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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
fd8fd4ca-62d4-4439-96e7-2e20fecdc855 | unsupervised-training-of-a-deep-clustering | 1904.01340 | null | http://arxiv.org/abs/1904.01340v1 | http://arxiv.org/pdf/1904.01340v1.pdf | Unsupervised training of a deep clustering model for multichannel blind source separation | We propose a training scheme to train neural network-based source separation
algorithms from scratch when parallel clean data is unavailable. In particular,
we demonstrate that an unsupervised spatial clustering algorithm is sufficient
to guide the training of a deep clustering system. We argue that previous work
on deep clustering requires strong supervision and elaborate on why this is a
limitation. We demonstrate that (a) the single-channel deep clustering system
trained according to the proposed scheme alone is able to achieve a similar
performance as the multi-channel teacher in terms of word error rates and (b)
initializing the spatial clustering approach with the deep clustering result
yields a relative word error rate reduction of 26 % over the unsupervised
teacher. | ['Reinhold Haeb-Umbach', 'Daniel Hasenklever', 'Lukas Drude'] | 2019-04-02 | null | null | null | null | ['unsupervised-spatial-clustering'] | ['time-series'] | [-8.28108639e-02 -4.24749628e-02 2.78645217e-01 -4.17551130e-01
-1.35234463e+00 -5.66835701e-01 5.80239415e-01 2.20683470e-01
-6.27019048e-01 2.60786593e-01 1.06983908e-01 -4.14391726e-01
-2.66125321e-01 -3.78598332e-01 -7.89729178e-01 -1.19672704e+00
-9.11073089e-02 5.76548636e-01 -6.87512010e-02 1.15890712e-01
1.08238377e-01 5.08651972e-01 -1.25407290e+00 7.31552839e-02
8.68002594e-01 5.50949335e-01 3.84120405e-01 8.66245270e-01
-5.97763062e-02 6.00557387e-01 -1.02463758e+00 5.51741868e-02
3.02077264e-01 -7.23335385e-01 -9.31032300e-01 -1.55037344e-01
2.11285800e-01 -2.46135771e-01 -4.55603063e-01 1.17359221e+00
6.90118730e-01 3.85271400e-01 5.53104699e-01 -9.10162270e-01
-4.50937748e-01 1.15639436e+00 -4.63025510e-01 3.29286009e-01
-5.27951121e-01 -2.29680464e-02 6.12062693e-01 -8.51845622e-01
3.55958521e-01 8.24476123e-01 6.56200349e-01 5.90889573e-01
-1.28105295e+00 -6.86299562e-01 7.06341788e-02 -8.21264908e-02
-1.77615714e+00 -8.68149161e-01 6.50420725e-01 -2.07376108e-01
1.13300967e+00 -1.47359744e-01 2.09482744e-01 1.12428868e+00
-1.41568571e-01 6.10613585e-01 9.55554605e-01 -6.16634309e-01
6.41483426e-01 -7.26730004e-02 4.41088378e-01 3.54248375e-01
2.49235108e-01 -1.86396971e-01 -4.22218829e-01 2.18316004e-01
5.24805188e-01 -2.75014460e-01 3.41382995e-02 -1.41989112e-01
-1.12015843e+00 8.84198070e-01 6.88110650e-01 7.55961537e-01
-2.17605084e-01 5.67154586e-01 3.31294298e-01 1.50002822e-01
3.21455747e-01 6.26149833e-01 -3.80019188e-01 -1.33126557e-01
-1.59677505e+00 -3.96457464e-01 5.75890958e-01 1.12185371e+00
7.03270376e-01 4.21828330e-01 2.33471170e-01 6.59323812e-01
1.00764140e-01 6.87077522e-01 5.83154261e-01 -1.06745672e+00
6.98678195e-02 -2.88743768e-02 -3.62932205e-01 -7.73519456e-01
-6.29480839e-01 -6.58568442e-01 -1.20019591e+00 5.22600450e-02
4.62510198e-01 -5.73131144e-01 -1.23258698e+00 1.79602432e+00
-1.06269561e-01 2.07189068e-01 2.41600007e-01 8.22864652e-01
7.29567587e-01 7.00330079e-01 -5.17329350e-02 -7.43547305e-02
8.24036062e-01 -8.73032928e-01 -8.57708693e-01 -7.55741820e-02
6.97571933e-01 -6.29549444e-01 6.67293966e-01 6.99254513e-01
-9.77235436e-01 -6.46272540e-01 -1.14136791e+00 6.39242679e-02
-5.57442844e-01 3.31288964e-01 6.98664606e-01 8.35708439e-01
-1.46419597e+00 6.49408162e-01 -1.31046438e+00 -5.58391392e-01
5.04546523e-01 6.58921480e-01 -3.26003462e-01 4.14770283e-02
-8.76338661e-01 5.49588561e-01 5.13568044e-01 1.40606567e-01
-1.35751450e+00 -3.65296632e-01 -6.68079674e-01 1.77656770e-01
-1.90006699e-02 -2.03896493e-01 8.95262420e-01 -9.91023183e-01
-1.51230943e+00 5.24450600e-01 -2.62672931e-01 -4.68450904e-01
-8.87985528e-02 -1.21562481e-01 -3.92114729e-01 5.72254360e-01
2.38579571e-01 9.50082064e-01 8.19290042e-01 -1.40427887e+00
-5.55722952e-01 -2.45609760e-01 -4.16288048e-01 1.06018797e-01
-4.67097610e-01 2.49971114e-02 -7.36123323e-01 -7.12851226e-01
1.32651433e-01 -8.18799615e-01 -4.45405781e-01 -7.70953536e-01
-7.62744069e-01 -6.30137622e-02 5.40879428e-01 -5.83023608e-01
8.99490833e-01 -2.28635406e+00 1.96046054e-01 7.08050132e-01
3.20033789e-01 1.90990344e-01 -1.81138173e-01 2.45126158e-01
-2.21388459e-01 1.80846974e-01 -3.96084309e-01 -6.13932073e-01
1.92407425e-02 6.16713613e-02 -1.53129190e-01 7.41678715e-01
4.32135053e-02 7.53904462e-01 -8.35972250e-01 -2.38469869e-01
1.77086517e-01 7.08245814e-01 -4.53434736e-01 2.31802359e-01
2.88726002e-01 5.38219631e-01 1.56118125e-01 2.18187898e-01
7.28808641e-01 -3.57226670e-01 5.36995828e-01 1.14211347e-02
5.64257354e-02 4.35587347e-01 -1.16170502e+00 2.27474236e+00
-3.14731300e-01 1.06263459e+00 5.84131479e-01 -1.25042951e+00
6.62558258e-01 3.93206894e-01 4.45224166e-01 -6.21449351e-01
2.22998828e-01 1.29837766e-01 1.70746848e-01 -1.03336193e-01
3.29383075e-01 -1.71613157e-01 -8.82034823e-02 7.68144369e-01
6.07876658e-01 1.94026619e-01 7.90813658e-03 6.49046183e-01
1.34893847e+00 -3.42571199e-01 -5.18894374e-01 -8.07069540e-01
6.67453930e-02 1.13197021e-01 1.51669651e-01 1.13077474e+00
-1.78229615e-01 6.98708713e-01 1.71882316e-01 -1.39006162e-02
-1.16153204e+00 -1.18455589e+00 -1.13237850e-01 1.04600608e+00
9.13701281e-02 -5.53553104e-01 -1.25530601e+00 -6.28221154e-01
-5.16238332e-01 3.88914615e-01 -4.91113603e-01 -7.27538317e-02
-5.89643717e-01 -1.02673304e+00 1.09267807e+00 6.84174120e-01
4.54562902e-01 -8.55630994e-01 -2.68342674e-01 -3.38071138e-02
-1.45568714e-01 -1.01541221e+00 -7.81048015e-02 1.12628758e+00
-7.70287275e-01 -9.47273433e-01 -4.63748693e-01 -1.05105233e+00
7.63372183e-01 4.15273398e-01 9.03777003e-01 3.23354632e-01
8.83773435e-03 2.51852721e-01 -3.10917258e-01 -7.68936798e-02
-3.14133614e-01 6.29613817e-01 2.52819777e-01 -4.68163490e-01
7.59465516e-01 -7.74795055e-01 -6.11526906e-01 -2.70285487e-01
-1.05250740e+00 -2.62466103e-01 6.42106175e-01 5.93795240e-01
2.52333611e-01 7.18390822e-01 7.01128244e-01 -7.60720611e-01
5.52071035e-01 -5.70517480e-01 -3.86497617e-01 -2.29599774e-01
-6.52103662e-01 1.00897841e-01 8.17968667e-01 -2.01053113e-01
-8.42267454e-01 4.10135746e-01 -5.46030521e-01 -4.01612282e-01
-8.61535251e-01 3.31624120e-01 -1.16807312e-01 2.39173993e-01
8.37008476e-01 2.43390769e-01 -1.46939546e-01 -6.61450565e-01
7.19498217e-01 8.12305868e-01 7.42818773e-01 -5.88306248e-01
9.30325747e-01 7.11718678e-01 -3.82647425e-01 -9.42090869e-01
-4.83704567e-01 -7.45308220e-01 -1.35971677e+00 1.09568380e-01
1.22808945e+00 -1.24285793e+00 -4.92363989e-01 3.31587732e-01
-1.04185736e+00 -8.03941548e-01 -1.85451567e-01 5.41327357e-01
-2.59620041e-01 1.02416828e-01 -8.74459803e-01 -5.97698450e-01
-2.15845346e-01 -1.23359931e+00 1.13942289e+00 7.34908739e-03
-2.74812747e-02 -1.22990000e+00 -7.94346929e-02 -6.31797090e-02
4.39642906e-01 -1.51571542e-01 6.83400869e-01 -9.25861359e-01
-3.97367805e-01 2.15204909e-01 -1.41867340e-01 3.24069113e-01
1.34688169e-01 -1.09856918e-01 -1.39601934e+00 -3.71430099e-01
2.69491613e-01 -2.68662065e-01 1.03156960e+00 6.02819502e-01
1.15313315e+00 2.10028086e-02 -3.90266895e-01 7.90797770e-01
1.58673894e+00 9.70982686e-02 5.50274432e-01 1.39167652e-01
1.04714739e+00 3.28940809e-01 -2.26529971e-01 -2.21581217e-02
2.19915658e-01 4.26252514e-01 5.56279980e-02 -3.87832761e-01
-3.95058244e-01 -1.22431424e-02 3.72426480e-01 1.29857707e+00
5.58154047e-01 3.70553918e-02 -1.13271654e+00 9.49073851e-01
-1.68887389e+00 -7.95161903e-01 -4.06112611e-01 1.78350604e+00
9.26738560e-01 6.84507862e-02 1.69561908e-01 2.93568015e-01
6.92791104e-01 -1.55211300e-01 -3.49533148e-02 -2.36392424e-01
-1.06880315e-01 5.09040594e-01 6.19627476e-01 7.25355327e-01
-1.30124509e+00 1.19351614e+00 7.89550781e+00 1.13369572e+00
-1.20710874e+00 3.72106552e-01 5.57519853e-01 -1.37141138e-01
-2.29919240e-01 -1.81599915e-01 -6.46021068e-01 4.11715388e-01
1.55663884e+00 5.63679814e-01 5.91969311e-01 5.85014045e-01
2.64406931e-02 -4.97928411e-02 -1.12243283e+00 1.00518060e+00
2.32931137e-01 -1.17702353e+00 -1.70783609e-01 -5.25332522e-03
8.11036587e-01 4.26549762e-01 4.99559231e-02 1.40405193e-01
7.52048254e-01 -1.43581426e+00 6.22709155e-01 4.11098190e-02
6.44979358e-01 -1.02267945e+00 5.79640985e-01 4.69039470e-01
-8.24464798e-01 2.20288962e-01 -3.84838134e-01 -2.30726730e-02
-1.15508452e-01 8.67113769e-01 -6.28521800e-01 3.95083755e-01
8.89223754e-01 4.38494414e-01 -7.51288056e-01 7.86575794e-01
-3.49317193e-01 1.11501503e+00 -5.88385344e-01 3.21085632e-01
5.60566127e-01 -1.86587811e-01 1.61350220e-01 1.82217586e+00
1.55297607e-01 1.98497381e-02 -1.37529328e-01 8.71507883e-01
-2.01502144e-01 -2.14808390e-01 -6.39565527e-01 1.38370067e-01
5.46176553e-01 1.31100619e+00 -1.48480999e+00 -5.16467571e-01
-1.62988842e-01 1.14702177e+00 4.88830835e-01 5.99470913e-01
-6.95549130e-01 -6.68291330e-01 5.22362530e-01 -4.90819842e-01
7.79060781e-01 -6.49722159e-01 -7.55193293e-01 -1.00213623e+00
-4.88510311e-01 -7.99148977e-01 7.44087771e-02 -6.13622427e-01
-1.05758977e+00 6.84049010e-01 -2.57269144e-01 -5.34972787e-01
-5.39970305e-03 -5.23982584e-01 -5.25973141e-01 8.69954228e-01
-1.30740714e+00 -7.72481680e-01 -9.66914073e-02 8.25480878e-01
3.58672351e-01 -2.51631856e-01 1.03124022e+00 6.51939392e-01
-8.25927734e-01 7.20844626e-01 6.19265020e-01 6.76900983e-01
6.51668072e-01 -1.60806668e+00 3.53637189e-01 1.39611602e+00
7.44791567e-01 1.03577256e+00 5.87745249e-01 -4.38595653e-01
-1.30816770e+00 -1.24510849e+00 4.91246134e-01 -8.07161868e-01
4.59352762e-01 -7.96644211e-01 -7.57173359e-01 6.41566396e-01
7.76559174e-01 -2.48511791e-01 9.42237377e-01 3.48766029e-01
-4.21833783e-01 -1.94624718e-02 -7.91571081e-01 3.03579241e-01
6.14917397e-01 -7.56415963e-01 -4.21674103e-01 4.40189928e-01
6.58701360e-01 -1.98436737e-01 -8.03578854e-01 -5.75072132e-02
-9.15884376e-02 -9.54659998e-01 8.23396385e-01 -2.96535283e-01
2.64682472e-01 -4.28287834e-01 -1.40310556e-01 -1.47560024e+00
-4.54656273e-01 -6.61425591e-01 8.32305998e-02 1.23213184e+00
4.84882653e-01 1.54168934e-01 8.08877051e-01 2.85519183e-01
-2.98093438e-01 -8.56404528e-02 -8.30135345e-01 -7.28049994e-01
6.45180583e-01 -7.01639235e-01 2.64233947e-01 1.15289402e+00
6.22251965e-02 5.58981478e-01 -4.75881658e-02 5.68659782e-01
6.08852088e-01 -2.29537368e-01 7.27003098e-01 -9.30604815e-01
-1.63581178e-01 -4.71392334e-01 5.38920090e-02 -1.28306758e+00
4.37642395e-01 -1.06533504e+00 6.66168094e-01 -1.45433879e+00
1.61605090e-01 -6.68387115e-01 -4.99204874e-01 3.77321959e-01
-8.42225701e-02 5.85729420e-01 3.78489420e-02 2.89684832e-01
-9.79441643e-01 3.22180569e-01 4.07817543e-01 -1.98548377e-01
-1.79363728e-01 -5.14219284e-01 -1.03768766e+00 5.26400924e-01
8.00773323e-01 -7.72624075e-01 -4.02448267e-01 -9.07165051e-01
3.17934602e-02 -4.98401940e-01 1.18595637e-01 -1.18889487e+00
6.96486413e-01 3.95157695e-01 4.86440063e-01 -4.40614700e-01
8.61744583e-02 -9.94678378e-01 -2.23659128e-01 2.66043276e-01
-4.10771370e-01 -2.27329567e-01 4.60120589e-01 5.17108381e-01
-1.52567029e-01 -1.14083514e-01 7.88619637e-01 6.66807685e-03
-5.40060580e-01 -6.05340265e-02 -8.67661059e-01 -5.32664992e-02
5.56313813e-01 9.75411236e-02 -2.25300252e-01 -3.79674822e-01
-8.05209577e-01 -2.34597847e-02 4.67804283e-01 4.74292934e-02
3.93762410e-01 -1.27355492e+00 -4.34795290e-01 3.04537147e-01
-3.21651220e-01 2.45407462e-01 -8.87691602e-02 8.75314176e-01
-6.14134610e-01 6.06478930e-01 1.50981724e-01 -7.68418074e-01
-7.56845772e-01 7.82246411e-01 4.20161754e-01 1.39991015e-01
-8.24159324e-01 1.07826197e+00 5.72655164e-02 -4.75942463e-01
6.19059205e-01 -2.21290499e-01 7.56860822e-02 -1.71830058e-01
5.40817916e-01 1.35332003e-01 5.66094041e-01 -6.06990516e-01
-4.70556706e-01 3.52240175e-01 1.40117332e-01 -4.04912531e-01
1.59493685e+00 -2.68195152e-01 -1.40455350e-01 3.73301148e-01
1.32349896e+00 1.25609666e-01 -1.06424201e+00 -6.29421175e-02
-1.05512729e-02 3.47503759e-02 3.51515383e-01 -7.28769124e-01
-1.19822717e+00 1.26048946e+00 6.72052562e-01 4.20570076e-01
1.21213353e+00 5.77304177e-02 6.57005489e-01 6.76319778e-01
1.96255185e-02 -1.32917380e+00 -3.31243165e-02 5.68208039e-01
1.21829055e-01 -1.28494895e+00 -2.71482050e-01 5.86054334e-03
-4.78557676e-01 9.95253265e-01 3.54994476e-01 -4.21300411e-01
1.07351637e+00 7.50975609e-01 3.85601014e-01 -6.00380778e-01
-4.37714845e-01 -4.01420146e-01 4.01565880e-02 8.85623753e-01
4.14313108e-01 -8.62499550e-02 3.83337587e-01 5.55998981e-01
-2.83726990e-01 -5.58055341e-01 4.90692317e-01 5.58481514e-01
-4.43705320e-01 -9.09779310e-01 -4.64314163e-01 1.93115339e-01
-5.70821166e-01 -5.18947542e-01 -4.03859466e-01 6.25241876e-01
5.46261221e-02 1.36362231e+00 3.08412135e-01 -5.02748609e-01
-2.41811290e-01 1.62845045e-01 3.01229000e-01 -8.35909784e-01
-6.79080129e-01 5.46444654e-01 -3.25250208e-01 -4.68925864e-01
-6.23872399e-01 -3.57712358e-01 -1.61638176e+00 -3.81015599e-01
-3.12304407e-01 5.84220827e-01 8.63040388e-01 1.21585369e+00
3.73023599e-01 6.91267490e-01 4.38786238e-01 -8.73970866e-01
-8.75691473e-02 -1.03782666e+00 -7.09638357e-01 4.30342793e-01
5.30007899e-01 -1.62960529e-01 -4.07246768e-01 2.81864703e-01] | [9.164896011352539, 3.2286343574523926] |
5b089d3f-a9e7-4203-b461-ce95700d331c | radar-de-parite-an-nlp-system-to-measure | 2304.09982 | null | https://arxiv.org/abs/2304.09982v1 | https://arxiv.org/pdf/2304.09982v1.pdf | Radar de Parité: An NLP system to measure gender representation in French news stories | We present the Radar de Parit\'e, an automated Natural Language Processing (NLP) system that measures the proportion of women and men quoted daily in six Canadian French-language media outlets. We outline the system's architecture and detail the challenges we overcame to address French-specific issues, in particular regarding coreference resolution, a new contribution to the NLP literature on French. We also showcase statistics covering over one year's worth of data (282,512 news articles). Our results highlight the underrepresentation of women in news stories, while also illustrating the application of modern NLP methods to measure gender representation and address societal issues. | ['Maite Taboada', 'Philipp Eibl', 'Prashanth Rao', 'Valentin-Gabriel Soumah'] | 2023-04-19 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [-1.90228820e-02 8.17171514e-01 -8.86443555e-01 -5.36459446e-01
-1.27302015e+00 -8.98473263e-01 1.02993977e+00 8.75692129e-01
-8.68408978e-01 1.19918931e+00 1.32125342e+00 -1.74723163e-01
-2.24321261e-01 -9.10587609e-01 -4.17997122e-01 2.10843980e-03
2.91706890e-01 8.69567156e-01 -2.57143617e-01 -6.67182386e-01
4.90984052e-01 2.29718223e-01 -1.05630350e+00 6.46459341e-01
7.53523111e-01 6.58383310e-01 -5.14539301e-01 5.15542507e-01
-4.93577510e-01 9.91814375e-01 -5.94468296e-01 -1.13657975e+00
-4.01085541e-02 -8.37526098e-02 -9.29141581e-01 -6.70021892e-01
8.28777015e-01 1.72505170e-01 -3.70676458e-01 1.10546696e+00
5.15814900e-01 -6.98410124e-02 5.23185194e-01 -6.77109122e-01
-3.88230860e-01 1.49849975e+00 -8.34540188e-01 7.39360631e-01
1.22311962e+00 -7.45419979e-01 1.27125001e+00 -6.29682541e-01
1.29675174e+00 1.87093556e+00 8.08689117e-01 3.08383733e-01
-1.05866659e+00 -8.76200259e-01 -6.65311217e-02 -3.69151115e-01
-1.54390681e+00 -7.89252818e-01 1.40377939e-01 -4.20203984e-01
6.07241750e-01 5.87076604e-01 6.00666344e-01 1.19038641e+00
3.58619899e-01 5.05761027e-01 1.05621052e+00 -4.49411273e-01
-5.58762789e-01 2.61882305e-01 2.32632250e-01 4.96113628e-01
2.98569888e-01 -2.67852455e-01 -9.52235758e-01 -7.43280351e-01
3.23327214e-01 -6.29980385e-01 2.03121677e-01 5.86511672e-01
-1.37043130e+00 1.12974668e+00 -1.21949062e-01 5.45554161e-01
-7.70001635e-02 -2.83707917e-01 6.92509592e-01 2.97280431e-01
8.27678621e-01 6.77563727e-01 -4.85113770e-01 -5.29877126e-01
-1.06738436e+00 7.67664433e-01 1.32844448e+00 9.79084313e-01
2.15206258e-02 -5.84681630e-01 -2.55343914e-01 1.04234695e+00
1.52023599e-01 7.44000196e-01 -1.73118278e-01 -1.06942165e+00
8.61692548e-01 5.17209828e-01 3.38236570e-01 -1.74902308e+00
-5.82497895e-01 1.40914600e-02 -4.06599253e-01 -9.04365063e-01
8.08835328e-01 -3.73551875e-01 6.87846765e-02 1.62076080e+00
-1.24958135e-01 -7.75820851e-01 -1.67049661e-01 5.65170467e-01
1.28314030e+00 5.37610412e-01 3.53862077e-01 -7.17025995e-01
1.87188232e+00 -1.30038902e-01 -8.90610278e-01 -3.09540182e-01
1.59955963e-01 -1.28670537e+00 6.33432329e-01 2.14644223e-01
-1.35105169e+00 -1.10244595e-01 -4.88594443e-01 -2.87341505e-01
-4.31659400e-01 1.26789406e-01 7.95923114e-01 9.53290939e-01
-3.82564813e-01 3.91915917e-01 -2.04380706e-01 -8.93482029e-01
3.10069859e-01 2.66108215e-01 -2.88866639e-01 3.10017496e-01
-1.34213638e+00 8.29415977e-01 1.93889260e-01 -3.86863053e-01
1.54598683e-01 -7.17289805e-01 -7.32668459e-01 -3.60458612e-01
7.51436770e-01 -6.33779317e-02 1.27808106e+00 -6.09625518e-01
-7.83345997e-01 1.62376785e+00 -4.56494004e-01 -2.89825261e-01
5.59480548e-01 -4.43922490e-01 -1.23692369e+00 7.51844123e-02
7.19414651e-01 5.22338748e-01 3.00192796e-02 -6.15499437e-01
-1.06480682e+00 -2.00631663e-01 5.70358662e-03 3.63187157e-02
-1.66289124e-03 1.06948340e+00 -7.48633146e-02 -8.13522041e-01
-2.80463677e-02 -6.55788600e-01 2.25168273e-01 -5.90577066e-01
-4.32875484e-01 -5.44748127e-01 1.07419007e-01 -9.49664950e-01
1.82321191e+00 -2.12846303e+00 -2.22897664e-01 4.25150722e-01
5.23286700e-01 -2.72819906e-01 2.68552512e-01 7.99930513e-01
2.88007528e-01 4.25469309e-01 1.45191088e-01 2.46502280e-01
2.56626368e-01 8.38426575e-02 -5.32445073e-01 6.27381384e-01
-3.75712901e-01 8.97161424e-01 -1.10042405e+00 -7.23959208e-01
-3.50639999e-01 7.66896009e-02 -4.86568362e-01 -8.59309673e-01
-6.48450181e-02 2.47307569e-01 -2.58239955e-01 9.03788686e-01
5.90234339e-01 3.48788381e-01 6.23273432e-01 -1.46613732e-01
-8.33644569e-01 6.75830126e-01 -7.46821404e-01 1.26935828e+00
2.15722814e-01 8.15248609e-01 5.58690548e-01 -2.83971250e-01
1.01329494e+00 -4.26453277e-02 3.70471686e-01 -6.02397978e-01
5.88820875e-01 4.31281507e-01 -1.98751446e-02 -1.58171400e-01
9.92061853e-01 -8.92155543e-02 -1.28352022e+00 4.16913569e-01
-1.41822055e-01 -6.53493777e-02 6.16175294e-01 4.48114157e-01
6.00437462e-01 -3.27867478e-01 7.38834083e-01 -1.02928162e+00
7.04335332e-01 2.38548949e-01 7.78192759e-01 8.97862017e-01
-4.29798663e-01 2.08815023e-01 1.12772262e+00 -6.42596602e-01
-7.80537486e-01 -9.17833745e-01 -5.13359785e-01 1.58662713e+00
-2.05059052e-01 -1.21487761e+00 -4.24651802e-01 -5.82218707e-01
1.78030193e-01 8.03534567e-01 -7.58470953e-01 3.36318433e-01
-7.73231626e-01 -9.50510383e-01 8.56966734e-01 2.95645483e-02
3.43343109e-01 -7.19260335e-01 -3.05786222e-01 5.36087938e-02
-9.98026848e-01 -1.35110235e+00 -4.75994140e-01 -3.38136882e-01
-3.26540470e-02 -1.36197722e+00 -2.31690198e-01 -8.23252678e-01
2.02451259e-01 -5.30978680e-01 1.58293152e+00 -4.61132526e-01
-8.00344497e-02 3.28170806e-01 -1.55177429e-01 -7.92247891e-01
-6.70733929e-01 4.49530929e-01 3.55647504e-01 -4.92206275e-01
1.27689159e+00 -8.80447999e-02 -3.34266536e-02 -1.96830854e-02
-8.65882263e-02 -4.93546307e-01 1.65614430e-02 4.35742408e-01
6.64772987e-02 -3.63276452e-01 5.31506896e-01 -1.64069808e+00
1.06341064e+00 -5.63512087e-01 -2.43118346e-01 5.32595590e-02
-1.38217315e-01 -4.85429257e-01 5.27310260e-02 6.71908632e-02
-1.09290826e+00 -6.46411717e-01 -2.82394856e-01 8.56361568e-01
1.67173922e-01 6.81007087e-01 2.40091845e-01 1.03752024e-01
9.71127450e-01 -4.77034688e-01 -7.47320578e-02 -3.66069049e-01
2.24602059e-01 9.09150481e-01 9.23508584e-01 -7.37055898e-01
3.13691914e-01 3.44496906e-01 -5.31585574e-01 -1.01661742e+00
-1.14901531e+00 -3.82569045e-01 -6.66826904e-01 -2.03946605e-01
7.13776886e-01 -1.03877652e+00 -1.30423045e+00 -1.88748911e-01
-9.19911087e-01 3.17490965e-01 -2.43971169e-01 5.40389419e-01
-3.64648461e-01 -8.14130232e-02 -9.76428390e-01 -8.24950218e-01
-3.15876931e-01 -2.36353233e-01 7.48690546e-01 9.52271093e-03
-1.18890667e+00 -6.07703447e-01 2.24503189e-01 4.02401060e-01
-7.14678224e-03 4.99702394e-01 8.63940835e-01 -8.57182980e-01
3.44040304e-01 -6.25035614e-02 -4.94187027e-01 -5.29991090e-01
-1.39242545e-01 9.33444872e-02 -2.99729675e-01 1.87591091e-02
-3.41686606e-01 -3.00024539e-01 6.07303143e-01 3.22492719e-01
3.51854026e-01 -5.62179387e-01 -5.64531982e-01 -1.96212351e-01
1.01478446e+00 -1.47191003e-01 4.11410093e-01 3.94793034e-01
1.76336929e-01 7.94423580e-01 8.48393500e-01 6.07651412e-01
5.88533223e-01 3.98849875e-01 -4.44559813e-01 5.23799360e-01
2.21174687e-01 -6.98629200e-01 2.45977864e-01 8.62640798e-01
-4.53023523e-01 1.38211578e-01 -1.03926611e+00 5.38599610e-01
-1.73999953e+00 -1.14431691e+00 -1.32106781e-01 1.61043155e+00
9.21831191e-01 3.75749230e-01 4.48467016e-01 -4.38183956e-02
8.71508718e-01 4.96056437e-01 2.40756512e-01 -8.15272212e-01
-4.73694503e-01 9.42581892e-02 7.75863409e-01 5.48570693e-01
-1.35330248e+00 1.12791765e+00 8.48658943e+00 8.28659236e-01
-3.73828113e-01 -3.76703404e-02 4.68918025e-01 -2.77852893e-01
-2.65083075e-01 -3.00586611e-01 -1.07900131e+00 3.08759689e-01
9.84167635e-01 -7.26904631e-01 3.11506212e-01 2.48067752e-01
2.36642689e-01 -3.29336822e-01 -6.87274098e-01 9.33372974e-01
3.35060745e-01 -1.50454271e+00 -6.09723032e-02 6.10177815e-02
5.28010666e-01 1.01700410e-01 -1.79250151e-01 5.45062959e-01
4.23736781e-01 -8.60282004e-01 1.05781209e+00 5.41481793e-01
1.15054202e+00 -1.13324296e+00 6.63181484e-01 2.73641229e-01
-8.98681760e-01 -3.22455078e-01 -4.43219036e-01 -3.73324394e-01
5.14044762e-01 7.46368825e-01 -4.70294148e-01 3.41046453e-01
8.31764221e-01 6.86532140e-01 -3.60862434e-01 3.26649696e-01
-2.10189462e-01 3.59509587e-01 -3.82391155e-01 -2.96115845e-01
2.24427328e-01 -1.89826995e-01 9.32672441e-01 1.56317806e+00
2.32904673e-01 6.01928353e-01 9.97020826e-02 4.69784379e-01
-3.52217972e-01 6.24420524e-01 -4.19396430e-01 -6.98929608e-01
8.79383504e-01 1.18994665e+00 -9.30727243e-01 -1.20050251e-01
-6.19830489e-01 2.49098450e-01 3.78690749e-01 -1.15692295e-01
-6.11556113e-01 -3.45515966e-01 3.79386783e-01 5.42225063e-01
-4.05677408e-01 -2.00995132e-01 -1.45510644e-01 -1.03157580e+00
-6.57406271e-01 -1.14978671e+00 8.97197187e-01 -3.63594681e-01
-1.29582357e+00 2.54589766e-01 2.14495614e-01 -4.37055528e-01
-1.98003724e-01 -3.71520340e-01 -2.61090044e-02 4.77127343e-01
-5.27197063e-01 -9.07596350e-01 3.32707256e-01 2.16188386e-01
1.72317699e-01 -5.46423256e-01 9.39268470e-01 5.72931468e-01
-1.59847796e-01 7.12387502e-01 -3.14411610e-01 1.41133785e-01
1.29341364e+00 -1.11317503e+00 3.87446225e-01 2.77412295e-01
6.74259063e-05 7.11571991e-01 9.45443273e-01 -1.01118565e+00
-9.25679684e-01 -3.61591548e-01 2.07497978e+00 -8.03308547e-01
1.02369392e+00 -6.53555691e-01 2.48401780e-02 9.84614134e-01
4.37452942e-01 -9.25358593e-01 1.00733864e+00 7.99937129e-01
-2.89644629e-01 2.04263464e-01 -1.44220746e+00 7.07274377e-01
1.28467464e+00 -7.45508611e-01 -1.09211743e+00 3.18722755e-01
3.91589373e-01 -5.98232388e-01 -1.04456854e+00 4.45194870e-01
9.45314705e-01 -7.08164811e-01 7.93293476e-01 -6.39532626e-01
6.21484578e-01 2.86864340e-01 -2.87584424e-01 -8.06537807e-01
-6.13647819e-01 -9.22953427e-01 2.98708797e-01 1.44556737e+00
6.63031578e-01 -4.83927697e-01 5.77187777e-01 5.27147055e-01
3.32073152e-01 -2.55854011e-01 -8.39675188e-01 1.36265785e-01
2.63602167e-01 -4.02959079e-01 4.32639450e-01 1.29790235e+00
6.24285638e-01 4.36092138e-01 -3.99623662e-01 -4.78873938e-01
1.12026751e-01 -4.57389839e-03 6.26238108e-01 -1.43552148e+00
2.23408848e-01 -3.58178973e-01 -1.22701019e-01 -5.39901316e-01
2.69794673e-01 -8.83211374e-01 -5.29623687e-01 -1.13690412e+00
4.68830854e-01 1.01499200e-01 3.65637064e-01 1.35341808e-01
4.52879459e-01 4.95956630e-01 2.79652625e-01 1.62278250e-01
-6.02833211e-01 -1.48553893e-01 8.67769480e-01 -5.10818884e-02
-1.38637409e-01 -6.37483075e-02 -1.59427404e+00 1.10907376e+00
4.21382010e-01 -4.83001977e-01 2.77267277e-01 -3.38063315e-02
1.16601276e+00 -1.82479188e-01 -2.13089243e-01 -4.83795673e-01
2.22258538e-01 -3.33673686e-01 5.20100772e-01 -8.16510499e-01
1.74743682e-02 -2.90828079e-01 1.86166074e-02 3.06275219e-01
-4.31552142e-01 2.23077103e-01 3.73828143e-01 2.32299253e-01
-3.36615950e-01 -7.71662220e-02 4.21225965e-01 -3.76892298e-01
-3.63665938e-01 -1.26354411e-01 -9.00560617e-01 5.83603382e-01
7.18474090e-01 2.98893124e-01 -6.82818472e-01 -4.46564794e-01
-1.02328050e+00 2.80746222e-01 1.77850619e-01 5.79930842e-01
-3.78398523e-02 -1.42906106e+00 -1.27925658e+00 5.10046817e-02
1.82961538e-01 -1.01839101e+00 -4.02506813e-02 8.63138676e-01
-6.40992045e-01 5.07442236e-01 -1.06123976e-01 1.80904850e-01
-1.27211940e+00 2.93184757e-01 -1.97805822e-01 -9.15010571e-02
-4.69535649e-01 8.08759391e-01 -2.94983119e-01 -7.36380577e-01
-3.21729109e-02 1.70031279e-01 -7.65889049e-01 9.13167238e-01
7.56773353e-01 7.32545435e-01 -4.66718256e-01 -1.18784726e+00
-7.61271298e-01 3.11515212e-01 -5.93394451e-02 -4.51209992e-01
1.17590392e+00 -6.84924603e-01 -6.87987924e-01 7.45086432e-01
8.58197927e-01 1.34244037e+00 -1.22646894e-02 -1.05560780e-01
3.54381502e-01 -5.00045896e-01 -4.27396566e-01 -9.19901967e-01
-4.88802373e-01 -2.14567810e-01 -2.82897770e-01 3.59273672e-01
4.65385109e-01 3.70242238e-01 7.18950272e-01 2.36293465e-01
4.57078248e-01 -1.68136299e+00 -8.14056277e-01 8.64563406e-01
1.01072454e+00 -8.03248167e-01 4.77778405e-01 -9.58444834e-01
-5.23412049e-01 1.06989527e+00 8.16311836e-02 2.26598065e-02
8.91959727e-01 3.35343182e-01 1.68446139e-01 -6.19760334e-01
-3.99837375e-01 2.48620972e-01 2.10751459e-01 1.18372954e-01
9.83585656e-01 3.08120698e-01 -1.57036662e+00 1.40806139e+00
-1.26812565e+00 -3.79082449e-02 5.88291168e-01 6.43368602e-01
-3.57227474e-01 -1.09371865e+00 -4.77199852e-01 9.49537277e-01
-1.20997071e+00 -1.07180618e-01 -8.56068432e-01 9.77440834e-01
2.54561782e-01 1.02004027e+00 3.15523714e-01 -2.54282027e-01
4.18571472e-01 -2.72747539e-02 6.76181078e-01 -5.73858440e-01
-8.49298894e-01 2.53644854e-01 1.16404462e+00 -5.07511258e-01
-8.98255408e-01 -1.26289487e+00 -9.04307485e-01 -1.15825772e+00
1.54163986e-01 4.65789497e-01 -5.63378856e-02 7.55758166e-01
4.33074571e-02 -1.75425503e-02 -9.84735563e-02 -3.26667517e-01
3.04802395e-02 -7.94414341e-01 -1.08955562e+00 -3.80555377e-03
-9.83350873e-02 -5.59689164e-01 -2.84886658e-01 -3.42631012e-01] | [9.194657325744629, 9.975492477416992] |
0a3d498d-fc16-46c2-8a1a-4c7efffeed18 | scalable-deep-graph-clustering-with-random | 2112.15530 | null | https://arxiv.org/abs/2112.15530v2 | https://arxiv.org/pdf/2112.15530v2.pdf | Scalable Deep Graph Clustering with Random-walk based Self-supervised Learning | Web-based interactions can be frequently represented by an attributed graph, and node clustering in such graphs has received much attention lately. Multiple efforts have successfully applied Graph Convolutional Networks (GCN), though with some limits on accuracy as GCNs have been shown to suffer from over-smoothing issues. Though other methods (particularly those based on Laplacian Smoothing) have reported better accuracy, a fundamental limitation of all the work is a lack of scalability. This paper addresses this open problem by relating the Laplacian smoothing to the Generalized PageRank and applying a random-walk based algorithm as a scalable graph filter. This forms the basis for our scalable deep clustering algorithm, RwSL, where through a self-supervised mini-batch training mechanism, we simultaneously optimize a deep neural network for sample-cluster assignment distribution and an autoencoder for a clustering-oriented embedding. Using 6 real-world datasets and 6 clustering metrics, we show that RwSL achieved improved results over several recent baselines. Most notably, we show that RwSL, unlike all other deep clustering frameworks, can continue to scale beyond graphs with more than one million nodes, i.e., handle web-scale. We also demonstrate how RwSL could perform node clustering on a graph with 1.8 billion edges using only a single GPU. | ['Rajiv Ramnath', 'Gagan Agrawal', 'Ruoming Jin', 'Dong Li', 'Xiang Li'] | 2021-12-31 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-1.11901045e-01 2.92937070e-01 -1.46845937e-01 -2.13875309e-01
-6.15595996e-01 -6.23232186e-01 5.86183608e-01 3.71892124e-01
-1.91656202e-01 4.02447283e-01 2.92573243e-01 -3.87186438e-01
-6.09339364e-02 -1.01870668e+00 -7.10663676e-01 -5.96872568e-01
-7.67258108e-01 5.93036830e-01 2.48337209e-01 1.13653652e-01
5.39644770e-02 2.33506456e-01 -9.58956659e-01 1.23298995e-01
7.26626933e-01 4.21036750e-01 -1.08662985e-01 8.75990748e-01
-1.60849243e-01 7.57077634e-01 -3.09763223e-01 -5.28458953e-01
8.02510157e-02 -1.79301634e-01 -8.85375023e-01 2.24215120e-01
6.74641013e-01 -2.02292666e-01 -6.32678330e-01 8.84391963e-01
4.74680185e-01 1.18105240e-01 6.34947777e-01 -1.58156502e+00
-8.13691437e-01 1.01315165e+00 -7.97990322e-01 -2.83119738e-01
1.43863901e-01 6.64552003e-02 1.42052925e+00 -4.28083628e-01
8.64672661e-01 1.35455656e+00 1.17589772e+00 1.75379485e-01
-1.62457740e+00 -4.46572930e-01 2.29096994e-01 1.06601929e-02
-1.40922320e+00 -8.23700149e-03 6.54037178e-01 -3.93418700e-01
1.14148843e+00 -8.04792866e-02 7.39844561e-01 1.04337060e+00
2.06180930e-01 6.35112107e-01 7.32325733e-01 -2.17203632e-01
2.96269417e-01 -3.40933233e-01 2.26938173e-01 9.90408182e-01
5.87300003e-01 -5.13040423e-01 -1.74329758e-01 -2.58625627e-01
6.70660198e-01 -8.63831341e-02 2.07747594e-02 -6.69222295e-01
-9.52639401e-01 1.12980270e+00 9.52995181e-01 2.03606620e-01
-5.83195724e-02 7.88431346e-01 4.14060384e-01 2.26522610e-01
7.04449117e-01 2.55562335e-01 -2.55504876e-01 1.56334206e-01
-9.03000295e-01 -2.03704312e-02 1.28357589e+00 7.86081970e-01
1.01520383e+00 -1.59186140e-01 1.88766688e-01 5.63102067e-01
4.21069801e-01 8.74805003e-02 1.21624917e-01 -9.86517429e-01
2.46597901e-01 9.19200122e-01 -5.69588959e-01 -1.46160984e+00
-6.73869491e-01 -5.65122306e-01 -1.11785698e+00 1.09654181e-01
3.63316357e-01 -2.63921052e-01 -8.58538926e-01 1.73503470e+00
2.43423149e-01 2.11777404e-01 -3.25536340e-01 6.10636175e-01
7.68749774e-01 4.06737834e-01 2.29945779e-02 3.58926266e-01
9.21178758e-01 -1.24094498e+00 -4.24804240e-01 -1.99345633e-01
9.38467979e-01 -4.06887889e-01 1.15466702e+00 2.82584667e-01
-7.98230827e-01 -2.32656188e-02 -1.03440785e+00 -2.01273456e-01
-6.53728902e-01 -2.85533905e-01 1.27454579e+00 6.00870192e-01
-1.86693299e+00 8.39576602e-01 -1.13304937e+00 -1.07635653e+00
5.10723114e-01 5.69820642e-01 -5.72949052e-01 1.20896302e-01
-7.89000273e-01 4.05788749e-01 2.87738174e-01 -2.76205033e-01
-6.15934551e-01 -5.67793310e-01 -8.06051850e-01 4.18405503e-01
3.47742945e-01 -7.38146126e-01 7.26433277e-01 -8.58280718e-01
-1.09836638e+00 7.32059956e-01 7.61478543e-02 -6.39211595e-01
1.96898147e-01 1.18637130e-01 -3.24528337e-01 1.66548863e-01
3.10114566e-02 7.94903576e-01 3.80158424e-01 -1.19979227e+00
-2.29053333e-01 -2.19133437e-01 9.19750333e-02 3.11761461e-02
-8.34894180e-01 -2.76958525e-01 -8.25642943e-01 -5.16976833e-01
6.36926740e-02 -1.02267408e+00 -4.75790679e-01 -4.63752002e-02
-6.78094864e-01 -4.13727760e-01 7.54714370e-01 -4.67331320e-01
1.39148009e+00 -1.86781681e+00 8.15645978e-02 5.57492554e-01
9.75489318e-01 -3.73696387e-02 -4.20436621e-01 8.26875210e-01
-2.50741793e-03 5.69317281e-01 -3.23769748e-01 -2.66839355e-01
1.91118419e-01 6.35035858e-02 3.36592287e-01 5.68262219e-01
1.04412824e-01 1.17642915e+00 -9.67251301e-01 -5.34621835e-01
6.95076063e-02 6.65927410e-01 -8.87002468e-01 -7.71763176e-02
-3.11069787e-01 -2.16229871e-01 -4.28197570e-02 3.28513920e-01
6.86188340e-01 -1.04652941e+00 7.26940453e-01 -1.62862837e-01
2.79765248e-01 1.17493019e-01 -1.08811963e+00 1.71589208e+00
-1.89988539e-01 8.46281171e-01 4.30967838e-01 -9.76856589e-01
5.25979996e-01 -1.19443551e-01 8.22512090e-01 -1.35087490e-01
1.71808228e-02 -3.11032515e-02 1.04624473e-01 -1.43745974e-01
1.56198025e-01 5.65841436e-01 1.32597759e-01 8.44144940e-01
1.76014334e-01 1.06721759e-01 4.52571839e-01 1.09387231e+00
2.00864577e+00 -1.93015814e-01 -7.33073801e-02 -4.80510175e-01
1.04907993e-02 8.95045847e-02 1.40079141e-01 7.35093832e-01
-9.94371548e-02 5.10392785e-01 9.03973520e-01 -4.87139523e-01
-9.52070475e-01 -1.10767591e+00 3.44270080e-01 1.21102595e+00
-6.21259995e-02 -9.61732745e-01 -9.81384397e-01 -9.65570211e-01
2.58737713e-01 2.83063263e-01 -6.35999680e-01 -1.80060565e-02
-4.43256199e-01 -9.84186172e-01 5.11285663e-01 4.39933836e-01
3.74751091e-01 -9.26390290e-01 2.36592919e-01 3.34768742e-01
1.37754023e-01 -1.13140810e+00 -5.91564834e-01 2.81612486e-01
-7.69163311e-01 -1.33973360e+00 -2.55770415e-01 -1.07625818e+00
8.24185073e-01 4.38123256e-01 1.60685813e+00 4.61107314e-01
-4.14914995e-01 5.88612914e-01 -2.41229594e-01 1.52457610e-01
-2.50936806e-01 6.49335265e-01 -1.25306606e-01 -2.10279822e-01
5.12018085e-01 -1.01816463e+00 -7.64759123e-01 -2.14030683e-01
-8.28560233e-01 -6.68962151e-02 6.14524245e-01 6.05618000e-01
1.66923523e-01 1.93273976e-01 4.46812272e-01 -1.42309809e+00
1.01210558e+00 -6.85173869e-01 -4.61386532e-01 1.07433923e-01
-8.58471036e-01 1.66678622e-01 5.07467985e-01 -1.71153516e-01
-4.43730563e-01 1.34518176e-01 -1.69999272e-01 -7.61118159e-02
3.38715538e-02 8.19845140e-01 1.39663860e-01 -3.36151898e-01
6.50055647e-01 -2.30739951e-01 2.56998748e-01 -2.48812988e-01
7.06558228e-01 3.94164324e-01 2.29134440e-01 -3.83947819e-01
9.87941086e-01 5.89757860e-01 1.62836790e-01 -8.11961353e-01
-6.27561748e-01 -5.09891450e-01 -7.33804226e-01 -1.47820652e-01
9.29124236e-01 -9.71305966e-01 -1.04684913e+00 3.04761708e-01
-8.57594848e-01 -8.95990908e-01 2.12853149e-01 2.10684791e-01
-1.97318122e-01 7.52670288e-01 -1.10518920e+00 -4.09379214e-01
-3.36253583e-01 -8.58633578e-01 9.61811423e-01 -1.74844414e-01
-2.75413603e-01 -1.52160311e+00 1.45572037e-01 2.64247376e-02
6.30796373e-01 4.40619022e-01 9.73921061e-01 -7.79545724e-01
-7.47973621e-01 -1.51018143e-01 -5.48068225e-01 -2.34178249e-02
1.18868627e-01 3.47586632e-01 -5.04812181e-01 -6.90526545e-01
-7.82485545e-01 -3.99271190e-01 1.11926854e+00 3.52848530e-01
1.18899655e+00 -3.75521630e-01 -6.74968302e-01 8.40196609e-01
1.57117474e+00 -4.90925670e-01 3.92223805e-01 1.82722211e-01
1.19993949e+00 2.69443959e-01 -3.03488582e-01 1.25189736e-01
8.56606722e-01 3.64458531e-01 5.80353796e-01 -3.85215908e-01
-4.10752863e-01 -2.47959659e-01 3.27996612e-01 1.20477116e+00
1.67810395e-01 -5.03828108e-01 -1.07603824e+00 4.49891925e-01
-1.95151436e+00 -7.48626351e-01 -5.32818735e-01 1.84297729e+00
5.80289304e-01 1.32534534e-01 3.33761573e-01 -1.51958376e-01
8.11270416e-01 3.09137732e-01 -7.41096079e-01 -9.35103670e-02
-1.17143951e-01 1.61080196e-01 7.04822123e-01 7.63489723e-01
-1.39350653e+00 1.13665247e+00 6.74725723e+00 6.62875712e-01
-8.21870863e-01 -7.78192747e-03 6.48080707e-01 8.13914165e-02
-3.03562462e-01 8.62133726e-02 -3.65150094e-01 3.95410568e-01
9.83377039e-01 -1.56075120e-01 9.06971097e-01 8.55908811e-01
-4.72112745e-02 7.56399333e-02 -1.05937028e+00 7.60212839e-01
2.31288094e-02 -1.35645628e+00 -9.66780912e-03 4.64355052e-01
9.05757427e-01 4.96777564e-01 -1.14230178e-01 3.50979805e-01
1.35694158e+00 -1.14483523e+00 -5.94562553e-02 8.95392150e-02
6.71398163e-01 -6.44409597e-01 5.20697713e-01 2.53634080e-02
-1.57016838e+00 1.49155140e-01 -3.37754577e-01 -1.38014285e-02
-8.58870372e-02 8.84111881e-01 -8.51812959e-01 3.25630009e-01
8.31450939e-01 9.80440795e-01 -9.80184734e-01 8.95709217e-01
-1.17344238e-01 8.98660302e-01 -4.94717181e-01 -1.86603338e-01
4.39002812e-01 -3.79098862e-01 2.71783262e-01 1.41478693e+00
1.24344870e-01 -2.29546070e-01 3.75826985e-01 7.43342578e-01
-5.23246408e-01 7.17809200e-02 -6.52364671e-01 -3.91980171e-01
4.56511617e-01 1.70354354e+00 -1.17457926e+00 -2.09593967e-01
-6.30571961e-01 1.11606026e+00 1.01657343e+00 5.50017834e-01
-7.22439408e-01 -5.74341357e-01 6.73955083e-01 3.50970030e-01
3.49680036e-01 -5.88037312e-01 -1.44913778e-01 -1.20565987e+00
-2.75254250e-01 -7.35466123e-01 5.57605207e-01 -4.66015339e-01
-1.76319587e+00 3.88371050e-01 -5.53719223e-01 -4.62706536e-01
4.82893139e-02 -6.43935561e-01 -7.72607684e-01 4.72680181e-01
-9.70811069e-01 -1.23588490e+00 -4.73672956e-01 5.71504712e-01
-2.65860446e-02 2.57090246e-03 7.98159599e-01 3.25204879e-01
-5.09851038e-01 6.15607738e-01 3.28126073e-01 5.25498152e-01
8.41597438e-01 -1.84191501e+00 8.75706613e-01 7.70931840e-01
3.47896814e-01 7.44715452e-01 3.98874253e-01 -9.08247948e-01
-1.70499718e+00 -1.41997123e+00 7.27911115e-01 -3.27603668e-01
1.06258941e+00 -8.54359448e-01 -8.60602140e-01 1.06912625e+00
5.76657355e-01 3.80419828e-02 5.87847114e-01 6.11169934e-01
-4.25374031e-01 -1.11947455e-01 -9.68393326e-01 9.01484191e-01
1.58352661e+00 -4.35506254e-01 2.20427915e-01 6.77375913e-01
8.90745819e-01 2.60023233e-02 -1.04974234e+00 1.42345920e-01
3.80174309e-01 -1.07265508e+00 8.61358285e-01 -5.22110939e-01
2.57410616e-01 -1.90515757e-01 1.63928255e-01 -1.52849829e+00
-8.57427061e-01 -6.96033537e-01 -2.64609993e-01 1.19329798e+00
3.22004408e-01 -6.27308428e-01 1.22790968e+00 3.93651009e-01
6.52423128e-02 -5.46165407e-01 -4.38040406e-01 -6.35850787e-01
1.71311453e-01 -1.21671923e-01 3.90449733e-01 1.14867568e+00
1.39300242e-01 7.06722558e-01 -1.58799943e-02 6.06958605e-02
9.89598811e-01 -1.84425220e-01 1.03647768e+00 -1.51195538e+00
-1.97668433e-01 -7.25028872e-01 -5.05693495e-01 -8.73228729e-01
4.06035393e-01 -1.46457589e+00 -2.33282745e-01 -2.16376567e+00
3.92454654e-01 -2.62579083e-01 -7.25417659e-02 5.74370563e-01
-2.62470663e-01 4.32093680e-01 -5.72070107e-02 2.20803125e-03
-1.18224025e+00 2.97880083e-01 8.43805611e-01 -3.35377514e-01
-1.46604657e-01 -3.96169573e-01 -8.24034810e-01 5.88935614e-01
8.66508365e-01 -2.54028738e-01 -3.89429778e-01 -5.28115809e-01
5.50819755e-01 -4.67193186e-01 2.50350982e-01 -1.18078852e+00
5.97694874e-01 3.75441849e-01 3.44329238e-01 -5.21932542e-01
-3.23941968e-02 -7.10386515e-01 2.37087131e-01 4.20745105e-01
-4.16658074e-01 2.40096346e-01 -1.13720894e-01 9.43183541e-01
3.45625952e-02 3.33336294e-01 5.00919044e-01 -1.30996883e-01
-5.46414375e-01 5.11162460e-01 -4.66791779e-01 1.56461075e-01
8.86345565e-01 7.52118379e-02 -5.16386867e-01 -6.49248540e-01
-8.61503601e-01 4.48073894e-01 6.80547595e-01 7.59691820e-02
1.98151872e-01 -1.34555769e+00 -7.72503793e-01 -4.07766376e-04
-1.89884365e-01 -8.02272037e-02 -8.38757306e-02 7.13102877e-01
-8.19239855e-01 1.14551939e-01 2.70042717e-01 -6.64223015e-01
-1.03774583e+00 7.17839360e-01 1.36576682e-01 -5.67447364e-01
-8.01054597e-01 8.63293052e-01 -5.50419912e-02 -7.58349717e-01
4.92712051e-01 -2.07307078e-02 1.86629444e-01 -1.13133542e-01
4.62052040e-02 3.78106117e-01 9.03503597e-02 -2.65448242e-01
-4.34529871e-01 4.21200454e-01 -1.11042969e-01 1.69956699e-01
1.61628222e+00 -1.03575945e-01 -5.38099110e-01 1.76732689e-01
1.49501288e+00 6.61263838e-02 -1.17062628e+00 -1.90676063e-01
3.19323212e-01 -3.93886026e-03 -2.60917033e-04 -5.56656599e-01
-1.33634198e+00 6.94346130e-01 1.49590671e-01 7.23593831e-01
7.55382717e-01 2.07030267e-01 7.60567963e-01 6.08412087e-01
2.21648216e-01 -9.92161930e-01 2.37366557e-01 4.88000810e-01
5.05043626e-01 -1.47464907e+00 2.67002702e-01 -6.33075178e-01
-3.67118418e-01 1.09399199e+00 6.32486999e-01 -5.53449988e-01
1.13830054e+00 5.23809016e-01 -9.45501551e-02 -5.17625809e-01
-8.78804982e-01 -4.03702319e-01 6.38579857e-03 6.71270490e-01
7.97291100e-01 2.22423881e-01 -1.67205378e-01 2.45470405e-01
-1.57061294e-01 -3.45076740e-01 5.06563842e-01 4.49773908e-01
-3.04875731e-01 -1.11960030e+00 1.89047277e-01 8.34974110e-01
-3.55393797e-01 -3.25595230e-01 -7.86595643e-01 8.83228838e-01
-3.55008751e-01 1.12123322e+00 2.13493869e-01 -6.01440847e-01
-2.21667975e-01 -1.97272643e-01 1.95003897e-01 -6.82306051e-01
-6.78195596e-01 -4.35126312e-02 1.14485987e-01 -8.05341661e-01
-2.63062179e-01 -3.57245147e-01 -1.22274399e+00 -8.67690444e-01
-1.11824624e-01 1.14219956e-01 5.13211608e-01 3.16703975e-01
6.80009604e-01 6.38694704e-01 4.21560884e-01 -9.99307275e-01
-1.85014158e-01 -9.04656827e-01 -7.44853914e-01 6.15886867e-01
-2.60939021e-02 -1.83808103e-01 -7.00217962e-01 -2.81832695e-01] | [6.984063625335693, 6.1527886390686035] |
d5048d48-ad98-4c23-8f1d-605ba5445872 | investigating-inner-properties-of-multimodal | 1711.05516 | null | http://arxiv.org/abs/1711.05516v2 | http://arxiv.org/pdf/1711.05516v2.pdf | Investigating Inner Properties of Multimodal Representation and Semantic Compositionality with Brain-based Componential Semantics | Multimodal models have been proven to outperform text-based approaches on
learning semantic representations. However, it still remains unclear what
properties are encoded in multimodal representations, in what aspects do they
outperform the single-modality representations, and what happened in the
process of semantic compositionality in different input modalities. Considering
that multimodal models are originally motivated by human concept
representations, we assume that correlating multimodal representations with
brain-based semantics would interpret their inner properties to answer the
above questions. To that end, we propose simple interpretation methods based on
brain-based componential semantics. First we investigate the inner properties
of multimodal representations by correlating them with corresponding
brain-based property vectors. Then we map the distributed vector space to the
interpretable brain-based componential space to explore the inner properties of
semantic compositionality. Ultimately, the present paper sheds light on the
fundamental questions of natural language understanding, such as how to
represent the meaning of words and how to combine word meanings into larger
units. | ['Cheng-qing Zong', 'Jiajun Zhang', 'Shaonan Wang', 'Nan Lin'] | 2017-11-15 | null | null | null | null | ['learning-semantic-representations'] | ['methodology'] | [ 3.79280597e-01 3.82824659e-01 -2.74166703e-01 -3.85585755e-01
-1.36618629e-01 -6.12607777e-01 1.04730034e+00 4.64508206e-01
-2.95281887e-01 4.75209266e-01 8.70384455e-01 -2.66136497e-01
-3.68119746e-01 -7.59929717e-01 -4.77521509e-01 -6.72140718e-01
1.63953409e-01 4.88645583e-01 -2.84174502e-01 -4.04167503e-01
4.35553640e-01 6.33351281e-02 -1.62939847e+00 6.64020658e-01
6.82916999e-01 7.75694847e-01 1.01524748e-01 4.34975982e-01
-6.89682424e-01 9.57052886e-01 -2.21934617e-01 -3.53733093e-01
-5.11727154e-01 -7.88069427e-01 -1.14992118e+00 1.72706410e-01
2.19660461e-01 1.52605563e-01 -7.87190646e-02 1.19565213e+00
-2.17169687e-01 2.19717786e-01 1.27769506e+00 -1.35030746e+00
-1.15598285e+00 9.07803893e-01 -1.64415777e-01 -5.57246841e-02
5.52493453e-01 -3.13012242e-01 1.41125011e+00 -8.24460983e-01
6.47539556e-01 1.70991778e+00 1.38915777e-01 7.25763440e-01
-1.52729368e+00 -2.21858714e-02 3.73586506e-01 3.22537869e-01
-1.13729179e+00 -1.51071683e-01 7.70822763e-01 -7.07411230e-01
6.98635757e-01 2.88130492e-01 5.97523749e-01 1.07418644e+00
2.86429599e-02 9.52953517e-01 1.19284308e+00 -5.16619384e-01
2.53857642e-01 1.32468566e-01 6.56175315e-01 7.12884963e-01
3.53107005e-01 -3.74920249e-01 -8.84696066e-01 -1.54006541e-01
7.50814557e-01 1.17963605e-01 -2.19446659e-01 -5.01036048e-01
-1.76075947e+00 1.19652081e+00 4.90386635e-01 7.00161397e-01
-4.79408592e-01 4.98256773e-01 2.96776623e-01 2.04658926e-01
2.12890118e-01 4.64558393e-01 -4.11107749e-01 1.63618594e-01
-4.23862457e-01 -1.43068999e-01 5.46893775e-01 7.15778649e-01
8.30538154e-01 -6.74140602e-02 2.26018772e-01 8.98075461e-01
5.87972283e-01 6.71111405e-01 7.26958752e-01 -8.77521336e-01
2.23256022e-01 7.01239824e-01 -2.40354434e-01 -9.33562458e-01
-4.54116732e-01 1.51826799e-01 -6.33308887e-01 -3.30750465e-01
4.10140038e-01 1.54555246e-01 -6.77839935e-01 2.07884789e+00
-2.34883711e-01 -1.58005178e-01 3.08452070e-01 1.06408453e+00
9.15756047e-01 4.96526569e-01 6.02787912e-01 1.72673687e-01
1.91882539e+00 -4.30298626e-01 -8.15380871e-01 -2.19213128e-01
5.96087873e-01 -3.61324817e-01 1.05823529e+00 -9.76946950e-02
-8.49318504e-01 -4.24292624e-01 -9.47429359e-01 -1.30462870e-01
-8.64604533e-01 -3.81360538e-02 9.24261510e-01 6.60493195e-01
-9.26563561e-01 2.82482177e-01 -5.98272026e-01 -9.63061452e-01
1.99536800e-01 2.12303977e-02 -5.73374271e-01 1.26162305e-01
-1.31217015e+00 1.14282811e+00 6.17455482e-01 -1.32319778e-01
-7.23793685e-01 -2.87324339e-01 -1.16456509e+00 4.84932289e-02
2.34317526e-01 -1.18401265e+00 9.35574234e-01 -1.24531853e+00
-1.17889655e+00 9.57982898e-01 -5.65590978e-01 -7.52493590e-02
-3.38273048e-01 1.52810767e-01 -2.49756217e-01 3.89705658e-01
-7.68203363e-02 8.41453731e-01 8.04691613e-01 -1.65363765e+00
-2.34655321e-01 -6.30888641e-01 1.70113862e-01 2.47332186e-01
-4.59320694e-01 -1.59097895e-01 3.81604433e-02 -6.08326674e-01
6.14386797e-01 -7.60072589e-01 1.12036586e-01 -1.26786068e-01
-3.00420493e-01 -2.86652565e-01 1.42076656e-01 -3.58878076e-01
7.49847889e-01 -2.13722658e+00 7.50899315e-01 3.24276000e-01
4.60435838e-01 -5.06012917e-01 -5.02969265e-01 7.36269474e-01
-1.78221047e-01 3.35233897e-01 -3.76119226e-01 -1.95948616e-01
4.13007140e-01 5.88626921e-01 -6.85596764e-01 3.79491121e-01
3.73775035e-01 1.17423809e+00 -1.07501042e+00 -4.71797734e-01
2.21608564e-01 5.08176982e-01 -4.79726017e-01 -2.17299685e-02
-2.10511506e-01 3.44678313e-01 -7.61521041e-01 6.83125615e-01
2.53858179e-01 -2.03676477e-01 5.06698430e-01 -3.84893000e-01
2.51644224e-01 1.45977855e-01 -6.25731468e-01 1.81675339e+00
-3.44038188e-01 9.19697642e-01 -1.91664711e-01 -1.46914721e+00
7.37558126e-01 3.31978798e-01 1.07865408e-02 -6.03772283e-01
3.31362695e-01 2.89633442e-02 1.35237873e-02 -5.45597613e-01
6.53627217e-01 -8.78571212e-01 -3.02438527e-01 7.29227304e-01
4.01996970e-01 -1.60857886e-01 -1.13375627e-01 3.53640497e-01
5.57697654e-01 -2.07022596e-02 4.07540858e-01 -3.00751507e-01
5.44267654e-01 -9.07182619e-02 -1.91647500e-01 4.76026714e-01
4.04054895e-02 5.44858336e-01 8.07037532e-01 -3.18997324e-01
-6.72157347e-01 -1.53081787e+00 -3.09397221e-01 1.48015440e+00
4.16253597e-01 -5.44574440e-01 -6.12257004e-01 -2.08348483e-01
-8.92287195e-02 9.44453239e-01 -9.95857537e-01 -2.65392572e-01
8.84611979e-02 -7.14425921e-01 6.17920578e-01 6.78494573e-01
3.76394354e-02 -1.05685008e+00 -5.49406767e-01 -1.44853756e-01
-4.91270453e-01 -9.57451582e-01 2.68516466e-02 2.29237199e-01
-1.09002447e+00 -9.49602604e-01 -6.56129956e-01 -6.07446551e-01
9.25829709e-01 2.53139764e-01 1.12117660e+00 1.04398198e-01
5.52865937e-02 9.94439662e-01 -5.07681906e-01 -3.37502867e-01
-3.12488854e-01 -2.04460651e-01 -4.28655092e-03 1.36765286e-01
5.62518001e-01 -4.59709674e-01 -1.59117267e-01 3.24637629e-02
-1.40429723e+00 1.03569388e-01 5.26308000e-01 9.19326782e-01
1.96864963e-01 -4.32879448e-01 5.16901791e-01 -7.69392014e-01
1.03141010e+00 -7.14612842e-01 1.89815834e-01 5.84654450e-01
-1.83041424e-01 6.52058303e-01 2.06374884e-01 -4.91165876e-01
-1.06681538e+00 -1.92999929e-01 2.95460761e-01 -5.76336309e-02
-3.29436034e-01 9.27008927e-01 -5.93978651e-02 4.69700217e-01
4.33752716e-01 5.35097003e-01 2.86722720e-01 -2.56461442e-01
1.24812818e+00 3.41573119e-01 3.52998167e-01 -1.09989393e+00
4.30684835e-01 6.79112315e-01 -7.50989839e-02 -1.13046491e+00
-7.64629781e-01 -4.00352597e-01 -8.89716029e-01 -6.44943714e-02
1.35495698e+00 -7.34009504e-01 -7.35409141e-01 -8.51719528e-02
-1.40800178e+00 2.31485918e-01 -2.20330164e-01 4.83014375e-01
-9.26049590e-01 5.79617620e-01 -4.95676696e-01 -8.92718971e-01
2.11318061e-02 -8.01203489e-01 1.20686483e+00 1.09157838e-01
-8.05053890e-01 -1.47064614e+00 6.77435398e-02 4.81455386e-01
3.03815633e-01 -1.67923361e-01 1.61269951e+00 -7.94655085e-01
-1.81089729e-01 9.97551531e-02 -4.70935553e-01 -1.02558462e-02
1.68176659e-03 -4.12824452e-01 -1.03055418e+00 1.95589438e-01
1.19626792e-02 -3.41690660e-01 1.12042439e+00 3.12814653e-01
8.32795143e-01 -1.47021534e-02 -1.12647459e-01 2.28357747e-01
1.33219600e+00 -7.23556951e-02 5.50418496e-01 1.57119647e-01
4.92439896e-01 1.22477984e+00 7.61981532e-02 2.34537080e-01
5.09047747e-01 1.03019774e-01 3.51727396e-01 1.33682102e-01
8.13123584e-02 -4.17372882e-01 4.58151817e-01 1.08205748e+00
-2.22530171e-01 -9.35384631e-02 -8.15480888e-01 5.49025834e-01
-2.02403831e+00 -1.01493979e+00 -1.07540756e-01 1.80432856e+00
6.35103226e-01 -4.28660601e-01 -7.37175867e-02 -2.65869409e-01
6.81033731e-01 1.63081005e-01 -1.59422651e-01 -4.34372425e-01
-4.61186320e-01 -4.86212671e-02 -3.33143324e-02 3.49193126e-01
-8.08558643e-01 9.44481075e-01 6.92030001e+00 4.02810633e-01
-7.99055815e-01 2.67531157e-01 2.48191565e-01 3.35825801e-01
-9.74638462e-01 2.10944340e-01 -8.40135142e-02 1.20962523e-02
7.78114259e-01 -6.33744523e-02 5.24546504e-01 4.73418683e-01
-1.36607885e-01 -2.03195393e-01 -1.50337362e+00 1.06192410e+00
6.12166405e-01 -1.12961388e+00 7.78114378e-01 1.00214005e-01
3.49039674e-01 -2.43004903e-01 -3.01288776e-02 1.60166889e-01
1.00476235e-01 -1.38156343e+00 9.07089949e-01 9.98921335e-01
3.98246169e-01 -3.01734537e-01 7.57087052e-01 1.19018868e-01
-1.13310623e+00 5.12253679e-02 -6.76489532e-01 -2.08016813e-01
1.55883819e-01 1.17115945e-01 -4.44514334e-01 4.54269558e-01
1.00831747e-01 6.45617247e-01 -6.49217069e-01 3.44751030e-01
-3.54492664e-01 4.09792542e-01 1.29971310e-01 -2.72121698e-01
1.02620557e-01 -3.47192675e-01 3.51145744e-01 1.23719013e+00
2.81721771e-01 6.71787187e-02 -2.60778785e-01 1.57493639e+00
1.33081928e-01 1.89177975e-01 -7.88584888e-01 -6.65717661e-01
2.00011551e-01 1.10342586e+00 -8.64493906e-01 -4.75510001e-01
-5.44007003e-01 9.56528187e-01 1.55390278e-01 6.02936983e-01
-5.42969823e-01 -2.82675847e-02 7.55106330e-01 -2.77530342e-01
-9.23139006e-02 -4.31519777e-01 -5.07750452e-01 -1.49398887e+00
-3.62632275e-01 -4.25389767e-01 3.50326151e-01 -1.08174002e+00
-1.76372743e+00 4.25232887e-01 3.68681222e-01 -9.34862196e-01
-8.28693211e-02 -9.99030113e-01 -5.83206475e-01 7.91702747e-01
-1.16121507e+00 -1.43745506e+00 1.36939296e-02 7.26356030e-01
3.63637149e-01 -2.47880876e-01 1.24466538e+00 -4.17543381e-01
-1.28525957e-01 9.57734808e-02 -3.18092965e-02 2.23786831e-01
3.16498458e-01 -1.22988856e+00 -2.16241241e-01 2.99559027e-01
5.70704639e-01 1.19423437e+00 5.98145306e-01 -4.07951146e-01
-1.73386192e+00 -4.19579089e-01 9.89868939e-01 -8.57324898e-01
1.15620458e+00 -2.23469436e-01 -8.83988142e-01 5.70677817e-01
4.74145979e-01 -3.19711894e-01 1.15078235e+00 5.44187248e-01
-7.74357140e-01 4.04115587e-01 -6.46132767e-01 7.16570854e-01
9.05753672e-01 -1.23801243e+00 -1.42907357e+00 6.25543892e-02
7.17919111e-01 5.70115387e-01 -7.14171290e-01 -1.28856093e-01
5.79645276e-01 -5.50909340e-01 1.04411888e+00 -1.16168213e+00
6.54370189e-01 -3.50740582e-01 -7.66486049e-01 -1.44396937e+00
-1.83738783e-01 3.00867885e-01 4.10599768e-01 9.37409520e-01
4.91648316e-01 -8.02188456e-01 1.85187429e-01 6.51601076e-01
2.47864276e-01 -2.31828451e-01 -9.14268017e-01 -5.09510219e-01
4.03634459e-01 -6.91184044e-01 4.58359003e-01 1.20550680e+00
7.23339021e-01 7.68451393e-01 -1.12991631e-01 -1.58571854e-01
5.61932802e-01 2.73836166e-01 2.85424292e-01 -1.36890125e+00
4.93692420e-02 -7.31940746e-01 -7.24830210e-01 -8.55540276e-01
6.55633688e-01 -1.31275940e+00 -5.45271114e-02 -1.50534499e+00
7.43772507e-01 1.73193589e-01 -4.79198903e-01 5.99865377e-01
-1.17468454e-01 8.50363150e-02 5.50041974e-01 2.47312710e-01
-7.33510911e-01 7.18204319e-01 1.17637110e+00 -3.75220776e-01
3.59099388e-01 -8.06693971e-01 -9.49963689e-01 8.30879152e-01
6.67601109e-01 -1.31271407e-01 -5.75297773e-01 -4.77069885e-01
6.52517080e-01 -1.06202823e-03 6.08842134e-01 -3.55155826e-01
1.35073185e-01 -5.11898398e-01 2.85270959e-01 -2.30955020e-01
4.28773314e-01 -8.35769594e-01 -1.19115464e-01 2.72065610e-01
-7.92727470e-01 5.92902042e-02 -2.10818648e-02 6.62088037e-01
-6.03928685e-01 -4.16536242e-01 3.53059113e-01 -2.17005029e-01
-9.40908074e-01 -3.54776293e-01 -1.02931023e+00 -2.65231729e-02
6.37905478e-01 -2.35409379e-01 -4.00213152e-01 -5.52953482e-01
-9.60665882e-01 -1.16061121e-01 2.99774617e-01 6.65598273e-01
8.18218768e-01 -1.51134419e+00 -4.10403609e-01 9.61228758e-02
5.39985418e-01 -8.10422599e-01 1.78242609e-01 8.45885456e-01
-1.89495519e-01 6.05249941e-01 -3.00532430e-01 -6.20059013e-01
-9.59050775e-01 5.31534016e-01 1.89551741e-01 4.95510131e-01
-2.30923250e-01 6.14411831e-01 7.51657009e-01 -5.72346389e-01
-6.28194585e-02 -5.38619697e-01 -4.78521675e-01 5.09643972e-01
3.81291151e-01 -5.16769895e-03 -5.41938841e-01 -1.06459582e+00
-4.75206971e-01 8.43201637e-01 4.86059219e-01 -4.77976054e-01
1.05535972e+00 -3.03950936e-01 -7.10067272e-01 1.07296312e+00
1.20220900e+00 -3.53730500e-01 -3.74799043e-01 -2.50334322e-01
4.17149663e-01 -1.87340870e-01 -2.17450604e-01 -6.07076526e-01
-8.01254749e-01 1.21246362e+00 2.53935844e-01 2.57207096e-01
8.46060693e-01 7.27234900e-01 3.15524668e-01 5.45921028e-01
2.38025784e-01 -9.47234929e-01 2.16583446e-01 8.00432682e-01
9.87530828e-01 -1.13020635e+00 -1.90956041e-01 -3.77128214e-01
-9.72890615e-01 1.51001918e+00 1.16023958e-01 1.44586667e-01
5.32899380e-01 -3.89637649e-01 -5.76205403e-02 -6.16846919e-01
-8.52761984e-01 -5.69241107e-01 6.26096249e-01 6.28286839e-01
8.51739764e-01 5.72361708e-01 -3.72816265e-01 7.76989639e-01
-2.53127486e-01 -3.73702765e-01 2.93127149e-01 6.53195739e-01
-5.23931026e-01 -1.00244093e+00 -6.03606462e-01 2.27475703e-01
2.12595072e-02 -1.94087356e-01 -8.14351201e-01 4.55129713e-01
7.80791342e-02 9.65687394e-01 1.70764104e-01 -3.53318036e-01
1.12131210e-02 6.32447183e-01 6.36949599e-01 -5.99781334e-01
-9.75600854e-02 -5.23973927e-02 7.95786530e-02 -3.33172053e-01
-8.54129672e-01 -6.92594230e-01 -1.56932986e+00 -1.51538134e-01
-8.54906719e-03 1.56836003e-01 7.46797860e-01 1.39706457e+00
2.83326954e-02 4.34086055e-01 -1.02397479e-01 -7.36003101e-01
-3.19023788e-01 -9.03516471e-01 -7.43740439e-01 8.17599475e-01
1.57613069e-01 -7.40627587e-01 -3.99181396e-01 3.65900099e-01] | [10.636983871459961, 2.03340482711792] |
965280f6-20bc-4ecd-9096-2805a7f59a84 | virtuoso-massive-multilingual-speech-text | 2210.15447 | null | https://arxiv.org/abs/2210.15447v2 | https://arxiv.org/pdf/2210.15447v2.pdf | Virtuoso: Massive Multilingual Speech-Text Joint Semi-Supervised Learning for Text-To-Speech | This paper proposes Virtuoso, a massively multilingual speech-text joint semi-supervised learning framework for text-to-speech synthesis (TTS) models. Existing multilingual TTS typically supports tens of languages, which are a small fraction of the thousands of languages in the world. One difficulty to scale multilingual TTS to hundreds of languages is collecting high-quality speech-text paired data in low-resource languages. This study extends Maestro, a speech-text joint pretraining framework for automatic speech recognition (ASR), to speech generation tasks. To train a TTS model from various types of speech and text data, different training schemes are designed to handle supervised (paired TTS and ASR data) and unsupervised (untranscribed speech and unspoken text) datasets. Experimental evaluation shows that 1) multilingual TTS models trained on Virtuoso can achieve significantly better naturalness and intelligibility than baseline ones in seen languages, and 2) they can synthesize reasonably intelligible and naturally sounding speech for unseen languages where no high-quality paired TTS data is available. | ['Bhuvana Ramabhadran', 'Andrew Rosenberg', 'Ankur Bapna', 'Yu Zhang', 'Gary Wang', 'Nobuyuki Morioka', 'Zhehuai Chen', 'Heiga Zen', 'Takaaki Saeki'] | 2022-10-27 | null | null | null | null | ['text-to-speech-synthesis'] | ['speech'] | [-3.89571674e-02 2.12831378e-01 -9.29761231e-02 -4.40715581e-01
-1.42720437e+00 -6.27437651e-01 7.91861236e-01 -4.56474662e-01
-7.69981965e-02 7.95643032e-01 6.77973390e-01 -9.86646473e-01
6.09125376e-01 -2.96967059e-01 -7.66685426e-01 -4.33970302e-01
3.67466331e-01 7.74394810e-01 -4.28576693e-02 -4.69117224e-01
-6.90493405e-01 1.89269483e-01 -1.27223265e+00 4.10494685e-01
1.17620552e+00 6.70182884e-01 7.58209109e-01 1.01030064e+00
-2.95465410e-01 8.58636677e-01 -8.51738453e-01 -4.45250481e-01
1.87679976e-01 -3.75414282e-01 -7.29238272e-01 5.39618172e-02
4.56916571e-01 -3.64901990e-01 -4.25577849e-01 7.09931612e-01
7.78776467e-01 1.44547045e-01 5.62466443e-01 -1.07618046e+00
-9.78532135e-01 1.22841716e+00 1.95836708e-01 4.44635898e-02
4.20461655e-01 3.20605576e-01 1.06716788e+00 -1.13310432e+00
4.38519597e-01 1.75840986e+00 1.50592640e-01 7.06648231e-01
-1.20581245e+00 -7.29953647e-01 -1.27106413e-01 -3.53592992e-01
-1.22851074e+00 -1.31791723e+00 4.41237539e-01 -1.31257370e-01
1.30098033e+00 5.88091195e-01 2.17180029e-01 1.64529443e+00
-4.90398370e-02 1.20542443e+00 1.11063457e+00 -7.03548491e-01
-6.84921220e-02 4.02919143e-01 -5.24051666e-01 3.44532847e-01
-4.74495769e-01 4.56654876e-01 -8.05055320e-01 3.23752642e-01
4.44887340e-01 -7.02710271e-01 -2.16949865e-01 5.06805480e-01
-1.73013163e+00 5.83536267e-01 -9.63726789e-02 4.79893953e-01
7.88104255e-03 -2.40396142e-01 6.85583711e-01 7.59869337e-01
7.68903911e-01 2.37517074e-01 -6.73692703e-01 -2.70070910e-01
-9.92852032e-01 -1.09504871e-01 7.08976269e-01 1.54066074e+00
4.72845346e-01 1.13477397e+00 -3.91901433e-01 1.41028488e+00
2.78845102e-01 1.47230375e+00 9.09144580e-01 -5.32763481e-01
9.99608457e-01 -1.87077627e-01 -9.18003023e-02 -8.59160870e-02
2.82175064e-01 -2.88070828e-01 -7.52780199e-01 -1.11989453e-01
2.87739486e-02 -4.67436850e-01 -1.08791602e+00 1.58462024e+00
8.69372189e-02 -2.72775829e-01 7.93078423e-01 5.10455132e-01
1.13035572e+00 1.44011962e+00 -1.48590878e-01 -4.53677416e-01
1.09802115e+00 -1.49005449e+00 -1.12788546e+00 -4.91907001e-01
6.60203457e-01 -1.37198997e+00 1.67408466e+00 1.81905329e-01
-1.36348641e+00 -8.92265141e-01 -8.44815254e-01 -3.93209219e-01
-4.61891621e-01 5.17817497e-01 7.67942667e-02 7.83817232e-01
-1.33851767e+00 8.73682946e-02 -6.78036451e-01 -2.31057659e-01
-3.71519119e-01 -5.46965487e-02 -3.19579035e-01 4.21078876e-02
-1.48010349e+00 9.78871524e-01 4.66006607e-01 -3.20112221e-02
-1.21549344e+00 -3.27869713e-01 -1.25740194e+00 -1.69429943e-01
2.87736118e-01 -3.94775808e-01 1.72477913e+00 -8.59711766e-01
-2.05625844e+00 6.00274444e-01 -4.80840445e-01 -4.05162781e-01
4.71873552e-01 -2.65922308e-01 -1.04707241e+00 -3.53819042e-01
1.06497914e-01 6.56336963e-01 9.34443414e-01 -1.16757286e+00
-6.47957444e-01 -8.78661722e-02 -6.95379138e-01 4.77323890e-01
-4.45361584e-01 4.37817484e-01 -2.03739390e-01 -1.10047066e+00
-1.61565155e-01 -8.03978980e-01 -1.60648767e-02 -6.25166714e-01
-8.41761529e-01 -4.38253880e-01 9.49922442e-01 -9.82045174e-01
1.09689498e+00 -1.91040754e+00 1.77861080e-01 -4.32882488e-01
-4.89250094e-01 5.54076970e-01 -4.29599941e-01 6.02588415e-01
5.86406849e-02 1.01480171e-01 1.23627953e-01 -8.23645353e-01
2.87270218e-01 4.27793473e-01 -8.39348197e-01 -1.94878936e-01
2.06201430e-02 9.49351430e-01 -9.34998989e-01 -4.15733427e-01
6.14245832e-01 2.95861304e-01 2.46915668e-02 7.20473528e-01
-3.23407680e-01 6.54309988e-01 -1.30102947e-01 5.65792143e-01
1.46345094e-01 4.31393474e-01 7.49543915e-03 9.67988893e-02
-3.75591457e-01 1.00198030e+00 -6.93669617e-01 1.55853832e+00
-1.06803131e+00 7.14274645e-01 -4.97299433e-02 -5.08962631e-01
1.12496436e+00 1.10215116e+00 -3.86441760e-02 -7.97321498e-01
-6.84718192e-02 8.57382715e-01 -2.10407108e-01 -3.69218200e-01
7.18038857e-01 -4.25200135e-01 -2.34354272e-01 2.50564009e-01
4.73988265e-01 -9.30491269e-01 1.98498398e-01 -1.12886000e-02
6.17586851e-01 -7.86507875e-02 1.85833693e-01 -1.70255244e-01
4.25682664e-01 -2.24755481e-01 3.18666101e-01 4.07515854e-01
-1.70585476e-02 6.51884258e-01 -2.48032510e-01 -7.14331716e-02
-1.27392721e+00 -1.55344725e+00 -6.10829750e-03 1.37586546e+00
-5.55002153e-01 -4.75695580e-01 -8.03077579e-01 -4.19183254e-01
-5.03119528e-01 1.23262298e+00 2.27451891e-01 1.36597827e-01
-6.18380427e-01 -8.63696039e-02 9.34917450e-01 3.73447597e-01
2.38667428e-01 -1.33923876e+00 6.33903444e-01 3.35165024e-01
-7.37798750e-01 -1.66452014e+00 -1.05702615e+00 1.58285603e-01
-5.25245190e-01 -6.98244646e-02 -1.03147089e+00 -1.18424118e+00
7.18109906e-02 1.81094423e-01 1.10943103e+00 -6.72856152e-01
3.34942669e-01 -8.09110031e-02 -4.42323804e-01 -4.75085020e-01
-1.62692237e+00 -9.22692567e-03 6.47698939e-01 -9.53933895e-02
-7.62666017e-02 -4.57763344e-01 3.55982363e-01 2.64760613e-01
-6.64982200e-01 3.44706297e-01 5.90624392e-01 7.47718573e-01
5.36645174e-01 -1.64018005e-01 7.59351969e-01 -5.25046349e-01
5.87690890e-01 -1.79907203e-01 -5.94118476e-01 5.62799752e-01
-2.88374543e-01 1.78155273e-01 1.13646805e+00 -5.22384048e-01
-1.22947824e+00 -1.86590105e-01 -5.44871867e-01 -5.41249871e-01
-2.45598540e-01 4.49857801e-01 -4.66017008e-01 6.12099707e-01
7.34641790e-01 6.16681159e-01 -3.31729174e-01 -5.88706613e-01
6.76180780e-01 1.52294850e+00 7.70948887e-01 -6.20294988e-01
9.80676115e-01 -2.88269997e-01 -7.65798569e-01 -1.29217100e+00
-7.24426925e-01 -3.23339373e-01 -4.57355380e-01 -7.92812482e-02
8.47583115e-01 -1.28350079e+00 -5.30752279e-02 5.31493008e-01
-1.36523974e+00 -7.14981377e-01 -3.06357354e-01 7.38788247e-01
-7.22878098e-01 2.49869034e-01 -8.14156830e-01 -9.52789247e-01
-5.94377935e-01 -1.46943474e+00 1.59362280e+00 -3.53703916e-01
-9.51363295e-02 -1.05529881e+00 1.19492300e-01 6.94959104e-01
5.25766134e-01 -3.75121027e-01 6.24152839e-01 -5.82681954e-01
-3.04903716e-01 2.03322858e-01 2.22422555e-01 8.98011982e-01
5.39302230e-01 8.67754295e-02 -1.06677973e+00 -4.74730492e-01
-7.42212906e-02 -8.43828380e-01 2.34441116e-01 2.24800676e-01
5.34669280e-01 -7.51289308e-01 1.33951306e-01 4.11215752e-01
6.53264940e-01 2.21706748e-01 3.99377257e-01 -4.03045863e-01
8.99875581e-01 5.80338061e-01 4.01573390e-01 3.96455601e-02
5.00599205e-01 8.11699569e-01 -2.62226582e-01 -4.36560154e-01
-8.23321640e-01 -6.04901135e-01 1.28277910e+00 2.17651868e+00
4.57098782e-01 -7.40450025e-01 -9.35514092e-01 6.99649751e-01
-1.38059592e+00 -5.41256666e-01 -1.35112733e-01 2.27094197e+00
1.31894648e+00 1.34331575e-02 4.72592525e-02 3.11960150e-02
9.34988916e-01 2.43749544e-01 -2.37738758e-01 -5.99488437e-01
-4.76049662e-01 4.30805683e-02 1.01521067e-01 9.02480066e-01
-7.36063659e-01 1.61948228e+00 6.32874298e+00 1.35531580e+00
-1.30786705e+00 2.39690363e-01 5.89532673e-01 8.32372457e-02
-5.93249083e-01 -1.60963550e-01 -7.74709642e-01 3.19211245e-01
1.69793379e+00 -3.87709349e-01 7.52812982e-01 7.54508495e-01
5.62336028e-01 5.64672232e-01 -1.27949989e+00 1.02540147e+00
-3.86492014e-02 -9.86992359e-01 5.78348815e-01 -2.91786015e-01
9.14179265e-01 4.96307462e-01 2.31831491e-01 6.38019979e-01
8.81837308e-01 -1.09568715e+00 1.29027116e+00 -1.77662894e-01
1.39554572e+00 -6.16790771e-01 2.54069865e-01 6.50336564e-01
-1.30208695e+00 3.05294216e-01 -1.38722643e-01 4.51485008e-01
4.59209383e-01 4.79461938e-01 -1.30224943e+00 6.85682714e-01
3.71663690e-01 4.95049238e-01 -2.81064987e-01 3.02984834e-01
-3.64496320e-01 1.03383541e+00 -3.01535964e-01 -8.88084173e-02
3.96458685e-01 2.01450121e-02 8.92098069e-01 1.47518158e+00
6.38859272e-01 -5.07074237e-01 4.80424255e-01 5.93890429e-01
-1.43043950e-01 5.61982632e-01 -7.83725977e-01 -5.44054806e-01
7.30731308e-01 7.81016946e-01 -1.33010820e-01 -5.85446417e-01
-3.03840607e-01 1.12669301e+00 2.12422818e-01 5.82116961e-01
-3.16290081e-01 -2.90288389e-01 3.19049925e-01 -1.82730556e-01
-1.26753211e-01 -4.95628834e-01 -9.09308717e-02 -1.33113015e+00
8.64042342e-02 -1.31733525e+00 -9.58814099e-02 -1.15737820e+00
-1.31341040e+00 1.40557826e+00 -2.80156612e-01 -1.30364239e+00
-8.81311774e-01 -4.20499712e-01 -3.01331013e-01 1.26570964e+00
-1.47495627e+00 -1.66918087e+00 3.98676217e-01 6.55924141e-01
1.28488147e+00 -8.14140677e-01 1.02256548e+00 2.06393883e-01
-4.57828552e-01 7.12238193e-01 3.20515901e-01 1.96764395e-01
8.08107138e-01 -1.45251238e+00 1.08848000e+00 1.06747246e+00
6.52286351e-01 2.81455666e-01 8.27695549e-01 -6.40129328e-01
-1.28685439e+00 -1.42788422e+00 1.38920665e+00 -4.42037702e-01
8.24200630e-01 -9.09685194e-01 -7.14864612e-01 9.00816321e-01
6.14134669e-01 -4.20568362e-02 4.67433512e-01 8.56164247e-02
-2.18527541e-01 -1.84089020e-01 -6.15787864e-01 1.04725146e+00
7.82287955e-01 -1.05338669e+00 -6.63255990e-01 5.78266561e-01
1.51742351e+00 -3.87430459e-01 -7.87168801e-01 2.23050386e-01
1.68296844e-01 -5.81718922e-01 6.16163492e-01 -4.89147931e-01
1.32108584e-01 -2.00908512e-01 -6.41909719e-01 -1.89740860e+00
2.28725001e-01 -1.37640071e+00 1.46628886e-01 1.49675739e+00
8.36613953e-01 -4.67593163e-01 1.26943931e-01 -5.25861792e-02
-8.69506955e-01 -1.24166824e-01 -1.09922230e+00 -1.36053491e+00
3.49508077e-01 -5.92303574e-01 6.83240831e-01 8.83231938e-01
2.16834322e-02 9.16756988e-01 -6.69175446e-01 2.48686597e-01
3.22308600e-01 -1.88054189e-01 8.33978415e-01 -5.76346695e-01
-2.97174126e-01 -1.50822267e-01 2.13721052e-01 -1.25693321e+00
4.53256130e-01 -1.17429090e+00 5.10595739e-01 -1.39695251e+00
-4.69019949e-01 -5.51393747e-01 2.95714587e-01 5.88387251e-01
-1.48933515e-01 -1.44264296e-01 7.13565201e-02 -8.40349123e-02
-1.39685959e-01 1.15938067e+00 1.45387042e+00 -1.10230327e-01
-2.50408083e-01 5.27926907e-02 -1.26018539e-01 3.53466690e-01
5.97275198e-01 -2.52665311e-01 -7.72278786e-01 -8.82416606e-01
-5.57943225e-01 6.42790675e-01 -2.43185908e-01 -9.67086732e-01
1.26076892e-01 -3.03569883e-01 -2.83858210e-01 -5.71063757e-01
5.40982783e-01 -4.45634425e-01 -1.61270909e-02 5.55083565e-02
-5.01593709e-01 -1.06873177e-01 1.79889858e-01 -1.57097168e-02
-4.60463583e-01 9.18301046e-02 7.75281072e-01 -1.36030298e-02
-1.84524029e-01 3.71332854e-01 -7.31912911e-01 2.98910797e-01
4.46743637e-01 1.12177007e-01 -2.45573372e-01 -6.63222611e-01
-5.85125625e-01 3.27513367e-02 -3.88625823e-02 9.92624760e-01
5.70810199e-01 -1.58246386e+00 -1.23243976e+00 3.73812079e-01
2.25308791e-01 -8.37539285e-02 -2.47594137e-02 1.44745842e-01
-3.22382785e-02 8.85265589e-01 2.10332751e-01 -5.16511440e-01
-1.14131534e+00 4.08317745e-01 2.56567091e-01 -2.57477276e-02
-3.62391442e-01 7.51277506e-01 1.82899296e-01 -1.10758245e+00
3.99005085e-01 -5.86842954e-01 2.98264265e-01 -4.20664936e-01
4.74697977e-01 1.82677180e-01 3.19156677e-01 -1.17871416e+00
3.29347537e-03 -3.88999805e-02 -1.03136867e-01 -9.44489539e-01
9.25545633e-01 -4.22143787e-01 1.74326152e-01 8.69611442e-01
1.24315977e+00 4.18772846e-01 -8.15911293e-01 -5.07288635e-01
-1.40534729e-01 -6.69800267e-02 2.48936981e-01 -8.80637288e-01
-6.33698165e-01 1.19953418e+00 1.26257524e-01 2.39738777e-01
7.05369532e-01 5.81025369e-02 1.45161092e+00 6.26531065e-01
3.37096870e-01 -1.27574909e+00 9.76163894e-02 9.28668320e-01
1.29419267e+00 -1.33903563e+00 -8.78678977e-01 -3.44267517e-01
-9.99925137e-01 8.97814214e-01 4.54666346e-01 5.64014852e-01
4.13034201e-01 5.96071005e-01 5.76968133e-01 3.93769532e-01
-1.08144128e+00 -1.44398421e-01 5.54893792e-01 7.49272943e-01
7.35154331e-01 5.15945494e-01 4.23805505e-01 2.90461659e-01
-1.07463121e+00 -5.22429824e-01 3.27629060e-01 5.21296620e-01
-4.62488264e-01 -1.28403771e+00 -3.90798748e-01 9.50843748e-03
-3.14159006e-01 -6.40128314e-01 -3.09786856e-01 1.90184489e-01
-3.04558814e-01 1.55744338e+00 -2.58676201e-01 -3.86018664e-01
3.55758518e-01 7.73220286e-02 1.45417199e-01 -9.74556386e-01
-3.54149848e-01 6.03084028e-01 5.41859925e-01 -1.37800992e-01
7.91103989e-02 -5.23320198e-01 -1.11965144e+00 -1.42751753e-01
-3.83693069e-01 3.13909888e-01 8.35532010e-01 1.11232889e+00
-9.15140659e-02 5.27140617e-01 1.09186172e+00 -6.57323360e-01
-6.75520241e-01 -1.24438715e+00 -5.51576555e-01 -1.33999838e-02
7.89117217e-01 -1.96257941e-02 -3.59062195e-01 3.43278319e-01] | [14.631162643432617, 7.001558303833008] |
91f90b1d-ba9e-4bfc-9e56-9b23339deb86 | neural-ordinary-differential-equation-value | null | null | https://openreview.net/forum?id=8WKd467B8H | https://openreview.net/pdf?id=8WKd467B8H | Neural Ordinary Differential Equation Value Networks for Parametrized Action Spaces | Action spaces equipped with parameter sets are a common occurrence in reinforcement learning applications. Solutions to problems of this class have been developed under different frameworks, such as parametrized action Markov decision processes (PAMDP) or hierarchical reinforcement learning (HRL). These approaches often require extensions or modifications to standard existing algorithms developed on standard MDPs. For this reason they can be unwieldy and, particularly in the case of HRL, computationally inefficient. We propose adopting a different parametrization scheme for state--action value networks based on neural ordinary differential equations (NODEs) as a scalable, plug--and--play approach for parametrized action spaces. NODEs value networks do not require extensive modification to existing algorithms nor the adoption of HRL methods. Our solution can directly be integrated into existing training algorithms and opens up new opportunities in single-agent and multi-agent settings with tight precision constraints on the action parameters such as robotics. | ['Jinkyoo Park', 'Hajime Asama', 'Atsushi Yamashita', 'Sanzhar Bakhtiyarov', 'Michael Poli', 'Stefano Massaroli'] | 2020-02-26 | null | null | null | iclr-workshop-deepdiffeq-2019-12 | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 1.20671913e-01 4.22985345e-01 -2.23162845e-01 1.02517754e-02
-4.15945947e-01 -5.24829209e-01 8.39597344e-01 1.16866492e-01
-7.28768647e-01 1.28009140e+00 -1.78764001e-01 -3.88197392e-01
-6.81089044e-01 -7.94673979e-01 -6.02951944e-01 -1.06587183e+00
-3.24446499e-01 7.04826474e-01 4.19720083e-01 -6.30122900e-01
2.24015206e-01 5.87289989e-01 -1.35583282e+00 -4.04877841e-01
6.44984841e-01 6.04994535e-01 1.88820496e-01 6.95028365e-01
-4.50514928e-02 8.65253806e-01 -2.43986592e-01 2.02164873e-02
4.69139755e-01 -3.57821614e-01 -9.18616295e-01 1.42793149e-01
-3.99575710e-01 -5.15335560e-01 -3.58378202e-01 8.21496308e-01
5.10833263e-01 5.74955940e-01 5.54133952e-01 -1.51182365e+00
-7.84038827e-02 7.09847033e-01 -1.91799104e-01 1.93484053e-02
2.45489702e-01 4.74785417e-01 6.98888779e-01 -1.97497487e-01
5.52821577e-01 1.48871171e+00 5.50705492e-01 8.19404662e-01
-1.41639543e+00 -7.26938471e-02 3.52752924e-01 1.99936464e-01
-8.26439202e-01 -1.34949073e-01 3.75304073e-01 -2.94911891e-01
9.89415109e-01 -1.72615111e-01 7.44766712e-01 1.23828292e+00
1.96510971e-01 9.23126817e-01 1.26836753e+00 -3.52956027e-01
7.80757904e-01 -2.37498954e-01 -3.05324495e-01 5.50701320e-01
2.53684312e-01 2.60276258e-01 -8.76917094e-02 -2.20628485e-01
1.21272480e+00 1.07686065e-01 1.82990104e-01 -1.05301583e+00
-1.22608542e+00 1.15919173e+00 1.60724685e-01 4.43816930e-03
-7.18964040e-01 5.08496404e-01 3.53239328e-01 5.09645700e-01
-2.02351250e-02 4.35346663e-01 -4.85239923e-01 -3.76674831e-01
-3.66078109e-01 9.84345555e-01 9.83664513e-01 6.69461906e-01
8.14924181e-01 1.33795843e-01 3.41121387e-03 4.97647464e-01
3.13481092e-01 1.37150362e-01 3.51903826e-01 -1.60758829e+00
3.25943142e-01 3.52503717e-01 6.88003480e-01 -4.05627936e-01
-6.14969909e-01 2.56065628e-03 -6.25862479e-01 5.36623716e-01
5.62090695e-01 -6.48751140e-01 -6.73191071e-01 1.67832136e+00
7.16933370e-01 1.27916560e-01 3.71391267e-01 6.49226248e-01
-1.24316886e-01 5.93730330e-01 1.27080709e-01 -2.01619089e-01
8.61902237e-01 -7.91390836e-01 -5.73357761e-01 2.11969446e-02
7.55040526e-01 -2.57272035e-01 8.50146472e-01 5.42406440e-01
-1.14252448e+00 -8.44527483e-02 -9.07752454e-01 3.45213532e-01
-4.25419927e-01 -4.42897737e-01 8.32354307e-01 4.80564982e-01
-1.27926421e+00 1.17032361e+00 -1.14242458e+00 -3.73531312e-01
9.97912064e-02 7.23387778e-01 -1.47559112e-02 2.08116561e-01
-1.23627782e+00 1.31809366e+00 5.62397420e-01 7.98092484e-02
-1.37416363e+00 -2.81157583e-01 -8.42265248e-01 -1.77855462e-01
9.77097273e-01 -5.71645319e-01 1.90276003e+00 -7.08669066e-01
-2.44438291e+00 1.55564338e-01 6.81773305e-01 -8.83603930e-01
7.98724592e-01 -2.06129774e-01 2.27133989e-01 1.85908839e-01
-4.58885655e-02 6.84355378e-01 8.20755959e-01 -9.13447261e-01
-4.25388694e-01 -1.18374517e-02 7.61060238e-01 5.54968417e-01
2.11603701e-01 -5.59867546e-02 3.76559973e-01 -2.54394650e-01
-1.74202576e-01 -1.21030509e+00 -1.04974151e+00 8.05842970e-03
-3.54445353e-02 -3.53735894e-01 6.74996495e-01 -8.72844681e-02
6.94236398e-01 -1.53426635e+00 6.59131527e-01 -9.04056728e-02
-2.47195929e-01 4.09124851e-01 -7.61550441e-02 1.00602663e+00
2.81763732e-01 -1.85888633e-01 -3.70939940e-01 -7.26691633e-02
5.66403151e-01 7.85029590e-01 -8.72839019e-02 4.68506277e-01
1.89125106e-01 6.09804690e-01 -1.08657098e+00 -4.22822952e-01
3.70232642e-01 2.74881512e-01 -5.91226518e-01 1.16115816e-01
-6.42754912e-01 6.17715657e-01 -7.31653333e-01 1.34670764e-01
1.82392016e-01 1.60586134e-01 3.07136118e-01 6.50615036e-01
-4.73912597e-01 3.44734758e-01 -1.60585034e+00 1.56248975e+00
-5.11625946e-01 1.43934563e-01 3.28143537e-01 -1.36243963e+00
6.25609815e-01 4.33623195e-01 6.42007947e-01 -2.31405854e-01
1.76437587e-01 1.29655883e-01 1.96992010e-01 -4.94486362e-01
4.27115172e-01 -4.71328437e-01 -2.46596947e-01 4.57713217e-01
2.33375266e-01 -4.37545300e-01 4.94776756e-01 -4.01190333e-02
1.22619236e+00 6.75150335e-01 5.38356543e-01 -1.50787368e-01
5.34428000e-01 1.39877703e-02 4.47612256e-01 8.52601409e-01
-3.23803425e-01 -5.92050031e-02 8.69448304e-01 -3.47843319e-01
-1.09488988e+00 -6.58460200e-01 -8.90141726e-02 1.13315630e+00
5.57794198e-02 -2.29142398e-01 -8.58459175e-01 -5.62603056e-01
-7.16849715e-02 4.30242687e-01 -4.44954574e-01 8.88835788e-02
-6.47032797e-01 -6.07691824e-01 3.84329736e-01 3.96697700e-01
4.33980525e-01 -1.42885983e+00 -1.00477600e+00 7.43246734e-01
4.28372622e-01 -9.89997029e-01 3.67873996e-01 5.54916084e-01
-9.81641173e-01 -8.35336030e-01 -8.24186504e-01 -4.59202856e-01
2.29891479e-01 -1.80081531e-01 6.84467852e-01 -2.92840123e-01
1.78885877e-01 6.73281908e-01 -3.20187986e-01 -3.44741136e-01
-6.09309316e-01 1.14807762e-01 3.60786468e-01 -1.10012725e-01
-9.97365639e-02 -6.57381713e-01 -5.26377618e-01 4.35439736e-01
-9.11100268e-01 -8.47298093e-03 4.48751837e-01 8.36985350e-01
3.12042713e-01 4.55921814e-02 6.88153744e-01 -6.89972579e-01
7.77732670e-01 -4.53859866e-01 -1.11300445e+00 -1.22890919e-01
-4.65076625e-01 4.57292438e-01 7.71219254e-01 -5.20976484e-01
-1.02378380e+00 1.88356429e-01 -1.21134058e-01 -1.85905904e-01
-3.74955535e-01 4.69260871e-01 1.10922297e-02 -2.72315830e-01
3.58419120e-01 -9.77106839e-02 2.73935914e-01 -3.89574915e-01
4.08814371e-01 3.26444149e-01 1.29811555e-01 -9.20475543e-01
5.87464750e-01 2.72613525e-01 4.57841754e-01 -6.89563096e-01
-3.42036247e-01 -8.64021927e-02 -5.93759060e-01 -1.25820473e-01
7.56488979e-01 -5.91682315e-01 -9.40747917e-01 5.47412395e-01
-8.18343699e-01 -1.04114211e+00 -4.53307897e-01 5.55945814e-01
-1.33122516e+00 4.19892222e-01 -7.44245827e-01 -1.06055009e+00
2.85581261e-01 -1.28030896e+00 7.75350988e-01 4.54097062e-01
3.03153098e-02 -1.17104805e+00 3.82516950e-01 -1.52481258e-01
5.09061217e-01 3.78601283e-01 7.21184373e-01 -5.81140637e-01
-5.93779504e-01 -1.07643921e-02 3.54878724e-01 1.77418351e-01
-1.49826646e-01 7.08024874e-02 -5.10162175e-01 -2.51331836e-01
-2.05910429e-01 -5.02608836e-01 3.28112483e-01 4.94835228e-01
7.76676595e-01 -3.67880851e-01 -1.04422055e-01 4.23608534e-02
1.41179657e+00 4.36101615e-01 4.47621405e-01 7.51511693e-01
3.04179519e-01 7.39266455e-01 7.07279980e-01 8.06416452e-01
3.59624773e-01 9.01151836e-01 5.97274601e-01 3.89207453e-01
5.98584175e-01 -1.31809682e-01 6.81214035e-01 2.31103033e-01
-3.43865424e-01 4.67624553e-02 -9.27665830e-01 4.80144978e-01
-2.31626129e+00 -9.62562442e-01 6.63948432e-02 2.16788268e+00
7.27245748e-01 4.15347926e-02 6.41470015e-01 8.32922831e-02
5.49559712e-01 -1.44532993e-02 -6.52270913e-01 -6.92894101e-01
3.16733688e-01 1.97383806e-01 5.28911591e-01 5.22362351e-01
-1.16927421e+00 9.01970863e-01 6.17692089e+00 5.26865482e-01
-7.90864527e-01 1.42332613e-01 9.35383141e-02 -8.11596774e-03
2.27103025e-01 2.08107442e-01 -8.42732251e-01 3.48813564e-01
1.18312943e+00 -1.02411024e-01 6.65010035e-01 9.68744755e-01
5.11730969e-01 -4.25063103e-01 -9.65762973e-01 5.99760830e-01
-6.75205171e-01 -1.00366509e+00 -4.85214472e-01 2.39760309e-01
6.69554889e-01 -4.61566038e-02 -1.86717883e-01 7.74813056e-01
9.74453449e-01 -7.54997373e-01 4.54826593e-01 2.23419517e-01
1.33525729e-01 -7.34785140e-01 6.34816289e-01 6.14841640e-01
-7.71014690e-01 -3.91780913e-01 -5.09205759e-01 -5.09165466e-01
3.50653678e-01 -1.27720445e-01 -5.60183167e-01 3.36791903e-01
3.23500633e-01 5.37679493e-01 9.40727368e-02 1.00413370e+00
-2.40183681e-01 3.65685403e-01 -5.59875429e-01 -4.92116868e-01
9.09779787e-01 -3.72962028e-01 5.23456931e-01 7.14877546e-01
9.12641361e-02 -5.20514026e-02 4.11245942e-01 4.01496977e-01
5.28562605e-01 -3.92283201e-02 -7.68754661e-01 -1.09829374e-01
-2.99398452e-02 1.16892505e+00 -9.11147475e-01 -2.86307544e-01
-2.91547567e-01 5.79757750e-01 3.68685752e-01 3.71912599e-01
-8.97280276e-01 -2.02690169e-01 8.25835705e-01 -1.39577404e-01
5.86865604e-01 -5.16991913e-01 2.33574703e-01 -1.07931471e+00
-3.36158991e-01 -9.12797391e-01 1.78540170e-01 -2.42835507e-01
-9.49196041e-01 1.67853251e-01 5.59087276e-01 -1.27731979e+00
-1.01614821e+00 -8.36915612e-01 -3.42430353e-01 3.64878923e-01
-1.35228515e+00 -6.33745134e-01 2.84969747e-01 7.13159382e-01
5.15085936e-01 -1.76582870e-03 8.64492953e-01 -1.42853662e-01
-6.67892396e-01 -1.70109451e-01 3.58433187e-01 -3.46102178e-01
1.34587944e-01 -1.48088419e+00 7.45636821e-02 6.47241533e-01
-2.41046086e-01 2.23781526e-01 9.43108201e-01 -2.88859189e-01
-1.41907132e+00 -7.73518920e-01 2.02358186e-01 -1.76897749e-01
1.06711006e+00 -2.61020690e-01 -7.16371119e-01 7.02958405e-01
3.11156273e-01 -1.75419450e-01 1.83397412e-01 -3.43972992e-04
3.54933798e-01 1.80846438e-01 -1.07330132e+00 8.61323893e-01
7.59893596e-01 -2.37597823e-01 -3.73596758e-01 3.46518934e-01
4.75591332e-01 -3.39370817e-01 -8.88424575e-01 1.70158654e-01
2.59449482e-01 -8.38506103e-01 9.22766566e-01 -8.68099749e-01
1.86136782e-01 -2.50997752e-01 -1.14539592e-02 -1.60198081e+00
-1.47911176e-01 -1.27364850e+00 -3.52509528e-01 8.12789500e-01
1.28604576e-01 -7.42047012e-01 6.52838886e-01 7.42428780e-01
-2.67335158e-02 -8.57663214e-01 -1.14843225e+00 -1.00586545e+00
3.50431055e-01 -3.23168904e-01 4.99835819e-01 5.53518057e-01
1.70731381e-01 6.08711243e-02 -4.68967736e-01 5.64973243e-02
6.04017317e-01 -1.59441531e-01 7.81765401e-01 -1.07466495e+00
-7.50300825e-01 -5.02000153e-01 -2.85025030e-01 -8.51617754e-01
2.85440058e-01 -3.95799279e-01 3.25811356e-01 -1.56060922e+00
-3.36690277e-01 -6.01123154e-01 -1.25764862e-01 6.66716278e-01
2.21162289e-01 -2.95674562e-01 3.09937537e-01 -1.36581764e-01
-7.35761344e-01 7.97627032e-01 1.18216753e+00 2.92912185e-01
-4.03677553e-01 3.10354799e-01 -7.60373399e-02 9.10762846e-01
1.08025301e+00 -5.53741813e-01 -5.34447551e-01 3.38332541e-02
2.97310352e-01 5.47167063e-01 3.62869114e-01 -1.10027158e+00
1.02286771e-01 -6.49015784e-01 -2.50527173e-01 -2.75760256e-02
4.22052324e-01 -7.22309649e-01 8.59008506e-02 7.52628446e-01
-4.98311996e-01 1.92712158e-01 -4.43161242e-02 6.77392364e-01
4.05858383e-02 -7.19611168e-01 7.95965016e-01 -4.89135981e-01
-7.35971272e-01 2.04143673e-01 -1.03494358e+00 -7.12475106e-02
1.23492742e+00 -1.21557958e-01 2.04013959e-02 -5.46963930e-01
-9.19264257e-01 2.53880471e-01 3.36962759e-01 1.29548952e-01
1.78446278e-01 -1.03028393e+00 -3.10613692e-01 -1.99660927e-01
-3.45319211e-01 1.89671233e-01 1.88078105e-01 1.04219782e+00
-5.89262664e-01 4.59262937e-01 -5.11544168e-01 -2.85348743e-01
-7.74895906e-01 6.04154289e-01 5.67583740e-01 -6.43255651e-01
-5.90344906e-01 2.52945602e-01 -2.66522691e-02 -5.84782302e-01
2.67094791e-01 -4.18865114e-01 -1.57262668e-01 8.93363431e-02
2.63240904e-01 4.63750064e-01 -3.33087265e-01 -3.65929782e-01
-1.69043764e-02 2.65348107e-01 5.97867444e-02 -5.67253649e-01
1.57881927e+00 -1.29447833e-01 2.62865275e-01 5.28204262e-01
5.02752960e-01 -7.02867150e-01 -1.87427473e+00 -9.34677497e-02
2.23045126e-01 7.68075883e-02 -6.50722859e-03 -3.06461930e-01
-5.10069191e-01 8.37788641e-01 3.63133967e-01 5.28059363e-01
6.71128869e-01 -3.75069439e-01 3.75952691e-01 9.19723272e-01
9.17779028e-01 -1.44218385e+00 2.24710498e-02 7.26403713e-01
6.64306164e-01 -1.04892755e+00 -1.83005825e-01 6.50366545e-02
-7.38366425e-01 1.20967698e+00 5.26671827e-01 -5.82643867e-01
3.89508247e-01 2.86553323e-01 -1.76508471e-01 1.51728451e-01
-9.10022736e-01 -6.43431187e-01 -4.92885500e-01 7.80941486e-01
9.92746931e-03 -1.35651836e-02 -5.28324962e-01 4.85845692e-02
2.45236397e-01 1.45222798e-01 8.91168952e-01 1.37068236e+00
-4.01204467e-01 -1.40814304e+00 -2.37024158e-01 5.98180555e-02
-4.09859270e-01 2.08825499e-01 -1.54343471e-02 1.06410778e+00
-2.11348176e-01 7.79140353e-01 -2.16099605e-01 8.49573091e-02
1.54266462e-01 1.62302494e-01 7.38036633e-01 -5.49864650e-01
-5.43952644e-01 2.75785401e-02 1.72615215e-01 -5.63451529e-01
-6.96737111e-01 -9.73436296e-01 -1.37507796e+00 -2.01809973e-01
-5.53638749e-02 1.04980119e-01 6.68943942e-01 1.14626551e+00
2.11515844e-01 2.21359402e-01 4.42796677e-01 -1.51494443e+00
-1.16357207e+00 -8.70282650e-01 -6.44641876e-01 -4.37887609e-02
2.69878030e-01 -1.15206873e+00 -1.15967132e-01 -4.16785121e-01] | [4.185216426849365, 2.1412529945373535] |
ded0511c-c520-4822-a158-de9823280144 | sample-attackability-in-natural-language | 2306.12043 | null | https://arxiv.org/abs/2306.12043v1 | https://arxiv.org/pdf/2306.12043v1.pdf | Sample Attackability in Natural Language Adversarial Attacks | Adversarial attack research in natural language processing (NLP) has made significant progress in designing powerful attack methods and defence approaches. However, few efforts have sought to identify which source samples are the most attackable or robust, i.e. can we determine for an unseen target model, which samples are the most vulnerable to an adversarial attack. This work formally extends the definition of sample attackability/robustness for NLP attacks. Experiments on two popular NLP datasets, four state of the art models and four different NLP adversarial attack methods, demonstrate that sample uncertainty is insufficient for describing characteristics of attackable/robust samples and hence a deep learning based detector can perform much better at identifying the most attackable and robust samples for an unseen target model. Nevertheless, further analysis finds that there is little agreement in which samples are considered the most attackable/robust across different NLP attack methods, explaining a lack of portability of attackability detection methods across attack methods. | ['Mark Gales', 'Vyas Raina'] | 2023-06-21 | null | null | null | null | ['adversarial-attack'] | ['adversarial'] | [ 1.69187412e-01 1.59055628e-02 -1.64071202e-01 -1.32841662e-01
-1.33674395e+00 -1.56691051e+00 1.22989297e+00 5.54722488e-01
-1.73818961e-01 5.07445216e-01 2.38963485e-01 -5.61823785e-01
-1.23833820e-01 -8.49210322e-01 -5.75363398e-01 -6.98826313e-01
-1.07225031e-01 6.96240962e-01 8.05959031e-02 -9.34010744e-02
2.56115705e-01 1.02885997e+00 -8.58265817e-01 4.40507174e-01
2.45813504e-01 6.25723422e-01 -9.37700987e-01 9.70269561e-01
-1.95981655e-02 6.34691298e-01 -1.14203608e+00 -6.75241709e-01
7.10573852e-01 -2.26462156e-01 -9.70766962e-01 -6.83911622e-01
6.26172245e-01 -4.31764454e-01 -2.89738089e-01 1.40484464e+00
6.71541512e-01 -1.44706249e-01 8.93455148e-01 -1.61459160e+00
-4.26812679e-01 1.02771354e+00 -1.55939907e-01 6.94330335e-01
6.84377193e-01 5.63498616e-01 8.47647011e-01 -3.00037414e-01
6.27724290e-01 1.82470453e+00 4.40434456e-01 8.26784551e-01
-1.19001710e+00 -9.09602165e-01 -1.49211228e-01 -1.46106288e-01
-1.03410208e+00 -6.58741653e-01 7.19989896e-01 -3.50263417e-01
7.40877688e-01 5.18743396e-01 9.96568054e-02 1.95226943e+00
4.67822760e-01 8.61601949e-01 1.21039641e+00 -2.53707647e-01
4.98160094e-01 1.88072592e-01 1.11757420e-01 2.87396252e-01
2.95260578e-01 6.48634911e-01 -2.42352337e-01 -1.04199123e+00
2.13574857e-01 -4.39276963e-01 -2.44770452e-01 9.10829082e-02
-8.31878662e-01 1.22423267e+00 2.96536237e-01 4.61144418e-01
-5.69902100e-02 -5.92038520e-02 8.79026294e-01 5.04736185e-01
3.44776303e-01 1.07075465e+00 -7.08819509e-01 -1.71154607e-02
-6.97674692e-01 6.38877690e-01 1.21994793e+00 7.18529880e-01
1.23288937e-01 2.48312131e-01 -1.46582037e-01 2.02669010e-01
1.94558710e-01 6.99252605e-01 1.28770158e-01 -6.81293070e-01
7.08252132e-01 2.37172559e-01 2.80948132e-01 -1.21657050e+00
-2.21369728e-01 2.94003375e-02 -3.46408963e-01 3.41459095e-01
7.69293189e-01 -6.39420688e-01 -7.47099578e-01 1.60834062e+00
2.72290915e-01 -1.46158174e-01 4.34574634e-01 5.19059062e-01
4.17153239e-01 8.26644897e-01 4.96463478e-01 3.40752527e-02
1.38163650e+00 -9.36239138e-02 -2.49870241e-01 -5.54078102e-01
5.03572702e-01 -8.66006494e-01 8.18648696e-01 3.72557521e-01
-8.38619173e-01 -1.01750180e-01 -1.23245382e+00 5.06769657e-01
-7.60249615e-01 -7.96718895e-01 6.19456410e-01 1.31754398e+00
-2.56729245e-01 6.66913092e-01 -5.74069798e-01 -8.61588493e-02
5.11778653e-01 7.72602186e-02 -5.33284783e-01 -1.63659021e-01
-1.78581977e+00 1.15468001e+00 5.36124945e-01 -3.60425502e-01
-1.55466628e+00 -9.20379996e-01 -7.91506529e-01 -5.85262701e-02
1.92039222e-01 -2.19644725e-01 1.07671297e+00 -9.98755693e-01
-1.13724291e+00 6.76187217e-01 3.47160429e-01 -7.20884144e-01
5.47435701e-01 5.51642329e-02 -7.73388207e-01 3.62947553e-01
-1.62024245e-01 2.64437079e-01 1.36312091e+00 -1.27844310e+00
-3.89206767e-01 -2.15077251e-01 2.48485297e-01 -1.02493644e-01
-3.29647958e-01 9.03011382e-01 5.01756370e-01 -6.85774088e-01
-4.39455122e-01 -7.59688377e-01 -1.36641458e-01 -1.41105950e-01
-7.14062929e-01 -1.44172564e-01 9.45785522e-01 -4.89131629e-01
9.08107400e-01 -2.05512667e+00 -2.42146060e-01 3.71759474e-01
1.78169057e-01 7.73354709e-01 -3.66556197e-01 8.06115568e-01
-1.82954043e-01 7.95974553e-01 -4.74161878e-02 2.79585838e-01
4.64880586e-01 4.78948429e-02 -1.16268861e+00 8.30767393e-01
5.30592859e-01 7.12119222e-01 -1.08574951e+00 -2.10129291e-01
1.59535617e-01 3.70949477e-01 -1.80986583e-01 8.90432820e-02
-2.37972021e-01 2.67282695e-01 -7.70433068e-01 8.50970209e-01
7.09005475e-01 5.64572811e-01 -1.15089536e-01 -3.05502582e-03
3.63234252e-01 3.57225448e-01 -1.05046666e+00 6.68540597e-01
-1.95783406e-01 6.78529084e-01 2.84986764e-01 -6.02779746e-01
9.61701989e-01 4.94353771e-01 -8.38313028e-02 -2.62898773e-01
3.85553449e-01 2.13673294e-01 3.07321966e-01 -2.91022956e-01
3.03953201e-01 -3.02930623e-01 -5.33055663e-01 6.54094815e-01
-3.17044593e-02 -3.17108721e-01 -5.27385287e-02 2.65318215e-01
1.64780247e+00 -4.24285203e-01 3.23752970e-01 -4.15630639e-01
4.85320210e-01 8.63082707e-02 7.15639293e-01 1.13215303e+00
-8.00061762e-01 3.16055208e-01 8.00995827e-01 -5.75331271e-01
-1.14818275e+00 -1.43150866e+00 -2.74537563e-01 8.46712053e-01
-2.84718961e-01 -1.82453617e-01 -8.07877362e-01 -1.15547180e+00
1.64320871e-01 9.30835605e-01 -5.39063573e-01 -4.67119843e-01
-4.99711841e-01 -7.13913739e-01 1.56778049e+00 3.43948662e-01
1.80292204e-01 -1.27801371e+00 -2.74056047e-01 1.06704518e-01
7.63576701e-02 -1.05397451e+00 -2.66090125e-01 1.59609586e-01
-2.79953420e-01 -1.28399670e+00 -1.14980690e-01 -3.63955706e-01
5.87624073e-01 -4.19217169e-01 1.11819160e+00 -1.90668017e-01
-3.80996287e-01 3.60011250e-01 -5.93713880e-01 -9.82313454e-01
-1.41178358e+00 -1.64498806e-01 2.87506729e-01 -1.88849136e-01
7.75345027e-01 -4.47403602e-02 -4.63492312e-02 1.80602252e-01
-1.13978851e+00 -1.21057582e+00 2.13386208e-01 2.78204381e-01
2.59231888e-02 6.20529115e-01 6.54537857e-01 -8.86510849e-01
1.25631225e+00 -7.26965070e-01 -4.33252573e-01 2.46489555e-01
-1.33715451e-01 4.66449447e-02 1.19339323e+00 -8.87735605e-01
-7.34532416e-01 -2.16902196e-01 -2.80939549e-01 -4.08413500e-01
-7.75252819e-01 1.56173393e-01 -5.20730376e-01 -2.57038057e-01
1.30365145e+00 -1.24451205e-01 -2.13161021e-01 -1.51214659e-01
4.60206002e-01 4.43775088e-01 1.36048540e-01 -9.60584104e-01
1.39045489e+00 3.50684494e-01 -6.64291009e-02 -8.02778840e-01
-6.83542728e-01 -1.43213689e-01 -3.40672880e-01 1.96779504e-01
5.98240256e-01 -4.28694993e-01 -5.68724215e-01 4.64676976e-01
-1.19065845e+00 -1.34152159e-01 -2.62715548e-01 2.94587612e-01
-3.56576502e-01 6.37689352e-01 -7.07023561e-01 -8.52763832e-01
-6.09212458e-01 -1.24217951e+00 8.03266287e-01 -2.33195260e-01
-7.52682030e-01 -1.08808541e+00 1.26880243e-01 1.50041133e-01
3.67227793e-01 7.07646847e-01 1.03115785e+00 -1.59514487e+00
1.08140474e-02 -1.07461333e+00 1.29246131e-01 5.06693721e-01
-1.27325997e-01 3.70949626e-01 -1.02265072e+00 -4.87138599e-01
3.24855387e-01 -6.19364202e-01 3.78627032e-01 4.95011695e-02
4.66730326e-01 -8.78622890e-01 -3.93483549e-01 3.13679546e-01
1.27044976e+00 1.77808508e-01 7.45545149e-01 3.03866386e-01
3.20021033e-01 7.72852421e-01 4.76459801e-01 1.49507806e-01
-5.40859818e-01 2.76723146e-01 4.87412423e-01 3.50530565e-01
4.41095829e-01 -3.24493378e-01 8.67047846e-01 -2.63810426e-01
7.86720634e-01 -7.40033925e-01 -1.24298692e+00 4.25704211e-01
-1.24455273e+00 -1.30172002e+00 1.90816879e-01 1.92914498e+00
8.27312052e-01 5.20641446e-01 1.11537382e-01 3.51912826e-01
5.66110551e-01 4.55097944e-01 -4.54631686e-01 -1.12312174e+00
-1.47286743e-01 3.13787878e-01 5.98300278e-01 5.72912753e-01
-1.53088474e+00 9.88324165e-01 7.08571911e+00 1.09732616e+00
-8.04804027e-01 -3.14436197e-01 5.40284216e-01 5.16945682e-02
-2.03149334e-01 3.25383544e-02 -1.03730547e+00 3.44644994e-01
1.37214148e+00 -5.21287143e-01 3.07849914e-01 8.65800798e-01
-2.05103625e-02 5.17454684e-01 -1.38151538e+00 2.94345677e-01
1.86853021e-01 -1.08714998e+00 2.75003523e-01 3.75774950e-02
3.16318423e-01 -1.39292836e-01 1.62960976e-01 3.50982308e-01
1.02567017e+00 -1.09388363e+00 5.73357999e-01 8.55798647e-02
1.59872279e-01 -1.06245697e+00 7.38556087e-01 5.72641671e-01
-6.93579018e-01 -1.93918288e-01 -4.87257212e-01 2.79553264e-01
4.76269834e-02 3.25338513e-01 -9.70866799e-01 2.83691227e-01
4.92518455e-01 1.85827002e-01 -4.43610728e-01 3.82229328e-01
-4.85372603e-01 8.67940605e-01 -3.54915887e-01 -7.99643695e-02
5.83463788e-01 4.37438846e-01 1.20351112e+00 1.34947610e+00
-2.77993798e-01 -6.20039292e-02 3.70252520e-01 9.88727927e-01
-1.28163835e-02 -1.04036167e-01 -1.05874431e+00 -5.42971194e-01
7.84420133e-01 1.25807929e+00 -7.24954188e-01 -9.03021991e-02
1.51494786e-01 4.95468110e-01 6.68333173e-02 1.25900611e-01
-7.31178343e-01 -4.15875763e-01 8.59136999e-01 -1.78484753e-01
-1.20824717e-01 -1.39040187e-01 -3.32360156e-02 -1.00422931e+00
-1.90843523e-01 -1.73309219e+00 6.47597969e-01 -2.63266325e-01
-2.06214976e+00 6.10301256e-01 1.96875364e-01 -1.10341585e+00
-4.71486181e-01 -9.41144228e-01 -9.67414081e-01 1.12243629e+00
-7.24701643e-01 -1.15255761e+00 5.66638112e-01 5.85076392e-01
3.19033146e-01 -5.14712989e-01 1.11091542e+00 -3.20031971e-01
-4.21948105e-01 7.86066532e-01 -1.99644819e-01 7.84057140e-01
5.43691635e-01 -1.16691840e+00 7.97661304e-01 1.26537502e+00
2.44474545e-01 9.10156965e-01 1.07703960e+00 -7.41181433e-01
-1.45041692e+00 -1.01799130e+00 5.60101986e-01 -1.13703239e+00
1.17140102e+00 -4.53764915e-01 -9.21896756e-01 5.95690608e-01
8.00807402e-02 -1.93693504e-01 7.52482891e-01 -1.95798967e-02
-1.07920349e+00 2.84591168e-01 -1.90681148e+00 7.75110185e-01
2.49128550e-01 -7.69308746e-01 -8.66535366e-01 5.40767848e-01
7.38639891e-01 -1.20371059e-01 -8.14843059e-01 2.59470880e-01
1.66783690e-01 -7.65024066e-01 1.29776120e+00 -1.00575662e+00
1.09506302e-01 -9.56881642e-02 -1.72980025e-01 -1.21318841e+00
-2.45440498e-01 -7.82788992e-01 -5.69591708e-02 1.56979704e+00
4.00004536e-01 -9.53850508e-01 7.46275187e-01 9.86317873e-01
2.59609461e-01 -4.47133541e-01 -9.05875802e-01 -9.55424309e-01
8.71038139e-01 -6.02007866e-01 7.55852282e-01 1.26211751e+00
-9.86915752e-02 6.01827651e-02 -6.66397363e-02 6.46779597e-01
8.10738862e-01 -4.67812270e-01 7.14246809e-01 -9.66082036e-01
-1.14951059e-01 -5.69924474e-01 -6.62183523e-01 -1.05127200e-01
5.46140850e-01 -9.06026840e-01 -1.26250356e-01 -8.04605603e-01
-2.24631101e-01 -3.69452596e-01 -7.45124891e-02 5.62589228e-01
-1.46517769e-01 -3.67472395e-02 3.12341481e-01 7.36826435e-02
4.17829156e-02 -6.13918528e-02 5.12030721e-01 -5.27934015e-01
1.41500920e-01 1.63098916e-01 -8.37831676e-01 9.53898549e-01
1.09622025e+00 -7.93271482e-01 -3.37532401e-01 -1.76128209e-01
2.35347047e-01 -3.14335018e-01 4.46487248e-01 -8.77736747e-01
-7.78399268e-03 -3.76328170e-01 5.45305371e-01 -2.14879662e-01
-3.78333293e-02 -8.61172020e-01 -2.06814006e-01 6.64041698e-01
-6.38652086e-01 -4.92388271e-02 5.16921043e-01 6.50815248e-01
1.42345473e-01 -6.70398295e-01 1.03639710e+00 -1.78006157e-01
-3.64622176e-01 5.49732685e-01 -5.29237270e-01 4.98035550e-01
1.12838042e+00 1.70492381e-01 -5.23044765e-01 -1.87208086e-01
-5.79192042e-01 7.40141701e-03 3.37481886e-01 6.38296902e-01
5.15770793e-01 -9.22459304e-01 -1.02158475e+00 2.35546947e-01
-5.05383313e-02 -6.26951635e-01 -1.05971940e-01 -9.44058001e-02
-5.60337126e-01 1.46338999e-01 -7.33331889e-02 -8.20023865e-02
-1.29750550e+00 1.18025208e+00 3.85766149e-01 -5.22154510e-01
-5.29719830e-01 8.05959404e-01 -2.36007541e-01 -4.47469771e-01
2.16327906e-01 1.75892547e-01 1.14083730e-01 1.05625270e-02
9.45230186e-01 4.59610760e-01 -1.92348167e-01 -7.56710887e-01
-5.45334220e-01 -2.44469550e-02 -3.89973968e-01 -5.74428849e-02
7.80779302e-01 5.26613176e-01 -2.54337881e-02 1.04527973e-01
1.31146502e+00 2.19985053e-01 -8.89244914e-01 -1.50557794e-02
9.79026034e-02 -6.60770714e-01 -2.65170753e-01 -8.24367940e-01
-6.91231847e-01 9.32498038e-01 1.39056191e-01 7.44081736e-01
6.28737450e-01 -1.94132373e-01 7.17158318e-01 4.05648172e-01
4.50709403e-01 -6.88540936e-01 -1.71103075e-01 5.53341031e-01
1.08133292e+00 -1.00792658e+00 1.54275611e-01 -1.70199022e-01
-6.43158555e-01 1.09874713e+00 3.85870039e-01 -4.41742837e-01
6.98179603e-01 5.32733202e-01 1.89918831e-01 -2.37768039e-01
-5.26525080e-01 5.15385449e-01 9.52889249e-02 9.89731789e-01
-4.88978736e-02 1.32043764e-01 4.05409075e-02 3.65421653e-01
-4.74184364e-01 -8.33755195e-01 4.97432351e-01 8.44193578e-01
-3.01209182e-01 -1.20977044e+00 -6.87798023e-01 3.62224698e-01
-9.30314362e-01 -1.96107656e-01 -1.07691479e+00 8.66035938e-01
-1.64594099e-01 1.33241701e+00 -2.66626209e-01 -4.47312564e-01
2.36248627e-01 2.42934629e-01 2.59139091e-01 -6.34498596e-01
-1.24219930e+00 -4.72663194e-01 2.47509196e-01 -4.55154985e-01
1.00880265e-01 -6.49737120e-01 -9.54972565e-01 -7.52667487e-01
-2.72948444e-01 2.64857441e-01 5.11777997e-01 8.75488222e-01
1.69529378e-01 -2.03066841e-02 7.22571313e-01 -7.92896748e-01
-1.20184886e+00 -6.75564110e-01 -5.41971326e-01 5.88214993e-01
3.69161129e-01 -1.77726001e-01 -9.77416635e-01 -3.51310432e-01] | [5.974251747131348, 7.972298622131348] |
70ab6c68-2753-4d4d-9079-26bd86df66e7 | negation-detection-for-clinical-text-mining | 2004.04980 | null | https://arxiv.org/abs/2004.04980v1 | https://arxiv.org/pdf/2004.04980v1.pdf | Negation Detection for Clinical Text Mining in Russian | Developing predictive modeling in medicine requires additional features from unstructured clinical texts. In Russia, there are no instruments for natural language processing to cope with problems of medical records. This paper is devoted to a module of negation detection. The corpus-free machine learning method is based on gradient boosting classifier is used to detect whether a disease is denied, not mentioned or presented in the text. The detector classifies negations for five diseases and shows average F-score from 0.81 to 0.93. The benefits of negation detection have been demonstrated by predicting the presence of surgery for patients with the acute coronary syndrome. | ['Ksenia Balabaeva', 'Anastasia Funkner', 'Sergey Kovalchuk'] | 2020-04-10 | null | null | null | null | ['negation-detection'] | ['natural-language-processing'] | [ 1.10804923e-01 3.05160522e-01 -6.23178005e-01 -5.92598140e-01
-3.97957355e-01 -2.11960778e-01 4.06482726e-01 9.88113642e-01
-8.34882557e-01 1.18132925e+00 2.48759866e-01 -9.59729552e-01
-2.77103454e-01 -7.62380600e-01 -8.65227133e-02 -4.02172565e-01
-6.54536560e-02 6.49451256e-01 -2.64382660e-01 -5.07873118e-01
2.25358635e-01 5.69825709e-01 -1.16710818e+00 7.54612625e-01
7.67628014e-01 5.15302658e-01 5.73676191e-02 5.98616302e-01
-1.42941877e-01 1.13349080e+00 -7.33976901e-01 -5.16643822e-01
-4.20624465e-02 -7.11877584e-01 -5.69543481e-01 -2.64856279e-01
-6.41706586e-02 -2.40311533e-01 -8.01855475e-02 1.05105066e+00
3.37610751e-01 -4.76081431e-01 7.34336317e-01 -8.00685346e-01
-1.47999644e-01 6.68871760e-01 1.02773711e-01 5.40974796e-01
6.71696961e-01 -3.22743863e-01 6.99122727e-01 -8.38584602e-01
8.48600328e-01 9.16938722e-01 4.75204706e-01 8.44568491e-01
-8.13953996e-01 -5.48216403e-01 -3.32326502e-01 1.29046977e-01
-1.03699613e+00 -1.84831232e-01 5.53794563e-01 -5.68657994e-01
1.19946516e+00 5.29636681e-01 7.90464222e-01 6.12419188e-01
9.74515736e-01 1.81333944e-01 1.05892158e+00 -8.59372795e-01
2.69637913e-01 3.95250589e-01 4.00468856e-01 8.37462068e-01
7.26712346e-01 3.85883288e-03 -4.08043534e-01 -6.33054554e-01
1.62071913e-01 7.59216547e-02 -2.67354995e-02 3.19784254e-01
-9.86495674e-01 8.51887047e-01 -1.18457608e-01 8.46023977e-01
-5.27994573e-01 -6.30374610e-01 8.70431662e-01 6.27882302e-01
4.34036136e-01 3.38538140e-01 -8.80896270e-01 -2.41430160e-02
-7.82458663e-01 2.39068568e-01 1.14278030e+00 5.74191988e-01
-9.12545994e-02 -1.30719598e-02 -6.79620123e-03 4.71543342e-01
3.07193816e-01 4.20525998e-01 7.38124490e-01 -3.84130448e-01
1.75447300e-01 9.01602507e-01 -5.18321432e-02 -9.75616932e-01
-7.02849865e-01 -3.07878643e-01 -9.06696856e-01 4.71113175e-02
2.80296862e-01 -4.18583453e-01 -7.29894817e-01 1.29077637e+00
2.38374814e-01 -4.46434051e-01 5.40506899e-01 5.28843701e-01
1.26609933e+00 1.57243863e-01 3.61506909e-01 -8.08395028e-01
1.53432131e+00 -3.25150073e-01 -1.41280282e+00 5.82876764e-02
1.18909860e+00 -7.49265850e-01 2.50075459e-01 8.69540930e-01
-9.29110944e-01 -1.40542448e-01 -9.44797397e-01 1.48151875e-01
-4.07894909e-01 2.50218779e-01 7.45503128e-01 6.34560525e-01
-7.91377485e-01 6.13686681e-01 -5.94867945e-01 -2.61456758e-01
2.31566414e-01 5.36767960e-01 -4.80506450e-01 4.34780195e-02
-1.45300961e+00 1.24819434e+00 5.90978682e-01 1.19884461e-01
-1.62879542e-01 -2.23635569e-01 -8.94383788e-01 -7.25966990e-02
2.91556939e-02 -8.54383290e-01 1.06144655e+00 -1.17793059e+00
-9.06937897e-01 1.40628302e+00 -3.21664006e-01 -5.65523446e-01
5.53674221e-01 4.37731631e-02 -9.80905831e-01 1.51996538e-01
1.21451214e-01 -1.84726771e-02 4.71368134e-01 -4.03338373e-01
-7.23316014e-01 -5.11763930e-01 -1.91925645e-01 -9.54601094e-02
-1.04952417e-01 4.20307964e-01 2.14759707e-01 -6.91984832e-01
2.20273271e-01 -4.94120449e-01 -4.36991245e-01 -2.41891474e-01
-2.45106667e-01 -4.75650579e-01 2.89158195e-01 -7.51473367e-01
1.45284903e+00 -1.93228805e+00 -4.75215703e-01 2.34780386e-01
1.85012966e-01 2.97355205e-01 5.71538389e-01 4.21701759e-01
-3.36586773e-01 2.98267633e-01 -2.83090078e-04 5.09797454e-01
-4.39257741e-01 5.90736330e-01 -1.22694425e-01 4.22894180e-01
2.79251188e-01 3.69807988e-01 -8.69843960e-01 -8.59972596e-01
1.05207950e-01 2.31592044e-01 -4.93160725e-01 -8.68894085e-02
3.71788368e-02 1.21819034e-01 -4.84831899e-01 8.67408335e-01
5.29857337e-01 -8.26420337e-02 6.57603562e-01 8.16672891e-02
1.49343058e-01 5.62017620e-01 -8.41803551e-01 1.03645205e+00
2.83753812e-01 1.87383816e-01 1.64156169e-01 -1.19835341e+00
9.53080356e-01 7.92368054e-01 3.36766660e-01 -4.95286226e-01
2.39908203e-01 4.15998131e-01 4.12723392e-01 -1.10512805e+00
1.18233226e-01 -7.13750720e-01 7.54290149e-02 -2.11483434e-01
5.85118644e-02 6.55002296e-02 4.86618042e-01 2.44207919e-01
1.27184176e+00 -2.43852943e-01 1.05737102e+00 -4.66183782e-01
8.83218288e-01 4.03807163e-01 9.93991971e-01 6.75084949e-01
-3.25363785e-01 -3.28893587e-02 7.55001783e-01 -6.81808889e-01
-6.35224760e-01 -4.78614807e-01 -5.73437572e-01 7.18230486e-01
-5.40313959e-01 -5.75531304e-01 -4.19285655e-01 -7.46664882e-01
7.01370090e-02 6.60832644e-01 -6.14818692e-01 -3.11282147e-02
-4.93940413e-01 -9.36612308e-01 3.16763043e-01 2.12189510e-01
-2.41299331e-01 -1.14809775e+00 -6.85517251e-01 4.40342039e-01
-2.35327762e-02 -7.95185506e-01 2.52026498e-01 7.25689054e-01
-1.40241730e+00 -1.30718315e+00 -3.30681354e-01 -1.04404926e+00
6.77960634e-01 -6.50904238e-01 1.21158504e+00 3.43586802e-01
-4.39785123e-01 3.99633758e-02 -3.02656531e-01 -8.35556924e-01
-9.41440403e-01 -3.00095409e-01 1.28584787e-01 -6.19602144e-01
9.06422734e-01 -9.23992544e-02 -2.37537384e-01 -2.61361778e-01
-7.49188244e-01 -3.05609673e-01 6.80559218e-01 1.07278585e+00
4.62740749e-01 -1.80935562e-01 9.11052644e-01 -1.46785247e+00
6.50691211e-01 -5.58301985e-01 -1.88359141e-01 -3.34149338e-02
-9.63917851e-01 -1.27927169e-01 4.26707536e-01 -2.31089950e-01
-7.26366699e-01 3.19975495e-01 -5.70827365e-01 2.51918346e-01
-5.20678043e-01 7.97625661e-01 1.52965635e-01 4.33633387e-01
8.31578672e-01 -9.49871019e-02 1.66814048e-02 -2.21355185e-01
-2.00512677e-01 6.35074079e-01 2.85868704e-01 6.89651072e-02
2.70279516e-02 2.17921615e-01 1.84653863e-01 -9.71129894e-01
-9.24404383e-01 -6.53022468e-01 -4.81996208e-01 -8.66921097e-02
6.99661255e-01 -8.37798834e-01 -6.20869756e-01 -5.86379953e-02
-1.02498019e+00 4.89916921e-01 -2.93885827e-01 1.01759326e+00
-2.23799661e-01 3.16863835e-01 -8.77585948e-01 -8.65010798e-01
-7.33829558e-01 -4.00312454e-01 3.73386294e-01 -2.28027463e-01
-6.19892955e-01 -9.01836514e-01 1.96606889e-01 1.78962588e-01
-3.49570252e-02 4.38405186e-01 1.22828019e+00 -1.36069691e+00
4.05290931e-01 -7.04368770e-01 4.10888970e-01 3.27464879e-01
2.10878953e-01 -4.12772112e-02 -7.78113067e-01 -2.32244786e-02
6.09300673e-01 -1.53123692e-01 6.77939296e-01 4.97059077e-01
7.00064957e-01 -4.21453148e-01 -4.58703637e-01 3.26053500e-02
1.34507966e+00 7.19574928e-01 4.11046356e-01 3.97165865e-01
4.19553928e-02 6.23115361e-01 8.45703721e-01 4.67409730e-01
-9.38665867e-02 -1.84251443e-02 3.09196115e-01 -5.57109155e-02
4.52492654e-01 2.21722186e-01 1.26539553e-02 7.69975007e-01
5.21008819e-02 -1.68238223e-01 -1.22346032e+00 4.48319942e-01
-1.58877039e+00 -8.27033639e-01 -5.61213613e-01 1.63281453e+00
1.00285780e+00 7.83740819e-01 -3.29419464e-01 5.51681161e-01
5.74034274e-01 -5.70518315e-01 -1.76095933e-01 -8.32799852e-01
-2.10174620e-01 2.22457811e-01 2.19107434e-01 4.31117296e-01
-1.10354137e+00 5.72096169e-01 7.11734915e+00 4.15477663e-01
-9.04031873e-01 -3.46694887e-02 6.93215370e-01 1.03535756e-01
7.62531608e-02 -1.70393482e-01 -6.64162338e-01 3.01162839e-01
1.17643702e+00 -4.35788929e-02 -4.58087832e-01 8.81525636e-01
4.88823116e-01 -3.69524449e-01 -1.09201527e+00 7.43822932e-01
2.11141899e-01 -1.23291242e+00 1.88072488e-01 -3.29277329e-02
3.27043474e-01 -2.52897777e-02 -4.63831842e-01 3.13136697e-01
-2.50405312e-01 -9.12988842e-01 4.26371932e-01 8.35257947e-01
5.33122778e-01 -7.29874492e-01 1.41111100e+00 6.64251685e-01
-4.08072978e-01 -1.32625490e-01 -3.12302709e-01 -5.45936406e-01
-4.06966209e-02 9.81852412e-01 -1.34736133e+00 4.86949950e-01
5.66065133e-01 4.79006648e-01 -2.00278983e-01 8.77978563e-01
-1.22698419e-01 9.15703833e-01 -2.01790586e-01 -2.76804835e-01
-7.97757320e-03 -1.03640638e-01 4.43718374e-01 1.55115187e+00
9.58304331e-02 5.93242228e-01 1.29446283e-01 1.66691110e-01
7.25587159e-02 7.77101994e-01 -9.38702643e-01 -8.43037851e-03
4.71046288e-03 9.19365823e-01 -8.11386466e-01 -6.83645129e-01
-4.31941569e-01 6.52170777e-01 -2.84749955e-01 -2.49422580e-01
-3.47552150e-01 -4.34862226e-01 -2.03341562e-02 3.87660205e-01
1.06628150e-01 5.16231000e-01 -5.24722815e-01 -1.06781149e+00
-6.60617501e-02 -1.05092764e+00 7.55686700e-01 -4.76384163e-01
-1.29443800e+00 6.83017254e-01 -3.62113357e-01 -9.95200098e-01
-5.47230601e-01 -8.86644542e-01 -3.42918634e-01 8.08006108e-01
-1.17925882e+00 -7.05276728e-01 3.30804646e-01 6.30605757e-01
2.33967721e-01 -4.48038995e-01 1.41261339e+00 3.01593781e-01
-1.70876756e-01 2.44293898e-01 -3.97516191e-02 4.95123178e-01
8.18238199e-01 -1.19258511e+00 -4.62317497e-01 2.73429066e-01
-3.79427284e-01 6.20175362e-01 9.22269762e-01 -9.43230569e-01
-8.68724883e-01 -7.33565927e-01 2.08018732e+00 -3.03188860e-01
4.49037880e-01 1.88693717e-01 -9.91692662e-01 3.67148906e-01
7.62932375e-02 -9.31704938e-02 1.18987238e+00 5.63019328e-02
1.36809908e-02 -4.84919213e-02 -1.49944127e+00 9.57805961e-02
3.12190682e-01 -3.59697521e-01 -1.16279769e+00 8.05166066e-01
2.09092006e-01 -2.93401986e-01 -9.50508773e-01 5.79373360e-01
2.86503196e-01 -5.36725402e-01 6.15077019e-01 -1.20665109e+00
4.35414374e-01 1.35913968e-01 -6.71466142e-02 -5.32055855e-01
-1.21290050e-01 -3.51831615e-01 -2.33486798e-02 6.70424938e-01
7.62727559e-01 -6.22183502e-01 8.24895203e-01 4.72471148e-01
4.81633991e-02 -7.70885050e-01 -9.44860518e-01 -2.74206549e-01
1.32800072e-01 -3.20246279e-01 -1.77845716e-01 1.39041388e+00
6.65050685e-01 6.23341143e-01 -4.95384075e-03 -2.41425201e-01
2.02002704e-01 -1.04892425e-01 -6.25850931e-02 -1.61158097e+00
7.26205334e-02 -1.59380347e-01 -7.39563227e-01 -1.07752681e-01
-9.98287275e-03 -9.02100265e-01 -3.24140221e-01 -1.65232682e+00
2.96062231e-01 4.28802557e-02 -4.68231827e-01 6.96662843e-01
6.88924342e-02 -1.49521366e-01 -3.16387147e-01 1.17711641e-01
-2.07940176e-01 -1.45252690e-01 8.50123227e-01 -1.21151358e-01
-7.56374747e-02 1.61760017e-01 -6.41543031e-01 1.30542183e+00
1.20102358e+00 -9.82134342e-01 -1.12062823e-02 3.42392534e-01
5.66021919e-01 5.81963539e-01 -2.83868518e-02 -6.91523135e-01
1.49732679e-01 -2.10632294e-01 6.16369486e-01 -9.39481676e-01
-9.92420316e-02 -1.01682615e+00 -7.50826076e-02 1.27725315e+00
-3.94579142e-01 4.12558198e-01 2.15177640e-01 5.58282733e-01
-4.66448396e-01 -7.42464483e-01 5.39416015e-01 -3.42809409e-01
-2.84606785e-01 -4.60253656e-01 -9.94375348e-01 -1.13614602e-02
1.04379618e+00 5.08744232e-02 1.15530379e-01 -1.52394980e-01
-1.43423831e+00 -3.41268666e-02 -5.09673432e-02 -4.23221327e-02
8.28238189e-01 -8.28650057e-01 -9.02882636e-01 1.32674009e-01
1.22887775e-01 -5.90353012e-01 -5.38189448e-02 8.13315630e-01
-8.60764861e-01 7.24333286e-01 -1.41803429e-01 -3.27781886e-01
-1.91214406e+00 7.81527221e-01 2.55068749e-01 -5.69149494e-01
-5.87161064e-01 4.93859291e-01 -4.97528374e-01 -3.93853486e-01
3.23787898e-01 -5.78732371e-01 -7.69404292e-01 2.19741896e-01
5.01832128e-01 7.34192580e-02 6.32731736e-01 -5.18962979e-01
-6.89273953e-01 -2.19360348e-02 -2.96307921e-01 1.34364456e-01
1.39422309e+00 4.58296001e-01 -5.75715005e-01 6.23909652e-01
7.85443604e-01 8.16594660e-02 2.22693682e-01 2.01030858e-02
4.38529938e-01 1.26400679e-01 1.09035954e-01 -1.26078141e+00
-5.22699416e-01 3.84282380e-01 7.39633083e-01 2.36099586e-01
1.12464702e+00 -2.44452223e-01 1.92994967e-01 9.90439951e-01
6.83020893e-03 -1.19444168e+00 -6.41607881e-01 5.40187776e-01
6.89102411e-01 -1.39656448e+00 3.09455514e-01 -4.06899363e-01
-6.52221322e-01 1.43067789e+00 7.58457631e-02 -5.11320531e-02
1.25005209e+00 6.56771541e-01 2.37106949e-01 -4.51076239e-01
-1.09798110e+00 2.72770017e-01 1.39827177e-01 3.24779093e-01
9.31466937e-01 2.28560567e-01 -1.33734119e+00 7.86921918e-01
-1.08993836e-02 5.54806709e-01 5.73440254e-01 1.29732025e+00
-4.91276234e-01 -1.13934994e+00 -4.64996517e-01 9.36581254e-01
-1.35175455e+00 -2.88822711e-01 -6.18740916e-01 7.94847548e-01
5.25156021e-01 9.61376607e-01 -2.44926080e-01 -1.19984381e-01
5.02939522e-01 4.33585554e-01 2.62086451e-01 -6.48722887e-01
-1.09763813e+00 2.66465545e-01 4.38481688e-01 -1.83905780e-01
-8.33971620e-01 -6.92926407e-01 -1.49038303e+00 -1.04919687e-01
-4.37217474e-01 6.16999686e-01 6.21381462e-01 8.16917837e-01
-8.88969600e-02 6.03583217e-01 -2.30459087e-02 2.19172537e-01
-5.55102944e-01 -9.33454335e-01 -7.06951976e-01 2.60344952e-01
5.14654815e-01 -2.29603261e-01 -3.85631770e-01 3.12997907e-01] | [8.447580337524414, 8.766731262207031] |
785eb824-eabb-4ae9-bd8a-602d8c28a850 | semi-supervised-learning-with-constraints-for | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Bauml_Semi-supervised_Learning_with_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Bauml_Semi-supervised_Learning_with_2013_CVPR_paper.pdf | Semi-supervised Learning with Constraints for Person Identification in Multimedia Data | We address the problem of person identification in TV series. We propose a unified learning framework for multiclass classification which incorporates labeled and unlabeled data, and constraints between pairs of features in the training. We apply the framework to train multinomial logistic regression classifiers for multi-class face recognition. The method is completely automatic, as the labeled data is obtained by tagging speaking faces using subtitles and fan transcripts of the videos. We demonstrate our approach on six episodes each of two diverse TV series and achieve state-of-the-art performance. | ['Rainer Stiefelhagen', 'Martin Bauml', 'Makarand Tapaswi'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['person-identification'] | ['computer-vision'] | [ 3.60092849e-01 -2.24868670e-01 -4.52810228e-01 -1.24692178e+00
-1.31499767e+00 -8.11966002e-01 5.96878588e-01 -4.28265691e-01
-3.90512437e-01 7.52276301e-01 -1.59026995e-01 1.62524804e-01
1.40769318e-01 -3.63487780e-01 -5.91655493e-01 -6.13592863e-01
-2.85341054e-01 5.63416243e-01 -2.42320940e-01 2.38263890e-01
-2.53072649e-01 2.24001825e-01 -1.72108805e+00 9.77083206e-01
1.75646558e-01 1.50221515e+00 -4.32948321e-01 9.30692971e-01
3.09245914e-01 9.27912414e-01 -6.10792160e-01 -8.26606452e-01
2.67173946e-01 -3.16610634e-01 -7.02685416e-01 2.99672306e-01
1.09796095e+00 -3.60844582e-01 -3.09778899e-01 7.27203250e-01
4.51104611e-01 -4.46852297e-02 1.17112947e+00 -1.67746830e+00
-5.34193873e-01 2.83238679e-01 -4.25102174e-01 2.52566278e-01
8.50001037e-01 -5.25388777e-01 1.03991115e+00 -8.72861505e-01
2.76464045e-01 1.61476433e+00 8.98916066e-01 9.25365925e-01
-1.34874558e+00 -1.06249595e+00 1.90924764e-01 2.40447819e-01
-1.58493876e+00 -8.33719254e-01 3.55992109e-01 -6.99272871e-01
5.76874554e-01 3.54751796e-01 4.30971891e-01 1.66011035e+00
-1.60632953e-01 8.13015699e-01 1.06502068e+00 -6.99039102e-01
-9.36961025e-02 5.04509985e-01 3.89114857e-01 8.82936656e-01
-3.00492525e-01 1.89275127e-02 -7.91832745e-01 -4.86611187e-01
6.10678017e-01 -2.13210329e-01 5.98850995e-02 -3.29078138e-02
-4.85368073e-01 8.67168546e-01 -2.99011558e-01 1.11666195e-01
1.72531396e-01 2.11517245e-01 3.83313656e-01 5.47182322e-01
6.77452981e-01 -3.13814789e-01 -4.50678408e-01 -9.77058988e-03
-1.05869019e+00 4.19784896e-02 9.91021335e-01 9.73636985e-01
5.69359660e-01 -1.42782688e-01 -2.32913092e-01 1.21536422e+00
3.11569840e-01 6.43606186e-01 3.51058841e-01 -1.00135314e+00
2.41630927e-01 1.23207338e-01 -4.92356941e-02 -5.17186761e-01
-3.07785183e-01 3.34802344e-02 -2.83719748e-01 -2.43124083e-01
4.87963170e-01 -3.01260293e-01 -8.32954168e-01 1.84514201e+00
9.64718387e-02 6.43414438e-01 -1.75393924e-01 3.63925695e-01
1.02797818e+00 4.81379867e-01 1.57830298e-01 -4.84195471e-01
1.22879779e+00 -8.90205622e-01 -6.51250780e-01 -3.86732407e-02
3.92527789e-01 -4.91671652e-01 3.56776714e-01 5.30826271e-01
-7.59309530e-01 -6.61228538e-01 -7.54672647e-01 2.26459846e-01
-2.35780731e-01 6.07295871e-01 3.73044223e-01 1.35794210e+00
-1.08157492e+00 3.85264814e-01 -4.62015927e-01 -3.93543124e-01
5.86781383e-01 7.46556401e-01 -5.44990420e-01 -1.65034682e-01
-9.45370853e-01 4.15948778e-01 -1.93902805e-01 5.74495159e-02
-1.23126447e+00 -3.18535626e-01 -9.65102315e-01 -2.25978822e-01
1.33518368e-01 -2.14514751e-02 1.61839080e+00 -1.44565845e+00
-1.31948590e+00 1.36958361e+00 -5.88940501e-01 -3.82235162e-02
4.27690476e-01 -8.30572098e-02 -8.60381842e-01 4.27476525e-01
2.72748888e-01 4.43424225e-01 1.35322714e+00 -1.10545480e+00
-8.91009331e-01 -5.73413491e-01 -2.48421002e-02 -1.44441620e-01
-7.79184997e-01 6.83293700e-01 -4.83990073e-01 -8.01544309e-01
-1.85255557e-01 -1.09874427e+00 3.47426802e-01 -8.65823776e-02
-3.67050946e-01 -4.46796298e-01 8.50888789e-01 -7.26700664e-01
8.58111739e-01 -2.16854072e+00 1.41330644e-01 3.34000647e-01
-1.34353921e-01 -3.01235139e-01 -2.41223410e-01 -4.21852730e-02
-1.63154125e-01 -4.25036345e-03 1.93736315e-01 -1.07212055e+00
9.04629230e-02 1.12194344e-01 -3.26219529e-01 6.37666047e-01
2.26623341e-01 5.02022922e-01 -5.41543901e-01 -6.10466301e-01
-1.25951856e-01 5.04115045e-01 -3.80616426e-01 4.27172393e-01
1.77314848e-01 6.36814773e-01 -3.34512681e-01 1.19092202e+00
3.59678775e-01 -1.78382903e-01 4.55339909e-01 -1.74750194e-01
3.31621349e-01 -7.77418688e-02 -1.15235066e+00 1.27362537e+00
-2.57475734e-01 8.37978184e-01 1.30151078e-01 -1.09694314e+00
6.05657160e-01 8.28191340e-01 5.17367601e-01 -1.86874211e-01
1.83028191e-01 1.06373668e-01 -6.30834341e-01 -7.15857089e-01
-8.73597153e-03 -1.61996603e-01 -3.11481863e-01 4.15329337e-01
7.90726781e-01 2.42459461e-01 2.01255068e-01 -1.98579669e-01
7.33997285e-01 1.67152107e-01 5.53266816e-02 1.81903299e-02
6.45612121e-01 -5.41839898e-01 4.73895401e-01 9.33370888e-01
-3.46011877e-01 5.72252691e-01 4.72336024e-01 -5.65788031e-01
-6.96249843e-01 -8.89714837e-01 -4.60342616e-01 1.85812211e+00
-3.10260177e-01 -4.25227731e-01 -7.61022747e-01 -8.98302555e-01
-3.07593873e-04 2.88352575e-02 -8.88158560e-01 3.03411305e-01
-6.62322283e-01 -7.92136669e-01 7.43347585e-01 6.79017782e-01
6.59335777e-02 -6.59699023e-01 8.16834718e-02 -1.38117969e-01
-2.57506341e-01 -1.57033718e+00 -4.58622426e-01 1.71794683e-01
-3.90622109e-01 -1.24951196e+00 -6.44422591e-01 -1.37058866e+00
5.48479140e-01 2.64098365e-02 1.04441977e+00 -3.34940106e-02
-5.49197197e-01 9.77743149e-01 -2.27231115e-01 -1.72488749e-01
-2.57531375e-01 -1.97354868e-01 5.23378909e-01 6.85475349e-01
5.78076065e-01 -3.29772681e-01 1.15248915e-02 6.43968523e-01
-3.15811485e-01 -7.14967847e-01 1.85871292e-02 8.03956747e-01
3.40515733e-01 -2.33933538e-01 6.60491526e-01 -6.81479096e-01
2.05356479e-01 -4.54595029e-01 -3.46536398e-01 7.00473607e-01
1.52074516e-01 -5.48400581e-01 3.53071511e-01 -8.72381091e-01
-8.72043848e-01 4.62439090e-01 1.95059367e-02 -3.62805456e-01
-4.11106974e-01 -3.20623741e-02 -1.42400682e-01 -5.88974774e-01
4.34589207e-01 3.81597271e-03 -4.16582644e-01 -3.89473170e-01
1.35745734e-01 1.10840797e+00 5.33122122e-01 -5.80224335e-01
4.91861612e-01 5.66184163e-01 -3.14066619e-01 -1.17585266e+00
-1.04667723e+00 -5.45171797e-01 -9.15927529e-01 -6.87340081e-01
9.75646973e-01 -1.39014041e+00 -1.08107007e+00 4.25484270e-01
-1.00723636e+00 -1.75503761e-01 2.42496490e-01 5.67211330e-01
-6.56998634e-01 2.11228207e-01 -9.59499240e-01 -1.37086892e+00
4.97965747e-03 -9.29303527e-01 1.29945159e+00 1.22736372e-01
-2.07374677e-01 -8.13909829e-01 1.01951234e-01 6.27197921e-01
-2.05488890e-01 -9.34719443e-02 5.54727495e-01 -1.00602937e+00
-2.49997973e-01 -6.36485219e-01 -5.18862531e-02 3.76513571e-01
3.51368301e-02 -1.18113637e-01 -1.67859590e+00 -5.47937751e-01
-2.64638802e-03 -1.23634720e+00 8.56673002e-01 4.99488771e-01
1.29384124e+00 -1.95852861e-01 -5.85215151e-01 5.14451921e-01
1.06161308e+00 1.92138359e-01 4.95201498e-02 -8.07140097e-02
5.00038981e-01 9.22845423e-01 2.89881885e-01 4.84177858e-01
4.00010526e-01 8.11926603e-01 -8.33946168e-02 2.99868107e-01
3.00356865e-01 -6.06494881e-02 5.34006953e-01 2.63278633e-01
-1.31499946e-01 -2.87178755e-01 -6.64497137e-01 3.44075531e-01
-1.59966350e+00 -1.16378188e+00 2.45746091e-01 2.05377173e+00
6.04058504e-01 -2.11080149e-01 6.32772207e-01 9.69143510e-02
1.02638352e+00 1.62096582e-02 -1.77178875e-01 -2.44879927e-02
-1.18311912e-01 1.68272510e-01 1.32151455e-01 3.73982280e-01
-1.84463179e+00 6.24653995e-01 8.21526241e+00 7.61346936e-01
-7.18309939e-01 3.07791859e-01 9.89410877e-01 -4.39396977e-01
5.24211049e-01 -6.13362849e-01 -1.24520004e+00 2.75318503e-01
1.15581107e+00 4.61877793e-01 5.41510642e-01 8.93535733e-01
-2.17321485e-01 1.54472813e-01 -1.63099754e+00 1.46095788e+00
8.47010255e-01 -8.12246680e-01 -2.97455758e-01 -1.17523074e-01
6.46130204e-01 -2.43166044e-01 2.88434833e-01 4.30164456e-01
1.48150206e-01 -1.21406627e+00 8.79061818e-01 2.36103058e-01
1.13369620e+00 -4.09554571e-01 3.70166034e-01 1.74922533e-02
-1.42663169e+00 -6.57549381e-01 -4.15216237e-02 2.19395176e-01
1.74607024e-01 -8.89367908e-02 -4.31167781e-01 2.79094160e-01
9.57039773e-01 9.26861823e-01 -5.45600951e-01 8.10185313e-01
-1.31130209e-02 6.08556628e-01 -1.64282113e-01 1.53026953e-01
-2.53001124e-01 -6.29886538e-02 1.61901161e-01 1.28003943e+00
2.44768366e-01 1.72221154e-01 6.56609237e-01 9.46749300e-02
-3.07864755e-01 1.22276656e-01 -5.35449624e-01 1.27270445e-01
2.95330524e-01 1.22355080e+00 -5.73867857e-01 -4.78102356e-01
-8.50392044e-01 9.98747408e-01 4.09959227e-01 3.61957192e-01
-8.29059422e-01 1.11212127e-01 4.52501923e-01 -2.70189375e-01
5.34930944e-01 1.58888280e-01 1.82205066e-01 -1.31127822e+00
2.60263737e-02 -8.37450862e-01 9.13952112e-01 -3.53816658e-01
-1.71440208e+00 1.01809680e+00 8.65582973e-02 -1.15857446e+00
-5.46382785e-01 -9.28275526e-01 -3.00129741e-01 4.23443288e-01
-1.16097438e+00 -1.48045087e+00 -2.10351959e-01 6.85924053e-01
6.58800483e-01 -6.42552733e-01 1.16417158e+00 7.93821573e-01
-7.35294998e-01 1.04300964e+00 1.43914655e-01 5.07150710e-01
1.06089699e+00 -9.68308628e-01 6.62737852e-03 3.25940251e-01
5.09881377e-01 1.25285298e-01 3.78134638e-01 -4.13924694e-01
-9.87875223e-01 -9.57016528e-01 8.61766696e-01 -7.65464544e-01
5.82519770e-01 -8.40159297e-01 -3.32951784e-01 1.17318380e+00
-1.79855481e-01 1.91147313e-01 1.45080090e+00 2.70244628e-01
-6.97040379e-01 -8.30379948e-02 -1.21117067e+00 -4.80811819e-02
9.75022852e-01 -1.02134299e+00 -3.20270360e-01 7.86153793e-01
2.40294829e-01 -7.42858052e-02 -8.34826887e-01 2.11940795e-01
1.27528203e+00 -5.74266672e-01 1.12013662e+00 -9.49461997e-01
7.21110851e-02 3.49486768e-01 -7.31077194e-01 -7.50049829e-01
-9.73155200e-02 -5.21472812e-01 -1.69223696e-01 1.33126283e+00
6.83834255e-01 -3.60575676e-01 7.28733599e-01 7.46241212e-01
3.97751391e-01 -3.15924197e-01 -1.29570270e+00 -8.68421197e-01
-2.37286106e-01 -6.18201435e-01 2.36368388e-01 8.56564045e-01
1.15048714e-01 3.46409857e-01 -9.44920838e-01 1.83572277e-01
8.58698666e-01 1.14974536e-01 4.80040938e-01 -1.50583279e+00
-4.69618291e-01 8.97636563e-02 -5.33448219e-01 -9.46698964e-01
8.39042604e-01 -6.28771603e-01 -7.68559948e-02 -5.50794959e-01
5.73181033e-01 -2.46849746e-01 -1.82246074e-01 6.16266072e-01
1.49214581e-01 1.13825977e+00 6.30207136e-02 2.41105214e-01
-9.57070529e-01 3.12527448e-01 2.81487823e-01 -4.31197345e-01
-1.43101504e-02 3.11754346e-01 -2.26172835e-01 7.98915982e-01
4.18770224e-01 -5.85317850e-01 1.64805315e-02 -3.53594333e-01
-4.54539180e-01 1.16776019e-01 4.10887331e-01 -8.01686466e-01
6.66347593e-02 1.53368413e-01 8.31972778e-01 -3.06774706e-01
1.07282078e+00 -7.87551165e-01 -7.56328553e-02 1.42783478e-01
-6.18358791e-01 -2.30637431e-01 1.10815711e-01 7.82029927e-01
-8.52443129e-02 -3.49148989e-01 6.14051402e-01 7.64467241e-03
-3.49678397e-01 3.59667361e-01 -4.38848853e-01 -7.62499198e-02
1.02157998e+00 -2.79569123e-02 -1.18139230e-01 -5.36088645e-01
-1.38166571e+00 1.99441403e-01 1.16758749e-01 5.24279296e-01
3.04198742e-01 -1.58632696e+00 -7.46528387e-01 4.68436450e-01
1.35555938e-01 -1.02943885e+00 1.34739280e-01 5.43960452e-01
2.16701373e-01 3.76632750e-01 -1.16964988e-01 -7.59325624e-01
-1.92952561e+00 3.18616956e-01 4.52603132e-01 1.18167609e-01
5.61056212e-02 1.11202192e+00 1.47028789e-01 -5.19327521e-01
7.71313846e-01 1.38824180e-01 -5.63589633e-01 4.90567327e-01
9.56670642e-01 3.12492520e-01 -1.31340861e-01 -1.06158543e+00
-5.65672934e-01 6.85406446e-01 -1.15191370e-01 -3.85678619e-01
1.25327122e+00 -1.58104405e-01 -7.72198290e-02 7.86293805e-01
1.53289056e+00 -3.47033083e-01 -1.02072489e+00 -2.12630764e-01
-1.61832243e-01 -5.79234660e-01 -3.25191110e-01 -4.71676677e-01
-7.86271095e-01 5.24419963e-01 8.57658327e-01 3.62255365e-01
8.72398674e-01 2.80917466e-01 2.66596109e-01 4.16958719e-01
4.81458753e-01 -1.11233056e+00 2.79575765e-01 3.29077952e-02
6.60914481e-01 -1.63281870e+00 -1.11161500e-01 -7.76721776e-01
-3.59500915e-01 1.12121606e+00 4.84302759e-01 9.50341895e-02
9.05495346e-01 4.39341702e-02 3.61862220e-02 9.62763131e-02
-6.42500162e-01 5.77567816e-02 4.70386505e-01 7.41726398e-01
5.18766224e-01 6.17884733e-02 2.27071732e-01 8.95160913e-01
-1.68078151e-02 -5.54898158e-02 2.64548380e-02 1.00498915e+00
-9.32211056e-02 -1.08450210e+00 -5.78038931e-01 6.94202244e-01
-7.97466934e-01 1.65230736e-01 -3.61914843e-01 4.06142265e-01
1.02889568e-01 1.44372702e+00 1.82108954e-01 -4.66136515e-01
2.78540570e-02 4.22471553e-01 8.18057537e-01 -6.30531669e-01
-4.14455295e-01 2.50341982e-01 4.55580115e-01 -4.31642443e-01
-8.20574820e-01 -9.89741445e-01 -4.04351473e-01 -2.36449726e-02
-3.71727198e-01 8.66622403e-02 3.86939794e-01 1.07370448e+00
-1.73914701e-01 7.72736892e-02 1.11908102e+00 -1.10387433e+00
-5.68444371e-01 -1.01519370e+00 -8.01978111e-01 3.04594189e-01
5.26796997e-01 -8.40780675e-01 -4.71965402e-01 5.08183002e-01] | [13.595293998718262, 1.2800146341323853] |
106d4f81-647c-47fc-a96f-5a6b8b7a426b | confluence-of-artificial-intelligence-and | 2012.08545 | null | https://arxiv.org/abs/2012.08545v2 | https://arxiv.org/pdf/2012.08545v2.pdf | Accelerated, Scalable and Reproducible AI-driven Gravitational Wave Detection | The development of reusable artificial intelligence (AI) models for wider use and rigorous validation by the community promises to unlock new opportunities in multi-messenger astrophysics. Here we develop a workflow that connects the Data and Learning Hub for Science, a repository for publishing AI models, with the Hardware Accelerated Learning (HAL) cluster, using funcX as a universal distributed computing service. Using this workflow, an ensemble of four openly available AI models can be run on HAL to process an entire month's worth (August 2017) of advanced Laser Interferometer Gravitational-Wave Observatory data in just seven minutes, identifying all four all four binary black hole mergers previously identified in this dataset and reporting no misclassifications. This approach combines advances in AI, distributed computing, and scientific data infrastructure to open new pathways to conduct reproducible, accelerated, data-driven discovery. | ['Ian Foster', 'Ben Blaiszik', 'Dawei Mu', 'Volodymyr Kindratenko', 'Daniel S. Katz', 'Maeve Heflin', 'Wei Wei', 'Ryan Chard', 'Maksim Levental', 'Minyang Tian', 'Xiaobo Huang', 'Asad Khan', 'E. A. Huerta'] | 2020-12-15 | null | null | null | null | ['gravitational-wave-detection'] | ['miscellaneous'] | [-3.99909973e-01 -2.16650605e-01 5.18789440e-02 -4.04755957e-02
-6.31158590e-01 -6.99562132e-01 7.82972038e-01 2.74755627e-01
-2.00137004e-01 5.69334030e-01 -1.15879506e-01 -8.56200278e-01
-3.97453278e-01 -6.07248604e-01 -5.59855938e-01 -8.19935560e-01
-4.65372771e-01 1.45775998e+00 2.08607584e-01 -4.39870059e-02
4.42404628e-01 9.64162886e-01 -1.62702096e+00 5.01386881e-01
4.98329461e-01 1.01098680e+00 4.99371812e-02 8.35362494e-01
5.00848889e-02 9.76951540e-01 -2.44328991e-01 -1.22272201e-01
6.67574465e-01 -1.31252378e-01 -9.98345435e-01 -8.69700491e-01
6.63597137e-02 2.71233350e-01 -1.40578732e-01 6.57756090e-01
7.37337589e-01 -1.85146723e-02 1.54348463e-02 -1.29098034e+00
-5.85068166e-01 5.75324774e-01 -2.86730945e-01 7.61128008e-01
-9.92636755e-02 9.14754033e-01 1.01724172e+00 -7.52626657e-01
7.57960558e-01 1.08789718e+00 6.01354122e-01 6.45685643e-02
-9.13482249e-01 -7.54393101e-01 -8.02804410e-01 5.18687487e-01
-1.15451634e+00 -6.47458613e-01 2.72374123e-01 -6.65758848e-01
1.60991597e+00 4.86360729e-01 9.92374361e-01 7.79967666e-01
3.95659685e-01 3.63210618e-01 9.16348815e-01 -4.52408820e-01
7.66802669e-01 -3.45335931e-01 -6.64185509e-02 8.13211739e-01
5.77842951e-01 3.85243833e-01 -1.07040727e+00 -6.41195893e-01
2.50888616e-01 -3.10893714e-01 1.85251817e-01 -2.08509907e-01
-1.52493942e+00 5.61718345e-01 2.01264322e-01 3.05539995e-01
-6.33334160e-01 8.69719535e-02 3.73381555e-01 2.65838712e-01
6.48085952e-01 1.33086801e+00 -1.31539273e+00 -6.12408280e-01
-4.57999438e-01 6.08579874e-01 8.57847333e-01 2.99778372e-01
8.27089906e-01 1.29458174e-01 4.29916918e-01 4.66744304e-01
2.60433316e-01 5.80921710e-01 6.89977348e-01 -1.14565158e+00
-2.22126737e-01 8.07996988e-01 8.77240598e-02 -6.77834153e-01
-8.32523644e-01 -5.92928112e-01 -3.89776260e-01 4.18489814e-01
9.70653296e-02 -3.03750157e-01 -5.81782997e-01 8.42126667e-01
8.98834229e-01 6.55785382e-01 2.35542834e-01 1.11324608e+00
8.92714679e-01 3.29940885e-01 2.59791434e-01 2.67922819e-01
1.31767535e+00 -9.01086986e-01 1.65361509e-01 -2.87991166e-01
1.09146690e+00 -5.48292160e-01 5.42463362e-01 6.05829775e-01
-9.24810588e-01 -3.07339102e-01 -1.03291416e+00 -2.10781202e-01
-6.44959807e-01 -3.73217344e-01 1.41554630e+00 2.23793432e-01
-1.00292718e+00 6.19748235e-01 -1.14186954e+00 -2.91194707e-01
4.67713207e-01 2.85436302e-01 -9.82558057e-02 2.29175150e-01
-7.59735823e-01 1.05783975e+00 4.72064286e-01 -4.31547791e-01
-1.10599983e+00 -1.42513633e+00 -1.58919156e-01 7.96352625e-02
-9.47305113e-02 -1.19061160e+00 1.29426420e+00 -7.04789102e-01
-1.24655318e+00 9.33367968e-01 2.53346294e-01 -8.03715110e-01
-1.43949002e-01 2.70495802e-01 -4.88060355e-01 1.92867015e-02
1.56809688e-01 4.33175147e-01 3.07501197e-01 -5.69105923e-01
-7.73420811e-01 -7.01517582e-01 -4.30430532e-01 -2.67714113e-01
1.58113353e-02 4.26200509e-01 1.20835803e-01 -2.84688890e-01
2.15499163e-01 -1.01522624e+00 -2.55702466e-01 -1.44671664e-01
3.50566685e-01 -4.66212124e-01 6.82413578e-01 -7.27956474e-01
4.15869474e-01 -1.81258869e+00 3.71895730e-01 -1.17929079e-01
3.81434202e-01 3.45965415e-01 -5.34609146e-02 3.34196806e-01
-2.36954406e-01 2.41039038e-01 -1.20382965e-01 -1.31375760e-01
-1.89412028e-01 7.38309324e-02 -2.15735003e-01 3.82146925e-01
9.46534127e-02 9.89993095e-01 -9.50719714e-01 1.46773964e-01
-6.15971386e-02 6.28385022e-02 -3.78466249e-01 1.49290636e-01
-6.48817360e-01 1.13506281e+00 -3.60261500e-01 8.97488952e-01
5.88703513e-01 -4.92807269e-01 9.07207355e-02 3.54890823e-01
-6.94097877e-01 4.53003645e-01 -7.03517437e-01 1.96448147e+00
-2.16585755e-01 4.08762783e-01 2.25376770e-01 -1.00512409e+00
8.52703810e-01 2.68403023e-01 9.00123298e-01 -7.05602825e-01
5.81483021e-02 4.67479408e-01 4.03547496e-01 -6.27954900e-01
-3.09930909e-02 -1.46455824e-01 3.28770190e-01 8.80428135e-01
3.77424002e-01 -2.76465237e-01 3.05794980e-02 2.26164814e-02
1.55923271e+00 5.49502447e-02 1.70575525e-03 -4.47926313e-01
8.03511664e-02 4.92574364e-01 6.97499931e-01 7.90992856e-01
-7.60033056e-02 1.18979573e-01 6.07116781e-02 -1.47729278e+00
-1.43436897e+00 -6.77693605e-01 -3.23903322e-01 1.28408480e+00
-5.69485247e-01 -5.69312751e-01 -3.91727090e-02 -1.84193879e-01
3.08416307e-01 6.19406700e-01 -1.59475967e-01 -1.74798161e-01
-2.57051643e-02 -1.18061817e+00 6.07737839e-01 -1.19247139e-01
3.73066068e-01 -1.20041335e+00 -6.96120381e-01 2.04344541e-02
9.11059827e-02 -6.84322238e-01 7.51358807e-01 -6.67037889e-02
-5.57634056e-01 -1.08090723e+00 1.39009461e-01 -2.74790287e-01
-4.80399318e-02 9.53981951e-02 1.30196226e+00 4.07288045e-01
-9.86451626e-01 1.13841265e-01 -3.07009339e-01 -1.20680130e+00
-5.65111458e-01 6.38794824e-02 3.81326139e-01 -2.31173247e-01
6.59268141e-01 -6.88986897e-01 -4.30760890e-01 -2.33086348e-01
-6.16327405e-01 4.95230198e-01 5.89387596e-01 4.09037441e-01
1.76433191e-01 -2.24117041e-01 7.87884176e-01 -2.87476271e-01
-9.50602069e-02 -1.30833316e+00 -1.21005130e+00 1.19922370e-01
-9.90520835e-01 2.22289413e-02 3.62213224e-01 1.67984769e-01
-8.86610985e-01 -7.92502053e-03 -1.64367016e-02 -3.38020355e-01
-2.27795422e-01 8.00516784e-01 3.95791888e-01 -4.86950755e-01
8.56643736e-01 -1.84722364e-01 2.14890331e-01 -7.35646665e-01
4.03296947e-01 8.38533640e-01 4.00493145e-01 -6.05136275e-01
5.34811318e-01 4.77257878e-01 2.57891625e-01 -7.83890188e-01
-6.47291303e-01 -6.91993594e-01 -3.80738378e-01 -1.74750369e-02
6.45101905e-01 -1.07561219e+00 -1.05214405e+00 3.03639233e-01
-9.76584375e-01 -3.08324277e-01 -3.28997016e-01 6.18929029e-01
-2.54053295e-01 -7.06899017e-02 -1.76158920e-01 -7.06302762e-01
-7.03785598e-01 -7.77384818e-01 1.00199938e+00 2.95802504e-01
-3.46038714e-02 -6.31751060e-01 5.99047542e-01 6.93170309e-01
7.73134828e-01 -3.66347805e-02 8.82901371e-01 -1.20091259e+00
-8.22892547e-01 -1.75677657e-01 -1.22735500e-01 -9.07294378e-02
-4.34012443e-01 1.28986165e-01 -9.54482496e-01 -1.61611095e-01
5.33890314e-02 -4.32364136e-01 6.77841723e-01 -8.82002488e-02
1.32765377e+00 -1.54844120e-01 -4.03431684e-01 6.55426204e-01
9.99061286e-01 1.39411137e-01 2.83121318e-01 6.00663483e-01
6.08462691e-01 2.29075432e-01 1.94031343e-01 7.75904417e-01
2.12990746e-01 3.37669462e-01 3.34909678e-01 -6.75431415e-02
1.15175299e-01 4.30902302e-01 -2.04468265e-01 7.79827476e-01
-4.27329481e-01 4.00048643e-01 -1.83432341e+00 4.33254272e-01
-1.87214530e+00 -1.09652984e+00 -3.81405652e-01 2.37934828e+00
7.40890086e-01 -2.59653091e-01 -1.91253260e-01 -4.60286468e-01
1.02499016e-01 -2.82603025e-01 -8.32751155e-01 -3.87471914e-01
-2.16989174e-01 7.45913327e-01 5.37852407e-01 2.54071444e-01
-7.61040032e-01 8.90089095e-01 6.96130991e+00 4.85859513e-01
-1.34432232e+00 8.09146345e-01 7.79624343e-01 -4.57493663e-01
-3.93139273e-01 5.00261068e-01 -3.40384096e-01 4.35723513e-01
1.87383854e+00 -6.57948077e-01 9.70228136e-01 9.52637613e-01
2.00886592e-01 4.94933501e-02 -5.38724303e-01 7.55826116e-01
-2.31827244e-01 -2.17360330e+00 -3.89535815e-01 -2.46258304e-02
6.85534418e-01 1.43998504e+00 -4.80768204e-01 3.05805504e-01
6.73687756e-01 -7.97137499e-01 2.71190763e-01 9.28460360e-01
3.88399035e-01 -5.80729544e-01 5.39153934e-01 4.48619425e-01
-3.97888452e-01 -3.38363379e-01 -6.14619911e-01 -7.14028776e-01
-4.17882293e-01 8.90765607e-01 -1.14991224e+00 7.24312782e-01
1.26697671e+00 5.34838021e-01 -8.60013902e-01 1.29964054e+00
2.61486441e-01 9.31022584e-01 -4.07351613e-01 1.70744270e-01
-1.72133014e-01 -1.92249030e-01 8.22066724e-01 6.80537820e-01
2.80723751e-01 2.94901311e-01 2.01669440e-01 8.23502004e-01
-7.57380426e-02 -1.76941156e-01 -6.42215490e-01 -3.90894264e-01
4.48202342e-01 1.76499414e+00 -5.52669764e-01 -5.72636008e-01
-3.65450144e-01 3.50447536e-01 3.62489164e-01 -1.69503257e-01
-4.41208780e-01 5.60926460e-02 8.61697376e-01 -1.80128247e-01
-1.28759275e-04 -5.92439950e-01 -5.56650400e-01 -1.15525627e+00
-6.11162126e-01 -1.09781218e+00 4.59427714e-01 -1.10258710e+00
-1.40434682e+00 2.24787816e-01 -2.77410120e-01 -6.00406587e-01
-6.83102384e-02 -5.02680898e-01 -6.92931533e-01 8.37317228e-01
-1.08902383e+00 -1.25911605e+00 -2.36207351e-01 -1.81544632e-01
1.95156142e-01 -8.61589015e-01 1.16446507e+00 2.11455941e-01
-6.26335919e-01 -2.91106433e-01 4.91995573e-01 -4.99716252e-01
4.87928241e-01 -9.29174781e-01 8.10377955e-01 4.46173429e-01
2.91713834e-01 4.37289625e-01 6.96681201e-01 -8.67090225e-01
-1.96824348e+00 -9.64090824e-01 8.56622696e-01 -7.04677939e-01
1.24432611e+00 -2.61643887e-01 -9.17723656e-01 8.97678137e-01
2.78220236e-01 2.11077198e-01 9.24314797e-01 3.63608956e-01
-2.71429867e-01 -1.94817618e-01 -9.07407105e-01 2.35295817e-02
8.85035992e-01 -2.62344390e-01 -5.67864776e-01 1.17343903e+00
7.73140252e-01 -1.15005702e-01 -1.03807497e+00 3.52989346e-01
3.50173056e-01 -6.28403783e-01 8.89013886e-01 -1.23094130e+00
2.93673903e-01 -4.53472704e-01 3.69937927e-01 -1.13266432e+00
-5.33624828e-01 -1.07306683e+00 3.00498959e-02 6.04686260e-01
4.96806234e-01 -9.67958748e-01 4.12997693e-01 6.44852102e-01
-5.51003814e-01 -4.50697958e-01 -1.41903269e+00 -5.12811482e-01
1.51976094e-01 -5.76108873e-01 9.16339636e-01 1.06356692e+00
9.82373580e-02 2.44231775e-01 4.42389175e-02 5.27654231e-01
4.36861932e-01 2.29091778e-01 8.15486252e-01 -1.68450129e+00
-5.91890991e-01 -4.01667863e-01 -6.05035961e-01 1.31174503e-02
7.20960926e-03 -1.40704477e+00 -4.42082465e-01 -1.03937566e+00
4.91161793e-02 -9.22296882e-01 -3.82535607e-01 1.01946437e+00
1.61781698e-01 3.73093843e-01 -2.02844471e-01 7.27262259e-01
-7.15385497e-01 3.34787428e-01 4.44686800e-01 -5.99470586e-02
2.20542550e-01 -4.53314543e-01 -5.58453560e-01 6.87438905e-01
8.44548464e-01 -6.91992342e-01 2.76968181e-01 -5.90088964e-01
5.48447549e-01 -3.14155459e-01 5.15455008e-01 -1.44236958e+00
4.47628438e-01 -3.36235613e-01 4.51241970e-01 -2.06572995e-01
3.43193300e-02 -3.45697224e-01 6.17309630e-01 4.61173564e-01
-4.77652512e-02 1.37036815e-01 5.58933139e-01 6.43456876e-02
4.12693858e-01 -1.95433617e-01 5.13526082e-01 -5.10986686e-01
-5.47398508e-01 3.98645818e-01 -3.15454841e-01 -1.68968901e-01
1.21783042e+00 7.91957140e-01 -7.70338833e-01 1.89513803e-01
-5.17320633e-01 2.29786694e-01 4.30583119e-01 5.50773978e-01
-7.65249133e-02 -6.59183145e-01 -1.07400930e+00 5.19959450e-01
1.33319452e-01 7.65028177e-03 2.09234923e-01 9.12006199e-01
-7.83751309e-01 6.50331318e-01 -2.84120649e-01 -4.92190719e-01
-6.47330642e-01 5.48615813e-01 4.21298027e-01 1.78098381e-01
-5.71126103e-01 7.58324146e-01 -5.68674862e-01 -9.18134809e-01
-4.54472512e-01 3.46200228e-01 2.14894786e-01 -2.47357354e-01
8.51372600e-01 3.76261204e-01 6.01094663e-01 -2.54034579e-01
-3.13824832e-01 -2.41163984e-01 1.82518929e-01 1.39607573e-02
1.77762425e+00 3.04376334e-01 -8.38462770e-01 8.01572025e-01
5.53129792e-01 -1.84884638e-01 -6.61593080e-01 1.39498100e-01
5.10601044e-01 -3.85952383e-01 3.53095978e-01 -1.32111979e+00
-8.22569370e-01 4.60309714e-01 6.67556643e-01 3.17369759e-01
6.33547306e-01 4.91167486e-01 5.10555804e-01 6.86645925e-01
5.02407432e-01 -7.98886657e-01 -3.19144785e-01 5.08520126e-01
9.96747494e-01 -1.16985619e+00 2.77322650e-01 4.10733759e-01
-2.77260572e-01 1.24996686e+00 5.41487932e-01 2.38140821e-01
4.99980152e-01 1.23628475e-01 -5.06957527e-03 -7.35960305e-01
-1.56723213e+00 -1.64416969e-01 -2.00661466e-01 4.01185334e-01
2.58196682e-01 1.52486518e-01 -4.57172364e-01 4.23610479e-01
-2.56928980e-01 1.13517962e-01 2.33960599e-01 9.25333321e-01
-4.36427683e-01 -1.14631820e+00 -2.61230171e-01 5.47431111e-01
-3.50239992e-01 -2.70105541e-01 -3.52331251e-01 1.69769108e-01
3.33600372e-01 7.54311562e-01 4.54244167e-01 -6.25932515e-02
-2.54900575e-01 5.25507987e-01 -7.77389575e-03 -6.06728554e-01
-6.36427164e-01 -4.80386972e-01 3.02488863e-01 -5.12447238e-01
-1.57633528e-01 -7.43957579e-01 -1.52934682e+00 -5.03232837e-01
-8.97958726e-02 4.37301189e-01 1.57870328e+00 1.14363611e+00
1.50653553e+00 2.01395363e-01 4.55351353e-01 -1.08037496e+00
-2.24328607e-01 -1.10243797e+00 -2.14058951e-01 7.59322494e-02
-1.18975766e-01 -5.78621209e-01 -5.64465344e-01 -1.74704999e-01] | [7.8827033042907715, 3.186253547668457] |
6cc1ff24-5557-4279-af4e-221d364a017d | utilizing-technical-data-to-discover-similar | 2301.04455 | null | https://arxiv.org/abs/2301.04455v1 | https://arxiv.org/pdf/2301.04455v1.pdf | Utilizing Technical Data to Discover Similar Companies in Dhaka Stock Exchange | Stock market investment have been an ideal form of investment for many years. Investing capitals smartly in stock market yields high profit returns. But there are many companies available in a market. Currently there are more than $345$ active companies who have stocks in Dhaka Stock Exchange (DSE). Analyzing all these companies is quite impossible. However, many companies tend to move together. This study aims at finding which companies in DSE have a close connection and move alongside each other. By analyzing this relation, the investors and traders will be able to analyze a lot of companies' statistics from a calculating just a handful number of companies. The conducted experiment yielded promising results. It was found that though the system was not given anything other than technical data, it was able to identify companies that show domain specific outcomes. In other words, a relation between technical data and fundamental data was discovered from the conducted experiment. | ['Mohammad Shafiul Alam', 'Tahsin Aziz', 'Tashreef Muhammad'] | 2023-01-11 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-9.20901179e-01 1.42215148e-01 -2.63701111e-01 -1.59452483e-02
-2.38472730e-01 -6.99630141e-01 4.39753443e-01 1.07179463e-01
-2.78399199e-01 8.60338032e-01 8.30678195e-02 -5.54014564e-01
1.40871610e-02 -1.28550148e+00 -3.23283374e-01 -5.22695482e-01
-7.89585859e-02 6.19045794e-01 4.11832154e-01 -6.64073527e-01
1.02956438e+00 6.49404764e-01 -1.21654487e+00 1.95576191e-01
2.76432395e-01 9.65992868e-01 1.40879303e-01 1.23492263e-01
-6.64956808e-01 1.28675950e+00 -8.24296772e-01 -6.47243500e-01
7.74489343e-01 -1.87182039e-01 -6.47422910e-01 -3.35888475e-01
-4.07095343e-01 -3.41349453e-01 4.95306641e-01 1.06160223e+00
2.78735489e-01 -4.91967320e-01 1.80320621e-01 -1.09099317e+00
-3.78329068e-01 1.06788349e+00 -7.91876376e-01 5.94728649e-01
2.82481741e-02 -1.98013410e-02 1.40726948e+00 -8.86903167e-01
6.29533589e-01 8.90297711e-01 3.57270777e-01 -1.21308982e-01
-6.45534515e-01 -1.17609799e+00 -1.91478148e-01 -2.54620969e-01
-9.95125115e-01 -1.45854114e-03 9.46357191e-01 -6.07589364e-01
9.58476126e-01 2.21498340e-01 1.01155591e+00 2.29741797e-01
3.85481417e-01 4.75203365e-01 1.23547482e+00 -2.48614207e-01
1.75589398e-01 8.50997984e-01 3.92047465e-02 -1.82572193e-02
9.58930612e-01 4.63702939e-02 -4.69023824e-01 3.28081027e-02
8.31520796e-01 2.47939304e-01 1.71672791e-01 1.95169806e-01
-1.16101205e+00 1.07641470e+00 2.13058889e-01 1.04384935e+00
-6.25408292e-01 -2.05503508e-01 2.61024117e-01 6.96483076e-01
1.81785867e-01 7.67371118e-01 -5.64318597e-01 -4.95243460e-01
-1.00083184e+00 2.29385376e-01 1.08733344e+00 5.87760866e-01
5.95862806e-01 2.11975694e-01 6.08612299e-01 -1.43466190e-01
3.31898242e-01 3.71314973e-01 7.59821057e-01 -7.16740966e-01
7.03811884e-01 1.18545401e+00 2.85881609e-01 -1.35236931e+00
-2.35124126e-01 -5.88005543e-01 -3.23615670e-01 5.23031235e-01
5.63572168e-01 -2.45422617e-01 -1.15002520e-01 8.03414404e-01
-2.87307024e-01 -2.66825646e-01 2.77330875e-01 5.74052572e-01
4.94141251e-01 6.08742654e-01 -4.82837111e-01 -1.91306427e-01
1.17872119e+00 -5.32495737e-01 -9.11102295e-01 7.57771060e-02
3.13104570e-01 -8.69407177e-01 6.36642933e-01 6.32453799e-01
-1.47111905e+00 -3.43882084e-01 -1.03654575e+00 6.29285276e-01
-6.50665760e-01 -2.92588860e-01 8.18730712e-01 8.79575729e-01
-8.20729196e-01 7.17987716e-01 -6.22551680e-01 -7.39789605e-02
3.29055488e-01 5.44804037e-01 2.47827508e-02 8.03316355e-01
-1.28681719e+00 1.04658973e+00 4.26918894e-01 -9.54548921e-03
-1.00222491e-01 -3.82686138e-01 -5.61737269e-02 8.70870203e-02
4.79437739e-01 1.48088425e-01 8.91429186e-01 -8.14731658e-01
-1.02603328e+00 7.26435244e-01 3.47200334e-01 -7.37689018e-01
5.78584969e-01 9.34242010e-02 -7.64129817e-01 4.57086600e-02
7.36370459e-02 -3.76782715e-02 -5.79078943e-02 -8.87780666e-01
-1.22025645e+00 -3.01008016e-01 8.94713495e-03 -2.47852772e-01
-4.25490141e-01 4.99372810e-01 3.59978914e-01 -6.60770059e-01
4.30392057e-01 -4.94928569e-01 -5.23455590e-02 -8.89333129e-01
-1.36868596e-01 -1.82203129e-02 4.80554640e-01 -4.42851931e-01
1.48276281e+00 -1.63168156e+00 -7.57271707e-01 5.50433874e-01
1.42543539e-01 1.40463963e-01 7.21717954e-01 9.14067388e-01
-2.20516115e-01 9.23543751e-01 4.79431808e-01 7.23800421e-01
1.64526507e-01 -2.09452733e-01 -5.46942174e-01 1.58684924e-01
6.82488084e-02 9.32392895e-01 -5.49481392e-01 -2.50035852e-01
-6.69791624e-02 -2.30090201e-01 -1.89193130e-01 -3.65077287e-01
5.45371436e-02 5.82494289e-02 -7.01180935e-01 1.00985479e+00
8.50449026e-01 -3.70235682e-01 1.53668478e-01 1.35652907e-02
-9.33286786e-01 3.84985924e-01 -1.42173517e+00 6.97476685e-01
1.58758879e-01 7.36384273e-01 -4.16588038e-01 -1.03742433e+00
1.45947230e+00 4.87574071e-01 7.78394818e-01 -5.98925889e-01
2.22919613e-01 8.74943078e-01 3.72832268e-01 -2.56299168e-01
7.00284243e-01 -7.41208136e-01 6.20062947e-02 6.92400396e-01
-5.13767779e-01 1.85458973e-01 5.00395060e-01 -1.51195645e-01
9.51438427e-01 -3.94840717e-01 4.95914698e-01 -3.85403931e-01
3.08964878e-01 2.80506164e-01 5.84595382e-01 2.26909757e-01
-9.21042040e-02 1.79109395e-01 7.67195106e-01 -4.65244681e-01
-9.03852642e-01 -6.34075582e-01 -7.01281279e-02 2.87316442e-01
1.71388239e-02 -9.65495873e-03 -1.86666951e-01 -4.09844182e-02
9.45184231e-02 5.54076254e-01 -4.50022280e-01 5.13779998e-01
-1.55667260e-01 -5.24909914e-01 2.33603939e-01 3.49774837e-01
9.97646511e-01 -1.23611212e+00 -1.11423850e+00 5.05786479e-01
3.19826782e-01 -6.12225890e-01 1.38470441e-01 2.99908161e-01
-1.18452823e+00 -8.87511790e-01 -7.84227371e-01 -2.92005837e-01
3.64421040e-01 1.09246008e-01 1.04885316e+00 8.83979574e-02
-8.17269459e-02 -4.84279484e-01 -2.80582070e-01 -1.10181725e+00
-3.94047260e-01 9.76243094e-02 -1.92672849e-01 -2.52477050e-01
7.69712448e-01 -2.92934209e-01 -4.79313821e-01 3.83990884e-01
-5.27454972e-01 -4.73271877e-01 7.60480225e-01 2.12103814e-01
1.04420803e-01 1.18169224e+00 9.30997610e-01 -9.78790045e-01
5.42146564e-01 -5.57325482e-01 -9.91845191e-01 2.11839586e-01
-9.90070403e-01 8.42723101e-02 1.72417816e-02 -2.17962265e-02
-9.13380802e-01 -5.00484586e-01 3.09387535e-01 2.65703741e-02
2.32673794e-01 1.05491507e+00 2.05173101e-02 2.41049364e-01
2.02739999e-01 -1.77450497e-02 1.69049963e-01 -4.97809500e-01
-3.96182954e-01 6.32557988e-01 3.16591263e-02 -9.99656022e-02
9.99639153e-01 4.64313895e-01 -1.13403477e-01 -2.96992600e-01
-3.58042538e-01 -4.76124793e-01 -4.06454951e-01 -2.74576813e-01
5.86809993e-01 -7.48939991e-01 -9.50313032e-01 4.35424060e-01
-6.43912673e-01 2.80850738e-01 -4.41475987e-01 8.34470451e-01
-8.93982574e-02 -4.01061058e-01 -3.82367283e-01 -1.30796814e+00
-3.56727429e-02 -7.07979918e-01 -6.11273982e-02 4.80628431e-01
-5.58172166e-01 -1.16018629e+00 8.66245925e-02 3.62310201e-01
6.01341307e-01 6.71328068e-01 3.91144902e-01 -1.27611482e+00
-1.09713864e+00 -7.09859669e-01 -8.90106857e-02 1.91785529e-01
7.22768664e-01 4.12302285e-01 -5.15713334e-01 -2.23293126e-01
5.68240285e-01 2.81632662e-01 3.08215052e-01 3.41782212e-01
-1.59604270e-02 -4.30417240e-01 -3.97015996e-02 -6.84299022e-02
1.67699289e+00 9.69178081e-01 8.20243239e-01 1.26201844e+00
6.55988930e-03 1.16442049e+00 9.79665577e-01 7.68473029e-01
1.00168169e-01 6.64010346e-02 2.02847004e-01 2.73874123e-02
6.13290668e-01 -1.08155161e-02 4.08425659e-01 7.05951750e-01
-5.71165442e-01 2.27993906e-01 -1.18087721e+00 5.18499732e-01
-1.44356918e+00 -1.51231527e+00 -3.29208881e-01 1.98162985e+00
6.84984624e-01 9.19299245e-01 4.92873073e-01 4.68689084e-01
6.41544342e-01 -1.38404205e-01 -3.77128065e-01 -2.60861665e-01
-2.75993109e-01 2.15804189e-01 1.05903375e+00 -9.27057639e-02
-4.01627660e-01 5.74810147e-01 6.30382586e+00 3.49959671e-01
-1.40294588e+00 -4.90496665e-01 5.67576408e-01 -1.41722977e-01
-6.31031275e-01 2.69566059e-01 -1.11311281e+00 7.18453050e-01
9.40372109e-01 -9.21029389e-01 -1.91373736e-01 7.21635640e-01
3.65863353e-01 -3.77481252e-01 -4.27245349e-01 5.07714629e-01
-5.51833928e-01 -1.97211719e+00 -7.41476417e-02 8.56462896e-01
7.67064631e-01 -2.90165871e-01 4.50204797e-02 -2.54592478e-01
3.97653520e-01 -9.36479867e-01 9.14343834e-01 7.62759924e-01
-1.98897034e-01 -1.01518023e+00 1.22920561e+00 4.01090056e-01
-1.15691507e+00 -2.91104287e-01 -2.82551974e-01 -3.89230400e-01
2.92324554e-02 5.59876561e-01 -7.15161562e-01 6.91420019e-01
8.95695627e-01 4.79734063e-01 -2.77821451e-01 8.96660209e-01
2.01536313e-01 5.61625957e-01 -2.86288798e-01 -3.59959781e-01
3.22999120e-01 -6.56670392e-01 1.03448324e-01 4.54735518e-01
9.09814537e-01 1.34422287e-01 -6.50419235e-01 9.45404112e-01
2.27544665e-01 2.99515992e-01 -7.96393216e-01 -7.54534006e-01
5.35418510e-01 8.41671169e-01 -1.39615047e+00 -5.83529100e-02
-7.52922416e-01 1.97430998e-01 -6.88948512e-01 -1.19008012e-02
-4.09219027e-01 -5.85512578e-01 3.05693805e-01 7.43944883e-01
4.50768262e-01 2.99848039e-02 -6.76301718e-01 -5.57026684e-01
-4.66804393e-02 -8.20636570e-01 2.29532599e-01 -5.78394055e-01
-1.00066710e+00 1.84944406e-01 -1.93449065e-01 -1.44287205e+00
-1.84407637e-01 -6.07180953e-01 -7.97309518e-01 1.04558241e+00
-1.32870889e+00 -6.10285938e-01 2.47023284e-01 2.76654959e-01
8.76251459e-02 -9.74127948e-01 3.76116395e-01 1.01252325e-01
-2.72092611e-01 5.39247580e-02 4.18644734e-02 3.64972711e-01
1.93890765e-01 -1.42911839e+00 2.89816111e-01 6.28601134e-01
2.08344579e-01 1.05173624e+00 8.14420521e-01 -1.01512361e+00
-9.03026283e-01 -1.66856736e-01 1.27528930e+00 -3.70782554e-01
1.29012620e+00 2.27566317e-01 -7.01673508e-01 6.25341833e-01
4.94104594e-01 -6.71532869e-01 7.42530644e-01 -1.80065274e-01
2.75562376e-01 -4.18624103e-01 -1.11531639e+00 3.34797144e-01
1.68936446e-01 -1.69675007e-01 -7.70113707e-01 -1.80615708e-01
3.20573300e-01 6.34700656e-02 -9.92464483e-01 -5.26983812e-02
4.35297936e-01 -1.45700240e+00 9.31934774e-01 -3.25618982e-01
2.96160579e-01 -1.51774839e-01 1.85734238e-02 -6.42747521e-01
4.74519506e-02 -8.63979995e-01 4.85698074e-01 1.75485563e+00
8.03069592e-01 -1.19370151e+00 1.18843985e+00 1.04225254e+00
4.00641680e-01 -5.11267900e-01 -4.55818594e-01 -1.09981322e+00
2.81999350e-01 9.55165178e-02 1.01356924e+00 1.10652483e+00
1.46844417e-01 3.68839763e-02 -5.38130058e-04 -2.80509740e-01
4.92252976e-01 5.02500057e-01 6.97213113e-01 -1.47396624e+00
-2.39028949e-02 -7.79756844e-01 -3.41286093e-01 -2.47579113e-01
-5.85165381e-01 -6.25342131e-01 -7.58286417e-01 -1.60195720e+00
3.28219160e-02 -5.59083760e-01 -6.60623908e-01 1.52802274e-01
3.91140580e-01 -2.78372079e-01 2.86184043e-01 7.56922364e-01
2.79665679e-01 -4.02044624e-01 1.35239816e+00 1.40212163e-01
-3.83828998e-01 5.46508133e-01 -1.27886605e+00 6.90313458e-01
1.24424922e+00 -3.29190165e-01 -3.24146569e-01 4.11230057e-01
1.03765082e+00 1.53077781e-01 -4.88756225e-02 -7.70889521e-01
4.25209701e-01 -4.97459650e-01 1.02124140e-01 -1.30382824e+00
-6.53817877e-02 -9.60599303e-01 8.73004794e-01 8.09576392e-01
1.44702047e-02 3.02691519e-01 -6.23416677e-02 -1.29471526e-01
-7.29214251e-01 -7.96155095e-01 4.11323577e-01 -5.20416915e-01
-6.66261733e-01 -2.32172370e-01 -2.37650976e-01 -8.05517063e-02
1.25213277e+00 -8.84107769e-01 -3.71818453e-01 -3.29530299e-01
-4.18663740e-01 9.34610963e-02 3.99153769e-01 -8.78922269e-03
2.44562700e-01 -1.30296600e+00 -5.28441846e-01 -8.49326774e-02
-2.98462272e-01 -2.27286980e-01 -1.98540181e-01 7.06913292e-01
-1.08073461e+00 7.41940200e-01 -6.39227152e-01 1.03222974e-01
-8.63961220e-01 4.29300189e-01 5.28954566e-02 -7.02983856e-01
-4.85989779e-01 8.86779070e-01 -3.06582481e-01 8.59553963e-02
-1.48846418e-01 -6.62573338e-01 -6.22473657e-01 1.03266454e+00
5.09160161e-01 5.58468044e-01 -1.21625304e-01 -4.32040572e-01
-4.37122703e-01 4.88250315e-01 -2.80073062e-02 -5.46807647e-01
1.85579598e+00 -9.84966382e-02 -2.57996708e-01 8.75468254e-01
8.36655080e-01 4.90124255e-01 -9.11226690e-01 8.48267823e-02
9.41757202e-01 -7.80360878e-01 -1.58740073e-01 -5.67985892e-01
-1.51466060e+00 6.56130493e-01 1.95978120e-01 1.02467501e+00
8.25163960e-01 -9.34906453e-02 7.58044541e-01 3.26104045e-01
5.80185652e-01 -1.48827302e+00 1.66674942e-01 3.00403181e-02
8.13486934e-01 -1.14346898e+00 2.52617657e-01 -2.07889359e-02
-7.71419406e-01 1.28662205e+00 8.66246894e-02 -2.78710872e-01
1.06434357e+00 5.78205109e-01 1.51200518e-01 -6.42025888e-01
-5.36909103e-01 -1.60551891e-02 -3.96701425e-01 3.25971067e-01
5.70284307e-01 -4.94868010e-02 -4.83395129e-01 8.79333556e-01
-8.06859016e-01 2.70097971e-01 1.04561222e+00 1.31127155e+00
-1.01773322e+00 -1.26316166e+00 -5.13415873e-01 4.19849694e-01
-1.06604218e+00 -1.37239099e-01 -4.78913963e-01 1.45404351e+00
-1.20550476e-01 8.11455667e-01 3.00432682e-01 -2.40943745e-01
5.69888413e-01 -2.08830550e-01 -7.65806064e-02 -4.30425972e-01
-9.02632773e-01 2.21064970e-01 -6.51304200e-02 -1.13674521e-01
-6.57076538e-01 -1.01218975e+00 -1.72540998e+00 -9.18457747e-01
-3.25969785e-01 5.29689550e-01 7.73814082e-01 4.17617708e-01
-3.57129931e-01 4.71333474e-01 9.99787271e-01 -8.72882754e-02
-4.66953576e-01 -7.78947473e-01 -1.44949770e+00 -1.08548321e-01
-7.82876015e-02 -5.09414732e-01 -5.71178198e-01 -3.47033404e-02] | [4.59636116027832, 4.23836612701416] |
b0b39e2c-6cab-40f2-b86c-ba2dae1aa911 | heuristic-algorithms-for-the-approximation-of | 2307.01639 | null | https://arxiv.org/abs/2307.01639v1 | https://arxiv.org/pdf/2307.01639v1.pdf | Heuristic Algorithms for the Approximation of Mutual Coherence | Mutual coherence is a measure of similarity between two opinions. Although the notion comes from philosophy, it is essential for a wide range of technologies, e.g., the Wahl-O-Mat system. In Germany, this system helps voters to find candidates that are the closest to their political preferences. The exact computation of mutual coherence is highly time-consuming due to the iteration over all subsets of an opinion. Moreover, for every subset, an instance of the SAT model counting problem has to be solved which is known to be a hard problem in computer science. This work is the first study to accelerate this computation. We model the distribution of the so-called confirmation values as a mixture of three Gaussians and present efficient heuristics to estimate its model parameters. The mutual coherence is then approximated with the expected value of the distribution. Some of the presented algorithms are fully polynomial-time, others only require solving a small number of instances of the SAT model counting problem. The average squared error of our best algorithm lies below 0.0035 which is insignificant if the efficiency is taken into account. Furthermore, the accuracy is precise enough to be used in Wahl-O-Mat-like systems. | ['Tamara Mchedlidze', 'Vera Chekan', 'Gregor Betz'] | 2023-07-04 | null | null | null | null | ['philosophy'] | ['miscellaneous'] | [ 2.52766013e-02 3.04976881e-01 -1.59287721e-01 -4.80838209e-01
-8.14047158e-01 -8.18958223e-01 4.45698738e-01 6.77451730e-01
-5.98898470e-01 8.34707975e-01 -4.68279243e-01 -6.85206413e-01
-2.56506145e-01 -1.17441154e+00 -5.79126716e-01 -7.88019359e-01
2.81703651e-01 1.27524018e+00 1.78772777e-01 -3.04325163e-01
3.58364075e-01 2.79647082e-01 -1.57901371e+00 3.85394245e-02
1.09668219e+00 1.06686926e+00 4.96965572e-02 6.19561732e-01
-1.18910186e-01 3.05340022e-01 -4.77706075e-01 -7.94204056e-01
3.36744964e-01 -4.61885810e-01 -9.00639594e-01 -1.44410312e-01
2.84937382e-01 2.25642696e-01 2.70545512e-01 1.60628402e+00
1.88508064e-01 7.59430006e-02 7.30422854e-01 -1.09153759e+00
1.75224282e-02 7.53770888e-01 -4.97150123e-01 -1.32565722e-01
2.85225391e-01 -3.48486900e-01 1.24569011e+00 -3.56053799e-01
5.17347991e-01 8.69142652e-01 3.22170913e-01 -1.06052659e-01
-1.34345281e+00 -4.33130741e-01 -8.28621238e-02 4.02696341e-01
-1.36704588e+00 -8.68244469e-02 3.79628360e-01 -1.97696447e-01
2.92477608e-01 7.23362386e-01 9.36813831e-01 2.04364195e-01
8.23893547e-02 4.40341800e-01 1.60610998e+00 -6.29586399e-01
5.83514214e-01 6.49756908e-01 4.67393875e-01 6.64637804e-01
7.85908043e-01 -3.38907748e-01 -1.42128542e-01 -2.84628600e-01
1.89608112e-01 -2.08323196e-01 -1.04868352e-01 -2.94688195e-01
-8.02791297e-01 9.75484014e-01 3.46480131e-01 7.30333805e-01
-1.33757219e-01 1.12052634e-01 -1.01911820e-01 4.59497213e-01
3.89352828e-01 3.78795832e-01 -3.92306298e-01 -1.26041368e-01
-1.15252960e+00 2.46174678e-01 1.24673843e+00 6.43889010e-01
8.51973474e-01 -6.40190661e-01 2.80519217e-01 2.41771966e-01
1.14100158e-01 7.75292754e-01 -2.68377084e-02 -7.35353887e-01
2.61925071e-01 7.73794472e-01 2.90048301e-01 -1.22021663e+00
-2.61225730e-01 -6.81837201e-01 -9.07424748e-01 6.92081228e-02
1.01476467e+00 -1.42864779e-01 -3.20099503e-01 1.56850052e+00
5.05138457e-01 -2.14448050e-01 -1.88899547e-01 8.05048168e-01
3.04083049e-01 6.02985919e-01 -4.88119900e-01 -4.84034926e-01
1.71145141e+00 -5.20332515e-01 -5.67877054e-01 -1.59149319e-01
8.06966126e-01 -9.74109054e-01 5.35294712e-01 7.56306052e-01
-1.14635623e+00 -3.98410447e-02 -7.75990844e-01 2.05640107e-01
-3.13024759e-01 2.17087418e-01 8.52208555e-01 1.04565752e+00
-9.01691079e-01 4.06629473e-01 -4.47210699e-01 -2.88770437e-01
-3.66616994e-02 6.76665545e-01 -5.32466054e-01 -1.94559306e-01
-1.10430515e+00 1.12227798e+00 2.81230628e-01 2.08711803e-01
1.60945281e-01 -5.31422645e-02 -4.62598950e-01 2.67548293e-01
5.33339322e-01 -7.78261900e-01 1.19878459e+00 -1.12860191e+00
-1.32357979e+00 9.74467635e-01 -5.60044408e-01 -5.57439268e-01
7.52161682e-01 3.57310385e-01 -1.17339473e-02 -1.55898497e-01
-7.26748481e-02 4.60086055e-02 3.60518515e-01 -1.10950387e+00
-6.36856377e-01 -5.46240866e-01 3.35143209e-01 -5.49961478e-02
-3.48168015e-02 -1.81463182e-01 -3.25097620e-01 9.01664644e-02
5.75959504e-01 -1.21647799e+00 -6.59889340e-01 -4.56806809e-01
-3.68689120e-01 -2.46563867e-01 -5.65720536e-02 -2.33027637e-01
1.18370092e+00 -1.71309829e+00 4.88749966e-02 7.97508538e-01
7.81590715e-02 -8.24675635e-02 1.94860488e-01 4.15160418e-01
1.30669445e-01 4.89673484e-03 -1.27903298e-01 -4.89210747e-02
1.51266739e-01 1.28982425e-01 3.63558903e-02 8.90981317e-01
-4.66838658e-01 5.63660681e-01 -7.50062108e-01 -5.38572729e-01
1.33754052e-02 -2.89539751e-02 -5.90696037e-01 -3.12370002e-01
-2.42855340e-01 -3.00105549e-02 -3.70311826e-01 1.83294669e-01
9.75252032e-01 -3.24736357e-01 4.42765892e-01 -1.63301274e-01
-2.82331109e-01 1.29020840e-01 -1.50218165e+00 1.05959678e+00
-5.19291580e-01 5.19649029e-01 -3.48803662e-02 -1.07456470e+00
9.17628586e-01 1.85188562e-01 3.96269649e-01 -3.48640740e-01
6.74839020e-01 6.83461308e-01 2.33859509e-01 -1.79882348e-02
6.54623985e-01 -4.27294582e-01 -2.30365440e-01 4.64844882e-01
-3.19025248e-01 -5.72437942e-01 7.20763385e-01 8.37852284e-02
9.28994298e-01 -7.10630834e-01 5.22139072e-01 -4.75990713e-01
6.96119606e-01 1.15091845e-01 4.02226478e-01 6.29967451e-01
2.78762132e-01 2.55125821e-01 8.38342071e-01 -3.51801425e-01
-7.93320298e-01 -5.08239865e-01 -2.08834186e-01 3.71817559e-01
3.04205269e-01 -5.16380906e-01 -8.80527735e-01 -5.56139886e-01
-1.67985171e-01 9.73510146e-01 -6.63323760e-01 1.06261082e-01
-4.19735722e-02 -7.15718389e-01 -2.36355752e-01 -1.60389885e-01
3.22934479e-01 -4.45714176e-01 -5.26233494e-01 1.14912353e-02
-4.07148421e-01 -9.71866727e-01 -5.91987558e-02 3.50035876e-01
-8.33246410e-01 -1.27300453e+00 -5.53103626e-01 -4.60004658e-01
9.36908245e-01 2.92505473e-01 1.13270020e+00 2.78095435e-02
1.48713470e-01 -1.13573112e-03 -1.67080700e-01 -3.62515032e-01
-2.68632650e-01 8.80151019e-02 -6.93225488e-02 -5.05268313e-02
6.36093557e-01 -3.77932072e-01 -3.12005311e-01 4.85319704e-01
-7.29798496e-01 -2.32604042e-01 5.60941517e-01 7.95464993e-01
7.71613300e-01 5.59306264e-01 -2.81677991e-01 -1.06973386e+00
4.97956723e-01 -3.09496433e-01 -1.17991614e+00 3.86142999e-01
-4.21536148e-01 1.69535935e-01 5.39758384e-01 -1.63116679e-01
-5.46986163e-01 3.26335222e-01 1.68336675e-01 3.76042664e-01
1.26495928e-01 5.98260045e-01 -1.18238464e-01 -5.52991498e-03
3.23844433e-01 -9.70592536e-03 -2.81352401e-01 -7.31916130e-02
4.02974457e-01 5.95725179e-01 2.91357726e-01 -4.68962491e-01
8.58140230e-01 3.24236214e-01 4.38784152e-01 -7.60131478e-01
-1.20802438e+00 -5.83584964e-01 -2.87125349e-01 -7.96419978e-02
5.00509202e-01 -5.39928138e-01 -1.19243681e+00 3.11288983e-02
-1.30132759e+00 -3.56833525e-02 -2.84026802e-01 5.23431420e-01
-5.38790107e-01 3.58377039e-01 -1.25316530e-01 -1.29120302e+00
-1.15264758e-01 -1.05719256e+00 7.67401218e-01 2.70540655e-01
-4.16725248e-01 -7.32495248e-01 1.67392910e-01 6.04964316e-01
8.38956311e-02 4.57649454e-02 8.27739894e-01 -6.86490774e-01
-5.28999805e-01 -7.93255687e-01 -1.02727160e-01 -2.42896620e-02
-2.49313161e-01 4.02272083e-02 -4.76660460e-01 -1.32480323e-01
1.96567595e-01 1.95024922e-01 7.08587945e-01 6.10454917e-01
8.98527682e-01 -2.65782386e-01 -2.32300118e-01 4.08171341e-02
1.47325289e+00 -2.13269621e-01 4.56927538e-01 3.16897303e-01
-6.09907173e-02 6.72715068e-01 8.46110225e-01 4.08438146e-01
5.18223643e-01 8.64588559e-01 4.69512433e-01 8.13499987e-02
6.86450779e-01 1.51913211e-01 7.07771853e-02 7.90230930e-01
-2.76928633e-01 -2.99939543e-01 -6.78472757e-01 3.47135723e-01
-1.87146962e+00 -1.14823925e+00 -8.63934696e-01 2.59656882e+00
6.67702854e-01 4.50790912e-01 5.65636195e-02 5.38464665e-01
7.11064279e-01 -3.38696003e-01 1.55368283e-01 -7.07278907e-01
-1.11488089e-01 3.28396618e-01 8.07056069e-01 9.34721410e-01
-7.72906065e-01 7.58058667e-01 5.36722088e+00 8.12926829e-01
-6.96708143e-01 5.25453091e-02 6.40176594e-01 5.86277507e-02
-4.57642496e-01 4.40967917e-01 -7.45366037e-01 6.01672292e-01
7.95356095e-01 -4.30672348e-01 2.91473418e-01 7.65033245e-01
8.53371844e-02 -9.19295132e-01 -8.49409699e-01 9.38737631e-01
-1.07279278e-01 -7.94810295e-01 -4.80397075e-01 4.27557141e-01
1.03263366e+00 -4.13809091e-01 5.02215810e-02 -3.67341936e-02
4.66041625e-01 -9.79172885e-01 6.22501314e-01 5.46254933e-01
4.86118942e-01 -1.05522132e+00 1.37851679e+00 5.55288255e-01
-9.63749826e-01 5.16759008e-02 -6.82814837e-01 -1.70763984e-01
2.21192539e-01 1.54080892e+00 -4.97039080e-01 5.04555643e-01
1.26399651e-01 -8.25814605e-02 -1.67953387e-01 1.33027625e+00
-2.57339150e-01 4.79559690e-01 -9.78189707e-01 -5.02919316e-01
3.28647405e-01 -6.81061268e-01 3.70640367e-01 8.14300776e-01
7.62407362e-01 1.11961856e-01 -1.86981827e-01 4.94779974e-01
9.40727256e-03 3.46130759e-01 -3.94556403e-01 4.38882932e-02
3.88356566e-01 1.44625854e+00 -1.03900504e+00 -3.60999554e-01
-5.72688878e-02 8.62434089e-01 -5.13181090e-02 -1.51529536e-01
-6.94438100e-01 -1.03283413e-01 2.83976346e-01 2.86520183e-01
3.45898837e-01 -2.38133222e-01 -4.92504656e-01 -8.76066506e-01
9.86995995e-02 -8.99266362e-01 2.47436434e-01 -3.19230765e-01
-9.70747709e-01 6.17538631e-01 -1.86416253e-01 -1.11230409e+00
-2.86258876e-01 -5.55202425e-01 -3.99855226e-01 7.66055286e-01
-1.12418854e+00 -4.96226996e-01 -1.47422165e-01 2.72594482e-01
-5.88007346e-02 1.86649978e-01 9.05559003e-01 8.89129788e-02
-3.01126450e-01 3.99379700e-01 3.43334913e-01 -4.50802386e-01
5.29830515e-01 -1.44914055e+00 -9.07071903e-02 6.13814473e-01
2.91620076e-01 5.76204300e-01 1.29320729e+00 -2.49530494e-01
-1.39544940e+00 -3.45122814e-01 1.64166188e+00 -3.61457855e-01
7.26944387e-01 -1.76991880e-01 -3.36472392e-01 2.78555959e-01
-7.09658265e-02 -2.70962119e-02 6.80132449e-01 4.21128780e-01
-1.01615228e-01 -2.67003298e-01 -1.09298790e+00 4.92647797e-01
4.62488383e-01 -1.75267220e-01 -2.20811516e-01 5.06164968e-01
-7.06940377e-03 -3.73386234e-01 -6.64972126e-01 -2.68409867e-02
5.81334054e-01 -1.27674830e+00 3.98662746e-01 -1.76774174e-01
2.80702025e-01 -4.89902377e-01 -3.01523983e-01 -1.26099515e+00
-8.88473094e-02 -6.12494230e-01 2.87405103e-02 9.56116915e-01
6.79584980e-01 -7.08974421e-01 8.94207478e-01 5.56760967e-01
6.66291773e-01 -7.72785604e-01 -1.11544836e+00 -5.93463302e-01
-3.53631787e-02 -4.58783418e-01 6.02851391e-01 8.19684505e-01
2.42969424e-01 3.84244233e-01 -8.54927599e-02 1.67728245e-01
6.36041999e-01 6.10991836e-01 9.02026415e-01 -1.62499666e+00
-5.21631479e-01 -6.62090361e-01 -5.21185458e-01 -9.95038807e-01
-1.71982981e-02 -6.89102471e-01 -1.22791447e-01 -1.61893117e+00
2.50290096e-01 -6.28736496e-01 -2.54271068e-02 -6.13731034e-02
-6.14777254e-03 3.11984550e-02 1.59070879e-01 -2.38061070e-01
-6.58633411e-01 9.58225355e-02 1.02208567e+00 -2.14875653e-01
-1.40469447e-02 7.40952790e-01 -7.73329377e-01 8.13329697e-01
9.47921336e-01 -7.51104891e-01 -9.26454961e-02 6.50756881e-02
9.66427565e-01 3.19455862e-01 -1.03283897e-02 -8.86654913e-01
5.42808294e-01 -9.22733992e-02 -8.46475214e-02 -8.23012829e-01
2.21200839e-01 -1.26400709e+00 4.33853328e-01 5.70291758e-01
-8.87528285e-02 2.46412233e-01 -2.07482263e-01 4.33045268e-01
-3.05916935e-01 -7.94396937e-01 7.30902076e-01 -1.00319713e-01
1.57539278e-01 -7.10724965e-02 -4.15335000e-01 -2.40598395e-01
1.06960225e+00 1.17613366e-02 -9.63113680e-02 -7.62118995e-01
-4.15468961e-01 1.67388394e-01 6.71024799e-01 -4.92818683e-01
-1.08427107e-02 -1.18612719e+00 -5.85950315e-01 -3.96558553e-01
-9.56126973e-02 -8.26804265e-02 2.53478408e-01 1.33758545e+00
-5.01644671e-01 6.80839777e-01 3.34795177e-01 -4.53553140e-01
-1.61646926e+00 3.61046314e-01 1.25757173e-01 -7.83288717e-01
5.78138195e-02 6.70089662e-01 -2.35544831e-01 -2.37768531e-01
-1.27637476e-01 -3.64795625e-01 -6.38784915e-02 3.37334275e-01
4.86240804e-01 4.85959172e-01 1.31164432e-01 -6.62777901e-01
-3.39947641e-01 6.59906983e-01 1.26330182e-01 -2.99910337e-01
1.19772601e+00 8.19355343e-03 -6.11657619e-01 2.85018533e-01
8.30064952e-01 5.30853271e-01 -3.96063656e-01 -7.49989972e-02
1.79004471e-03 -3.22549224e-01 1.38623476e-01 -4.98218328e-01
-8.86665404e-01 7.66799867e-01 3.07400793e-01 6.79324985e-01
1.03241694e+00 -1.41562968e-01 5.14113724e-01 6.55510366e-01
7.33675420e-01 -1.09820306e+00 -6.42197371e-01 5.44030011e-01
4.37141806e-01 -1.11155844e+00 3.71473819e-01 -6.46124423e-01
-3.81606460e-01 1.14161873e+00 -3.91223803e-02 -4.20127176e-02
5.79607129e-01 2.91686058e-01 -1.90040424e-01 -6.46154955e-02
-7.24637628e-01 -5.05245149e-01 3.69066358e-01 -3.91549058e-02
5.37714005e-01 4.70190555e-01 -1.16397846e+00 5.38541794e-01
-7.67603934e-01 -5.00596873e-02 5.54333329e-01 4.46197569e-01
-5.88695228e-01 -1.41144609e+00 -5.42313457e-01 3.75558168e-01
-3.74856085e-01 -5.93607314e-02 -4.93882358e-01 6.59059823e-01
1.99577138e-01 1.25458825e+00 -1.86018541e-01 -2.17120394e-01
2.93573648e-01 -1.43803418e-01 7.88031936e-01 -2.67587394e-01
-5.55153549e-01 -2.89224476e-01 1.53333873e-01 -2.49840289e-01
-4.29736108e-01 -6.41018927e-01 -8.75542343e-01 -8.37610900e-01
-9.11965787e-01 7.44784415e-01 9.36712742e-01 9.64244485e-01
-3.05625767e-01 6.24279156e-02 6.87553465e-01 -3.26812178e-01
-6.72646403e-01 -6.73706114e-01 -8.20299864e-01 -1.04853399e-01
-3.17248732e-01 -3.54583859e-01 -4.09365952e-01 -5.69268525e-01] | [6.55049991607666, 4.661287307739258] |
b61617ac-0d03-4c53-ab68-554db575aa65 | explicit-and-implicit-models-in-infrared-and | 2206.09581 | null | https://arxiv.org/abs/2206.09581v1 | https://arxiv.org/pdf/2206.09581v1.pdf | Explicit and implicit models in infrared and visible image fusion | Infrared and visible images, as multi-modal image pairs, show significant differences in the expression of the same scene. The image fusion task is faced with two problems: one is to maintain the unique features between different modalities, and the other is to maintain features at various levels like local and global features. This paper discusses the limitations of deep learning models in image fusion and the corresponding optimization strategies. Based on artificially designed structures and constraints, we divide models into explicit models, and implicit models that adaptively learn high-level features or can establish global pixel associations. Ten models for comparison experiments on 21 test sets were screened. The qualitative and quantitative results show that the implicit models have more comprehensive ability to learn image features. At the same time, the stability of them needs to be improved. Aiming at the advantages and limitations to be solved by existing algorithms, we discuss the main problems of multi-modal image fusion and future research directions. | ['Bin Sun', 'Zixuan Wang'] | 2022-06-20 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 1.59024879e-01 -4.97174084e-01 -1.25187263e-01 -4.38250422e-01
-9.56406772e-01 -1.43383041e-01 3.58329207e-01 -9.33650210e-02
-1.98681146e-01 5.79102457e-01 1.87801912e-01 1.94731399e-01
-4.94557112e-01 -7.29032755e-01 -1.91015363e-01 -1.28690827e+00
6.59082159e-02 -2.80048072e-01 -6.15523607e-02 -3.24047148e-01
7.97684640e-02 4.82660770e-01 -2.04308128e+00 6.02636755e-01
9.92361248e-01 1.42896020e+00 3.12959403e-01 5.66585362e-01
-1.77162498e-01 8.64408910e-01 -2.55675346e-01 -1.79300606e-02
3.64000142e-01 -4.35207695e-01 -5.93605936e-01 3.79199624e-01
6.07980788e-01 -8.33409801e-02 -2.80825853e-01 1.47234178e+00
8.40980947e-01 1.82238922e-01 4.41982627e-01 -1.37664878e+00
-9.11187589e-01 3.97592455e-01 -7.63688505e-01 4.42336909e-02
3.56294781e-01 -3.39386277e-02 6.41243219e-01 -7.96510935e-01
8.56684521e-02 1.18715620e+00 6.88521087e-01 3.63728106e-01
-9.55856323e-01 -6.79941416e-01 3.23188424e-01 5.81703842e-01
-1.50106108e+00 -4.84841019e-01 8.97081912e-01 -3.77764702e-01
4.67193097e-01 5.68882644e-01 4.68843639e-01 4.78906155e-01
5.12638390e-01 6.72788501e-01 1.32257462e+00 -5.22927284e-01
-1.16302304e-01 1.63280919e-01 2.37429172e-01 8.13160479e-01
4.53350367e-04 1.99249819e-01 -4.81422752e-01 -8.63962248e-02
5.28262496e-01 1.53639510e-01 -5.43820381e-01 -2.02093348e-01
-1.22036970e+00 7.49123454e-01 6.60644174e-01 7.78198600e-01
-4.56343770e-01 -2.55697012e-01 3.67011176e-03 3.69934618e-01
2.29977190e-01 1.49372220e-01 -4.85720575e-01 6.25165284e-01
-5.74840486e-01 -3.67067158e-02 1.85379133e-01 5.65684974e-01
1.29795337e+00 -3.29980701e-02 -2.17744663e-01 9.64949846e-01
5.28060853e-01 4.88863379e-01 5.15355647e-01 -9.97777164e-01
1.98528200e-01 7.14812219e-01 -5.73045760e-02 -1.25043523e+00
-4.32150632e-01 -2.09314778e-01 -1.24525452e+00 5.20695269e-01
5.94739914e-02 -2.07843736e-01 -1.05452585e+00 1.67403901e+00
2.10508391e-01 1.28472090e-01 1.86280608e-01 8.69171321e-01
1.41818702e+00 8.10887635e-01 1.28940150e-01 -6.28782630e-01
1.44299245e+00 -9.15588379e-01 -1.16587353e+00 -1.85420394e-01
3.63141298e-01 -1.01878870e+00 5.22105396e-01 8.95255134e-02
-1.06376684e+00 -1.05052292e+00 -1.01179945e+00 1.55257508e-02
-5.88209152e-01 3.74970883e-01 6.08347118e-01 5.68698585e-01
-1.32498133e+00 2.06103191e-01 -3.53239864e-01 -2.39257604e-01
5.41868210e-02 5.70275426e-01 -4.81636673e-01 -8.20647646e-03
-1.35802007e+00 8.91655564e-01 6.69900358e-01 3.26902747e-01
-3.53987873e-01 -4.08398867e-01 -1.07500255e+00 3.46796326e-02
9.64564830e-02 -8.87768388e-01 7.86482453e-01 -1.18563044e+00
-1.30209863e+00 8.32747161e-01 -2.66010106e-01 1.98585927e-01
1.34317294e-01 -1.85138185e-03 -8.30903888e-01 7.13353604e-02
-7.34366775e-02 6.09741926e-01 7.34555542e-01 -1.84434175e+00
-1.17707860e+00 -3.81607622e-01 1.28460126e-02 4.68823433e-01
-4.87905115e-01 1.08894981e-01 -4.66820836e-01 -4.36050653e-01
4.58023608e-01 -6.60782993e-01 -3.13819826e-01 4.40978073e-02
-9.90834087e-02 -1.18093930e-01 1.07492566e+00 -7.36474454e-01
1.06546354e+00 -2.15047622e+00 2.82899588e-01 2.37708345e-01
1.22116908e-01 3.04685563e-01 -3.20496827e-01 1.13378927e-01
-2.33798608e-01 -9.84466076e-02 -4.47019339e-02 -1.40215427e-01
-4.49746609e-01 8.81652758e-02 5.20063564e-02 4.58028167e-01
-1.02659419e-01 9.07693326e-01 -6.41112566e-01 -8.85708570e-01
6.75497651e-01 4.04822469e-01 7.16988416e-03 2.32974425e-01
3.04153800e-01 5.77509582e-01 -5.86418867e-01 9.26787496e-01
9.91452754e-01 -1.32018879e-01 -1.75757967e-02 -1.01923168e+00
-1.99251518e-01 -4.79219139e-01 -1.42491066e+00 1.65702808e+00
-1.83964312e-01 3.72247845e-01 3.50029349e-01 -1.23636591e+00
8.80578160e-01 3.85279000e-01 9.05969203e-01 -9.33236957e-01
4.66928631e-01 3.36420499e-02 -2.29789227e-01 -7.45905042e-01
5.75273275e-01 -2.09858611e-01 1.08968511e-01 3.56714837e-02
1.06329598e-01 -4.89877537e-02 6.59305528e-02 -2.75145471e-01
2.09123895e-01 -2.46728718e-01 3.15139860e-01 -1.01806849e-01
9.24630523e-01 -1.99096307e-01 7.43072331e-01 7.63944864e-01
-1.65941104e-01 6.08953536e-01 -2.79898375e-01 -4.22302246e-01
-5.67254186e-01 -8.97113025e-01 -1.93942696e-01 9.87345636e-01
7.62235880e-01 -8.96181092e-02 -2.56133229e-01 -3.75452936e-01
-1.45252079e-01 2.00865567e-01 -6.24443889e-01 -3.48240554e-01
-3.31552148e-01 -1.07395327e+00 -2.16155924e-04 4.21243727e-01
8.72264802e-01 -7.28760600e-01 -3.23649377e-01 -8.51373374e-02
-5.68787336e-01 -9.87831593e-01 -2.07763001e-01 2.37109195e-02
-7.86746621e-01 -1.13101161e+00 -6.81740522e-01 -1.06033790e+00
6.13682270e-01 9.19434428e-01 8.37978542e-01 3.25768054e-01
-8.85628909e-02 4.84831542e-01 -4.00493830e-01 -2.68826067e-01
-2.53840655e-01 -3.71946275e-01 -1.78044401e-02 3.62201720e-01
2.45171145e-01 -2.90542781e-01 -3.85229051e-01 4.49673027e-01
-1.10671592e+00 2.63806671e-01 6.90296352e-01 9.78578568e-01
6.16902649e-01 4.76387978e-01 3.48603368e-01 -7.07059205e-02
4.79356349e-01 -1.13609582e-01 -4.13006604e-01 8.66832674e-01
-3.75472605e-01 -1.22682303e-01 9.89530906e-02 -2.59499103e-01
-1.42020309e+00 2.17733368e-01 -2.58371606e-02 -3.79410952e-01
-2.80760884e-01 6.36061311e-01 -5.50919950e-01 -5.93054831e-01
4.59400296e-01 3.42321247e-01 2.13689044e-01 -3.79958272e-01
4.75839853e-01 6.93204999e-01 5.80994308e-01 -5.89437485e-01
7.44728088e-01 6.49916708e-01 -1.25186220e-01 -8.18381429e-01
-9.68416333e-01 -4.74901915e-01 -6.91440225e-01 -5.28378487e-01
1.00823414e+00 -1.15177715e+00 -5.91602445e-01 8.13240886e-01
-9.30056930e-01 1.93477467e-01 -1.33612454e-01 6.17748737e-01
-3.99239451e-01 6.74218476e-01 -5.13075888e-01 -7.35698581e-01
-2.31815994e-01 -1.26060081e+00 9.83118176e-01 7.05270588e-01
4.17636245e-01 -9.53989744e-01 2.80804019e-02 6.23633921e-01
4.59471554e-01 2.19298050e-01 6.75462723e-01 -2.99872626e-02
-4.31425184e-01 -8.91823843e-02 -3.13635707e-01 5.15256524e-01
7.19505250e-01 1.64933965e-01 -1.00463104e+00 -4.31107432e-01
3.12688679e-01 -2.69512743e-01 1.11610901e+00 6.69253170e-01
1.27023566e+00 -9.39323455e-02 -5.12687564e-01 6.13985121e-01
1.47084033e+00 4.33433652e-01 7.29070425e-01 3.25822473e-01
5.24698079e-01 5.93000233e-01 6.96370840e-01 1.04098298e-01
3.35593939e-01 4.89507437e-01 4.84549910e-01 -5.77524066e-01
-1.42765924e-01 1.96511298e-01 2.01699287e-01 1.01296747e+00
-3.16565961e-01 -1.00149043e-01 -5.92931986e-01 3.36620659e-01
-2.26238608e+00 -1.16370356e+00 -1.06748007e-01 1.90324605e+00
6.29360676e-01 -4.73074973e-01 -3.06361586e-01 2.83212457e-02
1.06312203e+00 2.54811347e-01 -3.48095387e-01 1.76099107e-01
-6.98496342e-01 -2.97196120e-01 2.69032300e-01 5.06483257e-01
-1.45305383e+00 4.23444688e-01 7.15084362e+00 7.87121654e-01
-1.14932525e+00 4.90466617e-02 6.96729600e-01 2.58013636e-01
-2.30800539e-01 -5.13645969e-02 -7.00169861e-01 2.64910817e-01
2.79738635e-01 4.02779207e-02 2.72457689e-01 2.68665403e-01
-3.91858965e-02 -3.09464067e-01 -7.20410764e-01 1.34542680e+00
1.94722950e-01 -1.26721370e+00 1.24752432e-01 -9.63219255e-02
1.03538084e+00 -4.25218605e-02 1.59247905e-01 4.01115716e-02
1.74031571e-01 -9.20705020e-01 6.50551558e-01 1.03312278e+00
4.27194029e-01 -5.11381924e-01 1.01355386e+00 2.52546459e-01
-1.51410007e+00 -3.39165717e-01 -3.78930688e-01 6.12169914e-02
2.06604913e-01 4.69858795e-01 3.32094222e-01 1.30117059e+00
7.92180717e-01 8.42971802e-01 -7.73943663e-01 1.11725879e+00
7.45624527e-02 -1.88805446e-01 -2.07172737e-01 5.07426023e-01
3.51224504e-02 -3.82610828e-01 2.31541097e-01 9.60544050e-01
5.38582742e-01 1.10984102e-01 4.36356574e-01 6.30980670e-01
3.60459864e-01 -5.72498143e-02 -4.43450570e-01 3.07445973e-01
3.53230864e-01 1.59662867e+00 -2.96369314e-01 -3.55225503e-01
-7.46796787e-01 5.93781352e-01 2.48167440e-02 5.28332710e-01
-8.08010638e-01 -5.33429943e-02 6.49835169e-01 -4.16495442e-01
-3.21999602e-02 -1.35976791e-01 -3.78760099e-01 -1.36606848e+00
-1.87886029e-01 -8.90068948e-01 7.77428389e-01 -1.05938244e+00
-1.49929881e+00 5.45119226e-01 1.50686473e-01 -1.32005334e+00
1.55084029e-01 -6.08174324e-01 -5.16997755e-01 9.92506921e-01
-1.59195697e+00 -1.41830575e+00 -8.46933365e-01 1.09018743e+00
2.52062470e-01 -3.14735860e-01 6.14354789e-01 4.46788490e-01
-4.52266008e-01 3.41313124e-01 2.80080914e-01 -5.42224236e-02
5.81797183e-01 -6.52344346e-01 -5.85137129e-01 7.73390710e-01
-1.57712139e-02 2.86163628e-01 5.30556858e-01 -3.01760286e-01
-1.20391917e+00 -8.99428308e-01 6.15136325e-01 -3.54861841e-02
2.58565992e-01 3.12658250e-01 -9.31746960e-01 4.38361585e-01
6.25356793e-01 1.83401946e-02 7.09735811e-01 6.06839322e-02
-2.42857821e-02 -4.41240251e-01 -1.21647680e+00 3.97952914e-01
6.39006436e-01 -4.79219407e-01 -5.80177665e-01 2.61732697e-01
4.75942552e-01 -3.11469078e-01 -9.71067667e-01 9.72150326e-01
5.55065334e-01 -1.19657898e+00 1.22694623e+00 -3.75526428e-01
7.06315786e-02 -6.67008817e-01 -5.33510387e-01 -1.30381119e+00
-7.36553907e-01 -4.92269211e-02 1.07054107e-01 1.41797173e+00
8.48188326e-02 -7.29660213e-01 2.04918161e-01 5.96192539e-01
7.35909715e-02 -5.51674664e-01 -9.75411773e-01 -4.76308852e-01
-1.94352135e-01 -6.24498874e-02 6.01985633e-01 1.13726449e+00
-3.29099923e-01 2.56294042e-01 -4.68427926e-01 4.03964967e-01
5.63825548e-01 3.85742009e-01 4.09270376e-01 -1.13685131e+00
1.83004424e-01 -6.02076054e-01 -4.39363956e-01 -6.55522108e-01
1.53747395e-01 -4.96995419e-01 -4.71921414e-02 -1.67478740e+00
5.90148926e-01 -3.96231532e-01 -9.19578612e-01 6.41999006e-01
-4.73934472e-01 2.83589482e-01 2.61350185e-01 2.57404894e-01
-5.89618385e-01 7.78128564e-01 1.22631526e+00 -5.53499281e-01
-8.03497657e-02 -9.70559493e-02 -9.33310628e-01 6.77507818e-01
1.03978240e+00 -2.33497215e-03 -2.70896047e-01 -5.17134905e-01
-4.66075614e-02 -1.17835624e-03 4.43963319e-01 -1.05113399e+00
4.89881784e-01 -5.15411198e-01 6.06464863e-01 -8.19962859e-01
5.16964257e-01 -1.05589521e+00 3.54845673e-01 4.09050345e-01
-1.47667706e-01 1.46333193e-02 2.58993328e-01 3.34904253e-01
-6.31554127e-01 -1.86264426e-01 1.13494992e+00 -2.06177562e-01
-1.19469786e+00 2.48424262e-01 -3.66208762e-01 -6.72605574e-01
1.17916620e+00 -4.31473017e-01 -3.52144718e-01 -3.21742594e-01
-7.95603216e-01 4.19920057e-01 3.59886080e-01 6.78225458e-01
7.17904389e-01 -1.73850548e+00 -7.11527526e-01 2.03221366e-01
2.43391678e-01 -3.42961729e-01 7.19296038e-01 8.98437083e-01
-9.75485742e-02 2.17656374e-01 -4.55626488e-01 -7.13937402e-01
-1.68278778e+00 6.87556148e-01 7.20373333e-01 -8.11936036e-02
-1.62670121e-01 6.72167242e-01 4.58647460e-01 -4.78981704e-01
1.42400026e-01 -9.07602832e-02 -5.05781829e-01 2.36047268e-01
5.88679254e-01 3.07785749e-01 1.85805112e-02 -1.04743612e+00
-3.49464595e-01 1.07193470e+00 1.07903816e-01 3.43382657e-01
1.18931496e+00 -5.90186894e-01 -5.74672580e-01 3.07801932e-01
1.05867445e+00 -3.03484023e-01 -9.49580371e-01 -5.20446956e-01
-4.64080572e-01 -6.40211225e-01 4.00748581e-01 -7.74920583e-01
-1.59296870e+00 5.22819996e-01 1.29769123e+00 2.70282090e-01
1.66360652e+00 1.13113895e-01 4.57569480e-01 2.35957190e-01
2.48048723e-01 -1.15262473e+00 -4.82032367e-04 2.50968486e-01
8.90403032e-01 -1.47872746e+00 1.38903096e-01 -4.74499196e-01
-3.87240410e-01 1.18201804e+00 7.37903059e-01 3.63388389e-01
7.71975338e-01 1.64019629e-01 4.22853768e-01 -2.33558014e-01
-4.79955614e-01 -4.84984279e-01 6.48930073e-01 6.48812056e-01
4.30426270e-01 -1.41080901e-01 -2.89687425e-01 1.88119009e-01
3.59142810e-01 -8.97077546e-02 -3.76542099e-02 9.62309659e-01
-7.73548901e-01 -9.38594103e-01 -9.28694844e-01 4.00100797e-01
-2.17776835e-01 3.03766489e-01 -1.15293339e-01 5.18193662e-01
5.63730657e-01 1.41451693e+00 -1.01032019e-01 -6.99743629e-01
1.91519648e-01 -1.84097946e-01 5.42851985e-01 1.05539691e-02
-4.08074528e-01 8.82933140e-02 -1.99169815e-01 -3.85033071e-01
-1.29278207e+00 -4.99608934e-01 -8.14522505e-01 -2.61787593e-01
-7.58490860e-01 5.53185008e-02 4.96550620e-01 9.84160006e-01
2.82127947e-01 6.65593922e-01 7.76021421e-01 -1.01827264e+00
-2.83843964e-01 -8.51573169e-01 -7.03130245e-01 4.54521269e-01
4.15849596e-01 -7.53445387e-01 -1.60367817e-01 2.43520424e-01] | [10.494902610778809, -1.7759027481079102] |
9698684d-b655-4864-8ec9-cbbf6caaf1df | descriptive-analysis-of-computational-methods | 2010.03378 | null | https://arxiv.org/abs/2010.03378v1 | https://arxiv.org/pdf/2010.03378v1.pdf | Descriptive analysis of computational methods for automating mammograms with practical applications | Mammography is a vital screening technique for early revealing and identification of breast cancer in order to assist to decrease mortality rate. Practical applications of mammograms are not limited to breast cancer revealing, identification ,but include task based lens design, image compression, image classification, content based image retrieval and a host of others. Mammography computational analysis methods are a useful tool for specialists to reveal hidden features and extract significant information in mammograms. Digital mammograms are mammography images available along with the conventional screen-film mammography to make automation of mammograms easier. In this paper, we descriptively discuss computational advancement in digital mammograms to serve as a compass for research and practice in the domain of computational mammography and related fields. The discussion focuses on research aiming at a variety of applications and automations of mammograms. It covers different perspectives on image pre-processing, feature extraction, application of mammograms, screen-film mammogram, digital mammogram and development of benchmark corpora for experimenting with digital mammograms. | ['Manish Joshi', 'Aparna Bhale'] | 2020-10-06 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [ 9.58353281e-01 2.95865148e-01 -2.01394483e-01 -5.91976523e-01
-4.81033862e-01 -1.51025951e-01 3.57912034e-01 6.08405054e-01
-3.59570354e-01 2.16396868e-01 1.20485365e-01 -7.70335376e-01
-3.42159182e-01 -8.87260675e-01 -2.69864351e-01 -6.35015130e-01
-3.26286882e-01 6.11089647e-01 4.24764901e-01 -5.49858138e-02
1.96859211e-01 6.82037890e-01 -1.69077182e+00 8.40532660e-01
5.43223955e-02 8.85722339e-01 4.72028971e-01 1.34607089e+00
3.68669331e-02 7.56627858e-01 -2.26277098e-01 -7.78225482e-01
-2.67980192e-02 -5.98346889e-01 -6.65202320e-01 5.32861233e-01
7.53479302e-02 -4.92284775e-01 -3.07057142e-01 1.17652857e+00
5.56986570e-01 -6.33609951e-01 9.15803194e-01 -6.33641183e-01
-5.02494872e-01 4.73592877e-01 -8.03727388e-01 7.84783781e-01
4.69439119e-01 -3.66761655e-01 1.44136578e-01 -7.75944650e-01
4.75874662e-01 1.22316873e+00 6.64447010e-01 5.60336351e-01
-5.36701798e-01 -2.82017887e-01 -1.03931844e+00 3.72887343e-01
-9.29977417e-01 -5.99678278e-01 1.26518384e-01 -3.06617230e-01
5.87059081e-01 1.16124249e+00 8.21560204e-01 2.64476866e-01
9.17722166e-01 8.04402292e-01 1.11736715e+00 -9.51803923e-01
4.10489477e-02 5.85557878e-01 6.07159674e-01 8.82034004e-01
6.72570050e-01 -5.28073832e-02 -3.12910795e-01 -1.64843097e-01
5.53831935e-01 3.25279385e-01 -5.88497985e-03 1.40904859e-01
-9.51576412e-01 6.88247085e-01 1.07922554e-01 7.19272137e-01
-4.31244284e-01 -2.85859376e-01 4.37733293e-01 3.59249294e-01
2.84161299e-01 4.28954549e-02 7.87722990e-02 2.77725726e-01
-1.00019062e+00 2.92588174e-01 6.41012430e-01 8.54965985e-01
2.31407046e-01 -3.42253864e-01 -2.94600934e-01 6.04894400e-01
2.55285293e-01 5.89712381e-01 1.00525439e+00 -5.60308278e-01
1.89286083e-01 7.02394843e-01 -3.24189603e-01 -1.33744800e+00
-6.51503623e-01 3.97358984e-02 -1.06316614e+00 -5.91881946e-02
1.47216037e-01 3.29806834e-01 -1.15828097e+00 3.50659013e-01
2.97219843e-01 -4.36656952e-01 5.74523909e-03 6.38191342e-01
1.33013368e+00 3.75339389e-01 -1.54643744e-01 -3.73577356e-01
2.25044656e+00 -2.08732113e-01 -1.05031800e+00 -1.55572474e-01
5.85532784e-01 -1.07219803e+00 4.34398741e-01 5.16275167e-01
-1.35928035e+00 -4.93176490e-01 -9.49915469e-01 -2.02281892e-01
-3.76138985e-01 4.75355446e-01 7.16836393e-01 1.00802886e+00
-7.43219018e-01 2.10668519e-01 -1.25704288e+00 -7.38714278e-01
8.23349774e-01 3.86619270e-01 -3.82345021e-01 -2.52574354e-01
-5.74281752e-01 8.56647789e-01 5.55590212e-01 -1.20319866e-01
-4.45265412e-01 -5.50709546e-01 -8.78652573e-01 -1.29638925e-01
2.48770416e-01 -2.64354587e-01 1.29778600e+00 -9.52879429e-01
-5.73344886e-01 1.66050899e+00 -3.70535672e-01 -1.03624403e+00
5.27142167e-01 1.83733050e-02 -4.65963483e-01 5.01945972e-01
4.02559601e-02 4.15448606e-01 7.30719447e-01 -5.46497226e-01
-9.42149103e-01 -4.44495887e-01 -6.53080881e-01 -3.14364135e-02
-4.87527877e-01 4.10070240e-01 -6.47630751e-01 -5.28861821e-01
4.58242238e-01 -6.48442924e-01 -3.14069659e-01 -7.99996592e-03
-2.69374400e-01 1.94191992e-01 1.03883302e+00 -8.68765473e-01
1.11764348e+00 -2.00301909e+00 -4.13114130e-01 2.21003413e-01
2.44467005e-01 1.40087783e-01 6.37393415e-01 1.60392672e-01
-3.71324979e-02 1.81400821e-01 -6.53748959e-02 2.04679415e-01
-5.13951719e-01 9.63277072e-02 3.07416469e-01 4.96545225e-01
3.91497374e-01 1.05309319e+00 -5.12657940e-01 -1.17133570e+00
3.53334516e-01 2.17646003e-01 -1.25911847e-01 -1.01980709e-01
2.87767500e-01 1.79846883e-01 -7.40515053e-01 1.19141936e+00
7.27460504e-01 -3.14809352e-01 2.48737156e-01 -4.91104096e-01
1.00684792e-01 -2.76911825e-01 -6.77270830e-01 9.60264444e-01
2.49155313e-01 8.96208167e-01 -2.18164116e-01 -1.04205668e+00
5.29404342e-01 4.16885674e-01 5.56136072e-01 -7.56086886e-01
3.92226100e-01 1.31267935e-01 2.56588161e-01 -1.18526387e+00
5.91236174e-01 9.49651748e-02 3.21215391e-01 2.74879068e-01
-2.64413029e-01 -2.35783920e-01 4.96002436e-01 9.25808176e-02
1.09640646e+00 -3.12249094e-01 1.14730787e+00 -4.15200353e-01
4.89909142e-01 3.37797672e-01 -5.83884344e-02 6.02629006e-01
7.43007138e-02 4.93580669e-01 3.30309093e-01 -6.39122486e-01
-1.24847662e+00 -3.49099040e-01 -7.43177176e-01 5.77732146e-01
-1.25195771e-01 -5.20701744e-02 -7.64609218e-01 -2.82999545e-01
-9.63626653e-02 2.35917211e-01 -9.16287839e-01 1.89083517e-01
-6.95119500e-01 -1.42935610e+00 1.44157961e-01 3.42799544e-01
8.33267927e-01 -8.04280400e-01 -1.19229925e+00 1.21730134e-01
1.99222639e-01 -9.47279811e-01 2.51065463e-01 7.51351267e-02
-1.24314141e+00 -1.41015303e+00 -9.97516155e-01 -9.33015049e-01
1.00476599e+00 6.02018058e-01 1.02914822e+00 5.55307090e-01
-1.21588397e+00 3.16269279e-01 -4.11649019e-01 -1.08415985e+00
-9.35347438e-01 -4.10239190e-01 -5.36782444e-01 -3.00974458e-01
8.51585448e-01 -5.79486750e-02 -5.42565465e-01 2.49894977e-01
-1.28698885e+00 1.86190903e-01 9.92317677e-01 8.08732927e-01
8.18196535e-01 5.83820283e-01 -8.61650258e-02 -1.41034698e+00
3.58215421e-01 -5.73083401e-01 -4.90678102e-01 3.09991747e-01
-3.20486754e-01 -5.07709563e-01 -1.85323089e-01 -3.65332067e-02
-9.49388146e-01 7.47403949e-02 8.47688392e-02 3.89306664e-01
-6.75960407e-02 5.83168805e-01 1.78605691e-01 -2.45175630e-01
8.36285353e-01 3.59067380e-01 1.31567687e-01 -6.94241345e-01
-4.37080085e-01 8.44450474e-01 7.85392582e-01 4.79334056e-01
3.38669509e-01 5.77399969e-01 5.28153360e-01 -1.27021551e+00
-2.21710548e-01 -6.59185469e-01 -6.76444411e-01 -1.16753027e-01
9.26759779e-01 -5.29930592e-01 -4.23299879e-01 1.90461382e-01
-7.13560402e-01 3.82245511e-01 3.28242257e-02 4.76752013e-01
-9.32143182e-02 3.02007496e-01 -6.64083421e-01 -8.01753700e-01
-3.65976244e-01 -9.79411542e-01 1.01213777e+00 3.84770095e-01
-1.17356665e-01 -7.30789006e-01 -5.55889666e-01 5.73809564e-01
2.37873122e-01 3.03015113e-01 9.50331569e-01 -6.37769639e-01
-1.78862482e-01 -8.86791408e-01 -3.10088009e-01 1.52379647e-02
3.36875916e-01 -1.12300247e-01 -9.34155941e-01 -5.94384149e-02
4.02648926e-01 -8.08062479e-02 9.60675836e-01 7.28905916e-01
1.55575347e+00 -1.69966623e-01 -8.70713055e-01 4.33080018e-01
1.40711284e+00 6.09041989e-01 9.57877219e-01 3.51829290e-01
5.63163385e-02 6.71914518e-01 7.57894754e-01 3.13482195e-01
-1.69545747e-02 1.16590425e-01 2.38986775e-01 -2.54159480e-01
-1.98827520e-01 6.15615845e-01 -4.26047862e-01 5.19780934e-01
-5.53513467e-01 -2.60670632e-01 -1.16653168e+00 2.79030085e-01
-1.19491756e+00 -8.19917858e-01 -6.38727963e-01 2.02151275e+00
7.01320112e-01 2.32375190e-01 -7.85575733e-02 7.57142663e-01
6.48585081e-01 -6.67542279e-01 5.31960092e-02 -5.81069179e-02
1.13172740e-01 3.86907190e-01 8.03138494e-01 -2.81024701e-03
-1.46103334e+00 2.83740968e-01 6.48446465e+00 6.12731874e-01
-1.12936628e+00 2.00527385e-01 1.15251827e+00 2.52714157e-01
2.52706707e-01 -5.91520071e-01 -9.01579797e-01 3.57779920e-01
8.82089555e-01 -3.47736254e-02 -4.62725073e-01 7.79715657e-01
1.74402237e-01 -9.98447061e-01 -1.19742346e+00 9.05046523e-01
6.62613884e-02 -1.48298037e+00 1.49645001e-01 1.36475578e-01
5.95074534e-01 -5.63400090e-01 8.15014467e-02 -4.14468378e-01
-5.10378063e-01 -1.16832173e+00 3.43696177e-01 5.95236301e-01
7.38599062e-01 -5.38743734e-01 1.26413882e+00 1.34136632e-01
-8.44337940e-01 1.00284144e-02 -6.16767764e-01 1.89101845e-01
-2.81648070e-01 4.37879294e-01 -1.26444805e+00 4.60427940e-01
9.75247502e-01 3.98236156e-01 -1.06810474e+00 1.57358336e+00
7.52061009e-01 6.05999112e-01 1.81098487e-02 -2.43135273e-01
7.71868303e-02 2.53861070e-01 4.12318468e-01 1.58460724e+00
6.19348049e-01 5.28295875e-01 -1.34449795e-01 1.38529718e-01
1.90277964e-01 2.34071732e-01 -5.41361749e-01 -2.84889728e-01
2.20140532e-01 1.12710667e+00 -1.42510343e+00 -5.25017142e-01
-4.24346924e-01 6.33070230e-01 -7.74551213e-01 -3.03335905e-01
-2.84727931e-01 -3.00164133e-01 -5.75995088e-01 5.60409606e-01
1.05672911e-01 2.28047222e-01 -3.93447340e-01 -4.00676489e-01
-6.79236203e-02 -1.20576787e+00 3.83738935e-01 -4.31106299e-01
-5.62825918e-01 5.76912582e-01 4.43314046e-01 -1.24263883e+00
-4.07297283e-01 -9.06097114e-01 -3.92540812e-01 4.49743122e-01
-1.15893137e+00 -1.25109768e+00 -5.86950541e-01 3.75839233e-01
8.03443134e-01 -5.75704396e-01 9.23279166e-01 3.09004396e-01
-1.05603866e-01 2.61687011e-01 2.86041766e-01 1.78660274e-01
3.92948925e-01 -1.21317613e+00 1.08253313e-02 5.72616935e-01
-2.00247303e-01 -9.67554562e-03 8.24472189e-01 -7.10499346e-01
-1.89640975e+00 -9.20288861e-01 7.60627270e-01 -2.80174404e-01
1.65973991e-01 8.58765095e-02 -7.53590167e-01 4.40588534e-01
2.92599276e-02 6.38392419e-02 6.97755873e-01 -8.18324983e-01
7.13555217e-01 -1.17544383e-01 -1.35180163e+00 3.01672101e-01
2.88632691e-01 1.72914892e-01 -4.00916219e-01 8.39558244e-01
4.67738546e-02 -6.65670097e-01 -7.34664917e-01 2.94631273e-01
6.74604952e-01 -9.76439893e-01 1.04607248e+00 -4.31126356e-01
7.21657276e-01 4.68150407e-01 5.50989807e-03 -2.43193775e-01
-2.33404845e-01 -3.79086912e-01 1.96820050e-01 8.16279292e-01
2.64964908e-01 -1.83957592e-01 9.45579052e-01 4.81079936e-01
1.30909920e-01 -6.49590731e-01 -8.19427192e-01 -1.02931805e-01
-4.96184528e-01 -3.87302816e-01 3.07930291e-01 4.41680908e-01
-3.02080691e-01 -3.61831754e-01 -1.44402444e-01 -8.98163244e-02
6.45717144e-01 -2.14076981e-01 5.59433520e-01 -1.10865664e+00
-4.24380273e-01 -3.00209999e-01 -1.09142947e+00 -3.09436142e-01
-1.04625428e+00 -6.89268887e-01 -1.83304742e-01 -1.31529629e+00
7.50836492e-01 -9.13950652e-02 1.69208348e-01 3.06926161e-01
-5.37233576e-02 8.64040077e-01 -1.99950129e-01 2.23376811e-01
-1.02525406e-01 -8.81430984e-01 1.36209929e+00 -2.48390317e-01
5.84933870e-02 5.12834907e-01 -5.87937474e-01 1.15145242e+00
5.72921991e-01 -7.28466928e-01 -3.25331509e-01 -1.51091903e-01
2.43763432e-01 3.48035038e-01 5.69199085e-01 -9.43472385e-01
3.32143813e-01 -4.21400815e-02 9.75076139e-01 -1.09270358e+00
2.78737307e-01 -8.27659905e-01 3.08132052e-01 9.44210887e-01
-2.77001083e-01 -1.18724272e-01 1.79734707e-01 6.18298292e-01
-2.09702224e-01 -9.83779550e-01 7.42351234e-01 -6.87079906e-01
-7.18567669e-01 -1.93499610e-01 -8.34956229e-01 -6.44148707e-01
1.27474141e+00 -6.85699761e-01 1.91639751e-01 -1.54942051e-02
-7.44519889e-01 -2.94164699e-02 3.22847366e-02 -1.06713392e-01
7.97999442e-01 -1.09161365e+00 -9.93065178e-01 2.99257159e-01
-8.36227015e-02 -1.21514365e-01 1.08587649e-02 1.10786867e+00
-1.20146739e+00 6.22053266e-01 -3.29941869e-01 -6.58696294e-01
-2.48944163e+00 2.08994985e-01 1.30807653e-01 -2.61841476e-01
-7.62320757e-01 9.34524298e-01 -7.15965033e-02 6.35121107e-01
3.86777043e-01 -6.73342049e-01 -7.24513769e-01 -7.45655447e-02
1.17563975e+00 3.36186826e-01 4.99544501e-01 -3.91355902e-01
-2.64671117e-01 4.71039265e-01 -3.74963224e-01 1.98174343e-01
1.42672026e+00 -1.31882995e-01 -3.24966311e-01 2.26743072e-01
1.08182323e+00 -2.53902614e-01 -1.79219589e-01 -2.17645988e-01
3.10256660e-01 -6.26398385e-01 3.61979932e-01 -6.65509582e-01
-8.96291673e-01 8.44006062e-01 1.16086662e+00 7.93792188e-01
1.50653017e+00 1.20536879e-01 2.40365505e-01 8.05855989e-01
4.81816977e-02 -9.34559226e-01 -1.03522681e-01 -3.16506505e-01
1.00760424e+00 -1.69158208e+00 7.77632654e-01 -4.84637290e-01
-5.56641817e-01 1.70318604e+00 -1.10019341e-01 1.13613330e-01
1.04548645e+00 6.53712749e-01 -3.19980651e-01 -4.61067110e-01
-2.81028479e-01 1.15487918e-01 8.38884890e-01 6.20878577e-01
9.03260350e-01 3.07419039e-02 -8.88335407e-01 5.77929616e-01
-2.16411740e-01 2.81803370e-01 3.15809697e-01 1.56751156e+00
-8.41103494e-01 -8.22230816e-01 -1.06408131e+00 1.21959090e+00
-1.36549163e+00 -4.72595682e-03 -4.13378477e-01 1.02487397e+00
3.01600307e-01 6.60179615e-01 1.19095899e-01 5.20274192e-02
-2.05732677e-02 -9.86717045e-02 6.99220598e-01 -6.02659285e-01
-4.11146313e-01 3.99058193e-01 3.42064887e-01 -1.37629405e-01
-7.08772421e-01 -7.14676976e-01 -9.73578334e-01 -1.10603951e-01
-3.70947659e-01 -7.51596987e-02 1.03808689e+00 8.06251824e-01
-3.48523855e-01 6.98127747e-01 6.79664388e-02 -8.18474352e-01
-3.14137757e-01 -1.10122180e+00 -6.43277287e-01 1.53756499e-01
2.31904224e-01 -6.37894124e-02 3.67934495e-01 9.17159855e-01] | [15.198606491088867, -2.5521717071533203] |
cb7522ac-9fa8-4a64-91a6-4ea3bf29e291 | st-pinn-a-self-training-physics-informed | 2306.09389 | null | https://arxiv.org/abs/2306.09389v1 | https://arxiv.org/pdf/2306.09389v1.pdf | ST-PINN: A Self-Training Physics-Informed Neural Network for Partial Differential Equations | Partial differential equations (PDEs) are an essential computational kernel in physics and engineering. With the advance of deep learning, physics-informed neural networks (PINNs), as a mesh-free method, have shown great potential for fast PDE solving in various applications. To address the issue of low accuracy and convergence problems of existing PINNs, we propose a self-training physics-informed neural network, ST-PINN. Specifically, ST-PINN introduces a pseudo label based self-learning algorithm during training. It employs governing equation as the pseudo-labeled evaluation index and selects the highest confidence examples from the sample points to attach the pseudo labels. To our best knowledge, we are the first to incorporate a self-training mechanism into physics-informed learning. We conduct experiments on five PDE problems in different fields and scenarios. The results demonstrate that the proposed method allows the network to learn more physical information and benefit convergence. The ST-PINN outperforms existing physics-informed neural network methods and improves the accuracy by a factor of 1.33x-2.54x. The code of ST-PINN is available at GitHub: https://github.com/junjun-yan/ST-PINN. | ['Jie Liu', 'Enqiang Zhoui', 'Zhichao Wang', 'Xinhai Chen', 'Junjun Yan'] | 2023-06-15 | null | null | null | null | ['pseudo-label', 'self-learning'] | ['miscellaneous', 'natural-language-processing'] | [-2.93692917e-01 -2.07664043e-01 -2.34688371e-01 -2.10613191e-01
-6.06462836e-01 -7.53195211e-02 1.73147291e-01 -2.39679530e-01
-8.33456218e-02 1.16854000e+00 -2.24423200e-01 -2.66077846e-01
-4.70193923e-01 -9.49833393e-01 -8.91929686e-01 -8.26802492e-01
-1.64125487e-01 5.29089451e-01 3.84092450e-01 -1.34056121e-01
3.31204832e-01 7.73179948e-01 -1.10183227e+00 -9.47058434e-04
1.25292730e+00 1.10636818e+00 1.04670428e-01 4.75580156e-01
-1.93992287e-01 8.68316174e-01 -6.36938512e-02 1.09127089e-01
2.43996218e-01 -2.56121188e-01 -7.77623117e-01 -6.40467644e-01
1.39590621e-01 -3.70671600e-01 -7.12513447e-01 9.22224343e-01
7.00067878e-01 4.45204765e-01 8.04833293e-01 -1.26521230e+00
-9.22188461e-01 4.50796902e-01 -5.63453317e-01 2.94156224e-01
-3.01486582e-01 1.76518008e-01 5.50875306e-01 -1.12472045e+00
3.41477901e-01 1.10346937e+00 1.12070763e+00 6.34231448e-01
-8.25583279e-01 -9.40143168e-01 -2.89783865e-01 1.47497132e-01
-1.40468884e+00 4.98085022e-02 1.02477145e+00 -4.98031497e-01
4.40416485e-01 8.48923251e-03 5.75892389e-01 9.67509270e-01
4.41708684e-01 7.17932582e-01 1.24228597e+00 -1.67296633e-01
5.24852991e-01 5.98702133e-02 4.11385387e-01 8.41196775e-01
2.05095857e-01 5.49416006e-01 -4.61978853e-01 -3.55183482e-01
1.39595163e+00 -1.83604002e-01 -2.52650082e-01 -3.02844733e-01
-8.45956504e-01 8.49560738e-01 8.05572510e-01 1.30159780e-01
-2.74226397e-01 4.24380302e-01 2.90268242e-01 -2.12809950e-01
7.75619209e-01 5.53705812e-01 -5.53997755e-01 8.70683193e-02
-7.85486341e-01 4.00545239e-01 8.07156742e-01 7.26528227e-01
8.14036787e-01 4.35755461e-01 -2.32043266e-01 8.57592404e-01
6.37094826e-02 5.09818852e-01 3.13959688e-01 -1.43698096e+00
-2.34787881e-01 4.71540272e-01 1.46742776e-01 -8.57025027e-01
-3.36228043e-01 -5.98422110e-01 -1.28199804e+00 4.60271388e-01
9.03535560e-02 -5.17774463e-01 -9.55098629e-01 1.45227802e+00
5.62193215e-01 8.19227636e-01 -2.07795620e-01 1.00553107e+00
1.10329819e+00 1.02502728e+00 7.16648176e-02 1.04925700e-01
9.23737586e-01 -1.18858325e+00 -2.84711808e-01 2.03021049e-01
5.72758019e-01 -4.77614701e-01 1.06460059e+00 2.75647223e-01
-8.45331788e-01 -8.41802239e-01 -9.76874948e-01 1.96813062e-01
-1.44065797e-01 -2.79969089e-02 8.30062151e-01 9.88923162e-02
-9.61366177e-01 1.41315281e+00 -9.07348514e-01 -2.93547194e-02
5.98915815e-01 3.02963972e-01 1.31757185e-01 2.38355756e-01
-1.29921412e+00 5.39694488e-01 5.09016931e-01 -1.13918625e-01
-7.47606993e-01 -1.35089684e+00 -5.13349652e-01 -9.98731479e-02
1.83800384e-01 -7.08129227e-01 1.27994382e+00 -6.12269819e-01
-1.80256212e+00 4.53561664e-01 1.47515893e-01 -1.67829812e-01
3.67702901e-01 -3.63148481e-01 -3.44789237e-01 1.82252511e-01
6.38657287e-02 5.76964617e-01 5.33712983e-01 -1.51642430e+00
-2.84470886e-01 3.09802026e-01 -1.56187043e-01 7.39384666e-02
-1.60522327e-01 -3.58242691e-01 -3.80045623e-01 -7.60774314e-01
-8.68877769e-02 -9.88037586e-01 -4.60224241e-01 3.83845270e-01
-4.52299744e-01 -4.35669929e-01 9.88913000e-01 -3.19536179e-01
6.23072267e-01 -1.91689003e+00 -2.31106639e-01 -6.03523180e-02
4.48763847e-01 6.40199244e-01 1.12385325e-01 2.38246694e-01
-7.74872303e-02 -1.74769778e-02 -5.15126884e-01 -1.90041587e-01
-3.41150425e-02 3.55788320e-01 -1.64963797e-01 3.89991313e-01
1.74336717e-01 1.00918007e+00 -9.42933381e-01 -6.23126984e-01
4.21574950e-01 6.16729319e-01 -3.42901289e-01 2.71136492e-01
-3.27099591e-01 1.00283360e+00 -8.95839036e-01 5.13599455e-01
9.78949070e-01 -7.22588301e-01 -5.76309502e-01 -2.34810203e-01
-2.16992959e-01 -3.07234433e-02 -1.11224043e+00 1.56799662e+00
-3.87411803e-01 4.94011492e-01 9.18822736e-03 -1.18588316e+00
8.58207285e-01 1.83818012e-01 8.23507488e-01 -5.45726895e-01
3.75862330e-01 3.91357601e-01 -1.18082024e-01 -4.81272101e-01
2.45794222e-01 -1.57749534e-01 2.80996799e-01 3.25984746e-01
6.31672144e-02 -4.84870821e-02 -1.01349838e-01 1.13666371e-01
9.96124625e-01 4.17253911e-01 1.67782437e-02 -7.26433873e-01
6.75408363e-01 1.95264682e-01 9.10801589e-01 8.79040003e-01
-2.47708082e-01 5.38523138e-01 -7.69363195e-02 -7.85368621e-01
-1.07505298e+00 -1.12786674e+00 -5.01882136e-01 8.39340270e-01
3.18725348e-01 3.43833980e-03 -6.14488900e-01 -2.25944519e-01
1.99975759e-01 6.44201100e-01 -4.30531830e-01 -3.09090674e-01
-8.01408589e-01 -6.76317692e-01 3.36654127e-01 6.06686473e-01
9.17242527e-01 -1.36666322e+00 -9.66805890e-02 3.16390395e-01
2.08442196e-01 -1.05017972e+00 -1.84575334e-01 -4.17013653e-02
-1.03786755e+00 -8.77948344e-01 -9.39132869e-01 -6.83536172e-01
4.83860403e-01 -1.80314168e-01 1.09038913e+00 1.14894055e-01
-3.81036878e-01 1.66786954e-01 -2.71352291e-01 -3.33938092e-01
-3.54122460e-01 2.75484055e-01 1.56178370e-01 -2.34887138e-01
-1.24023736e-01 -1.04167163e+00 -9.18253422e-01 1.08383089e-01
-3.47905755e-01 1.66617796e-01 6.82405233e-01 6.59991026e-01
7.43775487e-01 1.12821251e-01 4.75239992e-01 -9.36571240e-01
4.64268774e-01 -5.42555690e-01 -5.99127650e-01 -1.11252256e-01
-5.37233829e-01 6.72712922e-02 9.66397524e-01 -6.68414176e-01
-1.14670289e+00 -1.65298358e-01 -3.89178455e-01 -7.21752644e-01
-3.34113121e-01 3.48898172e-01 3.34388822e-01 -7.89074004e-01
6.60964906e-01 1.05005890e-01 -3.72751087e-01 -6.54815197e-01
4.82322946e-02 3.75217110e-01 7.41625726e-01 -1.13228321e+00
8.67298484e-01 2.19369099e-01 3.58164817e-01 -8.52985620e-01
-1.25846326e+00 -3.13129991e-01 -4.05933648e-01 -3.33629847e-01
6.81512415e-01 -7.59763777e-01 -8.54663193e-01 7.81904936e-01
-1.09781575e+00 -6.66553617e-01 -4.41755831e-01 6.36250496e-01
-4.07936156e-01 3.21276486e-01 -1.04832792e+00 -7.58975804e-01
-7.06548750e-01 -9.79601920e-01 7.43780971e-01 6.31250858e-01
-2.76911445e-02 -1.27822292e+00 3.29401232e-02 -5.51966056e-02
5.62612116e-01 4.84310955e-01 8.83986652e-01 -4.40790832e-01
-5.91782212e-01 9.03213695e-02 -4.96417224e-01 2.98199922e-01
1.66271046e-01 1.13716744e-01 -1.02726495e+00 -2.23385200e-01
2.38835365e-01 -4.41877186e-01 7.46628761e-01 7.45569587e-01
1.65886295e+00 -1.24563903e-01 -5.41370511e-01 9.22090769e-01
1.63353908e+00 2.44596928e-01 2.71292150e-01 5.30683585e-02
1.04128146e+00 4.27781306e-02 1.89876020e-01 4.27953422e-01
2.22356960e-01 1.94159403e-01 2.26159453e-01 -4.18124914e-01
-3.16359669e-01 -7.42236823e-02 -1.87264442e-01 1.09627676e+00
-1.13222189e-01 -2.12083578e-01 -1.02281451e+00 3.58881861e-01
-1.85658050e+00 -4.56911802e-01 -4.72861141e-01 1.70265985e+00
8.72945607e-01 2.14695752e-01 -3.47641349e-01 -8.18621367e-02
8.75563562e-01 -9.28101167e-02 -1.11961889e+00 -2.11103469e-01
2.38997102e-01 5.64769506e-01 5.03351450e-01 6.05658054e-01
-1.14619935e+00 8.12368155e-01 5.75922012e+00 1.03538001e+00
-1.25974894e+00 2.21588492e-01 7.54112363e-01 2.15059862e-01
2.72917211e-01 2.25421712e-02 -8.08808804e-01 6.06616676e-01
8.03902209e-01 -1.91190571e-01 1.16944514e-01 1.11091864e+00
3.44586283e-01 -3.01626660e-02 -9.33671057e-01 8.52746487e-01
-4.88829464e-01 -1.75560951e+00 7.63110146e-02 -3.98390032e-02
1.02102709e+00 3.21034282e-01 -9.37432349e-02 2.46319994e-01
5.53936541e-01 -8.79700541e-01 3.55696261e-01 8.62990558e-01
7.09527373e-01 -7.47133970e-01 8.49247992e-01 3.93145114e-01
-1.12361860e+00 2.90602356e-01 -4.42575037e-01 -2.49428257e-01
2.13485420e-01 1.02046549e+00 -9.14146230e-02 6.75214052e-01
9.01813149e-01 8.92418623e-01 -4.83690836e-02 1.27274609e+00
4.98326756e-02 1.00553751e+00 -3.65551114e-01 -8.82090107e-02
2.54518062e-01 -2.69076794e-01 5.73556125e-01 8.63317728e-01
4.10701126e-01 5.34055293e-01 2.66269594e-01 1.24351156e+00
-6.03950657e-02 -9.32254419e-02 -2.69533992e-01 1.64718702e-01
4.90728736e-01 1.33281648e+00 -7.38630831e-01 -4.57947820e-01
4.96143363e-02 6.32808864e-01 3.53775382e-01 3.24519485e-01
-9.76560891e-01 -4.02337581e-01 4.58034754e-01 -2.45021880e-02
7.43787587e-02 -3.23188961e-01 -4.36290860e-01 -8.31303477e-01
-3.95498604e-01 -2.33572826e-01 -3.86524200e-03 -1.06997085e+00
-1.79064870e+00 5.94805717e-01 6.43873662e-02 -1.06160367e+00
3.62655252e-01 -9.18487728e-01 -9.21515226e-01 9.25963044e-01
-1.37399399e+00 -8.05474758e-01 -5.55164695e-01 3.79918188e-01
3.18981230e-01 -1.25427559e-01 7.47024119e-01 3.02832127e-01
-6.20538592e-01 2.76945263e-01 4.93224919e-01 3.39767873e-01
4.20957446e-01 -1.06376743e+00 3.47956628e-01 4.15242642e-01
-5.25978625e-01 4.95566010e-01 7.60208786e-01 -7.82123446e-01
-1.17382109e+00 -1.19740045e+00 2.08622754e-01 -9.07588284e-03
6.31501079e-01 -1.13310844e-01 -1.23192561e+00 2.69409120e-01
2.53211766e-01 4.11417872e-01 3.94196987e-01 -2.75415480e-01
1.35660723e-01 -1.69493571e-01 -1.19897187e+00 3.69439244e-01
1.04837978e+00 -1.21983856e-01 -2.44175375e-01 6.69126093e-01
9.18322921e-01 -7.59157121e-01 -1.07739484e+00 8.19407582e-01
2.73834288e-01 -9.47121799e-01 1.06214738e+00 -8.44905227e-02
7.01247394e-01 -1.04858503e-01 1.79875284e-01 -1.22759938e+00
-5.79343498e-01 -4.17174190e-01 -3.60356003e-01 1.17345643e+00
2.83325389e-02 -9.07443464e-01 1.12995195e+00 4.92430985e-01
-4.59312111e-01 -1.29322398e+00 -7.51451194e-01 -9.93331015e-01
5.55944562e-01 -3.67478222e-01 5.02150834e-01 1.06251657e+00
-6.49226129e-01 -1.13496855e-02 -2.32815623e-01 3.61616224e-01
1.03368509e+00 1.43778011e-01 3.74638081e-01 -1.44258761e+00
-2.32994795e-01 -3.79049182e-01 3.64869535e-02 -9.47286010e-01
3.00443500e-01 -8.11667442e-01 1.40246032e-02 -1.66752863e+00
6.13827035e-02 -9.59348023e-01 -4.29293811e-01 3.89028311e-01
-9.54333395e-02 2.56729364e-01 -2.66596466e-01 4.47465807e-01
-2.99757749e-01 8.30796301e-01 1.73957443e+00 6.11238517e-02
-1.70114458e-01 -1.61117613e-02 -4.55499172e-01 9.14049327e-01
1.40422988e+00 -6.23992920e-01 -4.16447550e-01 -2.48656720e-01
-2.57795751e-01 -1.88720354e-03 6.55527592e-01 -1.51481724e+00
3.68905157e-01 -3.54646385e-01 6.07354283e-01 -8.21066082e-01
4.76145953e-01 -3.81179571e-01 8.38701520e-03 6.85420454e-01
-1.36672258e-01 -1.74853966e-01 4.57478255e-01 2.36232847e-01
1.70516908e-01 -3.44190747e-01 9.90638793e-01 -3.71693939e-01
-7.40243614e-01 7.65553534e-01 -2.81618863e-01 2.19487339e-01
7.77030528e-01 -4.75003123e-02 -1.19862810e-01 -8.82800147e-02
-3.58701438e-01 2.55193442e-01 2.12939262e-01 -3.16488072e-02
4.95960712e-01 -1.49199688e+00 -3.93764377e-01 2.00050026e-02
-3.68979514e-01 4.77123469e-01 5.38180530e-01 5.24241865e-01
-8.45570266e-01 2.26971969e-01 -2.55714804e-01 -7.33473063e-01
-4.10515159e-01 1.97813779e-01 6.68192267e-01 -2.08418354e-01
-8.31455767e-01 1.00913143e+00 1.67264715e-01 -9.72570181e-01
6.12442940e-02 -1.80390745e-01 1.18000746e-01 -7.91930139e-01
6.27739951e-02 7.06491768e-01 -2.39416689e-01 -3.67232144e-01
-1.20699264e-01 8.05479527e-01 4.71334420e-02 1.89666599e-01
1.56660616e+00 2.99620599e-01 -1.65545210e-01 4.34120774e-01
1.21830344e+00 -2.97715038e-01 -1.52968144e+00 -3.42824399e-01
-2.58508384e-01 5.21132164e-02 3.00231636e-01 -7.90445030e-01
-1.34041309e+00 7.85895944e-01 6.52975023e-01 -1.98100358e-01
8.75500798e-01 -1.70597151e-01 1.25016117e+00 3.37560713e-01
2.38804698e-01 -8.88300896e-01 1.83435649e-01 7.04395056e-01
6.69644177e-01 -1.19230986e+00 1.96359783e-01 -5.32142401e-01
-1.18606783e-01 1.08826351e+00 1.08035266e+00 -6.55344129e-01
1.13901031e+00 4.34513539e-01 6.36959029e-03 -2.79252619e-01
-3.33723933e-01 2.91943163e-01 2.64369190e-01 4.14261550e-01
3.11432302e-01 -1.15311004e-01 -1.50115967e-01 4.71866250e-01
-1.72476485e-01 3.71378839e-01 1.85084760e-01 9.87243891e-01
-3.75185698e-01 -1.05573142e+00 -1.47709385e-01 6.38074994e-01
-8.79632309e-02 -1.26296422e-02 1.57865509e-02 8.17066133e-01
2.79720366e-01 4.72451001e-01 -1.56090721e-01 -3.77128839e-01
-1.72545239e-02 -1.31782874e-01 3.13651592e-01 -3.44026923e-01
-1.77762851e-01 -2.53085226e-01 -3.53758961e-01 -3.82009149e-01
-4.82888013e-01 -2.04410717e-01 -1.78510332e+00 -6.36372447e-01
-7.54402056e-02 4.92615163e-01 3.52391154e-01 1.02454567e+00
6.10522866e-01 7.42927194e-01 5.13988376e-01 -1.29620659e+00
-3.15679431e-01 -7.72118866e-01 -6.25280201e-01 2.16137003e-02
2.49882281e-01 -9.94706154e-01 -4.03891146e-01 -2.69591749e-01] | [6.5172247886657715, 3.5213773250579834] |
39526e31-36de-42a1-817f-f7bf828a1bc6 | risk-aware-scene-sampling-for-dynamic | 2202.13510 | null | https://arxiv.org/abs/2202.13510v1 | https://arxiv.org/pdf/2202.13510v1.pdf | Risk-Aware Scene Sampling for Dynamic Assurance of Autonomous Systems | Autonomous Cyber-Physical Systems must often operate under uncertainties like sensor degradation and shifts in the operating conditions, which increases its operational risk. Dynamic Assurance of these systems requires designing runtime safety components like Out-of-Distribution detectors and risk estimators, which require labeled data from different operating modes of the system that belong to scenes with adverse operating conditions, sensors, and actuator faults. Collecting real-world data of these scenes can be expensive and sometimes not feasible. So, scenario description languages with samplers like random and grid search are available to generate synthetic data from simulators, replicating these real-world scenes. However, we point out three limitations in using these conventional samplers. First, they are passive samplers, which do not use the feedback of previous results in the sampling process. Second, the variables to be sampled may have constraints that are often not included. Third, they do not balance the tradeoff between exploration and exploitation, which we hypothesize is necessary for better search space coverage. We present a scene generation approach with two samplers called Random Neighborhood Search (RNS) and Guided Bayesian Optimization (GBO), which extend the conventional random search and Bayesian Optimization search to include the limitations. Also, to facilitate the samplers, we use a risk-based metric that evaluates how risky the scene was for the system. We demonstrate our approach using an Autonomous Vehicle example in CARLA simulation. To evaluate our samplers, we compared them against the baselines of random search, grid search, and Halton sequence search. Our samplers of RNS and GBO sampled a higher percentage of high-risk scenes of 83% and 92%, compared to 56%, 66% and 71% of the grid, random and Halton samplers, respectively. | ['Abhishek Dubey', 'Gabor Karsai', 'Yogesh Barve', 'Baiting Luo', 'Shreyas Ramakrishna'] | 2022-02-28 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 1.10978559e-01 3.13301325e-01 -8.12240392e-02 -1.02060549e-01
-5.80079138e-01 -5.05940735e-01 5.56893349e-01 -6.48402423e-02
-2.17863828e-01 9.73203838e-01 -2.13792428e-01 -4.92052466e-01
-3.16324204e-01 -9.37108397e-01 -6.78218961e-01 -7.12466300e-01
-2.80575454e-01 2.15598151e-01 6.74440920e-01 -7.61729386e-03
1.01008430e-01 5.09572029e-01 -1.82460999e+00 -2.93585360e-01
7.60227799e-01 1.01790071e+00 2.87656188e-01 5.85811853e-01
4.07980651e-01 3.89661223e-01 -7.99523950e-01 2.50402838e-01
5.08718371e-01 -2.20239013e-01 -8.28395635e-02 -9.72562358e-02
-3.29888970e-01 -6.03907287e-01 -1.02150038e-01 1.06967866e+00
2.05601901e-01 2.16413140e-01 6.17051184e-01 -1.79566824e+00
3.92346680e-01 3.33853930e-01 -4.81757492e-01 -1.46152273e-01
2.22219154e-01 8.53067219e-01 2.98296183e-01 -4.49280292e-01
4.04913485e-01 1.31499219e+00 3.74508739e-01 4.21188384e-01
-1.12870407e+00 -9.60425198e-01 2.48028114e-01 2.44946387e-02
-1.51910949e+00 -3.76847833e-01 6.20104849e-01 -6.40620589e-01
8.48567724e-01 3.00989002e-01 7.18483806e-01 1.31660438e+00
5.24315119e-01 3.59450400e-01 1.21208024e+00 -4.05057184e-02
9.15002167e-01 2.67061174e-01 3.85843799e-03 3.13244015e-01
6.47712409e-01 9.24544215e-01 -2.65508682e-01 -3.93830806e-01
3.86699051e-01 -1.39136732e-01 -3.25281084e-01 -5.67914426e-01
-8.56011927e-01 6.02055609e-01 5.26972748e-02 -4.20801044e-01
-3.43118906e-01 2.30029881e-01 2.72548437e-01 1.25863567e-01
-4.61563990e-02 6.34693503e-01 -3.35836172e-01 -2.18920484e-01
-6.06395602e-01 2.80678958e-01 8.15325439e-01 1.10328734e+00
8.22063863e-01 2.84062624e-01 -1.09659687e-01 1.81916729e-01
6.06237769e-01 6.82008266e-01 1.27562046e-01 -1.04988670e+00
3.14790756e-01 4.42232817e-01 5.65332532e-01 -6.17519915e-01
-3.74933183e-01 -1.40008211e-01 -5.35878539e-01 8.42971742e-01
1.51997030e-01 -4.60311294e-01 -1.07782257e+00 1.65406275e+00
4.58749801e-01 1.39185667e-01 2.65654445e-01 8.08148980e-01
-1.05345264e-01 6.96985722e-01 -2.08822578e-01 -4.71305877e-01
9.76448536e-01 -3.83583218e-01 -8.15132737e-01 -4.49465007e-01
3.23847860e-01 -3.97438198e-01 1.14945531e+00 5.50191760e-01
-6.59065485e-01 -3.41462046e-01 -1.66197693e+00 1.05715501e+00
-2.46114001e-01 -3.04673374e-01 2.84081459e-01 9.08572435e-01
-7.46359706e-01 4.21205521e-01 -1.01417542e+00 -2.90727437e-01
1.18294261e-01 4.51243296e-02 2.18431875e-01 1.45106882e-01
-1.14946532e+00 1.14073086e+00 3.42288852e-01 2.87241712e-02
-1.72028267e+00 -6.64742708e-01 -9.04024661e-01 -4.59029973e-02
9.22140062e-01 -2.48704299e-01 1.22087324e+00 -2.33839646e-01
-1.74266088e+00 -9.87497494e-02 1.81878537e-01 -5.50898433e-01
5.10607243e-01 -2.45327532e-01 -3.37526947e-01 -2.51458257e-01
-1.80432722e-01 3.61811072e-01 6.61558151e-01 -1.55275893e+00
-5.10538518e-01 4.35972363e-02 1.22832850e-01 7.27924854e-02
1.13585800e-01 -1.91734165e-01 -1.87578145e-02 -1.28182232e-01
-1.21406056e-01 -1.00804043e+00 -6.94141448e-01 -1.47647172e-01
-6.10876620e-01 2.73123294e-01 9.05338645e-01 -2.70547897e-01
1.17028499e+00 -2.33247113e+00 -2.78466403e-01 6.23723388e-01
-2.86238253e-01 2.31219158e-02 1.65528163e-01 6.80203676e-01
1.73109666e-01 2.29258806e-01 -1.90352231e-01 -2.31761914e-02
7.12748766e-02 3.11307967e-01 -5.70132256e-01 4.05817956e-01
2.23499179e-01 1.10585339e-01 -8.53371680e-01 -2.55069166e-01
4.78006601e-01 -9.87223014e-02 -4.46567655e-01 2.82277644e-01
-4.42579657e-01 4.59674597e-01 -5.44344008e-01 5.73460340e-01
6.40615165e-01 3.09081852e-01 1.29663035e-01 -4.84035164e-02
-1.50470898e-01 1.85771018e-01 -1.56249416e+00 8.61420035e-01
-4.19291645e-01 4.29504484e-01 1.49094671e-01 -4.46428210e-01
8.13028932e-01 1.83017239e-01 2.09989682e-01 -6.62386894e-01
1.52292728e-01 4.96451408e-02 -2.05643084e-02 -3.32929283e-01
4.85772669e-01 8.37191269e-02 -3.05897415e-01 2.63238698e-01
-4.54752147e-01 -7.60427237e-01 3.17324512e-02 -3.20927352e-02
1.39268017e+00 9.93213952e-02 3.37493807e-01 -3.29745770e-01
2.20814779e-01 5.49274199e-02 9.46197569e-01 7.76544571e-01
-2.29360417e-01 2.28287697e-01 8.15498054e-01 -3.52185406e-02
-8.20360661e-01 -1.16202140e+00 -1.25879198e-01 8.18458349e-02
6.84637904e-01 -3.77676249e-01 -5.95544100e-01 -5.27966559e-01
-6.82450682e-02 1.34599507e+00 -3.57789308e-01 -5.96730769e-01
-2.55879700e-01 -6.55825019e-01 4.23943877e-01 3.39084417e-01
2.57930130e-01 -6.67647302e-01 -1.41740751e+00 1.32603228e-01
3.29400718e-01 -9.16559219e-01 -1.79406583e-01 4.37581182e-01
-6.11029327e-01 -1.27825153e+00 1.38163000e-01 1.38983890e-01
7.79417932e-01 6.39582723e-02 6.62442684e-01 -2.55165130e-01
-3.14239144e-01 2.54754752e-01 -2.37238675e-01 -7.69076347e-01
-8.01935792e-01 -4.63768512e-01 3.39730591e-01 -2.16369912e-01
-1.73647612e-01 -4.07833368e-01 -4.35939282e-01 8.92546296e-01
-8.84418070e-01 -8.52397457e-02 6.24747455e-01 7.48582780e-01
5.48617482e-01 5.28829932e-01 3.79183769e-01 -4.55207109e-01
5.67270398e-01 -5.93885005e-01 -1.44932091e+00 6.10323995e-02
-9.41573083e-01 1.92088589e-01 4.82131660e-01 -7.03300238e-01
-9.01078105e-01 1.28782287e-01 2.75865346e-01 -6.10983372e-01
-1.52374595e-01 4.89756733e-01 -5.16054213e-01 2.28504121e-01
7.27778792e-01 -1.78177983e-01 2.58183330e-01 -7.79362395e-02
-4.92222048e-02 5.07210672e-01 8.74776617e-02 -5.48492849e-01
1.12951005e+00 3.48815024e-01 2.16401462e-02 -5.31306446e-01
-3.28713864e-01 -3.78943309e-02 1.56862110e-01 -5.97993195e-01
5.72051585e-01 -6.76586747e-01 -7.09065199e-01 4.48930591e-01
-7.55234241e-01 -6.65085256e-01 -6.04164779e-01 5.72518051e-01
-6.17705405e-01 4.60906848e-02 -4.48781885e-02 -1.46615577e+00
2.65874475e-01 -1.59371018e+00 9.50447857e-01 2.86928892e-01
-2.75120497e-01 -3.41492832e-01 -3.71604860e-02 -2.37789556e-01
5.21940649e-01 6.31896377e-01 7.56474376e-01 -3.81206036e-01
-9.67699707e-01 -2.32591018e-01 3.24401408e-01 2.42040619e-01
1.15516715e-01 2.37725616e-01 -8.53120208e-01 -6.33214116e-01
2.46218234e-01 -1.98652193e-01 2.76999831e-01 2.27199569e-01
9.30402637e-01 -4.46770519e-01 -6.35884523e-01 3.54165673e-01
1.42478371e+00 7.82379210e-01 7.27494478e-01 4.61664140e-01
1.95137295e-03 7.16666698e-01 1.38284636e+00 7.12494671e-01
1.02728955e-01 7.89197385e-01 1.00361478e+00 1.20608017e-01
3.68271858e-01 -2.77932614e-01 8.07886660e-01 -1.28161743e-01
4.58784848e-01 -1.97813123e-01 -9.29668486e-01 5.08374333e-01
-1.71186042e+00 -7.20417380e-01 1.10016376e-01 2.72937298e+00
7.26737022e-01 6.81430519e-01 1.55887520e-02 1.46450028e-01
6.21988595e-01 1.66533724e-03 -8.75419617e-01 -2.49992073e-01
3.77428859e-01 -2.36617044e-01 1.10662687e+00 5.92497349e-01
-7.49062657e-01 3.70137751e-01 6.29455805e+00 6.84777856e-01
-7.46004164e-01 -1.56505600e-01 4.46481586e-01 -2.40354583e-01
-2.11411968e-01 4.53264624e-01 -9.69844580e-01 6.39434874e-01
1.43841147e+00 -3.60801756e-01 3.82382661e-01 1.00630569e+00
5.98231018e-01 -7.40786433e-01 -1.27042365e+00 5.32384872e-01
-3.51070762e-01 -9.11720455e-01 -5.28476119e-01 3.09736907e-01
5.46153665e-01 -3.98066193e-02 -1.47977799e-01 3.70577157e-01
9.12474453e-01 -8.82824540e-01 1.10011637e+00 4.75602955e-01
4.81700510e-01 -7.69731224e-01 9.08637822e-01 6.27524674e-01
-8.55926573e-01 -2.58810848e-01 -2.36515962e-02 4.60077412e-02
6.04893506e-01 7.29149759e-01 -1.02651215e+00 4.52945173e-01
8.22204053e-01 -1.30976602e-01 -3.71172130e-01 1.13737440e+00
-2.40539849e-01 6.66648209e-01 -7.12529957e-01 -1.99008673e-01
-1.10023089e-01 1.72810245e-03 8.91910195e-01 5.48629999e-01
4.11941022e-01 -2.52914310e-01 4.38306630e-01 1.03428137e+00
8.36255491e-01 -7.27637112e-01 -8.62479210e-01 2.13112801e-01
9.26402986e-01 8.48126113e-01 -5.34829497e-01 -1.70227781e-01
-9.31692570e-02 3.80927883e-02 -5.49700677e-01 4.85050529e-01
-1.33527112e+00 -4.02629912e-01 9.33703363e-01 2.17486843e-01
-6.78715333e-02 -3.59491259e-01 -2.76889741e-01 -4.35330153e-01
3.66874523e-02 -9.88554418e-01 8.48668590e-02 -7.12592721e-01
-9.41202402e-01 5.21885574e-01 9.37179267e-01 -1.64277577e+00
-5.83946168e-01 -4.44263935e-01 -4.14245903e-01 8.01386595e-01
-1.22376442e+00 -4.21538621e-01 -4.74720508e-01 2.41416872e-01
5.89122415e-01 -1.80250015e-02 3.54932457e-01 -3.62816304e-02
-7.06258953e-01 2.49508038e-01 -6.16620295e-02 -5.11350632e-01
5.22877812e-01 -8.56877029e-01 2.41172552e-01 1.01370931e+00
-7.68992126e-01 4.61292654e-01 1.20222378e+00 -8.78139079e-01
-1.50806129e+00 -1.18369436e+00 -8.93805027e-02 -2.52237290e-01
7.49387026e-01 -5.17297566e-01 -6.68209493e-01 5.30672073e-01
-1.39250476e-02 -3.36441547e-01 4.60716113e-02 -2.57600814e-01
7.25452974e-02 -2.35906720e-01 -1.39717960e+00 9.14078474e-01
6.93891466e-01 -1.13815308e-01 -2.44737312e-01 2.43581846e-01
8.49896848e-01 -3.89861286e-01 -6.26424611e-01 9.04521883e-01
3.19343060e-01 -9.01713848e-01 6.45966172e-01 2.92103961e-02
-2.57757157e-01 -8.03175688e-01 -2.66341984e-01 -1.37681103e+00
2.03602061e-01 -8.45845580e-01 -9.81341600e-02 1.20335734e+00
5.97454190e-01 -1.16584933e+00 5.85584521e-01 6.97864890e-01
-3.67792785e-01 -7.55212188e-01 -1.18466771e+00 -1.35089612e+00
-3.22193474e-01 -5.95476687e-01 7.85701096e-01 3.85635495e-01
-1.71273947e-01 -1.18763626e-01 -2.05935448e-01 6.67929053e-01
6.29111171e-01 -3.69047433e-01 9.35028315e-01 -8.52998376e-01
-3.61300200e-01 -1.78655088e-01 -3.24827790e-01 -4.69679356e-01
-1.56476974e-01 8.86209831e-02 6.94963992e-01 -1.20610058e+00
-6.99682012e-02 -5.73488891e-01 -1.93925668e-02 3.18848133e-01
1.10516526e-01 -5.69671810e-01 -8.09511915e-02 1.55294463e-01
-1.42457157e-01 8.67300928e-01 6.56605065e-01 -2.11360484e-01
-3.90507609e-01 2.00625375e-01 -1.70584232e-01 8.41366112e-01
7.75169492e-01 -5.24423420e-01 -9.11211491e-01 1.47813903e-02
7.08350018e-02 2.91411281e-01 4.50008780e-01 -1.36919749e+00
1.23452045e-01 -7.74573743e-01 -1.59977809e-01 -7.69586325e-01
4.77479786e-01 -1.16258585e+00 7.64620781e-01 7.57995188e-01
2.73549892e-02 1.82930365e-01 4.64924663e-01 8.20436895e-01
-1.02785043e-01 -2.89395779e-01 8.10814261e-01 3.47979888e-02
-6.26311123e-01 -4.04114127e-02 -9.54197705e-01 -2.36208931e-01
1.65421999e+00 -5.11991560e-01 -4.04309541e-01 -4.01182622e-01
-3.03205490e-01 6.62416697e-01 7.67238200e-01 4.63978618e-01
6.52837276e-01 -9.63075995e-01 -1.82153076e-01 1.97571427e-01
2.83874333e-01 2.55230010e-01 6.92698434e-02 6.80876493e-01
-2.75318533e-01 8.13316032e-02 2.85100751e-02 -6.17936730e-01
-9.01195407e-01 6.94348276e-01 4.17908669e-01 -1.67288736e-01
-3.78887653e-01 3.27769071e-01 2.73324490e-01 -3.73356074e-01
3.59013140e-01 -4.69982535e-01 1.14891358e-01 -1.28207564e-01
2.41870522e-01 3.46588641e-01 -9.28043127e-02 1.09712407e-02
-5.67374945e-01 2.44138032e-01 3.71779650e-01 -5.04747927e-01
8.06892097e-01 -3.56899351e-02 3.05883139e-01 6.24372542e-01
4.59769845e-01 6.83449358e-02 -1.69571936e+00 3.78132254e-01
1.52820960e-01 -5.03269255e-01 -5.31534955e-04 -7.52957165e-01
-7.01316953e-01 3.00849557e-01 8.26290488e-01 2.37767249e-01
1.16336620e+00 -2.14536116e-01 3.04181397e-01 2.48132780e-01
1.09042621e+00 -1.22716331e+00 -2.89334152e-02 3.05052191e-01
9.38822567e-01 -8.52378666e-01 5.13578430e-02 -4.37869161e-01
-5.73340595e-01 5.38257122e-01 1.08964944e+00 4.37828153e-02
5.65359294e-01 7.71680474e-01 -1.20887809e-01 -4.55847941e-02
-1.12344146e+00 -7.46953674e-03 -3.85254174e-01 6.27719104e-01
-5.39618134e-01 -5.11587411e-02 -1.16177760e-01 4.38419759e-01
-1.73552316e-02 -1.94111675e-01 8.47281158e-01 1.17105436e+00
-4.99934733e-01 -8.53439867e-01 -7.24981070e-01 3.95983070e-01
2.54855722e-01 2.94683307e-01 -1.92257427e-02 1.01731968e+00
-2.99584176e-02 1.23622048e+00 -1.64689302e-01 -6.94512308e-01
6.53057575e-01 -9.82013196e-02 -9.64813158e-02 -6.12966239e-01
-1.42968178e-01 -1.37967721e-01 5.13945639e-01 -9.88918066e-01
1.16916880e-01 -8.50490034e-01 -1.16113913e+00 1.05291024e-01
-6.81601405e-01 1.76548094e-01 9.31173086e-01 7.44819701e-01
3.64514321e-01 8.97992015e-01 8.26138139e-01 -9.10898507e-01
-1.12566495e+00 -8.53324115e-01 -4.51103151e-01 -3.09175342e-01
3.20876628e-01 -1.14371049e+00 -7.66366065e-01 -2.98645526e-01] | [4.937016010284424, 2.0350987911224365] |
9e8a6eff-d150-45b2-beff-6e2526e2dc4b | duosearch-a-novel-search-engine-for-bulgarian | 2305.19392 | null | https://arxiv.org/abs/2305.19392v1 | https://arxiv.org/pdf/2305.19392v1.pdf | DuoSearch: A Novel Search Engine for Bulgarian Historical Documents | Search in collections of digitised historical documents is hindered by a two-prong problem, orthographic variety and optical character recognition (OCR) mistakes. We present a new search engine for historical documents, DuoSearch, which uses ElasticSearch and machine learning methods based on deep neural networks to offer a solution to this problem. It was tested on a collection of historical newspapers in Bulgarian from the mid-19th to the mid-20th century. The system provides an interactive and intuitive interface for the end-users allowing them to enter search terms in modern Bulgarian and search across historical spellings. This is the first solution facilitating the use of digitised historical documents in Bulgarian. | ['Milena Dobreva', 'Ivan Koychev', 'Suzan Hadzhieva', 'Angel Beshirov'] | 2023-05-30 | null | null | null | null | ['optical-character-recognition'] | ['computer-vision'] | [-4.01851982e-01 -6.06033325e-01 -1.15306489e-01 3.95040523e-04
-5.53312778e-01 -1.06736553e+00 1.10241532e+00 3.45587224e-01
-1.05964112e+00 8.49753678e-01 1.54588953e-01 -7.52318323e-01
-6.39149368e-01 -9.44149911e-01 -3.91118109e-01 -2.49054059e-01
-1.78673655e-01 8.27935278e-01 6.61933646e-02 -6.56449437e-01
8.98840129e-01 1.04862857e+00 -1.21786356e+00 7.14302212e-02
5.87399244e-01 7.03240275e-01 7.47408211e-01 9.99714255e-01
-3.04081678e-01 1.07597351e-01 -1.00339067e+00 -6.17097080e-01
3.22401494e-01 3.75303775e-01 -1.07307947e+00 -6.51984155e-01
6.89516366e-01 -5.86720109e-01 -8.42900276e-01 9.09769058e-01
5.95011413e-01 3.22491586e-01 5.89883387e-01 -3.95209104e-01
-1.11501634e+00 6.08554840e-01 2.09932283e-01 7.96091974e-01
2.62027413e-01 -2.08603889e-01 1.02830064e+00 -8.05700541e-01
8.56805444e-01 1.20482981e+00 7.11805046e-01 1.42790169e-01
-3.69808495e-01 -3.86232764e-01 -4.52749640e-01 5.96342206e-01
-1.42439711e+00 -2.26749033e-02 4.96555299e-01 -5.09306550e-01
1.30171299e+00 4.82954830e-01 8.48186910e-01 7.74065852e-01
2.67303675e-01 7.41488814e-01 4.06016320e-01 -8.95326078e-01
1.49844289e-01 1.29421413e-01 2.68410236e-01 3.63200396e-01
2.60899395e-01 1.49125963e-01 -1.11886650e-01 -3.60588133e-01
7.65083671e-01 2.30022427e-02 -1.84402123e-01 7.02041209e-01
-8.05373967e-01 8.91286194e-01 1.41427830e-01 8.62850964e-01
-2.38624543e-01 -1.37251958e-01 6.64807916e-01 3.63001078e-01
1.74180567e-01 8.64905655e-01 -4.52845871e-01 -4.54294890e-01
-1.47281003e+00 5.73423564e-01 8.96063983e-01 9.59588528e-01
5.67711517e-03 -5.02739400e-02 3.64402801e-01 1.02126586e+00
2.28255570e-01 4.23771203e-01 9.56297994e-01 -4.85724211e-01
6.21357679e-01 1.87508896e-01 2.64236748e-01 -1.07652330e+00
-2.26465806e-01 -7.50896633e-02 -4.48727876e-01 2.14220896e-01
5.19674242e-01 1.12776399e-01 -1.28476632e+00 8.39862645e-01
-1.39475167e-01 -7.89326072e-01 1.09021384e-02 6.94480956e-01
9.07251060e-01 1.05684245e+00 -2.59683818e-01 -9.12715960e-03
1.30679989e+00 -7.48529136e-01 -6.81617558e-01 4.85072769e-02
3.37170772e-02 -1.13169980e+00 8.40034366e-01 7.97366858e-01
-1.22573888e+00 -5.34432948e-01 -1.18326116e+00 -3.26865971e-01
-1.06069505e+00 3.08551073e-01 2.03566104e-01 6.15188956e-01
-9.64936018e-01 9.65537846e-01 -6.75220490e-01 -6.77860081e-01
-1.51689261e-01 3.73975158e-01 -1.47927940e-01 3.57381433e-01
-1.40927291e+00 1.06052947e+00 1.00914133e+00 5.16101979e-02
-4.65548962e-01 -3.17132562e-01 -4.73710001e-01 2.37843409e-01
1.19179964e-01 1.44663200e-01 1.34128678e+00 -5.46808720e-01
-1.23602068e+00 9.55704451e-01 5.04474878e-01 -1.58995152e-01
8.02639365e-01 -5.55867910e-01 -9.40480530e-01 1.90515757e-01
-1.33148685e-01 -3.09752282e-02 3.06276292e-01 -6.29825413e-01
-7.73451924e-01 -3.44197839e-01 -3.85925919e-01 4.09657992e-02
-3.87389123e-01 3.30732375e-01 -9.57503974e-01 -1.14935887e+00
3.61602311e-03 -6.80151403e-01 1.66381627e-01 -5.02494514e-01
-9.24561918e-02 -4.24191594e-01 7.47379303e-01 -1.30879390e+00
1.66381443e+00 -1.77998030e+00 -1.53077677e-01 3.81126761e-01
-3.65265757e-01 7.59496391e-01 2.48047471e-01 9.95061636e-01
8.67233649e-02 1.89893544e-01 1.50461704e-01 1.76795542e-01
-1.51572168e-01 2.24900231e-01 -2.84327745e-01 2.89036095e-01
-4.58967239e-01 4.37353343e-01 -8.86657178e-01 -6.05267346e-01
9.71657485e-02 1.65454507e-01 1.04181834e-01 -1.07293770e-01
-7.60847107e-02 -5.66513181e-01 -3.75617802e-01 9.52453613e-01
5.66579580e-01 2.55997598e-01 1.51838005e-01 4.14539903e-01
-5.00545919e-01 3.34246755e-01 -1.00500119e+00 1.59578443e+00
-2.79112875e-01 1.29210591e+00 -2.45593116e-01 -5.45126617e-01
9.95992005e-01 2.86286861e-01 -3.43098305e-02 -8.34821224e-01
1.74290612e-01 6.08693779e-01 -4.98929858e-01 -8.58746052e-01
1.51666784e+00 2.27153435e-01 2.62006223e-02 3.43510330e-01
1.18548803e-01 1.90321915e-02 5.85319281e-01 1.34769395e-01
7.19910860e-01 -2.90564839e-02 1.75298274e-01 -4.16700304e-01
4.52185899e-01 1.88509434e-01 4.88722660e-02 8.97031844e-01
3.19214791e-01 6.62601054e-01 -2.57123977e-01 -9.86100495e-01
-1.64784479e+00 -7.34021544e-01 -2.90382892e-01 1.03157496e+00
-2.14401424e-01 -2.58017868e-01 -5.44775307e-01 -1.79152504e-01
-1.01285748e-01 8.35273623e-01 -3.70575935e-01 3.02858323e-01
-9.94965434e-01 -3.69452953e-01 8.11452150e-01 2.79930413e-01
5.17267525e-01 -1.58970594e+00 -9.49320376e-01 6.06424093e-01
6.29866645e-02 -6.08143985e-01 -2.27949396e-01 2.52101094e-01
-7.84369946e-01 -6.70238674e-01 -1.42419875e+00 -9.23957467e-01
1.44854516e-01 -3.75661314e-01 8.91995370e-01 2.23909438e-01
-4.37151372e-01 1.03961378e-01 -5.19926012e-01 -4.05793697e-01
-5.59983552e-01 5.83359122e-01 -1.15831077e-01 -8.57497692e-01
5.68966925e-01 -2.48014092e-01 -5.46140671e-01 -2.35038787e-01
-1.07190216e+00 -6.06956303e-01 2.53941685e-01 5.79591095e-01
-1.01576239e-01 -2.06653643e-02 1.69824094e-01 -3.53524476e-01
9.52098906e-01 -5.05662978e-01 -1.20670426e+00 7.35125721e-01
-7.78675377e-01 -6.19333088e-02 6.36918664e-01 -4.32185620e-01
-8.85767221e-01 -3.83524626e-01 -4.75473315e-01 5.44997752e-02
-1.07389636e-01 7.05186725e-01 3.26911300e-01 7.91347027e-03
7.78068364e-01 7.17822552e-01 -6.87386215e-01 -1.09947491e+00
1.93213746e-01 1.26463413e+00 1.30325007e+00 -4.22246724e-01
3.86826605e-01 4.84421067e-02 -4.95787024e-01 -1.18348265e+00
1.92311883e-01 -6.44294918e-01 -6.09124601e-01 -4.13678050e-01
6.01391137e-01 -4.38445538e-01 -4.76230949e-01 7.24618733e-01
-1.33913743e+00 9.20234025e-02 2.04566047e-01 4.96604562e-01
-8.71710405e-02 7.57269621e-01 -8.29894006e-01 -9.75359023e-01
-9.49971437e-01 -7.19642818e-01 6.51674032e-01 5.00884354e-01
-1.10232659e-01 -8.78289878e-01 5.16765594e-01 -2.64634937e-01
1.85602322e-01 -4.81515825e-02 1.07616830e+00 -1.04031348e+00
-8.12871084e-02 -5.85307598e-01 -1.77252553e-02 1.78813890e-01
-1.20903067e-01 4.56983656e-01 -5.09806633e-01 -1.61843717e-01
-2.57005066e-01 1.65698379e-01 7.15358078e-01 1.70618102e-01
6.13316119e-01 -4.81824905e-01 -3.21920961e-01 4.34877038e-01
1.94670308e+00 1.14690030e+00 8.50168526e-01 1.40821552e+00
1.90851361e-01 1.62242234e-01 2.08065480e-01 4.60616440e-01
2.45505758e-02 3.23613942e-01 -1.05362907e-02 3.46596748e-01
3.24813336e-01 8.83264095e-03 -9.25728008e-02 5.96276402e-01
-3.04833710e-01 -7.26337910e-01 -1.47197711e+00 9.94072199e-01
-1.39705241e+00 -9.32626724e-01 7.90657550e-02 2.16128612e+00
8.95751119e-01 2.75436699e-01 8.90549123e-02 2.49018043e-01
8.40654552e-01 -2.34992310e-01 -2.03926191e-02 -7.90292621e-01
-1.23189576e-01 4.07409728e-01 7.27183223e-01 5.04699767e-01
-9.70457971e-01 1.05823565e+00 7.04214907e+00 9.27294075e-01
-1.25563884e+00 -3.40531439e-01 2.59828985e-01 -1.56245068e-01
-6.17236458e-02 -4.00158554e-01 -9.93819177e-01 7.85924613e-01
8.72423053e-01 -1.78503647e-01 2.54703462e-01 8.93216789e-01
-1.86349154e-01 -2.10046172e-01 -6.96129560e-01 8.72164488e-01
1.27495229e-01 -1.87905943e+00 1.95328653e-01 -5.36281727e-02
4.25709158e-01 2.82322079e-01 -1.30682573e-01 1.38784051e-01
3.21341008e-01 -8.87125015e-01 1.00186038e+00 5.56814671e-01
6.53662622e-01 -1.09964406e+00 5.64106107e-01 1.40435845e-01
-5.98019958e-01 -2.33552188e-01 -4.71365899e-01 1.27002999e-01
-1.16414921e-02 1.00618146e-01 -8.50401044e-01 3.38448197e-01
9.37082529e-01 -2.15353340e-01 -4.71447170e-01 1.60755515e+00
2.15279996e-01 2.64422327e-01 -6.30737782e-01 -5.36437154e-01
9.64956701e-01 -1.29252121e-01 5.37872910e-01 1.69487870e+00
7.99267352e-01 2.03488871e-01 -5.52704453e-01 2.29893222e-01
1.14893258e-01 3.09032410e-01 -3.55959088e-01 -3.13373625e-01
5.40616810e-01 7.54566908e-01 -9.55219507e-01 -4.49426770e-01
-1.37189686e-01 8.75888348e-01 6.72166944e-02 3.07556063e-01
-2.08236933e-01 -1.20612741e+00 -7.11259693e-02 8.09032694e-02
4.10754055e-01 -3.59530777e-01 -8.67723450e-02 -7.02109516e-01
1.78078234e-01 -8.82899165e-01 6.68572128e-01 -7.73784399e-01
-7.35901356e-01 1.01334262e+00 2.14094937e-01 -5.71206391e-01
-5.64750910e-01 -9.29193079e-01 -4.46446478e-01 7.94601738e-01
-8.29818189e-01 -8.05346549e-01 3.44639063e-01 2.37629563e-01
9.81819630e-01 -8.76587868e-01 7.36827433e-01 5.54588318e-01
6.55303430e-03 3.85266870e-01 1.19619930e+00 5.41046619e-01
3.54290485e-01 -1.29912210e+00 8.34806561e-01 6.89451218e-01
2.67787665e-01 5.94334066e-01 7.31474400e-01 -7.95307755e-01
-8.88317883e-01 -2.37199381e-01 1.57218015e+00 -8.41274261e-02
6.69997871e-01 -1.12105325e-01 -8.76828849e-01 4.22021091e-01
5.19492567e-01 -1.09588504e+00 4.45268631e-01 -2.31528878e-01
-5.24364337e-02 1.75671443e-01 -8.27980220e-01 6.32532001e-01
3.86753947e-01 -4.68771368e-01 -1.06741631e+00 4.85278279e-01
3.16303000e-02 -3.42356473e-01 -6.16055846e-01 -1.94559351e-01
1.00998294e+00 -4.81590122e-01 9.58365381e-01 -4.76923048e-01
3.41251493e-01 2.53651679e-01 4.02210131e-02 -1.05204403e+00
-2.71035165e-01 -1.02192760e+00 3.44196439e-01 9.49809253e-01
4.89282608e-01 -2.03389972e-01 5.36734402e-01 5.31312287e-01
-3.85078439e-03 -1.39413610e-01 -1.04415715e+00 -9.28537190e-01
3.48752379e-01 -4.60937202e-01 6.88368976e-01 8.94651651e-01
6.42498815e-03 -2.91078627e-01 -2.97382444e-01 -9.73675027e-02
6.31289706e-02 -3.08998674e-01 9.85162929e-02 -1.35782862e+00
-5.30734435e-02 -8.05227697e-01 -2.61331946e-01 -7.41111457e-01
-2.98571736e-01 -6.00980461e-01 -4.22527343e-02 -1.44630837e+00
-2.98679858e-01 -2.22296655e-01 -4.10015024e-02 1.22196324e-01
3.70748311e-01 2.69483477e-02 2.25612819e-01 5.21439314e-01
3.42737585e-02 -1.14604197e-01 7.01544523e-01 -2.41667300e-01
-1.62630603e-01 9.42359194e-02 -1.52295515e-01 5.49649656e-01
6.59871638e-01 -4.82983857e-01 -7.21327448e-03 -7.45634317e-01
7.40167916e-01 7.79127702e-02 7.26300816e-04 -1.11380684e+00
5.84421933e-01 -6.37818426e-02 6.60821497e-01 -1.24572074e+00
3.79154384e-02 -8.07814300e-01 2.53575414e-01 4.20912921e-01
-3.00069243e-01 7.38113940e-01 4.88226026e-01 2.26946875e-01
-1.94482401e-01 -1.09538627e+00 4.61789906e-01 -6.25512242e-01
-1.02394223e+00 -1.06641822e-01 -8.35509837e-01 -2.97508508e-01
6.03640258e-01 -4.71853703e-01 -1.91338465e-01 -2.29109526e-01
-7.30908215e-01 -1.63567346e-02 2.21862480e-01 4.65836555e-01
4.46225315e-01 -1.15564716e+00 -6.48678720e-01 1.12243108e-01
-1.27259180e-01 -3.65808278e-01 -2.26973608e-01 -2.94305056e-01
-1.75652373e+00 9.29904640e-01 -6.87157989e-01 9.01045054e-02
-1.48530698e+00 5.16578615e-01 1.23692855e-01 -3.12862158e-01
-6.55969203e-01 9.63983357e-01 -7.80042768e-01 -2.88795121e-02
7.14310944e-01 -1.12808138e-01 -5.68242192e-01 4.07656312e-01
6.80172503e-01 5.33905745e-01 4.20192808e-01 -4.46616977e-01
-2.55993843e-01 5.23491800e-01 -5.14155984e-01 -5.93712211e-01
1.42233503e+00 -1.03792302e-01 8.74027014e-02 1.84917301e-01
1.30773592e+00 -2.17512399e-01 -4.37281281e-01 -1.41830400e-01
3.20697874e-01 -4.72997040e-01 2.86329389e-01 -1.06480026e+00
-7.28175342e-01 6.13521159e-01 8.46334934e-01 5.04067898e-01
9.93451715e-01 -2.97545671e-01 8.99551809e-01 1.22433782e+00
7.62695819e-02 -1.82830703e+00 -5.60323298e-01 7.67820656e-01
9.36388373e-01 -8.40796113e-01 3.61706406e-01 6.61754549e-01
-2.62176514e-01 1.97364902e+00 -7.24466667e-02 5.87022724e-03
5.99905908e-01 8.56798515e-03 1.73158973e-01 -2.08062395e-01
-7.20058754e-02 1.94727525e-01 3.35270643e-01 1.84939191e-01
2.76009470e-01 -2.71897405e-01 -9.51810360e-01 2.04444408e-01
-6.10503614e-01 1.73524749e-02 3.75374109e-01 1.29517341e+00
-4.85544115e-01 -1.03332543e+00 -7.06565261e-01 1.43520802e-01
-1.11713326e+00 -5.07989705e-01 -5.23928106e-01 1.26421928e+00
-9.86104533e-02 4.56855208e-01 3.30941677e-01 -1.79587841e-01
-5.66054694e-02 1.56617180e-01 2.61008948e-01 6.80894181e-02
-1.04594004e+00 3.74313474e-01 1.54200613e-01 1.59122199e-01
1.74854577e-01 -6.48626208e-01 -8.02315354e-01 -4.99253213e-01
-3.75622869e-01 5.11211097e-01 1.09559405e+00 8.14780533e-01
-1.60673171e-01 -1.56639636e-01 6.39222339e-02 -7.57029653e-01
-4.13752139e-01 -1.17829072e+00 -5.98629296e-01 -1.29656009e-02
4.82808650e-01 -1.87544636e-02 -8.57296884e-02 2.24835694e-01] | [10.275880813598633, 10.282879829406738] |
4e0a5dc0-033f-401b-8dd4-e62829f4dd26 | keep-calm-and-explore-language-models-for | 2010.02903 | null | https://arxiv.org/abs/2010.02903v1 | https://arxiv.org/pdf/2010.02903v1.pdf | Keep CALM and Explore: Language Models for Action Generation in Text-based Games | Text-based games present a unique challenge for autonomous agents to operate in natural language and handle enormous action spaces. In this paper, we propose the Contextual Action Language Model (CALM) to generate a compact set of action candidates at each game state. Our key insight is to train language models on human gameplay, where people demonstrate linguistic priors and a general game sense for promising actions conditioned on game history. We combine CALM with a reinforcement learning agent which re-ranks the generated action candidates to maximize in-game rewards. We evaluate our approach using the Jericho benchmark, on games unseen by CALM during training. Our method obtains a 69% relative improvement in average game score over the previous state-of-the-art model. Surprisingly, on half of these games, CALM is competitive with or better than other models that have access to ground truth admissible actions. Code and data are available at https://github.com/princeton-nlp/calm-textgame. | ['Karthik Narasimhan', 'Matthew Hausknecht', 'Rohan Rao', 'Shunyu Yao'] | 2020-10-06 | null | https://aclanthology.org/2020.emnlp-main.704 | https://aclanthology.org/2020.emnlp-main.704.pdf | emnlp-2020-11 | ['action-generation', 'text-based-games'] | ['computer-vision', 'playing-games'] | [-1.37609601e-01 3.94152790e-01 -1.56853884e-01 4.49307896e-02
-1.10716486e+00 -7.25033343e-01 9.79215324e-01 -8.62614363e-02
-8.09801519e-01 9.03277397e-01 4.52513933e-01 -3.22366059e-01
-5.52413530e-06 -9.23687160e-01 -4.81701076e-01 -1.73796028e-01
-1.17098488e-01 9.52694356e-01 3.66828710e-01 -1.00179946e+00
3.36384892e-01 -2.83933163e-01 -1.34446216e+00 3.29506665e-01
7.14777410e-01 3.79319042e-01 2.29351282e-01 1.09648085e+00
2.13438675e-01 1.36784017e+00 -4.83222365e-01 -5.73120296e-01
7.36968875e-01 -7.64020741e-01 -1.05067885e+00 2.23406907e-02
-1.50778115e-01 -5.19431770e-01 -6.53897762e-01 1.15067995e+00
4.45415676e-01 5.66832602e-01 4.32105124e-01 -1.20555198e+00
-5.28244138e-01 9.50062513e-01 -2.82526225e-01 8.76476169e-02
6.23583317e-01 8.21757019e-01 1.43816578e+00 -4.13305759e-01
6.65969968e-01 1.28350806e+00 1.59757972e-01 1.04055977e+00
-9.81342673e-01 -3.28068107e-01 3.16270620e-01 6.45313226e-03
-1.08447278e+00 -2.63511300e-01 2.72789896e-01 -3.88144940e-01
1.22562432e+00 -3.95967513e-02 8.37248802e-01 1.06907952e+00
1.98326290e-01 1.13097036e+00 1.00233340e+00 -4.13244963e-01
4.88418907e-01 -6.35624588e-01 -3.28479648e-01 9.61630046e-01
-1.76083073e-01 6.31562173e-01 -6.30445480e-01 -1.04868494e-01
9.25606132e-01 -3.21331888e-01 2.04866722e-01 -3.56899887e-01
-9.79298592e-01 1.18372309e+00 1.76413551e-01 -2.41856035e-02
-6.08329594e-01 7.06408441e-01 2.69230604e-01 3.80258054e-01
2.94373959e-01 9.88737524e-01 -2.82845289e-01 -9.13325250e-01
-5.04099846e-01 8.57678950e-01 7.38581359e-01 6.60473347e-01
2.82676101e-01 7.65993819e-02 -3.12975675e-01 6.18940771e-01
3.09248149e-01 3.36825460e-01 5.85927546e-01 -1.52367043e+00
5.75178862e-01 5.24372339e-01 4.38738734e-01 -4.54935551e-01
-4.73696738e-01 -1.63823038e-01 -2.46020809e-01 5.74609399e-01
5.51475227e-01 -4.45692122e-01 -7.67007768e-01 2.05306363e+00
-5.24155051e-02 3.45029563e-01 4.10776854e-01 8.46416414e-01
4.69393045e-01 6.03001356e-01 2.90513247e-01 1.68200746e-01
1.01703894e+00 -1.12433779e+00 -2.00698078e-01 -9.21693981e-01
7.77005851e-01 -2.40462556e-01 1.44880593e+00 6.91151619e-01
-1.31240356e+00 -1.92791209e-01 -8.13473880e-01 1.67420670e-01
-4.08205204e-02 -1.52573332e-01 1.09463334e+00 5.11311114e-01
-1.32023871e+00 5.77622414e-01 -1.09359860e+00 -2.82451421e-01
3.51703197e-01 4.15378153e-01 -1.31579101e-01 5.11841141e-02
-1.37091327e+00 9.02015448e-01 8.73230338e-01 -3.98818076e-01
-1.50006580e+00 -1.68743208e-01 -1.02155387e+00 -1.41942918e-01
9.72892165e-01 -5.80349445e-01 2.09078836e+00 -7.01018989e-01
-2.14566183e+00 8.65577936e-01 4.96962994e-01 -9.42354858e-01
7.17921793e-01 -3.01778734e-01 2.15141058e-01 -1.20787203e-01
2.65648633e-01 9.98558164e-01 2.61142910e-01 -7.31278360e-01
-9.05991137e-01 1.61070481e-01 9.42654371e-01 7.56174028e-01
2.08255127e-01 -6.39683157e-02 -6.42316937e-01 -3.92562240e-01
-4.34421599e-01 -1.03310478e+00 -9.25809085e-01 -5.33527315e-01
-2.12062865e-01 -4.23098296e-01 -2.95060933e-01 -3.27556282e-01
1.07351100e+00 -1.74674129e+00 3.42343748e-01 -1.27760068e-01
2.85223782e-01 1.97073191e-01 -4.96384233e-01 5.83059311e-01
2.90525258e-01 1.68140847e-02 -8.76534134e-02 -3.39750499e-01
4.87384051e-01 4.31089282e-01 -1.65557966e-01 1.59189239e-01
-5.13108037e-02 1.36110389e+00 -1.37857413e+00 -1.15481719e-01
4.41762060e-01 -2.46314570e-01 -1.22687995e+00 8.01019967e-02
-8.28979135e-01 4.07341063e-01 -7.17053473e-01 1.70472786e-01
-6.01424724e-02 -2.06043527e-01 3.72982442e-01 9.71419871e-01
1.34311616e-01 5.91620743e-01 -1.11610878e+00 2.07766747e+00
-3.78503144e-01 2.72465467e-01 -3.00259203e-01 -7.04559863e-01
5.20217240e-01 1.72880873e-01 3.53244424e-01 -8.19216073e-01
3.53383392e-01 -1.52484905e-02 2.22087383e-01 4.09583002e-03
8.54316354e-01 -1.33039221e-01 -8.23500395e-01 5.40126622e-01
2.21189514e-01 -5.06305456e-01 6.67404473e-01 4.85869348e-01
1.28813362e+00 4.81379181e-01 6.84566021e-01 -1.35855183e-01
2.85060197e-01 3.19637090e-01 4.51808065e-01 1.42295325e+00
-3.44227254e-01 2.06171051e-01 8.85292172e-01 -4.79873955e-01
-8.54368925e-01 -1.01069748e+00 8.44529510e-01 1.60328698e+00
1.31111905e-01 -9.12453890e-01 -7.18359590e-01 -5.17029226e-01
-3.75626802e-01 9.38479125e-01 -5.59233963e-01 -3.44688237e-01
-4.87887442e-01 -5.66014290e-01 6.63232625e-01 4.67426717e-01
5.04783273e-01 -1.56976569e+00 -8.24567795e-01 4.57852185e-01
-2.05084145e-01 -9.34043169e-01 -5.31963706e-01 7.91253746e-02
-3.46339852e-01 -1.08298075e+00 -1.32323444e-01 -5.10083795e-01
1.62898019e-01 -2.03386366e-01 1.58257997e+00 1.62274346e-01
-9.58212540e-02 3.80999476e-01 -5.11076093e-01 -4.19772089e-01
-6.96411133e-01 1.70164794e-01 1.62824467e-01 -6.17907763e-01
3.78097087e-01 -4.56688344e-01 -3.72635782e-01 -2.50943266e-02
-5.96619487e-01 4.95232075e-01 3.77179444e-01 8.45870554e-01
3.42847437e-01 6.41013235e-02 3.07690591e-01 -8.48429382e-01
1.07405508e+00 -3.31905544e-01 -1.01337051e+00 -1.69571772e-01
-2.38945529e-01 3.24275106e-01 6.18597507e-01 -2.51117766e-01
-8.37487400e-01 1.57082543e-01 -2.24796072e-01 6.48882613e-02
-9.02203694e-02 4.92148101e-01 -1.32027352e-02 3.20657313e-01
1.02247870e+00 2.38383383e-01 -1.68072075e-01 2.76695620e-02
8.00984085e-01 1.91239908e-01 6.80819452e-01 -1.14998424e+00
6.47428393e-01 1.15924843e-01 -2.59016216e-01 -2.02172846e-01
-8.38487446e-01 -3.68015021e-01 -2.52303779e-01 -2.69789636e-01
8.46129775e-01 -1.09397471e+00 -1.08743584e+00 5.33080280e-01
-7.18687415e-01 -1.27811086e+00 -6.22095942e-01 4.16577011e-01
-1.33627486e+00 4.79230918e-02 -8.56991529e-01 -8.47709417e-01
-6.08428195e-02 -1.25462472e+00 9.46326256e-01 4.31679666e-01
-2.12103322e-01 -8.64814699e-01 3.35296482e-01 4.56862479e-01
6.32177666e-02 9.47394297e-02 2.26583004e-01 -6.86846852e-01
-5.12345552e-01 -2.70163640e-02 4.12641704e-01 -4.93656918e-02
4.49247919e-02 -5.70473671e-01 -4.99599785e-01 -2.76603132e-01
-4.29892600e-01 -9.45728958e-01 8.05401266e-01 4.70714152e-01
7.62128830e-01 -2.29865655e-01 1.08433373e-01 3.24423641e-01
1.17929649e+00 3.11645985e-01 7.20833778e-01 6.20561898e-01
2.40527883e-01 1.56628564e-01 8.98823500e-01 8.19949150e-01
4.36046034e-01 6.50076687e-01 6.49776280e-01 4.43249911e-01
3.22979838e-01 -7.10029721e-01 6.96134865e-01 1.35693625e-01
-2.60651886e-01 -6.33017659e-01 -1.04114473e+00 5.72011471e-01
-2.29738808e+00 -1.17249215e+00 4.61178154e-01 1.99472344e+00
9.75999177e-01 7.99314559e-01 5.01352966e-01 -3.91010761e-01
3.82276356e-01 2.99247265e-01 -6.08270466e-01 -3.02733749e-01
1.50622979e-01 6.42897666e-01 6.01141751e-01 1.18416524e+00
-1.22824395e+00 1.88160527e+00 6.06005764e+00 1.15448308e+00
-4.73661840e-01 7.47017041e-02 5.04165113e-01 -4.29334968e-01
2.67143119e-02 5.38624339e-02 -6.71552181e-01 1.04185894e-01
9.42453086e-01 -5.43671668e-01 9.81483161e-01 9.87371325e-01
3.75150412e-01 -1.62591726e-01 -8.23411942e-01 7.48072445e-01
-2.39649966e-01 -1.38463664e+00 -3.02397851e-02 2.47742295e-01
8.15250158e-01 2.80407548e-01 1.30013689e-01 1.00408268e+00
1.84780812e+00 -1.22993326e+00 7.49504030e-01 2.48019159e-01
5.66715479e-01 -9.32969630e-01 4.86088842e-01 6.92045331e-01
-1.09194410e+00 -1.75791264e-01 -4.02239889e-01 -8.69992733e-01
1.74404591e-01 -4.35621351e-01 -8.71689737e-01 4.40054834e-01
1.01193689e-01 6.90050125e-01 -4.73753393e-01 7.32778549e-01
-8.05201411e-01 5.99525213e-01 -2.52417475e-01 -2.50084251e-01
9.21744585e-01 -2.36543626e-01 4.74796474e-01 6.39738500e-01
-4.07314897e-02 5.57426333e-01 8.79320323e-01 7.67604709e-01
-1.44682169e-01 -4.57448624e-02 -6.45170987e-01 -3.53523523e-01
3.02436560e-01 1.01014948e+00 -8.19514215e-01 -3.18487793e-01
-2.48181969e-01 1.07269490e+00 4.49705750e-01 2.25307215e-02
-8.90064716e-01 1.10425226e-01 9.50236797e-01 -1.71562359e-01
-5.53772375e-02 -3.52922320e-01 1.33415967e-01 -1.20460165e+00
-4.01398063e-01 -1.18421710e+00 5.87269425e-01 -7.98325598e-01
-1.07901740e+00 4.81345683e-01 -6.50127307e-02 -1.10409808e+00
-9.90014255e-01 -7.85067379e-01 -6.47182584e-01 6.08334780e-01
-8.41868579e-01 -9.42255437e-01 1.72137976e-01 5.08051455e-01
9.79743719e-01 -5.19286692e-01 7.44471014e-01 -2.52113044e-01
-1.60667866e-01 3.64460766e-01 -2.56607622e-01 3.44093472e-01
3.74331206e-01 -1.62943220e+00 1.13392889e+00 1.04798090e+00
4.58658099e-01 2.22942859e-01 7.84421325e-01 -6.39310241e-01
-9.81393099e-01 -9.68588591e-01 3.13139766e-01 -9.29993570e-01
8.35931182e-01 -4.06771213e-01 -1.41529575e-01 1.02501178e+00
2.22140893e-01 -2.93640703e-01 3.74163866e-01 8.13764706e-02
3.55931669e-02 6.70673311e-01 -7.45103180e-01 1.17168534e+00
1.38362288e+00 -3.27946633e-01 -8.28845263e-01 5.06896496e-01
7.30481625e-01 -9.38548088e-01 -2.13820130e-01 -2.39938334e-01
9.65760276e-02 -9.86158490e-01 9.33733523e-01 -1.07283032e+00
6.96151733e-01 -2.85296381e-01 -2.38654271e-01 -1.61606848e+00
-4.07643139e-01 -1.10654318e+00 1.96086034e-01 4.59559172e-01
4.54221010e-01 -4.00504768e-01 1.15198994e+00 5.77700734e-01
-1.59375295e-01 -7.01410115e-01 -7.17834353e-01 -8.50109160e-01
4.21135247e-01 -8.17808568e-01 4.26772147e-01 4.83477443e-01
4.04681295e-01 3.72702748e-01 -6.43174410e-01 2.94837188e-02
4.03046757e-01 -1.63760990e-01 8.66501272e-01 -9.67694640e-01
-9.08869684e-01 -7.54878163e-01 -3.81280243e-01 -1.38000607e+00
4.13910210e-01 -8.95805120e-01 3.11893284e-01 -1.70068908e+00
1.55483648e-01 -6.50387481e-02 -6.65751845e-02 8.31987977e-01
-1.97663575e-01 2.76850432e-01 5.31580508e-01 -2.09376439e-01
-1.38422048e+00 6.34077072e-01 1.18916821e+00 -8.42337161e-02
-5.11714280e-01 6.84482604e-03 -1.16581047e+00 1.05903995e+00
1.16216266e+00 -2.76618630e-01 -7.12961435e-01 -3.61312658e-01
5.29103577e-01 3.74828309e-01 2.71371216e-01 -1.06136250e+00
3.00086439e-01 -7.72858679e-01 -2.09585592e-01 6.40605276e-06
5.45963705e-01 1.01778591e-02 -3.06652516e-01 7.63085663e-01
-6.24573946e-01 1.21389687e-01 2.92760581e-01 4.53418434e-01
4.72477861e-02 -2.33294770e-01 5.92122853e-01 -5.96653938e-01
-1.15436482e+00 3.81569535e-01 -8.50284517e-01 6.93941712e-01
9.60626900e-01 9.12312642e-02 -2.30816856e-01 -1.03850853e+00
-8.87105525e-01 6.40966296e-01 4.01087701e-01 5.19792736e-01
4.55028385e-01 -1.13722801e+00 -9.66491699e-01 -2.82556713e-01
1.14498839e-01 -1.11762971e-01 1.39028788e-01 8.35120529e-02
-6.54569149e-01 3.79063874e-01 -3.24719340e-01 6.49848282e-02
-9.18537378e-01 3.66011530e-01 7.52103925e-01 -9.36107874e-01
-6.01052582e-01 1.05107951e+00 3.29162449e-01 -7.53540158e-01
-9.19350162e-02 -3.04263055e-01 -6.63056448e-02 -5.91643572e-01
5.04643917e-01 -1.24932393e-01 -3.36493969e-01 -3.02464187e-01
-1.82667330e-01 1.62680186e-02 -5.67753203e-02 -7.26753592e-01
1.33353221e+00 2.41056830e-01 3.22432667e-01 1.28941685e-01
1.56220302e-01 -4.05690633e-02 -1.68754172e+00 -2.86958724e-01
1.41721489e-02 -3.71308476e-01 -1.24544717e-01 -9.35894847e-01
-7.24536359e-01 3.25606227e-01 1.42201334e-01 3.42294462e-02
8.00087333e-01 1.17425069e-01 4.16958958e-01 7.60541141e-01
9.49488103e-01 -1.10788035e+00 5.27936459e-01 1.19166851e+00
6.76747859e-01 -1.21431077e+00 -3.19554448e-01 1.49332494e-01
-1.16303360e+00 5.75735509e-01 9.63354766e-01 -5.60575485e-01
1.87900007e-01 2.63164461e-01 -5.97470850e-02 -3.39451492e-01
-1.18429804e+00 -7.15286314e-01 -2.28430077e-01 6.51722431e-01
1.97541252e-01 3.65248561e-01 -3.02680790e-01 8.43899608e-01
-8.82527053e-01 2.13633757e-02 9.16408777e-01 1.00517881e+00
-6.59469843e-01 -1.40541792e+00 -1.61725078e-02 3.22620809e-01
-4.62439328e-01 -3.56952697e-01 -5.12788773e-01 7.66698897e-01
-1.55873239e-01 1.11574900e+00 -2.83383764e-02 -4.94643420e-01
3.11967582e-01 -1.42409513e-02 6.45072639e-01 -1.06540143e+00
-5.00317395e-01 -2.50690132e-01 2.55821019e-01 -9.50708568e-01
-7.88953379e-02 -6.18450820e-01 -1.56848443e+00 -4.95817989e-01
2.56923825e-01 2.84194082e-01 -1.10970311e-01 1.03800607e+00
8.23695064e-02 5.23026764e-01 1.36085793e-01 -8.92193079e-01
-6.90902829e-01 -7.93764710e-01 -5.32773435e-01 3.62037361e-01
-4.79210503e-02 -7.53108621e-01 2.65456047e-02 -1.65212557e-01] | [3.740445375442505, 1.4127981662750244] |
9cd49219-56a1-4e99-b9db-359afd6081eb | efficient-blind-deblurring-under-high-noise | 1904.09154 | null | https://arxiv.org/abs/1904.09154v2 | https://arxiv.org/pdf/1904.09154v2.pdf | Efficient Blind Deblurring under High Noise Levels | The goal of blind image deblurring is to recover a sharp image from a motion blurred one without knowing the camera motion. Current state-of-the-art methods have a remarkably good performance on images with no noise or very low noise levels. However, the noiseless assumption is not realistic considering that low light conditions are the main reason for the presence of motion blur due to requiring longer exposure times. In fact, motion blur and high to moderate noise often appear together. Most works approach this problem by first estimating the blur kernel $k$ and then deconvolving the noisy blurred image. In this work, we first show that current state-of-the-art kernel estimation methods based on the $\ell_0$ gradient prior can be adapted to handle high noise levels while keeping their efficiency. Then, we show that a fast non-blind deconvolution method can be significantly improved by first denoising the blurry image. The proposed approach yields results that are equivalent to those obtained with much more computationally demanding methods. | ['Jérémy Anger', 'Gabriele Facciolo', 'Mauricio Delbracio'] | 2019-04-19 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 1.48214415e-01 -6.33144975e-01 5.07668555e-01 -2.67472472e-02
-4.89661187e-01 -4.95799839e-01 3.65494847e-01 -5.44154346e-01
-5.83615780e-01 8.92717600e-01 1.59495905e-01 -1.26951784e-01
-2.49255314e-01 -2.92924613e-01 -5.30818224e-01 -8.55664432e-01
2.71781415e-01 -8.36117119e-02 3.17491412e-01 2.07923427e-01
3.35911155e-01 2.90347964e-01 -1.51079476e+00 -1.36245966e-01
1.10994303e+00 7.65406847e-01 6.08241558e-01 9.12506759e-01
1.00954652e-01 9.72969651e-01 -4.85098004e-01 -1.80809572e-01
2.51694262e-01 -7.13706374e-01 -7.74801910e-01 3.39110643e-01
4.52847749e-01 -6.44976020e-01 -5.69614828e-01 1.53644466e+00
3.65476459e-01 3.00798625e-01 4.91540521e-01 -5.34042597e-01
-7.39966989e-01 1.55451179e-01 -7.16601551e-01 3.67450148e-01
3.01217884e-01 5.84785156e-02 2.84187078e-01 -9.08541620e-01
4.48183119e-01 9.03847992e-01 7.22984433e-01 3.88479322e-01
-1.33657920e+00 -1.51282996e-01 -2.61104971e-01 5.84069431e-01
-1.44309771e+00 -7.99551904e-01 7.76504636e-01 -5.12235105e-01
5.18138528e-01 4.05825824e-01 1.45487055e-01 6.71934187e-01
8.64590630e-02 3.34779769e-01 1.51350510e+00 -4.80280071e-01
2.92550474e-01 1.30356506e-01 2.10261434e-01 5.15402734e-01
5.59798658e-01 1.43827721e-02 -2.20828086e-01 -1.44739628e-01
8.03662479e-01 4.03089672e-02 -1.02774787e+00 -4.03394312e-01
-1.22825718e+00 4.22005564e-01 3.29419732e-01 6.31412029e-01
-5.37956119e-01 2.37812281e-01 -1.82595365e-02 1.78458422e-01
5.02261877e-01 3.04846317e-01 -1.22516684e-01 -2.71129817e-01
-1.45656073e+00 1.50865585e-01 7.73557067e-01 5.53929687e-01
8.18914950e-01 1.66314486e-02 -4.11324836e-02 8.95983219e-01
2.15188608e-01 4.81355876e-01 4.91697490e-01 -1.20124030e+00
-6.81947321e-02 -1.77256390e-01 8.08441937e-01 -7.62696326e-01
3.24706025e-02 -3.19171727e-01 -1.02516448e+00 4.85176712e-01
8.33842933e-01 -1.04292780e-01 -9.88373518e-01 1.49231982e+00
-6.31506443e-02 4.88782048e-01 1.69184014e-01 1.23994815e+00
3.60013902e-01 5.17807126e-01 -4.16331470e-01 -5.87469995e-01
1.46532881e+00 -9.42654550e-01 -1.19103980e+00 -4.55531210e-01
-3.85561809e-02 -1.27361667e+00 7.05607891e-01 4.53374833e-01
-1.27925158e+00 -5.79187870e-01 -9.56204236e-01 -4.02490944e-02
-1.06812073e-02 1.35131985e-01 2.88099647e-01 8.01402867e-01
-1.24315023e+00 7.74820864e-01 -7.16917336e-01 -4.42043334e-01
1.21355012e-01 1.25347555e-01 -2.73081452e-01 -6.35934055e-01
-9.50261414e-01 1.29204178e+00 -5.09251952e-02 4.05760437e-01
-7.37171531e-01 -5.06298184e-01 -7.02499747e-01 9.83103178e-03
2.82247633e-01 -7.93719590e-01 1.23666370e+00 -8.37882757e-01
-1.48759615e+00 6.12712204e-01 -5.64916134e-01 -3.82840276e-01
7.96555340e-01 -5.73393106e-01 -2.36805499e-01 3.79798353e-01
-2.20774740e-01 1.31422272e-02 1.38663256e+00 -1.48216462e+00
-1.50493622e-01 -7.30230883e-02 -1.22678965e-01 1.46027744e-01
1.35099724e-01 5.20868339e-02 -4.91667747e-01 -7.36277282e-01
7.49349743e-02 -7.37684965e-01 -1.48585930e-01 -4.66453806e-02
-1.13636656e-02 6.00796998e-01 8.80569875e-01 -9.26849961e-01
1.21696579e+00 -2.27364635e+00 1.67848393e-01 -3.39911491e-01
1.97983667e-01 5.62461257e-01 1.00392848e-01 2.87715346e-01
-1.97964281e-01 -3.17516923e-01 -6.07210875e-01 -3.73377591e-01
-2.36795247e-01 -1.08554831e-03 -1.89413995e-01 8.44419241e-01
-2.84296066e-01 5.39273143e-01 -8.82441163e-01 -1.59357339e-01
6.04929745e-01 6.99349046e-01 -2.67056763e-01 3.92051429e-01
2.97756881e-01 5.26140332e-01 1.83632702e-01 1.19780049e-01
1.16550720e+00 -2.18116492e-01 4.49129846e-04 -4.92009252e-01
-4.09027725e-01 -3.52620363e-01 -1.28883767e+00 1.61845469e+00
-3.96421522e-01 9.32929933e-01 7.80203342e-01 -8.95320058e-01
4.63878930e-01 4.21578377e-01 2.49571696e-01 -1.48396954e-01
9.48439166e-02 3.97152901e-01 -1.05924934e-01 -7.93689072e-01
5.18782973e-01 -4.64805812e-01 5.09116054e-01 1.47984877e-01
-2.23875269e-01 -3.54162514e-01 -3.27503169e-03 -3.69171239e-02
1.08991861e+00 1.02408476e-01 2.68574983e-01 -4.97670591e-01
6.85692608e-01 -3.04115742e-01 1.63515195e-01 9.03773546e-01
-3.25007051e-01 9.70337570e-01 -3.08895130e-02 -1.23113513e-01
-9.28668618e-01 -9.22889829e-01 -1.36813462e-01 3.90496373e-01
4.42326754e-01 1.91290807e-02 -1.31017280e+00 -1.04375541e-01
-2.71417141e-01 6.60116076e-01 -4.47415501e-01 6.65888786e-02
-4.57828492e-01 -1.11592340e+00 1.50923014e-01 3.76862027e-02
8.75639617e-01 -6.05452240e-01 -7.07068503e-01 1.46063685e-01
-4.51662958e-01 -1.27923787e+00 -6.86678231e-01 -1.35081619e-01
-7.23161757e-01 -1.02177036e+00 -1.39182436e+00 -7.16617405e-01
8.31294000e-01 8.00271809e-01 7.03037679e-01 3.13356891e-02
-3.63113910e-01 3.68954718e-01 -1.40534803e-01 2.42240027e-01
-3.99221748e-01 -6.90541267e-01 1.87360734e-01 3.97884667e-01
5.76264523e-02 -5.29723942e-01 -9.49906468e-01 3.27655584e-01
-1.17905760e+00 1.71326660e-02 5.80791414e-01 7.97238588e-01
-3.93055603e-02 4.96234030e-01 1.12590075e-01 -5.35150409e-01
5.49238384e-01 -1.09936018e-02 -6.47487879e-01 4.33194004e-02
-6.23412192e-01 2.05131859e-01 5.69201410e-01 -4.48634744e-01
-1.47420895e+00 5.86205255e-03 1.71595633e-01 -6.85698450e-01
-3.92257661e-01 2.17653081e-01 5.76698706e-02 -2.96714425e-01
6.26004040e-01 3.50473464e-01 1.09482065e-01 -7.80585051e-01
4.20425802e-01 7.53240824e-01 9.00634408e-01 -8.02638009e-02
9.31747675e-01 6.94870949e-01 -1.67940091e-02 -1.11362255e+00
-6.23157680e-01 -7.30113924e-01 -4.64187771e-01 -1.31771281e-01
8.77622128e-01 -8.66801918e-01 -5.53656936e-01 1.05034649e+00
-1.27451921e+00 -2.49660730e-01 -6.29283115e-02 7.95523345e-01
-5.39299071e-01 1.07309330e+00 -8.90818477e-01 -9.70458984e-01
-1.25145186e-02 -1.29896569e+00 4.74010259e-01 3.41621876e-01
5.84574975e-02 -1.00807023e+00 -1.61126244e-03 3.34210992e-01
9.11414981e-01 -2.37152070e-01 4.66234684e-01 2.40472496e-01
-6.14969492e-01 -2.16983557e-02 -6.00829184e-01 5.35750091e-01
5.19823670e-01 -4.13569540e-01 -1.27019119e+00 -3.86105001e-01
6.12975061e-01 3.73740882e-01 1.04536819e+00 8.83666873e-01
8.01472008e-01 -3.33361536e-01 -9.24005061e-02 6.73262000e-01
1.74878001e+00 1.67843082e-03 9.69384015e-01 1.49114758e-01
6.11093640e-01 4.05320734e-01 3.20672482e-01 1.48525551e-01
-3.44045758e-02 7.79851496e-01 3.43381524e-01 -2.40950838e-01
-4.83789027e-01 4.27173674e-01 3.44073474e-01 5.59137642e-01
-3.24307740e-01 -1.37043029e-01 -6.03635967e-01 8.29414010e-01
-1.90758228e+00 -1.16612995e+00 -4.76598859e-01 2.50701928e+00
8.43075335e-01 -2.36870438e-01 -3.26811820e-01 9.97985974e-02
9.20349717e-01 2.14521527e-01 -9.87508893e-02 2.63505671e-02
-7.38361329e-02 1.41094327e-01 7.68892050e-01 1.03296697e+00
-1.12503576e+00 5.75763583e-01 6.60773325e+00 6.47853673e-01
-1.03775847e+00 2.25186944e-01 1.74263537e-01 8.09550509e-02
9.30517986e-02 1.29115298e-01 -2.53957331e-01 7.58396566e-01
8.68984401e-01 -5.13446555e-02 9.13448274e-01 4.75037873e-01
6.45553648e-01 -7.09902763e-01 -6.32518291e-01 1.52518070e+00
2.06314519e-01 -8.86087537e-01 -4.33060616e-01 1.89701002e-02
6.47551298e-01 -2.34300241e-01 -2.77626831e-02 -3.90573323e-01
6.47919327e-02 -9.29271221e-01 5.70785284e-01 9.08001125e-01
6.63339436e-01 -3.85093212e-01 9.46745515e-01 3.23197484e-01
-8.50790083e-01 1.24772228e-01 -4.98195052e-01 -2.05905348e-01
4.77355361e-01 1.14346325e+00 -2.94436812e-01 6.53861284e-01
8.44206393e-01 6.00060582e-01 -4.30923343e-01 1.48998380e+00
-1.05162844e-01 5.29515803e-01 -5.00375628e-02 5.57096839e-01
1.79039706e-02 -6.18929029e-01 6.73971534e-01 1.33850670e+00
8.15942645e-01 2.19769180e-01 -4.91000444e-01 7.84202158e-01
1.55489445e-01 -3.68837714e-01 -4.92187142e-01 3.44756633e-01
-2.95000169e-02 1.11197817e+00 -6.25332057e-01 -4.35051650e-01
-5.24660051e-01 1.75929248e+00 -1.17648147e-01 7.52372622e-01
-7.70150065e-01 -4.53628093e-01 8.15392375e-01 -1.16150435e-02
6.50590539e-01 -3.89149427e-01 -7.54161850e-02 -1.61451340e+00
-1.17495015e-01 -5.61242044e-01 -8.86857212e-02 -1.17541397e+00
-1.23320711e+00 5.34142256e-01 -4.28975634e-02 -1.24802554e+00
-7.26199597e-02 -6.34181142e-01 -4.05789673e-01 1.26446509e+00
-1.75829554e+00 -6.71612620e-01 -5.14026105e-01 5.19129753e-01
6.17374122e-01 3.45088989e-01 6.23945832e-01 3.13958853e-01
-3.54274333e-01 -9.77393538e-02 4.83489931e-01 -1.46679834e-01
1.02110720e+00 -1.29774439e+00 1.27162889e-03 1.46459985e+00
-3.10101092e-01 7.46603787e-01 1.26935840e+00 -4.46221352e-01
-1.34367990e+00 -7.55835652e-01 7.94553995e-01 -2.53382683e-01
5.82285166e-01 -3.12865227e-02 -1.15969753e+00 2.04303384e-01
6.66903138e-01 2.61598736e-01 1.91093668e-01 -4.71840829e-01
-6.42554387e-02 -4.50888276e-02 -1.12174094e+00 4.29522783e-01
6.42284751e-01 -5.75213075e-01 -7.06236243e-01 -4.34024334e-02
1.99138552e-01 -3.03243905e-01 -4.88642246e-01 1.55466914e-01
4.19303775e-01 -1.34000599e+00 1.10230875e+00 1.05016261e-01
9.25328210e-02 -7.65574872e-01 -3.81157286e-02 -1.50307262e+00
-5.07421494e-01 -9.86363471e-01 -1.59153640e-01 9.14941311e-01
-1.39698952e-01 -5.95762789e-01 3.66556406e-01 5.81410706e-01
4.94715124e-02 1.47461161e-01 -8.32681298e-01 -7.71191776e-01
-3.51379037e-01 -1.12572826e-01 1.83990523e-02 9.66616988e-01
-1.27853557e-01 8.58344287e-02 -8.58731329e-01 3.89834017e-01
1.10040748e+00 7.49778524e-02 4.29772288e-01 -9.88281965e-01
-3.30965549e-01 -3.02979469e-01 -2.23658875e-01 -1.26647925e+00
-1.71515837e-01 -1.51575461e-01 5.26994467e-01 -1.48694670e+00
3.25666755e-01 1.70177501e-02 -4.35909033e-02 -1.36265397e-01
-5.24868846e-01 5.51010489e-01 1.06278516e-01 4.86766547e-01
-1.95437744e-01 1.62140384e-01 1.17036736e+00 3.72275002e-02
7.66850123e-03 -1.12450741e-01 -4.92389232e-01 8.93554747e-01
4.86084849e-01 -1.53862149e-01 -3.02176088e-01 -6.00508153e-01
-1.84324637e-01 6.89977556e-02 6.42491460e-01 -1.08514011e+00
4.07233894e-01 1.59205385e-02 3.04259479e-01 -1.54630408e-01
4.64636683e-01 -9.66340721e-01 5.45801163e-01 3.78228009e-01
9.28166434e-02 -2.90849596e-01 2.04157695e-01 6.70534074e-01
-2.80864030e-01 -6.61109567e-01 1.19189966e+00 -3.84738028e-01
-5.86093962e-01 -2.80605197e-01 -7.27857113e-01 -3.80018800e-01
6.95483148e-01 -2.39205614e-01 -5.04917443e-01 -7.50551462e-01
-7.85123110e-01 -5.61351478e-01 9.44678068e-01 -4.09836620e-02
5.14027596e-01 -9.04272318e-01 -5.98024428e-01 6.65147752e-02
-4.17717189e-01 -3.17223877e-01 6.10507667e-01 1.24494171e+00
-8.69977176e-01 2.51773864e-01 -5.91500849e-02 -3.62264574e-01
-1.28019035e+00 8.68073761e-01 5.31915307e-01 4.46623471e-03
-4.93470490e-01 6.51298106e-01 3.03898960e-01 4.70565438e-01
-8.92890021e-02 -2.49870196e-01 6.93128332e-02 -2.40540639e-01
8.68980169e-01 5.84491849e-01 1.35880634e-01 -7.14363992e-01
-3.60360324e-01 9.24091101e-01 1.30310073e-01 -3.25375706e-01
1.12966776e+00 -6.80923104e-01 -3.53409350e-01 8.47810432e-02
1.06238580e+00 4.11380261e-01 -1.43203032e+00 -7.32093602e-02
-2.37059593e-01 -8.14515173e-01 3.95529032e-01 -5.61128080e-01
-9.69750524e-01 8.07968199e-01 8.44940782e-01 3.16133708e-01
1.40284026e+00 -1.92833245e-01 5.51803946e-01 -3.16377357e-02
2.02531546e-01 -8.16352069e-01 -2.36959815e-01 2.32911870e-01
6.84911191e-01 -1.07270586e+00 2.07196712e-01 -6.17287397e-01
-2.33703628e-01 1.08062363e+00 -2.73166085e-03 -5.81264980e-02
6.41955674e-01 2.25368962e-01 2.62748450e-01 1.02423929e-01
-1.25114352e-01 -4.83071148e-01 2.51031548e-01 5.53159773e-01
3.78096730e-01 -3.12212974e-01 -4.15668070e-01 1.00927085e-01
3.07726771e-01 2.82446891e-01 9.50799406e-01 7.63653278e-01
-4.57937509e-01 -9.27414894e-01 -9.95219767e-01 5.90302935e-03
-7.56820917e-01 -2.84930766e-01 9.51151997e-02 3.00937742e-01
-6.21563382e-03 1.28110278e+00 -2.07300007e-01 8.91824216e-02
1.19630881e-01 4.26811501e-02 7.18700528e-01 -1.34858787e-01
-9.58792418e-02 4.12177712e-01 -1.56363353e-01 -4.96762544e-01
-8.36561739e-01 -5.81769943e-01 -6.76592946e-01 -4.16659474e-01
-3.65972430e-01 2.03705892e-01 5.40814698e-01 8.04788768e-01
2.41746992e-01 2.94639111e-01 3.11459273e-01 -1.09226656e+00
-2.97341824e-01 -1.11570239e+00 -8.71241271e-01 5.97022414e-01
8.76622677e-01 -4.88109022e-01 -9.01240945e-01 6.35219812e-01] | [11.601615905761719, -2.7164061069488525] |
915920f2-6217-4883-a0bb-cfccf64d588f | pandepth-joint-panoptic-segmentation-and | 2212.14180 | null | https://arxiv.org/abs/2212.14180v1 | https://arxiv.org/pdf/2212.14180v1.pdf | PanDepth: Joint Panoptic Segmentation and Depth Completion | Understanding 3D environments semantically is pivotal in autonomous driving applications where multiple computer vision tasks are involved. Multi-task models provide different types of outputs for a given scene, yielding a more holistic representation while keeping the computational cost low. We propose a multi-task model for panoptic segmentation and depth completion using RGB images and sparse depth maps. Our model successfully predicts fully dense depth maps and performs semantic segmentation, instance segmentation, and panoptic segmentation for every input frame. Extensive experiments were done on the Virtual KITTI 2 dataset and we demonstrate that our model solves multiple tasks, without a significant increase in computational cost, while keeping high accuracy performance. Code is available at https://github.com/juanb09111/PanDepth.git | ['Esa Rahtu', 'Juan Lagos'] | 2022-12-29 | null | null | null | null | ['panoptic-segmentation', 'depth-completion'] | ['computer-vision', 'computer-vision'] | [ 1.63173199e-01 -4.45862450e-02 -9.18267518e-02 -5.67183137e-01
-8.73392701e-01 -5.26457012e-01 3.83537948e-01 -1.20439939e-01
-3.89227033e-01 3.19634318e-01 -2.68855721e-01 -4.04116333e-01
2.16158882e-01 -8.99478376e-01 -7.23297596e-01 -5.55181503e-01
3.83684486e-01 6.37863696e-01 5.68692029e-01 4.33583856e-02
3.17724586e-01 2.92555243e-01 -1.75374496e+00 1.55535191e-01
7.69225001e-01 1.15288651e+00 7.89554715e-01 9.76376832e-01
-5.38566336e-02 6.55698836e-01 -1.87962666e-01 -1.63875148e-01
6.23454213e-01 6.74186870e-02 -8.14747274e-01 3.78727883e-01
6.58160746e-01 -4.89401937e-01 -3.40651661e-01 9.82773006e-01
1.67580739e-01 2.85033047e-01 3.13762039e-01 -1.36966813e+00
-5.61338756e-03 -1.74394980e-01 -8.06041062e-01 4.30380590e-02
9.05282572e-02 2.99731106e-01 9.40224707e-01 -8.20443571e-01
3.48088562e-01 1.13847852e+00 4.69588786e-01 2.53272265e-01
-1.06069207e+00 -7.04400778e-01 2.31533542e-01 1.31528407e-01
-1.30171680e+00 -3.95364255e-01 5.81489503e-01 -4.10712212e-01
1.12294400e+00 -5.48998406e-03 8.39073062e-01 6.80740356e-01
3.62634867e-01 9.49652016e-01 1.21407509e+00 2.17734471e-01
2.51953572e-01 -2.56765783e-01 2.20293820e-01 8.18095028e-01
1.70364350e-01 5.52344769e-02 -6.42079651e-01 3.88864160e-01
9.07095850e-01 3.94027680e-02 1.09230563e-01 -2.28924781e-01
-1.15652847e+00 8.43892276e-01 3.93078744e-01 -2.83384383e-01
-2.95295924e-01 5.80845714e-01 1.30689532e-01 6.88685998e-02
5.64077020e-01 -8.60389043e-03 -5.34766138e-01 -2.36097217e-01
-7.93619096e-01 4.78546232e-01 5.33436239e-01 1.18605781e+00
1.49615800e+00 -3.19581702e-02 3.54409963e-01 6.27291322e-01
2.39976361e-01 9.16929185e-01 8.33702311e-02 -1.75526571e+00
5.21886766e-01 3.86828423e-01 9.46007110e-03 -7.02225447e-01
-6.51044726e-01 1.05681568e-01 -4.29479808e-01 4.82169956e-01
3.96282434e-01 -1.35387674e-01 -1.42185819e+00 1.28132546e+00
4.00942147e-01 2.59160936e-01 1.67505190e-01 1.04889822e+00
8.55227768e-01 8.06072593e-01 4.12647277e-02 4.14359838e-01
1.29695451e+00 -1.23022842e+00 -3.81411254e-01 -1.12709475e+00
5.63464820e-01 -7.54423141e-01 9.16028678e-01 4.88260061e-01
-8.04285944e-01 -6.28478527e-01 -9.21278715e-01 -7.18922853e-01
-2.98381001e-01 -9.09251068e-03 9.15705800e-01 3.09842706e-01
-1.27159059e+00 1.65002197e-01 -1.05286622e+00 -4.21428740e-01
4.37703580e-01 1.08175144e-01 -2.32136473e-01 -3.21709454e-01
-6.26877546e-01 8.79212618e-01 4.37695682e-01 1.47006139e-02
-1.15239990e+00 -5.78264236e-01 -1.14658809e+00 -3.85345370e-01
4.31287795e-01 -7.87778854e-01 1.63389373e+00 -5.37396193e-01
-1.43891084e+00 9.84316528e-01 -5.11254430e-01 -5.24945617e-01
4.08947349e-01 -3.77017647e-01 1.70237288e-01 3.37391704e-01
3.77844572e-01 1.41696620e+00 7.31019020e-01 -1.26601267e+00
-8.96644294e-01 -5.60744166e-01 1.19993038e-01 5.56553900e-01
3.64032656e-01 -3.76005977e-01 -8.29545379e-01 -1.12424612e-01
6.76617503e-01 -1.07270217e+00 -6.65149629e-01 9.74125415e-02
-3.28114837e-01 2.30727699e-02 9.08789098e-01 -4.91596967e-01
2.66561031e-01 -1.93063354e+00 2.04342157e-01 -1.38386011e-01
3.33545208e-01 -2.40770414e-01 -5.43135591e-02 -8.16194192e-02
5.74755490e-01 -6.33172765e-02 -6.67356253e-01 -8.58519733e-01
-2.18491957e-01 5.33471763e-01 -2.26903796e-01 2.76162773e-01
8.14166591e-02 1.03656459e+00 -7.48768926e-01 -4.04312551e-01
7.96293557e-01 3.24114472e-01 -5.15911222e-01 1.64476842e-01
-6.35506809e-01 7.23675251e-01 -5.52049696e-01 8.02516103e-01
7.03207016e-01 -2.37204313e-01 -2.01761633e-01 1.01579428e-01
-2.23305911e-01 2.64829248e-01 -1.04301834e+00 2.18421984e+00
-5.48263133e-01 8.72122645e-01 1.97051436e-01 -8.67286563e-01
8.02539289e-01 -6.50542602e-02 5.10693789e-01 -1.06572831e+00
2.96574563e-01 1.36179313e-01 -4.35022146e-01 -4.01567131e-01
9.41638529e-01 -6.26064241e-02 -3.83360475e-01 3.86840463e-01
-1.86212048e-01 -1.12476778e+00 9.28889066e-02 1.50451943e-01
7.28646040e-01 4.55117673e-01 8.49254802e-02 -1.26760066e-01
-7.98115283e-02 6.62994206e-01 6.04597330e-01 5.29097795e-01
-3.26090217e-01 7.69577444e-01 2.50109494e-01 -4.50274736e-01
-1.03601921e+00 -1.07451880e+00 -1.38700143e-01 8.33206058e-01
7.56765902e-01 -1.71508625e-01 -5.21943331e-01 -1.42029315e-01
2.32249811e-01 6.50182247e-01 -3.81957352e-01 3.00011665e-01
-3.86135548e-01 -5.09962857e-01 4.56238180e-01 5.86236238e-01
7.63053715e-01 -9.18474853e-01 -1.20479083e+00 -6.72057550e-03
-5.34929693e-01 -1.76975596e+00 -6.84368610e-02 3.73418778e-01
-1.16316724e+00 -1.12363660e+00 -4.86859262e-01 -6.00383401e-01
3.32843393e-01 8.84623826e-01 1.06269908e+00 -1.14623129e-01
-4.30506319e-01 4.34542567e-01 -1.15440391e-01 -4.94919926e-01
4.80459295e-02 -1.28421988e-02 -3.14501107e-01 -3.20749342e-01
4.01976228e-01 -4.89061713e-01 -6.49700582e-01 3.10598969e-01
-7.67414153e-01 7.05468535e-01 4.57397997e-01 2.15317860e-01
1.11855781e+00 -3.00756931e-01 1.01311125e-01 -7.21431375e-01
-1.50949374e-01 -3.79406273e-01 -1.01190722e+00 -3.30918014e-01
-4.40260440e-01 -2.05093786e-01 2.11069748e-01 1.15984745e-01
-1.16087198e+00 3.52442712e-01 -2.04362676e-01 -5.88425398e-01
-4.74277347e-01 8.92854780e-02 7.33531117e-02 -8.59646648e-02
4.77765024e-01 2.89689481e-01 6.36104643e-02 -3.37373793e-01
5.13438046e-01 2.93482602e-01 6.60015821e-01 -5.52740037e-01
5.41177452e-01 9.56412852e-01 3.36480774e-02 -9.46896255e-01
-1.07129776e+00 -7.01164365e-01 -8.94582093e-01 -3.36770684e-01
1.22593486e+00 -1.43547225e+00 -4.97323871e-01 8.76422226e-01
-1.02196872e+00 -1.04674947e+00 1.24865258e-02 4.12190378e-01
-8.67762983e-01 2.61518002e-01 -4.66005117e-01 -4.64144915e-01
-1.49535790e-01 -1.39047110e+00 1.42433572e+00 3.26314002e-01
8.63263756e-02 -8.73484552e-01 -9.86201316e-02 8.81828725e-01
1.19182982e-01 2.03364760e-01 5.08350313e-01 5.82299680e-02
-1.24658287e+00 1.98591828e-01 -5.42087913e-01 5.01467846e-02
-1.19699165e-01 -1.08318537e-01 -1.29983890e+00 4.29003090e-02
-3.11269891e-02 -4.85101074e-01 1.33730519e+00 7.10637271e-01
1.21577394e+00 3.01548451e-01 -2.24741161e-01 1.07135534e+00
1.50916600e+00 9.56551954e-02 5.17753601e-01 4.40801412e-01
1.09481931e+00 7.02925563e-01 9.15201366e-01 4.02899295e-01
1.06419635e+00 3.50433499e-01 6.26868904e-01 -1.59714013e-01
-2.53416836e-01 -1.15596056e-01 1.69395924e-01 4.42983568e-01
1.90433159e-01 -1.84466422e-01 -1.25309479e+00 8.25091302e-01
-1.94103980e+00 -5.26354074e-01 -4.37945664e-01 1.81503379e+00
3.92549008e-01 2.46095330e-01 -7.29207769e-02 -1.57844961e-01
2.31269538e-01 3.50103140e-01 -9.30215299e-01 -3.69722068e-01
-1.21320948e-01 2.08573326e-01 8.84292483e-01 8.40252280e-01
-1.16178298e+00 1.46920919e+00 6.04233313e+00 4.20577496e-01
-1.11841917e+00 2.56464511e-01 7.93271661e-01 -3.91858369e-01
-3.13569814e-01 9.84217525e-02 -9.06706691e-01 1.22793145e-01
8.60353529e-01 -8.21486861e-02 3.24011773e-01 8.68405044e-01
2.87544847e-01 -9.20396864e-01 -7.67329931e-01 1.10846567e+00
-1.13887951e-01 -1.24663103e+00 -1.35826036e-01 1.59014881e-01
8.54183078e-01 8.61577988e-01 -1.50473583e-02 1.30061330e-02
4.83779907e-01 -7.16295302e-01 1.02542412e+00 2.15382025e-01
5.75639248e-01 -6.22002125e-01 2.45446831e-01 3.86175543e-01
-1.36831963e+00 -5.06412275e-02 -3.13761562e-01 -3.31144005e-01
3.74559760e-01 6.30025089e-01 -6.21658206e-01 3.66171151e-01
8.98518264e-01 1.02800047e+00 -6.03547931e-01 1.00104141e+00
-4.19275224e-01 8.65374207e-02 -5.89845538e-01 3.47803414e-01
4.25260335e-01 -5.05734801e-01 4.69560355e-01 9.13964629e-01
3.56437176e-01 2.39121675e-01 4.91935521e-01 8.37422669e-01
1.19888008e-01 -3.21827084e-01 -7.94569135e-01 2.97954917e-01
3.26666296e-01 1.30686200e+00 -1.03489828e+00 -4.39883977e-01
-4.30416942e-01 1.05955923e+00 1.19902030e-01 4.84561533e-01
-7.77656257e-01 -9.18181688e-02 1.16824222e+00 -1.21933699e-01
2.44291350e-01 -8.12964678e-01 -1.11679757e+00 -1.04503131e+00
-6.31365925e-02 -2.19255522e-01 1.24852784e-01 -1.06883299e+00
-5.61116457e-01 3.10831130e-01 1.35692939e-01 -1.01362550e+00
-1.83668613e-01 -6.33656979e-01 -3.33510071e-01 7.36096084e-01
-1.95472991e+00 -1.21402657e+00 -8.76007199e-01 6.16644502e-01
9.72915232e-01 3.90857339e-01 4.29875374e-01 1.33741736e-01
-5.69761813e-01 -2.05027416e-01 -2.02330351e-01 -2.44179592e-01
4.54853892e-01 -1.31532252e+00 8.40600729e-01 1.04258883e+00
-8.79560709e-02 -3.61174233e-02 5.56256294e-01 -6.13363981e-01
-1.36623323e+00 -1.30044019e+00 5.33067405e-01 -6.45244956e-01
4.38476861e-01 -2.99670905e-01 -7.09340811e-01 8.99421871e-01
6.21083053e-03 -1.48161992e-01 2.62223572e-01 -1.80308580e-01
-2.11018130e-01 3.62439007e-02 -9.37620878e-01 4.03914660e-01
1.07757270e+00 -5.77389956e-01 -2.54912794e-01 3.45724732e-01
9.36355293e-01 -9.41340268e-01 -5.95982969e-01 3.66965950e-01
4.19258803e-01 -1.10516798e+00 1.03250623e+00 3.24674129e-01
5.12980878e-01 -4.84568417e-01 -3.59134942e-01 -1.00656033e+00
7.93644339e-02 -9.11992788e-02 -3.71853113e-02 4.16911244e-01
4.60448414e-01 -7.24407852e-01 8.83396029e-01 7.59934485e-01
-4.75239247e-01 -4.75928873e-01 -9.49183285e-01 -4.82982278e-01
4.83175786e-03 -1.07188094e+00 3.50286543e-01 5.77729881e-01
-5.36375463e-01 2.71207094e-01 -1.83041900e-01 5.05664110e-01
9.22615945e-01 5.50848544e-01 1.05416489e+00 -1.23138690e+00
-5.98645955e-02 -4.59995121e-01 -2.33925655e-01 -1.43058264e+00
1.37005672e-01 -8.51101041e-01 4.01680946e-01 -1.95564485e+00
4.77376729e-02 -5.53963423e-01 1.81355536e-01 8.59925687e-01
9.30407923e-03 6.32131815e-01 1.92930564e-01 3.49446476e-01
-6.77808762e-01 5.77652097e-01 1.36007774e+00 -1.89933386e-02
-2.25458473e-01 -1.22559378e-02 -5.86966574e-01 7.97425866e-01
1.01277137e+00 -3.83730173e-01 -5.96237004e-01 -1.10112274e+00
-3.12192757e-02 1.90017030e-01 6.57196045e-01 -1.12717509e+00
2.86274493e-01 -3.18144053e-01 3.45520973e-01 -9.54250872e-01
9.87484992e-01 -4.49834466e-01 -4.92997020e-02 2.99963206e-01
2.44864210e-01 -1.21271284e-02 6.64258957e-01 4.91077214e-01
-1.90264866e-01 1.31345645e-01 8.19398880e-01 -3.26694012e-01
-1.53450060e+00 5.28708637e-01 -3.80572438e-01 -7.92337731e-02
1.16568208e+00 -4.50865507e-01 -3.39286953e-01 -1.68477595e-01
-6.06271684e-01 7.03750432e-01 7.51452267e-01 5.99110603e-01
8.25022638e-01 -8.06027949e-01 -4.58852887e-01 3.15240264e-01
2.47069612e-01 8.30505073e-01 4.40471262e-01 6.52723730e-01
-8.78170967e-01 4.68594044e-01 -3.07065874e-01 -1.11932254e+00
-1.03150046e+00 -2.02997718e-02 4.32818234e-01 2.09567353e-01
-9.86619830e-01 1.00316012e+00 5.13197780e-01 -5.83834469e-01
-4.27417457e-02 -7.49879539e-01 1.08988449e-01 2.17799265e-02
2.31471300e-01 2.77633548e-01 1.20073333e-01 -6.31978095e-01
-1.63364068e-01 8.92261267e-01 1.71260804e-01 -3.30061257e-01
1.16499090e+00 -4.57555801e-01 8.31101313e-02 4.53086853e-01
1.10802972e+00 -7.03404307e-01 -1.84381795e+00 -3.07624310e-01
-1.91027388e-01 -6.42021298e-01 4.61049408e-01 -5.11011243e-01
-1.17751122e+00 1.09580195e+00 4.47473884e-01 -2.62309730e-01
1.12284279e+00 1.00946024e-01 1.06857336e+00 4.99230683e-01
7.26216614e-01 -9.86531019e-01 1.39434040e-01 7.46596754e-01
4.62660253e-01 -1.69102931e+00 -2.79837176e-02 -4.16807950e-01
-9.96660113e-01 8.88471901e-01 8.78624856e-01 -7.52857924e-02
5.73714137e-01 3.36654335e-01 3.30896795e-01 -4.52923775e-01
-6.07530594e-01 -6.05048418e-01 -6.03310838e-02 5.33651412e-01
-1.32879287e-01 3.05922955e-01 4.78687644e-01 1.32871747e-01
-3.56588632e-01 -1.22094303e-01 6.03589118e-01 9.34184670e-01
-7.40667582e-01 -8.14985275e-01 -2.67888993e-01 2.86433339e-01
-1.70041863e-02 -6.31620036e-03 -9.97315720e-02 5.58602333e-01
-1.94047485e-02 1.07730556e+00 4.96467382e-01 -4.32840705e-01
1.35257363e-01 4.57630306e-03 4.30299640e-01 -7.89335966e-01
-4.59150411e-02 9.63486955e-02 5.41349016e-02 -9.03998911e-01
-4.19364035e-01 -7.31488407e-01 -1.65182960e+00 -3.75117809e-01
3.06585670e-01 -3.78712714e-01 9.86533821e-01 9.52368379e-01
3.70684773e-01 4.97596204e-01 5.34521520e-01 -1.34570611e+00
2.62820095e-01 -7.43303955e-01 -3.32097977e-01 -6.53285980e-02
5.20982802e-01 -6.12944961e-01 -1.60658732e-01 1.59290999e-01] | [8.381888389587402, -2.337484359741211] |
740697fd-2909-406c-99a8-1209fdd30906 | flowface-semantic-flow-guided-shape-aware | 2212.02797 | null | https://arxiv.org/abs/2212.02797v1 | https://arxiv.org/pdf/2212.02797v1.pdf | FlowFace: Semantic Flow-guided Shape-aware Face Swapping | In this work, we propose a semantic flow-guided two-stage framework for shape-aware face swapping, namely FlowFace. Unlike most previous methods that focus on transferring the source inner facial features but neglect facial contours, our FlowFace can transfer both of them to a target face, thus leading to more realistic face swapping. Concretely, our FlowFace consists of a face reshaping network and a face swapping network. The face reshaping network addresses the shape outline differences between the source and target faces. It first estimates a semantic flow (i.e., face shape differences) between the source and the target face, and then explicitly warps the target face shape with the estimated semantic flow. After reshaping, the face swapping network generates inner facial features that exhibit the identity of the source face. We employ a pre-trained face masked autoencoder (MAE) to extract facial features from both the source face and the target face. In contrast to previous methods that use identity embedding to preserve identity information, the features extracted by our encoder can better capture facial appearances and identity information. Then, we develop a cross-attention fusion module to adaptively fuse inner facial features from the source face with the target facial attributes, thus leading to better identity preservation. Extensive quantitative and qualitative experiments on in-the-wild faces demonstrate that our FlowFace outperforms the state-of-the-art significantly. | ['Xin Yu', 'Yu Ding', 'Lincheng Li', 'Bowen Ma', 'Zhimeng Zhang', 'Suzhen Wang', 'Tangjie Lv', 'Changjie Fan', 'Wei zhang', 'Hao Zeng'] | 2022-12-06 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 1.48709610e-01 5.27800679e-01 1.19518213e-01 -5.48910499e-01
-2.21325547e-01 -4.87394869e-01 4.13251877e-01 -6.72675908e-01
2.87820678e-02 3.70491385e-01 2.98960090e-01 4.06223416e-01
3.79777223e-01 -8.61498356e-01 -7.64464200e-01 -7.57669747e-01
3.10085475e-01 7.30201527e-02 -3.62168878e-01 -1.80943504e-01
-5.44491671e-02 7.39388049e-01 -1.79686618e+00 3.49376053e-01
7.70768821e-01 1.36534488e+00 -3.12110543e-01 1.63821459e-01
-1.56861931e-01 4.99464631e-01 -2.77237952e-01 -7.81640828e-01
5.26209950e-01 -6.93300247e-01 -7.28242755e-01 2.63924599e-01
9.10537064e-01 -5.04367173e-01 -4.18583423e-01 1.14483333e+00
4.05686110e-01 -2.05151383e-02 3.95978779e-01 -1.62320852e+00
-1.07993495e+00 1.34351343e-01 -7.21790969e-01 -3.12742472e-01
4.95279312e-01 1.66978508e-01 6.09016180e-01 -1.18271565e+00
6.13211274e-01 1.61954725e+00 7.47241497e-01 9.63391960e-01
-1.17250264e+00 -1.21055269e+00 4.95406501e-02 1.54915890e-02
-1.47128344e+00 -9.78682101e-01 1.07152307e+00 -3.60563695e-01
1.77179202e-01 6.22446910e-02 7.21084714e-01 9.12022114e-01
-1.77047569e-02 4.75402296e-01 8.44194174e-01 -2.36572132e-01
-8.18464905e-02 2.12635770e-01 -5.12542844e-01 8.62017691e-01
-1.74753442e-01 4.39141780e-01 -5.37588894e-01 -7.70270228e-02
9.79978144e-01 1.24591589e-01 -5.43623328e-01 -2.65210003e-01
-7.94435501e-01 7.04317689e-01 6.36261404e-01 2.60764301e-01
-4.88910019e-01 1.01454005e-01 1.96163177e-01 2.96067536e-01
5.55277228e-01 -6.23150654e-02 -2.28184506e-01 3.83351147e-01
-8.10473740e-01 1.60330430e-01 6.49348080e-01 7.23515332e-01
1.15992534e+00 8.34515467e-02 -2.30655774e-01 8.40629339e-01
5.61338961e-01 4.66933191e-01 3.29527587e-01 -1.18110490e+00
-7.00479699e-03 6.86380029e-01 -1.41649023e-01 -1.16382933e+00
9.06351954e-02 8.11811686e-02 -6.39149606e-01 5.39847612e-01
2.16885090e-01 -1.96795203e-02 -7.86827922e-01 2.03135514e+00
6.35122776e-01 5.09753406e-01 1.02688447e-01 8.52054060e-01
1.05246663e+00 3.07977170e-01 1.09107159e-01 -1.43145412e-01
1.48158228e+00 -1.02668667e+00 -9.34050262e-01 -2.46314421e-01
1.22778267e-01 -7.63397932e-01 7.12144315e-01 -2.35164255e-01
-1.33275425e+00 -7.62998819e-01 -8.37636709e-01 -2.15225562e-01
-2.09378332e-01 2.31392637e-01 2.50397354e-01 5.90102792e-01
-1.27786613e+00 6.56593859e-01 -4.51388657e-01 -1.84715793e-01
9.63434160e-01 4.85799581e-01 -1.02087784e+00 -1.71941556e-02
-1.05890799e+00 6.12902403e-01 7.78816640e-02 2.57254392e-01
-7.30689168e-01 -1.19136202e+00 -1.22909272e+00 2.48326749e-01
1.45505175e-01 -8.50573540e-01 8.97533357e-01 -1.86537576e+00
-1.74365568e+00 1.12645078e+00 -4.37860310e-01 1.20172016e-01
4.82281029e-01 -2.61152238e-02 -4.26811665e-01 1.66574344e-01
1.50495380e-01 1.13041866e+00 1.39637721e+00 -1.42191708e+00
-4.11054462e-01 -5.15020430e-01 -1.48585141e-01 4.78934869e-02
-4.79325563e-01 2.71701589e-02 -4.67825890e-01 -7.49199092e-01
-2.31856227e-01 -7.71000922e-01 4.14224595e-01 7.80292451e-01
-1.15960181e-01 7.11115897e-02 1.22938490e+00 -8.97343159e-01
7.99481630e-01 -2.56326818e+00 1.55101269e-01 3.11546363e-02
1.94721952e-01 4.13735598e-01 -6.09006763e-01 8.88074040e-02
-6.46644115e-01 -6.85104877e-02 -3.04312319e-01 -6.83660686e-01
-3.40839058e-01 6.50343820e-02 -3.15148652e-01 5.48697293e-01
5.92373252e-01 1.03683531e+00 -7.84094453e-01 -3.69933933e-01
1.50382757e-01 1.03889990e+00 -7.52595127e-01 3.91789407e-01
1.05004266e-01 6.22127175e-01 -1.65840372e-01 5.78023851e-01
1.26574731e+00 1.38234049e-01 2.47011017e-02 -5.75884283e-01
1.01897374e-01 -3.13433647e-01 -8.33904088e-01 1.60955322e+00
-4.55170631e-01 3.67991030e-01 5.20670056e-01 -5.08531928e-01
1.00981450e+00 4.37173098e-01 6.32564247e-01 -6.66022360e-01
4.84503478e-01 1.16276495e-01 -2.01340586e-01 -3.91694784e-01
-2.02418137e-02 -4.73173469e-01 5.09421349e-01 4.10905004e-01
3.27337056e-01 1.44786105e-01 -2.32067689e-01 -1.63219616e-01
4.77327377e-01 3.49243313e-01 6.17626049e-02 -2.15593532e-01
9.89981830e-01 -8.43377590e-01 7.44533241e-01 -3.09005857e-01
-3.87747467e-01 8.19498003e-01 4.23753887e-01 -4.70206589e-01
-7.41902649e-01 -1.06664205e+00 6.42492101e-02 9.56657469e-01
1.94634452e-01 -1.70091510e-01 -1.23876309e+00 -9.90424216e-01
3.15453500e-01 3.67060304e-01 -1.07045460e+00 -4.42303419e-01
-5.26073933e-01 -1.27597305e-03 4.58685815e-01 6.32263422e-01
7.65784442e-01 -1.23109865e+00 -2.46691793e-01 -2.17481516e-02
-1.31511271e-01 -1.01697361e+00 -1.15761030e+00 -7.80062258e-01
-4.58314657e-01 -1.32239544e+00 -6.38308346e-01 -9.61728215e-01
9.99503911e-01 7.04666898e-02 6.69419467e-01 3.03541422e-01
-2.00713605e-01 3.34703237e-01 -2.16421634e-01 -2.71515727e-01
-5.19347608e-01 -5.20477533e-01 1.83174964e-02 9.01457071e-01
6.07382804e-02 -6.71832144e-01 -7.42206097e-01 3.71426553e-01
-9.93216038e-01 8.80120397e-02 3.15498203e-01 7.32640266e-01
2.78252393e-01 -2.13495374e-01 5.26896238e-01 -6.88274026e-01
2.48299852e-01 -2.80262023e-01 -3.15724432e-01 3.48590046e-01
-3.05497676e-01 -1.02644481e-01 5.39881885e-01 -4.18193251e-01
-1.36357439e+00 1.79391116e-01 -2.42180139e-01 -1.04079390e+00
-1.36484265e-01 -2.48915568e-01 -6.91173017e-01 -4.23550695e-01
2.64475256e-01 3.58841494e-02 6.88626707e-01 -3.45497400e-01
5.53521812e-01 4.83412862e-01 8.11563313e-01 -3.62697691e-01
1.02599299e+00 9.66017962e-01 -1.37914181e-01 -5.25353253e-01
-5.84689915e-01 8.62057060e-02 -8.71461451e-01 -3.88448477e-01
8.93852890e-01 -8.26434612e-01 -1.03478026e+00 7.56061018e-01
-1.31886113e+00 3.35878842e-02 -5.17108440e-01 2.01555956e-02
-5.46064258e-01 1.45533532e-01 -3.05746824e-01 -5.39365649e-01
-4.07051861e-01 -1.00022268e+00 1.36870456e+00 5.35816789e-01
-9.31598693e-02 -1.12174642e+00 7.15484917e-02 3.04210991e-01
4.58127409e-01 4.74313974e-01 6.63944602e-01 -1.69343904e-01
-2.04516977e-01 -6.33252561e-02 -3.80541623e-01 5.06754696e-01
8.20392549e-01 8.72406885e-02 -1.36142242e+00 -4.05169308e-01
3.52841355e-02 -1.06987774e-01 7.20348477e-01 1.74202666e-01
9.54733372e-01 -5.49639761e-01 -3.35899413e-01 8.57751489e-01
1.00135672e+00 2.20709071e-01 8.61488700e-01 -3.31202626e-01
8.16734493e-01 1.09156954e+00 2.28704065e-01 3.17210078e-01
3.28631252e-01 4.31614250e-01 4.89451349e-01 -4.60342258e-01
-5.05712390e-01 -4.78788555e-01 5.29998243e-01 3.09720486e-01
1.58346459e-01 4.15241495e-02 -3.37701321e-01 5.09013236e-01
-1.53952968e+00 -8.35237145e-01 6.73633039e-01 1.95717859e+00
7.82916129e-01 -7.28634357e-01 -7.89807439e-02 -1.03156909e-01
1.00359738e+00 1.41533613e-01 -6.77578926e-01 -1.71694681e-01
1.23570070e-01 4.92006063e-01 -1.11586392e-01 5.90900719e-01
-9.58155692e-01 1.01190019e+00 5.23493433e+00 5.12720585e-01
-1.41532803e+00 1.53759137e-01 7.17495322e-01 -1.68069452e-01
-3.68936092e-01 1.21329520e-02 -3.53200555e-01 4.40914810e-01
5.32403886e-01 -3.18391800e-01 6.44945860e-01 7.08196044e-01
-4.50489558e-02 4.66166645e-01 -1.20186710e+00 1.07227242e+00
4.01887715e-01 -1.24736178e+00 2.12326720e-01 9.93726552e-02
5.56767941e-01 -6.19334519e-01 1.49535403e-01 -7.56650269e-02
2.02450659e-02 -1.21819305e+00 7.89244890e-01 6.95087969e-01
1.28406167e+00 -8.38948190e-01 6.37270808e-01 -2.52285123e-01
-1.44844317e+00 -1.15128547e-01 1.03835046e-01 1.75153494e-01
1.01230085e-01 2.13889912e-01 -2.67491937e-01 5.76003790e-01
7.79217422e-01 8.85930121e-01 -2.50065416e-01 4.73225504e-01
-2.19278440e-01 -8.51685777e-02 1.96238328e-02 7.99727857e-01
-3.68185043e-01 -3.28634471e-01 4.53206331e-01 5.79835653e-01
3.45332772e-01 4.06778157e-02 -1.44974023e-01 1.26831090e+00
-5.74159622e-01 1.04828633e-01 -5.39377987e-01 -7.45734498e-02
5.30799568e-01 1.47131574e+00 -2.84091264e-01 -1.79633126e-01
-4.65110242e-01 1.32661235e+00 3.11151385e-01 3.64040822e-01
-7.35922217e-01 -2.40719080e-01 1.33199036e+00 3.31547707e-01
3.24548304e-01 5.81651807e-01 7.69907162e-02 -1.05576503e+00
7.72609711e-02 -6.44441128e-01 1.32964373e-01 -7.18433022e-01
-1.44801557e+00 7.27042675e-01 -3.23560417e-01 -8.78717661e-01
-2.69927271e-02 -4.64182049e-01 -1.07756650e+00 1.11685514e+00
-1.59904349e+00 -1.64262557e+00 -6.39503956e-01 7.28971183e-01
2.71043360e-01 -2.07912624e-01 7.69836307e-01 3.28964829e-01
-5.25851786e-01 8.94739628e-01 -5.21058321e-01 4.92156148e-01
8.47921491e-01 -5.15996873e-01 4.13512349e-01 5.25189996e-01
-1.68617249e-01 6.99274540e-01 2.14183077e-01 -6.28052831e-01
-1.19247305e+00 -1.42542434e+00 6.60151601e-01 -1.05968058e-01
1.75900787e-01 -3.37072164e-01 -1.19760466e+00 8.00736010e-01
4.01376665e-01 5.90952933e-01 6.61944449e-01 -6.04274154e-01
-6.80114865e-01 -2.63897449e-01 -1.51723850e+00 4.21610922e-01
1.18968606e+00 -7.61638761e-01 -3.32008243e-01 -2.08845675e-01
5.24933755e-01 -1.97379991e-01 -8.67776930e-01 4.82316226e-01
9.03423250e-01 -1.07590485e+00 8.76496971e-01 -5.64671755e-01
4.80313152e-01 -3.37130636e-01 1.24510288e-01 -1.41003072e+00
-3.33682239e-01 -6.94634795e-01 1.12482622e-01 1.62076545e+00
-2.45673344e-01 -9.70293045e-01 6.99323237e-01 5.78542054e-01
1.23377547e-01 -5.29599190e-01 -1.02136099e+00 -5.16718149e-01
3.07497531e-01 1.55367702e-01 1.07239711e+00 1.05366230e+00
-3.18207800e-01 1.17737865e-02 -1.58638135e-01 -3.49955671e-02
5.38114965e-01 1.27596587e-01 6.03747249e-01 -1.27830100e+00
9.68308151e-02 -5.79732120e-01 -2.47022748e-01 -3.65009785e-01
9.33823347e-01 -9.69606161e-01 -9.15393010e-02 -8.23901057e-01
7.02958480e-02 -1.85288236e-01 -4.40951176e-02 8.66177082e-01
-1.75645396e-01 5.89745998e-01 3.34569484e-01 -7.57316947e-02
2.42986634e-01 1.11919332e+00 1.54868436e+00 -2.69606505e-02
-1.15534157e-01 -3.52935255e-01 -1.05512595e+00 8.06860447e-01
4.38895226e-01 -2.74171203e-01 -3.12813044e-01 -2.13912591e-01
-3.01045686e-01 9.29749832e-02 6.73962712e-01 -7.67978311e-01
2.62924060e-02 -9.63587016e-02 6.09825432e-01 1.65514033e-02
3.36411536e-01 -9.13512945e-01 3.05662960e-01 3.96632493e-01
-1.06549785e-01 -1.30351946e-01 3.83097738e-01 3.56289148e-01
-2.88482279e-01 4.07492146e-02 1.18134749e+00 7.13490844e-02
-4.02310610e-01 7.62176394e-01 2.27221519e-01 -1.91135406e-01
1.34656668e+00 -1.97809264e-01 -1.87542379e-01 -3.62582088e-01
-7.13893592e-01 8.11052397e-02 7.68867671e-01 8.05151999e-01
8.34414423e-01 -1.69546139e+00 -8.56718779e-01 1.00492597e+00
4.56607491e-02 -1.16311997e-01 4.16145027e-01 6.72172189e-01
-1.96201384e-01 -1.36295304e-01 -7.27293253e-01 -1.76272929e-01
-1.41993952e+00 5.61622679e-01 7.72675812e-01 4.38861161e-01
-3.55622083e-01 9.41235363e-01 8.53448451e-01 -4.59731936e-01
-2.45029107e-01 3.30227673e-01 -1.57654867e-01 1.35265723e-01
8.53335202e-01 2.51890063e-01 -1.62767574e-01 -1.21968782e+00
-4.34379429e-01 9.49536443e-01 -2.16243714e-02 2.68834401e-02
1.18675840e+00 -5.72314411e-02 -4.88689601e-01 -2.02572942e-01
1.65067983e+00 9.62381214e-02 -1.59500146e+00 -2.13807568e-01
-6.85952961e-01 -9.10231292e-01 8.67366642e-02 -5.28578699e-01
-1.91013336e+00 8.67425561e-01 6.50665164e-01 -4.50534165e-01
1.52813351e+00 4.39120904e-02 8.68590951e-01 -6.06390297e-01
1.45844832e-01 -6.70624733e-01 1.63139582e-01 8.09745565e-02
1.30921364e+00 -1.11234641e+00 -3.99407744e-01 -7.97684789e-01
-5.52710474e-01 1.16266727e+00 8.04133773e-01 -1.34863183e-01
8.52646053e-01 3.22997689e-01 9.22136754e-02 -1.48662746e-01
-4.80668068e-01 4.35471795e-02 4.77508426e-01 7.51202703e-01
1.84114695e-01 -2.02852204e-01 2.85705775e-01 5.22830367e-01
-2.95608729e-01 2.11615220e-01 1.12449512e-01 6.08422637e-01
7.52942339e-02 -1.17311358e+00 -4.60741550e-01 1.23326667e-01
-7.35347345e-02 8.93817842e-03 -5.78397751e-01 6.00844800e-01
5.37002861e-01 6.65901363e-01 5.52662492e-01 -4.68232453e-01
3.89162272e-01 2.09985003e-01 7.39565611e-01 -4.84982938e-01
-6.67467773e-01 -5.98457009e-02 -4.85207498e-01 -9.14671600e-01
-4.15492892e-01 -6.05090320e-01 -1.29447460e+00 -6.49057090e-01
-8.61216336e-02 -2.26683304e-01 4.60030854e-01 7.92044699e-01
7.33388007e-01 3.76047194e-01 8.89136374e-01 -1.08695221e+00
-1.09528840e-01 -7.11891234e-01 -5.91722608e-01 7.96258628e-01
6.99535429e-01 -8.61673832e-01 -4.03328687e-01 3.29859614e-01] | [12.753357887268066, -0.03902426362037659] |
05edf913-6172-4c61-a9ad-7af0c6cd624d | green-stability-assumption-unsupervised | 1802.00776 | null | http://arxiv.org/abs/1802.00776v1 | http://arxiv.org/pdf/1802.00776v1.pdf | Green Stability Assumption: Unsupervised Learning for Statistics-Based Illumination Estimation | In the image processing pipeline of almost every digital camera there is a
part dedicated to computational color constancy i.e. to removing the influence
of illumination on the colors of the image scene. Some of the best known
illumination estimation methods are the so called statistics-based methods.
They are less accurate than the learning-based illumination estimation methods,
but they are faster and simpler to implement in embedded systems, which is one
of the reasons for their widespread usage. Although in the relevant literature
it often appears as if they require no training, this is not true because they
have parameter values that need to be fine-tuned in order to be more accurate.
In this paper it is first shown that the accuracy of statistics-based methods
reported in most papers was not obtained by means of the necessary
cross-validation, but by using the whole benchmark datasets for both training
and testing. After that the corrected results are given for the best known
benchmark datasets. Finally, the so called green stability assumption is
proposed that can be used to fine-tune the values of the parameters of the
statistics-based methods by using only non-calibrated images without known
ground-truth illumination. The obtained accuracy is practically the same as
when using calibrated training images, but the whole process is much faster.
The experimental results are presented and discussed. The source code is
available at http://www.fer.unizg.hr/ipg/resources/color_constancy/. | ['Sven Lončarić', 'Nikola Banić'] | 2018-02-02 | null | null | null | null | ['color-constancy'] | ['computer-vision'] | [ 1.64778173e-01 -3.15969616e-01 1.39116511e-01 -2.69661963e-01
-3.35094243e-01 -4.86585766e-01 3.53257596e-01 1.97472841e-01
-6.58959508e-01 7.70351648e-01 -5.62419772e-01 -1.27177522e-01
-3.60627472e-02 -6.33731425e-01 -6.41800106e-01 -9.84164834e-01
2.76592195e-01 2.62557685e-01 4.87529188e-01 -6.76631108e-02
2.78863192e-01 6.31432235e-01 -2.01394510e+00 -1.26328796e-01
8.36156070e-01 8.89928401e-01 2.79912740e-01 6.01764977e-01
-1.30841985e-01 3.57764661e-01 -6.14516735e-01 -2.43498772e-01
2.85650879e-01 -8.09692681e-01 -4.44568008e-01 3.18733186e-01
3.50829870e-01 -1.39141217e-01 3.01042795e-01 1.21196198e+00
3.85070443e-01 1.88237399e-01 5.42234778e-01 -1.05137181e+00
-9.56044346e-02 5.24936989e-03 -4.14442986e-01 -1.56572267e-01
1.99383289e-01 5.75288497e-02 5.54162681e-01 -5.83020985e-01
4.99592751e-01 8.61887872e-01 2.96995670e-01 2.91739017e-01
-1.23455691e+00 -4.88778085e-01 -1.53641090e-01 4.96251911e-01
-1.33084452e+00 -3.52157801e-01 7.97974467e-01 -3.83708537e-01
3.27106476e-01 4.42032963e-01 7.88650692e-01 7.45036006e-01
1.27018884e-01 1.03645504e-01 1.86451232e+00 -9.84266877e-01
3.46878290e-01 6.82028174e-01 -4.87755947e-02 5.51253676e-01
7.28844047e-01 1.00691535e-01 -5.71850650e-02 2.87672192e-01
7.27709949e-01 -2.72776395e-01 -3.65014136e-01 -3.98104966e-01
-9.37313378e-01 6.91528022e-01 4.43949312e-01 7.54301608e-01
-2.57404268e-01 -7.92508200e-02 2.92489797e-01 2.37985447e-01
4.04709637e-01 3.79279137e-01 -3.73678893e-01 3.78900133e-02
-9.02637661e-01 -1.21914208e-01 7.34580815e-01 6.23620927e-01
1.06307590e+00 1.73468053e-01 1.60023928e-01 7.24390328e-01
1.03714995e-01 6.46232069e-01 3.73490304e-01 -7.83905983e-01
-1.45305529e-01 6.62657261e-01 3.25251907e-01 -9.61647213e-01
-3.93459499e-01 -3.87838483e-01 -6.38998270e-01 8.18147957e-01
8.75849903e-01 -1.76142901e-01 -9.37598348e-01 1.36986578e+00
3.22580814e-01 -6.22674599e-02 3.67541052e-02 9.77525711e-01
4.90772754e-01 5.56023240e-01 -1.50843993e-01 -3.60858798e-01
1.22142029e+00 -5.94830394e-01 -8.34345043e-01 -5.78078888e-02
1.69799805e-01 -1.25099599e+00 1.12273729e+00 7.87042618e-01
-8.14440429e-01 -6.17253006e-01 -1.20792282e+00 1.93163767e-01
-6.48631215e-01 5.25567591e-01 4.54063326e-01 1.11196363e+00
-9.76001322e-01 5.17403722e-01 -4.74253118e-01 -6.63289368e-01
-1.57413363e-01 1.18480824e-01 -3.14223588e-01 -8.98428038e-02
-8.40115309e-01 1.09327221e+00 5.04793227e-01 2.72226065e-01
-3.34972829e-01 -1.59156978e-01 -4.55201745e-01 -1.65391982e-01
3.99222165e-01 -1.05104789e-01 8.67281258e-01 -1.59436262e+00
-1.79455316e+00 9.62734103e-01 -1.37519270e-01 -1.53811455e-01
8.51131320e-01 -3.45323533e-02 -3.89596462e-01 1.75246432e-01
-3.81491870e-01 2.95847058e-01 9.59192276e-01 -1.61229539e+00
-4.01935756e-01 -1.28057040e-02 4.24252041e-02 -1.97375983e-01
-2.59997714e-02 -7.70221651e-02 -6.75949275e-01 -2.91331619e-01
5.82055487e-02 -1.04910314e+00 4.91935499e-02 -9.57974270e-02
-3.51568341e-01 2.46141478e-01 6.52735531e-01 -6.50969446e-01
7.52436459e-01 -2.07967734e+00 -2.31059760e-01 4.19062495e-01
-3.21705222e-01 3.57749641e-01 5.19130975e-02 3.36554557e-01
-2.67755359e-01 -2.60908425e-01 -1.05159916e-01 -8.99228528e-02
-1.34729296e-01 1.29300147e-01 1.14509195e-01 7.02513099e-01
2.07361802e-02 2.56749958e-01 -6.60462022e-01 -5.39239347e-01
8.81009042e-01 7.32175648e-01 -1.33018553e-01 2.89990548e-02
-2.33710051e-01 5.88038325e-01 -3.14137787e-02 2.83678293e-01
9.65982258e-01 1.74183995e-01 1.92512125e-01 -4.84497547e-01
-4.43162203e-01 -3.32955420e-01 -1.65212035e+00 1.09173226e+00
-5.10289729e-01 6.62458837e-01 -9.56335664e-02 -1.03071523e+00
1.11084676e+00 2.53298104e-01 4.64757353e-01 -8.39866161e-01
4.62352037e-01 5.19960284e-01 6.61365613e-02 -3.43943715e-01
3.04569006e-01 -1.21990517e-01 4.40072238e-01 1.37640342e-01
-1.29906759e-01 -4.08458441e-01 7.19115794e-01 -2.97528356e-01
3.25780243e-01 3.00742805e-01 2.53532320e-01 -3.68428797e-01
9.76743519e-01 1.85419679e-01 4.88674134e-01 3.67790312e-01
1.19405910e-01 6.47440851e-01 4.74624544e-01 -1.60994604e-01
-1.12547433e+00 -6.97884262e-01 -4.21050757e-01 5.79008341e-01
3.22780162e-01 -8.72180760e-02 -1.06205022e+00 -1.32082611e-01
-3.97370547e-01 7.38473535e-01 -5.38884521e-01 5.75255975e-02
-1.65639982e-01 -8.09993923e-01 -4.36444692e-02 -4.08678427e-02
6.62460148e-01 -8.76003623e-01 -9.12843049e-01 -7.38787726e-02
7.20883459e-02 -1.07344842e+00 2.31635347e-01 3.28904510e-01
-8.66074622e-01 -1.43844807e+00 -8.52512836e-01 -2.60721207e-01
9.18070495e-01 1.81537613e-01 9.07431841e-01 3.61094713e-01
-3.98897648e-01 3.90493661e-01 -5.24492919e-01 -6.09783471e-01
-5.10499179e-01 -7.90339038e-02 -2.31867686e-01 1.35305122e-01
9.60821435e-02 -1.16315819e-01 -5.51055074e-01 5.77473342e-01
-1.04016197e+00 1.05661780e-01 7.64298022e-01 6.45421624e-01
7.06598997e-01 3.01484048e-01 4.69070785e-02 -1.07219398e+00
9.30141509e-02 2.05694601e-01 -1.18066967e+00 3.05835098e-01
-7.55281508e-01 1.16766475e-01 8.06734979e-01 -3.46909791e-01
-1.09511387e+00 3.62319797e-01 -2.27185171e-02 -1.38992101e-01
-4.97189730e-01 -2.46978663e-02 -2.33248666e-01 -3.90941024e-01
5.96205354e-01 5.74669428e-02 -1.76613092e-01 -4.03060824e-01
1.29633278e-01 2.80576259e-01 3.13814223e-01 -3.57702494e-01
9.20436203e-01 4.37067389e-01 3.32067341e-01 -1.05835438e+00
-5.43044865e-01 -4.24100548e-01 -6.16262078e-01 -5.87761343e-01
7.99786985e-01 -3.56495678e-01 -4.95490372e-01 5.53293407e-01
-9.30080771e-01 -4.50195819e-01 -2.40470186e-01 5.59656620e-01
-4.68679011e-01 3.25827509e-01 -2.59462688e-02 -1.01174390e+00
2.45349407e-02 -1.28718579e+00 5.43096662e-01 5.51712811e-01
2.13854373e-01 -1.05405486e+00 -4.74777333e-02 1.04520202e-01
4.96778369e-01 4.00875211e-01 7.30671883e-01 -2.48714816e-02
-4.83963490e-01 -3.34698021e-01 -2.93078005e-01 7.58571506e-01
2.69060522e-01 4.76618320e-01 -1.12026501e+00 -1.68557495e-01
-7.19251186e-02 1.66355982e-01 6.51782274e-01 4.64842081e-01
9.23930585e-01 1.31566465e-01 1.38918176e-01 4.74906802e-01
2.12267017e+00 1.00149445e-01 8.91839862e-01 5.38329363e-01
2.58448720e-01 6.30288899e-01 8.96386683e-01 3.18164557e-01
-1.17147945e-01 8.27196360e-01 6.26211703e-01 -4.35766906e-01
-2.55131036e-01 3.15227777e-01 2.37093255e-01 3.26997310e-01
-4.22415942e-01 -1.66023254e-01 -6.26363695e-01 3.00123364e-01
-1.59886968e+00 -8.61432612e-01 -7.35700786e-01 2.83893585e+00
6.15223765e-01 1.57168373e-01 1.28146440e-01 6.25521898e-01
7.71764636e-01 -1.99547067e-01 -1.26527756e-01 -6.35784149e-01
-2.84274161e-01 3.49329859e-01 8.58771622e-01 6.67764843e-01
-9.84141767e-01 4.74108219e-01 5.56235886e+00 6.19168162e-01
-1.42670846e+00 9.65979770e-02 5.83101809e-01 2.13828564e-01
8.89687911e-02 1.62101611e-01 -4.34822679e-01 4.72344935e-01
8.29616666e-01 -5.03663532e-02 4.46238190e-01 6.35515153e-01
5.73158145e-01 -9.87913787e-01 -5.62478602e-01 1.14852059e+00
6.17495663e-02 -7.65544057e-01 -3.15030038e-01 -1.26960158e-01
7.98635960e-01 -3.26631814e-01 -6.11683056e-02 -3.46809149e-01
-3.17326188e-01 -7.50501812e-01 6.82253361e-01 7.68640935e-01
6.65734828e-01 -7.44825125e-01 1.09702981e+00 6.33342490e-02
-8.76111150e-01 7.76419938e-02 -5.83349109e-01 1.25009060e-01
-1.32480741e-01 1.06194103e+00 -4.22576994e-01 6.78389609e-01
6.61675453e-01 2.38917321e-01 -9.71029103e-01 1.51285613e+00
-5.46167076e-01 4.80717123e-01 -3.24319929e-01 -2.08955094e-01
5.22631705e-02 -6.16551518e-01 2.70454407e-01 1.18526435e+00
4.76290137e-01 -4.01978672e-01 -3.59656930e-01 7.59851158e-01
4.76853102e-01 4.25174952e-01 -2.75032401e-01 2.06592709e-01
3.78366979e-03 1.60873890e+00 -1.11103153e+00 -1.79117307e-01
-4.41197157e-01 1.09655011e+00 -3.47938806e-01 3.94848347e-01
-8.62212956e-01 -5.13963640e-01 2.25783363e-01 2.85813719e-01
1.30697802e-01 -2.44739577e-01 -1.29832074e-01 -8.83363664e-01
-1.27346098e-01 -5.80466807e-01 1.08104251e-01 -9.70381498e-01
-8.58555257e-01 5.05851507e-01 1.61384016e-01 -1.18464911e+00
-2.06009820e-01 -1.04067755e+00 -4.55651075e-01 9.83861089e-01
-1.66691136e+00 -8.41434062e-01 -7.29432940e-01 6.23230577e-01
2.07243577e-01 2.56897539e-01 8.34617734e-01 3.50770622e-01
-6.43603742e-01 2.26837054e-01 4.42256063e-01 -1.80763155e-01
1.03467417e+00 -1.40975666e+00 -3.66383225e-01 1.19931960e+00
2.21590288e-02 2.27845386e-01 1.16523826e+00 -2.29230970e-01
-1.21353531e+00 -5.90788841e-01 5.98871648e-01 8.37683007e-02
3.72917503e-01 -2.47650146e-01 -8.64064753e-01 1.00617141e-01
3.25249851e-01 -6.91356063e-02 2.98904628e-01 -3.13268363e-01
1.85568407e-02 -4.73189414e-01 -1.09233427e+00 2.46349081e-01
3.26785058e-01 -8.05723295e-02 -1.28204435e-01 3.35421771e-01
-1.98434182e-02 -2.54270732e-01 -6.03254139e-01 -2.62525864e-03
4.45705593e-01 -1.53950393e+00 7.52057195e-01 2.26015717e-01
4.77653928e-02 -6.55192912e-01 8.11700374e-02 -1.14242232e+00
4.76782657e-02 -2.16419369e-01 5.46958089e-01 1.31646669e+00
2.77661115e-01 -9.03732777e-01 5.27091801e-01 4.20355320e-01
2.40634590e-01 -2.31678650e-01 -7.37223029e-01 -8.30393493e-01
-3.07857275e-01 -4.12564009e-01 2.41468906e-01 7.28911996e-01
-5.68826675e-01 -4.20422330e-02 -2.21848652e-01 1.13759272e-01
7.58814156e-01 1.65204853e-01 8.94625604e-01 -1.37539971e+00
-2.58917004e-01 -3.92420173e-01 -4.12106246e-01 -9.75328162e-02
-1.66797638e-01 -2.65599072e-01 5.08767441e-02 -1.54613948e+00
-7.86101595e-02 -5.09842992e-01 -1.93794593e-01 3.47935975e-01
-1.99116059e-02 6.50114715e-01 1.32981509e-01 5.03425486e-03
-2.18864933e-01 9.64072421e-02 1.18270719e+00 1.17338501e-01
-1.75966024e-01 1.20585352e-01 -2.30068371e-01 6.46815598e-01
1.09632719e+00 -3.29797804e-01 -3.68171513e-01 -5.98776899e-02
3.37252408e-01 -6.69919789e-01 6.09604120e-01 -1.27492332e+00
-1.72560755e-02 -1.82799757e-01 5.42743742e-01 -3.27943683e-01
2.77766943e-01 -1.24991715e+00 4.29828048e-01 6.03317618e-01
2.20298275e-01 -1.27928391e-01 2.71667778e-01 1.76166028e-01
-1.29445180e-01 -6.87195420e-01 1.31406569e+00 -5.89041933e-02
-8.99104953e-01 -2.31980875e-01 -3.13084275e-01 -3.27892095e-01
1.00976527e+00 -4.77709979e-01 -1.40563995e-01 -4.38659161e-01
-5.06898224e-01 -5.17917156e-01 9.48305070e-01 -1.21087894e-01
2.38829941e-01 -9.35411572e-01 -4.86462474e-01 6.86867684e-02
2.47439981e-01 -4.78685141e-01 3.30860019e-01 1.11726248e+00
-9.72459197e-01 2.94591606e-01 -4.67702210e-01 -5.57157576e-01
-1.39233005e+00 5.70788682e-01 4.65705693e-01 1.64590210e-01
-3.34149837e-01 2.53645599e-01 -1.97549313e-01 4.84198406e-02
1.09130166e-01 -3.86652142e-01 -1.28313944e-01 -9.41887125e-02
4.52648640e-01 3.56554985e-01 3.58288556e-01 -6.38266206e-01
-3.61466259e-01 8.94260764e-01 6.13333762e-01 -7.85483792e-02
1.30828619e+00 -1.43539667e-01 -3.32417458e-01 6.09447062e-01
9.14942384e-01 2.97827244e-01 -1.07055581e+00 2.14001536e-01
-1.55733630e-01 -6.83835149e-01 3.30504656e-01 -8.41009557e-01
-1.22899163e+00 8.79234731e-01 1.04144824e+00 3.73158813e-01
1.58045602e+00 -3.89232427e-01 7.49912020e-03 1.47052124e-01
3.02092552e-01 -1.29468203e+00 -2.03125179e-01 1.00541078e-01
7.14198530e-01 -1.22922540e+00 3.68721515e-01 -6.34318531e-01
-4.89734232e-01 1.53126597e+00 3.42062145e-01 -1.97429568e-01
3.52587253e-01 2.19215080e-01 3.29639733e-01 1.99613556e-01
-1.44890457e-01 -6.35389388e-01 2.54734248e-01 5.87936044e-01
6.57085299e-01 -7.07082525e-02 -7.56982446e-01 -6.62315711e-02
6.32440224e-02 4.34161387e-02 6.43088460e-01 6.03355110e-01
-4.32489038e-01 -1.40298474e+00 -8.56198847e-01 1.65944204e-01
-2.96523005e-01 2.13797539e-01 -3.58838797e-01 1.17603374e+00
4.03722107e-01 1.09283674e+00 -1.77397609e-01 -1.46674767e-01
3.88091177e-01 1.33956177e-02 8.04971099e-01 -2.04890639e-01
-4.54613954e-01 7.14927539e-02 -8.37852508e-02 -6.08401716e-01
-8.41584921e-01 -5.65181732e-01 -1.00713658e+00 -4.56054211e-01
-4.95285928e-01 2.06691980e-01 1.18372107e+00 5.93469203e-01
-2.25009859e-01 4.72328514e-01 6.26022458e-01 -9.49214458e-01
1.42693684e-01 -7.51500368e-01 -6.96331739e-01 4.26432222e-01
1.58825647e-02 -7.85832942e-01 -6.26070738e-01 1.21675149e-01] | [10.251378059387207, -2.405738115310669] |
9d3c3fa4-43cb-4ed1-8b6d-acace1680fe3 | lhdr-hdr-reconstruction-for-legacy-content | 2211.11270 | null | https://arxiv.org/abs/2211.11270v1 | https://arxiv.org/pdf/2211.11270v1.pdf | LHDR: HDR Reconstruction for Legacy Content using a Lightweight DNN | High dynamic range (HDR) image is widely-used in graphics and photography due to the rich information it contains. Recently the community has started using deep neural network (DNN) to reconstruct standard dynamic range (SDR) images into HDR. Albeit the superiority of current DNN-based methods, their application scenario is still limited: (1) heavy model impedes real-time processing, and (2) inapplicable to legacy SDR content with more degradation types. Therefore, we propose a lightweight DNN-based method trained to tackle legacy SDR. For better design, we reform the problem modeling and emphasize degradation model. Experiments show that our method reached appealing performance with minimal computational cost compared with others. | ['Xiuhua Jiang', 'Cheng Guo'] | 2022-11-21 | null | null | null | null | ['hdr-reconstruction'] | ['computer-vision'] | [ 2.57992238e-01 -5.11487424e-01 7.61748776e-02 -1.97787479e-01
-6.44628853e-02 -4.01831307e-02 3.81302446e-01 -6.65408194e-01
-2.07484916e-01 7.70420611e-01 4.07985210e-01 -2.20541731e-01
6.89241514e-02 -7.94067681e-01 -4.60450530e-01 -6.73394680e-01
2.43141532e-01 -1.07546680e-01 2.40711570e-01 -3.07656467e-01
3.39910015e-02 8.65413308e-01 -1.68297255e+00 7.21569434e-02
1.05558026e+00 1.18795323e+00 4.32907671e-01 6.25438392e-01
6.49853423e-02 1.16557825e+00 -4.98349667e-01 -1.52283654e-01
5.43928146e-01 -4.55831707e-01 -3.19716126e-01 6.33666068e-02
2.68558711e-01 -1.08150864e+00 -1.05879152e+00 1.15228927e+00
8.58181655e-01 3.24972779e-01 2.03556716e-01 -1.04735744e+00
-1.10417831e+00 1.56218797e-01 -8.92726839e-01 2.79182017e-01
1.49579212e-01 2.46516898e-01 3.44086051e-01 -6.69567466e-01
6.74601078e-01 1.30799866e+00 5.78216732e-01 6.12349272e-01
-8.75231206e-01 -6.71278358e-01 -5.50787114e-02 4.82705742e-01
-1.15383792e+00 -5.06005108e-01 6.80750251e-01 1.35800004e-01
8.29182267e-01 1.36467636e-01 6.85676455e-01 1.17095733e+00
8.87356848e-02 6.31010592e-01 1.29949629e+00 -8.78540501e-02
4.42920625e-02 -3.74132216e-01 -1.39829114e-01 2.34748632e-01
1.33057535e-01 4.18768764e-01 -3.34615648e-01 4.26320434e-01
1.15891421e+00 1.89452305e-01 -5.03716052e-01 -5.24632109e-04
-9.79379356e-01 3.48147064e-01 4.58035976e-01 2.13902533e-01
-1.80499569e-01 2.61399329e-01 3.91709626e-01 5.00583708e-01
5.04416645e-01 -7.83476606e-02 -1.57152385e-01 -3.92510414e-01
-9.21143591e-01 -3.36737954e-03 2.83886373e-01 9.24436867e-01
4.28894758e-01 4.37703967e-01 6.84805140e-02 1.26369464e+00
2.38474116e-01 7.02962458e-01 3.90754640e-01 -1.27046371e+00
2.69321769e-01 2.71912694e-01 1.34364858e-01 -1.10645616e+00
-4.03739661e-01 -4.05971646e-01 -1.53846884e+00 5.84253967e-01
1.95788071e-01 1.18003249e-01 -1.11299360e+00 1.42212343e+00
9.67722200e-03 1.45800635e-01 -2.16179136e-02 1.42081797e+00
8.79937112e-01 1.06567168e+00 -1.17147855e-01 -3.93438458e-01
9.69186425e-01 -8.08771491e-01 -1.00105798e+00 -4.27085869e-02
-6.74148276e-02 -9.19652402e-01 1.19388628e+00 6.84910953e-01
-1.06375957e+00 -7.44879663e-01 -1.21040356e+00 -4.79800045e-01
2.27726046e-02 -2.06399914e-02 5.04064798e-01 7.18275964e-01
-1.36034977e+00 6.91894412e-01 -3.82793814e-01 -2.89564610e-01
3.97411793e-01 5.64184897e-02 -1.68283042e-02 -5.14484763e-01
-1.26099300e+00 7.72915840e-01 2.46485397e-01 3.16537112e-01
-1.00422001e+00 -5.65038085e-01 -5.24948478e-01 4.69955206e-02
3.27973694e-01 -6.86623514e-01 9.83983636e-01 -9.85938489e-01
-1.98075843e+00 6.34079158e-01 1.63792804e-01 -3.35173905e-01
7.87898362e-01 -3.74708295e-01 -7.86058664e-01 1.42788798e-01
-4.08879459e-01 4.01399761e-01 1.06052816e+00 -1.44349897e+00
-5.28119326e-01 -1.75251484e-01 3.52359638e-02 3.30306679e-01
-2.76430875e-01 2.27673054e-01 -7.04419851e-01 -8.27012777e-01
3.76721174e-01 -6.29599512e-01 -4.17239917e-03 4.88828689e-01
-3.42066139e-01 1.98432893e-01 1.29635966e+00 -8.47938597e-01
1.32913005e+00 -2.22706556e+00 -2.77104884e-01 -2.30451301e-01
4.93614644e-01 8.06347847e-01 -2.03391299e-01 3.10684800e-01
-1.53894529e-01 -1.97680175e-01 -7.61118755e-02 -1.64765060e-01
-8.10298547e-02 3.11454505e-01 -6.90296173e-01 5.65858781e-01
-3.07987958e-01 6.49140418e-01 -5.98569095e-01 -2.63150275e-01
6.56517208e-01 7.65067339e-01 -2.05572769e-01 2.47505143e-01
-8.03870037e-02 3.09188724e-01 5.89167587e-02 7.39071906e-01
1.37086368e+00 -7.70576969e-02 -1.20581426e-01 -4.85602915e-01
-2.09219128e-01 -2.61851132e-01 -1.22015107e+00 1.54564285e+00
-6.62895918e-01 9.23222542e-01 -1.49116069e-02 -4.88407582e-01
1.03005040e+00 -1.59207702e-01 3.24569106e-01 -1.21573746e+00
2.11389437e-01 2.28030622e-01 -1.14581212e-01 -7.20132530e-01
7.53152072e-01 -1.01823308e-01 2.70814121e-01 2.57199615e-01
-4.76424038e-01 1.78405985e-01 -1.28657997e-01 1.49334177e-01
1.06647253e+00 2.33418331e-01 1.55907750e-01 1.26901016e-01
4.72796589e-01 -5.76633990e-01 7.32929409e-01 5.45909405e-01
-4.50913638e-01 9.67949212e-01 3.09088200e-01 -5.92236519e-01
-1.49570954e+00 -1.27401841e+00 -1.08417220e-01 8.94322515e-01
7.12938547e-01 3.23387142e-03 -3.26630235e-01 -2.47066617e-02
-3.84733886e-01 5.47921658e-01 -2.11186409e-01 -1.22468233e-01
-7.58872986e-01 -6.51166499e-01 5.88194191e-01 4.99310106e-01
1.23309207e+00 -9.22754288e-01 -4.46391404e-01 -2.89496556e-02
-1.84445649e-01 -1.19529545e+00 -5.52339613e-01 -4.20246989e-01
-9.09544885e-01 -7.46915758e-01 -1.03073442e+00 -6.24258459e-01
2.40156457e-01 7.91468322e-01 1.05256522e+00 1.64447185e-02
-2.40273088e-01 -2.23699529e-02 -3.62983674e-01 2.72008646e-02
-3.93054783e-01 -5.18433094e-01 -5.29080222e-04 -5.65322414e-02
5.00018708e-02 -8.24573040e-01 -1.02299845e+00 3.49424303e-01
-1.21914589e+00 2.67042756e-01 7.35832512e-01 6.02883816e-01
4.80072230e-01 2.44756848e-01 5.08305311e-01 -6.01880014e-01
6.02138758e-01 -1.62107423e-01 -6.65074170e-01 1.01288907e-01
-8.75658751e-01 -3.01560521e-01 9.46706235e-01 -3.40411216e-01
-1.62623632e+00 -4.48090106e-01 -2.71409005e-01 -5.16258180e-01
-1.51360899e-01 -1.12555578e-01 -4.69565421e-01 -1.45868942e-01
4.23548132e-01 3.62415910e-01 8.35384354e-02 -5.04427433e-01
3.22315037e-01 9.29290771e-01 7.54951894e-01 -2.77289271e-01
8.08014452e-01 6.83324099e-01 1.35053918e-01 -8.95535588e-01
-5.54003119e-01 -8.47965106e-03 2.56916368e-03 -4.17847008e-01
6.71958506e-01 -1.18885517e+00 -7.98837245e-01 9.25099075e-01
-1.09430861e+00 -2.99641043e-01 -1.61390454e-01 2.78916270e-01
-5.54984629e-01 7.63742924e-01 -9.66821074e-01 -5.89698255e-01
-4.32839453e-01 -8.99320066e-01 7.36283898e-01 5.67517519e-01
4.45581704e-01 -5.93805432e-01 5.50257377e-02 2.95055956e-01
8.88766885e-01 1.37139693e-01 6.81726396e-01 3.50342274e-01
-9.19016302e-01 4.71327975e-02 -1.14152694e+00 6.03253305e-01
-2.95220632e-02 9.96809676e-02 -1.16167688e+00 -2.66251832e-01
1.11827925e-01 -1.80216596e-01 8.64140451e-01 5.57796180e-01
1.65605831e+00 -1.00686923e-01 6.33202940e-02 1.09188294e+00
1.88626420e+00 2.81950533e-01 1.44892490e+00 6.07665598e-01
8.59606922e-01 1.99905872e-01 6.17975354e-01 5.37458181e-01
6.26268312e-02 6.74751699e-01 4.78109539e-01 -4.28440660e-01
-6.99532092e-01 -1.67474359e-01 3.13886076e-01 7.72161841e-01
-1.71837896e-01 -5.92846692e-01 -5.15626431e-01 2.23007932e-01
-1.65255415e+00 -1.03994834e+00 -2.44616896e-01 1.90757859e+00
5.52055776e-01 -9.04102158e-03 -1.38285175e-01 2.84005795e-03
8.23667228e-01 5.35357237e-01 -9.60155308e-01 -4.11726952e-01
-6.12527549e-01 -1.32373586e-01 5.15791237e-01 8.19213390e-02
-6.45571768e-01 6.81758225e-01 6.09814692e+00 1.10069442e+00
-1.18194616e+00 8.22882876e-02 7.89346755e-01 -2.04730168e-01
-2.36982137e-01 -3.32349867e-01 -4.47212726e-01 7.60681927e-01
8.10993969e-01 -1.15473293e-01 6.33784294e-01 8.02720010e-01
5.93455374e-01 -1.90052867e-01 -4.69352573e-01 1.57402623e+00
1.89980745e-01 -1.00804710e+00 6.38619289e-02 -4.28041928e-02
7.88254082e-01 -1.14687602e-03 3.34110051e-01 1.24659285e-01
1.41303882e-01 -9.01840389e-01 4.85139668e-01 3.88607234e-01
1.33194923e+00 -7.65066385e-01 5.39099038e-01 -8.51192605e-03
-9.07777190e-01 -1.74019590e-01 -8.27361584e-01 -9.66552121e-04
3.44035476e-01 9.11837161e-01 -1.62177041e-01 3.65057588e-01
8.93990278e-01 9.49384332e-01 -3.86644274e-01 1.01323140e+00
-2.25590110e-01 2.97367334e-01 -1.05030529e-01 3.80164832e-01
-6.12985753e-02 -4.33421820e-01 5.67891538e-01 7.98979878e-01
5.88062525e-01 8.00075904e-02 -2.59272337e-01 6.40463412e-01
-2.98578680e-01 -2.78771073e-01 -6.59196258e-01 2.93197036e-01
4.15446103e-01 1.11471176e+00 -6.49510622e-01 -2.00445741e-01
-4.84525234e-01 1.30047154e+00 -6.70188144e-02 4.86330271e-01
-9.53386009e-01 -5.83591461e-01 7.58331597e-01 2.20181704e-01
1.48900554e-01 -2.52595395e-01 -3.72502118e-01 -1.26330388e+00
1.22138187e-01 -8.16830575e-01 2.75414437e-02 -1.33585882e+00
-1.28346801e+00 7.34313488e-01 -3.12238574e-01 -1.64749956e+00
1.60018191e-01 -5.32558680e-01 -3.77147973e-01 8.15710008e-01
-1.83629274e+00 -1.01130593e+00 -8.71586919e-01 5.54120660e-01
6.59400702e-01 -8.88281390e-02 2.79269099e-01 8.51565421e-01
-7.17579842e-01 3.05349171e-01 6.12049878e-01 -1.42702729e-01
9.64773655e-01 -9.79092538e-01 3.24283391e-01 1.36658752e+00
-5.36165535e-01 4.23659354e-01 9.48348165e-01 -4.91145283e-01
-1.35923100e+00 -1.12713444e+00 4.37758029e-01 2.19308197e-01
3.54369491e-01 1.87880844e-02 -9.33848977e-01 2.53722489e-01
2.73494154e-01 1.01112954e-01 2.90536374e-01 -3.52588356e-01
-3.98525804e-01 -6.40838087e-01 -1.28289711e+00 8.06001067e-01
1.33247268e+00 -4.68916625e-01 -3.80748287e-02 -3.90279181e-02
1.01206887e+00 -4.74414021e-01 -6.71282947e-01 2.14647144e-01
5.78333139e-01 -1.58762228e+00 1.04098332e+00 2.20954970e-01
8.21163416e-01 -6.75180078e-01 -2.35647887e-01 -1.06587183e+00
-2.70151824e-01 -6.66794300e-01 -4.41545725e-01 1.19216228e+00
-3.69913429e-01 -5.54795504e-01 5.99596381e-01 6.56052649e-01
-2.84255475e-01 -6.33084238e-01 -7.96798468e-01 -8.52076411e-01
-4.24803048e-01 -3.16722155e-01 6.75520837e-01 6.51016057e-01
-6.75022244e-01 1.81304395e-01 -1.19906044e+00 8.98120329e-02
7.69328415e-01 2.70675451e-01 6.06255889e-01 -8.70286226e-01
-3.25372249e-01 -2.16521353e-01 -3.56373549e-01 -1.44209075e+00
-2.55750060e-01 -2.23818183e-01 7.13952854e-02 -1.62715340e+00
1.50601864e-01 -4.55825299e-01 -1.92169011e-01 1.22996503e-02
-6.06159717e-02 5.44129431e-01 1.99972287e-01 3.86736006e-01
-7.43611991e-01 6.67943418e-01 1.50895798e+00 1.12849250e-01
-1.69015497e-01 -2.58895934e-01 -6.56616807e-01 5.06985426e-01
6.56107068e-01 -8.17766488e-02 -5.38482606e-01 -7.24728823e-01
3.05923492e-01 2.21078590e-01 2.92600900e-01 -1.26869094e+00
2.44846977e-02 -6.47724196e-02 6.95592046e-01 -6.76286519e-01
2.32314959e-01 -8.81600142e-01 4.04648185e-01 4.32488620e-01
-7.14800358e-02 -1.48594692e-01 -2.02249852e-03 6.82899773e-01
-2.43991345e-01 8.91675204e-02 1.23662591e+00 6.70572296e-02
-1.13775241e+00 6.30760968e-01 -2.21052393e-01 -1.63435459e-01
8.71047080e-01 -3.71068180e-01 -7.39800692e-01 -6.51502192e-01
-2.56037563e-01 -2.04775974e-01 9.22913790e-01 3.94086927e-01
9.25356865e-01 -1.29065800e+00 -4.43955243e-01 1.05766594e-01
-2.04881370e-01 -2.14496255e-02 9.81854022e-01 5.21682680e-01
-1.10429835e+00 8.56781304e-02 -5.48073113e-01 -4.10300076e-01
-1.13136292e+00 7.37543821e-01 1.72196001e-01 -1.18352152e-01
-1.09610510e+00 5.28393149e-01 2.23035589e-01 9.57526863e-02
2.55097032e-01 -4.13944572e-02 -1.24919102e-01 -3.13513577e-01
8.05820942e-01 8.50033700e-01 -4.75327000e-02 -4.56535637e-01
1.22031327e-02 5.95023811e-01 -1.47929862e-01 1.68253511e-01
1.38264287e+00 -8.25159788e-01 -1.48722813e-01 8.92545730e-02
1.30105925e+00 -4.09453660e-01 -1.55990958e+00 -2.47271016e-01
-6.76462471e-01 -9.39841330e-01 3.48773390e-01 -6.91161811e-01
-1.26882672e+00 7.99901247e-01 1.13114595e+00 -1.57569069e-02
1.80830956e+00 -5.42219222e-01 1.34592628e+00 2.04028457e-01
3.43653619e-01 -1.10404217e+00 1.41107827e-01 3.49742115e-01
7.22794414e-01 -1.17334652e+00 2.83604950e-01 -3.29051316e-01
-4.58143681e-01 1.36912572e+00 7.42731690e-01 -1.21637799e-01
4.42315012e-01 2.16443285e-01 2.72188157e-01 1.29171923e-01
-5.71310222e-01 6.57017529e-02 -2.71550953e-01 8.17454457e-01
4.87526879e-02 -3.61948788e-01 -2.65625209e-01 8.52552876e-02
2.00726226e-01 4.03201789e-01 8.93090129e-01 6.82627499e-01
-4.71665144e-01 -8.84905398e-01 -2.66927540e-01 4.47710663e-01
-3.67083669e-01 -8.16721618e-02 3.84901136e-01 4.75038350e-01
1.17693253e-01 9.09682691e-01 5.52126020e-03 -3.85042787e-01
2.64661461e-01 -7.02029645e-01 3.31757009e-01 2.82338653e-02
-6.53575920e-03 -3.88725102e-02 -1.44281536e-01 -9.16130602e-01
-4.52597141e-01 -1.72094926e-01 -8.75881732e-01 -8.10822725e-01
5.47674038e-02 -5.78385890e-01 6.71166956e-01 6.24671578e-01
4.66739804e-01 5.75641692e-01 8.44235361e-01 -6.92203164e-01
-3.59589875e-01 -7.69457877e-01 -9.04802024e-01 4.65474755e-01
4.33102101e-01 -4.39305723e-01 -3.13645631e-01 9.62602943e-02] | [10.900629997253418, -2.2299203872680664] |
74bc396a-1909-42e0-a90b-b2b93ce47d5b | automated-detection-of-covid-19-cases-from | 2109.02428 | null | https://arxiv.org/abs/2109.02428v1 | https://arxiv.org/pdf/2109.02428v1.pdf | Automated detection of COVID-19 cases from chest X-ray images using deep neural network and XGBoost | In late 2019 and after COVID-19 pandemic in the world, many researchers and scholars have tried to provide methods for detection of COVID-19 cases. Accordingly, this study focused on identifying COVID-19 cases from chest X-ray images. In this paper, a novel approach to diagnosing coronavirus disease from X-ray images was proposed. In the proposed method, DenseNet169 deep neural network was used to extract the features of X-ray images taken from the patients' chest and the extracted features were then given as input to the Extreme Gradient Boosting (XGBoost) algorithm so that it could perform the classification task. Evaluation of the proposed approach and its comparison with the methods presented in recent years revealed that the proposed method was more accurate and faster than the existing ones and had an acceptable performance in detection of COVID-19 cases from X-ray images. | ['Sharif Hasani', 'Hamid Nasiri'] | 2021-09-03 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 3.57673410e-03 -7.98669159e-01 6.96669798e-03 -4.21526849e-01
-1.61895454e-01 -2.79991150e-01 2.56508440e-01 3.62265050e-01
-6.53113961e-01 7.85929263e-01 -1.11559391e-01 -5.13641298e-01
-4.63423401e-01 -7.52310157e-01 -3.11941683e-01 -6.06852591e-01
-2.45043665e-01 7.36516595e-01 -9.83380824e-02 -9.69276875e-02
1.68646142e-01 8.80597949e-01 -1.35106385e+00 2.84413338e-01
7.46970832e-01 9.72076952e-01 6.75926447e-01 1.03032529e+00
1.74166039e-01 6.53987765e-01 -5.94116628e-01 1.58640489e-01
3.45741779e-01 -4.94676292e-01 -4.96834517e-01 -2.95912892e-01
3.90410237e-02 -9.13711607e-01 3.48265283e-02 5.82024157e-01
6.07024729e-01 3.31340879e-02 9.46189165e-01 -9.35973763e-01
-4.20983046e-01 -1.65838525e-01 -6.34490371e-01 9.36999261e-01
2.78222233e-01 -7.36078843e-02 4.09423143e-01 -7.72256494e-01
7.45426953e-01 8.84224057e-01 1.06731224e+00 4.38459009e-01
-4.50316876e-01 -7.36506939e-01 -5.16477048e-01 5.27346373e-01
-1.13059783e+00 5.50450087e-01 4.67133284e-01 -6.99915349e-01
1.00600505e+00 4.13107663e-01 9.32197511e-01 6.79311037e-01
5.36167443e-01 5.40347397e-01 1.27911091e+00 -3.00162077e-01
4.01862487e-02 1.19649172e-01 5.71052313e-01 8.02587926e-01
6.24549568e-01 3.07209879e-01 2.56977051e-01 -8.56816649e-01
4.52231973e-01 8.69564950e-01 -1.43374115e-01 1.53765589e-01
-8.46505642e-01 1.21545959e+00 6.88748896e-01 6.36842608e-01
-1.07663357e+00 -4.07398194e-01 6.23640120e-01 -3.97309177e-02
2.63580620e-01 2.95084387e-01 -6.25197828e-01 4.15263355e-01
-8.03597569e-01 4.80265439e-01 7.59872645e-02 2.00767532e-01
1.72442600e-01 -1.33358911e-01 -1.32648245e-01 4.18714911e-01
3.64450455e-01 7.87622094e-01 5.24061680e-01 -9.34464335e-02
3.37340891e-01 5.44496834e-01 1.10966668e-01 -1.02734578e+00
-8.00944507e-01 -3.64804834e-01 -9.02926922e-01 4.30733301e-02
-1.84131302e-02 -5.51547825e-01 -1.38880491e+00 1.10544050e+00
4.52514052e-01 8.58045593e-02 1.62664726e-02 1.29334509e+00
9.17518854e-01 8.83756518e-01 3.03374678e-01 -2.33425796e-01
1.79678977e+00 -3.56115133e-01 -6.54344261e-01 1.96254596e-01
4.18516189e-01 -6.71739519e-01 4.78054702e-01 4.11582977e-01
-5.03353536e-01 -6.42224252e-01 -1.03639269e+00 6.56508863e-01
-2.24006563e-01 4.50559437e-01 6.14028990e-01 7.58069754e-01
-5.67455709e-01 3.67730409e-01 -8.33989680e-01 -8.85951400e-01
5.13334453e-01 2.50656545e-01 4.99982350e-02 2.56562922e-02
-1.15982127e+00 9.73119974e-01 3.98301095e-01 7.35942721e-02
-8.08642864e-01 -4.40938085e-01 -2.10188791e-01 -9.54279304e-02
-1.73487857e-01 -9.12894964e-01 8.68073523e-01 -5.70642769e-01
-8.53608012e-01 7.80952334e-01 8.89682025e-02 -4.86373574e-01
3.38872612e-01 -3.15165609e-01 -4.35661912e-01 5.46373725e-01
-3.24174725e-02 4.15908128e-01 6.17861509e-01 -9.40505922e-01
-9.23590899e-01 -8.35326552e-01 -1.28101215e-01 4.57211286e-02
3.44647281e-02 6.32120311e-01 1.73210919e-01 -6.70623899e-01
-1.23442918e-01 -9.51351941e-01 -1.74233869e-01 -3.73162895e-01
-1.10003436e-02 -4.55156714e-01 9.92928088e-01 -9.62423921e-01
7.79645026e-01 -1.87369287e+00 -7.48863518e-01 3.67944777e-01
3.84477764e-01 9.16952312e-01 3.73179257e-01 5.79563975e-01
-8.43709037e-02 -7.72675350e-02 -7.80317560e-02 4.32010174e-01
-5.29982507e-01 2.04932336e-02 -1.73926219e-01 6.98598683e-01
7.22251236e-02 5.68752170e-01 -6.33678198e-01 -5.61395288e-01
2.87725687e-01 6.79189146e-01 -2.31698230e-01 7.35183775e-01
-1.71134949e-01 5.08363724e-01 -7.95363247e-01 6.66455209e-01
1.04786265e+00 -3.39526474e-01 5.51582836e-02 3.04400008e-02
1.50575206e-01 -2.53720403e-01 -5.88731289e-01 7.87300467e-01
-9.18097869e-02 4.89675611e-01 -1.59717966e-02 -1.00535274e+00
8.38825941e-01 7.18787193e-01 5.85155129e-01 -5.90157509e-01
7.63339937e-01 -4.45238687e-02 3.47037129e-02 -1.16380763e+00
7.00388849e-02 -3.01550180e-01 4.25972193e-01 6.18215144e-01
-4.51263636e-01 4.37333971e-01 1.42122194e-01 -2.69468158e-01
8.10697436e-01 -3.07333231e-01 4.97543246e-01 -4.77239788e-02
6.78929090e-01 6.53826833e-01 5.20113707e-01 9.53711808e-01
-2.47183308e-01 2.68265992e-01 -2.80106366e-01 -1.03993332e+00
-1.01399398e+00 -9.95081544e-01 -4.05436337e-01 8.36860538e-01
-8.55200738e-02 2.87386209e-01 -7.87814081e-01 -6.66217148e-01
-1.10481389e-01 4.26488698e-01 -6.17373168e-01 4.24085379e-01
-8.83975089e-01 -1.06174779e+00 5.00301242e-01 6.30216956e-01
6.33985221e-01 -1.11361825e+00 -1.34125698e+00 1.11404352e-01
1.34533867e-01 -5.73974490e-01 7.65484571e-02 6.28261268e-02
-1.16482615e+00 -1.29432929e+00 -9.79560971e-01 -9.35905397e-01
7.65810430e-01 3.31596881e-01 6.39011621e-01 5.69275737e-01
-9.85446513e-01 -1.74682930e-01 -5.26935101e-01 -9.90481436e-01
-3.86604697e-01 -1.39386207e-01 4.82885279e-02 -3.07062805e-01
6.82413936e-01 -3.46700139e-02 -1.09212554e+00 8.51281360e-03
-7.25150526e-01 -3.59603107e-01 5.73819876e-01 7.77428985e-01
2.32007071e-01 -1.02009520e-01 8.56331587e-01 -9.33982015e-01
8.61008525e-01 -7.04286635e-01 -6.77700818e-01 -1.10088244e-01
-8.44171107e-01 -5.46664178e-01 7.87522316e-01 1.53628383e-02
-1.06361663e+00 -7.03730732e-02 -6.32177234e-01 -3.30894887e-01
-3.54903817e-01 3.42557162e-01 7.17985511e-01 2.31515601e-01
7.28612542e-01 1.21666126e-01 -5.01212589e-02 -5.98439395e-01
-1.90922782e-01 1.13054895e+00 2.99233198e-01 1.24502748e-01
2.12372929e-01 5.46871960e-01 -8.28198493e-02 -7.00443089e-01
-5.67383945e-01 -7.52681971e-01 -1.00258142e-01 -2.53931701e-01
1.61949027e+00 -7.87742138e-01 -7.13809550e-01 5.46019256e-01
-1.23244607e+00 6.24906301e-01 4.51625288e-01 1.16849446e+00
-1.59025550e-01 1.78594753e-01 -7.41658568e-01 -8.16465557e-01
-1.09094393e+00 -1.00635946e+00 6.22805297e-01 2.76695520e-01
-1.67462289e-01 -7.31800795e-01 5.44140220e-01 6.01887524e-01
6.18466496e-01 3.28135461e-01 1.24798536e+00 -9.43910718e-01
-3.19733053e-01 -5.44169486e-01 -4.46149439e-01 2.61644363e-01
1.21610202e-01 -2.22219974e-01 -8.15819085e-01 -5.65106452e-01
6.82772338e-01 -2.07090393e-01 5.65316319e-01 6.28037453e-01
8.69059145e-01 -1.95652291e-01 -6.17720783e-01 6.87405407e-01
1.75653303e+00 9.54691350e-01 1.96962968e-01 3.36688399e-01
3.80220562e-01 -2.77404580e-02 7.92355835e-01 5.63275158e-01
8.26998055e-02 1.57740220e-01 4.50976551e-01 -4.87303466e-01
2.99525887e-01 3.99003655e-01 -5.49891531e-01 5.39941370e-01
-4.76661295e-01 -3.75421971e-01 -1.12652159e+00 4.34805155e-01
-1.41068983e+00 -1.06538534e+00 -3.95997405e-01 1.52139080e+00
2.85515845e-01 -1.59011737e-01 1.30362466e-01 -4.04429585e-02
9.77273285e-01 -2.23835692e-01 -3.67028445e-01 -6.64366722e-01
1.82769328e-01 3.11035663e-01 2.44577199e-01 -5.28218932e-02
-1.17512465e+00 -1.01054557e-01 6.91070843e+00 1.20300665e-01
-1.49616396e+00 4.97638226e-01 5.29159606e-01 6.47723302e-02
3.65270734e-01 -4.60877478e-01 -4.55720484e-01 5.19616604e-01
5.40550053e-01 1.29085407e-01 -8.65235031e-02 9.37820017e-01
2.72213906e-01 -1.45046830e-01 -4.97880131e-01 7.99452364e-01
2.44551331e-01 -1.51529634e+00 -2.06654340e-01 -1.88133150e-01
8.02447975e-01 5.43447733e-01 -3.67392689e-01 -5.70248701e-02
-1.07595377e-01 -7.12391436e-01 -1.04119986e-01 6.54222965e-01
4.38107103e-01 -8.95557344e-01 1.34930372e+00 4.64806229e-01
-7.66005874e-01 -1.84483901e-01 -1.61474153e-01 -1.11490026e-01
1.79453507e-01 2.91875690e-01 -1.68854034e+00 3.10581356e-01
9.08246875e-01 -1.25306755e-01 -2.04762593e-01 1.06187093e+00
1.67080060e-01 9.27618086e-01 -3.69097292e-01 -5.32597184e-01
4.06466097e-01 -1.12650931e-01 4.06628311e-01 1.45549476e+00
3.73631150e-01 2.57112652e-01 1.53529674e-01 5.42099655e-01
2.75727510e-01 4.05612886e-01 -7.05294669e-01 2.28702828e-01
5.90836667e-02 1.01547158e+00 -9.66559708e-01 -8.62082601e-01
-4.09828156e-01 4.18467432e-01 -2.07527116e-01 5.95783368e-02
-1.04282320e+00 -5.61237514e-01 2.48707518e-01 3.22228044e-01
7.19405174e-01 3.32923710e-01 -1.93614721e-01 -5.14715970e-01
-1.22832105e-01 -7.92280316e-01 7.27119565e-01 -9.60903585e-01
-1.18291330e+00 1.07990086e+00 1.43184796e-01 -1.04749107e+00
-2.90141791e-01 -6.58229113e-01 -8.12657118e-01 9.22659934e-01
-1.12639618e+00 -8.79407585e-01 -4.63639915e-01 3.53042513e-01
3.30238521e-01 -4.52626646e-01 8.14290047e-01 3.62972617e-01
-1.98751435e-01 6.40616715e-02 5.10586441e-01 1.31978408e-01
1.41607046e-01 -8.07567477e-01 1.63675323e-01 7.45398045e-01
-5.10350764e-01 8.49865139e-01 5.32649159e-01 -9.98274565e-01
-1.09994626e+00 -1.12323523e+00 9.47234154e-01 2.41581909e-02
4.98232804e-02 7.51892803e-03 -6.49094462e-01 4.22226876e-01
5.56160867e-01 -1.36713356e-01 6.47284925e-01 -5.57702363e-01
2.62761086e-01 -2.84011830e-02 -1.58225918e+00 -8.35520476e-02
4.10534918e-01 -2.64261395e-01 -1.10981929e+00 5.96947074e-01
2.05512315e-01 -2.05817670e-01 -5.54805577e-01 6.85679197e-01
5.52438080e-01 -9.32523072e-01 9.49756324e-01 -8.69465768e-01
9.11249444e-02 -2.84831554e-01 6.26711324e-02 -9.43702042e-01
-4.31312770e-01 3.01052060e-04 2.85320193e-01 3.97665024e-01
-2.70171668e-02 -6.13659382e-01 7.23193228e-01 -3.42027724e-01
1.39315456e-01 -1.01781392e+00 -5.70935905e-01 -3.92950475e-01
-4.06833708e-01 -1.15585476e-01 6.31736815e-01 8.66332352e-01
-5.50265014e-01 2.17306122e-01 -2.83799827e-01 2.29837507e-01
4.33841527e-01 4.61750567e-01 3.93491685e-01 -1.30442679e+00
-1.88945621e-01 1.70301959e-01 -5.77400446e-01 -3.32475454e-01
-4.77526784e-01 -7.93691397e-01 -2.52136230e-01 -1.82735896e+00
4.62778836e-01 -6.04308486e-01 -5.50972104e-01 3.65247065e-03
-4.32267249e-01 6.36074841e-01 2.52495497e-01 2.70089060e-01
5.22752143e-02 -1.98885188e-01 1.15785599e+00 2.76600779e-03
-7.00567290e-02 3.62559259e-01 -2.80400604e-01 8.09256613e-01
1.07272828e+00 -9.77979779e-01 -4.30656433e-01 -4.33096260e-01
2.28201285e-01 3.37569475e-01 3.18304390e-01 -1.04547596e+00
-1.01496883e-01 -1.10818736e-01 8.26188326e-01 -1.54294133e+00
7.57552087e-02 -8.13747823e-01 2.79657364e-01 1.02567017e+00
9.69416797e-02 6.09878123e-01 3.97535041e-02 2.95109719e-01
4.66855653e-02 -3.81679088e-01 7.23255336e-01 -1.33658260e-01
-2.46606961e-01 4.84078601e-02 -9.22004580e-01 -1.57159939e-01
1.43385971e+00 -1.67320088e-01 -2.33422950e-01 1.54465243e-01
-5.61283708e-01 -1.47790968e-01 9.79636386e-02 1.93498820e-01
8.18681598e-01 -1.01129830e+00 -9.60136533e-01 3.68646026e-01
-2.82815844e-02 -3.99662375e-01 4.24802005e-01 9.87554908e-01
-1.40323496e+00 6.51637912e-01 -6.56788588e-01 -9.05309558e-01
-1.73022056e+00 9.63959754e-01 1.67245805e-01 -2.39217833e-01
-6.31125510e-01 4.10963178e-01 6.53619245e-02 -1.19398855e-01
7.95052052e-02 -3.16249877e-01 -5.36064327e-01 -7.81690776e-02
6.15390539e-01 3.87826622e-01 3.06940287e-01 -6.12446129e-01
-7.29905605e-01 6.27961695e-01 -3.02121639e-01 4.47383255e-01
1.66239715e+00 2.07972407e-01 -1.59259200e-01 -1.26685081e-02
1.29723585e+00 -2.86705792e-01 -3.77456665e-01 1.18881837e-01
-3.88372838e-01 -4.04206693e-01 2.06112098e-02 -9.86699939e-01
-9.30246949e-01 6.84349537e-01 1.36841249e+00 1.95177093e-01
1.32805908e+00 -7.07446560e-02 1.08664286e+00 6.58864379e-01
6.66939989e-02 -8.95823061e-01 -4.67671007e-01 2.01819584e-01
4.76100177e-01 -1.03344786e+00 3.16093236e-01 1.07155941e-01
-4.41740245e-01 9.38233376e-01 3.41312557e-01 -4.56097513e-01
6.84813917e-01 1.14209779e-01 5.79860210e-01 -7.28890955e-01
-5.71675241e-01 1.72046766e-01 -7.71620646e-02 7.97010124e-01
2.97612697e-01 2.08030507e-01 -7.34005392e-01 2.14344382e-01
2.00633809e-01 5.17528772e-01 7.68264383e-02 1.36075974e+00
-5.76082885e-01 -5.01639068e-01 -8.73852611e-01 1.07256413e+00
-9.28464949e-01 4.19080406e-02 -1.24865226e-01 7.94701934e-01
6.60741448e-01 7.54072666e-01 1.24487847e-01 -2.04858765e-01
3.15952562e-02 1.72602218e-02 2.88997889e-01 -1.08399376e-01
-8.28568339e-01 -8.60038325e-02 3.20925787e-02 -5.46643063e-02
-6.33247077e-01 -3.80991161e-01 -1.51941466e+00 7.87411071e-03
-4.60776508e-01 3.84608001e-01 1.09015238e+00 8.69040370e-01
2.83327013e-01 4.80255634e-01 8.69924545e-01 -3.73603642e-01
-4.31649745e-01 -9.69029188e-01 -4.40942615e-01 4.53393847e-01
5.90849042e-01 -4.13655996e-01 -1.00472413e-01 7.80947730e-02] | [15.573984146118164, -1.7016030550003052] |
442d54e9-c091-45d0-b5c9-b1ccccd55d68 | volnet-estimating-human-body-part-volumes | 2107.02259 | null | https://arxiv.org/abs/2107.02259v1 | https://arxiv.org/pdf/2107.02259v1.pdf | VolNet: Estimating Human Body Part Volumes from a Single RGB Image | Human body volume estimation from a single RGB image is a challenging problem despite minimal attention from the research community. However VolNet, an architecture leveraging 2D and 3D pose estimation, body part segmentation and volume regression extracted from a single 2D RGB image combined with the subject's body height can be used to estimate the total body volume. VolNet is designed to predict the 2D and 3D pose as well as the body part segmentation in intermediate tasks. We generated a synthetic, large-scale dataset of photo-realistic images of human bodies with a wide range of body shapes and realistic poses called SURREALvols. By using Volnet and combining multiple stacked hourglass networks together with ResNeXt, our model correctly predicted the volume in ~82% of cases with a 10% tolerance threshold. This is a considerable improvement compared to state-of-the-art solutions such as BodyNet with only a ~38% success rate. | ['Torsten Schön', 'Vittorio Cozzolino', 'Fabian Leinen'] | 2021-07-05 | null | null | null | null | ['3d-pose-estimation', '3d-volumetric-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 8.36989135e-02 5.92330396e-01 1.22776791e-01 -2.50107944e-01
-5.50000727e-01 -2.18667835e-01 -3.75535265e-02 -1.68875590e-01
-4.66884404e-01 6.97175086e-01 2.14192197e-02 2.93737829e-01
4.86859798e-01 -6.78981781e-01 -6.89076662e-01 -2.43515581e-01
-2.12770075e-01 1.11329806e+00 3.97753924e-01 -2.30377600e-01
-3.70948106e-01 3.64232808e-01 -1.44995463e+00 -3.14873666e-01
5.64690232e-01 1.33810449e+00 -5.58814168e-01 7.44756997e-01
2.98686653e-01 1.16129100e-01 -4.50079471e-01 -6.30771756e-01
7.98142731e-01 -3.54724854e-01 -5.99200487e-01 3.57815117e-01
9.52844977e-01 -6.19474947e-01 -1.38714358e-01 4.58803982e-01
8.48699510e-01 -3.64571139e-02 4.10007507e-01 -1.10003793e+00
1.89415559e-01 1.92875534e-01 -7.34998286e-01 -1.96968332e-01
6.07993722e-01 4.62813735e-01 4.57215905e-01 -5.39671719e-01
8.15642536e-01 1.31915009e+00 1.22833169e+00 6.98690176e-01
-1.33490074e+00 -6.98802531e-01 -2.36390576e-01 -5.95776618e-01
-1.46434343e+00 -7.84314349e-02 6.44543827e-01 -2.94551551e-01
6.74050152e-01 2.98749566e-01 1.73474908e+00 6.48647428e-01
2.55645543e-01 6.24020576e-01 1.17133045e+00 -1.54638574e-01
-6.40340790e-04 -4.68737453e-01 -4.44590986e-01 1.02891719e+00
2.94743359e-01 -1.08637705e-01 -3.03470641e-01 -2.73704994e-02
1.19873977e+00 -1.68218330e-01 -3.59157808e-02 -7.55108118e-01
-1.17019999e+00 4.90598351e-01 7.45082676e-01 -3.52637917e-01
-3.92546654e-01 5.73822558e-01 2.81425744e-01 -2.60520428e-01
5.12750447e-01 2.76264459e-01 -5.54126620e-01 -5.86529560e-02
-1.11167562e+00 7.65394211e-01 9.25893724e-01 7.07507610e-01
5.52493930e-01 1.99684665e-01 -8.82418752e-02 5.05203426e-01
6.72218204e-01 8.92173707e-01 1.66280672e-01 -1.14071286e+00
5.16168296e-01 9.75173414e-01 -5.15006408e-02 -9.13524628e-01
-9.31666672e-01 -1.72020197e-01 -6.71649873e-01 3.39810282e-01
9.32605922e-01 -3.83617461e-01 -1.55795395e+00 1.36394143e+00
1.00457346e+00 -2.83572823e-01 -4.71615493e-01 1.23146486e+00
1.07448339e+00 4.25044149e-02 5.98368309e-02 2.04090133e-01
1.34963584e+00 -7.38613904e-01 -1.64538071e-01 -6.10122979e-01
1.51645914e-01 -3.94534677e-01 7.50975192e-01 2.88920969e-01
-1.43140316e+00 -3.22806001e-01 -9.83130038e-01 -2.44320452e-01
-2.87090283e-04 -8.16755593e-02 7.27275014e-01 7.92094469e-01
-9.01469588e-01 8.75190794e-01 -1.20519793e+00 -4.67659742e-01
4.39677447e-01 8.89416993e-01 -5.15300930e-01 1.12967744e-01
-7.54579306e-01 7.27731109e-01 3.09551358e-01 4.32513297e-01
-7.45367885e-01 -6.80005789e-01 -1.25010490e+00 -4.38743472e-01
4.53125954e-01 -1.24557078e+00 1.04604483e+00 -5.91844261e-01
-1.44561720e+00 1.28176820e+00 3.38621527e-01 -5.55108607e-01
1.11406052e+00 -3.89561504e-01 1.78730890e-01 1.63580030e-01
-2.29314238e-01 1.05861187e+00 8.30015481e-01 -9.71914232e-01
-9.06123314e-03 -9.04009283e-01 -3.49346548e-01 5.12450755e-01
3.21025819e-01 -2.37631679e-01 -9.60638165e-01 -3.72679710e-01
6.51380956e-01 -1.26310241e+00 -4.62728202e-01 5.48928320e-01
-7.12147832e-01 2.48286024e-01 3.64569545e-01 -1.19902909e+00
6.77359521e-01 -1.36144257e+00 2.60197341e-01 3.25145751e-01
4.31838274e-01 4.48286943e-02 3.15370262e-01 -6.53220564e-02
2.32292548e-01 1.35276899e-01 -3.13640386e-01 -5.64660370e-01
-6.32195994e-02 -1.04493536e-02 4.96202707e-01 8.03821266e-01
-2.06717849e-01 1.14746153e+00 -6.64785028e-01 -9.84377742e-01
5.82700610e-01 6.06374145e-01 -5.11988461e-01 2.14643568e-01
-1.58409446e-01 4.88917977e-01 -1.99255913e-01 1.16946912e+00
5.19118667e-01 -2.22846419e-01 4.42724407e-01 -3.95036072e-01
4.58050042e-01 -2.67723054e-01 -1.12798989e+00 1.97159374e+00
-6.48589283e-02 2.57755578e-01 2.93834388e-01 -2.49513030e-01
8.40056598e-01 1.65661573e-01 1.00280154e+00 -4.88416463e-01
5.43897033e-01 8.33919197e-02 9.18623209e-02 -1.68780029e-01
1.44984841e-01 -4.41140860e-01 -2.86492527e-01 8.13458264e-02
3.59707586e-02 -6.60776615e-01 1.04096420e-02 -9.89284813e-02
9.08586740e-01 6.70265019e-01 3.36386710e-01 3.61599959e-02
-3.29183675e-02 1.37246311e-01 6.65323138e-01 2.23742500e-01
-3.95329654e-01 1.07521558e+00 2.91478872e-01 -9.00801420e-01
-1.21829796e+00 -1.15788937e+00 1.20188355e-01 6.03803635e-01
2.52932280e-01 -5.01700938e-01 -1.04576433e+00 -6.09355509e-01
4.09336358e-01 1.92837715e-02 -8.88636410e-01 9.45427343e-02
-8.31017315e-01 -8.91108751e-01 4.73693848e-01 7.84664631e-01
7.44223475e-01 -1.07351720e+00 -1.05388570e+00 2.72123571e-02
-1.05318956e-01 -1.13249576e+00 -2.24969074e-01 -6.42368719e-02
-1.09439123e+00 -1.39224064e+00 -9.95352089e-01 -1.98340848e-01
6.43165529e-01 -7.60078490e-01 1.64808083e+00 1.88167065e-01
-7.75997162e-01 1.59562528e-01 5.84952384e-02 -4.60576475e-01
-1.55461475e-01 3.36345211e-02 2.01786403e-02 -5.34401894e-01
-2.11172059e-01 -5.70267975e-01 -1.13325071e+00 4.19650257e-01
-3.31031233e-01 3.77194881e-01 3.54065329e-01 3.23488712e-01
8.86952281e-01 -5.14683604e-01 -2.11052716e-01 -5.43093741e-01
7.37200379e-02 6.03576563e-02 -4.74561483e-01 -4.30731997e-02
-4.15839136e-01 -3.08137029e-01 1.04188137e-01 -4.42028433e-01
-5.35191476e-01 5.92559576e-01 -3.35838705e-01 -4.96900380e-01
-1.89719111e-01 -1.99720725e-01 -6.06621429e-02 -1.82296559e-01
6.53346241e-01 -5.92164248e-02 4.21136469e-01 -4.35803443e-01
2.42234200e-01 4.10456248e-02 6.66110396e-01 -4.12918210e-01
6.96637154e-01 5.88011980e-01 3.77552956e-01 -5.48722506e-01
-6.66629493e-01 -2.88667291e-01 -1.28210557e+00 -6.30999386e-01
1.05415428e+00 -1.04832733e+00 -1.05502176e+00 5.48383117e-01
-6.75466835e-01 -6.70045137e-01 -4.69790190e-01 3.95306826e-01
-7.63547182e-01 1.59211144e-01 -5.48225164e-01 -7.98675120e-01
-9.76491332e-01 -9.20014203e-01 1.61191034e+00 2.60336280e-01
-5.62434733e-01 -5.91199219e-01 -7.34461620e-02 7.13765562e-01
8.88970122e-03 1.20651972e+00 2.17470825e-01 -1.72360092e-01
-3.23203772e-01 -4.22504485e-01 -1.05040064e-02 1.14465863e-01
-2.42256731e-01 -1.34279475e-01 -6.83317482e-01 -2.56928682e-01
-6.06870413e-01 -6.69828534e-01 4.24832016e-01 7.85808921e-01
9.21399593e-01 -1.93290494e-03 -3.73971343e-01 9.82813179e-01
1.31596196e+00 -3.86876613e-01 6.49321795e-01 2.18461663e-01
1.11258554e+00 3.55677009e-01 5.94935715e-01 6.41254663e-01
5.76911271e-01 7.09549010e-01 7.94781744e-01 -4.68054920e-01
-4.40610230e-01 -4.19082791e-01 -1.76636621e-01 5.08135557e-01
-5.49398303e-01 1.27255380e-01 -1.25427139e+00 4.03314888e-01
-1.40428638e+00 -3.34270567e-01 -1.40287995e-01 2.32310796e+00
7.78472483e-01 2.53750354e-01 8.06563854e-01 -1.40054032e-01
4.94990736e-01 -1.03302933e-01 -1.00547314e+00 -1.42695969e-02
3.01409185e-01 8.55151266e-02 9.32428658e-01 -5.47575243e-02
-1.18044138e+00 9.06767368e-01 7.51764965e+00 3.82593960e-01
-9.58872616e-01 -1.59388006e-01 7.16219366e-01 -4.82084900e-01
2.48136625e-01 -3.49917173e-01 -7.09763229e-01 3.35302293e-01
7.89195657e-01 2.41186008e-01 1.69217020e-01 8.80555391e-01
-1.54266104e-01 -3.31662357e-01 -7.36726463e-01 1.19770288e+00
3.20619017e-01 -8.00511360e-01 -4.57587570e-01 3.68207574e-01
5.69769204e-01 2.28394434e-01 -4.24482167e-01 2.05510065e-01
3.66557270e-01 -1.12917781e+00 8.93697143e-01 6.32973492e-01
1.09022164e+00 -6.38928115e-01 6.31893635e-01 4.23758835e-01
-1.29136312e+00 2.84853369e-01 -1.40368879e-01 -4.00703475e-02
4.55764174e-01 2.85901695e-01 -1.03085625e+00 2.97463387e-01
1.03168571e+00 3.43776137e-01 -7.78972983e-01 1.21930528e+00
-2.11066395e-01 4.95893478e-01 -9.76352870e-01 2.71508694e-01
-3.42395484e-01 -5.26051857e-02 3.00779939e-01 7.11747885e-01
1.46887690e-01 1.43645227e-01 5.02846539e-01 8.06404173e-01
-2.71078199e-01 1.76893085e-01 -3.11244041e-01 2.43123397e-01
-5.54814264e-02 1.39703751e+00 -1.12987077e+00 -3.60160232e-01
2.41700143e-01 9.20919776e-01 9.94633585e-02 -1.15596510e-01
-9.64598715e-01 8.16810355e-02 4.12659883e-01 6.24466836e-01
9.30924565e-02 -2.17567444e-01 -3.08510453e-01 -9.12602663e-01
5.61716221e-02 -6.46671653e-01 3.17879677e-01 -9.27951217e-01
-7.90363252e-01 5.78381300e-01 1.46826774e-01 -1.08670175e+00
-5.32142580e-01 -5.78672647e-01 -1.02800727e-01 7.46226847e-01
-7.49579608e-01 -1.55445099e+00 -6.51223242e-01 3.85167599e-01
2.22718641e-01 5.17438293e-01 6.47573411e-01 -5.13613224e-02
-5.04352093e-01 4.21685994e-01 -5.53415239e-01 4.24620420e-01
5.40352285e-01 -1.49998081e+00 7.59114623e-01 2.92562276e-01
-2.64475465e-01 1.79882035e-01 8.04403067e-01 -1.10551727e+00
-1.36630654e+00 -1.00285029e+00 2.32196569e-01 -7.42361486e-01
1.76214017e-02 -5.49416840e-01 -3.87410522e-01 7.53682554e-01
-2.57599026e-01 5.63841701e-01 3.97214651e-01 2.60368790e-02
3.23440582e-02 -1.21155940e-01 -1.54807270e+00 4.91933614e-01
1.32277226e+00 2.09205389e-01 -3.69765371e-01 2.55472630e-01
4.29836243e-01 -1.33213210e+00 -1.42919719e+00 8.55145514e-01
1.16964734e+00 -1.01557660e+00 1.35149384e+00 -2.31373057e-01
4.53533143e-01 -1.38802230e-01 1.26985595e-01 -9.43639278e-01
2.07200333e-01 -2.27204055e-01 -3.63752276e-01 7.14141309e-01
9.97942463e-02 -1.34415179e-01 1.32475245e+00 1.28663099e+00
1.84206739e-01 -9.23629344e-01 -9.58057821e-01 -6.03466392e-01
-5.94574660e-02 -4.37928796e-01 3.98766756e-01 3.33908886e-01
-5.08803546e-01 -5.81709482e-02 -5.88915586e-01 -1.55634731e-01
9.59297836e-01 2.58410603e-01 1.28006029e+00 -1.39956844e+00
8.37341025e-02 -1.94563791e-02 -7.76848137e-01 -7.11946666e-01
-2.85534173e-01 -5.04343390e-01 2.73472339e-01 -1.70969808e+00
1.67913571e-01 -1.77736372e-01 3.52288842e-01 7.41175771e-01
7.69370347e-02 1.02345455e+00 4.57171470e-01 -1.30335102e-02
-5.10432780e-01 4.03667837e-01 1.64162111e+00 1.08228205e-02
-3.16233814e-01 -4.57839407e-02 -2.21730083e-01 1.10290658e+00
5.51457584e-01 -2.59489089e-01 -1.01965338e-01 -5.85046746e-02
9.83837619e-02 3.31567287e-01 6.89177990e-01 -1.45892441e+00
-1.84442565e-01 5.64832985e-02 1.22952306e+00 -9.23231363e-01
9.67308104e-01 -7.15773344e-01 4.71298933e-01 5.73381782e-01
5.47176274e-03 -4.68697064e-02 4.22957748e-01 3.01519275e-01
3.70981514e-01 4.98295248e-01 7.20769763e-01 -7.52489984e-01
-4.61521357e-01 5.29055595e-01 2.85260975e-01 1.38306737e-01
9.93021369e-01 -7.03798711e-01 2.39521489e-01 -2.49638349e-01
-1.17967868e+00 3.15602839e-01 8.79678369e-01 2.13346466e-01
6.76438093e-01 -1.18084121e+00 -5.58422506e-01 2.01584846e-01
-2.39002004e-01 5.20206571e-01 2.78158516e-01 8.51151705e-01
-9.35154855e-01 5.92189394e-02 -3.84643406e-01 -8.91428769e-01
-1.62266314e+00 -3.15009132e-02 8.00091147e-01 -3.94060820e-01
-9.49600995e-01 8.25237989e-01 -2.24997997e-01 -8.13224971e-01
8.61828856e-04 -4.47877020e-01 2.82847345e-01 -1.86558828e-01
1.03061758e-01 4.36956108e-01 -1.02018155e-01 -1.14777219e+00
-4.83559161e-01 9.26019490e-01 5.25390387e-01 -1.28761157e-01
1.08256412e+00 -2.26739779e-01 1.43500745e-01 2.99067020e-01
8.71381819e-01 -1.90770969e-01 -1.47785962e+00 3.07263970e-01
-5.49578190e-01 -4.63304847e-01 -9.82403755e-02 -1.09884632e+00
-1.46436059e+00 7.74213374e-01 9.63959277e-01 -2.58395612e-01
8.69928181e-01 1.57841414e-01 9.99300599e-01 -1.57296091e-01
6.87637508e-01 -1.16445744e+00 -1.31956294e-01 1.90914813e-02
9.91712868e-01 -1.35991168e+00 6.04064047e-01 -6.90454781e-01
-5.19601703e-01 8.41923594e-01 1.13694465e+00 -2.63144821e-01
4.62938309e-01 1.06806830e-01 2.11537704e-01 -4.44285035e-01
-1.46753132e-01 -3.61776471e-01 6.93775594e-01 5.56844115e-01
3.77455860e-01 1.27137959e-01 -1.03587009e-01 1.79817364e-01
-6.47498488e-01 1.07942373e-01 -1.13303639e-01 8.59375775e-01
-1.40989512e-01 -6.36086822e-01 -7.53692329e-01 6.11184895e-01
-6.04951441e-01 3.40960711e-01 -5.01083612e-01 1.22165954e+00
4.15556550e-01 2.48216316e-01 1.02857850e-01 -3.73519927e-01
6.15993023e-01 5.59425959e-03 7.36379087e-01 -6.45706415e-01
-7.12914586e-01 3.12849730e-01 8.22719783e-02 -1.02076089e+00
-4.18903410e-01 -5.32315552e-01 -1.44889379e+00 -2.14655325e-01
-1.05604470e-01 -5.80383122e-01 8.50211084e-01 8.40851963e-01
2.75904015e-02 4.82002169e-01 -2.36974388e-01 -1.50810027e+00
-1.66841194e-01 -1.05839026e+00 -8.21178734e-01 6.49005830e-01
-2.66416762e-02 -7.16716111e-01 -9.49578732e-03 7.35753253e-02] | [6.97276496887207, -0.9977486729621887] |
14772319-a297-49a2-845d-946be32524bb | devil-s-on-the-edges-selective-quad-attention | 2304.03495 | null | https://arxiv.org/abs/2304.03495v1 | https://arxiv.org/pdf/2304.03495v1.pdf | Devil's on the Edges: Selective Quad Attention for Scene Graph Generation | Scene graph generation aims to construct a semantic graph structure from an image such that its nodes and edges respectively represent objects and their relationships. One of the major challenges for the task lies in the presence of distracting objects and relationships in images; contextual reasoning is strongly distracted by irrelevant objects or backgrounds and, more importantly, a vast number of irrelevant candidate relations. To tackle the issue, we propose the Selective Quad Attention Network (SQUAT) that learns to select relevant object pairs and disambiguate them via diverse contextual interactions. SQUAT consists of two main components: edge selection and quad attention. The edge selection module selects relevant object pairs, i.e., edges in the scene graph, which helps contextual reasoning, and the quad attention module then updates the edge features using both edge-to-node and edge-to-edge cross-attentions to capture contextual information between objects and object pairs. Experiments demonstrate the strong performance and robustness of SQUAT, achieving the state of the art on the Visual Genome and Open Images v6 benchmarks. | ['Minsu Cho', 'Won Hwa Kim', 'Sanghyun Kim', 'Deunsol Jung'] | 2023-04-07 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Jung_Devils_on_the_Edges_Selective_Quad_Attention_for_Scene_Graph_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Jung_Devils_on_the_Edges_Selective_Quad_Attention_for_Scene_Graph_CVPR_2023_paper.pdf | cvpr-2023-1 | ['scene-graph-generation'] | ['computer-vision'] | [ 2.91234225e-01 7.41377324e-02 7.16030523e-02 -2.84537315e-01
-2.62479395e-01 -3.55425835e-01 4.84581679e-01 3.14740181e-01
-1.67232603e-01 4.10588384e-01 3.69101137e-01 1.05120361e-01
-3.28383774e-01 -8.85003030e-01 -7.76116192e-01 -5.69836318e-01
1.13355294e-02 3.78479451e-01 5.63256919e-01 -3.40784878e-01
1.95597008e-01 3.30130994e-01 -1.56321406e+00 3.94991159e-01
9.10172105e-01 9.66462553e-01 4.08635288e-01 5.10754108e-01
-2.80680120e-01 1.09822595e+00 -4.30062652e-01 -4.54314083e-01
8.05354677e-03 -4.46719170e-01 -1.00368345e+00 2.05828995e-01
5.96461236e-01 -5.63258454e-02 -5.20691037e-01 1.37744904e+00
3.63977671e-01 4.25378948e-01 3.58964860e-01 -1.52470350e+00
-1.15317380e+00 6.43787503e-01 -7.69213378e-01 3.78987938e-01
3.82247120e-01 2.85271525e-01 1.27586091e+00 -1.03801560e+00
7.64748812e-01 1.56106496e+00 9.84040797e-02 2.80725718e-01
-1.10476387e+00 -4.11186397e-01 6.84099317e-01 7.01295257e-01
-1.43670273e+00 -4.07381915e-02 9.77739751e-01 -4.02523220e-01
1.00364375e+00 2.73602784e-01 7.86683202e-01 1.10448933e+00
-6.35643899e-02 6.79491282e-01 6.90923512e-01 -1.93183169e-01
2.30396137e-01 -2.72775829e-01 4.00471300e-01 7.45263636e-01
2.83633590e-01 -1.17365241e-01 -3.85176003e-01 -4.97323126e-02
6.46719337e-01 8.57067928e-02 -4.04897422e-01 -5.48079371e-01
-1.35329020e+00 6.55463934e-01 1.13114047e+00 1.80559680e-01
-4.67844903e-01 3.02536368e-01 2.54069030e-01 -5.48155904e-02
2.26346646e-02 4.07416761e-01 -3.47312957e-01 4.41832185e-01
-3.00984513e-02 2.24891603e-01 2.90452063e-01 1.27185011e+00
8.95076752e-01 -3.10909301e-01 -7.00079203e-01 8.29727709e-01
1.57342300e-01 2.40911335e-01 8.47515464e-02 -7.14656413e-01
5.33081412e-01 1.06844497e+00 -1.12618640e-01 -1.64184952e+00
-3.56893659e-01 -4.36438590e-01 -8.29509199e-01 -7.70606920e-02
3.36604454e-02 1.61933318e-01 -1.06228817e+00 1.93725443e+00
5.85261583e-01 4.32377279e-01 -1.44688353e-01 1.09043038e+00
1.45093095e+00 3.53257746e-01 4.84847367e-01 1.00229187e-02
1.60059643e+00 -1.34188068e+00 -6.89964354e-01 -7.17843890e-01
1.65790990e-01 -5.52520692e-01 1.16400981e+00 -2.88536340e-01
-1.04446530e+00 -6.75882339e-01 -7.59256244e-01 -5.21032989e-01
-5.68552911e-01 -2.30431110e-01 6.33169532e-01 -2.99947590e-01
-8.01671207e-01 2.42194325e-01 -4.09118652e-01 -3.52112532e-01
6.66349173e-01 2.56425053e-01 -4.55949515e-01 -3.78763050e-01
-1.33336639e+00 5.75283587e-01 5.76276362e-01 2.91925699e-01
-7.97749341e-01 -4.52602774e-01 -1.21432745e+00 3.69057566e-01
8.69217277e-01 -1.02905250e+00 7.25090384e-01 -1.04477334e+00
-7.64477134e-01 1.00360370e+00 -6.40106946e-02 -1.07557945e-01
3.29867810e-01 -2.27727279e-01 -4.61244792e-01 3.07147443e-01
4.15394008e-01 7.87333190e-01 7.49356925e-01 -1.37909520e+00
-6.44103110e-01 -5.64698756e-01 2.56110221e-01 4.85299796e-01
1.06183507e-01 -3.01280320e-02 -1.15490329e+00 -6.23075366e-01
3.29279006e-01 -8.25154841e-01 -4.51304853e-01 2.81472644e-03
-6.64117813e-01 -4.36506331e-01 8.97265613e-01 -3.87763709e-01
1.05352473e+00 -2.18176579e+00 2.84692615e-01 7.46082738e-02
4.98890668e-01 1.57821298e-01 -4.83708709e-01 2.51484156e-01
-3.78996611e-01 2.51340084e-02 -1.13835752e-01 6.07675798e-02
-1.69532791e-01 2.88224787e-01 -4.63596992e-02 9.77589488e-02
5.40871263e-01 1.39529634e+00 -1.47202206e+00 -5.14896214e-01
2.17369258e-01 5.61276793e-01 -4.91514862e-01 2.64652699e-01
-4.11018908e-01 3.74713510e-01 -6.46902621e-01 6.91945434e-01
6.87842309e-01 -5.75089514e-01 6.43461570e-02 -4.75663245e-01
3.93561929e-01 2.37802088e-01 -1.09277129e+00 1.67734301e+00
-3.08001060e-02 4.63553548e-01 -9.78552550e-02 -7.38875389e-01
7.75730252e-01 -1.58864588e-01 1.82011142e-01 -8.23179960e-01
2.42176145e-01 -1.90820500e-01 7.98784941e-02 -8.04813862e-01
3.31695080e-01 4.88070101e-01 -4.53253230e-03 8.16232264e-02
1.08064853e-01 1.44672990e-01 2.45438695e-01 6.46879196e-01
1.11155939e+00 1.18323766e-01 3.34133327e-01 -3.39926243e-01
6.62134051e-01 -2.31350124e-01 7.93616116e-01 6.02946818e-01
-3.37871492e-01 5.24288476e-01 7.37402141e-01 -5.48254788e-01
-4.90642607e-01 -1.07591414e+00 3.44471276e-01 1.26815915e+00
8.90652537e-01 -5.32598197e-01 -5.62168419e-01 -9.13690805e-01
2.46973149e-02 6.10933721e-01 -1.00907624e+00 -5.44030070e-01
-5.16881287e-01 -4.62878883e-01 -2.59961128e-01 8.13648283e-01
5.72486699e-01 -1.52231526e+00 -5.13542533e-01 -4.25973460e-02
-3.48764747e-01 -1.25773394e+00 -7.27983415e-01 7.59362876e-02
-3.11283380e-01 -1.46922421e+00 -1.77404061e-01 -9.95761991e-01
1.07263982e+00 6.63301289e-01 1.58496952e+00 4.61999774e-01
-4.93744433e-01 2.26155192e-01 -3.16790462e-01 -1.93670824e-01
1.22572884e-01 -2.86169857e-01 -4.95655477e-01 2.43507698e-01
4.09603506e-01 -3.84378910e-01 -8.77032995e-01 3.43811482e-01
-7.41038322e-01 2.45212778e-01 4.62260634e-01 9.97692645e-01
8.48990679e-01 3.81395668e-02 4.66596246e-01 -1.12702000e+00
3.09424281e-01 -5.01447558e-01 -4.79284197e-01 4.84576404e-01
-1.94292784e-01 1.04033016e-01 3.71087432e-01 -1.45674706e-01
-1.03039706e+00 1.30439267e-01 1.04719728e-01 -4.84103411e-01
9.77935968e-04 3.04629326e-01 -7.28585184e-01 7.87719488e-02
5.37606180e-01 -2.22036820e-02 -6.68996274e-01 -5.64294308e-02
6.05856359e-01 9.38665792e-02 7.21840739e-01 -6.23413026e-01
5.46844840e-01 4.57975000e-01 7.35566691e-02 -5.03446221e-01
-1.16378117e+00 -4.86713529e-01 -6.77950621e-01 -2.46246770e-01
1.09178901e+00 -7.86883175e-01 -6.48008943e-01 2.34324113e-01
-1.07746077e+00 -1.93900257e-01 -4.02265996e-01 5.17361388e-02
-3.17554086e-01 1.92380875e-01 -4.86752510e-01 -4.71772641e-01
-2.39989012e-01 -1.24309742e+00 1.21644306e+00 5.28749526e-01
2.06387439e-03 -8.47599149e-01 -3.86315376e-01 4.41289514e-01
4.55867723e-02 5.28283358e-01 1.14125073e+00 -3.52381408e-01
-1.08790231e+00 1.11145109e-01 -7.67331123e-01 -1.03560016e-01
2.80967891e-01 4.44472283e-02 -9.02992368e-01 -1.84408560e-01
-4.51624930e-01 -2.67738849e-01 1.08128178e+00 2.76887804e-01
1.27191925e+00 -5.48403263e-02 -5.16290605e-01 6.43772781e-01
1.39622617e+00 1.85276583e-01 5.55730343e-01 -7.88827799e-03
1.17445266e+00 6.15029871e-01 6.43884182e-01 1.43725693e-01
4.74735081e-01 5.10328233e-01 8.32119763e-01 -2.35528350e-01
-4.46778983e-01 -4.65363204e-01 -3.20832253e-01 2.67772317e-01
2.00483009e-01 -3.47621143e-01 -8.52915764e-01 7.55782366e-01
-2.19247818e+00 -8.09401155e-01 -4.71699387e-01 2.08874202e+00
5.23054302e-01 2.36311913e-01 -1.89856201e-01 -2.67183185e-01
9.93824959e-01 2.70280659e-01 -7.74190724e-01 9.51533169e-02
-3.24377388e-01 -6.73535839e-02 5.64181544e-02 4.24585283e-01
-1.01523423e+00 1.21028066e+00 5.45562601e+00 3.47850263e-01
-6.57458961e-01 -1.30074278e-01 8.08430195e-01 5.79022765e-02
-3.81568253e-01 1.61987469e-01 -4.14940238e-01 2.97054470e-01
-1.59948822e-02 -3.90083790e-02 4.27451551e-01 8.40886652e-01
-2.52442747e-01 -1.61676586e-01 -1.04827023e+00 1.12279546e+00
2.40154400e-01 -1.12000680e+00 2.31719807e-01 -3.26060176e-01
8.74684155e-01 -7.52190277e-02 -1.14442796e-01 3.15753341e-01
5.34629941e-01 -9.01850820e-01 6.62261426e-01 4.93961126e-01
4.57455277e-01 -7.89093614e-01 5.39177179e-01 -3.21518742e-02
-1.51468611e+00 -1.18523814e-01 -4.01775539e-01 1.20879970e-01
-7.05015808e-02 5.04175663e-01 -4.22641933e-01 5.58418155e-01
9.43262577e-01 7.94978619e-01 -8.48587692e-01 9.19422150e-01
-6.94061577e-01 -4.66086827e-02 8.02698880e-02 2.61304863e-02
3.01327497e-01 -2.29836136e-01 5.44803441e-01 1.07105136e+00
-1.57175377e-01 2.85426944e-01 3.55601251e-01 1.07932222e+00
-3.16155881e-01 -3.19194049e-02 -4.97235715e-01 2.76814997e-01
4.92749959e-01 1.51262641e+00 -9.59490538e-01 -4.54291821e-01
-5.06652415e-01 1.14172339e+00 8.73928428e-01 5.75855374e-01
-8.45131636e-01 -3.19666773e-01 7.68036306e-01 -1.26171410e-01
3.84928226e-01 8.51022154e-02 -1.29415467e-01 -1.04925084e+00
-1.35829009e-03 -7.08018184e-01 9.15354729e-01 -1.22834015e+00
-1.47167468e+00 7.13627279e-01 -1.13249987e-01 -7.54020751e-01
3.64355415e-01 -4.36649829e-01 -6.52981341e-01 8.32175791e-01
-1.45937514e+00 -1.26997650e+00 -1.00684750e+00 8.00751388e-01
6.11086965e-01 3.27196777e-01 3.75018179e-01 1.11531816e-01
-6.63243175e-01 3.08553934e-01 -5.00475585e-01 2.94907510e-01
4.27002758e-01 -1.36751366e+00 6.76937759e-01 9.60676253e-01
3.04680109e-01 6.27976120e-01 3.57247323e-01 -8.02023590e-01
-1.23036444e+00 -1.40820515e+00 8.15030098e-01 -4.82045233e-01
5.37151158e-01 -5.62269449e-01 -1.05062509e+00 8.31706882e-01
2.32804880e-01 6.79311633e-01 2.64545411e-01 1.07129917e-01
-5.84958732e-01 6.17014319e-02 -7.55283058e-01 1.09701228e+00
1.73952854e+00 -5.27935982e-01 -5.84025681e-01 4.23376352e-01
8.75160336e-01 -5.19083381e-01 -3.40470612e-01 5.18526554e-01
2.01114640e-01 -1.09546065e+00 1.16777670e+00 -8.78932178e-01
4.15607989e-01 -5.48915446e-01 -6.94457889e-02 -1.27807641e+00
-8.96148443e-01 -5.61995029e-01 -1.33015458e-02 1.26781881e+00
2.65393406e-01 -3.54210973e-01 5.11243045e-01 5.57614625e-01
-1.54813781e-01 -6.92660987e-01 -5.05859911e-01 -3.93466026e-01
-4.23499227e-01 -1.91674039e-01 7.39067674e-01 1.05738389e+00
-1.97028130e-01 1.01966596e+00 -1.30237967e-01 3.23577940e-01
5.24295926e-01 6.44491434e-01 6.12456083e-01 -1.13962758e+00
-9.49122757e-02 -4.81147945e-01 -5.56310236e-01 -9.05474186e-01
1.83721721e-01 -8.41302395e-01 1.63738340e-01 -1.82645392e+00
4.14997011e-01 -2.96125084e-01 -4.80833709e-01 4.35095012e-01
-1.05196619e+00 8.94120559e-02 2.44837850e-01 -1.43071577e-01
-1.06849694e+00 6.40918374e-01 1.53124261e+00 -4.29094672e-01
-1.63115397e-01 -4.42693055e-01 -1.08954513e+00 8.16243529e-01
3.96687001e-01 -3.07412982e-01 -7.97531068e-01 -6.58377171e-01
3.26128274e-01 -2.48686895e-02 6.72731221e-01 -7.70648599e-01
2.05625474e-01 -3.34676117e-01 4.90082324e-01 -5.60517788e-01
1.51832476e-01 -9.32181418e-01 1.27982721e-01 2.71399081e-01
-4.48966205e-01 8.27474520e-02 1.09519102e-02 8.69851351e-01
-1.75533473e-01 1.94194496e-01 7.78373778e-01 -3.46047461e-01
-1.07809889e+00 7.00053394e-01 2.57304907e-01 4.06685412e-01
1.26864552e+00 2.69148666e-02 -6.39921665e-01 -1.23826809e-01
-8.80271971e-01 7.15032637e-01 3.47358972e-01 7.89564073e-01
7.21396685e-01 -1.53877664e+00 -6.73898220e-01 3.28692853e-01
6.65549874e-01 4.60412800e-01 5.21862864e-01 5.92024028e-01
-2.28946015e-01 1.21494807e-01 -2.12734625e-01 -6.59272432e-01
-1.35931003e+00 1.19808257e+00 3.85753125e-01 -2.62719505e-02
-7.50504076e-01 1.27336776e+00 8.95536482e-01 2.24024449e-02
1.99385121e-01 -4.72020715e-01 -3.87428552e-01 -1.01756573e-01
5.65119207e-01 1.24428801e-01 -7.39047751e-02 -9.48930681e-01
-5.04076123e-01 6.53121531e-01 -9.81772542e-02 4.16163623e-01
1.00635123e+00 -2.14319855e-01 -3.32170576e-01 2.10543141e-01
9.53295708e-01 -1.80695489e-01 -1.41556370e+00 -5.44757664e-01
-3.22052427e-02 -8.28090250e-01 -1.42035931e-01 -8.19158018e-01
-1.32964885e+00 7.40222275e-01 2.81000495e-01 1.80036023e-01
1.33392203e+00 2.98415601e-01 6.28367364e-01 1.68746814e-01
1.50973976e-01 -8.62191081e-01 4.37530726e-01 4.21043724e-01
1.21015680e+00 -1.30820608e+00 -1.35346606e-01 -9.74755585e-01
-7.27546096e-01 6.52557313e-01 1.09437311e+00 -1.23510882e-01
5.63866079e-01 6.24780543e-03 -1.33835554e-01 -6.00343406e-01
-9.28361654e-01 -6.71549022e-01 5.54121852e-01 7.15073407e-01
2.41290957e-01 -3.77869569e-02 -1.24709003e-01 5.28036118e-01
2.69648492e-01 -4.44947630e-01 2.81046405e-02 8.06538880e-01
-1.82841286e-01 -6.43777966e-01 -1.52469903e-01 3.16489488e-01
-6.10701777e-02 -2.38070786e-01 -7.93984294e-01 6.88371897e-01
4.01699007e-01 9.56375539e-01 1.36930108e-01 -1.87122762e-01
5.47329903e-01 -2.54277647e-01 2.65313387e-01 -5.93505085e-01
-6.17572129e-01 -3.27990279e-02 -1.05599873e-01 -8.07670653e-01
-2.34757900e-01 -4.59888071e-01 -1.50798523e+00 1.68900847e-01
-4.43766147e-01 -4.15561385e-02 4.25210893e-02 7.77517974e-01
6.69740677e-01 1.07777059e+00 4.34098572e-01 -4.53883708e-01
4.85979579e-02 -6.49032354e-01 -3.84590983e-01 1.11332333e+00
2.53076553e-01 -8.37300837e-01 4.24453802e-02 9.66361314e-02] | [10.250443458557129, 1.6046746969223022] |
99145e71-18d8-4694-bd48-95b830b7ee82 | multi-kernel-positional-embedding-convnext | 2301.06673 | null | https://arxiv.org/abs/2301.06673v2 | https://arxiv.org/pdf/2301.06673v2.pdf | Multi Kernel Positional Embedding ConvNeXt for Polyp Segmentation | Medical image segmentation is the technique that helps doctor view and has a precise diagnosis, particularly in Colorectal Cancer. Specifically, with the increase in cases, the diagnosis and identification need to be faster and more accurate for many patients; in endoscopic images, the segmentation task has been vital to helping the doctor identify the position of the polyps or the ache in the system correctly. As a result, many efforts have been made to apply deep learning to automate polyp segmentation, mostly to ameliorate the U-shape structure. However, the simple skip connection scheme in UNet leads to deficient context information and the semantic gap between feature maps from the encoder and decoder. To deal with this problem, we propose a novel framework composed of ConvNeXt backbone and Multi Kernel Positional Embedding block. Thanks to the suggested module, our method can attain better accuracy and generalization in the polyps segmentation task. Extensive experiments show that our model achieves the Dice coefficient of 0.8818 and the IOU score of 0.8163 on the Kvasir-SEG dataset. Furthermore, on various datasets, we make competitive achievement results with other previous state-of-the-art methods. | ['Hai-Dang Nguyen', 'Minh-Triet Tran', 'Nhat-Tan Bui', 'Quoc-Huy Trinh', 'Trong-Hieu Nguyen Mau'] | 2023-01-17 | null | null | null | null | ['polyp-segmentation'] | ['computer-vision'] | [-5.81318960e-02 1.76805988e-01 -1.70388788e-01 -2.41304785e-01
-5.46675146e-01 -2.89649576e-01 -1.06499411e-01 3.66534114e-01
-3.97279471e-01 1.52596742e-01 2.26952448e-01 -2.85393536e-01
-6.10828809e-02 -7.33751476e-01 -3.16420376e-01 -8.87636840e-01
6.60716668e-02 5.67274988e-02 3.68817121e-01 -3.39304134e-02
1.55203596e-01 -1.06938429e-01 -7.53691435e-01 2.66445339e-01
1.16229546e+00 1.00955784e+00 5.81218362e-01 2.99024612e-01
3.41804279e-03 4.29783523e-01 -2.02744856e-01 -5.19705951e-01
2.05034479e-01 -4.21978861e-01 -6.96942925e-01 6.90905973e-02
-2.67959654e-01 -3.90405536e-01 -4.45322275e-01 1.35434902e+00
5.44375539e-01 -1.44568101e-01 5.38520634e-01 -3.27728301e-01
-5.89018881e-01 6.87876284e-01 -8.45545053e-01 1.53071344e-01
3.06691676e-02 1.05769150e-02 7.90539563e-01 -5.32439530e-01
2.33443156e-01 8.12330067e-01 7.40108371e-01 3.38612914e-01
-6.69968843e-01 -3.75885665e-01 1.38398498e-01 1.61095127e-01
-1.23314118e+00 1.94642290e-01 4.25255477e-01 -2.18270719e-01
4.05285090e-01 2.29786336e-01 9.69490767e-01 7.01397181e-01
4.00241911e-01 9.66908991e-01 4.55542475e-01 -1.65451020e-01
-2.30396628e-01 2.75535643e-01 -3.26508284e-02 9.90024686e-01
4.92729872e-01 -1.90694749e-01 3.16349089e-01 2.81239688e-01
8.11002135e-01 3.00188392e-01 -4.91510838e-01 -2.95894861e-01
-1.21544087e+00 7.53631115e-01 1.00513422e+00 4.64085400e-01
-1.70618892e-01 -6.80532157e-02 6.10080302e-01 -8.78034458e-02
1.34597987e-01 4.82515007e-01 -2.00427249e-01 9.86344740e-02
-6.93645060e-01 -2.55757272e-01 6.11215770e-01 6.90271318e-01
1.48679927e-01 -3.66904289e-01 -1.07543409e-01 6.52355433e-01
3.37233275e-01 7.84065276e-02 9.95299637e-01 -4.78802085e-01
4.30820256e-01 1.11005878e+00 -6.04884364e-02 -1.15811837e+00
-5.57549179e-01 -8.92053127e-01 -1.17094183e+00 -3.11644733e-01
2.39636555e-01 -6.20860644e-02 -9.70051885e-01 1.24131048e+00
4.69606400e-01 2.06303746e-01 1.16575956e-01 1.05581117e+00
1.08499157e+00 6.21936917e-01 -7.04868808e-02 1.94581211e-01
1.68106461e+00 -1.35320246e+00 -5.95040858e-01 -1.71218842e-01
9.01253164e-01 -9.47493792e-01 8.67268026e-01 3.49517643e-01
-7.74772644e-01 -6.40935779e-01 -1.04082751e+00 -1.99046522e-01
-1.17666133e-01 7.44484186e-01 7.50716090e-01 5.61083853e-01
-7.01058805e-01 4.94198620e-01 -1.11281383e+00 -2.05233112e-01
5.45681417e-01 4.55155969e-01 -2.22015381e-01 -1.06497437e-01
-1.11113501e+00 6.37749135e-01 5.50461173e-01 3.87042642e-01
-4.60340768e-01 -5.87579668e-01 -8.93925786e-01 1.59956738e-01
2.51035422e-01 -7.43094087e-01 1.15385568e+00 -7.40358412e-01
-1.29601681e+00 7.70632029e-01 1.43570706e-01 -4.73960757e-01
5.65043211e-01 -3.65298428e-02 -1.91966847e-01 1.74591109e-01
5.21921962e-02 7.96926916e-01 4.52739000e-01 -7.78860033e-01
-9.61846709e-01 -3.88446450e-01 1.61476493e-01 4.19809312e-01
-4.44143027e-01 -5.19619226e-01 -8.40062201e-01 -6.61694288e-01
3.71240646e-01 -9.42688525e-01 -6.35830462e-01 1.72403991e-01
-4.69032884e-01 -1.36360396e-02 6.12522781e-01 -1.04235184e+00
1.38474202e+00 -2.55320454e+00 1.41052067e-01 7.81164691e-02
2.84018874e-01 5.01931787e-01 2.96156466e-01 -8.18132535e-02
1.54369557e-03 1.14217296e-01 -3.02951604e-01 -6.02024095e-03
-3.62726241e-01 1.36007294e-01 1.99699730e-01 5.98506153e-01
-7.99739435e-02 9.94843841e-01 -8.93571377e-01 -7.51851916e-01
3.31796885e-01 5.55462599e-01 -6.90072358e-01 2.08436504e-01
1.06188431e-01 6.66662395e-01 -7.72807658e-01 6.08170807e-01
6.27950668e-01 -6.76074564e-01 7.68302083e-02 -4.54717904e-01
-4.99093942e-02 9.51316133e-02 -9.42044139e-01 2.09374189e+00
-4.52226222e-01 1.41305432e-01 1.07695550e-01 -1.07239699e+00
6.38218164e-01 4.70392942e-01 5.78213394e-01 -6.44168437e-01
3.59640658e-01 5.05999029e-01 4.07466620e-01 -8.68551731e-01
1.76069289e-01 2.18431920e-01 -1.11090075e-02 -1.31654352e-01
-3.84344667e-01 1.19327493e-01 1.87859669e-01 -7.72956014e-02
7.89998114e-01 -1.15001440e-01 4.85205829e-01 -2.12794483e-01
8.08123350e-01 -1.51287299e-02 6.52168036e-01 1.38994917e-01
-2.75654018e-01 5.68263113e-01 4.38393354e-01 -4.79834318e-01
-7.79137015e-01 -6.49635077e-01 -4.02226388e-01 1.18607521e-01
6.47184432e-01 -1.93029463e-01 -7.89681196e-01 -8.41322601e-01
-2.37833351e-01 2.06292853e-01 -6.25619411e-01 -3.08016211e-01
-6.70335412e-01 -9.69250321e-01 2.19436288e-01 6.21994257e-01
8.72898698e-01 -6.79885447e-01 -6.22565210e-01 3.33736628e-01
-2.86842555e-01 -9.92428243e-01 -6.87387288e-01 -7.60003403e-02
-1.07540917e+00 -1.21098125e+00 -1.17504466e+00 -1.44331408e+00
1.08631468e+00 3.43683511e-01 5.82239509e-01 3.08085948e-01
-6.21997893e-01 -1.90675437e-01 -4.68350679e-01 5.63902557e-02
-1.92235023e-01 2.63997346e-01 -5.97731352e-01 -2.79002279e-01
1.36909798e-01 -2.02044457e-01 -1.31662273e+00 4.07320559e-01
-9.40349460e-01 3.54209155e-01 7.93699861e-01 1.07828391e+00
6.23713851e-01 2.56162763e-01 2.13883668e-01 -8.64455402e-01
3.85505229e-01 -3.57616484e-01 -5.72556436e-01 1.68284371e-01
-5.20586014e-01 -5.43643944e-02 7.56722212e-01 -2.09988698e-01
-8.59928250e-01 2.35036999e-01 -5.62573254e-01 -5.10711744e-02
1.02106556e-01 5.75248957e-01 5.15455417e-02 -1.03627518e-01
3.01415801e-01 2.43372649e-01 -1.87400971e-02 -3.47700834e-01
2.49654357e-03 7.77374268e-01 3.27020407e-01 1.14247739e-01
2.86213130e-01 3.34236532e-01 -1.07722305e-01 -5.39774120e-01
-7.44354665e-01 -6.51816428e-01 -3.60232323e-01 2.52356172e-01
1.04839969e+00 -1.04620838e+00 -5.74462354e-01 3.10857534e-01
-9.69938517e-01 1.92091286e-01 -7.48869777e-02 7.53086209e-01
-1.52223676e-01 7.03204691e-01 -8.77195299e-01 -1.17640711e-01
-7.01164722e-01 -1.81222928e+00 8.80302966e-01 6.54657423e-01
5.30251525e-02 -1.05746984e+00 -2.02613771e-01 1.77240968e-01
3.59406024e-01 2.95803130e-01 1.00682008e+00 -4.91683841e-01
-5.35586119e-01 -5.28101683e-01 -4.51828659e-01 2.71814793e-01
4.56622392e-01 -2.57597834e-01 -5.60133755e-01 -3.26775134e-01
2.03526661e-01 1.79183036e-01 1.03510296e+00 6.60173655e-01
1.58438671e+00 -7.07033426e-02 -7.10060000e-01 9.87540185e-01
1.37180305e+00 4.12551045e-01 4.93955702e-01 1.89619243e-01
6.82946980e-01 3.81674528e-01 6.61788821e-01 1.94547489e-01
4.60029066e-01 4.91009355e-01 6.89762771e-01 -3.92124534e-01
-2.28190169e-01 -2.39336997e-01 -6.77576987e-03 9.88394380e-01
2.73851752e-01 -1.95346788e-01 -6.30416453e-01 6.44140244e-01
-1.78659475e+00 -4.00354654e-01 -1.60716832e-01 2.00472999e+00
8.02034616e-01 3.11364271e-02 -2.65404820e-01 3.68476398e-02
6.82482541e-01 1.23970527e-02 -5.74009478e-01 -6.64517432e-02
2.93974578e-01 -2.57026911e-01 5.22417784e-01 3.10626000e-01
-1.34416330e+00 6.75382257e-01 5.20294476e+00 9.91114497e-01
-1.31269753e+00 -2.34812461e-02 8.79274964e-01 3.10048014e-01
-1.12806246e-01 -3.90367538e-01 -7.79581130e-01 6.44505918e-01
3.84614170e-01 9.85173658e-02 5.27982041e-02 8.27184737e-01
3.67415436e-02 -1.39775559e-01 -8.84601355e-01 1.09461451e+00
-7.78049305e-02 -1.11734641e+00 -1.55058399e-01 6.77162483e-02
7.35404313e-01 -1.18452132e-01 3.07987422e-01 8.96221027e-02
-2.48237088e-01 -9.06861424e-01 9.34793949e-02 1.48001730e-01
7.69791722e-01 -6.95439577e-01 1.16266179e+00 2.79914320e-01
-1.22227490e+00 -3.07783633e-02 -5.59609294e-01 5.51347733e-01
9.49869230e-02 7.29566813e-01 -1.04698014e+00 5.56121051e-01
4.78639275e-01 9.45942819e-01 -4.65979278e-01 1.65351760e+00
-4.32430357e-01 3.69937956e-01 -2.54276574e-01 -5.44102676e-02
4.16185319e-01 -2.77529120e-01 3.20488662e-01 1.17437768e+00
6.95792079e-01 7.29284585e-02 1.48915112e-01 5.09349585e-01
-1.08686842e-01 3.56111616e-01 -2.24416837e-01 8.86085350e-03
-6.27143532e-02 1.47036314e+00 -9.65142727e-01 -2.37578288e-01
-3.49211395e-01 1.15001512e+00 -1.13550879e-01 -7.59219825e-02
-1.03179049e+00 -5.00315607e-01 3.09505999e-01 2.19187737e-02
4.31722045e-01 8.72560218e-02 -2.61519045e-01 -1.24139273e+00
6.25861138e-02 -8.13452244e-01 3.50546747e-01 -2.23117933e-01
-8.32297385e-01 6.51683748e-01 -5.72655797e-01 -1.42849612e+00
5.82372472e-02 -7.18183339e-01 -6.08984053e-01 6.53542399e-01
-1.63109744e+00 -1.00229442e+00 -4.36121106e-01 2.60783195e-01
6.28820479e-01 1.30384162e-01 8.96214962e-01 5.93515217e-01
-6.73995435e-01 6.52630746e-01 2.23361835e-01 4.37125653e-01
6.51705265e-01 -1.28553057e+00 1.88255280e-01 6.38325393e-01
-1.37191102e-01 5.05474627e-01 2.39410460e-01 -5.10402262e-01
-1.18867612e+00 -1.10553944e+00 4.12971228e-01 7.28822425e-02
2.72133172e-01 2.74079815e-02 -7.57538736e-01 4.45987970e-01
1.53807119e-01 5.10771759e-02 6.52174532e-01 -2.72720277e-01
2.03643829e-01 -2.12547451e-01 -1.15410328e+00 6.19271576e-01
5.50028980e-01 -3.23276478e-03 -3.97626281e-01 4.86149132e-01
8.22975636e-01 -1.02440298e+00 -9.91569400e-01 5.36132216e-01
4.91646975e-01 -7.65733182e-01 9.65262592e-01 -4.42812294e-02
5.83590269e-01 -4.06648576e-01 2.36146271e-01 -1.28061342e+00
-3.23357910e-01 -5.63836135e-02 1.90168649e-01 6.28603995e-01
6.41337872e-01 -4.58507329e-01 8.48769665e-01 2.47751877e-01
-4.38964635e-01 -1.17547917e+00 -7.39376247e-01 -1.17016822e-01
-2.99548432e-02 4.62543406e-02 4.20661330e-01 7.59183526e-01
3.35954800e-02 -1.27583537e-02 -1.14942834e-01 3.92191797e-01
2.77315795e-01 2.77572185e-01 3.26630503e-01 -8.57366741e-01
-2.95393497e-01 -4.61570024e-01 -4.21541452e-01 -1.50483680e+00
-5.32726705e-01 -9.65263903e-01 -1.61690682e-01 -1.86667538e+00
2.96083778e-01 -4.01033431e-01 -2.87150890e-01 3.59624892e-01
-4.23183709e-01 2.99158134e-02 -2.34938622e-01 2.09945351e-01
-6.06619298e-01 3.99506956e-01 1.88777030e+00 -2.85720557e-01
-1.78144500e-01 2.96189010e-01 -9.28669155e-01 8.26148987e-01
8.41094434e-01 -3.58721614e-01 -2.26021409e-01 -6.75924957e-01
2.33828872e-01 1.12828650e-01 1.21653363e-01 -9.53225076e-01
3.84125918e-01 4.15203184e-01 5.74720562e-01 -6.36383176e-01
1.29340261e-01 -9.91109490e-01 -1.46936104e-01 1.09167206e+00
-1.21746443e-01 -1.23567842e-01 1.28610194e-01 5.18131077e-01
-5.46317399e-01 -3.98757607e-01 8.68299305e-01 -3.10831428e-01
-5.55190504e-01 3.99117649e-01 -5.85739277e-02 -1.23458177e-01
1.25000918e+00 -7.64148235e-02 -4.28120000e-03 2.28124857e-03
-7.19828010e-01 5.77580392e-01 2.60851592e-01 1.78226054e-01
7.13958740e-01 -1.07938409e+00 -5.60025513e-01 3.36259604e-01
-3.93393189e-02 5.85168302e-01 5.30018687e-01 1.41912448e+00
-9.89220142e-01 6.47930562e-01 2.08812147e-01 -6.74342215e-01
-1.18365169e+00 4.44244772e-01 6.73094511e-01 -5.67450166e-01
-9.22856212e-01 1.04664099e+00 4.86102521e-01 -2.72212982e-01
3.24200094e-01 -8.66611242e-01 -5.04970491e-01 -3.54430713e-02
5.84420323e-01 5.14441952e-02 7.83987194e-02 -2.79677302e-01
-1.18380971e-01 7.18879879e-01 -4.86047059e-01 5.63020170e-01
1.07788050e+00 -1.42495245e-01 -7.14971349e-02 -1.20966613e-01
1.19822085e+00 -7.70457834e-02 -9.94876266e-01 -1.02213122e-01
-4.00465310e-01 -3.68479282e-01 2.21642852e-01 -8.00501227e-01
-1.43701780e+00 1.04335940e+00 9.74335909e-01 1.83427915e-01
1.14317739e+00 -2.49970555e-01 1.26462185e+00 7.14117512e-02
9.90415961e-02 -7.04037368e-01 -1.01718232e-01 1.92307830e-01
5.44041455e-01 -1.47151470e+00 -7.71063194e-02 -6.39341176e-01
-6.89649940e-01 1.14803970e+00 5.25486052e-01 -2.15528309e-01
7.08827078e-01 7.56297112e-02 1.93004608e-01 -2.95261797e-02
-9.20221508e-02 1.28430188e-01 3.16807002e-01 1.89307109e-01
6.48704767e-01 1.19557366e-01 -6.69593036e-01 7.64963746e-01
-6.39721006e-02 -9.36912596e-02 2.30316371e-01 5.22681177e-01
-5.61365306e-01 -1.17220199e+00 -6.60006106e-02 4.59157497e-01
-9.00315702e-01 -1.86880797e-01 1.46775052e-01 7.68004656e-01
1.79681152e-01 6.72081351e-01 -2.10413888e-01 -1.85234040e-01
3.04251671e-01 -5.67585528e-01 2.33273670e-01 -5.58267772e-01
-5.77136874e-01 4.41588908e-01 -3.25391173e-01 -4.10743713e-01
-2.12083682e-01 -3.22263986e-01 -1.62496078e+00 1.91016465e-01
-5.47362030e-01 3.55195612e-01 6.13387465e-01 6.86148524e-01
1.48956284e-01 9.11174059e-01 5.44333756e-01 -3.35387826e-01
-6.35499835e-01 -7.52351284e-01 -4.35794830e-01 2.22736254e-01
2.20650509e-01 -2.30806231e-01 -1.82792008e-01 -2.39347845e-01] | [14.546889305114746, -2.720065116882324] |
cf2398c6-278f-4863-aa84-41c456573cb3 | continually-learning-from-existing-models | 2212.09097 | null | https://arxiv.org/abs/2212.09097v2 | https://arxiv.org/pdf/2212.09097v2.pdf | Continual Knowledge Distillation for Neural Machine Translation | While many parallel corpora are not publicly accessible for data copyright, data privacy and competitive differentiation reasons, trained translation models are increasingly available on open platforms. In this work, we propose a method called continual knowledge distillation to take advantage of existing translation models to improve one model of interest. The basic idea is to sequentially transfer knowledge from each trained model to the distilled model. Extensive experiments on Chinese-English and German-English datasets show that our method achieves significant and consistent improvements over strong baselines under both homogeneous and heterogeneous trained model settings and is robust to malicious models. | ['Yang Liu', 'Maosong Sun', 'Peng Li', 'Yuanchi Zhang'] | 2022-12-18 | null | null | null | null | ['nmt'] | ['computer-code'] | [-3.92178148e-02 -3.02130561e-02 -1.03261220e+00 -1.73785076e-01
-1.43600774e+00 -8.99045587e-01 9.43281651e-01 -3.23669076e-01
-4.44756925e-01 1.12328064e+00 2.91571051e-01 -6.65678620e-01
5.21352410e-01 -3.98007900e-01 -1.02801740e+00 -4.77944940e-01
5.46746075e-01 9.45075750e-01 1.53232381e-01 -1.74062744e-01
-3.10864989e-02 3.24405670e-01 -4.62911278e-01 4.19784725e-01
1.23640454e+00 1.87459096e-01 -3.39712828e-01 3.07117730e-01
8.52694288e-02 6.70582831e-01 -2.28797525e-01 -1.31665564e+00
6.50714099e-01 -4.00867254e-01 -1.11510181e+00 -7.17905611e-02
3.30175191e-01 -3.44516844e-01 -3.34368736e-01 1.17079425e+00
6.66896939e-01 -2.13492215e-01 5.75861335e-01 -8.99129808e-01
-1.14391875e+00 9.97440875e-01 -5.41582227e-01 2.73877889e-01
-1.17897347e-01 1.49016842e-01 1.12554514e+00 -1.30673218e+00
9.81860578e-01 9.94623661e-01 5.19859970e-01 5.86386323e-01
-1.31671298e+00 -7.02785075e-01 -5.49305929e-03 2.94404954e-01
-1.53328860e+00 -5.58306992e-01 3.84356201e-01 -2.16417968e-01
1.21769547e+00 6.67853579e-02 2.36256897e-01 1.61102939e+00
1.86979517e-01 1.05694127e+00 1.28580058e+00 -5.57640254e-01
-1.45107746e-01 7.73109555e-01 3.97472866e-02 4.76153105e-01
2.60961026e-01 1.37751505e-01 -6.09196305e-01 -5.88206351e-01
4.00674373e-01 -2.31901377e-01 -2.39668280e-01 -3.08396965e-01
-1.25466883e+00 7.36534536e-01 8.96784514e-02 3.15242559e-01
-3.86044323e-01 -8.02693963e-02 4.23095167e-01 4.96365845e-01
9.63146865e-01 4.19446170e-01 -7.85254776e-01 -3.35108966e-01
-9.52784061e-01 2.16723800e-01 9.26829755e-01 1.18453455e+00
8.43421042e-01 -2.54525483e-01 -2.16684192e-02 8.89342785e-01
1.20651245e-01 8.23224604e-01 6.53531194e-01 -6.08381629e-01
9.88828480e-01 3.43789518e-01 7.35591501e-02 -5.45362771e-01
2.17637271e-01 -5.64739227e-01 -5.97766995e-01 -5.95397294e-01
-7.38829076e-02 -1.66030973e-01 -9.15109038e-01 1.38340414e+00
3.13820422e-01 3.46909523e-01 2.97494382e-01 5.61867416e-01
4.38188285e-01 6.97510660e-01 -3.32239904e-02 -2.05535322e-01
8.79734755e-01 -1.53998828e+00 -6.94151640e-01 -3.52751434e-01
7.49490857e-01 -1.09165728e+00 8.68644059e-01 5.96023351e-02
-1.21180725e+00 -1.80643618e-01 -8.42641592e-01 -1.07951447e-01
-3.64094019e-01 1.17024761e-02 3.71343374e-01 5.85589468e-01
-1.05850947e+00 4.90221262e-01 -1.07123256e+00 -4.14787292e-01
5.28352916e-01 3.52728754e-01 -4.60055858e-01 -3.72953624e-01
-1.35731924e+00 1.41083944e+00 3.10141534e-01 -3.36362645e-02
-9.01230693e-01 -5.72059155e-01 -5.39147973e-01 -2.24834055e-01
2.65240848e-01 -7.20297217e-01 1.44760966e+00 -1.04228795e+00
-1.71577454e+00 8.18436086e-01 -2.84091741e-01 -4.29647505e-01
8.27572942e-01 -4.71524924e-01 -4.39323545e-01 -1.22213468e-01
2.06349324e-02 3.87466758e-01 8.22353363e-01 -1.33885443e+00
-4.56805259e-01 -1.79996133e-01 -1.31884590e-01 4.18833911e-01
-5.43304563e-01 3.49513143e-01 -9.60261941e-01 -5.50255537e-01
-1.60829335e-01 -1.24015498e+00 -2.94741124e-01 -5.45394838e-01
-6.56951666e-01 5.45650609e-02 6.86628520e-01 -1.02591681e+00
1.13632512e+00 -1.71497548e+00 3.33358884e-01 2.49263823e-01
-1.20895959e-01 6.28114223e-01 -2.91915953e-01 6.69988275e-01
2.39060387e-01 2.40949735e-01 -2.39109561e-01 -5.35887241e-01
3.90893668e-02 3.51081252e-01 -6.52072668e-01 4.15693492e-01
2.57885128e-01 1.30050659e+00 -8.24534476e-01 -5.71282685e-01
-1.75614551e-01 6.07356787e-01 -6.09650135e-01 7.97170103e-02
-2.23770618e-01 3.95761251e-01 -5.48047185e-01 7.61442602e-01
6.40481174e-01 -2.97259420e-01 4.02864456e-01 1.63751334e-01
2.04938114e-01 5.76146364e-01 -6.47980988e-01 1.59328854e+00
-2.89614379e-01 5.10859013e-01 -2.24635452e-01 -5.16694486e-01
7.99895167e-01 5.75200975e-01 2.86687165e-02 -4.64810371e-01
4.15027589e-02 5.81800997e-01 -6.61533326e-02 -2.69400775e-01
7.53671110e-01 -3.55514258e-01 2.32582822e-01 5.04860282e-01
1.20513260e-01 9.40279290e-03 3.00352294e-02 3.58115911e-01
8.89533758e-01 3.65910172e-01 1.22302614e-01 9.73550428e-04
3.67938280e-01 2.37657294e-01 6.69593930e-01 4.77420509e-01
-3.36326957e-01 4.05162692e-01 1.30833492e-01 -2.66847014e-01
-1.14364779e+00 -1.07432902e+00 8.96641463e-02 1.08082712e+00
-1.18111081e-01 -5.73811054e-01 -7.52949297e-01 -1.02176416e+00
-1.02473572e-01 6.87520981e-01 -4.09444749e-01 -1.73682943e-01
-7.99527526e-01 -8.86017084e-01 7.39459813e-01 5.35297155e-01
5.10564029e-01 -6.35187984e-01 5.19367754e-01 1.76821843e-01
-5.48763037e-01 -1.28449488e+00 -6.39511049e-01 -1.59695432e-01
-9.32289124e-01 -5.93804181e-01 -7.29784608e-01 -8.53093624e-01
5.86769044e-01 2.93301344e-01 1.22537971e+00 -1.86155915e-01
3.69016379e-01 1.19157232e-01 -3.03545356e-01 -3.10058415e-01
-7.75852621e-01 6.53040409e-01 2.00079337e-01 3.30971740e-02
7.59939373e-01 -3.22851866e-01 -2.10533828e-01 3.64142120e-01
-6.70984745e-01 1.56405360e-01 7.84432888e-01 9.16402996e-01
7.20915973e-01 -5.21266520e-01 4.74312812e-01 -1.16294849e+00
7.59288311e-01 -5.07263303e-01 -3.40373069e-01 6.90400898e-01
-9.84006405e-01 -3.09431888e-02 5.21014094e-01 -3.23922664e-01
-1.13249838e+00 -1.57366484e-01 4.40317020e-02 -5.98445475e-01
2.25340560e-01 7.47527361e-01 -1.19454719e-01 -8.34472775e-02
6.34281933e-01 5.18719792e-01 -3.58397424e-01 -5.41450143e-01
6.77172065e-01 9.28402841e-01 2.74527341e-01 -7.37946093e-01
1.14840329e+00 3.04562151e-01 -8.50305498e-01 -2.57944852e-01
-6.35532796e-01 -2.95077860e-01 -7.88519800e-01 2.78065115e-01
6.13351762e-01 -1.25564229e+00 3.03648591e-01 4.10903156e-01
-1.21116269e+00 -2.75009274e-01 1.69981476e-02 6.05926037e-01
-3.68069470e-01 3.26498121e-01 -1.05922401e+00 -7.98884779e-02
-6.39469385e-01 -1.38069904e+00 7.52974927e-01 -2.79683340e-02
-8.58891755e-02 -1.19909513e+00 6.05486929e-01 8.03753018e-01
5.73358953e-01 -4.55369622e-01 7.97362387e-01 -9.94836926e-01
-7.04743028e-01 -4.65663552e-01 8.03683978e-03 5.52644134e-01
1.70554429e-01 2.61446774e-01 -7.91784942e-01 -4.51213062e-01
-2.13327155e-01 -6.16454959e-01 6.80354834e-01 -1.89417079e-01
5.18712580e-01 -5.81170022e-01 -5.36763430e-01 4.51331586e-01
1.13229227e+00 -2.89387703e-01 6.13921225e-01 4.30619121e-01
6.71379268e-01 1.80425584e-01 3.76737475e-01 -5.96820675e-02
5.74602723e-01 6.87430024e-01 -1.92768991e-01 -1.26174986e-01
-5.04900850e-02 -4.95059848e-01 7.61791408e-01 1.62075472e+00
-1.86161399e-01 -7.87134692e-02 -1.01517177e+00 6.78051651e-01
-1.92550516e+00 -9.06380415e-01 1.15853816e-01 2.02497745e+00
1.38698077e+00 7.11952895e-02 -1.36607423e-01 -9.31076109e-01
9.20322061e-01 -2.92621460e-02 -4.91569072e-01 -5.50521731e-01
-4.31466520e-01 2.16761097e-01 6.36835873e-01 6.12495661e-01
-8.05496573e-01 1.54284930e+00 7.51966906e+00 9.13597763e-01
-9.96305406e-01 6.09828949e-01 5.42694807e-01 -2.08515823e-01
-4.90169585e-01 2.13803187e-01 -9.55227137e-01 1.89962655e-01
1.16181791e+00 -6.69361711e-01 4.14094329e-01 8.39476645e-01
-9.77080613e-02 6.01731598e-01 -9.84135985e-01 6.01670742e-01
1.48394197e-01 -1.42274201e+00 3.15236747e-01 2.23099604e-01
1.40055203e+00 7.80259371e-01 1.66315600e-01 6.12116992e-01
8.88561308e-01 -7.41406143e-01 4.01337445e-01 1.58931967e-02
5.74610233e-01 -6.46138310e-01 6.33412778e-01 4.43628430e-01
-5.78288138e-01 4.40231115e-01 -4.91633922e-01 2.25310266e-01
7.62826800e-02 1.97223946e-01 -1.10348260e+00 8.66394937e-01
3.54643464e-01 8.70092213e-01 -4.93270218e-01 6.17759347e-01
-5.41970670e-01 1.11795545e+00 -6.98125139e-02 3.12288553e-01
3.60848755e-01 -2.23113954e-01 2.70297974e-01 1.40510273e+00
1.68223426e-01 -1.69446558e-01 1.62375256e-01 7.40656078e-01
-7.44818270e-01 4.43157524e-01 -6.34197056e-01 -3.11837018e-01
5.22503376e-01 1.18631542e+00 -1.97525233e-01 -6.84342921e-01
-8.12368989e-01 1.33526003e+00 7.23844528e-01 5.67468047e-01
-8.96567106e-01 -1.50086191e-02 7.17935681e-01 -2.23004639e-01
3.68611783e-01 -1.71394989e-01 -5.73816635e-02 -1.79391325e+00
1.42437503e-01 -1.31971478e+00 3.24574918e-01 -4.19849664e-01
-1.49488926e+00 7.92771220e-01 -2.32797101e-01 -1.31474149e+00
-3.65740746e-01 -4.28811371e-01 -4.93771493e-01 1.10079181e+00
-1.75723612e+00 -1.59283292e+00 5.93223751e-01 6.94355905e-01
4.67455029e-01 -4.09997851e-01 1.02351940e+00 3.98739576e-01
-8.45891297e-01 9.45942342e-01 7.80563891e-01 3.63893062e-01
1.30982447e+00 -8.55259359e-01 7.58593142e-01 1.13485575e+00
4.05861646e-01 9.59848046e-01 5.47819316e-01 -8.32262576e-01
-1.58123624e+00 -1.15172470e+00 1.54887164e+00 -8.09851825e-01
1.03077352e+00 -3.52828771e-01 -9.61416364e-01 1.21942639e+00
7.34233737e-01 -4.94599761e-03 1.03431225e+00 2.62478381e-01
-9.55971181e-01 1.42915457e-01 -8.51065814e-01 6.84064269e-01
6.93478882e-01 -7.67662942e-01 -8.07529807e-01 6.65120065e-01
7.06824064e-01 -5.08673608e-01 -1.00560164e+00 2.82275259e-01
3.00601423e-01 -3.94472063e-01 7.07941413e-01 -1.14769959e+00
6.76491499e-01 -1.98064912e-02 -2.62756944e-01 -1.58183289e+00
-3.14661086e-01 -8.96679580e-01 -4.24649805e-01 1.19826484e+00
1.06717527e+00 -8.50924969e-01 6.87239707e-01 8.65271866e-01
2.42018085e-02 -6.16728187e-01 -9.69615936e-01 -1.05411232e+00
8.38891387e-01 -2.56344587e-01 5.44741511e-01 1.48398221e+00
4.19098020e-01 7.05087841e-01 -7.47274756e-01 1.66258126e-01
5.21012962e-01 2.68388450e-01 1.00896502e+00 -5.70477843e-01
-6.84907198e-01 -2.99915642e-01 -2.53358502e-02 -1.09390867e+00
4.73478377e-01 -1.33660674e+00 -2.18116626e-01 -1.47631526e+00
7.81935930e-01 -2.89653838e-01 -4.34145361e-01 5.95540822e-01
-6.25596881e-01 3.92237604e-01 1.03723191e-01 7.08369553e-01
-5.85443616e-01 6.61759079e-01 1.13260496e+00 -3.99709642e-01
6.30119815e-03 -1.48895204e-01 -7.86622882e-01 3.05194557e-01
7.59615779e-01 -6.08028173e-01 -2.65125304e-01 -1.07038498e+00
2.12917104e-01 -3.44638467e-01 -1.64942548e-01 -3.45844388e-01
2.91792512e-01 -2.47527018e-01 -3.19085196e-02 -2.65248984e-01
9.62422118e-02 -6.22421622e-01 1.11293100e-01 3.59272003e-01
-4.79642153e-01 2.39285171e-01 2.61738271e-01 5.03743112e-01
-2.75618345e-01 3.21626849e-02 5.34434736e-01 -3.94770131e-02
-4.08365250e-01 5.03273785e-01 -6.49455711e-02 1.61506325e-01
8.43948781e-01 3.70130122e-01 -6.84849024e-01 -3.02311212e-01
-4.29820985e-01 1.80586159e-01 6.67622089e-01 7.42896259e-01
2.60616243e-01 -1.56963456e+00 -1.27806258e+00 2.82492712e-02
4.20206815e-01 -4.80029196e-01 -1.34305567e-01 9.71625745e-01
-3.01272154e-01 5.70536256e-01 8.94733295e-02 -2.91451722e-01
-1.15348530e+00 5.56052506e-01 2.54046172e-01 -5.82981884e-01
-4.16184634e-01 7.20616400e-01 -1.21591486e-01 -9.78124559e-01
-1.07585624e-01 1.09822325e-01 4.76620704e-01 -5.11933684e-01
7.00781047e-01 1.89858541e-01 3.18155497e-01 -8.81326616e-01
-2.75600195e-01 2.33235016e-01 -8.55449617e-01 -4.19827014e-01
1.06204700e+00 -1.48323819e-01 -2.79738545e-01 2.41782710e-01
1.25868642e+00 1.15668125e-01 -9.54262376e-01 -9.11703944e-01
5.60957640e-02 -4.96266246e-01 -1.79475248e-01 -1.09677231e+00
-9.50554311e-01 8.57841372e-01 1.20950192e-01 -4.62169915e-01
7.26033330e-01 8.51920154e-03 9.19962585e-01 6.98341370e-01
6.32107437e-01 -1.23744273e+00 -4.00172979e-01 7.78448045e-01
4.81850445e-01 -1.37428367e+00 -7.47680441e-02 -3.06466013e-01
-8.75492215e-01 7.20001101e-01 6.00381970e-01 2.55892754e-01
3.97337794e-01 1.99956909e-01 5.66680312e-01 2.62825608e-01
-1.05579126e+00 8.62854794e-02 1.44555509e-01 4.42996085e-01
4.44755435e-01 2.00821817e-01 -4.52155203e-01 4.74139512e-01
3.26774865e-02 1.27480403e-01 3.20865899e-01 9.93501604e-01
-1.51908487e-01 -1.67956710e+00 -1.86096996e-01 1.97749794e-01
-7.63173461e-01 -5.69480181e-01 -8.48539412e-01 6.36995673e-01
-3.95233601e-01 8.25212836e-01 -4.84637529e-01 -5.77308834e-01
1.87706813e-01 4.56458986e-01 5.08457661e-01 -5.29591382e-01
-9.76519465e-01 1.20592423e-01 1.54714689e-01 -3.62253398e-01
-3.38656634e-01 -6.74992263e-01 -8.72825861e-01 -6.65703595e-01
-5.52656949e-01 3.85370642e-01 5.47085881e-01 9.63572264e-01
7.39082932e-01 -1.79868832e-01 5.87038040e-01 -4.81270164e-01
-1.00884020e+00 -9.70264494e-01 -1.39977455e-01 4.11688775e-01
-1.27469629e-01 -9.99423862e-02 -1.28483012e-01 3.09839070e-01] | [11.605687141418457, 10.30553150177002] |
9b21fe74-eb9f-4eed-b1e0-f75f8de37b66 | roof-material-classification-from-aerial | 2004.11482 | null | https://arxiv.org/abs/2004.11482v1 | https://arxiv.org/pdf/2004.11482v1.pdf | Roof material classification from aerial imagery | This paper describes an algorithm for classification of roof materials using aerial photographs. Main advantages of the algorithm are proposed methods to improve prediction accuracy. Proposed methods includes: method of converting ImageNet weights of neural networks for using multi-channel images; special set of features of second level models that are used in addition to specific predictions of neural networks; special set of image augmentations that improve training accuracy. In addition, complete flow for solving this problem is proposed. The following content is available in open access: solution code, weight sets and architecture of the used neural networks. The proposed solution achieved second place in the competition "Open AI Caribbean Challenge". | ['Roman Solovyev'] | 2020-04-23 | null | null | null | null | ['material-classification'] | ['computer-vision'] | [ 2.02686340e-01 -3.72700617e-02 2.06970394e-01 -5.36721289e-01
-3.50991115e-02 -2.65817136e-01 2.97475517e-01 -4.36295599e-01
-2.99276561e-01 8.21523309e-01 -6.85596466e-03 5.28209917e-02
-2.35408381e-01 -1.01299202e+00 -8.26800108e-01 -4.63244796e-01
-3.59352082e-01 3.05720747e-01 2.68378198e-01 -6.62545621e-01
4.37544674e-01 9.07368720e-01 -2.05168343e+00 7.70404518e-01
6.92592621e-01 1.07200670e+00 4.84614968e-01 8.21852684e-01
6.02073893e-02 8.04781437e-01 -4.47495729e-01 -9.73049272e-03
8.36685181e-01 1.23367690e-01 -1.20724416e+00 6.07959256e-02
9.76901531e-01 -4.26336825e-01 -2.03869894e-01 9.59251881e-01
3.69705021e-01 3.55119348e-01 1.10177112e+00 -1.19093788e+00
-7.40464449e-01 6.84537351e-01 -2.72148907e-01 4.76458400e-01
9.23464373e-02 -1.89750180e-01 7.77474344e-01 -6.54900908e-01
4.13880289e-01 7.95789063e-01 1.00046670e+00 3.02854270e-01
-9.96298909e-01 -5.46900570e-01 -4.53212149e-02 4.47167963e-01
-1.54924476e+00 -1.12103738e-01 3.58916759e-01 -6.38382316e-01
1.30842876e+00 2.90312827e-01 1.02133465e+00 6.10091567e-01
-3.19648534e-02 5.93064487e-01 1.14418614e+00 -5.26013792e-01
-6.69518709e-02 4.74856496e-01 3.80393207e-01 9.73765194e-01
-8.07855055e-02 1.56077087e-01 -1.97436363e-02 1.53756291e-01
8.72896552e-01 -4.12001669e-01 -8.43423679e-02 1.75895363e-01
-6.40048206e-01 6.94319844e-01 9.36710179e-01 4.03632641e-01
-4.83211100e-01 9.12489817e-02 2.39126578e-01 3.01421285e-01
3.02689731e-01 5.78935504e-01 -6.89286232e-01 6.45097196e-01
-8.79919529e-01 4.96235937e-01 4.35603559e-01 1.00458980e+00
8.35625887e-01 4.73410457e-01 8.26530159e-02 1.13639081e+00
2.15862334e-01 2.99616933e-01 5.53744972e-01 -1.06646395e+00
3.00614387e-01 7.10831642e-01 -8.49172473e-02 -9.45252955e-01
-7.34361053e-01 -3.49564999e-01 -7.49462724e-01 5.29913008e-01
7.94184133e-02 -2.08788037e-01 -1.24945319e+00 1.21499538e+00
4.08680784e-03 3.43729973e-01 2.75873572e-01 7.81002223e-01
1.18828881e+00 1.16746855e+00 -1.58529118e-01 3.23978961e-01
1.15455270e+00 -1.28590024e+00 -3.14063340e-01 -1.06780507e-01
3.34342420e-01 -7.97063410e-01 5.62449634e-01 4.11987722e-01
-1.03688896e+00 -9.02630746e-01 -1.12830853e+00 2.98850507e-01
-1.22797227e+00 6.28558934e-01 7.50318527e-01 7.06959009e-01
-1.30181825e+00 1.04091144e+00 -1.78497836e-01 -5.43131411e-01
4.00238484e-01 1.10067844e+00 -3.47520947e-01 3.73843431e-01
-9.75205898e-01 1.26634681e+00 9.85946417e-01 2.71862924e-01
-5.81816912e-01 -6.39583170e-01 -6.81257963e-01 5.42435097e-03
-2.18566313e-01 -4.11330640e-01 1.04672647e+00 -1.69429135e+00
-1.43724155e+00 1.01757514e+00 4.81238574e-01 -8.04556608e-01
9.43410322e-02 -2.22921073e-01 -3.80458385e-01 7.54995197e-02
-1.39477506e-01 1.35270810e+00 7.56035686e-01 -1.31103170e+00
-8.35829139e-01 -1.46051407e-01 2.18228593e-01 3.32187653e-01
-4.95683759e-01 -2.00983435e-01 1.79369911e-03 -7.12404430e-01
7.38912523e-02 -8.81247401e-01 -3.79043758e-01 -2.43831038e-01
-1.47675350e-01 2.44764350e-02 1.00563967e+00 -9.12624776e-01
8.02635789e-01 -1.64157581e+00 -6.15225397e-02 3.91319185e-01
-4.12171632e-01 5.46013176e-01 -1.85710490e-01 2.13532224e-01
-4.56313938e-01 7.98603240e-03 -3.04644972e-01 4.10976976e-01
-3.41684729e-01 1.60904557e-01 -2.16843143e-01 2.88062468e-02
1.32800058e-01 4.20980752e-01 -7.91412368e-02 -4.38624769e-01
7.66920149e-01 3.44216883e-01 -3.80284101e-01 -4.95227687e-02
-2.10583732e-02 1.31590039e-01 -1.20949157e-01 4.42917377e-01
9.35378313e-01 2.92606235e-01 -3.12810034e-01 -7.46746957e-01
-4.39738452e-01 -2.23843411e-01 -1.71571636e+00 1.35577142e+00
-3.89968336e-01 6.95628226e-01 -6.54849485e-02 -1.15116465e+00
1.10025549e+00 3.82888764e-01 3.65664423e-01 -6.25224635e-02
3.43426049e-01 2.85059847e-02 -3.45397480e-02 -7.32666671e-01
5.92760563e-01 1.66168436e-01 6.59446299e-01 -1.77910835e-01
6.55999362e-01 -3.86712044e-01 4.75861132e-01 -2.60858327e-01
4.75874037e-01 7.10713685e-01 1.40476838e-01 -6.00224316e-01
9.21432018e-01 3.96915704e-01 3.28051478e-01 4.54884380e-01
1.19671628e-01 5.15370965e-01 -3.87983650e-01 -1.14905047e+00
-1.27169716e+00 -5.40023923e-01 -4.02700096e-01 8.46772254e-01
-1.92634732e-01 -2.53577530e-01 -8.15056741e-01 -2.52229869e-01
-1.56716764e-01 3.72938633e-01 -8.09792280e-01 3.26801360e-01
-4.38025594e-01 -1.00644207e+00 6.53898418e-01 6.29253447e-01
1.13057017e+00 -1.11161971e+00 -6.45491183e-01 1.09035358e-01
2.41377912e-02 -1.10894275e+00 3.99456859e-01 3.46796900e-01
-1.29537892e+00 -1.14409745e+00 -6.06173694e-01 -1.20614231e+00
6.40412807e-01 8.81535187e-02 9.09606993e-01 5.01881540e-01
-7.68869817e-01 2.00590923e-01 -3.54275465e-01 -6.43923819e-01
9.50887725e-02 1.80391610e-01 -1.89887196e-01 -1.69661090e-01
1.89054698e-01 -6.54630840e-01 -3.42120230e-01 -6.05174117e-02
-5.78285694e-01 3.45488966e-01 6.79331124e-01 7.59567618e-01
1.83420733e-01 2.50492185e-01 1.71307057e-01 -7.66667306e-01
4.26184922e-01 -2.03224972e-01 -1.06134582e+00 2.72807419e-01
-2.90879399e-01 -6.55676648e-02 7.39326179e-01 1.05922565e-01
-1.11921072e+00 6.09631300e-01 -4.92577583e-01 -1.27280029e-02
-7.14532614e-01 2.02090472e-01 4.73096222e-01 -5.58422625e-01
7.36438811e-01 -3.03115845e-02 -3.94476146e-01 -4.80956078e-01
9.90987197e-02 7.32744098e-01 3.01341683e-01 -5.70248783e-01
6.93798840e-01 2.85335302e-01 3.12658817e-01 -1.25099146e+00
-8.65361154e-01 -3.72591197e-01 -1.25796354e+00 -5.50441802e-01
9.32241857e-01 -9.12172318e-01 -4.55774367e-01 5.72144210e-01
-1.00387406e+00 -2.21347988e-01 -1.55159265e-01 4.25040424e-01
-4.55902845e-01 -2.01596916e-01 -5.87671697e-01 -8.37884247e-01
-5.34791708e-01 -1.11084974e+00 5.56482315e-01 6.10723317e-01
3.40374649e-01 -9.69309926e-01 5.85389249e-02 4.51519787e-01
4.34855908e-01 2.42106095e-01 8.01016688e-01 -3.26621622e-01
-3.28423947e-01 -1.27171040e-01 -1.27455145e-01 9.23859715e-01
-2.60124326e-01 5.10114491e-01 -1.33009005e+00 1.26564071e-01
-6.04306638e-01 -6.37498260e-01 1.04326499e+00 8.37690234e-01
1.26720905e+00 -1.77392662e-01 -3.38383317e-01 7.19698489e-01
2.01441264e+00 3.30578387e-01 8.12693179e-01 6.17757559e-01
4.52300578e-01 7.97817230e-01 2.60232359e-01 1.51352435e-01
-2.31461450e-02 5.61482847e-01 5.54455459e-01 -2.63032705e-01
-2.58529149e-02 2.56259263e-01 4.61290553e-02 7.65985608e-01
-9.91640866e-01 2.16486305e-01 -1.13164043e+00 4.01316822e-01
-1.74052489e+00 -1.14687312e+00 -6.76638961e-01 1.68065536e+00
8.50558728e-02 3.70339155e-02 6.90351948e-02 3.66281241e-01
7.89215147e-01 -9.67318863e-02 5.70383221e-02 -7.73427367e-01
-3.74941081e-01 6.17762506e-01 8.85748148e-01 6.46858931e-01
-1.63169229e+00 1.23799980e+00 7.40316439e+00 6.27192795e-01
-1.02997708e+00 -2.78568715e-02 3.46863568e-01 3.71776760e-01
5.49465418e-01 3.01461518e-02 -9.07691896e-01 1.79898798e-01
9.19557095e-01 3.43142182e-01 5.55865467e-01 1.07887113e+00
8.21920261e-02 -2.60673761e-01 -6.13669693e-01 8.48857462e-01
2.35829383e-01 -1.86199594e+00 3.36836338e-01 -1.65630415e-01
9.64377463e-01 3.68311971e-01 -2.18986273e-01 2.16615647e-01
3.38032812e-01 -9.67356563e-01 6.02238595e-01 4.10516381e-01
4.49474514e-01 -8.14957201e-01 9.38401103e-01 3.63870412e-02
-1.42989612e+00 -5.67954063e-01 -8.61098707e-01 -4.99568999e-01
-3.52749497e-01 -1.06326789e-01 -6.61563337e-01 7.44619489e-01
1.20193291e+00 8.22908580e-01 -8.78391504e-01 1.43316603e+00
-1.91827655e-01 4.70367968e-01 -5.33972442e-01 2.42612049e-01
5.43945789e-01 -4.94752854e-01 1.17109150e-01 1.23001111e+00
4.74355191e-01 3.74072552e-01 1.84517696e-01 4.51335639e-01
5.87207004e-02 3.90859932e-01 -1.13557804e+00 4.42717850e-01
3.30457896e-01 1.42190516e+00 -6.96190476e-01 -5.20451963e-01
-4.21748161e-01 8.29627573e-01 1.93018883e-01 2.60548783e-03
-5.16960502e-01 -3.42744261e-01 2.63336480e-01 -2.02484932e-02
2.50440240e-01 2.16154858e-01 -2.56039172e-01 -9.28906322e-01
-5.09360433e-01 -3.84533525e-01 4.66486275e-01 -1.45057797e+00
-9.65802252e-01 9.09250259e-01 2.15325490e-01 -1.20821369e+00
4.48345780e-01 -1.24227130e+00 -9.15905714e-01 6.31545067e-01
-1.21605921e+00 -1.71311688e+00 -6.56900704e-01 6.54334128e-01
9.51770067e-01 -9.30535555e-01 1.28631163e+00 4.21028644e-01
-4.29188818e-01 -4.96639572e-02 3.19621265e-01 2.16941461e-01
1.86303675e-01 -1.11647618e+00 -5.22044441e-03 6.43256962e-01
4.66975905e-02 -1.63314007e-02 5.70625067e-01 -3.57195467e-01
-6.32815182e-01 -1.19063258e+00 3.90320003e-01 -2.39454545e-02
2.97161996e-01 1.32675394e-01 -4.81386900e-01 1.01249564e+00
7.16206491e-01 -2.84040809e-01 7.88797259e-01 -3.96681838e-02
1.50775611e-01 -3.59683216e-01 -1.22863245e+00 2.13019177e-01
5.05094767e-01 1.17121667e-01 -8.08489978e-01 8.65130723e-01
5.08094467e-02 -3.08147937e-01 -8.87332499e-01 4.20750767e-01
5.21678627e-01 -9.49345887e-01 1.14352274e+00 -6.00424886e-01
6.77901983e-01 -3.12921613e-01 -5.37292004e-01 -1.16551888e+00
-5.93627691e-01 -6.97513297e-02 3.17927301e-01 1.02801692e+00
6.75670385e-01 -2.57727325e-01 8.28848362e-01 6.54173270e-03
-4.28966552e-01 -6.84060156e-01 -6.46138668e-01 -3.84879977e-01
3.52749527e-02 -2.55567819e-01 3.10343802e-01 9.82799888e-01
-5.36919534e-01 3.38119507e-01 -5.56804001e-01 4.52698290e-01
5.06646574e-01 -1.50172308e-01 4.93991792e-01 -1.50460279e+00
5.33313937e-02 -2.80565113e-01 -7.27387309e-01 -3.63233656e-01
5.59958816e-02 -9.06524777e-01 -3.77531201e-01 -1.79812849e+00
2.56894380e-02 -4.67737049e-01 -2.05027103e-01 6.65608585e-01
4.75189149e-01 7.76164711e-01 4.94524747e-01 -9.84669477e-03
9.84304249e-02 3.10794652e-01 7.69536018e-01 -2.90634364e-01
-9.87505689e-02 1.39679775e-01 -8.12561214e-02 1.22584617e+00
1.15677440e+00 -2.78031170e-01 -3.27381551e-01 -7.17940450e-01
-7.21812025e-02 -3.38409513e-01 5.48543513e-01 -1.84895611e+00
4.68461886e-02 -1.39098674e-01 8.32908452e-01 -1.01188922e+00
8.11412990e-01 -1.29331481e+00 1.76457703e-01 5.43321609e-01
-2.83539414e-01 2.63290733e-01 5.60884535e-01 3.11215618e-03
-2.02060178e-01 -9.50973988e-01 1.20370626e+00 -6.64737582e-01
-1.06993902e+00 -1.46497890e-01 -8.38773966e-01 -6.27696335e-01
1.30731320e+00 -6.97889745e-01 -9.89794806e-02 2.57662800e-03
-1.24354196e+00 1.10936768e-01 -3.54556963e-02 3.70792449e-01
5.98972440e-01 -1.28144443e+00 -6.00445867e-01 1.04431547e-01
-3.68217945e-01 -3.45187664e-01 1.80010810e-01 5.66918850e-01
-1.22463298e+00 4.23604190e-01 -1.39954364e+00 -1.05911553e-01
-1.60274911e+00 8.92574042e-02 9.01071966e-01 -7.75994137e-02
-4.96232837e-01 9.46524143e-01 -2.44381949e-01 -6.95438862e-01
2.97196746e-01 1.33656207e-02 -9.33459759e-01 7.98971206e-02
2.74469286e-01 4.92876858e-01 2.01437756e-01 -8.22023511e-01
-1.44092053e-01 8.02071452e-01 9.68691520e-03 1.23076998e-01
2.06659317e+00 8.58490393e-02 -1.85914680e-01 4.17361110e-01
6.98515117e-01 -8.30522001e-01 -7.57243216e-01 1.27613127e-01
-1.07873395e-01 -3.58397782e-01 4.49072391e-01 -1.17082250e+00
-1.31631804e+00 1.02910089e+00 1.34092903e+00 -7.84840584e-02
1.28088689e+00 -8.31605554e-01 4.40007627e-01 9.62168694e-01
1.30668744e-01 -1.76730192e+00 -4.62460130e-01 7.21655369e-01
9.93534982e-01 -1.18988931e+00 3.34268183e-01 -4.88529027e-01
-5.20772994e-01 1.70834374e+00 1.10063100e+00 -7.66153991e-01
9.94322896e-01 3.66564989e-01 8.08895566e-04 -3.14114332e-01
-4.76586431e-01 -1.93940207e-01 2.79310644e-01 8.67692709e-01
4.71860766e-01 -5.39250411e-02 -2.08744988e-01 4.07575041e-01
-2.05403000e-01 -1.55745707e-02 6.23948336e-01 1.03688729e+00
-7.71381855e-01 -8.53432119e-01 -7.00072289e-01 7.70118535e-01
-3.69932741e-01 -3.05008858e-01 -5.27581155e-01 9.00277257e-01
7.92344987e-01 5.35136521e-01 -9.08908062e-03 -4.96099502e-01
4.47338849e-01 1.63074229e-02 5.54204047e-01 -4.89254981e-01
-1.06586981e+00 -3.02594334e-01 2.69090116e-01 -1.25650033e-01
-7.91669250e-01 -3.49740833e-01 -9.28385079e-01 -2.11410075e-01
-6.00636959e-01 2.27646038e-01 1.02076375e+00 8.93624604e-01
1.90905500e-02 6.41662836e-01 3.51143986e-01 -1.45918012e+00
1.36487901e-01 -1.21130538e+00 -5.59199870e-01 1.30809605e-01
-1.96309209e-01 -7.17180669e-01 1.13215327e-01 7.40834773e-01] | [9.460518836975098, -1.1133235692977905] |
aa867ba1-e6f5-49dc-aff7-89a057cb9680 | summarizing-community-based-question-answer | 2211.09892 | null | https://arxiv.org/abs/2211.09892v1 | https://arxiv.org/pdf/2211.09892v1.pdf | Summarizing Community-based Question-Answer Pairs | Community-based Question Answering (CQA), which allows users to acquire their desired information, has increasingly become an essential component of online services in various domains such as E-commerce, travel, and dining. However, an overwhelming number of CQA pairs makes it difficult for users without particular intent to find useful information spread over CQA pairs. To help users quickly digest the key information, we propose the novel CQA summarization task that aims to create a concise summary from CQA pairs. To this end, we first design a multi-stage data annotation process and create a benchmark dataset, CoQASUM, based on the Amazon QA corpus. We then compare a collection of extractive and abstractive summarization methods and establish a strong baseline approach DedupLED for the CQA summarization task. Our experiment further confirms two key challenges, sentence-type transfer and deduplication removal, towards the CQA summarization task. Our data and code are publicly available. | ['Xiaolan Wang', 'Yoshi Suhara', 'Ting-Yao Hsu'] | 2022-11-17 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 1.70130968e-01 3.08228973e-02 9.00395438e-02 -2.78989702e-01
-1.67109752e+00 -8.14833701e-01 4.94333059e-01 8.66896212e-01
-3.05917084e-01 8.02103639e-01 9.86447155e-01 -4.76405025e-02
-2.89154775e-03 -6.22544289e-01 -2.89540023e-01 -3.30060542e-01
2.41307393e-01 5.11666000e-01 1.69536144e-01 -7.12585449e-01
6.29430592e-01 -4.55391370e-02 -1.12217617e+00 5.99868357e-01
1.76551688e+00 7.86121845e-01 1.89500034e-01 6.65521085e-01
-4.89153683e-01 8.44369411e-01 -1.10635757e+00 -8.10894251e-01
-1.88022152e-01 -7.25199342e-01 -1.36630404e+00 -7.42294118e-02
6.53731823e-01 -2.47850761e-01 -2.34828502e-01 8.73444498e-01
7.60379732e-01 3.66056651e-01 5.80251455e-01 -9.85214174e-01
-6.41200662e-01 7.96817780e-01 -4.63640690e-01 2.70776987e-01
7.51413524e-01 3.51982974e-02 1.59455633e+00 -4.19501603e-01
5.47704220e-01 9.73557353e-01 5.29882133e-01 5.17774880e-01
-8.05128574e-01 -3.29788357e-01 -7.86459297e-02 1.19448707e-01
-9.96317625e-01 -5.99576950e-01 7.54685819e-01 1.23030715e-01
9.30679560e-01 8.74227881e-01 4.33220416e-01 7.06974685e-01
-1.75800905e-01 1.36417735e+00 4.47242767e-01 -2.80133545e-01
3.20688039e-01 -3.31264704e-01 4.07193124e-01 1.74442217e-01
3.76546443e-01 -1.28226137e+00 -6.81375623e-01 -4.21587646e-01
-1.58741213e-02 -2.71440148e-01 -3.63963068e-01 6.01965860e-02
-1.13491285e+00 8.71760547e-01 1.98477745e-01 2.87735492e-01
-5.44303954e-01 -1.33197963e-01 8.46952379e-01 3.22433501e-01
3.46509874e-01 1.10409749e+00 -2.70988822e-01 -3.26506943e-01
-9.23089266e-01 6.86023057e-01 9.80443120e-01 1.12158239e+00
4.42494303e-01 -5.54362714e-01 -8.19172025e-01 1.06879663e+00
3.39500122e-02 5.59732556e-01 5.92721939e-01 -1.21128380e+00
1.04177725e+00 8.01066577e-01 2.78939456e-01 -1.11083829e+00
-2.12046027e-01 -2.10786432e-01 -1.05717647e+00 -1.01967001e+00
8.04557428e-02 -1.09938778e-01 -1.82301089e-01 1.28073740e+00
2.35044464e-01 -4.42472309e-01 3.00156444e-01 6.70621037e-01
1.31053650e+00 9.30569053e-01 -1.13153368e-01 -5.58828175e-01
1.63381922e+00 -1.33810782e+00 -9.45381045e-01 -1.01967156e-01
9.81199503e-01 -9.37093914e-01 1.02848935e+00 1.60396025e-01
-1.23390698e+00 -2.73620099e-01 -8.57550681e-01 -6.47181630e-01
7.52085373e-02 -1.22828998e-01 4.06811178e-01 3.84719849e-01
-8.79048884e-01 4.46334928e-01 -4.67733383e-01 -5.03156781e-01
4.15773541e-01 -1.17294684e-01 -1.92277730e-01 -2.67194957e-01
-1.25248063e+00 5.43564558e-01 3.14936280e-01 -7.75493383e-02
-2.43931562e-01 -6.81685567e-01 -9.55526352e-01 2.90371329e-01
4.71116036e-01 -1.12671936e+00 1.77375650e+00 -4.39359844e-01
-1.30314541e+00 5.96996725e-01 -3.47789675e-01 -4.58076447e-01
2.27119252e-01 -5.22240162e-01 -2.46542126e-01 3.23668599e-01
5.15247405e-01 4.16644961e-01 3.74116302e-01 -8.73543262e-01
-7.02550530e-01 -3.43660355e-01 1.55574009e-01 6.24983132e-01
-3.33870828e-01 5.62751703e-02 -4.11552608e-01 -5.81018448e-01
-5.84386699e-02 -4.91046399e-01 -8.81311744e-02 -7.78375804e-01
-9.20543432e-01 -7.24017262e-01 4.97989058e-01 -1.04371119e+00
1.76094365e+00 -1.87969112e+00 1.48051038e-01 -1.43173173e-01
3.51844430e-01 3.75096381e-01 -4.36214745e-01 1.22973418e+00
4.81878459e-01 2.42961258e-01 -6.63359225e-01 -5.94540656e-01
1.30719900e-01 -1.43375218e-01 -4.83218938e-01 -2.36539826e-01
1.30546153e-01 9.55224454e-01 -1.32500196e+00 -7.04834998e-01
-4.27587777e-01 -2.30080739e-01 -5.37420213e-01 5.47779083e-01
-3.61403346e-01 4.29844379e-01 -7.69811094e-01 3.28292936e-01
5.94509542e-01 -1.88912228e-01 7.35493749e-02 -1.43088505e-01
-5.57727925e-02 8.89142036e-01 -7.01672673e-01 2.23219895e+00
-3.79882783e-01 5.08937597e-01 -1.54863270e-02 -9.46072936e-01
7.01505423e-01 2.41698891e-01 4.58103210e-01 -9.06389117e-01
-3.63914035e-02 2.60604948e-01 -2.66296297e-01 -6.34579897e-01
1.35022223e+00 3.74997765e-01 -6.54239357e-01 6.36272728e-01
-4.61857729e-02 -6.12935841e-01 8.20234895e-01 1.08242321e+00
1.39003003e+00 -4.54199672e-01 5.29690742e-01 -1.53881595e-01
6.93120658e-01 3.96288455e-01 4.27271962e-01 8.39534461e-01
-1.60335928e-01 8.84695411e-01 6.60410643e-01 5.11949919e-02
-9.52456295e-01 -7.58910835e-01 2.58911610e-01 9.02790904e-01
1.31338939e-01 -1.01994717e+00 -8.98436606e-01 -8.35210741e-01
-1.14501387e-01 1.01733756e+00 -2.51360625e-01 -2.57810324e-01
-6.90823793e-01 -4.10909951e-01 5.27441144e-01 1.81996211e-01
8.13754916e-01 -1.01032734e+00 -2.62750715e-01 2.15479150e-01
-1.21778274e+00 -8.88077855e-01 -1.13421857e+00 -5.76630533e-01
-7.45537102e-01 -8.77396226e-01 -7.06829429e-01 -8.44319463e-01
2.22228155e-01 8.62157464e-01 1.47062922e+00 1.34633914e-01
-1.91809312e-02 5.51429570e-01 -7.40669906e-01 -2.78786957e-01
-5.86780667e-01 6.69589460e-01 -3.02635103e-01 -2.85276979e-01
2.73375750e-01 -5.25998950e-01 -8.82362247e-01 -2.36415267e-01
-1.00561917e+00 1.25061378e-01 5.52822948e-01 5.62038839e-01
2.75584817e-01 -2.82632649e-01 1.10040963e+00 -1.01513278e+00
1.39690292e+00 -5.40843546e-01 4.08292040e-02 6.35341167e-01
-1.31883234e-01 1.14415690e-01 6.37364507e-01 2.44507298e-01
-1.07233918e+00 -3.80889475e-01 -4.78631705e-01 5.14832556e-01
8.37584138e-02 9.17727232e-01 -2.15951756e-01 7.11031020e-01
5.15920997e-01 4.31405991e-01 -1.59546480e-01 -5.41020215e-01
6.39337540e-01 1.07796669e+00 6.79911613e-01 -3.22001845e-01
4.61145520e-01 9.10148919e-02 -5.68312764e-01 -1.05227876e+00
-1.26670086e+00 -1.22665656e+00 -4.58306789e-01 1.96541734e-02
7.40475476e-01 -9.09068882e-01 -6.71172023e-01 3.53750795e-01
-1.51699352e+00 1.11970961e-01 -4.01276886e-01 -9.88692120e-02
-3.68400693e-01 9.98215735e-01 -5.63029468e-01 -5.63494503e-01
-1.24122155e+00 -5.97091019e-01 1.14773035e+00 5.30313909e-01
-6.10606194e-01 -6.85795605e-01 5.15942037e-01 1.02578378e+00
3.12014818e-01 -1.07507274e-01 8.40722322e-01 -1.02960706e+00
-2.98322082e-01 -1.29105836e-01 -1.24256559e-01 3.20919544e-01
2.59209484e-01 -1.69646144e-01 -5.11753261e-01 -2.24709347e-01
-5.82690127e-02 -7.18401909e-01 1.03963602e+00 1.11107677e-01
9.02014196e-01 -6.98206902e-01 8.33816156e-02 -1.45208716e-01
9.19499695e-01 -1.96950674e-01 7.31330574e-01 1.64535448e-01
5.77728212e-01 6.36258662e-01 7.53725588e-01 5.40026248e-01
9.73151684e-01 4.10799116e-01 1.67846993e-01 3.88259083e-01
-4.97830892e-03 -3.59252095e-01 2.20712751e-01 1.91023695e+00
2.06347212e-01 -4.85365212e-01 -6.46435142e-01 7.84635127e-01
-1.98440766e+00 -1.02328014e+00 -3.60307634e-01 2.07079697e+00
1.29967451e+00 -2.96191394e-01 1.85128942e-01 -1.48713529e-01
4.49718535e-01 3.19196433e-01 -4.35847282e-01 -3.39297026e-01
-3.46479118e-02 2.55279057e-02 -4.02664542e-02 2.42049545e-01
-9.19965863e-01 7.13372767e-01 5.34465027e+00 1.02010465e+00
-4.66027409e-01 3.83011885e-02 4.57716644e-01 1.10767998e-01
-6.92380846e-01 -2.55147338e-01 -5.16084313e-01 4.59211797e-01
9.41908479e-01 -6.38832033e-01 7.99886882e-02 5.50058842e-01
2.58398473e-01 -3.71913821e-01 -8.83906841e-01 8.85089993e-01
3.74067366e-01 -1.39911532e+00 4.30030018e-01 -4.74188507e-01
9.30582702e-01 -1.15151733e-01 -3.45179826e-01 6.86481237e-01
3.26514035e-01 -7.24142075e-01 4.05447870e-01 2.93375760e-01
3.55547965e-01 -7.06642807e-01 9.44130063e-01 5.10548353e-01
-9.59181607e-01 1.04562528e-01 -4.11782086e-01 3.59203340e-03
4.37785119e-01 5.99052370e-01 -5.19937992e-01 1.17018831e+00
5.84909260e-01 6.26960576e-01 -7.25300252e-01 1.32622230e+00
-2.38574728e-01 6.40250504e-01 -3.52340974e-02 -3.19184422e-01
2.00250521e-01 -3.78894478e-01 5.95390916e-01 1.13501740e+00
2.81850040e-01 3.42321515e-01 -1.22978881e-01 4.28503215e-01
-6.71488464e-01 6.37466371e-01 -2.11468413e-01 -2.53713638e-01
7.64934897e-01 1.45163763e+00 -4.44377214e-01 -4.70893294e-01
-3.45719963e-01 1.34907115e+00 4.90159839e-01 1.60044227e-02
-4.78033930e-01 -8.67454588e-01 4.42420781e-01 -3.92949641e-01
9.84926820e-02 -4.48642410e-02 -6.55066520e-02 -1.51878011e+00
4.02136028e-01 -1.07357574e+00 6.59462452e-01 -6.04734540e-01
-1.45741105e+00 3.01744699e-01 -2.15944842e-01 -1.10751367e+00
-3.26256216e-01 3.34933341e-01 -9.48531330e-01 6.95973158e-01
-1.40165901e+00 -9.48758245e-01 -4.18366849e-01 2.54511625e-01
8.93908679e-01 2.82598704e-01 6.95885539e-01 3.13076884e-01
-6.37509048e-01 4.85873967e-01 3.92060101e-01 1.98527023e-01
1.06025970e+00 -1.37086248e+00 6.91982806e-01 7.79545903e-01
1.11943975e-01 8.64965022e-01 7.18180418e-01 -6.01758242e-01
-1.30601144e+00 -1.04868329e+00 1.47181106e+00 -6.13459766e-01
4.93290901e-01 -2.10975066e-01 -1.15314591e+00 3.01657468e-01
8.04177880e-01 -7.69515932e-01 8.23672652e-01 1.70435116e-01
-9.30629894e-02 -2.66021758e-01 -8.05960178e-01 5.17885685e-01
8.44437182e-01 -4.28384215e-01 -1.11212146e+00 6.03860438e-01
1.19761491e+00 -2.27402732e-01 -8.21855843e-01 1.54652834e-01
1.30565926e-01 -7.70934761e-01 7.12918401e-01 -5.99982321e-01
6.54185295e-01 -2.10928380e-01 2.61436731e-01 -1.65551269e+00
1.96821129e-05 -8.90252948e-01 -1.79664820e-01 1.82015204e+00
3.70647639e-01 -4.05910820e-01 5.58838367e-01 7.17639387e-01
-5.21619081e-01 -3.60728741e-01 -8.10646117e-01 -5.22947133e-01
7.43914396e-02 6.94888295e-05 5.52156091e-01 8.33368361e-01
5.60809851e-01 9.53920782e-01 -8.11174884e-02 -1.66789070e-01
3.93790781e-01 4.36614901e-01 8.74106526e-01 -9.61807191e-01
9.34865549e-02 -5.15320659e-01 1.95304677e-01 -1.41422963e+00
-3.13661933e-01 -8.18383157e-01 1.49892822e-01 -2.20909977e+00
7.00885892e-01 -2.70292431e-01 1.80718735e-01 8.46049562e-02
-6.06922925e-01 -3.12994458e-02 1.57054827e-01 5.28122842e-01
-1.47698760e+00 9.34904158e-01 1.31955767e+00 -3.18738401e-01
-3.45282257e-01 7.20133409e-02 -1.18768001e+00 2.42300794e-01
7.20843911e-01 -2.91169912e-01 -4.47255343e-01 -5.36793172e-01
6.45963609e-01 3.19576353e-01 -2.58831382e-01 -7.40051568e-01
4.29270536e-01 -2.25052331e-02 -3.72283548e-01 -8.69734585e-01
-1.07504494e-01 -1.60133302e-01 -5.40563703e-01 1.97576866e-01
-6.95808589e-01 1.77652061e-01 -3.24942581e-02 4.97505277e-01
-6.84018910e-01 -4.74246532e-01 1.62853956e-01 -1.22783989e-01
-4.23392355e-01 6.26515299e-02 -3.88453186e-01 7.93053389e-01
5.88051140e-01 3.39620113e-01 -7.08000064e-01 -8.87864530e-01
-8.67150277e-02 9.36697483e-01 3.10766757e-01 3.19358915e-01
3.94131929e-01 -1.21535254e+00 -1.24957871e+00 -5.39907694e-01
3.52035522e-01 5.07031620e-01 7.66788125e-01 8.26358736e-01
-6.20806813e-01 7.87776649e-01 6.76496848e-02 -1.60086542e-01
-1.23718345e+00 3.03036571e-01 -2.74994463e-01 -6.55142546e-01
-3.63184482e-01 6.18361294e-01 -1.02591299e-01 -5.01315296e-01
3.14492024e-02 -4.28601086e-01 -5.90268135e-01 3.35455656e-01
8.90427470e-01 5.05894005e-01 2.79258281e-01 -5.00637054e-01
-4.85345721e-02 -8.04935247e-02 -2.96937555e-01 -3.96375135e-02
1.17229164e+00 -6.49168372e-01 -5.47448814e-01 1.93725541e-01
1.23766625e+00 3.14031869e-01 -5.49349368e-01 -2.14019388e-01
1.92431808e-01 -7.45918825e-02 -3.93127084e-01 -7.36564994e-01
-5.26632309e-01 7.57777393e-01 -3.64790231e-01 6.83045805e-01
1.07838261e+00 1.48230523e-01 1.50291193e+00 7.76087642e-01
7.46804848e-03 -1.16155374e+00 3.53629261e-01 6.35386884e-01
1.16022408e+00 -1.22348964e+00 4.65137400e-02 -3.96576613e-01
-8.97458553e-01 8.93771291e-01 4.07505274e-01 2.44560733e-01
-2.18304306e-01 -6.87524557e-01 -3.58349481e-03 -1.90969989e-01
-7.06406415e-01 -1.62288800e-01 3.86286169e-01 4.28771287e-01
3.83365124e-01 -1.37274235e-01 -7.25962877e-01 9.58809555e-01
-4.53453839e-01 -4.08483684e-01 8.28361988e-01 8.85922968e-01
-5.65670311e-01 -9.95866179e-01 -6.24253117e-02 6.73302412e-01
-5.17453730e-01 -3.30913961e-01 -6.09860122e-01 3.88429791e-01
-6.81257725e-01 1.30451930e+00 7.97287524e-02 -1.21385276e-01
5.02147436e-01 -2.26086900e-02 1.76660836e-01 -8.02813172e-01
-8.35330188e-01 -3.93797398e-01 5.80773830e-01 -4.01143432e-01
-5.51374078e-01 -9.41068590e-01 -1.11320710e+00 -3.50408196e-01
-2.61817366e-01 7.93747723e-01 4.12240535e-01 8.58684301e-01
7.14353383e-01 2.80899286e-01 7.67133296e-01 -3.10603648e-01
-6.19159639e-01 -1.34510195e+00 -3.03528011e-01 7.46172786e-01
2.63868988e-01 2.12979376e-01 -1.16400436e-01 2.98137069e-02] | [12.3326416015625, 9.28466796875] |
08a13143-ed58-46ac-a081-45cd2f13d305 | texttt-causalassembly-generating-realistic | 2306.10816 | null | https://arxiv.org/abs/2306.10816v1 | https://arxiv.org/pdf/2306.10816v1.pdf | $\texttt{causalAssembly}$: Generating Realistic Production Data for Benchmarking Causal Discovery | Algorithms for causal discovery have recently undergone rapid advances and increasingly draw on flexible nonparametric methods to process complex data. With these advances comes a need for adequate empirical validation of the causal relationships learned by different algorithms. However, for most real data sources true causal relations remain unknown. This issue is further compounded by privacy concerns surrounding the release of suitable high-quality data. To help address these challenges, we gather a complex dataset comprising measurements from an assembly line in a manufacturing context. This line consists of numerous physical processes for which we are able to provide ground truth causal relationships on the basis of a detailed study of the underlying physics. We use the assembly line data and associated ground truth information to build a system for generation of semisynthetic manufacturing data that supports benchmarking of causal discovery methods. To accomplish this, we employ distributional random forests in order to flexibly estimate and represent conditional distributions that may be combined into joint distributions that strictly adhere to a causal model over the observed variables. The estimated conditionals and tools for data generation are made available in our Python library $\texttt{causalAssembly}$. Using the library, we showcase how to benchmark several well-known causal discovery algorithms. | ['Mathias Drton', 'Martin Roth', 'Steffen Sonntag', 'Tim Pychynski', 'Tobias Windisch', 'Konstantin Göbler'] | 2023-06-19 | null | null | null | null | ['causal-discovery', 'benchmarking', 'benchmarking'] | ['knowledge-base', 'miscellaneous', 'robots'] | [ 4.02431935e-01 1.80346072e-01 6.52125292e-03 -4.45227742e-01
-5.94110310e-01 -6.23567224e-01 8.79496157e-01 5.27587414e-01
2.62167096e-01 1.23724115e+00 1.77071661e-01 -3.05274814e-01
-7.25860953e-01 -1.10886538e+00 -9.82784271e-01 -6.59784615e-01
-3.32035959e-01 8.21469665e-01 2.13405900e-02 1.68179289e-01
2.70241231e-01 4.88984227e-01 -1.75852942e+00 3.88366669e-01
7.82153308e-01 6.79402649e-01 2.89610416e-01 4.23728615e-01
-1.40959397e-01 3.02035451e-01 -5.79814672e-01 -2.94423491e-01
1.22478958e-02 -3.65908325e-01 -5.99820375e-01 -3.74715865e-01
8.99275616e-02 -1.91504788e-02 8.00746903e-02 7.59331048e-01
2.70569682e-01 -4.57217693e-02 8.49349439e-01 -1.50208211e+00
-4.11009610e-01 9.47263479e-01 -2.63377398e-01 -4.34627868e-02
5.07004201e-01 1.62143007e-01 9.01672304e-01 -6.38571143e-01
7.86400855e-01 1.30367613e+00 5.02255440e-01 -1.09277934e-01
-1.71241307e+00 -7.09559500e-01 -9.60480049e-02 -7.82948509e-02
-1.21345627e+00 -4.19293076e-01 3.79566759e-01 -8.20517838e-01
5.51335275e-01 3.69436085e-01 5.31513333e-01 1.53794158e+00
4.99101013e-01 2.46119708e-01 1.41526341e+00 -3.76112968e-01
5.23594022e-01 -1.39022255e-02 3.14234868e-02 5.18199801e-01
5.01153648e-01 5.53923488e-01 -8.79928529e-01 -5.75480759e-01
9.05912638e-01 -2.20300421e-01 -1.31008118e-01 -5.42915881e-01
-1.37531793e+00 7.50626028e-01 1.30271032e-01 3.62632498e-02
-5.23787618e-01 3.07999134e-01 1.63319722e-01 1.07414052e-01
3.33319664e-01 5.09275794e-01 -7.15362430e-01 -1.63516298e-01
-6.73016191e-01 5.39620101e-01 1.03383076e+00 8.79406571e-01
6.79842234e-01 -3.84760916e-01 -1.99594438e-01 5.59094131e-01
4.05376703e-01 4.68506008e-01 -3.73088509e-01 -8.68745029e-01
1.49302930e-01 3.89152527e-01 4.16304678e-01 -9.15967643e-01
-1.51156530e-01 -1.69701561e-01 -6.64870560e-01 3.66313189e-01
5.49345672e-01 -1.46399274e-01 -8.46948504e-01 1.46549141e+00
5.05167842e-01 4.01795357e-02 -1.86116949e-01 6.59010828e-01
3.52344126e-01 3.41890156e-01 3.40449035e-01 -4.20406282e-01
1.00973773e+00 5.24098286e-03 -8.18445504e-01 3.51560652e-01
3.19558606e-02 -9.12069678e-01 9.79985356e-01 7.36149848e-01
-6.87853158e-01 -3.10751140e-01 -9.70430017e-01 3.76416832e-01
-4.56581563e-01 -1.44639775e-01 1.05125272e+00 6.39182210e-01
-5.73614955e-01 9.44375396e-01 -9.20601010e-01 -3.90241116e-01
3.33135337e-01 4.00407106e-01 -3.57944041e-01 -2.00849727e-01
-1.25607371e+00 7.62057126e-01 4.08915550e-01 1.37597531e-01
-1.07769918e+00 -1.04040289e+00 -5.11058450e-01 -2.49831244e-01
6.08254969e-01 -9.83117163e-01 1.00757551e+00 -1.47218123e-01
-1.29822397e+00 2.53177851e-01 1.01337507e-01 -1.60923347e-01
6.96940362e-01 -3.17668617e-01 -3.74228984e-01 -3.13140988e-01
1.91658691e-01 8.28350484e-02 7.31700122e-01 -1.58049405e+00
-4.72188115e-01 -3.49148244e-01 -1.69473708e-01 -5.37590921e-01
3.84720743e-01 -7.35954707e-03 -1.34084433e-01 -5.15762329e-01
1.71256103e-02 -8.05195749e-01 -2.94491202e-01 -3.66721421e-01
-9.41423178e-01 2.93168854e-02 5.30980706e-01 -5.09668648e-01
9.02083278e-01 -1.74683785e+00 9.13343877e-02 2.62734830e-01
-1.63916610e-02 -5.10297835e-01 2.58877575e-01 8.98383856e-01
-2.34819412e-01 2.61392772e-01 -4.26378191e-01 -3.06234751e-02
1.52185813e-01 3.50050241e-01 -3.87287199e-01 3.98029834e-01
4.35264230e-01 5.03758192e-01 -9.29237604e-01 -2.91482925e-01
5.32331109e-01 3.12869906e-01 -3.30142975e-01 2.98388183e-01
-7.41930664e-01 7.18647301e-01 -5.06345749e-01 6.26669168e-01
5.36521852e-01 1.74506176e-02 2.84434646e-01 -6.24098144e-02
-3.98139149e-01 3.30663949e-01 -1.38333559e+00 1.74470747e+00
-3.40539783e-01 -1.53180920e-02 3.45336758e-02 -6.52332664e-01
9.85130966e-01 2.36090243e-01 6.13529742e-01 -2.66206264e-02
9.89380255e-02 6.04566783e-02 -1.57263026e-01 -4.29270238e-01
2.94372946e-01 -3.30536276e-01 -2.52338082e-01 2.98631102e-01
3.44430208e-02 -5.69617867e-01 1.64755851e-01 -1.85119007e-02
1.42965925e+00 4.19628173e-01 1.33640453e-01 -5.26535809e-01
9.84010696e-02 3.09957683e-01 4.76104409e-01 7.51240671e-01
3.84972662e-01 4.71006513e-01 7.04810500e-01 -2.65289634e-01
-9.37415302e-01 -1.59789586e+00 -3.17456424e-01 5.07519901e-01
-1.65262312e-01 -3.48750114e-01 -3.24223965e-01 -5.23166001e-01
3.53543222e-01 1.15848434e+00 -7.65979886e-01 2.86927577e-02
5.04755266e-02 -9.04863000e-01 2.49343425e-01 3.56359482e-02
5.94019098e-03 -8.76873076e-01 -2.97955543e-01 2.58299708e-01
1.39805570e-01 -7.59189367e-01 2.84929216e-01 5.20978987e-01
-6.53403103e-01 -1.33513868e+00 -5.10747032e-03 2.04582110e-01
3.41596037e-01 -4.96557176e-01 1.26683986e+00 -4.09436405e-01
-7.18954742e-01 1.96504757e-01 -1.60018608e-01 -7.35682905e-01
-8.09302926e-01 -1.52735651e-01 1.51570573e-01 -3.14079702e-01
5.82276210e-02 -9.16833103e-01 -2.66979098e-01 3.21713120e-01
-7.55435824e-01 -2.70172097e-02 6.57377601e-01 6.45117462e-01
7.33507931e-01 5.93229413e-01 4.71330702e-01 -9.85629797e-01
4.46711034e-01 -8.52666199e-01 -9.43561673e-01 1.02351569e-01
-7.56371379e-01 3.26071352e-01 2.98481554e-01 -1.24788433e-01
-1.34377468e+00 2.19899788e-01 1.37597740e-01 -2.28520244e-01
-6.26845658e-01 7.78016806e-01 -4.11887795e-01 7.06730902e-01
5.35613239e-01 -3.45765740e-01 -8.60151798e-02 -5.36638021e-01
6.18463695e-01 3.40911746e-01 5.63218534e-01 -1.05956447e+00
7.75590777e-01 4.46518213e-01 3.44902039e-01 -5.40066838e-01
-6.53102994e-01 -1.91505942e-02 -6.98220849e-01 -3.34308505e-01
8.22217226e-01 -5.94691515e-01 -8.20937693e-01 1.66525617e-01
-9.89481926e-01 -3.73826623e-01 -3.88998508e-01 5.29760242e-01
-7.25455701e-01 -2.08563298e-01 -2.37266004e-01 -1.02889121e+00
1.36165559e-01 -9.19434309e-01 1.04739368e+00 -2.48416260e-01
-5.80415606e-01 -7.93896973e-01 1.46312997e-01 1.55145586e-01
5.81902787e-02 7.27673352e-01 1.30955625e+00 -3.49569380e-01
-8.27313066e-01 -9.43363532e-02 -1.04134955e-01 7.38123134e-02
4.75306690e-01 5.58498979e-01 -8.02108884e-01 2.10581839e-01
-1.57810688e-01 -2.07370613e-02 5.05640328e-01 5.81619918e-01
1.05002832e+00 4.56595346e-02 -5.95508635e-01 2.22040322e-02
1.32350409e+00 -1.98372111e-01 6.37286544e-01 1.70079526e-02
5.87270737e-01 1.10390472e+00 8.25316012e-01 5.93926251e-01
1.48795664e-01 5.55455387e-01 6.63667023e-01 1.83200166e-01
5.86974621e-02 -4.70622838e-01 1.34469450e-01 3.78397167e-01
-1.49172381e-01 -1.98331892e-01 -9.96621847e-01 4.37452793e-01
-1.77240503e+00 -9.55111206e-01 -1.06567025e+00 2.39785480e+00
9.73327816e-01 1.48684323e-01 -9.11998302e-02 -5.25664091e-02
7.04874575e-01 -2.64967740e-01 -3.30623507e-01 -2.41218880e-01
9.92894992e-02 2.21579388e-01 7.07051218e-01 4.07653689e-01
-9.20632780e-01 3.74900639e-01 6.64359951e+00 5.80161333e-01
-6.40911460e-01 -9.60878748e-03 3.95825207e-01 -8.02323371e-02
-5.45014441e-01 3.16415787e-01 -6.45098984e-01 4.12900984e-01
1.28990114e+00 -1.13169365e-01 3.30423772e-01 5.22270024e-01
7.97296166e-01 -5.37223279e-01 -1.43394589e+00 5.19347548e-01
-7.35969305e-01 -1.49099743e+00 -1.17780827e-01 3.66313249e-01
7.45230138e-01 -1.20782599e-01 -2.09187284e-01 -8.02303329e-02
1.12433934e+00 -1.17538834e+00 7.12646127e-01 1.00997961e+00
6.33434594e-01 -6.12976432e-01 4.64083672e-01 1.38961211e-01
-5.21476567e-01 7.72179151e-03 -3.71785350e-02 -2.44899213e-01
3.50635886e-01 1.56176651e+00 -1.10838521e+00 1.09942257e+00
7.58413374e-01 3.99014801e-01 2.10772865e-02 7.63579428e-01
-4.05923963e-01 7.88358033e-01 -4.74071950e-01 1.74566045e-01
-5.43016970e-01 -2.29403093e-01 4.45928305e-01 7.96807230e-01
5.67213595e-01 -1.18562713e-01 9.99448672e-02 1.48045230e+00
2.44156137e-01 -1.61953390e-01 -7.22392619e-01 -2.72023678e-01
4.72558975e-01 1.07894659e+00 -7.39255786e-01 8.76618773e-02
-4.82733473e-02 5.79252362e-01 3.65944616e-02 6.21700138e-02
-7.94213474e-01 3.12641919e-01 9.33270335e-01 3.60641181e-01
1.27951831e-01 -5.48384726e-01 -5.45943558e-01 -6.57530248e-01
-2.85326630e-01 -7.54569471e-01 2.69733548e-01 -8.97460997e-01
-1.65699053e+00 -1.95141640e-02 4.96739388e-01 -5.91423035e-01
-3.24062288e-01 -4.63242978e-01 -2.35729590e-01 1.18556678e+00
-8.31146181e-01 -1.06026614e+00 -1.37308612e-01 3.47672075e-01
1.92932889e-01 1.95079863e-01 1.08703661e+00 2.13148305e-03
-5.45781851e-01 -3.63372684e-01 -8.26178864e-02 -5.37653983e-01
7.74996996e-01 -1.46172523e+00 2.37650499e-01 6.14032209e-01
-2.30733193e-02 8.91543806e-01 1.33657217e+00 -1.33721542e+00
-1.46792853e+00 -1.05967474e+00 5.54437578e-01 -8.52733552e-01
1.15524864e+00 -5.55581748e-01 -6.23117328e-01 5.53213239e-01
3.17148566e-02 -2.24136546e-01 7.55166233e-01 6.24119163e-01
-4.54733111e-02 1.07163146e-01 -1.04864371e+00 2.44936660e-01
1.07917106e+00 -2.22405970e-01 -6.88207388e-01 3.47927094e-01
5.99623621e-01 -5.28504327e-02 -1.38056612e+00 4.26121742e-01
3.13467145e-01 -1.04035020e+00 7.66940415e-01 -3.45038354e-01
9.20909822e-01 -6.05985105e-01 -2.09192708e-01 -1.35441101e+00
3.38651729e-03 -6.57913566e-01 1.12098262e-01 1.62404168e+00
5.07079482e-01 -5.30406654e-01 4.90960896e-01 7.52701342e-01
3.24851610e-02 -2.54536986e-01 -8.11852872e-01 -7.63309062e-01
-5.33294454e-02 -7.79879749e-01 8.24073672e-01 1.01043534e+00
-2.29664147e-01 3.04786980e-01 -1.12909064e-01 6.42930627e-01
9.50768530e-01 3.04950982e-01 8.67029428e-01 -1.57687557e+00
-4.34374958e-01 4.12186868e-02 -1.69727057e-01 -1.14325270e-01
-2.55350154e-02 -6.10683203e-01 2.29069039e-01 -1.56601012e+00
1.27620399e-01 -1.04293275e+00 -1.22651696e-01 4.65261459e-01
1.35940006e-02 -1.82315409e-01 -3.42739135e-01 5.23098446e-02
1.90765038e-01 5.86994827e-01 9.61110532e-01 1.28408656e-01
-8.31881687e-02 -2.30837036e-02 -3.04796219e-01 5.14173865e-01
6.71200693e-01 -6.98428690e-01 -5.00662744e-01 9.49082524e-02
4.14584726e-01 3.30305427e-01 8.41801822e-01 -7.84323394e-01
-1.16085425e-01 -4.76596236e-01 3.30428988e-01 -7.93522537e-01
1.06453978e-01 -8.67519379e-01 1.05454564e+00 2.52526075e-01
-3.79300624e-01 -1.71745196e-01 1.83457419e-01 8.73819053e-01
-8.23852569e-02 -7.47614726e-02 8.54943097e-02 -3.03686131e-02
-3.15743119e-01 7.47863874e-02 -2.51953751e-01 -4.08136040e-01
9.89252269e-01 1.22216865e-01 -1.32688805e-01 -1.01972237e-01
-9.75833893e-01 -9.23292618e-03 4.77830112e-01 4.87943470e-01
1.50639817e-01 -1.04769278e+00 -7.95043349e-01 -2.98140151e-03
9.82806832e-02 -4.79731476e-03 1.25329807e-01 7.02128232e-01
-1.11944787e-01 3.09406906e-01 -2.30355769e-01 -5.04898310e-01
-8.36874127e-01 7.08728909e-01 -1.33759886e-01 -1.66541502e-01
-3.34268630e-01 6.23876214e-01 8.15411732e-02 -4.99981314e-01
-5.11987507e-01 -5.78719258e-01 2.80497432e-01 1.08537875e-01
1.82505161e-01 4.46538866e-01 1.78037733e-01 1.05364554e-01
-2.83478916e-01 -3.08547080e-01 4.33458298e-01 -2.28858098e-01
1.75142038e+00 -6.44111931e-02 -4.69461560e-01 9.74654615e-01
5.84420621e-01 4.05193776e-01 -1.47415149e+00 5.39278209e-01
1.61262125e-01 -5.25478721e-01 -9.45144594e-02 -1.33505809e+00
-7.02931583e-01 4.13368851e-01 4.22657341e-01 5.88667274e-01
7.20534444e-01 1.76014423e-01 -2.78878529e-02 -2.72665620e-01
7.81744003e-01 -7.05585778e-01 -5.04936457e-01 1.52964994e-01
1.23142743e+00 -8.91506076e-01 2.49836460e-01 -1.03837132e+00
-1.26608191e-02 1.00822294e+00 6.51054308e-02 -2.86444859e-03
5.96823514e-01 6.56500518e-01 -2.31966943e-01 -5.41602671e-01
-8.23076248e-01 -5.19971438e-02 4.64994572e-02 8.70686889e-01
4.48405325e-01 3.87718022e-01 -2.70361632e-01 5.78077495e-01
-4.20663029e-01 2.20870867e-01 4.87883091e-01 6.87106967e-01
1.12713732e-01 -1.60391474e+00 -7.08812952e-01 6.89883232e-01
-3.32197696e-01 -8.84021297e-02 -3.04250062e-01 9.63639617e-01
2.51807272e-01 1.13817191e+00 -1.23354122e-01 -1.43128902e-01
5.18832803e-01 9.35775414e-02 5.81004322e-01 -6.99704230e-01
-2.23586783e-01 -1.89313799e-01 6.17320180e-01 -6.80160880e-01
-2.50719696e-01 -1.31161284e+00 -1.21224022e+00 -2.91263461e-01
-1.61291659e-01 -4.95137200e-02 1.02261567e+00 8.42340469e-01
2.78268039e-01 9.79084373e-01 7.05014542e-02 -8.17255080e-01
-4.71590042e-01 -8.20028007e-01 -6.82331502e-01 2.97073394e-01
-2.46533573e-01 -1.09336269e+00 -1.72390401e-01 4.05157089e-01] | [7.805120944976807, 5.2033610343933105] |
e93912c7-8bd5-4487-89ba-604f091f15f6 | deep-spatial-semantic-attention-for-fine | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Song_Deep_Spatial-Semantic_Attention_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Song_Deep_Spatial-Semantic_Attention_ICCV_2017_paper.pdf | Deep Spatial-Semantic Attention for Fine-Grained Sketch-Based Image Retrieval | Human sketches are unique in being able to capture both the spatial topology of a visual object, as well as its subtle appearance details. Fine-grained sketch-based image retrieval (FG-SBIR) importantly leverages on such fine-grained characteristics of sketches to conduct instance-level retrieval of photos. Nevertheless, human sketches are often highly abstract and iconic, resulting in severe misalignments with candidate photos which in turn make subtle visual detail matching difficult. Existing FG-SBIR approaches focus only on coarse holistic matching via deep cross-domain representation learning, yet ignore explicitly accounting for fine-grained details and their spatial context. In this paper, a novel deep FG-SBIR model is proposed which differs significantly from the existing models in that: (1) It is spatially aware, achieved by introducing an attention module that is sensitive to the spatial position of visual details; (2) It combines coarse and fine semantic information via a shortcut connection fusion block; and (3) It models feature correlation and is robust to misalignments between the extracted features across the two domains by introducing a novel higher order learnable energy function (HOLEF) based loss. Extensive experiments show that the proposed deep spatial-semantic attention model significantly outperforms the state-of-the-art.
| ['Yi-Zhe Song', 'Qian Yu', 'Jifei Song', 'Timothy M. Hospedales', 'Tao Xiang'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [-1.02816299e-02 -5.27978182e-01 -1.87596396e-01 -3.57560009e-01
-8.31816316e-01 -7.37337708e-01 1.02414858e+00 7.70231634e-02
1.49268191e-02 3.46089631e-01 4.61833388e-01 2.86607474e-01
-5.69457114e-01 -7.55657077e-01 -7.29545295e-01 -3.91874164e-01
1.78698421e-01 2.77576834e-01 2.55647391e-01 -1.95049152e-01
5.29427946e-01 8.61691713e-01 -1.47455299e+00 4.68901992e-01
6.30973041e-01 1.40290499e+00 2.59000957e-01 2.30660200e-01
-3.15614313e-01 4.78629053e-01 -4.22752261e-01 -5.16281903e-01
4.32260096e-01 -1.73641369e-01 -4.70896840e-01 -2.72316858e-03
1.08958042e+00 -5.19388735e-01 -6.73427761e-01 1.02478313e+00
3.12350094e-01 1.74537107e-01 7.02431440e-01 -1.21987033e+00
-1.27307379e+00 -1.02617415e-02 -6.74713850e-01 -3.96466479e-02
3.76680851e-01 1.40000522e-01 1.24878073e+00 -1.19944417e+00
6.59512818e-01 1.53994858e+00 6.12711370e-01 2.63936311e-01
-1.11777413e+00 -7.22172678e-01 5.31999171e-01 1.38277948e-01
-1.66418266e+00 -2.23475441e-01 1.33206296e+00 -3.46052200e-01
6.64244592e-01 2.73031265e-01 6.37908220e-01 1.12192130e+00
-1.08957380e-01 8.14126194e-01 1.08630240e+00 -2.42900819e-01
9.62908044e-02 -9.06065106e-02 -1.51068270e-01 7.72842228e-01
1.28714293e-01 2.05340773e-01 -6.57248020e-01 -1.36181056e-01
1.27885306e+00 7.08751500e-01 -6.99401349e-02 -5.98848403e-01
-1.24495983e+00 6.62816048e-01 9.05180156e-01 4.50429916e-01
-4.89681423e-01 3.95519733e-01 1.86091110e-01 1.71098679e-01
3.76624197e-01 4.31629062e-01 -8.87404904e-02 1.65942013e-01
-1.13849008e+00 4.65228975e-01 3.66869479e-01 9.83320117e-01
8.96107018e-01 -4.59821671e-02 -5.50314188e-01 1.05056906e+00
1.92901880e-01 4.15749431e-01 2.99614221e-01 -8.83320272e-01
3.57576817e-01 8.92716646e-01 7.56860226e-02 -1.49473691e+00
1.33880153e-01 -2.95588791e-01 -8.10536146e-01 2.87866294e-01
2.78469436e-02 7.04958558e-01 -1.00652766e+00 1.76370323e+00
4.36008610e-02 1.17763899e-01 -5.10887146e-01 1.22238171e+00
1.10584641e+00 2.87260711e-01 1.50933370e-01 5.37902534e-01
1.28568113e+00 -1.03060734e+00 -5.56860447e-01 -2.67131597e-01
-5.02863169e-01 -7.27230012e-01 1.23262644e+00 -7.56486654e-02
-1.23330259e+00 -7.39408970e-01 -1.15958476e+00 -5.69705427e-01
-7.63347805e-01 1.76244542e-01 6.65838182e-01 1.80166408e-01
-1.10158801e+00 5.92027009e-01 -2.41300881e-01 -2.93237269e-01
7.06697762e-01 1.13536000e-01 -6.06276572e-01 -4.48454529e-01
-9.87850726e-01 6.52419150e-01 -4.27137949e-02 1.28038406e-01
-7.00598300e-01 -8.90799105e-01 -9.60669279e-01 3.41176599e-01
2.31575936e-01 -8.63089979e-01 6.46818459e-01 -1.01056325e+00
-1.26474810e+00 8.47081363e-01 -1.58611998e-01 2.75553524e-01
6.44407213e-01 -7.42762312e-02 -1.79859698e-01 3.11818719e-01
2.72823215e-01 8.77820134e-01 1.14496827e+00 -1.44822347e+00
-2.24866301e-01 -6.03221416e-01 2.25684330e-01 1.82281375e-01
-1.96832150e-01 -2.06475452e-01 -7.59182692e-01 -1.21075439e+00
1.92685217e-01 -5.55815101e-01 1.07751310e-01 8.21249187e-01
-1.65302590e-01 -2.25202754e-01 1.01443231e+00 -4.90050316e-01
1.11235619e+00 -2.06000519e+00 1.19701356e-01 2.00677305e-01
2.71465302e-01 2.56006867e-01 -5.66328108e-01 8.05904984e-01
1.41391367e-01 9.10776481e-02 -6.43066615e-02 -3.06430638e-01
3.63516748e-01 1.16815463e-01 -5.66830754e-01 2.08005741e-01
5.18631160e-01 1.46241677e+00 -9.99666274e-01 -3.15963864e-01
4.90275085e-01 7.00508654e-01 -2.89999157e-01 2.06364378e-01
-9.54318047e-02 1.13445461e-01 -6.96112037e-01 1.05365694e+00
8.61295938e-01 -2.82258362e-01 -1.56442642e-01 -6.76501751e-01
3.12501262e-03 9.11096558e-02 -1.02217937e+00 2.03883839e+00
-4.71985400e-01 5.25174677e-01 1.30632579e-01 -7.73500144e-01
9.60418224e-01 -1.47626260e-02 3.36700320e-01 -1.10868418e+00
-2.51232117e-01 8.74697790e-02 -6.96434379e-01 -7.81383589e-02
5.18840790e-01 -1.12524889e-01 -1.88253239e-01 3.04227740e-01
1.74953807e-02 -1.35087147e-01 -2.59789258e-01 3.70899498e-01
9.25625086e-01 2.27476656e-01 1.63339049e-01 -1.28187388e-01
4.60419625e-01 -4.94551271e-01 3.55024815e-01 7.97688663e-01
-1.40852511e-01 8.91669154e-01 1.20914638e-01 -7.45194018e-01
-9.91694093e-01 -1.40112424e+00 2.47595794e-02 9.27630782e-01
7.37017035e-01 -4.17243391e-01 -2.46213362e-01 -5.95606983e-01
5.55062711e-01 4.68757898e-02 -8.55449200e-01 -9.02535319e-02
-3.78978699e-01 9.01355967e-02 3.38343590e-01 7.35146463e-01
6.68636978e-01 -1.13321531e+00 -2.07775980e-01 1.00902751e-01
2.92881746e-02 -1.11164856e+00 -9.66526449e-01 -4.37850147e-01
-5.43667912e-01 -1.03722239e+00 -9.35154438e-01 -6.57194495e-01
5.67638636e-01 7.95264363e-01 1.20245123e+00 3.87877494e-01
-6.80708885e-01 6.92721426e-01 -2.94012874e-01 -1.01714194e-01
3.17103475e-01 -3.07322949e-01 -2.88171053e-01 2.82932431e-01
3.05423260e-01 -6.82625592e-01 -1.00924528e+00 3.03294837e-01
-1.02045739e+00 -6.20667338e-02 8.15856814e-01 1.01585376e+00
7.12128401e-01 -1.44862354e-01 5.02449095e-01 -4.76695806e-01
6.19382381e-01 -2.35106871e-01 -3.96079689e-01 5.97399712e-01
-3.62626225e-01 -2.12198272e-02 4.03500706e-01 -3.75968486e-01
-8.76125276e-01 -1.99551374e-01 2.30085507e-01 -9.84736741e-01
-4.88225855e-02 6.79644272e-02 -1.29588604e-01 -4.36330616e-01
5.86783923e-02 3.02583605e-01 -2.12135404e-01 -6.26159489e-01
4.01282251e-01 2.31552675e-01 5.44058740e-01 -8.62300515e-01
9.63428974e-01 6.70303524e-01 7.24742785e-02 -7.18626916e-01
-6.00782931e-01 -4.89060968e-01 -6.63504422e-01 -9.72110927e-02
6.32726490e-01 -1.00844669e+00 -6.80473804e-01 4.18285549e-01
-1.07760024e+00 3.87199712e-03 -1.61584795e-01 8.83811563e-02
-4.26580042e-01 4.68616873e-01 -6.06980145e-01 -7.52067804e-01
-4.99069810e-01 -1.01968193e+00 1.83100474e+00 1.10148765e-01
2.10302323e-02 -9.23631787e-01 -2.14105666e-01 2.46341199e-01
6.47305429e-01 3.14370096e-01 9.10836041e-01 -1.26083896e-01
-1.12523437e+00 -3.33057493e-01 -9.88601625e-01 2.17039455e-02
2.07861036e-01 -1.52961882e-02 -9.51388180e-01 -3.31905305e-01
-5.14506280e-01 -2.76992261e-01 1.01794600e+00 2.17198014e-01
1.38105416e+00 -4.27222759e-01 -2.44985878e-01 6.36065125e-01
1.68498707e+00 -2.46678263e-01 5.92826843e-01 1.38540938e-01
9.89109218e-01 5.48625231e-01 5.47851920e-01 3.63313168e-01
4.78386194e-01 1.04117477e+00 3.84947568e-01 -3.05191487e-01
-4.65011179e-01 -6.81824446e-01 -7.40112662e-02 4.05537695e-01
7.22870529e-02 1.50596917e-01 -3.73179615e-01 6.70744002e-01
-1.93969309e+00 -1.17850840e+00 4.02805895e-01 2.07190990e+00
5.52296400e-01 -3.06961685e-01 1.16982702e-02 -1.71462491e-01
6.68500423e-01 6.30173862e-01 -5.87167561e-01 -2.77876139e-01
-2.58702606e-01 3.47834557e-01 1.92923397e-01 2.97345728e-01
-1.08932018e+00 9.29901481e-01 5.37083721e+00 1.09499621e+00
-1.06608903e+00 -6.26678169e-02 3.48840296e-01 -1.19019866e-01
-6.64201379e-01 -1.78460315e-01 -3.23791951e-01 5.58195472e-01
-1.93409353e-01 2.08704039e-01 6.45826697e-01 6.47011280e-01
-2.15078965e-01 2.65616953e-01 -1.12590826e+00 1.27882278e+00
2.24076167e-01 -1.59862471e+00 6.02348685e-01 -6.33316487e-02
8.37345302e-01 -2.38502339e-01 2.39971608e-01 9.83303636e-02
6.27849028e-02 -1.12836182e+00 9.64016020e-01 7.53858685e-01
1.07421064e+00 -6.61130011e-01 3.16388696e-01 -1.36438444e-01
-1.61069787e+00 -2.84831107e-01 -3.86884719e-01 -1.31847300e-02
-1.87133580e-01 3.11927587e-01 -1.21220082e-01 5.83451986e-01
7.85696864e-01 9.16237712e-01 -5.70663154e-01 8.66501629e-01
-1.01221278e-02 -1.94433808e-01 -7.10580796e-02 1.38528302e-01
4.69073743e-01 -2.43276075e-01 4.31786537e-01 1.20654762e+00
2.95021951e-01 1.80310234e-02 2.27277651e-01 1.30949533e+00
-2.26134881e-01 -2.08401576e-01 -6.88610554e-01 -1.23382367e-01
6.91477835e-01 1.31783998e+00 -4.80642349e-01 -1.85023874e-01
-3.95196825e-01 1.28817177e+00 5.81234813e-01 5.73655248e-01
-3.89705747e-01 -2.94763565e-01 1.00581515e+00 1.54236779e-01
7.32499063e-01 -1.21746957e-01 -4.04893816e-01 -1.12781143e+00
3.74508083e-01 -5.26428521e-01 3.10088515e-01 -9.81373191e-01
-1.94868982e+00 5.31288207e-01 -3.38204205e-01 -1.18846393e+00
1.75921187e-01 -5.44342875e-01 -8.04827332e-01 1.11085808e+00
-1.90999734e+00 -1.95298326e+00 -6.03428960e-01 8.15455139e-01
6.43034160e-01 -6.55419230e-02 7.02528000e-01 3.77059579e-01
-1.81352764e-01 8.21849823e-01 -1.21570528e-01 1.77225977e-01
7.69805908e-01 -1.09952927e+00 4.84444141e-01 4.33424383e-01
2.71232992e-01 8.68942082e-01 1.19885996e-01 -5.48592687e-01
-1.63605976e+00 -1.20770335e+00 7.84561217e-01 -4.88577813e-01
4.46748286e-01 -5.20768762e-01 -9.78491843e-01 2.42154300e-01
-1.15282230e-01 4.41984266e-01 2.17550203e-01 -1.02088049e-01
-1.06905723e+00 -3.34438384e-01 -1.21015382e+00 5.31983793e-01
1.26155090e+00 -1.09162009e+00 -4.97742683e-01 7.77085796e-02
3.96178216e-01 -1.32180840e-01 -8.57988536e-01 4.27476317e-01
1.13378727e+00 -1.04266775e+00 1.51990616e+00 -5.45649886e-01
5.29457510e-01 -2.73960382e-01 -4.11943167e-01 -8.87925446e-01
-7.15470254e-01 -3.20564181e-01 -2.16332562e-02 1.13256443e+00
-2.15179458e-01 -3.20079774e-01 4.64679450e-01 5.58923006e-01
1.53275475e-01 -8.37400794e-01 -8.27377379e-01 -6.92776978e-01
-5.13609648e-02 -1.11538917e-01 9.13618147e-01 1.03952110e+00
-3.17052245e-01 7.32679442e-02 -3.41715038e-01 5.73685542e-02
6.22411966e-01 7.65411854e-01 5.60621560e-01 -1.33286393e+00
-7.90189803e-02 -8.98729324e-01 -6.85609221e-01 -9.96613860e-01
2.15932295e-01 -6.57170773e-01 -2.22338244e-01 -1.56344604e+00
3.56257528e-01 -6.18155301e-01 -5.60221732e-01 3.41731071e-01
-2.89959729e-01 6.53104961e-01 4.40099925e-01 2.83474088e-01
-8.17378819e-01 7.43253827e-01 1.27363324e+00 -3.78457963e-01
2.33072683e-01 -3.73296171e-01 -7.24336565e-01 4.46291298e-01
2.24519193e-01 -6.27417862e-02 -2.99607843e-01 -5.25497496e-01
1.36528360e-02 -5.68629429e-03 1.07191825e+00 -7.12251902e-01
3.08068216e-01 -2.18055546e-01 8.82543087e-01 -5.60645401e-01
6.33142829e-01 -9.96270657e-01 6.58130720e-02 1.08981680e-03
-4.28864181e-01 1.18366189e-01 1.54845402e-01 9.12167847e-01
-3.66399020e-01 2.62980670e-01 5.68279505e-01 -3.51118684e-01
-9.19731677e-01 7.47784674e-01 2.46645287e-01 -5.25235273e-02
7.23093152e-01 -4.58888143e-01 -2.92158246e-01 -3.92764628e-01
-1.99318230e-01 -7.88106471e-02 8.23264718e-01 7.56411076e-01
8.93739522e-01 -1.76082289e+00 -4.50274110e-01 4.34151441e-01
5.10821819e-01 -3.02887321e-01 5.90632439e-01 3.69948298e-01
-5.69511726e-02 5.34899354e-01 -3.93137485e-01 -4.05450881e-01
-9.90159988e-01 7.46819377e-01 1.00777045e-01 -6.04259223e-02
-7.15565860e-01 8.43398750e-01 7.56038725e-01 -4.10327524e-01
2.76865929e-01 -6.87749162e-02 1.76216617e-01 -1.33751020e-01
5.73744237e-01 1.44578204e-01 -1.10330693e-02 -6.79874599e-01
-4.76069480e-01 1.18239093e+00 -3.11548673e-02 1.49550259e-01
1.28688300e+00 -1.45968169e-01 -4.73811999e-02 1.70905486e-01
1.41721249e+00 -2.93290298e-02 -1.50200832e+00 -6.18747830e-01
-2.59059042e-01 -1.06527030e+00 1.33252025e-01 -1.15522170e+00
-1.08064914e+00 1.02533579e+00 3.07138294e-01 -2.41737574e-01
1.20618176e+00 3.41420248e-02 8.56114864e-01 5.82226925e-02
3.96345347e-01 -8.69562089e-01 4.86913115e-01 1.52922258e-01
1.37444389e+00 -1.27923977e+00 1.21698081e-01 -2.56031603e-01
-4.42981988e-01 1.09832168e+00 5.30970752e-01 -4.49183613e-01
5.32437801e-01 -1.06084809e-01 -2.67749697e-01 -4.96706724e-01
-5.74681878e-01 -1.27942756e-01 9.65800464e-01 5.76207757e-01
1.40148848e-01 1.97715517e-02 -5.57528287e-02 7.24746525e-01
3.79628211e-01 -1.68627828e-01 -2.84528017e-01 8.00190449e-01
-1.62148863e-01 -9.80710387e-01 -1.44076332e-01 3.48741382e-01
-1.59056589e-01 -1.99966922e-01 -8.35761607e-01 6.98224306e-01
-3.63565274e-02 5.63419759e-01 1.45411596e-01 -1.52135491e-01
4.18503851e-01 -2.31899574e-01 7.21085608e-01 -2.45711759e-01
-4.56805170e-01 -1.40568912e-01 -4.04102206e-01 -8.62836897e-01
-2.46189311e-01 -2.60981232e-01 -6.26667798e-01 -3.20726007e-01
7.35864341e-02 -2.91597277e-01 7.27348626e-01 6.93424761e-01
7.33301938e-01 3.08759004e-01 6.37787342e-01 -1.30616975e+00
-3.99380505e-01 -6.25909090e-01 -7.06262350e-01 7.66219497e-01
4.85314012e-01 -1.07783544e+00 -1.38446927e-01 -3.88606519e-01] | [11.648605346679688, 0.5904399156570435] |
05b592a4-5178-4342-8eeb-b40bb7a09536 | wyweb-a-nlp-evaluation-benchmark-for | 2305.14150 | null | https://arxiv.org/abs/2305.14150v1 | https://arxiv.org/pdf/2305.14150v1.pdf | WYWEB: A NLP Evaluation Benchmark For Classical Chinese | To fully evaluate the overall performance of different NLP models in a given domain, many evaluation benchmarks are proposed, such as GLUE, SuperGLUE and CLUE. The fi eld of natural language understanding has traditionally focused on benchmarks for various tasks in languages such as Chinese, English, and multilingua, however, there has been a lack of attention given to the area of classical Chinese, also known as "wen yan wen", which has a rich history spanning thousands of years and holds signifi cant cultural and academic value. For the prosperity of the NLP community, in this paper, we introduce the WYWEB evaluation benchmark, which consists of nine NLP tasks in classical Chinese, implementing sentence classifi cation, sequence labeling, reading comprehension, and machine translation. We evaluate the existing pre-trained language models, which are all struggling with this benchmark. We also introduce a number of supplementary datasets and additional tools to help facilitate further progress on classical Chinese NLU. The github repository is https://github.com/baudzhou/WYWEB. | ['Yin Zhang', 'Xiaomi Zhong', 'Tianyu Wang', 'Qianglong Chen', 'Bo Zhou'] | 2023-05-23 | null | null | null | null | ['reading-comprehension'] | ['natural-language-processing'] | [-2.12047324e-02 -2.09857509e-01 -2.80377895e-01 -2.62716055e-01
-1.13805926e+00 -8.45713079e-01 5.33146858e-01 2.92797774e-01
-5.12664318e-01 1.04587162e+00 4.96571094e-01 -6.65784299e-01
3.42579454e-01 -5.79063356e-01 -4.93962765e-01 -3.35802794e-01
3.22905898e-01 4.08744842e-01 1.22357294e-01 -3.84532332e-01
4.08188909e-01 -2.28174567e-01 -9.20023084e-01 5.97546339e-01
1.40562737e+00 6.95130527e-01 4.27024871e-01 5.52940428e-01
-2.66601115e-01 7.65325189e-01 -6.44375980e-01 -6.22464776e-01
-3.71628165e-01 -7.26686835e-01 -1.18777871e+00 -5.95535219e-01
2.28618324e-01 9.03539080e-03 -6.66713808e-03 1.06582963e+00
6.86667442e-01 -3.11933346e-02 3.97119313e-01 -9.45712805e-01
-1.14767921e+00 1.04335988e+00 -2.45140389e-01 8.07526559e-02
7.04858005e-01 1.01423413e-01 1.32878542e+00 -9.12453830e-01
8.04281354e-01 1.04650450e+00 3.00331563e-01 5.96825480e-01
-6.28776014e-01 -7.53969967e-01 2.70523220e-01 3.66809756e-01
-1.19085395e+00 -2.23882511e-01 3.41710836e-01 -1.86265811e-01
1.11164665e+00 1.06351584e-01 1.90398499e-01 1.29179001e+00
5.44891432e-02 1.30961561e+00 9.99467671e-01 -6.79264307e-01
-9.51108038e-02 -4.97045033e-02 3.14564914e-01 5.54073870e-01
-1.77675605e-01 -3.64521623e-01 -6.89808309e-01 3.41201276e-01
5.62081113e-02 -2.45714605e-01 -6.28411472e-01 4.43503320e-01
-1.57807600e+00 6.89070761e-01 1.24477774e-01 6.43489778e-01
2.38070250e-01 -2.82872289e-01 6.69103861e-01 4.25346822e-01
6.61925375e-01 5.81972361e-01 -7.90879309e-01 -5.76504111e-01
-6.13665760e-01 1.25130460e-01 1.15979075e+00 1.33593416e+00
4.21147346e-01 -3.22620720e-01 -3.13619822e-01 9.74667847e-01
-9.44423601e-02 5.17275035e-01 6.46852791e-01 -6.04910433e-01
1.10438395e+00 6.12570763e-01 -6.02966659e-02 -6.12326384e-01
-2.39148676e-01 -5.96154779e-02 -8.80126536e-01 -6.25541389e-01
4.53047454e-01 -3.87176514e-01 -5.14057398e-01 1.60855591e+00
-3.70167010e-02 -2.92065181e-02 3.89596343e-01 6.83400571e-01
1.14656222e+00 9.72599864e-01 4.22505848e-02 9.74897295e-02
1.36410713e+00 -1.45013809e+00 -5.48116803e-01 -2.27141038e-01
1.01131320e+00 -1.03846419e+00 1.61143899e+00 3.11170131e-01
-1.02772665e+00 -3.95704776e-01 -8.15365076e-01 -6.25008821e-01
-6.58158541e-01 4.13961530e-01 4.32828933e-01 3.97843212e-01
-8.36434901e-01 3.48759860e-01 -8.15576077e-01 -5.40077329e-01
2.24272847e-01 -2.94028610e-01 -2.81086475e-01 -4.44776297e-01
-1.54241824e+00 1.06555545e+00 5.92261255e-01 1.09246574e-01
-6.48581862e-01 -8.51569116e-01 -9.05504584e-01 1.47663593e-01
3.86049509e-01 -9.38978642e-02 1.41399848e+00 -6.70918345e-01
-1.65675187e+00 1.05922389e+00 -3.04040492e-01 -2.48709857e-01
4.24754739e-01 -6.86500251e-01 -5.30694604e-01 -9.06672031e-02
2.58795351e-01 5.93684435e-01 -2.37635318e-02 -5.27965069e-01
-6.75271332e-01 -6.06913418e-02 9.10867006e-02 1.62396282e-01
-3.34428430e-01 5.67658603e-01 -7.40457416e-01 -6.18503213e-01
-3.92179221e-01 -8.10655713e-01 5.75214289e-02 -5.13954759e-01
-6.76180065e-01 -6.70731902e-01 4.75790322e-01 -9.75891173e-01
1.43838537e+00 -2.07253480e+00 2.18931008e-02 -4.42545444e-01
-1.52823016e-01 2.91810095e-01 -3.07809532e-01 8.47233951e-01
1.84713066e-01 5.64290643e-01 -3.91589761e-01 -1.30055368e-01
2.75646616e-02 6.73060641e-02 -3.33182484e-01 -8.23790953e-03
3.60686004e-01 1.09971118e+00 -1.13198054e+00 -3.57625067e-01
-2.01079324e-01 8.42258930e-02 -2.71331310e-01 2.80593604e-01
-5.86168408e-01 5.64358592e-01 -6.48646057e-01 7.93440878e-01
4.93998736e-01 -4.31211352e-01 2.03035504e-01 3.92590404e-01
-4.18108016e-01 8.73993397e-01 -4.70332831e-01 1.91185677e+00
-5.88726938e-01 6.38734400e-01 -9.88642722e-02 -7.68324196e-01
7.31035769e-01 5.26031137e-01 7.77496547e-02 -7.80360281e-01
-7.46225147e-03 3.99594098e-01 1.06866978e-01 -6.12995744e-01
6.44191802e-01 3.20737183e-01 -3.60104531e-01 6.05687201e-01
-2.60795187e-02 -2.14980453e-01 6.73716247e-01 3.48054320e-01
1.03811467e+00 3.47825974e-01 3.41291964e-01 -4.40779269e-01
7.53066182e-01 3.43830794e-01 5.21264553e-01 4.74108607e-01
-1.40368745e-01 7.07960188e-01 6.79829597e-01 -7.96303228e-02
-8.36498320e-01 -9.44686294e-01 -1.13847449e-01 1.27712286e+00
5.86536014e-03 -6.14912570e-01 -8.43798935e-01 -6.40590310e-01
-4.83074218e-01 9.01147366e-01 -3.14722776e-01 1.15081668e-01
-8.01720381e-01 -7.16437280e-01 8.25542629e-01 5.86136460e-01
9.38285351e-01 -1.43234932e+00 -2.22279966e-01 1.90951362e-01
-8.35609555e-01 -1.36637044e+00 -7.81984568e-01 -5.02052009e-02
-4.37302142e-01 -1.01379442e+00 -8.72465551e-01 -1.25956535e+00
3.55987072e-01 -4.04685475e-02 1.50606322e+00 2.30420977e-01
-3.37378797e-03 7.53369331e-02 -1.00640678e+00 -6.53203487e-01
-4.61981505e-01 5.98658621e-01 -2.75735229e-01 -5.37258446e-01
6.66518807e-01 6.55928776e-02 -3.27100724e-01 2.06986770e-01
-7.22372591e-01 3.43236864e-01 4.33398962e-01 8.13976169e-01
3.33843946e-01 -3.70759338e-01 4.99234915e-01 -1.07169163e+00
9.62806463e-01 -6.27395034e-01 -4.06715125e-01 8.00336897e-01
-1.94645599e-01 -7.21861571e-02 8.07145476e-01 -1.60262823e-01
-1.21648538e+00 -5.10537207e-01 -5.92548728e-01 5.72212517e-01
-2.51113653e-01 1.07404292e+00 -3.48753631e-01 4.09185410e-01
3.56158048e-01 3.04647863e-01 -4.03021157e-01 -7.67232358e-01
4.85117465e-01 8.18711579e-01 4.40903991e-01 -8.27391326e-01
2.20886737e-01 -1.41507357e-01 -6.51045561e-01 -9.75718498e-01
-1.22037625e+00 -4.04290706e-01 -7.88399279e-01 1.59963444e-01
1.02137387e+00 -8.14395666e-01 -6.22105658e-01 5.54149330e-01
-1.42609525e+00 -5.78575015e-01 3.41792017e-01 4.98202145e-01
-1.37025997e-01 4.88806844e-01 -1.03224349e+00 -2.04086170e-01
-6.37557447e-01 -1.12594485e+00 9.10339713e-01 2.17850208e-01
-2.75145859e-01 -1.15219235e+00 1.50995195e-01 4.12773788e-01
3.56217891e-01 -1.16267418e-02 1.23077857e+00 -6.54942453e-01
-3.97836387e-01 -1.23673212e-02 -3.00172299e-01 5.69487631e-01
-9.62598100e-02 2.39345189e-02 -6.10623002e-01 -7.94650614e-02
-2.04287738e-01 -8.59999359e-01 7.64368653e-01 3.82622071e-02
1.18195081e+00 -1.67634159e-01 -2.67596394e-01 2.66740978e-01
1.24724436e+00 2.26576358e-01 5.81945539e-01 4.47091073e-01
5.04104614e-01 6.28312290e-01 6.22585952e-01 -6.41289772e-03
7.70614028e-01 3.86361867e-01 -3.38445678e-02 2.87736896e-02
1.06053306e-02 -4.58978564e-01 4.86803293e-01 1.50729191e+00
-8.15582350e-02 -6.86546564e-01 -1.24983871e+00 5.59548676e-01
-1.75837338e+00 -6.03742242e-01 2.20433529e-02 1.85491586e+00
1.26177621e+00 -3.74820530e-02 -3.40259194e-01 -3.89498085e-01
5.05273402e-01 1.96546674e-01 -2.74388433e-01 -4.09520566e-01
-4.61789072e-01 2.31475174e-01 1.10450201e-01 4.65855002e-01
-1.07687533e+00 1.43683445e+00 5.52394819e+00 1.02947462e+00
-1.11837769e+00 9.91107598e-02 8.43379617e-01 3.03812385e-01
-2.05079079e-01 4.63325270e-02 -1.11735094e+00 5.70530951e-01
9.73896384e-01 -3.54682207e-01 2.16232821e-01 6.62380040e-01
2.29503438e-01 -6.87144771e-02 -1.08676159e+00 4.66838300e-01
2.21886873e-01 -1.38334608e+00 -5.98192476e-02 -4.54585016e-01
7.57562101e-01 5.24378240e-01 -9.81279835e-02 6.22087300e-01
5.24868608e-01 -1.00726819e+00 4.73390013e-01 1.73043132e-01
7.23695099e-01 -4.82490450e-01 8.47830355e-01 5.90375304e-01
-1.04599500e+00 3.61351728e-01 -4.12352592e-01 -9.15361121e-02
3.97609442e-01 3.30079228e-01 -3.79920930e-01 7.32939422e-01
8.10256243e-01 9.36771989e-01 -6.01496994e-01 8.48943055e-01
-9.25901711e-01 1.03771830e+00 -3.26493755e-02 -5.20282805e-01
5.19799471e-01 -4.25650895e-01 7.18392804e-02 1.53114462e+00
3.85894328e-01 8.91136378e-02 2.95479000e-01 6.31621063e-01
-4.51237679e-01 5.24248242e-01 -3.25735688e-01 -4.10635352e-01
5.07971227e-01 1.07518351e+00 -5.11488795e-01 -4.77727413e-01
-8.95515382e-01 8.77121866e-01 6.66684806e-01 4.45974708e-01
-6.97748542e-01 -5.06896019e-01 4.24228936e-01 -3.34149599e-01
-3.32316011e-02 -4.30060893e-01 -1.89346403e-01 -1.58159769e+00
1.14091046e-01 -1.15058494e+00 4.81017143e-01 -5.97901821e-01
-1.32062566e+00 7.20098257e-01 -2.01669097e-01 -1.02266705e+00
-1.41225681e-01 -7.99992859e-01 -7.91487932e-01 1.01703131e+00
-1.53865469e+00 -1.10548592e+00 -1.14391118e-01 2.51817465e-01
8.57005835e-01 -1.22848682e-01 9.50535834e-01 3.20250392e-01
-8.72926056e-01 5.76836407e-01 3.96163791e-01 6.26141012e-01
9.45723712e-01 -1.27839959e+00 7.16995835e-01 9.68387544e-01
2.50807285e-01 6.81464970e-01 3.18385959e-01 -5.38826942e-01
-1.20371091e+00 -1.03682697e+00 1.63992047e+00 -6.26221657e-01
9.90948558e-01 -6.19449377e-01 -1.00446844e+00 8.12248349e-01
7.26223052e-01 -5.06676078e-01 7.14056373e-01 2.47428149e-01
-1.49409264e-01 2.11746767e-01 -5.31278133e-01 7.44976938e-01
8.54866564e-01 -5.17583907e-01 -4.42187220e-01 6.75773561e-01
9.78284895e-01 -4.67802912e-01 -9.05402541e-01 2.25568891e-01
2.75914967e-01 -5.37276685e-01 6.08003914e-01 -8.30315769e-01
9.40727293e-01 9.81971398e-02 1.60394698e-01 -1.48451722e+00
-1.13665327e-01 -4.36616778e-01 4.20501083e-01 1.41510785e+00
1.03541636e+00 -5.34508944e-01 6.73053145e-01 4.63342935e-01
-3.88832897e-01 -9.04386520e-01 -5.34378767e-01 -7.69448817e-01
7.26701081e-01 -4.08008486e-01 5.07605672e-01 9.43611503e-01
4.56892103e-01 7.61349022e-01 -3.08106303e-01 -2.19616786e-01
3.09545211e-02 3.34591061e-01 5.51260948e-01 -6.16010487e-01
-1.48148581e-01 -6.24982715e-01 2.55302995e-01 -1.40642154e+00
3.09179544e-01 -1.21733546e+00 1.98218018e-01 -1.60301661e+00
2.29432330e-01 -2.30973378e-01 2.97631919e-02 5.64497173e-01
-5.59428811e-01 -6.92182183e-02 2.03867823e-01 1.91003993e-01
-8.19501758e-01 5.62121570e-01 1.45424736e+00 -1.86012700e-01
-5.98117337e-02 -9.66645554e-02 -6.86049402e-01 5.72892368e-01
1.02462852e+00 -2.56362498e-01 -1.67527974e-01 -1.25266290e+00
2.07198650e-01 -6.82852939e-02 -2.53722489e-01 -6.51741266e-01
2.87216485e-01 -1.71803683e-01 -4.65604961e-02 -5.25775969e-01
-9.61725041e-02 -1.00752629e-01 -5.76391101e-01 4.06845748e-01
-5.82881033e-01 4.96278435e-01 5.23820855e-02 -2.89412756e-02
-7.59763539e-01 -3.35536420e-01 5.50040245e-01 -4.04650152e-01
-8.82332623e-01 3.18865061e-01 -3.74173433e-01 6.77334070e-01
7.49044538e-01 3.83655488e-01 -6.91655099e-01 -3.67943257e-01
-2.89551258e-01 8.23545039e-01 1.68355748e-01 7.15328634e-01
4.02671069e-01 -8.41945052e-01 -1.19227612e+00 -1.55320317e-01
3.04321140e-01 -3.11534908e-02 1.90190151e-01 8.44863713e-01
-8.95521760e-01 9.08475041e-01 -5.81859499e-02 -1.44031256e-01
-1.02289891e+00 3.84806514e-01 -4.48939390e-02 -4.83820945e-01
-5.90183914e-01 9.27151382e-01 1.63518846e-01 -7.73906648e-01
1.82438508e-01 -5.13440013e-01 -3.89240801e-01 -2.77245790e-01
6.14792466e-01 2.81836748e-01 -4.05215435e-02 -4.48887676e-01
-2.95456082e-01 2.86337435e-01 -2.22574413e-01 9.07434151e-02
1.12005019e+00 -1.59132585e-01 -4.63114619e-01 3.54646504e-01
1.24685764e+00 -4.75607626e-03 -6.75449610e-01 -3.36038619e-01
5.52707851e-01 4.83176485e-03 -5.06227434e-01 -1.11985004e+00
-5.89371085e-01 1.38057661e+00 -1.11068763e-01 -1.52888363e-02
9.23621058e-01 1.55552611e-01 1.14335918e+00 5.53552628e-01
3.09875727e-01 -1.26066601e+00 -2.89461948e-02 1.29197979e+00
9.15888011e-01 -1.44716740e+00 -3.72115672e-01 -4.77260679e-01
-8.56798828e-01 9.61512208e-01 6.67097390e-01 2.02966571e-01
6.10240281e-01 2.27751166e-01 2.95158386e-01 1.46188691e-01
-8.65066350e-01 -1.78865388e-01 2.23279253e-01 3.60452443e-01
1.08534908e+00 2.33293563e-01 -7.75651634e-01 8.14144611e-01
-5.55547774e-01 -5.44385472e-03 3.29044402e-01 8.90099943e-01
-3.69885981e-01 -1.43429530e+00 1.84042573e-01 3.50365549e-01
-7.51181364e-01 -7.08766341e-01 -5.21013141e-01 9.10329342e-01
-9.09959674e-02 9.81967211e-01 -1.81792751e-01 -7.52545521e-02
2.10415691e-01 7.97515884e-02 2.29886979e-01 -9.14805293e-01
-5.55503666e-01 -1.91086024e-01 3.77349287e-01 -3.10994148e-01
-2.82450140e-01 -5.74953675e-01 -1.19508493e+00 -2.91745394e-01
-1.70020554e-02 4.92811829e-01 3.72909784e-01 1.06281745e+00
1.60155356e-01 1.30609915e-01 7.82902688e-02 -2.41572753e-01
-3.31242353e-01 -1.34473145e+00 -2.83820271e-01 2.13391811e-01
-1.94389790e-01 -9.77267623e-02 -8.69921371e-02 1.19007222e-01] | [10.881230354309082, 9.255338668823242] |
84dca1a5-5df0-40bf-84ec-64bd638e84ce | predicting-cardiovascular-disease-risk-using | 2305.05648 | null | https://arxiv.org/abs/2305.05648v1 | https://arxiv.org/pdf/2305.05648v1.pdf | Predicting Cardiovascular Disease Risk using Photoplethysmography and Deep Learning | Cardiovascular diseases (CVDs) are responsible for a large proportion of premature deaths in low- and middle-income countries. Early CVD detection and intervention is critical in these populations, yet many existing CVD risk scores require a physical examination or lab measurements, which can be challenging in such health systems due to limited accessibility. Here we investigated the potential to use photoplethysmography (PPG), a sensing technology available on most smartphones that can potentially enable large-scale screening at low cost, for CVD risk prediction. We developed a deep learning PPG-based CVD risk score (DLS) to predict the probability of having major adverse cardiovascular events (MACE: non-fatal myocardial infarction, stroke, and cardiovascular death) within ten years, given only age, sex, smoking status and PPG as predictors. We compared the DLS with the office-based refit-WHO score, which adopts the shared predictors from WHO and Globorisk scores (age, sex, smoking status, height, weight and systolic blood pressure) but refitted on the UK Biobank (UKB) cohort. In UKB cohort, DLS's C-statistic (71.1%, 95% CI 69.9-72.4) was non-inferior to office-based refit-WHO score (70.9%, 95% CI 69.7-72.2; non-inferiority margin of 2.5%, p<0.01). The calibration of the DLS was satisfactory, with a 1.8% mean absolute calibration error. Adding DLS features to the office-based score increased the C-statistic by 1.0% (95% CI 0.6-1.4). DLS predicts ten-year MACE risk comparable with the office-based refit-WHO score. It provides a proof-of-concept and suggests the potential of a PPG-based approach strategies for community-based primary prevention in resource-limited regions. | ['Diego Ardila', 'Goodarz Danaei', 'Yun Liu', 'Shruthi Prabhakara', 'Shravya Shetty', 'Greg S. Corrado', 'Yossi Matias', 'Cory Y. McLean', 'Babak Behsaz', 'Mariam Jabara', 'Sujay Kakarmath', 'Lauren Harrell', 'Christina Chen', 'Mayank Daswani', 'Sebastien Baur', 'Wei-Hung Weng'] | 2023-05-09 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 9.17729586e-02 8.39166045e-02 -4.23660666e-01 -3.08568835e-01
-8.66816223e-01 -1.85338333e-01 3.47919315e-01 5.90758741e-01
-2.06533954e-01 7.97242641e-01 4.68415499e-01 -8.14734340e-01
-8.87256116e-02 -1.23020339e+00 -2.09618762e-01 -3.00599873e-01
-4.92361546e-01 3.30488175e-01 7.45233074e-02 3.65797758e-01
-1.46856219e-01 2.20465600e-01 -7.71374404e-01 1.87921077e-01
9.11896646e-01 9.97039855e-01 -2.42044106e-01 9.01081085e-01
2.62315124e-01 3.75610948e-01 -2.74950951e-01 -2.27132276e-01
2.01389015e-01 -6.45777702e-01 -1.83215454e-01 -7.24601805e-01
5.60933590e-01 -4.98351604e-01 1.43505394e-01 4.10066873e-01
9.80539382e-01 -2.82231688e-01 7.11647570e-01 -8.55141580e-01
-6.22470796e-01 2.97781914e-01 -3.56142074e-01 1.03548445e-01
9.91276577e-02 9.36554745e-02 2.63997436e-01 -3.97195280e-01
1.49967477e-01 9.75213170e-01 1.36341965e+00 7.07500994e-01
-1.26477218e+00 -5.86966455e-01 -4.75459516e-01 -2.01556787e-01
-1.26182556e+00 -5.23315609e-01 1.37797192e-01 -6.47993267e-01
7.22841620e-01 6.05656326e-01 1.01833153e+00 7.20143437e-01
6.55287206e-01 -3.34809691e-01 1.20962942e+00 -4.36192542e-01
2.42800564e-01 1.01309344e-01 -1.16451993e-03 5.97768545e-01
6.38953984e-01 4.65317070e-01 1.66836947e-01 -8.47149789e-01
1.10180771e+00 3.84054005e-01 2.74946809e-01 2.37558968e-02
-1.05050898e+00 7.40415335e-01 9.33415219e-02 6.12667948e-02
-3.99968833e-01 1.74664557e-02 3.57792526e-01 1.47005394e-01
5.84849954e-01 -1.39961895e-02 -8.04016352e-01 -2.31813207e-01
-9.65372026e-01 5.41036278e-02 8.09882641e-01 -3.81296836e-02
-5.10521606e-02 -8.87587965e-02 -3.08635086e-01 7.95027256e-01
7.25378752e-01 9.97608662e-01 2.21452877e-01 -1.09216738e+00
1.47345304e-01 6.53675616e-01 1.85647801e-01 -1.01172829e+00
-8.41161311e-01 -5.31079590e-01 -1.37533212e+00 1.82718843e-01
7.72055686e-01 -6.36792362e-01 -7.04343975e-01 1.73238564e+00
2.84778148e-01 3.39912564e-01 -9.11395326e-02 7.31878221e-01
7.73728907e-01 2.24296510e-01 6.96716607e-01 -3.76355797e-01
1.60678625e+00 -1.17269345e-01 6.61101658e-03 -9.69144106e-02
5.35685003e-01 -3.52160513e-01 7.66797662e-01 1.12997465e-01
-8.22008789e-01 -5.62704563e-01 -7.68652320e-01 1.75845668e-01
-3.34609568e-01 1.12598017e-01 4.30653840e-01 1.68564796e+00
-9.33435678e-01 6.05143964e-01 -1.05208516e+00 -5.73072374e-01
5.67983389e-01 2.15548664e-01 4.83171195e-02 -4.85680485e-03
-1.31106329e+00 7.97983050e-01 -1.14460923e-01 -2.23554701e-01
-4.94956821e-01 -1.45448184e+00 -6.77669525e-01 -8.74507353e-02
-1.43512681e-01 -1.11540771e+00 4.95726913e-01 -3.16401303e-01
-1.52268863e+00 9.18313861e-01 -2.18945518e-01 -4.04160231e-01
5.04221976e-01 -3.66018862e-01 -7.97052801e-01 2.74809808e-01
3.08061838e-01 1.38810962e-01 3.16516787e-01 -3.98078382e-01
-3.64962846e-01 -7.89107084e-01 -5.19711971e-01 -2.04452962e-01
1.00518495e-01 1.65540710e-01 5.20745099e-01 -5.34258068e-01
-2.23672658e-01 -7.09585547e-01 -7.09965646e-01 3.33283544e-01
-7.73126259e-02 1.23119108e-01 1.91477627e-01 -1.12829614e+00
1.12120819e+00 -1.57978106e+00 -7.49789298e-01 3.52945507e-01
5.74715436e-01 6.53457165e-01 3.07222664e-01 1.08869724e-01
-5.39251529e-02 6.63111091e-01 -1.11364469e-01 1.85412943e-01
-2.50570595e-01 -2.99209356e-01 2.82383651e-01 4.33614135e-01
1.69699013e-01 1.02600622e+00 -8.81572664e-01 -3.83737087e-01
8.66721034e-01 8.38302672e-01 -2.96244234e-01 -3.20553511e-01
6.08335912e-01 5.05495131e-01 -4.20483530e-01 4.36370611e-01
6.14322484e-01 -1.93463326e-01 2.64387041e-01 2.07721338e-01
-3.07717651e-01 9.22287330e-02 -9.40397382e-01 1.29592884e+00
-2.63058394e-01 3.39679271e-01 -1.78614736e-01 -1.02706635e+00
1.43413913e+00 4.59237814e-01 4.81003821e-01 -7.73962736e-01
-2.42485479e-03 2.11585194e-01 3.10749948e-01 -3.53338867e-01
-8.06118071e-01 -6.72503769e-01 1.54548764e-01 1.83517396e-01
-5.93644679e-01 7.45488048e-01 -1.56171337e-01 6.08362444e-02
9.78015244e-01 -1.67962953e-01 7.00212121e-01 -6.61564052e-01
7.52746403e-01 -1.87363610e-01 6.96367145e-01 1.19420528e+00
-5.52997112e-01 6.94063604e-01 6.39042497e-01 -8.11634123e-01
-7.78503120e-01 -1.43112600e+00 -7.11592197e-01 5.46144664e-01
-5.38278759e-01 -4.53819245e-01 -1.92592770e-01 -4.23468500e-01
2.34295011e-01 2.82753080e-01 -5.46127677e-01 -2.65122682e-01
-4.52797681e-01 -1.43157732e+00 8.20978403e-01 4.84115928e-01
7.34595954e-01 -3.11388910e-01 -9.83477116e-01 5.65894365e-01
1.71404421e-01 -3.16398591e-01 1.44425541e-01 -5.04910350e-01
-1.09926021e+00 -1.23739636e+00 -1.42303884e+00 -2.60904789e-01
-2.56830119e-02 -4.45416540e-01 1.33612955e+00 6.82114363e-02
-4.04525638e-01 1.82733148e-01 -7.81223103e-02 -6.20365143e-01
-4.95584160e-01 -1.63540453e-01 6.92599788e-02 -2.01626569e-01
6.38898015e-01 -7.06951976e-01 -1.50133264e+00 6.80159181e-02
2.28603154e-01 -1.36728555e-01 6.01978004e-01 2.68757105e-01
4.61848021e-01 -5.50178587e-01 1.18270099e+00 -6.99052870e-01
3.55792530e-02 -5.96711755e-01 -3.38486105e-01 1.36821032e-01
-1.06816316e+00 -7.27736354e-01 3.72422010e-01 -3.36790562e-01
-6.59526885e-01 -2.45100513e-01 -7.81097859e-02 4.05533075e-01
-4.61746603e-01 4.76472884e-01 1.55203983e-01 3.29036564e-01
7.63033867e-01 -4.58991528e-02 3.78994673e-01 -6.96725607e-01
-2.34495580e-01 8.09370399e-01 1.96726516e-01 -5.06034315e-01
2.90233344e-01 3.03227305e-01 6.37845039e-01 -9.86274242e-01
-3.36359471e-01 -2.27632299e-01 -3.52127224e-01 4.84961793e-02
1.00067031e+00 -1.12947273e+00 -9.66642737e-01 4.01482165e-01
-4.66928750e-01 -6.23175085e-01 5.60583882e-02 8.43528986e-01
-1.26248062e-01 4.14075553e-01 -4.98775184e-01 -9.39424098e-01
-8.99975657e-01 -5.25132537e-01 7.30889022e-01 3.57545406e-01
-4.63256299e-01 -1.18658233e+00 4.23033893e-01 6.18402548e-02
8.51907492e-01 1.31290376e+00 9.61293340e-01 -4.60077584e-01
2.69334048e-01 -4.15696681e-01 -6.90958440e-01 2.31299326e-01
3.17826062e-01 -2.97383875e-01 -3.41274410e-01 6.66690320e-02
-2.91064143e-01 3.71181250e-01 3.82454753e-01 1.12194157e+00
6.32627547e-01 -1.83324069e-02 -5.72920561e-01 4.20557588e-01
1.43870759e+00 4.70003605e-01 7.42215931e-01 2.26004288e-01
4.45956290e-01 1.46308869e-01 -2.61648029e-01 4.37861651e-01
4.79852378e-01 5.22277057e-01 -1.58732548e-01 -3.43800932e-01
-3.17076594e-01 2.19682425e-01 2.20673695e-01 -1.52749076e-01
-4.49119538e-01 5.07036984e-01 -1.23776650e+00 3.23567301e-01
-1.30622709e+00 -8.02332163e-01 -9.34639871e-01 2.50895286e+00
7.24716187e-01 1.28026873e-01 6.33671463e-01 -9.88581255e-02
7.22263873e-01 -2.99106747e-01 -4.54540998e-01 -7.70524263e-01
1.51017800e-01 4.91584241e-01 4.13907677e-01 3.97594601e-01
-1.02748871e+00 -8.44772086e-02 6.13029289e+00 -1.92265138e-01
-1.00172913e+00 2.47153968e-01 1.15699530e+00 8.24420229e-02
4.81511921e-01 3.91098894e-02 -5.12251616e-01 7.46136069e-01
1.75484920e+00 -2.61790067e-01 3.28143798e-02 4.43928450e-01
6.41223907e-01 -2.30980426e-01 -9.11985815e-01 6.10178113e-01
-2.74029791e-01 -1.41893840e+00 -5.49739480e-01 2.66401976e-01
4.47524488e-01 1.65452227e-01 -1.47129402e-01 2.74233431e-01
2.02745155e-01 -8.74783695e-01 -1.02340057e-02 8.12453151e-01
1.47753072e+00 -4.11990851e-01 9.79492545e-01 1.30012408e-02
-1.12852895e+00 2.29347661e-01 -2.67011017e-01 -6.96058989e-01
4.25164431e-01 1.26903999e+00 -6.67014003e-01 3.59461755e-01
8.12872708e-01 6.36341333e-01 -6.70033932e-01 1.00456631e+00
3.53103459e-01 1.03751469e+00 -5.86827755e-01 1.03869885e-01
-1.32767543e-01 -4.39972989e-02 1.78157926e-01 1.16307473e+00
6.18869126e-01 2.24603698e-01 2.11538095e-02 9.00751770e-01
5.01335502e-01 2.03387722e-01 -2.29323685e-01 2.00782046e-01
3.17600697e-01 7.87110269e-01 -5.73805988e-01 -6.02214694e-01
-8.17380548e-01 2.72031516e-01 -4.99569684e-01 -2.77910754e-03
-5.44888675e-01 -4.63355631e-01 4.92129147e-01 7.32732236e-01
-2.84013778e-01 3.28148454e-01 -8.43771100e-01 -9.11990821e-01
-3.09792787e-01 -5.56229770e-01 7.69807458e-01 -1.59096062e-01
-1.22617376e+00 -1.18276693e-01 3.32359038e-02 -8.00476968e-01
-1.84555098e-01 -4.34062213e-01 -8.30784976e-01 1.52061248e+00
-1.30347025e+00 -9.23976004e-01 -2.92586088e-01 1.63557932e-01
-9.03801918e-02 -1.11912802e-01 1.37915277e+00 2.34815896e-01
-7.27824032e-01 5.57752669e-01 9.94404033e-02 3.77378196e-01
6.81595147e-01 -1.32418108e+00 2.64955342e-01 3.71961325e-01
-8.75622153e-01 8.59103560e-01 1.26013592e-01 -1.10933161e+00
-8.31979811e-01 -1.26522374e+00 1.33011615e+00 -6.32577062e-01
2.11421624e-01 5.38467988e-02 -2.88360864e-01 3.17600161e-01
-4.46543664e-01 1.09776326e-01 1.33502686e+00 4.49496359e-01
-1.77105770e-01 -3.15184981e-01 -1.63071609e+00 4.15129334e-01
6.19226277e-01 -8.60026702e-02 -5.27258098e-01 1.19983166e-01
1.23431392e-01 -2.54845619e-02 -1.55355108e+00 4.29045737e-01
1.44095898e+00 -1.18095493e+00 1.64335418e+00 -5.74462056e-01
3.67594302e-01 -3.25610191e-02 2.98396051e-02 -6.69177473e-01
-6.20766461e-01 -5.10058761e-01 -7.90055990e-02 1.21939039e+00
2.39678696e-01 -1.03167605e+00 6.69515848e-01 1.19414175e+00
2.01392129e-01 -6.62719667e-01 -1.16940260e+00 -5.77143908e-01
5.30203938e-01 -2.22971857e-01 5.59378147e-01 7.99071968e-01
7.23379478e-02 8.62642750e-02 -2.38360614e-01 1.15958154e-01
8.12095165e-01 -1.73217982e-01 4.47121680e-01 -1.87109709e+00
-7.18674958e-02 -3.11149627e-01 -3.29257071e-01 1.45720690e-01
-7.33334661e-01 -8.00251961e-01 -9.57898021e-01 -1.89477968e+00
4.28313643e-01 -5.87341309e-01 -7.20306337e-01 5.67175627e-01
-2.92856842e-01 1.91651061e-01 5.64151295e-02 -2.31196985e-01
1.71786875e-01 -1.53842688e-01 6.86104119e-01 7.72665516e-02
-8.03032815e-01 1.50072381e-01 -8.74510825e-01 5.99092543e-01
1.40246058e+00 -4.67388690e-01 -2.64189929e-01 1.00950614e-01
1.29318580e-01 3.52739662e-01 8.93455565e-01 -1.17801642e+00
-4.03392494e-01 -2.70963252e-01 1.27511597e+00 -2.12890461e-01
-5.29169738e-02 -3.96856338e-01 4.80115175e-01 1.35974216e+00
-1.35099456e-01 -3.75075221e-01 7.62023106e-02 3.20811450e-01
7.54767001e-01 3.12825024e-01 7.06648886e-01 -3.61940652e-01
2.53347550e-02 -2.04088129e-02 -7.44990647e-01 1.84445143e-01
6.13954365e-01 -2.67513394e-01 -5.62923253e-01 -5.43930009e-02
-1.04945302e+00 -1.62604287e-01 -3.46379966e-04 1.44530028e-01
2.57437974e-01 -1.25947738e+00 -1.10660470e+00 1.94695026e-01
-1.29571006e-01 -4.39870447e-01 3.32609773e-01 1.27696264e+00
-6.43578351e-01 4.72159505e-01 -2.65342087e-01 -3.02386045e-01
-9.22604978e-01 4.39272493e-01 5.90122521e-01 -1.96693078e-01
-1.03128767e+00 1.61819085e-01 -6.28366470e-02 -6.81692138e-02
-1.37163140e-02 -2.34543294e-01 -7.66270459e-02 -1.72456101e-01
1.08904850e+00 1.14605820e+00 -2.89488107e-01 -3.01652849e-01
-8.24229062e-01 6.56883419e-01 3.45711201e-01 5.29797934e-02
1.20504332e+00 -3.63219023e-01 -3.54818394e-03 3.67723495e-01
1.01020920e+00 -1.89655259e-01 -6.52971029e-01 2.12520331e-01
-2.07617462e-01 -1.09836183e-01 1.91634342e-01 -1.45057690e+00
-7.94153333e-01 8.43887150e-01 1.54854751e+00 2.28270844e-01
8.61466229e-01 -8.77821371e-02 7.65559912e-01 -1.26355201e-01
-2.08557546e-02 -5.76757610e-01 -6.68174148e-01 -5.27360924e-02
4.31066185e-01 -9.56450522e-01 1.56468183e-01 -2.57855028e-01
-1.06988832e-01 9.40279186e-01 1.49394900e-01 -2.06883714e-01
1.23500371e+00 -1.36626735e-01 2.15469211e-01 2.61031445e-02
-3.33557665e-01 6.60253018e-02 2.45560437e-01 1.28672695e+00
6.45766020e-01 7.13250875e-01 -8.68656695e-01 1.00566900e+00
7.90651515e-02 6.55090153e-01 1.91132605e-01 5.41600287e-01
-4.08973247e-01 -1.07673275e+00 -4.38566834e-01 1.20350027e+00
-9.24374282e-01 -1.25947565e-01 -2.68916905e-01 6.48245811e-01
5.23928523e-01 1.11548579e+00 3.62920351e-02 1.18311062e-01
2.10758552e-01 4.94698405e-01 3.97424996e-02 -5.11530757e-01
-5.43988705e-01 1.42428771e-01 3.30292046e-01 -4.50717092e-01
-5.79149067e-01 -9.03840899e-01 -7.97422051e-01 -3.69908720e-01
2.85741717e-01 -1.60996497e-01 2.45553598e-01 8.27286065e-01
6.67035937e-01 2.62394875e-01 2.20557630e-01 -4.16186124e-01
-2.18403757e-01 -9.99881685e-01 -6.62184536e-01 -7.86424130e-02
3.99502188e-01 -3.95956755e-01 -2.24137768e-01 5.41307516e-02] | [14.095785140991211, 3.07401442527771] |
3267484e-afc3-43b1-b39c-a84c2a66dea9 | npb-rec-non-parametric-assessment-of | 2208.03966 | null | https://arxiv.org/abs/2208.03966v1 | https://arxiv.org/pdf/2208.03966v1.pdf | NPB-REC: Non-parametric Assessment of Uncertainty in Deep-learning-based MRI Reconstruction from Undersampled Data | Uncertainty quantification in deep-learning (DL) based image reconstruction models is critical for reliable clinical decision making based on the reconstructed images. We introduce "NPB-REC", a non-parametric fully Bayesian framework for uncertainty assessment in MRI reconstruction from undersampled "k-space" data. We use Stochastic gradient Langevin dynamics (SGLD) during the training phase to characterize the posterior distribution of the network weights. We demonstrated the added-value of our approach on the multi-coil brain MRI dataset, from the fastmri challenge, in comparison to the baseline E2E-VarNet with and without inference-time dropout. Our experiments show that NPB-REC outperforms the baseline by means of reconstruction accuracy (PSNR and SSIM of $34.55$, $0.908$ vs. $33.08$, $0.897$, $p<0.01$) in high acceleration rates ($R=8$). This is also measured in regions of clinical annotations. More significantly, it provides a more accurate estimate of the uncertainty that correlates with the reconstruction error, compared to the Monte-Carlo inference time Dropout method (Pearson correlation coefficient of $R=0.94$ vs. $R=0.91$). The proposed approach has the potential to facilitate safe utilization of DL based methods for MRI reconstruction from undersampled data. Code and trained models are available in \url{https://github.com/samahkh/NPB-REC}. | ['Moti Freiman', 'Samah Khawaled'] | 2022-08-08 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [-1.56549767e-01 1.34558231e-01 2.08619133e-01 -5.99481702e-01
-1.37529695e+00 -6.21606829e-03 6.18953295e-02 5.37512191e-02
-9.00937021e-01 1.10906553e+00 1.43487945e-01 -3.73130590e-01
-4.75469768e-01 -4.27936941e-01 -1.05267084e+00 -8.82248580e-01
-5.85258424e-01 5.46855330e-01 -6.31814301e-02 3.82987767e-01
-2.69752536e-02 3.33212107e-01 -4.86192495e-01 9.37945247e-02
6.39288545e-01 1.09203696e+00 2.69047111e-01 3.25054586e-01
5.15422523e-01 7.57530510e-01 -1.66217729e-01 -3.68835002e-01
5.41440072e-03 -2.28605136e-01 -7.04879880e-01 -4.66430873e-01
-1.41226888e-01 -8.00933361e-01 -5.12510478e-01 1.29078722e+00
9.36491430e-01 2.41523236e-01 7.22907782e-01 -5.58739603e-01
8.82959664e-02 9.34359014e-01 -5.86707830e-01 7.18484223e-01
-2.13710949e-01 3.31282675e-01 2.85540074e-01 -6.93545341e-01
6.79348111e-01 8.16333413e-01 9.16551709e-01 6.20644152e-01
-1.57626390e+00 -9.61609662e-01 -3.79462987e-01 1.55060515e-01
-1.57677925e+00 -4.60623652e-01 3.92718643e-01 -5.52825153e-01
9.91685867e-01 -2.72685159e-02 4.24310654e-01 1.20127916e+00
9.04952049e-01 4.35822368e-01 1.46277928e+00 -2.60794833e-02
4.90169913e-01 -1.19028889e-01 3.43314767e-01 7.14816511e-01
4.01586592e-01 4.75980997e-01 -4.08717304e-01 -2.79328942e-01
9.85336721e-01 -3.21718544e-01 -4.30228710e-01 -1.29384026e-01
-9.84481573e-01 8.21155488e-01 5.47058642e-01 6.45049736e-02
-5.95030248e-01 7.95950234e-01 6.37776434e-01 -8.36052299e-02
5.31661570e-01 2.27749914e-01 -2.85526603e-01 -2.64683574e-01
-1.08516943e+00 2.58160174e-01 4.85495865e-01 7.49582410e-01
1.82201996e-01 3.43708426e-01 -3.49421114e-01 7.03647971e-01
4.45871472e-01 6.21201515e-01 3.48458737e-01 -1.21800554e+00
8.38513300e-02 -4.89971966e-01 1.14107810e-01 -5.46240091e-01
-7.21212626e-01 -8.06965768e-01 -1.13052309e+00 1.37024686e-01
3.51094097e-01 -3.61058891e-01 -8.99305820e-01 1.84814262e+00
-2.88602505e-02 3.96488816e-01 -2.81406403e-01 8.77307057e-01
7.41712272e-01 3.17367762e-01 1.07246324e-01 -4.06758010e-01
1.37114382e+00 -4.13411677e-01 -8.75568449e-01 3.70164029e-02
4.35066104e-01 -3.63985479e-01 8.37944567e-01 5.67152679e-01
-1.40198398e+00 -4.86794114e-02 -9.60566819e-01 5.23819327e-01
4.77821827e-01 -4.87630740e-02 4.28114384e-01 8.70434463e-01
-1.21337044e+00 9.15169001e-01 -1.51482129e+00 4.88753587e-01
8.96006048e-01 5.19531548e-01 -2.98644751e-02 -4.14741397e-01
-1.24916935e+00 9.70495462e-01 1.41422436e-01 2.15152830e-01
-1.57278991e+00 -1.18334532e+00 -6.17142737e-01 -2.25744858e-01
5.57082668e-02 -6.23680830e-01 1.25466788e+00 -2.74058342e-01
-1.36882484e+00 5.95715046e-01 1.21165976e-01 -9.64907706e-01
8.28254700e-01 -1.91530079e-01 -4.99413088e-02 4.30545896e-01
-4.70893681e-02 6.61909580e-01 5.46915352e-01 -8.64931047e-01
1.95041180e-01 -4.93216723e-01 -3.98771971e-01 -6.03141300e-02
4.97337461e-01 1.38422072e-01 -1.66222721e-01 -5.30685365e-01
1.64937511e-01 -1.02501965e+00 -6.21231437e-01 -1.04541697e-01
-3.64587307e-01 3.69882733e-01 -2.23102733e-01 -1.24695575e+00
8.86335790e-01 -1.87862039e+00 -3.13869476e-01 3.31767261e-01
3.24292034e-01 -1.37400344e-01 2.83436090e-01 -2.39025325e-01
-1.63382977e-01 4.94723860e-03 -6.56507015e-01 -1.80160478e-01
-2.33938172e-01 -1.04383819e-01 2.62488842e-01 8.40533853e-01
-5.40199541e-02 8.79770517e-01 -8.62572610e-01 -3.25797230e-01
1.08471036e-01 8.45873058e-01 -6.32510602e-01 -7.66797811e-02
2.92391658e-01 8.26960146e-01 -3.64064097e-01 2.31920823e-01
8.46363008e-01 -1.96437925e-01 3.58133078e-01 -4.09073204e-01
2.63244420e-01 1.25907706e-02 -1.04800010e+00 2.09546995e+00
-3.83486778e-01 3.35169762e-01 2.23915219e-01 -9.55360651e-01
5.60314894e-01 4.38322306e-01 6.44420207e-01 -6.42245233e-01
4.28007096e-01 3.51924062e-01 3.94768775e-01 -3.14031631e-01
-1.00121461e-01 -5.32703221e-01 4.64680642e-02 3.54140729e-01
2.34503791e-01 -2.65142590e-01 -1.39813945e-01 2.84677416e-01
1.38907790e+00 7.49235824e-02 -6.89181983e-02 -8.88554752e-01
1.41810244e-02 -2.24306121e-01 6.47333801e-01 1.01579285e+00
-6.07511580e-01 4.92792785e-01 4.41377789e-01 -1.30889773e-01
-9.77463186e-01 -1.44553232e+00 -8.90837729e-01 9.08304751e-02
-2.62854278e-01 1.07029617e-01 -1.03012109e+00 -2.88257360e-01
-2.18007579e-01 9.06228721e-01 -3.95352572e-01 -2.29175001e-01
-5.83896577e-01 -1.46001601e+00 5.88544667e-01 5.06242931e-01
3.52995276e-01 -8.22461486e-01 -7.51159668e-01 3.94134253e-01
-3.47196400e-01 -9.95351374e-01 -2.75728196e-01 2.89373338e-01
-1.33235145e+00 -7.34262705e-01 -8.19709957e-01 -9.02305767e-02
6.82963669e-01 -6.49408877e-01 1.10758603e+00 -3.18573773e-01
-6.16965592e-01 4.36764985e-01 -4.73661534e-03 -1.29577070e-01
-3.29153001e-01 -4.51431483e-01 2.69370377e-01 -5.07783055e-01
1.01262936e-02 -6.40591919e-01 -1.10623574e+00 -8.21223408e-02
-7.84443259e-01 -7.10283965e-02 4.89423752e-01 1.24344862e+00
8.87323260e-01 -8.37942734e-02 5.99197030e-01 -8.74018848e-01
6.13868952e-01 -5.15179038e-01 -6.97133303e-01 -1.53688520e-01
-8.78430843e-01 3.20841402e-01 5.30308485e-02 -3.37416142e-01
-1.07352161e+00 -1.24158740e-01 -3.86577278e-01 -4.53680664e-01
1.58110365e-01 6.52663827e-01 3.69772196e-01 -4.80883308e-02
7.10751295e-01 -1.98142454e-02 1.58538401e-01 -3.37043047e-01
5.95734268e-02 1.88446075e-01 3.58213812e-01 -7.46338964e-01
-5.99470660e-02 5.45970976e-01 1.12910703e-01 -4.63760227e-01
-2.50635505e-01 4.40381020e-02 -1.66963279e-01 -3.58521223e-01
8.30754459e-01 -1.05912399e+00 -7.40037620e-01 4.16071981e-01
-8.29204142e-01 -6.08457685e-01 -3.07613403e-01 1.26436603e+00
-8.54005277e-01 3.41975033e-01 -1.15576303e+00 -8.24039519e-01
-8.45511258e-01 -1.79977715e+00 4.94406700e-01 -1.20016642e-01
-1.28844991e-01 -8.42565775e-01 -6.46312535e-02 3.51753056e-01
6.67037785e-01 4.21795577e-01 8.89377594e-01 -4.20947671e-01
-4.92965102e-01 -1.20130837e-01 -2.41176471e-01 5.60990453e-01
-4.71191555e-01 -5.92398465e-01 -9.17395234e-01 -5.29873669e-01
4.93958503e-01 -3.23399991e-01 7.91948140e-01 1.42637181e+00
1.35418463e+00 1.38022816e-02 -1.13546863e-01 6.19444489e-01
1.68630171e+00 2.93583453e-01 8.26057613e-01 -1.96243469e-02
1.36714429e-01 2.05183566e-01 1.72146171e-01 6.71559811e-01
8.45128745e-02 4.83893842e-01 3.58312845e-01 3.05060595e-01
-2.73421854e-01 1.61474824e-01 2.28315622e-01 1.00671542e+00
5.29326349e-02 3.37511390e-01 -1.26992881e+00 4.62093174e-01
-1.43445897e+00 -4.69594240e-01 -1.91031903e-01 2.38652182e+00
1.06661332e+00 3.44149798e-01 -2.18394682e-01 -3.64503324e-01
7.05197215e-01 -2.83324450e-01 -7.23150969e-01 -2.25792587e-01
2.26762086e-01 6.89796686e-01 8.20172369e-01 6.14138722e-01
-7.10664928e-01 4.12826985e-01 5.91283655e+00 1.12659383e+00
-8.40033114e-01 8.77489269e-01 1.25583768e+00 -4.81104553e-01
-4.64030281e-02 -1.38666347e-01 -5.06269634e-01 6.69573009e-01
1.53717053e+00 1.84092054e-03 5.73669612e-01 4.67223197e-01
5.08101583e-01 -6.51906669e-01 -8.07393551e-01 1.13016331e+00
-1.53220698e-01 -1.44347537e+00 -5.84363341e-01 -1.01557085e-02
6.00823522e-01 5.89078248e-01 8.80680382e-02 1.62588239e-01
4.44661885e-01 -1.14707470e+00 6.20798707e-01 7.56896317e-01
1.20585883e+00 -8.75271022e-01 8.20244431e-01 2.27818504e-01
-4.72654015e-01 3.57646942e-01 -3.27855945e-01 2.63672233e-01
4.68126148e-01 1.21614397e+00 -8.44265163e-01 4.94854808e-01
1.09861350e+00 1.35590166e-01 4.35621031e-02 1.02651906e+00
-1.46381944e-01 7.78469265e-01 -3.98508042e-01 2.48557389e-01
1.98326316e-02 -1.61069825e-01 5.86343586e-01 1.18123174e+00
2.59859681e-01 2.92494506e-01 -2.32922047e-01 1.14287281e+00
3.55695188e-02 -1.64894208e-01 -1.15491688e-01 3.74643087e-01
2.46228740e-01 1.01618373e+00 -7.75197327e-01 -2.69595832e-01
1.18175663e-01 5.58254182e-01 1.55032471e-01 2.54993439e-01
-1.11048198e+00 -2.04310820e-01 1.53347418e-01 1.63775325e-01
-7.00559616e-02 -1.74436867e-01 -2.99834877e-01 -1.00606668e+00
-1.57521874e-01 -6.50476933e-01 2.25041032e-01 -8.90278220e-01
-1.15037024e+00 8.16632807e-01 4.45573300e-01 -6.50989056e-01
-2.50597149e-01 -6.07316554e-01 -3.85217406e-02 1.12660861e+00
-1.26707304e+00 -4.70380604e-01 1.55457854e-01 4.38407093e-01
1.37128562e-01 2.36583993e-01 8.17156553e-01 4.63370800e-01
-4.73797113e-01 5.94980955e-01 4.69342321e-01 1.17137179e-01
3.24990988e-01 -8.90907764e-01 3.80471647e-02 7.90815234e-01
-3.51278365e-01 5.91644883e-01 7.98878551e-01 -7.39346802e-01
-1.15595210e+00 -8.49588037e-01 1.35907635e-01 -2.48798847e-01
6.08327210e-01 -9.74209309e-02 -7.29080260e-01 4.84524161e-01
3.44074406e-02 3.15296441e-01 6.06135190e-01 -2.68497586e-01
-9.23594832e-03 -9.64968726e-02 -1.67038202e+00 1.87267661e-01
7.50775039e-01 -4.43213552e-01 -2.61711448e-01 4.21854615e-01
4.81664747e-01 -7.28411198e-01 -1.39341843e+00 6.92059040e-01
5.77509940e-01 -8.35846901e-01 9.76079583e-01 -2.48284653e-01
2.70321190e-01 1.69283852e-01 -1.93668813e-01 -1.21582544e+00
-1.92135632e-01 -3.86887342e-01 3.82577181e-02 5.68445921e-01
4.55574900e-01 -5.82785308e-01 7.56625414e-01 9.62546229e-01
-3.51758927e-01 -9.01556432e-01 -1.66028559e+00 -6.87623560e-01
5.01152158e-01 -9.62182581e-01 2.67961361e-02 6.73121333e-01
-1.72185808e-01 -3.86941105e-01 -2.82383770e-01 1.08822025e-01
1.13623774e+00 -5.97007871e-01 -2.63552397e-01 -7.00315356e-01
-5.78010738e-01 -1.77012563e-01 -2.52374679e-01 -4.22571003e-01
-5.67108579e-03 -1.07984281e+00 1.08404778e-01 -1.36032736e+00
6.28986418e-01 -6.72300756e-01 -6.19608104e-01 4.34289798e-02
1.12668365e-01 1.79865539e-01 -1.15413286e-01 1.19319707e-01
-1.88871741e-01 4.21051115e-01 1.14326704e+00 5.56976087e-02
1.93928331e-01 -1.08620614e-01 -2.11509928e-01 4.86302495e-01
9.41342652e-01 -9.18209016e-01 -3.26235592e-01 -4.70101744e-01
1.07854724e-01 6.10646069e-01 5.48895299e-01 -1.15225244e+00
3.11347248e-04 3.25497746e-01 4.05471474e-01 -4.03114617e-01
3.90667439e-01 -4.99564797e-01 6.46547556e-01 1.04976857e+00
-2.83339649e-01 2.15925146e-02 5.03801703e-01 4.48048353e-01
2.42666960e-01 -4.68152434e-01 1.10873830e+00 -4.42803234e-01
-1.20791472e-01 3.31095308e-01 -3.81880969e-01 2.51556069e-01
5.32438159e-01 1.77478194e-01 5.74767310e-03 -2.16897011e-01
-1.24609423e+00 -1.64844245e-01 -8.43331590e-02 -3.72381628e-01
8.15485775e-01 -1.05950129e+00 -8.35639536e-01 -8.81515443e-02
-3.49760413e-01 -1.90891046e-02 9.64978874e-01 1.53936577e+00
-8.24049115e-01 2.60514975e-01 -1.68016672e-01 -7.72261798e-01
-5.76444566e-01 9.00584087e-02 7.17701614e-01 -6.15952134e-01
-7.64786422e-01 1.05841708e+00 -7.60131329e-02 -2.46874630e-01
2.12090299e-01 -4.92014229e-01 4.40464109e-01 -4.10985768e-01
4.12230462e-01 4.96005476e-01 3.99936676e-01 -2.56457537e-01
-5.31666040e-01 6.46592751e-02 -2.12186217e-01 -7.29031205e-01
1.51423430e+00 5.28604202e-02 -8.28826576e-02 3.28540087e-01
1.08345509e+00 -5.40314794e-01 -1.43049622e+00 -1.64342716e-01
-1.32501274e-01 -1.18512407e-01 7.48299003e-01 -1.24550974e+00
-1.19446170e+00 9.51890767e-01 1.36610460e+00 -6.86379552e-01
9.29638386e-01 -1.98066711e-01 4.83506113e-01 -6.01087650e-03
7.33880043e-01 -9.98986483e-01 -2.17805982e-01 1.74982488e-01
8.20580423e-01 -1.02171504e+00 2.47910663e-01 6.82407292e-03
-6.54201686e-01 6.36106133e-01 1.70267165e-01 -3.65750581e-01
1.13900900e+00 6.47801816e-01 -2.03843608e-01 -3.73583823e-01
-4.15816933e-01 5.02793789e-01 2.64137238e-03 4.57594544e-01
5.09527087e-01 3.05364698e-01 -5.28681040e-01 8.71662199e-01
1.97724208e-01 3.45863909e-01 4.88894045e-01 8.63613427e-01
6.41524699e-03 -7.38788664e-01 -1.93267778e-01 8.24016333e-01
-8.45511258e-01 -4.81234461e-01 7.74415076e-01 5.61400712e-01
-3.34480464e-01 6.25054181e-01 -1.66537061e-01 2.98931021e-02
8.38233083e-02 6.28999248e-02 8.27258587e-01 -3.55611622e-01
-6.05852604e-01 3.92555445e-01 1.09547712e-01 -7.90895402e-01
-2.95748055e-01 -7.64050543e-01 -1.46595275e+00 -2.17766911e-01
-3.51435065e-01 1.14835978e-01 9.89576340e-01 5.98576486e-01
3.75481218e-01 8.34757447e-01 1.51647165e-01 -6.88200533e-01
-9.58383679e-01 -1.01605284e+00 -8.70797753e-01 8.19625556e-02
-6.23195395e-02 -7.64351666e-01 -3.38462949e-01 -3.47284347e-01] | [13.55212688446045, -2.356722593307495] |
a2b1e1c4-01f3-4c9d-a96d-ee7d64d1b4eb | generative-x-vectors-for-text-independent | 1809.06798 | null | http://arxiv.org/abs/1809.06798v1 | http://arxiv.org/pdf/1809.06798v1.pdf | Generative x-vectors for text-independent speaker verification | Speaker verification (SV) systems using deep neural network embeddings,
so-called the x-vector systems, are becoming popular due to its good
performance superior to the i-vector systems. The fusion of these systems
provides improved performance benefiting both from the discriminatively trained
x-vectors and generative i-vectors capturing distinct speaker characteristics.
In this paper, we propose a novel method to include the complementary
information of i-vector and x-vector, that is called generative x-vector. The
generative x-vector utilizes a transformation model learned from the i-vector
and x-vector representations of the background data. Canonical correlation
analysis is applied to derive this transformation model, which is later used to
transform the standard x-vectors of the enrollment and test segments to the
corresponding generative x-vectors. The SV experiments performed on the NIST
SRE 2010 dataset demonstrate that the system using generative x-vectors
provides considerably better performance than the baseline i-vector and
x-vector systems. Furthermore, the generative x-vectors outperform the fusion
of i-vector and x-vector systems for long-duration utterances, while yielding
comparable results for short-duration utterances. | ['Jichen Yang', 'Emre Yilmaz', 'Longting Xu', 'Haizhou Li', 'Rohan Kumar Das'] | 2018-09-17 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 1.51749820e-01 -3.90316486e-01 7.94003680e-02 -7.86382675e-01
-1.17189229e+00 -4.28433627e-01 7.54762352e-01 -2.37944096e-01
-2.22427621e-01 2.86288857e-01 5.85043311e-01 -4.05193746e-01
1.47984788e-01 -3.49734634e-01 -3.04534644e-01 -1.05659425e+00
1.70421407e-01 -9.58216563e-02 -3.55848551e-01 -1.31634176e-01
-8.70350227e-02 4.12221223e-01 -1.72914922e+00 -7.93746114e-02
4.94390398e-01 1.10962164e+00 -4.58279848e-02 7.12543905e-01
-8.34308788e-02 1.24334276e-01 -8.48596811e-01 -2.47719079e-01
1.24787122e-01 -4.94647294e-01 -1.70036212e-01 8.23253579e-03
5.96985817e-01 -3.12491298e-01 -7.03960776e-01 9.85033691e-01
6.55374229e-01 4.21127260e-01 6.97769284e-01 -1.24668217e+00
-9.31308389e-01 7.85506546e-01 -3.57899517e-01 1.82003334e-01
2.69572318e-01 -1.53748229e-01 1.03658068e+00 -1.09488165e+00
2.40261868e-01 1.46803355e+00 6.37446105e-01 6.55857444e-01
-1.20391369e+00 -8.84313822e-01 7.84032140e-03 4.68287468e-01
-1.65298581e+00 -6.56512976e-01 1.21717000e+00 -4.75718558e-01
9.59635675e-01 5.26886046e-01 2.43259490e-01 1.46534503e+00
1.07545257e-01 9.88786101e-01 8.82231355e-01 -5.70802689e-01
1.72806412e-01 3.48136634e-01 8.58514667e-01 3.64044189e-01
-2.53971517e-01 5.95660925e-01 -6.19757533e-01 -3.09542507e-01
3.06764960e-01 1.90465778e-01 -6.19485736e-01 7.88413920e-03
-9.10432160e-01 1.12183464e+00 1.76247060e-01 6.22834325e-01
-5.51538944e-01 -1.66305542e-01 4.41072017e-01 1.98443271e-02
3.51026148e-01 -1.80986792e-01 -2.76485354e-01 -5.02701551e-02
-1.14463925e+00 1.63242579e-01 6.40008032e-01 7.31559932e-01
6.26502395e-01 9.50436711e-01 -4.10928488e-01 1.11577177e+00
6.33635521e-01 6.77768946e-01 9.93698537e-01 -5.88119447e-01
4.93240476e-01 3.44333470e-01 -3.00028682e-01 -6.55295730e-01
-1.21090859e-01 -6.96401298e-01 -8.22809875e-01 -1.01660155e-01
8.56431723e-02 -9.47714671e-02 -1.06000841e+00 1.97331131e+00
2.09215552e-01 1.12892449e-01 3.89826894e-01 5.54220974e-01
1.07206988e+00 1.00936246e+00 -1.92596793e-01 -1.19293116e-01
1.28506041e+00 -6.20973527e-01 -1.08408165e+00 -4.62483540e-02
2.41345182e-01 -7.21268356e-01 8.83532703e-01 9.90660638e-02
-6.99691355e-01 -1.05719197e+00 -1.27924550e+00 2.54034877e-01
-5.15225470e-01 5.49436629e-01 3.73198874e-02 1.16128433e+00
-1.13080621e+00 1.42642379e-01 -5.27838230e-01 -1.67658165e-01
1.48620889e-01 4.64285135e-01 -5.78358233e-01 1.51473269e-01
-1.24070346e+00 6.70720756e-01 2.27869749e-01 9.75252315e-02
-8.46684396e-01 -6.45358920e-01 -1.15340924e+00 2.78971285e-01
-1.56177938e-01 6.34193495e-02 1.24041605e+00 -5.00384867e-01
-1.81947315e+00 3.47424179e-01 -4.65806514e-01 -2.22417593e-01
-2.42430195e-02 -2.52725363e-01 -8.09422374e-01 -1.04676053e-01
-1.18436091e-01 3.40778679e-01 9.91602004e-01 -1.37728739e+00
-4.53271300e-01 -7.19588578e-01 -6.69001698e-01 -1.65742591e-01
-5.56884170e-01 -1.34667312e-03 -3.21539909e-01 -6.68020904e-01
2.13266283e-01 -1.10251129e+00 1.79762870e-01 -6.71996117e-01
-7.87438631e-01 -4.98511910e-01 1.32730818e+00 -1.08469236e+00
1.27515626e+00 -2.67400551e+00 -4.23758253e-02 3.37367505e-01
1.21436074e-01 6.47699773e-01 -2.84913629e-01 3.73277426e-01
-4.52274084e-01 -3.20424527e-01 -9.80234221e-02 -5.07594228e-01
3.46784085e-01 2.52186149e-01 -4.56608027e-01 5.61493397e-01
2.45558545e-01 6.86467052e-01 -3.25949937e-01 -1.58081621e-01
3.93823653e-01 1.02349842e+00 -3.85417551e-01 3.49985212e-01
4.01478916e-01 1.91288128e-01 -2.12300234e-02 3.17395121e-01
7.86980391e-01 3.77388090e-01 5.60190901e-03 -3.69637758e-01
-3.84885706e-02 2.21589014e-01 -9.34497356e-01 1.23186278e+00
-3.65661025e-01 9.00794327e-01 1.19318105e-01 -8.74721110e-01
1.33204079e+00 6.56794786e-01 3.06420296e-01 -5.05399108e-01
3.81847203e-01 1.13123562e-02 1.59288734e-01 -1.89276055e-01
5.28256595e-01 -2.62748748e-01 -4.25289869e-02 4.31363910e-01
4.82495517e-01 -9.56212729e-03 -1.22303471e-01 1.59568921e-01
6.05527699e-01 -1.93480149e-01 1.65069476e-01 4.74438779e-02
1.00204480e+00 -8.86875033e-01 6.89953387e-01 4.86444652e-01
-2.35310271e-01 4.26956475e-01 1.24438651e-01 7.78904855e-02
-7.85559475e-01 -1.26395488e+00 -3.83341491e-01 1.03883958e+00
-2.27512419e-01 -3.14067215e-01 -7.95052409e-01 -5.53617477e-01
6.58809841e-02 1.17404068e+00 -8.16612124e-01 -3.88323486e-01
-5.65392852e-01 -5.34277737e-01 7.14401186e-01 8.15740764e-01
2.58250207e-01 -5.73039651e-01 9.66465920e-02 3.41858238e-01
-4.74711359e-02 -9.04257357e-01 -8.29319298e-01 3.89492601e-01
-4.73995179e-01 -5.38226128e-01 -8.78190041e-01 -7.62688756e-01
3.82875383e-01 2.46047869e-01 5.18454492e-01 -5.01494944e-01
-6.82336167e-02 4.10831839e-01 -3.61839175e-01 -4.71660107e-01
-7.58333981e-01 -3.78798157e-01 4.36987400e-01 4.77602869e-01
8.32491696e-01 -2.48791948e-01 -1.48038849e-01 3.04037601e-01
-8.29384983e-01 -5.61417460e-01 4.40822333e-01 1.25136483e+00
4.15470809e-01 -1.34342015e-01 6.66198969e-01 -3.32766265e-01
5.84061325e-01 -1.83326483e-01 -4.14174378e-01 1.05292462e-01
-6.37715876e-01 1.43389985e-01 2.88005322e-01 -7.02723503e-01
-1.19976187e+00 4.75894799e-03 -6.47018254e-01 -5.91812551e-01
-1.82609960e-01 2.76092619e-01 -6.51032090e-01 1.39245138e-01
5.11506200e-01 6.85011744e-01 5.12500070e-02 -6.98115230e-01
6.79149449e-01 1.17224085e+00 8.24083865e-01 -1.40291646e-01
8.05703342e-01 -1.21389344e-01 -6.72738731e-01 -1.25982988e+00
-3.77329618e-01 -6.96167648e-01 -5.01874030e-01 -1.33720890e-01
9.77213740e-01 -7.09671259e-01 -3.95265937e-01 5.02893984e-01
-1.25571454e+00 4.28795695e-01 -3.45042825e-01 8.91870379e-01
-1.12572595e-01 5.89517534e-01 -6.78917885e-01 -9.69785213e-01
-3.83390248e-01 -1.56603611e+00 1.22159803e+00 3.57678950e-01
-2.50086904e-01 -9.08466637e-01 2.81932056e-01 3.38666379e-01
5.99927843e-01 -3.72706890e-01 9.43626285e-01 -1.36262441e+00
3.75660174e-02 -6.51924849e-01 -1.33992434e-02 9.78715062e-01
2.71920413e-01 -3.57460938e-02 -1.49685645e+00 -4.21384305e-01
4.58751261e-01 2.89359987e-01 7.16170788e-01 5.79586804e-01
1.03407943e+00 -3.21487814e-01 -8.95320326e-02 6.40349627e-01
1.10644519e+00 6.77155972e-01 3.76394331e-01 -1.25086725e-01
6.53173029e-01 3.78768921e-01 6.09971918e-02 2.22471520e-01
1.84265599e-01 1.09047413e+00 -7.60922581e-02 -1.01379462e-01
-2.20812976e-01 -1.22498684e-01 6.50009692e-01 1.48597026e+00
1.78082570e-01 -7.04715624e-02 -7.26385474e-01 3.19489717e-01
-1.30390620e+00 -1.08837259e+00 1.02455057e-01 2.18011904e+00
6.80249572e-01 -2.77133822e-01 5.84805459e-02 6.42930508e-01
8.83958876e-01 2.02966347e-01 -4.49570745e-01 -3.91909897e-01
-1.96433246e-01 3.99294823e-01 6.56479225e-02 5.69401920e-01
-9.37491357e-01 4.29920375e-01 6.44095278e+00 7.36753464e-01
-1.57409859e+00 3.99006784e-01 3.06368947e-01 -1.76236890e-02
-2.25308955e-01 -5.19115210e-01 -1.17289591e+00 4.73404735e-01
1.47836876e+00 -3.11978161e-01 4.31748666e-02 9.26460445e-01
-8.05675704e-03 6.11172438e-01 -1.31790578e+00 1.33896315e+00
5.86784303e-01 -1.03665388e+00 -1.50181456e-02 4.15233701e-01
5.57024777e-01 -2.31923480e-02 4.33690935e-01 8.05673897e-01
9.13556293e-02 -7.46762276e-01 7.45712698e-01 1.90849245e-01
7.43403137e-01 -5.64251244e-01 1.06960571e+00 -1.53557181e-01
-9.90211010e-01 -3.24502587e-02 -1.55263409e-01 5.81139326e-01
1.59952939e-01 3.48260820e-01 -8.08080852e-01 7.01444268e-01
3.64805430e-01 3.90820831e-01 -3.55620682e-01 6.61628544e-01
6.67603035e-03 9.41902220e-01 -4.18267921e-02 8.04548562e-02
1.52342096e-01 -3.44072759e-01 6.91979110e-01 1.45530224e+00
4.60288286e-01 -2.76959836e-01 -2.02917814e-01 6.97479725e-01
1.42294383e-02 1.30749464e-01 -4.96903509e-01 -2.49599889e-01
4.08716142e-01 9.93948519e-01 2.52583232e-02 -5.77224493e-01
-5.30869246e-01 6.44039690e-01 -2.01258570e-01 7.57958651e-01
-8.87698770e-01 -6.47513628e-01 9.99674022e-01 -3.81333381e-01
6.66595280e-01 -2.50888050e-01 2.09013689e-02 -9.43040788e-01
-5.92803359e-02 -9.33419287e-01 -4.03674319e-02 -4.47734326e-01
-1.29560018e+00 1.34835839e+00 4.26613018e-02 -1.29583454e+00
-7.61506975e-01 -5.77153087e-01 -6.99250698e-01 1.29337537e+00
-1.32317448e+00 -1.07272315e+00 -1.74558058e-01 5.97060621e-01
5.56754589e-01 -8.77454937e-01 1.08994567e+00 3.41793954e-01
-8.63142729e-01 1.18679702e+00 6.91797793e-01 2.68549800e-01
4.67671543e-01 -1.11767018e+00 3.64565313e-01 9.66074824e-01
5.25011599e-01 8.71375859e-01 5.74119985e-01 -1.29445747e-01
-1.44088387e+00 -1.04298341e+00 9.54681635e-01 -2.30228066e-01
2.68330991e-01 -4.01425362e-01 -1.14615846e+00 7.12636173e-01
1.45397976e-01 -2.08392471e-01 1.39556408e+00 9.59457904e-02
-7.06703126e-01 -3.06368619e-01 -1.00939345e+00 1.87932149e-01
2.27935836e-01 -1.10864484e+00 -9.55355704e-01 -1.13201924e-01
6.16421640e-01 -2.53633335e-02 -8.25663984e-01 1.13624744e-01
5.24141133e-01 -8.07024240e-01 1.12386966e+00 -2.97285050e-01
-9.27632377e-02 -2.12498203e-01 -8.27467322e-01 -1.44132793e+00
-4.33858216e-01 -3.48196596e-01 -1.42992698e-02 1.64599144e+00
4.63886470e-01 -7.59267271e-01 5.92727304e-01 4.25088376e-01
-2.45999247e-01 -3.36004227e-01 -1.29862773e+00 -9.77767944e-01
-4.63810116e-02 -5.91598868e-01 8.68100762e-01 8.33305597e-01
-2.59801388e-01 5.58405161e-01 -2.88210452e-01 1.99972585e-01
6.66064382e-01 -4.82309498e-02 7.45894611e-01 -9.51863706e-01
-2.95313030e-01 -5.53967953e-01 -8.23245764e-01 -9.85677481e-01
4.87744033e-01 -9.70520973e-01 2.64192611e-01 -1.16222167e+00
-1.26660734e-01 -1.56817567e-02 -4.73429859e-01 2.82086641e-01
-4.18267787e-01 2.56172627e-01 2.43955061e-01 -5.21471426e-02
1.15261830e-01 1.10903752e+00 5.78255177e-01 -4.27954108e-01
-4.42784250e-01 1.54640839e-01 -4.97636735e-01 3.19874465e-01
4.96388525e-01 -1.07044369e-01 -3.16504359e-01 -3.92251424e-02
-8.89996409e-01 1.43737197e-01 -6.44940659e-02 -9.15212333e-01
-1.82070062e-02 4.20970857e-01 5.06803513e-01 -9.94538784e-01
9.94132936e-01 -5.21859646e-01 1.54716298e-01 2.29609311e-01
-3.65336537e-01 -5.86457103e-02 2.40297779e-01 4.93754506e-01
-6.75241768e-01 -6.91309869e-02 7.74616897e-01 4.88497198e-01
-3.73107433e-01 2.52471715e-01 -3.34500134e-01 -5.24620473e-01
7.58449256e-01 -4.16044593e-01 -9.53760073e-02 -6.49682462e-01
-7.19512701e-01 -5.11705160e-01 -1.72988996e-01 7.73287892e-01
8.27132106e-01 -1.76357424e+00 -9.09675777e-01 8.66684914e-01
1.39977559e-01 -7.95889676e-01 3.84149104e-01 5.56420922e-01
1.28319606e-01 7.42489576e-01 -1.04757190e-01 -8.65216017e-01
-1.74894202e+00 4.71636534e-01 5.65622076e-02 1.74200624e-01
-5.15934646e-01 9.12653744e-01 4.69898999e-01 -4.78131771e-01
4.90621030e-01 -3.64060104e-01 -3.25297326e-01 5.93668260e-02
8.85691047e-01 2.20459029e-01 1.76516294e-01 -1.41853952e+00
-5.44968009e-01 4.46532726e-01 -2.51053184e-01 -4.73451257e-01
1.47129679e+00 -3.03378236e-02 1.44371137e-01 5.38000703e-01
1.76660013e+00 1.82383060e-01 -7.92503774e-01 -5.53475857e-01
-8.72396529e-02 -4.41686720e-01 4.82839823e-01 -4.84313428e-01
-1.13424313e+00 1.11841512e+00 9.34387624e-01 7.50368312e-02
9.23948050e-01 -4.00707684e-02 7.00071394e-01 -6.71624541e-02
-8.49735439e-02 -6.83113694e-01 -2.72255570e-01 3.32541704e-01
9.76721108e-01 -9.90765750e-01 -3.42451572e-01 -8.89442563e-02
-6.64436519e-01 1.02590990e+00 2.34647810e-01 3.55767399e-01
8.19537103e-01 1.71148106e-02 5.63038290e-01 1.37108386e-01
-4.44072753e-01 1.74630396e-02 5.59120834e-01 7.19744325e-01
5.14334321e-01 4.08555746e-01 6.77716732e-02 8.13576996e-01
-5.64848900e-01 -4.33292955e-01 1.47618636e-01 6.08168244e-01
-1.04402721e-01 -1.42082703e+00 -8.98916602e-01 5.46630740e-01
-2.66020268e-01 1.00273140e-01 -1.95072442e-01 4.74555969e-01
-1.06143951e-01 1.21218264e+00 -4.31307517e-02 -9.38425422e-01
3.83077830e-01 6.03623211e-01 3.27779144e-01 -3.22482109e-01
-6.67875230e-01 2.88281292e-01 -1.16885424e-01 -2.90169239e-01
-1.19316645e-01 -8.05086672e-01 -8.32319438e-01 -1.40058488e-01
-7.31313169e-01 3.73542815e-01 1.39984679e+00 7.91520238e-01
1.04634278e-01 5.64399660e-01 9.24038827e-01 -1.12808955e+00
-1.02905619e+00 -1.24219263e+00 -8.92236173e-01 4.60821867e-01
6.58019185e-01 -5.68476379e-01 -6.34160340e-01 3.25439274e-02] | [14.355993270874023, 6.092092037200928] |
20ffab78-a837-4056-8ba6-aa620c9a7e4d | incorporating-dynamic-semantics-into-pre | 2203.16369 | null | https://arxiv.org/abs/2203.16369v2 | https://arxiv.org/pdf/2203.16369v2.pdf | Incorporating Dynamic Semantics into Pre-Trained Language Model for Aspect-based Sentiment Analysis | Aspect-based sentiment analysis (ABSA) predicts sentiment polarity towards a specific aspect in the given sentence. While pre-trained language models such as BERT have achieved great success, incorporating dynamic semantic changes into ABSA remains challenging. To this end, in this paper, we propose to address this problem by Dynamic Re-weighting BERT (DR-BERT), a novel method designed to learn dynamic aspect-oriented semantics for ABSA. Specifically, we first take the Stack-BERT layers as a primary encoder to grasp the overall semantic of the sentence and then fine-tune it by incorporating a lightweight Dynamic Re-weighting Adapter (DRA). Note that the DRA can pay close attention to a small region of the sentences at each step and re-weigh the vitally important words for better aspect-aware sentiment understanding. Finally, experimental results on three benchmark datasets demonstrate the effectiveness and the rationality of our proposed model and provide good interpretable insights for future semantic modeling. | ['Enhong Chen', 'Wei Wu', 'Qi Liu', 'Hongke Zhao', 'Mengdi Zhang', 'Kun Zhang', 'Kai Zhang'] | 2022-03-30 | null | https://aclanthology.org/2022.findings-acl.285 | https://aclanthology.org/2022.findings-acl.285.pdf | findings-acl-2022-5 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 1.09165005e-01 2.47031853e-01 -2.44355172e-01 -1.02825224e+00
-5.19329667e-01 -6.06303811e-01 5.30094445e-01 2.33518362e-01
-2.29757458e-01 2.53058434e-01 7.38741338e-01 -3.57258022e-01
5.83530776e-02 -1.09009850e+00 -6.26454532e-01 -3.09499890e-01
4.75481004e-01 3.13568950e-01 2.03305036e-01 -7.54926503e-01
3.00808012e-01 -1.34628773e-01 -1.12331510e+00 5.81854224e-01
7.39994407e-01 8.85284662e-01 2.93349534e-01 3.74262720e-01
-6.16463661e-01 7.24938154e-01 -4.65387523e-01 -9.81656492e-01
-8.84593353e-02 -2.31539503e-01 -9.24065948e-01 3.22577171e-02
-1.24612644e-01 -1.12106577e-01 2.24310696e-01 9.32529330e-01
2.61551291e-01 1.50589481e-01 2.49886855e-01 -1.03987896e+00
-6.49968565e-01 1.08218026e+00 -6.39926553e-01 2.49914393e-01
2.36484453e-01 1.23390704e-01 1.67321384e+00 -7.34660983e-01
4.84995127e-01 1.30635881e+00 6.47756815e-01 5.66679597e-01
-6.23306572e-01 -4.24235493e-01 9.87780929e-01 2.53671318e-01
-6.17214322e-01 -2.64279485e-01 1.16046703e+00 -4.79642488e-02
1.18011999e+00 2.96128839e-01 9.67975318e-01 8.85596335e-01
9.01633427e-02 1.10634041e+00 7.76061714e-01 -3.51328179e-02
2.52624989e-01 1.41684443e-01 5.20184040e-01 5.34859240e-01
2.74986386e-01 -7.08339214e-01 -6.76770270e-01 5.60042560e-02
-3.64768058e-02 1.43002987e-01 2.25761123e-02 -2.61776596e-01
-9.80483949e-01 9.28831577e-01 3.09802860e-01 2.40940735e-01
-4.15128410e-01 2.69558996e-01 7.87350714e-01 2.70458370e-01
7.92038381e-01 7.06373036e-01 -1.04998219e+00 -4.43709791e-01
-5.74449897e-01 1.76320925e-01 7.22926199e-01 9.30945933e-01
8.54870617e-01 -4.49020565e-02 -3.47986609e-01 8.88571024e-01
5.28002679e-01 4.12871718e-01 6.00553095e-01 -5.01388252e-01
5.60558319e-01 1.14834237e+00 -2.05084339e-01 -1.03104651e+00
-3.91931772e-01 -5.57676852e-01 -2.70514488e-01 -2.33594239e-01
-8.15603584e-02 -3.46583426e-02 -8.05443108e-01 1.67340994e+00
4.49362189e-01 -1.06488347e-01 1.87036037e-01 7.57511377e-01
6.46197379e-01 7.48357058e-01 3.02626163e-01 1.48584977e-01
1.66887891e+00 -1.41971076e+00 -6.62708163e-01 -8.92879963e-01
7.20223010e-01 -7.63978601e-01 1.56944442e+00 1.29713304e-02
-8.84437919e-01 -2.31124327e-01 -1.17894018e+00 -2.76037276e-01
-4.74902660e-01 -5.18233404e-02 9.49871540e-01 5.41836023e-01
-8.32294941e-01 2.18454495e-01 -9.36609030e-01 -2.63124913e-01
5.26000440e-01 1.82901397e-01 9.22107324e-02 -6.15453869e-02
-1.30192077e+00 5.73846340e-01 1.63472772e-01 1.39445812e-01
-6.04412079e-01 -8.42104733e-01 -1.12022960e+00 2.71853089e-01
5.00103891e-01 -1.08706474e+00 1.29656172e+00 -1.24744236e+00
-1.59655547e+00 7.64238238e-01 -7.11917520e-01 -3.62435162e-01
-4.48006466e-02 -4.47710276e-01 -3.79111111e-01 2.55400389e-02
1.80793747e-01 4.32128787e-01 8.80408823e-01 -1.23800671e+00
-6.32708251e-01 -5.57207167e-01 1.12076187e+00 5.45505881e-01
-6.58859193e-01 8.64752531e-02 -5.22416830e-01 -8.86575222e-01
-1.85328126e-01 -6.30152166e-01 -4.16925311e-01 -2.95073241e-01
-2.31153667e-01 -2.42701665e-01 5.11178195e-01 -4.43197995e-01
1.47258139e+00 -2.01961255e+00 1.35857359e-01 -8.79470855e-02
1.89731389e-01 2.44456753e-01 -3.59653324e-01 2.74848312e-01
5.28030060e-02 5.26748262e-02 -5.23748338e-01 -5.73026001e-01
4.81155282e-03 1.77418441e-01 -6.23484790e-01 -2.40245923e-01
3.79481614e-01 1.21071923e+00 -1.05748081e+00 -1.48019090e-01
1.40700471e-02 3.72529924e-01 -9.21329618e-01 1.09677024e-01
-6.39172971e-01 7.69087579e-03 -8.40083420e-01 6.97643340e-01
5.85444391e-01 -3.99090767e-01 1.87968642e-01 -4.83487576e-01
2.78325558e-01 8.31803083e-01 -6.89053118e-01 1.59563661e+00
-1.13813210e+00 3.81977350e-01 -1.41895562e-01 -9.12354171e-01
9.84132409e-01 -1.67356357e-01 2.98666239e-01 -8.45885158e-01
2.22686499e-01 2.08184193e-03 -1.79571822e-01 -4.23766583e-01
9.55283463e-01 -5.48602581e-01 -2.88299680e-01 7.16760278e-01
-1.13264486e-01 -2.08801821e-01 4.26763780e-02 3.42109352e-01
8.02703142e-01 -1.90293305e-02 4.90037024e-01 -2.58268893e-01
9.03639555e-01 6.62383214e-02 5.95887780e-01 2.90468931e-01
-2.05724239e-01 5.70750594e-01 7.03338623e-01 -4.88582402e-01
-8.01127732e-01 -5.89822114e-01 3.16023797e-01 1.38278365e+00
4.14052635e-01 -7.52693534e-01 -7.47141600e-01 -1.07611382e+00
-2.33546034e-01 1.12413359e+00 -7.76408970e-01 -5.39204895e-01
-6.32241249e-01 -1.06842434e+00 7.43685439e-02 7.59809375e-01
4.99264210e-01 -1.05470741e+00 -2.54340768e-01 2.00510681e-01
-2.76992828e-01 -9.41096842e-01 -6.17120445e-01 -4.43359464e-02
-8.03110778e-01 -8.78989816e-01 -2.79537141e-01 -7.71531463e-01
6.75176203e-01 5.73270679e-01 1.26467133e+00 9.43447128e-02
2.83278137e-01 3.38889748e-01 -7.49339104e-01 -5.28005362e-01
-2.09722385e-01 5.03885090e-01 -4.43782389e-01 2.16946408e-01
8.17771792e-01 -5.03565609e-01 -8.00056517e-01 1.13819472e-01
-1.03401518e+00 1.07340612e-01 3.43609542e-01 5.61901152e-01
6.37571812e-01 -1.26017198e-01 6.57616317e-01 -1.58259189e+00
8.57293844e-01 -5.15021801e-01 -1.41079605e-01 2.33417824e-01
-9.14402723e-01 -4.97542731e-02 8.00331235e-01 2.55009439e-02
-1.30485320e+00 -4.13351953e-01 -5.98622978e-01 2.25386634e-01
4.41461533e-01 8.23014319e-01 -4.45428312e-01 5.72900236e-01
2.63543893e-02 2.76040107e-01 -2.11385265e-01 -2.87647188e-01
6.66262090e-01 5.91550708e-01 -4.64897342e-02 -4.85767543e-01
6.22462332e-01 8.90684009e-01 -4.25203562e-01 -2.65145600e-01
-1.78779233e+00 -6.69158518e-01 -2.70796716e-01 1.03888363e-02
5.88260233e-01 -1.19761527e+00 -4.00611550e-01 5.26403189e-01
-1.06608701e+00 -1.93323106e-01 -5.52383363e-01 -6.21997118e-02
-3.89574885e-01 2.24543288e-01 -3.34790319e-01 -3.71938497e-01
-8.23013544e-01 -1.13502395e+00 1.24606276e+00 3.01106215e-01
-4.13564473e-01 -1.41287649e+00 1.10239893e-01 8.61889482e-01
5.97304821e-01 -3.47027987e-01 1.03374803e+00 -6.46724820e-01
-3.32140654e-01 -1.19070426e-01 -2.89288759e-01 4.20136541e-01
3.78568202e-01 -1.67990834e-01 -1.10301352e+00 -1.47060052e-01
3.90117466e-01 -1.00016020e-01 1.00457025e+00 1.31108582e-01
1.18453419e+00 -4.06766385e-01 -4.37213443e-02 5.94805896e-01
1.24295211e+00 6.53450340e-02 5.02194703e-01 8.69099617e-01
8.79357338e-01 4.75008339e-01 1.00621974e+00 4.44421083e-01
1.11886549e+00 3.42922896e-01 5.16748369e-01 1.56542376e-01
-2.99528807e-01 -4.27928209e-01 5.25336504e-01 1.39928031e+00
1.85786635e-01 -1.95396304e-01 -5.73257744e-01 8.38261306e-01
-1.67655754e+00 -5.91396451e-01 1.04215063e-01 1.55970454e+00
1.05379355e+00 4.56333697e-01 -1.41926229e-01 -6.23676740e-02
2.72014827e-01 9.98675704e-01 -6.78518355e-01 -7.09183455e-01
-2.27648556e-01 9.32045747e-03 1.09807864e-01 7.27756143e-01
-9.23085213e-01 1.25049746e+00 5.28775978e+00 7.97925770e-01
-1.15469646e+00 3.51221025e-01 6.67757332e-01 -8.36538821e-02
-1.28717673e+00 3.84553283e-01 -8.71583581e-01 4.15411234e-01
6.35920107e-01 -5.45860708e-01 4.03804481e-01 1.07664454e+00
1.61610916e-01 1.53323308e-01 -6.71570063e-01 4.91543531e-01
4.16150123e-01 -1.31421316e+00 6.52294338e-01 -5.40394306e-01
7.10862398e-01 -7.88386315e-02 5.80036081e-02 5.14742970e-01
1.99736714e-01 -5.56799114e-01 7.90850818e-01 3.28444809e-01
3.33811194e-01 -7.78170705e-01 8.99591088e-01 -9.22205001e-02
-1.31105983e+00 2.32091658e-02 -4.94234771e-01 -1.35728672e-01
3.26787204e-01 8.45298707e-01 -6.06701374e-01 5.48356175e-01
5.51411271e-01 1.30170512e+00 -6.86395109e-01 3.91197354e-01
-5.25319517e-01 8.14689815e-01 1.95300683e-01 -3.76826048e-01
4.27776635e-01 -1.76410466e-01 6.98689640e-01 1.05015445e+00
1.36610731e-01 -5.49798347e-02 -4.21586812e-01 5.59629560e-01
-2.08529383e-01 2.17299834e-01 -2.08695561e-01 -3.38249832e-01
1.23120382e-01 1.43819749e+00 -6.27017498e-01 -3.82399678e-01
-6.79518104e-01 9.21861708e-01 4.65466857e-01 2.73784161e-01
-7.69689500e-01 -4.07780677e-01 1.12536466e+00 3.71306092e-02
5.85892141e-01 5.94880730e-02 -7.91189671e-01 -1.44026816e+00
2.26030365e-01 -8.35141778e-01 4.86225188e-01 -8.79012823e-01
-1.12557220e+00 7.84265161e-01 -4.43749458e-01 -1.09738648e+00
1.11762099e-01 -6.02652431e-01 -8.55191886e-01 5.12969673e-01
-2.08775234e+00 -1.48162246e+00 -1.15754582e-01 1.56965807e-01
1.06658876e+00 -1.09480433e-01 5.49288332e-01 1.02165513e-01
-4.69834208e-01 5.59432209e-01 -2.60022074e-01 -2.05436036e-01
5.33845067e-01 -1.28885901e+00 7.81220257e-01 8.24068964e-01
-8.80025998e-02 9.09210861e-01 7.72927880e-01 -5.39551854e-01
-1.48826933e+00 -1.25613797e+00 1.25697935e+00 -5.89516342e-01
1.02224243e+00 -3.16239268e-01 -8.93867552e-01 8.09351861e-01
2.36620039e-01 -4.96325135e-01 8.12964976e-01 6.30254567e-01
-5.70865512e-01 -5.39961576e-01 -7.36199141e-01 6.83926463e-01
9.92865920e-01 -6.28278375e-01 -9.51425731e-01 8.63448307e-02
1.38698435e+00 -1.33663133e-01 -4.82860088e-01 4.91700709e-01
3.81536782e-01 -9.07221854e-01 8.55701447e-01 -6.97534025e-01
8.56467664e-01 -4.19913530e-01 -1.81079358e-01 -1.60350657e+00
-7.18598291e-02 -1.72208130e-01 -6.59704879e-02 1.50451946e+00
6.01061404e-01 -5.29015183e-01 8.55156243e-01 4.82168436e-01
-2.70745367e-01 -1.21786332e+00 -5.38225710e-01 -2.40200251e-01
-2.87616532e-02 -8.25189114e-01 1.14979577e+00 7.92685628e-01
1.51773497e-01 6.23190105e-01 -1.42746612e-01 9.41312313e-02
4.46304716e-02 7.03429401e-01 6.27678275e-01 -6.29935503e-01
-3.29950958e-01 -5.15339911e-01 -2.56587416e-01 -1.23796189e+00
2.59647667e-01 -1.02077699e+00 -1.49841264e-01 -1.70520973e+00
5.70061624e-01 -4.45555240e-01 -4.80046660e-01 5.17782688e-01
-8.52132142e-01 9.67529118e-02 7.08126202e-02 -1.01476461e-01
-9.99483883e-01 1.18569207e+00 1.29963291e+00 -5.29014051e-01
4.51026186e-02 2.17239305e-01 -1.54862368e+00 8.52164030e-01
8.32062542e-01 -4.92109954e-01 -8.45421433e-01 -9.12908316e-01
1.05482626e+00 -5.22846341e-01 -5.54233007e-02 -2.37693846e-01
-6.18083626e-02 -2.36058548e-01 -2.81063735e-01 -4.70390648e-01
2.29052633e-01 -8.01025689e-01 -4.68703270e-01 1.56222060e-01
-4.56212938e-01 1.41535074e-01 1.41794696e-01 6.83166325e-01
-6.10353112e-01 -3.60044986e-01 3.94063622e-01 -3.16365995e-02
-8.96770000e-01 2.36592382e-01 -1.23671979e-01 3.59593779e-01
7.99979508e-01 7.29091763e-02 -3.28849912e-01 -4.51647550e-01
-3.64560485e-01 4.53208148e-01 5.25712013e-01 7.80839324e-01
5.47477424e-01 -1.09314632e+00 -1.92320988e-01 1.25534788e-01
4.90992159e-01 2.20155701e-01 5.20869911e-01 5.27488530e-01
-3.79110485e-01 3.35232466e-01 2.56025910e-01 -1.68245640e-02
-1.05918825e+00 3.93452883e-01 2.07305044e-01 -4.67127711e-01
-4.49035883e-01 9.16369319e-01 3.42717946e-01 -8.56033623e-01
-2.69163936e-01 -3.55303645e-01 -5.54095864e-01 2.79788285e-01
5.39010406e-01 7.27602048e-03 1.56034634e-01 -5.42255700e-01
-4.58617121e-01 6.40799940e-01 -3.90680552e-01 1.58208638e-01
1.72168779e+00 -6.25449181e-01 -3.95412385e-01 6.51022673e-01
1.19453168e+00 1.69005737e-01 -1.12754643e+00 -3.74550760e-01
-4.31465581e-02 -3.90921295e-01 6.20979406e-02 -8.47214460e-01
-1.33547056e+00 9.59417582e-01 -1.62210926e-01 9.02719274e-02
1.22461390e+00 1.68113440e-01 1.22478902e+00 5.16221166e-01
1.35935217e-01 -1.07136011e+00 1.39602780e-01 7.40278959e-01
6.86881185e-01 -1.19503105e+00 -1.05338609e-02 -5.44874549e-01
-1.05867743e+00 9.63392496e-01 6.70237780e-01 -8.79513659e-03
7.33823001e-01 6.53559938e-02 3.54895890e-01 -3.97028059e-01
-9.92152214e-01 -6.52415156e-02 1.35342762e-01 2.05664769e-01
4.85825837e-01 1.95325330e-01 -3.99984270e-01 1.19859374e+00
-5.38505971e-01 -2.10144162e-01 5.19052505e-01 9.67917740e-01
-3.99767041e-01 -1.05751777e+00 3.71008039e-01 5.35777330e-01
-4.66682464e-01 -5.16751587e-01 -1.46114722e-01 2.86246657e-01
-2.61633933e-01 8.23444247e-01 -7.52972662e-02 -3.50426108e-01
4.91299123e-01 7.70947114e-02 1.13973371e-03 -7.57086337e-01
-6.92240536e-01 -2.69143194e-01 3.44768733e-01 -6.44089520e-01
-5.83921969e-01 -6.92721307e-01 -1.38032758e+00 -7.62589648e-02
-3.96026112e-02 6.01432547e-02 8.00610244e-01 1.20431554e+00
5.47365367e-01 9.36302125e-01 8.79985929e-01 -8.06499496e-02
-3.66649061e-01 -8.23219478e-01 -3.77338976e-01 5.07970870e-01
2.74528146e-01 -3.51775020e-01 -3.47791731e-01 7.42085651e-03] | [11.489680290222168, 6.62973690032959] |
275f5bcd-2065-49c9-b110-732e236d2c1d | transductive-learning-with-string-kernels-for | 1811.01734 | null | http://arxiv.org/abs/1811.01734v1 | http://arxiv.org/pdf/1811.01734v1.pdf | Transductive Learning with String Kernels for Cross-Domain Text Classification | For many text classification tasks, there is a major problem posed by the
lack of labeled data in a target domain. Although classifiers for a target
domain can be trained on labeled text data from a related source domain, the
accuracy of such classifiers is usually lower in the cross-domain setting.
Recently, string kernels have obtained state-of-the-art results in various text
classification tasks such as native language identification or automatic essay
scoring. Moreover, classifiers based on string kernels have been found to be
robust to the distribution gap between different domains. In this paper, we
formally describe an algorithm composed of two simple yet effective
transductive learning approaches to further improve the results of string
kernels in cross-domain settings. By adapting string kernels to the test set
without using the ground-truth test labels, we report significantly better
accuracy rates in cross-domain English polarity classification. | ['Radu Tudor Ionescu', 'Andrei M. Butnaru'] | 2018-11-02 | null | null | null | null | ['cross-domain-text-classification', 'native-language-identification'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.75400066e-01 -1.96629480e-01 -6.40957415e-01 -6.40207171e-01
-1.38755929e+00 -1.02248931e+00 6.79786265e-01 5.66502631e-01
-4.32134032e-01 1.02096355e+00 4.35158573e-02 -4.58671391e-01
-6.93782046e-02 -7.50009179e-01 -4.75448072e-01 -4.21674073e-01
5.25965214e-01 7.23574400e-01 4.19023633e-01 -2.22502455e-01
2.83750683e-01 4.99407984e-02 -1.08947027e+00 3.90260816e-01
1.26682222e+00 8.99612427e-01 -2.53401637e-01 5.59021056e-01
-2.24663347e-01 7.02168167e-01 -5.62235236e-01 -8.62063348e-01
-2.10100841e-02 -5.43519258e-01 -1.06667495e+00 -1.19698651e-01
5.51992714e-01 2.56844103e-01 -7.92334750e-02 1.22267437e+00
3.46569598e-01 -1.15807816e-01 1.20811176e+00 -1.19348550e+00
-8.46158266e-01 7.32494652e-01 -4.44138795e-01 -8.94550011e-02
5.63154042e-01 -3.50086153e-01 1.24445355e+00 -6.78775251e-01
5.12921095e-01 9.56220329e-01 8.40229392e-01 3.28783810e-01
-1.57540905e+00 -8.64542902e-01 -1.24991536e-02 1.36504680e-01
-7.65846908e-01 -2.17053294e-01 8.72215331e-01 -5.48059940e-01
5.28322995e-01 1.72147602e-02 1.20722182e-01 1.41101992e+00
1.50566539e-02 7.15999424e-01 1.71533251e+00 -8.24097633e-01
1.79846257e-01 6.74104452e-01 6.38965309e-01 5.33731103e-01
1.30678266e-01 -1.85177952e-01 -6.04430258e-01 -6.14116073e-01
1.46297544e-01 -6.60732746e-01 -1.84577659e-01 -5.97336590e-01
-1.19562721e+00 1.17793477e+00 -2.80583394e-03 4.72872436e-01
1.45933896e-01 -4.88629639e-01 8.70117724e-01 6.31606579e-01
1.03230917e+00 6.40968919e-01 -8.26037288e-01 -2.04230607e-01
-8.39841127e-01 1.54439613e-01 1.19330680e+00 6.67350709e-01
5.23774505e-01 -3.87391448e-01 -4.92543690e-02 1.18387616e+00
-1.90287247e-01 4.12348747e-01 7.14142799e-01 -3.87006462e-01
7.97905087e-01 8.05984676e-01 1.31332213e-02 -8.49268913e-01
-1.98937938e-01 -8.47663805e-02 -4.38192040e-01 -1.31457642e-01
9.56389606e-01 -1.33521542e-01 -5.53914607e-01 1.87946928e+00
2.24069670e-01 -2.02958241e-01 4.79951948e-01 4.33124512e-01
6.91225350e-01 4.83552098e-01 9.15309191e-02 -3.69817428e-02
1.30880177e+00 -7.99352050e-01 -3.70786905e-01 -3.34067106e-01
1.24123657e+00 -7.75957465e-01 1.29527843e+00 2.71972626e-01
-6.92382455e-01 -2.51569688e-01 -9.31071639e-01 -2.88640894e-02
-6.42759085e-01 4.35023218e-01 4.85973656e-01 9.63345408e-01
-6.26461506e-01 3.50236714e-01 -2.52793282e-01 -6.62269771e-01
3.19647640e-01 2.30256408e-01 -7.29120493e-01 -1.33573517e-01
-1.43033957e+00 1.13294196e+00 4.57301289e-01 -7.03229010e-01
-2.43428096e-01 -7.23263621e-01 -6.85455263e-01 -2.45120842e-02
7.13364109e-02 -1.56445161e-01 1.33179438e+00 -1.34989035e+00
-1.52675498e+00 1.47870588e+00 1.09126866e-01 -2.88858533e-01
4.57834631e-01 3.82084846e-02 -4.44135010e-01 1.21474572e-01
4.13627505e-01 1.34048522e-01 8.17361653e-01 -8.79282176e-01
-5.41189134e-01 -4.77364153e-01 -1.48537889e-01 9.53122601e-02
-7.48206854e-01 2.18460739e-01 1.98423881e-02 -5.88456511e-01
-2.88904846e-01 -1.08424640e+00 3.61860693e-01 -3.41367781e-01
-2.25436285e-01 -6.16188109e-01 7.87259698e-01 -5.81183970e-01
9.97004151e-01 -2.05824232e+00 -1.03775524e-02 7.49447709e-03
-1.37911901e-01 3.80445182e-01 -2.42192540e-02 1.85422823e-01
-3.41515571e-01 -1.22714557e-01 -1.94433704e-01 1.84599943e-02
-7.91987404e-03 -7.99465775e-02 -5.08436978e-01 4.45195496e-01
2.57880002e-01 7.19557285e-01 -9.02921796e-01 -6.01610124e-01
-2.42635116e-01 -1.73151031e-01 -1.41534105e-01 3.15724947e-02
5.28798513e-02 5.19570597e-02 -3.30274850e-01 5.34001768e-01
4.97884840e-01 -2.90252894e-01 3.75007361e-01 -1.71823315e-02
3.15602571e-01 6.30694151e-01 -7.55328417e-01 1.43217218e+00
-4.50627953e-01 8.57728720e-01 -6.30818382e-02 -1.40422249e+00
9.81552362e-01 2.55801678e-01 3.08150858e-01 -5.06448269e-01
1.67935744e-01 4.65894043e-01 7.71185607e-02 -1.16364695e-01
3.27774972e-01 -4.62606519e-01 -4.77965772e-01 6.67575002e-01
3.39002699e-01 -3.18215519e-01 9.81537253e-02 1.54154673e-01
9.98300731e-01 -8.29665065e-02 3.46023589e-01 -4.50380176e-01
5.95892608e-01 5.21732748e-01 1.82325765e-01 4.77749914e-01
-3.01481783e-01 3.74218464e-01 8.31381381e-01 -1.27007496e-02
-1.05810297e+00 -1.01963925e+00 -6.41275406e-01 1.52971971e+00
-1.23761050e-01 -2.03618467e-01 -7.85265028e-01 -1.27678382e+00
1.66247025e-01 5.95677733e-01 -8.25123966e-01 -3.50930244e-01
-1.41855597e-01 -1.13152754e+00 1.04186714e+00 7.15979695e-01
3.86075139e-01 -5.93448699e-01 -7.18051195e-02 1.07325591e-01
-2.42255300e-01 -1.42458367e+00 -3.56041223e-01 6.44888937e-01
-8.66304874e-01 -1.18097293e+00 -6.57780230e-01 -1.12623417e+00
5.24486363e-01 7.82145411e-02 1.19619894e+00 -6.32924318e-01
1.01053670e-01 3.78506184e-01 -5.31620681e-01 -4.31900561e-01
-7.27393687e-01 5.12169540e-01 2.26301298e-01 -8.37816745e-02
8.63660097e-01 -5.87306917e-02 1.52194872e-01 4.27178651e-01
-6.32002711e-01 -2.64700860e-01 2.72708088e-01 1.29152238e+00
9.26216841e-02 -5.56903780e-02 9.01816845e-01 -1.17508340e+00
1.07933640e+00 -4.88088191e-01 -4.53140527e-01 4.75547880e-01
-7.59397626e-01 5.35974987e-02 7.82869220e-01 -9.54850912e-01
-1.02934217e+00 -4.79015075e-02 3.85597020e-01 1.09228678e-01
-2.08202124e-01 5.46969116e-01 3.11778896e-02 -2.09655270e-01
1.08234417e+00 -9.19159129e-03 -1.68690681e-02 -2.16874182e-01
1.63355127e-01 1.01123452e+00 2.41636589e-01 -9.62887466e-01
5.99384665e-01 7.79456869e-02 -2.18404099e-01 -5.78124642e-01
-1.29542708e+00 -4.86847192e-01 -7.73084342e-01 8.18605870e-02
7.88153529e-01 -8.96464527e-01 -2.28355765e-01 6.46685839e-01
-8.13736200e-01 -4.03629631e-01 -6.57521635e-02 5.35923660e-01
-5.35796404e-01 3.90810430e-01 -5.66026032e-01 -4.63961065e-01
-1.74368560e-01 -8.87201011e-01 8.41321886e-01 9.31847170e-02
-4.50873673e-01 -1.50055563e+00 4.20171827e-01 5.88556230e-01
-5.13087027e-02 -1.81045428e-01 1.33795750e+00 -1.44029295e+00
4.80430216e-01 -5.59202969e-01 -3.73351157e-01 6.46667898e-01
2.98172534e-01 -2.29097709e-01 -1.13518119e+00 -1.35862142e-01
-2.59108692e-01 -9.13432419e-01 7.38000810e-01 9.04315040e-02
7.98647463e-01 1.30777419e-01 -4.18217659e-01 8.89259428e-02
1.10680223e+00 -9.39812213e-02 1.77719772e-01 4.38114494e-01
4.88808781e-01 8.67653906e-01 6.69557571e-01 -3.15612629e-02
3.04112375e-01 5.10592282e-01 -4.25842941e-01 -1.72014505e-01
2.15414241e-01 -2.85154343e-01 4.43849951e-01 5.65606415e-01
4.00169402e-01 -1.76369444e-01 -1.08083403e+00 6.74677551e-01
-1.80212331e+00 -7.56893814e-01 -2.43515626e-01 2.26267838e+00
1.31483507e+00 2.69096047e-01 4.07905459e-01 2.93322921e-01
8.35445046e-01 -2.56153613e-01 -4.90186781e-01 -4.91579920e-01
-3.08183581e-01 3.87879550e-01 5.78371167e-01 3.50176960e-01
-1.37730980e+00 1.01535499e+00 7.23705435e+00 9.08076644e-01
-1.09392476e+00 1.69378266e-01 6.53549731e-01 2.16620341e-01
4.35713120e-02 -5.30939586e-02 -8.38430226e-01 4.79645222e-01
9.45287347e-01 -2.97896445e-01 -8.17278698e-02 9.55672860e-01
-4.74399030e-01 -3.88223320e-01 -1.14136422e+00 9.07261431e-01
1.93214774e-01 -9.28797185e-01 -2.53892004e-01 -9.61616859e-02
8.71712923e-01 -6.27120882e-02 1.98841155e-01 5.12397230e-01
5.55848122e-01 -9.33807909e-01 3.49367023e-01 -1.74274072e-01
1.01686442e+00 -7.33487606e-01 7.18086183e-01 3.66841912e-01
-6.16329372e-01 1.73545964e-02 -3.47506702e-01 -6.01295643e-02
-4.82845694e-01 7.35249758e-01 -1.02162480e+00 2.67531991e-01
5.67928851e-01 6.68300986e-01 -7.63405442e-01 5.46464980e-01
-8.18148777e-02 9.32675958e-01 -1.04969949e-01 -3.34664702e-01
9.50579569e-02 -3.71656477e-01 1.54972941e-01 1.31604457e+00
6.77145272e-02 -1.96403325e-01 2.03149438e-01 5.07839799e-01
-2.60688394e-01 4.68803704e-01 -8.52813780e-01 -2.46899158e-01
1.45230696e-01 1.21790981e+00 -6.45211041e-01 -3.87893498e-01
-6.58985198e-01 1.04375803e+00 6.38088405e-01 2.96354681e-01
-7.18885064e-01 -5.94385684e-01 5.53345084e-01 3.20027508e-02
2.16249377e-03 -5.14540523e-02 -6.63237453e-01 -1.36218524e+00
1.18737623e-01 -9.19259369e-01 5.38421571e-01 -5.53690493e-01
-1.95421469e+00 2.89282531e-01 7.46974871e-02 -1.15997708e+00
-1.64502457e-01 -1.25557947e+00 -3.93839985e-01 9.91088450e-01
-1.26125014e+00 -1.13406384e+00 7.59160444e-02 6.66746438e-01
2.35890031e-01 -4.45935369e-01 1.08647144e+00 1.35616764e-01
-2.79515803e-01 7.92476594e-01 6.33050263e-01 5.03796697e-01
1.59495592e+00 -1.42702138e+00 -1.75728630e-02 4.94268417e-01
1.03214271e-01 3.55143398e-01 4.83170778e-01 -5.79348087e-01
-9.93001938e-01 -7.94858694e-01 1.03698587e+00 -7.73168862e-01
1.20851898e+00 -5.17808974e-01 -1.08101583e+00 6.93704665e-01
-3.24752741e-02 -1.43678427e-01 1.22527897e+00 6.57204151e-01
-9.31559861e-01 4.43675146e-02 -1.16383040e+00 3.42714041e-01
5.39525032e-01 -1.06943822e+00 -7.89205551e-01 5.09441733e-01
1.14981785e-01 -3.25953782e-01 -9.49038029e-01 2.35851228e-01
5.02661049e-01 -6.83051229e-01 7.19266951e-01 -1.10103989e+00
6.99400783e-01 1.15636244e-01 -1.39331207e-01 -1.64685988e+00
-2.36002162e-01 -1.21018931e-03 2.50217617e-01 1.38198161e+00
6.95784807e-01 -7.61983991e-01 8.15961182e-01 7.30971456e-01
3.52452546e-01 -4.81510848e-01 -8.95682991e-01 -8.85955632e-01
8.46422851e-01 -3.53930324e-01 5.56965880e-02 1.50182843e+00
6.89388394e-01 7.92940259e-01 -3.05499323e-02 -2.03312784e-01
5.09883344e-01 3.56683999e-01 5.66263020e-01 -1.70441985e+00
-1.79044664e-01 -4.75603580e-01 -4.42933857e-01 -6.45847440e-01
1.01662457e+00 -1.20968091e+00 -1.46216482e-01 -1.07695580e+00
5.58901191e-01 -6.18282855e-01 -1.36033222e-01 4.36351180e-01
-4.63245779e-01 3.93617392e-01 -2.38225460e-01 4.68087718e-02
-3.96361291e-01 2.52281487e-01 9.32126105e-01 -3.17894280e-01
-8.20378587e-02 -3.34841609e-02 -7.79659569e-01 5.89413643e-01
1.06273496e+00 -5.95891476e-01 -4.35128450e-01 -1.08180627e-01
2.52358049e-01 -2.66153902e-01 2.05923691e-01 -7.00175226e-01
2.36190837e-02 -1.32964715e-01 4.52528059e-01 -8.17160830e-02
1.50974452e-01 -5.52899003e-01 -5.39179027e-01 1.79233924e-01
-6.28572524e-01 -7.17351660e-02 2.77484208e-01 5.82383692e-01
-2.73202896e-01 -5.40556014e-01 1.05361056e+00 1.82472259e-01
-3.40853393e-01 -2.21288741e-01 -4.42469865e-01 6.43147111e-01
1.10848081e+00 -1.76883325e-01 -4.52535450e-01 -1.76339298e-01
-3.74368221e-01 -1.08407520e-01 7.25030601e-01 4.91508096e-01
1.68725569e-02 -1.28892601e+00 -7.44029999e-01 -8.41809288e-02
5.10015786e-01 -5.57896435e-01 -2.10629165e-01 8.54251742e-01
-8.35036412e-02 6.17837608e-01 -1.48446172e-01 -4.96396631e-01
-1.49749613e+00 4.92889136e-01 2.75132537e-01 -6.41692877e-01
-8.74607414e-02 7.47607708e-01 2.48110384e-01 -9.97640908e-01
-1.56619892e-01 -2.12140471e-01 -1.01487055e-01 4.03117001e-01
2.59800881e-01 1.75248995e-01 3.69212478e-01 -5.30726671e-01
-3.07894498e-01 5.03149450e-01 -2.45342582e-01 -2.68894285e-01
8.94468069e-01 3.56183589e-01 4.03060690e-02 6.98950589e-01
1.25212216e+00 2.36272201e-01 -6.00988388e-01 -4.26991761e-01
1.29190281e-01 -2.43981734e-01 6.15715282e-03 -1.01209390e+00
-3.45882207e-01 7.38449275e-01 3.64551753e-01 3.77766073e-01
8.87622476e-01 1.17920496e-01 4.32775944e-01 2.78515190e-01
3.24382752e-01 -1.41115451e+00 9.20122042e-02 7.79276073e-01
3.50317478e-01 -1.66165245e+00 -5.46320751e-02 -5.08513272e-01
-8.80341291e-01 1.20916605e+00 5.26818514e-01 2.25704834e-02
6.54292524e-01 1.60886928e-01 2.10197315e-01 1.00493193e-01
-5.65683067e-01 5.70205599e-03 3.87061149e-01 8.21970224e-01
8.44039142e-01 2.14846313e-01 -3.93095702e-01 7.01331675e-01
-1.66216969e-01 -2.02457622e-01 3.75017911e-01 6.82830751e-01
-1.83861837e-01 -1.50106514e+00 -5.56221783e-01 7.44745135e-01
-6.09749079e-01 -1.73709869e-01 -9.26417708e-01 6.51387393e-01
-3.31333071e-01 9.18505907e-01 -2.04126045e-01 -3.64827693e-01
1.87424108e-01 6.02737844e-01 7.02725172e-01 -6.34907961e-01
-6.78145468e-01 -4.45923239e-01 3.65976065e-01 1.26667395e-01
-5.59544206e-01 -8.34490716e-01 -9.07065332e-01 -1.91938639e-01
-5.44097543e-01 4.50103372e-01 4.28963691e-01 9.47037160e-01
2.20145851e-01 9.88734439e-02 4.90139961e-01 -3.01055461e-01
-8.90673697e-01 -1.11598158e+00 -7.83026457e-01 6.19302571e-01
5.05387038e-02 -7.59446979e-01 -4.55241293e-01 1.69176489e-01] | [10.426767349243164, 8.92202377319336] |
ba799f81-0143-4f2f-bd69-bdf1383cb036 | simple-sorting-criteria-help-find-the-causal | 2303.18211 | null | https://arxiv.org/abs/2303.18211v1 | https://arxiv.org/pdf/2303.18211v1.pdf | Simple Sorting Criteria Help Find the Causal Order in Additive Noise Models | Additive Noise Models (ANM) encode a popular functional assumption that enables learning causal structure from observational data. Due to a lack of real-world data meeting the assumptions, synthetic ANM data are often used to evaluate causal discovery algorithms. Reisach et al. (2021) show that, for common simulation parameters, a variable ordering by increasing variance is closely aligned with a causal order and introduce var-sortability to quantify the alignment. Here, we show that not only variance, but also the fraction of a variable's variance explained by all others, as captured by the coefficient of determination $R^2$, tends to increase along the causal order. Simple baseline algorithms can use $R^2$-sortability to match the performance of established methods. Since $R^2$-sortability is invariant under data rescaling, these algorithms perform equally well on standardized or rescaled data, addressing a key limitation of algorithms exploiting var-sortability. We characterize and empirically assess $R^2$-sortability for different simulation parameters. We show that all simulation parameters can affect $R^2$-sortability and must be chosen deliberately to control the difficulty of the causal discovery task and the real-world plausibility of the simulated data. We provide an implementation of the sortability measures and sortability-based algorithms in our library CausalDisco (https://github.com/CausalDisco/CausalDisco). | ['Sebastian Weichwald', 'Antoine Chambaz', 'Christof Seiler', 'Myriam Tami', 'Alexander G. Reisach'] | 2023-03-31 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 1.73366874e-01 2.58417670e-02 -5.56179345e-01 -3.09951305e-01
-6.75187647e-01 -6.61557257e-01 1.03223419e+00 2.51721919e-01
-3.44078720e-01 1.02003658e+00 5.60227692e-01 -5.89430153e-01
-7.06567764e-01 -9.73630071e-01 -1.00980532e+00 -5.26998818e-01
-6.69361472e-01 2.95797318e-01 1.38857201e-01 1.73133135e-01
1.63121253e-01 2.50858694e-01 -1.35572827e+00 -1.23598725e-01
7.53276765e-01 4.77129132e-01 1.00310072e-01 5.36676586e-01
2.87892908e-01 4.63166505e-01 -3.77511322e-01 -2.45967865e-01
3.17067921e-01 -8.73729944e-01 -6.60607278e-01 -6.63410187e-01
1.69224918e-01 6.62548393e-02 -4.78269428e-01 1.10121405e+00
3.22078168e-01 1.18520170e-01 9.69499826e-01 -1.43949640e+00
-3.71549636e-01 9.70727503e-01 -4.60311592e-01 4.28396285e-01
2.99952775e-01 4.92912084e-01 1.27759647e+00 -3.58340502e-01
7.16321170e-01 1.59988570e+00 6.09056771e-01 4.30901423e-02
-1.78927302e+00 -7.95016766e-01 1.10140525e-01 -7.54015073e-02
-1.33428109e+00 -1.95430964e-01 3.94536227e-01 -7.21412480e-01
4.80593085e-01 4.75881279e-01 4.25001264e-01 1.18269455e+00
3.97063404e-01 5.86938597e-02 1.34485841e+00 -3.08473319e-01
4.13715899e-01 -4.37373847e-01 2.05952093e-01 4.84992832e-01
5.82712591e-01 7.38175929e-01 -6.25947297e-01 -4.86594796e-01
9.12227035e-01 -2.46143475e-01 -2.61186510e-01 -2.06887215e-01
-1.27108407e+00 8.86124074e-01 3.13784599e-01 -7.27573484e-02
-2.47982025e-01 8.10395598e-01 2.73840100e-01 2.76073545e-01
2.64290661e-01 6.96646035e-01 -6.29783690e-01 -3.25304627e-01
-5.76768994e-01 5.76525927e-01 4.22119349e-01 6.40563667e-01
4.31625158e-01 -1.74695045e-01 -2.23631591e-01 5.26225030e-01
3.13880503e-01 7.60719478e-01 1.99066892e-01 -1.09219313e+00
2.22171471e-01 5.02714336e-01 2.93960780e-01 -9.04307306e-01
-6.70523524e-01 -2.53946304e-01 -8.31402481e-01 -3.42812538e-02
8.50815415e-01 -2.06822976e-01 -8.48119140e-01 2.50488520e+00
1.62343845e-01 2.55743742e-01 -3.43620002e-01 8.11631262e-01
3.41768771e-01 3.19845557e-01 3.09427321e-01 -5.24610639e-01
1.36112571e+00 1.34863764e-01 -4.61751848e-01 -1.93917155e-01
4.20247376e-01 -5.46274304e-01 1.35844851e+00 -1.55242812e-03
-8.85660589e-01 -7.45761767e-02 -8.05165768e-01 4.40404236e-01
-9.93597228e-03 -6.52833641e-01 8.93724680e-01 4.92632449e-01
-5.74497581e-01 8.00058484e-01 -8.98707628e-01 -2.81736344e-01
2.87015289e-01 7.48870075e-02 -2.01319247e-01 1.20966762e-01
-1.54424429e+00 5.95172822e-01 7.98124671e-02 -2.69762754e-01
-1.17347240e+00 -1.24710989e+00 -5.20122051e-01 1.68119475e-01
5.28324425e-01 -8.27781975e-01 1.06444693e+00 -3.32159340e-01
-7.83450782e-01 4.50498641e-01 -1.93804294e-01 -4.07469690e-01
5.83392143e-01 -5.97891398e-02 -3.30675185e-01 -4.20973003e-01
2.53380626e-01 1.76602349e-01 3.12546194e-01 -1.19253218e+00
-3.95093530e-01 -2.67624527e-01 -7.31781051e-02 -6.96235970e-02
2.64694691e-01 1.52370587e-01 8.56468338e-04 -7.31940567e-01
1.74312964e-01 -8.47014070e-01 -4.46122199e-01 -2.42105618e-01
-5.68010688e-01 -3.57711077e-04 -7.23890588e-03 -1.57858491e-01
1.47540653e+00 -1.94344294e+00 7.20194131e-02 3.37176085e-01
2.84294695e-01 -4.53302205e-01 -1.29263625e-01 5.63488603e-01
-4.04172212e-01 6.01305842e-01 -6.34456217e-01 2.68810958e-01
1.74233228e-01 1.83942661e-01 -2.52326936e-01 6.99272692e-01
9.85762477e-02 7.36194968e-01 -1.11178708e+00 -1.63683772e-01
-7.10160937e-04 1.51461083e-02 -6.52375162e-01 1.66316152e-01
-3.63887519e-01 3.41851562e-01 -2.32868940e-01 1.36116832e-01
3.38784277e-01 -2.21638262e-01 4.70821500e-01 6.77162036e-02
-3.20742875e-01 5.95478177e-01 -1.43231058e+00 9.56623077e-01
-3.65764290e-01 4.60539967e-01 -2.10601047e-01 -9.95017946e-01
6.22599125e-01 1.20903440e-01 4.53122705e-01 -5.52580655e-01
-4.78134975e-02 2.23589957e-01 4.56382304e-01 -2.62275696e-01
3.27011384e-02 -4.39049631e-01 -3.17694753e-01 4.97980118e-01
-3.29854727e-01 -2.81191230e-01 8.60918388e-02 1.93188623e-01
1.56177962e+00 -1.81661323e-01 5.68935156e-01 -8.13042819e-01
-4.28554505e-01 9.75602269e-02 9.01937306e-01 1.14829457e+00
2.55446341e-02 4.16803658e-01 1.15330637e+00 -1.49318486e-01
-1.19009197e+00 -1.40081871e+00 -5.49551725e-01 7.44441271e-01
2.57866345e-02 -4.30798948e-01 -3.75916034e-01 -3.83404613e-01
3.52321506e-01 9.96360838e-01 -1.00776207e+00 -3.57677579e-01
-3.96329522e-01 -1.42826140e+00 6.99699402e-01 3.95434886e-01
1.01799533e-01 -6.59783125e-01 -5.77973485e-01 7.73108974e-02
1.90915093e-02 -5.06191909e-01 -2.44362488e-01 4.28018093e-01
-6.37074709e-01 -1.43183005e+00 -4.63243201e-02 1.97846636e-01
4.18344676e-01 -1.08047917e-01 1.24316752e+00 -1.76578373e-01
-2.36466944e-01 -7.06898496e-02 -9.58888978e-02 -3.87704968e-01
-5.11108994e-01 -2.73293912e-01 2.29098484e-01 -5.04386008e-01
5.97730428e-02 -9.26610231e-01 -5.79222679e-01 3.63465339e-01
-8.19117069e-01 -1.60037309e-01 2.08363444e-01 8.53850722e-01
5.04618168e-01 2.15507969e-01 6.19005144e-01 -8.83814454e-01
5.80178440e-01 -7.43767321e-01 -1.01296079e+00 -9.99860317e-02
-9.13427174e-01 3.67074639e-01 4.12848055e-01 -4.77518380e-01
-6.98086023e-01 -3.96658957e-01 2.81799227e-01 -5.23724370e-02
9.48991477e-02 9.66312289e-01 -2.90340304e-01 6.87537193e-01
9.35683489e-01 -3.44824761e-01 -2.33079791e-01 -3.76308262e-01
5.19272149e-01 1.32754162e-01 5.06738245e-01 -7.27546930e-01
7.33914256e-01 4.44191962e-01 4.30431694e-01 -3.85067433e-01
-6.15325809e-01 1.00465521e-01 -3.04191768e-01 1.96821522e-02
5.61720669e-01 -7.10638404e-01 -9.87070382e-01 7.77509138e-02
-7.15687871e-01 -8.35329652e-01 -3.85365069e-01 8.73201489e-01
-5.56521118e-01 -1.39039969e-02 -1.98813424e-01 -9.24917936e-01
2.77770638e-01 -1.15839767e+00 6.88611448e-01 -1.26392931e-01
-5.17686784e-01 -1.03841400e+00 3.33495378e-01 -2.07104996e-01
1.26458749e-01 4.72089529e-01 1.37721455e+00 -4.65199620e-01
-3.56535941e-01 2.20657319e-01 -2.69113779e-01 -3.31126899e-01
1.29607931e-01 2.71237165e-01 -6.13858819e-01 4.99622561e-02
-3.23042601e-01 9.89059731e-02 9.36154068e-01 9.38116789e-01
9.95526433e-01 -5.95127404e-01 -4.75881040e-01 5.13476133e-01
1.18255460e+00 1.96756244e-01 5.43698311e-01 2.75022715e-01
5.21377981e-01 5.61727464e-01 2.59613663e-01 5.72415113e-01
3.10512930e-01 7.20737219e-01 4.89916235e-01 1.17227979e-01
-2.17136834e-03 -6.46213710e-01 9.02799740e-02 1.21902056e-01
7.72380456e-02 -3.74521494e-01 -1.24093223e+00 5.53239524e-01
-1.85273778e+00 -1.14771259e+00 -5.92406213e-01 2.64867663e+00
1.16320634e+00 4.05287683e-01 2.62092292e-01 -8.16130564e-02
6.26817048e-01 9.53447446e-02 -5.42969525e-01 -1.09613165e-01
-2.87053585e-01 2.74089035e-02 1.05419207e+00 8.82315695e-01
-7.53282309e-01 5.33047438e-01 7.01932955e+00 7.44714320e-01
-7.62446463e-01 -3.27017531e-02 6.07537270e-01 -2.08122954e-01
-7.39484251e-01 3.76340210e-01 -3.27838570e-01 7.04070330e-01
1.28068125e+00 -7.35632479e-01 3.83390367e-01 4.20220941e-01
8.27833951e-01 -3.20176512e-01 -1.29593241e+00 3.72685075e-01
-7.87656665e-01 -1.26219785e+00 -3.13399494e-01 3.01024258e-01
6.91393435e-01 1.49846852e-01 -1.79051876e-03 -4.25939448e-02
1.24503088e+00 -1.30200660e+00 7.15127826e-01 5.15093088e-01
9.64443743e-01 -5.41135311e-01 5.13144791e-01 1.45383224e-01
-9.24262345e-01 -6.41870573e-02 -2.26209581e-01 -4.48580950e-01
2.24939749e-01 1.23694909e+00 -6.69994712e-01 4.79894251e-01
7.26087451e-01 3.68426025e-01 -3.34871829e-01 7.87081242e-01
-3.97875905e-01 1.23550498e+00 -6.10572457e-01 1.55547202e-01
-1.06136449e-01 -8.37569535e-02 7.53412604e-01 1.05577040e+00
2.43515685e-01 3.59770656e-01 -1.78691983e-01 1.10236406e+00
-1.89534366e-01 -2.88115710e-01 -5.62706947e-01 -2.13733643e-01
9.83621597e-01 4.76798743e-01 -5.81681550e-01 -1.16812661e-01
-2.15063676e-01 1.43870711e-01 -5.44556081e-02 3.83750826e-01
-8.87252092e-01 2.62923203e-02 1.04786980e+00 2.78155655e-01
-2.38174826e-01 -1.40527070e-01 -5.74201524e-01 -9.09083962e-01
-3.13744575e-01 -8.68520021e-01 6.84650302e-01 -6.34647667e-01
-1.34518659e+00 -6.30319566e-02 6.33974552e-01 -7.67211497e-01
-2.26553977e-01 -3.21046621e-01 -4.86394435e-01 9.65879917e-01
-6.84868038e-01 -5.48512697e-01 1.23621318e-02 4.19558994e-02
1.82512626e-01 2.96197414e-01 5.73182762e-01 -6.56930730e-02
-7.75474072e-01 3.89019608e-01 3.12503189e-01 -2.00241044e-01
7.09838569e-01 -1.45096004e+00 5.21013200e-01 9.30576324e-01
-6.07183054e-02 1.02720428e+00 1.34503031e+00 -1.02057374e+00
-1.17313993e+00 -1.01125777e+00 5.68088531e-01 -6.63724124e-01
1.15599895e+00 -4.29247409e-01 -7.82372415e-01 8.19001019e-01
-2.29230836e-01 -4.72059585e-02 4.29937124e-01 6.09565079e-01
-5.85429668e-01 -2.58225556e-02 -9.74174976e-01 9.20802057e-01
1.45543897e+00 -3.00481975e-01 -4.41532046e-01 2.64599234e-01
8.82871091e-01 3.06145009e-02 -1.06918037e+00 7.58997738e-01
6.74607933e-01 -8.32041264e-01 7.58541465e-01 -8.48202944e-01
6.79160416e-01 -2.98809230e-01 -4.17176366e-01 -1.47794557e+00
-2.81648785e-01 -6.39007270e-01 3.37897986e-01 1.10192978e+00
6.93621159e-01 -7.67034709e-01 2.59142727e-01 6.73482180e-01
2.46786132e-01 -3.32744062e-01 -1.13199306e+00 -8.82272899e-01
4.98129606e-01 -7.94047892e-01 6.12863660e-01 1.17217982e+00
2.92027388e-02 2.78054774e-01 -1.36809587e-01 3.55822325e-01
8.25331390e-01 1.00921772e-01 6.28288329e-01 -1.32817495e+00
-4.47488606e-01 -7.40744174e-01 -2.55104423e-01 -5.49643099e-01
-5.49897812e-02 -7.27588177e-01 -2.86784954e-03 -1.23579276e+00
5.27600586e-01 -6.89432144e-01 -2.58845896e-01 5.34431517e-01
-5.16661882e-01 -1.17569894e-01 -4.05406915e-02 1.57665923e-01
1.38880685e-01 3.98374587e-01 7.32518792e-01 1.41488582e-01
-3.34420264e-01 -2.87398398e-01 -7.00586736e-01 7.93503642e-01
7.61713266e-01 -8.52639854e-01 -5.89529693e-01 -1.20089471e-01
6.64149761e-01 2.83595413e-01 7.90244937e-01 -4.08612400e-01
-2.12001547e-01 -9.10429299e-01 3.98983359e-02 -1.99267834e-01
-2.94574171e-01 -3.51473689e-01 7.91785836e-01 4.99101102e-01
-8.09117377e-01 2.68494427e-01 1.95065960e-01 5.26316226e-01
2.58465737e-01 -6.06312789e-02 6.27934217e-01 7.20507503e-02
-1.05305806e-01 -4.77281474e-02 -3.53992522e-01 4.04745787e-01
6.21266425e-01 4.59663838e-01 -6.87758148e-01 -4.52935129e-01
-3.71928841e-01 2.23048881e-01 5.07733822e-01 2.60885864e-01
-1.60383787e-02 -1.10039663e+00 -9.23874021e-01 -2.04799995e-01
1.21681429e-01 -2.46581823e-01 4.95035388e-02 8.93054128e-01
-1.12725250e-01 2.68227279e-01 3.12127739e-01 -3.58340532e-01
-8.27637911e-01 7.00444460e-01 4.25297052e-01 -1.95173755e-01
-3.45679820e-01 6.75295293e-01 5.39835334e-01 -2.37925991e-01
-1.46315843e-01 -3.06523561e-01 3.79050463e-01 6.04977831e-02
4.20697898e-01 5.46581864e-01 -4.29057509e-01 -2.46850923e-02
-5.49984694e-01 1.38802618e-01 5.20488441e-01 -5.84577739e-01
1.37137163e+00 -1.52122676e-01 -1.37125105e-01 7.24953651e-01
7.90476203e-01 2.36450970e-01 -1.28807580e+00 4.07599770e-02
1.08809911e-01 -4.43232805e-01 -6.91677853e-02 -1.03959167e+00
-6.71812177e-01 3.47030371e-01 3.69954258e-01 4.58066702e-01
8.04815233e-01 2.29105398e-01 -2.31380790e-01 -5.10300137e-02
3.42019826e-01 -5.78403473e-01 -3.14716786e-01 1.43236846e-01
9.71569717e-01 -9.92208719e-01 2.88623214e-01 -3.40224355e-01
-2.59440571e-01 3.87333721e-01 3.12647998e-01 -2.96305232e-02
6.58200979e-01 5.12966096e-01 -1.64485395e-01 -3.07906687e-01
-1.04274714e+00 -1.67990789e-01 3.24476101e-02 3.47768843e-01
4.99566406e-01 4.40665841e-01 -8.86003315e-01 5.86906791e-01
-5.26603222e-01 -2.20860273e-01 6.16793275e-01 4.19750452e-01
-2.42326185e-01 -8.83129060e-01 -5.36726296e-01 6.96546614e-01
-4.19971585e-01 -2.90654391e-01 -2.23156184e-01 1.12638783e+00
-1.14814654e-01 1.04905450e+00 2.56490052e-01 -3.23236614e-01
3.81465793e-01 -3.18616331e-02 1.74613237e-01 -3.80821466e-01
-1.84773669e-01 1.07559808e-01 3.02930176e-01 -6.65994763e-01
-3.70934725e-01 -9.51236367e-01 -1.19138193e+00 -6.01321936e-01
1.04586836e-02 1.78350881e-01 3.53052735e-01 8.34392190e-01
2.40109637e-01 6.09548867e-01 5.74844778e-01 -2.53858685e-01
-4.48769987e-01 -1.00383985e+00 -5.47592282e-01 4.25481856e-01
1.98563308e-01 -1.05438280e+00 -7.68025219e-01 2.40741111e-02] | [7.92240571975708, 5.321852207183838] |
2fc6e643-3a59-492b-ba12-3ab235e5c51a | unsupervised-learning-on-3d-point-clouds-by | 2202.02543 | null | https://arxiv.org/abs/2202.02543v2 | https://arxiv.org/pdf/2202.02543v2.pdf | Unsupervised Learning on 3D Point Clouds by Clustering and Contrasting | Learning from unlabeled or partially labeled data to alleviate human labeling remains a challenging research topic in 3D modeling. Along this line, unsupervised representation learning is a promising direction to auto-extract features without human intervention. This paper proposes a general unsupervised approach, named \textbf{ConClu}, to perform the learning of point-wise and global features by jointly leveraging point-level clustering and instance-level contrasting. Specifically, for one thing, we design an Expectation-Maximization (EM) like soft clustering algorithm that provides local supervision to extract discriminating local features based on optimal transport. We show that this criterion extends standard cross-entropy minimization to an optimal transport problem, which we solve efficiently using a fast variant of the Sinkhorn-Knopp algorithm. For another, we provide an instance-level contrasting method to learn the global geometry, which is formulated by maximizing the similarity between two augmentations of one point cloud. Experimental evaluations on downstream applications such as 3D object classification and semantic segmentation demonstrate the effectiveness of our framework and show that it can outperform state-of-the-art techniques. | ['Mohammed Bennamoun', 'Jian Zhang', 'Qiang Wu', 'Litao Yu', 'Guofeng Mei'] | 2022-02-05 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [ 7.06116334e-02 3.07642639e-01 -2.26041615e-01 -7.07717299e-01
-1.02335572e+00 -5.26200593e-01 5.29651463e-01 2.50206232e-01
-2.06245407e-01 3.21215004e-01 -4.76224869e-01 -2.33136520e-01
-4.00721222e-01 -7.07231104e-01 -8.05386782e-01 -9.38942730e-01
-1.11249778e-04 6.87339246e-01 1.82143912e-01 1.01106785e-01
3.86318743e-01 8.81219745e-01 -1.51079857e+00 -5.42297997e-02
1.14845932e+00 1.19902539e+00 4.29291397e-01 1.52615920e-01
-3.02631646e-01 7.01246440e-01 -9.75047424e-02 4.10579406e-02
5.44102967e-01 -1.05754711e-01 -1.01738989e+00 7.09398508e-01
1.51124910e-01 -6.94306493e-02 1.69326365e-01 9.10727978e-01
3.53328854e-01 4.21539307e-01 1.06711674e+00 -1.16700411e+00
-3.65379691e-01 7.32246488e-02 -8.21798861e-01 -7.43571818e-02
-1.75941624e-02 -1.34138064e-03 9.49610949e-01 -1.04290926e+00
5.62612414e-01 1.04221427e+00 4.75323915e-01 2.51690954e-01
-1.17729056e+00 -1.53635234e-01 3.08280140e-01 3.98554327e-03
-1.49546397e+00 -7.35786557e-02 1.06108868e+00 -6.47332728e-01
7.77080894e-01 1.66103512e-01 3.27978820e-01 4.48410034e-01
-2.96902269e-01 1.08789515e+00 1.05891740e+00 -4.82609749e-01
5.55159628e-01 3.26665640e-01 4.24984889e-03 8.25378895e-01
-1.19137473e-01 -1.45344958e-01 -3.01151454e-01 -2.78706737e-02
7.07551658e-01 -6.59019174e-03 -3.86393704e-02 -1.00589371e+00
-9.98303294e-01 8.56599927e-01 5.45910478e-01 3.59672196e-02
-2.31249601e-01 -2.80764848e-01 4.48531918e-02 2.10820604e-02
1.04266691e+00 1.51270598e-01 -6.44412577e-01 2.30471700e-01
-1.11431921e+00 1.25934273e-01 6.41531408e-01 1.27290022e+00
1.32623589e+00 -2.28057653e-01 1.84594899e-01 8.80555868e-01
5.19458890e-01 3.15101594e-01 -1.86515804e-02 -1.07186806e+00
3.89941335e-01 8.88133764e-01 3.10552735e-02 -8.12559307e-01
-4.26602244e-01 -4.47270215e-01 -7.06425071e-01 2.85099864e-01
1.85193777e-01 -2.27575395e-02 -9.47689831e-01 1.23821723e+00
8.30848813e-01 5.23441494e-01 -2.89876819e-01 1.04213583e+00
4.43772048e-01 5.62641799e-01 -9.85138267e-02 -2.85111547e-01
8.76960993e-01 -9.45542991e-01 -2.37617910e-01 6.60943240e-02
9.79715586e-01 -5.20006418e-01 6.26006961e-01 1.10133886e-01
-1.01163554e+00 -5.03116369e-01 -8.37665379e-01 -3.42304297e-02
-4.26304400e-01 8.29651803e-02 6.86385930e-01 6.65364385e-01
-8.15005481e-01 7.48785436e-01 -1.03526580e+00 -2.52504736e-01
8.50445449e-01 5.01213253e-01 -2.53577590e-01 2.10454147e-02
-5.19327283e-01 6.93661213e-01 3.24878275e-01 1.41030118e-01
-7.25682795e-01 -8.87086332e-01 -1.04600823e+00 -2.69809663e-01
5.92840016e-01 -7.41508842e-01 9.53440368e-01 -3.73850137e-01
-1.63723683e+00 1.22200310e+00 -2.10609779e-01 -2.43970022e-01
4.58041996e-01 -1.88083604e-01 2.08322734e-01 4.80478972e-01
3.57426167e-01 6.90338135e-01 8.89184892e-01 -1.68258214e+00
-6.20837331e-01 -6.55026734e-01 -1.84807137e-01 3.75924647e-01
-2.01020569e-01 -2.23828137e-01 -4.92481500e-01 -3.75671864e-01
6.29936576e-01 -8.62678647e-01 -7.92500615e-01 1.64458856e-01
-5.31420469e-01 -3.78861040e-01 1.05345964e+00 -2.87294835e-01
6.13533676e-01 -2.03328729e+00 2.86288381e-01 6.68703675e-01
2.62438744e-01 3.00213438e-03 2.47234911e-01 1.78547204e-02
8.29533711e-02 1.34913966e-01 -8.53116572e-01 -6.20608449e-01
1.49323707e-02 2.50446796e-01 -1.58072747e-02 8.61827850e-01
6.22882009e-01 8.68250072e-01 -9.34268475e-01 -7.21480846e-01
7.49683559e-01 2.37085730e-01 -6.14565253e-01 1.94680646e-01
-3.49861652e-01 1.03336263e+00 -8.64085555e-01 8.11344206e-01
1.09079504e+00 -3.79862577e-01 -1.26812682e-01 -1.04469128e-01
-2.51475155e-01 6.22418784e-02 -1.22934926e+00 2.22548437e+00
-4.85496610e-01 8.28549862e-02 1.52260438e-01 -1.66278577e+00
1.12907577e+00 -9.83066950e-03 1.02140665e+00 -2.03597397e-01
2.12003261e-01 2.95280844e-01 -7.75611043e-01 -5.02591431e-01
2.09323764e-01 -1.24417342e-01 -7.44330660e-02 3.59195441e-01
1.99796990e-01 -5.96977055e-01 -1.79014459e-01 1.36807501e-01
8.24607074e-01 5.64572692e-01 1.26704618e-01 -3.73240203e-01
6.69236660e-01 2.84128375e-02 5.39880991e-01 6.54336810e-01
-1.10965870e-01 5.95761240e-01 1.03076227e-01 -1.47370666e-01
-6.82510018e-01 -1.05445898e+00 -2.57843167e-01 6.60096228e-01
5.83679199e-01 -2.99053788e-01 -9.16752458e-01 -1.07581890e+00
2.14539263e-02 5.99053800e-01 -3.58728200e-01 7.12807104e-02
-5.10190904e-01 -7.08033562e-01 3.38627771e-02 5.04580200e-01
5.40573835e-01 -4.82270181e-01 -4.44872171e-01 5.94922751e-02
-7.66972303e-02 -1.25724924e+00 -1.24014169e-01 5.08309960e-01
-1.15795624e+00 -7.62836456e-01 -5.93334019e-01 -9.58130419e-01
9.83277202e-01 4.39431489e-01 8.80886793e-01 2.19921069e-03
-3.64646226e-01 4.71734792e-01 -4.34875548e-01 -2.38695130e-01
1.76667809e-01 2.90585250e-01 -5.34834787e-02 2.39269137e-01
2.74950117e-01 -7.98623204e-01 -4.85760778e-01 6.60996437e-01
-7.26126611e-01 -1.93650857e-01 5.69678307e-01 4.99437094e-01
1.27213669e+00 1.58805564e-01 3.69635195e-01 -8.79457057e-01
-7.83551310e-04 -7.69343019e-01 -6.27234221e-01 1.49219200e-01
-4.01511639e-01 9.24736485e-02 3.56415153e-01 -1.72399636e-02
-1.15463805e+00 6.55221820e-01 -5.81420073e-03 -7.84205914e-01
-4.91050899e-01 2.10677892e-01 -6.31964803e-01 -3.38911504e-01
2.27970243e-01 3.26788634e-01 -4.26083133e-02 -6.38366818e-01
7.64097273e-01 7.18320549e-01 3.21225464e-01 -8.83499861e-01
1.21451175e+00 9.43804979e-01 2.44347796e-01 -1.01219606e+00
-1.10120749e+00 -9.88926530e-01 -1.29340625e+00 -3.50801826e-01
1.09865975e+00 -7.24412739e-01 -6.16751850e-01 2.13803142e-01
-9.53107417e-01 -2.08556786e-01 -5.06633341e-01 6.59767509e-01
-1.06505501e+00 4.61493760e-01 -2.17373133e-01 -9.47853208e-01
1.00497261e-01 -1.07024467e+00 1.64918828e+00 1.27883404e-01
2.68314451e-01 -1.19426191e+00 -1.26636371e-01 6.76554680e-01
-1.66562691e-01 3.97707850e-01 8.29356313e-01 -5.50065517e-01
-9.37217772e-01 -7.57234776e-03 -2.73693144e-01 4.99612898e-01
1.62490532e-01 -1.97435200e-01 -9.60486889e-01 -7.13209584e-02
3.47325444e-01 -1.92325264e-01 7.88178682e-01 3.90370488e-01
1.57928550e+00 8.24416205e-02 -6.11321092e-01 9.09391820e-01
1.27165246e+00 -2.90631473e-01 3.23616832e-01 1.89910848e-02
9.40716088e-01 7.82356799e-01 9.20755446e-01 5.18612623e-01
5.11684477e-01 5.32488286e-01 6.06286287e-01 -1.08891673e-01
1.96816444e-01 -2.55336642e-01 -1.51011376e-02 9.58206117e-01
-3.12444419e-01 -4.43490921e-04 -7.21411824e-01 4.66899335e-01
-1.97092736e+00 -6.66612446e-01 -2.78632134e-01 1.91350961e+00
3.27448159e-01 -3.03647574e-02 3.09505332e-02 1.61666557e-01
7.83732116e-01 4.33930643e-02 -5.71415722e-01 -7.42407143e-03
8.01844597e-02 1.89003751e-01 6.50745451e-01 3.88404608e-01
-1.38917553e+00 1.06842053e+00 4.99237537e+00 1.10036337e+00
-8.14977467e-01 1.38413921e-01 6.75452054e-01 2.08272099e-01
-2.22833171e-01 7.53389001e-02 -7.55413234e-01 3.73680949e-01
4.15403277e-01 1.68625027e-01 1.71359554e-01 1.03877568e+00
4.40064818e-01 -7.84783140e-02 -1.05339813e+00 9.59778965e-01
2.82605980e-02 -1.26313138e+00 -8.17685276e-02 3.95328194e-01
9.15729642e-01 -3.51340249e-02 -6.88002408e-02 2.52417345e-02
1.94018915e-01 -6.43222451e-01 7.47448742e-01 5.27269244e-01
5.06558120e-01 -8.30805600e-01 4.17273879e-01 7.46001482e-01
-1.46649837e+00 1.67585298e-01 -4.98621851e-01 2.26657063e-01
4.42190379e-01 9.71624851e-01 -6.91020310e-01 1.03749526e+00
6.99891269e-01 1.08831596e+00 -2.28016987e-01 1.11677861e+00
-1.92663029e-01 4.57243413e-01 -6.56533301e-01 2.27036074e-01
4.34149832e-01 -5.26308179e-01 7.87978113e-01 1.00474620e+00
2.39141375e-01 2.71724373e-01 5.00091255e-01 1.03673756e+00
-1.37276277e-01 2.98571318e-01 -7.23121881e-01 3.88770550e-01
2.57671088e-01 1.37732220e+00 -1.10064447e+00 -2.53986176e-02
-2.63340145e-01 1.05110967e+00 4.35666203e-01 2.14236751e-01
-6.51320398e-01 -3.49950306e-02 4.04148996e-01 5.53989932e-02
6.10760152e-01 -4.55134064e-01 -5.88049173e-01 -1.08805823e+00
2.80466855e-01 2.57192366e-02 2.86214411e-01 -5.08290648e-01
-1.52273965e+00 1.90667689e-01 1.73694968e-01 -1.64003229e+00
4.00709100e-02 -5.61504781e-01 -6.43338442e-01 5.57893753e-01
-1.86697817e+00 -1.20073581e+00 -2.73062378e-01 7.89128304e-01
6.57820225e-01 8.62225592e-02 5.21081150e-01 2.59233803e-01
-4.23175901e-01 2.02512547e-01 6.97163045e-02 -5.99381812e-02
2.39974275e-01 -1.31995642e+00 -1.18188053e-01 5.11156857e-01
2.13129863e-01 3.05250466e-01 1.22740008e-01 -5.82261562e-01
-1.26716828e+00 -1.40753198e+00 5.11462808e-01 -4.83638495e-01
3.53934526e-01 -3.12701344e-01 -7.30455637e-01 4.94775623e-01
-3.11476618e-01 3.34335804e-01 7.47194588e-01 -2.58928359e-01
3.84128392e-02 4.87358347e-02 -1.40571344e+00 1.99784622e-01
1.39011288e+00 -3.12088579e-01 -4.56772745e-01 6.63607061e-01
6.50979996e-01 -3.39597255e-01 -1.04066980e+00 5.99483073e-01
-9.38281417e-02 -8.66308272e-01 1.09451854e+00 -4.47217226e-01
2.35419124e-01 -4.90799159e-01 -3.03430438e-01 -1.20632744e+00
-1.71752051e-02 -7.00978875e-01 -1.77318618e-01 1.33155227e+00
3.76624763e-01 -3.23964149e-01 1.09028482e+00 5.25093079e-01
-4.47249681e-01 -9.64822769e-01 -8.91854346e-01 -9.15116072e-01
1.37586474e-01 -6.63838506e-01 4.85094637e-01 1.13067544e+00
-3.11712265e-01 1.17389843e-01 -1.15189597e-01 5.92297614e-01
1.06255352e+00 3.22457731e-01 7.65174270e-01 -1.40077162e+00
-2.20654178e-02 -1.84137061e-01 -6.54550254e-01 -1.64552641e+00
5.21882594e-01 -1.40840507e+00 2.40574107e-01 -1.51450026e+00
-1.04927674e-01 -7.39796221e-01 -1.72428980e-01 1.91421434e-01
9.67718437e-02 -4.81836591e-03 3.06650307e-02 2.41169959e-01
-8.28726113e-01 8.58923316e-01 1.38569832e+00 -2.04316080e-01
-2.72641003e-01 2.86878914e-01 -4.64255899e-01 9.71912086e-01
5.97685933e-01 -5.18670142e-01 -4.23158884e-01 -3.89824182e-01
-2.49207303e-01 -2.05130219e-01 4.45959270e-01 -8.29486072e-01
3.00960481e-01 -3.40902418e-01 1.91317528e-01 -1.03431630e+00
3.41579854e-01 -1.06000757e+00 -4.90606040e-01 -1.79114729e-01
-1.50013804e-01 -6.41477585e-01 -1.60898715e-01 7.92888224e-01
-1.17257521e-01 -3.72484505e-01 7.78434217e-01 -3.20835173e-01
-7.39379466e-01 6.30042255e-01 -1.75924841e-02 4.88892309e-02
1.36693966e+00 -2.87299335e-01 1.92100972e-01 -5.75205833e-02
-8.51419389e-01 5.44848800e-01 3.87942463e-01 -2.56135799e-02
7.68026233e-01 -1.20692766e+00 -6.00850105e-01 1.10061519e-01
1.04983076e-01 7.40701139e-01 2.66096622e-01 9.46956277e-01
-4.02878284e-01 1.76287934e-01 1.61832511e-01 -1.17404711e+00
-7.00189829e-01 3.74069363e-01 2.70522177e-01 -1.50051326e-01
-7.56170511e-01 9.90167856e-01 9.80781168e-02 -8.78915548e-01
1.47714443e-03 -1.10815667e-01 -1.06526713e-03 -1.21806506e-02
5.90098202e-02 4.72568631e-01 3.05730343e-01 -7.31570542e-01
-4.96171355e-01 9.51568186e-01 1.88076310e-02 8.12491700e-02
1.69491816e+00 -4.78252918e-01 -9.16893110e-02 3.62933338e-01
1.46068549e+00 -2.60708034e-01 -1.54078877e+00 -2.42974639e-01
2.03205213e-01 -5.73148727e-01 2.20722556e-01 -2.59702086e-01
-1.24973130e+00 9.61100876e-01 3.86811882e-01 2.09416136e-01
8.43802869e-01 5.44574261e-01 7.32530296e-01 4.60889995e-01
5.19128621e-01 -1.27412355e+00 -9.76420417e-02 2.44311094e-01
4.03425395e-01 -1.39769995e+00 -2.39038114e-02 -9.70611095e-01
-4.19088870e-01 9.45935428e-01 5.48569202e-01 -1.99150518e-01
1.10142267e+00 5.11585847e-02 -1.34993121e-01 -4.18392599e-01
-2.47255906e-01 -5.55278778e-01 2.41552651e-01 6.52260840e-01
-1.78266495e-01 1.05043845e-02 1.74237508e-02 3.75092238e-01
-8.46468732e-02 -2.29718342e-01 5.74045302e-03 1.17463589e+00
-4.72297192e-01 -1.01847327e+00 -2.28667527e-01 4.52456713e-01
3.73944230e-02 3.08235615e-01 -2.80183375e-01 5.90888739e-01
3.52684200e-01 9.38512444e-01 1.01859547e-01 -3.13380569e-01
1.72727734e-01 2.95330156e-02 5.86569309e-01 -8.74848366e-01
-8.38142261e-02 1.77766383e-01 -3.10002506e-01 -6.04130685e-01
-9.79753733e-01 -8.20711553e-01 -1.55222189e+00 2.30655164e-01
-5.93355298e-01 1.26299322e-01 8.77194643e-01 1.24760306e+00
2.76589781e-01 2.20090359e-01 1.24299383e+00 -1.40514433e+00
-3.96969110e-01 -5.29458880e-01 -7.73160696e-01 2.88290590e-01
1.57713313e-02 -1.03738773e+00 -5.02755284e-01 1.40590250e-01] | [8.03272533416748, -3.1550307273864746] |
58c732ad-679f-49eb-8765-4478599e01eb | sc-ml-self-supervised-counterfactual-metric | 2304.01647 | null | https://arxiv.org/abs/2304.01647v1 | https://arxiv.org/pdf/2304.01647v1.pdf | SC-ML: Self-supervised Counterfactual Metric Learning for Debiased Visual Question Answering | Visual question answering (VQA) is a critical multimodal task in which an agent must answer questions according to the visual cue. Unfortunately, language bias is a common problem in VQA, which refers to the model generating answers only by associating with the questions while ignoring the visual content, resulting in biased results. We tackle the language bias problem by proposing a self-supervised counterfactual metric learning (SC-ML) method to focus the image features better. SC-ML can adaptively select the question-relevant visual features to answer the question, reducing the negative influence of question-irrelevant visual features on inferring answers. In addition, question-irrelevant visual features can be seamlessly incorporated into counterfactual training schemes to further boost robustness. Extensive experiments have proved the effectiveness of our method with improved results on the VQA-CP dataset. Our code will be made publicly available. | ['Zhenyu Lu', 'Zhongfeng Chen', 'Ziheng Wu', 'Xu Yang', 'ShiYang Yan', 'Xinyao Shu'] | 2023-04-04 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 1.72149986e-01 1.01136394e-01 -8.06989148e-02 -2.71887362e-01
-7.94516504e-01 -7.30764627e-01 7.95976341e-01 -1.96537152e-01
-4.81091350e-01 6.07333064e-01 5.20590901e-01 -5.26879549e-01
8.18152130e-02 -7.41223276e-01 -6.88389063e-01 -7.21883297e-01
2.68478602e-01 1.34819135e-01 1.86517656e-01 -2.09372789e-01
4.87115711e-01 2.16561958e-01 -1.52299321e+00 6.83248103e-01
9.42949891e-01 7.42313802e-01 2.93674707e-01 5.50264776e-01
-4.45582837e-01 1.22490668e+00 -7.82811403e-01 -7.44475365e-01
1.67303890e-01 -8.25454295e-01 -7.48211861e-01 1.45281926e-01
6.09583199e-01 -2.85308331e-01 -3.61729503e-01 1.09109437e+00
5.27714372e-01 4.48345877e-02 5.96355021e-01 -1.55469227e+00
-9.33647573e-01 8.87241587e-02 -5.91943324e-01 3.44937384e-01
8.05681050e-01 7.42388904e-01 1.10613775e+00 -1.03888738e+00
6.75723433e-01 1.80720115e+00 -2.45831837e-03 8.99484754e-01
-1.30262399e+00 -6.83534086e-01 3.68835956e-01 6.92138374e-01
-1.09423482e+00 -4.51822311e-01 1.32811058e+00 -2.65623212e-01
4.15223628e-01 6.17120922e-01 5.06038547e-01 1.27726138e+00
1.22546203e-01 1.12138689e+00 1.39365518e+00 -3.22917134e-01
3.94320458e-01 2.81776071e-01 -3.29229206e-01 8.63619030e-01
3.69391181e-02 2.83414960e-01 -6.20790064e-01 -1.68871030e-01
4.08672392e-01 -1.52017713e-01 -4.72463399e-01 -7.56755233e-01
-1.28967047e+00 1.16132915e+00 7.10208237e-01 4.70706597e-02
-3.62624764e-01 -5.13109639e-02 1.87520221e-01 3.40722710e-01
1.69556066e-01 5.04136741e-01 -1.59309775e-01 3.28740627e-01
-4.47265714e-01 4.66179013e-01 3.46721143e-01 6.87759340e-01
8.26034129e-01 2.34669261e-02 -9.18973744e-01 6.25610828e-01
3.79353970e-01 7.11533189e-01 5.37939548e-01 -1.27494216e+00
7.07924426e-01 8.82920444e-01 2.73501128e-01 -1.11966145e+00
-1.79820254e-01 -5.24517894e-02 -5.03327429e-01 4.30870473e-01
7.30436683e-01 -1.18913054e-01 -7.66322434e-01 1.94689906e+00
5.69139838e-01 -2.79800296e-01 1.70780972e-01 1.31650400e+00
1.05773818e+00 5.14502823e-01 3.60336423e-01 -2.56067604e-01
1.53557277e+00 -8.52135420e-01 -9.17402804e-01 -4.96213138e-01
5.03627479e-01 -6.45939946e-01 1.76515770e+00 -1.43568933e-01
-7.38633156e-01 -4.60696042e-01 -8.96670222e-01 -1.93288475e-01
-2.80854166e-01 -1.66771542e-02 5.25220931e-01 6.71732485e-01
-7.96080410e-01 -2.32230037e-01 4.72400263e-02 -1.19003512e-01
6.51215613e-01 -3.52566466e-02 -2.07875520e-01 -2.98656106e-01
-1.30882204e+00 8.59962881e-01 2.01115236e-01 1.60439089e-02
-1.11100626e+00 -4.26302642e-01 -9.55787897e-01 -8.26762766e-02
7.20143616e-01 -9.27704871e-01 1.18489647e+00 -1.40468884e+00
-1.41233659e+00 8.56323421e-01 -5.22315562e-01 -2.08998829e-01
6.96834505e-01 2.64993042e-01 -2.61150181e-01 5.50042450e-01
2.53362060e-01 1.03875530e+00 1.17494631e+00 -1.83476198e+00
-4.17680830e-01 -5.49266100e-01 4.85403597e-01 4.45132941e-01
7.70177022e-02 -3.96162480e-01 -3.50429595e-01 -6.31642580e-01
-6.52739629e-02 -7.34147072e-01 -1.09686300e-01 1.52443856e-01
-2.44045556e-01 -5.14304399e-01 9.24453139e-01 -5.50870657e-01
7.99521744e-01 -2.04273677e+00 6.04056679e-02 -1.71670884e-01
1.28219113e-01 8.94572958e-02 -5.25410414e-01 1.03954293e-01
3.33742686e-02 -1.11880675e-01 -2.99981326e-01 1.53418347e-01
-1.03661595e-02 1.32373318e-01 -5.90481162e-01 4.53856349e-01
3.90737057e-01 1.37810254e+00 -9.70679164e-01 -9.90667462e-01
1.51351482e-01 1.52575448e-01 -4.29830939e-01 5.46389759e-01
-5.26079774e-01 4.30161387e-01 -3.86349618e-01 6.97882354e-01
8.63012552e-01 -1.77876756e-01 1.02751955e-01 -2.84370273e-01
2.05668911e-01 5.11373207e-02 -6.22291803e-01 1.42380488e+00
-4.44320261e-01 5.54434955e-01 1.38311908e-01 -8.71155143e-01
7.40987837e-01 1.18496865e-01 -8.91288146e-02 -1.25965619e+00
1.09109007e-01 -3.13067615e-01 7.08442181e-02 -1.03493941e+00
2.03382269e-01 -2.48540282e-01 -2.90018949e-03 3.48900974e-01
-1.86762020e-01 -1.98942989e-01 1.20498590e-01 4.39532012e-01
5.25864542e-01 2.14480758e-01 2.78435498e-01 3.99299972e-02
7.03003466e-01 2.19190180e-01 3.94804269e-01 7.63007104e-01
-6.78021014e-01 4.50367898e-01 6.30931258e-01 -1.42559513e-01
-7.04266667e-01 -1.21693575e+00 3.62277478e-01 1.24673426e+00
4.13302720e-01 1.26429975e-01 -7.36788809e-01 -1.21827996e+00
1.34763801e-02 1.20039380e+00 -9.62015629e-01 -3.19346845e-01
-4.26284522e-01 -4.47246343e-01 1.71218306e-01 1.98144689e-01
4.93843973e-01 -1.42513430e+00 -7.79277921e-01 -3.01307589e-01
-5.88496208e-01 -7.90253818e-01 -6.05459154e-01 -3.29239011e-01
-6.33916616e-01 -1.30432522e+00 -9.36815143e-01 -6.53416157e-01
7.04039156e-01 7.08161414e-01 1.12245417e+00 1.04962192e-01
8.80564004e-02 6.74569368e-01 -2.37694904e-01 -3.59081596e-01
-3.41056436e-01 -4.94241416e-01 -5.08615673e-01 2.55385190e-01
4.58049357e-01 4.43102717e-02 -9.10406172e-01 1.62389353e-01
-9.67397153e-01 1.74690932e-01 6.14220858e-01 8.33596528e-01
2.79454947e-01 -6.22925341e-01 8.45071852e-01 -8.66767466e-01
9.21195328e-01 -4.60602224e-01 -5.12349248e-01 4.55716968e-01
-3.03023428e-01 1.75716072e-01 5.82867742e-01 -5.31041443e-01
-1.45458794e+00 3.61737609e-02 1.54371008e-01 -3.43123972e-01
-1.13056205e-01 2.82983948e-02 -5.68935156e-01 7.02826157e-02
5.50128341e-01 4.33629096e-01 1.21657457e-02 2.81304633e-03
8.04624856e-01 3.81767303e-01 3.78470957e-01 -3.35447043e-01
7.83533275e-01 8.19568515e-01 -4.71059941e-02 -5.71433187e-01
-9.76399243e-01 -2.92689651e-01 -2.01554403e-01 -4.80819106e-01
1.06929064e+00 -7.47589946e-01 -9.72914338e-01 -1.39188975e-01
-1.31166899e+00 -2.36548726e-02 -7.60696679e-02 1.64238244e-01
-5.29040456e-01 6.22362494e-01 1.89208817e-02 -1.09905803e+00
-7.95640722e-02 -1.03409135e+00 8.45203638e-01 3.73462051e-01
-3.95196006e-02 -7.59748816e-01 1.00690894e-01 9.08335686e-01
1.37424871e-01 -1.63574647e-02 1.12002397e+00 -4.86748904e-01
-7.27468014e-01 2.65602767e-01 -6.55787468e-01 -6.13644533e-02
-1.08740076e-01 -4.65370923e-01 -1.21659327e+00 -2.09529772e-01
1.75452918e-01 -5.25235951e-01 1.08903027e+00 1.20133050e-01
1.19012654e+00 -5.22189319e-01 -1.79062977e-01 8.89504179e-02
1.25997758e+00 2.67281175e-01 4.46713507e-01 1.63415611e-01
6.68539166e-01 1.02908123e+00 8.62792790e-01 1.24185324e-01
7.09483624e-01 5.51116109e-01 8.58281374e-01 -1.68442219e-01
-3.41600269e-01 -4.73258466e-01 4.07627940e-01 1.14354643e-03
3.55043173e-01 -2.71508306e-01 -6.13649130e-01 7.25438893e-01
-1.80593789e+00 -1.15748107e+00 -2.93671519e-01 2.01001906e+00
7.76315451e-01 -2.34219521e-01 1.22633487e-01 -4.16574739e-02
6.64029062e-01 3.66369009e-01 -7.16583610e-01 -1.77745566e-01
-3.64553601e-01 -2.33290896e-01 8.30540061e-02 4.59372044e-01
-9.77728784e-01 8.50143492e-01 6.01873112e+00 6.65482938e-01
-9.07956243e-01 2.61600137e-01 6.13175273e-01 -1.90362558e-01
-8.50360930e-01 4.33437899e-02 -1.89965546e-01 5.02722681e-01
3.59405011e-01 7.84495659e-03 3.49326283e-01 6.77291632e-01
4.37360734e-01 -4.55001026e-01 -9.99494970e-01 1.00955594e+00
4.40773815e-01 -1.27571404e+00 5.28555214e-01 -1.80244625e-01
5.28053820e-01 -4.92425561e-01 2.81399667e-01 5.02518415e-01
2.93471247e-01 -9.40053999e-01 8.38135004e-01 5.57560742e-01
5.29050708e-01 -8.09499562e-01 5.78577518e-01 2.70973951e-01
-8.12555492e-01 -3.08285654e-01 -4.12040591e-01 -4.03327076e-03
-8.46194029e-02 1.52757153e-01 -8.46352518e-01 4.31136131e-01
6.83564425e-01 3.23568918e-02 -9.87045825e-01 8.06022227e-01
-5.67823470e-01 5.51553667e-01 4.34942096e-01 -3.23908210e-01
9.44532827e-02 -8.64607841e-02 6.29500151e-01 9.07356501e-01
-1.29868776e-01 1.42704332e-02 -1.26457587e-01 1.11142445e+00
-2.89035160e-02 2.59353280e-01 -7.81223595e-01 2.11088434e-01
2.57155567e-01 1.03215420e+00 -4.18460876e-01 -3.45217556e-01
-5.87396443e-01 1.07176077e+00 5.23925543e-01 8.08804810e-01
-7.82816827e-01 -3.29849511e-01 2.52093792e-01 5.82714286e-03
3.43783885e-01 1.62710011e-01 -2.14417368e-01 -1.25638640e+00
-1.24589335e-02 -1.09976852e+00 6.98146343e-01 -1.13046575e+00
-1.37697816e+00 3.24674845e-01 -5.85163906e-02 -1.16634154e+00
-1.39975160e-01 -6.15464151e-01 -5.29538870e-01 7.02095687e-01
-1.69363642e+00 -1.22588992e+00 -2.54130155e-01 7.36280739e-01
5.95326424e-01 -3.48887630e-02 4.77301657e-01 -1.53033137e-01
-3.67395639e-01 3.93912792e-01 -3.42608660e-01 2.39202101e-02
7.86646247e-01 -1.24145103e+00 -2.78027475e-01 7.54039109e-01
2.07901135e-01 3.89319599e-01 8.45041454e-01 -3.96831095e-01
-1.41500092e+00 -9.37463462e-01 8.74261439e-01 -6.16342247e-01
3.71188223e-01 -3.23821723e-01 -9.64911163e-01 3.21951658e-01
7.06770837e-01 -2.69082516e-01 6.98640049e-01 -2.21972167e-01
-6.75322175e-01 -1.55917376e-01 -1.26135683e+00 1.00261211e+00
7.31320322e-01 -7.25646436e-01 -1.06648314e+00 1.92875147e-01
7.63586342e-01 9.23115760e-02 -5.37668495e-03 1.22518077e-01
2.95956105e-01 -1.15482008e+00 9.96856153e-01 -6.55715108e-01
4.47364688e-01 -6.20127559e-01 -8.94232765e-02 -1.27229261e+00
-2.20289499e-01 -2.10651115e-01 -1.01312906e-01 1.11604333e+00
2.63822079e-01 -4.18772995e-01 5.94090939e-01 2.81616211e-01
4.65547979e-01 -3.60528648e-01 -1.22109032e+00 -3.42609614e-01
-1.45325065e-02 -1.88099533e-01 5.83614528e-01 9.65398252e-01
-1.49145707e-01 6.43591583e-01 -3.90782386e-01 1.03114314e-01
7.10027635e-01 6.08548105e-01 9.13638592e-01 -7.74246454e-01
-5.02942577e-02 -4.71801400e-01 1.03374608e-01 -1.02123022e+00
3.72054964e-01 -8.45968962e-01 6.52347207e-02 -1.51224911e+00
5.05676568e-01 2.79818594e-01 -2.00479999e-01 2.57858276e-01
-6.91950023e-01 2.28424594e-01 4.10629630e-01 -1.19603882e-02
-9.10550833e-01 9.24997032e-01 1.89486802e+00 -5.26973546e-01
1.47200841e-02 -2.33637303e-01 -8.49198878e-01 5.83309472e-01
7.95505822e-01 -3.49685699e-01 -5.88677883e-01 -3.50808084e-01
1.05831221e-01 1.77098840e-01 8.08282912e-01 -4.35813427e-01
-8.30383971e-02 -5.26784778e-01 6.94757223e-01 -6.32235110e-01
4.02728140e-01 -7.84596741e-01 -5.01906335e-01 6.14249885e-01
-4.93455738e-01 -1.42003987e-02 1.04856074e-01 8.39287639e-01
-1.59592435e-01 -2.47255623e-01 7.29192555e-01 -3.66852164e-01
-7.38524556e-01 -6.90065771e-02 -4.32828665e-01 3.22243661e-01
1.12305486e+00 -6.74544424e-02 -5.11016250e-01 -7.79455602e-01
-2.06215233e-01 4.58116442e-01 3.46462131e-01 5.26243448e-01
9.92438555e-01 -1.50249946e+00 -7.95714617e-01 -5.21406671e-03
6.04358017e-01 -4.05860871e-01 5.48186898e-01 5.54683387e-01
-1.27529204e-01 6.27173603e-01 -6.91845044e-02 -5.45948327e-01
-1.24077511e+00 1.11033762e+00 3.38863790e-01 1.05933957e-02
-1.01072364e-01 7.96737731e-01 8.67594719e-01 -3.87959272e-01
1.19687021e-01 8.20876732e-02 -6.57597840e-01 1.15061894e-01
5.40701807e-01 2.81399757e-01 -4.04380769e-01 -7.54675210e-01
-5.99842846e-01 3.22460473e-01 6.53548613e-02 -3.81731570e-01
7.42309630e-01 -3.90976071e-01 7.83819556e-02 3.16404909e-01
9.70543981e-01 2.14476347e-01 -1.38882017e+00 -1.85240582e-01
-1.45167345e-02 -7.23587275e-01 -1.19067706e-01 -9.78039742e-01
-8.26797128e-01 9.49117899e-01 7.72114038e-01 1.64854228e-01
1.11248553e+00 3.50786597e-01 2.47233078e-01 2.82775074e-01
8.48191231e-03 -9.29629922e-01 6.62886858e-01 7.16856942e-02
1.37520957e+00 -1.58640456e+00 -1.36401936e-01 -2.08048627e-01
-1.08251786e+00 6.71943903e-01 9.51605320e-01 8.50571766e-02
3.41724604e-02 -3.85703146e-01 6.25017583e-01 -3.33900511e-01
-8.71609807e-01 -3.69470686e-01 4.84172314e-01 8.15739155e-01
3.68042976e-01 4.24308106e-02 -4.05919880e-01 4.63849127e-01
6.49520978e-02 -3.72808337e-01 3.58760834e-01 6.81714177e-01
-2.09131613e-01 -8.23094964e-01 -6.70720160e-01 1.27966329e-01
-1.74128771e-01 -1.63894910e-02 -8.16376150e-01 7.10853815e-01
1.30821899e-01 1.39092803e+00 1.73714466e-03 -1.29458889e-01
3.59181136e-01 1.71109602e-01 5.90010047e-01 -2.13396445e-01
-3.59176427e-01 -1.24521188e-01 -1.14130527e-01 -6.31066680e-01
-5.39108336e-01 -5.38330853e-01 -1.15097141e+00 1.59403253e-02
-9.06875506e-02 1.76877573e-01 2.79764205e-01 8.26080501e-01
3.82072061e-01 3.97647649e-01 6.61060691e-01 -3.52156609e-01
-5.83960474e-01 -8.77493620e-01 -4.02822718e-02 7.94016540e-01
6.26974404e-01 -7.20760703e-01 -5.58019400e-01 -9.08259153e-02] | [10.787108421325684, 1.7193355560302734] |
8d881117-84a2-431a-aec2-7f8195e00230 | improving-retrieval-augmented-large-language | 2307.03027 | null | https://arxiv.org/abs/2307.03027v1 | https://arxiv.org/pdf/2307.03027v1.pdf | Improving Retrieval-Augmented Large Language Models via Data Importance Learning | Retrieval augmentation enables large language models to take advantage of external knowledge, for example on tasks like question answering and data imputation. However, the performance of such retrieval-augmented models is limited by the data quality of their underlying retrieval corpus. In this paper, we propose an algorithm based on multilinear extension for evaluating the data importance of retrieved data points. There are exponentially many terms in the multilinear extension, and one key contribution of this paper is a polynomial time algorithm that computes exactly, given a retrieval-augmented model with an additive utility function and a validation set, the data importance of data points in the retrieval corpus using the multilinear extension of the model's utility function. We further proposed an even more efficient ({\epsilon}, {\delta})-approximation algorithm. Our experimental results illustrate that we can enhance the performance of large language models by only pruning or reweighting the retrieval corpus, without requiring further training. For some tasks, this even allows a small model (e.g., GPT-JT), augmented with a search engine API, to outperform GPT-3.5 (without retrieval augmentation). Moreover, we show that weights based on multilinear extension can be computed efficiently in practice (e.g., in less than ten minutes for a corpus with 100 million elements). | ['Ce Zhang', 'Sebastian Schelter', 'Meng Cao', 'Shaopeng Wei', 'Samantha Biegel', 'Stefan Grafberger', 'Xiaozhong Lyu'] | 2023-07-06 | null | null | null | null | ['imputation', 'retrieval', 'imputation', 'question-answering', 'imputation'] | ['computer-vision', 'methodology', 'miscellaneous', 'natural-language-processing', 'time-series'] | [-1.01825267e-01 1.04631081e-01 -1.93504453e-01 -1.58470735e-01
-1.45842719e+00 -5.75922072e-01 1.82420030e-01 6.05507255e-01
-9.18718636e-01 6.99197054e-01 -4.65271845e-02 -5.13292372e-01
-4.57025766e-01 -8.17095399e-01 -9.35378671e-01 -4.08845693e-01
-1.34340063e-01 8.67010236e-01 1.11843251e-01 -2.22737998e-01
3.38125348e-01 1.06100902e-01 -1.49968112e+00 1.46141082e-01
9.88433480e-01 1.13579214e+00 3.05202305e-01 6.87962174e-01
-5.24679065e-01 5.98138094e-01 -6.27050579e-01 -5.88845074e-01
2.83944547e-01 2.91056901e-01 -7.02455640e-01 -3.66421700e-01
2.44122446e-01 -2.30731443e-01 -1.71426982e-01 7.17853785e-01
3.36932540e-01 2.92253017e-01 6.08842373e-01 -1.25588572e+00
-5.51587403e-01 7.56427526e-01 -3.80135357e-01 -1.87466905e-01
3.65268528e-01 -2.74265140e-01 1.18434191e+00 -1.14472997e+00
4.49772149e-01 1.22314739e+00 6.29601419e-01 3.48210074e-02
-1.36624634e+00 -6.68643177e-01 2.12401927e-01 -1.28866360e-01
-1.57416809e+00 -2.58835912e-01 4.35423523e-01 -1.86010689e-01
1.20274687e+00 5.87290525e-01 2.91169852e-01 2.50789046e-01
-9.53431353e-02 8.96152854e-01 7.07360983e-01 -6.15649283e-01
1.44225210e-01 2.57017791e-01 5.12869060e-01 6.65831864e-01
3.09410810e-01 -1.53415382e-01 -4.87844557e-01 -8.33448887e-01
2.48018891e-01 -7.08825141e-02 -1.05759621e-01 -3.40640306e-01
-9.75537062e-01 8.70695889e-01 1.48481950e-01 1.38011634e-01
-2.51197278e-01 4.31765050e-01 2.51618207e-01 6.09322190e-01
8.35068464e-01 5.74536562e-01 -8.37869346e-01 -2.16038033e-01
-9.00996447e-01 2.75344163e-01 8.35983455e-01 1.04658675e+00
1.02182364e+00 -3.99784505e-01 -1.76144853e-01 1.11368287e+00
2.74850607e-01 7.48607934e-01 4.71878588e-01 -1.00671601e+00
7.19654799e-01 7.06004977e-01 1.83345750e-01 -7.62571692e-01
-3.33139002e-01 -1.75792143e-01 -6.26684844e-01 -3.53662074e-01
4.72302884e-01 -1.86468795e-01 -5.15193462e-01 1.92887306e+00
3.89142781e-01 -3.12399138e-02 1.52224660e-01 5.15792787e-01
3.37941676e-01 6.85684860e-01 8.93392712e-02 -3.29549074e-01
1.33614421e+00 -5.83679974e-01 -5.70735812e-01 -9.69454646e-02
1.27069950e+00 -9.87677217e-01 1.26373446e+00 3.84096205e-01
-1.20434570e+00 -2.66839713e-01 -7.13805318e-01 -4.46320891e-01
-3.91413450e-01 3.86879630e-02 8.85591149e-01 5.35720408e-01
-8.78414035e-01 3.54681700e-01 -5.14292121e-01 5.25325127e-02
-5.59936129e-02 4.93996263e-01 -3.28009337e-01 -1.88240811e-01
-1.47142136e+00 8.77448201e-01 4.50779229e-01 -9.98303890e-02
-7.89866000e-02 -9.17929530e-01 -8.55516315e-01 1.89259022e-01
6.80311263e-01 -6.87338591e-01 1.27215827e+00 -3.99440110e-01
-9.75616097e-01 4.16813761e-01 -4.22412813e-01 -4.79374170e-01
2.94331700e-01 -3.13153446e-01 -2.26241648e-02 -5.05941827e-03
3.47256660e-02 4.44808155e-01 5.83548725e-01 -1.08551335e+00
-5.34561753e-01 -4.99580860e-01 2.75754631e-01 2.00014517e-01
-6.49057746e-01 1.54656127e-01 -9.97873366e-01 -6.79717064e-01
1.90899298e-01 -9.85433638e-01 -4.41047579e-01 -2.21523736e-02
-8.49040151e-02 -5.60840130e-01 1.60109326e-01 -6.94797039e-01
1.58656907e+00 -1.96206176e+00 -2.85598952e-02 5.90683579e-01
1.80221125e-01 7.88334757e-02 -4.26862538e-01 4.47694153e-01
4.04823005e-01 4.29064751e-01 -2.98698723e-01 -4.96789455e-01
2.91719228e-01 3.65018398e-01 -3.30251247e-01 1.77375406e-01
-4.40491326e-02 8.88784409e-01 -8.14112842e-01 -6.22002721e-01
-1.60812348e-01 3.90541971e-01 -7.56507337e-01 -4.56679761e-02
-4.27410305e-01 -4.26560879e-01 -4.97869283e-01 4.28335667e-01
5.67888677e-01 -3.30097675e-01 1.63462654e-01 5.12018949e-02
1.38382450e-01 3.40349585e-01 -1.50444460e+00 1.55669665e+00
-1.00265622e+00 3.74949008e-01 -1.26384813e-02 -6.05427146e-01
7.19267190e-01 1.94945991e-01 4.85159785e-01 -6.89274609e-01
-3.50973368e-01 4.46375489e-01 -3.00759584e-01 -3.39989036e-01
9.69669461e-01 -2.38290969e-02 -3.15638930e-01 8.98860753e-01
-1.59674197e-01 7.62725919e-02 4.56208795e-01 4.50738788e-01
9.89974558e-01 -2.48305544e-01 4.18496989e-02 -4.10131365e-02
3.56393635e-01 -1.25839278e-01 3.97861987e-01 1.03994107e+00
5.98544419e-01 3.90771389e-01 5.74856937e-01 -1.58873916e-01
-8.90267670e-01 -6.39251947e-01 -2.74050444e-01 1.39044940e+00
-2.27743655e-01 -8.67365181e-01 -5.18281341e-01 -4.87104744e-01
3.13292652e-01 7.40018010e-01 -2.87483484e-01 -1.80822179e-01
-5.39195061e-01 -9.08639908e-01 5.71687639e-01 5.00321031e-01
3.58251520e-02 -8.31795573e-01 -2.19883472e-01 3.65962833e-01
-3.93279791e-01 -8.99747193e-01 -6.79617345e-01 3.22210610e-01
-1.00866604e+00 -7.97288954e-01 -5.05233347e-01 -5.07857323e-01
6.98254883e-01 2.87440956e-01 1.24594057e+00 4.37950432e-01
1.58460885e-01 4.96431589e-01 -2.79370040e-01 -3.75229567e-01
-1.80234045e-01 2.51005471e-01 5.25600985e-02 -4.40600127e-01
4.80386823e-01 -3.40465248e-01 -2.60321379e-01 1.94903046e-01
-1.16424453e+00 -3.01239580e-01 6.30235374e-01 9.92726207e-01
8.10313463e-01 -7.24929869e-02 6.92658007e-01 -9.97707903e-01
8.78589153e-01 -6.26482666e-01 -9.28868890e-01 6.72543824e-01
-9.16272521e-01 6.78900003e-01 3.96758854e-01 -6.38301015e-01
-6.94803238e-01 -2.41326049e-01 -7.91126266e-02 -2.55336076e-01
4.89971250e-01 1.22306812e+00 -6.52483571e-03 9.07217041e-02
5.24658740e-01 1.41829073e-01 -5.13350740e-02 -6.78025246e-01
5.30743122e-01 7.47776866e-01 1.71451584e-01 -1.02831745e+00
7.18401670e-01 1.58747181e-03 -3.49992141e-02 -6.46919191e-01
-7.91227460e-01 -8.41742277e-01 -2.08552852e-01 3.23705018e-01
4.76989821e-02 -9.32183087e-01 -9.27586794e-01 -1.25712398e-02
-1.09700620e+00 -1.93498567e-01 -3.04822922e-01 5.21193743e-01
-3.35111827e-01 5.26619673e-01 -4.47051376e-01 -1.24844885e+00
-6.67970896e-01 -8.43513727e-01 1.22234809e+00 -3.62641633e-01
-1.02176763e-01 -9.01488423e-01 1.87569298e-02 5.70943356e-01
3.07143211e-01 -3.61550689e-01 1.21523535e+00 -8.56499135e-01
-7.04899728e-01 -5.47021270e-01 -1.65187195e-01 3.97491664e-01
-3.14219803e-01 -2.53707200e-01 -7.50315130e-01 -1.98520735e-01
-1.47324145e-01 -4.03766870e-01 8.77634883e-01 1.18984409e-01
1.29673028e+00 -5.73520601e-01 -9.91083235e-02 3.15656334e-01
1.27677584e+00 -1.51338950e-01 4.30478424e-01 3.36773694e-01
4.12454247e-01 6.29885972e-01 9.85539615e-01 4.18689340e-01
4.31624681e-01 7.50385940e-01 2.00137552e-02 1.40333623e-01
2.53300697e-01 -2.83402890e-01 3.29873502e-01 1.02888405e+00
1.31897733e-01 -1.98937625e-01 -8.17355633e-01 6.09152257e-01
-1.93020594e+00 -6.57960355e-01 -1.34722829e-01 2.72904038e+00
1.15563190e+00 2.03487381e-01 -2.19529390e-01 1.15650639e-01
2.25335807e-01 -3.57807040e-01 -5.97881794e-01 -3.25689405e-01
-1.91600561e-01 5.05958438e-01 6.29689515e-01 8.85918021e-01
-6.51532650e-01 7.62884736e-01 6.32530165e+00 1.17951763e+00
-4.02208120e-01 4.62361090e-02 3.69051993e-01 -2.63778478e-01
-7.46946454e-01 1.35561779e-01 -7.98674405e-01 4.12132680e-01
1.13996327e+00 -5.46608329e-01 6.05509818e-01 8.46229494e-01
-1.09473594e-01 -2.16905192e-01 -9.58041728e-01 6.91147447e-01
8.38102475e-02 -1.04666579e+00 1.27698839e-01 2.85686702e-01
4.45036143e-01 3.93443787e-03 7.65386820e-02 6.85976624e-01
2.81381667e-01 -7.15847969e-01 4.78049070e-01 5.22215664e-01
6.62763178e-01 -9.56733823e-01 8.48829985e-01 7.65750885e-01
-1.08366430e+00 -9.89010371e-03 -7.02118456e-01 1.29755169e-01
3.13252099e-02 8.58415902e-01 -7.83074498e-01 6.39354527e-01
5.11265457e-01 -1.31757140e-01 -6.02799892e-01 8.25906217e-01
1.48245141e-01 5.12277186e-01 -8.54657471e-01 -2.20069572e-01
5.73990727e-03 -2.57859111e-01 3.13884169e-01 1.00752664e+00
5.93756974e-01 1.57583371e-01 9.54252109e-02 3.75601619e-01
-4.32703406e-01 4.74324197e-01 -4.91115659e-01 2.48734038e-02
6.24270022e-01 1.17508936e+00 -5.03305607e-02 -4.69843745e-01
-4.49294269e-01 3.78734857e-01 5.59892833e-01 3.65357518e-01
-6.31537914e-01 -3.65589052e-01 4.66774255e-01 1.10724844e-01
8.25656876e-02 -1.67617172e-01 -1.30096093e-01 -1.12842464e+00
4.52255428e-01 -9.55997586e-01 5.95546305e-01 -8.54518950e-01
-1.33131409e+00 4.73646224e-01 1.55980721e-01 -8.68078709e-01
-5.72065949e-01 -4.57251996e-01 2.30248734e-01 1.18093097e+00
-1.48704493e+00 -8.21337402e-01 2.73112744e-01 4.88991559e-01
1.72159910e-01 1.92494076e-02 9.43090260e-01 3.91080976e-01
-2.18165144e-01 9.39497590e-01 4.42850262e-01 -1.67652145e-01
7.77036428e-01 -1.37408912e+00 1.01380743e-01 4.77839112e-01
3.25044572e-01 1.13345313e+00 5.23037970e-01 -4.98761684e-01
-1.58540499e+00 -8.83507013e-01 1.50093961e+00 -6.17614567e-01
8.06450963e-01 -3.39225501e-01 -1.11748183e+00 6.24418378e-01
-1.76383838e-01 -1.70235604e-01 9.56790030e-01 6.54860258e-01
-4.39238042e-01 -2.16339722e-01 -1.12175322e+00 6.60572052e-01
8.26693177e-01 -5.72215736e-01 -5.55343091e-01 5.16449273e-01
9.08411205e-01 -3.87367398e-01 -1.13372731e+00 3.54027957e-01
7.09418237e-01 -2.15300873e-01 1.14070117e+00 -7.48364866e-01
2.69626174e-02 -1.77626833e-01 -3.92200679e-01 -9.19851840e-01
6.28805384e-02 -5.95549226e-01 -5.69595277e-01 9.79890466e-01
7.84227610e-01 -7.73856997e-01 5.64238429e-01 1.20914543e+00
2.14427456e-01 -8.22251439e-01 -1.00119209e+00 -7.90312707e-01
2.65123934e-01 -9.20908928e-01 6.43376529e-01 4.88487780e-01
-8.80828034e-03 2.66467541e-01 -3.14021051e-01 1.13578826e-01
5.26294768e-01 2.35647097e-01 8.18785548e-01 -1.45001733e+00
-4.34106588e-01 -6.03991561e-02 1.84089184e-01 -1.19249570e+00
2.78345287e-01 -1.04011297e+00 -2.15274423e-01 -1.38147402e+00
3.31338286e-01 -1.02350843e+00 -5.29854238e-01 6.23116314e-01
-4.64001507e-01 5.19933216e-02 2.41953805e-01 9.45499092e-02
-8.40942323e-01 3.89026612e-01 9.17664587e-01 -1.41843647e-01
-2.32030585e-01 -1.34938583e-03 -9.18634593e-01 5.84676921e-01
7.26397991e-01 -8.18158805e-01 -3.90286952e-01 -6.19674146e-01
8.86958122e-01 2.84727752e-01 1.06059641e-01 -4.76264983e-01
4.80427474e-01 3.41487974e-02 3.72855999e-02 -7.78056920e-01
4.31594312e-01 -9.07167435e-01 1.36573017e-01 2.32383162e-01
-7.02482283e-01 3.14154476e-01 2.95115352e-01 5.88094115e-01
-2.32447863e-01 -7.35618651e-01 2.30143368e-02 1.04106441e-01
-1.57025978e-01 2.03891873e-01 -2.27881268e-01 1.03739679e-01
3.60122502e-01 1.87078744e-01 -2.57780254e-01 -3.20752949e-01
-4.25499529e-01 6.03386581e-01 1.76970139e-01 1.81921706e-01
5.18277884e-01 -1.37118721e+00 -6.44866943e-01 3.19183879e-02
2.67142594e-01 1.40799209e-01 7.31194690e-02 7.05850542e-01
-4.67930734e-02 6.88536048e-01 6.10220432e-01 -3.16058904e-01
-1.25170803e+00 6.49615467e-01 -1.28504485e-01 -9.14030075e-01
-8.58797207e-02 5.12724161e-01 -1.61873311e-01 -6.76937103e-01
2.75811106e-01 -4.94337738e-01 1.14210799e-01 2.11775333e-01
5.07277131e-01 3.46599400e-01 3.32844973e-01 -1.94147736e-01
-1.60056904e-01 4.20131356e-01 -2.47295678e-01 -4.36605394e-01
1.17948961e+00 6.56740740e-02 -3.72816354e-01 3.85594100e-01
1.29958022e+00 2.94933587e-01 -6.35353267e-01 -6.22847617e-01
1.66213900e-01 -4.21189547e-01 1.31809413e-01 -7.86857903e-01
-7.90122628e-01 5.79567671e-01 2.68445104e-01 2.48162806e-01
1.21052241e+00 -1.18932888e-01 7.98289657e-01 1.09606767e+00
6.12902880e-01 -1.12195623e+00 -1.97276682e-01 5.34657061e-01
8.64021659e-01 -1.16419208e+00 -3.05472333e-02 -1.19662069e-01
-3.32481831e-01 8.14191401e-01 2.39677340e-01 2.88215369e-01
6.74322188e-01 2.78813928e-01 -7.49402568e-02 8.59642476e-02
-1.12761998e+00 -2.73703933e-01 3.80026460e-01 2.34760642e-01
4.74550694e-01 -5.01106419e-02 -6.16075933e-01 7.19509423e-01
-2.40304023e-01 -5.04666008e-02 7.95721039e-02 8.93446684e-01
-4.24680203e-01 -1.48781633e+00 -4.11741525e-01 7.86340892e-01
-5.52307904e-01 -6.07080102e-01 -4.92777377e-02 7.79463947e-01
-2.06658766e-01 1.10418558e+00 -9.82809067e-02 -1.15764678e-01
4.29196954e-01 3.03481877e-01 2.59112448e-01 -4.76369590e-01
-6.77598715e-01 1.68355197e-01 2.25846350e-01 -3.26327771e-01
-1.50078639e-01 -6.13192081e-01 -1.22266102e+00 -3.26479435e-01
-6.13179922e-01 6.13247931e-01 8.43035340e-01 7.56717145e-01
4.61238354e-01 1.16101757e-01 4.54841346e-01 -1.34749010e-01
-9.12636280e-01 -1.13222253e+00 -3.70243043e-01 1.59082815e-01
9.28972811e-02 -4.32090104e-01 -4.55277681e-01 -3.23186845e-01] | [11.329200744628906, 7.644433975219727] |
2b02c65d-4316-46ea-8910-02584b54600a | learning-deep-mri-reconstruction-models-from | 2301.02613 | null | https://arxiv.org/abs/2301.02613v1 | https://arxiv.org/pdf/2301.02613v1.pdf | Learning Deep MRI Reconstruction Models from Scratch in Low-Data Regimes | Magnetic resonance imaging (MRI) is an essential diagnostic tool that suffers from prolonged scan times. Reconstruction methods can alleviate this limitation by recovering clinically usable images from accelerated acquisitions. In particular, learning-based methods promise performance leaps by employing deep neural networks as data-driven priors. A powerful approach uses scan-specific (SS) priors that leverage information regarding the underlying physical signal model for reconstruction. SS priors are learned on each individual test scan without the need for a training dataset, albeit they suffer from computationally burdening inference with nonlinear networks. An alternative approach uses scan-general (SG) priors that instead leverage information regarding the latent features of MRI images for reconstruction. SG priors are frozen at test time for efficiency, albeit they require learning from a large training dataset. Here, we introduce a novel parallel-stream fusion model (PSFNet) that synergistically fuses SS and SG priors for performant MRI reconstruction in low-data regimes, while maintaining competitive inference times to SG methods. PSFNet implements its SG prior based on a nonlinear network, yet it forms its SS prior based on a linear network to maintain efficiency. A pervasive framework for combining multiple priors in MRI reconstruction is algorithmic unrolling that uses serially alternated projections, causing error propagation under low-data regimes. To alleviate error propagation, PSFNet combines its SS and SG priors via a novel parallel-stream architecture with learnable fusion parameters. Demonstrations are performed on multi-coil brain MRI for varying amounts of training data. PSFNet outperforms SG methods in low-data regimes, and surpasses SS methods with few tens of training samples. | ['Tolga Çukur', 'Muzaffer Özbey', 'Şaban Öztürk', 'Salman UH Dar'] | 2023-01-06 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 5.89692771e-01 1.16901398e-01 -1.31034970e-01 -5.77144682e-01
-1.15784311e+00 -2.30990246e-01 5.11784613e-01 -1.26429573e-01
-6.86262190e-01 4.57590669e-01 3.00716728e-01 -5.14149308e-01
-2.86813259e-01 -3.16395849e-01 -7.66457677e-01 -8.61065924e-01
-3.36340666e-01 6.27647936e-01 3.41412485e-01 1.59877062e-01
-1.03804357e-01 6.67964995e-01 -8.39231253e-01 2.69372672e-01
4.78960454e-01 8.28429401e-01 5.67565799e-01 6.07410431e-01
3.07581246e-01 7.63092697e-01 -2.47725755e-01 1.43455669e-01
4.29165691e-01 -1.20533295e-01 -5.28057933e-01 -1.15639985e-01
2.00103238e-01 -7.06664205e-01 -4.73733068e-01 8.34618926e-01
9.07217681e-01 7.95713589e-02 5.49679399e-01 -8.46097469e-01
1.60725296e-01 5.96071422e-01 -4.64626014e-01 5.41323841e-01
-9.05300304e-02 6.67665303e-01 3.89989495e-01 -8.25517476e-01
4.56452131e-01 6.34808838e-01 1.09790528e+00 3.81124288e-01
-1.57307291e+00 -5.66299081e-01 -1.89502388e-01 -7.42901936e-02
-1.22045958e+00 -6.11876726e-01 6.45081341e-01 -4.47760522e-01
9.86237407e-01 2.26152897e-01 6.51726604e-01 9.79894757e-01
6.56278610e-01 6.27312362e-01 1.39066076e+00 -8.79682675e-02
2.94636250e-01 -3.66640687e-01 1.54726222e-01 5.31154275e-01
-4.93409252e-03 2.42062807e-01 -5.71756065e-01 -4.35932845e-01
1.19001985e+00 1.66373432e-01 -6.45311058e-01 -3.01714957e-01
-1.26107168e+00 7.05911338e-01 2.99134731e-01 1.70903355e-01
-7.67730594e-01 2.08728373e-01 3.43182564e-01 2.39089161e-01
2.09687725e-01 4.37073648e-01 -3.84719163e-01 -3.64368297e-02
-1.49642777e+00 -1.01863928e-02 6.22859597e-01 7.05945373e-01
5.06338120e-01 1.69614568e-01 -2.75608450e-01 7.74309158e-01
1.48170099e-01 4.14473444e-01 6.27744973e-01 -1.00151658e+00
8.33101198e-02 -1.75710842e-01 -1.84905723e-01 -5.90160906e-01
-9.02059913e-01 -9.14752662e-01 -1.05842638e+00 1.48281381e-01
2.87104636e-01 -2.11117849e-01 -1.23130405e+00 1.86382604e+00
3.13521415e-01 3.47223341e-01 -3.15422058e-01 1.07315314e+00
5.77669024e-01 3.36393833e-01 2.95670461e-02 -3.75295669e-01
1.47209847e+00 -6.95922017e-01 -4.32429463e-01 -3.43177080e-01
3.53554606e-01 -5.01747429e-01 1.00777507e+00 6.76953375e-01
-1.32352638e+00 -2.15856835e-01 -1.11098516e+00 1.90187648e-01
3.03117871e-01 -7.24119917e-02 6.44952357e-01 6.73003674e-01
-1.39307427e+00 7.47331738e-01 -1.39308190e+00 2.04076499e-01
4.49857742e-01 7.57500470e-01 -2.34936953e-01 -3.46932888e-01
-8.88781011e-01 9.49430943e-01 2.24923223e-01 1.47595286e-01
-1.22380376e+00 -1.30359960e+00 -7.59120703e-01 -1.21398285e-01
3.55621397e-01 -7.87343383e-01 1.31685793e+00 -7.06255078e-01
-1.62703216e+00 4.50056851e-01 1.30554453e-01 -6.46785021e-01
6.70470893e-01 -1.58607922e-02 -3.72276336e-01 6.74974680e-01
-3.81379835e-02 7.52191663e-01 1.19304895e+00 -9.59537566e-01
2.16533557e-01 -8.41910392e-02 -2.06670702e-01 1.27741948e-01
9.33932662e-02 -4.71323691e-02 -3.72990936e-01 -6.86999142e-01
4.61082339e-01 -1.03459239e+00 -4.41890419e-01 7.97833428e-02
-6.49075210e-02 3.20782632e-01 4.38947797e-01 -8.70794535e-01
7.80232549e-01 -1.97670162e+00 -2.30749831e-01 3.96772921e-01
5.54712415e-01 5.02491184e-02 -1.29257455e-01 -1.41025782e-02
-3.17348897e-01 -4.29038137e-01 -6.47624135e-01 -1.67276159e-01
-3.47397894e-01 3.08084488e-01 -1.32663995e-01 6.38531446e-01
1.23866856e-01 7.88366199e-01 -9.68276560e-01 -2.90049434e-01
1.63640946e-01 5.68548262e-01 -8.34899783e-01 2.95781285e-01
5.01473211e-02 8.91594112e-01 -1.42727762e-01 3.65767002e-01
7.10334003e-01 -6.09465241e-01 5.24827778e-01 -7.48432875e-01
1.28528625e-01 2.07405254e-01 -9.12676811e-01 2.01490569e+00
-5.83352506e-01 1.73933566e-01 5.85359275e-01 -1.17950559e+00
4.52134728e-01 5.04076421e-01 8.65280092e-01 -6.56899631e-01
1.02863409e-01 3.08917493e-01 1.72457993e-01 -5.35973907e-01
1.35434359e-01 -7.02528179e-01 1.91195935e-01 6.78249002e-01
2.45557994e-01 -4.26795930e-01 -1.69528380e-01 2.29232505e-01
1.64229548e+00 1.41696915e-01 -8.36896971e-02 -4.65897828e-01
1.97484866e-02 -1.61328107e-01 6.21466279e-01 1.13472617e+00
-6.46016523e-02 9.13273573e-01 2.65064925e-01 -4.77498055e-01
-1.06383121e+00 -1.40262341e+00 -6.19382679e-01 7.49107122e-01
-3.31590950e-01 -2.19038233e-01 -5.56287050e-01 -5.00182688e-01
-3.00509840e-01 4.96765137e-01 -2.83238620e-01 -5.26215918e-02
-8.24761450e-01 -1.33798432e+00 5.10888517e-01 5.42517960e-01
2.66791284e-01 -8.41150880e-01 -9.57136571e-01 6.02831006e-01
-1.67751044e-01 -1.03731787e+00 -3.74992102e-01 6.49416924e-01
-1.21720219e+00 -8.43861222e-01 -9.61313725e-01 -1.63258106e-01
7.54806459e-01 2.26628892e-02 1.12395597e+00 -8.29622746e-02
-3.47575366e-01 4.60157126e-01 5.16590439e-02 1.85805019e-02
-2.92220205e-01 -4.10801843e-02 1.82100430e-01 -3.51726323e-01
-2.34302863e-01 -1.02406883e+00 -9.37022865e-01 2.06984863e-01
-1.10947645e+00 3.65593642e-01 8.84995759e-01 1.31795144e+00
6.63303137e-01 -2.88972914e-01 6.53606892e-01 -8.67975950e-01
3.73017579e-01 -7.59946048e-01 -5.02205312e-01 -3.79907750e-02
-5.59716880e-01 3.29785705e-01 4.47302520e-01 -5.83478630e-01
-9.68964875e-01 -4.99182194e-02 -3.83145005e-01 -6.42995894e-01
1.14582367e-01 7.38137722e-01 3.66570950e-01 -2.06789702e-01
7.11178958e-01 2.86740750e-01 4.38828588e-01 -4.75900650e-01
9.37241167e-02 2.21633837e-01 7.80824900e-01 -8.99907053e-01
2.10745990e-01 5.24938226e-01 6.97441846e-02 -5.78382075e-01
-6.59604073e-01 -2.63082981e-01 -3.25138211e-01 -2.64356017e-01
5.93074441e-01 -9.17574763e-01 -5.09338558e-01 3.90968978e-01
-7.43479609e-01 -6.47255242e-01 -3.01884800e-01 9.79322195e-01
-6.84707522e-01 2.67270356e-01 -1.01023006e+00 -3.14751178e-01
-4.75827485e-01 -1.50691175e+00 9.03380573e-01 -1.84137583e-01
-2.23698005e-01 -8.38718832e-01 -1.53135657e-01 2.24965334e-01
7.96323776e-01 9.14537087e-02 9.36612844e-01 -5.96019328e-01
-6.24144256e-01 -3.30754765e-03 -2.62103528e-01 3.55326116e-01
-1.44148730e-02 -8.83501470e-01 -8.55569303e-01 -7.17898071e-01
6.99403107e-01 -3.10352862e-01 6.76855862e-01 8.93983364e-01
1.42240322e+00 -1.24587260e-01 -9.44500789e-02 9.11544859e-01
1.31603706e+00 7.18692131e-03 4.93798614e-01 8.62250291e-03
4.83204275e-01 2.63127685e-01 -5.90497553e-02 3.27204555e-01
1.82126001e-01 4.64978963e-01 1.48764580e-01 -2.92280704e-01
-4.03298765e-01 1.00142986e-01 1.65999070e-01 1.05203426e+00
2.49644350e-02 3.24221015e-01 -1.16622257e+00 3.07416558e-01
-1.55040693e+00 -6.27381861e-01 1.20043188e-01 2.23175931e+00
1.16664052e+00 2.34751061e-01 -4.17648293e-02 -8.04506615e-02
3.72038543e-01 3.45042497e-02 -8.78231823e-01 2.36381412e-01
1.50046989e-01 5.43611884e-01 7.64266074e-01 3.80993158e-01
-7.65374422e-01 2.55157322e-01 6.82846832e+00 8.52939367e-01
-1.46355450e+00 7.53863096e-01 7.19652414e-01 -3.86903703e-01
-2.59803832e-01 -4.25886223e-03 -2.94961363e-01 4.73920703e-01
1.19534159e+00 1.90145329e-01 4.59713340e-01 4.89925086e-01
2.28764683e-01 -4.70640093e-01 -1.04396236e+00 1.05557346e+00
1.38353314e-02 -1.47468460e+00 -3.55374515e-01 -5.37575521e-02
4.63907242e-01 7.55163074e-01 -1.31195765e-02 1.13074757e-01
3.62786651e-01 -9.90588844e-01 5.43077409e-01 6.18046045e-01
1.10179698e+00 -4.34444189e-01 7.05169857e-01 4.17631745e-01
-6.08168960e-01 1.51928097e-01 -4.36144657e-02 1.42797589e-01
4.18027759e-01 1.02701497e+00 -1.02485061e+00 5.95281482e-01
4.72688317e-01 3.84995639e-01 -2.18739793e-01 9.74079132e-01
-6.16539503e-03 9.04818952e-01 -6.94460809e-01 7.42870212e-01
1.83527902e-01 -6.92242458e-02 6.42107129e-01 1.18073416e+00
2.91057438e-01 3.79385710e-01 3.45068634e-01 7.67940104e-01
1.20947607e-01 -4.05660063e-01 -3.18550169e-01 4.15252596e-01
2.42358983e-01 1.30407524e+00 -8.33137631e-01 -4.79155779e-01
-1.92787260e-01 5.55935383e-01 1.19268551e-01 2.75860429e-01
-6.59692526e-01 2.89409637e-01 -4.05093879e-02 5.32269537e-01
8.32180679e-02 -4.14947420e-01 -3.70758802e-01 -1.09408712e+00
-1.68574899e-01 -9.19033349e-01 3.05487722e-01 -8.15773785e-01
-1.36312914e+00 5.94261229e-01 4.34288651e-01 -1.09404790e+00
-3.18607241e-01 -4.92596984e-01 -3.77303779e-01 9.05040085e-01
-1.46545780e+00 -9.11700666e-01 -7.31788278e-02 6.32845044e-01
2.34705880e-01 9.93068218e-02 6.52183592e-01 3.49546403e-01
-2.25773111e-01 3.16289693e-01 6.39893487e-02 -7.04849735e-02
6.78005099e-01 -9.70856726e-01 -2.64921840e-02 6.58104062e-01
-2.45594651e-01 7.47461677e-01 6.56939924e-01 -6.91182435e-01
-1.57330418e+00 -7.76606441e-01 1.75693855e-01 -4.31099124e-02
6.37457669e-01 -6.88719600e-02 -1.05430472e+00 6.27011061e-01
9.26718190e-02 4.32885289e-01 7.31859744e-01 5.76267615e-02
-1.50321454e-01 -1.93377972e-01 -1.34427476e+00 3.74561548e-01
8.27618778e-01 -6.46848083e-01 -6.55200958e-01 5.45969605e-01
4.00983870e-01 -7.71201432e-01 -1.17453444e+00 5.94827831e-01
5.16544282e-01 -8.34670126e-01 9.65785265e-01 5.14840037e-02
1.90382123e-01 -2.60923475e-01 1.15385570e-01 -1.26805341e+00
-3.51901084e-01 -6.20735407e-01 -1.80341884e-01 4.30944681e-01
1.02238968e-01 -7.87075937e-01 7.43222356e-01 8.33009541e-01
-5.25158286e-01 -8.74035180e-01 -1.16332972e+00 -7.69462824e-01
2.38597039e-02 -7.05644190e-01 4.13614929e-01 9.09731865e-01
-2.20985055e-01 5.82770593e-02 -2.50557721e-01 1.94853052e-01
9.30990756e-01 -2.08524436e-01 2.58801520e-01 -7.61586308e-01
-8.19605887e-01 -2.43270785e-01 -8.48556906e-02 -1.02064884e+00
-6.73946962e-02 -1.15746641e+00 2.67593682e-01 -1.05537343e+00
3.73855203e-01 -8.18612754e-01 -4.85643327e-01 5.97821891e-01
-4.58732508e-02 3.69582146e-01 1.23527452e-01 3.77171129e-01
-2.12145135e-01 3.34428668e-01 1.18696547e+00 2.69743451e-03
-2.27623116e-02 -2.79922813e-01 -3.27603042e-01 6.96395099e-01
5.83550692e-01 -6.28266037e-01 -6.29705191e-01 -5.81423521e-01
5.59475496e-02 7.41857529e-01 6.46024644e-01 -1.24656403e+00
5.42915583e-01 1.91022500e-01 6.06222987e-01 -6.19033694e-01
3.72248083e-01 -6.40925229e-01 6.04216218e-01 6.51879668e-01
-1.29058763e-01 1.34190068e-01 3.38717580e-01 3.65012437e-01
8.24596267e-03 -2.12084413e-01 9.75141525e-01 -1.99850827e-01
-1.44046068e-01 4.56323057e-01 -4.72593099e-01 9.77365747e-02
3.62033159e-01 -1.18380755e-01 1.37329802e-01 -2.11147562e-01
-1.16899931e+00 7.29079694e-02 2.14726523e-01 -1.39364243e-01
6.92440987e-01 -1.03899109e+00 -6.65028274e-01 5.14523327e-01
-4.66950446e-01 2.49718487e-01 6.67181253e-01 1.53427863e+00
-3.87333721e-01 6.98627681e-02 -2.40061939e-01 -1.08857429e+00
-6.51865780e-01 1.99679226e-01 4.82107073e-01 -6.50272071e-01
-9.96301711e-01 7.59785891e-01 2.31688648e-01 -5.17454088e-01
-3.03219080e-01 -2.09276944e-01 4.35113877e-01 -3.17045480e-01
5.64415634e-01 7.18043372e-03 5.11389017e-01 -1.08909227e-01
-3.01758856e-01 8.15288946e-02 -3.88805062e-01 -4.33218509e-01
1.67730510e+00 1.66681647e-01 1.24101192e-02 1.73902646e-01
1.09070480e+00 -3.71089071e-01 -1.74090552e+00 -4.28680599e-01
-1.59353837e-01 -1.59844607e-01 6.89692318e-01 -1.04273760e+00
-1.15727687e+00 7.10461378e-01 7.18335092e-01 -4.25070763e-01
1.28549218e+00 -2.63693213e-01 7.53738046e-01 1.50152639e-01
7.02334583e-01 -8.72384727e-01 -6.94774615e-04 3.32502395e-01
8.17082524e-01 -1.01902199e+00 3.04572582e-01 -5.17840683e-02
-4.11043614e-01 1.11897075e+00 1.35441691e-01 -2.06558183e-01
8.53761017e-01 7.61317968e-01 -2.07979605e-01 -4.84822452e-01
-5.54592669e-01 4.17758048e-01 1.65154651e-01 4.29009914e-01
2.51900822e-01 1.09825395e-01 -1.46445289e-01 5.94092309e-01
-3.64752226e-02 4.15291131e-01 3.01697850e-01 1.09796798e+00
-2.49742065e-03 -1.03987396e+00 -4.41710204e-01 9.39817846e-01
-5.09068072e-01 -5.46273172e-01 7.92928874e-01 5.06270170e-01
-9.99351442e-02 4.56979185e-01 2.86262203e-02 5.22499606e-02
-1.04856178e-01 -5.39578907e-02 9.28462923e-01 -5.29696167e-01
-7.81955540e-01 3.90094876e-01 -3.98662500e-02 -7.43750215e-01
-3.86946559e-01 -8.42257202e-01 -1.25262642e+00 -1.09558932e-01
-2.08875537e-01 -1.40506532e-02 6.95592523e-01 1.05485058e+00
4.06585753e-01 7.22705960e-01 4.55085307e-01 -1.15899789e+00
-8.99008334e-01 -8.79081070e-01 -4.85009164e-01 1.04179502e-01
3.93994868e-01 -5.97288132e-01 -1.74243808e-01 -6.87924400e-02] | [13.524965286254883, -2.384619951248169] |
23df7b94-ee91-4339-8cc1-8879e912b844 | bottom-up-higher-resolution-networks-for | 1908.10357 | null | https://arxiv.org/abs/1908.10357v3 | https://arxiv.org/pdf/1908.10357v3.pdf | HigherHRNet: Scale-Aware Representation Learning for Bottom-Up Human Pose Estimation | Bottom-up human pose estimation methods have difficulties in predicting the correct pose for small persons due to challenges in scale variation. In this paper, we present HigherHRNet: a novel bottom-up human pose estimation method for learning scale-aware representations using high-resolution feature pyramids. Equipped with multi-resolution supervision for training and multi-resolution aggregation for inference, the proposed approach is able to solve the scale variation challenge in bottom-up multi-person pose estimation and localize keypoints more precisely, especially for small person. The feature pyramid in HigherHRNet consists of feature map outputs from HRNet and upsampled higher-resolution outputs through a transposed convolution. HigherHRNet outperforms the previous best bottom-up method by 2.5% AP for medium person on COCO test-dev, showing its effectiveness in handling scale variation. Furthermore, HigherHRNet achieves new state-of-the-art result on COCO test-dev (70.5% AP) without using refinement or other post-processing techniques, surpassing all existing bottom-up methods. HigherHRNet even surpasses all top-down methods on CrowdPose test (67.6% AP), suggesting its robustness in crowded scene. The code and models are available at https://github.com/HRNet/Higher-HRNet-Human-Pose-Estimation. | ['Bin Xiao', 'Bowen Cheng', 'Thomas S. Huang', 'Lei Zhang', 'Jingdong Wang', 'Honghui Shi'] | 2019-08-27 | higherhrnet-scale-aware-representation | null | null | cvpr-2020-6 | ['2d-human-pose-estimation'] | ['computer-vision'] | [-4.87202644e-01 -5.82943968e-02 5.29639900e-01 -1.74618855e-01
-9.17249739e-01 -1.45139217e-01 2.33909249e-01 -1.38868559e-02
-7.69714952e-01 5.88573098e-01 3.16821396e-01 7.69958079e-01
2.82907695e-01 -6.72734201e-01 -8.47191095e-01 -2.98958033e-01
-1.46772176e-01 7.24898994e-01 5.84445357e-01 -6.21415377e-01
-3.49362165e-01 4.45734650e-01 -1.75829077e+00 3.27523530e-01
5.09615898e-01 8.95456970e-01 5.94903715e-02 1.03311312e+00
6.49891019e-01 2.27065310e-01 -9.00533736e-01 -8.04301560e-01
4.52437669e-01 1.09875150e-01 -3.54084939e-01 -1.00481436e-01
1.22284913e+00 -3.80240649e-01 -5.58566034e-01 7.39234090e-01
1.18255508e+00 2.42097870e-01 2.92029709e-01 -1.02609062e+00
-4.63026941e-01 -2.33280230e-02 -7.87227869e-01 3.40420336e-01
8.59825075e-01 2.24507600e-01 7.42094457e-01 -1.37995279e+00
5.49150407e-01 1.63753176e+00 1.18419886e+00 5.39740801e-01
-8.42739522e-01 -5.42962611e-01 2.62676686e-01 8.43236968e-02
-1.83354950e+00 -7.86993653e-02 2.11050615e-01 -4.33508247e-01
1.10760736e+00 1.47225484e-01 9.80231881e-01 1.15523183e+00
1.96107939e-01 8.50588441e-01 1.00914407e+00 -5.11540696e-02
-1.88415870e-01 -1.86237410e-01 -1.16977066e-01 8.11418056e-01
5.63372016e-01 1.04868338e-01 -9.31508839e-01 5.51453382e-02
1.04570067e+00 1.41829699e-01 -1.29324079e-01 -2.10694835e-01
-1.28758085e+00 5.33580065e-01 1.00659049e+00 2.36199461e-02
-5.65652728e-01 4.09523964e-01 3.46909344e-01 -9.70864743e-02
4.92145956e-01 4.31762636e-01 -6.58619881e-01 -5.43075651e-02
-8.90689492e-01 8.63250494e-01 2.85859585e-01 1.06349838e+00
3.82087439e-01 -7.60119706e-02 -6.25753045e-01 8.20866227e-01
4.46794368e-02 9.26186621e-01 3.26419383e-01 -7.13362157e-01
7.46455610e-01 4.21674192e-01 4.10184324e-01 -1.08748519e+00
-9.68189836e-01 -8.93817365e-01 -7.77377188e-01 7.83511475e-02
8.31464589e-01 -3.84486258e-01 -1.07179594e+00 1.58178234e+00
4.52984035e-01 -5.64527251e-02 -2.69604951e-01 1.27645981e+00
1.04146469e+00 5.01272202e-01 9.51633379e-02 3.08160454e-01
1.97214460e+00 -1.19645131e+00 -4.17741686e-01 -5.32237291e-01
2.04796642e-01 -4.31283772e-01 9.78463233e-01 5.80415845e-01
-1.07530379e+00 -1.12772548e+00 -9.17822003e-01 -2.95643806e-01
-3.75433683e-01 6.02537811e-01 5.98505676e-01 6.46360159e-01
-9.50886965e-01 5.11359274e-01 -6.12974584e-01 -5.39478362e-01
4.26173657e-01 4.77832884e-01 -6.95438981e-01 -8.75597373e-02
-1.23764324e+00 8.61158729e-01 9.07759294e-02 3.11604291e-01
-6.07977033e-01 -7.67736018e-01 -1.10521054e+00 -2.37290859e-01
4.84466434e-01 -1.16268778e+00 1.15705943e+00 -2.72035420e-01
-1.33941817e+00 6.87199175e-01 -2.05421388e-01 -5.72547972e-01
1.12486041e+00 -1.16177678e+00 -2.79470354e-01 2.70756543e-01
3.65117878e-01 9.54808235e-01 8.83228600e-01 -9.47929442e-01
-8.87657642e-01 -7.02920377e-01 1.36349220e-02 5.49513161e-01
-2.33807624e-03 -6.86395243e-02 -9.99330997e-01 -8.08938324e-01
6.36594370e-02 -1.02224612e+00 -4.60111022e-01 -7.94249326e-02
-3.81534606e-01 -2.33634681e-01 2.74717987e-01 -1.00927043e+00
9.46986437e-01 -1.71336162e+00 2.00789064e-01 -3.07382699e-02
3.95764053e-01 2.32567802e-01 -3.92864645e-02 1.17726862e-01
1.83982596e-01 -3.67492229e-01 1.89485699e-01 -8.34376037e-01
3.28620486e-02 -3.55391115e-01 2.91341126e-01 7.68977940e-01
6.81247935e-02 1.08658123e+00 -5.81601381e-01 -4.37496871e-01
4.67857838e-01 1.00802314e+00 -6.59934938e-01 9.76554155e-02
2.83295602e-01 4.91854519e-01 -7.42058083e-02 7.34450698e-01
7.66797960e-01 -6.16379119e-02 -4.20181870e-01 -3.43381733e-01
-3.16497535e-02 -3.73249501e-01 -1.45778883e+00 1.71250641e+00
-3.25259417e-01 5.83079100e-01 -1.87026441e-01 -1.53570801e-01
7.01572418e-01 7.41657913e-02 1.90125972e-01 -6.31461382e-01
1.19087070e-01 8.12265575e-02 -4.28462088e-01 -2.07457066e-01
7.08349049e-01 1.07656822e-01 -4.67017114e-01 -4.92830962e-01
1.47646993e-01 1.61403418e-01 3.47441286e-01 1.87215190e-02
7.92555451e-01 2.86715120e-01 3.54387641e-01 1.07962303e-02
5.47710359e-01 -3.91756535e-01 5.57121694e-01 8.32415938e-01
-3.53753328e-01 1.16943288e+00 5.82709201e-02 -7.64324248e-01
-8.66632879e-01 -1.07494164e+00 2.01544296e-02 1.46755695e+00
1.15259424e-01 -6.30089343e-01 -9.29465950e-01 -5.63128531e-01
2.78212190e-01 4.11861017e-02 -9.25480604e-01 3.81716192e-01
-8.70837629e-01 -6.71156764e-01 6.68951809e-01 9.38268006e-01
7.42812037e-01 -7.66106546e-01 -7.52786219e-01 5.20489663e-02
-4.41697240e-01 -1.49426210e+00 -6.19857132e-01 -3.86718094e-01
-2.96973586e-01 -1.02006555e+00 -1.47756445e+00 -4.55608279e-01
5.28257251e-01 1.70573935e-01 1.05352080e+00 -2.14512035e-01
-5.18919528e-01 4.36069578e-01 -2.46697992e-01 -4.92849469e-01
4.65099961e-01 1.66339934e-01 4.12635654e-01 -1.06336564e-01
2.07141846e-01 -1.92774937e-01 -9.87177730e-01 4.39724654e-01
5.80709949e-02 -1.32422447e-01 6.43581629e-01 5.25175750e-01
6.45966351e-01 -5.73207103e-02 3.48196745e-01 -3.53351057e-01
2.08777755e-01 1.70202062e-01 -5.74002385e-01 2.61948183e-02
6.57511363e-03 -2.78376788e-01 5.34436762e-01 -3.01419228e-01
-8.17267418e-01 2.83915520e-01 -3.39519739e-01 -4.24696267e-01
-2.69489199e-01 -2.98130482e-01 -2.88844824e-01 -7.77706206e-02
9.43381786e-01 -9.10218433e-03 -2.08862960e-01 -5.12880743e-01
1.88854530e-01 2.97353327e-01 8.29381824e-01 -3.55281681e-01
1.00038540e+00 6.60763919e-01 2.11742185e-02 -1.07204008e+00
-1.14599180e+00 -7.11347640e-01 -1.00630641e+00 -4.01406646e-01
1.22124720e+00 -1.61020911e+00 -9.71371710e-01 6.73019171e-01
-1.06445062e+00 -1.70831904e-01 -3.49830329e-01 5.25436699e-01
-4.90399092e-01 1.28577903e-01 -7.05803275e-01 -7.57371724e-01
-4.80549276e-01 -9.80247200e-01 1.65516162e+00 2.65057892e-01
-2.56651938e-01 -5.06685078e-01 -9.37996432e-02 7.57153213e-01
2.34544739e-01 5.13337374e-01 -4.53001201e-01 -3.66032720e-01
-2.88908780e-01 -5.78239262e-01 -1.73488230e-01 7.78959244e-02
-4.32037950e-01 -5.11993051e-01 -9.81529593e-01 -7.48412192e-01
-6.47400320e-01 -2.35111594e-01 9.17052984e-01 7.52717018e-01
7.72054553e-01 -8.00073743e-02 -2.45278925e-01 8.21177006e-01
1.11279988e+00 -7.30730534e-01 4.82654601e-01 4.27433342e-01
1.11003673e+00 6.17376864e-01 8.21235716e-01 6.52579606e-01
7.77317464e-01 1.21247613e+00 1.51055515e-01 -2.85946578e-01
-4.44138378e-01 -4.53972518e-01 3.47376049e-01 4.97473106e-02
-6.68896854e-01 8.32169652e-02 -8.61651897e-01 2.99601018e-01
-1.85111010e+00 -9.60363746e-01 -1.80548787e-01 2.17601800e+00
3.31782281e-01 2.12663844e-01 9.38808262e-01 -5.71870543e-02
7.91755736e-01 3.58883440e-02 -4.17170614e-01 2.71043599e-01
-1.46227032e-01 8.48078802e-02 7.62093961e-01 5.00058651e-01
-1.54561615e+00 1.10334551e+00 5.52869415e+00 7.75404394e-01
-6.39271855e-01 2.67441511e-01 2.89425403e-01 -4.64898825e-01
6.55336022e-01 -6.88878179e-01 -1.48578405e+00 2.08392635e-01
5.29476523e-01 3.49192202e-01 4.89309877e-02 1.06065309e+00
1.12321220e-01 -1.51610538e-01 -8.54183793e-01 1.48089123e+00
3.00107956e-01 -9.51254964e-01 -1.95112213e-01 5.69779687e-02
5.92735469e-01 8.29126034e-03 5.16713946e-04 7.08681226e-01
-4.91112731e-02 -1.18182456e+00 1.00598967e+00 4.75039393e-01
8.18091214e-01 -1.15830481e+00 1.18730950e+00 2.41988629e-01
-1.79454744e+00 -2.29778022e-01 -5.28691828e-01 -1.91617876e-01
2.17122421e-01 4.32721794e-01 -6.20989263e-01 5.22659838e-01
1.21069515e+00 5.21036267e-01 -1.15940225e+00 1.08425021e+00
-3.12305093e-01 -9.05945003e-02 -4.36053425e-01 8.76125321e-02
-1.68296471e-01 5.93243361e-01 5.43069839e-01 1.76776958e+00
2.92899072e-01 3.61711606e-02 5.24604082e-01 1.81767970e-01
7.44919777e-02 8.74895602e-02 -2.20694765e-01 8.08290899e-01
2.25850165e-01 1.25266469e+00 -5.56967139e-01 -3.77955228e-01
-1.09666795e-01 1.30179358e+00 4.47196752e-01 2.11669818e-01
-1.05174959e+00 -3.48775178e-01 6.83960438e-01 6.64007723e-01
5.63941061e-01 -2.97620684e-01 -2.14950014e-02 -1.19608629e+00
2.05571264e-01 -8.79106283e-01 4.57671583e-01 -6.70692205e-01
-1.15212369e+00 7.39336550e-01 2.79799163e-01 -1.19043875e+00
-2.55722910e-01 -8.31624269e-01 -3.72276679e-02 7.31858075e-01
-1.02500236e+00 -1.56000054e+00 -7.25241125e-01 7.32523680e-01
6.97817087e-01 -1.66399211e-01 5.90151727e-01 3.03616643e-01
-4.69667196e-01 1.13384533e+00 -6.32987380e-01 4.18470263e-01
7.99914122e-01 -1.46954823e+00 7.36539721e-01 9.89339828e-01
-1.05250686e-01 6.07099891e-01 1.01341581e+00 -8.23626578e-01
-1.07134521e+00 -1.12665093e+00 7.53932238e-01 -8.98743808e-01
9.22667384e-02 -7.49094605e-01 -3.89073551e-01 5.29909492e-01
-1.97153404e-01 4.96161431e-01 4.06162888e-01 3.51781666e-01
-4.94559258e-01 -2.06315860e-01 -1.24493015e+00 6.26450539e-01
1.30531085e+00 -1.24621712e-01 -5.94572067e-01 3.67708951e-01
7.04020381e-01 -9.17694867e-01 -9.43903804e-01 4.90470141e-01
6.70184970e-01 -1.14216447e+00 1.46782875e+00 -9.03069228e-02
7.47838467e-02 -4.81270641e-01 -1.41774118e-01 -1.07314420e+00
-6.84373140e-01 -5.40630519e-01 -4.20248210e-01 6.36533380e-01
2.06015348e-01 -5.71989477e-01 9.77083147e-01 2.42040545e-01
1.32494405e-01 -8.04082632e-01 -9.86813188e-01 -8.74430716e-01
-1.75522014e-01 -3.47952783e-01 3.47882807e-01 1.89057514e-01
-3.60167265e-01 4.52362329e-01 -6.69629335e-01 5.93362153e-01
1.07378721e+00 -3.74551475e-01 1.31601715e+00 -1.15682423e+00
-4.72410381e-01 -1.13148004e-01 -8.35360110e-01 -1.23611879e+00
-3.22916299e-01 -2.43663490e-01 -1.26526710e-02 -1.71524072e+00
1.52730167e-01 2.47347653e-01 1.21489629e-01 3.10668886e-01
-4.45455194e-01 8.11795056e-01 6.59764528e-01 7.14342818e-02
-9.18358147e-01 4.35325176e-01 1.28111446e+00 2.33890861e-01
-2.01276466e-01 2.06428200e-01 -5.36628366e-01 1.00374341e+00
7.29055107e-01 -1.00692250e-01 -1.66185461e-02 -1.01917520e-01
1.50744885e-01 -8.00216645e-02 7.93787301e-01 -1.92216074e+00
3.31094354e-01 4.01167929e-01 1.22949815e+00 -9.26770449e-01
9.00919914e-01 -4.55419987e-01 -6.18071780e-02 7.47334599e-01
-3.47173586e-02 3.74540776e-01 2.30658457e-01 5.39659023e-01
2.85953023e-02 4.20550525e-01 8.38103175e-01 -3.07746768e-01
-8.01507950e-01 5.27233005e-01 7.46727735e-02 2.41490543e-01
9.03624713e-01 -4.35592204e-01 -1.38009608e-01 -3.55384618e-01
-1.12838614e+00 2.41251305e-01 2.64053792e-01 5.98904252e-01
7.83006191e-01 -1.22956967e+00 -1.09638953e+00 1.07317910e-01
1.73805907e-01 2.42548063e-01 6.19908631e-01 7.81850100e-01
-6.22093976e-01 5.02283514e-01 -2.36193612e-01 -5.85436344e-01
-1.48500097e+00 2.15396032e-01 5.50947189e-01 -5.38344264e-01
-9.17443991e-01 1.28452504e+00 2.66137242e-01 -4.64169234e-01
2.56954223e-01 -1.14401095e-01 -2.06811458e-01 1.64668113e-01
8.77993822e-01 6.32181168e-01 -3.08326688e-02 -1.19979918e+00
-7.79203713e-01 1.09401000e+00 -7.00822249e-02 -1.48253635e-01
1.22767150e+00 -9.19516459e-02 4.53363955e-01 9.26277786e-02
9.53848958e-01 2.43240654e-01 -1.69746065e+00 -3.81636024e-02
-5.67123711e-01 -5.39582789e-01 -4.15983796e-01 -7.08076000e-01
-9.14696455e-01 7.69251287e-01 9.03075218e-01 -4.19741392e-01
7.40103185e-01 4.06672470e-02 7.12242961e-01 4.13456917e-01
7.00520575e-01 -1.32121253e+00 5.32198191e-01 5.66895962e-01
1.25167692e+00 -1.27371955e+00 3.68167430e-01 -5.48870623e-01
-8.08742583e-01 8.24343324e-01 1.06886232e+00 -4.19772238e-01
3.89866352e-01 1.14717402e-01 -4.23611440e-02 -3.82995069e-01
-2.43290439e-01 -6.49361372e-01 7.21472204e-01 8.96746933e-01
4.04950768e-01 2.28533670e-01 1.20222270e-01 7.25293934e-01
-8.88155341e-01 -2.35068604e-01 3.08632888e-02 4.92524922e-01
-5.05066693e-01 -3.03129375e-01 -8.86207223e-01 7.79846609e-02
-5.07688105e-01 1.19450018e-01 -1.57578915e-01 1.00703919e+00
5.38764477e-01 8.04027855e-01 -1.03005663e-01 -5.11561275e-01
1.08281839e+00 -3.28768075e-01 6.38423681e-01 -4.95695442e-01
-7.65276670e-01 1.13521822e-01 1.15903437e-01 -9.98604417e-01
-1.07095242e-02 -8.28770876e-01 -1.05515170e+00 -3.26624006e-01
-7.08943559e-03 -1.75796539e-01 2.86969423e-01 6.52751863e-01
3.61532390e-01 7.12715864e-01 -4.94612008e-03 -1.63686216e+00
-2.92662263e-01 -1.03507352e+00 -2.59886801e-01 5.15048623e-01
3.22828919e-01 -8.86828423e-01 -5.50505742e-02 -3.42101663e-01] | [7.22398042678833, -0.7391964793205261] |
f38d5088-52d2-4574-82a3-0562404309c4 | a-reinforcement-learning-environment-for-2 | 2107.07373 | null | https://arxiv.org/abs/2107.07373v2 | https://arxiv.org/pdf/2107.07373v2.pdf | A Reinforcement Learning Environment for Mathematical Reasoning via Program Synthesis | We convert the DeepMind Mathematics Dataset into a reinforcement learning environment by interpreting it as a program synthesis problem. Each action taken in the environment adds an operator or an input into a discrete compute graph. Graphs which compute correct answers yield positive reward, enabling the optimization of a policy to construct compute graphs conditioned on problem statements. Baseline models are trained using Double DQN on various subsets of problem types, demonstrating the capability to learn to correctly construct graphs despite the challenges of combinatorial explosion and noisy rewards. | ['Alok Singh', 'Johnny Ye', 'Joseph Palermo'] | 2021-07-15 | a-reinforcement-learning-environment-for-3 | https://openreview.net/forum?id=-GU1sfGnM5K | https://openreview.net/pdf?id=-GU1sfGnM5K | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 1.70796275e-01 4.90746796e-01 -3.33204508e-01 -3.85902494e-01
-6.79957807e-01 -9.45767164e-01 5.73552549e-01 4.01782304e-01
-3.28860521e-01 6.50869370e-01 -2.26014424e-02 -9.02161241e-01
1.36147007e-01 -1.59748137e+00 -1.29334712e+00 -1.86053187e-01
-4.73642200e-01 5.75375497e-01 1.31072011e-02 -1.51087344e-02
2.22300828e-01 4.09624636e-01 -1.13481677e+00 2.42614210e-01
7.13017941e-01 6.72791898e-01 6.77654818e-02 1.31131411e+00
-1.30606949e-01 1.31492710e+00 -7.06060410e-01 -3.00849557e-01
6.12830400e-01 -2.35072836e-01 -1.03476322e+00 -8.42342079e-02
4.00487989e-01 -7.03976750e-01 -3.56410384e-01 1.30423319e+00
-1.12525165e-01 2.51419008e-01 3.93797278e-01 -1.52630377e+00
-8.32989335e-01 1.04113936e+00 -1.55641465e-02 1.22710638e-01
6.47249758e-01 7.95665920e-01 1.51870382e+00 -3.00054818e-01
6.26000047e-01 1.37582517e+00 2.31130049e-02 4.49000865e-01
-1.69582582e+00 -2.73964494e-01 -3.04815620e-02 -1.74888387e-01
-8.00448895e-01 -3.74532230e-02 3.87568295e-01 -4.85322505e-01
1.62497294e+00 5.29524460e-02 9.37076926e-01 8.17355335e-01
3.39618564e-01 9.35683608e-01 1.01785314e+00 -3.10675770e-01
8.20912778e-01 -3.74053180e-01 1.48396954e-01 1.21801448e+00
-2.40487698e-02 7.16133475e-01 -2.36513019e-01 -2.38933742e-01
9.15166259e-01 -1.85751572e-01 1.61952853e-01 -4.89276916e-01
-1.02263629e+00 9.62386608e-01 5.10054588e-01 -3.15676808e-01
-1.99784622e-01 9.95789647e-01 4.16566074e-01 7.03029454e-01
-2.64400452e-01 1.07659531e+00 -3.43641281e-01 -1.63695291e-01
-4.99924839e-01 9.77494657e-01 9.24589634e-01 1.05483913e+00
8.65644991e-01 5.59609056e-01 -4.26074296e-01 -1.57522053e-01
-6.36318699e-02 6.59176350e-01 -2.21779600e-01 -1.46067965e+00
7.58231521e-01 8.20639729e-01 1.07801154e-01 -5.54244757e-01
-3.53870839e-01 -2.10227311e-01 -1.02749234e-03 5.91452241e-01
3.25829476e-01 -4.93805319e-01 -7.48479307e-01 1.69590664e+00
1.81761280e-01 8.61669928e-02 2.67240107e-01 7.90859401e-01
3.00306767e-01 9.43782747e-01 2.61525571e-01 -1.66402105e-03
7.50197887e-01 -8.57217252e-01 -3.35626513e-01 -4.38400388e-01
8.94496441e-01 -3.45224962e-02 1.16542447e+00 7.16299295e-01
-1.47698092e+00 -3.06218982e-01 -1.14942610e+00 -8.50096866e-02
-1.00023210e-01 -4.13266063e-01 1.10928249e+00 3.48022759e-01
-1.48322880e+00 7.75170088e-01 -7.12980032e-01 5.66838980e-01
3.27851683e-01 6.01354837e-01 -8.27425420e-02 -3.41640323e-01
-1.04387236e+00 8.88854265e-01 7.07143486e-01 -1.91357613e-01
-1.74945343e+00 -7.77259827e-01 -1.21900785e+00 2.51904339e-01
6.72024071e-01 -5.62546849e-01 1.78884745e+00 -1.08955097e+00
-1.65510297e+00 6.33706391e-01 3.06495458e-01 -8.67936492e-01
3.59570742e-01 1.72839448e-01 -9.06007737e-02 1.74436763e-01
-6.26217946e-02 7.28909254e-01 9.05073464e-01 -6.38272166e-01
-7.03616023e-01 -1.87456787e-01 8.42429757e-01 4.23668064e-02
3.38300258e-01 -2.52212007e-02 1.68761045e-01 -3.46117795e-01
-4.56289530e-01 -9.45396066e-01 -4.55970168e-01 -2.11397365e-01
-4.30778414e-01 -5.61687827e-01 3.37585717e-01 -4.94558334e-01
8.50347519e-01 -2.06152248e+00 6.59480274e-01 3.10599446e-01
4.51961100e-01 -2.96860307e-01 -2.79154271e-01 2.07638726e-01
-6.01464175e-02 3.14794421e-01 -2.59475231e-01 3.06526452e-01
4.38476026e-01 5.08522749e-01 -3.68841261e-01 3.95813674e-01
7.51611650e-01 1.23193371e+00 -1.29197478e+00 -1.76974446e-01
9.53237433e-03 -5.44747651e-01 -1.11528361e+00 5.78370690e-01
-1.41041386e+00 -2.25693602e-02 -5.88767469e-01 5.50432503e-01
4.44655687e-01 -2.83127695e-01 5.53697348e-01 7.34293044e-01
1.80837482e-01 5.02673268e-01 -1.17290044e+00 1.65382206e+00
-4.63785261e-01 3.72459561e-01 7.65717924e-02 -1.03608549e+00
7.06268311e-01 1.78606734e-02 -9.35458168e-02 -8.45696807e-01
-8.52417052e-02 1.92950740e-01 1.98836178e-01 -3.61404657e-01
6.71568930e-01 -1.72521174e-01 -5.39912820e-01 7.12658107e-01
2.86116362e-01 -7.76204765e-01 3.54197770e-01 5.71575701e-01
1.65959334e+00 2.39399254e-01 -1.22779958e-01 -3.01790208e-01
6.03071451e-02 2.37214997e-01 5.16165137e-01 8.95873964e-01
6.50400072e-02 -2.75869548e-01 1.35585356e+00 -6.33235812e-01
-1.04142666e+00 -1.38218105e+00 4.65348601e-01 1.36798751e+00
-3.67259443e-01 -4.67962414e-01 -6.53626859e-01 -9.36200559e-01
2.60415167e-01 1.18241191e+00 -4.64884192e-01 -3.39710623e-01
-8.46674502e-01 -1.69250146e-01 5.90191722e-01 6.09945118e-01
1.50869012e-01 -1.21352625e+00 -8.56267452e-01 3.85216177e-01
4.31372762e-01 -8.71066332e-01 -4.02894318e-01 7.07514822e-01
-8.84683907e-01 -1.29236317e+00 -7.74789229e-02 -9.60422039e-01
7.28522897e-01 -3.87125373e-01 1.55534947e+00 2.52842039e-01
-9.40378681e-02 4.15201038e-01 1.33509457e-01 -2.09621891e-01
-8.63267362e-01 -5.07015325e-02 -3.97623301e-01 -7.75561154e-01
5.18210232e-02 -5.32773495e-01 -1.38543814e-01 -4.59119111e-01
-7.10169375e-01 -1.55428313e-02 1.17593080e-01 7.71516562e-01
3.46112162e-01 1.63929313e-02 3.62498641e-01 -9.05539572e-01
9.42589402e-01 -4.92194474e-01 -1.56489503e+00 1.00180350e-01
-4.46430296e-01 8.60525489e-01 1.08401823e+00 -1.58732921e-01
-7.30170310e-01 3.12056988e-01 4.29622501e-01 -2.73445070e-01
-4.80088731e-03 5.94410419e-01 3.09862141e-02 3.00463259e-01
9.77238894e-01 1.18709944e-01 -1.26022473e-01 2.72952944e-01
6.93714917e-01 -3.08280021e-01 5.76702058e-01 -1.72234786e+00
7.88751304e-01 -3.20716590e-01 3.79451931e-01 -3.33391763e-02
-5.64265370e-01 4.12234485e-01 -1.00511014e-01 5.70537895e-02
7.29712307e-01 -6.75944507e-01 -1.17191386e+00 1.06107913e-01
-1.23964477e+00 -1.02055383e+00 -6.64514542e-01 -3.07335537e-02
-9.13480997e-01 -2.18224272e-01 -8.54135275e-01 -7.26695716e-01
4.48543532e-03 -1.64917052e+00 7.72750080e-01 1.58106685e-01
-1.00880764e-01 -8.97717714e-01 9.91026685e-02 -2.53441989e-01
1.52345270e-01 4.41093743e-01 1.66800416e+00 -5.24259269e-01
-1.15908039e+00 1.07447103e-01 -6.30251393e-02 4.61546719e-01
-4.52901691e-01 2.95717090e-01 -5.41930318e-01 -2.14807801e-02
-2.65953183e-01 -8.43176126e-01 3.91422659e-01 5.77231869e-02
1.42943466e+00 -5.76725900e-01 2.47054040e-01 4.61026460e-01
1.38892221e+00 1.29774213e-01 2.50181317e-01 1.39366701e-01
3.42321306e-01 1.73574969e-01 1.91282809e-01 3.42979729e-01
2.24017218e-01 1.82054043e-02 8.50457013e-01 5.15100777e-01
2.74642050e-01 -6.27506077e-01 6.23741090e-01 6.48061857e-02
2.21109435e-01 1.15681045e-01 -1.30473769e+00 4.54382867e-01
-1.71222425e+00 -6.86114788e-01 -4.03954126e-02 1.83726394e+00
1.24082065e+00 5.59755802e-01 1.70256972e-01 -3.52614559e-02
2.61734903e-01 1.28161937e-01 -7.28015423e-01 -9.67377901e-01
1.89394295e-01 1.08047664e+00 5.36252618e-01 8.12130511e-01
-7.29026735e-01 1.16129577e+00 7.36518478e+00 2.61136383e-01
-9.71140981e-01 -4.26979601e-01 5.10050178e-01 -1.35655195e-01
-6.75475240e-01 2.41142213e-01 -3.30820382e-01 -1.20081715e-02
1.24442518e+00 -4.43748355e-01 1.40196347e+00 1.08382082e+00
-2.72922218e-01 -2.63498574e-01 -1.86480892e+00 3.02537471e-01
-6.54587865e-01 -1.58975065e+00 -1.10832646e-01 -1.45469218e-01
8.17062199e-01 -7.20814764e-02 1.63201198e-01 9.34227824e-01
1.51170504e+00 -1.39712906e+00 8.40982974e-01 2.49207288e-01
5.79196930e-01 -1.07916903e+00 6.33134544e-02 4.72114652e-01
-7.05149233e-01 -2.26631433e-01 -8.88976082e-02 -7.68030524e-01
-3.48875970e-01 1.96406245e-01 -1.38370550e+00 -1.13462023e-02
1.40545398e-01 5.52162111e-01 -4.68522459e-01 4.86549646e-01
-9.69791770e-01 4.93962646e-01 -2.41867736e-01 -3.92795920e-01
5.79174340e-01 -2.36512482e-01 2.65364856e-01 9.54159856e-01
9.44179669e-03 1.74648359e-01 5.91121197e-01 1.77004838e+00
-4.44136500e-01 -5.36978126e-01 -6.89573109e-01 -4.87315357e-01
2.92123288e-01 9.99589682e-01 -3.47167850e-01 -5.89772940e-01
-2.07432836e-01 5.99039078e-01 8.63284886e-01 5.35583735e-01
-8.26823294e-01 -8.62613320e-02 4.60535288e-01 -2.85987347e-01
1.49484411e-01 -5.09117484e-01 -7.27039516e-01 -8.61626446e-01
6.88805105e-03 -1.33921778e+00 4.37748551e-01 -7.63084710e-01
-7.94370294e-01 -1.06275596e-01 9.77045577e-03 -3.66315454e-01
-4.92612600e-01 -7.37178266e-01 -6.59267962e-01 1.03647959e+00
-1.23001921e+00 -3.33253741e-01 3.19416583e-01 5.10555506e-01
2.44352758e-01 -1.66286990e-01 9.62058723e-01 -4.45891261e-01
-4.82407540e-01 3.26678097e-01 -5.28335631e-01 2.38473967e-01
-4.34465036e-02 -1.93950474e+00 7.27286398e-01 1.04453969e+00
-9.14221033e-02 5.60524225e-01 6.58953726e-01 -5.53466201e-01
-2.31170940e+00 -1.05996394e+00 3.50019068e-01 -4.76304084e-01
9.85773325e-01 -3.79557788e-01 -6.23577118e-01 1.07895970e+00
4.07468021e-01 4.16637987e-01 4.54961769e-02 3.07255201e-02
-5.66846609e-01 1.42474726e-01 -1.22283447e+00 6.68922961e-01
1.02974772e+00 -7.80138731e-01 -8.39459062e-01 5.23322463e-01
1.21610248e+00 -1.01215625e+00 -8.65932465e-01 -1.53670430e-01
-1.50552079e-01 -4.83922035e-01 6.60563111e-01 -1.26993179e+00
1.18040979e+00 -1.89388409e-01 -3.93517882e-01 -1.50756991e+00
-1.81689933e-01 -8.89394045e-01 -6.20078087e-01 5.90014815e-01
5.30006647e-01 -3.81021887e-01 7.96238840e-01 1.02578032e+00
-1.72749743e-01 -6.70182467e-01 -6.79022074e-01 -4.63588595e-01
4.24170136e-01 -6.53520942e-01 8.47981393e-01 7.14922845e-01
3.58218014e-01 4.50542152e-01 2.45870411e-01 2.44089931e-01
5.31969905e-01 4.47234124e-01 6.30986810e-01 -6.33696616e-01
-1.03763020e+00 -6.03162766e-01 4.18030694e-02 -8.28068316e-01
6.66821599e-01 -1.50009859e+00 1.21563151e-01 -1.13755918e+00
-1.24809094e-01 -5.37176669e-01 6.68469444e-02 7.65760660e-01
6.24385476e-03 -7.76013494e-01 2.69368678e-01 -4.82235044e-01
-7.19527066e-01 2.32503325e-01 1.33846486e+00 -4.76945341e-01
6.99387491e-02 -1.97882190e-01 -6.33293927e-01 4.21130598e-01
7.17961550e-01 -4.48656350e-01 -5.02215505e-01 -6.74982607e-01
9.93883610e-01 9.87182677e-01 5.06962717e-01 -7.00293064e-01
1.30640551e-01 -9.31112051e-01 1.71827331e-01 7.09447786e-02
-2.96712518e-01 -6.59244180e-01 -2.06409648e-01 7.53107309e-01
-9.28335369e-01 6.52676225e-01 3.25035751e-01 4.79423970e-01
-4.18893918e-02 -3.44004184e-01 5.79297960e-01 -4.48403388e-01
-7.70059943e-01 1.57346696e-01 -4.63050455e-01 5.88639438e-01
1.06460536e+00 5.02448499e-01 -4.26919788e-01 -2.25738302e-01
-6.68195128e-01 5.69109678e-01 4.67689693e-01 9.35269520e-02
5.66453516e-01 -1.06266963e+00 -3.62452388e-01 1.89465821e-01
-1.71374172e-01 6.26911342e-01 -3.63949388e-01 3.39786373e-02
-8.56763482e-01 6.65474609e-02 -3.39193642e-01 -2.47383177e-01
-5.18916070e-01 8.49710941e-01 5.13169408e-01 -7.67770529e-01
-3.79607081e-01 7.25779116e-01 -6.87358528e-02 -7.67584682e-01
2.09547088e-01 -1.12845933e+00 4.91039366e-01 -6.81747139e-01
3.37023348e-01 2.06787646e-01 6.53102295e-03 5.93910456e-01
-9.68741812e-03 -4.39704746e-01 1.75789893e-01 -1.68845326e-01
1.49942684e+00 9.80000138e-01 -2.63934374e-01 1.53296247e-01
9.65725660e-01 -3.44802260e-01 -1.31686819e+00 -1.58707544e-01
2.07045823e-01 -1.42925918e-01 1.32630914e-01 -1.02564931e+00
-8.50251079e-01 8.28385711e-01 1.95395090e-02 3.49195182e-01
8.24828386e-01 -2.26267591e-01 4.25013125e-01 8.83922517e-01
6.21181726e-01 -1.19494891e+00 3.97489309e-01 9.39021587e-01
8.08497071e-01 -8.83355260e-01 -1.24528363e-01 8.66403878e-02
-4.10380125e-01 1.29969645e+00 8.83301556e-01 -8.86760235e-01
6.10777587e-02 8.78256857e-01 -7.32324541e-01 -1.36478409e-01
-1.05290997e+00 1.82177469e-01 -2.13218942e-01 5.60084403e-01
8.42164829e-02 4.82832879e-01 1.51403576e-01 4.90641028e-01
-4.38413650e-01 2.20005363e-01 8.23323488e-01 1.23104990e+00
-2.97798067e-01 -1.09380138e+00 -1.84994474e-01 4.92197543e-01
-2.64591604e-01 -7.21281320e-02 -2.44158894e-01 7.71865308e-01
-2.03445390e-01 3.81857216e-01 1.96995884e-01 -1.88211739e-01
3.54254454e-01 1.45888582e-01 8.80977035e-01 -1.04880214e+00
-9.63215649e-01 -5.55533171e-01 3.25260967e-01 -7.66172886e-01
3.14396322e-01 -4.74618554e-01 -1.88342381e+00 -4.13270116e-01
2.74570346e-01 -1.01994038e-01 2.90398389e-01 8.11952353e-01
1.15079753e-01 8.58104885e-01 4.00373876e-01 -3.29954743e-01
-1.28845394e+00 -2.90560335e-01 -5.11141837e-01 2.87157267e-01
5.56429923e-01 -1.57037750e-01 1.91797242e-02 -1.37110934e-01] | [8.6090726852417, 7.248565673828125] |
fd2a3d8e-f379-4f86-a1c3-30b24b293133 | is-centralized-training-with-decentralized | 2305.17352 | null | https://arxiv.org/abs/2305.17352v1 | https://arxiv.org/pdf/2305.17352v1.pdf | Is Centralized Training with Decentralized Execution Framework Centralized Enough for MARL? | Centralized Training with Decentralized Execution (CTDE) has recently emerged as a popular framework for cooperative Multi-Agent Reinforcement Learning (MARL), where agents can use additional global state information to guide training in a centralized way and make their own decisions only based on decentralized local policies. Despite the encouraging results achieved, CTDE makes an independence assumption on agent policies, which limits agents to adopt global cooperative information from each other during centralized training. Therefore, we argue that existing CTDE methods cannot fully utilize global information for training, leading to an inefficient joint-policy exploration and even suboptimal results. In this paper, we introduce a novel Centralized Advising and Decentralized Pruning (CADP) framework for multi-agent reinforcement learning, that not only enables an efficacious message exchange among agents during training but also guarantees the independent policies for execution. Firstly, CADP endows agents the explicit communication channel to seek and take advices from different agents for more centralized training. To further ensure the decentralized execution, we propose a smooth model pruning mechanism to progressively constraint the agent communication into a closed one without degradation in agent cooperation capability. Empirical evaluations on StarCraft II micromanagement and Google Research Football benchmarks demonstrate that the proposed framework achieves superior performance compared with the state-of-the-art counterparts. Our code will be made publicly available. | ['Mingli Song', 'Jie Song', 'Yanhao Huang', 'Tongya Zheng', 'KaiXuan Chen', 'Yunpeng Qing', 'Shunyu Liu', 'Yihe Zhou'] | 2023-05-27 | null | null | null | null | ['multi-agent-reinforcement-learning', 'starcraft-ii', 'starcraft'] | ['methodology', 'playing-games', 'playing-games'] | [-5.51219285e-01 9.25763622e-02 -6.75022066e-01 -1.41790891e-02
-4.50222075e-01 -3.95389199e-01 4.56716865e-01 1.11311719e-01
-8.34211826e-01 1.27711713e+00 -2.32812271e-01 -4.30355906e-01
-2.89757729e-01 -8.19437504e-01 -5.90634406e-01 -1.09842074e+00
-3.48414212e-01 7.44628549e-01 2.96781123e-01 -3.55290949e-01
-9.15627256e-02 2.51928091e-01 -1.15417159e+00 -3.22909206e-01
1.07345963e+00 7.62140691e-01 4.04564649e-01 4.50318396e-01
2.81829417e-01 1.37432790e+00 -8.69273305e-01 2.05965756e-04
4.29880559e-01 -1.62078112e-01 -6.80074692e-01 3.36401701e-01
-2.36632347e-01 -9.56234932e-01 -4.33286399e-01 1.10079896e+00
3.84093881e-01 3.19317162e-01 1.57034338e-01 -1.36898708e+00
-1.26893729e-01 9.40648437e-01 -8.14010084e-01 6.57410026e-02
-6.25215322e-02 4.41262007e-01 9.01419699e-01 -9.58517417e-02
4.97065812e-01 1.14416516e+00 1.39482051e-01 6.11746371e-01
-9.55061615e-01 -8.03634584e-01 9.79981601e-01 3.69764328e-01
-1.07478762e+00 -6.47819340e-02 6.92550182e-01 1.03363860e-03
7.33252645e-01 3.51421386e-02 8.21232617e-01 8.22044015e-01
3.82550210e-02 1.04198456e+00 1.26967156e+00 -2.42677972e-01
5.16494811e-01 1.90332994e-01 -2.32323259e-01 1.14541495e+00
2.90846020e-01 4.48622584e-01 -4.37597960e-01 -4.12759155e-01
9.99142289e-01 -8.81637856e-02 -6.06868006e-02 -4.94282156e-01
-1.15137255e+00 9.68586683e-01 4.05275971e-01 6.10976182e-02
-7.15147078e-01 3.55817407e-01 5.17053664e-01 7.88521469e-01
7.67338872e-02 2.85631984e-01 -4.15567130e-01 -1.10803105e-01
-3.01847607e-01 4.09701377e-01 7.85708785e-01 7.51855016e-01
1.05033481e+00 4.17009771e-01 1.00553352e-02 6.26340806e-01
4.30656701e-01 4.49984908e-01 4.00963545e-01 -1.39033782e+00
7.17944682e-01 6.46127105e-01 2.90882826e-01 -1.01281297e+00
-2.70256460e-01 -5.75745285e-01 -8.30983996e-01 6.17917836e-01
-4.34952416e-02 -8.06890726e-01 -2.30338588e-01 1.76126516e+00
8.42926800e-01 2.69017845e-01 4.51282322e-01 1.04889953e+00
2.57703602e-01 4.35646176e-01 5.74214160e-02 -5.38961112e-01
9.10152972e-01 -1.18503094e+00 -4.33770269e-01 -1.43416509e-01
7.25819528e-01 -1.24161720e-01 5.37383139e-01 7.11987555e-01
-8.82820547e-01 -2.28891522e-01 -1.09720826e+00 7.05019653e-01
6.58116639e-02 1.94322810e-01 8.79673719e-01 3.62789929e-01
-8.49089146e-01 4.88915503e-01 -1.24386275e+00 9.19369683e-02
3.29108685e-01 6.87082946e-01 -1.13770209e-01 3.43792200e-01
-1.06875789e+00 7.87799180e-01 5.81788719e-01 -2.60366071e-02
-1.60128415e+00 -1.68997332e-01 -5.20775676e-01 -1.52209699e-01
1.21135855e+00 -5.74908078e-01 1.61033857e+00 -7.75972903e-01
-2.17705083e+00 -3.22518349e-02 4.25771683e-01 -7.28387356e-01
6.11428440e-01 -2.22362429e-01 1.27713969e-02 3.96860570e-01
9.60595012e-02 4.96943802e-01 7.95465410e-01 -1.51508343e+00
-1.18841827e+00 -9.61402580e-02 6.90413237e-01 7.73989260e-01
-6.96883321e-01 -3.26812327e-01 -3.03428948e-01 -4.75578845e-01
-2.89075911e-01 -9.33419645e-01 -7.80790448e-01 -4.11240906e-01
-2.47579724e-01 -4.93060231e-01 1.03401864e+00 -1.98166266e-01
1.18134272e+00 -1.73863041e+00 4.72192973e-01 4.00506258e-01
4.51911569e-01 2.67083228e-01 -1.09975591e-01 5.23782790e-01
7.41738021e-01 -1.71731457e-01 1.29170895e-01 -4.69315857e-01
9.25656855e-02 8.48449707e-01 -2.43120134e-01 4.68608826e-01
-4.77100223e-01 5.16505659e-01 -1.12795329e+00 -7.26447344e-01
2.38596305e-01 -1.82174772e-01 -8.36727321e-01 2.21395954e-01
-5.12363374e-01 6.97839797e-01 -1.24046648e+00 5.13821721e-01
3.48589480e-01 -2.89456218e-01 7.51226425e-01 3.52631122e-01
-2.14083463e-01 -6.18936345e-02 -1.37658250e+00 1.52143621e+00
-4.67649400e-01 1.44167449e-02 8.18742633e-01 -1.35382915e+00
7.38334417e-01 5.12143731e-01 8.97193551e-01 -6.10971034e-01
1.43307760e-01 6.75694570e-02 -1.51816802e-02 -1.67419031e-01
3.22154969e-01 2.96778321e-01 -1.09626211e-01 6.34027064e-01
-4.70343381e-02 9.26521048e-02 5.27812421e-01 4.75390226e-01
1.10356653e+00 1.41369123e-02 2.87848175e-01 -8.77525508e-02
7.19596207e-01 1.39378011e-01 1.17560756e+00 1.03410780e+00
-3.37943316e-01 -7.33556926e-01 3.75860542e-01 -4.26412851e-01
-5.98673999e-01 -6.26295924e-01 6.65463388e-01 1.19544709e+00
5.13341606e-01 -6.05843604e-01 -5.16676247e-01 -1.16941679e+00
2.26260230e-01 2.80646384e-01 -3.46862257e-01 8.47159326e-02
-8.56481791e-01 -5.23019731e-01 3.51560712e-01 1.84739217e-01
9.82448220e-01 -1.06966925e+00 -9.67127621e-01 5.18401861e-01
-5.69245778e-02 -1.01147282e+00 -2.79726595e-01 1.56815961e-01
-8.80266964e-01 -1.27181327e+00 -4.32345271e-01 -5.62762141e-01
7.06438184e-01 3.15500081e-01 7.04213917e-01 4.67584819e-01
2.04539493e-01 6.40629590e-01 -4.49582934e-01 -2.33502224e-01
-6.26911521e-01 1.80217043e-01 5.07422984e-01 -2.23201960e-01
-1.19386867e-01 -7.19830632e-01 -7.22806096e-01 5.74243248e-01
-6.42007649e-01 1.39840633e-01 8.79304826e-01 9.92057145e-01
3.44852984e-01 3.32389206e-01 8.33132982e-01 -7.05595911e-01
7.72374332e-01 -3.48966390e-01 -1.05791771e+00 2.80373991e-01
-8.48869264e-01 1.64386537e-02 1.14957201e+00 -4.06095415e-01
-1.17770517e+00 -1.88277617e-01 3.81601810e-01 -6.02169514e-01
-6.46901727e-02 4.91461188e-01 3.20516974e-01 -3.49938035e-01
3.96374196e-01 4.12713587e-01 2.20919475e-01 -2.72152781e-01
2.44927928e-01 4.78415191e-01 2.63082206e-01 -1.10811615e+00
8.52033913e-01 3.37690800e-01 -2.78656855e-02 -5.78364432e-01
-3.91278386e-01 -2.32051298e-01 -7.01158196e-02 -3.41595888e-01
3.26810509e-01 -1.11571085e+00 -1.52627301e+00 3.67586970e-01
-8.09662879e-01 -5.76100647e-01 -8.67662728e-02 6.25745118e-01
-6.10700369e-01 6.23119056e-01 -8.04179192e-01 -8.73374641e-01
-2.79216886e-01 -1.24169087e+00 3.32358658e-01 4.37255472e-01
4.10553038e-01 -9.65177119e-01 3.32176477e-01 3.33111852e-01
2.70329654e-01 -3.75576830e-03 5.83797872e-01 -5.17399311e-01
-8.46324086e-01 2.78242677e-01 3.29743862e-01 3.26135844e-01
1.70859084e-01 -2.87512004e-01 -2.41626546e-01 -8.67829382e-01
-6.22263774e-02 -9.04210985e-01 5.33060491e-01 1.12884305e-01
7.59425223e-01 -7.89316416e-01 -5.08418381e-01 2.51050800e-01
1.25547469e+00 2.96317905e-01 -1.31563723e-01 7.01784790e-01
4.10961002e-01 3.14548612e-01 1.01000464e+00 1.09053814e+00
6.08560920e-01 5.71598172e-01 7.82269597e-01 6.03439398e-02
4.32083488e-01 -3.33439261e-01 6.86371684e-01 9.83814180e-01
-2.34524712e-01 2.96778083e-02 -4.02864158e-01 2.73228556e-01
-2.44466186e+00 -8.10302615e-01 6.05170131e-01 1.95617104e+00
1.07166588e+00 -7.10368752e-02 3.02383780e-01 -2.29417995e-01
4.93666798e-01 1.21850267e-01 -8.29477310e-01 -6.35411367e-02
1.35169834e-01 -3.60440612e-01 6.17962778e-01 5.53068340e-01
-1.06755400e+00 1.09368348e+00 5.28526831e+00 9.70789015e-01
-9.33500707e-01 2.91035682e-01 3.18364382e-01 -1.75464764e-01
1.03692688e-01 1.31559387e-01 -9.20653939e-01 4.00384992e-01
5.52111387e-01 -1.91007450e-01 8.00727725e-01 1.18584716e+00
4.50410843e-01 -3.77425641e-01 -7.24434555e-01 7.07464516e-01
-4.97379929e-01 -1.33838010e+00 -5.37171289e-02 2.48785108e-01
1.06611896e+00 1.53808728e-01 -2.10060343e-01 6.64327800e-01
1.20244491e+00 -4.89616185e-01 5.82332611e-01 1.19958758e-01
1.48499221e-01 -1.05705726e+00 4.29048508e-01 9.86319125e-01
-1.24349403e+00 -7.20041573e-01 -5.34351051e-01 -2.45271266e-01
4.24481221e-02 1.65332332e-02 -8.72246563e-01 9.86257851e-01
5.49226403e-01 5.81720293e-01 -2.13381886e-01 7.40205944e-01
-4.43006337e-01 4.88476455e-01 -3.52727234e-01 -2.61582524e-01
7.06267297e-01 -4.64634389e-01 7.50815690e-01 5.09515285e-01
-1.07823811e-01 3.08453053e-01 1.25494480e+00 4.18226361e-01
8.87050256e-02 8.23828802e-02 -4.17464644e-01 -1.20278392e-02
7.85395801e-01 1.33717477e+00 -4.49992180e-01 -5.59613049e-01
-4.65356201e-01 5.30585706e-01 8.44399035e-01 5.33492863e-01
-7.30346084e-01 -4.02196497e-02 6.89154387e-01 -4.84867662e-01
3.97818625e-01 -4.28874195e-01 3.65915030e-01 -1.23289156e+00
-5.60671538e-02 -1.33961892e+00 5.26394010e-01 6.84764087e-02
-1.02680779e+00 4.42597270e-01 -2.21261717e-02 -1.25184703e+00
-6.29561663e-01 -1.78084075e-01 -7.28473485e-01 2.18762323e-01
-1.71885443e+00 -8.56881261e-01 1.15083698e-02 8.71014178e-01
4.74085867e-01 -7.22472668e-01 7.74478853e-01 6.99357539e-02
-8.70359659e-01 4.80737597e-01 2.27326304e-01 7.28615373e-02
3.52154553e-01 -1.18441772e+00 -5.36054850e-01 6.05928838e-01
-1.12410178e-02 4.67469245e-01 6.51665866e-01 -6.66912615e-01
-1.50251758e+00 -8.82125080e-01 -2.19297484e-01 4.10410672e-01
6.47052586e-01 1.77327558e-01 -5.72163820e-01 6.96816504e-01
4.43410218e-01 -2.97162116e-01 3.30580622e-01 5.61770424e-03
1.94213271e-01 -4.04380947e-01 -7.69303799e-01 7.86067009e-01
5.76557696e-01 4.65989299e-02 -2.18390897e-01 4.14482802e-01
7.01547921e-01 -2.65916616e-01 -8.18868697e-01 2.80033588e-01
2.77146883e-02 -9.74048913e-01 5.80430627e-01 -5.63361406e-01
-8.60878974e-02 -5.14534295e-01 1.05524413e-01 -1.54065382e+00
-8.76073614e-02 -1.11553836e+00 -5.30419707e-01 8.83509457e-01
8.09755027e-02 -9.23776686e-01 1.18457615e+00 3.23107392e-01
-1.59477144e-01 -9.79033053e-01 -1.12107921e+00 -8.19330752e-01
8.17649905e-03 -1.85735598e-02 4.25161213e-01 7.33886182e-01
4.04435635e-01 1.18163057e-01 -7.14317322e-01 4.16714579e-01
8.33030045e-01 2.02518716e-01 1.17140234e+00 -8.36955845e-01
-8.64884973e-01 -3.51355106e-01 2.82338951e-02 -1.52440953e+00
3.93588603e-01 -4.92439479e-01 -6.94323778e-02 -1.26644206e+00
7.18959048e-02 -8.48340034e-01 -3.19622695e-01 6.56530976e-01
1.16520673e-01 -4.06762600e-01 4.36078250e-01 3.38180184e-01
-1.35886252e+00 1.05381441e+00 1.55712104e+00 -1.15713164e-01
-4.23670262e-01 1.51957929e-01 -5.39289355e-01 8.37613344e-01
9.96078193e-01 -3.95659506e-01 -6.74666405e-01 -3.79874110e-01
2.82221921e-02 5.52241921e-01 2.70617425e-01 -9.81313467e-01
6.85815692e-01 -6.50977612e-01 -1.25671908e-01 -2.91221201e-01
2.80921489e-01 -8.47897887e-01 -3.14398348e-01 1.03433883e+00
-1.98296651e-01 1.18709467e-01 -1.31266758e-01 9.05312359e-01
-2.55933374e-01 -1.49793908e-01 5.30952513e-01 -3.64779323e-01
-7.79611290e-01 4.59320188e-01 -6.37191117e-01 -2.23793060e-01
1.39565539e+00 2.87176698e-01 -4.64301586e-01 -6.36790812e-01
-6.49213433e-01 1.23779690e+00 2.74827331e-01 8.20142776e-02
5.19165874e-01 -1.14488935e+00 -6.14703357e-01 -1.46198571e-01
-4.80319887e-01 1.31421369e-02 2.57656395e-01 9.60481107e-01
-1.64529115e-01 4.77189898e-01 -1.39732882e-01 -4.36786890e-01
-1.27093995e+00 4.72048044e-01 4.24378753e-01 -8.07401836e-01
-5.50109982e-01 4.04980868e-01 5.42573910e-03 -6.09670579e-01
5.05260706e-01 -1.43720150e-01 -2.75657952e-01 -3.94933224e-01
3.17014515e-01 5.25487244e-01 -3.71610254e-01 -5.70643097e-02
-6.31740689e-02 1.85623378e-01 -4.63942319e-01 -2.68932313e-01
1.18112159e+00 -3.39287370e-01 1.09771332e-02 -8.25453848e-02
5.03060400e-01 -2.28272423e-01 -1.68622959e+00 -6.12135530e-01
-2.34277502e-01 -4.44054067e-01 1.86995506e-01 -7.32174337e-01
-1.33942270e+00 2.43692875e-01 4.90183607e-02 2.53594965e-01
8.62771749e-01 -2.30666071e-01 5.85208416e-01 1.08442342e+00
9.82373416e-01 -1.59039056e+00 4.18347418e-01 4.90127504e-01
6.90252125e-01 -1.22015965e+00 2.88567781e-01 -1.24626480e-01
-1.01482916e+00 9.34820473e-01 1.28534210e+00 -2.12818339e-01
3.38469774e-01 1.05477728e-01 -2.63082399e-03 -1.03289895e-01
-1.24119985e+00 -6.42774031e-02 -4.68468904e-01 4.58355933e-01
-3.17353815e-01 1.07909307e-01 -2.71336436e-01 4.98718262e-01
1.77277610e-01 -1.77512970e-02 4.48205620e-01 1.22193158e+00
-7.79259741e-01 -1.36620367e+00 -3.02266300e-01 3.06353182e-01
-1.53940395e-01 4.62102890e-01 -6.41686097e-02 9.61831748e-01
-1.45291686e-02 1.13072944e+00 -2.52980918e-01 -2.39807874e-01
5.51097319e-02 -5.20844877e-01 3.33937973e-01 -4.69845682e-01
-6.53768361e-01 3.28244478e-01 1.33717895e-01 -6.50617719e-01
-4.81880724e-01 -3.90060604e-01 -1.47849274e+00 -4.79807019e-01
-2.38891065e-01 7.02794611e-01 2.14865997e-01 9.61751044e-01
3.88896555e-01 6.35664701e-01 9.82501566e-01 -7.18575060e-01
-1.29009342e+00 -8.47961009e-01 -5.64620435e-01 -1.76499918e-01
3.73130888e-01 -9.55781519e-01 -1.78043872e-01 -5.20511627e-01] | [3.7710211277008057, 2.0598902702331543] |
4766f81d-0d4b-43f7-89c4-872f3feb4d8c | snp2vec-scalable-self-supervised-pre-training | 2204.06699 | null | https://arxiv.org/abs/2204.06699v1 | https://arxiv.org/pdf/2204.06699v1.pdf | SNP2Vec: Scalable Self-Supervised Pre-Training for Genome-Wide Association Study | Self-supervised pre-training methods have brought remarkable breakthroughs in the understanding of text, image, and speech. Recent developments in genomics has also adopted these pre-training methods for genome understanding. However, they focus only on understanding haploid sequences, which hinders their applicability towards understanding genetic variations, also known as single nucleotide polymorphisms (SNPs), which is crucial for genome-wide association study. In this paper, we introduce SNP2Vec, a scalable self-supervised pre-training approach for understanding SNP. We apply SNP2Vec to perform long-sequence genomics modeling, and we evaluate the effectiveness of our approach on predicting Alzheimer's disease risk in a Chinese cohort. Our approach significantly outperforms existing polygenic risk score methods and all other baselines, including the model that is trained entirely with haploid sequences. We release our code and dataset on https://github.com/HLTCHKUST/snp2vec. | ['Pascale Fung', 'Nancy Y. Ip', 'Xiaopu Zhou', 'Tiffany T. W. Mak', 'Zihan Liu', 'Tiezheng Yu', 'Samuel Cahyawijaya'] | 2022-04-14 | null | https://aclanthology.org/2022.bionlp-1.14 | https://aclanthology.org/2022.bionlp-1.14.pdf | bionlp-acl-2022-5 | ['genome-understanding'] | ['medical'] | [ 2.02556670e-01 1.96388125e-01 -2.10576788e-01 -6.36114359e-01
-8.59180748e-01 -4.19547170e-01 1.42392144e-01 2.10722163e-01
-2.63135731e-01 9.08985794e-01 6.42475426e-01 -3.88691187e-01
1.64529786e-01 -6.51296198e-01 -7.76332557e-01 -5.98422825e-01
-1.36648947e-02 5.10170102e-01 -1.77819207e-01 -1.16610311e-01
-7.60166571e-02 -5.23872554e-01 -1.29010844e+00 1.79937169e-01
1.15549052e+00 2.00421870e-01 2.21360445e-01 8.30013573e-01
-7.46527240e-02 1.36844411e-01 -2.66113937e-01 -8.90597284e-01
-2.25854158e-01 -6.29923344e-01 -5.94055712e-01 -5.34326732e-01
4.16919082e-01 -2.21442655e-01 -3.43989491e-01 1.10217988e+00
9.02343750e-01 -5.93962729e-01 4.63697523e-01 -6.33826077e-01
-1.05832684e+00 9.04981136e-01 -5.10002315e-01 2.01853320e-01
2.08903328e-01 6.09940477e-03 1.41615391e+00 -4.60111529e-01
7.08487689e-01 1.12578571e+00 9.51638520e-01 7.65770733e-01
-9.95305717e-01 -4.72489268e-01 -2.51897544e-01 3.97798032e-01
-1.26732767e+00 -1.87059268e-01 2.46144593e-01 -5.74621975e-01
9.50758636e-01 2.11255416e-01 5.98280489e-01 1.41596568e+00
3.27587128e-01 6.39018476e-01 5.52410305e-01 -4.57981229e-01
1.21900449e-02 -5.07210433e-01 5.44314563e-01 6.59165621e-01
3.29778165e-01 1.20815121e-01 -3.42302889e-01 -5.95172584e-01
6.32617235e-01 -1.24741934e-01 -1.10669710e-01 5.15850969e-02
-1.14154184e+00 9.60341454e-01 7.42277922e-03 2.05210969e-01
-2.51239717e-01 2.01256439e-01 3.46934915e-01 1.32464543e-01
5.60582817e-01 1.07188813e-01 -6.43197834e-01 -2.71431476e-01
-5.10193527e-01 2.86063608e-02 5.68060338e-01 6.50251865e-01
4.66021895e-01 -1.34340882e-01 -2.87128836e-01 1.22375512e+00
3.28592747e-01 5.14778376e-01 8.65507603e-01 -4.66554880e-01
6.53451607e-02 3.80398184e-01 -1.65870264e-01 -6.13933742e-01
-6.20084345e-01 -4.01092052e-01 -7.17234433e-01 -4.19009537e-01
7.22476915e-02 -4.88539308e-01 -9.61339295e-01 1.83250463e+00
4.15601999e-01 6.69713557e-01 1.52371779e-01 7.28562176e-01
9.50218201e-01 4.95966196e-01 2.67123431e-01 2.11812675e-01
1.60679948e+00 -7.68834412e-01 -5.03890991e-01 -8.15744847e-02
7.77479172e-01 -3.41077060e-01 7.13283837e-01 1.72516763e-01
-6.45333111e-01 -6.25721216e-01 -8.46491992e-01 -3.42426062e-01
-1.91169322e-01 6.21594377e-02 9.60836411e-01 1.27628529e+00
-8.95207226e-01 3.27746511e-01 -9.88794506e-01 -5.74810088e-01
5.83425581e-01 1.34736478e-01 -3.36617619e-01 -1.97254017e-01
-1.50599730e+00 5.74639738e-01 4.36681896e-01 -9.98427048e-02
-7.33702183e-01 -1.01236904e+00 -7.88980007e-01 2.62059361e-01
-4.46915030e-02 -7.70794392e-01 8.34014654e-01 -3.92428994e-01
-1.12633836e+00 8.45467567e-01 -2.93837667e-01 -5.27197063e-01
3.81156541e-02 -5.80525935e-01 -7.05227673e-01 -2.02160835e-01
-4.54463661e-02 6.29638493e-01 1.56140044e-01 -4.77458209e-01
-1.83687136e-01 -5.93547344e-01 -3.95264000e-01 -1.33360639e-01
-2.76232809e-01 1.97772887e-02 -5.74384749e-01 -5.49219906e-01
-4.26071167e-01 -8.46746206e-01 -5.33243537e-01 -1.26230851e-01
-5.57166398e-01 -5.68941087e-02 1.44607663e-01 -1.00812685e+00
1.02315569e+00 -2.04776645e+00 -8.00439194e-02 -7.41686001e-02
2.18395099e-01 6.42738879e-01 -5.76711357e-01 4.30535227e-01
-3.05831999e-01 5.30669093e-01 -2.60424227e-01 2.58127600e-01
-7.74914771e-03 1.52540371e-01 1.31071016e-01 2.54464537e-01
4.65480804e-01 1.24348485e+00 -9.84355390e-01 -1.43376723e-01
-1.23892665e-01 9.00057137e-01 -8.75399292e-01 8.03343430e-02
-4.66692656e-01 4.44008112e-01 -6.18742228e-01 3.11830908e-01
8.16015244e-01 -2.17990503e-01 8.59189332e-01 6.87925890e-02
-1.56813245e-02 1.94546103e-01 -4.34704155e-01 1.75622416e+00
5.93896471e-02 3.20351750e-01 -5.80883622e-01 -1.06015754e+00
8.40928733e-01 9.05794501e-02 4.07061249e-01 -4.81326491e-01
-1.03350371e-01 -2.02964589e-01 3.00036520e-01 -8.28309059e-01
-1.59955576e-01 1.09200083e-01 9.85567737e-03 3.22467625e-01
1.34143814e-01 5.56502402e-01 9.43425819e-02 1.53860196e-01
1.21205580e+00 3.76303583e-01 4.45102692e-01 -1.07119076e-01
3.45327049e-01 -7.57973939e-02 9.96644318e-01 6.05122745e-01
-5.24239950e-02 7.26147115e-01 8.43094647e-01 -4.83231813e-01
-1.12280357e+00 -8.03041935e-01 -3.14465910e-01 1.03068125e+00
-4.35038716e-01 -6.53580308e-01 -6.71451747e-01 -7.33501256e-01
1.95029929e-01 4.22148257e-01 -8.00230324e-01 -6.09090785e-03
-2.82985091e-01 -1.77434409e+00 1.03695786e+00 6.46283686e-01
1.42929420e-01 -5.05069435e-01 -1.72136113e-01 2.61125892e-01
-1.16366364e-01 -6.87978923e-01 -4.45125729e-01 1.05737574e-01
-5.06133735e-01 -1.39079821e+00 -8.51495683e-01 -5.76674938e-01
3.64747554e-01 -1.87486127e-01 1.11076593e+00 1.26581609e-01
-9.37110186e-01 8.79050195e-02 -6.11191869e-01 -7.07049072e-01
-2.75051028e-01 -1.72983065e-01 -6.42547905e-02 -2.48195887e-01
8.73777390e-01 -4.42608953e-01 -6.55780971e-01 -7.34690875e-02
-6.71117008e-01 -1.82489619e-01 7.85930812e-01 1.11651707e+00
6.11372292e-01 -1.58323392e-01 8.45781863e-01 -1.29746294e+00
3.28945190e-01 -6.64076030e-01 -4.55778152e-01 6.11643076e-01
-3.76881748e-01 1.50842234e-01 2.83614308e-01 1.10243581e-01
-1.11865723e+00 -3.47582959e-02 -8.57043743e-01 4.51980829e-01
-2.99336642e-01 9.94328976e-01 -4.57738727e-01 1.94878593e-01
4.37784702e-01 3.39242011e-01 1.48205519e-01 -5.13979435e-01
5.20708203e-01 8.04911613e-01 3.05091709e-01 -1.87020615e-01
2.35967398e-01 1.15218498e-01 -3.66282493e-01 -1.07911682e+00
-7.20803559e-01 -4.60070461e-01 -8.45777810e-01 4.02394742e-01
1.15638053e+00 -1.06260443e+00 -7.35185146e-01 4.92643207e-01
-6.93004608e-01 -3.23475540e-01 5.29353440e-01 6.40438557e-01
-2.25268364e-01 6.95365250e-01 -5.77191293e-01 -5.75205386e-01
-4.07632351e-01 -6.74326420e-01 1.14222932e+00 6.01479150e-02
-2.03759834e-01 -1.10178673e+00 7.74002075e-01 6.18461668e-01
3.96298505e-02 -4.11101654e-02 1.23740780e+00 -8.87715995e-01
-1.35444701e-01 -1.29869059e-01 -1.86332583e-01 1.80535614e-01
9.77391377e-02 -1.20356478e-01 -8.74810696e-01 5.25098331e-02
-6.22822762e-01 -4.24095154e-01 1.30140102e+00 4.46215034e-01
1.19780815e+00 7.42183924e-02 -4.29323912e-01 8.85486782e-01
1.33598959e+00 3.29494365e-02 1.04455233e+00 -4.75580357e-02
8.03963661e-01 2.46476248e-01 4.51149851e-01 4.62283581e-01
5.17606139e-01 5.57388842e-01 1.62449032e-01 -1.38453007e-01
-4.10231315e-02 -1.90165445e-01 2.50205606e-01 7.49580383e-01
-2.16305375e-01 -5.41326702e-01 -1.06420362e+00 4.26305056e-01
-1.66337550e+00 -9.35322821e-01 -3.66276652e-01 2.04446697e+00
9.94231880e-01 -5.08324683e-01 1.13335997e-01 -4.47581202e-01
8.71953726e-01 1.49811834e-01 -7.27093697e-01 -2.78467029e-01
-2.15431452e-01 4.67497259e-01 3.62092972e-01 2.83668399e-01
-1.29499590e+00 9.46980476e-01 7.07376719e+00 3.28893661e-01
-6.99971914e-01 2.63289899e-01 7.03546703e-01 1.23619616e-01
-7.22689450e-01 -4.06425565e-01 -9.65942204e-01 5.68033397e-01
1.26774740e+00 -1.37649894e-01 -2.21235380e-01 5.73147237e-01
8.32853690e-02 3.51044744e-01 -8.47716987e-01 6.74867153e-01
2.10867003e-01 -1.28042448e+00 -1.81755543e-01 8.82017985e-02
5.36989689e-01 4.72967029e-01 6.23407811e-02 2.24776104e-01
5.72315276e-01 -1.06539357e+00 -4.04883146e-01 4.37439442e-01
7.16951609e-01 -7.60089695e-01 1.03018582e+00 -1.79168299e-01
-6.77752018e-01 2.11081266e-01 -1.03388691e+00 3.14121157e-01
1.44503415e-01 1.21578455e+00 -9.10728693e-01 7.36789405e-01
5.69415450e-01 8.56680214e-01 -6.02100670e-01 9.73564327e-01
-2.67987818e-01 9.90014970e-01 1.13515466e-01 -1.32433236e-01
-1.67362332e-01 -2.67041832e-01 1.08455203e-01 1.33193362e+00
3.67170542e-01 1.65560365e-01 -2.52515730e-02 7.86503971e-01
-1.56264342e-02 4.23400283e-01 -1.64217904e-01 -4.58691686e-01
2.31275514e-01 9.20034587e-01 -1.96731463e-01 -2.74461836e-01
-7.88021863e-01 9.10449028e-01 6.19266450e-01 1.35811746e-01
-9.20302570e-01 -4.32861477e-01 1.07319295e+00 -4.41560149e-01
5.92428446e-01 -1.22622540e-02 -7.88734406e-02 -1.26784861e+00
-3.52579594e-01 -9.57841396e-01 6.32550061e-01 -3.59933048e-01
-1.48093224e+00 1.15205973e-01 -6.06022298e-01 -6.57932520e-01
-1.60364583e-01 -5.74655294e-01 -5.90389252e-01 8.94104064e-01
-1.40618944e+00 -1.04770029e+00 -7.30101466e-02 1.92526113e-05
2.72000104e-01 -2.48323917e-01 1.18646777e+00 2.96421558e-01
-1.01141131e+00 9.26138401e-01 4.62528557e-01 2.39174962e-01
9.15878773e-01 -1.31737888e+00 1.16038978e+00 5.16986609e-01
1.14597537e-01 9.06879187e-01 3.28460693e-01 -1.18166947e+00
-1.45792878e+00 -1.51299465e+00 9.52092648e-01 -5.47803104e-01
6.18244052e-01 -5.41982591e-01 -9.76414382e-01 8.02793682e-01
1.90273300e-01 -4.04926896e-01 1.62895727e+00 7.54020512e-01
-6.27354085e-01 2.24207401e-01 -8.24771702e-01 4.84689027e-01
1.17690361e+00 -3.49571377e-01 -5.75666368e-01 5.11756241e-01
7.38975465e-01 1.73566699e-01 -1.14039624e+00 2.20353320e-01
9.66467381e-01 -8.62119973e-01 1.19266093e+00 -1.04716194e+00
6.65564001e-01 5.58613427e-02 -2.31007803e-02 -1.48849440e+00
-7.26492226e-01 -1.38921037e-01 3.09289068e-01 1.05049527e+00
7.71748841e-01 -7.96645343e-01 8.45954716e-01 1.92078471e-01
-2.04932094e-01 -3.66597027e-01 -5.61661541e-01 -7.09306419e-01
1.90219328e-01 -3.35796118e-01 6.93101823e-01 1.16069746e+00
1.51948199e-01 1.45693034e-01 -7.06430614e-01 3.06513280e-01
6.95137799e-01 -6.02379777e-02 5.89304149e-01 -1.24005651e+00
-5.42503178e-01 -1.19721830e-01 -7.81579494e-01 -5.92733741e-01
1.94359973e-01 -1.08405960e+00 1.39583573e-01 -1.40261447e+00
8.21459591e-01 3.02677099e-02 -5.39274156e-01 6.46445453e-01
-9.24832523e-01 3.18378896e-01 -5.87600231e-01 -4.33050722e-01
-1.98330924e-01 4.33412224e-01 7.49308527e-01 -2.04883531e-01
-7.67078921e-02 -3.50420326e-01 -1.07670319e+00 2.82126456e-01
1.19249022e+00 -2.15435103e-01 -3.91469032e-01 -6.76695049e-01
3.20852488e-01 -2.62976319e-01 3.74333501e-01 -7.24478245e-01
-1.99092343e-01 5.35553508e-02 6.24203980e-01 -4.21857506e-01
9.66205001e-02 1.48957327e-01 1.15477324e-01 7.31797934e-01
-6.24103606e-01 -3.19232881e-01 6.23383112e-02 9.03825462e-01
4.73924577e-02 -2.42537975e-01 4.13692832e-01 -1.66544214e-01
-7.18324661e-01 4.06108290e-01 -3.70251656e-01 2.05207262e-02
7.06259549e-01 3.36929709e-01 -8.12177956e-01 -2.47342303e-01
-8.98571074e-01 3.57132763e-01 1.79522231e-01 2.98200399e-01
5.69060445e-01 -9.53193605e-01 -1.32332146e+00 4.00068849e-01
2.66804963e-01 -7.59325683e-01 8.85770142e-01 7.57835209e-01
-5.29727876e-01 7.60718107e-01 -2.83822566e-01 -2.51849681e-01
-1.36330140e+00 6.65398777e-01 -7.66452327e-02 1.75385296e-01
-6.46764219e-01 1.17035496e+00 4.60534781e-01 -6.25967681e-01
4.86941822e-02 8.75213444e-02 -5.62751055e-01 -9.38233510e-02
9.22302067e-01 1.38334468e-01 -1.71139225e-01 -3.32866818e-01
-4.73754048e-01 5.14580131e-01 -3.35062265e-01 2.42692783e-01
1.55589962e+00 1.71617210e-01 -3.48732799e-01 2.73856163e-01
1.26798105e+00 -2.41968334e-02 -5.49945056e-01 7.80811757e-02
1.14193171e-01 -1.75368309e-01 -7.76211694e-02 -8.42794418e-01
-9.92147863e-01 8.66571248e-01 7.62698889e-01 -4.30904239e-01
6.13306344e-01 1.55111179e-01 9.10609066e-01 3.42386812e-01
2.25630403e-01 -7.48278439e-01 -4.52255011e-01 5.10608673e-01
3.89507532e-01 -1.22120655e+00 -9.01196804e-03 -6.59476697e-01
-6.69737160e-01 8.03530276e-01 4.32691187e-01 1.74322918e-01
6.17379367e-01 1.01215400e-01 4.63278860e-01 -1.70693576e-01
-1.04328465e+00 -4.58241850e-01 3.81028175e-01 1.10948658e+00
1.13280880e+00 4.09819216e-01 -6.83883846e-01 9.29748595e-01
-3.30826432e-01 1.36229083e-01 4.59097862e-01 5.96922100e-01
-4.47785705e-01 -1.76226866e+00 5.27697690e-02 8.51349294e-01
-7.63994873e-01 -6.48625851e-01 -6.15778744e-01 3.26446354e-01
1.45912096e-01 7.99495161e-01 4.76239510e-02 -4.68700290e-01
-2.02305034e-01 4.49686795e-01 3.17704082e-01 -8.42501223e-01
-1.53583154e-01 3.98186535e-01 5.42154789e-01 -5.67426622e-01
-2.69142419e-01 -8.71285915e-01 -9.48410034e-01 1.35344341e-01
-5.43960743e-02 2.61945337e-01 5.21338165e-01 5.15843689e-01
7.80885875e-01 7.20068157e-01 2.80301630e-01 -8.17163289e-02
-3.11590880e-01 -9.41933632e-01 -5.91871321e-01 5.59250377e-02
-1.32156983e-01 -2.51803994e-01 2.95350671e-01 3.90930474e-02] | [6.163951873779297, 5.706696510314941] |
a9130ce0-9db8-4527-893e-a84b84079f92 | autonovel-automatically-discovering-and | 2106.15252 | null | https://arxiv.org/abs/2106.15252v1 | https://arxiv.org/pdf/2106.15252v1.pdf | AutoNovel: Automatically Discovering and Learning Novel Visual Categories | We tackle the problem of discovering novel classes in an image collection given labelled examples of other classes. We present a new approach called AutoNovel to address this problem by combining three ideas: (1) we suggest that the common approach of bootstrapping an image representation using the labelled data only introduces an unwanted bias, and that this can be avoided by using self-supervised learning to train the representation from scratch on the union of labelled and unlabelled data; (2) we use ranking statistics to transfer the model's knowledge of the labelled classes to the problem of clustering the unlabelled images; and, (3) we train the data representation by optimizing a joint objective function on the labelled and unlabelled subsets of the data, improving both the supervised classification of the labelled data, and the clustering of the unlabelled data. Moreover, we propose a method to estimate the number of classes for the case where the number of new categories is not known a priori. We evaluate AutoNovel on standard classification benchmarks and substantially outperform current methods for novel category discovery. In addition, we also show that AutoNovel can be used for fully unsupervised image clustering, achieving promising results. | ['Andrew Zisserman', 'Andrea Vedaldi', 'Sébastien Ehrhardt', 'Sylvestre-Alvise Rebuffi', 'Kai Han'] | 2021-06-29 | null | null | null | null | ['image-clustering', 'novel-class-discovery', 'novel-class-discovery'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 5.13894737e-01 3.33655596e-01 -6.31721169e-02 -4.96220171e-01
-7.82213807e-01 -6.24083698e-01 7.28390574e-01 4.19553727e-01
-5.26784539e-01 8.11434209e-01 -2.97128428e-02 -4.04010080e-02
-3.04750055e-01 -5.91620922e-01 -7.84698725e-01 -9.46402669e-01
9.23832506e-02 9.80585515e-01 3.67510796e-01 2.92932749e-01
3.45181912e-01 4.36690688e-01 -2.17478967e+00 2.78623879e-01
6.62660837e-01 9.77274776e-01 1.71599835e-01 6.18283927e-01
-2.07599588e-02 6.24394417e-01 -5.14142454e-01 -1.51812196e-01
2.67297596e-01 -6.61261261e-01 -1.32525337e+00 7.69906938e-01
2.51826555e-01 2.64981568e-01 3.59582961e-01 1.05345345e+00
1.01878062e-01 2.04127148e-01 1.10633790e+00 -1.01780927e+00
-1.13910429e-01 4.14126515e-01 -5.19446492e-01 -1.47414356e-01
-1.54751912e-01 -1.02400340e-01 1.02297986e+00 -7.21868515e-01
6.46950543e-01 1.03786230e+00 6.35273814e-01 3.91919345e-01
-1.56031203e+00 -3.47566605e-01 6.76364750e-02 3.22570354e-01
-1.55260348e+00 -5.57868361e-01 6.02222085e-01 -7.22808123e-01
3.42825919e-01 3.81093383e-01 2.62009501e-01 5.22525609e-01
-4.95668054e-01 6.16657376e-01 1.26618445e+00 -8.18409979e-01
5.95508873e-01 5.48974991e-01 5.09847403e-01 4.44186211e-01
6.56526685e-02 1.39888197e-01 -2.95420643e-02 -2.32418239e-01
1.97974980e-01 -3.00085396e-01 1.00319227e-02 -8.49477053e-01
-1.22172642e+00 9.01152790e-01 3.76429200e-01 4.95672047e-01
-3.29924822e-01 -1.91401482e-01 5.07444441e-01 4.19380851e-02
6.13731146e-01 6.37373328e-01 -5.58794916e-01 3.32659990e-01
-9.17826653e-01 -1.47372246e-01 6.34531438e-01 6.49476945e-01
1.18281627e+00 -3.70592892e-01 2.18943015e-01 1.01149440e+00
3.68520804e-02 1.67890608e-01 7.63678551e-01 -9.22688305e-01
-1.45714104e-01 5.78905523e-01 -8.08889046e-03 -5.34212410e-01
-3.58657628e-01 -4.55892146e-01 -7.05472231e-01 2.38905594e-01
3.85659188e-01 1.04936168e-01 -1.19569242e+00 1.52760577e+00
3.55663598e-01 1.23443671e-01 1.01535343e-01 2.98620999e-01
5.76928735e-01 3.88317943e-01 -6.78808391e-02 -5.55779636e-01
9.31956470e-01 -8.03118348e-01 -3.36017936e-01 -1.17095605e-01
6.16348207e-01 -5.34494519e-01 7.44235396e-01 5.43607533e-01
-6.97271466e-01 -8.03514004e-01 -8.69283557e-01 4.80915219e-01
-5.02248108e-01 3.77033561e-01 5.04704475e-01 5.48758805e-01
-1.08382809e+00 6.62600577e-01 -5.53786278e-01 -5.24681985e-01
4.78349447e-01 5.33574343e-01 -4.26258296e-01 -1.28747523e-01
-5.43493927e-01 6.99084103e-01 1.21686625e+00 -1.47674337e-01
-8.42734456e-01 -2.04042196e-02 -6.11480296e-01 -9.65822712e-02
5.27892649e-01 -3.46363097e-01 1.03939092e+00 -1.36884034e+00
-1.14401186e+00 1.07060373e+00 -1.93315014e-01 -3.21619689e-01
2.33884051e-01 3.70845050e-01 -1.49657220e-01 2.48813957e-01
3.38079304e-01 8.18648696e-01 8.66933048e-01 -1.83083320e+00
-9.71214056e-01 -3.48432988e-01 -4.01181996e-01 8.24043602e-02
-2.53047466e-01 -3.56369287e-01 -3.27139378e-01 -3.43461454e-01
3.59129727e-01 -9.11921322e-01 -3.59065533e-01 -4.08272594e-01
-3.39366317e-01 -5.09850025e-01 7.41969943e-01 -2.28683293e-01
7.31609046e-01 -2.39041209e+00 1.77504048e-01 4.58179414e-01
2.36376569e-01 3.04035962e-01 -9.02230218e-02 6.17681667e-02
-3.67623627e-01 4.33256477e-02 -4.59998250e-01 -2.15331376e-01
-3.24965507e-01 5.81495643e-01 -2.36897722e-01 3.33747476e-01
1.83243141e-01 6.13607049e-01 -1.05621231e+00 -5.52294850e-01
3.55986297e-01 -2.48406269e-02 -4.56192791e-01 9.53139737e-02
-2.06709236e-01 6.63956463e-01 4.34419662e-02 1.39329806e-01
5.13523221e-01 -3.99274349e-01 4.17126209e-01 3.08991093e-02
2.00342506e-01 1.14711277e-01 -1.41978061e+00 1.13229907e+00
-2.24074155e-01 3.32019240e-01 -3.08603883e-01 -1.51927924e+00
8.91032100e-01 1.93821922e-01 4.84302223e-01 -8.86143222e-02
3.95571217e-02 2.71694839e-01 -1.58048332e-01 -4.33222502e-01
1.55896842e-01 -4.85868543e-01 7.81564787e-02 4.99987364e-01
5.62090516e-01 -1.19502723e-01 3.62227619e-01 3.57247144e-02
8.17999959e-01 -8.76970962e-02 4.61979121e-01 -2.50974774e-01
3.93161714e-01 6.07735179e-02 3.76297653e-01 9.56574500e-01
2.44502947e-01 5.04469812e-01 4.13320243e-01 -4.11450475e-01
-9.64365602e-01 -1.09199691e+00 -4.52981144e-01 1.02286160e+00
-7.95867816e-02 -2.06780970e-01 -6.71404481e-01 -1.16976607e+00
-2.16207713e-01 7.53056586e-01 -7.83035755e-01 -1.37173533e-01
-7.19148666e-02 -8.84378374e-01 7.78618976e-02 3.54518592e-01
2.57826596e-01 -1.00547743e+00 -2.62631685e-01 1.00654915e-01
-4.17204738e-01 -8.92445564e-01 6.40416369e-02 6.47152781e-01
-1.00032139e+00 -1.34325850e+00 -3.42689216e-01 -1.11638975e+00
1.11137795e+00 2.28377998e-01 1.10176146e+00 1.50291741e-01
-1.12436414e-01 3.27577055e-01 -6.69086754e-01 -2.56486028e-01
-6.57317936e-01 -1.19930329e-02 5.72868027e-02 2.95215786e-01
3.87442589e-01 -4.38540876e-01 -3.25513706e-02 5.79818666e-01
-9.30884063e-01 -1.90101922e-01 5.79427898e-01 1.00334775e+00
6.45272136e-01 8.25973034e-01 7.92254806e-01 -1.16134632e+00
7.41987675e-02 -4.17126983e-01 -3.90195936e-01 2.70185143e-01
-6.03339255e-01 2.03949228e-01 4.94767785e-01 -5.14820874e-01
-7.17161477e-01 5.39452791e-01 2.99981423e-02 -2.76757479e-01
-5.54886818e-01 4.23382252e-01 -1.58368081e-01 5.54149263e-02
8.71018708e-01 1.93326980e-01 1.31061137e-01 -6.07262790e-01
5.26051939e-01 9.24945533e-01 7.29685903e-01 -4.44364041e-01
6.67228520e-01 5.86870790e-01 -6.27661571e-02 -9.40573037e-01
-1.07776511e+00 -9.61777568e-01 -1.26542270e+00 -1.67652607e-01
7.19026983e-01 -7.09351003e-01 -3.53270441e-01 2.40334660e-01
-8.46563995e-01 -1.10014863e-01 -8.25856984e-01 5.51791310e-01
-6.89278781e-01 7.06509650e-01 5.18097915e-02 -8.46378684e-01
2.95610070e-01 -1.01444328e+00 9.92672741e-01 -1.78431764e-01
-3.55864435e-01 -1.00067413e+00 5.78852147e-02 4.35089231e-01
-1.90339178e-01 2.37314925e-01 1.07231867e+00 -1.19573355e+00
-2.51608402e-01 -5.07218421e-01 -1.97404280e-01 8.75699282e-01
3.34551215e-01 -2.37935588e-01 -1.15538538e+00 -3.57945681e-01
1.03382371e-01 -4.85219657e-01 1.13311708e+00 2.80243516e-01
1.24755812e+00 -2.94331849e-01 -5.54585159e-01 7.82799870e-02
1.16862512e+00 1.51522368e-01 4.18467820e-01 9.78334770e-02
5.41551471e-01 9.32273686e-01 5.01048326e-01 1.05980493e-01
1.88112289e-01 6.04947448e-01 2.81427652e-01 -9.79231521e-02
-7.97230378e-03 -9.19837411e-03 -1.03980690e-01 8.20925057e-01
5.16275726e-02 1.98843814e-02 -8.14649463e-01 7.42362201e-01
-2.05864429e+00 -7.25336671e-01 -1.27487183e-01 2.62685537e+00
8.58507276e-01 1.86372846e-01 4.02383208e-01 4.66308624e-01
9.45196331e-01 -4.77934927e-01 -3.31130475e-01 -1.12379283e-01
2.58074980e-02 3.60338569e-01 4.45115477e-01 4.28522438e-01
-1.46995568e+00 6.71202064e-01 6.77146101e+00 9.04150546e-01
-6.02327168e-01 3.64244580e-02 7.22213805e-01 2.50910878e-01
8.20308253e-02 3.37687582e-01 -6.40500963e-01 3.83261114e-01
8.22018564e-01 4.75453632e-03 1.31288096e-01 8.08929324e-01
-1.77088350e-01 -3.58466119e-01 -1.31835544e+00 7.29561985e-01
2.76251823e-01 -9.23074663e-01 2.07573056e-01 3.31580728e-01
8.87095213e-01 -2.08007425e-01 -2.20527962e-01 3.07857573e-01
5.90842664e-01 -8.40386212e-01 4.08304602e-01 4.17411327e-01
6.15222275e-01 -8.67872417e-01 9.14144456e-01 6.29717171e-01
-6.68339074e-01 -2.03600228e-01 -4.44081932e-01 -3.24989259e-02
-3.66109788e-01 7.50670373e-01 -1.17014134e+00 6.01278365e-01
5.86206496e-01 7.46242762e-01 -1.02687299e+00 1.40106511e+00
-2.69859433e-01 7.12499559e-01 -3.12468559e-01 2.20064297e-01
2.62022465e-01 -3.34655084e-02 2.33958110e-01 8.81736934e-01
-6.36350587e-02 -2.31211469e-01 5.15426576e-01 3.55732352e-01
2.82774977e-02 9.10769030e-02 -6.07063115e-01 1.45754099e-01
8.56690928e-02 1.17762554e+00 -1.26585591e+00 -5.86679637e-01
5.25576249e-02 8.31419706e-01 3.41113746e-01 1.77526996e-01
-1.76714182e-01 -3.52479964e-01 -1.52837917e-01 -4.11500745e-02
6.37163162e-01 1.29670888e-01 -1.33191705e-01 -1.03330731e+00
-8.50051492e-02 -6.59920633e-01 5.71509421e-01 -6.08221114e-01
-1.43160141e+00 5.76423943e-01 2.48938903e-01 -1.26314497e+00
-4.82211798e-01 -5.39236665e-01 -2.51621276e-01 4.55467790e-01
-1.20363796e+00 -8.32431138e-01 6.28168732e-02 2.60160774e-01
3.73504758e-01 -1.27494514e-01 9.71490979e-01 3.52491662e-02
-1.45772398e-01 1.55211732e-01 5.30116916e-01 1.83551118e-01
6.48743629e-01 -1.39382267e+00 2.41226181e-02 7.08831370e-01
4.74098623e-01 3.25906605e-01 6.81126893e-01 -4.30156559e-01
-4.05671090e-01 -1.25310719e+00 8.65902424e-01 -6.44155741e-01
3.85010362e-01 -4.27571028e-01 -1.00219452e+00 6.07103348e-01
-1.29501522e-01 1.08834310e-02 9.38640356e-01 3.23931426e-01
-3.71588141e-01 7.61430785e-02 -1.17227781e+00 6.18597232e-02
5.76757669e-01 -2.39783302e-01 -6.97818339e-01 6.87081635e-01
3.92026991e-01 2.39861399e-01 -5.97431958e-01 4.71185744e-01
2.54022658e-01 -7.90737450e-01 8.21749449e-01 -5.86183310e-01
2.45307535e-01 -5.29311240e-01 -2.22457424e-01 -1.42166030e+00
-2.98767716e-01 -1.24901488e-01 2.79276758e-01 1.23610890e+00
3.46881330e-01 -4.15715516e-01 7.68088877e-01 8.05105641e-02
2.36830994e-01 -4.92636859e-01 -9.82217610e-01 -9.45447445e-01
3.56469192e-02 -2.21149966e-01 2.53538877e-01 1.02592814e+00
-1.92458168e-01 7.60872960e-01 -3.10340375e-01 -2.17041913e-02
8.54638934e-01 1.48152173e-01 8.16191077e-01 -1.78251243e+00
-4.10206318e-01 -2.78005481e-01 -8.35665226e-01 -7.02730000e-01
2.68937647e-01 -1.09074020e+00 4.48088109e-01 -1.34472322e+00
6.03258312e-01 -4.51325446e-01 -3.45268875e-01 6.34250700e-01
-1.57886133e-01 5.22418380e-01 2.58100126e-02 3.97419900e-01
-9.91020560e-01 3.03246707e-01 7.71872640e-01 -6.10350184e-02
-1.37097359e-01 2.83684999e-01 -7.48428524e-01 9.55495775e-01
5.54288566e-01 -6.61339879e-01 -3.66337061e-01 1.42588556e-01
-1.07704781e-01 -3.87722492e-01 4.52192664e-01 -1.06823921e+00
7.80926347e-02 6.61024749e-02 4.71354902e-01 -6.21926367e-01
1.95793193e-02 -9.83681500e-01 3.72209474e-02 3.23003232e-01
-5.02914906e-01 -6.34078979e-01 -1.92929078e-02 7.47031510e-01
-1.41652822e-01 -7.62081206e-01 1.14193749e+00 -3.23183239e-01
-7.46173620e-01 -7.74927065e-02 -4.14199352e-01 -6.77044690e-02
1.05001390e+00 -3.60680848e-01 7.85602182e-02 -2.27653727e-01
-1.19212401e+00 2.66403675e-01 5.49970627e-01 1.64819568e-01
4.21446264e-01 -1.21318400e+00 -6.30950332e-01 3.44839483e-01
4.61172223e-01 1.61244631e-01 -1.68586627e-01 5.86641908e-01
-1.37423262e-01 1.96057156e-01 1.24426283e-01 -9.68175471e-01
-1.43065119e+00 8.37262094e-01 1.22872278e-01 -2.80596465e-01
-3.14179182e-01 6.34747624e-01 3.51972669e-01 -8.81736159e-01
2.29155362e-01 2.91476190e-01 -4.23438936e-01 3.41438890e-01
2.84378886e-01 2.21857056e-01 3.32091480e-01 -7.75801122e-01
-1.95132643e-01 4.10488904e-01 -4.21666056e-01 1.18281813e-02
1.45620489e+00 -1.47386998e-01 -3.91682923e-01 8.56596708e-01
1.41656756e+00 -3.61613482e-01 -9.48712766e-01 -5.38801074e-01
4.05116618e-01 -4.18551177e-01 -1.06596604e-01 -8.54502261e-01
-6.22964025e-01 6.63908601e-01 7.30360210e-01 2.92044967e-01
1.19871640e+00 4.12621617e-01 8.44287276e-02 5.87961137e-01
3.89486879e-01 -1.15493083e+00 1.69320747e-01 4.17653263e-01
4.50028926e-01 -1.30101705e+00 1.06671289e-01 -2.59830177e-01
-6.68588996e-01 1.04710865e+00 6.79323971e-02 -1.65019953e-03
6.89695060e-01 -2.54375517e-01 3.53154689e-02 -3.01458210e-01
-6.36664331e-01 -5.99228203e-01 4.90249038e-01 8.04516494e-01
6.71480522e-02 3.30599472e-02 -1.07117392e-01 1.51761502e-01
-6.12798817e-02 -1.91949122e-02 5.01880348e-01 8.99482310e-01
-6.81685865e-01 -1.29012012e+00 -3.44903767e-01 8.02098274e-01
-2.67274141e-01 1.40885085e-01 -7.73385286e-01 5.89934707e-01
4.95910257e-01 1.08992827e+00 8.68873019e-03 -3.46960068e-01
9.49363410e-02 3.06541026e-01 3.16327095e-01 -1.17642558e+00
-1.76146671e-01 2.86334157e-01 -1.07015772e-02 2.56747007e-02
-9.05132234e-01 -6.66168094e-01 -7.61369824e-01 3.64211231e-01
-8.69761169e-01 7.42772341e-01 6.35750651e-01 1.13872540e+00
1.29477968e-02 3.13944489e-01 1.10467553e+00 -9.62190390e-01
-4.75157857e-01 -8.33931923e-01 -7.53903627e-01 5.97754300e-01
2.90837616e-01 -7.63410509e-01 -6.65134311e-01 3.72757286e-01] | [9.50295639038086, 3.0110929012298584] |
68e13bd1-71db-4527-b941-c2100c5412c1 | improving-depression-estimation-from-facial | 2212.06400 | null | https://arxiv.org/abs/2212.06400v1 | https://arxiv.org/pdf/2212.06400v1.pdf | Improving Depression estimation from facial videos with face alignment, training optimization and scheduling | Deep learning models have shown promising results in recognizing depressive states using video-based facial expressions. While successful models typically leverage using 3D-CNNs or video distillation techniques, the different use of pretraining, data augmentation, preprocessing, and optimization techniques across experiments makes it difficult to make fair architectural comparisons. We propose instead to enhance two simple models based on ResNet-50 that use only static spatial information by using two specific face alignment methods and improved data augmentation, optimization, and scheduling techniques. Our extensive experiments on benchmark datasets obtain similar results to sophisticated spatio-temporal models for single streams, while the score-level fusion of two different streams outperforms state-of-the-art methods. Our findings suggest that specific modifications in the preprocessing and training process result in noticeable differences in the performance of the models and could hide the actual originally attributed to the use of different neural network architectures. | ['Miguel Bordallo López', 'Le Nguyen', 'Constantino Álvarez Casado', 'Manuel Lage Cañellas'] | 2022-12-13 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [ 1.01560920e-01 -2.08340243e-01 -2.77649701e-01 -7.99865544e-01
-1.99473947e-01 -1.46015093e-01 8.94204497e-01 2.16356441e-02
-7.09738851e-01 3.63975853e-01 3.79280686e-01 -4.39059660e-02
-1.13500878e-02 -4.45642531e-01 -5.03129840e-01 -4.19245213e-01
-5.20493686e-01 3.45892549e-01 -1.05421074e-01 -4.27401572e-01
-1.19019032e-01 8.75741780e-01 -1.50198960e+00 4.52518314e-01
-1.00844070e-01 1.12803102e+00 -7.15091109e-01 5.96565306e-01
-1.51406810e-01 8.21401238e-01 -7.76923478e-01 -2.43493035e-01
4.34089094e-01 -4.31340903e-01 -4.27292466e-01 4.20989171e-02
8.24697673e-01 -7.28000760e-01 -4.04464215e-01 4.51154053e-01
8.95483673e-01 1.84320912e-01 3.55722576e-01 -1.37204766e+00
-3.90775830e-01 1.94263816e-01 -6.25978112e-01 3.95472765e-01
4.46607530e-01 1.28855288e-01 3.11768889e-01 -6.11069679e-01
7.15571225e-01 1.28203976e+00 1.00298953e+00 7.64515996e-01
-1.32680392e+00 -9.75876093e-01 1.30270049e-01 1.48999795e-01
-1.41387975e+00 -1.00150478e+00 6.50624216e-01 -2.90320963e-01
1.31372297e+00 -8.92084315e-02 9.27438080e-01 1.38712943e+00
8.96439329e-02 4.24113870e-01 1.03136563e+00 -1.57264322e-01
2.34488979e-01 -1.61883011e-02 -1.51910216e-01 7.94729233e-01
-1.03616841e-01 1.19970255e-01 -7.81263649e-01 -2.19981611e-01
7.27427721e-01 8.45009089e-03 8.08676630e-02 -4.44696486e-01
-7.36620426e-01 7.63445914e-01 3.46725136e-01 4.32581544e-01
-5.64907610e-01 2.69171774e-01 7.16431499e-01 4.07845527e-01
7.07657993e-01 2.98309654e-01 -4.74014908e-01 -3.74563724e-01
-1.48559916e+00 3.92402858e-01 4.69205409e-01 2.87141263e-01
4.51200038e-01 4.64189976e-01 -1.83615908e-01 6.46691322e-01
-3.40509824e-02 5.37844598e-02 4.89782572e-01 -1.14205921e+00
7.25579858e-02 4.38888550e-01 -2.65689701e-01 -1.21904659e+00
-8.92624736e-01 -4.62259054e-01 -6.20582521e-01 3.62991005e-01
2.74311900e-01 -5.38985550e-01 -1.19266725e+00 1.92506003e+00
1.68440670e-01 5.97359061e-01 -2.48132199e-01 7.86516011e-01
7.72883058e-01 1.00097761e-01 2.78740436e-01 -1.19732479e-02
8.85816336e-01 -8.00251603e-01 -5.38522959e-01 -9.20400321e-02
8.53761971e-01 -4.60463285e-01 5.95503211e-01 2.59936839e-01
-1.25622427e+00 -5.55935144e-01 -1.06602371e+00 1.15831442e-01
-4.18665528e-01 2.01718044e-02 7.81351209e-01 8.40686500e-01
-1.48627341e+00 6.92650139e-01 -9.43494141e-01 -6.82009161e-01
6.91954434e-01 7.54115522e-01 -6.82200730e-01 1.62567645e-01
-7.93096364e-01 9.56081271e-01 -7.23728985e-02 4.42303903e-02
-7.94987261e-01 -8.41269791e-01 -9.25952077e-01 1.15586780e-02
-9.12934169e-02 -5.58997810e-01 1.15210128e+00 -1.51014960e+00
-1.53468871e+00 9.03419256e-01 -1.69650450e-01 -5.83288670e-01
4.00286019e-01 -1.17376022e-01 -3.93774927e-01 2.01062039e-01
-3.28582585e-01 1.18520868e+00 7.52616823e-01 -7.40076184e-01
-2.18514398e-01 -5.24214506e-01 3.70662361e-02 2.28282869e-01
-6.48850858e-01 3.58023137e-01 -3.91052574e-01 -5.40256441e-01
-2.44106591e-01 -7.28305221e-01 -3.55898440e-01 1.06551200e-01
2.92461812e-01 1.44479066e-01 8.74322534e-01 -5.23193598e-01
1.07608545e+00 -1.99049628e+00 -3.49676870e-02 2.22903028e-01
8.65529329e-02 4.49764073e-01 -7.42309988e-01 2.32519303e-02
-5.65863729e-01 2.28026062e-01 2.06159502e-01 -6.28684759e-01
-2.03582957e-01 7.77166039e-02 9.10752788e-02 6.97953761e-01
3.47490370e-01 7.43686259e-01 -4.47631925e-01 -3.52656990e-01
3.18694979e-01 8.05328548e-01 -8.67631614e-01 -1.02151580e-01
-3.89262699e-02 3.67591053e-01 1.82947218e-02 7.24021792e-01
6.05522215e-01 1.28517672e-03 1.71017706e-01 -4.14431393e-02
1.93196788e-01 2.82530963e-01 -8.28361034e-01 1.91303790e+00
-2.66249031e-01 8.34698558e-01 1.01714790e-01 -1.18675804e+00
6.25220895e-01 4.74784553e-01 1.09627068e+00 -1.15457106e+00
3.44474912e-01 -3.49657506e-01 7.30935186e-02 -3.75005633e-01
4.42997456e-01 -5.50400838e-02 2.31526256e-01 3.33235145e-01
3.74706030e-01 2.44748712e-01 4.15706970e-02 -7.93639421e-02
1.29311371e+00 3.16759259e-01 -1.44145578e-01 7.64363969e-04
6.05294155e-03 -1.36828482e-01 5.98098993e-01 5.06990314e-01
-5.02795517e-01 6.51743710e-01 6.90016747e-01 -6.53249562e-01
-9.88054752e-01 -6.37455285e-01 -6.11998476e-02 1.34682548e+00
-4.32277232e-01 -6.24133229e-01 -7.68137932e-01 -5.41497588e-01
-7.79947415e-02 2.92295396e-01 -1.08841705e+00 -2.21819699e-01
-5.57836592e-01 -1.10114098e+00 1.07988191e+00 7.05144346e-01
4.01944876e-01 -8.39167416e-01 -7.59365082e-01 2.37600461e-01
2.65938878e-01 -1.34277105e+00 1.93624318e-01 1.37953430e-01
-9.17243302e-01 -7.78733850e-01 -7.27914572e-01 -3.55880678e-01
3.96325469e-01 3.03957723e-02 1.13596308e+00 2.64750361e-01
-1.96282327e-01 4.41287637e-01 -2.36817777e-01 -5.30121148e-01
3.32777388e-02 2.65052497e-01 2.55490005e-01 -2.00656056e-01
5.37629187e-01 -6.69741392e-01 -5.34643531e-01 7.94545223e-04
-7.66872406e-01 -3.94827090e-02 3.44144225e-01 5.86617589e-01
6.92918375e-02 -4.04423207e-01 4.20522034e-01 -4.91438210e-01
6.94442511e-01 -5.19566834e-01 -2.65912801e-01 -1.63678285e-02
-5.52475512e-01 -6.41523078e-02 2.85351574e-01 -5.10327458e-01
-7.50719607e-01 1.88332144e-02 -2.36219317e-01 -1.03540897e+00
-4.81374472e-01 3.90901715e-01 4.27817911e-01 -3.68228942e-01
7.38141596e-01 -9.95858237e-02 3.58141422e-01 -2.26294279e-01
5.20684123e-02 1.69047847e-01 1.78213432e-01 -3.27857763e-01
3.38518471e-01 5.36918700e-01 9.87603888e-02 -7.92036712e-01
-3.97962779e-01 -2.36344278e-01 -6.38568461e-01 -2.23673150e-01
9.09014285e-01 -1.01080465e+00 -6.28156126e-01 5.97122192e-01
-1.01175165e+00 -6.52844667e-01 -2.54697651e-01 3.65884155e-01
-3.25387657e-01 -4.48338017e-02 -6.44983590e-01 -4.91823167e-01
-3.00330400e-01 -1.00631213e+00 1.07390010e+00 2.32832327e-01
-4.74254191e-01 -1.06262410e+00 3.46589416e-01 2.21031353e-01
9.53187466e-01 3.94816041e-01 6.13072991e-01 -7.93366849e-01
-7.09033236e-02 -1.63550153e-01 -1.09946653e-01 9.06440020e-02
2.11310059e-01 3.35709244e-01 -1.13091409e+00 -2.23420277e-01
-3.24414670e-01 -5.70085049e-01 6.92317128e-01 5.86153448e-01
1.22524500e+00 -1.09348491e-01 -2.00683936e-01 9.19248939e-01
1.00856090e+00 9.84940305e-02 6.77947342e-01 3.66385013e-01
4.56770062e-01 5.17427862e-01 3.25754806e-02 5.82480550e-01
2.53195196e-01 9.65446770e-01 3.31128240e-01 -6.04759753e-01
-1.74499124e-01 2.29788631e-01 6.10506952e-01 2.55299836e-01
-2.94768721e-01 1.40184835e-01 -8.92672002e-01 4.47446138e-01
-1.64951777e+00 -1.20516253e+00 6.00596905e-01 1.83235037e+00
6.42106831e-01 -5.21800667e-02 3.72902215e-01 -1.57315299e-01
2.26986840e-01 4.97695923e-01 -4.70435619e-01 -7.81745195e-01
-1.31593049e-01 8.13950360e-01 4.65773761e-01 1.27472907e-01
-9.33996618e-01 9.51229572e-01 7.73503113e+00 4.97163147e-01
-1.90368450e+00 -1.99335758e-02 8.48341763e-01 -9.20503557e-01
1.44418120e-01 -4.51214671e-01 -5.44664145e-01 1.44273072e-01
1.37813938e+00 1.17681876e-01 4.81659323e-01 5.40850997e-01
5.34862220e-01 -9.18235108e-02 -1.08520031e+00 9.44149911e-01
3.84543151e-01 -1.43492889e+00 -6.69533089e-02 7.98386559e-02
5.76900899e-01 3.56843531e-01 2.59907633e-01 3.51541787e-01
1.44969925e-01 -1.34131956e+00 5.64887762e-01 5.03677070e-01
6.39351189e-01 -6.31960452e-01 5.98392487e-01 -2.01377422e-01
-8.79093170e-01 -7.16935024e-02 1.23210989e-01 -2.88030922e-01
-1.44313082e-01 1.13788113e-01 -9.32958364e-01 1.97378635e-01
8.32372725e-01 6.63264334e-01 -8.53842020e-01 7.19294965e-01
3.28208894e-01 6.56512380e-01 -5.01930177e-01 1.79754123e-01
2.91300416e-01 1.48872137e-01 1.48565173e-01 1.33833659e+00
1.55918911e-01 -1.38852194e-01 -9.17573646e-02 3.83451372e-01
-1.24046840e-01 1.62620619e-01 -7.32457876e-01 -5.84754907e-02
-9.76584107e-03 1.28997219e+00 -5.95456779e-01 -3.15006137e-01
-6.85811281e-01 7.80360579e-01 2.98092753e-01 4.74363476e-01
-1.06830537e+00 8.48774165e-02 1.21289432e+00 4.02354270e-01
2.53367513e-01 -4.04124558e-01 -3.57268065e-01 -9.96619105e-01
-2.24134222e-01 -1.11841524e+00 3.16254139e-01 -8.12669039e-01
-6.98733211e-01 5.61156571e-01 8.32780376e-02 -6.80454850e-01
-4.96339083e-01 -5.20190895e-01 -6.49711549e-01 5.17646015e-01
-1.47252262e+00 -9.97354925e-01 -3.06731969e-01 8.18021119e-01
3.30087036e-01 -3.08209002e-01 9.69869375e-01 4.51145202e-01
-8.43422413e-01 9.19476390e-01 -3.76315743e-01 4.68107611e-01
9.14173305e-01 -6.22448623e-01 2.65370727e-01 6.81872249e-01
1.13784194e-01 5.18350542e-01 4.34876561e-01 -2.95305341e-01
-1.01962161e+00 -9.73780632e-01 5.28925180e-01 -3.12618434e-01
4.17092264e-01 -3.07693809e-01 -6.51841223e-01 7.18089879e-01
5.83641171e-01 7.61666894e-02 7.95364380e-01 4.01278734e-01
-2.67525494e-01 -3.00102621e-01 -1.23431313e+00 6.63872004e-01
1.06513834e+00 -4.27085012e-01 -2.80788764e-02 1.48921952e-01
2.11566225e-01 -4.80331510e-01 -7.89946377e-01 6.59550607e-01
8.40050578e-01 -1.15127456e+00 9.12486851e-01 -1.13427055e+00
4.69448924e-01 2.01746404e-01 -8.05970058e-02 -1.19232190e+00
-2.84029067e-01 -2.44827434e-01 1.57327913e-02 9.35145378e-01
3.05796534e-01 -3.43015283e-01 1.19543970e+00 8.47785652e-01
1.56293601e-01 -9.44179416e-01 -8.03488255e-01 -3.62569958e-01
-3.62854339e-02 -4.25572872e-01 7.12928355e-01 1.22118902e+00
-1.83543265e-01 2.54881233e-02 -2.36749575e-01 -1.42195970e-01
4.73532379e-02 -4.39709574e-01 8.72693002e-01 -1.02997172e+00
6.13655448e-02 -5.92762351e-01 -7.69062161e-01 -2.99922377e-01
4.67691660e-01 -6.14893198e-01 -4.43926454e-01 -1.11172175e+00
1.02523401e-01 -3.61838311e-01 -4.02138352e-01 1.17483008e+00
2.93777794e-01 5.06217659e-01 1.91031978e-01 -2.03185394e-01
-5.43888092e-01 5.81979573e-01 7.14239061e-01 -1.08238701e-02
-3.89993727e-01 -3.38539928e-01 -6.42604530e-01 7.96885669e-01
9.50968623e-01 -3.09990138e-01 -4.35149550e-01 -7.61603236e-01
1.32749036e-01 -4.36477810e-02 5.09742379e-01 -1.39581275e+00
2.54312813e-01 -9.58259404e-02 9.49724078e-01 -3.71555127e-02
7.32990265e-01 -7.79198170e-01 8.07172135e-02 3.45454872e-01
-3.49957645e-01 5.61310768e-01 8.41244519e-01 4.95072752e-02
-1.27119422e-01 2.57893145e-01 7.52733767e-01 -5.85435890e-02
-7.08690047e-01 4.18593496e-01 -5.34333944e-01 -3.08994710e-01
9.95866597e-01 -4.90391940e-01 -1.17439993e-01 -6.59453571e-01
-8.66433859e-01 -1.92049310e-01 5.67048907e-01 6.34912610e-01
3.20549160e-01 -1.29194736e+00 -6.66074693e-01 3.90705734e-01
-1.52466804e-01 -4.96000141e-01 1.47242397e-01 8.99216235e-01
-4.76512194e-01 3.01147580e-01 -6.96536124e-01 -6.58370733e-01
-1.54620612e+00 1.03170201e-01 8.03875208e-01 -1.70166895e-01
-1.23379864e-01 7.12129116e-01 -3.22571397e-01 -4.41045076e-01
3.93869579e-01 -1.16537571e-01 -1.16121516e-01 3.24726522e-01
3.38453382e-01 2.44163424e-01 5.24655998e-01 -6.97645307e-01
-5.25346577e-01 3.97802174e-01 -1.64177373e-01 -3.23878914e-01
1.53167093e+00 1.69408768e-01 1.90872535e-01 7.33531499e-03
1.24910140e+00 -3.30416381e-01 -1.18321848e+00 6.47816136e-02
-3.08074772e-01 -1.58908471e-01 3.55575085e-01 -6.71227932e-01
-1.51889431e+00 8.90996397e-01 1.08926392e+00 -3.72568488e-01
1.47648251e+00 -4.31390673e-01 5.39784551e-01 2.98690468e-01
7.77105317e-02 -1.00697386e+00 1.01227596e-01 4.80878443e-01
5.29767573e-01 -1.00222361e+00 1.93279117e-01 1.76458836e-01
-4.61944073e-01 9.87640738e-01 8.88247550e-01 -4.76697795e-02
7.29473352e-01 6.05753481e-01 3.34089011e-01 -2.91794300e-01
-1.02162981e+00 -1.29486665e-01 2.96852235e-02 4.89596844e-01
7.69194603e-01 -3.89178574e-01 5.30531518e-02 2.24893942e-01
-1.54566139e-01 4.88858610e-01 1.80564031e-01 1.12601459e+00
1.01312116e-01 -1.16167033e+00 -8.68897438e-02 4.18532848e-01
-8.10903370e-01 1.40284188e-02 -4.32687223e-01 1.00358474e+00
3.00739557e-01 7.03352034e-01 6.16727352e-01 -5.77765644e-01
2.96537906e-01 2.79613465e-01 8.19456458e-01 -4.30932075e-01
-9.54250634e-01 -7.85531700e-02 1.80233449e-01 -9.90068316e-01
-8.97153378e-01 -8.04947853e-01 -1.05064178e+00 -6.23647749e-01
-6.85241213e-03 -3.20207566e-01 8.01161945e-01 9.88414764e-01
9.25925970e-01 5.34635603e-01 3.70064676e-01 -1.13601172e+00
-6.43411800e-02 -9.71163869e-01 -2.00088710e-01 4.76146519e-01
3.66087705e-01 -6.97187304e-01 1.27836019e-01 -4.49413154e-03] | [13.516631126403809, 1.6988434791564941] |
bbefb0ff-d30c-4a0e-95b7-fac73fc4aa9e | global-distance-distributions-separation-for | 2006.00752 | null | https://arxiv.org/abs/2006.00752v3 | https://arxiv.org/pdf/2006.00752v3.pdf | Global Distance-distributions Separation for Unsupervised Person Re-identification | Supervised person re-identification (ReID) often has poor scalability and usability in real-world deployments due to domain gaps and the lack of annotations for the target domain data. Unsupervised person ReID through domain adaptation is attractive yet challenging. Existing unsupervised ReID approaches often fail in correctly identifying the positive samples and negative samples through the distance-based matching/ranking. The two distributions of distances for positive sample pairs (Pos-distr) and negative sample pairs (Neg-distr) are often not well separated, having large overlap. To address this problem, we introduce a global distance-distributions separation (GDS) constraint over the two distributions to encourage the clear separation of positive and negative samples from a global view. We model the two global distance distributions as Gaussian distributions and push apart the two distributions while encouraging their sharpness in the unsupervised training process. Particularly, to model the distributions from a global view and facilitate the timely updating of the distributions and the GDS related losses, we leverage a momentum update mechanism for building and maintaining the distribution parameters (mean and variance) and calculate the loss on the fly during the training. Distribution-based hard mining is proposed to further promote the separation of the two distributions. We validate the effectiveness of the GDS constraint in unsupervised ReID networks. Extensive experiments on multiple ReID benchmark datasets show our method leads to significant improvement over the baselines and achieves the state-of-the-art performance. | ['Wen-Jun Zeng', 'Xin Jin', 'Zhibo Chen', 'Cuiling Lan'] | 2020-06-01 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/392_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123520715.pdf | eccv-2020-8 | ['unsupervised-person-re-identification'] | ['computer-vision'] | [-3.62534374e-01 -1.48535743e-01 -2.97209561e-01 -7.10523427e-01
-4.44746912e-01 -3.44470412e-01 6.27139032e-01 3.25348049e-01
-6.45978153e-01 9.46628213e-01 1.43897265e-01 7.90094286e-02
-3.97793621e-01 -6.56384408e-01 -4.74429429e-01 -8.76035988e-01
-1.11570610e-02 9.34443951e-01 2.08143845e-01 -7.81453960e-03
-2.51195095e-02 2.05434501e-01 -1.51829576e+00 -2.41926670e-01
1.33407938e+00 7.14969337e-01 -1.40194653e-03 1.43801868e-01
-1.60242960e-01 1.99704155e-01 -7.55611777e-01 -7.13417649e-01
6.73051715e-01 -1.50502190e-01 -4.84055579e-01 -2.99768418e-01
4.18185681e-01 -3.79482120e-01 -2.41498917e-01 1.30191147e+00
7.42930591e-01 2.98978299e-01 6.79050088e-01 -1.56571579e+00
-7.05562770e-01 4.75067526e-01 -1.20630515e+00 1.37507871e-01
1.68119907e-01 -1.04715921e-01 6.60450578e-01 -6.10799968e-01
4.00378287e-01 1.24829483e+00 7.63767540e-01 6.26065731e-01
-1.27257228e+00 -1.13818300e+00 5.71189761e-01 2.55612910e-01
-1.80553210e+00 -2.96222210e-01 6.09270096e-01 -3.89331132e-01
4.80472893e-01 9.11233574e-02 2.83695370e-01 1.09247148e+00
-2.81655252e-01 8.09822857e-01 6.70358896e-01 -2.36948416e-01
1.44909367e-01 6.19328141e-01 4.77624714e-01 1.03854060e-01
5.20975769e-01 1.40802458e-01 -6.74466193e-01 -3.12370777e-01
4.46003407e-01 1.87717512e-01 -1.88858993e-02 -6.01904273e-01
-6.16415381e-01 5.60606539e-01 3.19461614e-01 -1.54441461e-01
-1.75892919e-01 -3.78576040e-01 4.01898652e-01 8.49602520e-02
5.35111666e-01 6.62514195e-02 -4.55050200e-01 -2.36038789e-02
-7.93219626e-01 5.15783608e-01 6.78390503e-01 1.17205703e+00
7.76953697e-01 -4.80223686e-01 -3.30413908e-01 1.22300470e+00
1.45177558e-01 4.68846053e-01 5.11207283e-01 -3.23187798e-01
7.07638383e-01 6.22861028e-01 3.68416309e-01 -1.03582633e+00
-2.33961269e-01 -5.11388302e-01 -9.88804340e-01 -1.36234120e-01
8.53359878e-01 -2.49197245e-01 -8.61299634e-01 1.98613250e+00
7.06264198e-01 1.17165796e-01 -1.45400316e-01 8.29878151e-01
5.48013628e-01 2.47726098e-01 3.65475237e-01 5.97466566e-02
1.22649503e+00 -6.39576197e-01 -5.00291407e-01 -4.34215963e-01
4.56245095e-01 -2.84194738e-01 6.68440402e-01 1.53638914e-01
-8.47462595e-01 -5.40315866e-01 -9.68955219e-01 1.12796880e-01
-3.01549762e-01 3.34662169e-01 3.19121242e-01 5.24484634e-01
-6.97801769e-01 4.38505799e-01 -6.66538298e-01 -4.32412118e-01
4.54189420e-01 5.99172056e-01 -3.90971184e-01 -1.70923024e-01
-1.15874708e+00 4.91868258e-01 3.68619084e-01 -5.14637008e-02
-2.67101169e-01 -9.41745937e-01 -6.65882409e-01 4.01929282e-02
8.11233819e-02 -6.34893417e-01 9.39365864e-01 -8.12193274e-01
-1.04650247e+00 9.85753894e-01 -2.28440195e-01 -5.78454792e-01
9.36792374e-01 -2.52835184e-01 -4.91136253e-01 -1.89354181e-01
4.46280569e-01 8.70984375e-01 7.52529502e-01 -1.31279480e+00
-1.11208987e+00 -5.72334945e-01 -3.67745548e-01 2.65316337e-01
-6.22149229e-01 -9.83931571e-02 -6.34515226e-01 -6.59124672e-01
-2.43534483e-02 -7.12593913e-01 1.53501108e-01 -1.84280381e-01
-5.44236064e-01 -5.75068414e-01 5.18909931e-01 -6.22966945e-01
1.19790888e+00 -2.28272939e+00 -2.79923975e-01 5.61797380e-01
3.71593624e-01 2.22508505e-01 -8.20268318e-02 -2.79879617e-03
-1.70769766e-01 -1.86788708e-01 9.97071415e-02 -7.66597450e-01
8.91303830e-03 2.16189981e-01 -1.93851173e-01 5.57229996e-01
4.02337871e-02 4.39567566e-01 -1.10982084e+00 -3.92098874e-01
-1.61389276e-01 2.78466552e-01 -3.63717258e-01 3.30152422e-01
3.26418877e-01 5.32637656e-01 -2.94361174e-01 4.64381844e-01
1.16022992e+00 -2.23022085e-02 3.14123690e-01 -9.50575918e-02
2.25401655e-01 2.35703677e-01 -1.39176297e+00 1.10098326e+00
1.51893824e-01 2.60049671e-01 1.23258188e-01 -1.03134716e+00
1.20618665e+00 -2.61615723e-01 5.17504454e-01 -9.51042712e-01
3.85536887e-02 -3.85315940e-02 -1.95353031e-02 -1.39302611e-01
6.51575089e-01 -2.24668771e-01 9.99925286e-02 3.92378807e-01
-1.13504328e-01 7.27793634e-01 3.07336628e-01 1.80955172e-01
4.75738376e-01 -8.67811069e-02 -1.00006141e-01 -3.95796776e-01
4.24208105e-01 -3.12559962e-01 1.01053047e+00 8.47741783e-01
-4.22239929e-01 6.26653850e-01 4.42111790e-01 -3.27071130e-01
-8.43344271e-01 -1.32485271e+00 -2.80518234e-01 1.22992039e+00
7.01880693e-01 -2.03969300e-01 -6.05359495e-01 -1.00484884e+00
4.45271492e-01 5.24690449e-01 -7.20814347e-01 -3.04260403e-01
-4.92111087e-01 -1.06473124e+00 5.39001524e-01 7.25947976e-01
4.72315609e-01 -3.61237496e-01 1.18111223e-01 2.37688832e-02
-2.29197979e-01 -9.14650500e-01 -6.59282267e-01 9.21411961e-02
-4.94300693e-01 -1.10028470e+00 -9.49225187e-01 -8.27052712e-01
1.19887054e+00 3.80243808e-01 9.12062764e-01 -5.50438538e-02
-7.48158172e-02 1.88022211e-01 -9.70821232e-02 -6.54065847e-01
-1.29980827e-02 3.02592739e-02 5.93859553e-01 2.98631657e-02
7.97693968e-01 -7.14872658e-01 -6.82252884e-01 7.85510719e-01
-5.53635001e-01 -3.01662028e-01 2.76666015e-01 8.88041019e-01
4.61338520e-01 3.31568182e-01 9.72257376e-01 -7.89804935e-01
6.56108022e-01 -7.62180388e-01 -5.68879664e-01 2.87814975e-01
-8.92231584e-01 -4.45487387e-02 4.98604566e-01 -7.96125591e-01
-1.18448567e+00 -2.57878304e-01 2.75787979e-01 -2.01309830e-01
-1.89608529e-01 1.59010068e-01 -2.82992393e-01 3.72206956e-01
6.74827456e-01 6.77936301e-02 9.56475139e-02 -5.21795571e-01
1.28552362e-01 7.65331089e-01 9.84097362e-01 -8.04352880e-01
1.01343608e+00 5.74311793e-01 -4.94455576e-01 -5.26258647e-01
-7.09973216e-01 -9.25444186e-01 -7.03522205e-01 1.08307257e-01
4.10651237e-01 -1.13224900e+00 -5.96223891e-01 6.75366879e-01
-8.04050028e-01 -1.27854794e-01 -3.11731309e-01 3.77106756e-01
1.09310269e-01 5.46115696e-01 -3.88583213e-01 -1.05323040e+00
-4.53637570e-01 -7.18301475e-01 8.34503710e-01 7.32527196e-01
-3.34337562e-01 -9.83614624e-01 -1.03230774e-01 2.04291627e-01
1.83768421e-01 -1.63991421e-01 7.38831937e-01 -1.08488655e+00
-1.23160191e-01 -5.85002303e-01 -5.66255808e-01 2.32852533e-01
3.26775134e-01 -3.30296665e-01 -9.06443417e-01 -5.32682478e-01
-5.40166318e-01 -9.13048089e-02 7.80913651e-01 3.92495424e-01
1.03073502e+00 -6.07048906e-02 -6.64280772e-01 5.24408817e-01
8.46208930e-01 9.06028133e-03 4.38081503e-01 6.02825940e-01
6.52943015e-01 8.71798098e-01 7.65133739e-01 7.15712309e-01
8.15802336e-01 6.17115855e-01 -2.67943852e-02 -9.08031613e-02
8.60640332e-02 -5.92069685e-01 2.33660251e-01 1.36522636e-01
1.99273020e-01 -9.11390558e-02 -8.13621640e-01 7.87978828e-01
-1.96568751e+00 -9.33194458e-01 1.67376325e-01 2.76051474e+00
1.06764627e+00 3.08878124e-01 7.45438695e-01 -2.04996243e-01
1.13815141e+00 -2.95295566e-01 -8.36092114e-01 2.07482487e-01
-7.37061501e-02 -3.07988971e-01 6.93556786e-01 2.09151983e-01
-1.11718905e+00 5.75256109e-01 5.74917650e+00 8.76284420e-01
-9.40337062e-01 -2.58549184e-01 6.02903128e-01 -2.31850147e-01
1.18108448e-02 -2.74151802e-01 -1.39561737e+00 8.28685045e-01
5.58874011e-01 -3.87996763e-01 1.11206979e-01 9.02905524e-01
6.72247484e-02 -1.98339537e-01 -1.23327219e+00 1.23748720e+00
-1.92280725e-01 -8.36240768e-01 -1.36666775e-01 1.39009580e-01
7.25093007e-01 -9.77640226e-03 1.73379362e-01 4.57682639e-01
5.69013596e-01 -6.84890985e-01 8.76801491e-01 4.68289882e-01
6.81117356e-01 -1.07526457e+00 8.17216337e-01 5.43328166e-01
-1.12752533e+00 -3.03470790e-01 -5.43618977e-01 1.44612178e-01
-5.49502335e-02 6.88034356e-01 -7.73932517e-01 5.75196147e-01
1.05301630e+00 7.61620402e-01 -6.05813920e-01 1.17486298e+00
1.66639760e-02 3.23940486e-01 -5.28967738e-01 3.56951118e-01
-2.99383819e-01 -3.35717559e-01 5.47174156e-01 1.14092159e+00
-2.50591133e-02 -2.75098085e-01 2.89743751e-01 9.97343600e-01
-2.09364012e-01 -1.89398021e-01 -4.01484109e-02 3.20849836e-01
9.86163199e-01 1.00363672e+00 -4.62804615e-01 -3.06764305e-01
-1.43297911e-01 7.87916958e-01 4.49051648e-01 5.43817878e-01
-7.98733652e-01 -4.16678965e-01 1.03564322e+00 2.46461675e-01
1.98309004e-01 7.21667111e-02 -3.85715544e-01 -1.07325757e+00
4.11157310e-01 -5.62535226e-01 7.68852651e-01 -5.09272330e-03
-2.15603423e+00 1.46125317e-01 3.17473233e-01 -1.25513208e+00
-2.07948968e-01 -2.00100452e-01 -7.59329498e-01 1.12481952e+00
-1.83641028e+00 -9.57592607e-01 -4.29041743e-01 5.29676020e-01
1.37416413e-02 -1.86572045e-01 3.30348164e-01 7.14132011e-01
-9.43050086e-01 1.50280881e+00 2.98731059e-01 4.00677025e-01
1.35254848e+00 -1.28640401e+00 3.45691800e-01 8.42608571e-01
-5.22014678e-01 9.59695220e-01 6.39618039e-01 -8.54020536e-01
-9.06878173e-01 -1.14981854e+00 6.13555908e-01 -3.96940142e-01
3.09128016e-01 -4.46735650e-01 -1.07021177e+00 4.43030447e-01
-3.92568082e-01 -6.21441975e-02 7.59030581e-01 3.80786985e-01
-5.64915419e-01 -4.45910484e-01 -1.43092084e+00 2.18426660e-01
9.87378836e-01 -3.08831185e-01 -3.36629093e-01 2.06968427e-01
1.83054581e-01 -3.83620530e-01 -6.12813175e-01 2.87515342e-01
6.30771160e-01 -9.33671534e-01 1.13785994e+00 -5.44543386e-01
-5.19629680e-02 -4.43796724e-01 2.81840384e-01 -1.22960675e+00
-2.77657807e-01 -4.83220488e-01 -3.78100164e-02 1.91195738e+00
2.10599199e-01 -8.69930744e-01 9.48622823e-01 1.12918794e+00
3.47692639e-01 -4.24393922e-01 -8.20886552e-01 -1.05201387e+00
7.04264417e-02 -1.35913566e-01 8.69091392e-01 9.86295700e-01
4.94974144e-02 1.39333665e-01 -3.71191114e-01 4.26078916e-01
8.89996707e-01 -1.48660839e-01 1.08383083e+00 -1.58808839e+00
-1.24787815e-01 -5.11244118e-01 -4.44366038e-01 -1.29931426e+00
2.07811192e-01 -7.41913617e-01 1.02078676e-01 -1.26544642e+00
5.49936652e-01 -9.63135004e-01 -4.68636632e-01 4.35638905e-01
-4.97877747e-01 -2.11330235e-01 -1.03942715e-02 3.87456745e-01
-6.95630968e-01 6.22456849e-01 5.68930924e-01 -1.73790663e-01
-5.67141116e-01 9.80178341e-02 -1.13551390e+00 6.06193841e-01
6.16457283e-01 -4.73894954e-01 -4.45948541e-01 -3.37045848e-01
-5.70562929e-02 -6.43035769e-01 2.56494075e-01 -8.94710362e-01
6.05545640e-01 3.03184409e-02 7.53679872e-01 -6.09965086e-01
2.54087299e-02 -5.50522387e-01 -1.93712845e-01 -8.43563452e-02
-1.69687808e-01 -1.60686567e-01 1.51449755e-01 8.34817231e-01
-3.58920172e-02 1.90656900e-01 9.82993960e-01 2.89241105e-01
-5.03258049e-01 4.85656738e-01 2.43771642e-01 1.45620137e-01
8.87579083e-01 -4.64145392e-01 -3.77705038e-01 -3.19322199e-01
-3.10847282e-01 9.54760253e-01 4.68145341e-01 6.32161319e-01
3.80894125e-01 -1.22073126e+00 -6.84087813e-01 4.26510394e-01
3.20821941e-01 3.95737648e-01 4.13223028e-01 9.32535708e-01
5.99529222e-02 -8.80827680e-02 -9.96259302e-02 -5.09625852e-01
-1.29146254e+00 2.64181167e-01 4.49375272e-01 -3.24226797e-01
-5.67434371e-01 1.02223396e+00 6.02100015e-01 -6.74692929e-01
7.60513306e-01 1.71061888e-01 -2.52847373e-01 1.45768240e-01
1.01912451e+00 5.68761528e-01 -1.50592521e-01 -4.81661201e-01
-4.07066464e-01 1.52598798e-01 -5.81232548e-01 3.38413328e-01
1.35223365e+00 -3.81562084e-01 -4.02669162e-02 -1.06557228e-01
9.19126511e-01 5.35383821e-02 -1.40999007e+00 -3.91811520e-01
1.49022862e-01 -6.17252767e-01 -4.97060001e-01 -7.25444913e-01
-8.98685753e-01 5.66483498e-01 6.59224272e-01 1.18573092e-01
9.56119597e-01 -2.90247649e-02 9.24084961e-01 7.42311478e-02
1.84019089e-01 -1.52294755e+00 -1.84533477e-01 3.85289282e-01
2.54577905e-01 -1.24749243e+00 8.73487368e-02 -4.22903210e-01
-6.35842204e-01 8.97568285e-01 9.09526765e-01 1.04068354e-01
4.91601765e-01 2.45167881e-01 -1.44791082e-02 1.90136522e-01
-1.86758965e-01 -4.86360453e-02 4.11462635e-01 1.10596955e+00
6.85747936e-02 1.92653656e-01 -6.13035820e-02 1.07554102e+00
-3.13553542e-01 -3.00560951e-01 1.47543222e-01 6.56152725e-01
-6.60937876e-02 -1.32378328e+00 -4.63503659e-01 5.20599961e-01
-2.25780427e-01 1.81614250e-01 -4.28204834e-01 7.28627205e-01
2.50264555e-01 8.83133471e-01 3.04946750e-01 -3.51463199e-01
4.39442813e-01 -5.99360978e-03 2.82843411e-01 -3.44054431e-01
-2.28514448e-01 -5.85734956e-02 -7.90965557e-02 -6.42255470e-02
-8.27368349e-02 -7.78688371e-01 -1.41353488e+00 -4.77940738e-01
-4.50232178e-01 2.87267536e-01 3.73908311e-01 9.17517960e-01
5.88555574e-01 4.41709235e-02 6.36466026e-01 -6.50399148e-01
-8.55384171e-01 -8.08594406e-01 -9.05839324e-01 6.88390195e-01
3.39542538e-01 -9.44113314e-01 -2.47127339e-01 -1.89190686e-01] | [14.741352081298828, 1.0656429529190063] |
5c1cf30f-f7f5-498d-b549-b03d363a1f5d | paraphrasing-textual-entailment-and-semantic | 2208.05387 | null | https://arxiv.org/abs/2208.05387v1 | https://arxiv.org/pdf/2208.05387v1.pdf | Paraphrasing, textual entailment, and semantic similarity above word level | This dissertation explores the linguistic and computational aspects of the meaning relations that can hold between two or more complex linguistic expressions (phrases, clauses, sentences, paragraphs). In particular, it focuses on Paraphrasing, Textual Entailment, Contradiction, and Semantic Similarity. In Part I: "Similarity at the Level of Words and Phrases", I study the Distributional Hypothesis (DH) and explore several different methodologies for quantifying semantic similarity at the levels of words and short phrases. In Part II: "Paraphrase Typology and Paraphrase Identification", I focus on the meaning relation of paraphrasing and the empirical task of automated Paraphrase Identification (PI). In Part III: "Paraphrasing, Textual Entailment, and Semantic Similarity", I present a novel direction in the research on textual meaning relations, resulting from joint research carried out on on paraphrasing, textual entailment, contradiction, and semantic similarity. | ['Venelin Kovatchev'] | 2022-08-10 | null | null | null | null | ['paraphrase-identification'] | ['natural-language-processing'] | [ 2.82577842e-01 -1.19331278e-01 -4.52848405e-01 -5.22206604e-01
-3.61879230e-01 -9.37517762e-01 6.37590230e-01 7.90985584e-01
3.89892026e-03 3.58958453e-01 9.63248372e-01 -5.12248516e-01
-5.84050834e-01 -5.63765764e-01 -9.36810151e-02 -1.30275384e-01
5.82957745e-01 5.98967671e-02 -3.31072360e-01 -4.70299035e-01
1.22060299e+00 3.77217680e-01 -1.36292505e+00 7.74904191e-01
6.39124095e-01 7.02198386e-01 1.00668006e-01 2.60682255e-01
-8.43304157e-01 1.17508328e+00 -6.96307719e-01 -7.46555090e-01
-3.50354880e-01 -9.31456864e-01 -1.66108239e+00 -1.92138813e-02
3.72956425e-01 3.87784123e-01 -7.62028992e-02 1.20286584e+00
2.60022193e-01 1.04524814e-01 7.32845068e-01 -1.18087816e+00
-9.73888993e-01 8.03930044e-01 -1.10390253e-01 5.59644520e-01
1.60405159e+00 -4.60444987e-01 1.25515509e+00 -9.03910160e-01
6.72270179e-01 1.53055525e+00 9.18050706e-01 1.68576136e-01
-1.07317173e+00 -2.94782013e-01 -2.57916451e-01 4.49822068e-01
-1.25456214e+00 -2.88731277e-01 7.42964506e-01 -4.38977689e-01
1.46241140e+00 5.43066680e-01 7.25212395e-01 1.04377139e+00
4.74508613e-01 4.92973238e-01 1.04035115e+00 -7.16300666e-01
9.83546898e-02 3.57456863e-01 8.11219394e-01 2.67683357e-01
2.46742830e-01 -4.58097279e-01 -4.32342619e-01 -3.27425182e-01
1.34190470e-01 -1.48046315e-01 -1.04075648e-01 2.14847833e-01
-8.77678633e-01 1.08060372e+00 -1.86935421e-02 1.21787775e+00
-5.78567013e-02 -2.75723457e-01 9.53497112e-01 9.14250672e-01
2.76490510e-01 8.56083453e-01 3.97718027e-02 -3.68286133e-01
-8.90900493e-01 3.04767072e-01 1.18250418e+00 1.12921405e+00
4.75779444e-01 -5.21433294e-01 -7.17132390e-02 1.07018399e+00
2.69669175e-01 3.78880292e-01 1.11557925e+00 -1.14179564e+00
6.19360924e-01 8.04764450e-01 -1.01726964e-01 -1.44021738e+00
-9.16868150e-02 7.69849941e-02 -3.16593260e-01 -8.25837314e-01
-5.74413277e-02 3.23037207e-01 3.62520246e-03 1.51317942e+00
-2.15211332e-01 -5.04986882e-01 3.67667049e-01 4.56125826e-01
1.21822810e+00 4.88156885e-01 1.47326514e-01 -6.05983853e-01
1.55932748e+00 -6.31488860e-01 -8.34077001e-01 -3.24057966e-01
1.09303868e+00 -1.15023243e+00 1.69781911e+00 -2.59653330e-01
-1.51624978e+00 -3.73161435e-01 -8.30744088e-01 -3.89199227e-01
-5.14517486e-01 -2.83811957e-01 2.57026136e-01 4.78601903e-01
-7.85016119e-01 7.27468193e-01 1.07323296e-01 -9.17429209e-01
-1.35252357e-01 -3.18631113e-01 -1.87765017e-01 1.52920596e-02
-1.13312054e+00 1.34645987e+00 2.71353483e-01 -3.47448438e-01
4.45054233e-01 -7.71001577e-01 -1.11010635e+00 1.39583513e-01
-1.33218840e-01 -9.29542363e-01 9.83972609e-01 -1.14858627e+00
-9.68158305e-01 1.71714318e+00 -6.53106868e-01 -3.68886404e-02
-1.44681498e-01 -8.15172568e-02 -4.21285480e-01 1.42849728e-01
6.22041225e-01 -7.17218220e-02 3.08846533e-01 -8.82030249e-01
-2.52530932e-01 -6.08592212e-01 1.48842826e-01 3.93036693e-01
-3.71449590e-01 8.20995986e-01 4.54721153e-01 -8.91934752e-01
3.79160404e-01 -6.01486742e-01 5.16738594e-01 -4.91812438e-01
-1.53115869e-01 -7.87641406e-01 5.41086257e-01 -6.83973134e-01
1.69967020e+00 -2.23073363e+00 1.96938455e-01 2.15931848e-01
-7.36039951e-02 -2.57213980e-01 2.60107145e-02 1.19728434e+00
-5.02785683e-01 2.80256450e-01 -3.37532610e-01 3.06749530e-03
3.26346695e-01 5.13950229e-01 -5.42092919e-01 1.35171622e-01
-3.68427813e-01 1.17212951e+00 -8.74195218e-01 -7.24904060e-01
-6.17858544e-02 -2.88922817e-01 -1.24181241e-01 -5.98770119e-02
2.94968307e-01 -3.42212051e-01 -1.16559304e-01 6.40815437e-01
2.98464000e-01 -2.30209008e-01 3.42351347e-01 -2.33261347e-01
-1.24234520e-01 7.18448222e-01 -5.41356206e-01 1.66568208e+00
-7.61732221e-01 9.26598907e-01 -3.40204537e-01 -1.28634572e+00
8.40180576e-01 3.23543906e-01 9.59598180e-03 -7.54362166e-01
2.59753823e-01 2.45979875e-01 -2.07530990e-01 -1.02247822e+00
6.79848671e-01 -7.62079120e-01 -5.72200179e-01 7.28412032e-01
-3.83603245e-01 -8.29543531e-01 4.44296867e-01 3.89429361e-01
9.72876251e-01 -4.17148978e-01 9.43524659e-01 -7.03950822e-01
8.25984716e-01 2.22635061e-01 -1.59813553e-01 4.23811018e-01
4.37667593e-02 9.48446244e-02 6.63304865e-01 -9.08261836e-02
-1.08069611e+00 -1.22500253e+00 -3.46331805e-01 8.68155658e-01
5.13992846e-01 -7.05337107e-01 -5.98077953e-01 -2.07303226e-01
1.80296764e-01 1.30638134e+00 -4.49456036e-01 -3.30572814e-01
-4.61773217e-01 -1.98426262e-01 7.94129372e-01 6.43518686e-01
2.75167525e-01 -1.05682826e+00 -4.11351413e-01 -1.61105603e-01
-6.20585024e-01 -1.16680324e+00 -2.89494395e-01 -1.96800902e-01
-9.85398650e-01 -9.08717394e-01 -2.18732730e-01 -1.11473525e+00
3.54883194e-01 2.36771032e-01 1.39002049e+00 8.03053658e-03
-1.46923885e-02 4.35852945e-01 -7.35826671e-01 -8.14381763e-02
-7.22350597e-01 -3.35861683e-01 -2.48936161e-01 -8.00904453e-01
7.83872902e-01 -7.16319919e-01 1.73760634e-02 1.32515030e-02
-9.31713760e-01 -4.64774311e-01 1.49632454e-01 6.50957704e-01
4.91599068e-02 -1.14673950e-01 6.41581297e-01 -6.74516022e-01
1.57958496e+00 -8.38375688e-01 4.87956047e-01 6.12667382e-01
-7.31764674e-01 -2.82129440e-02 5.87735116e-01 -1.91047743e-01
-1.16456127e+00 -8.98137391e-01 -3.04654390e-01 7.87079632e-02
-2.28097647e-01 6.93706751e-01 6.62175044e-02 7.93980807e-02
7.74153948e-01 3.02292138e-01 2.58829355e-01 -2.26362482e-01
4.24325764e-01 1.05604970e+00 6.02434039e-01 -6.62774861e-01
3.52372915e-01 2.51066804e-01 -7.31151402e-02 -9.21721041e-01
-1.15686834e+00 -9.41308379e-01 -2.53616661e-01 -1.13544036e-02
4.54658240e-01 -6.13613605e-01 -4.31960076e-01 -1.46616474e-01
-1.16944957e+00 2.49639645e-01 -7.06775963e-01 3.36508930e-01
-8.32429349e-01 1.23831463e+00 -9.05403316e-01 -1.47043675e-01
-4.73768771e-01 -4.54341412e-01 8.52078497e-01 -8.41482878e-02
-1.45383608e+00 -1.60014963e+00 1.19525425e-01 6.05563462e-01
1.87273666e-01 8.32411274e-02 1.67806685e+00 -8.00193965e-01
7.45271742e-01 -1.07006803e-01 -3.37349474e-01 3.06542426e-01
4.69073027e-01 -1.54835701e-01 -1.70405954e-01 7.54936039e-02
7.96588540e-01 -5.97267568e-01 1.24277547e-01 4.42990102e-02
6.30390882e-01 -6.64561450e-01 -7.28971884e-02 7.22753927e-02
1.42008495e+00 1.09428853e-01 6.36392236e-01 3.63321245e-01
8.17771442e-03 1.09930503e+00 7.84777224e-01 3.99882078e-01
1.19796775e-01 3.42638969e-01 -4.08524036e-01 5.73037982e-01
-1.75563604e-01 -2.52605617e-01 3.01375002e-01 9.70035911e-01
4.40369457e-01 -2.75389906e-02 -7.09186614e-01 6.85071588e-01
-1.79364681e+00 -1.29795599e+00 -4.90920871e-01 1.65916681e+00
9.18207586e-01 -1.75293371e-01 -2.93006562e-02 3.69142979e-01
6.48272812e-01 1.59917429e-01 7.81679526e-02 -1.37441468e+00
-3.14343691e-01 2.59599447e-01 -3.35945129e-01 6.31118596e-01
-2.54365236e-01 9.66038525e-01 7.04177523e+00 1.06374705e+00
-5.96687257e-01 -1.35327578e-01 -1.52685475e-02 2.37597004e-01
-8.56281996e-01 2.51801521e-01 -2.19212621e-01 6.03017688e-01
6.65137649e-01 -6.28423750e-01 1.76608935e-01 6.99085057e-01
2.66642064e-01 -4.33546454e-01 -1.40628791e+00 1.06113434e+00
6.31630242e-01 -1.28943264e+00 5.08893847e-01 -5.04730999e-01
6.43646121e-01 -7.54110932e-01 -1.52905121e-01 -3.33531685e-02
-3.21289420e-01 -9.31539357e-01 5.60892940e-01 2.65497655e-01
4.27788585e-01 -5.30679226e-01 1.10707664e+00 2.40194306e-01
-1.08707631e+00 -1.52903125e-01 -5.21360993e-01 -5.84261358e-01
2.44734064e-01 3.42070043e-01 -1.08711950e-01 6.18690670e-01
3.83075923e-01 1.04052746e+00 -4.08412158e-01 4.23786223e-01
-3.54068398e-01 9.65426117e-02 -1.58878544e-03 -4.98757422e-01
1.84313178e-01 -4.16995585e-01 5.61484933e-01 1.67936862e+00
1.83970660e-01 2.87921160e-01 -3.95801842e-01 1.15502024e+00
3.24017167e-01 4.39622134e-01 -9.13765252e-01 -1.51314050e-01
8.68562877e-01 6.74866378e-01 -5.92624724e-01 -6.48851514e-01
-4.05747741e-01 1.17927432e+00 2.99235228e-02 9.08746943e-02
-6.12838745e-01 -6.58024967e-01 4.88940716e-01 -2.57866399e-04
-2.86018908e-01 1.45542815e-01 -9.78728712e-01 -1.03896236e+00
2.80323207e-01 -7.00522423e-01 3.53647470e-01 -1.03075361e+00
-1.85875273e+00 6.23906441e-02 6.85158491e-01 -7.52581656e-01
-4.00300920e-01 -4.48041171e-01 -1.07908642e+00 8.25079858e-01
-6.24022424e-01 -8.41015756e-01 -1.93151563e-01 6.16767943e-01
7.63178349e-01 -2.06792690e-02 8.69370937e-01 -1.35777384e-01
3.07859620e-03 5.83289802e-01 -1.07098222e-01 -2.32534498e-01
4.14823830e-01 -8.19795787e-01 1.76551864e-01 2.70639598e-01
-1.50699228e-01 1.05683482e+00 9.28503931e-01 -5.80496550e-01
-8.69843721e-01 -3.62586290e-01 2.19703817e+00 -3.97125959e-01
1.16685343e+00 3.39760810e-01 -7.37042129e-01 5.19257724e-01
3.69464695e-01 -1.21678197e+00 1.14669108e+00 2.84546521e-02
-4.72300768e-01 3.61729860e-01 -1.36179352e+00 7.59913206e-01
1.26037133e+00 -1.25412548e+00 -1.82248104e+00 6.79363608e-01
7.39917278e-01 2.05220819e-01 -1.13354850e+00 1.53304875e-01
5.89221835e-01 -9.57575500e-01 1.03865767e+00 -8.12219024e-01
9.14231300e-01 2.78707504e-01 -3.78865004e-01 -9.30102944e-01
-4.13251936e-01 -2.62963057e-01 4.34472919e-01 1.18628013e+00
2.25459754e-01 -5.27745783e-01 4.98484224e-01 7.65523255e-01
-2.63201892e-01 -5.86913407e-01 -8.29085112e-01 -1.05128860e+00
4.72307384e-01 -2.54035741e-01 2.92465985e-01 1.37272763e+00
1.34177625e+00 7.88053334e-01 4.68475342e-01 -6.99702501e-01
2.33841345e-01 6.38259590e-01 1.37611136e-01 -7.81966448e-01
-1.07230261e-01 -7.06507921e-01 -4.40933555e-01 -8.64383101e-01
8.10939074e-01 -1.21495545e+00 -5.50026119e-01 -1.58933866e+00
4.92574692e-01 3.43139768e-01 5.59500217e-01 7.38443881e-02
8.54736120e-02 -2.22887117e-02 2.74789393e-01 4.94039059e-01
-3.25527966e-01 2.07060874e-01 7.46095061e-01 7.48489127e-02
-4.58889417e-02 -2.07201555e-01 -8.97896826e-01 1.01773512e+00
9.11442280e-01 -4.38969702e-01 -5.80767035e-01 -1.55560046e-01
7.67410934e-01 2.92661309e-01 3.39851528e-01 -3.36845487e-01
2.03134686e-01 -4.54593480e-01 -2.56198585e-01 -3.13387215e-01
4.13308330e-02 -7.32843399e-01 2.50229925e-01 5.96587360e-01
-8.69467258e-01 8.11293840e-01 5.69071099e-02 1.37955531e-01
-6.03333831e-01 -1.24681866e+00 5.56559384e-01 -3.37242395e-01
-4.07520503e-01 -8.47688854e-01 -8.21760893e-01 5.78891873e-01
8.88872921e-01 -8.82688344e-01 -2.09453031e-01 -4.19297963e-01
-7.79023826e-01 -1.17665708e-01 5.69181204e-01 4.18001562e-01
7.98623562e-01 -1.30759215e+00 -4.39306676e-01 -3.01515490e-01
3.11010987e-01 -8.46244097e-01 -3.52306366e-02 8.04915488e-01
-4.03705299e-01 5.47170579e-01 -1.21059217e-01 4.57042828e-02
-1.65293217e+00 6.01771772e-01 6.21684939e-02 -1.10347241e-01
-2.80506283e-01 6.50137424e-01 -2.70110238e-02 -3.89796756e-02
-1.31785706e-01 -4.08763498e-01 -2.03033030e-01 2.71136582e-01
1.04084834e-01 5.78127265e-01 -3.42806101e-01 -7.52160788e-01
-3.73998404e-01 9.14304435e-01 1.68075055e-01 1.75740533e-02
5.36348879e-01 -6.66951358e-01 -9.51359987e-01 7.96319008e-01
1.72620392e+00 -9.69512761e-02 5.86494505e-01 -7.41765648e-02
5.50615191e-01 -5.21276414e-01 -7.20400870e-01 -5.56083798e-01
8.25823024e-02 4.78095651e-01 -2.16425046e-01 5.80650032e-01
8.46440315e-01 4.92404908e-01 9.39142644e-01 6.40601039e-01
-3.27806287e-02 -1.36296940e+00 2.33208030e-01 8.19306076e-01
1.20380807e+00 -6.28854096e-01 6.19851872e-02 -6.39157116e-01
-6.25567913e-01 1.18858755e+00 1.98972031e-01 -7.43449181e-02
5.69257855e-01 1.58655062e-01 -3.19472909e-01 -5.42252004e-01
-3.91900837e-01 2.40144357e-01 8.40402171e-02 2.17102289e-01
8.23980749e-01 -1.25391096e-01 -1.55163860e+00 5.96481740e-01
-9.67531562e-01 -2.66295403e-01 5.21616578e-01 8.90707850e-01
-5.10616064e-01 -9.41202581e-01 -1.98874220e-01 3.84369791e-01
-4.99196261e-01 -4.87404406e-01 -1.21143270e+00 7.05042183e-01
-1.32786885e-01 1.35546041e+00 2.23943353e-01 -3.40536088e-01
2.22380459e-01 2.05843285e-01 7.11300969e-01 -6.07805371e-01
-1.09077108e+00 -7.42173553e-01 6.16470873e-01 -5.33363640e-01
-6.79851890e-01 -5.60968637e-01 -1.30432296e+00 -8.69593978e-01
-9.72192734e-02 5.66351414e-01 4.10867602e-01 1.33199978e+00
1.42921552e-01 -1.39038146e-01 4.25044060e-01 8.15136172e-03
-8.88788462e-01 -9.56364155e-01 -5.95291555e-01 1.04108727e+00
-3.66345972e-01 -9.26450789e-02 -7.02034950e-01 -9.74098295e-02] | [10.832481384277344, 9.142715454101562] |
0650f689-f9cc-43b6-9e65-7a708aed67f6 | the-stable-entropy-hypothesis-and-entropy | 2302.06784 | null | https://arxiv.org/abs/2302.06784v1 | https://arxiv.org/pdf/2302.06784v1.pdf | The Stable Entropy Hypothesis and Entropy-Aware Decoding: An Analysis and Algorithm for Robust Natural Language Generation | State-of-the-art language generation models can degenerate when applied to open-ended generation problems such as text completion, story generation, or dialog modeling. This degeneration usually shows up in the form of incoherence, lack of vocabulary diversity, and self-repetition or copying from the context. In this paper, we postulate that ``human-like'' generations usually lie in a narrow and nearly flat entropy band, and violation of these entropy bounds correlates with degenerate behavior. Our experiments show that this stable narrow entropy zone exists across models, tasks, and domains and confirm the hypothesis that violations of this zone correlate with degeneration. We then use this insight to propose an entropy-aware decoding algorithm that respects these entropy bounds resulting in less degenerate, more contextual, and "human-like" language generation in open-ended text generation settings. | ['Jackie C. K. Cheung', 'Jason Weston', 'Doina Precup', "Timothy J. O'Donnell", 'Kushal Arora'] | 2023-02-14 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 2.52056897e-01 5.63694179e-01 -1.76416710e-01 -1.92644119e-01
-4.34193820e-01 -7.00252056e-01 9.36632037e-01 9.18862745e-02
-1.92512944e-02 1.13073635e+00 7.33905733e-01 -2.41570100e-01
-7.81093165e-03 -6.38813198e-01 -5.07589936e-01 -3.71069282e-01
3.12458307e-01 7.02566206e-01 -1.13994896e-01 -4.73450333e-01
2.38998502e-01 -3.66324991e-01 -1.27078438e+00 1.53879941e-01
1.26032615e+00 4.49715942e-01 4.92791906e-02 6.87356651e-01
-2.16270953e-01 6.85682893e-01 -7.50857711e-01 -6.38817251e-01
7.04501718e-02 -7.96777964e-01 -8.59123230e-01 -2.99713105e-01
1.50597095e-02 -3.28734696e-01 -2.59774536e-01 9.11443830e-01
4.80207264e-01 1.63764492e-01 1.00695789e+00 -1.15242112e+00
-8.84384394e-01 1.28950965e+00 -1.66401580e-01 -1.93027318e-01
7.14899004e-01 2.64555514e-01 1.16360581e+00 -5.03300548e-01
9.58120823e-01 1.21746254e+00 7.71904707e-01 9.00797665e-01
-1.43555868e+00 -4.53956991e-01 -4.35232855e-02 -5.65401077e-01
-1.49499726e+00 -5.63738465e-01 6.27427101e-01 -4.86191273e-01
9.43738878e-01 3.58820230e-01 8.22501540e-01 1.74824858e+00
3.78339648e-01 6.46770537e-01 8.25275183e-01 -4.10863906e-01
5.12313783e-01 2.64248163e-01 -1.02033824e-01 5.13023853e-01
6.60602987e-01 2.20578700e-01 -9.38060522e-01 -2.92783082e-01
5.06139755e-01 -7.80298471e-01 -3.68435979e-01 -2.18949780e-01
-1.15081310e+00 8.68733048e-01 -5.72872199e-02 3.88397485e-01
-1.65547431e-02 -3.64424102e-02 3.23604852e-01 2.85571814e-01
6.41761184e-01 1.05493855e+00 -1.28011465e-01 -6.70272827e-01
-1.09567106e+00 6.76884055e-01 1.26827610e+00 1.02177644e+00
4.47664380e-01 6.46303445e-02 -3.56896073e-01 8.51681530e-01
2.05597207e-01 1.92815959e-01 9.93847609e-01 -9.35377002e-01
4.71803963e-01 4.33443695e-01 1.36307895e-01 -6.26243651e-01
-4.19903696e-01 -4.17836040e-01 -7.20016062e-01 -2.20974639e-01
5.32710791e-01 -5.41128874e-01 -5.07983625e-01 2.39581656e+00
-1.38298124e-01 -1.86570317e-01 3.05135638e-01 6.84729934e-01
5.74157417e-01 4.60749686e-01 2.70627216e-02 -4.12829816e-01
1.06916869e+00 -5.48977137e-01 -8.10673118e-01 -4.02209759e-01
7.97936797e-01 -6.52404845e-01 1.29044855e+00 -4.45144884e-02
-1.12059104e+00 -1.29324466e-01 -1.04664981e+00 -1.78111747e-01
-1.61448538e-01 -3.68819922e-01 6.76688373e-01 1.08147621e+00
-9.28594112e-01 8.51438284e-01 -4.37591940e-01 -4.53619838e-01
-1.36858923e-03 4.17523971e-03 -5.81186302e-02 3.66482019e-01
-1.20059776e+00 7.77027667e-01 5.04127741e-01 -3.93797934e-01
-4.07561630e-01 -5.81732810e-01 -6.42781794e-01 7.82884359e-02
1.42187968e-01 -1.12800872e+00 1.30498147e+00 -8.50784302e-01
-1.56119382e+00 6.17065668e-01 -6.95662871e-02 -4.08298045e-01
7.53332913e-01 -2.58978188e-01 -2.02236757e-01 -5.13684273e-01
-7.12106973e-02 6.64513230e-01 5.69716930e-01 -1.25184166e+00
1.48453880e-02 -2.88284868e-01 -1.80991530e-01 4.58712280e-01
-2.88618833e-01 -4.80744511e-01 1.60235807e-01 -7.41875589e-01
5.48866056e-02 -1.01244199e+00 3.53395008e-02 -6.21900320e-01
-7.48210907e-01 -1.62713125e-01 2.97856122e-01 -4.37309891e-01
1.66232705e+00 -1.81024981e+00 9.07494202e-02 8.38597268e-02
2.99204856e-01 -1.47037655e-01 -5.90602532e-02 6.69061780e-01
3.53842586e-01 5.90684831e-01 -7.90423378e-02 -5.86792231e-01
4.70574409e-01 3.92741263e-02 -3.97286266e-01 5.21230958e-02
-1.08813152e-01 8.52280021e-01 -9.83709276e-01 -5.37015855e-01
-2.81424075e-01 9.01561172e-04 -8.00454974e-01 1.12729669e-01
-6.94061279e-01 3.59404951e-01 -1.82461336e-01 3.97811174e-01
2.59422421e-01 -2.10096762e-01 4.13275570e-01 4.45595622e-01
-1.43550888e-01 5.38660109e-01 -7.03425467e-01 1.70568728e+00
-1.76371858e-01 7.75595665e-01 -3.71757776e-01 -2.95908868e-01
9.42548633e-01 1.31631017e-01 -4.38049845e-02 -4.15731907e-01
1.38664186e-01 2.88189739e-01 3.04118782e-01 -3.77574205e-01
1.14016962e+00 -3.91562432e-01 -1.30956367e-01 8.09510231e-01
6.38471171e-02 -4.97176945e-01 2.33058617e-01 2.28336468e-01
9.93547499e-01 -1.10486820e-02 2.40850255e-01 -3.77134532e-01
-1.45638555e-01 -2.32893944e-01 4.35659945e-01 1.10866046e+00
-9.93979350e-02 6.74912632e-01 8.01414728e-01 8.43529999e-02
-1.37504137e+00 -1.06565022e+00 -2.33423878e-02 1.01600599e+00
2.04726160e-01 -8.02130580e-01 -1.02656233e+00 -2.63368905e-01
-2.00458035e-01 1.23327589e+00 -3.85152847e-01 -5.50416231e-01
-3.20256531e-01 -6.48630083e-01 9.37568843e-01 1.94646254e-01
3.47896814e-01 -7.91294158e-01 -3.67504418e-01 2.03024670e-01
-7.60965765e-01 -8.49888980e-01 -5.33017814e-01 -1.74833760e-02
-7.11094081e-01 -4.59016055e-01 -5.86372614e-01 -3.84377539e-01
2.91936576e-01 -2.37885237e-01 1.30782163e+00 -1.62415713e-01
2.44206980e-01 -3.70388068e-02 -3.21736634e-01 -2.57841855e-01
-1.13648653e+00 5.14896393e-01 1.45554632e-01 -4.14528161e-01
6.56579733e-02 -6.51397169e-01 -4.53186840e-01 1.26163632e-01
-8.57508123e-01 2.99817204e-01 2.79650509e-01 9.67222452e-01
-8.25464819e-03 -2.92441785e-01 7.42772043e-01 -5.84810317e-01
1.46179390e+00 -5.97438216e-01 -1.99035108e-01 5.08277535e-01
-9.17358935e-01 5.75805604e-01 6.68118119e-01 -6.65019214e-01
-1.16663969e+00 -3.86632890e-01 5.37442490e-02 -4.65071239e-02
-1.11111929e-03 5.71746290e-01 2.38030627e-02 4.29289788e-01
9.22855318e-01 3.34421456e-01 8.74498412e-02 -2.85188586e-01
6.00789547e-01 9.10736084e-01 3.12995344e-01 -7.06661165e-01
4.54179019e-01 4.01583724e-02 -3.35166782e-01 -6.90634429e-01
-5.93468130e-01 9.44041237e-02 -1.04545049e-01 -2.10042953e-01
5.81071913e-01 -7.51811445e-01 -4.62002069e-01 3.22595894e-01
-1.06829929e+00 -5.54210901e-01 -5.66611350e-01 2.78577685e-01
-8.65894079e-01 4.84827787e-01 -4.90745723e-01 -1.06961870e+00
-4.52509403e-01 -8.58710170e-01 9.71528709e-01 4.03834194e-01
-1.12276471e+00 -1.06988418e+00 4.92249429e-01 3.82973254e-01
3.66781533e-01 2.55818255e-02 1.11301267e+00 -7.45983362e-01
-2.73711860e-01 -1.87821444e-02 2.18263850e-01 -6.64553195e-02
-1.05803488e-02 -1.38315908e-03 -6.82000697e-01 -1.08988516e-01
-2.49238700e-01 -4.51653808e-01 7.09894240e-01 1.23376317e-01
6.06969059e-01 -6.01808965e-01 -2.06109196e-01 6.07236624e-01
9.56602931e-01 6.35658428e-02 5.08382022e-01 -3.00862454e-02
1.63095057e-01 4.99619633e-01 1.28601745e-01 7.91794002e-01
4.95717168e-01 6.61940634e-01 -8.54198337e-02 3.72582376e-01
-4.91072005e-03 -9.21028137e-01 5.51640034e-01 7.94170976e-01
5.06221205e-02 -8.87162328e-01 -9.83515203e-01 4.77339953e-01
-1.87593293e+00 -1.10140443e+00 1.73973098e-01 2.23387599e+00
1.25299633e+00 2.51216322e-01 2.00778648e-01 -1.58809304e-01
6.02379084e-01 2.42172822e-01 -4.74513918e-01 -8.64322245e-01
-3.83654177e-01 -5.69668300e-02 2.49077752e-01 6.99278057e-01
-5.19836366e-01 1.07352495e+00 7.20449972e+00 9.15908754e-01
-8.91296148e-01 -4.53285575e-02 7.44238734e-01 -2.13592619e-01
-7.77691483e-01 1.21661983e-01 -5.70999742e-01 7.78396070e-01
9.33981776e-01 -7.06324756e-01 6.22242570e-01 5.47154427e-01
5.72064072e-02 -1.04294680e-01 -1.29603505e+00 7.42581010e-01
6.73787296e-02 -1.27487791e+00 1.90076768e-01 2.51689732e-01
8.43435764e-01 -2.30578661e-01 5.96349277e-02 3.28552991e-01
6.07955217e-01 -1.26211929e+00 1.06522095e+00 4.04497027e-01
9.07265425e-01 -3.83716583e-01 3.65687668e-01 7.49705493e-01
-7.32677877e-01 -1.95276402e-02 -1.53343573e-01 -2.36919850e-01
3.19809169e-01 4.53007460e-01 -1.18100810e+00 2.63442785e-01
-4.11197767e-02 6.79113157e-03 -3.50606769e-01 6.30230904e-01
3.13091986e-02 5.18941104e-01 -3.18750679e-01 -3.83160025e-01
-7.04311207e-02 -2.88874775e-01 8.73064280e-01 9.94030058e-01
4.82808679e-01 -9.30430815e-02 -3.17410946e-01 1.29156363e+00
-1.22147538e-01 -4.69776690e-02 -7.86240339e-01 -5.13922870e-01
6.08455598e-01 4.79461282e-01 -5.76618075e-01 -1.84851885e-01
2.45163664e-02 1.13736260e+00 1.76963419e-01 2.78315336e-01
-8.32299352e-01 -1.99160427e-01 7.58829176e-01 6.87450618e-02
-1.35274619e-01 -6.87122121e-02 -9.23525691e-01 -1.30359709e+00
6.43136203e-02 -8.93902183e-01 -6.23023175e-02 -6.08959913e-01
-1.19058073e+00 6.10039532e-01 -4.60823551e-02 -7.59497106e-01
-8.27644765e-01 -1.58981159e-02 -6.29388690e-01 5.84611833e-01
-6.50677085e-01 -7.28338838e-01 -8.25455114e-02 2.14540526e-01
5.22263825e-01 -2.71586984e-01 7.48301923e-01 -2.25231037e-01
-4.96156037e-01 1.02305496e+00 3.57122362e-01 -1.37280464e-01
5.22574663e-01 -1.21388960e+00 6.89814627e-01 5.58965206e-01
-7.49325892e-03 7.87557125e-01 1.00470257e+00 -1.01847124e+00
-9.81725633e-01 -7.52111733e-01 1.29551363e+00 -4.87228483e-01
5.21292150e-01 -6.53420031e-01 -5.57502031e-01 5.34569383e-01
3.63868177e-01 -8.96949589e-01 7.33449519e-01 3.73945534e-01
-4.63495225e-01 5.03937483e-01 -1.07389796e+00 1.07922649e+00
1.38329279e+00 -4.93873417e-01 -5.51352441e-01 4.11969841e-01
1.01646090e+00 -3.72781396e-01 -6.76615179e-01 1.75399557e-01
7.79093385e-01 -9.87333238e-01 5.97083628e-01 -5.99307954e-01
7.22546458e-01 3.84220034e-01 -1.33625016e-01 -1.48506844e+00
-9.75758508e-02 -1.39335227e+00 -1.18850783e-01 1.26649225e+00
7.20541418e-01 -4.71894950e-01 6.06475651e-01 7.72144377e-01
-1.21824376e-01 -6.25698149e-01 -9.33541477e-01 -8.97063315e-01
5.66997707e-01 -1.40464246e-01 8.11992407e-01 8.75871956e-01
7.70759225e-01 7.00119078e-01 -3.81649911e-01 -5.96965432e-01
4.24626410e-01 -2.81777307e-02 6.39662683e-01 -1.14252186e+00
-3.69630784e-01 -8.16930354e-01 9.99876484e-02 -1.27359521e+00
1.82793111e-01 -7.73732543e-01 2.05655038e-01 -1.16206026e+00
2.47827977e-01 -5.65658331e-01 3.80005866e-01 -3.52294650e-03
-4.02119756e-02 -1.81750447e-01 3.08358580e-01 1.98212817e-01
-5.20614564e-01 7.20908701e-01 1.00341547e+00 1.34601146e-01
-4.70524937e-01 -1.03746682e-01 -1.10664046e+00 4.97609824e-01
9.17315364e-01 -3.21363002e-01 -4.98349607e-01 -2.51010060e-01
7.31082439e-01 3.71792972e-01 4.33942713e-02 -9.19553638e-01
-7.29044005e-02 -2.19150558e-01 8.47353861e-02 -1.96376313e-02
4.31063443e-01 -9.39960182e-02 4.01667655e-01 3.34598422e-01
-6.87993526e-01 -1.74672022e-01 -1.66447699e-01 3.37260157e-01
2.24362478e-01 -2.49652445e-01 5.94128072e-01 -1.30172029e-01
8.37927759e-02 -2.61439919e-01 -5.37523925e-01 4.92385685e-01
7.77441680e-01 -4.68041539e-01 -5.49377441e-01 -8.42461526e-01
-4.94044989e-01 1.71012685e-01 7.43293345e-01 5.92740774e-01
2.43357643e-01 -1.09549868e+00 -7.83021450e-01 2.00931907e-01
-4.89896014e-02 -3.31884950e-01 6.82210997e-02 3.87868047e-01
-3.49206448e-01 4.83497739e-01 3.46606821e-02 -3.09114665e-01
-9.11750019e-01 -9.02350321e-02 3.58601481e-01 -5.25442004e-01
-2.33624265e-01 7.24043787e-01 6.68414161e-02 -4.29355413e-01
9.26619768e-02 -3.34174037e-01 2.71667689e-01 -5.60453199e-02
4.97865938e-02 3.96901369e-01 -2.49786332e-01 -3.88973415e-01
-5.91181032e-02 -5.97414039e-02 8.33573639e-02 -6.43322289e-01
7.67474353e-01 -3.46305102e-01 1.68091133e-01 4.76495922e-01
6.62141383e-01 4.80451360e-02 -1.04210186e+00 2.92036235e-02
-9.59093198e-02 -2.73418635e-01 -3.49267572e-01 -9.25272226e-01
-2.36464918e-01 3.98970902e-01 1.68009177e-01 6.05645895e-01
6.61058009e-01 1.11626744e-01 8.57757151e-01 3.16458404e-01
2.64633894e-01 -1.34621060e+00 1.35769337e-01 8.55427384e-01
8.98679852e-01 -8.87740016e-01 -2.13794336e-01 -4.13067378e-02
-8.75589967e-01 7.16821373e-01 7.24849105e-01 4.15969461e-01
3.09766203e-01 1.81711823e-01 -2.92935520e-01 1.22779205e-01
-1.26285911e+00 8.33691433e-02 -5.09098917e-02 5.90933979e-01
6.56995237e-01 3.66389334e-01 -6.78164065e-01 7.86238730e-01
-1.12611580e+00 -1.45377904e-01 7.62007535e-01 5.98721385e-01
-4.30978149e-01 -1.00607574e+00 -1.68869868e-01 2.76495248e-01
-2.51518339e-01 -3.45128596e-01 -1.01276517e+00 6.29576564e-01
1.52938932e-01 1.12891650e+00 1.97868213e-01 -5.09859383e-01
-1.99307114e-01 5.19994795e-01 4.56155151e-01 -3.67038846e-01
-6.09992087e-01 -1.93723455e-01 4.33486760e-01 -1.64337829e-01
1.78338245e-01 -6.19126439e-01 -9.68470752e-01 -8.66876483e-01
-6.36122882e-01 8.83811899e-03 4.64155674e-01 9.29065347e-01
5.74993670e-01 9.67791900e-02 2.80916303e-01 -3.43572438e-01
-7.30743349e-01 -1.22440112e+00 -4.77747619e-01 5.61641514e-01
2.67011464e-01 -3.89783114e-01 -4.48372513e-01 -1.59588829e-01] | [11.586286544799805, 9.20119857788086] |
8562f0ca-6513-4b47-ba85-e3d0fc017bb1 | a-graph-based-method-for-soccer-action | 2211.12334 | null | https://arxiv.org/abs/2211.12334v1 | https://arxiv.org/pdf/2211.12334v1.pdf | A Graph-Based Method for Soccer Action Spotting Using Unsupervised Player Classification | Action spotting in soccer videos is the task of identifying the specific time when a certain key action of the game occurs. Lately, it has received a large amount of attention and powerful methods have been introduced. Action spotting involves understanding the dynamics of the game, the complexity of events, and the variation of video sequences. Most approaches have focused on the latter, given that their models exploit the global visual features of the sequences. In this work, we focus on the former by (a) identifying and representing the players, referees, and goalkeepers as nodes in a graph, and by (b) modeling their temporal interactions as sequences of graphs. For the player identification, or player classification task, we obtain an accuracy of 97.72% in our annotated benchmark. For the action spotting task, our method obtains an overall performance of 57.83% average-mAP by combining it with other audiovisual modalities. This performance surpasses similar graph-based methods and has competitive results with heavy computing methods. Code and data are available at https://github.com/IPCV/soccer_action_spotting. | ['Gloria Haro', 'Coloma Ballester', 'Alejandro Cartas'] | 2022-11-22 | null | null | null | null | ['action-spotting'] | ['computer-vision'] | [ 3.43790166e-02 -4.05187130e-01 -2.56736994e-01 2.15632051e-01
-5.93859076e-01 -6.46176994e-01 6.01255655e-01 2.21344680e-01
-3.04345250e-01 2.31755957e-01 3.09714437e-01 2.24847257e-01
4.27177213e-02 -5.42872250e-01 -4.53729689e-01 -6.08640909e-01
-2.51928359e-01 2.78219402e-01 8.54691446e-01 -2.83753365e-01
3.90583009e-01 2.29111791e-01 -1.71475720e+00 6.56148493e-01
2.28532135e-01 1.08107996e+00 2.18580499e-01 9.10137177e-01
1.26911476e-01 1.48710704e+00 -5.78501344e-01 -4.66353863e-01
1.11215062e-01 -6.49643362e-01 -6.87084794e-01 4.95889112e-02
2.29259983e-01 8.80058184e-02 -9.69589174e-01 9.41065371e-01
4.22166944e-01 4.27546442e-01 2.00337991e-01 -1.72397852e+00
5.00761308e-02 6.67332172e-01 -8.27800870e-01 5.40477455e-01
9.18843508e-01 8.34405348e-02 1.11931694e+00 -7.51173496e-01
8.43800128e-01 9.74539936e-01 5.09123921e-01 2.53478974e-01
-8.07427824e-01 -6.51608527e-01 7.69920722e-02 1.08146834e+00
-1.52380681e+00 -3.77106488e-01 7.27566838e-01 -7.01533437e-01
7.57193327e-01 2.73784548e-01 9.54387903e-01 8.46477747e-01
7.36980736e-02 1.10420310e+00 6.49626493e-01 -2.65234441e-01
1.28351092e-01 -4.51880276e-01 -1.01566441e-01 6.05969131e-01
-3.49406004e-01 -1.63259208e-01 -1.19433153e+00 -9.20210499e-03
6.24670804e-01 1.37393221e-01 -3.14597934e-01 -2.24912137e-01
-1.18933833e+00 5.53378701e-01 2.69911289e-01 2.59392709e-01
-4.03971761e-01 3.26927304e-01 8.16133022e-01 1.93445578e-01
3.86729896e-01 2.40989089e-01 1.61870331e-01 -8.60487342e-01
-8.93361151e-01 4.37125683e-01 6.91971421e-01 7.22312391e-01
3.29660296e-01 -1.71957627e-01 -2.99947798e-01 6.41235769e-01
-2.27919921e-01 9.28353518e-02 2.02210739e-01 -9.43630278e-01
6.15345120e-01 8.46573830e-01 9.52473190e-03 -1.42682922e+00
-4.09238189e-01 2.46326942e-02 -4.41875309e-01 1.35927960e-01
7.86574066e-01 1.42874286e-01 -5.43059349e-01 1.39351642e+00
2.77583867e-01 5.78005791e-01 -4.90285248e-01 1.11952877e+00
8.47058356e-01 6.86356425e-01 -5.34097292e-02 -3.80871631e-02
1.53711987e+00 -1.09000242e+00 -6.68946981e-01 -2.45117739e-01
3.78206074e-01 -7.04443753e-01 9.25197899e-01 5.35246611e-01
-1.00445557e+00 -4.47001666e-01 -8.63355279e-01 1.84509814e-01
-1.52160853e-01 1.42069861e-01 4.95818049e-01 3.09600264e-01
-8.51270676e-01 6.27682865e-01 -1.10854542e+00 -2.81222135e-01
1.65228114e-01 -2.26976201e-02 -4.85121369e-01 -9.47685763e-02
-1.06703937e+00 5.45063674e-01 4.21822995e-01 -8.01163688e-02
-9.89440084e-01 -4.30364102e-01 -8.04269910e-01 -1.32758440e-02
9.20528710e-01 1.13795735e-02 1.14590073e+00 -9.24628079e-01
-1.30489576e+00 8.15945506e-01 2.32312828e-02 -3.39449614e-01
7.03077495e-01 -1.73519611e-01 -3.67441446e-01 4.47051287e-01
3.27835865e-02 8.59496668e-02 6.01137221e-01 -7.38834023e-01
-9.61243331e-01 -4.13995326e-01 3.14254045e-01 4.56853837e-01
-1.64004669e-01 3.59917253e-01 -1.05471337e+00 -6.18992090e-01
4.66562957e-02 -1.03728640e+00 -7.58123472e-02 -1.41202331e-01
-3.13132077e-01 -2.84480751e-01 6.47133827e-01 -8.95114005e-01
1.63151419e+00 -2.42587113e+00 3.01908076e-01 3.30173708e-02
4.92000878e-01 1.82530880e-01 3.82393636e-02 8.79141450e-01
-5.11626229e-02 -2.51949161e-01 2.31175572e-01 -2.27735475e-01
-5.13245398e-03 -5.03279008e-02 -5.64268976e-02 6.47964954e-01
-2.87895709e-01 7.73533106e-01 -1.08732021e+00 -5.40153921e-01
3.22446704e-01 3.07573974e-01 -4.30625439e-01 2.01182425e-01
4.65211272e-02 5.00505686e-01 -2.69288719e-01 5.91554165e-01
1.79838330e-01 -1.78497136e-01 4.90213275e-01 -3.18853445e-02
-9.49672461e-02 1.02216139e-01 -1.48700273e+00 1.66439319e+00
5.95635436e-02 1.02103591e+00 -7.42290393e-02 -1.05433679e+00
5.78528404e-01 3.53837281e-01 7.99848258e-01 -7.41159916e-01
1.69817973e-02 -8.92263874e-02 -2.73724319e-03 -5.23996115e-01
6.78412616e-01 3.08683485e-01 -3.30887645e-01 4.41568762e-01
9.68411192e-02 3.30160081e-01 8.46158028e-01 4.95608002e-01
1.54243875e+00 1.79524094e-01 3.76648307e-01 1.67037308e-01
2.50137717e-01 -3.35826576e-02 5.14007688e-01 7.26811051e-01
-2.82732904e-01 6.31333292e-01 8.99386764e-01 -4.24531579e-01
-7.85272837e-01 -8.25775683e-01 5.53795755e-01 1.39761412e+00
4.89451051e-01 -9.87022758e-01 -6.67540729e-01 -3.49252850e-01
-1.23174183e-01 1.78370371e-01 -7.55055845e-01 -1.47904843e-01
-7.95907676e-01 -2.38024354e-01 3.95548463e-01 6.46830559e-01
4.90672678e-01 -1.14093864e+00 -6.28097951e-01 2.45058507e-01
-6.06921673e-01 -1.43010068e+00 -6.16716325e-01 -2.85655856e-01
-4.08021778e-01 -1.46719134e+00 -6.33254349e-01 -5.28819621e-01
1.02400437e-01 3.33829641e-01 1.17872500e+00 1.00852109e-01
-4.18362409e-01 5.31279743e-01 -7.80144215e-01 -1.92163840e-01
-6.75209835e-02 -1.08287200e-01 -9.56673175e-02 3.55803460e-01
1.94556445e-01 -5.07916868e-01 -6.94343865e-01 7.35656798e-01
-6.45642638e-01 2.21844405e-01 1.49264261e-01 4.10785884e-01
5.89370549e-01 1.57599285e-01 -6.26540631e-02 -5.10581553e-01
4.56601769e-01 -4.95620430e-01 -3.60259682e-01 2.61023998e-01
-8.95683616e-02 -4.72062379e-01 3.85721505e-01 -5.41573644e-01
-5.97130597e-01 3.30177397e-01 1.24958768e-01 -6.40347600e-01
-7.08251074e-02 4.22144264e-01 -2.36848332e-02 -3.73101495e-02
5.36780238e-01 3.26756954e-01 -2.30153516e-01 -3.80111009e-01
6.42837435e-02 3.68844509e-01 6.38354659e-01 -3.48667800e-01
4.70555007e-01 5.86558759e-01 -5.57458885e-02 -7.62751698e-01
-6.24156713e-01 -9.98628199e-01 -6.41464472e-01 -1.10929310e+00
7.93509960e-01 -9.18055654e-01 -1.14912927e+00 6.78782463e-01
-8.14685106e-01 -4.46712226e-01 -1.33763924e-01 5.13382256e-01
-6.54213250e-01 5.12980878e-01 -5.82006872e-01 -8.76762867e-01
1.62929609e-01 -9.28268969e-01 1.03466308e+00 1.85370922e-01
-2.99061030e-01 -7.61271894e-01 2.02423587e-01 4.71312314e-01
-3.76967043e-02 6.01994574e-01 1.62689924e-01 -5.39610207e-01
-4.74834830e-01 -5.82787454e-01 -8.18112791e-02 -1.83294192e-01
-1.11899547e-01 2.05918364e-02 -6.41131580e-01 -2.60304689e-01
-4.45879281e-01 -1.27402857e-01 8.48867357e-01 3.86761934e-01
1.14938498e+00 8.38944316e-02 -2.76669621e-01 4.16466027e-01
1.03874552e+00 3.67407084e-01 6.95063055e-01 1.36793301e-01
7.20742226e-01 4.60190684e-01 8.80063832e-01 8.89251411e-01
3.13640445e-01 1.24658787e+00 6.57007098e-01 1.74024731e-01
-1.12899631e-01 -4.97393578e-01 5.46034515e-01 6.39097273e-01
-7.62172997e-01 -3.56968671e-01 -1.10507989e+00 5.20770788e-01
-2.48888755e+00 -1.48410904e+00 -1.62815601e-01 2.06069303e+00
4.00598198e-01 2.03449741e-01 8.32147837e-01 4.82754439e-01
1.06205869e+00 5.20449400e-01 -4.03464139e-01 1.28981948e-01
4.89116758e-02 -1.54722128e-02 4.71136808e-01 1.85668066e-01
-1.24312294e+00 9.50054049e-01 5.54082966e+00 1.23948252e+00
-7.96258986e-01 1.01456583e-01 3.05471599e-01 -3.53916258e-01
4.45682198e-01 -1.63506821e-01 -5.55093110e-01 6.11846149e-01
6.06923878e-01 -3.70837510e-01 5.42534053e-01 6.96990967e-01
3.24407697e-01 -3.45596582e-01 -9.33299005e-01 1.32846379e+00
2.81769514e-01 -1.32880807e+00 -2.77434707e-01 -1.06927961e-01
3.41474563e-01 -1.83322147e-01 -1.96816117e-01 2.65829116e-01
3.00803393e-01 -1.05660176e+00 1.08428490e+00 6.12054825e-01
6.93020046e-01 -7.03156292e-01 6.06933594e-01 2.73934036e-01
-1.91035390e+00 -1.35763779e-01 -3.61761637e-02 -5.17591894e-01
3.35236758e-01 9.95376408e-02 -5.33040643e-01 6.04799390e-01
9.03122842e-01 1.20728791e+00 -5.16664982e-01 1.43001366e+00
-4.57601994e-01 7.42588401e-01 -1.50738895e-01 -1.34928524e-01
7.28400797e-02 -1.25807792e-01 6.79663360e-01 1.15601158e+00
2.09972501e-01 2.37672895e-01 5.18006623e-01 3.26303422e-01
1.32738069e-01 1.86969966e-01 -3.39667439e-01 -4.25228417e-01
2.61602819e-01 1.22760701e+00 -1.14494514e+00 -1.60753652e-01
-3.80662560e-01 8.92503560e-01 2.36241505e-01 1.13846488e-01
-1.12989748e+00 -3.15431058e-01 7.41781533e-01 3.83000076e-01
3.40678662e-01 -4.15051788e-01 6.54624701e-02 -1.15344524e+00
9.38338563e-02 -7.82478631e-01 8.56764436e-01 -7.94802487e-01
-8.34818482e-01 4.36947733e-01 -6.62619844e-02 -1.46406221e+00
-1.64624706e-01 -5.40213346e-01 -6.50155962e-01 2.50690430e-01
-6.13695383e-01 -8.71388257e-01 -7.00381875e-01 8.67988646e-01
6.44590735e-01 -1.16425224e-01 3.45284700e-01 4.68678832e-01
-6.21451437e-01 3.63071471e-01 -2.24594116e-01 5.25497377e-01
4.67781305e-01 -1.14944935e+00 3.72152627e-01 9.90018785e-01
6.51326180e-01 -1.93301126e-01 8.81727636e-01 -5.66911280e-01
-1.37308562e+00 -6.26202464e-01 6.45248890e-01 -4.02329892e-01
1.03923023e+00 -5.90884328e-01 -6.31826162e-01 5.79807222e-01
-2.84645278e-02 -1.09340958e-01 4.57462817e-01 -1.91042828e-03
-1.34822100e-01 5.28156161e-02 -4.31095660e-01 7.10584939e-01
1.46070683e+00 -6.26155376e-01 -1.40464172e-01 4.82944906e-01
1.43377587e-01 -6.75600350e-01 -6.42029583e-01 -2.29595192e-02
5.96194386e-01 -1.06008923e+00 9.82157946e-01 -6.29489124e-01
5.31329215e-01 -3.06178063e-01 3.51549834e-02 -1.12997401e+00
-3.40148985e-01 -6.76135361e-01 -1.44014403e-01 8.95444572e-01
-5.02932854e-02 -9.09754634e-02 9.95660365e-01 3.29457521e-01
-4.22740309e-03 -4.97220457e-01 -1.10581851e+00 -7.50134170e-01
-6.96596324e-01 -5.80781460e-01 2.32448503e-01 8.63049686e-01
5.36848843e-01 5.82776666e-02 -6.91301167e-01 -1.07758150e-01
4.46859539e-01 2.34450936e-01 9.49443936e-01 -9.77326334e-01
-4.45497096e-01 -5.36395431e-01 -1.06470346e+00 -9.78087842e-01
-1.00873532e-02 -8.12757969e-01 -1.08886473e-01 -1.44661033e+00
4.21320111e-01 1.40539035e-01 -3.31610262e-01 5.74579418e-01
-1.65357962e-01 5.07202327e-01 5.01684844e-01 3.06224167e-01
-1.25969458e+00 1.97316229e-01 1.05039454e+00 1.41026163e-02
-9.48938578e-02 1.70285717e-01 -1.62257865e-01 7.34555125e-01
8.15282822e-01 -3.79366487e-01 -1.41296238e-01 -7.36328587e-02
4.63923097e-01 4.60936457e-01 5.50504267e-01 -1.23026776e+00
5.79507232e-01 -3.43499213e-01 -1.25134423e-01 -3.20044667e-01
6.16848290e-01 -4.51966017e-01 4.13230479e-01 5.15050471e-01
-2.38323197e-01 2.54169405e-01 9.95251462e-02 8.24399710e-01
-5.52872241e-01 3.30439769e-02 4.87919718e-01 -1.42498031e-01
-1.41024470e+00 3.80186379e-01 -5.61646640e-01 2.80475557e-01
1.27969825e+00 -3.19785267e-01 -1.43824860e-01 -8.25755060e-01
-1.10756338e+00 1.69216111e-01 3.07837039e-01 5.42604446e-01
5.79061091e-01 -1.33687353e+00 -7.97468364e-01 -2.29470894e-01
2.72166461e-01 -6.25626445e-01 5.63651800e-01 1.14336228e+00
-6.99302554e-01 5.46248257e-02 -3.86918038e-01 -6.03433967e-01
-1.82387233e+00 3.42186600e-01 2.85236776e-01 -3.75480741e-01
-7.09718406e-01 8.57506812e-01 7.56104887e-02 2.51443088e-01
2.84363955e-01 2.51646876e-01 -5.74478447e-01 3.67802411e-01
7.53312528e-01 8.04390192e-01 -9.45489258e-02 -8.84878159e-01
-5.17249942e-01 4.05372113e-01 2.68540114e-01 5.52434660e-02
1.29513967e+00 1.23672903e-01 1.95484146e-01 5.88909507e-01
7.60788500e-01 -8.88830237e-03 -1.40153491e+00 -2.93416321e-01
9.52213183e-02 -8.04786444e-01 -1.25944659e-01 -4.18134183e-01
-1.31075346e+00 8.50937724e-01 2.28157371e-01 4.78379726e-01
1.08094299e+00 2.21793920e-01 6.02625608e-01 -1.09833561e-01
5.04135072e-01 -1.05239952e+00 3.94343019e-01 4.77363288e-01
8.78471613e-01 -1.01658356e+00 -2.05290481e-03 -5.52522838e-01
-8.57154310e-01 1.04715300e+00 4.18961287e-01 -2.64324278e-01
4.66987580e-01 3.69056374e-01 -1.36346206e-01 -3.91319185e-01
-7.25322723e-01 -4.23247784e-01 5.09598613e-01 3.71724397e-01
1.72576129e-01 2.67509043e-01 -2.91807115e-01 6.08112216e-01
-2.24146754e-01 -2.38602180e-02 4.79481071e-01 8.53641629e-01
-3.35277945e-01 -7.92241931e-01 -1.01885952e-01 2.55215019e-01
-5.37230015e-01 1.52121201e-01 -6.04719102e-01 7.63291657e-01
-8.53648931e-02 1.04241335e+00 1.92671984e-01 -7.15621412e-01
6.19300961e-01 -1.03929818e-01 4.38446790e-01 -5.13912916e-01
-6.38454080e-01 -1.16077147e-01 2.12386802e-01 -1.06405413e+00
-4.66386914e-01 -9.44515944e-01 -9.16766465e-01 -6.55016959e-01
-8.33434314e-02 1.51612222e-01 3.15898806e-01 7.17714667e-01
2.51331866e-01 4.64072853e-01 3.68627876e-01 -1.05289829e+00
6.01717047e-02 -7.50906169e-01 -9.27481294e-01 6.90945387e-01
-2.95939654e-01 -9.53745484e-01 -3.91359299e-01 2.80186255e-02] | [7.906529903411865, 0.14123082160949707] |
72d14a7f-8ffe-49ed-b941-e6137cf64a84 | resset-a-recurrent-model-for-sequence-of-sets | 1802.00948 | null | http://arxiv.org/abs/1802.00948v1 | http://arxiv.org/pdf/1802.00948v1.pdf | Resset: A Recurrent Model for Sequence of Sets with Applications to Electronic Medical Records | Modern healthcare is ripe for disruption by AI. A game changer would be
automatic understanding the latent processes from electronic medical records,
which are being collected for billions of people worldwide. However, these
healthcare processes are complicated by the interaction between at least three
dynamic components: the illness which involves multiple diseases, the care
which involves multiple treatments, and the recording practice which is biased
and erroneous. Existing methods are inadequate in capturing the dynamic
structure of care. We propose Resset, an end-to-end recurrent model that reads
medical record and predicts future risk. The model adopts the algebraic view in
that discrete medical objects are embedded into continuous vectors lying in the
same space. We formulate the problem as modeling sequences of sets, a novel
setting that have rarely, if not, been addressed. Within Resset, the bag of
diseases recorded at each clinic visit is modeled as function of sets. The same
hold for the bag of treatments. The interaction between the disease bag and the
treatment bag at a visit is modeled in several, one of which as residual of
diseases minus the treatments. Finally, the health trajectory, which is a
sequence of visits, is modeled using a recurrent neural network. We report
results on over a hundred thousand hospital visits by patients suffered from
two costly chronic diseases -- diabetes and mental health. Resset shows
promises in multiple predictive tasks such as readmission prediction,
treatments recommendation and diseases progression. | ['Truyen Tran', 'Svetha Venkatesh', 'Phuoc Nguyen'] | 2018-02-03 | null | null | null | null | ['readmission-prediction'] | ['medical'] | [ 3.82866889e-01 3.81009221e-01 -2.10740834e-01 -1.71694994e-01
-6.19064510e-01 -3.18070918e-01 3.70489478e-01 3.94018233e-01
-3.29259336e-02 7.32825637e-01 1.07605386e+00 -1.91221565e-01
-5.76960444e-01 -7.87503183e-01 -7.97772229e-01 -9.45097446e-01
-2.00611547e-01 1.20692778e+00 -6.12421989e-01 -7.79403746e-02
-4.31212187e-01 -1.03289731e-01 -1.10770977e+00 5.53978562e-01
6.49003804e-01 6.74673021e-01 -6.66219881e-03 6.41912758e-01
7.03216121e-02 1.21264005e+00 -3.66536558e-01 -4.61471081e-01
-5.26347421e-02 -5.47740519e-01 -5.50840259e-01 -6.34323806e-02
-3.88840646e-01 -2.96867013e-01 -3.99956405e-01 6.91045403e-01
4.64267999e-01 -3.37561131e-01 9.38101411e-01 -1.02587497e+00
-6.91797495e-01 8.67222846e-01 -3.48660350e-01 -1.10629104e-01
4.73738432e-01 3.14265341e-02 8.60371649e-01 -3.51168603e-01
4.67406899e-01 1.23827958e+00 1.00594318e+00 7.90807128e-01
-1.41412461e+00 -1.75554410e-01 1.47623196e-01 -1.71461552e-01
-9.28231597e-01 -2.89662451e-01 3.38910103e-01 -9.01605785e-01
1.01161742e+00 6.40156448e-01 8.72271717e-01 1.70322192e+00
9.04167295e-01 6.74575388e-01 5.80077469e-01 -6.76487759e-02
3.80511492e-01 -3.84568982e-02 6.27079248e-01 2.84604281e-01
2.92185664e-01 5.44137135e-02 -2.03062162e-01 -9.33509111e-01
4.44220215e-01 1.19722688e+00 -3.60732526e-01 -2.34828405e-02
-1.07090259e+00 8.05022538e-01 3.98985081e-04 1.29975706e-01
-1.07761264e+00 1.13694638e-01 4.20293152e-01 -1.49843786e-02
7.76973426e-01 1.24422267e-01 -7.45397866e-01 -3.40126336e-01
-7.06622899e-01 2.19051450e-01 1.01980948e+00 6.64157271e-01
5.34630660e-03 -4.35726911e-01 -3.42093080e-01 5.99004090e-01
1.27567366e-01 5.83239973e-01 7.31816590e-01 -3.58473331e-01
3.52037638e-01 7.52606213e-01 2.59200603e-01 -1.15857708e+00
-7.67765999e-01 -4.00081128e-01 -1.60399902e+00 -6.97911084e-01
1.25657290e-01 -3.68321627e-01 -1.16674542e+00 1.82195890e+00
6.51031360e-02 5.15694022e-01 2.26149783e-01 2.22324088e-01
6.25526011e-01 7.76787221e-01 1.08102880e-01 -7.89467692e-01
1.53987920e+00 -6.55599415e-01 -1.16282058e+00 -4.26636972e-02
1.00263786e+00 -1.49554789e-01 4.46800411e-01 3.94364893e-01
-1.48490548e+00 -3.01720407e-02 -3.07482004e-01 1.27689674e-01
-1.68787897e-01 -1.12939730e-01 4.91386592e-01 1.98475555e-01
-1.15091753e+00 4.18381572e-01 -1.21559429e+00 -5.20944238e-01
4.51548010e-01 3.76330912e-01 6.48666695e-02 -1.67183444e-01
-1.16914713e+00 4.70276862e-01 -1.52166650e-01 2.04557762e-01
-6.42691433e-01 -9.77936625e-01 -5.15097797e-01 3.22608441e-01
4.37885582e-01 -1.46124721e+00 1.09790540e+00 -7.29890406e-01
-8.23873699e-01 6.17973745e-01 -5.46231627e-01 -6.29327536e-01
4.59041595e-01 -2.86776692e-01 -5.05821645e-01 -4.80145454e-01
3.70809659e-02 -2.16161862e-01 5.49632490e-01 -9.09972489e-01
-5.07643282e-01 -7.48050451e-01 -5.19081116e-01 2.79830784e-01
-1.92926839e-01 -3.10106456e-01 -2.12964922e-01 -6.11223638e-01
-1.04577996e-01 -1.16229188e+00 -5.71773231e-01 -7.43618727e-01
-6.07552588e-01 -1.74714506e-01 -9.85002238e-03 -8.88945878e-01
1.79998922e+00 -2.17246127e+00 4.94944930e-01 -3.19619179e-02
6.76112175e-01 -2.52416432e-01 1.72934487e-01 7.52332866e-01
-3.82253319e-01 3.05777788e-01 -1.72547877e-01 -5.65975130e-01
-3.10212463e-01 3.68352324e-01 -5.97457767e-01 3.70550573e-01
-1.60114944e-01 1.19127584e+00 -9.41022992e-01 8.84855315e-02
-4.14613008e-01 4.18626726e-01 -6.36751652e-01 2.06511170e-01
-1.64074093e-01 1.31031111e-01 -6.22560740e-01 4.61458504e-01
3.10778379e-01 -8.70066047e-01 2.60496765e-01 1.89971909e-01
3.46524388e-01 4.32106793e-01 -7.96929240e-01 1.34213293e+00
1.18891224e-01 -2.21826971e-01 -3.48865211e-01 -8.67218316e-01
2.74763763e-01 7.94725776e-01 1.15198445e+00 -3.61024171e-01
-7.73330107e-02 -1.56896606e-01 -2.24149064e-03 -6.64203644e-01
1.97018191e-01 -4.04122949e-01 -3.71257395e-01 5.57583094e-01
-4.41521406e-01 4.65336353e-01 -3.74329120e-01 1.93195015e-01
1.70703948e+00 -5.13269961e-01 5.06906986e-01 6.10541888e-02
-2.15586826e-01 7.75608867e-02 9.64384675e-01 1.01015258e+00
3.91235612e-02 6.43340647e-01 6.49033189e-01 -8.07939470e-01
-8.26718032e-01 -1.16627049e+00 -1.20297141e-01 5.46624303e-01
-2.73402452e-01 -5.02781928e-01 -5.31203866e-01 -3.12080115e-01
1.31300539e-01 5.74558735e-01 -1.16712618e+00 -5.40583313e-01
-4.32286769e-01 -1.39085257e+00 3.46812725e-01 4.43205416e-01
-3.10009092e-01 -1.21255040e+00 -4.11983758e-01 6.66496813e-01
-3.64821404e-01 -5.38860500e-01 -6.50423288e-01 1.98494315e-01
-8.91837776e-01 -1.04919565e+00 -6.73848391e-01 -5.74414670e-01
4.53057915e-01 -1.40509456e-01 1.42461550e+00 -8.40427205e-02
-5.44701040e-01 7.28909254e-01 -6.96317554e-02 -6.94696188e-01
-4.23933208e-01 -1.19198963e-01 5.13981879e-01 2.23956212e-01
5.42302728e-01 -5.14073253e-01 -7.82205939e-01 -2.86137760e-01
-1.03426838e+00 3.14084262e-01 2.75075644e-01 9.63666201e-01
6.16016150e-01 2.71864757e-02 5.30162394e-01 -1.36988485e+00
5.64594150e-01 -1.23962772e+00 1.88536510e-01 4.93911088e-01
-6.36385977e-01 -2.06222504e-01 3.87395203e-01 -4.88796443e-01
-7.43545413e-01 -2.39079967e-02 2.22453892e-01 -2.67964453e-01
-2.51155347e-01 7.88400590e-01 7.39273876e-02 1.38295007e+00
3.27128917e-01 3.11956018e-01 3.54159847e-02 -4.66098636e-01
7.22520351e-02 7.87692070e-01 2.35147998e-01 6.24694228e-02
2.05767527e-01 4.95925099e-01 -2.56079137e-01 -5.93647897e-01
-8.55249166e-01 -5.47204673e-01 -1.27911404e-01 1.83942303e-01
1.20865440e+00 -1.10981393e+00 -9.43641663e-01 5.19129694e-01
-1.10308659e+00 -2.42049173e-01 -6.25077426e-01 4.28263307e-01
-5.73154986e-01 -3.23976457e-01 -1.12783396e+00 -1.02863431e+00
-4.93048847e-01 -9.96413171e-01 1.29596269e+00 -9.85725448e-02
-6.96934342e-01 -1.40413988e+00 5.48353732e-01 1.89433977e-01
4.96414490e-02 3.37488562e-01 1.39612138e+00 -7.76002586e-01
-3.97970736e-01 -3.42234790e-01 3.53419751e-01 -1.64041296e-01
8.29060435e-01 -2.85711735e-01 -5.15425086e-01 -3.84257615e-01
5.38814664e-01 1.82123512e-01 6.72847688e-01 1.09562504e+00
1.11132514e+00 -6.81762099e-01 -9.29573178e-01 5.79947114e-01
1.16877806e+00 8.62122715e-01 6.77588880e-01 -1.89500093e-01
9.51187432e-01 6.35861754e-01 6.10272028e-02 7.20711231e-01
6.51535332e-01 3.82267892e-01 2.14645058e-01 -3.64182442e-01
6.23211443e-01 -2.00435445e-01 3.77177238e-01 1.22223413e+00
1.38627127e-01 -6.53709829e-01 -1.34134912e+00 6.33442640e-01
-2.27664447e+00 -1.30075514e+00 -2.50859350e-01 2.08603287e+00
8.82940948e-01 -1.12081148e-01 2.46825144e-01 -2.02957794e-01
3.80514771e-01 -2.62154490e-01 -8.44301820e-01 -3.15635979e-01
1.67469252e-02 -1.90031394e-01 3.07472140e-01 3.26164514e-01
-8.23499203e-01 3.77624184e-01 6.43732023e+00 1.66563645e-01
-7.32896864e-01 -4.87205014e-02 1.21682250e+00 -4.14262116e-01
-4.42280442e-01 -3.17355841e-01 -8.04558516e-01 7.95363069e-01
1.32030225e+00 -2.44430512e-01 4.42468554e-01 2.92160064e-01
5.51561236e-01 1.55942291e-01 -1.51752245e+00 1.02570033e+00
2.62165572e-02 -1.18798482e+00 2.42586419e-01 4.44256872e-01
7.79112399e-01 9.03918520e-02 3.48526150e-01 1.56449705e-01
7.92103350e-01 -1.19093955e+00 2.19461441e-01 1.36747289e+00
6.64659500e-01 -4.72334892e-01 6.66904986e-01 7.30341494e-01
-6.75912499e-01 -3.74717385e-01 9.86587182e-02 -1.01417296e-01
3.64178181e-01 7.87779331e-01 -8.55437398e-01 4.91704613e-01
8.52448761e-01 1.27558112e+00 -6.47948980e-02 6.30373776e-01
6.96634293e-01 9.26642239e-01 2.58970521e-02 4.21029776e-01
-4.07276154e-02 -4.75781173e-01 4.11296666e-01 1.00490677e+00
5.38375556e-01 4.51347917e-01 2.25291863e-01 5.45980096e-01
1.08134290e-02 -1.62890747e-01 -1.08629584e+00 -2.17099771e-01
-6.50844648e-02 6.80768907e-01 -4.24352735e-01 -5.79332113e-01
-2.45760754e-01 1.02855945e+00 3.97262275e-02 5.45611680e-01
-7.08886683e-01 4.22093093e-01 9.47530806e-01 3.29234660e-01
-9.34623107e-02 4.95145559e-01 -3.89489383e-01 -1.37904537e+00
-3.57313640e-02 -1.01622462e+00 5.17545581e-01 -6.23086691e-01
-1.70394158e+00 6.00188673e-01 -2.24415839e-01 -1.06368601e+00
-7.43241251e-01 -9.06721577e-02 -1.99929699e-01 9.30985928e-01
-8.24604452e-01 -6.91082537e-01 9.98374373e-02 5.56340158e-01
4.55941528e-01 -6.90211058e-02 1.23211813e+00 2.10842311e-01
-9.41447973e-01 2.14092240e-01 5.76331079e-01 -7.06392601e-02
4.64243799e-01 -1.16273081e+00 3.82614613e-01 1.64979428e-01
-3.15397441e-01 6.45992994e-01 6.25829041e-01 -1.18724966e+00
-1.23078454e+00 -1.25234556e+00 1.37025297e+00 -9.82841790e-01
5.04918933e-01 -3.86116058e-01 -1.11527622e+00 1.09524345e+00
-2.82380909e-01 -4.63820606e-01 9.99139547e-01 1.13877922e-01
2.00651452e-01 1.14821672e-01 -8.83304775e-01 7.01920450e-01
1.09624469e+00 -5.26190579e-01 -5.60300708e-01 6.25727415e-01
1.05137122e+00 -2.08529532e-01 -7.32010305e-01 4.55473870e-01
4.16458189e-01 -4.78604585e-01 9.87099349e-01 -1.32743382e+00
9.02557850e-01 3.69747430e-01 -2.25640051e-02 -1.43466735e+00
-6.88886225e-01 -6.06267571e-01 -3.77108514e-01 9.92508411e-01
2.93427914e-01 -7.19250977e-01 6.41768277e-01 9.42727268e-01
1.91285118e-01 -1.02160954e+00 -6.38783336e-01 -3.54834884e-01
-1.36635333e-01 -8.93498585e-02 7.39199877e-01 1.14859331e+00
1.47910833e-01 5.81732035e-01 -7.37259567e-01 1.11213438e-01
3.72613847e-01 -3.54876518e-02 2.64762044e-01 -1.61041403e+00
-6.49536908e-01 -2.08754271e-01 -3.57187092e-02 -6.88512921e-01
-2.04526618e-01 -6.82468832e-01 -6.70981407e-03 -1.83105946e+00
8.25890124e-01 -3.83717626e-01 -5.30850708e-01 4.09414560e-01
-4.03517127e-01 -5.95672190e-01 -1.37714431e-01 5.31712472e-01
-3.44052643e-01 4.10413414e-01 7.35173821e-01 -1.83818668e-01
-6.94840193e-01 4.03795809e-01 -9.42653894e-01 9.44156706e-01
5.40817201e-01 -8.72850478e-01 -4.78341252e-01 -2.58808941e-01
6.66937292e-01 8.48798096e-01 6.76681697e-02 -4.56687689e-01
1.95464626e-01 -2.28091478e-01 1.43714741e-01 -5.34091890e-01
4.73937660e-01 -9.93660748e-01 8.19036245e-01 7.95872152e-01
-5.37409246e-01 7.06428945e-01 -1.93371177e-02 1.07624197e+00
7.40408599e-02 4.10811305e-01 -3.71575989e-02 -3.01037848e-01
4.24963772e-01 6.34255409e-01 -6.23385787e-01 1.58710852e-02
9.11690652e-01 1.90187275e-01 -1.06933787e-01 -5.48380375e-01
-1.21987295e+00 3.15049350e-01 9.19825360e-02 3.47824514e-01
6.25735343e-01 -1.12481189e+00 -7.03774393e-01 1.28415957e-01
-6.80429861e-02 1.88753545e-01 5.46232164e-01 9.20491874e-01
4.65040840e-02 4.04407740e-01 3.52828920e-01 -4.84191895e-01
-1.36992991e+00 1.03865790e+00 3.27275902e-01 -9.95312989e-01
-7.54612923e-01 4.85984713e-01 6.91109776e-01 -1.25796154e-01
4.69279468e-01 -5.87707222e-01 -4.13413495e-01 4.49763149e-01
7.53648221e-01 2.82011688e-01 -2.22885266e-01 -4.73504335e-01
-2.00238481e-01 2.56559551e-01 -4.15658146e-01 1.24218889e-01
1.92550838e+00 -2.83538103e-01 -4.19072956e-01 1.31886148e+00
1.00826550e+00 -4.18960989e-01 -6.38519526e-01 -3.22409362e-01
9.68835503e-03 3.87081712e-01 -2.39214405e-01 -1.07461393e+00
-6.16079569e-01 5.00845730e-01 6.12219274e-01 6.24484181e-01
1.17058313e+00 5.47919646e-02 9.03302431e-01 -1.04027823e-01
2.33023599e-01 -4.19699460e-01 -2.82235175e-01 4.00445879e-01
7.13874280e-01 -9.77514684e-01 -3.94592792e-01 -1.00073315e-01
-5.85930824e-01 4.65962768e-01 1.20358076e-03 -8.62495787e-03
1.08503711e+00 3.04885715e-01 -1.54146478e-01 -4.31140929e-01
-1.37296259e+00 2.39742935e-01 2.14973882e-01 9.95000675e-02
3.08640242e-01 6.62657917e-01 1.26453852e-02 1.13218272e+00
-1.34697467e-01 3.94889683e-01 6.21794343e-01 7.55639076e-01
3.60470086e-01 -7.92649150e-01 -3.11291277e-01 1.16779184e+00
-7.65120685e-01 -4.10583168e-01 -2.28004918e-01 1.23387732e-01
1.67811275e-01 7.44474351e-01 4.03511375e-01 -4.80290502e-01
5.64831734e-01 4.06628489e-01 -1.23831868e-01 -8.45963478e-01
-7.86375880e-01 1.08329274e-01 -2.35408336e-01 -5.04896939e-01
-5.19707277e-02 -1.02368534e+00 -1.07905281e+00 -3.14105034e-01
1.39116511e-01 9.90640819e-02 1.91644445e-01 9.96615589e-01
6.63012207e-01 1.04829693e+00 4.66881096e-01 -3.19349647e-01
-6.61378205e-01 -7.79897153e-01 -8.01221669e-01 6.44588411e-01
9.43342984e-01 -2.87441224e-01 -3.72268736e-01 3.67084771e-01] | [7.831291198730469, 6.132724761962891] |
b4289162-02b4-4866-b600-fdc873438408 | learning-local-descriptors-by-optimizing-the | 1603.09095 | null | https://arxiv.org/abs/1603.09095v6 | https://arxiv.org/pdf/1603.09095v6.pdf | Learning Local Descriptors by Optimizing the Keypoint-Correspondence Criterion: Applications to Face Matching, Learning from Unlabeled Videos and 3D-Shape Retrieval | Current best local descriptors are learned on a large dataset of matching and non-matching keypoint pairs. However, data of this kind is not always available since detailed keypoint correspondences can be hard to establish. On the other hand, we can often obtain labels for pairs of keypoint bags. For example, keypoint bags extracted from two images of the same object under different views form a matching pair, and keypoint bags extracted from images of different objects form a non-matching pair. On average, matching pairs should contain more corresponding keypoints than non-matching pairs. We describe an end-to-end differentiable architecture that enables the learning of local keypoint descriptors from such weakly-labeled data. Additionally, we discuss how to improve the method by incorporating the procedure of mining hard negatives. We also show how can our approach be used to learn convolutional features from unlabeled video signals and 3D models. Our implementation is available at https://github.com/nenadmarkus/wlrn | ['Jörgen Ahlberg', 'Igor S. Pandžić', 'Nenad Markuš'] | 2016-03-30 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [-3.20129007e-01 -2.02465042e-01 -3.49843413e-01 -4.81654584e-01
-1.18731892e+00 -6.71471238e-01 3.16519648e-01 3.05444986e-01
-2.51121819e-01 5.54099798e-01 5.59587665e-02 2.57876337e-01
-1.29936382e-01 -6.94641113e-01 -1.20620775e+00 -5.99432826e-01
-1.95435718e-01 3.62756848e-01 2.83253998e-01 2.56550252e-01
1.69023737e-01 7.16091812e-01 -1.39070761e+00 2.84386605e-01
2.79080898e-01 1.29207551e+00 5.69034100e-01 4.81047392e-01
-9.61499885e-02 3.95336241e-01 -2.60381103e-01 -1.52214661e-01
7.49128759e-01 -1.56532511e-01 -5.94447136e-01 7.24510103e-02
1.25533772e+00 -6.36975527e-01 -5.35818338e-01 1.07308066e+00
3.06819141e-01 1.22489832e-01 3.02041382e-01 -1.47367799e+00
-7.10263252e-01 2.66770482e-01 -5.51832676e-01 1.71413273e-01
3.07785422e-01 -2.15150341e-01 1.22652745e+00 -1.08859313e+00
7.39663959e-01 1.04029238e+00 4.78295058e-01 3.82499397e-01
-1.07283783e+00 -8.79714072e-01 3.35732192e-01 2.69753963e-01
-1.42736518e+00 -4.84255731e-01 8.49932492e-01 -2.56512374e-01
5.51834166e-01 7.80770630e-02 8.43164802e-01 8.34028959e-01
-6.62288666e-02 1.05436122e+00 8.32566977e-01 -2.26830244e-01
2.06444412e-02 -2.28841826e-02 1.68276787e-01 9.34056878e-01
7.54632950e-02 1.95624113e-01 -6.87033534e-01 -3.69666219e-01
1.02930856e+00 6.60293102e-01 -4.44883764e-01 -1.00834513e+00
-1.55085158e+00 7.07621396e-01 7.57964849e-01 1.97271794e-01
-3.06161910e-01 4.18126494e-01 1.66572064e-01 3.98671240e-01
2.82124549e-01 2.18802735e-01 -5.08355319e-01 3.51681225e-02
-7.83545852e-01 5.36644638e-01 5.51253438e-01 1.36146367e+00
1.23683941e+00 -5.80367565e-01 1.54264212e-01 6.52918518e-01
2.57736623e-01 4.63279039e-01 2.17763260e-01 -1.01465642e+00
4.15877670e-01 4.88866925e-01 2.76050538e-01 -1.17064595e+00
-1.57384217e-01 1.11260600e-01 -5.53542376e-01 1.13211237e-01
6.12524390e-01 2.58524150e-01 -9.19932067e-01 1.57144070e+00
3.00107330e-01 7.10786581e-01 -3.79776448e-01 1.02704751e+00
8.97602737e-01 5.02903581e-01 -3.15977097e-01 4.36739177e-02
1.16787159e+00 -1.02512085e+00 -3.41667771e-01 -2.58074254e-01
5.07792294e-01 -6.52132392e-01 6.75101578e-01 1.68408900e-01
-1.07657325e+00 -5.30377984e-01 -8.16678584e-01 -2.45407522e-01
-4.21581507e-01 1.00884460e-01 6.17761135e-01 -1.40525773e-01
-8.95351946e-01 6.43805087e-01 -9.72534776e-01 -3.02753031e-01
3.19904715e-01 4.05799389e-01 -8.32191944e-01 -2.56081671e-01
-9.16816235e-01 6.89736485e-01 1.12277947e-01 2.10500300e-01
-9.28485572e-01 -7.82767355e-01 -1.02563524e+00 -1.84501126e-01
4.38531280e-01 -4.16979700e-01 1.11630893e+00 -8.05812776e-01
-7.84263015e-01 1.12593651e+00 -2.62184709e-01 -2.19848186e-01
4.91367072e-01 -4.05052871e-01 -1.00184523e-01 3.62458169e-01
2.83217013e-01 9.07669067e-01 8.46026838e-01 -1.24748719e+00
-8.83131266e-01 -5.31670153e-01 2.66103506e-01 9.57428962e-02
-2.74775643e-02 -9.39364061e-02 -7.06922770e-01 -6.13149166e-01
4.32995379e-01 -9.52134550e-01 8.24616011e-03 6.96027637e-01
-2.95354694e-01 -1.99905634e-01 7.61845231e-01 -3.79142851e-01
6.64629817e-01 -2.26164627e+00 1.55651182e-01 2.23828763e-01
3.39864850e-01 -2.18513911e-03 -3.54875684e-01 2.71489531e-01
-1.32512540e-01 -1.82216287e-01 1.07790351e-01 -2.16571078e-01
3.71979289e-02 2.48175189e-01 -3.90264243e-01 6.44955158e-01
1.77552521e-01 9.32669044e-01 -1.16390061e+00 -3.95036668e-01
2.63980091e-01 1.82832226e-01 -3.34227979e-01 2.56064028e-01
-1.18309394e-01 1.80165887e-01 -5.55193007e-01 9.06666279e-01
6.62866473e-01 -3.89549941e-01 -2.56788522e-01 -5.11614799e-01
1.62309989e-01 5.18434756e-02 -1.42200756e+00 1.95164943e+00
-3.11831832e-01 6.14979744e-01 1.16176642e-01 -1.05768394e+00
8.19939435e-01 1.76374093e-01 7.18337953e-01 -4.94699478e-01
-4.28912230e-02 3.04173082e-01 -4.95117933e-01 -3.77880216e-01
5.17830670e-01 -2.60052532e-02 4.52356935e-02 5.03130436e-01
2.85360664e-01 -1.28518000e-01 1.19378977e-01 1.07861839e-01
1.01055801e+00 1.38045192e-01 3.20579141e-01 5.38058840e-02
2.41566896e-01 -2.28935838e-01 6.93246126e-01 7.33713388e-01
-3.74285519e-01 7.96514332e-01 2.35167265e-01 -6.72322750e-01
-1.00344753e+00 -1.21390629e+00 -3.06254655e-01 7.78895199e-01
5.79619765e-01 -5.78610063e-01 -3.25269669e-01 -8.34708095e-01
4.31980401e-01 -9.10160393e-02 -5.56370139e-01 1.39103895e-02
-5.22980213e-01 2.23672409e-02 1.75272703e-01 7.33800530e-01
2.54461169e-01 -7.80814230e-01 -4.94859517e-01 1.13968290e-01
-7.08089471e-02 -1.11442852e+00 -7.59165823e-01 4.20057714e-01
-8.23585033e-01 -1.37328672e+00 -8.94642472e-01 -9.97006297e-01
9.14829195e-01 7.89359510e-01 1.10715580e+00 3.86804312e-01
-2.49680430e-01 7.20357835e-01 -3.73281509e-01 -1.65109560e-01
3.30081359e-02 -2.81995356e-01 1.07451648e-01 -1.33559972e-01
7.37518549e-01 -4.40747499e-01 -6.91404581e-01 5.85780919e-01
-7.84012139e-01 -2.12363303e-01 4.20727968e-01 9.69307005e-01
1.10092390e+00 -2.63013631e-01 4.01764691e-01 -4.81070578e-01
1.00762546e-01 -4.51597393e-01 -7.69331098e-01 3.44945371e-01
-3.20736282e-02 -1.00007467e-02 3.35415721e-01 -5.48919797e-01
-2.39414051e-01 3.66155535e-01 2.15155438e-01 -1.12140989e+00
-3.62389952e-01 2.46023923e-01 -2.71347594e-02 -1.83510944e-01
3.48526776e-01 -7.67238112e-03 1.46049289e-02 -4.48509187e-01
4.82499361e-01 6.32031739e-01 4.23029363e-01 -7.07743764e-01
8.76258552e-01 8.02991211e-01 -1.70685306e-01 -6.63146079e-01
-8.40066373e-01 -8.53572786e-01 -9.32047665e-01 -4.68511460e-03
7.00820923e-01 -1.10158944e+00 -3.29195589e-01 3.64267528e-01
-1.18762636e+00 -3.19876552e-01 -3.27341944e-01 5.93574464e-01
-7.67623723e-01 4.18767154e-01 -4.53560770e-01 -3.84054035e-01
1.14794318e-02 -1.12834156e+00 1.51452601e+00 2.30009079e-01
-1.29758552e-01 -8.43345821e-01 1.34174312e-02 2.16073096e-01
-1.85419828e-01 3.40138972e-02 6.08293295e-01 -6.93167865e-01
-1.08892393e+00 -4.47164416e-01 -1.84967086e-01 3.51776421e-01
2.86902159e-01 -1.42124996e-01 -8.75098944e-01 -4.56977695e-01
-3.29809219e-01 -5.46432972e-01 8.71942461e-01 4.82670426e-01
1.07022250e+00 3.85198481e-02 -6.34859502e-01 6.95556104e-01
1.33859015e+00 5.93256839e-02 3.23118895e-01 2.77427018e-01
7.93862879e-01 5.61418831e-01 8.46752822e-01 1.61691666e-01
3.98593307e-01 7.33338594e-01 3.82451832e-01 -3.48025411e-02
8.65218043e-02 -3.65484059e-01 2.22317562e-01 6.27406776e-01
2.54857630e-01 2.09544122e-01 -7.85640955e-01 8.11314762e-01
-2.00059652e+00 -1.00566804e+00 1.37520567e-01 2.28264618e+00
6.02078915e-01 5.21709537e-03 3.37087400e-02 -2.31738970e-01
8.51817250e-01 1.06748097e-01 -6.22799218e-01 1.53516173e-01
6.15614057e-02 1.78069863e-02 5.65676868e-01 4.59612459e-01
-1.35570467e+00 8.20406795e-01 5.42416000e+00 4.82436687e-01
-9.91190016e-01 -3.39279771e-02 3.46265644e-01 -1.68339878e-01
-3.48473430e-01 1.96690872e-01 -9.66817200e-01 3.92060161e-01
3.32094371e-01 2.93440018e-02 2.21055970e-01 9.45402980e-01
-1.84728771e-01 -2.67637670e-01 -1.59861445e+00 1.40494132e+00
1.43476680e-01 -1.34794044e+00 -1.24313585e-01 1.19275779e-01
6.64633811e-01 2.39279732e-01 -1.60374269e-01 -2.30306690e-03
1.35970458e-01 -6.76737189e-01 7.51342475e-01 6.40316010e-01
5.60150921e-01 -5.86067319e-01 3.98240924e-01 3.34996283e-01
-1.29573834e+00 -1.65814608e-02 -7.30498910e-01 1.68911129e-01
-3.21290158e-02 4.91153538e-01 -4.68440861e-01 5.04575968e-01
9.02808964e-01 1.10287321e+00 -4.33990210e-01 1.32469475e+00
-1.47942394e-01 -8.41936767e-02 -4.59602982e-01 1.40059337e-01
2.38921300e-01 -3.72185051e-01 3.85061443e-01 8.48833263e-01
4.23003107e-01 9.67240557e-02 4.53427494e-01 8.60667527e-01
-1.98115349e-01 -2.36614421e-01 -1.01800275e+00 6.87003089e-03
7.22485244e-01 1.37239587e+00 -6.69230044e-01 -3.29048067e-01
-8.97722900e-01 9.61664617e-01 4.97015417e-01 2.91172057e-01
-6.68963432e-01 -3.45656514e-01 1.05439425e+00 1.55917436e-01
4.47331250e-01 -3.20304275e-01 3.73716682e-01 -1.39009249e+00
2.60897398e-01 -7.11410880e-01 3.87219161e-01 -1.00890577e+00
-1.61502016e+00 2.53616333e-01 1.23290986e-01 -1.53403950e+00
-2.22156480e-01 -8.05957854e-01 -3.94334197e-01 7.26430953e-01
-1.47148788e+00 -1.20013547e+00 -6.24390185e-01 7.95393765e-01
2.72923052e-01 3.71488333e-02 8.12285542e-01 2.87464887e-01
-1.88877583e-02 4.52481747e-01 2.53397167e-01 4.37973112e-01
9.03560698e-01 -1.20284545e+00 3.52492452e-01 5.99959850e-01
5.34451067e-01 6.05653226e-01 1.92059755e-01 -3.59921634e-01
-1.56214845e+00 -8.99775624e-01 6.26178503e-01 -5.58982313e-01
7.46965528e-01 -5.53047538e-01 -9.76590455e-01 9.45685863e-01
-2.76778460e-01 6.92942381e-01 6.15116656e-01 4.69471253e-02
-5.24673522e-01 -2.79611617e-01 -8.37917984e-01 5.87138593e-01
1.18779218e+00 -8.80140424e-01 -5.76938927e-01 4.82915014e-01
4.47692424e-01 -6.52539432e-01 -8.71146560e-01 1.80364192e-01
5.82040846e-01 -8.13656688e-01 1.12885845e+00 -6.16616547e-01
1.27136394e-01 -6.04255915e-01 -4.41163540e-01 -1.28022218e+00
-8.52734149e-02 -3.54313433e-01 -9.54095051e-02 9.94055033e-01
-3.30671296e-02 -4.33003902e-01 9.04826880e-01 6.33085787e-01
1.76291093e-02 -8.05112183e-01 -9.53746617e-01 -1.01901948e+00
6.99824616e-02 -4.18189824e-01 5.80151379e-01 9.31963086e-01
-1.80677976e-02 1.00466041e-02 -2.55327702e-01 3.33928674e-01
6.55219734e-01 6.89249575e-01 1.01047146e+00 -1.15510190e+00
-1.89462781e-01 -1.49644390e-01 -9.48616326e-01 -1.46098292e+00
5.03873467e-01 -1.04644084e+00 1.02243610e-01 -1.37389016e+00
2.78016776e-01 -8.01826000e-01 -3.66525382e-01 6.56624794e-01
5.77003472e-02 3.81091952e-01 3.16902727e-01 2.41096914e-01
-5.30628502e-01 2.99938738e-01 1.33048260e+00 -3.25932682e-01
-2.39810795e-02 -1.38837725e-01 -4.80214208e-01 7.11263835e-01
6.25150263e-01 -4.58757818e-01 -2.54174501e-01 -5.04916072e-01
1.78966209e-01 1.54005542e-01 8.22149217e-01 -7.76665211e-01
4.03975099e-01 -1.41087621e-01 5.62888384e-01 -6.94523454e-01
3.90525728e-01 -1.13146901e+00 6.48445934e-02 1.78293269e-02
-2.23133460e-01 9.72327068e-02 2.20508408e-02 7.72961557e-01
-4.19477731e-01 -3.06058198e-01 6.53348804e-01 -3.81244779e-01
-9.25122499e-01 7.18151867e-01 3.46610785e-01 -4.64636385e-02
1.25342274e+00 -4.49161291e-01 -2.22686246e-01 -3.09189379e-01
-6.99836254e-01 3.47679436e-01 7.53515005e-01 5.77259004e-01
9.82819676e-01 -1.65538359e+00 -4.15821850e-01 4.19129401e-01
4.14244324e-01 3.31194997e-01 3.28700133e-02 7.86915064e-01
-4.05646980e-01 3.00309956e-01 -3.05322081e-01 -8.32131803e-01
-1.41681218e+00 6.08436406e-01 4.94305789e-01 3.00959378e-01
-8.10763359e-01 8.99525225e-01 5.14563441e-01 -6.22117758e-01
5.02516568e-01 -5.13032675e-01 2.61431187e-01 1.68884411e-01
7.29000330e-01 9.04682726e-02 1.49826631e-01 -6.64798379e-01
-4.20059234e-01 9.22824204e-01 -4.21808153e-01 1.64556861e-01
1.46980071e+00 -5.47107719e-02 -1.50894627e-01 6.41823888e-01
1.69865322e+00 -8.64818022e-02 -1.52336025e+00 -5.76716304e-01
-8.92949477e-02 -8.82990837e-01 -6.71465695e-02 -2.65598863e-01
-1.19823694e+00 8.03093791e-01 6.58612251e-01 -3.07209603e-02
9.20042276e-01 3.95629048e-01 8.16869199e-01 6.46257639e-01
7.30613232e-01 -9.60524619e-01 1.52433410e-01 3.60290676e-01
8.61812532e-01 -1.43209243e+00 -1.18183671e-02 -1.90867856e-01
-3.16070110e-01 1.22578657e+00 5.26622772e-01 -4.58729237e-01
9.67685759e-01 1.49760216e-01 1.88678324e-01 -3.92200768e-01
-6.68522716e-01 -1.39741540e-01 3.62281501e-01 5.99010527e-01
1.82648540e-01 -7.21932799e-02 2.18656346e-01 1.25153810e-01
3.17605436e-02 6.84744772e-03 3.03905219e-01 1.14751899e+00
-3.11646163e-01 -1.12833059e+00 -3.68016809e-01 5.90384901e-01
-2.53104091e-01 -3.70847136e-02 -2.87538171e-01 7.92207003e-01
1.34344818e-02 5.91176093e-01 4.90135074e-01 -3.17337126e-01
2.90373713e-01 -8.45167488e-02 7.69927382e-01 -6.94120646e-01
-7.88163394e-02 8.25791061e-02 -3.30412537e-01 -7.91716754e-01
-6.17171705e-01 -6.78774893e-01 -1.29499280e+00 -3.08429331e-01
-4.25997406e-01 5.68010025e-02 3.63567501e-01 6.81654513e-01
3.11708003e-01 -5.57466298e-02 5.26777446e-01 -1.06424904e+00
-4.19821262e-01 -5.64956367e-01 -9.51450288e-01 6.37484550e-01
7.46280313e-01 -8.87443066e-01 -3.42835367e-01 2.60088593e-01] | [7.9529643058776855, -2.4755403995513916] |
ad0ecee0-76a1-4321-b577-d296a90e549f | practical-fast-and-robust-point-cloud | 2111.04228 | null | https://arxiv.org/abs/2111.04228v1 | https://arxiv.org/pdf/2111.04228v1.pdf | Practical, Fast and Robust Point Cloud Registration for 3D Scene Stitching and Object Localization | 3D point cloud registration ranks among the most fundamental problems in remote sensing, photogrammetry, robotics and geometric computer vision. Due to the limited accuracy of 3D feature matching techniques, outliers may exist, sometimes even in very large numbers, among the correspondences. Since existing robust solvers may encounter high computational cost or restricted robustness, we propose a novel, fast and highly robust solution, named VOCRA (VOting with Cost function and Rotating Averaging), for the point cloud registration problem with extreme outlier rates. Our first contribution is to employ the Tukey's Biweight robust cost to introduce a new voting and correspondence sorting technique, which proves to be rather effective in distinguishing true inliers from outliers even with extreme (99%) outlier rates. Our second contribution consists in designing a time-efficient consensus maximization paradigm based on robust rotation averaging, serving to seek inlier candidates among the correspondences. Finally, we apply Graduated Non-Convexity with Tukey's Biweight (GNC-TB) to estimate the correct transformation with the inlier candidates obtained, which is then used to find the complete inlier set. Both standard benchmarking and realistic experiments with application to two real-data problems are conducted, and we show that our solver VOCRA is robust against over 99% outliers and more time-efficient than the state-of-the-art competitors. | ['Lei Sun'] | 2021-11-08 | null | null | null | null | ['3d-feature-matching'] | ['computer-vision'] | [-5.40453494e-02 -4.36365962e-01 2.93214172e-01 -2.93030869e-02
-1.09834599e+00 -4.90086228e-01 5.74168086e-01 1.55030444e-01
-2.45889217e-01 4.94610429e-01 -2.55857110e-01 -1.17701344e-01
-3.69295806e-01 -5.49238384e-01 -8.53113592e-01 -7.32454658e-01
-1.91032782e-01 7.96851635e-01 -5.41831693e-03 -2.03446358e-01
6.37613297e-01 1.09113801e+00 -1.59758842e+00 -5.08988082e-01
1.25476706e+00 8.48159730e-01 -3.51363234e-02 3.32850337e-01
1.47416592e-01 -2.68452168e-02 -2.98593998e-01 -1.24974377e-01
8.37940991e-01 -1.23971239e-01 -6.43754303e-01 3.90892744e-01
8.37418914e-01 -1.04772620e-01 2.48742253e-01 1.20311940e+00
6.56579614e-01 3.81659269e-01 5.75922549e-01 -1.23854995e+00
-3.13148797e-02 3.23021039e-02 -9.89302874e-01 -2.59611964e-01
5.86262822e-01 -8.53484571e-02 8.42277586e-01 -1.63924003e+00
6.70742869e-01 9.59189236e-01 1.06565809e+00 -2.68055149e-03
-1.23491323e+00 -4.83440787e-01 -6.03656992e-02 -4.37288992e-02
-1.79903185e+00 -4.43257034e-01 7.43640065e-01 -4.52318937e-01
7.31603086e-01 7.99520910e-01 6.26006365e-01 2.67829627e-01
-4.74373959e-02 2.88259119e-01 8.73826146e-01 -3.18500251e-01
2.37924606e-01 -4.39613283e-01 -2.49971617e-02 4.08705831e-01
5.13162553e-01 1.21839307e-01 -3.34563702e-01 -7.27072716e-01
7.52225757e-01 2.40004092e-01 -4.04937178e-01 -6.85981512e-01
-1.55383265e+00 7.47088552e-01 5.77064514e-01 1.51717126e-01
-6.16583288e-01 1.22616462e-01 5.96706755e-02 3.57341170e-01
6.84390247e-01 5.42457104e-01 -4.06817228e-01 1.52882561e-01
-1.13717079e+00 4.97295648e-01 7.08808362e-01 1.08202243e+00
1.01748919e+00 -7.42359906e-02 2.68483222e-01 7.60737598e-01
3.23466063e-01 8.25921237e-01 1.15664914e-01 -8.50241721e-01
4.57817137e-01 6.73016369e-01 2.86689728e-01 -1.38691175e+00
-5.84248900e-01 -4.33021635e-01 -9.77250576e-01 3.04556668e-01
2.78407425e-01 1.37164429e-01 -7.47481585e-01 1.23600876e+00
6.45084262e-01 5.98379910e-01 -1.29952624e-01 1.04433811e+00
4.85343009e-01 4.38536942e-01 -5.91085494e-01 -5.26908875e-01
9.61693823e-01 -3.23551029e-01 -1.73264116e-01 -4.79854904e-02
6.40077829e-01 -1.24691463e+00 5.59504807e-01 3.24352831e-01
-1.06661618e+00 -2.26717532e-01 -9.80416954e-01 1.68339163e-01
5.39435931e-02 -4.76728752e-02 4.21269894e-01 2.71875560e-01
-9.49084520e-01 8.24988306e-01 -7.90430725e-01 -4.61329162e-01
1.43567622e-01 6.43033504e-01 -5.61132431e-01 -1.03065759e-01
-4.55864877e-01 1.07995307e+00 4.96895574e-02 4.19900924e-01
-2.43447244e-01 -8.39613140e-01 -8.06842208e-01 -3.44374359e-01
4.21500057e-01 -8.49904180e-01 7.56763399e-01 -7.31204033e-01
-1.20650959e+00 9.30863142e-01 -4.20652807e-01 -1.63960308e-01
8.78872573e-01 -2.00342700e-01 1.95714068e-02 -2.06746131e-01
2.36184344e-01 2.16110006e-01 8.58475924e-01 -1.45404935e+00
-3.97451222e-01 -6.49995506e-01 -6.36372924e-01 1.83552563e-01
2.66080290e-01 7.24117085e-02 -3.64825457e-01 -5.74470401e-01
1.29284573e+00 -1.17657495e+00 -6.77269161e-01 -2.20426813e-01
-3.99205625e-01 -1.01939075e-01 6.19756937e-01 -4.94052619e-01
7.77922332e-01 -2.16873574e+00 2.31870249e-01 1.10791278e+00
6.59702271e-02 -5.04972450e-02 -1.39658060e-02 4.42718267e-01
-2.57741868e-01 -1.70488745e-01 -5.35205126e-01 -3.94676685e-01
-2.34621421e-01 1.53731376e-01 -4.13823664e-01 1.25100434e+00
1.26133608e-02 4.94723171e-01 -9.99439657e-01 -1.71453103e-01
5.09138644e-01 2.28791118e-01 -5.00541329e-01 -2.24975869e-01
2.55085498e-01 5.05388916e-01 -2.94037133e-01 8.67037892e-01
1.16865706e+00 1.21726103e-01 -2.15323731e-01 -1.82173923e-01
-4.34275448e-01 -2.34271631e-01 -2.00849009e+00 1.56569684e+00
-1.24548569e-01 2.57885456e-01 2.06971854e-01 -7.22495854e-01
1.21242237e+00 6.24894276e-02 9.68196094e-01 -2.18927413e-01
1.46173298e-01 9.18093741e-01 -2.90181220e-01 -4.68313433e-02
8.03207636e-01 1.79763362e-02 5.65003827e-02 1.19053774e-01
-5.03241479e-01 -6.30672872e-01 3.54391113e-02 -4.01336364e-02
9.68705058e-01 6.84270309e-03 4.23979104e-01 -5.46755791e-01
8.23909700e-01 1.88801304e-01 7.54239023e-01 5.77118814e-01
4.19693589e-02 1.18663156e+00 -2.56069247e-02 -4.93001342e-01
-1.05021846e+00 -7.24228024e-01 -2.28388622e-01 4.09223109e-01
4.25107479e-01 -1.45137236e-01 -2.17320710e-01 -1.94738686e-01
3.53183061e-01 3.22615385e-01 -1.48476034e-01 1.12668194e-01
-7.57675588e-01 -7.61040151e-01 1.45475762e-02 1.87428191e-01
1.14313826e-01 -5.73525846e-01 -4.02267098e-01 3.06969911e-01
-4.21592668e-02 -9.25328672e-01 -3.81673098e-01 -3.94965038e-02
-1.29759324e+00 -1.25596404e+00 -6.54099464e-01 -6.13729239e-01
1.11681473e+00 6.28974020e-01 1.06140935e+00 3.41166705e-01
-2.10654914e-01 2.09176317e-01 -3.31591666e-01 -2.58701950e-01
-6.92690760e-02 -1.54241607e-01 3.56987715e-01 5.53543605e-02
1.52476117e-01 -5.66704988e-01 -4.94019240e-01 5.73030829e-01
-6.67287648e-01 -4.48055774e-01 3.60572070e-01 7.41552711e-01
9.32594597e-01 -2.50102639e-01 7.83265568e-03 -6.12968206e-01
2.99364626e-01 -3.06323886e-01 -1.10822248e+00 -5.73159866e-02
-4.07111764e-01 -1.85665265e-01 4.33802158e-01 -1.58223420e-01
-2.75371313e-01 6.36443973e-01 -6.34363145e-02 -7.28351831e-01
1.57873362e-01 5.02613962e-01 -5.67059673e-04 -7.35517204e-01
7.10458279e-01 3.72684859e-02 -7.88827613e-03 -3.81722867e-01
3.41761947e-01 4.05036449e-01 7.67816484e-01 -5.83648682e-01
1.40676796e+00 8.76686573e-01 4.79687244e-01 -1.00703299e+00
-3.69968951e-01 -1.11126423e+00 -8.14389169e-01 -6.58926591e-02
2.76263863e-01 -9.61217582e-01 -6.06876612e-01 5.57767272e-01
-1.20699525e+00 2.81153977e-01 -3.02123666e-01 5.90866625e-01
-6.45323575e-01 7.09259212e-01 3.60528864e-02 -7.31664360e-01
-4.40507770e-01 -1.24060011e+00 1.21771169e+00 -8.82512555e-02
-1.28395557e-01 -5.69483817e-01 2.65366375e-01 4.93069626e-02
1.68863103e-01 7.66808987e-01 2.32057989e-01 -6.39751017e-01
-6.75398946e-01 -5.46334505e-01 -7.21931607e-02 8.88961107e-02
-5.37254252e-02 4.50841367e-01 -5.28983712e-01 -6.07716024e-01
4.02026810e-02 3.16149265e-01 6.13112569e-01 4.17413205e-01
8.11344504e-01 -1.96527559e-02 -4.41885084e-01 8.92246425e-01
1.65740120e+00 -3.48756254e-01 6.64415896e-01 4.89208728e-01
7.69947350e-01 3.29449087e-01 8.88066351e-01 6.76263034e-01
1.30121350e-01 7.56490767e-01 7.16809332e-01 -1.40500337e-01
3.66883516e-01 2.56825686e-01 1.63274899e-01 1.06459546e+00
-5.65125048e-01 3.60336095e-01 -1.00957692e+00 5.88629425e-01
-2.08060765e+00 -7.82250643e-01 -8.05875182e-01 2.80382609e+00
2.65248865e-01 -3.15156698e-01 -4.75795940e-02 3.27169359e-01
8.89878511e-01 -1.09562978e-01 -2.67124802e-01 -1.05712317e-01
-2.44433522e-01 2.46785805e-01 1.07616341e+00 5.93471229e-01
-1.16718161e+00 7.08418667e-01 5.55889416e+00 4.94133860e-01
-9.69743669e-01 -9.47839767e-02 -5.88699281e-02 1.34976685e-01
-2.08121359e-01 3.41003239e-01 -5.39055645e-01 2.08487973e-01
3.56384516e-01 -8.15218687e-02 2.80476034e-01 9.15408552e-01
2.82307982e-01 -1.51600718e-01 -9.56831038e-01 1.32744169e+00
1.88635319e-01 -1.20139623e+00 -1.69327170e-01 2.67173469e-01
1.11810243e+00 3.58787328e-01 -3.12678427e-01 -3.33025366e-01
2.71543324e-01 -7.06361890e-01 6.88251793e-01 4.57792729e-01
5.48437476e-01 -9.79600966e-01 8.36797178e-01 2.65499264e-01
-1.24811733e+00 1.17836528e-01 -8.29521596e-01 3.81638743e-02
1.14313073e-01 1.01701081e+00 -5.73421240e-01 1.12507570e+00
5.75525165e-01 8.39235365e-01 -3.00162852e-01 1.66195631e+00
3.49244736e-02 -6.91188946e-02 -8.03261518e-01 3.02516431e-01
1.26057893e-01 -7.56715000e-01 1.07042515e+00 7.43179083e-01
8.77964616e-01 3.43725353e-01 4.04190063e-01 4.67874318e-01
4.09704261e-02 3.74852031e-01 -6.52099609e-01 8.66865218e-01
6.73258305e-01 1.30185866e+00 -7.86168754e-01 -1.07711598e-01
-1.16739415e-01 7.80182183e-01 -6.01450354e-02 1.00457571e-01
-5.78370869e-01 -2.36242086e-01 7.63081789e-01 3.84741165e-02
1.27139136e-01 -4.99692231e-01 -5.40126801e-01 -1.17639828e+00
3.20946455e-01 -8.97755802e-01 2.42643937e-01 -4.71807092e-01
-1.23997474e+00 4.37285632e-01 -1.58522055e-01 -1.82036877e+00
-8.85355398e-02 -3.87062341e-01 -7.89923608e-01 7.58320153e-01
-1.44599855e+00 -9.14882898e-01 -6.55720532e-01 7.40700424e-01
2.72151709e-01 1.03167139e-01 5.69875598e-01 3.76597047e-01
-3.33465695e-01 2.91929871e-01 4.65241671e-01 -1.68453425e-01
7.34580874e-01 -1.09302914e+00 4.31421012e-01 1.08531547e+00
7.95652643e-02 6.35015726e-01 9.81418967e-01 -6.23010814e-01
-1.83458412e+00 -1.04817998e+00 9.82811987e-01 -4.70813572e-01
5.72510898e-01 -2.81263310e-02 -9.27603483e-01 6.19922280e-01
-4.56107318e-01 2.98414111e-01 1.33715153e-01 -5.11978716e-02
-1.11926608e-01 -2.33472139e-01 -1.22254789e+00 5.36557496e-01
9.46782768e-01 -9.45311040e-02 -7.07935214e-01 5.63241959e-01
2.05094427e-01 -7.87529171e-01 -7.80950904e-01 8.19280565e-01
2.80030072e-01 -9.49689031e-01 1.22597432e+00 -7.80632347e-02
-7.30677545e-02 -8.02018642e-01 -2.52823442e-01 -1.10282934e+00
-1.50261089e-01 -9.22726810e-01 3.99770200e-01 1.10793042e+00
1.61184862e-01 -8.64780188e-01 7.21404731e-01 4.19714659e-01
-3.42965961e-01 -4.15973395e-01 -1.33443892e+00 -1.00264323e+00
-1.75062433e-01 -5.43749571e-01 7.09177136e-01 1.21840382e+00
-3.86165261e-01 -2.68820286e-01 -3.28191102e-01 5.47389746e-01
9.14784670e-01 3.95357490e-01 1.36055148e+00 -1.73266196e+00
2.50949830e-01 -4.02455598e-01 -7.44992197e-01 -7.94345975e-01
-1.69700589e-02 -8.02563667e-01 3.34282458e-01 -1.30870700e+00
-2.08766803e-01 -7.70968497e-01 7.76706338e-02 1.92343637e-01
-1.19460359e-01 3.95308554e-01 1.89178050e-01 5.74948907e-01
-2.77051270e-01 4.43838924e-01 8.66460264e-01 5.46889715e-02
-5.12141883e-01 3.61784697e-01 -1.92541674e-01 9.87428010e-01
3.56190920e-01 -5.63017547e-01 3.66905421e-01 -3.31856787e-01
4.04516459e-01 -8.80583450e-02 3.91671509e-01 -1.10998917e+00
4.28771019e-01 -6.64838552e-02 2.15849176e-01 -1.14497447e+00
2.02161461e-01 -1.31421816e+00 5.71959496e-01 4.87040997e-01
4.49029386e-01 3.35624129e-01 -1.48637578e-01 3.80140930e-01
-2.73999602e-01 -2.38146007e-01 7.55973816e-01 -9.88758951e-02
-4.80199277e-01 4.63514984e-01 1.42145932e-01 -3.08679581e-01
1.09575510e+00 -5.59528291e-01 -4.91872430e-02 -9.55673829e-02
-4.65637237e-01 3.55822057e-01 9.30713832e-01 1.07562393e-01
6.80420458e-01 -1.37638843e+00 -1.13410592e+00 2.98058212e-01
2.10535944e-01 4.13868159e-01 -4.97070327e-02 1.14969766e+00
-1.00582957e+00 -2.91269124e-02 1.22283652e-01 -1.18572891e+00
-1.34045231e+00 2.49528840e-01 1.85590371e-01 8.20677653e-02
-6.05680108e-01 7.85137296e-01 -4.96916831e-01 -7.11979151e-01
2.43590232e-02 -4.54023153e-01 3.35488468e-01 8.05467665e-02
1.43537283e-01 8.98456335e-01 5.14557123e-01 -1.00673461e+00
-7.66245067e-01 1.32512307e+00 4.23712492e-01 1.42034188e-01
1.38007617e+00 -1.30506670e-02 -5.40851653e-01 4.29154374e-02
1.00317144e+00 3.85414898e-01 -9.23076868e-01 -1.67720214e-01
3.22505176e-01 -8.76802802e-01 -2.32203573e-01 7.82226622e-02
-1.05254102e+00 3.45692813e-01 3.92331898e-01 5.41011766e-02
1.01646423e+00 -2.18573019e-01 6.74255967e-01 4.86638695e-01
6.02462530e-01 -9.87737000e-01 -5.20723641e-01 6.88971043e-01
1.05610216e+00 -1.31187856e+00 6.44958496e-01 -6.43685699e-01
-1.89698711e-01 1.11673200e+00 1.80754721e-01 -7.71026611e-01
5.32484770e-01 -5.78826144e-02 1.52793095e-01 -2.55017221e-01
-6.44265264e-02 -3.12766671e-01 3.78665090e-01 2.39432871e-01
2.85110082e-02 -9.61733609e-02 -5.53034484e-01 -2.72296369e-01
-3.76059264e-01 -3.49657714e-01 4.23727393e-01 8.83284450e-01
-4.68405396e-01 -9.36500072e-01 -1.08212590e+00 3.88244957e-01
-3.02176416e-01 7.57114664e-02 -2.80781448e-01 8.03632081e-01
-3.92568894e-02 8.09813917e-01 8.73907432e-02 -1.50245130e-01
6.78644001e-01 -2.35095933e-01 2.51120746e-01 -3.33834678e-01
-8.27553928e-01 4.10293400e-01 -2.90036798e-01 -8.11930835e-01
-6.28521681e-01 -9.75503266e-01 -1.24216652e+00 -4.96856928e-01
-7.64677703e-01 2.19252169e-01 8.81201863e-01 9.05652404e-01
3.55817556e-01 -2.36827329e-01 1.02947748e+00 -1.26911914e+00
-7.56686509e-01 -6.91655874e-01 -5.28526366e-01 5.41175783e-01
3.64674628e-01 -5.99297762e-01 -6.89563572e-01 -3.82074565e-01] | [7.732126235961914, -2.8319382667541504] |
9a521c3f-2d5e-4fa9-9852-f09417feaf45 | neural-pipeline-for-zero-shot-data-to-text | null | null | https://openreview.net/forum?id=Lz2fD-uQyeh | https://openreview.net/pdf?id=Lz2fD-uQyeh | Neural Pipeline for Zero-Shot Data-to-Text Generation | In data-to-text (D2T) generation, training on in-domain data leads to overfitting to the data representation and repeating training data noise. We examine how to avoid finetuning the pretrained language models (PLMs) on D2T generation datasets while still taking advantage of surface realization capabilities of PLMs. Inspired by pipeline approaches, we propose to generate text by rephrasing single-item templates using a sequence of modules trained on general-domain text-based operations—ordering, aggregation, and paragraph compression. We train PLMs for performing these operations on a synthetic corpus WikiFluent which we build from English Wikipedia. Our experiments on two major triple-to-text datasets—WebNLG and E2E—show that our approach enables D2T generation from RDF triples in zero-shot settings. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['data-to-text-generation'] | ['natural-language-processing'] | [ 3.41475070e-01 1.03904772e+00 -6.27874658e-02 -2.88843870e-01
-1.03124559e+00 -3.80893379e-01 1.07586300e+00 3.01264614e-01
-7.38066658e-02 9.44446921e-01 6.55564666e-01 -3.15227777e-01
1.57310903e-01 -1.37414432e+00 -1.27967918e+00 9.99135673e-02
1.94813550e-01 1.01662457e+00 1.19729765e-01 -5.48935175e-01
2.49804676e-01 -1.00338988e-01 -1.87080359e+00 9.60581303e-01
1.31132245e+00 6.54570222e-01 9.69816744e-02 5.31712830e-01
-9.56945777e-01 8.15591812e-01 -5.72294474e-01 -7.85538256e-01
4.51367468e-01 -3.71562421e-01 -8.79356861e-01 1.21234737e-01
6.56816781e-01 -1.45439431e-01 -2.15112120e-01 5.62711298e-01
4.31987703e-01 2.04919819e-02 9.12485063e-01 -9.10337329e-01
-8.51189435e-01 1.29181671e+00 -2.18469858e-01 -5.04275143e-01
4.82632548e-01 1.55909061e-01 1.13066900e+00 -1.33832765e+00
1.16956365e+00 1.26944160e+00 7.21255600e-01 7.92471170e-01
-1.56270564e+00 -1.72179908e-01 -4.33421791e-01 -3.35925400e-01
-1.18890798e+00 -8.81070614e-01 7.44204074e-02 -2.99206257e-01
1.65984929e+00 1.64350957e-01 5.25082171e-01 1.21647322e+00
-1.64636627e-01 8.88854086e-01 5.51595211e-01 -6.92533851e-01
3.11428368e-01 1.03009254e-01 -2.56182015e-01 5.89189827e-01
7.43646383e-01 -2.16048598e-01 -8.97817850e-01 -2.66076118e-01
5.16108930e-01 -7.27470219e-01 2.51859725e-01 -3.68981391e-01
-1.23222232e+00 7.17310071e-01 -7.86122307e-02 -3.82884964e-02
-2.64149994e-01 2.62065619e-01 5.53424716e-01 3.92190725e-01
4.72906470e-01 6.27532601e-01 -5.42615473e-01 -1.79638684e-01
-9.63119268e-01 7.08477020e-01 1.14805543e+00 1.79287493e+00
9.05244589e-01 1.14077471e-01 -5.96508920e-01 8.46891642e-01
7.78303221e-02 7.20766723e-01 5.36531448e-01 -7.22365856e-01
1.16393483e+00 7.59456992e-01 3.06743562e-01 -2.06429616e-01
-2.55998224e-01 1.29786432e-01 -5.02311826e-01 -3.43319684e-01
3.96546036e-01 -4.10212010e-01 -1.25190187e+00 1.34721184e+00
1.38501972e-01 -4.63965267e-01 4.90782708e-01 2.81598598e-01
9.44761157e-01 6.00752950e-01 2.73139477e-01 1.47027552e-01
1.00478733e+00 -5.85621595e-01 -4.44815695e-01 -4.12856340e-01
1.34098577e+00 -5.59228897e-01 1.25082684e+00 2.21026734e-01
-1.51885486e+00 -4.36621994e-01 -8.88369799e-01 -6.23191953e-01
-6.11364901e-01 1.40481040e-01 5.72058678e-01 7.13943243e-01
-8.24595690e-01 6.95898712e-01 -5.35942733e-01 -6.68315411e-01
4.23966885e-01 4.45532948e-02 -3.49374652e-01 -3.89390886e-02
-1.22673404e+00 7.71606684e-01 8.74118268e-01 -3.38547796e-01
-7.34532297e-01 -1.00337589e+00 -8.82597983e-01 -5.85635602e-02
5.04020393e-01 -1.09636438e+00 1.28567982e+00 -3.76983464e-01
-1.47819614e+00 8.03807497e-01 -2.23417968e-01 -6.89903855e-01
3.92610818e-01 -1.44511104e-01 -2.82266766e-01 -2.09716290e-01
1.82044491e-01 8.83945823e-01 7.22311616e-01 -1.00734472e+00
-5.75863540e-01 -1.14128500e-01 -1.79619789e-01 1.69076279e-01
-3.63873541e-01 -3.19001615e-01 -2.07827583e-01 -6.17480993e-01
-1.14664845e-01 -6.36075914e-01 -2.45230179e-02 -3.79384905e-01
-8.68413389e-01 -3.42303008e-01 4.08081949e-01 -5.51836967e-01
1.25015807e+00 -1.48841441e+00 1.45338267e-01 2.57166952e-01
-1.26464665e-01 1.15374222e-01 -3.99458051e-01 1.02747762e+00
-8.28011148e-03 5.83884060e-01 -2.12302759e-01 -3.21022242e-01
5.42952001e-01 3.20236117e-01 -4.76170957e-01 -4.64945793e-01
4.08029199e-01 1.24828053e+00 -8.22786629e-01 -6.40492022e-01
-1.49179652e-01 -1.83002368e-01 -1.02422249e+00 2.71914721e-01
-1.15601563e+00 -1.26658037e-01 -3.69608432e-01 3.94203156e-01
4.79144722e-01 -1.66532978e-01 5.64257741e-01 -3.01218033e-01
-1.76745370e-01 7.96997070e-01 -1.00740993e+00 2.25153208e+00
-5.62711775e-01 1.70685723e-01 -5.87000549e-01 -7.60368645e-01
1.02470493e+00 1.63275734e-01 1.76879704e-01 -8.94899487e-01
-3.47592473e-01 4.38938528e-01 -4.42564100e-01 -9.40691233e-01
1.05320513e+00 -1.46073699e-01 -4.72084433e-01 7.07135677e-01
5.14310956e-01 -5.23702681e-01 8.76232684e-01 5.62238395e-01
1.14746594e+00 5.61051667e-01 4.35226485e-02 -3.16538841e-01
-7.35641271e-02 2.64157712e-01 1.99389949e-01 9.27589834e-01
1.07374310e+00 5.82106590e-01 6.12818241e-01 -2.28480980e-01
-1.73811710e+00 -1.05547440e+00 6.88169226e-02 9.72494364e-01
-5.93807459e-01 -1.04815471e+00 -7.46090233e-01 -8.56530249e-01
2.32751921e-01 1.59741175e+00 -4.53942269e-01 -1.46148175e-01
-8.23965669e-01 -8.75490069e-01 9.58326757e-01 3.67920667e-01
3.17344189e-01 -7.88122892e-01 -1.67966366e-01 4.09243494e-01
-3.34358245e-01 -9.57573771e-01 -4.49186772e-01 2.50758260e-01
-7.58470654e-01 -7.17178702e-01 -3.25425178e-01 -5.62766373e-01
7.19122827e-01 -1.28227383e-01 1.70028734e+00 -9.36329514e-02
-2.79027760e-01 1.96284518e-01 -3.23952585e-01 -5.16017377e-01
-1.02699077e+00 6.01596653e-01 -1.69866219e-01 -4.65957731e-01
4.81687933e-01 -4.90201086e-01 -1.50209263e-01 -1.89492524e-01
-1.21880281e+00 5.17587423e-01 4.90070015e-01 7.33979583e-01
4.36527640e-01 -3.07888269e-01 7.16870129e-01 -1.25437248e+00
6.59310639e-01 -6.33690476e-01 -5.10740161e-01 6.57260180e-01
-5.73438764e-01 7.26931095e-01 5.76812208e-01 4.53688353e-02
-1.46110117e+00 -2.65046686e-01 -1.82478949e-02 3.47895660e-02
7.09451959e-02 6.36291921e-01 -3.51576269e-01 5.85690796e-01
1.08741164e+00 2.52219558e-01 -3.35310489e-01 -5.32953918e-01
1.02137017e+00 5.30906022e-01 1.97067156e-01 -1.32597113e+00
8.67857635e-01 1.62881657e-01 -1.76755205e-01 -8.17258954e-01
-7.67985761e-01 2.85919458e-01 -5.43003678e-01 2.31378168e-01
7.01013327e-01 -8.61453056e-01 -2.89181113e-01 2.58278906e-01
-1.07467973e+00 -6.58695579e-01 -9.02778566e-01 -9.90561992e-02
-7.85568833e-01 2.53900766e-01 -5.16261220e-01 -5.24548888e-01
-6.20364964e-01 -5.88266611e-01 1.31051600e+00 -4.86576743e-02
-2.16236219e-01 -9.31857288e-01 1.03270970e-01 2.04806656e-01
3.67587030e-01 -9.66638476e-02 1.38553333e+00 -8.94485235e-01
-7.70731330e-01 1.11308493e-01 -2.08146855e-01 1.46204636e-01
-4.45424356e-02 2.34649666e-02 -6.69168413e-01 1.53231442e-01
-5.99211752e-01 -7.28615224e-01 7.51152635e-01 -2.65430152e-01
1.00441718e+00 -6.62980795e-01 -2.75876313e-01 5.19539475e-01
1.37931526e+00 -3.05371165e-01 1.01833749e+00 3.05172622e-01
6.57821298e-01 6.51549220e-01 4.45849597e-01 6.16191387e-01
7.05207825e-01 5.86059034e-01 1.05063334e-01 3.35151464e-01
-4.44131464e-01 -1.05547440e+00 4.00943190e-01 6.60268247e-01
7.69086331e-02 -5.60724676e-01 -9.38452303e-01 7.82715201e-01
-1.82364798e+00 -1.08527672e+00 -2.57972747e-01 2.21505189e+00
1.33170569e+00 1.73324600e-01 -4.60354500e-02 -3.08189988e-01
3.05913299e-01 -1.85260311e-01 -3.90202701e-01 -4.41236049e-01
-4.83619064e-01 5.53820610e-01 6.62269056e-01 8.05207565e-02
-5.88136911e-01 1.06238317e+00 6.42845964e+00 1.04363823e+00
-5.87731421e-01 -8.47423375e-02 3.51659685e-01 -3.09272557e-01
-1.09281123e+00 2.20319912e-01 -1.18332124e+00 3.48151535e-01
1.27375078e+00 -6.28504395e-01 5.51640868e-01 5.25822699e-01
1.88785091e-01 7.93371722e-02 -1.38899422e+00 4.07044619e-01
-1.77967479e-03 -1.93890738e+00 7.33645618e-01 -4.80396636e-02
9.57817852e-01 6.06705025e-02 -3.01040083e-01 6.96347713e-01
9.20009315e-01 -1.03414595e+00 9.72501516e-01 7.85185814e-01
9.77440357e-01 -4.33569074e-01 1.65276378e-01 3.01367968e-01
-8.40880096e-01 4.97422069e-02 -5.34604907e-01 3.43675673e-01
2.22188398e-01 8.33421767e-01 -1.35990250e+00 8.75396907e-01
3.41304451e-01 4.95854616e-01 -4.33567256e-01 4.89085495e-01
-1.81751147e-01 3.66824478e-01 -4.45361495e-01 4.92504202e-02
-1.68488428e-01 -9.14374590e-02 3.51678222e-01 1.36931038e+00
9.13035154e-01 -1.96808800e-01 -9.14429352e-02 1.16208291e+00
-5.48056483e-01 1.52302504e-01 -8.59438181e-01 -4.19435948e-01
6.28896058e-01 1.12816477e+00 -1.02965996e-01 -6.66575909e-01
-5.71098328e-01 7.00167179e-01 5.76474309e-01 2.78979510e-01
-5.59425414e-01 -5.18754244e-01 3.86223286e-01 5.35351038e-01
5.46355426e-01 -2.60580212e-01 -4.02889311e-01 -1.54858649e+00
1.56792194e-01 -1.05639255e+00 2.76991695e-01 -9.51543272e-01
-1.37073481e+00 2.48622298e-01 1.71201527e-01 -9.25121367e-01
-6.57504320e-01 -3.21418047e-01 -4.14351791e-01 8.31260622e-01
-1.21897006e+00 -1.38464630e+00 -6.01631701e-02 3.46410334e-01
4.34714109e-01 -8.55192021e-02 8.67587864e-01 1.94145739e-01
-3.20878327e-01 6.19297862e-01 3.09612229e-02 6.66136518e-02
6.99531615e-01 -1.43046510e+00 1.23159349e+00 6.48739040e-01
-8.01162422e-02 8.51160944e-01 6.75489724e-01 -1.11122465e+00
-1.77365124e+00 -1.42016685e+00 1.26522291e+00 -5.50196052e-01
9.39049065e-01 -7.71710694e-01 -7.20873475e-01 9.96446133e-01
3.57325733e-01 -4.09233570e-01 5.15086293e-01 9.54283215e-03
-5.53739309e-01 1.40269905e-01 -1.05660391e+00 8.02768826e-01
1.58650041e+00 -4.27315593e-01 -7.55631626e-01 5.34770250e-01
8.28906655e-01 -5.42978227e-01 -1.26328433e+00 1.46193236e-01
3.42810839e-01 -6.94414496e-01 8.32099140e-01 -9.45848703e-01
8.03134680e-01 -1.60020813e-01 -3.12336355e-01 -1.52868295e+00
2.37601340e-01 -9.01551902e-01 -2.44125962e-01 1.40673590e+00
9.36240315e-01 -4.03420955e-01 8.90759766e-01 9.11501527e-01
-4.55977768e-01 -9.89258215e-02 -7.64014721e-01 -7.51243532e-01
4.18776006e-01 -3.71903926e-01 1.13430488e+00 5.64596057e-01
4.40225303e-01 5.56382775e-01 -3.52863014e-01 -5.16617417e-01
6.95212841e-01 -1.12019867e-01 1.28829229e+00 -1.04852092e+00
-2.81532496e-01 4.93225418e-02 4.33835834e-01 -1.23010480e+00
-1.20188415e-01 -1.43829417e+00 2.04686951e-02 -1.72618449e+00
9.81677994e-02 -7.11588681e-01 6.72989309e-01 5.67872524e-01
4.50696349e-02 -2.18477845e-01 6.55705035e-02 -2.37102807e-02
-3.93054128e-01 7.61252224e-01 1.19436562e+00 -1.29937172e-01
5.35389520e-02 -5.29320538e-01 -7.27683306e-01 3.00837934e-01
5.69376707e-01 -4.38551068e-01 -4.52078849e-01 -9.26870883e-01
8.81728768e-01 2.68211395e-01 1.03255883e-01 -8.42430592e-01
1.64525405e-01 -3.25142294e-01 2.74006814e-01 -5.14394939e-01
1.07668422e-01 -3.64233702e-01 3.02122384e-01 -9.89701599e-02
-6.34054542e-01 -1.93902597e-01 2.19311073e-01 2.14423865e-01
2.39352599e-01 -6.13675654e-01 2.10641935e-01 -4.80491430e-01
-4.59376812e-01 1.72344178e-01 -3.00481915e-01 5.24504304e-01
4.15693045e-01 -7.10529536e-02 -9.77951646e-01 -8.31277203e-03
-6.13890469e-01 1.55752867e-01 6.61574900e-01 3.69559199e-01
3.47840846e-01 -1.41138852e+00 -9.60867703e-01 3.81271809e-01
5.17803609e-01 8.26201066e-02 2.06714123e-02 3.31101805e-01
-2.84565270e-01 6.74031734e-01 1.63531825e-02 -1.24965847e-01
-5.22322178e-01 4.04633462e-01 1.85637772e-01 -4.95829225e-01
-5.70674419e-01 3.42382312e-01 -1.44896060e-01 -8.49946022e-01
-3.30463797e-01 -5.89109242e-01 2.66895711e-01 1.58433802e-02
2.43842363e-01 2.95638680e-01 4.51935977e-01 1.12990633e-01
2.69875377e-01 4.58472595e-03 -7.80202821e-02 -2.73376107e-01
1.37037086e+00 3.00344843e-02 -1.46393999e-01 1.79118708e-01
8.65821540e-01 1.89478084e-01 -8.28841746e-01 -1.99498534e-01
5.66619456e-01 -5.86035661e-02 -4.17278618e-01 -8.52347136e-01
-5.01339555e-01 5.38808703e-01 -2.52134204e-01 3.80366534e-01
5.60743749e-01 -3.37545238e-02 1.01941407e+00 9.77878511e-01
5.75131893e-01 -1.51106346e+00 1.39801297e-02 7.26400554e-01
8.92800510e-01 -8.61551940e-01 -1.25881702e-01 -2.79179037e-01
-6.23412669e-01 1.25759482e+00 7.26145804e-01 6.35950193e-02
2.41013646e-01 4.03668195e-01 -4.14512247e-01 -2.66242623e-01
-1.34572923e+00 -2.73841828e-01 1.00419052e-01 7.85027385e-01
5.78348458e-01 -1.48262918e-01 -2.03338355e-01 4.83628392e-01
-4.24930096e-01 4.83769596e-01 8.08564126e-01 9.71145034e-01
-6.25448287e-01 -1.66393781e+00 -4.69277352e-02 9.73836958e-01
-1.26523674e-01 -5.33721447e-01 -1.93958864e-01 8.93121004e-01
-3.51536907e-02 7.79879987e-01 1.15826771e-01 -2.14421138e-01
5.23210824e-01 5.09311080e-01 6.37641847e-01 -1.04502845e+00
-5.62400877e-01 -2.87748963e-01 1.10847664e+00 -3.04836154e-01
-2.05912113e-01 -8.02065670e-01 -1.38952887e+00 -3.63965452e-01
7.07394630e-02 2.14725867e-01 6.74400210e-01 7.99875021e-01
7.76337564e-01 3.45336556e-01 7.13659227e-02 -5.62035978e-01
-6.10765517e-01 -1.21760643e+00 -3.70829523e-01 6.63187742e-01
-3.87396932e-01 -2.20269918e-01 9.02293921e-02 2.33445644e-01] | [11.394323348999023, 8.7838773727417] |
a884bfb3-0d59-4651-b517-47a19e9cdcc7 | fitvid-high-capacity-pixel-level-video | null | null | https://openreview.net/forum?id=iim-R8xu0TG | https://openreview.net/pdf?id=iim-R8xu0TG | FitVid: High-Capacity Pixel-Level Video Prediction | An agent that is capable of predicting what happens next can perform a variety of tasks through planning with no additional training. Furthermore, such an agent can internally represent the complex dynamics of the real-world and therefore can acquire a representation useful for a variety of visual perception tasks. This makes predicting the future frames of a video, conditioned on the observed past and potentially future actions, an interesting task which remains exceptionally challenging despite many recent advances. Existing video prediction models have shown promising results on simple narrow benchmarks but they generate low quality predictions on real-life datasets with more complicated dynamics or broader domain. There is a growing body of evidence that underfitting on the training data is one of the primary causes for the low quality predictions. In this paper, we argue that the inefficient use of parameters in the current video models is the main reason for underfitting. Therefore, we introduce a new architecture, named FitVid, which is capable of fitting the common benchmarks so well that it begins to suffer from overfitting -- while having similar parameter count as the current state-of-the-art models. We analyze the consequences of overfitting, illustrating how it can produce unexpected outcomes such as generating high quality output by repeating the training data, and how it can be mitigated using existing image augmentation techniques. As a result, FitVid outperforms the current state-of-the-art models across four different video prediction benchmarks on four different metrics. | ['Dumitru Erhan', 'Chelsea Finn', 'Sergey Levine', 'Suraj Nair', 'Mohammad Taghi Saffar', 'Mohammad Babaeizadeh'] | 2021-09-29 | null | null | null | null | ['image-augmentation', 'video-prediction'] | ['computer-vision', 'computer-vision'] | [ 3.20720553e-01 2.00340763e-01 -2.95075297e-01 -2.33153880e-01
-1.27407327e-01 -2.36010209e-01 1.03754878e+00 -2.31251314e-01
-3.44310254e-01 7.05232084e-01 3.59450310e-01 -2.93511122e-01
2.33945549e-01 -6.04981422e-01 -1.05795944e+00 -4.33702916e-01
-1.99707553e-01 4.67374444e-01 6.28012717e-01 -3.55337381e-01
-6.04854971e-02 2.53946662e-01 -1.80929732e+00 4.38183546e-01
4.79033321e-01 9.90803659e-01 1.92650735e-01 6.41267836e-01
1.14553913e-01 1.23810148e+00 -4.85558510e-01 -4.27381963e-01
3.58621955e-01 -4.53125358e-01 -7.12914169e-01 2.96135783e-01
4.58299160e-01 -5.75629532e-01 -6.23971939e-01 6.77935243e-01
-2.34744437e-02 1.03868507e-01 4.56004292e-01 -1.52220023e+00
-3.16411823e-01 2.16837391e-01 -2.26496473e-01 3.06983322e-01
4.68483806e-01 6.63567364e-01 7.79569566e-01 -5.35212755e-01
8.52810025e-01 1.24483848e+00 7.19956577e-01 7.76489258e-01
-1.33445430e+00 -4.74340022e-01 4.58159626e-01 4.65027690e-01
-9.60607052e-01 -4.53112781e-01 6.21320128e-01 -5.08756161e-01
1.28355598e+00 2.08763599e-01 9.63063657e-01 1.55390370e+00
3.66028190e-01 7.80788779e-01 9.75409985e-01 -1.03941463e-01
4.25773412e-02 -2.98440475e-02 -3.69089395e-01 6.43168092e-01
2.73418557e-02 4.63417411e-01 -5.06432891e-01 -7.53994146e-03
8.41704965e-01 7.00042993e-02 -3.62832308e-01 -5.09563625e-01
-1.43985534e+00 5.81571937e-01 3.12614709e-01 6.98802993e-02
-4.26052660e-01 3.70143443e-01 4.60069388e-01 3.72312278e-01
2.97376961e-01 6.79618061e-01 -5.48604906e-01 -4.69433606e-01
-7.31458902e-01 4.75162685e-01 6.56592250e-01 7.62524843e-01
5.56849420e-01 3.10308099e-01 -5.88240661e-02 3.96863580e-01
1.49138585e-01 1.65849730e-01 6.63232923e-01 -1.27687824e+00
3.71818811e-01 5.85196853e-01 3.61028701e-01 -1.02055299e+00
-4.80917156e-01 -3.22751194e-01 -8.06155384e-01 4.89305913e-01
6.14427745e-01 -9.28225443e-02 -1.13072181e+00 1.90595877e+00
-4.59126383e-02 5.05856991e-01 9.19594839e-02 8.59983325e-01
4.84971523e-01 9.86277163e-01 2.15379983e-01 -2.27721661e-01
9.22927439e-01 -1.23679471e+00 -4.91307318e-01 -7.61852801e-01
4.04570520e-01 -3.82492393e-01 1.13967729e+00 4.97230381e-01
-9.11117971e-01 -8.01097333e-01 -1.09846675e+00 4.09393042e-01
-1.48241103e-01 -2.63356924e-01 9.38951612e-01 3.10128450e-01
-1.05071115e+00 8.55461419e-01 -1.08498836e+00 -4.51969266e-01
3.42296571e-01 2.96201855e-01 -6.29678130e-01 -7.74742886e-02
-8.94875586e-01 1.11970723e+00 3.79071474e-01 -3.16411465e-01
-1.01837730e+00 -7.41159260e-01 -7.81943679e-01 4.86570485e-02
5.08930743e-01 -7.79887140e-01 1.32278848e+00 -1.18897748e+00
-1.36759377e+00 5.04000306e-01 -2.33881418e-02 -8.63518655e-01
6.72764838e-01 -2.45159730e-01 -4.90544826e-01 -3.39563787e-02
-2.18255907e-01 1.02237296e+00 1.04123724e+00 -1.21472168e+00
-6.49890363e-01 -2.02274621e-02 2.33147591e-01 1.60469249e-01
-9.50197577e-02 -2.91731894e-01 -5.96707582e-01 -6.87212944e-01
-2.45332360e-01 -1.10044515e+00 -4.33188230e-01 -3.74161266e-02
5.83619662e-02 -1.15295183e-02 7.52721608e-01 -3.20328623e-01
1.14506185e+00 -2.04743600e+00 1.79089427e-01 -2.49356523e-01
1.55652836e-01 3.84597063e-01 -3.15863788e-01 3.32099497e-01
-6.33884668e-02 2.81900585e-01 1.37568220e-01 -4.14595127e-01
-2.55658805e-01 5.78414500e-01 -4.79546696e-01 1.77273303e-01
4.07049388e-01 8.30645561e-01 -8.93011689e-01 -2.03115568e-01
3.68958622e-01 4.43980604e-01 -6.43955708e-01 3.59944850e-01
-6.59555912e-01 7.03535914e-01 -1.82805628e-01 2.62310535e-01
1.77994654e-01 -4.53019023e-01 1.57475367e-01 -1.02250986e-01
1.38733268e-01 1.48502141e-01 -9.43119228e-01 1.54819393e+00
-7.21825808e-02 8.24465454e-01 -4.19282734e-01 -9.40108418e-01
7.66240537e-01 3.71611029e-01 5.92272639e-01 -8.87438118e-01
-7.29915649e-02 -6.49664700e-02 2.78976113e-01 -6.03674412e-01
4.65706885e-01 -2.54325598e-01 1.81824699e-01 3.02369744e-01
4.17901352e-02 -3.24660323e-06 3.31287682e-01 1.58855125e-01
1.40774572e+00 4.97147977e-01 1.89149767e-01 2.51563430e-01
2.33282313e-01 2.89895564e-01 8.91149163e-01 8.38971972e-01
-3.39621514e-01 6.09669745e-01 4.67141926e-01 -9.64247406e-01
-1.44712698e+00 -7.74608970e-01 3.16201478e-01 8.92020166e-01
3.26326527e-02 -6.33965194e-01 -5.01680911e-01 -6.30743086e-01
-1.27757773e-01 6.55649722e-01 -6.38253808e-01 -3.48116636e-01
-6.09344482e-01 -7.19492972e-01 3.36628079e-01 6.20053351e-01
3.43058825e-01 -1.38525450e+00 -1.07516980e+00 4.04277503e-01
-2.02052474e-01 -1.56861639e+00 1.26872271e-01 -5.23244925e-02
-9.42310393e-01 -1.07214093e+00 -4.82865334e-01 -3.70392114e-01
3.61608893e-01 2.68482774e-01 1.61787188e+00 2.54377455e-01
2.91599110e-02 5.05455196e-01 -5.05293310e-01 -4.46700215e-01
-8.30261886e-01 -2.31479302e-01 1.95204005e-01 -1.25964358e-01
2.12466151e-01 -6.47602081e-01 -5.59884548e-01 3.41169775e-01
-8.66240799e-01 6.23204827e-01 5.72887301e-01 8.69921744e-01
4.82385874e-01 1.91802941e-02 4.01860058e-01 -7.45850563e-01
3.05683643e-01 -4.32807982e-01 -4.92042065e-01 4.47827391e-02
-5.15607417e-01 1.40050590e-01 8.88269126e-01 -7.04928994e-01
-9.72112596e-01 1.73774749e-01 1.16646037e-01 -7.09528089e-01
-3.30737412e-01 3.58716726e-01 9.81794447e-02 1.28954887e-01
6.09596014e-01 2.66152650e-01 2.79633254e-01 -3.02602172e-01
6.22332133e-02 3.80535796e-02 6.50224924e-01 -2.85771430e-01
7.22282767e-01 5.93377948e-01 1.35663033e-01 -5.77230215e-01
-8.89924467e-01 -1.31814823e-01 -3.05899829e-01 -4.07316238e-01
6.72784269e-01 -1.00611484e+00 -6.24823213e-01 4.87779379e-01
-1.12253845e+00 -7.29976118e-01 -4.39632446e-01 5.06970286e-01
-8.82711768e-01 2.38678530e-01 -4.38789427e-01 -5.98818302e-01
7.04222992e-02 -1.33268774e+00 6.46028161e-01 1.35255173e-01
-3.18990976e-01 -7.82756269e-01 1.16859935e-01 2.35022500e-01
4.54104185e-01 4.19582039e-01 7.59502411e-01 -4.92914408e-01
-7.81723022e-01 -8.67950320e-02 -1.69516727e-02 4.57406133e-01
-2.33968064e-01 9.32258889e-02 -7.57816792e-01 -2.80513436e-01
-4.77633551e-02 -2.48673484e-01 9.81230617e-01 2.31589407e-01
1.13783944e+00 -3.18158239e-01 -2.63390422e-01 6.03418291e-01
1.31470203e+00 3.00546914e-01 8.71523142e-01 6.47130430e-01
5.03240466e-01 3.47206324e-01 7.41970301e-01 4.29959595e-01
4.64662224e-01 9.21353340e-01 9.80239987e-01 1.97379831e-02
-1.60497725e-01 -2.94020653e-01 5.98083735e-01 4.81660604e-01
-3.81409585e-01 -3.59905601e-01 -9.93926764e-01 5.04154623e-01
-2.24174953e+00 -1.34409571e+00 9.83951986e-02 2.07643366e+00
4.46878672e-01 4.92576033e-01 3.38526517e-01 2.05125272e-01
2.07419872e-01 2.78402776e-01 -6.40709817e-01 -2.73273170e-01
-3.45753953e-02 -8.39168504e-02 3.74022186e-01 1.25167266e-01
-1.15487099e+00 9.00847137e-01 6.84137154e+00 4.69050020e-01
-1.18212700e+00 -9.71127227e-02 7.86234021e-01 1.57351624e-02
3.91725861e-02 9.24201608e-02 -6.05824709e-01 4.98577446e-01
1.03160310e+00 4.35059369e-02 4.61379379e-01 8.28244567e-01
3.61836642e-01 -1.16538055e-01 -1.34012091e+00 9.19569254e-01
9.60385278e-02 -1.73021054e+00 2.98890173e-01 1.00302152e-01
5.96400082e-01 2.11876854e-01 4.39993478e-02 5.39872050e-01
2.10263729e-01 -1.28915954e+00 7.84832954e-01 7.22012699e-01
5.38759291e-01 -3.88436407e-01 6.66415632e-01 6.65024459e-01
-8.26594114e-01 -3.62470955e-01 -2.85486132e-01 -5.36153853e-01
3.21797490e-01 1.47057116e-01 -9.32828903e-01 1.63167834e-01
6.73057199e-01 9.21097100e-01 -7.68764496e-01 1.17079806e+00
6.79947110e-03 6.79671228e-01 -3.16921771e-01 2.82571673e-01
4.23391789e-01 1.13582417e-01 5.60538948e-01 9.82600451e-01
4.49297100e-01 -9.23526734e-02 2.69241750e-01 5.58836281e-01
1.11696951e-01 -3.39630872e-01 -7.80014873e-01 1.02562634e-02
1.95949554e-01 8.48761916e-01 -4.84869480e-01 -4.16719764e-01
-7.31653392e-01 7.48326063e-01 3.77374172e-01 3.07361364e-01
-1.04155195e+00 5.27491391e-01 7.41244376e-01 3.36716384e-01
3.66340518e-01 -2.36583680e-01 4.32691276e-02 -1.25958407e+00
2.40769871e-02 -1.25905597e+00 2.56456554e-01 -9.33568895e-01
-9.45582867e-01 6.61131144e-01 -1.00912631e-01 -1.48345816e+00
-6.10758781e-01 -6.62072539e-01 -6.21773243e-01 2.18178645e-01
-1.45992041e+00 -1.09933913e+00 -4.78954285e-01 5.27689695e-01
7.58023739e-01 -3.67365330e-01 8.31210673e-01 8.94628558e-03
-5.17974138e-01 2.19894126e-01 -3.10935885e-01 -8.21460187e-02
5.68112314e-01 -1.02948380e+00 7.08333433e-01 8.70296478e-01
4.23917979e-01 -4.69741551e-03 1.08156919e+00 -4.27583098e-01
-1.33467937e+00 -1.10215998e+00 4.34432417e-01 -6.02349102e-01
6.74854457e-01 -2.50314288e-02 -9.40195441e-01 8.76815617e-01
2.94964388e-02 1.56847909e-01 4.64415044e-01 -2.02626452e-01
-3.70842993e-01 -3.41466255e-02 -8.18278790e-01 7.98883498e-01
1.24858403e+00 -1.23683222e-01 -4.83707607e-01 1.68187916e-01
7.44563341e-01 -3.95181984e-01 -8.14264238e-01 5.74172556e-01
5.92079341e-01 -1.50318170e+00 1.12924778e+00 -8.51771295e-01
7.69045651e-01 -2.21272543e-01 1.44134862e-02 -1.33198464e+00
-4.24605489e-01 -7.55417287e-01 -5.92470586e-01 8.29622567e-01
3.81416947e-01 -2.55799949e-01 1.03954363e+00 7.17704892e-01
-2.28731446e-02 -7.40839541e-01 -8.60850871e-01 -8.35756898e-01
-2.06895083e-01 -6.65944755e-01 5.87326467e-01 6.81393564e-01
-1.24423457e-02 3.99311006e-01 -8.31799448e-01 -1.28443718e-01
3.13169926e-01 -1.10408030e-01 1.23526442e+00 -1.10967350e+00
-4.63925064e-01 -3.41923088e-01 -8.99063110e-01 -1.14744258e+00
1.09712586e-01 -2.82513678e-01 -1.66021466e-01 -1.31310880e+00
1.54626593e-01 -3.74785066e-01 -2.14970499e-01 5.36009550e-01
-1.25392765e-01 1.79546878e-01 8.49453926e-01 2.97609627e-01
-7.68973172e-01 4.95627463e-01 1.44029307e+00 -1.95604526e-02
-1.42442569e-01 5.55051155e-02 -4.36214417e-01 1.04840577e+00
8.53771150e-01 -1.94012865e-01 -4.61455762e-01 -4.34098452e-01
3.50473553e-01 1.52121544e-01 6.56389117e-01 -1.52176392e+00
9.32321418e-03 -3.30371350e-01 4.35486168e-01 -2.45844498e-01
7.05933452e-01 -1.08963239e+00 3.97234470e-01 4.83223975e-01
-1.28120527e-01 3.33603472e-01 3.55771333e-01 8.07406366e-01
-2.20463559e-01 6.73506781e-02 6.15424931e-01 -3.29468727e-01
-1.24590170e+00 2.84419715e-01 -5.34884751e-01 -7.26577826e-03
1.18303430e+00 -4.42595750e-01 -4.88531917e-01 -7.23973870e-01
-8.05499911e-01 1.25613376e-01 7.77410448e-01 8.04024339e-01
5.52949369e-01 -1.14670122e+00 -5.27742743e-01 2.92089075e-01
2.52147093e-02 -1.76942557e-01 1.05586782e-01 7.50642002e-01
-4.07675594e-01 2.80422509e-01 -5.58307409e-01 -6.74112856e-01
-1.17842507e+00 7.59104788e-01 3.01798612e-01 -5.58883548e-01
-9.37756121e-01 5.08304358e-01 2.79627442e-01 -2.15453543e-02
2.73902327e-01 -3.93675536e-01 -3.04399878e-01 -2.70722330e-01
4.85443324e-01 1.06565230e-01 -2.18289822e-01 -7.83875167e-01
-1.33486152e-01 2.21000612e-01 -7.94057027e-02 1.51598558e-01
1.58442926e+00 -2.13613808e-02 3.74668300e-01 3.09715837e-01
6.45311952e-01 -4.68835771e-01 -1.88712788e+00 -1.25698531e-02
-2.15553105e-01 -5.43075383e-01 -2.03414634e-01 -7.46618271e-01
-1.04761410e+00 7.68285811e-01 3.97385120e-01 3.62143219e-01
1.13142383e+00 -1.62608370e-01 7.16195941e-01 3.19030881e-01
4.87644494e-01 -9.49499309e-01 2.97755957e-01 6.12704873e-01
9.21106160e-01 -1.42454743e+00 1.14920884e-01 -3.03690404e-01
-1.06751835e+00 9.63421643e-01 9.10483122e-01 -2.20671564e-01
3.69346827e-01 2.34843761e-01 -6.59464598e-02 -6.03541024e-02
-1.34601235e+00 -9.57701355e-02 1.50261104e-01 8.31244826e-01
1.22537613e-01 -1.52096406e-01 2.32169718e-01 4.17178541e-01
-2.08528817e-01 2.70180076e-01 7.63614655e-01 7.86282539e-01
-2.67501622e-01 -1.03521419e+00 -2.68046886e-01 5.88753998e-01
-3.82196933e-01 2.53206074e-01 -1.53672919e-01 9.36186910e-01
1.35164127e-01 7.96477437e-01 2.14330271e-01 -6.78619266e-01
2.96139687e-01 8.47088024e-02 4.73184198e-01 -4.21716988e-01
-4.29491609e-01 -1.80218190e-01 1.82431340e-01 -9.74261880e-01
-7.31057286e-01 -6.30175829e-01 -9.97761905e-01 -3.31786752e-01
5.69971800e-02 -2.95983613e-01 1.70697093e-01 9.33402002e-01
5.48487902e-01 6.08319104e-01 2.25593984e-01 -1.31699526e+00
-5.09650230e-01 -8.35977495e-01 -7.30684623e-02 8.83047819e-01
4.19524431e-01 -8.86248767e-01 -2.22335339e-01 3.21681023e-01] | [8.411170959472656, 0.3922211229801178] |
5f8e58eb-392b-4d3f-979d-dc0bbbebe1e6 | power-system-anomaly-detection-and | 2209.12629 | null | https://arxiv.org/abs/2209.12629v2 | https://arxiv.org/pdf/2209.12629v2.pdf | Power System Anomaly Detection and Classification Utilizing WLS-EKF State Estimation and Machine Learning | Power system state estimation is being faced with different types of anomalies. These might include bad data caused by gross measurement errors or communication system failures. Sudden changes in load or generation can be considered as anomaly depending on the implemented state estimation method. Additionally, considering power grid as a cyber physical system, state estimation becomes vulnerable to false data injection attacks. The existing methods for anomaly classification cannot accurately classify (discriminate between) the above mentioned three types of anomalies, especially when it comes to discrimination between sudden load changes and false data injection attacks. This paper presents a new algorithm for detecting anomaly presence, classifying the anomaly type and identifying the origin of the anomaly, i.e., measurements that contain gross errors in case of bad data, or buses associated with loads experiencing a sudden change, or state variables targeted by false data injection attack. The algorithm combines analytical and machine learning (ML) approaches. The first stage exploits an analytical approach to detect anomaly presence by combining $\chi^2$-test and anomaly detection index. The second stage utilizes ML for classification of anomaly type and identification of its origin, with particular reference to discrimination between sudden load changes and false data injection attacks. The proposed ML based method is trained to be independent of the network configuration which eliminates retraining of the algorithm after network topology changes. The results obtained by implementing the proposed algorithm on IEEE 14 bus test system demonstrate the accuracy and effectiveness of the proposed algorithm. | ['Vladimir Terzija', 'Elena Gryazina', 'Victor Levi', 'Dragan Ćetenović', 'Mile Mitrovic', 'Sajjad Asefi'] | 2022-09-26 | null | null | null | null | ['anomaly-classification'] | ['computer-vision'] | [ 6.21017069e-02 -3.34486783e-01 1.57280490e-02 1.20686352e-01
-1.15068004e-01 -7.77667046e-01 5.68503976e-01 8.43181193e-01
6.08689710e-02 9.78025854e-01 -4.62111562e-01 -6.84603870e-01
-3.51280391e-01 -9.50812399e-01 -2.17420042e-01 -8.96668553e-01
-5.05779326e-01 5.73344350e-01 3.35604459e-01 -6.10321797e-02
4.64076728e-01 9.63208556e-01 -1.68438935e+00 -8.64435732e-02
8.45337629e-01 1.08110225e+00 -4.96589482e-01 7.63011992e-01
-1.36841282e-01 4.85785514e-01 -1.34071159e+00 5.41653275e-01
3.87757838e-01 -5.66561878e-01 -4.46772724e-01 2.63213720e-02
-3.43885928e-01 -3.18421096e-01 1.27178818e-01 1.26189506e+00
3.44867826e-01 -8.75432566e-02 6.61681712e-01 -1.95918822e+00
2.10985124e-01 2.97602564e-01 -4.28502381e-01 1.08166420e+00
7.08840966e-01 2.94664383e-01 4.24855649e-01 -4.96189058e-01
9.92811099e-02 7.36367106e-01 4.08925146e-01 -1.71462297e-02
-1.30576730e+00 -8.06059480e-01 -7.27661774e-02 4.80834097e-01
-1.31514668e+00 -1.68279201e-01 1.04981256e+00 -4.93579149e-01
1.06806433e+00 4.28926796e-01 5.39355814e-01 5.28246760e-01
6.32306635e-01 2.54985154e-01 1.19001031e+00 -2.93012828e-01
7.06600487e-01 1.94896892e-01 4.98229116e-01 1.40084386e-01
7.21000910e-01 2.20714822e-01 1.26601681e-01 -5.12704849e-01
1.81382030e-01 -9.39880311e-02 -2.85347134e-01 1.25564486e-01
-7.12648571e-01 5.82306683e-01 -4.30025049e-02 7.35957444e-01
-7.09652007e-01 -3.35179389e-01 7.37823188e-01 7.60716319e-01
9.04326960e-02 2.48486117e-01 -5.84575295e-01 -2.67248094e-01
-8.11804295e-01 -1.90068781e-01 9.56356466e-01 5.88039041e-01
3.48847628e-01 1.08296204e+00 3.00985932e-01 5.19382991e-02
1.66479260e-01 5.89388669e-01 9.78548169e-01 1.74282432e-01
5.31085841e-02 8.86663795e-01 1.23961698e-02 -1.13343251e+00
-5.37147164e-01 -7.16122389e-01 -7.68047035e-01 6.72846556e-01
4.75891113e-01 -3.43663841e-01 -5.66859424e-01 1.29473293e+00
3.59992176e-01 4.67672557e-01 3.65659535e-01 2.62491107e-01
1.21403180e-01 8.91594291e-01 -2.81448275e-01 -7.30927289e-01
1.03177381e+00 8.10259301e-03 -1.13621426e+00 2.28827447e-01
6.72321737e-01 -6.98978484e-01 2.57420003e-01 7.91562676e-01
-6.49000108e-01 -3.18705082e-01 -1.31363654e+00 1.18348885e+00
-8.97479415e-01 -1.62992150e-01 1.20297536e-01 8.39049637e-01
-6.02042019e-01 5.14278829e-01 -7.92930186e-01 -4.47780609e-01
-2.49707788e-01 5.03303528e-01 -2.54651099e-01 5.48885643e-01
-1.28745568e+00 1.00016189e+00 9.83221650e-01 2.25326166e-01
-5.33991635e-01 -4.96375263e-01 -6.17824137e-01 7.41819739e-02
2.46170893e-01 2.19280928e-01 6.97745740e-01 -1.08707237e+00
-1.11839819e+00 1.30040914e-01 1.24686845e-01 -8.87275875e-01
4.08883780e-01 3.59435588e-01 -1.44642115e+00 1.77555874e-01
-1.92916915e-01 -7.93083668e-01 7.76409030e-01 -8.93241704e-01
-7.29004323e-01 -4.49346602e-01 -4.28476125e-01 -3.10172141e-01
5.39454892e-02 -3.03289980e-01 1.02515972e+00 -3.83093089e-01
3.62318754e-01 -4.45981264e-01 5.09584583e-02 -6.36932671e-01
-5.26750326e-01 -1.74090620e-02 1.62936711e+00 -6.19300306e-01
1.49548197e+00 -1.81256449e+00 -6.61063969e-01 1.16194344e+00
-2.78210729e-01 5.29584706e-01 4.42450345e-01 8.79705906e-01
-6.24516666e-01 -1.32479385e-01 -1.03416800e-01 4.52889591e-01
-3.32139991e-03 4.66546476e-01 -4.21079576e-01 7.89491713e-01
2.62787342e-01 6.13230802e-02 -8.13423574e-01 1.47831604e-01
6.36875808e-01 -4.50671092e-03 9.04514119e-02 1.92772731e-01
9.28615481e-02 5.15730321e-01 -4.99366850e-01 5.45719683e-01
4.78554070e-01 2.60637224e-01 2.49082260e-02 -2.51493871e-01
-2.22161755e-01 4.19307649e-02 -1.95044243e+00 3.34124476e-01
-3.30188543e-01 5.84107280e-01 -7.52081051e-02 -1.53398013e+00
1.13089299e+00 8.32662821e-01 5.48390925e-01 -5.40414691e-01
2.96781540e-01 5.91858685e-01 4.05127764e-01 -2.24323601e-01
-3.82024199e-02 2.19841361e-01 6.95625469e-02 6.58794641e-01
2.77676936e-02 2.18693614e-01 4.96199548e-01 -6.70717210e-02
1.17542446e+00 -4.70996410e-01 9.15239275e-01 -4.06449616e-01
1.11718845e+00 -1.11446083e-01 7.76719689e-01 4.54897791e-01
-1.83365464e-01 -4.35863495e-01 7.79539704e-01 -4.46423113e-01
-7.22469270e-01 -1.11482954e+00 -3.08209687e-01 3.46313417e-02
-1.13182306e-01 9.02147666e-02 -1.37333959e-01 -7.47049212e-01
5.13169654e-02 1.18608809e+00 -3.19198430e-01 -6.07760549e-01
-4.15318102e-01 -1.09613347e+00 4.36570346e-01 1.56163111e-01
5.91620803e-01 -9.32568550e-01 -7.63490140e-01 5.75730264e-01
5.38797617e-01 -6.90652430e-01 4.46398497e-01 5.45893848e-01
-9.31062400e-01 -1.57323635e+00 2.59902686e-01 -2.85882503e-01
9.10025418e-01 -3.99923176e-01 9.45483804e-01 3.72339070e-01
-4.10039306e-01 3.77740622e-01 -3.69966626e-01 -5.00951052e-01
-9.95040834e-01 -4.80253845e-01 4.79759663e-01 1.58157095e-01
5.17561734e-01 -8.74198616e-01 -2.71559626e-01 3.57955456e-01
-1.05052733e+00 -9.72706258e-01 3.48274529e-01 6.83276594e-01
8.46154764e-02 1.00594759e+00 1.12790251e+00 -8.74880075e-01
7.76095569e-01 -9.89801764e-01 -1.13648105e+00 4.00354601e-02
-1.11089921e+00 -7.69662187e-02 1.11870492e+00 -3.16902459e-01
-6.01271868e-01 -1.34785116e-01 -2.29948714e-01 -2.87733003e-02
-7.57408679e-01 5.39171696e-01 -2.54089713e-01 -1.83489755e-01
4.73888189e-01 4.81812388e-01 1.52022198e-01 -3.51556510e-01
-4.50667590e-01 6.13387585e-01 3.42692703e-01 -1.77255809e-01
1.15547764e+00 1.43554226e-01 3.65542293e-01 -1.01450026e+00
-2.05009967e-01 -3.50135088e-01 -5.04328370e-01 -3.30949217e-01
2.90147662e-01 -4.87522066e-01 -8.39194655e-01 7.95233369e-01
-7.85250366e-01 2.84037232e-01 -2.53400624e-01 4.82150644e-01
-4.04746123e-02 3.87099147e-01 -3.93468857e-01 -1.15584862e+00
-4.01307434e-01 -9.90605295e-01 8.60823318e-02 3.28170091e-01
-1.58076718e-01 -1.40679896e+00 5.87896779e-02 -4.63856280e-01
6.15423799e-01 7.88630426e-01 9.73848283e-01 -1.62314737e+00
-1.02431498e-01 -1.12370884e+00 3.41942221e-01 5.17500699e-01
5.02088487e-01 3.86591256e-01 -6.21726334e-01 -7.56465137e-01
3.01580340e-01 3.52313548e-01 -3.25297415e-01 5.20215034e-02
4.93671298e-01 -4.81542319e-01 -1.55882239e-01 7.43375644e-02
1.80443525e+00 9.97031569e-01 5.90460777e-01 3.60108644e-01
1.64735526e-01 -1.04975045e-01 5.66814303e-01 6.70581281e-01
-2.85050452e-01 5.15002012e-01 6.06910229e-01 -1.76361963e-01
6.48415446e-01 2.74097830e-01 5.11839509e-01 5.64310193e-01
3.16661716e-01 -1.84677452e-01 -8.70912492e-01 2.96737522e-01
-1.43561089e+00 -1.25661325e+00 -3.57707649e-01 2.42236567e+00
3.93911034e-01 6.87483311e-01 2.36186832e-01 1.27819252e+00
7.88331568e-01 -1.96534589e-01 -6.21543825e-01 -7.94998884e-01
-1.91511557e-01 3.17690670e-01 3.49760413e-01 4.92724031e-01
-7.90128052e-01 4.44282331e-02 4.72576141e+00 5.40370166e-01
-1.36205792e+00 -2.11110324e-01 2.38212049e-01 4.55971211e-01
3.40269297e-01 -1.03873365e-01 -6.50865734e-01 7.92953849e-01
1.24081051e+00 -4.70085502e-01 -2.77014021e-02 6.91831470e-01
4.75711554e-01 -4.52151269e-01 -9.61545944e-01 5.17036498e-01
3.01908869e-02 -7.13418543e-01 9.79042030e-04 1.74533110e-03
5.25584817e-01 -3.69216383e-01 -4.15896773e-01 2.23263904e-01
-5.90823926e-02 -5.25242686e-01 3.25766593e-01 4.27458405e-01
-1.87563837e-01 -1.02312863e+00 1.32431829e+00 3.91017348e-01
-1.22000694e+00 -4.28590804e-01 2.24129543e-01 -3.80224496e-01
3.34836334e-01 8.55160058e-01 -9.69831705e-01 8.62179518e-01
3.01723987e-01 4.91591185e-01 -5.76416433e-01 1.20219219e+00
-1.88243493e-01 1.02566218e+00 -6.66703463e-01 1.27326742e-01
1.32914037e-01 -2.04719082e-01 1.06107438e+00 8.07106614e-01
3.44446778e-01 -1.42005250e-01 3.55838895e-01 5.67072093e-01
8.43778610e-01 -2.81917714e-02 -7.82925189e-01 2.31753856e-01
7.42889166e-01 1.03961122e+00 -1.05079818e+00 -3.70454371e-01
-1.74330294e-01 4.17672902e-01 -4.97024804e-01 4.02713835e-01
-4.63125229e-01 -6.97522104e-01 3.73818159e-01 5.10684997e-02
-6.51392788e-02 2.01657787e-02 -1.49498522e-01 -9.09990907e-01
6.05050623e-02 -6.94881856e-01 7.13825822e-01 -2.93455720e-01
-1.28430367e+00 5.63563764e-01 1.89946428e-01 -1.60847235e+00
-8.39627266e-01 -4.57467437e-01 -1.42971623e+00 8.94409239e-01
-1.10349619e+00 -5.74893773e-01 7.48355389e-02 7.45703697e-01
2.85190523e-01 -3.58490676e-01 8.77160311e-01 3.13754290e-01
-8.62805545e-01 4.41372156e-01 2.27887467e-01 3.41489375e-01
-8.38907585e-02 -1.35888231e+00 -1.94574133e-01 1.52441251e+00
-1.40323490e-01 1.57244399e-01 1.17440104e+00 -7.49155581e-01
-1.13253272e+00 -7.38886595e-01 4.91981268e-01 1.51653036e-01
9.45529878e-01 -1.74069941e-01 -1.16911209e+00 6.21926427e-01
1.76940113e-01 1.53382927e-01 6.18674457e-01 -3.91043097e-01
7.69004151e-02 -2.64808744e-01 -1.73208463e+00 1.84154272e-01
-1.72353566e-01 -1.56148940e-01 -6.91590786e-01 1.77338317e-01
-3.36817652e-01 4.07916419e-02 -1.11597002e+00 3.75666291e-01
-5.06584086e-02 -1.11826205e+00 6.36974752e-01 -3.74062687e-01
-5.19505203e-01 -7.44906962e-01 1.10046461e-01 -1.54484951e+00
1.85717233e-02 -5.15758693e-01 -3.30796748e-01 1.29823422e+00
3.75021398e-01 -1.38854027e+00 4.70651090e-01 2.03465298e-01
-7.41908401e-02 -3.89580995e-01 -1.20913541e+00 -9.14352238e-01
-3.58880609e-01 -2.59545565e-01 7.14036286e-01 1.29163635e+00
3.93527329e-01 -1.15103431e-01 1.14879809e-01 7.53904223e-01
5.99744380e-01 -3.05832382e-02 5.32928586e-01 -1.32079124e+00
-1.79066971e-01 -5.10698736e-01 -1.16955280e+00 1.87408179e-01
-5.01524322e-02 -4.28991854e-01 -3.30447763e-01 -9.92383897e-01
-8.90398383e-01 -2.18513086e-01 -4.94231641e-01 3.62401098e-01
6.38061240e-02 1.99767992e-01 -1.72544301e-01 2.63892617e-02
1.74009681e-01 2.51250207e-01 2.12466553e-01 2.25868523e-01
-1.30690306e-01 6.10646009e-01 3.45483065e-01 8.11871111e-01
1.26422167e+00 -2.77589977e-01 -4.26271141e-01 4.88989323e-01
1.48327768e-01 2.85160959e-01 3.07110071e-01 -1.54328120e+00
1.44296661e-01 -1.64641403e-02 5.11655867e-01 -8.89418900e-01
-3.60594779e-01 -1.49092722e+00 4.71239060e-01 1.24111354e+00
2.21583724e-01 8.34556520e-01 4.39804554e-01 6.61132932e-01
-3.70588601e-01 -3.55979711e-01 7.47211874e-01 2.36664146e-01
-8.25937986e-01 -2.36457497e-01 -9.10982072e-01 -3.09394956e-01
1.57091081e+00 -5.56564629e-01 -3.15952599e-01 -3.39729697e-01
-9.83510852e-01 2.73510724e-01 1.00350104e-01 1.84870929e-01
3.78057122e-01 -1.22949433e+00 -4.73627150e-01 7.00626731e-01
3.47675271e-02 -6.34273708e-01 -4.34069186e-02 9.16272819e-01
-4.82121080e-01 1.48117006e-01 -4.24559444e-01 -4.94987726e-01
-1.33385742e+00 4.48573083e-01 6.34634972e-01 -2.64780372e-01
-3.03172410e-01 -7.35049248e-02 -1.02158487e+00 2.26469129e-01
7.06070140e-02 -2.01453820e-01 -6.05711162e-01 1.78778395e-01
4.77488518e-01 7.16810644e-01 5.87095916e-01 -5.44585645e-01
-2.95102596e-01 2.89963722e-01 -5.38046658e-02 3.22396517e-01
8.74369860e-01 -7.92571455e-02 -2.61672378e-01 8.94618750e-01
8.54390264e-01 -1.58281654e-01 -6.63116693e-01 -1.06052600e-01
2.96309352e-01 -2.77737468e-01 -9.22904909e-03 -1.00824046e+00
-1.18499947e+00 6.38567030e-01 9.90891278e-01 1.15249693e+00
1.30800855e+00 -6.90688908e-01 3.46544713e-01 1.49506167e-01
4.36388403e-01 -1.03729475e+00 -4.22326326e-01 2.81634569e-01
3.41593236e-01 -8.96058500e-01 -1.12017468e-02 -2.09770957e-03
-3.78044359e-02 1.43040919e+00 5.68454027e-01 -4.67102170e-01
1.11846089e+00 4.30872589e-01 -3.81612740e-02 -1.92120060e-01
-7.73830056e-01 1.37858555e-01 1.06697783e-01 6.83418691e-01
4.96462286e-02 4.14896980e-02 -7.46609092e-01 9.51908156e-02
-8.64715204e-02 -1.91059738e-01 1.00594294e+00 1.35586774e+00
-4.62024987e-01 -6.94829166e-01 -7.45126069e-01 9.68148828e-01
-6.19422734e-01 3.73024166e-01 -2.60408549e-03 1.01123130e+00
1.05452858e-01 1.11609316e+00 4.75747615e-01 -1.91432029e-01
5.59726834e-01 3.13512743e-01 -2.06767246e-01 -2.58597553e-01
-7.84124434e-01 -2.20983177e-01 -8.50601718e-02 -4.66516435e-01
4.85591069e-02 -7.55740702e-01 -1.26482558e+00 -3.04540575e-01
-6.09140337e-01 7.82717824e-01 7.76013553e-01 1.26011920e+00
-5.21449558e-02 6.41800284e-01 9.65610981e-01 -4.60882723e-01
-6.41590118e-01 -8.79568517e-01 -9.76510227e-01 5.46285272e-01
4.44314212e-01 -8.29310894e-01 -1.11128676e+00 -2.98717678e-01] | [6.226116180419922, 2.5562891960144043] |
69121545-3e78-4671-8797-48a940089cc2 | universal-adversarial-attack-on-deep-learning | 2109.07142 | null | https://arxiv.org/abs/2109.07142v1 | https://arxiv.org/pdf/2109.07142v1.pdf | Universal Adversarial Attack on Deep Learning Based Prognostics | Deep learning-based time series models are being extensively utilized in engineering and manufacturing industries for process control and optimization, asset monitoring, diagnostic and predictive maintenance. These models have shown great improvement in the prediction of the remaining useful life (RUL) of industrial equipment but suffer from inherent vulnerability to adversarial attacks. These attacks can be easily exploited and can lead to catastrophic failure of critical industrial equipment. In general, different adversarial perturbations are computed for each instance of the input data. This is, however, difficult for the attacker to achieve in real time due to higher computational requirement and lack of uninterrupted access to the input data. Hence, we present the concept of universal adversarial perturbation, a special imperceptible noise to fool regression based RUL prediction models. Attackers can easily utilize universal adversarial perturbations for real-time attack since continuous access to input data and repetitive computation of adversarial perturbations are not a prerequisite for the same. We evaluate the effect of universal adversarial attacks using NASA turbofan engine dataset. We show that addition of universal adversarial perturbation to any instance of the input data increases error in the output predicted by the model. To the best of our knowledge, we are the first to study the effect of the universal adversarial perturbation on time series regression models. We further demonstrate the effect of varying the strength of perturbations on RUL prediction models and found that model accuracy decreases with the increase in perturbation strength of the universal adversarial attack. We also showcase that universal adversarial perturbation can be transferred across different models. | ['Venkataramana Runkana', 'Sagar Srinivas', 'Sri Harsha Nistala', 'Pradeep Rathore', 'Arghya Basak'] | 2021-09-15 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [ 3.22290361e-01 -2.64419109e-01 4.24185008e-01 2.17095509e-01
-3.97876799e-01 -9.33943689e-01 2.37922221e-01 2.90132344e-01
9.70451534e-02 5.57273746e-01 -7.12698698e-01 -6.20513737e-01
-4.02114660e-01 -1.03126597e+00 -1.04850972e+00 -8.45660806e-01
-3.55587065e-01 3.08190495e-01 7.43695721e-02 -3.46766680e-01
1.15471713e-01 1.04672813e+00 -1.31515074e+00 1.40209466e-01
3.63475204e-01 1.15385759e+00 -3.24245393e-01 9.09593880e-01
5.01392901e-01 6.99680805e-01 -1.09840035e+00 6.10040687e-02
5.93170583e-01 -1.85951814e-01 -4.68056321e-01 -1.95395231e-01
-1.93223536e-01 -2.77469277e-01 -6.34703994e-01 8.99634421e-01
6.07297301e-01 2.75112510e-01 6.12198949e-01 -1.36786938e+00
-2.84637958e-01 3.27722222e-01 -7.96095002e-03 2.09349498e-01
2.64203161e-01 4.98906583e-01 3.29377562e-01 -1.33449361e-01
2.94592917e-01 8.15682828e-01 6.07004821e-01 2.45131016e-01
-1.09470916e+00 -5.99171340e-01 -2.57317096e-01 -1.07276738e-02
-1.09070599e+00 1.72321036e-01 1.06330431e+00 -3.62115800e-01
9.16213989e-01 5.51222265e-01 2.10096180e-01 1.11666954e+00
1.07010329e+00 1.02622963e-01 8.42447937e-01 -3.36171448e-01
3.27779591e-01 -1.15279123e-01 -2.15323806e-01 2.82319218e-01
-7.89996311e-02 7.47929037e-01 1.68326467e-01 -2.54576206e-01
9.79755282e-01 3.24696302e-01 -2.02972561e-01 4.33115885e-02
-9.46188390e-01 6.03993416e-01 4.81788933e-01 2.33701065e-01
-5.73396027e-01 3.79342765e-01 7.15474606e-01 9.86494005e-01
3.33165638e-02 1.03980517e+00 -9.00019586e-01 -6.76134452e-02
-3.40864837e-01 8.46284032e-02 7.88411796e-01 7.17547715e-01
9.90515724e-02 6.76872611e-01 -8.83198157e-02 3.77446800e-01
-3.34338218e-01 7.62496471e-01 4.21652853e-01 -7.53689349e-01
2.27625832e-01 1.96321309e-01 1.35938808e-01 -9.85879302e-01
-4.70903277e-01 -4.46245462e-01 -8.01495969e-01 7.13598073e-01
4.37908232e-01 -5.46039224e-01 -9.94424284e-01 1.23553860e+00
5.01229465e-02 1.79502413e-01 2.39006743e-01 4.02093053e-01
-6.74130172e-02 6.27806842e-01 -1.78117529e-01 -2.53869891e-01
8.21746230e-01 -2.39083186e-01 -7.03332603e-01 2.56438971e-01
3.25335413e-01 -9.69909489e-01 9.60448623e-01 6.86138809e-01
-7.78449118e-01 -6.79624081e-01 -1.39112234e+00 7.42168844e-01
-6.87609613e-01 -4.26484197e-01 4.53225017e-01 9.26652670e-01
-1.33190826e-01 1.31823754e+00 -9.93041873e-01 4.75392416e-02
1.08546093e-01 6.74420893e-01 -3.90975147e-01 3.98022495e-02
-1.37439060e+00 9.97635067e-01 2.15298638e-01 2.12399095e-01
-1.32549810e+00 -9.44687486e-01 -5.58132410e-01 -1.73878357e-01
3.50461066e-01 -2.13421986e-01 1.12263048e+00 -7.69782603e-01
-1.44928217e+00 3.59501764e-02 4.81194526e-01 -7.43951619e-01
7.55615950e-01 -3.63760203e-01 -9.96829152e-01 -8.23286474e-02
-5.87264180e-01 -4.01927054e-01 1.18088424e+00 -1.04211020e+00
-1.43679261e-01 -1.30828381e-01 2.05624655e-01 -5.76286435e-01
-2.55820841e-01 -1.20949835e-01 4.85599875e-01 -8.58875394e-01
-1.39476895e-01 -1.03832710e+00 -2.96944380e-01 -3.36962372e-01
-5.22870004e-01 3.64915878e-01 1.21058977e+00 -6.03493810e-01
9.19821024e-01 -2.10161805e+00 -1.99346870e-01 4.64720696e-01
-3.23051989e-01 3.67248774e-01 1.10963486e-01 8.35690975e-01
-5.13857782e-01 -8.03643093e-03 -2.03250706e-01 3.91454995e-01
-9.24483612e-02 3.76719564e-01 -8.20194662e-01 6.56221986e-01
4.59499598e-01 7.36335754e-01 -8.21705818e-01 3.44593555e-01
5.21100581e-01 4.03693110e-01 -2.07819775e-01 2.72708893e-01
-2.30312243e-01 5.54735959e-01 -5.15807807e-01 6.39692903e-01
3.19961280e-01 3.87537688e-01 -2.05824837e-01 -2.56915390e-01
2.80693710e-01 -2.49917805e-01 -1.01535869e+00 7.66912758e-01
-7.19568372e-01 4.54579741e-01 -3.90704513e-01 -1.02952325e+00
1.01180613e+00 6.44429147e-01 6.08144760e-01 -5.03759086e-01
4.51932490e-01 2.28792593e-01 2.95996279e-01 -2.36664757e-01
2.13139474e-01 -5.61496198e-01 -4.00782764e-01 4.08278614e-01
-1.72426581e-01 -4.24133658e-01 -2.99807161e-01 -3.72018874e-01
1.41178119e+00 -1.81982666e-01 -4.55121286e-02 -1.38529867e-01
5.74342489e-01 -2.19030201e-01 3.61678690e-01 5.76842070e-01
7.63184279e-02 3.56846005e-01 5.98403335e-01 -6.14451230e-01
-1.21554363e+00 -1.21070790e+00 5.52162826e-02 5.72332561e-01
-1.34765282e-01 3.58597748e-02 -3.30463856e-01 -8.96720588e-01
3.57640445e-01 1.02194309e+00 -6.58664644e-01 -8.31423879e-01
-7.72477210e-01 -4.74100977e-01 8.95523667e-01 8.75235438e-01
5.48988804e-02 -1.03896892e+00 -6.40127242e-01 4.27302778e-01
6.51864111e-01 -1.04888606e+00 -2.19369963e-01 7.06397831e-01
-8.87377143e-01 -1.29998076e+00 -1.06892623e-01 -3.12572569e-01
6.90227151e-01 -4.42990929e-01 8.27410340e-01 4.38868850e-02
-6.32941127e-01 5.54362714e-01 -9.53100175e-02 -9.03810561e-01
-9.77088630e-01 -3.97155523e-01 4.51608330e-01 -2.41938770e-01
-1.58363804e-01 -6.64451361e-01 -2.41824657e-01 5.97494960e-01
-1.10664964e+00 -9.38019693e-01 8.38912502e-02 8.18734229e-01
5.12907922e-01 9.82324064e-01 8.51429343e-01 -7.65305221e-01
7.51942396e-01 -4.40233022e-01 -6.85293555e-01 1.02848513e-03
-5.48575938e-01 -1.16351321e-02 1.41569436e+00 -8.35223734e-01
-5.25499761e-01 -1.24068543e-01 -1.61248177e-01 -9.48289454e-01
-1.01367369e-01 4.33550537e-01 -2.21463382e-01 -3.54287624e-01
5.98206520e-01 -5.51631711e-02 1.88841131e-02 -2.23234057e-01
-2.55629029e-02 3.68599176e-01 6.39212072e-01 -6.28781796e-01
1.16216886e+00 1.80487126e-01 6.36665761e-01 -8.40317965e-01
-2.18433410e-01 1.03388742e-01 -4.48562831e-01 -3.96427691e-01
3.74600589e-01 -3.68167579e-01 -1.00793219e+00 6.58451200e-01
-8.75612736e-01 -4.06543344e-01 -6.47341251e-01 2.68312186e-01
-7.72852778e-01 4.04319346e-01 -6.75949514e-01 -7.98975050e-01
-2.26043478e-01 -1.08475399e+00 4.50486124e-01 -2.74949819e-01
-1.25060141e-01 -1.00799072e+00 -3.63381475e-01 -1.12124972e-01
5.46739280e-01 1.14752650e+00 1.10245585e+00 -9.44821119e-01
-2.87385672e-01 -1.13934743e+00 6.44085944e-01 9.43025470e-01
4.92427081e-01 1.80383250e-01 -8.14518392e-01 -5.64733207e-01
5.18402994e-01 -1.64980933e-01 1.08237080e-01 8.26201364e-02
1.19238961e+00 -2.15951443e-01 5.56432968e-03 3.06917459e-01
1.35830796e+00 6.67991459e-01 8.04939747e-01 2.63493985e-01
6.30989254e-01 3.35515618e-01 8.99251044e-01 4.06211615e-01
-8.85918379e-01 4.02510554e-01 8.16520989e-01 4.55567800e-02
4.33061719e-01 1.45120965e-02 5.15151143e-01 3.99225861e-01
-3.03179711e-01 -2.46159732e-01 -8.33792567e-01 2.12469921e-01
-1.22086561e+00 -8.65782201e-01 -3.80421579e-02 2.62709260e+00
4.43061441e-01 7.09705770e-01 -2.21758604e-01 7.59068668e-01
7.19900191e-01 -2.32654646e-01 -8.20731699e-01 -7.42422819e-01
1.61793724e-01 3.85576278e-01 9.36764419e-01 2.73897707e-01
-1.02960837e+00 5.23323774e-01 6.33412170e+00 6.50941432e-01
-1.27821624e+00 -9.70363095e-02 3.98223996e-01 -2.64464647e-01
1.01910390e-01 -3.81802350e-01 -2.25503296e-01 4.72615093e-01
1.31260228e+00 -2.80795097e-01 5.09194851e-01 8.16637397e-01
1.35412559e-01 4.87620205e-01 -1.37627959e+00 5.86766660e-01
-3.80358070e-01 -9.78497565e-01 -1.10005736e-01 1.39589086e-01
7.04260707e-01 -4.35308874e-01 1.62859470e-01 3.33137929e-01
3.31107348e-01 -1.27319992e+00 3.29799712e-01 5.78095675e-01
5.06607175e-01 -1.40202355e+00 1.09110880e+00 1.27684921e-01
-9.11840975e-01 -5.11530697e-01 -1.63454503e-01 -3.56766433e-02
2.80467778e-01 4.41704005e-01 -6.98771358e-01 7.69514680e-01
2.73122251e-01 2.61443723e-02 -1.27191275e-01 5.13683558e-01
-8.68560895e-02 6.56221747e-01 -4.09624428e-01 2.94468433e-01
8.48074853e-02 1.50140315e-01 6.65434718e-01 5.44846356e-01
4.11293209e-01 -2.78997868e-01 1.78502575e-01 4.87652361e-01
7.97024518e-02 -3.55892420e-01 -8.78760397e-01 -4.17076111e-01
4.01249111e-01 5.65978885e-01 -5.83273172e-01 1.59397170e-01
1.89019125e-02 8.73327255e-01 -3.25899571e-01 3.33714277e-01
-1.00231278e+00 -6.69019938e-01 7.98711181e-01 2.53509074e-01
4.12352294e-01 -4.26828563e-01 -2.18484208e-01 -3.01952749e-01
4.68604490e-02 -9.69485104e-01 3.47323179e-01 -5.14531195e-01
-1.45737016e+00 5.18917620e-01 -1.95711464e-01 -1.52664661e+00
-5.15887737e-01 -7.48067260e-01 -7.90119052e-01 9.25100923e-01
-9.40296233e-01 -1.02697814e+00 8.29423070e-02 8.68303776e-01
4.37653929e-01 -4.58049536e-01 9.15045023e-01 6.76372722e-02
-4.92097437e-01 7.27337301e-01 3.37348253e-01 7.16270357e-02
5.99919081e-01 -1.02135301e+00 3.81530911e-01 1.02502871e+00
-3.58972341e-01 5.91066420e-01 1.23186350e+00 -8.31613004e-01
-1.71583378e+00 -1.33547354e+00 1.65865663e-02 -5.68740845e-01
1.16571748e+00 -5.22363819e-02 -9.38965440e-01 7.48932838e-01
-9.86335650e-02 1.50715917e-01 6.61736906e-01 -3.46254706e-01
-6.06065728e-02 -2.00578630e-01 -1.52428579e+00 5.24386585e-01
3.14342052e-01 -7.76433110e-01 -3.65239143e-01 5.26220798e-01
1.00164139e+00 -4.88575965e-01 -1.61125040e+00 5.91101468e-01
3.94801497e-01 -6.69041038e-01 1.19856656e+00 -8.72633278e-01
5.32681823e-01 -2.86665380e-01 -2.44033933e-01 -1.26848292e+00
-1.23179061e-02 -8.23887825e-01 -4.07675594e-01 1.00235343e+00
2.41169572e-01 -1.01327252e+00 2.93141395e-01 5.64571261e-01
-2.41145045e-01 -8.15665126e-01 -9.91331995e-01 -1.34035504e+00
3.83356810e-01 -7.01381266e-01 9.33149040e-01 8.60707760e-01
-1.92408875e-01 -3.55193824e-01 -4.03197527e-01 8.46843183e-01
2.49355197e-01 -1.17423095e-01 6.01357698e-01 -9.40634370e-01
-5.44188440e-01 3.56132127e-02 -8.01317275e-01 -2.63095032e-02
-1.20500468e-01 -2.90964544e-01 1.02829151e-01 -8.12229574e-01
-8.61960173e-01 -4.52456981e-01 -8.67544889e-01 3.24208349e-01
3.84397781e-03 4.84434292e-02 2.70764798e-01 -2.90437900e-02
3.11994761e-01 3.48420084e-01 1.18066978e+00 -2.02987120e-01
-1.09493047e-01 4.70555097e-01 -1.07024528e-01 4.63304013e-01
1.16042137e+00 -5.93939424e-01 -5.83743334e-01 4.59256805e-02
9.22104120e-02 2.22340420e-01 4.87883925e-01 -1.32531023e+00
5.76161547e-03 -1.01314552e-01 4.81392592e-01 -3.88891041e-01
2.33331859e-01 -1.33063793e+00 4.72636193e-01 8.11933875e-01
2.03556195e-02 3.10338378e-01 6.92837477e-01 7.57645726e-01
-6.05418608e-02 -2.18482628e-01 6.85284257e-01 7.83012807e-02
-2.22410947e-01 2.25201592e-01 -3.77038926e-01 -5.73443949e-01
1.37606633e+00 -2.22260654e-01 -1.35728210e-01 -1.32054314e-01
-7.96153307e-01 -1.25122845e-01 5.06949484e-01 5.20285189e-01
4.63372082e-01 -1.18842113e+00 -3.83570164e-01 4.06777382e-01
-2.89018989e-01 -2.23817050e-01 3.05850923e-01 5.86040735e-01
-5.01345098e-01 2.48352170e-01 -4.10395205e-01 -1.59889475e-01
-1.18201768e+00 1.37137973e+00 5.70191801e-01 -3.45308512e-01
-5.40644765e-01 3.65300328e-01 -3.21105421e-01 -3.02007049e-01
-3.29626650e-02 -4.16785061e-01 2.64082670e-01 -3.41229767e-01
3.21690410e-01 6.21031225e-01 5.03505707e-01 -8.28031376e-02
-1.33275256e-01 4.04357672e-01 -4.37068641e-02 3.63569438e-01
1.22503781e+00 3.92131150e-01 8.87238309e-02 6.60995662e-01
1.07616925e+00 -8.62629861e-02 -1.23725545e+00 3.27977747e-01
-3.02789122e-01 -3.10025156e-01 1.36067653e-02 -8.06982338e-01
-1.27591157e+00 7.05368280e-01 7.80885279e-01 7.24472582e-01
1.28224027e+00 -5.68862736e-01 9.24090505e-01 5.34940600e-01
5.94116986e-01 -8.67555082e-01 2.35823262e-02 4.38385814e-01
1.03775775e+00 -7.49912977e-01 -1.84973460e-02 -2.80649841e-01
-5.11760294e-01 1.24132168e+00 3.31737578e-01 -4.68234032e-01
7.60042548e-01 6.69915557e-01 1.66663036e-01 -4.49389592e-02
-4.73084182e-01 5.76639950e-01 5.90312816e-02 7.42809653e-01
-6.80762604e-02 1.11138180e-01 1.84212208e-01 5.38763642e-01
-1.44998327e-01 -1.20722063e-01 5.81360877e-01 1.34065783e+00
5.60157821e-02 -1.09207177e+00 -8.02495062e-01 5.06085157e-01
-6.47164941e-01 2.28700161e-01 -4.75799441e-02 8.87922108e-01
-4.51850183e-02 1.09260714e+00 -1.24225788e-01 -6.00519359e-01
6.92180932e-01 1.96566954e-01 4.05711770e-01 -1.62970632e-01
-9.37162697e-01 -4.19151217e-01 6.26100823e-02 -5.88453889e-01
3.14054132e-01 -4.81825501e-01 -1.31264472e+00 -4.74690318e-01
-2.36314893e-01 -1.33554876e-01 7.16711938e-01 8.99068654e-01
2.64728040e-01 1.34515953e+00 1.04119980e+00 -8.07223558e-01
-1.03606200e+00 -8.54476035e-01 -8.58284116e-01 3.36579621e-01
4.42857474e-01 -7.03588903e-01 -7.21283376e-01 5.15407771e-02] | [5.444273948669434, 7.475541591644287] |
7662ad04-5cbf-475e-a1ca-e87386c7c6ee | image-based-automatic-dial-meter-reading-in | 2201.02850 | null | https://arxiv.org/abs/2201.02850v2 | https://arxiv.org/pdf/2201.02850v2.pdf | Image-based Automatic Dial Meter Reading in Unconstrained Scenarios | The replacement of analog meters with smart meters is costly, laborious, and far from complete in developing countries. The Energy Company of Parana (Copel) (Brazil) performs more than 4 million meter readings (almost entirely of non-smart devices) per month, and we estimate that 850 thousand of them are from dial meters. Therefore, an image-based automatic reading system can reduce human errors, create a proof of reading, and enable the customers to perform the reading themselves through a mobile application. We propose novel approaches for Automatic Dial Meter Reading (ADMR) and introduce a new dataset for ADMR in unconstrained scenarios, called UFPR-ADMR-v2. Our best-performing method combines YOLOv4 with a novel regression approach (AngReg), and explores several postprocessing techniques. Compared to previous works, it decreased the Mean Absolute Error (MAE) from 1,343 to 129 and achieved a meter recognition rate (MRR) of 98.90% -- with an error tolerance of 1 Kilowatt-hour (kWh). | ['David Menotti', 'Rayson Laroca', 'Gabriel Salomon'] | 2022-01-08 | null | null | null | null | ['dial-meter-reading', 'image-based-automatic-meter-reading'] | ['computer-vision', 'computer-vision'] | [-4.29954045e-02 -2.06836179e-01 2.47884676e-01 -5.20704627e-01
-1.12169409e+00 -5.95116138e-01 7.13592649e-01 -6.11206926e-02
-1.77922294e-01 9.15387213e-01 8.47721398e-02 -6.16668165e-01
1.21929586e-01 -1.22268581e+00 -2.08915994e-01 -7.30618000e-01
4.64845270e-01 5.95710516e-01 -2.38335878e-01 -5.24763949e-02
4.39161539e-01 5.72852731e-01 -8.83489609e-01 -2.64282405e-01
1.47272980e+00 1.14731550e+00 2.81841099e-01 7.02307701e-01
1.14844561e-01 1.16169298e+00 -1.07590556e+00 -3.38974148e-01
4.16904539e-01 -6.04971528e-01 -5.52403569e-01 -1.73733644e-02
2.67922312e-01 -8.50670576e-01 -1.30373299e-01 1.00054336e+00
7.54668057e-01 -3.33815105e-02 8.57550561e-01 -1.06768751e+00
-1.09414411e+00 7.13455260e-01 -6.22228265e-01 1.68396756e-01
5.55570960e-01 3.19286734e-02 8.29419255e-01 -8.56587589e-01
-2.32948169e-01 5.17316580e-01 8.01633537e-01 -1.90738842e-01
-1.06555498e+00 -4.57955867e-01 -4.81584996e-01 5.04835188e-01
-1.84370708e+00 -4.48777854e-01 5.87136030e-01 -1.87383220e-01
1.29898477e+00 6.33873105e-01 8.01349819e-01 3.89099300e-01
3.81550118e-02 6.16289675e-01 1.38908160e+00 -5.54038882e-01
5.00386894e-01 1.84160352e-01 -1.54339939e-01 1.13711908e-01
3.23296249e-01 -4.40592527e-01 1.61244068e-03 2.37881735e-01
5.87411702e-01 -1.49701223e-01 -2.40140378e-01 2.93767005e-01
-9.23642159e-01 8.09697092e-01 2.20724776e-01 6.36639535e-01
-5.25144517e-01 -1.74796462e-01 -1.70555800e-01 1.62950471e-01
2.40867376e-01 3.77544016e-01 -3.18200231e-01 -4.85134393e-01
-1.48266494e+00 -4.77702230e-01 9.37199056e-01 8.59855890e-01
6.97515070e-01 6.08145952e-01 2.26147667e-01 7.46769845e-01
2.65233368e-01 1.16109872e+00 7.57921398e-01 -7.82173514e-01
5.89435756e-01 6.92728817e-01 4.87994671e-01 -8.06204736e-01
-5.78351259e-01 8.66653323e-02 -1.21415722e+00 9.00219977e-02
3.66631866e-01 -8.36548284e-02 -5.21143913e-01 6.78865731e-01
-6.83044121e-02 -2.13384286e-01 -1.98937893e-01 5.66278994e-01
4.75375950e-01 8.01266074e-01 -2.98688203e-01 -2.64434040e-01
1.08034968e+00 -8.94006848e-01 -6.79988623e-01 -2.33251695e-02
3.18178415e-01 -7.01748490e-01 1.08995605e+00 6.97663486e-01
-1.14800620e+00 -4.16531682e-01 -1.41195953e+00 7.21228346e-02
-6.15163147e-01 3.37675691e-01 7.87253864e-03 1.09601641e+00
-1.07037401e+00 6.06391847e-01 -5.23876667e-01 -2.89498121e-01
1.87877357e-01 1.70551836e-01 1.35070309e-01 2.27986440e-01
-7.84638643e-01 1.52296245e+00 -2.18879834e-01 1.00968711e-01
-3.61781776e-01 -6.95540488e-01 -6.51244819e-01 7.22201690e-02
-2.27321744e-01 -2.05205366e-01 1.15007484e+00 -5.02070427e-01
-1.77312648e+00 5.27434766e-01 -5.10002710e-02 -8.12731683e-01
8.94995570e-01 -3.64348888e-01 -8.66556406e-01 1.66932195e-01
1.06725402e-01 2.35278174e-01 6.06713295e-01 -9.24099147e-01
-5.23141146e-01 -1.91881478e-01 -4.03875262e-01 -7.54247233e-02
-3.03659942e-02 -3.70519012e-01 2.95431614e-01 -4.51995194e-01
1.04236431e-01 -4.66694802e-01 2.19687889e-03 -6.46017432e-01
-3.84634703e-01 -2.46807989e-02 6.29215598e-01 -1.49752235e+00
9.68316019e-01 -1.71510708e+00 -6.73255444e-01 4.81075853e-01
-1.48099884e-01 4.44760829e-01 3.80875170e-01 3.51821542e-01
-8.36049989e-02 -3.45145278e-02 -3.96030456e-01 -1.38516396e-01
2.43051365e-01 7.58498013e-02 -2.37496883e-01 6.78548813e-01
-3.54766965e-01 1.21877205e+00 -7.44403720e-01 -5.40304035e-02
1.28333402e+00 6.56946480e-01 4.36667949e-01 -6.99746981e-02
3.08449060e-01 3.88566732e-01 -1.22720554e-01 9.71371770e-01
1.10975039e+00 -2.73185670e-01 1.44137532e-01 -3.53391737e-01
-2.67303139e-01 4.40763533e-01 -1.39000297e+00 1.04972458e+00
-6.99737966e-01 7.93626606e-01 -5.07558286e-01 -9.28969204e-01
1.43581831e+00 9.12770480e-02 8.21547270e-01 -1.40433407e+00
1.07694060e-01 2.51711726e-01 -7.86366403e-01 -3.24192852e-01
7.75950909e-01 5.50783575e-02 9.96134505e-02 2.60117710e-01
-6.00313485e-01 -4.33143646e-01 1.17439687e-01 -7.11638555e-02
9.70722854e-01 -9.16342735e-02 8.34213912e-01 -3.46747875e-01
6.25565052e-01 -2.29165554e-01 1.83180407e-01 6.72542095e-01
-1.55409008e-01 6.54882133e-01 -2.65210658e-01 -3.16109359e-01
-1.19818854e+00 -1.14670968e+00 -4.67703283e-01 3.19909096e-01
2.17336401e-01 -5.05804829e-02 -8.05863500e-01 -7.06034154e-02
1.79607421e-01 1.53900051e+00 1.21017911e-01 3.38248193e-01
-5.20810306e-01 -9.62210476e-01 6.53750181e-01 8.00565720e-01
1.26582527e+00 -7.11939514e-01 -5.96542120e-01 4.61278409e-01
-3.45812261e-01 -1.33928764e+00 -2.35119626e-01 1.15162380e-01
-6.28869116e-01 -1.01174390e+00 -7.82911181e-01 -3.89045298e-01
4.75630492e-01 2.19466716e-01 1.47660983e+00 -1.05654523e-01
-2.34790012e-01 3.36907715e-01 -4.46352750e-01 1.67263695e-03
-3.62855047e-01 -2.76092112e-01 8.67375880e-02 -1.00716785e-01
7.34839022e-01 -7.88400292e-01 -6.88958824e-01 5.73653877e-01
-2.96718508e-01 -3.01234037e-01 4.97275889e-01 2.53846616e-01
5.37094176e-01 4.82110642e-02 8.46445918e-01 -4.80431437e-01
4.03234661e-01 -1.55757293e-01 -1.13684845e+00 2.06440926e-01
-1.48335719e+00 -3.72271270e-01 9.40714717e-01 -1.37524590e-01
-8.55676115e-01 1.29608303e-01 -3.63456398e-01 3.81782264e-01
-1.52771905e-01 -2.59736359e-01 -5.06809175e-01 -1.52464077e-01
4.03531492e-01 5.34078479e-01 -7.16671526e-01 -7.66468227e-01
4.55313891e-01 1.12214398e+00 9.54945743e-01 4.80111353e-02
1.03340912e+00 2.83474654e-01 -1.34839475e-01 -1.13776183e+00
-2.93853194e-01 -4.71903771e-01 -7.48565793e-01 -2.38404527e-01
8.97141278e-01 -1.09699726e+00 -9.41182673e-01 7.77180314e-01
-5.55105150e-01 -3.31447899e-01 -6.19751930e-01 5.28428257e-01
-2.83557832e-01 6.88158095e-01 -5.36874413e-01 -1.01620042e+00
-6.58259451e-01 -7.99734533e-01 7.44496822e-01 5.69504619e-01
-3.36088657e-01 -8.66766155e-01 6.00665361e-02 6.64009035e-01
7.84755051e-01 7.80300274e-02 5.22404194e-01 -5.10831594e-01
-6.87339664e-01 -3.34876984e-01 -4.77248162e-01 7.31450796e-01
3.36634070e-01 -3.89056206e-01 -1.17483759e+00 -4.53867428e-02
3.13532770e-01 1.82392299e-01 4.17069137e-01 5.10859549e-01
8.39440107e-01 -2.81813115e-01 1.40414059e-01 4.37063843e-01
1.83409667e+00 4.25508171e-01 1.45155978e+00 6.76469386e-01
5.22445202e-01 -4.70923007e-01 3.36686969e-01 8.62063348e-01
8.39163303e-01 5.57828248e-01 3.79635990e-01 -1.15737863e-01
2.24161431e-01 -2.33995184e-01 4.36458021e-01 1.10101759e+00
-2.85851628e-01 -2.32666120e-01 -3.75666887e-01 5.58100998e-01
-1.41772461e+00 -9.21288192e-01 -6.45421743e-01 2.32004023e+00
5.27653396e-01 -1.70476675e-01 1.72430605e-01 5.74854970e-01
6.26373529e-01 5.61840534e-02 -6.58727348e-01 -2.87027329e-01
-3.21506321e-01 2.31306627e-01 1.13377500e+00 8.06149065e-01
-7.58843243e-01 3.69697213e-01 6.93823910e+00 4.52142507e-01
-6.44223452e-01 9.91994422e-03 5.90658784e-01 3.64746541e-01
-6.54361174e-02 -2.57828653e-01 -6.41150832e-01 8.06106925e-01
1.24721682e+00 -2.80629247e-01 6.33745611e-01 1.03197050e+00
5.19532740e-01 -7.67713487e-01 -9.99471724e-01 1.27489841e+00
2.92208046e-01 -1.07813132e+00 -3.61033320e-01 1.19571067e-01
8.59436154e-01 1.29552856e-01 -3.65849614e-01 -1.74232665e-02
4.81587946e-01 -1.09306073e+00 5.97885132e-01 7.83432484e-01
7.57728338e-01 -6.25055313e-01 9.43752050e-01 3.05591494e-01
-1.60754371e+00 1.49569243e-01 -6.50458932e-01 -9.57995728e-02
4.65818077e-01 1.17057741e+00 -8.76499236e-01 6.80896699e-01
7.43817806e-01 5.77351868e-01 -6.97057009e-01 9.68731165e-01
-5.41409731e-01 5.73331594e-01 -5.66912532e-01 -1.33123308e-01
-2.55360544e-01 -7.95706391e-01 -4.30229045e-02 1.26441646e+00
5.88978171e-01 -6.71042055e-02 -1.42368868e-01 6.56229496e-01
-7.80441090e-02 8.24755710e-03 -2.74019241e-01 4.81779367e-01
6.68224394e-01 1.38917100e+00 -6.78272963e-01 -4.98577774e-01
-3.94739926e-01 1.29704332e+00 -4.70483899e-01 9.49750543e-02
-1.09177899e+00 -5.43968201e-01 1.29140243e-01 -1.69063985e-01
3.75370711e-01 -8.42118487e-02 -8.83943439e-01 -1.26381052e+00
2.90555924e-01 -6.94356382e-01 2.58158632e-02 -1.13087833e+00
-1.36474967e+00 -4.86371107e-02 -3.07139754e-01 -1.26196814e+00
-4.29597855e-01 -4.62343752e-01 -7.10445881e-01 9.05698061e-01
-1.65575230e+00 -5.80667138e-01 -7.51055181e-01 6.60493255e-01
5.75038195e-01 -4.66028415e-02 9.77377892e-01 6.25487387e-01
-1.50169894e-01 6.30572021e-01 7.51268804e-01 3.03670526e-01
2.93430001e-01 -1.71251535e+00 6.85415745e-01 8.63398373e-01
1.48154393e-01 -2.63701022e-01 5.21553814e-01 -3.27484131e-01
-1.35654652e+00 -9.58080292e-01 1.11526632e+00 -4.41253603e-01
6.25322759e-01 9.04001482e-03 -7.18241513e-01 8.61551702e-01
4.38226283e-01 -5.95853269e-01 5.44399381e-01 -6.35735512e-01
-1.36328191e-01 -6.33138299e-01 -1.73145592e+00 2.16284901e-01
3.70349556e-01 -5.42393744e-01 -5.85596919e-01 3.77446502e-01
-3.55103612e-02 -6.38408288e-02 -1.28622222e+00 -2.43963692e-02
4.11190152e-01 -1.27162349e+00 1.07035851e+00 7.18638718e-01
-2.75159806e-01 -5.74973583e-01 -4.78864133e-01 -1.20652092e+00
-1.03831492e-01 -7.35474885e-01 -1.63913026e-01 1.55847025e+00
3.68671238e-01 -1.10016429e+00 4.61509556e-01 7.14364052e-01
-9.89465602e-03 -1.76914960e-01 -1.01772332e+00 -9.21642125e-01
-7.35884346e-03 -6.04878247e-01 1.18218231e+00 8.35638225e-01
6.22449294e-02 3.06250155e-01 -6.81738019e-01 3.11230898e-01
6.04814529e-01 8.46852660e-02 7.70048797e-01 -9.61689651e-01
-2.47513637e-01 -2.60973603e-01 -5.09617031e-01 -1.54547060e+00
-4.91871864e-01 -5.29407978e-01 -2.41197646e-02 -2.07749438e+00
-1.26713768e-01 -1.65982157e-01 -7.52060814e-03 4.00272876e-01
3.32830340e-01 5.75378060e-01 1.90551728e-01 1.57022312e-01
-2.06280231e-01 3.67007554e-01 7.29998171e-01 -5.13829708e-01
2.67054196e-02 1.93099901e-01 -6.06363654e-01 9.56535578e-01
1.12640047e+00 -7.33601302e-02 -8.46424606e-03 -2.66009092e-01
2.26702601e-01 -1.78618729e-01 1.06542230e-01 -1.00193167e+00
9.60121155e-02 1.67320639e-01 1.02248979e+00 -1.23755777e+00
2.25337982e-01 -1.01590872e+00 3.56781453e-01 4.46455300e-01
5.28794169e-01 4.09259826e-01 -4.48836505e-01 -3.76047269e-02
1.24078497e-01 -5.00439584e-01 9.33626056e-01 -9.88062471e-02
-6.29748881e-01 -4.66147900e-01 -6.52773917e-01 -2.28838846e-01
9.60575998e-01 -3.30534071e-01 -5.93859553e-01 -6.94435418e-01
-3.81435901e-01 1.51988894e-01 5.36959648e-01 -1.18816473e-01
4.28706348e-01 -1.28695512e+00 -6.45228982e-01 1.92997918e-01
-4.09878582e-01 -2.99264580e-01 -9.56475735e-02 8.58142078e-01
-9.50458705e-01 2.47430101e-01 1.47867903e-01 -5.97711563e-01
-8.51095736e-01 1.44616351e-01 5.91027439e-01 -3.76165032e-01
-1.13303185e+00 2.40996286e-01 -6.98854208e-01 -1.37230933e-01
4.40608934e-02 -5.36712587e-01 -9.67585742e-02 1.72680825e-01
8.89583945e-01 1.15188384e+00 5.22518277e-01 -6.21134222e-01
-4.46044713e-01 8.43589485e-01 4.46492434e-01 6.65488988e-02
1.33037043e+00 -4.85733241e-01 1.67638306e-02 4.02708650e-01
9.21011567e-01 2.30351508e-01 -8.71874809e-01 3.46304685e-01
7.50152394e-02 -4.61081803e-01 -2.75382936e-01 -1.13869035e+00
-1.00649500e+00 4.62173223e-01 8.12470973e-01 7.60043204e-01
1.40905261e+00 -2.31263459e-01 1.08921289e+00 5.71207106e-01
4.97293979e-01 -1.51958561e+00 -3.77235442e-01 2.70023227e-01
4.37761128e-01 -1.08068419e+00 4.87092108e-01 5.45223393e-02
-5.42233944e-01 1.03705812e+00 -3.75562124e-02 -1.89820141e-01
5.54685056e-01 4.97952849e-01 2.48269901e-01 3.11320275e-01
1.89490333e-01 -1.97118595e-02 -1.09039359e-01 9.26769614e-01
1.55226722e-01 5.22460938e-01 -1.60958990e-02 4.10834104e-01
-6.63480997e-01 -5.35036139e-02 8.35249662e-01 7.23446906e-01
-6.10804498e-01 -8.02298188e-01 -5.42262137e-01 4.83383566e-01
-3.05945337e-01 -2.30169132e-01 -1.01423673e-01 7.94952810e-01
-9.96977910e-02 1.52221155e+00 1.64916605e-01 -3.45365644e-01
5.00169575e-01 1.13909326e-01 4.48546708e-01 2.51423240e-01
-4.08314526e-01 -6.75086901e-02 1.62493940e-02 -2.34611794e-01
-5.52638531e-01 -7.46512234e-01 -1.35235107e+00 -1.00939357e+00
-4.88956690e-01 -3.27122919e-02 9.07754481e-01 1.26339400e+00
-1.64556116e-01 3.39579850e-01 1.29828930e+00 -8.05317104e-01
-6.75825298e-01 -1.07930493e+00 -1.08407569e+00 3.77225131e-02
5.75789362e-02 -2.43503228e-02 -5.68357646e-01 1.95227504e-01] | [11.385509490966797, 2.668581247329712] |
5e5c6b81-b8cf-483a-bb7f-793584c89f77 | compact-representation-for-image | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Zhang_Compact_Representation_for_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Zhang_Compact_Representation_for_2014_CVPR_paper.pdf | Compact Representation for Image Classification: To Choose or to Compress? | In large scale image classification, features such as Fisher vector or VLAD have achieved state-of-the-art results. However, the combination of large number of examples and high dimensional vectors necessitates dimensionality reduction, in order to reduce its storage and CPU costs to a reasonable range. In spite of the popularity of various feature compression methods, this paper argues that feature selection is a better choice than feature compression. We show that strong multicollinearity among feature dimensions may not exist, which undermines feature compression's effectiveness and renders feature selection a natural choice. We also show that many dimensions are noise and throwing them away is helpful for classification. We propose a supervised mutual information (MI) based importance sorting algorithm to choose features. Combining with 1-bit quantization, MI feature selection has achieved both higher accuracy and less computational cost than feature compression methods such as product quantization and BPBC. | ['Jianfei Cai', 'Jianxin Wu', 'Yu Zhang'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['feature-compression'] | ['computer-vision'] | [ 2.80345529e-01 -6.13858342e-01 -3.59800369e-01 -5.69449902e-01
-8.09694052e-01 -2.83225417e-01 4.28378314e-01 2.01088563e-01
-4.54155415e-01 7.10443377e-01 1.92634165e-01 -2.23411560e-01
-8.02339315e-01 -9.40814018e-01 3.31687368e-02 -8.55965257e-01
-3.24594468e-01 1.96407095e-01 5.79072610e-02 -8.03638101e-02
7.04581797e-01 6.48256361e-01 -2.24363208e+00 1.08012967e-02
7.43150234e-01 1.30436015e+00 2.42835939e-01 3.89827728e-01
-3.10674280e-01 3.00495684e-01 -3.77864301e-01 -3.79960656e-01
4.99065429e-01 -4.48344141e-01 -8.84012222e-01 6.73466027e-02
-2.97717787e-02 -3.46307516e-01 -4.03558373e-01 1.17855763e+00
5.02581835e-01 1.23328656e-01 8.56651902e-01 -1.46090329e+00
-4.99021083e-01 4.30976868e-01 -5.79285502e-01 1.40797481e-01
2.93419689e-01 -2.06657499e-01 1.27604198e+00 -7.49313354e-01
4.01283145e-01 1.14419484e+00 5.23946047e-01 8.31922665e-02
-1.01811194e+00 -4.90952849e-01 -4.48103935e-01 4.92684215e-01
-1.85174012e+00 -2.31436118e-01 8.41570675e-01 -1.40818983e-01
1.00732493e+00 7.75827765e-01 9.13703442e-01 3.59613568e-01
2.70029992e-01 5.58940828e-01 9.64639008e-01 -6.15934849e-01
2.89881647e-01 8.20047595e-03 1.97919831e-01 6.61602974e-01
5.54476619e-01 4.12604474e-02 -5.47076643e-01 -6.18238270e-01
5.68233192e-01 3.40368658e-01 -2.07737789e-01 -3.53271663e-01
-1.03357136e+00 1.33043396e+00 1.69966117e-01 3.60555202e-01
-4.26578641e-01 5.62982857e-02 4.33555096e-01 6.98424637e-01
2.02885196e-01 3.40162456e-01 -3.20736319e-01 -3.41095358e-01
-9.02595580e-01 2.68109709e-01 6.57031775e-01 6.76486313e-01
9.16140437e-01 -2.16767609e-01 4.60653529e-02 8.58806252e-01
2.20870316e-01 5.00062943e-01 8.03336918e-01 -9.71113205e-01
5.34794256e-02 7.86921322e-01 -2.19929412e-01 -1.56561911e+00
-3.34800720e-01 -2.51137078e-01 -1.02059782e+00 1.47258967e-01
4.44559567e-02 3.65714401e-01 -5.75416803e-01 1.41490591e+00
1.75774693e-01 -3.99593800e-01 3.83738317e-02 9.46445346e-01
3.09209645e-01 3.92027557e-01 -2.77827889e-01 -6.53147995e-01
1.32831144e+00 -3.48193586e-01 -7.30002165e-01 3.39304984e-01
7.85266340e-01 -9.17695940e-01 9.36059117e-01 5.07715285e-01
-7.58944988e-01 -3.45681667e-01 -1.14422941e+00 1.70302898e-01
-3.43173057e-01 -2.27144465e-01 1.37903535e+00 9.28905249e-01
-8.10765803e-01 8.68766963e-01 -6.22535646e-01 -2.31106013e-01
6.23766065e-01 5.83952785e-01 -5.21462977e-01 -3.35717440e-01
-1.11244071e+00 5.95611572e-01 1.80113137e-01 -3.95111769e-01
-2.12753881e-02 -3.01625937e-01 -7.90276170e-01 9.08015296e-02
1.05875237e-02 -4.86934364e-01 5.28314054e-01 -5.41982532e-01
-1.00198233e+00 3.47825319e-01 -1.76688626e-01 -4.37097222e-01
2.03680485e-01 1.99594232e-03 -5.44604421e-01 4.79671061e-01
4.32362854e-02 6.74430192e-01 9.66671467e-01 -5.92081964e-01
-8.76116097e-01 -5.15460491e-01 -2.32231930e-01 3.40311080e-01
-8.89576733e-01 3.12862359e-02 -6.77090138e-02 -5.10971725e-01
6.26060963e-01 -7.75932908e-01 -3.72316033e-01 4.42781746e-02
-2.37162963e-01 -3.79981041e-01 7.78903186e-01 -8.08383152e-02
1.68508506e+00 -2.23250222e+00 -2.42825225e-01 5.64805865e-01
2.61931717e-01 -1.76755756e-01 -7.98437074e-02 4.71829802e-01
2.67224330e-02 3.49043012e-01 -1.15229793e-01 6.86429664e-02
-6.27189800e-02 3.43913347e-01 -6.30982639e-03 7.78003395e-01
2.12922528e-01 4.84105468e-01 -5.97260714e-01 -8.41294348e-01
2.09269747e-01 6.76381826e-01 -7.44221926e-01 -2.30511695e-01
4.94249821e-01 -2.41030604e-01 -6.83687150e-01 8.14180374e-01
8.90484989e-01 -4.59884614e-01 -1.61621600e-01 -3.03356290e-01
-2.17222478e-02 1.63627699e-01 -1.42920470e+00 1.26126337e+00
2.86297426e-02 4.10951167e-01 -4.76471394e-01 -1.01140285e+00
9.24015760e-01 4.26438786e-02 8.89351308e-01 -6.87078953e-01
1.75221846e-01 2.18399808e-01 1.57805473e-01 -3.32573354e-01
5.94962895e-01 7.67717836e-03 -6.57022372e-02 4.54516858e-01
-1.03751212e-01 -2.86055505e-01 4.07292634e-01 1.53972998e-01
1.09444618e+00 -4.96943474e-01 6.48570180e-01 -3.73789638e-01
4.94244933e-01 8.47824514e-02 4.92023468e-01 5.81440508e-01
-1.18582048e-01 8.00430715e-01 3.01498860e-01 -4.89658356e-01
-9.24966335e-01 -5.70588768e-01 -6.63762212e-01 7.78557897e-01
1.73074156e-01 -7.94737995e-01 -3.94262046e-01 -5.47849536e-01
2.84561217e-01 2.66295075e-01 -4.71003801e-01 -2.39527434e-01
-1.77063152e-01 -1.05352962e+00 2.08096221e-01 6.92167357e-02
3.27039808e-01 -5.92947602e-01 -9.15778518e-01 2.12854609e-01
-3.55264843e-02 -4.66248393e-01 -2.12035790e-01 5.46383142e-01
-8.53273571e-01 -1.16356230e+00 -3.87629658e-01 -4.37346816e-01
7.00155914e-01 7.08520770e-01 8.35638285e-01 2.03674510e-01
-7.21103549e-01 1.06332004e-01 -8.17599952e-01 -9.97663662e-02
-6.54396042e-02 -1.62316021e-02 3.53091300e-01 -1.65252700e-01
7.52576053e-01 -6.19894028e-01 -8.16030324e-01 1.50336176e-01
-1.02401972e+00 -4.49706852e-01 7.01026320e-01 1.16145289e+00
6.66319191e-01 7.80935943e-01 4.66995478e-01 -4.90141571e-01
8.18600476e-01 -1.92063317e-01 -4.21285957e-01 -1.33424364e-02
-1.21421480e+00 1.51423305e-01 4.98300433e-01 -3.27881455e-01
-4.36223716e-01 1.15982525e-01 -9.93897021e-02 -1.78074285e-01
1.13226250e-01 4.75222737e-01 1.75559502e-02 -4.67087895e-01
6.26024246e-01 3.94735456e-01 2.93916136e-01 -3.75122994e-01
7.94596821e-02 9.01486397e-01 -1.45320550e-01 1.36522427e-01
6.19456410e-01 3.50761324e-01 3.16960692e-01 -8.65678728e-01
-3.66068631e-01 -4.55844045e-01 -5.30973971e-01 3.37174118e-01
4.60543543e-01 -6.05598271e-01 -6.51432455e-01 7.17759505e-02
-6.11976981e-01 6.83310270e-01 -3.95882905e-01 7.71286428e-01
-6.02020919e-01 4.43219781e-01 -4.65255409e-01 -7.62106001e-01
-2.95094997e-01 -1.30041659e+00 7.72973120e-01 4.04076092e-02
-2.50764787e-01 -4.31085259e-01 -3.32598001e-01 1.20428778e-01
5.84647477e-01 -9.23815891e-02 1.03567064e+00 -7.08509684e-01
-3.45203817e-01 -5.20879745e-01 -1.97815418e-01 1.51733413e-01
4.99769479e-01 -1.38757169e-01 -7.35763609e-01 -3.02499413e-01
8.50723311e-02 -2.29204491e-01 8.23191524e-01 3.17858219e-01
1.26978993e+00 -3.51402253e-01 -3.95169348e-01 6.00267172e-01
1.52007115e+00 2.76108593e-01 5.96361816e-01 3.46170247e-01
2.71118402e-01 4.18024629e-01 8.93235624e-01 1.11839056e+00
2.17003584e-01 3.10314685e-01 2.63956875e-01 2.24945992e-01
1.16167434e-01 1.31087646e-01 -1.69503868e-01 1.00388563e+00
2.21915185e-01 -5.18061295e-02 -5.10190964e-01 3.82547140e-01
-1.43396783e+00 -9.01273549e-01 -1.18814319e-01 2.28684783e+00
8.53705466e-01 -3.23565826e-02 6.90326616e-02 8.55571091e-01
4.76235062e-01 4.27803881e-02 -3.22006822e-01 -3.12433124e-01
-3.50603700e-01 -7.03586638e-02 5.60275793e-01 1.98766053e-01
-1.14821184e+00 5.85892200e-01 6.98365879e+00 1.21864605e+00
-1.08926308e+00 -1.76750094e-01 7.92800725e-01 -4.83848378e-02
-3.88975531e-01 -1.28806233e-01 -7.71492660e-01 5.33971488e-01
7.79341280e-01 -4.11712319e-01 2.89628327e-01 9.07720804e-01
-3.02816153e-01 -1.85591117e-01 -6.69444144e-01 1.49503803e+00
4.14872393e-02 -1.18024921e+00 2.91141003e-01 4.28062141e-01
4.16208804e-01 -1.82204232e-01 1.77388996e-01 -1.10814586e-01
-7.33402371e-02 -9.99936938e-01 1.97031707e-01 1.30089223e-01
8.02900016e-01 -1.19890010e+00 9.12060499e-01 6.55300319e-02
-1.17514586e+00 -3.44282806e-01 -8.66362691e-01 -9.48475525e-02
5.14763519e-02 1.01329207e+00 -4.29631531e-01 4.18698400e-01
7.07752287e-01 4.08901423e-01 -7.03078151e-01 1.16947508e+00
4.20985729e-01 2.73569345e-01 -6.11506224e-01 -1.86910406e-01
1.61148682e-01 -2.96024919e-01 3.25918496e-01 9.00816858e-01
7.01718569e-01 1.63049236e-01 1.27424225e-01 1.51953116e-01
3.04628700e-01 5.29566646e-01 -8.21832001e-01 -3.56676280e-01
7.76896954e-01 1.13595819e+00 -8.96982551e-01 -2.81559080e-01
-6.25967443e-01 1.11704981e+00 -4.56218142e-03 -2.66088456e-01
-3.37282985e-01 -7.85278320e-01 9.29237604e-01 -1.21554717e-01
4.44470674e-01 -1.37740970e-01 -2.48007894e-01 -9.72016394e-01
-8.20157230e-02 -7.13583827e-01 4.77044731e-01 -3.30308340e-02
-1.29436219e+00 6.53617084e-01 -1.27130345e-01 -1.75862586e+00
-5.20846903e-01 -3.13020527e-01 4.14322913e-02 5.69474339e-01
-1.43451762e+00 -6.76226735e-01 2.05970909e-02 7.51305997e-01
2.63318688e-01 -3.36066157e-01 1.04127491e+00 3.50473166e-01
-5.53522222e-02 7.63278604e-01 3.13162744e-01 -3.61296475e-01
5.07976472e-01 -9.27863181e-01 -7.85628185e-02 3.70304406e-01
1.10851847e-01 5.25137365e-01 6.46917343e-01 -5.49252033e-01
-1.98666000e+00 -6.99434876e-01 1.06034780e+00 7.36946538e-02
5.11475146e-01 7.17059523e-02 -7.43888915e-01 -2.39643365e-01
-2.32250333e-01 1.71749771e-01 1.05295384e+00 -1.88153282e-01
-2.18497619e-01 -2.05941468e-01 -1.48348367e+00 5.66396952e-01
1.19801402e+00 -3.52452874e-01 -3.39772493e-01 3.36220682e-01
5.88777065e-01 3.95910114e-01 -1.06466293e+00 4.43124443e-01
6.72694802e-01 -1.04130542e+00 1.03637087e+00 -2.41991520e-01
1.49020776e-01 -1.58234790e-01 -7.39226818e-01 -1.02770674e+00
-6.51468158e-01 -4.29240286e-01 1.10066362e-01 1.19383705e+00
2.14003474e-01 -6.27634227e-01 7.24741280e-01 5.98890543e-01
4.58449781e-01 -8.65220606e-01 -1.18283367e+00 -8.93689990e-01
-2.59600073e-01 -4.35365409e-01 7.34244525e-01 9.85225022e-01
2.56082773e-01 1.91403270e-01 -3.17803949e-01 -3.14338565e-01
8.09231102e-01 5.00972807e-01 3.02600265e-01 -1.60911798e+00
1.03039801e-01 -4.88123327e-01 -9.87265944e-01 -7.10942209e-01
-6.69025555e-02 -8.37604403e-01 -3.71176630e-01 -1.14275479e+00
3.38556826e-01 -6.95426047e-01 -4.37585801e-01 4.15616304e-01
4.78840023e-02 6.35577202e-01 -5.91049083e-02 3.06557506e-01
-4.71030951e-01 5.89985192e-01 1.10254991e+00 -3.74546088e-02
-2.29995683e-01 -2.49770403e-01 -9.75484252e-01 5.05672693e-01
6.66131973e-01 -5.09984434e-01 -3.74932081e-01 1.03240587e-01
2.16709599e-01 9.08166841e-02 -3.03842545e-01 -9.64809775e-01
1.66362688e-01 -1.93581417e-01 7.91500747e-01 -6.75744295e-01
3.83323848e-01 -9.88164783e-01 1.03143506e-01 5.95612466e-01
-2.26367071e-01 1.90966338e-01 -4.15919214e-01 4.34827119e-01
-5.82049489e-01 -4.65886623e-01 5.99121749e-01 -1.10532425e-01
-7.21635580e-01 3.30274403e-01 -3.40416729e-01 -4.10422444e-01
7.47994065e-01 -5.59144437e-01 5.30177951e-02 -2.67700791e-01
-3.16265792e-01 -4.02468396e-03 4.81959581e-01 2.30472729e-01
9.56429362e-01 -1.58435225e+00 -5.78019261e-01 7.13679612e-01
1.97991133e-01 -4.94879007e-01 1.87047482e-01 7.11651325e-01
-2.89727449e-01 4.84222710e-01 -2.43400410e-01 -4.68590498e-01
-1.51729035e+00 5.84087133e-01 -4.71397638e-01 4.64245602e-02
-7.91521490e-01 7.97796309e-01 -1.78922758e-01 6.13588877e-02
8.06582198e-02 1.23964407e-01 -3.87495935e-01 4.81681794e-01
9.71295178e-01 5.00090063e-01 9.57307816e-02 -6.67597413e-01
-5.49750805e-01 7.30797589e-01 -1.26809537e-01 -1.10875089e-02
1.43909633e+00 -4.95998085e-01 -2.41493776e-01 6.80720806e-02
1.76549685e+00 -1.80504143e-01 -7.93593764e-01 -6.67785481e-02
-2.89488621e-02 -1.12115645e+00 3.93578500e-01 -2.31930241e-01
-9.92460310e-01 5.88253736e-01 9.29666877e-01 4.43891793e-01
1.38722205e+00 -5.20801023e-02 7.03235090e-01 6.28981054e-01
7.99625158e-01 -1.13094854e+00 -2.43023187e-01 4.71172869e-01
6.26808345e-01 -1.42237484e+00 3.25762331e-01 -5.59308171e-01
-6.01842880e-01 1.17295146e+00 6.64789155e-02 -1.84811726e-01
9.45949197e-01 4.19352174e-01 -3.12114537e-01 1.65072177e-02
-7.86868155e-01 -4.18763876e-01 1.90312147e-01 4.78911012e-01
3.01595509e-01 8.57429504e-02 -9.17656839e-01 5.75576782e-01
-5.97450614e-01 -1.39714861e-02 1.92844808e-01 1.10155630e+00
-7.15157151e-01 -1.23874533e+00 -4.07836676e-01 1.13977861e+00
-6.71674728e-01 -2.43024305e-01 3.82799692e-02 5.62686682e-01
-2.15708613e-02 1.02582741e+00 1.61526337e-01 -6.85201228e-01
-1.44888744e-01 8.97056684e-02 4.72102374e-01 -3.05884015e-02
-3.54042053e-01 2.41650879e-01 -8.76801834e-02 -6.75691247e-01
-3.41023386e-01 -6.36967242e-01 -1.15760267e+00 -4.67266053e-01
-7.17898130e-01 3.83755177e-01 8.70308101e-01 6.51932538e-01
4.85038161e-01 1.77624356e-02 1.00330818e+00 -6.33762300e-01
-9.95610476e-01 -7.08046556e-01 -1.05406570e+00 4.47872370e-01
3.42153102e-01 -8.57454062e-01 -6.32739961e-01 -2.03052059e-01] | [8.384576797485352, 3.9275076389312744] |
0926c326-b71c-4747-9533-bdab0aecf1c8 | adversarial-training-for-satire-detection | 1902.11145 | null | http://arxiv.org/abs/1902.11145v2 | http://arxiv.org/pdf/1902.11145v2.pdf | Adversarial Training for Satire Detection: Controlling for Confounding Variables | The automatic detection of satire vs. regular news is relevant for downstream
applications (for instance, knowledge base population) and to improve the
understanding of linguistic characteristics of satire. Recent approaches build
upon corpora which have been labeled automatically based on article sources. We
hypothesize that this encourages the models to learn characteristics for
different publication sources (e.g., "The Onion" vs. "The Guardian") rather
than characteristics of satire, leading to poor generalization performance to
unseen publication sources. We therefore propose a novel model for satire
detection with an adversarial component to control for the confounding variable
of publication source. On a large novel data set collected from German news
(which we make available to the research community), we observe comparable
satire classification performance and, as desired, a considerable drop in
publication classification performance with adversarial training. Our analysis
shows that the adversarial component is crucial for the model to learn to pay
attention to linguistic properties of satire. | ['Roman Klinger', 'Robert McHardy', 'Heike Adel'] | 2019-02-28 | adversarial-training-for-satire-detection-1 | https://aclanthology.org/N19-1069 | https://aclanthology.org/N19-1069.pdf | naacl-2019-6 | ['knowledge-base-population', 'satire-detection'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.04352556e-01 7.96566606e-02 -6.51653707e-01 -2.77940899e-01
-9.33375657e-01 -1.09329379e+00 9.49111462e-01 4.58295405e-01
-4.39136118e-01 4.53966379e-01 7.72602260e-01 -4.87848461e-01
8.99024904e-02 -7.66215026e-01 -8.62683773e-01 -5.58050931e-01
3.47567767e-01 4.30283546e-01 -2.34241530e-01 -5.40801644e-01
4.78513300e-01 3.39395195e-01 -1.00003219e+00 4.61403549e-01
7.18963981e-01 6.55361474e-01 -1.82661235e-01 2.81614542e-01
-9.75399017e-02 9.17288661e-01 -9.05152738e-01 -8.76108944e-01
3.13129127e-01 -3.59507680e-01 -3.87889326e-01 -2.65416682e-01
8.30595016e-01 8.01592693e-03 -2.59057254e-01 1.24836040e+00
4.45872366e-01 -1.96105957e-01 7.86943078e-01 -9.59849179e-01
-1.00728631e+00 1.46461225e+00 -4.28972960e-01 4.86519158e-01
5.06792180e-02 2.09212881e-02 1.34571540e+00 -3.43775898e-01
9.51398194e-01 1.40957701e+00 7.81832933e-01 5.68970919e-01
-1.42140532e+00 -8.24766517e-01 2.25642502e-01 -1.17122442e-01
-1.11765254e+00 -6.32305741e-01 1.01514077e+00 -6.79138958e-01
1.09961614e-01 5.36607206e-01 6.37565926e-02 1.82992518e+00
2.98052996e-01 7.34561324e-01 1.35918176e+00 -2.45621502e-01
2.29996771e-01 6.87270939e-01 1.92751974e-01 5.04010379e-01
2.57073104e-01 3.40152569e-02 -6.60474360e-01 -4.67897832e-01
3.60843778e-01 -3.09332222e-01 -3.34610224e-01 4.77431901e-02
-9.69031811e-01 1.14683390e+00 3.42157006e-01 4.73561764e-01
-9.89083350e-02 -1.17344551e-01 5.72366059e-01 5.71321905e-01
6.98989689e-01 1.06807303e+00 -5.93306124e-01 1.62387237e-01
-1.00678802e+00 4.23272103e-01 1.10877430e+00 8.29342008e-01
-1.80403795e-02 -1.72828138e-01 -4.12727773e-01 7.51659095e-01
6.15490936e-02 5.59015453e-01 6.70796275e-01 -5.94791412e-01
4.63099539e-01 4.97894108e-01 2.60384250e-02 -1.24163342e+00
-2.41064817e-01 -7.92219281e-01 -5.25113821e-01 5.88908978e-02
8.36348534e-01 -9.79615524e-02 -5.68936765e-01 1.87594664e+00
-2.14871258e-01 -5.10194823e-02 3.87059182e-01 7.92943478e-01
1.03173196e+00 3.75468105e-01 3.84965688e-01 2.55058501e-02
1.57354593e+00 -4.83869135e-01 -3.69389325e-01 -2.23191425e-01
8.19629848e-01 -9.47531164e-01 1.21585798e+00 4.54694569e-01
-8.14807296e-01 -6.24514967e-02 -9.90913451e-01 -1.73180982e-01
-6.13722146e-01 6.75806254e-02 6.14627421e-01 6.13271832e-01
-3.77304912e-01 8.16787720e-01 -3.88192832e-01 -2.62643993e-01
6.09855473e-01 -1.46898508e-01 -3.63760106e-02 -9.65326838e-03
-1.36884069e+00 9.86127913e-01 -6.26295514e-04 -4.16787595e-01
-5.87791145e-01 -9.34813440e-01 -6.73980236e-01 2.37278305e-02
3.37001890e-01 -1.44605160e-01 1.06082177e+00 -1.62526011e+00
-1.19989944e+00 1.09678209e+00 2.21456140e-01 -5.97992957e-01
6.11876667e-01 -6.97076023e-02 -5.87133288e-01 -8.92372429e-02
2.21311301e-01 6.92382231e-02 8.96211743e-01 -1.34195137e+00
-2.72785455e-01 -3.69773626e-01 1.43413126e-01 -9.36940089e-02
-4.68994498e-01 2.82767236e-01 1.53800651e-01 -1.04554999e+00
-1.75298482e-01 -9.70459998e-01 2.20827937e-01 -1.79172859e-01
-5.06449223e-01 -2.16927931e-01 4.25907701e-01 -7.77733862e-01
1.02472174e+00 -2.30486655e+00 -7.01703429e-02 2.52795219e-01
1.95703417e-01 -1.61960274e-01 -1.65960178e-01 1.12797365e-01
5.60308881e-02 5.25219977e-01 2.59275138e-01 4.78782542e-02
1.09752201e-01 2.86106393e-03 -6.20008290e-01 6.54412031e-01
8.69732872e-02 7.54888415e-01 -1.03968251e+00 -2.20842078e-01
-4.57830280e-01 2.27724627e-01 -6.72334671e-01 -1.39988154e-01
-3.13260555e-01 2.76173413e-01 -5.87925553e-01 8.10971975e-01
2.23213464e-01 -6.41782433e-02 4.03662659e-02 -2.69175380e-01
-2.60548592e-02 6.28375590e-01 -4.00146127e-01 1.20613945e+00
-5.34408867e-01 6.78642809e-01 2.56279141e-01 -8.09303224e-01
7.93851614e-01 2.61354327e-01 2.24077284e-01 -6.13045394e-01
5.75397313e-01 2.59377152e-01 3.82952303e-01 -2.06618100e-01
4.99188751e-01 -5.17107725e-01 -5.10746121e-01 2.58929849e-01
-1.33483842e-01 1.44078061e-01 2.02007275e-02 5.22109628e-01
9.86216903e-01 -2.38242134e-01 -9.50416625e-02 -7.89602697e-01
2.86668897e-01 3.63730907e-01 8.01941156e-01 9.06845570e-01
-4.50006779e-03 2.84395933e-01 8.08068335e-01 -4.61282730e-02
-8.83622468e-01 -8.35809827e-01 -4.21308041e-01 1.36531687e+00
-9.22776461e-02 -3.06479394e-01 -3.98007751e-01 -9.84030962e-01
3.59139234e-01 1.42983866e+00 -6.99797034e-01 -4.18587238e-01
-3.72021943e-01 -6.42041087e-01 7.50112236e-01 3.40716898e-01
-5.98985218e-02 -9.81198847e-01 -3.32220763e-01 -2.14087740e-02
-9.48137343e-02 -9.83476400e-01 -5.95206141e-01 2.99942046e-01
-5.19431233e-01 -1.02911055e+00 -5.89147508e-01 -5.32225609e-01
4.51376170e-01 -2.51380116e-01 1.40264308e+00 7.98455626e-02
6.81761727e-02 3.71885061e-01 -4.15340543e-01 -7.88483441e-01
-1.04149032e+00 1.46281809e-01 -5.01200296e-02 -1.71726540e-01
5.53031325e-01 -4.77887005e-01 -2.49384835e-01 1.78869307e-01
-6.67860091e-01 -2.71848351e-01 4.87017900e-01 6.81635916e-01
5.51348999e-02 -2.93673985e-02 6.94013357e-01 -1.50790131e+00
6.96168840e-01 -8.79769921e-01 -5.94582379e-01 3.36257555e-02
-6.23868108e-01 2.63389587e-01 8.16726685e-01 -9.61520553e-01
-8.98124695e-01 -3.77866417e-01 1.20000109e-01 -3.82838190e-01
-1.17223430e-02 5.62446654e-01 -1.51961744e-01 1.33593604e-01
1.01102936e+00 -2.35780925e-01 -1.29458174e-01 -7.59258986e-01
4.44923311e-01 7.00018346e-01 4.98310059e-01 -7.72100866e-01
8.43770027e-01 3.34132969e-01 -2.60723233e-01 -5.99003792e-01
-1.30995584e+00 -4.12442148e-01 -2.24066064e-01 2.14681342e-01
8.00242901e-01 -1.03207147e+00 -5.90881944e-01 2.14124448e-03
-1.05621159e+00 -2.98076689e-01 -1.58887714e-01 3.19753617e-01
-2.33726382e-01 3.21068130e-02 -8.19366753e-01 -5.23470759e-01
-4.03381258e-01 -7.86408722e-01 6.13993764e-01 -7.44000673e-02
-6.55070901e-01 -1.19377995e+00 -2.11004198e-01 6.50787175e-01
3.26077044e-01 3.27462196e-01 1.38806796e+00 -1.41493547e+00
2.88226139e-02 -2.96715826e-01 -1.75469946e-02 3.28777879e-01
-7.71999732e-02 -1.69306457e-01 -1.02122867e+00 -3.79703864e-02
2.48390734e-01 -4.03539717e-01 9.93680000e-01 -5.90916052e-02
9.19722557e-01 -8.61411810e-01 -2.16845810e-01 3.09096038e-01
1.21406984e+00 -1.89635158e-01 3.98148984e-01 3.93758208e-01
5.57524800e-01 6.01937830e-01 2.74631798e-01 1.27425551e-01
-6.85170991e-03 3.70896846e-01 4.65187579e-02 6.97908625e-02
2.94452184e-03 -3.58209372e-01 7.35840797e-01 4.64569241e-01
2.33521819e-01 -2.80291885e-01 -6.87230110e-01 5.66121578e-01
-1.40532506e+00 -1.22423708e+00 -7.00048804e-02 2.19042087e+00
1.23554051e+00 4.44421560e-01 9.10270736e-02 -1.28247902e-01
4.63470757e-01 2.84452680e-02 -3.36189836e-01 -6.10909581e-01
-2.80977219e-01 6.82457685e-02 9.95099247e-01 5.02167583e-01
-1.04119837e+00 9.96976733e-01 6.47990751e+00 6.74790263e-01
-1.42183673e+00 5.81564456e-02 5.95510185e-01 -1.73494652e-01
-4.87208813e-01 -1.97872370e-01 -8.01038802e-01 5.48681974e-01
8.93350899e-01 -3.94361466e-01 3.77571583e-01 1.02119839e+00
1.40145823e-01 1.83898658e-01 -1.42127466e+00 5.53209484e-01
4.56426948e-01 -1.20555627e+00 1.04612194e-01 3.90536850e-03
7.28608251e-01 1.47229806e-01 1.54239044e-01 4.15850639e-01
6.31378770e-01 -1.03179669e+00 1.03906655e+00 1.74252093e-01
3.32182080e-01 -7.24786162e-01 7.27554262e-01 4.19126481e-01
-2.22949341e-01 -1.33352071e-01 -3.31348568e-01 6.28202483e-02
-3.39243412e-01 6.13154948e-01 -6.90228522e-01 8.26348066e-02
2.56390899e-01 6.41977310e-01 -7.90521324e-01 4.38398033e-01
-2.82900155e-01 9.42489624e-01 -5.56014571e-03 -1.53400088e-02
1.13717757e-01 3.01416386e-02 7.71400750e-01 1.20492411e+00
1.82239190e-01 -6.17583320e-02 3.00502956e-01 1.15540254e+00
-5.06629765e-01 3.31728011e-01 -6.27671182e-01 -6.09046161e-01
2.36547217e-01 1.38652086e+00 -6.05101228e-01 -3.84423137e-01
-4.52651501e-01 9.11692619e-01 1.55937195e-01 2.48286828e-01
-5.81442773e-01 -1.63958579e-01 3.99702758e-01 5.04658759e-01
-2.29115225e-02 1.68672934e-01 -6.18223310e-01 -1.10524428e+00
-2.92874157e-01 -1.25511444e+00 5.03188252e-01 -4.69808549e-01
-1.86611998e+00 1.70616791e-01 -2.26694256e-01 -6.85285270e-01
1.31929860e-01 -6.95377767e-01 -6.79340720e-01 1.03446925e+00
-1.19804657e+00 -1.19429481e+00 3.67548138e-01 2.39061221e-01
4.72876281e-01 -2.98781544e-01 6.86547041e-01 1.12958290e-01
-2.74826139e-01 7.91863978e-01 1.20644212e-01 5.42011857e-01
1.03430426e+00 -1.24082625e+00 -3.96259613e-02 4.95791286e-01
3.49776953e-01 8.67812335e-01 9.78901505e-01 -7.95584261e-01
-1.35303748e+00 -1.13714719e+00 1.26495290e+00 -1.12552178e+00
1.16579902e+00 -4.29207861e-01 -8.62940609e-01 7.30316222e-01
1.36876881e-01 -6.06798053e-01 8.93324673e-01 6.94758773e-01
-1.01849270e+00 9.96790268e-03 -1.22543561e+00 7.18693912e-01
8.09787452e-01 -4.72281098e-01 -8.43391359e-01 5.37189841e-01
4.49254423e-01 -5.48891686e-02 -8.05770338e-01 -6.42888546e-02
5.25074184e-01 -4.66412544e-01 7.59534061e-01 -8.10499549e-01
1.02768004e+00 2.36812592e-01 -1.66860372e-01 -1.38006783e+00
-4.64852333e-01 -4.25855488e-01 2.18170568e-01 1.45682466e+00
7.49580145e-01 -4.01828021e-01 5.02255738e-01 6.27533197e-01
3.56226340e-02 -4.19170499e-01 -5.48372388e-01 -8.14316392e-01
5.66704392e-01 -1.30725622e-01 9.26821083e-02 1.50381303e+00
6.68487921e-02 8.54533553e-01 -2.59253234e-01 1.55743152e-01
4.53667462e-01 3.36435556e-01 3.67597282e-01 -1.54170489e+00
-4.94394153e-01 -6.58913732e-01 -1.45826191e-01 -5.07572353e-01
5.58195293e-01 -1.44730580e+00 5.19452803e-02 -9.83804584e-01
4.15469021e-01 -6.12815261e-01 -4.73460406e-01 4.63956386e-01
-2.59282365e-02 3.54543984e-01 3.45141590e-02 1.29665747e-01
-3.45205635e-01 3.02145313e-02 1.12388146e+00 -2.39050165e-01
-2.00334549e-01 -3.66641730e-02 -1.47998035e+00 9.99949813e-01
8.72477353e-01 -6.92956805e-01 -2.14831054e-01 -1.83795750e-01
5.02299905e-01 -3.96595359e-01 5.88015676e-01 -4.82194215e-01
-2.68016636e-01 -4.70281273e-01 3.64596367e-01 -2.80320644e-03
6.72708498e-03 -6.59036815e-01 -3.31755102e-01 3.76841217e-01
-9.59069788e-01 -1.18470296e-01 2.51217097e-01 5.85529447e-01
3.31373513e-01 -2.53662020e-01 8.34509969e-01 -2.20896140e-01
-9.94355008e-02 -2.34639063e-01 -2.91914701e-01 5.12297571e-01
3.90904963e-01 5.17929554e-01 -5.23658931e-01 -4.04271930e-01
-5.39716601e-01 -8.78004655e-02 6.27097785e-01 7.51220107e-01
-1.11164384e-01 -1.13304710e+00 -1.04899561e+00 -1.49918988e-01
2.17315912e-01 -6.02513254e-01 -5.07871866e-01 7.83571362e-01
-7.94515666e-03 4.81892973e-02 1.62836611e-01 1.02426209e-01
-1.34734368e+00 7.90997684e-01 -3.99072543e-02 -1.12801358e-01
-4.79925334e-01 7.27635264e-01 1.98965088e-01 -1.13543697e-01
1.43773124e-01 -2.07993016e-01 -2.02811152e-01 2.79235244e-01
4.02821690e-01 1.79347321e-01 -9.31500942e-02 -7.35835612e-01
-2.45884866e-01 -7.91247189e-02 -3.48641396e-01 -7.79729057e-03
1.18704426e+00 3.41615230e-01 5.18579111e-02 7.11183012e-01
1.15042663e+00 9.05987680e-01 -4.50919092e-01 -2.56666750e-01
3.25617582e-01 -2.73458987e-01 2.93082178e-01 -1.37119257e+00
-8.69673610e-01 4.25980300e-01 2.61663079e-01 1.72739744e-01
4.89797831e-01 2.48771280e-01 5.74664414e-01 2.91904390e-01
3.11237257e-02 -1.20966971e+00 -1.34890273e-01 3.92761797e-01
1.14461327e+00 -1.17369020e+00 9.98532474e-02 -3.51643413e-01
-7.82623589e-01 7.73236692e-01 1.71184719e-01 -1.84775800e-01
4.48992342e-01 3.21075350e-01 5.22256076e-01 -2.58805752e-01
-5.73742747e-01 1.52463406e-01 4.65106428e-01 1.91799894e-01
7.61798382e-01 9.06431824e-02 -6.98266327e-01 1.05290639e+00
-5.21700203e-01 -4.72011387e-01 5.45848548e-01 4.67598140e-01
-2.84889579e-01 -9.38197255e-01 -3.26837599e-01 6.49992406e-01
-1.12749994e+00 -5.37743211e-01 -1.15123284e+00 6.04428530e-01
2.75144160e-01 8.24308157e-01 1.55881876e-02 -1.31536394e-01
1.67582542e-01 3.69310826e-01 3.24046552e-01 -6.95046604e-01
-1.09109330e+00 2.02886690e-03 4.28919673e-01 -1.37405261e-01
-1.34543508e-01 -6.03923619e-01 -8.75375152e-01 -4.59687233e-01
-2.41686463e-01 3.29628944e-01 5.49193025e-01 8.93911958e-01
2.73945481e-01 3.50394458e-01 3.96969646e-01 -2.03994170e-01
-1.00712013e+00 -8.83513808e-01 -7.04806983e-01 1.06116056e+00
9.21483859e-02 -4.05491501e-01 -6.34585679e-01 1.62180662e-01] | [8.980667114257812, 10.242315292358398] |
c2c58f14-d1ed-4565-85bc-d59932450d8d | fusion-of-simple-models-for-native-language | null | null | https://aclanthology.org/W17-5048 | https://aclanthology.org/W17-5048.pdf | Fusion of Simple Models for Native Language Identification | In this paper we describe the approaches we explored for the 2017 Native Language Identification shared task. We focused on simple word and sub-word units avoiding heavy use of hand-crafted features. Following recent trends, we explored linear and neural networks models to attempt to compensate for the lack of rich feature use. Initial efforts yielded f1-scores of 82.39{\%} and 83.77{\%} in the development and test sets of the fusion track, and were officially submitted to the task as team L2F. After the task was closed, we carried on further experiments and relied on a late fusion strategy for combining our simple proposed approaches with modifications of the baselines provided by the task. As expected, the i-vectors based sub-system dominates the performance of the system combinations, and results in the major contributor to our achieved scores. Our best combined system achieves 90.1{\%} and 90.2{\%} f1-score in the development and test sets of the fusion track, respectively. | ['Ramon Astudillo', 'Fabio Kepler', 'Alberto Abad'] | 2017-09-01 | null | null | null | ws-2017-9 | ['native-language-identification'] | ['natural-language-processing'] | [ 4.11557779e-02 -2.18446003e-04 -3.28746103e-02 -3.80854726e-01
-1.21790469e+00 -7.45466411e-01 9.69247103e-01 1.01908699e-01
-1.04288602e+00 7.00314224e-01 2.19482392e-01 -4.83576745e-01
6.63767532e-02 -1.47912055e-01 -4.62381750e-01 -4.90524232e-01
2.16176003e-01 3.64075691e-01 1.93514507e-02 -2.14917988e-01
1.21899031e-01 8.57497156e-02 -1.50925398e+00 3.15909475e-01
7.51330018e-01 7.85811365e-01 1.58254281e-01 8.12251210e-01
-6.24864586e-02 4.40721095e-01 -7.99078882e-01 -7.29476511e-01
4.45744872e-01 -5.26585728e-02 -6.03676617e-01 -3.76447618e-01
7.27266133e-01 -3.37501377e-01 -3.23352396e-01 9.91663694e-01
7.54334509e-01 1.72378510e-01 7.80952394e-01 -1.01114929e+00
-5.12687147e-01 9.46421623e-01 -3.95654470e-01 1.47475585e-01
3.59760761e-01 2.32855722e-01 1.00623047e+00 -1.32833123e+00
3.82088542e-01 1.05414855e+00 8.60632837e-01 6.65947437e-01
-1.17012966e+00 -1.05829203e+00 1.43667206e-01 -2.54738271e-01
-1.65213358e+00 -9.62478936e-01 2.02153996e-01 -5.14954388e-01
1.48714483e+00 8.37099627e-02 1.36411831e-01 1.17343080e+00
-2.66588897e-01 7.93322504e-01 1.20842171e+00 -8.02161217e-01
-1.59060895e-01 5.75557649e-01 5.12673140e-01 4.53932643e-01
3.76740366e-01 8.48717317e-02 -4.84287292e-01 -6.48216158e-02
3.59304845e-01 -3.87670279e-01 -2.69929409e-01 3.85143638e-01
-1.17404377e+00 7.73315430e-01 3.62488963e-02 6.95282698e-01
-1.55154690e-01 -1.35600865e-01 3.83438677e-01 1.80023640e-01
4.69260961e-01 3.99092704e-01 -7.09372282e-01 -4.13351476e-01
-1.25088930e+00 2.96170622e-01 7.06014752e-01 8.84189188e-01
4.60176736e-01 2.67670602e-02 -5.58111012e-01 1.16499710e+00
1.32299855e-01 4.55689520e-01 5.37392080e-01 -6.65079534e-01
6.91508055e-01 4.86177593e-01 1.19821385e-01 -4.62170988e-01
-4.11763132e-01 -6.97495103e-01 -5.09036183e-01 -1.09102428e-01
7.40522742e-01 -3.49026889e-01 -1.16057789e+00 1.97506642e+00
-1.43833682e-01 -2.76225895e-01 6.08041435e-02 3.19943994e-01
8.68768275e-01 4.83374149e-01 3.64493251e-01 -2.00519666e-01
1.30338264e+00 -8.67484391e-01 -7.09458113e-01 -8.53791311e-02
9.81030047e-01 -1.10711205e+00 9.13392186e-01 2.67652541e-01
-1.15290594e+00 -8.19655716e-01 -1.15569460e+00 2.22742707e-01
-4.69764292e-01 4.22924370e-01 2.10872680e-01 1.00187266e+00
-1.46209383e+00 5.59203088e-01 -4.59343731e-01 -4.75211173e-01
1.75381824e-01 6.28786147e-01 -5.56194007e-01 -1.25781521e-01
-1.11424232e+00 9.74660993e-01 5.43700874e-01 -1.19174197e-01
-5.08893967e-01 -6.13210917e-01 -6.44167483e-01 -1.32818237e-01
2.10840791e-01 -1.97966561e-01 1.45450115e+00 -5.56654572e-01
-1.19097376e+00 1.02770185e+00 -1.23425223e-01 -3.50017220e-01
3.78906578e-01 -4.63893354e-01 -5.32668948e-01 -3.80944371e-01
2.05801666e-01 8.32541466e-01 3.68426949e-01 -9.57654893e-01
-9.49500561e-01 -1.17338225e-01 -1.12485625e-01 1.49368212e-01
-7.05944359e-01 3.51168722e-01 -5.10894001e-01 -6.70842588e-01
-2.85370916e-01 -1.02012873e+00 1.21549018e-01 -7.80995190e-01
-3.33612502e-01 -5.72868288e-01 3.21399868e-01 -1.03751636e+00
1.53791201e+00 -2.17579293e+00 -2.38206521e-01 1.13239914e-01
1.54326752e-01 8.65344763e-01 -2.16794789e-01 4.61892456e-01
-2.19844460e-01 4.09990221e-01 4.51983558e-03 -7.19015241e-01
1.01653427e-01 -2.75850564e-01 -1.69785228e-04 2.27139264e-01
1.39382377e-01 6.54118598e-01 -6.18734181e-01 -7.83119798e-02
4.67747636e-03 5.52213252e-01 -3.89531285e-01 1.97063655e-01
2.38349348e-01 6.18760027e-02 5.18565886e-02 4.78080451e-01
6.23683274e-01 2.83267379e-01 -7.54546747e-02 -9.82571021e-02
-3.56684685e-01 6.72428429e-01 -1.11832297e+00 1.57785392e+00
-3.63077313e-01 5.69921315e-01 2.10503116e-01 -6.24264300e-01
9.84441876e-01 5.61192870e-01 3.23622376e-01 -6.60041690e-01
2.46839434e-01 5.59783161e-01 3.96500319e-01 4.53972211e-03
6.43632114e-01 7.85804614e-02 -1.17883123e-01 4.29041743e-01
4.93857443e-01 1.29608959e-01 3.21981758e-01 2.76194602e-01
9.38803256e-01 8.45299140e-02 3.47037882e-01 -5.03857613e-01
6.61910176e-01 -1.75853178e-01 3.04162472e-01 9.90845740e-01
-4.63360608e-01 7.05312550e-01 9.57288444e-02 -2.14828774e-01
-9.88475919e-01 -7.84924507e-01 -1.20356210e-01 1.26314175e+00
-5.40595531e-01 -7.76141167e-01 -9.68949914e-01 -6.83876276e-01
-1.96434379e-01 7.94014335e-01 -4.36913878e-01 -8.05902034e-02
-4.94799972e-01 -6.18586183e-01 9.58379865e-01 5.35184085e-01
3.30787301e-01 -7.00918257e-01 4.57432382e-02 2.26793095e-01
-7.03684539e-02 -1.12858486e+00 -6.51898980e-01 2.87567228e-01
-2.21933812e-01 -7.93301761e-01 -1.10913217e+00 -8.20200145e-01
1.88615143e-01 2.60063559e-01 9.51765537e-01 3.97155993e-02
6.06937483e-02 1.20958902e-01 -3.83395821e-01 -5.38190067e-01
-4.29725111e-01 3.88468295e-01 4.01649803e-01 -1.62737086e-01
5.96749365e-01 -1.46691380e-02 -3.46073329e-01 1.75819010e-01
-4.03782248e-01 9.68016405e-03 6.50702238e-01 8.39621842e-01
2.34157994e-01 -2.14189410e-01 4.97666270e-01 -7.48057127e-01
7.65418649e-01 -1.91659749e-01 -3.34666163e-01 1.61272958e-01
-7.03634262e-01 -9.26890075e-02 4.08683062e-01 -6.28493667e-01
-7.89604127e-01 5.43658100e-02 -4.01244044e-01 -1.49115980e-01
-2.49501199e-01 2.83677787e-01 -2.25414455e-01 1.03587113e-01
5.84971786e-01 8.33869204e-02 -3.19422245e-01 -6.24819040e-01
2.89028376e-01 1.15994823e+00 5.38808405e-01 -5.12135267e-01
6.46280050e-01 -3.06969374e-01 -6.69930935e-01 -6.78760290e-01
-6.40582263e-01 -5.70530593e-01 -5.07514000e-01 1.05044723e-01
8.03329766e-01 -1.23167288e+00 -5.43710351e-01 6.19939268e-01
-1.10888553e+00 -1.75167710e-01 -1.03361875e-01 7.50908971e-01
-6.33955225e-02 1.21202640e-01 -2.69142389e-01 -1.06544805e+00
-5.88561058e-01 -1.30321670e+00 8.92285407e-01 2.28939846e-01
-6.04964018e-01 -8.00372899e-01 1.46582246e-01 4.94335204e-01
7.83960581e-01 -2.04649553e-01 5.16390145e-01 -1.29747784e+00
2.96950221e-01 -4.68717813e-01 -3.35152954e-01 7.08409905e-01
2.95028817e-02 -1.04952402e-01 -1.40532064e+00 -5.48662603e-01
-1.68340698e-01 -2.71598160e-01 8.99200439e-01 1.44760862e-01
6.42956614e-01 -2.05001414e-01 -2.12454006e-01 3.54866505e-01
1.03207982e+00 2.29675010e-01 1.47511065e-01 1.94296703e-01
5.59924960e-01 7.04768181e-01 1.53983399e-01 7.80313537e-02
5.65734267e-01 9.49960709e-01 -2.15747863e-01 3.58296260e-02
-4.19068158e-01 -3.46539974e-01 7.78528512e-01 1.02905476e+00
-2.69039124e-02 -3.02725822e-01 -1.26412165e+00 7.26304471e-01
-1.61821806e+00 -6.30440891e-01 -1.78765878e-01 2.37827015e+00
8.20424378e-01 4.38764095e-01 2.79134274e-01 1.65640503e-01
8.01904261e-01 -2.64859676e-01 -9.12370980e-02 -5.55309057e-01
-3.20214391e-01 4.30937022e-01 6.17980123e-01 5.61027408e-01
-1.37262762e+00 1.04049981e+00 6.52497721e+00 1.22571743e+00
-9.96799529e-01 1.94863513e-01 4.99105215e-01 -2.22016305e-01
-1.13177918e-01 -1.49540603e-01 -1.34961128e+00 4.09949511e-01
1.47795749e+00 -2.05797210e-01 2.55112350e-01 4.15192723e-01
-2.91881889e-01 -6.17835037e-02 -1.10245037e+00 1.03379631e+00
1.00306585e-01 -7.82521069e-01 -5.03667332e-02 1.58446565e-01
7.90069878e-01 3.88906360e-01 1.03940628e-01 7.44845927e-01
4.30608630e-01 -1.24167752e+00 7.12467134e-01 1.78278059e-01
1.15117073e+00 -6.73790812e-01 9.45980728e-01 4.51587141e-01
-1.16254008e+00 6.89962432e-02 5.44165820e-02 -2.41629273e-01
-2.16335222e-01 3.72796357e-01 -9.59311008e-01 5.31336129e-01
4.81987029e-01 3.41704637e-01 -6.53555810e-01 7.34437704e-01
-5.67752905e-02 8.27644765e-01 -4.61933523e-01 -5.61299175e-02
2.54314810e-01 3.53333920e-01 4.86957163e-01 1.59193182e+00
3.27391088e-01 -4.13444549e-01 2.73598135e-01 2.55922288e-01
-3.10512573e-01 3.37757498e-01 -5.63476026e-01 -1.75793260e-01
4.92397994e-01 1.27583480e+00 -3.98426592e-01 -5.05503476e-01
-5.84141552e-01 6.62554324e-01 5.53886771e-01 2.98274249e-01
-5.75385213e-01 -4.87206101e-01 6.12430632e-01 4.85699475e-02
2.52452880e-01 -2.55356222e-01 -3.73332560e-01 -1.14884305e+00
-7.46790469e-02 -1.03663552e+00 2.97406912e-01 -2.73044854e-01
-1.02646172e+00 1.02660906e+00 1.49454594e-01 -9.01997805e-01
-6.05498493e-01 -7.60966957e-01 -3.82885873e-01 1.40582037e+00
-1.07562232e+00 -1.15558755e+00 1.12844996e-01 4.40681815e-01
4.48008895e-01 -7.45322943e-01 1.07218564e+00 4.79025275e-01
-6.63821757e-01 1.26590991e+00 1.77290156e-01 1.78236470e-01
1.07967317e+00 -1.12628722e+00 5.47539592e-01 1.13734674e+00
3.38689148e-01 9.37140942e-01 5.83054423e-01 -5.20182729e-01
-1.00173104e+00 -8.83487463e-01 1.39875460e+00 -4.95398194e-01
5.80164194e-01 -5.96389294e-01 -7.37265706e-01 4.63375777e-01
4.84193653e-01 -4.15903896e-01 8.09070349e-01 2.77526557e-01
-4.51135069e-01 9.06816050e-02 -1.15968764e+00 4.70455498e-01
8.62990618e-01 -7.25032747e-01 -3.70703906e-01 8.74560699e-02
6.49757147e-01 -2.16578871e-01 -8.14353704e-01 5.34655809e-01
6.69447124e-01 -6.27757967e-01 7.24253595e-01 -3.59431744e-01
1.92782003e-02 -1.45337075e-01 -4.09881741e-01 -1.31111848e+00
-5.19820273e-01 -5.63501775e-01 4.72893678e-02 1.70614421e+00
7.47528076e-01 -4.72111136e-01 3.30440640e-01 7.04269290e-01
-1.28396198e-01 -5.60174704e-01 -9.51227248e-01 -6.53726935e-01
3.05399269e-01 -6.44812226e-01 1.87682539e-01 7.55320907e-01
-3.16455692e-01 4.90732312e-01 -2.70801067e-01 -3.14216435e-01
4.50164497e-01 -7.01841831e-01 7.25318253e-01 -9.86403465e-01
-2.34595239e-01 -7.65977859e-01 -4.99195233e-02 -5.81865132e-01
1.14689372e-01 -1.16838670e+00 -1.07470773e-01 -1.15445161e+00
2.46908203e-01 -3.88636500e-01 -5.87559938e-01 7.61994123e-01
-4.63905185e-01 6.86914861e-01 4.47427779e-01 2.83075094e-01
-2.81724066e-01 1.52541757e-01 5.98647118e-01 -1.43337354e-01
-1.40311196e-01 -6.21149577e-02 -1.19826877e+00 4.96500254e-01
9.49457586e-01 -2.55437016e-01 -1.26779437e-01 -7.02336371e-01
-1.40563965e-01 -5.58671355e-01 -1.48783877e-01 -9.74504530e-01
2.68758953e-01 3.03136170e-01 3.98738891e-01 -3.50082606e-01
2.87459701e-01 -4.07145023e-01 5.26600629e-02 4.22050893e-01
-3.44959378e-01 1.87915906e-01 5.50176084e-01 -2.17514783e-01
-1.77539259e-01 -1.73319712e-01 6.32150114e-01 5.40500395e-02
-4.01552677e-01 -2.13321280e-02 -3.10382813e-01 -6.01450764e-02
8.37690711e-01 -1.97722495e-01 -1.61338612e-01 -3.61820906e-01
-4.83186930e-01 1.56925112e-01 2.64882714e-01 7.59353518e-01
9.69709083e-02 -1.19708836e+00 -1.16358435e+00 3.03157955e-01
1.52728707e-01 -4.52316344e-01 -1.18856058e-01 7.04816937e-01
-9.65025201e-02 8.11326087e-01 -9.84732285e-02 -2.50273138e-01
-1.34986448e+00 8.35178345e-02 1.12031475e-01 -5.13066173e-01
-1.00060932e-01 1.07367229e+00 1.07226558e-01 -5.69265485e-01
5.23751676e-01 1.33145139e-01 -3.32771152e-01 1.75398603e-01
8.13282430e-01 4.47755396e-01 3.95092249e-01 -1.10458314e+00
-6.62760317e-01 3.46800119e-01 -5.50180733e-01 -4.73614991e-01
1.16289866e+00 -5.53324213e-03 1.53571159e-01 4.29183662e-01
1.34102881e+00 2.38774866e-01 -7.30077028e-01 -3.73139292e-01
-3.36267538e-02 -4.57220711e-02 1.64054871e-01 -1.27157605e+00
-6.61868095e-01 8.69082928e-01 6.76603675e-01 1.07922129e-01
8.07478011e-01 -2.07695067e-01 5.95323563e-01 1.54841408e-01
1.87284380e-01 -1.11417818e+00 -5.88526368e-01 9.50318515e-01
6.84689820e-01 -1.17381155e+00 -9.11532491e-02 -2.22667202e-01
-3.15406770e-01 9.88055944e-01 5.72864294e-01 4.96276431e-02
3.83019149e-01 4.30590361e-01 2.48645946e-01 4.77046967e-01
-7.68823802e-01 -3.09037507e-01 7.08877563e-01 5.09306312e-01
1.08234525e+00 2.49564737e-01 -6.05883062e-01 1.00449431e+00
-3.59424055e-01 -1.67942777e-01 8.85643438e-02 6.96594834e-01
-2.78839797e-01 -1.30480838e+00 -4.03127521e-01 4.52942252e-01
-8.37963343e-01 -4.15447354e-01 -5.80369413e-01 7.62653649e-01
4.16727096e-01 1.10197091e+00 -9.26586166e-02 -8.67627442e-01
3.87950778e-01 4.71888959e-01 3.81106496e-01 -5.03047645e-01
-1.11191511e+00 9.42001268e-02 2.77237684e-01 -3.83301198e-01
1.69250276e-02 -7.08012223e-01 -8.80786717e-01 -2.55270481e-01
-3.93219650e-01 1.73216358e-01 8.03645968e-01 9.44581091e-01
2.76253134e-01 3.19544822e-01 2.84479976e-01 -8.04846346e-01
-6.23245358e-01 -1.56964481e+00 -1.32983059e-01 4.26889151e-01
3.11900675e-01 -5.59130132e-01 -2.24670380e-01 -1.67512655e-01] | [10.40861988067627, 10.534111022949219] |
27b83f5d-83c8-44fc-95a6-aa3f1b0c132e | inter-slice-context-residual-learning-for-3d | 2011.14155 | null | https://arxiv.org/abs/2011.14155v1 | https://arxiv.org/pdf/2011.14155v1.pdf | Inter-slice Context Residual Learning for 3D Medical Image Segmentation | Automated and accurate 3D medical image segmentation plays an essential role in assisting medical professionals to evaluate disease progresses and make fast therapeutic schedules. Although deep convolutional neural networks (DCNNs) have widely applied to this task, the accuracy of these models still need to be further improved mainly due to their limited ability to 3D context perception. In this paper, we propose the 3D context residual network (ConResNet) for the accurate segmentation of 3D medical images. This model consists of an encoder, a segmentation decoder, and a context residual decoder. We design the context residual module and use it to bridge both decoders at each scale. Each context residual module contains both context residual mapping and context attention mapping, the formal aims to explicitly learn the inter-slice context information and the latter uses such context as a kind of attention to boost the segmentation accuracy. We evaluated this model on the MICCAI 2018 Brain Tumor Segmentation (BraTS) dataset and NIH Pancreas Segmentation (Pancreas-CT) dataset. Our results not only demonstrate the effectiveness of the proposed 3D context residual learning scheme but also indicate that the proposed ConResNet is more accurate than six top-ranking methods in brain tumor segmentation and seven top-ranking methods in pancreas segmentation. Code is available at https://git.io/ConResNet | ['Yong Xia', 'Yan Wang', 'Yutong Xie', 'Jianpeng Zhang'] | 2020-11-28 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [ 2.64229834e-01 3.62358004e-01 -2.92568207e-01 -4.57040399e-01
-9.82543707e-01 -1.73820183e-01 4.76734132e-01 2.21658468e-01
-4.67596650e-01 3.26896310e-01 4.68641907e-01 -4.99644548e-01
1.63142178e-02 -4.28782344e-01 -6.07408464e-01 -6.50675118e-01
6.00934029e-02 5.92455029e-01 4.20130938e-01 -1.44290389e-03
3.58143859e-02 4.37603325e-01 -8.09483647e-01 3.93432051e-01
1.15386891e+00 1.19692063e+00 6.67562664e-01 5.82323670e-01
-3.94360781e-01 9.52352107e-01 -4.03215319e-01 -6.53181896e-02
1.31232530e-01 -4.39945459e-01 -8.16341639e-01 -1.58354379e-02
2.14204505e-01 -2.80499965e-01 -3.36412877e-01 1.18723977e+00
6.67789817e-01 -3.34476531e-01 5.39217830e-01 -7.64643550e-01
-4.85398084e-01 7.74296522e-01 -4.09146160e-01 4.28865016e-01
1.24977112e-01 1.36274651e-01 6.06027901e-01 -6.15163863e-01
5.42767584e-01 8.27785552e-01 7.04556346e-01 6.96483791e-01
-6.77784324e-01 -4.68969524e-01 3.29914212e-01 2.05666244e-01
-1.13698554e+00 1.51772425e-01 6.45035863e-01 -4.05247152e-01
7.63294756e-01 1.68179542e-01 1.03664839e+00 1.07407868e+00
7.04141796e-01 1.08116853e+00 9.44260001e-01 -1.13674909e-01
1.64418161e-01 -2.22689196e-01 4.37846899e-01 6.30684733e-01
6.86577559e-02 9.66465548e-02 5.39216623e-02 2.15956241e-01
8.46386611e-01 2.13054404e-01 -4.66268122e-01 -1.84166729e-01
-1.39678264e+00 6.91925466e-01 1.14729667e+00 5.36490917e-01
-4.59709913e-01 2.51135856e-01 4.08325255e-01 -7.88810626e-02
4.15710002e-01 1.67268470e-01 -4.08897430e-01 1.96351320e-01
-8.12743783e-01 -1.18675776e-01 6.21455133e-01 9.23153102e-01
3.49874003e-03 -2.25336209e-01 -4.18777436e-01 7.93154001e-01
3.66029590e-01 1.48828685e-01 9.10685122e-01 -5.47636807e-01
1.72321513e-01 8.65535796e-01 -4.26818281e-01 -4.57948029e-01
-9.54586506e-01 -7.30459213e-01 -9.71929550e-01 9.82444659e-02
2.78517425e-01 1.09240204e-01 -1.58911431e+00 1.64324820e+00
3.32466096e-01 3.77455592e-01 -9.42566916e-02 1.27101743e+00
1.38421297e+00 2.52562284e-01 3.90510917e-01 9.05341133e-02
1.48439646e+00 -1.24267232e+00 -4.78317648e-01 -6.42352402e-02
8.57056439e-01 -6.93057179e-01 8.61643493e-01 1.56439707e-01
-1.05778408e+00 -3.53824615e-01 -9.64844644e-01 -2.47898057e-01
-2.87296146e-01 9.29930508e-02 5.47901809e-01 3.43743235e-01
-1.24485111e+00 2.68977314e-01 -1.04388857e+00 -3.74112189e-01
7.11595714e-01 4.69947964e-01 -2.50956059e-01 -1.71004042e-01
-1.20562506e+00 1.04588568e+00 3.53666782e-01 1.98794499e-01
-1.09496343e+00 -8.27943742e-01 -8.08988869e-01 -1.06873013e-01
2.93509394e-01 -9.19135630e-01 1.42357278e+00 -9.65415239e-01
-1.44863605e+00 1.08218801e+00 1.65047526e-01 -5.97097576e-01
6.26730800e-01 4.26818430e-02 -2.02014938e-01 3.49180728e-01
1.25738561e-01 9.50292289e-01 4.11048740e-01 -1.00823975e+00
-6.94705904e-01 -4.64278668e-01 3.50013115e-02 4.32624757e-01
2.33113170e-01 -4.02515292e-01 -7.39394486e-01 -7.99987137e-01
4.54453319e-01 -9.13828135e-01 -6.21273994e-01 -5.35026826e-02
-4.15837497e-01 -2.27540433e-01 6.35348678e-01 -8.53675842e-01
7.84497201e-01 -1.91877246e+00 3.12091321e-01 4.18086909e-02
3.40114802e-01 1.78668708e-01 -7.46889710e-02 -4.17389095e-01
-1.52039722e-01 1.08585298e-01 -5.96816182e-01 -7.66994134e-02
-1.87811166e-01 3.02449554e-01 2.27798730e-01 4.64163721e-01
-7.98172206e-02 1.17986548e+00 -8.38946939e-01 -4.09210920e-01
3.86886090e-01 6.58893347e-01 -5.46351969e-01 2.36448824e-01
-3.30635071e-01 7.37750649e-01 -5.64733148e-01 8.10157359e-01
5.61407089e-01 -4.43277836e-01 1.38242081e-01 -3.76925051e-01
1.66344345e-02 2.44384393e-01 -5.03767729e-01 2.10067725e+00
-2.88513064e-01 1.99964106e-01 1.71874598e-01 -1.00738609e+00
5.70538223e-01 4.39852089e-01 8.97672296e-01 -1.04318893e+00
4.15154397e-01 4.61803049e-01 1.36229575e-01 -5.78509390e-01
1.32448554e-01 -2.01033607e-01 -8.88015255e-02 2.78638285e-02
-2.48428341e-02 -2.78127223e-01 -9.05659571e-02 1.62553310e-01
1.02179253e+00 9.21597704e-02 3.67622018e-01 -4.44963396e-01
6.75389469e-01 1.84304416e-01 7.73037970e-01 4.71514463e-01
-5.47442853e-01 6.75030470e-01 5.79317153e-01 -4.96015728e-01
-6.08553350e-01 -8.71005177e-01 -3.17763299e-01 7.06542253e-01
6.42224669e-01 -1.09049134e-01 -8.02734137e-01 -1.23870850e+00
-2.12893207e-02 6.31989300e-01 -8.79236639e-01 -1.56735629e-01
-7.11211920e-01 -7.69363225e-01 3.39992225e-01 7.76353776e-01
6.07596755e-01 -9.72787440e-01 -7.29390681e-01 2.02381656e-01
-1.87423944e-01 -1.08953929e+00 -7.58503139e-01 5.21317184e-01
-1.23664415e+00 -1.31862926e+00 -1.07828653e+00 -1.05969810e+00
8.26394260e-01 6.12183064e-02 1.11097503e+00 1.14603952e-01
-4.04576570e-01 3.55832577e-01 -4.29408282e-01 -2.30580017e-01
-4.56375808e-01 3.71248603e-01 -5.82676470e-01 -7.28623092e-01
4.22434211e-01 -3.66210014e-01 -8.87955129e-01 3.92420620e-01
-9.91886675e-01 5.08093536e-01 9.53080475e-01 9.83727634e-01
1.01429141e+00 -4.06447619e-01 4.85205263e-01 -1.10697031e+00
4.07688171e-01 -5.70890725e-01 -5.37045121e-01 1.81995690e-01
-4.74581659e-01 -2.23736480e-01 3.79508972e-01 -1.16384074e-01
-7.04537272e-01 1.10157527e-01 -7.63413548e-01 -3.50353241e-01
-2.72323728e-01 6.58271730e-01 -8.84212554e-02 -7.71518648e-02
3.61923397e-01 9.28708762e-02 -5.14860302e-02 -4.38659489e-01
4.05723043e-02 3.83387595e-01 4.72234845e-01 -3.16460371e-01
7.11468160e-02 1.36730045e-01 8.15246999e-02 -2.92621166e-01
-7.74280787e-01 -2.67512292e-01 -5.71845710e-01 -9.94584560e-02
1.27222097e+00 -8.59074831e-01 -4.47973996e-01 3.42889667e-01
-8.42972517e-01 -7.78718770e-01 -4.09467936e-01 4.64823484e-01
-4.80036646e-01 2.24440947e-01 -8.80892158e-01 5.06863557e-02
-6.24114633e-01 -1.89706087e+00 1.08206689e+00 2.95744985e-01
1.16001040e-01 -1.10341918e+00 -2.08351478e-01 3.21114093e-01
5.53710580e-01 5.05225956e-01 1.04062319e+00 -7.63080180e-01
-6.46910727e-01 -5.24638705e-02 -4.29852545e-01 1.32766426e-01
1.53434142e-01 -5.73847413e-01 -7.46554971e-01 -1.74923122e-01
-4.18285728e-02 1.33599602e-02 1.13232529e+00 9.75671291e-01
1.46640706e+00 8.04810226e-02 -5.25758147e-01 9.67486084e-01
1.37407172e+00 3.84228408e-01 4.32636827e-01 1.50567859e-01
8.61808658e-01 2.05769598e-01 2.20487341e-01 -6.72060549e-02
6.74283385e-01 4.61200833e-01 8.95937920e-01 -3.34074944e-01
-6.40805721e-01 6.82191402e-02 3.25762387e-03 1.10390925e+00
1.94758967e-01 -7.19133541e-02 -1.20472693e+00 5.85293770e-01
-1.61474848e+00 -3.59936833e-01 -1.30032450e-01 1.78733969e+00
7.76008308e-01 1.36337802e-01 -1.24602661e-01 -2.34306335e-01
7.24603474e-01 -5.23765311e-02 -8.54564190e-01 -2.80281961e-01
1.31625667e-01 5.08264825e-02 5.28412640e-01 5.19051671e-01
-1.24187207e+00 8.00953388e-01 5.45881414e+00 7.13583291e-01
-1.30538130e+00 3.34997922e-01 8.63240778e-01 1.26399726e-01
-3.79515618e-01 -3.46334159e-01 -5.38351655e-01 4.88325268e-01
5.31214893e-01 2.73991227e-01 -2.34413035e-02 8.48398924e-01
-1.72105152e-02 -1.81521654e-01 -1.14486694e+00 8.92909527e-01
-4.40541655e-02 -1.30582631e+00 7.69975409e-02 4.57090959e-02
6.68191969e-01 3.81806403e-01 -2.41730623e-02 3.90405715e-01
1.55373409e-01 -1.10453308e+00 5.71406543e-01 4.91245687e-01
8.44765306e-01 -6.56341076e-01 1.05439508e+00 1.43323988e-01
-1.09494436e+00 -1.76973082e-02 -1.98016018e-01 6.31155550e-01
1.28380075e-01 7.29535758e-01 -9.08963203e-01 5.90589225e-01
5.73759019e-01 9.32542086e-01 -4.47193027e-01 1.35695636e+00
-3.97442371e-01 3.60327035e-01 -1.73522979e-01 1.46881327e-01
4.86416936e-01 1.63230840e-02 6.84858322e-01 1.27133596e+00
2.85817295e-01 2.69366831e-01 3.44700903e-01 6.00721478e-01
-4.08422619e-01 1.22225799e-01 -8.53847116e-02 3.26016873e-01
-1.20505661e-01 1.28954554e+00 -1.04101324e+00 -3.94479394e-01
-3.62314969e-01 1.00145268e+00 3.35834138e-02 1.91640452e-01
-9.36341465e-01 -3.70033383e-02 5.09780765e-01 -2.93698944e-02
1.76334456e-01 9.45655406e-02 -6.29939020e-01 -1.14617503e+00
-4.03940707e-01 -7.19213128e-01 7.36064315e-01 -6.76238716e-01
-1.24843252e+00 8.57492149e-01 -2.88157880e-01 -1.09177315e+00
1.24757633e-01 -7.82722056e-01 -4.35826004e-01 7.76435316e-01
-1.90510499e+00 -1.20843399e+00 -6.85415685e-01 5.67800701e-01
7.03480124e-01 9.90412235e-02 6.08998060e-01 3.66124719e-01
-4.59205389e-01 5.11011004e-01 -2.79825687e-01 1.33693874e-01
6.39167726e-01 -1.50221503e+00 4.27593999e-02 5.61113656e-01
-3.79706681e-01 2.01690614e-01 1.78128988e-01 -7.34773219e-01
-1.23895609e+00 -1.31090343e+00 4.15108204e-01 -2.15375334e-01
2.01303095e-01 1.88851841e-02 -7.70483553e-01 6.77192628e-01
2.94203043e-01 2.79745132e-01 7.02654660e-01 -3.59345615e-01
1.25951648e-01 2.65844483e-02 -1.38016164e+00 5.39688885e-01
1.06109846e+00 -1.18572325e-01 -6.86056316e-01 2.59351462e-01
1.01428950e+00 -1.10006547e+00 -1.06538272e+00 8.68617773e-01
4.06840414e-01 -7.55229950e-01 9.28477883e-01 -1.44828171e-01
5.45381486e-01 -2.51801699e-01 -2.17466950e-01 -1.25286686e+00
-2.13078901e-01 7.90357664e-02 1.17837713e-04 4.18984473e-01
3.81179422e-01 -5.66446781e-01 7.72628188e-01 6.30929530e-01
-9.04560447e-01 -1.22531497e+00 -9.14039850e-01 -2.55734384e-01
4.38225329e-01 -4.57873404e-01 5.94784379e-01 9.42663968e-01
-1.26361802e-01 -6.72126096e-03 2.34949470e-01 2.01186955e-01
5.27311146e-01 1.35265052e-01 1.75766885e-01 -9.62532759e-01
5.38451504e-03 -9.55180109e-01 -3.92932892e-01 -1.17418432e+00
-1.01621456e-01 -1.43963277e+00 5.05783036e-03 -2.04806685e+00
3.59376341e-01 -5.13664424e-01 -6.72380805e-01 5.33205211e-01
-3.00551176e-01 2.82234550e-01 1.11557320e-01 1.55179366e-01
-6.63463354e-01 4.06606883e-01 1.75316429e+00 -3.34411800e-01
-1.98173180e-01 3.60640325e-02 -8.01502287e-01 7.13611364e-01
7.78436303e-01 -3.23238343e-01 -3.06511968e-01 -6.02181673e-01
-1.91558957e-01 3.59272063e-01 4.59160715e-01 -9.68590438e-01
4.46376860e-01 1.57913864e-01 7.53361046e-01 -9.00849283e-01
-1.41772315e-01 -1.08119154e+00 -1.76330402e-01 7.36318052e-01
-4.56856459e-01 -1.19469456e-01 3.16662282e-01 2.98583806e-01
-2.60825276e-01 -1.55052260e-01 9.34181690e-01 -3.75683606e-01
-7.65703976e-01 6.80670261e-01 -3.28730285e-01 8.97588804e-02
1.10390198e+00 -1.13230966e-01 -9.06390771e-02 5.36138490e-02
-1.15296900e+00 5.33442616e-01 2.71421522e-01 2.37458602e-01
8.06816399e-01 -1.24514437e+00 -6.73814654e-01 5.17121665e-02
-1.17256977e-02 5.12525916e-01 5.03037095e-01 1.35691559e+00
-8.61019373e-01 5.09548724e-01 -7.43532106e-02 -9.70768511e-01
-9.04113233e-01 3.52889806e-01 7.37509310e-01 -5.91365814e-01
-1.01332545e+00 9.47786927e-01 5.45014918e-01 -4.81442213e-01
5.56683123e-01 -9.51074719e-01 -3.70462358e-01 -2.48825639e-01
4.45869118e-01 -3.63979191e-02 2.07491636e-01 -5.15969098e-01
-4.64304149e-01 5.84198356e-01 -3.68653893e-01 3.16269696e-01
1.30277896e+00 -3.56271639e-02 2.54561678e-02 -6.40791282e-02
1.17225981e+00 -4.95870173e-01 -1.33708298e+00 -1.60076648e-01
1.25971511e-01 -9.90372598e-02 4.56821889e-01 -1.38116789e+00
-1.49934292e+00 9.26413774e-01 9.16182041e-01 -9.37230512e-02
1.49359572e+00 9.57937837e-02 1.14172435e+00 -1.49101436e-01
2.14314282e-01 -7.22338021e-01 -4.89238799e-02 5.75861871e-01
8.45803797e-01 -1.37644696e+00 -1.59914508e-01 -3.18824112e-01
-7.65276074e-01 1.13452446e+00 6.64994836e-01 -2.30390266e-01
7.73103297e-01 4.30600196e-01 2.43067831e-01 -2.66742319e-01
-4.67886925e-01 -3.38106692e-01 4.96579111e-01 2.96549231e-01
5.93926370e-01 1.57148898e-01 -4.09225315e-01 9.77019250e-01
1.26207069e-01 1.65748626e-01 2.06546828e-01 7.90661454e-01
-2.96164423e-01 -1.03438234e+00 -6.93975985e-02 5.24628282e-01
-6.26172781e-01 -2.09306534e-02 -1.35966524e-01 9.02359545e-01
2.37644270e-01 4.40908223e-01 -1.31201431e-01 -2.57473499e-01
4.14924711e-01 -3.88120599e-02 4.69863534e-01 -5.25848150e-01
-9.47739124e-01 3.61339509e-01 -2.31115565e-01 -6.51844084e-01
-3.68038893e-01 -3.90618384e-01 -1.55866253e+00 1.20765232e-01
-8.65689814e-02 -1.34091591e-02 8.13766837e-01 8.25208843e-01
2.11609423e-01 1.17546558e+00 3.79484892e-01 -7.44302988e-01
-1.38886899e-01 -9.15252447e-01 -3.04372370e-01 2.57844836e-01
2.30300292e-01 -5.15407860e-01 -4.42584194e-02 -1.46968752e-01] | [14.618496894836426, -2.5244925022125244] |
18621ccf-c2d7-4318-96e7-954f4b558b17 | electronic-structure-properties-from-atom | 2206.14087 | null | https://arxiv.org/abs/2206.14087v1 | https://arxiv.org/pdf/2206.14087v1.pdf | Electronic-structure properties from atom-centered predictions of the electron density | The electron density of a molecule or material has recently received major attention as a target quantity of machine-learning models. A natural choice to construct a model that yields transferable and linear-scaling predictions is to represent the scalar field using a multi-centered atomic basis analogous to that routinely used in density fitting approximations. However, the non-orthogonality of the basis poses challenges for the learning exercise, as it requires accounting for all the atomic density components at once. We devise a gradient-based approach to directly minimize the loss function of the regression problem in an optimized and highly sparse feature space. In so doing, we overcome the limitations associated with adopting an atom-centered model to learn the electron density over arbitrarily complex datasets, obtaining extremely accurate predictions. The enhanced framework is tested on 32-molecule periodic cells of liquid water, presenting enough complexity to require an optimal balance between accuracy and computational efficiency. We show that starting from the predicted density a single Kohn-Sham diagonalization step can be performed to access total energy components that carry an error of just 0.1 meV/atom with respect to the reference density functional calculations. Finally, we test our method on the highly heterogeneous QM9 benchmark dataset, showing that a small fraction of the training data is enough to derive ground-state total energies within chemical accuracy. | ['Michele Ceriotti', 'Mariana Rossi', 'Alan M. Lewis', 'Andrea Grisafi'] | 2022-06-28 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [ 2.58760244e-01 -3.34045216e-02 -2.03434631e-01 -2.05786094e-01
-9.83435631e-01 -6.39192984e-02 5.45510113e-01 3.79082769e-01
-6.84341669e-01 1.43327117e+00 -1.08838402e-01 -2.12858409e-01
9.83065516e-02 -8.66556227e-01 -6.68775737e-01 -1.25359643e+00
-4.56881747e-02 6.72120154e-01 -1.22939460e-02 -1.08504191e-01
4.51942921e-01 7.96414137e-01 -1.45926476e+00 7.34703094e-02
1.05064297e+00 9.52809334e-01 9.38699916e-02 2.33412445e-01
5.71438670e-02 6.26398981e-01 -1.82577476e-01 -2.44400546e-01
1.08395025e-01 -5.78629136e-01 -9.52004731e-01 -1.66232083e-02
2.79495865e-01 3.99406515e-02 -1.93045899e-01 9.94276583e-01
4.50079441e-01 4.52057749e-01 1.24506855e+00 -6.28266335e-01
-2.56907016e-01 1.27017051e-01 -3.21696579e-01 -1.66422993e-01
2.15648711e-01 1.28145188e-01 1.12205338e+00 -9.68431234e-01
7.23542094e-01 7.57941067e-01 7.09377825e-01 4.27303940e-01
-1.91982889e+00 -3.19737881e-01 -4.61615771e-01 1.64798096e-01
-1.67456400e+00 -3.07447672e-01 6.00601673e-01 -5.69428861e-01
1.31337380e+00 1.19920723e-01 7.37040639e-01 7.84567773e-01
6.19577825e-01 -1.06851809e-01 1.05499828e+00 -6.96165800e-01
6.36951745e-01 3.06300610e-01 -9.45091061e-03 5.25029004e-01
5.72210252e-01 9.19755828e-03 -4.65356857e-01 -5.19962192e-01
2.75733262e-01 -1.96101695e-01 -3.56179029e-01 -8.76542509e-01
-9.67561603e-01 1.14767957e+00 5.41832209e-01 2.29841590e-01
-6.04592502e-01 1.77285299e-01 3.53057623e-01 -2.19478533e-01
5.00717700e-01 7.23703444e-01 -3.74548256e-01 -7.33009577e-02
-1.02103019e+00 4.07223463e-01 9.54302490e-01 3.46922159e-01
1.15846670e+00 3.50239426e-02 3.01785618e-01 4.21433240e-01
1.11068770e-01 3.29682738e-01 2.03267157e-01 -8.94296587e-01
1.59356251e-01 2.64554054e-01 3.32510620e-01 -6.99336290e-01
-3.07366699e-01 -2.29092315e-01 -1.02740991e+00 3.78212810e-01
4.68630999e-01 -1.98072538e-01 -4.69878584e-01 1.55418956e+00
5.46969473e-01 -4.22320038e-01 3.58491540e-02 6.51165724e-01
2.09149793e-01 9.19365406e-01 1.28920764e-01 -7.39610136e-01
1.03925884e+00 -5.38968325e-01 -3.79217714e-01 6.12756126e-02
1.13177693e+00 -5.73131204e-01 8.00378144e-01 5.78956068e-01
-1.22500539e+00 -1.40059352e-01 -1.05419803e+00 -1.57967150e-01
-2.13485926e-01 -1.34049132e-01 8.66395891e-01 4.72866476e-01
-7.13135302e-01 1.39590764e+00 -8.09667408e-01 -4.09548581e-02
8.26077536e-02 7.21268535e-01 -8.20098042e-01 5.11454605e-02
-1.04325008e+00 1.34266257e+00 5.90473533e-01 -1.98274732e-01
-3.68672490e-01 -1.04413140e+00 -8.06594491e-01 1.79109380e-01
1.33237377e-01 -6.36577785e-01 1.12685204e+00 -7.18727350e-01
-1.46670401e+00 4.94003296e-01 -3.21239710e-01 -4.61252093e-01
1.50501668e-01 3.26011240e-01 -6.32724911e-02 2.10878566e-01
-4.92900908e-02 4.78479058e-01 5.42154849e-01 -1.05847096e+00
1.10317580e-02 -2.90080428e-01 -3.00282508e-01 1.05187520e-01
-3.15962553e-01 -3.14612269e-01 1.85234785e-01 -4.33560386e-02
1.28855392e-01 -1.05506003e+00 -5.64297080e-01 -2.68033475e-01
-3.43086481e-01 8.04781020e-02 1.91873714e-01 -4.89047498e-01
1.11295331e+00 -1.72024465e+00 4.84771907e-01 4.82181311e-01
1.56787127e-01 2.19706502e-02 3.53151709e-01 8.86370838e-01
-3.68651748e-01 -1.02653634e-02 -5.85176766e-01 -3.88744354e-01
-1.63966700e-01 1.11288689e-02 -3.37767787e-03 7.05244958e-01
1.71278968e-01 5.74885249e-01 -5.51813483e-01 -3.77660930e-01
1.97070614e-01 9.48760867e-01 -1.10328913e+00 3.69049311e-02
-1.58214509e-01 4.16999817e-01 -3.68862271e-01 1.77251101e-01
5.84154606e-01 -4.90950733e-01 4.65806872e-01 -4.87279147e-01
-4.26740408e-01 6.67291701e-01 -1.15470564e+00 1.59459937e+00
-4.21362013e-01 8.78932849e-02 -5.52312620e-02 -1.19444132e+00
7.92673409e-01 2.40643740e-01 9.29071844e-01 -4.40757126e-01
-3.79690900e-02 6.03947222e-01 4.38406229e-01 3.81689630e-02
4.09514844e-01 -9.08846617e-01 6.21554330e-02 2.01805696e-01
3.43356311e-01 -5.76128185e-01 2.62944609e-01 2.04638317e-02
8.64932656e-01 -7.79437795e-02 7.60921240e-01 -8.72399449e-01
6.71617448e-01 2.04034314e-01 2.03770995e-01 3.35706085e-01
3.24109137e-01 5.31978250e-01 6.91982985e-01 -4.73036796e-01
-1.33956742e+00 -6.30937815e-01 -7.81613767e-01 4.83001858e-01
-3.20201606e-01 -6.62882626e-01 -8.79924893e-01 -1.26926973e-01
1.50610685e-01 6.74891710e-01 -6.07473910e-01 -5.01874149e-01
-3.00120950e-01 -1.26169074e+00 -1.22354247e-01 -7.56843667e-03
-4.48951032e-03 -9.30566192e-01 -3.04289550e-01 3.13264847e-01
2.68951446e-01 -6.61042571e-01 -1.99519530e-01 7.70325422e-01
-7.85634458e-01 -8.08961213e-01 -5.35575032e-01 -1.61718667e-01
5.71978033e-01 -1.04141623e-01 1.00693882e+00 -1.79554690e-02
-4.70161438e-01 9.27293226e-02 1.41142145e-01 5.18046059e-02
-5.11899650e-01 3.13451201e-01 3.35818321e-01 -1.57324761e-01
3.67193788e-01 -7.42396593e-01 -6.44001603e-01 -2.76674002e-01
-5.55998027e-01 -1.07268646e-01 3.87864530e-01 9.77846682e-01
6.62837625e-01 -1.20919086e-01 5.09410381e-01 -7.84618497e-01
2.77288824e-01 -6.09687924e-01 -8.03649664e-01 -1.01754636e-01
-7.84438252e-01 4.55597728e-01 9.13255394e-01 -1.17420681e-01
-8.12840223e-01 3.67790520e-01 -4.48725909e-01 1.23005100e-01
9.06817913e-02 5.86143196e-01 -1.25949562e-01 -5.42610228e-01
7.96237230e-01 2.13944718e-01 8.79286304e-02 -5.48664749e-01
1.06864668e-01 5.21057129e-01 2.39152297e-01 -8.81639063e-01
5.07499993e-01 2.24091679e-01 7.82540560e-01 -1.03378546e+00
-7.35116184e-01 -3.47813964e-01 -8.41469228e-01 2.80543476e-01
8.45512867e-01 -7.58937716e-01 -1.19432580e+00 -5.22409566e-02
-1.06737173e+00 -7.02471733e-02 -5.38438082e-01 7.50343204e-01
-7.05581248e-01 5.12732625e-01 -5.36310732e-01 -8.34901929e-01
-4.08811629e-01 -1.41483486e+00 9.72865939e-01 -2.26382107e-01
-3.66350502e-01 -1.10377598e+00 3.61231834e-01 1.51488125e-01
4.01337236e-01 1.75679326e-01 1.24213827e+00 -4.73461896e-01
-4.25159633e-01 -2.41712570e-01 8.37012455e-02 2.61923254e-01
1.78794339e-01 9.06969979e-03 -8.71932626e-01 -3.78657848e-01
1.60519704e-01 -4.95380521e-01 9.11443293e-01 3.16279441e-01
1.16484201e+00 -1.42490208e-01 -2.73798168e-01 5.34917891e-01
1.52622914e+00 1.16188541e-01 5.00699639e-01 6.45443201e-02
7.19173253e-01 4.26346183e-01 1.99809015e-01 5.79659343e-01
9.70621258e-02 8.49442542e-01 9.38367099e-02 7.32463524e-02
3.59239280e-01 -1.43669277e-01 1.82060063e-01 6.98306322e-01
-3.04179788e-01 2.60050911e-02 -1.02010131e+00 1.38202518e-01
-1.40705466e+00 -1.13380778e+00 -2.68161178e-01 2.58628702e+00
1.18680501e+00 7.11661577e-02 1.49783641e-01 3.85088027e-01
1.06608994e-01 -1.26617193e-01 -5.39894462e-01 -4.75787163e-01
2.34541193e-01 5.14921606e-01 4.40190703e-01 8.66783917e-01
-7.95237005e-01 6.08872175e-01 7.00537109e+00 8.02693605e-01
-1.33341682e+00 -1.55551732e-01 7.53950953e-01 -1.85989708e-01
-2.31929719e-01 1.83604807e-01 -9.66367781e-01 5.97869635e-01
1.37773693e+00 -3.04106236e-01 4.71370161e-01 9.68905449e-01
2.19718680e-01 -3.72407883e-01 -1.10354495e+00 8.75199080e-01
-3.05593431e-01 -1.61155879e+00 1.53488576e-01 4.74387944e-01
7.45515049e-01 -4.20446880e-02 -1.58706561e-01 1.70696199e-01
-3.20137680e-01 -1.30162942e+00 3.56910348e-01 4.11828071e-01
9.57303822e-01 -8.85814667e-01 4.44530249e-01 4.14179713e-01
-7.82377839e-01 3.04734021e-01 -6.00378454e-01 -3.03654820e-01
1.38447896e-01 7.67817974e-01 -6.72689080e-01 3.53168011e-01
1.75131321e-01 4.23917770e-01 -1.25177294e-01 8.46442878e-01
5.33916414e-01 3.73057812e-01 -4.86728191e-01 -8.33444148e-02
3.18830013e-01 -9.49753225e-01 6.68736994e-02 9.49645758e-01
6.31696343e-01 9.18079764e-02 3.20115350e-02 8.67315710e-01
-1.96818441e-01 4.61265802e-01 -6.91191912e-01 -1.07070506e-01
7.25240037e-02 1.31674147e+00 -4.36296344e-01 -1.40480623e-01
-4.80594754e-01 7.59688616e-01 6.52866125e-01 1.87180951e-01
-4.94296402e-01 -9.75551307e-02 7.24790633e-01 4.93287921e-01
3.98472995e-01 -3.29963446e-01 1.46822765e-01 -1.20884132e+00
-1.10669136e-01 -7.35347033e-01 -9.18887854e-02 -4.02278721e-01
-1.16747034e+00 4.27636057e-01 1.96188405e-01 -7.16426313e-01
-5.38940072e-01 -1.06797051e+00 -4.79403496e-01 1.24601424e+00
-1.45337605e+00 -4.88606930e-01 1.38650864e-01 2.85832882e-01
-2.03540772e-01 8.47236291e-02 1.15914869e+00 3.80481720e-01
-5.89973390e-01 1.58267766e-01 4.55657452e-01 -5.79815447e-01
4.55793142e-01 -1.44987166e+00 -2.00365596e-02 2.94785589e-01
-8.14037174e-02 8.33673894e-01 9.98733997e-01 -4.94421422e-01
-1.71571386e+00 -7.57191658e-01 1.00437808e+00 -1.92922384e-01
7.40305424e-01 -4.80383962e-01 -1.24386132e+00 4.00819659e-01
-2.10237168e-02 4.27615285e-01 8.63864779e-01 6.08177781e-02
-3.86687112e-03 -4.91414918e-03 -1.19277465e+00 3.83423239e-01
7.03335643e-01 -5.71346402e-01 -1.95094019e-01 7.32940376e-01
3.38885993e-01 -1.81712911e-01 -1.17171454e+00 3.73227268e-01
2.64633119e-01 -1.16218209e+00 9.66551185e-01 -7.23462522e-01
3.22243512e-01 -1.51753658e-02 -2.83072859e-01 -1.26145458e+00
-3.20819318e-01 -7.11808801e-01 -1.28931016e-01 7.53356099e-01
4.36642379e-01 -6.57383144e-01 9.07794714e-01 1.01356626e+00
-4.33534496e-02 -1.13243747e+00 -1.15018821e+00 -4.96491402e-01
6.39438033e-01 -1.04626611e-01 2.09098741e-01 7.03415751e-01
4.44184750e-01 5.70127189e-01 -4.12241012e-01 -2.06774190e-01
5.21653175e-01 6.83406889e-02 4.41662908e-01 -1.41229343e+00
-4.23057318e-01 -1.55528113e-01 -3.29592675e-01 -8.02843690e-01
4.84912187e-01 -1.05893266e+00 -1.88216954e-01 -1.02684021e+00
4.56703544e-01 -4.05977458e-01 -7.10252523e-02 1.82379335e-01
1.30780414e-01 3.20734501e-01 -3.80582623e-02 5.56936488e-02
-2.70233840e-01 9.50416744e-01 1.09405231e+00 1.03073597e-01
-7.99527094e-02 -7.27968067e-02 -5.11072457e-01 7.20442593e-01
6.76365972e-01 -5.45540094e-01 -2.09995344e-01 4.01515961e-01
3.98576051e-01 2.06192732e-01 2.37583995e-01 -1.10161400e+00
9.31646153e-02 -2.04517186e-01 3.19693714e-01 -2.54162282e-01
8.25171530e-01 -8.28790665e-01 3.53739321e-01 5.80822408e-01
-1.33795097e-01 -2.24301100e-01 1.57929659e-01 1.14082262e-01
-1.99409083e-01 -7.78435707e-01 1.01170826e+00 -1.69206098e-01
-4.46500145e-02 4.00159985e-01 -3.74015331e-01 -2.34454095e-01
8.55395138e-01 6.56477213e-02 1.80220485e-01 -1.10722534e-01
-5.52559495e-01 -4.63584602e-01 8.72781813e-01 -5.10962725e-01
2.19087750e-01 -1.33464956e+00 -3.34816217e-01 1.59582049e-01
-5.91251664e-02 -1.97264880e-01 2.17527688e-01 8.32577884e-01
-7.95921743e-01 7.52735972e-01 -8.05379450e-02 -4.16705251e-01
-8.70884776e-01 6.39480770e-01 7.44220853e-01 -3.69052142e-01
-4.25823689e-01 2.83423960e-01 9.42424014e-02 -3.33125502e-01
-3.23288649e-01 -1.72096401e-01 2.54187167e-01 7.60028288e-02
4.16423649e-01 2.76091933e-01 4.39187527e-01 -9.15721893e-01
-3.50475788e-01 5.44751823e-01 -1.19041748e-01 7.55016133e-02
1.41778445e+00 4.03431982e-01 -1.99400693e-01 3.87181878e-01
1.51403916e+00 1.03643343e-01 -1.24861562e+00 5.42590171e-02
-6.28995523e-02 -1.84529975e-01 3.43814313e-01 -2.38117725e-01
-4.76668805e-01 9.40023124e-01 2.34867096e-01 1.66563898e-01
5.04530609e-01 -3.24246943e-01 4.79871035e-01 8.95745933e-01
4.37710971e-01 -7.26065636e-01 -2.08114102e-01 3.17085922e-01
6.59655750e-01 -1.40258396e+00 3.79130214e-01 -3.69004756e-01
-5.16715169e-01 1.14209890e+00 2.82191008e-01 -7.23048225e-02
5.51350772e-01 1.20984197e-01 -6.64666891e-01 -1.71371222e-01
-6.08005047e-01 9.17928368e-02 4.70190972e-01 2.77902484e-01
8.31742346e-01 -2.02512264e-01 -3.37155342e-01 1.61208525e-01
-1.73799649e-01 -1.50673628e-01 5.70204020e-01 7.14497983e-01
-4.99516010e-01 -1.45693219e+00 -1.15922078e-01 4.75343734e-01
-4.75980312e-01 -3.74701709e-01 -6.34864196e-02 9.33702528e-01
-1.59677356e-01 5.26778936e-01 -1.31892398e-01 1.19109452e-01
5.56291454e-02 4.38196719e-01 6.59920573e-01 -7.89439440e-01
-3.09438080e-01 -1.25776306e-01 -7.38480687e-02 -3.96803021e-01
-3.74401659e-01 -6.45841718e-01 -1.42262983e+00 -6.24284804e-01
-3.57594043e-01 7.83419430e-01 6.06976807e-01 9.79491949e-01
2.44932234e-01 5.16700074e-02 4.80612546e-01 -1.43412757e+00
-9.19328868e-01 -8.01146626e-01 -8.80448163e-01 3.85682255e-01
3.68403226e-01 -7.93409050e-01 -6.06349051e-01 -3.69835377e-01] | [5.303282260894775, 5.219146251678467] |
7552c22a-5f4e-4db8-b1b4-1971a5e5358d | evaluation-of-the-potential-of-near-infrared | 2301.08252 | null | https://arxiv.org/abs/2301.08252v1 | https://arxiv.org/pdf/2301.08252v1.pdf | Evaluation of the potential of Near Infrared Hyperspectral Imaging for monitoring the invasive brown marmorated stink bug | The brown marmorated stink bug (BMSB), Halyomorpha halys, is an invasive insect pest of global importance that damages several crops, compromising agri-food production. Field monitoring procedures are fundamental to perform risk assessment operations, in order to promptly face crop infestations and avoid economical losses. To improve pest management, spectral cameras mounted on Unmanned Aerial Vehicles (UAVs) and other Internet of Things (IoT) devices, such as smart traps or unmanned ground vehicles, could be used as an innovative technology allowing fast, efficient and real-time monitoring of insect infestations. The present study consists in a preliminary evaluation at the laboratory level of Near Infrared Hyperspectral Imaging (NIR-HSI) as a possible technology to detect BMSB specimens on different vegetal backgrounds, overcoming the problem of BMSB mimicry. Hyperspectral images of BMSB were acquired in the 980-1660 nm range, considering different vegetal backgrounds selected to mimic a real field application scene. Classification models were obtained following two different chemometric approaches. The first approach was focused on modelling spectral information and selecting relevant spectral regions for discrimination by means of sparse-based variable selection coupled with Soft Partial Least Squares Discriminant Analysis (s-Soft PLS-DA) classification algorithm. The second approach was based on modelling spatial and spectral features contained in the hyperspectral images using Convolutional Neural Networks (CNN). Finally, to further improve BMSB detection ability, the two strategies were merged, considering only the spectral regions selected by s-Soft PLS-DA for CNN modelling. | ['Alessandro Ulrici', 'Peter Offermans', 'Lara Maistrello', 'Aravind Krishnaswamy Rangarajan', 'Camilla Menozzi', 'Bas Boom', 'Rosalba Calvini', 'Veronica Ferrari'] | 2023-01-19 | null | null | null | null | ['scene-classification', 'variable-selection'] | ['computer-vision', 'methodology'] | [ 6.75262034e-01 -5.35322666e-01 -8.67492706e-02 1.74213365e-01
8.51752535e-02 -8.19985271e-01 9.98385549e-02 2.16973305e-01
-2.83171058e-01 7.56893873e-01 -5.53354681e-01 -4.22186404e-01
-5.60222983e-01 -1.12475204e+00 -3.02848965e-01 -1.01406634e+00
-2.09884733e-01 2.09376514e-01 2.28825007e-02 -5.42919040e-01
-1.46637574e-01 1.08395886e+00 -1.79947996e+00 2.30888531e-01
9.35155332e-01 1.02535129e+00 8.92358720e-01 3.29500616e-01
2.32170463e-01 -1.65953994e-01 -4.18117225e-01 8.71666223e-02
5.01931429e-01 -5.85156083e-02 -2.83856511e-01 3.34418446e-01
-1.70831475e-02 -2.45831758e-01 3.31429720e-01 1.50572443e+00
1.33589655e-01 -2.75295936e-02 6.35918081e-01 -1.03349757e+00
-3.46621007e-01 5.07950485e-01 -8.96688521e-01 -5.75766340e-03
2.41354242e-01 4.68530774e-01 5.44830143e-01 -3.86457622e-01
2.81431824e-01 1.00763130e+00 5.15680194e-01 1.05946849e-03
-1.52991688e+00 -5.46690643e-01 -7.72791728e-02 4.81424689e-01
-1.44540906e+00 8.77596512e-02 6.38647437e-01 -5.73859155e-01
8.04226041e-01 3.87128145e-01 1.19975436e+00 7.88424432e-01
3.92498463e-01 3.01343381e-01 1.21205473e+00 -2.05635503e-01
2.77039230e-01 6.09785132e-02 2.33403333e-02 4.68121737e-01
8.05046141e-01 4.45867479e-01 -1.76240243e-02 -2.35662207e-01
4.87701416e-01 2.10997954e-01 -5.99816084e-01 -4.68558341e-01
-8.71622801e-01 8.09710205e-01 9.31259453e-01 4.93102700e-01
-9.19882357e-01 -5.56253970e-01 6.63890615e-02 1.52952060e-01
3.03687483e-01 5.73588073e-01 -7.68043518e-01 5.14226735e-01
-9.05899882e-01 -1.72014818e-01 6.19169593e-01 3.05134922e-01
1.06810296e+00 2.03282624e-01 2.83318907e-01 8.12472820e-01
4.80463475e-01 1.06658995e+00 4.04570967e-01 -4.42498684e-01
-1.83025926e-01 7.35723734e-01 -1.46147758e-02 -1.35438550e+00
-5.42266488e-01 -3.03509653e-01 -8.86239290e-01 2.93934137e-01
1.08393602e-01 -1.67848855e-01 -6.99998379e-01 1.30518687e+00
2.20888868e-01 -1.91674486e-01 9.32251662e-02 1.16864824e+00
5.08054078e-01 9.78964090e-01 1.22472994e-01 -4.60304409e-01
1.56700218e+00 -1.47962153e-01 -3.99448216e-01 -6.75662905e-02
5.22599220e-01 -6.52602315e-01 7.54917443e-01 7.77434170e-01
-3.74148875e-01 -3.66773486e-01 -1.19389820e+00 6.60220921e-01
-9.69041348e-01 5.97967446e-01 6.40737712e-01 6.74768806e-01
-8.28974009e-01 6.89513445e-01 -7.57490754e-01 -8.13955128e-01
3.27955246e-01 5.26779592e-01 -6.93252504e-01 -9.84973311e-02
-8.49964440e-01 8.67735207e-01 7.29124784e-01 6.86300695e-01
-7.94755518e-01 -1.74755156e-01 -7.73658037e-01 2.34353885e-01
3.69728714e-01 -2.07722649e-01 4.08993423e-01 -1.27964544e+00
-1.58920455e+00 9.53752041e-01 1.14041179e-01 -3.60059977e-01
-2.61041015e-01 -1.45062402e-01 -3.42387557e-01 2.65890986e-01
7.94647038e-02 4.28225815e-01 6.87198579e-01 -1.26593387e+00
-3.84554148e-01 -9.97702718e-01 -2.62879699e-01 7.23853568e-03
-4.49708760e-01 -9.39368270e-03 6.29028141e-01 -3.43635142e-01
3.81956130e-01 -1.02382827e+00 -2.52465695e-01 -5.49897738e-02
-4.39449102e-01 4.38047767e-01 1.15275717e+00 -7.60833144e-01
4.59888965e-01 -2.09931540e+00 4.12717521e-01 4.45566803e-01
-4.33260322e-01 9.68933284e-01 -4.34363753e-01 2.84241915e-01
-2.36302331e-01 -1.62791222e-01 -4.73889619e-01 4.84027117e-01
-4.83752638e-01 3.54521237e-02 6.73076212e-02 5.30575037e-01
3.27680498e-01 3.45266134e-01 -7.71821201e-01 4.14379239e-02
7.46268332e-01 6.00797832e-01 1.27704352e-01 -1.41268387e-01
-3.13315928e-01 3.27888489e-01 -2.55106747e-01 1.12927151e+00
1.17980039e+00 3.10056806e-01 2.20948994e-01 -5.91513038e-01
-5.53143620e-01 -6.39209032e-01 -1.01896906e+00 1.19160354e+00
-2.56275415e-01 3.91214401e-01 7.12687254e-01 -1.22557461e+00
1.13251293e+00 2.44507119e-01 5.47240257e-01 -2.69981563e-01
2.86789030e-01 3.42167675e-01 -2.11502656e-01 -5.45709491e-01
2.72136807e-01 3.67644615e-02 6.65962815e-01 -3.50443393e-01
8.15492645e-02 -2.12319762e-01 1.02232121e-01 -5.47128260e-01
4.14000779e-01 1.87378690e-01 7.00392723e-01 -5.77137232e-01
7.28946149e-01 5.48284352e-01 3.55157077e-01 2.10651904e-02
-2.38124847e-01 -9.20834243e-02 1.17242090e-01 -5.57577968e-01
-2.47814894e-01 -7.10902154e-01 -3.65146220e-01 7.33644664e-01
1.69407412e-01 3.82600725e-01 -5.09174585e-01 -8.31272081e-02
5.58527336e-02 7.59530008e-01 -1.74595460e-01 -2.45712757e-01
1.58797443e-01 -1.46230733e+00 1.24809891e-01 -1.04082935e-01
6.92902863e-01 -1.17666352e+00 -9.19464886e-01 3.23884666e-01
1.38066232e-01 -8.10480535e-01 6.08645618e-01 6.98576927e-01
-1.05370939e+00 -1.24309826e+00 -6.30150437e-01 -4.56240207e-01
3.61874342e-01 8.44764173e-01 5.50855517e-01 -3.69360775e-01
-8.48126590e-01 1.53691202e-01 -5.59112608e-01 -4.82143134e-01
-2.25679547e-01 -1.28410727e-01 -1.32641613e-01 1.21561408e-01
8.57105196e-01 -5.24451971e-01 -4.27632064e-01 3.03375244e-01
-9.96385574e-01 -2.47025311e-01 8.67595196e-01 8.87682617e-01
7.24975228e-01 4.98724401e-01 -1.35932807e-02 -3.96429360e-01
1.70929492e-01 -4.53971863e-01 -1.16096604e+00 3.86876196e-01
-5.55243045e-02 -5.91947556e-01 5.95031679e-01 -3.35704565e-01
-7.62695432e-01 5.48379660e-01 2.74406433e-01 -1.57001898e-01
-8.39901328e-01 1.04976761e+00 -5.20549536e-01 -3.97373110e-01
7.94815719e-01 1.98519647e-01 1.13883063e-01 -1.73218206e-01
-1.38243005e-01 7.62145400e-01 2.64402151e-01 -4.71245870e-03
6.52788341e-01 4.40027297e-01 3.76592875e-01 -1.71750665e+00
-4.81580764e-01 -6.19544923e-01 -6.74861133e-01 -2.54576594e-01
1.15387106e+00 -8.90988886e-01 -7.61558950e-01 7.12737262e-01
-1.02245724e+00 -2.50418514e-01 1.19358875e-01 8.21343362e-01
-2.18887046e-01 3.37245911e-01 -4.06772107e-01 -8.75534832e-01
-4.14191812e-01 -1.23129940e+00 9.22349393e-01 3.53889436e-01
1.77345678e-01 -5.94062924e-01 -9.49323177e-02 2.59338289e-01
3.17925155e-01 5.42088807e-01 8.84725332e-01 -2.94283539e-01
-2.84181356e-01 -2.49055490e-01 -4.50409263e-01 5.21812379e-01
5.03789961e-01 5.46836317e-01 -9.59341168e-01 -2.86168516e-01
1.56622037e-01 -7.30674788e-02 8.10562909e-01 8.93452883e-01
6.81445539e-01 2.61387322e-02 -3.97996485e-01 7.72797108e-01
1.90346324e+00 6.02613330e-01 4.27074552e-01 3.87344122e-01
4.14100260e-01 9.11507845e-01 9.92081761e-01 5.47603190e-01
-5.05052507e-01 5.18810689e-01 1.07460475e+00 -3.16358328e-01
3.96219760e-01 3.42814267e-01 5.69928229e-01 1.97396968e-02
-2.99304008e-01 -3.01700950e-01 -7.87550926e-01 2.94883043e-01
-1.47579348e+00 -9.75308478e-01 -4.14879650e-01 2.43578506e+00
1.81279853e-01 -3.71111721e-01 2.85709579e-03 2.73916274e-01
8.44102859e-01 -1.09751068e-01 -3.52916420e-01 -2.62719601e-01
-5.87376118e-01 2.94970274e-01 9.44049001e-01 2.21676573e-01
-1.37931454e+00 9.00060475e-01 4.61421728e+00 3.87587041e-01
-1.65103209e+00 -2.74975866e-01 3.10973197e-01 1.88922450e-01
3.15270424e-01 2.65834499e-02 -5.83986044e-01 1.97334439e-01
5.72142959e-01 4.13533419e-01 8.66644979e-01 7.09200025e-01
4.10822928e-01 -4.51843441e-01 -3.16432267e-01 7.74213374e-01
-1.94271371e-01 -6.80211902e-01 6.00789078e-02 2.53326654e-01
4.73789811e-01 1.09138295e-01 -3.13191861e-02 -2.74244815e-01
1.35126412e-01 -5.69597065e-01 1.20526008e-01 2.85157025e-01
5.58453321e-01 -6.21588051e-01 1.05466115e+00 2.69265890e-01
-1.22679162e+00 -3.72414112e-01 -8.18098128e-01 -7.02921227e-02
-3.93834971e-02 7.72935569e-01 -5.28000414e-01 6.90608859e-01
8.83751512e-01 9.29419875e-01 -5.25244951e-01 9.82689500e-01
-9.06208232e-02 5.07397056e-01 -4.19058472e-01 -8.97600129e-02
2.98063934e-01 -7.58643687e-01 7.99291313e-01 8.71312201e-01
7.62234211e-01 1.44058079e-01 2.23374784e-01 1.05752468e+00
7.39090800e-01 2.86972761e-01 -9.16962147e-01 -3.23808193e-01
9.75431576e-02 1.42691135e+00 -1.10700667e+00 4.42663506e-02
-1.90330684e-01 9.09281015e-01 -5.74436128e-01 3.63432139e-01
-5.05559266e-01 -2.41703585e-01 5.71312368e-01 -8.72617364e-02
2.84040689e-01 -2.28071973e-01 2.64484197e-01 -1.05688620e+00
-5.01777947e-01 -9.33808923e-01 1.78317010e-01 -9.41485345e-01
-1.21326315e+00 3.37519169e-01 1.62216067e-01 -1.12194979e+00
3.69882464e-01 -1.23235106e+00 -2.54920810e-01 1.04750228e+00
-1.54850781e+00 -1.22790873e+00 -6.96626961e-01 5.41434228e-01
2.08344221e-01 -1.87748045e-01 1.31678653e+00 -2.35927343e-01
-7.34693944e-01 -2.83068001e-01 2.91758448e-01 -4.87512499e-01
3.45511347e-01 -8.58794391e-01 -6.98756874e-01 6.52710438e-01
-1.20896168e-01 1.47734642e-01 5.97061753e-01 -7.95086503e-01
-1.36693132e+00 -1.04170918e+00 3.16536486e-01 7.39264548e-01
5.72412133e-01 3.18981230e-01 -6.63177252e-01 4.51074570e-01
1.77124925e-02 -3.93289119e-01 8.78174126e-01 -2.20131591e-01
1.22352697e-01 -4.13967580e-01 -1.51368642e+00 3.59551430e-01
1.79173589e-01 -1.64600089e-01 -6.71844855e-02 6.50462866e-01
2.57806510e-01 2.67755389e-01 -7.92795777e-01 6.72108591e-01
5.61433494e-01 -9.37385798e-01 1.12414241e+00 -1.50356337e-01
1.62233412e-01 -4.92954284e-01 -5.33140779e-01 -1.45552385e+00
-5.28584063e-01 -2.58430809e-01 5.85463643e-01 8.08889806e-01
7.04550147e-02 -7.33556688e-01 5.52743673e-01 -1.76304460e-01
8.50691181e-03 -1.99859142e-02 -6.76068604e-01 -8.12988102e-01
-5.28122187e-01 -1.82036348e-02 4.87688035e-01 7.26921678e-01
-2.31732801e-01 -8.68901536e-02 -1.28820166e-01 8.79019022e-01
5.78052342e-01 6.79090479e-03 3.83585721e-01 -1.75744629e+00
-1.96112141e-01 -4.88711268e-01 -7.32419968e-01 -1.34777695e-01
1.40472412e-01 -6.39544010e-01 -1.09005980e-01 -1.19166768e+00
-1.09429136e-01 8.36547986e-02 -1.94446728e-01 6.69975817e-01
7.51300082e-02 3.81277591e-01 1.29623026e-01 7.91324452e-02
4.81537819e-01 2.86001414e-01 9.82875705e-01 -4.16250736e-01
-6.47099018e-01 3.61009568e-01 -3.10044676e-01 7.69706607e-01
7.75223970e-01 -2.68319458e-01 -4.21768486e-01 -2.71395028e-01
-2.34766696e-02 2.55175173e-01 8.31894994e-01 -1.17619503e+00
-2.39755526e-01 -4.38052684e-01 3.12478691e-01 -7.31155097e-01
5.43335021e-01 -1.49949598e+00 5.49616277e-01 1.09118176e+00
1.94735751e-01 -3.40350747e-01 7.29595363e-01 2.79931605e-01
-2.19748676e-01 -6.13382518e-01 8.70383084e-01 -2.70754546e-01
-7.26888955e-01 6.15307316e-02 -1.01165748e+00 -1.09359586e+00
1.42301059e+00 -5.69873691e-01 -2.66026944e-01 8.39815885e-02
-7.48254955e-01 -1.94276989e-01 4.81845021e-01 -2.34514058e-01
4.72485393e-01 -6.75195098e-01 -5.30016840e-01 5.30909419e-01
1.16009332e-01 -3.18194389e-01 3.98581743e-01 7.88067341e-01
-1.24755394e+00 4.51541960e-01 -8.59732568e-01 -6.64751410e-01
-1.47870588e+00 7.64282823e-01 2.54695207e-01 -6.69474304e-02
9.76166576e-02 7.61325896e-01 -1.24170542e-01 -3.38659763e-01
-9.68302265e-02 -3.50822449e-01 -7.16307402e-01 3.55515987e-01
3.47387671e-01 3.32205087e-01 1.27511755e-01 -9.37735677e-01
-3.19682240e-01 5.90496778e-01 5.73503673e-01 2.81942248e-01
1.61208165e+00 1.02803379e-01 -6.13473356e-01 2.31463253e-01
7.04550385e-01 -3.34386945e-01 -8.49569619e-01 2.41259158e-01
-4.34687585e-02 -3.27499539e-01 4.10097331e-01 -1.02217388e+00
-1.13527679e+00 9.95137513e-01 9.72334087e-01 6.03987038e-01
1.69375014e+00 -6.12668991e-01 1.65671125e-01 6.71143174e-01
6.30315542e-01 -6.48381948e-01 -5.58523715e-01 1.00651793e-01
7.81595469e-01 -1.30029750e+00 9.42318141e-02 -6.88013434e-01
-4.22435075e-01 1.56455147e+00 3.12819183e-01 3.91816609e-02
6.67591035e-01 -6.58200085e-02 -2.44968012e-02 -3.24386746e-01
-1.63138956e-01 -3.54540557e-01 1.28235640e-02 1.04069602e+00
2.91118622e-01 5.35780728e-01 -1.28008887e-01 2.18684554e-01
8.30789730e-02 6.78774342e-02 2.96929598e-01 7.99637914e-01
-6.08226478e-01 -7.96960950e-01 -8.87648880e-01 4.95825619e-01
-4.36619557e-02 -4.83085997e-02 -5.54575622e-01 6.76034033e-01
3.02683383e-01 1.07794225e+00 -2.98742861e-01 -4.51135755e-01
2.61036694e-01 -2.33780771e-01 3.21126491e-01 -6.12882793e-01
-4.72344249e-01 3.74928981e-01 -4.79905844e-01 -4.46701229e-01
-9.09755468e-01 -7.95539439e-01 -3.83514404e-01 -2.75591582e-01
-8.47262025e-01 -8.56623650e-02 1.16062701e+00 7.61021078e-01
9.03545972e-03 1.43627077e-01 7.20660329e-01 -1.25365007e+00
-3.84218723e-01 -1.00405717e+00 -1.30809319e+00 -1.96599007e-01
5.22627775e-03 -6.99113131e-01 -3.99260998e-01 1.14014708e-02] | [9.509947776794434, -1.665880799293518] |
f99db249-1b79-4094-bc3a-56145deed006 | fedora-flying-event-dataset-for-reactive | 2305.14392 | null | https://arxiv.org/abs/2305.14392v1 | https://arxiv.org/pdf/2305.14392v1.pdf | FEDORA: Flying Event Dataset fOr Reactive behAvior | The ability of living organisms to perform complex high speed manoeuvers in flight with a very small number of neurons and an incredibly low failure rate highlights the efficacy of these resource-constrained biological systems. Event-driven hardware has emerged, in recent years, as a promising avenue for implementing complex vision tasks in resource-constrained environments. Vision-based autonomous navigation and obstacle avoidance consists of several independent but related tasks such as optical flow estimation, depth estimation, Simultaneous Localization and Mapping (SLAM), object detection, and recognition. To ensure coherence between these tasks, it is imperative that they be trained on a single dataset. However, most existing datasets provide only a selected subset of the required data. This makes inter-network coherence difficult to achieve. Another limitation of existing datasets is the limited temporal resolution they provide. To address these limitations, we present FEDORA, a first-of-its-kind fully synthetic dataset for vision-based tasks, with ground truths for depth, pose, ego-motion, and optical flow. FEDORA is the first dataset to provide optical flow at three different frequencies - 10Hz, 25Hz, and 50Hz | ['Kaushik Roy', 'Manish Nagaraj', 'Wachirawit Ponghiran', 'Adarsh Kosta', 'Amogh Joshi'] | 2023-05-22 | null | null | null | null | ['simultaneous-localization-and-mapping', 'autonomous-navigation'] | ['computer-vision', 'computer-vision'] | [ 1.67983919e-02 -4.94668752e-01 1.99569285e-01 -1.46761134e-01
4.96526770e-02 -6.24519527e-01 6.87998056e-01 -3.06015998e-01
-7.31143534e-01 8.57097387e-01 -1.42203376e-01 9.46991146e-02
4.75861598e-03 -4.03445840e-01 -4.81560141e-01 -4.07746524e-01
-3.53798509e-01 1.64268494e-01 5.68814218e-01 -2.19985411e-01
2.92432278e-01 6.04785502e-01 -2.05369878e+00 -1.81452557e-01
4.08538818e-01 9.10340786e-01 5.14923871e-01 1.06543195e+00
3.99972171e-01 7.79231012e-01 -4.85129595e-01 3.69944096e-01
4.61583853e-01 -4.53858584e-01 -4.91377980e-01 -1.96775332e-01
7.35109091e-01 -5.47877073e-01 -6.44043207e-01 5.81266880e-01
6.16238117e-01 2.61857837e-01 3.10672343e-01 -1.21427250e+00
-2.05313921e-01 -3.97497386e-01 -1.51260182e-01 5.50808907e-01
3.09515476e-01 7.10307539e-01 6.33990467e-01 -9.03125226e-01
8.27979088e-01 1.03629446e+00 5.16719997e-01 7.30370402e-01
-1.17476034e+00 -3.81551176e-01 -2.34447062e-01 9.92776453e-02
-1.19505167e+00 -7.90259063e-01 3.27133417e-01 -7.61954427e-01
1.36910546e+00 -2.19840467e-01 1.02437413e+00 1.02159011e+00
5.41378975e-01 1.09033376e-01 8.71423960e-01 -4.18012552e-02
5.14208972e-01 -3.54189783e-01 -2.65796185e-01 8.42887104e-01
4.73638564e-01 4.59614545e-01 -1.23586667e+00 1.36459574e-01
1.09942293e+00 -2.93164223e-01 -5.46459973e-01 -4.53895271e-01
-1.77257895e+00 5.23770511e-01 5.72055280e-01 -1.78512931e-01
-3.31325769e-01 7.05723584e-01 1.63241357e-01 3.84902000e-01
-3.00885774e-02 7.68446207e-01 -5.24455488e-01 -4.63987231e-01
-7.12562382e-01 3.19899321e-01 8.72990608e-01 8.31293106e-01
8.07764888e-01 4.35207099e-01 3.83345544e-01 1.50993899e-01
3.91406983e-01 7.48205960e-01 5.54822862e-01 -1.44713104e+00
6.68804273e-02 3.95703502e-02 4.00597751e-01 -1.03267026e+00
-6.64822280e-01 -3.14814508e-01 -5.87239206e-01 7.17653871e-01
6.45396829e-01 -3.44645619e-01 -7.61630833e-01 1.88243842e+00
2.92707950e-01 3.21447194e-01 2.01103657e-01 1.30317855e+00
5.27412236e-01 3.82105142e-01 -3.31159681e-01 -8.16457570e-02
1.01462102e+00 -6.92333758e-01 -4.50703084e-01 -8.23073208e-01
3.16206157e-01 -6.09780848e-01 9.13978517e-01 4.59417224e-01
-7.19214022e-01 -4.75298852e-01 -1.28970218e+00 4.43733484e-02
-2.57948548e-01 -1.64719313e-01 1.07874167e+00 3.66980582e-01
-1.19543302e+00 6.40291631e-01 -1.30064797e+00 -6.99997067e-01
3.53265345e-01 3.13776284e-01 -6.82910025e-01 3.45168561e-02
-8.36464107e-01 1.01393652e+00 6.61403313e-02 2.53124200e-02
-1.12383139e+00 -6.42525613e-01 -8.61098349e-01 -4.04277384e-01
-1.27458960e-01 -1.01277459e+00 1.01661158e+00 -4.88770068e-01
-1.58422041e+00 7.40830958e-01 -1.59242880e-02 -5.63035905e-01
4.53519315e-01 -4.09522235e-01 2.08176747e-02 3.09172690e-01
6.49956316e-02 1.23584294e+00 5.86389482e-01 -6.73461616e-01
-7.30251849e-01 -4.85526055e-01 -9.90343466e-02 3.42975378e-01
-1.44379497e-01 -4.16248173e-01 -1.78791434e-01 -5.95709532e-02
2.23489061e-01 -1.00078690e+00 -1.10762261e-01 7.90931106e-01
6.05796725e-02 4.17676270e-01 7.81021655e-01 -1.72209293e-01
3.12123835e-01 -2.10322475e+00 1.70569256e-01 -5.87550342e-01
2.15104297e-01 5.63146137e-02 -1.48624554e-01 2.86250025e-01
4.70642656e-01 -2.26474747e-01 -5.44599593e-02 -3.82972121e-01
-4.79232728e-01 3.18599343e-01 -8.42216387e-02 9.29020405e-01
2.99708247e-01 6.37953818e-01 -1.08730721e+00 -4.03741688e-01
3.90885651e-01 6.27500296e-01 -6.89437628e-01 3.27124111e-02
-7.74008408e-02 9.48005855e-01 -1.68299958e-01 5.92854321e-01
2.40075022e-01 -1.29106015e-01 -2.15460166e-01 -1.71196178e-01
-5.08670449e-01 1.87481895e-01 -1.09546328e+00 2.25074673e+00
-3.04664880e-01 1.39104450e+00 1.37466788e-01 -5.54790795e-01
8.75799716e-01 1.61878943e-01 7.06142724e-01 -7.91375697e-01
2.55621880e-01 3.89702469e-01 2.29584426e-01 -5.31283915e-01
4.84662443e-01 5.52714355e-02 3.11856776e-01 1.49043351e-01
1.32501647e-01 -5.52598894e-01 3.96788239e-01 -4.30385098e-02
1.49840164e+00 4.55374032e-01 1.29470453e-01 -3.58802915e-01
5.91298379e-02 3.22442263e-01 7.56686628e-01 6.30714297e-01
-7.61566937e-01 8.24564755e-01 -5.79952672e-02 -6.72491848e-01
-1.12986934e+00 -1.11701715e+00 -1.65418983e-01 4.41490501e-01
2.93820411e-01 -5.82203083e-02 -3.36682379e-01 7.80250654e-02
6.34494945e-02 -2.02238001e-02 -3.84822696e-01 -1.19788341e-01
-5.37230074e-01 -6.08642280e-01 5.59276342e-01 1.76063597e-01
7.89313853e-01 -1.00398409e+00 -1.90648270e+00 3.14037174e-01
-1.11667238e-01 -1.46435523e+00 -4.22764868e-02 3.39593798e-01
-7.47739613e-01 -1.20899177e+00 -5.08503139e-01 -5.72696269e-01
3.64108682e-01 5.77937961e-01 1.03820753e+00 -1.90223902e-01
-6.45702600e-01 3.49019885e-01 -8.35570991e-02 -3.74984503e-01
2.06226572e-01 -3.22797269e-01 4.42965984e-01 -2.95300901e-01
-1.30101591e-02 -7.89251685e-01 -9.59865093e-01 3.17276776e-01
-4.84739184e-01 4.98484410e-02 3.55714977e-01 8.75576615e-01
5.72766125e-01 -4.83807296e-01 3.46495986e-01 -1.10840984e-01
1.83459133e-01 -1.77641556e-01 -8.77062559e-01 -3.67201447e-01
-3.14319760e-01 -4.56396379e-02 6.36742353e-01 -4.63097692e-01
-5.95659137e-01 4.07741129e-01 3.91885430e-01 -4.24859971e-01
-1.40704304e-01 3.36490095e-01 4.53168392e-01 -3.99131507e-01
1.13850367e+00 2.63194740e-01 3.50378126e-01 2.25244641e-01
1.69997945e-01 2.87218928e-01 9.61020052e-01 -8.64471197e-02
4.75097865e-01 9.28691924e-01 5.25827825e-01 -1.15915060e+00
-4.67349857e-01 -3.79395247e-01 -7.23752260e-01 -4.49829847e-01
7.27404237e-01 -1.25813317e+00 -8.61880779e-01 7.53007650e-01
-1.13677931e+00 -5.57458639e-01 -3.13566297e-01 1.09350193e+00
-7.62860596e-01 1.57272518e-01 -3.96733433e-01 -7.04370379e-01
-2.10463509e-01 -1.08341205e+00 7.68449664e-01 3.80503863e-01
-3.14529955e-01 -7.27879047e-01 2.65068442e-01 5.09106740e-02
6.35997653e-01 5.39736032e-01 7.92638510e-02 1.71481147e-01
-8.68446290e-01 -1.05389357e-01 -1.67627603e-01 4.87200217e-03
1.94642857e-01 2.91181713e-01 -1.06452501e+00 -5.78825831e-01
-7.43704587e-02 -6.76283479e-01 8.45882475e-01 5.27835429e-01
3.25857192e-01 2.49680191e-01 -1.34768844e-01 1.01888239e+00
1.37827396e+00 5.63242026e-02 4.17854428e-01 3.58177304e-01
4.86407816e-01 6.17024779e-01 5.38653672e-01 6.19767785e-01
3.80736530e-01 5.72710097e-01 7.89247572e-01 2.42159784e-01
-3.33838165e-01 1.84441030e-01 3.29929650e-01 5.14971375e-01
-6.23662490e-03 -3.38330418e-01 -9.01861310e-01 6.17960095e-01
-1.69058180e+00 -7.81924665e-01 1.52815571e-02 2.42734790e+00
4.09853756e-01 1.69529065e-01 -1.84202418e-01 7.05418736e-02
2.61621743e-01 8.83658007e-02 -8.45301807e-01 4.58542109e-02
-2.94526190e-01 -1.45709679e-01 5.41191697e-01 4.58274961e-01
-1.06326473e+00 8.97941411e-01 6.41706896e+00 -2.29514241e-01
-1.40482974e+00 -2.11154222e-01 2.07606182e-02 -3.94467145e-01
1.46472573e-01 6.09559007e-02 -6.77163541e-01 4.05938238e-01
1.13547194e+00 7.45713860e-02 6.33894205e-01 6.15521967e-01
4.57051128e-01 -6.10017538e-01 -1.13945401e+00 1.24427569e+00
-1.24984356e-02 -1.36101735e+00 -2.92087257e-01 1.38492018e-01
5.31677723e-01 6.75028324e-01 -8.03447142e-02 -2.88407445e-01
2.39384234e-01 -8.22428644e-01 8.03503692e-01 6.18617535e-01
8.82497787e-01 -4.53889132e-01 4.30169523e-01 4.30763841e-01
-1.12542915e+00 -4.95505659e-03 -5.75628281e-01 -5.40571094e-01
2.38997623e-01 5.75843871e-01 -6.26460552e-01 -3.17033526e-04
7.60857105e-01 1.01042533e+00 -3.52375031e-01 1.26151621e+00
-5.85947484e-02 1.40596882e-01 -5.97610295e-01 -1.58583388e-01
-2.98210867e-02 7.33909607e-02 7.02285945e-01 8.64863753e-01
4.24533278e-01 -1.17816851e-01 1.15613505e-01 8.11242104e-01
3.18422139e-01 -4.99115527e-01 -8.89497042e-01 -1.27441324e-02
8.06507885e-01 1.25313032e+00 -7.88169563e-01 -5.67674590e-03
-4.87765700e-01 9.46977913e-01 1.75912365e-01 1.19503997e-01
-6.32108927e-01 -5.03895581e-01 1.23275340e+00 -1.36737689e-01
3.98720130e-02 -1.07305813e+00 -2.05754086e-01 -1.18872285e+00
-1.08248122e-01 -4.09647226e-01 -9.10436586e-02 -9.16378975e-01
-6.75792575e-01 5.88981450e-01 -5.14956236e-01 -1.26988828e+00
-5.30117273e-01 -5.89571416e-01 -2.09969580e-01 8.76964092e-01
-1.64845788e+00 -6.70691669e-01 -1.01759505e+00 6.43392324e-01
3.52853298e-01 -2.40434662e-01 8.89684200e-01 1.14058740e-01
-4.44256902e-01 -7.16244131e-02 -1.29119039e-01 -9.43081975e-02
7.64647126e-01 -1.03566122e+00 7.23361313e-01 1.19726312e+00
1.01177663e-01 4.00748581e-01 1.02267194e+00 -4.70687985e-01
-1.81217301e+00 -9.10648465e-01 5.65433562e-01 -5.62517643e-01
4.79234636e-01 -3.55947316e-01 -5.84901512e-01 3.89376074e-01
-1.35977864e-01 5.62159896e-01 9.48909074e-02 -4.99640971e-01
-1.31670579e-01 -1.46228343e-01 -1.00615990e+00 6.75015450e-01
1.11734390e+00 -5.43437183e-01 -1.53250709e-01 2.39145800e-01
6.62658274e-01 -5.77082932e-01 -5.82921207e-01 3.15949649e-01
8.15804482e-01 -1.25891483e+00 9.79612470e-01 -9.88686234e-02
3.28224838e-01 -5.88039219e-01 -2.67309695e-01 -1.24783933e+00
-2.85995007e-01 -7.34853685e-01 -1.34124592e-01 6.86377883e-01
1.01490013e-01 -7.17563272e-01 1.00322950e+00 2.64387429e-01
-3.28988224e-01 -2.28550315e-01 -1.02976024e+00 -9.18183804e-01
-3.25858086e-01 -2.30332226e-01 -6.43620491e-02 5.95341623e-01
-1.37964979e-01 2.79650360e-01 -3.11506927e-01 2.03175247e-01
7.39059150e-01 -1.33428127e-01 9.68498468e-01 -1.27266884e+00
-1.91708595e-01 -1.59819916e-01 -9.09670472e-01 -1.04436612e+00
-1.85601622e-01 -3.29295874e-01 4.26094204e-01 -1.51919687e+00
-4.44153816e-01 -1.43249065e-01 1.06123149e-01 2.84150153e-01
2.45502755e-01 3.06623399e-01 -3.01651713e-02 5.06486356e-01
-3.85673642e-01 5.04446089e-01 1.24143875e+00 2.38838613e-01
-1.82068855e-01 -3.91974241e-01 -1.03355981e-01 5.56960642e-01
6.81517363e-01 -1.29073873e-01 -6.81248248e-01 -7.96276450e-01
2.10645869e-02 1.68213755e-01 6.51682794e-01 -1.93701804e+00
6.25808775e-01 -1.98346451e-01 5.09232938e-01 -2.79068828e-01
8.01176608e-01 -4.43811536e-01 8.07564110e-02 9.23614085e-01
-1.63174998e-02 2.89850473e-01 3.22120935e-01 5.85105598e-01
-1.99558958e-01 3.96798491e-01 9.94532049e-01 -3.12420636e-01
-1.28921795e+00 3.12637568e-01 -6.71545446e-01 3.62387538e-01
8.47654045e-01 -4.09349322e-01 -7.13102996e-01 -2.18530968e-01
-1.67007044e-01 8.81089419e-02 6.63918674e-01 3.08270514e-01
8.37947488e-01 -7.67282486e-01 -5.80706239e-01 4.79183942e-01
1.80191889e-01 3.17050308e-01 7.09957629e-02 6.90588951e-01
-1.10672915e+00 5.79382658e-01 -9.42615032e-01 -9.36023891e-01
-9.56978679e-01 1.19025074e-01 6.06287479e-01 4.78302062e-01
-6.79015040e-01 1.05725336e+00 -3.08065563e-02 -1.83752880e-01
4.85978974e-03 -1.73180118e-01 -1.65637843e-02 -2.17964530e-01
4.83235300e-01 2.92566091e-01 -6.27980530e-02 -4.90826011e-01
-6.73606157e-01 6.24784291e-01 4.73418593e-01 -4.41297054e-01
1.14530838e+00 -3.42526793e-01 1.46168083e-01 5.10093570e-01
7.20132947e-01 -4.83195335e-01 -2.04393244e+00 4.25979495e-02
-3.06674987e-01 -5.45513451e-01 2.70883411e-01 -3.89086604e-01
-8.19713831e-01 1.06692898e+00 6.80509627e-01 -1.20260015e-01
9.19874907e-01 -5.07951140e-01 5.21318436e-01 7.13674366e-01
8.38630736e-01 -8.92450273e-01 2.73970813e-01 9.06398535e-01
7.10839033e-01 -1.28000247e+00 2.85086147e-02 -1.39433056e-01
-1.48881182e-01 1.08529103e+00 9.75597680e-01 -2.02284530e-01
4.58420038e-01 3.51306140e-01 1.53251633e-01 -4.76073548e-02
-1.17535555e+00 -2.82141507e-01 -1.82864353e-01 1.05215096e+00
3.10943305e-01 -3.67941290e-01 2.20886558e-01 -3.94603997e-01
-3.61223847e-01 1.33810267e-01 9.18174624e-01 1.34539008e+00
-7.91657865e-01 -4.59318787e-01 4.22866317e-03 7.64064118e-02
-1.37151405e-01 1.34916976e-02 -2.56881863e-01 7.57946253e-01
-1.40015632e-01 9.93168473e-01 1.52224615e-01 -4.83251661e-01
3.10124904e-01 -3.23798716e-01 7.14149773e-01 -4.77379382e-01
-1.19446166e-01 -2.72120029e-01 -2.19016429e-02 -9.13571596e-01
-4.44473594e-01 -6.28481507e-01 -1.33812320e+00 -4.38626766e-01
-4.51098867e-02 -3.07081640e-01 9.75980580e-01 7.09732473e-01
6.78322434e-01 5.95921695e-01 3.07790101e-01 -1.25932598e+00
-1.91532180e-01 -7.60242820e-01 -4.82028484e-01 2.29324669e-01
7.37580001e-01 -8.74576926e-01 -5.84591091e-01 2.09580615e-01] | [8.593239784240723, -1.3269559144973755] |
ddb303e6-cddc-492f-834a-8f52a62bd5a7 | nonparallel-high-quality-audio-super | 2210.15887 | null | https://arxiv.org/abs/2210.15887v2 | https://arxiv.org/pdf/2210.15887v2.pdf | Nonparallel High-Quality Audio Super Resolution with Domain Adaptation and Resampling CycleGANs | Neural audio super-resolution models are typically trained on low- and high-resolution audio signal pairs. Although these methods achieve highly accurate super-resolution if the acoustic characteristics of the input data are similar to those of the training data, challenges remain: the models suffer from quality degradation for out-of-domain data, and paired data are required for training. To address these problems, we propose Dual-CycleGAN, a high-quality audio super-resolution method that can utilize unpaired data based on two connected cycle consistent generative adversarial networks (CycleGAN). Our method decomposes the super-resolution method into domain adaptation and resampling processes to handle acoustic mismatch in the unpaired low- and high-resolution signals. The two processes are then jointly optimized within the CycleGAN framework. Experimental results verify that the proposed method significantly outperforms conventional methods when paired data are not available. Code and audio samples are available from https://chomeyama.github.io/DualCycleGAN-Demo/. | ['Kentaro Tachibana', 'Ryuichi Yamamoto', 'Reo Yoneyama'] | 2022-10-28 | null | null | null | null | ['audio-super-resolution', 'audio-super-resolution'] | ['audio', 'music'] | [ 3.22139680e-01 -1.35063946e-01 7.53383785e-02 -7.38353878e-02
-1.65616667e+00 -5.24150968e-01 2.96193182e-01 -8.13044250e-01
1.73062950e-01 7.53050268e-01 5.64196050e-01 2.85169303e-01
1.78435475e-01 -6.91720068e-01 -7.90235937e-01 -6.67890906e-01
7.41557330e-02 2.02779584e-02 6.97767511e-02 -3.15848798e-01
-3.50171059e-01 -1.13494925e-01 -1.63916624e+00 5.32502830e-01
9.66972411e-01 8.29692125e-01 2.12220013e-01 1.05491769e+00
2.07178891e-01 5.28396487e-01 -5.64752400e-01 -1.05763428e-01
4.01093304e-01 -1.03170550e+00 -3.04861814e-01 -2.97867000e-01
4.50886875e-01 -5.05529940e-01 -4.50266838e-01 1.26021600e+00
9.26241994e-01 2.12825850e-01 2.58776695e-01 -1.10935497e+00
-8.09052527e-01 8.71604681e-01 -4.63105887e-01 2.93849975e-01
4.67645407e-01 -4.71271500e-02 7.74138510e-01 -1.14589310e+00
3.13860714e-01 1.34980643e+00 8.05980861e-01 7.96243310e-01
-1.20517683e+00 -1.36125529e+00 -2.09962860e-01 2.11986691e-01
-1.48875201e+00 -8.89432132e-01 1.06546640e+00 -2.04671502e-01
1.48013651e-01 1.11171536e-01 3.81429195e-01 1.48322332e+00
-1.99394196e-01 1.98724553e-01 1.04617476e+00 -2.33031183e-01
8.31456035e-02 -8.60641301e-02 -5.83375573e-01 8.51506889e-02
-1.97014377e-01 4.54015732e-01 -9.88384783e-01 -1.60942510e-01
1.11618054e+00 -3.04405183e-01 -5.62714577e-01 1.63189068e-01
-9.12878573e-01 5.50290406e-01 2.34536037e-01 4.65233713e-01
-2.86802262e-01 5.26479520e-02 2.77347803e-01 3.94412428e-01
6.22444570e-01 1.32485926e-01 -1.14568360e-01 -1.06330886e-01
-1.09809089e+00 2.86376595e-01 3.16875786e-01 9.98804629e-01
4.73683864e-01 8.61795783e-01 9.73704234e-02 1.04704416e+00
1.35986611e-01 6.16526961e-01 7.74036705e-01 -1.27177978e+00
6.30519986e-01 -3.22995216e-01 1.17215618e-01 -8.65029216e-01
1.35673150e-01 -5.39681315e-01 -1.24458206e+00 7.28268847e-02
2.72122055e-01 -3.37471426e-01 -7.53070235e-01 2.03798151e+00
3.51739883e-01 8.79432440e-01 2.76675999e-01 1.13079643e+00
1.09214556e+00 9.14317191e-01 -1.87334314e-01 -2.25347936e-01
1.22466612e+00 -8.07305157e-01 -1.17174149e+00 -1.13107599e-01
-2.95167655e-01 -9.08583164e-01 1.12283182e+00 3.77191633e-01
-1.51567149e+00 -1.13426638e+00 -1.32354760e+00 -1.28678456e-01
1.43967003e-01 -1.20146222e-01 -2.16821507e-01 4.01846468e-01
-9.91374195e-01 6.89698279e-01 -6.91696405e-01 3.83016855e-01
2.72068888e-01 5.77787533e-02 -1.66806325e-01 -4.65673022e-02
-1.74837971e+00 1.55946925e-01 1.64928809e-01 4.68540229e-02
-1.20317233e+00 -1.04698622e+00 -9.42368209e-01 -4.71521989e-02
2.21826017e-01 -5.51612020e-01 1.46375883e+00 -1.11354709e+00
-1.76476312e+00 3.01707149e-01 -1.02664657e-01 -3.84424269e-01
5.41720927e-01 -4.10599351e-01 -9.93894100e-01 2.82109290e-01
-4.28520478e-02 2.11895451e-01 1.48648202e+00 -1.44924259e+00
-5.67732036e-01 -1.01657532e-01 -4.27005678e-01 2.18105778e-01
-1.90054402e-01 -1.31787002e-01 -4.12708342e-01 -1.08522856e+00
8.61096606e-02 -6.18353248e-01 -1.05395995e-01 -3.79417837e-01
-3.03309411e-01 4.79464054e-01 9.42845643e-01 -1.14268768e+00
1.08674061e+00 -2.39194417e+00 1.71216473e-01 -1.73140779e-01
5.03618978e-02 3.19645166e-01 -3.18100005e-01 6.60142973e-02
-2.95824200e-01 1.80998996e-01 -2.38275409e-01 -4.41927791e-01
-2.26348847e-01 -1.68855548e-01 -6.28150284e-01 2.42140129e-01
2.48921871e-01 4.64946687e-01 -8.65659595e-01 -4.57688004e-01
1.82892382e-02 1.18695498e+00 -4.96375710e-01 6.00374162e-01
9.49664265e-02 1.13853967e+00 -2.47972548e-01 5.05327940e-01
7.42055357e-01 1.55410171e-01 7.07294270e-02 -3.98805916e-01
1.26974121e-01 3.90081167e-01 -1.47462428e+00 1.87134576e+00
-4.97765392e-01 4.01185691e-01 4.64061677e-01 -5.17705917e-01
9.45469558e-01 7.21007705e-01 3.44567508e-01 -8.01790774e-01
-6.69714585e-02 2.83725113e-01 -2.80241102e-01 -2.56345451e-01
5.96628189e-01 -4.58563387e-01 -9.99676958e-02 1.98356658e-01
2.04089940e-01 -3.39361340e-01 -8.43354389e-02 -1.17418140e-01
7.83326387e-01 2.74540633e-01 4.84225750e-02 2.97279060e-01
4.27866787e-01 -6.14789903e-01 1.09311986e+00 3.68203908e-01
-1.62059620e-01 1.18727458e+00 9.04630870e-02 1.88777238e-01
-1.43799436e+00 -1.25279069e+00 1.01175129e-01 9.25668836e-01
-1.92535426e-02 -3.49764347e-01 -1.07355857e+00 2.66207624e-02
-4.74670887e-01 3.72478992e-01 -3.42788696e-01 -1.31089613e-01
-7.84468114e-01 -3.11005980e-01 9.29658592e-01 4.93049622e-01
6.06540203e-01 -7.43827939e-01 -1.06881872e-01 3.34289283e-01
-6.71476483e-01 -1.35290885e+00 -7.70838320e-01 -2.76421636e-01
-7.93315291e-01 -6.08534276e-01 -9.01142418e-01 -7.56876230e-01
1.28273711e-01 1.95363835e-01 1.02755642e+00 -3.83106470e-01
4.74239625e-02 1.05935931e-01 -4.48754430e-01 -1.84035152e-01
-7.65528858e-01 -7.01617897e-02 3.44147533e-01 2.24632442e-01
-8.17076638e-02 -1.26102352e+00 -5.83217084e-01 3.28379422e-01
-9.43947554e-01 1.84132680e-02 4.28195417e-01 8.98639977e-01
1.00590432e+00 2.60730833e-01 9.98270392e-01 -5.97142816e-01
6.03428841e-01 -5.47917128e-01 -4.83511865e-01 -3.21577728e-01
-3.35606784e-01 -1.57301366e-01 1.00603330e+00 -7.19056785e-01
-1.28465140e+00 -8.71298835e-02 -3.41408849e-01 -9.43235457e-01
-1.65958390e-01 1.75225869e-01 -5.17646372e-01 3.30920845e-01
6.32453382e-01 2.79675454e-01 -1.79395735e-01 -7.79361725e-01
3.77274364e-01 7.08850682e-01 1.21139205e+00 -2.69105762e-01
1.29094803e+00 3.12447637e-01 -3.26949298e-01 -7.13332474e-01
-6.43127739e-01 -1.52606340e-02 -3.91225010e-01 -1.01701446e-01
7.28457272e-01 -1.56510997e+00 -6.15089498e-02 6.00443125e-01
-8.47361922e-01 -4.12816346e-01 -4.56409872e-01 6.37345612e-01
-6.95929229e-01 2.53286034e-01 -7.76474237e-01 -8.07039261e-01
-3.97724509e-01 -8.95819902e-01 9.17891502e-01 4.20900375e-01
-6.15997659e-03 -6.90806866e-01 3.30473959e-01 3.39386463e-01
4.17775482e-01 4.12738115e-01 4.46087450e-01 -2.36561999e-01
-4.61079150e-01 -3.11051350e-04 2.10880578e-01 5.62665999e-01
2.97468156e-01 -9.72894430e-02 -1.34710717e+00 -4.02608275e-01
2.36223325e-01 -2.90626585e-01 5.87361693e-01 3.88945282e-01
1.12792480e+00 -5.25493979e-01 2.07892463e-01 9.05905306e-01
1.36854327e+00 5.50060906e-02 7.51204312e-01 -1.83893248e-01
6.88908517e-01 3.63631696e-01 7.06905782e-01 5.14132202e-01
2.41054833e-01 8.06933939e-01 3.33769590e-01 -7.68453553e-02
-6.91523314e-01 -7.53552198e-01 7.43225396e-01 1.42350495e+00
-1.08025856e-01 9.23441052e-02 -4.53427136e-01 6.71661496e-01
-1.35184956e+00 -1.28154695e+00 3.03638615e-02 2.04949021e+00
1.30189264e+00 -5.57528473e-02 2.25562155e-01 5.45090854e-01
1.11583161e+00 3.18295687e-01 -7.08613634e-01 -5.52051216e-02
-1.51459202e-01 4.47252810e-01 1.27856299e-01 6.07643962e-01
-8.10580015e-01 6.01094067e-01 5.91515827e+00 1.05809867e+00
-1.08760047e+00 5.42091608e-01 3.52483273e-01 -3.76925260e-01
-3.91993314e-01 -4.00211155e-01 -5.22131860e-01 6.11928880e-01
1.30280542e+00 -2.95183688e-01 7.55989909e-01 5.85786223e-01
3.30951303e-01 3.97501498e-01 -7.72785008e-01 1.03365839e+00
-4.84347269e-02 -9.73654151e-01 -1.85527116e-01 -3.18315059e-01
8.64170730e-01 -2.86542386e-01 5.92864275e-01 2.64756143e-01
2.84281880e-01 -1.01659024e+00 8.15792799e-01 4.39850658e-01
1.54737878e+00 -1.00307822e+00 3.38603020e-01 1.27158120e-01
-1.68924534e+00 -2.70316545e-02 -2.05682784e-01 2.36355707e-01
1.81816921e-01 5.46200573e-01 -5.62069952e-01 7.17437327e-01
9.57332730e-01 6.65015101e-01 -8.59557390e-02 5.39014935e-01
-4.01293099e-01 8.66763651e-01 -1.29756723e-02 8.23912859e-01
-3.53248864e-01 1.03504984e-02 8.70541453e-01 1.03542960e+00
6.70278549e-01 2.44654208e-01 -1.93339869e-01 1.13281226e+00
-2.57003963e-01 -2.82931596e-01 -4.60754693e-01 5.35689294e-02
9.07045066e-01 1.11282945e+00 7.03310743e-02 -6.59060106e-02
-2.66614169e-01 9.17787373e-01 -1.78072482e-01 5.07287204e-01
-9.62589025e-01 -5.18555582e-01 8.64547431e-01 6.49849698e-02
3.88052732e-01 -5.19553162e-02 -1.63346455e-01 -1.20304656e+00
8.20516422e-02 -1.32465124e+00 2.55611539e-01 -9.55424488e-01
-1.10749841e+00 7.84347892e-01 -2.47513354e-01 -1.78528214e+00
-6.98680580e-01 1.68668836e-01 -4.38756436e-01 1.00550401e+00
-1.62646961e+00 -1.21245742e+00 -4.74289834e-01 8.14959705e-01
7.07884252e-01 -1.48072556e-01 8.80987167e-01 6.17687821e-01
-3.76633108e-01 8.60737026e-01 3.03404242e-01 2.01395825e-01
9.38378930e-01 -9.83804524e-01 4.60606039e-01 1.15473557e+00
-7.93312956e-03 1.73542753e-01 9.77069438e-01 -4.76407468e-01
-1.12537503e+00 -1.33204865e+00 3.55464131e-01 -9.37057212e-02
4.68361199e-01 -4.29037899e-01 -1.33416033e+00 4.37762439e-01
2.18506604e-01 1.19525298e-01 7.52188504e-01 -3.64083529e-01
-5.90145648e-01 -3.03554088e-01 -1.19549489e+00 4.18841332e-01
9.59952176e-01 -9.68028843e-01 -5.50578356e-01 -3.59393686e-01
1.17882800e+00 -7.21954465e-01 -1.22037280e+00 3.77821237e-01
5.17649829e-01 -1.02465141e+00 1.11098444e+00 -1.77013084e-01
7.09160864e-01 -6.05397165e-01 -4.07859832e-01 -1.56175268e+00
-1.63184643e-01 -9.63209927e-01 -3.59897405e-01 1.62646723e+00
2.14769349e-01 -5.07491767e-01 3.35608810e-01 7.22621903e-02
-1.68712974e-01 -9.45364758e-02 -1.05054379e+00 -8.62368166e-01
1.18802987e-01 -3.78475666e-01 9.06457067e-01 1.04717112e+00
-1.10496297e-01 2.83604711e-01 -7.88579345e-01 6.17335498e-01
9.10517871e-01 6.62995577e-02 7.34916806e-01 -8.69801581e-01
-5.58103561e-01 -5.95726855e-02 1.14328722e-02 -6.67543888e-01
1.17506504e-01 -4.79790002e-01 2.61466116e-01 -1.02147388e+00
-1.56346664e-01 -1.44936264e-01 -2.91311353e-01 1.33291304e-01
-2.02295333e-01 6.40836477e-01 2.31117874e-01 2.24585637e-01
-1.91433340e-01 7.36776114e-01 1.30550659e+00 2.90453453e-02
-3.70097995e-01 -6.56395853e-02 -7.83695996e-01 7.01260924e-01
8.56510639e-01 -5.14778256e-01 -4.84070271e-01 -3.45145762e-01
-1.04269259e-01 6.02819204e-01 3.98658812e-01 -1.35405755e+00
1.22656047e-01 -2.43432559e-02 3.16187561e-01 -3.90491605e-01
6.16948247e-01 -7.47912467e-01 7.10821748e-01 1.52040273e-01
-5.32103002e-01 -2.53129750e-01 3.06620896e-01 6.03674531e-01
-6.27044737e-01 2.76920944e-01 1.22117305e+00 4.21409756e-02
-3.15710783e-01 3.49240482e-01 -1.10989191e-01 3.36191416e-01
4.26707119e-01 2.53857709e-02 -2.20422924e-01 -8.55903566e-01
-8.27532589e-01 -1.16363980e-01 3.69348735e-01 5.64700305e-01
7.96269834e-01 -1.80104887e+00 -1.13597655e+00 2.36285016e-01
-2.15923667e-01 3.53567660e-01 7.39596009e-01 4.92797285e-01
1.21314846e-01 -2.04100162e-02 -2.64941394e-01 -2.63456643e-01
-1.21800029e+00 4.10900205e-01 4.96172488e-01 -9.93478894e-02
-5.99634767e-01 7.50573874e-01 1.57105371e-01 -1.53514475e-01
1.06187299e-01 -1.69818938e-01 -1.11112028e-01 5.75175323e-03
9.99073088e-01 5.14596701e-01 -1.84360072e-01 -8.65615308e-01
-3.02315131e-02 7.04787970e-01 2.64612257e-01 -5.43009996e-01
1.40614522e+00 -4.11660463e-01 2.34851927e-01 7.03087628e-01
1.16371071e+00 2.85214990e-01 -1.54651642e+00 -6.11748099e-01
-7.12790012e-01 -5.57558954e-01 1.47466108e-01 -4.77860242e-01
-1.31008303e+00 8.58796358e-01 7.54860938e-01 1.50309786e-01
1.65497863e+00 -3.35810602e-01 1.10189641e+00 -4.20454383e-01
1.61053881e-01 -1.10777354e+00 2.88368165e-01 2.56597072e-01
1.12028527e+00 -8.37413311e-01 -2.18841791e-01 -2.98740417e-01
-5.29132187e-01 7.47406662e-01 4.84315366e-01 -2.69232601e-01
4.53300416e-01 5.42203903e-01 2.00265333e-01 4.98273015e-01
-7.46419013e-01 -6.15525767e-02 2.93261528e-01 1.04068613e+00
2.32471481e-01 -1.24399453e-01 1.88737556e-01 1.13705254e+00
-5.62000394e-01 -5.19013219e-03 6.41425908e-01 4.50624019e-01
-1.94090277e-01 -9.95650887e-01 -8.35717618e-01 -4.46614400e-02
-7.81567931e-01 -1.16534509e-01 -1.47653520e-01 4.04049397e-01
8.17638189e-02 1.13819981e+00 -4.03196737e-02 -7.82639682e-01
3.80929798e-01 4.29930985e-02 2.93945789e-01 -3.01096112e-01
-3.27880919e-01 5.59320569e-01 -8.54412019e-02 -5.36611021e-01
-3.87095988e-01 -6.90810502e-01 -1.20356023e+00 -3.92066479e-01
-2.37310976e-01 1.89967528e-01 3.94793481e-01 5.01949906e-01
4.49104190e-01 1.00809574e+00 9.76973057e-01 -1.14931965e+00
-5.73374927e-01 -1.04700327e+00 -7.35453367e-01 3.18401098e-01
9.43553329e-01 -2.28769556e-01 -5.46623051e-01 5.51169991e-01] | [15.364683151245117, 6.014230728149414] |
6b112748-4225-494c-ad6a-6549c6dfcb34 | brain-intelligence-go-beyond-artificial | 1706.01040 | null | http://arxiv.org/abs/1706.01040v1 | http://arxiv.org/pdf/1706.01040v1.pdf | Brain Intelligence: Go Beyond Artificial Intelligence | Artificial intelligence (AI) is an important technology that supports daily
social life and economic activities. It contributes greatly to the sustainable
growth of Japan's economy and solves various social problems. In recent years,
AI has attracted attention as a key for growth in developed countries such as
Europe and the United States and developing countries such as China and India.
The attention has been focused mainly on developing new artificial intelligence
information communication technology (ICT) and robot technology (RT). Although
recently developed AI technology certainly excels in extracting certain
patterns, there are many limitations. Most ICT models are overly dependent on
big data, lack a self-idea function, and are complicated. In this paper, rather
than merely developing next-generation artificial intelligence technology, we
aim to develop a new concept of general-purpose intelligence cognition
technology called Beyond AI. Specifically, we plan to develop an intelligent
learning model called Brain Intelligence (BI) that generates new ideas about
events without having experienced them by using artificial life with an imagine
function. We will also conduct demonstrations of the developed BI intelligence
learning model on automatic driving, precision medical care, and industrial
robots. | ['Min Chen', 'Seiichi Serikawa', 'Yujie Li', 'Hyoungseop Kim', 'Huimin Lu'] | 2017-06-04 | null | null | null | null | ['artificial-life', 'industrial-robots'] | ['miscellaneous', 'robots'] | [-1.52444541e-01 5.16189396e-01 -4.93556075e-02 -1.90123275e-01
4.39498097e-01 -7.61064813e-02 6.87988400e-01 -1.97908700e-01
-2.08774582e-01 8.44430149e-01 5.02580479e-02 -4.13138002e-01
-2.16810659e-01 -1.17872965e+00 -3.44156802e-01 -3.41332614e-01
1.65755526e-04 5.71218789e-01 6.97237104e-02 -5.10570347e-01
4.47366804e-01 4.34402734e-01 -1.55074334e+00 -9.32527110e-02
1.33987629e+00 7.81109929e-01 4.86475080e-01 2.14024276e-01
-4.49815392e-01 1.29349089e+00 -7.38188505e-01 -7.76086226e-02
-8.35808590e-02 -4.89662349e-01 -9.44938719e-01 -1.59121096e-01
-8.35714579e-01 1.04420111e-01 -2.06447288e-01 9.76398349e-01
1.77902609e-01 -2.54818410e-01 6.05546534e-01 -1.49238825e+00
-8.99277925e-01 5.27331352e-01 -3.89417976e-01 -6.42618611e-02
3.55383843e-01 1.15870237e-01 2.06099451e-01 -4.97065634e-01
5.18860579e-01 1.19140720e+00 5.27030230e-01 7.28517950e-01
-3.37785035e-01 -8.12899292e-01 -7.21637011e-02 7.02072203e-01
-1.03419638e+00 5.04019707e-02 9.74231064e-01 -3.46186310e-01
8.69238496e-01 -9.07026231e-02 8.96359205e-01 6.52384222e-01
6.64106309e-01 1.01735175e+00 1.02254736e+00 -6.86645031e-01
5.02762735e-01 3.20186049e-01 3.37912560e-01 7.43710220e-01
4.35312539e-01 1.98445052e-01 -1.57443121e-01 4.24156517e-01
7.21373618e-01 3.12731177e-01 2.80511472e-02 -1.32603452e-01
-1.58678281e+00 8.71790528e-01 6.03559196e-01 8.77757311e-01
-6.62836194e-01 1.81713570e-02 1.81517974e-01 3.95512313e-01
7.75621757e-02 6.47502482e-01 -6.66095555e-01 -5.03713787e-01
9.69196185e-02 -8.55176002e-02 1.04748344e+00 8.90322030e-01
8.94232512e-01 3.14160585e-01 3.19895774e-01 7.33284056e-01
1.83100581e-01 6.08498752e-01 1.10283864e+00 -1.13506687e+00
3.29170772e-03 1.10739815e+00 -3.43838930e-02 -1.07185543e+00
-8.41267347e-01 -3.40831459e-01 -1.01468658e+00 2.66856134e-01
-1.02525316e-01 -6.33378029e-01 -8.93823445e-01 1.43067992e+00
2.66826199e-03 -1.15546428e-01 5.22530794e-01 7.48486638e-01
9.46300209e-01 9.35800850e-01 5.03559746e-02 -3.05408742e-02
1.34131408e+00 -1.11111808e+00 -1.11072326e+00 -3.78555030e-01
6.85511708e-01 -2.09357873e-01 8.81705403e-01 7.39661694e-01
-9.08582509e-01 -9.04196203e-01 -1.13475966e+00 1.81724757e-01
-1.04408729e+00 -1.61403537e-01 1.21205902e+00 7.33594060e-01
-8.93133283e-01 2.72598892e-01 -5.69225132e-01 -7.85711288e-01
1.95323557e-01 4.85715419e-01 -1.30552635e-01 4.13773246e-02
-1.46337593e+00 1.34473062e+00 4.98233706e-01 -3.01949948e-01
-3.22726071e-01 -1.79948777e-01 -7.92612135e-01 -8.10858756e-02
1.46095395e-01 -8.36788654e-01 1.19056952e+00 -1.27720141e+00
-1.78030705e+00 4.59907830e-01 2.10995078e-01 -5.91425180e-01
-2.37560064e-01 -1.96223646e-01 -9.56237316e-01 7.91346580e-02
1.43467456e-01 8.21502149e-01 3.00322205e-01 -1.18515050e+00
-8.38915825e-01 -6.30179465e-01 -9.88095626e-02 1.43809244e-01
-7.04923332e-01 -7.71238580e-02 -2.33517244e-01 -5.36742568e-01
2.60169208e-02 -9.67658937e-01 -3.73391241e-01 -4.08089519e-01
1.41487420e-01 -5.21164298e-01 1.07050431e+00 -3.13829064e-01
9.80285108e-01 -2.06160069e+00 -2.49892864e-02 8.04337189e-02
1.87951714e-01 4.50390458e-01 3.40306386e-02 2.53756046e-01
1.86638400e-01 -1.02122188e-01 1.42701771e-02 5.91464341e-01
-1.37215942e-01 5.99863589e-01 -1.75678298e-01 -3.40058595e-01
7.84142986e-02 9.34731781e-01 -1.12399757e+00 -3.58181447e-01
4.70871925e-01 2.37108514e-01 -3.13644737e-01 -5.80273848e-03
-1.68523043e-01 3.66693527e-01 -9.01782632e-01 6.29777908e-01
2.41374105e-01 -3.61188948e-01 -1.86166331e-01 2.67103106e-01
-1.77651465e-01 -6.21165708e-02 -7.17395663e-01 1.66919661e+00
-4.92010415e-01 5.43139279e-01 -2.80404657e-01 -1.37920678e+00
1.19333398e+00 4.34022278e-01 6.74786747e-01 -1.13187027e+00
3.12518358e-01 1.54670045e-01 2.58524716e-01 -9.71556962e-01
1.99751586e-01 -2.72325099e-01 -1.45644963e-01 4.05133158e-01
-1.11247301e-01 -5.94251633e-01 9.46010128e-02 -3.74912284e-02
1.19628584e+00 3.84929329e-02 4.44179147e-01 -6.43008873e-02
5.88866174e-01 4.87831831e-01 7.21702814e-01 3.17675024e-01
-3.23032588e-01 1.18867069e-01 1.53229814e-02 -8.83660555e-01
-6.51541710e-01 -8.34560394e-01 1.21644564e-01 7.01267123e-01
3.83217216e-01 -2.31121808e-01 -7.23053098e-01 -4.65814531e-01
-3.96996796e-01 1.04227841e+00 -1.60885826e-01 -3.73903722e-01
-3.20926696e-01 -6.53329432e-01 3.05902690e-01 3.71513337e-01
1.39104843e+00 -1.79673147e+00 -7.74682581e-01 4.58034128e-01
-2.33210146e-01 -9.75291908e-01 3.55134726e-01 2.00092167e-01
-8.24203491e-01 -7.23551095e-01 -5.40282011e-01 -1.14410484e+00
4.31691200e-01 3.98186028e-01 9.24551904e-01 -1.73283145e-01
-2.49704197e-01 3.75432879e-01 -6.01759851e-01 -1.45700192e+00
-4.87782806e-01 1.63686290e-01 2.47951269e-01 -4.86858606e-01
9.25285995e-01 -4.65373963e-01 -4.92087007e-01 3.27866882e-01
-7.38601029e-01 3.92104268e-01 1.01635349e+00 5.25069237e-01
-2.16544703e-01 7.94239938e-01 1.30059206e+00 -5.66790938e-01
7.06274271e-01 -7.12858081e-01 -3.46053764e-02 7.65933394e-02
-7.05792427e-01 -3.46484706e-02 6.08270168e-01 -2.84378737e-01
-1.30134130e+00 -8.36225823e-02 -1.00002222e-01 1.59406126e-01
-4.92433190e-01 6.24445200e-01 -8.63232389e-02 2.89512351e-02
6.65163457e-01 2.64619559e-01 3.57696116e-01 -2.16426328e-02
1.74471676e-01 1.39861226e+00 4.87608343e-01 -3.82102877e-01
4.56491768e-01 2.35473216e-01 -2.08356008e-01 -1.02863717e+00
-4.12072986e-01 -7.91415423e-02 -2.90485740e-01 -4.52496469e-01
9.92646277e-01 -6.54570639e-01 -9.04282331e-01 4.33931649e-01
-1.16589594e+00 -1.58996418e-01 -1.67722031e-01 6.49034798e-01
-5.67611516e-01 1.35150880e-01 -4.76484984e-01 -6.65456831e-01
-3.95189226e-01 -7.97996163e-01 4.94038612e-01 6.63559139e-01
-2.74846017e-01 -9.14954185e-01 8.14840198e-02 5.16031682e-01
5.63750684e-01 -3.58557925e-02 9.42673326e-01 -3.73039216e-01
-4.23601270e-01 -2.00369477e-01 -2.86034614e-01 3.97460520e-01
3.61998707e-01 -4.31806326e-01 -7.61051238e-01 4.64046061e-01
4.00508493e-01 -2.15518862e-01 4.63649601e-01 1.75654396e-01
1.08789206e+00 -1.98170319e-01 -6.82780921e-01 -8.51323530e-02
9.45360065e-01 1.30711222e+00 8.82137954e-01 5.62647462e-01
3.58415991e-01 8.82230222e-01 7.07627416e-01 1.35204807e-01
8.17299545e-01 2.33980164e-01 5.12784980e-02 -2.21078143e-01
7.36243576e-02 -1.37740402e-02 3.91391367e-01 1.23265851e+00
-4.54501182e-01 1.63400754e-01 -1.13601887e+00 5.16820610e-01
-2.09098840e+00 -1.02649522e+00 1.56659968e-02 1.67566025e+00
6.91622019e-01 2.26921424e-01 -2.06471533e-01 3.72686803e-01
4.19154435e-01 -5.35089850e-01 -6.62625432e-01 -6.12580955e-01
8.47056210e-02 9.69808102e-02 1.29413884e-02 -4.17047590e-02
-8.09965968e-01 1.07287955e+00 6.33831882e+00 5.22760510e-01
-1.13414502e+00 -1.66022584e-01 6.30990803e-01 6.17094219e-01
-7.13775083e-02 -3.39351237e-01 -3.26880455e-01 4.82512087e-01
1.01746833e+00 -3.19141418e-01 4.25774992e-01 1.18564236e+00
2.32521281e-01 -2.23428812e-02 -5.41850209e-01 1.08615589e+00
8.91830549e-02 -1.22477174e+00 1.95261799e-02 7.87598938e-02
6.35358393e-01 -9.52480659e-02 -9.89912003e-02 7.22091675e-01
5.44320166e-01 -8.60140145e-01 2.72270650e-01 5.46356499e-01
1.54557556e-01 -8.58549535e-01 8.60248268e-01 5.90357244e-01
-7.05742478e-01 -6.04382277e-01 -4.46193218e-01 -8.03045273e-01
-2.23332364e-03 6.06162429e-01 -8.34632814e-01 4.10411537e-01
8.56544256e-01 6.44229114e-01 -3.34764183e-01 8.30772102e-01
-2.93352067e-01 4.08704221e-01 -4.04648967e-02 -5.27262688e-01
2.56272793e-01 -2.57776797e-01 1.51704147e-01 7.81262636e-01
4.82906997e-01 5.50819159e-01 5.30500375e-02 6.87146842e-01
3.41364563e-01 -8.78490806e-02 -1.11389387e+00 -2.20790297e-01
2.04127029e-01 1.12805748e+00 -8.76228094e-01 -4.93471533e-01
-7.38987684e-01 7.93817341e-01 -8.87596905e-02 1.67202905e-01
-6.50252819e-01 -7.63323247e-01 5.45709968e-01 -1.56591401e-01
-1.88098684e-01 -2.83264220e-01 -7.98030853e-01 -8.70135963e-01
-4.93971556e-01 -9.44105983e-01 8.99673700e-02 -1.05401218e+00
-1.05765140e+00 6.17502034e-01 -1.15330271e-01 -1.10493743e+00
-6.17686510e-01 -7.42645264e-01 -6.48023486e-01 2.20884636e-01
-1.37438154e+00 -1.05934739e+00 -4.84772056e-01 6.59539700e-01
7.42065966e-01 -6.38132513e-01 1.02098370e+00 1.05568103e-01
-4.46606070e-01 -1.37464508e-01 -1.68154344e-01 -3.08512268e-03
3.77293706e-01 -1.02814853e+00 8.03699568e-02 4.53200221e-01
-9.82847661e-02 3.75272542e-01 6.79062247e-01 -4.83155161e-01
-1.47185290e+00 -7.67317176e-01 9.62639511e-01 -1.88946113e-01
5.51352620e-01 4.81657125e-02 -5.13889849e-01 5.82780659e-01
4.94297951e-01 -6.87560916e-01 6.43766940e-01 2.06096680e-03
3.13447207e-01 -3.65808725e-01 -1.29105246e+00 7.72266150e-01
8.72309625e-01 1.26087308e-01 -1.01113558e+00 3.90870750e-01
1.07405961e+00 3.39729398e-01 -6.63031757e-01 6.11584723e-01
3.88085544e-01 -8.66351902e-01 8.74409258e-01 -2.97558993e-01
3.57596636e-01 -4.66783494e-01 3.16908509e-01 -1.30325222e+00
-6.47434413e-01 -6.52044833e-01 -1.11755490e-01 8.13586652e-01
2.67561704e-01 -1.02638602e+00 7.73253918e-01 8.29271972e-01
-4.90267098e-01 -4.52186823e-01 -3.71322036e-01 -8.64097834e-01
-1.58348177e-02 -5.73318660e-01 7.69598186e-01 1.02889740e+00
8.56030047e-01 7.11742520e-01 5.19608110e-02 -1.06371030e-01
2.62920409e-01 -6.55435845e-02 8.55740309e-01 -1.53157079e+00
7.55623132e-02 -4.44563478e-01 -7.55475163e-01 -7.37296760e-01
1.10969506e-01 -2.94185013e-01 -1.53270796e-01 -2.00267220e+00
5.10355495e-02 -3.86608362e-01 -4.67317730e-01 7.08820641e-01
1.45399407e-01 1.81960195e-01 4.65200543e-02 1.27923295e-01
-4.21273351e-01 4.93211329e-01 1.49224341e+00 -2.08383501e-01
-4.47462678e-01 2.24299468e-02 -1.05539024e+00 1.11147869e+00
1.28441572e+00 1.49898887e-01 -5.87137699e-01 -2.25740254e-01
3.28677893e-01 -7.41474424e-03 -4.55676839e-02 -1.54049051e+00
5.06303787e-01 -3.22115660e-01 4.13992196e-01 -3.80949914e-01
9.52900052e-02 -1.07618165e+00 1.13857910e-01 8.64902914e-01
1.78221449e-01 -3.17448862e-02 -1.20628057e-02 1.10477716e-01
-4.50438499e-01 -1.06637754e-01 4.87255424e-01 -3.36972028e-01
-1.30149698e+00 -2.28969797e-01 -8.60925496e-01 -4.99614805e-01
1.54621351e+00 -3.91306162e-01 -3.95196348e-01 -5.09448171e-01
-3.21147174e-01 4.26410079e-01 2.41677001e-01 8.88158858e-01
6.82270229e-01 -1.22459674e+00 -2.86606550e-01 3.83628637e-01
1.82871401e-01 -3.01378705e-02 -1.03589818e-01 3.36460143e-01
-6.99372470e-01 8.48801136e-01 -6.13875389e-01 -1.16439581e-01
-7.61845529e-01 9.59492028e-01 -1.54735848e-01 -6.68006092e-02
-7.44357049e-01 1.88395113e-01 3.46018881e-01 -5.59353948e-01
2.44259506e-01 -2.64782697e-01 -6.69004977e-01 -3.63084793e-01
6.55311406e-01 4.60507333e-01 -3.22379172e-01 -2.40776598e-01
-1.15326837e-01 5.50597250e-01 1.46594524e-01 -7.52441585e-02
1.23249066e+00 -1.61526293e-01 -1.46801546e-01 5.22183597e-01
5.16424239e-01 -5.58691919e-01 -4.45538282e-01 2.17327297e-01
2.16328427e-01 1.81923598e-01 1.51074246e-01 -1.10317421e+00
-9.45029438e-01 8.15256357e-01 5.34042001e-01 3.93842459e-01
1.19344473e+00 -1.44110173e-02 1.14357829e+00 8.73865604e-01
1.03874719e+00 -1.43651080e+00 4.29853857e-01 6.59433663e-01
7.92904615e-01 -1.23333681e+00 -2.77133584e-01 -2.07548767e-01
-8.81603599e-01 1.13002884e+00 7.73101270e-01 7.70259649e-02
9.46954250e-01 2.30215266e-01 2.89312690e-01 -3.65318596e-01
-5.77139616e-01 -4.19678301e-01 -1.09770320e-01 1.00015008e+00
4.31407630e-01 1.09581478e-01 -4.50804949e-01 7.18050003e-01
-4.83363211e-01 6.47446692e-01 3.43259007e-01 1.10023737e+00
-1.07058811e+00 -1.00776160e+00 -4.27123696e-01 3.69596124e-01
-1.87633127e-01 3.10495585e-01 -3.94155204e-01 7.27302849e-01
5.00807285e-01 1.34025896e+00 1.35084257e-01 -6.55164421e-01
9.44403931e-02 3.75874564e-02 2.45475650e-01 -4.42445606e-01
-2.56977201e-01 -5.38085759e-01 1.05435856e-01 -3.25391203e-01
-5.57258785e-01 -1.71939492e-01 -1.87791955e+00 -2.85649300e-01
3.05405427e-02 2.51191854e-01 1.03109074e+00 1.03086889e+00
5.46654940e-01 6.40631914e-01 6.49321377e-01 -6.41420066e-01
1.95543632e-01 -1.15785563e+00 -3.26227278e-01 1.49188358e-02
-3.40562314e-01 -6.99964941e-01 -1.36326738e-02 9.01588425e-02] | [9.159613609313965, 6.386629104614258] |
dcff279d-4e38-40a8-9bc1-d4f06a8236e6 | adversarial-transfer-learning-for-punctuation | 2004.00248 | null | https://arxiv.org/abs/2004.00248v1 | https://arxiv.org/pdf/2004.00248v1.pdf | Adversarial Transfer Learning for Punctuation Restoration | Previous studies demonstrate that word embeddings and part-of-speech (POS) tags are helpful for punctuation restoration tasks. However, two drawbacks still exist. One is that word embeddings are pre-trained by unidirectional language modeling objectives. Thus the word embeddings only contain left-to-right context information. The other is that POS tags are provided by an external POS tagger. So computation cost will be increased and incorrect predicted tags may affect the performance of restoring punctuation marks during decoding. This paper proposes adversarial transfer learning to address these problems. A pre-trained bidirectional encoder representations from transformers (BERT) model is used to initialize a punctuation model. Thus the transferred model parameters carry both left-to-right and right-to-left representations. Furthermore, adversarial multi-task learning is introduced to learn task invariant knowledge for punctuation prediction. We use an extra POS tagging task to help the training of the punctuation predicting task. Adversarial training is utilized to prevent the shared parameters from containing task specific information. We only use the punctuation predicting task to restore marks during decoding stage. Therefore, it will not need extra computation and not introduce incorrect tags from the POS tagger. Experiments are conducted on IWSLT2011 datasets. The results demonstrate that the punctuation predicting models obtain further performance improvement with task invariant knowledge from the POS tagging task. Our best model outperforms the previous state-of-the-art model trained only with lexical features by up to 9.2% absolute overall F_1-score on test set. | ['Jian-Hua Tao', 'Zhengkun Tian', 'Jiangyan Yi', 'Cunhang Fan', 'Ye Bai'] | 2020-04-01 | null | null | null | null | ['punctuation-restoration'] | ['natural-language-processing'] | [ 7.46799335e-02 8.74551162e-02 -2.19023138e-01 -3.00662488e-01
-9.67917800e-01 -5.28541982e-01 2.27241665e-01 9.12222043e-02
-7.15731621e-01 6.53309703e-01 3.45213145e-01 -3.56649846e-01
4.98451144e-01 -6.59367681e-01 -8.87333393e-01 -6.54208243e-01
6.50061443e-02 6.42328784e-02 3.16797316e-01 -3.40228647e-01
-5.99862412e-02 4.14168164e-02 -9.64044034e-01 4.17268217e-01
1.21208644e+00 7.35124946e-01 7.04875588e-01 3.14023376e-01
-6.24872625e-01 5.24145544e-01 -7.56657600e-01 -4.59119648e-01
1.26060024e-01 -1.35757402e-01 -5.24259448e-01 -2.62075990e-01
3.11972518e-02 -7.03280047e-02 -4.26791221e-01 1.22800577e+00
7.03547180e-01 -6.74729496e-02 4.52300340e-01 -8.90877366e-01
-1.23262596e+00 8.95285010e-01 -3.03932637e-01 1.85038060e-01
-1.08426549e-01 -1.41198009e-01 1.01898324e+00 -1.17674291e+00
2.31349334e-01 9.56868291e-01 7.68180192e-01 6.88643813e-01
-8.25427115e-01 -7.14807749e-01 2.95266420e-01 3.09605449e-01
-1.47678804e+00 -6.44480139e-02 6.90907300e-01 -1.45577908e-01
1.13666928e+00 5.74803650e-02 1.07348077e-01 1.06991065e+00
6.00758672e-01 8.29629362e-01 9.81166184e-01 -4.45460796e-01
-1.87763292e-02 1.66478693e-01 7.35151097e-02 6.42840147e-01
4.05963324e-02 -1.39106408e-01 -3.65714818e-01 1.40375227e-01
5.51382720e-01 1.28655151e-01 -2.50713140e-01 4.45375685e-03
-9.82557178e-01 5.47249496e-01 4.19596344e-01 6.44074559e-01
-1.27374202e-01 5.79342581e-02 4.86086816e-01 2.10895538e-01
4.82249439e-01 2.25584686e-01 -8.71964455e-01 -1.76050439e-01
-6.20620608e-01 -3.63159865e-01 3.13927382e-01 1.03728712e+00
9.73259568e-01 4.18859810e-01 -3.28401029e-01 1.15299499e+00
3.37982655e-01 5.80223680e-01 1.04089355e+00 -7.88050219e-02
7.13989735e-01 4.01066363e-01 -7.23851547e-02 -5.53124964e-01
-1.78283244e-01 -6.23256385e-01 -6.55488491e-01 -1.39430851e-01
6.48489818e-02 -4.25103396e-01 -1.26978326e+00 1.61894357e+00
-1.67504579e-01 4.90702301e-01 2.56574154e-01 6.77991807e-01
5.73228538e-01 9.44432557e-01 4.50414896e-01 1.06044106e-01
1.50482512e+00 -1.13421822e+00 -1.09571445e+00 -5.94816685e-01
9.66117859e-01 -1.14718544e+00 1.22712517e+00 1.19861858e-02
-8.70019555e-01 -8.40697289e-01 -1.16834021e+00 -1.96926862e-01
-5.87980926e-01 2.30988234e-01 1.68834239e-01 5.46605110e-01
-7.14841008e-01 5.34640014e-01 -8.53523672e-01 8.99243876e-02
2.06257015e-01 3.69947135e-01 -2.71567971e-01 -4.02424932e-02
-1.60718286e+00 1.01255214e+00 5.37109256e-01 7.11318552e-02
-5.58360875e-01 -6.90658271e-01 -1.09711659e+00 2.70822108e-01
-3.99655849e-03 -7.38860741e-02 1.23526025e+00 -1.04337227e+00
-1.33535433e+00 5.13189197e-01 -2.92005479e-01 -3.55290592e-01
9.80260298e-02 -3.81495982e-01 -7.87702024e-01 -2.06726313e-01
1.66057885e-01 4.09535259e-01 6.23485982e-01 -1.08513522e+00
-7.69452691e-01 -7.66882999e-03 -3.53589684e-01 2.89199889e-01
-7.41693735e-01 1.03981122e-02 -5.58073163e-01 -1.06240332e+00
2.90049277e-02 -7.96943605e-01 -8.27329755e-02 -4.60342318e-01
-9.03183296e-02 -4.34586346e-01 8.89838576e-01 -9.38174129e-01
1.44421387e+00 -2.26219010e+00 -2.61202842e-01 -2.62643725e-01
-3.69461685e-01 8.15528095e-01 -4.13911492e-01 2.63074249e-01
-2.51084745e-01 5.19868612e-01 -1.28493652e-01 -7.40337729e-01
7.71046523e-03 5.47929645e-01 -4.78784412e-01 1.70301363e-01
5.91185749e-01 9.55842853e-01 -9.15063977e-01 -4.05950844e-01
1.59806356e-01 5.94989359e-01 -4.75224614e-01 3.40232611e-01
7.26080537e-02 1.13515317e-01 -3.08378279e-01 4.36525822e-01
9.06129479e-01 2.31304169e-01 1.09986283e-01 -3.63944122e-03
-8.99616554e-02 8.03888440e-01 -1.01667655e+00 1.83123147e+00
-9.00156260e-01 7.31704757e-02 -4.19884413e-01 -8.74764383e-01
1.00354147e+00 5.95792413e-01 8.74445140e-02 -8.08017910e-01
1.61025181e-01 2.09029853e-01 6.18834868e-02 -3.50031406e-01
5.82483649e-01 -3.83719236e-01 -4.28882301e-01 8.06029364e-02
3.17492366e-01 3.46794873e-01 -2.34878376e-01 -1.95574507e-01
1.07005715e+00 2.27024660e-01 -2.32562143e-02 -9.05623287e-02
5.87662816e-01 -1.84867695e-01 1.26012611e+00 2.69229233e-01
-1.57090425e-01 7.72144556e-01 1.75612256e-01 -1.74344555e-01
-1.01102781e+00 -1.26594687e+00 -1.87365368e-01 1.10392606e+00
9.33246538e-02 -4.84033674e-01 -7.21518517e-01 -1.01253402e+00
-2.30242416e-01 8.36307824e-01 -5.16879916e-01 -3.63942176e-01
-9.35791433e-01 -6.44276798e-01 6.89351082e-01 8.64462018e-01
4.15564477e-01 -1.30424333e+00 3.57851326e-01 4.74016726e-01
-2.13867038e-01 -1.22169006e+00 -9.65333819e-01 4.71252352e-01
-8.55009198e-01 -5.87838113e-01 -7.36980379e-01 -1.48654473e+00
7.92005599e-01 5.89181595e-02 6.65608704e-01 -4.30103615e-02
2.52638847e-01 -1.70399830e-01 -6.88573182e-01 -3.46295178e-01
-4.50032055e-01 2.06735656e-01 1.44537687e-01 -1.05852030e-01
6.30218446e-01 -4.71963048e-01 -2.63897419e-01 3.29965591e-01
-9.81998622e-01 -2.61467665e-01 7.85622060e-01 1.00043023e+00
8.18522155e-01 -1.40006781e-01 8.06366384e-01 -7.17278957e-01
3.55116963e-01 -3.76642555e-01 -2.76423335e-01 1.51402444e-01
-4.45020825e-01 2.64640003e-01 1.12147236e+00 -5.98200023e-01
-1.13948607e+00 -1.73372310e-02 -4.23194796e-01 -4.74361956e-01
1.05711959e-01 4.69334573e-01 -6.50075793e-01 2.46394426e-01
3.38015676e-01 4.26294655e-01 -3.83999318e-01 -7.36265361e-01
1.85877562e-01 9.87508118e-01 3.82539690e-01 -3.18694472e-01
8.94160330e-01 -8.09039548e-02 -5.51869392e-01 -5.01135111e-01
-8.69600177e-01 -3.25731397e-01 -6.54284537e-01 3.11226726e-01
8.95787776e-01 -1.11303926e+00 -1.20699219e-01 5.45954406e-01
-1.26961601e+00 -1.83225438e-01 -1.43993571e-01 3.66580635e-01
-1.66837841e-01 5.91268718e-01 -6.92632139e-01 -4.97869104e-01
-5.10937512e-01 -1.12536240e+00 7.17346013e-01 3.63897562e-01
1.38160020e-01 -1.10484147e+00 -9.03862193e-02 1.96039841e-01
3.66833597e-01 -4.09510553e-01 8.97855520e-01 -8.99469316e-01
-3.71971905e-01 -1.43581167e-01 6.42599817e-03 9.82389390e-01
3.88987511e-01 -5.44960499e-01 -9.40143764e-01 -1.41580030e-01
-1.51432216e-01 -1.30336182e-02 7.91795373e-01 -1.31704435e-01
1.20225477e+00 -2.82940120e-01 -2.53307194e-01 5.33996463e-01
1.34462047e+00 3.00960273e-01 1.00382042e+00 5.39311111e-01
9.05819535e-01 5.11865355e-02 7.17528939e-01 4.43782136e-02
4.61222708e-01 4.57583219e-01 2.37061843e-01 7.50336647e-02
-5.67841470e-01 -7.52631724e-01 8.07474077e-01 1.37340724e+00
4.51604098e-01 -3.10174406e-01 -9.30964172e-01 8.97259951e-01
-1.61864936e+00 -5.97341120e-01 -1.04489634e-02 2.24536467e+00
1.32330906e+00 3.35102111e-01 -6.22539520e-01 1.48332208e-01
9.22455728e-01 1.41755342e-01 -1.08521178e-01 -7.78232694e-01
-4.68641743e-02 6.27321303e-01 5.98637879e-01 5.72286189e-01
-1.21870661e+00 1.48028052e+00 4.61657715e+00 1.06797671e+00
-1.16758204e+00 4.51603174e-01 1.96457043e-01 3.84223819e-01
-4.98787016e-01 1.08872093e-01 -1.16549706e+00 8.48976910e-01
1.06564808e+00 9.71985683e-02 5.43744639e-02 8.13940287e-01
4.63812277e-02 2.79991657e-01 -7.00678766e-01 6.77412987e-01
2.09289476e-01 -9.12841022e-01 7.09241629e-02 -2.11299241e-01
6.20170474e-01 1.75474316e-01 7.10441843e-02 7.45840013e-01
1.45669058e-01 -7.86589921e-01 6.42295659e-01 3.37788165e-01
7.92997360e-01 -1.06853271e+00 1.00751281e+00 3.15766662e-01
-1.21197629e+00 1.92110941e-01 -5.68562746e-01 6.21121284e-03
2.12666675e-01 5.83049476e-01 -1.01198304e+00 5.10948956e-01
4.65447158e-01 5.72581589e-01 -3.97522062e-01 9.14810956e-01
-8.01357388e-01 8.74331832e-01 -9.23597738e-02 -1.14721373e-01
3.29128563e-01 3.43149491e-02 2.88558573e-01 1.40817130e+00
5.07335842e-01 -2.56342024e-01 2.51493007e-01 3.81387204e-01
-3.32479268e-01 1.74287990e-01 -2.11907774e-01 1.07297651e-01
6.95077360e-01 1.08966017e+00 -2.74610043e-01 -1.41435117e-01
-5.82714975e-01 1.39846039e+00 4.59834605e-01 1.32910132e-01
-9.79500592e-01 -8.44119370e-01 9.34144855e-01 7.32825324e-02
7.30113328e-01 -4.08202499e-01 -4.65632498e-01 -1.22279394e+00
1.42551392e-01 -4.26876754e-01 4.45346273e-02 -6.58515751e-01
-1.27135980e+00 6.42884195e-01 -6.09632373e-01 -1.35099816e+00
8.90327245e-02 -5.48086107e-01 -8.33719969e-01 1.21201932e+00
-2.26363730e+00 -1.36981213e+00 1.80836037e-01 2.35388458e-01
6.38770282e-01 -1.42936274e-01 1.07401216e+00 6.10605896e-01
-7.07017362e-01 1.16424096e+00 2.51950592e-01 7.42329001e-01
1.07775903e+00 -1.38057113e+00 4.93614405e-01 1.20580900e+00
-1.66804753e-02 4.80926037e-01 2.89721757e-01 -7.67659247e-01
-1.00638986e+00 -1.44775844e+00 1.70263982e+00 -3.08353096e-01
7.82277584e-01 -5.40712655e-01 -1.16386914e+00 8.86997342e-01
3.03797930e-01 3.22862685e-01 8.67675602e-01 3.31387445e-02
-3.01469803e-01 -1.66455388e-01 -8.49875808e-01 6.21858895e-01
7.72271693e-01 -5.17509580e-01 -1.14791930e+00 3.01651098e-02
8.91495705e-01 -4.36029375e-01 -7.55790591e-01 3.49980295e-01
-5.78488363e-03 -3.12642813e-01 7.10972667e-01 -4.42021310e-01
4.38086867e-01 -5.29446304e-01 -1.46784678e-01 -1.59591269e+00
-4.15349424e-01 -4.18919683e-01 1.68994680e-01 1.73643923e+00
6.72788858e-01 -4.93423760e-01 4.92597103e-01 1.92862645e-01
-8.66717398e-01 -6.85029864e-01 -1.07058358e+00 -9.83811677e-01
5.00768185e-01 -4.32974398e-01 5.17356098e-01 9.92250681e-01
9.94371846e-02 4.21429366e-01 -3.10180873e-01 5.11751533e-01
1.79039538e-01 -3.06941926e-01 1.38672069e-01 -7.91034579e-01
5.87438270e-02 1.67432830e-01 -3.58780175e-01 -1.09730327e+00
4.15492684e-01 -1.30491424e+00 2.47219533e-01 -1.46101201e+00
-1.72859758e-01 -8.78108501e-01 -8.70607436e-01 1.13341916e+00
-5.66431880e-01 1.79226190e-01 2.08788335e-01 4.72440310e-02
-3.28171700e-01 8.12814832e-01 1.21123910e+00 -3.36349756e-02
-1.16035096e-01 -2.00699702e-01 -6.04068100e-01 6.73242629e-01
1.23107219e+00 -8.61152470e-01 -1.18502691e-01 -8.90236557e-01
-3.31707502e-04 -2.76491702e-01 1.84742175e-02 -1.02797365e+00
2.05748588e-01 9.01109502e-02 4.57003653e-01 -4.06638116e-01
3.33547026e-01 -8.14807296e-01 -6.03790998e-01 3.35560381e-01
1.14777789e-01 1.35778889e-01 5.51060975e-01 3.38306069e-01
-1.76722199e-01 -5.44959903e-01 6.82128072e-01 1.40491381e-01
-9.04507756e-01 1.89521044e-01 -5.12453318e-01 1.46806940e-01
8.59922647e-01 -6.50364161e-02 -1.20068207e-01 1.90829054e-01
-6.99480057e-01 2.11916506e-01 2.26381466e-01 8.29339325e-01
4.50789332e-01 -1.62920094e+00 -4.99148726e-01 4.32023168e-01
1.33143112e-01 -4.95021930e-03 2.79224545e-01 3.37431848e-01
-2.15544432e-01 3.04254591e-01 -1.54438436e-01 -1.51675522e-01
-1.06505096e+00 5.58900058e-01 3.01987559e-01 -3.60624254e-01
-3.46615285e-01 8.59390736e-01 4.43738922e-02 -6.07276976e-01
1.54312342e-01 -4.87444013e-01 -1.44614369e-01 1.76596418e-02
5.92672467e-01 -1.31750956e-01 3.64496857e-01 -5.91808200e-01
-4.58472133e-01 4.56306934e-01 -4.62251723e-01 1.26464680e-01
1.34061456e+00 -1.19090371e-01 1.81983650e-01 2.17924759e-01
1.33305943e+00 2.70373166e-01 -1.14620614e+00 -4.54947203e-01
-2.00567972e-02 -1.18176997e-01 2.49373630e-01 -1.06438363e+00
-8.53914380e-01 1.19073224e+00 6.97927117e-01 -1.69760868e-01
8.97350132e-01 -1.89968139e-01 1.29540098e+00 8.38622451e-02
2.37167850e-01 -1.28931415e+00 7.35750645e-02 1.02751338e+00
6.56082809e-01 -1.06347656e+00 -5.54249108e-01 -4.59610075e-01
-8.92400026e-01 9.39775825e-01 8.27469826e-01 -2.99858093e-01
6.42027438e-01 5.23806572e-01 2.82182008e-01 6.19927108e-01
-3.49593788e-01 -3.67345244e-01 1.74201474e-01 6.62917137e-01
5.32871664e-01 1.20951787e-01 -5.16312897e-01 1.21575046e+00
-1.93526164e-01 -2.51665771e-01 4.92501199e-01 9.65328634e-01
-6.13012671e-01 -1.79906714e+00 -2.09383726e-01 5.82925342e-02
-7.44489908e-01 -5.97848654e-01 6.14478886e-02 3.92026573e-01
4.56211776e-01 8.41524005e-01 2.33256996e-01 -3.81540745e-01
4.00997609e-01 4.86643463e-01 1.48452640e-01 -9.46974158e-01
-7.29910016e-01 4.82236110e-02 -8.90849158e-02 -1.40957117e-01
1.83534548e-01 -5.58826864e-01 -1.80182123e+00 4.20199260e-02
-3.97978157e-01 3.31812382e-01 6.33414567e-01 8.72946322e-01
6.33735538e-01 8.97614121e-01 7.88801730e-01 -4.76066351e-01
-5.74453831e-01 -1.23564172e+00 -5.04249573e-01 1.89484835e-01
1.47383451e-01 -3.88231397e-01 -2.59022444e-01 9.83396992e-02] | [14.115266799926758, 7.292847156524658] |
a489e46a-b91b-450f-8fd0-676bdf5e8e18 | complete-solution-for-vehicle-re-id-in | 2212.04126 | null | https://arxiv.org/abs/2212.04126v1 | https://arxiv.org/pdf/2212.04126v1.pdf | Complete Solution for Vehicle Re-ID in Surround-view Camera System | Vehicle re-identification (Re-ID) is a critical component of the autonomous driving perception system, and research in this area has accelerated in recent years. However, there is yet no perfect solution to the vehicle re-identification issue associated with the car's surround-view camera system. Our analysis identifies two significant issues in the aforementioned scenario: i) It is difficult to identify the same vehicle in many picture frames due to the unique construction of the fisheye camera. ii) The appearance of the same vehicle when seen via the surround vision system's several cameras is rather different. To overcome these issues, we suggest an integrative vehicle Re-ID solution method. On the one hand, we provide a technique for determining the consistency of the tracking box drift with respect to the target. On the other hand, we combine a Re-ID network based on the attention mechanism with spatial limitations to increase performance in situations involving multiple cameras. Finally, our approach combines state-of-the-art accuracy with real-time performance. We will soon make the source code and annotated fisheye dataset available. | ['Jing Song', 'Xiaoquan Wang', 'Fan Wang', 'Tianhao Xu', 'Zizhang Wu'] | 2022-12-08 | null | null | null | null | ['vehicle-re-identification'] | ['computer-vision'] | [-2.48508677e-01 -2.98815995e-01 3.36414874e-02 -3.92957449e-01
-5.02225637e-01 -6.22520626e-01 5.62343776e-01 -2.12227702e-01
-5.76454461e-01 3.15197021e-01 -2.80486345e-01 -2.70454466e-01
1.94356769e-01 -2.59613723e-01 -7.33354092e-01 -6.59461260e-01
4.23215777e-01 1.69528723e-01 8.63376617e-01 -2.21554130e-01
4.94671643e-01 6.81640327e-01 -1.98877311e+00 -2.11584270e-01
4.60243285e-01 8.73926640e-01 6.36281013e-01 6.61899507e-01
1.71255499e-01 5.35221457e-01 -4.14848387e-01 -4.40899432e-01
3.93448979e-01 -1.74662039e-01 -2.74136156e-01 1.37146581e-02
9.20411408e-01 -6.31945252e-01 -3.72061968e-01 1.50058532e+00
1.99751571e-01 8.35270807e-02 3.81549001e-01 -1.79489291e+00
-4.13914144e-01 -6.91428185e-02 -7.70951509e-01 5.52125394e-01
2.13557318e-01 8.23823214e-02 3.36610258e-01 -9.04697239e-01
5.86512089e-01 1.20320153e+00 7.80708969e-01 7.01640248e-01
-7.70877361e-01 -9.25049245e-01 2.09307507e-01 7.67969072e-01
-1.63467610e+00 -7.20430970e-01 7.50376701e-01 -5.02615690e-01
6.23784542e-01 -1.75197311e-02 3.81286412e-01 8.79916370e-01
2.82789469e-01 3.65367919e-01 5.37217796e-01 -1.91079155e-01
1.84974575e-03 5.30185163e-01 3.82865340e-01 2.93007940e-01
4.93926913e-01 4.25657362e-01 -1.48322478e-01 2.36201897e-01
3.42502326e-01 1.89942688e-01 -4.24898528e-02 -7.77289689e-01
-9.16838765e-01 6.07272625e-01 8.61879811e-03 1.34965092e-01
1.57067418e-01 1.32205263e-01 3.80196273e-01 1.92101181e-01
2.26122975e-01 1.54692262e-01 2.39208061e-03 -1.38239264e-01
-7.10250497e-01 2.71135151e-01 4.16644096e-01 1.29211318e+00
8.23269546e-01 4.04252624e-03 5.22246420e-01 4.49370921e-01
3.82862985e-01 6.33327425e-01 2.20297068e-01 -9.68210638e-01
5.30215979e-01 3.07158321e-01 2.57976502e-01 -1.24506998e+00
-2.84088731e-01 -3.26693282e-02 -4.32977170e-01 8.17228019e-01
6.52699351e-01 -9.70331430e-02 -4.15473521e-01 1.56354916e+00
3.28215420e-01 4.30599630e-01 2.30112582e-01 1.05731952e+00
8.16623449e-01 3.56598586e-01 -9.83439684e-02 5.04883938e-02
1.47967339e+00 -1.01352596e+00 -8.68618906e-01 -4.35647607e-01
5.50051451e-01 -7.55636632e-01 2.97629595e-01 9.43426266e-02
-7.32191682e-01 -8.92089903e-01 -1.36947083e+00 1.86965331e-01
-5.50138593e-01 7.46436566e-02 -4.30043526e-02 7.83653736e-01
-1.16187465e+00 1.06902078e-01 -5.57572782e-01 -8.22624505e-01
-1.10312417e-01 2.70678788e-01 -6.99996352e-01 -1.13333367e-01
-1.05329704e+00 1.42830110e+00 2.44899526e-01 1.15835212e-01
-7.02453673e-01 -4.70438749e-01 -8.53388190e-01 -1.73242748e-01
5.14222562e-01 -3.37559313e-01 1.09910488e+00 -1.11151540e+00
-1.22289455e+00 6.63276851e-01 -4.65063840e-01 -5.09504735e-01
6.94823503e-01 -1.06905460e-01 -8.83701444e-01 -9.85642970e-02
2.08713129e-01 6.08788550e-01 9.70854044e-01 -1.53553486e+00
-1.15587544e+00 -5.32720327e-01 -2.18373522e-01 2.59125203e-01
9.32237953e-02 2.19700843e-01 -5.48398435e-01 -1.30527124e-01
-7.99289197e-02 -1.08948863e+00 1.49109617e-01 9.52350944e-02
2.51189526e-02 -2.59400457e-01 1.49772763e+00 -4.29473937e-01
1.07907224e+00 -2.64747524e+00 -3.40029478e-01 -6.60739541e-02
3.65196228e-01 5.64535856e-01 -9.81533080e-02 2.45896325e-01
-2.71118969e-01 -2.01189235e-01 3.61225873e-01 -3.26938033e-01
-4.18938845e-01 -3.84604298e-02 -2.67589748e-01 7.87885427e-01
-1.28510743e-01 5.05388021e-01 -9.40908849e-01 -3.12650830e-01
5.16365826e-01 2.57718205e-01 -2.83729173e-02 8.21494311e-02
5.08623481e-01 1.89245179e-01 -1.67702377e-01 3.66717577e-01
1.20389032e+00 2.68257856e-01 -1.53973714e-01 -3.11167717e-01
-6.31056368e-01 -3.20687830e-01 -1.41600621e+00 1.08209240e+00
-9.89124272e-03 1.23956239e+00 1.49869099e-01 -7.15151489e-01
9.49357688e-01 1.87012687e-01 2.59538054e-01 -6.69121146e-01
1.25262558e-01 2.36675918e-01 6.11986704e-02 -6.18406415e-01
9.71291125e-01 1.95226029e-01 1.94478080e-01 2.11839274e-01
-2.45095298e-01 2.76282847e-01 1.34059444e-01 1.93169508e-02
7.68800616e-01 -1.27793446e-01 3.37092847e-01 -1.61738232e-01
7.06332386e-01 6.32682294e-02 5.49563944e-01 6.92746758e-01
-8.32935989e-01 6.54289603e-01 1.37576252e-01 -6.42994821e-01
-1.15841639e+00 -6.73659503e-01 -1.49701461e-01 6.81816876e-01
9.54708338e-01 -1.53502941e-01 -7.73128092e-01 -3.84081662e-01
1.37346894e-01 7.63506293e-01 -6.00049078e-01 -1.46449268e-01
-6.22809887e-01 -1.35935932e-01 5.68266869e-01 6.04658842e-01
6.32048428e-01 -4.16689008e-01 -1.01818705e+00 4.15870734e-02
-1.09517023e-01 -1.41382694e+00 -7.65696228e-01 -1.06036015e-01
-4.78244424e-01 -1.26829505e+00 -5.22657871e-01 -7.20796704e-01
6.92021251e-01 1.35888445e+00 6.37467802e-01 6.37010187e-02
7.70831630e-02 4.21823829e-01 -2.19582558e-01 -5.14860690e-01
-6.34080827e-01 -3.55339468e-01 2.72111088e-01 9.98663008e-02
7.86126792e-01 1.37710929e-01 -5.37201405e-01 8.87255549e-01
-3.89498442e-01 -1.68370828e-01 1.97919592e-01 3.38932782e-01
2.77659863e-01 2.70437926e-01 3.27963501e-01 -3.18093687e-01
3.24164480e-01 -4.13714170e-01 -1.05276644e+00 1.88367724e-01
-6.86879694e-01 -2.63646245e-01 4.59838122e-01 -4.15440202e-01
-9.64332819e-01 3.62896383e-01 1.38228372e-01 -6.31760001e-01
-4.14250493e-01 -2.47140184e-01 -1.63802192e-01 -3.55526865e-01
4.03277993e-01 1.28824010e-01 4.67873871e-01 -2.16839388e-01
2.30953142e-01 8.62265468e-01 6.84154510e-01 3.53680998e-01
7.63620794e-01 6.52964890e-01 -9.50183719e-02 -9.77051497e-01
-1.99040994e-01 -8.27136159e-01 -6.85697675e-01 -5.79046667e-01
1.08106852e+00 -1.13640428e+00 -8.68454158e-01 7.32730567e-01
-1.42031944e+00 2.61999905e-01 1.45323738e-01 7.05942512e-01
-2.78040618e-01 6.86328471e-01 8.56284723e-02 -8.49159300e-01
1.65278718e-01 -1.39232290e+00 8.87239814e-01 3.77208471e-01
-5.29633202e-02 -6.91857874e-01 1.47742599e-01 3.19820255e-01
2.62042522e-01 -1.53543979e-01 3.77798170e-01 -5.28346598e-01
-6.35401547e-01 -2.59455949e-01 -4.49948162e-01 2.64771312e-01
1.84956528e-02 1.39389679e-01 -1.00783551e+00 -3.81716758e-01
9.82846394e-02 4.78497952e-01 7.44890034e-01 3.68211865e-01
5.54835737e-01 1.73485383e-01 -7.98828006e-01 4.34059948e-01
1.37770271e+00 7.30455935e-01 6.65709257e-01 5.54243505e-01
6.59760833e-01 8.14773202e-01 8.08841944e-01 9.79696959e-03
7.49363661e-01 1.01291847e+00 6.54345453e-01 8.88322294e-03
-3.05220902e-01 -1.50264829e-01 3.55382055e-01 2.83005536e-01
1.44036025e-01 -2.31495202e-01 -8.90908957e-01 6.28428578e-01
-1.87452996e+00 -1.43141556e+00 -4.00658518e-01 2.38304830e+00
-2.61415064e-01 -2.19883001e-03 2.59336889e-01 -5.82231730e-02
1.14835012e+00 -4.99513187e-02 -5.77345788e-01 -2.58534342e-01
8.21803231e-03 -1.19721270e+00 1.02174413e+00 5.14432967e-01
-1.13618124e+00 7.02288985e-01 6.71891594e+00 4.52968836e-01
-1.18958044e+00 8.94282386e-02 2.27829844e-01 2.49423996e-01
3.47492024e-02 -4.57281340e-03 -1.41509771e+00 7.45191753e-01
7.67538488e-01 -2.02497229e-01 2.72699744e-01 9.59571481e-01
1.53606072e-01 -4.53976601e-01 -1.20113337e+00 1.34900975e+00
6.75209165e-01 -9.99175310e-01 -4.80082572e-01 2.03571469e-01
3.83940250e-01 2.17852563e-01 6.22359999e-02 -8.35964680e-02
-6.81123734e-02 -5.29035032e-01 9.61130023e-01 3.47711176e-01
6.58635080e-01 -7.48553514e-01 9.15692389e-01 4.02552158e-01
-1.45768535e+00 -2.19575807e-01 -4.57785219e-01 1.64428473e-01
1.47834972e-01 -9.98805538e-02 -8.11499417e-01 3.01169038e-01
9.36795592e-01 7.84458756e-01 -8.00619841e-01 1.12581325e+00
3.54253083e-01 -8.61647800e-02 -1.37412831e-01 9.03507918e-02
3.28514986e-02 -2.41812319e-01 7.59076774e-01 8.48253012e-01
4.43278551e-01 -1.70512065e-01 -2.10499316e-01 6.69073164e-01
4.75037724e-01 -4.35696632e-01 -1.16551197e+00 4.61924702e-01
5.95957041e-01 1.04479146e+00 -7.11440563e-01 -9.75307077e-02
-8.45264792e-01 6.44003153e-01 4.00219038e-02 4.01038289e-01
-1.03442693e+00 -2.90569037e-01 9.42799032e-01 2.29555041e-01
4.97429043e-01 -2.65359223e-01 2.25991160e-02 -8.15412283e-01
-4.03222106e-02 -4.82970744e-01 9.43867024e-03 -9.55092430e-01
-9.82891798e-01 7.15792418e-01 2.70848662e-01 -1.68424475e+00
-3.16250592e-01 -5.05816758e-01 -4.47330087e-01 6.83610678e-01
-1.68152738e+00 -8.55775416e-01 -5.32541037e-01 4.21984613e-01
6.65841103e-01 -5.20449698e-01 3.72734457e-01 4.49889570e-01
-5.47573507e-01 7.38224149e-01 4.45047915e-01 -9.16185677e-02
1.00640988e+00 -7.96761096e-01 5.39754927e-01 1.22796428e+00
-3.38744730e-01 3.01305145e-01 8.28522027e-01 -5.40228605e-01
-1.43794405e+00 -9.68793929e-01 9.10156250e-01 -7.82384694e-01
5.40878415e-01 -2.77898997e-01 -8.47671390e-01 6.99071407e-01
2.33290493e-01 -1.47782797e-02 2.75124162e-01 -3.11348468e-01
-2.34587505e-01 -4.53147173e-01 -9.97229397e-01 5.12190580e-01
6.57687724e-01 -4.40914869e-01 -4.70623910e-01 -1.80098057e-01
2.88800001e-01 -4.30317879e-01 -2.50449568e-01 6.35017976e-02
7.82845557e-01 -1.07263935e+00 8.72454226e-01 1.06314733e-03
-2.24724025e-01 -9.14514065e-01 -1.96016952e-01 -1.04913437e+00
-4.15114313e-01 -2.81024873e-01 2.59216249e-01 1.27177179e+00
3.99411432e-02 -9.71671045e-01 5.54501772e-01 7.16804743e-01
-9.22171026e-02 -1.48608655e-01 -1.23836637e+00 -9.43768680e-01
-2.17616811e-01 -4.28384811e-01 5.63529313e-01 4.45976973e-01
-9.53590944e-02 2.90051103e-01 -5.86276412e-01 4.46814656e-01
6.36169851e-01 -1.52367249e-01 1.16121733e+00 -1.29164016e+00
3.42449963e-01 -2.69071221e-01 -8.89928818e-01 -1.11799824e+00
1.98853910e-01 -2.99872994e-01 3.23299110e-01 -9.85494196e-01
3.41325402e-01 -2.82572836e-01 1.67118814e-02 -8.86677429e-02
-4.98619536e-03 4.37023714e-02 5.47749341e-01 3.87017220e-01
-6.69347823e-01 2.62934744e-01 8.80528092e-01 -7.17465952e-02
8.26399475e-02 1.06925838e-01 -6.75927758e-01 6.83079779e-01
6.26941919e-01 -5.53669870e-01 -3.93795341e-01 -5.88158846e-01
1.10770566e-02 1.11361090e-02 6.62002921e-01 -1.16654384e+00
8.16785157e-01 9.26629677e-02 1.06412165e-01 -1.10017693e+00
3.46531689e-01 -1.41358089e+00 5.00416517e-01 2.77894557e-01
1.21195346e-01 5.87242961e-01 3.84210765e-01 8.62017453e-01
-3.32759529e-01 -5.43079197e-01 9.19881165e-01 1.80748012e-02
-1.38442147e+00 -5.99109493e-02 -7.30145276e-01 -2.13769883e-01
1.64923370e+00 -8.07295740e-01 -6.19140327e-01 -3.93336892e-01
9.62034389e-02 2.88251996e-01 7.97322989e-01 8.84664118e-01
6.07469797e-01 -1.23953235e+00 -4.98660445e-01 4.83171493e-01
4.19851303e-01 -3.59798998e-01 4.35788125e-01 7.94652045e-01
-5.06172538e-01 6.45131946e-01 -3.37036699e-01 -7.60114551e-01
-1.73260128e+00 9.51295316e-01 5.26176810e-01 4.09075439e-01
-6.69884741e-01 3.46916139e-01 4.52425510e-01 1.19095854e-02
2.16096565e-01 5.42128868e-02 -5.05858660e-01 6.35420531e-02
9.08255577e-01 8.14138412e-01 2.03198828e-02 -1.37574744e+00
-6.57094300e-01 9.44549561e-01 -1.99620709e-01 3.07377242e-02
7.61064649e-01 -7.94684589e-01 2.83978432e-01 3.04023564e-01
1.11561716e+00 -1.47116527e-01 -1.59290743e+00 -5.69099672e-02
-3.06107819e-01 -7.02241123e-01 8.89186114e-02 -1.97714925e-01
-9.22454119e-01 7.18298733e-01 1.18592119e+00 1.34378672e-01
6.65593147e-01 -1.72149837e-01 5.31599760e-01 3.49712878e-01
2.79294759e-01 -1.21370625e+00 -2.72421837e-01 5.70278347e-01
6.19720876e-01 -1.64338934e+00 -6.82285652e-02 -2.62845099e-01
-7.95132816e-01 1.27298915e+00 7.92812943e-01 6.17814399e-02
6.63470209e-01 3.08756620e-01 3.21933359e-01 1.03020966e-02
-4.69716340e-01 -1.28921404e-01 -2.62408629e-02 8.83755207e-01
-1.88693389e-01 -2.64527768e-01 1.01771474e-01 1.37057722e-01
1.74580589e-01 -2.78854430e-01 7.37760961e-01 5.70660651e-01
-4.35268432e-01 -6.82024360e-01 -7.04251230e-01 -1.12382948e-01
1.43873719e-02 2.43269220e-01 -1.75088808e-01 1.02722228e+00
2.24075496e-01 1.27264428e+00 3.74953598e-01 -6.13444328e-01
4.62654710e-01 -1.23659842e-01 1.25377908e-01 -1.73501849e-01
-1.39024213e-01 -1.36334404e-01 -2.59959668e-01 -5.29422998e-01
-5.67195833e-01 -9.19935286e-01 -8.45784187e-01 -5.23336112e-01
-5.58274806e-01 -9.91159827e-02 9.19693887e-01 8.27923656e-01
4.88013506e-01 2.26082548e-01 6.21876657e-01 -9.87878680e-01
-4.46915954e-01 -5.55294752e-01 -5.84723592e-01 2.73948401e-01
7.83945918e-01 -9.05149460e-01 -4.48828727e-01 1.45711169e-01] | [7.968259334564209, -1.2705531120300293] |
66f4bdeb-e931-48a3-be67-a409b214e4e1 | towards-controlled-and-diverse-generation-of | 2107.11781 | null | https://arxiv.org/abs/2107.11781v1 | https://arxiv.org/pdf/2107.11781v1.pdf | Towards Controlled and Diverse Generation of Article Comments | Much research in recent years has focused on automatic article commenting. However, few of previous studies focus on the controllable generation of comments. Besides, they tend to generate dull and commonplace comments, which further limits their practical application. In this paper, we make the first step towards controllable generation of comments, by building a system that can explicitly control the emotion of the generated comments. To achieve this, we associate each kind of emotion category with an embedding and adopt a dynamic fusion mechanism to fuse this embedding into the decoder. A sentence-level emotion classifier is further employed to better guide the model to generate comments expressing the desired emotion. To increase the diversity of the generated comments, we propose a hierarchical copy mechanism that allows our model to directly copy words from the input articles. We also propose a restricted beam search (RBS) algorithm to increase intra-sentence diversity. Experimental results show that our model can generate informative and diverse comments that express the desired emotions with high accuracy. | ['Houfeng Wang', 'Linhao Zhang'] | 2021-07-25 | null | null | null | null | ['comment-generation'] | ['natural-language-processing'] | [ 1.52099982e-01 1.54564112e-01 -1.42520115e-01 -5.39876699e-01
-6.54523611e-01 -3.68692160e-01 4.85633165e-01 9.99362767e-02
-1.42782718e-01 6.85493290e-01 7.52729237e-01 -8.16319361e-02
5.60197473e-01 -7.87197351e-01 -2.65346974e-01 -6.06077373e-01
6.70292556e-01 -2.80823056e-02 -1.11810938e-01 -5.07819533e-01
5.51205397e-01 -8.77771974e-02 -1.46282089e+00 5.95692515e-01
1.35573459e+00 5.70191503e-01 4.41062897e-01 6.32022083e-01
-4.85128224e-01 8.30895722e-01 -8.36347878e-01 -6.50807798e-01
-2.17528373e-01 -8.98188412e-01 -3.57632101e-01 3.18520099e-01
-5.02614379e-01 -1.57498017e-01 2.83506453e-01 9.92184579e-01
8.45612347e-01 6.00466803e-02 7.80908167e-01 -1.15862834e+00
-8.86101186e-01 1.12136555e+00 -5.04485011e-01 -3.13992053e-01
6.30954742e-01 -4.29804577e-03 1.17550874e+00 -1.03433597e+00
6.80723369e-01 1.30774927e+00 1.63896829e-01 7.99030721e-01
-8.42188776e-01 -6.97394192e-01 4.08697128e-01 3.22173387e-02
-1.11062050e+00 -4.60240126e-01 1.23054767e+00 -1.68823436e-01
6.01210058e-01 4.02014643e-01 7.35359311e-01 1.11151910e+00
2.12468237e-01 1.08281350e+00 8.08120430e-01 -5.95237076e-01
3.66147339e-01 6.35311365e-01 -1.29006013e-01 3.29665750e-01
-1.26571044e-01 -6.15060627e-01 -6.25119030e-01 -6.57031238e-02
2.34558433e-01 -1.04689345e-01 -4.31255013e-01 -1.05018253e-02
-1.09556627e+00 1.07276392e+00 2.58681625e-01 2.09126472e-01
-4.85963225e-01 -1.98533848e-01 4.73789573e-01 1.27403602e-01
7.76540041e-01 5.83114803e-01 -1.85934439e-01 -2.18388468e-01
-8.43557894e-01 2.65163243e-01 9.11537111e-01 1.03644478e+00
5.16723931e-01 -1.31529421e-01 -5.81828415e-01 1.08808124e+00
5.93466818e-01 4.39812988e-01 5.86943448e-01 -4.97613192e-01
5.25058746e-01 7.66134202e-01 2.67563999e-01 -1.26061606e+00
-4.44663540e-02 -3.81274611e-01 -8.40500712e-01 -1.68803379e-01
-4.42451745e-01 -6.36993110e-01 -4.92272198e-01 1.52586508e+00
2.03960091e-01 -2.44589731e-01 3.45862329e-01 1.01611853e+00
7.56338775e-01 1.22435939e+00 -2.27617640e-02 -4.35493171e-01
1.19797897e+00 -1.08614874e+00 -9.97881293e-01 -1.93501815e-01
7.27588475e-01 -9.61870670e-01 1.39425182e+00 2.73413688e-01
-9.84776616e-01 -3.74493599e-01 -1.01083076e+00 -7.22470880e-02
-3.93954590e-02 4.39343512e-01 2.54943371e-01 4.83339190e-01
-7.47352898e-01 -1.13504209e-01 -4.05500084e-01 -1.31375045e-01
5.17980233e-02 8.00947621e-02 1.84670001e-01 1.72611281e-01
-1.36287141e+00 6.26344621e-01 2.71930337e-01 7.03620315e-02
-2.00988501e-01 -4.73910958e-01 -8.07908356e-01 7.41308555e-02
1.67232335e-01 -6.61736369e-01 1.25177276e+00 -1.36942291e+00
-1.89228010e+00 3.59911144e-01 -4.64639187e-01 -5.62197678e-02
2.67391443e-01 -9.39222425e-02 -3.92887175e-01 -5.53369289e-03
-8.20910633e-02 1.00242782e+00 7.30618179e-01 -1.47389662e+00
-6.30222201e-01 1.20661527e-01 8.97409543e-02 5.13240576e-01
-8.90660703e-01 1.72808573e-01 -5.71536958e-01 -9.38099742e-01
-2.77329504e-01 -9.65482056e-01 -4.05766129e-01 -3.00035119e-01
-7.14319766e-01 -3.41613770e-01 4.74242419e-01 -4.40695047e-01
1.71630359e+00 -2.30770683e+00 3.32020789e-01 1.47157133e-01
1.71859395e-02 1.20848063e-02 -1.34772465e-01 6.15152180e-01
2.90998131e-01 3.07836503e-01 -1.70746431e-01 -4.57490176e-01
2.12430328e-01 6.39706925e-02 -4.69869971e-01 -2.53559917e-01
2.61190921e-01 8.40950906e-01 -7.87279785e-01 -7.00301409e-01
-1.61438078e-01 3.94221306e-01 -9.47834253e-01 4.97941971e-01
-4.91414189e-01 1.66502073e-01 -7.58307517e-01 1.24098174e-01
6.13686562e-01 -2.01077357e-01 1.13382831e-01 -1.54599056e-01
-1.79742590e-01 2.78502375e-01 -9.40038085e-01 1.32144642e+00
-7.80264139e-01 2.24419221e-01 -9.20646191e-02 -4.80305940e-01
1.36683786e+00 2.82412410e-01 1.92471370e-01 -4.26914960e-01
2.47332111e-01 -9.00124013e-03 -7.65524134e-02 -6.02552950e-01
1.02837610e+00 -1.47520304e-01 -3.33720356e-01 7.45473564e-01
-4.36466396e-01 -2.81823337e-01 2.51661897e-01 4.49693561e-01
5.85973024e-01 -1.66976780e-01 1.59495354e-01 4.51571457e-02
6.68956339e-01 -1.27684265e-01 4.89830375e-01 3.80710900e-01
1.51704252e-01 6.50541902e-01 6.22068822e-01 8.37321728e-02
-9.42731440e-01 -5.86294889e-01 3.44352812e-01 1.08949196e+00
2.57621229e-01 -6.28129900e-01 -7.73613870e-01 -5.48090100e-01
-3.89784068e-01 1.14960682e+00 -4.92705286e-01 -3.14485282e-01
-1.95025146e-01 -6.84969664e-01 1.92236453e-01 2.65750647e-01
3.77034485e-01 -1.23141336e+00 -4.11474705e-01 3.33380401e-01
-7.11647630e-01 -7.17558563e-01 -6.84770703e-01 -1.86307028e-01
-3.97931397e-01 -4.51553971e-01 -8.66343260e-01 -8.92252207e-01
1.03956878e+00 1.41304687e-01 8.05930197e-01 1.47900328e-01
3.21615547e-01 -4.09311391e-02 -1.02021492e+00 -5.16730785e-01
-6.56221569e-01 4.47168648e-01 -3.31368834e-01 3.33567202e-01
3.25212419e-01 -2.68828243e-01 -5.05724728e-01 3.72748822e-02
-1.14596879e+00 5.77958763e-01 7.42390573e-01 8.04885805e-01
1.98931441e-01 -1.21570207e-01 8.56837213e-01 -8.00043404e-01
1.54351437e+00 -6.51195228e-01 -1.05961584e-01 3.28747392e-01
-5.80045283e-01 2.45173693e-01 9.34099317e-01 -5.09727776e-01
-1.39205456e+00 -9.89744514e-02 -4.31514829e-01 1.20211905e-02
1.09068893e-01 8.48611116e-01 -2.10451871e-01 4.88615900e-01
4.14332896e-01 3.40558290e-01 -5.73862530e-02 -1.51787981e-01
4.54687208e-01 1.27538216e+00 -5.77480644e-02 -3.49571079e-01
5.62373340e-01 -1.73834357e-02 -6.62179887e-01 -4.96955752e-01
-8.18008602e-01 -2.21614182e-01 -2.00022087e-01 -5.02989829e-01
8.34301710e-01 -9.79751527e-01 -4.95642215e-01 2.50356793e-01
-1.29654753e+00 -1.08450800e-01 1.34148628e-01 3.15067977e-01
-3.20289940e-01 2.43095741e-01 -5.93162835e-01 -9.89192724e-01
-6.83500648e-01 -1.21079111e+00 1.15849781e+00 3.77166867e-01
-5.48235476e-01 -7.59871900e-01 1.48779824e-02 2.21436784e-01
4.32091117e-01 -6.75287098e-02 8.19374204e-01 -4.50335830e-01
-2.63527155e-01 -1.09758548e-01 -1.30235389e-01 2.51697212e-01
8.30815211e-02 2.72501558e-01 -4.90447640e-01 7.31178001e-02
1.36405928e-03 -3.68937314e-01 6.61509275e-01 1.22295851e-02
1.05108118e+00 -5.92524052e-01 -2.01728284e-01 2.12305367e-01
1.04697335e+00 2.51321167e-01 6.77222490e-01 1.19536154e-01
6.11070812e-01 6.58232570e-01 9.11276579e-01 9.24168825e-01
7.77801573e-01 5.62054813e-01 1.50373429e-01 -8.87907222e-02
1.99405193e-01 -3.29241514e-01 7.27348626e-01 1.46240032e+00
3.44314188e-01 -7.97619879e-01 -4.94461983e-01 4.48829830e-01
-1.86958027e+00 -9.31588411e-01 -5.25206216e-02 1.58319211e+00
1.21899414e+00 1.77030846e-01 -1.52116269e-01 1.15464829e-01
7.85353601e-01 2.79458493e-01 -3.36512893e-01 -8.19695175e-01
6.72189668e-02 -2.08155796e-01 -6.55452013e-02 4.69638914e-01
-6.05477095e-01 9.33780551e-01 5.67070770e+00 7.77833521e-01
-1.36025965e+00 -2.21495360e-01 7.42764056e-01 -1.07438646e-01
-1.12293482e+00 -2.19974324e-01 -8.08218062e-01 7.68849194e-01
6.98018789e-01 -6.27299488e-01 1.87106445e-01 7.86694229e-01
6.97158575e-01 6.18054308e-02 -6.58610702e-01 8.16154063e-01
4.53954011e-01 -1.13045609e+00 3.46123934e-01 -3.13542634e-01
8.95730913e-01 -6.54468954e-01 -3.64318467e-03 2.83190548e-01
2.47944117e-01 -6.95858359e-01 7.95503259e-01 6.44333839e-01
4.46165591e-01 -1.13379121e+00 7.76271164e-01 6.14936352e-01
-9.06850219e-01 2.49882769e-02 -3.04752707e-01 -2.21485295e-03
4.71840858e-01 8.97055089e-01 -9.90125299e-01 4.29520816e-01
1.69298649e-01 7.63927698e-01 -5.18813372e-01 6.99569762e-01
-4.65530097e-01 6.87902749e-01 -4.08777744e-02 -9.17913377e-01
1.52401373e-01 -2.17048824e-01 3.67475957e-01 1.37117898e+00
6.09554648e-01 1.50800034e-01 3.42757940e-01 8.05352390e-01
-2.69117773e-01 7.16174126e-01 -3.32400948e-01 -1.47916436e-01
5.62790215e-01 1.36826432e+00 -5.20919263e-01 -4.12794888e-01
-1.79852426e-01 1.15064108e+00 2.14606971e-01 2.25045085e-01
-9.89114106e-01 -6.03769541e-01 3.50816518e-01 -2.05930367e-01
2.74284303e-01 -1.70148984e-02 -4.49785292e-01 -1.32279480e+00
1.45754576e-01 -1.01580107e+00 -1.60099298e-01 -1.10181332e+00
-1.17396772e+00 6.22511864e-01 -3.63087833e-01 -1.23824048e+00
-4.44758147e-01 3.07719316e-02 -8.21083724e-01 8.06246042e-01
-1.32069993e+00 -9.84950483e-01 -2.66218930e-01 2.93837935e-01
7.70094991e-01 3.03987553e-03 7.03144133e-01 1.66085914e-01
-6.10612214e-01 7.28220344e-01 -1.56792119e-01 -1.49803283e-02
8.86812150e-01 -9.83162582e-01 7.15041310e-02 8.55565429e-01
-1.94611907e-01 6.64777279e-01 9.98302639e-01 -6.52039707e-01
-1.25573123e+00 -1.18351161e+00 1.27020824e+00 1.05768874e-01
3.97595167e-01 -4.84796911e-01 -7.18329430e-01 2.29893357e-01
6.34479344e-01 -7.65542924e-01 9.75042403e-01 1.39753036e-02
3.03917285e-02 -5.98426769e-03 -7.98201025e-01 1.16641307e+00
6.15560055e-01 -3.16558152e-01 -4.36685860e-01 2.36333668e-01
1.02416301e+00 -2.38851473e-01 -5.16544223e-01 1.47293940e-01
4.48589683e-01 -8.07446718e-01 3.84718210e-01 -1.92425832e-01
1.03698719e+00 -3.76315206e-01 1.34636849e-01 -1.81375420e+00
-2.07313105e-01 -5.34347713e-01 1.79515123e-01 1.74288559e+00
7.74620950e-01 -2.93229192e-01 6.19139671e-01 5.87109566e-01
-2.89855808e-01 -9.11147237e-01 -1.89322919e-01 -8.95071477e-02
-2.27910817e-01 -2.80244172e-01 7.90623903e-01 7.28595793e-01
4.76997495e-01 5.79143405e-01 -6.68201447e-01 -1.49536118e-01
-7.32338801e-02 3.20787340e-01 8.32024038e-01 -7.37252474e-01
-1.26554862e-01 -4.68622386e-01 1.14666216e-01 -1.28162622e+00
2.80848891e-01 -8.44235957e-01 5.18312752e-01 -1.62664437e+00
2.85703748e-01 -4.81384903e-01 1.57785900e-02 2.95040607e-01
-5.61190546e-01 3.59787010e-02 2.77729183e-01 4.79015596e-02
-6.71927691e-01 9.19069886e-01 1.36422312e+00 1.49124423e-02
-3.47439229e-01 -1.13161065e-01 -1.35709608e+00 3.29543680e-01
1.09610569e+00 -3.63597333e-01 -6.52626097e-01 -3.55378896e-01
6.29400313e-01 2.84099746e-02 -1.41763031e-01 -6.63323581e-01
2.04251871e-01 -3.88981164e-01 1.68994710e-01 -6.09294116e-01
1.98783353e-01 -5.09261608e-01 -2.20295265e-02 2.08265349e-01
-9.34718430e-01 1.89454019e-01 -2.47007072e-01 3.14741254e-01
-4.93523657e-01 -3.79385173e-01 5.35472751e-01 9.34933499e-02
-2.62949735e-01 5.36601506e-02 -7.35420823e-01 -1.72323704e-01
1.09988642e+00 1.08994186e-01 -2.99262013e-02 -9.15098548e-01
-2.67110139e-01 4.42060202e-01 5.71668446e-01 7.26766229e-01
8.07088912e-01 -1.39276159e+00 -8.46778810e-01 7.84655064e-02
2.53846675e-01 -2.31377035e-01 2.41264794e-02 4.70039099e-01
-2.71350652e-01 1.05715126e-01 4.94544320e-02 -1.86183408e-01
-1.29846799e+00 4.56987411e-01 -1.50038332e-01 -2.52738982e-01
-2.07118601e-01 7.65276611e-01 -9.14193913e-02 -4.15266037e-01
-1.74624640e-02 -1.26141936e-01 -5.32233357e-01 3.26361954e-01
6.25339866e-01 -1.88134551e-01 -2.34872386e-01 -5.24262309e-01
-4.27134112e-02 1.32345200e-01 -1.90310135e-01 -3.06808770e-01
1.18344617e+00 -3.84832859e-01 -2.61575699e-01 4.85310316e-01
1.16956234e+00 4.41114902e-01 -8.45312893e-01 1.38462126e-01
-1.74191162e-01 -2.68003225e-01 -3.49406339e-02 -7.58295834e-01
-9.32391346e-01 7.80349135e-01 -4.79657277e-02 3.09021801e-01
1.29677391e+00 -1.32555589e-01 1.02255416e+00 1.86998934e-01
-3.87522653e-02 -1.18508124e+00 2.08357409e-01 5.76023877e-01
9.73188221e-01 -9.39001739e-01 -2.55429327e-01 -5.04769385e-01
-1.27606606e+00 1.17919707e+00 7.26604462e-01 1.01345867e-01
4.16586637e-01 3.83724689e-01 3.48133504e-01 1.50243148e-01
-1.17556643e+00 5.11985570e-02 1.30431935e-01 2.88980335e-01
8.51995647e-01 2.60360986e-02 -8.26577425e-01 8.21832180e-01
-4.33771968e-01 9.91199538e-02 8.07167649e-01 7.37641156e-01
-7.16853082e-01 -1.37514031e+00 -1.92562744e-01 4.28037763e-01
-2.58862019e-01 -2.41120860e-01 -5.60010016e-01 3.76186334e-02
-4.80763167e-02 1.26665986e+00 -4.02462184e-02 -5.58054447e-01
1.33186266e-01 7.83197731e-02 -1.39842838e-01 -6.90137863e-01
-6.32021010e-01 2.11953372e-01 2.50102252e-01 -1.82498246e-02
-3.93658817e-01 -5.59148073e-01 -1.35881674e+00 -2.73473591e-01
-4.98385638e-01 5.54302096e-01 5.63523471e-01 8.29911113e-01
6.12826049e-01 6.47125661e-01 1.11036646e+00 -6.42027438e-01
-3.55709165e-01 -1.06603420e+00 -3.17624331e-01 4.79660928e-01
-2.64183767e-02 -2.33091146e-01 -3.69603127e-01 2.07220614e-01] | [12.069561958312988, 8.999163627624512] |
e945c55b-a7d9-4a09-b111-b60f3933c172 | predicting-individual-responses-to-vasoactive | 1901.10400 | null | http://arxiv.org/abs/1901.10400v1 | http://arxiv.org/pdf/1901.10400v1.pdf | Predicting Individual Responses to Vasoactive Medications in Children with Septic Shock | Objective: Predict individual septic children's personalized physiologic
responses to vasoactive titrations by training a Recurrent Neural Network (RNN)
using EMR data.
Materials and Methods: This study retrospectively analyzed EMR of patients
admitted to a pediatric ICU from 2009 to 2017. Data included charted time
series vitals, labs, drugs, and interventions of children with septic shock
treated with dopamine, epinephrine, or norepinephrine. A RNN was trained to
predict responses in heart rate (HR), systolic blood pressure (SBP), diastolic
blood pressure (DBP) and mean arterial pressure (MAP) to 8,640 titrations
during 652 septic episodes and evaluated on a holdout set of 3,883 titrations
during 254 episodes. A linear regression model using titration data as its sole
input was also developed and compared to the RNN model. Evaluation methods
included the correlation coefficient between actual physiologic responses and
RNN predictions, mean absolute error (MAE), and area under the receiver
operating characteristic curve (AUC).
Results: The actual physiologic responses displayed significant variability
and were more accurately predicted by the RNN model than by titration alone
(r=0.20 vs r=0.05, p<0.01). The RNN showed MAE and AUC improvements over the
linear model. The RNN's MAEs associated with dopamine and epinephrine were 1-3%
lower than the linear regression model MAE for HR, SBP, DBP, and MAP. Across
all vitals vasoactives, the RNN achieved 1-19% AUC improvement over the linear
model.
Conclusion: This initial attempt in pediatric critical care to predict
individual physiologic responses to vasoactive dose changes in children with
septic shock demonstrated an RNN model showed some improvement over a linear
model. While not yet clinically applicable, further development may assist
clinical administration of vasoactive medications in children with septic
shock. | ['Randall Wetzel', 'Jessica Asencio', 'David Ledbetter', 'Cameron Carlin', 'Nicole Fronda', 'Melissa Aczon', 'Barry Markovitz'] | 2019-01-15 | null | null | null | null | ['holdout-set'] | ['computer-vision'] | [-1.27502397e-01 -3.90166044e-02 5.72511852e-02 -3.36655796e-01
-4.98496294e-01 -4.29465503e-01 -4.61721539e-01 8.13611269e-01
-4.61672306e-01 8.22507381e-01 6.00418270e-01 -9.06493604e-01
-3.24449390e-01 -5.03309309e-01 -4.20895875e-01 -4.89711076e-01
-3.55675370e-01 7.88933635e-01 -3.77374381e-01 -2.76485160e-02
-1.64049491e-02 6.65813327e-01 -7.02748716e-01 3.70286912e-01
6.73806489e-01 9.46707189e-01 -5.41061938e-01 1.14886820e+00
1.01658953e-02 1.60690558e+00 -6.60087287e-01 4.65650320e-01
3.45935524e-01 -5.81950307e-01 -2.24083826e-01 -8.20791900e-01
3.94597232e-01 -7.07245350e-01 -4.26765651e-01 -6.83778599e-02
1.16349852e+00 1.64830238e-01 6.51350200e-01 -7.96225905e-01
-1.61996999e-04 9.83892977e-01 2.01187897e-02 3.42796952e-01
1.86683014e-01 6.41353369e-01 2.01315701e-01 -2.40062341e-01
-3.11717689e-01 7.54490077e-01 1.13048804e+00 8.30521822e-01
-1.40671372e+00 -6.40966177e-01 -1.42597601e-01 -5.89779973e-01
-1.11796594e+00 1.33909201e-02 -2.38436032e-02 -8.09859216e-01
1.09243977e+00 3.37846465e-02 1.15129852e+00 6.86953545e-01
6.34669304e-01 -1.09851435e-01 8.45352709e-01 6.80240914e-02
2.00268328e-02 2.04525933e-01 2.49032900e-01 3.90628994e-01
1.17112987e-01 3.45180899e-01 -1.84564874e-01 -2.92965144e-01
9.14945245e-01 5.51975191e-01 -6.84849977e-01 2.13040978e-01
-8.18558991e-01 6.88370407e-01 6.97182817e-03 -1.91966131e-01
-9.85468686e-01 6.43687174e-02 7.60557473e-01 4.34914231e-01
7.52989277e-02 9.91233766e-01 -1.06224453e+00 -5.77304304e-01
-5.22307634e-01 -2.46881932e-01 1.11254871e+00 6.14208162e-01
-1.38225481e-01 3.12030286e-01 -1.18952185e-01 1.04584956e+00
-9.24017131e-02 7.31027722e-01 6.36693835e-01 -8.96519303e-01
3.19199950e-01 4.25819129e-01 3.02221298e-01 -5.44061124e-01
-1.49333978e+00 -2.96509296e-01 -9.00880933e-01 -1.33909941e-01
2.92005330e-01 -9.20668542e-01 -6.72915161e-01 1.61596525e+00
-3.71088497e-02 -6.46215826e-02 4.03888166e-01 5.57418048e-01
8.81513834e-01 4.15636271e-01 4.96985942e-01 -7.48028040e-01
7.75525808e-01 -5.13575852e-01 -3.41650128e-01 4.31826204e-01
8.60409379e-01 -3.68528366e-01 3.24705243e-01 4.95554864e-01
-1.45829415e+00 -1.98224053e-01 -5.62796831e-01 6.63792253e-01
-7.44133303e-03 -3.72931123e-01 1.68914720e-01 5.34331024e-01
-1.11847210e+00 1.10973120e+00 -9.73832190e-01 -6.86847344e-02
7.98279196e-02 5.05315721e-01 3.72414403e-02 3.73010844e-01
-1.31041551e+00 1.36664557e+00 1.73304856e-01 6.83981692e-03
-5.27867794e-01 -1.54524875e+00 -6.82097137e-01 4.19574350e-01
-2.36419395e-01 -1.17938137e+00 1.17047489e+00 -1.74378246e-01
-1.43200684e+00 1.66896567e-01 1.79995254e-01 -6.02340996e-01
3.37691218e-01 -4.82893050e-01 -1.72187448e-01 3.68202090e-01
-4.61781293e-01 3.06675166e-01 -4.16018516e-02 -5.64496756e-01
-4.62632477e-01 -1.36770681e-01 -5.13119578e-01 3.07538152e-01
9.42016318e-02 -5.65378070e-02 6.33223057e-01 -1.77022085e-01
1.83201820e-01 -7.21018076e-01 -5.42347729e-01 -4.12330389e-01
1.04839943e-01 3.02131802e-01 7.44661540e-02 -7.54565835e-01
1.35808301e+00 -1.70735896e+00 -4.44801420e-01 2.42950141e-01
6.13868237e-01 2.50406086e-01 2.29825109e-01 6.02043748e-01
-8.60886633e-01 5.57967834e-02 1.90929830e-01 3.74731064e-01
-3.79194885e-01 -9.10660103e-02 -2.64342517e-01 3.89536440e-01
-1.53903216e-01 9.40700352e-01 -9.79812205e-01 -3.60242985e-02
9.18420374e-01 5.78485608e-01 -5.08692086e-01 6.58985853e-01
2.62070656e-01 7.17137933e-01 1.82787627e-02 3.84672463e-01
1.21956371e-01 -2.59851545e-01 2.63107643e-02 1.47295356e-01
-4.39676732e-01 4.56512690e-01 -5.50922096e-01 8.37046862e-01
-4.53371495e-01 6.36026204e-01 -3.90094936e-01 -5.33299387e-01
1.40120149e+00 9.83867228e-01 1.52678514e+00 -4.79144633e-01
5.70724964e-01 2.21351624e-01 3.25376391e-01 -7.62249827e-01
-3.69815916e-01 -9.07875419e-01 4.43380386e-01 5.27377605e-01
-5.60007572e-01 -3.44589114e-01 -3.61389399e-01 -7.73182884e-02
1.33007216e+00 -4.39261705e-01 4.73649502e-01 -4.22173619e-01
3.16191703e-01 1.16361007e-02 4.46396679e-01 9.31541741e-01
-2.23194838e-01 7.20009446e-01 4.95898664e-01 -9.67588246e-01
-8.37495804e-01 -1.21227729e+00 -4.59849119e-01 5.27247548e-01
-6.64699793e-01 1.91125885e-01 -2.26048052e-01 3.18030044e-02
3.18075687e-01 9.40263689e-01 -6.41462386e-01 -3.77155960e-01
-7.11960733e-01 -7.76052117e-01 7.36024201e-01 9.48862791e-01
-3.57247263e-01 -9.85547304e-01 -1.13748288e+00 1.09098446e+00
1.13613389e-01 -5.96788526e-01 -3.26119125e-01 5.83132565e-01
-1.24000609e+00 -1.17389202e+00 -6.54853344e-01 -2.05565631e-01
2.14168042e-01 -1.84680000e-01 9.71964717e-01 -2.92482316e-01
-5.07391751e-01 7.90678680e-01 -2.70073377e-02 -1.20350397e+00
-7.53088772e-01 -2.09152654e-01 6.48846745e-01 -9.22980607e-01
6.57951772e-01 -8.82237792e-01 -1.25472212e+00 1.64749622e-01
-4.48239684e-01 -1.33056566e-01 1.23345301e-01 7.26313233e-01
4.80938494e-01 -8.16420138e-01 1.00347185e+00 -7.10548759e-01
7.66129553e-01 -9.19007838e-01 -6.23074234e-01 1.69820506e-02
-1.31382596e+00 -2.44178191e-01 1.17555654e+00 -5.68192661e-01
-2.32094929e-01 -3.43018383e-01 1.66834682e-01 -8.84178460e-01
2.32163724e-02 2.54539639e-01 9.89557803e-01 6.26690745e-01
9.14288759e-01 7.09859505e-02 4.42311883e-01 -1.42169118e-01
-2.42461875e-01 5.17665803e-01 6.37134969e-01 -6.68925405e-01
7.62899499e-03 -3.06593597e-01 4.99693215e-01 -1.67165354e-01
-4.50036973e-01 -4.63442326e-01 -4.91888791e-01 -2.55143672e-01
7.62655020e-01 -9.91919219e-01 -1.56245363e+00 1.19233966e-01
-7.14967012e-01 -5.35815120e-01 -6.65262997e-01 1.22263193e+00
-6.10621512e-01 -8.42748731e-02 -9.03452635e-01 -8.97094250e-01
-1.15368354e+00 -8.99147213e-01 -8.65820330e-03 2.41029277e-01
-7.39429891e-01 -1.33137393e+00 6.35439396e-01 -6.63898587e-02
8.15597475e-01 8.67881894e-01 1.13729036e+00 -1.18549061e+00
1.82831973e-01 -3.63317668e-01 4.56280038e-02 6.03656054e-01
7.54592478e-01 3.51270825e-01 -3.66983503e-01 -2.08788931e-01
1.67985842e-01 -5.64270327e-03 6.19048290e-02 9.07505214e-01
1.04552007e+00 -3.41369003e-01 3.05996448e-01 6.57106876e-01
1.57117689e+00 1.33954787e+00 2.72451669e-01 -7.41916448e-02
4.68216807e-01 3.79706651e-01 9.08340048e-03 9.95459974e-01
2.84094959e-01 -2.40132421e-01 9.59109589e-02 4.55797603e-03
5.39494753e-01 -2.35614121e-01 1.79269284e-01 7.86003709e-01
-1.82387367e-01 -4.48451377e-02 -1.39857399e+00 3.02432626e-01
-1.29950786e+00 -7.12164283e-01 -4.19725120e-01 2.28122497e+00
7.55882800e-01 -1.99805334e-01 2.97732383e-01 -3.42882037e-01
5.66508830e-01 -4.34100091e-01 -9.62673724e-01 -1.45181322e+00
1.29640207e-01 7.50211895e-01 6.40814960e-01 3.51530075e-01
-2.87026972e-01 4.36854988e-01 7.38051414e+00 -6.06760383e-01
-1.38775599e+00 -5.52445471e-01 7.79406965e-01 -5.84105194e-01
2.92864889e-01 -5.74676573e-01 -2.88830429e-01 5.46406388e-01
1.76619589e+00 -2.56198436e-01 5.46751320e-01 7.65325427e-01
8.85345161e-01 1.04615167e-01 -1.58275115e+00 1.04873359e+00
-2.51338691e-01 -1.36707425e+00 -4.87851441e-01 -4.58607525e-01
4.76434946e-01 3.94333601e-01 -8.59619677e-02 4.92815346e-01
5.29580712e-01 -1.39443982e+00 -2.37810701e-01 8.65958154e-01
1.23334587e+00 -4.30963337e-01 1.03754592e+00 -3.42451558e-02
-6.37852609e-01 -2.18320116e-01 -1.26783341e-01 -1.32668823e-01
-2.74497420e-02 2.67817020e-01 -1.56259584e+00 -7.92452395e-02
4.03347641e-01 2.76660532e-01 3.12443525e-01 7.71335840e-01
1.57758489e-01 7.96202183e-01 -6.80298328e-01 -4.81813736e-02
1.25899553e-01 -1.75216317e-01 2.63233215e-01 1.22138119e+00
-5.30353598e-02 1.20633495e+00 -3.50319266e-01 6.81353688e-01
2.97868609e-01 6.19266510e-01 -7.65103340e-01 2.34382004e-01
4.21877801e-01 6.98901653e-01 -2.42134966e-02 -5.82002342e-01
-3.71240824e-01 -1.80629149e-01 -1.22026615e-01 3.38479340e-01
-7.07354188e-01 -1.71832353e-01 6.86021507e-01 3.47802341e-01
-4.10048813e-01 5.57279885e-01 -1.11124027e+00 -7.07137525e-01
-6.51066482e-01 -9.36551630e-01 6.77031279e-01 -8.23555768e-01
-9.30302083e-01 4.68616903e-01 -5.36754765e-02 -1.24910891e+00
-7.49078870e-01 -4.50253874e-01 -7.86507070e-01 1.47162092e+00
-8.28650832e-01 2.96286523e-01 -3.39841634e-01 1.52998820e-01
2.52976090e-01 -1.69997513e-02 1.19596159e+00 5.19289598e-02
-6.50575578e-01 4.77428734e-01 1.06016593e-02 1.23644084e-01
7.01286256e-01 -1.24966705e+00 -3.44534576e-01 1.91816017e-02
-1.28114092e+00 1.19486880e+00 7.03450799e-01 -7.42656767e-01
-9.43638682e-01 -9.76198673e-01 3.23381722e-01 -6.69176817e-01
6.09157979e-01 7.69887090e-01 -9.78179872e-01 6.56260252e-01
-1.34758046e-02 -3.31931233e-01 1.34847212e+00 -2.08790138e-01
-4.59685251e-02 -1.70448586e-01 -1.37639475e+00 3.59410197e-01
2.82109410e-01 -1.51729450e-01 -6.70227826e-01 1.64300114e-01
7.05426097e-01 -7.33699083e-01 -2.00658274e+00 9.45297480e-01
8.44218016e-01 -6.81275606e-01 6.84867740e-01 -1.20253873e+00
5.70741892e-01 7.99483657e-02 -3.80203896e-03 -1.59656620e+00
-2.58526534e-01 -7.72209585e-01 1.24915205e-02 3.44133943e-01
5.42874932e-01 -1.33605778e+00 5.44372559e-01 1.67957020e+00
-4.76110041e-01 -1.52739131e+00 -5.10583401e-01 -2.32987344e-01
5.93182027e-01 -1.48270056e-01 7.59902894e-01 8.97819161e-01
6.42288566e-01 3.57063025e-01 -2.99833089e-01 8.63422453e-02
1.55301765e-01 -3.09890330e-01 4.06071693e-01 -9.21028674e-01
-2.40399718e-01 -5.65079868e-01 -1.78000420e-01 -5.24302006e-01
-5.66627204e-01 -7.08990395e-01 -2.59913594e-01 -1.80808091e+00
6.57344535e-02 -5.34014523e-01 -9.63125825e-01 6.79722428e-01
-2.42451549e-01 -7.86592007e-01 8.83135572e-02 -8.61783624e-02
3.65642697e-01 1.49910331e-01 9.22256470e-01 2.67425776e-01
-1.17227399e+00 1.91283226e-01 -6.36710167e-01 2.89470285e-01
1.14965665e+00 -6.28775895e-01 -6.28299534e-01 6.04452565e-03
-8.68044049e-02 1.23079562e+00 -2.95919478e-01 -8.11634183e-01
2.80742019e-01 -5.76729596e-01 7.94116437e-01 -7.91403472e-01
-1.22294694e-01 -7.49811113e-01 5.62885463e-01 1.02471066e+00
-4.13041294e-01 7.76587188e-01 7.20738113e-01 -2.65620381e-01
4.87321280e-02 1.88720047e-01 9.96314228e-01 -1.68893829e-01
4.71726418e-01 6.15884364e-01 -7.02023745e-01 3.73660356e-01
7.46325135e-01 -1.52038977e-01 -5.64252913e-01 -6.78320467e-01
-9.63224292e-01 5.14413893e-01 5.11711873e-02 1.36180803e-01
8.18789184e-01 -4.92276460e-01 -8.72923911e-01 7.36631230e-02
-1.90402806e-01 3.16370845e-01 5.10906816e-01 1.31484139e+00
-1.17713404e+00 6.61731899e-01 5.93850911e-02 -5.39075077e-01
-8.99333298e-01 6.78631842e-01 1.02338445e+00 1.25008672e-01
-7.32573509e-01 6.07098758e-01 2.21584663e-01 -1.33103222e-01
3.00515056e-01 -4.59193647e-01 -2.05782261e-02 -2.60622501e-01
4.98828351e-01 7.40034163e-01 -2.94550002e-01 2.17573121e-01
-1.90479413e-01 4.11868840e-01 -3.49798761e-02 3.06566298e-01
1.34461296e+00 1.58779204e-01 -1.40557632e-01 6.22306526e-01
1.17752445e+00 -4.22490299e-01 -9.45987105e-01 1.12187721e-01
-5.59686005e-01 9.73738432e-02 -5.50846040e-01 -1.13048673e+00
-9.95403826e-01 6.93573415e-01 8.22011411e-01 6.82390109e-02
1.18098485e+00 -3.76475483e-01 5.82382977e-01 3.78564566e-01
-3.65798548e-02 -8.72306764e-01 -3.08797687e-01 3.54932934e-01
6.75456107e-01 -7.98899412e-01 1.19642783e-02 3.78059268e-01
-8.73187482e-01 1.55229878e+00 7.97824621e-01 -6.42607978e-04
7.27556467e-01 2.72167832e-01 8.44932437e-01 -9.11362842e-02
-1.26304233e+00 8.65055561e-01 -4.84658331e-01 2.00880602e-01
5.03755212e-01 4.56746757e-01 -1.60542265e-01 4.16982174e-01
-3.33132923e-01 1.58327475e-01 1.01853526e+00 5.96246302e-01
-5.06046534e-01 -2.84414381e-01 -1.92520022e-01 1.01078105e+00
-5.31107903e-01 -4.59135443e-01 1.32628515e-01 6.25698328e-01
-3.40769947e-01 9.54992831e-01 5.06394878e-02 -3.39949638e-01
7.59758294e-01 6.58250093e-01 2.04874054e-01 -6.29134953e-01
-1.14954686e+00 -1.22032009e-01 -1.20572336e-01 -3.87894034e-01
1.11358859e-01 -6.81909204e-01 -1.69178283e+00 -3.43379706e-01
1.37464121e-01 2.57430941e-01 7.42128849e-01 3.14071506e-01
1.76148221e-01 9.13270175e-01 1.13756418e+00 7.66490251e-02
-9.07936931e-01 -1.05067098e+00 -4.57492322e-01 -1.35155961e-01
8.67339551e-01 -1.93249404e-01 -8.57520401e-01 -5.09320557e-01] | [8.030557632446289, 6.18095588684082] |
1803d502-3d3e-484b-aab7-58f5d3c65b09 | exploring-edge-tpu-for-network-intrusion | 2103.16295 | null | https://arxiv.org/abs/2103.16295v1 | https://arxiv.org/pdf/2103.16295v1.pdf | Exploring Edge TPU for Network Intrusion Detection in IoT | This paper explores Google's Edge TPU for implementing a practical network intrusion detection system (NIDS) at the edge of IoT, based on a deep learning approach. While there are a significant number of related works that explore machine learning based NIDS for the IoT edge, they generally do not consider the issue of the required computational and energy resources. The focus of this paper is the exploration of deep learning-based NIDS at the edge of IoT, and in particular the computational and energy efficiency. In particular, the paper studies Google's Edge TPU as a hardware platform, and considers the following three key metrics: computation (inference) time, energy efficiency and the traffic classification performance. Various scaled model sizes of two major deep neural network architectures are used to investigate these three metrics. The performance of the Edge TPU-based implementation is compared with that of an energy efficient embedded CPU (ARM Cortex A53). Our experimental evaluation shows some unexpected results, such as the fact that the CPU significantly outperforms the Edge TPU for small model sizes. | ['Marius Portmann', 'Raja Jurdak', 'Mohanad Sarhan', 'Siamak Layeghy', 'Seyedehfaezeh Hosseininoorbin'] | 2021-03-30 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [-2.45838076e-01 -2.12196916e-01 -4.91897374e-01 4.37220000e-02
4.45663154e-01 -2.02487215e-01 2.84069270e-01 -2.55352855e-01
-4.81050074e-01 2.07964972e-01 -6.88278675e-01 -1.17660975e+00
-1.25628427e-01 -1.14322841e+00 -4.24785525e-01 -5.38532555e-01
-1.14697069e-01 3.90318602e-01 2.51949102e-01 1.26922309e-01
1.60781872e-02 7.43447661e-01 -1.57327414e+00 -2.59659588e-01
7.19925091e-02 1.92340446e+00 -5.18407643e-01 7.44890809e-01
-2.68229321e-02 8.36975813e-01 -7.42308378e-01 -3.33686084e-01
4.60049629e-01 8.77907947e-02 -4.47778344e-01 -5.73274255e-01
2.41853163e-01 -5.83689630e-01 -7.95498550e-01 7.32024431e-01
7.68348098e-01 -2.71652509e-02 4.00658250e-01 -1.90558290e+00
-1.71846896e-01 4.61648464e-01 -2.59810448e-01 8.49381268e-01
-4.00875688e-01 1.34105027e-01 3.88514847e-01 -6.93712980e-02
1.46346420e-01 9.87562239e-01 8.77388656e-01 5.02980471e-01
-3.66068304e-01 -9.69157338e-01 -3.58144999e-01 7.50015914e-01
-1.30968583e+00 -5.86926758e-01 8.40678632e-01 1.70739546e-01
1.66466391e+00 1.34197399e-01 6.24245465e-01 1.29892623e+00
8.15851510e-01 4.68601376e-01 7.96087444e-01 -4.97154057e-01
1.00161397e+00 -1.71195101e-02 5.22201300e-01 8.61768425e-01
8.28067839e-01 5.23019195e-01 -1.97918639e-01 -4.62165810e-02
4.21787888e-01 1.83067888e-01 5.60644925e-01 2.46147171e-01
-3.01243782e-01 6.24612451e-01 5.36100209e-01 4.39650714e-01
-5.90767920e-01 7.48622715e-01 1.00077367e+00 2.48294994e-01
3.67386520e-01 -1.90456823e-01 -9.72056627e-01 -5.71598232e-01
-5.59648275e-01 -6.67970300e-01 1.24793756e+00 1.00175929e+00
4.78335023e-01 5.95664740e-01 3.12088132e-01 8.72500092e-02
4.66767251e-01 2.62945533e-01 3.56218964e-01 -6.18123472e-01
1.39634442e-02 6.60826623e-01 -4.89445359e-01 -6.31889224e-01
-7.93794274e-01 -2.35189453e-01 -7.73490965e-01 4.98835683e-01
-1.51525037e-02 -8.59001637e-01 -8.62345934e-01 1.26263499e+00
3.33633840e-01 5.82175374e-01 1.95827469e-01 2.31391907e-01
1.02898467e+00 4.51750249e-01 3.65039170e-01 2.31592298e-01
1.70748961e+00 -1.13362026e+00 -4.73400712e-01 -2.16571450e-01
7.25783110e-01 -1.47783846e-01 4.10000235e-01 8.86690989e-02
-7.63299882e-01 -5.31402349e-01 -1.35298193e+00 7.20722079e-02
-1.17399490e+00 -4.91654575e-02 1.07667315e+00 1.57551682e+00
-1.19901991e+00 5.35721838e-01 -1.09897625e+00 -6.80932820e-01
4.80108231e-01 7.28720129e-01 3.72580767e-01 1.56973138e-01
-1.05622184e+00 7.59637475e-01 5.79179645e-01 -4.61414605e-01
-6.34279966e-01 -4.06884283e-01 -4.48625654e-01 4.67808872e-01
1.29019693e-01 -8.23246777e-01 1.16544521e+00 -7.31665373e-01
-1.63483012e+00 5.42266548e-01 4.70032156e-01 -7.36202419e-01
2.19527960e-01 2.83927441e-01 -8.87758791e-01 7.86502007e-03
-7.46570587e-01 6.79738283e-01 7.19113648e-01 -4.36417907e-01
-8.41183782e-01 -3.45310509e-01 3.84694427e-01 -4.10971493e-01
-7.17985749e-01 -1.50845349e-01 -1.49250552e-01 -2.29514748e-01
-5.40141404e-01 -9.71431851e-01 6.59905300e-02 -1.79538187e-02
-1.20897606e-01 -7.36070096e-01 1.74433982e+00 -2.39834785e-02
1.08936858e+00 -1.93408573e+00 -1.04121459e+00 7.77469128e-02
1.89744070e-01 5.94632089e-01 8.83454680e-02 -2.64415201e-02
5.31110503e-02 1.80249587e-01 4.45386976e-01 -1.58259138e-01
2.78082848e-01 4.98098016e-01 1.93885818e-01 2.72010684e-01
-3.41847837e-01 9.24810767e-01 -4.80437666e-01 -2.90757000e-01
5.26154280e-01 9.01619911e-01 -4.07909185e-01 -1.71335861e-01
9.97797772e-02 -1.62199885e-01 -5.57422101e-01 1.11697876e+00
4.94502217e-01 -1.72207847e-01 1.63708068e-02 -6.02188349e-01
-1.15813836e-01 3.81069690e-01 -6.55995667e-01 1.13156605e+00
-7.43033826e-01 9.24199283e-01 -8.51028413e-02 -1.13483441e+00
7.16582179e-01 5.19303203e-01 7.59365201e-01 -1.23665059e+00
1.04904222e+00 2.59830862e-01 2.86725909e-01 -3.53141367e-01
5.13802618e-02 2.87583947e-01 2.25698128e-01 8.33706558e-01
8.96992758e-02 7.50552177e-01 9.45027471e-02 -3.18736970e-01
1.76597142e+00 -2.05286577e-01 1.87868983e-01 -6.41984701e-01
7.87385553e-02 -8.98037013e-03 1.68882266e-01 7.48361766e-01
-8.30355585e-01 -5.84918141e-01 3.44273895e-01 -1.17097282e+00
-8.74322176e-01 -6.99456275e-01 -4.24559355e-01 1.04812884e+00
1.16380736e-01 -2.52301872e-01 -1.09298933e+00 -9.53564882e-01
-3.55245382e-01 7.73081660e-01 -4.45320070e-01 -4.29202735e-01
-3.92630100e-01 -1.15745389e+00 8.19877803e-01 8.04398000e-01
1.13769948e+00 -1.18449140e+00 -1.68844652e+00 1.64846659e-01
4.72731322e-01 -1.45090652e+00 1.43741310e-01 7.85897434e-01
-1.08705890e+00 -1.31874907e+00 5.36722004e-01 -6.96074128e-01
3.15278292e-01 2.68840581e-01 1.26041579e+00 3.35104942e-01
-5.26080012e-01 7.47580528e-01 -3.66357714e-01 -1.00947320e+00
-6.11957461e-02 2.16133744e-01 4.23963904e-01 -5.03435612e-01
1.32249260e+00 -8.42068970e-01 -8.22741449e-01 1.88497156e-01
-7.91500568e-01 -3.26559842e-01 7.94266522e-01 3.17268133e-01
1.88811481e-01 7.86584914e-01 5.71183085e-01 -4.20342505e-01
2.86907524e-01 -7.95624793e-01 -5.80977023e-01 7.62237061e-04
-1.08100188e+00 -1.45986199e-01 7.17323720e-01 -5.09294808e-01
-5.09410679e-01 -2.12803945e-01 -3.30759615e-01 -3.68667364e-01
-1.85646787e-01 -1.64308935e-01 -2.72755802e-01 -6.11553490e-01
5.81921697e-01 -1.47602618e-01 -4.97917503e-01 -2.52914846e-01
-3.02433074e-01 1.02230036e+00 1.75374120e-01 -4.09089506e-01
3.35829228e-01 5.00423789e-01 3.98459435e-01 -1.00331604e+00
-1.80846572e-01 3.48402448e-02 -2.41162375e-01 -3.08381408e-01
6.28944457e-01 -6.40142977e-01 -1.30605495e+00 3.07414442e-01
-7.65906155e-01 -4.92351443e-01 -2.05010951e-01 3.83091748e-01
-1.62443563e-01 -8.48266557e-02 -7.55629539e-01 -6.31846607e-01
-1.17958069e+00 -1.19964111e+00 7.65015006e-01 4.75801855e-01
5.50833493e-02 -1.34892905e+00 -1.36650562e-01 1.55969962e-01
8.21345091e-01 2.47535780e-01 1.01132584e+00 -9.67667103e-01
-4.16677237e-01 -4.96765703e-01 -5.21883965e-01 2.70219117e-01
-1.22879699e-01 2.11722329e-02 -1.14745092e+00 -2.60607749e-01
3.83594006e-01 8.77212659e-02 5.33669710e-01 5.31720579e-01
1.32918012e+00 -2.47952059e-01 -6.40487731e-01 8.47532332e-01
1.73120320e+00 1.01385403e+00 7.63226569e-01 5.88752866e-01
5.28313756e-01 -2.74284869e-01 -1.36316314e-01 3.31359863e-01
8.19493830e-02 5.03287852e-01 9.23033655e-01 2.98119672e-02
-6.88769817e-02 3.26058030e-01 4.62076336e-01 7.58445323e-01
9.26269814e-02 -7.31511235e-01 -1.01045310e+00 9.21434537e-02
-1.41001153e+00 -5.42103350e-01 3.97033617e-02 1.74028945e+00
-1.94020465e-01 5.65857887e-01 1.46242827e-01 5.02512753e-01
5.46914160e-01 -1.03284486e-01 -1.00888062e+00 -1.08559334e+00
4.59242910e-01 6.69439495e-01 9.24850523e-01 -5.13902530e-02
-1.13245869e+00 6.04888856e-01 6.79990625e+00 9.14001882e-01
-1.38419425e+00 5.12036502e-01 7.09232688e-01 -1.43177405e-01
4.70320225e-01 -4.12276655e-01 -7.38516569e-01 8.03948939e-01
1.98873472e+00 5.00673652e-02 3.77981842e-01 1.50484157e+00
-1.18027359e-01 -8.13589618e-02 -1.12520206e+00 1.14676106e+00
-1.79846212e-01 -1.19398689e+00 -3.01432222e-01 3.72003317e-01
2.90202588e-01 6.57298505e-01 -6.46860525e-02 6.18637145e-01
2.09879130e-01 -7.46770442e-01 3.84409219e-01 -1.84417576e-01
6.88556015e-01 -1.32314730e+00 1.01853919e+00 3.38014849e-02
-1.34674144e+00 -3.98345619e-01 -4.97486085e-01 -4.39190060e-01
-3.50972980e-01 6.46512449e-01 -6.23211026e-01 -2.60789003e-02
1.11407483e+00 1.94762364e-01 -4.95367855e-01 6.24694824e-01
2.06283331e-01 6.41497970e-01 -7.42689550e-01 -5.43776274e-01
3.54335159e-01 -6.25891089e-02 1.69705510e-01 1.27164400e+00
4.47789788e-01 3.51086929e-02 -8.00226256e-03 2.85952836e-01
-4.83857930e-01 -3.59789431e-01 -7.09468663e-01 -2.56890524e-03
7.28001118e-01 1.63433003e+00 -1.20197904e+00 -5.02779722e-01
-6.66703820e-01 6.75474942e-01 -2.60788321e-01 -1.53749391e-01
-1.16498363e+00 -4.49246258e-01 8.66133451e-01 -7.08269626e-02
1.10387139e-01 -1.84142161e-02 -6.87341511e-01 -6.60484493e-01
-1.58457503e-01 -3.23730528e-01 3.18417788e-01 -5.01414895e-01
-8.24298680e-01 8.27632129e-01 -1.66275352e-01 -7.47699082e-01
1.53625473e-01 -1.08570731e+00 -9.62630928e-01 -8.55029821e-02
-1.40966630e+00 -8.05503786e-01 -4.70273376e-01 5.41335225e-01
3.72537524e-01 -4.24760044e-01 8.14875662e-01 9.29329097e-01
-1.24016333e+00 9.04785156e-01 9.30485427e-02 1.14725485e-01
-2.91525066e-01 -8.53601754e-01 1.01198483e+00 7.13198960e-01
-5.97325973e-02 5.40821366e-02 4.33926463e-01 -3.93724144e-01
-1.89501727e+00 -1.24422503e+00 2.37438321e-01 -4.22973692e-01
6.50520861e-01 -1.69243410e-01 -1.52275145e-01 8.76200855e-01
3.24523151e-01 3.72068942e-01 7.38986135e-01 -3.67709607e-01
-4.63559590e-02 -3.25780183e-01 -1.70696330e+00 6.95893288e-01
1.18583786e+00 -2.01602295e-01 1.12798542e-01 4.07995343e-01
3.95971388e-01 -1.86858755e-02 -8.46386254e-01 3.95338297e-01
6.88137293e-01 -9.13782537e-01 7.82504678e-01 -4.73481357e-01
-3.83362114e-01 3.60389687e-02 -2.06284374e-01 -8.15105975e-01
-1.98230192e-01 -3.19525093e-01 -1.04799092e+00 9.67499375e-01
-1.35195166e-01 -8.75544608e-01 1.26495707e+00 4.00779486e-01
-1.94048882e-01 -8.56298625e-01 -1.42561233e+00 -1.13504159e+00
-4.22584385e-01 -7.63224185e-01 6.03563070e-01 6.70256555e-01
-3.82538475e-02 5.53014994e-01 1.12163864e-01 4.69373725e-02
6.05246842e-01 -4.68606114e-01 3.71300846e-01 -1.55320084e+00
1.51486516e-01 -4.41362977e-01 -9.93884146e-01 -4.51898336e-01
-8.90563242e-03 -5.14749348e-01 -4.28716302e-01 -1.17448819e+00
-2.04550475e-01 -5.50530612e-01 -7.06283808e-01 7.42155612e-01
6.13209069e-01 3.65455478e-01 -2.35335395e-01 -1.70727924e-01
-6.63561583e-01 -8.19054060e-03 3.18402499e-01 -1.42242566e-01
5.48339188e-02 -7.13594928e-02 -4.86512154e-01 9.84980047e-01
1.31762600e+00 -7.07601607e-01 -5.37167847e-01 -4.55383658e-01
4.86229956e-02 -4.52158391e-01 1.55016884e-01 -1.81200969e+00
5.99290252e-01 2.89263576e-01 4.06374961e-01 -1.86465845e-01
1.69537708e-01 -1.59828949e+00 2.63499886e-01 1.13666475e+00
7.22997129e-01 5.89208782e-01 5.42176843e-01 2.63143480e-01
5.42376995e-01 -1.01882748e-01 9.54752445e-01 2.12439746e-01
-1.05353689e+00 6.52275145e-01 -6.34441733e-01 -2.84098625e-01
1.44069767e+00 -4.65829551e-01 -6.50433123e-01 2.84353793e-01
-2.13823825e-01 -4.17996943e-01 2.29744941e-01 4.21738684e-01
3.22785288e-01 -1.24849951e+00 1.57281046e-03 3.06122363e-01
-3.37344021e-01 -5.56286693e-01 2.12931875e-02 6.68288410e-01
-9.45251942e-01 7.79305160e-01 -6.29480422e-01 -3.47442061e-01
-1.10545230e+00 9.39801574e-01 7.55061746e-01 -2.43541166e-01
-7.69125342e-01 4.97999519e-01 -3.93683493e-01 -5.26833422e-02
6.14635348e-01 -9.72683057e-02 1.38864636e-01 -1.92386031e-01
6.30560935e-01 1.05847514e+00 5.95767260e-01 -1.07322317e-02
-6.88153863e-01 3.71235728e-01 1.43111378e-01 5.20450771e-01
1.18654847e+00 1.54107526e-01 8.01060349e-02 -1.56271771e-01
1.25702167e+00 -8.11845183e-01 -9.57021356e-01 1.25687018e-01
1.53863251e-01 3.29819649e-01 8.08413923e-01 -7.12155759e-01
-1.77268243e+00 7.34227479e-01 1.52085853e+00 4.79593664e-01
1.47065473e+00 -4.97408062e-01 1.32733893e+00 8.17508578e-01
6.26789868e-01 -1.25843346e+00 3.12075242e-02 6.47781432e-01
-3.28586161e-01 -1.03687263e+00 -1.84865639e-01 -6.65620193e-02
4.16596383e-01 1.39861965e+00 9.28298235e-01 -1.66780487e-01
1.20418596e+00 9.49883103e-01 -1.92951620e-01 -4.68651414e-01
-7.46053517e-01 -1.17464520e-01 -4.57077503e-01 7.77742922e-01
-9.10648406e-02 1.80084392e-01 -5.43333367e-02 2.35345259e-01
-4.33048159e-02 2.03510925e-01 4.48257595e-01 1.38651764e+00
-5.23604393e-01 -7.94169486e-01 -1.32430106e-01 6.63455784e-01
-8.65354776e-01 -1.28563449e-01 -2.13560045e-01 8.43121171e-01
4.82654154e-01 1.17752945e+00 6.64891064e-01 -9.17305052e-01
5.37237264e-02 9.40374881e-02 3.47463846e-01 -3.99697348e-02
-7.16048062e-01 -7.33496010e-01 6.10665679e-02 -5.50162733e-01
4.48037460e-02 1.50286397e-02 -1.17094970e+00 -1.18877053e+00
-2.21982434e-01 -2.71388069e-02 1.46568763e+00 1.02378905e+00
5.64532399e-01 1.12776959e+00 5.66776097e-01 -9.05804276e-01
1.87234685e-01 -7.33577430e-01 -4.58539367e-01 -4.58406180e-01
-2.25700662e-02 -7.48132288e-01 -5.39039314e-01 -5.42566061e-01] | [8.293167114257812, 2.744737148284912] |
41c19346-782b-4ae4-b55f-c56c3b403724 | cnn-based-real-time-2d-3d-deformable | 2212.07692 | null | https://arxiv.org/abs/2212.07692v2 | https://arxiv.org/pdf/2212.07692v2.pdf | CNN-based real-time 2D-3D deformable registration from a single X-ray projection | Purpose: The purpose of this paper is to present a method for real-time 2D-3D non-rigid registration using a single fluoroscopic image. Such a method can find applications in surgery, interventional radiology and radiotherapy. By estimating a three-dimensional displacement field from a 2D X-ray image, anatomical structures segmented in the preoperative scan can be projected onto the 2D image, thus providing a mixed reality view. Methods: A dataset composed of displacement fields and 2D projections of the anatomy is generated from the preoperative scan. From this dataset, a neural network is trained to recover the unknown 3D displacement field from a single projection image. Results: Our method is validated on lung 4D CT data at different stages of the lung deformation. The training is performed on a 3D CT using random (non domain-specific) diffeomorphic deformations, to which perturbations mimicking the pose uncertainty are added. The model achieves a mean TRE over a series of landmarks ranging from 2.3 to 5.5 mm depending on the amplitude of deformation. Conclusion: In this paper, a CNN-based method for real-time 2D-3D non-rigid registration is presented. This method is able to cope with pose estimation uncertainties, making it applicable to actual clinical scenarios, such as lung surgery, where the C-arm pose is planned before the intervention. | ['Stéphane Cotin', 'Jean-Louis Dillenseger', 'François Lecomte'] | 2022-12-15 | null | null | null | null | ['mixed-reality'] | ['computer-vision'] | [ 1.97261542e-01 6.31333649e-01 1.34431822e-02 -2.83288568e-01
-7.64303684e-01 -5.76684952e-01 4.52104658e-01 1.09977357e-01
-6.44744217e-01 3.66764486e-01 8.02749097e-02 -2.59261221e-01
-2.72903174e-01 -6.37377620e-01 -8.86353970e-01 -7.28090942e-01
-7.90365860e-02 1.17902637e+00 2.06491947e-01 -2.36997291e-01
-1.20595589e-01 1.07642722e+00 -8.88086557e-01 9.70672234e-04
4.67453897e-01 5.90875685e-01 3.33093464e-01 6.12267733e-01
2.20034301e-01 6.89728409e-02 -2.61910081e-01 9.05753151e-02
5.00197649e-01 -2.87187040e-01 -7.28504002e-01 6.84752986e-02
3.94676119e-01 -3.99329662e-01 -1.97502226e-01 8.86550963e-01
8.26725602e-01 3.16469409e-02 4.56430703e-01 -1.89035386e-01
4.17915136e-02 4.50506270e-01 -1.26650229e-01 -7.20767453e-02
4.63682681e-01 5.02808997e-03 3.72112705e-03 -8.32176626e-01
9.19174373e-01 7.39841521e-01 9.39994097e-01 7.71265626e-01
-1.22507548e+00 -1.18358061e-01 -5.34551919e-01 -4.09259468e-01
-1.12968135e+00 2.03288063e-01 6.66969419e-01 -7.32737243e-01
4.83157307e-01 5.22337079e-01 8.64222050e-01 6.95964992e-01
9.88987923e-01 -5.30776568e-03 1.37932467e+00 -4.99045640e-01
2.17577800e-01 6.81812614e-02 -3.02148879e-01 6.59426570e-01
1.84233531e-01 5.05736172e-01 2.06491873e-01 -2.38789007e-01
1.17161679e+00 2.17971161e-01 -7.26618886e-01 -6.27760231e-01
-1.21463513e+00 4.93301570e-01 8.66079628e-01 8.18742692e-01
-6.25268877e-01 1.14022054e-01 1.24385446e-01 -1.54062077e-01
3.56607884e-01 3.48869532e-01 -1.96225286e-01 -7.46650398e-02
-8.86021256e-01 4.94183414e-02 7.30539858e-01 5.44896781e-01
3.17282826e-01 -1.45985201e-01 -1.79369077e-01 2.73752779e-01
4.89661098e-01 4.87651706e-01 1.05533469e+00 -5.38076162e-01
1.99522287e-01 3.90157253e-01 -9.24609005e-02 -6.17604256e-01
-9.77699995e-01 -4.87981170e-01 -8.85874033e-01 3.94336939e-01
5.24470985e-01 -7.25537464e-02 -1.29449046e+00 1.46205544e+00
6.85837030e-01 3.38427454e-01 -1.24667667e-01 1.23017311e+00
8.79971206e-01 -6.54444620e-02 -4.30876553e-01 -2.86156893e-01
1.25891888e+00 -4.17384565e-01 -6.10272586e-01 -1.31053999e-01
6.12822115e-01 -7.77618945e-01 6.60634875e-01 1.51059568e-01
-1.49987793e+00 -3.85233879e-01 -5.98108590e-01 3.71028543e-01
9.22198892e-02 6.57243058e-02 -2.28312463e-01 6.80766404e-01
-9.36514199e-01 9.81195152e-01 -1.46241653e+00 -1.44842789e-01
9.26632136e-02 8.51693094e-01 -6.86410189e-01 5.73042147e-02
-1.13368094e+00 1.40343058e+00 4.42335427e-01 5.52301824e-01
-6.19436383e-01 -8.35510790e-01 -6.71166122e-01 -1.41044185e-01
1.44254401e-01 -7.69719541e-01 1.11280000e+00 -5.58875501e-01
-1.87999928e+00 1.20859337e+00 1.38065785e-01 -4.22181517e-01
9.43773389e-01 -4.96476190e-03 -4.25239429e-02 5.55884801e-02
-2.80408382e-01 1.05689876e-01 4.94098455e-01 -1.23287237e+00
3.38213056e-01 -7.80658245e-01 -2.66426891e-01 3.97613317e-01
3.02670985e-01 -2.21235901e-01 -5.60239196e-01 -5.88619411e-01
6.58062160e-01 -1.55441844e+00 -6.62524164e-01 -6.51380420e-03
-4.50127065e-01 7.81998634e-01 3.62452745e-01 -9.12313700e-01
5.86234808e-01 -1.76979756e+00 5.48081398e-01 4.69810903e-01
2.73000032e-01 1.15044445e-01 3.57182741e-01 5.07320389e-02
-4.57044691e-01 -1.03071541e-01 -6.21450245e-01 -2.91030824e-01
-4.82082695e-01 2.23423928e-01 2.30152041e-01 8.71194839e-01
-4.43564057e-01 1.13747072e+00 -7.02501476e-01 -1.88557729e-01
3.99917096e-01 6.90507889e-01 -3.20385337e-01 2.85191029e-01
7.43128881e-02 1.34366548e+00 -3.27268332e-01 1.80925176e-01
8.22727501e-01 8.30612332e-02 3.02045792e-01 -2.57681519e-01
-1.14547886e-01 7.66428513e-03 -9.43973720e-01 2.19145942e+00
-7.36486197e-01 1.70360073e-01 1.33622155e-01 -6.72483146e-01
7.79988229e-01 6.91916108e-01 1.06444335e+00 -4.70037401e-01
6.14960611e-01 5.67380726e-01 3.40099603e-01 -4.60889369e-01
-1.26440763e-01 -6.05806470e-01 1.09755158e-01 3.40953141e-01
-3.42001021e-02 -9.25289094e-01 -4.38388109e-01 -3.71065021e-01
9.15428162e-01 1.54609963e-01 2.82013208e-01 -3.47708642e-01
6.81364417e-01 -1.28656149e-01 1.46689072e-01 3.79371226e-01
1.65280506e-01 1.12406075e+00 1.66249141e-01 -6.49821758e-01
-9.04574871e-01 -9.80334580e-01 -5.21339893e-01 -3.13275516e-01
5.79277650e-02 3.65690559e-01 -9.54090476e-01 -6.43507719e-01
-1.12455152e-01 5.01403570e-01 -9.09748137e-01 -2.60926485e-01
-1.17458713e+00 -6.54889941e-01 1.69004276e-01 2.49588236e-01
-4.06105369e-02 -8.37837696e-01 -8.32218111e-01 3.24171096e-01
-1.28811318e-02 -1.06007302e+00 -3.75917226e-01 3.17446589e-01
-1.52824664e+00 -1.03448617e+00 -8.91908646e-01 -4.51838583e-01
1.01893485e+00 -3.31210405e-01 1.05228114e+00 -1.44865051e-01
-3.59994829e-01 5.54710150e-01 1.24950156e-01 -3.57246011e-01
-1.02297795e+00 -2.41302073e-01 7.67240599e-02 -2.67117888e-01
-4.03870225e-01 -5.36744475e-01 -5.37063837e-01 3.83536816e-01
-1.01148796e+00 -4.74963672e-02 5.26977181e-01 7.60206401e-01
1.02198064e+00 -3.04684818e-01 -2.11308822e-01 -1.08999979e+00
5.06726086e-01 -2.47694895e-01 -7.98752785e-01 2.55740024e-02
-4.30760264e-01 7.93147758e-02 6.33300364e-01 -6.36169612e-01
-9.86639321e-01 6.07219219e-01 -4.03880179e-01 -7.32447386e-01
-3.30463827e-01 5.29523373e-01 1.15856171e-01 -5.54593801e-01
1.05076897e+00 2.49973051e-02 4.53171223e-01 -3.38044167e-01
1.47249252e-01 -1.17621180e-02 5.47373891e-01 -3.66516769e-01
1.03122997e+00 6.27603948e-01 5.76625884e-01 -2.97085404e-01
-5.42333961e-01 -2.50435859e-01 -1.20531559e+00 -3.62953931e-01
9.80151057e-01 -5.56178510e-01 -5.22801936e-01 1.93154097e-01
-1.09294593e+00 -4.95779365e-01 -8.63574207e-01 1.31852591e+00
-8.27116013e-01 3.17448229e-01 -4.68274772e-01 -6.96109533e-02
-4.47255075e-01 -1.74085617e+00 1.14721417e+00 3.25531587e-02
1.61571614e-02 -1.35378218e+00 5.46364546e-01 -1.13776382e-02
5.56098998e-01 7.03927577e-01 5.52531600e-01 -6.44737184e-01
-4.89239395e-01 -5.33161759e-01 6.32676423e-01 1.51115581e-01
2.20470920e-01 -3.82168680e-01 -6.85146928e-01 -3.41611505e-01
7.25528538e-01 1.70494497e-01 1.76433012e-01 8.75385761e-01
1.06279731e+00 6.85681179e-02 -2.85886228e-01 8.39260280e-01
1.55332804e+00 1.05048098e-01 6.09854698e-01 4.69822325e-02
7.45233595e-01 2.90150970e-01 4.71865445e-01 1.22757874e-01
-2.17110198e-02 1.07653296e+00 9.50290859e-01 -1.67810366e-01
-1.49569899e-01 1.91361174e-01 -8.41193125e-02 6.95842922e-01
-4.19667900e-01 2.01351300e-01 -1.17547524e+00 1.79937646e-01
-1.30022836e+00 -5.27239501e-01 -3.45429629e-01 2.82329988e+00
6.76659524e-01 1.52705694e-02 -3.15095723e-01 -1.95762932e-01
7.55195618e-01 -2.54873246e-01 -4.12226737e-01 -1.87657326e-01
4.83621925e-01 6.13759995e-01 7.73434699e-01 7.98445165e-01
-7.14175999e-01 1.89162910e-01 5.58567905e+00 2.48062968e-01
-1.77064300e+00 2.96135545e-01 2.38194019e-01 2.77338233e-02
-3.10354650e-01 -1.23856202e-01 -3.80189925e-01 3.31443548e-01
9.59706366e-01 -8.37405771e-02 1.12499736e-01 5.77822685e-01
2.27088153e-01 -9.76106599e-02 -1.04441679e+00 6.69531941e-01
1.29187569e-01 -1.33737123e+00 -4.67476577e-01 2.99300700e-01
5.82151473e-01 2.94037491e-01 -3.89657803e-02 -1.50773332e-01
-2.67796636e-01 -1.09697843e+00 4.58085150e-01 8.21782231e-01
1.19221711e+00 -3.00964922e-01 8.26183736e-01 7.86625862e-01
-7.24811196e-01 3.81431818e-01 1.11951545e-01 6.35752141e-01
5.43350816e-01 6.51191533e-01 -1.37168407e+00 9.70235705e-01
2.90076971e-01 2.68633068e-01 -1.84838250e-01 1.07819724e+00
-1.19996503e-01 2.68792778e-01 -5.44018388e-01 4.88000780e-01
-1.61278933e-01 -2.91723728e-01 7.70964384e-01 7.59159982e-01
7.08826780e-01 2.72515953e-01 -9.96203423e-02 7.86072373e-01
-1.08787388e-01 2.35957857e-02 -6.61378145e-01 6.83301866e-01
-5.78246601e-02 1.24175262e+00 -8.53922367e-01 3.30446474e-02
8.45818818e-02 8.01823080e-01 -1.67222574e-01 -1.58581525e-01
-8.63488853e-01 3.34535867e-01 -1.71177208e-01 5.29218853e-01
-1.03843600e-01 -1.30362019e-01 -1.87213168e-01 -1.07281280e+00
1.23693123e-01 -3.80003065e-01 3.25645991e-02 -7.76415825e-01
-6.77976727e-01 9.53141212e-01 2.52886176e-01 -1.45782578e+00
-7.30222523e-01 -6.23534977e-01 -7.36298382e-01 1.11932170e+00
-1.11617160e+00 -7.81333327e-01 -6.44996166e-01 3.51295173e-01
-1.01064062e-02 1.03037111e-01 1.10466194e+00 2.47848898e-01
8.41840282e-02 3.91292460e-02 1.96642116e-01 9.25027113e-03
6.75758064e-01 -1.39790833e+00 3.02706301e-01 4.58297879e-01
-8.45743120e-02 5.94684482e-01 6.22464180e-01 -7.29268491e-01
-1.29946482e+00 -8.75681162e-01 4.85389411e-01 -6.67644620e-01
1.74549907e-01 -1.66900419e-02 -9.17474389e-01 8.58328819e-01
-1.78901091e-01 8.95004690e-01 3.82798284e-01 -7.05599666e-01
4.84147340e-01 1.59485936e-01 -1.64996576e+00 2.02641770e-01
6.13639832e-01 -3.71302783e-01 -5.34535825e-01 5.20062804e-01
4.28659230e-01 -1.77413082e+00 -1.43973148e+00 6.19255900e-01
4.88548815e-01 -9.10564065e-01 1.14558280e+00 -2.64160931e-01
1.03564732e-01 -1.09405629e-01 4.13912535e-01 -1.53432488e+00
9.07758251e-02 -6.30686522e-01 6.86593428e-02 8.40293989e-02
2.34552566e-03 -7.95083225e-01 8.90532136e-01 7.63392448e-01
-3.49188000e-01 -9.24209177e-01 -1.51378191e+00 -5.69897711e-01
3.95887733e-01 -3.25657070e-01 2.36470029e-01 8.05234969e-01
-4.79735941e-01 -2.99896002e-01 2.02095091e-01 4.00755316e-01
3.43402267e-01 -2.10133679e-02 5.25825977e-01 -1.43265092e+00
-4.29571420e-01 -1.58068180e-01 -5.00168502e-01 -7.47506678e-01
8.59976411e-02 -1.09421134e+00 1.93119347e-01 -1.36765563e+00
-1.40570015e-01 -5.25744379e-01 1.15345970e-01 2.59007886e-03
2.14842394e-01 7.80455768e-02 7.38733858e-02 4.49791610e-01
5.69359303e-01 2.31091812e-01 1.81684732e+00 2.80495107e-01
-3.11280191e-01 5.50229847e-01 1.05770946e-01 8.24939013e-01
6.35434210e-01 -7.97159851e-01 -1.95047438e-01 -1.98360369e-01
-1.01027235e-01 7.70474911e-01 4.04116571e-01 -1.05117214e+00
1.63415760e-01 1.72646821e-01 2.05705211e-01 -4.85728592e-01
4.34569508e-01 -1.53748035e+00 7.59030402e-01 9.99798238e-01
-1.33853212e-01 1.09779209e-01 3.44864339e-01 3.48640591e-01
-2.48637304e-01 -6.05654955e-01 9.43726897e-01 -4.45891857e-01
3.20068419e-01 5.87071300e-01 -7.72695318e-02 -7.01265335e-02
9.98101890e-01 -2.85313815e-01 3.43110710e-01 -7.68915787e-02
-1.39054644e+00 -6.12200499e-01 5.81717014e-01 -7.55018070e-02
6.71065867e-01 -1.34988034e+00 -6.46176100e-01 3.10671389e-01
-2.80096829e-01 7.77774453e-01 5.07629931e-01 1.42645633e+00
-9.57163513e-01 3.81024063e-01 -2.07750738e-01 -1.05201769e+00
-1.18943250e+00 5.17115831e-01 1.31625068e+00 -7.51573920e-01
-7.44092643e-01 5.17236531e-01 1.24760062e-01 -8.36699307e-01
-1.57841802e-01 -7.14110017e-01 -2.08736941e-01 -5.87850988e-01
1.15880277e-03 -1.74752444e-01 7.06079662e-01 -1.02636671e+00
-4.52785909e-01 1.14367211e+00 2.86294281e-01 -2.94571757e-01
1.29908085e+00 1.35676146e-01 -2.56815180e-02 2.69861400e-01
1.21055734e+00 3.26692224e-01 -1.01256466e+00 -1.50944799e-01
-3.51489991e-01 -4.47653145e-01 1.34649679e-01 -7.14939833e-01
-1.38549960e+00 7.89220691e-01 9.91602123e-01 -1.59222245e-01
1.04320228e+00 -1.16102368e-01 4.39545512e-01 -3.35418843e-02
3.26296449e-01 -4.30032343e-01 -1.98405206e-01 3.42529476e-01
1.19860268e+00 -9.27744746e-01 1.52515113e-01 -4.48143810e-01
-4.01456684e-01 1.52616775e+00 2.07816660e-01 -1.43374741e-01
8.83327544e-01 4.88558084e-01 2.06752315e-01 -3.30930382e-01
1.10463895e-01 1.41169548e-01 6.52419209e-01 3.95637989e-01
4.21368748e-01 4.05307487e-02 -4.42074597e-01 -5.11722304e-02
-7.64387548e-02 2.07368851e-01 7.20580041e-01 1.01782668e+00
1.36693075e-01 -1.24489522e+00 -7.50026941e-01 -5.03682792e-02
-6.21757984e-01 2.71495819e-01 1.17575146e-01 9.11452651e-01
1.04559831e-01 -3.28124836e-02 -4.20046262e-02 1.23914197e-01
7.52779663e-01 -1.03870206e-01 8.79672825e-01 -8.20690095e-01
-1.04630435e+00 2.07044870e-01 -2.99280703e-01 -7.50547349e-01
-2.97600359e-01 -5.58614612e-01 -1.62449706e+00 3.56861562e-01
-4.14553672e-01 -1.18980454e-02 1.26503193e+00 7.96675384e-01
1.07876606e-01 6.83170140e-01 5.77166438e-01 -1.32507229e+00
-6.00200415e-01 -6.79218650e-01 -4.26735640e-01 4.78870690e-01
2.54137009e-01 -5.57178497e-01 -4.13789451e-01 -1.51049882e-01] | [13.82205867767334, -2.767116069793701] |
8ba3b5a5-1fd1-4bf7-a217-996519743ef8 | c-saw-a-framework-for-graph-sampling-and | 2009.09103 | null | https://arxiv.org/abs/2009.09103v1 | https://arxiv.org/pdf/2009.09103v1.pdf | C-SAW: A Framework for Graph Sampling and Random Walk on GPUs | Many applications require to learn, mine, analyze and visualize large-scale graphs. These graphs are often too large to be addressed efficiently using conventional graph processing technologies. Many applications have requirements to analyze, transform, visualize and learn large scale graphs. These graphs are often too large to be addressed efficiently using conventional graph processing technologies. Recent literatures convey that graph sampling/random walk could be an efficient solution. In this paper, we propose, to the best of our knowledge, the first GPU-based framework for graph sampling/random walk. First, our framework provides a generic API which allows users to implement a wide range of sampling and random walk algorithms with ease. Second, offloading this framework on GPU, we introduce warp-centric parallel selection, and two optimizations for collision migration. Third, towards supporting graphs that exceed GPU memory capacity, we introduce efficient data transfer optimizations for out-of-memory sampling, such as workload-aware scheduling and batched multi-instance sampling. In its entirety, our framework constantly outperforms the state-of-the-art projects. First, our framework provides a generic API which allows users to implement a wide range of sampling and random walk algorithms with ease. Second, offloading this framework on GPU, we introduce warp-centric parallel selection, and two novel optimizations for collision migration. Third, towards supporting graphs that exceed the GPU memory capacity, we introduce efficient data transfer optimizations for out-of-memory and multi-GPU sampling, such as workload-aware scheduling and batched multi-instance sampling. Taken together, our framework constantly outperforms the state of the art projects in addition to the capability of supporting a wide range of sampling and random walk algorithms. | ['Hang Liu', 'Xiaoye S. Li', 'Adolfy Hoisie', 'Lingda Li', 'Santosh Pandey'] | 2020-09-18 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [-1.23801455e-01 -2.49502674e-01 -1.24622829e-01 -1.26713822e-02
-4.92713094e-01 -6.39429331e-01 3.55950594e-01 5.27904332e-01
-2.72567540e-01 2.97553092e-01 -1.80235237e-01 -8.00533712e-01
-1.33700535e-01 -1.50583065e+00 -5.05464911e-01 -3.71101469e-01
-5.39039016e-01 7.24560261e-01 8.90456378e-01 -1.46906763e-01
2.72369444e-01 8.32875609e-01 -1.77642524e+00 2.53569901e-01
6.80628240e-01 5.64545989e-01 2.27883428e-01 9.10100162e-01
-5.09087563e-01 1.69489846e-01 -3.34436953e-01 -1.21705413e-01
2.39654273e-01 -2.35030457e-01 -8.78966033e-01 -6.82184566e-03
5.13477862e-01 -2.05844834e-01 1.07836679e-01 8.27681005e-01
7.31367350e-01 1.45470083e-01 1.48112237e-01 -1.44860530e+00
2.07421124e-01 4.71109092e-01 -9.44023371e-01 1.42214134e-01
4.60659742e-01 3.84511977e-01 8.42485487e-01 -7.78558850e-01
8.50781262e-01 1.07331419e+00 6.44897521e-01 -9.72735137e-02
-1.25276828e+00 -4.72796947e-01 2.82888889e-01 6.90488890e-02
-1.29168892e+00 -2.62310803e-02 7.56797433e-01 -3.46302211e-01
1.33438230e+00 8.21732998e-01 1.13012624e+00 4.18811977e-01
1.79375917e-01 5.42669237e-01 8.05268943e-01 -3.42347473e-01
4.02851909e-01 -3.11038941e-01 3.41824055e-01 1.12215352e+00
3.49378735e-01 -9.46701393e-02 -4.49000329e-01 -4.36821163e-01
8.46718252e-01 -6.21769093e-02 -1.46130219e-01 -6.82202816e-01
-1.40213656e+00 6.92786992e-01 3.61661315e-01 1.49768516e-01
-1.94825217e-01 4.33843613e-01 6.26561701e-01 1.82507098e-01
3.86747390e-01 3.73324454e-01 -3.13811004e-01 -4.25543040e-01
-1.28426325e+00 1.97896197e-01 1.01436508e+00 1.23175764e+00
1.43620050e+00 -8.35246146e-02 -1.88524693e-01 5.21279693e-01
-7.96999186e-02 4.38486278e-01 1.11634523e-01 -7.08891869e-01
4.35164690e-01 9.54720497e-01 -3.98822516e-01 -1.20147085e+00
-9.01114106e-01 -3.80993962e-01 -1.09059322e+00 3.14344764e-01
1.87122300e-01 -2.92018764e-02 -7.21426547e-01 1.20189202e+00
8.17560971e-01 2.14452609e-01 -4.77958679e-01 7.46604741e-01
7.32101023e-01 7.05078065e-01 -3.18111807e-01 -3.40652257e-01
1.60940444e+00 -1.21510673e+00 -1.82490170e-01 -1.45185918e-01
9.61406767e-01 -9.21367586e-01 1.55262220e+00 4.20939773e-01
-1.00484800e+00 -3.64168137e-01 -1.00963449e+00 -1.01259276e-01
-4.00197417e-01 -1.68735072e-01 1.04254901e+00 7.00858295e-01
-1.22295487e+00 6.67272627e-01 -1.29395580e+00 -5.91401637e-01
1.25708431e-01 3.58793020e-01 -3.21233571e-01 -1.09446526e-01
-3.39834005e-01 2.85437882e-01 4.12854046e-01 -3.49410206e-01
-3.15118551e-01 -1.03114605e+00 -7.38741875e-01 1.76954776e-01
6.60570383e-01 -9.72816110e-01 7.81724632e-01 -3.55082363e-01
-1.06892431e+00 5.95882356e-01 -2.17194289e-01 -2.19780415e-01
4.61192966e-01 2.41319507e-01 -9.05152261e-02 1.19698428e-01
4.23049890e-02 2.15387017e-01 4.20847923e-01 -9.55933213e-01
-2.50551522e-01 -3.60270917e-01 -3.62576284e-02 2.57440209e-01
-5.02025664e-01 -1.84432253e-01 -8.24125707e-01 -4.41034198e-01
-1.18938677e-01 -9.23258305e-01 -4.27367359e-01 -7.58461952e-02
-4.68679696e-01 1.03380485e-02 1.28357244e+00 -7.10862204e-02
1.62731016e+00 -1.98745549e+00 1.45407971e-02 5.07771075e-01
6.39767647e-01 9.79331359e-02 8.65468383e-02 8.01114798e-01
-4.00965149e-03 1.01461612e-01 -1.66519657e-01 -2.66378164e-01
7.37248212e-02 2.34605730e-01 -1.34235159e-01 3.26321632e-01
-2.03187987e-01 6.73216760e-01 -1.12877047e+00 -4.24031466e-01
3.89716923e-01 3.58332306e-01 -7.25899100e-01 3.93926613e-02
-3.96401793e-01 2.97885053e-02 -4.24660265e-01 4.13388997e-01
9.80402470e-01 -5.17337918e-01 3.97754043e-01 -3.90337199e-01
-3.30746502e-01 2.78254854e-03 -1.55318892e+00 1.66140318e+00
-4.98747796e-01 2.75669307e-01 1.55519575e-01 -7.52176166e-01
8.15642238e-01 -2.98010141e-01 4.83311981e-01 -4.53807294e-01
-2.87898153e-01 3.09452415e-01 -3.53797436e-01 5.76543957e-02
7.62597978e-01 3.15268576e-01 -1.43604670e-02 1.09024799e+00
-3.34841251e-01 -4.81462955e-01 7.57474244e-01 6.29267693e-01
1.75618863e+00 5.45626022e-02 3.03944796e-01 -4.99420613e-01
2.74984926e-01 2.64536887e-01 8.98315571e-03 4.93909180e-01
4.06493396e-01 3.16140026e-01 8.34588051e-01 -8.29015076e-01
-8.65170419e-01 -9.94483590e-01 4.11437094e-01 1.34249771e+00
7.33501390e-02 -1.19821346e+00 -7.24256277e-01 -6.92307234e-01
-6.04449324e-02 5.13674140e-01 -2.07739592e-01 2.73821682e-01
-6.37641370e-01 -6.58131242e-01 1.11607485e-01 4.19647664e-01
4.93830174e-01 -9.01202559e-01 -8.94183695e-01 4.12962705e-01
3.89469177e-01 -9.05836761e-01 -6.85974240e-01 -2.03054175e-02
-1.11096239e+00 -1.25639272e+00 -2.78343588e-01 -4.30344433e-01
8.57628047e-01 7.32036650e-01 1.46609724e+00 5.36230505e-01
-7.73504555e-01 5.50633609e-01 -1.83519423e-01 -5.44413961e-02
-7.97494277e-02 4.54802901e-01 -4.63620842e-01 -3.69604796e-01
-9.75590423e-02 -9.41702008e-01 -6.67922020e-01 4.51237828e-01
-9.27698553e-01 4.54640687e-01 9.59288105e-02 5.03929198e-01
1.00258207e+00 1.30821839e-01 -1.64558083e-01 -1.14170825e+00
8.36618364e-01 -3.12132806e-01 -9.06095743e-01 2.34451920e-01
-6.16509974e-01 2.14187250e-01 7.31606543e-01 -3.01834404e-01
-3.94400656e-01 5.54771312e-02 -3.50119919e-02 -4.68250990e-01
2.23616183e-01 5.33335567e-01 -6.15202896e-02 -9.06675383e-02
7.32843161e-01 8.75624716e-02 -5.23315147e-02 -2.52023160e-01
5.34017324e-01 5.01117259e-02 -1.66849904e-02 -9.23203051e-01
6.66750908e-01 4.30695921e-01 4.32553262e-01 -7.33876944e-01
-2.54145741e-01 -6.63760185e-01 -4.57533985e-01 -2.88223058e-01
6.65427804e-01 -4.67786133e-01 -6.72391415e-01 3.00208211e-01
-8.95260274e-01 -8.28324080e-01 -2.63756424e-01 1.54827893e-01
-5.64987004e-01 4.58626598e-01 -4.37993705e-01 -5.60988367e-01
-6.67643368e-01 -1.49265337e+00 1.38167107e+00 4.14567580e-03
-4.21555996e-01 -1.06618547e+00 -1.86828393e-02 -1.42355070e-01
6.35676384e-01 6.60184562e-01 7.86430597e-01 -3.30991775e-01
-1.03206563e+00 -7.65467808e-02 -4.69250023e-01 -6.26891434e-01
-2.03240407e-03 4.05198216e-01 -2.75830299e-01 -7.37777948e-01
-6.75891519e-01 1.72940299e-01 5.87930918e-01 -5.45272045e-02
1.46644998e+00 -1.64960414e-01 -8.13377321e-01 9.26895142e-01
1.43144643e+00 -1.91153795e-01 5.06708801e-01 2.80114561e-01
1.06024051e+00 2.07018867e-01 5.38986981e-01 4.73457038e-01
5.26249290e-01 8.62255096e-01 5.34770250e-01 -2.44148970e-01
-2.72308886e-01 -3.76901776e-02 1.00936003e-01 1.03054667e+00
-2.00555578e-01 -7.23906904e-02 -1.21169603e+00 4.06858355e-01
-1.89423561e+00 -7.67001867e-01 -7.14600742e-01 2.39222717e+00
3.61712247e-01 1.53668299e-01 5.23099959e-01 1.93692625e-01
4.73113984e-01 4.33602482e-01 -4.21624333e-01 -7.92365313e-01
4.68968511e-01 6.86794043e-01 5.85935712e-01 5.48901916e-01
-7.55992293e-01 9.20151114e-01 4.85797739e+00 1.12374365e+00
-1.21830833e+00 9.20963939e-03 3.07240725e-01 -2.19847351e-01
-5.90852737e-01 2.98227966e-01 -7.71824241e-01 2.73094743e-01
9.19687688e-01 -5.83977997e-01 6.40173912e-01 1.18843126e+00
1.91202573e-02 -1.93980902e-01 -1.09397435e+00 9.61261153e-01
-1.42288581e-01 -1.63846195e+00 -9.56582278e-02 4.36727881e-01
5.56351542e-01 9.08825994e-02 -5.11186659e-01 1.07194424e-01
5.34004390e-01 -7.64820039e-01 3.41670156e-01 1.33906722e-01
9.81169760e-01 -1.01779270e+00 3.41839224e-01 2.16093108e-01
-1.88571560e+00 5.27976692e-01 -2.01789215e-01 -1.32688537e-01
2.52406240e-01 1.26761997e+00 -6.38775170e-01 8.57137442e-01
5.89665055e-01 4.42101717e-01 -6.69713199e-01 9.71717298e-01
-3.34345214e-02 4.24692810e-01 -6.07706964e-01 -1.61205918e-01
1.33055404e-01 -3.14944386e-01 4.88689691e-01 1.37548566e+00
6.76158488e-01 -1.67151868e-01 6.83052003e-01 4.32807744e-01
-7.73170672e-04 4.58302557e-01 -7.65740693e-01 5.76834455e-02
5.50268173e-01 1.46239758e+00 -1.37869298e+00 -4.51406956e-01
-3.12029958e-01 9.75161076e-01 6.15982115e-01 1.48041993e-01
-7.41051733e-01 -7.33067989e-01 6.02596223e-01 7.39006996e-01
3.43377799e-01 -6.37949407e-01 -2.70723879e-01 -1.00162935e+00
3.35610099e-02 -8.67518604e-01 6.67613804e-01 -5.49549997e-01
-9.64955151e-01 6.37777269e-01 8.04335251e-02 -1.03359520e+00
-1.81012824e-01 -3.69978070e-01 -9.26758707e-01 7.73840010e-01
-1.04048884e+00 -1.08469367e+00 -8.02405059e-01 8.27572286e-01
2.90272355e-01 5.02023706e-03 1.04785812e+00 2.24172905e-01
-3.80959451e-01 3.93527091e-01 -2.19734654e-01 -2.70473480e-01
4.77418244e-01 -1.19556642e+00 8.48633885e-01 8.60470712e-01
5.05270623e-02 7.98462808e-01 4.53459710e-01 -7.64374733e-01
-2.13530779e+00 -1.11345863e+00 2.52080232e-01 7.23106936e-02
8.23957384e-01 -5.85699379e-01 -9.81596470e-01 6.28960788e-01
1.35124866e-02 5.57746351e-01 6.61699355e-01 3.58066857e-01
-3.33390325e-01 -3.01134497e-01 -1.02996981e+00 7.94529736e-01
1.23421955e+00 -2.90825427e-01 2.23746195e-01 7.60961831e-01
5.74421346e-01 -8.40452433e-01 -9.43954825e-01 2.28902280e-01
3.77667546e-01 -1.14617562e+00 9.48963225e-01 -7.22966120e-02
3.44976336e-02 -5.30887663e-01 2.81437114e-02 -1.24267423e+00
-3.19611698e-01 -9.69600916e-01 -3.72367859e-01 1.00762224e+00
1.24969684e-01 -9.02708352e-01 1.07183921e+00 1.97498724e-01
-1.29821196e-01 -1.04247248e+00 -7.38898098e-01 -7.61196196e-01
-3.55113059e-01 -6.89804375e-01 8.70914221e-01 1.03304029e+00
-4.34967652e-02 2.88796455e-01 -1.16163166e-02 1.38393492e-01
6.61601007e-01 6.70574367e-01 1.53754759e+00 -1.01903987e+00
-6.63762629e-01 -6.32630110e-01 -4.25229549e-01 -9.15157080e-01
-3.76238585e-01 -1.09339213e+00 -5.14354169e-01 -2.03163671e+00
2.20414083e-02 -4.60905194e-01 3.27445954e-01 4.96742725e-01
-1.46996215e-01 2.94186890e-01 2.86919206e-01 2.97748540e-02
-5.30461490e-01 8.40070192e-03 1.25975513e+00 1.62310228e-01
-3.20965081e-01 -1.00219343e-02 -4.10582781e-01 5.35277307e-01
8.08113456e-01 -4.45161998e-01 -7.15595305e-01 -4.42187190e-01
4.22798604e-01 9.18719918e-02 1.95825621e-01 -1.08252537e+00
3.48246783e-01 -2.01829955e-01 4.31247428e-02 -8.61480176e-01
1.68569639e-01 -5.87597668e-01 5.33166468e-01 7.08786488e-01
5.01384318e-01 6.45743906e-01 6.29529655e-01 4.22833651e-01
4.08128686e-02 2.85331547e-01 4.89519238e-01 -4.34547514e-01
-5.44597387e-01 6.68004394e-01 -1.23844855e-01 7.53253922e-02
1.43923032e+00 -4.52857882e-01 -5.35725534e-01 -2.06867978e-01
-6.89870775e-01 3.67846817e-01 8.54199946e-01 1.15700103e-01
3.97864908e-01 -1.31155169e+00 -3.96898836e-01 4.45169836e-01
3.68707930e-03 2.63766795e-01 3.40865612e-01 7.88200200e-01
-1.14413774e+00 -7.33366981e-02 -9.44882110e-02 -6.99428797e-01
-1.71567500e+00 8.47379088e-01 -1.61322013e-01 -8.06625307e-01
-7.31834173e-01 7.34347761e-01 6.58649430e-02 -4.86133754e-01
-4.11458135e-01 -3.12023729e-01 5.82643569e-01 8.26825872e-02
5.83838701e-01 7.13475883e-01 3.89288753e-01 -5.88530265e-02
-6.15603805e-01 7.76895285e-01 1.00040138e-01 2.16686919e-01
1.47376156e+00 2.55834281e-01 -5.30806780e-01 2.46967807e-01
1.10921693e+00 4.71528202e-01 -8.34487498e-01 1.82422623e-01
-2.03396335e-01 -5.96916258e-01 2.27061100e-02 -4.19981927e-01
-1.18623304e+00 9.84349430e-01 2.06581220e-01 5.55326521e-01
1.35164833e+00 -2.87830383e-01 7.58931637e-01 5.93547821e-02
9.16523755e-01 -8.74354005e-01 -6.21869899e-02 4.58954602e-01
6.73437238e-01 -6.86160862e-01 6.11208558e-01 -9.57760274e-01
-2.66522735e-01 1.41438651e+00 6.30314112e-01 -1.74482360e-01
6.85747743e-01 8.46164942e-01 -3.68600726e-01 -4.55618560e-01
-6.52506948e-01 -1.86102495e-01 2.16307834e-01 6.04535639e-01
5.54501534e-01 3.52254868e-01 -4.21397060e-01 -6.15083352e-02
-6.00997031e-01 -9.58304107e-02 5.18621564e-01 1.03764641e+00
-3.25177401e-01 -1.57725275e+00 -2.93477625e-01 6.13560319e-01
1.61146671e-01 -1.97456762e-01 2.80617941e-02 8.41797769e-01
-3.57458651e-01 4.67394888e-01 2.34847441e-01 -2.60318458e-01
4.73729432e-01 -2.58493066e-01 4.86410350e-01 -8.12458992e-01
-8.54791522e-01 -2.28963848e-02 2.88854867e-01 -9.38018858e-01
2.27697998e-01 -2.79373139e-01 -1.25014377e+00 -8.22812617e-01
-3.29638628e-04 6.53467104e-02 6.92617297e-01 3.05779755e-01
9.18171704e-01 6.53241694e-01 2.25232393e-01 -1.14990222e+00
-1.49136130e-02 -4.47338372e-01 -5.10984480e-01 -4.40121302e-03
-2.64129639e-01 -4.02160466e-01 -1.18133798e-01 -4.43871558e-01] | [7.042637825012207, 5.507956027984619] |
3b848623-18de-4949-a873-62090e7ebed7 | rs-gv-at-semeval-2021-task-1-sense-relative | null | null | https://aclanthology.org/2021.semeval-1.82 | https://aclanthology.org/2021.semeval-1.82.pdf | RS\_GV at SemEval-2021 Task 1: Sense Relative Lexical Complexity Prediction | We present the technical report of the system called RS{\_}GV at SemEval-2021 Task 1 on lexical complexity prediction of English words. RS{\_}GV is a neural network using hand-crafted linguistic features in combination with character and word embeddings to predict target words{'} complexity. For the generation of the hand-crafted features, we set the target words in relation to their senses. RS{\_}GV predicts the complexity well of biomedical terms but it has problems with the complexity prediction of very complex and very simple target words. | ['Gayatri Venugopal', 'Regina Stodden'] | 2021-08-01 | null | null | null | semeval-2021 | ['lexical-complexity-prediction'] | ['natural-language-processing'] | [ 9.93417948e-03 2.63999641e-01 -2.91205883e-01 -5.24505734e-01
-3.20220619e-01 -1.76593721e-01 5.75978160e-01 6.83836997e-01
-9.32183266e-01 5.47461808e-01 5.76627612e-01 -8.54168713e-01
-2.32906967e-01 -8.16355705e-01 -2.67159611e-01 -3.66530381e-02
-2.65131623e-01 5.71403682e-01 -1.33967176e-01 -6.74073994e-01
3.20676893e-01 3.91167462e-01 -1.48553026e+00 4.10576791e-01
3.36218387e-01 1.03466749e+00 4.50795323e-01 7.67685175e-01
-4.47271228e-01 3.66855711e-01 -2.55192131e-01 -2.14599550e-01
1.06614642e-01 -1.62716135e-01 -1.03594041e+00 -5.91990471e-01
9.00238529e-02 2.61866450e-01 -3.41039598e-01 9.52300787e-01
5.21802366e-01 2.69897431e-01 1.09309542e+00 -8.45441163e-01
-8.02607656e-01 9.82375681e-01 5.58804162e-02 7.77437091e-01
5.70183516e-01 7.70478472e-02 1.47144008e+00 -1.10749435e+00
8.64690661e-01 1.44848657e+00 8.70071769e-01 6.98624909e-01
-9.80209708e-01 -5.08040011e-01 8.14613551e-02 4.52689111e-01
-1.55938053e+00 -4.78600174e-01 4.65925038e-01 -5.76587915e-01
2.09650373e+00 2.22819358e-01 4.79289234e-01 1.07193291e+00
7.20153570e-01 3.41906220e-01 8.78139615e-01 -7.58921862e-01
1.34370700e-01 1.09872960e-01 6.16731465e-01 8.42291474e-01
1.22939520e-01 3.10888886e-01 -3.12008530e-01 -2.57145703e-01
2.45519906e-01 -4.51026291e-01 1.21017627e-01 4.58685994e-01
-1.00876868e+00 1.20876503e+00 2.33672723e-01 6.39335811e-01
-2.28076786e-01 4.17959653e-02 8.72916877e-01 3.22101355e-01
6.24072254e-01 8.22578907e-01 -1.01035666e+00 8.29459876e-02
-7.36311853e-01 3.20432365e-01 7.45437503e-01 1.00284481e+00
3.75333905e-01 7.16306865e-02 -2.56258279e-01 9.27119911e-01
4.09621805e-01 1.56289339e-01 9.30452108e-01 -1.68150127e-01
3.34946752e-01 4.65059489e-01 -1.65290743e-01 -8.19345891e-01
-1.13868654e+00 -4.68615517e-02 -6.11886442e-01 -1.12477630e-01
2.15477198e-02 -4.34948280e-02 -1.00633979e+00 1.62267411e+00
-2.17014432e-01 -3.72883677e-01 9.23982486e-02 2.61022180e-01
1.49023592e+00 7.07198441e-01 7.43315220e-01 -3.93480033e-01
1.46641123e+00 -7.52110541e-01 -8.26286674e-01 -2.78455764e-01
9.33481216e-01 -6.23203337e-01 1.11584663e+00 4.47302073e-01
-1.03118670e+00 -7.56832838e-01 -1.03247476e+00 -2.43109182e-01
-1.07792830e+00 -3.21688175e-01 5.57532668e-01 7.67368257e-01
-1.28463292e+00 1.02574360e+00 -4.41417187e-01 -1.90642655e-01
1.05617374e-01 6.40738189e-01 -4.95693684e-01 2.81396359e-01
-1.75768924e+00 1.60798633e+00 8.06491792e-01 -3.10651839e-01
-4.97710854e-01 -8.26257765e-01 -1.29529428e+00 8.56844932e-02
-1.06726952e-01 -4.77869719e-01 8.17322671e-01 -3.95325571e-01
-1.25575995e+00 1.05435526e+00 4.70497459e-02 -3.40207398e-01
-8.72988403e-02 7.62371868e-02 -1.12448573e+00 -3.12576413e-01
9.44863558e-02 6.53502464e-01 4.02435988e-01 -4.76143211e-01
-4.07432735e-01 -1.23244569e-01 -2.76322335e-01 1.00059606e-01
-6.65652752e-01 1.48624301e-01 4.40055458e-03 -7.88476825e-01
-4.50147748e-01 -6.54678166e-01 -4.22748953e-01 -5.08216441e-01
-2.81396657e-01 -9.25652862e-01 -1.27997369e-01 -6.43150151e-01
1.69054711e+00 -1.91653228e+00 5.96782044e-02 2.48045713e-01
4.36792046e-01 4.70154196e-01 -5.45929790e-01 5.73006988e-01
-6.00708723e-01 4.72906202e-01 1.60927519e-01 3.28744389e-02
1.60047382e-01 4.75668907e-02 1.64898574e-01 3.03773373e-01
2.40017548e-01 9.26566422e-01 -1.07167673e+00 -5.17148495e-01
3.73179734e-01 3.18714291e-01 -6.31387711e-01 8.50039795e-02
-3.58760893e-01 -3.43809873e-01 -5.00322640e-01 2.94655532e-01
2.63337761e-01 4.13888954e-02 4.93794866e-02 -2.40787938e-01
-7.21047148e-02 7.30635226e-01 -8.14993620e-01 1.28771114e+00
-7.14874506e-01 4.52753574e-01 -7.98678696e-01 -9.45424974e-01
9.05516207e-01 3.25287461e-01 3.75521779e-01 -7.55770266e-01
5.02585948e-01 2.47073263e-01 1.88928246e-01 -5.33939600e-01
4.44617063e-01 -6.18668497e-01 -3.46134394e-01 2.09842682e-01
4.15901244e-01 -1.87028021e-01 3.36693257e-01 -2.13724524e-01
1.28650558e+00 -1.95163563e-01 9.83961642e-01 -7.98965931e-01
7.17248738e-01 -3.39188613e-02 3.16393018e-01 4.97981995e-01
-1.55743912e-01 3.19186717e-01 9.61573198e-02 -9.33844805e-01
-1.21579623e+00 -1.01787436e+00 -4.90232110e-01 1.41194797e+00
-3.81516010e-01 -1.01299214e+00 -4.65320289e-01 -3.74444932e-01
-4.77575278e-03 1.12443984e+00 -8.74560297e-01 -4.00125653e-01
-3.86368990e-01 -5.54206967e-01 8.53459537e-01 6.72474742e-01
-3.52782577e-01 -1.58839762e+00 -3.76800656e-01 6.45337343e-01
6.70660511e-02 -9.18488741e-01 -5.23767054e-01 7.24889874e-01
-4.26220477e-01 -7.33986855e-01 -1.73712805e-01 -1.03270602e+00
1.61645964e-01 -5.89551508e-01 1.51327765e+00 7.39291236e-02
-7.78548896e-01 -2.16822758e-01 -5.53161383e-01 -9.39177632e-01
-6.79775655e-01 1.27135515e-01 5.27421474e-01 -7.98597455e-01
7.78851151e-01 -3.56678575e-01 -3.34629148e-01 -9.73701179e-02
-9.01585460e-01 -3.77170205e-01 3.88474643e-01 9.02017236e-01
5.50041974e-01 -1.18549950e-02 5.86861253e-01 -1.07931709e+00
9.95275855e-01 -4.12132889e-01 -2.12788776e-01 3.50621566e-02
-8.77402842e-01 3.74568343e-01 8.70565891e-01 -6.56808734e-01
-1.58968136e-01 -9.49952379e-02 -1.01022899e+00 -9.60101783e-02
-1.65738508e-01 7.45767951e-01 -6.54309914e-02 3.23281616e-01
9.77970064e-01 2.21015587e-01 -3.55140865e-01 -4.89769042e-01
6.96027994e-01 7.04978526e-01 1.81497648e-01 -2.37525791e-01
3.37209523e-01 -3.40734720e-01 -1.05203815e-01 -1.08759832e+00
-6.85793519e-01 -3.74466598e-01 -5.02907276e-01 2.96581030e-01
9.73819196e-01 -6.80950463e-01 -7.22887516e-01 1.96975898e-02
-1.10922766e+00 -1.21730566e-01 -2.80070156e-01 3.68211240e-01
-6.87245905e-01 2.68470883e-01 -8.01783800e-01 -5.54981589e-01
-7.59933650e-01 -9.15443242e-01 8.58639061e-01 -1.70105979e-01
-1.04511392e+00 -1.24373496e+00 1.11493304e-01 -1.73768535e-01
3.37931097e-01 3.25186849e-02 1.55520666e+00 -1.24283648e+00
5.62066674e-01 -2.63719797e-01 1.20282464e-01 4.90447015e-01
1.84397802e-01 -2.17541113e-01 -7.28933036e-01 -7.19284490e-02
-1.76215813e-01 -1.88339770e-01 7.85943627e-01 4.35570657e-01
1.27589834e+00 -2.72879928e-01 -3.39983612e-01 4.36703831e-01
1.37948036e+00 1.97621435e-01 4.19340461e-01 2.61827439e-01
3.64132464e-01 3.66672575e-01 2.89182514e-01 5.20322561e-01
4.38316345e-01 3.59167755e-01 2.72834972e-02 2.21947089e-01
1.57605223e-02 -1.62279308e-01 2.89325833e-01 1.09061515e+00
9.91261378e-02 -4.92439061e-01 -1.09034359e+00 5.04350960e-01
-9.89904046e-01 -7.04208016e-01 -1.10685211e-02 1.60959721e+00
1.15279949e+00 5.35109997e-01 4.03877385e-02 2.59643644e-01
4.29412842e-01 2.17813984e-01 -1.48276806e-01 -1.28548384e+00
9.80303138e-02 9.35987592e-01 4.20442462e-01 7.11301684e-01
-8.87983561e-01 1.13225389e+00 7.37952471e+00 1.14540410e+00
-8.89667451e-01 8.85423794e-02 5.12062371e-01 1.74596742e-01
-4.71299052e-01 -5.08513689e-01 -1.19193459e+00 2.74025738e-01
1.65281332e+00 -4.60684121e-01 2.05919683e-01 7.71146655e-01
9.45934728e-02 2.97890693e-01 -1.49035347e+00 8.24373126e-01
1.00232795e-01 -1.50346386e+00 2.26367891e-01 -1.13387056e-01
2.08298892e-01 2.42687225e-01 -7.15493560e-02 5.43559909e-01
2.72182435e-01 -1.58762252e+00 5.72905421e-01 4.70583081e-01
1.21244550e+00 -7.72964120e-01 8.31437051e-01 3.36408257e-01
-1.11752450e+00 8.18897858e-02 -8.37550163e-01 -2.22200975e-01
2.05420796e-02 7.21485317e-01 -9.64410126e-01 2.75303304e-01
3.51975977e-01 4.64833647e-01 -4.51345414e-01 4.51243907e-01
6.65410906e-02 5.19759834e-01 -2.68317491e-01 -5.98809123e-01
2.22172722e-01 2.39292905e-01 3.06688607e-01 1.75751853e+00
1.04823768e-01 2.38320321e-01 1.43227100e-01 6.06047809e-01
-8.58974643e-03 7.53399253e-01 -7.99680054e-01 -5.34518898e-01
3.05558294e-01 1.01501405e+00 -7.27879524e-01 -5.46290934e-01
-2.05272257e-01 6.80488229e-01 3.76667410e-01 -2.74496794e-01
-4.87657160e-01 -7.87108302e-01 6.64910257e-01 1.19631931e-01
1.66459441e-01 -1.37864426e-01 -4.43588674e-01 -8.26262116e-01
-4.36970502e-01 -6.73031151e-01 5.60320854e-01 -4.12863374e-01
-1.80639088e+00 1.19059289e+00 -5.57544157e-02 -8.94113123e-01
-2.73282886e-01 -1.36292684e+00 -3.36200148e-01 1.18764317e+00
-1.30669785e+00 -7.38955438e-01 4.85177636e-01 6.15692496e-01
8.41208220e-01 -4.39908653e-01 1.47121155e+00 4.86252047e-02
-2.62774657e-02 8.36440563e-01 2.86611449e-03 4.59378632e-03
4.35366690e-01 -1.30954075e+00 7.63447642e-01 -1.97309881e-01
5.42016048e-03 6.72667682e-01 9.26827133e-01 -7.60917842e-01
-1.12897158e+00 -1.17481124e+00 1.81117165e+00 -7.32464969e-01
8.67009163e-01 -4.35246587e-01 -5.15154898e-01 4.51192111e-01
6.35625049e-02 7.67470896e-02 1.22065699e+00 2.84167469e-01
-3.48487973e-01 2.64521450e-01 -1.07731605e+00 5.53715050e-01
9.87476468e-01 -6.79187894e-01 -1.34715736e+00 7.19346583e-01
1.05562222e+00 -1.63121372e-01 -1.16504133e+00 5.08975208e-01
5.07317245e-01 -3.86103302e-01 1.14795005e+00 -1.17145634e+00
6.69869065e-01 2.56115973e-01 -2.49075472e-01 -1.46876836e+00
-6.63264632e-01 -1.19072422e-01 1.91784009e-01 5.58282614e-01
8.26433659e-01 -2.59966582e-01 4.39416796e-01 2.42879570e-01
-2.77426958e-01 -8.33114743e-01 -1.04352319e+00 -9.16159272e-01
7.28288531e-01 -7.94045568e-01 3.33039582e-01 9.10485983e-01
5.48164666e-01 6.94153905e-01 -1.63948074e-01 -1.71478242e-01
-2.78131035e-03 -3.95890415e-01 -1.94988668e-01 -1.29997134e+00
-2.37051278e-01 -6.12642348e-01 -7.71763980e-01 -3.31866533e-01
4.51745957e-01 -1.26348019e+00 3.57161537e-02 -1.42168760e+00
-1.20467775e-01 -1.41611844e-01 -6.52866721e-01 5.27665198e-01
-2.54962236e-01 1.59558505e-01 3.07479296e-02 -5.00294864e-01
-3.47706884e-01 2.96486139e-01 8.21544945e-01 -1.60378426e-01
-1.45168975e-01 -5.45089655e-02 -7.79638112e-01 7.08108306e-01
6.09073222e-01 -4.11343277e-01 -2.10440993e-01 -1.34914219e-01
5.57458162e-01 -4.78037223e-02 -2.16077268e-01 -7.12771475e-01
-2.56067693e-01 -2.15455994e-01 5.21737218e-01 -5.81080496e-01
1.67205155e-01 -4.29098725e-01 -5.85117601e-02 6.65445268e-01
-7.72681892e-01 5.67519665e-01 4.41031724e-01 2.19875589e-01
1.15104422e-01 -5.35358131e-01 7.60696232e-01 -4.05282348e-01
-6.81608200e-01 1.93043292e-01 -7.79811144e-01 2.16962203e-01
6.38453007e-01 -5.24322242e-02 1.40635699e-01 6.49458542e-02
-1.25947285e+00 1.65993616e-01 -1.26509786e-01 8.23760211e-01
7.58660674e-01 -1.27663171e+00 -6.92129493e-01 2.24732101e-01
4.10062045e-01 -6.44107044e-01 -4.73314226e-02 1.70113787e-01
-5.59881330e-01 7.56697357e-01 -3.52437168e-01 1.55403987e-01
-1.07472670e+00 1.14204848e+00 4.63946611e-02 -5.51292598e-01
-5.27672589e-01 1.09577191e+00 -1.11879699e-01 -5.97568035e-01
1.00127630e-01 -5.15095234e-01 -7.28761077e-01 2.24752948e-01
8.30608785e-01 1.55933410e-01 3.55168521e-01 -6.26325607e-01
-7.53295541e-01 2.70537138e-01 -1.96330681e-01 1.53757259e-01
1.64837217e+00 8.10185075e-02 -1.64609447e-01 6.82189107e-01
1.66354167e+00 -3.57469350e-01 1.36520818e-01 -3.39785755e-01
4.10542101e-01 1.56976491e-01 1.68352798e-01 -9.50126767e-01
-5.47701359e-01 8.47269595e-01 8.75294864e-01 -1.49763390e-01
7.53603339e-01 6.28079847e-02 7.89893508e-01 7.60022998e-01
2.34662592e-01 -1.45098817e+00 -1.00695632e-01 9.77869153e-01
7.10858285e-01 -8.73790324e-01 5.71874678e-02 -4.01654065e-01
-4.24785107e-01 1.40379620e+00 3.92803043e-01 -2.39925966e-01
1.15542662e+00 2.34787405e-01 -1.28717944e-01 -3.41410995e-01
-1.01368105e+00 -2.32616693e-01 6.13705337e-01 5.80804169e-01
7.67233074e-01 3.75971854e-01 -9.56620753e-01 1.13755918e+00
-6.03223860e-01 -7.55598247e-02 3.45452070e-01 6.72335863e-01
-5.62793970e-01 -1.20079935e+00 -3.79599668e-02 9.81054962e-01
-4.98418957e-01 -8.09523404e-01 -1.80507883e-01 6.96798682e-01
4.85399097e-01 8.20297122e-01 -1.29069522e-01 -7.81283557e-01
5.99233210e-01 4.19612408e-01 2.92651594e-01 -1.15743971e+00
-8.19501221e-01 -1.18119217e-01 4.61939782e-01 -4.54747647e-01
1.57808065e-02 -3.75973284e-01 -1.38580704e+00 -3.00675016e-02
-2.47965649e-01 9.48502943e-02 6.11970127e-01 1.27199304e+00
2.51214728e-02 5.49058914e-01 4.78339076e-01 -6.29106402e-01
-7.10167527e-01 -1.23953032e+00 -7.08736897e-01 5.03916740e-01
4.58224751e-02 -4.37303901e-01 -2.12451905e-01 -1.16845034e-01] | [10.612018585205078, 10.40668773651123] |
3824c70f-29ac-4b06-9ff0-ae745b49f039 | cacti-a-framework-for-scalable-multi-task | 2212.05711 | null | https://arxiv.org/abs/2212.05711v2 | https://arxiv.org/pdf/2212.05711v2.pdf | CACTI: A Framework for Scalable Multi-Task Multi-Scene Visual Imitation Learning | Large-scale training have propelled significant progress in various sub-fields of AI such as computer vision and natural language processing. However, building robot learning systems at a comparable scale remains challenging. To develop robots that can perform a wide range of skills and adapt to new scenarios, efficient methods for collecting vast and diverse amounts of data on physical robot systems are required, as well as the capability to train high-capacity policies using such datasets. In this work, we propose a framework for scaling robot learning, with specific focus on multi-task and multi-scene manipulation in kitchen environments, both in simulation and in the real world. Our proposed framework, CACTI, comprises four stages that separately handle: data collection, data augmentation, visual representation learning, and imitation policy training, to enable scalability in robot learning . We make use of state-of-the-art generative models as part of the data augmentation stage, and use pre-trained out-of-domain visual representations to improve training efficiency. Experimental results demonstrate the effectiveness of our approach. On a real robot setup, CACTI enables efficient training of a single policy that can perform 10 manipulation tasks involving kitchen objects, and is robust to varying layouts of distractors. In a simulated kitchen environment, CACTI trains a single policy to perform 18 semantic tasks across 100 layout variations for each individual task. We will release the simulation task benchmark and augmented datasets in both real and simulated environments to facilitate future research. | ['Vikash Kumar', 'Aravind Rajeswaran', 'Shuran Song', 'Vincent Moens', 'Homanga Bharadhwaj', 'Zhao Mandi'] | 2022-12-12 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 1.59832239e-01 -2.05069721e-01 -1.95515957e-02 -3.09188068e-01
-3.54988039e-01 -6.09940052e-01 5.05439341e-01 -1.21867642e-01
-5.67272007e-01 5.37809014e-01 -3.78244668e-01 -2.74782836e-01
-5.22852801e-02 -6.34426296e-01 -1.15836251e+00 -5.83905041e-01
-2.38709807e-01 9.24086392e-01 3.64184141e-01 -2.60998696e-01
9.94379744e-02 6.71484768e-01 -1.72299528e+00 2.44544446e-02
6.45598948e-01 6.77920341e-01 1.07809091e+00 7.57048190e-01
1.42202172e-02 7.53729522e-01 -7.91228652e-01 3.61434311e-01
3.69556278e-01 -6.24855533e-02 -7.13092506e-01 2.83056349e-01
1.07689641e-01 -3.18401128e-01 -2.99188703e-01 7.06306398e-01
5.07486045e-01 4.74954963e-01 7.00059414e-01 -1.56244743e+00
-7.74030089e-01 5.92238307e-01 -4.67417955e-01 -1.99571848e-01
4.42520455e-02 5.21060765e-01 3.91037613e-01 -6.35215223e-01
5.55451274e-01 1.56898379e+00 2.44019687e-01 6.93815410e-01
-1.07123566e+00 -8.38161230e-01 3.04300487e-01 4.96680364e-02
-9.42388594e-01 -2.83762366e-01 5.89661062e-01 -4.17781562e-01
1.27136850e+00 -5.11336684e-01 5.22790074e-01 1.43615842e+00
3.69817048e-01 9.26685452e-01 1.03441906e+00 -2.61708081e-01
4.26447004e-01 -1.55572686e-02 -3.40462565e-01 7.86228597e-01
2.47951657e-01 2.93083161e-01 -1.66336581e-01 2.21003145e-01
1.24754071e+00 8.63436908e-02 1.33619830e-01 -9.82295394e-01
-1.45089376e+00 7.52141714e-01 7.64556706e-01 2.11993232e-01
-3.35014135e-01 7.13132203e-01 6.07259810e-01 3.04019839e-01
-2.28250459e-01 7.80342996e-01 -5.67810655e-01 -1.62879512e-01
-6.80612102e-02 4.14389700e-01 7.97011495e-01 1.63174570e+00
7.51810908e-01 3.18639100e-01 8.57045203e-02 1.11239314e+00
1.84294522e-01 7.97680020e-01 6.11780822e-01 -1.21915400e+00
5.84281802e-01 3.88244539e-01 8.76286253e-02 -6.71672225e-01
-6.19606555e-01 -1.19350538e-01 -6.33151591e-01 4.95697737e-01
1.10269159e-01 -2.08028540e-01 -1.43163371e+00 1.75183475e+00
3.74799073e-01 -7.17672557e-02 3.31159651e-01 9.39453125e-01
4.90428001e-01 7.65350342e-01 2.08026737e-01 3.15340757e-01
1.12149870e+00 -1.35496938e+00 -4.93291557e-01 -7.27341652e-01
7.48187900e-01 -4.43146467e-01 1.36073232e+00 3.10677648e-01
-7.69281328e-01 -9.12702203e-01 -1.05039835e+00 4.36873809e-02
-4.65528220e-01 2.64597446e-01 7.41472840e-01 1.03914380e-01
-9.68866706e-01 3.17542374e-01 -1.19842303e+00 -7.76100874e-01
4.15473193e-01 4.79469150e-01 -4.64041173e-01 -4.56403077e-01
-6.52751744e-01 1.20904422e+00 7.56607533e-01 -1.26075611e-01
-1.71607339e+00 -3.30789864e-01 -1.17601764e+00 -3.09688020e-02
7.30971098e-01 -6.33201599e-01 1.57103312e+00 -7.78401971e-01
-1.78684044e+00 4.67198461e-01 4.18745577e-01 -2.75565416e-01
3.97766493e-02 -4.53187466e-01 3.63393012e-03 -1.01973854e-01
1.36793435e-01 1.01596713e+00 8.98875654e-01 -1.84575391e+00
-5.82613766e-01 -1.98995054e-01 1.13199338e-01 3.55120957e-01
-1.08834289e-01 -1.51112705e-01 -5.12398481e-01 -3.72040063e-01
-2.72549272e-01 -1.28135681e+00 -4.95297402e-01 -9.70252752e-02
-4.46074381e-02 -2.19473422e-01 9.91056263e-01 -1.67492583e-01
3.46215099e-01 -2.10687709e+00 3.76181126e-01 -9.02927220e-02
-2.62289554e-01 2.19509229e-01 -4.91619319e-01 3.72062832e-01
2.48896062e-01 -3.35103899e-01 -1.33445174e-01 -4.48955745e-01
1.27155855e-01 8.36234510e-01 -1.75623789e-01 1.74164057e-01
3.21946681e-01 8.34198475e-01 -1.08137476e+00 -3.11608702e-01
4.90719557e-01 6.40794486e-02 -4.71562475e-01 4.55543309e-01
-4.47010964e-01 6.31010413e-01 -5.01791954e-01 6.37374580e-01
3.08407336e-01 -2.02676177e-01 1.43217728e-01 2.71735340e-01
8.99831653e-02 -1.49613142e-01 -1.02233529e+00 2.16784739e+00
-9.05797184e-01 3.08172375e-01 3.11908007e-01 -1.13969302e+00
1.19383788e+00 -3.89670357e-02 3.50975186e-01 -7.99133539e-01
3.14693511e-01 1.64679065e-01 2.14707404e-01 -6.38532698e-01
6.43278480e-01 1.68471470e-01 -4.70337480e-01 2.78133392e-01
4.68470454e-01 -6.77385986e-01 3.05587739e-01 -5.42396344e-02
1.18776059e+00 5.34129083e-01 2.35591114e-01 2.68418137e-02
-1.02110906e-02 4.39235896e-01 2.69827545e-01 8.34749997e-01
-2.40858078e-01 1.82881013e-01 3.95416236e-03 -3.36421221e-01
-1.36508524e+00 -9.87904906e-01 2.19864771e-01 1.55256391e+00
4.10579383e-01 6.63174838e-02 -4.66280341e-01 -5.26312172e-01
2.63607025e-01 8.51096570e-01 -4.87547427e-01 -3.67361635e-01
-8.58406901e-01 -4.23961490e-01 4.03853595e-01 9.04598117e-01
5.76422632e-01 -1.81151807e+00 -1.21885455e+00 2.59598613e-01
3.30238402e-01 -1.36567831e+00 1.74099728e-02 6.11775577e-01
-7.21058905e-01 -9.91263509e-01 -5.35478592e-01 -1.30814803e+00
7.85862029e-01 5.23540914e-01 8.14733267e-01 -2.18846247e-01
-4.19816554e-01 6.71097934e-01 -4.66818213e-01 -5.64471245e-01
-8.19654644e-01 1.89009413e-01 2.26330325e-01 -7.70814240e-01
-1.19721055e-01 -5.25094151e-01 -9.77218524e-02 4.58164692e-01
-8.02572966e-01 1.02941588e-01 1.19091105e+00 1.08075905e+00
4.76867676e-01 -8.72038454e-02 5.83324850e-01 -6.29377306e-01
6.56921923e-01 -6.00677311e-01 -8.45352054e-01 2.39976317e-01
-3.69595826e-01 1.61324635e-01 7.92144120e-01 -8.13712537e-01
-1.06926692e+00 3.01428497e-01 1.78686112e-01 -7.88112462e-01
-5.89929879e-01 4.92956221e-01 -7.05878884e-02 6.98581338e-02
7.75762677e-01 1.73294574e-01 8.48663971e-02 -3.49173576e-01
6.13365173e-01 5.34023285e-01 8.20005476e-01 -1.06503487e+00
7.17827678e-01 5.59785441e-02 -3.45856622e-02 -7.32417464e-01
-3.51690501e-01 -3.79517943e-01 -6.92360520e-01 -1.98385566e-02
6.37936115e-01 -9.03870881e-01 -7.75940537e-01 6.54159367e-01
-1.02266526e+00 -1.23714101e+00 -1.54009238e-01 4.32978868e-01
-9.99823809e-01 -1.44947857e-01 -4.77119654e-01 -6.85624957e-01
-1.39392465e-01 -1.43270671e+00 1.21175277e+00 3.09720516e-01
1.76050931e-01 -7.64882088e-01 2.07920209e-01 7.25664943e-02
4.16878462e-01 7.57532865e-02 9.07824636e-01 -5.33289135e-01
-6.35668039e-01 -4.14987057e-02 -1.96630374e-01 3.10957611e-01
2.31108844e-01 -2.55134523e-01 -7.73293138e-01 -6.74393952e-01
-3.11144114e-01 -1.10570037e+00 5.49641550e-01 2.15972587e-01
1.26429546e+00 1.13698550e-01 -6.44174218e-01 4.84749973e-01
1.23348260e+00 4.97371942e-01 4.06340510e-01 4.93121177e-01
9.10309255e-01 4.56613481e-01 9.88334835e-01 3.11915874e-01
4.09587711e-01 6.80306911e-01 7.23416090e-01 4.10513766e-02
5.88349216e-02 -3.41798007e-01 4.72409934e-01 5.33099294e-01
1.59682512e-01 -3.40016454e-01 -9.99652147e-01 7.34564424e-01
-1.98952770e+00 -6.32131875e-01 4.92957026e-01 1.88843966e+00
5.60987294e-01 1.54683068e-01 -7.50707090e-02 -2.82330751e-01
4.35756892e-01 -2.27456629e-01 -9.51310873e-01 -4.44196135e-01
4.36550915e-01 1.67077214e-01 5.99415898e-01 1.17236756e-01
-1.14077747e+00 1.41636205e+00 6.01509809e+00 6.80092037e-01
-1.13588369e+00 5.82804680e-02 1.41356319e-01 1.42384574e-01
3.91395152e-01 -2.04199731e-01 -7.04966545e-01 1.23254314e-01
8.58030379e-01 2.27217346e-01 6.78033769e-01 1.54909706e+00
-1.18244857e-01 -4.13214803e-01 -1.24476397e+00 9.07422841e-01
1.15742810e-01 -9.80595589e-01 -2.26726010e-01 -1.11038826e-01
7.13095069e-01 3.79154116e-01 7.73639679e-02 9.36053514e-01
9.90802586e-01 -1.10082150e+00 7.20116615e-01 -1.45987477e-02
6.57941878e-01 -6.18063092e-01 6.27659440e-01 6.14111423e-01
-1.14690638e+00 -4.24964428e-01 -5.18380761e-01 5.09335697e-02
1.20186228e-02 -1.84730649e-01 -1.44671285e+00 4.24666107e-01
9.11600471e-01 7.12248802e-01 -3.69227886e-01 9.63127434e-01
-3.44888180e-01 1.55353308e-01 -3.72291952e-01 -2.06606939e-01
2.81599551e-01 1.72912106e-01 1.96508944e-01 9.82163310e-01
2.97931015e-01 -2.19093367e-01 7.29037642e-01 7.42914796e-01
1.31049519e-02 -3.16789746e-01 -9.56891954e-01 -9.62702185e-02
6.97505653e-01 1.30112934e+00 -7.71867871e-01 -2.37131715e-01
-2.77168781e-01 1.05195916e+00 7.71197855e-01 4.12639320e-01
-8.59637499e-01 -1.74812183e-01 5.83895087e-01 -4.59007829e-01
4.14668232e-01 -8.58192325e-01 -1.73493668e-01 -6.55116439e-01
-1.64042294e-01 -9.62848783e-01 2.74701379e-02 -8.78572643e-01
-1.13162923e+00 4.73840058e-01 3.60740185e-01 -9.06801581e-01
-4.80595559e-01 -1.05836594e+00 -4.35819536e-01 5.49163699e-01
-1.40545452e+00 -1.36236823e+00 -6.58058345e-01 5.72818458e-01
9.71045613e-01 -3.55516344e-01 9.33809221e-01 7.12531898e-03
-5.93149126e-01 3.45243931e-01 9.69156697e-02 1.00288037e-02
7.35422850e-01 -1.15617871e+00 5.76954603e-01 4.18112427e-01
-1.94863062e-02 4.30692315e-01 6.16970658e-01 -6.50365651e-01
-1.71147275e+00 -1.29693711e+00 -2.87388504e-01 -4.68848467e-01
5.86563170e-01 -6.53125465e-01 -8.73357832e-01 9.57691908e-01
1.02809802e-01 7.40326429e-03 1.77026480e-01 9.69602540e-02
-7.55873621e-02 8.73225257e-02 -1.02324712e+00 5.65979600e-01
1.23029840e+00 -3.73761915e-02 -5.58142662e-01 3.47601235e-01
8.45049083e-01 -8.19153786e-01 -7.35689282e-01 4.17893499e-01
3.50523412e-01 -3.81884426e-01 8.99701297e-01 -4.97744799e-01
4.10940021e-01 -2.44438842e-01 -1.34633735e-01 -1.75929153e+00
-3.80450368e-01 -2.90905654e-01 -7.76928067e-02 8.92432332e-01
2.09672987e-01 -4.81012195e-01 5.11692226e-01 2.67988801e-01
-7.49771357e-01 -6.64674938e-01 -5.42805970e-01 -9.85150099e-01
6.20754808e-02 -3.74949008e-01 4.76367533e-01 5.78863740e-01
-2.94581463e-04 5.27578771e-01 -3.63638431e-01 2.38193080e-01
2.95517385e-01 6.93827718e-02 1.42162347e+00 -9.20383751e-01
-4.18010056e-01 -2.53635556e-01 -2.97418892e-01 -1.10051966e+00
4.18765813e-01 -7.12948680e-01 8.09367239e-01 -1.75066090e+00
2.95706373e-02 -9.34286356e-01 -8.30212515e-03 8.23308527e-01
5.65196760e-02 -3.20610464e-01 3.59967679e-01 2.20201030e-01
-6.75274432e-01 9.47625756e-01 1.55740345e+00 -1.58452839e-01
-3.20529044e-01 -1.78099275e-01 -3.21981549e-01 6.21105194e-01
9.23397243e-01 -3.77191395e-01 -8.28258157e-01 -8.08758616e-01
-3.35039347e-01 -1.51956096e-01 4.01824474e-01 -1.45379555e+00
1.42009109e-01 -5.52989602e-01 4.95119154e-01 -3.21746767e-01
5.87564290e-01 -1.06932271e+00 -1.58898458e-01 5.66067994e-01
-2.66040146e-01 2.35590443e-01 6.13697588e-01 6.60706282e-01
1.89334422e-01 -1.68755978e-01 7.45920718e-01 -2.93783814e-01
-1.27967131e+00 1.92387968e-01 -3.32456797e-01 -9.72630531e-02
1.45562780e+00 1.15760066e-01 -5.54940641e-01 -5.26418984e-02
-4.50018227e-01 7.37487912e-01 6.60807967e-01 9.96138930e-01
7.90482700e-01 -1.12842822e+00 -3.79925132e-01 2.46414751e-01
2.93669820e-01 5.75342774e-01 1.21488057e-01 3.10580343e-01
-5.28098941e-01 2.50280738e-01 -6.91585839e-01 -8.17392945e-01
-8.71258318e-01 1.02675521e+00 1.01805933e-01 -2.54317760e-01
-6.78960562e-01 6.83578908e-01 4.22116816e-01 -8.45109344e-01
3.94959211e-01 -7.49932170e-01 -4.64399382e-02 -6.25278890e-01
1.81419015e-01 9.44454968e-02 -2.81536400e-01 -3.89437377e-01
-1.65860757e-01 2.99955875e-01 -3.76211777e-02 -4.07573879e-02
1.23877525e+00 1.41147479e-01 3.14629942e-01 5.60326040e-01
7.50214875e-01 -7.93078780e-01 -1.59634066e+00 -1.46793514e-01
-1.47559509e-01 -2.80383915e-01 -2.42827982e-01 -9.08714473e-01
-8.05996478e-01 8.31673205e-01 6.05819404e-01 -1.69608146e-01
7.73026288e-01 1.87119722e-01 6.48396134e-01 1.00566900e+00
9.96388793e-01 -1.21522498e+00 7.65893340e-01 7.88103282e-01
1.12757123e+00 -1.39299297e+00 -2.56466389e-01 -1.54349372e-01
-1.04544616e+00 8.25112104e-01 1.33504307e+00 -3.63924503e-01
2.15376779e-01 4.51537877e-01 5.06848209e-02 -1.51081398e-01
-7.82484710e-01 -2.28504777e-01 -4.25135717e-02 1.08483577e+00
7.95381330e-03 8.28978270e-02 5.08474946e-01 2.99675256e-01
-5.43264449e-02 2.21672691e-02 3.80214423e-01 1.57402837e+00
-7.76163995e-01 -1.00388885e+00 -4.04358566e-01 2.61121690e-01
3.53076428e-01 4.67631876e-01 -8.11735839e-02 1.07933021e+00
5.36941923e-02 7.84470022e-01 9.76012498e-02 -3.77522796e-01
5.77179492e-01 -7.51260445e-02 8.11275303e-01 -9.67874467e-01
-3.30779850e-01 -1.62468225e-01 -2.98201200e-02 -6.17794096e-01
-3.19712937e-01 -5.34395039e-01 -1.70954120e+00 3.89817804e-02
-9.78423506e-02 -2.02006727e-01 9.41041768e-01 8.87996137e-01
4.18699682e-01 9.80903447e-01 3.09159756e-01 -1.51144087e+00
-7.43529856e-01 -1.25751531e+00 -3.71347427e-01 3.33677083e-01
1.45631894e-01 -1.29308045e+00 1.82917461e-01 3.10498066e-02] | [4.541669845581055, 0.8522148132324219] |
a25a5e9f-d86d-4c80-84df-e26e72e82094 | prospects-and-applications-of-incoherent | 2304.09922 | null | https://arxiv.org/abs/2304.09922v1 | https://arxiv.org/pdf/2304.09922v1.pdf | Prospects and Applications of Incoherent Light in Non-contact Wireless Sensing Systems | The increasing demand for wireless sensing systems has led to the exploration of alternative technologies to overcome the spectrum scarcity of traditional approaches based on radio frequency (RF) waves or microwaves. Incoherent light sources such as light-emitting diodes (LED), paired with light sensors, have the potential to become an attractive option for wireless sensing due to their energy efficiency, longer lifespan, and lower cost. Although coherent light or laser may present safety risks to human eyes and skin, incoherent visible and infrared light has low intensity, and does not harm the human body. Incoherent light has the potential to supersede other wireless sensing technologies, namely RF, laser and camera, by providing many additional benefits including easy implementation, wide bandwidth, reusable frequency, minimum interference, enhanced privacy and simpler data processing. However, the application of incoherent light in the wireless sensing domain is still in its infancy and is an emerging research topic. This study explores the enormous potential and benefits of incoherent visible and infrared light in wireless sensing through various indoor and outdoor applications including speed estimation of vehicles, human vitals monitoring, blood glucose sensing, gesture recognition, occupancy estimation and structural health monitoring. | ["John F. O'Hara", 'Sabit Ekin', 'Md Zobaer Islam'] | 2023-04-19 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 6.61979437e-01 -4.11972314e-01 -2.23451704e-01 -1.11421920e-01
-2.82823369e-02 -4.38982755e-01 1.92614079e-01 -1.89521983e-01
-6.46003187e-01 1.20683694e+00 -1.00669697e-01 -7.40263388e-02
-1.10830128e-01 -8.66002679e-01 6.58991784e-02 -1.20869946e+00
2.01763794e-01 -5.71525753e-01 -1.09043038e-02 1.50301069e-01
4.32427553e-03 2.74358004e-01 -1.41609764e+00 -5.98159969e-01
8.61497581e-01 1.19815111e+00 1.25423625e-01 4.65827316e-01
3.35159115e-02 1.60821736e-01 -4.04833168e-01 3.13165158e-01
2.42460489e-01 -3.72928709e-01 4.29865122e-02 -6.41340435e-01
-1.40153900e-01 -2.89250374e-01 -1.17624000e-01 7.30014026e-01
8.10819805e-01 -5.15267625e-03 2.15560049e-01 -9.26930964e-01
-6.56344891e-01 6.06273226e-02 -1.07112622e+00 2.24267989e-01
7.82919943e-01 2.75168628e-01 1.22168675e-01 -5.15615165e-01
1.10171124e-01 6.63488209e-01 5.56649685e-01 8.13475668e-01
-8.38188469e-01 -1.15490782e+00 -4.30171072e-01 2.28176266e-01
-1.37334156e+00 -6.47406936e-01 6.09950006e-01 1.39168873e-01
8.32747877e-01 5.59595704e-01 7.15391874e-01 1.05917609e+00
5.41856945e-01 -1.28105476e-01 1.55691385e+00 -6.49907410e-01
4.46641713e-01 1.36490643e-01 -5.23226685e-04 6.53545856e-01
6.72865033e-01 4.44103152e-01 -7.41758287e-01 -7.90393353e-02
3.00853670e-01 3.41790199e-01 -5.16184807e-01 1.53684914e-01
-1.04605687e+00 2.75004953e-01 3.44396502e-01 6.06956363e-01
-4.86923993e-01 4.91700917e-01 -2.64269084e-01 2.12955326e-02
5.03571033e-01 -1.04320254e-02 -2.17958950e-02 -1.42243475e-01
-2.83935815e-01 -5.85040808e-01 5.14745414e-01 5.60282111e-01
4.68760729e-01 -9.40979198e-02 3.70971896e-02 6.91455781e-01
8.46580386e-01 1.27350199e+00 1.51614174e-01 -6.59692705e-01
1.89755455e-01 2.78793693e-01 3.44832718e-01 -6.50584221e-01
-8.20002735e-01 -3.00294518e-01 -9.41966951e-01 2.57165670e-01
9.57083479e-02 -6.16001010e-01 -7.49901354e-01 1.60964930e+00
3.63861769e-01 3.44010115e-01 2.51940995e-01 8.49884272e-01
9.77329612e-01 5.66856205e-01 4.45893019e-01 -6.06797874e-01
1.50644195e+00 -2.01486975e-01 -7.92127073e-01 -2.59671122e-01
-4.66995649e-02 -8.63243818e-01 4.01391596e-01 1.41688406e-01
-7.65535712e-01 -3.94817293e-01 -1.02026474e+00 2.60123640e-01
-3.66183698e-01 -2.20270187e-01 7.66401947e-01 1.46977735e+00
-7.66944945e-01 -2.84036160e-01 -9.08522367e-01 -6.95622921e-01
5.02834320e-01 2.77466744e-01 1.00456938e-01 -4.22669560e-01
-1.02307510e+00 7.49028802e-01 -1.80680171e-01 1.77845865e-01
4.93943170e-02 -7.47087836e-01 -5.40004671e-01 -6.82938099e-02
1.56216875e-01 -9.17976499e-01 5.07786214e-01 -1.51261061e-01
-2.03863764e+00 3.94509375e-01 -3.01167369e-01 -2.16522351e-01
2.83040218e-02 -1.98147699e-01 -7.47828841e-01 2.75270194e-01
-1.47755727e-01 3.03079933e-01 5.30755162e-01 -7.35672474e-01
-4.60215420e-01 -6.08259559e-01 -3.08850139e-01 5.90857305e-03
-2.35463291e-01 -4.53396179e-02 4.19687599e-01 -1.22839831e-01
1.94814265e-01 -9.85358953e-01 -1.72767445e-01 2.11177379e-01
-2.46829376e-01 -4.05971080e-01 9.76270080e-01 8.15305933e-02
5.84336460e-01 -2.19497228e+00 -7.66801715e-01 3.39190573e-01
1.41667858e-01 2.54033685e-01 1.19480558e-01 2.53513366e-01
4.64311540e-01 -5.12232296e-02 -4.77431789e-02 2.92247266e-01
-5.75398147e-01 -9.54146087e-02 3.01152974e-01 8.70768905e-01
-4.40835506e-01 8.83010805e-01 -1.07804811e+00 -8.58998597e-02
5.33633888e-01 8.38877022e-01 1.46397769e-01 -4.46237266e-01
4.41786945e-01 9.17700648e-01 -6.45512998e-01 9.71590579e-01
8.71208608e-01 -1.03418522e-01 -5.88552374e-03 -2.78045803e-01
-5.91588140e-01 -2.35003263e-01 -9.58926737e-01 1.54845870e+00
-6.40974104e-01 5.99623144e-01 3.13199013e-02 -7.14324176e-01
1.08739400e+00 5.24495482e-01 7.62122154e-01 -1.46871281e+00
1.02652255e-02 2.15295255e-01 -4.80890691e-01 -1.10602987e+00
-1.03323288e-01 -6.62307739e-01 2.02623889e-01 5.59362829e-01
-6.34817183e-01 3.76082361e-01 -1.67339563e-01 -4.93464619e-01
1.17611551e+00 2.48740420e-01 3.68828356e-01 -1.29096687e-01
3.46022964e-01 -4.74824518e-01 3.12667966e-01 4.93722677e-01
-1.10259794e-01 -7.49297738e-02 -9.25158203e-01 -2.12286010e-01
-1.98041331e-02 -1.19220710e+00 -2.96029538e-01 7.87752807e-01
9.10093486e-01 3.77956152e-01 1.37839183e-01 -3.36201396e-03
3.90496891e-04 3.68209839e-01 -3.53720263e-02 -1.49230257e-01
-3.67208540e-01 -1.02995646e+00 4.46871459e-01 2.58074969e-01
1.17397141e+00 -6.55691624e-01 -1.38070071e+00 3.54424626e-01
-2.00784996e-01 -1.23493600e+00 1.06405593e-01 5.09671234e-02
-7.67498136e-01 -9.75087166e-01 -8.61426592e-01 -3.73371512e-01
4.53863502e-01 9.11470592e-01 6.87440038e-01 -1.13057546e-01
-8.79783034e-01 9.18254793e-01 -2.52100706e-01 -8.93684983e-01
2.02027127e-01 -4.64394957e-01 2.29809374e-01 2.11779475e-02
8.63063574e-01 -7.04367340e-01 -1.44240510e+00 2.00074121e-01
-3.26057881e-01 -4.46530312e-01 6.87881529e-01 1.30165100e-01
3.04380566e-01 6.45995066e-02 7.76621938e-01 -5.59854150e-01
4.17096972e-01 -6.43490791e-01 -4.78671819e-01 1.01055928e-01
-7.46997476e-01 -2.97756404e-01 3.44534546e-01 -1.80790290e-01
-1.42402720e+00 -1.80869736e-02 2.08347030e-02 6.21615469e-01
-4.82263714e-01 2.68430889e-01 1.25380844e-01 -4.82882053e-01
8.51271629e-01 2.23208629e-02 -8.90863035e-03 -3.21064055e-01
1.73398405e-01 8.49582016e-01 4.03557301e-01 1.42001376e-01
8.96102667e-01 9.99541044e-01 4.95015711e-01 -1.50059605e+00
-4.25519884e-01 -7.17751443e-01 1.18136637e-01 -4.24763024e-01
1.02946675e+00 -1.04618204e+00 -1.21310842e+00 1.61406294e-01
-7.57105052e-01 1.96322367e-01 -1.31124705e-01 1.08861589e+00
2.29536835e-02 4.85503703e-01 -1.15238473e-01 -1.25473142e+00
-5.78020334e-01 -3.96522105e-01 5.08367181e-01 8.46788585e-01
-3.69413853e-01 -6.91499531e-01 6.61537796e-02 4.51402992e-01
8.85469437e-01 6.67445064e-01 6.07151270e-01 7.21995473e-01
-6.60588980e-01 -3.40436280e-01 -2.83660620e-01 -2.19496563e-01
8.30855250e-01 -3.97360027e-01 -1.28619242e+00 -9.04751495e-02
-9.40683410e-02 5.12832217e-02 8.23317528e-01 7.80060768e-01
6.22266769e-01 3.95476036e-02 -8.55883658e-01 6.56546772e-01
1.79319572e+00 6.02420270e-01 6.32087886e-01 -8.41082819e-03
2.52552807e-01 2.28377014e-01 3.78529102e-01 4.69364464e-01
2.59019211e-02 2.06148699e-01 4.73156601e-01 -4.03615981e-01
-6.31476864e-02 3.72307628e-01 5.73920049e-02 8.35767463e-02
-7.93752789e-01 -3.70691568e-01 -3.27060491e-01 -4.06026877e-02
-1.14902484e+00 -1.34124517e+00 -5.10092914e-01 2.37589526e+00
3.21457684e-01 -2.67696738e-01 8.81445780e-02 1.63044691e-01
6.06504500e-01 -1.06108293e-01 -5.76123536e-01 -1.41790971e-01
9.19252932e-02 5.85581422e-01 7.81980455e-01 4.13681865e-02
-6.63154900e-01 5.83390072e-02 6.68790483e+00 8.03126469e-02
-1.13264608e+00 4.05657798e-01 6.70464113e-02 -7.93346912e-02
-3.51261497e-01 -1.96969941e-01 -4.71598148e-01 6.98323786e-01
6.90164864e-01 2.26044908e-01 3.24051023e-01 2.20402077e-01
4.67044264e-01 -9.67289865e-01 -7.26712465e-01 1.14875937e+00
-2.97311991e-01 -9.47875440e-01 -6.54843986e-01 3.29109043e-01
5.09521723e-01 1.31011486e-01 2.29005247e-01 -3.02535564e-01
-2.99594849e-01 -7.65479088e-01 -2.32931059e-02 5.84095418e-01
1.17609453e+00 -5.45462847e-01 5.57881355e-01 1.99649885e-01
-1.14340639e+00 -1.94290385e-01 -4.06339914e-01 -4.38186705e-01
1.07789725e-01 9.51447785e-01 -1.63150102e-01 3.14635158e-01
9.21668053e-01 4.02941495e-01 9.41561237e-02 1.09633505e+00
-3.00684050e-02 5.15471935e-01 -7.15660572e-01 -6.24500871e-01
-1.14398725e-01 -3.12380552e-01 4.83523250e-01 8.81950676e-01
6.30113423e-01 5.54497600e-01 -1.60273850e-01 8.10814738e-01
2.02554926e-01 -1.87037975e-01 -8.67454410e-01 4.15530592e-01
5.75312316e-01 1.08527851e+00 -9.25223827e-01 9.80941579e-02
-8.74286592e-01 4.51811701e-01 -9.11083877e-01 6.62939250e-01
-6.68231249e-01 -5.16744912e-01 4.53091204e-01 -1.33503927e-02
-2.19992831e-01 -5.13501823e-01 -3.79719853e-01 -5.73571980e-01
1.98455527e-02 6.28116131e-02 3.19825143e-01 -5.36128044e-01
-9.11879897e-01 7.22763389e-02 -1.16344728e-01 -1.43657100e+00
4.92524564e-01 -2.87251741e-01 -6.18209779e-01 8.92142653e-01
-2.07621694e+00 -5.56484759e-01 -1.10125327e+00 6.24222696e-01
4.31411296e-01 1.52536988e-01 9.50495243e-01 2.07803637e-01
-3.20304960e-01 2.19579220e-01 3.22575510e-01 -5.30195773e-01
6.45545721e-01 -6.18418753e-01 -4.68917876e-01 6.11648738e-01
-2.36037880e-01 5.86983800e-01 5.63896239e-01 -3.09895784e-01
-1.79079795e+00 -5.96704006e-01 3.43532115e-01 8.16070884e-02
5.14565185e-02 6.35068640e-02 -1.28796101e-01 -8.25942755e-02
2.92951614e-01 3.10072601e-01 1.25035393e+00 -1.56441331e-01
1.80899296e-02 -6.11138880e-01 -1.55107880e+00 2.32094079e-01
9.81243610e-01 -5.27523346e-02 -5.64721925e-03 3.40853214e-01
7.11125974e-03 1.05301470e-01 -6.26160562e-01 3.68827790e-01
1.25897801e+00 -9.54421163e-01 1.17898047e+00 4.35095608e-01
-2.92146802e-01 -1.32445037e-01 9.52917710e-02 -9.51991141e-01
-4.85567749e-01 -7.80135632e-01 4.27797325e-02 1.12642658e+00
2.11778268e-01 -1.33371854e+00 6.98360801e-01 6.18446887e-01
3.07827413e-01 -2.86975801e-01 -1.27736628e+00 -8.41109455e-01
-8.29333246e-01 -2.37642631e-01 5.34968078e-01 5.13306141e-01
3.44522029e-01 4.22371209e-01 -3.92103016e-01 1.49962291e-01
1.03143096e+00 6.63225576e-02 2.22666353e-01 -1.57275784e+00
6.72805533e-02 3.64051647e-02 -2.17435420e-01 -9.21532691e-01
-3.68191451e-01 -4.83774334e-01 1.12292424e-01 -1.89214087e+00
-5.75513244e-02 -7.66496181e-01 -5.77931345e-01 3.48771393e-01
-2.37128860e-03 1.00757253e+00 -2.56818771e-01 7.51873245e-03
-3.88718843e-01 1.30114689e-01 1.17966259e+00 -2.99855143e-01
-5.37497938e-01 4.95618343e-01 -6.22028649e-01 5.85230589e-01
9.63990033e-01 -4.59337324e-01 -3.91843468e-01 -3.24111551e-01
3.94634813e-01 -6.57439828e-02 3.76672566e-01 -1.35034549e+00
3.14598709e-01 -3.74363244e-01 6.57917082e-01 1.24543831e-02
4.48395640e-01 -1.30455291e+00 3.57665092e-01 1.07543564e+00
3.62956554e-01 -5.11470914e-01 -1.23095550e-01 9.41955149e-01
4.46333259e-01 -4.18200269e-02 8.52720916e-01 -3.71857494e-01
-8.50092947e-01 -5.43452986e-02 -9.10315692e-01 -5.25782704e-01
1.35972917e+00 -9.45640624e-01 -8.04236650e-01 -1.50379255e-01
-2.75770247e-01 -2.95998193e-02 -1.48937637e-02 2.42820770e-01
6.66194558e-01 -9.08262670e-01 -3.09544653e-01 2.09677637e-01
9.51007977e-02 -5.30052245e-01 3.89103889e-01 1.08466268e+00
-2.55288303e-01 4.53938097e-01 -2.83456534e-01 -7.63606071e-01
-1.44975495e+00 2.23767415e-01 -2.62411386e-02 5.40014803e-01
-8.85043561e-01 6.98665261e-01 -2.71955222e-01 5.84604561e-01
4.85137515e-02 -1.62362292e-01 -2.26888403e-01 -3.46573859e-01
5.61205447e-01 9.56245124e-01 -2.68166959e-01 -1.25146791e-01
-9.14992213e-01 1.22816718e+00 8.18193018e-01 3.30550462e-01
1.26115358e+00 -6.26802921e-01 3.49616148e-02 2.22129479e-01
7.76042640e-01 3.01416725e-01 -7.92490423e-01 1.23468274e-02
-3.15451294e-01 -4.15121257e-01 2.05428645e-01 -8.89798820e-01
-7.07845926e-01 5.03689766e-01 1.41640866e+00 6.02144480e-01
1.39450908e+00 -2.93805101e-03 8.61828864e-01 4.31203961e-01
8.64522278e-01 -9.49604213e-01 -8.72469097e-02 -2.94069648e-01
1.49762049e-01 -1.24712920e+00 3.00239533e-01 -6.73455238e-01
3.23791146e-01 1.19037962e+00 2.93329418e-01 3.04160327e-01
7.84113765e-01 4.53488857e-01 3.32536042e-01 -1.25547692e-01
-1.62474290e-01 -4.25472736e-01 -1.48460716e-01 1.17001188e+00
6.28647327e-01 7.72837400e-02 -7.35640228e-01 -1.68504760e-01
3.28859687e-01 4.90906179e-01 3.17163616e-01 1.21364021e+00
-6.98034048e-01 -8.41836035e-01 -5.11606514e-01 6.21295989e-01
-7.77230382e-01 4.74601448e-01 2.48492166e-01 5.18103182e-01
4.26516235e-01 1.72406292e+00 -3.09743196e-01 -7.23294765e-02
2.08814725e-01 -8.15676246e-03 6.93737626e-01 -3.57670397e-01
1.30727306e-01 5.64616285e-02 -9.67274234e-02 -5.44537902e-01
-1.34296453e+00 -1.47415832e-01 -1.18297040e+00 -2.14987602e-02
-6.18503571e-01 7.19561949e-02 1.17855418e+00 1.20877182e+00
4.67152387e-01 3.36908609e-01 7.08016872e-01 -4.16176319e-01
-4.02503610e-02 -7.73622990e-01 -7.49942124e-01 -2.14891300e-01
5.17447531e-01 -7.28732765e-01 -1.62132561e-01 -1.22050755e-01] | [6.611507415771484, 0.6265912055969238] |
5f9cdb86-28e9-4a46-9713-d0c8e848e850 | adversarially-regularized-policy-learning | 2109.07627 | null | https://arxiv.org/abs/2109.07627v3 | https://arxiv.org/pdf/2109.07627v3.pdf | Adversarially Regularized Policy Learning Guided by Trajectory Optimization | Recent advancement in combining trajectory optimization with function approximation (especially neural networks) shows promise in learning complex control policies for diverse tasks in robot systems. Despite their great flexibility, the large neural networks for parameterizing control policies impose significant challenges. The learned neural control policies are often overcomplex and non-smooth, which can easily cause unexpected or diverging robot motions. Therefore, they often yield poor generalization performance in practice. To address this issue, we propose adVErsarially Regularized pOlicy learNIng guided by trajeCtory optimizAtion (VERONICA) for learning smooth control policies. Specifically, our proposed approach controls the smoothness (local Lipschitz continuity) of the neural control policies by stabilizing the output control with respect to the worst-case perturbation to the input state. Our experiments on robot manipulation show that our proposed approach not only improves the sample efficiency of neural policy learning but also enhances the robustness of the policy against various types of disturbances, including sensor noise, environmental uncertainty, and model mismatch. | ['Ye Zhao', 'Tuo Zhao', 'Simiao Zuo', 'Zhigen Zhao'] | 2021-09-16 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 6.55458495e-02 2.03625575e-01 -5.53982437e-01 9.93151292e-02
-5.00631094e-01 -5.77608705e-01 3.69164705e-01 -1.63695395e-01
-4.42744434e-01 1.13346529e+00 -1.02816895e-01 -4.84809905e-01
-2.67888099e-01 -5.29721260e-01 -1.25777078e+00 -1.09188926e+00
-1.44929975e-01 -4.98362854e-02 -7.75921866e-02 -1.95155278e-01
1.33118734e-01 6.23394310e-01 -1.01593280e+00 -5.37222147e-01
9.02291775e-01 9.17060137e-01 1.55710861e-01 3.39095592e-01
3.67512584e-01 8.10037136e-01 -5.70595503e-01 3.19316149e-01
3.62906188e-01 -2.91292872e-02 -2.79055029e-01 -1.29358485e-01
2.24141881e-01 -3.64837438e-01 -6.23845696e-01 1.49279976e+00
4.82748449e-01 7.85804808e-01 5.74501336e-01 -1.31815755e+00
-6.96691632e-01 6.11665130e-01 -2.56118625e-01 -2.86455870e-01
-2.55056709e-01 6.29950643e-01 4.09559458e-01 -4.24279690e-01
3.34764451e-01 1.54836881e+00 7.40601361e-01 8.79208565e-01
-1.04903519e+00 -7.85380244e-01 5.94293892e-01 -1.53392658e-01
-1.09130991e+00 -2.74593949e-01 5.05546093e-01 -3.25457305e-01
5.75198948e-01 -1.08700112e-01 3.14498633e-01 1.38833177e+00
5.29793203e-01 7.32941628e-01 7.19904304e-01 8.47598240e-02
5.15173674e-01 -1.18940324e-01 -4.62865382e-01 6.76989377e-01
3.14109474e-01 7.56179094e-01 3.33748132e-01 -2.84290045e-01
1.08801162e+00 8.25666040e-02 -5.70676267e-01 -4.28288698e-01
-1.32586801e+00 6.28681600e-01 6.79487824e-01 -2.58331299e-01
-5.40620029e-01 6.73653543e-01 5.53832471e-01 5.54533362e-01
-1.26082987e-01 8.78904462e-01 -6.19573057e-01 -1.52903229e-01
9.60369408e-02 6.87773883e-01 6.05370760e-01 1.18555355e+00
2.84175903e-01 9.87843096e-01 -1.29985243e-01 6.92263603e-01
1.41197994e-01 7.65080333e-01 5.38954914e-01 -1.35847259e+00
7.13293314e-01 1.55316144e-01 6.36596262e-01 -1.19879508e+00
-4.22183007e-01 -2.87906200e-01 -1.11811364e+00 7.51748621e-01
5.38048029e-01 -7.89024413e-01 -9.08524275e-01 1.98863804e+00
2.45137051e-01 1.95832670e-01 3.25856358e-01 8.95711601e-01
-9.77468193e-02 8.34239185e-01 8.96684080e-02 -2.98396677e-01
4.26296592e-01 -7.49719679e-01 -8.36308718e-01 -2.85623133e-01
2.70342141e-01 -1.88208625e-01 1.31276512e+00 3.10800642e-01
-9.25826550e-01 -4.54874605e-01 -1.09555256e+00 6.23261988e-01
-1.49073437e-01 1.36022836e-01 8.42697993e-02 6.79519102e-02
-5.28548241e-01 8.93373013e-01 -1.07426369e+00 2.72160713e-02
2.84875780e-01 5.24701715e-01 -6.52600601e-02 1.44933194e-01
-1.08956337e+00 9.63112950e-01 7.21599936e-01 2.11888343e-01
-1.19884455e+00 -7.44315624e-01 -7.91561782e-01 -4.61366661e-02
1.04388940e+00 -2.74465442e-01 1.38055480e+00 -8.00831258e-01
-2.20432186e+00 -2.74762452e-01 2.88948506e-01 -5.36056042e-01
5.09756088e-01 -3.61114353e-01 -1.84076995e-01 -1.65179208e-01
-3.05933416e-01 4.51411009e-01 1.25894856e+00 -1.11701167e+00
-5.41145027e-01 -9.48230103e-02 -2.82099098e-02 2.04693109e-01
-2.27399886e-01 -5.20232260e-01 3.84365842e-02 -9.16915715e-01
-2.42016882e-01 -1.15451229e+00 -7.54417419e-01 2.55867839e-01
-3.14657509e-01 -1.06470048e-01 9.17763770e-01 -2.20611408e-01
9.45118546e-01 -2.14990115e+00 2.94977218e-01 2.72803247e-01
-2.95521021e-01 4.30555046e-01 -2.28803217e-01 2.99628705e-01
2.12162569e-01 6.76531270e-02 -2.18048960e-01 2.13466465e-01
1.42862067e-01 4.28541332e-01 -9.07656789e-01 5.95894039e-01
2.51604050e-01 7.64751613e-01 -1.13199258e+00 1.49769947e-01
1.92221344e-01 2.40804851e-01 -3.91690522e-01 3.57003897e-01
-6.69145584e-01 7.69506454e-01 -9.03023303e-01 5.49085140e-01
1.34982511e-01 2.20643923e-01 -1.32538006e-01 2.15290576e-01
-8.71407241e-02 -2.10999578e-01 -1.13789725e+00 1.14500177e+00
-4.28049147e-01 3.91903669e-01 5.93214750e-01 -1.26514018e+00
1.01798427e+00 4.29844856e-01 3.88211131e-01 -2.50925660e-01
4.74261343e-01 1.79262653e-01 8.21254309e-03 -5.13134062e-01
1.62851214e-01 1.35338949e-02 -1.08038321e-01 -3.67064662e-02
-2.00718522e-01 -3.84411693e-01 -3.37260693e-01 -5.02222657e-01
8.21433008e-01 2.56225348e-01 3.78155380e-01 -1.84248686e-01
4.84648168e-01 9.16822106e-02 9.07049417e-01 7.31075346e-01
-4.07363266e-01 4.98469137e-02 4.26641494e-01 -1.60792187e-01
-1.15128016e+00 -7.43706286e-01 3.36679161e-01 8.64400983e-01
1.88584924e-01 3.33727837e-01 -5.43925643e-01 -4.25772130e-01
4.18669462e-01 7.94911742e-01 -4.36145127e-01 -7.29203224e-01
-1.00952351e+00 -3.37969750e-01 6.51789606e-01 7.93979526e-01
2.08553538e-01 -1.14657950e+00 -5.33852696e-01 5.41346312e-01
4.26140308e-01 -1.02592230e+00 -6.74150765e-01 1.15386687e-01
-8.18444967e-01 -1.14007974e+00 -7.11192787e-01 -7.48025775e-01
8.40711057e-01 -1.97067097e-01 2.52297282e-01 -3.33347678e-01
1.56519905e-01 3.84538323e-01 1.86790630e-01 -5.59207618e-01
-6.36098802e-01 -8.42581317e-02 7.20053732e-01 -1.49509355e-01
-4.06990290e-01 -3.66310120e-01 -1.43029109e-01 5.17066538e-01
-7.55562723e-01 -5.74092805e-01 4.08864141e-01 1.13202775e+00
7.50839889e-01 2.74667501e-01 9.85847414e-01 -4.57092762e-01
1.20157683e+00 -5.82774580e-01 -1.37255955e+00 9.81723741e-02
-7.05686092e-01 2.69449949e-01 1.49248934e+00 -1.22754419e+00
-8.91798675e-01 1.27695292e-01 1.96920171e-01 -9.71878350e-01
-1.61150604e-01 3.79512906e-01 -5.10383062e-02 -2.30732709e-01
7.57556379e-01 1.27380535e-01 5.12742281e-01 -1.30979016e-01
5.07325292e-01 3.00639838e-01 9.33891714e-01 -9.29251313e-01
1.01663077e+00 9.93034095e-02 3.30598384e-01 -6.03993654e-01
-3.77013922e-01 6.76292181e-02 -1.72699839e-01 -2.03470796e-01
4.84339356e-01 -7.17597425e-01 -1.03379583e+00 5.06676018e-01
-8.95050108e-01 -6.99345708e-01 -3.63933593e-01 5.99227905e-01
-9.93797541e-01 5.02480492e-02 -3.85738522e-01 -1.06469345e+00
-2.20509663e-01 -1.24265313e+00 4.41992939e-01 4.55472350e-01
5.73799685e-02 -9.77323771e-01 -1.35333017e-01 -6.24871790e-01
6.07532918e-01 7.92387187e-01 9.69694495e-01 -2.48026222e-01
-3.53620231e-01 -2.17845038e-01 2.97497064e-01 5.55457592e-01
1.85316578e-01 -5.05355410e-02 -3.68791282e-01 -6.98394537e-01
1.09965540e-01 -6.95531070e-01 3.28256130e-01 5.26838660e-01
1.15258002e+00 -7.66026139e-01 -3.55396718e-01 6.42471075e-01
1.22425532e+00 6.48320913e-01 8.61487836e-02 4.53267366e-01
8.05285871e-01 4.54783738e-01 9.04107034e-01 3.22135210e-01
-2.53176004e-01 2.68015325e-01 7.32540011e-01 3.43380004e-01
4.28718925e-01 -4.14925873e-01 8.11431706e-01 3.11337531e-01
1.28418386e-01 -5.64996004e-02 -6.21484101e-01 5.37615180e-01
-2.15215063e+00 -7.78996170e-01 5.10489643e-01 2.19318962e+00
8.02443981e-01 1.18886523e-01 3.54214273e-02 -2.17724949e-01
7.69130766e-01 4.16008681e-02 -1.38835871e+00 -3.93963993e-01
1.17886238e-01 -3.74214828e-01 1.00595438e+00 4.73024637e-01
-9.86819625e-01 9.13776815e-01 5.98334694e+00 7.46781111e-01
-1.55102885e+00 -5.31412125e-01 2.52974093e-01 -5.82041778e-02
1.17810749e-01 -5.65353870e-01 -6.81543171e-01 4.65506077e-01
8.30013573e-01 -3.07260990e-01 8.49437773e-01 1.35845292e+00
5.35184562e-01 4.72801685e-01 -9.12044585e-01 6.54748976e-01
-5.25092781e-01 -1.23202920e+00 -2.66185552e-01 -1.87348783e-01
9.51081097e-01 9.68023092e-02 4.38789904e-01 6.26160562e-01
7.01218843e-01 -1.04262757e+00 7.67414510e-01 4.89430547e-01
6.68816805e-01 -1.03798509e+00 5.57081580e-01 7.18188405e-01
-7.91943789e-01 -7.76907206e-01 -6.16236806e-01 9.29283053e-02
2.25968003e-01 -1.51314721e-01 -8.31952333e-01 7.54583776e-02
2.63368368e-01 7.06268609e-01 1.53600961e-01 9.43065047e-01
-4.26152170e-01 6.38605595e-01 -3.93190056e-01 -5.52740097e-01
7.02786446e-01 -2.24749282e-01 1.04130816e+00 5.84216893e-01
2.60126263e-01 -6.35586530e-02 6.00701451e-01 8.61637712e-01
7.42149577e-02 -2.63594061e-01 -1.10337973e+00 -2.60557264e-01
6.59157217e-01 5.90419173e-01 -5.80491610e-02 -6.26798570e-02
1.65736079e-02 2.96137929e-01 5.60373127e-01 8.40110540e-01
-7.94302404e-01 -5.63045204e-01 9.92371261e-01 -4.81415629e-01
3.06388974e-01 -5.89079857e-01 -2.11171880e-02 -7.67404318e-01
1.09199040e-01 -1.00716710e+00 -6.35775328e-02 -2.65264839e-01
-1.24408865e+00 2.79482424e-01 7.41071999e-02 -1.38822973e+00
-7.63066411e-01 -7.41993845e-01 -6.33476019e-01 5.89411139e-01
-1.27246118e+00 -4.54306036e-01 2.38530114e-01 5.90228796e-01
5.53102016e-01 -5.16638815e-01 6.14199221e-01 -1.68325171e-01
-8.06461453e-01 5.95449865e-01 7.56172717e-01 -5.96221536e-02
6.74864829e-01 -1.23407269e+00 1.64640963e-01 7.19121456e-01
-6.27199769e-01 6.07344329e-01 8.06813836e-01 -6.37029707e-01
-1.67919290e+00 -1.71135783e+00 -2.92156309e-01 -2.96705663e-02
1.02012873e+00 5.48027679e-02 -1.07852519e+00 6.78486764e-01
-2.16949835e-01 1.56511217e-01 -2.75306135e-01 -5.42667449e-01
-2.32371818e-02 -1.52187020e-01 -1.24388325e+00 1.11543334e+00
5.86458206e-01 -1.60947934e-01 -3.89945894e-01 1.02824576e-01
1.15111566e+00 -6.72024548e-01 -9.46170270e-01 6.42735183e-01
4.17111665e-01 5.43098105e-03 9.25321758e-01 -1.15420151e+00
3.18107866e-02 -3.56519610e-01 -4.74015251e-02 -1.85726333e+00
-1.86496735e-01 -1.18723297e+00 -4.96331096e-01 7.09621787e-01
2.97532886e-01 -8.88914585e-01 7.17532575e-01 5.49126208e-01
-3.31236184e-01 -1.10368061e+00 -8.55172455e-01 -1.26871824e+00
5.18791735e-01 -1.50191531e-01 4.71990138e-01 7.28541195e-01
-5.32683358e-02 -1.51691232e-02 -6.35508299e-01 5.85829437e-01
4.80012804e-01 -3.21206272e-01 8.19357753e-01 -5.63815892e-01
-1.86663270e-01 -6.50051713e-01 -1.41521078e-02 -1.02372038e+00
6.04038119e-01 -3.43812346e-01 7.15964735e-01 -9.62434769e-01
-9.14385676e-01 -4.62636143e-01 -3.43737036e-01 4.01552588e-01
-2.76810199e-01 -6.71419740e-01 1.25398383e-01 8.12332481e-02
-1.13060683e-01 9.19385672e-01 1.47223115e+00 -3.65864307e-01
-4.60718423e-01 6.00639403e-01 -3.16956937e-01 8.60019863e-01
1.13111126e+00 -3.73987526e-01 -8.32847297e-01 -5.18693328e-01
-1.56777993e-01 2.98767984e-01 1.06234543e-01 -1.05906451e+00
1.48038268e-01 -8.20866287e-01 -8.12569037e-02 -1.15591131e-01
1.69646859e-01 -1.15725410e+00 -2.38293514e-01 8.68707716e-01
-6.75194263e-01 1.60318241e-01 4.04881507e-01 1.25626779e+00
-1.74979225e-01 -1.84999526e-01 1.03071976e+00 9.72873624e-03
-5.12997329e-01 6.60506248e-01 -5.38353980e-01 6.50377851e-03
9.62871432e-01 3.40385646e-01 -1.33684576e-01 -5.05243599e-01
-5.16818702e-01 6.98956370e-01 2.71960557e-01 6.75294459e-01
5.89247227e-01 -1.32465875e+00 -2.89942861e-01 7.25880042e-02
-1.66455910e-01 2.97975451e-01 -3.22340846e-01 4.82058585e-01
-3.61041352e-02 3.98452878e-01 -1.95983171e-01 -3.58645648e-01
-6.55332446e-01 8.01991642e-01 5.50822794e-01 1.29929811e-01
-6.67477787e-01 5.51387250e-01 -7.23880306e-02 -7.23048747e-01
8.39731991e-01 -8.75146270e-01 -2.18345057e-02 -6.80541813e-01
4.21655387e-01 5.06540895e-01 -4.45669264e-01 -9.74937677e-02
1.29952312e-01 4.43965018e-01 1.62795454e-01 8.85727853e-02
1.06096506e+00 1.08308427e-01 2.28653267e-01 4.30132210e-01
1.00669551e+00 -3.46693009e-01 -1.98559821e+00 -2.65186816e-01
1.00862324e-01 -1.26213595e-01 -6.95622340e-02 -6.15207136e-01
-9.21235323e-01 6.05123997e-01 4.48966771e-01 5.12351356e-02
6.98169351e-01 -6.91201746e-01 7.76247859e-01 1.22865772e+00
4.71333802e-01 -1.22779536e+00 1.11105949e-01 9.53505218e-01
1.09869516e+00 -1.18237841e+00 -1.80318162e-01 1.39292911e-01
-7.91390896e-01 1.18688357e+00 1.02605569e+00 -7.42732406e-01
4.68563348e-01 2.65550047e-01 9.48185697e-02 3.23172867e-01
-7.36107171e-01 3.36326927e-01 2.17350945e-01 8.91049087e-01
-2.48823926e-01 -9.56115592e-03 9.12497416e-02 2.56546110e-01
2.44493425e-01 -1.51289538e-01 3.69067401e-01 9.80860889e-01
-7.10646331e-01 -6.80165946e-01 -6.06731713e-01 6.62707910e-02
-3.84869456e-01 4.35389847e-01 4.10666876e-02 1.08396673e+00
-3.45488369e-01 6.25959754e-01 -2.46827632e-01 -2.54227936e-01
6.24004006e-01 -2.74000436e-01 3.09002995e-01 -4.19062197e-01
-2.97985345e-01 -3.86940129e-02 -1.18669666e-01 -7.68350959e-01
7.75956735e-02 -4.31364089e-01 -1.54320312e+00 8.61058980e-02
-2.31476024e-01 -2.05190107e-02 5.35904229e-01 8.39432716e-01
3.02991897e-01 5.58084369e-01 6.95111752e-01 -9.48617697e-01
-1.83070076e+00 -9.60478485e-01 -2.31809244e-01 7.42907226e-02
9.16000366e-01 -1.01857233e+00 -3.71354491e-01 -1.62374243e-01] | [4.762088298797607, 2.1828415393829346] |
8241561e-1916-4e8d-821e-1ff904ff8e97 | multilingual-multimodal-machine-translation | null | null | https://aclanthology.org/W19-6809 | https://aclanthology.org/W19-6809.pdf | Multilingual Multimodal Machine Translation for Dravidian Languages utilizing Phonetic Transcription | null | ['John P. McCrae', 'Manel Zarrouk', 'Sridevy S', 'Bernardo Stearns', 'Ruba Priyadharshini', 'Arun Jayapal', 'Mihael Arcan', 'Bharathi Raja Chakravarthi'] | 2019-08-01 | null | null | null | ws-2019-8 | ['multimodal-machine-translation'] | ['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] | [-7.376296520233154, 3.715768575668335] |
89986952-0524-43ce-a50d-4f9a93cdcb19 | task-wise-split-gradient-boosting-trees-for | 2108.07107 | null | https://arxiv.org/abs/2108.07107v1 | https://arxiv.org/pdf/2108.07107v1.pdf | Task-wise Split Gradient Boosting Trees for Multi-center Diabetes Prediction | Diabetes prediction is an important data science application in the social healthcare domain. There exist two main challenges in the diabetes prediction task: data heterogeneity since demographic and metabolic data are of different types, data insufficiency since the number of diabetes cases in a single medical center is usually limited. To tackle the above challenges, we employ gradient boosting decision trees (GBDT) to handle data heterogeneity and introduce multi-task learning (MTL) to solve data insufficiency. To this end, Task-wise Split Gradient Boosting Trees (TSGB) is proposed for the multi-center diabetes prediction task. Specifically, we firstly introduce task gain to evaluate each task separately during tree construction, with a theoretical analysis of GBDT's learning objective. Secondly, we reveal a problem when directly applying GBDT in MTL, i.e., the negative task gain problem. Finally, we propose a novel split method for GBDT in MTL based on the task gain statistics, named task-wise split, as an alternative to standard feature-wise split to overcome the mentioned negative task gain problem. Extensive experiments on a large-scale real-world diabetes dataset and a commonly used benchmark dataset demonstrate TSGB achieves superior performance against several state-of-the-art methods. Detailed case studies further support our analysis of negative task gain problems and provide insightful findings. The proposed TSGB method has been deployed as an online diabetes risk assessment software for early diagnosis. | ['Guang Ning', 'Weiqing Wang', 'Yufang Bi', 'Yong Yu', 'Wei-Wei Tu', 'Mian Li', 'Tiange Wang', 'Yu Xu', 'Min Xu', 'Jieli Lu', 'Yanru Qu', 'Jian Shen', 'Xiawei Guo', 'Weinan Zhang', 'Zhiyun Zhao', 'Zhenghui Wang', 'Mingcheng Chen'] | 2021-08-16 | null | null | null | null | ['diabetes-prediction'] | ['medical'] | [ 8.54650587e-02 -2.69615203e-01 -5.01325727e-01 -6.35773063e-01
-8.16376925e-01 3.96234661e-01 1.28482282e-01 5.59381723e-01
-1.56173736e-01 8.26814353e-01 1.98802173e-01 -4.37957168e-01
-4.51467991e-01 -6.40465319e-01 -4.14522648e-01 -8.88930619e-01
-1.71866551e-01 7.06964731e-01 -7.28072599e-02 -7.19749406e-02
5.15098125e-02 -7.87717924e-02 -1.38977575e+00 4.77880061e-01
1.54046714e+00 1.13057494e+00 7.15622902e-02 1.93440065e-01
-2.36767635e-01 8.63378942e-01 -3.54618847e-01 -5.81136346e-01
2.06701860e-01 -4.50292259e-01 -5.21732807e-01 -1.79957580e-02
1.91634342e-01 -1.72882050e-01 5.46399653e-01 7.65108407e-01
9.46976960e-01 -5.30263245e-01 6.76028252e-01 -1.66947281e+00
-2.89133061e-02 3.80177975e-01 -7.03283846e-01 1.87168360e-01
2.46591941e-02 -2.57346243e-01 6.79050744e-01 -5.77863455e-01
3.93231839e-01 1.10293138e+00 1.01437151e+00 4.52998519e-01
-1.26983368e+00 -8.33386958e-01 1.90468207e-01 4.05022681e-01
-1.03723764e+00 -2.79710203e-01 7.92538822e-01 -8.54269743e-01
5.10618925e-01 3.09395373e-01 6.89669311e-01 1.04903650e+00
4.82383788e-01 7.50080764e-01 1.35675097e+00 -4.00261372e-01
2.54032344e-01 1.59792662e-01 3.98866028e-01 6.30050421e-01
2.76854157e-01 3.62491384e-02 -3.05986822e-01 -5.87845147e-01
4.36142772e-01 2.39020929e-01 7.35739842e-02 -6.74694300e-01
-1.25811458e+00 8.24989974e-01 3.14111650e-01 2.95458101e-02
-5.98629117e-01 -3.58551621e-01 8.57082725e-01 5.36565363e-01
8.98245931e-01 -1.85715824e-01 -8.99242640e-01 2.31131718e-01
-4.85980183e-01 4.35113877e-01 8.09930146e-01 7.34146535e-01
3.94539565e-01 -2.91422278e-01 -3.49382550e-01 1.11340773e+00
2.94410825e-01 1.42478496e-01 5.73444307e-01 -3.15083802e-01
7.25168288e-01 8.96105230e-01 -1.01835169e-01 -7.77501822e-01
-7.48308778e-01 -8.18856359e-01 -1.40996039e+00 8.99444893e-02
5.85642099e-01 -3.02485913e-01 -9.04895663e-01 1.63029015e+00
8.41214299e-01 2.42780045e-01 4.72046286e-02 8.57655823e-01
7.91875601e-01 7.95929208e-02 5.56762099e-01 -4.45137084e-01
1.69715691e+00 -7.08823800e-01 -7.09252954e-01 -2.88791731e-02
1.17644238e+00 -4.48251247e-01 7.66011953e-01 7.39750147e-01
-6.01766825e-01 -3.57549250e-01 -6.74501240e-01 2.29115143e-01
-1.48954973e-01 2.59276927e-01 7.33227849e-01 8.59166145e-01
-4.58583206e-01 2.16923535e-01 -6.46039367e-01 -3.84214193e-01
5.36234140e-01 1.71419516e-01 -1.19305238e-01 -3.65468055e-01
-1.26276648e+00 9.03940856e-01 2.79611319e-01 3.98851782e-02
-5.52353621e-01 -1.16960692e+00 -5.56833327e-01 -2.96677500e-01
3.21327090e-01 -1.22800589e+00 1.03400612e+00 -9.71376121e-01
-1.04817212e+00 8.77620459e-01 -9.61198285e-02 -4.29623485e-01
9.55102801e-01 -1.05460815e-01 -4.29040790e-01 -3.73332202e-01
1.97193682e-01 2.64364425e-02 4.79434550e-01 -1.05607712e+00
-9.53613937e-01 -1.01492655e+00 -6.28540456e-01 1.87766016e-01
-1.46488845e-01 2.81177554e-02 1.65620878e-01 -8.36845517e-01
1.50656402e-01 -4.89726871e-01 -6.01175845e-01 -1.93606198e-01
-4.31256175e-01 -4.71905082e-01 4.59375262e-01 -9.47248042e-01
1.46082032e+00 -1.78232872e+00 6.61957636e-02 7.72699788e-02
4.07882690e-01 1.99761420e-01 1.13638572e-01 1.17218152e-01
-2.56832063e-01 -2.75411773e-02 -4.37955046e-03 -2.18158811e-01
-4.15116876e-01 1.99347228e-01 2.11242035e-01 1.58680886e-01
9.95607525e-02 6.26289904e-01 -6.74014091e-01 -6.70285225e-01
2.20238239e-01 2.53095269e-01 -6.18450940e-01 1.57622233e-01
-6.45290315e-02 6.41594589e-01 -6.75935328e-01 1.01325750e+00
8.62223446e-01 -2.83345371e-01 4.35323298e-01 -3.66928726e-01
-2.40035187e-02 1.95021451e-01 -9.97509718e-01 1.49089193e+00
-3.57772887e-01 -4.11899477e-01 6.85450481e-03 -1.68351603e+00
1.05162811e+00 2.91378438e-01 8.66938353e-01 -8.48873675e-01
-5.16342930e-03 4.73993301e-01 1.73352182e-01 -7.80086815e-01
-4.49839175e-01 -6.23897851e-01 3.28323655e-02 5.18220700e-02
-1.91840306e-01 7.40101278e-01 7.08945021e-02 -1.98863268e-01
1.01300073e+00 -7.57422820e-02 7.13864625e-01 -3.93267512e-01
3.78930926e-01 9.44667235e-02 1.22304869e+00 6.22903585e-01
-4.91196215e-01 3.21467727e-01 6.74539745e-01 -9.38303947e-01
-7.66435087e-01 -5.66008449e-01 -5.46196342e-01 1.30487049e+00
-1.94477737e-01 -2.60228485e-01 -3.43275815e-01 -1.12486005e+00
4.65465188e-01 2.36417398e-01 -6.51574910e-01 1.28808945e-01
-4.36163276e-01 -1.78155982e+00 1.98550761e-01 2.78915763e-01
5.01907587e-01 -6.03693128e-01 -6.05974197e-01 5.59962273e-01
-3.95272821e-01 -8.72208297e-01 7.40790665e-02 3.82570028e-01
-1.22717988e+00 -1.20907855e+00 -9.64863241e-01 -7.55797267e-01
3.63430917e-01 -2.93083098e-02 1.45770502e+00 -2.55643371e-02
-4.34405148e-01 -2.54377037e-01 -5.46993971e-01 -6.94024444e-01
-3.83292377e-01 1.90312698e-01 -1.30023152e-01 1.96084604e-01
6.43120110e-01 -6.52280688e-01 -8.23927283e-01 4.98124123e-01
-3.99886966e-01 4.37658727e-01 8.25374365e-01 1.08978319e+00
5.15111625e-01 -6.29192516e-02 1.27547300e+00 -1.24855614e+00
3.43194038e-01 -8.55056882e-01 -2.97516167e-01 4.62208629e-01
-1.21234727e+00 -9.36933234e-02 2.60204881e-01 -3.60597968e-01
-9.03315544e-01 1.06360519e-03 -1.09841265e-01 2.02994123e-02
-1.07863151e-01 7.98260570e-01 -2.88427532e-01 2.22385719e-01
5.61116993e-01 2.30430230e-01 2.93627203e-01 -8.57169390e-01
-3.54138672e-01 8.47269535e-01 -3.18084776e-01 -5.11286974e-01
-2.69406885e-02 9.71868932e-02 7.01325759e-02 -2.84530282e-01
-8.63375366e-01 -4.65236783e-01 -4.22715724e-01 -2.24951170e-02
6.56884193e-01 -1.22587812e+00 -6.73901737e-01 7.33654499e-01
-7.11671114e-01 -2.22381532e-01 1.29448712e-01 5.59607506e-01
-3.46934050e-01 1.71898842e-01 -4.01498586e-01 -7.59688437e-01
-7.41280258e-01 -1.07782090e+00 9.91722345e-01 -2.35892981e-01
2.27783620e-01 -1.04208457e+00 1.19786620e-01 7.46896505e-01
3.36152017e-01 6.72070920e-01 1.45101142e+00 -8.16590309e-01
-5.17104045e-02 3.64362448e-02 -4.83132690e-01 3.73358518e-01
2.53735900e-01 -4.57519203e-01 -5.36667705e-01 -2.88989633e-01
3.72109413e-02 -2.20334396e-01 7.73428798e-01 7.66677141e-01
1.14692688e+00 -2.15336949e-01 -5.06816566e-01 4.80779380e-01
1.50249565e+00 2.62980402e-01 3.54054868e-01 4.56466138e-01
6.59760833e-01 7.27595508e-01 8.00311804e-01 9.36881781e-01
8.72955441e-01 9.32345271e-01 2.35729396e-01 -5.37654519e-01
-1.06072053e-01 1.06692918e-01 -3.21464688e-02 7.34274983e-01
-2.68398285e-01 2.51582146e-01 -1.25410950e+00 2.96155661e-01
-2.11252785e+00 -2.95659453e-01 -5.76398373e-01 2.37749982e+00
1.03839946e+00 5.70877977e-02 5.29951870e-01 2.80113369e-01
6.93322897e-01 -3.88260722e-01 -5.95829248e-01 -3.05662334e-01
1.64060041e-01 -2.00874045e-01 3.72039586e-01 2.89851930e-02
-1.18072045e+00 2.79713988e-01 5.41021967e+00 7.97125518e-01
-1.10012388e+00 4.72121507e-01 1.00293565e+00 1.10655613e-01
1.86239779e-01 -2.26387739e-01 -7.84969747e-01 7.68453538e-01
6.61579370e-01 -2.44701147e-01 -1.71139777e-01 8.44745278e-01
3.66456032e-01 -9.45670009e-02 -1.10757160e+00 9.71129835e-01
-1.92416191e-01 -8.28634918e-01 1.51075721e-01 1.60130799e-01
3.46866131e-01 -6.54947460e-02 -2.78163552e-01 5.53976595e-01
1.69902191e-01 -6.46935463e-01 2.17398822e-01 3.44970137e-01
6.94321156e-01 -6.58877194e-01 1.00066447e+00 4.71456945e-01
-9.59988058e-01 -4.08942372e-01 -1.66702986e-01 -1.13256112e-01
-3.76738757e-02 1.48382795e+00 -1.03527582e+00 1.11612964e+00
7.64002562e-01 8.59716892e-01 -5.61153650e-01 1.26215506e+00
4.76812154e-01 5.51757455e-01 -3.68479416e-02 2.63523996e-01
-2.50616968e-01 -2.43392497e-01 2.84439236e-01 1.10420144e+00
4.76833403e-01 -1.15315333e-01 5.31276643e-01 2.86483377e-01
3.14766765e-01 6.68171585e-01 -3.37565213e-01 6.08232796e-01
5.23320474e-02 1.02219462e+00 -2.79876798e-01 -4.53641742e-01
-4.42156196e-01 4.87263352e-01 1.17540278e-01 6.37133643e-02
-6.81597412e-01 4.46928553e-02 5.11566699e-01 2.69436657e-01
-2.49226019e-02 4.66781408e-01 -6.55568421e-01 -1.28951252e+00
4.60707881e-02 -1.12841606e+00 9.04395819e-01 -2.31329471e-01
-1.77661192e+00 3.82477671e-01 -8.34385455e-02 -1.41901612e+00
-8.94750655e-02 -3.61895829e-01 -9.36671197e-02 7.91368663e-01
-1.88993835e+00 -1.31373894e+00 -4.78663117e-01 7.00268865e-01
5.63807011e-01 -1.99952707e-01 9.23104286e-01 8.78390253e-01
-9.68352020e-01 6.34458661e-01 1.59535944e-01 -1.11027107e-01
9.06338990e-01 -1.21072948e+00 -2.27435276e-01 2.08518058e-01
-7.43334055e-01 1.12392873e-01 3.85916650e-01 -8.33361089e-01
-1.15302575e+00 -1.29492903e+00 1.02400649e+00 -9.63182673e-02
2.16696978e-01 -3.27303022e-01 -8.09893489e-01 4.06395465e-01
-3.21943015e-01 1.54255070e-02 1.11194730e+00 7.24414408e-01
-3.71581197e-01 -7.25052476e-01 -1.40201974e+00 1.41011842e-03
1.01150620e+00 3.04132581e-01 -4.16327775e-01 5.27856886e-01
2.74811983e-01 -3.60413551e-01 -1.17235065e+00 1.06006646e+00
8.58925819e-01 -1.08039284e+00 9.70088542e-01 -7.95176387e-01
4.61573064e-01 9.89714339e-02 -1.49795443e-01 -1.25987983e+00
-3.08896005e-01 -7.37770647e-02 -6.90746084e-02 1.12801826e+00
3.20531189e-01 -8.79781485e-01 6.99375391e-01 4.98415828e-01
4.63918969e-02 -1.10193312e+00 -1.20354772e+00 -7.60361433e-01
1.50120631e-01 -1.60581678e-01 4.81004655e-01 1.38076890e+00
-1.07289180e-01 4.04489398e-01 -6.26416922e-01 -5.32539710e-02
9.71970260e-01 2.06245422e-01 6.70981348e-01 -1.87267220e+00
-6.54291287e-02 -1.67354718e-01 -4.12376881e-01 -4.60592896e-01
-3.96560609e-01 -8.33948314e-01 -4.12237883e-01 -1.53120208e+00
6.52110755e-01 -1.11044407e+00 -7.03495800e-01 5.58483541e-01
-4.34133500e-01 -1.46643743e-01 -1.06887959e-01 1.67390764e-01
-4.24451023e-01 3.52079332e-01 1.23849928e+00 -9.40568149e-02
-3.13336164e-01 4.84789431e-01 -8.05398285e-01 4.31114078e-01
7.63877153e-01 -6.14247680e-01 -3.53156626e-01 -2.00697154e-01
1.49287745e-01 3.02324653e-01 3.39117914e-01 -8.68360877e-01
-1.97840601e-01 -2.74994731e-01 3.05514932e-01 -5.03695011e-01
-1.58836469e-01 -8.95557404e-01 3.78715366e-01 9.74175572e-01
-2.64307082e-01 -6.05868250e-02 -1.72431186e-01 5.84881186e-01
-1.84854507e-01 2.40986481e-01 6.11874163e-01 -2.56308652e-02
-3.11122268e-01 1.76072791e-01 -4.00275253e-02 -4.57920209e-02
1.15818024e+00 8.45859647e-02 -1.90730438e-01 2.34316036e-01
-9.35552835e-01 5.29500663e-01 -2.06184506e-01 3.64399105e-01
3.20124835e-01 -1.44088364e+00 -1.31648314e+00 2.83105701e-01
4.70251232e-01 -4.86671068e-02 5.38460314e-01 1.47410858e+00
-1.25992045e-01 1.29634827e-01 -2.18961895e-01 -9.04627800e-01
-1.61073339e+00 5.64980447e-01 2.97576427e-01 -8.34214151e-01
-7.24040151e-01 3.63504350e-01 3.14702213e-01 -5.33791125e-01
4.34879422e-01 -5.62182188e-01 -3.02691489e-01 4.64899957e-01
4.46849823e-01 4.81085986e-01 6.09115362e-01 -2.72367839e-02
-4.50377882e-01 4.63997304e-01 -1.84517547e-01 5.67135096e-01
1.43720269e+00 -1.62917674e-01 -6.53832778e-02 4.43195134e-01
9.48508680e-01 -6.11968458e-01 -7.08237886e-01 -4.56491590e-01
2.47046500e-01 -3.91285300e-01 5.07370234e-02 -1.32665384e+00
-1.01967514e+00 7.03656912e-01 1.18044579e+00 2.50546515e-01
1.45788872e+00 -2.75617480e-01 6.83557630e-01 -2.01705787e-02
3.92797917e-01 -8.95315588e-01 -2.14024037e-01 -4.10232227e-03
7.94788182e-01 -1.62485564e+00 2.96183745e-04 -5.20442247e-01
-8.28589559e-01 8.12383413e-01 4.48852897e-01 2.95061052e-01
8.35814714e-01 1.34408265e-01 2.90403962e-01 7.52495416e-03
-8.64034534e-01 -6.68266341e-02 1.07041663e-02 6.91238761e-01
4.21650052e-01 2.75736868e-01 -8.66450369e-01 9.41494644e-01
2.22976983e-01 7.52287328e-01 -5.16394302e-02 8.36832404e-01
-1.99977964e-01 -1.38295877e+00 -3.01633418e-01 8.58763158e-01
-6.38824761e-01 -4.83779646e-02 1.04116686e-01 5.27383268e-01
5.51427841e-01 8.21725249e-01 -2.88159460e-01 -4.47861880e-01
4.36567783e-01 3.75022560e-01 3.04394931e-01 -2.37111330e-01
-7.06446707e-01 2.71702707e-01 2.93352067e-01 -4.13929254e-01
-6.15776062e-01 -6.10009074e-01 -7.51578569e-01 3.23209912e-02
-1.65875882e-01 9.00526568e-02 7.99300611e-01 8.57795000e-01
6.84846997e-01 6.91473305e-01 8.61646414e-01 -3.18216711e-01
-7.81117439e-01 -9.86913264e-01 -5.16985595e-01 4.76042956e-01
3.94720554e-01 -7.50375330e-01 -9.72151384e-02 -3.66394930e-02] | [14.255770683288574, 3.118441581726074] |
2365bf93-b926-4c6d-92e5-ef7872dad425 | fastwave-accelerating-autoregressive | 2002.04971 | null | https://arxiv.org/abs/2002.04971v1 | https://arxiv.org/pdf/2002.04971v1.pdf | FastWave: Accelerating Autoregressive Convolutional Neural Networks on FPGA | Autoregressive convolutional neural networks (CNNs) have been widely exploited for sequence generation tasks such as audio synthesis, language modeling and neural machine translation. WaveNet is a deep autoregressive CNN composed of several stacked layers of dilated convolution that is used for sequence generation. While WaveNet produces state-of-the art audio generation results, the naive inference implementation is quite slow; it takes a few minutes to generate just one second of audio on a high-end GPU. In this work, we develop the first accelerator platform~\textit{FastWave} for autoregressive convolutional neural networks, and address the associated design challenges. We design the Fast-Wavenet inference model in Vivado HLS and perform a wide range of optimizations including fixed-point implementation, array partitioning and pipelining. Our model uses a fully parameterized parallel architecture for fast matrix-vector multiplication that enables per-layer customized latency fine-tuning for further throughput improvement. Our experiments comparatively assess the trade-off between throughput and resource utilization for various optimizations. Our best WaveNet design on the Xilinx XCVU13P FPGA that uses only on-chip memory, achieves 66 faster generation speed compared to CPU implementation and 11 faster generation speed than GPU implementation. | ['Farinaz Koushanfar', 'Shehzeen Hussain', 'Mojan Javaheripi', 'Ryan Kastner', 'Paarth Neekhara'] | 2020-02-09 | null | null | null | null | ['audio-generation'] | ['audio'] | [ 2.78361917e-01 -1.71746954e-01 4.23743986e-02 -1.63683996e-01
-4.69462961e-01 -2.81068146e-01 3.90714467e-01 -2.85839260e-01
-2.64042646e-01 3.50029171e-01 4.69987988e-01 -7.60001004e-01
4.27793086e-01 -9.10702765e-01 -6.76548779e-01 -5.36543906e-01
-9.78830382e-02 1.27973789e-02 -3.73160839e-02 -3.17178458e-01
-1.04337417e-01 4.77441281e-01 -1.59267843e+00 7.14285553e-01
1.87728658e-01 1.09340501e+00 6.02679141e-02 1.39885306e+00
4.62052040e-02 1.12473810e+00 -7.34610736e-01 -1.96574196e-01
1.37927830e-01 -3.48502189e-01 -5.37688017e-01 -5.61057746e-01
3.19802254e-01 -6.01402700e-01 -5.68485320e-01 5.26948512e-01
1.00310242e+00 -7.15774074e-02 5.66727817e-01 -1.24245167e+00
-3.22435617e-01 1.25607860e+00 -4.62416798e-01 3.95185530e-01
-2.24411208e-02 3.04867208e-01 7.02174306e-01 -7.21416235e-01
7.38281682e-02 1.27031779e+00 9.12996948e-01 4.78969991e-01
-9.68332589e-01 -1.16643131e+00 -3.86319607e-01 -8.66871029e-02
-1.32692742e+00 -7.36412466e-01 7.22274482e-01 -4.27406073e-01
1.59567499e+00 2.56392717e-01 7.83218443e-01 1.41452384e+00
6.30841970e-01 5.40872395e-01 4.55048472e-01 -4.70949173e-01
4.10306394e-01 -3.61471325e-01 5.95386028e-02 6.58755898e-01
-1.22603863e-01 1.30513683e-01 -9.89509404e-01 -2.70818442e-01
1.22829783e+00 -4.26143259e-01 -4.75750007e-02 5.78662038e-01
-1.20915723e+00 7.63741791e-01 3.28731745e-01 -9.50501040e-02
-4.83371764e-01 1.19742191e+00 1.07106543e+00 3.64087343e-01
3.02464694e-01 3.28868628e-01 -5.50830364e-01 -6.51496768e-01
-1.01358986e+00 4.60526675e-01 6.52630687e-01 1.12062073e+00
2.36824885e-01 1.31829417e+00 -2.77220279e-01 8.44309628e-01
1.48809925e-01 2.42044613e-01 8.06633234e-01 -6.44265831e-01
4.59404618e-01 4.47431616e-02 -6.28564417e-01 -7.44641602e-01
-5.25467098e-01 -8.27694595e-01 -1.36486185e+00 2.35411495e-01
-1.98634401e-01 -8.75214219e-01 -8.26991200e-01 1.30814731e+00
2.92799119e-02 5.10494590e-01 1.31413072e-01 8.11812878e-01
1.10891926e+00 1.07625282e+00 1.25352934e-01 2.03812525e-01
1.48180521e+00 -1.18826592e+00 -2.72102058e-01 -1.72219142e-01
7.33407199e-01 -1.10106754e+00 1.15869951e+00 4.94846284e-01
-1.13302815e+00 -8.35323572e-01 -1.44156897e+00 -4.54342276e-01
1.34620622e-01 7.33992457e-01 9.05673981e-01 9.43679214e-01
-1.26755726e+00 5.48145950e-01 -9.72769201e-01 1.24529704e-01
5.95396720e-02 6.33704841e-01 1.02737084e-01 7.67242551e-01
-1.00830078e+00 -4.92368126e-03 5.26974320e-01 1.88347653e-01
-7.47554064e-01 -1.29961717e+00 -7.34000146e-01 4.39401418e-01
-3.24340701e-01 -1.15257931e+00 1.48865163e+00 -8.89809012e-01
-2.33585286e+00 2.05529004e-01 7.97529370e-02 -1.10875165e+00
1.06147468e-01 -4.52586323e-01 -4.92710531e-01 -2.61541337e-01
-4.28725690e-01 9.09960687e-01 8.90299559e-01 -1.85078532e-01
-4.57250386e-01 1.69677451e-01 -3.11504275e-01 -6.63972169e-04
-3.79569322e-01 3.42323601e-01 -2.33155578e-01 -1.08429694e+00
-3.03631991e-01 -1.11686158e+00 -4.68887836e-01 -3.27610016e-01
-4.05319095e-01 1.26082554e-01 7.65325308e-01 -3.81077617e-01
1.24826169e+00 -2.02912426e+00 -2.58406192e-01 8.02610815e-02
2.88343232e-04 2.51738757e-01 -4.20272984e-02 3.21937263e-01
-3.66147935e-01 -1.45303264e-01 3.07743639e-01 -2.49841154e-01
-3.80621590e-02 -1.42197579e-01 -9.92437720e-01 1.57215610e-01
2.29630783e-01 8.91063929e-01 -4.71889794e-01 1.73531135e-03
1.47273853e-01 7.85757959e-01 -1.17021155e+00 1.41368225e-01
-2.28166461e-01 1.46911189e-01 -1.97830811e-01 3.25139195e-01
4.19662982e-01 9.76798683e-02 -8.90519693e-02 -3.17256212e-01
-3.21253151e-01 4.59362388e-01 -1.00236535e+00 1.63754666e+00
-1.00542903e+00 1.18822718e+00 -3.84251803e-01 -5.30646801e-01
1.14965200e+00 4.54042196e-01 8.16662982e-02 -3.46778482e-01
4.26598936e-01 2.18366712e-01 9.52567160e-02 -5.40996827e-02
1.10075331e+00 2.46613964e-01 -2.39355490e-01 2.87389934e-01
1.30966932e-01 -3.85394692e-01 -1.25677049e-01 -3.81383225e-02
1.11536026e+00 1.06528625e-01 2.42997169e-01 -3.15206826e-01
1.53904215e-01 -7.63824955e-02 2.99452513e-01 6.60715759e-01
6.34008825e-01 4.88895059e-01 6.61884308e-01 -6.79364622e-01
-1.33366704e+00 -7.33672440e-01 2.92940974e-01 1.61381805e+00
-8.40525210e-01 -8.27236414e-01 -8.77912223e-01 5.27039111e-01
-6.69008613e-01 7.47806191e-01 -1.37775540e-01 -1.35654420e-01
-8.38561296e-01 -8.33658278e-01 1.45272458e+00 8.00803781e-01
4.73041654e-01 -1.24200487e+00 -1.19019461e+00 6.32602394e-01
5.59525788e-01 -1.18369615e+00 -3.34960312e-01 1.57162145e-01
-8.14048469e-01 -7.55340457e-02 -4.16079611e-01 -7.03170478e-01
1.38355315e-01 -1.94275633e-01 1.33366513e+00 -1.55649737e-01
-4.81886089e-01 -2.71297723e-01 -2.34915838e-01 -7.55069137e-01
-4.78855640e-01 3.69622558e-01 2.07905874e-01 -4.25697893e-01
-2.84466892e-01 -9.42768812e-01 -6.51704013e-01 -2.87143588e-01
-6.40773177e-01 5.41065753e-01 5.02770245e-01 1.01169169e+00
3.96055937e-01 -1.37885302e-01 2.63950109e-01 -7.47868001e-01
9.14318383e-01 -2.12482795e-01 -7.51579165e-01 -3.65559906e-01
5.86219272e-03 1.81407258e-01 9.86832738e-01 -6.42067194e-01
-9.70183969e-01 2.84031361e-01 -8.70762587e-01 -5.90374112e-01
3.84435534e-01 5.74745297e-01 5.16216159e-01 -9.65065695e-03
9.14467454e-01 8.93389434e-02 -3.31428885e-01 -2.03365237e-01
3.29818934e-01 6.90639734e-01 8.53527784e-01 -6.42412782e-01
3.60103309e-01 4.00308408e-02 1.24593452e-01 -1.21490383e+00
-2.11953714e-01 2.94825565e-02 -4.30055931e-02 -8.60669836e-02
6.32678509e-01 -1.46288812e+00 -1.04195726e+00 5.64813077e-01
-1.40016282e+00 -6.09378219e-01 -1.63838223e-01 5.52518308e-01
-6.01363063e-01 -2.25125939e-01 -9.46530104e-01 -7.13205159e-01
-1.41207016e+00 -1.36719716e+00 1.32445049e+00 1.67613931e-03
-5.67941725e-01 -5.52629530e-01 -8.62106979e-02 -2.36441091e-01
6.50736034e-01 1.29897714e-01 9.89971280e-01 -1.64932132e-01
-6.65838540e-01 -3.84039357e-02 -1.55829102e-01 1.68427750e-01
-5.24351716e-01 4.43590730e-01 -1.07229710e+00 -1.34916350e-01
-2.88616806e-01 -1.83233857e-01 7.16367185e-01 6.14695251e-01
1.58168697e+00 -4.00074333e-01 4.74852137e-02 1.18742287e+00
1.21142602e+00 1.44091576e-01 8.29768002e-01 -8.55061784e-02
8.44736218e-01 -2.62179300e-02 1.89186037e-01 8.96223128e-01
-1.80870980e-01 8.23303521e-01 1.29458457e-01 -1.37040809e-01
-2.31870651e-01 -1.37675762e-01 5.96657872e-01 1.42127085e+00
-1.49696380e-01 -2.24543989e-01 -9.51182723e-01 1.66678950e-01
-1.46529639e+00 -7.95621932e-01 -4.21142966e-01 1.93173015e+00
7.84100175e-01 1.70438454e-01 -8.84181708e-02 3.59085649e-01
3.03396042e-02 9.65369344e-02 -6.00279383e-02 -1.04172981e+00
2.96465963e-01 1.14510214e+00 6.18023157e-01 2.95200944e-01
-8.79378498e-01 1.08970582e+00 6.51988268e+00 1.13126874e+00
-1.65680552e+00 -1.53260333e-02 8.49916637e-01 -4.78549778e-01
-8.24961960e-02 -2.36774892e-01 -1.18493915e+00 2.16504037e-01
1.56679749e+00 -2.25859731e-01 1.53741762e-01 1.28947484e+00
3.35772783e-01 3.99740279e-01 -9.53521848e-01 1.16541278e+00
-1.91852868e-01 -1.94118619e+00 1.54663011e-01 -7.77882114e-02
6.17903829e-01 2.52221823e-01 2.47044668e-01 5.75123012e-01
3.82251531e-01 -1.19968855e+00 8.12355936e-01 1.13006018e-01
1.08758235e+00 -1.32266259e+00 5.58355391e-01 -3.22316512e-02
-1.38488460e+00 -2.23743804e-02 -4.37398076e-01 -3.43432814e-01
8.58757868e-02 6.03203654e-01 -1.33769095e+00 1.84465311e-02
6.26703620e-01 2.17757508e-01 -1.74923286e-01 6.99133337e-01
2.59080380e-01 1.03691041e+00 -2.34059140e-01 -2.71007776e-01
4.74535167e-01 1.13379441e-01 2.98754305e-01 1.70112169e+00
7.66835868e-01 -7.43504465e-02 -2.31237784e-01 4.84748065e-01
-1.72638312e-01 1.61118865e-01 -2.29298547e-01 6.54160753e-02
4.21707153e-01 1.23462141e+00 -4.67915207e-01 -4.17823464e-01
-1.38302548e-02 8.13689351e-01 1.14409961e-01 2.14652047e-02
-1.19446957e+00 -6.63788319e-01 1.02325952e+00 -6.76275715e-02
9.21045989e-02 -5.52718639e-01 -6.15081549e-01 -7.44766235e-01
-3.45412493e-01 -9.15878832e-01 -1.11345716e-01 -1.03351247e+00
-4.32539403e-01 1.13805628e+00 -3.03753257e-01 -1.20347524e+00
-1.08101201e+00 -5.17045319e-01 -6.12341344e-01 1.05914235e+00
-9.56572473e-01 -1.13854897e+00 -4.01681960e-01 2.54626900e-01
9.84866977e-01 -7.04889297e-01 1.02222705e+00 3.08092833e-01
-5.44261396e-01 9.43857968e-01 -8.75282213e-02 -9.20159218e-04
3.71048450e-02 -7.84938693e-01 1.26526213e+00 7.72362530e-01
1.77733704e-01 6.81062758e-01 8.91359091e-01 -4.04721648e-01
-1.67869842e+00 -1.44307935e+00 4.50343698e-01 4.63952720e-01
8.31774056e-01 -5.16230166e-01 -6.51856065e-01 6.26701951e-01
3.75346094e-01 -1.21021405e-01 6.82957709e-01 -1.58945154e-02
-2.42139950e-01 8.64994130e-04 -4.02825773e-01 1.15153086e+00
8.55501890e-01 -3.97515595e-01 3.53112131e-01 3.68133187e-01
1.24857450e+00 -1.07600045e+00 -8.89080584e-01 2.38307253e-01
7.42436349e-01 -8.30955565e-01 1.08126915e+00 -3.12792033e-01
8.68087053e-01 -4.15619351e-02 1.09801954e-02 -1.11231494e+00
-2.51332670e-01 -1.21122396e+00 -1.32219091e-01 9.74180639e-01
3.86169791e-01 -2.15622932e-01 1.08249319e+00 -3.52482731e-03
-7.12543964e-01 -8.91021490e-01 -7.63233483e-01 -3.61584663e-01
-2.34884873e-01 -9.57238257e-01 7.87043810e-01 3.34927917e-01
-2.96274364e-01 6.29055202e-01 -7.14986801e-01 1.06899478e-01
1.41444549e-01 -3.77985090e-01 1.16047239e+00 -6.51332855e-01
-6.90768659e-01 -4.12437767e-01 -4.98086184e-01 -1.20691216e+00
6.29068688e-02 -5.96709192e-01 -3.50162350e-02 -7.77134657e-01
-3.59793931e-01 -3.73896688e-01 3.49649370e-01 3.68964434e-01
3.22514623e-01 4.67859149e-01 2.61363294e-03 -1.98983878e-01
3.07351351e-01 5.10456681e-01 7.01528847e-01 2.07589403e-01
-4.18433070e-01 6.83132792e-03 -3.64103049e-01 7.32265115e-01
8.75922918e-01 -1.83476955e-01 -7.15479612e-01 -7.58802116e-01
3.62023801e-01 3.15405488e-01 3.25966626e-01 -1.55180728e+00
3.17593962e-01 2.36931831e-01 5.14410794e-01 -6.73764527e-01
8.18344891e-01 -2.59535551e-01 6.23462319e-01 5.24156451e-01
-4.44312990e-01 5.20061493e-01 5.89734614e-01 5.94028011e-02
-6.05108514e-02 -1.88322440e-01 7.01780021e-01 2.17742339e-01
-7.02169299e-01 3.03148270e-01 -7.32937038e-01 -2.91060567e-01
4.63884413e-01 -1.02730744e-01 -1.92026421e-01 -3.38866115e-01
-3.53236854e-01 -4.94318515e-01 -2.02028796e-01 3.54291290e-01
7.88153350e-01 -1.27523851e+00 -9.39240515e-01 3.49564344e-01
-3.42458844e-01 5.00877127e-02 5.39788723e-01 3.06735039e-01
-1.37148631e+00 7.66915083e-01 -3.90986532e-01 -5.54978251e-01
-1.62644541e+00 6.91954419e-02 3.34970474e-01 -2.85382032e-01
-8.83163691e-01 1.17409968e+00 1.49630263e-01 -2.97271367e-02
-3.12889554e-02 -6.61020398e-01 2.19881702e-02 -2.94407070e-01
7.17724323e-01 3.95858616e-01 4.13168013e-01 -2.88587928e-01
-6.55289963e-02 1.93265095e-01 2.86240522e-02 -2.31046483e-01
1.19463205e+00 6.47377253e-01 -1.42006785e-01 2.28575662e-01
1.27674842e+00 -2.34420955e-01 -9.72352862e-01 1.24969892e-01
-5.13791203e-01 -4.80739772e-02 4.61725056e-01 -1.40230343e-01
-1.23381877e+00 9.96075869e-01 5.21151245e-01 -3.60458672e-01
1.21911848e+00 -5.45940101e-01 9.44275320e-01 4.99689072e-01
1.59094870e-01 -7.64640808e-01 1.51485831e-01 9.24493611e-01
1.02069509e+00 -1.97436079e-01 1.96131572e-01 -5.23163974e-01
-6.99807286e-01 1.39509416e+00 5.12186885e-01 -4.04453784e-01
4.89923656e-01 1.13078082e+00 -1.73833489e-01 1.04974039e-01
-1.34634268e+00 3.57396215e-01 2.04267621e-01 4.19887751e-01
9.07570660e-01 2.29810447e-01 5.83220161e-02 5.29306233e-01
-1.30916274e+00 1.47144318e-01 6.85337067e-01 6.31536186e-01
9.77650192e-03 -7.65569508e-01 -2.29107320e-01 4.94206518e-01
-6.21518552e-01 -6.70702696e-01 2.92823046e-01 4.77056116e-01
-2.67305285e-01 5.32096267e-01 7.01979101e-01 -6.16805553e-01
1.82398371e-02 -3.86731178e-01 3.05216253e-01 -4.24105614e-01
-1.22905648e+00 3.08548540e-01 4.06978488e-01 -2.54846454e-01
1.61426887e-01 -2.87207723e-01 -1.08596349e+00 -6.72747850e-01
2.94060986e-02 -2.38537833e-01 1.03682065e+00 5.05725086e-01
6.53701246e-01 1.28267145e+00 3.89273196e-01 -1.09948874e+00
-3.36731821e-01 -8.68056893e-01 -2.46544197e-01 -6.13522708e-01
9.48344171e-02 -1.31411985e-01 3.59219819e-01 2.94503629e-01] | [15.236981391906738, 5.765880584716797] |
92880aef-a1dd-457e-90a8-8b1872f062ff | top-k-extreme-contextual-bandits-with-arm | 2102.07800 | null | https://arxiv.org/abs/2102.07800v1 | https://arxiv.org/pdf/2102.07800v1.pdf | Top-$k$ eXtreme Contextual Bandits with Arm Hierarchy | Motivated by modern applications, such as online advertisement and recommender systems, we study the top-$k$ extreme contextual bandits problem, where the total number of arms can be enormous, and the learner is allowed to select $k$ arms and observe all or some of the rewards for the chosen arms. We first propose an algorithm for the non-extreme realizable setting, utilizing the Inverse Gap Weighting strategy for selecting multiple arms. We show that our algorithm has a regret guarantee of $O(k\sqrt{(A-k+1)T \log (|\mathcal{F}|T)})$, where $A$ is the total number of arms and $\mathcal{F}$ is the class containing the regression function, while only requiring $\tilde{O}(A)$ computation per time step. In the extreme setting, where the total number of arms can be in the millions, we propose a practically-motivated arm hierarchy model that induces a certain structure in mean rewards to ensure statistical and computational efficiency. The hierarchical structure allows for an exponential reduction in the number of relevant arms for each context, thus resulting in a regret guarantee of $O(k\sqrt{(\log A-k+1)T \log (|\mathcal{F}|T)})$. Finally, we implement our algorithm using a hierarchical linear function class and show superior performance with respect to well-known benchmarks on simulated bandit feedback experiments using extreme multi-label classification datasets. On a dataset with three million arms, our reduction scheme has an average inference time of only 7.9 milliseconds, which is a 100x improvement. | ['Inderjit Dhillon', 'Daniel Hill', 'Dean Foster', 'Rahul Kidambi', 'Lexing Ying', 'Alexander Rakhlin', 'Rajat Sen'] | 2021-02-15 | null | null | null | null | ['extreme-multi-label-classification'] | ['methodology'] | [ 2.14651182e-01 2.24707037e-01 -6.46868110e-01 -4.73737359e-01
-1.54432535e+00 -9.90462601e-01 -1.48464993e-01 2.00988278e-01
-6.91138685e-01 8.80647242e-01 -2.97343642e-01 -8.77084553e-01
-7.50109076e-01 -9.27768886e-01 -1.34509122e+00 -7.45593369e-01
-3.93404216e-01 8.49398553e-01 -2.53928632e-01 -2.79570818e-02
1.21148258e-01 2.31790915e-01 -1.41628075e+00 -1.54277040e-02
7.54196465e-01 1.67058957e+00 -3.13143104e-01 7.65022278e-01
-2.33360514e-01 5.12130558e-01 -4.26054537e-01 -8.05221438e-01
6.06429040e-01 -4.69387144e-01 -1.00558960e+00 -8.66928175e-02
6.45269513e-01 -3.03046227e-01 -3.28660280e-01 1.21074879e+00
2.87727147e-01 2.45474219e-01 3.11053813e-01 -8.11959982e-01
-4.91854995e-01 1.28678823e+00 -9.21886206e-01 1.48501962e-01
1.78896770e-01 -1.56423464e-01 1.59937513e+00 -2.58863777e-01
1.48741737e-01 1.09029329e+00 2.71294326e-01 3.93432617e-01
-1.47571266e+00 -9.36371207e-01 7.32461929e-01 -1.94490641e-01
-9.50195670e-01 -1.46576717e-01 3.59890610e-01 -9.86052901e-02
6.19702458e-01 8.41560900e-01 6.03323221e-01 2.59903252e-01
-5.69226384e-01 9.31523144e-01 1.22069943e+00 -6.28770888e-01
6.09236538e-01 -3.91547531e-02 6.64769351e-01 8.30179513e-01
2.25737751e-01 1.81247488e-01 -4.21934187e-01 -5.14597118e-01
5.42307198e-01 3.11626971e-01 7.82271754e-03 -6.22517802e-02
-7.32770860e-01 1.10597503e+00 5.38889945e-01 -8.86509344e-02
-1.71357289e-01 8.21141362e-01 1.39261618e-01 5.83598614e-01
4.79708165e-01 1.84857935e-01 -6.73244417e-01 -1.27210885e-01
-8.51538062e-01 4.12799239e-01 7.39843130e-01 1.10752594e+00
8.59376967e-01 -3.50250721e-01 -3.24347079e-01 8.10937345e-01
8.90140906e-02 8.63053024e-01 3.51812467e-02 -1.34164584e+00
9.33489382e-01 4.57837433e-01 6.84234679e-01 -3.94025266e-01
-2.30789959e-01 -6.36831462e-01 -6.43094897e-01 -3.84276696e-02
8.22262287e-01 -2.66550481e-01 -7.71032035e-01 2.10589170e+00
4.78093863e-01 -3.90060037e-01 -4.51658696e-01 7.39442348e-01
4.84913029e-03 4.48186845e-01 -2.71111012e-01 -6.69218063e-01
1.49172544e+00 -9.54413354e-01 -1.54757112e-01 -5.32154799e-01
7.86012173e-01 -3.82806122e-01 1.25042498e+00 6.31362617e-01
-1.58760047e+00 9.88341421e-02 -8.03331971e-01 2.34625950e-01
-1.55182198e-01 -2.81170845e-01 1.23366070e+00 1.06376517e+00
-7.29140460e-01 4.48972523e-01 -3.94857466e-01 4.37555552e-01
5.39995313e-01 8.24378133e-01 1.64467636e-02 -2.92929649e-01
-8.16327989e-01 7.88183883e-02 1.59017816e-02 2.94013154e-02
-6.70367897e-01 -5.48314691e-01 -3.92186344e-01 2.41773218e-01
6.94838345e-01 -4.29741114e-01 1.30084932e+00 -9.22038794e-01
-1.30554557e+00 4.59709197e-01 -1.77060649e-01 -5.71046352e-01
5.46000659e-01 -1.54971510e-01 1.21318959e-01 -2.49612816e-02
-8.17498714e-02 2.36110061e-01 6.42645419e-01 -7.04522908e-01
-1.00602973e+00 -7.13851452e-01 5.55098355e-01 5.38421795e-02
-3.83653313e-01 -3.44060622e-02 -2.59144962e-01 -4.14830923e-01
3.60579163e-01 -1.29698789e+00 -5.84189415e-01 -2.64847666e-01
-1.37336925e-01 -3.16474289e-01 -2.03685328e-01 -1.63081765e-01
1.33200324e+00 -2.02439356e+00 -1.40253708e-01 6.69679463e-01
2.49244109e-01 -2.53934741e-01 1.12324674e-02 1.03457727e-01
2.10920811e-01 2.07665175e-01 -1.39799669e-01 -9.98032317e-02
3.09576720e-01 -6.56515220e-03 -5.25022388e-01 4.69352394e-01
-7.13666677e-01 7.43468165e-01 -6.35294020e-01 -7.09581226e-02
-2.27378741e-01 -1.59449100e-01 -8.79822314e-01 -4.20238264e-02
-6.58461750e-01 -1.45770773e-01 -7.84999311e-01 4.60532159e-01
6.95675552e-01 -7.57997572e-01 2.81795830e-01 3.58746469e-01
3.37988257e-01 3.89644772e-01 -1.45874763e+00 1.51259911e+00
-4.86069024e-01 -1.22707024e-01 1.50022030e-01 -1.11400974e+00
3.96799892e-01 -9.34057683e-02 5.60859501e-01 -7.94483125e-01
7.25963712e-02 4.42495286e-01 -2.01868519e-01 1.16070405e-01
1.98877707e-01 -2.97791123e-01 -4.33171540e-01 1.00407493e+00
-3.66900474e-01 2.55217761e-01 2.74497092e-01 2.98178107e-01
1.26004922e+00 -3.34530950e-01 -2.22468898e-01 -7.13128597e-02
8.56141299e-02 -1.30817696e-01 5.32613218e-01 1.44073486e+00
-4.96746935e-02 -2.86295991e-02 7.47083426e-01 -7.19163120e-01
-7.92170703e-01 -9.30760384e-01 2.85130516e-02 1.84520781e+00
2.42789954e-01 -2.98560634e-02 -8.43962848e-01 -8.18243206e-01
3.67892951e-01 5.86314797e-01 -8.55163395e-01 2.48006344e-01
-5.04178584e-01 -9.99311566e-01 3.15350384e-01 4.99532104e-01
2.17914343e-01 -6.08753741e-01 -4.17219698e-01 2.95441478e-01
-4.75811064e-02 -6.79132104e-01 -8.57156575e-01 6.82168722e-01
-1.04270661e+00 -7.27991521e-01 -5.49442291e-01 -4.56227720e-01
6.90638900e-01 2.19758868e-01 9.97148037e-01 -8.87326300e-02
-1.15565479e-01 1.90948933e-01 -2.05115736e-01 -3.55131149e-01
1.77972108e-01 -3.46978270e-02 -9.97135118e-02 -1.97557241e-01
4.09773409e-01 -4.99315023e-01 -1.14282811e+00 1.54333621e-01
-6.57213271e-01 -2.21734002e-01 6.28639996e-01 9.30400312e-01
8.86620045e-01 -2.46667042e-01 6.06062353e-01 -1.21853554e+00
2.86474347e-01 -3.14180642e-01 -1.19399786e+00 4.25126851e-01
-9.31317687e-01 2.96424687e-01 7.23360717e-01 -6.89883709e-01
-6.05513096e-01 9.48980823e-02 8.74070004e-02 -3.95093784e-02
4.12725598e-01 3.57190967e-01 3.36095363e-01 -5.67266606e-02
7.56424427e-01 2.41120175e-01 -1.73094049e-01 -4.22905385e-01
5.98770797e-01 5.99042058e-01 8.68112594e-02 -1.07781041e+00
3.35708082e-01 3.71794403e-01 1.20657831e-01 -1.05435878e-01
-1.28700495e+00 -1.04441553e-01 3.68941724e-01 2.16026112e-01
5.82556129e-02 -8.06591690e-01 -1.76260436e+00 -1.55086890e-01
-4.47977185e-01 -5.73128939e-01 -5.17938077e-01 5.04791796e-01
-8.14692378e-01 2.00406641e-01 -7.73555338e-01 -1.33794332e+00
-6.51722312e-01 -1.00903308e+00 7.01843202e-01 1.98241919e-01
1.72541440e-01 -3.45397830e-01 -2.98894554e-01 9.04689789e-01
2.59295136e-01 -1.90680102e-01 1.21476698e+00 -6.23595595e-01
-9.11646664e-01 -3.52673054e-01 -1.94195643e-01 9.54557657e-02
-3.19713861e-01 -8.28260720e-01 -6.49593174e-01 -5.71273863e-01
-2.02797040e-01 -6.21988595e-01 8.83124828e-01 6.14600658e-01
1.62968361e+00 -8.21112335e-01 -2.38737494e-01 3.99946123e-01
1.26542747e+00 2.86086380e-01 2.63673276e-01 1.38957784e-01
1.32720068e-01 1.81479916e-01 7.56309688e-01 6.63962960e-01
7.92119950e-02 4.72445250e-01 6.32853985e-01 1.32908300e-01
4.57647711e-01 -5.61543778e-02 2.09404856e-01 2.35265136e-01
1.10968031e-01 -1.75123423e-01 -4.76630747e-01 5.77767968e-01
-1.79529941e+00 -7.03379631e-01 2.24802062e-01 2.78847718e+00
1.18049085e+00 5.68014503e-01 4.72594321e-01 -1.47762755e-02
5.12669325e-01 -4.36738729e-02 -1.01323390e+00 -7.22668648e-01
1.76090404e-01 5.92628300e-01 1.04953599e+00 6.88533604e-01
-7.95189977e-01 6.92988098e-01 5.82973146e+00 1.31351542e+00
-7.54052639e-01 1.17382735e-01 1.17145836e+00 -1.04215991e+00
-4.81750965e-01 1.32058650e-01 -9.16029453e-01 6.86413050e-01
1.09431124e+00 1.93168316e-02 1.19006193e+00 1.09859180e+00
-2.82938898e-01 -2.56791621e-01 -1.38451481e+00 9.63380039e-01
-4.20258582e-01 -1.32512045e+00 -2.29113415e-01 4.72182065e-01
8.30470622e-01 3.27851176e-02 4.89786178e-01 3.62710327e-01
1.05015767e+00 -1.07811594e+00 7.28535414e-01 -1.12738147e-01
9.99928176e-01 -1.09460986e+00 2.28051364e-01 6.52540505e-01
-8.82357657e-01 -6.55652702e-01 -3.65692228e-01 -1.06946133e-01
-2.12082446e-01 7.85910428e-01 -5.09637713e-01 2.41294920e-01
9.38938856e-01 -3.18583339e-01 1.92788735e-01 6.53003097e-01
2.91544557e-01 6.72886789e-01 -1.00242209e+00 -4.91738617e-01
4.39805627e-01 -3.56992781e-01 1.92573611e-02 8.13560605e-01
2.68906146e-01 5.84155321e-01 2.16890469e-01 3.63933891e-01
-5.85356176e-01 2.81560749e-01 -1.30177680e-02 -1.79416277e-02
6.82542205e-01 9.15609598e-01 -6.63276136e-01 -3.26358974e-01
-1.70654431e-01 5.83522975e-01 7.40923107e-01 2.94364661e-01
-1.02254343e+00 -3.69481564e-01 6.04921520e-01 -1.74786672e-02
7.13303089e-01 1.76111415e-01 -3.62143308e-01 -6.47796690e-01
3.11048597e-01 -7.37207294e-01 8.16924274e-01 -1.11908324e-01
-1.21271074e+00 2.69877702e-01 -2.83357382e-01 -7.14294612e-01
-7.97337219e-02 -5.27808905e-01 5.65561093e-02 7.87045658e-01
-1.14585328e+00 -8.51354599e-01 3.07490140e-01 5.91044307e-01
1.68506950e-01 1.28799841e-01 7.48654425e-01 2.63215303e-01
-3.44045371e-01 1.24307966e+00 6.21611714e-01 -1.31790787e-01
2.38446072e-01 -1.26352465e+00 8.15980434e-02 2.59306937e-01
2.49869302e-01 6.45745873e-01 6.39703393e-01 -1.62119165e-01
-1.55133355e+00 -8.09227824e-01 4.69056457e-01 -3.12959045e-01
4.91032481e-01 -4.25378829e-01 -2.44663894e-01 9.49266136e-01
-4.72398192e-01 4.04724389e-01 1.08038521e+00 6.07803106e-01
-7.45462179e-01 -7.12162256e-01 -1.34688938e+00 6.02025032e-01
1.32151175e+00 -3.44310015e-01 1.20996766e-01 6.95556223e-01
8.11878562e-01 -5.57139933e-01 -9.78166938e-01 2.22762331e-01
9.43050504e-01 -7.74697661e-01 8.07433128e-01 -7.86389172e-01
1.20847926e-01 2.33504429e-01 -4.06220764e-01 -7.75843322e-01
-3.21447521e-01 -9.08767939e-01 -5.87010860e-01 7.88077772e-01
9.60706651e-01 -8.55305016e-01 1.25066185e+00 1.00398779e+00
4.71828818e-01 -1.16854298e+00 -1.36826158e+00 -5.59373081e-01
5.67284703e-01 -4.36456650e-01 8.51175964e-01 5.45830369e-01
1.24114923e-01 1.17882505e-01 -4.07576710e-01 -5.93133084e-02
7.11650133e-01 8.07306528e-01 5.01618028e-01 -9.61732924e-01
-1.09392285e+00 -4.52485204e-01 3.33471924e-01 -1.53000033e+00
-2.67491311e-01 -8.10848057e-01 -1.76051259e-01 -1.03954256e+00
6.54519439e-01 -1.17463803e+00 -8.51699412e-01 5.80621362e-01
-1.93806887e-02 9.27842315e-03 1.19810484e-01 -7.15079904e-02
-9.28098440e-01 1.38523087e-01 1.00243986e+00 -2.24779040e-01
-1.42346352e-01 4.14084792e-01 -1.28865206e+00 4.73871678e-01
4.90393788e-01 -7.57875860e-01 -3.21500331e-01 -4.95881677e-01
8.33811343e-01 7.22349644e-01 -5.23664989e-02 -3.92380416e-01
1.75896510e-01 -4.96119320e-01 2.19440803e-01 -5.34553587e-01
4.02289122e-01 -8.66839886e-01 -1.17522776e-01 6.03932261e-01
-9.37014043e-01 -2.53438979e-01 -9.33129713e-02 9.13966775e-01
4.54730004e-01 -1.85196444e-01 6.21587276e-01 -2.37678766e-01
4.48132366e-01 5.00049353e-01 1.80573389e-01 1.46510854e-01
1.10310984e+00 4.67304047e-03 -6.07508361e-01 -4.42820460e-01
-5.65396547e-01 5.18388629e-01 1.03537738e-01 -2.45834842e-01
1.20087996e-01 -1.15308535e+00 -4.39273506e-01 -4.93159480e-02
-1.15164980e-01 2.57477343e-01 4.18067276e-01 6.91245854e-01
-1.18138917e-01 4.50408489e-01 5.35512626e-01 -4.35222030e-01
-9.58823919e-01 8.42682064e-01 2.34905258e-01 -6.40749633e-01
-2.00988173e-01 1.35058451e+00 5.11194319e-02 -1.87316492e-01
6.02925241e-01 -2.63414681e-01 5.12970507e-01 -1.91245660e-01
6.73531413e-01 5.26244044e-01 -5.49738249e-03 1.29362449e-01
-2.17623547e-01 2.93605506e-01 -6.86187208e-01 -4.32883799e-01
1.34173143e+00 -3.27133238e-02 -7.87342116e-02 7.90223405e-02
1.04251766e+00 1.28468692e-01 -1.31460893e+00 -4.93942380e-01
-8.95301998e-02 -5.01171649e-01 -6.32469729e-02 -1.04465210e+00
-1.02825058e+00 6.19375527e-01 8.46408725e-01 5.22927344e-01
1.06859171e+00 2.63908088e-01 8.68090570e-01 7.12683916e-01
9.62344646e-01 -1.27985919e+00 -1.30331472e-01 2.48446077e-01
3.46145421e-01 -1.07169318e+00 6.90899044e-02 -1.37661740e-01
-2.99352378e-01 6.10601008e-01 2.43970484e-01 -1.81471884e-01
5.80693722e-01 6.34436458e-02 -5.48535943e-01 -3.24661881e-02
-8.36333096e-01 6.79185539e-02 -1.06789179e-01 -2.42682904e-01
2.70326376e-01 4.69424605e-01 -4.33628768e-01 9.75732446e-01
-3.19111139e-01 -7.90014938e-02 1.15515932e-01 8.15214813e-01
-8.10559809e-01 -1.17210221e+00 -2.66518503e-01 1.03261983e+00
-9.92017388e-01 -1.17301263e-01 -3.91255543e-02 1.48151919e-01
-2.28900582e-01 1.16348135e+00 3.82263549e-02 -8.16008672e-02
1.64011151e-01 4.65215929e-02 6.86267555e-01 -3.28969985e-01
-7.59603560e-01 2.62147456e-01 -1.28297089e-02 -5.93588591e-01
-6.98592979e-03 -4.67120409e-01 -1.26651549e+00 -6.91106617e-01
-5.28048635e-01 6.22431099e-01 6.58729017e-01 8.73534083e-01
3.12176615e-01 7.24125728e-02 9.83006418e-01 -2.09694535e-01
-1.20105684e+00 -7.40944743e-01 -9.36321378e-01 2.15522677e-01
3.66629601e-01 -4.33487833e-01 -4.01368260e-01 -5.33910573e-01] | [4.758172035217285, 3.5068204402923584] |
ab5181f2-ce40-4b8b-bf76-80b799cb8564 | communication-robust-multi-agent-learning-by | 2305.05116 | null | https://arxiv.org/abs/2305.05116v1 | https://arxiv.org/pdf/2305.05116v1.pdf | Communication-Robust Multi-Agent Learning by Adaptable Auxiliary Multi-Agent Adversary Generation | Communication can promote coordination in cooperative Multi-Agent Reinforcement Learning (MARL). Nowadays, existing works mainly focus on improving the communication efficiency of agents, neglecting that real-world communication is much more challenging as there may exist noise or potential attackers. Thus the robustness of the communication-based policies becomes an emergent and severe issue that needs more exploration. In this paper, we posit that the ego system trained with auxiliary adversaries may handle this limitation and propose an adaptable method of Multi-Agent Auxiliary Adversaries Generation for robust Communication, dubbed MA3C, to obtain a robust communication-based policy. In specific, we introduce a novel message-attacking approach that models the learning of the auxiliary attacker as a cooperative problem under a shared goal to minimize the coordination ability of the ego system, with which every information channel may suffer from distinct message attacks. Furthermore, as naive adversarial training may impede the generalization ability of the ego system, we design an attacker population generation approach based on evolutionary learning. Finally, the ego system is paired with an attacker population and then alternatively trained against the continuously evolving attackers to improve its robustness, meaning that both the ego system and the attackers are adaptable. Extensive experiments on multiple benchmarks indicate that our proposed MA3C provides comparable or better robustness and generalization ability than other baselines. | ['Yang Yu', 'Zhongzhang Zhang', 'Feng Chen', 'Lei Yuan'] | 2023-05-09 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-2.96765342e-02 1.60752043e-01 1.41430482e-01 6.41951799e-01
-4.38387156e-01 -7.45959401e-01 6.59741998e-01 2.54197586e-02
-3.55922729e-01 8.54840517e-01 -1.25231862e-01 -2.84565210e-01
-1.76169500e-01 -1.02445161e+00 -7.54949152e-01 -1.28206146e+00
-4.39338595e-01 1.36428699e-01 1.56352788e-01 -7.63738751e-01
9.80948880e-02 1.02105081e-01 -9.58307564e-01 -3.34448010e-01
1.16761363e+00 6.79520190e-01 -3.78684327e-02 6.48194134e-01
3.44150037e-01 7.84726262e-01 -1.33157706e+00 -2.73504496e-01
5.53044081e-01 -5.27325392e-01 -4.33094531e-01 2.94677150e-02
-5.91686666e-01 -3.35998386e-01 -6.08436763e-01 1.32367694e+00
7.17250526e-01 1.27106503e-01 4.45316434e-01 -1.77128851e+00
-6.29409373e-01 1.17837906e+00 -6.13915086e-01 -2.45913342e-02
2.92582870e-01 5.74111104e-01 5.83254635e-01 1.55720711e-01
2.32443199e-01 1.47894585e+00 2.36116007e-01 9.87511039e-01
-1.16337955e+00 -9.99087393e-01 5.56479990e-01 8.80557895e-02
-1.14642453e+00 -3.49859633e-02 9.06029701e-01 -6.54191673e-02
3.04402500e-01 1.67064756e-01 7.13645160e-01 1.61914945e+00
3.69082242e-01 6.79487169e-01 1.13535786e+00 -1.08984821e-01
5.35521448e-01 2.25008741e-01 -4.96282071e-01 5.17565548e-01
2.52893955e-01 7.02813387e-01 -7.48877227e-02 -5.51766515e-01
5.13869286e-01 -1.90881506e-01 -4.00208831e-01 -2.08603233e-01
-1.03018200e+00 8.71723413e-01 4.99648422e-01 8.82974546e-03
-4.89506155e-01 1.23105668e-01 4.12106425e-01 9.69251871e-01
1.20123543e-01 7.08582342e-01 -3.63647789e-01 3.00745550e-03
1.83662847e-01 1.69690028e-01 9.35571849e-01 5.91056645e-01
4.79620546e-01 5.46034992e-01 1.55382201e-01 3.48541498e-01
2.95057148e-01 4.68025565e-01 5.02724826e-01 -7.49163449e-01
4.58061457e-01 8.61106455e-01 8.16905499e-03 -1.13678610e+00
-3.36639524e-01 -6.95982993e-01 -9.93185401e-01 4.86604661e-01
1.33801967e-01 -9.17188227e-01 -2.74017781e-01 2.01948619e+00
5.24462223e-01 4.74235922e-01 6.75330997e-01 6.09876692e-01
3.27469349e-01 6.35484874e-01 -6.35541156e-02 -5.50775409e-01
7.75733352e-01 -8.50977540e-01 -4.94755238e-01 -3.79900187e-02
3.89750719e-01 -4.14653689e-01 5.75463176e-01 1.61914036e-01
-9.88304734e-01 -1.80545449e-01 -1.23029208e+00 1.31391418e+00
-3.95087391e-01 -5.71626186e-01 4.07780170e-01 9.18948829e-01
-6.86562002e-01 3.10939252e-01 -7.22630084e-01 -3.26935947e-02
2.68608838e-01 6.98262751e-01 -3.50687616e-02 3.60662192e-01
-1.42349827e+00 7.13635862e-01 5.13784766e-01 -1.92101777e-01
-1.47720134e+00 -4.58934754e-01 -5.86156368e-01 -1.21564999e-01
9.02162433e-01 -7.26813018e-01 8.35860372e-01 -1.16418576e+00
-2.00409675e+00 -1.26436383e-01 8.32361937e-01 -5.55606544e-01
6.85990810e-01 3.65301847e-01 -3.46164703e-01 4.45600972e-02
-1.84890583e-01 1.82291269e-01 1.12812459e+00 -1.72628903e+00
-6.71994746e-01 -7.12393373e-02 7.68974781e-01 4.80711579e-01
-7.22797871e-01 -1.82968348e-01 2.79558212e-01 -8.44482124e-01
-5.05188286e-01 -1.23333585e+00 -5.83365381e-01 -5.79772890e-01
-5.07954836e-01 -2.36785978e-01 1.14730239e+00 -1.30870745e-01
1.07578862e+00 -2.00123048e+00 5.84203005e-01 3.98163140e-01
2.55528301e-01 4.76075679e-01 -6.25325203e-01 7.05742538e-01
1.41806766e-01 2.94288903e-01 -7.63974264e-02 3.69193777e-02
1.48413461e-02 3.00095588e-01 -2.42001444e-01 5.13621986e-01
5.74476160e-02 7.24920332e-01 -1.06982744e+00 -1.02350518e-01
-7.95216337e-02 3.20546418e-01 -5.56986153e-01 5.42428613e-01
-2.08621979e-01 9.83070135e-01 -9.88471150e-01 4.90182310e-01
5.66815794e-01 3.61852869e-02 2.56402493e-01 3.90007377e-01
9.10464004e-02 -4.04572159e-01 -1.23814547e+00 9.09360170e-01
-4.49217439e-01 -5.56133837e-02 4.24280643e-01 -1.28396595e+00
9.02291775e-01 3.88628185e-01 5.74102104e-01 -5.21831453e-01
4.36184376e-01 4.77424487e-02 6.62579119e-01 -3.25655490e-01
9.10278484e-02 3.02999735e-01 -2.77217060e-01 7.14361906e-01
-1.92771316e-01 -3.88494469e-02 -3.55786495e-02 3.53165537e-01
1.26514721e+00 -2.85774320e-01 1.38502359e-01 -7.17180222e-02
9.14399087e-01 -2.47306973e-01 7.20636487e-01 1.05677247e+00
-4.47065532e-01 -2.56179959e-01 6.73118591e-01 -1.58342838e-01
-7.76352227e-01 -8.52727473e-01 5.18437386e-01 1.05280566e+00
6.38070047e-01 -2.45651975e-01 -8.19893777e-01 -1.07793760e+00
-2.19485294e-02 3.84104848e-01 -7.20265806e-01 -8.02372038e-01
-7.16609180e-01 -1.06067169e+00 7.37417340e-01 -8.64539593e-02
7.79186249e-01 -1.06247830e+00 -3.86519313e-01 4.22754467e-01
4.83054593e-02 -8.43500555e-01 -4.74470735e-01 -8.44305679e-02
-1.28516689e-01 -1.43973136e+00 -4.99532819e-01 -6.78220868e-01
7.01450586e-01 1.25967056e-01 4.37069118e-01 5.00246882e-01
5.63278273e-02 6.99492812e-01 -6.59736633e-01 -5.15223026e-01
-1.21145248e+00 9.50367898e-02 4.75569457e-01 1.62725583e-01
-2.20288545e-01 -6.74997985e-01 -5.42063475e-01 4.01724577e-01
-9.30102170e-01 -4.78638887e-01 6.43102705e-01 1.11891460e+00
-2.36620665e-01 5.95956028e-01 1.22945428e+00 -3.69142026e-01
1.21897554e+00 -7.41840541e-01 -7.83474028e-01 3.06202769e-01
-4.57136929e-01 -5.69619238e-02 1.43563128e+00 -1.12068558e+00
-8.69102836e-01 -1.40732691e-01 9.37497020e-02 -2.60106117e-01
3.52309272e-02 1.74673319e-01 -3.99297148e-01 -6.90607071e-01
6.20160758e-01 5.45360208e-01 4.43509400e-01 2.38483012e-01
2.60161549e-01 6.37574375e-01 -3.39889079e-02 -7.48910666e-01
1.56448817e+00 2.70267606e-01 1.76271826e-01 -6.54500186e-01
-1.01662785e-01 3.41637462e-01 6.75905049e-02 -3.92112315e-01
3.98153961e-01 -7.94606268e-01 -1.33009148e+00 8.26809347e-01
-1.00890696e+00 -3.23446244e-01 -1.77873835e-01 3.20157319e-01
-4.53713447e-01 4.49962437e-01 -4.27677691e-01 -8.49536300e-01
-3.09337020e-01 -1.53355038e+00 3.89206707e-01 5.32825351e-01
3.76378715e-01 -1.13790083e+00 1.85967594e-01 7.23332837e-02
6.92377806e-01 4.84777004e-01 8.34767580e-01 -9.24821913e-01
-6.78937852e-01 -1.33894105e-02 4.38983679e-01 2.63565451e-01
1.84232444e-01 -1.11442327e-01 -3.55653852e-01 -8.50817025e-01
1.45555288e-01 -5.61068654e-01 3.14329058e-01 -1.11715212e-01
9.00490344e-01 -8.94008934e-01 -3.35644364e-01 4.25309092e-01
1.07555723e+00 5.79009235e-01 3.48977983e-01 6.77969933e-01
3.66517931e-01 6.13727033e-01 3.78695637e-01 7.50449300e-01
5.15264869e-01 4.01093721e-01 9.58766341e-01 8.03158656e-02
4.50055897e-01 -2.21139818e-01 8.61119807e-01 6.52688563e-01
-3.45418788e-03 -3.79472286e-01 -3.46678913e-01 5.60162216e-02
-1.90343153e+00 -9.39884067e-01 5.41348159e-01 2.07149696e+00
7.41106987e-01 -8.24700517e-04 4.17465538e-01 1.96694192e-02
8.42345834e-01 4.92786430e-02 -8.27157080e-01 -1.79653496e-01
-3.88843179e-01 -2.61967868e-01 3.90359223e-01 3.43533546e-01
-9.93836343e-01 8.42149615e-01 5.46605968e+00 9.54831064e-01
-1.10817897e+00 -2.31333193e-03 5.69192171e-01 6.81134686e-02
-1.23140812e-01 4.25486006e-02 -4.99256581e-01 7.25988507e-01
6.03216887e-01 -5.50054193e-01 9.31665421e-01 6.20212257e-01
3.25364947e-01 2.53334939e-01 -6.64919913e-01 4.85763311e-01
-1.29530907e-01 -9.06712890e-01 4.39145900e-02 3.27071518e-01
1.00238717e+00 -3.23373824e-01 3.26516300e-01 5.98472834e-01
8.65284383e-01 -7.09862649e-01 3.35236073e-01 1.52265042e-01
1.17362618e-01 -1.16062927e+00 6.59340501e-01 6.74940050e-01
-1.07962847e+00 -7.15582669e-01 -2.07550362e-01 1.07562765e-02
-1.57026768e-01 -3.62796873e-01 -3.52912933e-01 7.70745516e-01
3.42401564e-01 2.57636279e-01 -5.77203929e-01 7.37692833e-01
-2.69965678e-01 4.16534126e-01 -1.54045492e-01 -4.13580626e-01
3.97053927e-01 -2.95350015e-01 1.12827408e+00 4.66496259e-01
9.72623155e-02 1.42219707e-01 7.90299118e-01 5.72172046e-01
-7.23807216e-02 6.65038377e-02 -7.94296324e-01 4.58916230e-03
8.00575674e-01 1.28037512e+00 -3.87484580e-01 -1.17576443e-01
-2.00086236e-01 7.15661347e-01 3.98496479e-01 5.53075612e-01
-1.08503306e+00 -3.80184233e-01 6.96089208e-01 -5.59847116e-01
-1.47137176e-02 3.57537828e-02 3.13495427e-01 -9.99505162e-01
-2.26020157e-01 -1.52336442e+00 3.50566834e-01 5.64535968e-02
-1.47408736e+00 7.51288891e-01 -3.97443801e-01 -1.35373139e+00
-2.67981201e-01 -1.50543928e-01 -1.06234479e+00 2.66345322e-01
-1.39991450e+00 -1.11373365e+00 -8.46020579e-02 9.31808770e-01
3.05692255e-01 -8.79273057e-01 6.63487673e-01 -1.24128014e-02
-1.06007588e+00 1.00557435e+00 2.05265462e-01 1.00782767e-01
3.58999670e-01 -1.07010210e+00 -1.77277744e-01 8.40739191e-01
-3.86606932e-01 5.77090025e-01 7.00983942e-01 -5.36686659e-01
-1.83773589e+00 -1.28361046e+00 -4.35222685e-01 -1.33979805e-02
1.12810075e+00 -2.69920141e-01 -5.88025033e-01 3.78837943e-01
5.65176189e-01 -2.40930408e-01 4.34563607e-01 -4.33219701e-01
-5.21572344e-02 -1.44321024e-01 -1.23458183e+00 1.13621211e+00
8.24908733e-01 -1.97316468e-01 -2.33429179e-01 4.19350326e-01
9.52974141e-01 -1.32552072e-01 -7.54860222e-01 2.49409616e-01
9.63906422e-02 -5.70268929e-01 8.73531640e-01 -6.79133356e-01
-6.13153018e-02 -3.04778725e-01 6.99981526e-02 -2.00154972e+00
7.27333128e-02 -1.34349358e+00 -2.25836381e-01 1.15666938e+00
1.68720737e-01 -1.25227761e+00 6.60836339e-01 -1.73965800e-05
1.75738201e-01 -5.83069086e-01 -1.08050132e+00 -1.14006388e+00
4.92847890e-01 3.41704458e-01 6.87703669e-01 1.07657170e+00
2.90905863e-01 1.93876088e-01 -5.89852691e-01 6.12612963e-01
7.44473338e-01 -1.99647948e-01 1.09552360e+00 -7.75031149e-01
-5.26346326e-01 -6.10185444e-01 -7.92766586e-02 -8.00270140e-01
6.21293187e-01 -5.61019003e-01 -6.41558543e-02 -7.75948226e-01
-9.87551212e-02 -5.87007642e-01 -2.20671877e-01 2.67061263e-01
-3.85879815e-01 -2.26885334e-01 3.34776044e-01 8.59829858e-02
-7.42140770e-01 1.05561626e+00 1.31418669e+00 -3.99433464e-01
-4.99633282e-01 3.48969251e-01 -9.23277020e-01 6.36245131e-01
9.78092015e-01 -4.84816611e-01 -4.82237130e-01 -6.76012365e-04
-1.95088927e-02 2.95477718e-01 3.22558761e-01 -9.15846705e-01
4.82397705e-01 -4.44911182e-01 -2.21188888e-01 1.66235596e-01
8.08599964e-02 -9.73353386e-01 -5.55149801e-02 1.16104698e+00
-2.59475023e-01 1.45095006e-01 1.92991979e-02 8.81051362e-01
-2.96021048e-02 -5.04992604e-02 8.84392977e-01 -1.82383448e-01
-2.66152740e-01 4.30154949e-01 -7.99843013e-01 7.02780113e-02
1.51489329e+00 2.31488615e-01 -7.90362716e-01 -6.96171045e-01
-3.78517002e-01 8.44105661e-01 3.54522407e-01 5.32201886e-01
6.03926897e-01 -1.06480646e+00 -9.17244136e-01 7.06609488e-02
-1.77655250e-01 -4.22683418e-01 2.07802042e-01 5.75067461e-01
6.38680011e-02 -2.64273316e-01 -3.15612733e-01 -2.05351681e-01
-1.14028239e+00 7.96840549e-01 6.92084253e-01 -3.28372926e-01
-2.35437468e-01 5.05259931e-01 1.76889375e-01 -6.19688451e-01
3.82940143e-01 4.71255064e-01 -5.10037303e-01 -1.14439540e-01
3.64102006e-01 4.86856908e-01 -6.54259205e-01 -5.09461462e-01
-1.01325586e-01 3.56643051e-01 -1.06474437e-01 5.87539747e-02
1.21260774e+00 -2.80421942e-01 -9.05491784e-02 -2.76603371e-01
7.59945333e-01 7.92546570e-02 -1.30869877e+00 -2.09226936e-01
-3.74374479e-01 -2.41037369e-01 -2.65748948e-01 -5.65644085e-01
-1.22301710e+00 2.49538630e-01 4.91479754e-01 8.58004570e-01
1.21470618e+00 -3.37913901e-01 6.19414032e-01 5.41543603e-01
4.73624200e-01 -1.01557827e+00 6.96030855e-01 4.99849677e-01
7.67135561e-01 -1.31680417e+00 -2.87882000e-01 -1.95390865e-01
-5.71998596e-01 9.81117547e-01 1.13786709e+00 -3.58379006e-01
5.12133539e-01 3.34901959e-01 6.24184236e-02 1.50623187e-01
-8.82943094e-01 -6.69481680e-02 -3.98126006e-01 9.85321939e-01
-5.08312821e-01 -6.43652827e-02 -2.87567943e-01 4.78443593e-01
-7.70393610e-02 -7.36779988e-01 1.04708135e+00 8.62880528e-01
-3.62953335e-01 -1.42347980e+00 -5.19495428e-01 4.56226543e-02
-4.03072357e-01 2.30430245e-01 -4.60027575e-01 7.78575599e-01
1.92958787e-01 1.42706037e+00 -3.82943422e-01 -6.68767095e-01
2.73128688e-01 -6.60559177e-01 9.06151533e-02 -1.80163890e-01
-9.28303182e-01 -2.17507407e-03 -3.28533977e-01 -3.39753777e-01
-3.24504822e-01 -3.37515712e-01 -9.86292064e-01 -4.94425088e-01
-5.54933548e-01 5.23731530e-01 2.99468875e-01 7.72087634e-01
2.93401957e-01 7.78879821e-01 1.64827383e+00 -4.38784510e-01
-1.18773389e+00 -7.76337028e-01 -4.23654437e-01 2.47348770e-01
5.32289624e-01 -7.14963973e-01 -7.50647902e-01 -6.27145767e-01] | [3.8058900833129883, 2.336461305618286] |
c8bbdf48-85b2-446e-be12-d2d6467edc18 | dumb-a-benchmark-for-smart-evaluation-of | 2305.13026 | null | https://arxiv.org/abs/2305.13026v1 | https://arxiv.org/pdf/2305.13026v1.pdf | DUMB: A Benchmark for Smart Evaluation of Dutch Models | We introduce the Dutch Model Benchmark: DUMB. The benchmark includes a diverse set of datasets for low-, medium- and high-resource tasks. The total set of eight tasks include three tasks that were previously not available in Dutch. Instead of relying on a mean score across tasks, we propose Relative Error Reduction (RER), which compares the DUMB performance of models to a strong baseline which can be referred to in the future even when assessing different sets of models. Through a comparison of 14 pre-trained models (mono- and multi-lingual, of varying sizes), we assess the internal consistency of the benchmark tasks, as well as the factors that likely enable high performance. Our results indicate that current Dutch monolingual models under-perform and suggest training larger Dutch models with other architectures and pre-training objectives. At present, the highest performance is achieved by DeBERTaV3 (large), XLM-R (large) and mDeBERTaV3 (base). In addition to highlighting best strategies for training larger Dutch models, DUMB will foster further research on Dutch. A public leaderboard is available at https://dumbench.nl. | ['Malvina Nissim', 'Martijn Wieling', 'Wietse de Vries'] | 2023-05-22 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [-2.28841364e-01 1.08050473e-01 -1.30511343e-01 -4.81034428e-01
-1.25422657e+00 -6.51597023e-01 1.08481538e+00 -4.12364528e-02
-9.43109870e-01 8.37636650e-01 6.77988946e-01 -3.45736623e-01
1.05732551e-03 -2.05510259e-01 -7.77891636e-01 -1.71824589e-01
2.37409949e-01 8.01622689e-01 9.07411128e-02 -3.24989945e-01
-1.85227036e-01 9.43472311e-02 -1.14676416e+00 6.09815776e-01
8.47131073e-01 3.48133236e-01 4.71915752e-01 4.97554958e-01
2.49407560e-01 4.31895614e-01 -4.70777303e-01 -9.25988972e-01
3.16519141e-01 -8.95078629e-02 -8.33899736e-01 -3.75099152e-01
7.19210565e-01 -2.97818184e-01 -4.85496558e-02 5.08931816e-01
8.74486685e-01 1.75250657e-02 7.07705677e-01 -1.06145155e+00
-8.04797828e-01 8.72397304e-01 -9.20971483e-02 4.34301406e-01
4.51071501e-01 4.15022731e-01 1.12901378e+00 -1.16618574e+00
1.07200146e+00 1.48050690e+00 7.98497021e-01 5.12726068e-01
-1.31181312e+00 -7.83923686e-01 4.62946564e-01 2.09158927e-01
-1.22635150e+00 -1.08530319e+00 8.79074261e-02 -4.87562925e-01
1.75358534e+00 1.60491690e-01 3.58664155e-01 1.96127689e+00
1.32994905e-01 8.83381248e-01 1.23218381e+00 -5.37357450e-01
-3.56898338e-01 3.37308019e-01 2.76710540e-02 2.50979573e-01
3.91941130e-01 2.88221896e-01 -7.77934909e-01 4.59969081e-02
1.19004034e-01 -6.98572755e-01 -3.51562440e-01 -1.93306699e-01
-1.73277771e+00 6.35920823e-01 1.30346805e-01 5.05945861e-01
-2.50045627e-01 -1.48854256e-01 6.69727683e-01 5.84063768e-01
5.73525012e-01 6.23704612e-01 -1.10948229e+00 -4.53430742e-01
-9.34136987e-01 5.20434082e-01 8.52899909e-01 1.08033884e+00
4.55584943e-01 3.34493490e-03 -2.68823415e-01 1.22538877e+00
3.20315361e-01 5.37640810e-01 5.67533553e-01 -9.47349131e-01
1.00944221e+00 2.78956592e-01 7.95696601e-02 -3.64666283e-01
-5.33615530e-01 -2.36604095e-01 -4.06014293e-01 -4.72059771e-02
8.02923977e-01 -2.91538715e-01 -8.31509233e-01 1.86336911e+00
-8.64403844e-02 -4.76221234e-01 1.20106712e-01 7.20834136e-01
8.91693234e-01 5.99121273e-01 1.71124667e-01 4.50591221e-02
1.03543794e+00 -1.21876311e+00 -5.57555139e-01 -5.10930300e-01
1.11549282e+00 -1.19928288e+00 1.38432050e+00 5.80418885e-01
-1.22520876e+00 -6.94547176e-01 -8.49004507e-01 -3.39910269e-01
-6.73679650e-01 2.86581427e-01 3.65967959e-01 5.75239480e-01
-1.42217970e+00 3.48283380e-01 -8.84543836e-01 -9.23092365e-01
-1.25233993e-01 8.11252519e-02 -5.99560738e-01 -2.01194987e-01
-1.20127535e+00 1.49405396e+00 3.82077962e-01 1.64064676e-01
-9.79885817e-01 -8.91819477e-01 -9.17590916e-01 -4.83485520e-01
1.11820303e-01 -4.98029441e-01 1.53987992e+00 -8.41561675e-01
-1.45527816e+00 1.17829931e+00 -1.63300321e-01 -3.84076655e-01
9.44219887e-01 -5.12953579e-01 -5.76149642e-01 -5.33540666e-01
1.89853743e-01 1.00920355e+00 2.58677244e-01 -1.02340806e+00
-6.88457906e-01 -2.41209511e-02 9.40426514e-02 3.43061745e-01
-1.32528782e-01 4.11216289e-01 -4.84575152e-01 -6.61945820e-01
-4.24105346e-01 -9.46833551e-01 -2.76644249e-02 -5.84096611e-01
-3.23442608e-01 -3.24860841e-01 2.87557751e-01 -9.03093934e-01
1.19370127e+00 -1.99050379e+00 1.62285447e-01 -2.92342037e-01
-3.65180895e-02 2.82550961e-01 -6.90708280e-01 8.41783702e-01
-1.32787079e-01 4.05216694e-01 -5.24449609e-02 -3.89379472e-01
3.00641477e-01 1.85093120e-01 1.44568935e-01 4.07753736e-01
2.54688025e-01 1.05609894e+00 -7.11215258e-01 -2.64585674e-01
1.59338623e-01 3.71272326e-01 -6.61442220e-01 -7.55930226e-03
-1.24929518e-01 2.99498558e-01 4.65118773e-02 5.19819856e-01
5.65048218e-01 5.18562198e-02 1.40909106e-01 -1.34665385e-01
-1.84736818e-01 7.91111112e-01 -9.77533579e-01 1.91069019e+00
-7.62874305e-01 8.59817743e-01 1.82177097e-01 -4.96841967e-01
6.28604293e-01 4.02595907e-01 2.50071920e-02 -1.08818138e+00
-3.04728866e-01 7.22546101e-01 6.26614809e-01 -4.34244007e-01
5.41830719e-01 3.35839301e-01 -7.33053759e-02 2.05357537e-01
1.81852818e-01 -4.95216101e-02 6.83218241e-01 1.04695603e-01
1.06317866e+00 4.45669532e-01 3.54081869e-01 -6.56336963e-01
4.06869382e-01 5.99498339e-02 4.36287254e-01 6.80154741e-01
-3.48400652e-01 4.66836989e-01 3.06933999e-01 -3.69935989e-01
-1.24464297e+00 -1.10688210e+00 -5.72629809e-01 1.43854702e+00
-4.77580220e-01 -6.53941453e-01 -6.07670784e-01 -5.82405508e-01
1.22084245e-01 8.84596229e-01 -5.98641038e-01 2.97145754e-01
-7.40862966e-01 -7.31031239e-01 6.38459921e-01 4.25540209e-01
1.54423401e-01 -1.27785194e+00 -2.98493654e-01 2.99884886e-01
-2.28056267e-01 -1.29564488e+00 -4.91476774e-01 1.56150401e-01
-4.93392408e-01 -8.20812166e-01 -9.06814575e-01 -9.19785798e-01
4.06825058e-02 -3.49006318e-02 1.62815845e+00 -2.20421225e-01
2.58515000e-01 2.37135023e-01 -4.70476925e-01 -4.58877504e-01
-6.52216434e-01 6.99901104e-01 -3.41873942e-03 -6.86744213e-01
6.13396108e-01 -1.49631247e-01 -3.26957107e-01 3.41234922e-01
-3.11981529e-01 1.95477024e-01 8.07346225e-01 8.32529604e-01
2.80738354e-01 -7.61152208e-01 5.50058603e-01 -1.00343752e+00
6.61357462e-01 -4.42377061e-01 -3.83275449e-01 1.96561232e-01
-5.87723911e-01 -2.63096511e-01 4.20937955e-01 -3.15522969e-01
-1.10793650e+00 -4.47847366e-01 -3.15267354e-01 1.19939663e-01
-2.28154585e-01 5.44932067e-01 -1.44365281e-01 4.27810878e-01
7.40750253e-01 -1.30073994e-01 -3.34694713e-01 -6.96728706e-01
4.28497195e-01 7.73182750e-01 6.75489381e-02 -7.79564023e-01
2.89260417e-01 4.50077374e-03 -8.23747277e-01 -8.65428448e-01
-5.52102208e-01 -2.48766720e-01 -7.63528466e-01 1.73937932e-01
7.79371679e-01 -1.39052331e+00 -2.15062629e-02 4.65056211e-01
-1.27110195e+00 -1.02350187e+00 -1.36181721e-02 6.61450803e-01
-3.02085578e-01 -5.96381463e-02 -7.01135993e-01 -3.25225979e-01
-9.39684585e-02 -1.50315106e+00 1.24825680e+00 -6.12281024e-01
-7.57249832e-01 -1.28930473e+00 2.76412100e-01 4.62862819e-01
4.78102416e-01 1.97170153e-01 8.55521977e-01 -8.89889121e-01
-1.30762786e-01 1.32335559e-01 -2.04019561e-01 3.71389508e-01
-9.49582458e-02 2.00415570e-02 -9.76785481e-01 -5.64150393e-01
-4.90879714e-01 -7.95811594e-01 8.65887403e-01 3.92756373e-01
4.18358445e-01 -8.61699581e-02 -3.48575950e-01 4.67344463e-01
1.12880814e+00 -3.13120373e-02 3.55636090e-01 7.37408280e-01
5.99401057e-01 6.76105320e-01 4.23257619e-01 -5.96613484e-03
9.66699600e-01 8.16981912e-01 1.28973508e-02 -1.60982087e-01
-5.40824056e-01 -3.48101072e-02 7.82918990e-01 1.39641118e+00
-4.35388461e-02 -3.65026504e-01 -1.16322100e+00 8.71351838e-01
-1.52429962e+00 -6.42754078e-01 -2.68522203e-01 2.19336581e+00
9.14792597e-01 3.52931559e-01 3.03449631e-01 -2.95065999e-01
4.34208035e-01 2.62148291e-01 -3.85260642e-01 -9.39339161e-01
-3.56012285e-01 1.99037462e-01 2.68731982e-01 6.27677441e-01
-8.36925030e-01 1.27453506e+00 7.37407351e+00 8.37466180e-01
-8.66709352e-01 4.58764046e-01 5.18356800e-01 -3.76879066e-01
-3.16740960e-01 -7.46291727e-02 -9.66615915e-01 3.85164976e-01
1.43727982e+00 -2.14184985e-01 5.38549423e-01 5.05464315e-01
3.07755589e-01 2.64641196e-02 -1.56404221e+00 5.02873123e-01
-1.31206354e-02 -6.65747225e-01 8.08285624e-02 2.15943202e-01
8.58168364e-01 9.44769800e-01 -1.50518194e-02 8.52315247e-01
6.42893016e-01 -1.13641238e+00 1.15582585e+00 1.42065391e-01
8.65151882e-01 -3.58986437e-01 7.37892032e-01 2.22118109e-01
-1.02933002e+00 1.18699484e-01 -4.73099142e-01 -1.58800647e-01
1.53020501e-01 2.95328438e-01 -7.15836465e-01 6.01209700e-01
8.39100063e-01 9.64937031e-01 -8.13600123e-01 6.44509196e-01
-2.35246807e-01 6.94719374e-01 -3.25093418e-01 9.67216399e-03
5.05960166e-01 9.69803799e-03 2.91815758e-01 1.81658089e+00
2.90864497e-01 -5.76468587e-01 2.20388532e-01 4.97471958e-01
-1.29633531e-01 5.61768413e-01 -7.10207403e-01 3.57134193e-02
5.84236562e-01 1.18001699e+00 -9.92735848e-02 -2.02391237e-01
-1.00128865e+00 7.34859943e-01 7.62234867e-01 4.36883241e-01
-7.23978877e-01 -1.64014974e-03 6.58367872e-01 2.04413787e-01
4.01115082e-02 -4.02387649e-01 -6.25462085e-02 -1.06840432e+00
9.18782279e-02 -1.43401992e+00 3.24248224e-01 -8.30244064e-01
-1.32189786e+00 7.32745588e-01 1.59889653e-01 -9.54258084e-01
-4.36276317e-01 -9.34669256e-01 -7.19989389e-02 1.20556223e+00
-1.81312454e+00 -1.38572598e+00 -7.87932724e-02 4.07570034e-01
9.99607742e-01 -3.19617689e-01 9.03048337e-01 4.40293312e-01
-6.19321287e-01 8.69738698e-01 1.78589106e-01 7.38816485e-02
1.39859498e+00 -1.33095133e+00 8.63611341e-01 7.25939274e-01
1.83823720e-01 6.03916705e-01 5.24469018e-01 -5.67293108e-01
-1.03318298e+00 -8.90668094e-01 1.43737102e+00 -8.77393663e-01
9.78041530e-01 -5.82610965e-01 -6.56786501e-01 1.10032105e+00
8.27185512e-01 -4.53467667e-01 4.13798749e-01 5.87767422e-01
-3.63722444e-01 -1.21679038e-01 -5.76670170e-01 6.71300292e-01
1.39667439e+00 -3.14371079e-01 -7.08792925e-01 4.86493349e-01
6.22248292e-01 -5.81413984e-01 -1.06705439e+00 4.09790248e-01
7.77645886e-01 -1.11854994e+00 6.32182419e-01 -5.74994802e-01
5.81452966e-01 3.48155707e-01 -3.75665963e-01 -1.83531511e+00
-3.40358555e-01 -4.50656295e-01 3.71722698e-01 1.30486369e+00
1.16576660e+00 -8.24191153e-01 3.14389676e-01 2.35930011e-01
-4.15771604e-01 -5.67614317e-01 -1.09034264e+00 -9.88751948e-01
7.91733384e-01 -6.09652340e-01 3.95547062e-01 1.04296172e+00
-2.53523320e-01 3.42243403e-01 -1.66872323e-01 -2.65754759e-01
2.17148468e-01 -4.30935055e-01 9.71622109e-01 -1.07682133e+00
-1.36281386e-01 -7.04614580e-01 -3.09246741e-02 -7.40771294e-01
3.02679777e-01 -1.43171084e+00 -7.67984614e-02 -1.66267407e+00
1.55787513e-01 -3.98086160e-01 -2.04532802e-01 7.35637188e-01
-2.20178291e-01 1.15969874e-01 3.28472197e-01 3.19550633e-01
-5.18505037e-01 2.51765966e-01 1.29512465e+00 -8.80880840e-03
-1.20230995e-01 -4.13570315e-01 -6.45667315e-01 6.38290942e-01
7.41312623e-01 -3.55660528e-01 -2.45830834e-01 -1.24225330e+00
1.82023510e-01 -6.04688168e-01 3.55544575e-02 -8.72720897e-01
-2.07304195e-01 1.29647076e-01 2.82783598e-01 -2.26303726e-01
2.98014671e-01 -5.29228866e-01 9.81166437e-02 4.26143825e-01
-3.82870406e-01 6.76891863e-01 3.81390572e-01 -1.28553405e-01
-2.21343413e-01 -2.82735936e-02 5.24263859e-01 -3.83404791e-01
-6.87620640e-01 -1.76267531e-02 -2.73784757e-01 3.67428571e-01
6.70654655e-01 -9.89922583e-02 -5.67889214e-01 -1.42343730e-01
-6.34105563e-01 4.24256921e-01 3.67515475e-01 9.26137686e-01
9.23292413e-02 -1.33465827e+00 -1.34071922e+00 -5.06944507e-02
5.45478702e-01 -3.65540117e-01 -8.24710876e-02 9.44521368e-01
-5.29171228e-01 8.13578308e-01 -8.86750966e-02 -4.58358407e-01
-8.91359210e-01 1.82600647e-01 2.41025418e-01 -6.22926950e-01
-5.05036891e-01 6.20013177e-01 7.27044120e-02 -9.13259864e-01
1.31396487e-01 -3.58031839e-01 -8.15746263e-02 2.65418410e-01
2.33753830e-01 5.13799191e-01 4.64315802e-01 -8.63227606e-01
-5.88305116e-01 3.31353724e-01 -3.75196666e-01 -4.59513515e-01
1.27648795e+00 -1.06693938e-01 1.72076017e-01 7.27329433e-01
1.12146592e+00 1.93566516e-01 -9.85449672e-01 -1.54696941e-01
2.83235699e-01 -7.58495256e-02 -8.39208513e-02 -1.34401631e+00
-7.48302877e-01 7.87905693e-01 4.60979193e-01 -1.78456560e-01
7.27098584e-01 -7.70387352e-02 5.84111154e-01 5.08428276e-01
3.66694272e-01 -1.22978020e+00 -2.34116852e-01 7.69950807e-01
1.34745872e+00 -1.38385844e+00 -5.85445985e-02 -1.04888707e-01
-8.49570394e-01 7.76307762e-01 8.05416465e-01 1.04162015e-01
5.07580876e-01 4.91492331e-01 4.54433501e-01 1.42157704e-01
-1.13338172e+00 -3.59534293e-01 3.87493193e-01 5.94632566e-01
1.17786610e+00 6.30702078e-02 -7.95795143e-01 5.89506805e-01
-6.62068009e-01 -1.55887276e-01 3.91769737e-01 6.95428312e-01
-4.52817343e-02 -1.32819819e+00 -1.47666514e-01 3.80245715e-01
-5.76149821e-01 -5.65407872e-01 -4.40529585e-01 1.58369064e+00
2.21850082e-01 1.08081591e+00 5.62284887e-03 -2.22612843e-01
8.28594506e-01 1.56217173e-01 4.62193280e-01 -6.07959449e-01
-9.07064617e-01 2.22166210e-01 7.80467153e-01 -6.69091225e-01
-4.24412727e-01 -9.57257748e-01 -4.24445629e-01 -2.85513908e-01
-1.76738799e-02 -2.10400566e-01 6.16415262e-01 7.03453004e-01
2.48923421e-01 4.23426211e-01 -5.62089495e-02 -8.95426333e-01
-5.15007675e-01 -1.57676220e+00 -1.75063670e-01 5.19770682e-01
2.61593089e-02 -6.20229721e-01 -3.62831205e-01 -1.59804821e-02] | [10.935775756835938, 10.027222633361816] |
6c261f2e-1fae-4b32-a907-489f309ed9d0 | towards-a-robust-framework-for-nerf | 2305.18079 | null | https://arxiv.org/abs/2305.18079v3 | https://arxiv.org/pdf/2305.18079v3.pdf | Towards a Robust Framework for NeRF Evaluation | Neural Radiance Field (NeRF) research has attracted significant attention recently, with 3D modelling, virtual/augmented reality, and visual effects driving its application. While current NeRF implementations can produce high quality visual results, there is a conspicuous lack of reliable methods for evaluating them. Conventional image quality assessment methods and analytical metrics (e.g. PSNR, SSIM, LPIPS etc.) only provide approximate indicators of performance since they generalise the ability of the entire NeRF pipeline. Hence, in this paper, we propose a new test framework which isolates the neural rendering network from the NeRF pipeline and then performs a parametric evaluation by training and evaluating the NeRF on an explicit radiance field representation. We also introduce a configurable approach for generating representations specifically for evaluation purposes. This employs ray-casting to transform mesh models into explicit NeRF samples, as well as to "shade" these representations. Combining these two approaches, we demonstrate how different "tasks" (scenes with different visual effects or learning strategies) and types of networks (NeRFs and depth-wise implicit neural representations (INRs)) can be evaluated within this framework. Additionally, we propose a novel metric to measure task complexity of the framework which accounts for the visual parameters and the distribution of the spatial data. Our approach offers the potential to create a comparative objective evaluation framework for NeRF methods. | ['David R Bull', 'Nantheera Anantrasirichai', 'Adrian Azzarelli'] | 2023-05-29 | null | null | null | null | ['neural-rendering', 'image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 3.38566333e-01 -2.66900361e-01 4.52014536e-01 -3.91418040e-01
-6.80640817e-01 -5.33749342e-01 8.37933421e-01 1.80765659e-01
-3.31328183e-01 5.21806896e-01 9.67336297e-02 -3.98586303e-01
-2.97236979e-01 -9.84355569e-01 -6.75588727e-01 -4.90094423e-01
-1.46318719e-01 -7.74520338e-02 3.12941968e-01 -3.79091173e-01
4.53256905e-01 1.04220295e+00 -1.97601247e+00 3.50792527e-01
7.58834481e-01 1.05700064e+00 3.75817358e-01 9.78829145e-01
-5.98120801e-02 6.98142707e-01 -9.59985316e-01 -1.98483825e-01
2.30194643e-01 -2.35691980e-01 -5.79045892e-01 -4.58805293e-01
7.36674070e-01 -3.58360201e-01 -1.16402589e-01 7.60812581e-01
7.98335731e-01 4.16835129e-01 7.64661431e-01 -1.08402562e+00
-5.84408879e-01 -1.14213213e-01 -3.30453277e-01 2.69198060e-01
5.35750151e-01 2.69638151e-01 5.75446725e-01 -7.64871657e-01
5.90616405e-01 1.34847581e+00 8.78919959e-01 2.65961379e-01
-1.38879752e+00 -3.16921204e-01 -5.87989837e-02 3.08622164e-03
-1.17626011e+00 -3.20030987e-01 6.46277845e-01 -4.49898750e-01
1.02039289e+00 6.51651919e-01 5.53956151e-01 9.12225783e-01
2.78414011e-01 3.25339437e-01 1.56175029e+00 -5.13774753e-01
4.41058874e-01 3.03741455e-01 1.83561817e-02 4.21447545e-01
-4.75274883e-02 4.31185842e-01 -4.09122169e-01 -9.71939787e-02
1.13659441e+00 -4.03903574e-01 -5.22944808e-01 -3.67759526e-01
-8.92520905e-01 4.55065876e-01 6.31234229e-01 2.30502456e-01
-4.57201004e-01 4.63257045e-01 3.90199304e-01 2.54237056e-01
5.06269157e-01 4.82636154e-01 -2.49438614e-01 -3.11179116e-04
-9.91956413e-01 4.42130327e-01 5.55373967e-01 4.43544507e-01
7.69228041e-01 2.45388165e-01 -5.17533302e-01 9.98632431e-01
2.52990156e-01 2.36804441e-01 1.66852370e-01 -1.22028911e+00
-7.66586363e-02 3.06610167e-01 1.81483969e-01 -1.19734478e+00
-3.90609473e-01 -3.89636368e-01 -6.55619442e-01 1.16514206e+00
2.80341297e-01 1.94091886e-01 -9.59835589e-01 1.57179570e+00
2.79589504e-01 7.16317296e-02 -6.82522878e-02 9.46271956e-01
8.95524502e-01 8.43265653e-01 1.73690856e-01 3.11619103e-01
1.19807005e+00 -5.68920135e-01 -4.71275508e-01 2.28793025e-01
3.26569527e-01 -7.23068833e-01 1.41949391e+00 7.12511361e-01
-1.41960108e+00 -8.26526582e-01 -1.12559736e+00 -1.70774058e-01
-7.00223267e-01 -7.42796883e-02 4.65357721e-01 9.00822401e-01
-1.61331058e+00 9.64584887e-01 -4.56338644e-01 4.20109443e-02
2.46708691e-01 1.66738078e-01 -7.47188330e-02 -2.31404845e-02
-1.01394832e+00 1.05166948e+00 2.46228173e-01 2.24964336e-01
-9.48830009e-01 -8.59688044e-01 -7.30565310e-01 1.23183176e-01
-2.10968360e-01 -7.22717047e-01 1.08302820e+00 -1.21363449e+00
-1.54670489e+00 7.11187959e-01 2.28465900e-01 -1.76391795e-01
5.62350690e-01 5.84213948e-03 -4.20372844e-01 1.32499889e-01
-2.72388190e-01 8.27548265e-01 7.14056671e-01 -1.74593985e+00
-3.68884414e-01 -1.48944259e-01 5.09590507e-01 2.23868236e-01
-5.57226911e-02 1.15834244e-01 -1.99755684e-01 -7.37474203e-01
-1.67955562e-01 -2.74917662e-01 -2.48338267e-01 4.14009064e-01
-6.52510002e-02 7.94397891e-02 6.54310405e-01 -7.36039221e-01
8.97638023e-01 -2.22457242e+00 -1.86707944e-01 4.61322963e-01
2.31214896e-01 3.17941934e-01 -3.39332253e-01 1.63087174e-01
-3.28655422e-01 1.83520496e-01 -2.41615191e-01 -2.95713633e-01
1.02270357e-01 1.46620795e-01 -5.76032996e-02 3.38328093e-01
2.15069324e-01 6.89033687e-01 -7.97575653e-01 -1.11012600e-01
5.79157352e-01 1.30785346e+00 -4.26072955e-01 5.10316268e-02
-1.39192954e-01 2.66805977e-01 -4.84722555e-02 3.33892643e-01
9.07143056e-01 5.65617457e-02 -1.32898763e-01 -2.65805244e-01
-3.73838782e-01 1.60581440e-01 -1.39263475e+00 1.52923453e+00
-8.31498861e-01 9.58168328e-01 2.48886403e-02 -5.84632277e-01
1.01404834e+00 1.93783417e-01 -2.21741870e-02 -1.28001928e+00
2.23596841e-01 2.00881526e-01 -3.85731250e-01 -2.99147010e-01
7.85248518e-01 1.90327857e-02 3.64771932e-01 9.70460251e-02
-7.24254409e-03 -3.67807858e-02 -6.55456036e-02 -2.26915598e-01
1.01697743e+00 5.91233313e-01 9.47614461e-02 -2.70277590e-01
4.06612903e-01 -1.73169747e-01 -6.00058325e-02 7.48451173e-01
8.51671174e-02 9.19062138e-01 4.07199830e-01 -4.74020332e-01
-1.09280705e+00 -1.32748938e+00 -2.60147601e-01 1.06886864e+00
1.26347110e-01 -1.50338680e-01 -8.19393694e-01 -2.87728280e-01
-3.47263902e-01 8.84292483e-01 -7.30638444e-01 -5.05676353e-03
-5.07875144e-01 -5.16976655e-01 5.34772694e-01 4.18822467e-01
3.93101543e-01 -1.19971168e+00 -1.12457931e+00 -2.02278662e-02
1.52124166e-01 -7.36151814e-01 2.38034725e-01 2.21718773e-01
-9.96719480e-01 -8.06545973e-01 -9.43615615e-01 -3.70435029e-01
4.84908402e-01 3.06474984e-01 1.43685257e+00 1.03151754e-01
-4.14705455e-01 4.73278284e-01 -2.33594283e-01 -2.16595069e-01
-4.76940840e-01 -4.84953523e-01 -5.25253773e-01 -2.02622265e-01
-3.38027954e-01 -6.86249018e-01 -9.75529194e-01 2.19490170e-01
-1.20289075e+00 1.93470325e-02 5.45511127e-01 4.22785968e-01
7.10297644e-01 -3.97542641e-02 1.68238133e-01 -6.54663861e-01
8.92722189e-01 -2.54922152e-01 -6.03306890e-01 9.89953727e-02
-5.32306552e-01 -6.32649362e-02 5.62605321e-01 -3.14598352e-01
-1.20163608e+00 -2.75511473e-01 -4.27456528e-01 -4.70663548e-01
-5.17828763e-01 4.17869717e-01 -1.02421671e-01 -4.49799091e-01
1.16365981e+00 -6.65454343e-02 -3.18653613e-01 -5.05244255e-01
4.09960568e-01 2.66789824e-01 6.96534216e-01 -5.16003788e-01
6.84882283e-01 2.92218387e-01 9.90512595e-02 -7.56829321e-01
-1.31594628e-01 -2.04014838e-01 -2.05924943e-01 -6.24157965e-01
7.41754115e-01 -6.20508552e-01 -5.84281206e-01 2.77896374e-01
-1.39503229e+00 -4.87575442e-01 -5.00991762e-01 3.74740779e-01
-5.35856843e-01 2.05452323e-01 -3.36698979e-01 -1.08258593e+00
-1.82277977e-01 -1.20685828e+00 1.15940511e+00 1.55605569e-01
-1.37713179e-01 -9.84635413e-01 2.86687315e-01 -1.23955213e-01
6.89122379e-01 7.11547852e-01 9.50779736e-01 8.74394551e-02
-6.37827039e-01 -1.91282909e-02 -6.19893432e-01 6.29468977e-01
-3.38540554e-01 1.14169940e-01 -1.49094045e+00 -1.71594392e-03
-6.41143918e-02 -1.08329281e-01 6.19111836e-01 6.70659125e-01
1.26805079e+00 -9.06259790e-02 5.61679639e-02 7.06562579e-01
1.88647795e+00 2.60655791e-01 1.28900492e+00 4.98624891e-01
5.72139084e-01 8.51497710e-01 2.38768488e-01 3.02976876e-01
1.06128836e-02 7.97973871e-01 6.37027979e-01 -5.65258026e-01
-4.65181857e-01 4.18066196e-02 2.64236897e-01 2.47861356e-01
-5.68765461e-01 -3.65694880e-01 -7.79312551e-01 1.93813100e-01
-1.34586668e+00 -8.79624367e-01 -3.76229554e-01 2.43033624e+00
4.22013313e-01 1.12184668e-02 -6.82456642e-02 4.45483685e-01
5.49699306e-01 1.63698494e-01 -1.45022854e-01 -9.35931981e-01
-1.04020767e-01 7.57316649e-01 4.30891424e-01 5.08170784e-01
-7.74375200e-01 3.20239305e-01 6.68708372e+00 8.37889731e-01
-1.23412871e+00 1.53340340e-01 7.82645106e-01 1.06541783e-01
-6.82638824e-01 -2.11533859e-01 -2.41212472e-01 1.06384628e-01
1.07562172e+00 3.69193584e-01 4.70669806e-01 7.47066438e-01
4.15847123e-01 -3.32005411e-01 -9.47995603e-01 1.06497240e+00
-5.76555822e-03 -1.27718651e+00 1.49151713e-01 4.04225625e-02
5.80355525e-01 -1.66450799e-01 1.88888431e-01 1.05874330e-01
6.19840696e-02 -1.35296452e+00 9.59026515e-01 8.70567381e-01
1.00712860e+00 -7.92098939e-01 3.94623458e-01 3.04399598e-02
-1.05400264e+00 1.71374410e-01 -2.44176462e-01 4.73277643e-02
4.84778658e-02 5.10449469e-01 -4.14448470e-01 6.55567288e-01
8.97937775e-01 7.85028115e-02 -7.30735362e-01 1.38042796e+00
6.04455918e-02 2.19107032e-01 -1.94032222e-01 4.21676002e-02
2.00390384e-01 -6.39413223e-02 4.00979459e-01 1.39866436e+00
4.95903552e-01 -2.71872759e-01 -3.84968162e-01 1.07181549e+00
2.04098672e-01 3.44317883e-01 -5.89518428e-01 3.71727645e-01
5.38291223e-02 1.23907578e+00 -8.77639353e-01 -8.20798129e-02
-2.61123002e-01 1.02044296e+00 9.84494388e-02 6.49708807e-01
-8.26354861e-01 -4.84900266e-01 5.71491599e-01 3.64768654e-01
3.77935581e-02 -1.21054724e-01 -2.73920983e-01 -6.16513669e-01
9.65039060e-02 -8.30880463e-01 -9.27308574e-02 -1.38983655e+00
-9.09058154e-01 9.09293175e-01 2.26541057e-01 -1.07252312e+00
-2.29684532e-01 -8.32120717e-01 -5.53931177e-01 1.23762512e+00
-1.69651020e+00 -7.58572459e-01 -7.44258106e-01 4.41638201e-01
2.06891537e-01 2.18385592e-01 8.81331921e-01 4.43735808e-01
-1.41904041e-01 4.24829721e-01 7.91355595e-02 -2.66094685e-01
2.53009230e-01 -1.26415002e+00 4.80703562e-01 6.86344385e-01
-2.34834719e-02 3.53649765e-01 7.74958193e-01 -2.04208717e-01
-1.02557611e+00 -9.17183638e-01 4.33119684e-01 -3.51555526e-01
7.70543441e-02 -3.24753046e-01 -1.03612578e+00 6.54748753e-02
2.16140553e-01 3.02227512e-02 3.19814235e-01 -1.93166196e-01
-4.85163242e-01 -7.46392533e-02 -1.36475682e+00 7.48037219e-01
9.13159609e-01 -5.14204383e-01 -1.47988960e-01 1.32103842e-02
4.89736646e-01 -3.79640937e-01 -8.31745148e-01 5.02828479e-01
6.03630543e-01 -1.77284431e+00 1.37932885e+00 -1.93951069e-03
5.75612247e-01 -4.29940432e-01 -3.36456567e-01 -1.25822330e+00
-2.98926443e-01 -2.54467249e-01 1.79773998e-02 1.04319096e+00
2.13757768e-01 -5.52840829e-01 5.52445769e-01 2.71601379e-01
-2.43846983e-01 -7.63926327e-01 -7.26243615e-01 -6.43734694e-01
-1.01305485e-01 -7.85680890e-01 5.70804417e-01 7.56609678e-01
-8.21003318e-01 -9.38456804e-02 -1.01420078e-02 1.75996527e-01
3.89485925e-01 -3.35597634e-01 6.01949751e-01 -1.28262949e+00
-4.10947263e-01 -8.07496846e-01 -5.67299128e-01 -7.44143188e-01
-2.44690046e-01 -6.97496176e-01 8.50243866e-02 -1.78033781e+00
-2.62188584e-01 -5.93117416e-01 -2.98842609e-01 5.30695207e-02
2.09690899e-01 5.91151953e-01 9.52897519e-02 -5.58453016e-02
-4.26604003e-02 2.97783881e-01 1.28677070e+00 1.22364759e-01
-1.41840637e-01 -3.43255907e-01 -3.94422740e-01 6.09160841e-01
7.73915648e-01 -1.81731299e-01 -6.90079749e-01 -4.93576586e-01
4.30300206e-01 9.70697924e-02 1.08516359e+00 -1.42221713e+00
-1.78205490e-01 9.98213142e-02 7.71298945e-01 -4.17388558e-01
4.73062187e-01 -8.03693712e-01 5.59018672e-01 1.84390068e-01
-3.35937649e-01 2.35782653e-01 3.86747360e-01 3.12337488e-01
-1.28248036e-01 -4.37265605e-01 8.77077937e-01 -1.52688667e-01
-8.30454409e-01 -4.96469252e-02 -2.71819711e-01 -1.59855351e-01
6.09861970e-01 -6.55896246e-01 -2.62300551e-01 -2.62202770e-01
-5.42588770e-01 -3.86054873e-01 5.85510492e-01 1.67999163e-01
8.40038061e-01 -1.25134325e+00 -6.27301097e-01 1.33048683e-01
1.40736308e-02 -6.35691285e-02 4.72086400e-01 4.14987803e-01
-9.15032685e-01 -1.30853221e-01 -4.75716501e-01 -4.83261704e-01
-1.13349712e+00 4.57239330e-01 7.17320621e-01 -1.09094977e-01
-4.85658497e-01 5.64404488e-01 3.98940742e-01 -4.47908163e-01
3.33781451e-01 -3.87875646e-01 -3.85490209e-01 -7.65681267e-04
5.76500118e-01 5.43165267e-01 3.77997696e-01 -5.41273713e-01
-1.34746209e-02 4.99028146e-01 4.26332593e-01 -3.44292343e-01
1.19301188e+00 1.17798701e-01 1.13135748e-01 3.37301254e-01
1.24408722e+00 -1.14651032e-01 -1.21921754e+00 3.07618201e-01
-3.27922523e-01 -5.15577555e-01 4.14997756e-01 -1.09675896e+00
-9.84061837e-01 1.09383905e+00 1.11642265e+00 3.23970735e-01
1.54143226e+00 -3.56338948e-01 2.93508679e-01 2.98579992e-03
2.49619082e-01 -9.55643594e-01 -8.10500756e-02 2.63127387e-01
1.16444457e+00 -6.96242988e-01 -1.40362635e-01 -4.23657477e-01
-1.03820235e-01 1.10706246e+00 2.95539021e-01 -2.13053539e-01
4.45173413e-01 4.51332599e-01 1.98091984e-01 -2.31177673e-01
-3.90060812e-01 -1.07019559e-01 5.14690280e-01 9.89821613e-01
5.57438552e-01 -1.74822390e-01 -7.00883269e-02 1.87787390e-03
-2.21565083e-01 5.16541451e-02 3.38419080e-01 8.55955362e-01
-3.35788250e-01 -8.12526405e-01 -6.06368899e-01 2.52845705e-01
-3.60165566e-01 -1.39754668e-01 -8.78348276e-02 8.66996884e-01
2.34367728e-01 5.46330929e-01 5.67554720e-02 -3.29695582e-01
7.39615202e-01 -2.74604678e-01 6.82836890e-01 -2.27314413e-01
-9.73099649e-01 -9.78035014e-03 1.42599896e-01 -8.11699867e-01
-4.93250906e-01 -3.11947495e-01 -9.02531266e-01 -3.58391941e-01
1.62710361e-02 -2.74674058e-01 1.05693305e+00 5.36942601e-01
1.38240546e-01 9.19827700e-01 4.66381043e-01 -1.38867557e+00
-1.31813839e-01 -6.17708743e-01 -3.96491319e-01 3.50407869e-01
1.75488919e-01 -6.69625938e-01 -2.70838410e-01 -2.13090852e-01] | [10.972745895385742, -2.2146472930908203] |
e803e257-4bbd-4ade-bd5e-633f95d29269 | improving-translation-of-out-of-vocabulary | null | null | https://aclanthology.org/2022.amta-research.11 | https://aclanthology.org/2022.amta-research.11.pdf | Improving Translation of Out Of Vocabulary Words using Bilingual Lexicon Induction in Low-Resource Machine Translation | Dictionary-based data augmentation techniques have been used in the field of domain adaptation to learn words that do not appear in the parallel training data of a machine translation model. These techniques strive to learn correct translations of these words by generating a synthetic corpus from in-domain monolingual data utilising a dictionary obtained from bilingual lexicon induction. This paper applies these techniques to low resource machine translation, where there is often a shift in distribution of content between the parallel data and any monolingual data. English-Pashto machine learning systems are trained using a novel approach that introduces monolingual data to existing joint learning techniques for bilingual word embeddings, combined with word-for-word back-translation to improve the translation of words that do not or rarely appear in the parallel training data. Improvements are made both in terms of BLEU, chrF and word translation accuracy for an En->Ps model, compared to a baseline and when combined with back-translation. | ['Antonio Valerio Micele Barone', 'Barry Hadow', 'Alexandra Birch', 'Jonas Waldendorf'] | null | null | null | null | amta-2022-9 | ['word-translation'] | ['natural-language-processing'] | [ 2.22441778e-01 -1.09310605e-01 -7.23290145e-01 -3.70391369e-01
-9.39314365e-01 -8.16293299e-01 9.90286589e-01 3.28490883e-01
-8.03087652e-01 1.06246400e+00 5.47204256e-01 -8.60735118e-01
4.58714157e-01 -6.58846140e-01 -8.10861886e-01 -3.46937299e-01
5.43689430e-01 1.20732820e+00 -1.93422914e-01 -8.01698864e-01
-6.61440119e-02 1.24101348e-01 -8.63332987e-01 3.23059440e-01
8.82230759e-01 -7.31125027e-02 3.44828933e-01 2.96534926e-01
-4.09796596e-01 -5.48208272e-03 -3.73428255e-01 -5.06726205e-01
5.18924594e-01 -7.93734074e-01 -7.17205226e-01 -1.09847583e-01
4.76922959e-01 7.57837519e-02 -2.32446834e-01 9.56648529e-01
5.98313153e-01 -1.87654257e-01 7.27166474e-01 -5.88226974e-01
-9.89540339e-01 7.37624168e-01 -3.74503464e-01 5.49900353e-01
4.15779948e-01 -1.75749063e-01 1.03635120e+00 -1.22566605e+00
8.97821367e-01 1.13282490e+00 5.77725649e-01 4.21477348e-01
-1.63198531e+00 -7.39053547e-01 -3.03883970e-01 1.82383254e-01
-1.40134192e+00 -4.59780365e-01 4.75813061e-01 -3.11441869e-01
1.57275295e+00 -5.58334403e-02 4.57941204e-01 1.32652712e+00
4.08439726e-01 4.43693250e-01 1.37271059e+00 -1.29101145e+00
-1.57732442e-01 5.36940038e-01 -3.04393530e-01 3.29035014e-01
1.47142276e-01 3.91917706e-01 -6.47668898e-01 -1.69690400e-02
6.33367002e-01 -4.11617607e-01 1.34333968e-01 -2.85747647e-01
-1.39127159e+00 1.13238561e+00 -1.63274240e-02 7.43904829e-01
-2.22760320e-01 -4.03151423e-01 5.09026706e-01 8.33538055e-01
8.49604309e-01 5.51941574e-01 -9.32622135e-01 -7.51287639e-02
-8.65046680e-01 2.74592996e-01 6.68720901e-01 1.14034426e+00
9.90725577e-01 1.78162511e-02 -2.47836262e-02 1.06905949e+00
1.35419905e-01 8.12573731e-01 1.24854100e+00 -7.15859458e-02
7.09754646e-01 5.50147951e-01 -1.18143884e-02 -3.62532943e-01
1.50901347e-01 -3.43779474e-01 -2.19500706e-01 -1.60339445e-01
1.52471185e-01 -1.85805380e-01 -1.07538080e+00 1.70938313e+00
2.91335464e-01 -2.13043615e-01 5.45845211e-01 4.09457207e-01
3.29022497e-01 7.59261668e-01 1.31291509e-01 -2.51987010e-01
1.09214497e+00 -1.04305065e+00 -8.15971851e-01 -6.25013053e-01
9.68306243e-01 -1.45456493e+00 1.30247211e+00 -4.12409939e-02
-7.61969268e-01 -8.39262187e-01 -1.08971500e+00 -1.47858605e-01
-7.99886763e-01 1.13785975e-01 1.96139976e-01 6.59362495e-01
-9.53765094e-01 4.13356096e-01 -5.32877922e-01 -6.98588192e-01
-4.80418615e-02 4.81896639e-01 -6.72568798e-01 -2.96470731e-01
-1.41629541e+00 1.53771460e+00 7.40389109e-01 -6.59701824e-01
-6.45233452e-01 -6.22683942e-01 -9.53412056e-01 -5.19159496e-01
-3.27218801e-01 -4.26994950e-01 1.17698348e+00 -1.31945932e+00
-1.42391586e+00 1.10243082e+00 -3.13357145e-01 -3.94062549e-01
2.00167015e-01 -1.34156480e-01 -7.67250955e-01 -5.96353590e-01
3.95307779e-01 6.00882053e-01 7.05854595e-01 -8.66274357e-01
-7.76288688e-01 -2.55792886e-01 -4.35496479e-01 5.75862408e-01
-2.29881033e-01 3.07779998e-01 -6.94646463e-02 -8.53262484e-01
-2.44781539e-01 -1.03090656e+00 -2.52963215e-01 -7.81843185e-01
1.31214052e-01 -3.00933391e-01 5.94702244e-01 -9.65357065e-01
1.10375059e+00 -1.85363615e+00 2.92595983e-01 1.06018908e-01
-6.13359153e-01 5.92763484e-01 -5.86818874e-01 9.28618610e-01
-3.34159225e-01 -9.48673561e-02 -1.10330291e-01 -2.71704644e-01
-1.80967882e-01 7.71689594e-01 -4.10939425e-01 4.04643804e-01
2.94534206e-01 9.77756798e-01 -9.47536647e-01 -2.84680009e-01
1.69167802e-01 2.64262766e-01 -3.78954977e-01 3.05545360e-01
-2.07235038e-01 5.93357265e-01 1.18648790e-01 1.94402531e-01
3.26362044e-01 5.98472357e-01 4.05507952e-01 1.83631703e-01
-2.12982774e-01 8.89187813e-01 -5.84777534e-01 1.94278526e+00
-8.69068623e-01 5.91295600e-01 -5.77043116e-01 -9.00407135e-01
1.17073238e+00 6.96534574e-01 1.07789589e-02 -1.01917851e+00
1.41879663e-01 8.25365245e-01 2.44294092e-01 -3.43712777e-01
5.67055047e-01 -7.32319653e-01 -9.24845710e-02 4.79418039e-01
5.55506408e-01 -2.24090084e-01 3.71301025e-01 -2.64492720e-01
7.15442836e-01 5.19127667e-01 5.59065342e-01 -4.53679472e-01
4.95350808e-01 6.13646507e-01 4.20136184e-01 1.67564452e-01
2.78298914e-01 3.93514782e-01 -1.00778103e-01 -4.85552639e-01
-1.77838767e+00 -1.05712926e+00 -2.40519464e-01 1.27965629e+00
-2.80372739e-01 -2.89326668e-01 -5.61652839e-01 -8.19238007e-01
-2.84298837e-01 1.01862383e+00 -4.99489456e-01 -1.54153988e-01
-9.89614189e-01 -7.91134655e-01 4.94709015e-01 3.26536447e-01
-2.50231951e-01 -1.04527652e+00 1.36763930e-01 5.72532713e-01
-2.63887882e-01 -1.15383780e+00 -6.45620823e-01 5.04672527e-01
-1.06816196e+00 -5.69527686e-01 -5.56055963e-01 -1.28615820e+00
7.40883827e-01 -2.42641643e-02 1.36665881e+00 -3.52924258e-01
-5.50642004e-03 -1.63743049e-01 -6.20836377e-01 -6.15980685e-01
-1.07819474e+00 3.73642832e-01 3.96825284e-01 -3.05953830e-01
1.23892689e+00 -4.66690451e-01 7.71593079e-02 1.85634091e-01
-1.01360655e+00 -9.33874026e-02 8.80643368e-01 1.07382345e+00
5.99500954e-01 -6.31442487e-01 5.96207619e-01 -8.84569168e-01
8.22709739e-01 -5.79879224e-01 -3.51315349e-01 2.72302836e-01
-1.04039538e+00 3.89042109e-01 6.12123728e-01 -8.28639507e-01
-6.08665049e-01 -3.45567577e-02 -2.17356279e-01 -9.05329362e-02
-1.72143266e-01 5.14160573e-01 -8.78304243e-02 1.60886377e-01
1.07972527e+00 3.92603278e-01 -4.51749424e-03 -5.93361557e-01
6.13297582e-01 1.07869208e+00 2.89076686e-01 -5.40131092e-01
8.64022315e-01 -1.96409032e-01 -5.11659324e-01 -5.75441241e-01
-5.54249227e-01 -6.71742737e-01 -1.07476544e+00 4.44777995e-01
6.34416878e-01 -1.18545091e+00 7.69937515e-01 -1.06206723e-01
-1.23132718e+00 -2.87827700e-01 -5.28478146e-01 8.93109739e-01
-5.02381861e-01 8.13894123e-02 -2.73426503e-01 -8.21489468e-02
-2.68124908e-01 -1.16598403e+00 9.08394158e-01 -1.84919849e-01
-6.22662127e-01 -1.39013386e+00 8.83720398e-01 2.34601498e-01
4.30635184e-01 -3.41858476e-01 1.30434322e+00 -1.24117315e+00
1.89652979e-01 -3.44331771e-01 1.66924819e-01 6.64071739e-01
5.75120866e-01 -4.93684977e-01 -4.69627172e-01 -6.07728004e-01
-2.69794077e-01 -2.21288264e-01 1.35541618e-01 -1.64869443e-01
-2.30573967e-01 -3.82173777e-01 -1.40157208e-01 4.32902575e-01
1.65413845e+00 1.93980098e-01 3.86268288e-01 5.47856748e-01
4.99510437e-01 4.46119338e-01 5.24823904e-01 -2.42648765e-01
2.20458344e-01 9.34488833e-01 -2.80825287e-01 -2.21592650e-01
-4.56719548e-01 -5.83004534e-01 7.42131293e-01 1.34157360e+00
2.26943925e-01 8.33186507e-02 -1.13024521e+00 1.27459967e+00
-1.51477540e+00 -4.06457275e-01 -2.19954059e-01 2.25685096e+00
1.48932803e+00 -3.30981947e-02 3.05023771e-02 -1.11387193e-01
5.74793816e-01 -1.18419029e-01 8.69372711e-02 -1.04527724e+00
-2.46659219e-01 1.00986516e+00 8.74069691e-01 8.48743618e-01
-8.29826534e-01 1.52988112e+00 6.55698633e+00 9.13229048e-01
-1.13200176e+00 7.14835584e-01 5.95597960e-02 1.91159263e-01
-4.94204670e-01 2.23644927e-01 -9.26791489e-01 2.01088652e-01
1.44495404e+00 -2.31700882e-01 5.30180097e-01 5.05796909e-01
1.48308739e-01 2.25652516e-01 -1.41178930e+00 6.74748421e-01
3.44560534e-01 -9.56285775e-01 3.23364466e-01 3.77455987e-02
1.23903191e+00 4.35260892e-01 -1.04493648e-02 4.99751478e-01
5.06342113e-01 -1.05575311e+00 4.10726935e-01 -1.33189604e-01
1.07893240e+00 -8.23033512e-01 9.70341623e-01 3.71508598e-01
-6.79077208e-01 5.57573915e-01 -5.46612918e-01 -1.24994837e-01
2.88943145e-02 1.97747827e-01 -1.41312361e+00 5.38377047e-01
1.75951406e-01 5.68822801e-01 -4.04437631e-01 5.06401837e-01
-4.72392201e-01 7.29497731e-01 -2.19064355e-01 5.02983518e-02
4.13312525e-01 -4.50726002e-01 4.49873090e-01 1.54549503e+00
4.29944605e-01 -6.79835975e-01 1.70560464e-01 4.26729590e-01
-1.10777408e-01 8.59102011e-01 -9.57999408e-01 -2.33904526e-01
2.20384613e-01 7.13338554e-01 -2.29228899e-01 -4.42167193e-01
-6.66437447e-01 1.33403635e+00 2.41826490e-01 2.64967829e-01
-3.99145484e-01 -2.96200722e-01 7.32445478e-01 2.32180580e-01
2.42152452e-01 -4.56028849e-01 -2.97208250e-01 -1.10444582e+00
-5.02842180e-02 -1.37112725e+00 -1.29068971e-01 -2.85589814e-01
-1.30449867e+00 1.06151545e+00 7.70989992e-03 -1.25580466e+00
-7.21463859e-01 -5.71023524e-01 -3.12718183e-01 1.54937112e+00
-1.38015270e+00 -1.48782194e+00 5.87084532e-01 5.13811171e-01
8.75531375e-01 -8.35046470e-01 1.32111156e+00 4.94107395e-01
-4.12170626e-02 7.13863194e-01 4.58561867e-01 2.79397577e-01
1.15362394e+00 -1.22561681e+00 9.42171872e-01 8.59483421e-01
8.19579601e-01 7.00843513e-01 8.82775128e-01 -8.12104881e-01
-8.25986624e-01 -1.11116695e+00 1.86330509e+00 -7.69634962e-01
8.79690349e-01 -4.93040890e-01 -7.04531729e-01 9.60859776e-01
8.14054072e-01 -2.26402521e-01 1.02056468e+00 2.31222332e-01
-4.40100104e-01 8.78191069e-02 -9.03083026e-01 6.99765384e-01
5.71163774e-01 -8.18655968e-01 -1.23463631e+00 6.09057844e-01
4.44636315e-01 -1.34144634e-01 -6.02243900e-01 1.66586682e-01
2.68601149e-01 -1.10388488e-01 7.03683436e-01 -1.10613477e+00
3.71343821e-01 -1.78273559e-01 -3.73269111e-01 -1.87510073e+00
-3.93463857e-02 -5.15822113e-01 4.07446474e-01 1.01238263e+00
9.88194168e-01 -4.75451618e-01 4.76078063e-01 -2.56001174e-01
-2.37312540e-03 -3.96782190e-01 -1.13815844e+00 -1.00550890e+00
7.73360074e-01 -2.22825515e-03 3.85136276e-01 1.19561255e+00
6.28722757e-02 1.01265049e+00 -3.58003050e-01 -2.09574461e-01
2.30226040e-01 -1.58864424e-01 6.66920185e-01 -8.81519139e-01
-2.75364637e-01 -9.99547318e-02 -2.87004620e-01 -8.46171081e-01
3.17367196e-01 -1.43135071e+00 9.95579585e-02 -1.41491318e+00
2.37830821e-02 -4.62044716e-01 -1.02775931e-01 5.31842172e-01
-1.78290412e-01 6.63173497e-01 -1.36691496e-01 3.21600050e-01
1.56399682e-01 5.01204550e-01 9.49602842e-01 -1.62728075e-02
-2.54072458e-01 -3.09263617e-01 -3.36873531e-01 3.07733893e-01
8.75207603e-01 -9.75746632e-01 -3.30399781e-01 -7.67901421e-01
2.09348217e-01 -3.16972286e-01 -3.69070888e-01 -7.63141215e-01
-2.21672490e-01 -1.21580325e-01 1.51698992e-01 -3.51054072e-02
-1.39335841e-02 -8.92741859e-01 7.68036097e-02 4.30649847e-01
-2.69572049e-01 8.75262916e-01 3.80192846e-01 1.93882659e-01
-4.34234947e-01 -3.83783728e-01 8.31672370e-01 -1.98062524e-01
-5.41804731e-01 -7.94586167e-02 -5.65478146e-01 1.62005927e-02
9.72759902e-01 -2.09795281e-01 2.28617147e-01 2.02639382e-02
-6.91450596e-01 -3.12854409e-01 6.44830942e-01 8.62267673e-01
1.43455863e-01 -1.66238356e+00 -1.17734587e+00 6.59023345e-01
4.85280991e-01 -5.00180662e-01 -6.33955181e-01 6.70088530e-01
-7.07042396e-01 6.49540126e-01 -5.53160965e-01 -2.88758874e-01
-1.11952496e+00 5.00240862e-01 1.56011149e-01 -4.75594848e-01
-1.85710028e-01 6.02449179e-01 -3.48669022e-01 -1.17943132e+00
-2.36957446e-01 -5.96702993e-02 9.12345853e-03 -4.32096981e-02
1.68841839e-01 -1.74547032e-01 4.59181309e-01 -1.13339365e+00
-1.64259747e-01 4.97509867e-01 -3.04507852e-01 -6.68727398e-01
1.22325063e+00 -1.86621919e-01 -1.49395898e-01 6.83052719e-01
1.33759844e+00 5.00408590e-01 -3.89924377e-01 -7.90399849e-01
3.05390328e-01 -3.78049463e-01 -8.97402540e-02 -1.01295424e+00
-3.96108359e-01 8.44710767e-01 9.09221351e-01 -5.46230078e-01
6.79073036e-01 -5.71339391e-02 8.74100387e-01 1.91469669e-01
4.60128546e-01 -1.33679807e+00 -3.66344541e-01 9.30532277e-01
6.40008748e-01 -1.44952297e+00 -2.94838458e-01 5.74949868e-02
-5.66104889e-01 1.06639314e+00 2.62560278e-01 -2.63174266e-01
4.23463196e-01 2.78435618e-01 6.84575975e-01 2.63114452e-01
-5.55518448e-01 -2.25777149e-01 3.71909738e-01 7.25238264e-01
8.68200719e-01 2.66822129e-01 -9.96949196e-01 -1.20966388e-02
-2.74212629e-01 -2.43705556e-01 2.00926378e-01 8.30323219e-01
-1.97872400e-01 -2.25559044e+00 -2.15575621e-01 1.11408822e-01
-6.61538601e-01 -6.90913498e-01 -5.80152988e-01 8.14949334e-01
4.44294453e-01 7.79228389e-01 1.21293403e-01 -3.70194435e-01
2.91901708e-01 6.09717429e-01 5.85799098e-01 -1.17066145e+00
-7.55755842e-01 1.31565765e-01 1.56675473e-01 -5.95438629e-02
-2.89252907e-01 -7.31062591e-01 -8.53878796e-01 -5.87687455e-02
-2.54011869e-01 5.93880951e-01 1.00102007e+00 1.26589763e+00
1.22758700e-02 2.62478501e-01 4.54166561e-01 -4.95008796e-01
-6.02306366e-01 -1.45063412e+00 -1.04924262e-01 4.49530214e-01
1.24996752e-01 -2.80293912e-01 9.19171795e-02 2.05102727e-01] | [11.267314910888672, 10.226995468139648] |
d3be8e16-4786-4b8d-83c5-e617c4655907 | a-domain-knowledge-inspired-music-embedding | 2212.00973 | null | https://arxiv.org/abs/2212.00973v1 | https://arxiv.org/pdf/2212.00973v1.pdf | A Domain-Knowledge-Inspired Music Embedding Space and a Novel Attention Mechanism for Symbolic Music Modeling | Following the success of the transformer architecture in the natural language domain, transformer-like architectures have been widely applied to the domain of symbolic music recently. Symbolic music and text, however, are two different modalities. Symbolic music contains multiple attributes, both absolute attributes (e.g., pitch) and relative attributes (e.g., pitch interval). These relative attributes shape human perception of musical motifs. These important relative attributes, however, are mostly ignored in existing symbolic music modeling methods with the main reason being the lack of a musically-meaningful embedding space where both the absolute and relative embeddings of the symbolic music tokens can be efficiently represented. In this paper, we propose the Fundamental Music Embedding (FME) for symbolic music based on a bias-adjusted sinusoidal encoding within which both the absolute and the relative attributes can be embedded and the fundamental musical properties (e.g., translational invariance) are explicitly preserved. Taking advantage of the proposed FME, we further propose a novel attention mechanism based on the relative index, pitch and onset embeddings (RIPO attention) such that the musical domain knowledge can be fully utilized for symbolic music modeling. Experiment results show that our proposed model: RIPO transformer which utilizes FME and RIPO attention outperforms the state-of-the-art transformers (i.e., music transformer, linear transformer) in a melody completion task. Moreover, using the RIPO transformer in a downstream music generation task, we notice that the notorious degeneration phenomenon no longer exists and the music generated by the RIPO transformer outperforms the music generated by state-of-the-art transformer models in both subjective and objective evaluations. | ['D. Herremans', 'J. Kang', 'Z. Guo'] | 2022-12-02 | null | null | null | null | ['music-generation', 'music-modeling', 'music-generation'] | ['audio', 'music', 'music'] | [ 1.49786100e-01 -2.68259436e-01 -8.17323625e-02 -4.02622558e-02
-5.49298644e-01 -7.44893014e-01 5.04880011e-01 -4.42380421e-02
-1.53204992e-01 3.41910303e-01 2.70508915e-01 1.98253989e-01
-5.83672464e-01 -6.04431868e-01 -6.10455751e-01 -8.33280206e-01
3.71933840e-02 2.66496778e-01 -6.03489019e-02 -4.10160482e-01
2.30608135e-01 1.22579299e-01 -1.61852443e+00 1.13755330e-01
8.74952257e-01 1.15456963e+00 1.96053162e-01 3.11877310e-01
-1.57444015e-01 6.31889164e-01 -5.07751942e-01 -3.18561584e-01
2.16069236e-01 -6.31015360e-01 -2.75855392e-01 -3.80766898e-01
2.82549113e-01 -1.15689144e-01 -4.44861919e-01 1.15005922e+00
6.65623784e-01 3.45622599e-01 6.96069241e-01 -1.25282276e+00
-9.18791354e-01 1.07986462e+00 -3.14760566e-01 -2.75736749e-01
2.44932026e-01 -8.22230726e-02 1.61514723e+00 -1.07233584e+00
2.13241413e-01 1.18210757e+00 7.50737846e-01 2.09228858e-01
-1.38354635e+00 -9.46463227e-01 -2.37754416e-02 2.89299458e-01
-1.69780743e+00 -2.81110615e-01 1.30890489e+00 -4.56993341e-01
6.48341119e-01 4.23943788e-01 8.28189969e-01 1.00330424e+00
-2.52786987e-02 6.98944747e-01 8.44689548e-01 -3.75000060e-01
2.88046971e-02 1.83172133e-02 -2.79881865e-01 3.26719791e-01
-2.33044863e-01 1.99830428e-01 -8.40774477e-01 4.82165366e-02
9.97621715e-01 -2.90088989e-02 -3.85667294e-01 -3.47178727e-01
-1.71051705e+00 5.87577999e-01 5.68005145e-01 5.85645318e-01
-3.65617722e-01 3.78720552e-01 5.48429191e-01 2.51871288e-01
-3.86170931e-02 7.08408833e-01 -1.61426619e-01 -2.72291750e-01
-9.90301073e-01 2.71877915e-01 3.73966783e-01 8.67891312e-01
4.36148673e-01 4.97848451e-01 -3.49066079e-01 1.06053543e+00
2.42600903e-01 5.59407413e-01 8.79907787e-01 -7.87405550e-01
2.52021641e-01 5.31097114e-01 -5.96960681e-03 -1.17658424e+00
-1.65205643e-01 -7.30296969e-01 -1.00862527e+00 4.82588373e-02
2.79910266e-01 4.59042490e-01 -4.49582279e-01 2.38789415e+00
-4.98615019e-02 4.51144099e-01 7.65583068e-02 9.23796773e-01
7.23686099e-01 6.70994520e-01 -1.31077483e-01 -4.89913262e-02
1.36405480e+00 -8.83053720e-01 -9.12548482e-01 2.24048674e-01
1.15442634e-01 -8.96237791e-01 1.47240317e+00 4.26086515e-01
-9.37870562e-01 -9.73489046e-01 -1.24180102e+00 -2.51860350e-01
-2.17605591e-01 3.77692848e-01 3.89048845e-01 3.04258257e-01
-5.86354434e-01 9.66426849e-01 -5.73540807e-01 -1.06762931e-01
-6.46110624e-02 3.17054391e-01 -1.85925245e-01 4.75947142e-01
-1.36245537e+00 4.30111289e-01 5.00726521e-01 1.52991310e-01
-7.06008434e-01 -7.46844649e-01 -7.15782344e-01 4.17207450e-01
1.76527083e-01 -7.18951702e-01 1.09272575e+00 -1.02602470e+00
-1.64003670e+00 5.39681256e-01 1.26682848e-01 -1.86482534e-01
1.61370620e-01 -3.44545960e-01 -3.83374423e-01 -2.29962803e-02
6.66462928e-02 5.37954748e-01 1.00861585e+00 -9.63490963e-01
-3.37099731e-01 -1.66111350e-01 6.12044595e-02 3.90073091e-01
-7.27483511e-01 -3.02834392e-01 -1.82937503e-01 -1.39263701e+00
1.80725217e-01 -1.02315807e+00 2.87928849e-01 2.90002506e-02
-5.52911758e-01 -2.91154087e-01 5.76675296e-01 -5.54844737e-01
1.67191195e+00 -2.64550114e+00 6.09225035e-01 1.93054661e-01
5.45616075e-02 -2.86349817e-03 -2.08170831e-01 5.66104889e-01
-3.12987179e-01 5.70290349e-02 -1.69265583e-01 -1.66452959e-01
5.78891337e-01 8.67495835e-02 -7.30055332e-01 9.68310237e-02
1.50184661e-01 9.60929215e-01 -9.75430906e-01 -2.86806881e-01
1.78425580e-01 6.79391146e-01 -6.39400542e-01 2.10415736e-01
-1.61959320e-01 5.29796779e-01 -3.60280544e-01 5.99978983e-01
2.53724247e-01 1.08086199e-01 -4.21651490e-02 -5.92829823e-01
-1.59145281e-01 4.61072415e-01 -1.24976861e+00 1.98923886e+00
-3.15063059e-01 4.23275620e-01 -2.89132655e-01 -7.83984900e-01
1.05371630e+00 6.16015553e-01 5.50774395e-01 -5.65491974e-01
8.28200504e-02 5.01938343e-01 2.85546929e-01 -6.41364232e-02
5.44259250e-01 -5.69314182e-01 -1.38314292e-01 4.88262802e-01
1.25060668e-02 -6.12995327e-02 -1.21209556e-02 -9.61959362e-02
6.42315924e-01 3.69907975e-01 2.69837677e-01 -7.82938004e-02
5.41336060e-01 -3.67951304e-01 6.04149818e-01 3.57546687e-01
8.38363394e-02 6.41567111e-01 3.38594526e-01 -1.53229102e-01
-1.08898020e+00 -1.14643478e+00 -6.82037100e-02 1.27937579e+00
6.27285689e-02 -6.60106778e-01 -5.65953434e-01 2.15491988e-02
1.07504591e-01 7.22992837e-01 -6.02811158e-01 -4.73894566e-01
-5.94952524e-01 -2.32123122e-01 8.78846407e-01 7.67428517e-01
3.03315252e-01 -1.42206275e+00 -3.98382008e-01 4.62487251e-01
-3.76559854e-01 -9.43231463e-01 -9.26094055e-01 3.66871133e-02
-7.54439831e-01 -6.50225222e-01 -7.18369842e-01 -6.09898984e-01
2.41383966e-02 -4.09327559e-02 7.18929172e-01 -1.43054694e-01
1.81965232e-02 2.22852662e-01 -4.33621615e-01 -3.75934333e-01
-9.84038264e-02 -6.60075620e-02 3.81691337e-01 4.90201682e-01
1.04232300e-02 -1.17086279e+00 -4.41943049e-01 2.95544356e-01
-1.02377450e+00 2.18839660e-01 7.46213317e-01 1.03587842e+00
8.19038332e-01 -3.24035250e-02 9.28041756e-01 -3.09834689e-01
6.19843721e-01 -1.05905823e-01 -1.24324515e-01 4.39388342e-02
-3.65524381e-01 2.05034584e-01 7.36559987e-01 -9.41398978e-01
-5.76150537e-01 -2.92254686e-01 -1.07390173e-01 -1.01789629e+00
2.49508321e-01 8.97886455e-01 -3.29465419e-01 2.42183238e-01
3.48513484e-01 5.29408693e-01 -1.43510684e-01 -6.80836260e-01
3.24047565e-01 5.09699583e-01 8.12440991e-01 -7.33768046e-01
9.33916032e-01 1.55152544e-01 1.23079203e-01 -7.37685859e-01
-6.87262774e-01 -2.95148969e-01 -5.50471544e-01 -1.71940941e-02
7.75423706e-01 -7.52205670e-01 -7.90717542e-01 3.91117603e-01
-1.00181341e+00 8.32299739e-02 -6.97952330e-01 5.76293290e-01
-9.25347447e-01 3.68771732e-01 -4.80190098e-01 -8.76585662e-01
-4.30241317e-01 -1.01500463e+00 1.12798011e+00 -7.40456879e-02
-5.48929691e-01 -6.49322569e-01 3.46548744e-02 2.29124241e-02
3.60882759e-01 4.22213703e-01 1.46696067e+00 -6.02262557e-01
-5.86219013e-01 -9.68743637e-02 -1.33302351e-02 2.76679754e-01
2.10698441e-01 -1.24739937e-01 -1.15382171e+00 -1.99624047e-01
-1.02823898e-01 -2.24481359e-01 7.47857213e-01 2.06791714e-01
9.71053302e-01 -2.15685949e-01 1.40550420e-01 7.13515937e-01
1.12178922e+00 2.05282494e-01 5.70273757e-01 2.15886295e-01
8.18225980e-01 3.08404088e-01 5.46400130e-01 4.47798908e-01
1.28708228e-01 1.21747816e+00 2.99375445e-01 1.58558771e-01
-3.23947847e-01 -7.01720119e-01 6.50446951e-01 1.44402242e+00
-3.81501794e-01 1.49059176e-01 -5.17648697e-01 4.19367880e-01
-1.98685813e+00 -1.06060171e+00 2.97964603e-01 2.56748462e+00
9.84503448e-01 2.50077434e-02 8.21524933e-02 7.02402115e-01
5.72159708e-01 1.05852373e-01 -6.41880393e-01 -1.96342349e-01
-2.51694739e-01 4.27550346e-01 -2.27234304e-01 1.39389887e-01
-9.87315416e-01 7.17345893e-01 4.82978773e+00 1.12293994e+00
-1.20006824e+00 -5.76830618e-02 -2.85171002e-01 -8.29950273e-02
-4.18247610e-01 -6.72764555e-02 -1.41003743e-01 3.51025939e-01
6.24450028e-01 -4.55921322e-01 8.75515819e-01 5.85572124e-01
1.38652518e-01 7.18555033e-01 -1.41078174e+00 1.21485198e+00
-8.73458534e-02 -9.42249060e-01 3.89138252e-01 3.33856903e-02
3.52376968e-01 -4.82320130e-01 4.23548132e-01 6.16081119e-01
-6.10124290e-01 -1.16237760e+00 1.36562788e+00 7.02062905e-01
1.14103198e+00 -6.97403371e-01 4.71350104e-01 2.54663885e-01
-1.66235828e+00 -2.64561713e-01 -1.54247910e-01 6.47669286e-03
7.15246052e-03 2.36560881e-01 -3.87137890e-01 8.21009278e-01
5.23802519e-01 1.10628176e+00 -3.81264329e-01 8.56309056e-01
-2.04100519e-01 6.51902616e-01 -2.02766806e-01 1.63027197e-01
1.68879226e-01 -3.95067900e-01 9.39075053e-01 8.29872251e-01
6.92028701e-01 -6.18578680e-02 -3.08581274e-02 1.25459337e+00
3.84037979e-02 2.90936768e-01 -4.24779177e-01 -5.81121802e-01
5.19241035e-01 1.03621292e+00 -3.17299902e-01 -1.66910902e-01
-1.19830422e-01 7.88306832e-01 7.61382421e-03 4.20777589e-01
-8.37346852e-01 -5.87901831e-01 7.51092732e-01 3.02793141e-02
4.71596748e-01 -1.75321057e-01 -1.72956809e-01 -1.20476794e+00
5.69324307e-02 -1.00881183e+00 3.10312416e-02 -9.16172564e-01
-1.35099363e+00 6.15048289e-01 -1.44144446e-01 -1.76964235e+00
-3.61425698e-01 -5.24727643e-01 -5.49593210e-01 9.54072118e-01
-1.29287088e+00 -1.41948020e+00 -9.61188674e-02 6.69547081e-01
2.81556010e-01 -3.24393243e-01 1.17068672e+00 5.35412669e-01
-2.50991225e-01 8.82536471e-01 5.47083542e-02 2.33772188e-01
8.61516893e-01 -1.29359639e+00 -1.03275515e-02 2.53782153e-01
6.89647615e-01 9.40187752e-01 5.54745495e-01 -2.12119803e-01
-1.48004138e+00 -8.79849851e-01 8.21915448e-01 -2.32844427e-01
8.46392930e-01 -3.77491266e-01 -9.89782512e-01 5.52249074e-01
1.48840109e-02 -3.85339409e-01 8.59023631e-01 1.49490699e-01
-7.05946922e-01 -3.31825823e-01 -5.02095580e-01 6.90010190e-01
9.81949091e-01 -9.32107210e-01 -7.37878561e-01 -1.34120643e-01
9.24952149e-01 -2.72087812e-01 -9.13514435e-01 5.66635430e-01
1.05639744e+00 -8.14167440e-01 1.04875290e+00 -4.52639461e-01
4.06200171e-01 -6.80733085e-01 -4.88660395e-01 -1.18180001e+00
-5.30559480e-01 -6.08898044e-01 -5.21122962e-02 1.45125043e+00
1.32367030e-01 -4.52284038e-01 5.90068810e-02 -1.01770386e-01
-3.26310903e-01 -6.28134191e-01 -1.16655624e+00 -9.49067712e-01
-4.29336764e-02 -4.89685893e-01 7.64921367e-01 1.13478732e+00
5.27841933e-02 5.97884536e-01 -5.68702281e-01 -5.21141067e-02
3.97986561e-01 5.85470438e-01 6.03217542e-01 -1.39322400e+00
-5.71408033e-01 -8.14249992e-01 -6.72579110e-01 -8.38044047e-01
1.06546462e-01 -1.21058536e+00 -2.35147044e-01 -9.78475571e-01
1.97743699e-01 -4.09629434e-01 -9.62182820e-01 5.42076170e-01
-4.81599383e-02 2.33122185e-01 6.83528066e-01 4.01177615e-01
-1.57653034e-01 1.07755196e+00 1.30390859e+00 -2.55631983e-01
-2.20483333e-01 -2.22236633e-01 -5.91625810e-01 6.89867020e-01
5.57478368e-01 -3.60975295e-01 -3.74945939e-01 -2.41324767e-01
4.91888344e-01 6.47503659e-02 5.13361692e-01 -1.20228577e+00
2.03871638e-01 6.15144260e-02 1.51667356e-01 -4.46455717e-01
7.58980870e-01 -8.29392254e-01 3.70414168e-01 3.21792960e-01
-4.69554216e-01 9.68098789e-02 2.94325948e-01 6.40004456e-01
-6.26481175e-01 -5.00358120e-02 5.25728881e-01 2.85894930e-01
-3.93408149e-01 1.53843284e-01 1.28119379e-01 -5.35883605e-02
5.61066747e-01 -3.08204800e-01 6.17348850e-02 -3.93605441e-01
-7.78003812e-01 -2.83317894e-01 2.06006244e-01 5.99445045e-01
6.32710040e-01 -2.02736020e+00 -7.58550823e-01 3.17771137e-01
3.33668828e-01 -4.40477639e-01 2.64659375e-01 1.10100472e+00
5.86729236e-02 3.29495162e-01 -3.43560845e-01 -5.29989243e-01
-9.72534180e-01 6.86620712e-01 1.34055451e-01 -2.00777605e-01
-6.99999392e-01 4.54758376e-01 5.45229316e-01 -5.19177854e-01
3.75906646e-01 -6.15678728e-01 -2.36274198e-01 1.73738137e-01
3.16355199e-01 1.28099248e-01 -3.25939000e-01 -8.07548940e-01
-3.06064397e-01 1.07966554e+00 4.87668306e-01 -3.94970715e-01
1.27231920e+00 1.59392968e-01 -2.00519711e-01 1.03931189e+00
8.90476942e-01 4.35808241e-01 -7.16840506e-01 -2.42629737e-01
-1.89290032e-01 -4.32991594e-01 -1.50043115e-01 -7.66266882e-01
-8.48374069e-01 1.18503869e+00 4.30121124e-01 1.56064779e-01
1.29367566e+00 -2.81945974e-01 8.42970550e-01 1.08430482e-01
3.28847498e-01 -7.27519035e-01 1.94401503e-01 5.69000840e-01
1.35921717e+00 -6.65423691e-01 -3.44069123e-01 -1.16738096e-01
-5.61907351e-01 1.07023907e+00 2.80596733e-01 -8.65442976e-02
5.00173628e-01 -7.41620064e-02 -9.65725780e-02 6.02321103e-02
-6.55918241e-01 -1.23771437e-01 8.18196833e-01 2.61804283e-01
6.02515757e-01 2.02517658e-01 -1.86315730e-01 1.29520595e+00
-7.58237600e-01 -1.74298346e-01 1.53994188e-01 3.30767900e-01
-2.49396116e-01 -1.08090162e+00 -4.39952195e-01 -8.09419714e-03
-2.42850229e-01 -3.47587407e-01 -4.27580088e-01 7.14318275e-01
4.30509210e-01 5.82534313e-01 -8.40684772e-02 -6.61877930e-01
6.11413956e-01 2.95294136e-01 5.26846170e-01 -4.24246073e-01
-6.31267130e-01 4.82639790e-01 -3.09789956e-01 -4.08097088e-01
-3.81620944e-01 -3.71616602e-01 -1.31133783e+00 -4.89211679e-02
-2.99455583e-01 1.06469467e-01 3.93583298e-01 5.78488529e-01
2.17283294e-01 8.28523993e-01 5.36811411e-01 -9.91971612e-01
-7.18190193e-01 -1.15540326e+00 -1.00742650e+00 6.21133626e-01
3.32055390e-01 -8.62624049e-01 -3.48914683e-01 8.02406967e-02] | [15.96082592010498, 5.491605758666992] |
712f199e-c39f-491f-9571-dc8e8807b054 | global-sensitivity-analysis-in-probabilistic | 2110.03749 | null | https://arxiv.org/abs/2110.03749v1 | https://arxiv.org/pdf/2110.03749v1.pdf | Global sensitivity analysis in probabilistic graphical models | We show how to apply Sobol's method of global sensitivity analysis to measure the influence exerted by a set of nodes' evidence on a quantity of interest expressed by a Bayesian network. Our method exploits the network structure so as to transform the problem of Sobol index estimation into that of marginalization inference. This way, we can efficiently compute indices for networks where brute-force or Monte Carlo based estimators for variance-based sensitivity analysis would require millions of costly samples. Moreover, our method gives exact results when exact inference is used, and also supports the case of correlated inputs. The proposed algorithm is inspired by the field of tensor networks, and generalizes earlier tensor sensitivity techniques from the acyclic to the cyclic case. We demonstrate the method on three medium to large Bayesian networks that cover the areas of project risk management and reliability engineering. | ['Manuele Leonelli', 'Rafael Ballester-Ripoll'] | 2021-10-07 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 1.19893730e-01 7.90769607e-03 -1.11229599e-01 -3.51036608e-01
-1.59173831e-01 -5.15045166e-01 3.27844858e-01 -3.74724492e-02
-2.17668876e-01 7.63175607e-01 1.50587246e-01 -5.81422269e-01
-9.58482563e-01 -1.22823012e+00 -7.29488492e-01 -6.27155304e-01
-6.66440904e-01 5.05180478e-01 4.50880677e-01 -1.80655539e-01
6.06042266e-01 8.03131938e-01 -8.48417580e-01 -1.29910275e-01
3.35070819e-01 8.58235776e-01 -3.09191197e-01 5.47850490e-01
2.95346528e-01 8.17936063e-01 -3.71716827e-01 -6.21325970e-01
9.35887173e-02 -2.74789650e-02 -8.18657994e-01 -3.06150436e-01
6.55175978e-03 -3.95735681e-01 -2.08278477e-01 1.21389592e+00
3.20662767e-01 1.32412776e-01 8.95707428e-01 -1.34076464e+00
-1.38204351e-01 1.24182355e+00 -7.17018664e-01 5.05190372e-01
2.35044390e-01 -2.56732166e-01 1.12997448e+00 -6.65871143e-01
3.53549004e-01 1.61875081e+00 7.45173633e-01 -1.97024480e-01
-1.62908268e+00 -4.41703081e-01 2.02674121e-02 1.16409436e-01
-1.21675622e+00 -1.34422749e-01 6.91372752e-01 -8.84423733e-01
4.66224670e-01 1.60655752e-01 3.80342722e-01 9.07498658e-01
4.11550522e-01 3.49804908e-01 1.22679222e+00 -1.41499862e-01
6.07122779e-01 -2.17549920e-01 5.42687535e-01 6.74748838e-01
6.55165017e-01 2.31360689e-01 -4.95081663e-01 -6.79930568e-01
7.87559211e-01 2.97211036e-02 -3.94411646e-02 -1.43923417e-01
-1.19136727e+00 9.17392731e-01 5.62493801e-02 1.83610260e-01
-2.79309154e-01 8.31440330e-01 3.08863491e-01 3.33109528e-01
4.94621515e-01 2.62797654e-01 -3.80133510e-01 -1.47292726e-02
-8.18975925e-01 2.06176236e-01 1.14407730e+00 6.58567846e-01
6.98248804e-01 -1.18148051e-01 -3.14259350e-01 6.00115061e-01
6.20268404e-01 5.59884369e-01 -7.44194508e-01 -1.35194004e+00
4.10342216e-01 2.64505565e-01 9.49350670e-02 -1.42569935e+00
-3.67269605e-01 -5.21791637e-01 -7.15483069e-01 1.34870484e-01
7.32953310e-01 -2.04156920e-01 -3.53739589e-01 1.86288631e+00
1.67558834e-01 -1.03587016e-01 -5.86121261e-01 5.82034886e-01
-1.28465772e-01 5.42007744e-01 -3.28548729e-01 -3.90431464e-01
1.21476877e+00 -1.27532035e-01 -7.60442078e-01 2.21200466e-01
4.10627007e-01 -5.53791463e-01 6.06239438e-01 6.79965794e-01
-1.01580644e+00 2.37473220e-01 -9.70928848e-01 6.49704814e-01
-3.14653724e-01 -3.13625246e-01 7.80540466e-01 1.02911103e+00
-9.96429801e-01 1.02207434e+00 -8.48853588e-01 -1.55987486e-03
4.16903824e-01 3.22461933e-01 5.29047251e-02 -2.87642062e-01
-1.34311330e+00 9.29579794e-01 1.68665335e-01 4.87704188e-01
-1.26951277e+00 -7.13983715e-01 -3.73324662e-01 2.51743644e-01
7.69229591e-01 -4.13426667e-01 9.46400762e-01 -4.55901437e-02
-1.44973576e+00 -3.27063322e-01 3.65144968e-01 3.11624669e-02
3.18648398e-01 1.76261604e-01 1.67749487e-02 1.22514307e-01
1.53287007e-02 -2.93261558e-01 7.85316586e-01 -8.86724114e-01
-3.71322967e-02 -4.56192106e-01 5.31449139e-01 -5.49254656e-01
-3.19539100e-01 3.35999131e-01 -8.71145725e-02 -4.79723155e-01
3.88574779e-01 -6.90893829e-01 -5.76086879e-01 -1.78122088e-01
-6.81421578e-01 -1.14557333e-01 2.61224478e-01 -6.65051877e-01
1.26588953e+00 -1.76790905e+00 3.37467462e-01 1.23885369e+00
2.30625242e-01 -7.42550731e-01 1.57793805e-01 8.17119420e-01
-1.22805573e-01 3.56271446e-01 -5.44399440e-01 2.70001203e-01
2.97705650e-01 1.12953253e-01 -2.40177318e-01 7.44106948e-01
-5.02527766e-02 4.62979943e-01 -8.59095275e-01 -6.83584392e-01
-1.83452889e-01 1.23353243e-01 -7.49117494e-01 -1.64333984e-01
2.20389534e-02 4.30789515e-02 -5.49985230e-01 6.82147145e-01
5.42284846e-01 -2.52320945e-01 5.44070363e-01 -4.51252341e-01
1.74393654e-01 5.40654659e-02 -1.59162283e+00 1.16495800e+00
-4.79514182e-01 3.20974857e-01 1.64944082e-01 -9.72828269e-01
6.24496579e-01 2.60609388e-01 3.72636318e-01 6.32273331e-02
2.80270845e-01 5.98984845e-02 1.13600709e-01 -2.76181757e-01
2.86441874e-02 -2.51735777e-01 -3.66687715e-01 9.97736812e-01
7.63123035e-02 -2.55798757e-01 5.31887054e-01 6.54522598e-01
1.52829266e+00 -1.54110312e-01 -3.40413973e-02 -6.64793849e-01
1.23334780e-01 -4.76879478e-01 5.69329441e-01 8.95568013e-01
1.41546726e-01 -1.72814399e-01 1.44323790e+00 3.62565219e-02
-8.87822747e-01 -1.15196049e+00 -3.67016017e-01 1.01855719e+00
-4.50904310e-01 -1.40225410e-01 -7.28638589e-01 -6.94016397e-01
2.91944444e-01 5.26470900e-01 -7.47920692e-01 1.10818386e-01
-2.39692047e-01 -1.23865867e+00 5.05080581e-01 5.35058260e-01
1.57443956e-01 -3.78388375e-01 -6.18149899e-02 2.08028108e-01
-4.21677791e-02 -8.00776720e-01 -2.51368254e-01 -5.21036945e-02
-1.24619389e+00 -1.09910500e+00 -4.19431061e-01 1.35726810e-01
7.46248901e-01 -3.71129751e-01 8.06162894e-01 -3.86043012e-01
-1.04313530e-01 3.88877064e-01 1.04283877e-01 5.54140583e-02
-4.54486221e-01 -4.12664860e-01 2.13076279e-01 -2.51682941e-02
-2.01372802e-01 -9.93967712e-01 -1.85487926e-01 6.72218978e-01
-9.89633381e-01 -6.38587117e-01 4.97439712e-01 7.83271790e-01
-1.29490078e-03 5.65122962e-01 5.57552814e-01 -8.47527742e-01
9.35773551e-01 -6.27594471e-01 -1.18470705e+00 5.27785420e-01
-8.88182163e-01 4.17529613e-01 9.16635543e-02 -2.59429038e-01
-1.05653644e+00 -1.92925587e-01 6.59072280e-01 7.01235011e-02
6.64407432e-01 1.26977563e+00 -7.98566192e-02 -2.64384806e-01
5.87111771e-01 -6.51629090e-01 -1.38879418e-01 -4.38246965e-01
4.47530240e-01 2.72756904e-01 2.57563114e-01 -1.09721661e+00
7.20301032e-01 4.92244989e-01 9.21667397e-01 -2.51765668e-01
-4.66588259e-01 1.19009212e-01 -4.53827173e-01 -5.10490716e-01
3.45182836e-01 -3.27780724e-01 -1.31694460e+00 2.95513928e-01
-1.27163255e+00 -4.46491595e-03 -1.62709057e-02 6.21371388e-01
-3.33459526e-01 4.98811364e-01 -8.14749599e-01 -1.17035735e+00
4.17263657e-01 -1.11514640e+00 3.71816486e-01 -3.45655292e-01
-1.18419696e-02 -1.22889769e+00 3.10351610e-01 1.54879779e-01
4.96540606e-01 2.88250506e-01 1.03721476e+00 -3.76517504e-01
-8.16666543e-01 -5.38854301e-01 -3.74739200e-01 4.24714148e-01
-1.94204375e-01 4.94643897e-01 -6.90091074e-01 -1.20303854e-01
6.02230839e-02 2.17839450e-01 6.90954745e-01 5.54420173e-01
9.25291955e-01 -4.85965490e-01 -2.09525198e-01 2.38603447e-02
1.42637765e+00 -1.65048629e-01 6.14165425e-01 -1.44864321e-01
4.91218567e-01 1.08766651e+00 4.74524647e-01 5.95639825e-01
5.57709634e-02 5.97852409e-01 5.99125445e-01 5.18633544e-01
5.45439959e-01 7.11619994e-03 3.55997711e-01 8.28411996e-01
-4.70616847e-01 -9.18232352e-02 -1.05919552e+00 6.20466113e-01
-1.92477822e+00 -9.70735848e-01 -1.45407647e-01 2.06491709e+00
8.45816314e-01 3.05656016e-01 1.53245240e-01 1.91800475e-01
1.03477800e+00 -1.55566663e-01 1.55205298e-02 -3.73889655e-01
2.40353763e-01 1.56517163e-01 9.59438503e-01 6.52430117e-01
-6.14547431e-01 2.13882968e-01 7.52041340e+00 8.07092667e-01
-1.85553730e-01 4.08805579e-01 5.25531292e-01 4.80079316e-02
-4.39917624e-01 5.18729329e-01 -3.64400029e-01 2.95460850e-01
1.21933591e+00 -1.12042449e-01 5.91050029e-01 6.97926283e-01
2.89514780e-01 -4.48320061e-01 -1.20589530e+00 2.58384556e-01
-5.39488137e-01 -1.04303432e+00 -4.98428583e-01 3.34843278e-01
6.86696589e-01 -3.06617796e-01 -2.49833204e-02 -3.26517910e-01
8.47282410e-01 -7.87069201e-01 4.57168937e-01 8.78389239e-01
3.68469894e-01 -8.37325335e-01 7.31657028e-01 -6.61549941e-02
-7.20965862e-01 -3.47620696e-01 -3.24697733e-01 -1.74827784e-01
3.48964214e-01 1.63996959e+00 -9.77203727e-01 2.77154207e-01
5.79572558e-01 4.45422500e-01 -1.95503742e-01 9.86380458e-01
-3.60980928e-01 8.24261010e-01 -6.57648563e-01 -2.91878343e-01
-1.43364295e-01 -3.76622736e-01 7.40139127e-01 1.20078492e+00
1.80508047e-01 -2.24427387e-01 -4.98493910e-01 1.34988582e+00
2.09044844e-01 -2.14161187e-01 -6.66842580e-01 -1.57229722e-01
5.75058699e-01 1.20845103e+00 -9.26022947e-01 -1.34825721e-01
-1.03018776e-01 2.55391486e-02 -1.23402342e-01 5.62895477e-01
-6.23325467e-01 -7.19821572e-01 2.63486147e-01 -5.93231842e-02
3.69763911e-01 -3.71434152e-01 -4.26801145e-01 -7.15259910e-01
-4.29073982e-02 -6.93498909e-01 3.67943823e-01 -5.06533802e-01
-1.52790165e+00 -5.97499274e-02 7.44409859e-01 -6.89714968e-01
-3.75012994e-01 -9.36810791e-01 -5.92405498e-01 7.91644156e-01
-9.36180651e-01 -6.41789913e-01 3.47199976e-01 3.18127394e-01
-5.01065910e-01 1.04749143e-01 3.11587065e-01 3.18429798e-01
-8.44869494e-01 1.94925815e-01 1.16338275e-01 2.54001338e-02
2.47445673e-01 -1.14750540e+00 8.74431804e-02 1.16902304e+00
-3.90762955e-01 9.17172730e-01 9.56221640e-01 -8.58551919e-01
-1.47673893e+00 -5.88195860e-01 5.39713442e-01 -4.93193030e-01
1.30269861e+00 -3.20885420e-01 -4.02720004e-01 6.14362419e-01
-2.85298288e-01 -1.20113064e-02 4.16698813e-01 5.27472198e-01
-7.18471229e-01 -3.83128792e-01 -1.25988984e+00 5.08334339e-01
1.02684057e+00 -5.97675204e-01 -2.52033800e-01 5.01683652e-01
5.28797030e-01 1.88634872e-01 -1.56185079e+00 4.09275770e-01
5.13077080e-01 -8.18574250e-01 9.07270312e-01 -5.12428403e-01
4.09723103e-01 -3.18505496e-01 -6.04872227e-01 -9.83562469e-01
-3.72141659e-01 -7.21009135e-01 -2.03143120e-01 1.08507061e+00
5.48937261e-01 -9.18480039e-01 2.56212950e-01 7.52773523e-01
6.11956753e-02 -4.31851536e-01 -1.13222206e+00 -9.69128311e-01
-1.81665927e-01 -7.65945077e-01 7.40393460e-01 9.03472185e-01
4.00228322e-01 -6.03998117e-02 -1.26441598e-01 3.42777371e-01
1.30762482e+00 -1.35918722e-01 1.96931630e-01 -1.40811098e+00
-7.98603594e-01 -4.93842721e-01 -5.84898293e-01 -4.82798159e-01
6.78870454e-02 -7.11930096e-01 8.99752136e-03 -1.12416220e+00
5.75881481e-01 -3.86487156e-01 -3.31610590e-01 4.66530204e-01
2.50691265e-01 1.23702981e-01 -1.54708430e-01 1.56744495e-02
-2.03850836e-01 1.76503539e-01 9.90116239e-01 -9.56959128e-02
2.77280301e-01 2.59404898e-01 -6.32728636e-01 6.78059995e-01
5.02262652e-01 -9.41437125e-01 -3.56863797e-01 7.47268051e-02
1.25598514e+00 4.38046306e-01 6.59893870e-01 -3.56589258e-01
3.33328068e-01 -4.61039811e-01 -1.72309086e-01 -5.09626269e-01
1.97305158e-01 -7.24524915e-01 2.93595880e-01 5.02019227e-01
-3.14586908e-01 9.36280340e-02 -1.56901762e-01 6.96472883e-01
1.24952964e-01 -6.62543893e-01 4.23332572e-01 7.69146606e-02
4.52263691e-02 8.78576562e-02 -4.71386313e-01 -1.25536129e-01
4.50291306e-01 3.01432669e-01 -6.80366993e-01 -5.14175057e-01
-9.21202183e-01 9.39302668e-02 8.36943388e-02 -4.52840209e-01
6.92013323e-01 -1.30079782e+00 -7.13660479e-01 -4.59326029e-01
-3.88232082e-01 -4.24876988e-01 4.48191836e-02 1.39739299e+00
-3.77148688e-01 2.57249683e-01 -1.42346937e-02 -4.54086572e-01
-6.58231020e-01 4.06185329e-01 2.88794279e-01 -3.46083969e-01
1.19608790e-01 6.42877400e-01 -2.06585482e-01 -3.18321675e-01
1.35629009e-02 -4.13268954e-01 1.21722132e-01 1.19641960e-01
2.50277966e-01 1.29415572e+00 1.60927884e-02 8.13865066e-02
-4.24585521e-01 4.72205400e-01 2.20450968e-01 -8.31296206e-01
1.44519329e+00 -6.41818196e-02 -1.08834457e+00 7.07262218e-01
1.08886552e+00 -7.42367352e-04 -8.80459368e-01 -1.75548479e-01
2.37923190e-01 -4.19888824e-01 3.63283157e-01 -6.57694757e-01
-1.16749120e+00 8.51709247e-01 2.39163175e-01 4.75791425e-01
8.69245231e-01 4.42397781e-02 -1.17727891e-01 7.97949910e-01
4.50044990e-01 -1.21384680e+00 1.06165662e-01 4.12636489e-01
8.78844857e-01 -4.98521268e-01 5.34137607e-01 -5.60349584e-01
-6.96902499e-02 1.15454590e+00 -9.11296010e-02 -2.09930278e-02
1.20776200e+00 4.23114061e-01 -8.12976658e-01 -5.26471734e-01
-7.14847863e-01 3.29705507e-01 1.02746703e-01 2.80689329e-01
1.66431695e-01 2.57127821e-01 -2.20925599e-01 3.94214630e-01
1.43993646e-01 -6.28618598e-02 1.04701555e+00 8.83929491e-01
-1.77216128e-01 -9.02567983e-01 -6.42558873e-01 5.65419614e-01
-5.09336531e-01 -2.17383295e-01 1.97684485e-03 5.32124519e-01
-4.13068861e-01 1.04846334e+00 -1.76979244e-01 -4.42688435e-01
7.05613866e-02 -6.15225220e-03 6.55188322e-01 -3.40375721e-01
-1.68601960e-01 2.33745463e-02 4.11736429e-01 -8.25615644e-01
-6.01679742e-01 -1.00748086e+00 -6.68631434e-01 -6.75895154e-01
-6.67435884e-01 7.64490962e-02 1.06980109e+00 9.15627360e-01
-5.17609008e-02 5.25088429e-01 8.96344781e-01 -5.37042141e-01
-9.16344762e-01 -1.15562189e+00 -9.83658493e-01 -4.33651865e-01
-2.07546316e-02 -9.53741372e-01 -7.94896960e-01 -3.33518833e-01] | [7.420628070831299, 4.90974760055542] |
ff7ef679-d34c-43a6-beaf-61e9649e22a4 | explanation-based-weakly-supervised-learning | 2006.09562 | null | https://arxiv.org/abs/2006.09562v1 | https://arxiv.org/pdf/2006.09562v1.pdf | Explanation-based Weakly-supervised Learning of Visual Relations with Graph Networks | Visual relationship detection is fundamental for holistic image understanding. However, localizing and classifying (subject, predicate, object) triplets constitutes a hard learning objective due to the combinatorial explosion of possible relationships, their long-tail distribution in natural images, and an expensive annotation process. This paper introduces a novel weakly-supervised method for visual relationship detection that relies only on image-level predicate annotations. A graph neural network is trained to classify the predicates in an image from the graph representation of all objects, implicitly encoding an inductive bias for pairwise relationships. We then frame relationship detection as the explanation of such a predicate classifier, i.e. we reconstruct a complete relationship by recovering the subject and the object of a predicted predicate. Using this novel technique and minimal labels, we present comparable results to recent fully-supervised and weakly-supervised methods on three diverse and challenging datasets: HICO-DET for human-object interaction, Visual Relationship Detection for generic object-to-object relationships, and UnRel for unusual relationships. | ['Hossein Azizpour', 'Federico Baldassarre', 'Josephine Sullivan', 'Kevin Smith'] | 2020-06-16 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6336_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123730613.pdf | eccv-2020-8 | ['visual-relationship-detection'] | ['computer-vision'] | [ 6.86330318e-01 5.25560915e-01 -4.87172961e-01 -5.39243877e-01
-3.27418536e-01 -7.33815134e-01 6.52481139e-01 6.42904341e-01
-2.30729692e-02 5.39718330e-01 1.09544098e-01 -1.75859213e-01
-6.25928715e-02 -3.64428192e-01 -1.02447808e+00 -5.26070416e-01
-3.21765721e-01 9.98884141e-01 4.10405636e-01 8.24728757e-02
-1.58347487e-02 4.89300072e-01 -1.75750494e+00 6.38031602e-01
1.61579013e-01 9.85491395e-01 2.20311105e-01 6.93705618e-01
2.18655333e-01 1.12563956e+00 -2.65043110e-01 -4.12384629e-01
-1.78999454e-02 -5.67998946e-01 -1.11133695e+00 5.14855683e-01
8.92358780e-01 1.30707815e-01 -2.26831809e-01 9.60437775e-01
7.92090967e-02 6.11104332e-02 8.79124105e-01 -1.86090970e+00
-7.91973948e-01 5.10271847e-01 -7.38870800e-01 1.47032708e-01
7.00471818e-01 -1.37185410e-01 1.74840915e+00 -7.77456403e-01
1.10344028e+00 1.24877024e+00 3.25208396e-01 2.00293377e-01
-1.57705224e+00 -2.74463058e-01 2.73785383e-01 5.37989140e-01
-1.44489038e+00 -2.87184656e-01 8.39333296e-01 -7.46394157e-01
1.19464779e+00 2.71052361e-01 7.03687668e-01 9.59228218e-01
-4.73574549e-02 8.52796197e-01 8.46180797e-01 -4.72923607e-01
-4.86391783e-02 2.52642304e-01 2.55956858e-01 9.73504841e-01
1.57806501e-01 3.85041647e-02 -8.04423034e-01 -1.36627123e-01
4.22323734e-01 -2.38065571e-01 -3.25355411e-01 -1.19236374e+00
-1.22372305e+00 3.95795882e-01 7.28025496e-01 -2.18694285e-02
-2.35011280e-01 1.73832104e-01 2.66936600e-01 1.46348730e-01
3.45581084e-01 2.81380087e-01 -4.65688974e-01 3.45321566e-01
-4.86002117e-01 1.73239365e-01 9.04571235e-01 1.22224164e+00
9.24902439e-01 -5.59825480e-01 -2.35783294e-01 5.61154187e-01
3.08786273e-01 3.75562534e-02 -2.58857995e-01 -7.17613101e-01
3.90127987e-01 8.61456275e-01 -8.20844024e-02 -1.31270587e+00
-5.32292128e-01 -2.71374822e-01 -6.00476444e-01 1.02856345e-01
3.89124185e-01 5.34278691e-01 -8.37156713e-01 1.61385286e+00
4.68418568e-01 -1.60617203e-01 -5.53204790e-02 9.07498896e-01
1.24303901e+00 1.62166759e-01 2.64725894e-01 -2.39041507e-01
1.43339610e+00 -1.03574300e+00 -5.64787269e-01 -4.41271663e-01
4.71321195e-01 -5.95610559e-01 8.72022629e-01 5.57414666e-02
-8.79063368e-01 -4.70059663e-01 -7.71924973e-01 -4.27024096e-01
-5.02059937e-01 7.01470524e-02 9.14711237e-01 1.36746438e-02
-9.46919084e-01 3.31867874e-01 -4.02043283e-01 -6.13445520e-01
9.52422440e-01 4.40133810e-01 -7.95135140e-01 -2.55546067e-02
-4.88079160e-01 7.69854963e-01 8.75323176e-01 -1.57080684e-02
-8.18693697e-01 -3.49975258e-01 -1.02594447e+00 1.15271524e-01
5.85220575e-01 -7.85253882e-01 7.09257960e-01 -1.16864717e+00
-6.07705474e-01 1.64834249e+00 -3.04020703e-01 -3.68867934e-01
2.37199470e-01 3.25887613e-02 -7.93601200e-02 3.28388244e-01
1.75135404e-01 9.88248765e-01 8.57648075e-01 -1.81456339e+00
-4.86278653e-01 -3.44167382e-01 6.03105240e-02 3.64313006e-01
1.48884118e-01 9.47505683e-02 -5.40214121e-01 -3.36932272e-01
5.33006489e-01 -9.18467283e-01 1.10870786e-01 3.23399335e-01
-7.72151887e-01 -5.93176603e-01 1.02299154e+00 -4.93535638e-01
4.72259104e-01 -2.23763084e+00 2.67854661e-01 2.12908015e-01
6.33595586e-01 -7.96822980e-02 -1.19551145e-01 2.65746742e-01
-5.67457616e-01 -7.52172396e-02 -3.09724182e-01 -4.10123199e-01
-5.09545468e-02 5.71848571e-01 -2.60835856e-01 6.12149954e-01
5.50232172e-01 1.17138326e+00 -1.15725803e+00 -9.69697118e-01
1.55809298e-01 2.89363563e-01 -4.21711922e-01 4.22705442e-01
-4.75183070e-01 5.48885822e-01 4.95489985e-02 9.35991406e-01
4.40664530e-01 -7.38782525e-01 3.92386109e-01 -7.30866849e-01
2.96661407e-01 6.54181466e-02 -7.63197839e-01 1.40115845e+00
7.87731633e-02 8.96333456e-01 -1.93636999e-01 -1.37112498e+00
8.53013158e-01 1.69946507e-01 4.70755190e-01 -6.38695002e-01
5.14079034e-02 -6.39903098e-02 -7.81958830e-03 -6.94843173e-01
1.64683193e-01 -7.10140318e-02 1.77251622e-01 1.53522953e-01
3.47816139e-01 -8.16316530e-02 3.27561289e-01 6.07554018e-01
9.26446617e-01 5.58023155e-01 5.91341078e-01 -9.97986868e-02
4.08560872e-01 9.44267139e-02 3.14326733e-01 7.67956793e-01
-2.48570800e-01 7.79380858e-01 8.80735576e-01 -5.10883570e-01
-8.39378536e-01 -1.21631265e+00 3.50064561e-02 1.19964302e+00
6.84320450e-01 -4.92096364e-01 -1.92327090e-02 -9.24296081e-01
1.12304650e-01 4.81833547e-01 -7.27802038e-01 -9.88387689e-02
-4.55929667e-01 -1.65827870e-01 3.41175799e-03 4.94042307e-01
1.62111506e-01 -1.27278483e+00 -3.96147519e-01 -2.75405228e-01
-3.44028831e-01 -1.70364261e+00 -2.43687943e-01 5.52132368e-01
-3.87234747e-01 -1.49562120e+00 9.89977717e-02 -1.24636889e+00
1.03584993e+00 1.76305413e-01 1.64507663e+00 2.77061701e-01
-5.21842718e-01 6.56996787e-01 -1.26418307e-01 -1.88872859e-01
-1.51474953e-01 -6.17401481e-01 -1.81265697e-01 1.49296343e-01
1.88192174e-01 -3.96223545e-01 -3.02993029e-01 2.38349348e-01
-4.24310684e-01 2.56066889e-01 4.63485420e-01 8.76443267e-01
1.05266094e+00 9.32959765e-02 -1.07267730e-01 -9.37858343e-01
-9.73686017e-03 -3.38524789e-01 -4.14901495e-01 6.47915363e-01
-1.35166302e-01 1.09630957e-01 2.97183394e-01 -4.82147545e-01
-9.08113837e-01 6.08585000e-01 5.64430296e-01 -4.52878058e-01
-3.46384168e-01 1.58726990e-01 -3.20318490e-01 -1.33543015e-01
6.43250406e-01 -4.40073013e-03 -2.03926921e-01 -8.41168910e-02
6.91445470e-01 4.33066972e-02 8.38561714e-01 -4.51125950e-01
8.66469800e-01 7.76826084e-01 4.63542044e-01 -7.65586376e-01
-1.31069720e+00 -8.86191547e-01 -1.12593913e+00 -3.70730579e-01
1.13452125e+00 -8.81336868e-01 -1.00129056e+00 -2.21996948e-01
-1.38967884e+00 -2.43154407e-01 -4.01589841e-01 4.17453170e-01
-8.00616801e-01 3.75989795e-01 -5.78079410e-02 -5.75042009e-01
3.62953395e-02 -6.78601921e-01 1.32086289e+00 3.89116295e-02
-4.01896268e-01 -1.06524217e+00 -1.57487109e-01 6.30811155e-01
-3.89097750e-01 5.84893346e-01 1.10495067e+00 -7.15596795e-01
-7.52449930e-01 -7.02780634e-02 -8.61488461e-01 1.41617313e-01
1.71643160e-02 4.19038795e-02 -8.91942739e-01 -6.49006367e-02
-4.84432697e-01 -7.20540524e-01 8.58052731e-01 6.37380779e-02
1.29185438e+00 -2.47241959e-01 -7.26800084e-01 4.23177361e-01
1.24080384e+00 -1.88268721e-01 1.57543808e-01 -1.11360475e-02
1.28575206e+00 1.09600604e+00 6.85387969e-01 1.61640376e-01
4.64564770e-01 7.49751747e-01 7.69484103e-01 -3.51813167e-01
-4.96714056e-01 -2.45202854e-01 9.53024440e-03 -5.41877449e-02
-1.17189921e-01 -4.66276795e-01 -1.06498671e+00 6.31027520e-01
-2.14646554e+00 -9.80851710e-01 -5.72093129e-01 1.93935502e+00
6.95490003e-01 1.89179480e-01 1.56521112e-01 -6.75443411e-02
8.34694505e-01 9.11698639e-02 -3.50371242e-01 -1.65642425e-01
-3.64303589e-01 -3.07526849e-02 2.73660094e-01 3.38618279e-01
-1.34754479e+00 1.07503390e+00 6.06008673e+00 3.82111043e-01
-6.07286990e-01 -6.80816025e-02 5.16532779e-01 2.02989310e-01
-1.28719077e-01 3.86686146e-01 -6.60364270e-01 -2.18651578e-01
2.66313493e-01 1.12228297e-01 2.47798041e-01 7.54372954e-01
-2.30123982e-01 -2.85307288e-01 -1.68354380e+00 1.20649302e+00
5.36534786e-01 -1.18640649e+00 1.90035462e-01 -4.26643416e-02
5.37303269e-01 -2.12751895e-01 -1.60137430e-01 -1.01141118e-01
1.74519822e-01 -1.06756139e+00 8.78967583e-01 4.96797621e-01
5.34840345e-01 -4.02359307e-01 5.21092415e-01 1.86947718e-01
-1.39651787e+00 1.52597055e-01 -1.42074466e-01 -1.10347100e-01
8.94904286e-02 4.47176725e-01 -1.10642159e+00 4.64564562e-01
7.94338167e-01 1.15188313e+00 -8.43563855e-01 8.17243457e-01
-5.09083509e-01 3.03959697e-01 -3.32105011e-01 1.98563233e-01
-1.08379439e-01 1.46873510e-02 5.62111259e-01 1.07350636e+00
-4.96595800e-01 -5.11374958e-02 5.84364474e-01 9.88908470e-01
-1.05820440e-01 -1.21254824e-01 -8.74935985e-01 -9.24045220e-02
-4.56870534e-02 1.35753906e+00 -1.09623504e+00 -2.67398864e-01
-3.77847314e-01 1.02061033e+00 7.56802082e-01 3.61014605e-01
-7.61316240e-01 6.75537884e-02 2.52118349e-01 1.16512880e-01
3.95837218e-01 -2.22260624e-01 -2.27928013e-01 -8.62605512e-01
2.54984349e-01 -5.58841705e-01 7.60573328e-01 -1.04591012e+00
-1.33970892e+00 3.54442507e-01 1.69298261e-01 -1.14274585e+00
-2.54330840e-02 -7.58357644e-01 -3.19721282e-01 4.31349009e-01
-1.49318969e+00 -1.67638731e+00 -5.89665234e-01 6.55819416e-01
1.47052988e-01 1.49863735e-01 7.34860241e-01 -3.34675796e-02
-1.95261121e-01 2.78255254e-01 -7.54674256e-01 2.88650692e-02
6.54369056e-01 -1.44085896e+00 -2.70850863e-02 6.37931406e-01
7.73721039e-01 4.46490854e-01 8.84079516e-01 -6.46536469e-01
-1.22619462e+00 -1.00235450e+00 1.21103394e+00 -8.99603844e-01
8.98444951e-01 -7.88018048e-01 -9.29449916e-01 9.36126649e-01
2.28228211e-01 6.92376137e-01 6.10412598e-01 2.58168995e-01
-7.37711072e-01 8.57182369e-02 -7.74046779e-01 6.24240220e-01
1.59333313e+00 -8.34715307e-01 -4.90834147e-01 9.77615416e-01
6.91271722e-01 -4.51622993e-01 -6.41805232e-01 5.31323433e-01
4.29821610e-01 -9.77912009e-01 1.20995665e+00 -7.38439143e-01
4.42316502e-01 -5.28770208e-01 -1.17686428e-01 -6.04680538e-01
-1.86534211e-01 -3.83913189e-01 -3.11046392e-01 1.10499978e+00
4.67278779e-01 -1.57058105e-01 7.31729984e-01 3.51745337e-01
1.15515389e-01 -6.12253189e-01 -6.93674266e-01 -7.17926323e-01
-7.43764222e-01 -2.07061276e-01 -1.68656737e-01 1.03199923e+00
-3.49141210e-02 7.14427352e-01 -3.08760464e-01 4.62418318e-01
1.01743209e+00 5.27606964e-01 8.11623812e-01 -1.17838514e+00
-5.40280104e-01 -1.69094831e-01 -9.20656741e-01 -8.64354730e-01
5.64764440e-01 -1.03688967e+00 2.89982021e-01 -1.63798380e+00
6.48055494e-01 -3.08736384e-01 7.35695427e-03 9.20984268e-01
-1.67608619e-01 5.82610726e-01 5.93316182e-02 3.06050509e-01
-1.20384896e+00 4.19730753e-01 1.17469883e+00 -4.16502625e-01
-5.73675847e-03 -2.63085753e-01 -4.73460346e-01 8.76027703e-01
4.42389905e-01 -5.13132274e-01 -4.87647831e-01 -1.11020982e-01
5.39310813e-01 -6.50842264e-02 1.06462169e+00 -5.48982084e-01
3.13959002e-01 -2.03047737e-01 3.82301897e-01 -8.16054165e-01
3.74488413e-01 -1.15064347e+00 8.29064324e-02 2.43682563e-01
-3.99784595e-01 -2.05293685e-01 1.50473770e-02 9.32639062e-01
-3.34018379e-01 1.37724817e-01 5.75675488e-01 1.42426612e-02
-1.00307107e+00 2.24324957e-01 2.08815336e-01 1.00437537e-01
1.29662716e+00 -2.48687893e-01 -5.62072396e-01 -3.03643793e-01
-1.10583937e+00 2.58636296e-01 3.38981628e-01 5.03510356e-01
7.74055719e-01 -1.18002522e+00 -5.40743649e-01 3.45107988e-02
7.81857491e-01 6.95773289e-02 -6.36491701e-02 7.74667978e-01
-2.99604774e-01 9.07844231e-02 -1.10562995e-01 -1.12983310e+00
-1.86432588e+00 7.65297055e-01 1.54873192e-01 -2.25838833e-02
-7.77986646e-01 1.18236434e+00 6.14023685e-01 -2.60583341e-01
2.57635295e-01 -8.50386545e-02 -4.02997524e-01 1.18236884e-01
1.92598160e-02 -1.85851410e-01 -2.64188647e-01 -1.18999720e+00
-7.87073553e-01 6.24967694e-01 5.76093281e-03 3.12642664e-01
1.17278421e+00 1.10860011e-02 -6.03721201e-01 5.88475466e-01
1.37110043e+00 -2.37000614e-01 -1.08335280e+00 -4.02672827e-01
1.70666873e-01 -4.60593164e-01 -3.27490896e-01 -5.60215592e-01
-7.29954660e-01 6.91379547e-01 3.23778152e-01 2.95581937e-01
9.43641424e-01 9.43789363e-01 -2.45188363e-02 4.64829087e-01
1.69351190e-01 -6.15069389e-01 5.85043669e-01 3.27226013e-01
1.06064570e+00 -1.69635570e+00 3.66983205e-01 -1.09970140e+00
-7.88544774e-01 9.58351195e-01 7.51064777e-01 8.63861367e-02
5.53219378e-01 -6.20634146e-02 -2.85522282e-01 -7.25252688e-01
-7.30648994e-01 -6.15916610e-01 7.88882434e-01 8.16137373e-01
4.34699565e-01 9.21877548e-02 4.29902822e-02 -1.53127849e-01
1.83519197e-03 -5.36655843e-01 5.56305647e-02 8.52643371e-01
-1.21729061e-01 -9.02836978e-01 -6.10218905e-02 3.35586965e-01
-9.53773558e-02 -2.11251974e-01 -9.19035137e-01 7.16805756e-01
2.96467692e-01 9.30330694e-01 2.96288818e-01 -1.11880019e-01
1.93816602e-01 -1.86360434e-01 7.50853002e-01 -8.56909454e-01
-2.64126420e-01 -1.99307650e-01 3.91065061e-01 -7.51423240e-01
-1.02759242e+00 -6.05518758e-01 -1.48355055e+00 3.35879564e-01
-4.13120776e-01 -2.44352490e-01 2.97265589e-01 1.10538363e+00
1.52936876e-01 5.06360471e-01 2.90254712e-01 -7.63701499e-01
2.93235630e-01 -3.29167753e-01 -4.36659843e-01 1.05470800e+00
3.74124378e-01 -9.42225814e-01 -4.46155161e-01 4.59903240e-01] | [10.27157211303711, 1.6368355751037598] |
a4d8fb8b-6f7a-4e2b-beb4-91cedb03067a | hifa-high-fidelity-text-to-3d-with-advanced | 2305.18766 | null | https://arxiv.org/abs/2305.18766v2 | https://arxiv.org/pdf/2305.18766v2.pdf | HiFA: High-fidelity Text-to-3D with Advanced Diffusion Guidance | Automatic text-to-3D synthesis has achieved remarkable advancements through the optimization of 3D models. Existing methods commonly rely on pre-trained text-to-image generative models, such as diffusion models, providing scores for 2D renderings of Neural Radiance Fields (NeRFs) and being utilized for optimizing NeRFs. However, these methods often encounter artifacts and inconsistencies across multiple views due to their limited understanding of 3D geometry. To address these limitations, we propose a reformulation of the optimization loss using the diffusion prior. Furthermore, we introduce a novel training approach that unlocks the potential of the diffusion prior. To improve 3D geometry representation, we apply auxiliary depth supervision for NeRF-rendered images and regularize the density field of NeRFs. Extensive experiments demonstrate the superiority of our method over prior works, resulting in advanced photo-realism and improved multi-view consistency. | ['Junzhe Zhu', 'Peiye Zhuang'] | 2023-05-30 | null | null | null | null | ['text-to-3d'] | ['computer-vision'] | [ 2.85361707e-01 9.45819169e-02 4.82874960e-02 -4.05937374e-01
-6.36062801e-01 -4.90947247e-01 9.14659619e-01 -3.36257637e-01
3.11549827e-02 3.82894307e-01 6.20570719e-01 -1.59965053e-01
1.43777877e-01 -9.38407362e-01 -7.60841727e-01 -5.97324610e-01
3.79262060e-01 1.10643998e-01 5.19776717e-03 -1.29384831e-01
1.88256696e-01 6.58675909e-01 -1.38715875e+00 8.84089619e-02
1.05893993e+00 9.31848705e-01 2.42122039e-01 6.03890717e-01
-3.75902466e-02 7.50849664e-01 -4.82149780e-01 -4.89011973e-01
5.15156627e-01 -3.58021170e-01 -3.74331146e-01 2.46649936e-01
6.63571715e-01 -8.07191908e-01 -5.52826703e-01 9.48944271e-01
5.42210281e-01 2.00600073e-01 8.23219419e-01 -1.02899504e+00
-1.01822805e+00 -1.09096117e-01 -7.53454387e-01 -1.26495451e-01
3.73386025e-01 6.99330047e-02 9.28393960e-01 -9.24485505e-01
6.17145181e-01 1.42032826e+00 5.99367857e-01 5.74143410e-01
-1.39656126e+00 -4.56457138e-01 3.30590487e-01 -1.51973203e-01
-1.26472211e+00 -2.62675941e-01 1.04621732e+00 -4.95806038e-01
1.05687702e+00 1.13473184e-01 7.35160708e-01 1.11835563e+00
1.17447592e-01 8.19435418e-01 1.24373412e+00 -3.28256816e-01
1.04058132e-01 9.12338030e-03 -5.51933408e-01 7.48487949e-01
-6.69711530e-02 2.43668169e-01 -5.59720874e-01 -3.57900485e-02
1.34031820e+00 -1.49941295e-01 -3.83314997e-01 -5.06363511e-01
-8.27959239e-01 7.80001581e-01 6.38910532e-01 -2.55699068e-01
-3.74944538e-01 2.07097456e-01 -1.13787733e-01 -2.83018917e-01
1.05002081e+00 2.59026170e-01 -4.33242060e-02 4.54664193e-02
-8.53831768e-01 3.16111147e-01 3.76535147e-01 8.67153943e-01
8.95889759e-01 2.36765131e-01 -1.35959297e-01 1.06308889e+00
6.01836503e-01 6.90388024e-01 2.13135034e-02 -1.30926073e+00
4.04831529e-01 6.22613251e-01 8.44143927e-02 -9.90408957e-01
-1.09485090e-01 -3.62395704e-01 -8.06544662e-01 4.56097692e-01
2.87042223e-02 2.26451278e-01 -1.26956105e+00 1.70808709e+00
5.65620363e-01 1.39099702e-01 -1.71537194e-02 9.79145050e-01
6.75858617e-01 6.64117873e-01 -1.60875261e-01 1.87649429e-01
8.39025676e-01 -1.05044448e+00 -6.00440800e-01 -2.46277213e-01
2.90787965e-01 -7.44773507e-01 1.04066312e+00 1.71756744e-01
-1.39267743e+00 -3.48371595e-01 -9.71899569e-01 -4.77298617e-01
2.18750592e-02 -5.45668304e-02 6.64600015e-01 5.62124014e-01
-1.14121568e+00 5.54223359e-01 -1.01778281e+00 2.25623380e-02
6.02195024e-01 -1.24640400e-02 -7.60557726e-02 -3.25818270e-01
-8.39254975e-01 8.60296786e-01 -1.45701006e-01 -4.70032655e-02
-1.02413619e+00 -1.01434588e+00 -1.04947913e+00 -6.41564429e-02
-4.68924493e-02 -1.10585380e+00 1.09643233e+00 -8.16017270e-01
-1.67524254e+00 8.75232399e-01 -1.85265392e-01 -6.80373609e-02
5.62628567e-01 -2.55548477e-01 4.74993475e-02 2.92296320e-01
9.42838490e-02 9.10304308e-01 8.26384008e-01 -1.62628877e+00
-3.09288919e-01 -3.34909737e-01 3.21168333e-01 5.91247201e-01
-3.48907858e-01 -4.83144701e-01 -7.44982839e-01 -7.95955420e-01
2.59864986e-01 -6.71590030e-01 -3.31601441e-01 3.70581150e-01
-4.76045400e-01 2.70348221e-01 6.31036103e-01 -6.75129652e-01
9.71761942e-01 -2.09415030e+00 3.09498906e-01 9.13336203e-02
3.76555055e-01 1.86908673e-02 -1.85278282e-01 2.86047786e-01
1.49873719e-01 1.36852607e-01 -3.03762138e-01 -8.04234803e-01
2.09838729e-02 3.92318547e-01 -3.74134958e-01 3.19882900e-01
3.85434896e-01 8.48853528e-01 -8.11475575e-01 -2.66421169e-01
5.66586435e-01 1.13967705e+00 -8.10538113e-01 3.10881734e-01
-3.71580392e-01 6.99248552e-01 -4.52475697e-01 4.92209077e-01
1.00488079e+00 -5.09989977e-01 -1.78431049e-01 -4.34614748e-01
-7.41751939e-02 3.41617435e-01 -8.75542223e-01 1.89844203e+00
-7.28597343e-01 4.19574052e-01 1.66401151e-03 -5.08784592e-01
7.37414777e-01 3.62927280e-02 5.78307629e-01 -8.77750993e-01
2.94688027e-02 1.19532242e-01 -4.71454352e-01 -2.14562312e-01
5.70091605e-01 -2.55132228e-01 5.09745896e-01 2.76022643e-01
-1.21492438e-01 -7.70736217e-01 -1.82189032e-01 2.98082143e-01
7.45601356e-01 7.34688938e-01 -1.22124434e-01 1.34464756e-01
2.20265135e-01 -2.16156423e-01 4.13400292e-01 4.68492031e-01
2.78358251e-01 1.22912359e+00 2.62589335e-01 -2.37089396e-01
-1.33966506e+00 -1.17738354e+00 -1.65365532e-01 4.01725352e-01
2.34058589e-01 -4.21126395e-01 -7.48937309e-01 -6.49054229e-01
-9.71591696e-02 8.15391600e-01 -5.77187002e-01 -3.55184898e-02
-3.45099390e-01 -8.57835650e-01 4.19529289e-01 5.86453259e-01
6.40384138e-01 -3.27394187e-01 -4.93321329e-01 -1.21321101e-02
-2.97801495e-01 -1.39668500e+00 -4.47101057e-01 -1.39189079e-01
-9.98326778e-01 -7.30976164e-01 -1.11653888e+00 -2.72865206e-01
8.13180685e-01 5.24269462e-01 1.26090038e+00 8.85762833e-03
-2.58659750e-01 5.85223436e-01 -1.39298901e-01 -3.04607838e-01
-5.01433253e-01 -2.32849777e-01 -2.31570035e-01 -7.75389671e-02
-2.07656488e-01 -9.13263619e-01 -1.05320382e+00 3.89246285e-01
-1.06166124e+00 5.78606904e-01 4.43402231e-01 6.30146503e-01
7.92996049e-01 -2.74296939e-01 4.60296273e-02 -6.48919940e-01
2.54734218e-01 -2.03664660e-01 -6.66650832e-01 5.08398525e-02
-8.28774631e-01 1.33308366e-01 4.10488069e-01 -3.16715211e-01
-1.46747386e+00 -5.46528473e-02 -2.65199304e-01 -7.84300566e-01
5.41924536e-02 1.50231972e-01 -1.40099734e-01 -2.72088408e-01
5.26428699e-01 1.28246114e-01 -7.72423819e-02 -3.98681611e-01
5.07523835e-01 2.33505264e-01 3.62680852e-01 -6.02588475e-01
8.94968271e-01 8.36490333e-01 2.25049019e-01 -7.23440707e-01
-1.10121977e+00 -1.32305413e-01 -3.86742622e-01 -2.38845408e-01
9.19024706e-01 -1.28891170e+00 -2.48975843e-01 5.61639607e-01
-1.21180379e+00 -6.06881142e-01 -2.86174178e-01 4.55381811e-01
-6.05413735e-01 5.97535431e-01 -5.78361571e-01 -8.73970270e-01
-1.31090075e-01 -1.09023547e+00 1.55700111e+00 2.34634608e-01
7.51913711e-02 -1.34818244e+00 1.24035537e-01 5.13667166e-01
4.72234368e-01 3.25185359e-01 1.03257453e+00 4.22372729e-01
-9.79086101e-01 1.53574973e-01 -4.80876029e-01 6.38634861e-01
1.13450751e-01 -2.48621311e-03 -1.34637380e+00 -7.02461526e-02
2.84849741e-02 -3.15704137e-01 8.11219931e-01 6.64920509e-01
1.10857081e+00 -1.08855598e-01 -2.60034978e-01 1.21333182e+00
1.44450283e+00 -1.24736167e-01 8.95704210e-01 2.00140476e-01
1.09593856e+00 5.51420450e-01 2.54605770e-01 5.14211595e-01
6.94908798e-01 7.13725984e-01 6.75146759e-01 -4.71571773e-01
-5.52153766e-01 -6.07048690e-01 1.75663710e-01 7.32259870e-01
-3.40554386e-01 -4.85919625e-01 -5.65646410e-01 3.41020733e-01
-1.46020603e+00 -6.29379630e-01 -3.94442268e-02 2.05712914e+00
6.45830274e-01 -1.10380299e-01 -3.34994227e-01 -2.12051600e-01
2.58763164e-01 4.89770383e-01 -7.52930522e-01 1.07216705e-02
-3.82751882e-01 1.06403410e-01 4.35447723e-01 8.30499291e-01
-7.28124321e-01 9.06175494e-01 6.76317167e+00 7.16795027e-01
-1.16341627e+00 6.63003745e-03 8.10379088e-01 -1.44823506e-01
-1.12220204e+00 1.44732565e-01 -7.40507841e-01 1.49862662e-01
3.29491198e-01 3.14804703e-01 5.44603109e-01 5.29376686e-01
3.51774096e-01 -1.83000743e-01 -9.28786993e-01 1.06679857e+00
3.28004271e-01 -1.37418175e+00 3.19369704e-01 3.32555592e-01
1.24525499e+00 1.41475409e-01 2.68788666e-01 -1.49858713e-01
4.11880404e-01 -9.65874076e-01 8.30564916e-01 5.96017480e-01
9.09692407e-01 -4.97559130e-01 1.10376485e-01 1.67745456e-01
-9.84127223e-01 3.11974078e-01 -3.65034789e-01 1.58953130e-01
5.24750710e-01 7.60016680e-01 -5.80126107e-01 7.15808272e-01
6.29529119e-01 8.96431684e-01 -3.65178019e-01 6.51833475e-01
-4.78969753e-01 1.97039023e-01 -3.84719163e-01 4.07600373e-01
1.86103016e-01 -5.00343502e-01 5.60310245e-01 8.15834522e-01
5.09650528e-01 9.27574635e-02 -8.79933834e-02 1.38270986e+00
-5.85353486e-02 -8.58568326e-02 -7.86173165e-01 4.20375355e-02
2.14568332e-01 1.20341051e+00 -5.49737811e-01 -8.18968266e-02
-5.61851859e-01 1.31323993e+00 2.77705282e-01 6.97677970e-01
-8.79285634e-01 1.09065652e-01 7.17038035e-01 4.17507052e-01
3.07390839e-01 -4.56580728e-01 -4.02680695e-01 -1.24904060e+00
7.40590543e-02 -5.79021990e-01 -9.56453979e-02 -1.24088311e+00
-1.34354568e+00 4.56342399e-01 1.12650715e-01 -1.12616527e+00
-1.68236986e-01 -5.79315126e-01 -5.50243735e-01 1.04628468e+00
-1.90317631e+00 -1.42394280e+00 -4.96777534e-01 4.79184747e-01
3.76792520e-01 2.52921015e-01 6.52942359e-01 3.55522543e-01
-2.91671425e-01 3.29929888e-01 5.06641790e-02 -3.46374393e-01
6.21636987e-01 -1.20872390e+00 6.45806849e-01 8.26205850e-01
8.97739232e-02 4.81550694e-01 4.23624963e-01 -6.67623878e-01
-1.28068340e+00 -1.05416584e+00 4.15734977e-01 -6.34104848e-01
2.37603918e-01 -3.24245930e-01 -7.70768642e-01 5.24627447e-01
1.56742513e-01 -4.84595634e-02 5.56631505e-01 -2.02625573e-01
-3.79146636e-01 1.27118573e-01 -9.85110939e-01 7.50319839e-01
1.22771406e+00 -6.99309528e-01 3.33550796e-02 2.14946732e-01
6.13924801e-01 -8.01309049e-01 -8.40000153e-01 4.67904687e-01
4.60733056e-01 -1.26977789e+00 1.31884778e+00 -7.81869441e-02
8.10342252e-01 -3.18620026e-01 -3.80106896e-01 -1.49457979e+00
-1.67954564e-01 -5.52384257e-01 -2.32656240e-01 1.13695562e+00
1.98053598e-01 -4.04720485e-01 7.93763816e-01 6.71521604e-01
-3.60429287e-01 -7.82884717e-01 -7.35825181e-01 -5.27744710e-01
1.00581326e-01 -4.71057355e-01 4.39142436e-01 9.09654140e-01
-7.11057365e-01 3.59552920e-01 -4.80330080e-01 3.39249551e-01
8.33614647e-01 -3.15385349e-02 8.12817812e-01 -1.02461064e+00
-3.62331510e-01 -4.04190153e-01 -1.31131068e-01 -1.74428022e+00
2.95718703e-02 -8.69484723e-01 -3.51379290e-02 -1.86557055e+00
1.37556523e-01 -3.86088789e-01 2.39603817e-01 1.33794516e-01
-2.06159621e-01 4.85863239e-01 7.12664472e-03 2.23981410e-01
-9.94836092e-02 1.08410168e+00 1.79382086e+00 7.64811933e-02
-9.69464704e-02 -3.39688748e-01 -5.87943256e-01 8.14706683e-01
4.12941873e-01 -1.96498111e-01 -7.04864323e-01 -1.00176299e+00
4.46330190e-01 -6.30423129e-02 6.63929760e-01 -7.38034070e-01
-9.17850062e-02 -1.36733025e-01 6.69142663e-01 -7.30599523e-01
7.41521060e-01 -6.00839496e-01 8.81174654e-02 -1.22454017e-01
-9.20518935e-02 -9.92714986e-02 1.98564336e-01 7.37361312e-01
-1.53963163e-01 4.80559170e-02 8.47356200e-01 -1.19729877e-01
-3.89352947e-01 5.56877375e-01 7.38799898e-03 1.46668911e-01
5.76969564e-01 -3.55167836e-01 -2.45367482e-01 -6.60791695e-01
-3.70639920e-01 9.23940018e-02 9.11432326e-01 2.43617520e-01
8.12398493e-01 -1.44381762e+00 -4.68286365e-01 3.89811993e-01
4.50568050e-02 4.63089973e-01 5.76060116e-01 7.62380183e-01
-6.15735650e-01 1.57841034e-02 -7.92069454e-03 -7.20736504e-01
-9.37355280e-01 1.58850670e-01 4.87988144e-01 -9.10905451e-02
-8.61640871e-01 8.77721786e-01 7.74949729e-01 -5.05907953e-01
1.63212806e-01 -3.09202343e-01 1.58164471e-01 -3.95604163e-01
3.30356359e-01 1.74150571e-01 -6.87691867e-02 -5.99191129e-01
-7.96191841e-02 9.58003342e-01 8.70696157e-02 -3.79076958e-01
1.41047335e+00 -4.05064821e-01 1.63772568e-01 1.04204923e-01
1.19070566e+00 1.85144588e-01 -2.04842973e+00 -1.99715376e-01
-7.16670275e-01 -8.16616178e-01 4.88021672e-01 -7.19155490e-01
-1.12702119e+00 1.12321985e+00 5.01996934e-01 -1.59193009e-01
1.10392463e+00 -2.20055893e-01 8.79904389e-01 -3.64466920e-03
1.41066954e-01 -8.94613802e-01 2.76092172e-01 5.12899935e-01
7.85174906e-01 -1.21778250e+00 1.06081717e-01 -6.63931727e-01
-6.45047605e-01 9.66012597e-01 5.16515136e-01 9.45393927e-03
6.00677907e-01 2.67029315e-01 2.21523225e-01 -2.54166514e-01
-5.24155438e-01 -4.98669669e-02 6.25053525e-01 7.60215700e-01
4.49706286e-01 -3.71495277e-01 2.01713413e-01 -4.25159484e-02
-2.34767236e-02 -1.71206221e-01 2.96439111e-01 7.11879253e-01
5.59505597e-02 -1.06811631e+00 -1.83140159e-01 -1.78241730e-03
-2.93573081e-01 -3.23815912e-01 -3.22539598e-01 5.71771979e-01
-1.06426343e-01 6.41005158e-01 1.01953939e-01 -3.09893698e-01
3.79063755e-01 -2.13985547e-01 8.77006888e-01 -5.55491090e-01
-1.32528320e-01 4.33521241e-01 -1.24096841e-01 -6.85567260e-01
-6.20934308e-01 -3.52256149e-01 -1.01033449e+00 -2.24215701e-01
-3.01396489e-01 -3.42411816e-01 8.17035794e-01 6.61428392e-01
5.53615749e-01 5.74787617e-01 7.80152202e-01 -1.17276359e+00
-4.01903689e-01 -6.47541761e-01 -4.58704561e-01 4.38400894e-01
3.23965132e-01 -8.36115360e-01 -4.53293949e-01 7.25122727e-03] | [9.278450965881348, -3.142446756362915] |
bb4a4ce8-3f59-4931-b6ec-6f85ac9eb6f8 | hybrid-local-global-transformer-for-image | 2109.07100 | null | https://arxiv.org/abs/2109.07100v3 | https://arxiv.org/pdf/2109.07100v3.pdf | Complementary Feature Enhanced Network with Vision Transformer for Image Dehazing | Conventional CNNs-based dehazing models suffer from two essential issues: the dehazing framework (limited in interpretability) and the convolution layers (content-independent and ineffective to learn long-range dependency information). In this paper, firstly, we propose a new complementary feature enhanced framework, in which the complementary features are learned by several complementary subtasks and then together serve to boost the performance of the primary task. One of the prominent advantages of the new framework is that the purposively chosen complementary tasks can focus on learning weakly dependent complementary features, avoiding repetitive and ineffective learning of the networks. We design a new dehazing network based on such a framework. Specifically, we select the intrinsic image decomposition as the complementary tasks, where the reflectance and shading prediction subtasks are used to extract the color-wise and texture-wise complementary features. To effectively aggregate these complementary features, we propose a complementary features selection module (CFSM) to select the more useful features for image dehazing. Furthermore, we introduce a new version of vision transformer block, named Hybrid Local-Global Vision Transformer (HyLoG-ViT), and incorporate it within our dehazing networks. The HyLoG-ViT block consists of the local and the global vision transformer paths used to capture local and global dependencies. As a result, the HyLoG-ViT introduces locality in the networks and captures the global and long-range dependencies. Extensive experiments on homogeneous, non-homogeneous, and nighttime dehazing tasks reveal that the proposed dehazing network can achieve comparable or even better performance than CNNs-based dehazing models. | ['Long Xu', 'Hongyu Li', 'Jia Li', 'Dong Zhao'] | 2021-09-15 | null | null | null | null | ['image-dehazing', 'intrinsic-image-decomposition'] | ['computer-vision', 'computer-vision'] | [ 1.44198254e-01 -3.68138701e-01 2.72423208e-01 -3.02431613e-01
-3.22331071e-01 -1.16544656e-01 7.11979568e-01 -3.73053700e-01
-3.06306750e-01 3.47962022e-01 1.79730654e-01 1.42300099e-01
-3.86208653e-01 -1.09392262e+00 -6.79178834e-01 -1.22289932e+00
2.74198234e-01 -4.16534662e-01 6.04153275e-01 -5.82072794e-01
2.37878770e-01 4.86476868e-01 -2.00608802e+00 3.50020945e-01
1.02257502e+00 1.33335507e+00 4.52353448e-01 5.85562170e-01
1.74780264e-01 1.00693762e+00 -3.27895731e-01 6.92635775e-02
3.40471923e-01 -4.89013582e-01 -4.55735683e-01 2.08145529e-02
4.03207421e-01 -4.52685654e-01 -2.87758738e-01 9.59522367e-01
5.47896385e-01 2.08546624e-01 4.91204590e-01 -1.20479727e+00
-8.90211642e-01 1.39643764e-02 -7.22512782e-01 1.81526899e-01
-2.20014110e-01 2.48289898e-01 9.15249228e-01 -1.03315890e+00
1.52460579e-02 1.05127895e+00 6.60485625e-01 3.72224838e-01
-7.71183312e-01 -6.42552435e-01 4.17820930e-01 4.37865287e-01
-1.51296604e+00 -5.04860342e-01 1.05976725e+00 -1.75162494e-01
8.41365755e-01 2.82602191e-01 5.73815405e-01 8.19719195e-01
3.18592578e-01 6.30208790e-01 1.24092650e+00 -4.08265471e-01
5.05572855e-02 1.50297895e-01 2.63317943e-01 9.62870598e-01
1.58120859e-02 3.81314367e-01 -7.60303199e-01 2.01477796e-01
7.31705487e-01 3.48328501e-01 -4.70506489e-01 -1.56531725e-02
-9.69206095e-01 6.56950355e-01 6.51787579e-01 2.85972446e-01
-3.24417472e-01 8.22715610e-02 9.68591943e-02 2.77542412e-01
5.06755352e-01 5.35287820e-02 -3.94832432e-01 4.36584562e-01
-7.45609343e-01 3.72773744e-02 4.02885228e-01 7.11575747e-01
1.33156419e+00 1.44712403e-01 -3.18186790e-01 9.61157501e-01
2.99780905e-01 3.62831384e-01 4.89597589e-01 -5.58524370e-01
1.63859129e-01 5.78888297e-01 -7.65005276e-02 -1.29306817e+00
-4.21635598e-01 -5.67244947e-01 -1.10518706e+00 4.28299844e-01
-7.97802731e-02 -8.47010314e-03 -1.16856492e+00 1.58561325e+00
1.28440827e-01 2.92726010e-01 3.57943103e-02 1.02398324e+00
9.69845653e-01 8.25997472e-01 -7.62828365e-02 8.40019062e-02
1.20690060e+00 -1.41274405e+00 -5.68109930e-01 -2.50934720e-01
2.09390029e-01 -8.07396770e-01 1.14261651e+00 3.27038646e-01
-9.40854788e-01 -9.78964090e-01 -1.26064003e+00 -2.86133975e-01
-6.61972821e-01 2.85950154e-01 3.62900913e-01 2.82247007e-01
-1.28989065e+00 5.31640828e-01 -5.43928325e-01 -1.15327694e-01
3.04244906e-01 2.47590736e-01 -9.47799012e-02 -2.83567607e-01
-1.19866633e+00 8.03866684e-01 3.34138960e-01 5.30901253e-01
-1.06778276e+00 -5.07600904e-01 -8.27870071e-01 1.94338784e-01
3.64965737e-01 -6.18315876e-01 6.98144615e-01 -1.27722192e+00
-1.55563319e+00 5.71291208e-01 -1.80031769e-02 -8.99566710e-02
1.55114591e-01 -1.71533316e-01 -4.37151223e-01 2.28760645e-01
-1.45805746e-01 4.82519835e-01 1.46194327e+00 -1.39991713e+00
-8.62916589e-01 -3.04575413e-01 2.34181508e-01 3.84570301e-01
-6.85712397e-01 -1.14383735e-01 -5.45050323e-01 -8.80876601e-01
-1.02126472e-01 -4.55004573e-01 1.07270606e-01 1.96026742e-01
-1.75808296e-01 7.85020739e-03 1.13796246e+00 -7.09020495e-01
1.17618620e+00 -2.42172885e+00 9.38655362e-02 7.28586242e-02
5.66161335e-01 4.81660008e-01 -4.60890144e-01 2.81147569e-01
-1.40921175e-01 -8.63390192e-02 -3.11350167e-01 -3.25107038e-01
-2.30324537e-01 3.26657146e-01 -1.58736169e-01 3.77998501e-01
5.44561684e-01 7.90955901e-01 -7.20400631e-01 -2.93997705e-01
5.74894309e-01 7.76400924e-01 -3.15840095e-01 4.52886701e-01
-4.58212607e-02 2.00541854e-01 -3.80006671e-01 7.09483147e-01
1.02476561e+00 5.36296703e-02 -5.24780393e-01 -4.68631327e-01
-4.24590886e-01 -4.88136187e-02 -8.43471169e-01 1.35445940e+00
-6.71536684e-01 3.37380409e-01 1.24041870e-01 -1.05212986e+00
1.01982379e+00 5.10931388e-02 2.13495642e-01 -8.42222273e-01
1.44614175e-01 2.30105668e-02 -3.12819034e-01 -6.46754503e-01
2.52667695e-01 -1.59879610e-01 4.93695289e-01 1.49966225e-01
2.04778463e-01 1.61345139e-01 -3.23636889e-01 -1.52964368e-01
8.84887040e-01 3.27877730e-01 1.39832884e-01 -3.83007795e-01
1.02723861e+00 -1.99552372e-01 6.54445648e-01 5.45126379e-01
-1.75105572e-01 8.30564559e-01 1.57292217e-01 -7.31445491e-01
-6.63438141e-01 -9.06605840e-01 1.52502045e-01 1.16961443e+00
4.82696801e-01 -3.26924831e-01 -6.53065503e-01 -7.39752412e-01
-3.08597445e-01 4.78074968e-01 -8.66131186e-01 -5.70319653e-01
-4.28972334e-01 -9.15334046e-01 2.42454231e-01 4.51264530e-01
1.30169022e+00 -9.48997498e-01 -4.29278910e-01 -8.48329738e-02
-1.25598624e-01 -8.75645041e-01 -5.01294911e-01 3.10345322e-01
-4.54125762e-01 -9.82154012e-01 -6.92152143e-01 -9.37938213e-01
5.58915257e-01 1.08258605e+00 7.47507155e-01 3.35650086e-01
-1.36947334e-01 2.15162069e-01 -4.66662616e-01 -4.59323585e-01
1.47262767e-01 8.55445191e-02 -3.44476461e-01 5.51042020e-01
3.38172168e-01 -6.79569960e-01 -8.47264647e-01 3.58108521e-01
-1.17382205e+00 2.47733191e-01 7.51271963e-01 1.17969334e+00
5.17296553e-01 4.42619473e-01 3.71308416e-01 -6.71121955e-01
4.05479074e-01 -3.02554131e-01 -4.29986835e-01 3.92040074e-01
-6.67698741e-01 -7.10363761e-02 9.53065634e-01 -2.84134775e-01
-1.42045820e+00 1.00936005e-02 -1.83026984e-01 -6.46249533e-01
-1.32589430e-01 3.85269046e-01 -4.24850374e-01 -6.45442903e-01
5.39581656e-01 5.87096035e-01 -2.34142989e-02 -4.72979099e-01
1.99937537e-01 5.52121222e-01 4.45546895e-01 -3.83207411e-01
1.08849192e+00 7.53730178e-01 -3.60546932e-02 -1.05323958e+00
-9.03348982e-01 -4.79841709e-01 -6.35242522e-01 -2.69632906e-01
9.85081375e-01 -1.03323805e+00 -5.53123772e-01 1.04723811e+00
-1.09197807e+00 -2.90387928e-01 -2.18507573e-01 2.37552866e-01
-2.38966510e-01 3.44579697e-01 -4.60086823e-01 -6.81401670e-01
-3.27074468e-01 -1.08530343e+00 1.17097449e+00 5.79000831e-01
5.41442752e-01 -1.02755737e+00 -2.12358743e-01 1.19080447e-01
7.00128555e-01 1.45385996e-01 9.08406913e-01 -1.19219840e-01
-6.91268265e-01 2.19766125e-01 -6.48096859e-01 8.44705582e-01
3.55976939e-01 7.08280802e-02 -1.42862058e+00 -1.96100667e-01
2.93168545e-01 -2.01424032e-01 1.33710861e+00 3.56879920e-01
1.25501549e+00 -7.38434494e-02 1.74809061e-02 1.08937991e+00
1.67257237e+00 1.06385179e-01 8.39305937e-01 5.26955843e-01
1.02100694e+00 8.33637357e-01 4.88385499e-01 3.29758167e-01
3.97357315e-01 3.16928625e-01 6.70349360e-01 -7.30883300e-01
-2.64728129e-01 -4.25027944e-02 3.98949146e-01 1.01078093e+00
-4.16353732e-01 -1.67803213e-01 -5.63621223e-01 5.38131773e-01
-1.95824230e+00 -6.39509976e-01 7.67402649e-02 2.05085897e+00
5.35024047e-01 -1.19594820e-01 -1.00694619e-01 8.56105611e-03
6.87281013e-01 4.92600173e-01 -4.10285801e-01 -2.48929918e-01
-3.52728605e-01 2.88698643e-01 5.03585637e-01 4.20900762e-01
-1.27086341e+00 9.11502242e-01 5.33212423e+00 9.52651918e-01
-1.16866624e+00 1.48936003e-01 6.24038160e-01 1.73832223e-01
-2.36774519e-01 -9.57672596e-02 -6.00946724e-01 2.67747492e-01
4.80149627e-01 2.81885326e-01 6.03255928e-01 5.89947522e-01
1.83601528e-01 -8.33988041e-02 -6.84118748e-01 1.07686234e+00
3.35411131e-01 -1.10493839e+00 1.53529465e-01 -1.55522496e-01
7.29318380e-01 -8.49873945e-02 1.84799165e-01 2.07258791e-01
6.97263703e-02 -9.40202057e-01 7.79372931e-01 7.38512337e-01
5.63498974e-01 -8.69280398e-01 8.29828978e-01 1.41417518e-01
-1.47514939e+00 -2.15830848e-01 -5.88383675e-01 -1.61887985e-02
-3.75462681e-01 5.64219952e-01 3.70367914e-02 9.23528135e-01
1.13480842e+00 1.04819655e+00 -6.64668858e-01 7.93915808e-01
-3.44734460e-01 3.94745857e-01 -9.44021195e-02 2.44371802e-01
4.93768334e-01 -3.15739512e-01 1.92246616e-01 1.08018696e+00
1.85175061e-01 1.20308757e-01 -2.50870595e-03 8.37217093e-01
1.56666532e-01 -1.57055452e-01 -4.98287588e-01 2.94366658e-01
1.27586603e-01 1.41466784e+00 -4.23407555e-01 -2.02431068e-01
-8.39782000e-01 1.20089602e+00 3.20049942e-01 6.01575375e-01
-8.56096447e-01 -7.91065693e-01 7.09376216e-01 -1.52354121e-01
4.41642493e-01 -5.16231433e-02 -1.10005066e-01 -1.20230126e+00
1.11124925e-01 -7.92336702e-01 2.93751180e-01 -9.11218226e-01
-1.51733255e+00 7.20850170e-01 -1.72525197e-01 -1.18298304e+00
3.73650461e-01 -7.89702177e-01 -8.78699601e-01 1.11584663e+00
-2.36802459e+00 -1.70034850e+00 -1.03151488e+00 1.16264021e+00
6.04248881e-01 -5.11538722e-02 6.06515288e-01 3.03189307e-01
-8.02357376e-01 3.77636522e-01 1.17112763e-01 2.73128562e-02
8.68548632e-01 -1.05238247e+00 1.81383621e-02 1.09735990e+00
-3.34667623e-01 4.91336256e-01 1.87834889e-01 -2.75527179e-01
-1.13780034e+00 -1.43427706e+00 5.61933219e-01 5.47236763e-02
5.45267522e-01 -2.68321306e-01 -9.11386073e-01 4.11598146e-01
2.37883642e-01 2.41780266e-01 3.79112840e-01 -3.11934680e-01
-5.36270916e-01 -6.81307316e-01 -8.09649706e-01 5.44848502e-01
8.75756562e-01 -6.91997111e-01 -3.72926056e-01 2.76564777e-01
9.01750922e-01 -5.18385991e-02 -4.98368144e-01 3.12389493e-01
5.68844140e-01 -1.39432728e+00 1.01400399e+00 -4.68001403e-02
5.55724442e-01 -5.73636830e-01 -2.42336556e-01 -1.15454626e+00
-7.16428995e-01 -5.57024360e-01 7.89999142e-02 1.36304402e+00
5.29970080e-02 -8.75025034e-01 3.08512598e-01 1.72289521e-01
-3.33184421e-01 -7.30035365e-01 -5.46539128e-01 -6.82120919e-01
-1.68094695e-01 -1.02983609e-01 7.81220853e-01 9.98542786e-01
-6.89669311e-01 2.82199591e-01 -6.59880877e-01 4.58890498e-01
5.81071496e-01 2.77835019e-02 4.45817351e-01 -1.08255899e+00
-3.64255488e-01 -2.98518091e-01 -2.30753526e-01 -9.50740099e-01
-2.43178353e-01 -4.83879209e-01 2.39212826e-01 -1.37288034e+00
9.65212211e-02 -3.18592072e-01 -8.29058349e-01 7.16708541e-01
-3.52931291e-01 2.18177497e-01 9.00679752e-02 4.20634717e-01
-2.70839274e-01 8.60926569e-01 1.30442667e+00 -2.41953060e-01
-2.58138299e-01 -9.90170911e-02 -9.00637567e-01 7.54103363e-01
8.10482502e-01 -1.64640009e-01 -6.28687084e-01 -7.36950457e-01
-4.39489372e-02 -4.47548479e-01 6.95956945e-01 -1.05416214e+00
2.68268704e-01 -3.12633485e-01 4.47018147e-01 -3.79897177e-01
3.25774163e-01 -8.31465721e-01 -2.03084469e-01 1.45872816e-01
-8.05224851e-02 -1.57791764e-01 1.17314290e-02 5.52177846e-01
-5.20586729e-01 -4.09738682e-02 9.04380143e-01 -1.76067635e-01
-1.03469121e+00 5.21454155e-01 -2.32206538e-01 -4.68359858e-01
9.83655572e-01 -3.57543945e-01 -6.35703802e-01 -1.98720962e-01
-3.63757342e-01 7.14549497e-02 2.74785012e-01 2.71882653e-01
8.91208649e-01 -1.19715428e+00 -5.41197181e-01 5.57400525e-01
1.97920367e-01 2.51890779e-01 4.43752289e-01 8.55574489e-01
-5.94135940e-01 1.01730034e-01 -4.22958791e-01 -4.15463388e-01
-1.13174129e+00 6.53032839e-01 5.28837085e-01 -8.85031223e-02
-6.64080203e-01 9.58395004e-01 8.73487592e-01 -1.17035165e-01
1.24281630e-01 -2.56570011e-01 -4.20264721e-01 -1.16928726e-01
7.18442202e-01 3.97033125e-01 2.34674141e-01 -5.57294548e-01
-2.23471373e-01 8.36051404e-01 -1.56916872e-01 2.45840475e-01
1.54852271e+00 -3.78139377e-01 -4.72468376e-01 1.93543598e-01
1.23934627e+00 -1.34546533e-01 -1.50859499e+00 -4.57435310e-01
-4.75521833e-01 -5.44197083e-01 5.44380844e-01 -6.17808163e-01
-1.43243313e+00 1.25345492e+00 6.92909539e-01 1.99908644e-01
2.03100538e+00 -4.15220976e-01 8.05274844e-01 1.84370726e-01
6.26864806e-02 -9.69415009e-01 1.99661553e-01 4.87597048e-01
8.41965854e-01 -1.05646718e+00 -1.00733310e-01 -5.43319941e-01
-4.77463841e-01 1.27416265e+00 8.96030605e-01 -2.83573985e-01
7.78440714e-01 1.37428313e-01 1.46375746e-01 -2.89374918e-01
-6.25664830e-01 -4.86514568e-01 4.24714625e-01 8.41624558e-01
5.59287779e-02 -3.14462721e-01 9.58228335e-02 5.83741724e-01
2.59174198e-01 -2.29119763e-01 1.14985153e-01 8.63981783e-01
-5.70553422e-01 -8.58098090e-01 -3.81260663e-01 2.00841695e-01
-1.84013806e-02 -2.63060987e-01 -3.90238047e-01 7.31105208e-01
6.31264806e-01 9.47936893e-01 -1.21084243e-01 -9.33461130e-01
2.30801806e-01 -2.84799784e-01 2.09072173e-01 -3.28104675e-01
-7.61067092e-01 2.13802949e-01 -2.28738829e-01 -6.93000972e-01
-6.96477354e-01 -1.41208008e-01 -7.53503799e-01 -3.35629225e-01
-4.95683014e-01 -2.52315640e-01 3.59808266e-01 1.01155794e+00
3.27379793e-01 9.11548257e-01 1.00897574e+00 -9.15774286e-01
-1.89972222e-01 -7.61051834e-01 -7.79999256e-01 1.87354624e-01
6.85713768e-01 -7.04217315e-01 -4.86132503e-01 8.49420726e-02] | [10.958629608154297, -2.937124729156494] |
45a454da-ab21-40ce-8691-a985abbb8c80 | semi-supervised-vector-quantization-in-visual | 2207.06738 | null | https://arxiv.org/abs/2207.06738v1 | https://arxiv.org/pdf/2207.06738v1.pdf | Semi-supervised Vector-Quantization in Visual SLAM using HGCN | In this paper, two semi-supervised appearance based loop closure detection technique, HGCN-FABMAP and HGCN-BoW are introduced. Furthermore an extension to the current state of the art localization SLAM algorithm, ORB-SLAM, is presented. The proposed HGCN-FABMAP method is implemented in an off-line manner incorporating Bayesian probabilistic schema for loop detection decision making. Specifically, we let a Hyperbolic Graph Convolutional Neural Network (HGCN) to operate over the SURF features graph space, and perform vector quantization part of the SLAM procedure. This part previously was performed in an unsupervised manner using algorithms like HKmeans, kmeans++,..etc. The main Advantage of using HGCN, is that it scales linearly in number of graph edges. Experimental results shows that HGCN-FABMAP algorithm needs far more cluster centroids than HGCN-ORB, otherwise it fails to detect loop closures. Therefore we consider HGCN-ORB to be more efficient in terms of memory consumption, also we conclude the superiority of HGCN-BoW and HGCN-FABMAP with respect to other algorithms. | ['Ali Mohades Khorasani', 'Saeed Shiry Ghidary', 'Amir Zarringhalam'] | 2022-07-14 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [-1.24706902e-01 2.38922283e-01 6.62273169e-02 -1.84115693e-01
-4.96661752e-01 -2.84892827e-01 7.46555924e-01 4.40161765e-01
-4.39318955e-01 6.07873857e-01 1.20156091e-02 -3.55355114e-01
-3.36860687e-01 -7.79176533e-01 -8.12282145e-01 -4.73335624e-01
-3.88589472e-01 8.13042879e-01 4.52692509e-01 -1.57560706e-01
5.90634346e-01 7.47208774e-01 -1.53721237e+00 -6.06325269e-01
5.61001003e-01 9.32774961e-01 1.57849472e-02 7.96276569e-01
-4.65131504e-03 5.41430116e-01 5.35820276e-02 6.53253049e-02
2.05610529e-01 -2.06656516e-01 -7.48842239e-01 -2.20689178e-01
7.02321649e-01 1.36898920e-01 -4.14498657e-01 1.12574625e+00
3.27531785e-01 4.18635875e-01 9.91082668e-01 -1.46815825e+00
-2.81359792e-01 1.74768254e-01 -6.70687079e-01 -1.84334099e-01
5.31433940e-01 -3.59171033e-01 8.39457512e-01 -7.90655613e-01
9.46704209e-01 1.34690630e+00 1.17266238e+00 -1.68509573e-01
-1.19296491e+00 -4.84465539e-01 -1.45300344e-01 1.33803532e-01
-1.86432445e+00 -7.46650100e-02 7.16600776e-01 -3.92559022e-01
1.08118343e+00 1.06766764e-02 7.04250872e-01 4.44945306e-01
6.60462797e-01 4.71479982e-01 9.38428700e-01 -7.53230333e-01
6.08409047e-01 -2.38123775e-01 2.96440542e-01 1.44725358e+00
7.53867507e-01 9.43693891e-02 -7.20493197e-01 -4.30568784e-01
6.87941790e-01 -2.32334822e-01 -2.94315219e-02 -1.14907873e+00
-1.14786398e+00 1.07523429e+00 9.04089570e-01 1.80341005e-01
-2.00789765e-01 8.23042750e-01 4.43016022e-01 8.81912187e-02
2.30245143e-01 1.30285621e-01 4.70570065e-02 1.14096083e-01
-1.06864893e+00 2.59182751e-01 7.32117355e-01 1.43289936e+00
1.53819788e+00 -4.08549666e-01 3.21416974e-01 1.66568801e-01
8.17195654e-01 5.41272402e-01 3.40887487e-01 -8.88313890e-01
-1.03862889e-01 6.50125504e-01 3.00672092e-02 -1.53444695e+00
-8.74861002e-01 -7.76294172e-02 -5.95580697e-01 3.16668332e-01
-6.30436018e-02 1.50425956e-01 -1.00172532e+00 1.24745214e+00
4.79660034e-01 1.17976766e-03 -3.82351992e-03 8.47390473e-01
6.54947817e-01 4.41062301e-01 -1.19033493e-01 2.95823544e-01
1.07904291e+00 -1.04045594e+00 -8.13061714e-01 -1.47628516e-01
1.09316933e+00 -8.76654387e-01 5.50304115e-01 3.93876344e-01
-2.27346346e-01 -5.14620066e-01 -1.32388222e+00 -1.27876163e-01
-5.81255853e-01 4.47775543e-01 1.05558395e+00 4.15329933e-01
-1.41322124e+00 5.87759912e-01 -1.02971041e+00 -1.20075703e+00
1.43949658e-01 4.66256440e-01 -8.43790352e-01 -1.33584114e-02
-6.77077711e-01 8.29978228e-01 9.34603930e-01 1.29115134e-01
-6.04228795e-01 2.83072710e-01 -1.39307761e+00 -4.14382756e-01
2.77881715e-02 -4.99408156e-01 8.06572974e-01 -6.75543725e-01
-1.22758842e+00 7.20868230e-01 -1.92401335e-01 -6.47570252e-01
4.81564909e-01 -2.31310487e-01 -1.14034219e-02 2.14733630e-01
2.01317787e-01 9.35257852e-01 7.42489517e-01 -1.32462609e+00
-7.10373223e-01 -3.78942877e-01 -4.03431952e-01 3.40910912e-01
5.07764041e-01 -3.36034149e-01 -4.18236226e-01 -5.63477278e-02
1.01202703e+00 -1.28864920e+00 -1.90707237e-01 2.93128341e-01
-4.39551711e-01 -5.15748858e-01 1.19328856e+00 -3.42524081e-01
8.88912439e-01 -1.95064354e+00 -1.47402883e-01 6.38170421e-01
1.54588819e-01 -1.92436501e-01 2.52767563e-01 8.25511932e-01
2.81305432e-01 -2.13643044e-01 -2.05738619e-01 -4.52137232e-01
1.98748857e-01 5.02967179e-01 1.67832553e-01 1.46716094e+00
-1.68293089e-01 6.16552234e-01 -1.02511394e+00 -9.83216703e-01
6.58281446e-01 2.97751397e-01 -3.79026294e-01 -7.30120838e-02
-6.36980981e-02 -9.26571786e-02 -2.56586730e-01 7.23461807e-01
9.32644069e-01 -6.82456717e-02 2.41529010e-02 -1.74562052e-01
-4.60501254e-01 -1.93883613e-01 -1.41601801e+00 2.47229123e+00
-1.73316270e-01 1.02489316e+00 -6.73883408e-02 -8.33966911e-01
1.31556189e+00 -1.46935984e-01 2.46163368e-01 -2.01425046e-01
3.55929464e-01 2.88681269e-01 -4.39181983e-01 -6.70297295e-02
1.00027025e+00 2.55428761e-01 -9.78803933e-02 1.26286626e-01
5.31464636e-01 8.28857627e-03 7.27297971e-03 5.09902894e-01
1.04396641e+00 7.70479739e-01 6.89679503e-01 -7.55773306e-01
5.69757938e-01 5.66081703e-01 2.71570504e-01 9.11506712e-01
-5.36254585e-01 5.33903301e-01 2.65167832e-01 -6.01677358e-01
-7.84990549e-01 -1.05574214e+00 2.97343675e-02 6.84916019e-01
6.96352124e-01 -7.05040216e-01 -6.21734738e-01 -7.19628215e-01
1.67464927e-01 3.49141717e-01 -6.45697832e-01 -5.63332252e-02
-3.41043621e-01 -3.65393281e-01 3.90691549e-01 -3.26025160e-03
5.93808353e-01 -9.52277958e-01 -9.61648524e-01 2.52537578e-01
5.57417348e-02 -6.60637081e-01 6.73406050e-02 3.67448181e-01
-6.47285640e-01 -1.23256278e+00 -1.57680869e-01 -9.34660256e-01
7.84577370e-01 3.73944789e-01 7.02826977e-01 -2.47160289e-02
-4.77898896e-01 4.76676911e-01 -6.67134285e-01 -3.82043183e-01
-8.58521983e-02 2.15410843e-01 2.23661825e-01 -5.89857996e-02
5.19677460e-01 -4.91317689e-01 -5.00255525e-01 -4.49705161e-02
-3.80585194e-01 -1.88089088e-01 7.14527965e-01 6.08247578e-01
7.76226282e-01 7.39818513e-02 -2.91590214e-01 -7.92819500e-01
1.88112110e-01 -2.29099929e-01 -9.35896754e-01 1.81119144e-02
-9.24933553e-01 3.26677024e-01 1.37352630e-01 1.18032515e-01
-5.00345230e-01 7.60558069e-01 3.99545021e-03 -4.85285431e-01
-1.13813832e-01 4.68796998e-01 2.94256538e-01 -9.27974224e-01
7.63935804e-01 1.39844432e-01 6.30969629e-02 -2.52726108e-01
5.78225017e-01 7.04836547e-01 5.34750700e-01 -1.46807000e-01
9.70880330e-01 6.84373856e-01 4.65977043e-01 -9.51969624e-01
-1.57993406e-01 -1.11105967e+00 -9.07351375e-01 -4.10375178e-01
9.30369496e-01 -9.60473955e-01 -8.03783596e-01 2.78722525e-01
-1.15345585e+00 -1.41280964e-01 2.59768575e-01 6.59813762e-01
-8.35523844e-01 6.33645952e-01 -2.13399068e-01 -1.07933950e+00
-3.53664756e-01 -8.39997709e-01 1.38773298e+00 1.50519192e-01
-5.95294423e-02 -9.23185706e-01 4.48736310e-01 -1.78100020e-01
8.14064741e-02 6.72299266e-01 3.74271989e-01 -6.31177187e-01
-6.97260022e-01 -4.50615972e-01 -3.75903010e-01 -1.49707824e-01
-1.82362497e-01 4.33993749e-02 -6.23924375e-01 -5.38092196e-01
-5.41419744e-01 -2.61149079e-01 7.34532297e-01 3.68361354e-01
3.84697586e-01 3.91769707e-02 -6.75542474e-01 8.22460413e-01
2.09504366e+00 -6.58617839e-02 5.55915117e-01 7.46297300e-01
8.84733617e-01 2.67671525e-01 1.06340325e+00 3.69241029e-01
4.17251408e-01 4.91878986e-01 7.80841351e-01 -1.10278741e-01
-1.04143552e-01 -5.90954185e-01 1.51362240e-01 8.05707693e-01
1.85077310e-01 -1.72512591e-01 -1.10042262e+00 4.83766258e-01
-2.29631138e+00 -4.00137216e-01 -4.64310437e-01 2.15655112e+00
-1.62184343e-01 2.81414110e-02 -2.41000086e-01 1.18444689e-01
7.38736570e-01 1.41791120e-01 4.85313684e-02 -4.11992043e-01
-4.60419133e-02 -4.25692052e-02 1.30316293e+00 7.98175991e-01
-1.56808817e+00 1.34355891e+00 5.39685535e+00 6.07635617e-01
-8.87818933e-01 1.85904890e-01 -3.78582805e-01 8.10462892e-01
1.99562311e-01 6.38398647e-01 -6.48776829e-01 -3.83980572e-02
6.34067476e-01 2.38202140e-01 3.03612441e-01 1.11574602e+00
-1.75055563e-01 -7.75566757e-01 -6.85461223e-01 1.38226879e+00
4.55857754e-01 -1.15323269e+00 -5.13584400e-03 -1.13488339e-01
5.05379438e-01 3.70867193e-01 -6.64610028e-01 2.33046949e-01
4.17489827e-01 -6.42309427e-01 1.00838542e+00 3.55380386e-01
5.16765952e-01 -9.44334507e-01 9.43459749e-01 1.60646319e-01
-1.61612022e+00 3.24812800e-01 -7.25177824e-01 -1.01094417e-01
6.87773572e-03 2.83201724e-01 -1.05389082e+00 1.03915095e+00
6.99207902e-01 6.43540144e-01 -9.10591960e-01 1.30401158e+00
-3.94572735e-01 1.93086132e-01 -6.02531672e-01 -3.23034406e-01
7.03089714e-01 -2.49028906e-01 4.11504120e-01 1.37321067e+00
4.26788181e-01 -4.55756456e-01 5.35971105e-01 6.18059576e-01
4.69573408e-01 3.23661178e-01 -8.94577205e-01 2.56395698e-01
4.68071789e-01 1.32554328e+00 -1.17359877e+00 -2.87433505e-01
-2.45333500e-02 1.13432682e+00 4.60410982e-01 1.20081171e-01
-8.05339992e-01 -9.18916941e-01 1.04907535e-01 -4.37842123e-02
1.54988691e-01 -5.88695645e-01 2.64134288e-01 -7.58966267e-01
-4.79360461e-01 -1.76576629e-01 3.79074275e-01 -1.18988979e+00
-6.40593529e-01 4.86839265e-01 -6.13972768e-02 -1.16714358e+00
-4.72295135e-02 -7.91740239e-01 -1.35036886e-01 6.35248184e-01
-1.38409734e+00 -1.41127610e+00 -6.57578766e-01 6.40686333e-01
1.66760758e-01 1.70534641e-01 7.87124038e-01 -9.00475085e-02
-4.12042588e-02 2.21488342e-01 2.20604584e-01 1.85135335e-01
7.71904111e-01 -1.28536594e+00 3.17563564e-01 1.03380024e+00
4.85802799e-01 4.36629385e-01 8.93406272e-01 -9.51346636e-01
-1.74248171e+00 -1.31570220e+00 9.31126654e-01 -2.03918993e-01
5.51532507e-01 -6.08066082e-01 -4.21383858e-01 9.48490083e-01
5.98489530e-02 2.48633102e-01 7.96016753e-02 5.33582307e-02
-2.32907519e-01 5.54348268e-02 -1.10527182e+00 3.22078764e-01
9.32761192e-01 -7.27238834e-01 -6.28710747e-01 5.60227036e-01
5.77511072e-01 -5.35398662e-01 -6.39124155e-01 2.95843810e-01
4.42128330e-01 -1.16159272e+00 4.80730504e-01 2.20327541e-01
-5.20221770e-01 -9.40973639e-01 -4.00241792e-01 -9.42257941e-01
-4.79489774e-01 -7.06263721e-01 3.03842705e-02 8.42604697e-01
-5.36908070e-03 -7.85525203e-01 9.50378001e-01 -4.54153299e-01
-3.00320685e-01 -1.05581902e-01 -1.19919431e+00 -8.12950611e-01
-4.62443620e-01 -3.59555006e-01 1.52448997e-01 7.77284920e-01
-1.62417606e-01 2.80517507e-02 -1.68252826e-01 6.66298270e-01
1.01788974e+00 7.97347277e-02 1.08737576e+00 -1.40377820e+00
2.35370651e-01 2.10203771e-02 -1.35947120e+00 -6.78916335e-01
-4.17424254e-02 -9.20472026e-01 5.02154708e-01 -1.86987650e+00
-1.20690636e-01 -5.32040834e-01 -2.19220176e-01 4.30679590e-01
3.65951210e-01 3.57844949e-01 -8.95441547e-02 3.37965012e-01
-9.93934393e-01 4.87451941e-01 5.26261270e-01 -2.68341135e-02
-1.87126771e-01 -4.87553447e-01 2.79870033e-01 6.73895657e-01
5.90623677e-01 -6.81793213e-01 -2.48409092e-01 4.74727787e-02
1.03947677e-01 -2.28426322e-01 3.76835614e-01 -1.60934627e+00
6.65758014e-01 3.44768912e-01 3.05234522e-01 -1.14195728e+00
1.79681242e-01 -8.02761376e-01 4.86155242e-01 8.01731169e-01
2.37721622e-01 3.52147162e-01 -2.74986774e-02 9.67555642e-01
-2.10764110e-01 -2.83694893e-01 6.37524962e-01 -1.52883872e-01
-1.36284411e+00 1.62366182e-01 -5.08463442e-01 -5.33672571e-01
9.01907206e-01 -3.41859639e-01 4.21316437e-02 -1.81370929e-01
-5.12535810e-01 -1.42832762e-02 9.19343114e-01 2.67277390e-01
5.66468000e-01 -1.40110207e+00 -1.37692198e-01 2.92088211e-01
6.56713545e-01 1.37320206e-01 2.37212423e-03 1.02680278e+00
-1.69243360e+00 5.29095531e-01 -1.65743276e-01 -1.07181191e+00
-1.13340163e+00 7.35889316e-01 2.02396140e-01 1.79414496e-01
-7.48338223e-01 7.78259158e-01 -2.91837215e-01 -7.34471917e-01
4.16994244e-01 -3.40903699e-01 5.95171973e-02 -2.13190258e-01
-3.38173322e-02 3.71598959e-01 1.17887534e-01 -1.11259890e+00
-6.52746260e-01 9.94347811e-01 4.10151482e-01 -1.64834112e-01
7.42828548e-01 -2.39293352e-01 -4.79640543e-01 4.31447357e-01
1.19978380e+00 -2.05922142e-01 -8.32297325e-01 1.46664694e-01
3.13800156e-01 -1.90877661e-01 3.73451322e-01 -2.87874609e-01
-4.64405894e-01 5.89136779e-01 1.06992781e+00 -3.44668150e-01
8.37994397e-01 -4.40008603e-02 7.93433562e-02 6.58071935e-01
1.07998407e+00 -1.11211491e+00 -4.33106720e-01 5.23366511e-01
6.06537580e-01 -1.24427915e+00 4.40920442e-01 -3.15217137e-01
-8.60866308e-02 1.28956318e+00 3.98736209e-01 -7.03513384e-01
6.36803448e-01 -2.22765714e-01 9.63774994e-02 -5.47805429e-01
-1.01183012e-01 -4.42836374e-01 1.16407856e-01 5.33943355e-01
9.71831381e-02 8.59774798e-02 -5.35965800e-01 -3.97276759e-01
-1.95096225e-01 -9.02461559e-02 3.91806036e-01 1.37183607e+00
-9.46004212e-01 -7.41088569e-01 -6.24130726e-01 1.26288578e-01
3.38457853e-01 1.26695484e-01 -4.06408966e-01 1.38650155e+00
4.66001064e-01 7.06946075e-01 -7.54608260e-03 -5.90055823e-01
-2.01951023e-02 -9.46361497e-02 4.90269542e-01 -2.90297210e-01
-1.39158279e-01 1.58241883e-01 -2.42305156e-02 -8.91523898e-01
-5.53753078e-01 -6.15316272e-01 -1.53843117e+00 -1.51154861e-01
-5.85150719e-01 3.73813570e-01 1.20844853e+00 6.04597330e-01
3.04609209e-01 7.35536218e-02 1.55881524e-01 -1.17851615e+00
-1.09351821e-01 -1.07083893e+00 -8.60400319e-01 8.43838081e-02
3.11881781e-01 -1.13241184e+00 -4.69612032e-01 -4.05469656e-01] | [7.368973731994629, -2.028442621231079] |
fe8588a3-10f8-4121-89fe-3e9ee48c998b | logical-reasoning-for-natural-language | 2305.13214 | null | https://arxiv.org/abs/2305.13214v1 | https://arxiv.org/pdf/2305.13214v1.pdf | Logical Reasoning for Natural Language Inference Using Generated Facts as Atoms | State-of-the-art neural models can now reach human performance levels across various natural language understanding tasks. However, despite this impressive performance, models are known to learn from annotation artefacts at the expense of the underlying task. While interpretability methods can identify influential features for each prediction, there are no guarantees that these features are responsible for the model decisions. Instead, we introduce a model-agnostic logical framework to determine the specific information in an input responsible for each model decision. This method creates interpretable Natural Language Inference (NLI) models that maintain their predictive power. We achieve this by generating facts that decompose complex NLI observations into individual logical atoms. Our model makes predictions for each atom and uses logical rules to decide the class of the observation based on the predictions for each atom. We apply our method to the highly challenging ANLI dataset, where our framework improves the performance of both a DeBERTa-base and BERT baseline. Our method performs best on the most challenging examples, achieving a new state-of-the-art for the ANLI round 3 test set. We outperform every baseline in a reduced-data setting, and despite using no annotations for the generated facts, our model predictions for individual facts align with human expectations. | ['Marek Rei', 'Oana-Maria Camburu', 'Haim Dubossarsky', 'Pasquale Minervini', 'Joe Stacey'] | 2023-05-22 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 3.88933599e-01 8.98703516e-01 -5.32872796e-01 -6.77365601e-01
-7.73284137e-01 -4.81778562e-01 9.85525191e-01 1.52242765e-01
-1.03808083e-01 7.68480062e-01 5.94393671e-01 -4.40839410e-01
-2.41120070e-01 -7.03087330e-01 -1.05376422e+00 -2.45866328e-01
2.09020212e-01 1.17201793e+00 1.47710089e-03 4.91133444e-02
4.41869982e-02 -2.98527535e-02 -1.55178142e+00 9.13535774e-01
7.87386239e-01 1.20348465e+00 -2.15007201e-01 5.16220152e-01
-1.84180021e-01 1.42500842e+00 -4.35885429e-01 -6.37771487e-01
3.77472118e-02 -2.78206676e-01 -1.09079528e+00 -2.31843561e-01
7.16734171e-01 -1.93112552e-01 5.73318079e-02 6.30538702e-01
-1.63896114e-01 -1.90084755e-01 1.00684154e+00 -1.30323029e+00
-8.45013380e-01 1.16679323e+00 -1.86536074e-01 6.72038421e-02
4.25776631e-01 1.10211857e-01 1.85479999e+00 -9.38714981e-01
7.02714980e-01 1.45253873e+00 7.23792672e-01 5.18289626e-01
-1.35745931e+00 -4.57472593e-01 5.43350697e-01 4.54446554e-01
-1.00748742e+00 -6.09801948e-01 3.97358656e-01 -4.41449344e-01
1.29230535e+00 1.24760047e-01 5.26655495e-01 1.31575942e+00
1.62780017e-01 1.13525963e+00 8.62611115e-01 -5.56485236e-01
2.08038911e-01 -4.17940021e-02 4.60034639e-01 8.40383112e-01
4.22741890e-01 -4.41945298e-03 -8.74819696e-01 -4.26373892e-02
2.00551808e-01 -2.59198904e-01 -9.51720029e-02 8.51578936e-02
-1.22560513e+00 7.35487103e-01 2.88931549e-01 -3.05961035e-02
-4.57352638e-01 3.78287017e-01 1.83931872e-01 5.81790060e-02
5.14631808e-01 8.46981227e-01 -9.91512299e-01 1.14956209e-02
-7.97212780e-01 6.29500687e-01 1.21008396e+00 8.48879635e-01
4.93073404e-01 -3.87220025e-01 -4.06285882e-01 4.21855897e-01
3.94088238e-01 1.65529236e-01 2.25950778e-01 -1.13124979e+00
3.40284854e-01 9.11820233e-01 4.00931388e-02 -7.37513661e-01
-4.59104329e-01 -5.51244736e-01 -6.61944866e-01 3.16815600e-02
5.87517858e-01 4.90117185e-02 -9.67530370e-01 1.73635948e+00
2.92966124e-02 7.78369382e-02 1.24721706e-01 5.28658092e-01
7.33467758e-01 6.41468644e-01 2.40198717e-01 1.21256776e-01
1.32677734e+00 -1.05385768e+00 -5.40238321e-01 -6.41328514e-01
7.79285610e-01 -2.20308319e-01 1.22621262e+00 8.09035361e-01
-9.88096833e-01 -3.08357805e-01 -9.44765747e-01 -4.28631842e-01
-1.49697900e-01 1.29580945e-01 7.76152730e-01 1.97546259e-02
-8.42241704e-01 3.49339873e-01 -6.87842727e-01 -2.34900266e-02
5.88353992e-01 2.64548451e-01 -2.65572846e-01 -4.98644374e-02
-1.18497145e+00 1.11863387e+00 7.03572214e-01 -1.59404017e-02
-8.81129146e-01 -1.07488620e+00 -8.37692797e-01 4.00726408e-01
7.60805309e-01 -1.02414656e+00 1.45641184e+00 -1.01396799e+00
-1.09481156e+00 9.57254827e-01 -5.89250147e-01 -1.10156870e+00
3.96533817e-01 -4.80839223e-01 -1.43630758e-01 -1.98704332e-01
2.03540519e-01 8.47048104e-01 5.82586765e-01 -1.40072453e+00
-9.26770270e-01 -1.26810923e-01 4.02350038e-01 -1.19128771e-01
1.12751745e-01 -1.95305690e-01 -3.67009670e-01 -3.37508649e-01
2.72062123e-02 -8.68842006e-01 4.08055931e-02 1.71645191e-02
-8.37789536e-01 -6.65206969e-01 5.25512815e-01 -4.04476464e-01
1.12573099e+00 -1.77775395e+00 6.74465224e-02 -2.87533067e-02
4.15718019e-01 -3.48892435e-02 8.64218026e-02 5.10148518e-03
-1.36749536e-01 5.94386935e-01 -2.05676839e-01 -6.89937413e-01
3.43936265e-01 4.72689688e-01 -8.34965050e-01 -1.58007160e-01
6.06472194e-01 1.03569567e+00 -8.27461660e-01 -4.85896289e-01
1.31497188e-02 6.10897169e-02 -9.77423191e-01 2.43142605e-01
-1.02616239e+00 1.70280322e-01 -3.09845090e-01 4.14871484e-01
1.10890776e-01 -6.74404562e-01 1.61697403e-01 -2.91330755e-01
2.91705281e-01 8.62401128e-01 -8.31154048e-01 1.30816078e+00
-5.50322056e-01 5.54542720e-01 -3.98795575e-01 -9.97353435e-01
6.56927228e-01 2.95277983e-01 2.82627698e-02 -4.23855692e-01
-3.01168233e-01 1.12590879e-01 3.40153724e-01 -5.61728776e-01
2.63584912e-01 -3.69579881e-01 -6.85878694e-02 6.33131266e-01
1.50680870e-01 -1.56148016e-01 3.18735123e-01 3.96248490e-01
1.07127738e+00 3.03932697e-01 5.61300695e-01 -2.58514881e-01
3.96745741e-01 1.70782954e-01 7.10199416e-01 1.07095492e+00
2.09570557e-01 5.24652958e-01 8.91409814e-01 -1.01531518e+00
-1.03482378e+00 -8.30207467e-01 2.62270309e-02 1.14351058e+00
-2.21852124e-01 -7.41082549e-01 -4.26946104e-01 -9.75582719e-01
5.14802672e-02 1.49965572e+00 -9.82840836e-01 -6.22885348e-03
-3.52083474e-01 -5.00305712e-01 5.28882146e-01 6.01703584e-01
2.59678632e-01 -1.21545696e+00 -4.00948793e-01 2.27977097e-01
-4.02110726e-01 -1.31594157e+00 1.79290980e-01 3.36889088e-01
-5.03979802e-01 -1.07173717e+00 5.36454916e-01 -3.56209248e-01
5.63756287e-01 -3.93610001e-01 1.81667399e+00 4.30326432e-01
-4.26659435e-02 -4.61704433e-02 -2.14758664e-01 -9.15666699e-01
-7.88851857e-01 3.52415413e-01 -2.73190369e-03 -1.22052707e-01
4.44439888e-01 -3.88136655e-01 -2.78451025e-01 7.06666335e-02
-7.98384845e-01 5.45640528e-01 4.38175350e-01 9.65793669e-01
6.29985690e-01 9.69583839e-02 4.37505662e-01 -1.31325221e+00
4.01561409e-01 -5.52638054e-01 -3.35497618e-01 5.22175848e-01
-7.11374938e-01 7.31055498e-01 1.06460893e+00 3.33429314e-02
-1.24407208e+00 -1.22402973e-01 1.70227382e-02 3.31973694e-02
-3.35241497e-01 8.71037483e-01 -1.91666439e-01 8.52123976e-01
8.87711108e-01 -1.20277166e-01 -4.34598684e-01 -3.20099056e-01
5.75423062e-01 2.93858826e-01 6.83688343e-01 -1.06486499e+00
5.99278569e-01 4.60321486e-01 1.15084812e-01 -2.43864998e-01
-2.00764680e+00 7.68854022e-02 -6.30045414e-01 2.26985782e-01
7.25691378e-01 -1.01306844e+00 -8.26683283e-01 1.05028160e-01
-1.42555964e+00 -4.62387264e-01 -3.11106324e-01 -2.85138283e-02
-5.05230010e-01 -1.19764499e-01 -3.05969328e-01 -8.44059527e-01
-3.36586028e-01 -1.01183212e+00 1.04785061e+00 -6.84914812e-02
-9.80872989e-01 -1.16272748e+00 -1.89384773e-01 6.69393182e-01
5.84926456e-02 8.76132622e-02 1.50300992e+00 -1.00209618e+00
-7.22972929e-01 -1.18482687e-01 -1.97064668e-01 2.93048143e-01
-1.25033140e-01 2.68319845e-01 -1.17553198e+00 3.27751219e-01
-1.85119778e-01 -7.41918206e-01 9.97993052e-01 2.54800767e-01
1.54966640e+00 -8.32459450e-01 -4.13261324e-01 2.62491882e-01
9.90745902e-01 -2.40755126e-01 5.02242744e-01 3.22596461e-01
6.16104603e-01 5.38556516e-01 4.37486202e-01 1.49839669e-01
6.54700577e-01 6.33912325e-01 5.68179190e-01 4.59284671e-02
-1.02238068e-02 -5.95031500e-01 2.61626720e-01 5.22527844e-02
-1.61632553e-01 -6.13769889e-01 -1.20167446e+00 6.60959065e-01
-2.09806466e+00 -1.09080553e+00 -1.21516824e-01 1.84893358e+00
1.19314730e+00 6.66900754e-01 -2.89827108e-01 1.35820299e-01
-6.38152808e-02 4.89568673e-02 -6.57811522e-01 -2.31489167e-01
-1.90161407e-01 2.18619302e-01 6.83102831e-02 8.24605048e-01
-9.32252109e-01 1.09136891e+00 6.35519457e+00 7.74295986e-01
-7.44636595e-01 -1.20938018e-01 1.01218498e+00 -1.67199492e-01
-6.42981172e-01 2.52726585e-01 -1.00511956e+00 2.20569193e-01
9.31618392e-01 -4.68757749e-02 4.12677854e-01 8.84761870e-01
2.49483213e-01 -2.80004263e-01 -1.86269248e+00 5.25850117e-01
5.59222251e-02 -1.64631438e+00 5.43219686e-01 -7.88942128e-02
7.04049170e-01 -7.36869276e-02 -1.42137483e-01 3.45018864e-01
7.12603211e-01 -1.61575198e+00 1.05447769e+00 6.22748852e-01
4.81938601e-01 -3.86961430e-01 6.36008441e-01 5.84122717e-01
-5.89911103e-01 -1.75964072e-01 -9.53976735e-02 -6.39222920e-01
8.27952698e-02 9.01092768e-01 -1.23566139e+00 3.22894216e-01
3.93858105e-01 9.05896485e-01 -6.27664745e-01 3.84052902e-01
-9.88044143e-01 1.11791360e+00 -3.50735575e-01 7.97571465e-02
1.87305421e-01 2.50339568e-01 4.17351514e-01 1.11474442e+00
-1.00552455e-01 5.21613955e-02 1.49115198e-03 1.40041280e+00
-3.73398691e-01 -4.72375572e-01 -4.43517536e-01 -1.21868096e-01
4.91947740e-01 9.86299455e-01 -2.93829262e-01 -6.38671100e-01
-2.25942746e-01 5.98238170e-01 7.18861938e-01 1.77220017e-01
-7.89431512e-01 2.09600806e-01 6.57838345e-01 6.62775487e-02
9.39994082e-02 4.87360656e-02 -8.41756701e-01 -1.23764491e+00
1.07506603e-01 -1.11074030e+00 5.52379251e-01 -1.03866088e+00
-1.38907945e+00 3.92706901e-01 9.96001735e-02 -5.32595098e-01
-7.04805374e-01 -9.08424139e-01 -6.11533403e-01 6.60574138e-01
-1.39759672e+00 -1.34620869e+00 -9.47652087e-02 4.72543426e-02
6.29114628e-01 3.29560451e-02 1.00041521e+00 -3.36677939e-01
-4.49620515e-01 4.20724154e-01 -4.24782693e-01 1.55454665e-01
5.91961801e-01 -1.56768298e+00 7.31651664e-01 8.76764655e-01
5.61475635e-01 8.41654122e-01 9.36076105e-01 -5.12148201e-01
-1.12372088e+00 -9.44536924e-01 1.35616982e+00 -1.02976072e+00
7.51395285e-01 -3.66052866e-01 -8.35971951e-01 1.30581665e+00
1.15081787e-01 -6.11786060e-02 7.13574409e-01 8.30438852e-01
-7.23747313e-01 1.34963080e-01 -7.86871791e-01 6.59133673e-01
1.29339671e+00 -3.49191725e-01 -1.01970315e+00 4.27313119e-01
8.50904405e-01 -3.34682077e-01 -6.86102629e-01 3.54762077e-01
7.68599570e-01 -9.32642937e-01 8.74006391e-01 -1.15316617e+00
1.28004587e+00 -4.18963522e-01 -3.72779556e-02 -1.33914125e+00
-3.50404412e-01 -3.25326294e-01 -3.35504204e-01 1.05217731e+00
1.27037239e+00 -4.66313869e-01 6.00323498e-01 1.17513919e+00
-1.67450324e-01 -1.00352585e+00 -6.18576765e-01 -3.63766193e-01
-4.87693585e-02 -8.35040092e-01 6.20853841e-01 7.31226563e-01
3.42118703e-02 9.50942218e-01 -2.84852356e-01 1.28609598e-01
5.74084818e-01 3.13737959e-01 8.36937129e-01 -1.36459374e+00
-6.69511497e-01 -4.13270473e-01 -3.56509835e-02 -1.01539445e+00
5.98787546e-01 -1.07046330e+00 1.19294517e-01 -1.76055515e+00
4.22603190e-01 -4.80370402e-01 -9.95111391e-02 1.13160074e+00
-5.00367761e-01 -4.15050238e-02 2.31844723e-01 1.09125465e-01
-6.97503865e-01 2.85216421e-01 8.84023964e-01 -3.08030695e-01
2.22052589e-01 -1.57323897e-01 -1.16404510e+00 1.03872156e+00
5.65752625e-01 -3.89147222e-01 -5.12508333e-01 -7.69078732e-01
8.56669009e-01 -4.04762834e-01 6.02377415e-01 -7.53843486e-01
1.16397060e-01 -3.00902516e-01 4.02644575e-01 -3.08781475e-01
2.08467588e-01 -6.45956039e-01 1.58067793e-01 2.74596512e-01
-9.37355995e-01 -2.97408611e-01 1.18647456e-01 4.52504873e-01
-7.37330392e-02 -1.23320006e-01 4.79234725e-01 -9.96231213e-02
-5.15992939e-01 4.72815298e-02 -5.20651452e-02 4.45324093e-01
4.33684051e-01 1.28860235e-01 -5.52915394e-01 -5.26960552e-01
-6.37226045e-01 2.90439755e-01 2.14647338e-01 2.85738647e-01
3.19883764e-01 -8.38746727e-01 -8.90083194e-01 3.59658934e-02
2.76898891e-01 6.26166105e-01 -2.41101086e-01 6.01783574e-01
-1.33386120e-01 5.80729425e-01 2.02108964e-01 -4.73144263e-01
-9.25834477e-01 3.12271923e-01 3.84588361e-01 -8.20008039e-01
-3.96787286e-01 9.56527650e-01 5.67052603e-01 -4.29065585e-01
1.09556489e-01 -8.25931668e-01 8.24512392e-02 -1.79401502e-01
5.85884631e-01 -9.60481688e-02 4.53347620e-03 -1.72160476e-01
-2.99951881e-01 4.35514189e-02 -3.20249468e-01 1.21217472e-02
1.46425068e+00 2.17211634e-01 -1.09040610e-01 6.00014925e-01
6.99690759e-01 -7.73974359e-02 -1.23725653e+00 -2.54275560e-01
1.98460132e-01 -1.50803581e-01 -9.91916433e-02 -1.50481117e+00
-6.31871998e-01 8.58145297e-01 -4.45257723e-01 1.46697268e-01
7.59156942e-01 3.09617698e-01 5.09550154e-01 7.07853615e-01
1.80875391e-01 -7.33220518e-01 -1.54462010e-01 9.14539337e-01
1.15295196e+00 -1.44450736e+00 1.50415555e-01 -4.17599559e-01
-7.56567299e-01 1.04312897e+00 6.86389565e-01 4.62700129e-02
1.52116820e-01 3.76980662e-01 -1.08250402e-01 -3.42709243e-01
-1.46859908e+00 1.81202829e-01 4.14574504e-01 2.75753438e-01
6.58127606e-01 2.21943974e-01 1.29936904e-01 9.16726053e-01
-5.82784951e-01 2.00456381e-01 3.52708280e-01 2.99607038e-01
-3.35861892e-01 -9.39818203e-01 -5.35846986e-02 6.91189885e-01
-4.49579954e-01 -4.58997041e-01 -6.64103091e-01 7.96914995e-01
2.07782164e-01 8.80052030e-01 -4.83413339e-02 -2.67083615e-01
1.51897714e-01 5.14118135e-01 3.42561275e-01 -7.37129331e-01
-3.54183435e-01 -6.26310229e-01 6.58064902e-01 -5.59646547e-01
-3.90160292e-01 -5.67199051e-01 -1.49151897e+00 -1.76225349e-01
3.28661688e-02 -7.33411079e-03 2.98262805e-01 1.54849637e+00
4.87083465e-01 4.13330644e-01 1.61347911e-02 -4.65187788e-01
-5.19919097e-01 -8.96697521e-01 9.66585986e-03 7.19924212e-01
1.92310274e-01 -6.13394797e-01 -3.41770798e-01 4.31166381e-01] | [9.529972076416016, 6.950017929077148] |
5532790a-93fb-4ebd-a340-a8998a76053e | a-simple-multi-modality-transfer-learning | 2203.04287 | null | https://arxiv.org/abs/2203.04287v2 | https://arxiv.org/pdf/2203.04287v2.pdf | A Simple Multi-Modality Transfer Learning Baseline for Sign Language Translation | This paper proposes a simple transfer learning baseline for sign language translation. Existing sign language datasets (e.g. PHOENIX-2014T, CSL-Daily) contain only about 10K-20K pairs of sign videos, gloss annotations and texts, which are an order of magnitude smaller than typical parallel data for training spoken language translation models. Data is thus a bottleneck for training effective sign language translation models. To mitigate this problem, we propose to progressively pretrain the model from general-domain datasets that include a large amount of external supervision to within-domain datasets. Concretely, we pretrain the sign-to-gloss visual network on the general domain of human actions and the within-domain of a sign-to-gloss dataset, and pretrain the gloss-to-text translation network on the general domain of a multilingual corpus and the within-domain of a gloss-to-text corpus. The joint model is fine-tuned with an additional module named the visual-language mapper that connects the two networks. This simple baseline surpasses the previous state-of-the-art results on two sign language translation benchmarks, demonstrating the effectiveness of transfer learning. With its simplicity and strong performance, this approach can serve as a solid baseline for future research. Code and models are available at: https://github.com/FangyunWei/SLRT. | ['Stephen Lin', 'Zhirong Wu', 'Xiao Sun', 'Fangyun Wei', 'Yutong Chen'] | 2022-03-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Chen_A_Simple_Multi-Modality_Transfer_Learning_Baseline_for_Sign_Language_Translation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Chen_A_Simple_Multi-Modality_Transfer_Learning_Baseline_for_Sign_Language_Translation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['sign-language-recognition', 'sign-language-translation'] | ['computer-vision', 'computer-vision'] | [ 1.08136535e-02 -8.95656124e-02 -4.84498411e-01 -6.33747458e-01
-1.17504764e+00 -6.38440430e-01 9.13808703e-01 -9.04269457e-01
-7.30929315e-01 6.07726932e-01 5.17328262e-01 -3.05562496e-01
4.53957081e-01 -3.54979187e-01 -9.14850771e-01 -5.75730979e-01
3.00183952e-01 7.55178094e-01 2.15520963e-01 -2.39367604e-01
-2.79586911e-01 -2.30938625e-02 -1.30769479e+00 4.92288888e-01
6.44393504e-01 6.82498574e-01 6.23251349e-02 7.26643622e-01
7.21779186e-03 7.36191809e-01 -4.24685627e-01 -3.94039005e-01
4.04348880e-01 -7.79026866e-01 -8.43551755e-01 -2.35126823e-01
1.13573122e+00 -9.10349905e-01 -5.84534705e-01 6.76924109e-01
8.49239469e-01 -1.48230970e-01 6.22577965e-01 -1.52357686e+00
-9.69460487e-01 5.48768282e-01 -1.71049252e-01 -2.74143368e-01
1.73100919e-01 6.33969486e-01 1.11102808e+00 -1.00407505e+00
1.18275177e+00 1.17866921e+00 4.61161345e-01 9.59330797e-01
-7.53761053e-01 -8.37407649e-01 1.83354139e-01 1.98308319e-01
-1.03833246e+00 -4.70090687e-01 4.45769548e-01 -2.90686250e-01
1.05409503e+00 -1.22799791e-01 7.82303631e-01 1.64209390e+00
-2.55960763e-01 1.31215143e+00 1.24314940e+00 -3.81623715e-01
-1.59055069e-01 -2.81321704e-01 -1.06870934e-01 6.92603528e-01
-6.15198389e-02 2.53535420e-01 -9.26200986e-01 1.15532592e-01
7.95923054e-01 -3.58618140e-01 -3.70416820e-01 -4.58355695e-01
-1.41760957e+00 5.61863542e-01 3.99572283e-01 1.67618707e-01
-1.10350475e-01 5.00667751e-01 6.48696244e-01 7.69642055e-01
1.03829212e-01 -1.10040903e-01 -5.99858284e-01 -3.89345199e-01
-7.25657940e-01 2.49902055e-01 8.70090902e-01 1.36702609e+00
4.17126983e-01 -7.91364443e-03 -1.87813133e-01 7.32145488e-01
5.62869847e-01 1.10322821e+00 4.46374327e-01 -7.77312458e-01
9.09814775e-01 3.21910292e-01 -8.45509917e-02 4.79925759e-02
-1.03623070e-01 -6.59910068e-02 -3.29442799e-01 4.59819734e-01
8.57942760e-01 -2.73839086e-01 -1.46778226e+00 1.77232301e+00
1.51769072e-01 7.85149336e-02 1.71178188e-02 1.27065647e+00
9.21983659e-01 3.86198580e-01 2.22677678e-01 4.56325680e-01
1.22580814e+00 -1.38495648e+00 -4.54366058e-01 -3.27516139e-01
7.94433177e-01 -8.64943504e-01 1.55606043e+00 4.13499065e-02
-1.01060712e+00 -2.34146446e-01 -7.25581110e-01 -4.78955448e-01
-3.67001086e-01 5.11531234e-01 3.52010548e-01 6.17767386e-02
-1.11188626e+00 1.76179394e-01 -1.02120209e+00 -1.01178837e+00
4.14809704e-01 2.01572806e-01 -4.74307269e-01 -4.56184804e-01
-9.91355777e-01 1.18629324e+00 4.44567427e-02 1.56360120e-01
-9.75305259e-01 -5.76165855e-01 -7.08742261e-01 -5.19043267e-01
5.75833656e-02 -9.23962474e-01 1.64617336e+00 -1.00692952e+00
-1.85723424e+00 1.25782168e+00 4.30099741e-02 -1.10456094e-01
1.08829939e+00 -3.32922369e-01 -2.82875746e-01 7.27084503e-02
-7.14293122e-02 8.77115786e-01 9.06899273e-01 -1.04995573e+00
-6.88501477e-01 -1.49606779e-01 3.41120549e-02 4.05606896e-01
-2.48256363e-02 3.29617798e-01 -7.69385636e-01 -5.72418571e-01
-2.82072842e-01 -1.19312036e+00 1.94138542e-01 4.52432394e-01
8.41898471e-03 -2.44454458e-01 8.89513254e-01 -1.07869518e+00
7.62244642e-01 -2.04177213e+00 3.50288987e-01 -1.62083167e-03
-1.35774508e-01 5.00811398e-01 -7.82867730e-01 5.70145845e-01
1.77767262e-01 -3.14253479e-01 -2.98778653e-01 -5.81588507e-01
2.72076547e-01 5.10553598e-01 -2.07014948e-01 4.87274915e-01
1.17063254e-01 1.39968288e+00 -1.00146544e+00 -4.66107756e-01
1.14524975e-01 5.91621518e-01 -3.94851267e-01 2.65545905e-01
-3.74135256e-01 6.68552339e-01 -2.57814944e-01 7.76499033e-01
2.81495750e-01 -1.84540838e-01 9.44477618e-02 -1.76572397e-01
5.53623587e-02 7.32986212e-01 -5.85518420e-01 2.13612008e+00
-5.29169023e-01 9.54137325e-01 1.45635575e-01 -7.37845004e-01
6.90457165e-01 4.12249774e-01 3.44040155e-01 -9.58103001e-01
2.30883047e-01 6.95074797e-01 -1.21821987e-03 -7.63213933e-01
6.01155050e-02 -1.99969009e-01 -8.42885002e-02 8.18709433e-01
3.67945164e-01 -2.93663144e-01 2.89043486e-01 4.05175574e-02
1.00353348e+00 7.38889575e-01 -1.75632104e-01 2.23689809e-01
1.83174595e-01 1.21126547e-01 3.15309107e-01 3.35267961e-01
-3.53051126e-01 6.58796906e-01 1.58350497e-01 -2.51076669e-01
-1.18181884e+00 -1.19799161e+00 2.42060810e-01 1.35592246e+00
-2.38387093e-01 -2.14420393e-01 -6.83978140e-01 -9.38526988e-01
1.22960053e-01 6.05599344e-01 -4.81150359e-01 -6.61344826e-02
-1.04370522e+00 -1.68059796e-01 1.16686594e+00 9.14402068e-01
6.36281192e-01 -1.29064298e+00 -4.98615801e-01 -2.29534477e-01
-3.90405297e-01 -1.25048792e+00 -8.56037438e-01 -1.47097230e-01
-6.11761332e-01 -1.01090634e+00 -1.16501057e+00 -1.15243733e+00
6.17190659e-01 -1.60134301e-01 1.22072423e+00 4.73021604e-02
-3.96469124e-02 5.85770309e-01 -4.23722804e-01 -2.56293714e-01
-7.14253545e-01 1.30892843e-01 -1.84402257e-01 -2.68085629e-01
6.56417727e-01 -4.32088733e-01 -3.50098252e-01 3.68260384e-01
-7.64620364e-01 1.97999209e-01 7.65339732e-01 9.89711463e-01
3.18146169e-01 -1.23468840e+00 2.77639300e-01 -1.97758853e-01
5.55174828e-01 -2.38133054e-02 -5.67784667e-01 5.58820188e-01
-3.31862003e-01 9.94649529e-02 3.40963364e-01 -6.45106792e-01
-9.02603030e-01 9.75314900e-02 5.68600558e-02 -5.15689492e-01
-4.77024764e-02 4.59776223e-01 -1.65660456e-01 9.81295202e-03
5.26285708e-01 3.73616636e-01 3.18165660e-01 -4.34233367e-01
7.85174310e-01 8.49265397e-01 6.76122665e-01 -6.11708999e-01
8.82252038e-01 3.87199193e-01 -3.06903332e-01 -5.12625873e-01
-3.99254918e-01 -4.10502791e-01 -9.18755114e-01 -3.02552193e-01
7.76811182e-01 -1.07308912e+00 -3.50318402e-01 8.25442135e-01
-1.03559601e+00 -1.21508527e+00 -3.65039438e-01 7.38190711e-01
-9.22711790e-01 3.31374317e-01 -7.75563657e-01 -8.34958032e-02
-4.25153136e-01 -9.74153757e-01 1.39067078e+00 -1.71351448e-01
-2.60702521e-01 -9.13012087e-01 2.65251756e-01 5.92072070e-01
4.85899746e-01 2.61334442e-02 5.59808791e-01 -2.85477102e-01
-6.70952678e-01 -1.76213995e-01 -4.06918734e-01 7.40231872e-01
1.50887802e-01 -1.01245036e-02 -7.03323066e-01 -4.90107805e-01
-7.66462326e-01 -9.04288471e-01 9.39426482e-01 6.75069094e-02
1.38928667e-01 -9.34445560e-02 -1.31471798e-01 6.95054710e-01
1.00819433e+00 -2.40692079e-01 5.47934830e-01 2.95760095e-01
7.84466505e-01 3.62394899e-01 6.44396663e-01 6.59395847e-03
8.55482399e-01 8.76947284e-01 9.23117902e-03 -2.81256169e-01
-9.14273739e-01 -5.82425594e-01 7.97679424e-01 1.16094244e+00
-4.18820292e-01 -1.38756886e-01 -1.04601705e+00 8.34449351e-01
-1.99594975e+00 -8.16673934e-01 1.92959398e-01 2.03972960e+00
1.04677999e+00 -4.30913508e-01 2.90614218e-01 -4.16683316e-01
1.97876677e-01 9.42118391e-02 -5.41340649e-01 -2.52466679e-01
-2.02589542e-01 1.25435904e-01 6.35369837e-01 5.79559863e-01
-1.02449727e+00 1.42046380e+00 5.84646225e+00 4.52315062e-01
-1.43486023e+00 2.33878255e-01 -2.61190534e-01 -4.65531260e-01
-1.82187378e-01 4.90652286e-02 -7.81847537e-01 2.74372369e-01
8.69054854e-01 4.26495299e-02 6.97867155e-01 5.79532325e-01
2.29318544e-01 2.68949270e-01 -1.39362395e+00 8.10398340e-01
3.54437292e-01 -8.71154785e-01 2.08634481e-01 1.77812994e-01
7.64007688e-01 9.57196116e-01 -5.50308526e-02 5.37500501e-01
6.87922120e-01 -8.51086378e-01 9.06159103e-01 3.67941320e-01
1.25029457e+00 -7.40668327e-02 5.89265943e-01 1.31735101e-01
-1.12482059e+00 2.15795174e-01 3.25592831e-02 6.63835704e-02
6.86607778e-01 -1.64131910e-01 -7.35946178e-01 3.09608757e-01
6.21366858e-01 9.24063563e-01 -2.79994488e-01 9.36813116e-01
-8.28261673e-01 7.81307280e-01 -6.15633905e-01 4.82223891e-02
5.61792850e-01 -2.71305978e-01 3.66769552e-01 1.40311372e+00
3.29955459e-01 -3.05510372e-01 5.05343415e-02 5.80346823e-01
-2.86517918e-01 1.69318080e-01 -6.99951053e-01 -2.63711959e-01
8.22822377e-02 6.79185092e-01 -2.03530505e-01 -6.85669363e-01
-8.34312856e-01 1.26821852e+00 3.36069286e-01 7.75929570e-01
-9.15032387e-01 -1.56352326e-01 8.27499270e-01 -1.42702639e-01
4.22820300e-01 -3.77226293e-01 -5.30460924e-02 -1.50744951e+00
4.06231254e-01 -1.29878855e+00 2.69582599e-01 -8.11083734e-01
-1.30641651e+00 3.26462865e-01 -3.69497985e-02 -1.66437459e+00
-5.15135288e-01 -9.19605076e-01 -4.31429297e-01 8.31929266e-01
-1.73056376e+00 -1.92940795e+00 -4.43658859e-01 9.57274020e-01
4.48134214e-01 -1.80569544e-01 7.37378359e-01 5.46065927e-01
-2.32136771e-02 8.80760908e-01 3.39170843e-02 6.83812678e-01
1.31487334e+00 -9.47682142e-01 6.87159598e-01 7.44049907e-01
3.53688210e-01 3.48735392e-01 2.88933128e-01 -5.82341611e-01
-1.29169476e+00 -9.85031009e-01 1.28815985e+00 -8.04061294e-01
9.04047489e-01 -3.95412445e-01 -5.24733424e-01 1.14119959e+00
4.16463256e-01 9.81500447e-02 3.13678771e-01 -2.39037097e-01
-8.22478592e-01 8.43460709e-02 -6.78799331e-01 8.80236268e-01
1.53702343e+00 -9.51905072e-01 -6.40775144e-01 5.01342654e-01
4.27581102e-01 -7.50977397e-01 -8.53184581e-01 2.34935552e-01
1.20518982e+00 -4.54927146e-01 7.23117948e-01 -8.60707223e-01
5.70558488e-01 -2.98580706e-01 -2.26140946e-01 -1.35062206e+00
1.62768155e-01 -5.62978089e-01 5.83106689e-02 9.69576776e-01
5.82087040e-01 -6.02147400e-01 6.84096456e-01 5.23137510e-01
-2.44759247e-01 -4.56779599e-01 -1.09016716e+00 -1.10505104e+00
4.62175220e-01 -3.44375610e-01 3.53585899e-01 9.73741651e-01
3.80406119e-02 3.82932991e-01 -5.47452271e-01 -2.18539044e-01
6.18814290e-01 9.72797945e-02 1.29392743e+00 -6.71601713e-01
-3.75327915e-01 -6.29360020e-01 -2.70185560e-01 -1.37931395e+00
1.90753594e-01 -1.16647279e+00 2.26484671e-01 -1.92198181e+00
-4.88907248e-02 -9.14305598e-02 -2.73561813e-02 9.97374952e-01
1.58677176e-01 3.07482123e-01 4.70164448e-01 2.74138540e-01
-3.29399407e-01 7.19022751e-01 1.60597742e+00 -2.74727464e-01
-3.25785875e-02 -1.65335402e-01 -5.18699028e-02 5.11602998e-01
7.30589032e-01 -3.33303362e-01 -2.28834078e-01 -1.10731244e+00
-2.05516875e-01 -3.51371229e-01 4.80078995e-01 -5.87390959e-01
2.84878582e-01 -5.56101948e-02 -2.50940740e-01 -3.36551458e-01
3.68280321e-01 -7.71115065e-01 -3.34230572e-01 3.08201909e-01
-3.22458923e-01 -1.45121679e-01 1.96128607e-01 1.75039038e-01
-3.84106874e-01 2.89977819e-01 6.08214855e-01 -2.44250651e-02
-8.04251552e-01 2.65543401e-01 -1.92826495e-01 2.24804431e-01
6.32092237e-01 -1.14982791e-01 -6.81277156e-01 -6.46422803e-01
-4.29824322e-01 4.40666139e-01 5.95295310e-01 8.65409374e-01
4.34427828e-01 -1.45010746e+00 -9.72780883e-01 2.39772424e-01
4.85092372e-01 -1.32780194e-01 -1.65933236e-01 1.04924870e+00
-5.77736020e-01 5.43338418e-01 -5.40488124e-01 -6.07036233e-01
-1.42136025e+00 -1.04445525e-01 3.37404966e-01 -1.54498160e-01
-8.16334784e-01 9.00077522e-01 -3.71394493e-02 -1.01994169e+00
3.35312814e-01 -8.50182474e-01 4.79478478e-01 -2.87793219e-01
2.19787806e-01 1.63829699e-01 -2.35519841e-01 -8.71841133e-01
-3.59571606e-01 9.55545664e-01 1.59939155e-01 -5.21470189e-01
1.21132386e+00 9.09744352e-02 5.17449575e-03 3.41250092e-01
1.24323928e+00 -2.82962769e-01 -1.41169095e+00 -5.36943972e-01
-5.74398227e-02 -6.88470900e-02 -3.08528602e-01 -1.14896607e+00
-1.06708884e+00 1.04650986e+00 4.85646039e-01 -9.81479943e-01
1.04382265e+00 2.13310748e-01 8.67554009e-01 5.36026001e-01
4.72551167e-01 -1.28946412e+00 -1.10763693e-02 7.43019879e-01
1.28624713e+00 -1.42880571e+00 -2.95938194e-01 3.10191642e-02
-7.93329775e-01 8.28097403e-01 5.45144975e-01 -1.42442584e-01
4.12566423e-01 4.12290305e-01 8.38015318e-01 -1.61608141e-02
-7.11100221e-01 -3.44828099e-01 4.80289727e-01 6.67545140e-01
4.42541361e-01 2.15163052e-01 -3.42190713e-01 1.19225264e-01
-3.03487986e-01 6.20847762e-01 6.53524101e-02 1.08710468e+00
7.57473633e-02 -1.38310349e+00 -2.20635422e-02 6.38637692e-02
1.69168755e-01 -1.44377336e-01 -7.26450980e-01 1.12200809e+00
-1.14019038e-02 5.25213540e-01 -1.17155574e-01 -1.84486508e-01
6.27681255e-01 4.26961303e-01 7.59047687e-01 -2.82094002e-01
-5.08476555e-01 -1.58644021e-02 3.80009115e-01 -7.67991841e-01
-6.09437346e-01 -7.75297999e-01 -1.34770823e+00 -1.70200970e-02
5.64159155e-02 -5.05452275e-01 7.04351723e-01 1.11227310e+00
3.00737828e-01 5.84089048e-02 -2.01781482e-01 -7.20303833e-01
-7.57245660e-01 -1.27083194e+00 -2.37119302e-01 6.46138728e-01
2.53659964e-01 -3.89221370e-01 -2.04209849e-01 4.57590938e-01] | [9.315308570861816, -6.6402268409729] |
0448985d-6d22-4ff1-9fdf-ce313cf2affa | evaluation-of-speech-representations-for-mos | 2306.09979 | null | https://arxiv.org/abs/2306.09979v1 | https://arxiv.org/pdf/2306.09979v1.pdf | Evaluation of Speech Representations for MOS prediction | In this paper, we evaluate feature extraction models for predicting speech quality. We also propose a model architecture to compare embeddings of supervised learning and self-supervised learning models with embeddings of speaker verification models to predict the metric MOS. Our experiments were performed on the VCC2018 dataset and a Brazilian-Portuguese dataset called BRSpeechMOS, which was created for this work. The results show that the Whisper model is appropriate in all scenarios: with both the VCC2018 and BRSpeech- MOS datasets. Among the supervised and self-supervised learning models using BRSpeechMOS, Whisper-Small achieved the best linear correlation of 0.6980, and the speaker verification model, SpeakerNet, had linear correlation of 0.6963. Using VCC2018, the best supervised and self-supervised learning model, Whisper-Large, achieved linear correlation of 0.7274, and the best model speaker verification, TitaNet, achieved a linear correlation of 0.6933. Although the results of the speaker verification models are slightly lower, the SpeakerNet model has only 5M parameters, making it suitable for real-time applications, and the TitaNet model produces an embedding of size 192, the smallest among all the evaluated models. The experiment results are reproducible with publicly available source-code1 . | ['Arlindo R. Galvão Filho', 'Anderson S. Soares', 'Lucas R. S. Gris', 'Arnaldo Cândido Júnior', 'Edresson Casanova', 'Frederico S. Oliveira'] | 2023-06-16 | null | null | null | null | ['speaker-verification'] | ['speech'] | [-5.18385053e-01 1.09207623e-01 -1.99544504e-02 -5.05123675e-01
-8.63034010e-01 -3.66803467e-01 5.52623332e-01 1.77633747e-01
-4.06022757e-01 4.13554013e-01 4.59990591e-01 -3.27469438e-01
-7.36679882e-02 -4.27377909e-01 -2.81071782e-01 -4.90317196e-01
-3.22268575e-01 2.50192702e-01 5.24798892e-02 -3.67204130e-01
9.53338668e-02 2.29073763e-01 -1.88085866e+00 1.14939779e-01
7.53589809e-01 1.09074676e+00 1.48372486e-01 9.70836639e-01
-1.05333947e-01 6.97279453e-01 -9.33041275e-01 -6.09291553e-01
-3.64912525e-02 -3.71478111e-01 -5.92968702e-01 -3.94749492e-01
2.51536459e-01 1.44594893e-01 -2.80898452e-01 8.68559182e-01
8.94793272e-01 2.71460023e-02 6.23425722e-01 -1.51187325e+00
-7.66895950e-01 1.09530830e+00 4.01908085e-02 2.16278896e-01
6.38087571e-01 -6.80334195e-02 1.22424972e+00 -1.09204197e+00
2.18700707e-01 1.30888367e+00 5.85540235e-01 8.00819397e-01
-8.58931899e-01 -9.45048630e-01 -3.53626579e-01 4.86437589e-01
-1.67397904e+00 -8.39053154e-01 6.59511089e-01 -2.21250370e-01
9.79963422e-01 4.96041745e-01 3.84178311e-01 1.15613627e+00
-8.80190134e-02 6.28732026e-01 1.21795726e+00 -6.22156858e-01
4.84267116e-01 7.18495548e-01 2.69485593e-01 6.08773291e-01
-3.16793293e-01 4.79964346e-01 -7.99552858e-01 -2.36443415e-01
1.54753640e-01 -4.49168861e-01 -2.14995027e-01 2.70983577e-01
-1.13767099e+00 8.94589126e-01 1.66343525e-01 8.33930612e-01
-9.39965099e-02 -3.30892444e-01 3.91988069e-01 7.74616778e-01
3.52772593e-01 2.85914838e-01 -5.14643848e-01 -5.83548307e-01
-1.14758217e+00 -3.26188296e-01 1.09896195e+00 9.67622221e-01
6.52159631e-01 4.43431526e-01 6.58851117e-02 1.22944927e+00
5.26019752e-01 7.74132013e-01 9.80421782e-01 -7.23549128e-01
3.38404417e-01 2.39557177e-01 -1.41732410e-01 -8.80673349e-01
-4.71901238e-01 -2.92753041e-01 -7.12002218e-01 6.68508234e-03
1.89259842e-01 -2.02018172e-01 -4.94347095e-01 1.40949690e+00
-1.71340749e-01 1.12272903e-01 5.59277594e-01 6.38856888e-01
1.27351367e+00 9.79536891e-01 -1.86398104e-01 -1.37587428e-01
1.14743435e+00 -1.08400357e+00 -9.53384459e-01 2.12641410e-03
6.02586865e-01 -9.99789655e-01 1.13811576e+00 3.08657527e-01
-8.55281472e-01 -9.40473855e-01 -1.23153353e+00 3.44735652e-01
-4.94856477e-01 1.56426907e-01 4.37251888e-02 1.29530680e+00
-1.51915598e+00 3.49755824e-01 -5.34375250e-01 -4.75597471e-01
-1.13221683e-01 2.65943974e-01 -2.99099594e-01 2.94664413e-01
-1.30418420e+00 1.11565769e+00 2.31578853e-02 -1.31757379e-01
-9.34563637e-01 -4.30050999e-01 -8.50564539e-01 2.99021125e-01
-3.86859745e-01 2.30805695e-01 1.10988450e+00 -7.45005190e-01
-1.95715261e+00 5.86369634e-01 -3.67804110e-01 -7.96566904e-01
1.63792372e-01 2.45763481e-01 -1.42576134e+00 -3.89130320e-03
-2.85553098e-01 7.02214718e-01 6.90720320e-01 -1.14764059e+00
-4.26667780e-01 -2.37684045e-02 -1.24886766e-01 -2.73534834e-01
-8.28831851e-01 3.08417886e-01 -2.44394783e-02 -2.75580436e-01
-9.84206647e-02 -8.12997222e-01 1.60965666e-01 -5.23823977e-01
-1.92911223e-01 -3.72150630e-01 8.41420710e-01 -6.69181347e-01
1.52262843e+00 -2.33934307e+00 -1.63419753e-01 1.88144118e-01
-1.51295543e-01 4.83915985e-01 -4.50576931e-01 6.64552569e-01
-1.76633745e-01 2.55076975e-01 -5.95732778e-02 -7.19001174e-01
2.77016044e-01 1.10139161e-01 2.57446188e-02 3.34770262e-01
1.69416606e-01 6.26035452e-01 -6.34204566e-01 -5.55210173e-01
4.41875756e-01 7.24482596e-01 -5.88590264e-01 3.25051397e-01
3.62197757e-01 1.43006325e-01 1.03176318e-01 5.71516275e-01
6.65675342e-01 2.30459169e-01 -2.27384083e-02 2.93733994e-03
-3.55911553e-01 5.13033152e-01 -1.24950111e+00 1.46887648e+00
-1.01509535e+00 9.74263787e-01 4.00292948e-02 -6.65456057e-01
1.65563107e+00 8.26644540e-01 2.92065918e-01 -5.80207527e-01
1.59698948e-01 5.32201171e-01 1.62798777e-01 -5.04052460e-01
5.28504431e-01 -3.21746953e-02 3.65375588e-03 2.89444506e-01
4.86070722e-01 -2.62914151e-01 2.19487756e-01 2.10927397e-01
8.88446212e-01 -6.59212351e-01 1.35682300e-01 -3.76428336e-01
1.04849350e+00 -5.77817023e-01 2.96431541e-01 2.39163190e-01
-6.93287611e-01 5.61109722e-01 2.95279682e-01 3.44059914e-02
-6.83143914e-01 -1.05960023e+00 -3.34487379e-01 8.58434260e-01
-7.97282085e-02 -6.72352374e-01 -7.83779025e-01 -5.30092895e-01
-1.72488913e-01 9.03771639e-01 -3.64070952e-01 -1.30885154e-01
-1.12616099e-01 -1.56396464e-01 8.97629261e-01 3.61785918e-01
2.99822062e-01 -1.15904570e+00 1.19851336e-01 1.14420883e-01
-2.83088267e-01 -1.13772106e+00 -6.20310247e-01 1.10108964e-01
-4.60477799e-01 -7.38563418e-01 -6.54310107e-01 -1.20660067e+00
1.11833163e-01 -3.23221743e-01 9.18551505e-01 -1.18758775e-01
1.18248202e-01 3.67326081e-01 -5.75407743e-01 -3.93271655e-01
-9.69869971e-01 -7.58888796e-02 5.21113873e-01 2.70367265e-01
5.55337250e-01 -4.68233496e-01 -1.08610898e-01 4.96821016e-01
-5.13966382e-01 -6.26595080e-01 1.62181109e-01 8.25256348e-01
1.04845420e-01 1.25831053e-01 1.05544686e+00 -1.08484186e-01
5.14415681e-01 -2.40093678e-01 -5.05940497e-01 7.93910548e-02
-6.64906502e-01 -1.30866230e-01 7.60523379e-01 -3.88515741e-01
-7.37925589e-01 -2.25907430e-01 -7.23982751e-01 -3.58052045e-01
-2.03973502e-01 3.79922062e-01 -3.48091334e-01 7.63489529e-02
6.78210795e-01 2.01750293e-01 4.02533226e-02 -5.78155994e-01
5.69741987e-02 1.51211548e+00 2.01710403e-01 -1.23467371e-01
7.72060454e-01 -6.03469498e-02 -7.77008295e-01 -1.44222629e+00
-3.39824915e-01 -5.68243206e-01 -4.92616802e-01 -2.96973228e-01
8.21691215e-01 -9.52738464e-01 -6.58575892e-01 4.85285580e-01
-1.12122655e+00 -1.09354720e-01 -3.54069531e-01 8.03390026e-01
-4.55888480e-01 1.55348420e-01 -5.63766778e-01 -1.25814760e+00
-3.22218210e-01 -1.20344651e+00 8.60819221e-01 1.95451900e-01
-3.56107205e-01 -1.24127316e+00 2.53410637e-01 3.88033062e-01
7.12099731e-01 -3.06690991e-01 5.66088557e-01 -1.18723381e+00
1.36403561e-01 -2.17981294e-01 1.43199950e-01 1.04644895e+00
4.16885078e-01 8.99222121e-02 -1.47028542e+00 -3.99084091e-01
1.61976330e-02 -2.86258280e-01 4.11528081e-01 -4.46473584e-02
1.03620470e+00 -1.09580472e-01 3.02489907e-01 2.20479757e-01
1.01011229e+00 3.36959600e-01 5.79394698e-01 1.26814485e-01
1.78698957e-01 6.33115411e-01 4.25903141e-01 2.31165543e-01
6.74872398e-01 6.36941552e-01 1.63346782e-01 8.06861967e-02
-4.23981756e-01 -3.11709106e-01 9.50957179e-01 1.94290555e+00
2.31096566e-01 -1.92053378e-01 -1.00741458e+00 4.51744109e-01
-1.20048726e+00 -9.94582415e-01 -2.60706007e-01 2.19952965e+00
7.99592316e-01 1.21108569e-01 2.04766527e-01 8.15750301e-01
6.60454273e-01 3.44861180e-01 5.35707735e-02 -9.86496925e-01
-3.31403583e-01 5.23255646e-01 -5.76696591e-03 8.56582344e-01
-1.04357469e+00 7.24050641e-01 6.67445278e+00 5.83482504e-01
-1.19987929e+00 2.02931449e-01 2.27152199e-01 3.49094979e-02
-1.35387238e-02 -3.42698038e-01 -9.09714520e-01 7.06806958e-01
1.99360216e+00 -3.03067505e-01 4.11002547e-01 7.20223963e-01
3.47433954e-01 3.21477890e-01 -1.28903031e+00 1.23113048e+00
4.33676869e-01 -8.21962535e-01 -2.52373278e-01 -1.42691001e-01
6.40944004e-01 1.82077065e-01 7.29816779e-02 6.07559025e-01
1.35570586e-01 -1.01283753e+00 6.79446220e-01 1.67368189e-01
7.38140166e-01 -8.38054836e-01 1.05838752e+00 2.30398893e-01
-1.18932331e+00 -3.61966938e-01 -4.61564302e-01 1.01090908e-01
3.74725647e-02 7.08458900e-01 -8.90503764e-01 4.72193450e-01
8.18655193e-01 7.17674017e-01 -7.45215118e-01 9.34155643e-01
-9.59655344e-02 1.08391178e+00 -1.00989796e-01 -4.95829314e-01
-3.03316154e-02 1.04874432e-01 4.60740864e-01 1.49336576e+00
3.78329039e-01 -5.26479483e-01 -1.62014559e-01 3.48339438e-01
2.64906084e-05 4.95361775e-01 -2.25744337e-01 -1.34248272e-01
7.94110358e-01 1.08201671e+00 3.21478955e-02 -3.67210686e-01
-3.93749028e-01 7.15976655e-01 3.88994589e-02 2.90696234e-01
-1.01618421e+00 -6.83400214e-01 6.45939469e-01 -1.36045918e-01
3.64581019e-01 -9.28445235e-02 -2.37019733e-01 -1.14529383e+00
-1.72428995e-01 -1.03515697e+00 9.99207497e-02 -5.91795743e-01
-1.31742752e+00 1.03058136e+00 -2.40814522e-01 -1.43950653e+00
-3.46504301e-01 -6.25615656e-01 -7.23205864e-01 1.01736057e+00
-1.53533816e+00 -6.35560393e-01 -4.59317565e-02 6.00814342e-01
5.00129759e-01 -7.55920112e-01 1.37095654e+00 5.62965989e-01
-5.31412244e-01 1.04161239e+00 2.46179521e-01 2.16333985e-01
6.99110508e-01 -1.19127762e+00 3.21424067e-01 3.99472505e-01
5.82616508e-01 4.13390726e-01 5.47421813e-01 7.34671503e-02
-1.21149433e+00 -9.68387902e-01 1.58238399e+00 -2.35297799e-01
8.64976227e-01 -4.87305582e-01 -9.05196726e-01 2.47422576e-01
5.35551250e-01 -4.01388593e-02 9.99466538e-01 1.72765553e-01
-6.37443900e-01 -4.06800151e-01 -1.34887946e+00 1.83368728e-01
6.59590542e-01 -1.11925662e+00 -6.29887104e-01 1.41170651e-01
6.90997362e-01 9.43681374e-02 -1.30345249e+00 -2.13182066e-02
4.56562459e-01 -9.60523546e-01 6.58949494e-01 -8.05947259e-02
1.02962755e-01 -2.57027745e-01 -6.42510355e-01 -1.62077272e+00
-1.73846986e-02 -4.62400913e-01 -2.54796237e-01 1.71755326e+00
9.80261385e-01 -1.00765252e+00 5.27875721e-01 1.26897797e-01
-5.86253516e-02 -3.04646045e-01 -1.26325750e+00 -1.27889085e+00
2.79992908e-01 -6.10589921e-01 9.05029178e-01 9.84928608e-01
2.41986871e-01 2.48162195e-01 -1.82042435e-01 2.41206065e-01
3.63725275e-01 -3.28995407e-01 7.02706575e-01 -1.12221003e+00
-2.14135766e-01 -3.00882995e-01 -8.42456639e-01 -8.25565815e-01
3.41375709e-01 -1.05210257e+00 3.43522839e-02 -1.29882014e+00
-2.20191166e-01 -3.96888763e-01 -5.38919032e-01 2.90251225e-01
1.48197278e-01 1.13057479e-01 3.47008795e-01 -1.66085914e-01
-1.17848717e-01 6.33495808e-01 7.60099471e-01 -5.01720786e-01
-3.78766477e-01 7.15552568e-02 -4.15101171e-01 4.16789412e-01
1.13232875e+00 -3.60037863e-01 -3.03291738e-01 -2.43100584e-01
-5.91776252e-01 -6.52531758e-02 8.27873796e-02 -1.33303368e+00
3.54661793e-01 6.25811517e-01 5.59518300e-02 -5.18090665e-01
6.94556355e-01 -7.62398541e-01 -2.13999674e-01 4.41248685e-01
-4.40458864e-01 -4.89935242e-02 2.22887963e-01 9.51356962e-02
-6.07767999e-01 -4.25251454e-01 8.15482080e-01 5.92019737e-01
-4.80081677e-01 2.17554346e-02 -6.44041419e-01 -9.21364035e-03
6.39497757e-01 -7.80098839e-03 -1.91469073e-01 -6.95974350e-01
-8.46840918e-01 1.15340814e-01 1.24268092e-01 9.68884885e-01
7.58000851e-01 -1.46022582e+00 -9.82869089e-01 5.27095735e-01
3.28867048e-01 -7.72016585e-01 -5.30066341e-02 6.86807871e-01
-1.57018021e-01 5.25630116e-01 1.37573674e-01 -7.18371332e-01
-1.45609605e+00 2.43357092e-01 3.06065321e-01 6.37036264e-02
-1.01762386e-02 6.72174513e-01 -4.60707039e-01 -1.11977565e+00
6.48245633e-01 -3.86162758e-01 -4.14501190e-01 1.69398248e-01
7.66111553e-01 6.67756140e-01 4.69888091e-01 -1.14691508e+00
-6.83985710e-01 3.90498906e-01 3.15714836e-01 -3.82380217e-01
1.22381151e+00 -7.96247572e-02 3.13865244e-02 8.52905691e-01
1.73566556e+00 3.15725356e-01 -4.36921090e-01 -1.11724354e-01
1.74598917e-01 -1.84914291e-01 2.87581950e-01 -9.18684125e-01
-1.03838527e+00 1.02875459e+00 1.13328850e+00 4.80411112e-01
9.94028449e-01 -8.16705525e-02 6.22604966e-01 2.90414125e-01
5.22516847e-01 -1.04916751e+00 -1.79126300e-02 6.81842864e-01
9.35050964e-01 -1.27929997e+00 -5.30639350e-01 -9.23136026e-02
-8.16091955e-01 9.74478006e-01 4.34247434e-01 2.11698934e-01
1.21688259e+00 3.83587927e-01 5.48488438e-01 2.14867607e-01
-7.55866051e-01 -1.11461163e-01 3.07313740e-01 8.63782823e-01
6.74671292e-01 3.49718302e-01 -1.77321032e-01 7.09883451e-01
-9.98146236e-01 -3.73320997e-01 5.04730582e-01 5.50127268e-01
-3.63405973e-01 -1.30494273e+00 -4.63497490e-01 1.20720193e-01
-1.93116114e-01 -4.56632711e-02 -5.67845404e-01 5.85861266e-01
-6.83049262e-02 1.85755193e+00 2.51775920e-01 -8.63302231e-01
6.70998216e-01 3.53432894e-01 -3.73424962e-02 -4.58439499e-01
-8.10910344e-01 -1.86137378e-01 2.97356635e-01 -3.05921465e-01
-4.73489851e-01 -5.43763101e-01 -1.28478730e+00 -4.58295166e-01
-6.06861532e-01 6.63578033e-01 1.06517172e+00 4.96405631e-01
1.72327861e-01 4.47324753e-01 1.38511598e+00 -5.53935468e-01
-6.11887097e-01 -1.39020324e+00 -6.38943493e-01 7.84448534e-02
5.03778875e-01 -3.08531255e-01 -8.06501985e-01 -1.68281585e-01] | [14.270849227905273, 6.166194438934326] |
f36af5c3-9359-4523-aa5d-35c6a86849f2 | provable-burer-monteiro-factorization-for-a | 1606.01316 | null | http://arxiv.org/abs/1606.01316v3 | http://arxiv.org/pdf/1606.01316v3.pdf | Provable Burer-Monteiro factorization for a class of norm-constrained matrix problems | We study the projected gradient descent method on low-rank matrix problems
with a strongly convex objective. We use the Burer-Monteiro factorization
approach to implicitly enforce low-rankness; such factorization introduces
non-convexity in the objective. We focus on constraint sets that include both
positive semi-definite (PSD) constraints and specific matrix norm-constraints.
Such criteria appear in quantum state tomography and phase retrieval
applications.
We show that non-convex projected gradient descent favors local linear
convergence in the factored space. We build our theory on a novel descent
lemma, that non-trivially extends recent results on the unconstrained problem.
The resulting algorithm is Projected Factored Gradient Descent, abbreviated as
ProjFGD, and shows superior performance compared to state of the art on quantum
state tomography and sparse phase retrieval applications. | ['Sujay Sanghavi', 'Dohyung Park', 'Srinadh Bhojanapalli', 'Constantine Caramanis', 'Anastasios Kyrillidis'] | 2016-06-04 | null | null | null | null | ['quantum-state-tomography'] | ['medical'] | [ 3.36713940e-01 1.00778401e-01 -1.96338370e-01 -2.30645284e-01
-1.08819973e+00 -6.52772307e-01 5.41597426e-01 -5.04021466e-01
-6.65352345e-01 9.62606728e-01 4.46379095e-01 -6.11713350e-01
-3.95786136e-01 -5.96849203e-01 -6.26204669e-01 -9.25261140e-01
-1.61139026e-01 6.26108050e-01 -1.87191829e-01 -5.09632647e-01
3.91591012e-01 5.00089943e-01 -6.36410296e-01 1.46855250e-01
8.43282938e-01 4.28098232e-01 -1.48545533e-01 6.89236045e-01
4.08960223e-01 5.31514525e-01 3.73543203e-02 -2.65270382e-01
7.40956485e-01 -4.80853289e-01 -9.97404099e-01 -4.51200195e-02
3.66021901e-01 -2.14089334e-01 -7.88734078e-01 1.30989516e+00
4.73552763e-01 6.65221829e-03 4.93535221e-01 -9.48586762e-01
-5.82055330e-01 5.65701723e-01 -4.92152095e-01 2.96099395e-01
5.27853251e-01 -2.17697456e-01 1.42882454e+00 -1.20459831e+00
8.80009711e-01 8.84975076e-01 6.21161044e-01 4.23614234e-01
-1.69993103e+00 -4.47273731e-01 -1.76913008e-01 2.94926763e-01
-1.39041889e+00 -7.02240825e-01 5.18357933e-01 -4.97226089e-01
1.04603052e+00 3.58157545e-01 7.38355875e-01 5.13444901e-01
2.04029605e-01 6.32929087e-01 1.42561042e+00 -7.36633182e-01
1.65176123e-01 -1.27798617e-01 2.77240604e-01 8.42823565e-01
4.83703136e-01 3.66109908e-01 -9.87911105e-01 -6.54468417e-01
6.06148601e-01 -4.35118407e-01 -6.76958919e-01 -8.56323779e-01
-1.63073432e+00 9.97313797e-01 2.46869922e-01 2.48133928e-01
-3.50140184e-01 3.82806808e-01 1.28579497e-01 5.69853127e-01
3.01569939e-01 4.46649164e-01 -2.44471699e-01 -1.86554432e-01
-9.77097571e-01 1.43753022e-01 1.16158438e+00 7.33924627e-01
1.24032795e+00 1.00032702e-01 -6.17309399e-02 1.67786568e-01
3.64569902e-01 9.21981037e-01 -1.25425383e-01 -9.18708265e-01
5.49074888e-01 -2.15636894e-01 4.71537262e-01 -6.67019665e-01
-4.30022538e-01 -3.99178743e-01 -8.50405216e-01 -1.42196724e-02
2.29033500e-01 -2.63037503e-01 -7.30154932e-01 1.74329269e+00
3.50318253e-01 1.58053815e-01 6.65698647e-02 1.19739842e+00
3.37873429e-01 7.19087362e-01 -7.27125287e-01 -7.54204333e-01
7.82908380e-01 -7.16616333e-01 -9.26357687e-01 -5.02846688e-02
7.85639226e-01 -7.25248575e-01 7.58010089e-01 4.27941263e-01
-1.06136096e+00 1.80282906e-01 -1.13557303e+00 -6.04095235e-02
2.17654496e-01 1.17539227e-01 1.25874329e+00 1.06587374e+00
-1.31349051e+00 8.69409323e-01 -9.31376159e-01 5.17592095e-02
-1.10012218e-01 7.73859501e-01 -5.42241633e-01 -9.60727632e-02
-9.75049794e-01 7.10730016e-01 3.57604504e-01 3.34946334e-01
-6.68370306e-01 -7.68268168e-01 -6.30200088e-01 -1.92984939e-01
2.43837997e-01 -9.94183362e-01 9.75220680e-01 -3.46673220e-01
-1.68869793e+00 9.32204425e-01 -3.17737073e-01 -3.38143587e-01
2.66997576e-01 -1.79737821e-01 -2.56738693e-01 2.55398721e-01
5.31099886e-02 -7.77716842e-03 9.45308745e-01 -8.65917563e-01
-2.43147952e-03 -2.76646376e-01 3.23236823e-01 3.39261323e-01
-1.40657842e-01 -4.61700745e-02 -2.19842464e-01 4.83612791e-02
7.91124046e-01 -1.41265213e+00 -5.37330747e-01 -5.12112021e-01
-4.71794307e-01 1.36493430e-01 4.64716405e-01 -3.27303380e-01
1.30026758e+00 -1.86804712e+00 6.44840777e-01 6.42156541e-01
4.31430399e-01 -7.99750462e-02 -9.34711173e-02 8.91951740e-01
-3.37708414e-01 -2.19990924e-01 -3.32817376e-01 1.36083290e-02
2.03741327e-01 1.38126895e-01 -4.68258470e-01 1.20933831e+00
-1.21327139e-01 7.19888747e-01 -1.08951092e+00 5.25798313e-02
-1.61390267e-02 1.96871206e-01 -1.19931614e+00 -3.76547694e-01
1.38513312e-01 5.66106498e-01 -5.45996547e-01 3.46127301e-01
9.00410593e-01 -4.99085188e-01 3.78689170e-01 -6.31529629e-01
-3.78605545e-01 6.29786074e-01 -1.56873095e+00 2.14753771e+00
-2.22846627e-01 5.43450356e-01 6.21086955e-01 -1.04929030e+00
2.33071133e-01 2.23837867e-01 8.37090373e-01 -3.93673688e-01
-1.27312198e-01 5.09120166e-01 3.06997001e-01 -1.49076328e-01
8.98626804e-01 -5.57684660e-01 -3.06245200e-02 8.05429697e-01
2.06431851e-01 -4.12552357e-01 3.77678096e-01 9.71123993e-01
1.11150014e+00 8.23851451e-02 2.03838542e-01 -1.02512646e+00
5.44845045e-01 -1.06096447e-01 5.16138315e-01 8.32269907e-01
3.00684720e-02 2.93693811e-01 4.95462388e-01 -1.57171547e-01
-1.05697381e+00 -1.02004433e+00 -4.41887140e-01 1.08331394e+00
9.71204937e-02 -1.10127056e+00 -2.24281669e-01 -3.76476228e-01
-1.39467031e-01 2.51715213e-01 -3.71286184e-01 -1.75368004e-02
-4.31502223e-01 -1.32066238e+00 2.39271283e-01 -9.32561755e-02
2.97327876e-01 -1.34624362e-01 1.82205841e-01 6.96365535e-02
9.61050838e-02 -1.04990494e+00 -5.70814371e-01 4.67841297e-01
-7.67182052e-01 -8.94710422e-01 -4.81024176e-01 -4.10938591e-01
9.62703943e-01 2.13708013e-01 1.05767298e+00 -4.68246967e-01
-4.25530076e-02 7.35672653e-01 -9.34896097e-02 2.40100503e-01
-1.45354122e-01 -2.37452701e-01 7.13418722e-01 5.85744716e-02
1.24495186e-01 -6.21113360e-01 -5.53330123e-01 -2.39882693e-02
-8.60144734e-01 1.16351284e-01 3.51626843e-01 1.28720248e+00
6.73984945e-01 -3.93530101e-01 -1.93876158e-02 -1.18571758e+00
6.70983136e-01 -1.38208404e-01 -8.21833134e-01 3.25752422e-02
-6.96894526e-01 6.59244180e-01 3.88124049e-01 1.62728261e-02
-7.96969771e-01 2.50122994e-01 -4.20621485e-02 -1.24581397e-01
8.80702615e-01 7.44333744e-01 1.57752633e-01 -8.85064602e-01
8.91702235e-01 2.30370268e-01 -4.64447588e-01 -2.20875144e-02
7.55129516e-01 3.69762391e-01 2.12703183e-01 -9.57823575e-01
1.38854957e+00 8.62065792e-01 6.01610184e-01 -8.67046714e-01
-9.92515266e-01 -6.76049352e-01 -6.67390525e-01 1.06870912e-01
4.44845498e-01 -1.07667696e+00 -7.18784451e-01 -5.99582158e-02
-7.89837480e-01 -2.11710230e-01 -4.52176690e-01 9.97554839e-01
-8.20099175e-01 7.59557486e-01 -8.72333646e-01 -8.95302653e-01
-3.49258989e-01 -1.05040956e+00 1.07141721e+00 -2.38751501e-01
3.54304612e-01 -9.72974300e-01 6.09440386e-01 2.09202945e-01
2.46591076e-01 -2.51678318e-01 5.23358762e-01 -2.44608521e-01
-7.86392450e-01 1.46029830e-01 -2.12928444e-01 1.84692353e-01
-2.67140508e-01 -3.50556314e-01 -7.31842697e-01 -7.55401969e-01
2.63208658e-01 -3.83123666e-01 1.19679391e+00 2.22443029e-01
5.25970757e-01 -2.57285684e-01 -1.77567005e-01 1.01052213e+00
1.54118085e+00 -5.07062018e-01 5.60855627e-01 8.03419854e-03
7.22976029e-01 1.99639350e-02 5.18327415e-01 6.10016406e-01
-1.40073359e-01 4.75941509e-01 2.95984428e-02 1.30167201e-01
3.24132621e-01 -1.35877416e-01 5.07694364e-01 1.44028330e+00
-3.75039011e-01 1.86557740e-01 -7.63245106e-01 2.22223356e-01
-2.09151244e+00 -1.01284838e+00 -4.78350043e-01 2.30982733e+00
7.37622976e-01 -2.03802079e-01 -1.66876763e-01 1.11431509e-01
4.90933597e-01 3.55263531e-01 -4.28836495e-01 -3.43362629e-01
-2.47091651e-01 6.74591064e-01 1.02223539e+00 1.04852140e+00
-1.03151321e+00 1.04271853e+00 7.56581116e+00 8.27502310e-01
-8.31076920e-01 4.98121440e-01 -1.98031619e-01 -1.63707525e-01
-8.51803899e-01 5.15339017e-01 -7.02640057e-01 -1.33187383e-01
7.15648830e-01 -3.15436840e-01 8.82295370e-01 4.81492966e-01
1.09877013e-01 9.18543190e-02 -1.10028327e+00 1.41590774e+00
-6.90791681e-02 -1.62546074e+00 -2.82011390e-01 4.69023526e-01
1.29655755e+00 4.55959827e-01 1.87819287e-01 1.31408617e-01
3.42549682e-01 -7.06091344e-01 3.41219962e-01 2.34924406e-01
1.10288966e+00 -7.16047049e-01 1.59466609e-01 3.76059674e-02
-1.01798546e+00 1.37087241e-01 -4.72808003e-01 -2.51615018e-01
4.75903273e-01 9.81478751e-01 -4.99357045e-01 6.47539020e-01
-3.88217866e-02 7.83816993e-01 1.62883967e-01 8.41611624e-01
-3.07120204e-01 5.10380268e-01 -7.92659402e-01 1.28399506e-01
4.16024655e-01 -1.15227246e+00 8.84252846e-01 1.09388304e+00
3.03205788e-01 4.60795850e-01 2.93184787e-01 7.89974153e-01
-1.04298748e-01 1.31118178e-01 -6.19210780e-01 -2.94337153e-01
3.63362543e-02 1.26037443e+00 -5.09247482e-01 -1.33470327e-01
-5.84902763e-01 1.09822524e+00 1.71559706e-01 5.98972023e-01
-4.33574587e-01 -1.66994736e-01 7.42345810e-01 -7.15899318e-02
3.00251782e-01 -7.02021301e-01 -3.56027000e-02 -2.00802636e+00
-4.97349575e-02 -7.23042965e-01 2.45451838e-01 -3.93113136e-01
-1.18144071e+00 2.14081004e-01 -2.14673221e-01 -8.98023069e-01
-3.20932001e-01 -8.52173865e-01 -3.58334094e-01 9.38871801e-01
-1.45098746e+00 -8.92732322e-01 3.86238158e-01 9.22795594e-01
-4.25777197e-01 6.25627860e-02 8.35156500e-01 4.96858537e-01
-3.75779271e-01 2.07672596e-01 5.97159088e-01 -2.14697570e-01
6.02175415e-01 -1.54004526e+00 3.08585055e-02 1.23388088e+00
4.66318429e-01 1.04974818e+00 7.39947915e-01 -6.49756849e-01
-2.47060204e+00 -5.52071929e-01 6.99310482e-01 -2.70296603e-01
1.28039300e+00 -4.38278764e-01 -1.53547332e-01 8.89270544e-01
1.38713390e-01 1.89392135e-01 8.56299341e-01 6.02891684e-01
-4.56178874e-01 1.60985187e-01 -9.29966688e-01 2.72127658e-01
1.17050791e+00 -1.26756334e+00 -2.41379380e-01 1.14843786e+00
2.12652370e-01 -5.52734196e-01 -7.72453368e-01 3.34961623e-01
2.66088098e-01 -8.27661037e-01 9.96807516e-01 -8.35325658e-01
-1.79390192e-01 -5.10202467e-01 -4.63933319e-01 -1.24659193e+00
-7.69241631e-01 -1.57167089e+00 -2.71266997e-01 2.77486920e-01
5.30542076e-01 -7.17838407e-01 9.83872771e-01 5.67642510e-01
-4.02555496e-01 -5.03363311e-01 -1.05367720e+00 -7.52965331e-01
1.91719592e-01 -3.11158180e-01 -2.19651699e-01 1.11459517e+00
6.61195338e-01 7.17039526e-01 -7.01860547e-01 3.71503532e-01
8.05053711e-01 3.93119007e-01 5.33193111e-01 -8.12632799e-01
-7.86782324e-01 -9.98530686e-02 -4.35093969e-01 -1.43064189e+00
1.62635043e-01 -1.37421989e+00 -1.70421660e-01 -1.27104676e+00
4.58505839e-01 -4.02620196e-01 -1.80202901e-01 1.26860112e-01
-3.00661344e-02 4.10904497e-01 1.78781718e-01 3.58551741e-01
-8.23561668e-01 6.71876073e-01 1.28455281e+00 -2.63148937e-02
-2.15254083e-01 -5.77632189e-02 -5.14627814e-01 4.09813881e-01
3.82037222e-01 -3.45356166e-01 -1.51718393e-01 -4.00240481e-01
1.28307521e+00 1.87075153e-01 -6.44067153e-02 -7.95469522e-01
4.60660964e-01 -5.86244836e-02 -1.44233331e-01 -5.03210545e-01
6.79011762e-01 -3.94504547e-01 2.10349187e-01 5.61523139e-01
1.53214522e-02 -1.91009969e-01 -2.60436118e-01 5.80740094e-01
-7.98108205e-02 -4.35671121e-01 5.33051312e-01 -1.09642066e-01
-4.18919891e-01 6.18124902e-01 -2.40400359e-01 1.48274079e-01
4.61961538e-01 2.58016527e-01 -2.74395317e-01 -5.37815988e-01
-1.02664518e+00 -7.14427829e-02 5.59382677e-01 -5.25950670e-01
3.89635593e-01 -1.51016617e+00 -9.12807107e-01 1.41357556e-01
-8.92456099e-02 -5.42582750e-01 1.32643208e-01 1.33596659e+00
-6.41386330e-01 6.31651819e-01 2.31516004e-01 -6.23800755e-01
-9.74120975e-01 3.46104473e-01 2.37417638e-01 -3.77783269e-01
-4.61209178e-01 1.00365758e+00 1.05066620e-01 -5.35710514e-01
-3.44082475e-01 -1.31541058e-01 5.09123504e-01 -3.69091965e-02
4.37675297e-01 2.90528536e-01 1.70110896e-01 -7.66727567e-01
-3.70068759e-01 5.31695127e-01 -9.89039149e-03 -6.12195909e-01
1.39127636e+00 -3.47594582e-02 -6.35190129e-01 1.50746778e-01
1.35285437e+00 7.14176357e-01 -8.20741236e-01 -5.15993774e-01
-2.19745114e-01 -3.69968385e-01 3.72189760e-01 -2.05391899e-01
-7.99011528e-01 6.48077548e-01 3.40871751e-01 1.31566361e-01
6.98513567e-01 -2.33815789e-01 6.21221066e-01 1.21555305e+00
8.49974751e-01 -9.32902217e-01 -2.50755370e-01 1.00373065e+00
4.55185592e-01 -1.01653397e+00 4.78589833e-01 -5.01298785e-01
-1.51430652e-01 1.05945575e+00 -1.07591420e-01 -5.06669998e-01
7.94313312e-01 2.53561139e-01 -5.10493219e-01 -2.61735767e-01
-4.51270670e-01 -3.65572691e-01 3.73916298e-01 3.62633854e-01
4.17128801e-01 3.00223082e-01 -7.17950702e-01 -7.02666352e-03
-4.58253354e-01 -2.02539086e-01 7.92769611e-01 9.89255011e-01
-3.78860921e-01 -1.51507497e+00 -3.92226696e-01 5.11349916e-01
-4.22377467e-01 -5.65211654e-01 -2.65359759e-01 2.91571587e-01
-2.73828447e-01 7.69472361e-01 -4.73781765e-01 -3.41128320e-01
-1.46765620e-01 -2.42132992e-02 1.33121693e+00 -7.79808402e-01
-3.60752016e-01 2.17537567e-01 3.88872951e-01 -8.17013979e-01
-7.62869775e-01 -7.07798541e-01 -1.01827812e+00 -5.66428185e-01
-7.10986614e-01 7.38570690e-01 6.80965364e-01 8.78813744e-01
2.08701149e-01 1.00571588e-01 6.66808784e-01 -9.34310734e-01
-8.49608064e-01 -7.41520047e-01 -8.71873736e-01 1.13958173e-01
5.07028341e-01 -3.26951593e-01 -5.24205267e-01 -4.30824548e-01] | [6.472716808319092, 4.655330181121826] |
eeb45128-0897-412f-b74b-5460fdb8f071 | incorporating-bert-into-neural-machine-1 | 2002.06823 | null | https://arxiv.org/abs/2002.06823v1 | https://arxiv.org/pdf/2002.06823v1.pdf | Incorporating BERT into Neural Machine Translation | The recently proposed BERT has shown great power on a variety of natural language understanding tasks, such as text classification, reading comprehension, etc. However, how to effectively apply BERT to neural machine translation (NMT) lacks enough exploration. While BERT is more commonly used as fine-tuning instead of contextual embedding for downstream language understanding tasks, in NMT, our preliminary exploration of using BERT as contextual embedding is better than using for fine-tuning. This motivates us to think how to better leverage BERT for NMT along this direction. We propose a new algorithm named BERT-fused model, in which we first use BERT to extract representations for an input sequence, and then the representations are fused with each layer of the encoder and decoder of the NMT model through attention mechanisms. We conduct experiments on supervised (including sentence-level and document-level translations), semi-supervised and unsupervised machine translation, and achieve state-of-the-art results on seven benchmark datasets. Our code is available at \url{https://github.com/bert-nmt/bert-nmt}. | ['Tie-Yan Liu', 'Di He', 'Wengang Zhou', 'Jinhua Zhu', 'Houqiang Li', 'Tao Qin', 'Lijun Wu', 'Yingce Xia'] | 2020-02-17 | null | https://openreview.net/forum?id=Hyl7ygStwB | https://openreview.net/pdf?id=Hyl7ygStwB | iclr-2020-1 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 5.44751704e-01 3.23198974e-01 -4.27662015e-01 -6.26680076e-01
-1.02974367e+00 -4.50217634e-01 7.48853207e-01 -6.07478246e-03
-4.00909036e-01 7.98720956e-01 5.05663097e-01 -9.98245060e-01
2.51719505e-01 -6.38789833e-01 -9.66699064e-01 -2.58690566e-01
4.67316210e-01 5.98347425e-01 -1.99151352e-01 -3.62717837e-01
1.02484994e-01 -1.58322722e-01 -7.27835178e-01 5.80574274e-01
1.06337166e+00 6.19749367e-01 2.98882306e-01 5.12806892e-01
-3.95070791e-01 5.71272731e-01 -4.11278069e-01 -5.76745152e-01
8.22452456e-02 -7.96242118e-01 -1.22296417e+00 -3.62249762e-01
-6.30027056e-02 -2.20150694e-01 -3.22346002e-01 9.35592055e-01
4.31793272e-01 -8.78354982e-02 7.52267241e-01 -9.98160481e-01
-1.19540989e+00 1.22749054e+00 -5.57694256e-01 3.52407366e-01
2.29138017e-01 1.79405451e-01 1.28577137e+00 -9.77812529e-01
4.25158679e-01 1.53646529e+00 2.80976862e-01 8.07498217e-01
-1.19804657e+00 -7.26163864e-01 2.86850750e-01 2.21893758e-01
-1.12192678e+00 -5.58765173e-01 4.90006715e-01 -1.80129841e-01
1.19090950e+00 1.16269402e-01 2.23562017e-01 1.32974863e+00
3.88172746e-01 1.31735396e+00 1.21649575e+00 -6.57213569e-01
-1.38410076e-01 2.99999025e-02 2.99590528e-01 5.67291796e-01
-4.20472361e-02 -1.89572275e-02 -5.42099416e-01 1.38099536e-01
5.80016553e-01 -1.63347051e-01 -3.73516560e-01 1.95374236e-01
-1.58152831e+00 9.23070073e-01 4.92272109e-01 3.92884195e-01
-2.81242102e-01 3.55609268e-01 4.41188544e-01 7.28811800e-01
7.33267844e-01 4.53229457e-01 -8.06875348e-01 -4.19366419e-01
-6.45165086e-01 -1.69695660e-01 6.61641181e-01 9.85353112e-01
7.44784474e-01 -1.36708856e-01 -2.65553623e-01 1.01605594e+00
2.85584241e-01 4.83677566e-01 7.90698647e-01 -4.41146910e-01
9.22055006e-01 6.59122050e-01 -1.71623871e-01 -3.77206087e-01
-1.29107520e-01 -4.22079414e-01 -9.95877862e-01 -3.73725712e-01
1.47554830e-01 -2.53306270e-01 -9.67386007e-01 2.04415178e+00
-5.13847135e-02 -1.54436538e-02 2.16243297e-01 9.56623375e-01
7.24311769e-01 1.03456700e+00 -1.98708981e-01 -2.27909073e-01
1.19389760e+00 -1.46882081e+00 -7.66099870e-01 -4.93421972e-01
9.62080598e-01 -7.12614596e-01 1.34383583e+00 8.73987228e-02
-1.08483040e+00 -5.71683645e-01 -9.17631030e-01 -4.40626353e-01
-4.27175641e-01 2.17314899e-01 5.16658783e-01 2.08502337e-01
-1.10627091e+00 5.10287404e-01 -9.10045445e-01 -4.44767356e-01
3.66895884e-01 4.30338979e-01 -2.71119863e-01 -1.19244128e-01
-1.55986106e+00 1.06937814e+00 1.97431162e-01 2.18431398e-01
-7.72866011e-01 -2.13762343e-01 -8.24253082e-01 7.48992562e-02
2.82369614e-01 -1.05704927e+00 1.56577098e+00 -1.15615821e+00
-1.72405982e+00 7.44810104e-01 -6.25152886e-01 -6.50079727e-01
2.93222219e-01 -6.10791683e-01 -1.97624683e-01 -1.58591658e-01
2.52570331e-01 8.70718598e-01 6.75272346e-01 -8.15266907e-01
-1.33036777e-01 -2.49192566e-01 2.67890334e-01 2.37566337e-01
-2.92354256e-01 7.69246370e-02 -3.90559554e-01 -6.14478111e-01
-7.09932148e-02 -9.54717219e-01 -1.99841514e-01 -2.51990110e-01
-5.12906075e-01 -5.44637084e-01 4.37041849e-01 -8.37714195e-01
1.32359374e+00 -1.97486997e+00 5.36600530e-01 -3.93182278e-01
-3.24704386e-02 2.56271511e-01 -3.97628933e-01 8.37664485e-01
-1.06507421e-01 4.90292400e-01 -4.24250096e-01 -4.03326809e-01
7.93634653e-02 2.07508564e-01 -5.09366035e-01 1.14347138e-01
4.08446461e-01 1.53129041e+00 -8.24033737e-01 -1.67949021e-01
-7.11249486e-02 2.95253813e-01 -4.07595575e-01 1.58417627e-01
-5.73384047e-01 5.44580638e-01 -7.83067226e-01 4.61719036e-01
3.47703248e-01 -5.70450306e-01 -7.00478330e-02 2.25197002e-01
6.91593364e-02 9.33773994e-01 -4.05664742e-01 1.97207081e+00
-6.73969567e-01 7.31157064e-01 -9.21823457e-02 -1.07106531e+00
9.03264344e-01 4.91870522e-01 -2.50447243e-02 -6.95084095e-01
1.89809352e-01 4.26136941e-01 2.52906084e-01 -3.79680485e-01
3.99282217e-01 -1.11117907e-01 -3.78254876e-02 9.18334126e-01
6.97098821e-02 1.06368884e-01 -1.74397435e-02 3.16609293e-01
9.85957921e-01 2.72662014e-01 3.88396740e-01 -3.65432531e-01
4.32473302e-01 2.53494605e-02 3.30666900e-01 4.84178782e-01
1.33252248e-01 5.32890677e-01 6.56576455e-01 -1.21986873e-01
-9.24696505e-01 -7.94477105e-01 1.17026731e-01 1.20803845e+00
1.80353019e-02 -5.74596643e-01 -8.62306058e-01 -8.70818019e-01
-1.79657519e-01 9.19541955e-01 -6.47255838e-01 -3.88688087e-01
-7.13991642e-01 -5.98809242e-01 5.98797023e-01 6.18187487e-01
6.13577366e-01 -1.02577162e+00 -1.41891807e-01 2.73834318e-01
-5.06722331e-01 -7.79044330e-01 -7.26983488e-01 2.91325659e-01
-1.01747906e+00 -5.26518762e-01 -6.00179434e-01 -8.10537100e-01
5.89419782e-01 3.16345006e-01 1.13502038e+00 1.26602635e-01
2.29220614e-01 7.27414340e-02 -7.08021164e-01 -4.23619866e-01
-5.62297642e-01 6.04372919e-01 -1.14118822e-01 -9.37312171e-02
4.28828657e-01 -5.88238239e-01 -5.30684233e-01 4.44629550e-01
-8.82017136e-01 6.41959310e-01 8.35384250e-01 1.23311520e+00
3.02057981e-01 -6.54803157e-01 8.54199469e-01 -9.72248614e-01
9.14701223e-01 -5.12724638e-01 -1.61655053e-01 3.92335922e-01
-5.24271011e-01 3.27909827e-01 7.86304116e-01 -3.84846061e-01
-7.74346232e-01 -4.76725608e-01 -4.21240985e-01 -2.11215466e-01
1.04710296e-01 7.92431772e-01 -1.64730951e-01 4.80903625e-01
5.66546142e-01 3.60159934e-01 8.65416415e-03 -6.89080656e-01
5.73612750e-01 1.03969824e+00 1.02401301e-01 -7.81754136e-01
7.02011883e-01 -3.64285782e-02 -5.60596287e-01 -4.04551089e-01
-9.88322198e-01 -2.20042154e-01 -7.48733819e-01 3.06836694e-01
8.95552039e-01 -7.93346584e-01 -1.98691428e-01 5.71064577e-02
-1.42565203e+00 -5.49836397e-01 -4.16562669e-02 4.43040639e-01
-5.63516021e-01 1.94256246e-01 -7.43047535e-01 -5.02453268e-01
-6.65430367e-01 -1.38116515e+00 1.36804831e+00 8.23909976e-03
-3.03790033e-01 -1.03193033e+00 -1.26553595e-01 6.49243832e-01
5.18027186e-01 -3.01228493e-01 1.30072296e+00 -7.14043558e-01
-6.00713432e-01 9.93544832e-02 -2.67599583e-01 4.76386547e-01
1.83487415e-01 -4.56964791e-01 -9.80446517e-01 -3.84619832e-01
-1.80666393e-03 -4.49501604e-01 1.12470841e+00 1.75730392e-01
1.12973070e+00 -4.53491271e-01 -3.50322068e-01 5.82091272e-01
1.18884051e+00 7.36235157e-02 5.59663296e-01 3.32133651e-01
6.35222793e-01 4.86614406e-01 4.72226202e-01 -6.37316406e-02
4.01780337e-01 6.65010810e-01 2.01254040e-01 -3.06211840e-02
-1.21222548e-01 -4.79943365e-01 7.59023368e-01 1.32724011e+00
4.74067852e-02 -5.92902184e-01 -9.59411442e-01 4.03637558e-01
-2.02395296e+00 -5.38948953e-01 1.69054359e-01 1.84582424e+00
1.04054129e+00 2.04552472e-01 -2.60260224e-01 -1.65247485e-01
6.22600257e-01 1.73373893e-01 -6.48184299e-01 -9.10110831e-01
-5.33212498e-02 2.63548255e-01 2.52112150e-01 5.71349740e-01
-8.32894921e-01 1.36069930e+00 5.39060450e+00 8.76975298e-01
-1.13489735e+00 4.54186440e-01 8.51323724e-01 2.21037179e-01
-5.36551714e-01 1.45276725e-01 -6.13569081e-01 2.84885585e-01
1.13425851e+00 -2.96048045e-01 6.20636404e-01 4.91355747e-01
2.48652950e-01 3.30873102e-01 -1.43609726e+00 7.05670834e-01
-1.03004031e-01 -1.30885935e+00 2.95199037e-01 2.53370721e-02
6.64883554e-01 3.29715610e-01 1.12300171e-02 7.12568343e-01
2.50591248e-01 -1.28539169e+00 5.09714842e-01 2.38185614e-01
9.56691921e-01 -4.22198951e-01 7.49627829e-01 7.08018482e-01
-8.36840034e-01 1.43528059e-01 -4.17394817e-01 -2.11495727e-01
2.04852700e-01 3.78858179e-01 -1.00714326e+00 7.94792414e-01
2.13348180e-01 9.88426805e-01 -4.56956893e-01 4.42501009e-01
-6.25850737e-01 1.06069493e+00 -6.35032877e-02 -2.82816678e-01
5.03624558e-01 -2.33839273e-01 4.79678810e-01 1.33534563e+00
1.95736289e-01 -5.91339543e-03 6.64065778e-02 8.63214254e-01
-5.18409014e-01 2.85631567e-01 -6.37682319e-01 -4.79314834e-01
2.06143498e-01 8.50520074e-01 -3.43696266e-01 -4.99973625e-01
-5.16011477e-01 1.29520237e+00 4.92779106e-01 5.95488012e-01
-8.43860567e-01 -1.91914111e-01 6.55476093e-01 -5.62943853e-02
5.46262190e-02 -4.47351545e-01 -4.35202718e-01 -1.47678804e+00
1.73186168e-01 -1.16920257e+00 1.34481549e-01 -8.50139618e-01
-1.16378915e+00 7.07342803e-01 -5.06081022e-02 -1.15112233e+00
-4.07255590e-01 -6.47973061e-01 -6.14681780e-01 1.09865093e+00
-1.46863985e+00 -1.19062591e+00 2.42410243e-01 3.33054662e-01
1.17649698e+00 -9.91372094e-02 9.41966236e-01 5.50507866e-02
-6.30617380e-01 7.02172458e-01 3.27695847e-01 3.25325102e-01
7.52733111e-01 -1.12431240e+00 9.18112040e-01 8.02920878e-01
3.79599869e-01 9.27643239e-01 3.46775979e-01 -4.65334207e-01
-1.62606227e+00 -1.12479377e+00 1.15783584e+00 -4.93225485e-01
7.60552824e-01 -6.82189107e-01 -9.12839472e-01 9.50951219e-01
7.57302940e-01 -5.31077325e-01 5.64982653e-01 1.67562455e-01
-3.86919886e-01 1.11511789e-01 -5.32290637e-01 8.31135809e-01
1.19063079e+00 -6.21344745e-01 -8.51483107e-01 5.00946403e-01
1.39038885e+00 -2.56644636e-01 -6.64281547e-01 3.37265819e-01
2.75433749e-01 -4.44055796e-01 6.19742155e-01 -8.09086084e-01
7.80661166e-01 -7.87374005e-02 -1.88426092e-01 -1.68584752e+00
-2.99398750e-01 -7.95352578e-01 5.20425988e-03 1.11777115e+00
9.05940294e-01 -9.52046335e-01 3.65341425e-01 1.41399994e-01
-4.28754687e-01 -1.23892546e+00 -1.02222526e+00 -6.58733606e-01
6.13417804e-01 -3.81291568e-01 7.51221061e-01 8.86463761e-01
1.48562774e-01 1.02295625e+00 -3.90569061e-01 -1.54533193e-01
1.04619905e-01 2.72784740e-01 8.18890512e-01 -7.70305157e-01
-5.61781347e-01 -6.56883121e-01 -2.34014960e-03 -1.86886883e+00
2.92839289e-01 -1.32193005e+00 -1.29589187e-02 -1.94140756e+00
3.51786107e-01 -7.45112523e-02 -3.03251952e-01 6.65693879e-01
-3.17489564e-01 -6.07589819e-03 9.50902402e-02 2.97221273e-01
-4.10287142e-01 9.05042708e-01 1.58842456e+00 -4.06903505e-01
1.19776547e-01 -4.12053578e-02 -9.92318332e-01 2.90973336e-01
1.09230852e+00 -4.10046577e-01 -4.23404872e-01 -1.11676240e+00
2.06459999e-01 2.27970585e-01 -6.83830772e-03 -4.84667331e-01
-1.00912757e-01 -1.32517278e-01 1.64245903e-01 -4.34557289e-01
3.60517144e-01 -3.14656705e-01 -3.58945310e-01 3.92389953e-01
-8.00181210e-01 3.55392843e-01 1.50216877e-01 4.20490414e-01
-4.36915159e-01 -1.88358903e-01 4.56701785e-01 -1.99635923e-01
-4.11792666e-01 3.25411648e-01 -2.99163967e-01 7.54102245e-02
4.29724187e-01 3.57436612e-02 -4.58636761e-01 -5.93157411e-01
-4.87969607e-01 5.69788277e-01 2.99943596e-01 8.25993359e-01
5.76453447e-01 -1.11737525e+00 -9.85148489e-01 1.92627773e-01
1.11309372e-01 2.80454382e-02 -1.70339540e-01 1.08382595e+00
-1.02659695e-01 7.32365072e-01 1.62387475e-01 -6.09268367e-01
-9.44934666e-01 5.15651524e-01 1.77250370e-01 -4.00831759e-01
-5.18401027e-01 8.92773628e-01 5.80919206e-01 -7.36923277e-01
-3.77201587e-02 -5.55052698e-01 4.25911322e-02 -3.63910466e-01
3.75242233e-01 -9.52373892e-02 -2.55937269e-03 -4.02076393e-01
-2.34913692e-01 3.92923236e-01 -3.45623702e-01 -3.47883493e-01
1.16727531e+00 -3.43334585e-01 -2.86471218e-01 6.91053033e-01
1.35233247e+00 -4.14299756e-01 -8.28633368e-01 -5.30584574e-01
5.66859357e-02 -1.59472570e-01 -2.59443074e-01 -9.71543491e-01
-6.61012530e-01 1.45779681e+00 1.57060966e-01 -1.53434828e-01
1.06650507e+00 8.01972896e-02 1.05289841e+00 6.42587900e-01
4.73914385e-01 -7.02211916e-01 -9.17508546e-03 8.70104313e-01
1.10017085e+00 -1.31809890e+00 -3.79979044e-01 -1.95883483e-01
-6.54408813e-01 1.05586684e+00 6.13791227e-01 9.07125920e-02
4.14120555e-01 -1.21624246e-02 1.39049754e-01 2.02398658e-01
-1.30992925e+00 -3.20191979e-02 2.19709978e-01 1.59537360e-01
9.23906744e-01 2.29140654e-01 -5.62572420e-01 5.34569025e-01
-3.89677495e-01 -4.32615057e-02 3.26692492e-01 7.99819112e-01
-4.21047240e-01 -1.49258578e+00 -5.94091900e-02 5.34630120e-01
-4.21856552e-01 -4.77672994e-01 -7.48558760e-01 5.42473912e-01
-1.98466599e-01 9.69841063e-01 -3.63660961e-01 -5.57113588e-01
5.37181795e-02 4.11571115e-01 2.97457993e-01 -9.50044394e-01
-6.20681584e-01 1.41473085e-01 3.32820505e-01 -3.79618526e-01
-2.29609087e-01 -5.39333642e-01 -1.18456542e+00 -1.13791622e-01
-3.74112785e-01 4.16781425e-01 6.37736738e-01 1.01645219e+00
3.63595277e-01 4.93860960e-01 5.93345404e-01 -4.18366283e-01
-8.37464571e-01 -1.51897275e+00 1.25768304e-01 1.81996331e-01
4.23744112e-01 -3.90794903e-01 -1.58744648e-01 -4.75447252e-03] | [11.64008617401123, 9.924751281738281] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.