.02466571e-01
1.07540274e+00 4.11028825e-02 -3.77269119e-01 1.93900540e-01
4.11499977e-01 -6.60751045e-01 7.95806646e-01 -1.46918702e+00
8.82698536e-01 -1.07883416e-01 -2.59003818e-01 -1.55684006e+00
-5.88608623e-01 -2.18017846e-01 1.86832219e-01 9.23699200e-01
8.15506816e-01 -8.46718907e-01 3.71519655e-01 6.53912723e-01
-1.54686794e-01 -7.13290215e-01 -5.99660993e-01 -4.13906902e-01
5.05509198e-01 -1.21619508e-01 5.34295678e-01 1.34794509e+00
1.54261529e-01 7.98451424e-01 5.33332750e-02 -1.70155346e-01
-6.62007332e-02 2.53102869e-01 3.31802905e-01 -1.24359381e+00
-1.69172227e-01 -9.20847297e-01 -3.34815860e-01 -1.12395012e+00
6.49780035e-01 -1.33484590e+00 -1.34044915e-01 -1.58261991e+00
4.17594492e-01 -7.77605399e-02 -5.67006707e-01 8.11163604e-01
-1.76303625e-01 2.54215091e-01 1.66989967e-01 1.52149916e-01
-2.37578541e-01 4.87134814e-01 9.72182691e-01 -5.85232615e-01
1.33025646e-01 -3.15119326e-01 -1.11461782e+00 6.57261789e-01
9.60247576e-01 -2.70510733e-01 -3.61387283e-01 -1.35256803e+00
6.72753215e-01 -4.86530423e-01 3.51587534e-01 -6.30899429e-01
-1.54787712e-02 -2.78732687e-01 4.90711451e-01 -2.50709683e-01
-1.56072341e-02 -5.79068601e-01 -4.20209527e-01 4.99669224e-01
-9.64700103e-01 4.01814431e-01 4.20682311e-01 3.53196949e-01
-5.00436544e-01 -6.85247406e-02 6.29366577e-01 -4.95603323e-01
-7.22515285e-01 1.46281451e-01 -7.85441995e-01 2.12042883e-01
3.35857511e-01 8.09883326e-02 -7.20014513e-01 -5.15607357e-01
-4.85942483e-01 1.93624139e-01 5.10657132e-01 7.35458434e-01
5.88630259e-01 -1.33485174e+00 -6.86936855e-01 1.66933432e-01
2.97909737e-01 -1.10811777e-01 -1.76535353e-01 2.25706294e-01
-4.28548694e-01 7.38335431e-01 -3.50236207e-01 -3.23015422e-01
-6.00197196e-01 4.93013382e-01 5.51255941e-01 -1.14878558e-01
-1.66221172e-01 7.83589423e-01 6.89897597e-01 -5.50708771e-01
-6.46256283e-02 -7.27999508e-01 -2.27252245e-01 -4.48193960e-02
5.12404621e-01 -5.03506780e-01 -2.05423310e-02 -6.08760476e-01
-2.84563243e-01 3.80178303e-01 -3.44979465e-01 1.69321790e-01
1.38110614e+00 1.31357551e-01 -3.85334671e-01 6.85352325e-01
1.23346639e+00 -1.77413538e-01 -5.76642871e-01 -5.20017028e-01
2.24780232e-01 -3.25469822e-02 -3.51109177e-01 -6.77567840e-01
-9.63927925e-01 1.45533633e+00 1.30243242e-01 3.75102669e-01
8.77349675e-01 1.56535164e-01 6.20244145e-01 8.92809510e-01
4.45811786e-02 -7.99877644e-01 2.28457123e-01 1.01506948e+00
1.02888668e+00 -1.26293182e+00 -3.67503226e-01 -4.06503072e-03
-3.45681190e-01 1.48631799e+00 7.17429757e-01 -6.66048974e-02
3.84210706e-01 -6.28400296e-02 1.86080396e-01 -1.24181256e-01
-1.05224669e+00 -1.52464300e-01 4.01152194e-01 3.77566814e-01
1.14219403e+00 1.04508273e-01 -2.21081421e-01 7.50949085e-01
-2.68114597e-01 -3.19173694e-01 5.05482018e-01 5.86465776e-01
-5.19112229e-01 -1.14879298e+00 -1.31774634e-01 3.99018615e-01
-4.35952842e-01 -4.60160464e-01 -9.65195894e-01 5.09263277e-01
-2.60541379e-01 6.28487408e-01 3.31866831e-01 -4.83517945e-01
1.27417549e-01 4.25347805e-01 2.12258771e-01 -1.07608593e+00
-4.75678086e-01 -4.14368272e-01 -4.73696962e-02 -2.48649970e-01
-2.84627676e-01 -3.92472923e-01 -1.40920377e+00 -1.65615693e-01
2.57747024e-01 6.91439733e-02 2.54640788e-01 1.20914185e+00
4.65914965e-01 5.84807873e-01 3.78508031e-01 -6.04916871e-01
-5.46312869e-01 -1.01830661e+00 -8.19944665e-02 7.15882182e-01
4.21982169e-01 -3.45550984e-01 -1.68602973e-01 -1.63963325e-02] | [10.61548900604248, 8.846811294555664] |
98ad51a4-eaad-4c17-9fde-6fbb52a3cdd2 | rqdia-regularizing-q-value-distributions-with | null | null | https://openreview.net/forum?id=rqcLsG8Kme9 | https://openreview.net/pdf?id=rqcLsG8Kme9 | rQdia: Regularizing Q-Value Distributions With Image Augmentation | rQdia (pronounced “Arcadia”) regularizes Q-value distributions with augmented images in pixel-based deep reinforcement learning. With a simple auxiliary loss, that equalizes these distributions via MSE, rQdia boosts DrQ and SAC on 9/12 and 10/12 tasks respectively in the MuJoCo Continuous Control Suite from pixels, and Data-Efficient Rainbow on 18/26 Atari Arcade environments. Gains are measured in both sample efficiency and longer-term training. Moreover, the addition of rQdia finally propels model-free continuous control from pixels over the state encoding baseline. Additional results, namely more random seeds, pending. | ['Chenliang Xu', 'Jing Bi', 'Samuel Lerman'] | 2021-09-29 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [-1.22315660e-01 -2.23886650e-02 -4.33315128e-01 -2.90809602e-01
-1.16930389e+00 -4.79286432e-01 8.93758237e-01 -2.83423662e-01
-9.81996715e-01 1.00837231e+00 1.35261742e-02 -3.38185102e-01
-5.89147322e-02 -7.33837903e-01 -1.06953394e+00 -7.94022858e-01
-4.99784321e-01 2.47323319e-01 6.91418201e-02 -3.52109790e-01
1.75885022e-01 3.72978926e-01 -1.05544603e+00 -7.60621130e-02
6.86057925e-01 1.21344399e+00 1.01855621e-01 8.36551249e-01
5.11791825e-01 1.22365129e+00 -4.29595649e-01 -1.15951665e-01
8.56722653e-01 -3.37289810e-01 -5.84088922e-01 1.47191286e-01
9.66973424e-01 -9.18550968e-01 -7.57125378e-01 9.19414580e-01
6.80676699e-01 5.35355508e-01 2.25385800e-01 -1.28016376e+00
-4.76487577e-01 5.95538497e-01 -7.05020249e-01 4.15047526e-01
-3.65201205e-01 1.20055890e+00 1.13430560e+00 -2.81334281e-01
5.91611683e-01 1.39890945e+00 3.85454684e-01 5.87445557e-01
-1.72578239e+00 -6.13632858e-01 3.60685378e-01 -3.06273228e-03
-9.57382977e-01 -6.11545920e-01 3.53614658e-01 -1.18785836e-01
1.32645071e+00 -2.13297173e-01 4.70768243e-01 1.11016989e+00
1.07564054e-01 1.17465580e+00 1.43740821e+00 -1.39582977e-01
7.19480574e-01 -4.21715260e-01 -1.74549997e-01 7.57982075e-01
-1.63278535e-01 8.51285458e-01 -2.19864294e-01 2.50989407e-01
1.09734118e+00 -4.46223199e-01 3.35074961e-02 -6.54295385e-01
-9.90827739e-01 9.87940252e-01 7.27378964e-01 -3.61764789e-01
-4.68126297e-01 1.15968704e+00 4.47230846e-01 3.58916104e-01
1.65558934e-01 7.46181488e-01 -5.59588492e-01 -4.98465508e-01
-8.44353497e-01 6.37755990e-01 2.31329441e-01 1.02271974e+00
7.55458951e-01 7.77873933e-01 -6.07011974e-01 5.42884409e-01
8.41361582e-02 1.03886604e+00 3.75871748e-01 -1.83420002e+00
8.15459132e-01 -1.84402615e-01 4.82535392e-01 -4.00116533e-01
-4.95459080e-01 -6.84510887e-01 -4.51654941e-01 9.97445941e-01
4.92751151e-01 -6.78415656e-01 -1.40740514e+00 2.19150352e+00
1.74976468e-01 1.74242198e-01 1.23561718e-01 1.00282013e+00
-1.20979182e-01 7.40740180e-01 1.22225083e-01 2.68278390e-01
1.02961016e+00 -1.23096919e+00 -3.20334524e-01 -7.19941080e-01
2.43442193e-01 4.57638651e-02 1.30109847e+00 5.02423882e-01
-1.31477678e+00 -6.17736816e-01 -9.88680482e-01 5.26485182e-02
3.53154843e-03 1.38417780e-01 9.15214241e-01 4.03932452e-01
-1.35541272e+00 7.52777696e-01 -1.40596533e+00 1.33111566e-01
9.26726699e-01 5.14470041e-01 -1.24849416e-01 -4.63418849e-03
-1.11379957e+00 8.76353264e-01 2.42472112e-01 -1.42628491e-01
-1.85500824e+00 -9.29848731e-01 -8.41430187e-01 1.03363171e-01
5.90427637e-01 -5.10845184e-01 1.59808135e+00 -1.01478982e+00
-1.87897885e+00 4.88485336e-01 4.34721977e-01 -1.32709682e+00
7.06944048e-01 -4.41719055e-01 -1.21246569e-01 3.84732932e-01
1.25244290e-01 1.40985250e+00 9.14242864e-01 -8.62982929e-01
-8.01804304e-01 -2.50453353e-01 5.10853589e-01 2.56863803e-01
1.25868797e-01 -3.87023270e-01 -5.21312654e-01 -5.97430646e-01
-6.20464563e-01 -1.05406153e+00 -6.83300495e-01 3.30584466e-01
-7.10449219e-02 -2.58544879e-03 3.83474171e-01 -7.05051720e-01
7.55166173e-01 -2.11497498e+00 8.25408697e-02 7.18891323e-02
5.41506819e-02 5.74482307e-02 -7.57756770e-01 7.32863396e-02
-3.44209932e-02 -1.41634509e-01 -4.85026091e-01 -2.17953727e-01
2.94475883e-01 4.49925184e-01 -2.08570153e-01 6.39171302e-01
5.69035232e-01 9.78688419e-01 -9.06681538e-01 -1.95779607e-01
2.50948220e-01 5.18043637e-02 -9.79151905e-01 -2.38866154e-02
-7.68112898e-01 2.82780945e-01 -3.36312652e-01 2.90182829e-01
6.82498932e-01 -1.10542268e-01 2.33819522e-03 1.73126444e-01
-6.38422295e-02 1.57120526e-01 -9.77373481e-01 2.20698285e+00
-5.82635522e-01 5.53600907e-01 4.23995048e-01 -9.19166803e-01
5.05380750e-01 -1.34349763e-01 3.08703035e-01 -1.42960334e+00
-8.88875592e-03 -2.70447433e-01 4.79287468e-02 -1.93076625e-01
6.53932273e-01 -1.07795850e-03 -1.18051760e-01 1.63012937e-01
1.37803465e-01 -3.58616263e-01 5.29867768e-01 2.82516778e-01
1.43984747e+00 6.53744817e-01 -1.82357118e-01 -2.88261712e-01
-1.21656515e-01 1.09678380e-01 6.05160952e-01 8.99085939e-01
-6.50555372e-01 4.57232207e-01 9.29129541e-01 -1.32907286e-01
-1.20297205e+00 -1.39547527e+00 -1.56239821e-02 1.13924944e+00
-1.46138951e-01 -7.61949420e-02 -5.20775020e-01 -9.81076717e-01
3.52611631e-01 1.01489437e+00 -8.80583942e-01 -3.03730398e-01
-6.58202410e-01 -5.66827595e-01 6.89450145e-01 6.89099491e-01
9.16477323e-01 -1.07015812e+00 -8.10893953e-01 2.30366752e-01
3.15464348e-01 -8.47126603e-01 -4.23184156e-01 6.30679250e-01
-9.81494188e-01 -8.14902604e-01 -6.72440946e-01 -2.35157564e-01
2.26445153e-01 -1.01187117e-01 1.45423555e+00 -3.99566889e-01
-1.11356534e-01 1.94986269e-01 -2.74154544e-01 6.44538999e-02
-1.58200786e-01 9.72860027e-03 -1.92788169e-01 -4.83941495e-01
-1.85886130e-01 -3.26460570e-01 -7.49764085e-01 -2.36577671e-02
-7.65972018e-01 -2.32801706e-01 5.80348969e-01 1.24880564e+00
5.55724502e-01 -2.95277029e-01 4.54508841e-01 -6.85812056e-01
2.43802860e-01 -4.39170659e-01 -1.35129678e+00 -2.53042281e-01
-8.99722219e-01 6.08589888e-01 7.12458014e-01 -2.31629372e-01
-1.02176249e+00 -1.08693708e-02 -1.52823880e-01 -8.44622612e-01
-1.65291950e-02 1.03448570e-01 2.17034712e-01 2.74637580e-01
9.62346256e-01 1.10347323e-01 3.55433702e-01 -1.95752412e-01
8.30982745e-01 1.76598318e-02 6.63752198e-01 -8.41468513e-01
5.74201703e-01 4.27120358e-01 -1.24762341e-01 -2.37246662e-01
-5.13569653e-01 4.66495678e-02 1.96061552e-01 6.83068335e-02
9.39156651e-01 -1.32196712e+00 -6.14250362e-01 5.00377774e-01
-5.07035673e-01 -1.35536730e+00 -8.55503500e-01 4.42187130e-01
-1.01802218e+00 -2.12620944e-02 -7.82692611e-01 -6.16647482e-01
-6.85872659e-02 -1.23491216e+00 1.00014460e+00 3.41889888e-01
2.97099948e-01 -7.48708308e-01 2.80014277e-01 1.39152110e-01
5.89224458e-01 1.74244702e-01 5.68525195e-01 -1.38035538e-02
-6.73596680e-01 8.30977410e-02 -1.65715456e-01 8.23564053e-01
-2.45118216e-01 -1.13772765e-01 -9.04677451e-01 -5.95692515e-01
-6.60698533e-01 -9.79749203e-01 1.12016666e+00 6.23113513e-01
1.13508844e+00 -2.63567656e-01 1.62864119e-01 6.81012988e-01
1.51322126e+00 3.65759581e-01 8.18022609e-01 5.09030044e-01
3.37152928e-01 -3.20530473e-03 8.55335712e-01 8.11586857e-01
1.58999294e-01 6.35857284e-01 9.76535082e-01 -2.72485971e-01
-4.16763246e-01 -3.66412908e-01 8.10595870e-01 -2.00865895e-01
2.16271773e-01 -1.87019005e-01 -6.28372431e-01 6.48113370e-01
-1.71017444e+00 -1.10210085e+00 2.35468283e-01 1.85970223e+00
1.13516521e+00 3.26863974e-01 3.95769417e-01 -3.28203917e-01
4.11431551e-01 3.41876715e-01 -1.16151428e+00 -3.03246647e-01
-2.64299721e-01 7.35352576e-01 1.02326488e+00 5.71694970e-01
-1.26193511e+00 1.36254632e+00 6.74700737e+00 1.16701281e+00
-1.07159781e+00 1.72695249e-01 1.01753247e+00 -5.63849986e-01
-1.74132347e-01 -5.21111935e-02 -5.84297061e-01 4.49472994e-01
1.23581946e+00 2.82984048e-01 8.91821742e-01 1.19775236e+00
4.61177796e-01 -4.33559030e-01 -9.26706851e-01 5.28978169e-01
-3.79468828e-01 -1.47066295e+00 -5.26537001e-01 1.28170550e-01
1.02565575e+00 6.24159157e-01 4.99385864e-01 8.73830378e-01
1.23319244e+00 -8.79309356e-01 1.05841589e+00 3.19484442e-01
8.24109793e-01 -9.62106466e-01 3.19166988e-01 9.19131190e-03
-5.72344303e-01 -3.83633286e-01 -4.02806491e-01 -7.22550526e-02
-1.09228984e-01 2.16735616e-01 -6.58366144e-01 1.92541569e-01
6.44494772e-01 6.97235882e-01 -6.35976255e-01 7.76027381e-01
-5.41771352e-01 9.11158979e-01 -3.57701480e-01 1.68068424e-01
8.23692381e-01 -2.53140152e-01 3.33151639e-01 9.48555350e-01
-2.69218326e-01 -2.32148141e-01 1.42489552e-01 1.13197494e+00
-1.39412791e-01 -6.82229936e-01 -2.58176029e-01 1.38309700e-02
3.12613457e-01 1.18218184e+00 -3.62544477e-01 -4.60594028e-01
-7.21944273e-02 9.42370117e-01 4.61359352e-01 4.91221964e-01
-1.26603699e+00 -2.97985524e-01 1.11306024e+00 -3.51086929e-02
6.55845165e-01 -2.72882491e-01 -2.53204048e-01 -8.93290281e-01
-3.97866160e-01 -1.09416008e+00 1.56614348e-01 -8.28506052e-01
-1.06692779e+00 2.40206078e-01 4.80429009e-02 -1.00681496e+00
-4.69221234e-01 -6.94626093e-01 -2.77622908e-01 7.46052444e-01
-1.69747627e+00 -7.22939610e-01 1.80225492e-01 6.66465938e-01
5.92360735e-01 -1.93781435e-01 4.88173634e-01 3.37928623e-01
-9.09405351e-01 8.14781129e-01 4.59674507e-01 1.75635368e-01
5.90633571e-01 -1.64673746e+00 7.64205098e-01 1.00645530e+00
-1.38202757e-01 8.99042264e-02 6.27312481e-01 -1.69281930e-01
-1.50218761e+00 -1.25251603e+00 -2.44651705e-01 -2.74598032e-01
8.32196236e-01 -1.80612072e-01 -2.86076695e-01 7.19964623e-01
5.65682411e-01 3.03726196e-01 -2.00794473e-01 -1.73960403e-01
-3.94017041e-01 -4.18008685e-01 -1.23605847e+00 6.96333230e-01
8.32275033e-01 -1.97537467e-01 -1.71225101e-01 3.57470930e-01
7.25448310e-01 -4.56261069e-01 -7.39366055e-01 -4.05802913e-02
8.57123882e-02 -8.35808814e-01 8.97919595e-01 -8.93227100e-01
7.96618700e-01 -2.64350027e-01 -3.51003379e-01 -1.58078730e+00
-3.67823601e-01 -7.47453511e-01 4.71851183e-03 7.38283515e-01
3.90570164e-01 -3.49474758e-01 1.14316607e+00 4.10608500e-01
-3.68294716e-01 -5.51519454e-01 -9.89885390e-01 -1.06648767e+00
6.41671658e-01 -4.70127851e-01 1.93894252e-01 5.18749714e-01
-4.84372467e-01 1.68346316e-01 -3.94533098e-01 3.39428149e-02
7.10642517e-01 -2.14515999e-01 8.12019765e-01 -2.22402558e-01
-9.08579290e-01 -5.41955769e-01 -1.09650262e-01 -1.38995492e+00
1.67355910e-01 -6.79721713e-01 3.70464325e-01 -1.26801085e+00
-2.11316988e-01 -7.62593865e-01 -3.50181401e-01 9.03194904e-01
-5.21394983e-02 8.76954719e-02 4.15610969e-01 -1.67345986e-01
-8.37806225e-01 1.04054856e+00 1.27916527e+00 -3.90081316e-01
6.02257513e-02 -2.47736648e-01 -4.43800747e-01 3.04160386e-01
8.90357077e-01 -3.48523021e-01 -4.89283681e-01 -6.86581433e-01
-7.45020285e-02 4.46467370e-01 5.50503671e-01 -1.26961935e+00
-2.64217466e-01 -4.81183738e-01 2.74531603e-01 -1.53565347e-01
4.42224205e-01 -5.76529145e-01 -3.43772411e-01 7.34752238e-01
-5.92065573e-01 2.86925703e-01 5.48535585e-01 6.47504985e-01
1.96706690e-02 4.70049940e-02 1.24493635e+00 -1.86255485e-01
-9.60158408e-01 3.29287261e-01 -4.26559895e-01 5.26414931e-01
1.04720962e+00 2.29398832e-01 -4.89741504e-01 -4.17964965e-01
-5.39023519e-01 6.61434293e-01 2.65789002e-01 7.45225921e-02
2.41367683e-01 -1.25524676e+00 -7.40921795e-01 4.61332537e-02
-2.18022943e-01 4.95183356e-02 2.24085182e-01 6.55183494e-01
-7.86136329e-01 1.03262581e-01 -5.75952053e-01 -4.43974316e-01
-4.30249244e-01 4.07621861e-01 6.88819587e-01 -5.41410506e-01
-4.16669339e-01 1.01813817e+00 -1.74067616e-02 -3.63595963e-01
2.98445344e-01 -4.26457167e-01 3.70758057e-01 -1.98664352e-01
2.77591348e-01 4.86733615e-01 -6.25796318e-02 1.98207304e-01
-2.02830642e-01 -1.17017381e-01 -3.45083699e-02 -7.66539216e-01
1.34478462e+00 2.42424197e-02 4.56062555e-01 -2.05449015e-02
1.02190125e+00 -3.16821128e-01 -2.66094875e+00 9.45518315e-02
-2.64524370e-01 -2.30758667e-01 4.76538569e-01 -1.45893919e+00
-1.68942726e+00 9.49710250e-01 1.09330964e+00 -3.03380817e-01
1.14583254e+00 -4.48922813e-01 5.63319445e-01 5.16676605e-01
2.81466782e-01 -1.35224402e+00 2.61215329e-01 6.14386678e-01
8.47467303e-01 -1.23630083e+00 -6.53767884e-02 6.11229241e-01
-9.83036578e-01 6.45168424e-01 9.64181781e-01 -7.61512697e-01
2.55645424e-01 5.83818793e-01 -5.11749312e-02 1.51448727e-01
-1.04015672e+00 -3.56120467e-01 -5.27513683e-01 5.81146061e-01
-5.08364961e-02 1.74669176e-02 8.84256139e-02 1.41563222e-01
1.08922862e-01 2.87657771e-02 4.30657864e-01 1.03132439e+00
-4.38258469e-01 -9.66468513e-01 -1.13873892e-01 4.06154722e-01
-3.89435500e-01 -2.68150508e-01 3.98918450e-01 1.17382967e+00
-7.11270943e-02 8.22524130e-01 5.52536786e-01 -1.97769672e-01
2.49308243e-01 -3.89865279e-01 5.26691258e-01 -2.23376989e-01
-7.93786287e-01 6.43696040e-02 2.38501236e-01 -1.02122378e+00
-6.09412603e-02 -7.14081049e-01 -1.57229638e+00 -5.09763420e-01
-1.40864328e-01 2.54080836e-02 5.60367942e-01 5.27764142e-01
6.50451541e-01 6.99631333e-01 7.87105501e-01 -7.29405701e-01
-1.35068738e+00 -9.84626949e-01 -5.66283464e-01 2.77860433e-01
3.68937105e-01 -5.59421480e-01 -2.09484100e-01 -1.63302138e-01] | [4.190040588378906, 1.5809446573257446] |
45873268-157a-4f1c-a606-72f262ae7c47 | cs-mlgcn-multiplex-graph-convolutional | 2210.08811 | null | https://arxiv.org/abs/2210.08811v1 | https://arxiv.org/pdf/2210.08811v1.pdf | CS-MLGCN : Multiplex Graph Convolutional Networks for Community Search in Multiplex Networks | Community Search (CS) is one of the fundamental tasks in network science and has attracted much attention due to its ability to discover personalized communities with a wide range of applications. Given any query nodes, CS seeks to find a densely connected subgraph containing query nodes. Most existing approaches usually study networks with a single type of proximity between nodes, which defines a single view of a network. However, in many applications such as biological, social, and transportation networks, interactions between objects span multiple aspects, yielding networks with multiple views, called multiplex networks. Existing CS approaches in multiplex networks adopt pre-defined subgraph patterns to model the communities, which cannot find communities that do not have such pre-defined patterns in real-world networks. In this paper, we propose a query-driven graph convolutional network in multiplex networks, CS-MLGCN, that can capture flexible community structures by learning from the ground-truth communities in a data-driven fashion. CS-MLGCN first combines the local query-dependent structure and global graph embedding in each type of proximity and then uses an attention mechanism to incorporate information on different types of relations. Experiments on real-world graphs with ground-truth communities validate the quality of the solutions we obtain and the efficiency of our model. | ['Farnoosh Hashemi', 'Ali Behrouz'] | 2022-10-17 | null | null | null | null | ['community-search'] | ['graphs'] | [-0.08506904 0.0367949 -0.41213173 -0.16823293 0.1469012 -0.8577075
0.4956891 0.41481018 0.1877222 0.29252484 0.20757274 -0.14063318
-0.6012214 -1.3368144 -0.5142181 -0.4187539 -0.45879415 0.7088097
0.63771105 -0.27572972 0.01164853 0.39384824 -0.8861392 0.16728164
0.81235206 0.55293393 0.1686451 0.42376295 -0.11140826 0.5193599
-0.16078474 -0.3739839 0.2275072 -0.30860507 -0.87969476 0.2868828
0.30686352 0.27612346 -1.0636995 1.1430995 0.27792633 -0.24298471
0.3482545 -1.625912 -0.981887 0.80102813 -0.53992695 0.3257278
0.3831536 0.23550662 1.6177189 -0.53369987 0.93208516 1.4531952
0.49166805 0.22076117 -1.5540993 -0.6410566 0.37867296 0.20454426
-1.561998 0.02238633 0.9448391 -0.6044707 0.5972101 0.1169754
1.0343229 0.7270543 -0.05094709 0.5014517 0.76075333 0.10695755
-0.00559107 -0.17030838 -0.00999682 0.7995681 0.57282364 -0.16135503
-0.46044117 -0.2598328 0.9866536 0.49479905 -0.46057963 -1.0200845
-1.4749124 1.0637975 1.3365238 0.58921224 -0.1014818 0.25290093
0.3082669 0.5153926 0.39087212 0.3884157 -0.08448967 0.6173587
-0.7016996 -0.00803287 0.92532045 1.1567336 0.9304369 -0.48769325
0.02949985 0.5733254 0.47348273 0.29140505 -0.08230156 -0.63224345
0.45760217 1.3972194 -0.3903019 -1.8784692 -0.3232209 -0.6622387
-1.3558656 -0.49081594 0.08895037 0.14497757 -0.5133033 1.7450047
0.39720926 0.26907772 -0.17198385 0.9587774 0.97333735 0.40279716
-0.49760363 0.16291164 0.98604476 -0.93996775 -0.23554315 -0.16560198
0.32105902 -0.19337915 0.59740543 -0.31641582 -0.521531 -0.05542891
-0.79457045 0.24275288 -0.31727335 -0.53163546 0.7289991 0.21987322
-1.4259332 0.40251678 -0.56112754 -0.89991057 0.5854356 0.391536
-0.5190141 -0.40848324 -1.1180537 0.18540625 0.40589765 0.09343605
-0.82651746 -0.39343643 -0.7751122 0.51498145 0.8831598 -0.75804406
0.421243 -0.7835073 -0.64740705 0.886877 -0.12685381 -0.09480596
0.08138569 0.5666009 -0.5396669 0.44753233 0.35471565 0.5448529
0.2747287 -1.2296329 -0.27783546 -0.22968143 0.5378947 0.05512628
-0.45419472 -0.16735198 -0.6784677 -0.3905516 0.48485395 -1.0704969
-0.37959948 0.3049939 -0.79768074 -0.34404752 0.9056979 0.07272289
1.4017062 -1.8775489 0.5494713 0.8134191 1.2033825 -0.12012713
-0.48290002 0.9488986 0.04522356 0.57258 -0.214445 0.11267034
-0.18030205 0.33719262 0.0723532 0.4280377 0.3034751 1.0643102
-1.3477277 -0.54682714 -0.2947725 0.18933316 -0.7651717 -0.04464987
-0.28072134 0.23950963 -0.6439076 0.58465445 0.48399085 -1.220535
0.8479436 -0.13057162 0.19621813 -0.0656259 -1.2448994 1.4145195
0.16885793 0.7779996 0.40739048 -1.2086812 0.81609976 0.24173617
0.7545517 -0.27519926 -0.3227762 0.06531979 0.30798247 -0.37161058
0.08414892 0.28916767 0.09701464 0.6712656 0.01069349 0.29941294
0.4359536 0.9780947 1.6060786 -0.84467787 0.12446821 -0.37364954
0.3500561 -0.1847585 0.68080425 0.5608428 -0.19524248 0.51820004
1.0251828 -0.36085963 -0.8557542 -1.0205867 0.2647184 0.7871021
0.5847969 -0.63073987 -0.22540942 -0.6195089 0.27887833 -0.49188766
-0.572529 -0.15872152 -0.33558896 -0.48167813 0.22728525 0.15462105
0.29023996 -0.9723555 0.15242368 0.18220896 -0.41597047 -1.1473626
-0.9753822 -0.3675931 -0.8565682 -1.8035975 -0.4456907 -1.132391
0.96532726 0.8027674 1.34667 0.7975533 -0.147868 0.4941276
-0.31117883 0.2827206 -0.08960844 0.38768238 -0.01199331 0.45319277
0.23430137 -0.9569217 -0.62397337 0.36553183 -1.0118973 -0.1315518
0.6142734 0.81990767 0.39275485 0.21428907 0.59229916 -0.9834198
0.8235861 -0.8849138 -0.49274087 0.40865242 -0.63678145 0.02790589
0.25768843 -0.41062093 -0.13625638 -0.27997482 0.72709924 -0.6132534
0.3556586 1.1060569 -0.33845586 -0.25888893 0.4800494 0.15908955
0.15485884 -0.26025015 0.30296478 0.2486798 0.08974958 -0.4424449
1.0174154 0.5244763 0.20439535 -0.7359444 -0.55245817 -0.9026899
-0.66682255 -0.27394095 0.5138405 -0.8235222 -0.8855535 0.2145728
-1.0064538 -0.03400371 0.15795116 0.09281004 -0.02162714 0.51808214
-0.7613919 -0.363577 -0.182741 -0.8874818 0.7371926 0.16487879
0.11636213 -1.3065263 0.19836414 0.08029842 0.48266783 0.4633029
1.0067593 -0.62923443 -1.3402416 -0.07063743 -0.6040924 -0.21591265
0.4203254 0.11627919 -0.355772 -0.6868035 -0.9401848 -0.06368721
0.94324267 0.09775743 0.9022457 -0.5618411 -0.8919936 0.52507746
1.5713065 -0.3229949 0.3112879 -0.10402933 0.9638713 0.5948891
-0.14995001 0.13676347 0.60331815 0.33172426 0.86889267 -0.17292877
0.2541669 -0.4079158 -0.02663792 1.0244169 0.00557532 -0.499292
-0.97209907 0.8226652 -2.0382798 -1.3012456 -0.10435171 1.822306
0.78787667 0.16665798 0.2822044 -0.29692313 1.1738418 0.4416274
-0.6864321 0.3047923 -0.41930386 -0.2407211 0.20998882 0.47586125
-0.8890483 0.727665 5.497443 0.33431366 -0.9828703 -0.15744981
0.39758244 0.07969208 -0.7548698 0.27081427 -0.23615067 0.39758325
0.3425254 -0.50780016 0.7593609 0.5242691 -0.00880301 0.43080366
-1.172035 0.7919671 -0.2147753 -1.7716215 0.17679007 0.44758236
1.0246323 0.31941643 -0.24616629 0.01313124 0.591778 -1.0718215
0.31403825 0.5993175 0.8219688 -0.46052352 0.53726524 0.45573616
-1.7536267 -0.07458686 -0.42907527 0.17250332 -0.12425668 0.71355045
-0.84853905 0.75176156 0.71177334 1.2867906 -0.69649434 1.36986
0.05566022 0.5705655 -0.48859254 -0.24582058 0.3081586 -0.47490498
0.8060892 1.018056 0.14717917 -0.07139756 0.5484594 1.2802505
-0.5152491 -0.14864948 -0.90262496 -0.6164668 0.8926129 1.2233802
-0.98746175 0.01602997 -0.48974544 0.75056463 0.66417474 0.6092493
-0.3189469 -0.39382064 0.7552321 0.52265775 0.39710113 -0.20485143
0.30975306 -1.3750991 0.05401988 -0.86707306 0.55629957 -0.45680702
-1.7839066 0.48667914 -0.1964834 -1.1268766 0.3523253 -0.23081724
-0.9481748 0.67799145 -1.2164333 -1.2389566 -0.4679099 0.61708623
-0.07589219 -0.05851879 0.6057236 0.25488052 -0.50647694 0.36851865
0.2517377 0.6163675 0.4121464 -1.2614652 0.39673504 0.7953669
0.27293572 0.9416222 0.18933277 -0.8038448 -1.3264097 -1.2513043
0.99980783 -0.18458183 0.8769469 -0.5721976 -1.0146068 0.69880205
0.16666926 0.45802987 0.5380084 0.3171991 -0.6153687 -0.3353766
-0.8677955 0.5251448 1.4955761 -0.81140065 -0.01005511 0.4601403
1.0464925 0.09903453 -1.0217694 0.15457535 0.2759648 -0.8177828
0.9737063 -0.80263233 0.4222945 -0.4275086 0.05391327 -1.4925159
-0.8003473 -0.5949404 -0.09914225 1.0942967 0.36796448 -0.84277344
0.8895471 0.09976256 0.28163093 -0.87965393 -0.8484132 -0.5383341
-0.24778336 0.1923514 0.911661 1.3907431 0.02399877 0.6964308
-0.00805226 0.47167346 0.84887743 0.5611928 0.65086544 -1.7712281
-0.15493819 -0.7415439 -0.72890466 -1.0579919 -0.08338255 -1.2292631
-0.41257748 -1.9957482 0.5766403 -0.8344439 -0.23978011 0.24640438
-0.15081772 -0.10892022 0.14020039 0.35418943 -1.0594563 0.5354896
1.4989504 -0.6099492 -0.00811817 -0.04995093 -0.86307704 0.25942492
0.44940838 -0.46026018 -0.577982 -0.35129723 0.7368453 0.08105407
0.55363107 -0.70344454 0.6329752 -0.34022364 -0.07907383 -0.4268539
0.03391504 -0.9390387 0.43838143 0.4677722 -0.36822906 0.03119704
-0.5142919 1.3669739 -0.33473328 0.30349582 0.39790893 -0.25930893
-0.514592 0.99246985 0.01719027 0.21875523 0.95224565 -0.19740766
-0.48580804 -0.58000535 -0.79150975 0.98898095 0.70905155 0.47815993
0.720765 -1.5106269 -0.86943126 0.10750041 0.44664496 0.3850881
0.01467309 0.7252156 -0.44247976 0.2740811 0.1240323 -0.85355407
-1.1898123 0.71286374 0.5368178 -0.43688807 -0.5274603 0.7486155
0.34155127 -0.7541447 0.0913085 -0.11731732 -0.314965 -0.10297891
-0.0125276 0.14159675 -0.4565398 -0.6830841 -0.4534881 0.43519762
0.11983889 0.3658021 1.4263604 -0.2046718 -0.7213908 0.32730538
1.1891632 -0.11238917 -0.845376 -0.9811408 0.26467404 -0.5568213
-0.2637455 -0.25056222 -1.6631838 0.7443527 -0.0079508 0.7651648
0.79630864 0.507231 0.53992814 0.615743 0.43872657 -0.50238013
0.48681265 0.40940318 0.92109776 -1.2569672 -0.10796864 -0.80742997
-0.26387745 1.0598336 0.6932644 -0.14721952 1.051148 -0.2641575
-0.57713413 -0.78374165 -0.9577061 -0.27605245 0.40575677 0.65903026
0.06046319 0.2631271 0.1979924 0.35894448 0.27747083 -0.37861717
0.51933724 0.7428917 -0.50629383 -1.0356956 0.06786926 0.7962143
0.03938549 -0.10634877 -0.9472163 0.66609347 -0.10142895 1.0181766
0.09961119 -0.6194857 0.1722413 -0.45739305 0.10613912 -0.9971602
-0.4183515 -0.11840801 -0.02670298 -0.5989502 -0.6316101 -0.44813287
-1.0607457 -0.7839388 -0.4253195 0.23461547 -0.2807683 0.6068945
0.7064104 0.3981615 0.8783165 -0.43456396 -0.10923489 -0.7421887
-0.7659425 0.60109437 0.540845 -0.5536127 -0.3901285 -0.19821724] | [7.212950706481934, 6.024726390838623] |
9ea1ed1f-27d3-4f2e-8954-7b5e04d69a6c | contrastive-losses-are-natural-criteria-for | 2211.10056 | null | https://arxiv.org/abs/2211.10056v1 | https://arxiv.org/pdf/2211.10056v1.pdf | Contrastive Losses Are Natural Criteria for Unsupervised Video Summarization | Video summarization aims to select the most informative subset of frames in a video to facilitate efficient video browsing. Unsupervised methods usually rely on heuristic training objectives such as diversity and representativeness. However, such methods need to bootstrap the online-generated summaries to compute the objectives for importance score regression. We consider such a pipeline inefficient and seek to directly quantify the frame-level importance with the help of contrastive losses in the representation learning literature. Leveraging the contrastive losses, we propose three metrics featuring a desirable key frame: local dissimilarity, global consistency, and uniqueness. With features pre-trained on the image classification task, the metrics can already yield high-quality importance scores, demonstrating competitive or better performance than past heavily-trained methods. We show that by refining the pre-trained features with a lightweight contrastively learned projection module, the frame-level importance scores can be further improved, and the model can also leverage a large number of random videos and generalize to test videos with decent performance. Code available at https://github.com/pangzss/pytorch-CTVSUM. | ['Hajime Nagahara', 'Mayu Otani', 'Yuta Nakashima', 'Zongshang Pang'] | 2022-11-18 | null | null | null | null | ['unsupervised-video-summarization'] | ['computer-vision'] | [ 2.94646740e-01 -1.26754031e-01 -5.14111817e-01 -3.45205933e-01
-1.18967772e+00 -4.38340485e-01 5.86389303e-01 4.32188585e-02
-2.52006412e-01 6.45349681e-01 6.86235607e-01 1.91812977e-01
-9.48354229e-02 -3.09167773e-01 -7.88197756e-01 -5.69696665e-01
-3.56419683e-01 1.30231455e-01 3.40871364e-02 1.19619787e-01
4.26711440e-01 2.21037269e-01 -1.93385470e+00 4.16515738e-01
7.09929883e-01 1.13279104e+00 1.38964608e-01 8.70037138e-01
3.79422516e-01 9.88166690e-01 -2.33666509e-01 -2.55391568e-01
4.14878488e-01 -6.28921270e-01 -7.47699261e-01 1.88069224e-01
7.76946127e-01 -7.56451607e-01 -4.32867110e-01 9.18413281e-01
4.59513277e-01 5.34956634e-01 5.66970885e-01 -1.05365312e+00
-3.58359367e-01 4.53544885e-01 -4.81964022e-01 3.76030356e-01
6.13445103e-01 3.69632781e-01 1.62005639e+00 -9.98548865e-01
8.28678071e-01 1.03266823e+00 7.20093727e-01 3.22649062e-01
-1.24810016e+00 -4.54343021e-01 1.46791741e-01 5.22206962e-01
-1.11521518e+00 -7.70598650e-01 6.69540763e-01 -3.44026208e-01
6.68121994e-01 3.51895094e-01 7.01887429e-01 1.09755528e+00
-2.40232095e-01 9.27122951e-01 6.01908505e-01 -1.30030945e-01
1.79573834e-01 -1.13623552e-01 -3.54376389e-03 9.09610033e-01
2.47959957e-01 -3.90609093e-02 -9.03083444e-01 -7.04860687e-02
6.06191933e-01 9.03529823e-02 -6.45893276e-01 -5.28031826e-01
-1.20105231e+00 7.97299147e-01 4.69108194e-01 -5.51829897e-02
-4.55308825e-01 2.76392549e-01 5.38413346e-01 2.78459042e-01
4.98077184e-01 6.56109631e-01 -3.77270401e-01 -5.30548573e-01
-1.41794145e+00 4.12879407e-01 6.41412139e-01 8.32734048e-01
8.28224003e-01 -1.29164591e-01 -4.32041168e-01 7.66437650e-01
-5.93639761e-02 2.79539198e-01 2.97759145e-01 -1.57077897e+00
2.93359399e-01 2.38068059e-01 7.61699770e-03 -9.75557804e-01
1.41110048e-02 -4.55547780e-01 -4.47367638e-01 1.84189335e-01
3.82100582e-01 3.32084708e-02 -7.15642214e-01 1.83740747e+00
2.13867471e-01 3.07745337e-01 -3.56776327e-01 1.03878057e+00
4.02776331e-01 6.82842016e-01 -6.35173991e-02 -3.17366689e-01
1.03377056e+00 -1.15604866e+00 -2.23555893e-01 1.48996159e-01
4.42377925e-01 -5.98441362e-01 1.31364143e+00 4.78981197e-01
-1.50870502e+00 -4.44065094e-01 -1.03629279e+00 -1.94974542e-01
3.13537896e-01 4.45012487e-02 5.03794849e-01 2.75765359e-01
-1.12192690e+00 1.01989114e+00 -9.11768198e-01 -3.28346491e-01
7.76725411e-01 2.19198987e-01 -3.05223376e-01 -2.53264576e-01
-6.21720850e-01 5.84381759e-01 1.50119692e-01 -3.78298700e-01
-1.13417614e+00 -9.98195827e-01 -5.99160671e-01 2.39965796e-01
4.93164450e-01 -9.07043338e-01 1.18324792e+00 -1.14706373e+00
-1.30295002e+00 5.63918471e-01 -2.20408052e-01 -4.71223056e-01
7.01154590e-01 -4.92486835e-01 2.04290822e-01 8.40703726e-01
1.25492990e-01 8.39188516e-01 1.07454050e+00 -9.67620313e-01
-7.42134213e-01 5.85631169e-02 1.70807734e-01 3.37853283e-01
-5.40022075e-01 -1.62256375e-01 -7.29689896e-01 -8.57670009e-01
-1.91916287e-01 -8.33644927e-01 -4.48776409e-02 3.30638349e-01
-1.76265851e-01 1.55047970e-02 4.66881067e-01 -8.49529266e-01
1.23187387e+00 -2.16887164e+00 3.57099980e-01 1.36915967e-01
4.48671073e-01 6.64307252e-02 -3.38937044e-01 2.39877328e-01
1.11252770e-01 -2.76398305e-02 -2.98268646e-01 -2.94583738e-01
-5.63842766e-02 -2.77485996e-01 -1.61784828e-01 5.02497077e-01
4.34197396e-01 7.53923655e-01 -1.13695776e+00 -4.86615688e-01
1.83474556e-01 4.79921907e-01 -1.33298457e+00 2.87269890e-01
-2.29250461e-01 3.33044946e-01 -3.42887491e-01 6.28316402e-01
1.75741762e-01 -5.97477615e-01 1.73347965e-01 -4.66905773e-01
2.27741271e-01 4.22078758e-01 -9.89174604e-01 1.84279883e+00
-3.16821814e-01 8.75238776e-01 -8.69707465e-02 -1.08799410e+00
4.35191453e-01 -4.53483351e-02 7.73461938e-01 -2.81197965e-01
-3.19567919e-02 6.65767640e-02 -2.60596424e-01 -4.10452873e-01
6.42846107e-01 2.64975786e-01 3.12760800e-01 5.36553085e-01
2.94468254e-01 1.10226639e-01 3.50294024e-01 4.21804070e-01
1.44659019e+00 4.93840098e-01 3.14859152e-01 -2.16975436e-01
2.48750925e-01 -1.08805649e-01 5.02971232e-01 6.89306676e-01
-1.42111689e-01 1.01440549e+00 5.49153149e-01 -2.97588885e-01
-1.38480580e+00 -1.06577563e+00 -1.57365333e-02 1.37507594e+00
5.45545295e-02 -7.68819869e-01 -6.21497631e-01 -6.77187622e-01
1.67324618e-02 4.68401760e-01 -4.65234727e-01 -2.74008691e-01
-5.15195251e-01 -3.43688488e-01 2.17119992e-01 4.01043206e-01
1.64765388e-01 -7.74426460e-01 -6.63033783e-01 7.42691532e-02
-2.52491504e-01 -7.79665887e-01 -6.88207030e-01 -1.16305821e-01
-8.34810734e-01 -1.01755297e+00 -9.17600393e-01 -3.55637908e-01
6.57002091e-01 3.87556046e-01 1.25106573e+00 3.70300382e-01
-3.00824523e-01 5.82742095e-01 -5.29986918e-01 2.95957655e-01
-9.06677768e-02 1.65927872e-01 7.42810071e-02 -1.63659547e-02
4.95241815e-03 -9.39060688e-01 -1.03793514e+00 1.07508689e-01
-7.51148820e-01 1.18602753e-01 5.11980355e-01 9.56796706e-01
6.29440427e-01 -3.46108556e-01 4.62136954e-01 -7.34683633e-01
3.26513886e-01 -6.15394473e-01 -2.21826598e-01 8.82419646e-02
-6.26355886e-01 2.20480874e-01 6.84075058e-01 -4.53258693e-01
-6.47448599e-01 -2.05211639e-01 4.01340201e-02 -9.27027106e-01
2.56273746e-01 4.40601736e-01 3.34819295e-02 1.17955096e-01
6.48165226e-01 2.50642121e-01 1.43165737e-01 -2.70948023e-01
6.07358336e-01 1.71144888e-01 5.49603641e-01 -7.57578671e-01
6.80969059e-01 4.14777488e-01 -1.87502325e-01 -6.76922560e-01
-8.15482676e-01 -4.10913110e-01 -3.08174610e-01 -3.52778643e-01
3.14393967e-01 -1.10867405e+00 -5.26497722e-01 -4.30210866e-02
-7.09454596e-01 -3.07613701e-01 -5.47978342e-01 5.83209276e-01
-8.27999294e-01 5.90992332e-01 -5.67297101e-01 -5.67082763e-01
-3.16574305e-01 -1.06340075e+00 1.08132720e+00 1.22327551e-01
-3.92193586e-01 -6.85982823e-01 -2.79213265e-02 2.80167669e-01
3.39376122e-01 1.25650704e-01 6.32950604e-01 -5.30467153e-01
-8.79811049e-01 -1.62045971e-01 -1.61753133e-01 5.72012901e-01
1.93337891e-02 2.60976970e-01 -9.31738019e-01 -4.55404758e-01
-2.91115850e-01 -5.57056487e-01 1.37867737e+00 5.66940665e-01
1.39719081e+00 -6.17731571e-01 3.26122455e-02 9.11125064e-01
1.23539734e+00 -3.95603716e-01 4.47412997e-01 1.24718785e-01
6.66637957e-01 4.34081972e-01 6.13616168e-01 8.00908148e-01
3.64849865e-01 6.05268896e-01 2.38616347e-01 3.55066508e-01
-3.55692357e-01 -2.81494856e-01 6.23680592e-01 7.82135963e-01
-3.34032238e-01 -1.00880720e-01 -4.88461345e-01 5.33296466e-01
-1.89925969e+00 -1.40591168e+00 5.38982093e-01 2.29364419e+00
9.42835450e-01 2.25192919e-01 4.42124128e-01 -7.34832138e-02
6.55773342e-01 4.71421421e-01 -8.04452479e-01 -5.83289284e-03
2.30137948e-02 8.29711854e-02 4.32916969e-01 4.57644224e-01
-1.10518312e+00 6.47329926e-01 6.25759840e+00 8.68600130e-01
-1.02253509e+00 -7.80101866e-03 8.28065395e-01 -8.44348967e-01
-6.09901547e-01 8.87659341e-02 -4.40914094e-01 6.86408341e-01
9.90840673e-01 -3.65187198e-01 7.25374639e-01 8.85307133e-01
2.51603931e-01 -4.06198716e-03 -1.36701453e+00 1.02330911e+00
1.32936105e-01 -1.63675559e+00 2.45502591e-01 3.83356586e-02
8.58717144e-01 7.73986951e-02 2.21923023e-01 2.44302779e-01
1.46550745e-01 -8.51913571e-01 8.31729174e-01 7.46631026e-01
8.12638283e-01 -7.32372761e-01 3.15668374e-01 -1.07078470e-01
-1.07874918e+00 -2.15806067e-01 -4.22682554e-01 1.19275786e-02
1.57649428e-01 6.82276249e-01 -5.35724163e-01 3.99587631e-01
6.28694475e-01 1.23118508e+00 -5.31975985e-01 1.16738486e+00
-1.26074895e-01 8.24570060e-01 -3.94174129e-01 8.23142231e-02
1.65503040e-01 -1.97405323e-01 7.29388952e-01 1.17984128e+00
4.87165809e-01 -5.40538542e-02 1.12163760e-01 6.83221638e-01
-4.03321862e-01 9.95304510e-02 -3.88738394e-01 -1.48639172e-01
6.13076448e-01 1.22754490e+00 -4.22687441e-01 -5.15097201e-01
-4.12358969e-01 9.82697070e-01 3.99562627e-01 3.46635163e-01
-9.01580274e-01 -1.46115988e-01 9.27835822e-01 2.74589688e-01
6.46762311e-01 -1.00450911e-01 -1.42004073e-01 -1.49375057e+00
2.19443709e-01 -1.16922688e+00 3.92340750e-01 -6.51111662e-01
-1.19649041e+00 3.78468573e-01 -1.80204455e-02 -1.46807349e+00
-4.71777469e-01 -3.77132893e-01 -6.66291356e-01 4.02124971e-01
-1.39950240e+00 -9.61282432e-01 -3.96053970e-01 4.28369075e-01
6.96164489e-01 -2.12509304e-01 3.06681454e-01 2.79733866e-01
-6.53103650e-01 7.84543037e-01 9.67799202e-02 -1.82236254e-01
8.39425445e-01 -1.13388777e+00 1.79548234e-01 9.17526484e-01
2.15350866e-01 5.20827353e-01 8.77310216e-01 -2.85218745e-01
-1.43314755e+00 -9.47752893e-01 3.48806083e-01 -4.20845151e-01
6.43238664e-01 -1.31000869e-03 -7.40631104e-01 5.84472775e-01
8.03867448e-03 1.95523411e-01 5.98578215e-01 1.86852917e-01
-6.00622594e-01 -1.85133040e-01 -8.30340147e-01 7.41698265e-01
1.39041674e+00 -6.06776059e-01 -4.42178011e-01 3.03027362e-01
6.53432131e-01 -1.08187608e-01 -8.61302853e-01 1.94057405e-01
8.44655752e-01 -1.19820619e+00 1.06962407e+00 -6.72682643e-01
9.36205268e-01 -1.77872211e-01 -3.80450010e-01 -1.00012720e+00
-3.54489237e-01 -7.96700537e-01 -4.61361796e-01 1.07543314e+00
3.18769485e-01 -2.41676196e-01 8.05476487e-01 5.67967534e-01
-1.73774347e-01 -8.70530665e-01 -6.87583506e-01 -6.88616693e-01
-2.89242864e-01 -3.28400671e-01 3.43446910e-01 6.09855771e-01
1.45377945e-02 2.47578874e-01 -6.21039450e-01 -7.21568093e-02
8.11973333e-01 1.04324006e-01 8.12618852e-01 -1.04273784e+00
-5.52558541e-01 -7.23697305e-01 -5.20513117e-01 -1.26241004e+00
6.27610832e-02 -9.17371511e-01 -6.50241151e-02 -1.24506736e+00
5.74154019e-01 -1.81126475e-01 -5.94708383e-01 2.83708781e-01
-3.86476845e-01 2.83977717e-01 3.91990095e-01 4.88479286e-01
-1.15829229e+00 6.89464390e-01 1.03354204e+00 -2.48007998e-02
-7.61381090e-02 -1.78874195e-01 -8.49651158e-01 5.93096018e-01
6.87911272e-01 -2.82825112e-01 -5.34725487e-01 -2.98031718e-01
1.46873310e-01 -1.28226265e-01 4.19725180e-01 -1.19371116e+00
1.72667116e-01 -1.25934571e-01 4.73164082e-01 -2.91031957e-01
4.56507862e-01 -3.30097795e-01 -2.53114663e-02 2.77307898e-01
-7.74660051e-01 -7.74432346e-02 -1.80522054e-01 6.87520146e-01
-1.38075262e-01 -1.60739198e-01 7.84994304e-01 -2.94581987e-02
-7.55649209e-01 4.50434744e-01 -7.03664497e-03 2.34706849e-01
8.12041283e-01 -3.70333433e-01 -1.40693471e-01 -7.01276958e-01
-4.10498023e-01 4.90848236e-02 8.40434372e-01 1.35560662e-01
7.52573669e-01 -1.35165119e+00 -8.41660798e-01 8.90354663e-02
3.91301326e-02 -2.28714615e-01 2.52734393e-01 9.87140536e-01
-5.17317295e-01 -3.27799954e-02 -3.75368029e-01 -6.31853819e-01
-9.87664998e-01 2.38240808e-01 4.42922954e-03 -1.97437391e-01
-7.61537790e-01 9.54903066e-01 1.99685723e-01 1.54501900e-01
2.89246947e-01 -1.42844826e-01 9.16035622e-02 1.99340448e-01
6.90844297e-01 5.16306996e-01 -1.30893171e-01 -3.86644065e-01
-3.31640780e-01 3.43054116e-01 -1.59082934e-01 -1.19434409e-01
1.66845906e+00 -2.40168691e-01 3.02758276e-01 1.03898972e-01
1.53341019e+00 1.00496680e-01 -1.94904578e+00 -1.35459065e-01
-4.04030830e-02 -8.00656796e-01 1.31423414e-01 -4.35742289e-01
-1.20111811e+00 5.82850397e-01 3.45290363e-01 -1.06474422e-01
1.27204037e+00 -6.62586838e-02 7.42814839e-01 4.29629415e-01
2.45931998e-01 -9.66831326e-01 4.77356881e-01 1.39091089e-01
8.50811124e-01 -1.10700095e+00 3.39977473e-01 -1.78970248e-02
-7.99342930e-01 1.08425450e+00 2.27132425e-01 -4.71150190e-01
5.16928315e-01 4.04396728e-02 -4.56254363e-01 1.00350613e-02
-1.18772650e+00 -1.78889081e-01 4.82419342e-01 4.17496532e-01
4.65142757e-01 -1.18913114e-01 -8.86370987e-02 3.42042625e-01
-1.76252082e-01 -2.77252775e-02 5.32705843e-01 7.72300184e-01
-6.12175643e-01 -8.03970158e-01 7.31803402e-02 9.74486887e-01
-4.60904270e-01 -2.90839761e-01 1.06605105e-01 4.10910845e-01
-3.67392153e-01 6.55455828e-01 8.49437043e-02 -5.00327528e-01
9.01757702e-02 -1.66019723e-01 5.33217132e-01 -4.64478552e-01
-2.40469381e-01 3.66527252e-02 1.65288761e-01 -1.05480242e+00
-5.57352960e-01 -8.71776402e-01 -7.77058899e-01 -6.90370381e-01
-8.16263407e-02 3.87958065e-03 2.42938444e-01 6.38286412e-01
5.85360765e-01 1.84191450e-01 8.11789334e-01 -1.27970326e+00
-7.34493434e-01 -7.12943017e-01 -3.28305721e-01 6.77153587e-01
5.20893693e-01 -6.71864569e-01 -5.69162250e-01 3.48384798e-01] | [10.201072692871094, 0.4518008232116699] |
29f4d872-36f1-4c0f-b89f-52671f709e6a | a-question-entailment-approach-to-question | 1901.08079 | null | http://arxiv.org/abs/1901.08079v1 | http://arxiv.org/pdf/1901.08079v1.pdf | A Question-Entailment Approach to Question Answering | One of the challenges in large-scale information retrieval (IR) is to develop
fine-grained and domain-specific methods to answer natural language questions.
Despite the availability of numerous sources and datasets for answer retrieval,
Question Answering (QA) remains a challenging problem due to the difficulty of
the question understanding and answer extraction tasks. One of the promising
tracks investigated in QA is to map new questions to formerly answered
questions that are `similar'. In this paper, we propose a novel QA approach
based on Recognizing Question Entailment (RQE) and we describe the QA system
and resources that we built and evaluated on real medical questions. First, we
compare machine learning and deep learning methods for RQE using different
kinds of datasets, including textual inference, question similarity and
entailment in both the open and clinical domains. Second, we combine IR models
with the best RQE method to select entailed questions and rank the retrieved
answers. To study the end-to-end QA approach, we built the MedQuAD collection
of 47,457 question-answer pairs from trusted medical sources, that we introduce
and share in the scope of this paper. Following the evaluation process used in
TREC 2017 LiveQA, we find that our approach exceeds the best results of the
medical task with a 29.8% increase over the best official score. The evaluation
results also support the relevance of question entailment for QA and highlight
the effectiveness of combining IR and RQE for future QA efforts. Our findings
also show that relying on a restricted set of reliable answer sources can bring
a substantial improvement in medical QA. | ['Dina Demner-Fushman', 'Asma Ben Abacha'] | 2019-01-23 | null | null | null | null | ['question-similarity'] | ['natural-language-processing'] | [ 2.86638170e-01 2.78894603e-01 5.19070029e-02 -4.05313492e-01
-1.93783152e+00 -7.45761216e-01 5.12887359e-01 6.85937226e-01
-6.80120647e-01 6.80587649e-01 7.18084276e-01 -3.07365775e-01
-7.32060552e-01 -5.71273506e-01 -5.55961967e-01 -1.06945552e-01
2.62087464e-01 9.66261566e-01 3.22299808e-01 -7.06332028e-01
3.48994195e-01 -9.10157990e-03 -1.30744493e+00 1.15329969e+00
1.06575716e+00 1.19298327e+00 -2.22928613e-01 8.09719443e-01
-1.80002138e-01 1.19421268e+00 -6.38712168e-01 -7.38003075e-01
-1.42131239e-01 -4.72629160e-01 -1.78860855e+00 -6.14363730e-01
6.20473444e-01 -1.88797295e-01 -1.21557161e-01 4.86954868e-01
9.23222721e-01 -6.35587424e-02 5.44880867e-01 -6.55582070e-01
-7.28315711e-01 2.39998102e-01 1.57562792e-01 4.47504133e-01
1.07828903e+00 -9.63938609e-02 1.24828506e+00 -7.49788225e-01
9.48562741e-01 1.04204714e+00 6.01348102e-01 6.38965964e-01
-8.47311318e-01 -1.91399992e-01 -6.20097995e-01 5.13451993e-01
-1.13741338e+00 -4.81717020e-01 3.20665032e-01 -1.14694059e-01
1.10020125e+00 6.34132981e-01 -2.42136359e-01 6.97484851e-01
2.48564720e-01 8.02028835e-01 1.15399933e+00 -6.18487716e-01
2.47177601e-01 1.39172047e-01 5.21210313e-01 6.10788107e-01
-1.77293867e-01 -2.48631746e-01 -2.37757415e-01 -6.89018607e-01
-2.27963686e-01 -3.42377573e-01 -3.86197746e-01 2.23686650e-01
-1.16173720e+00 1.07715511e+00 5.54297328e-01 4.66214746e-01
-4.79410082e-01 -2.98574984e-01 6.04162693e-01 6.23930335e-01
4.27603066e-01 1.15133321e+00 -8.13673794e-01 7.13215470e-02
-9.28870976e-01 5.22936463e-01 1.19288850e+00 5.12127101e-01
3.58841389e-01 -1.10542917e+00 -9.59499657e-01 8.62243474e-01
1.50328472e-01 5.30848145e-01 4.02823597e-01 -1.18240893e+00
5.68697214e-01 8.50784004e-01 1.03874162e-01 -9.82718647e-01
-5.69728017e-01 -3.33513647e-01 -5.12500703e-01 -5.74438870e-01
4.87356335e-01 7.83804711e-03 -6.76538467e-01 1.38925910e+00
3.57567310e-01 -3.55436236e-01 2.82812834e-01 7.26191998e-01
1.60641730e+00 4.95007336e-01 1.03179656e-01 1.72743946e-03
2.05575705e+00 -7.93005824e-01 -9.17440534e-01 -1.26876682e-02
9.13623989e-01 -1.06921434e+00 1.03001666e+00 2.02375561e-01
-1.11966205e+00 -3.80117774e-01 -6.06942773e-01 -4.70179856e-01
-3.39305103e-01 1.72894284e-01 2.52197683e-01 4.00568008e-01
-1.14126563e+00 2.79035509e-01 -2.61915028e-01 -3.28742474e-01
3.00197482e-01 2.04887152e-01 -2.99410969e-01 -5.98907292e-01
-1.71889400e+00 1.29136431e+00 9.68675613e-02 -8.22749510e-02
-7.54007936e-01 -8.06167245e-01 -6.34901345e-01 5.75764850e-02
4.42431152e-01 -1.03149128e+00 1.51705933e+00 -4.87266362e-01
-9.57277656e-01 1.13421214e+00 -3.27685684e-01 -5.47335446e-01
2.12352902e-01 -3.85977358e-01 -4.73472416e-01 8.39385688e-01
2.94915587e-01 5.94024003e-01 2.35606223e-01 -8.19542706e-01
-4.24428523e-01 -3.76145333e-01 4.25123513e-01 1.53570801e-01
-7.64579102e-02 3.79336387e-01 -3.77678931e-01 -2.75563356e-02
-9.89966542e-02 -7.30993569e-01 -2.08948776e-01 -2.58808464e-01
-2.99145013e-01 -7.81322956e-01 6.76229084e-03 -1.11126578e+00
1.46347940e+00 -1.65604091e+00 -1.29503369e-01 1.72011796e-02
3.74713182e-01 3.45945656e-01 -2.22101659e-01 8.40166211e-01
3.87679525e-02 1.02841988e-01 -2.77377605e-01 3.95016366e-04
2.71441303e-02 1.24026284e-01 -4.19134289e-01 1.62871592e-02
4.54470336e-01 1.23882329e+00 -9.56064582e-01 -9.21606064e-01
-3.57525200e-01 2.17755884e-01 -5.77887833e-01 4.19183493e-01
-4.79740798e-01 4.23431307e-01 -6.78012729e-01 7.65514731e-01
3.81501615e-01 -5.93157411e-01 -8.81227776e-02 -4.03494835e-01
5.20928085e-01 8.79695594e-01 -6.22589707e-01 1.82471144e+00
-4.53380913e-01 2.88803011e-01 -1.16875872e-01 -7.68102944e-01
7.23398566e-01 7.48917460e-01 3.93426716e-01 -1.21745181e+00
-3.45228687e-02 5.54560304e-01 -1.16464674e-01 -1.03230238e+00
4.93692100e-01 -1.82178125e-01 -2.67661572e-01 5.24138570e-01
1.06184818e-01 -1.88571021e-01 2.72274286e-01 5.34743130e-01
1.47371900e+00 -1.64375305e-01 3.79689693e-01 -2.99822897e-01
9.61334825e-01 3.94533306e-01 -2.66023763e-02 8.57153535e-01
-1.87348187e-01 6.52898192e-01 4.13888752e-01 -2.88803399e-01
-4.67866510e-01 -8.61848831e-01 -4.37383175e-01 1.26558316e+00
-3.48864555e-01 -3.28490585e-01 -7.14355648e-01 -9.43650901e-01
-1.69483066e-01 6.26548290e-01 -6.76878273e-01 -1.92412660e-01
-6.72001362e-01 -4.73477662e-01 8.60830665e-01 5.40589774e-03
2.16562465e-01 -1.23794234e+00 -5.34687936e-01 1.51582196e-01
-1.01340771e+00 -1.07179081e+00 -4.65116084e-01 -1.91774979e-01
-8.24201107e-01 -1.35169518e+00 -9.43686008e-01 -7.06983864e-01
2.44056940e-01 -1.36049077e-01 1.83074915e+00 3.05884719e-01
-2.49665827e-01 8.28445256e-01 -6.89011753e-01 -3.35702032e-01
-6.21302605e-01 4.11614895e-01 -5.72748721e-01 -3.68886024e-01
7.05325723e-01 2.09921449e-01 -1.11022890e+00 2.65520394e-01
-1.16011441e+00 -5.73024631e-01 6.19639099e-01 7.38432944e-01
4.89724338e-01 -6.36883676e-01 1.06397974e+00 -9.24758315e-01
1.21845603e+00 -6.57149315e-01 -3.12629901e-02 8.18597615e-01
-7.49547124e-01 3.51019084e-01 2.02864602e-01 1.50043100e-01
-9.80947495e-01 -3.95822585e-01 -7.63572812e-01 3.36096346e-01
-1.40916735e-01 8.08176994e-01 2.37217426e-01 1.87670812e-01
1.18908751e+00 -8.83038342e-03 -1.35718556e-02 -4.97646749e-01
5.41332245e-01 7.86112905e-01 3.19894075e-01 -6.33864641e-01
2.15600818e-01 3.17362875e-01 -1.53401226e-01 -2.60611624e-01
-1.39605391e+00 -8.65178704e-01 -2.40276575e-01 3.88428569e-03
1.02312505e+00 -6.70371652e-01 -9.32824492e-01 -3.28756213e-01
-1.30628252e+00 2.88134992e-01 -3.74852151e-01 2.56182611e-01
-3.70863557e-01 6.09223962e-01 -8.20252061e-01 -5.13732612e-01
-1.11539340e+00 -1.03255522e+00 1.36885142e+00 -1.93303861e-02
-5.77435970e-01 -9.68050599e-01 7.06195235e-01 1.12762296e+00
5.52636802e-01 6.26779422e-02 1.10573900e+00 -1.06047559e+00
-2.91458130e-01 -3.75011057e-01 -2.54912198e-01 1.71626776e-01
5.62135987e-02 -6.43381476e-01 -9.28283393e-01 -2.01397032e-01
2.34113693e-01 -9.26976085e-01 9.32159245e-01 1.50198296e-01
8.64255726e-01 -2.67553985e-01 -1.39040411e-01 -1.23083435e-01
1.22601545e+00 -1.61679342e-01 7.61472881e-01 3.93564790e-01
1.47907630e-01 1.09585786e+00 8.78364205e-01 9.78800431e-02
7.60617435e-01 4.15885746e-01 2.42579371e-01 -7.27686957e-02
-5.39514795e-02 2.36227941e-02 -5.76109998e-02 7.84685314e-01
1.95687771e-01 -1.22190475e-01 -1.11723065e+00 6.94446445e-01
-1.54027009e+00 -7.23716378e-01 -2.31962249e-01 2.08606696e+00
1.39017582e+00 -1.49074823e-01 -2.06740260e-01 -9.91669446e-02
1.50822997e-01 -2.32827753e-01 -2.57912934e-01 -5.30533731e-01
9.38152522e-03 8.66853058e-01 -9.89218149e-03 4.98756260e-01
-8.94108474e-01 4.77218807e-01 6.15315151e+00 8.67863357e-01
-5.14283299e-01 2.74047852e-01 7.63944983e-01 1.23310074e-01
-4.69299853e-01 -2.71855056e-01 -7.62719750e-01 6.18096590e-02
1.36644042e+00 7.80407041e-02 -2.46865936e-02 3.98083121e-01
-1.49728835e-01 -1.36342049e-01 -1.40607905e+00 4.71471727e-01
4.25430477e-01 -1.46051383e+00 1.43293172e-01 -3.47781062e-01
4.95336533e-01 8.02891254e-02 -5.68985716e-02 6.25472724e-01
1.24631003e-01 -1.22199440e+00 -3.08800824e-02 8.46103311e-01
6.31899238e-01 -4.51889902e-01 1.32271612e+00 3.10431153e-01
-7.16805756e-01 1.01329468e-01 -2.87433177e-01 2.54244745e-01
3.70019674e-02 4.64617431e-01 -1.15893066e+00 9.67786133e-01
7.32860923e-01 1.34347528e-01 -8.73537660e-01 9.35251057e-01
-1.69699192e-01 5.83319366e-01 -1.57034367e-01 -1.22140050e-01
3.11481774e-01 3.74684632e-01 8.78802389e-02 1.28401721e+00
2.59534363e-02 2.35033303e-01 -9.84516889e-02 5.89719713e-01
-3.54516149e-01 4.89137173e-01 -2.43907690e-01 6.30116612e-02
2.34514907e-01 1.24399793e+00 -2.71241307e-01 -4.35997069e-01
-2.59768724e-01 7.16337800e-01 2.64142126e-01 1.06313840e-01
-5.47445297e-01 -3.51695776e-01 1.36245759e-02 -9.59236473e-02
-1.21602975e-02 4.86640066e-01 -2.89006834e-03 -9.53571916e-01
2.55398691e-01 -1.53065443e+00 1.22400928e+00 -5.48978508e-01
-1.71653211e+00 9.51428294e-01 -6.73636198e-02 -1.12318480e+00
-5.39228082e-01 -5.20395756e-01 -1.13426462e-01 9.17621017e-01
-1.91613162e+00 -1.03485703e+00 -7.30201378e-02 8.08067322e-01
3.52087915e-01 1.45389289e-01 1.14109683e+00 6.13257051e-01
2.51015425e-01 6.02857769e-01 5.05130067e-02 1.16601706e-01
1.15735888e+00 -1.16415358e+00 -7.36854225e-02 1.46717787e-01
8.95068273e-02 8.55787873e-01 6.58844292e-01 -4.61663246e-01
-1.35941792e+00 -6.54891193e-01 1.59877551e+00 -1.20820308e+00
4.62268919e-01 6.06640577e-02 -1.15499914e+00 1.94417536e-01
5.89392483e-01 -2.78896868e-01 1.08522749e+00 2.06388190e-01
-4.02151912e-01 -3.54795642e-02 -1.33962226e+00 1.92741781e-01
3.53838861e-01 -9.18014526e-01 -1.26298201e+00 6.97791755e-01
1.06892943e+00 -3.37677807e-01 -1.26256216e+00 7.18554378e-01
3.73628765e-01 -6.92948699e-01 1.26019740e+00 -9.65993404e-01
7.24645853e-01 -3.10148001e-02 -2.55122364e-01 -8.86987805e-01
-2.53786482e-02 -1.43351614e-01 2.47620102e-02 9.67065632e-01
7.85829842e-01 -4.03641045e-01 5.05960882e-01 8.71020496e-01
-8.87902603e-02 -1.07253683e+00 -1.11611271e+00 -1.28044322e-01
5.70161521e-01 -2.02103764e-01 3.81012052e-01 7.95743585e-01
1.30952820e-01 6.53895915e-01 4.87248972e-02 -1.05847642e-01
1.92257166e-01 4.42717850e-01 2.80853987e-01 -1.05242991e+00
-3.96458775e-01 -7.95578063e-02 2.78423697e-01 -8.55076313e-01
-1.30125523e-01 -1.15513229e+00 1.64714113e-01 -2.06483531e+00
4.58891690e-01 -9.06070694e-02 -3.73601854e-01 2.07422167e-01
-5.84933400e-01 1.23299420e-01 -6.95919022e-02 1.90517753e-01
-1.16970825e+00 3.11118364e-01 1.26530552e+00 -2.89364040e-01
1.48206189e-01 1.57499183e-02 -8.46152902e-01 2.73117304e-01
5.69864392e-01 -6.04957700e-01 -2.90465117e-01 -5.57923198e-01
7.75967836e-01 4.02752101e-01 1.78639174e-01 -5.42455077e-01
3.97632509e-01 2.80470043e-01 2.40518302e-02 -7.29799390e-01
1.39026076e-01 -5.89543998e-01 -6.14934802e-01 6.58609331e-01
-9.36477780e-01 1.51356488e-01 2.03590184e-01 3.33142072e-01
-5.06357372e-01 -5.07446110e-01 3.50089461e-01 -3.73381615e-01
-1.96077943e-01 -1.08036678e-02 -3.11963260e-01 7.49248147e-01
3.90692532e-01 4.31212664e-01 -6.29798830e-01 -5.33305407e-01
-5.56129634e-01 6.42127633e-01 -2.89306462e-01 3.63665193e-01
8.34824085e-01 -9.48681653e-01 -1.30922318e+00 -5.86641312e-01
5.42220891e-01 -2.11010784e-01 4.71872896e-01 8.97730052e-01
-4.55861479e-01 1.04455006e+00 2.87340820e-01 -5.94256103e-01
-1.40555549e+00 5.80302775e-01 3.21352184e-01 -1.10867751e+00
-6.77783191e-02 8.05353343e-01 -9.81556177e-02 -9.48248386e-01
1.24105692e-01 -3.30194771e-01 -6.28509998e-01 1.74434945e-01
8.14755440e-01 2.15558320e-01 5.80817282e-01 -2.86187828e-01
-4.04685348e-01 2.88294464e-01 -3.40053916e-01 -2.04392485e-02
1.09095538e+00 -8.61690268e-02 -5.50995946e-01 5.27061000e-02
1.29869306e+00 -1.32701546e-01 -1.25310391e-01 -5.93499482e-01
5.58214188e-01 -7.29573593e-02 -1.14896871e-01 -1.40682364e+00
-4.13164794e-01 9.50912833e-01 8.15225601e-01 1.34528205e-01
1.14475691e+00 5.61999381e-01 1.07736266e+00 9.15552318e-01
1.05068147e-01 -9.75632548e-01 4.07089323e-01 6.81004167e-01
1.11111510e+00 -1.50787139e+00 -1.19218685e-01 -1.07605539e-01
-5.14994025e-01 9.59507465e-01 2.46609405e-01 2.16974884e-01
4.07466233e-01 -3.33736360e-01 3.23347986e-01 -8.64391983e-01
-8.21897209e-01 -2.83768266e-01 9.75041211e-01 1.62192300e-01
8.98550212e-01 -3.48022044e-01 -7.14441419e-01 5.86589217e-01
-5.45505285e-02 2.37692475e-01 4.61466834e-02 9.22078013e-01
-3.33186358e-01 -1.16548610e+00 -4.23926443e-01 6.76138878e-01
-1.07449186e+00 -4.86638159e-01 -4.90761220e-01 3.21395785e-01
-1.40816659e-01 1.54947591e+00 -4.64659065e-01 -4.07385938e-02
4.48776305e-01 2.11670637e-01 3.80891949e-01 -6.24756515e-01
-1.19202554e+00 -4.48671907e-01 5.75853527e-01 -6.44650817e-01
-5.01604617e-01 -3.04250330e-01 -1.22020316e+00 1.29590824e-01
-2.79467195e-01 6.18667006e-01 4.33551788e-01 1.19164240e+00
6.36791348e-01 5.43068409e-01 3.98095906e-01 3.39971840e-01
-9.17549908e-01 -9.83938217e-01 1.70819387e-01 7.55142987e-01
3.72100919e-01 -8.37078691e-02 -1.77231222e-01 -1.63113743e-01] | [8.788662910461426, 8.579341888427734] |
3ac55d30-5459-479c-ba15-e45d8543116a | captainglove-capacitive-and-inertial-fusion | 2306.04319 | null | https://arxiv.org/abs/2306.04319v1 | https://arxiv.org/pdf/2306.04319v1.pdf | CaptAinGlove: Capacitive and Inertial Fusion-Based Glove for Real-Time on Edge Hand Gesture Recognition for Drone Control | We present CaptAinGlove, a textile-based, low-power (1.15Watts), privacy-conscious, real-time on-the-edge (RTE) glove-based solution with a tiny memory footprint (2MB), designed to recognize hand gestures used for drone control. We employ lightweight convolutional neural networks as the backbone models and a hierarchical multimodal fusion to reduce power consumption and improve accuracy. The system yields an F1-score of 80% for the offline evaluation of nine classes; eight hand gesture commands and null activity. For the RTE, we obtained an F1-score of 67% (one user). | ['Paul Lukowicz', 'Bo Zhou', 'Lala Ray', 'Daniel Geißler', 'Sungho Suh', 'Hymalai Bello'] | 2023-06-07 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.21422291e-01 -3.39400172e-01 -3.71657759e-01 -1.24827027e-01
-2.90179491e-01 -8.15711141e-01 -3.45320702e-02 -5.97711504e-01
-7.35249996e-01 4.64409649e-01 3.01186014e-02 -5.14680386e-01
-1.53277904e-01 -4.52356756e-01 -4.96185869e-01 -5.58931231e-01
-1.09946288e-01 -3.07130575e-01 1.08029116e-02 1.90920115e-03
-2.58102030e-01 7.65894711e-01 -1.69716012e+00 3.49461049e-01
3.40666890e-01 1.55897617e+00 -4.72667277e-01 1.19900954e+00
6.00722730e-01 4.85081345e-01 -7.68248081e-01 -6.21524453e-01
4.69923705e-01 3.79323930e-01 -4.86420244e-01 -7.20797658e-01
6.78885400e-01 -9.51655149e-01 -6.86920524e-01 3.71178538e-01
1.07028985e+00 -2.76351660e-01 7.58917853e-02 -1.42408049e+00
2.77116969e-02 9.78506431e-02 -8.50844383e-02 -4.24607456e-01
4.91412431e-01 6.91333413e-01 6.46255076e-01 -6.38460279e-01
3.59838516e-01 5.66630960e-01 7.32628047e-01 9.22292233e-01
-8.65962267e-01 -1.15131986e+00 -3.61515403e-01 -2.51044303e-01
-1.50661099e+00 -3.76635432e-01 4.14481282e-01 2.04645321e-02
1.26884162e+00 6.63635612e-01 9.87983644e-01 1.87080836e+00
3.99210244e-01 6.83050990e-01 9.87925410e-01 1.21210389e-01
2.30886087e-01 -3.33596706e-01 -6.59282133e-02 7.17058480e-01
5.88480830e-01 3.23397309e-01 -9.24303710e-01 -2.75065988e-01
8.39627445e-01 3.02566200e-01 -2.62525469e-01 -2.33494952e-01
-1.23821104e+00 8.25906470e-02 5.04580259e-01 -7.11634010e-02
-4.02238667e-01 7.93295264e-01 4.00444925e-01 8.97024572e-02
-2.10483238e-01 2.58908778e-01 -6.65226638e-01 -8.60541701e-01
-6.44046247e-01 -2.58182231e-02 1.22072446e+00 1.25501776e+00
-9.36725959e-02 1.00089848e-01 -3.91526937e-01 5.66753410e-02
3.26309413e-01 1.00755191e+00 1.85957938e-01 -6.24346256e-01
6.23918533e-01 6.09864354e-01 5.44880331e-01 -4.87080693e-01
-5.85901856e-01 -1.43923879e-01 -6.58488035e-01 4.95036662e-01
2.77165055e-01 -8.52207303e-01 -1.09750938e+00 1.37601995e+00
1.05786860e-01 -2.79495686e-01 -1.96521506e-01 1.16110027e+00
1.11097658e+00 -4.91284393e-02 4.69748348e-01 4.11853760e-01
1.27585042e+00 -6.70066297e-01 -6.74406946e-01 1.52153626e-01
1.93952218e-01 -5.57283163e-01 1.40606916e+00 7.26087928e-01
-6.46697223e-01 -4.81687009e-01 -1.18904388e+00 1.00117423e-01
-4.86713558e-01 5.88956237e-01 9.06192303e-01 1.46442807e+00
-8.48774493e-01 5.05745649e-01 -1.01419556e+00 -4.81526047e-01
6.35038972e-01 1.05225408e+00 -3.07368487e-01 5.13612092e-01
-4.95390356e-01 6.14270926e-01 1.23039052e-01 1.41284883e-01
-9.43630159e-01 -3.28122735e-01 -1.98223799e-01 -1.06138162e-01
3.03580761e-01 -5.41252196e-01 9.54723954e-01 -4.49332476e-01
-2.00264883e+00 5.76859474e-01 1.52158290e-01 -1.26849458e-01
8.39732349e-01 -8.15080166e-01 -5.79276562e-01 1.12921871e-01
-8.60776126e-01 7.31790602e-01 9.61799562e-01 -8.78086925e-01
-6.56217158e-01 -2.44258747e-01 -5.99669144e-02 -3.10677439e-02
-6.29170120e-01 -6.16662167e-02 -2.18256369e-01 -5.28747559e-01
-3.53696078e-01 -1.07179666e+00 3.55102539e-01 2.03685522e-01
-1.84724987e-01 1.90165527e-02 1.08551490e+00 -8.73319507e-01
1.31745112e+00 -1.87958300e+00 -2.96508044e-01 4.40134674e-01
4.59955871e-01 7.24629879e-01 -5.62252700e-02 1.06633544e-01
2.84811944e-01 6.58686534e-02 2.65320152e-01 -1.67371646e-01
1.97754860e-01 1.71333969e-01 -1.90185174e-01 3.80907416e-01
-3.61481272e-02 1.37391782e+00 -5.08354902e-01 1.79941635e-04
5.11187255e-01 7.74381161e-01 -1.38269871e-01 5.34323812e-01
3.55459243e-01 2.63919592e-01 -2.05920264e-01 1.41706312e+00
4.34887677e-01 3.33180636e-01 1.55977070e-01 -4.89065677e-01
-6.26250803e-02 2.65218243e-02 -1.09841752e+00 1.79317796e+00
-4.36169297e-01 8.48417997e-01 3.29658628e-01 3.09409976e-01
7.51381755e-01 5.00416636e-01 5.38152456e-01 -4.87959355e-01
9.03983772e-01 1.10372089e-01 -2.46079877e-01 -9.33071733e-01
4.26996231e-01 5.03455162e-01 -2.40617990e-01 3.76310945e-01
1.46933466e-01 3.46213013e-01 -4.63795245e-01 -2.74536937e-01
1.40939307e+00 5.55575788e-01 -2.38616392e-01 -3.16897258e-02
-2.06890836e-01 -4.01265830e-01 -9.38363466e-03 5.98350465e-01
-4.07074988e-01 4.65781957e-01 -1.44541621e-01 -8.17581236e-01
-7.09131539e-01 -1.02096307e+00 4.19532508e-01 1.28092849e+00
2.32986540e-01 -3.69878352e-01 -1.01565075e+00 -7.13948131e-01
-2.49427576e-02 2.59670854e-01 -3.81624550e-01 -1.19040580e-02
-5.94126046e-01 -2.80787826e-01 1.31380773e+00 1.04147053e+00
1.04674101e+00 -8.58652949e-01 -1.81753755e+00 -7.07019400e-03
-1.23866089e-01 -1.03988779e+00 -3.91887188e-01 3.63201618e-01
-6.72389865e-01 -9.95513499e-01 -5.33884048e-01 -3.60194504e-01
2.59234697e-01 -2.68443733e-01 8.78552020e-01 -1.25637978e-01
-5.36642790e-01 8.78587186e-01 -3.64310712e-01 -5.45799613e-01
6.68504894e-01 2.54169524e-01 2.37784192e-01 1.84144247e-02
7.00463712e-01 -3.99770349e-01 -8.79907668e-01 4.74646360e-01
-3.26923847e-01 -3.26318383e-01 6.92982554e-01 6.99129343e-01
1.51696116e-01 -7.83134997e-01 -1.13469981e-01 1.99633390e-01
8.00363362e-01 2.09239960e-01 -5.62847972e-01 3.85280699e-01
-5.31787276e-01 -2.13701010e-01 4.03128326e-01 -1.09333372e+00
-5.94148517e-01 5.05583048e-01 -2.41938680e-01 -5.47970533e-01
-4.60565418e-01 -3.23351979e-01 -3.95444602e-01 -8.44330549e-01
6.18401825e-01 -3.15827527e-03 -3.49632263e-01 -3.31985116e-01
3.65992576e-01 1.38764894e+00 7.21235633e-01 -4.12851185e-01
5.31569481e-01 2.66575336e-01 -4.44761328e-02 -6.83311462e-01
4.37152907e-02 -2.12827325e-01 -6.56698406e-01 -6.08435571e-01
8.52952898e-01 -1.03658438e+00 -1.95803916e+00 5.18260837e-01
-1.02028906e+00 -7.87215829e-01 -8.46021250e-02 5.43621421e-01
-1.64345354e-01 -1.59640878e-01 -2.17030361e-01 -1.16709578e+00
-1.19099975e+00 -8.52783322e-01 1.28154325e+00 5.26566684e-01
-6.22970045e-01 -2.48955250e-01 -2.38879547e-01 3.72830361e-01
5.88987827e-01 5.69396973e-01 4.00067829e-02 -1.96401164e-01
-5.09571493e-01 -6.41447544e-01 -1.46285981e-01 9.19312797e-03
2.21930351e-02 -2.21277669e-01 -1.57439876e+00 -6.35212541e-01
-4.68665451e-01 -5.32238841e-01 3.78685385e-01 2.91811004e-02
1.15638864e+00 -4.53239262e-01 -4.66389775e-01 9.76213098e-01
1.30488861e+00 3.18359375e-01 7.78243959e-01 -2.34111756e-01
9.96247888e-01 -1.22035816e-02 3.09423417e-01 6.12907469e-01
-3.47208939e-02 4.39305991e-01 6.61441028e-01 -2.58124918e-01
-9.74548310e-02 -3.31060767e-01 4.32019770e-01 1.20306365e-01
-7.00468659e-01 -8.18195403e-01 -7.76075125e-01 5.66506684e-01
-1.62379444e+00 -4.00763452e-01 8.30264296e-03 2.32733226e+00
3.16307962e-01 -2.11316168e-01 4.85214859e-01 8.04138631e-02
3.66511703e-01 -1.72736779e-01 -6.49649322e-01 -5.30639291e-01
2.22835734e-01 8.62220705e-01 1.09173739e+00 2.84105744e-02
-1.47083902e+00 8.25690448e-01 6.96923733e+00 6.38792932e-01
-1.41846931e+00 6.97707012e-02 5.25585003e-02 -7.92210102e-01
5.33391356e-01 -6.94688439e-01 -5.43017209e-01 5.56268811e-01
1.10530007e+00 7.32224047e-01 8.35856318e-01 1.15973830e+00
-2.90493220e-01 1.10867620e-01 -9.84521925e-01 1.49793267e+00
4.84944263e-04 -8.89241636e-01 -1.13781713e-01 3.12381536e-01
2.26424694e-01 2.86470801e-01 1.19485997e-01 -6.52256459e-02
2.01765046e-01 -1.39199746e+00 6.42210007e-01 3.28759938e-01
1.80380309e+00 -1.03477979e+00 1.01792312e+00 3.07675414e-02
-1.52665532e+00 -1.84529498e-01 9.53620747e-02 -1.94680318e-01
-3.04099411e-01 -3.13165814e-01 -6.38403654e-01 1.64743796e-01
1.21930742e+00 -8.81091505e-02 -4.16235745e-01 6.93976223e-01
-3.30267578e-01 6.40865743e-01 -6.55353427e-01 -9.30068791e-01
-4.09693755e-02 4.47903335e-01 4.71680462e-01 1.42843688e+00
1.86671302e-01 3.20611835e-01 -2.36431703e-01 4.87596989e-01
-3.44121516e-01 -3.12081248e-01 -5.16297460e-01 9.63191909e-04
2.66241193e-01 1.39639080e+00 -5.24461746e-01 -1.11911595e-02
1.87352952e-02 1.41374350e+00 -3.87520134e-01 2.68521786e-01
-9.18933392e-01 -1.29837298e+00 1.04740083e+00 -5.99975660e-02
3.86943191e-01 -4.19575930e-01 -5.91444671e-01 -1.08269703e+00
3.49903643e-01 -6.24659300e-01 -8.89330134e-02 -6.16844475e-01
-8.60774159e-01 4.97250646e-01 -4.58181292e-01 -1.47088122e+00
8.82516876e-02 -1.19704974e+00 -3.34952295e-01 6.95911050e-01
-9.06266212e-01 -1.69046187e+00 -1.17252457e+00 7.49849856e-01
3.79386954e-02 -2.18992576e-01 1.27816594e+00 4.61308122e-01
-3.79186004e-01 1.29450965e+00 -2.12962747e-01 6.71045840e-01
5.70189416e-01 -9.73166168e-01 3.85752469e-01 6.56847298e-01
-1.09593123e-01 9.02286232e-01 1.19608194e-01 -6.76916778e-01
-2.10791373e+00 -1.01010668e+00 3.98128033e-01 -8.90748858e-01
2.22530961e-01 -6.85137331e-01 -1.02387786e-01 5.63397586e-01
2.38833576e-01 1.24335937e-01 8.91650558e-01 -1.33614376e-01
-3.13789517e-01 -2.96894103e-01 -1.60388649e+00 9.23509836e-01
1.43200946e+00 -6.14021242e-01 -3.53124142e-01 -3.11591953e-01
2.82421052e-01 -6.60065114e-01 -1.20430696e+00 4.83054101e-01
1.81047344e+00 -5.15659630e-01 8.16745281e-01 -6.26743495e-01
-1.52927682e-01 -3.80978495e-01 -7.43552148e-02 -7.47684300e-01
-6.39779447e-03 -1.31497645e+00 -5.00455201e-01 7.26072311e-01
1.39470518e-01 -5.27230442e-01 1.15106082e+00 1.18995655e+00
3.21000814e-01 -7.56214321e-01 -1.23392940e+00 -8.04296255e-01
-7.19096184e-01 -6.42037690e-01 7.90812671e-01 3.23215574e-01
2.95255631e-01 5.63965738e-02 -6.72241569e-01 6.61841035e-02
4.59496945e-01 -2.95116067e-01 1.05879092e+00 -1.09276819e+00
7.48808682e-02 -1.55096516e-01 -5.64804435e-01 -1.02467656e+00
-6.08983636e-01 -2.04967812e-01 2.67701596e-02 -9.66338575e-01
-3.22442859e-01 1.71824187e-01 -2.99931884e-01 1.05006325e+00
4.14259911e-01 6.75283253e-01 2.06345022e-01 -3.39400351e-01
-7.30150878e-01 1.90650523e-01 5.79440475e-01 -2.45044351e-01
-2.97629923e-01 -1.86985433e-01 -2.47604400e-01 4.80175495e-01
9.21454847e-01 -3.06571573e-01 7.49995410e-02 -4.55001414e-01
1.25464842e-01 -4.62762982e-01 6.70189738e-01 -1.41159904e+00
3.57052296e-01 2.57709585e-02 8.76592815e-01 -4.56655085e-01
5.00720859e-01 -1.51045549e+00 3.05655807e-01 8.68901551e-01
1.65402740e-02 -1.18715994e-01 6.09831870e-01 1.00269742e-01
4.22717363e-01 5.71535707e-01 2.53925562e-01 2.83483505e-01
-5.38152933e-01 2.64917523e-01 -4.38351601e-01 -6.29237294e-01
9.13519681e-01 -5.57108164e-01 -3.94455016e-01 -1.92305297e-01
-3.51262271e-01 3.34809558e-03 2.02977538e-01 4.77364987e-01
8.28847110e-01 -1.22831535e+00 -7.07433224e-02 5.92139006e-01
1.99768096e-01 -3.04885983e-01 -9.60365236e-02 5.46567678e-01
-8.55316043e-01 4.33778912e-01 -7.51631439e-01 -4.89099979e-01
-1.78244281e+00 -7.23766237e-02 3.29646528e-01 2.38485709e-01
-3.56792480e-01 8.74346077e-01 -9.83703852e-01 -3.60772550e-01
9.45609331e-01 -6.20725334e-01 9.64322239e-02 -2.25642651e-01
6.04985654e-01 9.42974031e-01 2.22954080e-01 -1.54598355e-01
-7.29096293e-01 8.28098655e-01 6.61220193e-01 -2.25883067e-01
1.25641823e+00 3.48321438e-01 4.70111638e-01 5.41052818e-02
8.88483405e-01 -3.84703502e-02 -1.44584787e+00 6.28390193e-01
-3.31089109e-01 -4.62808341e-01 3.55107598e-02 -1.45820630e+00
-9.09208894e-01 7.77388215e-01 1.59515285e+00 -2.84388453e-01
1.23535573e+00 -5.55656552e-01 1.09092200e+00 9.51134205e-01
6.54068708e-01 -1.20087361e+00 -1.84902295e-01 3.13911170e-01
6.56611204e-01 -1.06520116e+00 -1.39144883e-01 2.73101460e-02
-4.53912556e-01 1.42800987e+00 5.73069692e-01 1.94616355e-02
4.71000850e-01 8.48168314e-01 5.37779033e-02 -2.05797091e-01
-2.94259667e-01 -1.73345223e-01 5.68120599e-01 9.84163821e-01
1.92064404e-01 5.03314972e-01 1.07208766e-01 7.05734968e-01
-2.01451093e-01 4.29103285e-01 -3.47649872e-01 1.20880663e+00
2.22630304e-04 -4.93908525e-01 -1.45204961e-01 4.24570352e-01
-3.85310650e-01 1.70054436e-01 -9.95172858e-01 7.61186182e-01
4.89744365e-01 1.03893781e+00 -1.80602947e-04 -1.62732184e+00
6.28868282e-01 1.80370867e-01 5.49022794e-01 1.77669778e-01
-1.42531216e+00 -9.50212106e-02 1.03563257e-01 -1.22470069e+00
-1.78922474e-01 7.56522175e-03 -7.89278746e-01 -7.28478432e-01
-1.75525874e-01 -6.97428226e-01 1.08361673e+00 6.17227733e-01
7.16303647e-01 4.55049366e-01 2.87227154e-01 -9.58815277e-01
-1.28839508e-01 -1.11520255e+00 -2.71833390e-01 3.41907330e-02
3.50063175e-01 -4.25580502e-01 -1.04887262e-01 -1.34375826e-01] | [6.554998397827148, -0.16574181616306305] |
c9599a47-12ff-402d-ba6c-6935cfdc7ac3 | machine-learning-in-asset-management-part-1 | null | null | https://jfds.pm-research.com/content/2/1/10/tab-article-info | https://vixra.org/pdf/1912.0492v1.pdf | Machine Learning in Asset Management—Part 1: Portfolio Construction—Trading Strategies | This is the first in a series of arti-cles dealing with machine learning in asset management. Asset management can be broken into the following tasks: (1) portfolio construction, (2) risk management, (3) capital management, (4) infra-structure and deployment, and (5) sales and marketing. This article focuses on portfolio construction using machine learning. Historically, algorithmic trading could be more narrowly defined as the automation of sell-side trade execution, but since the introduction of more advanced algorithms, the definition has grown to include idea generation, alpha factor design, asset allocation, position sizing, and the testing of strategies. Machine learning, from the vantage of a decision-making tool, can help in all these areas. | ['Derek Snow'] | 2020-02-10 | null | null | null | journal-of-financial-data-science-2020-2 | ['algorithmic-trading'] | ['time-series'] | [-8.58649686e-02 -3.61864008e-02 -1.86957344e-01 -1.62950203e-01
-3.98478180e-01 -8.47732425e-01 5.67584097e-01 4.51717943e-01
-1.52300581e-01 6.14418030e-01 7.49175102e-02 -1.16219306e+00
-5.39491713e-01 -1.01902521e+00 -2.19159096e-01 -5.54672539e-01
-5.19414097e-02 1.03146851e+00 -1.00556791e-01 1.34538058e-02
7.89011359e-01 7.58784890e-01 -9.91240561e-01 -4.96644862e-02
2.47686505e-01 1.68950319e+00 -4.66215909e-01 4.27890062e-01
-4.83988285e-01 1.07306027e+00 -5.89997470e-01 -8.06563139e-01
5.17920077e-01 -4.09974039e-01 -4.34503734e-01 6.85676783e-02
-5.42981386e-01 -5.45268059e-01 5.07617354e-01 6.22494459e-01
4.81854938e-02 -1.27573445e-01 7.20449090e-01 -1.21139753e+00
1.45710066e-01 9.94152784e-01 -6.48307145e-01 4.86917906e-02
-1.12988792e-01 -3.01076658e-02 1.33291125e+00 -4.93578255e-01
-1.22919142e-01 8.61005425e-01 4.85979766e-01 -3.16978842e-01
-1.02297127e+00 -7.38150477e-01 -6.42374158e-02 -2.25515604e-01
-8.09725165e-01 1.49555923e-02 5.89563131e-01 -9.29973841e-01
9.09857273e-01 1.80186659e-01 9.32203293e-01 2.14342877e-01
5.42356908e-01 4.99756515e-01 1.09383893e+00 -7.02552736e-01
6.67535245e-01 2.11381227e-01 -9.44776759e-02 -1.93771031e-02
5.78272402e-01 4.21236634e-01 -3.13370265e-02 -2.98976272e-01
7.54824162e-01 -3.75915878e-02 2.93477237e-01 -2.81629562e-01
-7.67950416e-01 1.44096625e+00 -1.95849627e-01 2.25540161e-01
-3.68468493e-01 4.60672267e-02 3.74759257e-01 7.99281418e-01
2.80024946e-01 1.17013454e+00 -7.33599782e-01 -2.99796671e-01
-1.19396210e+00 3.32600713e-01 1.26565111e+00 6.14622355e-01
2.59906679e-01 3.78514796e-01 2.69551665e-01 3.68491292e-01
5.27418375e-01 1.70043066e-01 2.46621177e-01 -1.14116836e+00
6.29471660e-01 3.93217862e-01 3.61971736e-01 -4.82240617e-01
-4.87889469e-01 -6.45621777e-01 -2.90329188e-01 5.16394675e-01
4.67468172e-01 -5.28761446e-01 -6.39835060e-01 1.04437542e+00
-9.45180357e-02 -4.83898073e-01 -3.07688802e-01 3.12467247e-01
-1.15552194e-01 5.26649415e-01 -3.61568779e-02 -5.25565147e-01
9.93266165e-01 -9.14437473e-01 -4.10773337e-01 -3.20679136e-03
2.54850000e-01 -8.51159692e-01 7.49695068e-03 9.39404070e-01
-1.36942923e+00 -5.49375042e-02 -1.24573636e+00 6.25964880e-01
-4.69088137e-01 -3.20559114e-01 1.24172258e+00 8.39002907e-01
-6.44843578e-01 7.98155546e-01 -4.82270271e-01 2.64519334e-01
5.44802606e-01 5.81322849e-01 1.75592378e-01 4.32227403e-01
-1.00945985e+00 1.09567714e+00 4.16003525e-01 -1.36048391e-01
-4.91030425e-01 -7.20752060e-01 -4.49328363e-01 3.91795754e-01
7.95631230e-01 -6.22218490e-01 1.45628786e+00 -1.10519016e+00
-1.12971008e+00 3.07458729e-01 9.62758362e-01 -7.42016971e-01
5.10844946e-01 -2.82326818e-01 -1.65409133e-01 -4.36880559e-01
-1.42154649e-01 -5.35697192e-02 9.18023825e-01 -6.59294784e-01
-1.14441562e+00 -3.34672004e-01 5.94047979e-02 6.05862066e-02
6.86065108e-02 1.85145780e-01 3.18856478e-01 -1.07885671e+00
1.01972848e-01 -7.42487013e-01 -1.67229325e-01 -5.29940367e-01
-7.01651126e-02 1.12837121e-01 7.40288347e-02 -7.30291724e-01
1.41787851e+00 -1.58693016e+00 1.85583681e-02 6.05017841e-01
-7.52306059e-02 5.09795882e-02 6.48460448e-01 7.23566592e-01
-5.51220357e-01 2.33332559e-01 1.03953250e-01 1.75511479e-01
2.76029140e-01 -2.79321522e-01 -5.74679911e-01 8.79214257e-02
7.48554394e-02 8.55998695e-01 -5.03799260e-01 -7.16975182e-02
2.29968548e-01 -5.46949625e-01 -1.73670337e-01 3.78665118e-03
-3.13317329e-01 -5.24388216e-02 -6.31096244e-01 1.13784945e+00
1.51627064e-01 -1.42264338e-02 2.47680441e-01 2.55593777e-01
-3.79133791e-01 3.25978726e-01 -1.24886429e+00 7.62452543e-01
-3.35814983e-01 3.99638861e-01 -6.06509931e-02 -1.04850554e+00
7.87119627e-01 7.51857609e-02 7.39440739e-01 -6.38385832e-01
2.02227280e-01 4.33346808e-01 4.14482772e-01 3.12888995e-02
4.36900824e-01 -5.35687685e-01 -2.34581590e-01 1.11804354e+00
-3.42469215e-01 1.13472082e-01 9.82044041e-02 -2.89816052e-01
1.17352259e+00 3.71741727e-02 5.96623480e-01 3.66700999e-02
-5.81627749e-02 1.52074546e-01 4.30837482e-01 3.78657460e-01
1.83678210e-01 1.06215261e-01 8.98111701e-01 -4.00995612e-01
-1.25500393e+00 -1.15779769e+00 -1.33601874e-01 9.38643932e-01
-5.21067083e-01 9.96388420e-02 -6.09636843e-01 -5.05935848e-01
1.01100290e+00 9.13647234e-01 -4.65214640e-01 3.13707590e-01
-5.40319920e-01 -8.28194916e-01 5.33369146e-02 7.03931451e-01
1.07473888e-01 -1.11850941e+00 -9.37066376e-01 5.89132726e-01
7.50336289e-01 -4.51371193e-01 -2.43420973e-01 8.07685077e-01
-8.76965284e-01 -1.16964579e+00 -5.01246214e-01 -1.45368531e-01
1.58535734e-01 -2.52636343e-01 1.30083299e+00 -7.09467903e-02
-2.01443061e-01 2.40758166e-01 -4.75259721e-01 -1.08834505e+00
-2.55379707e-01 -5.02402559e-02 -5.25722876e-02 -7.29344711e-02
2.09415242e-01 -5.07744312e-01 -2.38635138e-01 3.09468627e-01
-6.63693786e-01 -2.91527450e-01 1.00198865e+00 8.19207370e-01
3.66941303e-01 7.28173733e-01 7.92689979e-01 -8.97319615e-01
4.28708047e-01 -3.86454999e-01 -9.16708529e-01 5.81871152e-01
-1.10281479e+00 5.34698516e-02 7.47041553e-02 -1.86696663e-01
-1.00836015e+00 -3.38421047e-01 7.78357834e-02 -1.88339710e-01
2.43289396e-01 8.37260723e-01 -2.30809107e-01 8.04935619e-02
5.94419576e-02 -1.86013933e-02 6.06986508e-02 -5.13954759e-01
6.01387136e-02 5.06103277e-01 1.34456187e-01 -2.59421706e-01
1.03775716e+00 -9.63376835e-02 3.88797730e-01 -5.16575217e-01
-6.28543317e-01 6.70832619e-02 -7.13334501e-01 -1.62115812e-01
5.66906333e-01 -5.27778447e-01 -5.42501330e-01 3.66147846e-01
-3.66272777e-01 -4.51150328e-01 -6.29127026e-01 4.29622740e-01
-7.23711431e-01 -2.94031113e-01 -3.01973492e-01 -1.10686493e+00
-2.80188829e-01 -8.30143154e-01 4.17257011e-01 1.41126215e-01
-6.07821941e-01 -1.18284476e+00 -3.29796784e-02 7.77615190e-01
3.65968555e-01 5.91669023e-01 1.30180371e+00 -1.19694877e+00
-8.77830029e-01 -6.92254841e-01 1.23268478e-01 5.11866450e-01
3.04263860e-01 -1.37078881e-01 -5.36142409e-01 -2.30515227e-01
3.56193244e-01 -2.28118867e-01 4.60659146e-01 5.51926196e-01
7.97210574e-01 -2.19888255e-01 -1.90337956e-01 5.34751654e-01
1.20991266e+00 1.16924894e+00 6.47024274e-01 9.80329216e-01
1.25407025e-01 1.06432867e+00 1.31400847e+00 6.41068995e-01
-1.12734050e-01 6.69943273e-01 3.67587209e-01 3.33333224e-01
5.81000030e-01 -1.56494036e-01 9.46379229e-02 2.27031365e-01
5.39813377e-02 -1.65744409e-01 -9.65807974e-01 1.59345523e-01
-1.86470699e+00 -1.22672462e+00 4.64204967e-01 2.05319238e+00
2.99096107e-01 9.62525785e-01 6.45733118e-01 4.12394345e-01
3.32170516e-01 8.71499479e-02 -6.79034352e-01 -7.53888726e-01
3.05515260e-01 1.68574899e-01 7.50055194e-01 4.62365486e-02
-1.13205492e+00 3.94704521e-01 6.35819435e+00 7.13184357e-01
-7.69838691e-01 -2.50376433e-01 1.02609789e+00 -3.52984250e-01
-5.57444930e-01 2.20616654e-01 -9.08462644e-01 6.77780449e-01
5.47449827e-01 -1.62799954e-01 4.84129369e-01 9.17112887e-01
5.62380739e-02 -2.55544126e-01 -1.09862351e+00 6.36721969e-01
-1.02076225e-01 -1.57169425e+00 -2.34419674e-01 4.18249607e-01
6.12143338e-01 -5.39740443e-01 6.05229847e-02 3.05201083e-01
2.18952507e-01 -1.26298094e+00 1.03912067e+00 7.43477225e-01
2.81786770e-01 -1.31339753e+00 1.04445565e+00 -4.72393148e-02
-9.29545462e-01 -7.50386357e-01 1.49382710e-01 -3.29279564e-02
5.04677892e-01 5.65237284e-01 -5.83397269e-01 5.41972399e-01
5.16061306e-01 2.36869782e-01 -1.05558060e-01 1.11130524e+00
6.05461448e-02 4.31698710e-01 -1.29156098e-01 -1.32677350e-02
2.00821400e-01 -7.40277469e-01 8.19450691e-02 6.11616135e-01
2.75206268e-01 8.80895481e-02 2.26257741e-02 6.68519557e-01
3.10265392e-01 -2.41912887e-01 -1.61217541e-01 -6.17027760e-01
4.97601390e-01 1.00181544e+00 -9.91306067e-01 1.70646444e-01
-4.28882182e-01 2.40497768e-01 -3.54750127e-01 3.74303013e-02
-6.35289907e-01 -6.98102355e-01 5.36700666e-01 6.88474059e-01
7.98159361e-01 -5.36696352e-02 -9.46950674e-01 -3.17302436e-01
4.62346748e-02 -1.10154486e+00 5.10149479e-01 -4.22349960e-01
-1.10943389e+00 -2.28431560e-02 1.06055193e-01 -9.28422511e-01
-7.43952870e-01 -8.49268615e-01 -9.06315804e-01 8.46836686e-01
-1.00925386e+00 -7.61831284e-01 3.37045759e-01 -2.47106090e-01
3.09443444e-01 -9.78094280e-01 4.45415109e-01 8.74584913e-02
-2.89527684e-01 4.61648166e-01 5.02915144e-01 1.17661655e-01
2.10690171e-01 -1.51223683e+00 2.08545014e-01 3.18711311e-01
-4.63832542e-02 2.45241180e-01 4.31039959e-01 -1.11394322e+00
-1.20909762e+00 -8.18971097e-01 6.56111896e-01 -3.82044792e-01
1.13442826e+00 -3.30310166e-02 -4.44831789e-01 6.64890885e-01
-1.06700510e-01 -1.10766017e+00 9.03459430e-01 3.09677273e-01
2.15116933e-01 -6.02594197e-01 -9.74521637e-01 1.89341456e-01
2.55997449e-01 -1.91957816e-01 -4.34127659e-01 1.11190893e-01
3.72848362e-01 -2.27506220e-01 -1.16742957e+00 1.87691256e-01
9.95933771e-01 -8.86680961e-01 9.48368669e-01 -7.40264595e-01
8.28632265e-02 1.75290450e-01 -3.75249535e-02 -7.86709309e-01
-4.82812077e-01 -9.37586069e-01 -3.14477056e-01 1.45280528e+00
7.63034523e-01 -7.23622262e-01 9.72524703e-01 6.02951050e-01
-6.17920980e-03 -1.18145657e+00 -8.80034924e-01 -7.16098130e-01
1.76267937e-01 -4.89764541e-01 8.95702660e-01 6.98136866e-01
-4.51598205e-02 1.88878298e-01 -1.78393275e-01 -5.38340747e-01
8.09914470e-01 5.72959840e-01 4.80490804e-01 -1.36202204e+00
-9.62366939e-01 -1.04039586e+00 -8.64896178e-02 -3.85852396e-01
-4.16502893e-01 -4.18564826e-01 -4.33804095e-01 -1.35293221e+00
-1.74254745e-01 -6.60077035e-01 -4.52791512e-01 2.68401206e-01
2.67463982e-01 -4.96201247e-01 3.34545076e-01 2.96721667e-01
-6.01290837e-02 -2.35070027e-02 7.59775519e-01 -3.77615616e-02
-2.99575657e-01 8.86254191e-01 -1.16371608e+00 7.25819111e-01
1.02546990e+00 -2.80696779e-01 -3.75033289e-01 1.46136761e-01
7.09421396e-01 4.60422754e-01 -9.61111858e-02 -4.54470158e-01
-1.13692567e-01 -6.49154186e-01 6.27667785e-01 -7.29863048e-01
1.14132114e-01 -8.55439305e-01 6.32534087e-01 7.75794566e-01
-2.74980903e-01 5.53117752e-01 2.66262013e-02 3.63853097e-01
-8.28479752e-02 -7.30878234e-01 5.42419374e-01 -1.65852115e-01
-2.62224823e-01 2.67681107e-02 -3.70844275e-01 -8.07528496e-02
1.32922280e+00 -3.16356272e-01 -1.57559559e-01 -4.21265393e-01
-8.01207066e-01 4.45736319e-01 4.51176316e-01 1.78133503e-01
2.50364691e-01 -1.12812161e+00 -3.43369126e-01 3.66356745e-02
-4.42522585e-01 -1.25864714e-01 -2.64326900e-01 5.67657053e-01
-6.78975284e-01 7.96864212e-01 -1.94830880e-01 1.03537314e-01
-7.72181153e-01 4.11637664e-01 4.56577927e-01 -8.98381829e-01
-5.31130098e-02 7.28415489e-01 -3.23218922e-03 4.56467927e-01
3.52214664e-01 1.39505342e-01 -3.67993087e-01 7.74649739e-01
4.93581444e-01 8.02238643e-01 1.10825755e-01 -7.85899907e-02
-1.88301086e-01 3.78767043e-01 -2.15096012e-01 -3.67295057e-01
1.65629768e+00 3.21146160e-01 -7.39899948e-02 5.02072752e-01
5.12895584e-01 -3.52689981e-01 -1.06582046e+00 2.41900429e-01
7.54417121e-01 -4.11718667e-01 1.27249882e-01 -1.10891843e+00
-1.06426287e+00 7.66901374e-01 2.13452712e-01 6.43662572e-01
1.14952517e+00 -3.26008320e-01 6.55808508e-01 1.80040941e-01
5.67739487e-01 -1.63130951e+00 2.24084347e-01 2.19119102e-01
7.84216762e-01 -7.55463541e-01 3.79884362e-01 3.09868902e-01
-7.64368415e-01 1.18535638e+00 2.43022189e-01 5.58048151e-02
1.14377260e+00 5.35776675e-01 -3.91917884e-01 -1.56027451e-01
-6.52691007e-01 3.47416736e-02 3.06155384e-01 2.35636592e-01
2.07084373e-01 6.65555373e-02 5.19080199e-02 1.02248311e+00
-4.89704669e-01 -1.84021920e-01 3.48677188e-01 1.18313920e+00
-8.55052292e-01 -1.39265239e+00 -5.90845942e-01 1.25727057e+00
-6.72036111e-01 -1.31814592e-02 -4.59761322e-01 1.14369798e+00
3.59468639e-01 6.96377397e-01 3.58258456e-01 -2.78732687e-01
5.77728450e-01 1.61740020e-01 5.07900596e-01 -6.95708156e-01
-8.20166171e-01 3.52975100e-01 4.22692597e-01 -2.54423499e-01
1.05036266e-01 -1.29605901e+00 -7.60674298e-01 -4.49294806e-01
-5.40334940e-01 2.16246963e-01 6.72537327e-01 1.16858244e+00
-3.46371979e-01 7.93712676e-01 9.58205283e-01 -5.56093633e-01
-1.13829350e+00 -5.92734218e-01 -1.27150273e+00 -2.55659968e-01
-1.93804875e-02 -9.31225836e-01 -3.37472767e-01 -4.04906332e-01] | [4.616764068603516, 4.108184337615967] |
28355508-beb7-4117-8621-7eb07c4e081c | improving-the-gap-in-visual-speech | 2305.14203 | null | https://arxiv.org/abs/2305.14203v1 | https://arxiv.org/pdf/2305.14203v1.pdf | Improving the Gap in Visual Speech Recognition Between Normal and Silent Speech Based on Metric Learning | This paper presents a novel metric learning approach to address the performance gap between normal and silent speech in visual speech recognition (VSR). The difference in lip movements between the two poses a challenge for existing VSR models, which exhibit degraded accuracy when applied to silent speech. To solve this issue and tackle the scarcity of training data for silent speech, we propose to leverage the shared literal content between normal and silent speech and present a metric learning approach based on visemes. Specifically, we aim to map the input of two speech types close to each other in a latent space if they have similar viseme representations. By minimizing the Kullback-Leibler divergence of the predicted viseme probability distributions between and within the two speech types, our model effectively learns and predicts viseme identities. Our evaluation demonstrates that our method improves the accuracy of silent VSR, even when limited training data is available. | ['Shigeo Morishima', 'Qi Feng', 'Keitaro Tanaka', 'Sara Kashiwagi'] | 2023-05-23 | null | null | null | null | ['metric-learning', 'metric-learning', 'visual-speech-recognition'] | ['computer-vision', 'methodology', 'speech'] | [ 4.11507636e-01 4.52517495e-02 -3.60842794e-01 -3.30662996e-01
-1.06150627e+00 -6.38866663e-01 8.04521620e-01 -2.74714679e-01
-2.63639987e-01 4.16909337e-01 5.42340994e-01 -4.02398974e-01
4.78851534e-02 -2.00973704e-01 -5.49737453e-01 -7.10711837e-01
3.26987803e-01 1.77851334e-01 8.55649933e-02 -4.72293720e-02
2.83375323e-01 4.10789460e-01 -1.99998665e+00 4.09381121e-01
8.64538789e-01 1.04627132e+00 5.29289365e-01 8.98637474e-01
-3.64880860e-01 5.85297823e-01 -7.81099141e-01 -2.32155845e-01
3.61577034e-01 -5.50396860e-01 -7.10722387e-01 4.04865630e-02
8.08721125e-01 -4.76469323e-02 -5.71476161e-01 1.11323726e+00
6.20721579e-01 2.67938197e-01 1.08831322e+00 -1.34685981e+00
-6.80020928e-01 4.56407428e-01 -2.75212258e-01 1.84089959e-01
3.88320297e-01 -3.43073234e-02 1.12919474e+00 -8.78746688e-01
5.35446107e-01 1.22655678e+00 3.91027987e-01 1.01964557e+00
-1.27100194e+00 -4.15566295e-01 2.15774551e-01 5.04905164e-01
-1.48824918e+00 -1.02040350e+00 6.61267042e-01 -5.82749844e-01
9.11996067e-01 5.05734205e-01 4.91271555e-01 1.38599122e+00
-3.22413832e-01 1.01357317e+00 1.15444219e+00 -4.06064630e-01
1.78404748e-01 2.48290852e-01 -3.30697417e-01 3.93793404e-01
-4.29771245e-01 3.23310465e-01 -1.08644319e+00 2.53078789e-01
3.03505063e-01 -4.68740016e-01 -4.59109187e-01 -5.73057711e-01
-1.15565681e+00 6.43649518e-01 -6.94938675e-02 3.53727072e-01
5.22215106e-02 -2.26378247e-01 4.58167456e-02 3.31818074e-01
3.70567530e-01 1.73632845e-01 -1.84676081e-01 -6.03691816e-01
-1.10563636e+00 -1.39067113e-01 6.70669913e-01 7.70363152e-01
4.59083557e-01 1.14444263e-01 -2.18756691e-01 1.11541259e+00
4.25129056e-01 7.20348060e-01 7.86781967e-01 -6.42152905e-01
6.22734308e-01 3.04463744e-01 -1.36201233e-01 -8.26950014e-01
7.37823695e-02 2.94132624e-02 -2.33348444e-01 2.98323095e-01
4.83439445e-01 3.40607315e-01 -1.06605923e+00 1.86716163e+00
2.71060206e-02 3.60814691e-01 4.47124124e-01 9.31507111e-01
8.73643100e-01 6.17774606e-01 -1.64043888e-01 -2.61122435e-01
8.37262869e-01 -8.57255280e-01 -6.45540297e-01 -2.58229136e-01
4.17607874e-01 -8.59104097e-01 1.37556028e+00 4.83411178e-02
-1.13153720e+00 -3.30035985e-01 -1.14138544e+00 -4.95525375e-02
-3.84826869e-01 2.47670352e-01 1.36735156e-01 7.97081649e-01
-1.25026131e+00 3.75609845e-01 -7.76341975e-01 -3.71717304e-01
2.98397541e-01 2.49560431e-01 -4.18473899e-01 1.48936182e-01
-8.30002785e-01 9.62299168e-01 7.50999674e-02 -1.64482057e-01
-9.36980426e-01 -6.20424151e-01 -9.24047530e-01 4.70739938e-02
2.32678980e-01 -1.64925501e-01 1.26606715e+00 -1.09708762e+00
-1.85964406e+00 1.12353468e+00 -4.74070191e-01 -3.21215540e-01
5.87810934e-01 3.70116025e-01 -4.50870693e-01 4.15200591e-02
-3.16117764e-01 7.36862600e-01 1.09121168e+00 -1.39096391e+00
-6.72473669e-01 -2.38710269e-01 -2.74101675e-01 5.22145629e-01
-4.34840143e-01 -1.89232093e-03 -3.35111022e-01 -6.77810431e-01
2.19096825e-01 -7.39839375e-01 5.73253274e-01 1.61914513e-01
-2.18016639e-01 -3.41110229e-01 7.26412892e-01 -6.60286009e-01
1.05934262e+00 -2.51290226e+00 5.26302576e-01 2.15322990e-02
1.13202646e-01 4.76446062e-01 -3.41742754e-01 1.89025439e-02
5.74922450e-02 3.48370709e-02 -2.46788368e-01 -6.36489153e-01
1.85123354e-01 2.90223688e-01 -4.62056577e-01 5.20546257e-01
-1.59014314e-01 7.06681907e-01 -8.39284360e-01 -4.37089026e-01
3.17485183e-01 5.46791732e-01 -1.00441568e-01 4.50668365e-01
-4.46501859e-02 2.42628366e-01 2.24472642e-01 4.88151044e-01
6.14849627e-01 1.52545616e-01 1.69575289e-01 -1.19135961e-01
-1.14372872e-01 7.52802610e-01 -9.29147720e-01 1.73971510e+00
-6.07460976e-01 1.16963267e+00 5.09953648e-02 -1.01328933e+00
9.32950675e-01 3.51924390e-01 3.15456361e-01 -1.00430965e+00
-3.84676158e-02 2.54677147e-01 1.82279386e-02 -4.38598841e-01
1.85601816e-01 -2.42545277e-01 3.26479703e-01 2.96443135e-01
8.65620747e-02 -5.31226434e-02 -9.15884227e-02 -7.40494058e-02
6.88150048e-01 -1.23122253e-01 1.25408828e-01 1.63670406e-02
5.25266290e-01 -5.86424649e-01 3.91150117e-01 5.03499568e-01
-4.32364732e-01 7.80007303e-01 2.88440138e-01 9.70839262e-02
-7.78248966e-01 -1.80659914e+00 -1.50118381e-01 1.04448593e+00
3.00728351e-01 -3.35280061e-01 -7.97361493e-01 -6.61124527e-01
3.60879488e-02 8.30529928e-01 -6.08774781e-01 -4.50013071e-01
-2.84505099e-01 -5.88402823e-02 6.74408436e-01 5.07899344e-01
-3.48252356e-02 -8.41151059e-01 -4.18597341e-01 -3.39240819e-01
-3.76617640e-01 -1.18763089e+00 -7.62863219e-01 -6.84971660e-02
-4.17944252e-01 -8.99859905e-01 -8.11597943e-01 -9.71805513e-01
3.72343808e-01 3.26341927e-01 7.45181620e-01 -2.53223270e-01
-2.10609734e-01 5.50187767e-01 -2.65484303e-01 -2.46131346e-01
-9.08039808e-01 -2.81051416e-02 3.44564199e-01 3.25426996e-01
4.46268380e-01 -3.36804688e-01 -4.38244700e-01 5.94845891e-01
-5.01906097e-01 1.29045054e-01 3.27897578e-01 6.53068483e-01
5.77482641e-01 -2.79752821e-01 3.08062941e-01 -1.35582551e-01
5.11509061e-01 -3.35206628e-01 -5.48817396e-01 5.55693150e-01
-5.99640846e-01 2.28648990e-01 3.04918200e-01 -8.14969599e-01
-7.79664695e-01 -1.23127796e-01 -3.46153486e-03 -6.50799096e-01
-2.48845890e-01 -7.10922573e-03 -4.08914715e-01 -1.40904412e-01
4.70502466e-01 6.07323229e-01 3.69831860e-01 -5.82326293e-01
4.54004049e-01 1.41802824e+00 6.38204575e-01 -1.93484470e-01
5.54786623e-01 4.27111745e-01 -3.33740652e-01 -1.26125634e+00
-4.33593243e-01 -5.41432261e-01 -4.06596094e-01 -1.78722322e-01
8.92766535e-01 -5.34124374e-01 -7.12899387e-01 5.02238512e-01
-8.56698334e-01 -5.15023828e-01 -3.96809042e-01 6.94201946e-01
-1.02968299e+00 4.77587909e-01 -1.02987960e-01 -9.31689620e-01
3.07996254e-02 -1.47886872e+00 1.06393361e+00 2.69359183e-02
-1.83462307e-01 -9.57628667e-01 1.22058272e-01 4.10436869e-01
3.28711003e-01 -3.17824990e-01 8.60908210e-01 -7.49198556e-01
-4.71277982e-01 1.15430236e-01 -7.99471438e-02 4.95468915e-01
6.57474041e-01 -8.17011222e-02 -1.17111564e+00 -3.17727685e-01
-8.80657583e-02 -1.76543608e-01 9.52850997e-01 4.05157775e-01
8.81376445e-01 -4.50123549e-01 -3.09901267e-01 8.04335773e-01
6.89111948e-01 2.38682494e-01 4.27344948e-01 4.88366000e-02
6.23225093e-01 9.18992043e-01 4.27930355e-01 6.33196533e-02
3.39350611e-01 1.09461451e+00 2.53724217e-01 2.54498661e-01
-6.90777898e-01 -5.53909898e-01 6.01258874e-01 9.64737892e-01
3.16029221e-01 -5.06286263e-01 -7.80361414e-01 6.98908269e-01
-1.50921023e+00 -9.73341286e-01 6.79358363e-01 2.60996461e+00
8.21981668e-01 -7.87746683e-02 1.64164111e-01 1.79622337e-01
7.23759770e-01 3.38980883e-01 -5.27357817e-01 -4.62337941e-01
-3.74734551e-01 -4.88916524e-02 3.40273052e-01 7.20352411e-01
-8.49444509e-01 1.11116922e+00 6.93027067e+00 9.14843500e-01
-1.53584480e+00 -6.54990831e-03 1.05247132e-01 -4.56709027e-01
-4.64179784e-01 -2.26494268e-01 -5.68858743e-01 5.60644329e-01
1.09209716e+00 -2.32196420e-01 7.77578712e-01 4.62245673e-01
2.40864992e-01 2.64031470e-01 -1.31443632e+00 1.56234312e+00
5.30124247e-01 -1.28515303e+00 1.21288553e-01 3.00261490e-02
4.15455699e-01 1.68967709e-01 6.38813734e-01 -2.61328705e-02
-1.42351955e-01 -1.24174809e+00 8.65251184e-01 5.02573788e-01
1.11899114e+00 -5.19444942e-01 1.52142614e-01 2.79216945e-01
-1.20825481e+00 -3.89073305e-02 -2.79884994e-01 3.50834429e-01
-5.98829798e-02 -3.06360155e-01 -1.18707228e+00 4.61211242e-02
3.75279307e-01 6.55499876e-01 -3.91942620e-01 9.91798460e-01
-1.52462989e-01 5.37076712e-01 -1.67415112e-01 -5.26388362e-02
-1.21885814e-01 -5.81614114e-02 9.74181771e-01 8.77054989e-01
3.32623392e-01 -3.70560110e-01 9.04809404e-03 8.35942209e-01
-2.47000158e-02 2.71123201e-01 -7.56952226e-01 -4.01545763e-01
6.22171223e-01 6.09489501e-01 -2.86283344e-01 -1.01458341e-01
-4.32236880e-01 1.14322615e+00 3.23079884e-01 4.95221525e-01
-4.36086476e-01 -1.36908486e-01 1.15600908e+00 1.53107112e-02
3.18427533e-01 -2.06728294e-01 -1.76557317e-01 -1.05165040e+00
2.12730393e-01 -9.42307949e-01 2.12283731e-01 -6.76966906e-01
-1.11363876e+00 6.04930282e-01 -9.93628502e-02 -1.20529819e+00
-4.61350650e-01 -7.49072194e-01 -2.79963702e-01 9.46932316e-01
-1.55048513e+00 -1.05783570e+00 -2.86165476e-02 6.31168067e-01
8.36940825e-01 -6.39686286e-01 7.42367864e-01 4.34731133e-02
-9.36715230e-02 1.26042485e+00 2.97888637e-01 1.80194397e-02
6.80471241e-01 -1.44152355e+00 5.02875030e-01 7.14061141e-01
7.48770237e-01 4.83820200e-01 5.98670244e-01 -3.94557178e-01
-1.23391044e+00 -6.14537418e-01 8.86834681e-01 -5.07372856e-01
4.79046404e-01 -6.59748018e-01 -8.72906506e-01 3.59536380e-01
-5.43704741e-02 -4.63396274e-02 8.35754752e-01 2.91463658e-02
-6.99007094e-01 -7.15196803e-02 -9.41189408e-01 8.38892817e-01
1.14445567e+00 -1.21179724e+00 -8.09403896e-01 8.60303864e-02
7.20444977e-01 -4.17806447e-01 -5.26351929e-01 1.74687207e-01
6.62960827e-01 -7.66446054e-01 9.23082411e-01 -6.45380676e-01
-1.79622158e-01 -2.72880375e-01 -6.53423905e-01 -1.57547760e+00
3.08446199e-01 -7.26762652e-01 -2.80797064e-01 1.10613120e+00
5.06608069e-01 -5.36609650e-01 7.85464346e-01 1.97257787e-01
-1.33274466e-01 -5.91674745e-01 -1.39274466e+00 -9.42023933e-01
1.43909886e-01 -5.00879526e-01 5.18097341e-01 7.03317046e-01
1.13824353e-01 8.27375054e-02 -2.52927601e-01 2.39593893e-01
4.19235617e-01 1.05176225e-01 4.87587839e-01 -1.05413306e+00
-2.84236193e-01 -8.20782542e-01 -9.21375096e-01 -1.22940373e+00
6.44065320e-01 -1.30696416e+00 2.85723716e-01 -1.44369900e+00
8.03767815e-02 -2.15398818e-01 -3.48551929e-01 3.40259492e-01
-3.19407135e-02 8.03248584e-02 3.38115036e-01 1.60219356e-01
-2.38825023e-01 7.73504496e-01 8.57374072e-01 -3.95997286e-01
-3.35348815e-01 1.75903916e-01 -3.89845729e-01 6.28458977e-01
5.79496264e-01 -2.64551759e-01 -6.69254601e-01 -4.96954054e-01
-1.93675309e-01 2.32331571e-03 1.97413370e-01 -9.06646729e-01
3.11480463e-01 -1.50118977e-01 -2.13412449e-01 -4.11577702e-01
5.76578498e-01 -6.72837853e-01 -3.73443931e-01 8.26123431e-02
-7.14716792e-01 -1.95841312e-01 6.94839880e-02 5.95951200e-01
-2.16613159e-01 -2.99671981e-02 9.03645754e-01 3.95395368e-01
-7.01452732e-01 2.34180897e-01 -4.22209233e-01 2.84063160e-01
9.77060914e-01 -5.55036128e-01 -2.39899680e-01 -5.07617056e-01
-8.16791236e-01 -3.82915810e-02 5.75198948e-01 1.00351894e+00
9.40607846e-01 -1.28933787e+00 -5.53796887e-01 5.32449365e-01
4.91398603e-01 -5.11212707e-01 1.39672145e-01 6.91338301e-01
-6.24007434e-02 3.81398886e-01 -7.16079324e-02 -8.22476447e-01
-1.72546279e+00 5.20150661e-01 6.82904661e-01 5.06769836e-01
-5.27755499e-01 1.06283534e+00 1.81906804e-01 -4.77691323e-01
7.48453736e-01 -2.30147734e-01 -1.91053823e-01 1.95346981e-01
5.75061858e-01 4.03894484e-01 5.01222722e-02 -1.16148961e+00
-4.15068954e-01 5.53765953e-01 -1.34807870e-01 -4.31678891e-01
8.47104788e-01 -3.83891225e-01 4.13924247e-01 8.69132876e-01
1.60876358e+00 2.51542211e-01 -1.50467265e+00 -1.49215296e-01
-1.14405550e-01 -7.94862807e-01 5.59474565e-02 -7.34382689e-01
-7.93042660e-01 1.20684826e+00 1.08278716e+00 1.06079774e-02
8.69807899e-01 3.86882812e-01 7.33957052e-01 3.12776044e-02
1.48620391e-02 -1.13945711e+00 1.29966676e-01 2.71206498e-01
6.68180346e-01 -1.28537095e+00 -6.36893511e-01 -1.54419377e-01
-8.06855977e-01 8.01569819e-01 3.09428751e-01 5.77913821e-01
2.96687722e-01 2.00038552e-01 4.80390251e-01 2.06706896e-01
-7.13425398e-01 -4.74995792e-01 6.75900102e-01 1.05299866e+00
2.68111169e-01 1.70388862e-01 -3.99680734e-02 4.34383154e-02
-5.86413383e-01 -4.98592734e-01 2.24147797e-01 6.89095557e-01
-3.42042685e-01 -1.03749394e+00 4.51325215e-02 2.55259663e-01
-1.98608190e-01 -1.86494365e-01 -5.84712446e-01 3.60037953e-01
-2.08300352e-01 1.04996240e+00 3.57513011e-01 -5.34939766e-01
2.49495074e-01 2.20240563e-01 6.28445208e-01 -5.73869467e-01
1.23982906e-01 -1.40883371e-01 -1.65561795e-01 -5.49961269e-01
-2.08627835e-01 -6.80708110e-01 -1.01169765e+00 8.01121593e-02
-1.45022914e-01 4.93347757e-02 1.04258037e+00 1.01104224e+00
3.16537499e-01 1.45726904e-01 7.55074978e-01 -5.21783113e-01
-9.57007051e-01 -7.18154311e-01 -6.19665384e-01 5.52318394e-01
8.21216941e-01 -8.83198977e-01 -7.77648747e-01 -7.12535828e-02] | [14.33244800567627, 5.054844379425049] |
8f7882e9-9d04-4aa8-9d28-134af752e556 | regularized-weight-aggregation-in-networked | 2301.12617 | null | https://arxiv.org/abs/2301.12617v1 | https://arxiv.org/pdf/2301.12617v1.pdf | Regularized Weight Aggregation in Networked Federated Learning for Glioblastoma Segmentation | In federated learning (FL), the global model at the server requires an efficient mechanism for weight aggregation and a systematic strategy for collaboration selection to manage and optimize communication payload. We introduce a practical and cost-efficient method for regularized weight aggregation and propose a laborsaving technique to select collaborators per round. We illustrate the performance of our method, regularized similarity weight aggregation (RegSimAgg), on the Federated Tumor Segmentation (FeTS) 2022 challenge's federated training (weight aggregation) problem. Our scalable approach is principled, frugal, and suitable for heterogeneous non-IID collaborators. Using FeTS2021 evaluation criterion, our proposed algorithm RegSimAgg stands at 3rd position in the final rankings of FeTS2022 challenge in the weight aggregation task. Our solution is open sourced at: \url{https://github.com/dskhanirfan/FeTS2022} | ['Mojtaba Jafaritadi', 'Suleiman A. Khan', 'Elina Kontio', 'Esa Alhoniemi', 'Mohammad Ayyaz Azeem', 'Muhammad Irfan Khan'] | 2023-01-30 | null | null | null | null | ['tumor-segmentation'] | ['computer-vision'] | [-1.79959044e-01 2.67445028e-01 -3.88771206e-01 -3.83195192e-01
-1.23810506e+00 -2.54460424e-01 5.00218332e-01 2.75379390e-01
-4.38090593e-01 9.10177350e-01 5.47444046e-01 -5.65776944e-01
-6.95852876e-01 -7.26633072e-01 -1.30374566e-01 -9.33613777e-01
-2.15564236e-01 6.96769357e-01 8.10045749e-02 -3.26443389e-02
-1.49712518e-01 5.14715016e-01 -8.41421187e-01 4.53701168e-01
8.77387881e-01 1.14575112e+00 2.20278457e-01 9.12230492e-01
-3.12953383e-01 8.15544248e-01 -5.00906229e-01 -7.60175824e-01
3.76429677e-01 -2.18400493e-01 -1.27001190e+00 -4.03103560e-01
2.54056185e-01 2.11372301e-01 -1.58396125e-01 8.06809008e-01
8.64680648e-01 -1.11134022e-01 4.57130581e-01 -1.32054412e+00
2.39577115e-01 1.11676204e+00 -2.95660943e-01 3.16088021e-01
-6.19164743e-02 1.03584915e-01 1.04986227e+00 -7.89970338e-01
1.03014255e+00 7.40219355e-01 7.98261642e-01 6.97569788e-01
-9.32786584e-01 -7.35662222e-01 8.14382061e-02 2.05165029e-01
-1.08078969e+00 -5.66276193e-01 4.16535169e-01 -3.23471546e-01
6.33344710e-01 9.49377179e-01 6.37844741e-01 1.03565395e+00
1.18034281e-01 1.04892313e+00 1.00775588e+00 -4.20604140e-01
4.71733958e-01 1.06271845e-03 4.50388581e-01 8.53499293e-01
3.65273356e-01 -1.67261899e-01 -1.01497483e+00 -9.10072744e-01
5.74649312e-02 1.23286292e-01 -3.09423596e-01 -2.15258479e-01
-1.36806870e+00 7.40382969e-01 3.65713358e-01 4.92527574e-01
-4.86618787e-01 1.95926964e-01 4.93290097e-01 5.96545637e-01
8.81786942e-01 2.28194818e-01 -6.36349380e-01 -1.39694378e-01
-1.06573987e+00 1.02887452e-01 7.74726570e-01 5.57798386e-01
3.54871243e-01 -7.17006743e-01 -6.14650071e-01 5.88674366e-01
1.26325190e-01 4.76225913e-02 6.71665549e-01 -1.14929938e+00
2.86240339e-01 5.60779572e-01 7.08615035e-02 -2.67020613e-01
-6.99017465e-01 -7.15290785e-01 -1.02911663e+00 1.35187879e-01
6.53815627e-01 -4.11661357e-01 -1.91138580e-01 1.66051757e+00
8.41955721e-01 2.03798428e-01 -7.56434398e-03 7.93965638e-01
9.88649011e-01 -1.64763585e-01 1.87863663e-01 -2.52582848e-01
1.36626911e+00 -1.16788805e+00 -4.39388812e-01 6.63605750e-01
1.22380936e+00 -7.18581140e-01 5.02370298e-01 4.20931578e-01
-1.15336239e+00 1.06975518e-01 -7.49640465e-01 3.74565482e-01
-2.30157137e-01 -9.54624861e-02 1.21855295e+00 7.54308641e-01
-1.37738955e+00 9.36067343e-01 -9.38654959e-01 -4.10843164e-01
7.44172692e-01 5.45635879e-01 -4.99903113e-01 1.07412755e-01
-1.13579822e+00 5.96385717e-01 -7.43473917e-02 -5.11538625e-01
-7.59953916e-01 -1.10235786e+00 -1.27538383e-01 -5.58966771e-02
1.34532139e-01 -1.15746844e+00 1.24960160e+00 -7.36404359e-01
-1.24327505e+00 8.94548953e-01 1.04172938e-01 -6.18555367e-01
9.04670596e-01 3.15900415e-01 -2.64012069e-01 2.52257794e-01
-2.23413080e-01 2.51996517e-01 3.43676239e-01 -5.49752712e-01
-8.31813216e-01 -4.18572515e-01 -4.11072761e-01 2.64897943e-01
-8.96923184e-01 1.94441378e-01 -1.93770558e-01 -6.53063893e-01
-1.64975822e-01 -6.66617811e-01 -7.29483724e-01 4.32199985e-02
-5.11931837e-01 -5.05879104e-01 3.12841386e-01 -3.52334231e-01
1.27652514e+00 -1.86777270e+00 1.22308947e-01 5.71080983e-01
7.69236565e-01 1.30028963e-01 -3.35059464e-01 3.71425688e-01
1.06911466e-01 -4.32847030e-02 1.28247827e-01 -6.63376808e-01
5.01595587e-02 -2.31005237e-01 2.93728024e-01 5.18539310e-01
-6.47576332e-01 8.29661608e-01 -1.13871324e+00 -7.60470569e-01
-2.92562783e-01 2.47487739e-01 -3.49824995e-01 1.65383682e-01
7.77152274e-03 2.82297283e-01 -9.54586685e-01 9.09907043e-01
4.77312386e-01 -5.28397679e-01 4.94095206e-01 -2.23479524e-01
-3.88560481e-02 5.96780665e-02 -9.46854353e-01 2.05179381e+00
-3.01226258e-01 7.41370469e-02 4.33712602e-01 -6.30905807e-01
6.36941552e-01 5.03555179e-01 1.33047700e+00 -1.41881704e-01
2.20031545e-01 4.71024513e-01 -3.55511695e-01 -1.68578893e-01
1.92461148e-01 2.72262067e-01 8.85607898e-02 9.96989429e-01
3.92666578e-01 3.95437479e-01 1.57071769e-01 8.20819616e-01
2.00473642e+00 -5.07086039e-01 2.15074405e-01 -5.20390987e-01
4.40183461e-01 -1.61405802e-01 4.37889427e-01 9.53968465e-01
-5.42811096e-01 2.63987809e-01 2.90547520e-01 -8.46551597e-01
-3.65880400e-01 -7.13481009e-01 -8.12522769e-02 1.30051482e+00
-2.69222468e-01 -6.48180306e-01 -7.75579691e-01 -1.41414201e+00
2.17706755e-01 3.54810655e-01 -7.16179311e-01 1.43919259e-01
-1.80920318e-01 -1.03569067e+00 5.53112268e-01 -1.28618598e-01
3.08096427e-02 -7.55303502e-01 -5.57857394e-01 2.31091052e-01
-2.07898796e-01 -5.14033318e-01 -6.92813516e-01 3.85353744e-01
-8.41641486e-01 -1.40080404e+00 -8.92566562e-01 -2.94038560e-02
7.22550094e-01 2.03207776e-01 1.20111525e+00 3.40077668e-01
-5.53410053e-01 5.94923973e-01 -4.05339181e-01 -5.06573915e-01
-2.28141680e-01 4.59045559e-01 -1.62610814e-01 2.78519928e-01
2.18443781e-01 -4.31990415e-01 -1.10177743e+00 2.64320463e-01
-3.01422566e-01 1.89546108e-01 3.55181366e-01 9.21789110e-01
5.00192106e-01 -5.53179502e-01 6.33612871e-01 -1.32972956e+00
6.47682667e-01 -5.89175045e-01 -3.44130754e-01 5.96999943e-01
-9.23697770e-01 -1.39106154e-01 3.44335616e-01 1.20120883e-01
-8.50652039e-01 2.44827822e-01 -1.82037093e-02 -2.17137709e-01
1.39631853e-01 2.25171596e-01 2.69821197e-01 -5.90938568e-01
9.71494377e-01 -3.62810582e-01 3.41793031e-01 -6.43745065e-01
2.50406772e-01 6.71261609e-01 -1.53848007e-01 -6.78240001e-01
3.91125411e-01 6.11893773e-01 1.54115364e-01 2.11040061e-02
-8.94443870e-01 -4.07548338e-01 -1.24431305e-01 -3.90481710e-01
2.85767466e-01 -7.52385139e-01 -9.62619662e-01 2.31964305e-01
-9.41286445e-01 -4.06364590e-01 -7.11862028e-01 5.22902608e-01
-4.13400888e-01 4.63823564e-02 -6.58417284e-01 -4.81595933e-01
-1.38029099e+00 -8.84761870e-01 7.10251629e-01 4.10008542e-02
-3.48449826e-01 -1.08996689e+00 3.74224693e-01 7.27857232e-01
9.82501090e-01 5.32378137e-01 4.63959605e-01 -1.12683344e+00
-2.53008693e-01 -2.68429786e-01 -7.28414804e-02 -2.62395173e-01
-4.04748395e-02 -1.68117732e-01 -9.06323850e-01 -4.10582244e-01
-4.83068705e-01 -4.91122723e-01 9.25374508e-01 1.88036725e-01
1.49057174e+00 -3.76674175e-01 -7.18416095e-01 5.29673219e-01
1.06744146e+00 -4.80645269e-01 -1.53619081e-01 1.65562227e-01
2.76060224e-01 4.08295184e-01 4.52832550e-01 9.67320621e-01
3.74635130e-01 5.77399909e-01 5.06357670e-01 -2.74603784e-01
-4.81210887e-01 2.98401237e-01 -1.33114578e-02 8.71356070e-01
-3.59047413e-01 -1.87622502e-01 -9.27859128e-01 4.23954338e-01
-2.03639460e+00 -8.15952897e-01 -3.97322327e-02 2.12541556e+00
1.01089382e+00 -2.49304399e-01 1.72738448e-01 -1.66410550e-01
5.79895079e-01 2.94655487e-02 -4.38680410e-01 -2.38870025e-01
-8.53296518e-02 2.25731269e-01 5.55524051e-01 3.94007206e-01
-7.84746706e-01 4.64323550e-01 5.83728647e+00 1.37923229e+00
-6.72638774e-01 7.86736250e-01 1.01028657e+00 -7.52415121e-01
-5.52756548e-01 -2.64402717e-01 -6.50625944e-01 4.64920044e-01
1.19779992e+00 -5.33450484e-01 3.02113652e-01 5.39073467e-01
-4.66232896e-02 1.34517044e-01 -8.67985129e-01 8.61211538e-01
-4.71015900e-01 -1.96361244e+00 -1.70877188e-01 3.07552814e-01
8.02823365e-01 5.77317715e-01 -1.00107856e-01 -1.25904188e-01
9.37003911e-01 -7.38412559e-01 2.99508721e-01 8.06748092e-01
7.15886652e-01 -7.65200675e-01 6.50461555e-01 1.26951918e-01
-7.58173645e-01 -1.81311682e-01 -2.67210364e-01 5.63572764e-01
-6.16969615e-02 1.11948371e+00 -7.01038837e-01 9.38219965e-01
5.41526258e-01 3.86511475e-01 -5.57228386e-01 1.05255318e+00
4.39323068e-01 5.83446026e-01 -3.18112493e-01 -2.02719316e-01
1.26031041e-01 3.24214250e-01 5.83198607e-01 1.18232751e+00
4.89329606e-01 1.55433774e-01 9.99365970e-02 -1.60897911e-01
-4.72066075e-01 4.48135287e-01 -9.15692449e-02 1.64656416e-01
6.44558311e-01 1.95959461e+00 -5.51160634e-01 -3.06550652e-01
-4.80518378e-02 4.23013389e-01 7.08047569e-01 -7.98787922e-02
-4.78216112e-01 -1.92732781e-01 6.28346682e-01 4.40349169e-02
-1.77631021e-01 3.95532459e-01 -3.27397197e-01 -9.37143266e-01
-4.18489635e-01 -7.84035146e-01 1.33364058e+00 -1.54778838e-01
-1.43411744e+00 9.51799035e-01 -3.87072980e-01 -1.28227019e+00
1.01644672e-01 -2.64044732e-01 -8.02407265e-01 5.50545216e-01
-1.26142490e+00 -1.18036497e+00 -4.41656649e-01 9.44171786e-01
1.91817179e-01 -6.83047712e-01 1.24707484e+00 4.58491236e-01
-4.89757627e-01 1.26788986e+00 1.61732137e-01 -2.69336849e-01
7.92559683e-01 -1.19951093e+00 -8.77595469e-02 2.72293717e-01
5.27408309e-02 2.21891612e-01 5.22719681e-01 -3.79953146e-01
-1.46604955e+00 -1.20149159e+00 1.13384855e+00 -3.93050611e-01
7.98099041e-01 -2.00528741e-01 -1.43870488e-01 1.41294613e-01
4.36262816e-01 5.30912220e-01 1.39112735e+00 4.45272744e-01
-1.09605893e-01 -5.75122774e-01 -1.60276401e+00 4.05588716e-01
1.05522633e+00 8.82510617e-02 2.80949354e-01 1.30764234e+00
6.70330942e-01 -3.80776346e-01 -1.21551323e+00 2.70211101e-01
4.20509815e-01 -9.10375297e-01 7.10770488e-01 -8.13300729e-01
-9.97990742e-03 1.79830074e-01 -2.63078306e-02 -1.24695301e+00
-4.84865308e-01 -1.26283932e+00 -4.22731638e-01 7.04329252e-01
7.17545569e-01 -7.87904918e-01 1.22025776e+00 7.68670976e-01
-1.07184298e-01 -1.54593301e+00 -1.39558542e+00 -5.90025783e-01
-9.38613489e-02 -2.26443246e-01 8.28077912e-01 9.91976559e-01
2.73856103e-01 -5.57593644e-01 -1.41999871e-01 -2.23730102e-01
1.08862936e+00 1.36654168e-01 4.61818278e-01 -1.31616271e+00
-5.44419825e-01 -5.13886690e-01 -2.85223812e-01 3.21217673e-03
6.45813644e-02 -1.25220299e+00 -7.17070758e-01 -1.38986659e+00
3.94966334e-01 -7.95873702e-01 -1.18735349e+00 8.28422487e-01
8.75333920e-02 2.57391363e-01 2.18646020e-01 3.66478384e-01
-1.32049251e+00 1.38297990e-01 1.12731349e+00 -1.62794530e-01
2.48972386e-01 2.89231390e-01 -7.98315883e-01 2.95346200e-01
9.28281665e-01 -6.04790807e-01 -2.04532057e-01 -2.72497714e-01
2.40147516e-01 2.99558252e-01 1.83487013e-01 -8.44624937e-01
7.45572150e-01 -2.30527699e-01 3.10322285e-01 -3.81241500e-01
1.07751973e-01 -6.44292176e-01 6.78051472e-01 8.39472592e-01
-6.49152100e-01 -1.46308452e-01 -4.16602135e-01 3.24546635e-01
2.22543553e-01 -2.11052179e-01 5.90731740e-01 -2.22167715e-01
9.77497622e-02 8.23331237e-01 -1.45738184e-01 -5.28748408e-02
1.12346113e+00 2.72739768e-01 -5.59028804e-01 7.34083652e-02
-1.02268720e+00 5.71438074e-01 1.17853060e-01 -5.04619777e-02
2.62307972e-01 -1.14853859e+00 -9.87769783e-01 -1.04094557e-01
-3.75499651e-02 -4.41042870e-01 6.32448494e-01 1.33049881e+00
-1.61138475e-01 1.07936867e-01 -2.25806725e-03 -1.14406273e-01
-1.45118225e+00 2.11868241e-01 5.18131912e-01 -8.96016598e-01
-4.97104347e-01 1.50841379e+00 -3.86879474e-01 -5.90859532e-01
5.09342492e-01 3.50658208e-01 1.52528221e-02 2.03433886e-01
7.12493837e-01 7.35356271e-01 6.26467645e-01 -1.96888540e-02
-4.84640986e-01 -1.21493235e-01 -4.21978652e-01 4.13756892e-02
1.49548185e+00 1.96352333e-01 -1.99711323e-01 -3.10224481e-02
1.00934672e+00 2.62197293e-02 -8.42705965e-01 -4.05687213e-01
7.69892931e-02 -2.85908669e-01 4.07633841e-01 -1.11633885e+00
-1.59949517e+00 1.12613715e-01 7.81771600e-01 1.16995454e-01
1.12073839e+00 6.00536428e-02 7.97207177e-01 3.16439867e-01
5.93027651e-01 -1.07712173e+00 -3.31237942e-01 1.09328605e-01
6.50591075e-01 -9.57862675e-01 2.11356640e-01 -1.69069871e-01
-4.92449641e-01 1.09592187e+00 3.53542030e-01 1.77170098e-01
9.48559821e-01 4.49100584e-01 1.15669072e-01 -4.75718886e-01
-1.53426635e+00 1.35232478e-01 3.76199149e-02 1.17407501e-01
4.62354153e-01 5.19418359e-01 -6.24706328e-01 7.96180308e-01
-2.27972809e-02 1.93481166e-02 9.68264043e-02 7.81996906e-01
-3.45806569e-01 -1.46091139e+00 -1.01865150e-01 8.30605865e-01
-5.34490705e-01 -3.32111269e-02 -4.96548295e-01 7.02842996e-02
1.36914104e-01 6.82843804e-01 -4.53713983e-01 -2.88958728e-01
9.68957767e-02 2.69908682e-02 3.11410666e-01 -2.43363798e-01
-1.39882684e+00 -1.21238731e-01 2.96961665e-01 -1.04021549e+00
-3.59545261e-01 -6.55598342e-01 -9.80358243e-01 -4.67666477e-01
-1.61751151e-01 6.98471844e-01 7.95139432e-01 5.43698251e-01
8.13003540e-01 3.99154574e-01 1.13789487e+00 -3.80604267e-01
-7.80651689e-01 -6.59449875e-01 -6.09464586e-01 2.80914366e-01
-1.25966460e-01 -3.83518875e-01 -3.32942754e-01 -6.60217226e-01] | [6.069920063018799, 6.4277496337890625] |
532eb426-2e8e-41d3-90bb-6c5492edd8e6 | playing-it-safe-information-constrains | 2304.08976 | null | https://arxiv.org/abs/2304.08976v2 | https://arxiv.org/pdf/2304.08976v2.pdf | Playing it safe: information constrains collective betting strategies | Every interaction of a living organism with its environment involves the placement of a bet. Armed with partial knowledge about a stochastic world, the organism must decide its next step or near-term strategy, an act that implicitly or explicitly involves the assumption of a model of the world. Better information about environmental statistics can improve the bet quality, but in practice resources for information gathering are always limited. We argue that theories of optimal inference dictate that ``complex'' models are harder to infer with bounded information and lead to larger prediction errors. Thus, we propose a principle of ``playing it safe'' where, given finite information gathering capacity, biological systems should be biased towards simpler models of the world, and thereby to less risky betting strategies. In the framework of Bayesian inference, we show that there is an optimally safe adaptation strategy determined by the Bayesian prior. We then demonstrate that, in the context of stochastic phenotypic switching by bacteria, implementation of our principle of ``playing it safe'' increases fitness (population growth rate) of the bacterial collective. We suggest that the principle applies broadly to problems of adaptation, learning and evolution, and illuminates the types of environments in which organisms are able to thrive. | ['Vijay Balasubramanian', 'Philipp Fleig'] | 2023-04-18 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 6.18719935e-01 2.27208763e-01 2.63288748e-02 -2.53863316e-02
-8.89003798e-02 -3.51217359e-01 5.95831692e-01 1.45942941e-01
-6.97708666e-01 1.13970876e+00 -2.01461807e-01 -4.72887754e-01
-4.01670039e-01 -1.04526663e+00 -8.35179210e-01 -1.14901996e+00
-5.55701591e-02 5.44412911e-01 2.04591274e-01 -1.79111853e-01
4.56540197e-01 9.70421806e-02 -1.46466160e+00 -2.57213563e-01
4.92931753e-01 6.22340977e-01 5.86291730e-01 1.00193799e+00
1.57423556e-01 5.90710223e-01 -5.79239130e-01 -4.56077814e-01
1.60054669e-01 -7.71645248e-01 -6.82368815e-01 -1.59423918e-01
-4.31159407e-01 1.32281974e-01 2.10160956e-01 1.06244969e+00
2.88696885e-01 1.00583099e-01 8.57264698e-01 -7.19570696e-01
-4.85522330e-01 7.67970502e-01 -2.38541931e-01 7.95133784e-03
1.62066370e-02 6.56014562e-01 8.60547125e-01 -1.24153696e-01
5.55542588e-01 1.38411844e+00 5.81312776e-01 5.53598464e-01
-1.70035541e+00 -2.89218068e-01 2.32720017e-01 -9.14207846e-02
-1.24114120e+00 -6.33272469e-01 4.89011437e-01 -7.06704795e-01
8.14806879e-01 4.55926895e-01 9.68252003e-01 1.09057331e+00
8.37537467e-01 2.37239987e-01 1.09679806e+00 -3.53247285e-01
6.82807505e-01 -4.08424623e-02 -3.10985118e-01 4.49485242e-01
9.20207977e-01 4.77650791e-01 -7.35440493e-01 -4.06412780e-01
6.29180133e-01 -1.30369052e-01 -3.74384731e-01 -2.17806995e-01
-1.01701331e+00 5.33982277e-01 -5.83939627e-03 1.44011885e-01
-4.18016434e-01 4.77747738e-01 3.67804803e-02 2.30938941e-01
2.52734631e-01 9.24145639e-01 -7.55872905e-01 -1.28078863e-01
-3.33265126e-01 2.04515702e-04 1.12979555e+00 3.68953466e-01
7.74492800e-01 -4.23592746e-01 3.89541000e-01 4.99755710e-01
6.60200357e-01 7.05052614e-01 5.60150445e-02 -1.27051950e+00
-1.88564077e-01 1.94467157e-01 5.19998372e-01 -8.37960899e-01
-2.00799018e-01 -5.08303225e-01 -6.40659332e-01 3.00951481e-01
6.95544302e-01 -3.26030433e-01 -4.20815170e-01 2.22904825e+00
2.76571959e-01 -1.22818403e-01 1.18613489e-01 4.14778799e-01
-1.63608611e-01 6.01271033e-01 -6.44004252e-03 -7.53584385e-01
9.95883942e-01 -3.98509204e-02 -5.52229524e-01 -3.07180792e-01
3.32443237e-01 -4.94230807e-01 8.52065206e-01 6.71785712e-01
-1.07298529e+00 -5.09146303e-02 -8.56081247e-01 4.73241568e-01
-1.90519750e-01 -7.79109657e-01 7.58903980e-01 9.02278602e-01
-1.08453906e+00 9.51120079e-01 -1.07838976e+00 -5.56787550e-01
3.73394966e-01 2.19484553e-01 -4.45589563e-03 3.18592727e-01
-8.15480113e-01 1.10894430e+00 3.48401636e-01 7.79388919e-02
-1.00102174e+00 -4.73153472e-01 -3.77477616e-01 -1.58969745e-01
6.34213924e-01 -9.16684449e-01 9.57591355e-01 -9.32574034e-01
-1.60075569e+00 6.61122739e-01 -8.67512301e-02 -2.55824536e-01
4.31276560e-01 -3.28481868e-02 3.32762420e-01 -2.76574820e-01
-2.34896600e-01 3.71274412e-01 5.40585876e-01 -1.23635972e+00
-4.01034117e-01 -5.45415521e-01 4.95006740e-02 1.32530883e-01
-6.12045601e-02 -1.87936828e-01 1.93446368e-01 -2.37331659e-01
7.86241964e-02 -1.05070233e+00 -5.60714543e-01 7.22777247e-02
-3.65616590e-01 -1.86460465e-01 1.06214516e-01 -3.98099780e-01
1.01759505e+00 -1.97788286e+00 3.67763460e-01 1.76576301e-01
3.54443714e-02 -4.00944173e-01 6.50999323e-03 5.07302880e-01
4.34564531e-01 4.95079577e-01 -4.65314865e-01 7.87150040e-02
5.72667718e-02 4.65855062e-01 -4.58790213e-02 6.44758701e-01
1.35099038e-01 3.30605447e-01 -9.88607228e-01 -3.74717444e-01
-2.31680542e-01 3.69186997e-01 -8.27724934e-01 1.80198044e-01
-4.53079015e-01 5.91243327e-01 -5.90724051e-01 3.54681522e-01
2.77767807e-01 -4.29764062e-01 5.60959995e-01 4.80198056e-01
-4.15735215e-01 4.37373310e-01 -1.04112709e+00 1.14164090e+00
-2.66981542e-01 4.60849941e-01 1.50682315e-01 -9.54183877e-01
6.90714061e-01 5.58021432e-03 2.36402050e-01 1.03564776e-01
3.37628305e-01 2.91290790e-01 6.06631517e-01 -3.06457609e-01
-1.45172462e-01 -3.64630878e-01 6.78953677e-02 7.29926288e-01
-2.57115752e-01 -7.13718116e-01 -3.22675295e-02 -1.60221741e-01
1.31889248e+00 2.94219911e-01 7.24063933e-01 -7.49869466e-01
6.65825233e-02 -1.26958892e-01 8.60571027e-01 1.26778269e+00
-3.46270561e-01 -1.20371200e-01 5.68969488e-01 -3.86768490e-01
-1.05273461e+00 -9.85875905e-01 -5.09917080e-01 9.90834236e-01
1.97336420e-01 2.52668485e-02 -7.97197700e-01 5.33227064e-02
-8.40282738e-02 8.64299059e-01 -8.55866253e-01 -1.34536847e-01
-3.31122071e-01 -1.33198893e+00 1.73650369e-01 -1.62501648e-01
2.24594995e-01 -9.30436492e-01 -1.36646318e+00 4.09549475e-01
8.97822157e-02 -2.03412563e-01 -7.31485412e-02 6.79873526e-01
-8.41380835e-01 -9.73531783e-01 -2.90959030e-01 -7.30247283e-03
6.75420403e-01 -1.74616694e-01 1.16672492e+00 3.16412926e-01
-2.49597475e-01 3.10327262e-01 -1.11586653e-01 -5.88667870e-01
-7.46046007e-01 -2.72235811e-01 3.04440320e-01 -3.47660303e-01
2.56984860e-01 -6.64513409e-01 -7.00337410e-01 2.66738832e-01
-7.40961730e-01 1.12140864e-01 5.14085531e-01 1.03748238e+00
4.18711394e-01 2.33778745e-01 2.91709304e-01 -7.05652475e-01
3.09105903e-01 -5.01617312e-01 -7.69576073e-01 4.48341966e-01
-5.22257388e-01 2.77642578e-01 3.67086053e-01 -5.29870510e-01
-1.20633698e+00 -1.56796351e-01 1.38210952e-01 6.35440409e-01
-1.08456947e-01 4.48013067e-01 -2.76127785e-01 9.45017859e-02
7.39767432e-01 1.10967100e-01 3.41468565e-02 -3.33165079e-01
2.41893455e-02 4.41817939e-01 -6.96084127e-02 -1.00840604e+00
3.83712232e-01 5.73030055e-01 4.92255181e-01 -9.46740150e-01
-6.30243063e-01 3.02194387e-01 -4.65123981e-01 -3.91609490e-01
7.53709197e-01 -3.96156847e-01 -1.16513717e+00 4.83826131e-01
-9.96699095e-01 -7.13910222e-01 -4.02014732e-01 4.26266819e-01
-9.59619761e-01 4.48375307e-02 -2.02498212e-01 -1.48840559e+00
3.00250441e-01 -1.12551022e+00 7.72495329e-01 2.10615456e-01
-4.13314581e-01 -9.83713210e-01 5.42073548e-01 6.32242933e-02
5.45768619e-01 2.48149142e-01 9.26351190e-01 -5.69155216e-01
-8.13932717e-01 3.19858462e-01 4.10238951e-01 -3.54348570e-02
3.02878022e-01 4.88129675e-01 -8.88106644e-01 -2.16113493e-01
4.61587876e-01 -1.86957940e-01 9.26024497e-01 6.70642614e-01
7.91292131e-01 -5.96721649e-01 -4.96664166e-01 4.46413428e-01
1.39795470e+00 5.00806451e-01 3.03463876e-01 1.81779772e-01
1.00287661e-01 1.16280198e+00 2.51413643e-01 6.61714494e-01
1.37514204e-01 4.71033454e-01 3.75337481e-01 6.11534774e-01
4.74841565e-01 -5.22158183e-02 4.12314594e-01 9.69851613e-01
-3.19410592e-01 -4.56676692e-01 -9.68801081e-01 3.94291162e-01
-1.69969380e+00 -1.06545532e+00 3.21979940e-01 2.58126783e+00
1.38526678e+00 2.61923164e-01 -6.34416938e-02 -8.75812769e-02
6.31106555e-01 -3.30228537e-01 -9.60953891e-01 -4.91970956e-01
-2.55468845e-01 -3.44813988e-02 5.11036754e-01 9.07212138e-01
-7.43816972e-01 4.56307352e-01 7.31563377e+00 6.29261017e-01
-7.70242751e-01 -6.92725601e-03 1.07065892e+00 5.55903129e-02
-4.92575824e-01 2.47077584e-01 -6.71842515e-01 5.78696609e-01
1.24602294e+00 -2.24578574e-01 7.20068634e-01 4.83578116e-01
2.90422231e-01 -6.11070275e-01 -1.41489029e+00 2.38662854e-01
-4.72428262e-01 -1.09345102e+00 -2.23596811e-01 3.26782674e-01
6.45479560e-01 -1.57238483e-01 1.25983274e-02 -2.29311973e-01
1.22807169e+00 -1.03065062e+00 7.94393778e-01 8.68117630e-01
1.72765136e-01 -5.35898447e-01 5.11095345e-01 8.25018227e-01
-5.57270885e-01 -8.89171138e-02 -5.02237976e-01 -5.23142099e-01
-4.77290899e-02 6.97622478e-01 -7.89888322e-01 -1.54195905e-01
6.65336967e-01 3.55020821e-01 -2.05302969e-01 8.32905531e-01
-1.58709452e-01 5.59906542e-01 -6.68751061e-01 -5.79486072e-01
-3.27683568e-01 -3.53716612e-01 6.32737279e-01 8.97026241e-01
4.91526395e-01 3.55517954e-01 -7.51275197e-02 1.01576722e+00
4.29280400e-01 -8.89820009e-02 -7.05380559e-01 -1.62175626e-01
7.51150548e-01 4.33643997e-01 -9.03765023e-01 -2.20703453e-01
6.99679926e-02 3.40610802e-01 1.22383177e-01 2.25631550e-01
-3.36415887e-01 5.07128388e-02 7.33018637e-01 -3.28680277e-01
1.38320044e-01 -1.34636432e-01 -4.88188952e-01 -8.62065732e-01
-5.39380252e-01 -8.21634233e-01 3.54794979e-01 -4.29865092e-01
-1.36272144e+00 -9.47963446e-02 1.52947484e-02 -1.96937308e-01
-1.51065141e-01 -7.28761733e-01 -3.51460367e-01 6.94498003e-01
-1.00734770e+00 -2.15612605e-01 2.81735718e-01 2.75828000e-02
4.67324406e-01 2.47286707e-01 7.08524346e-01 -6.45975530e-01
-6.03497624e-01 6.53018616e-03 4.76020217e-01 -6.29122138e-01
2.44498119e-01 -1.21546566e+00 1.01837628e-01 6.75314546e-01
-2.15444669e-01 8.83760393e-01 1.30133379e+00 -9.34986413e-01
-1.67611730e+00 -7.12214351e-01 6.19964421e-01 -7.90301383e-01
8.81552994e-01 -2.99475163e-01 -7.20593393e-01 3.90396178e-01
1.10455535e-01 -7.14316368e-01 9.65075195e-01 1.79866150e-01
-7.90532604e-02 -3.56732830e-02 -1.29826164e+00 9.37891364e-01
1.07657039e+00 -2.51966178e-01 -4.31312472e-01 3.95178288e-01
7.06180274e-01 3.51510197e-01 -8.36901248e-01 3.43535691e-01
8.37696612e-01 -9.75151002e-01 7.78346896e-01 -5.95332742e-01
1.65674701e-01 -3.22648913e-01 -5.14897645e-01 -1.20339990e+00
-3.79483968e-01 -9.64071393e-01 2.43523702e-01 6.92349911e-01
5.11297584e-01 -9.74747598e-01 3.86756361e-01 5.92850268e-01
3.36204797e-01 -5.87176025e-01 -1.12138188e+00 -8.63880932e-01
4.81493592e-01 -1.42820820e-01 4.95410502e-01 7.01534510e-01
9.64692309e-02 -8.63360018e-02 -1.54149160e-01 1.65473327e-01
9.96537566e-01 -1.84115320e-01 5.98507762e-01 -1.47643244e+00
-8.59163761e-01 -6.14581645e-01 -1.32911280e-01 -8.10698569e-01
-2.57021546e-01 -1.43800557e-01 7.70961344e-01 -8.35994601e-01
6.69313908e-01 -4.30144042e-01 -3.97064209e-01 4.39313008e-03
-2.22057715e-01 -5.14910594e-02 -1.40913054e-01 2.70445883e-01
-4.57323909e-01 1.89632803e-01 1.10260284e+00 1.36381000e-01
-6.10255264e-02 -1.98462635e-01 -7.38564849e-01 9.33344722e-01
7.73558676e-01 -6.14769518e-01 -2.39654481e-01 -1.42247960e-01
9.20775354e-01 7.78817236e-02 1.95064873e-01 -5.49374461e-01
8.94355699e-02 -9.60814953e-01 1.45449474e-01 1.86454616e-02
4.10540789e-01 -5.48375666e-01 8.45836401e-01 1.15923083e+00
-5.55397153e-01 -2.88272649e-01 -2.38965437e-01 7.04936028e-01
6.33482397e-01 -4.35714930e-01 9.88466501e-01 -4.70148832e-01
1.35076359e-01 -1.15686305e-01 -8.66145849e-01 4.08780761e-02
9.05984461e-01 -1.84166491e-01 -3.91923875e-01 -2.31661081e-01
-6.73943460e-01 -2.56913411e-03 1.06791472e+00 -3.77611488e-01
2.53278017e-01 -6.02396250e-01 -7.50802636e-01 -1.32821411e-01
-2.89491415e-01 -2.06344292e-01 -1.70485988e-01 7.75242925e-01
-6.83217227e-01 3.85718703e-01 -5.60227968e-02 -5.45757711e-01
-8.48877430e-01 5.01834989e-01 4.74560976e-01 -2.07412592e-03
4.25008051e-02 1.26397932e+00 6.82126343e-01 -7.88305979e-03
-1.13065138e-01 -1.87001318e-01 1.50994524e-01 -1.13693625e-01
4.00423259e-01 3.35739374e-01 -5.42247474e-01 -3.29101413e-01
-4.22453910e-01 4.17975932e-01 1.58415213e-01 -3.23048562e-01
1.37872326e+00 -3.79750520e-01 -5.37236691e-01 9.24911439e-01
3.35459709e-01 3.12517509e-02 -1.64819717e+00 -3.63344164e-03
1.00687899e-01 -7.41972864e-01 -1.28789410e-01 -7.43304908e-01
-3.07352185e-01 6.76636100e-01 1.81569561e-01 7.33110547e-01
9.12665963e-01 1.30018041e-01 -6.02411516e-02 9.68648493e-01
8.10060441e-01 -1.04573679e+00 2.54746340e-02 2.32767001e-01
7.34046698e-01 -9.47820544e-01 2.64104679e-02 -4.52603363e-02
-7.37646967e-02 8.67408872e-01 1.70672327e-01 8.73954073e-02
7.98423946e-01 4.44230616e-01 -2.96442449e-01 1.49772530e-02
-1.35003078e+00 -2.49013044e-02 -2.76316494e-01 6.91305757e-01
2.56058425e-01 3.37842762e-01 -2.69425809e-01 1.65960819e-01
-2.11441800e-01 -2.79003501e-01 5.95783055e-01 7.77158797e-01
-1.06383824e+00 -1.05307245e+00 -6.88438892e-01 3.79084438e-01
-5.23027837e-01 -1.11458279e-01 -6.42569900e-01 5.05124688e-01
3.99060428e-01 1.05236483e+00 2.61437092e-02 1.15690574e-01
-3.40421230e-01 6.46692142e-02 6.79318428e-01 -5.94599485e-01
-2.00179026e-01 1.45815864e-01 9.13521051e-02 -3.82476628e-01
-4.76667166e-01 -1.31635666e+00 -6.47178531e-01 -5.97298205e-01
-5.29247820e-01 2.44300410e-01 5.74612916e-01 9.55340564e-01
1.26882285e-01 3.49952102e-01 4.50541645e-01 -6.60343945e-01
-6.82676375e-01 -7.09512115e-01 -5.30179918e-01 -4.85376567e-02
3.10079724e-01 -6.89795792e-01 -7.15326071e-01 2.77851433e-01] | [6.000269889831543, 4.249303817749023] |
0e0157ae-2e91-4226-b137-be3122ef5ba4 | uncertainty-quantification-in-imaging-and | 2004.00227 | null | https://arxiv.org/abs/2004.00227v3 | https://arxiv.org/pdf/2004.00227v3.pdf | Uncertainty quantification in imaging and automatic horizon tracking: a Bayesian deep-prior based approach | In inverse problems, uncertainty quantification (UQ) deals with a probabilistic description of the solution nonuniqueness and data noise sensitivity. Setting seismic imaging into a Bayesian framework allows for a principled way of studying uncertainty by solving for the model posterior distribution. Imaging, however, typically constitutes only the first stage of a sequential workflow, and UQ becomes even more relevant when applied to subsequent tasks that are highly sensitive to the inversion outcome. In this paper, we focus on how UQ trickles down to horizon tracking for the determination of stratigraphic models and investigate its sensitivity with respect to the imaging result. As such, the main contribution of this work consists in a data-guided approach to horizon tracking uncertainty analysis. This work is fundamentally based on a special reparameterization of reflectivity, known as "deep prior". Feasible models are restricted to the output of a convolutional neural network with a fixed input, while weights and biases are Gaussian random variables. Given a deep prior model, the network parameters are sampled from the posterior distribution via a Markov chain Monte Carlo method, from which the conditional mean and point-wise standard deviation of the inferred reflectivities are approximated. For each sample of the posterior distribution, a reflectivity is generated, and the horizons are tracked automatically. In this way, uncertainty on model parameters naturally translates to horizon tracking. As part of the validation for the proposed approach, we verified that the estimated confidence intervals for the horizon tracking coincide with geologically complex regions, such as faults. | ['Gabrio Rizzuti', 'Felix J. Herrmann', 'Ali Siahkoohi'] | 2020-04-01 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 2.65288740e-01 1.97690830e-01 4.17800367e-01 -3.50638062e-01
-8.72437656e-01 -4.63724613e-01 7.17407405e-01 -3.55866030e-02
-5.90368390e-01 8.34626079e-01 7.10488260e-02 -2.21614406e-01
-5.00184000e-01 -1.19922197e+00 -8.58452380e-01 -9.34613764e-01
-1.45621672e-01 6.77216411e-01 2.13621780e-01 1.82317972e-01
4.63525116e-01 7.25580812e-01 -1.44430649e+00 -1.00516409e-01
7.65789628e-01 1.10559642e+00 4.86818776e-02 4.32005256e-01
-6.12037592e-02 4.93878096e-01 -4.69497889e-01 6.38228804e-02
-3.48045602e-02 -1.86056510e-01 -7.32715964e-01 -1.25883475e-01
6.06723167e-02 -5.79132378e-01 8.58519748e-02 1.18866420e+00
3.26952696e-01 2.98951954e-01 1.05185962e+00 -5.92197776e-01
5.06808572e-02 8.01882148e-01 -3.24075878e-01 -8.09206907e-03
-1.58451404e-02 6.35817368e-03 7.79982626e-01 -9.89888787e-01
3.48503709e-01 1.00171530e+00 7.40075886e-01 -4.48133089e-02
-1.34487820e+00 -1.32282719e-01 -5.92528954e-02 -4.40693386e-02
-1.56029463e+00 -4.49013352e-01 6.64847493e-01 -8.39974880e-01
5.26990294e-01 7.38613168e-03 6.06776834e-01 8.60461175e-01
1.59911275e-01 4.15699393e-01 1.19641340e+00 -4.42992449e-01
7.90050983e-01 -5.94652332e-02 1.97912440e-01 3.16382110e-01
3.59577715e-01 2.92307973e-01 -4.60559756e-01 -4.41703834e-02
8.15106094e-01 -3.46447259e-01 -4.82104123e-01 -2.61293232e-01
-8.62969577e-01 8.39841783e-01 3.44058752e-01 1.20498136e-01
-4.38190997e-01 3.79219860e-01 5.49376346e-02 -2.19734445e-01
4.08053815e-01 2.56734669e-01 7.17071518e-02 -1.34407701e-02
-1.32107806e+00 2.84034401e-01 8.40491176e-01 4.26347226e-01
9.21936333e-01 2.65009075e-01 -1.41147316e-01 3.96239221e-01
7.60976434e-01 9.45821583e-01 -2.11482942e-02 -1.02302372e+00
2.60854159e-02 5.74779138e-02 4.86662358e-01 -7.22216964e-01
-4.21810061e-01 -4.65428233e-01 -7.27060497e-01 6.60270631e-01
7.84115255e-01 -1.93085194e-01 -8.64159167e-01 1.42244458e+00
2.22188085e-01 8.64777863e-02 2.59031594e-01 9.47977245e-01
4.53069478e-01 6.25912070e-01 -2.05543146e-01 -9.25597623e-02
1.34810948e+00 -1.23303242e-01 -5.16836882e-01 -2.93309629e-01
1.51995376e-01 -4.57783312e-01 7.17797220e-01 4.01742488e-01
-7.87141323e-01 -2.19606444e-01 -1.03337550e+00 4.36464936e-01
-4.30617370e-02 7.76135698e-02 3.02340418e-01 6.03809655e-01
-8.19700420e-01 9.33258772e-01 -1.04942322e+00 -6.09814599e-02
1.41819462e-01 -2.94907275e-03 6.55930489e-02 1.18941955e-01
-1.45020950e+00 1.04296064e+00 6.60318255e-01 8.60538900e-01
-1.25397074e+00 -5.71351051e-01 -8.16985607e-01 2.65475303e-01
2.82951802e-01 -3.79976988e-01 1.02936029e+00 -5.00252604e-01
-1.65699661e+00 4.92330998e-01 1.23272963e-01 -5.44684112e-01
9.42329645e-01 -2.37070873e-01 -7.45651424e-02 1.79355010e-01
-2.80350223e-02 2.15030640e-01 1.10364330e+00 -1.25484931e+00
-1.75365403e-01 -6.44387081e-02 -1.18150651e-01 -2.76362807e-01
1.07102744e-01 -2.50317365e-01 -3.36910486e-01 -8.34890530e-02
2.97155321e-01 -5.91686130e-01 -3.22145969e-01 -7.59785026e-02
-2.83840358e-01 3.78189594e-01 1.09544262e-01 -7.00399458e-01
1.01388180e+00 -2.03771782e+00 9.99236107e-02 7.83310056e-01
-6.50923476e-02 -2.31635183e-01 4.74911928e-01 3.91453207e-01
1.77821502e-01 -9.67409238e-02 -9.87592757e-01 -1.41317904e-01
7.77733028e-02 6.64527714e-02 -3.94815356e-01 6.86719358e-01
3.11542392e-01 3.02656680e-01 -7.41451621e-01 -3.07383597e-01
5.04265010e-01 5.56142569e-01 -2.59451210e-01 2.01214090e-01
-4.04898942e-01 5.80139399e-01 -4.52524900e-01 3.54920238e-01
1.09254503e+00 -5.12017310e-02 1.77280307e-01 -3.27397794e-01
-4.66689557e-01 -7.18296766e-02 -1.53197026e+00 1.32059145e+00
-6.69633806e-01 3.79294097e-01 1.05500534e-01 -8.02657187e-01
1.25095701e+00 1.20377183e-01 7.31733516e-02 -2.47967958e-01
2.84776717e-01 4.59794551e-01 -1.03322856e-01 -5.36493421e-01
5.28036356e-01 -2.28810847e-01 9.40118134e-02 4.40203130e-01
-1.29804984e-01 -4.62106615e-01 -1.45563856e-02 -1.70217287e-02
6.26124322e-01 5.87181389e-01 1.42881483e-01 -5.30279100e-01
4.69140977e-01 -1.17141470e-01 2.40992635e-01 9.67225611e-01
2.06600636e-01 1.00940323e+00 7.89410114e-01 -2.13597625e-01
-1.05627215e+00 -1.22125375e+00 -6.31320596e-01 1.64027154e-01
-1.38125330e-01 2.55247951e-01 -7.54379630e-01 -1.24282055e-01
-1.99548975e-02 8.61091256e-01 -8.32567453e-01 -1.10155031e-01
-1.76080361e-01 -1.13471532e+00 5.03215730e-01 3.55428934e-01
5.50288796e-01 -9.31381524e-01 -9.90004778e-01 3.70533556e-01
2.51637176e-02 -8.32654238e-01 3.57131094e-01 3.64111602e-01
-1.01409781e+00 -8.35324347e-01 -8.07290375e-01 3.04117203e-01
4.54973847e-01 -5.24198949e-01 9.04423594e-01 -4.36859846e-01
5.79284951e-02 2.29333282e-01 -1.57369673e-01 -1.53026685e-01
-4.27312881e-01 -1.22482046e-01 -2.38989159e-01 3.84512544e-01
-1.51928380e-01 -6.94526136e-01 -5.89677632e-01 7.58665279e-02
-1.01997960e+00 -2.42668301e-01 3.93129349e-01 6.87313855e-01
4.32748348e-01 1.45607973e-02 2.41938710e-01 -7.48841465e-01
4.53505099e-01 -5.90442657e-01 -1.17532825e+00 2.43028954e-01
-4.99944270e-01 4.34245825e-01 2.06333011e-01 -1.20185852e-01
-1.55166554e+00 2.74175592e-02 -3.16979945e-01 -2.56582141e-01
-1.67650461e-01 1.01303017e+00 -1.12092987e-01 5.57124615e-02
9.31618571e-01 -4.18919697e-02 5.53557798e-02 -3.55677664e-01
2.90224224e-01 5.32893419e-01 6.58704817e-01 -9.33992326e-01
5.55826962e-01 7.53302753e-01 2.92456836e-01 -9.53542352e-01
-8.13073695e-01 -1.26124099e-01 -5.79664409e-01 -7.41268873e-01
7.16919422e-01 -6.71721280e-01 -6.79058671e-01 6.41407430e-01
-1.05398881e+00 -3.30320686e-01 -3.98203373e-01 8.13837528e-01
-4.14755523e-01 4.97129589e-01 -3.18621516e-01 -1.45655012e+00
-2.61168480e-01 -1.13274360e+00 9.50036228e-01 2.57097811e-01
-1.08338855e-01 -1.05854726e+00 7.32293427e-02 -6.84676394e-02
4.33295280e-01 3.91381979e-01 5.85473955e-01 -1.51859805e-01
-6.54887199e-01 -3.51164371e-01 -2.87147880e-01 4.82464403e-01
-3.49858165e-01 3.57925773e-01 -1.54100919e+00 7.15429559e-02
1.66029766e-01 -5.00818230e-02 1.17218888e+00 8.05860400e-01
7.78712511e-01 2.34468415e-01 6.61312267e-02 6.11282706e-01
1.69534469e+00 -1.20570697e-01 9.00356889e-01 5.50379932e-01
2.78567255e-01 6.97035253e-01 5.03837585e-01 8.85467649e-01
-1.68696314e-01 3.04893702e-01 8.97200048e-01 2.98318148e-01
2.40368620e-01 9.60196331e-02 1.61432832e-01 1.73327774e-01
-4.09241319e-01 -1.05922826e-01 -1.13468826e+00 5.19791722e-01
-1.68221664e+00 -8.67065728e-01 -4.23669159e-01 2.67476463e+00
5.58648586e-01 3.48480284e-01 -3.81372869e-01 4.51885685e-02
7.21272349e-01 2.07224607e-01 -3.59100789e-01 -2.34665088e-02
-5.69985248e-03 1.43236950e-01 6.80989623e-01 9.28947806e-01
-9.05262351e-01 4.55382705e-01 5.87455845e+00 5.44911861e-01
-1.01243472e+00 -1.41545340e-01 3.02559137e-01 2.94003159e-01
-6.66512907e-01 2.93474704e-01 -7.62735128e-01 4.99456644e-01
1.00126946e+00 2.98945040e-01 3.15756023e-01 5.89272916e-01
4.90840077e-01 -8.04376960e-01 -8.68597984e-01 5.25567949e-01
-3.46968681e-01 -1.38103461e+00 -3.07652414e-01 1.06661789e-01
5.52784443e-01 3.83430682e-02 -1.62773639e-01 -9.79469344e-02
2.07769603e-01 -9.20052826e-01 1.13734639e+00 1.31265557e+00
7.14029789e-01 -7.87324369e-01 1.09938490e+00 1.53074384e-01
-6.81100547e-01 1.46729335e-01 -3.48947078e-01 -1.15154088e-01
5.62543750e-01 1.32229245e+00 -7.62721658e-01 7.00993299e-01
6.97154462e-01 3.06139976e-01 -1.14639916e-01 1.08258426e+00
-4.17662531e-01 8.33411038e-01 -6.77439690e-01 1.77310079e-01
1.88166738e-01 -4.56922919e-01 5.15198052e-01 1.12614608e+00
6.70477152e-01 -2.63583511e-01 -3.07877332e-01 1.37427187e+00
2.93929994e-01 -2.97679335e-01 -4.68583316e-01 1.70174479e-01
4.71035182e-01 1.15521884e+00 -9.50802445e-01 -1.41117185e-01
6.20598458e-02 3.01953912e-01 1.18063062e-01 4.37802255e-01
-6.22222602e-01 -1.72743946e-01 7.08234534e-02 -2.01565586e-02
3.34658831e-01 -1.10891089e-01 -3.97192001e-01 -7.90231645e-01
-8.40986799e-03 -4.23887342e-01 1.88571632e-01 -9.51088488e-01
-9.67468083e-01 3.53991628e-01 4.25957978e-01 -1.18680120e+00
-4.23800141e-01 -7.02513397e-01 -6.92272604e-01 1.36187541e+00
-1.51780677e+00 -7.76333988e-01 -4.45865959e-01 4.62183952e-02
-1.95136711e-01 2.94095635e-01 7.57660925e-01 1.21417586e-02
-3.81226003e-01 -1.13786682e-01 6.17613137e-01 -7.88189545e-02
3.50626826e-01 -1.12709212e+00 9.53114107e-02 1.09230268e+00
-3.60563487e-01 3.43202680e-01 1.12544501e+00 -6.50990069e-01
-8.94229531e-01 -7.32789934e-01 3.72401506e-01 -8.52885693e-02
9.72672820e-01 -4.51921932e-02 -1.21741998e+00 4.65806872e-01
-2.30690867e-01 1.72058582e-01 6.69650435e-02 -5.19889370e-02
1.90249402e-02 -1.75182626e-01 -1.16458356e+00 2.15541482e-01
2.81412750e-01 -6.37664497e-01 -4.39267695e-01 5.24371937e-02
1.66500852e-01 -4.55398649e-01 -1.06191587e+00 5.03437161e-01
7.13590205e-01 -1.14487994e+00 7.53188491e-01 -7.81495422e-02
6.29033566e-01 -5.82927883e-01 -3.76102775e-01 -1.13139343e+00
1.09419465e-01 -2.08523199e-01 -5.23217171e-02 1.22898591e+00
3.42024177e-01 -5.38662553e-01 6.82956338e-01 4.75653857e-01
-2.16492534e-01 -2.68936515e-01 -1.16190767e+00 -5.81795275e-01
8.87943432e-02 -8.04968119e-01 4.59384948e-01 4.53488588e-01
-6.53617203e-01 -2.13179484e-01 -2.94305503e-01 9.32125270e-01
9.43843603e-01 1.77647367e-01 3.15992266e-01 -1.32839537e+00
-4.43755805e-01 -2.96966314e-01 -9.05742049e-02 -7.06736267e-01
-1.67369887e-01 -4.93841320e-01 5.86153746e-01 -1.32920170e+00
-1.80759609e-01 -3.95759016e-01 -2.36681979e-02 -1.12373240e-01
1.50778428e-01 5.33191748e-02 -2.20847473e-01 2.05130070e-01
1.47127613e-01 5.70497096e-01 9.22832072e-01 -5.73868416e-02
6.18926287e-02 2.54907608e-01 1.59915701e-01 9.20310199e-01
7.77574480e-01 -5.03294706e-01 -1.65006980e-01 -5.59450865e-01
6.77669287e-01 4.67556089e-01 6.84430480e-01 -1.12339091e+00
1.13432981e-01 -7.54490495e-02 2.23387972e-01 -6.74817502e-01
2.80281812e-01 -8.24767947e-01 4.89971340e-01 3.81926566e-01
-1.84131786e-02 -6.42480373e-01 -5.85403643e-04 4.93175983e-01
-2.28065446e-01 -1.21158695e+00 9.48878586e-01 -2.27135360e-01
-7.47647643e-01 -1.79598674e-01 -5.44580102e-01 -1.01617232e-01
4.92557406e-01 -1.86551526e-01 5.03494078e-03 -3.06271911e-01
-1.06513262e+00 2.63701137e-02 2.58673608e-01 -5.46097994e-01
5.55095911e-01 -7.77760208e-01 -7.57282317e-01 4.07329090e-02
3.90206948e-02 4.39437002e-01 6.13600850e-01 8.94793868e-01
-7.29937851e-01 -3.64633906e-03 2.66349986e-02 -9.07951295e-01
-4.00536388e-01 1.34095058e-01 8.89194489e-01 -1.45401046e-01
-5.32775700e-01 7.08000958e-01 -2.07806036e-01 -3.66152465e-01
5.69633469e-02 -4.32951301e-01 -4.01751488e-01 4.28051472e-01
4.49531257e-01 5.47683835e-01 1.84228212e-01 -3.38852793e-01
-1.93721488e-01 5.16945064e-01 3.47943485e-01 -5.96690893e-01
1.29858422e+00 -1.38501585e-01 -2.57702917e-01 7.77798295e-01
5.79876602e-01 -1.28939018e-01 -1.63778317e+00 -2.06737742e-01
6.31088093e-02 -2.20556542e-01 3.69337738e-01 -6.56600058e-01
-9.22029734e-01 1.07294369e+00 5.96153796e-01 1.14939265e-01
7.23092735e-01 -9.00475606e-02 -1.41496167e-01 3.63328636e-01
3.38474959e-01 -9.69331741e-01 -4.08754557e-01 6.42523944e-01
8.70684505e-01 -1.05323160e+00 1.69926867e-01 -1.10807493e-01
-3.49054486e-01 1.47224104e+00 1.03105724e-01 -3.88118019e-03
7.43185818e-01 4.03071672e-01 -8.14109072e-02 -2.66908020e-01
-1.70786381e-02 -1.14897192e-01 -1.71710681e-02 4.78854001e-01
2.29852840e-01 3.92992236e-02 -3.52294631e-02 3.97311985e-01
-1.45365521e-01 1.94905281e-01 7.42690563e-01 8.02498817e-01
-7.80780852e-01 -5.35180211e-01 -9.48516130e-01 1.86210737e-01
-1.10264674e-01 -1.45021632e-01 4.05174404e-01 5.69888473e-01
-2.20487323e-02 6.24667227e-01 2.92312026e-01 1.75823823e-01
1.49330646e-01 -6.02867268e-03 3.26235503e-01 -3.19020241e-01
1.96293071e-02 8.05715546e-02 3.40786368e-01 -1.94347158e-01
-3.73485982e-01 -8.27257395e-01 -1.34717143e+00 -7.77529627e-02
-2.65773684e-01 3.63152474e-01 8.15487802e-01 1.02461028e+00
-3.36859345e-01 7.03405619e-01 3.56386602e-01 -1.14428449e+00
-8.09300482e-01 -1.15802300e+00 -8.73617113e-01 -2.89065033e-01
2.64690191e-01 -8.06161106e-01 -7.60207415e-01 -2.03032330e-01] | [6.8192853927612305, 3.3954017162323] |
59622d76-77bb-4152-a052-69d94fa13c39 | on-the-robustness-of-chatgpt-an-adversarial | 2302.12095 | null | https://arxiv.org/abs/2302.12095v4 | https://arxiv.org/pdf/2302.12095v4.pdf | On the Robustness of ChatGPT: An Adversarial and Out-of-distribution Perspective | ChatGPT is a recent chatbot service released by OpenAI and is receiving increasing attention over the past few months. While evaluations of various aspects of ChatGPT have been done, its robustness, i.e., the performance to unexpected inputs, is still unclear to the public. Robustness is of particular concern in responsible AI, especially for safety-critical applications. In this paper, we conduct a thorough evaluation of the robustness of ChatGPT from the adversarial and out-of-distribution (OOD) perspective. To do so, we employ the AdvGLUE and ANLI benchmarks to assess adversarial robustness and the Flipkart review and DDXPlus medical diagnosis datasets for OOD evaluation. We select several popular foundation models as baselines. Results show that ChatGPT shows consistent advantages on most adversarial and OOD classification and translation tasks. However, the absolute performance is far from perfection, which suggests that adversarial and OOD robustness remains a significant threat to foundation models. Moreover, ChatGPT shows astounding performance in understanding dialogue-related texts and we find that it tends to provide informal suggestions for medical tasks instead of definitive answers. Finally, we present in-depth discussions of possible research directions. | ['Xing Xie', 'Yue Zhang', 'Binxin Jiao', 'Xiubo Geng', 'Wei Ye', 'Haojun Huang', 'Linyi Yang', 'Yidong Wang', 'Runkai Zheng', 'Hao Chen', 'Wenxin Hou', 'Xixu Hu', 'Jindong Wang'] | 2023-02-22 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [-1.55107349e-01 4.67731237e-01 3.65967937e-02 -2.69575506e-01
-1.10256457e+00 -8.29250991e-01 7.07535326e-01 4.47943294e-03
-1.87321499e-01 8.70505273e-01 5.46807408e-01 -5.20926356e-01
-3.70211154e-02 -4.43051457e-01 -4.94465262e-01 -4.11071867e-01
5.44337481e-02 6.21408820e-01 -2.60346243e-03 -7.11664975e-01
2.54157912e-02 2.47127898e-02 -5.60682178e-01 4.38899487e-01
9.15228486e-01 5.91933548e-01 -6.02301717e-01 8.68119597e-01
3.15741032e-01 1.49610150e+00 -1.27365029e+00 -1.32754457e+00
1.64726719e-01 -3.44981164e-01 -1.40364897e+00 -4.10189748e-01
2.66758025e-01 -6.48565769e-01 -5.36132336e-01 8.56467664e-01
1.06777143e+00 -4.24143597e-02 5.50870836e-01 -1.61435962e+00
-9.53765988e-01 1.11778724e+00 -7.41469339e-02 3.64958733e-01
6.00721419e-01 7.44920075e-01 1.20766068e+00 -3.25564295e-01
7.22126007e-01 1.40375102e+00 9.73488748e-01 7.46694565e-01
-8.97790432e-01 -5.89130878e-01 -1.18362512e-02 2.16894951e-02
-9.42911923e-01 -4.42709118e-01 3.84042561e-01 -3.29992175e-01
1.09006238e+00 4.88586813e-01 6.57045245e-02 1.84428525e+00
5.31115890e-01 9.96094942e-01 8.62040877e-01 -6.87749982e-02
5.79698682e-02 2.45395288e-01 1.28936954e-02 4.21068460e-01
-2.38629848e-01 -7.67424628e-02 -5.48433125e-01 -5.54033995e-01
1.10216193e-01 -5.95043957e-01 -3.47719848e-01 4.40753698e-01
-1.15703261e+00 9.47402954e-01 2.54344463e-01 2.33425409e-01
-7.32329637e-02 1.38640612e-01 9.50083554e-01 7.09929883e-01
8.31094861e-01 7.52013445e-01 -2.69883335e-01 -7.41801322e-01
-3.02572608e-01 6.34382784e-01 1.25592232e+00 9.74769831e-01
-1.47524387e-01 -8.56855810e-02 -4.00336951e-01 8.77769947e-01
-1.71523303e-01 3.36170703e-01 2.56196767e-01 -1.24650967e+00
9.93158877e-01 1.54295877e-01 -3.67115736e-02 -1.19937265e+00
-2.85377413e-01 -1.21964365e-01 -6.74643636e-01 -4.33427803e-02
6.50960743e-01 -4.81221557e-01 -2.62328953e-01 1.73423374e+00
-3.96981649e-02 -2.82240361e-01 3.70793402e-01 8.04833114e-01
1.12247813e+00 6.18165135e-01 7.37142861e-02 2.07052201e-01
1.08120775e+00 -1.03818464e+00 -8.86904716e-01 -3.98357987e-01
8.06196749e-01 -9.35411930e-01 1.08157527e+00 4.45213556e-01
-1.24266922e+00 1.32180387e-02 -7.57893264e-01 -1.57133177e-01
-2.45962098e-01 -3.88141334e-01 4.89299744e-01 7.30366826e-01
-7.42319405e-01 6.87417388e-01 -6.07580900e-01 -3.13999265e-01
2.29717121e-01 9.08659324e-02 -4.58179623e-01 -1.34606466e-01
-1.80894589e+00 1.50202978e+00 -5.89041822e-02 -9.92546976e-02
-8.00630033e-01 -7.34487951e-01 -6.79493845e-01 -1.98855683e-01
3.59247953e-01 -5.52409530e-01 1.81112421e+00 -5.62607944e-01
-1.53590620e+00 6.40771031e-01 5.84178746e-01 -7.24210083e-01
1.15439689e+00 -3.34543049e-01 -3.90399873e-01 2.01552257e-01
1.13159038e-01 4.60479617e-01 5.71478486e-01 -5.89191437e-01
-1.04312636e-01 2.69632488e-02 4.89549369e-01 3.57853681e-01
-2.48509228e-01 4.01028305e-01 -2.68277466e-01 -8.26567948e-01
-7.08313286e-01 -1.16382420e+00 -6.75562362e-04 -7.42266774e-02
-9.30331767e-01 -3.38586330e-01 6.10361159e-01 -8.51292849e-01
1.18510962e+00 -1.98751724e+00 1.32823316e-02 -2.75877506e-01
2.41455391e-01 3.07353407e-01 -8.44347104e-02 8.04087877e-01
-7.22910762e-02 4.53145444e-01 -2.33772025e-01 -2.61716306e-01
1.38300970e-01 7.35466853e-02 -3.95012826e-01 4.55572695e-01
4.41810876e-01 1.07260132e+00 -9.13877547e-01 -5.10326862e-01
-4.36584242e-02 2.07441315e-01 -7.12050438e-01 2.81936258e-01
-1.30810097e-01 3.50089908e-01 -5.17410576e-01 6.89954996e-01
2.93171883e-01 -6.73371479e-02 -1.53003529e-01 1.99088007e-01
2.82742321e-01 6.64367199e-01 -4.69847769e-01 1.44264054e+00
-2.61805087e-01 8.85968804e-01 2.87947711e-02 -6.03128970e-01
6.04716778e-01 6.36148572e-01 1.78342476e-01 -5.11296511e-01
3.40711743e-01 1.54370949e-01 3.91027868e-01 -5.94871402e-01
6.81197822e-01 -1.60497651e-01 -5.79219162e-01 7.66319752e-01
-4.81493175e-02 -4.30639118e-01 -1.11957155e-01 6.76748276e-01
1.60817599e+00 -3.07405442e-01 8.89729336e-02 -8.21010619e-02
8.76440108e-02 2.18280867e-01 2.50342935e-01 7.69806743e-01
-4.74588275e-01 7.97185779e-01 9.40694749e-01 -2.02261165e-01
-7.35542357e-01 -8.93404841e-01 -4.15983833e-02 9.12318230e-01
-1.43951595e-01 -4.15004969e-01 -1.13301384e+00 -1.13931608e+00
5.73226735e-02 1.00734353e+00 -6.99676573e-01 -4.53796089e-01
-4.16893750e-01 -7.02788651e-01 1.51999569e+00 3.93869400e-01
4.10151929e-01 -1.23373771e+00 -3.53875279e-01 1.47431299e-01
-6.82206154e-01 -1.13216591e+00 -8.02693903e-01 2.46512946e-02
-3.91698927e-01 -1.06811130e+00 -8.24407578e-01 -4.26906943e-01
1.55675784e-01 -1.73472892e-02 1.34903967e+00 1.44878775e-02
-9.32337120e-02 5.35683572e-01 -6.06453001e-01 -6.44476414e-01
-1.18148327e+00 2.27900133e-01 1.02927707e-01 -5.01196384e-01
1.05693802e-01 -3.31953138e-01 -3.17568302e-01 6.50528371e-01
-7.29779363e-01 -3.03974837e-01 3.13074261e-01 9.38355803e-01
-2.60729909e-01 -2.33088240e-01 9.16582823e-01 -1.16706932e+00
1.29327655e+00 -7.87754893e-01 4.24344577e-02 7.49213845e-02
-3.75564516e-01 -2.67675549e-01 6.81784093e-01 -4.38227564e-01
-9.60648477e-01 -7.09915698e-01 -5.40243566e-01 -3.59121710e-01
5.53838238e-02 4.82293069e-01 -5.78685999e-02 1.13659136e-01
1.02163517e+00 -2.66253293e-01 2.64509887e-01 -3.43478471e-01
3.93405795e-01 9.49748278e-01 5.74763775e-01 -6.33854866e-01
7.20233440e-01 -1.65496767e-02 -7.70014405e-01 -6.29313231e-01
-6.13832593e-01 1.32611781e-01 8.91280770e-02 -1.80744737e-01
8.51184905e-01 -6.96542144e-01 -8.89284432e-01 6.07086420e-01
-1.35456944e+00 -6.72015786e-01 -1.44167081e-01 2.21262559e-01
-4.69256103e-01 5.23229182e-01 -1.18858385e+00 -5.71790457e-01
-6.24397337e-01 -1.41401505e+00 7.16936231e-01 -2.12449759e-01
-9.22108173e-01 -1.15089715e+00 9.55052823e-02 7.29187727e-01
3.46360505e-01 3.46372098e-01 8.54461968e-01 -1.09341025e+00
-8.51924345e-02 -6.09526157e-01 7.07630739e-02 6.08874500e-01
5.63881174e-02 -5.43136895e-03 -1.04392004e+00 -6.36364147e-02
1.26558781e-01 -8.31138194e-01 4.03514653e-01 3.05265072e-03
8.29079866e-01 -7.13264108e-01 1.74877383e-02 1.90657198e-01
5.23636818e-01 1.18613996e-01 6.86903119e-01 3.79891306e-01
3.86605531e-01 7.70738900e-01 8.30668688e-01 4.62860286e-01
4.03574109e-01 5.55251241e-01 4.90434736e-01 2.06668451e-01
-8.43255371e-02 -2.63104916e-01 6.97936833e-01 8.15068364e-01
2.45545626e-01 -8.49509418e-01 -1.06872630e+00 4.20181245e-01
-1.67155957e+00 -1.09875035e+00 -3.29360515e-02 1.67671156e+00
1.14346647e+00 4.30390060e-01 1.41225472e-01 5.42889349e-02
6.09946966e-01 1.53479099e-01 -6.62353754e-01 -9.62785900e-01
-1.23294458e-01 -2.81661183e-01 2.05952168e-01 3.50978702e-01
-8.38743329e-01 7.72991776e-01 7.10772181e+00 7.64725685e-01
-8.62342894e-01 2.32036769e-01 1.00074875e+00 4.87032682e-02
-2.27248028e-01 -3.40099514e-01 -3.18545163e-01 6.40811265e-01
8.87970328e-01 -3.72122973e-01 3.15560818e-01 9.29078341e-01
-7.45891780e-02 5.21422997e-02 -1.25610328e+00 5.63627660e-01
4.53243218e-02 -1.20480549e+00 -2.40297779e-01 2.77897250e-02
5.73627174e-01 1.98444590e-01 3.34019691e-01 4.79784846e-01
8.07849884e-01 -1.15919304e+00 8.33768845e-01 1.87415127e-02
6.19332552e-01 -7.58386791e-01 1.05534470e+00 5.51355600e-01
-3.37423444e-01 4.99207638e-02 -2.85032183e-01 -8.66943151e-02
1.21047921e-01 2.32304201e-01 -1.10847425e+00 4.97525364e-01
7.42877841e-01 5.89051783e-01 -4.00031477e-01 4.92735535e-01
-2.73326665e-01 8.24234843e-01 -1.38205484e-01 -9.52520669e-02
4.73749816e-01 2.30225444e-01 7.01842308e-01 1.33161855e+00
-4.52230275e-02 2.00636387e-01 -1.37538269e-01 8.42885494e-01
-4.35250014e-01 -1.11519158e-01 -9.07513976e-01 -2.91260242e-01
4.67866153e-01 9.64819551e-01 -1.49123847e-01 -4.48946580e-02
-3.16357344e-01 1.03213716e+00 2.34645858e-01 -9.39478949e-02
-1.14230371e+00 -5.52057683e-01 8.74050617e-01 -3.17126572e-01
-3.14417452e-01 2.63013691e-01 -2.29767323e-01 -9.28559184e-01
3.26221995e-02 -1.51105356e+00 5.31154990e-01 -8.30597878e-01
-1.63734806e+00 8.12580585e-01 -1.28995985e-01 -1.29085243e+00
-4.99302030e-01 -3.77210438e-01 -8.95276904e-01 8.19817424e-01
-9.27365601e-01 -8.83508742e-01 7.45528489e-02 5.76106668e-01
6.02170050e-01 -1.90303013e-01 9.63473737e-01 1.62879646e-01
-6.62670732e-01 1.20817435e+00 -6.27911910e-02 5.92666209e-01
1.13453364e+00 -1.12729871e+00 9.70794797e-01 5.85228682e-01
-2.98738897e-01 6.07325613e-01 1.08017838e+00 -7.32867420e-01
-1.39463186e+00 -1.04852581e+00 8.18409085e-01 -1.19758642e+00
9.15617049e-01 -2.10066870e-01 -8.17094147e-01 8.02840292e-01
5.28031528e-01 -2.97455430e-01 5.71184993e-01 4.02352912e-03
-4.31422621e-01 2.43087426e-01 -1.35342932e+00 7.67522216e-01
7.77253687e-01 -6.73305154e-01 -8.65706682e-01 7.19996750e-01
1.11281586e+00 -7.11327791e-01 -1.10072482e+00 -1.52825168e-03
4.55540955e-01 -9.36300278e-01 8.50582778e-01 -7.24115551e-01
9.68141854e-01 3.81618857e-01 1.81038044e-02 -1.64641118e+00
-9.31035727e-03 -1.09920442e+00 2.84490496e-01 1.55602705e+00
7.07129717e-01 -7.68334329e-01 6.14812255e-01 1.09742403e+00
-3.28647703e-01 -6.03772819e-01 -9.05627251e-01 -9.03818429e-01
4.97851163e-01 -6.62217259e-01 3.57392222e-01 1.22791541e+00
5.83722770e-01 4.67214108e-01 -8.38209569e-01 -1.21382862e-01
1.97249144e-01 -4.72947031e-01 9.47714627e-01 -6.27096593e-01
-4.41714913e-01 -3.16391975e-01 -1.96458489e-01 -5.38461149e-01
2.83178568e-01 -9.68479693e-01 8.45833495e-02 -1.43482554e+00
-6.04506321e-02 -3.75350893e-01 1.45206705e-01 6.81612313e-01
-3.30657870e-01 2.43740261e-01 3.17367733e-01 2.03767180e-01
-4.37843293e-01 4.25791711e-01 1.14680362e+00 -4.75196391e-01
8.90194345e-03 3.08695078e-01 -1.03272855e+00 6.00053012e-01
1.01881981e+00 -6.26964271e-01 -3.67292821e-01 -6.11273646e-01
1.20004579e-01 2.92488426e-01 2.66202211e-01 -6.04763329e-01
-6.92965370e-03 -5.81652410e-02 -2.89226055e-01 -1.25615701e-01
3.42786163e-01 -4.77225870e-01 -1.84676170e-01 5.78198910e-01
-7.32627809e-01 3.14671665e-01 2.35503554e-01 3.85509610e-01
-8.52834061e-02 -2.38192648e-01 5.57729542e-01 -1.01603903e-01
1.46350292e-02 8.16662386e-02 -6.64592743e-01 7.26702929e-01
7.04017162e-01 1.17960416e-01 -7.76709437e-01 -1.01766205e+00
-3.78720194e-01 4.96588469e-01 2.61656612e-01 6.67804599e-01
3.86734188e-01 -1.01134491e+00 -1.06681383e+00 -3.90812606e-01
2.37686217e-01 -2.92131484e-01 2.95975298e-01 9.17243659e-01
-5.55606782e-01 3.38006198e-01 4.37565893e-03 -2.23647088e-01
-1.30388772e+00 4.58579510e-01 3.33094627e-01 -3.51818919e-01
-7.01078653e-01 1.01189089e+00 9.57022682e-02 -7.35579610e-01
4.61735219e-01 -2.13123769e-01 1.17872596e-01 -8.63021910e-02
6.15839362e-01 5.56695640e-01 1.83831513e-01 -4.38254625e-01
-3.73997062e-01 -2.34320223e-01 -3.11768174e-01 -1.61614805e-01
1.14007151e+00 1.24785736e-01 2.58401521e-02 3.36350054e-01
1.17769516e+00 4.66755778e-02 -8.17193985e-01 -6.28331453e-02
-7.15087205e-02 -2.01380789e-01 -4.80481088e-01 -1.26558363e+00
-8.77594829e-01 1.00193226e+00 -2.88164755e-03 4.47258592e-01
5.83394110e-01 1.98672581e-02 9.60379899e-01 5.87265551e-01
1.56480804e-01 -1.03241599e+00 1.99629053e-01 6.89015865e-01
1.32755971e+00 -1.11181653e+00 -2.55720820e-02 -2.92667657e-01
-1.29425728e+00 8.26330304e-01 6.94715023e-01 1.52559102e-01
2.88340867e-01 3.15869600e-01 3.85227948e-01 4.49865079e-03
-1.03270578e+00 5.19322634e-01 -1.69380605e-02 6.22787595e-01
4.22901899e-01 4.70750360e-03 2.63571180e-02 7.51781762e-01
-6.72217846e-01 -3.53798240e-01 8.16743076e-01 7.74733067e-01
1.49369091e-01 -9.85472798e-01 -4.03460830e-01 2.77939826e-01
-9.41418588e-01 -2.42998600e-01 -9.39006329e-01 9.58364189e-01
-4.42806512e-01 1.29410923e+00 -3.95608723e-01 -7.40116954e-01
5.22811830e-01 6.75226003e-02 -2.64816731e-02 -5.30660212e-01
-1.34835756e+00 -4.12315518e-01 8.08101714e-01 -4.65100318e-01
1.35568082e-01 -5.64356804e-01 -8.79506707e-01 -9.24467921e-01
-4.39882070e-01 3.84885609e-01 3.75071168e-01 8.51271093e-01
2.59308606e-01 5.19114494e-01 5.39581835e-01 -5.03495157e-01
-1.07240021e+00 -1.08796453e+00 -2.02838913e-01 3.66235852e-01
1.47487223e-01 -2.49098018e-01 -4.80104715e-01 -4.07137185e-01] | [6.221107482910156, 8.151312828063965] |
4bae97a0-39b9-4e39-9944-0747e1d4db18 | fuxi-a-cascade-machine-learning-forecasting | 2306.12873 | null | https://arxiv.org/abs/2306.12873v2 | https://arxiv.org/pdf/2306.12873v2.pdf | FuXi: A cascade machine learning forecasting system for 15-day global weather forecast | Over the past few years, due to the rapid development of machine learning (ML) models for weather forecasting, state-of-the-art ML models have shown superior performance compared to the European Centre for Medium-Range Weather Forecasts (ECMWF)'s high-resolution forecast (HRES) in 10-day forecasts at a spatial resolution of 0.25 degree. However, the challenge remains to perform comparably to the ECMWF ensemble mean (EM) in 15-day forecasts. Previous studies have demonstrated the importance of mitigating the accumulation of forecast errors for effective long-term forecasts. Despite numerous efforts to reduce accumulation errors, including autoregressive multi-time step loss, using a single model is found to be insufficient to achieve optimal performance in both short and long lead times. Therefore, we present FuXi, a cascaded ML weather forecasting system that provides 15-day global forecasts with a temporal resolution of 6 hours and a spatial resolution of 0.25 degree. FuXi is developed using 39 years of the ECMWF ERA5 reanalysis dataset. The performance evaluation, based on latitude-weighted root mean square error (RMSE) and anomaly correlation coefficient (ACC), demonstrates that FuXi has comparable forecast performance to ECMWF EM in 15-day forecasts, making FuXi the first ML-based weather forecasting system to accomplish this achievement. | ['Hao Li', 'Yuan Qi', 'Yinghui Xu', 'Yuan Cheng', 'Feng Zhang', 'Xiaohui Zhong', 'Lei Chen'] | 2023-06-22 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [-5.03339529e-01 -4.14303750e-01 2.16680676e-01 -3.58833760e-01
-5.22145271e-01 -4.97610927e-01 1.00264728e+00 -1.11926096e-02
-1.76286265e-01 1.17903841e+00 1.74246997e-01 -1.10864127e+00
-2.41991088e-01 -9.80224371e-01 -2.61582946e-03 -8.35025847e-01
-6.10171199e-01 4.78253551e-02 -1.64390057e-01 -6.94761932e-01
2.34282613e-01 7.78603613e-01 -1.44072330e+00 -2.40284443e-01
1.17323029e+00 9.65286791e-01 -5.88206984e-02 8.95442307e-01
-1.81354582e-01 5.14498532e-01 -4.99270976e-01 3.17073971e-01
4.78412926e-01 -2.92328089e-01 -3.61450195e-01 -3.43177080e-01
2.70532608e-01 -3.70714396e-01 2.78540462e-01 3.01350087e-01
6.47001565e-01 5.82920492e-01 5.73965371e-01 -9.15539861e-01
2.02125101e-03 -2.97209751e-02 -3.38436604e-01 4.61711228e-01
-9.36864167e-02 1.63200889e-02 7.29153514e-01 -1.08143342e+00
1.59071013e-01 9.90821362e-01 9.83721018e-01 -9.52515081e-02
-1.33711135e+00 -7.01741219e-01 5.43193743e-02 -4.89127487e-01
-1.27383721e+00 -3.76380265e-01 2.20158294e-01 -1.03897607e+00
1.51419830e+00 2.69549400e-01 6.03178382e-01 3.18805426e-01
1.11636996e+00 -3.11466366e-01 1.41223383e+00 -4.83715028e-01
3.08673859e-01 -6.86863661e-02 -3.54171102e-03 5.10675982e-02
2.45448142e-01 9.18728471e-01 -1.38669774e-01 -2.57799685e-01
4.75106627e-01 -1.19255260e-01 -2.81107664e-01 4.49603677e-01
-9.20091867e-01 1.10460854e+00 3.68958801e-01 4.46145713e-01
-8.27408135e-01 -4.37616557e-01 1.93545893e-01 6.55778289e-01
1.30138576e+00 6.37949288e-01 -1.03292644e+00 1.25528337e-03
-1.52959645e+00 7.33880281e-01 7.90382981e-01 1.05515309e-01
4.05310333e-01 1.01830840e+00 2.65456021e-01 5.49915493e-01
4.42964226e-01 1.08991671e+00 1.50810152e-01 -7.20432997e-01
2.19881043e-01 1.13942124e-01 7.56267905e-01 -1.24293995e+00
-7.08207905e-01 -9.34234083e-01 -1.18793154e+00 8.68000507e-01
1.78880170e-01 -8.66175234e-01 -8.11591983e-01 1.15547681e+00
7.47024342e-02 9.33139548e-02 5.51315427e-01 8.53941083e-01
1.95731461e-01 1.44727683e+00 2.54434019e-01 -5.75722575e-01
7.31774628e-01 -6.75122321e-01 -9.85025883e-01 -3.76250148e-01
1.03390253e+00 -1.00984550e+00 3.94798338e-01 1.36376008e-01
-5.77088654e-01 -7.82862842e-01 -8.69604945e-01 7.44573295e-01
-8.97629023e-01 8.12042430e-02 4.87394750e-01 3.51767480e-01
-1.14661980e+00 8.16901743e-01 -6.59766436e-01 -2.39545509e-01
-5.48529327e-01 -1.45695806e-01 -2.53540784e-01 4.72782761e-01
-1.74043536e+00 1.43058574e+00 4.12576765e-01 6.49586141e-01
-3.28151375e-01 -1.12589216e+00 -9.12508547e-01 1.17880560e-01
-5.03851116e-01 -4.23447758e-01 1.07464004e+00 -7.07856476e-01
-1.47287428e+00 1.15894876e-01 -3.43227714e-01 -6.71372592e-01
2.76295394e-01 -4.46467102e-01 -1.19647634e+00 -3.65592092e-01
-4.86676730e-02 1.75891936e-01 5.70985436e-01 -1.02881598e+00
-1.11933422e+00 -1.17929637e-01 -4.56425756e-01 9.12725627e-02
4.35395926e-01 1.99955091e-01 7.67035246e-01 -8.58371675e-01
6.10374734e-02 -9.13366914e-01 -5.65736532e-01 -6.49139762e-01
3.73288691e-01 1.20731048e-01 8.08116913e-01 -1.24623883e+00
1.59396458e+00 -1.92268801e+00 -3.92437488e-01 3.38427424e-01
-4.24207687e-01 4.77580935e-01 1.49216861e-01 8.05456340e-01
-2.73125917e-01 1.75546989e-01 -4.42321390e-01 -2.73254156e-01
-3.37186903e-01 4.88801450e-01 -1.29201186e+00 4.92185891e-01
2.88509399e-01 2.79642701e-01 -5.81671298e-01 3.37799728e-01
7.34376788e-01 4.41046357e-01 5.75749539e-02 4.48623657e-01
2.09871586e-02 7.18956709e-01 1.36237681e-01 1.93909749e-01
1.18020546e+00 2.11029336e-01 -1.24512710e-01 5.72901428e-01
-9.04423952e-01 2.28023827e-01 -1.16831410e+00 7.57271945e-01
-6.00146651e-01 7.13085771e-01 2.70710230e-01 -4.11449552e-01
1.32865536e+00 6.61507010e-01 9.97304395e-02 -6.84315503e-01
-3.89786243e-01 6.10010207e-01 1.10676385e-01 -2.81615555e-01
4.89005536e-01 -4.34325457e-01 3.06864619e-01 2.61471361e-01
-3.22914898e-01 -3.40100706e-01 7.53399432e-02 -1.52995840e-01
1.23685028e-03 9.00860652e-02 2.63786525e-01 -8.34174037e-01
5.29375255e-01 8.77272412e-02 6.61163330e-01 5.45221686e-01
2.00877175e-01 3.35534155e-01 -1.14410464e-03 -9.51839685e-01
-8.81816268e-01 -6.42573118e-01 -7.39332914e-01 9.27080274e-01
-6.98759437e-01 -2.70123452e-01 4.56999838e-02 1.03761849e-03
3.81348521e-01 1.22841811e+00 -4.63424742e-01 4.89240378e-01
-1.98334545e-01 -1.28705168e+00 3.51193845e-01 3.61089766e-01
4.92725909e-01 -7.09467888e-01 -6.80141389e-01 7.19398022e-01
2.53600299e-01 -6.50339663e-01 5.58741093e-01 1.15724273e-01
-1.22539151e+00 -4.05217469e-01 -8.07193041e-01 4.41093221e-02
2.91780561e-01 5.26609793e-02 1.18690026e+00 -1.74311623e-01
3.92341226e-01 -1.51646072e-02 -2.60858625e-01 -5.36674261e-01
-1.86509788e-01 -8.73603299e-02 2.62061149e-01 -2.39791974e-01
9.72543284e-03 -6.24364853e-01 -4.00705099e-01 1.92841873e-01
-5.64009964e-01 1.91336088e-02 2.18525991e-01 8.07913244e-01
1.01848409e-01 1.29100531e-01 8.74300122e-01 -2.76710629e-01
5.79199016e-01 -7.17248499e-01 -1.28872168e+00 3.52894664e-02
-1.27261174e+00 -3.66493404e-01 5.85570037e-01 3.27005386e-01
-1.50042284e+00 -4.05033320e-01 -3.26588184e-01 1.38785467e-01
-5.80516756e-01 1.31269741e+00 8.23056161e-01 -2.88963243e-02
4.58230674e-01 3.16903234e-01 -8.89501721e-02 -7.23549366e-01
9.08738375e-02 5.62045515e-01 3.92353803e-01 -2.26110652e-01
7.57354796e-01 2.48966198e-02 3.16329956e-01 -1.01710188e+00
-7.30346262e-01 -5.03121614e-01 -7.07062900e-01 -2.87894994e-01
8.45840871e-01 -1.30052042e+00 -8.53855256e-03 7.40419745e-01
-8.87893677e-01 -5.32050073e-01 3.92170250e-01 8.98521841e-01
5.82421720e-02 -2.18251720e-01 -3.39595467e-01 -1.33833194e+00
-6.59948707e-01 -6.15216434e-01 5.85005701e-01 1.55027494e-01
-3.07757586e-01 -1.57272911e+00 6.73524976e-01 -4.97632891e-01
1.23275936e+00 8.42294455e-01 6.85623109e-01 -1.71558470e-01
1.07841687e-02 -4.52144712e-01 -9.55610201e-02 4.22050387e-01
2.24537879e-01 4.30912167e-01 -9.99082685e-01 -4.26306039e-01
-1.91308856e-01 1.94125473e-01 8.41592371e-01 7.24928141e-01
3.31640810e-01 -3.25377524e-01 -3.52622271e-02 7.38592684e-01
1.59039176e+00 3.99630606e-01 3.98641557e-01 6.09145820e-01
1.18908055e-01 6.32535219e-01 7.32597768e-01 5.84009290e-01
8.68213847e-02 1.83275133e-01 1.52043253e-01 -5.58653533e-01
5.56467772e-01 2.71761656e-01 7.07096010e-02 6.85342550e-01
-5.85770428e-01 -5.93465008e-03 -1.53239441e+00 4.57762927e-01
-1.61694920e+00 -1.06486011e+00 -4.98897016e-01 2.27805328e+00
3.41663599e-01 2.45236054e-01 -4.52222258e-01 6.29300028e-02
4.46359068e-02 7.69592285e-01 1.77456826e-01 -1.01620102e+00
-2.18141705e-01 8.41377154e-02 8.43567550e-01 1.05107903e+00
-1.27694666e+00 6.24137700e-01 6.78318882e+00 7.04348907e-02
-1.54593158e+00 -6.96346983e-02 5.20957351e-01 3.18959564e-01
-3.46449204e-02 5.71051575e-02 -1.16974545e+00 4.74655598e-01
1.83549094e+00 -2.64966935e-01 1.57982901e-01 7.53553212e-01
9.50805902e-01 -4.07597035e-01 -1.22782521e-01 2.42172912e-01
-5.10433197e-01 -1.51567888e+00 -3.68989259e-02 1.05174922e-03
1.21903455e+00 2.90644437e-01 -2.22012535e-01 6.60376012e-01
2.31212482e-01 -1.15524185e+00 2.90109277e-01 9.71816480e-01
7.10272491e-01 -1.11633146e+00 1.36482167e+00 5.53466797e-01
-1.42239988e+00 -7.78846666e-02 -4.63071555e-01 -9.40300286e-01
3.70304376e-01 1.17865062e+00 -4.90532637e-01 9.78415370e-01
8.32473516e-01 6.62653804e-01 -8.17456916e-02 8.66016567e-01
-8.32954943e-02 1.07883608e+00 -3.56479853e-01 7.45544612e-01
8.21385622e-01 -4.28399682e-01 5.61240435e-01 1.28615999e+00
7.47181356e-01 4.96864587e-01 2.11239889e-01 9.60332602e-02
7.28712738e-01 1.11237176e-01 -7.70319164e-01 3.07775557e-01
3.04089248e-01 8.98978591e-01 -5.26595023e-03 -4.66732562e-01
-5.60685456e-01 2.20658287e-01 -3.44620138e-01 7.35103428e-01
-2.84558088e-01 -5.07450461e-01 1.18271410e+00 9.00564492e-02
1.80792809e-01 -7.72680521e-01 -5.42204261e-01 -8.36543143e-01
-3.87995839e-01 -6.62127852e-01 2.42131501e-01 -8.48164141e-01
-9.32788849e-01 8.16131473e-01 4.21557166e-02 -1.31260407e+00
-7.96883345e-01 -4.97834027e-01 -1.08616805e+00 2.01784801e+00
-2.06311989e+00 -8.36947560e-01 -1.33219576e-02 3.49856634e-03
3.20948452e-01 -2.49131471e-01 1.52577138e+00 -8.70301053e-02
-2.87904650e-01 -2.45491922e-01 9.47254479e-01 -2.28825733e-01
7.07938552e-01 -1.28611052e+00 6.54870033e-01 9.75467741e-01
-8.09342980e-01 6.43665314e-01 9.71253455e-01 -9.02898371e-01
-8.70137334e-01 -1.42072809e+00 1.70506930e+00 1.50073878e-02
6.28196895e-01 1.89030349e-01 -1.24119246e+00 8.22313130e-01
4.28410381e-01 2.52902247e-02 5.71611583e-01 3.42772543e-01
6.31179735e-02 -4.16509360e-01 -8.78359795e-01 2.70278662e-01
-1.93539023e-01 -3.13409179e-01 -7.75399625e-01 1.45549297e-01
3.06733429e-01 -3.23171496e-01 -1.37571001e+00 1.15763164e+00
7.24259973e-01 -8.43251824e-01 4.73750472e-01 -6.64426386e-01
2.26922616e-01 -4.79625672e-01 -1.66127369e-01 -2.11686182e+00
-6.85855746e-01 -4.35206175e-01 7.20166638e-02 9.78092730e-01
6.15505219e-01 -1.29979360e+00 -8.57499391e-02 6.13080919e-01
-6.87883198e-02 -5.74910164e-01 -9.48100746e-01 -8.99094284e-01
5.11746347e-01 -4.38796103e-01 6.13970876e-01 1.17935097e+00
-4.77460772e-01 -1.15449779e-01 -6.47804320e-01 8.76659393e-01
4.24316138e-01 4.73356485e-01 6.45753562e-01 -1.74830949e+00
2.28030547e-01 -4.51836705e-01 1.41204163e-01 -3.56002957e-01
5.14876917e-02 -1.34118468e-01 -1.02063514e-01 -1.37614739e+00
-8.65306199e-01 -3.20349842e-01 -3.37249607e-01 3.94880086e-01
-2.09960282e-01 -2.52028525e-01 -5.45782223e-02 1.91026181e-01
6.43555224e-01 5.69724679e-01 7.43668616e-01 4.11973149e-01
-3.85725021e-01 2.32703686e-01 2.75135249e-01 5.26513457e-01
1.05617857e+00 -2.48911723e-01 -2.25040987e-02 -3.68811756e-01
3.07250202e-01 3.96989554e-01 -3.26882303e-02 -1.25722301e+00
-1.97403848e-01 -5.52437425e-01 9.26375091e-01 -1.09329581e+00
-7.13776946e-02 -6.28392160e-01 5.70693612e-01 5.33740878e-01
1.80775188e-02 5.24448812e-01 8.89567316e-01 2.49116067e-02
-6.80485964e-01 2.58611649e-01 7.98219383e-01 1.15118235e-01
-9.86106038e-01 1.08024180e-01 -9.50241208e-01 -6.20768428e-01
8.95703793e-01 2.73063928e-01 -2.86316127e-01 -4.03073370e-01
-8.23073864e-01 5.47796309e-01 1.92086875e-01 1.86040893e-01
1.79682106e-01 -8.73930275e-01 -1.18937743e+00 6.14409268e-01
-2.74077982e-01 -4.60686445e-01 3.25728476e-01 6.92945659e-01
-6.80767298e-01 1.00767696e+00 -2.97773033e-01 -3.02534789e-01
-8.30565453e-01 1.47250399e-01 8.53392661e-01 -4.30168599e-01
-5.73199272e-01 6.35220528e-01 -1.62604004e-01 -7.34744310e-01
-3.38433236e-01 -3.26986223e-01 -3.81685168e-01 4.34327126e-01
9.39167798e-01 3.41424823e-01 3.32452804e-01 -7.28951454e-01
-1.68945119e-01 5.99914551e-01 5.48207283e-01 -2.66215414e-01
1.32124722e+00 -3.30578506e-01 -1.03631072e-01 8.42779338e-01
7.52097666e-01 -3.80173512e-02 -1.32468450e+00 7.77250156e-02
8.18442181e-02 -5.15804946e-01 7.77330756e-01 -1.12401628e+00
-7.80176640e-01 9.15326655e-01 6.48446202e-01 2.44435772e-01
9.39514816e-01 -1.04177547e+00 6.73869252e-01 2.88625062e-01
-9.18556452e-02 -8.30646515e-01 -1.17866015e+00 1.23276758e+00
1.31428826e+00 -1.30414665e+00 1.35420144e-01 2.49375284e-01
-4.40522283e-01 1.39816463e+00 1.32448688e-01 1.17447032e-02
1.27164459e+00 3.49184036e-01 6.32399857e-01 2.14544564e-01
-9.85912561e-01 -2.28128918e-02 3.86251986e-01 9.06566996e-03
7.16786444e-01 4.65232879e-01 -4.38121319e-01 1.85922056e-01
-2.26195559e-01 3.15703377e-02 1.87864080e-01 9.97117341e-01
-7.15820849e-01 -4.85701352e-01 -9.09979403e-01 4.41462487e-01
-5.60671926e-01 -3.31521600e-01 5.07915080e-01 6.37842000e-01
-1.23525798e-01 1.12118554e+00 3.94569188e-01 2.69751512e-02
-1.31108621e-02 5.97398460e-01 -5.82620740e-01 -1.90219715e-01
-6.10229254e-01 1.87500894e-01 3.83334070e-01 -3.33897531e-01
-3.86300802e-01 -5.18586457e-01 -9.31132436e-01 -7.90706754e-01
-5.87182306e-02 6.77634656e-01 7.67986476e-01 1.20799160e+00
3.75768960e-01 3.95273089e-01 9.32432771e-01 -1.34007168e+00
-4.70938146e-01 -1.34863842e+00 -9.82745290e-01 -4.02015507e-01
8.25566769e-01 -6.71193182e-01 -8.65255654e-01 -2.02250600e-01] | [6.518044948577881, 2.93401837348938] |
ce97253a-70a1-4a1a-a64b-fb17a33d2d05 | how-good-are-variational-autoencoders-at | 2304.10767 | null | https://arxiv.org/abs/2304.10767v1 | https://arxiv.org/pdf/2304.10767v1.pdf | How good are variational autoencoders at transfer learning? | Variational autoencoders (VAEs) are used for transfer learning across various research domains such as music generation or medical image analysis. However, there is no principled way to assess before transfer which components to retrain or whether transfer learning is likely to help on a target task. We propose to explore this question through the lens of representational similarity. Specifically, using Centred Kernel Alignment (CKA) to evaluate the similarity of VAEs trained on different datasets, we show that encoders' representations are generic but decoders' specific. Based on these insights, we discuss the implications for selecting which components of a VAE to retrain and propose a method to visually assess whether transfer learning is likely to help on classification tasks. | ['Marek Grzes', 'Lisa Bonheme'] | 2023-04-21 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 4.49752748e-01 2.17330977e-01 -5.80029935e-02 -1.19556025e-01
-7.44821370e-01 -5.51846862e-01 6.20971560e-01 -2.56135702e-01
-3.64018559e-01 5.91546893e-01 5.30979693e-01 -4.45231766e-01
-2.66130805e-01 -6.03186190e-01 -9.82780874e-01 -6.23764694e-01
2.70077825e-01 2.57792383e-01 -1.42020255e-01 -3.36098932e-02
2.32694775e-01 2.93846637e-01 -1.28309715e+00 5.43105721e-01
7.23053813e-01 5.74526250e-01 4.50141966e-01 4.16729689e-01
1.11725226e-01 6.26634359e-01 -7.00608790e-01 -4.25863385e-01
8.79710726e-03 -8.17761183e-01 -1.02211010e+00 -2.88806856e-02
5.06718993e-01 -2.33611286e-01 -2.30342790e-01 9.95286047e-01
5.41638613e-01 2.74147630e-01 1.40555620e+00 -1.05490375e+00
-9.68050241e-01 5.30982435e-01 -1.69475421e-01 3.88966084e-01
-2.40796767e-02 3.71945789e-03 1.08179188e+00 -8.53740811e-01
8.78806651e-01 1.22384679e+00 5.32046497e-01 8.77067387e-01
-1.45905614e+00 -5.83829999e-01 -1.79468393e-01 3.27009887e-01
-1.14373493e+00 -5.92848480e-01 8.95249248e-01 -7.15389788e-01
5.81322551e-01 6.91440701e-02 4.18045968e-01 1.58102560e+00
2.42583901e-01 8.27849984e-01 1.12820780e+00 -4.05497283e-01
2.73868263e-01 4.26123828e-01 -1.95719555e-01 5.43464959e-01
1.94416866e-01 7.78279454e-02 -6.29685640e-01 -2.10354980e-02
9.56824243e-01 -4.62884754e-01 -3.10966671e-01 -4.52473164e-01
-1.21357429e+00 1.16381681e+00 4.49012816e-01 4.25783932e-01
-4.27178949e-01 9.77791101e-02 4.45438325e-01 3.83705765e-01
3.93910915e-01 7.70646274e-01 -2.73021638e-01 -2.10994966e-02
-8.53505850e-01 3.18585075e-02 1.69125557e-01 5.66370964e-01
6.03741705e-01 3.37448776e-01 -2.94627994e-01 9.01713908e-01
2.31146604e-01 5.04414700e-02 5.83670139e-01 -9.98798072e-01
9.56774652e-02 2.49678895e-01 -2.77455628e-01 -5.56751966e-01
-6.19378686e-02 -3.85868847e-01 -6.51905477e-01 5.29736161e-01
3.00452411e-01 -2.93732762e-01 -8.18613708e-01 1.88079107e+00
-5.76094016e-02 4.14358437e-01 2.03455180e-01 7.23010659e-01
7.97251403e-01 5.13959408e-01 4.37420458e-02 7.98880085e-02
1.23096800e+00 -6.25240326e-01 -5.26758850e-01 -1.59538478e-01
7.07728505e-01 -6.45631969e-01 1.10883164e+00 2.94110119e-01
-1.01709855e+00 -7.93985724e-01 -1.21755600e+00 -1.93170026e-01
-2.06718236e-01 5.93607798e-02 2.73684412e-01 5.86665094e-01
-8.98406148e-01 9.66830492e-01 -7.70910203e-01 -4.30175304e-01
6.11247897e-01 1.39995903e-01 -3.53200257e-01 1.84750795e-01
-9.62808132e-01 1.29783106e+00 4.14734602e-01 -1.74701765e-01
-1.11864293e+00 -6.56915069e-01 -7.69806802e-01 6.59242421e-02
-1.63131192e-01 -9.55767632e-01 1.07630098e+00 -1.29014719e+00
-1.54280913e+00 9.09777641e-01 1.20475935e-02 -3.96693259e-01
3.62254530e-01 -1.18563212e-01 -1.92985862e-01 1.00624330e-01
8.62199441e-02 9.57568169e-01 9.45901394e-01 -1.24509084e+00
-3.80145967e-01 -2.88156152e-01 -8.53793174e-02 3.54817837e-01
-4.40188855e-01 -1.14825793e-01 -5.24057746e-02 -8.15377235e-01
-1.63100570e-01 -1.00214326e+00 2.17308298e-01 9.17442814e-02
-1.11426018e-01 -4.26708877e-01 6.99591339e-01 -8.27582896e-01
7.40251362e-01 -2.31577897e+00 5.76784909e-01 5.68677336e-02
9.49655399e-02 2.27168411e-01 -3.14551532e-01 2.66117811e-01
-1.24924168e-01 9.60561037e-02 -2.51248538e-01 -1.16908215e-01
-3.86878140e-02 2.02385172e-01 -4.47517246e-01 4.73939538e-01
3.58879864e-01 1.00213993e+00 -7.99363434e-01 -3.90587896e-01
1.63364843e-01 7.82404542e-01 -5.47445357e-01 2.44578212e-01
-5.98150715e-02 8.88685286e-01 -2.69512951e-01 2.38067672e-01
2.98409760e-01 -2.59480536e-01 1.38260201e-01 -2.31933236e-01
2.77698249e-01 3.56566191e-01 -6.48824513e-01 1.76132941e+00
-4.08982247e-01 1.20219505e+00 -1.16269134e-01 -1.35952210e+00
7.81815052e-01 3.22315365e-01 1.89780757e-01 -6.72970414e-01
1.98213294e-01 8.76772106e-02 4.20401096e-01 -3.63576710e-01
1.22723319e-01 -4.94867265e-01 2.25057930e-01 5.52662075e-01
4.75212604e-01 1.92625467e-02 -3.40436161e-01 -1.25378743e-02
9.29953516e-01 5.06581783e-01 1.52157158e-01 -2.43038431e-01
1.55648991e-01 1.04647484e-02 4.06251460e-01 5.54860830e-01
-2.08494902e-01 7.00177491e-01 4.16465819e-01 -1.02994703e-01
-1.17724419e+00 -1.27744865e+00 -4.02293116e-01 1.17840457e+00
-2.19300598e-01 -2.82703668e-01 -8.24030101e-01 -4.74940330e-01
-2.40763068e-01 1.09058440e+00 -9.72326756e-01 -4.21532005e-01
-2.65454412e-01 -4.50029641e-01 6.24703467e-01 6.21753991e-01
1.18086942e-01 -1.34687543e+00 -7.70951390e-01 -3.13499868e-02
3.55536565e-02 -8.70431423e-01 -2.65245140e-01 6.11014888e-02
-1.08597815e+00 -9.74381506e-01 -1.07305253e+00 -8.82941902e-01
6.28899932e-01 2.23896280e-01 9.82271969e-01 -2.00891748e-01
-5.49172647e-02 7.80669212e-01 -3.55251819e-01 -5.00629246e-01
-7.50042140e-01 -1.12664411e-02 1.06186591e-01 -1.13805510e-01
1.87310919e-01 -6.69101775e-01 -7.83260047e-01 2.04421580e-01
-8.10471177e-01 1.12206275e-02 7.51140356e-01 9.24911499e-01
4.69038486e-01 -2.51198292e-01 5.56632221e-01 -9.11825001e-01
1.00354314e+00 -5.53693593e-01 -2.24429801e-01 3.64649624e-01
-6.40912950e-01 1.98254064e-01 3.92440856e-01 -5.25498152e-01
-9.87077236e-01 -2.97188491e-01 -7.79390335e-02 -8.18234444e-01
-2.09980205e-01 5.51507235e-01 1.11134656e-01 2.10561901e-01
9.20760572e-01 1.17549017e-01 1.86282516e-01 -3.52542728e-01
4.52892154e-01 5.58152378e-01 5.21725833e-01 -7.65430391e-01
6.92845047e-01 3.84799451e-01 -2.58437008e-01 -8.82798553e-01
-4.23127502e-01 -1.30920902e-01 -6.34509861e-01 -1.30543277e-01
1.28590131e+00 -8.28875124e-01 -4.53588754e-01 -1.18463218e-01
-1.13436985e+00 -4.41796690e-01 -3.22570652e-01 7.78183758e-01
-7.42541313e-01 9.58712623e-02 -2.98695445e-01 -4.31444675e-01
-3.04892868e-01 -1.24853539e+00 7.12855160e-01 1.48904234e-01
-5.76200247e-01 -1.02132761e+00 3.85136992e-01 4.19692695e-01
1.75926834e-01 -5.80549240e-02 1.28792822e+00 -7.53023267e-01
-2.68093199e-01 2.78920412e-01 2.29788641e-03 6.14934862e-01
2.49424756e-01 -1.14135884e-01 -1.20266187e+00 -4.68718052e-01
-1.48149863e-01 -3.17351192e-01 8.34939837e-01 6.64614677e-01
1.08984494e+00 -1.53935729e-02 -3.14807534e-01 7.07723379e-01
1.03028393e+00 2.93338656e-01 9.64895844e-01 3.36241961e-01
6.42609596e-01 8.90501678e-01 1.16681278e-01 -1.38357446e-01
-2.61732005e-02 6.83107853e-01 5.22347018e-02 2.09802076e-01
-2.24328354e-01 -3.62291962e-01 7.25648224e-01 9.76778269e-01
-2.79860437e-01 8.19104314e-02 -6.19292557e-01 6.39150858e-01
-1.54858482e+00 -8.67786527e-01 3.18409413e-01 2.14674997e+00
5.37830889e-01 6.95483461e-02 1.59004971e-01 -2.01997310e-01
6.77786171e-01 3.39122377e-02 -5.73428810e-01 -5.66551805e-01
9.75783449e-03 5.22755504e-01 7.38394856e-02 1.49418518e-01
-7.29258418e-01 8.36149991e-01 6.91699505e+00 7.87346423e-01
-1.00935471e+00 1.73415810e-01 4.26841438e-01 1.31253570e-01
-4.51132327e-01 4.02612016e-02 -2.79979050e-01 3.11534315e-01
1.13698208e+00 -9.09751281e-03 4.02133673e-01 6.08581960e-01
-8.44928250e-03 2.25666791e-01 -1.46539521e+00 8.38661909e-01
6.04440458e-02 -1.55092883e+00 1.88527986e-01 1.10091567e-01
6.81449890e-01 -4.55177128e-02 4.38144326e-01 5.10492265e-01
2.71481067e-01 -1.30967259e+00 4.99332517e-01 5.02908468e-01
8.06776702e-01 -5.57967007e-01 4.17948812e-01 8.88117924e-02
-6.45095110e-01 3.38177979e-01 -5.73474944e-01 1.28328890e-01
-1.63071439e-01 -1.41456023e-01 -1.26000428e+00 1.92486092e-01
5.41643620e-01 5.20432234e-01 -5.39724886e-01 8.77850115e-01
-2.55328864e-01 9.04732466e-01 2.85855591e-01 6.19391873e-02
2.54885480e-03 -2.73926973e-01 5.69914222e-01 9.85048652e-01
5.59777439e-01 -1.49977371e-01 -4.50658709e-01 1.17325473e+00
-9.90602672e-02 1.49448931e-01 -8.08481932e-01 -1.98267326e-01
2.50315845e-01 7.82224476e-01 -7.33534336e-01 -1.94623426e-01
-5.60296416e-01 1.34807897e+00 3.20726603e-01 4.60419118e-01
-5.34715295e-01 -1.62345931e-01 7.39297926e-01 -7.16615915e-02
5.19014835e-01 1.60143897e-02 -2.13813573e-01 -1.14687467e+00
-2.72995502e-01 -9.28904176e-01 2.59572625e-01 -1.11244690e+00
-1.36126161e+00 4.02764052e-01 1.21343240e-01 -1.28325462e+00
-3.71742666e-01 -9.41328108e-01 -9.28997636e-01 8.23314369e-01
-1.07063305e+00 -1.13827646e+00 1.94617420e-01 7.01954842e-01
5.17859876e-01 -4.83474523e-01 1.03519452e+00 -3.04454472e-02
-3.82274061e-01 6.49398208e-01 3.86234254e-01 4.03261095e-01
7.13440001e-01 -1.29258955e+00 1.62825063e-01 4.03635800e-01
7.65371144e-01 5.81840217e-01 8.35919738e-01 -4.57764655e-01
-1.14397490e+00 -6.86501801e-01 2.33196408e-01 -6.30808413e-01
3.49700600e-01 -1.83966905e-01 -1.19693041e+00 1.00851178e+00
5.57471931e-01 -1.58180758e-01 9.89554107e-01 3.99070084e-01
-3.81908417e-01 1.93372011e-01 -8.06686938e-01 6.67044997e-01
7.58070230e-01 -9.77283180e-01 -1.02721381e+00 -1.82586419e-03
5.00477433e-01 -1.27266109e-01 -9.43440616e-01 2.93304652e-01
7.04265535e-01 -8.06542993e-01 9.78715003e-01 -9.54740465e-01
7.68423498e-01 -1.57612073e-03 -2.30785236e-01 -1.74202132e+00
-4.70927775e-01 -1.66728064e-01 1.77520141e-01 1.12871683e+00
4.00076538e-01 -4.49152082e-01 8.46800923e-01 2.21396267e-01
-1.62653521e-01 -4.65439916e-01 -1.02940297e+00 -5.98618567e-01
8.38524580e-01 -4.13899392e-01 9.40978453e-02 1.19168162e+00
-4.11561355e-02 8.40242922e-01 -3.50908637e-01 -4.63695005e-02
3.74560803e-01 -3.61015424e-02 8.10482740e-01 -1.32286930e+00
-3.82686734e-01 -6.06667757e-01 -5.20243466e-01 -5.80777526e-01
3.83957863e-01 -1.29588389e+00 -1.76215410e-01 -1.47861671e+00
2.75810510e-01 7.62405246e-02 -6.83432937e-01 2.36243740e-01
-9.42666009e-02 1.30395472e-01 2.19532922e-01 3.09946567e-01
-1.60029650e-01 6.33077562e-01 1.27382863e+00 -2.33264223e-01
-2.04971284e-01 -8.89746919e-02 -7.78498650e-01 6.30174756e-01
6.94138288e-01 -4.17385221e-01 -6.30672812e-01 -3.05383742e-01
1.33213818e-01 4.93171401e-02 5.27973294e-01 -7.78682113e-01
-9.62423719e-03 -1.35927901e-01 6.27500832e-01 -1.92431584e-01
2.58308768e-01 -4.32644188e-01 8.48984718e-02 3.98103923e-01
-5.17788172e-01 -1.60853520e-01 5.53146720e-01 5.32657206e-01
-1.37560606e-01 -6.57171845e-01 7.84676313e-01 -4.22120430e-02
-7.93865025e-01 4.62217703e-02 -6.30156040e-01 1.47272483e-01
7.95197070e-01 -2.81694114e-01 -6.61367327e-02 -6.38233662e-01
-8.59636903e-01 -2.31350303e-01 3.45482141e-01 3.95156950e-01
7.92070329e-01 -1.47208977e+00 -1.01839817e+00 1.81660254e-03
1.51964366e-01 -4.08314615e-01 3.07519585e-01 5.95789909e-01
-2.93085605e-01 3.17865461e-01 -6.07815504e-01 -4.57572728e-01
-1.19346118e+00 4.93997484e-01 4.04030830e-01 1.34782851e-01
-7.57100463e-01 9.41271305e-01 7.26092577e-01 -2.88415462e-01
3.29480991e-02 -4.16704128e-03 -3.35558653e-01 2.12381318e-01
2.53119648e-01 3.26434582e-01 -7.56192431e-02 -7.28493214e-01
-9.90199670e-02 4.59154993e-01 -1.93697304e-01 -4.41296548e-01
1.44038653e+00 1.26998797e-01 2.17000082e-01 5.89813352e-01
1.29691732e+00 -9.93526056e-02 -1.49745774e+00 -2.11007982e-01
-1.16161704e-01 -1.45130157e-01 3.77881713e-02 -6.32622600e-01
-9.52122986e-01 1.23142374e+00 8.48037422e-01 -9.07333046e-02
8.24075222e-01 2.36548558e-01 2.74476707e-01 2.54027814e-01
-6.90979287e-02 -1.11193681e+00 1.83720276e-01 1.18102670e-01
1.08415556e+00 -1.11855733e+00 -3.34430523e-02 1.71440765e-01
-9.31188822e-01 9.68849897e-01 3.92609954e-01 -3.26255739e-01
5.68446577e-01 -3.39065850e-01 2.58969665e-02 -2.53777921e-01
-6.89936578e-01 -1.42585531e-01 7.74493158e-01 9.19480801e-01
5.57000697e-01 3.97415981e-02 -7.68329799e-02 3.77234817e-01
-3.48664165e-01 -1.65167138e-01 4.99783486e-01 6.39454484e-01
-2.77008951e-01 -1.00509191e+00 -3.76873106e-01 6.12763762e-01
-3.80094200e-01 -3.74457017e-02 -4.17786479e-01 6.68861687e-01
1.36104390e-01 5.30864537e-01 2.12914005e-01 -4.24609095e-01
1.01305835e-01 4.08277661e-01 8.30438852e-01 -7.43069232e-01
-2.02786207e-01 1.78373367e-01 -2.66398519e-01 1.23203732e-02
-5.23286700e-01 -7.64761269e-01 -8.56239855e-01 3.04143014e-03
-2.53642976e-01 1.99761167e-02 5.27230263e-01 1.12731850e+00
2.85919845e-01 7.65932322e-01 1.15970626e-01 -6.87715948e-01
-5.14019132e-01 -1.08206403e+00 -5.83805799e-01 4.19286013e-01
2.43788585e-01 -8.72843862e-01 -2.07011327e-01 1.47691429e-01] | [9.788146018981934, 2.522564172744751] |
206ad65b-d715-446d-9eb8-ac1c85bdd4d1 | neural-voice-puppetry-audio-driven-facial | 1912.05566 | null | https://arxiv.org/abs/1912.05566v2 | https://arxiv.org/pdf/1912.05566v2.pdf | Neural Voice Puppetry: Audio-driven Facial Reenactment | We present Neural Voice Puppetry, a novel approach for audio-driven facial video synthesis. Given an audio sequence of a source person or digital assistant, we generate a photo-realistic output video of a target person that is in sync with the audio of the source input. This audio-driven facial reenactment is driven by a deep neural network that employs a latent 3D face model space. Through the underlying 3D representation, the model inherently learns temporal stability while we leverage neural rendering to generate photo-realistic output frames. Our approach generalizes across different people, allowing us to synthesize videos of a target actor with the voice of any unknown source actor or even synthetic voices that can be generated utilizing standard text-to-speech approaches. Neural Voice Puppetry has a variety of use-cases, including audio-driven video avatars, video dubbing, and text-driven video synthesis of a talking head. We demonstrate the capabilities of our method in a series of audio- and text-based puppetry examples, including comparisons to state-of-the-art techniques and a user study. | ['Matthias Nießner', 'Justus Thies', 'Christian Theobalt', 'Mohamed Elgharib', 'Ayush Tewari'] | 2019-12-11 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2619_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123610698.pdf | eccv-2020-8 | ['talking-head-generation', 'talking-face-generation'] | ['computer-vision', 'computer-vision'] | [ 4.23326135e-01 4.76323992e-01 2.97511697e-01 -3.43136817e-01
-9.37975228e-01 -5.11143804e-01 7.92398751e-01 -1.02499461e+00
2.32844248e-01 4.31762010e-01 6.28168702e-01 2.35588863e-01
4.80534077e-01 -3.56170684e-01 -1.01682949e+00 -6.48190379e-01
4.45351191e-02 4.69669104e-01 -9.43287760e-02 -9.09256637e-02
-1.90073326e-01 6.24156415e-01 -1.99468744e+00 6.99154079e-01
-1.76806033e-01 9.00200069e-01 -8.86154026e-02 1.27924871e+00
1.55240208e-01 1.09717560e+00 -8.57352078e-01 -3.03373933e-01
3.52156788e-01 -6.19275868e-01 -5.25098026e-01 4.97364461e-01
8.42038512e-01 -8.87940943e-01 -8.18242550e-01 4.65638518e-01
7.00160265e-01 1.70831382e-01 5.28801501e-01 -1.85521555e+00
-6.09265685e-01 6.29440665e-01 -2.70587444e-01 -5.03475666e-01
9.32240903e-01 5.07905841e-01 5.94835162e-01 -9.38120842e-01
9.61244047e-01 2.00267100e+00 6.15435064e-01 1.19362152e+00
-1.33260024e+00 -9.57567513e-01 -1.80164769e-01 -2.06947058e-01
-1.25508952e+00 -1.11464822e+00 7.50313163e-01 -4.07311052e-01
5.32191753e-01 6.70064688e-02 9.01112676e-01 1.85014927e+00
1.15488255e-02 8.68759871e-01 5.37222326e-01 -2.77250916e-01
6.65801689e-02 -1.94050550e-01 -8.80088568e-01 6.23176157e-01
-8.63195479e-01 4.22161013e-01 -1.06132019e+00 -2.88077861e-01
1.08520603e+00 -2.34934628e-01 -3.56714398e-01 -2.14408472e-01
-1.36779058e+00 5.82311034e-01 -1.21222183e-01 -2.02811524e-01
-3.39683861e-01 8.87170017e-01 3.45601827e-01 3.07535142e-01
2.16846153e-01 6.53758198e-02 1.09625265e-01 -4.26101446e-01
-1.36783195e+00 8.37483287e-01 9.00610507e-01 1.13439620e+00
1.98060319e-01 8.19812477e-01 -2.32993320e-01 4.41764802e-01
3.99219722e-01 5.02335787e-01 3.67485136e-01 -1.97040391e+00
1.29319699e-02 -1.21002153e-01 2.21129134e-01 -7.27568388e-01
4.47312966e-02 3.20084929e-01 -2.72292793e-01 7.91121244e-01
2.55033135e-01 -3.19025487e-01 -8.20423126e-01 1.88665175e+00
3.16492379e-01 5.59099853e-01 1.47522286e-01 9.54976618e-01
1.01917136e+00 1.07684350e+00 -3.59259516e-01 -2.70091385e-01
1.04173672e+00 -1.05236936e+00 -7.47720420e-01 2.31936857e-01
-6.86918944e-02 -6.84317589e-01 1.14195836e+00 5.12879670e-01
-1.52770972e+00 -5.09112179e-01 -7.33385503e-01 -1.63336813e-01
2.75121182e-01 -6.60303831e-02 4.59210835e-02 4.17232692e-01
-1.47938180e+00 6.30654335e-01 -7.07882762e-01 -3.54292303e-01
1.54625386e-01 1.86862528e-01 -7.08972454e-01 3.15502346e-01
-8.03025007e-01 2.03432903e-01 -9.89174694e-02 -2.41370454e-01
-1.81373322e+00 -9.29558575e-01 -8.94147456e-01 8.67288411e-02
1.99934036e-01 -9.43465292e-01 1.99494803e+00 -1.62483621e+00
-2.07553864e+00 8.79454732e-01 -2.10793838e-01 -3.25936198e-01
8.82914543e-01 -1.71025828e-01 -3.27050686e-01 7.59156346e-01
7.55615234e-02 1.25821853e+00 1.72138906e+00 -1.30816376e+00
-3.42569768e-01 7.30112046e-02 -4.27330099e-03 2.47998342e-01
4.31972109e-02 3.78952444e-01 -3.61733288e-01 -8.78039956e-01
-4.66935337e-01 -9.97669220e-01 2.67823875e-01 7.54115343e-01
-5.24590671e-01 1.61584437e-01 1.47149575e+00 -4.53401327e-01
6.17944419e-01 -2.35695124e+00 2.41764292e-01 -2.57870883e-01
2.53531814e-01 -1.61233366e-01 -4.31552827e-01 6.69406056e-01
-4.25904393e-01 -4.50264066e-02 -1.41847692e-02 -7.75087178e-01
-5.04845530e-02 -5.07143140e-02 -6.29139662e-01 3.17111731e-01
2.52097309e-01 6.77770078e-01 -7.98409700e-01 -4.47517663e-01
-5.49691692e-02 1.02550542e+00 -7.51538396e-01 4.54037219e-01
-5.39427757e-01 5.88514030e-01 -2.00702972e-03 6.87651575e-01
5.65213934e-02 8.27378035e-02 -3.33548449e-02 1.83414260e-03
-6.43631816e-03 -7.94754131e-04 -9.35493410e-01 1.77985859e+00
-3.47279310e-01 1.09659767e+00 6.77808940e-01 -3.10760915e-01
7.88430214e-01 1.00153244e+00 4.69507515e-01 9.04409308e-03
2.22138315e-01 -3.70045155e-02 -3.83804023e-01 -6.46278501e-01
5.29948235e-01 -4.01310861e-01 4.31628712e-02 8.19749355e-01
1.48155779e-01 -5.45444250e-01 -1.99744776e-01 2.87853092e-01
1.07868516e+00 7.70384371e-01 -2.52917558e-01 1.68236613e-01
1.00270614e-01 -3.54233533e-01 1.69167772e-01 3.01937282e-01
-1.00619748e-01 1.21013141e+00 7.26097226e-01 -3.79725844e-01
-1.36301041e+00 -1.23112822e+00 6.06134355e-01 1.15490127e+00
-5.24560869e-01 -5.38422287e-01 -1.16885817e+00 -2.84460813e-01
-9.95321199e-02 7.07970858e-01 -5.91931701e-01 -3.81901525e-02
-6.93569243e-01 4.19345170e-01 1.01619363e+00 3.56814086e-01
4.15063426e-02 -1.44955266e+00 -6.79366887e-01 2.75873691e-01
-2.64688194e-01 -1.34346259e+00 -9.69261348e-01 -6.14073634e-01
-5.23521841e-01 -8.86601746e-01 -1.04932725e+00 -7.74090588e-01
4.96907294e-01 1.69820964e-01 9.49126005e-01 -1.28237993e-01
-2.83082724e-01 9.45701182e-01 8.06132890e-03 -2.25242466e-01
-1.11381519e+00 -5.69248438e-01 6.60518229e-01 4.85856086e-01
-3.41914773e-01 -9.97078896e-01 -3.89181465e-01 2.36180350e-01
-9.82839882e-01 2.78401375e-01 -1.11829512e-01 7.28670359e-01
1.62491724e-01 -3.40523362e-01 4.20038968e-01 -2.67168343e-01
6.55313373e-01 -1.75147668e-01 -4.24431086e-01 -2.03945875e-01
3.79263073e-01 -2.89264709e-01 6.75395668e-01 -9.72990394e-01
-1.03539085e+00 2.97598660e-01 -2.10047644e-02 -1.44371915e+00
-1.20768830e-01 -2.04153135e-01 -2.33332098e-01 2.16380641e-01
7.59078145e-01 2.08482072e-01 4.85570550e-01 -1.95200786e-01
5.39223194e-01 6.96605861e-01 1.25353789e+00 -6.83876693e-01
9.31601644e-01 6.70175612e-01 -1.81923777e-01 -1.16301894e+00
-1.30648389e-01 5.00967145e-01 -3.92579675e-01 -6.75172508e-01
8.33634615e-01 -1.12408602e+00 -1.31648505e+00 7.45131075e-01
-1.48042655e+00 -6.53694332e-01 -6.01027429e-01 1.25074327e-01
-1.28050065e+00 2.31529903e-02 -5.15581548e-01 -8.44545901e-01
-1.45771369e-01 -1.42159164e+00 1.56656408e+00 9.81651321e-02
-6.27468705e-01 -4.84387934e-01 -1.54679805e-01 3.07625771e-01
1.78315774e-01 6.21902943e-01 6.59505188e-01 -1.85374588e-01
-7.30790138e-01 5.15832528e-02 2.90075779e-01 1.87468112e-01
7.06334785e-02 7.31057346e-01 -1.24448776e+00 -2.40676180e-01
-2.73739606e-01 -6.29528999e-01 2.52985567e-01 4.15436149e-01
8.93770814e-01 -8.03678513e-01 -1.05259381e-01 6.99685276e-01
5.43759942e-01 1.35183975e-01 4.54488337e-01 -3.63547772e-01
5.81797123e-01 9.06766355e-01 1.57576457e-01 5.51890612e-01
1.81113526e-01 7.07192183e-01 5.23524165e-01 2.23908633e-01
-4.86666828e-01 -8.89709353e-01 9.12802637e-01 2.58455098e-01
-2.31189523e-02 -2.50024974e-01 -5.81036866e-01 5.11320770e-01
-1.55066228e+00 -1.28938174e+00 3.85979235e-01 1.80013084e+00
7.40288913e-01 -1.53106213e-01 5.53752661e-01 1.78169906e-01
8.47827196e-01 2.02032626e-01 -6.43229127e-01 -4.48640257e-01
8.67229551e-02 5.05956560e-02 -3.01414251e-01 5.35571098e-01
-6.19655311e-01 1.05561328e+00 7.02939701e+00 4.20733899e-01
-1.41030598e+00 -8.97890106e-02 3.77468914e-01 -8.20998311e-01
-5.22511661e-01 -1.94652542e-01 -4.18633401e-01 2.06836879e-01
1.12016332e+00 -4.44904268e-01 6.98211551e-01 7.06720710e-01
7.47015536e-01 4.24110860e-01 -1.59175289e+00 1.18012357e+00
2.76270598e-01 -1.67426562e+00 3.39560598e-01 -4.61979071e-03
4.44079399e-01 -4.85647380e-01 3.74604672e-01 1.45470947e-01
3.32288980e-01 -1.27802908e+00 1.34989882e+00 4.52816397e-01
1.56239295e+00 -6.36719584e-01 -3.75955552e-01 2.13601947e-01
-8.87048841e-01 -1.79432780e-01 3.00675213e-01 3.72001156e-02
5.68101645e-01 -2.23383129e-01 -9.90389526e-01 -2.88522452e-01
8.37586522e-01 4.24036920e-01 2.35935569e-01 3.99926484e-01
-2.14997157e-01 4.53251451e-01 -2.76205659e-01 3.48799318e-01
1.10257909e-01 2.41137803e-01 1.11007392e+00 1.03934979e+00
5.93967319e-01 1.91215366e-01 -1.67770877e-01 1.00759053e+00
-4.49861556e-01 -2.23638490e-01 -1.04089022e+00 -3.95141900e-01
4.21243191e-01 1.01197064e+00 -2.87275970e-01 -2.73291290e-01
-7.10252076e-02 1.11214912e+00 -1.89692006e-01 5.31938195e-01
-8.66464913e-01 -2.21185520e-01 9.86151040e-01 5.76076031e-01
1.70733884e-01 -2.32150644e-01 4.88773257e-01 -9.68890011e-01
-1.88161686e-01 -1.27067125e+00 -1.57428786e-01 -1.65541196e+00
-7.71929741e-01 9.08283293e-01 1.34042546e-01 -1.36647761e+00
-1.02243304e+00 -1.56087056e-01 -8.01059663e-01 5.10236740e-01
-7.75553942e-01 -1.28443754e+00 -3.64892691e-01 7.44724870e-01
9.82584059e-01 -5.29729545e-01 7.49395072e-01 -4.61302176e-02
-1.95626579e-02 4.97859865e-01 -4.20484811e-01 4.88997921e-02
8.66409004e-01 -6.84070766e-01 7.99766362e-01 5.36366105e-01
3.19690034e-02 2.26848885e-01 9.94219244e-01 -4.27005470e-01
-1.47097230e+00 -9.63488400e-01 4.57245350e-01 -3.46275687e-01
5.15983820e-01 -6.30185664e-01 -5.38336992e-01 9.49661613e-01
4.59392518e-01 7.73356017e-03 4.84638065e-01 -8.93253267e-01
-4.68807191e-01 -5.31924926e-02 -1.22278023e+00 1.10622263e+00
1.10115147e+00 -7.42757142e-01 -4.54661489e-01 1.24707475e-01
1.15244281e+00 -5.69277227e-01 -5.13879299e-01 -4.72883098e-02
1.13001215e+00 -1.20673168e+00 8.72997940e-01 -6.17158890e-01
8.54960799e-01 -2.44713128e-01 -1.21935129e-01 -1.16652703e+00
9.25833434e-02 -1.67941833e+00 -2.50359029e-01 1.20709729e+00
1.51630165e-02 -3.53066884e-02 8.50378633e-01 6.10911548e-01
-4.75117266e-02 -2.68667579e-01 -9.20820475e-01 -3.63068700e-01
-1.86045274e-01 -5.11006415e-01 6.17803872e-01 6.24925375e-01
-1.22412659e-01 8.56912732e-02 -8.30756664e-01 1.97732896e-02
5.90888321e-01 -2.62616336e-01 1.20601058e+00 -8.97201359e-01
-4.40156907e-01 -1.40657648e-01 -2.51317441e-01 -9.48510826e-01
5.94787478e-01 -6.26182139e-01 1.09372914e-01 -1.04179263e+00
-2.89259464e-01 3.36017370e-01 6.99009538e-01 4.53160822e-01
5.54564595e-01 4.68863457e-01 6.93370461e-01 2.40006685e-01
8.54115784e-02 8.18964601e-01 1.35239470e+00 -4.35372405e-02
-3.98244768e-01 6.81341663e-02 -4.48344469e-01 9.46145892e-01
3.46225291e-01 -4.66179579e-01 -5.63482583e-01 -3.90001297e-01
-3.26234028e-02 9.22599196e-01 7.29412854e-01 -9.44698870e-01
7.77621195e-02 -6.47853389e-02 4.10332233e-01 -3.43011081e-01
1.10480165e+00 -7.18653679e-01 6.36884928e-01 1.90042287e-01
-6.15156889e-01 1.53877735e-01 2.22528428e-01 4.46869105e-01
-7.11518526e-02 2.19911560e-01 8.68357658e-01 -2.87663609e-01
-7.91529864e-02 4.47720498e-01 -8.61759782e-01 -8.54112431e-02
9.61263895e-01 -4.70258802e-01 -5.50805107e-02 -1.46786404e+00
-9.54724193e-01 -9.45241228e-02 5.78498900e-01 7.94909239e-01
1.12939072e+00 -1.59627533e+00 -9.06055152e-01 4.97897387e-01
-1.74452528e-01 -1.26344711e-01 1.99261025e-01 2.83632398e-01
-6.77182436e-01 -1.69334590e-01 -3.59921634e-01 -8.02254021e-01
-1.55015218e+00 3.73061657e-01 5.03257692e-01 6.83801115e-01
-9.14519966e-01 7.57782280e-01 5.96898675e-01 -2.79848218e-01
4.78145391e-01 -1.93624850e-02 3.14732969e-01 -8.12662542e-02
8.09934855e-01 2.64876723e-01 -5.88677168e-01 -8.65256846e-01
-3.18884254e-02 3.77685815e-01 3.08566868e-01 -8.38738918e-01
1.18897414e+00 1.27976499e-02 2.69814998e-01 5.23286045e-01
1.24542904e+00 1.07073836e-01 -1.89872229e+00 3.37035246e-02
-8.66398871e-01 -3.96770179e-01 -3.01433206e-01 -4.06908154e-01
-1.16017747e+00 1.01807249e+00 1.23148292e-01 -2.42104009e-01
1.02520633e+00 -8.78888071e-02 9.49139595e-01 3.33677024e-01
1.97012231e-01 -6.65872633e-01 5.08490503e-01 3.01255375e-01
1.37269819e+00 -5.64899743e-01 -4.46291357e-01 -4.72144634e-02
-8.01796317e-01 1.41752458e+00 4.77635980e-01 -2.10789874e-01
5.19304335e-01 8.39526355e-01 3.75333637e-01 3.36063989e-02
-1.19735706e+00 5.48213780e-01 -1.32668570e-01 8.86181831e-01
1.31465301e-01 -2.66693890e-01 7.21674502e-01 3.43053609e-01
-6.90736234e-01 2.55911320e-01 1.01406884e+00 7.75652349e-01
-1.33827969e-01 -6.71766222e-01 -7.68881142e-01 -2.58186609e-01
-5.16245246e-01 4.93335724e-02 -5.89780271e-01 7.07786798e-01
-3.55333716e-01 8.05847824e-01 2.70671338e-01 -3.62283796e-01
1.56792641e-01 3.00720930e-01 6.50207341e-01 -4.06545430e-01
-5.73831439e-01 4.55333680e-01 1.02671251e-01 -8.51393938e-01
-4.77458715e-01 -7.07296550e-01 -1.27932811e+00 -4.76690561e-01
3.62399817e-01 -1.48424864e-01 5.53758979e-01 5.07350683e-01
4.46497500e-01 3.19651991e-01 7.83913374e-01 -1.65204346e+00
-1.78161696e-01 -6.79624975e-01 -4.14498776e-01 2.61279762e-01
8.96031857e-01 -3.48484248e-01 -3.94324422e-01 7.14498222e-01] | [13.227287292480469, -0.4366714656352997] |
fa40f5e6-dcde-4653-8626-ecf6a758ddcc | solar-second-order-loss-and-attention-for | 2001.08972 | null | https://arxiv.org/abs/2001.08972v5 | https://arxiv.org/pdf/2001.08972v5.pdf | SOLAR: Second-Order Loss and Attention for Image Retrieval | Recent works in deep-learning have shown that second-order information is beneficial in many computer-vision tasks. Second-order information can be enforced both in the spatial context and the abstract feature dimensions. In this work, we explore two second-order components. One is focused on second-order spatial information to increase the performance of image descriptors, both local and global. It is used to re-weight feature maps, and thus emphasise salient image locations that are subsequently used for description. The second component is concerned with a second-order similarity (SOS) loss, that we extend to global descriptors for image retrieval, and is used to enhance the triplet loss with hard-negative mining. We validate our approach on two different tasks and datasets for image retrieval and image matching. The results show that our two second-order components complement each other, bringing significant performance improvements in both tasks and lead to state-of-the-art results across the public benchmarks. Code available at: http://github.com/tonyngjichun/SOLAR | ['Krystian Mikolajczyk', 'Vassileios Balntas', 'Tony Ng', 'Yurun Tian'] | 2020-01-24 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4963_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123700256.pdf | eccv-2020-8 | ['patch-matching'] | ['computer-vision'] | [ 6.58320114e-02 -4.32613105e-01 -3.75451446e-01 -3.92315239e-01
-7.99411774e-01 -3.09282005e-01 8.23802292e-01 4.94393289e-01
-6.23438239e-01 2.27657899e-01 1.32362172e-01 -2.66067740e-02
-4.44335788e-01 -7.07132161e-01 -5.05815089e-01 -7.36205816e-01
-1.84608951e-01 6.18556961e-02 5.34041941e-01 -2.43612349e-01
4.69036341e-01 7.94735074e-01 -1.92247367e+00 3.99812430e-01
4.18279767e-01 1.40806115e+00 3.27824861e-01 1.44402713e-01
6.91465139e-02 6.08858824e-01 -3.14825773e-01 -2.21618578e-01
4.82263833e-01 -9.69825089e-02 -6.91030324e-01 -1.18289106e-01
5.91772854e-01 -2.26455092e-01 -3.22731733e-01 1.06086838e+00
5.63687682e-01 2.16480806e-01 5.56599915e-01 -1.26656938e+00
-5.08123636e-01 1.23561502e-01 -8.11220229e-01 3.38346034e-01
1.40271574e-01 2.31310520e-02 1.26777399e+00 -1.10063982e+00
5.64171195e-01 1.25929677e+00 6.69382870e-01 2.23521829e-01
-1.06475830e+00 -5.82843244e-01 1.03953645e-01 6.28179908e-01
-1.50276899e+00 -3.78253996e-01 9.34310019e-01 -5.51937282e-01
8.58901620e-01 4.12280589e-01 4.02063251e-01 6.22317433e-01
1.45821303e-01 8.09953034e-01 1.01361179e+00 -5.46622157e-01
-6.47812337e-02 4.89060059e-02 2.83313036e-01 5.98617435e-01
2.04952471e-02 3.31817478e-01 -4.98803347e-01 -8.98544565e-02
5.59762061e-01 3.25964630e-01 -1.55071318e-01 -5.62086940e-01
-1.08040631e+00 9.65577006e-01 9.29485738e-01 5.42859793e-01
-5.66003442e-01 2.34927267e-01 4.90344524e-01 2.27225363e-01
5.95307291e-01 4.70174700e-01 -2.59332776e-01 -6.31705150e-02
-9.05159891e-01 3.95807058e-01 4.66413110e-01 5.74237049e-01
9.38829720e-01 -5.42779565e-01 -5.34733117e-01 1.21675968e+00
2.03493103e-01 4.02710170e-01 5.08142948e-01 -7.85737634e-01
3.39761615e-01 6.35223269e-01 -2.10040033e-01 -1.32480955e+00
-5.38728356e-01 -4.77278382e-01 -6.81174934e-01 2.27854386e-01
2.26571620e-01 4.14466977e-01 -9.76615071e-01 1.71997619e+00
2.59231359e-01 5.99428117e-02 -4.31880087e-01 1.14228225e+00
9.24736440e-01 4.13488835e-01 4.01519202e-02 3.12276065e-01
1.54320216e+00 -1.12868965e+00 -2.87945181e-01 -1.19656965e-01
5.13127387e-01 -8.32257092e-01 9.65061188e-01 2.04045087e-01
-1.08394682e+00 -5.85045457e-01 -9.26551282e-01 -3.04248571e-01
-6.64913893e-01 3.31681728e-01 5.49935639e-01 1.69460014e-01
-1.26653039e+00 8.24775577e-01 -5.81913948e-01 -4.13428009e-01
4.64284718e-01 1.80253729e-01 -3.48838627e-01 -2.22067088e-01
-1.20084715e+00 8.47977042e-01 1.08739346e-01 -1.64860174e-01
-3.24199587e-01 -9.11234558e-01 -8.53133500e-01 2.97154695e-01
1.24247342e-01 -4.05716568e-01 9.30512905e-01 -7.37756550e-01
-9.74989712e-01 1.22352433e+00 -1.33569032e-01 -3.88576120e-01
5.77598512e-01 -1.25551358e-01 -9.31421742e-02 3.40912282e-01
2.68175513e-01 9.43614006e-01 7.22134769e-01 -1.22746980e+00
-7.45324552e-01 -5.74349463e-01 9.79600772e-02 1.59354404e-01
-5.88723779e-01 7.90676996e-02 -8.76908958e-01 -9.53720093e-01
9.73767191e-02 -1.02908576e+00 -1.68765962e-01 3.23907644e-01
-3.37351382e-01 -4.87422556e-01 6.58068120e-01 -5.53930461e-01
1.18836999e+00 -2.21340418e+00 1.96700737e-01 3.65294963e-01
1.52175307e-01 4.14907992e-01 -3.83564383e-01 5.77824712e-01
-2.53419757e-01 9.45994034e-02 -1.12603813e-01 -6.93077028e-01
5.57250120e-02 -7.42907003e-02 -2.92106092e-01 3.69771570e-01
3.80567253e-01 9.97252285e-01 -6.53474092e-01 -3.63264799e-01
2.68311620e-01 5.75219631e-01 -3.57076883e-01 -5.80087118e-02
2.35672519e-01 8.81367326e-02 -3.27744126e-01 5.28842330e-01
7.35872686e-01 -4.37570482e-01 -1.49341837e-01 -3.67009193e-01
-1.57723650e-01 2.80960858e-01 -9.68937993e-01 1.65105462e+00
-4.71229017e-01 5.75006366e-01 -3.07818174e-01 -1.05316472e+00
7.45453358e-01 -1.65774286e-01 7.08595157e-01 -1.22265935e+00
-5.54322451e-02 1.72247842e-01 -2.90222287e-01 -3.41716051e-01
4.09283191e-01 3.09124410e-01 -9.82224476e-03 2.73395598e-01
-9.17874724e-02 1.18265055e-01 3.55461448e-01 1.24445535e-01
8.74604881e-01 -1.17450066e-01 3.49779338e-01 -5.67828596e-01
6.58099711e-01 -2.05507889e-01 3.02672029e-01 7.39785075e-01
5.74200880e-03 7.60782897e-01 4.87239808e-01 -6.14543855e-01
-8.72807443e-01 -7.77376175e-01 -1.82225183e-01 1.15176022e+00
3.59726250e-01 -4.19248432e-01 -4.07405525e-01 -5.88322401e-01
3.25750560e-01 2.34914109e-01 -8.55996370e-01 -2.90604949e-01
-4.56251800e-01 -3.30235183e-01 2.57410884e-01 7.29778290e-01
4.81373906e-01 -1.05442250e+00 -8.39027524e-01 -9.86182988e-02
-7.31004449e-03 -9.56611097e-01 -6.50628507e-01 1.68354332e-01
-6.22781575e-01 -1.08854806e+00 -1.01304710e+00 -6.72839224e-01
4.81614798e-01 5.72853446e-01 1.14569938e+00 4.69867080e-01
-7.10472524e-01 3.82314950e-01 -5.14017463e-01 -3.42210770e-01
2.16764376e-01 1.56571761e-01 -2.45390579e-01 5.23156896e-02
5.00372648e-01 -4.99460697e-01 -8.81945372e-01 3.03115547e-01
-1.07847118e+00 -2.39238933e-01 7.52256334e-01 9.97322977e-01
6.82168067e-01 -3.00771683e-01 2.24374741e-01 -4.86422360e-01
4.84355956e-01 -3.21919024e-01 -7.05769479e-01 2.32016519e-01
-5.93705058e-01 1.86505511e-01 4.04412389e-01 -3.83182049e-01
-4.82879639e-01 8.67706984e-02 -2.10848004e-01 -6.10965610e-01
-5.26835136e-02 6.62626624e-01 2.45109469e-01 -3.74468803e-01
3.95305395e-01 2.22336248e-01 -7.79847950e-02 -6.00943089e-01
1.71991497e-01 3.28268409e-01 1.38549909e-01 -3.87085646e-01
6.54632866e-01 5.09593368e-01 2.77387053e-01 -8.03196132e-01
-7.22654641e-01 -8.98217916e-01 -5.59925139e-01 1.18636303e-01
7.25522935e-01 -6.60555542e-01 -6.50087655e-01 4.71693754e-01
-1.00267458e+00 -1.95227474e-01 -3.07126552e-01 3.50431353e-01
-4.05204773e-01 4.23713297e-01 -4.58699793e-01 -4.94115829e-01
-2.62287915e-01 -1.30108440e+00 1.36173546e+00 1.45585403e-01
1.55586496e-01 -9.37764287e-01 1.50041729e-01 3.48427705e-02
5.00659943e-01 6.41289353e-02 8.68044555e-01 -6.49562180e-01
-5.84200323e-01 -2.36991495e-01 -6.28429472e-01 2.41195112e-01
-1.28942847e-01 -2.20482454e-01 -9.92718518e-01 -4.28512871e-01
-2.26656318e-01 -3.78001541e-01 1.52113616e+00 3.68732482e-01
1.41246092e+00 -1.44701153e-01 -4.64389533e-01 7.29808033e-01
1.55819869e+00 -4.57776971e-02 7.79305518e-01 5.61377883e-01
5.40329814e-01 6.24497712e-01 8.21787834e-01 5.02315938e-01
4.62171674e-01 1.36884868e+00 4.98879790e-01 -4.93101507e-01
-3.45773846e-01 -1.70226339e-02 -1.37875915e-01 3.50710213e-01
6.24310859e-02 -8.03841874e-02 -9.94423985e-01 7.51950145e-01
-2.05268502e+00 -8.76243711e-01 1.12773791e-01 2.17863607e+00
6.52877569e-01 -3.08818728e-01 2.22057223e-01 1.28835188e-02
5.77056766e-01 3.05260062e-01 -4.67105985e-01 -4.96368408e-02
9.33597889e-03 2.06819773e-01 4.93721813e-01 3.25824410e-01
-1.48941278e+00 7.11619377e-01 5.19095612e+00 1.09614217e+00
-1.40115941e+00 1.13230228e-01 7.08675504e-01 -9.42430124e-02
-1.57954216e-01 -1.79765910e-01 -6.27860248e-01 4.09431309e-01
3.44280779e-01 6.71060458e-02 2.70130962e-01 7.78288662e-01
-1.29969552e-01 3.00656701e-03 -1.13838863e+00 1.21268547e+00
2.05243960e-01 -1.34178746e+00 1.26630720e-02 2.48548567e-01
5.96908331e-01 2.92680949e-01 3.18985701e-01 9.60529149e-02
-2.52524644e-01 -8.35147262e-01 8.30374122e-01 4.70522642e-01
8.26122880e-01 -7.64526010e-01 8.20722878e-01 2.80978978e-02
-1.34189284e+00 -2.25753978e-01 -4.00918931e-01 2.20890433e-01
-1.29026845e-01 5.83054483e-01 -2.92559832e-01 5.46060681e-01
1.00552416e+00 8.39026570e-01 -8.93379807e-01 1.38255703e+00
2.68942360e-02 2.05379748e-03 -2.83867002e-01 -1.16150975e-02
4.73661542e-01 4.87733353e-03 4.17900413e-01 1.19153106e+00
3.83338600e-01 -3.57723385e-02 2.86616564e-01 8.00755143e-01
-6.25965297e-02 2.10508913e-01 -5.62368929e-01 2.69875735e-01
3.59235406e-01 1.37347507e+00 -6.93406999e-01 -1.64698437e-01
-2.93512791e-01 1.03558362e+00 3.63868594e-01 2.91743129e-01
-8.07144940e-01 -6.21847630e-01 7.20606685e-01 1.39273435e-01
5.12187183e-01 -5.33597544e-02 -1.51325380e-02 -1.04639030e+00
3.72769654e-01 -6.42830551e-01 4.77410883e-01 -4.06036794e-01
-1.44685686e+00 7.02923179e-01 5.84935397e-02 -1.27418256e+00
-2.36409172e-01 -8.31537306e-01 -4.22836483e-01 7.99908996e-01
-2.07456851e+00 -1.34970272e+00 -5.44634461e-01 6.67319298e-01
3.77585918e-01 -7.50321969e-02 6.71124816e-01 7.33142078e-01
-3.03166598e-01 8.53666425e-01 1.77094176e-01 2.33198181e-02
8.39812636e-01 -1.04037166e+00 3.06535721e-01 6.18896961e-01
2.87245333e-01 6.05876267e-01 3.11371922e-01 -1.71823367e-01
-1.14240623e+00 -9.28927243e-01 9.15926933e-01 -2.11775884e-01
5.12224913e-01 -3.02463979e-01 -8.21087062e-01 2.01908320e-01
2.08472237e-02 3.50272238e-01 4.35039788e-01 -5.80879375e-02
-6.89124227e-01 -2.96965748e-01 -1.00022173e+00 3.76789540e-01
8.76836360e-01 -5.77516258e-01 -9.00993943e-02 3.75060827e-01
6.14343703e-01 -1.95644930e-01 -7.90472865e-01 4.96186256e-01
6.65672600e-01 -1.09115613e+00 1.27021027e+00 -4.81830835e-01
4.79204297e-01 -1.63516417e-01 -2.29841620e-01 -1.12861097e+00
-5.34136593e-01 -2.71853805e-01 1.74400479e-01 1.02666342e+00
4.03353274e-01 -7.08815336e-01 5.81894636e-01 4.21437889e-01
-1.15691990e-01 -1.07107997e+00 -8.84881616e-01 -9.26341832e-01
2.03961581e-02 -1.76620021e-01 4.74014163e-01 9.07241821e-01
-2.51508504e-01 8.73472244e-02 -2.63289630e-01 -2.06014328e-02
3.66360962e-01 3.36029559e-01 5.29467940e-01 -1.23546481e+00
-1.05929166e-01 -8.66496742e-01 -7.58940935e-01 -1.15981805e+00
6.54868037e-02 -8.22422445e-01 2.27598590e-03 -1.40763831e+00
4.93835896e-01 -5.33568323e-01 -7.04851389e-01 6.31190836e-01
-1.69719428e-01 6.91145837e-01 5.06070316e-01 3.28429580e-01
-6.69984162e-01 4.56229717e-01 6.77164614e-01 -1.89559102e-01
-9.77880284e-02 -1.65770855e-02 -6.04887426e-01 5.13419211e-01
8.24869335e-01 -4.01964128e-01 -2.69181192e-01 -5.56877196e-01
1.00700803e-01 -4.62440550e-01 8.38653266e-01 -1.06271291e+00
3.94760489e-01 7.89523707e-04 2.10663900e-01 -6.25765562e-01
5.03303111e-01 -9.04956758e-01 -2.90076971e-01 5.30266404e-01
-5.27840495e-01 2.00876862e-01 2.67698765e-01 3.80222797e-01
-4.75984663e-01 -2.77594179e-01 8.77976656e-01 1.00496635e-02
-9.54980135e-01 4.29371417e-01 3.92180532e-01 -1.57405421e-01
1.02917492e+00 4.55652364e-02 -3.44409168e-01 -4.85188991e-01
-4.26547408e-01 1.99604332e-01 4.75285053e-01 7.33445585e-01
8.17856491e-01 -1.54298782e+00 -6.14250422e-01 2.48763368e-01
6.15625978e-01 -5.95003307e-01 2.78904021e-01 9.84614730e-01
-4.69617277e-01 6.74238205e-01 -3.10458720e-01 -7.79480696e-01
-1.41569948e+00 6.97277009e-01 2.26285055e-01 -4.72156405e-01
-5.04620790e-01 9.91117597e-01 4.51248765e-01 -1.22804120e-01
4.79427248e-01 -2.47539505e-01 -3.84045184e-01 2.61187792e-01
6.77869380e-01 1.01379149e-01 2.47218028e-01 -7.73199916e-01
-6.96243227e-01 1.01953495e+00 -2.15770990e-01 5.65784052e-02
1.41477168e+00 -2.15838812e-02 -3.29487056e-01 2.00908169e-01
1.74542522e+00 -2.33310014e-01 -9.83563721e-01 -5.26620209e-01
2.10222557e-01 -7.15889871e-01 2.03671798e-01 -7.68335819e-01
-1.28430998e+00 1.08525479e+00 1.00263321e+00 4.17117864e-01
1.42983758e+00 1.15519822e-01 6.70237958e-01 3.42729181e-01
2.49570057e-01 -7.81603873e-01 2.24936903e-01 5.35239935e-01
1.12058413e+00 -1.37736833e+00 6.34428188e-02 -2.62189835e-01
-5.93029261e-01 1.04242074e+00 4.13429558e-01 -1.61231562e-01
7.47749448e-01 -1.33976564e-01 -6.38745651e-02 -3.80848646e-01
-5.15200973e-01 -5.47654629e-01 8.92851055e-01 3.79016548e-01
4.36950982e-01 -2.63792604e-01 -4.10523653e-01 3.00602227e-01
2.93557167e-01 -2.65492082e-01 -2.56734639e-01 7.18523502e-01
-2.21971959e-01 -1.22252619e+00 -1.52510747e-01 3.85065109e-01
-5.08850157e-01 -3.49886954e-01 -4.49437469e-01 7.42494583e-01
7.91726857e-02 6.83246672e-01 1.57971472e-01 -2.83248484e-01
3.91512424e-01 -3.20811272e-01 4.16260749e-01 -3.88030559e-01
-5.15973985e-01 2.65924856e-02 -2.14882642e-01 -9.81014669e-01
-3.59383017e-01 -4.97608721e-01 -8.44086707e-01 -1.02909677e-01
-2.27623150e-01 -6.82302401e-04 8.04944992e-01 4.99937981e-01
6.75928295e-01 3.82339716e-01 6.38163805e-01 -1.06656659e+00
-6.45905197e-01 -7.35565245e-01 -3.46857607e-01 7.28060782e-01
5.31088889e-01 -8.40935290e-01 -2.29810864e-01 -2.26819471e-01] | [10.624214172363281, 0.6810358762741089] |
0b156a0c-3ea3-4485-b0fd-93797da7af00 | real-robot-challenge-using-deep-reinforcement | 2109.15233 | null | https://arxiv.org/abs/2109.15233v3 | https://arxiv.org/pdf/2109.15233v3.pdf | Solving the Real Robot Challenge using Deep Reinforcement Learning | This paper details our winning submission to Phase 1 of the 2021 Real Robot Challenge; a challenge in which a three-fingered robot must carry a cube along specified goal trajectories. To solve Phase 1, we use a pure reinforcement learning approach which requires minimal expert knowledge of the robotic system, or of robotic grasping in general. A sparse, goal-based reward is employed in conjunction with Hindsight Experience Replay to teach the control policy to move the cube to the desired x and y coordinates of the goal. Simultaneously, a dense distance-based reward is employed to teach the policy to lift the cube to the z coordinate (the height component) of the goal. The policy is trained in simulation with domain randomisation before being transferred to the real robot for evaluation. Although performance tends to worsen after this transfer, our best policy can successfully lift the real cube along goal trajectories via an effective pinching grasp. Our approach outperforms all other submissions, including those leveraging more traditional robotic control techniques, and is the first pure learning-based method to solve this challenge. | ['David Cordova Bulens', 'Qiang Wang', 'Stephen J. Redmond', "Noel O'Connor", 'Kevin McGuinness', 'Francisco Roldan Sanchez', 'Robert McCarthy'] | 2021-09-30 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [-6.13554940e-02 3.53472918e-01 -2.62750220e-02 -1.29265189e-02
-7.49404788e-01 -1.06537127e+00 3.69485140e-01 -1.67845711e-01
-5.21412313e-01 9.42281485e-01 -2.16729417e-01 -3.47344846e-01
-5.45505643e-01 -3.99936318e-01 -1.24531174e+00 -9.14577782e-01
-5.12394369e-01 8.98529053e-01 1.34880826e-01 -5.25305986e-01
7.27548242e-01 7.11725116e-01 -1.29084301e+00 -1.83266714e-01
4.43460256e-01 9.21115041e-01 7.95858324e-01 8.66469085e-01
3.05747181e-01 7.70760119e-01 -2.05433026e-01 4.28115577e-01
6.78817451e-01 1.31538123e-01 -1.09885204e+00 -8.30742717e-02
-1.52249470e-01 -6.49008274e-01 -2.01225549e-01 7.99263716e-01
4.83195633e-01 1.69100299e-01 6.66584074e-01 -1.50161755e+00
-2.64889836e-01 6.28828049e-01 -5.25179803e-01 -4.94614840e-01
6.07790351e-01 6.09128058e-01 7.28277862e-01 -3.91740173e-01
8.29604566e-01 1.35432291e+00 3.48554671e-01 6.14495158e-01
-9.23177838e-01 -2.95849293e-01 1.87494874e-01 -1.82324335e-01
-8.78767490e-01 3.20656635e-02 3.72147322e-01 -3.91623706e-01
1.04894626e+00 -3.89049321e-01 5.05148470e-01 9.91112471e-01
4.73939061e-01 7.20401287e-01 9.74896967e-01 -4.27326977e-01
5.42127848e-01 -3.46219629e-01 -6.21603787e-01 4.71947461e-01
-7.20089078e-02 5.05988359e-01 -6.69734329e-02 -1.00054346e-01
1.00666738e+00 -1.72766596e-01 -2.18901783e-02 -1.28281808e+00
-1.51151562e+00 5.01600564e-01 8.56933892e-01 -5.17973416e-02
-1.02407730e+00 7.14786470e-01 2.14619175e-01 4.91152138e-01
-5.61294913e-01 1.04114664e+00 -8.30092549e-01 -4.60131258e-01
-1.15472935e-01 1.13807738e+00 1.04227710e+00 1.42946649e+00
4.86285716e-01 -2.40448341e-01 4.80424147e-03 3.11625361e-01
1.28189594e-01 5.24944723e-01 -1.31273061e-01 -1.80476844e+00
7.21632361e-01 3.04533005e-01 9.05796468e-01 -4.13455725e-01
-5.97077668e-01 2.55814135e-01 -9.30000767e-02 1.01705062e+00
6.48023248e-01 -6.71595395e-01 -8.62749040e-01 1.47881484e+00
4.33869928e-01 -4.59259659e-01 4.17964876e-01 1.26132286e+00
-1.58489764e-01 5.60592413e-01 -3.97622325e-02 1.43156245e-01
8.65937114e-01 -8.51040721e-01 -4.69050594e-02 -1.79487124e-01
5.39409876e-01 -6.46772742e-01 9.76525426e-01 8.60423326e-01
-1.17725599e+00 -1.37824655e-01 -1.04223430e+00 3.22338730e-01
-1.00652903e-01 -1.24469295e-01 7.22352743e-01 -1.07034095e-01
-9.20855045e-01 1.16294634e+00 -1.03594124e+00 -3.03529680e-01
2.03454196e-01 6.07709646e-01 -4.08765525e-01 -4.76028860e-01
-6.33635581e-01 1.37511241e+00 5.06074250e-01 4.50595608e-03
-1.30077720e+00 -6.03218496e-01 -6.47058070e-01 -2.56712496e-01
7.60767341e-01 -2.78403342e-01 1.86033261e+00 -2.51813829e-01
-1.98427105e+00 4.50103790e-01 6.35240436e-01 -3.33926588e-01
6.46476269e-01 -5.25550306e-01 6.26677215e-01 2.49530017e-01
2.26765454e-01 1.18811476e+00 8.52053165e-01 -1.44895601e+00
-6.03493989e-01 -1.76515713e-01 4.79474634e-01 5.10120690e-01
5.10520339e-01 -3.24878186e-01 -1.46074817e-02 -1.93679295e-02
1.43805072e-01 -1.34261155e+00 -4.20530766e-01 -3.47225778e-02
-4.10975903e-01 -5.11669219e-01 6.07898295e-01 -1.72662094e-01
-8.87215734e-02 -1.87760150e+00 7.44067013e-01 2.18873754e-01
-1.97456807e-01 -2.54594415e-01 -2.96241134e-01 9.01305854e-01
1.75740182e-01 -5.33958018e-01 8.49579945e-02 3.15937787e-01
2.56878316e-01 2.56442517e-01 -5.44758379e-01 4.41877306e-01
3.33919734e-01 9.15748179e-01 -1.38094103e+00 -2.22376585e-02
3.81547548e-02 5.56713417e-02 -7.13945925e-01 4.32422966e-01
-8.46531987e-01 6.17883742e-01 -7.68154144e-01 5.41064978e-01
4.63753015e-01 1.05500527e-01 3.43154222e-01 2.52379209e-01
-2.84512639e-01 9.75987762e-02 -9.80815589e-01 2.00623226e+00
-2.83459544e-01 -9.60158408e-02 4.74687070e-01 -8.85415316e-01
1.09372830e+00 2.40738869e-01 7.55437493e-01 -3.07982624e-01
4.18286957e-02 3.41382056e-01 9.63935778e-02 -6.29453182e-01
4.53722358e-01 5.71002699e-02 -3.64840865e-01 4.98449475e-01
1.63813657e-03 -1.12872744e+00 4.25117165e-02 -3.12227453e-03
1.38014817e+00 1.18938422e+00 -7.84946010e-02 -2.89284706e-01
-7.78632835e-02 6.53496206e-01 4.01582010e-02 8.03049982e-01
6.16151979e-03 3.86785269e-01 7.18720913e-01 -3.30683976e-01
-1.43357539e+00 -1.13910842e+00 4.63413507e-01 1.11046505e+00
2.65242964e-01 2.56609321e-01 -6.80851400e-01 -5.98550558e-01
5.29662967e-01 8.39290857e-01 -5.38111389e-01 5.71006127e-02
-7.61025608e-01 1.15065098e-01 4.94729765e-02 6.28294528e-01
1.88096344e-01 -1.60372567e+00 -1.31970155e+00 4.62759942e-01
2.64681131e-01 -8.08709741e-01 -2.19682276e-01 7.03260899e-01
-7.32336044e-01 -1.39794409e+00 -9.77851510e-01 -9.49139774e-01
6.86212122e-01 -9.19435471e-02 5.17348886e-01 -2.42526621e-01
-3.81850481e-01 7.23560274e-01 -5.71253121e-01 -4.82970476e-01
-2.69752115e-01 1.36485472e-01 2.61204660e-01 -9.56016958e-01
-3.66108380e-02 -4.40259129e-01 -6.10846460e-01 2.54614890e-01
-3.67737383e-01 -2.31535330e-01 7.95561731e-01 6.27171040e-01
2.33979091e-01 5.31509193e-03 4.06514317e-01 6.25014724e-03
9.53349233e-01 -3.87805313e-01 -9.74176347e-01 -2.91180480e-02
-3.22767437e-01 2.37387046e-01 5.94352782e-01 -7.91209936e-01
-5.58376551e-01 6.30535662e-01 3.39716166e-01 -4.60059971e-01
-1.44509956e-01 1.48699373e-01 2.03674316e-01 -5.23180775e-02
6.87964380e-01 -3.00309178e-03 4.29592520e-01 -3.83851498e-01
5.72328687e-01 3.05806786e-01 6.29746377e-01 -1.27257907e+00
5.52292764e-01 -8.38705450e-02 2.12546542e-01 -1.57506585e-01
-4.08468425e-01 -7.22112134e-02 -7.93568373e-01 -9.00247991e-02
5.76742172e-01 -4.95286047e-01 -1.58120835e+00 3.51688892e-01
-1.19767392e+00 -9.91009653e-01 -3.79666507e-01 5.05017400e-01
-1.62589347e+00 -1.41701698e-01 -4.15211827e-01 -9.15889621e-01
-1.54660419e-01 -1.27085209e+00 1.22774398e+00 1.51580617e-01
-1.23690836e-01 -1.04330368e-01 2.83670798e-02 -3.81249785e-01
4.66486812e-01 2.94465810e-01 8.01547289e-01 -2.95525998e-01
-5.59373975e-01 -4.28767204e-01 -1.33460522e-01 9.63803902e-02
-9.08494461e-03 -3.32109004e-01 -1.81070149e-01 -6.17582679e-01
-3.55192125e-01 -1.15690565e+00 7.16527179e-02 3.09003294e-01
8.93505752e-01 -8.99900869e-02 -4.16405469e-01 -1.96518414e-02
1.20392179e+00 3.89148861e-01 5.97409010e-01 7.21478641e-01
1.97551191e-01 7.36558855e-01 1.23318541e+00 5.33721864e-01
3.54595333e-01 5.91847599e-01 9.35425580e-01 7.71523297e-01
4.08113152e-01 -3.07749957e-01 2.76741952e-01 -1.07876763e-01
-2.50688881e-01 2.08794773e-01 -9.14855421e-01 4.80857015e-01
-1.97901309e+00 -5.48205316e-01 4.44884360e-01 2.11037302e+00
8.10357749e-01 1.93225667e-01 2.86848992e-01 -2.61333305e-02
3.74771267e-01 -5.15155077e-01 -1.09245849e+00 -4.34642285e-01
6.69048548e-01 4.66046408e-02 7.73234665e-01 4.86663848e-01
-9.13048148e-01 1.22946787e+00 6.65563536e+00 2.61500955e-01
-1.03626359e+00 -5.20849466e-01 -2.08129108e-01 -1.66460544e-01
4.57961112e-01 1.01592839e-01 -6.80741131e-01 9.99225006e-02
5.55131018e-01 -6.30584508e-02 1.15885985e+00 1.27293479e+00
4.41221036e-02 -4.12399173e-01 -1.42602324e+00 4.73394603e-01
-6.26260281e-01 -8.78528953e-01 -6.08332574e-01 -1.29143707e-02
4.36433613e-01 7.72881880e-02 -8.75344649e-02 6.93024695e-01
1.13762641e+00 -1.10905075e+00 8.26066136e-01 6.59900784e-01
6.98497474e-01 -8.36415052e-01 4.35892433e-01 7.26497889e-01
-6.72127187e-01 -5.15376985e-01 -4.86256480e-01 -3.56877446e-01
1.61857679e-01 -3.41799498e-01 -1.37993026e+00 3.34957063e-01
7.00328946e-01 2.27659926e-01 3.76701206e-01 9.59927320e-01
-4.57845420e-01 -1.12312526e-01 -4.86795247e-01 -6.08131945e-01
6.54541552e-01 2.49256957e-02 5.06963551e-01 4.58800763e-01
2.16776490e-01 2.65955955e-01 6.05791807e-01 8.32166195e-01
3.35540831e-01 -2.89540768e-01 -7.39782751e-01 -1.31694138e-01
4.46744859e-01 1.01804531e+00 -3.86624455e-01 1.98828578e-01
3.97882819e-01 8.58344555e-01 5.04806936e-01 3.12182933e-01
-3.98747951e-01 -7.07418203e-01 4.58612651e-01 -7.96432421e-02
7.20306814e-01 -5.84427238e-01 8.59910101e-02 -3.52434367e-01
1.10660851e-01 -7.83772886e-01 -2.52720594e-01 -1.27162695e+00
-1.07373989e+00 2.73309827e-01 2.12723345e-01 -1.22945631e+00
-9.20785546e-01 -9.04067457e-01 -3.16962570e-01 9.78586078e-01
-1.30628407e+00 -8.70688021e-01 -1.16529077e-01 4.65678781e-01
2.42033780e-01 -2.29082868e-01 9.90093887e-01 -5.25677621e-01
2.52790332e-01 8.21204334e-02 -1.00939117e-01 -2.04002842e-01
6.55024350e-01 -1.38644183e+00 2.73030430e-01 -7.43434057e-02
-8.71373773e-01 5.12948275e-01 8.45669270e-01 -7.66233802e-01
-2.06225777e+00 -6.87291741e-01 3.65403742e-02 -4.68157589e-01
8.05349827e-01 -2.77078032e-01 -4.84298915e-01 8.05720210e-01
-6.40471466e-03 -6.32649362e-02 -3.49974513e-01 -1.53071091e-01
-9.10449028e-02 1.07654415e-01 -1.39431608e+00 5.32630026e-01
7.54707754e-01 2.25495875e-01 -8.31478238e-01 5.29094815e-01
1.03240263e+00 -9.26496923e-01 -9.18590486e-01 4.23523754e-01
7.86827385e-01 -1.75975800e-01 8.53508472e-01 -9.26574469e-01
6.89944685e-01 -2.21388847e-01 -3.87419254e-01 -1.58580923e+00
-2.77155548e-01 -9.60835993e-01 -2.70684827e-02 5.16651094e-01
2.60286301e-01 -4.17018056e-01 9.12403524e-01 4.52844441e-01
-3.47637460e-02 -9.08932865e-01 -8.42701674e-01 -8.96549225e-01
6.83289587e-01 -9.94152948e-02 5.18249512e-01 3.91017586e-01
5.87983072e-01 3.43632288e-02 -1.99045047e-01 2.61346638e-01
7.16209829e-01 1.61758065e-01 9.53703284e-01 -8.47305179e-01
-2.50287682e-01 -3.33497375e-01 -2.35254187e-02 -1.17481482e+00
6.17323555e-02 -6.79129004e-01 8.25277090e-01 -1.54265845e+00
-2.78130800e-01 -8.81898820e-01 2.22101152e-01 6.96537316e-01
3.94080192e-01 -5.74626684e-01 5.12347758e-01 1.84551343e-01
-5.33784747e-01 3.86276871e-01 1.70318317e+00 1.26469105e-01
-4.47318882e-01 1.16766244e-01 -4.26096529e-01 4.11470056e-01
1.00686336e+00 -2.60750383e-01 -3.06146860e-01 -3.37318867e-01
1.17093354e-01 5.96220493e-01 3.60989243e-01 -7.68581271e-01
3.36695224e-01 -6.19270861e-01 3.30483198e-01 -5.29009759e-01
3.47060084e-01 -1.05566704e+00 -4.30390805e-01 8.19489658e-01
-5.33858836e-01 1.33998886e-01 2.64215678e-01 5.00080764e-01
5.50750494e-01 -3.83777767e-01 4.60907370e-01 -4.01136816e-01
-4.87943888e-01 2.50508189e-02 -4.44332093e-01 -1.84838027e-01
1.46510613e+00 1.76260352e-01 -6.21119402e-02 -3.64559174e-01
-8.46221566e-01 7.42349148e-01 5.84895194e-01 6.27430499e-01
5.74378848e-01 -1.07924557e+00 -5.51015615e-01 -1.71413973e-01
1.06189761e-03 3.83614928e-01 -1.92880556e-01 5.16083956e-01
-5.56974769e-01 4.69671190e-01 -5.84240198e-01 -7.03591406e-01
-5.81347406e-01 8.74290049e-01 3.47052902e-01 -8.90780389e-02
-8.28437090e-01 6.23429418e-01 -4.31976914e-01 -1.05395663e+00
6.19951963e-01 -3.18515539e-01 1.17582001e-01 -7.75998294e-01
4.07309562e-01 1.86699301e-01 -1.49816126e-01 3.14722955e-01
-3.18007082e-01 4.05706793e-01 -1.30964294e-01 -4.34622169e-01
1.65402663e+00 2.69581676e-01 3.99930589e-03 1.47238791e-01
7.00138628e-01 -6.17934465e-01 -1.94088638e+00 4.86663245e-02
1.68983206e-01 -2.87860841e-01 -4.49741781e-01 -1.13912392e+00
-4.59239811e-01 6.15446448e-01 3.25613648e-01 -6.02887757e-03
5.16929984e-01 7.73571804e-03 3.89295042e-01 1.07821417e+00
1.12719858e+00 -1.18305004e+00 3.39960426e-01 8.13912630e-01
1.47714913e+00 -1.08983564e+00 -3.98227647e-02 2.55461838e-02
-9.12916183e-01 1.26581800e+00 7.43689656e-01 -8.95756662e-01
3.02926093e-01 6.00005090e-01 -1.72616988e-01 -1.87515080e-01
-6.99235260e-01 1.30412206e-01 -1.54153362e-01 9.43492770e-01
-4.12896685e-02 8.67260844e-02 1.94152132e-01 6.55098036e-02
-5.26979387e-01 2.11688146e-01 3.67421925e-01 1.60564911e+00
-8.64650428e-01 -9.95011866e-01 -5.17492414e-01 2.54662842e-01
3.23453210e-02 4.14218009e-01 -2.51673162e-01 8.60154748e-01
-3.12984705e-01 7.34434485e-01 6.68747947e-02 -1.29354343e-01
6.80409908e-01 -6.24857359e-02 1.19469929e+00 -6.55637503e-01
-3.13877732e-01 -1.17240220e-01 -4.63213474e-02 -1.08348751e+00
7.26557598e-02 -8.05976868e-01 -1.78870392e+00 -1.88031331e-01
1.74608335e-01 7.30009452e-02 1.17852771e+00 8.54889035e-01
2.87070364e-01 2.98077136e-01 6.12781048e-01 -1.76766884e+00
-1.44866347e+00 -9.47180212e-01 -5.53917468e-01 -1.45505648e-02
4.85571086e-01 -9.60588932e-01 -6.17596321e-02 -4.21413571e-01] | [4.594848155975342, 0.8425165414810181] |
a7a4cba0-b868-4508-901c-2a43d734debe | conformalizing-machine-translation-evaluation | 2306.06221 | null | https://arxiv.org/abs/2306.06221v1 | https://arxiv.org/pdf/2306.06221v1.pdf | Conformalizing Machine Translation Evaluation | Several uncertainty estimation methods have been recently proposed for machine translation evaluation. While these methods can provide a useful indication of when not to trust model predictions, we show in this paper that the majority of them tend to underestimate model uncertainty, and as a result they often produce misleading confidence intervals that do not cover the ground truth. We propose as an alternative the use of conformal prediction, a distribution-free method to obtain confidence intervals with a theoretically established guarantee on coverage. First, we demonstrate that split conformal prediction can ``correct'' the confidence intervals of previous methods to yield a desired coverage level. Then, we highlight biases in estimated confidence intervals, both in terms of the translation language pairs and the quality of translations. We apply conditional conformal prediction techniques to obtain calibration subsets for each data subgroup, leading to equalized coverage. | ['André F. T. Martins', 'Chrysoula Zerva'] | 2023-06-09 | null | null | null | null | ['conformal-prediction', 'machine-translation', 'conformal-prediction'] | ['computer-vision', 'natural-language-processing', 'reasoning'] | [ 1.99992046e-01 4.50017422e-01 -5.99694550e-01 -5.78516483e-01
-1.60819077e+00 -9.98319805e-01 7.25095093e-01 4.19747084e-01
2.36800965e-02 1.23617494e+00 1.66653320e-01 -5.09184301e-01
-1.37446091e-01 -7.53655374e-01 -8.74311626e-01 -3.84457707e-01
2.76759982e-01 9.54682887e-01 2.09484488e-01 1.93336532e-01
1.87669158e-01 1.68882787e-01 -1.17805135e+00 1.62040785e-01
1.34370434e+00 9.11063850e-01 -3.66425008e-01 4.10324037e-01
5.41414544e-02 3.08468014e-01 -6.68297470e-01 -9.66239631e-01
1.45869106e-02 -5.78232169e-01 -5.45822799e-01 -4.88830417e-01
1.22340627e-01 -3.90239209e-02 4.05446291e-01 1.32436430e+00
1.93086237e-01 -5.15854180e-01 1.26789033e+00 -1.21265030e+00
-5.02014577e-01 1.02943027e+00 -4.46309447e-01 -1.68087810e-01
3.09936166e-01 -5.03331602e-01 8.83267701e-01 -8.27526271e-01
5.63812971e-01 9.60188091e-01 1.03815627e+00 2.08387986e-01
-1.60378349e+00 -6.01295054e-01 -3.48947525e-01 -3.35267335e-01
-1.44036031e+00 -5.58220625e-01 4.60170716e-01 -5.84008276e-01
2.95639902e-01 4.33295965e-01 3.34569216e-01 1.08366561e+00
6.32380843e-01 3.04349035e-01 1.39507556e+00 -7.77324140e-01
4.02081311e-01 6.60752356e-01 5.26128383e-03 3.30600202e-01
6.18722737e-01 5.29268146e-01 -3.59646320e-01 -5.26357412e-01
5.84323883e-01 -5.28151453e-01 -3.38182002e-01 -5.93476057e-01
-1.37385619e+00 9.09181893e-01 -1.49897719e-02 1.58763513e-01
5.29341623e-02 2.06537172e-01 1.98344812e-01 3.11026037e-01
9.19725895e-01 3.00836474e-01 -5.82208276e-01 -1.56037211e-01
-1.02886617e+00 2.93551356e-01 8.23425829e-01 1.17843378e+00
4.98505980e-01 -4.29650187e-01 -4.62612778e-01 5.54325700e-01
2.76778519e-01 8.17373276e-01 -5.13747372e-02 -8.69434893e-01
4.85661447e-01 1.16051197e-01 6.06524110e-01 -6.82115257e-01
-9.55394059e-02 -5.47021091e-01 -5.96464396e-01 1.12513475e-01
6.30918145e-01 -9.21098664e-02 -5.91826439e-01 1.78639400e+00
-1.36461230e-02 -2.59712785e-01 6.18626215e-02 6.09987557e-01
-5.26395552e-02 3.54670256e-01 -6.37751445e-02 -5.21916330e-01
9.63015199e-01 -5.43515682e-01 -6.37596548e-01 2.61263847e-01
7.37665057e-01 -1.01516747e+00 1.12242556e+00 3.60175908e-01
-9.87952232e-01 -1.98944211e-01 -1.10447264e+00 3.46406013e-01
1.40481256e-02 1.04019038e-01 3.97481084e-01 1.00813890e+00
-7.88449764e-01 9.13183510e-01 -7.17444777e-01 -1.69351753e-02
1.33198038e-01 6.93703145e-02 -3.93959582e-02 1.32513359e-01
-1.14084196e+00 1.07966769e+00 2.44761363e-01 -1.31395027e-01
-5.62750995e-01 -7.26289332e-01 -5.25540471e-01 -2.48806411e-03
8.90111737e-03 -3.96871477e-01 1.37144077e+00 -1.06555402e+00
-1.25600195e+00 5.34590364e-01 -8.30001384e-02 -3.76066297e-01
8.30192804e-01 -6.99202809e-03 -3.83680344e-01 -2.93185085e-01
1.35953248e-01 4.86278594e-01 6.04698241e-01 -1.28557408e+00
-3.27452540e-01 -1.93395153e-01 -2.86550224e-01 -1.28869578e-01
1.13945283e-01 6.67404383e-02 -2.91240096e-01 -7.36963928e-01
2.25597501e-01 -1.03667307e+00 -1.89497530e-01 -2.66508013e-01
-5.36913037e-01 2.15222612e-02 -1.35718167e-01 -5.93332767e-01
1.28758895e+00 -1.81506252e+00 2.49567941e-01 4.81818378e-01
-1.75191686e-01 -3.55523229e-01 2.65826136e-01 2.96465337e-01
1.18955337e-01 3.46390754e-01 -6.24013007e-01 -4.04973626e-01
1.72663048e-01 1.65222421e-01 -4.85665351e-01 6.34356618e-01
1.75097436e-01 7.61107683e-01 -6.12015843e-01 -6.16685688e-01
3.46537083e-02 4.42853332e-01 -4.74649608e-01 -1.04685798e-01
-4.50706005e-01 5.33438742e-01 -1.51153326e-01 3.87938172e-01
7.42839336e-01 -1.36461586e-01 3.44268858e-01 3.32353748e-02
-3.56463604e-02 2.96542078e-01 -8.27338457e-01 1.41780269e+00
-2.31795341e-01 4.73895907e-01 -4.84176546e-01 -6.21442378e-01
9.41353381e-01 2.35051543e-01 1.14458539e-01 -1.93251684e-01
1.26019523e-01 5.02590954e-01 -1.53519154e-01 2.61936456e-01
3.25956255e-01 -4.53053325e-01 -3.28776032e-01 4.33390439e-01
4.15307097e-03 -3.79683048e-01 -3.29731941e-01 -1.77210853e-01
4.76240039e-01 4.06166345e-01 2.10198611e-01 -8.35382938e-01
1.63875699e-01 1.55049428e-01 6.01589680e-01 6.68538988e-01
2.10357551e-02 1.07269657e+00 9.59851265e-01 -1.51627839e-01
-1.51996422e+00 -1.18026483e+00 -7.31137335e-01 3.29466522e-01
-4.56726030e-02 -2.98756987e-01 -1.10217559e+00 -8.43227327e-01
-1.17653213e-01 1.07878840e+00 -7.97537982e-01 -8.20091814e-02
-6.58567026e-02 -8.13329637e-01 4.76716787e-01 4.92436647e-01
-1.50763365e-02 -3.32719415e-01 -2.58627981e-01 8.04159977e-03
-3.94553065e-01 -8.03604662e-01 -4.14451092e-01 2.33177468e-01
-8.46400023e-01 -9.85585153e-01 -8.25832307e-01 -1.77829504e-01
6.73103750e-01 -3.65690649e-01 1.34847856e+00 -2.47946382e-01
5.84230304e-01 -5.01698814e-02 -2.77265966e-01 -5.72697759e-01
-9.13556278e-01 4.94836532e-02 9.06795561e-02 -3.28314811e-01
2.93770820e-01 -4.00675446e-01 -2.34144181e-01 5.87321579e-01
-6.26992166e-01 3.28208543e-02 4.56214577e-01 9.10794556e-01
8.02602291e-01 -1.41876385e-01 5.83066463e-01 -1.03293908e+00
7.00099945e-01 -3.56019944e-01 -8.92492115e-01 6.93008959e-01
-1.31713641e+00 4.33259636e-01 2.87068665e-01 -3.44239116e-01
-9.08902168e-01 -2.51688004e-01 5.43448590e-02 -2.27206558e-01
1.07823387e-01 7.55883574e-01 1.21025192e-02 8.93925205e-02
9.85610485e-01 -2.13809893e-01 -2.56520472e-02 -4.34512615e-01
1.82168111e-01 7.19265580e-01 3.51452589e-01 -9.02508378e-01
5.76281369e-01 1.35019019e-01 -3.04020904e-02 5.52172214e-02
-8.07397544e-01 1.19212732e-01 -5.98279655e-01 1.31020613e-03
6.01141274e-01 -7.65575886e-01 -1.62291229e-01 -9.66528058e-02
-1.15126264e+00 -2.15597957e-01 -3.66776526e-01 6.80621624e-01
-8.76943111e-01 3.52291167e-01 -1.91136047e-01 -9.84417915e-01
-9.13817063e-02 -1.19803250e+00 1.26973581e+00 -1.53437823e-01
-4.05365825e-01 -1.00086713e+00 3.75648916e-01 -1.06877752e-01
3.04702461e-01 3.33811969e-01 1.17880678e+00 -5.91682076e-01
-2.16765195e-01 -2.40400240e-01 -1.81396559e-01 1.86782658e-01
-4.73105256e-03 2.63245314e-01 -8.78980041e-01 -1.76601544e-01
-9.67131704e-02 6.70828391e-03 5.51587462e-01 5.56936800e-01
8.53521645e-01 -2.98664451e-01 -4.62615132e-01 2.01783136e-01
1.35491836e+00 -1.34571835e-01 6.65637553e-01 -1.81447547e-02
-9.66570713e-03 6.81394100e-01 8.14103723e-01 2.09347203e-01
3.66814807e-02 9.99109328e-01 -4.27086465e-02 2.03740433e-01
2.83567552e-02 -6.43865585e-01 2.96228915e-01 6.27891719e-01
-1.03457652e-01 -6.61187619e-02 -1.04911089e+00 4.23181295e-01
-1.85261393e+00 -5.83027601e-01 -3.00560631e-02 2.90942454e+00
1.04364312e+00 4.11005706e-01 2.10871056e-01 -6.00305945e-02
8.05720389e-01 -4.35708493e-01 -2.03409180e-01 -4.09180045e-01
-2.13155568e-01 -2.49453057e-02 7.81232536e-01 8.63018990e-01
-5.79883575e-01 5.50419092e-01 7.73448515e+00 8.44070554e-01
-5.73851228e-01 2.98299789e-01 8.99733782e-01 2.28152320e-01
-1.16215658e+00 4.46045935e-01 -6.33492649e-01 5.85031450e-01
1.23640323e+00 -2.85264909e-01 1.12364240e-01 7.12434709e-01
-6.63968697e-02 -2.17538804e-01 -1.13831913e+00 5.51474929e-01
-1.20827116e-01 -1.20328486e+00 -1.36000589e-01 1.43027946e-01
9.14467096e-01 -2.07560390e-01 3.33952568e-02 9.06770229e-02
3.23486060e-01 -1.01381683e+00 9.81739283e-01 7.91035533e-01
1.20970058e+00 -1.09602046e+00 1.00048959e+00 4.41967309e-01
-5.21866798e-01 2.25584969e-01 -5.47859907e-01 2.61208981e-01
6.24842271e-02 1.15477633e+00 -7.03903735e-01 6.10388935e-01
3.95427734e-01 9.48291793e-02 -4.18615580e-01 8.54259670e-01
-1.92813531e-01 7.39121974e-01 -3.97239000e-01 -1.16568387e-01
-2.12739930e-01 -4.33936805e-01 4.66970503e-01 9.52169418e-01
7.50901759e-01 -2.56996512e-01 -4.08676982e-01 1.10820222e+00
1.69065922e-01 1.19383149e-01 -6.84948981e-01 -4.55426127e-02
5.82623720e-01 5.50415576e-01 -7.09447026e-01 -1.53807208e-01
-3.24574709e-01 7.21898913e-01 2.50730932e-01 2.46849969e-01
-1.07038462e+00 -1.15222842e-01 2.61434764e-01 8.32953006e-02
2.66555287e-02 -2.14395598e-01 -6.03338480e-01 -1.35808015e+00
7.33636469e-02 -7.87534297e-01 9.26337615e-02 -7.07035244e-01
-1.06326938e+00 6.07970119e-01 4.69681114e-01 -1.26036215e+00
-5.39143682e-01 -5.99227428e-01 -2.09959000e-01 1.13833964e+00
-1.07798123e+00 -8.98010731e-01 3.72204065e-01 -8.48276466e-02
1.12405129e-01 -5.47173806e-03 1.09931660e+00 -5.00567928e-02
-1.58330634e-01 8.09204280e-01 4.74917591e-01 -3.21683913e-01
9.33214247e-01 -1.17970216e+00 2.77243316e-01 7.51997769e-01
3.20010155e-01 4.51600313e-01 1.23171329e+00 -8.31399262e-01
-8.78822207e-01 -7.73363590e-01 1.16870964e+00 -8.37767243e-01
4.73594815e-01 -3.24278861e-01 -6.40721440e-01 7.49559045e-01
-1.75357193e-01 -2.87309676e-01 6.47173822e-01 6.50490522e-01
-4.97096717e-01 3.39307357e-03 -1.32428622e+00 4.34148788e-01
7.01057851e-01 -3.76458347e-01 -4.47526127e-01 3.78805071e-01
8.72732759e-01 -3.08909535e-01 -1.08783031e+00 7.23522305e-01
7.99045801e-01 -1.17030680e+00 5.68460286e-01 -4.81773943e-01
4.74692225e-01 -2.44303599e-01 -6.69016361e-01 -1.32390630e+00
-6.79243505e-02 -4.27730501e-01 6.06580041e-02 1.25531638e+00
1.14063048e+00 -7.21522987e-01 6.44001007e-01 7.72644222e-01
7.65124932e-02 -7.57911205e-01 -1.17118049e+00 -8.47486019e-01
5.92335105e-01 -8.04785728e-01 9.45773959e-01 7.62212694e-01
2.74544537e-01 1.22010778e-03 -3.54225963e-01 1.08967617e-01
8.28551114e-01 2.59094089e-01 3.78481299e-01 -1.29827559e+00
-4.28791851e-01 -3.54205012e-01 -3.83838624e-01 -5.83665192e-01
4.49736677e-02 -6.01817131e-01 1.23323523e-01 -1.05017936e+00
4.60106969e-01 -8.75517130e-01 -2.36148953e-01 1.85245164e-02
-3.07141393e-02 3.02438259e-01 -3.97029519e-01 3.10805351e-01
-1.11087948e-01 3.16673517e-01 8.56678903e-01 8.59019160e-02
1.49896368e-01 3.50975096e-01 -7.38224745e-01 5.53838193e-01
8.40701580e-01 -7.04288542e-01 -5.22541642e-01 -2.24408776e-01
8.90768528e-01 3.51658642e-01 3.20148826e-01 -8.03791583e-01
-2.24650949e-01 -2.03497380e-01 4.13589329e-01 -6.26909852e-01
-9.09678265e-02 -7.20510244e-01 6.89193189e-01 2.08349720e-01
-5.64649642e-01 2.28669904e-02 2.01176658e-01 7.16904998e-01
-1.35222077e-01 -4.45931077e-01 6.41968131e-01 3.38102967e-01
3.61326039e-01 -4.07597423e-02 -1.90072320e-02 -2.48081282e-01
6.48622155e-01 1.08860664e-01 -2.88621783e-02 -5.41684449e-01
-5.66917658e-01 -1.05366170e-01 8.80856395e-01 1.34636164e-01
2.23150283e-01 -1.62080503e+00 -8.94076645e-01 -2.63485070e-02
2.80967951e-01 -3.31508994e-01 -1.46972522e-01 1.07406271e+00
-3.27186018e-01 6.47731602e-01 1.72455266e-01 -7.06792474e-01
-7.95041382e-01 7.06328273e-01 3.55059803e-01 -1.66412950e-01
-2.54239768e-01 5.19802928e-01 -5.63629456e-02 -7.16666281e-01
1.04398495e-02 -4.93982345e-01 2.91286796e-01 -2.17115834e-01
3.04809302e-01 1.15649417e-01 2.24253133e-01 -3.42548221e-01
-2.51177669e-01 7.07048476e-01 3.61013293e-01 -7.01056778e-01
7.90535271e-01 -2.78350890e-01 -5.79090556e-03 7.44617879e-01
8.40331018e-01 4.75690067e-01 -1.11242032e+00 -3.74660781e-03
1.67370021e-01 -6.09423697e-01 -1.65437177e-01 -1.10920918e+00
-3.66346419e-01 8.85896444e-01 5.66977382e-01 2.25925356e-01
9.27871048e-01 8.60377178e-02 1.68409962e-02 -8.02796483e-02
7.16557324e-01 -8.73349905e-01 -7.35598922e-01 -8.03276524e-03
9.52880740e-01 -1.41408193e+00 2.19542161e-01 -5.93603969e-01
-6.61250651e-01 1.02354491e+00 -1.17005132e-01 2.16712490e-01
5.59483409e-01 4.11671400e-01 -1.47061810e-01 3.58850509e-01
-6.13689184e-01 2.53910720e-01 5.15302837e-01 7.17932522e-01
6.63811505e-01 4.43264306e-01 -6.86301708e-01 9.10910308e-01
-3.71458620e-01 8.14799406e-03 2.77889818e-01 2.55199224e-01
-3.25665891e-01 -1.35693800e+00 -4.45042193e-01 3.39489460e-01
-4.67872143e-01 -1.22152537e-01 -3.59419286e-01 7.15416253e-01
-1.87511370e-01 8.76005054e-01 4.07456644e-02 -3.87615561e-01
9.30576921e-02 2.17694044e-01 8.25837135e-01 -2.21097514e-01
2.45372392e-02 1.32988200e-01 5.56745291e-01 -1.75531864e-01
-2.56698757e-01 -8.52245748e-01 -7.01996148e-01 -5.21647215e-01
-7.16285884e-01 7.16152072e-01 7.55202770e-01 8.96780133e-01
3.59022230e-01 -5.81359752e-02 4.89653468e-01 -3.98849726e-01
-1.01471519e+00 -1.14350247e+00 -3.82420540e-01 -1.03854006e-02
6.94945501e-03 -8.08631837e-01 -4.31113780e-01 -2.97422230e-01] | [7.898689270019531, 4.4471259117126465] |
2c67c9bc-0d7c-43d9-9d15-e20fc6283ac1 | improvement-of-e-commerce-recommendation | null | null | https://www.dbc.wroc.pl/publication/148614 | https://sciendo.com/en/article/10.15611/eada.2020.3.03?tab=anteprima-pdf | Improvement of e-commerce recommendation systems with deep hybrid collaborative filtering with content: A case study | This paper presents a proposition to utilize flexible neural network architecture called Deep Hybrid Collaborative Filtering with Content (DHCF) as a product recommendation engine. Its main goal is to provide better shopping suggestions for customers on the e-commerce platform. The system was tested on 2018 Amazon Reviews Dataset, using repeated cross validation and compared with other approaches: collaborative filtering (CF) and deep collaborative filtering (DCF) in terms of mean squared error (MSE), mean absolute error (MAE) and mean absolute percentage error (MAPE). DCF and DHCF were proved to be significantly better than the CF. DHCF proved to be better than DCF in terms of MAE and MAPE, it also scored the best on separate test data. The significance of the differences was checked by means of a Friedman test, followed by post-hoc comparisons to control p-value. The experiment shows that DHCF can outperform other approaches considered in the study, with more robust scores. | ['Michał Górnik', 'Filip Wójcik'] | 2020-11-24 | null | null | null | econometrics-ekonometria-advances-in-applied | ['collaborative-filtering'] | ['miscellaneous'] | [-6.10332131e-01 -3.93280029e-01 1.94592252e-01 -5.46371162e-01
6.45418093e-02 -6.13689184e-01 6.74784720e-01 3.31341535e-01
-6.03551507e-01 5.26931703e-01 2.65697479e-01 -4.27385837e-01
-7.33053982e-01 -9.62200522e-01 -3.69715154e-01 -4.53592420e-01
-2.97040790e-01 8.75395313e-02 -1.00838505e-01 -5.25413692e-01
7.61928499e-01 2.77079672e-01 -1.71380723e+00 3.73735458e-01
7.89084792e-01 1.29753172e+00 1.54942989e-01 1.68307424e-01
-2.64122099e-01 1.67838916e-01 -5.11170268e-01 -8.87089074e-01
5.36911488e-01 1.27891498e-02 -6.05112672e-01 -4.13134038e-01
2.94220477e-01 -2.09595576e-01 8.76777172e-02 1.00626600e+00
5.04739344e-01 8.68687570e-01 4.59395736e-01 -9.12801266e-01
-1.21434426e+00 6.25592887e-01 -9.60107669e-02 1.08126760e-01
4.55964953e-01 -2.05337644e-01 8.28654349e-01 -1.30501497e+00
7.24952400e-01 1.11956358e+00 8.89367223e-01 1.08543225e-01
-9.60614204e-01 -6.43701375e-01 3.34446788e-01 9.84393880e-02
-1.11270034e+00 2.90496469e-01 3.01008999e-01 -2.13428974e-01
1.21460640e+00 9.29400474e-02 6.58449769e-01 5.71473956e-01
6.67411208e-01 5.45202971e-01 1.18283296e+00 -2.89642900e-01
4.35382366e-01 5.20578086e-01 5.14740467e-01 3.14782590e-01
2.98394293e-01 2.73173660e-01 -1.28146648e-01 4.41957414e-02
6.37808383e-01 4.44556504e-01 1.01392947e-01 2.39690602e-01
-6.72350407e-01 1.16717982e+00 6.78451180e-01 7.78137147e-01
-6.87944114e-01 -5.74870527e-01 4.30652559e-01 8.45331967e-01
5.38360000e-01 5.77785075e-01 -7.46688008e-01 2.37448558e-01
-7.03818381e-01 4.75781351e-01 9.93783116e-01 4.79070842e-01
1.00743622e-01 6.68935701e-02 -6.17880709e-02 7.60772109e-01
6.93559229e-01 2.73134828e-01 8.02707255e-01 -7.27933645e-01
-1.88781977e-01 4.61110920e-01 2.64798522e-01 -1.37043202e+00
-4.00012255e-01 -9.73955929e-01 -7.34733701e-01 3.60011369e-01
3.51687133e-01 -4.69767988e-01 -4.88075823e-01 1.37122250e+00
1.91955522e-01 -3.00341427e-01 -1.25674620e-01 9.91379261e-01
1.00396204e+00 7.41629839e-01 1.71054319e-01 -2.08981588e-01
8.17825675e-01 -1.02294028e+00 -8.30080032e-01 3.02514315e-01
4.21784341e-01 -1.16981637e+00 7.52721667e-01 1.00222278e+00
-9.79034901e-01 -1.18236864e+00 -9.66154456e-01 4.21054870e-01
-8.03642154e-01 2.00878277e-01 8.82465005e-01 8.32829893e-01
-1.00814831e+00 9.60772455e-01 -2.67254591e-01 -2.05924094e-01
1.39808536e-01 5.61947346e-01 -3.86376381e-01 5.27379327e-02
-1.27298641e+00 9.03723061e-01 2.54770666e-01 2.04870760e-01
-3.73099744e-01 -6.60377800e-01 -2.72895038e-01 2.06186339e-01
-4.43147533e-02 -4.90035474e-01 9.17689621e-01 -1.27818274e+00
-1.67299688e+00 -7.18018636e-02 2.97280490e-01 -6.28887832e-01
4.25003618e-01 -6.56973541e-01 -8.27391505e-01 -5.20981193e-01
-3.49660575e-01 4.81877357e-01 4.23985392e-01 -1.08195281e+00
-1.06610513e+00 -4.15436655e-01 3.00937861e-01 -1.08070470e-01
-2.46323273e-01 9.19501036e-02 1.21521704e-01 -5.66956043e-01
-4.86676656e-02 -7.81877637e-01 -3.04451287e-01 -3.93057436e-01
1.19458735e-01 -7.18052745e-01 6.88750744e-01 -7.61434495e-01
1.42304707e+00 -1.93502641e+00 -3.27694625e-01 5.45933783e-01
1.45708010e-01 7.74758220e-01 -1.89913973e-01 5.57420015e-01
9.21070576e-02 6.34044409e-02 4.70398605e-01 4.03967090e-02
2.08518475e-01 -1.70209464e-02 1.12139747e-01 2.86764503e-01
-9.46522504e-02 6.39239550e-01 -4.87695575e-01 1.71814397e-01
3.48065227e-01 5.69973648e-01 -5.66474974e-01 1.78354811e-02
-1.06599219e-01 4.44428204e-03 -2.35385507e-01 3.78636092e-01
1.06926167e+00 3.09109576e-02 -2.34318040e-02 -4.02466327e-01
-2.88629532e-01 -4.25386429e-02 -1.70211840e+00 1.30313158e+00
-6.53676093e-01 3.66871357e-01 2.27201246e-02 -6.61939144e-01
1.37135708e+00 -5.66846691e-02 1.99817792e-01 -1.06969631e+00
5.32742798e-01 3.85896236e-01 3.12784404e-01 -1.51340052e-01
6.95522726e-01 2.20694795e-01 4.55067277e-01 3.57133687e-01
1.09181874e-01 7.48034477e-01 2.24846900e-01 2.47662246e-01
7.33125925e-01 -5.00751138e-02 -1.36681080e-01 -5.44300616e-01
6.71374142e-01 -3.65463942e-01 1.58470809e-01 8.42081606e-01
-1.24627188e-01 5.83921745e-02 -2.44126707e-01 -5.32064974e-01
-8.11486483e-01 -8.70063722e-01 -1.48376867e-01 9.28283513e-01
1.09603010e-01 -2.81491309e-01 -5.35889983e-01 -8.32793236e-01
2.69031584e-01 7.86453485e-01 -6.34893477e-01 1.30030140e-01
-7.96611309e-02 -4.64591444e-01 -2.54347444e-01 5.35370827e-01
7.57740259e-01 -1.00346863e+00 -3.07087809e-01 2.79063106e-01
5.69789767e-01 -4.21271503e-01 -1.98523313e-01 8.29270929e-02
-8.45068693e-01 -7.99206436e-01 -6.63522959e-01 -5.70972204e-01
2.89643347e-01 4.36869055e-01 9.33249712e-01 1.58080488e-01
1.17033206e-01 3.13077345e-02 -8.41900289e-01 -4.14210290e-01
-4.56369258e-02 -5.82358614e-02 6.30110651e-02 -7.48785511e-02
7.45255828e-01 -1.72646940e-01 -9.22508597e-01 5.38469672e-01
-7.32932150e-01 -6.44989073e-01 5.67148745e-01 7.75908709e-01
4.22888517e-01 5.33896148e-01 9.15407598e-01 -1.03299308e+00
1.29786968e+00 -7.12959111e-01 -7.19965279e-01 -4.12407108e-02
-1.47919440e+00 -4.97090966e-01 7.07777798e-01 -3.31500649e-01
-1.17907619e+00 -5.01989126e-01 -5.14986455e-01 -1.32607281e-01
-2.13514730e-01 8.14985275e-01 2.96877623e-01 -4.12140340e-02
6.22437119e-01 -2.33055741e-01 -8.85999948e-02 -6.91957712e-01
4.23913956e-01 8.01036060e-01 1.94733173e-01 1.25987828e-01
9.49705392e-02 -1.61456764e-01 -2.17046961e-01 -3.54537398e-01
-6.44263268e-01 -4.67000067e-01 -5.98896742e-01 -3.24909359e-01
6.40941739e-01 -5.70028365e-01 -9.20532286e-01 3.44936699e-01
-7.22439229e-01 2.80964971e-01 2.20223829e-01 8.40991318e-01
2.52682924e-01 1.10149503e-01 -6.34419680e-01 -8.66614997e-01
-7.19147980e-01 -8.42021704e-01 5.93951419e-02 5.72887361e-01
-3.19104046e-01 -1.12122393e+00 -2.54623801e-01 3.61806989e-01
1.00907350e+00 5.23388088e-02 8.47070813e-01 -1.39952743e+00
2.26742268e-01 -6.83029115e-01 -1.77515492e-01 8.19621146e-01
-2.16467142e-01 3.19952071e-01 -6.49435043e-01 -3.48152727e-01
-5.02090389e-03 2.61971325e-01 5.30222774e-01 6.20686173e-01
8.96443605e-01 -2.08316490e-01 1.42890006e-01 6.65216595e-02
1.78166604e+00 7.93406069e-01 7.76603043e-01 5.49434543e-01
5.52344732e-02 5.72141409e-01 8.01460564e-01 4.65824038e-01
8.08381736e-02 3.79939348e-01 4.31317568e-01 1.68053493e-01
-7.15702176e-02 1.91829354e-01 2.04143122e-01 6.87788427e-01
-1.93882272e-01 -3.06829661e-01 -4.13227677e-01 1.54973030e-01
-1.84135997e+00 -9.67323482e-01 -5.90037823e-01 2.21664476e+00
-1.19258642e-01 3.35453659e-01 3.83336037e-01 4.64891493e-01
5.41245759e-01 -4.87950057e-01 -2.24102587e-01 -1.18376851e+00
-6.61726221e-02 6.18005574e-01 2.89077520e-01 2.30870843e-01
-1.09113133e+00 5.16035140e-01 5.81887245e+00 7.77196527e-01
-1.11793911e+00 3.18686366e-01 3.87587041e-01 1.41061366e-01
-1.69132367e-01 -4.31667775e-01 -7.21419990e-01 6.52933896e-01
1.43084049e+00 -3.23911384e-02 4.16838348e-01 8.29286993e-01
4.33806926e-02 -3.80530447e-01 -5.84610105e-01 6.03326976e-01
-1.05741531e-01 -1.39183486e+00 -6.25670552e-02 1.76792651e-01
1.13361716e+00 1.72796756e-01 2.29332790e-01 6.75211012e-01
3.48491043e-01 -7.21443594e-01 4.35574949e-01 9.49517190e-01
-2.46252269e-01 -1.14088786e+00 1.61085224e+00 1.26985699e-01
-6.27429426e-01 -5.66468060e-01 -5.51815510e-01 -2.78006881e-01
-1.96948335e-01 7.48739183e-01 -9.20939893e-02 7.24824071e-01
9.13604140e-01 4.62741315e-01 -3.71638805e-01 1.41938305e+00
2.58983582e-01 5.53133726e-01 -1.02161236e-01 -6.13625765e-01
5.09450555e-01 -5.50389946e-01 9.30187628e-02 1.11287868e+00
5.32514632e-01 -5.71685433e-02 3.91624942e-02 6.60306036e-01
1.96618989e-01 7.72659540e-01 -1.17292069e-01 -1.32161483e-01
4.92830306e-01 1.21841753e+00 -8.95598888e-01 -2.47606114e-01
-5.76994956e-01 6.40613377e-01 -3.78921211e-01 -7.63046965e-02
-7.48622596e-01 -6.84559464e-01 3.09286535e-01 3.10848579e-02
6.17170513e-01 8.22146460e-02 -2.43185595e-01 -6.55452430e-01
-3.50601822e-01 -7.75021553e-01 1.87227920e-01 -5.95538735e-01
-1.57970607e+00 7.04093993e-01 -2.43874043e-01 -1.22426617e+00
7.38243237e-02 -7.22843230e-01 -5.06247282e-01 1.17654037e+00
-8.09817731e-01 -7.78181911e-01 -2.02918500e-01 5.28569460e-01
5.16386032e-01 -6.32685781e-01 9.60766613e-01 7.43876636e-01
-4.58292067e-01 6.43252969e-01 7.06923604e-01 -2.35056847e-01
4.39064771e-01 -1.05074871e+00 -2.33552065e-02 5.57800829e-01
-4.84031849e-02 1.02837610e+00 7.44207263e-01 -6.35379255e-01
-1.40595615e+00 -7.82738864e-01 8.59102488e-01 1.44133884e-02
4.44829494e-01 1.81249201e-01 -6.20412886e-01 1.23756640e-01
4.11613703e-01 -2.89268911e-01 9.46410358e-01 6.20163321e-01
-2.40576744e-01 -2.96941072e-01 -1.70531821e+00 1.65570021e-01
4.97665882e-01 -5.05146757e-02 -2.67589003e-01 1.97507784e-01
4.02056038e-01 1.04055166e-01 -1.53000998e+00 3.08530360e-01
1.09929013e+00 -1.41001153e+00 6.11829758e-01 -4.48344857e-01
5.63662231e-01 -2.71560162e-01 -4.18535054e-01 -1.41715944e+00
-8.28203619e-01 2.27071971e-01 1.56987816e-01 1.36352479e+00
7.77189136e-01 -7.25801110e-01 5.17543375e-01 6.05800450e-01
-1.97004437e-01 -1.09527469e+00 -4.64521974e-01 -6.00040793e-01
2.51193732e-01 -4.13075566e-01 6.94288790e-01 8.78593922e-01
-4.18022394e-01 1.78530112e-01 -2.28407398e-01 -2.80808777e-01
-1.04473811e-03 9.61876139e-02 5.97127974e-01 -1.50182593e+00
-2.85060793e-01 -7.04634905e-01 -2.44164094e-01 -3.03931624e-01
-3.17404509e-01 -7.19322145e-01 -5.10777771e-01 -1.79112470e+00
-3.62391979e-01 -1.63241550e-01 -8.02911818e-01 1.23594411e-01
2.41758823e-01 1.68603390e-01 4.07470942e-01 2.91137751e-02
-4.49260801e-01 -2.58886814e-03 1.25046253e+00 2.76769213e-02
-3.33781123e-01 5.85948288e-01 -8.43088627e-01 4.46016103e-01
7.21859515e-01 -1.63807437e-01 -5.06401777e-01 -1.91572696e-01
5.98660886e-01 -1.92644909e-01 -3.69915962e-01 -1.01269209e+00
1.64332479e-01 1.82299659e-01 9.32651758e-01 -7.05433786e-01
-2.56516695e-01 -1.00277758e+00 5.08706987e-01 5.25727034e-01
-4.68008995e-01 4.47225422e-01 1.92160860e-01 4.14901078e-01
-2.67064571e-01 -5.03524303e-01 5.51694691e-01 -7.90335424e-03
-8.74040961e-01 -1.87296793e-01 -5.06545722e-01 -1.02946174e+00
9.89926517e-01 -3.00494790e-01 -1.62910372e-01 -4.65467155e-01
-1.19439173e+00 -5.45776002e-02 -1.75889358e-01 7.27452457e-01
5.19532740e-01 -1.21786726e+00 -3.66962790e-01 2.44992837e-01
-1.23637348e-01 -8.17627728e-01 5.57521045e-01 7.68836081e-01
-4.19769526e-01 6.09810114e-01 -2.75662303e-01 1.00621454e-01
-1.20883405e+00 6.16615653e-01 1.77815050e-01 -2.07264930e-01
-1.49480805e-01 1.05965865e+00 -6.41163409e-01 -5.70206106e-01
4.96306211e-01 -2.40786940e-01 -6.37628317e-01 3.42771411e-01
6.34319067e-01 8.65197361e-01 5.65753162e-01 -5.39047837e-01
-8.88248608e-02 2.56603360e-01 -5.01241863e-01 2.76770890e-01
1.46171737e+00 -1.58464998e-01 -6.81723133e-02 -2.06748359e-02
1.12963367e+00 8.53176713e-02 -6.80979311e-01 -1.61393166e-01
-3.07261758e-02 -6.48608506e-01 6.57713592e-01 -1.75820577e+00
-1.40667498e+00 3.71332765e-01 1.54253173e+00 6.75764382e-01
1.11367369e+00 -7.89013445e-01 7.94086337e-01 4.82140481e-01
3.84881407e-01 -1.39475143e+00 -3.57288897e-01 6.95746124e-01
7.37554491e-01 -1.26282060e+00 -1.20270267e-01 8.92601758e-02
-6.42030835e-01 1.33884084e+00 4.65809464e-01 -6.33006215e-01
1.54514384e+00 -6.17234455e-03 1.05269533e-02 -1.87184945e-01
-6.95570529e-01 2.18925141e-02 3.77238601e-01 2.27506593e-01
9.55681324e-01 1.41406968e-01 -1.27752483e+00 9.10676003e-01
-2.42984593e-01 3.95992219e-01 7.56746009e-02 7.82746911e-01
-5.30531466e-01 -1.08990598e+00 -2.86775380e-02 7.65662193e-01
-6.85671329e-01 1.93328261e-01 -2.06760928e-01 7.77540326e-01
6.96752965e-01 1.49575949e+00 2.21033275e-01 -9.88991499e-01
4.25107688e-01 -1.56607375e-01 2.20746115e-01 -2.19768077e-01
-1.49205554e+00 2.34151483e-01 1.35563657e-01 -4.41953778e-01
-4.71163213e-01 -4.36784357e-01 -9.54851031e-01 -7.88757801e-01
-8.16056609e-01 5.05323529e-01 1.25521374e+00 9.74752963e-01
5.02032042e-01 4.81013507e-01 8.60295177e-01 -4.83658314e-01
-5.42207420e-01 -1.28079271e+00 -6.54385805e-01 5.05785823e-01
-2.03596264e-01 -6.50866568e-01 -3.22230965e-01 -5.73645711e-01] | [10.023388862609863, 5.744769096374512] |
034961e3-f2a2-4233-863f-92f9a15cdc13 | joint-multi-frame-detection-and-segmentation | 1906.10886 | null | https://arxiv.org/abs/1906.10886v1 | https://arxiv.org/pdf/1906.10886v1.pdf | Joint Multi-frame Detection and Segmentation for Multi-cell Tracking | Tracking living cells in video sequence is difficult, because of cell morphology and high similarities between cells. Tracking-by-detection methods are widely used in multi-cell tracking. We perform multi-cell tracking based on the cell centroid detection, and the performance of the detector has high impact on tracking performance. In this paper, UNet is utilized to extract inter-frame and intra-frame spatio-temporal information of cells. Detection performance of cells in mitotic phase is improved by multi-frame input. Good detection results facilitate multi-cell tracking. A mitosis detection algorithm is proposed to detect cell mitosis and the cell lineage is built up. Another UNet is utilized to acquire primary segmentation. Jointly using detection and primary segmentation, cells can be fine segmented in highly dense cell population. Experiments are conducted to evaluate the effectiveness of our method, and results show its state-of-the-art performance. | ['Zibin Zhou', 'Peng Gao', 'Fei Wang', 'Chengkang He', 'Wenjuan Xi', 'Huaying Chen'] | 2019-06-26 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [-8.42157304e-02 -7.63249815e-01 -6.75759017e-02 5.03195167e-01
-5.90053618e-01 -6.98702216e-01 2.12418199e-01 2.91048467e-01
-7.46553481e-01 8.68963540e-01 -3.02012891e-01 2.19867066e-01
4.74697202e-01 -7.31793225e-01 -1.65451407e-01 -1.22339821e+00
1.03092059e-01 3.45761925e-01 8.88440251e-01 2.37223864e-01
3.89601469e-01 7.16129541e-01 -9.80960429e-01 -2.98150666e-02
3.31279248e-01 3.92471731e-01 6.15837276e-01 1.47588634e+00
-1.08905271e-01 6.86161339e-01 -5.35435259e-01 2.22732350e-01
-2.27723420e-02 -4.41571712e-01 -3.97378027e-01 9.10253078e-02
-2.29293033e-01 -1.97893590e-01 -1.31574377e-01 8.26419592e-01
6.89603329e-01 -2.71719933e-01 8.60373020e-01 -1.06182945e+00
4.80586141e-02 6.08962253e-02 -1.14894819e+00 8.29457283e-01
4.23191071e-01 1.53189927e-01 2.32856795e-01 -8.81028235e-01
9.82135952e-01 1.01389468e+00 6.04361653e-01 5.70864677e-01
-8.91223490e-01 -5.31041265e-01 -2.14951247e-01 -1.99948013e-01
-1.64999115e+00 -2.13069007e-01 3.04924935e-01 -4.73211467e-01
4.64312494e-01 2.17820913e-01 8.40458691e-01 3.65462750e-01
7.62888908e-01 8.40226233e-01 9.74858582e-01 -4.76419151e-01
-1.09465942e-01 -4.01746124e-01 3.58985513e-02 9.03658748e-01
3.88296217e-01 7.33877420e-02 -3.91380817e-01 -2.68235411e-02
1.42809725e+00 2.71671176e-01 -5.38645446e-01 -5.16819172e-02
-1.74430597e+00 2.93198317e-01 -8.66712704e-02 9.27141070e-01
-4.07490879e-02 3.04480165e-01 4.65208679e-01 -2.79198915e-01
3.63100730e-02 -1.71806395e-01 -2.22282559e-01 -2.41942391e-01
-1.18910706e+00 1.17685318e-01 4.59898531e-01 9.06350017e-01
5.68584979e-01 -1.24962874e-01 -3.46516490e-01 4.23508763e-01
3.78988266e-01 4.79679614e-01 7.41839349e-01 -7.88723588e-01
-4.16886657e-01 8.31565440e-01 9.47349370e-02 -8.96127045e-01
-6.62648678e-01 -1.39940724e-01 -1.03987277e+00 2.57656872e-01
1.01144409e+00 -2.58079499e-01 -8.06174040e-01 1.12699997e+00
7.09695458e-01 5.30984402e-01 3.42477933e-02 8.36707532e-01
6.70148194e-01 7.16603518e-01 1.18920277e-03 -8.60866606e-01
1.62816262e+00 -8.16834748e-01 -9.65677142e-01 2.32418656e-01
8.88377547e-01 -8.91204596e-01 2.53591835e-01 2.85886154e-02
-1.06951630e+00 -5.97064316e-01 -7.39939630e-01 1.28290430e-01
-2.49236524e-01 2.49170214e-01 2.66725898e-01 2.90293336e-01
-7.33430147e-01 2.21503526e-01 -1.20582068e+00 -5.02607644e-01
5.16788542e-01 4.38753098e-01 -2.66928494e-01 1.99103758e-01
-5.25151968e-01 4.24085498e-01 6.73194051e-01 -1.32150739e-01
-7.96069324e-01 -6.89880431e-01 -6.42029583e-01 -3.11917573e-01
3.38465758e-02 -9.48358655e-01 1.15394258e+00 -2.31537551e-01
-1.27887762e+00 1.07411110e+00 -6.00281358e-01 -9.09895152e-02
6.26423478e-01 2.49756575e-01 -2.84926534e-01 5.63697815e-01
2.18650430e-01 7.58075118e-01 2.34298289e-01 -1.28318632e+00
-1.37807322e+00 -1.88890651e-01 -5.62730253e-01 1.31952181e-01
3.18670683e-02 3.75478953e-01 -1.08158708e+00 -5.56150615e-01
8.18044469e-02 -8.47788095e-01 -2.72128433e-01 6.69922829e-02
-2.71185964e-01 7.99580961e-02 1.40494168e+00 -4.16149348e-01
1.38506234e+00 -1.92038882e+00 2.11592689e-02 -2.92326391e-01
2.69359559e-01 2.63058305e-01 3.98479342e-01 -1.58597846e-02
5.25325298e-01 -3.34210768e-02 7.07031861e-02 -2.58384734e-01
-6.94429040e-01 8.69826972e-03 3.34944308e-01 7.88284123e-01
-2.82012254e-01 9.16614771e-01 -8.98498356e-01 -1.40878046e+00
3.66454661e-01 5.90279162e-01 -1.08703271e-01 5.12771383e-02
2.09377334e-01 8.77356708e-01 -5.30761778e-01 1.12708116e+00
6.54430747e-01 -5.20445466e-01 -5.20211495e-02 -2.69022197e-01
-4.73349690e-01 -8.78639281e-01 -1.26050687e+00 1.30597699e+00
1.87277138e-01 5.06100237e-01 1.96942583e-01 -1.71044111e-01
6.79395914e-01 4.68751013e-01 7.95325041e-01 -5.31212389e-02
4.60684538e-01 4.35838819e-01 -5.82277812e-02 -2.67363429e-01
5.54108620e-01 4.31283377e-02 1.79552421e-01 1.52394846e-01
-2.50004739e-01 -7.01979324e-02 7.92967558e-01 1.42582983e-01
9.07685101e-01 2.67629415e-01 6.46894574e-01 -4.91461277e-01
9.18352962e-01 6.66976571e-02 1.03757548e+00 3.46444100e-01
-7.35922933e-01 7.66991854e-01 1.02525562e-01 -3.97686511e-01
-8.90721738e-01 -5.56177378e-01 -1.34082481e-01 7.02784538e-01
7.52621472e-01 -2.13093646e-02 -7.39941657e-01 -3.13316822e-01
-7.27305934e-02 -3.18374932e-01 -5.43694317e-01 4.85103697e-01
-1.03137362e+00 -1.13401246e+00 5.03235817e-01 3.91813308e-01
4.16078836e-01 -9.94816244e-01 -1.10602224e+00 4.85805303e-01
-1.58803016e-01 -1.31459570e+00 -6.82897210e-01 6.57632276e-02
-9.29721475e-01 -1.20002031e+00 -1.37853837e+00 -1.32040000e+00
8.28803003e-01 5.89412689e-01 6.05120063e-01 5.73234499e-01
-5.32845199e-01 6.04359852e-03 -1.33954898e-01 -1.59232423e-01
-3.33805621e-01 -2.07429915e-03 -3.36594582e-02 -1.44832000e-01
1.81087449e-01 -1.73487589e-01 -7.71539032e-01 6.92970216e-01
-7.32454062e-01 2.20472544e-01 2.70390660e-01 7.45439172e-01
1.16072035e+00 1.73913747e-01 2.48494416e-01 -8.50741506e-01
-1.24251217e-01 2.65107527e-02 -8.09376597e-01 2.25922018e-01
9.01008099e-02 -5.35594404e-01 6.35164857e-01 -6.45850658e-01
-9.42324519e-01 4.11723912e-01 2.05925405e-01 -3.75523984e-01
-2.50032604e-01 1.71587721e-01 -3.95351509e-03 -4.90968496e-01
1.23020567e-01 4.69576508e-01 -3.12307253e-02 3.55329290e-02
-2.87652701e-01 2.94032514e-01 9.10168052e-01 -1.43380851e-01
5.02190650e-01 9.31911707e-01 3.44508141e-01 -8.77489209e-01
-4.21860427e-01 -9.50516403e-01 -8.80584061e-01 -5.64722300e-01
8.31756234e-01 -9.42953944e-01 -1.02318025e+00 9.60082173e-01
-1.08737791e+00 -3.59657973e-01 9.93331373e-02 7.17422068e-01
-5.06052136e-01 6.06183350e-01 -1.28242767e+00 -8.13189685e-01
-3.50276440e-01 -1.10482717e+00 1.18400896e+00 9.70735490e-01
-6.43041916e-03 -1.34475756e+00 2.34989271e-01 -7.00234026e-02
4.38166000e-02 5.31004608e-01 2.78602570e-01 -2.20346138e-01
-9.90591407e-01 -1.43767953e-01 -8.19520950e-02 -8.09891284e-01
1.98079303e-01 7.32329667e-01 -7.97988653e-01 -4.89154637e-01
-2.57143617e-01 3.57260287e-01 8.94376218e-01 1.08373463e+00
7.39301920e-01 3.84247124e-01 -1.47025514e+00 8.73781264e-01
1.84402156e+00 4.36666697e-01 6.08181179e-01 5.33936501e-01
7.63209224e-01 1.50037423e-01 9.90046263e-01 4.05836940e-01
1.29953414e-01 4.48783576e-01 1.01685271e-01 -2.44333580e-01
-3.30861509e-01 3.84217411e-01 2.49583557e-01 9.42127168e-01
-2.10991070e-01 -4.67067808e-01 -1.08381867e+00 7.03159213e-01
-1.82798636e+00 -1.03904915e+00 -6.21173501e-01 1.66579926e+00
8.77338648e-01 1.38766691e-01 4.91238415e-01 1.73751622e-01
1.19554007e+00 -2.81820118e-01 -4.59670186e-01 3.55710953e-01
-4.80482638e-01 -4.38313454e-01 6.63907647e-01 6.41097188e-01
-1.30258048e+00 9.36629415e-01 6.81056929e+00 1.17893147e+00
-9.82118249e-01 -6.80810288e-02 7.69701123e-01 1.33126050e-01
2.55229443e-01 -1.25898704e-01 -1.57005107e+00 7.36954987e-01
2.07620412e-01 -3.13319206e-01 -3.65256339e-01 1.66798323e-01
2.63454944e-01 -8.09720874e-01 -8.43921304e-01 1.18598723e+00
-3.13337713e-01 -1.57211673e+00 -1.19071260e-01 4.04843271e-01
7.81294942e-01 -4.54317063e-01 -4.61012602e-01 -1.22573659e-01
1.17421806e-01 -5.57774484e-01 4.46065009e-01 6.55863762e-01
7.49631524e-01 -7.35511124e-01 9.10151899e-01 5.72962344e-01
-1.89667594e+00 1.41721711e-01 -2.75894642e-01 1.44395187e-01
7.06963956e-01 7.21462011e-01 -6.91762686e-01 5.13303101e-01
3.24232727e-01 8.30710173e-01 -5.23627162e-01 1.60111690e+00
4.64347333e-01 2.44318053e-01 -5.25765002e-01 -1.41717419e-01
-1.78349763e-01 -3.21087152e-01 6.17979169e-01 1.82166374e+00
6.14242136e-01 2.74080753e-01 4.92184371e-01 4.58847374e-01
1.18721589e-01 4.74922694e-02 -5.12437880e-01 2.05586344e-01
6.78801656e-01 1.84705365e+00 -1.66132069e+00 -5.87281644e-01
-3.00199628e-01 8.97477567e-01 -2.55901292e-02 9.85016525e-02
-9.74887908e-01 -3.36927503e-01 2.06208631e-01 4.31541987e-02
2.93560326e-01 -3.73348743e-01 -2.80232757e-01 -7.68252373e-01
-5.99313676e-01 -2.86265701e-01 5.03685772e-01 -5.87668777e-01
-7.09119856e-01 7.77809545e-02 -4.02932346e-01 -1.32618690e+00
7.08093643e-02 -4.60783392e-01 -7.53579199e-01 6.05477273e-01
-1.39291048e+00 -1.35298073e+00 -4.04645383e-01 3.40661705e-01
5.67913413e-01 1.11194730e-01 5.00387847e-01 4.98135209e-01
-9.12102401e-01 3.37724000e-01 1.61910325e-01 4.56841648e-01
5.90932906e-01 -1.15582764e+00 -1.72112077e-01 1.06872630e+00
-2.25194126e-01 4.40775722e-01 4.83007550e-01 -9.87831652e-01
-1.20380390e+00 -1.08408237e+00 5.91561913e-01 -4.74947184e-01
3.41825724e-01 -1.24263708e-02 -7.34108150e-01 4.06098634e-01
1.43137366e-01 4.82915759e-01 8.51475298e-01 -8.18310559e-01
7.45417595e-01 2.79576719e-01 -9.94416535e-01 6.78492129e-01
6.73399508e-01 -1.34281546e-01 5.55795655e-02 3.77719142e-02
3.65458608e-01 -9.39552069e-01 -1.05735350e+00 3.71762455e-01
7.74876237e-01 -8.02466512e-01 8.81277561e-01 4.15534884e-01
-4.56611579e-03 -1.10453570e+00 3.14105958e-01 -7.31781542e-01
-4.96875852e-01 -6.26253366e-01 -3.76849115e-01 1.39460838e+00
7.85019323e-02 -1.71027154e-01 1.12468886e+00 -1.94252521e-01
-1.24599837e-01 -7.64879584e-01 -8.57646585e-01 -7.02092052e-01
-1.56544074e-01 5.02187490e-01 2.75467455e-01 9.16976511e-01
-1.14238001e-02 3.87762100e-01 -1.22772828e-02 9.80031788e-02
6.98329329e-01 2.37017587e-01 8.99105906e-01 -1.22002816e+00
7.06470981e-02 -5.09390175e-01 -4.60385561e-01 -1.17863762e+00
-3.40424657e-01 -4.44892198e-01 -5.80464564e-02 -1.31258798e+00
7.09520578e-01 -8.95386562e-02 -1.60290569e-01 1.05747320e-01
-5.03582001e-01 5.39793134e-01 1.87870830e-01 4.98566091e-01
-9.28287923e-01 -6.30454123e-02 1.87746119e+00 1.36752814e-01
-2.32218578e-01 -3.75701398e-01 -2.88934652e-02 8.39895606e-01
5.87606728e-01 -3.68664116e-01 2.73904711e-01 2.86381617e-02
-3.16338390e-01 3.83837581e-01 2.27101371e-01 -1.45867348e+00
9.00653839e-01 -2.22503711e-02 8.34252775e-01 -1.28866768e+00
4.12958533e-01 -6.19118094e-01 4.42265749e-01 1.00599718e+00
2.32613355e-01 7.25871027e-02 1.42129809e-01 5.81345797e-01
-3.52131754e-01 -1.91666767e-01 1.21196175e+00 -4.03294027e-01
-7.12481201e-01 4.22605902e-01 -7.80944407e-01 -5.85271418e-02
1.52885306e+00 -1.07986712e+00 -3.11513305e-01 2.82248169e-01
-5.53813636e-01 4.28357244e-01 9.09721375e-01 -3.24531734e-01
4.52384025e-01 -1.40881777e+00 -5.03844559e-01 6.85853288e-02
2.00471492e-03 2.27739230e-01 1.18025884e-01 1.30092025e+00
-1.22335267e+00 4.12918866e-01 -2.96351850e-01 -1.08507180e+00
-1.75497305e+00 4.22570378e-01 4.19293225e-01 -4.56836611e-01
-4.55456764e-01 9.36943293e-01 2.95164078e-01 2.05316409e-01
-1.30086198e-01 -4.76961851e-01 -4.86481041e-01 1.40196398e-01
7.73191512e-01 6.20372713e-01 -5.19863188e-01 -9.43022907e-01
-4.36826915e-01 1.32772791e+00 -1.86296165e-01 1.63830817e-01
6.42187476e-01 -4.68844354e-01 -1.49305552e-01 6.17395937e-01
8.83550942e-01 1.76086612e-02 -1.34900177e+00 1.23453498e-01
-1.75568506e-01 -5.95430136e-01 -2.22234298e-02 -1.95596203e-01
-1.11773109e+00 6.79910004e-01 6.69855416e-01 1.19125746e-01
1.06263566e+00 -3.51598710e-01 1.14406443e+00 -1.49021462e-01
5.64176202e-01 -9.67030644e-01 3.68553251e-02 5.71867764e-01
-5.44192940e-02 -1.20065033e+00 1.70999244e-01 -6.57039106e-01
-1.70950547e-01 1.35951626e+00 8.57067347e-01 3.55775952e-02
5.81895590e-01 9.29599345e-01 2.07795247e-01 -1.74845606e-01
-5.10594487e-01 -1.94743723e-01 -4.60001320e-01 5.26378512e-01
6.12745941e-01 -2.20585182e-01 -3.69458169e-01 5.57373352e-02
4.73434120e-01 3.75442475e-01 6.05583251e-01 1.20329773e+00
-8.84455383e-01 -8.92324984e-01 -8.38954747e-01 1.37661070e-01
-8.27385604e-01 4.68845636e-01 2.58192439e-02 8.94816101e-01
1.20519869e-01 5.26458979e-01 1.11604773e-01 9.31948572e-02
-2.63368875e-01 -3.58785152e-01 4.96152759e-01 -3.53902280e-01
-5.20060956e-01 7.98903763e-01 -4.05956447e-01 -2.22181737e-01
-7.04149365e-01 -8.95798504e-01 -1.95006037e+00 -4.43128109e-01
-7.16440499e-01 9.74112824e-02 -4.49912697e-02 6.55264556e-01
4.52480055e-02 6.58608019e-01 3.76187086e-01 -9.73720193e-01
5.47385633e-01 -6.02444053e-01 -9.19433057e-01 1.83378980e-01
5.59328973e-01 -5.06694376e-01 -2.31127188e-01 7.62131751e-01] | [14.515825271606445, -3.223742961883545] |
bfe5e058-c3f0-4f46-bb20-5e9f3edfdafb | underwater-fish-detection-using-deep-learning | 1811.01494 | null | http://arxiv.org/abs/1811.01494v1 | http://arxiv.org/pdf/1811.01494v1.pdf | Underwater Fish Detection using Deep Learning for Water Power Applications | Clean energy from oceans and rivers is becoming a reality with the
development of new technologies like tidal and instream turbines that generate
electricity from naturally flowing water. These new technologies are being
monitored for effects on fish and other wildlife using underwater video.
Methods for automated analysis of underwater video are needed to lower the
costs of analysis and improve accuracy. A deep learning model, YOLO, was
trained to recognize fish in underwater video using three very different
datasets recorded at real-world water power sites. Training and testing with
examples from all three datasets resulted in a mean average precision (mAP)
score of 0.5392. To test how well a model could generalize to new datasets, the
model was trained using examples from only two of the datasets and then tested
on examples from all three datasets. The resulting model could not recognize
fish in the dataset that was not part of the training set. The mAP scores on
the other two datasets that were included in the training set were higher than
the scores achieved by the model trained on all three datasets. These results
indicate that different methods are needed in order to produce a trained model
that can generalize to new data sets such as those encountered in real world
applications. | ['Shari Matzner', 'Wenwei Xu'] | 2018-11-05 | null | null | null | null | ['fish-detection'] | ['computer-vision'] | [-2.77138166e-02 -3.89389470e-02 7.28684664e-01 -4.47286546e-01
-2.98603982e-01 -6.70175970e-01 1.88479081e-01 1.21603578e-01
-8.13754141e-01 7.21703708e-01 7.35526681e-02 -6.58773705e-02
-3.08472142e-02 -1.07857096e+00 -7.89022028e-01 -7.45108128e-01
-6.21058822e-01 2.37031564e-01 4.43052918e-01 -2.17740655e-01
3.16252708e-01 5.96838176e-01 -1.87088120e+00 1.51144385e-01
7.75513589e-01 5.06890476e-01 2.69550025e-01 9.17197764e-01
-1.82365403e-02 3.92853737e-01 -8.25862765e-01 -2.35240042e-01
6.72487438e-01 -5.24888933e-01 -3.41915965e-01 -2.87569880e-01
7.31568575e-01 -7.17103601e-01 -9.47446749e-02 1.00234103e+00
5.04930317e-01 1.83023140e-01 3.47556651e-01 -8.51016760e-01
-2.05478936e-01 5.33326030e-01 -1.15929790e-01 3.71517748e-01
1.29250601e-01 3.91106606e-01 6.45584404e-01 -5.25520921e-01
4.12792444e-01 1.00398791e+00 9.20601726e-01 5.73742330e-01
-1.00450015e+00 -9.70983863e-01 -2.98958451e-01 -7.67141432e-02
-9.86112058e-01 -4.61194992e-01 1.86425410e-02 -4.02011752e-01
1.09619927e+00 2.09345911e-02 1.16997993e+00 2.89584041e-01
4.54849690e-01 1.16235644e-01 9.78969395e-01 -1.22285999e-01
4.91181374e-01 8.71619135e-02 -3.91966313e-01 4.36642885e-01
7.20436752e-01 3.10896605e-01 -4.18363333e-01 3.10636219e-02
5.56374013e-01 -1.72147840e-01 -2.90253580e-01 1.41662762e-01
-6.07180297e-01 8.25823724e-01 6.10158205e-01 2.99544275e-01
-2.32397422e-01 2.08731204e-01 4.29936379e-01 4.65597957e-01
3.70540559e-01 8.00139308e-01 -7.98321366e-01 -2.51610428e-01
-8.58703792e-01 2.97120482e-01 1.15610421e+00 6.09401047e-01
9.93679821e-01 4.03298974e-01 7.88102806e-01 7.50581980e-01
3.43196899e-01 9.12333012e-01 7.05496550e-01 -1.01451230e+00
1.71547368e-01 6.95784032e-01 1.09571442e-01 -6.42483592e-01
-7.26577997e-01 -1.68236494e-02 -8.44635069e-02 5.76483607e-01
5.01868248e-01 -7.91125417e-01 -1.25356734e+00 1.17141294e+00
1.28277630e-01 3.21555525e-01 6.62934124e-01 8.11999738e-01
1.16680121e+00 9.72776532e-01 -1.60595644e-02 1.18417241e-01
8.72940958e-01 -2.25854263e-01 -3.35854173e-01 -3.43695283e-01
9.06557858e-01 -3.52068722e-01 5.78030765e-01 4.02795076e-01
-6.13276243e-01 -3.45577747e-01 -1.34651828e+00 4.42817152e-01
-7.92994022e-01 -2.13402450e-01 5.60288250e-01 6.85112119e-01
-1.11916459e+00 8.54376495e-01 -9.57149863e-01 -6.31144702e-01
3.23296696e-01 5.95597923e-01 -7.60213673e-01 -2.45701484e-02
-1.18082345e+00 1.22137237e+00 1.66567236e-01 4.29985136e-01
-1.17227995e+00 -4.86782640e-01 -1.01307583e+00 5.30237183e-02
-3.67032290e-01 2.30775461e-01 1.09800744e+00 -8.94742429e-01
-1.30886412e+00 4.16475266e-01 4.07904118e-01 -4.09152418e-01
4.35421616e-01 -3.35178077e-01 -3.20493698e-01 3.25329043e-02
2.22291097e-01 5.43121099e-01 2.96500564e-01 -1.17630541e+00
-1.09536171e+00 -2.31376290e-01 5.32159060e-02 1.85236096e-01
-2.71021962e-01 -2.25520819e-01 1.25205114e-01 -6.36972412e-02
1.00000687e-01 -1.05214083e+00 -8.68303031e-02 9.16090757e-02
2.26762965e-01 1.84903350e-02 6.64227545e-01 -5.53503096e-01
5.17950118e-01 -1.91990447e+00 -1.72955811e-01 1.63694680e-01
-4.91329879e-01 5.27034283e-01 -2.50941753e-01 6.42008066e-01
3.46945435e-01 2.96468765e-01 -2.21565217e-01 1.65063828e-01
-4.89385873e-01 7.30026960e-01 2.63618469e-01 6.38898671e-01
2.64629632e-01 1.71945186e-03 -1.33179319e+00 2.13245526e-02
1.37583360e-01 1.83703288e-01 -4.67728317e-01 2.30476201e-01
4.02011946e-02 3.07248861e-01 -4.15748768e-02 5.43519020e-01
6.45949006e-01 5.99251986e-01 3.72001380e-02 1.34183601e-01
-4.93773848e-01 -2.13981181e-01 -9.82702434e-01 1.19160497e+00
-5.51809371e-01 1.06309009e+00 -6.71109781e-02 -5.84516764e-01
1.13264620e+00 2.84145266e-01 3.01862627e-01 -6.63902581e-01
-2.14930344e-02 5.24658382e-01 5.45421898e-01 -1.27567255e+00
5.71844816e-01 -5.68527281e-01 2.12089375e-01 1.42216627e-02
1.01271324e-01 -4.06220943e-01 3.13101232e-01 -2.31730983e-01
9.77414370e-01 1.71503678e-01 -3.34810883e-01 -6.06995702e-01
6.70638084e-02 2.72349596e-01 6.00640655e-01 5.86735547e-01
-1.69172451e-01 3.81009847e-01 2.10906282e-01 -1.05918109e+00
-1.14564216e+00 -5.92556834e-01 -5.21515548e-01 9.72120345e-01
2.30910927e-01 -8.52511078e-02 -4.83673871e-01 -4.08764452e-01
2.65951842e-01 5.29085994e-01 -7.21506596e-01 -3.83545719e-02
-2.40472898e-01 -8.49388242e-01 7.86288440e-01 5.75499356e-01
3.61567855e-01 -1.16738355e+00 -1.27553189e+00 3.56341958e-01
3.44476581e-01 -7.16063678e-01 4.23050106e-01 5.80809772e-01
-1.11447799e+00 -1.10210836e+00 -3.66931170e-01 -6.49769127e-01
6.83302939e-01 -2.33307071e-02 6.99253082e-01 2.90711910e-01
-2.00748995e-01 -2.92795021e-02 -6.22020006e-01 -8.01740050e-01
-5.98460853e-01 -2.47950807e-01 3.19078773e-01 -4.63093072e-01
7.26353407e-01 -3.67127985e-01 -4.75403994e-01 2.23363966e-01
-1.05588961e+00 -5.60905039e-01 6.09208703e-01 7.34027684e-01
-8.35022852e-02 1.08141162e-01 8.94603074e-01 -5.63136935e-01
4.72717807e-02 -7.64238119e-01 -7.71758139e-01 -1.65352091e-01
-1.14262439e-01 6.15379773e-02 5.59965611e-01 -3.94846767e-01
-6.37331545e-01 3.33371043e-01 -3.05759460e-01 -3.04906219e-02
-1.17814966e-01 9.05268848e-01 2.41053179e-01 -1.12774983e-01
8.74557197e-01 -7.42666377e-03 2.71195203e-01 -4.13190424e-01
-3.83745134e-01 7.88280189e-01 2.98111051e-01 -2.84617990e-02
7.55028367e-01 3.76204491e-01 1.39838839e-02 -1.50763559e+00
-4.57716286e-01 -2.02134654e-01 -5.70433438e-01 -4.65132207e-01
7.53953993e-01 -1.00005615e+00 -2.72299498e-01 7.24224627e-01
-6.56200826e-01 -6.40432060e-01 -5.09962924e-02 9.71762240e-01
1.50167137e-01 4.40752506e-02 -8.13633353e-02 -9.03226137e-01
-1.55905977e-01 -1.09099114e+00 7.25618482e-01 7.71880507e-01
7.18026087e-02 -9.55269396e-01 4.44873393e-01 -3.22821200e-01
5.63155532e-01 2.43612796e-01 2.35816598e-01 -9.23117995e-01
1.48362550e-03 -5.28230608e-01 1.94721147e-01 6.47726238e-01
4.67688769e-01 3.92884970e-01 -9.48710322e-01 -4.56673682e-01
-1.32577300e-01 -4.36057895e-01 8.65623593e-01 5.66096827e-02
2.52363712e-01 -2.70885140e-01 -9.44384485e-02 4.17691678e-01
1.60507894e+00 4.72715795e-01 7.72622168e-01 6.26147807e-01
3.24112356e-01 7.57867277e-01 3.92751932e-01 1.79805905e-01
1.50059938e-01 1.49203926e-01 7.97092497e-01 6.52598664e-02
4.04469311e-01 1.46315191e-02 5.74558437e-01 1.62902340e-01
-2.45107353e-01 -5.10132387e-02 -1.00676799e+00 1.04895997e+00
-1.30882716e+00 -1.00915551e+00 -3.53403956e-01 2.29000950e+00
4.20265675e-01 1.61446594e-02 -1.26223713e-01 4.41253074e-02
3.98593783e-01 -4.22428548e-01 -6.96240723e-01 -6.92699969e-01
2.17411146e-01 1.94302738e-01 8.55682492e-01 4.78210568e-01
-1.08980131e+00 7.42576063e-01 6.82892179e+00 -2.73760289e-01
-1.33363426e+00 -4.79036629e-01 5.24761006e-02 -2.34612003e-02
-4.53716815e-02 -4.70885187e-02 -8.34599972e-01 4.00494903e-01
1.23372865e+00 -6.93435967e-02 8.56209621e-02 7.57728934e-01
3.92881185e-01 -4.83287305e-01 -1.02186406e+00 2.74627239e-01
1.70812115e-01 -9.00422812e-01 -1.26702562e-01 1.36909649e-01
8.02752614e-01 5.74164212e-01 -6.54338539e-01 4.02218819e-01
5.82417727e-01 -9.94962156e-01 4.42644626e-01 4.75652099e-01
5.07963359e-01 -4.89285648e-01 1.55424571e+00 3.06976050e-01
-8.91744852e-01 -1.92198411e-01 -8.73465419e-01 -6.34065509e-01
-1.34857640e-01 1.11490950e-01 -1.01437092e+00 1.47551194e-01
1.31650853e+00 7.33908176e-01 -6.02863252e-01 1.48663807e+00
-2.85224263e-02 5.72854042e-01 -9.75879133e-01 -4.60176229e-01
3.18217725e-01 -1.66296497e-01 3.10190737e-01 1.22622240e+00
6.45144165e-01 1.25918925e-01 -2.87316561e-01 7.48880655e-02
-3.48476693e-03 -7.77492821e-02 -9.54135001e-01 3.46837193e-02
5.98344386e-01 1.01678932e+00 -6.08989537e-01 -3.15664291e-01
-5.60201883e-01 3.12265307e-01 -6.29143324e-03 3.10124829e-02
-5.03697217e-01 -5.82053065e-01 9.04110312e-01 -1.93108749e-02
2.99939901e-01 -1.18533142e-01 1.44282803e-01 -8.35446477e-01
-3.70748460e-01 -3.97387296e-01 1.46767616e-01 -7.08523154e-01
-1.46065044e+00 5.82205951e-01 1.87865376e-01 -1.62220895e+00
-1.41880557e-01 -6.60656691e-01 -7.71804214e-01 1.07557309e+00
-1.54388118e+00 -5.84020257e-01 -7.53535628e-01 -1.42454579e-01
4.30262715e-01 -1.06674448e-01 9.93105471e-01 1.66564628e-01
-1.15022302e-01 2.17397735e-01 4.91477549e-01 4.31199312e-01
4.88165826e-01 -1.22998309e+00 2.36915588e-01 7.16767132e-01
-1.68268085e-01 3.85268003e-01 8.97313952e-01 -7.31632113e-01
-1.40704751e+00 -1.03071105e+00 5.47048986e-01 -2.36989111e-01
7.88663089e-01 1.28798962e-01 -1.08657718e+00 6.71275020e-01
1.01554103e-01 -4.78721634e-02 1.01793635e+00 -3.72786403e-01
2.67716706e-01 -2.63894349e-01 -1.45880377e+00 7.66102448e-02
4.40945059e-01 8.72630533e-03 -6.68934226e-01 7.27949888e-02
-1.01797311e-02 -5.05468309e-01 -1.19834507e+00 5.08604467e-01
1.22241914e+00 -7.74354577e-01 3.03925335e-01 -6.99492157e-01
6.54928029e-01 -5.48484087e-01 -5.77349551e-02 -1.71820676e+00
9.34548602e-02 9.01612192e-02 6.21577561e-01 9.70757663e-01
6.71646476e-01 -6.91210985e-01 7.36852109e-01 5.89205682e-01
-3.80775422e-01 -4.32198167e-01 -1.03019941e+00 -8.23394120e-01
4.86391455e-01 -1.03805080e-01 3.66119802e-01 6.91567421e-01
8.31530914e-02 -4.62383814e-02 -3.01517785e-01 7.36469448e-01
4.27845538e-01 -2.36585230e-01 1.03727555e+00 -1.52316594e+00
1.47486508e-01 1.00571655e-01 -1.07394588e+00 -4.74348575e-01
-2.80561119e-01 -4.10917103e-01 6.33511007e-01 -2.11249566e+00
-2.93982267e-01 -4.40178454e-01 2.53750861e-01 8.44933808e-01
-2.81389579e-02 6.65822148e-01 1.03744633e-01 2.45539129e-01
3.71782988e-01 4.55390990e-01 8.67143869e-01 -5.11864526e-03
-4.06967372e-01 -4.97347303e-02 -3.57812583e-01 8.24738145e-01
8.45012128e-01 -7.09359407e-01 -1.73177347e-02 -7.55239725e-01
2.56405324e-01 -2.20970616e-01 -3.15016583e-02 -1.47772217e+00
2.58349419e-01 -6.01158142e-02 6.99261427e-01 -1.68335661e-01
1.39070496e-01 -1.16997051e+00 4.47147518e-01 9.07269239e-01
1.53464796e-02 -1.69762839e-02 6.21968925e-01 4.48834211e-01
-1.76191702e-02 -7.74590373e-01 9.04691935e-01 -4.47005272e-01
-1.08505929e+00 -2.29256079e-01 -8.17457080e-01 -2.94699103e-01
1.10450530e+00 -3.52235347e-01 -5.31953752e-01 -2.61287987e-01
-8.33634257e-01 5.81840098e-01 5.94993830e-01 4.42289412e-01
6.37074411e-01 -6.44783437e-01 -1.04222488e+00 3.81180137e-01
8.13485608e-02 2.82178283e-01 2.74245199e-02 2.09301963e-01
-1.48845756e+00 -2.93448538e-01 -5.69997966e-01 -7.86296904e-01
-1.08006227e+00 -2.22397804e-01 8.65800142e-01 3.80188495e-01
-7.42574394e-01 7.04425573e-01 -3.23874563e-01 -4.23201233e-01
-6.01366237e-02 -5.65749526e-01 -5.92386663e-01 1.82240233e-01
7.36686051e-01 3.17252189e-01 -7.86560681e-03 -6.60426974e-01
-3.41330230e-01 7.36027300e-01 2.64984459e-01 -9.72932111e-03
1.90661812e+00 2.45333269e-01 9.37078148e-02 6.38469994e-01
1.00487626e+00 -8.97770631e-04 -1.54217553e+00 4.24096406e-01
-1.64060429e-01 -5.85029244e-01 5.71664236e-02 -7.47995496e-01
-1.25612009e+00 7.36757219e-01 1.04429746e+00 5.13314247e-01
8.55907679e-01 -3.00870627e-01 4.12550360e-01 7.65184343e-01
3.51015478e-01 -8.69967043e-01 -3.16769928e-01 5.81531346e-01
6.67649806e-01 -1.32983840e+00 2.42052212e-01 4.07687038e-01
-6.38261735e-01 1.41564608e+00 7.59878874e-01 -4.41252768e-01
5.82101047e-01 4.67566699e-01 5.28887272e-01 -3.53515029e-01
-3.73744249e-01 -2.47289166e-01 -4.58717734e-01 7.31061995e-01
2.39384562e-01 -2.29053199e-01 -2.76148438e-01 1.35739192e-01
-3.23655397e-01 2.98154205e-01 1.20653272e+00 1.20994377e+00
-9.80166793e-01 -7.12308645e-01 -1.84973136e-01 8.22286963e-01
-4.91395921e-01 1.42928705e-01 -3.51732135e-01 8.71672928e-01
1.82585299e-01 1.05234694e+00 3.45601141e-01 -6.15317643e-01
4.79922265e-01 -1.28091453e-02 8.86655971e-03 -7.37855434e-01
-6.56202734e-01 -3.52272779e-01 3.71702462e-01 -6.15346655e-02
-8.42426300e-01 -7.75254309e-01 -1.47695816e+00 -2.24410981e-01
-5.89124858e-01 6.96388841e-01 1.03321850e+00 7.69982159e-01
-1.41410559e-01 3.14474821e-01 6.78641617e-01 -1.50479639e+00
-3.24554294e-01 -1.53230953e+00 -6.82804167e-01 2.00558528e-01
4.34130043e-01 -5.39447844e-01 -9.09820318e-01 2.02156737e-01] | [8.484566688537598, -1.2336680889129639] |
188f7f43-96fb-4d1c-8443-9c0ec3137546 | spectral-clustering-via-orthogonalization | 2305.10356 | null | https://arxiv.org/abs/2305.10356v1 | https://arxiv.org/pdf/2305.10356v1.pdf | Spectral Clustering via Orthogonalization-Free Methods | Graph Signal Filter used as dimensionality reduction in spectral clustering usually requires expensive eigenvalue estimation. We analyze the filter in an optimization setting and propose to use four orthogonalization-free methods by optimizing objective functions as dimensionality reduction in spectral clustering. The proposed methods do not utilize any orthogonalization, which is known as not well scalable in a parallel computing environment. Our methods theoretically construct adequate feature space, which is, at most, a weighted alteration to the eigenspace of a normalized Laplacian matrix. We numerically hypothesize that the proposed methods are equivalent in clustering quality to the ideal Graph Signal Filter, which exploits the exact eigenvalue needed without expensive eigenvalue estimation. Numerical results show that the proposed methods outperform Power Iteration-based methods and Graph Signal Filter in clustering quality and computation cost. Unlike Power Iteration-based methods and Graph Signal Filter which require random signal input, our methods are able to utilize available initialization in the streaming graph scenarios. Additionally, numerical results show that our methods outperform ARPACK and are faster than LOBPCG in the streaming graph scenarios. We also present numerical results showing the scalability of our methods in multithreading and multiprocessing implementations to facilitate parallel spectral clustering. | ['Haizhao Yang', 'Qiyuan Pang'] | 2023-05-16 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [ 3.70503850e-02 1.00115508e-01 3.29900116e-01 3.40826482e-01
-4.88269776e-01 -6.94248259e-01 -6.74304040e-03 1.27111124e-02
-2.73930877e-01 1.08495623e-01 1.54253349e-01 -2.03290552e-01
-6.57982230e-01 -5.75414956e-01 -3.96158367e-01 -8.72140646e-01
-7.03458965e-01 3.00887674e-01 -4.28275950e-02 -1.58852056e-01
2.23140717e-02 4.31818813e-01 -1.26208067e+00 -3.16176653e-01
8.08107853e-01 8.00107300e-01 1.00474201e-01 1.08385801e+00
2.01449677e-01 4.16036621e-02 -3.06988835e-01 7.57673532e-02
5.71819425e-01 -4.66593146e-01 -5.28185070e-01 6.21905148e-01
1.96492538e-01 1.75696254e-01 -4.68283951e-01 1.26449347e+00
7.83744931e-01 3.02010447e-01 4.24576789e-01 -1.38119745e+00
-1.23122826e-01 9.76368129e-01 -7.95449674e-01 1.46802187e-01
5.66641092e-01 -9.03313011e-02 8.96597564e-01 -1.04461992e+00
5.61036348e-01 9.80764091e-01 9.83528018e-01 -7.84944147e-02
-1.61840308e+00 -5.32905042e-01 -1.08865604e-01 3.19330692e-01
-1.81923163e+00 -4.49005991e-01 9.96205032e-01 -3.51827145e-01
6.46863103e-01 5.64029157e-01 1.00842202e+00 4.42226380e-01
-4.05243307e-01 5.09447396e-01 9.24829185e-01 -6.13895297e-01
3.40682119e-01 -2.16865480e-01 3.77291262e-01 8.50737333e-01
4.61447805e-01 -1.61182806e-01 -6.66998029e-01 -6.57454491e-01
4.29420412e-01 -2.81230599e-01 -6.95995390e-01 -5.47278345e-01
-1.57979476e+00 7.86903620e-01 9.19211423e-04 2.14275330e-01
-5.58600008e-01 2.96710461e-01 2.17117608e-01 5.55619299e-01
3.68589520e-01 2.79941380e-01 1.01531252e-01 -1.76744834e-02
-1.30994737e+00 -9.97009426e-02 1.23670733e+00 9.24716592e-01
8.37550938e-01 4.28960383e-01 2.86064208e-01 5.03338933e-01
3.55243057e-01 8.29554975e-01 3.92468542e-01 -1.28841221e+00
2.05384955e-01 3.73970538e-01 -2.40112066e-01 -1.46303892e+00
-1.04769003e+00 -6.74176037e-01 -1.28903353e+00 -2.90806860e-01
4.04207498e-01 -3.12864691e-01 -2.77280688e-01 1.42466509e+00
4.20619547e-01 6.52934909e-01 2.89595183e-02 9.91406322e-01
3.93279970e-01 5.22421300e-01 -6.45609438e-01 -8.86559963e-01
1.26336217e+00 -5.07657111e-01 -8.55131388e-01 3.91690671e-01
6.15030348e-01 -9.79514658e-01 9.31136608e-01 7.45200396e-01
-1.03479218e+00 -2.19870538e-01 -1.25421500e+00 5.93810260e-01
9.06381831e-02 2.39878893e-01 6.23161554e-01 1.22881150e+00
-1.29987121e+00 7.43269265e-01 -9.23775375e-01 -5.91648042e-01
-1.00163139e-01 3.48619848e-01 -2.42363617e-01 3.97979200e-01
-1.00587749e+00 1.02264486e-01 2.39610493e-01 4.69104620e-03
-3.82005572e-01 -7.79688895e-01 -5.86573601e-01 2.15164512e-01
3.32705557e-01 -5.60520351e-01 3.80945951e-01 -6.95908666e-01
-1.63313258e+00 3.90334606e-01 -1.45258963e-01 -5.74122727e-01
3.23984861e-01 1.42446488e-01 -3.20919096e-01 6.66002870e-01
-1.03865579e-01 -5.25779510e-03 1.19849241e+00 -9.89078701e-01
-1.03545338e-01 -1.46062151e-01 -3.84176165e-01 3.56313497e-01
-7.35370517e-01 -4.52094257e-01 -3.65691543e-01 -7.24512756e-01
8.06137204e-01 -1.07770085e+00 -5.47765911e-01 -4.12458092e-01
-6.63061142e-01 1.66050956e-01 8.18014681e-01 -4.89158690e-01
1.64216375e+00 -2.31247354e+00 3.63572806e-01 1.01995182e+00
6.03459954e-01 -3.52304548e-01 -7.34836757e-02 8.82647574e-01
-3.61835748e-01 -1.23563455e-02 -1.97418228e-01 -4.21997346e-03
1.39702529e-01 -1.44576281e-01 -1.01033747e-01 1.27435994e+00
-5.25396585e-01 2.33813718e-01 -8.42073381e-01 -4.37516510e-01
2.30780214e-01 2.82232404e-01 -8.59153152e-01 -1.50913641e-01
4.07434344e-01 8.81930888e-02 -1.22964934e-01 2.76774019e-01
7.82566190e-01 -4.77033585e-01 5.96532464e-01 -7.25412786e-01
-9.98194609e-03 -3.74460608e-01 -2.17292237e+00 1.80482841e+00
-2.58800298e-01 6.58645511e-01 7.45463014e-01 -1.50981021e+00
5.43801785e-01 3.37653846e-01 1.07198811e+00 2.32928827e-01
-6.66801110e-02 9.94547158e-02 1.57616019e-01 -3.39479834e-01
3.42104644e-01 1.13888688e-01 1.71666831e-01 8.87778878e-01
1.46252453e-01 -1.78369910e-01 6.92883193e-01 8.84575129e-01
1.31846714e+00 -4.76416290e-01 2.58767068e-01 -9.37589645e-01
6.14096880e-01 -2.33010098e-01 3.53140712e-01 6.36512041e-01
-7.77356699e-02 4.08336550e-01 5.15223980e-01 1.39190644e-01
-8.48186433e-01 -9.42270696e-01 4.38456936e-03 7.85459161e-01
2.67198056e-01 -1.01997948e+00 -9.84783649e-01 -1.37114167e-01
-2.71784157e-01 2.24491313e-01 -1.97498903e-01 1.78493366e-01
-4.64472532e-01 -1.07628167e+00 5.08909523e-01 -1.45778030e-01
7.93772489e-02 -1.35338783e-01 -5.30047476e-01 2.93134302e-01
-2.39055604e-01 -1.05901659e+00 -8.16701353e-01 4.44855429e-02
-9.25878108e-01 -1.33587325e+00 -5.36622226e-01 -4.11707729e-01
7.32868493e-01 6.51742280e-01 7.98551679e-01 6.45653009e-02
-4.01232034e-01 1.15249908e+00 -4.36387092e-01 5.41466661e-02
-2.45589420e-01 1.49925038e-01 4.83176291e-01 5.81920445e-01
4.19241488e-02 -1.10823143e+00 -7.49040663e-01 2.72539914e-01
-8.73959064e-01 -1.68482557e-01 2.51855224e-01 7.24494934e-01
4.81418848e-01 5.28926611e-01 2.70133674e-01 -6.71449900e-01
1.11459780e+00 -4.78000015e-01 -5.74229896e-01 -6.27221121e-03
-7.46043861e-01 1.85930863e-01 8.45347643e-01 -4.02538389e-01
-3.30148309e-01 5.84344387e-01 3.37901980e-01 -6.01440787e-01
2.55942822e-01 5.25856972e-01 1.52138337e-01 -3.78620446e-01
7.72661984e-01 3.30788255e-01 1.59013867e-01 -4.22206223e-01
7.77824759e-01 3.14405054e-01 2.89715976e-01 -5.06377935e-01
1.30119061e+00 8.52791846e-01 4.40039128e-01 -1.52975655e+00
-9.32362601e-02 -8.65784943e-01 -5.76483667e-01 -5.64932585e-01
6.41658127e-01 -6.62985981e-01 -1.05594194e+00 -2.79292818e-02
-8.33458126e-01 1.98380828e-01 -3.02382439e-01 7.99418986e-01
-6.77647471e-01 1.11080694e+00 -5.32450676e-01 -9.56639171e-01
-2.59546250e-01 -8.70557725e-01 9.29235399e-01 -2.14841828e-01
-1.47479147e-01 -1.02307570e+00 2.53975037e-02 -1.15378454e-01
1.54290214e-01 2.48769417e-01 4.91109014e-01 -3.75126123e-01
-3.40806097e-01 -1.11397035e-01 6.08263798e-02 -1.74955815e-01
-9.12595093e-02 8.07871446e-02 -5.41001260e-01 -7.67328620e-01
4.39470750e-04 1.68162659e-01 6.46961391e-01 6.38151526e-01
9.22981977e-01 -2.72933573e-01 -3.79784346e-01 8.58397007e-01
1.43700325e+00 -3.53538245e-01 1.42981589e-01 -2.19848827e-01
6.30297482e-01 6.00133538e-01 9.64414254e-02 8.21745217e-01
5.58384182e-03 5.53871036e-01 1.79947168e-01 -1.81905776e-01
5.62655414e-03 1.76293209e-01 4.05567527e-01 1.54836202e+00
-1.91247627e-01 6.26946092e-02 -7.59556532e-01 4.08062905e-01
-1.99402201e+00 -1.07394373e+00 -7.37642646e-01 2.17940664e+00
3.66528451e-01 -3.15960735e-01 5.00285804e-01 7.01789975e-01
8.60732198e-01 3.99926379e-02 -2.53389686e-01 -2.83678100e-02
-1.26142904e-01 1.42818213e-01 9.39923882e-01 7.16858566e-01
-1.10021961e+00 3.49776685e-01 6.63688755e+00 8.01267207e-01
-6.70384169e-01 2.32979611e-01 -1.56971544e-01 -2.01406196e-01
-2.27082133e-01 6.03342503e-02 -2.35741630e-01 2.81636387e-01
1.01248336e+00 -6.77646756e-01 9.00562704e-01 6.27704144e-01
6.25599086e-01 9.62809175e-02 -8.60745609e-01 1.45769465e+00
-2.63904817e-02 -1.19291556e+00 -3.91706228e-01 3.36957991e-01
5.45501888e-01 -1.69457808e-01 -1.63568705e-01 -3.94825608e-01
1.45099044e-01 -6.89649165e-01 4.24217016e-01 4.97794449e-01
3.31647336e-01 -8.06650221e-01 3.08513224e-01 2.41076261e-01
-1.67111349e+00 9.83790681e-02 -2.73485124e-01 4.72901501e-02
4.93333429e-01 1.16383147e+00 -5.60800552e-01 8.18815947e-01
5.37062228e-01 6.26893222e-01 -3.46851438e-01 9.69987094e-01
2.49952480e-01 1.06026864e+00 -9.60314810e-01 5.91604002e-02
1.30045876e-01 -7.66550243e-01 1.11856639e+00 1.22522449e+00
6.65595651e-01 2.41373107e-01 5.72620809e-01 4.35986072e-01
1.56533375e-01 5.75710654e-01 -4.57750291e-01 -1.88196823e-01
4.64630127e-01 1.44001758e+00 -1.24578905e+00 -2.68363327e-01
-4.85344797e-01 9.43940878e-01 -1.90305844e-01 6.48431063e-01
-5.82837105e-01 -6.87778592e-01 4.45517361e-01 2.64026105e-01
2.45771497e-01 -6.56135321e-01 -1.04308017e-02 -1.29765868e+00
-1.22253917e-01 -8.57605219e-01 5.53384125e-01 -2.08626330e-01
-1.14126325e+00 3.38703275e-01 5.03004203e-03 -1.41830361e+00
-1.72301888e-01 -4.68415558e-01 -3.10231507e-01 4.96427119e-01
-8.19042444e-01 -6.16513371e-01 -3.16648126e-01 9.92824495e-01
2.08655611e-01 -1.75241157e-01 5.81596375e-01 4.77524340e-01
-4.90771621e-01 3.46297026e-01 3.96467775e-01 -1.38005152e-01
5.08486092e-01 -1.32968795e+00 -1.22054443e-01 1.24375606e+00
4.51609641e-01 7.28150725e-01 1.00302100e+00 -4.83871460e-01
-2.06198573e+00 -6.87469482e-01 1.15904287e-01 1.60813540e-01
1.03668463e+00 -4.00415510e-01 -6.47282124e-01 3.47059280e-01
5.09673536e-01 -2.79267784e-02 9.74818051e-01 5.60983308e-02
-2.21130829e-02 -1.33469045e-01 -7.92052686e-01 4.70864743e-01
1.21147585e+00 -3.12051952e-01 -9.82568413e-02 8.44478965e-01
4.05085623e-01 -8.58679563e-02 -1.15663278e+00 5.35019934e-02
2.45758623e-01 -1.04270303e+00 9.61150467e-01 -1.33275911e-01
-4.65439081e-01 -6.87600195e-01 -2.63711274e-01 -1.37219226e+00
-5.28175950e-01 -1.46346617e+00 -2.64493376e-01 8.58331919e-01
2.68484741e-01 -8.02406192e-01 9.53817070e-01 1.32705541e-02
2.40652770e-01 -2.91783541e-01 -9.91917908e-01 -9.81673121e-01
-4.28710103e-01 -6.71038687e-01 1.89184442e-01 1.26741540e+00
4.91287708e-01 3.87496769e-01 -2.91263908e-01 6.09198928e-01
1.32267666e+00 2.22011209e-01 7.09865153e-01 -1.31074631e+00
-6.94472730e-01 -5.23906529e-01 -5.97397685e-01 -8.61292422e-01
1.09919757e-01 -1.07958066e+00 -4.55403328e-01 -1.23205030e+00
-3.36349234e-02 -7.30184913e-02 1.64164156e-01 -2.06124976e-01
6.31465465e-02 3.15717459e-01 3.01379383e-01 1.93575054e-01
-4.30049062e-01 3.08534592e-01 7.09898591e-01 -9.20231193e-02
-3.82428139e-01 -6.22551478e-02 -2.43699834e-01 7.57032990e-01
4.73530173e-01 -3.05446923e-01 -7.24739671e-01 1.28020152e-01
5.11257052e-01 1.15872987e-01 -6.14864491e-02 -1.16901875e+00
6.08964443e-01 2.73051322e-01 -6.65488318e-02 -4.84853953e-01
2.64516443e-01 -1.02919805e+00 4.03349251e-01 6.91854835e-01
1.33481115e-01 1.34532675e-01 -1.68412343e-01 9.90968406e-01
-9.25030280e-03 -1.91741675e-01 6.75768256e-01 1.21306904e-01
-3.36220324e-01 2.34051928e-01 -7.35430837e-01 6.75820373e-03
9.01832104e-01 -4.28073823e-01 2.11303622e-01 -9.51405346e-01
-1.27413750e+00 2.57502973e-01 3.27369362e-01 -2.99595237e-01
4.44408447e-01 -1.30073798e+00 -9.09239054e-01 9.57911387e-02
-3.28194201e-01 -5.99468112e-01 2.07807109e-01 1.50590420e+00
-6.76759124e-01 8.99573416e-02 3.49593610e-01 -8.51253986e-01
-1.36292422e+00 6.43786371e-01 2.15553135e-01 -1.52880996e-01
-6.49328470e-01 5.42413354e-01 -9.30209756e-02 -3.13915491e-01
1.29606307e-01 -6.04781508e-02 8.79557356e-02 3.12197268e-01
2.48390153e-01 9.86175776e-01 1.01805896e-01 -4.89847660e-01
-3.71307641e-01 8.01644027e-01 6.98679626e-01 -4.36535358e-01
1.29919219e+00 -5.03379941e-01 -4.11235273e-01 2.02222839e-01
1.18713534e+00 5.01475811e-01 -6.15103066e-01 6.94392854e-03
7.75097609e-02 -2.68496066e-01 2.54311234e-01 2.38183543e-01
-1.14267635e+00 5.61873257e-01 7.68970966e-01 9.92938519e-01
1.53491902e+00 -2.68833786e-01 5.10345399e-01 5.53893626e-01
2.31832385e-01 -1.26277435e+00 -1.25541821e-01 4.09424864e-02
7.10732222e-01 -5.55511057e-01 4.58178371e-01 -9.37279761e-01
-6.23928979e-02 1.37577343e+00 4.78875861e-02 -3.72739613e-01
1.10231841e+00 4.04740334e-01 -1.55381173e-01 -2.39034414e-01
-5.23497403e-01 -4.44388628e-01 1.99071899e-01 6.81882381e-01
3.64203125e-01 9.21798199e-02 -7.41803527e-01 2.85653800e-01
-4.94640440e-01 -7.39898145e-01 8.04461062e-01 3.53145540e-01
-4.15646821e-01 -9.45015788e-01 -8.83199751e-01 5.33638358e-01
-2.55869776e-01 -1.30768999e-01 -2.65298367e-01 6.47218347e-01
-4.38976884e-01 1.21955407e+00 -2.05710799e-01 -3.57094198e-01
3.07837963e-01 -4.11311686e-02 4.12319124e-01 -3.29378217e-01
-3.37386757e-01 5.54651618e-01 8.92658234e-02 -9.19834673e-01
-7.30393231e-01 -7.47179389e-01 -1.19906271e+00 -3.12106669e-01
-5.22536993e-01 6.35435581e-01 5.85553586e-01 5.38025379e-01
3.55716676e-01 3.81455243e-01 8.37006569e-01 -1.00949419e+00
-4.27188903e-01 -7.00110912e-01 -1.02594388e+00 3.11234117e-01
4.41344567e-02 -3.88492942e-01 -6.64002776e-01 1.64541289e-01] | [7.072559356689453, 4.854666709899902] |
8304f694-1940-4a3a-a8f7-4b5103818909 | multilingual-bidirectional-unsupervised | 2209.02821 | null | https://arxiv.org/abs/2209.02821v4 | https://arxiv.org/pdf/2209.02821v4.pdf | Multilingual Bidirectional Unsupervised Translation Through Multilingual Finetuning and Back-Translation | We propose a two-stage approach for training a single NMT model to translate unseen languages both to and from English. For the first stage, we initialize an encoder-decoder model to pretrained XLM-R and RoBERTa weights, then perform multilingual fine-tuning on parallel data in 40 languages to English. We find this model can generalize to zero-shot translations on unseen languages. For the second stage, we leverage this generalization ability to generate synthetic parallel data from monolingual datasets, then bidirectionally train with successive rounds of back-translation. Our approach, which we EcXTra (English-centric Crosslingual (X) Transfer), is conceptually simple, only using a standard cross-entropy objective throughout. It is also data-driven, sequentially leveraging auxiliary parallel data and monolingual data. We evaluate unsupervised NMT results for 7 low-resource languages, and find that each round of back-translation training further refines bidirectional performance. Our final single EcXTra-trained model achieves competitive translation performance in all translation directions, notably establishing a new state-of-the-art for English-to-Kazakh (22.9 > 10.4 BLEU). Our code is available at https://github.com/manestay/EcXTra . | ['Mohammad Sadegh Rasooli', 'Chris Callison-Burch', 'Ajay Patel', 'Bryan Li'] | 2022-09-06 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [ 1.60881311e-01 1.06471367e-02 -5.39290547e-01 -4.48495805e-01
-1.71131980e+00 -9.37976122e-01 9.33165908e-01 -5.85988462e-01
-4.95591968e-01 1.21876907e+00 4.54173505e-01 -9.18554366e-01
7.04964459e-01 -4.68577057e-01 -1.07844555e+00 -3.35147113e-01
4.42281693e-01 9.68962848e-01 -4.75168467e-01 -5.97264171e-01
-2.81666636e-01 3.49591710e-02 -4.85396147e-01 4.23865885e-01
1.09857273e+00 1.73035696e-01 1.41433820e-01 5.87540865e-01
4.53340448e-02 3.74039561e-01 -3.01781535e-01 -1.01005781e+00
4.36737448e-01 -8.24765801e-01 -1.05220962e+00 -2.15748176e-01
2.47175559e-01 -3.08036953e-01 -3.09770912e-01 7.66498089e-01
7.73242235e-01 -2.76900202e-01 3.97160292e-01 -6.78503275e-01
-1.33690548e+00 1.19874322e+00 -4.21195000e-01 6.78613409e-02
1.79051995e-01 3.70606512e-01 1.08309352e+00 -1.39484799e+00
9.40036774e-01 1.11744785e+00 6.03255808e-01 9.11621630e-01
-1.55319178e+00 -8.21177542e-01 -2.19433993e-01 -1.82888076e-01
-1.28843760e+00 -8.94039989e-01 3.35968912e-01 -1.72338039e-01
1.24823987e+00 2.16752633e-01 4.47153091e-01 1.80159032e+00
2.32915014e-01 8.87392223e-01 1.39739943e+00 -6.09116316e-01
-1.47152409e-01 2.11422980e-01 -4.29045677e-01 3.31202179e-01
-5.47702946e-02 3.73022497e-01 -5.34106433e-01 1.20516993e-01
3.93982887e-01 -4.00498807e-01 -2.05390647e-01 2.31573671e-01
-1.72180891e+00 7.98084438e-01 3.28939885e-01 3.23444217e-01
-1.66066214e-01 -7.89076876e-05 4.14763808e-01 8.52917314e-01
6.56803608e-01 5.01669645e-01 -7.48668134e-01 -2.79654294e-01
-1.06065118e+00 -1.18009701e-01 6.48261189e-01 1.32668626e+00
8.27504575e-01 9.55260471e-02 -1.86635166e-01 1.01913893e+00
-1.42047890e-02 9.02832508e-01 7.68187702e-01 -4.61232543e-01
1.01284373e+00 -2.30715051e-02 -1.24538340e-01 1.54956216e-02
2.49230466e-03 -6.29670799e-01 -7.90865779e-01 -4.72905606e-01
1.32407308e-01 -4.85458553e-01 -1.13030326e+00 1.93081152e+00
4.70237471e-02 -4.03164327e-01 4.67618018e-01 8.34993243e-01
5.58771312e-01 9.18205380e-01 -1.09284692e-01 -1.50556833e-01
1.08734512e+00 -1.35255635e+00 -6.05220854e-01 -4.71483767e-01
9.11519349e-01 -1.16945016e+00 1.48004282e+00 -1.01720775e-02
-1.29102671e+00 -5.37753105e-01 -9.16099906e-01 -5.12882531e-01
-2.78745741e-01 5.54555833e-01 3.55339766e-01 3.90165716e-01
-1.16430259e+00 4.02909338e-01 -9.67002690e-01 -5.28020442e-01
3.91902290e-02 2.20175385e-01 -4.48581457e-01 -2.87136942e-01
-1.56988776e+00 1.14742494e+00 5.99245071e-01 1.35592163e-01
-1.13561285e+00 -5.75137556e-01 -8.13377738e-01 -4.18201953e-01
1.41075300e-02 -9.38868880e-01 1.26636946e+00 -1.16259837e+00
-1.90670109e+00 1.05450380e+00 -3.47295552e-01 -3.77989978e-01
8.34510207e-01 -3.58189017e-01 -6.91254318e-01 -4.23019856e-01
3.86032552e-01 8.03093314e-01 5.42843580e-01 -1.12410080e+00
-2.59114265e-01 -4.81071211e-02 -3.48465264e-01 3.89736861e-01
-2.41436079e-01 2.18663350e-01 -7.36709595e-01 -8.61312270e-01
-1.32475838e-01 -1.28198707e+00 -7.08207786e-02 -6.85560048e-01
-8.50938618e-01 1.88020006e-01 1.26608044e-01 -1.12707400e+00
1.14440525e+00 -1.83408439e+00 4.34183568e-01 -8.63679945e-02
-3.19749922e-01 1.42116874e-01 -5.07992208e-01 7.59033740e-01
-4.46482860e-02 2.68337697e-01 -4.60958987e-01 -6.95194840e-01
-3.34234759e-02 1.89765334e-01 -5.61585248e-01 2.47549236e-01
3.98828864e-01 1.51758730e+00 -9.02506173e-01 -2.07375944e-01
-1.34596750e-01 4.31431890e-01 -4.14418131e-01 1.46043777e-01
-1.02403700e-01 8.24660540e-01 -3.25370505e-02 6.43881679e-01
5.61459363e-01 -1.48853183e-01 4.45504695e-01 4.20255475e-02
-9.63718891e-02 7.74907112e-01 -2.96230137e-01 2.16522908e+00
-8.41877759e-01 6.49957061e-01 -4.25413102e-01 -4.61405128e-01
8.63547742e-01 4.78397936e-01 -3.62311378e-02 -8.94630015e-01
-5.79232164e-02 8.88301730e-01 -1.13406293e-02 -1.23402022e-01
6.23065829e-01 -3.37244600e-01 -2.90000230e-01 5.95864117e-01
4.28712815e-01 -2.66229689e-01 1.50865689e-01 1.08311549e-01
6.84628189e-01 5.67641377e-01 1.12083271e-01 -3.15387905e-01
2.71359235e-01 2.10139230e-01 4.95161653e-01 5.61329544e-01
2.25917950e-01 7.54241645e-01 -1.19083732e-01 -3.73885542e-01
-1.50479996e+00 -1.45065641e+00 4.29200940e-02 1.07771146e+00
-2.63233453e-01 -4.72558409e-01 -7.26592243e-01 -8.29091668e-01
-2.74157852e-01 1.03740871e+00 -4.18484777e-01 -2.20994353e-01
-8.91311467e-01 -9.73448098e-01 9.29256737e-01 2.42763251e-01
3.61855179e-01 -9.07469571e-01 3.53368014e-01 2.66146451e-01
-6.78384781e-01 -1.14989460e+00 -8.91711056e-01 3.20944071e-01
-9.00542855e-01 -2.74220437e-01 -9.77431059e-01 -9.64919329e-01
5.59130847e-01 -1.42783642e-01 1.47992098e+00 -3.46687227e-01
2.76145399e-01 -1.77488610e-01 -2.42118746e-01 -7.76551664e-02
-9.31978822e-01 7.25784183e-01 2.41311684e-01 -2.11154521e-01
4.90498930e-01 -4.41887110e-01 -3.26364070e-01 2.57674724e-01
-5.66189945e-01 5.17025769e-01 7.74088025e-01 1.09697652e+00
8.61646295e-01 -8.03593159e-01 5.37271619e-01 -8.33042502e-01
7.18636692e-01 -5.07482946e-01 -4.63511467e-01 5.60847223e-01
-5.41809082e-01 1.60903230e-01 9.39835787e-01 -4.59582359e-01
-9.11484063e-01 -2.02822864e-01 -2.77310401e-01 -2.49792695e-01
1.98653266e-01 4.61477280e-01 -1.25511736e-01 3.90917063e-01
7.57564664e-01 6.62035882e-01 -3.71391475e-01 -4.60673094e-01
8.57202232e-01 9.88609672e-01 5.86142302e-01 -8.40939581e-01
1.09130478e+00 1.40814826e-01 -5.94136000e-01 -3.98666501e-01
-7.05738485e-01 7.93195218e-02 -6.82468057e-01 3.01096737e-01
8.38322341e-01 -1.44639981e+00 6.50277436e-02 2.78510392e-01
-1.12757659e+00 -7.59706438e-01 -3.27046812e-01 7.18450367e-01
-6.62556350e-01 -3.89937013e-02 -1.10362649e+00 -1.99663296e-01
-7.96236753e-01 -1.27213418e+00 1.17667365e+00 -3.23283374e-01
-3.01059067e-01 -1.13332593e+00 4.11182284e-01 5.50958574e-01
5.07145405e-01 -2.22403035e-01 7.79629290e-01 -6.93000078e-01
-5.01359761e-01 8.16312432e-02 -3.53170233e-03 4.98041987e-01
2.65911818e-01 -2.93899596e-01 -7.51423955e-01 -7.02885509e-01
-2.56660730e-01 -7.80716538e-01 7.38699198e-01 2.69754766e-03
5.24767816e-01 -3.34623635e-01 -7.02126622e-02 1.01467705e+00
1.27011275e+00 -2.32057691e-01 4.89592999e-01 4.13798183e-01
7.29782045e-01 1.50921956e-01 3.07861626e-01 -1.78368330e-01
7.05147147e-01 7.15125501e-01 -2.97930151e-01 -4.55309063e-01
-4.32264805e-01 -7.18675017e-01 9.50147092e-01 1.74829197e+00
-1.02554165e-01 -2.77732432e-01 -9.75625396e-01 7.35866189e-01
-1.42195499e+00 -6.20953798e-01 8.35702717e-02 2.22225308e+00
1.48625088e+00 2.61299070e-02 -8.01963210e-02 -6.33288503e-01
5.67963004e-01 -1.89164847e-01 -5.09751081e-01 -6.51125908e-01
-5.18458366e-01 4.58804667e-01 6.27628744e-01 8.31996083e-01
-6.71400726e-01 1.73778057e+00 6.18488979e+00 9.34965551e-01
-1.36484802e+00 4.97860998e-01 8.29697669e-01 -3.14047873e-01
-7.87247539e-01 2.83692747e-01 -8.53916109e-01 4.70430344e-01
1.34033084e+00 -1.85449213e-01 9.18266833e-01 3.44263077e-01
1.66184127e-01 6.41230047e-01 -1.25462008e+00 6.30863786e-01
-1.42128896e-02 -1.21291375e+00 3.99566919e-01 2.46979203e-02
1.23381937e+00 8.47690880e-01 2.01423854e-01 6.88529134e-01
6.97582722e-01 -9.74566519e-01 7.38303781e-01 1.22227892e-01
1.48251152e+00 -6.79369330e-01 4.28145975e-01 2.86168337e-01
-7.68212855e-01 4.48098630e-01 -3.63478065e-01 2.85305560e-01
3.82945448e-01 4.91120845e-01 -8.71454418e-01 8.72781038e-01
4.38531011e-01 7.21803308e-01 -3.82593155e-01 2.24405721e-01
-5.95453143e-01 9.01809633e-01 -2.44274840e-01 3.11632395e-01
3.29614878e-01 -3.70930582e-01 4.65815932e-01 1.49379003e+00
6.43644094e-01 -4.44622040e-01 4.08006869e-02 8.74453425e-01
-6.55245662e-01 3.49659294e-01 -6.34243786e-01 -1.53139785e-01
4.76578832e-01 9.72822785e-01 -1.99621305e-01 -5.97028077e-01
-4.99852568e-01 1.69099915e+00 6.21844471e-01 7.43262231e-01
-8.36155534e-01 -2.71477073e-01 5.01990378e-01 -2.82750189e-01
1.41495198e-01 -3.13455492e-01 -1.91015795e-01 -1.87492418e+00
2.09406912e-01 -1.29015541e+00 7.14950114e-02 -6.53597474e-01
-1.23581672e+00 1.03882837e+00 -3.71212631e-01 -1.27523148e+00
-5.45077443e-01 -3.37857217e-01 -3.45728070e-01 1.38255119e+00
-1.59631979e+00 -1.71652329e+00 4.15046394e-01 5.65888464e-01
8.19325686e-01 -3.96602452e-01 9.97139037e-01 4.09644067e-01
-5.43063939e-01 1.10846722e+00 5.53307950e-01 3.47990125e-01
1.18819678e+00 -1.13402689e+00 1.23859930e+00 1.07737207e+00
4.33129132e-01 9.16026115e-01 3.73773575e-01 -7.23640800e-01
-1.46958911e+00 -1.36562562e+00 1.58775628e+00 -8.39565217e-01
8.17596674e-01 -7.70255089e-01 -4.84926403e-01 1.22389853e+00
7.01528668e-01 -1.44908011e-01 6.61096275e-01 1.87562123e-01
-4.42537665e-01 1.01529270e-01 -7.07911015e-01 1.01618254e+00
1.19185758e+00 -9.03115988e-01 -6.14340544e-01 4.74188954e-01
9.88811731e-01 -5.08408129e-01 -1.05340326e+00 2.91305184e-01
4.91026580e-01 -4.27702844e-01 7.57272899e-01 -9.68680203e-01
7.04297245e-01 -9.64254215e-02 -4.11570668e-01 -1.85754311e+00
-1.77745908e-01 -1.13856828e+00 1.28951341e-01 1.10327935e+00
1.24109161e+00 -7.29419827e-01 3.08810502e-01 -4.93363962e-02
-2.46205807e-01 -7.07817972e-01 -9.44804788e-01 -1.11139011e+00
7.93084681e-01 -2.45910481e-01 6.26899838e-01 1.20997179e+00
-1.39242858e-01 8.44417512e-01 -7.51646161e-01 -1.20338298e-01
4.06956673e-01 2.06054404e-01 9.20487642e-01 -3.33110660e-01
-6.01303935e-01 -1.83448598e-01 3.54794770e-01 -1.33000302e+00
1.93718016e-01 -1.65953600e+00 -4.24251221e-02 -1.38641262e+00
3.54886860e-01 -2.55718321e-01 -2.17855588e-01 5.16673505e-01
-4.20921028e-01 5.75417876e-01 1.91961378e-01 4.88269955e-01
-1.67117938e-01 7.47457385e-01 1.38181043e+00 -1.37582913e-01
-2.33372197e-01 -2.30866835e-01 -6.03345811e-01 8.36264715e-02
9.60445702e-01 -5.29157639e-01 -2.37847745e-01 -1.25428069e+00
2.39656702e-01 -1.07506081e-01 -1.20266952e-01 -6.30744338e-01
-1.73283949e-01 -7.44631141e-02 4.29858536e-01 -2.87825853e-01
1.69178829e-01 -3.71329427e-01 2.57292092e-01 3.74172270e-01
-4.29062575e-01 5.28308213e-01 2.48099282e-01 -5.08849230e-03
-2.46411234e-01 2.70263702e-01 5.07068634e-01 -1.31539866e-01
-2.04952702e-01 2.78852701e-01 -6.50700331e-02 4.14498299e-01
4.48723733e-01 1.51620716e-01 -3.31746429e-01 -3.26552391e-01
-7.38730073e-01 2.13167667e-01 6.51366651e-01 7.45002747e-01
1.99349135e-01 -1.65335906e+00 -1.41291916e+00 3.41011345e-01
3.22319508e-01 -3.76363128e-01 -2.12650359e-01 8.96304786e-01
-3.54743958e-01 4.45014298e-01 -4.82418016e-02 -4.91548330e-01
-6.39413059e-01 3.97528857e-01 2.32196420e-01 -4.59938169e-01
-3.66674036e-01 7.30405450e-01 -3.97722945e-02 -1.20041585e+00
-4.64808702e-01 -7.49683157e-02 6.75238073e-01 -3.79609793e-01
1.46352038e-01 2.84503978e-02 1.94613546e-01 -7.43246436e-01
-1.50907725e-01 4.95476246e-01 -3.55078191e-01 -7.84285307e-01
1.10596645e+00 -3.24081540e-01 -1.42784178e-01 5.21710038e-01
1.43010211e+00 4.13364291e-01 -8.41274679e-01 -4.22937900e-01
-3.50423418e-02 -1.00256637e-01 -3.61160964e-01 -1.29677176e+00
-7.82547414e-01 9.00444627e-01 1.62978098e-01 -5.54742992e-01
1.07818556e+00 -2.68214702e-04 1.05535197e+00 4.52286214e-01
4.19028014e-01 -1.11716139e+00 -4.00352776e-01 8.70875597e-01
8.66288424e-01 -1.47558367e+00 -3.68590415e-01 5.93117513e-02
-7.06741393e-01 9.25725222e-01 3.70139986e-01 2.29087725e-01
1.89361215e-01 2.51319528e-01 5.23032486e-01 3.65091652e-01
-8.70892942e-01 4.84957062e-02 2.71510243e-01 2.72413969e-01
8.97211313e-01 4.30689067e-01 -4.33885038e-01 3.44648063e-01
-7.33410299e-01 5.29665686e-02 1.29053414e-01 5.78896761e-01
1.79110080e-01 -1.57582891e+00 -6.33717924e-02 7.88907483e-02
-5.54464817e-01 -8.61648500e-01 -5.26485920e-01 7.77082086e-01
-2.13832393e-01 7.71859646e-01 -1.33784920e-01 -5.51318884e-01
1.46085277e-01 2.65238762e-01 4.32587296e-01 -5.34579694e-01
-7.08041966e-01 1.64102629e-01 3.69274467e-01 -3.91812593e-01
1.06092058e-01 -5.98289311e-01 -8.15861225e-01 -4.31585491e-01
-1.48871139e-01 2.54157722e-01 7.35623240e-01 7.93193161e-01
5.49655497e-01 2.22744927e-01 7.96243191e-01 -5.31826317e-01
-7.30081975e-01 -1.26607883e+00 2.44153649e-01 4.19795632e-01
3.56304467e-01 2.27963746e-01 -1.28744736e-01 8.46832991e-02] | [11.65914249420166, 10.24752140045166] |
1ccb0fbc-69d7-49c8-b081-5cfd76d91968 | principles-and-guidelines-for-evaluating | 2306.16740 | null | https://arxiv.org/abs/2306.16740v1 | https://arxiv.org/pdf/2306.16740v1.pdf | Principles and Guidelines for Evaluating Social Robot Navigation Algorithms | A major challenge to deploying robots widely is navigation in human-populated environments, commonly referred to as social robot navigation. While the field of social navigation has advanced tremendously in recent years, the fair evaluation of algorithms that tackle social navigation remains hard because it involves not just robotic agents moving in static environments but also dynamic human agents and their perceptions of the appropriateness of robot behavior. In contrast, clear, repeatable, and accessible benchmarks have accelerated progress in fields like computer vision, natural language processing and traditional robot navigation by enabling researchers to fairly compare algorithms, revealing limitations of existing solutions and illuminating promising new directions. We believe the same approach can benefit social navigation. In this paper, we pave the road towards common, widely accessible, and repeatable benchmarking criteria to evaluate social robot navigation. Our contributions include (a) a definition of a socially navigating robot as one that respects the principles of safety, comfort, legibility, politeness, social competency, agent understanding, proactivity, and responsiveness to context, (b) guidelines for the use of metrics, development of scenarios, benchmarks, datasets, and simulators to evaluate social navigation, and (c) a design of a social navigation metrics framework to make it easier to compare results from different simulators, robots and datasets. | ['Roberto Martin-Martin', 'Alexander Toshev', 'Naoki Yokoyama', 'Peng Xu', 'Xuesu Xiao', 'Marynel Vazquez', 'Nathan Tsoi', 'Peter Trautman', 'Ada V. Taylor', 'Peter Stone', 'Phani Teja Singamaneni', 'Soeren Pirk', 'Reuth Mirksy', 'Luis J. Manso', 'Tsang-Wei Edward Lee', 'Haresh Karnan', 'Jonathan P. How', 'Justin Hart', 'Sehoon Ha', 'Michael Everett', 'Hao-Tien Lewis Chiang', 'Rohan Chandra', 'Joydeep Biswas', 'Abhijat Biswas', 'Aniket Bera', 'Rachid Alami', 'Alexandre Alahi', 'Fei Xia', 'Chengshu Li', "Claudia Perez-D'Arpino", 'Anthony Francis'] | 2023-06-29 | null | null | null | null | ['benchmarking', 'social-navigation', 'benchmarking', 'robot-navigation'] | ['miscellaneous', 'robots', 'robots', 'robots'] | [-6.14921264e-02 3.24142337e-01 1.84730917e-01 -5.03184438e-01
-9.48924944e-02 -6.92536592e-01 8.76107156e-01 1.26641124e-01
-9.22233939e-01 7.62007415e-01 3.64095062e-01 -3.34448963e-01
-3.47576499e-01 -4.04724628e-01 -2.51137376e-01 -3.84283155e-01
-2.01324999e-01 3.06687266e-01 1.46242395e-01 -8.43404770e-01
3.93999577e-01 3.21811557e-01 -1.92180562e+00 -4.98545498e-01
8.87508750e-01 4.22342509e-01 3.80703926e-01 5.58144510e-01
2.46107206e-01 9.20968115e-01 -2.96217501e-01 -1.95605919e-01
1.02167681e-01 -5.66867471e-01 -1.13999319e+00 -3.68646950e-01
-1.12906374e-01 -2.09133163e-01 1.20615371e-01 9.21545029e-01
4.90423113e-01 5.83883047e-01 8.54887545e-01 -1.70036268e+00
-6.94770396e-01 6.35666907e-01 2.04524323e-01 -5.69405973e-01
9.99278486e-01 4.31445241e-01 8.93313050e-01 -8.20064694e-02
1.00640297e+00 1.57639873e+00 7.76900828e-01 7.50742376e-01
-7.34391868e-01 -2.80890226e-01 1.21045180e-01 -5.37851006e-02
-7.09584117e-01 -3.76127064e-01 5.52507006e-02 -5.29817522e-01
9.87108350e-01 6.94021285e-02 7.55584240e-01 1.35204041e+00
2.46038333e-01 4.47484702e-01 1.00109589e+00 -3.24351639e-01
4.37321991e-01 1.81992441e-01 1.70709789e-01 3.15658569e-01
2.83896983e-01 1.98085353e-01 -2.84821279e-02 -1.68700650e-01
2.60740817e-01 -3.02692622e-01 -1.00587130e-01 -1.00842178e+00
-1.47592986e+00 7.16240525e-01 5.27682781e-01 2.19483972e-01
-5.88010490e-01 2.11718500e-01 7.76964843e-01 4.31880444e-01
-2.93986797e-01 9.12938118e-01 -1.62176833e-01 -8.16421211e-01
5.36539912e-01 5.18583417e-01 1.23223853e+00 9.21564698e-01
4.99548227e-01 -3.22752416e-01 4.52672392e-01 9.40770626e-01
6.07429266e-01 5.76536298e-01 5.41738451e-01 -1.80563545e+00
1.05659319e-02 4.72863466e-01 5.92554927e-01 -1.28822958e+00
-8.17511320e-01 1.20425582e-01 -6.17951527e-02 5.86723924e-01
3.86667699e-01 -3.43205601e-01 -3.13618153e-01 1.94119620e+00
2.57291526e-01 -8.15304577e-01 6.98063374e-01 8.26327860e-01
5.91656268e-01 5.53788245e-01 2.81724721e-01 2.27520898e-01
1.05127406e+00 -1.06087017e+00 -4.42912072e-01 -5.05271435e-01
1.12535572e+00 -4.43353385e-01 1.33855188e+00 1.39376596e-01
-6.78403676e-01 -1.32966936e-01 -1.01274514e+00 -1.61092326e-01
-5.64464390e-01 -6.71046615e-01 6.66776717e-01 6.93374813e-01
-1.37498009e+00 3.61695200e-01 -4.26264614e-01 -1.33337748e+00
-3.31214994e-01 3.77021819e-01 -6.21075213e-01 -8.22687242e-03
-1.02665162e+00 1.56958413e+00 1.60210252e-01 -4.95133549e-02
-6.03401005e-01 3.52361947e-01 -1.07634878e+00 -4.10123795e-01
6.77180812e-02 -6.09280884e-01 1.45504010e+00 -1.01575911e+00
-1.74414599e+00 7.29618311e-01 3.03115487e-01 -2.45838523e-01
5.69393933e-01 -2.14104474e-01 -1.09504677e-01 -4.06785160e-01
5.68349838e-01 1.03657687e+00 7.44260177e-02 -1.50646996e+00
-9.36461568e-01 -1.24181248e-01 3.85092586e-01 7.27398932e-01
-4.94651347e-02 1.68636665e-01 -3.63393016e-02 1.42554909e-01
-1.90511551e-02 -1.44218731e+00 -7.11644888e-01 -1.62365541e-01
4.23298031e-02 -1.67478457e-01 5.08288622e-01 -1.55285046e-01
2.81391770e-01 -2.23637986e+00 2.71489143e-01 9.45048481e-02
-5.40964911e-03 -1.77778959e-01 -3.75555664e-01 6.85985029e-01
3.76124531e-01 -1.43222138e-01 -1.17901340e-01 -1.27977341e-01
5.20364821e-01 2.41073355e-01 -1.15574328e-02 4.73771513e-01
-5.44807076e-01 6.00818932e-01 -1.60491896e+00 -1.39356196e-01
4.58738148e-01 4.87108260e-01 -6.90541565e-01 -1.24346733e-01
2.89942205e-01 5.18493891e-01 -2.55401999e-01 2.98365504e-01
5.32272086e-02 2.87731141e-01 1.92909390e-01 5.40258408e-01
-3.68787825e-01 2.93586224e-01 -9.04469788e-01 1.38782716e+00
-5.89885235e-01 8.81021142e-01 2.63503939e-01 -2.93777823e-01
1.03068650e+00 -2.97858864e-02 3.01883638e-01 -8.99012208e-01
3.46318305e-01 5.35580516e-01 1.72992751e-01 -6.95206642e-01
1.01010406e+00 3.70133758e-01 -4.85697299e-01 6.62326097e-01
-3.82024199e-01 -5.76960802e-01 7.31256083e-02 6.12136722e-03
1.18836164e+00 3.92833203e-01 3.70733947e-01 -4.30486828e-01
3.05891573e-01 2.62117684e-01 2.17399314e-01 8.93847883e-01
-9.37227011e-01 -2.96751633e-02 3.17250788e-01 -2.49554619e-01
-9.48931873e-01 -9.47976768e-01 2.17467949e-01 1.51530159e+00
8.15286934e-01 -2.07825184e-01 -8.47288311e-01 -3.40952903e-01
-6.61699846e-02 1.11766481e+00 -5.14713287e-01 -1.40008271e-01
-5.27777433e-01 -3.09315473e-01 5.08182287e-01 1.45008624e-01
3.65557700e-01 -1.53305399e+00 -1.23874879e+00 3.11724003e-02
-2.72059888e-01 -9.70677137e-01 9.85378772e-02 2.97335535e-01
-4.55700427e-01 -1.01510823e+00 -3.94972980e-01 -1.23103690e+00
5.66484928e-01 6.96126044e-01 7.73965180e-01 1.74259488e-03
3.27908844e-01 8.52935791e-01 -7.41056323e-01 -3.13629448e-01
-8.24651599e-01 1.98808871e-02 3.11135530e-01 -8.49417031e-01
2.04724208e-01 -3.96847308e-01 -6.06236100e-01 8.53942633e-01
-4.58721489e-01 -3.17894012e-01 3.89622927e-01 5.18371344e-01
-3.33097875e-01 -3.10281515e-01 6.29791677e-01 -3.49277675e-01
1.25016105e+00 -8.53163064e-01 -7.98409730e-02 -8.64989907e-02
-8.15180659e-01 -3.32647890e-01 3.58724147e-01 -9.56565961e-02
-8.11645567e-01 -1.37017041e-01 -5.42227030e-02 8.31760287e-01
-4.40919042e-01 3.03519160e-01 4.07742932e-02 -1.38982639e-01
1.20887160e+00 -2.00679347e-01 5.88292241e-01 2.74669230e-01
6.89558566e-01 9.62016463e-01 3.53958309e-01 -5.30796766e-01
4.04883951e-01 4.75600183e-01 -4.22240734e-01 -1.00582564e+00
1.95000067e-01 -2.82498062e-01 -2.70657063e-01 -5.31919360e-01
6.93511128e-01 -6.22780383e-01 -1.02079105e+00 4.31126744e-01
-8.13313782e-01 -9.54540670e-01 -3.08972776e-01 5.77866375e-01
-1.04014671e+00 5.13086081e-01 -4.31365043e-01 -9.92340922e-01
2.89482847e-02 -1.56959236e+00 7.35362351e-01 1.69354960e-01
-1.00424612e+00 -8.82251441e-01 2.33438075e-01 3.08801621e-01
7.61689901e-01 2.26686373e-01 6.63921177e-01 -5.23374736e-01
-1.84775561e-01 -2.81310439e-01 -2.00818405e-01 2.20195875e-01
2.56375968e-01 -1.70001127e-02 -5.25215685e-01 -6.62095174e-02
-1.13030747e-01 -6.24099731e-01 2.72663385e-02 1.22251987e-01
-6.70458749e-02 2.57251617e-02 -4.40238386e-01 7.97022358e-02
7.78401434e-01 8.36477101e-01 6.48003459e-01 1.14899182e+00
3.32903475e-01 1.41851509e+00 1.05393815e+00 -7.95610026e-02
1.14557958e+00 5.30993760e-01 6.30868495e-01 2.95709401e-01
4.63904291e-01 1.14147060e-01 6.17022097e-01 5.10496497e-01
4.57134992e-02 -3.24477136e-01 -1.13259804e+00 5.98299742e-01
-2.10982180e+00 -7.70306051e-01 2.00559665e-02 1.85605013e+00
1.16048835e-01 1.41712185e-02 2.80198097e-01 -9.69939753e-02
5.93647480e-01 3.44897844e-02 -4.38748360e-01 -9.07110155e-01
7.80431852e-02 -8.87573242e-01 4.19158816e-01 7.39132643e-01
-8.03102851e-01 1.16105676e+00 6.46563339e+00 6.66173473e-02
-7.51580536e-01 -1.16935201e-01 2.76506007e-01 4.07212049e-01
-2.23696366e-01 -3.20605258e-03 -2.65423179e-01 1.71512216e-01
7.82871068e-01 -7.71540627e-02 6.00610256e-01 1.29825103e+00
3.66569340e-01 -4.29778397e-01 -1.13913524e+00 7.11616516e-01
8.81242678e-02 -4.84803289e-01 -4.30264622e-01 -2.52352674e-02
3.21731746e-01 3.69323730e-01 1.01358615e-01 6.12057805e-01
9.09173727e-01 -1.00202799e+00 1.02465463e+00 1.87945262e-01
6.17133416e-02 -4.64975923e-01 8.47355962e-01 1.62952229e-01
-6.90499783e-01 -3.97688568e-01 -8.72462913e-02 -4.87978607e-01
3.33919883e-01 -5.90662360e-01 -8.25244069e-01 8.10038075e-02
8.80551338e-01 6.80968404e-01 -2.18953744e-01 9.45113480e-01
3.54507565e-02 -2.67538577e-01 -2.10612833e-01 -8.68082225e-01
6.32132471e-01 -5.47703981e-01 6.19347215e-01 9.16129231e-01
2.65113264e-01 -1.77298024e-01 5.10466227e-04 2.63956517e-01
6.12357676e-01 4.64070708e-01 -1.15437293e+00 3.68449353e-02
5.66854596e-01 8.53056669e-01 -9.41323340e-01 4.19442266e-01
-1.98309571e-01 8.36933672e-01 9.00973156e-02 2.67533869e-01
-6.71792746e-01 -5.71430802e-01 9.12153780e-01 -3.13941091e-01
-4.46370751e-01 -5.05444467e-01 -2.16581553e-01 -3.00626516e-01
-2.34111249e-01 -9.60645020e-01 2.39555035e-02 -8.50098968e-01
-8.94815803e-01 8.34413171e-01 -1.50986716e-01 -1.30131233e+00
-5.87225318e-01 -5.80366313e-01 -2.04653397e-01 3.53632659e-01
-1.01152968e+00 -8.83179247e-01 -5.98380029e-01 -1.21589042e-02
2.86979884e-01 -1.71659499e-01 9.57143128e-01 -9.93207023e-02
-2.02227071e-01 2.96357870e-01 2.47394606e-01 -3.19980085e-01
8.15566123e-01 -1.02073908e+00 7.17200994e-01 1.18430912e-01
-7.64557838e-01 8.16975772e-01 1.01314175e+00 -5.90626836e-01
-1.33235967e+00 -5.58006287e-01 4.59646255e-01 -6.77995861e-01
7.04521596e-01 -1.34007886e-01 -2.49938607e-01 6.93646133e-01
2.37429738e-01 -9.31552768e-01 5.54580510e-01 2.55265445e-01
-1.94986183e-02 2.14821428e-01 -1.32483435e+00 1.38353491e+00
1.24700356e+00 -2.91568905e-01 -5.66431761e-01 1.94773510e-01
8.29470873e-01 -1.84045464e-01 -4.69227821e-01 1.25541702e-01
9.91807103e-01 -1.22144520e+00 7.57955194e-01 -7.70097598e-02
1.12700596e-01 -2.60327339e-01 -5.08486450e-01 -1.39011407e+00
-1.58534899e-01 -7.62985885e-01 9.34094131e-01 8.81442904e-01
6.39897645e-01 -1.27069545e+00 5.89852870e-01 9.18496311e-01
-4.46221143e-01 -3.20371568e-01 -6.51657283e-01 -7.12353170e-01
5.03932685e-02 -3.62745047e-01 5.35559416e-01 9.24005151e-01
7.03574717e-01 3.25079590e-01 -1.69955179e-01 -1.84531301e-01
3.07663500e-01 -6.34999931e-01 1.32547724e+00 -1.28035080e+00
3.38563591e-01 -7.98829496e-01 -7.29966283e-01 -9.44158077e-01
4.15889740e-01 -5.60710967e-01 6.41089022e-01 -1.95936394e+00
-1.33588389e-01 -7.93841302e-01 2.13891700e-01 7.87982717e-03
2.64689803e-01 -8.66834819e-02 1.71393991e-01 2.39285499e-01
-9.42921698e-01 5.51721454e-01 9.85004246e-01 2.97093719e-01
-5.21084189e-01 -2.99816608e-01 -9.20238554e-01 9.59947169e-01
8.44200015e-01 -1.41053319e-01 -6.32088721e-01 -3.18329811e-01
5.19770980e-01 -2.04208851e-01 3.12018543e-01 -1.21530986e+00
4.87113267e-01 -3.24022174e-01 -4.18869793e-01 3.75206359e-02
5.27951121e-01 -9.11545575e-01 1.49445727e-01 6.81421101e-01
-4.06687617e-01 2.82081604e-01 -7.26858675e-02 4.72172529e-01
1.24908030e-01 -3.30146164e-01 5.63487113e-01 -1.06824316e-01
-1.07110906e+00 -6.10137820e-01 -9.53726947e-01 -9.00269672e-02
1.42452848e+00 -7.65447855e-01 -5.56143582e-01 -9.37232137e-01
-3.30738157e-01 6.19972944e-01 1.38079405e+00 8.16852033e-01
2.97067076e-01 -8.26609790e-01 -2.58831322e-01 -1.85289621e-01
3.84978771e-01 -2.08923757e-01 1.03327028e-01 4.03999239e-01
-1.04794967e+00 2.68468708e-01 -6.06796563e-01 -4.60093051e-01
-8.89590979e-01 1.53189987e-01 4.55912977e-01 6.34945989e-01
-4.30028260e-01 5.57170808e-01 3.10455374e-02 -1.30428326e+00
5.30098677e-01 -1.22979328e-01 -5.10717511e-01 -1.05693368e-02
7.34735802e-02 7.06838310e-01 -3.88707250e-01 -1.13883221e+00
-3.90019655e-01 4.22426432e-01 1.82550043e-01 -4.85751867e-01
1.07328498e+00 -6.16115093e-01 -2.97675937e-01 6.04206622e-01
7.60318995e-01 -1.77924305e-01 -9.25319791e-01 3.92289639e-01
3.55564743e-01 -5.48861884e-02 -5.65452874e-01 -7.41491795e-01
-1.34650707e-01 1.36786863e-01 7.56754994e-01 5.22553265e-01
4.46598560e-01 -1.31394118e-01 5.80190241e-01 1.04054749e+00
9.74284351e-01 -1.35480595e+00 7.03491718e-02 1.03584719e+00
8.90043139e-01 -1.27599013e+00 -4.08981085e-01 -1.26165256e-01
-9.97160912e-01 7.45504856e-01 6.41564786e-01 1.34367436e-01
5.09019077e-01 -9.31822211e-02 5.88467717e-01 -7.51184225e-02
-3.93921018e-01 -1.29101291e-01 -6.04411781e-01 1.11289668e+00
2.43209541e-01 1.82826236e-01 -3.01186293e-01 7.59162083e-02
-7.60196328e-01 -3.05362254e-01 8.61412048e-01 1.24854887e+00
-7.57085085e-01 -8.84664714e-01 -2.52859294e-01 1.41227484e-01
1.72458172e-01 4.91855711e-01 -6.77004457e-01 1.08857405e+00
-2.76654363e-01 1.54228270e+00 -1.76534191e-01 -5.80001473e-01
7.02290893e-01 -3.18185717e-01 -6.76419735e-02 -4.75529701e-01
-5.42936862e-01 -7.25872159e-01 6.02037430e-01 -6.78167939e-01
-4.06211644e-01 -9.26873744e-01 -1.38018930e+00 -4.19918537e-01
-4.02655862e-02 4.88949627e-01 1.19073915e+00 6.46564722e-01
4.25063163e-01 -4.28519547e-02 4.58533376e-01 -1.27118278e+00
-4.15267169e-01 -9.14006233e-01 -1.79334506e-01 3.57388496e-01
2.24977672e-01 -8.90035927e-01 -7.03927100e-01 -5.74831426e-01] | [4.822906970977783, 0.9852423071861267] |
1e391622-2420-4891-9bab-d11ccf1de402 | a-comparative-analysis-of-latent-regressor | 2302.13678 | null | https://arxiv.org/abs/2302.13678v1 | https://arxiv.org/pdf/2302.13678v1.pdf | A Comparative Analysis Of Latent Regressor Losses For Singing Voice Conversion | Previous research has shown that established techniques for spoken voice conversion (VC) do not perform as well when applied to singing voice conversion (SVC). We propose an alternative loss component in a loss function that is otherwise well-established among VC tasks, which has been shown to improve our model's SVC performance. We first trained a singer identity embedding (SIE) network on mel-spectrograms of singer recordings to produce singer-specific variance encodings using contrastive learning. We subsequently trained a well-known autoencoder framework (AutoVC) conditioned on these SIEs, and measured differences in SVC performance when using different latent regressor loss components. We found that using this loss w.r.t. SIEs leads to better performance than w.r.t. bottleneck embeddings, where converted audio is more natural and specific towards target singers. The inclusion of this loss component has the advantage of explicitly forcing the network to reconstruct with timbral similarity, and also negates the effect of poor disentanglement in AutoVC's bottleneck embeddings. We demonstrate peculiar diversity between computational and human evaluations on singer-converted audio clips, which highlights the necessity of both. We also propose a pitch-matching mechanism between source and target singers to ensure these evaluations are not influenced by differences in pitch register. | ['Simon Dixon', "Brendan O'Connor"] | 2023-02-27 | null | null | null | null | ['voice-conversion', 'voice-conversion'] | ['audio', 'speech'] | [ 3.40307169e-02 -8.57407153e-02 3.15592080e-01 -1.24811582e-01
-9.24698830e-01 -7.81057537e-01 5.79709232e-01 -2.30576500e-01
-6.70487821e-01 4.87857819e-01 7.37712026e-01 -1.44295141e-01
-1.50288031e-01 -4.37032789e-01 -4.99140620e-01 -6.34507895e-01
-1.14675857e-01 2.96826363e-01 -9.38730780e-03 -3.61244321e-01
-2.25455210e-01 3.96405429e-01 -1.75334287e+00 2.60118812e-01
5.34314513e-01 6.88713372e-01 1.24957241e-01 1.15652347e+00
1.85071513e-01 4.48220193e-01 -9.57231641e-01 -3.82141083e-01
2.89872974e-01 -8.28723252e-01 -7.34288275e-01 -3.05457771e-01
6.13693893e-01 2.42808536e-02 -1.80165380e-01 6.43818080e-01
9.62838948e-01 4.40505296e-01 8.08421552e-01 -6.67461276e-01
-5.84151566e-01 6.54406428e-01 1.49396986e-01 2.53425270e-01
3.96223038e-01 1.11361854e-01 1.42516804e+00 -1.01894784e+00
5.16772568e-01 1.19394493e+00 1.04496551e+00 7.83491015e-01
-1.66695654e+00 -5.69937408e-01 -4.92839575e-01 5.95378391e-02
-1.26959431e+00 -8.09318304e-01 8.97717774e-01 -2.57799029e-01
8.85612011e-01 6.78900838e-01 7.03243613e-01 1.14800048e+00
-1.48790583e-01 4.17247772e-01 1.00672114e+00 -6.90622330e-01
7.85738826e-02 3.63168985e-01 -4.26130027e-01 7.72721469e-02
-4.62720484e-01 5.94317198e-01 -6.03937685e-01 -1.14530073e-02
6.20578110e-01 -7.81797230e-01 -6.86861336e-01 -9.92716998e-02
-1.05475855e+00 8.82078230e-01 2.05688760e-01 7.00326920e-01
-1.78225473e-01 1.24084711e-01 6.52101815e-01 7.05363512e-01
3.71392667e-01 1.01021957e+00 -2.23906726e-01 -2.66781330e-01
-1.10700035e+00 2.34591693e-01 9.64653254e-01 2.68458456e-01
1.49308041e-01 6.50430799e-01 -3.16399485e-01 1.31042409e+00
-6.59033880e-02 1.60312444e-01 8.73051524e-01 -1.10400712e+00
1.17693454e-01 -2.43214622e-01 -1.69200256e-01 -7.89391220e-01
-1.46227285e-01 -7.24783897e-01 -4.39860821e-01 3.29726577e-01
4.18229520e-01 -2.89002843e-02 -4.25905526e-01 2.05884552e+00
3.60780172e-02 2.87638992e-01 1.19748868e-01 1.07679772e+00
5.45293868e-01 6.89443111e-01 -1.72535405e-01 -2.41800338e-01
1.06032920e+00 -7.40763366e-01 -8.99229169e-01 1.22377001e-01
1.78291917e-01 -1.08531034e+00 1.83701980e+00 3.69931221e-01
-1.28526247e+00 -1.01598096e+00 -1.26996708e+00 -1.45863116e-01
-2.33699739e-01 6.70572594e-02 2.64627427e-01 8.65944147e-01
-1.22248697e+00 1.11178517e+00 -4.96490866e-01 -3.92223038e-02
-1.97226614e-01 4.07010823e-01 -3.83070439e-01 5.88063061e-01
-1.45895576e+00 1.04929733e+00 2.43824914e-01 -1.40531912e-01
-8.53535414e-01 -9.20900285e-01 -6.53231204e-01 1.85334161e-01
-1.26068205e-01 -4.44498718e-01 1.57727206e+00 -1.20069766e+00
-1.99309433e+00 6.75128400e-01 2.14168146e-01 -5.22807121e-01
4.61314172e-01 -2.67447412e-01 -7.54543245e-01 1.10072315e-01
-2.76747704e-01 5.47023356e-01 1.13945508e+00 -1.11916351e+00
-3.34810197e-01 4.68485095e-02 -3.11992049e-01 3.54520023e-01
-6.84852779e-01 3.70623656e-02 5.80371730e-02 -9.04001653e-01
-4.54384297e-01 -7.78422773e-01 2.75822788e-01 -2.85383642e-01
-6.82196841e-02 -2.35670835e-01 6.38728380e-01 -8.10027242e-01
1.53456187e+00 -2.16434407e+00 3.92629206e-01 4.79821935e-02
-1.78821787e-01 6.33556664e-01 -4.45744872e-01 6.38807952e-01
-2.98919797e-01 1.54827032e-02 -2.55176425e-01 -6.32514834e-01
8.50958973e-02 2.29496345e-01 -5.54669619e-01 3.49007249e-01
1.98846161e-01 4.69204664e-01 -6.16541266e-01 -2.09497988e-01
1.54201031e-01 8.87778461e-01 -9.35856998e-01 4.04117137e-01
5.54015189e-02 3.95156235e-01 5.51108122e-01 2.07157414e-02
2.12492213e-01 6.13485038e-01 6.10865392e-02 -2.86537170e-01
-3.24349374e-01 7.89710522e-01 -9.92447793e-01 1.53143251e+00
-8.82353306e-01 8.89708877e-01 1.88586935e-01 -7.94069648e-01
9.73219931e-01 8.63346756e-01 2.93206125e-02 -4.21723515e-01
6.12336546e-02 5.52289307e-01 4.56143945e-01 -8.60129222e-02
5.00507534e-01 -7.26347446e-01 2.04669401e-01 3.43617380e-01
6.60629630e-01 -5.61176479e-01 -1.29436329e-01 -3.76914352e-01
7.03124344e-01 1.79466996e-02 -2.46968865e-02 -3.27792287e-01
6.15081191e-01 -4.66058761e-01 2.99649924e-01 5.45311034e-01
-1.00836791e-01 9.06723678e-01 2.76782125e-01 1.26760185e-01
-1.14421809e+00 -1.24520576e+00 -1.54974833e-01 1.20047700e+00
-5.37756264e-01 -4.94325668e-01 -9.20383573e-01 -3.02420080e-01
-1.76101401e-01 1.14259470e+00 -3.54880154e-01 -4.78480726e-01
-8.19945514e-01 -1.56134292e-01 1.04718173e+00 4.04662073e-01
-2.16289997e-01 -1.26023602e+00 -4.19747323e-01 3.96027803e-01
-5.04487939e-03 -7.08102345e-01 -9.09590542e-01 3.82074773e-01
-7.25409091e-01 -6.82348013e-01 -8.61946225e-01 -8.97519112e-01
-8.45420584e-02 -2.10264355e-01 1.17918563e+00 -3.25269431e-01
-1.54765800e-01 6.38395607e-01 -2.45662853e-01 -3.74585956e-01
-8.57235670e-01 1.18947715e-01 5.22869945e-01 -8.99813026e-02
1.83059908e-02 -8.90132964e-01 -3.60774994e-01 2.35272035e-01
-9.72886980e-01 -4.29339051e-01 1.42082334e-01 1.14011323e+00
3.85076851e-01 -9.75280162e-03 8.12034488e-01 -5.52070141e-01
1.22064006e+00 -4.11826894e-02 -2.80095726e-01 -1.22939013e-01
-6.07721090e-01 9.33072418e-02 9.75024462e-01 -7.68210232e-01
-9.20420408e-01 -2.34532237e-01 -6.11879230e-01 -7.08527863e-01
-4.37448882e-02 2.64539719e-01 -5.83579503e-02 2.01379612e-01
9.02894497e-01 4.03869562e-02 2.49660611e-01 -7.47352064e-01
4.25372064e-01 8.45174372e-01 9.08858538e-01 -4.45266545e-01
8.38487566e-01 -2.34109908e-02 -3.72395694e-01 -9.77324784e-01
-4.54243124e-01 -4.29057151e-01 -4.17901278e-01 -1.51781023e-01
7.34044075e-01 -5.94430923e-01 -6.96226776e-01 1.09921470e-01
-1.14446723e+00 -3.48341018e-01 -8.82353485e-01 9.52357590e-01
-8.27999771e-01 1.89133689e-01 -6.96245193e-01 -8.30230236e-01
-4.22616631e-01 -9.51683760e-01 6.29714966e-01 -1.46043599e-01
-5.59602022e-01 -1.14318871e+00 5.84819496e-01 -1.89159736e-02
7.23852575e-01 -8.57240781e-02 8.96553576e-01 -8.10644150e-01
2.33124271e-01 -8.24954286e-02 4.22986209e-01 1.18303597e+00
2.92196631e-01 -4.84795757e-02 -1.54584873e+00 -3.61153662e-01
3.29917073e-01 -3.14877689e-01 9.48777974e-01 2.04552487e-01
7.69783318e-01 -4.89922762e-01 5.33540726e-01 6.09400034e-01
1.03362060e+00 1.43836826e-01 5.95519543e-01 -7.40162283e-02
4.39155757e-01 7.88236737e-01 1.59362406e-01 -6.56037033e-03
-2.77139246e-01 1.03480911e+00 -2.30440479e-02 -1.94015667e-01
-7.05938339e-01 -5.62827885e-01 7.10160196e-01 1.56101513e+00
-3.18551362e-01 -6.05344996e-02 -5.39622426e-01 6.12236381e-01
-1.18653893e+00 -1.04362118e+00 2.15892524e-01 2.39413667e+00
1.26674187e+00 1.33153439e-01 2.59835839e-01 6.03880525e-01
5.29742658e-01 2.86458790e-01 -1.28808543e-01 -8.75348449e-01
-1.49195850e-01 9.96059120e-01 1.12246640e-01 7.76500762e-01
-8.27274263e-01 9.25105631e-01 6.33121967e+00 1.05453491e+00
-1.22258723e+00 3.17607194e-01 3.79931182e-02 -2.58130431e-01
-4.92383480e-01 -1.47784248e-01 -4.16578680e-01 2.08451033e-01
1.35067201e+00 -1.36088897e-02 8.33731830e-01 6.44205511e-01
2.40493894e-01 4.88712132e-01 -1.31133258e+00 8.44300210e-01
2.01529548e-01 -1.05501211e+00 -7.19148144e-02 -1.51473820e-01
4.25549716e-01 -2.86127985e-01 4.57594782e-01 5.17167687e-01
-7.47092068e-02 -1.32892108e+00 9.51789260e-01 2.94184566e-01
1.06542623e+00 -8.91508102e-01 6.08701646e-01 8.27546865e-02
-1.00090933e+00 1.05282322e-01 -2.94723451e-01 -6.26598969e-02
-3.75913493e-02 1.10747024e-01 -1.24930322e+00 2.71010488e-01
3.85179877e-01 1.75840691e-01 -2.33809084e-01 7.62217820e-01
-1.39450699e-01 1.06327224e+00 -1.77148074e-01 1.43006086e-01
7.66369402e-02 -1.27537891e-01 8.97643030e-01 1.50709939e+00
3.00995559e-01 -4.65231031e-01 -4.69649643e-01 8.98061275e-01
-2.68201288e-02 3.28101903e-01 -5.16977787e-01 -2.33252749e-01
5.80315709e-01 8.74287486e-01 -1.90361649e-01 6.58968240e-02
-1.34752184e-01 1.13685179e+00 1.19996309e-01 2.87350625e-01
-5.86517930e-01 -5.61722457e-01 8.80785525e-01 1.77811012e-01
3.75305891e-01 -1.86653793e-01 4.82011326e-02 -7.96027124e-01
-1.97056457e-01 -1.05428588e+00 1.20714031e-01 -5.41233361e-01
-1.32760572e+00 9.19756174e-01 -2.27051511e-01 -1.22717667e+00
-5.42581439e-01 -5.80089629e-01 -7.82011449e-01 1.18831027e+00
-1.35532701e+00 -8.60678494e-01 3.49930167e-01 5.61331451e-01
6.21746838e-01 -3.44302744e-01 1.28239071e+00 5.09374499e-01
-1.93321481e-01 9.01737869e-01 2.71323975e-02 -1.16675854e-01
9.56239402e-01 -1.48492503e+00 3.01284462e-01 5.04417479e-01
7.27849185e-01 6.38386846e-01 8.99284601e-01 5.18108793e-02
-9.11839843e-01 -7.92900741e-01 1.07349265e+00 -4.16796476e-01
5.20334721e-01 -5.08225501e-01 -1.02810156e+00 2.75145620e-01
4.18159992e-01 -1.64225996e-01 9.11078811e-01 2.55941987e-01
-4.43554163e-01 -2.96326816e-01 -6.79658115e-01 6.11728549e-01
7.72118986e-01 -1.25506628e+00 -9.60826159e-01 -1.04912817e-01
9.24672961e-01 -1.09306633e-01 -8.42268884e-01 2.79099703e-01
7.09677339e-01 -1.13568914e+00 1.18570173e+00 -7.40373731e-01
2.18532294e-01 -1.25134468e-01 -1.96312949e-01 -1.69748521e+00
-1.69191867e-01 -7.44605541e-01 1.64574459e-01 1.52984560e+00
4.30979609e-01 -5.16868114e-01 1.68247432e-01 -2.56059803e-02
-5.45262635e-01 -4.05932248e-01 -1.09069622e+00 -1.09993088e+00
3.99282902e-01 -5.98896444e-01 3.44669908e-01 9.07712340e-01
-1.02305099e-01 4.63915706e-01 -5.15772939e-01 -2.40903825e-01
4.66903225e-02 -2.57373452e-01 5.45386016e-01 -1.18898141e+00
-7.65339375e-01 -5.96572697e-01 -9.89187285e-02 -6.53345048e-01
2.90689886e-01 -1.02638745e+00 1.42099923e-02 -9.64730442e-01
-5.47492385e-01 -2.02631742e-01 -6.39987826e-01 1.75539047e-01
4.76887822e-02 4.79597509e-01 4.58020717e-01 2.23785087e-01
2.50303596e-01 7.92201042e-01 1.23348415e+00 1.32127255e-01
-5.50333381e-01 2.69252837e-01 -3.54403764e-01 6.98441744e-01
8.86851966e-01 -4.71282780e-01 -4.37787116e-01 -8.92342329e-02
-1.48549244e-01 1.09005660e-01 2.79476196e-01 -1.24745309e+00
2.63573602e-02 4.66808945e-01 1.63377449e-01 -7.49770775e-02
7.73308396e-01 -6.23605728e-01 2.02977911e-01 4.31861937e-01
-7.40630865e-01 -2.15795755e-01 3.21762085e-01 2.08076984e-01
-5.95612228e-01 -4.05733764e-01 8.96010756e-01 7.23758116e-02
-1.59839600e-01 -2.60726959e-01 -3.25514317e-01 1.32157907e-01
2.62839407e-01 -1.86356664e-01 3.74573052e-01 -5.69168746e-01
-7.93199658e-01 -5.31937242e-01 2.10613310e-01 4.92448598e-01
3.93822581e-01 -1.49471927e+00 -8.09113741e-01 4.06063020e-01
-1.74896315e-01 -6.75474584e-01 3.12935077e-02 7.93391407e-01
-3.13492477e-01 3.47158760e-01 -1.71620637e-01 -2.69135028e-01
-1.37069547e+00 3.11547339e-01 5.16145408e-01 -1.05432682e-02
-4.26536173e-01 9.71617997e-01 3.38043161e-02 -6.64781511e-01
4.63986397e-01 -2.59511799e-01 -2.72746712e-01 3.92883092e-01
1.89159647e-01 4.38198209e-01 1.93769291e-01 -7.74396896e-01
-1.47175431e-01 4.69754905e-01 2.88920492e-01 -6.86931968e-01
1.17519629e+00 1.48611188e-01 3.45239937e-01 8.12439263e-01
1.37762618e+00 7.33671069e-01 -1.07386100e+00 9.89347175e-02
-4.48372103e-02 -2.77935714e-01 2.74830580e-01 -7.66150951e-01
-7.63019264e-01 1.08986306e+00 6.49622440e-01 2.68235415e-01
1.09996939e+00 -2.10553244e-01 7.71280229e-01 -7.78543502e-02
-1.04147725e-01 -1.25431287e+00 1.86287329e-01 5.00563145e-01
1.35739517e+00 -6.99054718e-01 -4.27459866e-01 -2.50720978e-02
-7.49611974e-01 1.19902992e+00 2.29325250e-01 -3.64920914e-01
5.45826674e-01 1.78346694e-01 2.46713206e-01 1.43507481e-01
-6.38538718e-01 -2.50858754e-01 5.04861951e-01 6.17774844e-01
6.90281630e-01 1.76539257e-01 -4.25306559e-01 6.88690662e-01
-8.57069016e-01 -4.27379161e-01 2.25850388e-01 1.55225754e-01
-1.22633837e-01 -1.33675086e+00 -4.78935003e-01 2.80147772e-02
-5.15572727e-01 -2.07013696e-01 -5.19807398e-01 8.47186327e-01
2.67574430e-01 8.91940415e-01 2.40256011e-01 -6.18907094e-01
5.66203415e-01 3.93389165e-01 4.40949112e-01 -7.15406477e-01
-1.06540835e+00 3.80968839e-01 2.52842754e-01 -2.34333277e-01
-2.55918950e-01 -5.46903610e-01 -9.08814669e-01 4.44848873e-02
-4.60900456e-01 5.69612205e-01 7.50787079e-01 3.50677639e-01
3.54584046e-02 8.15122604e-01 7.94162512e-01 -9.11618412e-01
-9.87382114e-01 -1.16399217e+00 -8.11956048e-01 5.39926350e-01
5.51620305e-01 -4.27122623e-01 -8.18705142e-01 6.71597421e-02] | [15.462530136108398, 6.018396377563477] |
f174b5e6-0119-4b6d-ad58-791f49a92150 | gentus-simulating-user-behaviour-and-language | 2208.10817 | null | https://arxiv.org/abs/2208.10817v1 | https://arxiv.org/pdf/2208.10817v1.pdf | GenTUS: Simulating User Behaviour and Language in Task-oriented Dialogues with Generative Transformers | User simulators (USs) are commonly used to train task-oriented dialogue systems (DSs) via reinforcement learning. The interactions often take place on semantic level for efficiency, but there is still a gap from semantic actions to natural language, which causes a mismatch between training and deployment environment. Incorporating a natural language generation (NLG) module with USs during training can partly deal with this problem. However, since the policy and NLG of USs are optimised separately, these simulated user utterances may not be natural enough in a given context. In this work, we propose a generative transformer-based user simulator (GenTUS). GenTUS consists of an encoder-decoder structure, which means it can optimise both the user policy and natural language generation jointly. GenTUS generates both semantic actions and natural language utterances, preserving interpretability and enhancing language variation. In addition, by representing the inputs and outputs as word sequences and by using a large pre-trained language model we can achieve generalisability in feature representation. We evaluate GenTUS with automatic metrics and human evaluation. Our results show that GenTUS generates more natural language and is able to transfer to an unseen ontology in a zero-shot fashion. In addition, its behaviour can be further shaped with reinforcement learning opening the door to training specialised user simulators. | ['Milica Gašić', 'Michael Heck', 'Carel van Niekerk', 'Nurul Lubis', 'Shutong Feng', 'Christian Geishauser', 'Hsien-Chin Lin'] | 2022-08-23 | null | https://aclanthology.org/2022.sigdial-1.28 | https://aclanthology.org/2022.sigdial-1.28.pdf | sigdial-acl-2022-9 | ['task-oriented-dialogue-systems'] | ['natural-language-processing'] | [ 1.98806107e-01 6.89548373e-01 3.15285981e-01 -4.26678717e-01
-3.50276619e-01 -6.23284757e-01 9.64279950e-01 -1.15262598e-01
-4.18156445e-01 8.08202744e-01 3.29772413e-01 -3.53988826e-01
1.99721202e-01 -1.14953256e+00 -8.24513793e-01 -2.15192407e-01
2.18343809e-01 8.97336841e-01 3.69130313e-01 -1.03131926e+00
-1.27003327e-01 2.17811525e-01 -1.69475114e+00 4.57409173e-01
1.18607116e+00 3.85596663e-01 7.29044080e-01 8.46351147e-01
-4.41761523e-01 8.59805584e-01 -8.24476123e-01 -1.80748880e-01
6.63594455e-02 -6.46346331e-01 -1.20681238e+00 -2.52953544e-02
-3.69877875e-01 -3.34569514e-01 3.15770619e-02 8.03058624e-01
6.22125506e-01 2.23803550e-01 4.68049228e-01 -1.25515902e+00
-4.66547906e-01 7.78938830e-01 3.10405999e-01 -4.42132622e-01
8.18652272e-01 6.39356256e-01 9.05367494e-01 -1.50480911e-01
6.83440030e-01 1.85285807e+00 1.82529852e-01 1.02052891e+00
-1.10806859e+00 -2.55001247e-01 -2.58999579e-02 -1.62875101e-01
-9.76843596e-01 -2.79753715e-01 3.73191804e-01 -3.43334138e-01
1.20433509e+00 2.82998055e-01 6.32587612e-01 1.40349197e+00
8.60999748e-02 9.34111297e-01 1.03747475e+00 -7.18232036e-01
3.39639246e-01 6.48060739e-01 -3.00544292e-01 6.69665694e-01
-3.61488134e-01 4.00335997e-01 -9.56082419e-02 5.35418242e-02
6.66259348e-01 -2.89197236e-01 -1.49320066e-01 -3.49742055e-01
-1.00486028e+00 1.10783732e+00 3.01160723e-01 4.02589560e-01
-2.61108696e-01 4.21554875e-03 5.66672623e-01 6.79106832e-01
-4.63011675e-02 1.08668327e+00 -5.00630379e-01 -2.67843306e-01
-2.51735777e-01 4.94643033e-01 1.22363532e+00 1.05617428e+00
7.93860853e-01 1.87596440e-01 -6.17588282e-01 1.02986777e+00
3.58817816e-01 6.19767010e-01 9.57142532e-01 -6.78537250e-01
1.52942359e-01 8.38971496e-01 5.13733253e-02 -5.00635445e-01
-4.31206942e-01 -1.20299548e-01 -4.72195715e-01 1.57516420e-01
1.65457204e-01 -5.25970817e-01 -8.04762661e-01 2.03349566e+00
3.72909635e-01 3.10510341e-02 5.96722722e-01 8.14010024e-01
7.87002385e-01 8.60295594e-01 2.85139203e-01 9.09874961e-02
1.53350377e+00 -9.16391492e-01 -6.14046514e-01 -4.00013000e-01
9.18235958e-01 -4.24013585e-01 1.43368590e+00 1.16031773e-01
-9.70202684e-01 -7.35925257e-01 -7.47530878e-01 2.01590329e-01
-5.66436946e-01 -2.58922815e-01 4.53109920e-01 6.29086137e-01
-1.20612860e+00 5.42745650e-01 -4.99591440e-01 -4.98624772e-01
-2.14095131e-01 4.22276497e-01 -1.43341556e-01 1.43665329e-01
-1.98776639e+00 1.18190217e+00 8.07473779e-01 -3.84795755e-01
-9.27599549e-01 -3.71642083e-01 -1.26203108e+00 2.07544155e-02
5.47566712e-01 -9.56716895e-01 1.80708480e+00 -9.93558407e-01
-2.17546201e+00 4.56891716e-01 2.07991704e-01 -7.04780638e-01
5.36414444e-01 1.18493572e-01 -3.81871581e-01 -2.88587481e-01
6.37936741e-02 8.98185968e-01 6.32841468e-01 -1.14760578e+00
-6.96446121e-01 2.48846412e-02 5.78685343e-01 5.26889801e-01
-1.88664287e-01 -1.47349358e-01 -2.11909398e-01 -3.56012493e-01
-7.45239973e-01 -9.56731796e-01 -5.43522418e-01 -6.16156220e-01
-1.24769397e-01 -3.87683094e-01 5.63262343e-01 -4.51848835e-01
1.23337138e+00 -1.90909266e+00 3.36777121e-01 1.76237985e-01
-1.95423409e-01 6.33829296e-01 -3.97049367e-01 7.37743616e-01
1.69372007e-01 -1.34346634e-01 -2.51112640e-01 -7.25429952e-02
3.43568802e-01 6.82004392e-01 -1.06546961e-01 -4.25884455e-01
2.84057736e-01 1.07759249e+00 -1.25396609e+00 -2.45412350e-01
4.08546686e-01 2.58839309e-01 -8.87907565e-01 8.27468216e-01
-7.73219764e-01 6.16057158e-01 -4.95632589e-01 -8.30377340e-02
1.37758672e-01 9.28039942e-03 3.08835506e-01 2.35193625e-01
6.09219931e-02 4.80347216e-01 -1.22457671e+00 1.73658001e+00
-1.26275969e+00 1.33936763e-01 -2.84173310e-01 -9.46263492e-01
1.05329680e+00 4.46483940e-01 -3.57556231e-02 -1.11525154e+00
1.88560948e-01 1.39665410e-01 1.92860708e-01 -7.96645582e-01
5.47403932e-01 -4.21746045e-01 -5.01104414e-01 5.54963052e-01
3.57743710e-01 -5.00365913e-01 2.54429460e-01 2.17463404e-01
8.58116210e-01 3.52790684e-01 5.70574045e-01 -8.18652883e-02
7.84269154e-01 -9.73624587e-02 1.09975703e-01 8.20902765e-01
2.50972331e-01 2.62671232e-01 5.58155954e-01 -1.61235660e-01
-9.99210238e-01 -7.06885517e-01 2.60351479e-01 1.40576577e+00
-7.35551268e-02 -4.53937799e-01 -1.18668592e+00 -7.73928404e-01
-2.28630230e-01 1.35854578e+00 -3.23636055e-01 -6.12287343e-01
-4.77237880e-01 -2.75423080e-01 5.95676184e-01 2.47868881e-01
4.63283002e-01 -1.73172832e+00 -1.00966620e+00 6.12651825e-01
-1.21322989e-01 -1.17058277e+00 -3.51010531e-01 1.75245076e-01
-4.98536766e-01 -7.78649032e-01 -5.21640003e-01 -6.54436529e-01
3.78282130e-01 -2.03505784e-01 1.19655776e+00 2.96649765e-02
-9.83663425e-02 2.73602366e-01 -6.68586135e-01 -5.37066936e-01
-1.53372872e+00 3.14056337e-01 -9.82831642e-02 -1.54633194e-01
4.64757048e-02 -3.38616401e-01 -2.82683223e-01 3.04455400e-01
-1.26225221e+00 4.73575890e-01 5.22890568e-01 9.90910351e-01
-1.47117466e-01 -5.54376394e-02 6.99153721e-01 -1.19403446e+00
1.08798802e+00 -3.31942260e-01 -6.04192495e-01 3.00656319e-01
-5.41845858e-01 7.67055631e-01 1.09094071e+00 -2.67760843e-01
-1.23689306e+00 3.04056033e-02 -5.66433251e-01 4.63363621e-03
-5.49794018e-01 3.30328852e-01 -3.21778834e-01 3.61860335e-01
1.05564129e+00 2.79614329e-01 4.14634407e-01 -1.98686495e-01
5.55559635e-01 9.96051311e-01 1.08102463e-01 -8.30176115e-01
6.38207495e-01 -2.53757447e-01 -5.09662747e-01 -8.12175035e-01
-4.79601413e-01 -1.79070994e-01 -1.82383657e-01 -9.93459951e-03
7.01206744e-01 -7.15096712e-01 -7.06290543e-01 2.68131226e-01
-1.17044365e+00 -7.44449198e-01 -5.12508512e-01 7.62387067e-02
-7.68855274e-01 1.83267578e-01 -2.79903561e-01 -9.33182836e-01
-3.01061481e-01 -1.51294184e+00 1.25962055e+00 2.11201146e-01
-5.83356082e-01 -1.17170954e+00 1.31617129e-01 1.41277775e-01
6.12075329e-01 -7.17766676e-03 8.43160391e-01 -9.88131762e-01
-1.54982924e-01 1.07359020e-02 1.67705730e-01 4.87486124e-01
1.72955707e-01 -3.83313298e-01 -8.86916637e-01 -2.25849926e-01
-1.92823187e-01 -4.68736529e-01 1.25225320e-01 -6.49874359e-02
7.65025556e-01 -6.48290813e-01 -4.04289365e-02 7.36847073e-02
1.06367350e+00 4.86317009e-01 8.81365240e-01 3.32826018e-01
3.89135659e-01 8.23044300e-01 5.51569998e-01 2.95672476e-01
5.44195712e-01 9.45670009e-01 1.74263343e-01 -1.45752043e-01
-2.83976030e-02 -4.54658628e-01 5.43747067e-01 4.18578118e-01
1.61720335e-01 -2.94593662e-01 -8.67257595e-01 3.64094555e-01
-1.77689838e+00 -8.06704044e-01 2.43552595e-01 2.11291814e+00
1.25993752e+00 7.00563639e-02 2.67931610e-01 -6.29607067e-02
4.63887334e-01 -1.01425670e-01 -2.26440683e-01 -9.42079365e-01
2.84234285e-01 5.13104439e-01 1.98830679e-01 8.54334950e-01
-6.76402569e-01 1.30263257e+00 5.35019588e+00 6.73822403e-01
-1.14256310e+00 -4.35491875e-02 2.51311868e-01 4.81419325e-01
-4.87580836e-01 -5.29515184e-02 -7.04327404e-01 4.42140162e-01
1.27141321e+00 -2.19830006e-01 5.84255695e-01 8.24883461e-01
3.71425599e-01 1.53423563e-01 -1.06735635e+00 5.18828630e-01
-2.44255409e-01 -1.00417483e+00 2.70803928e-01 -2.71139681e-01
5.11961639e-01 -2.59096146e-01 -2.90316880e-01 9.53022599e-01
7.48906493e-01 -1.09660351e+00 5.70716321e-01 3.71669918e-01
7.07679570e-01 -7.71225750e-01 8.76440763e-01 7.84713209e-01
-6.99252963e-01 -6.89032227e-02 -1.27634242e-01 -1.17410831e-01
1.39123335e-01 -7.25057200e-02 -1.62598872e+00 6.99993253e-01
1.20608822e-01 1.95652261e-01 -4.58579838e-01 5.06167293e-01
-5.05570352e-01 2.88495749e-01 -1.27944931e-01 -5.40130019e-01
5.49945354e-01 -1.33389890e-01 5.05998552e-01 1.30620539e+00
1.75770432e-01 3.16624381e-02 3.10047358e-01 9.17322934e-01
2.64532477e-01 2.00969636e-01 -8.69874656e-01 -1.93112776e-01
2.20040411e-01 1.02277708e+00 -2.52704471e-01 -4.05587167e-01
-1.77324533e-01 1.25891507e+00 1.11773834e-01 2.46299863e-01
-7.62158751e-01 -4.27021414e-01 6.11764252e-01 1.92140937e-01
-8.32625553e-02 2.58534461e-01 1.26198903e-01 -9.69561100e-01
-4.43928659e-01 -1.43729758e+00 1.46277815e-01 -6.93362594e-01
-8.94601285e-01 8.79750907e-01 1.34503633e-01 -9.85545218e-01
-1.07469273e+00 -5.06218016e-01 -3.90660077e-01 9.61937129e-01
-1.32334661e+00 -9.80117202e-01 -1.14518739e-01 3.75550985e-01
7.77173460e-01 -4.39872622e-01 1.24736667e+00 2.85774358e-02
-1.86729625e-01 5.98844051e-01 -3.90998960e-01 -9.83715057e-02
4.77410048e-01 -1.51068890e+00 7.57569790e-01 3.69074076e-01
-6.82046488e-02 5.51192999e-01 1.07582211e+00 -4.21310812e-01
-1.15808868e+00 -1.18995118e+00 8.41016531e-01 -2.86540687e-01
3.76216173e-01 -5.11260211e-01 -9.07190681e-01 4.07366008e-01
3.77747208e-01 -6.30895197e-01 4.36108232e-01 -8.18108022e-03
4.21602391e-02 2.82801747e-01 -1.18356705e+00 9.29801643e-01
9.77428257e-01 -3.43671978e-01 -7.73172081e-01 3.07048142e-01
1.00543451e+00 -7.01093316e-01 -6.19727314e-01 2.24338189e-01
2.33450472e-01 -8.78008008e-01 8.27516437e-01 -8.26851726e-01
4.27647144e-01 -1.42143056e-01 6.67449459e-02 -1.86047244e+00
-8.59055743e-02 -8.60298574e-01 1.84600562e-01 1.05857563e+00
5.72392702e-01 -7.31474638e-01 3.19683671e-01 5.84303975e-01
-2.19499394e-01 -5.45314133e-01 -4.23186183e-01 -7.00014412e-01
-7.46310735e-03 -3.40510637e-01 1.05820155e+00 6.22536719e-01
1.67763308e-01 9.93407071e-01 -4.26960856e-01 -3.05805132e-02
-9.90117490e-02 -1.71500668e-01 1.06696773e+00 -1.07961559e+00
-7.48654366e-01 -3.59242499e-01 -1.59336716e-01 -1.00521111e+00
3.41519147e-01 -1.01554143e+00 3.31414461e-01 -1.59248376e+00
-2.80781090e-01 -3.96443546e-01 1.90805599e-01 5.40042520e-01
-2.11990550e-01 -5.73624074e-01 1.46870688e-01 -3.07717443e-01
-4.31256860e-01 7.59736955e-01 1.63845849e+00 6.24113567e-02
-4.96475577e-01 6.25768006e-02 -7.25476027e-01 4.25507039e-01
7.66746819e-01 -1.73758224e-01 -8.60961318e-01 -1.50795043e-01
4.96484414e-02 3.34629595e-01 2.49501452e-01 -8.46153498e-01
-2.58263648e-02 -3.31136763e-01 -2.24571466e-01 3.00048590e-01
4.46614511e-02 -6.95231259e-01 4.98468727e-02 6.32335007e-01
-4.33767676e-01 -1.62525058e-01 3.68626773e-01 1.94144130e-01
-1.59757718e-01 -5.41891098e-01 8.52234364e-01 -4.19491947e-01
-9.14315939e-01 -8.46863687e-02 -5.98348200e-01 1.05796628e-01
9.13593352e-01 -9.20415148e-02 1.08196363e-01 -6.33575201e-01
-5.30898213e-01 5.06888628e-01 2.62802929e-01 7.72606075e-01
4.24772918e-01 -1.08001637e+00 -6.51244104e-01 5.31842172e-01
3.27839464e-01 2.17252538e-01 1.66618839e-01 9.23479944e-02
-4.37801957e-01 5.19219279e-01 -2.84227818e-01 -4.32004660e-01
-9.09802854e-01 3.74953777e-01 5.98312497e-01 -5.19432843e-01
-4.69473332e-01 5.31654060e-01 3.31300676e-01 -1.22303486e+00
6.77037612e-02 -2.39202768e-01 -5.37562311e-01 -1.21997088e-01
5.17874241e-01 -2.03157067e-01 -5.51286973e-02 -3.21480185e-01
1.00611016e-01 -4.58330438e-02 5.13635576e-02 -3.34263176e-01
1.18845248e+00 3.94506715e-02 2.06125185e-01 2.41662100e-01
9.30504382e-01 -3.86976540e-01 -9.20368969e-01 -2.24696070e-01
2.14163903e-02 -1.02427170e-01 -3.75441134e-01 -9.00515676e-01
-5.92819452e-01 7.30173945e-01 3.98111045e-01 4.49721515e-01
8.11689258e-01 -7.60420039e-02 8.43332171e-01 5.32374203e-01
5.91784835e-01 -1.20461965e+00 2.09178358e-01 8.78158808e-01
1.05451095e+00 -1.15365577e+00 -7.33385146e-01 -1.30612910e-01
-1.26972449e+00 8.92244637e-01 8.01774085e-01 7.03460574e-02
-1.65395588e-02 2.27788091e-01 9.32330117e-02 -6.87747076e-02
-8.49052250e-01 -4.07745242e-01 2.25498229e-01 7.76289523e-01
4.75339472e-01 2.16107249e-01 -2.57874519e-01 5.56242049e-01
-6.65491104e-01 7.21244365e-02 4.27028060e-01 8.18716407e-01
-4.94467378e-01 -1.70175445e+00 -1.72537774e-01 1.25758410e-01
1.36331484e-01 -8.25378951e-03 -2.44056880e-01 7.61690199e-01
3.37618701e-02 8.27802062e-01 -2.68340349e-01 -4.07133639e-01
8.66219223e-01 3.73335123e-01 4.11749899e-01 -1.19333875e+00
-8.68055105e-01 -2.28605077e-01 4.72420633e-01 -5.19568563e-01
7.26288259e-02 -3.77538949e-01 -1.51724243e+00 -1.57800950e-02
-1.96404800e-01 4.21993315e-01 5.26854932e-01 1.10550272e+00
3.07543099e-01 9.71071601e-01 7.39193022e-01 -4.11992699e-01
-9.47839320e-01 -9.96698141e-01 -1.72398210e-01 6.70342207e-01
9.97611694e-03 -5.03385365e-01 6.46410808e-02 -2.15022042e-02] | [13.074106216430664, 8.033596992492676] |
bb722414-5cca-44a4-b0b4-aff27d722ec2 | using-classifier-features-to-determine | null | null | https://aclanthology.org/N18-4009 | https://aclanthology.org/N18-4009.pdf | Using Classifier Features to Determine Language Transfer on Morphemes | The aim of this thesis is to perform a Native Language Identification (NLI) task where we identify an English learner{'}s native language background based only on the learner{'}s English writing samples. We focus on the use of English grammatical morphemes across four proficiency levels. The outcome of the computational task is connected to a position in second language acquisition research that holds all learners acquire English grammatical morphemes in the same order, regardless of native language background. We use the NLI task as a tool to uncover cross-linguistic influence on the developmental trajectory of morphemes. We perform a cross-corpus evaluation across proficiency levels to increase the reliability and validity of the linguistic features that predict the native language background. We include native English data to determine the different morpheme patterns used by native versus non-native English speakers. Furthermore, we conduct a human NLI task to determine the type and magnitude of language transfer cues used by human raters versus the classifier. | ['ra', 'Alex Lavrentovich'] | 2018-06-01 | null | null | null | naacl-2018-6 | ['cross-corpus', 'native-language-identification'] | ['computer-vision', 'natural-language-processing'] | [ 8.76792073e-02 -1.39336765e-01 -4.07030821e-01 -3.02607834e-01
-6.16928577e-01 -8.87826204e-01 4.99355018e-01 5.11343837e-01
-8.33395302e-01 2.26769149e-01 3.96661848e-01 -9.59576190e-01
-2.04641491e-01 -5.95236778e-01 -5.45369565e-01 -5.65993786e-02
3.29266489e-01 5.91208003e-02 -1.34493142e-01 -2.68821895e-01
6.94149733e-01 3.63645911e-01 -1.56637943e+00 1.87865779e-01
1.23945606e+00 2.17001662e-01 5.75518847e-01 3.95366013e-01
-2.92140692e-01 7.64268458e-01 -5.73163629e-01 -3.65973324e-01
2.95245916e-01 -6.84117079e-01 -8.66243958e-01 -1.70940161e-01
9.11236107e-01 -1.86212838e-01 -1.43836156e-01 1.27368128e+00
1.90196559e-01 8.11671838e-02 8.20996881e-01 -4.86651093e-01
-8.49485099e-01 1.02664912e+00 -1.46502376e-01 5.10748327e-01
8.86784434e-01 6.17242232e-02 9.18990493e-01 -8.45584333e-01
7.79957294e-01 1.35240746e+00 5.22610545e-01 4.87745315e-01
-1.12454164e+00 -1.13930500e+00 2.25304857e-01 -7.23220110e-02
-1.33611846e+00 -8.34782958e-01 4.70468313e-01 -1.01631796e+00
6.20977879e-01 -2.29298547e-01 8.04832160e-01 1.05656731e+00
1.33322641e-01 4.80188966e-01 1.88309085e+00 -1.03234529e+00
-1.52594164e-01 7.14275897e-01 4.76772010e-01 7.31466711e-01
2.40678802e-01 3.78922999e-01 -1.05096948e+00 5.17530501e-01
4.56447095e-01 -7.12647080e-01 -3.31898779e-01 3.46438885e-01
-1.22788620e+00 7.63401806e-01 -1.97756633e-01 9.96311486e-01
5.57199260e-03 -5.08594215e-01 4.13285553e-01 8.56742382e-01
1.94418579e-01 7.77259231e-01 -2.66500711e-01 -5.87092102e-01
-8.33059251e-01 -1.30346075e-01 6.85710967e-01 8.24369311e-01
4.24380481e-01 2.90249676e-01 8.89216736e-02 1.24246967e+00
4.93645847e-01 5.98074794e-01 9.98258471e-01 -3.75146598e-01
6.24840021e-01 5.06489575e-01 -5.02996564e-01 -6.16040111e-01
7.50424787e-02 -1.17640711e-01 2.74426844e-02 4.65447992e-01
8.24140251e-01 -5.41789643e-02 -6.99144423e-01 1.99682784e+00
-6.49078637e-02 -7.52661586e-01 4.46284078e-02 3.18226367e-01
6.05594695e-01 7.83676982e-01 5.65269470e-01 -3.42791975e-01
1.24453413e+00 -2.70837247e-01 -3.53035390e-01 -1.50050491e-01
1.08631921e+00 -7.46832371e-01 1.60115671e+00 5.41343272e-01
-1.31600988e+00 -9.48634624e-01 -1.16014361e+00 -1.37232197e-02
-6.57294929e-01 2.17624560e-01 5.98073661e-01 1.33113253e+00
-1.03169858e+00 6.20798826e-01 -4.76164907e-01 -3.61064404e-01
-2.40623206e-01 3.19402575e-01 -7.18780220e-01 1.77858844e-01
-1.07230067e+00 1.19035423e+00 4.55691516e-01 -3.77471298e-01
-6.25455678e-01 -8.94476533e-01 -1.03861570e+00 -3.08743775e-01
-5.11238039e-01 2.63492823e-01 1.08827710e+00 -1.62857878e+00
-1.46088862e+00 1.73869944e+00 2.16570962e-02 2.83459067e-01
3.97345245e-01 8.94079953e-02 -4.84650910e-01 -1.24537975e-01
8.10632050e-01 3.36037815e-01 3.08181345e-01 -8.76268864e-01
-1.20500171e+00 -6.23497784e-01 -2.91713476e-01 4.02885288e-01
-5.76923907e-01 4.62084800e-01 5.26390821e-02 -8.10126662e-01
1.29992977e-01 -7.23386824e-01 6.63471460e-01 -5.02925813e-01
7.66945183e-02 -9.22033966e-01 -7.66073316e-02 -1.06300616e+00
1.44562566e+00 -2.25943470e+00 1.00085683e-01 4.77422446e-01
-9.98962671e-02 -1.19008929e-01 -3.59200239e-02 3.08604449e-01
-1.34836778e-01 5.39729357e-01 1.52143031e-01 -5.93534112e-02
1.63880140e-01 -4.05068606e-01 6.91212118e-02 5.32490969e-01
-2.35198811e-02 5.68785429e-01 -7.59744287e-01 -3.20365399e-01
-8.45234916e-02 5.85795529e-02 -2.90423751e-01 2.41727278e-01
3.95123452e-01 4.74038929e-01 1.05284393e-01 5.25979161e-01
-1.08722155e-03 7.38602638e-01 1.79953173e-01 5.61306238e-01
-8.37897778e-01 9.88792598e-01 -8.80897820e-01 1.44566953e+00
-7.65710115e-01 9.79249775e-01 2.93468267e-01 -6.69065952e-01
8.75808120e-01 1.94594070e-01 -2.67740041e-01 -9.23919201e-01
2.93646216e-01 8.80109191e-01 8.56271565e-01 -3.44653308e-01
3.68485868e-01 -3.51522207e-01 -1.90061569e-01 4.46893722e-01
1.13065839e-01 1.28264520e-02 5.42918265e-01 -3.16724777e-01
6.30469441e-01 2.05205083e-01 3.29102337e-01 -1.23565269e+00
4.09056574e-01 4.30858582e-02 4.98609424e-01 8.90931189e-01
-1.67544663e-01 -1.56509668e-01 1.97922915e-01 4.48374376e-02
-6.33189619e-01 -1.28608656e+00 -5.11819839e-01 1.63919914e+00
-1.96078539e-01 -2.26779711e-02 -6.99782252e-01 -2.38843217e-01
-6.35508150e-02 1.31168759e+00 -5.28746426e-01 -3.06021661e-01
-7.41065919e-01 -1.99809521e-01 4.79568571e-01 3.81464094e-01
1.00495545e-02 -1.25130260e+00 -3.06073785e-01 7.10662752e-02
1.99289978e-01 -8.29849243e-01 -4.33595270e-01 1.91442683e-01
-4.87090141e-01 -7.56274581e-01 -3.67952079e-01 -1.30589998e+00
4.30382878e-01 -2.43357912e-01 9.74932313e-01 3.15239370e-01
3.42214224e-03 6.01562917e-01 -3.95342469e-01 -6.38626635e-01
-1.12610459e+00 4.86960590e-01 3.98215860e-01 -5.73398352e-01
8.77447009e-01 -4.15883273e-01 2.69512475e-01 -2.27729410e-01
-4.89139855e-01 -1.21759012e-01 3.00694376e-01 2.93784678e-01
1.25589326e-01 1.04875252e-01 4.56532776e-01 -8.52986097e-01
8.34931493e-01 -5.16864181e-01 -5.56834877e-01 3.55782956e-01
-7.04514742e-01 -2.45354205e-01 6.30797863e-01 -7.74160743e-01
-9.41608727e-01 -2.36923158e-01 -2.89072514e-01 4.26802814e-01
-3.78740817e-01 6.71684980e-01 -3.45925659e-01 -1.39066651e-01
6.50449634e-01 2.22498313e-01 -5.70631996e-02 -4.07225013e-01
-1.96068838e-01 8.57015967e-01 6.27884746e-01 -1.10017109e+00
8.03211272e-01 -5.06464243e-01 -4.25833493e-01 -1.33236194e+00
-5.79751372e-01 -1.64065421e-01 -1.06939542e+00 -2.42666930e-01
8.46613407e-01 -1.22789598e+00 -6.28532469e-01 6.76385581e-01
-7.40370512e-01 -6.79863155e-01 -7.61899054e-02 1.00143957e+00
-3.20639819e-01 8.56998339e-02 -7.56760657e-01 -7.62079835e-01
1.24699019e-01 -1.42431951e+00 3.74501199e-01 1.12653807e-01
-6.71821773e-01 -1.17652571e+00 6.85315058e-02 2.05576882e-01
-1.09170247e-02 -1.47004694e-01 1.42036808e+00 -9.51531231e-01
-2.40708247e-01 1.29300833e-01 1.65939093e-01 2.49715686e-01
2.78718859e-01 9.90269408e-02 -6.02944553e-01 -3.93065691e-01
2.41788160e-02 -3.76622945e-01 4.33648318e-01 1.86382800e-01
2.46469304e-01 -5.57656214e-03 7.11758509e-02 6.95880055e-01
1.38645160e+00 4.47414279e-01 1.09809913e-01 5.05522847e-01
7.69621074e-01 1.03373301e+00 2.60634333e-01 -4.16184634e-01
4.82028186e-01 3.81686568e-01 -7.83426881e-01 6.05286181e-01
-2.83031464e-01 -5.81846952e-01 1.09985960e+00 1.44739079e+00
1.70132756e-01 -4.95069660e-02 -1.34358537e+00 9.46237206e-01
-7.03900874e-01 -6.40107691e-01 -1.03816964e-01 2.48623943e+00
1.19382179e+00 1.15012191e-01 2.74133474e-01 3.12373966e-01
5.82007468e-01 -2.76250929e-01 3.90392989e-02 -7.76072621e-01
-6.05659466e-03 5.65724432e-01 4.44559842e-01 9.68793929e-01
-6.55849874e-01 1.34252429e+00 6.27300882e+00 8.11778009e-01
-1.20309722e+00 -3.66200596e-01 3.57010186e-01 3.22575539e-01
-4.10180777e-01 -2.65332848e-01 -1.18415976e+00 5.46968579e-01
1.29297435e+00 -1.70104980e-01 5.26317060e-01 3.77203912e-01
1.83771234e-02 -2.51418263e-01 -1.43643308e+00 6.98736668e-01
1.43270686e-01 -5.96056700e-01 3.83353867e-02 3.48186120e-02
4.35979128e-01 -2.49205038e-01 8.30267668e-02 5.14295697e-01
3.02802771e-01 -1.24545479e+00 1.27203774e+00 1.55700326e-01
1.27837384e+00 -5.22795379e-01 8.07285160e-02 3.63649249e-01
-9.48630571e-01 -1.13621227e-01 -6.93979934e-02 -5.98269701e-01
-2.26644874e-01 -4.02039677e-01 -7.59854913e-01 -3.07534695e-01
5.67872167e-01 2.37289891e-01 -8.72437000e-01 3.69416445e-01
-4.19837326e-01 1.19257164e+00 -2.11908802e-01 -3.37741077e-01
-1.02200441e-03 -3.07819873e-01 4.77839321e-01 1.35977995e+00
4.70960647e-01 8.81272405e-02 1.50367707e-01 7.01968729e-01
1.57118589e-01 9.11005378e-01 -9.22401428e-01 -3.94726664e-01
8.71748447e-01 6.08894408e-01 -7.95439065e-01 -2.56138071e-02
-7.77491450e-01 9.93464589e-01 6.86404169e-01 1.04795136e-01
2.43459314e-01 -5.01214564e-01 6.02751374e-01 3.47504377e-01
-2.97814518e-01 -4.19138730e-01 -5.45872927e-01 -8.28793883e-01
-7.41916746e-02 -1.07439899e+00 8.96147490e-02 -1.83756649e-01
-1.38218164e+00 1.37450114e-01 6.46995455e-02 -5.07570028e-01
-3.14733207e-01 -1.31141627e+00 -5.31741619e-01 1.45482779e+00
-9.24204469e-01 -8.83504927e-01 1.25618562e-01 3.60698342e-01
5.17391801e-01 -7.87328899e-01 7.30384350e-01 5.98730668e-02
-5.48347950e-01 1.11087620e+00 8.07855465e-03 5.34678042e-01
5.80860734e-01 -1.56984425e+00 2.08643228e-01 9.75597620e-01
2.36979112e-01 1.03548241e+00 3.18098664e-01 -9.93483543e-01
-1.00018430e+00 -4.16611254e-01 1.38195956e+00 -6.40287638e-01
9.36742663e-01 -4.40519542e-01 -6.02503419e-01 8.96267772e-01
4.11234558e-01 -8.68882060e-01 9.32715416e-01 3.34566921e-01
-1.83944374e-01 1.55310988e-01 -7.97568262e-01 7.44463384e-01
1.16320789e+00 -1.01209652e+00 -8.93997014e-01 -2.75594573e-02
4.54385340e-01 -1.08489364e-01 -1.19984901e+00 2.84297705e-01
7.11062074e-01 -9.16981757e-01 4.87518281e-01 -5.15382886e-01
6.02783561e-01 7.30600730e-02 1.93094090e-02 -1.44266200e+00
-5.29117525e-01 -4.84577328e-01 8.91371727e-01 1.59574270e+00
6.19124770e-01 -5.39475799e-01 3.25833410e-01 6.64053798e-01
-1.13302603e-01 -4.17029321e-01 -9.08910453e-01 -5.65942407e-01
8.54558527e-01 -3.33859593e-01 2.84209382e-02 1.18519831e+00
2.35372409e-01 3.68097901e-01 5.65317333e-01 1.93796996e-02
1.75829321e-01 -2.28371099e-01 4.63287115e-01 -1.43751514e+00
-4.73604091e-02 -9.91774440e-01 -5.55400133e-01 -5.88169336e-01
7.46152580e-01 -1.42810988e+00 5.87718636e-02 -8.18267584e-01
-1.25664011e-01 -5.01292348e-01 -9.48702246e-02 2.35111862e-01
-3.64426762e-01 -6.49148226e-02 4.86067832e-01 1.41547490e-02
1.48990275e-02 9.59667787e-02 6.22743428e-01 1.70715153e-01
-6.22290492e-01 -4.63750632e-03 -8.37732494e-01 7.43051469e-01
9.37649786e-01 -3.49062771e-01 -4.44982409e-01 -5.71168900e-01
2.59045392e-01 -6.55610114e-02 -1.99722141e-01 -1.00055027e+00
3.90993543e-02 -2.59696335e-01 5.29212117e-01 8.45733881e-02
-4.48710680e-01 -5.47249675e-01 -3.60898465e-01 2.61316568e-01
-3.83023113e-01 6.52744889e-01 4.62994158e-01 -6.33192480e-01
-1.55565068e-01 -9.24410939e-01 9.33166564e-01 -3.21661144e-01
-6.37935162e-01 2.55414285e-03 -8.36300433e-01 4.64923143e-01
8.57486963e-01 -6.55592680e-01 2.03128666e-01 -1.53991012e-02
-3.42463553e-01 -1.16453260e-01 7.92307496e-01 6.31182611e-01
5.37416339e-02 -1.16186512e+00 -8.43195438e-01 5.20765424e-01
1.09830223e-01 -6.20336950e-01 -3.46915185e-01 4.21480596e-01
-7.01393723e-01 2.02413872e-01 -4.31538492e-01 -3.94025147e-02
-1.22181869e+00 3.42099667e-01 2.74959534e-01 5.41803986e-02
-1.47991374e-01 1.08193922e+00 2.65098959e-01 -7.14974999e-01
1.52085602e-01 1.27534032e-01 -3.48181069e-01 2.98251688e-01
4.88671452e-01 6.23491228e-01 -2.62488574e-01 -1.06464720e+00
-4.13290933e-02 6.22480094e-01 -2.63396770e-01 -5.17720997e-01
7.90460408e-01 -1.69570670e-01 -2.22079977e-01 1.27382982e+00
1.15917158e+00 1.14761245e+00 -5.13484597e-01 8.98649842e-02
2.84092695e-01 -2.36622348e-01 -2.73323268e-01 -7.83796787e-01
-2.65739858e-01 7.85440266e-01 5.93657017e-01 -5.04322723e-02
6.15833104e-01 -2.53173560e-01 8.21279734e-02 2.49988288e-02
1.34261087e-01 -1.47739971e+00 -7.35947192e-01 7.98463941e-01
7.17582941e-01 -1.13273072e+00 -1.97891802e-01 -2.13597789e-01
-3.00544530e-01 1.06268251e+00 9.21270251e-01 1.03242055e-01
6.35995269e-01 8.93004611e-03 3.35831434e-01 -2.29352787e-02
-1.52241752e-01 -1.53822795e-01 4.42098737e-01 4.69939023e-01
1.22155929e+00 2.66707957e-01 -9.72405076e-01 6.80798292e-01
-1.25083590e+00 -4.69682693e-01 3.79782468e-01 6.58379853e-01
-4.19163406e-01 -1.32833338e+00 -3.81031811e-01 5.26871264e-01
-8.12146425e-01 -4.78591442e-01 -5.73571980e-01 1.15854251e+00
4.73118067e-01 9.57704067e-01 1.61318213e-01 -3.96907032e-01
2.57986546e-01 5.75942039e-01 7.92574763e-01 -9.89414036e-01
-1.07249951e+00 -5.83448887e-01 -5.55940159e-02 1.19855493e-01
-1.10372409e-01 -1.09541857e+00 -9.83331084e-01 -2.55076945e-01
-9.08883363e-02 2.05673799e-01 4.14803177e-01 9.63934124e-01
-4.80253816e-01 1.03953183e-01 2.52311409e-01 -3.59470248e-01
-4.41253901e-01 -1.32816911e+00 -7.62382329e-01 3.93029958e-01
1.16071902e-01 -3.94476950e-01 -5.48651159e-01 -6.60342649e-02] | [10.703539848327637, 10.140847206115723] |
f824a596-b1c1-4045-8084-33fa48ac51f0 | understanding-plasticity-in-neural-networks | 2303.01486 | null | https://arxiv.org/abs/2303.01486v2 | https://arxiv.org/pdf/2303.01486v2.pdf | Understanding plasticity in neural networks | Plasticity, the ability of a neural network to quickly change its predictions in response to new information, is essential for the adaptability and robustness of deep reinforcement learning systems. Deep neural networks are known to lose plasticity over the course of training even in relatively simple learning problems, but the mechanisms driving this phenomenon are still poorly understood. This paper conducts a systematic empirical analysis into plasticity loss, with the goal of understanding the phenomenon mechanistically in order to guide the future development of targeted solutions. We find that loss of plasticity is deeply connected to changes in the curvature of the loss landscape, but that it typically occurs in the absence of saturated units or divergent gradient norms. Based on this insight, we identify a number of parameterization and optimization design choices which enable networks to better preserve plasticity over the course of training. We validate the utility of these findings in larger-scale learning problems by applying the best-performing intervention, layer normalization, to a deep RL agent trained on the Arcade Learning Environment. | ['Will Dabney', 'Razvan Pascanu', 'Bernardo Avila Pires', 'Evgenii Nikishin', 'Zeyu Zheng', 'Clare Lyle'] | 2023-03-02 | null | null | null | null | ['atari-games'] | ['playing-games'] | [ 4.88455266e-01 4.89124889e-03 -3.28594521e-02 -2.10618630e-01
-3.11939213e-02 -5.32620668e-01 4.09328997e-01 1.98893636e-01
-6.98281884e-01 7.06129372e-01 2.36998454e-01 -1.51606023e-01
-3.91607374e-01 -6.11804605e-01 -1.09068453e+00 -8.75456929e-01
-1.30179390e-01 -5.47194146e-02 2.14114860e-01 -3.85524392e-01
4.21322614e-01 6.74360633e-01 -1.17829692e+00 -5.84075339e-02
6.43444598e-01 7.44330347e-01 2.59967059e-01 2.53204674e-01
3.59727830e-01 7.42476046e-01 -3.03615957e-01 -2.24945903e-01
4.66278225e-01 -5.63858151e-01 -7.62503743e-01 -3.98125350e-01
2.50672638e-01 -1.79598451e-01 -4.81426209e-01 8.66195977e-01
6.44980669e-01 3.90760392e-01 5.42303264e-01 -7.22178936e-01
-5.86989760e-01 6.06079161e-01 -1.29065484e-01 7.20690548e-01
-3.20466787e-01 5.39673328e-01 1.15482545e+00 -4.58878726e-01
5.14576435e-01 9.97853696e-01 8.52470040e-01 7.55025864e-01
-1.54999399e+00 -4.97139752e-01 2.32440934e-01 -5.76839074e-02
-9.11094427e-01 -6.44818842e-01 5.51733553e-01 -3.85646850e-01
9.09801602e-01 -3.09064239e-01 9.52137053e-01 8.66712451e-01
6.86935246e-01 4.46078092e-01 1.11174405e+00 -1.11845843e-01
3.56204152e-01 -1.26101226e-01 -7.41745681e-02 6.72867596e-01
2.27330893e-01 4.53431010e-01 -7.00108826e-01 1.96229726e-01
8.35047841e-01 -2.18300939e-01 -4.86550450e-01 -5.47690213e-01
-8.00470591e-01 6.69118226e-01 8.48523200e-01 2.96352655e-01
-2.81059831e-01 4.56608087e-01 3.06977153e-01 6.58221900e-01
1.27211794e-01 1.13222909e+00 -5.78813016e-01 -3.01191539e-01
-6.56653345e-01 2.19588757e-01 3.38221669e-01 9.12503749e-02
7.81330466e-01 3.76740664e-01 6.56977743e-02 9.02365804e-01
1.63960889e-01 3.48682292e-02 8.91473472e-01 -9.58398998e-01
2.58381724e-01 7.76134908e-01 -2.80013889e-01 -8.42109442e-01
-5.56034088e-01 -7.88546920e-01 -4.50925112e-01 6.22604489e-01
5.62021375e-01 -3.87620896e-01 -5.05217195e-01 2.20929003e+00
3.96667607e-02 -2.96191424e-01 7.32206181e-02 8.04913223e-01
2.03331843e-01 2.85322815e-01 2.09573016e-01 3.92866367e-03
5.79492569e-01 -5.75548112e-01 -1.25155255e-01 -6.04525566e-01
7.09718168e-01 -9.04306993e-02 1.34325004e+00 7.52013549e-02
-1.11645913e+00 -2.66716719e-01 -1.26666844e+00 1.24643050e-01
-2.63871729e-01 -3.17577839e-01 5.56440890e-01 5.52970648e-01
-9.75267887e-01 1.28899789e+00 -9.08106506e-01 -5.00025988e-01
7.16694534e-01 6.81879342e-01 -2.87921339e-01 1.94623739e-01
-1.06811500e+00 1.30213177e+00 3.46173882e-01 2.10542813e-01
-9.27061617e-01 -1.02567172e+00 -3.76311988e-01 3.99914324e-01
-2.18893643e-02 -6.86254084e-01 1.09927166e+00 -1.38993990e+00
-1.67731023e+00 5.84270060e-01 2.67458349e-01 -6.39517665e-01
3.21547717e-01 -1.71703264e-01 3.30008835e-01 -2.49546930e-01
-5.09885907e-01 7.69550383e-01 6.45470500e-01 -1.03558385e+00
-2.54193693e-01 -4.88250643e-01 -1.23780360e-02 4.88948196e-01
-5.04118264e-01 -4.07574445e-01 1.50485516e-01 -6.28538132e-01
-5.03842346e-03 -1.00826192e+00 -2.15397164e-01 4.55488488e-02
1.88830927e-01 7.18861520e-02 3.00060362e-01 -3.89147043e-01
8.35570514e-01 -2.13566136e+00 3.83568585e-01 2.42973208e-01
8.37935060e-02 2.75138885e-01 -3.99453908e-01 4.07573462e-01
-7.43454546e-02 2.95762904e-02 -2.81448841e-01 8.58470872e-02
-1.16635099e-01 2.29947716e-01 -6.00919366e-01 5.23081779e-01
3.50039154e-01 1.00430071e+00 -7.34444737e-01 2.61574686e-01
-2.30503038e-01 3.33257645e-01 -7.33785510e-01 2.65685618e-02
-2.70492285e-01 4.51889038e-01 -1.89060286e-01 1.91813573e-01
3.43558118e-02 9.87880025e-03 2.47763664e-01 1.29004121e-01
-1.71196774e-01 4.38860923e-01 -6.04871094e-01 1.30294859e+00
-3.55093300e-01 9.88313437e-01 -8.77531916e-02 -1.21166003e+00
8.76461446e-01 -3.73969190e-02 4.03638005e-01 -9.65310395e-01
2.92219043e-01 1.48034275e-01 6.23423636e-01 -1.71817169e-01
2.22484723e-01 -3.23692799e-01 3.12982649e-01 5.41465521e-01
3.06319539e-02 7.39261834e-03 1.70184821e-01 -2.24630877e-01
1.30712795e+00 7.23181218e-02 -4.41986974e-03 -3.89273971e-01
1.69242434e-02 2.97106570e-04 6.45013154e-01 8.49065781e-01
-3.20395261e-01 2.91513473e-01 6.94753945e-01 -2.67590404e-01
-1.12595761e+00 -9.66700673e-01 -9.36916843e-02 1.37681794e+00
-9.11114588e-02 3.68144810e-02 -5.97159386e-01 -2.42898077e-01
1.18682630e-01 4.67935890e-01 -8.41274679e-01 -9.63635087e-01
-7.03931570e-01 -9.83410239e-01 6.73133016e-01 4.71891254e-01
5.18720865e-01 -1.38207817e+00 -8.85511696e-01 2.54997611e-01
5.42271137e-01 -5.75662315e-01 -2.09387496e-01 5.35174072e-01
-1.21250069e+00 -1.09663844e+00 -5.44757724e-01 -7.30763137e-01
8.17641497e-01 8.09311271e-02 7.21879601e-01 3.52298200e-01
-1.27136633e-01 3.72764856e-01 -6.21939311e-03 -1.86925903e-01
-2.91283756e-01 3.89161080e-01 1.78619698e-01 -4.05390412e-01
-1.92529425e-01 -8.72258484e-01 -6.59938991e-01 2.90837139e-01
-9.67612982e-01 -9.55152437e-02 6.91168129e-01 9.88381982e-01
4.81305629e-01 2.12027624e-01 7.64847875e-01 -7.40165114e-01
1.07918632e+00 -4.77977872e-01 -5.23611486e-01 7.36055970e-02
-1.05519688e+00 4.53360885e-01 8.38387668e-01 -6.26475334e-01
-9.08998251e-01 -1.07217230e-01 -1.34537265e-01 7.97769278e-02
2.79327422e-01 7.42426157e-01 1.67722479e-02 -5.42712450e-01
8.05028260e-01 3.00140798e-01 7.56268278e-02 -3.36072892e-01
1.69228718e-01 -1.03586487e-01 2.34284088e-01 -6.21264756e-01
5.44583619e-01 8.31773579e-02 3.08761280e-02 -6.29278004e-01
-6.45445287e-01 2.07535297e-01 -4.03709024e-01 -3.11615020e-01
2.50018984e-01 -6.39021218e-01 -6.61186218e-01 4.84318525e-01
-4.23202604e-01 -1.22177410e+00 -4.85103875e-01 2.83345163e-01
-6.46844447e-01 -2.51364768e-01 -4.50549722e-01 -1.94044024e-01
-3.32068741e-01 -9.58242118e-01 3.42814364e-02 4.53149289e-01
-1.77555889e-01 -1.06674719e+00 6.11759007e-01 5.97772412e-02
5.65136671e-01 -1.07299149e-01 1.40196955e+00 -4.44382548e-01
-2.90531039e-01 2.45926753e-01 7.37173334e-02 3.90810102e-01
1.62881054e-02 1.09248944e-01 -6.13277078e-01 -3.81151378e-01
-1.59021318e-01 -4.53339368e-01 1.30157220e+00 3.18716466e-01
1.06593108e+00 -3.70456815e-01 -6.23529889e-02 7.76883125e-01
1.29735625e+00 5.01361117e-02 5.41450858e-01 7.17929244e-01
3.14658701e-01 5.18163621e-01 -2.14784313e-03 5.92921413e-02
5.73900379e-02 4.46175128e-01 4.37013716e-01 2.49717519e-01
-1.72595993e-01 -3.07050943e-01 5.85147202e-01 3.48238289e-01
1.02124792e-02 1.43582061e-01 -9.29891884e-01 4.65038270e-01
-1.69939351e+00 -1.08667183e+00 7.16996312e-01 2.22240043e+00
1.08867407e+00 4.94480252e-01 1.39606297e-01 6.59701973e-02
3.33991468e-01 9.47870985e-02 -1.15858316e+00 -6.46931350e-01
-2.25803390e-01 1.30328029e-01 5.61225951e-01 2.99515158e-01
-4.89526540e-01 1.11387575e+00 6.93811464e+00 3.10904741e-01
-1.71253026e+00 -3.36222649e-01 6.30597413e-01 -3.70824069e-01
-1.41907111e-01 -1.59854576e-01 -5.57317734e-01 2.12195814e-01
1.09397328e+00 -1.68889478e-01 1.02155530e+00 6.85692310e-01
3.48863125e-01 -1.48854688e-01 -9.65383053e-01 2.34529108e-01
-3.24911743e-01 -1.46695912e+00 -1.86615109e-01 -1.56089142e-01
7.97995567e-01 4.15018111e-01 6.18059039e-01 3.25242311e-01
5.15467048e-01 -1.11000919e+00 6.61400318e-01 3.98373067e-01
2.61070132e-01 -7.36419678e-01 3.05106848e-01 2.97674477e-01
-4.70560551e-01 -5.42991281e-01 -3.12861830e-01 -2.91140676e-01
-2.96548665e-01 1.97540045e-01 -9.04626489e-01 -5.94530880e-01
5.51673710e-01 6.10575795e-01 -8.37030530e-01 1.21742845e+00
-2.82518744e-01 7.51591563e-01 -9.56790969e-02 -5.45591600e-02
5.16133271e-02 -1.16586924e-01 3.64332527e-01 7.82518566e-01
1.60479367e-01 -1.11601790e-02 -6.07879579e-01 9.00713146e-01
-4.85578507e-01 -1.27145678e-01 -4.41825718e-01 -3.56754392e-01
4.71591413e-01 8.76972377e-01 -6.81161761e-01 3.07054818e-01
-3.98979411e-02 6.67092383e-01 9.47224617e-01 4.65125412e-01
-4.71142828e-01 -8.71421546e-02 8.52666855e-01 1.72783509e-01
3.41523528e-01 -4.29791659e-01 -5.79359829e-01 -6.56996667e-01
-1.13382891e-01 -7.87728369e-01 5.64938188e-02 -5.24542928e-01
-1.03697014e+00 2.93004006e-01 -4.42532539e-01 -3.93396825e-01
-5.31384908e-02 -5.84522307e-01 -8.51047218e-01 5.17048836e-01
-1.38750339e+00 -5.14921308e-01 1.39346078e-01 4.73018140e-01
2.46629238e-01 -1.65533245e-01 6.28786743e-01 3.50758582e-02
-1.07266700e+00 6.62132800e-01 6.32028461e-01 -8.16567838e-02
5.54785728e-01 -1.07481277e+00 2.96345651e-01 7.14498281e-01
8.53674933e-02 9.04759943e-01 8.51029396e-01 -4.31894273e-01
-1.39448071e+00 -9.91001189e-01 2.71004230e-01 -1.28344163e-01
6.49249911e-01 -2.27956355e-01 -1.01412582e+00 5.85250258e-01
-1.72629595e-01 -7.78642818e-02 7.16472208e-01 3.02559614e-01
-4.20881599e-01 -4.02648389e-01 -1.06289124e+00 8.97756696e-01
9.86101627e-01 -5.17479300e-01 -4.58312899e-01 -1.68930098e-01
4.28319633e-01 -1.60179541e-01 -7.63944507e-01 3.28490108e-01
7.39729881e-01 -8.85327876e-01 7.73141265e-01 -8.19292605e-01
4.19672459e-01 -6.39876276e-02 5.84828071e-02 -1.84054315e+00
-4.85674918e-01 -4.64461803e-01 1.97792165e-02 8.73999059e-01
6.92135870e-01 -6.51880920e-01 8.80223751e-01 6.95427239e-01
-1.08307056e-01 -1.06377137e+00 -9.28472579e-01 -5.07697284e-01
6.31030858e-01 5.10595478e-02 1.53934807e-01 6.63753033e-01
1.42892003e-01 1.62174374e-01 7.82035291e-02 -1.46413133e-01
5.29548675e-02 -2.66849160e-01 4.26199704e-01 -9.92971420e-01
-9.43766013e-02 -9.16831017e-01 -2.51177460e-01 -7.54835486e-01
2.91265547e-01 -7.44332492e-01 2.72131473e-01 -1.18414664e+00
2.35782638e-02 -5.67116261e-01 -5.32238483e-01 6.31048977e-01
-1.57417729e-01 1.29399225e-01 3.18074495e-01 2.96457350e-01
-3.84666085e-01 8.64082098e-01 1.18912935e+00 1.15800329e-01
-5.96409976e-01 1.21192761e-01 -9.91334379e-01 6.34910047e-01
1.28362668e+00 -5.60807526e-01 -5.04813850e-01 -5.02430260e-01
6.13155603e-01 -4.85217035e-01 2.55652964e-01 -1.22707200e+00
2.81435065e-02 -3.61085594e-01 5.45279443e-01 4.58228230e-01
9.49928686e-02 -6.63251758e-01 -1.58353508e-01 7.25952506e-01
-8.69401872e-01 2.98126847e-01 4.84638870e-01 5.07251680e-01
3.00464988e-01 -2.74499387e-01 1.08592534e+00 8.62652734e-02
-7.03565836e-01 2.72267938e-01 -6.75162554e-01 3.38385850e-01
7.11645246e-01 -2.06684917e-01 -3.76828730e-01 -8.94183144e-02
-4.57003087e-01 1.48990192e-02 6.19631350e-01 4.26648617e-01
4.21032578e-01 -9.93980944e-01 -3.24728906e-01 1.51046008e-01
-2.98048526e-01 -3.68869662e-01 8.32737535e-02 7.54973888e-01
-4.68062967e-01 2.23489717e-01 -6.86505258e-01 6.70803897e-03
-8.78553391e-01 1.66352898e-01 9.14483070e-01 -1.45321280e-01
-4.81598288e-01 8.52543294e-01 7.55168917e-03 -3.02644759e-01
3.47585052e-01 -3.14867318e-01 -2.32080087e-01 -1.64873913e-01
2.60095656e-01 1.67170450e-01 6.11873008e-02 -2.22705722e-01
-1.00641340e-01 3.19670588e-01 -2.95085222e-01 -2.40131021e-02
1.60259914e+00 -1.12742469e-01 9.38156098e-02 3.22072685e-01
8.65692914e-01 -3.35323453e-01 -1.99710858e+00 -1.99804902e-01
-2.07711887e-02 -1.42249152e-01 1.93746418e-01 -8.54828417e-01
-1.39572573e+00 8.19509268e-01 6.54291272e-01 -1.52646583e-02
9.85261798e-01 -2.75897026e-01 4.51104611e-01 7.34313607e-01
-5.03276810e-02 -1.26433766e+00 3.88334870e-01 7.77249277e-01
9.34680045e-01 -9.51313496e-01 2.97625531e-02 5.78017056e-01
-5.78822136e-01 1.09797192e+00 7.89915323e-01 -3.46548051e-01
4.11370248e-01 -4.74920869e-02 -2.18547449e-01 -6.02623150e-02
-1.02242446e+00 2.15754863e-02 -2.96050422e-02 5.37526369e-01
3.79365623e-01 -2.99321383e-01 -1.03786163e-01 3.24269652e-01
-4.01877046e-01 -1.79855078e-01 4.71840292e-01 7.71629095e-01
-8.43044400e-01 -9.52322960e-01 1.44093141e-01 5.10460973e-01
-3.88657093e-01 -1.33998767e-01 -4.82902348e-01 6.70312524e-01
-1.89012140e-02 4.30269569e-01 -1.04983645e-02 -3.35992306e-01
2.65963376e-01 6.13828339e-02 7.11828113e-01 -5.10328770e-01
-7.68686473e-01 -6.29506826e-01 -2.51375556e-01 -4.03101295e-01
-3.78075838e-02 -7.84946084e-01 -1.41512048e+00 -5.12239277e-01
-1.48162708e-01 -2.29578495e-01 5.89134634e-01 1.04643154e+00
3.75021189e-01 6.25312448e-01 5.24262488e-01 -5.01485884e-01
-6.59915030e-01 -6.73776031e-01 -4.05993372e-01 2.95200109e-01
4.24817950e-01 -6.98731720e-01 -3.61724168e-01 -1.17604129e-01] | [8.42251205444336, 3.2987732887268066] |
c540d395-49fd-4230-80a9-9df49b47e8e0 | jet-tagging-in-the-lund-plane-with-graph | 2012.08526 | null | https://arxiv.org/abs/2012.08526v2 | https://arxiv.org/pdf/2012.08526v2.pdf | Jet tagging in the Lund plane with graph networks | The identification of boosted heavy particles such as top quarks or vector bosons is one of the key problems arising in experimental studies at the Large Hadron Collider. In this article, we introduce LundNet, a novel jet tagging method which relies on graph neural networks and an efficient description of the radiation patterns within a jet to optimally disentangle signatures of boosted objects from background events. We apply this framework to a number of different benchmarks, showing significantly improved performance for top tagging compared to existing state-of-the-art algorithms. We study the robustness of the LundNet taggers to non-perturbative and detector effects, and show how kinematic cuts in the Lund plane can mitigate overfitting of the neural network to model-dependent contributions. Finally, we consider the computational complexity of this method and its scaling as a function of kinematic Lund plane cuts, showing an order of magnitude improvement in speed over previous graph-based taggers. | ['Huilin Qu', 'Frédéric A. Dreyer'] | 2020-12-15 | null | null | null | null | ['jet-tagging'] | ['graphs'] | [-2.40073711e-01 -2.75903032e-03 -2.02545121e-01 -4.21202004e-01
-4.61284339e-01 -5.80715060e-01 8.68336976e-01 3.32111835e-01
-3.52534145e-01 5.96958756e-01 5.70294112e-02 -7.04861104e-01
-2.45185927e-01 -8.57553780e-01 -6.24797583e-01 -8.50147545e-01
-4.35596049e-01 1.30264175e+00 7.48574078e-01 -1.85819596e-01
-1.28695011e-01 8.00901294e-01 -9.40456390e-01 3.61987680e-01
-2.64375359e-01 1.11000252e+00 -5.10847926e-01 6.72433496e-01
1.55016080e-01 5.65342724e-01 -3.20188731e-01 -6.30780041e-01
4.75771993e-01 -4.60372835e-01 -5.29524326e-01 -8.24900627e-01
3.26174974e-01 2.21882001e-01 -6.64917231e-01 1.44055855e+00
7.62397468e-01 1.47035405e-01 4.84809399e-01 -1.12259543e+00
-1.63409740e-01 1.00107169e+00 -1.00822054e-01 8.40950608e-01
-6.43160880e-01 -6.73402846e-02 1.06677067e+00 -7.34148145e-01
8.47657084e-01 1.32535565e+00 8.71475220e-01 2.89156139e-02
-1.45522344e+00 -8.93651009e-01 -6.75431192e-01 1.36153668e-01
-9.69951868e-01 -3.29862565e-01 6.59149706e-01 -5.75473309e-01
1.20703030e+00 7.21324608e-02 6.53288901e-01 7.27903426e-01
8.48348618e-01 -3.18820700e-02 8.83230329e-01 -7.25629628e-01
2.19634488e-01 1.29046785e-02 4.56967503e-01 6.90180361e-01
5.65572262e-01 8.56117666e-01 -5.02988696e-01 -3.56961846e-01
7.62576580e-01 -5.61485112e-01 4.56784964e-01 -1.32028782e+00
-1.22209716e+00 1.32336736e+00 5.53565025e-01 2.61466056e-01
2.19407901e-01 5.04205048e-01 9.49588656e-01 -4.06043194e-02
7.28865206e-01 7.87129223e-01 -6.91121101e-01 2.95388550e-01
-6.89154506e-01 4.36412185e-01 6.20832920e-01 7.26292551e-01
1.16315879e-01 3.45937163e-01 -3.28038096e-01 4.90927517e-01
3.20331752e-01 4.58605409e-01 -7.16789365e-02 -4.97228235e-01
3.00229102e-01 1.80530444e-01 -6.77091777e-02 -7.16849208e-01
-1.21437168e+00 -9.66659129e-01 -7.84502387e-01 4.30782050e-01
5.21225929e-01 -3.31507146e-01 -9.13365424e-01 1.33206677e+00
6.29730225e-01 -4.40921068e-01 -1.68691367e-01 9.65820014e-01
1.00520563e+00 3.60863715e-01 2.01472536e-01 -6.91956282e-02
1.36440313e+00 -1.00636637e+00 -2.98994571e-01 -2.14499816e-01
6.41572297e-01 -6.67187870e-01 -2.82079309e-01 2.35075653e-01
-1.05381477e+00 -5.28556585e-01 -1.08312774e+00 1.19768210e-01
-3.85081857e-01 -8.39252695e-02 1.36102331e+00 7.39912093e-01
-7.70237088e-01 1.10535085e+00 -9.25735176e-01 -2.56000042e-01
-2.19952673e-01 7.95875967e-01 -1.65040731e-01 4.91695106e-01
-1.34190083e+00 9.03571486e-01 1.00364304e+00 -3.44122872e-02
-4.41122353e-01 -7.65299797e-01 -6.73727691e-01 1.49087980e-01
1.54719681e-01 -4.03698236e-01 1.70378470e+00 -2.28462011e-01
-9.84771669e-01 7.37971663e-01 4.89844799e-01 -7.45208085e-01
1.33270890e-01 3.39207411e-01 -9.76281404e-01 -1.98519416e-02
-6.81785718e-02 1.78817406e-01 2.28043452e-01 -6.58813357e-01
-3.91325295e-01 -1.11427002e-01 -2.60552734e-01 -3.50801259e-01
3.43345970e-01 8.78964782e-01 -2.59618759e-01 -4.53519523e-01
3.61655205e-01 -9.55662489e-01 -2.91717052e-01 -6.46294177e-01
-4.42713171e-01 -3.34481597e-02 2.01843262e-01 -5.41633785e-01
5.14491141e-01 -1.82502174e+00 -1.99798960e-02 3.85303795e-01
4.57375497e-01 2.93867737e-01 -2.84503750e-03 4.56121534e-01
-4.70578343e-01 -2.00298697e-01 5.71125329e-01 3.45950216e-01
4.26245034e-01 -8.83226022e-02 1.89452559e-01 6.36344969e-01
-8.62938315e-02 1.01634288e+00 -8.39392602e-01 -5.78026958e-02
1.45537838e-01 2.26584181e-01 -5.32836497e-01 -2.66873926e-01
-2.39307627e-01 3.84882718e-01 -3.24438065e-01 2.43425697e-01
7.93883145e-01 9.42828655e-02 4.90131915e-01 -4.21006292e-01
-3.34005296e-01 7.26454854e-01 -9.06339943e-01 9.81531858e-01
6.44982904e-02 4.98425364e-01 2.42758945e-01 -1.19055438e+00
9.04433310e-01 1.00357741e-01 4.41458374e-01 -1.20985281e+00
4.18291390e-01 8.39427635e-02 4.77674365e-01 2.80851781e-01
7.14816213e-01 -6.74379647e-01 -3.02424818e-01 3.01658392e-01
5.58694184e-01 3.44931424e-01 2.84803629e-01 2.99625576e-01
1.24957943e+00 -2.22708389e-01 2.63418704e-01 -5.99834442e-01
-2.17529312e-01 2.54973531e-01 6.14728928e-01 1.14166975e+00
-1.88805591e-02 1.43585473e-01 8.03065956e-01 -9.63969290e-01
-1.23993707e+00 -9.59437311e-01 -3.84791166e-01 1.28388286e+00
4.94594034e-03 -5.57319939e-01 -2.17678517e-01 -7.86890090e-01
2.23571166e-01 8.97670209e-01 -3.25675160e-01 -3.75740975e-01
-7.47513890e-01 -1.49413669e+00 4.24964011e-01 7.66846657e-01
-3.79171185e-02 -1.09260297e+00 -1.71609521e-01 2.42611676e-01
5.84121764e-01 -1.22170436e+00 2.76182562e-01 1.21889675e+00
-5.53258896e-01 -9.94044363e-01 1.43411338e-01 -4.24906582e-01
3.36225837e-01 -9.06843171e-02 1.61951041e+00 -1.06148392e-01
-3.53973657e-01 -2.85148829e-01 -2.34800547e-01 -5.21047831e-01
-9.72864866e-01 -2.21688986e-01 3.67984265e-01 -7.21660018e-01
5.89180470e-01 -1.87643878e-02 -1.01538360e-01 3.07081729e-01
-2.83810019e-01 -4.28079218e-02 4.97544229e-01 9.17380989e-01
3.30650330e-01 5.33681214e-01 5.88960387e-03 -7.79267311e-01
-5.31117106e-03 -3.95940781e-01 -1.43450201e+00 -2.60877699e-01
-3.83587480e-01 4.48906869e-01 4.64928567e-01 -3.40530016e-02
-6.87891245e-01 -8.75137001e-03 -8.39153454e-02 -2.64766160e-02
1.65730283e-01 1.09356619e-01 2.39129495e-02 -8.25495481e-01
4.76071686e-01 -5.96867800e-01 -5.88550687e-01 -8.42258990e-01
2.98852831e-01 -3.35669853e-02 6.37884974e-01 -4.84428823e-01
9.67776477e-01 -1.34682328e-01 7.23078668e-01 -1.76532120e-01
-9.81833637e-01 -6.59012973e-01 -9.00644779e-01 -9.50399712e-02
9.75844145e-01 -6.87055171e-01 -6.57264531e-01 5.11647686e-02
-1.16593063e+00 -1.08068779e-01 -1.86513722e-01 8.55738699e-01
-2.61215359e-01 -3.08588725e-02 -7.17647672e-01 -5.03771544e-01
-2.08649412e-01 -9.69773769e-01 7.92459130e-01 -4.00968194e-02
3.96299243e-01 -1.04048109e+00 2.30081826e-01 7.82220140e-02
2.40364566e-01 1.73480645e-01 1.44950843e+00 -1.45883965e+00
-4.79976445e-01 -3.66152436e-01 -6.71751499e-01 1.08225979e-02
-7.84956515e-01 -5.63128531e-01 -5.07067978e-01 -5.80009758e-01
-1.42828375e-01 -2.09772121e-02 1.18931723e+00 4.36438799e-01
5.94458342e-01 1.95998102e-02 -6.80793881e-01 6.10636652e-01
1.37102008e+00 1.89993903e-01 1.88463852e-02 4.61138219e-01
6.75534189e-01 1.24842145e-01 3.67763132e-01 4.92280312e-02
-3.88615429e-01 1.26378536e+00 5.11221290e-01 -1.78077400e-01
-2.67834753e-01 2.16139667e-02 1.75228402e-01 9.70822632e-01
-1.46320403e-01 -2.04155073e-01 -1.05461133e+00 1.07948504e-01
-1.67991221e+00 -8.21525931e-01 -7.07160830e-01 2.18256783e+00
3.05722922e-01 7.76335716e-01 2.24908248e-01 -2.53032625e-01
8.93293500e-01 -4.87561431e-03 -3.70507687e-02 -8.97723317e-01
3.26099366e-01 7.13980317e-01 1.28908193e+00 2.17611358e-01
-1.35968053e+00 7.70470858e-01 7.47621489e+00 7.58510828e-01
-7.46153951e-01 5.99193633e-01 1.45941094e-01 -4.16715950e-01
9.95445848e-02 3.12286913e-02 -1.13953030e+00 1.87636882e-01
1.12146306e+00 4.51967753e-02 3.83983552e-01 8.07935417e-01
-2.61569619e-01 2.01524422e-01 -1.15177417e+00 4.84913230e-01
-6.44455180e-02 -1.53785443e+00 -5.36081970e-01 7.18310699e-02
5.75798452e-01 8.28015029e-01 -6.82194531e-01 7.78036594e-01
8.67527783e-01 -6.76234841e-01 8.84211481e-01 3.22649162e-03
4.45436984e-01 -9.13437724e-01 1.10685647e+00 -9.13959346e-04
-1.05392039e+00 1.43324092e-01 -9.57597375e-01 -1.04310967e-01
2.03101039e-01 9.67882276e-01 -1.27710056e+00 7.47025073e-01
6.25408530e-01 -7.18736351e-02 -7.70551741e-01 1.35150850e+00
-5.42332530e-02 9.21192646e-01 -3.44142139e-01 -1.86784089e-01
2.77604252e-01 -8.25812146e-02 7.26156235e-01 1.49267089e+00
3.95096362e-01 -2.72726238e-01 4.05810416e-01 6.98038280e-01
-3.82611491e-02 -1.56572565e-01 -5.91369450e-01 -2.45438263e-01
-7.45854676e-02 1.50059748e+00 -1.28267813e+00 -5.88453077e-02
-3.64972949e-01 -2.45239623e-02 2.49892056e-01 1.11824721e-01
-7.17585504e-01 -2.02708095e-01 2.93400198e-01 -8.86137504e-03
8.27554703e-01 -3.91342968e-01 6.64530620e-02 -6.78672612e-01
-2.96114802e-01 -4.47579533e-01 4.83608633e-01 -7.08021998e-01
-1.08647895e+00 2.64337301e-01 2.50691503e-01 -8.14405739e-01
-2.70510018e-01 -1.20729268e+00 -4.82177019e-01 6.94469333e-01
-9.95464027e-01 -1.12056661e+00 4.43165123e-01 -2.31974751e-01
-9.95017365e-02 -6.37410462e-01 6.61576033e-01 2.98972160e-01
-3.34424108e-01 2.82649219e-01 7.14830756e-01 2.82386750e-01
5.08973956e-01 -1.42987669e+00 1.13372934e+00 8.12494993e-01
4.32117999e-01 7.17353597e-02 1.04227507e+00 -1.01289403e+00
-1.57446837e+00 -1.13111293e+00 9.43157196e-01 -2.78659999e-01
1.23165250e+00 -1.06851840e+00 -3.98251772e-01 7.79073298e-01
8.74135047e-02 7.19280005e-01 1.78719059e-01 5.55280924e-01
-2.32063755e-01 5.42168580e-02 -8.41338813e-01 4.47951034e-02
1.06445384e+00 -1.80303872e-01 -3.23886961e-01 9.04284596e-01
6.68877184e-01 -5.22034585e-01 -7.17136383e-01 6.86103880e-01
3.56984526e-01 -1.09561110e+00 1.22246671e+00 -1.13648796e+00
-2.04292089e-01 2.84695886e-02 1.53612748e-01 -1.43389261e+00
-1.09649420e+00 -4.99258846e-01 2.22959369e-01 9.61121917e-01
1.46071926e-01 -4.13125187e-01 6.80465460e-01 -1.34429798e-01
-1.30890027e-01 -1.72803909e-01 -1.22106969e+00 -1.14110434e+00
3.80529404e-01 -5.92466652e-01 3.42864901e-01 6.96620226e-01
-6.71282187e-02 5.31258464e-01 -3.91440749e-01 4.02403891e-01
7.26255894e-01 6.07594065e-02 1.48444548e-01 -1.63536859e+00
-6.35897040e-01 -2.24909529e-01 -9.72719133e-01 -1.98054284e-01
1.81916598e-02 -1.68131101e+00 1.36934668e-01 -1.16020918e+00
3.66275132e-01 -5.27123928e-01 -7.61050761e-01 3.68586421e-01
5.52123725e-01 6.88701570e-01 9.34062228e-02 -3.28557640e-01
-9.29360867e-01 -3.63579355e-02 7.87927210e-01 -3.07039022e-01
4.63961899e-01 -7.04232976e-02 -1.49565101e-01 9.11517262e-01
7.07844138e-01 -1.05814505e+00 3.57687235e-01 -1.69989750e-01
8.06014061e-01 3.62743467e-01 7.93305337e-01 -1.32027936e+00
6.62215203e-02 2.21037209e-01 9.26473200e-01 -7.77217984e-01
1.73852876e-01 -3.34771633e-01 1.75877720e-01 3.39633912e-01
-3.23241651e-01 1.49739236e-01 5.36101103e-01 3.55243295e-01
6.99202567e-02 -7.20852554e-01 7.93185949e-01 -1.85917348e-01
-4.96277004e-01 -2.50623617e-02 -3.56792063e-01 9.45037156e-02
7.39387095e-01 6.99175179e-01 -4.61752594e-01 1.99341163e-01
-6.14969194e-01 -3.48357819e-02 1.62869394e-02 3.86279881e-01
-2.85789460e-01 -1.74004388e+00 -3.85594845e-01 3.29825580e-01
-5.97477779e-02 -7.96755016e-01 -3.10906261e-01 8.25251341e-01
-5.22491455e-01 1.08025181e+00 -2.52069175e-01 -3.56779158e-01
-1.39285791e+00 8.88501585e-01 4.19421315e-01 -7.15308785e-01
-7.51378298e-01 8.12414587e-01 6.14346899e-02 -7.01261342e-01
9.61849838e-03 -2.07922682e-01 5.10459058e-02 -2.15019554e-01
2.16968998e-01 2.88176477e-01 7.89373100e-01 -6.67474508e-01
-5.43109298e-01 1.07376546e-01 3.56209911e-02 2.57575780e-01
1.19743335e+00 6.49695456e-01 -2.99755007e-01 9.28315669e-02
1.04665256e+00 1.73188284e-01 -4.39441472e-01 -2.34297335e-01
2.42400840e-01 1.79574881e-02 6.15631163e-01 -1.14498639e+00
-1.10456085e+00 7.62846291e-01 7.31980801e-01 7.08374262e-01
2.68643379e-01 4.24247473e-01 6.53587520e-01 6.42580092e-01
5.09621084e-01 -8.73924136e-01 -5.39602458e-01 6.44314945e-01
5.45093298e-01 -9.69516516e-01 2.17201158e-01 -3.14121336e-01
1.01861591e-02 1.30249310e+00 2.45470196e-01 6.58544376e-02
3.58410388e-01 5.46503663e-01 -1.20063677e-01 -6.92309260e-01
-6.20262206e-01 -7.53959194e-02 6.86047614e-01 3.38138908e-01
2.26942644e-01 2.68971115e-01 -3.15602958e-01 6.24750972e-01
-3.38940501e-01 -4.10369217e-01 2.18027875e-01 6.18233263e-01
-5.58810830e-01 -1.38001621e+00 -4.98643756e-01 6.53696239e-01
-6.02555215e-01 -2.93335736e-01 -2.61351943e-01 1.06416035e+00
3.83985519e-01 5.07078767e-01 6.63884506e-02 -2.02440560e-01
1.73791528e-01 4.73096669e-01 7.03427255e-01 -6.89886332e-01
-7.96214461e-01 1.14583634e-01 7.00215578e-01 -5.05184174e-01
-2.11588562e-01 -6.19354010e-01 -1.12801945e+00 4.31991965e-02
-7.05397010e-01 5.61817765e-01 9.28820193e-01 1.09161997e+00
-9.29082260e-02 8.81519556e-01 1.66384071e-01 -1.14178741e+00
-7.25822985e-01 -6.43458784e-01 -8.78769994e-01 -1.74324214e-02
-5.72997592e-02 -1.19027996e+00 -2.56568670e-01 -8.32277954e-01] | [15.698441505432129, 2.919255018234253] |
ea6274f4-47d6-43a5-8a15-3d451e6d69d3 | guide-the-learner-controlling-product-of | 2302.02852 | null | https://arxiv.org/abs/2302.02852v1 | https://arxiv.org/pdf/2302.02852v1.pdf | Guide the Learner: Controlling Product of Experts Debiasing Method Based on Token Attribution Similarities | Several proposals have been put forward in recent years for improving out-of-distribution (OOD) performance through mitigating dataset biases. A popular workaround is to train a robust model by re-weighting training examples based on a secondary biased model. Here, the underlying assumption is that the biased model resorts to shortcut features. Hence, those training examples that are correctly predicted by the biased model are flagged as being biased and are down-weighted during the training of the main model. However, assessing the importance of an instance merely based on the predictions of the biased model may be too naive. It is possible that the prediction of the main model can be derived from another decision-making process that is distinct from the behavior of the biased model. To circumvent this, we introduce a fine-tuning strategy that incorporates the similarity between the main and biased model attribution scores in a Product of Experts (PoE) loss function to further improve OOD performance. With experiments conducted on natural language inference and fact verification benchmarks, we show that our method improves OOD results while maintaining in-distribution (ID) performance. | ['Mohammad Taher Pilehvar', 'Hossein Amirkhani', 'Ali Modarressi'] | 2023-02-06 | null | null | null | null | ['fact-verification'] | ['natural-language-processing'] | [ 3.87570083e-01 5.68633854e-01 -6.14468038e-01 -8.81637096e-01
-6.37630761e-01 -5.58590174e-01 7.60984421e-01 3.64680529e-01
-1.38058946e-01 5.08174479e-01 2.71179616e-01 -3.70551229e-01
5.12797236e-02 -8.27211916e-01 -8.27587903e-01 -4.71818596e-01
4.52140033e-01 5.30184925e-01 2.52799302e-01 9.39579234e-02
5.12363851e-01 2.82266289e-01 -1.44519830e+00 5.82376838e-01
7.79966593e-01 1.14351809e+00 -2.30892807e-01 2.30019242e-01
-1.32369906e-01 1.11736774e+00 -7.91840434e-01 -9.47978675e-01
4.03838098e-01 -3.47687364e-01 -6.50798202e-01 8.94712433e-02
7.11317837e-01 -1.93954557e-01 -1.74355451e-02 1.19004786e+00
2.16279015e-01 -7.19422773e-02 9.67966318e-01 -1.49616599e+00
-5.69230139e-01 7.29057372e-01 -5.69879949e-01 1.76165506e-01
4.23757359e-02 9.53265056e-02 1.33762646e+00 -8.80648613e-01
6.76144719e-01 1.21543133e+00 5.77051461e-01 4.14792567e-01
-1.56936800e+00 -8.71876359e-01 5.14712274e-01 -1.61218181e-01
-1.17105985e+00 -4.28648442e-01 6.03970945e-01 -6.06595039e-01
5.94270587e-01 2.20383424e-03 6.24946393e-02 1.07246470e+00
1.69294283e-01 7.27327049e-01 1.18597817e+00 -3.38542879e-01
3.26141119e-01 8.19071054e-01 3.90340030e-01 4.91093963e-01
5.05090237e-01 1.73951253e-01 -6.74353838e-01 -4.97561783e-01
1.45051479e-01 -2.78041691e-01 -1.58999056e-01 -7.47718275e-01
-7.54232883e-01 9.80922639e-01 4.93442237e-01 1.65338181e-02
-5.40143073e-01 -8.36387500e-02 2.53409773e-01 2.68315613e-01
6.54636085e-01 8.83938015e-01 -6.45920575e-01 1.46630198e-01
-1.05854428e+00 6.85084820e-01 8.41462672e-01 7.29027867e-01
7.47931302e-01 -4.58001703e-01 -5.98722816e-01 7.67981529e-01
3.67594540e-01 1.50129378e-01 5.09327710e-01 -9.95603979e-01
3.16630602e-01 8.81961763e-01 3.07111114e-01 -9.39920545e-01
8.50594267e-02 -5.98445237e-01 -2.76008934e-01 4.76953179e-01
6.48774683e-01 1.59368172e-01 -8.05806160e-01 1.69149578e+00
3.83202314e-01 -4.57824498e-01 -1.81154981e-01 7.42784679e-01
2.65849590e-01 2.81291634e-01 3.92106920e-01 1.31425336e-01
1.04577720e+00 -7.72839904e-01 -4.48804975e-01 -2.34300941e-01
7.42127120e-01 -6.84726298e-01 1.09784055e+00 5.30486763e-01
-7.91738272e-01 -5.35139322e-01 -1.22157633e+00 3.73007171e-02
-3.08108807e-01 -1.23156518e-01 5.79909921e-01 4.96936917e-01
-2.63119131e-01 8.26573133e-01 -3.17223370e-01 -1.98865924e-02
6.64498270e-01 1.34534761e-01 -7.16597885e-02 -1.65609330e-01
-1.17747211e+00 1.00726974e+00 2.16454431e-01 -4.21813697e-01
-1.01710927e+00 -9.24269617e-01 -5.33597231e-01 1.24280125e-01
3.48994941e-01 -5.15151620e-01 1.30979443e+00 -1.25594187e+00
-9.56557214e-01 9.88415956e-01 -2.61760622e-01 -6.86303139e-01
9.33732092e-01 -3.04207116e-01 -3.65604937e-01 -3.71631056e-01
4.48945314e-01 6.22365355e-01 1.10885024e+00 -1.29073858e+00
-9.30822194e-01 -3.46382171e-01 1.86661139e-01 -6.22126088e-02
-7.30511397e-02 -1.27883956e-01 2.73433384e-02 -7.21900761e-01
-6.63814172e-02 -1.08268976e+00 8.98834839e-02 2.14015916e-01
-4.93111104e-01 -5.63996375e-01 5.64671516e-01 -3.75017136e-01
1.42271638e+00 -2.16344523e+00 -3.57285112e-01 3.85200799e-01
2.75971055e-01 2.79429525e-01 9.84890237e-02 3.21766548e-02
-2.51956522e-01 2.31256634e-01 -1.46462858e-01 -9.71108750e-02
1.53365925e-01 -6.80631548e-02 -7.32424319e-01 2.34239131e-01
4.52059418e-01 4.17010635e-01 -7.53125310e-01 -1.97440132e-01
-1.52489781e-01 2.37611355e-03 -5.75039446e-01 2.90374339e-01
-3.48575413e-01 5.68945520e-02 -1.95624977e-01 4.86672014e-01
7.23259091e-01 -2.69873708e-01 1.25510022e-01 -3.05861294e-01
2.19375640e-01 8.44909191e-01 -1.24004829e+00 8.44714642e-01
-4.78789330e-01 5.13892174e-01 -3.51917684e-01 -7.93268085e-01
9.57172692e-01 1.24619796e-03 2.21391252e-04 -4.78553206e-01
-5.86808473e-02 3.73678118e-01 2.26612821e-01 -3.82413030e-01
5.17686605e-01 -3.33048582e-01 4.99373488e-02 4.82767373e-01
3.83005776e-02 2.63995063e-02 1.87955480e-02 2.49581516e-01
7.75803268e-01 2.33544454e-01 3.65554363e-01 -1.71123356e-01
2.94967622e-01 1.47302449e-01 6.15491927e-01 9.25759196e-01
-3.28628600e-01 5.13523102e-01 8.05258751e-01 -2.06612542e-01
-1.03206921e+00 -1.00535762e+00 -3.73995632e-01 1.18334448e+00
-1.57354042e-01 -3.37531507e-01 -4.57499892e-01 -1.44259572e+00
6.20370686e-01 1.39772177e+00 -9.23502803e-01 -4.41852421e-01
-4.02563550e-02 -5.58912218e-01 5.15739679e-01 6.57803893e-01
3.29170555e-01 -5.24566472e-01 -3.03149790e-01 2.26878464e-01
3.36703728e-03 -8.24087322e-01 -2.94589043e-01 3.41681868e-01
-6.88425362e-01 -1.10749662e+00 -6.11831188e-01 -2.89935946e-01
7.97639549e-01 1.39601901e-01 1.31589007e+00 -1.39716081e-02
1.68419495e-01 -1.15337521e-01 -8.59542340e-02 -6.73379600e-01
-6.95379436e-01 9.26174447e-02 1.02127362e-02 6.53019845e-02
1.02637696e+00 -4.91049550e-02 -3.33377272e-01 6.03447676e-01
-5.58391511e-01 -1.93139642e-01 5.51277101e-01 8.23862314e-01
4.80285794e-01 2.18488052e-01 7.71364093e-01 -1.43257868e+00
5.16465306e-01 -6.26651883e-01 -6.63211524e-01 2.54807979e-01
-1.30571175e+00 3.25625062e-01 3.63857746e-01 -6.28699481e-01
-1.22895014e+00 -1.99639753e-01 7.38820285e-02 -2.93128073e-01
-1.26684576e-01 3.34792882e-01 -2.56445467e-01 2.85681546e-01
8.09478223e-01 -2.29153544e-01 -1.09979495e-01 -3.98351789e-01
1.91499926e-02 8.26074839e-01 7.00070933e-02 -6.46015823e-01
7.66577303e-01 3.12819004e-01 -2.16193333e-01 -2.31965691e-01
-1.62214398e+00 -2.97154486e-01 -2.09665745e-01 -3.89387645e-02
4.03965592e-01 -9.28600729e-01 -3.53094876e-01 1.41342610e-01
-8.14152002e-01 -7.73863718e-02 -3.09341043e-01 4.82060611e-01
-3.93985882e-02 -2.08630458e-01 -4.99295220e-02 -8.21124315e-01
7.69220069e-02 -9.11962211e-01 9.16219592e-01 1.10471413e-01
-8.83894444e-01 -7.59810388e-01 -1.21031022e-02 3.62321317e-01
4.19127196e-01 -3.00345235e-02 1.15433466e+00 -1.08766115e+00
-3.51037323e-01 -6.94741070e-01 -2.80080259e-01 7.82979667e-01
5.58317155e-02 1.59214914e-01 -1.28603935e+00 8.86577740e-03
1.09684676e-01 -3.00896883e-01 9.34520781e-01 1.92281380e-01
1.21023762e+00 -2.28161067e-01 -3.35230827e-01 5.85661307e-02
1.21063328e+00 -6.74711838e-02 3.82311732e-01 3.21076244e-01
4.84164894e-01 9.15866494e-01 9.43021178e-01 2.17167377e-01
3.61652195e-01 6.57114923e-01 2.23914966e-01 1.36041626e-01
-7.84390941e-02 -7.16767848e-01 4.50100899e-01 -1.44638404e-01
3.33045244e-01 -1.69859573e-01 -6.28465772e-01 6.48581862e-01
-1.70739198e+00 -8.94057393e-01 -1.27955496e-01 2.48952603e+00
7.63439536e-01 8.00470710e-01 1.70170456e-01 1.84469506e-01
6.68241322e-01 7.08765090e-02 -7.90336549e-01 -3.48068088e-01
-5.49023636e-02 -1.08737960e-01 5.90643048e-01 4.00244296e-01
-9.59475815e-01 5.44552743e-01 6.07917786e+00 6.31520033e-01
-9.78784323e-01 -1.69006791e-02 8.30993652e-01 -7.36801475e-02
-5.72191179e-01 1.22482605e-01 -1.25407827e+00 6.54431462e-01
6.15439057e-01 -3.24064404e-01 -8.35125372e-02 1.33883953e+00
-8.55661184e-02 -2.93148667e-01 -1.58952558e+00 4.25365388e-01
8.98829997e-02 -1.10369992e+00 2.17097208e-01 3.75221968e-01
7.87363529e-01 -1.52857602e-01 2.44501874e-01 6.67255998e-01
4.02417004e-01 -7.59064913e-01 9.33280587e-01 3.04632366e-01
2.17736110e-01 -6.05933785e-01 9.11988378e-01 4.15048271e-01
-5.53581417e-01 3.49200040e-04 -4.61814910e-01 -5.36216646e-02
-3.07528347e-01 1.03196228e+00 -1.12478316e+00 1.04297675e-01
5.60155272e-01 5.05771935e-01 -6.82795942e-01 6.92624331e-01
-5.24478197e-01 8.85525167e-01 -1.86518356e-02 8.72825757e-02
8.36331695e-02 3.27255815e-01 4.18940812e-01 9.30747747e-01
-7.03224167e-02 -3.25130045e-01 -7.39744902e-02 1.20460832e+00
-3.18966001e-01 -2.20407754e-01 -6.25379026e-01 -1.56980902e-01
4.80261683e-01 1.06875300e+00 -1.00593746e-01 -5.60983360e-01
-3.60978752e-01 4.94176775e-01 3.45496565e-01 1.11148886e-01
-9.34736490e-01 -3.51839721e-01 7.05103159e-01 4.60276932e-01
4.24539447e-01 4.10987586e-01 -5.73745370e-01 -1.02563572e+00
1.27068460e-01 -8.06004882e-01 5.00930965e-01 -7.00805545e-01
-1.69249237e+00 3.76695275e-01 -1.74135938e-01 -1.19087148e+00
-2.16605082e-01 -5.92943668e-01 -2.95439452e-01 1.00371480e+00
-1.75686038e+00 -6.00181758e-01 -1.52167588e-01 6.06760643e-02
2.94365764e-01 -5.24065867e-02 5.98430872e-01 1.79901198e-01
-3.49492669e-01 8.09894085e-01 -3.33172232e-01 1.29476152e-02
9.91318762e-01 -1.42739606e+00 1.40065461e-01 7.29410529e-01
-7.85581097e-02 8.03264141e-01 8.98198783e-01 -7.09346116e-01
-5.08063257e-01 -1.19355309e+00 1.39154708e+00 -7.98289120e-01
5.68652630e-01 -1.86741292e-01 -1.10166144e+00 6.49837077e-01
-1.87628925e-01 -1.54206812e-01 8.31000686e-01 3.31215739e-01
-9.00316715e-01 -1.92934364e-01 -1.43768179e+00 3.49058837e-01
7.20326543e-01 -5.99992633e-01 -8.63079011e-01 1.70877367e-01
3.41440022e-01 -1.96454704e-01 -6.98093116e-01 2.00672776e-01
6.08801067e-01 -8.74876440e-01 6.13765180e-01 -9.03726816e-01
8.45825374e-01 -4.34752822e-01 -4.09873456e-01 -1.32524061e+00
-3.97020221e-01 -1.34574667e-01 -8.02715123e-02 1.47586274e+00
7.12275267e-01 -5.92104018e-01 7.51253843e-01 1.04329026e+00
3.25308323e-01 -7.74909317e-01 -3.49520177e-01 -7.85790861e-01
-5.15870228e-02 -3.31043392e-01 6.95971012e-01 1.00185335e+00
-1.94804072e-01 4.82988268e-01 -1.88054517e-01 2.52648175e-01
6.34502947e-01 3.01826060e-01 8.47709119e-01 -1.32269228e+00
-2.92315900e-01 -3.37045163e-01 -1.21666834e-01 -1.00401580e+00
2.91629374e-01 -8.92364800e-01 1.33770838e-01 -1.25222349e+00
2.29632810e-01 -6.09610438e-01 -4.42682475e-01 4.77909684e-01
-4.37576443e-01 8.45523551e-02 9.91277397e-02 3.59690070e-01
-3.47545922e-01 2.23196059e-01 8.19156468e-01 -7.92127326e-02
2.46412046e-02 3.56380552e-01 -1.11724472e+00 9.21873987e-01
5.48551381e-01 -9.77248371e-01 -2.70567566e-01 -1.76662698e-01
3.48314673e-01 -4.79870319e-01 5.16700506e-01 -5.49618959e-01
-1.35533204e-02 3.49988304e-02 4.92471933e-01 -4.34041351e-01
1.05219251e-02 -8.62033367e-01 -1.14550412e-01 3.85972708e-01
-1.07516062e+00 -1.96462244e-01 -1.35295853e-01 7.47602522e-01
-2.11052686e-01 -4.19346005e-01 9.67460275e-01 -6.31123111e-02
-4.15743977e-01 -1.31781682e-01 -1.10226527e-01 2.14024767e-01
8.33365858e-01 -1.89174935e-01 -4.21248496e-01 -1.67284578e-01
-5.26295781e-01 9.15209502e-02 5.40358365e-01 5.35960913e-01
1.53713971e-01 -1.14757681e+00 -7.09054232e-01 2.20714048e-01
6.61802888e-01 -4.21363890e-01 -1.95719793e-01 6.11783147e-01
1.90709040e-01 4.99748677e-01 1.16585813e-01 -3.85457993e-01
-1.01894665e+00 4.95733887e-01 5.03874302e-01 -4.37125832e-01
-1.17584042e-01 8.98179054e-01 3.38292778e-01 -4.83124435e-01
3.36336046e-01 -8.57172459e-02 1.24130383e-01 1.58773601e-01
5.31677485e-01 4.40530449e-01 2.70210654e-01 -4.26338732e-01
-4.51247424e-01 -3.27050425e-02 -5.29615700e-01 -6.48331419e-02
1.20785952e+00 1.47549352e-02 1.38225198e-01 6.47594810e-01
1.06008828e+00 2.36923620e-01 -1.28627634e+00 -4.78240907e-01
2.15607375e-01 -8.31793308e-01 2.43630722e-01 -1.25959802e+00
-8.90124619e-01 7.53415942e-01 3.87502789e-01 2.27732614e-01
6.22944415e-01 -1.11872397e-01 3.70071590e-01 1.33071437e-01
3.70193630e-01 -1.13895774e+00 -9.60706845e-02 1.88256297e-02
6.99549139e-01 -1.41237795e+00 3.33400108e-02 -3.01674902e-01
-1.01541018e+00 9.35617924e-01 8.64371717e-01 -2.63112456e-01
6.25435412e-01 -6.86045513e-02 -2.95652468e-02 -1.75822929e-01
-8.78381729e-01 1.25822350e-01 4.57115769e-01 4.02052045e-01
5.95786512e-01 7.57000148e-02 -1.96862102e-01 6.80984437e-01
2.40144376e-02 2.27785751e-01 3.57297659e-01 7.07286656e-01
-2.85747826e-01 -8.97797585e-01 -1.21337704e-01 9.22095954e-01
-5.21174431e-01 6.40091300e-02 -5.90202630e-01 6.93633139e-01
2.75329798e-01 8.59263122e-01 7.31761307e-02 -2.57532418e-01
5.91362178e-01 3.39312941e-01 3.14491093e-01 -7.90227711e-01
-6.56478524e-01 -3.45638454e-01 3.28165621e-01 -5.63981712e-01
-2.83618093e-01 -6.37756765e-01 -8.04087818e-01 -2.28327826e-01
-5.48148930e-01 7.47308061e-02 6.17723703e-01 9.00107801e-01
4.53269809e-01 4.65377837e-01 6.59698486e-01 -8.34495574e-02
-1.12431586e+00 -8.73374581e-01 -5.36436379e-01 6.98652089e-01
2.32557744e-01 -8.36677194e-01 -6.34442449e-01 -2.28501990e-01] | [9.040609359741211, 4.624279022216797] |
50076b8b-7993-42a6-a553-00beb0365a49 | native-language-identification-using | null | null | https://aclanthology.org/C12-1027 | https://aclanthology.org/C12-1027.pdf | Native Language Identification using Recurring $n$-grams -- Investigating Abstraction and Domain Dependence | null | ['Detmar Meurers', 'Serhiy Bykh'] | 2012-12-01 | native-language-identification-using-1 | https://aclanthology.org/C12-1027 | https://aclanthology.org/C12-1027.pdf | coling-2012-12 | ['native-language-identification'] | ['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] | [-7.340203762054443, 3.6877286434173584] |
a3af051f-d8b4-412d-bf8b-6055529d8fb0 | conversational-text-to-sql-an-odyssey-into | 2302.11054 | null | https://arxiv.org/abs/2302.11054v1 | https://arxiv.org/pdf/2302.11054v1.pdf | Conversational Text-to-SQL: An Odyssey into State-of-the-Art and Challenges Ahead | Conversational, multi-turn, text-to-SQL (CoSQL) tasks map natural language utterances in a dialogue to SQL queries. State-of-the-art (SOTA) systems use large, pre-trained and finetuned language models, such as the T5-family, in conjunction with constrained decoding. With multi-tasking (MT) over coherent tasks with discrete prompts during training, we improve over specialized text-to-SQL T5-family models. Based on Oracle analyses over n-best hypotheses, we apply a query plan model and a schema linking algorithm as rerankers. Combining MT and reranking, our results using T5-3B show absolute accuracy improvements of 1.0% in exact match and 3.4% in execution match over a SOTA baseline on CoSQL. While these gains consistently manifest at turn level, context dependent turns are considerably harder. We conduct studies to tease apart errors attributable to domain and compositional generalization, with the latter remaining a challenge for multi-turn conversations, especially in generating SQL with unseen parse trees. | ['Dilek Hakkani-Tur', 'Lu Zeng', 'Sree Hari Krishnan Parthasarathi'] | 2023-02-21 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [ 1.96619585e-01 5.49389362e-01 -1.08097434e-01 -9.04453456e-01
-1.76406109e+00 -9.51658964e-01 7.34257638e-01 3.31880212e-01
-1.31129399e-01 4.73462939e-01 7.03411877e-01 -6.41903877e-01
1.38905719e-01 -6.63044155e-01 -1.10040176e+00 2.27002561e-01
1.47624344e-01 1.21726012e+00 2.17757732e-01 -7.09332705e-01
-2.46600639e-02 -1.99356318e-01 -1.10208261e+00 1.29862523e+00
9.34588850e-01 8.22340727e-01 -7.18010962e-02 9.50267076e-01
-7.53966331e-01 1.16544676e+00 -8.44799697e-01 -8.11589301e-01
-1.02551311e-01 -1.48598224e-01 -1.21178389e+00 -3.15624386e-01
8.44573081e-01 -2.16579080e-01 1.06625468e-01 3.76106590e-01
4.70350683e-01 -2.70168424e-01 1.25621289e-01 -1.16454268e+00
-1.97266877e-01 1.15717041e+00 7.22860619e-02 -1.72710881e-01
9.51601982e-01 2.84126639e-01 1.41162086e+00 -7.73981035e-01
7.59313405e-01 1.66782999e+00 8.66499424e-01 3.97738636e-01
-1.63082707e+00 -3.16256523e-01 3.66225466e-02 -1.50075451e-01
-1.02346003e+00 -8.37776303e-01 1.98829547e-01 -3.25528532e-01
1.85973895e+00 7.05266118e-01 8.38357061e-02 1.26530099e+00
6.38290346e-02 1.08003354e+00 1.07507133e+00 -4.84514832e-01
-1.93551078e-01 2.86859006e-01 1.22437797e-01 7.18959391e-01
-3.39207143e-01 -2.68533349e-01 -8.32941353e-01 -4.88419950e-01
8.43487009e-02 -4.59643006e-01 2.04337183e-02 -1.45786777e-01
-1.26110673e+00 7.47840703e-01 -1.78917274e-02 -4.84037399e-02
-2.94660062e-01 -1.77010428e-02 8.36182714e-01 6.11578166e-01
3.55889767e-01 9.02994812e-01 -9.61551905e-01 -6.25504434e-01
-8.27955961e-01 8.56390357e-01 1.53148925e+00 1.58504391e+00
7.53848314e-01 -3.04959774e-01 -6.37008488e-01 1.05209816e+00
-1.20348603e-01 6.11237407e-01 3.15692216e-01 -1.07483232e+00
1.25444400e+00 7.29607761e-01 7.28968233e-02 -4.39123273e-01
-3.38984430e-01 -3.08822840e-01 2.93881614e-02 -6.26754940e-01
5.78958035e-01 -1.79129779e-01 -6.90593064e-01 1.75829673e+00
1.43041566e-01 -7.05679417e-01 4.51229453e-01 4.06887889e-01
6.27962112e-01 6.02375746e-01 1.65613770e-01 -9.42629352e-02
1.59828937e+00 -9.82302904e-01 -5.67122102e-01 -5.62177181e-01
1.02023542e+00 -9.82612669e-01 1.72459006e+00 3.46548080e-01
-1.30571330e+00 -4.61655825e-01 -5.81961274e-01 -4.38438594e-01
-1.93303093e-01 -5.85031584e-02 5.99751949e-01 5.54031193e-01
-9.55541968e-01 1.31914675e-01 -7.82564878e-01 -6.67057157e-01
-1.82743311e-01 -1.47471204e-01 -3.03165186e-02 -6.62741661e-02
-1.23435462e+00 9.40365076e-01 1.51355013e-01 -3.74689728e-01
-8.92865241e-01 -8.89286637e-01 -8.39203894e-01 -1.59543484e-01
7.71984994e-01 -5.86380184e-01 2.16821122e+00 -3.39284390e-01
-1.54917943e+00 8.67079914e-01 -7.28201270e-01 -6.77798331e-01
5.96553385e-01 -5.29537201e-01 -3.58451933e-01 -1.80348828e-01
5.35507560e-01 4.63258117e-01 9.51597393e-02 -9.19044793e-01
-6.74306393e-01 -2.18823925e-01 1.88491106e-01 2.50401348e-01
1.19609699e-01 1.88462198e-01 -3.56709927e-01 -2.69819140e-01
1.94430556e-02 -8.67185056e-01 1.13066711e-01 -7.67506838e-01
-7.46396065e-01 -6.81941688e-01 4.40802366e-01 -9.56850648e-01
1.55909419e+00 -1.69112372e+00 1.50322809e-03 -9.06005576e-02
-1.05686478e-01 -1.84421688e-01 -1.09767154e-01 1.05456090e+00
2.05720454e-01 2.65368223e-01 -1.31732866e-01 -2.99225122e-01
3.38585824e-01 1.46472797e-01 -7.12940514e-01 -4.05552864e-01
4.08163011e-01 1.17232811e+00 -5.58742881e-01 -5.09831548e-01
-2.17738628e-01 -3.76280189e-01 -8.86554778e-01 6.99989676e-01
-1.16332829e+00 -2.49577016e-02 -3.59142423e-01 6.51337504e-01
1.62297487e-01 -2.17533723e-01 4.44750637e-01 -1.99059069e-01
-1.13468088e-01 1.27234459e+00 -5.59932172e-01 1.88775396e+00
-8.82037222e-01 4.12572891e-01 5.08726425e-02 -5.45900404e-01
7.71254420e-01 3.39283258e-01 7.50703290e-02 -9.80631649e-01
-6.08366609e-01 4.48382914e-01 -1.86409652e-01 -1.01365435e+00
7.15915680e-01 -1.38228610e-02 -6.49337947e-01 4.60996151e-01
5.25525585e-02 -4.56443220e-01 1.55444592e-01 6.12855673e-01
1.29606795e+00 6.20282330e-02 1.23314904e-02 -1.38596939e-02
6.90698922e-02 6.59311295e-01 3.33938926e-01 9.82944608e-01
2.69711107e-01 4.77193236e-01 8.92687023e-01 -3.31357598e-01
-1.25397182e+00 -8.01375985e-01 -1.01366465e-03 1.83031678e+00
-4.69290435e-01 -7.33006716e-01 -8.63654494e-01 -8.89074385e-01
3.15107048e-01 1.43599761e+00 -1.59642577e-01 2.14048609e-01
-1.04641068e+00 -4.53082591e-01 1.17096138e+00 4.16084766e-01
3.08016032e-01 -9.12601471e-01 -3.49687040e-01 6.86442852e-01
-6.95979297e-01 -1.46989584e+00 -5.72298884e-01 3.72286052e-01
-6.30514205e-01 -9.57922280e-01 -5.66489026e-02 -3.85913014e-01
-1.36367515e-01 -2.44602889e-01 1.77330625e+00 -2.93507993e-01
2.11447906e-02 4.81160074e-01 -2.87050635e-01 -4.47057933e-01
-9.58033502e-01 4.83735681e-01 -4.56794649e-01 -3.66399229e-01
5.81467807e-01 -1.30563661e-01 -3.83578651e-02 3.93471062e-01
-5.51715612e-01 2.54361898e-01 6.25100136e-01 9.34564590e-01
2.55745023e-01 -7.55568326e-01 4.77133006e-01 -1.52740407e+00
8.71399879e-01 -5.25458634e-01 -4.20901209e-01 7.25332439e-01
-8.26007366e-01 3.66035700e-01 5.10052919e-01 -1.66458905e-01
-1.27889347e+00 -2.21500650e-01 -1.63523316e-01 -3.32665779e-02
-1.94844082e-01 6.72752440e-01 -2.99680948e-01 5.55560827e-01
1.09594214e+00 3.48136872e-01 7.10990503e-02 -4.48512644e-01
6.32373571e-01 7.19854593e-01 5.64313769e-01 -1.04585397e+00
3.11824739e-01 -1.69673949e-01 -7.44479954e-01 -5.87585986e-01
-9.77004230e-01 -4.18514103e-01 -1.58505812e-01 1.17130049e-01
8.64770293e-01 -1.05239511e+00 -7.69289017e-01 1.24708556e-01
-1.40312386e+00 -8.06658447e-01 6.59456775e-02 -8.70151445e-02
-7.34459817e-01 1.31295279e-01 -7.72015035e-01 -7.23708570e-01
-4.48448598e-01 -1.09915519e+00 1.31699324e+00 -2.79589504e-01
-7.47632504e-01 -8.55643809e-01 9.53960344e-02 9.88411546e-01
6.60967112e-01 -1.72572359e-01 1.48371637e+00 -1.30534518e+00
-8.18993211e-01 -6.12464622e-02 -9.73453671e-02 9.87867713e-02
-2.95634776e-01 -1.68029338e-01 -9.41870451e-01 -1.07110202e-01
-8.32514539e-02 -1.04809976e+00 5.27942896e-01 -1.44165128e-01
8.91067386e-01 -7.82692850e-01 -2.73246855e-01 4.43081588e-01
9.68743980e-01 8.77284408e-02 2.65616179e-01 2.51015186e-01
5.21489263e-01 8.68560255e-01 7.77450204e-01 1.94188684e-01
1.15051842e+00 8.13851595e-01 2.09802985e-01 3.06126416e-01
-4.97566611e-02 -8.10646176e-01 4.35154140e-01 7.96441674e-01
7.45285094e-01 -4.63186711e-01 -1.36884844e+00 5.40829122e-01
-1.68388450e+00 -7.80330181e-01 -1.61300585e-01 1.99181259e+00
1.39889789e+00 3.23138416e-01 1.99319392e-01 -5.33001363e-01
4.61135894e-01 1.73597068e-01 -5.16048670e-01 -6.41748130e-01
-2.16527045e-01 1.62867367e-01 3.43792260e-01 8.30478489e-01
-7.42553413e-01 1.22468054e+00 6.49495554e+00 7.33984590e-01
-1.05981326e+00 1.90680340e-01 5.56339562e-01 -6.32623658e-02
-8.41042399e-01 2.72970498e-01 -1.14320910e+00 2.12404102e-01
1.34729326e+00 -1.23550229e-01 6.57749951e-01 8.71528566e-01
-5.15421107e-02 7.08453804e-02 -1.58448231e+00 5.34269869e-01
-2.15034392e-02 -1.28253043e+00 6.28904030e-02 -2.94842720e-01
4.36180890e-01 4.25722510e-01 -2.62238711e-01 1.08896565e+00
6.01610124e-01 -9.99875546e-01 8.96732330e-01 3.25642496e-01
8.63082111e-01 -2.54973888e-01 5.16808987e-01 6.17661059e-01
-7.65500605e-01 3.56305316e-02 2.86417012e-03 2.33455434e-01
1.48162812e-01 3.41587454e-01 -1.55249047e+00 3.94472390e-01
4.96559948e-01 7.59187341e-02 -6.06652737e-01 3.81554455e-01
-1.95916221e-02 7.59525716e-01 -4.38667685e-01 -3.65061462e-01
1.53193772e-01 2.85499156e-01 5.43248415e-01 1.68056047e+00
3.35894637e-02 4.94069420e-02 3.65496546e-01 9.56993282e-01
-1.63565978e-01 9.21502858e-02 -4.61378187e-01 -2.67474055e-01
8.91913235e-01 8.92443955e-01 2.02211559e-01 -6.79785013e-01
-4.06021863e-01 8.00047874e-01 5.91531038e-01 4.15141195e-01
-5.39514840e-01 -4.44293141e-01 5.20857275e-01 2.66068012e-01
5.63393869e-02 -4.36912104e-02 -2.81600416e-01 -1.09984529e+00
4.70322609e-01 -1.63445044e+00 5.61126828e-01 -6.57674313e-01
-1.21127248e+00 6.93868458e-01 2.37842321e-01 -5.66425920e-01
-9.58752036e-01 -2.70138323e-01 -3.50685656e-01 9.21457410e-01
-1.20884323e+00 -1.20899403e+00 -2.37854179e-02 3.24477255e-01
9.44316447e-01 -1.79277033e-01 1.02743196e+00 2.27288440e-01
-3.59802514e-01 8.82630229e-01 -2.57933229e-01 1.61082730e-01
1.03154039e+00 -1.62414491e+00 9.44059670e-01 4.25722122e-01
3.59424762e-02 9.89143252e-01 8.15973580e-01 -7.69583583e-01
-1.68701065e+00 -1.12212074e+00 1.60036731e+00 -1.13356149e+00
6.84582412e-01 -8.58280122e-01 -1.13444352e+00 1.15362930e+00
4.92207289e-01 -4.53990459e-01 4.67973709e-01 7.73707807e-01
-7.89185524e-01 -1.86817124e-01 -8.19612086e-01 3.35292876e-01
9.41797853e-01 -1.04103363e+00 -8.04029286e-01 6.57667398e-01
1.27598834e+00 -9.59298491e-01 -1.03585899e+00 4.17738706e-01
5.96969604e-01 -1.05218756e+00 6.03320420e-01 -9.96542513e-01
4.07164842e-01 2.54726857e-01 -4.97731209e-01 -1.07805979e+00
4.02672112e-01 -9.72956717e-01 2.80913591e-01 1.30147946e+00
1.03560734e+00 -5.06440163e-01 7.62186170e-01 8.23178172e-01
-5.42263508e-01 -7.15165019e-01 -7.77070582e-01 -6.79166675e-01
1.06528185e-01 -7.29197145e-01 7.34421432e-01 7.78715909e-01
3.34036380e-01 7.77269006e-01 -3.11794002e-02 6.24571219e-02
3.41990083e-01 2.10795090e-01 1.23754692e+00 -6.99263096e-01
-5.45072973e-01 -2.35700652e-01 4.75657314e-01 -1.53118718e+00
1.47708297e-01 -8.73489201e-01 3.27785790e-01 -1.23256814e+00
-1.72587745e-02 -6.81847274e-01 4.74599332e-01 5.87190628e-01
-1.49131671e-01 -6.62361503e-01 5.38758337e-02 1.82744592e-01
-8.89390111e-01 4.37827289e-01 6.15767300e-01 -3.13748606e-02
-3.19102883e-01 2.24702790e-01 -6.46710098e-01 1.74492165e-01
4.83932287e-01 -4.08655554e-01 -3.91894609e-01 -9.80612934e-01
3.30071032e-01 9.25780833e-01 1.54542223e-01 -6.23633623e-01
4.53552961e-01 -1.86264798e-01 -3.54663491e-01 -6.53292656e-01
3.99882674e-01 -1.84991449e-01 -6.54214919e-02 1.23262160e-01
-1.05994475e+00 4.00048226e-01 3.82392704e-01 3.50747824e-01
-3.67297083e-01 1.93852317e-02 2.07490742e-01 -3.39276463e-01
-2.25157171e-01 -2.07194656e-01 -2.74334639e-01 7.83171535e-01
3.53508323e-01 2.01957405e-01 -7.81688631e-01 -6.31009877e-01
-6.05814695e-01 6.92160308e-01 5.64652085e-02 5.17609954e-01
2.65123367e-01 -9.04008031e-01 -8.52307260e-01 2.09436134e-01
6.59206331e-01 1.98650241e-01 -6.58604652e-02 7.11105168e-01
-4.49506283e-01 1.07452440e+00 2.73382217e-01 -7.92456985e-01
-1.07665265e+00 5.73049746e-02 3.90274137e-01 -4.29567844e-01
-1.21291831e-01 1.12981510e+00 -6.65409714e-02 -1.20774043e+00
4.46195483e-01 -7.04412580e-01 4.75421697e-01 -8.04561377e-02
4.66630980e-02 2.39952981e-01 4.61665988e-01 4.33193520e-02
-3.19190264e-01 -8.33489820e-02 -3.14503849e-01 -5.14089286e-01
9.78974462e-01 -8.06814507e-02 -1.98887944e-01 5.64327061e-01
1.13076270e+00 2.67393827e-01 -8.50050271e-01 -6.15871847e-01
5.37798166e-01 -8.28294009e-02 -5.20335376e-01 -1.41078842e+00
-1.64036021e-01 7.95531631e-01 -7.58890510e-02 4.90252793e-01
5.73971868e-01 1.99960470e-01 9.89106417e-01 8.47394764e-01
4.03679341e-01 -9.17780042e-01 1.31152332e-01 7.89136469e-01
1.13196969e+00 -1.32538116e+00 -5.36100149e-01 -3.18942130e-01
-9.50391710e-01 1.00255716e+00 8.95250916e-01 4.33550000e-01
-1.11113107e-02 2.51464725e-01 4.06874597e-01 -3.24025750e-01
-1.55509329e+00 7.57053792e-02 -7.61944503e-02 1.95491433e-01
7.84184694e-01 1.29249722e-01 1.35137975e-01 5.10346532e-01
-7.79608846e-01 -3.57401967e-01 2.03931645e-01 8.34827721e-01
-3.67952645e-01 -1.23685575e+00 -1.12092055e-01 5.92384040e-01
-3.92725855e-01 -3.60398561e-01 -4.52206284e-01 8.34094226e-01
-4.22181159e-01 1.18626285e+00 7.81557560e-02 -4.84137982e-01
9.26909328e-01 7.39485800e-01 1.71000734e-01 -1.03453648e+00
-1.14050007e+00 1.11890612e-02 1.14189017e+00 -9.04209554e-01
2.28245080e-01 -9.17264521e-01 -1.06809199e+00 -3.72884452e-01
-2.17965052e-01 4.83948231e-01 6.87039196e-01 7.71169722e-01
6.64141417e-01 2.26313546e-01 3.66307110e-01 4.66879643e-02
-1.24840331e+00 -1.27710199e+00 2.08483890e-01 4.14426267e-01
3.49601328e-01 1.03818953e-01 6.68970728e-03 -1.41206205e-01] | [10.066301345825195, 7.867446422576904] |
a2ccc9b1-b27c-44c8-81d2-3f94bad9f163 | searching-for-the-holy-grail-of-sponsorship | 2305.09473 | null | https://arxiv.org/abs/2305.09473v1 | https://arxiv.org/pdf/2305.09473v1.pdf | Searching for the "Holy Grail" of sponsorship-linked marketing: A generalizable sponsorship ROI model | Marketers routinely allocate a significant portion of their budget to sponsorship. However, isolating the return on investment from such efforts has remained a challenge. Thus, a dataset of more than 5,800 sponsorships is analyzed using survival analysis approaches that utilizes the sponsor's renewal of the sponsorship as a proxy for positive ROI. In addition to its contribution to the sponsorship-linked marketing literature, the resulting model can be utilized by managers to generate predicted values relative to the sponsor's probability of renewal and the ultimate duration of time the sponsor will remain with the property, representing a novel managerial contribution. | ['Jonathan A. Jensen'] | 2023-04-11 | null | null | null | null | ['marketing', 'survival-analysis'] | ['miscellaneous', 'miscellaneous'] | [ 1.67158574e-01 4.53265399e-01 -7.40753293e-01 -1.50042385e-01
-5.29413521e-01 -3.70539218e-01 3.37753743e-01 5.37630260e-01
-3.66241097e-01 4.78640646e-01 4.16499853e-01 -6.63747787e-01
-4.24960017e-01 -7.85842180e-01 -4.77642953e-01 -4.39138651e-01
5.11588752e-02 4.73512448e-02 -3.92004520e-01 8.57948512e-02
6.99060023e-01 4.42468762e-01 -1.13156104e+00 -1.31609827e-01
1.78216964e-01 9.80778039e-01 3.99548039e-02 1.18641235e-01
3.36185783e-01 7.79652774e-01 -8.91727626e-01 -1.51023895e-01
2.33347550e-01 -2.55946785e-01 -2.46413663e-01 -4.88323011e-02
-2.40701120e-02 -8.20874095e-01 -1.86727345e-01 3.15135807e-01
1.28004491e-01 -4.14403267e-02 5.84878802e-01 -1.37118089e+00
-5.50301611e-01 8.89226496e-01 -5.27571976e-01 5.19803226e-01
3.42502654e-01 2.49518454e-01 1.50127184e+00 -4.10616159e-01
9.53183353e-01 9.79024649e-01 4.30596054e-01 -3.39803755e-01
-1.45960164e+00 -8.77484202e-01 6.36001769e-03 -4.08450454e-01
-1.04547656e+00 -4.68200952e-01 6.37186825e-01 -7.67053425e-01
7.39020824e-01 -8.82186890e-02 1.00233960e+00 7.98477709e-01
8.63711834e-01 1.71159565e-01 1.04930794e+00 -3.06767374e-01
2.23203301e-02 3.24179202e-01 6.21362962e-02 -2.74765819e-01
5.80078959e-01 5.63853145e-01 -5.40655673e-01 -4.02467370e-01
6.92912936e-01 1.54539689e-01 7.10385069e-02 -8.89146775e-02
-1.12880719e+00 1.00091612e+00 2.26760253e-01 1.81783691e-01
-6.18192494e-01 6.42442644e-01 -1.57733381e-01 2.30535910e-01
2.76525557e-01 7.16610193e-01 -1.76562116e-01 -2.80713022e-01
-1.14444864e+00 5.15134692e-01 8.05039227e-01 6.83667779e-01
5.25947869e-01 -1.16075143e-01 -1.01377100e-01 -1.34972511e-02
4.86816049e-01 3.25144917e-01 -8.36721882e-02 -1.13529563e+00
2.44105712e-01 7.03214288e-01 5.72565556e-01 -1.00434721e+00
-1.92426562e-01 -9.34704542e-01 4.66813803e-01 2.83141583e-01
3.51466626e-01 -1.30657226e-01 -3.04627627e-01 1.43686569e+00
-2.60048866e-01 -4.53462094e-01 -1.48249771e-02 3.50797623e-01
1.49622291e-01 4.34003025e-01 1.15311675e-01 -4.07899082e-01
1.04080272e+00 -6.00245953e-01 -4.88676101e-01 -2.00043127e-01
4.18781906e-01 -8.68977845e-01 8.07795644e-01 3.41947794e-01
-1.05370927e+00 -6.43734410e-02 -9.48208869e-01 7.96557784e-01
1.54349998e-01 -4.15353268e-01 1.02328098e+00 4.33983654e-01
-6.45131767e-01 6.76832259e-01 -7.27725744e-01 -3.61219347e-01
7.90018916e-01 1.98389634e-01 3.19123119e-01 -1.62667073e-02
-7.97417283e-01 8.28223169e-01 -7.68127590e-02 -1.03966437e-01
-1.22085774e+00 -9.25636649e-01 -1.24713704e-01 3.19651157e-01
2.96793759e-01 -7.43305922e-01 1.06468952e+00 -7.02178121e-01
-3.32638085e-01 2.81409264e-01 3.05685669e-01 -4.59016234e-01
3.82320344e-01 2.29069918e-01 -2.69706100e-01 -2.02706326e-02
6.01534545e-01 5.32369614e-01 5.43357253e-01 -9.62716877e-01
-9.14223790e-01 -6.61890924e-01 2.81616718e-01 2.77136583e-02
-5.34091830e-01 1.74962968e-01 5.88164032e-01 -7.32005417e-01
9.84706134e-02 -8.61523271e-01 -4.75910366e-01 -6.73458338e-01
-3.39790061e-02 1.63727417e-03 4.11710382e-01 -3.80305469e-01
1.37203872e+00 -2.19141960e+00 -4.10347104e-01 1.54151563e-02
1.82260334e-01 -8.89125586e-01 8.31037611e-02 1.03763831e+00
1.29483506e-01 3.75088096e-01 2.71463096e-01 3.42080206e-01
-2.00732127e-02 -3.75936389e-01 -2.34594777e-01 4.12086844e-01
2.77660906e-01 6.77983820e-01 -5.91902852e-01 -7.41608813e-02
-1.51769921e-01 3.60306464e-02 -2.44180545e-01 -1.48723558e-01
3.82269233e-01 9.05556604e-02 -6.28493845e-01 1.23266828e+00
2.80767322e-01 -3.01970094e-01 2.33579740e-01 2.82029599e-01
-2.56962657e-01 2.19149500e-01 -5.12375474e-01 7.72111237e-01
-5.62209077e-02 6.08801663e-01 2.44952291e-01 -6.96029305e-01
1.10827589e+00 -2.79167276e-02 5.47824204e-01 -4.44254726e-01
3.27941805e-01 1.66327700e-01 3.86240244e-01 -1.15278922e-02
9.12878335e-01 -7.31792569e-01 -1.72184333e-01 7.69591033e-01
-6.00478292e-01 2.32941449e-01 -5.64777888e-02 4.21895564e-01
1.57221901e+00 7.52139762e-02 2.25973219e-01 -2.28143036e-01
-2.67078638e-01 6.49048030e-01 5.83379269e-01 3.42382044e-01
-3.28066289e-01 -1.04757525e-01 7.73393452e-01 2.40020722e-01
-8.74953449e-01 -1.00317705e+00 -3.81017268e-01 6.56804442e-01
1.76821630e-02 1.26260683e-01 -3.10059756e-01 -2.69143015e-01
8.02560627e-01 1.09436285e+00 -5.52737713e-01 -1.71278521e-01
-2.03194395e-01 -5.23653269e-01 2.23312095e-01 7.06438363e-01
-6.68725669e-02 -5.77775657e-01 -1.19274902e+00 5.65966547e-01
1.58807233e-01 -2.23063126e-01 -5.04840314e-01 4.15076882e-01
-1.22211993e+00 -1.18630755e+00 -6.25694394e-01 -1.35746762e-01
5.61449707e-01 4.94680107e-01 1.07547641e+00 -1.56654045e-01
-3.19923699e-01 7.21868157e-01 -1.91744626e-01 -5.66933095e-01
-2.68661499e-01 -7.77034611e-02 -9.73691791e-02 -6.85171664e-01
6.70975089e-01 -5.35768569e-01 -5.99524200e-01 3.55938762e-01
-5.36247611e-01 -6.38975203e-01 7.25979149e-01 6.94361508e-01
4.20901448e-01 3.57744873e-01 1.46745014e+00 -6.06370509e-01
8.79804492e-01 -8.31104159e-01 -5.30035794e-01 -7.77834132e-02
-1.39847350e+00 -6.44997776e-01 -2.58609742e-01 -5.72749734e-01
-1.03271973e+00 -4.49599028e-01 7.30622351e-01 2.49508210e-02
1.46760285e-01 1.07930338e+00 1.72885448e-01 1.83418870e-01
4.67766494e-01 -6.13520741e-01 3.88875455e-01 -2.17708126e-01
3.27487960e-02 4.21058863e-01 1.29252434e-01 -2.35723585e-01
6.10661626e-01 3.88202757e-01 -2.41718907e-02 -3.23506892e-01
-5.57513773e-01 -1.43836692e-01 -1.57169923e-01 -5.93191266e-01
3.97402853e-01 -9.87835348e-01 -9.03692901e-01 -3.96345258e-01
-5.27267992e-01 -2.63002198e-02 -5.03501654e-01 1.15511298e+00
-4.80177790e-01 -1.59267828e-01 -7.51228130e-04 -9.34372663e-01
2.15969548e-01 -1.07453120e+00 -1.02961123e-01 2.64233857e-01
-7.28205264e-01 -8.00696671e-01 -3.22441518e-01 1.01309061e+00
4.32761401e-01 9.65357006e-01 9.09744740e-01 -5.86922050e-01
-1.19642389e+00 -6.51989460e-01 -3.06038111e-02 2.76703864e-01
1.59254685e-01 -3.28784212e-02 -2.13375524e-01 -5.60356021e-01
4.29632068e-01 6.85641393e-02 6.78560972e-01 1.11835897e+00
1.20717324e-01 -3.29025015e-02 -6.34143412e-01 -3.00006978e-02
1.44685340e+00 7.69036174e-01 3.46321344e-01 1.04608321e+00
-1.74895987e-01 1.39129186e+00 1.59738553e+00 9.76423085e-01
2.59142786e-01 3.39029104e-01 5.44759750e-01 3.06645095e-01
2.79457003e-01 -4.32145834e-01 4.70539480e-01 7.96649680e-02
9.95898098e-02 -7.04046562e-02 -8.67117703e-01 7.15354025e-01
-1.35374177e+00 -9.16868150e-01 -2.77970672e-01 2.20022082e+00
2.66365290e-01 7.10482717e-01 3.83512914e-01 -1.64605230e-02
7.54180074e-01 1.21441849e-01 -9.96310830e-01 -3.57533127e-01
-1.07068136e-01 -3.17577541e-01 1.18060255e+00 -4.71361056e-02
-1.56204745e-01 4.37002271e-01 7.98617935e+00 5.64131439e-01
-5.95702767e-01 -1.45148605e-01 1.17439222e+00 -5.96218824e-01
-8.18543494e-01 1.08080065e+00 -1.02191329e+00 2.64617652e-01
1.25092566e+00 -8.63708735e-01 2.23462135e-01 8.32674205e-01
6.78087771e-01 -7.10510731e-01 -1.20116150e+00 3.58483903e-02
-2.09775660e-02 -1.29217041e+00 -5.12316763e-01 9.39549506e-01
4.89003569e-01 -6.43532813e-01 3.13973963e-01 -7.96995163e-02
2.38643214e-01 -9.98781919e-01 1.06969631e+00 5.91786683e-01
6.09683275e-01 -1.03677750e+00 6.96405470e-01 2.00152606e-01
-6.38276815e-01 -7.13126421e-01 -5.46348169e-02 -8.33351612e-01
6.50923848e-01 5.87815464e-01 -1.23753715e+00 7.65336975e-02
6.74767256e-01 3.63200843e-01 -3.56234998e-01 3.67261142e-01
5.01486100e-02 1.00558043e+00 1.41703397e-01 1.77412584e-01
2.43032724e-01 -4.17922884e-01 4.20033723e-01 1.13017410e-01
4.29259598e-01 -7.23849237e-02 -2.50008881e-01 1.16371787e+00
9.33000371e-02 1.13384493e-01 -6.36632442e-01 -1.12571740e+00
9.27014589e-01 1.12652349e+00 -9.75165784e-01 1.60825066e-02
-3.86198491e-01 -1.84059769e-01 -3.65429997e-01 2.92434007e-01
-5.80000281e-01 -7.32976556e-01 2.50729591e-01 8.38656068e-01
2.06895694e-01 -2.87444908e-02 -6.95167601e-01 -8.63988027e-02
-3.37833427e-02 -5.86427867e-01 -1.01573681e-02 -7.04719305e-01
-8.62788796e-01 -1.14453606e-01 8.87483656e-02 -9.34626222e-01
1.57590985e-01 5.64063862e-02 -4.82510030e-01 1.09528673e+00
-1.34291172e+00 -1.03755569e+00 -6.33838624e-02 -3.51004988e-01
-7.28558078e-02 -3.60226303e-01 4.63221036e-02 -1.29240409e-01
-3.14878881e-01 3.47468615e-01 9.74805206e-02 -6.75472021e-01
7.38398969e-01 -8.22314978e-01 -1.74750343e-01 4.24665749e-01
-6.52399838e-01 9.24406886e-01 5.35641432e-01 -1.12253952e+00
-1.30265665e+00 -5.20847023e-01 8.50914240e-01 -5.66175938e-01
1.03851759e+00 1.93888843e-01 -3.18678349e-01 8.75635684e-01
4.26025614e-02 -1.00831473e+00 1.01492786e+00 3.52709413e-01
4.97140259e-01 -4.40417469e-01 -1.16685534e+00 2.97125489e-01
7.23405302e-01 3.70048471e-02 -6.61639273e-01 -3.90054211e-02
8.63886774e-01 3.39207798e-01 -1.84160316e+00 -2.35904127e-01
6.70547247e-01 -8.78190696e-01 9.15559769e-01 -1.38558641e-01
8.25692832e-01 2.48931974e-01 -4.06954914e-01 -5.88828564e-01
-4.84923154e-01 -5.81799269e-01 5.69870830e-01 1.59179580e+00
4.37322080e-01 -7.56761730e-01 1.01207840e+00 8.72157395e-01
-1.82183221e-01 -1.11149621e+00 -9.90626693e-01 -8.83146942e-01
2.04016417e-02 -7.76549708e-03 9.16845500e-01 7.57000566e-01
2.16120742e-02 1.07620724e-01 -7.10342824e-02 -2.63336807e-01
8.67946506e-01 1.42362982e-01 8.77179861e-01 -1.50095737e+00
-1.21821508e-01 -5.48005223e-01 -1.77813068e-01 -6.26833618e-01
-1.89803511e-01 -7.41078854e-01 -3.66272360e-01 -1.62942064e+00
6.09788001e-01 -9.26811695e-01 -4.76188034e-01 2.78737754e-01
4.36130911e-01 -3.42417687e-01 1.58166245e-01 8.06259751e-01
3.45335513e-01 3.10774714e-01 1.25571120e+00 -2.95228343e-02
-6.10664666e-01 6.94170892e-02 -1.57568216e+00 4.91624057e-01
5.46454668e-01 -6.51879430e-01 -2.56953269e-01 2.60129690e-01
6.71931386e-01 8.10495257e-01 5.00525832e-01 -3.88938189e-01
-1.61670431e-01 -8.49968791e-01 1.90900743e-01 -1.00382650e+00
1.09451916e-03 -8.56060743e-01 9.91699696e-01 8.41333449e-01
-6.23591781e-01 1.49695814e-01 -1.49665549e-01 8.85108650e-01
-8.61941352e-02 -5.45685649e-01 3.59574050e-01 7.58584365e-02
2.57774860e-01 -2.83262342e-01 -4.00956303e-01 -1.53814539e-01
1.53911102e+00 -7.73473561e-01 -3.15957725e-01 -1.26494139e-01
-6.05972171e-01 1.85233325e-01 9.25228238e-01 2.22236857e-01
4.93095845e-01 -1.38403368e+00 -6.29439831e-01 -5.91737688e-01
1.62719801e-01 -5.55029452e-01 2.83372939e-01 1.09632242e+00
-2.61814535e-01 6.25571668e-01 -2.63078362e-01 -5.21434322e-02
-9.30083513e-01 5.96501172e-01 -1.40110061e-01 -1.85163189e-02
-5.46809554e-01 5.94601750e-01 4.51184005e-01 7.15358496e-01
-2.21635908e-01 -3.21621336e-02 -1.60732493e-01 6.90659761e-01
1.65534131e-02 1.05591416e+00 -5.36027372e-01 -5.91196299e-01
-4.96465772e-01 -1.80515230e-01 -2.74433643e-01 -3.50422025e-01
1.74643493e+00 -4.32452857e-01 -1.82972878e-01 6.79528177e-01
7.08963454e-01 8.24545100e-02 -1.48928332e+00 1.02008604e-01
2.73133248e-01 -1.06619918e+00 3.36065531e-01 -8.39033782e-01
-1.01478970e+00 2.21481025e-01 3.70920807e-01 7.76534975e-02
8.06829989e-01 1.37509078e-01 5.96302867e-01 -1.04662850e-01
1.80302620e-01 -1.32237673e+00 4.78104085e-01 -7.09377170e-01
8.40661764e-01 -8.19441438e-01 3.53500426e-01 -1.19597362e-02
-4.56458449e-01 7.16592371e-01 1.87504172e-01 -8.59179571e-02
7.28589594e-01 -1.10512488e-01 -3.60774100e-01 -4.05601621e-01
-7.97919095e-01 5.68650603e-01 -4.80017543e-01 5.75009286e-01
5.58140039e-01 4.36961502e-01 -1.07891619e+00 1.08966911e+00
-2.32251152e-01 6.36627972e-02 1.48576915e+00 1.14844692e+00
-5.91417611e-01 -1.08769977e+00 -5.03675222e-01 1.13137352e+00
-7.70877600e-01 -1.75710827e-01 -2.44321182e-01 1.01260304e+00
-2.05804640e-03 1.37809670e+00 -9.32755601e-03 -2.11729720e-01
4.14335668e-01 -3.81177723e-01 -5.39566614e-02 -6.04185343e-01
-7.79556870e-01 6.29086018e-01 2.38261834e-01 -1.73361957e-01
-6.44459009e-01 -1.51714838e+00 -1.11019254e+00 -6.12711132e-01
-8.56657445e-01 1.56379901e-02 7.85003901e-01 4.51778680e-01
1.90406024e-01 7.49280512e-01 1.03923011e+00 -2.22245395e-01
-1.01064229e+00 -8.87168944e-01 -1.28723228e+00 -7.23383725e-02
-2.41965726e-01 -1.04026175e+00 -8.79044652e-01 -1.34931192e-01] | [9.170112609863281, 5.871751308441162] |
e1915189-0b57-4e1b-8d07-d6018a2b6160 | sugar-efficient-subgraph-level-training-via | 2202.00075 | null | https://arxiv.org/abs/2202.00075v3 | https://arxiv.org/pdf/2202.00075v3.pdf | SUGAR: Efficient Subgraph-level Training via Resource-aware Graph Partitioning | Graph Neural Networks (GNNs) have demonstrated a great potential in a variety of graph-based applications, such as recommender systems, drug discovery, and object recognition. Nevertheless, resource-efficient GNN learning is a rarely explored topic despite its many benefits for edge computing and Internet of Things (IoT) applications. To improve this state of affairs, this work proposes efficient subgraph-level training via resource-aware graph partitioning (SUGAR). SUGAR first partitions the initial graph into a set of disjoint subgraphs and then performs local training at the subgraph-level. We provide a theoretical analysis and conduct extensive experiments on five graph benchmarks to verify its efficacy in practice. Our results show that SUGAR can achieve up to 33 times runtime speedup and 3.8 times memory reduction on large-scale graphs. We believe SUGAR opens a new research direction towards developing GNN methods that are resource-efficient, hence suitable for IoT deployment. | ['Radu Marculescu', 'Mengtian Yang', 'Yuedong Yang', 'Zihui Xue'] | 2022-01-31 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [-3.22382078e-02 4.82488945e-02 -6.35900199e-01 -1.29186705e-01
7.53953308e-02 -2.64993101e-01 2.35966712e-01 2.26529926e-01
-1.67010233e-01 6.77906752e-01 -1.77965626e-01 -9.16471899e-01
-3.25154632e-01 -1.33656049e+00 -5.34882903e-01 -6.55755937e-01
-4.85115588e-01 4.63046640e-01 3.41971725e-01 6.85182363e-02
-9.09087285e-02 6.00799084e-01 -1.26190138e+00 -3.29154223e-01
8.84610057e-01 9.65865195e-01 -1.46550700e-01 3.28626961e-01
-1.85023233e-01 5.78570604e-01 -2.08215326e-01 -4.74229068e-01
5.40746987e-01 -4.09190029e-01 -5.36611974e-01 6.55471254e-03
9.28842947e-02 7.75018781e-02 -5.62163055e-01 1.26669145e+00
4.71290290e-01 2.22582504e-01 5.67690320e-02 -1.42731977e+00
-4.15104508e-01 9.50626075e-01 -6.97960675e-01 2.05851525e-01
-1.04098916e-01 -3.88475694e-02 9.75521743e-01 -2.28930697e-01
4.61704016e-01 9.26095366e-01 7.28996456e-01 3.06519687e-01
-7.59867370e-01 -8.34900320e-01 3.11098397e-01 3.30438614e-01
-1.29517484e+00 1.15432277e-01 9.37534690e-01 1.12543426e-01
1.11720133e+00 2.44254977e-01 1.04126132e+00 5.69772840e-01
2.24715188e-01 4.88171071e-01 8.43190074e-01 -1.26481831e-01
5.20901442e-01 -3.05310905e-01 8.11482072e-02 7.68782914e-01
8.88551950e-01 -1.23599984e-01 -1.73109591e-01 -7.57061020e-02
9.00984883e-01 2.04663098e-01 -1.81404337e-01 -5.53508162e-01
-8.83966804e-01 9.21526611e-01 9.11806345e-01 3.34384263e-01
-4.74470884e-01 3.93859148e-01 6.41842186e-01 2.74887592e-01
4.69519138e-01 2.40370363e-01 -3.68547171e-01 -3.81659716e-02
-6.21898711e-01 -3.47671151e-01 1.01712155e+00 9.58127975e-01
6.92618847e-01 3.89879912e-01 4.35907841e-01 5.40493965e-01
2.50183523e-01 3.95465553e-01 3.06033671e-01 -3.19133073e-01
3.12829554e-01 1.03618753e+00 -6.11549914e-01 -1.28895140e+00
-7.98414052e-01 -7.01213896e-01 -1.34195650e+00 -3.72833341e-01
2.07554847e-01 -1.22632071e-01 -7.36443520e-01 1.45403385e+00
6.38212860e-01 6.66102827e-01 -2.78411478e-01 8.44843626e-01
9.40867126e-01 4.79768634e-01 6.07314296e-02 -1.16867356e-01
1.32033563e+00 -1.07132196e+00 -3.93464506e-01 -1.87474489e-01
8.69233787e-01 -1.81471854e-01 6.80160344e-01 2.66788095e-01
-6.98672354e-01 -2.96741307e-01 -9.74834681e-01 3.24874371e-01
-4.76640046e-01 -2.72406667e-01 1.31906307e+00 1.11184728e+00
-1.21499908e+00 6.40700936e-01 -8.96546006e-01 -5.44964910e-01
6.22465193e-01 6.92234039e-01 -1.77609622e-01 -1.05100714e-01
-9.21393991e-01 1.74539834e-01 5.78675151e-01 8.54203105e-03
-5.56985080e-01 -2.91783869e-01 -8.69796336e-01 3.42928708e-01
6.60866618e-01 -8.29118431e-01 6.04923248e-01 -6.00944698e-01
-1.25112391e+00 4.09495860e-01 3.03940058e-01 -7.37876534e-01
5.67882136e-02 1.95168585e-01 -6.52343273e-01 5.28194383e-02
-2.80848354e-01 3.25515568e-01 5.81377149e-01 -5.34638762e-01
-2.84953475e-01 -4.53205884e-01 2.43113965e-01 4.05112915e-02
-7.59085417e-01 -1.84158489e-01 -7.18149304e-01 -4.98377889e-01
2.85740882e-01 -1.21629822e+00 -5.16929269e-01 -4.52102661e-01
-5.39863169e-01 -3.47706199e-01 7.28696227e-01 -2.96559066e-01
1.39253891e+00 -1.81907749e+00 -1.29424617e-01 6.07508540e-01
6.54439151e-01 4.08292770e-01 -1.09813355e-01 3.73093307e-01
8.55541155e-02 1.93183124e-01 1.18595980e-01 1.33416116e-01
-7.45142326e-02 1.00041971e-01 2.20529228e-01 4.49179351e-01
-3.96906525e-01 1.22395694e+00 -9.55469489e-01 -1.29030392e-01
3.01604629e-01 5.34962296e-01 -5.79669237e-01 -2.43649349e-01
-2.67895639e-01 2.14331746e-01 -6.48515701e-01 7.18607008e-01
7.09663332e-01 -1.02382040e+00 6.87959254e-01 -1.79158360e-01
3.43958706e-01 3.27755332e-01 -1.19155097e+00 1.17061830e+00
-3.40445876e-01 4.44440901e-01 -2.45230258e-01 -1.17262888e+00
9.70179677e-01 -6.96914867e-02 6.86564028e-01 -8.01038206e-01
4.95930791e-01 4.72504124e-02 2.78306007e-01 -2.92993709e-02
3.10045153e-01 1.66525811e-01 1.75235178e-02 5.32582581e-01
-1.87841326e-01 3.85070175e-01 3.45148414e-01 2.26517305e-01
1.54624426e+00 -2.55917311e-01 4.35723215e-01 -4.20169920e-01
4.01279300e-01 -2.04074293e-01 2.85290122e-01 5.94865561e-01
-2.45354801e-01 -7.06019700e-02 3.70054990e-01 -8.86901021e-01
-8.12447369e-01 -6.04756534e-01 3.53490621e-01 9.67967391e-01
3.08111042e-01 -6.82991564e-01 -7.11571634e-01 -8.84707451e-01
-8.30453038e-02 2.77612418e-01 -3.11414063e-01 -2.52589911e-01
-4.22617346e-01 -8.80902290e-01 2.60036498e-01 5.89343369e-01
5.38266718e-01 -1.13883328e+00 -3.64261150e-01 3.03466260e-01
2.57187605e-01 -1.32533669e+00 -3.54321927e-01 6.37506694e-02
-1.28371775e+00 -1.15461111e+00 -3.27586472e-01 -8.18799734e-01
9.34187591e-01 7.19358742e-01 1.26169991e+00 5.41839123e-01
-5.24791777e-02 2.49457151e-01 -3.26158166e-01 -1.91434011e-01
-1.73020333e-01 4.38411355e-01 8.99670199e-02 -1.89752027e-01
4.59456533e-01 -9.04704213e-01 -6.66592181e-01 4.46898490e-01
-8.10863316e-01 2.70508928e-03 5.97238123e-01 5.01656294e-01
7.41662860e-01 5.86592436e-01 5.98971248e-01 -1.09426999e+00
6.18577242e-01 -5.33883870e-01 -1.09728301e+00 1.43144563e-01
-9.15769637e-01 2.93210447e-02 8.76235187e-01 -3.32387596e-01
-4.65273082e-01 -7.66127557e-02 -1.07027970e-01 -3.09526056e-01
1.30700812e-01 7.74826288e-01 -2.02206686e-01 -4.40670729e-01
4.14338946e-01 -2.67331321e-02 -3.18796933e-01 -1.30247444e-01
2.37566128e-01 3.82186145e-01 4.47263829e-02 -2.33173937e-01
7.02275515e-01 4.33166921e-01 3.58046621e-01 -8.51543665e-01
-5.79394281e-01 -6.25565171e-01 -3.40857841e-02 -2.68895596e-01
5.04680157e-01 -8.21771860e-01 -1.06314325e+00 8.45783874e-02
-6.54847920e-01 -6.22134626e-01 -9.11112060e-04 4.55387920e-01
-5.68196550e-02 4.28288192e-01 -5.39630592e-01 -5.05557060e-01
-7.26143241e-01 -9.88639534e-01 7.25015998e-01 5.04449785e-01
1.34412274e-01 -1.18227780e+00 -4.21440415e-02 1.90203145e-01
4.80529666e-01 7.99137726e-02 8.06993306e-01 -9.27790761e-01
-9.72347081e-01 -6.17076159e-02 -5.73773563e-01 -2.11311597e-02
1.46181673e-01 -4.45978522e-01 -3.87531370e-01 -5.60260355e-01
-3.09781969e-01 2.93488577e-02 7.88977444e-01 3.73882532e-01
1.35970342e+00 -2.81454474e-01 -6.25287890e-01 8.68885338e-01
1.78088427e+00 2.32553884e-01 6.52291358e-01 2.65165687e-01
1.13569713e+00 1.77087635e-02 1.63181141e-01 3.57682496e-01
3.44641209e-01 3.83995444e-01 8.13914716e-01 -1.08013861e-01
-1.14207283e-01 -1.14954494e-01 8.33304375e-02 1.06920040e+00
-3.15832675e-01 -7.08649278e-01 -8.09518695e-01 3.95652801e-01
-1.85834742e+00 -7.05453813e-01 -1.01315044e-01 2.20508862e+00
1.74623892e-01 2.97237426e-01 2.74672747e-01 1.09875247e-01
8.86635542e-01 7.90270343e-02 -8.49038899e-01 -3.50150526e-01
1.45870104e-01 3.22708011e-01 7.68357992e-01 -1.08298156e-02
-9.01467919e-01 9.67467904e-01 5.48812771e+00 7.80481815e-01
-1.17600369e+00 6.59134313e-02 6.65427387e-01 3.10997397e-01
-3.89200389e-01 5.65383472e-02 -6.40271664e-01 4.09290671e-01
1.14634418e+00 -4.02952433e-01 7.86335289e-01 1.14272797e+00
-7.20827281e-02 2.55996048e-01 -6.88248694e-01 1.13070273e+00
-1.97746575e-01 -1.46548808e+00 6.56229705e-02 4.44746971e-01
1.02585793e+00 5.65336883e-01 -3.25519115e-01 3.54352862e-01
5.13020277e-01 -7.77035713e-01 -8.35294798e-02 -1.26148239e-01
5.59255421e-01 -1.08903086e+00 8.65616739e-01 2.11697549e-01
-1.67254341e+00 -9.18435454e-02 -5.54707170e-01 -1.58301920e-01
5.88542856e-02 9.49347436e-01 -9.09257233e-01 8.81599724e-01
6.87966108e-01 5.51928699e-01 -5.05102277e-01 1.27338886e+00
-1.95411101e-01 7.87231445e-01 -5.15041590e-01 -5.14618695e-01
1.73767269e-01 -5.71231544e-01 3.98776531e-01 8.92087996e-01
4.09981638e-01 9.24430192e-02 3.45143586e-01 5.47501445e-01
-6.19726181e-01 4.37670767e-01 -7.10283637e-01 -3.96048218e-01
5.20606101e-01 1.44696426e+00 -1.63298130e+00 -3.15830916e-01
-5.37429810e-01 6.73030555e-01 3.72229606e-01 1.05088644e-01
-1.02563274e+00 -5.03586709e-01 3.68216842e-01 6.71973526e-02
5.78188896e-01 -2.67812431e-01 -6.46760687e-02 -1.06010056e+00
-5.30737899e-02 -8.61111999e-01 5.75357795e-01 -4.32145804e-01
-1.06975234e+00 7.32537270e-01 -2.37472236e-01 -1.01490879e+00
3.79240923e-02 -5.91696501e-01 -5.48594534e-01 1.41287312e-01
-1.36744666e+00 -9.12354589e-01 -5.82465231e-01 4.00744587e-01
3.16092372e-02 4.52352464e-02 5.66904008e-01 4.80542958e-01
-5.47812402e-01 5.34082413e-01 -2.73963269e-02 2.28583023e-01
3.80299687e-02 -9.93688941e-01 9.60098505e-01 9.21462238e-01
5.45156121e-01 5.73324144e-01 4.17110354e-01 -8.85137022e-01
-1.75431967e+00 -1.34913731e+00 6.26353025e-01 3.12759727e-01
7.48044014e-01 -2.46024817e-01 -6.52776659e-01 6.75053418e-01
2.58388836e-02 3.21944118e-01 5.36419034e-01 2.79302716e-01
-2.50292093e-01 -3.55163485e-01 -1.06501949e+00 7.76706457e-01
1.42054737e+00 -2.43955225e-01 3.17737639e-01 6.10148966e-01
8.03856492e-01 -4.49089259e-01 -1.01138175e+00 6.12107575e-01
4.21702325e-01 -1.04541802e+00 8.34755898e-01 -4.20930862e-01
-2.45833158e-01 -1.73378497e-01 1.51508272e-01 -1.13751352e+00
-3.61922383e-01 -8.32744241e-01 -4.70443726e-01 8.16075087e-01
2.20252365e-01 -1.14741313e+00 1.26993716e+00 1.75832510e-01
-6.75601736e-02 -9.87361312e-01 -6.70115769e-01 -1.01265848e+00
-4.62678581e-01 -4.90016431e-01 9.79968786e-01 9.53753531e-01
-2.84849629e-02 6.04436517e-01 -2.23477468e-01 1.14784822e-01
5.53110719e-01 3.53814542e-01 8.79402220e-01 -1.41758990e+00
-2.36566618e-01 -7.00328708e-01 -8.61379504e-01 -1.00453198e+00
3.99105623e-02 -1.11164784e+00 -4.65235323e-01 -1.84838831e+00
1.30336985e-01 -6.22199059e-01 -4.62401927e-01 5.74237287e-01
-4.85679656e-02 4.44892853e-01 -8.22437927e-02 -2.23958373e-01
-1.05457783e+00 3.27880830e-01 1.09032130e+00 -1.80833042e-01
-2.76726127e-01 1.52542830e-01 -8.11519682e-01 5.92681289e-01
1.20200038e+00 -5.85047543e-01 -7.21555769e-01 -1.44924968e-01
5.98133862e-01 -1.43996835e-01 7.60866404e-02 -1.21906447e+00
3.08914840e-01 5.44942059e-02 -7.38033727e-02 -5.25320530e-01
-1.73449129e-01 -9.51182544e-01 4.65776116e-01 6.70386791e-01
3.27186793e-01 2.17654616e-01 3.22576314e-01 7.38453269e-01
1.20348446e-01 2.34711021e-01 4.82207537e-01 4.31283563e-02
-7.57105410e-01 8.27316940e-01 1.92747146e-01 -1.31492121e-02
1.06217611e+00 -2.84184575e-01 -2.87731439e-01 -5.69488168e-01
-4.07206714e-01 1.56391248e-01 4.92419630e-01 2.42435262e-01
4.86773312e-01 -1.22690892e+00 -3.10336888e-01 2.79056489e-01
4.60291430e-02 -8.01857412e-02 3.16991538e-01 8.61559510e-01
-5.62799811e-01 5.24319947e-01 -4.35383655e-02 -6.82147861e-01
-1.12938190e+00 8.16195488e-01 9.79208052e-02 -6.85169160e-01
-8.27934206e-01 7.49798536e-01 -2.60040611e-02 -4.50837940e-01
3.00993383e-01 -2.00864330e-01 5.63656129e-02 -6.06060743e-01
3.27619761e-01 6.88139200e-01 3.70657086e-01 -4.69672352e-01
-4.11938399e-01 3.36188197e-01 6.98622689e-02 6.09635413e-01
1.35915124e+00 6.69091567e-02 -2.47210473e-01 -2.09070966e-01
1.06063557e+00 5.44134248e-03 -8.29868257e-01 -1.54669657e-01
8.06479082e-02 -1.26832098e-01 6.08968846e-02 -3.44058543e-01
-1.61478078e+00 4.23337579e-01 2.92935073e-01 7.18946636e-01
1.42726791e+00 1.21387981e-01 1.14432549e+00 6.44953310e-01
7.42170095e-01 -8.14054251e-01 -7.20960274e-02 3.68067622e-01
1.35663062e-01 -9.88357127e-01 3.11725795e-01 -6.16240442e-01
-3.63610387e-02 9.13595915e-01 5.39504707e-01 -3.09411645e-01
9.24188077e-01 1.62405059e-01 -4.44029719e-01 -5.13974190e-01
-5.14983177e-01 -3.46122712e-01 3.56915742e-01 4.50573683e-01
1.30976915e-01 3.81414562e-01 -1.91654846e-01 3.45524490e-01
-1.35248169e-01 -5.55575863e-02 1.90214977e-01 5.80531657e-01
-2.74645329e-01 -1.22976148e+00 2.47107565e-01 7.81982601e-01
-3.84431005e-01 -2.31723875e-01 -2.23195434e-01 9.06208992e-01
-1.91643611e-01 8.19344103e-01 -2.53642276e-02 -6.87565565e-01
-5.42358635e-03 -3.81535441e-01 5.53517342e-01 -4.48505491e-01
-4.57756251e-01 -1.52306259e-02 1.60999149e-01 -7.60063350e-01
-3.57971311e-01 -1.25367522e-01 -1.40827370e+00 -5.48842371e-01
-6.43866003e-01 2.26495698e-01 8.00617099e-01 8.41903090e-01
8.48360896e-01 7.67397046e-01 4.77542073e-01 -6.81465924e-01
-1.22922949e-01 -5.41882634e-01 -5.47735393e-01 1.87017843e-01
-2.47225195e-01 -5.22796214e-01 -3.30468565e-01 -6.24318123e-01] | [7.021803855895996, 5.758882999420166] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.