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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2ea4023a-6737-486c-b042-3e2504f103eb | gaussian-membership-inference-privacy | 2306.07273 | null | https://arxiv.org/abs/2306.07273v1 | https://arxiv.org/pdf/2306.07273v1.pdf | Gaussian Membership Inference Privacy | We propose a new privacy notion called $f$-Membership Inference Privacy ($f$-MIP), which explicitly considers the capabilities of realistic adversaries under the membership inference attack threat model. By doing so $f$-MIP offers interpretable privacy guarantees and improved utility (e.g., better classification accuracy). Our novel theoretical analysis of likelihood ratio-based membership inference attacks on noisy stochastic gradient descent (SGD) results in a parametric family of $f$-MIP guarantees that we refer to as $\mu$-Gaussian Membership Inference Privacy ($\mu$-GMIP). Our analysis additionally yields an analytical membership inference attack that offers distinct advantages over previous approaches. First, unlike existing methods, our attack does not require training hundreds of shadow models to approximate the likelihood ratio. Second, our analytical attack enables straightforward auditing of our privacy notion $f$-MIP. Finally, our analysis emphasizes the importance of various factors, such as hyperparameters (e.g., batch size, number of model parameters) and data specific characteristics in controlling an attacker's success in reliably inferring a given point's membership to the training set. We demonstrate the effectiveness of our method on models trained across vision and tabular datasets. | ['Gjergji Kasneci', 'Martin Pawelczyk', 'Tobias Leemann'] | 2023-06-12 | null | null | null | null | ['inference-attack', 'membership-inference-attack'] | ['adversarial', 'computer-vision'] | [ 2.21756268e-02 4.87038791e-02 -7.27059543e-02 -5.18064618e-01
-1.21652317e+00 -1.11206102e+00 2.67620951e-01 3.04768067e-02
-5.71835995e-01 6.73552752e-01 -3.85151953e-01 -9.99405324e-01
-1.51434526e-01 -7.81651318e-01 -1.08227921e+00 -6.87305987e-01
-5.07781148e-01 5.10932207e-02 -1.88261852e-01 1.41816929e-01
2.64766455e-01 6.29064977e-01 -8.42827201e-01 -3.11011225e-01
7.02557683e-01 1.33815992e+00 -4.87747878e-01 6.69803679e-01
4.16908860e-01 3.19393903e-01 -9.00435150e-01 -1.00513697e+00
8.72942090e-01 4.08391319e-02 -5.35176694e-01 -4.17139858e-01
6.19883358e-01 -8.30018103e-01 -4.01509672e-01 1.40442324e+00
1.80949688e-01 1.21330425e-01 4.44738895e-01 -1.45372784e+00
-3.28795671e-01 4.88071978e-01 -5.20112634e-01 8.72303024e-02
1.41547292e-01 3.23520273e-01 9.83181536e-01 -2.80535519e-01
4.43861097e-01 1.06199181e+00 7.98802137e-01 5.15457630e-01
-1.53010154e+00 -1.07986140e+00 5.96629828e-02 -2.54677713e-01
-1.71335495e+00 -3.89050096e-01 4.34907317e-01 -2.87750900e-01
4.51338351e-01 6.45036459e-01 4.94353613e-03 8.06924880e-01
-8.79711807e-02 7.05700636e-01 1.21929669e+00 -9.39536467e-02
4.19823050e-01 4.78193253e-01 1.99724451e-01 7.34109819e-01
5.68901241e-01 3.61100137e-01 -3.01934540e-01 -1.06878185e+00
6.70002520e-01 -2.61317641e-01 -3.96475732e-01 -5.44227779e-01
-4.44879174e-01 9.59400356e-01 1.54520169e-01 -5.76536894e-01
1.49504930e-01 5.71831763e-01 4.47939247e-01 3.83019745e-01
3.13145369e-01 3.83290410e-01 -4.76071179e-01 4.41376902e-02
-6.15622282e-01 4.82519746e-01 1.03638494e+00 1.07102156e+00
8.47256541e-01 -2.47364715e-02 1.65638804e-01 2.22280532e-01
2.79704332e-01 7.09287763e-01 -1.70874577e-02 -1.14925981e+00
6.55262709e-01 -1.18905343e-01 4.88293558e-01 -1.20275772e+00
2.06326202e-01 -3.39778513e-01 -4.61058497e-01 3.79014015e-01
5.75065672e-01 -5.69428086e-01 -3.42527270e-01 2.25594091e+00
2.65163273e-01 6.35585636e-02 2.40282983e-01 6.13038838e-01
1.30085081e-01 3.63933593e-01 3.28885689e-02 -8.93687531e-02
1.21528268e+00 -1.48455068e-01 -3.02961677e-01 3.57283354e-02
5.21768630e-01 -2.23668739e-01 1.08822978e+00 3.88807684e-01
-1.08793569e+00 1.54935822e-01 -1.10916567e+00 1.02144495e-01
-3.89135480e-01 -2.19483048e-01 1.15545321e+00 1.46581256e+00
-9.86061811e-01 3.66481632e-01 -9.19199824e-01 1.77854374e-01
8.33612025e-01 7.84337103e-01 -4.59744066e-01 2.18263403e-01
-1.19717205e+00 4.01817739e-01 1.35262221e-01 -1.55482337e-01
-1.17096138e+00 -7.04247057e-01 -8.25404227e-01 1.04656577e-01
3.77465546e-01 -5.35427332e-01 1.20160949e+00 -4.17607635e-01
-1.13189864e+00 8.39659572e-01 -1.72275230e-01 -1.05726123e+00
7.99293578e-01 2.06480958e-02 -7.30143860e-02 2.70310938e-01
-9.57949460e-02 1.97092474e-01 9.32157874e-01 -1.41126800e+00
-6.50806725e-01 -7.29964495e-01 5.80436587e-01 1.60940215e-01
-4.02697057e-01 4.12044190e-02 -2.53714740e-01 -5.09151161e-01
-5.05955070e-02 -8.31957459e-01 -4.94694501e-01 3.32595944e-01
-6.33492768e-01 2.62408823e-01 9.26174402e-01 -5.05717695e-01
1.13725889e+00 -2.43673968e+00 -5.58244467e-01 8.59200656e-01
3.13352764e-01 2.74659842e-01 4.14064854e-01 1.43749461e-01
4.46807504e-01 5.30147433e-01 -3.43447596e-01 -7.02105880e-01
3.10359865e-01 1.94025263e-01 -7.70687580e-01 8.95723581e-01
-4.10566270e-01 6.12190783e-01 -6.65454388e-01 -2.37735271e-01
2.14111116e-02 2.69602984e-01 -6.34848475e-01 9.96646062e-02
-3.15216362e-01 1.91735566e-01 -6.03256583e-01 6.10059500e-01
1.08361197e+00 -1.74497530e-01 2.30849385e-01 5.18264771e-02
2.70427674e-01 1.29468784e-01 -1.16147065e+00 1.25027657e+00
-2.61103243e-01 3.38274360e-01 5.47061086e-01 -6.53588593e-01
6.53789878e-01 7.21743479e-02 4.21999097e-02 7.05804899e-02
2.51705110e-01 1.40277520e-01 -3.87135893e-01 4.27982323e-02
5.70276797e-01 -1.11069193e-03 -3.57280195e-01 7.31393874e-01
-2.52493680e-01 -3.57693247e-02 -4.34399664e-01 3.44579071e-01
9.56177413e-01 -2.80203044e-01 1.04877487e-01 -4.56898421e-01
3.11546296e-01 -3.50985110e-01 6.68888628e-01 1.41808283e+00
-5.32240152e-01 3.37173611e-01 6.06777012e-01 -1.82457566e-01
-6.65895164e-01 -1.29889727e+00 -2.70731777e-01 8.52934420e-01
3.55437994e-01 -3.44173014e-01 -9.01916385e-01 -9.48041379e-01
5.70679188e-01 8.01071048e-01 -5.64975083e-01 -3.61490473e-02
-2.66496301e-01 -9.50110614e-01 1.33223093e+00 1.68867841e-01
6.12094820e-01 -1.89786226e-01 -3.16086560e-01 -3.26756150e-01
5.06382361e-02 -1.00789225e+00 -7.31047988e-01 2.81828288e-02
-7.76871383e-01 -1.01575661e+00 1.89809546e-01 -2.24961877e-01
9.53853667e-01 2.27517948e-01 7.15023279e-01 -1.55469447e-01
-2.14852512e-01 7.42422760e-01 4.55164053e-02 -6.19127452e-01
-3.99915546e-01 -1.41652480e-01 1.73289388e-01 1.63589269e-01
3.84666353e-01 -6.90324783e-01 -7.10847974e-01 1.42921478e-01
-9.61254835e-01 -6.95293784e-01 1.37395233e-01 5.98397195e-01
6.00674152e-01 1.71222419e-01 2.60635436e-01 -1.36496067e+00
9.10020232e-01 -5.09703994e-01 -1.26516438e+00 2.56175101e-01
-8.02726090e-01 -6.45494089e-02 8.65317464e-01 -3.49399239e-01
-8.34241509e-01 -1.04673415e-01 -6.48515522e-02 -6.48538530e-01
1.21778153e-01 1.64909989e-01 -3.30712855e-01 -6.68265045e-01
7.36643136e-01 3.64722788e-01 2.89841294e-01 -3.03027213e-01
5.41143715e-01 5.63002944e-01 8.39078903e-01 -1.02387094e+00
9.38674808e-01 8.68182242e-01 2.74119973e-01 -4.57709014e-01
-3.62858087e-01 -7.20771551e-02 1.89459510e-02 3.84228110e-01
2.76421964e-01 -7.41663635e-01 -1.53100634e+00 4.99756813e-01
-7.94718623e-01 -1.64054751e-01 -2.75510579e-01 2.19532743e-01
-6.51888013e-01 9.02033508e-01 -6.38814449e-01 -1.28978515e+00
-5.04526317e-01 -1.28204203e+00 6.95881248e-01 -2.75291651e-02
-2.19462905e-02 -1.09758627e+00 -4.21344697e-01 3.55956137e-01
2.78816223e-01 5.36769807e-01 7.13070095e-01 -8.44527125e-01
-7.92071939e-01 -6.41285658e-01 -1.30965710e-01 4.92170393e-01
-4.30518128e-02 -2.14640558e-01 -1.11361313e+00 -6.87452674e-01
3.97262692e-01 -3.82820278e-01 5.43305635e-01 3.30675513e-01
1.72289634e+00 -1.12890089e+00 -2.06716523e-01 1.18966508e+00
1.36868775e+00 6.60659745e-02 5.37326455e-01 2.22883955e-01
3.62890184e-01 2.17048794e-01 6.21456087e-01 8.26906919e-01
2.64126688e-01 2.85479277e-01 5.96329331e-01 1.66432977e-01
7.40972877e-01 -4.28687364e-01 2.42408857e-01 -3.52428168e-01
3.30308855e-01 -9.48262289e-02 -3.55375439e-01 1.88021883e-01
-1.55359375e+00 -8.58190835e-01 5.07492781e-01 2.83146167e+00
1.05132127e+00 1.85679466e-01 1.04116850e-01 -2.64016837e-01
4.67691451e-01 1.63014606e-01 -7.75852978e-01 -6.38511240e-01
-7.11485147e-02 2.85984665e-01 1.27375484e+00 8.66171777e-01
-1.21415424e+00 7.68166482e-01 6.64208412e+00 1.14550662e+00
-7.60767043e-01 1.11199878e-01 1.09016633e+00 -2.57411927e-01
-5.72248995e-01 2.12123692e-01 -9.35109198e-01 5.94019532e-01
7.59422243e-01 -5.23763359e-01 6.93372965e-01 1.11414647e+00
-1.66789174e-01 2.12885384e-02 -1.07453477e+00 9.30506766e-01
-5.63050434e-02 -1.38050997e+00 -1.70609877e-01 5.34369171e-01
3.16081434e-01 -1.75586477e-01 7.30153024e-01 -5.22906110e-02
1.14270198e+00 -9.39713776e-01 5.54626048e-01 6.57301471e-02
1.12464404e+00 -1.18428385e+00 3.57048273e-01 3.96824449e-01
-6.82741165e-01 -1.43521890e-01 -4.81560916e-01 2.68836886e-01
7.18654040e-03 3.38783741e-01 -9.68834698e-01 4.94333714e-01
6.66009128e-01 -2.58440584e-01 -1.93505049e-01 4.19760495e-01
-1.69437408e-01 6.69778764e-01 -9.19340491e-01 8.59242752e-02
2.04802111e-01 -3.13485861e-01 6.96973324e-01 1.26971877e+00
1.33425295e-02 1.83666676e-01 -4.06886917e-03 9.37462509e-01
-3.18025291e-01 -2.37840712e-01 -6.42232656e-01 6.25176057e-02
9.65247929e-01 8.70778382e-01 -5.82225211e-02 -1.94787551e-02
1.80399027e-02 8.72978866e-01 2.55747110e-01 3.23537797e-01
-6.68750346e-01 -4.76939678e-01 1.38657939e+00 -1.52936727e-01
2.25568846e-01 -3.69027346e-01 -3.61848861e-01 -1.13859272e+00
8.97746682e-02 -1.00343800e+00 7.54515171e-01 -5.02474420e-02
-1.43561733e+00 1.96484700e-01 2.24837195e-02 -9.19501364e-01
-3.17883760e-01 -5.24695933e-01 -4.34915006e-01 1.08771133e+00
-1.13669753e+00 -1.08022273e+00 3.95002037e-01 8.29149127e-01
-4.12549853e-01 -1.04170948e-01 1.09122813e+00 -6.06116280e-02
-4.62504089e-01 1.67004275e+00 3.27307045e-01 3.56360883e-01
3.80305558e-01 -1.30000293e+00 6.09809339e-01 1.27792025e+00
6.26998395e-02 1.36320329e+00 7.60813296e-01 -5.19534230e-01
-1.61310220e+00 -9.52414572e-01 2.77288854e-01 -6.70696855e-01
6.04457796e-01 -5.52396774e-01 -5.60229897e-01 1.02593219e+00
-5.45184374e-01 1.61064982e-01 1.13682497e+00 1.51874542e-01
-9.36409414e-01 -3.86094689e-01 -2.03286362e+00 7.22761452e-01
6.14061475e-01 -8.57970715e-01 7.79878646e-02 2.30782673e-01
8.94575775e-01 -6.19296670e-01 -8.39597821e-01 2.57650614e-01
7.18068480e-01 -9.08853054e-01 1.13348222e+00 -7.05693126e-01
-3.31096917e-01 -2.30021954e-01 -6.48641706e-01 -5.09128988e-01
2.27296501e-01 -1.24401891e+00 -5.16780972e-01 9.74488616e-01
4.42793310e-01 -1.13101721e+00 1.11177313e+00 1.35404682e+00
4.94191855e-01 -5.84109426e-01 -1.24381912e+00 -7.99830317e-01
2.56511301e-01 -6.89549387e-01 9.07466829e-01 9.27484751e-01
-1.59348384e-01 -6.70121193e-01 -5.90798378e-01 8.00300717e-01
1.21382701e+00 -1.43665105e-01 1.12271643e+00 -6.14650071e-01
-7.37692416e-01 -1.19869098e-01 -4.25953180e-01 -1.13304734e+00
1.60489172e-01 -5.88212073e-01 -3.60705048e-01 -4.79846984e-01
-6.06407151e-02 -8.99588466e-01 -2.94880420e-01 4.67051953e-01
-1.38046592e-01 2.07682282e-01 1.84125081e-01 2.30700776e-01
-4.39930230e-01 1.52869970e-01 5.99739850e-01 8.48225653e-02
-8.51165429e-02 5.61697900e-01 -1.13240159e+00 6.38323307e-01
8.29581380e-01 -3.28692943e-01 -6.00101113e-01 -1.32854104e-01
6.00558035e-02 7.40337893e-02 5.59278488e-01 -5.82360864e-01
3.40989113e-01 -1.69965401e-01 9.80119631e-02 -2.10599601e-01
5.37013292e-01 -7.82707512e-01 1.74111128e-01 4.58850354e-01
-3.91492695e-01 -2.12649047e-01 1.51931375e-01 7.99334943e-01
1.83429435e-01 -2.87937284e-01 7.60203242e-01 -1.47880465e-01
-2.65550554e-01 6.47624075e-01 -1.44370168e-01 1.44100249e-01
9.58589554e-01 -2.22549707e-01 -4.04415548e-01 -6.17738724e-01
-6.14510536e-01 3.42078209e-01 6.95851862e-01 -2.46981859e-01
5.37904739e-01 -7.54497766e-01 -4.12076145e-01 3.30915570e-01
3.10346368e-03 4.85153049e-02 6.30510151e-02 3.52238506e-01
-5.24626791e-01 7.78270066e-02 4.08181250e-01 -2.64178723e-01
-1.40187037e+00 4.96749580e-01 3.81062776e-01 -2.35229522e-01
-3.32909048e-01 1.12850952e+00 2.50874132e-01 -4.86132771e-01
5.59036613e-01 7.08306767e-03 6.66466892e-01 -5.41597307e-01
6.40209794e-01 4.08580005e-01 -2.99047500e-01 -3.93504143e-01
-1.39769614e-01 -7.12876618e-02 -4.36049610e-01 -4.59621161e-01
7.38138378e-01 -2.46183857e-01 -1.72035009e-01 -1.57844514e-01
1.42284453e+00 3.84046286e-01 -1.53506136e+00 -2.24447936e-01
-4.30896133e-01 -1.00711596e+00 -2.61650980e-01 -6.52985036e-01
-8.80821288e-01 5.99106371e-01 4.58336949e-01 7.33484030e-02
9.69536901e-01 -1.81049764e-01 6.88263774e-01 5.88362753e-01
7.97407627e-01 -7.41157293e-01 -6.56269610e-01 7.08468677e-03
3.10610801e-01 -1.05524957e+00 1.33892864e-01 -4.61309701e-01
-4.21305746e-01 5.89673042e-01 3.59053552e-01 5.16656367e-03
8.08060646e-01 3.83168340e-01 2.37221885e-02 1.94527254e-01
-3.48018587e-01 4.88038361e-01 -2.36566588e-01 6.87533915e-01
-4.56172138e-01 2.52961487e-01 -1.68665603e-01 9.44152832e-01
-4.70452249e-01 -2.59200186e-01 3.88570011e-01 8.98931861e-01
-1.34735227e-01 -1.19721341e+00 -4.19588953e-01 5.18474996e-01
-1.13127887e+00 -8.16658661e-02 -1.74549207e-01 6.76289737e-01
-8.57386887e-02 9.20294702e-01 -1.77076533e-01 -4.63034928e-01
-3.24117839e-02 -3.00008327e-01 1.87641770e-01 -3.41968149e-01
-7.02919066e-01 -3.34144890e-01 -5.54413386e-02 -7.00295150e-01
1.41650587e-01 -7.16457427e-01 -1.07081878e+00 -8.75418842e-01
-2.90946543e-01 5.02796888e-01 7.48012781e-01 5.18478274e-01
4.70343053e-01 -5.08125484e-01 1.07093728e+00 -2.62750447e-01
-1.48326814e+00 -3.22700799e-01 -9.21804190e-01 4.00510162e-01
3.66247892e-01 -1.59101918e-01 -8.96690905e-01 -3.03754240e-01] | [5.940936088562012, 7.022603988647461] |
194445b5-d1ea-4383-9ac4-acb67477e028 | efficient-rule-learning-with-template | 2003.06071 | null | https://arxiv.org/abs/2003.06071v2 | https://arxiv.org/pdf/2003.06071v2.pdf | Towards Learning Instantiated Logical Rules from Knowledge Graphs | Efficiently inducing high-level interpretable regularities from knowledge graphs (KGs) is an essential yet challenging task that benefits many downstream applications. In this work, we present GPFL, a probabilistic rule learner optimized to mine instantiated first-order logic rules from KGs. Instantiated rules contain constants extracted from KGs. Compared to abstract rules that contain no constants, instantiated rules are capable of explaining and expressing concepts in more details. GPFL utilizes a novel two-stage rule generation mechanism that first generalizes extracted paths into templates that are acyclic abstract rules until a certain degree of template saturation is achieved, then specializes the generated templates into instantiated rules. Unlike existing works that ground every mined instantiated rule for evaluation, GPFL shares groundings between structurally similar rules for collective evaluation. Moreover, we reveal the presence of overfitting rules, their impact on the predictive performance, and the effectiveness of a simple validation method filtering out overfitting rules. Through extensive experiments on public benchmark datasets, we show that GPFL 1.) significantly reduces the runtime on evaluating instantiated rules; 2.) discovers much more quality instantiated rules than existing works; 3.) improves the predictive performance of learned rules by removing overfitting rules via validation; 4.) is competitive on knowledge graph completion task compared to state-of-the-art baselines. | ['Yu Guan', 'Yulong Gu', 'Paolo Missier'] | 2020-03-13 | null | null | null | null | ['inductive-knowledge-graph-completion'] | ['knowledge-base'] | [ 3.42669845e-01 9.47060525e-01 -7.16822624e-01 -2.90976048e-01
-6.98073387e-01 -7.68036783e-01 5.71426153e-01 1.67794719e-01
2.27444723e-01 8.88026118e-01 1.52059495e-01 -6.96877182e-01
-5.70868492e-01 -1.19999337e+00 -1.23520064e+00 -2.36990571e-01
-3.10817808e-01 7.69181788e-01 5.96711874e-01 -9.03581232e-02
-3.07487577e-01 3.58672649e-01 -1.61820972e+00 6.24678254e-01
9.75202441e-01 8.63091767e-01 -3.06527913e-01 2.19506547e-01
-1.69366255e-01 9.39929187e-01 -1.62234217e-01 -9.21909511e-01
2.12378100e-01 -6.78652432e-03 -1.13655043e+00 -9.47948322e-02
5.19303858e-01 1.65983751e-01 -2.21476525e-01 8.30280721e-01
-2.90281683e-01 -3.15044969e-02 6.79056048e-01 -1.24439037e+00
-4.96774137e-01 1.24534118e+00 1.12550845e-02 -1.10380225e-01
2.85523504e-01 1.20422363e-01 1.63484502e+00 -7.32550502e-01
9.06241000e-01 1.30832314e+00 7.02640653e-01 5.65380991e-01
-1.53841960e+00 -5.99171519e-01 4.79023993e-01 3.28441262e-02
-1.44713545e+00 -3.19931209e-01 4.44823980e-01 -2.60893285e-01
1.27545881e+00 3.80760908e-01 5.04065275e-01 1.24791551e+00
-5.40722944e-02 7.92222917e-01 7.36055374e-01 -3.96600634e-01
2.25029543e-01 2.47475971e-02 5.78078091e-01 1.23729873e+00
1.00386453e+00 1.36075005e-01 -6.33441150e-01 -4.69614804e-01
3.62682521e-01 -4.43041593e-01 -2.41312981e-01 -2.86635399e-01
-9.00113285e-01 7.48713553e-01 1.74359292e-01 -4.81579378e-02
-2.51506358e-01 1.84538186e-01 1.94816083e-01 4.93678041e-02
8.08521882e-02 6.09784067e-01 -9.61199164e-01 2.35645860e-01
-5.58908582e-01 3.80665183e-01 1.11984658e+00 1.28289199e+00
9.95980978e-01 -1.50077552e-01 -5.01238704e-01 5.23508728e-01
7.40733072e-02 4.33887094e-01 -3.27833854e-02 -7.18649149e-01
4.69879538e-01 1.25198257e+00 -1.86303437e-01 -7.03190804e-01
-4.22402948e-01 -5.57770967e-01 -4.84996229e-01 -2.36719772e-01
2.20349237e-01 -5.12587391e-02 -1.05938280e+00 1.63576317e+00
2.99261779e-01 1.92123279e-01 2.55057693e-01 4.32631850e-01
9.57072079e-01 4.66778904e-01 3.52407783e-01 -7.53365830e-02
1.38600028e+00 -6.49609447e-01 -4.08806980e-01 -4.48465854e-01
6.28641546e-01 1.90757308e-02 1.08581841e+00 5.94905138e-01
-8.16645920e-01 -2.16887966e-01 -1.09570098e+00 1.47240311e-01
-4.76454258e-01 3.23605910e-02 1.45519543e+00 6.75493419e-01
-6.30771816e-01 6.50430679e-01 -6.98254406e-01 2.28546560e-02
6.97466969e-01 5.67134380e-01 -5.67139566e-01 -2.13243321e-01
-1.49700081e+00 6.50972426e-01 1.02568555e+00 -2.72008568e-01
-7.64355898e-01 -1.17780113e+00 -1.20039546e+00 1.47196814e-01
1.17060232e+00 -1.09403896e+00 9.93188918e-01 -5.80719888e-01
-1.30146790e+00 5.64565659e-01 -3.61313313e-01 -6.12649202e-01
9.24371183e-02 -1.19566001e-01 -7.44092405e-01 -2.04873934e-01
3.84842195e-02 3.72555166e-01 5.61627746e-01 -1.25713670e+00
-7.14625120e-01 -1.86056241e-01 3.80256414e-01 -2.33578578e-01
-1.28695235e-01 -5.39561093e-01 -9.51046050e-01 -3.54318917e-01
9.04669687e-02 -8.78612518e-01 -1.36769712e-01 -8.16360474e-01
-9.57625151e-01 -6.81078315e-01 4.26434636e-01 -1.08590476e-01
1.40885103e+00 -1.77849364e+00 8.64788443e-02 6.95096374e-01
3.27111065e-01 2.97414780e-01 1.13421991e-01 1.88568234e-01
-1.26963064e-01 4.11427230e-01 -2.42160484e-01 9.97394323e-02
3.31988901e-01 6.97037578e-01 -5.64277649e-01 -2.11155146e-01
4.36969727e-01 1.23787463e+00 -1.01191890e+00 -4.88924563e-01
-1.34561416e-02 8.85369722e-03 -6.10116839e-01 -1.61344379e-01
-8.83844078e-01 -2.15269327e-01 -5.17152369e-01 7.43934274e-01
2.63138652e-01 -5.18914044e-01 6.48306370e-01 -3.23327899e-01
5.16383708e-01 5.46063244e-01 -1.33449113e+00 1.31500614e+00
-2.96755195e-01 -5.85385077e-02 -6.15608692e-01 -7.28417873e-01
8.81061375e-01 3.92860770e-02 -1.17330616e-02 -1.65409625e-01
-2.76589125e-01 9.40734148e-02 -3.11317146e-01 -3.88223201e-01
1.51598096e-01 -1.85111359e-01 -2.16325164e-01 6.26877397e-02
3.10887963e-01 -8.11427310e-02 6.09970093e-01 5.69500148e-01
1.44998741e+00 3.24716866e-01 4.82295156e-01 -2.07146600e-01
5.76625526e-01 2.03277811e-01 1.01201022e+00 1.01034582e+00
4.89793211e-01 -1.47978678e-01 8.90667498e-01 -8.06695640e-01
-2.25690052e-01 -1.20520425e+00 -8.51846561e-02 9.23317611e-01
-1.20360442e-01 -1.08616090e+00 -3.18336964e-01 -1.31600392e+00
4.91957903e-01 1.02616525e+00 -6.68241560e-01 -3.21002185e-01
-2.50217408e-01 -8.18891168e-01 9.07847166e-01 5.25916338e-01
2.63735026e-01 -9.93146062e-01 -2.60717608e-02 4.70566638e-02
-1.03672348e-01 -1.52219450e+00 -6.60786033e-02 4.15840179e-01
-8.10978591e-01 -1.60151935e+00 5.57068110e-01 -4.47177947e-01
6.51407897e-01 -3.01889598e-01 1.55009747e+00 2.68847849e-02
-1.43391669e-01 2.46544018e-01 -2.41113499e-01 -6.69098020e-01
-4.72674668e-01 3.05672377e-01 5.18995477e-03 -1.48285612e-01
5.99361360e-01 -6.49292827e-01 1.18397158e-02 1.35339543e-01
-9.44168270e-01 1.71134844e-01 8.92674446e-01 6.26580358e-01
1.01410246e+00 5.57670236e-01 4.84552115e-01 -1.72410226e+00
5.24678588e-01 -3.36289555e-01 -7.90429235e-01 5.94080567e-01
-1.09941936e+00 7.75700927e-01 8.15015674e-01 -1.36797264e-01
-1.08994818e+00 2.92487502e-01 3.52297872e-01 -2.50986397e-01
-2.32719943e-01 7.14940608e-01 -3.94544005e-01 2.24757046e-01
1.01577938e+00 -5.46965674e-02 -5.25379896e-01 -1.07448868e-01
5.10975659e-01 6.39488474e-02 6.02942586e-01 -1.20846391e+00
1.12487018e+00 2.81272888e-01 4.79710758e-01 -3.57781827e-01
-1.29002225e+00 -1.15015782e-01 -2.37804562e-01 2.65255898e-01
2.61628866e-01 -8.08598995e-01 -9.05164659e-01 -3.24685961e-01
-9.29169059e-01 -3.85789126e-01 -4.15558875e-01 1.41454861e-01
-2.54932106e-01 2.48065874e-01 -4.25743580e-01 -6.76694930e-01
-3.42438847e-01 -7.45292544e-01 9.60682392e-01 -1.14403896e-01
-4.21308756e-01 -1.02684355e+00 4.44488600e-02 3.91365975e-01
-3.05105925e-01 4.87891614e-01 1.36906743e+00 -7.68312633e-01
-7.02317357e-01 -1.11760408e-01 1.16472393e-01 2.44098872e-01
1.54874111e-02 -6.30634325e-03 -8.83178711e-01 2.99284488e-01
-9.27687287e-01 -2.43589148e-01 1.15328324e+00 3.03347576e-02
1.31446445e+00 -6.74771547e-01 -7.21358716e-01 6.46686137e-01
1.16257727e+00 -2.13359207e-01 6.08938098e-01 1.74917877e-01
6.22406423e-01 4.10506189e-01 3.37705314e-01 1.17663227e-01
4.87157643e-01 4.97568578e-01 3.17006320e-01 1.94995582e-01
-9.31988358e-02 -7.57175505e-01 4.14610893e-01 1.27009436e-01
-2.99816340e-01 -4.60170582e-02 -1.00306702e+00 5.56274235e-01
-2.12033200e+00 -8.79999936e-01 -3.94705534e-01 2.05406165e+00
1.24342132e+00 6.42619967e-01 2.87847314e-02 3.55577409e-01
4.48907912e-01 -2.48493090e-01 -4.58598197e-01 -3.08240682e-01
-3.94163936e-01 8.34527314e-01 6.05846465e-01 7.30281472e-01
-1.06897736e+00 1.34206033e+00 5.53169727e+00 9.53702271e-01
-5.34136713e-01 -2.28009254e-01 2.67707646e-01 -1.62954326e-03
-6.93965614e-01 4.88174617e-01 -1.20169485e+00 1.26613721e-01
7.47491777e-01 -4.75724339e-02 5.44856906e-01 9.85048175e-01
-3.61436278e-01 2.59328902e-01 -1.40173745e+00 4.99437362e-01
-2.43216828e-01 -1.69697046e+00 6.97530270e-01 -2.06467812e-03
8.51813793e-01 -2.94854045e-01 -2.23113984e-01 8.20681930e-01
8.70333910e-01 -1.23541117e+00 5.60701370e-01 6.38054729e-01
6.92352653e-01 -8.50583613e-01 5.80407500e-01 1.04967773e-01
-1.17389965e+00 -1.31082624e-01 -2.47967362e-01 -4.56043109e-02
-2.58113801e-01 1.24398363e+00 -1.17194629e+00 1.24600160e+00
4.64166820e-01 5.70754826e-01 -7.44237959e-01 3.95279497e-01
-1.01491129e+00 9.89324629e-01 -3.66638869e-01 8.77711922e-02
7.19521567e-02 1.21765792e-01 3.96577477e-01 1.44839931e+00
-1.74075082e-01 2.53974617e-01 6.31513819e-02 1.08245039e+00
-1.93397164e-01 -1.14357963e-01 -6.59307420e-01 -5.98926321e-02
4.02513385e-01 1.33654702e+00 -3.36505890e-01 -3.07171345e-01
-2.86960840e-01 5.03474116e-01 6.26646042e-01 5.79239905e-01
-7.28833020e-01 -4.01514828e-01 5.19224882e-01 2.78845161e-01
5.33304214e-01 2.80734032e-01 -2.23449022e-01 -1.06925225e+00
2.29984790e-01 -7.65154243e-01 1.07652557e+00 -4.82032239e-01
-1.18988788e+00 4.62479591e-01 1.89038560e-01 -5.09566307e-01
-2.71733105e-01 -8.75953197e-01 -4.01907176e-01 5.35752535e-01
-1.44931948e+00 -1.34976649e+00 -2.34985352e-01 7.28010058e-01
4.92536165e-02 -1.03866026e-01 9.87496018e-01 -1.63592756e-01
-6.06460869e-01 7.62151718e-01 -7.97622085e-01 7.16174990e-02
2.89604038e-01 -1.45745111e+00 3.03790987e-01 8.14022601e-01
5.15123487e-01 1.02380192e+00 5.62265873e-01 -9.06343222e-01
-1.53407538e+00 -1.60074651e+00 9.04610038e-01 -8.27300787e-01
6.07219219e-01 -2.29812577e-01 -8.01972151e-01 1.21391785e+00
-4.84718084e-01 3.86208177e-01 7.87054121e-01 8.99475873e-01
-1.00971913e+00 -2.97197342e-01 -8.36317718e-01 7.33696580e-01
1.54784954e+00 -2.41433322e-01 -7.63441145e-01 3.19869101e-01
7.14948416e-01 -2.47258916e-01 -8.76755536e-01 9.82838929e-01
5.34710288e-01 -6.91924989e-01 9.21546876e-01 -1.26429284e+00
2.52669394e-01 -4.50300157e-01 -3.10635474e-02 -1.01261544e+00
-6.34614706e-01 -1.02186561e+00 -9.94485199e-01 1.03367126e+00
1.02183127e+00 -6.11934841e-01 8.15846682e-01 6.18239939e-01
-2.21099466e-01 -1.15731585e+00 -5.89049160e-01 -1.04242456e+00
-2.08421022e-01 -7.82636404e-01 8.04065526e-01 8.04547191e-01
3.72683793e-01 7.42048144e-01 4.58415411e-02 6.27362847e-01
7.01227903e-01 4.11531270e-01 8.39577198e-01 -1.60474682e+00
-5.58061838e-01 -3.78647059e-01 -4.60913092e-01 -5.61452687e-01
7.05992460e-01 -1.44719028e+00 -3.87600482e-01 -1.63815188e+00
2.18136907e-01 -4.51128155e-01 -2.46570081e-01 1.24254954e+00
-4.06862885e-01 -1.64322674e-01 -3.50884646e-01 -2.50474215e-01
-8.09655547e-01 2.01119632e-01 9.60412264e-01 -2.11094588e-01
-2.12216318e-01 -2.85785701e-02 -1.13998580e+00 7.53306568e-01
6.05655968e-01 -3.57908338e-01 -6.78644121e-01 8.26531947e-02
7.94446826e-01 -3.13120425e-01 4.70275879e-01 -7.23719954e-01
1.24080643e-01 -4.81593341e-01 1.50111616e-01 -1.29985884e-01
-5.42604774e-02 -5.21462977e-01 4.28854734e-01 3.09966028e-01
-1.69866994e-01 -5.68250895e-01 3.95205259e-01 8.46750200e-01
1.49702117e-01 -1.09032065e-01 3.06071877e-01 -5.97925372e-02
-7.62594759e-01 3.14266026e-01 5.25955632e-02 3.08384567e-01
8.81267905e-01 -3.74326333e-02 -3.39258492e-01 4.49979957e-03
-7.82544494e-01 2.58969426e-01 -1.77075360e-02 3.14273745e-01
4.92796302e-01 -1.18308592e+00 -5.39927304e-01 6.56554252e-02
5.85399270e-01 4.06115323e-01 -3.38338524e-01 5.02700090e-01
-5.24688028e-02 7.08048880e-01 5.56588829e-01 -2.43978575e-01
-8.67805421e-01 6.40139520e-01 2.28136018e-01 -7.11934566e-01
-5.76926470e-01 7.96676159e-01 2.09466487e-01 -5.62385023e-01
1.52943522e-01 -6.89163744e-01 2.29049828e-02 -5.22460461e-01
3.18569452e-01 2.44457558e-01 4.18523073e-01 -2.94013461e-03
-5.71623862e-01 2.01910198e-01 -1.75529450e-01 3.08898956e-01
1.33236003e+00 6.89132869e-01 -5.39132431e-02 1.76084220e-01
3.97106022e-01 2.84018934e-01 -9.89175797e-01 -3.69451582e-01
3.74033988e-01 2.17668992e-02 -7.57109523e-02 -1.22679043e+00
-9.32157993e-01 4.68720607e-02 -4.44593996e-01 -8.62256391e-04
8.69319081e-01 4.61691946e-01 4.45792258e-01 8.79322410e-01
4.67008621e-01 -9.28796887e-01 -3.14206660e-01 6.77768052e-01
5.86222708e-01 -8.92390430e-01 1.92981958e-01 -1.12090981e+00
-6.16144598e-01 1.04744065e+00 5.54980457e-01 1.08959131e-01
4.09409732e-01 3.47123921e-01 -5.76886177e-01 -5.48837066e-01
-1.14780414e+00 -4.02958393e-01 7.63467848e-01 5.68542600e-01
6.87273368e-02 2.64902294e-01 -1.86795592e-02 1.50618744e+00
-3.83694738e-01 8.01777616e-02 7.65603259e-02 5.87088823e-01
-3.01739901e-01 -1.15244770e+00 -5.80066033e-02 7.18884051e-01
-3.37422818e-01 -3.00369263e-01 -7.14625180e-01 9.94160950e-01
4.21062887e-01 8.75022173e-01 -6.18227959e-01 -4.61451501e-01
6.04920983e-01 5.15131176e-01 7.08989024e-01 -9.00746584e-01
-4.16387439e-01 -7.06963241e-01 6.26189888e-01 -6.80186093e-01
3.24700447e-03 -4.31988418e-01 -1.53045499e+00 -4.51802611e-02
-3.75439227e-01 3.90187711e-01 5.09569831e-02 1.11359334e+00
7.18703270e-01 5.87444484e-01 4.85506058e-02 1.62317038e-01
-3.26463431e-01 -5.53106964e-01 -4.67481077e-01 5.70324063e-01
-1.13563314e-01 -7.99473763e-01 -3.08178723e-01 1.56163916e-01] | [8.969869613647461, 7.516846656799316] |
12ba881a-7e5f-44e1-ac9d-3812d27a1ea2 | convolutional-neural-network-array-for-sign | 2004.11836 | null | https://arxiv.org/abs/2004.11836v1 | https://arxiv.org/pdf/2004.11836v1.pdf | Convolutional Neural Network Array for Sign Language Recognition using Wearable IMUs | Advancements in gesture recognition algorithms have led to a significant growth in sign language translation. By making use of efficient intelligent models, signs can be recognized with precision. The proposed work presents a novel one-dimensional Convolutional Neural Network (CNN) array architecture for recognition of signs from the Indian sign language using signals recorded from a custom designed wearable IMU device. The IMU device makes use of tri-axial accelerometer and gyroscope. The signals recorded using the IMU device are segregated on the basis of their context, such as whether they correspond to signing for a general sentence or an interrogative sentence. The array comprises of two individual CNNs, one classifying the general sentences and the other classifying the interrogative sentence. Performances of individual CNNs in the array architecture are compared to that of a conventional CNN classifying the unsegregated dataset. Peak classification accuracies of 94.20% for general sentences and 95.00% for interrogative sentences achieved with the proposed CNN array in comparison to 93.50% for conventional CNN assert the suitability of the proposed approach. | ['Rinki Gupta', 'Karush Suri'] | 2020-04-21 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 3.98023754e-01 -1.24111705e-01 -2.00212643e-01 -4.07089829e-01
-3.59218985e-01 -5.59856296e-01 6.15645051e-01 -4.93072003e-01
-8.64918232e-01 4.80366915e-01 3.23094368e-01 -2.74541825e-01
1.14703029e-01 -2.82323450e-01 -5.08478105e-01 -8.15171421e-01
3.88448238e-01 3.07179708e-02 -4.72864695e-02 -2.16780864e-02
6.32602945e-02 5.38215578e-01 -1.47999799e+00 5.14183231e-02
1.97292402e-01 1.31953299e+00 -1.53154537e-01 1.13151884e+00
2.69225817e-02 4.46246356e-01 -1.07446885e+00 1.54555991e-01
2.41413593e-01 -5.00448406e-01 -2.83504337e-01 -1.43612534e-01
7.59513557e-01 -5.30642152e-01 -2.76553005e-01 7.07846105e-01
9.63268459e-01 -2.98111081e-01 5.04169405e-01 -7.55645037e-01
-4.11187559e-01 3.44775677e-01 -1.04981281e-01 3.28958929e-01
4.21738654e-01 -2.33106986e-02 7.42896080e-01 -6.64350331e-01
6.79246366e-01 5.89878678e-01 8.13699305e-01 6.70865476e-01
-5.34979939e-01 -6.19661093e-01 -3.09736669e-01 -1.19335897e-01
-1.07548738e+00 -3.09461206e-01 8.37168753e-01 -5.51275432e-01
1.03427649e+00 3.88114750e-01 9.40176904e-01 9.96576607e-01
3.17110807e-01 8.59505475e-01 1.21486509e+00 -5.91405869e-01
2.91754574e-01 -2.75741726e-01 4.85909253e-01 6.20893955e-01
5.14839172e-01 -1.86004475e-01 -6.62907243e-01 2.22873405e-01
8.21887553e-01 -5.35049476e-02 -5.47071621e-02 1.09351687e-01
-1.33833671e+00 2.68463314e-01 4.28290725e-01 7.71756589e-01
-6.23101056e-01 4.20651376e-01 3.75353873e-01 1.94928423e-01
3.78686897e-02 2.26790354e-01 -3.46397817e-01 -4.66160864e-01
-8.90330613e-01 1.23148203e-01 7.56475985e-01 7.74783254e-01
-1.13653265e-01 4.08548027e-01 1.26697078e-01 4.53782469e-01
5.08413434e-01 7.95100272e-01 8.36038172e-01 -1.86545387e-01
6.22641325e-01 8.08444798e-01 6.73654377e-02 -7.09657907e-01
-6.70799315e-01 -4.47648436e-01 -7.49436855e-01 1.59254551e-01
7.41392672e-01 -7.48105705e-01 -1.18866968e+00 1.19763756e+00
6.93952814e-02 5.59526607e-02 2.51051635e-01 1.29453945e+00
1.17389548e+00 1.87607482e-01 -1.09963283e-01 4.54070508e-01
1.45751810e+00 -2.37891093e-01 -7.85770178e-01 -1.90935675e-02
5.02860606e-01 -5.46566308e-01 7.99701452e-01 3.26950580e-01
-5.19123971e-01 -4.78482693e-01 -1.55768335e+00 -1.69955250e-02
-4.00300294e-01 7.80442715e-01 3.78660321e-01 1.04962885e+00
-8.17103148e-01 1.12976268e-01 -8.72739851e-01 -6.65269375e-01
1.75776646e-01 8.62598300e-01 -1.94508672e-01 5.46537280e-01
-7.83902824e-01 5.69557786e-01 1.86432675e-01 7.06991434e-01
3.82314213e-02 1.69946611e-01 -5.21963120e-01 -3.50349844e-01
-5.18145382e-01 -2.63372868e-01 1.13130307e+00 -9.74763870e-01
-1.95531082e+00 9.75613892e-01 4.89783357e-04 -4.67402726e-01
6.59608483e-01 -3.53798270e-01 -5.57610869e-01 7.87258744e-02
-3.18647444e-01 3.22997838e-01 6.30860388e-01 -5.65589070e-01
-5.84320486e-01 -4.32983816e-01 -2.61539131e-01 1.32000357e-01
-2.85536438e-01 8.54788572e-02 3.16166654e-02 -4.91320252e-01
7.49019325e-01 -1.00226891e+00 4.03261214e-01 -1.24081612e-01
-5.15202880e-01 -1.60993382e-01 1.16203010e+00 -7.20652223e-01
9.26371515e-01 -1.77162588e+00 -1.79875806e-01 1.41405776e-01
1.76584974e-01 8.70385170e-01 3.18102270e-01 2.04799190e-01
-1.94154512e-02 -1.77221477e-01 1.13708533e-01 4.96999687e-03
-3.85953784e-02 3.92064750e-01 -5.69916926e-02 6.24033749e-01
2.09718466e-01 1.08843100e+00 -6.73058152e-01 -2.03118369e-01
2.56413996e-01 5.61719656e-01 3.46925668e-02 2.05662623e-02
2.46987909e-01 6.57462776e-01 -5.86778402e-01 1.06125689e+00
1.02205366e-01 9.53876004e-02 1.49787754e-01 -1.05114408e-01
-2.75088698e-01 3.60401899e-01 -1.08370435e+00 1.29251921e+00
-1.14003018e-01 1.31287718e+00 -3.48685198e-02 -8.63994122e-01
1.26869702e+00 7.39142358e-01 1.08906396e-01 -7.07307100e-01
7.86412060e-01 6.10741675e-01 1.87536791e-01 -9.54670191e-01
3.04649591e-01 -2.63953116e-02 -1.03363484e-01 4.13087875e-01
-2.92890854e-02 1.49336696e-01 -5.59245050e-02 -6.05384350e-01
9.51800406e-01 2.13609681e-01 1.46450430e-01 1.19248461e-02
4.20677602e-01 -2.65233576e-01 1.73598602e-01 6.98046803e-01
-4.57932353e-01 7.01928496e-01 1.24925204e-01 -9.29647982e-01
-7.64008999e-01 -8.04694116e-01 -1.39585100e-02 6.29789472e-01
-1.81233078e-01 3.72549444e-01 -6.78489029e-01 -3.23279470e-01
-1.18373282e-01 -1.89547073e-02 -4.63975489e-01 3.24447095e-01
-1.01758456e+00 -6.54047191e-01 1.15655303e+00 8.69716167e-01
1.06950223e+00 -1.08297086e+00 -1.51145208e+00 1.32636413e-01
1.26139358e-01 -1.12392807e+00 1.09709814e-01 2.36644685e-01
-9.52756643e-01 -8.81749392e-01 -1.00706255e+00 -9.70616221e-01
4.93098468e-01 -4.39240932e-01 3.90385360e-01 -1.02651067e-01
7.05659296e-03 2.54530996e-01 -5.16058862e-01 -8.25500250e-01
-2.61594690e-02 1.74418613e-01 2.07498178e-01 4.80438173e-01
7.58093834e-01 -4.51137662e-01 -5.98159730e-01 -2.81712580e-02
-5.63323498e-01 -1.49233162e-01 5.90808749e-01 6.24128282e-01
1.31393895e-01 -9.90608990e-01 3.40965599e-01 -5.52919731e-02
9.30209160e-01 -8.65448732e-03 -2.63336241e-01 6.05803765e-02
-2.54680574e-01 -1.52214384e-02 1.72366709e-01 -6.14207327e-01
-4.21581298e-01 3.16527605e-01 -2.10505694e-01 1.48498788e-01
-3.97484481e-01 4.31455344e-01 1.84311032e-01 -2.22313404e-01
6.68021083e-01 4.21771884e-01 1.80654209e-02 -4.12831038e-01
-3.90627794e-02 1.47885811e+00 9.24920678e-01 -1.03606530e-01
2.61762440e-01 3.37168545e-01 7.77224302e-02 -1.27139950e+00
-2.32830886e-02 -4.68551219e-01 -8.90260279e-01 -8.17528188e-01
9.96563971e-01 -5.75370669e-01 -8.91162038e-01 1.17438054e+00
-1.11640000e+00 3.49879498e-03 -2.94408537e-02 8.63033950e-01
-2.02459395e-01 4.73060533e-02 -4.72731620e-01 -1.17971063e+00
-5.82245827e-01 -9.96448815e-01 1.20417261e+00 5.75141966e-01
-6.86893761e-01 -6.48762703e-01 -5.79735301e-02 2.97137946e-01
4.80087072e-01 7.76054025e-01 3.15739185e-01 -8.14178705e-01
-1.63285628e-01 -1.35103834e+00 -5.95853925e-02 2.87603021e-01
2.08450109e-01 -1.28685579e-01 -1.10434353e+00 8.85172486e-02
-1.00952365e-01 -2.35272005e-01 5.48756361e-01 4.27966326e-01
4.03770208e-02 -2.08954751e-01 -1.62680056e-02 4.07231688e-01
1.28290749e+00 6.53401017e-01 5.81107378e-01 4.89494264e-01
6.14747524e-01 -9.10098944e-03 -1.74632654e-01 1.23263195e-01
1.61998898e-01 5.59931815e-01 1.30616307e-01 -1.06128473e-02
-2.68589824e-01 5.23388907e-02 2.49126509e-01 7.99461305e-01
-4.86162484e-01 -2.91128546e-01 -1.17313182e+00 5.60158193e-01
-1.56589663e+00 -6.20437205e-01 -3.64579737e-01 1.82921362e+00
2.93476850e-01 5.22745959e-02 1.70767024e-01 7.58148491e-01
6.02777064e-01 1.11670822e-01 -4.28633243e-01 -6.95124090e-01
-3.80725116e-01 4.75753576e-01 7.44691670e-01 3.56266141e-01
-1.29954338e+00 6.02740705e-01 5.87878609e+00 -1.61328286e-01
-2.04733968e+00 -1.99442983e-01 1.50661856e-01 -1.80921793e-01
6.85534537e-01 -6.75096154e-01 -7.51246929e-01 3.77335936e-01
9.89649117e-01 5.50067663e-01 -1.57483593e-01 6.04079008e-01
2.81273395e-01 -3.44092757e-01 -8.56456637e-01 1.06488502e+00
2.41859108e-01 -1.10941064e+00 -9.45855454e-02 -5.75490296e-02
5.72038293e-01 2.65230030e-01 5.64642716e-03 -2.65096605e-01
-4.76381183e-01 -7.69189894e-01 8.65751743e-01 7.31878698e-01
9.27847028e-01 -1.69260487e-01 1.22664583e+00 3.35837781e-01
-1.06042171e+00 -3.03125586e-02 3.23746115e-01 -8.56929302e-01
8.18605497e-02 -3.76196727e-02 -1.17929864e+00 1.60894394e-01
4.40560699e-01 4.36314672e-01 -3.57756525e-01 8.40585887e-01
-3.13415885e-01 1.08311582e+00 -7.31157601e-01 -9.94487941e-01
5.23825586e-01 2.14614738e-02 6.12744570e-01 1.37182736e+00
3.84795964e-01 8.85929763e-02 -4.89871651e-01 3.07003170e-01
1.77210003e-01 8.24974105e-02 -6.43805325e-01 -3.60642731e-01
-1.39955264e-02 6.56390011e-01 -7.79232264e-01 -5.10636389e-01
-2.32930988e-01 1.01077616e+00 -3.39421153e-01 4.80707824e-01
-2.63383478e-01 -7.30333269e-01 5.11997283e-01 -9.95334759e-02
3.22433263e-01 -6.49327338e-01 -9.35393691e-01 -1.07537377e+00
6.17421448e-01 -4.34719324e-01 -2.29366757e-02 -7.04032421e-01
-5.73594868e-01 5.94166100e-01 -4.01101410e-01 -1.51206195e+00
-6.31742239e-01 -1.07957268e+00 -5.49305737e-01 9.76545691e-01
-1.05305779e+00 -1.20856225e+00 -4.81320411e-01 3.11901420e-01
2.57521152e-01 -2.91416138e-01 9.05678451e-01 1.09269328e-01
-5.44035614e-01 6.01103783e-01 7.92609602e-02 1.04160595e+00
3.06342572e-01 -1.00557387e+00 4.20524001e-01 1.03082335e+00
1.95488334e-01 7.08060741e-01 4.85823184e-01 -4.73406702e-01
-1.36741650e+00 -6.99435174e-01 1.44087744e+00 -3.54167044e-01
3.43949199e-01 -3.82457733e-01 -1.51894793e-01 5.48967540e-01
8.42179731e-03 7.11517932e-05 5.92370272e-01 -4.56172556e-01
-3.28242570e-01 -8.55189264e-02 -1.17569840e+00 5.23010373e-01
7.76787758e-01 -6.39034569e-01 -8.49166274e-01 1.74777377e-02
-1.84917450e-01 -7.56409764e-01 -4.48772311e-01 1.41175106e-01
1.56176698e+00 -5.61724663e-01 5.16672909e-01 -6.30274951e-01
3.12360495e-01 -3.12953472e-01 -2.15062976e-01 -7.30949998e-01
2.91530788e-01 -4.36241865e-01 -1.38147652e-01 6.32122099e-01
3.66401166e-01 -1.02759528e+00 1.09078920e+00 6.91893935e-01
4.94324379e-02 -7.30845034e-01 -1.39381993e+00 -7.60854542e-01
-4.16753978e-01 -5.03678024e-01 3.73235494e-01 4.84494001e-01
1.19042635e-01 3.92920792e-01 -3.76672804e-01 3.98592129e-02
2.21192703e-01 -1.77217424e-01 8.04529071e-01 -1.19830322e+00
1.24359481e-01 -5.61466038e-01 -1.28593576e+00 -1.11963665e+00
-5.35149753e-01 -6.23200834e-01 1.36217792e-02 -1.54995239e+00
-6.75982893e-01 2.51112014e-01 -1.03662550e-01 4.54865545e-01
2.83650100e-01 7.45715559e-01 1.33091182e-01 2.00904101e-01
-1.08255103e-01 -1.82694405e-01 8.97363663e-01 -1.92039669e-01
-3.54922861e-01 4.54904251e-02 1.87076125e-02 6.11610174e-01
9.11716640e-01 -7.77719729e-03 2.35649928e-01 -6.02319121e-01
2.29340985e-01 -1.11184083e-01 5.40459454e-01 -1.52092791e+00
4.40211713e-01 6.15780294e-01 6.41402483e-01 -5.65866828e-01
2.68732697e-01 -6.98237598e-01 -6.60292879e-02 8.92598391e-01
-2.96203017e-01 1.10410012e-01 2.80885380e-02 2.64642417e-01
-3.73326510e-01 2.38193795e-01 4.24899012e-01 2.50080317e-01
-7.05427408e-01 -4.47998405e-01 -7.32726038e-01 -5.75989187e-01
5.25747120e-01 -1.01337302e+00 7.93950856e-02 -4.84312117e-01
-8.08693051e-01 -3.93228024e-01 -1.57017082e-01 4.50450957e-01
6.06577516e-01 -1.36495686e+00 -4.82740909e-01 4.21511173e-01
5.89300413e-03 -1.86751887e-01 -3.56679112e-01 9.25340295e-01
-9.53767836e-01 1.03391600e+00 -4.00719345e-01 -7.07570016e-01
-1.54268777e+00 -5.90469003e-01 5.28340876e-01 1.70442373e-01
-5.89542747e-01 9.22979474e-01 -8.09708357e-01 -2.60424286e-01
3.70674461e-01 -1.08661771e+00 -3.46826136e-01 1.67842895e-01
4.58893716e-01 2.81990647e-01 2.77325481e-01 -9.37064767e-01
-4.96526331e-01 9.98859584e-01 3.85279447e-01 -4.34108347e-01
1.25984895e+00 3.01155567e-01 3.23477656e-01 7.11090922e-01
1.12237918e+00 -4.06789184e-01 -8.88390243e-01 -1.83046430e-01
1.01650283e-01 1.47920111e-02 -1.94751352e-01 -1.17332160e+00
-8.09210241e-01 9.05422747e-01 1.06635964e+00 -1.01751750e-02
1.05833447e+00 -4.34537649e-01 8.40434968e-01 7.22867906e-01
3.57134610e-01 -1.20080435e+00 -3.57905120e-01 7.69161046e-01
9.82410610e-01 -9.10073996e-01 -3.80615383e-01 3.49948734e-01
-1.60439357e-01 1.47845471e+00 8.94362628e-02 -6.36294186e-01
5.30490041e-01 1.81479663e-01 7.25793839e-01 -3.42856646e-01
1.75682232e-01 -2.07319453e-01 5.56665599e-01 5.27539611e-01
7.74772406e-01 4.16176945e-01 -9.79421735e-01 4.32235062e-01
-5.05238116e-01 4.62315500e-01 3.24971080e-01 1.46481776e+00
-2.43339255e-01 -4.99229223e-01 -5.31484425e-01 5.82491577e-01
-4.65166897e-01 1.84256509e-01 -1.02575111e+00 8.73449564e-01
3.14511180e-01 9.43420827e-01 1.71719313e-01 -7.51538694e-01
5.37619472e-01 7.32993126e-01 3.78412992e-01 -1.84214652e-01
-9.32024539e-01 1.71944802e-03 3.80665928e-01 -2.17619181e-01
-5.38583159e-01 -8.36473286e-01 -1.15420353e+00 3.44056964e-01
-2.21358404e-01 -3.87516856e-01 1.26908970e+00 1.28665674e+00
1.66096538e-01 3.09295595e-01 7.30853109e-03 -1.02008331e+00
-3.15584809e-01 -1.36915123e+00 -3.92965466e-01 3.62020694e-02
7.15575039e-01 -3.73504162e-01 -7.67813176e-02 1.35487348e-01] | [9.069401741027832, -6.361203670501709] |
355d43ce-9af8-4d65-ae13-7da976db5a95 | context-aware-feature-generation-for-zero | 2008.06893 | null | https://arxiv.org/abs/2008.06893v1 | https://arxiv.org/pdf/2008.06893v1.pdf | Context-aware Feature Generation for Zero-shot Semantic Segmentation | Existing semantic segmentation models heavily rely on dense pixel-wise annotations. To reduce the annotation pressure, we focus on a challenging task named zero-shot semantic segmentation, which aims to segment unseen objects with zero annotations. This task can be accomplished by transferring knowledge across categories via semantic word embeddings. In this paper, we propose a novel context-aware feature generation method for zero-shot segmentation named CaGNet. In particular, with the observation that a pixel-wise feature highly depends on its contextual information, we insert a contextual module in a segmentation network to capture the pixel-wise contextual information, which guides the process of generating more diverse and context-aware features from semantic word embeddings. Our method achieves state-of-the-art results on three benchmark datasets for zero-shot segmentation. Codes are available at: https://github.com/bcmi/CaGNet-Zero-Shot-Semantic-Segmentation. | ['Liqing Zhang', 'Zihan Zhao', 'Siyuan Zhou', 'Li Niu', 'Zhangxuan Gu'] | 2020-08-16 | null | null | null | null | ['zero-shot-segmentation'] | ['computer-vision'] | [ 2.88899511e-01 2.38688529e-01 -3.03515315e-01 -6.86174452e-01
-7.82615423e-01 -3.11415017e-01 3.27094972e-01 2.18787968e-01
-5.34860671e-01 2.30219468e-01 3.69321287e-01 1.13030553e-01
2.73919195e-01 -1.00138950e+00 -6.61485970e-01 -6.12010598e-01
4.81901944e-01 1.60034329e-01 6.57326519e-01 -1.67138338e-01
2.08326682e-01 -1.77566290e-01 -1.48819721e+00 6.99465200e-02
9.50091243e-01 8.98329854e-01 4.78651971e-01 4.21547055e-01
-4.62271512e-01 2.30067268e-01 -1.86740071e-01 -2.18350694e-01
9.12685618e-02 -4.11378890e-01 -1.05420566e+00 2.72357821e-01
1.39108330e-01 -3.34513001e-02 -1.10633589e-01 1.23725569e+00
2.32751340e-01 4.16427463e-01 5.74758351e-01 -1.08703625e+00
-8.63337398e-01 6.67609036e-01 -5.12159705e-01 1.17247753e-01
2.01900024e-02 2.23073438e-01 1.38183331e+00 -8.96588027e-01
7.22440362e-01 1.14129937e+00 4.07645494e-01 7.38701701e-01
-1.05764377e+00 -4.07242298e-01 4.26572025e-01 2.27645546e-01
-1.35703647e+00 -8.57243463e-02 8.00845861e-01 -5.81506908e-01
5.37186265e-01 -5.57970954e-03 7.43662715e-01 9.23618972e-01
-3.61773878e-01 1.02263856e+00 5.22566438e-01 -3.83486480e-01
4.22072560e-01 -1.10576957e-01 3.89152080e-01 5.90725243e-01
1.25353962e-01 -4.82137382e-01 -2.90119261e-01 3.45543951e-01
6.01003826e-01 3.74354452e-01 -7.99119323e-02 -4.80020732e-01
-1.08047128e+00 9.83210385e-01 7.77634144e-01 5.34656584e-01
-1.62139520e-01 2.26111025e-01 4.10547554e-01 -2.94596016e-01
5.89221954e-01 4.34336931e-01 -6.24118984e-01 -1.16149433e-01
-7.25451589e-01 -9.50658694e-02 3.83047789e-01 1.02641702e+00
1.17320740e+00 -2.60363430e-01 -5.63021123e-01 1.01302576e+00
3.09866130e-01 1.58292800e-01 6.30706906e-01 -8.25046122e-01
2.64228851e-01 8.63847435e-01 -7.26382658e-02 -5.41354179e-01
-2.99731404e-01 -2.50646800e-01 -4.03784662e-01 -2.29357034e-01
1.71063274e-01 -1.46213815e-01 -1.62478507e+00 1.59976685e+00
6.05621994e-01 5.92946172e-01 -4.65849750e-02 1.16112816e+00
1.04475033e+00 6.64600790e-01 4.51812178e-01 3.38218868e-01
1.46010804e+00 -1.45461512e+00 -4.14693058e-01 -3.83702368e-01
7.61555910e-01 -4.32403982e-01 1.51024127e+00 -1.38176754e-01
-5.44159651e-01 -5.00592828e-01 -1.00321496e+00 -4.37805623e-01
-6.30259633e-01 -2.12714046e-01 3.81743908e-01 3.77578169e-01
-6.12095356e-01 3.60162050e-01 -8.51016760e-01 -4.34962332e-01
7.56187618e-01 3.41660380e-02 -7.11143538e-02 -3.22234005e-01
-1.21246696e+00 3.77484471e-01 6.59837782e-01 -2.49443576e-01
-8.76809299e-01 -7.70299077e-01 -1.17934668e+00 1.81447133e-01
7.29915023e-01 -4.37528819e-01 1.22639799e+00 -8.14976573e-01
-1.15939343e+00 8.63365173e-01 -2.41056845e-01 -2.44002357e-01
1.22958161e-01 -1.98144749e-01 -8.55108798e-02 2.42590472e-01
5.20312846e-01 1.10987973e+00 6.29195333e-01 -1.26802862e+00
-6.97104454e-01 -3.83810282e-01 1.44886971e-01 1.28185779e-01
-5.01242816e-01 -3.49637389e-01 -9.85902667e-01 -7.32493758e-01
1.01093180e-01 -7.23419666e-01 -6.10094547e-01 -4.81573455e-02
-5.37143707e-01 -4.48908627e-01 8.71130049e-01 -4.07861561e-01
1.09830415e+00 -2.32652974e+00 1.09054014e-01 -1.15758821e-01
1.53559923e-01 3.59320700e-01 -2.99089462e-01 5.27111515e-02
2.66465008e-01 3.71745259e-01 -7.45775104e-01 -4.23858881e-01
1.24340594e-01 2.86743820e-01 4.14218009e-02 4.66834344e-02
3.88379127e-01 1.21725857e+00 -1.16184938e+00 -7.09797800e-01
3.25566918e-01 4.65485215e-01 -7.09496200e-01 8.83726999e-02
-6.18821561e-01 5.18506467e-01 -7.91161418e-01 7.23129570e-01
4.66449708e-01 -3.45845282e-01 -2.18422309e-01 -1.11609377e-01
9.50316142e-04 -1.65796369e-01 -7.55780458e-01 2.23675537e+00
-4.27567095e-01 3.20438981e-01 -3.12583476e-01 -1.12554300e+00
7.97471881e-01 9.67341885e-02 4.72402453e-01 -6.77879989e-01
5.67881405e-01 4.94114980e-02 -3.18968296e-01 -4.01394874e-01
6.71232939e-01 -2.65860558e-03 -5.18759847e-01 1.93348989e-01
3.78373802e-01 -1.85531512e-01 3.97305518e-01 3.20172131e-01
8.31333637e-01 1.77223980e-01 4.99784797e-02 -2.28634715e-01
3.53399366e-01 1.84792757e-01 9.79337633e-01 4.55828458e-01
-3.60389858e-01 1.02473438e+00 4.98948127e-01 -1.12672642e-01
-8.67213309e-01 -1.07480025e+00 -7.29996264e-02 1.33419502e+00
5.54481268e-01 -4.62891757e-01 -1.09993279e+00 -6.97218299e-01
-7.43164718e-02 7.55767822e-01 -8.77889037e-01 -2.14933276e-01
-1.86444551e-01 -5.05621135e-01 1.66704416e-01 7.86677718e-01
4.28905904e-01 -1.09362888e+00 -6.07864082e-01 3.36960465e-01
-1.77935019e-01 -1.19970763e+00 -6.89059138e-01 5.09446785e-02
-6.91815794e-01 -9.92124379e-01 -9.17926967e-01 -9.69149649e-01
8.75175953e-01 3.84635448e-01 8.81693900e-01 5.43087861e-03
-6.08471870e-01 3.14968407e-01 -8.61571610e-01 -3.13249946e-01
1.63050696e-01 3.67813885e-01 -5.77361524e-01 1.33753061e-01
6.54389799e-01 -2.55252123e-01 -8.78351748e-01 2.70333350e-01
-1.13991344e+00 1.87017009e-01 3.39182407e-01 6.44635975e-01
1.04856443e+00 -3.68130296e-01 6.36568010e-01 -1.16023719e+00
3.64254892e-01 -5.74846029e-01 -4.33553100e-01 1.55700237e-01
-3.48248333e-01 -5.77226281e-02 4.92232829e-01 -2.73058712e-01
-9.94644761e-01 3.33097339e-01 -1.59966454e-01 -4.76631075e-01
-2.96181321e-01 2.70332873e-01 -4.31719750e-01 2.69273609e-01
3.81331831e-01 1.28218025e-01 -4.71977681e-01 -6.10671580e-01
9.84185874e-01 6.91403627e-01 5.39032042e-01 -4.94988292e-01
4.75874811e-01 5.29522121e-01 -4.40160006e-01 -8.02099049e-01
-1.33419716e+00 -9.75307763e-01 -8.23341548e-01 4.06330936e-02
1.54467237e+00 -9.28215623e-01 1.21677533e-01 5.29174268e-01
-8.13082159e-01 -6.65822685e-01 -6.04686439e-01 1.17129944e-01
-5.45619786e-01 2.41914749e-01 -4.09558564e-01 -1.58096254e-01
-3.15454245e-01 -1.21595407e+00 1.28935933e+00 6.76473498e-01
-1.18136004e-01 -9.92591918e-01 7.50751123e-02 5.57775199e-01
1.14949852e-01 1.49046630e-01 6.84860051e-01 -7.17802584e-01
-6.01652443e-01 -1.17346771e-01 -5.67753971e-01 4.96069074e-01
2.22659945e-01 -1.76564053e-01 -8.99702847e-01 1.45610496e-02
-5.94516754e-01 -3.01302522e-01 1.23056889e+00 4.30641919e-01
1.47240460e+00 -3.18382792e-02 -2.98169166e-01 6.87824190e-01
1.54945076e+00 -5.99837452e-02 4.85475183e-01 7.45216906e-02
1.23871303e+00 3.32694709e-01 9.56208646e-01 4.37179476e-01
5.96806347e-01 4.69923377e-01 3.26774061e-01 -9.41910669e-02
-2.48416945e-01 -4.21988189e-01 -2.72921532e-01 6.77373886e-01
3.84165138e-01 -3.20198476e-01 -9.52329576e-01 1.15857303e+00
-1.93228185e+00 -4.38584983e-01 -2.83384528e-02 1.78175938e+00
8.80313456e-01 1.68148667e-01 -2.04150137e-02 -1.17570885e-01
8.79137754e-01 4.57071334e-01 -7.21232057e-01 -2.44569853e-01
3.19087535e-01 2.39018664e-01 5.25821269e-01 4.35159683e-01
-1.26168013e+00 1.64361024e+00 4.79763746e+00 9.40998495e-01
-9.30430293e-01 6.26664817e-01 6.07512355e-01 1.08254896e-02
-4.31776911e-01 7.24668577e-02 -8.01120818e-01 7.21256852e-01
5.48022032e-01 -2.00731680e-01 -8.39748532e-02 9.62091029e-01
9.34149325e-02 -1.81404337e-01 -6.51965916e-01 6.29453063e-01
1.62968978e-01 -1.16380119e+00 5.49996048e-02 -2.56616682e-01
1.13035214e+00 1.51526153e-01 -1.75579354e-01 3.54491949e-01
4.22985882e-01 -7.16870070e-01 5.18040836e-01 3.49260092e-01
8.36827219e-01 -6.22069538e-01 6.14849150e-01 -6.29198924e-02
-1.46669376e+00 4.90730815e-02 -4.06971127e-01 2.19568074e-01
4.45397317e-01 8.02201629e-01 -5.35483718e-01 3.14048588e-01
6.60589099e-01 9.70008910e-01 -5.07618189e-01 1.23514652e+00
-5.49790144e-01 7.00629890e-01 -8.39155093e-02 6.70570508e-02
6.19005263e-01 -1.19706884e-01 1.78548485e-01 1.28650546e+00
2.20935494e-01 2.89672107e-01 4.56250340e-01 8.83469462e-01
-3.46199274e-01 1.57803297e-01 -2.67622590e-01 -1.39494881e-01
4.10591543e-01 1.33185947e+00 -1.17385495e+00 -5.35533309e-01
-4.95241702e-01 1.10838175e+00 1.95816442e-01 3.25737685e-01
-8.38687718e-01 -6.52517378e-01 8.04692268e-01 -9.45793092e-02
5.99851608e-01 -1.10578910e-01 -4.77238268e-01 -1.08263922e+00
-1.09944992e-01 -2.95200404e-02 3.47631574e-01 -7.39214003e-01
-1.04760873e+00 2.09617659e-01 -8.00707564e-02 -1.09181631e+00
2.56586879e-01 -3.67195129e-01 -8.57933044e-01 4.39556867e-01
-1.47577131e+00 -1.23728669e+00 -6.15341842e-01 4.67857331e-01
1.10483098e+00 3.21471959e-01 5.16654491e-01 2.61361867e-01
-6.30520880e-01 4.09058928e-01 -1.31275862e-01 3.02531540e-01
5.75926840e-01 -1.27463543e+00 6.54604018e-01 9.35205996e-01
8.86804834e-02 1.79218218e-01 4.46698368e-01 -6.23159707e-01
-7.62822509e-01 -1.61444008e+00 6.23820126e-01 -2.73171961e-01
6.91454530e-01 -4.12721187e-01 -1.02494979e+00 5.80740690e-01
-1.58051848e-01 4.45234895e-01 6.24591529e-01 1.13912392e-02
-3.23711544e-01 1.51807114e-01 -9.98654068e-01 5.19295812e-01
1.24125373e+00 -4.57385123e-01 -6.92744732e-01 2.34100834e-01
1.40329039e+00 -1.39287710e-01 -7.24921048e-01 2.53634930e-01
2.07385182e-01 -4.29505527e-01 8.89302492e-01 -4.23042983e-01
5.70693672e-01 -3.40495646e-01 -4.51479070e-02 -1.30962515e+00
-2.86028951e-01 -4.06480357e-02 2.78649807e-01 1.23185241e+00
4.97115076e-01 -4.89051670e-01 8.25508595e-01 3.95977259e-01
-4.94988382e-01 -7.48525918e-01 -7.03337312e-01 -7.19925404e-01
1.27207279e-01 -3.61800581e-01 5.98053634e-01 8.67060244e-01
-1.37572393e-01 2.45356292e-01 -1.92847271e-02 1.11795997e-03
5.18877327e-01 8.97346288e-02 4.87665772e-01 -1.20244515e+00
1.30610630e-01 -2.94440657e-01 -6.31278396e-01 -9.69694555e-01
2.63819218e-01 -9.95158076e-01 5.16908705e-01 -2.02542973e+00
3.39049309e-01 -3.55019599e-01 -6.89498067e-01 6.91515088e-01
-4.42496032e-01 5.96313655e-01 2.42281288e-01 -1.24770716e-01
-1.24571216e+00 7.48994708e-01 1.42209852e+00 -2.15979800e-01
-1.72931999e-01 -3.20711046e-01 -8.07701111e-01 7.05885649e-01
1.02623212e+00 -5.67421794e-01 -6.00247085e-01 -5.80606997e-01
-1.92628518e-01 -3.81385446e-01 3.75751019e-01 -9.34476078e-01
1.48764849e-01 -2.58258522e-01 -1.11920394e-01 -4.84072596e-01
3.34358007e-01 -4.88673747e-01 -4.56086814e-01 2.03307435e-01
-3.31295341e-01 -5.83788097e-01 -6.05634823e-02 6.72214389e-01
-3.06465238e-01 -4.15313810e-01 7.81951070e-01 -2.95750260e-01
-1.40425837e+00 5.95967412e-01 -2.30655462e-01 5.65625012e-01
1.21803117e+00 -1.62987024e-01 -3.12451422e-01 1.57607064e-01
-9.98108625e-01 6.77170515e-01 7.17316508e-01 8.00045431e-01
5.21403193e-01 -1.16045475e+00 -3.46567690e-01 8.58570188e-02
6.05593562e-01 5.27242601e-01 5.04105270e-01 5.88787854e-01
-3.93271744e-01 2.25591615e-01 -1.20209470e-01 -5.82456350e-01
-9.91345406e-01 4.53332603e-01 5.47920493e-03 1.35724768e-01
-6.94810987e-01 1.11558723e+00 4.62052643e-01 -4.27593559e-01
3.10264956e-02 -4.28165585e-01 -2.57150501e-01 2.16785699e-01
2.83957511e-01 1.01590239e-01 -2.46545866e-01 -6.85274243e-01
-2.61968642e-01 7.78402865e-01 -2.07842104e-02 6.38381764e-02
1.29537082e+00 -2.16540769e-01 1.24320291e-01 4.89918798e-01
1.35768116e+00 -4.55573052e-01 -1.56052315e+00 -3.88366103e-01
1.21683009e-01 -5.17064333e-01 1.68946967e-01 -5.98206282e-01
-1.39627540e+00 1.01189840e+00 4.94396359e-01 -1.40497342e-01
9.42408144e-01 4.16314393e-01 1.24955964e+00 1.22468481e-02
2.09352449e-01 -1.38133252e+00 1.80008784e-01 5.69513142e-01
3.50098312e-01 -1.40797365e+00 -3.38859200e-01 -6.77254856e-01
-7.54536569e-01 5.63628435e-01 8.13632905e-01 -1.96703255e-01
6.77805543e-01 -1.91886678e-01 1.40576571e-01 -2.04762608e-01
-3.06656599e-01 -8.67297232e-01 1.96247846e-01 5.57528257e-01
2.02613264e-01 3.76834393e-01 -3.95501643e-01 9.07423377e-01
6.79201558e-02 9.62994322e-02 4.33285654e-01 9.41052198e-01
-8.71975780e-01 -1.09138989e+00 1.37413576e-01 4.25615847e-01
-2.42927223e-01 -1.17142506e-01 -2.16116071e-01 3.61729801e-01
3.14211726e-01 8.84971142e-01 1.78820759e-01 -2.88499624e-01
1.95517689e-01 7.48381913e-02 1.15615912e-02 -1.21004903e+00
-3.64640318e-02 -3.46356370e-02 -2.59337157e-01 -6.32970452e-01
-3.05764556e-01 -4.46074545e-01 -1.71454394e+00 3.37195545e-01
-1.07816860e-01 8.35347101e-02 6.24269009e-01 9.37460721e-01
4.21338916e-01 8.23530316e-01 2.97170222e-01 -8.46084833e-01
7.45111704e-02 -8.56937170e-01 -5.53259790e-01 7.19262362e-01
3.67428698e-02 -7.17412114e-01 -2.37848848e-01 1.48168385e-01] | [9.65347957611084, 1.0164978504180908] |
2ed9b638-8cc2-462b-9403-7ca8e66adfff | contrastive-hierarchical-discourse-graph-for | 2306.00177 | null | https://arxiv.org/abs/2306.00177v1 | https://arxiv.org/pdf/2306.00177v1.pdf | Contrastive Hierarchical Discourse Graph for Scientific Document Summarization | The extended structural context has made scientific paper summarization a challenging task. This paper proposes CHANGES, a contrastive hierarchical graph neural network for extractive scientific paper summarization. CHANGES represents a scientific paper with a hierarchical discourse graph and learns effective sentence representations with dedicated designed hierarchical graph information aggregation. We also propose a graph contrastive learning module to learn global theme-aware sentence representations. Extensive experiments on the PubMed and arXiv benchmark datasets prove the effectiveness of CHANGES and the importance of capturing hierarchical structure information in modeling scientific papers. | ['Jiawei Zhang', 'Xiao Liu', 'Haopeng Zhang'] | 2023-05-31 | null | null | null | null | ['scientific-article-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.92264664e-01 1.11837530e+00 -4.69522178e-01 -8.18527117e-02
-6.14532590e-01 -3.94185543e-01 6.20177329e-01 9.92938280e-01
1.13771237e-01 8.24744761e-01 1.31823516e+00 -2.45086372e-01
-2.49633998e-01 -6.95356250e-01 -1.07632565e+00 -2.10285991e-01
-2.71155417e-01 4.33710039e-01 -1.99109942e-01 5.65411709e-03
7.30998576e-01 3.07341307e-01 -9.18546021e-01 5.14829338e-01
1.24963403e+00 2.66054362e-01 1.60266936e-01 1.14245558e+00
-6.19195819e-01 1.43262982e+00 -1.41341341e+00 -3.54313046e-01
-3.13549697e-01 -5.99196076e-01 -1.16672647e+00 3.30273174e-02
1.03793657e+00 1.37862608e-01 -8.34670603e-01 8.47204924e-01
6.33385241e-01 3.07482213e-01 9.73944902e-01 -7.76202619e-01
-1.16296244e+00 1.53250706e+00 -4.65186238e-01 5.50311744e-01
6.53634131e-01 -3.18596095e-01 1.43614590e+00 -1.84507772e-01
1.19741488e+00 1.43863189e+00 5.75674415e-01 4.64751542e-01
-9.57696319e-01 -1.55184180e-01 4.29798931e-01 3.07708472e-01
-6.77254736e-01 -1.38582468e-01 1.22797060e+00 -3.44290406e-01
1.39177155e+00 6.79661393e-01 7.68675506e-01 1.37895584e+00
8.34657788e-01 9.07981813e-01 3.62855464e-01 -1.28097281e-01
2.08082646e-01 -6.23786211e-01 9.09517288e-01 7.68461645e-01
9.29810464e-01 -8.55228841e-01 -6.59515798e-01 -2.89139628e-01
1.08145230e-01 -1.62567616e-01 -3.81591141e-01 1.07101329e-01
-1.03078675e+00 7.40537941e-01 8.19600046e-01 3.74916494e-01
-4.45582241e-01 2.05659851e-01 9.47024107e-01 1.76983684e-01
6.96620524e-01 1.26413286e+00 1.95533693e-01 4.64741766e-01
-9.76858854e-01 3.51095736e-01 1.00500500e+00 1.04173625e+00
2.08022937e-01 4.72776592e-01 -1.14310598e+00 5.35997272e-01
-2.65246630e-01 2.59732474e-02 3.48246515e-01 -9.15573418e-01
7.13233948e-01 1.16897225e+00 -4.76177007e-01 -1.36451709e+00
-7.61901259e-01 -8.90108228e-01 -1.36994874e+00 -7.10471809e-01
-5.70945382e-01 2.13635519e-01 -7.19841242e-01 1.18804002e+00
-1.16069332e-01 6.65810108e-02 3.01400274e-01 3.41121674e-01
2.25982308e+00 9.24561918e-01 1.86648190e-01 -5.48200548e-01
1.13214159e+00 -1.07298744e+00 -1.30059290e+00 8.17413479e-02
5.80283284e-01 -1.21611193e-01 7.68634379e-01 6.93110004e-02
-1.35855746e+00 -2.92723060e-01 -1.31809211e+00 -7.35842943e-01
-3.78760189e-01 -1.37477294e-01 6.24798656e-01 -1.70998305e-01
-1.23527730e+00 8.66914868e-01 -3.38859349e-01 -2.75271833e-01
8.77108634e-01 1.70722663e-01 -2.05009520e-01 1.73134312e-01
-1.00353444e+00 7.10106134e-01 6.56059384e-01 -2.07335860e-01
-3.74247968e-01 -9.58987892e-01 -1.00557709e+00 6.41681790e-01
1.72398463e-01 -1.42001271e+00 8.12573075e-01 -3.22555542e-01
-1.28428507e+00 6.29283309e-01 -2.64836669e-01 -8.06016982e-01
1.72287017e-01 -1.75312627e-02 -2.68730000e-02 5.29512525e-01
1.62580043e-01 3.26374829e-01 4.81455028e-01 -1.25683916e+00
1.40358716e-01 -4.99599457e-01 -9.01506841e-02 4.16343093e-01
-4.92222637e-01 -2.21498698e-01 -2.26364613e-01 -9.54434037e-01
-3.87027860e-01 -2.76232392e-01 -3.31180364e-01 -1.02071798e+00
-1.25317085e+00 -9.86542642e-01 4.16096091e-01 -1.08494294e+00
1.50804710e+00 -1.25575113e+00 8.99956405e-01 -3.05806220e-01
1.07336295e+00 -1.39446437e-01 -3.30997914e-01 6.27673090e-01
7.28942901e-02 4.24486369e-01 -1.68957815e-01 -3.41073334e-01
-7.73947537e-02 -2.01647937e-01 -5.40422082e-01 -2.74675656e-02
1.96994931e-01 1.53204942e+00 -1.12402999e+00 -6.78174555e-01
-2.53205031e-01 1.86512694e-01 -3.10317338e-01 3.90614986e-01
-4.92367893e-01 4.47583944e-02 -6.92143559e-01 3.34353864e-01
5.78836024e-01 -5.69886744e-01 4.69816208e-01 -4.39595520e-01
1.93707317e-01 4.74489063e-01 -2.60723889e-01 1.53179610e+00
1.56621963e-01 6.97044611e-01 -2.30887190e-01 -1.19834590e+00
8.68467033e-01 -1.26190186e-02 2.85414070e-01 -7.27165937e-01
3.49083245e-02 -3.48672509e-01 -4.42967385e-01 -4.92518753e-01
9.60990846e-01 6.17061615e-01 -3.43438357e-01 2.65467644e-01
2.72291034e-01 -3.47819775e-01 5.19100726e-01 1.14017439e+00
1.60010850e+00 -1.48963094e-01 5.02064347e-01 -5.97153127e-01
3.99416298e-01 -7.33801872e-02 3.10646355e-01 1.22340214e+00
3.04023713e-01 5.66316366e-01 1.12289226e+00 -4.70614076e-01
-8.07981908e-01 -8.53359163e-01 4.65834349e-01 9.88408744e-01
-1.57455415e-01 -1.03432024e+00 -7.16803610e-01 -8.69502902e-01
1.83525190e-01 1.08306456e+00 -9.36486244e-01 -5.23970008e-01
-7.16884613e-01 -5.05494654e-01 4.91025716e-01 3.15533012e-01
1.51078358e-01 -1.20469713e+00 -6.12592958e-02 -1.38602802e-03
-1.33331478e-01 -8.66377890e-01 -7.02903867e-01 -3.13470997e-02
-1.06656766e+00 -9.98367369e-01 -5.29889166e-01 -9.67796385e-01
6.90102279e-01 3.90954524e-01 1.53289330e+00 -5.60232066e-02
-3.74262273e-01 4.85927731e-01 -2.57118195e-01 -5.24043441e-01
-7.01114237e-01 6.09541774e-01 -3.29402983e-01 -6.43184781e-01
-1.57727569e-01 -4.65954781e-01 -4.27238464e-01 -1.00617933e+00
-8.81254911e-01 2.76827663e-01 5.09556711e-01 6.75200045e-01
3.59545887e-01 -3.99860382e-01 9.54770148e-01 -1.35354805e+00
1.49433899e+00 -4.45635378e-01 -1.33194655e-01 5.73813498e-01
-4.68346059e-01 4.93564457e-01 7.89120138e-01 9.28951129e-02
-8.17668796e-01 -6.90462232e-01 3.24147195e-01 -3.31401646e-01
3.47333729e-01 9.50972974e-01 -1.20261654e-01 2.92481661e-01
6.24069095e-01 4.31781411e-01 -1.59670040e-01 -3.07726324e-01
7.06973910e-01 3.19113374e-01 6.20418847e-01 -3.03627193e-01
3.54540020e-01 4.87589613e-02 3.11910450e-01 -1.07065058e+00
-1.15169120e+00 -2.55897313e-01 -4.33028579e-01 -1.01502948e-01
9.51001227e-01 -7.21158743e-01 -9.07276154e-01 -1.31855637e-01
-1.57201886e+00 2.10160732e-01 -4.86204803e-01 -2.15488747e-01
-2.80187815e-01 9.04874742e-01 -8.26225638e-01 -3.38263988e-01
-1.40226507e+00 -6.00265861e-01 1.08867860e+00 3.53098214e-01
-4.43477422e-01 -1.15216112e+00 2.32831404e-01 6.11979127e-01
1.15147710e-01 7.75951087e-01 1.15877986e+00 -7.36607790e-01
-4.73851204e-01 -4.76492383e-02 -3.52345526e-01 -1.29974976e-01
2.36298263e-01 -6.08000830e-02 -5.61638236e-01 -3.33530486e-01
-3.74264657e-01 -1.81139499e-01 1.69247890e+00 7.05884218e-01
1.65069270e+00 -1.13575816e+00 -4.09215093e-01 5.53574026e-01
9.29258347e-01 -3.07431757e-01 4.35195625e-01 1.85571775e-01
1.47973132e+00 5.82581937e-01 -3.91184479e-01 3.81958038e-01
7.46494651e-01 3.58046009e-03 2.44919762e-01 2.51262505e-02
-5.77115893e-01 -2.54561931e-01 8.23196247e-02 1.73984122e+00
-4.53859009e-02 -7.22383916e-01 -6.58444762e-01 3.79597038e-01
-1.86629736e+00 -1.21809793e+00 -2.97822505e-01 1.42884529e+00
1.06460297e+00 1.16573845e-03 -1.16865687e-01 -4.01736319e-01
5.84982514e-01 6.34875059e-01 -5.62551200e-01 -8.75241041e-01
-7.29858637e-01 1.57654583e-01 4.40318108e-01 5.07532835e-01
-1.06437039e+00 9.28986251e-01 6.46994829e+00 6.24148369e-01
-6.78776085e-01 -2.54441798e-01 6.02074385e-01 -8.64568502e-02
-9.07782018e-01 -1.85194269e-01 -2.93898880e-01 2.31270015e-01
8.61878455e-01 -1.04160058e+00 -1.04742542e-01 4.02356565e-01
7.12698400e-02 4.22678441e-01 -1.05747676e+00 7.98038602e-01
6.65399551e-01 -2.48491383e+00 1.26151061e+00 -2.74203628e-01
7.33938932e-01 -9.23944190e-02 -3.48852158e-01 4.97547984e-01
4.77057755e-01 -1.13784814e+00 1.79410279e-01 8.95391107e-01
3.09016287e-01 -7.22135544e-01 6.46950781e-01 -4.89103347e-02
-7.42531717e-01 1.54050037e-01 -5.99203169e-01 8.93930066e-03
-2.41313115e-01 6.40931368e-01 -1.02774882e+00 1.33623493e+00
3.75084519e-01 1.42381430e+00 -1.11356866e+00 7.68075526e-01
-2.04615980e-01 8.75563860e-01 4.78090495e-01 -5.01216948e-01
9.71890986e-02 -2.03466237e-01 1.11414468e+00 1.85282159e+00
-1.84941068e-02 6.71194820e-03 2.34544054e-01 9.48602259e-01
-9.58782911e-01 3.26154500e-01 -8.87460411e-01 -7.25522995e-01
1.93366781e-01 1.24246657e+00 -5.88547349e-01 -5.12057304e-01
9.83653963e-02 8.23478401e-01 8.55736136e-01 4.53097790e-01
-9.91066769e-02 -6.22635186e-01 -1.86903074e-01 -2.14131758e-01
1.18528143e-01 -4.12303954e-02 -6.17338538e-01 -1.35819435e+00
-1.79340839e-01 -6.04911685e-01 9.14453387e-01 -7.78497458e-01
-1.32208192e+00 5.33620834e-01 -1.62261114e-01 -7.02629745e-01
9.59052667e-02 -1.73722237e-01 -1.07709074e+00 3.32756042e-01
-1.18973398e+00 -1.35183418e+00 -1.97452322e-01 -9.86873060e-02
7.68098712e-01 -4.26414549e-01 8.24873686e-01 -5.83901048e-01
-9.03911293e-01 6.33778930e-01 1.69368193e-01 -3.98966223e-02
3.25071841e-01 -1.66569853e+00 6.33224547e-01 5.07245302e-01
-1.35372773e-01 7.78772116e-01 1.03042006e+00 -1.24495780e+00
-1.97723937e+00 -1.44529045e+00 9.53501344e-01 -5.11704981e-01
6.48618460e-01 -4.00503814e-01 -1.26747918e+00 6.74481988e-01
1.23906565e+00 -7.44777262e-01 7.54793525e-01 2.25787789e-01
-4.02570426e-01 5.58970198e-02 -6.89636528e-01 7.36195922e-01
1.28209865e+00 -2.01122209e-01 -1.16293538e+00 8.33596468e-01
1.42162061e+00 -2.50166297e-01 -1.07252932e+00 4.40952420e-01
-1.09890692e-01 -1.79737732e-01 7.71042764e-01 -1.03906202e+00
9.89861548e-01 2.03375086e-01 5.15869737e-01 -1.48876297e+00
-8.91317666e-01 -8.21535528e-01 -9.82857943e-01 1.23014438e+00
1.45156503e-01 -2.87120849e-01 6.79849744e-01 -9.38890874e-02
-5.38228333e-01 -6.60084069e-01 -7.27151871e-01 -4.45989460e-01
2.81786978e-01 5.24479330e-01 2.19096839e-01 1.11509192e+00
4.24826473e-01 1.23019040e+00 -1.58858046e-01 -2.23519832e-01
8.99586022e-01 4.49217200e-01 5.57109833e-01 -1.57461667e+00
1.51774049e-01 -9.82363760e-01 -4.12869036e-01 -4.02184695e-01
8.86049151e-01 -1.66471350e+00 -6.65442526e-01 -2.84120893e+00
7.68894255e-01 7.54636049e-01 -3.40441704e-01 2.29865566e-01
-7.59117186e-01 -6.21592462e-01 2.35035317e-03 7.23642036e-02
-1.30688179e+00 9.29417551e-01 1.34201372e+00 -1.04349291e+00
-5.05552925e-02 -4.95613307e-01 -1.23375142e+00 1.32205471e-01
5.35063207e-01 -2.06518576e-01 -4.53278095e-01 -3.38164002e-01
4.26214814e-01 6.08878545e-02 -2.07371917e-02 -4.87464637e-01
5.42152286e-01 -1.67860501e-02 4.92394656e-01 -9.10086095e-01
-2.48788178e-01 2.99077839e-01 -3.27880412e-01 5.05491078e-01
-1.12066567e+00 -2.13124999e-03 3.75288516e-01 8.36540937e-01
-9.65587944e-02 2.08508253e-01 1.73279628e-01 -3.22836727e-01
-2.29891449e-01 1.73374876e-01 -4.40083146e-01 4.23517078e-01
4.08548146e-01 2.71278806e-02 -9.17210877e-01 -3.96865696e-01
-3.93074781e-01 6.16875708e-01 9.89530608e-02 7.03256845e-01
1.00743699e+00 -1.08790159e+00 -1.32010186e+00 -4.70219582e-01
1.48859099e-01 8.36551655e-03 3.95777971e-01 2.73561835e-01
-6.02767050e-01 4.71250087e-01 -9.99253988e-02 -2.11785153e-01
-1.23988962e+00 7.04685807e-01 -1.83735713e-01 -6.05522752e-01
-1.11638284e+00 7.31278062e-01 3.78255248e-01 -3.86050642e-01
2.56174624e-01 -4.66542035e-01 -9.72707331e-01 2.79213548e-01
4.94359761e-01 6.02235854e-01 1.20522991e-01 -1.88357726e-01
-1.13544039e-01 1.51598886e-01 -5.88461995e-01 7.73026824e-01
1.61806273e+00 1.99554235e-01 -8.90397727e-01 5.16254067e-01
1.15002072e+00 2.03496478e-02 -2.88261354e-01 -6.11605197e-02
4.44649994e-01 4.18451786e-01 1.68683857e-01 -6.47865891e-01
-5.70335686e-01 3.80053818e-01 -2.71340579e-01 4.32827055e-01
6.59613669e-01 2.02853277e-01 6.48400962e-01 9.05204773e-01
-4.07542288e-01 -1.11398423e+00 2.68032908e-01 7.27443397e-01
1.59864926e+00 -8.02541971e-01 7.03503549e-01 -3.41518730e-01
-5.48405588e-01 1.28793037e+00 4.47167724e-01 -3.81156385e-01
-1.58112533e-02 -7.55768865e-02 -5.33203840e-01 -7.44292021e-01
-9.50828016e-01 3.95424694e-01 9.14468825e-01 3.47843975e-01
6.09973729e-01 3.09663117e-01 -5.47109485e-01 8.73572707e-01
-3.96498799e-01 -3.21886957e-01 1.14915872e+00 6.92471743e-01
-4.25492644e-01 -3.08752239e-01 1.45083919e-01 1.03969145e+00
-3.85580868e-01 -1.11080661e-01 -1.26573396e+00 2.29649201e-01
-8.26926768e-01 7.82393515e-01 -1.94855537e-02 -9.19409990e-02
4.41179782e-01 -9.23207253e-02 5.27936995e-01 -8.20036352e-01
-1.00250709e+00 -2.74016857e-01 2.90429711e-01 -3.48655641e-01
-2.46554524e-01 -2.96468377e-01 -1.35287499e+00 -1.92374304e-01
2.49719799e-01 6.02159739e-01 2.60812700e-01 4.51073706e-01
1.20011687e+00 1.55620456e+00 3.87118399e-01 -8.23885024e-01
-4.01453763e-01 -1.19539046e+00 -3.60178649e-01 5.73885858e-01
4.85074729e-01 -6.73095211e-02 -3.16525877e-01 -1.21862575e-01] | [12.592289924621582, 9.551342964172363] |
ecc49ef7-8b42-4313-bb8a-f5a4a0e5a881 | hand-object-interaction-and-precise | 1511.03814 | null | http://arxiv.org/abs/1511.03814v2 | http://arxiv.org/pdf/1511.03814v2.pdf | Hand-Object Interaction and Precise Localization in Transitive Action Recognition | Action recognition in still images has seen major improvement in recent years
due to advances in human pose estimation, object recognition and stronger
feature representations produced by deep neural networks. However, there are
still many cases in which performance remains far from that of humans. A major
difficulty arises in distinguishing between transitive actions in which the
overall actor pose is similar, and recognition therefore depends on details of
the grasp and the object, which may be largely occluded. In this paper we
demonstrate how recognition is improved by obtaining precise localization of
the action-object and consequently extracting details of the object shape
together with the actor-object interaction. To obtain exact localization of the
action object and its interaction with the actor, we employ a coarse-to-fine
approach which combines semantic segmentation and contextual features, in
successive stages. We focus on (but are not limited) to face-related actions, a
set of actions that includes several currently challenging categories. We
present an average relative improvement of 35% over state-of-the art and
validate through experimentation the effectiveness of our approach. | ['Shimon Ullman', 'Amir Rosenfeld'] | 2015-11-12 | null | null | null | null | ['action-recognition-in-still-images'] | ['computer-vision'] | [ 5.82788348e-01 7.40210572e-03 -8.22751038e-03 -2.93611258e-01
-5.95031977e-01 -5.04716754e-01 7.17423201e-01 -6.64918199e-02
-4.58131850e-01 5.42835951e-01 3.57136965e-01 4.36145306e-01
-1.47357881e-01 -2.43155986e-01 -6.00939929e-01 -6.07713461e-01
-2.10912600e-02 7.39605069e-01 4.91978705e-01 5.72754256e-02
2.90549308e-01 9.79940057e-01 -1.60251176e+00 5.70367575e-01
3.34121078e-01 1.04643881e+00 -1.28039867e-01 6.78657889e-01
1.66577980e-01 6.22393429e-01 -7.43659914e-01 -1.92588896e-01
2.45689720e-01 -4.58489627e-01 -1.09934115e+00 6.43593609e-01
6.80397749e-01 -5.39447546e-01 -5.96972518e-02 8.04796219e-01
2.15195447e-01 1.03459843e-01 8.29141498e-01 -1.12129581e+00
-1.51808023e-01 2.28064477e-01 -7.78576970e-01 3.56814228e-02
6.83062613e-01 9.38381851e-02 8.57432067e-01 -7.71140039e-01
8.76020133e-01 1.53371751e+00 5.00805140e-01 5.36121428e-01
-1.15682912e+00 -2.94928879e-01 2.01866418e-01 2.34481514e-01
-1.26145864e+00 -5.21014750e-01 7.17006564e-01 -4.78021830e-01
1.00870597e+00 8.57917070e-02 5.39449275e-01 1.02615774e+00
8.90792236e-02 1.06190836e+00 1.10663521e+00 -5.69536686e-01
1.82990089e-01 -1.88979968e-01 -5.08474894e-02 5.74625134e-01
9.54027846e-03 5.56240678e-02 -3.51066232e-01 4.45199050e-02
9.51154709e-01 3.96886617e-02 -1.94745567e-02 -7.54371047e-01
-1.15210783e+00 3.61056596e-01 3.99845600e-01 5.71419358e-01
-7.24412024e-01 4.33131725e-01 2.80971259e-01 -2.73299307e-01
8.02380443e-02 2.28993297e-01 -4.56562936e-01 -3.86399090e-01
-9.77239728e-01 3.57437670e-01 8.33489537e-01 7.49041557e-01
4.96242046e-01 -1.78328246e-01 -1.74862906e-01 5.18742740e-01
1.41300708e-01 2.68756986e-01 -1.99417751e-02 -1.04952264e+00
3.17593962e-01 7.24935055e-01 2.29671955e-01 -9.39176500e-01
-3.87747347e-01 -1.36794329e-01 -4.00266260e-01 3.99065912e-01
9.62053835e-01 9.69474763e-02 -1.06256652e+00 1.37767243e+00
6.15577042e-01 -1.52682155e-01 -1.11670010e-01 9.02664185e-01
4.27304208e-01 1.49836510e-01 3.43194216e-01 -3.58343869e-02
1.45840335e+00 -9.21216130e-01 -5.53577840e-01 -4.24798667e-01
4.53695148e-01 -7.22550452e-01 6.56376600e-01 4.82778430e-01
-1.10459566e+00 -6.35725558e-01 -8.03106487e-01 -1.37040997e-02
-3.04956734e-01 4.93256450e-01 6.19507432e-01 4.95048851e-01
-7.03779221e-01 8.97247672e-01 -8.55760694e-01 -7.71955788e-01
7.52918482e-01 6.77571058e-01 -7.50746965e-01 -1.62834320e-02
-7.38254964e-01 1.03525662e+00 4.22857225e-01 1.50011554e-01
-9.12726581e-01 -2.37772435e-01 -8.32791448e-01 -2.63462868e-02
8.24159980e-01 -3.22820514e-01 1.21588647e+00 -1.34004569e+00
-1.43334472e+00 9.13040578e-01 -5.49475700e-02 -3.60592335e-01
7.05616891e-01 -5.36102355e-01 -9.85168070e-02 3.99489820e-01
-9.59896073e-02 8.43238771e-01 1.13877857e+00 -1.21161926e+00
-7.10725188e-01 -7.18475640e-01 2.87114382e-01 1.47844508e-01
1.33809388e-01 4.27752674e-01 -5.50506294e-01 -4.97876763e-01
1.57628044e-01 -1.09601021e+00 -6.60874695e-02 2.69984066e-01
-2.21215457e-01 -4.09938633e-01 9.51481700e-01 -6.65233910e-01
7.32189119e-01 -2.01189971e+00 2.70456553e-01 -1.14075989e-02
-7.05351457e-02 4.50897664e-01 -1.54679283e-01 5.21807849e-01
-2.26427719e-01 -2.21532688e-01 -6.84304163e-02 -2.82600671e-01
-1.91341471e-02 1.81784093e-01 -3.49865556e-02 8.02451670e-01
3.57801974e-01 9.92192745e-01 -6.48561954e-01 -5.51177740e-01
3.66384089e-01 4.61355567e-01 -2.88466901e-01 1.61924094e-01
-2.29891926e-01 6.14775181e-01 -7.12434113e-01 9.05947149e-01
4.04769897e-01 -2.41638417e-03 2.01276287e-01 -3.38535130e-01
1.60850182e-01 -5.70368096e-02 -1.44201279e+00 1.66744268e+00
-7.36818165e-02 3.47298831e-01 2.57350296e-01 -9.97269094e-01
6.45170748e-01 2.64453739e-01 7.41221189e-01 -3.78500372e-01
3.67412597e-01 8.14175904e-02 2.04646483e-01 -5.14633715e-01
2.86289841e-01 -7.18684793e-02 1.54826436e-02 4.66238111e-01
1.39809862e-01 2.18059383e-02 2.87880242e-01 -1.24893971e-01
8.80258381e-01 6.90681815e-01 7.41180658e-01 -1.16971649e-01
5.86764991e-01 -3.31916921e-02 1.82540238e-01 5.27409434e-01
-5.28828621e-01 5.29315531e-01 5.86358488e-01 -4.93188381e-01
-7.19896734e-01 -7.92846382e-01 2.30216887e-02 1.08410251e+00
1.23276010e-01 -1.55895084e-01 -1.00674450e+00 -1.03133774e+00
1.73586626e-02 4.73715276e-01 -8.52420092e-01 2.34904289e-02
-8.58954966e-01 -2.10106477e-01 2.86682099e-01 1.03253412e+00
4.48556304e-01 -1.38179195e+00 -9.84405458e-01 1.70023069e-01
1.58389270e-01 -1.33612072e+00 -3.11886162e-01 2.26375600e-03
-8.91360939e-01 -1.36097443e+00 -7.63317645e-01 -3.82823408e-01
7.46103346e-01 -3.56095694e-02 9.97994542e-01 2.59204768e-02
-6.31616414e-01 6.78275943e-01 -4.15711313e-01 -3.85559797e-01
-2.74797559e-01 -2.23536700e-01 -3.71379927e-02 2.56889939e-01
4.12634701e-01 -2.90362448e-01 -4.85182732e-01 5.12998462e-01
-7.62859046e-01 -2.03166887e-01 8.32798839e-01 4.71545905e-01
4.03933942e-01 2.05531027e-02 6.12194352e-02 -5.99124014e-01
7.22137541e-02 1.25775382e-01 -2.38363624e-01 2.78956592e-01
3.43865976e-02 2.22661328e-02 1.37628913e-01 -4.62706536e-01
-1.02854908e+00 7.43177593e-01 3.59824374e-02 -3.40414762e-01
-9.01142061e-01 -1.57860041e-01 -3.94819200e-01 -1.15805410e-01
4.83207822e-01 -1.16531223e-01 2.05136091e-03 -5.21738887e-01
3.33209723e-01 5.22180855e-01 2.65795290e-01 -5.42429328e-01
5.17887414e-01 7.75493503e-01 2.70882607e-01 -8.63662064e-01
-8.57745826e-01 -5.11133254e-01 -1.29622030e+00 -4.62770820e-01
1.07541311e+00 -4.76411372e-01 -9.11548495e-01 5.41820586e-01
-1.24285293e+00 -2.75351480e-02 -4.03192222e-01 4.92492884e-01
-7.49446630e-01 4.07176644e-01 -4.13440704e-01 -9.06956971e-01
4.39998247e-02 -1.27084887e+00 1.63246238e+00 5.11090606e-02
-4.65886712e-01 -5.97607851e-01 -3.34035248e-01 5.85796177e-01
1.29914477e-01 3.76567662e-01 5.41734934e-01 -8.22238445e-01
-6.56628132e-01 -4.76792723e-01 -2.10867375e-01 3.81245285e-01
1.79220214e-01 -4.83597182e-02 -8.89638245e-01 -2.41258174e-01
-1.45509928e-01 -4.33505833e-01 7.41405964e-01 4.74629134e-01
9.58852708e-01 8.13517421e-02 -5.64841509e-01 1.73297264e-02
1.08991647e+00 3.22944850e-01 7.06385016e-01 5.71894720e-02
6.97667718e-01 1.05491960e+00 8.96523714e-01 2.38793850e-01
-1.74004585e-01 1.02490222e+00 4.78429109e-01 9.29015409e-03
-2.45384440e-01 2.43029781e-02 3.17098916e-01 -8.13992545e-02
-5.15182614e-01 -4.25617322e-02 -8.34376037e-01 4.59637702e-01
-1.73502123e+00 -9.28587437e-01 9.40794963e-03 2.03160477e+00
4.11926955e-01 1.77735284e-01 4.96814460e-01 2.70617098e-01
6.12100303e-01 1.43805653e-01 -4.81696099e-01 -2.03162819e-01
3.99292350e-01 2.19683617e-01 1.95045382e-01 3.22738945e-01
-1.48061264e+00 1.08933997e+00 6.68612003e+00 8.73128533e-01
-8.57772470e-01 -1.31665871e-01 3.39552552e-01 5.19887023e-02
5.89599252e-01 -2.58383155e-02 -8.65591168e-01 -3.22162174e-02
4.00987387e-01 3.77678782e-01 1.48835719e-01 1.07756460e+00
-2.93560307e-02 -4.75119889e-01 -1.54380810e+00 8.00288320e-01
3.00514668e-01 -7.76298642e-01 -9.63604301e-02 6.61693215e-02
5.43522120e-01 -6.04566514e-01 -3.21535587e-01 1.12780727e-01
-1.16353989e-01 -1.07083368e+00 7.40885973e-01 5.00486016e-01
5.60166955e-01 -6.90631390e-01 6.48402631e-01 2.47503847e-01
-1.33905303e+00 -1.38882667e-01 -1.05312273e-01 -1.07159898e-01
1.78923551e-02 1.18947335e-01 -7.37944603e-01 4.51017231e-01
5.97731471e-01 6.64662123e-01 -4.75359499e-01 6.81808054e-01
-4.25175518e-01 2.05432490e-01 -1.77468538e-01 -9.21284128e-03
2.24153712e-01 -5.83692826e-02 5.87619901e-01 1.13441026e+00
-1.17888130e-01 2.51116902e-01 4.42889959e-01 7.30402768e-01
2.42090166e-01 3.63331214e-02 -4.42129940e-01 -2.85086930e-01
-1.55847490e-01 1.15683913e+00 -1.03294754e+00 -2.34299481e-01
-2.78906465e-01 1.18827605e+00 1.87071040e-01 2.82250792e-01
-8.64783585e-01 -1.89103350e-01 4.87352431e-01 2.36962423e-01
7.49393940e-01 -3.21997374e-01 1.67825937e-01 -7.19701886e-01
2.89748847e-01 -9.54162002e-01 2.89326519e-01 -7.13469565e-01
-8.26391459e-01 4.86721665e-01 3.07162434e-01 -1.11247981e+00
-4.04289514e-01 -8.26204300e-01 -2.68160492e-01 5.75520754e-01
-8.50070000e-01 -1.57240081e+00 -2.78759390e-01 4.57100362e-01
6.82038605e-01 1.21573001e-01 7.41115868e-01 2.50844043e-02
-2.53640324e-01 2.86306679e-01 -2.85763413e-01 2.39092037e-01
4.44464028e-01 -9.89283383e-01 4.18371744e-02 6.04111314e-01
4.41891611e-01 4.16591853e-01 7.18066096e-01 -6.75367951e-01
-1.25838697e+00 -7.29626775e-01 7.81202853e-01 -7.62175322e-01
5.48827767e-01 -4.63563889e-01 -6.03116870e-01 7.36784279e-01
-7.91261420e-02 1.03263699e-01 1.55793414e-01 3.60591859e-02
-1.72187254e-01 -7.30908588e-02 -1.10776889e+00 4.72675860e-01
1.17870736e+00 -4.09921676e-01 -8.66643429e-01 3.47760767e-01
1.67870224e-02 -3.66726547e-01 -5.79838753e-01 4.94218230e-01
8.08410347e-01 -1.18250155e+00 1.06851661e+00 -8.78969193e-01
3.35710138e-01 -2.58319259e-01 -1.22251764e-01 -8.49627614e-01
-1.47616059e-01 -2.57367343e-01 -1.11444637e-01 1.01952112e+00
-1.27137795e-01 -2.11458564e-01 1.05528450e+00 5.25289416e-01
3.43601525e-01 -7.91576922e-01 -9.02641356e-01 -8.18027914e-01
-3.07616025e-01 -3.47551614e-01 1.52962908e-01 3.92456084e-01
-1.80883348e-01 2.29746252e-01 -2.93414354e-01 -5.51241934e-02
5.11189997e-01 3.35969955e-01 7.87964165e-01 -1.29289198e+00
-3.15758377e-01 -3.35778415e-01 -8.14505637e-01 -1.06780422e+00
2.37388909e-01 -3.11936647e-01 1.17973380e-01 -1.44596446e+00
3.09368461e-01 2.11789161e-01 -8.84437039e-02 5.72388470e-01
-2.10589040e-02 3.93937677e-01 2.29780510e-01 1.23081543e-01
-7.85525382e-01 3.69027138e-01 1.41514337e+00 -3.31697278e-02
5.87347569e-03 4.01195809e-02 -1.94196880e-01 1.07968998e+00
4.14315641e-01 -4.37608242e-01 -6.57428131e-02 -8.91681910e-02
-3.76215398e-01 1.20401621e-01 5.99080086e-01 -1.27994752e+00
-3.38236094e-02 -1.47627890e-01 7.46335804e-01 -6.37726843e-01
7.02515304e-01 -1.20376945e+00 2.08804533e-01 6.37400329e-01
-3.54026943e-01 -3.41638327e-01 1.78192869e-01 6.01433992e-01
-1.04236513e-01 -3.15718234e-01 9.35882449e-01 -3.23112786e-01
-8.07174385e-01 1.31515339e-01 -2.96970576e-01 -2.48405755e-01
1.45142651e+00 -4.87224221e-01 1.98449776e-01 -3.15948039e-01
-1.23610914e+00 -2.95475721e-01 4.20432985e-01 4.54235435e-01
3.53037804e-01 -1.08003104e+00 -5.81836104e-01 7.63846263e-02
6.88991398e-02 -3.06038499e-01 1.71901569e-01 1.14236796e+00
-3.58694345e-01 5.53867400e-01 -3.33767384e-01 -7.01193213e-01
-1.85349262e+00 6.11743748e-01 3.06188554e-01 -2.20134154e-01
-4.69314128e-01 8.39516938e-01 1.91295922e-01 4.74676937e-02
4.46807086e-01 -3.51170927e-01 -3.46188426e-01 1.76616639e-01
5.72948992e-01 5.77337086e-01 -1.27091810e-01 -1.02888513e+00
-6.61313653e-01 8.49664450e-01 -9.67264995e-02 1.52193800e-01
1.26786399e+00 2.20634848e-01 -6.09976090e-02 1.89411312e-01
9.61848915e-01 -1.07449442e-01 -1.48515129e+00 -1.32325113e-01
9.87327397e-02 -7.38854825e-01 -2.71023214e-01 -8.24388981e-01
-1.03043580e+00 1.01040077e+00 5.98361135e-01 4.24250588e-02
1.08036697e+00 3.79208505e-01 4.48964685e-01 3.53388041e-01
6.15565062e-01 -1.16231155e+00 3.98705035e-01 3.56207192e-01
1.21732390e+00 -1.14809966e+00 3.97245556e-01 -5.86796701e-01
-6.64609730e-01 1.20189440e+00 6.36694193e-01 -3.24969083e-01
4.73496169e-01 1.93839863e-01 -1.18288480e-01 -4.11943227e-01
-1.89865142e-01 -4.30171251e-01 6.82311773e-01 4.16638255e-01
2.95543641e-01 -6.78446516e-02 -2.09032908e-01 2.98276097e-01
3.77607226e-01 1.06227444e-02 -6.20864853e-02 1.18405604e+00
-4.73130494e-01 -1.19824111e+00 -5.38898349e-01 3.17742556e-01
-5.88522911e-01 3.07744205e-01 -9.18379128e-01 1.17903793e+00
3.64901036e-01 7.76390016e-01 3.39029394e-02 -6.71963617e-02
6.48746312e-01 1.60343200e-01 9.53636885e-01 -5.96994877e-01
-5.27342498e-01 2.32950777e-01 2.17083588e-01 -1.17759681e+00
-7.99557149e-01 -1.06033671e+00 -1.30650544e+00 1.74994826e-01
-3.98080796e-01 -2.70644307e-01 6.57194495e-01 1.28440750e+00
1.35332331e-01 4.22797352e-01 1.46340132e-01 -1.40832984e+00
-6.45286143e-01 -1.01843238e+00 -6.76713467e-01 6.31680727e-01
5.37802204e-02 -1.01698816e+00 -2.15602294e-01 1.60393894e-01] | [7.936575889587402, 0.2556283473968506] |
428d2ce1-4c3d-449b-b460-189bcfeaa7b9 | multi-scale-embedded-cnn-for-music-tagging | 1906.06746 | null | https://arxiv.org/abs/1906.06746v1 | https://arxiv.org/pdf/1906.06746v1.pdf | Multi-scale Embedded CNN for Music Tagging (MsE-CNN) | Convolutional neural networks (CNN) recently gained notable attraction in a variety of machine learning tasks: including music classification and style tagging. In this work, we propose implementing intermediate connections to the CNN architecture to facilitate the transfer of multi-scale/level knowledge between different layers. Our novel model for music tagging shows significant improvement in comparison to the proposed approaches in the literature, due to its ability to carry low-level timbral features to the last layer. | ['Nima Hamidi', 'Mohsen Vahidzadeh', 'Stephen Baek'] | 2019-06-16 | null | null | null | null | ['music-classification'] | ['music'] | [ 1.44170016e-01 -3.12779784e-01 -2.82670826e-01 -1.17209613e-01
-4.60406929e-01 -6.40427530e-01 4.14705515e-01 1.66820616e-01
-5.29643059e-01 4.55199182e-01 3.84152561e-01 1.18869498e-01
-3.71248163e-02 -6.23247862e-01 -4.39325929e-01 -2.09133908e-01
-1.68356806e-01 5.12350313e-02 -3.40624174e-05 -1.06421851e-01
2.03534976e-01 6.38623059e-01 -1.43192601e+00 6.45159960e-01
2.82339901e-01 1.28521097e+00 7.63042946e-04 6.50149703e-01
-2.42064133e-01 1.09124291e+00 -6.79558039e-01 -3.62614274e-01
-4.39566411e-02 -3.09455633e-01 -1.10649228e+00 -4.78861302e-01
5.52466273e-01 5.02230152e-02 -3.40518385e-01 9.00909185e-01
7.09478617e-01 2.85902768e-01 3.15904200e-01 -8.22329283e-01
-6.05808377e-01 9.43075061e-01 -2.16439679e-01 2.65545577e-01
2.53008753e-01 -3.60028118e-01 1.29871023e+00 -7.11335719e-01
4.86438781e-01 9.55740452e-01 1.17223549e+00 5.21484733e-01
-8.72047663e-01 -8.73303950e-01 -9.48465243e-02 4.18076485e-01
-1.44078934e+00 -3.16031188e-01 1.08637452e+00 -4.13549662e-01
1.00815058e+00 -2.23088376e-02 6.81541562e-01 9.93442833e-01
-1.34061873e-01 9.07724738e-01 9.65858877e-01 -4.29538935e-01
2.53399201e-02 -2.57195026e-01 -8.51479769e-02 5.49881995e-01
-2.44651079e-01 -2.67633885e-01 -1.13191521e+00 -4.29835655e-02
1.14691114e+00 -3.83246422e-01 1.35420710e-01 8.11044276e-02
-1.17353868e+00 6.18348956e-01 7.39756823e-01 7.05924809e-01
-3.38643253e-01 5.29440939e-01 9.16672349e-01 3.33665639e-01
4.73012179e-01 8.29553366e-01 -6.58178866e-01 -5.58948994e-01
-1.19485152e+00 -6.05498627e-02 6.07744098e-01 6.05892956e-01
1.76741689e-01 3.92853618e-01 -3.92234921e-01 1.20365691e+00
-8.12204629e-02 -1.33324251e-01 6.03722036e-01 -8.82304907e-01
2.93790996e-01 4.65029120e-01 -3.46757233e-01 -8.23318422e-01
-4.92901534e-01 -1.01119375e+00 -1.12864900e+00 1.10067993e-01
2.21394613e-01 1.78333428e-02 -4.70516920e-01 1.76516449e+00
-2.72542000e-01 6.10649288e-01 -3.29045616e-02 7.72950292e-01
1.15023720e+00 3.74128670e-01 1.55382946e-01 2.01386020e-01
1.19451237e+00 -8.64105225e-01 -6.76348150e-01 1.12583682e-01
2.54433304e-01 -1.13423574e+00 1.09257841e+00 5.63172579e-01
-1.18862414e+00 -1.15182698e+00 -1.02783203e+00 -1.27554357e-01
-4.67021137e-01 6.37479246e-01 7.72525132e-01 3.56360227e-01
-1.21752501e+00 1.03990209e+00 -4.24483359e-01 -3.53080362e-01
8.28999758e-01 7.65618801e-01 -1.76215723e-01 5.87463796e-01
-1.13782012e+00 5.03866851e-01 6.28249884e-01 1.72883630e-01
-5.46984851e-01 -5.33577144e-01 -6.22642934e-01 1.84178129e-01
-1.33213803e-01 -6.78321123e-01 1.36724639e+00 -1.24589789e+00
-2.06503606e+00 1.00916886e+00 8.25720578e-02 -6.17482245e-01
1.72478408e-01 -6.04315341e-01 -3.97863716e-01 -5.97856455e-02
-2.79877603e-01 9.23891783e-01 6.42955840e-01 -7.30662346e-01
-5.76124728e-01 -8.94091092e-03 4.73596632e-01 1.21446483e-01
-7.62706518e-01 1.76533118e-01 -4.78319943e-01 -1.29613447e+00
-2.50423774e-02 -9.17468727e-01 1.03361368e-01 -3.86128604e-01
-2.96420097e-01 -4.18779641e-01 5.88935792e-01 -5.10926485e-01
1.30796766e+00 -2.21531940e+00 3.66041154e-01 -1.60732359e-01
-2.93817651e-02 5.38331687e-01 -2.30190784e-01 2.81110287e-01
7.29223201e-03 -8.75703916e-02 1.51948616e-01 -5.61054945e-01
4.11911868e-02 1.92357358e-02 -3.43996733e-01 9.29743350e-02
1.65223792e-01 1.05893838e+00 -8.16509485e-01 -2.68324792e-01
2.28729308e-01 4.59446907e-01 -5.48081875e-01 2.24111751e-01
-7.69628137e-02 5.85128427e-01 -1.23527519e-01 4.96383756e-01
2.74432302e-01 -2.36398671e-02 2.08385617e-01 -3.61540228e-01
-1.31240875e-01 8.02564502e-01 -1.03994548e+00 2.26689792e+00
-8.10379028e-01 9.37859178e-01 -1.10484347e-01 -8.71396482e-01
8.87043059e-01 7.14231551e-01 5.60139298e-01 -2.99349308e-01
4.21095371e-01 1.76719412e-01 6.15382232e-02 -1.03895767e-02
6.41937971e-01 -2.72530138e-01 -3.90049100e-01 1.73211843e-01
6.41550779e-01 8.12281817e-02 -1.61632165e-01 -2.92246431e-01
8.70045781e-01 2.80825764e-01 1.81456745e-01 -1.59623086e-01
7.90535450e-01 -2.48625621e-01 5.24003088e-01 7.50804961e-01
-1.45662829e-01 5.53225577e-01 1.15519419e-01 -5.24462044e-01
-9.51876819e-01 -8.26763868e-01 -7.90851638e-02 1.35390186e+00
-2.50859439e-01 -6.84263647e-01 -6.08758211e-01 -4.03538197e-01
-1.23875923e-01 -1.47787869e-01 -6.09335601e-01 -1.50574341e-01
-7.34055042e-01 -2.57476032e-01 1.20975637e+00 9.85903084e-01
8.08826804e-01 -1.64176738e+00 -3.55711102e-01 4.27780241e-01
-2.95884795e-02 -1.07591712e+00 -2.48944044e-01 5.81081212e-01
-1.16250408e+00 -8.15794408e-01 -6.78066254e-01 -1.13904500e+00
-9.97690558e-02 -1.89277768e-01 1.20577002e+00 -1.66435912e-01
-2.28258967e-01 -4.71590944e-02 -4.39115673e-01 -5.47613144e-01
-1.65257230e-01 8.94216835e-01 7.07854852e-02 -7.30235428e-02
2.35281929e-01 -8.60978007e-01 -4.59988326e-01 -1.49805129e-01
-7.94957101e-01 -5.46257794e-02 4.41308439e-01 7.79842615e-01
5.17973781e-01 3.09146680e-02 8.95682633e-01 -6.64190710e-01
6.32995963e-01 2.67316960e-02 -2.11529464e-01 -1.89269006e-01
-1.29268423e-01 -6.25178590e-02 7.58241355e-01 -4.32938218e-01
-7.05838442e-01 1.19899884e-01 -5.34939349e-01 -4.95991558e-01
-4.47471648e-01 5.96275389e-01 8.27688202e-02 -2.48285070e-01
1.70199543e-01 -2.39694789e-01 -5.53024650e-01 -1.12256837e+00
5.11622012e-01 6.73301458e-01 8.53704453e-01 -5.16066909e-01
6.69987798e-01 2.71872193e-01 1.55510068e-01 -5.52517951e-01
-1.38624644e+00 -5.87394536e-01 -1.19105482e+00 -1.36011735e-01
9.83340442e-01 -9.32636917e-01 -9.30215299e-01 6.93693221e-01
-1.16781914e+00 -3.66442412e-01 -6.46425784e-01 4.83526379e-01
-7.27863848e-01 -1.45276263e-01 -1.07403576e+00 -6.04901433e-01
-4.70042586e-01 -6.65047109e-01 8.99048865e-01 3.28827232e-01
-5.83626151e-01 -1.16659367e+00 1.67202935e-01 2.15047272e-03
5.32980502e-01 8.17026123e-02 9.74569976e-01 -3.74465823e-01
-1.51966363e-01 -5.70605472e-02 -1.85449407e-01 5.32271922e-01
1.15728043e-01 -4.22988981e-01 -1.47183764e+00 -2.31633171e-01
-4.89231318e-01 -5.93816638e-01 1.14120269e+00 5.29287755e-01
1.50504708e+00 2.83619672e-01 1.71045005e-01 9.59934950e-01
1.27656400e+00 1.40545785e-01 4.32592750e-01 6.25014484e-01
8.32900405e-01 3.13612656e-03 1.00408792e-01 3.07957232e-01
-5.40954620e-02 9.56670880e-01 1.64802313e-01 -4.30190593e-01
-6.43630564e-01 -1.91267133e-01 1.86522990e-01 1.27466905e+00
-3.88334423e-01 1.52969971e-01 -5.53297460e-01 5.06025195e-01
-1.81567919e+00 -1.02800167e+00 1.92225009e-01 1.87773871e+00
9.11309183e-01 2.56284893e-01 3.88602257e-01 4.66649145e-01
4.24838275e-01 3.32280248e-01 -2.65860915e-01 -6.99735641e-01
-3.83528203e-01 9.57231641e-01 1.05395138e-01 7.05965683e-02
-1.63487697e+00 1.31955183e+00 7.39799500e+00 9.44749951e-01
-1.29020870e+00 3.77182007e-01 1.93070963e-01 -1.43798813e-01
3.24665755e-01 -3.74069631e-01 -4.21622604e-01 -4.94916886e-02
8.78914356e-01 3.38119417e-01 4.28280830e-01 4.50256586e-01
-4.97245044e-02 3.99415076e-01 -9.30661023e-01 1.06367135e+00
4.20081094e-02 -1.54488957e+00 2.28587642e-01 -3.10634851e-01
8.81043077e-01 3.14064510e-02 2.93063521e-01 3.46063435e-01
8.13777000e-02 -1.16216803e+00 8.33119810e-01 5.95780253e-01
9.75619078e-01 -1.25351477e+00 1.05046904e+00 -2.81291157e-01
-1.71715641e+00 -1.46783322e-01 -1.95890114e-01 -5.12181520e-01
-1.47333182e-02 2.79609472e-01 -6.15347803e-01 5.90740860e-01
8.38659942e-01 1.12657070e+00 -5.68905950e-01 1.17590094e+00
-6.41943514e-02 9.71911669e-01 9.91154164e-02 3.15869033e-01
2.97123134e-01 1.54717281e-01 3.88836533e-01 1.45820940e+00
2.00118706e-01 -3.35026950e-01 1.20060541e-01 6.68621719e-01
-5.50102949e-01 3.72554392e-01 -3.32277745e-01 -1.45216748e-01
1.65667936e-01 1.23645031e+00 -8.43617022e-01 -4.45535094e-01
-3.00312638e-01 1.21429181e+00 5.13162732e-01 6.93562701e-02
-4.67084706e-01 -6.20899141e-01 9.37834024e-01 -2.69061804e-01
4.86330509e-01 -3.61043125e-01 -6.79628551e-01 -1.03344226e+00
-2.52836198e-01 -3.97822440e-01 5.23174167e-01 -7.92152166e-01
-1.30403376e+00 7.95960367e-01 -5.20715892e-01 -1.27637947e+00
-5.25438339e-02 -8.02644253e-01 -6.61181033e-01 8.20065439e-01
-1.46374190e+00 -1.38766801e+00 -4.15841378e-02 5.98174036e-01
6.14471078e-01 -5.40940106e-01 1.29827952e+00 7.34536827e-01
-2.89620340e-01 7.66001582e-01 6.02578791e-03 5.38568914e-01
8.53033185e-01 -1.38992918e+00 4.97419596e-01 3.09454978e-01
8.37035477e-01 5.36009073e-01 2.20951349e-01 -2.89930463e-01
-8.91359329e-01 -1.12291253e+00 1.01086867e+00 -1.57984957e-01
7.86986291e-01 -3.79530311e-01 -6.02446318e-01 4.03564423e-01
4.21794444e-01 -2.93767691e-01 9.89554286e-01 7.29555070e-01
-5.97294807e-01 -1.50858611e-01 -5.52641034e-01 2.85707861e-01
1.14611995e+00 -1.14998627e+00 -5.85796356e-01 -2.12718770e-01
3.71368229e-01 -3.47252280e-01 -1.14379966e+00 4.91698414e-01
8.72417092e-01 -6.47633910e-01 1.15094817e+00 -6.88590467e-01
4.65523511e-01 -2.43278712e-01 -9.10168663e-02 -1.23834383e+00
-5.58525562e-01 -5.10571837e-01 -2.30406627e-01 1.43301570e+00
1.52577877e-01 1.26544699e-01 7.71951377e-01 -6.22404814e-01
-3.17215830e-01 -6.59296453e-01 -1.04916441e+00 -6.60431564e-01
1.01892285e-01 -6.98567986e-01 5.79803228e-01 1.04121184e+00
9.28578228e-02 5.21940231e-01 -6.46621287e-01 -1.71095252e-01
9.35585499e-02 4.18497622e-01 4.15383548e-01 -1.78311980e+00
-4.41215962e-01 -8.75110865e-01 -7.82170057e-01 -8.32014143e-01
4.40589160e-01 -1.29516244e+00 -2.54230946e-01 -1.45269811e+00
2.40939617e-01 -3.58633906e-01 -1.16005778e+00 6.89108193e-01
1.36841804e-01 1.28736782e+00 4.37766492e-01 1.87786296e-01
-8.50352705e-01 3.60051543e-01 1.05248082e+00 -1.75477505e-01
-3.47207546e-01 8.00204128e-02 -3.80876571e-01 9.58068013e-01
1.01069820e+00 -2.92714685e-01 1.42676067e-02 -4.85197306e-01
3.39846700e-01 -2.60062397e-01 3.69596332e-01 -1.56306744e+00
-6.04998656e-02 4.09077913e-01 6.33083224e-01 -5.92053831e-01
4.73815590e-01 -5.27388513e-01 4.97750491e-02 2.43911117e-01
-7.78540134e-01 -1.22270226e-01 5.66238225e-01 4.85312790e-02
-6.94132745e-01 -1.34509161e-01 6.91766083e-01 -2.97995731e-02
-8.14706087e-01 1.36955529e-01 -4.39081341e-01 -2.40372688e-01
1.53280169e-01 1.14854224e-01 1.66737586e-01 -2.88596302e-01
-1.15852523e+00 -6.09988153e-01 2.31109276e-01 7.98903763e-01
2.88344175e-01 -1.77702785e+00 -2.84377456e-01 -2.13593175e-03
8.22983757e-02 -5.32039762e-01 5.18076718e-02 6.99402690e-01
-2.66097009e-01 6.21562958e-01 -6.05651915e-01 -3.83128732e-01
-1.27094889e+00 8.45645368e-02 3.07619661e-01 -2.39624038e-01
-6.18035674e-01 1.00368893e+00 -2.64248610e-01 -3.26594800e-01
6.34494483e-01 -3.62329006e-01 -4.69915152e-01 2.46225968e-01
2.90878505e-01 2.05792442e-01 2.13895753e-01 -7.26602077e-01
-2.89543301e-01 7.72976398e-01 2.42647573e-01 -8.80839601e-02
1.48285031e+00 1.72329694e-01 -2.15802982e-01 1.14242601e+00
1.10074854e+00 -4.20559198e-02 -9.08042014e-01 -3.22549313e-01
3.64138424e-01 1.52792325e-02 3.16992998e-01 -9.51867402e-01
-1.11908734e+00 1.14905238e+00 7.81110168e-01 9.64461640e-02
1.28576934e+00 1.16516158e-01 1.01973486e+00 4.03941154e-01
3.16139042e-01 -1.28898013e+00 1.89189151e-01 9.11819458e-01
5.35405338e-01 -8.54201615e-01 -2.66586453e-01 1.18451327e-01
-3.16506267e-01 1.36188030e+00 4.06405985e-01 -4.47737485e-01
7.35692143e-01 3.08035791e-01 3.63645285e-01 -1.90596297e-01
-5.08483112e-01 -5.25395751e-01 8.40934992e-01 3.41645747e-01
1.20556045e+00 6.79311752e-02 -1.95935577e-01 8.03440988e-01
-4.00543302e-01 2.45769158e-01 8.30250084e-02 6.41309321e-01
-2.78563380e-01 -1.41462123e+00 -2.04336364e-02 2.72981465e-01
-8.78964961e-01 -2.10819915e-01 -6.96051359e-01 3.95948231e-01
5.68830073e-01 7.59268939e-01 2.35011041e-01 -5.15597105e-01
3.48372042e-01 3.61212224e-01 6.49106264e-01 -7.38021076e-01
-1.33575845e+00 3.37327570e-01 5.87300956e-02 -4.78080064e-01
-9.74573553e-01 -4.63788867e-01 -1.06313455e+00 8.60959217e-02
-1.48490280e-01 1.83506265e-01 5.57363808e-01 9.12688553e-01
2.81372875e-01 1.16260993e+00 4.05511290e-01 -1.16869736e+00
6.87696859e-02 -1.22090816e+00 -7.53596604e-01 4.70904976e-01
2.37142622e-01 -6.70826077e-01 2.44920328e-01 1.84264377e-01] | [15.786416053771973, 5.2700276374816895] |
51a8a8ec-2516-4e72-ab70-c036cc68c7cf | sleep-quality-prediction-in-caregivers-using | null | null | https://www.sciencedirect.com/science/article/pii/S001048251930160X | https://www.sciencedirect.com/science/article/pii/S001048251930160X | Sleep quality prediction in caregivers using physiological signals | Most caregivers of people with dementia (CPWD) experience a high degree of stress due to the demands of providing care, especially when addressing unpredictable behavioral and psychological symptoms of dementia. Such challenging responsibilities make caregivers susceptible to poor sleep quality with detrimental effects on their overall health. Hence, monitoring caregivers’ sleep quality can provide important CPWD stress assessment. Most current sleep studies are based on polysomnography, which is expensive and potentially disrupts the caregiving routine. To address these issues, we propose a clinical decision support system to predict sleep quality based on trends of physiological signals in the deep sleep stage. This system utilizes four raw physiological signals using a wearable device (E4 wristband): heart rate variability, electrodermal activity, body movement, and skin temperature. To evaluate the performance of the proposed method, analyses were conducted on a two-week period of sleep monitored on eight CPWD. The best performance is achieved using the random forest classifier with an accuracy of 75% for sleep quality, and 73% for restfulness, respectively. We found that the most important features to detect these measures are sleep efficiency (ratio of amount of time asleep to the amount of time in bed) and skin temperature. The results from our sleep analysis system demonstrate the capability of using wearable sensors to measure sleep quality and restfulness in CPWD. | ['Jennifer C. Hughes', 'Reza Sadeghi', 'Tanvi Banerjee', 'Larry W. Lawhorne'] | 2019-05-20 | null | null | null | computers-in-biology-and-medicine-2019-5 | ['heart-rate-variability', 'sleep-quality-prediction-1'] | ['medical', 'medical'] | [-2.02469751e-01 -3.87622952e-01 -2.08386287e-01 -4.10067737e-01
2.16363501e-02 -1.32819235e-01 -2.58997798e-01 8.65950435e-02
-6.25615597e-01 1.11316264e+00 4.49020505e-01 -1.02599762e-01
-3.46496776e-02 -3.74106050e-01 4.25495803e-01 -8.20935309e-01
-1.79491177e-01 -1.16876699e-01 -7.97008798e-02 4.40009497e-02
-1.14947081e-01 1.85851797e-01 -1.15064335e+00 -3.43857706e-02
1.15639186e+00 1.24661267e+00 3.93216461e-01 2.11661667e-01
4.33238208e-01 3.09539557e-01 -9.04165924e-01 1.22777745e-01
-2.03523248e-01 -4.67061788e-01 -2.56420970e-01 -5.26962541e-02
-5.56945682e-01 -4.54026610e-01 -2.74521504e-02 8.16485584e-01
8.73231053e-01 3.13428231e-02 1.30079940e-01 -1.15848184e+00
-1.44774556e-01 -1.33333728e-01 1.25561759e-01 1.03201067e+00
5.72905958e-01 1.06932446e-01 3.48805040e-01 -4.15699095e-01
-3.01216125e-01 5.50372660e-01 9.59783673e-01 6.92200840e-01
-1.29343951e+00 -6.81540191e-01 -3.36247504e-01 6.52957439e-01
-1.23731995e+00 -1.07847607e+00 5.83840847e-01 -3.27346742e-01
1.23573649e+00 3.19096744e-01 1.41401649e+00 1.24264002e+00
1.00905573e+00 -1.87361136e-01 1.39506245e+00 -3.38331833e-02
9.82766211e-01 1.54834881e-01 3.77067327e-01 2.38679975e-01
5.90607822e-01 -2.44599789e-01 -5.50152779e-01 -5.34161806e-01
1.45116493e-01 8.23887825e-01 -7.93168783e-01 3.43161076e-01
-8.82135868e-01 4.32065725e-01 3.19035560e-01 4.29933488e-01
-4.70853120e-01 -2.95073718e-01 3.01118106e-01 1.80772379e-01
4.38364655e-01 9.06114727e-02 -4.43956435e-01 -4.74066675e-01
-9.87152576e-01 -3.46642494e-01 7.80071378e-01 3.74898285e-01
1.44856483e-01 -3.80731761e-01 -2.74625540e-01 7.91544437e-01
5.57845592e-01 8.45965147e-01 1.04371953e+00 -8.59372139e-01
-2.59388871e-02 6.81615651e-01 4.35706407e-01 -4.88533914e-01
-9.87360537e-01 -2.65558034e-01 -9.11383331e-01 3.41077037e-02
1.51034929e-02 -2.91287422e-01 -2.88980752e-01 1.56730437e+00
-1.74287587e-01 -3.02255481e-01 -7.49114854e-03 8.77334356e-01
6.24375224e-01 -3.87061499e-02 3.37826431e-01 -1.13721728e+00
1.77839208e+00 -4.28840876e-01 -1.21565533e+00 -6.19839966e-01
-2.20044963e-02 -2.03455910e-01 1.17392194e+00 3.05589736e-01
-8.60903859e-01 -2.28389591e-01 -1.20388639e+00 3.90721440e-01
6.80386135e-03 4.20890227e-02 2.43466154e-01 1.02581179e+00
-1.20883322e+00 5.60568333e-01 -1.60392249e+00 -9.54319537e-01
4.38520551e-01 4.32243139e-01 -2.53245920e-01 5.96905798e-02
-9.07647848e-01 1.17664742e+00 -3.65614355e-01 2.67113417e-01
-3.82315874e-01 -3.77671182e-01 -4.18076545e-01 1.57429263e-01
-4.11872327e-01 -1.10048985e+00 1.02953124e+00 -4.89473224e-01
-1.50485361e+00 6.58859909e-01 -7.23043978e-01 -1.61043763e-01
2.13226397e-02 -3.98473114e-01 -7.12406933e-01 3.20008606e-01
2.84348637e-01 -1.93061098e-01 4.98759598e-01 -2.33910456e-01
1.93680257e-01 -1.13160825e+00 -4.35410708e-01 6.16066437e-03
-5.62193930e-01 2.14181602e-01 6.30918145e-01 -1.51801556e-01
2.50076532e-01 -8.36251795e-01 4.18968685e-03 4.37468179e-02
2.20088363e-02 9.99152884e-02 5.86048126e-01 -6.68487191e-01
1.33092213e+00 -2.11171460e+00 -5.34680843e-01 -2.53680766e-01
4.32859659e-01 2.29145158e-02 7.25003481e-01 1.84845895e-01
2.12677702e-01 1.14664976e-02 -1.72112614e-01 -4.86255378e-01
-1.70386344e-01 4.13546532e-01 2.09816739e-01 7.97159076e-01
-3.77056152e-01 8.13883066e-01 -7.26998925e-01 -1.51392490e-01
1.12588309e-01 5.54364026e-01 -2.31583223e-01 5.96379638e-01
4.79149193e-01 3.26598704e-01 -2.03918010e-01 7.76543677e-01
3.79807860e-01 -3.10688972e-01 3.23815048e-01 -1.65177211e-01
-1.84807554e-01 6.86644971e-01 -1.78123742e-01 1.47899425e+00
-1.39152646e-01 3.21698129e-01 9.37918574e-02 -4.38551784e-01
1.05288386e+00 5.28244913e-01 4.56375748e-01 -8.60918462e-01
3.71922493e-01 5.74245378e-02 8.59878585e-02 -1.01578975e+00
-3.78967166e-01 -6.71786785e-01 2.08067879e-01 4.02260214e-01
-3.19945097e-01 4.64590162e-01 -2.04601571e-01 -2.89207190e-01
1.69912267e+00 -4.43998218e-01 9.42628503e-01 -6.08896673e-01
2.21566647e-01 -6.03793442e-01 8.42151046e-01 1.38170660e-01
-1.13240767e+00 2.44547963e-01 2.99643934e-01 -3.43751967e-01
-3.16620946e-01 -1.26651478e+00 -1.97067618e-01 5.26647866e-01
-1.11260019e-01 -4.47423160e-01 -6.84791505e-01 -1.52106822e-01
-2.08470210e-01 5.99730670e-01 -4.66791898e-01 -7.30883121e-01
7.55429119e-02 -1.14998555e+00 2.29605049e-01 6.38498783e-01
7.74595737e-01 -9.77398396e-01 -1.43514025e+00 2.50165403e-01
-4.15676653e-01 -8.13988209e-01 -3.29799026e-01 6.77172601e-01
-1.40949690e+00 -1.03515387e+00 -4.42034274e-01 -4.03412670e-01
4.22053307e-01 4.56012219e-01 7.08983898e-01 2.24868357e-02
-2.01988921e-01 3.89371544e-01 -2.55636126e-01 -2.06390187e-01
2.95249194e-01 -9.15589854e-02 6.10082150e-01 -3.36928934e-01
9.22221482e-01 -1.52510619e+00 -1.21802342e+00 4.54502940e-01
-1.93313092e-01 -5.16088843e-01 4.34694558e-01 2.97892839e-01
3.42882335e-01 -1.34459332e-01 8.81829858e-01 3.78147364e-02
9.76190686e-01 -7.85177231e-01 1.09986603e-01 -2.67431699e-02
-1.12292957e+00 -3.45974624e-01 5.54221272e-01 -1.68639943e-01
-7.81187177e-01 -4.31684226e-01 1.14800982e-01 1.04683094e-01
-3.62588316e-01 7.30118307e-04 -5.89946389e-01 5.18727005e-01
6.96291864e-01 1.07123956e-01 3.82228434e-01 -3.57166827e-01
-7.43936718e-01 9.85141516e-01 4.51006025e-01 5.60727417e-02
-6.93521425e-02 3.74771565e-01 -3.09120566e-01 -8.92989218e-01
-7.81430304e-01 -6.80868506e-01 -4.84558731e-01 -2.25033581e-01
1.15854073e+00 -9.47192490e-01 -1.11596024e+00 3.37644368e-01
-6.97113514e-01 -4.00145292e-01 1.74850617e-02 7.20569015e-01
-1.98662668e-01 1.98025599e-01 -3.30870509e-01 -1.08694816e+00
-1.11858010e+00 -8.26635599e-01 5.21214485e-01 3.49538118e-01
-9.33109105e-01 -9.55455124e-01 3.79545838e-01 4.74261463e-01
8.05813551e-01 3.76223415e-01 6.76963329e-01 -2.42416665e-01
2.56043434e-01 2.83254776e-02 3.41854781e-01 3.87096494e-01
8.81512344e-01 -7.28195727e-01 -9.99220133e-01 -3.79045099e-01
1.13573563e+00 2.15698835e-02 1.47018597e-01 6.74539268e-01
4.45867300e-01 -3.90980273e-01 -2.87178069e-01 5.17562091e-01
1.26666462e+00 6.23356044e-01 8.47667813e-01 5.33774316e-01
-4.46886532e-02 -7.05546066e-02 6.83288127e-02 8.26070845e-01
6.03901267e-01 5.59403300e-01 4.74164069e-01 4.39382553e-01
1.94223002e-01 3.89842540e-01 1.00986886e+00 7.33198643e-01
-3.11936766e-01 -2.64626816e-02 -6.93777084e-01 1.85545295e-01
-1.55594802e+00 -8.77590001e-01 -1.91326261e-01 2.18016148e+00
6.65105045e-01 -2.42020078e-02 3.44995260e-01 3.54140162e-01
4.56091017e-01 -1.61595017e-01 -8.22249830e-01 -3.51760596e-01
1.24892294e-01 4.23425734e-01 8.50227475e-02 -7.36491680e-02
-4.36120600e-01 7.91124776e-02 6.80299473e+00 -6.41179264e-01
-1.05153835e+00 4.16976869e-01 2.87881225e-01 -7.44921625e-01
2.22738504e-01 -3.26057762e-01 -3.67683858e-01 1.15661073e+00
1.57125783e+00 1.32390922e-02 7.10267663e-01 9.53615546e-01
1.07003832e+00 -7.35529006e-01 -1.05499315e+00 1.14272463e+00
1.14626467e-01 -4.05746162e-01 -9.57189202e-01 6.62720427e-02
-1.03796162e-01 1.57983266e-02 -4.88640785e-01 -9.64665413e-02
-5.81109464e-01 -8.90207171e-01 3.43436867e-01 9.74125028e-01
6.83110297e-01 -3.05722594e-01 9.70200956e-01 1.80127636e-01
-8.13010693e-01 -2.98159599e-01 -1.95009112e-01 -7.27863193e-01
2.74620444e-01 9.68575180e-01 -8.32402647e-01 -2.85306603e-01
1.18019676e+00 6.88597202e-01 -5.29863715e-01 1.08037233e+00
-5.18360138e-01 7.12652206e-01 -3.50768149e-01 -1.34912908e-01
-7.74798453e-01 -1.89620048e-01 2.66037732e-01 6.27022922e-01
5.76597869e-01 5.34729123e-01 -3.13381165e-01 9.64327395e-01
3.17025185e-01 -3.92575324e-01 -3.54613841e-01 3.33048433e-01
5.20800054e-01 1.22661459e+00 -7.16548026e-01 1.47958100e-02
-4.48463261e-01 9.69321668e-01 -3.60569619e-02 1.16989411e-01
-4.72019315e-01 3.83486114e-02 1.01312590e+00 3.53500068e-01
-3.65682781e-01 -4.38060224e-01 -7.05106735e-01 -1.16359198e+00
3.88760954e-01 -4.44205523e-01 2.43190289e-01 -9.07560229e-01
-1.02459562e+00 4.67611521e-01 -3.16399008e-01 -1.01992786e+00
3.00675899e-01 -8.73354673e-02 -1.00045156e+00 8.33034337e-01
-1.15286505e+00 -1.32738173e-01 -1.05706739e+00 9.26260889e-01
1.81736588e-01 6.78266510e-02 1.24685621e+00 3.74225885e-01
-8.85099173e-01 3.59891266e-01 -2.82364011e-01 -5.86889505e-01
6.74366355e-01 -1.01671147e+00 -3.35441649e-01 6.10266209e-01
-8.64428222e-01 9.23587143e-01 4.79666471e-01 -6.03008747e-01
-1.25584412e+00 -6.78119540e-01 1.03777063e+00 -4.34679687e-01
2.99237609e-01 -2.77505368e-01 -5.67850947e-01 3.24785560e-01
-1.31254354e-02 -1.20464027e-01 1.37314665e+00 -1.19359940e-01
4.95078772e-01 -5.41422665e-01 -1.60740483e+00 2.02468738e-01
9.01665449e-01 -4.69471812e-01 -1.09081411e+00 2.76919007e-02
3.12208384e-01 3.23558450e-01 -1.16556573e+00 1.31790951e-01
7.98567295e-01 -1.37329817e+00 4.80384827e-01 -1.92502309e-02
-6.58542737e-02 -7.04289898e-02 -1.76482499e-01 -1.22137558e+00
-4.89870965e-01 -5.61700583e-01 -3.24904442e-01 1.05964506e+00
-2.36084342e-01 -1.00469303e+00 5.00647664e-01 1.27894032e+00
-3.69242907e-01 -6.59636021e-01 -9.71728921e-01 -6.98624611e-01
-7.66981006e-01 -3.39032412e-02 3.67006034e-01 4.18269992e-01
9.42954779e-01 5.75559914e-01 -4.51143906e-02 -2.16142219e-02
3.10981810e-01 -3.96328300e-01 5.47085935e-03 -1.40500093e+00
1.41488537e-01 1.13701873e-01 -4.93595004e-01 -1.68195397e-01
-9.00485292e-02 -6.82105541e-01 -1.26078010e-01 -1.85098290e+00
4.29289758e-01 -7.37430379e-02 -5.83469510e-01 5.42699337e-01
-1.55629605e-01 -3.83912809e-02 -4.57755029e-01 3.10130358e-01
-3.34509701e-01 8.02322328e-01 6.92023456e-01 1.63362503e-01
-7.46523142e-01 3.14051390e-01 -9.21195567e-01 5.48788607e-01
1.08638239e+00 -4.54316527e-01 -6.66809320e-01 -8.49867687e-02
2.19932869e-02 2.15066090e-01 3.27141523e-01 -1.37064028e+00
1.10633045e-01 2.78944820e-02 7.36403763e-01 -4.83676419e-02
6.01808012e-01 -1.05837762e+00 4.38656569e-01 8.36167276e-01
3.76377344e-01 2.28711352e-01 1.84087604e-02 2.47545972e-01
2.70225555e-01 1.57175750e-01 8.20703030e-01 -1.19297899e-01
4.23446745e-02 -1.36917476e-02 -8.29463482e-01 -2.21510544e-01
8.51318657e-01 -5.20170450e-01 -5.75140715e-01 -1.24320574e-01
-1.02132285e+00 -9.52837020e-02 5.81283748e-01 4.03615348e-02
6.48462117e-01 -1.25208461e+00 1.12583682e-01 3.84905726e-01
-7.31706619e-02 -3.91024709e-01 1.90506265e-01 1.66855896e+00
-4.99495529e-02 3.03439379e-01 -6.93894207e-01 -4.05992419e-01
-1.13929963e+00 4.44482356e-01 5.78441083e-01 1.12149648e-01
-7.53877044e-01 2.75636047e-01 -4.67716902e-01 7.32636511e-01
2.37342164e-01 -7.64521241e-01 -2.41708696e-01 1.99527010e-01
9.23914790e-01 8.53190362e-01 4.64918047e-01 -5.76012172e-02
-9.60938871e-01 1.65900186e-01 3.90182585e-01 1.70278475e-01
1.49816728e+00 -4.66337651e-01 -5.02437711e-01 6.20675564e-01
9.24208045e-01 -1.78890318e-01 -9.26671207e-01 4.69457358e-01
-2.71388721e-02 -2.27908604e-02 6.08720705e-02 -9.18652117e-01
-8.20372641e-01 6.55967355e-01 1.39961505e+00 3.09093148e-01
1.52246988e+00 -1.30581245e-01 1.13569975e+00 3.20803970e-01
5.67417383e-01 -8.71050000e-01 6.37412891e-02 -2.46308103e-01
7.30353177e-01 -9.26723659e-01 1.52690977e-01 1.98726848e-01
-2.87277907e-01 1.13188291e+00 3.48867327e-01 6.58356771e-02
1.02469742e+00 1.84129968e-01 7.20143840e-02 -2.09924608e-01
-6.48919165e-01 1.65743798e-01 -2.76361853e-01 7.49658823e-01
4.29067701e-01 2.22111881e-01 -6.25172317e-01 1.40501451e+00
-5.04798114e-01 4.22829628e-01 4.33198959e-01 9.80525076e-01
-7.97895312e-01 -7.38607943e-01 -3.68639082e-01 9.16845143e-01
-5.68411887e-01 9.69474763e-02 -1.81889087e-01 1.71531111e-01
1.50739551e-01 1.61071491e+00 -6.83870018e-02 -3.95823091e-01
5.25848866e-01 5.73094964e-01 2.83219963e-01 -6.77490354e-01
-5.48717260e-01 -1.56390920e-01 1.13130510e-01 -6.96379542e-01
-8.01565170e-01 -7.97043860e-01 -1.18854606e+00 -1.56822905e-01
1.26486048e-01 8.12403858e-02 5.39339662e-01 1.11610246e+00
4.51193273e-01 4.47967529e-01 4.15091246e-01 -6.85781777e-01
-1.17561698e-01 -1.35723341e+00 -9.84097958e-01 -5.37070744e-02
6.07235909e-01 -9.00560021e-01 -6.42997086e-01 9.78134293e-03] | [13.560246467590332, 3.3915083408355713] |
1dcef78b-d14d-4758-9f40-d5c268a936c2 | structure-aware-face-clustering-on-a-large-1 | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Shen_Structure-Aware_Face_Clustering_on_a_Large-Scale_Graph_With_107_Nodes_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Shen_Structure-Aware_Face_Clustering_on_a_Large-Scale_Graph_With_107_Nodes_CVPR_2021_paper.pdf | Structure-Aware Face Clustering on a Large-Scale Graph With 107 Nodes | Face clustering is a promising method for annotating unlabeled face images. Recent supervised approaches have boosted the face clustering accuracy greatly, however their performance is still far from satisfactory. These methods can be roughly divided into global-based and local-based ones. Global-based methods suffer from the limitation of training data scale, while local-based ones are difficult to grasp the whole graph structure information and usually take a long time for inference. Previous approaches fail to tackle these two challenges simultaneously. To address the dilemma of large-scale training and efficient inference, we propose the STructure-AwaRe Face Clustering (STAR-FC) method. Specifically, we design a structure-preserved subgraph sampling strategy to explore the power of large-scale training data, which can increase the training data scale from 10^5 to 10^7. During inference, the STAR-FC performs efficient full-graph clustering with two steps: graph parsing and graph refinement. And the concept of node intimacy is introduced in the second step to mine the local structural information. The STAR-FC gets 91.97 pairwise F-score on partial MS1M within 310s which surpasses the state-of-the-arts. Furthermore, we are the first to train on very large-scale graph with 20M nodes, and achieve superior inference results on 12M testing data. Overall, as a simple and effective method, the proposed STAR-FC provides a strong baseline for large-scale face clustering. Code is available at https://sstzal.github.io/STAR-FC/. | ['Jie zhou', 'Jiwen Lu', 'Dalong Du', 'Guan Huang', 'Zheng Zhu', 'Wanhua Li', 'Shuai Shen'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['face-clustering'] | ['computer-vision'] | [-8.83313268e-02 2.18771681e-01 -3.74038309e-01 -6.04835153e-01
-8.63668561e-01 -3.43848944e-01 3.83269072e-01 -2.13289469e-01
-6.36387318e-02 3.84044439e-01 4.33174595e-02 -1.15237996e-01
-1.32478669e-01 -7.30921626e-01 -5.90198934e-01 -8.21016133e-01
-1.10099494e-01 6.78535402e-01 1.67086974e-01 1.96689948e-01
-3.27382050e-02 4.20384258e-01 -1.50486290e+00 1.14719413e-01
7.26899266e-01 8.28781426e-01 1.51171848e-01 2.62371212e-01
-2.29522362e-01 5.01185060e-01 -2.99517244e-01 -6.87932730e-01
6.06388338e-02 -4.79163438e-01 -9.30972695e-01 3.69596779e-01
6.39836669e-01 -8.39768201e-02 -2.12211818e-01 1.17745769e+00
5.05285859e-01 -1.33023441e-01 4.32596415e-01 -1.51519227e+00
-2.85490453e-01 6.51737332e-01 -1.00487781e+00 -1.53668121e-01
2.45064139e-01 -3.73382457e-02 9.28261936e-01 -1.05701160e+00
6.85393035e-01 1.48046017e+00 7.31393516e-01 7.79974163e-01
-1.03782392e+00 -1.03039765e+00 1.92172170e-01 1.76362231e-01
-1.78097761e+00 -7.15226114e-01 8.71714294e-01 -1.80956975e-01
6.21784449e-01 1.38476312e-01 4.53714907e-01 8.46484005e-01
-4.59767908e-01 6.82220995e-01 1.16386330e+00 -2.51708210e-01
2.06758708e-01 -2.25565061e-01 -5.82817607e-02 1.13311386e+00
2.07598850e-01 -2.30493695e-01 -5.60663521e-01 -8.47558454e-02
7.46887565e-01 -8.70799273e-02 -1.55281276e-01 -1.19120724e-01
-7.35546589e-01 9.42028880e-01 5.89067638e-01 3.13746035e-01
-3.83341610e-02 7.61109814e-02 2.56870240e-01 1.32244810e-01
5.13375461e-01 3.90146896e-02 -3.57731044e-01 1.38185844e-01
-1.25894833e+00 -9.83529091e-02 7.94682443e-01 9.96931255e-01
8.73622656e-01 -2.03514442e-01 4.04982045e-02 9.42086160e-01
5.27313411e-01 3.43753397e-01 5.06636053e-02 -1.11830258e+00
3.56595188e-01 7.07226336e-01 -5.70300996e-01 -1.27424932e+00
-3.77846241e-01 -3.66084605e-01 -1.10891306e+00 -1.19549215e-01
5.81765652e-01 -1.72148451e-01 -1.01843822e+00 1.82580292e+00
8.07647705e-01 6.85901225e-01 -3.84599268e-01 6.97309554e-01
9.80917990e-01 3.49734217e-01 -2.14165747e-02 -4.94164616e-01
1.49546897e+00 -1.05230820e+00 -5.96030653e-01 -1.10277452e-01
6.66186929e-01 -6.04121923e-01 9.44023728e-01 2.99283206e-01
-9.29377735e-01 -4.76968437e-01 -6.32913053e-01 1.96638823e-01
-1.36096522e-01 1.52912304e-01 8.23138714e-01 7.90520489e-01
-1.36637068e+00 4.36430156e-01 -8.20351481e-01 -4.45505589e-01
9.06821311e-01 5.76572061e-01 -6.00949287e-01 -5.71372271e-01
-7.45173216e-01 1.95133731e-01 3.36830080e-01 1.79362386e-01
-8.06954443e-01 -4.51069295e-01 -8.21232080e-01 2.30232114e-03
8.08948278e-01 -4.41764116e-01 7.41166711e-01 -6.25577152e-01
-1.29649425e+00 9.30108964e-01 -4.13071632e-01 4.70669754e-02
1.04466058e-01 2.05882311e-01 -3.78237277e-01 4.20817912e-01
1.49556413e-01 9.38345730e-01 9.17247653e-01 -1.29889333e+00
-3.31456065e-01 -7.54677653e-01 -1.24560989e-01 9.66857560e-03
-6.46173954e-01 1.56523913e-01 -1.32569790e+00 -5.11740625e-01
2.51127571e-01 -1.03019011e+00 -1.57019928e-01 -5.77341765e-02
-4.01733696e-01 -5.34871340e-01 7.56977797e-01 -4.27196085e-01
1.26189423e+00 -2.03794074e+00 7.12976381e-02 3.38281780e-01
4.64668185e-01 3.45366627e-01 -4.02627915e-01 2.17098787e-01
7.00782165e-02 3.20407987e-01 -2.57784575e-01 -7.85663724e-01
-1.08450092e-01 2.66189188e-01 2.78453708e-01 3.49709958e-01
1.60342067e-01 1.06839633e+00 -7.94193506e-01 -9.59339619e-01
-3.98144349e-02 6.00535691e-01 -7.42713332e-01 1.65883318e-01
-8.11973065e-02 5.17433703e-01 -3.63479853e-01 8.23597372e-01
1.03473353e+00 -7.12782681e-01 8.61010075e-01 -2.41325721e-01
4.23866749e-01 -4.19142395e-02 -1.40700102e+00 1.77763379e+00
-1.33257553e-01 2.18544245e-01 5.37234128e-01 -1.02909470e+00
7.04165101e-01 2.05551103e-01 5.59589505e-01 -4.00928140e-01
1.25554889e-01 -1.45370765e-02 -1.52472496e-01 -3.17687511e-01
-6.58047795e-02 -5.78792300e-03 1.55417934e-01 3.34975481e-01
2.43437260e-01 1.66454300e-01 3.10982853e-01 5.23057103e-01
1.15493965e+00 3.30447033e-02 4.03747447e-02 -2.91127056e-01
5.77381253e-01 -4.15439308e-01 7.41676927e-01 3.03270638e-01
-1.54416472e-01 4.84238088e-01 6.47916555e-01 -1.76166251e-01
-4.32144850e-01 -7.92568266e-01 -4.35070843e-02 1.08155465e+00
1.06353998e-01 -1.08081675e+00 -1.21872389e+00 -1.14349127e+00
-6.63661882e-02 7.79459551e-02 -5.62616587e-01 4.73701321e-02
-5.12262166e-01 -9.15452302e-01 5.94447851e-01 4.37980801e-01
5.86689234e-01 -1.01185250e+00 2.41247475e-01 -1.77475363e-01
-4.18621957e-01 -1.45361936e+00 -6.16088212e-01 -2.96067864e-01
-8.36595654e-01 -1.29783213e+00 -2.45462552e-01 -1.00962520e+00
1.08958519e+00 3.42854142e-01 1.26863873e+00 6.81977093e-01
-3.49499047e-01 1.91864699e-01 -2.63257921e-01 1.22060545e-01
-5.89887798e-02 1.43606290e-01 2.86310576e-02 1.01942122e-01
4.65587169e-01 -7.62139499e-01 -6.67535007e-01 5.62761784e-01
-5.46152174e-01 -4.36016060e-02 7.41331577e-01 6.10976279e-01
6.64839387e-01 3.89589906e-01 5.23603737e-01 -1.11103415e+00
1.36291534e-01 -4.15851653e-01 -4.16871548e-01 3.06447923e-01
-7.43801355e-01 -4.89386693e-02 4.02859092e-01 -2.86969393e-01
-9.01229680e-01 3.18365216e-01 -3.40514392e-01 -4.23449636e-01
-3.08221936e-01 4.50007588e-01 -5.74767888e-01 -2.44003221e-01
2.37308234e-01 -3.56748514e-02 5.09405434e-02 -5.23898423e-01
2.97510982e-01 4.77209717e-01 3.91910881e-01 -6.67241395e-01
9.78080511e-01 5.36807358e-01 1.26878455e-01 -9.37068701e-01
-8.03943574e-01 -5.03385842e-01 -7.90417731e-01 -3.46531421e-01
8.33471358e-01 -9.93398726e-01 -9.84358311e-01 5.02744138e-01
-7.73751378e-01 -4.20972824e-01 2.13060364e-01 1.07844047e-01
-5.47802746e-02 7.00424850e-01 -7.61355221e-01 -6.67861938e-01
-2.37450510e-01 -1.04307723e+00 1.34993565e+00 1.39659733e-01
3.09040286e-02 -9.52495277e-01 -3.38136405e-01 8.20552588e-01
4.72021997e-02 2.19147041e-01 6.39798820e-01 -4.40820843e-01
-5.88233411e-01 4.91976663e-02 -4.60716009e-01 4.03238162e-02
1.30935475e-01 -1.13406740e-01 -8.76939714e-01 -5.89237511e-01
-2.16347694e-01 -5.16222894e-01 7.80246735e-01 2.55447954e-01
1.51260591e+00 -2.58339345e-01 -6.23932719e-01 6.82540536e-01
1.33659971e+00 -3.91878515e-01 5.93008459e-01 -3.84728134e-01
1.03976536e+00 5.69245517e-01 6.23480916e-01 3.45509470e-01
5.96444964e-01 6.90296948e-01 3.90345961e-01 -1.50035232e-01
-3.65375310e-01 -3.92766476e-01 3.02842796e-01 1.02807164e+00
-1.82615280e-01 -2.18948752e-01 -8.21471334e-01 3.17915767e-01
-1.65708303e+00 -7.79330850e-01 -1.64487496e-01 2.04255819e+00
8.56600583e-01 1.57578930e-01 3.01885068e-01 8.11132789e-02
8.88827801e-01 1.51507646e-01 -3.56600136e-01 2.89595634e-01
1.64632782e-01 2.31481344e-01 1.08839452e-01 6.11529231e-01
-1.02025175e+00 1.24358571e+00 5.50298119e+00 1.24308574e+00
-8.26420844e-01 2.59893864e-01 7.41467953e-01 1.32567836e-02
-8.37064907e-02 7.89604932e-02 -1.10922337e+00 5.00228405e-01
9.53472853e-01 4.10419822e-01 5.98461688e-01 8.05525124e-01
-8.99422988e-02 1.40321087e-02 -9.03636158e-01 1.20791256e+00
2.34051690e-01 -1.07713401e+00 -4.72629163e-03 3.86838347e-01
6.31589472e-01 -1.91274956e-02 -1.96826041e-01 3.14121693e-01
2.52503932e-01 -1.23956609e+00 2.03665793e-02 6.64660037e-02
1.07172239e+00 -8.15737367e-01 6.00047350e-01 4.04722393e-01
-1.72646964e+00 7.27493092e-02 -3.77996802e-01 1.79471016e-01
-6.94270432e-02 7.46652365e-01 -5.60580611e-01 6.19286478e-01
8.64107311e-01 7.13117599e-01 -8.22368860e-01 5.57244956e-01
-3.14064890e-01 9.50493097e-01 -5.06674886e-01 2.40087539e-01
-4.01329808e-02 -3.87209773e-01 1.25866219e-01 9.79327321e-01
7.13040903e-02 3.26710194e-01 5.41460276e-01 4.55326825e-01
-5.76189041e-01 7.42553771e-02 -3.97131503e-01 -1.29728422e-01
7.37032175e-01 1.78018284e+00 -1.23842788e+00 -1.85798690e-01
-5.61002791e-01 1.04083073e+00 6.47750437e-01 1.89224958e-01
-9.75832760e-01 -7.48402253e-02 4.56062257e-01 2.43753016e-01
5.09062469e-01 -2.02847034e-01 9.39026251e-02 -1.11528933e+00
2.15840898e-02 -1.15215313e+00 6.72226131e-01 -4.04032290e-01
-1.26413095e+00 6.65258706e-01 -4.05870713e-02 -6.69359028e-01
3.59878987e-02 -5.66700101e-01 -4.92055982e-01 1.82843402e-01
-1.29393816e+00 -1.56020284e+00 -5.94785571e-01 8.78135085e-01
3.71401101e-01 -5.41503355e-02 8.34169805e-01 5.73643625e-01
-7.73529410e-01 9.70070601e-01 -2.11268708e-01 4.34492201e-01
7.21784532e-01 -1.09558320e+00 2.96820432e-01 8.32960963e-01
4.81741220e-01 7.03938663e-01 1.66228786e-01 -7.66459346e-01
-1.49812078e+00 -1.17427945e+00 7.14819670e-01 -5.63349485e-01
3.40412140e-01 -7.61187673e-01 -9.05173063e-01 5.53104222e-01
-1.14403836e-01 4.81553465e-01 6.64885700e-01 5.04193604e-01
-6.23455226e-01 -2.89226234e-01 -1.20810258e+00 4.98007149e-01
1.58007836e+00 -5.75791359e-01 -4.73302864e-02 4.70455676e-01
5.99556088e-01 -1.46974683e-01 -1.19777727e+00 4.38318938e-01
3.54043782e-01 -1.00768638e+00 9.52654779e-01 -2.53071994e-01
1.77473828e-01 -2.45588869e-01 -4.14650589e-02 -7.96170115e-01
-3.67431492e-01 -7.28388965e-01 -2.80958295e-01 1.67264020e+00
2.96974331e-01 -6.12433791e-01 1.27743185e+00 3.57582420e-01
1.20958991e-01 -1.03463054e+00 -8.31601679e-01 -6.57748044e-01
-3.75910029e-02 -2.81463921e-01 7.57602990e-01 1.08272016e+00
-1.82203859e-01 6.20816886e-01 -2.94577122e-01 3.28543603e-01
9.80927169e-01 2.33784422e-01 8.83465767e-01 -1.44020605e+00
-2.21677333e-01 -2.99664497e-01 -3.26711446e-01 -1.10617495e+00
4.33198214e-01 -1.07832360e+00 -1.22996852e-01 -1.41088748e+00
5.07965803e-01 -3.88817459e-01 -9.04637426e-02 7.73355246e-01
-3.76398593e-01 7.12580144e-01 6.08517379e-02 1.42414674e-01
-8.92299831e-01 4.51378047e-01 1.23340595e+00 7.20855668e-02
1.37235790e-01 -1.69659570e-01 -7.75130153e-01 7.15712428e-01
7.83282876e-01 -5.25210083e-01 -6.21704698e-01 -1.34314775e-01
-6.51225373e-02 -9.44364667e-02 2.30645329e-01 -9.32015836e-01
3.91698629e-01 -1.61380824e-02 3.72819036e-01 -4.67744946e-01
3.56369436e-01 -7.53549278e-01 1.55597478e-01 3.40491414e-01
1.23070292e-01 -1.27459019e-01 -7.80398026e-02 6.54904723e-01
-1.98661417e-01 1.86432034e-01 8.44960749e-01 -1.08917497e-01
-7.14302480e-01 8.78934681e-01 2.82962829e-01 3.09013546e-01
8.74325931e-01 -4.41126749e-02 -1.60095364e-01 -3.50820422e-01
-6.39087498e-01 3.66468549e-01 4.83549774e-01 3.21871012e-01
5.86834788e-01 -1.26527739e+00 -7.01352656e-01 2.20600337e-01
-8.22895095e-02 1.30554587e-01 2.56904930e-01 1.01232171e+00
-2.34853640e-01 5.43103963e-02 2.48634145e-01 -7.76093900e-01
-1.66967010e+00 6.81607962e-01 7.06669390e-02 -3.62299085e-01
-4.96522039e-01 1.08530140e+00 1.10636465e-01 -5.83798051e-01
2.41788998e-01 1.98263630e-01 -1.41594589e-01 9.96993110e-02
4.44330096e-01 2.04343587e-01 -7.92644694e-02 -8.27551484e-01
-6.01056397e-01 1.00411916e+00 6.54648319e-02 2.00991720e-01
1.31626058e+00 -1.95850879e-01 -5.13481259e-01 -1.09638467e-01
1.34712732e+00 8.02725255e-02 -1.09669375e+00 -2.09634990e-01
1.19469337e-01 -5.26520550e-01 -7.42867664e-02 -5.47179580e-01
-1.58609521e+00 8.08389068e-01 3.50754410e-01 -1.19140893e-02
1.36696732e+00 3.29845458e-01 7.71994710e-01 1.91431075e-01
4.41923708e-01 -9.94747519e-01 3.12904745e-01 1.57789871e-01
6.58922076e-01 -1.47417760e+00 2.27746829e-01 -1.09737706e+00
-4.55000103e-01 7.68960834e-01 8.62310648e-01 2.28912264e-01
9.14388478e-01 3.05190921e-01 -1.25629544e-01 -6.32793665e-01
-6.05934918e-01 -2.36750141e-01 4.12758291e-01 4.25760329e-01
3.98188174e-01 8.90611019e-03 7.70692825e-02 4.72785234e-01
-5.55072352e-02 -1.99286610e-01 -1.79497719e-01 4.83142257e-01
-2.40849748e-01 -1.33687055e+00 -1.64382100e-01 4.57261473e-01
-4.57552701e-01 2.70004570e-02 -4.63894010e-01 8.62682998e-01
1.74143031e-01 1.30917835e+00 -1.14919603e-01 -5.69763958e-01
-1.03023782e-01 2.44456530e-02 6.37502968e-01 -6.07450604e-01
-2.00852841e-01 2.43153140e-01 1.19340539e-01 -8.61786485e-01
-6.21079564e-01 -5.95108867e-01 -1.38675344e+00 -6.97914302e-01
-5.39717436e-01 3.62787694e-01 4.47737217e-01 9.09875333e-01
5.86875081e-01 1.97063357e-01 7.98229277e-01 -6.90939069e-01
-1.48271233e-01 -9.64007497e-01 -5.49110532e-01 4.83617872e-01
-1.80724695e-01 -5.70964336e-01 -3.65222275e-01 -1.52589707e-02] | [13.509328842163086, 0.9461907744407654] |
f004fdd8-098a-4df6-9cb9-d0a931667c24 | remote-task-oriented-grasp-area-teaching-by | 2303.10195 | null | https://arxiv.org/abs/2303.10195v1 | https://arxiv.org/pdf/2303.10195v1.pdf | Remote Task-oriented Grasp Area Teaching By Non-Experts through Interactive Segmentation and Few-Shot Learning | A robot operating in unstructured environments must be able to discriminate between different grasping styles depending on the prospective manipulation task. Having a system that allows learning from remote non-expert demonstrations can very feasibly extend the cognitive skills of a robot for task-oriented grasping. We propose a novel two-step framework towards this aim. The first step involves grasp area estimation by segmentation. We receive grasp area demonstrations for a new task via interactive segmentation, and learn from these few demonstrations to estimate the required grasp area on an unseen scene for the given task. The second step is autonomous grasp estimation in the segmented region. To train the segmentation network for few-shot learning, we built a grasp area segmentation (GAS) dataset with 10089 images grouped into 1121 segmentation tasks. We benefit from an efficient meta learning algorithm for training for few-shot adaptation. Experimental evaluation showed that our method successfully detects the correct grasp area on the respective objects in unseen test scenes and effectively allows remote teaching of new grasp strategies by non-experts. | ['Eckehard Steinbach', 'Shaobo Zhou', 'Sudarshan Rajagopalan', 'Furkan Kaynar'] | 2023-03-17 | null | null | null | null | ['interactive-segmentation'] | ['computer-vision'] | [ 4.36891019e-01 1.38381585e-01 6.28461763e-02 -5.63367009e-01
-7.03536868e-01 -6.16886437e-01 1.65899694e-01 1.42538734e-02
-5.45264840e-01 4.09556806e-01 -6.01468801e-01 4.32465598e-02
-2.36019507e-01 -7.63461113e-01 -1.14084697e+00 -8.86968911e-01
-3.48826855e-01 1.08894408e+00 6.10375822e-01 -2.17151895e-01
6.27408385e-01 7.45351195e-01 -1.85124469e+00 3.09328824e-01
1.13340235e+00 8.63897264e-01 1.40700483e+00 8.38988245e-01
-1.61235586e-01 2.73497045e-01 -5.38173437e-01 4.82557088e-01
5.29223144e-01 -7.72709325e-02 -8.97852600e-01 2.42677689e-01
1.57330513e-01 -7.47223616e-01 6.37760088e-02 1.01583457e+00
1.89986408e-01 5.16047299e-01 9.65366244e-01 -1.17724240e+00
-2.75388777e-01 9.93555963e-01 -2.19564468e-01 8.81190300e-02
3.53258729e-01 5.76432168e-01 4.99427646e-01 -8.33148241e-01
6.90169573e-01 1.22263646e+00 4.08120751e-02 7.02500939e-01
-9.16402817e-01 -2.89604843e-01 8.46417472e-02 2.67439127e-01
-7.75904417e-01 9.76090357e-02 6.32570684e-01 -4.89806116e-01
9.43637609e-01 -1.86684579e-01 8.52591157e-01 1.12510896e+00
8.79870802e-02 1.08913636e+00 1.27542353e+00 -4.84464586e-01
7.10192919e-01 -5.25473021e-02 1.39463007e-01 8.26934159e-01
-5.06334342e-02 1.62054852e-01 -1.45872846e-01 4.41077292e-01
1.06607234e+00 2.15229571e-01 -6.51094168e-02 -1.07921147e+00
-1.38869202e+00 4.90181088e-01 8.10343444e-01 5.21017909e-01
-7.44058013e-01 1.35575995e-01 2.64712870e-01 6.58410132e-01
-2.33806461e-01 8.40490997e-01 -8.26232910e-01 -2.84330435e-02
-4.92549092e-01 1.26774460e-01 9.62789536e-01 1.50032508e+00
9.16439652e-01 -9.54614058e-02 -2.07301393e-01 4.37577605e-01
-6.56200796e-02 6.15471184e-01 2.92406976e-01 -1.35332477e+00
3.49002093e-01 4.62963849e-01 4.30762351e-01 -2.24084407e-01
-3.13494027e-01 3.92623931e-01 -7.25011528e-02 6.45989478e-01
7.15870976e-01 -1.04203478e-01 -1.52553535e+00 1.14862609e+00
3.76473159e-01 -3.89887750e-01 -2.04780195e-02 8.41635466e-01
5.02773702e-01 4.65081632e-01 2.20713705e-01 -4.14715558e-02
1.09437454e+00 -1.26641548e+00 -2.78597236e-01 -2.96036422e-01
4.73069400e-01 -2.93600172e-01 1.32577395e+00 6.88670695e-01
-9.01188850e-01 -8.81599545e-01 -9.63653266e-01 2.81324834e-01
-6.02793813e-01 2.51542538e-01 8.61016572e-01 2.33816624e-01
-8.31494689e-01 1.02889514e+00 -9.74045694e-01 -6.88563764e-01
4.42514300e-01 4.43615735e-01 -1.26473993e-01 -3.09257567e-01
-5.04424572e-01 1.08096230e+00 1.03475988e+00 4.93030362e-02
-1.69874299e+00 -2.97331750e-01 -7.91810095e-01 2.64834523e-01
8.99919152e-01 -2.62442261e-01 1.29997516e+00 -7.31067419e-01
-1.58531213e+00 8.10677469e-01 4.37227398e-01 -2.22957835e-01
4.98522162e-01 -3.36864471e-01 6.10953510e-01 5.07450819e-01
2.01693803e-01 1.13109446e+00 1.29143405e+00 -1.74096739e+00
-7.15842307e-01 -4.50678349e-01 2.35548750e-01 3.03286135e-01
1.42759725e-01 -4.18705732e-01 -5.93362860e-02 -1.10680453e-01
4.46428895e-01 -9.11489189e-01 -1.22397333e-01 -8.31381828e-02
-1.12549655e-01 -5.65558314e-01 8.60282898e-01 -3.92677158e-01
-8.85530785e-02 -1.84910119e+00 6.90612674e-01 7.70769194e-02
-1.33267820e-01 1.41985044e-01 -3.51240933e-01 3.24228585e-01
1.82118326e-01 -5.55249095e-01 -2.72243679e-01 1.33140311e-01
3.45348530e-02 2.95493960e-01 -2.95539409e-01 1.20348021e-01
1.52605376e-03 9.71099854e-01 -1.27775216e+00 -4.79212463e-01
5.91050386e-01 -1.62082225e-01 -2.49855891e-01 8.66821706e-01
-6.57682180e-01 7.85271406e-01 -7.82916427e-01 1.13041830e+00
4.17860836e-01 1.38913468e-01 1.93570480e-01 1.05687445e-02
-7.05257133e-02 -4.21669424e-01 -8.46006930e-01 2.31949067e+00
-5.16736865e-01 1.27252758e-01 4.03956860e-01 -1.25689864e+00
1.16271341e+00 6.25655204e-02 4.32180107e-01 -2.53449172e-01
4.05482829e-01 5.93170762e-01 1.25381380e-01 -1.12869561e+00
1.48093596e-01 5.99005818e-02 -7.12764729e-03 2.94701308e-01
6.38772666e-01 -9.48595703e-01 2.23860636e-01 -2.05647439e-01
1.08947492e+00 7.87672460e-01 2.71726754e-02 -2.86289990e-01
-3.57321575e-02 3.70403618e-01 -9.97599438e-02 1.08676064e+00
-5.55976927e-01 2.64464706e-01 -1.10758357e-01 -5.57096064e-01
-9.25943434e-01 -1.26520276e+00 2.43962258e-01 1.52045238e+00
5.28922737e-01 6.26476228e-01 -8.75446439e-01 -6.93658113e-01
1.65974407e-03 8.95412326e-01 -6.08153224e-01 7.65917897e-02
-7.79102087e-01 -1.26134694e-01 -2.91027337e-01 7.18051791e-01
5.65763354e-01 -2.06896043e+00 -1.64930868e+00 2.25853547e-01
8.36478919e-02 -8.12593818e-01 1.85585469e-01 8.49625647e-01
-1.10058963e+00 -1.30854356e+00 -1.07924652e+00 -1.42303097e+00
9.99653459e-01 5.81550002e-01 6.47239685e-01 5.18373065e-02
-6.73873603e-01 9.23506320e-01 -7.24314809e-01 -5.55494070e-01
-4.43072498e-01 1.73741579e-01 5.78428581e-02 -8.10171425e-01
4.33827102e-01 -5.37797034e-01 -3.79573852e-01 2.10737959e-01
-5.25358260e-01 -1.25877246e-01 1.18001258e+00 8.03399920e-01
4.03596729e-01 -7.31825083e-02 5.63350379e-01 -3.38402480e-01
5.47830284e-01 -1.24863751e-01 -6.05557203e-01 7.45290697e-01
-2.88118690e-01 1.12025224e-01 2.57085651e-01 -5.41310191e-01
-1.11073542e+00 4.09260482e-01 2.69939154e-01 -4.61935252e-01
-7.00304329e-01 1.04476914e-01 -7.20285345e-03 -1.39766037e-01
5.63989222e-01 2.12336764e-01 -3.06190271e-02 -2.83583522e-01
4.59845394e-01 4.89637166e-01 5.69601476e-01 -1.05205655e+00
5.83569109e-01 9.08332020e-02 -3.26129615e-01 -8.23526978e-01
-5.56636333e-01 -7.53401697e-01 -1.35337329e+00 -4.30357426e-01
1.11551213e+00 -4.83024538e-01 -1.00860095e+00 6.68806076e-01
-1.14471507e+00 -1.18420577e+00 -5.08394539e-01 3.84180665e-01
-1.28062642e+00 1.74588457e-01 -4.25193131e-01 -8.98282111e-01
-2.12078720e-01 -1.43277621e+00 1.20699453e+00 3.01493376e-01
2.40927115e-01 -3.95391971e-01 -2.86430806e-01 1.99808896e-01
2.52771258e-01 7.96542168e-02 8.91045630e-01 -7.92657256e-01
-8.60709906e-01 -7.35395849e-02 -2.50519156e-01 1.23403110e-01
1.90954819e-01 -3.85975182e-01 -8.66728544e-01 -6.92324102e-01
2.95699060e-01 -1.02967131e+00 9.47493017e-01 5.29352844e-01
1.25829518e+00 2.16852278e-01 -4.81601775e-01 1.82687432e-01
1.30051064e+00 5.11568487e-01 4.54835594e-01 4.12914872e-01
6.15124941e-01 8.78527939e-01 1.53013492e+00 1.09934568e-01
-2.56596029e-01 3.49330157e-01 7.29534686e-01 4.58596408e-01
1.65312216e-01 7.83898309e-02 2.86079437e-01 5.03626227e-01
-4.95505840e-01 -3.86952125e-02 -9.26472247e-01 6.22249365e-01
-1.73172641e+00 -8.66449237e-01 4.00176406e-01 1.92872179e+00
5.39342940e-01 2.42319331e-01 1.18760904e-02 -1.59802571e-01
5.56867838e-01 -2.02978641e-01 -9.75605965e-01 -2.37886131e-01
5.07418752e-01 2.40501150e-01 3.98360252e-01 3.01414400e-01
-1.00914240e+00 1.43621838e+00 5.53349066e+00 5.92237353e-01
-6.67717576e-01 -1.47637114e-01 6.45395666e-02 3.87727082e-01
2.80836910e-01 -1.57465070e-01 -5.11833847e-01 8.73249322e-02
3.22910011e-01 3.37786883e-01 8.29751253e-01 1.38868916e+00
-3.12734187e-01 -6.93544507e-01 -1.47420681e+00 5.54089427e-01
3.31798531e-02 -6.69100404e-01 -1.17063485e-01 -4.39401329e-01
6.09369159e-01 4.33939360e-02 -1.92283928e-01 8.79383326e-01
3.89417917e-01 -8.26256394e-01 5.40315568e-01 4.61360484e-01
5.97871780e-01 -3.16456020e-01 5.37570715e-01 8.41713965e-01
-9.43549156e-01 -5.45383096e-01 -6.59616351e-01 7.58079812e-02
3.79543868e-03 -2.87427127e-01 -1.40099573e+00 2.44885951e-01
7.85505176e-01 2.32515246e-01 -2.65851945e-01 1.03717315e+00
-4.38931048e-01 -5.89711592e-02 -1.40547842e-01 -4.48264599e-01
3.90553296e-01 -3.32532257e-01 6.41301751e-01 7.34229445e-01
3.11888099e-01 3.54739994e-01 5.44406950e-01 1.06875479e+00
3.97688925e-01 -3.17635596e-01 -7.19862998e-01 -2.13785819e-03
5.76369226e-01 1.04670382e+00 -1.30707908e+00 -4.84993041e-01
1.22297063e-01 1.31932044e+00 4.03557271e-01 3.30773711e-01
-3.71120393e-01 -5.80735564e-01 -3.23365867e-01 -3.83830726e-01
6.46745920e-01 -4.15539771e-01 8.54057744e-02 -8.68675947e-01
-7.13330954e-02 -8.55101347e-01 6.31126575e-03 -1.07966423e+00
-8.89103174e-01 2.79496133e-01 3.57750684e-01 -9.40608084e-01
-2.13865384e-01 -1.04984856e+00 -4.94167387e-01 4.75917011e-01
-1.30658972e+00 -1.19309449e+00 -9.10609782e-01 3.13506126e-01
1.36982203e+00 -3.02600890e-01 8.69857371e-01 -5.02976954e-01
9.36439335e-02 -1.47934139e-01 -1.97729215e-01 -1.39119402e-01
5.35715938e-01 -1.60196364e+00 9.61214229e-02 3.40727061e-01
-1.57427967e-01 4.59819347e-01 9.07685518e-01 -8.46037924e-01
-1.64476454e+00 -5.07979572e-01 -1.03513911e-01 -4.89328921e-01
4.44021434e-01 -2.66880721e-01 -9.92338061e-01 7.94218481e-01
3.36858593e-02 -3.25546056e-01 -9.68666077e-02 -1.01402812e-01
1.33982018e-01 1.93957880e-01 -1.31750810e+00 3.33862722e-01
9.50652063e-01 -2.65909173e-02 -1.25755179e+00 6.01462603e-01
6.84680820e-01 -6.17367148e-01 -6.94455564e-01 6.06438875e-01
6.06327295e-01 -7.82148361e-01 9.04372633e-01 -7.19780684e-01
3.38363439e-01 1.34963453e-01 -2.33615696e-01 -1.42007470e+00
-3.74172404e-02 -2.83749521e-01 -8.23318139e-02 4.87717420e-01
5.07556610e-02 -1.93897396e-01 9.05758023e-01 4.80135471e-01
-4.45876569e-01 -5.78144491e-01 -4.92276639e-01 -1.03056610e+00
1.10974990e-01 7.50959367e-02 2.50323743e-01 5.07307768e-01
1.29180014e-01 -1.39735237e-01 2.12025538e-01 1.92638323e-01
8.72110784e-01 6.24045253e-01 1.02567172e+00 -1.37934494e+00
-2.38979816e-01 -1.95922956e-01 -1.01457804e-01 -1.31427503e+00
1.92138195e-01 -6.96834445e-01 1.00878525e+00 -1.72073627e+00
3.93159658e-01 -6.44131541e-01 -5.55442199e-02 5.74332237e-01
-7.20700547e-02 -5.01760304e-01 2.90509522e-01 3.34210455e-01
-7.24350393e-01 4.39428955e-01 1.81172693e+00 -3.92328441e-01
-5.62375724e-01 2.26788342e-01 1.60518974e-01 7.04064667e-01
8.65959585e-01 -2.95773655e-01 -4.28627282e-01 -4.01034415e-01
-4.25793201e-01 4.08669829e-01 4.06466424e-01 -1.21727586e+00
1.73627064e-01 -4.44588482e-01 2.74493963e-01 -6.76025629e-01
4.06728089e-01 -1.09927630e+00 -5.81905365e-01 9.57578361e-01
-4.19343829e-01 -4.49015319e-01 1.45530835e-01 9.06935155e-01
3.12834144e-01 -8.71494174e-01 4.65909839e-01 -7.81678200e-01
-1.40889096e+00 6.78057149e-02 -2.88846493e-01 -3.34076136e-01
1.49159503e+00 -4.75411296e-01 -9.67952833e-02 7.52814710e-02
-1.32551718e+00 5.17006159e-01 5.50453544e-01 3.51833373e-01
8.73205602e-01 -6.42789721e-01 -3.31144333e-01 1.97294727e-01
1.28830001e-01 3.12634528e-01 2.19364718e-01 5.63551843e-01
-6.17058158e-01 1.21531025e-01 -9.49907005e-01 -8.09720755e-01
-1.02796412e+00 9.68979895e-01 1.24447159e-01 1.72408685e-01
-6.94427907e-01 1.13771141e+00 1.07299136e-02 -6.77846014e-01
6.65862918e-01 -7.05127716e-01 -5.27458191e-02 -1.30899653e-01
1.51926339e-01 4.51727092e-01 -2.71351069e-01 2.57889330e-02
1.18827119e-01 4.76943493e-01 -7.96601623e-02 3.41019817e-02
1.55078328e+00 1.26489505e-01 1.38598997e-02 8.28211486e-01
7.99026132e-01 -7.94198573e-01 -1.80581605e+00 1.65749177e-01
7.01657832e-02 -4.05556530e-01 -6.00747705e-01 -1.03455734e+00
-6.04121447e-01 1.05359030e+00 7.28150725e-01 2.45287433e-01
7.81447411e-01 2.30630800e-01 4.77584898e-01 1.26492310e+00
1.25205147e+00 -1.43359900e+00 7.67537475e-01 5.09425461e-01
1.08143365e+00 -1.76879251e+00 -1.18373170e-01 -5.09588301e-01
-6.05403721e-01 1.36413980e+00 1.02550709e+00 -4.92050856e-01
2.37622514e-01 -1.14259444e-01 -1.15600385e-01 -7.51460433e-01
-1.41230747e-01 -9.72216651e-02 8.42306390e-03 9.89675462e-01
-2.84116298e-01 1.52072683e-01 2.30214044e-01 7.48423999e-03
1.25183761e-01 -4.17041965e-02 4.81570750e-01 1.43459427e+00
-1.30295563e+00 -6.21077120e-01 -1.55625656e-01 5.36877215e-01
3.96512985e-01 4.39677894e-01 -3.13144386e-01 9.04043794e-01
4.80091758e-02 6.35319591e-01 -3.18678953e-02 -9.85660404e-02
2.28427097e-01 2.22704917e-01 1.29099834e+00 -1.24899805e+00
-3.65728885e-01 -1.40011817e-01 -4.24562514e-01 -6.45528495e-01
-4.93179351e-01 -7.60069728e-01 -1.65272987e+00 4.80224758e-01
-4.93375778e-01 7.30577409e-02 9.13378835e-01 1.10688806e+00
-2.11702526e-01 7.17730463e-01 4.77989793e-01 -1.71397603e+00
-1.10370600e+00 -1.25687420e+00 -7.75505066e-01 2.56861001e-01
5.53952381e-02 -1.09944475e+00 -3.74641567e-01 3.18967327e-02] | [5.76169490814209, -0.766557514667511] |
da654107-0e10-4485-a334-0d5372475bb2 | novel-predator-prey-model-admitting-exact | 2208.02457 | null | https://arxiv.org/abs/2208.02457v2 | https://arxiv.org/pdf/2208.02457v2.pdf | Novel predator-prey model admitting exact analytical solution | The Lotka-Volterra predator-prey model still represents the paradigm for the description of the competition in population dynamics. Despite its extreme simplicity, it does not admit an analytical solution, and for this reason, numerical integration methods are usually adopted to apply it to various fields of science. The aim of the present work is to investigate the existence of new predator-prey models sharing the broad features of the standard Lotka-Volterra model and, at the same time, offer the advantage of possessing exact analytical solutions. To this purpose, a general Hamiltonian formalism, which is suitable for treating a large class of predator-prey models in population dynamics within the same framework, has been developed as a first step. The only existing model having the property of admitting a simple exact analytical solution, is identified within the above class of models. The solution of this special predator-prey model is obtained explicitly, in terms of known elementary functions, and its main properties are studied. Finally, the generalization of this model, based on the concept of power-law competition, as well as its extension to the case of $N$-component competition systems, are considered. | ['G. Kaniadakis'] | 2022-08-04 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-2.86207408e-01 -4.15015072e-01 1.90682039e-01 1.74498379e-01
2.47001931e-01 -3.67670298e-01 5.56826711e-01 4.47648793e-01
-5.81247330e-01 9.57652271e-01 -7.39293814e-01 -8.15155432e-02
-6.04397297e-01 -8.28287601e-01 -2.40652859e-01 -1.21128488e+00
-3.49271357e-01 4.43118155e-01 2.83826947e-01 -7.08469868e-01
-2.61057280e-02 7.44502306e-01 -1.89108968e+00 -7.58460522e-01
8.83784115e-01 9.36526239e-01 1.36157140e-01 1.71808228e-01
-2.87035443e-02 2.90279359e-01 -5.15288651e-01 -3.48021835e-01
4.53158207e-02 -8.02263856e-01 -2.88976520e-01 -2.91344468e-02
-5.25325358e-01 1.12474822e-01 -6.41315058e-02 9.74118650e-01
3.16243500e-01 1.50719762e-01 9.75132465e-01 -1.20894229e+00
-4.63566482e-01 3.46996546e-01 -2.23839507e-01 3.83105457e-01
1.84911609e-01 -1.01172570e-02 8.16065729e-01 -4.69751686e-01
4.42731500e-01 1.06565702e+00 3.67143989e-01 4.89562869e-01
-1.42674196e+00 -3.15337360e-01 -1.39851213e-01 1.86436191e-01
-1.91916144e+00 7.12895393e-02 6.32939339e-01 -5.31358659e-01
8.33876848e-01 1.90670878e-01 1.21619964e+00 6.33351922e-01
7.14085817e-01 2.18308151e-01 1.20185924e+00 -4.35116708e-01
5.81007421e-01 2.28832945e-01 2.25790784e-01 2.24038705e-01
9.11951602e-01 8.61749500e-02 7.25240037e-02 -4.07740951e-01
4.77119505e-01 4.15315330e-02 -3.42685640e-01 -7.04763532e-01
-4.85064059e-01 8.89782608e-01 4.48547602e-01 9.55185115e-01
-3.64555568e-01 -1.86532632e-01 2.06040338e-01 9.14803073e-02
3.17718625e-01 2.03362286e-01 1.00552216e-01 3.93305153e-01
-5.08525074e-01 7.32056975e-01 1.16649425e+00 4.19598401e-01
6.50089502e-01 3.15237999e-01 4.06727999e-01 3.47820401e-01
3.08976263e-01 7.81809270e-01 4.12564695e-01 -3.63430083e-01
-2.31050625e-01 7.21037269e-01 2.79639959e-01 -8.33985567e-01
-4.56761390e-01 -5.74510634e-01 -9.30039823e-01 4.83150929e-01
4.42448467e-01 7.94676617e-02 -9.91270319e-02 1.75264919e+00
4.02659714e-01 -1.33563876e-01 4.51503277e-01 7.07101464e-01
3.51508051e-01 6.88112020e-01 1.15537569e-01 -8.96889031e-01
1.18941653e+00 -1.00527272e-01 -6.85394764e-01 4.45283175e-01
3.13253045e-01 -5.28560221e-01 5.61031640e-01 3.52409154e-01
-8.74933422e-01 -3.94448154e-02 -8.24719131e-01 6.41355038e-01
-6.46142900e-01 -1.80833921e-01 3.04331392e-01 5.06309986e-01
-1.08340609e+00 5.52966177e-01 -5.69261909e-01 -7.81608522e-01
-4.12812263e-01 3.62359017e-01 -1.74980313e-01 1.53277189e-01
-1.24302900e+00 1.08654118e+00 2.23582163e-01 3.38747442e-01
-4.65240926e-01 -1.80500105e-01 -6.41918242e-01 4.66314346e-01
2.82954961e-01 -4.98751611e-01 9.71271634e-01 -1.01246047e+00
-1.39717233e+00 6.63590789e-01 9.68150273e-02 -4.13100749e-01
4.87124741e-01 6.29103959e-01 -2.81381667e-01 1.90149583e-02
-5.21377362e-02 2.88607534e-02 4.23842818e-01 -1.03921807e+00
-1.88237950e-01 -5.03632963e-01 2.29858637e-01 7.37543777e-02
-2.40015104e-01 -3.16034481e-02 4.28339541e-01 -3.39498699e-01
-3.52689147e-01 -7.78515756e-01 -3.47177267e-01 -1.65845305e-01
1.83342472e-01 -2.78660417e-01 5.61706007e-01 9.01199654e-02
1.15513885e+00 -2.08967137e+00 6.18217587e-01 3.84295195e-01
-3.41640450e-02 4.00883704e-01 1.74104229e-01 1.15540850e+00
9.98775885e-02 -2.61694103e-01 -6.13382161e-01 4.03710939e-02
-8.94294009e-02 3.39341700e-01 -6.16807379e-02 7.69013703e-01
2.71987796e-01 3.36806476e-01 -6.26050174e-01 -6.24243915e-01
3.43419880e-01 6.92276478e-01 -5.32630324e-01 -1.45132601e-01
-1.07664958e-01 3.20830196e-01 -5.62911272e-01 1.72694728e-01
6.75121367e-01 9.35513079e-02 3.16670120e-01 3.15198720e-01
-9.13971364e-01 -4.29348826e-01 -1.32765865e+00 6.00913048e-01
-7.38465488e-02 1.70009822e-01 5.59836030e-01 -1.41091442e+00
1.04203928e+00 5.40071666e-01 8.80827725e-01 -4.19944376e-01
5.83853841e-01 7.26924777e-01 6.17754996e-01 -1.53016225e-01
1.05822265e-01 -5.14508903e-01 -2.32585028e-01 1.23934828e-01
-9.79410410e-02 -7.49084279e-02 6.82947814e-01 -2.69567966e-01
6.70320809e-01 -3.69683921e-01 8.28958273e-01 -8.79422247e-01
1.01890481e+00 1.30208582e-01 2.91276813e-01 2.55113989e-01
-2.12703228e-01 -2.26602241e-01 5.88225424e-01 -3.06158960e-01
-8.57957840e-01 -8.70840907e-01 -5.50489724e-01 2.09641472e-01
4.06654209e-01 3.44624445e-02 -8.07002366e-01 4.90260094e-01
2.97600657e-01 4.50196028e-01 -5.60241163e-01 -2.90764302e-01
-4.54152226e-01 -1.16981721e+00 2.69112259e-01 -3.28841209e-01
3.31733346e-01 -1.10748446e+00 -9.71100569e-01 4.87836242e-01
1.42395243e-01 -5.55165529e-01 3.20079863e-01 3.01061749e-01
-8.51896524e-01 -1.23396277e+00 -1.11022055e+00 -4.81575966e-01
2.54300624e-01 3.05508703e-01 6.23435080e-01 2.67241657e-01
-5.16797483e-01 4.70686167e-01 -2.56962329e-01 -2.53012031e-01
-6.46716833e-01 -1.85936391e-01 3.44079286e-01 1.85005516e-01
4.36501116e-01 -6.00628912e-01 -3.12168270e-01 3.99015397e-01
-1.13609982e+00 -7.02663064e-01 4.18911904e-01 6.97638810e-01
5.04781783e-01 4.17136639e-01 8.82777035e-01 -8.82352069e-02
6.58779740e-01 -5.59635878e-01 -1.24440825e+00 -4.54173908e-02
-4.99337196e-01 -1.28479317e-01 1.18176734e+00 -2.57668316e-01
-7.37790525e-01 2.05957312e-02 -2.67934054e-01 2.62018405e-02
-1.15247425e-02 4.52082932e-01 -3.16643178e-01 -5.18929303e-01
4.20037657e-01 5.56359231e-01 2.55152017e-01 -6.45040154e-01
-1.07370615e-01 4.44704354e-01 1.18101202e-01 -1.94324061e-01
6.34761453e-01 4.28773731e-01 5.72737515e-01 -1.15243781e+00
1.39259910e-02 -4.75209624e-01 -5.17394066e-01 -2.21350849e-01
8.07852745e-01 -4.22807723e-01 -1.19601166e+00 7.01729655e-01
-1.07631910e+00 2.00189068e-03 -5.68340421e-01 5.75636446e-01
-8.36474597e-01 4.42405462e-01 -4.73697364e-01 -1.49261749e+00
-6.70865700e-02 -1.08937347e+00 4.70745027e-01 3.50410134e-01
3.41308683e-01 -8.91586781e-01 3.66964728e-01 -4.78142053e-01
4.20915961e-01 1.49565145e-01 9.16819155e-01 -4.94083852e-01
-4.08116221e-01 -2.91687459e-01 2.20907077e-01 1.27250835e-01
-9.10347700e-02 1.34439781e-01 -2.15842918e-01 -2.63193846e-01
4.02664840e-01 3.67074013e-01 5.82613170e-01 5.96541584e-01
1.49883449e-01 5.33477850e-02 -3.70207697e-01 2.09571458e-02
1.84973466e+00 5.08807361e-01 2.23396808e-01 1.64523706e-01
-1.28647625e-01 6.73765898e-01 5.40353954e-01 6.53659940e-01
2.23367766e-01 9.55401599e-01 6.41831577e-01 3.80973071e-01
5.62307954e-01 3.20625037e-01 2.13317648e-01 7.42117465e-01
-3.18274528e-01 -4.72428083e-01 -6.06976211e-01 3.70704681e-01
-1.82508779e+00 -1.25968254e+00 -5.17695606e-01 2.51697922e+00
9.88152847e-02 -1.43535048e-01 6.11571968e-01 4.72119957e-01
7.35377789e-01 -1.43465251e-01 -2.40993977e-01 -5.06692767e-01
-4.89398271e-01 -5.00468574e-02 4.44638610e-01 6.04614794e-01
-7.50865579e-01 4.54963446e-01 6.33663607e+00 6.56466365e-01
-1.21982956e+00 1.19906671e-01 -1.43517792e-01 2.19074503e-01
-1.78789790e-03 9.50882793e-04 -8.83883119e-01 5.96467614e-01
8.83060932e-01 -6.50577784e-01 5.06398559e-01 5.84338427e-01
4.78936195e-01 -5.41354835e-01 -3.49912345e-01 7.59654164e-01
-8.94788802e-02 -6.00457191e-01 -1.01838596e-01 5.71515858e-01
2.79642999e-01 -3.54388803e-01 -8.04902092e-02 5.24436273e-02
-2.19027579e-01 -6.06468678e-01 5.18138647e-01 4.14765656e-01
1.52905732e-01 -8.69343102e-01 9.54009473e-01 7.64937639e-01
-1.37943733e+00 -3.55946988e-01 -6.08019114e-01 -5.75443864e-01
5.43877661e-01 3.85221064e-01 1.46468818e-01 5.69071293e-01
3.31852138e-01 4.60764319e-01 -1.57637447e-01 1.56967878e+00
4.49546844e-01 1.99721813e-01 -6.52734220e-01 -5.81632793e-01
3.02837789e-01 -7.31199324e-01 7.32343078e-01 7.63407111e-01
8.60083878e-01 2.14977875e-01 -1.88191861e-01 9.99459147e-01
6.24920547e-01 8.29129457e-01 -8.33925068e-01 -9.13191959e-02
1.26012728e-01 1.00101054e+00 -8.60712290e-01 -1.94831878e-01
-2.88496941e-01 4.00942087e-01 3.34405042e-02 3.41585763e-02
-6.53180957e-01 -2.41675124e-01 7.47303784e-01 3.60325456e-01
4.24832165e-01 -2.55164683e-01 4.05357748e-01 -1.22890496e+00
-1.94611832e-01 -3.24047774e-01 2.99017996e-01 -6.99380115e-02
-1.20565975e+00 6.99649632e-01 6.03301823e-01 -1.39854634e+00
-5.45970201e-01 -8.68160963e-01 -5.55481195e-01 9.20620143e-01
-1.24032128e+00 -6.79073095e-01 1.34051174e-01 7.25110114e-01
3.76004241e-02 -5.16805090e-02 9.02540147e-01 5.24937510e-01
-6.81577623e-01 5.91332279e-02 8.17182720e-01 -6.62022710e-01
3.74720152e-03 -1.02408588e+00 -2.54658014e-01 5.40905297e-01
-4.56674516e-01 5.52501798e-01 1.33249652e+00 -5.93407989e-01
-1.36680567e+00 -3.84610236e-01 1.05777347e+00 2.67308652e-01
9.43617821e-01 -2.81890005e-01 -9.31783497e-01 5.77953979e-02
8.40414315e-02 -9.80838686e-02 4.50885922e-01 -5.12699842e-01
1.97736412e-01 -1.30297542e-01 -1.16125941e+00 5.35581648e-01
6.03597701e-01 -1.36391865e-02 -4.46269453e-01 7.97010362e-02
8.52886140e-02 2.79018283e-01 -8.73406529e-01 8.30510333e-02
5.65941751e-01 -1.17676651e+00 9.53737855e-01 -2.25016505e-01
-2.63574630e-01 -3.22373480e-01 -7.30946288e-02 -1.22696149e+00
-2.55612910e-01 -3.06893378e-01 2.69427747e-01 9.61586297e-01
-1.12712160e-01 -1.28257406e+00 7.43980482e-02 5.57106249e-02
1.35326549e-01 -6.37554049e-01 -1.69956195e+00 -1.24497807e+00
4.13735509e-01 3.55320930e-01 2.52768070e-01 4.48789418e-01
3.45111728e-01 1.60997316e-01 -4.34502661e-01 -1.54596299e-01
8.94516528e-01 1.31107196e-01 3.31882507e-01 -1.74392033e+00
-5.21811426e-01 -7.42323339e-01 -7.84078658e-01 -5.02615988e-01
2.89686590e-01 -5.09804189e-01 4.89724614e-02 -1.17379642e+00
-7.14214146e-02 -2.53883123e-01 -2.16981962e-01 -3.58853608e-01
1.34631515e-01 2.83218741e-01 3.75372082e-01 4.74112213e-01
-1.03065960e-01 6.53311729e-01 7.50173986e-01 1.95011735e-01
-2.54165590e-01 4.00717556e-01 -3.98833990e-01 5.93932211e-01
7.15013742e-01 -5.60287654e-01 -3.84260535e-01 3.59977663e-01
1.68379754e-01 1.89737484e-01 5.58457911e-01 -1.04851592e+00
2.55805608e-02 -1.86578065e-01 -3.54506940e-01 -2.77899295e-01
6.28555834e-01 -1.16099572e+00 7.22127736e-01 1.04856098e+00
1.89131066e-01 2.25311607e-01 5.56765404e-03 6.60056889e-01
-3.45358789e-01 -8.25756490e-01 1.01078033e+00 -2.21033946e-01
-3.85103792e-01 3.03234328e-02 -1.13255775e+00 -2.62162745e-01
1.34483099e+00 -2.72577852e-01 -2.57867992e-01 -1.93746075e-01
-8.62039924e-01 -1.13502465e-04 7.51353443e-01 -2.37187967e-01
7.85948560e-02 -1.12177885e+00 -2.85210460e-01 -1.28179282e-01
1.31165847e-01 -6.98356271e-01 5.86140752e-01 1.11971486e+00
-6.55843318e-01 6.58637226e-01 -6.40794575e-01 -3.98281991e-01
-1.07459283e+00 1.10022295e+00 6.66939855e-01 -3.77059340e-01
-4.34688091e-01 -1.01714484e-01 4.37961400e-01 2.12073773e-02
-2.26820737e-01 -9.18724835e-02 -4.17394459e-01 6.59370795e-02
4.09100354e-01 3.63308102e-01 -3.27007532e-01 -1.42444491e+00
-4.78263110e-01 8.78721297e-01 8.18915784e-01 -7.10911453e-02
1.15321970e+00 -3.47618937e-01 -4.32421893e-01 5.99958599e-01
9.13851500e-01 -7.28961676e-02 -7.21870661e-01 7.77346864e-02
-1.41621768e-01 -2.66715810e-02 -3.15869957e-01 -2.05888495e-01
-5.73390543e-01 9.32010472e-01 4.81670231e-01 1.00421894e+00
1.08780241e+00 -2.78785210e-02 3.42140079e-01 3.25739831e-01
6.51516795e-01 -5.66309929e-01 -6.24533296e-01 2.12290078e-01
8.46753895e-01 -7.96040893e-01 -3.72525334e-01 -5.34888029e-01
-1.07704699e-01 1.17328846e+00 3.06525230e-01 -5.87481141e-01
7.84320116e-01 3.78699780e-01 -3.17832917e-01 4.25277986e-02
-5.30941725e-01 -7.17803299e-01 -2.79087394e-01 6.18849516e-01
2.94569314e-01 1.95742518e-01 -1.47597444e+00 4.05495793e-01
2.64860749e-01 1.64641738e-01 7.27689683e-01 7.98778653e-01
-6.01752877e-01 -1.13137293e+00 -4.91028279e-01 7.69130811e-02
-3.92340362e-01 2.54745364e-01 -5.03857732e-01 1.25108027e+00
4.66790944e-02 8.80593479e-01 5.69504872e-03 1.74445450e-01
5.11966944e-01 -1.05178945e-01 4.32199955e-01 -1.72550350e-01
-7.21078753e-01 2.73368597e-01 -3.69558334e-01 3.15066101e-03
-6.75098121e-01 -6.04198039e-01 -9.98845875e-01 -5.43856919e-01
-5.46554327e-01 9.73072827e-01 4.75631714e-01 8.96388352e-01
-2.04952255e-01 4.11948413e-02 5.42630613e-01 -1.00413203e+00
-8.51644158e-01 -7.06042349e-01 -1.27129352e+00 4.64460813e-02
9.79037508e-02 -1.11593604e+00 -5.57953417e-01 -5.64722717e-01] | [5.971914768218994, 4.303834438323975] |
63185bc8-a137-4915-beb8-47e54f0f6fa4 | mutual-information-regularized-offline | 2210.07484 | null | https://arxiv.org/abs/2210.07484v1 | https://arxiv.org/pdf/2210.07484v1.pdf | Mutual Information Regularized Offline Reinforcement Learning | Offline reinforcement learning (RL) aims at learning an effective policy from offline datasets without active interactions with the environment. The major challenge of offline RL is the distribution shift that appears when out-of-distribution actions are queried, which makes the policy improvement direction biased by extrapolation errors. Most existing methods address this problem by penalizing the policy for deviating from the behavior policy during policy improvement or making conservative updates for value functions during policy evaluation. In this work, we propose a novel MISA framework to approach offline RL from the perspective of Mutual Information between States and Actions in the dataset by directly constraining the policy improvement direction. Intuitively, mutual information measures the mutual dependence of actions and states, which reflects how a behavior agent reacts to certain environment states during data collection. To effectively utilize this information to facilitate policy learning, MISA constructs lower bounds of mutual information parameterized by the policy and Q-values. We show that optimizing this lower bound is equivalent to maximizing the likelihood of a one-step improved policy on the offline dataset. In this way, we constrain the policy improvement direction to lie in the data manifold. The resulting algorithm simultaneously augments the policy evaluation and improvement by adding a mutual information regularization. MISA is a general offline RL framework that unifies conservative Q-learning (CQL) and behavior regularization methods (e.g., TD3+BC) as special cases. Our experiments show that MISA performs significantly better than existing methods and achieves new state-of-the-art on various tasks of the D4RL benchmark. | ['Shuicheng Yan', 'Min Lin', 'Zhongwen Xu', 'Bingyi Kang', 'Xiao Ma'] | 2022-10-14 | null | null | null | null | ['d4rl'] | ['robots'] | [-1.12527698e-01 1.69092983e-01 -8.25722218e-01 -6.35029972e-02
-7.48712003e-01 -7.01312006e-01 4.92859662e-01 2.46138260e-01
-8.18462729e-01 8.86555314e-01 2.76033849e-01 -5.16856492e-01
-2.99856484e-01 -6.37995780e-01 -9.51953173e-01 -9.10130143e-01
-2.37357050e-01 5.02176583e-01 -1.25212327e-01 -1.14960812e-01
2.28221714e-01 4.26417738e-01 -1.12097979e+00 -9.28896591e-02
9.53796208e-01 8.88588309e-01 1.15500353e-01 6.15996003e-01
5.11968210e-02 1.06545210e+00 -6.24877334e-01 1.98211923e-01
4.14941162e-01 -4.41638499e-01 -7.30095088e-01 6.30755126e-02
-7.10923523e-02 -7.61630237e-01 -4.96684402e-01 1.12409115e+00
3.83104235e-01 5.71659446e-01 4.40029353e-01 -1.21154618e+00
-2.55333900e-01 6.83428466e-01 -4.17563528e-01 1.06536061e-01
2.56739974e-01 6.28275335e-01 9.75493729e-01 -1.80955440e-01
3.75405312e-01 1.47650933e+00 1.67073578e-01 6.55363917e-01
-1.49000919e+00 -5.08438528e-01 7.39373386e-01 7.29923099e-02
-8.76837909e-01 -2.14031518e-01 7.19222367e-01 -3.15679371e-01
8.01151931e-01 5.48527688e-02 7.14115202e-01 1.08989835e+00
4.50735614e-02 1.30055285e+00 1.37598264e+00 -2.08952248e-01
8.23619246e-01 7.80797079e-02 -1.56656161e-01 5.07098079e-01
-8.39960724e-02 6.84331834e-01 -3.00100565e-01 -3.71854633e-01
7.65433609e-01 -7.14166164e-02 -1.05168382e-02 -6.36877060e-01
-8.70444775e-01 7.15754032e-01 1.53087184e-01 -4.14128974e-02
-4.36029106e-01 3.56469393e-01 4.18062449e-01 5.28075278e-01
3.51131797e-01 5.85599422e-01 -7.15467811e-01 -6.36455178e-01
-3.22175443e-01 5.16256928e-01 7.39922762e-01 6.68921888e-01
8.75063360e-01 1.55417830e-01 -6.04358792e-01 5.33897638e-01
1.75653309e-01 6.53099060e-01 4.14532453e-01 -1.34154475e+00
5.37722230e-01 3.69361997e-01 6.18157208e-01 -5.74568689e-01
-2.51437068e-01 -3.16726387e-01 -2.17245013e-01 3.50570649e-01
6.83292210e-01 -5.13512015e-01 -7.01320231e-01 2.19865370e+00
5.33389807e-01 1.89515337e-01 1.23896867e-01 7.93571830e-01
-2.41652712e-01 5.88053346e-01 2.13967785e-01 -5.86239934e-01
5.23613870e-01 -6.86574519e-01 -7.98474073e-01 -2.22569421e-01
7.59865046e-01 -1.90323219e-01 1.34471858e+00 3.92712712e-01
-9.85292852e-01 -3.77961516e-01 -6.88932657e-01 5.11729836e-01
2.17003990e-02 -1.69088319e-01 4.83109266e-01 1.95170641e-01
-7.96347141e-01 9.14177358e-01 -1.14472318e+00 -8.57822895e-02
3.46450329e-01 4.82201427e-01 3.96597832e-02 3.17531794e-01
-1.08826995e+00 9.74322021e-01 4.49284524e-01 -3.25493068e-01
-1.41127706e+00 -7.42308438e-01 -6.43629134e-01 -1.14853390e-01
1.10838377e+00 -2.84876853e-01 1.75217831e+00 -1.12073517e+00
-2.17671561e+00 2.20626593e-01 9.82431620e-02 -7.99359977e-01
6.90026939e-01 -5.22801101e-01 -1.86103314e-01 -1.13753177e-01
-1.49321020e-01 3.37442666e-01 1.09778214e+00 -1.13523555e+00
-7.68845260e-01 -4.66576576e-01 2.81347841e-01 5.39749742e-01
-2.66616166e-01 -4.08859342e-01 -1.95307001e-01 -4.07578260e-01
-5.03706634e-01 -1.02615070e+00 -4.86030370e-01 -2.38103405e-01
-1.16169147e-01 -4.00886893e-01 8.83958817e-01 -2.52436876e-01
1.34196162e+00 -2.00216556e+00 1.85562313e-01 3.31434458e-01
-1.06461264e-01 3.43088388e-01 -2.76756495e-01 4.33133841e-01
2.04294533e-01 -4.27963287e-02 -8.03908855e-02 -1.41423896e-01
2.19685376e-01 6.10044420e-01 -6.36090279e-01 7.45235384e-01
-1.38283491e-01 7.45129406e-01 -1.25490487e+00 -9.94590670e-02
2.68266052e-01 -2.86317039e-02 -8.48405540e-01 6.20846093e-01
-7.22246051e-01 1.09305263e+00 -8.74023616e-01 9.84301642e-02
3.94898057e-01 -8.40001367e-03 4.97168779e-01 2.49467388e-01
-2.57106960e-01 4.52446014e-01 -1.08246517e+00 1.56674063e+00
-5.55728495e-01 3.06191668e-02 1.03327520e-01 -1.22458053e+00
7.02610075e-01 7.15172365e-02 9.83934402e-01 -8.59800100e-01
1.26277283e-01 -1.95784330e-01 1.74247213e-02 -5.26210606e-01
1.15890786e-01 4.70267329e-03 -7.85594955e-02 7.45722651e-01
-3.03356666e-02 -8.07233080e-02 4.39132974e-02 -1.39285019e-02
1.01809239e+00 5.01287162e-01 4.60906595e-01 -2.05504492e-01
4.02188122e-01 -2.22758919e-01 8.73699248e-01 1.08649290e+00
-5.84279895e-01 -4.69057083e-01 8.18214476e-01 -1.84583604e-01
-7.08472133e-01 -9.96576190e-01 2.52693117e-01 1.36697614e+00
7.71017224e-02 -3.16096634e-01 -6.14307642e-01 -1.21161079e+00
4.45150346e-01 9.83362973e-01 -6.24384940e-01 -5.44692934e-01
-6.08222961e-01 -3.07204336e-01 1.17068186e-01 4.56502199e-01
4.78681266e-01 -1.02221346e+00 -6.80941403e-01 3.66148740e-01
1.41766965e-01 -7.26498067e-01 -6.26450598e-01 2.02704653e-01
-9.84004080e-01 -9.72270131e-01 -2.63561755e-01 -1.50681064e-01
5.59354544e-01 -1.48033947e-01 9.13299441e-01 -4.40087080e-01
2.70209134e-01 8.25688481e-01 -9.61062759e-02 -2.99025476e-01
-4.94248301e-01 -1.64699882e-01 4.67627138e-01 8.69302079e-02
1.30058259e-01 -4.61951315e-01 -6.55006111e-01 2.70543694e-01
-7.67587006e-01 -3.56190264e-01 3.13211113e-01 9.48853314e-01
7.53194332e-01 -4.82582040e-02 6.84116423e-01 -8.47869575e-01
8.55786800e-01 -4.94395018e-01 -1.16139650e+00 2.43964538e-01
-9.18864965e-01 5.65150201e-01 9.35421109e-01 -6.98646545e-01
-1.10942113e+00 6.27333820e-02 1.05932400e-01 -5.59507549e-01
-1.07864358e-01 3.16462368e-01 -1.19803928e-01 2.34381869e-01
4.65010732e-01 1.50562048e-01 2.42251411e-01 -4.43033040e-01
6.30178750e-01 4.11964297e-01 2.43467480e-01 -1.12929392e+00
4.61936891e-01 4.43452299e-01 -8.33780617e-02 -4.45259362e-01
-1.06355631e+00 -4.00478184e-01 -2.08547086e-01 -3.64755213e-01
4.93198395e-01 -6.37688756e-01 -1.30818021e+00 1.95271969e-01
-4.99860525e-01 -1.02705979e+00 -7.30581880e-01 6.02697134e-01
-1.05376363e+00 2.12619230e-01 -3.32401127e-01 -1.21948588e+00
4.82631922e-02 -1.16187549e+00 6.89433575e-01 2.20825046e-01
1.26629978e-01 -1.11254239e+00 4.77013737e-01 -1.17268413e-01
2.16989726e-01 -6.20918758e-02 7.83726215e-01 -5.64849555e-01
-2.77133346e-01 2.47566149e-01 3.98476690e-01 5.31583071e-01
2.18314424e-01 -4.24608320e-01 -7.71384120e-01 -7.43005812e-01
1.54434130e-01 -6.33809030e-01 7.08419979e-01 4.70525861e-01
1.52160335e+00 -8.31817925e-01 -9.20700133e-02 4.71284330e-01
1.13319612e+00 5.56303263e-01 2.46320188e-01 2.63950318e-01
3.95072103e-01 3.80874902e-01 1.13255191e+00 1.08768642e+00
2.52107918e-01 5.54103494e-01 5.87351322e-01 2.20435902e-01
5.19687712e-01 -7.81835794e-01 8.42516303e-01 3.71992528e-01
2.27911055e-01 2.38300804e-02 -5.37470222e-01 2.22825274e-01
-2.24064398e+00 -9.23987806e-01 6.86369777e-01 2.65109968e+00
1.21019304e+00 3.74089740e-02 4.30483788e-01 -3.92359346e-01
2.73374766e-01 2.06905767e-01 -1.37011325e+00 -3.61681014e-01
3.08387101e-01 1.31192043e-01 8.31158400e-01 6.89950168e-01
-1.12652290e+00 1.07802653e+00 6.48573303e+00 7.89584458e-01
-1.09133506e+00 -6.28202185e-02 6.43570602e-01 -1.46186978e-01
-1.75323948e-01 9.55093279e-02 -7.32935548e-01 4.86906797e-01
9.26622272e-01 -1.68713093e-01 1.00378263e+00 9.66853499e-01
6.44104302e-01 -2.77163863e-01 -1.28197622e+00 7.46884942e-01
-4.76891577e-01 -9.65874016e-01 -3.84958029e-01 8.83997679e-02
9.14740920e-01 -6.36151480e-03 2.97003984e-01 7.92985499e-01
8.86610925e-01 -6.99756444e-01 5.92750132e-01 5.88821173e-01
5.55742860e-01 -9.59777653e-01 2.08835304e-01 7.93764055e-01
-7.80883193e-01 -6.33900285e-01 -2.47257173e-01 -1.98361993e-01
-2.24767104e-01 2.18160778e-01 -7.82487571e-01 4.14073914e-02
3.68835777e-01 8.38948727e-01 -4.36138622e-02 5.82648158e-01
-5.05383253e-01 7.50048339e-01 -2.43720308e-01 -6.79308996e-02
4.09791797e-01 -6.09444201e-01 7.18569577e-01 6.92193925e-01
-7.32832700e-02 -1.09713152e-01 8.54045808e-01 9.08586323e-01
-2.44193338e-03 -3.11215520e-02 -6.98139846e-01 -2.30860308e-01
4.43308979e-01 8.38423252e-01 -8.23674053e-02 -3.34859550e-01
-1.76810458e-01 7.06849813e-01 6.97248697e-01 6.74686015e-01
-7.92326391e-01 2.07741961e-01 1.03405643e+00 -3.60954523e-01
1.97078586e-01 -2.74121135e-01 1.50076434e-01 -1.07021475e+00
-1.29285440e-01 -1.13831413e+00 5.37100017e-01 -5.41378781e-02
-1.16726959e+00 -1.11955099e-01 8.72448534e-02 -1.23877549e+00
-5.22688627e-01 -3.28562081e-01 -2.89826989e-01 5.30671835e-01
-1.44083309e+00 -4.16697264e-01 4.43910539e-01 7.26756513e-01
3.95764619e-01 -3.84743176e-02 5.99658787e-01 -1.69519752e-01
-7.44686425e-01 4.94804323e-01 5.04418135e-01 -1.89303368e-01
7.64478803e-01 -1.51095033e+00 -1.49929687e-01 5.35413742e-01
-1.51765183e-01 4.88872141e-01 6.71495736e-01 -6.77777946e-01
-1.64496863e+00 -1.10377896e+00 -1.64306760e-02 -2.31551394e-01
9.36676025e-01 -1.49484217e-01 -7.07524300e-01 7.90294528e-01
-5.86645603e-02 8.53172094e-02 3.56240362e-01 4.75057364e-02
-1.87001035e-01 -3.16956937e-01 -1.00606608e+00 7.72431731e-01
8.46824169e-01 -5.83345890e-01 -3.47568601e-01 4.65218723e-01
7.27975845e-01 -3.34320664e-01 -8.29125822e-01 3.05535316e-01
2.55214036e-01 -6.36057675e-01 6.83487177e-01 -1.23131597e+00
-1.17897935e-01 -1.52762324e-01 -3.55105964e-03 -1.60747302e+00
-1.35924920e-01 -1.23684919e+00 -6.63771927e-01 6.85974300e-01
6.48289770e-02 -6.76668823e-01 7.94427454e-01 6.32139444e-01
7.86722228e-02 -9.65853274e-01 -8.28364670e-01 -9.95625556e-01
3.49398285e-01 -4.68048841e-01 5.44664204e-01 6.34740949e-01
2.71243066e-01 -6.20727241e-02 -5.74338794e-01 5.78982979e-02
6.43969655e-01 1.83593199e-01 9.44036424e-01 -5.18380523e-01
-7.80915380e-01 -4.63405669e-01 1.77990839e-01 -1.60793924e+00
6.12891912e-01 -7.17785060e-01 2.09798217e-01 -1.06741345e+00
2.95544472e-02 -5.79219997e-01 -6.01518750e-01 4.78792816e-01
-2.42048323e-01 -8.60429168e-01 2.81486303e-01 1.43980950e-01
-8.69655252e-01 1.01330996e+00 1.51090932e+00 9.67503246e-03
-7.56019831e-01 3.60110641e-01 -4.10431504e-01 7.37484455e-01
1.06802726e+00 -4.66491014e-01 -7.45822489e-01 -1.44840062e-01
1.10866219e-01 4.44111049e-01 7.30321482e-02 -6.37132823e-01
1.59610771e-02 -8.41935694e-01 -1.74842756e-02 -4.59088296e-01
6.96328282e-02 -7.45364845e-01 -4.04284149e-01 6.68968201e-01
-9.44102645e-01 -8.56401026e-02 7.26501793e-02 1.01286793e+00
1.22502141e-01 9.72010717e-02 8.57620001e-01 -9.49235260e-02
-6.29928172e-01 5.86103857e-01 -5.68519115e-01 4.57661003e-01
9.95688319e-01 4.04840767e-01 -1.99450269e-01 -6.91980839e-01
-7.89561093e-01 6.82617843e-01 2.48737201e-01 2.09985584e-01
4.69989121e-01 -1.17612481e+00 -2.89664179e-01 1.80854470e-01
-6.07603043e-02 -2.40626723e-01 1.43041546e-02 8.84064972e-01
1.86933935e-01 3.64387482e-01 -3.67334597e-02 -3.68530661e-01
-7.35994875e-01 6.58411086e-01 5.73924899e-01 -7.10303366e-01
-4.84356284e-01 4.59056824e-01 2.26050749e-01 -6.28627002e-01
5.95862627e-01 -4.82560426e-01 -1.74249515e-01 -2.31228709e-01
4.42936420e-01 4.98999149e-01 -3.21916610e-01 -1.21632583e-01
-5.78314140e-02 9.85860750e-02 -3.50774638e-02 -4.37302947e-01
1.08328903e+00 -4.97702993e-02 1.79534674e-01 5.80585539e-01
1.10573030e+00 -1.37331024e-01 -2.10509062e+00 -4.64097470e-01
2.41875798e-01 -4.67258990e-01 1.21184185e-01 -9.80152488e-01
-9.20406520e-01 4.46493626e-01 6.44805014e-01 -6.22067787e-03
9.94145334e-01 -1.54008687e-01 4.87879246e-01 6.90217555e-01
5.33636630e-01 -1.76902986e+00 2.90857047e-01 7.43987620e-01
6.96870506e-01 -1.31105149e+00 -7.60076046e-02 4.52275783e-01
-9.19873118e-01 7.80694425e-01 8.16023290e-01 -2.72042394e-01
6.59378767e-01 1.51125014e-01 -8.95383060e-02 1.75085276e-01
-9.56187189e-01 -3.84032458e-01 8.96397531e-02 5.02439559e-01
-6.59529306e-03 4.79783118e-01 -2.73561865e-01 2.37228781e-01
2.97786117e-01 -2.04769224e-02 1.65347487e-01 1.27189326e+00
-5.09277523e-01 -1.26760626e+00 -2.24564865e-01 3.31770986e-01
-3.16871464e-01 3.75360161e-01 -2.33236983e-01 6.89428926e-01
-1.69372365e-01 9.42623138e-01 2.39477251e-02 -2.26229295e-01
2.92133868e-01 -8.97510424e-02 5.69734156e-01 -4.83814090e-01
-4.31882560e-01 2.68569320e-01 -1.09741919e-01 -1.23411989e+00
-1.97337031e-01 -7.46464133e-01 -1.43776095e+00 -5.26277944e-02
-9.65341106e-02 1.55028880e-01 4.43558484e-01 1.03036988e+00
3.94284129e-01 4.28918719e-01 1.05799329e+00 -4.66433018e-01
-1.61722803e+00 -7.71916330e-01 -5.97466707e-01 4.41078663e-01
5.90707779e-01 -8.67150247e-01 -3.44202697e-01 -4.36897844e-01] | [4.0716423988342285, 2.2066099643707275] |
4c781e21-6ad8-40e6-b5e0-aeafdf4f1dbf | sms-spam-filtering-using-probabilistic-topic | 1606.05554 | null | http://arxiv.org/abs/1606.05554v1 | http://arxiv.org/pdf/1606.05554v1.pdf | SMS Spam Filtering using Probabilistic Topic Modelling and Stacked Denoising Autoencoder | In This paper we present a novel approach to spam filtering and demonstrate
its applicability with respect to SMS messages. Our approach requires minimum
features engineering and a small set of la- belled data samples. Features are
extracted using topic modelling based on latent Dirichlet allocation, and then
a comprehensive data model is created using a Stacked Denoising Autoencoder
(SDA). Topic modelling summarises the data providing ease of use and high
interpretability by visualising the topics using word clouds. Given that the
SMS messages can be regarded as either spam (unwanted) or ham (wanted), the SDA
is able to model the messages and accurately discriminate between the two
classes without the need for a pre-labelled training set. The results are
compared against the state-of-the-art spam detection algorithms with our
proposed approach achieving over 97% accuracy which compares favourably to the
best reported algorithms presented in the literature. | ['A. Stephen McGough', 'Peter Matthews', 'Toby Breckon', 'Noura Al Moubayed'] | 2016-06-17 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [ 2.04648033e-01 1.93721145e-01 4.04255956e-01 -5.57938278e-01
-4.05707896e-01 -5.87719418e-02 1.39109135e+00 3.16476643e-01
-4.48244035e-01 3.96549940e-01 3.09789330e-01 -3.19231987e-01
-2.54857570e-01 -9.23364818e-01 -6.36615679e-02 -1.06016695e+00
-6.76305667e-02 1.09408212e+00 4.54392910e-01 -3.23010325e-01
3.73989612e-01 3.98683906e-01 -2.05792737e+00 7.07249999e-01
7.97758520e-01 1.05507958e+00 1.31667122e-01 6.65883005e-01
-4.80565339e-01 3.14987630e-01 -1.25255203e+00 -4.71192241e-01
2.81492740e-01 -1.53491512e-01 -7.73569286e-01 5.10404587e-01
3.04444402e-01 -1.82643428e-01 7.11493567e-02 8.04812729e-01
2.39441648e-01 3.03877980e-01 8.85883451e-01 -1.05586851e+00
-5.47819659e-02 -2.27614045e-02 -6.42500371e-02 2.63313819e-02
2.64770150e-01 -1.67186573e-01 5.53676128e-01 -8.10728312e-01
6.31627262e-01 1.50161469e+00 6.78963900e-01 4.02082920e-01
-1.46170342e+00 -4.67822850e-01 7.67925084e-02 2.18212962e-01
-8.84259582e-01 -1.59144476e-01 3.95844042e-01 -5.42228281e-01
8.83886158e-01 3.67411226e-01 6.51481330e-01 1.17737937e+00
8.45944807e-02 7.18038976e-01 1.46510804e+00 -4.25109118e-01
6.18176818e-01 6.78878963e-01 3.99632812e-01 8.57539028e-02
2.63888210e-01 -1.16710886e-01 -5.09511650e-01 -8.17342341e-01
1.51991606e-01 2.12976396e-01 1.22703448e-01 -4.70729083e-01
-5.73754191e-01 1.21692073e+00 6.53653443e-02 4.89053369e-01
-6.50450408e-01 -3.90845001e-01 6.90282822e-01 3.79766583e-01
9.28459108e-01 -2.04563513e-02 -1.67719007e-01 -4.34297286e-02
-1.55858135e+00 1.36884227e-01 1.05887496e+00 4.58836317e-01
3.97244304e-01 4.40441370e-02 1.10947497e-01 6.89832568e-01
7.24547088e-01 1.36760548e-01 7.19677448e-01 -2.69062281e-01
5.87940775e-02 7.73469210e-01 2.54340917e-01 -1.12158537e+00
-6.42722398e-02 -9.30593312e-02 -6.82252347e-01 5.60404956e-01
3.37092191e-01 1.38489664e-01 -1.13108003e+00 7.96505630e-01
4.22797531e-01 7.83415958e-02 7.31116682e-02 6.69267654e-01
8.70522916e-01 8.00879776e-01 2.53860056e-01 -2.10477725e-01
1.47727585e+00 -3.73371720e-01 -8.03728521e-01 7.46096969e-02
2.49886289e-01 -9.15629268e-01 4.43853498e-01 8.91193271e-01
-8.26806188e-01 -4.72790152e-01 -8.51502895e-01 1.75382704e-01
-8.79747689e-01 -7.32305124e-02 4.46457446e-01 1.16048384e+00
-1.08298790e+00 6.11148596e-01 -7.90408731e-01 -7.33043134e-01
2.83629745e-01 6.42932534e-01 -1.93243474e-01 2.88412452e-01
-1.27200031e+00 9.94375229e-01 4.47679192e-01 6.26835227e-02
-8.35973382e-01 -1.80931553e-01 -5.14439702e-01 -6.74566776e-02
-1.35863319e-01 -3.23026061e-01 9.68510032e-01 -1.09789252e+00
-1.40615976e+00 9.55390692e-01 -2.25605413e-01 -8.92207742e-01
6.45025730e-01 -5.36308847e-02 -5.11755049e-01 4.19659168e-01
-1.30720571e-01 3.61560911e-01 1.60640669e+00 -1.59180939e+00
-8.59689474e-01 -3.08613241e-01 -3.25134665e-01 -2.11465001e-01
-4.77046520e-01 5.06085336e-01 -1.41180609e-03 -6.13341510e-01
3.02794695e-01 -5.77395022e-01 -4.49105874e-02 -4.38525021e-01
9.83492844e-03 -3.97931606e-01 1.40856814e+00 -1.02892399e+00
1.32835007e+00 -1.93600261e+00 -2.44826954e-02 4.82616037e-01
1.82813033e-01 6.38934135e-01 5.25047660e-01 8.34417820e-01
-2.45738298e-01 -2.23556355e-01 -2.43872821e-01 -6.09299123e-01
1.98628336e-01 3.91130298e-01 -5.89460671e-01 7.17486143e-01
7.40282536e-02 6.51907995e-02 -5.72095454e-01 -4.49725151e-01
6.82431221e-01 8.09462249e-01 -2.42949706e-02 1.39384836e-01
-1.58183545e-01 -3.34671251e-02 -1.99497476e-01 3.33158970e-01
9.22149241e-01 2.87396401e-01 2.16757506e-01 2.84046978e-01
-2.41037514e-02 4.15941507e-01 -1.37950218e+00 7.67015815e-01
-3.91546190e-01 8.02270234e-01 3.07351172e-01 -1.26912498e+00
1.25319374e+00 5.39069772e-01 1.21605791e-01 -2.31012762e-01
4.63459402e-01 2.27855638e-01 -2.85535246e-01 -1.33591950e-01
6.27059102e-01 -4.01706815e-01 1.24856949e-01 5.53323269e-01
3.70723844e-01 1.34976253e-01 1.55551523e-01 3.57190549e-01
6.07955039e-01 -1.92808092e-01 1.43293604e-01 -5.98132312e-01
6.74045682e-01 5.72966971e-02 1.77111905e-02 7.97637105e-01
-3.68759669e-02 3.29682648e-01 6.66495502e-01 -6.14880681e-01
-1.04870653e+00 -8.87826204e-01 -4.71782982e-01 1.10427892e+00
-1.66719124e-01 -3.38228315e-01 -9.97553587e-01 -8.65891993e-01
5.91031909e-02 8.35613370e-01 -8.22855771e-01 1.76493526e-01
-3.40038836e-01 -1.12092936e+00 1.38407364e-01 4.10090061e-03
2.00087219e-01 -9.15014744e-01 -5.71300268e-01 1.93010017e-01
7.65070096e-02 -4.50053543e-01 5.71355641e-01 2.14772493e-01
-1.22916555e+00 -9.29728389e-01 -6.60597503e-01 -7.75042832e-01
7.03027010e-01 1.99143752e-01 1.15710008e+00 1.73462436e-01
-4.17017847e-01 1.41402796e-01 -5.04533350e-01 -5.73919773e-01
-8.39840829e-01 -1.87570557e-01 6.71921521e-02 2.88815886e-01
1.04964292e+00 -4.10040706e-01 -6.63312197e-01 3.02278191e-01
-1.13159525e+00 -4.83665228e-01 2.68215925e-01 9.21218693e-01
-1.63288876e-01 3.94351870e-01 4.85917240e-01 -1.05469382e+00
9.35050905e-01 -6.60432994e-01 -5.02762914e-01 -2.80266821e-01
-5.25738657e-01 -3.88345391e-01 4.47890818e-01 -8.73776078e-02
-1.26205528e+00 -7.87154585e-02 -3.75621468e-02 9.80698913e-02
-8.50495458e-01 -1.83895901e-01 4.21711355e-02 1.85233951e-02
7.86944866e-01 2.36024842e-01 3.62893075e-01 -6.92413151e-01
3.43498647e-01 1.51957583e+00 8.84112492e-02 -4.19289470e-02
4.57163125e-01 1.02719474e+00 -2.99546480e-01 -1.27906132e+00
-1.56624496e-01 -1.18308783e+00 -9.55624759e-01 -2.51745552e-01
6.37289584e-01 -5.66866040e-01 -6.07224345e-01 5.83364129e-01
-9.39440370e-01 4.24089789e-01 -2.16158032e-01 1.04603924e-01
-5.05314350e-01 8.32411289e-01 -5.11770725e-01 -1.33543515e+00
-5.21535456e-01 -9.25361216e-01 1.23286259e+00 -2.94703007e-01
-5.19739270e-01 -1.05404222e+00 -2.16542512e-01 4.13532048e-01
4.54714477e-01 -5.00814505e-02 9.80434895e-01 -1.38035429e+00
-1.11547470e-01 -8.35675418e-01 -8.99483636e-02 5.79586506e-01
-1.11114293e-01 -2.61580735e-01 -1.15396535e+00 -4.04009312e-01
5.33823729e-01 1.08966909e-01 1.21914852e+00 6.47001565e-02
7.16693759e-01 -8.85644853e-02 -3.82125348e-01 -2.00014725e-01
1.14227200e+00 -2.02194937e-02 6.66408479e-01 7.80336559e-01
-2.23451167e-01 8.76077533e-01 5.67338228e-01 4.04503465e-01
-2.60396302e-01 3.72949660e-01 6.05840683e-01 -1.05058383e-02
1.98465452e-01 4.07819390e-01 1.99788168e-01 6.08475626e-01
3.39233637e-01 -2.43382528e-01 -8.67932916e-01 7.79541492e-01
-1.88358378e+00 -9.24766958e-01 -6.75902128e-01 2.21940255e+00
4.26800400e-01 2.78329074e-01 4.58560526e-01 5.67380905e-01
8.79437983e-01 9.38483402e-02 3.01411897e-01 -9.12025511e-01
1.79206014e-01 3.71599555e-01 3.19980741e-01 6.24102473e-01
-1.63198614e+00 7.61837780e-01 6.82490301e+00 1.13671303e+00
-6.77790940e-01 1.72206566e-01 2.98441052e-01 -5.89851141e-02
1.96741775e-01 4.72356705e-03 -8.18638802e-01 8.45926464e-01
1.35998809e+00 2.09074080e-01 -2.24608660e-01 8.00628901e-01
3.47062826e-01 -5.23876369e-01 -4.10043329e-01 4.34250712e-01
4.55601037e-01 -8.82059693e-01 2.23193198e-01 1.34931549e-01
6.24732614e-01 -2.51069367e-01 2.28909682e-02 1.37835786e-01
3.21772903e-01 -8.86116028e-01 2.99788803e-01 8.48832965e-01
-1.16305957e-02 -8.10042858e-01 1.33607733e+00 5.82002461e-01
-4.09436405e-01 -1.69750780e-01 -4.50856686e-01 -7.73908570e-02
2.30620921e-01 7.85631597e-01 -1.15702105e+00 5.22310972e-01
7.80248284e-01 1.66784927e-01 -5.20613968e-01 1.11583495e+00
9.65307057e-02 8.91407728e-01 -5.85724235e-01 -2.19246462e-01
4.93409365e-01 -3.88594896e-01 6.94705486e-01 1.52719283e+00
5.38365692e-02 -4.98942018e-01 5.52913137e-02 3.53681803e-01
6.74802005e-01 2.17160419e-01 -3.06981653e-01 5.25146015e-02
-7.76433870e-02 1.05568957e+00 -1.06713510e+00 -9.33993995e-01
-2.42753267e-01 1.14034939e+00 -3.73099267e-01 1.52955379e-03
-8.46619979e-02 -4.20005411e-01 5.01664758e-01 3.28060716e-01
5.89383662e-01 4.83306460e-02 -2.41027072e-01 -8.24867904e-01
-1.70344993e-01 -8.23653162e-01 3.49508166e-01 -4.08981621e-01
-1.54201365e+00 7.52079368e-01 2.02294022e-01 -1.18272567e+00
-4.39926565e-01 -9.55990791e-01 -5.10191202e-01 1.15617859e+00
-9.56362307e-01 -1.16302538e+00 -8.80055279e-02 3.64017971e-02
6.65700555e-01 -5.75768590e-01 1.16202939e+00 -5.15083782e-02
1.13194725e-02 -2.36403599e-01 8.48150730e-01 -2.97773719e-01
5.84358096e-01 -1.50893462e+00 6.31045640e-01 5.12727201e-01
-1.77030966e-01 8.43147874e-01 1.17829287e+00 -9.29338574e-01
-5.26014805e-01 -8.20142210e-01 1.29532385e+00 -5.98109424e-01
7.56914437e-01 -6.55743003e-01 -1.12359083e+00 6.61567301e-02
5.01143754e-01 -5.40228426e-01 9.52191710e-01 1.87775999e-01
-1.22555159e-01 4.25478667e-01 -1.29246247e+00 1.54033393e-01
1.11386918e-01 -4.81214076e-01 -1.12316728e+00 5.56442857e-01
-1.64838254e-01 1.07846901e-01 -5.53330660e-01 -1.00641966e-01
3.93787891e-01 -1.31836545e+00 8.41122329e-01 -9.02454019e-01
1.04991324e-01 -7.98896030e-02 8.86211842e-02 -1.15463471e+00
-1.09876506e-01 -2.34436661e-01 -2.15746120e-01 9.55369234e-01
-3.00961882e-02 -5.11888623e-01 1.01801789e+00 1.93472490e-01
4.76285547e-01 -4.77737129e-01 -1.07249284e+00 -4.95379448e-01
4.09195684e-02 -4.52009261e-01 3.35547417e-01 7.93099046e-01
5.98978475e-02 -1.11131206e-01 -3.54321033e-01 6.52294680e-02
6.84760690e-01 9.33501963e-03 6.64090753e-01 -1.91155005e+00
1.35985985e-01 -3.03884923e-01 -9.37843561e-01 -6.02762401e-01
8.56698006e-02 -6.93298697e-01 -1.24341168e-01 -1.59572363e+00
-4.11703400e-02 6.17834143e-02 7.35424086e-02 -6.45219237e-02
7.57036284e-02 3.03151906e-01 2.07770728e-02 2.43917018e-01
-5.70027113e-01 3.45757365e-01 1.67884037e-01 5.63921258e-02
3.23787071e-02 4.34231967e-01 -3.94315869e-01 1.04109120e+00
8.67718875e-01 -4.67297703e-01 3.33491936e-02 1.45877346e-01
-8.73893797e-02 -4.94640112e-01 5.18764377e-01 -8.09101045e-01
2.29911700e-01 5.24546623e-01 5.45872331e-01 -9.44683731e-01
6.06798530e-01 -1.02294469e+00 -2.41162211e-01 4.75761324e-01
-8.56858790e-02 -4.50654179e-01 6.55694976e-02 9.03240681e-01
-2.37095803e-01 -7.77209520e-01 7.51441360e-01 -2.93900877e-01
-7.10366189e-01 -4.02857363e-01 -1.17390144e+00 -6.74658120e-01
1.02500904e+00 -4.85935837e-01 -5.38929217e-02 -6.93111420e-01
-1.34811270e+00 -4.93028667e-04 3.80882174e-01 4.37672794e-01
3.31272930e-01 -9.26290512e-01 -4.47294086e-01 5.57478070e-01
-7.32743070e-02 -4.32101667e-01 4.52062547e-01 5.15901566e-01
-6.54215157e-01 7.05902994e-01 -2.68093459e-02 -5.78493297e-01
-1.63766837e+00 3.62215817e-01 5.36329038e-02 -2.35051420e-02
-6.32104218e-01 6.19089842e-01 -3.11497539e-01 -5.36886871e-01
4.52951670e-01 3.20436805e-01 -6.28273785e-01 6.22560203e-01
5.71696341e-01 7.06903815e-01 5.90730548e-01 -1.01045084e+00
-2.11067706e-01 -1.18226394e-01 -3.60998958e-01 -4.38075252e-02
1.24899054e+00 1.14556225e-02 -3.45079452e-01 4.07808691e-01
1.08796239e+00 -3.04180175e-01 -7.32640207e-01 -8.29484835e-02
3.72758478e-01 -7.07308650e-01 3.60235602e-01 -6.42600596e-01
-1.73337564e-01 1.12442291e+00 1.01482320e+00 1.01087141e+00
7.99879730e-01 -2.16391683e-01 5.78322947e-01 3.47824633e-01
2.86224801e-02 -1.52506578e+00 -3.78688604e-01 3.55549902e-01
7.14344919e-01 -1.29773319e+00 1.20716505e-01 -4.93659794e-01
-6.89973593e-01 1.27082086e+00 -7.85603076e-02 -5.78477442e-01
8.59813333e-01 -6.18165731e-02 2.17293918e-01 -4.30979043e-01
-6.86203420e-01 -3.46826524e-01 2.16404676e-01 9.03065205e-01
3.15441340e-01 5.08703701e-02 -6.80776656e-01 3.49911332e-01
-2.58401394e-01 -4.01948333e-01 3.16793054e-01 9.87979949e-01
-9.58324075e-01 -1.20578587e+00 -7.99676239e-01 8.55657220e-01
-6.73794568e-01 1.11821376e-01 -7.21184194e-01 7.85244107e-01
1.15291834e-01 1.14029503e+00 2.96473205e-01 -3.86112705e-02
1.57013431e-01 6.96154594e-01 -1.70120224e-01 -6.72055662e-01
-1.03856468e+00 2.99168259e-01 1.59020960e-01 -1.77028030e-01
-4.39948797e-01 -8.84724855e-01 -6.72261178e-01 -1.65379584e-01
-7.20327795e-01 5.91276944e-01 1.15765345e+00 1.03535771e+00
2.62566030e-01 1.97308674e-01 5.22183776e-01 -9.73322511e-01
-5.21753430e-01 -1.29862630e+00 -6.61581755e-01 5.54489255e-01
1.99328184e-01 -7.33236134e-01 -6.92614377e-01 7.70822167e-02] | [7.93793249130249, 10.008736610412598] |
fa3b6ce5-0bd3-4ff2-954d-17ffc41ead4e | learning-adaptive-control-flow-in | null | null | https://openreview.net/forum?id=v8IbnUesFpE | https://openreview.net/pdf?id=v8IbnUesFpE | Learning Adaptive Control Flow in Transformers for Improved Systematic Generalization | Despite successes across a broad range of applications, Transformers have limited capability in systematic generalization. The situation is especially frustrating for algorithmic tasks, where they often fail to find intuitive solutions that can be simply expressed in terms of attention patterns. Here we propose two modifications to the Transformer architecture, copy gate and geometric attention, which facilitate learning such intuitive and interpretable solutions to algorithmic problems. Our novel Transformer, called Transformer Control Flow (TCF) achieves 100% length generalization accuracy on the classic compositional table lookup task. The resulting attention and gating patterns are interpretable, demonstrating that the model implements adaptive control flow. | ['Jürgen Schmidhuber', 'Kazuki Irie', 'Róbert Csordás'] | 2021-10-08 | null | null | null | neurips-workshop-aiplans-2021-12 | ['systematic-generalization'] | ['reasoning'] | [ 2.96869040e-01 1.98808312e-01 -1.94594741e-01 -4.30558145e-01
-4.21522737e-01 -8.69727254e-01 4.98731285e-01 2.86552131e-01
8.48917067e-02 8.21095526e-01 3.89522195e-01 -8.30014586e-01
-4.30522084e-01 -7.56566703e-01 -6.16054118e-01 -2.49614894e-01
-9.15685818e-02 7.96531737e-01 5.23856608e-03 -5.61894834e-01
4.89671618e-01 6.05452657e-01 -1.37491810e+00 5.49569964e-01
1.01885474e+00 1.03396261e+00 1.14793792e-01 6.60682857e-01
-5.27990222e-01 9.33628559e-01 -4.30538654e-01 -4.86664295e-01
4.04460728e-02 -3.72711062e-01 -1.14887774e+00 -3.56159896e-01
6.03669047e-01 1.87580641e-02 -9.55847949e-02 7.34644711e-01
7.87546858e-02 2.12430969e-01 6.53000772e-01 -1.24012649e+00
-1.19359434e+00 8.47592413e-01 -1.27017677e-01 4.88755018e-01
3.87222052e-01 2.52572507e-01 1.64481831e+00 -8.16928625e-01
5.03677189e-01 1.52552581e+00 8.60599816e-01 8.21964145e-01
-1.73858821e+00 -4.77476299e-01 5.67708313e-01 2.98844755e-01
-9.43228960e-01 -2.42632613e-01 4.50022697e-01 -1.90068901e-01
1.32624328e+00 4.41276044e-01 8.03030610e-01 8.06279898e-01
4.76745665e-01 8.62432599e-01 9.82728422e-01 -1.79393426e-01
2.39799470e-01 -1.58789843e-01 2.98856795e-01 7.83677280e-01
2.49521241e-01 -4.76194583e-02 -5.96365690e-01 9.40348208e-02
5.89806557e-01 9.87959430e-02 -3.43555808e-01 -4.17179137e-01
-9.72266614e-01 8.33471835e-01 1.13800967e+00 2.37146392e-01
-8.86021107e-02 3.91013026e-01 4.07992303e-01 6.39874339e-01
-4.58125435e-02 1.42252600e+00 -5.76039732e-01 -5.46199083e-02
-3.23960841e-01 4.77223903e-01 7.61686623e-01 1.09221053e+00
5.80759943e-01 2.58259803e-01 -1.03987962e-01 2.70563185e-01
-3.61051485e-02 3.18035096e-01 4.71913964e-01 -6.76167309e-01
3.75812382e-01 8.76033902e-01 -2.13664323e-01 -9.73382473e-01
-5.71776450e-01 -6.91086411e-01 -7.35019445e-01 -2.19924022e-02
4.47527111e-01 3.11247826e-01 -8.95083070e-01 1.81478202e+00
-1.01632245e-01 -5.67505777e-01 7.32829049e-02 6.85455918e-01
6.22291267e-01 7.92003036e-01 4.07352656e-01 8.77670497e-02
1.16253233e+00 -8.99079740e-01 -6.37520373e-01 -5.93186557e-01
7.92939961e-01 -3.50056052e-01 1.77328575e+00 4.63752538e-01
-1.24972165e+00 -3.40982020e-01 -1.08402848e+00 -5.75581849e-01
-6.44036710e-01 -4.44636077e-01 1.09006333e+00 5.73817790e-01
-1.11397111e+00 5.59317052e-01 -4.80664313e-01 -3.01740259e-01
8.13798964e-01 6.25629485e-01 -1.45386264e-01 -8.08522925e-02
-9.32454109e-01 9.96334672e-01 3.39790791e-01 1.37303412e-01
-8.00250292e-01 -1.12180698e+00 -8.55897248e-01 5.31681359e-01
4.44366306e-01 -9.52402055e-01 1.65026224e+00 -9.11523521e-01
-1.10293162e+00 6.25210524e-01 -3.88462067e-01 -6.17226899e-01
1.48433641e-01 -3.23181063e-01 -1.58405617e-01 -2.31769249e-01
-1.16449641e-02 8.93028677e-01 6.73305392e-01 -1.09498465e+00
-4.85265106e-01 -3.24539393e-01 2.85838872e-01 1.17122650e-01
-2.85474658e-01 -3.25121999e-01 1.61246121e-01 -8.88880372e-01
8.87754783e-02 -5.85759938e-01 -4.12405223e-01 2.49352172e-01
-4.29176569e-01 -2.45556623e-01 8.82907987e-01 -3.00802141e-01
1.50951111e+00 -1.74092448e+00 4.60592061e-01 2.08292112e-01
6.68175161e-01 1.52193829e-01 9.54158902e-02 2.84928054e-01
-2.20550075e-01 5.36104858e-01 -4.23172563e-01 9.05182213e-02
2.74581462e-01 4.51545864e-01 -8.96046162e-01 -2.60800481e-01
4.75300670e-01 1.55462277e+00 -1.06183004e+00 -7.23682642e-02
-1.05048247e-01 -7.79777318e-02 -1.14615536e+00 9.69335809e-02
-6.87454343e-01 1.34184182e-01 -4.25057143e-01 6.64999425e-01
4.21162657e-02 -5.52278578e-01 3.06251287e-01 3.28734852e-02
2.11656243e-01 7.34378994e-01 -6.37426317e-01 1.53919792e+00
-7.13132143e-01 9.21311915e-01 -3.37563306e-01 -8.55980039e-01
7.34549046e-01 -3.95860635e-02 -3.31947684e-01 -7.52523541e-01
2.71634362e-03 2.40746841e-01 3.50429386e-01 -3.40719342e-01
6.58702374e-01 -5.85870683e-01 -2.50825018e-01 7.11989343e-01
-1.57537252e-01 -3.93396586e-01 1.01650365e-01 3.56082797e-01
1.10088587e+00 -6.23101071e-02 5.09963572e-01 -6.54155493e-01
4.36545163e-01 8.73262808e-02 2.55413949e-01 7.93237984e-01
1.66178912e-01 3.43972534e-01 8.63914609e-01 -1.03874981e+00
-1.00019538e+00 -1.31182396e+00 5.63955940e-02 1.30178559e+00
-1.26500383e-01 -8.70511591e-01 -5.47450781e-01 -5.36262810e-01
1.30844742e-01 8.65329027e-01 -7.94489026e-01 -4.92230892e-01
-8.00383925e-01 -4.10698682e-01 4.11078185e-01 1.02934623e+00
4.31868523e-01 -1.23033774e+00 -8.51577401e-01 2.97002465e-01
9.65017006e-02 -6.65494025e-01 -5.92373967e-01 6.13268673e-01
-1.20996964e+00 -9.68971908e-01 1.21143311e-01 -7.48329759e-01
6.29468501e-01 -8.40372220e-02 1.53388679e+00 3.37965250e-01
-4.33734417e-01 1.81654274e-01 2.94449218e-02 -3.29466522e-01
-4.16763335e-01 5.04104495e-01 -1.38883591e-01 -4.52419400e-01
2.66898692e-01 -6.61039352e-01 -3.85543466e-01 1.43384799e-01
-7.87469327e-01 6.31600320e-02 4.11604077e-01 1.04878700e+00
2.88693756e-01 -2.44338229e-01 7.97814250e-01 -1.05691385e+00
1.02490067e+00 -2.97683865e-01 -3.50789517e-01 2.98800260e-01
-7.63622582e-01 6.90514207e-01 1.21545589e+00 -3.11400443e-01
-8.03928077e-01 -2.30295390e-01 -1.04061671e-01 -2.07268134e-01
1.92712843e-02 3.96755368e-01 -2.51490861e-01 -1.93259314e-01
9.26870942e-01 1.44481316e-01 -1.20477043e-01 -4.24535036e-01
5.75748205e-01 4.35117492e-03 5.51715732e-01 -9.66964006e-01
6.86310172e-01 5.84240966e-02 1.84475884e-01 -4.82403517e-01
-9.25331056e-01 3.45411718e-01 -3.89336169e-01 2.83395112e-01
7.63781071e-01 -2.25907132e-01 -9.34220076e-01 -1.22923456e-01
-9.89032686e-01 -4.24462676e-01 -6.32478356e-01 -3.91699731e-01
-6.73713267e-01 -2.04143152e-01 -5.58420122e-01 -5.10087252e-01
-3.71897638e-01 -1.22611344e+00 7.69452989e-01 1.42097875e-01
-8.72151256e-01 -1.39823699e+00 -4.86727923e-01 -2.26338148e-01
8.12959373e-01 6.26008287e-02 1.92333782e+00 -6.83923244e-01
-9.63761747e-01 1.64538354e-01 -2.13431060e-01 -1.50651544e-01
1.44017056e-01 -2.33441234e-01 -9.66438770e-01 -1.00139074e-01
-3.14712226e-01 -4.93940413e-01 8.55472445e-01 1.64437056e-01
1.67771280e+00 -5.23346961e-01 -4.31463450e-01 7.05964863e-01
1.14968395e+00 3.56648773e-01 5.16837895e-01 1.32760495e-01
7.16130555e-01 4.09195274e-01 -6.10351786e-02 6.50324896e-02
2.54544705e-01 5.52419543e-01 4.17382836e-01 6.13688529e-02
-1.01226931e-02 -7.74613559e-01 -1.18935686e-02 2.96772093e-01
8.32647458e-02 -3.03990245e-01 -1.13011038e+00 3.05124879e-01
-1.59262943e+00 -9.63753045e-01 1.86537698e-01 1.84326720e+00
7.69652784e-01 5.40735900e-01 -3.29326987e-02 4.31413949e-01
2.14373544e-01 -1.12435808e-02 -6.96238577e-01 -9.99192834e-01
8.47877655e-03 6.50377870e-01 9.12604108e-02 7.11079061e-01
-7.48301744e-01 1.08737302e+00 8.26811981e+00 7.00848818e-01
-1.06189835e+00 -3.15672576e-01 8.34831655e-01 -4.14196253e-02
-7.70195961e-01 -4.42876779e-02 -6.24303818e-01 -5.01492247e-02
8.88017952e-01 -4.46005583e-01 8.07096183e-01 6.77347839e-01
-1.66804910e-01 3.71666700e-01 -1.62068474e+00 8.41145873e-01
-2.74209291e-01 -1.75934958e+00 7.75735438e-01 -3.80201072e-01
4.22495574e-01 -5.93529403e-01 4.96570319e-01 5.07008553e-01
4.62565839e-01 -1.67928207e+00 7.79180646e-01 2.30126977e-01
8.01206470e-01 -6.39210463e-01 9.48491022e-02 -9.62864459e-02
-1.23268080e+00 -7.45891690e-01 -2.32658207e-01 -6.13910258e-01
-1.57275438e-01 -1.84907883e-01 -8.23597610e-01 -3.42562539e-03
5.66404998e-01 8.88164103e-01 -8.35958362e-01 7.54351735e-01
-5.16156137e-01 4.72984314e-01 -1.11038238e-01 -3.98460209e-01
4.96493101e-01 2.25083828e-01 3.83617878e-01 1.05023110e+00
1.23613440e-01 2.50140220e-01 -1.05501123e-01 1.26129627e+00
-2.25744560e-01 -7.72285089e-02 -8.14214051e-01 -2.71674395e-01
3.98187935e-01 8.95552874e-01 -8.60565543e-01 -2.82449424e-01
-6.50864765e-02 4.91341680e-01 7.55595982e-01 4.63081598e-01
-5.00145733e-01 -3.06657642e-01 9.03527856e-01 3.49191815e-01
3.66178781e-01 -6.28922358e-02 -9.69893038e-01 -1.04091144e+00
-8.61553326e-02 -1.16512036e+00 6.72456861e-01 -1.03606629e+00
-1.16144955e+00 6.31004512e-01 7.14462399e-02 -7.27741420e-01
-2.35672712e-01 -1.06824172e+00 -7.23847151e-01 7.59783983e-01
-9.20182168e-01 -9.88454640e-01 -4.52353209e-02 3.57074589e-01
7.03926682e-01 -1.87095348e-02 9.36603367e-01 -5.30101322e-02
-3.12816501e-01 8.29411924e-01 -2.89269120e-01 -3.98335397e-01
2.94817016e-02 -1.67090452e+00 8.89375091e-01 4.66422051e-01
2.50900567e-01 1.09687221e+00 8.71888578e-01 -1.45683780e-01
-1.75889492e+00 -8.82950962e-01 8.53810728e-01 -6.85769320e-01
7.30309367e-01 -5.98928928e-01 -1.07927525e+00 1.12567806e+00
2.00914398e-01 -2.34659865e-01 5.55439055e-01 4.43988085e-01
-8.70031476e-01 -2.08801746e-01 -1.09162664e+00 1.02442241e+00
1.42642295e+00 -5.19282639e-01 -8.93124461e-01 9.65068638e-02
9.71119463e-01 -3.01634699e-01 -5.20750880e-01 2.00812921e-01
5.72331905e-01 -7.51945913e-01 8.98986936e-01 -1.43336594e+00
6.78628206e-01 -1.79986447e-01 -1.66181773e-01 -1.50916994e+00
-5.81341743e-01 -9.23585474e-01 -3.64426486e-02 6.77248180e-01
6.84107244e-01 -8.74696970e-01 5.22306263e-01 5.65837085e-01
-4.60767120e-01 -1.15666282e+00 -6.71814680e-01 -5.76485336e-01
4.31777120e-01 -4.23500985e-01 9.18159783e-01 7.51289964e-01
3.70964110e-01 8.53731036e-01 7.58029073e-02 -3.06289166e-01
1.58906654e-01 4.22851205e-01 5.18901110e-01 -1.19565964e+00
-3.06286097e-01 -1.08174217e+00 -3.48831534e-01 -1.34743237e+00
1.51209921e-01 -1.22520232e+00 -1.96038648e-01 -1.38661063e+00
-9.00939405e-02 -5.07921636e-01 -1.18817382e-01 5.88791072e-01
-1.44983873e-01 1.00852534e-01 2.91074902e-01 -4.81848754e-02
-5.00622272e-01 5.77543199e-01 1.38481164e+00 -3.22407275e-01
-1.19337380e-01 -3.35033059e-01 -1.33558619e+00 4.77322072e-01
7.14033782e-01 -2.24198923e-01 -7.91116714e-01 -6.19974136e-01
7.42308438e-01 -1.48233309e-01 3.64658684e-01 -9.67756748e-01
4.66513596e-02 -2.05505490e-01 4.19656694e-01 -2.60456413e-01
1.22892270e-02 -7.60443747e-01 1.33869022e-01 6.51855826e-01
-8.97367179e-01 8.62003088e-01 7.96165705e-01 4.74732548e-01
2.25661602e-02 1.37513757e-01 5.08131325e-01 -2.63030201e-01
-7.33172417e-01 2.65400469e-01 -5.29180050e-01 5.39380133e-01
6.62686467e-01 -2.69421399e-01 -5.14737487e-01 -3.37511152e-01
-8.69063437e-01 4.83668506e-01 4.37263668e-01 3.98267418e-01
7.64162540e-01 -1.31896794e+00 -2.34679103e-01 4.52680707e-01
1.94942117e-01 6.00798056e-02 -1.29576564e-01 4.80900019e-01
-7.30717361e-01 6.75098300e-01 -3.75168085e-01 -4.35113549e-01
-7.71829009e-01 9.17535305e-01 6.15248382e-01 -3.63880575e-01
-7.37518549e-01 9.51332569e-01 6.79597437e-01 -2.48670265e-01
1.26997873e-01 -9.63238180e-01 2.34785020e-01 -1.91644937e-01
6.00265741e-01 2.40217112e-02 -1.99382827e-02 7.33780637e-02
-2.38155693e-01 6.13380373e-01 -4.13163990e-01 2.20854923e-01
1.21155691e+00 9.64570418e-02 -2.07224160e-01 4.37247962e-01
8.59873593e-01 -3.90951037e-01 -8.75222206e-01 1.48720834e-02
2.73968846e-01 -4.36277002e-01 -2.97876716e-01 -9.62464631e-01
-7.46503353e-01 1.27564466e+00 8.86163563e-02 4.59501833e-01
1.21921146e+00 7.52870599e-03 7.35194325e-01 9.43005741e-01
1.99823141e-01 -2.94317544e-01 2.40567461e-01 7.42571712e-01
1.17827690e+00 -7.01383889e-01 -1.71365693e-01 -2.87097096e-01
-3.76898289e-01 1.26123428e+00 7.76379526e-01 -1.22730218e-01
3.71941984e-01 3.04358572e-01 -1.61790133e-01 -2.79467702e-01
-1.37435436e+00 1.66054107e-02 2.90890634e-01 7.46287167e-01
5.17434716e-01 -1.18019141e-01 1.47099376e-01 5.09168446e-01
-7.20542192e-01 -3.04008305e-01 4.90195870e-01 8.62486303e-01
-5.52742124e-01 -7.43647754e-01 -1.03330366e-01 5.70042074e-01
-1.86111838e-01 -3.91729742e-01 -5.43374479e-01 9.53566909e-01
-3.29639912e-01 3.61916512e-01 2.40960911e-01 -3.85937393e-01
4.00166124e-01 3.80335361e-01 6.99229360e-01 -5.86162686e-01
-9.61930215e-01 -6.10508382e-01 6.66442933e-03 -6.19701982e-01
2.27491453e-01 -2.78577507e-01 -1.41661155e+00 -6.44498348e-01
2.73614705e-01 2.20775604e-01 -1.10204525e-01 8.24926853e-01
3.81125480e-01 6.96133137e-01 3.17548625e-02 -5.85084617e-01
-7.06255198e-01 -5.51132798e-01 -1.58039898e-01 3.86542469e-01
5.37808180e-01 -6.89017057e-01 -1.79935634e-01 -2.86166760e-04] | [9.476909637451172, 7.159472465515137] |
eb06d6fc-5bc3-4047-86ef-e9cd49658481 | towards-global-optimality-in-cooperative-marl | 2207.11143 | null | https://arxiv.org/abs/2207.11143v3 | https://arxiv.org/pdf/2207.11143v3.pdf | Towards Global Optimality in Cooperative MARL with the Transformation And Distillation Framework | Decentralized execution is one core demand in cooperative multi-agent reinforcement learning (MARL). Recently, most popular MARL algorithms have adopted decentralized policies to enable decentralized execution and use gradient descent as their optimizer. However, there is hardly any theoretical analysis of these algorithms taking the optimization method into consideration, and we find that various popular MARL algorithms with decentralized policies are suboptimal in toy tasks when gradient descent is chosen as their optimization method. In this paper, we theoretically analyze two common classes of algorithms with decentralized policies -- multi-agent policy gradient methods and value-decomposition methods to prove their suboptimality when gradient descent is used. In addition, we propose the Transformation And Distillation (TAD) framework, which reformulates a multi-agent MDP as a special single-agent MDP with a sequential structure and enables decentralized execution by distilling the learned policy on the derived ``single-agent" MDP. This approach uses a two-stage learning paradigm to address the optimization problem in cooperative MARL, maintaining its performance guarantee. Empirically, we implement TAD-PPO based on PPO, which can theoretically perform optimal policy learning in the finite multi-agent MDPs and shows significant outperformance on a large set of cooperative multi-agent tasks. | ['Chongjie Zhang', 'Jianhao Wang', 'Chenghao Li', 'Jianing Ye'] | 2022-07-12 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [-6.57832980e-01 1.40382405e-02 -6.16899908e-01 8.25228915e-02
-8.43391955e-01 -3.91909957e-01 4.23438072e-01 3.41852382e-02
-7.93188095e-01 1.37100315e+00 4.63602841e-02 -5.16139686e-01
-3.34086210e-01 -4.32889700e-01 -7.24302471e-01 -1.16708314e+00
-3.92329097e-01 8.02561224e-01 9.38562974e-02 -4.96854842e-01
2.27100495e-02 1.64090648e-01 -1.13124764e+00 -2.82435287e-02
1.12523389e+00 8.27302933e-01 4.75752085e-01 8.96679819e-01
2.64237076e-01 1.24518681e+00 -7.81316340e-01 2.80634575e-02
4.37678397e-01 -1.76857352e-01 -8.29933226e-01 2.03959733e-01
-2.39889801e-01 -7.22270548e-01 -2.97642350e-01 9.90023851e-01
7.80602217e-01 3.07918936e-01 4.47637171e-01 -1.71796751e+00
-2.93444216e-01 8.64034235e-01 -9.64340150e-01 2.00177282e-01
-3.99265578e-03 4.29480463e-01 1.09983158e+00 -2.08486095e-01
4.24544096e-01 1.42200923e+00 2.94481397e-01 6.78179264e-01
-1.03691995e+00 -3.17035079e-01 6.48125172e-01 3.97977829e-01
-7.92876840e-01 -7.41733015e-02 6.38916194e-01 -1.89303339e-01
1.13626754e+00 -1.74801812e-01 5.56709230e-01 7.61485994e-01
5.92571616e-01 1.24418354e+00 1.49044478e+00 -3.50022227e-01
5.57439685e-01 -7.37151727e-02 -2.67316461e-01 7.80175328e-01
3.81328583e-01 1.82269618e-01 -4.43972230e-01 -5.96776009e-01
4.61407304e-01 -1.05792008e-01 6.88251480e-02 -4.54634905e-01
-1.41551208e+00 8.91046762e-01 9.50788781e-02 1.49061661e-02
-7.63138235e-01 5.15084326e-01 7.30171621e-01 9.13603306e-01
3.78607482e-01 3.95851612e-01 -6.84814692e-01 -8.32059085e-02
-4.14528549e-01 6.31826639e-01 1.09232485e+00 7.12321997e-01
8.91622961e-01 4.39364433e-01 -1.72354907e-01 4.98907417e-01
3.06889206e-01 6.94786370e-01 6.63176417e-01 -1.49704635e+00
6.63655281e-01 1.05786465e-01 6.55457675e-01 -8.11246157e-01
-5.06078005e-01 -2.64054745e-01 -9.08237934e-01 6.45297050e-01
1.92647710e-01 -8.84011686e-01 6.42726570e-02 1.68902159e+00
7.44280279e-01 -3.79166529e-02 6.45363569e-01 9.13452506e-01
-2.91898586e-02 6.11619532e-01 -2.43246064e-01 -8.29544842e-01
1.00621676e+00 -1.41761732e+00 -7.26804018e-01 -1.73904989e-02
9.59160924e-01 -3.37993234e-01 8.02383125e-01 5.80805421e-01
-1.03389680e+00 -1.28522933e-01 -1.01897192e+00 5.86791098e-01
1.10264905e-01 2.25958899e-01 4.65035766e-01 1.99559703e-01
-1.16181016e+00 6.68355644e-01 -9.20985699e-01 9.22670290e-02
3.54092778e-03 5.01169145e-01 3.69312949e-02 2.87868232e-01
-9.53997195e-01 8.73309433e-01 6.59135282e-01 -8.43497589e-02
-1.66281545e+00 -2.01122731e-01 -3.25084746e-01 -8.25029984e-02
1.06015110e+00 -8.32947612e-01 1.81931794e+00 -8.98032725e-01
-2.06267643e+00 2.40845725e-01 2.16202945e-01 -6.03870034e-01
6.87990427e-01 -3.43640000e-01 7.54514709e-02 3.20039272e-01
2.14462340e-01 9.94938314e-02 1.19531631e+00 -1.28014183e+00
-1.24848628e+00 -1.86593868e-02 6.91907883e-01 7.78113782e-01
-3.82107109e-01 -2.23578602e-01 3.80136281e-01 -2.98364401e-01
-6.48693502e-01 -9.48573768e-01 -6.99802637e-01 -6.64223313e-01
-4.86102365e-02 -6.95799291e-01 8.05371106e-01 -3.53842050e-01
1.03686523e+00 -1.85783279e+00 4.84703511e-01 -9.95730758e-02
6.03575349e-01 7.83287063e-02 -2.54838020e-01 7.06848025e-01
5.35098314e-01 -2.11555839e-01 2.42555365e-01 -2.79481441e-01
5.21905303e-01 5.32001376e-01 -2.86872774e-01 6.23632073e-01
-3.29454482e-01 7.12906003e-01 -1.26104236e+00 -6.89695656e-01
-1.15576603e-01 -4.75748271e-01 -6.72060490e-01 4.53324080e-01
-7.58380413e-01 3.15229833e-01 -9.72513080e-01 3.95052582e-01
4.08272117e-01 -1.90426603e-01 7.85687923e-01 1.97705224e-01
-2.44164243e-01 -1.68937922e-01 -1.27027047e+00 1.49773097e+00
-5.02120733e-01 1.88456446e-01 7.65591860e-01 -1.41481113e+00
5.79154849e-01 5.23237526e-01 1.05123401e+00 -4.22100067e-01
-6.53073266e-02 3.53376210e-01 1.06146447e-01 -5.91383398e-01
3.91818732e-01 1.57918051e-01 -1.45780221e-01 8.93367946e-01
-1.45547062e-01 -1.05001099e-01 4.62384522e-01 3.44066292e-01
1.34600282e+00 2.09031567e-01 4.38253373e-01 -4.33445007e-01
7.06652224e-01 3.09351683e-01 9.00747061e-01 1.13903975e+00
-5.10417163e-01 -5.00025749e-01 8.70695710e-01 -4.31719840e-01
-1.01225352e+00 -5.94842494e-01 7.04594731e-01 1.55784917e+00
1.00599162e-01 -2.90768325e-01 -5.52703381e-01 -1.09509778e+00
2.90323466e-01 2.79890954e-01 -2.76887774e-01 1.91755414e-01
-7.44152248e-01 -1.00260854e+00 2.62390167e-01 -2.30605397e-02
7.43553996e-01 -1.04703557e+00 -9.30164397e-01 6.29834354e-01
-9.12681147e-02 -9.64062154e-01 -5.39284348e-01 1.59607172e-01
-6.82015181e-01 -1.24895322e+00 -7.30782330e-01 -7.79038429e-01
5.69455802e-01 3.36381912e-01 9.51390386e-01 -6.98653609e-02
2.97651440e-01 9.68567729e-01 -4.21584159e-01 -2.84920961e-01
-6.51688039e-01 3.92640382e-01 6.67291164e-01 1.75790454e-03
-2.47682333e-01 -5.19480407e-01 -7.01757669e-01 2.80129761e-01
-7.07676888e-01 -5.17937914e-02 8.19690049e-01 1.14769626e+00
4.70774084e-01 1.73966020e-01 9.57053721e-01 -5.45563638e-01
1.31743455e+00 -4.97395843e-01 -9.29444194e-01 5.29584885e-01
-8.15412343e-01 3.83379579e-01 1.24215281e+00 -5.77129006e-01
-1.08199275e+00 -2.13911533e-01 3.41204047e-01 -5.66092968e-01
1.94311813e-01 5.19348860e-01 2.61565298e-01 5.29394969e-02
5.27937472e-01 6.14660025e-01 5.22527456e-01 -2.54685819e-01
4.72290277e-01 6.02489293e-01 2.11791769e-02 -1.36693192e+00
5.67842364e-01 3.54282588e-01 1.47797465e-01 -3.81679296e-01
-6.82318091e-01 -2.31826559e-01 1.17857358e-03 -2.66088277e-01
3.99899900e-01 -9.26089585e-01 -1.71481216e+00 6.15818739e-01
-1.05559802e+00 -9.00590956e-01 -1.98879465e-01 7.29983211e-01
-1.12643111e+00 5.87652564e-01 -7.39583075e-01 -8.45366001e-01
-5.24292648e-01 -1.32867503e+00 6.07597291e-01 7.77033642e-02
6.49710238e-01 -1.07210743e+00 5.60639024e-01 -7.70700052e-02
4.29781854e-01 4.04628515e-02 7.42392719e-01 -5.90908766e-01
-3.50007445e-01 3.73675793e-01 1.69747487e-01 4.46749389e-01
6.46685287e-02 -2.73506969e-01 -3.53206486e-01 -1.19464350e+00
1.08832002e-01 -8.24649394e-01 3.24001163e-01 3.63736629e-01
6.80980623e-01 -8.41256618e-01 -2.07207739e-01 2.62295604e-01
1.52807093e+00 1.81950316e-01 -2.46755451e-01 7.78106570e-01
4.03239071e-01 3.42170149e-01 1.02461374e+00 1.01058531e+00
7.24067152e-01 3.62204969e-01 7.66114891e-01 3.29654902e-01
4.59773690e-01 2.69258041e-02 9.83017087e-01 1.10860288e+00
-2.16265142e-01 3.86743322e-02 -6.58571541e-01 2.28253171e-01
-2.80401278e+00 -8.93349767e-01 4.64427948e-01 1.90869606e+00
9.93945360e-01 -1.99031994e-01 5.57811081e-01 -5.48621178e-01
5.11504650e-01 2.60477215e-01 -1.02371299e+00 -4.78754669e-01
-8.61499310e-02 -4.28773582e-01 5.83950579e-01 4.37349349e-01
-9.94706750e-01 9.84415829e-01 6.22366428e+00 9.84852970e-01
-9.33877587e-01 5.87046921e-01 3.06798905e-01 -1.59359381e-01
9.97290984e-02 -2.36305688e-02 -9.11877811e-01 4.68087226e-01
6.25144184e-01 -7.21880257e-01 9.15550947e-01 1.19521892e+00
7.49102354e-01 -2.64524400e-01 -7.99644053e-01 8.92846227e-01
-4.40602034e-01 -1.15283501e+00 -2.73335844e-01 1.49227217e-01
1.10444963e+00 2.88041115e-01 -9.06346813e-02 6.53104007e-01
1.14824963e+00 -4.73195702e-01 5.39075792e-01 2.15569034e-01
2.15516955e-01 -8.49097013e-01 5.50385535e-01 8.92653465e-01
-1.14163077e+00 -5.92863977e-01 -6.04253411e-01 -3.66904855e-01
-1.08882591e-01 2.23481387e-01 -6.58846140e-01 7.50991046e-01
6.34015858e-01 5.66180766e-01 6.17208444e-02 6.90640450e-01
-2.44061351e-01 3.31722677e-01 -3.02741110e-01 -4.71775830e-01
7.41972923e-01 -4.64872837e-01 7.30947912e-01 6.26995444e-01
-3.97053408e-03 -1.76198497e-01 1.02963793e+00 2.79855460e-01
2.69675460e-02 1.87233493e-01 -4.46775079e-01 -7.60556385e-03
3.96547228e-01 1.32399321e+00 -4.18873489e-01 -4.91487056e-01
-4.59175020e-01 6.73404217e-01 8.96817446e-01 6.23950839e-01
-8.16661358e-01 -2.30446368e-01 8.99032950e-01 -4.75842953e-01
2.62233973e-01 -6.56879783e-01 4.92554963e-01 -1.44023871e+00
6.90876544e-02 -1.35942984e+00 4.08717126e-01 -2.01755539e-01
-1.36586416e+00 3.97582471e-01 4.36510481e-02 -1.20289516e+00
-8.41751635e-01 -4.50963646e-01 -5.75987041e-01 3.14928472e-01
-1.70197463e+00 -8.35314691e-01 3.41155261e-01 8.04317236e-01
6.34372890e-01 -6.63858533e-01 4.63692605e-01 -1.26741705e-02
-8.49850178e-01 3.48306477e-01 1.00406516e+00 -2.29874134e-01
6.44711375e-01 -1.42287159e+00 -4.14728373e-01 6.52647376e-01
-5.11828661e-01 1.61862206e-02 7.31606305e-01 -5.12678802e-01
-1.94506860e+00 -9.31105077e-01 -4.30073626e-02 1.60074517e-01
1.06608415e+00 3.35899070e-02 -3.08190227e-01 6.59194410e-01
6.67150378e-01 -1.41686052e-01 8.29401836e-02 -1.41456664e-01
1.98681071e-01 -3.97068232e-01 -8.91481936e-01 6.18895411e-01
6.57716274e-01 -6.06001392e-02 -3.68895650e-01 7.54383266e-01
9.52790022e-01 -4.30955529e-01 -9.28268373e-01 9.55457762e-02
1.30819812e-01 -7.27850497e-01 5.74818015e-01 -9.11895096e-01
1.77759737e-01 -3.59726697e-01 -9.89342332e-02 -1.84216213e+00
-1.87473118e-01 -1.33000672e+00 -7.46721685e-01 7.08379269e-01
3.84679176e-02 -1.03554714e+00 7.41577983e-01 1.17874995e-01
-1.63790986e-01 -9.51562703e-01 -1.06199014e+00 -1.17726326e+00
2.35203117e-01 1.18827708e-01 4.37096506e-01 7.08786130e-01
2.36729085e-01 2.55526245e-01 -7.73624420e-01 1.47848979e-01
9.72106218e-01 2.13622779e-01 1.10142446e+00 -3.86824042e-01
-9.41555679e-01 -4.67882216e-01 4.09952015e-01 -1.33050418e+00
6.17116570e-01 -7.25393236e-01 3.17785982e-03 -1.40862036e+00
-5.18285786e-04 -5.72903991e-01 -2.35058978e-01 4.49598193e-01
4.13491130e-02 -7.16063142e-01 5.32225966e-01 6.21251345e-01
-1.28954124e+00 9.25722599e-01 1.48921323e+00 -3.23296189e-01
-4.20161456e-01 1.65546656e-01 -4.00344133e-01 6.99554563e-01
1.06460035e+00 -5.61168373e-01 -4.85360354e-01 -4.66986954e-01
3.16315979e-01 4.09882247e-01 -4.73698042e-02 -6.41542196e-01
3.70031923e-01 -9.08984780e-01 -4.78391230e-01 -2.93804228e-01
-2.78713088e-03 -6.60292387e-01 -3.61584455e-01 1.05582094e+00
-3.40938479e-01 6.00223601e-01 -3.20798337e-01 7.66874909e-01
-7.21627325e-02 -3.75691384e-01 4.95409817e-01 -3.80820930e-01
-8.03444386e-01 4.27580535e-01 -6.73331320e-01 2.76419014e-01
1.42804313e+00 5.34958601e-01 -4.96693283e-01 -3.62514853e-01
-7.50204504e-01 1.03888917e+00 1.16166048e-01 -1.22960374e-01
4.76783365e-01 -1.13626194e+00 -9.11449492e-01 -4.36808914e-01
-5.02256155e-01 -1.11789882e-01 9.39726159e-02 1.13127828e+00
-4.33567315e-01 2.06655160e-01 -1.69404611e-01 -3.78481209e-01
-9.57069218e-01 6.89531028e-01 5.78282118e-01 -1.10944653e+00
-4.55524504e-01 1.08044952e-01 -1.46122336e-01 -4.05519873e-01
2.12236419e-01 -3.34397964e-02 3.50602828e-02 1.22078462e-02
2.03817487e-01 7.36152291e-01 -3.80718797e-01 4.63196188e-02
-6.97894320e-02 1.19584933e-01 -2.84852386e-01 -3.89795870e-01
1.34490609e+00 -3.40262532e-01 -3.53073716e-01 1.93992376e-01
8.06376755e-01 -3.67411196e-01 -1.51600409e+00 -6.60374105e-01
-3.54878455e-02 -2.70272940e-01 -1.26616329e-01 -3.07926625e-01
-9.99779582e-01 3.49655509e-01 2.27937579e-01 4.71499741e-01
8.32574010e-01 -2.47955158e-01 6.15764141e-01 1.20812738e+00
8.99133921e-01 -1.73494351e+00 5.50724149e-01 8.56947005e-01
7.85800338e-01 -1.28680027e+00 1.41187489e-01 3.09004068e-01
-9.91755009e-01 1.14211190e+00 9.72239614e-01 -4.54373062e-01
5.86326182e-01 2.29655936e-01 -1.97362691e-01 4.98194806e-02
-1.38340831e+00 -2.98261017e-01 -5.23823678e-01 6.02536678e-01
-2.34759808e-01 3.18785310e-01 -6.82960868e-01 4.21687573e-01
9.64558125e-02 2.10670102e-02 6.54410064e-01 1.27153134e+00
-6.71206474e-01 -1.11969066e+00 -3.93779725e-01 4.37726408e-01
-3.30674887e-01 3.45030665e-01 1.83739632e-01 7.15414107e-01
-2.70951718e-01 9.00919318e-01 -2.83247977e-01 -1.69385016e-01
1.80572271e-02 -3.65580797e-01 4.89282817e-01 -3.34393144e-01
-7.56093085e-01 1.83924153e-01 -1.49374651e-02 -5.05133569e-01
-5.65672994e-01 -4.49856997e-01 -1.27314126e+00 -6.35031641e-01
-4.69854102e-02 4.86313194e-01 4.31015402e-01 1.10846257e+00
4.33300436e-01 4.20406282e-01 1.24206018e+00 -8.47159326e-01
-1.73720443e+00 -6.76425278e-01 -7.92691350e-01 2.72075329e-02
4.55768526e-01 -6.28071129e-01 -2.41293758e-01 -3.96519929e-01] | [3.7733094692230225, 2.0581250190734863] |
3fcdcdce-0f55-48e1-bcc3-701cbeb638c6 | pedestrian-attribute-recognition-in-video | 2106.06485 | null | https://arxiv.org/abs/2106.06485v1 | https://arxiv.org/pdf/2106.06485v1.pdf | Pedestrian Attribute Recognition in Video Surveillance Scenarios Based on View-attribute Attention Localization | Pedestrian attribute recognition in surveillance scenarios is still a challenging task due to inaccurate localization of specific attributes. In this paper, we propose a novel view-attribute localization method based on attention (VALA), which relies on the strong relevance between attributes and views to capture specific view-attributes and to localize attribute-corresponding areas by attention mechanism. A specific view-attribute is composed by the extracted attribute feature and four view scores which are predicted by view predictor as the confidences for attribute from different views. View-attribute is then delivered back to shallow network layers for supervising deep feature extraction. To explore the location of a view-attribute, regional attention is introduced to aggregate spatial information of the input attribute feature in height and width direction for constraining the image into a narrow range. Moreover, the inter-channel dependency of view-feature is embedded in the above two spatial directions. An attention attribute-specific region is gained after fining the narrow range by balancing the ratio of channel dependencies between height and width branches. The final view-attribute recognition outcome is obtained by combining the output of regional attention with the view scores from view predictor. Experiments on three wide datasets (RAP, RAPv2, PETA, and PA-100K) demonstrate the effectiveness of our approach compared with state-of-the-art methods. | ['Linlin Ou', 'Xinyi Yu', 'Weichen Chen'] | 2021-06-11 | null | null | null | null | ['pedestrian-attribute-recognition'] | ['computer-vision'] | [ 2.17334218e-02 -1.02648355e-01 3.23297717e-02 -8.85262847e-01
-6.32233143e-01 -3.47535014e-01 5.49795270e-01 -4.59437966e-02
-2.74125814e-01 4.83690709e-01 4.34024304e-01 1.05600528e-01
4.33900580e-02 -9.47981417e-01 -8.34225833e-01 -8.38098347e-01
1.77409753e-01 2.26467207e-01 4.58921224e-01 -7.30145648e-02
2.06260815e-01 4.44714099e-01 -1.70330954e+00 6.27490997e-01
6.23123050e-01 1.61967182e+00 2.09124058e-01 3.12706232e-01
-6.67401180e-02 3.96427006e-01 -4.64188337e-01 -4.66070265e-01
2.64781147e-01 -1.80160344e-01 -1.71256512e-01 1.07829362e-01
7.74132252e-01 -4.15979087e-01 -2.32424483e-01 9.40912902e-01
5.04288554e-01 8.07674974e-03 7.42929399e-01 -1.31220543e+00
-6.20803356e-01 2.36053616e-01 -6.15470469e-01 3.93241405e-01
2.77707011e-01 3.90468061e-01 9.84739482e-01 -1.10553420e+00
3.13558340e-01 1.30354726e+00 3.46237332e-01 2.14295328e-01
-9.29804087e-01 -5.41397750e-01 1.02756953e+00 6.53740704e-01
-1.39433444e+00 -2.41060242e-01 1.03538477e+00 -3.47750336e-01
9.99903619e-01 2.59707302e-01 9.79641795e-01 1.21742511e+00
3.38510841e-01 8.41368735e-01 9.62868273e-01 5.67587242e-02
2.27062013e-02 2.71630466e-01 -9.00217611e-03 7.41517067e-01
-3.27089243e-02 7.40315467e-02 -6.90322220e-01 7.85382912e-02
4.57355887e-01 1.80916935e-01 -1.98814198e-01 -6.10441446e-01
-1.13319731e+00 6.55907154e-01 7.70533502e-01 -4.29964423e-01
-6.41276062e-01 -2.54276127e-01 3.12437832e-01 -9.13153291e-02
2.68154114e-01 9.70718116e-02 -6.77647471e-01 1.86487213e-01
-3.22442591e-01 -4.80863824e-02 3.76679957e-01 1.26067793e+00
8.19616199e-01 -1.50529280e-01 -5.99953294e-01 6.04695797e-01
5.13932943e-01 7.25835383e-01 -1.38248980e-01 -7.01866210e-01
8.17059278e-01 1.06961465e+00 5.48036098e-02 -8.28534782e-01
-3.79101396e-01 -5.62332153e-01 -8.53433490e-01 2.55908370e-01
3.36286664e-01 -8.00278932e-02 -9.44611669e-01 1.63392675e+00
4.56417412e-01 -1.61124747e-02 6.24207268e-03 1.13524008e+00
1.03459024e+00 6.32294297e-01 4.02180731e-01 9.84021649e-02
1.68158042e+00 -7.75815070e-01 -2.03642026e-01 -5.39092541e-01
7.06414431e-02 -4.61114675e-01 9.21343565e-01 1.48998231e-01
-6.92214012e-01 -9.27859426e-01 -9.67522323e-01 9.02887732e-02
-6.11449659e-01 5.15104711e-01 4.21133012e-01 3.74164373e-01
-6.77886248e-01 -8.12873468e-02 -2.17680320e-01 -4.13118005e-01
7.61801600e-01 3.65223020e-01 -5.18438518e-01 -6.59003034e-02
-1.27006590e+00 7.45334268e-01 1.25862688e-01 3.71177912e-01
-9.39637125e-01 -6.31095231e-01 -1.05401492e+00 3.18588763e-01
3.74563068e-01 -9.45888519e-01 4.62841779e-01 -9.73391771e-01
-1.24919832e+00 7.57828474e-01 -2.89159298e-01 -1.43371388e-01
4.83622849e-01 -4.21097249e-01 -4.59539235e-01 7.55939074e-03
3.48631829e-01 8.12214851e-01 7.80617297e-01 -1.26482701e+00
-9.99402285e-01 -7.87913620e-01 5.60493395e-02 6.14402771e-01
-1.92054719e-01 -3.06419760e-01 -5.54765522e-01 -3.10858220e-01
2.73971587e-01 -5.09004474e-01 -6.94515482e-02 1.22717403e-01
-5.98243594e-01 -2.14524105e-01 7.67758608e-01 -5.19478619e-01
9.31129456e-01 -2.14063144e+00 2.03078330e-01 2.45621324e-01
1.13939457e-01 9.94564146e-02 -2.94585079e-02 1.76882111e-02
-5.88316135e-02 -2.81932950e-01 3.27281505e-02 1.15364924e-01
-2.76027083e-01 1.97717100e-02 -2.73579538e-01 5.65834224e-01
5.03363609e-01 7.75974691e-01 -7.58797526e-01 -6.40753388e-01
4.92447346e-01 6.02434576e-01 -2.99165517e-01 4.26150203e-01
-4.65510264e-02 2.84961641e-01 -8.12053502e-01 9.08639371e-01
8.01860511e-01 1.60064504e-01 -2.15336919e-01 -5.69456875e-01
-2.19453871e-01 -1.43610269e-01 -1.02551723e+00 1.33190322e+00
-2.78097361e-01 4.95124757e-01 -3.18219066e-02 -7.56182194e-01
1.21197867e+00 -9.11189914e-02 2.10102051e-01 -6.22297466e-01
1.63493440e-01 -3.84666622e-01 -2.49254674e-01 -7.17110574e-01
1.23411223e-01 4.82385397e-01 -2.92977303e-01 -4.56539355e-02
1.03325238e-02 4.51570958e-01 -3.17830920e-01 4.83208448e-02
4.71930325e-01 3.50498110e-01 3.22186172e-01 -2.60155171e-01
1.02451551e+00 -3.22683245e-01 6.42872632e-01 7.04590201e-01
-6.06851101e-01 5.65385938e-01 7.52708972e-01 -1.00885165e+00
-8.94951522e-01 -1.16201484e+00 1.08505888e-02 1.38082027e+00
4.25948530e-01 -1.62815914e-01 -6.13875151e-01 -1.18444979e+00
-2.01478787e-02 7.02607632e-01 -1.07470286e+00 -2.96746224e-01
-5.96623123e-01 -4.52152222e-01 6.00950420e-02 1.01645792e+00
9.05706227e-01 -1.21544909e+00 -8.72513831e-01 -1.84337180e-02
-2.43994057e-01 -1.16846049e+00 -5.19565105e-01 1.49270162e-01
-3.61251712e-01 -1.00900233e+00 -5.49141705e-01 -5.28657794e-01
7.18937337e-01 2.06079796e-01 1.07978129e+00 -1.19426094e-01
-4.82668877e-02 2.77606308e-01 -1.73277885e-01 -4.88206893e-01
2.98097193e-01 1.34235948e-01 -8.78195390e-02 7.28318751e-01
7.25797594e-01 -3.97032171e-01 -9.57407415e-01 5.82637787e-01
-3.92023250e-02 9.01366919e-02 1.04141080e+00 8.71964991e-01
8.06711197e-01 -4.67973143e-01 4.67419893e-01 -5.42194903e-01
-2.51636971e-02 -4.56349581e-01 -7.35934496e-01 3.48490655e-01
-1.89063922e-01 1.00990109e-01 6.91585898e-01 -2.40648940e-01
-1.11961889e+00 4.62698281e-01 -2.81110648e-02 -3.50510269e-01
-6.93937540e-01 -2.62101255e-02 -1.16239476e+00 4.97348070e-01
2.31961578e-01 4.04768050e-01 -3.00065838e-02 1.10586196e-01
3.17950815e-01 4.18501437e-01 5.31425536e-01 -3.09807867e-01
3.12278926e-01 4.88907427e-01 9.36340094e-02 -5.44544041e-01
-1.04828095e+00 -3.20063561e-01 -9.25397933e-01 -6.35757923e-01
1.36123991e+00 -8.97192359e-01 -8.78486812e-01 4.22317356e-01
-1.06143594e+00 1.30909875e-01 -8.30567181e-02 4.43236023e-01
-3.97509873e-01 5.92934228e-02 1.13858052e-01 -7.70779133e-01
-2.39068717e-01 -1.38980269e+00 1.22928059e+00 4.46010917e-01
1.79966718e-01 -4.52542871e-01 -6.39814615e-01 1.21541634e-01
2.11290866e-01 2.68676102e-01 7.83261299e-01 -7.16749609e-01
-7.20926285e-01 -3.15701425e-01 -6.05337501e-01 1.86544526e-02
-2.52797246e-01 -9.76565406e-02 -1.25458968e+00 -6.24030344e-02
-4.42805111e-01 -9.56834927e-02 1.18442917e+00 7.12288857e-01
1.25530958e+00 -1.21285632e-01 -5.84468305e-01 8.47339094e-01
1.22799206e+00 3.87993008e-02 4.84270930e-01 4.16670203e-01
8.49097669e-01 5.84102571e-01 7.96505332e-01 5.18949032e-01
5.79309821e-01 8.24177206e-01 8.03653717e-01 -8.16021860e-02
2.45460924e-02 -2.06040889e-01 2.82564729e-01 -7.89077133e-02
-2.39689648e-01 -2.42473513e-01 -6.20664001e-01 4.22206849e-01
-1.88184714e+00 -1.08307827e+00 -1.95891738e-01 2.35616446e+00
8.24522451e-02 4.44268316e-01 2.14356124e-01 -3.19703609e-01
7.99464285e-01 2.18670011e-01 -7.44201601e-01 -2.20177591e-01
-3.81664306e-01 -6.08004749e-01 4.46774960e-01 3.34547639e-01
-1.60177541e+00 8.26969326e-01 4.97694921e+00 4.60434258e-01
-7.32759953e-01 -3.96138877e-01 1.04508042e+00 -7.85950478e-03
-5.22176139e-02 -2.33263001e-01 -1.28564727e+00 5.55567622e-01
4.01608884e-01 6.03774250e-01 3.97322066e-02 1.13854539e+00
-1.45451263e-01 -8.25865045e-02 -1.10418999e+00 8.29155803e-01
2.77730912e-01 -8.55733752e-01 3.87607425e-01 9.53114033e-02
1.89780518e-01 -2.97542781e-01 3.10290456e-01 3.93156767e-01
1.04705326e-01 -9.04561758e-01 7.80885994e-01 7.65742838e-01
9.08473969e-01 -9.75758970e-01 1.05312467e+00 7.92286620e-02
-1.64332867e+00 -5.57929635e-01 -6.54224217e-01 1.81557611e-01
1.31165951e-01 1.88968480e-01 -4.93548751e-01 5.47862291e-01
9.71236587e-01 8.50422442e-01 -7.34232306e-01 7.37080932e-01
-4.40335900e-01 4.36590850e-01 -1.71383992e-01 6.30105734e-02
2.77892232e-01 -1.24240227e-01 6.52956009e-01 1.19843268e+00
2.48651609e-01 2.86448836e-01 1.23065852e-01 8.70936096e-01
3.11306834e-01 -1.68535680e-01 -6.78774357e-01 9.11309540e-01
6.77761674e-01 1.39583254e+00 -5.46214640e-01 -4.38918501e-01
-6.83878303e-01 9.87141907e-01 3.46581936e-01 5.36018670e-01
-1.04030573e+00 -1.22221410e-01 6.82240546e-01 1.28901750e-01
9.57355797e-01 2.54749358e-01 -4.11056578e-01 -7.63938904e-01
1.77927375e-01 -4.27560747e-01 6.97691917e-01 -7.64662445e-01
-1.32679164e+00 9.23838973e-01 1.08097173e-01 -1.45108926e+00
-3.41807455e-02 -6.44731343e-01 -7.13283539e-01 1.21775925e+00
-1.34534812e+00 -1.62344110e+00 -7.71861434e-01 7.38700807e-01
7.48140097e-01 -5.87535143e-01 7.67406762e-01 -5.76010086e-02
-6.12206101e-01 8.02850425e-01 -5.41978419e-01 2.09883541e-01
7.40995705e-01 -1.19111693e+00 3.93863291e-01 7.04946160e-01
-2.19376564e-01 1.87800467e-01 3.91951144e-01 -5.88409901e-01
-9.99394715e-01 -1.22285748e+00 7.28648305e-01 -8.28086495e-01
4.92788777e-02 -5.22841275e-01 -7.31058240e-01 5.34447730e-01
2.69060642e-01 4.23736811e-01 5.97682178e-01 2.52793748e-02
-6.15621448e-01 -5.46316147e-01 -8.68972182e-01 1.87800735e-01
9.93388951e-01 -2.44318768e-01 -4.28935528e-01 -2.35256538e-01
4.35516119e-01 -1.73649445e-01 -5.38738966e-01 3.77967268e-01
8.32191825e-01 -1.11702275e+00 1.27894521e+00 -6.42671943e-01
3.94810557e-01 -3.73868227e-01 -4.63201672e-01 -1.26271665e+00
-7.94442713e-01 9.45393443e-02 5.83895929e-02 1.37468362e+00
4.76886332e-01 -4.11965668e-01 5.08376896e-01 5.03742754e-01
-3.42390060e-01 -1.13534594e+00 -1.04764199e+00 -1.91372037e-01
-4.16664302e-01 1.10669911e-03 7.90117204e-01 5.38414240e-01
-5.78636646e-01 5.99985480e-01 -2.27944836e-01 6.03121996e-01
7.58594215e-01 2.88322687e-01 7.24086702e-01 -1.29795754e+00
8.25630873e-02 -4.70678627e-01 -6.93463564e-01 -1.03397870e+00
6.81583956e-02 -4.31120485e-01 -6.46488667e-02 -1.58100224e+00
3.44351232e-01 -1.64883807e-01 -4.68547493e-01 3.92574966e-01
-6.39024317e-01 5.85214496e-02 1.45092234e-01 -2.83726398e-02
-7.29379773e-01 6.34562850e-01 1.05609322e+00 -2.61870980e-01
-1.47483453e-01 1.94386631e-01 -5.78552246e-01 8.45272660e-01
6.22585893e-01 -1.18894704e-01 -3.43310148e-01 -2.59887159e-01
-9.99962240e-02 4.18783501e-02 7.47375965e-01 -1.12435412e+00
2.91556329e-01 -6.57141656e-02 1.41459918e+00 -1.03392315e+00
5.07656217e-01 -9.87096548e-01 -4.28351849e-01 7.35227242e-02
-3.35247487e-01 3.43396693e-01 1.29001990e-01 7.81230211e-01
-2.34005973e-02 5.01287520e-01 7.29835212e-01 -1.26793385e-01
-1.14851785e+00 3.90303880e-01 -1.86825290e-01 -2.45619603e-02
1.24445498e+00 -5.49419582e-01 -1.34488463e-01 -3.11865628e-01
-7.89008617e-01 4.58482236e-01 6.38252348e-02 4.96904135e-01
1.03771079e+00 -1.43638861e+00 -9.73493755e-01 7.80408382e-01
3.74940902e-01 -6.52565137e-02 6.22790217e-01 6.07635558e-01
4.15287428e-02 2.07899690e-01 -6.57753408e-01 -9.10733402e-01
-1.29784286e+00 8.62307370e-01 3.83825481e-01 7.01155141e-02
-6.88401639e-01 1.10812283e+00 8.76840413e-01 -9.90623683e-02
2.77546912e-01 -1.00023039e-01 -8.29803824e-01 1.71215653e-01
6.06475413e-01 1.39332324e-01 -3.19578081e-01 -1.12250423e+00
-7.31848180e-01 8.22767258e-01 -1.45351306e-01 1.30869240e-01
1.14185417e+00 -3.05792511e-01 2.44130701e-01 1.35556636e-02
8.79222810e-01 -1.52633354e-01 -2.00818467e+00 -6.77368119e-02
-6.80680037e-01 -5.67719638e-01 -1.74840942e-01 -9.70421493e-01
-1.27241051e+00 1.08096552e+00 9.04970884e-01 -2.44700357e-01
1.25026095e+00 2.88120538e-01 2.69158363e-01 5.49097024e-02
3.97174247e-02 -9.06813025e-01 -5.64284064e-02 2.50736147e-01
9.39052403e-01 -1.31634200e+00 2.04940587e-02 -3.66870403e-01
-1.06658220e+00 1.00598919e+00 1.41418636e+00 1.18210062e-01
6.03153825e-01 3.50525277e-03 7.53909200e-02 -2.91673750e-01
-5.22519350e-01 -1.68175578e-01 5.79408884e-01 9.04814899e-01
1.66659996e-01 8.63242224e-02 4.23335820e-01 9.50051188e-01
1.17205709e-01 -6.23342097e-01 -1.64864212e-02 4.04157758e-01
-5.22284448e-01 -4.66114014e-01 -4.26703662e-01 5.17206311e-01
-1.71532974e-01 -8.10239017e-02 -5.32038987e-01 6.93714797e-01
5.69465280e-01 7.35957205e-01 2.62135744e-01 -5.23797810e-01
5.76961756e-01 -4.48070057e-02 3.49877439e-02 -8.16897303e-02
-4.78041738e-01 4.95283157e-02 9.59545225e-02 -8.17598820e-01
-8.65005776e-02 -7.36900568e-01 -7.72204399e-01 1.35660738e-01
-1.28111273e-01 -1.87877044e-01 1.62958652e-01 6.82905555e-01
5.62286377e-01 6.35431111e-01 5.37283659e-01 -7.76769876e-01
-3.06543320e-01 -8.61384571e-01 -1.69707343e-01 4.16383654e-01
4.05667096e-01 -9.78087902e-01 -6.39929716e-03 -8.28190222e-02] | [14.403976440429688, 0.9907699823379517] |
e3ef052c-7c52-4a25-aad4-7b4a12767278 | quantum-machine-learning-and-quantum | 2004.12076 | null | https://arxiv.org/abs/2004.12076v2 | https://arxiv.org/pdf/2004.12076v2.pdf | Quantum machine learning and quantum biomimetics: A perspective | Quantum machine learning has emerged as an exciting and promising paradigm inside quantum technologies. It may permit, on the one hand, to carry out more efficient machine learning calculations by means of quantum devices, while, on the other hand, to employ machine learning techniques to better control quantum systems. Inside quantum machine learning, quantum reinforcement learning aims at developing "intelligent" quantum agents that may interact with the outer world and adapt to it, with the strategy of achieving some final goal. Another paradigm inside quantum machine learning is that of quantum autoencoders, which may allow one for employing fewer resources in a quantum device via a training process. Moreover, the field of quantum biomimetics aims at establishing analogies between biological and quantum systems, to look for previously inadvertent connections that may enable useful applications. Two recent examples are the concepts of quantum artificial life, as well as of quantum memristors. In this Perspective, we give an overview of these topics, describing the related research carried out by the scientific community. | ['Lucas Lamata'] | 2020-04-25 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 1.52094990e-01 2.04936668e-01 8.02130848e-02 1.35804638e-01
5.72334826e-02 -4.24642444e-01 8.38625014e-01 2.71793276e-01
-5.47748148e-01 9.81775641e-01 -3.78820926e-01 -2.95259029e-01
9.26050842e-02 -1.45624411e+00 -7.21090555e-01 -1.16012907e+00
-2.33956408e-02 4.88819033e-01 4.72072922e-02 -6.61196709e-01
5.98897398e-01 4.43483323e-01 -1.44457948e+00 4.84494455e-02
6.93789184e-01 6.05062187e-01 7.44239613e-02 6.07355118e-01
3.40189599e-02 5.74910939e-01 -2.52877384e-01 -4.06426460e-01
3.15047503e-02 -7.14375138e-01 -8.56192529e-01 -2.07121342e-01
-3.59879732e-01 1.12291828e-01 -6.16946399e-01 1.41606677e+00
2.30328277e-01 1.73574343e-01 8.16725016e-01 -4.70423669e-01
-8.04938078e-01 5.55411339e-01 3.03300798e-01 2.48146653e-01
3.07043552e-01 3.56297374e-01 9.69724894e-01 -5.91132455e-02
5.50150871e-01 9.22215283e-01 5.65751642e-02 9.64852691e-01
-1.49162424e+00 -2.43405595e-01 -1.00644541e+00 6.68840826e-01
-1.31432760e+00 -2.49360442e-01 5.52136064e-01 -2.79131085e-01
1.18681061e+00 -5.03947549e-02 7.69431591e-01 7.94744194e-01
8.24938715e-01 3.94272149e-01 1.30924594e+00 -7.93522775e-01
8.11298549e-01 1.94988906e-01 -1.31025881e-01 8.60914171e-01
1.40867054e-01 9.61166859e-01 -4.78088945e-01 -9.12644267e-02
6.39511585e-01 -1.67419299e-01 -1.16084583e-01 -1.15693673e-01
-1.28036535e+00 9.73613322e-01 6.77555799e-01 8.14434946e-01
-5.21865487e-01 2.85251677e-01 2.61834174e-01 3.45400661e-01
-1.47591189e-01 9.97757196e-01 -2.29998440e-01 -1.58862874e-01
-2.64455706e-01 -1.03888318e-01 8.81588578e-01 2.74460405e-01
1.38942170e+00 -2.17848513e-02 3.15886199e-01 6.26706779e-02
2.26127077e-02 5.71537614e-01 5.22735834e-01 -9.34542179e-01
-2.38206118e-01 4.10462916e-01 -2.01478917e-02 -3.02004099e-01
-5.88850081e-01 1.86755657e-01 -9.83454049e-01 4.04052287e-01
2.73582727e-01 -3.71884890e-02 -3.42435569e-01 1.45410502e+00
2.38546386e-01 1.54974326e-01 5.19985676e-01 6.70577765e-01
2.02640295e-01 7.77714908e-01 -1.58218428e-01 -4.49365288e-01
1.28261077e+00 -4.09210056e-01 -5.58783829e-01 3.95673782e-01
9.38364744e-01 -5.16696751e-01 5.24275601e-01 4.94917572e-01
-7.10995436e-01 -3.17834854e-01 -1.06003487e+00 4.09698784e-01
-8.27688873e-01 -5.22658765e-01 1.24106812e+00 9.89190638e-01
-9.64008510e-01 1.37607813e+00 -1.01104355e+00 -5.03230155e-01
1.16597027e-01 6.88467145e-01 -2.66873181e-01 3.58630627e-01
-1.49560010e+00 1.19106030e+00 8.62146378e-01 -1.68481305e-01
-7.77625978e-01 -6.82565197e-02 -3.87036026e-01 8.95107165e-03
1.23302028e-01 -1.07859588e+00 8.39563847e-01 -7.09244609e-01
-2.18597388e+00 8.56904447e-01 5.02639450e-02 -7.31569946e-01
-2.62903094e-01 3.56916606e-01 -4.75595027e-01 1.91477552e-01
-2.59117961e-01 5.87000489e-01 7.26418197e-01 -3.80201966e-01
-2.82129675e-01 -7.38495767e-01 2.01908350e-01 -1.03206411e-01
-4.48488146e-01 -4.32582974e-01 5.63637793e-01 1.26499727e-01
9.76771638e-02 -1.20144403e+00 -3.54137927e-01 -5.82770109e-01
-1.94384605e-01 -7.22718179e-01 3.76846701e-01 2.77986318e-01
8.40060592e-01 -1.78734970e+00 6.34446859e-01 1.56285837e-01
9.85402986e-02 2.65948147e-01 7.95687214e-02 8.71022880e-01
4.40958142e-02 7.25022182e-02 -7.76002929e-03 4.16920960e-01
-1.98633656e-01 4.16553229e-01 -5.81305660e-02 6.04357123e-01
2.94969708e-01 1.02662945e+00 -9.11770225e-01 -1.15128599e-01
3.26332867e-01 2.46006504e-01 -6.29783213e-01 -6.63775653e-02
-2.67421991e-01 9.03197050e-01 -8.67467344e-01 1.56324506e-01
1.29471645e-01 -4.62109387e-01 1.10567778e-01 4.68022861e-02
-4.20073271e-01 3.28552246e-01 -7.75603950e-01 1.51243281e+00
-4.21545327e-01 4.65120971e-01 -4.87496436e-01 -1.29693079e+00
7.12924778e-01 4.64057833e-01 3.65616977e-01 -1.16904080e+00
4.44904089e-01 5.35582662e-01 3.31494570e-01 -5.10661960e-01
3.62751633e-01 -5.20814419e-01 6.46218807e-02 6.33522511e-01
1.05974637e-01 -5.28658271e-01 3.22922170e-01 -2.00125486e-01
1.30839407e+00 -6.46817461e-02 3.47005844e-01 -1.65644109e-01
9.18046951e-01 -3.17868032e-02 4.31934521e-02 7.83872128e-01
-3.41521472e-01 -1.43552497e-01 7.33695403e-02 -8.58105123e-01
-1.29776704e+00 -1.24659777e+00 -4.01896060e-01 6.67387426e-01
3.58150840e-01 -3.98971960e-02 -9.32396829e-01 8.21714997e-02
-3.34123462e-01 4.52802151e-01 -5.73017120e-01 -5.65455556e-01
-4.30444241e-01 -8.57305050e-01 4.60375249e-01 -9.68468562e-02
4.20156032e-01 -1.47334468e+00 -7.66591847e-01 4.06353533e-01
2.13046148e-01 -8.23362589e-01 4.35935557e-01 5.23274183e-01
-1.33849108e+00 -7.91591108e-01 -3.74320269e-01 -3.89532655e-01
3.40710282e-01 -1.70461491e-01 8.94505382e-01 1.94815934e-01
-2.68255323e-01 6.47185445e-02 -5.51525176e-01 -1.82217270e-01
-1.09105277e+00 2.30465442e-01 6.69810116e-01 -1.90620482e-01
6.19910240e-01 -6.44541562e-01 -5.04743934e-01 -1.78537712e-01
-1.03261971e+00 -1.26740411e-01 7.16957450e-01 1.07202160e+00
4.16890234e-01 1.54911652e-01 4.50577676e-01 -5.52508771e-01
4.30249453e-01 -1.86254397e-01 -7.33016372e-01 2.56092191e-01
-5.98771632e-01 7.08911479e-01 1.00706625e+00 -1.17193080e-01
-6.20669603e-01 -9.79907662e-02 -2.56784260e-01 2.23143518e-01
-2.16875851e-01 2.73245811e-01 2.42057502e-01 -4.62780505e-01
1.09499419e+00 5.46892166e-01 5.34289004e-03 3.17007415e-02
5.27699113e-01 7.13145018e-01 1.40346155e-01 -4.33367014e-01
7.61102438e-01 7.06546903e-01 5.47649920e-01 -1.23776817e+00
-4.70706999e-01 -1.61648035e-01 -8.80704224e-01 -1.86036609e-03
1.12934256e+00 -3.11646909e-01 -1.28146291e+00 3.86726648e-01
-1.17137682e+00 -1.03734255e-01 -3.50865662e-01 6.36028886e-01
-1.03756320e+00 1.83021232e-01 -9.09943819e-01 -9.02630866e-01
-2.43213490e-01 -1.07305110e+00 7.17387259e-01 7.90888429e-01
3.39537561e-02 -1.14243257e+00 3.83920252e-01 2.37646401e-01
2.69383967e-01 -1.74463883e-01 1.03854322e+00 -5.02451777e-01
-9.79239643e-01 -2.61039346e-01 2.08529398e-01 3.03648263e-01
6.79526478e-03 -1.64886743e-01 -9.76371109e-01 -3.40908825e-01
4.41024918e-03 -4.57032531e-01 7.05465078e-01 1.03036344e-01
6.38535321e-01 1.14560619e-01 -2.64841944e-01 2.53769815e-01
1.30130196e+00 5.71714163e-01 9.81520176e-01 2.84643978e-01
2.76351392e-01 3.21901441e-01 1.27773270e-01 2.77555913e-01
-2.21477523e-02 5.93340278e-01 4.21067923e-01 4.06778932e-01
3.58369023e-01 1.44025823e-02 1.89641342e-01 1.16213369e+00
-5.41411221e-01 2.73233175e-01 -7.56816149e-01 -1.63214102e-01
-1.40699852e+00 -1.18859875e+00 -8.47373158e-02 2.40081191e+00
5.70655048e-01 8.56717899e-02 -9.57449526e-02 1.81378752e-01
7.06852794e-01 -2.37784177e-01 -4.53015924e-01 -8.32572460e-01
-2.52421349e-01 7.72678494e-01 3.60782295e-01 1.74601257e-01
-9.04606998e-01 1.16725326e+00 6.03727913e+00 7.01426566e-01
-1.34964788e+00 7.08666369e-02 2.12959155e-01 5.84067047e-01
-1.20463431e-01 4.70573694e-01 -4.88472313e-01 4.00099128e-01
1.51646388e+00 -1.74481466e-01 1.22338927e+00 3.75820518e-01
1.07398583e-03 -3.36808980e-01 -1.18314159e+00 9.49135005e-01
-5.21114051e-01 -1.80895543e+00 -7.14786723e-02 3.58214706e-01
9.38072920e-01 2.78877407e-01 -1.51171416e-01 2.44325072e-01
-3.05998176e-01 -8.21445048e-01 3.19218218e-01 8.39034498e-01
2.63525307e-01 -7.39489615e-01 6.73165858e-01 5.95944166e-01
-6.67385459e-01 -2.45463386e-01 -6.18671060e-01 -6.20087564e-01
-4.48840857e-02 3.11327577e-01 -5.24445474e-01 3.45143288e-01
4.16637480e-01 4.43074733e-01 -3.22006345e-01 9.05891299e-01
-8.35129991e-02 2.98668355e-01 -3.28612298e-01 -9.51255918e-01
1.89493418e-01 -8.27982664e-01 4.37822551e-01 4.73699152e-01
2.08820507e-01 2.07518101e-01 -3.67287159e-01 1.32812905e+00
9.74213257e-02 1.36093974e-01 -9.15857673e-01 -7.51471460e-01
2.28243977e-01 1.13978672e+00 -9.44039643e-01 -3.25514019e-01
-1.78560078e-01 1.05947399e+00 1.91357896e-01 4.34463657e-02
-5.35400212e-01 -5.04328549e-01 2.85812199e-01 -2.27221340e-01
-5.92234246e-02 -4.00355518e-01 -2.64892876e-01 -1.29766250e+00
-6.28489017e-01 -4.64316219e-01 -1.97957590e-01 -5.39456367e-01
-9.51942027e-01 4.67345893e-01 -7.28392720e-01 -1.04506850e+00
-4.22891617e-01 -9.07298028e-01 -4.98815686e-01 7.12699354e-01
-1.08725119e+00 -6.52747571e-01 4.23055381e-01 4.54815030e-01
-1.78892091e-01 -3.92051816e-01 1.35376608e+00 9.42118764e-02
-4.09146845e-01 7.45398998e-02 6.75283790e-01 -9.68238562e-02
1.26268610e-01 -1.17351389e+00 1.44931436e-01 4.74934846e-01
5.61285555e-01 6.40266597e-01 9.35357034e-01 -2.02458143e-01
-1.89897835e+00 -3.54994237e-01 7.48809576e-01 -3.23582470e-01
1.01589847e+00 -4.05158103e-02 -8.35267603e-01 -2.54417919e-02
1.43216416e-01 -1.91783711e-01 6.22739315e-01 1.95059940e-01
-1.10222116e-01 -9.78721455e-02 -8.96883309e-01 6.47293270e-01
5.35529733e-01 -1.08959925e+00 -5.13257623e-01 6.06154203e-01
3.62193674e-01 3.62520628e-02 -9.78918374e-01 1.61305636e-01
5.89906752e-01 -1.23047066e+00 6.55850887e-01 -7.50007510e-01
3.68195713e-01 -3.62023801e-01 -6.44218922e-02 -1.35125005e+00
-1.84484750e-01 -8.75073373e-01 -1.11416476e-02 4.49236125e-01
3.98241848e-01 -9.53406751e-01 9.32262301e-01 2.18646154e-01
-9.34660714e-03 -5.29778659e-01 -1.30803168e+00 -8.50946903e-01
5.41671693e-01 -1.73703238e-01 2.96473205e-01 6.83764994e-01
4.02072161e-01 3.65454853e-01 -1.19766064e-01 1.74312666e-02
5.54641247e-01 1.14898533e-01 4.26193058e-01 -1.11434424e+00
-6.28203094e-01 -6.02766156e-01 -8.70196164e-01 -7.35583246e-01
2.94276208e-01 -1.02216959e+00 -5.90510331e-02 -9.34532404e-01
7.47533143e-02 -2.50652432e-01 -5.12816906e-01 -2.60901432e-02
2.38285348e-01 4.39472318e-01 -3.67838331e-02 3.32012922e-01
-5.70610762e-01 5.54736257e-01 1.44012809e+00 -9.40529183e-02
-3.16334009e-01 8.97621587e-02 -2.09702164e-01 5.19732237e-01
8.41987967e-01 -5.23133636e-01 1.87250659e-01 -8.58753994e-02
8.59515250e-01 3.42627019e-01 2.39704221e-01 -1.36382020e+00
5.13291895e-01 -8.59136879e-02 7.57395923e-02 -3.66376750e-02
2.80665964e-01 -4.04885173e-01 -5.61957322e-02 1.13681102e+00
1.54366326e-02 -2.82921374e-01 -3.67375612e-01 5.03559589e-01
-1.58503905e-01 -5.16114831e-01 1.10845935e+00 -2.95855165e-01
-6.86318696e-01 1.47511080e-01 -8.62218380e-01 -3.85198742e-01
1.15594876e+00 -2.11318105e-01 -4.57470059e-01 7.66807869e-02
-8.38799834e-01 -2.02390179e-01 5.66598535e-01 1.84348762e-01
3.20847034e-01 -1.03254378e+00 -1.17089480e-01 1.43956229e-01
1.62254378e-01 -7.25037813e-01 3.67745608e-01 6.95626199e-01
-8.29788566e-01 7.98882842e-01 -5.84685028e-01 -5.01709461e-01
-5.12199283e-01 8.67163002e-01 5.78477144e-01 -4.71014157e-02
-6.13273792e-02 5.27359664e-01 -7.99889192e-02 -3.15660536e-01
-5.52762985e-01 1.83863491e-01 -1.00928314e-01 -3.80639702e-01
5.35034716e-01 2.87500441e-01 1.02417484e-01 -5.22163749e-01
-2.90859729e-01 5.15197456e-01 2.11368278e-01 -3.84628847e-02
1.15451157e+00 -3.16857509e-02 -6.30025268e-01 6.04032457e-01
1.08361745e+00 -5.10996163e-01 -6.08839571e-01 -7.04030227e-03
3.75893414e-01 2.98988670e-01 1.20607875e-01 -4.97642696e-01
-3.74368221e-01 1.21264088e+00 8.67017150e-01 9.11297202e-01
9.61779356e-01 1.21520802e-01 8.33549440e-01 1.18180394e+00
1.10925949e+00 -1.40081000e+00 1.01903945e-01 8.04808021e-01
9.09184515e-02 -1.40263951e+00 -3.24451566e-01 9.23053920e-02
4.07713018e-02 1.77647066e+00 2.31064670e-02 -4.04366493e-01
7.33715355e-01 -1.59727439e-01 -2.63030887e-01 -2.10903108e-01
-6.89020872e-01 -5.01191676e-01 -1.48793608e-02 4.67983514e-01
5.33437550e-01 5.86032867e-01 -5.30742347e-01 -2.42754027e-01
-9.61634517e-02 1.95055634e-01 8.34230542e-01 7.79182792e-01
-7.08864689e-01 -1.66178858e+00 -3.06311578e-01 3.51529211e-01
-2.80567199e-01 -7.20203072e-02 -1.37021765e-01 3.33923578e-01
2.60926902e-01 8.00809681e-01 -4.26300839e-02 -5.82708478e-01
-5.65633923e-03 5.32710366e-02 1.00913262e+00 -6.55905426e-01
-2.57454515e-01 -3.38928044e-01 -3.76669437e-01 -3.89612108e-01
-5.29867589e-01 -5.55577695e-01 -1.52269483e+00 -6.81689322e-01
-6.17703736e-01 4.29678023e-01 8.87717247e-01 1.42340422e+00
1.68391332e-01 3.01443309e-01 7.18578696e-01 -8.41227710e-01
-6.78498983e-01 -6.94134116e-01 -9.10933077e-01 2.38788888e-01
-7.32869096e-03 -6.06319189e-01 -3.13124396e-02 -1.20894298e-01] | [5.524615287780762, 4.925134658813477] |
cd62af12-1764-4022-8c47-acc7946a0694 | multimodal-representation-learning-via | 2103.04537 | null | https://arxiv.org/abs/2103.04537v5 | https://arxiv.org/pdf/2103.04537v5.pdf | Multimodal Representation Learning via Maximization of Local Mutual Information | We propose and demonstrate a representation learning approach by maximizing the mutual information between local features of images and text. The goal of this approach is to learn useful image representations by taking advantage of the rich information contained in the free text that describes the findings in the image. Our method trains image and text encoders by encouraging the resulting representations to exhibit high local mutual information. We make use of recent advances in mutual information estimation with neural network discriminators. We argue that the sum of local mutual information is typically a lower bound on the global mutual information. Our experimental results in the downstream image classification tasks demonstrate the advantages of using local features for image-text representation learning. | ['William M. Wells', 'Polina Golland', 'Steven Horng', 'Seth Berkowitz', 'Keegan Quigley', 'Miriam Cha', 'Daniel Moyer', 'Ruizhi Liao'] | 2021-03-08 | null | null | null | null | ['mutual-information-estimation'] | ['methodology'] | [ 5.90500653e-01 2.29579195e-01 -3.12257707e-01 -7.63625503e-01
-9.48590338e-01 -3.67537439e-01 1.05586398e+00 2.27994159e-01
-5.26198566e-01 5.45679629e-01 6.07365549e-01 -1.32611739e-02
-1.60389364e-01 -6.04024529e-01 -6.47183061e-01 -6.98316276e-01
-9.65477824e-02 1.76292434e-01 -2.76462823e-01 2.21969381e-01
4.60771650e-01 2.48101979e-01 -1.59071422e+00 7.17133284e-01
5.63512146e-01 9.18732941e-01 5.24489164e-01 9.23199534e-01
-2.55403757e-01 1.37143254e+00 -4.53601480e-01 -1.11576639e-01
-1.10652233e-02 -8.39428008e-01 -9.40287769e-01 2.59098321e-01
3.15226018e-01 -3.34851623e-01 -6.52478814e-01 9.62836683e-01
2.41228238e-01 2.96013564e-01 1.32783949e+00 -1.03792655e+00
-8.78085494e-01 5.97322166e-01 -5.41301668e-01 4.56363589e-01
9.04812515e-02 -1.04895905e-01 1.56571519e+00 -6.95085585e-01
6.76640868e-01 1.05204010e+00 3.17615181e-01 3.87543231e-01
-1.28709507e+00 -4.86680001e-01 -1.95761379e-02 -2.77340680e-01
-1.27302408e+00 -5.51212072e-01 3.96929502e-01 -6.43936872e-01
1.11635160e+00 2.24231891e-02 1.46719009e-01 8.34126890e-01
5.46520889e-01 9.92984056e-01 9.08143342e-01 -6.42651856e-01
-1.79813668e-01 4.24809843e-01 3.44516989e-03 9.26909149e-01
-3.83403786e-02 7.16886893e-02 -7.59041071e-01 -1.33302892e-02
9.19492364e-01 1.91577688e-01 -1.37420297e-01 -1.28467992e-01
-1.19262147e+00 1.08185709e+00 6.72692120e-01 7.01002896e-01
-1.30680084e-01 5.88197768e-01 3.08448821e-01 4.18013722e-01
7.30029821e-01 3.42141300e-01 -7.30588436e-02 -4.00114618e-02
-9.03531432e-01 -1.79727286e-01 6.35533810e-01 9.05419827e-01
9.15704250e-01 -2.17439145e-01 -2.11770087e-01 8.83351803e-01
5.00508964e-01 4.02702034e-01 7.16172993e-01 -8.33678305e-01
3.23009282e-01 4.26866531e-01 -2.82046139e-01 -1.06888032e+00
-7.94675872e-02 -2.09917396e-01 -6.48002207e-01 1.48772508e-01
1.36925548e-01 3.96950021e-02 -8.08953345e-01 1.78007245e+00
-4.54818040e-01 -2.13430792e-01 1.25477776e-01 3.94097805e-01
8.64035428e-01 5.14417231e-01 3.38321239e-01 -3.16915140e-02
9.73507166e-01 -6.79191113e-01 -8.38207722e-01 -2.40450814e-01
9.54221785e-01 -8.60251844e-01 5.80633163e-01 -3.43616515e-01
-9.42322314e-01 -3.00034702e-01 -8.69108319e-01 -1.89935699e-01
-4.32387859e-01 2.29837835e-01 6.32404327e-01 2.20000714e-01
-1.15352178e+00 7.76245356e-01 -5.28551936e-01 -4.66102302e-01
6.43190563e-01 3.90466034e-01 -7.07790077e-01 1.23457655e-01
-7.97594249e-01 8.38514805e-01 3.79940480e-01 -4.10326511e-01
-5.77805698e-01 -3.81178677e-01 -9.98048127e-01 1.32451534e-01
-3.43619555e-01 -6.87496841e-01 1.20277297e+00 -1.34310234e+00
-1.09669411e+00 1.28181863e+00 -3.40367258e-01 -3.03327084e-01
2.52731025e-01 1.54548317e-01 6.39140159e-02 4.40633208e-01
2.07816958e-01 9.52184260e-01 8.05636823e-01 -1.14221561e+00
-4.32961553e-01 -3.34039062e-01 -1.66686833e-01 3.99632424e-01
-2.96005011e-01 1.44297071e-02 -1.69323474e-01 -6.14241064e-01
9.18644518e-02 -6.77012622e-01 -1.46253988e-01 2.55148977e-01
-1.25136405e-01 -2.61783928e-01 4.96274918e-01 -4.99307960e-01
7.58911490e-01 -2.24884224e+00 -1.68245777e-01 2.83492297e-01
4.01398748e-01 -3.81514579e-01 -2.57228881e-01 6.37558103e-01
-1.89189658e-01 2.19682738e-01 -4.72930931e-02 -4.05458450e-01
-2.60198563e-01 3.67715418e-01 -3.51332873e-01 5.70608318e-01
4.28096622e-01 1.08904898e+00 -8.43171060e-01 -7.12927580e-01
1.49144337e-01 5.69427490e-01 -3.52964759e-01 2.57012039e-01
2.30557215e-03 3.52331340e-01 -5.51361203e-01 -4.06677043e-03
2.24964663e-01 -7.45803595e-01 1.74590886e-01 -1.96787000e-01
7.35240281e-02 5.43394983e-01 -4.99562770e-01 1.71661520e+00
-4.73554432e-01 1.26824677e+00 -2.25623250e-01 -1.15574455e+00
7.37411737e-01 2.96341240e-01 5.32216072e-01 -5.96292853e-01
2.83387661e-01 -1.50459260e-01 -1.46842971e-01 -4.85157341e-01
1.80059925e-01 -3.53017986e-01 3.69847238e-01 8.99657071e-01
4.60827976e-01 -8.61711055e-02 -1.14935018e-01 6.41343653e-01
9.26254928e-01 -4.01692837e-02 4.53713775e-01 -3.71738166e-01
7.51157776e-02 -4.32759851e-01 -2.90113330e-01 1.08922935e+00
-9.85812172e-02 7.19816387e-01 5.26858211e-01 -3.41582559e-02
-1.17844296e+00 -9.03351843e-01 -2.61766016e-01 1.17878556e+00
-1.64382726e-01 -6.01554930e-01 -3.53540927e-01 -7.23659217e-01
-1.27510756e-01 5.21240056e-01 -1.02504587e+00 -1.87056810e-01
-1.30717874e-01 -4.24208492e-01 4.17704374e-01 7.99001575e-01
2.59946316e-01 -9.80321646e-01 -3.21764559e-01 -1.93569243e-01
-1.68280602e-01 -1.02915645e+00 -4.67182428e-01 4.89966154e-01
-8.39067280e-01 -7.55613625e-01 -6.91951275e-01 -9.90644872e-01
7.89012194e-01 5.55348098e-01 1.09178889e+00 1.60392955e-01
-6.47409201e-01 8.40496182e-01 -1.20827682e-01 -1.46683082e-01
-5.65134943e-01 -2.09056944e-01 -4.29592431e-01 -1.30867511e-01
3.04802567e-01 -4.41394508e-01 -4.55080628e-01 1.01606689e-01
-9.94956315e-01 2.10576624e-01 7.65801907e-01 1.12052405e+00
4.97783065e-01 -1.22223586e-01 4.62257802e-01 -8.21554899e-01
6.43263042e-01 -5.58455586e-01 -1.90148592e-01 1.77227780e-01
-4.96972919e-01 5.87524951e-01 -1.00516923e-01 -3.68605435e-01
-1.07951677e+00 1.50005281e-01 6.92470074e-02 -1.58848211e-01
-8.66423771e-02 6.46850586e-01 3.58348489e-01 1.39636183e-02
6.39806867e-01 3.99851829e-01 2.88193285e-01 -9.85452905e-02
6.33694351e-01 7.94514120e-01 2.39367634e-01 -3.63600373e-01
1.73248842e-01 4.62760299e-01 -8.17675143e-03 -9.19950366e-01
-1.12197018e+00 -7.94439912e-01 -5.88704288e-01 -1.36813089e-01
1.05393994e+00 -1.07759440e+00 -7.11817145e-01 -1.08039565e-01
-1.11987948e+00 -1.51447982e-01 -4.78325695e-01 6.68169856e-01
-9.62087631e-01 3.33929777e-01 -6.71680510e-01 -8.73876274e-01
-1.52546495e-01 -9.43947017e-01 1.27731133e+00 -4.89595719e-02
-2.25285351e-01 -1.37697589e+00 1.29102215e-01 8.68402869e-02
2.97032505e-01 1.20983571e-02 8.12738061e-01 -9.33764040e-01
-4.40774828e-01 -3.45639110e-01 -7.03202784e-01 3.23410273e-01
3.01982611e-01 -1.32594377e-01 -1.14171004e+00 -2.90360004e-02
-6.49465844e-02 -5.57457030e-01 1.44165444e+00 5.87242544e-01
1.04999030e+00 -4.30054307e-01 -3.82752568e-01 2.07585126e-01
1.40503633e+00 -2.35866815e-01 5.91539979e-01 3.05327438e-02
4.61369336e-01 8.02514017e-01 3.48282009e-02 4.78274018e-01
9.73975062e-02 3.59017402e-01 2.00968057e-01 -2.44666524e-02
-2.11164862e-01 -3.15252721e-01 3.06506008e-01 7.54651666e-01
1.53463066e-01 -1.87869549e-01 -6.07739210e-01 6.09852552e-01
-1.96016562e+00 -1.26121438e+00 7.54569918e-02 1.92181885e+00
9.38475609e-01 -2.44025253e-02 -2.82701373e-01 -4.59806204e-01
6.21179104e-01 1.88686922e-01 -2.18714252e-01 -2.42018729e-01
-1.78949684e-01 6.88971356e-02 4.89087224e-01 5.15936494e-01
-1.20089746e+00 7.36704350e-01 8.16270733e+00 6.30998790e-01
-5.80478072e-01 -1.05879664e-01 7.57978797e-01 -2.54764669e-02
-2.41418079e-01 -1.27028301e-01 -5.15098989e-01 9.34192985e-02
1.17318726e+00 -2.98337877e-01 1.85731754e-01 7.89177477e-01
-6.94734529e-02 -1.16984472e-01 -1.37446260e+00 1.10809672e+00
2.49412537e-01 -1.53882933e+00 3.11887175e-01 1.72639459e-01
9.78855789e-01 3.88172686e-01 2.94822693e-01 -2.02057853e-01
6.85209692e-01 -1.33523405e+00 1.60009950e-01 7.44242847e-01
7.35520720e-01 -6.29570007e-01 5.72871625e-01 2.77799308e-01
-9.79155481e-01 3.63154747e-02 -4.71380711e-01 4.94080223e-02
-4.21814829e-01 6.10020816e-01 -1.00624883e+00 1.95242658e-01
3.87113810e-01 1.15793192e+00 -3.88478756e-01 7.83003569e-01
-2.04969883e-01 3.08884382e-01 -2.14685813e-01 -1.02156907e-01
3.94118965e-01 -1.17525443e-01 1.96383089e-01 1.63752818e+00
7.28263631e-02 -3.68889212e-03 1.46310106e-01 1.21674919e+00
-5.05761623e-01 3.25621188e-01 -1.22004962e+00 -5.65495014e-01
-6.23990037e-02 1.22330534e+00 -6.63643956e-01 -5.63890278e-01
-4.80303973e-01 1.04445684e+00 4.97367501e-01 2.07831085e-01
-1.28165260e-01 -3.39655757e-01 5.26994348e-01 -2.59221435e-01
3.03456575e-01 -2.24410817e-01 -2.79374301e-01 -1.18448663e+00
-1.23133205e-01 -4.20109719e-01 3.40698242e-01 -9.21657145e-01
-1.57012606e+00 4.17430520e-01 -4.04687896e-02 -9.73430097e-01
-5.08396208e-01 -7.29162455e-01 -4.90679473e-01 1.01831996e+00
-1.23267937e+00 -1.22639966e+00 3.17269261e-03 6.33212566e-01
5.94837248e-01 -1.38847262e-01 9.56902623e-01 -2.48424038e-01
-8.53794888e-02 5.00632167e-01 5.36369264e-01 3.57877493e-01
6.96715534e-01 -1.28749549e+00 -5.77265695e-02 2.20213979e-01
5.38162827e-01 8.49451840e-01 5.31322181e-01 -5.83872974e-01
-9.31130707e-01 -7.77630568e-01 9.75113809e-01 -5.24546564e-01
9.80280161e-01 -2.70496607e-01 -7.37458289e-01 1.17179215e+00
5.82817495e-01 -9.26080160e-03 1.12267005e+00 1.66335106e-01
-7.33311594e-01 4.89081204e-01 -9.28140521e-01 5.38160741e-01
7.10647941e-01 -1.08346033e+00 -5.37065327e-01 7.54633367e-01
4.66077864e-01 1.94840237e-01 -8.15891683e-01 -1.67806730e-01
7.32472003e-01 -6.58764184e-01 6.60026848e-01 -7.31969237e-01
8.46693039e-01 2.93693036e-01 -4.80650604e-01 -1.10286438e+00
-2.35500515e-01 -3.87695312e-01 3.40676278e-01 1.02617073e+00
5.39871335e-01 -4.58528221e-01 5.17416298e-01 5.81698298e-01
3.34433436e-01 -3.96316260e-01 -7.99756765e-01 -4.72919255e-01
2.59742826e-01 -2.79878348e-01 -1.14942752e-01 9.66691434e-01
5.74312150e-01 5.81997931e-01 -3.12201023e-01 -3.69066805e-01
5.60607910e-01 6.95819706e-02 3.96819413e-01 -1.15018141e+00
-3.88973445e-01 -4.89492595e-01 -7.49832749e-01 -1.29432213e+00
5.47893882e-01 -1.35914874e+00 2.73985922e-01 -1.44527102e+00
9.83783007e-01 -5.00784181e-02 -5.13080359e-01 4.44047660e-01
2.58944798e-02 3.28046978e-01 1.46417201e-01 5.61611891e-01
-7.30011761e-01 4.60486442e-01 1.20884812e+00 -2.55010426e-01
1.19780332e-01 -2.54138827e-01 -7.40046203e-01 6.60351813e-01
6.99104905e-01 -5.71128607e-01 -3.29182535e-01 -3.70335460e-01
3.07519525e-01 -8.26960579e-02 3.30254406e-01 -6.05209708e-01
2.81315804e-01 -1.53759815e-04 6.76156282e-01 -5.11137009e-01
3.29886496e-01 -6.70896590e-01 -4.16702271e-01 1.80935830e-01
-1.41145432e+00 -2.53339857e-01 3.01180389e-02 7.99549222e-01
-2.86246628e-01 -4.42120880e-01 9.00680661e-01 -4.36311543e-01
-4.17919636e-01 7.24068582e-02 -6.49245858e-01 -2.43789956e-01
6.38729930e-01 8.64163190e-02 -2.51464427e-01 -8.90047133e-01
-9.01856184e-01 -1.27488434e-01 2.36324310e-01 2.39281639e-01
7.48191059e-01 -1.30128086e+00 -7.66441345e-01 2.67133057e-01
3.74533802e-01 -8.13755393e-01 -9.96659249e-02 8.31522763e-01
-8.58905166e-03 6.69253707e-01 -1.66827396e-01 -6.08589232e-01
-1.39978480e+00 3.30873787e-01 3.89533877e-01 -4.04839754e-01
-8.21589291e-01 7.48894274e-01 8.34869385e-01 -1.76234171e-01
-6.09961003e-02 3.74691896e-02 -1.76506504e-01 -1.46538734e-01
8.49639297e-01 -7.77718648e-02 -2.13680461e-01 -9.15377855e-01
-1.03120193e-01 3.94767404e-01 -5.18990576e-01 -4.63522494e-01
1.27434063e+00 -3.84842724e-01 -2.44854629e-01 5.83114088e-01
1.88453722e+00 -2.96990007e-01 -1.02518189e+00 -4.44146693e-01
-1.58030428e-02 -5.27608573e-01 4.36807662e-01 -5.76200545e-01
-8.86102676e-01 1.00055325e+00 6.81173742e-01 1.83757812e-01
7.23496258e-01 5.60595810e-01 1.90953299e-01 6.62459314e-01
-1.31484807e-01 -9.79956627e-01 4.69869614e-01 4.39835697e-01
8.46192479e-01 -1.52393389e+00 9.08295512e-02 -2.28948057e-01
-7.85063744e-01 1.29662144e+00 4.69735935e-02 -2.99172908e-01
8.25240850e-01 2.48411715e-01 -1.14286378e-01 -4.48109895e-01
-9.48284209e-01 -4.80297893e-01 5.37633657e-01 5.29567897e-01
1.01686430e+00 -1.41159415e-01 -2.96474397e-02 -3.69231508e-04
4.33753878e-02 -1.61281019e-01 2.41692215e-01 9.69914019e-01
-7.46004581e-01 -9.03652906e-01 9.33234543e-02 7.30958581e-01
-7.11664021e-01 -3.22755694e-01 -5.76129794e-01 6.10417902e-01
-4.49520320e-01 9.78880465e-01 3.02586257e-01 -9.89829004e-02
-4.08938527e-01 1.54770806e-01 7.20752478e-01 -6.90369248e-01
-2.85121322e-01 1.48231164e-01 -4.42379415e-02 -4.32135910e-01
-6.73158288e-01 -5.47968686e-01 -1.30065703e+00 -1.17165586e-02
-5.83404839e-01 -1.19322456e-01 9.35184479e-01 1.10773230e+00
2.91834444e-01 4.13488865e-01 6.40567839e-01 -8.28314006e-01
-4.32653099e-01 -9.22275722e-01 -7.91806400e-01 5.74815392e-01
4.21380341e-01 -4.34617132e-01 -4.03227121e-01 4.63040829e-01] | [9.540130615234375, 2.67230486869812] |
2121940a-100e-476c-ab99-0ba00695b175 | saral-a-low-resource-cross-lingual-domain | null | null | https://aclanthology.org/P19-3004 | https://aclanthology.org/P19-3004.pdf | SARAL: A Low-Resource Cross-Lingual Domain-Focused Information Retrieval System for Effective Rapid Document Triage | With the increasing democratization of electronic media, vast information resources are available in less-frequently-taught languages such as Swahili or Somali. That information, which may be crucially important and not available elsewhere, can be difficult for monolingual English speakers to effectively access. In this paper we present an end-to-end cross-lingual information retrieval (CLIR) and summarization system for low-resource languages that 1) enables English speakers to search foreign language repositories of text and audio using English queries, 2) summarizes the retrieved documents in English with respect to a particular information need, and 3) provides complete transcriptions and translations as needed. The SARAL system achieved the top end-to-end performance in the most recent IARPA MATERIAL CLIR+summarization evaluations. Our demonstration system provides end-to-end open query retrieval and summarization capability, and presents the original source text or audio, speech transcription, and machine translation, for two low resource languages. | ['Banriskhem Kayang Khonglah', 'Chester Palen-Michel', 'Marjorie Freedman', 'Thamme Gowda', 'Scott Miller', 'Constantine Lignos', 'Jayadev Billa', 'Srikanth Madikeri', 'Elizabeth Boschee', 'Michael Pust', 'Jonathan May', 'Joel Barry'] | 2019-07-01 | null | null | null | acl-2019-7 | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [-2.17111576e-02 2.14615837e-02 -3.20695937e-01 9.88590866e-02
-2.33261800e+00 -8.98766756e-01 5.45335233e-01 6.81752980e-01
-7.22561240e-01 8.69477153e-01 9.43582654e-01 -2.73984253e-01
-3.03249627e-01 -1.81852892e-01 -3.87521446e-01 -3.31151001e-02
1.71691179e-01 8.21307957e-01 -3.05880271e-02 -6.25615656e-01
2.89886236e-01 1.95216700e-01 -1.36298382e+00 4.34166431e-01
1.14206398e+00 6.76996529e-01 3.99814427e-01 8.00171077e-01
-1.47370279e-01 5.53409934e-01 -7.53620028e-01 -2.57317960e-01
-2.27431536e-01 -3.46923381e-01 -1.25086784e+00 -3.01996171e-01
5.78631938e-01 -2.37925410e-01 -3.72845918e-01 7.28489637e-01
1.11520708e+00 4.40357812e-03 5.40218115e-01 -4.18587387e-01
-6.62272692e-01 1.00838542e+00 -3.90568942e-01 3.60475838e-01
1.07297957e+00 -3.94647628e-01 9.72057104e-01 -9.33699131e-01
1.02209711e+00 1.23106480e+00 3.60468984e-01 1.97026491e-01
-9.66465950e-01 -6.86963677e-01 -3.16543370e-01 4.34056260e-02
-1.81847870e+00 -1.08841658e+00 4.56725746e-01 2.14667507e-02
1.32438505e+00 5.39932311e-01 3.42589110e-01 9.38528538e-01
8.27428401e-02 8.41681778e-01 3.36696357e-01 -7.45465279e-01
-1.48895711e-01 2.79751718e-01 -1.08342394e-01 1.76864922e-01
9.19456482e-02 -7.21127927e-01 -9.70514834e-01 -1.47713214e-01
-1.19793750e-01 -5.70123792e-01 -2.26189494e-01 5.55823863e-01
-1.49695122e+00 5.56345940e-01 -1.21311888e-01 4.71193373e-01
-5.09338439e-01 -4.32444006e-01 9.02007699e-01 7.29194582e-01
9.10893977e-01 5.13093948e-01 -3.70312989e-01 -3.23470324e-01
-1.36184025e+00 1.91323802e-01 9.29516017e-01 1.25413084e+00
5.65372944e-01 -1.42450377e-01 -1.80558532e-01 1.54667771e+00
2.47344345e-01 1.18220127e+00 7.24678993e-01 -8.28293502e-01
9.21015501e-01 7.62258694e-02 -1.56821296e-01 -7.17580259e-01
-1.54343992e-01 -4.03807461e-01 -7.13150501e-01 -9.65711832e-01
-2.65195757e-01 -2.34721377e-01 -3.41772527e-01 1.36650145e+00
1.81557178e-01 -8.80527675e-01 6.62703633e-01 4.65290517e-01
1.44264257e+00 1.27205014e+00 -2.29425982e-01 -8.80202889e-01
1.50840926e+00 -7.45253444e-01 -1.12448454e+00 -2.88385093e-01
5.39470911e-01 -1.64747095e+00 9.51490045e-01 1.66330144e-01
-1.47815180e+00 -2.28385925e-01 -8.47525179e-01 -6.14861071e-01
-4.41132814e-01 2.51534164e-01 1.04465947e-01 4.07686591e-01
-1.34917068e+00 -5.71158249e-03 -6.45826340e-01 -8.26665163e-01
-1.59268543e-01 8.18342045e-02 -6.23731077e-01 -2.12772027e-01
-1.27648687e+00 7.14399755e-01 3.74717265e-01 -2.21590757e-01
-4.47251946e-01 -6.68494403e-01 -8.50789607e-01 -1.39970690e-01
2.60404080e-01 -2.94765353e-01 1.60882950e+00 -3.43541712e-01
-1.38611329e+00 9.52025890e-01 -3.37581366e-01 -2.60029584e-01
3.35106999e-01 -3.69037509e-01 -5.58941066e-01 6.88806117e-01
6.36897624e-01 6.72807813e-01 3.62347901e-01 -7.04747975e-01
-7.88804173e-01 -4.03555870e-01 -4.79756922e-01 7.84663141e-01
-4.24705505e-01 8.73637557e-01 -9.12366867e-01 -8.59318018e-01
1.71005800e-01 -5.77338040e-01 2.34686449e-01 -6.29403889e-01
-6.66502297e-01 -3.73288155e-01 7.81183660e-01 -1.43079531e+00
1.67892182e+00 -2.08836412e+00 -2.79395431e-02 -2.76195128e-02
-4.10446852e-01 7.91477710e-02 -5.86645864e-02 1.12166274e+00
3.46064121e-01 3.47716391e-01 2.00845242e-01 -1.81280464e-01
-7.30378833e-03 -1.14213020e-01 -5.89595914e-01 2.85579264e-01
-3.14027846e-01 5.79653502e-01 -8.69563282e-01 -9.96038973e-01
-6.72899783e-02 5.35412192e-01 -4.34907854e-01 -9.26424637e-02
1.62529349e-01 4.51672554e-01 -4.40348715e-01 8.79068196e-01
6.23523956e-03 3.26667219e-01 -1.39200494e-01 2.27144714e-02
-5.26879013e-01 8.74226034e-01 -9.50698555e-01 2.08051991e+00
-7.00790942e-01 9.88314211e-01 2.69764930e-01 -5.18486798e-01
6.89976156e-01 8.23307097e-01 5.23282945e-01 -9.87869084e-01
-1.65276557e-01 6.82334661e-01 -6.25247300e-01 -5.15408456e-01
1.05440950e+00 4.14926827e-01 -6.20898604e-01 5.76598763e-01
2.01617017e-01 -6.42943144e-01 6.58862591e-01 4.98434246e-01
7.59846628e-01 -2.66064852e-01 4.80062515e-01 -3.42191905e-01
6.09453380e-01 4.18888360e-01 -2.11697053e-02 6.96006477e-01
1.85952812e-01 5.93308508e-01 -1.58121631e-01 2.01135725e-01
-1.05965722e+00 -8.38757277e-01 -2.56637692e-01 1.57689691e+00
-2.25312859e-01 -7.14173436e-01 -6.12706065e-01 1.16638936e-01
-5.21323502e-01 8.93656194e-01 2.41650000e-01 3.71053740e-02
-6.13230765e-01 8.50846320e-02 9.26444173e-01 -2.95751870e-01
3.96466106e-01 -1.25050223e+00 -1.94247782e-01 3.87377888e-01
-1.08753574e+00 -1.04680514e+00 -1.03341949e+00 1.50238900e-02
-6.16512060e-01 -3.35747600e-01 -1.10020709e+00 -1.12267649e+00
4.69925366e-02 3.69112492e-01 9.78195667e-01 -5.38134515e-01
-2.23732188e-01 7.42083609e-01 -5.30335844e-01 -6.03908777e-01
-6.20718837e-01 6.37237072e-01 1.68146789e-01 -4.66134995e-01
2.70798206e-01 -3.03785294e-01 -1.39723346e-01 -1.71309784e-01
-9.12634492e-01 -1.51931494e-01 4.57870901e-01 4.38885540e-01
8.58564913e-01 -1.29876122e-01 9.35815513e-01 -4.88774568e-01
1.26689088e+00 -4.01662380e-01 -2.79523998e-01 6.09361947e-01
-3.87641788e-01 3.20821404e-02 3.90996873e-01 -1.32113472e-01
-1.04659808e+00 -3.04742306e-01 -3.93890977e-01 1.69131592e-01
7.19268546e-02 1.24587250e+00 4.45055775e-02 4.84544814e-01
7.01925576e-01 3.65631074e-01 -3.51308435e-01 -7.10670650e-01
3.19113523e-01 1.60081613e+00 7.14925051e-01 -4.39396620e-01
4.57282126e-01 -9.31095332e-03 -6.95897877e-01 -1.50289643e+00
-6.68278217e-01 -1.05299473e+00 -4.96646017e-01 -2.52385408e-01
6.60044551e-01 -1.27632070e+00 -2.14262053e-01 1.35171354e-01
-1.23631406e+00 3.13090324e-01 -3.85587811e-01 7.38085628e-01
-4.57378805e-01 3.02233338e-01 -5.37686288e-01 -7.40229607e-01
-1.04848838e+00 -9.74183619e-01 1.53149354e+00 6.23046979e-02
-6.07479453e-01 -6.51002347e-01 3.26369375e-01 7.04056382e-01
3.90087932e-01 -4.61325705e-01 7.75464952e-01 -9.36856568e-01
1.17713772e-02 -5.40562630e-01 5.10049015e-02 1.15076706e-01
1.12974830e-01 -7.47261047e-02 -6.37121379e-01 -4.51269239e-01
-4.44668472e-01 -7.88282871e-01 5.30580103e-01 3.92254263e-01
4.16877031e-01 -7.47598767e-01 -7.76122436e-02 1.12413332e-01
1.01401281e+00 1.64596513e-01 4.83397156e-01 2.93390781e-01
1.31216288e-01 7.38700807e-01 6.20441079e-01 4.17562097e-01
5.24896085e-01 8.63229275e-01 -5.68170190e-01 1.65671900e-01
-3.16602856e-01 -5.16513646e-01 6.85085654e-01 1.90949523e+00
3.46226752e-01 -1.56397834e-01 -8.86689544e-01 9.04706001e-01
-1.57737172e+00 -1.02116585e+00 4.30032790e-01 2.12354875e+00
1.33523786e+00 -3.80327523e-01 1.24133462e-02 -1.51883423e-01
5.90188146e-01 5.43155372e-02 -3.47161710e-01 -5.32790184e-01
-2.77346849e-01 2.33949050e-01 5.20015180e-01 7.41862297e-01
-7.32497394e-01 1.08147633e+00 6.45841026e+00 1.36296022e+00
-1.09931231e+00 2.95602590e-01 2.50155210e-01 -2.76383489e-01
-4.26936537e-01 -3.69270027e-01 -8.98558497e-01 5.61786406e-02
1.38827026e+00 -1.05252016e+00 5.31230271e-01 6.15265965e-01
5.28379381e-01 -1.78486153e-01 -8.02570403e-01 1.20513904e+00
3.25694174e-01 -1.27740383e+00 3.86411309e-01 -2.72101432e-01
5.45512736e-01 4.73952353e-01 1.27303591e-02 6.05839431e-01
-3.62793654e-02 -7.51997113e-01 9.34411705e-01 1.85792565e-01
1.30601037e+00 -9.71562684e-01 5.97806334e-01 5.90670466e-01
-1.19095433e+00 4.24529701e-01 -3.00834417e-01 5.28620362e-01
2.61801511e-01 5.24075665e-02 -8.08252931e-01 9.19799507e-01
8.16019177e-01 4.78485584e-01 -6.07206762e-01 9.76147890e-01
-5.81606403e-02 5.67893207e-01 -6.00721836e-01 -1.88741103e-01
3.11904520e-01 -8.48387368e-03 9.07605946e-01 1.67303312e+00
5.73318005e-01 -6.18747808e-03 2.82359600e-01 3.38071734e-02
-3.23073030e-01 1.09434795e+00 -8.15892279e-01 -3.67260128e-01
7.57624686e-01 8.94410074e-01 -5.78033209e-01 -3.59295785e-01
-2.91824162e-01 8.81000638e-01 4.19174097e-02 4.81273413e-01
-6.86085746e-02 -1.00664842e+00 1.33220673e-01 -6.97693527e-02
-2.41477519e-01 -1.54709369e-01 2.23510742e-01 -1.27784681e+00
1.52740493e-01 -1.38567770e+00 6.04190350e-01 -8.62926245e-01
-8.97163630e-01 7.57725656e-01 2.40525424e-01 -1.26644218e+00
-1.05466866e+00 1.56421989e-01 1.30407438e-01 9.56195354e-01
-1.22386336e+00 -1.09533525e+00 5.08128762e-01 7.15905070e-01
1.17851496e+00 -5.18940747e-01 1.03310251e+00 7.67245352e-01
-2.57857889e-01 6.85149848e-01 5.62115669e-01 -1.34400696e-01
1.05346191e+00 -6.68453157e-01 -2.47588620e-01 7.99872696e-01
3.36720377e-01 6.88324928e-01 7.73037314e-01 -5.99231362e-01
-1.61057317e+00 -9.22924697e-01 1.82514966e+00 -1.60054684e-01
7.62422323e-01 -1.89327091e-01 -5.60113966e-01 6.09517038e-01
8.15164566e-01 -8.29607844e-01 8.50211799e-01 2.28225276e-01
-1.74123198e-01 -3.62582982e-01 -9.39037383e-01 7.10175931e-01
6.64982557e-01 -1.11206019e+00 -8.25365424e-01 7.89777100e-01
1.17510521e+00 -4.39515054e-01 -7.80291200e-01 1.19227454e-01
5.16613066e-01 -4.01331820e-02 9.78259265e-01 -3.70319188e-01
1.03524357e-01 -4.47029844e-02 -4.43046719e-01 -1.40164566e+00
1.81025743e-01 -1.11524975e+00 4.66431558e-01 1.52080333e+00
9.08314645e-01 -2.94296503e-01 -1.06084254e-02 2.08710551e-01
-3.97271872e-01 -9.79061127e-02 -1.26491141e+00 -5.82068503e-01
3.46812904e-02 -6.21733844e-01 1.31392583e-01 7.46303141e-01
5.44918299e-01 8.20104003e-01 -3.78770769e-01 7.24048447e-03
2.87662357e-01 -7.37379119e-02 5.85867882e-01 -9.24050510e-01
1.25409171e-01 -4.81454521e-01 4.46766093e-02 -8.64640355e-01
1.40527487e-01 -1.16855013e+00 4.40537967e-02 -1.90426004e+00
2.63046145e-01 1.16680332e-01 2.09714338e-01 3.72625649e-01
3.18028659e-01 2.99866050e-01 -4.38418798e-03 6.10439956e-01
-1.00146616e+00 5.03904045e-01 9.14064467e-01 -4.93838131e-01
-6.44064724e-01 -3.58457342e-02 -9.97242153e-01 2.89877534e-01
4.97953475e-01 -4.16587293e-01 -3.74981374e-01 -7.10047483e-01
4.25143600e-01 4.16202128e-01 -5.51596999e-01 -8.15172732e-01
6.36021972e-01 3.18346396e-02 -6.86616153e-02 -1.31359625e+00
9.95725319e-02 -4.78879720e-01 8.52014348e-02 1.28539324e-01
-6.88973606e-01 3.62476319e-01 5.01706421e-01 -5.29645048e-02
-8.10229897e-01 -2.46430635e-01 4.37123150e-01 -1.77493878e-02
-3.13432813e-01 1.96749084e-02 -9.25225258e-01 5.40740609e-01
3.70359659e-01 1.07664108e-01 -1.79752201e-01 -1.07209444e+00
-4.61700767e-01 4.05253887e-01 2.24531703e-02 5.13938546e-01
5.13681591e-01 -1.22685742e+00 -1.27415812e+00 -3.64075303e-02
4.19847399e-01 -2.44160533e-01 1.00472592e-01 6.73107505e-01
-7.30191648e-01 1.17403269e+00 6.11233711e-03 -3.70250165e-01
-1.34613788e+00 1.06268421e-01 -2.50076532e-01 -3.86796474e-01
-5.26988924e-01 6.49734676e-01 -3.29559505e-01 -6.00030482e-01
5.04822969e-01 6.59651533e-02 -4.14980739e-01 6.25446558e-01
9.37685966e-01 3.10034424e-01 4.74821001e-01 -1.13409650e+00
-2.27311850e-01 3.90603542e-01 -3.42609018e-01 -1.03903151e+00
1.22089148e+00 -8.99134994e-01 -2.61876136e-01 6.97414339e-01
1.36296666e+00 6.35531306e-01 3.23544033e-02 -5.47749221e-01
9.38017368e-02 5.18708071e-03 3.44284147e-01 -8.09797287e-01
-4.04416531e-01 6.45835638e-01 2.19203264e-01 1.13738142e-01
1.27185333e+00 3.13133627e-01 8.98094237e-01 1.05830491e+00
3.88873965e-01 -1.66404116e+00 -3.03969592e-01 9.63807940e-01
1.26217520e+00 -9.89676476e-01 1.39524385e-01 1.56644374e-01
-5.56882262e-01 9.76119757e-01 -1.69394061e-01 7.06816852e-01
4.76618022e-01 -9.50740948e-02 3.53919446e-01 -1.11548804e-01
-7.24386811e-01 -1.28341407e-01 5.96583426e-01 4.19995964e-01
7.86226332e-01 -4.01919186e-02 -4.93266791e-01 3.50215793e-01
-7.31187701e-01 -4.14471209e-01 3.18602681e-01 8.14838171e-01
-5.45452714e-01 -8.71468663e-01 -5.51985621e-01 2.45616391e-01
-1.08864164e+00 -5.20988524e-01 -4.98557806e-01 5.69386184e-01
-6.68084860e-01 1.34368503e+00 -7.16150627e-02 8.33637118e-02
4.65734601e-01 1.35705143e-01 9.88067240e-02 -6.93045437e-01
-6.16041899e-01 7.55314291e-01 5.13993680e-01 -3.05695891e-01
-2.50069708e-01 -9.45387006e-01 -9.95999753e-01 -3.13110441e-01
-1.50762573e-01 6.50083780e-01 1.00990784e+00 7.73438275e-01
3.92566353e-01 2.46428087e-01 5.70189118e-01 -8.45082939e-01
-2.55215138e-01 -1.31717098e+00 -4.19231266e-01 -9.39645320e-02
3.61273050e-01 1.77361518e-01 -1.32063210e-01 1.93731457e-01] | [14.345773696899414, 7.288815975189209] |
ffb6ab9f-a9ea-4819-b8c7-d7074f026423 | state-of-the-art-in-human-scanpath-prediction | 2102.12239 | null | https://arxiv.org/abs/2102.12239v2 | https://arxiv.org/pdf/2102.12239v2.pdf | State-of-the-Art in Human Scanpath Prediction | The last years have seen a surge in models predicting the scanpaths of fixations made by humans when viewing images. However, the field is lacking a principled comparison of those models with respect to their predictive power. In the past, models have usually been evaluated based on comparing human scanpaths to scanpaths generated from the model. Here, instead we evaluate models based on how well they predict each fixation in a scanpath given the previous scanpath history. This makes model evaluation closely aligned with the biological processes thought to underly scanpath generation and allows to apply established saliency metrics like AUC and NSS in an intuitive and interpretable way. We evaluate many existing models of scanpath prediction on the datasets MIT1003, MIT300, CAT2000 train and CAT200 test, for the first time giving a detailed picture of the current state of the art of human scanpath prediction. We also show that the discussed method of model benchmarking allows for more detailed analyses leading to interesting insights about where and when models fail to predict human behaviour. The MIT/Tuebingen Saliency Benchmark will implement the evaluation of scanpath models as detailed here, allowing researchers to score their models on the established benchmark datasets MIT300 and CAT2000. | ['Matthias Bethge', 'Matthias Kümmerer'] | 2021-02-24 | null | null | null | null | ['scanpath-prediction'] | ['computer-vision'] | [ 3.96728516e-01 1.24558896e-01 -2.01449826e-01 -4.53045756e-01
3.62892523e-02 -4.25480753e-01 7.06510484e-01 3.79778683e-01
-5.52900016e-01 6.49626672e-01 6.92271441e-02 -3.77809376e-01
-3.63119006e-01 -4.95958745e-01 -6.61382973e-01 -5.49759746e-01
-6.34856448e-02 6.43777430e-01 7.52913177e-01 -2.30502620e-01
9.15155590e-01 2.63805985e-01 -2.07238531e+00 2.09538236e-01
6.91355824e-01 5.66767633e-01 5.17408490e-01 7.85681844e-01
3.01297843e-01 4.70684379e-01 -5.48487544e-01 -5.58966458e-01
5.12059815e-02 -8.30236375e-01 -9.41936135e-01 -3.95931393e-01
5.34742355e-01 2.40610436e-01 4.62059267e-02 8.27354550e-01
4.16820437e-01 -2.33718082e-02 8.24136257e-01 -1.18833590e+00
-6.07096493e-01 5.66543221e-01 -6.54044747e-01 9.34821427e-01
5.26454449e-01 3.55461270e-01 1.24239206e+00 -4.62646484e-01
9.56879020e-01 1.27598989e+00 4.32288140e-01 4.02371585e-01
-1.36781847e+00 -4.22458291e-01 -4.32706513e-02 8.70647669e-01
-8.94147217e-01 -1.85934842e-01 2.26644352e-01 -4.27418977e-01
1.06149876e+00 4.07250673e-01 1.17604315e+00 1.21262586e+00
5.18097401e-01 7.71151364e-01 1.53788149e+00 -5.56530058e-01
1.84510842e-01 2.24055752e-01 4.46847856e-01 4.48085129e-01
3.48586828e-01 2.14565948e-01 -7.66017616e-01 2.90381670e-01
7.62070060e-01 -3.66592139e-01 -1.85233459e-01 -3.49571288e-01
-1.44764376e+00 6.09384477e-01 6.61267698e-01 2.66182303e-01
-2.09341422e-01 -1.10702053e-01 2.75287807e-01 2.12329492e-01
2.27134913e-01 8.27582240e-01 -3.61940712e-01 -1.46799028e-01
-1.31766903e+00 5.70656836e-01 5.16676486e-01 6.63096309e-01
6.93587184e-01 -2.87222981e-01 -2.29191378e-01 5.65500915e-01
2.13492393e-01 4.69176680e-01 7.61115730e-01 -7.95518279e-01
6.88022077e-02 4.46629733e-01 1.01592511e-01 -9.81960952e-01
-4.30287004e-01 -3.67375642e-01 -3.03673476e-01 3.90540868e-01
5.20463347e-01 4.05789971e-01 -9.14416134e-01 1.52237916e+00
-1.64158225e-01 -1.46918267e-01 -5.36549032e-01 9.39149559e-01
5.17221272e-01 3.09161276e-01 2.35558510e-01 -8.23234953e-03
1.21312559e+00 -1.00577939e+00 -3.10118616e-01 -2.39749879e-01
4.42121744e-01 -7.26665378e-01 1.31205750e+00 5.32774627e-01
-1.24674547e+00 -7.55521357e-01 -9.60710704e-01 4.34762090e-02
-4.66987759e-01 -1.86855748e-01 5.99142730e-01 4.72652316e-01
-1.34959114e+00 7.55687118e-01 -7.88495839e-01 -9.04640913e-01
4.60354954e-01 2.34351128e-01 -9.27194133e-02 2.19000056e-01
-1.03149939e+00 1.41499102e+00 4.36825514e-01 -2.89010763e-01
-8.94770563e-01 -7.23116696e-01 -4.37464386e-01 1.03339091e-01
5.71420193e-02 -7.38610685e-01 1.19053054e+00 -1.30875278e+00
-8.75580311e-01 1.38933194e+00 -4.97023493e-01 -8.68587911e-01
4.01480585e-01 1.52963519e-01 -3.13557208e-01 1.14772141e-01
6.65512830e-02 1.34022081e+00 6.72887981e-01 -9.27057385e-01
-5.76519251e-01 -2.02030703e-01 8.21612403e-02 2.00408161e-01
3.08349431e-01 3.37387592e-01 -1.16069444e-01 -3.72270644e-01
-1.12390131e-01 -9.85773027e-01 -4.08342667e-02 -7.37979934e-02
-4.38007057e-01 -2.45963022e-01 2.60638773e-01 -3.68874520e-01
1.28849566e+00 -1.68510497e+00 2.15176865e-01 8.33944157e-02
1.88125104e-01 3.68460804e-01 -1.24686711e-01 4.07002300e-01
-4.66304481e-01 2.19795123e-01 -1.61095485e-02 -1.78003069e-02
-5.91507331e-02 1.13031743e-02 -2.98264861e-01 1.90076083e-01
1.55372232e-01 1.08471608e+00 -8.72009158e-01 -5.85542440e-01
1.57539681e-01 1.49170086e-01 -5.25694191e-01 -3.64599563e-02
-1.84943751e-01 3.78066450e-01 -1.21774792e-03 5.75227022e-01
3.88122648e-01 -4.63081330e-01 -2.79202282e-01 2.47877777e-01
-2.88884670e-01 3.65677595e-01 -4.77495879e-01 1.29154372e+00
8.14145654e-02 1.10326850e+00 -6.94142163e-01 -6.50642991e-01
6.88112974e-01 -1.70062616e-01 -7.56543353e-02 -1.12735319e+00
1.37613446e-01 2.65828431e-01 6.40455067e-01 -2.62724966e-01
4.94377136e-01 -5.64675732e-03 3.44186217e-01 5.84296644e-01
2.09033713e-01 7.10218176e-02 5.38656056e-01 1.07096471e-01
1.02730167e+00 2.54312396e-01 6.12979293e-01 -4.59497511e-01
4.67030793e-01 2.44474560e-01 -7.08353147e-02 8.21250200e-01
-3.85203272e-01 7.88115561e-01 6.31445110e-01 -5.51274836e-01
-1.27304554e+00 -1.17706788e+00 -3.02212626e-01 1.13512933e+00
1.96936116e-01 -2.93908268e-01 -1.14142168e+00 -3.72002572e-01
-4.07198519e-01 1.06339097e+00 -1.19099271e+00 -2.75640309e-01
-3.89145583e-01 -8.49436581e-01 3.31265122e-01 3.65951777e-01
3.58883202e-01 -1.73126674e+00 -1.32966781e+00 -3.05273086e-01
1.95658393e-03 -5.37999094e-01 1.86565593e-02 1.40275165e-01
-8.69352400e-01 -1.37392414e+00 -6.54677331e-01 -4.50805008e-01
5.27568638e-01 4.89316702e-01 1.53531516e+00 4.03253168e-01
-4.39925790e-01 1.68144509e-01 -3.13315183e-01 -8.90651345e-01
-1.84748054e-01 2.42204383e-01 -1.01618029e-01 -5.16761482e-01
7.11151659e-01 -5.05012512e-01 -8.80687118e-01 7.00091720e-01
-6.94367945e-01 3.97046000e-01 8.09113085e-01 5.79828382e-01
6.64695978e-01 -4.77258235e-01 2.83471137e-01 -9.22673881e-01
3.30145001e-01 -5.76722383e-01 -6.20293379e-01 2.04679877e-01
-1.03845727e+00 -3.62943895e-02 1.73144385e-01 -1.98712200e-01
-9.27085936e-01 -2.50492096e-01 3.33694555e-02 -6.28183559e-02
-4.38044459e-01 3.14207017e-01 3.39501649e-01 -2.92661656e-02
9.11806166e-01 3.65206301e-01 -1.95519045e-01 -3.24182093e-01
1.49671867e-01 2.75328755e-02 3.95918041e-01 1.26360985e-03
6.98780417e-01 3.35471392e-01 1.65714607e-01 -8.37347269e-01
-9.14364040e-01 -4.56856787e-01 -9.80984211e-01 -3.80434334e-01
8.86283159e-01 -4.02441025e-01 -6.57150090e-01 2.89730728e-01
-1.00785875e+00 -4.36558336e-01 -1.43646643e-01 2.38918424e-01
-9.04176533e-01 6.28728420e-02 -7.51137957e-02 -6.46886289e-01
2.47418694e-03 -9.60980833e-01 8.63082647e-01 5.59269369e-01
-7.13580251e-01 -1.13657868e+00 3.38498414e-01 3.38919669e-01
4.05454248e-01 1.26826063e-01 1.04455030e+00 -7.10350454e-01
-6.29943311e-01 -2.14819908e-02 -3.38744462e-01 -4.33510076e-03
-3.54037076e-01 2.35102400e-01 -1.04290843e+00 2.37067584e-02
-1.88658193e-01 -8.86843130e-02 1.04736769e+00 6.74697280e-01
1.08095539e+00 1.24799162e-01 -6.22639239e-01 3.25697899e-01
1.38557923e+00 -7.75943100e-02 9.52722847e-01 7.33775496e-01
3.03645402e-01 8.50515604e-01 1.01280153e+00 -1.28698602e-01
9.41033810e-02 7.22162783e-01 7.67363906e-01 6.67509809e-02
-1.62658140e-01 -3.49763930e-01 1.53878525e-01 1.43915206e-01
-3.91686589e-01 -4.10429507e-01 -1.06884301e+00 6.35300696e-01
-1.72050726e+00 -1.26853061e+00 -5.58981836e-01 2.40646791e+00
3.36155057e-01 5.66964805e-01 4.99382943e-01 1.91449866e-01
7.76429594e-01 1.31766692e-01 -4.01816905e-01 -5.61869323e-01
-1.78812221e-01 2.18164757e-01 4.17460471e-01 2.17335850e-01
-7.09858537e-01 8.95258725e-01 7.78680468e+00 6.75461352e-01
-9.41583693e-01 1.20608464e-01 7.57847965e-01 -4.21697319e-01
-1.20283641e-01 8.70960802e-02 -8.34250510e-01 6.17794573e-01
1.42306662e+00 -3.18470955e-01 3.69019121e-01 6.39878213e-01
5.03508151e-01 -7.99616575e-01 -1.30653548e+00 7.91474581e-01
1.71863452e-01 -1.17173278e+00 -6.41758228e-03 1.25936955e-01
6.97730601e-01 2.66200036e-01 3.83937836e-01 1.61064193e-01
5.46789542e-02 -1.25474381e+00 9.09898937e-01 7.59888828e-01
2.33202532e-01 -4.95439172e-01 5.60843647e-01 3.39908838e-01
-4.26576555e-01 -2.37940736e-02 -7.59587526e-01 -2.97823936e-01
1.30710676e-01 2.77874500e-01 -9.74106193e-01 1.64892256e-01
8.53527844e-01 7.71523058e-01 -1.46779799e+00 1.62939024e+00
-3.22989732e-01 7.98311651e-01 9.17246714e-02 -2.40993917e-01
2.77654290e-01 3.10352705e-02 6.30434036e-01 1.19947588e+00
2.29681596e-01 -3.80690634e-01 -6.44412398e-01 1.18379080e+00
5.04509866e-01 1.34578243e-01 -4.57476735e-01 1.37814239e-01
2.11179733e-01 1.19934881e+00 -1.10641599e+00 -1.20853335e-01
-2.23663643e-01 7.41596222e-01 2.60428697e-01 1.84623357e-02
-1.07356489e+00 -3.44787650e-02 3.16137403e-01 5.72023511e-01
1.89569592e-01 1.76940247e-01 -4.34946269e-01 -6.81428969e-01
-2.16593042e-01 -6.20506644e-01 1.64254725e-01 -1.38897991e+00
-9.80789602e-01 6.96845412e-01 4.21316922e-01 -9.92776275e-01
-2.71196246e-01 -7.42526889e-01 -8.97337019e-01 1.03902924e+00
-1.45527601e+00 -8.68871152e-01 -4.23809052e-01 1.41628608e-01
6.66979015e-01 -3.98701988e-02 5.75577378e-01 -2.60666639e-01
-4.55117822e-01 1.62221789e-01 -2.82129258e-01 -4.69626963e-01
8.43500793e-01 -1.46291399e+00 7.14204371e-01 5.94960690e-01
3.54656130e-01 6.09423101e-01 1.32175088e+00 -5.33989966e-01
-3.79901499e-01 -6.07329488e-01 1.04973769e+00 -1.00281191e+00
6.90801978e-01 -1.19239613e-01 -9.49046016e-01 5.71954727e-01
4.78710860e-01 -4.71071303e-01 5.62779069e-01 2.44392172e-01
-1.70855403e-01 2.17510194e-01 -8.28751802e-01 6.87489986e-01
1.22361004e+00 -3.34043838e-02 -9.02466893e-01 3.20370346e-01
3.82762700e-01 -9.29889828e-02 -2.02614263e-01 -7.81814530e-02
6.67100072e-01 -1.87836754e+00 8.15576196e-01 -7.02120602e-01
7.68725336e-01 -1.01371534e-01 2.95260459e-01 -1.32360005e+00
-6.67209506e-01 -2.09728062e-01 2.85651058e-01 8.25076818e-01
5.82777679e-01 -4.36814845e-01 7.76715875e-01 1.07601374e-01
-1.07986346e-01 -7.89891660e-01 -6.45888805e-01 -6.57179177e-01
-1.25729680e-01 -1.43197954e-01 3.94519955e-01 4.58990097e-01
-1.23596378e-01 2.37264231e-01 -7.08228722e-02 -3.08869958e-01
4.76938039e-01 -8.66626650e-02 6.30856395e-01 -1.55540621e+00
-3.23601514e-01 -8.92298639e-01 -7.67589092e-01 -5.54414570e-01
1.38147578e-01 -6.96909130e-01 -8.43460578e-03 -1.29896533e+00
6.22171283e-01 1.95711404e-02 -5.80306947e-01 2.66815752e-01
-2.81900257e-01 5.16572654e-01 1.38713464e-01 5.16234040e-01
-8.28502953e-01 -5.29301688e-02 1.21811295e+00 1.91010848e-01
1.67422920e-01 1.76004320e-01 -6.97591364e-01 7.96967685e-01
8.79359663e-01 -4.80573416e-01 -7.20488012e-01 -2.34540880e-01
6.18853211e-01 -4.35439616e-01 8.33121479e-01 -1.33076644e+00
5.63629530e-02 -2.37609088e-01 5.65798700e-01 -5.76052129e-01
1.56123698e-01 -4.03092146e-01 6.11042194e-02 5.59238195e-01
-5.72187662e-01 4.15462077e-01 1.08591154e-01 6.00388944e-01
-7.13182241e-02 -4.99195397e-01 9.88800526e-01 -3.47006112e-01
-9.48913455e-01 -5.23981415e-02 -3.84837925e-01 -2.65848227e-02
9.47921932e-01 -6.75170243e-01 -5.05301595e-01 -3.11171710e-01
-7.67536998e-01 -2.87203677e-02 9.04248297e-01 5.42009294e-01
4.00867462e-01 -9.09230590e-01 -4.37635481e-01 3.75742912e-02
2.57402092e-01 -6.41134143e-01 2.32102334e-01 1.35456824e+00
-6.30007923e-01 8.48220706e-01 -9.68710124e-01 -6.94041491e-01
-1.50105417e+00 8.62155557e-01 2.21476242e-01 -1.86746746e-01
-7.97219276e-02 7.74800003e-01 4.65205222e-01 2.12611198e-01
-2.76013821e-01 -1.95420206e-01 -4.64934140e-01 1.42595634e-01
5.35661578e-01 5.01653850e-01 -6.11358695e-02 -7.55920649e-01
-2.41289258e-01 3.77613187e-01 -1.10037081e-01 -1.40731409e-01
1.32334626e+00 -2.65560091e-01 -1.51067451e-01 7.40924358e-01
5.73062539e-01 -3.77013922e-01 -1.16126323e+00 3.41872275e-01
2.35672116e-01 -6.34047210e-01 -3.23775262e-01 -1.12848294e+00
-5.67970455e-01 1.13495100e+00 5.90753019e-01 3.01506042e-01
1.02352560e+00 5.73718213e-02 2.73287594e-01 2.63248622e-01
4.57136452e-01 -9.20388877e-01 2.23805130e-01 3.71729314e-01
7.75353968e-01 -1.22063398e+00 -9.36626643e-02 -2.52549082e-01
-8.25789928e-01 9.33296859e-01 6.42128110e-01 -3.44203591e-01
2.38537848e-01 -4.55036759e-01 -2.38198116e-01 -3.89061183e-01
-9.97519791e-01 -2.95664340e-01 5.49956083e-01 8.60290885e-01
5.58444262e-01 -9.77208614e-02 -3.93484950e-01 2.01136306e-01
-5.18961489e-01 -1.03126355e-01 8.05408001e-01 5.28701544e-01
-9.43233430e-01 -1.02746880e+00 -2.50352263e-01 6.56349301e-01
-3.27320009e-01 -1.77031964e-01 -6.93055034e-01 1.05005443e+00
1.55788884e-01 6.01680577e-01 7.98647776e-02 -1.62304491e-01
1.48167983e-01 2.54891515e-01 8.23904932e-01 -7.57297635e-01
-5.42179585e-01 -2.63632476e-01 -7.83503875e-02 -5.81900656e-01
-5.79340518e-01 -9.71095860e-01 -8.44594061e-01 -3.80357981e-01
-1.02653719e-01 -1.93858873e-02 5.30645788e-01 8.64257216e-01
3.98887247e-02 3.39263946e-01 1.39422959e-03 -1.10973775e+00
-1.39630526e-01 -9.22487080e-01 -6.92482471e-01 4.62106079e-01
2.82622874e-03 -9.47244644e-01 -3.49391699e-01 1.63052961e-01] | [10.02005672454834, 1.5453217029571533] |
eaf8e101-5bcb-4556-8ca9-6eb3d9d3d0db | what-can-i-do-here-leveraging-deep-3d | 1812.00889 | null | http://arxiv.org/abs/1812.00889v1 | http://arxiv.org/pdf/1812.00889v1.pdf | What can I do here? Leveraging Deep 3D saliency and geometry for fast and scalable multiple affordance detection | This paper develops and evaluates a novel method that allows for the
detection of affordances in a scalable and multiple-instance manner on visually
recovered pointclouds. Our approach has many advantages over alternative
methods, as it is based on highly parallelizable, one-shot learning that is
fast in commodity hardware. The approach is hybrid in that it uses a geometric
representation together with a state-of-the-art deep learning method capable of
identifying 3D scene saliency. The geometric component allows for a compact and
efficient representation, boosting the performance of the deep network
architecture which proved insufficient on its own. Moreover, our approach
allows not only to predict whether an input scene affords or not the
interactions, but also the pose of the objects that allow these interactions to
take place. Our predictions align well with crowd-sourced human judgment as
they are preferred with 87% probability, show high rates of improvement with
almost four times (4x) better performance over a deep learning-only baseline
and are seven times (7x) faster than previous art. | ['Walterio Mayol-Cuevas', 'Eduardo Ruiz'] | 2018-12-03 | null | null | null | null | ['multiple-affordance-detection', 'affordance-detection'] | ['computer-vision', 'computer-vision'] | [ 1.11280113e-01 -4.64935862e-02 1.71185955e-01 -1.82576999e-01
-5.28747320e-01 -3.18012387e-01 7.51484096e-01 3.82099450e-01
-5.47417402e-01 3.26896012e-01 3.54212373e-01 -1.97937950e-01
-1.19335979e-01 -5.76616824e-01 -8.47664118e-01 -4.63029236e-01
-1.33234948e-01 6.68166697e-01 9.01759267e-01 -4.04652774e-01
4.36679006e-01 8.26164007e-01 -2.19652772e+00 2.65861243e-01
3.70410770e-01 1.07795835e+00 5.75940490e-01 7.45707393e-01
2.41142422e-01 8.55918884e-01 -3.68386120e-01 3.99424508e-02
3.23321521e-01 2.04355165e-01 -6.56124353e-01 -1.88042328e-01
8.76419485e-01 -7.01882184e-01 -1.83942810e-01 4.26755875e-01
6.66715086e-01 2.84493774e-01 6.84852064e-01 -1.12416136e+00
-8.05323645e-02 -2.50654370e-01 -6.08005583e-01 3.92673612e-01
6.07309997e-01 4.29432094e-01 1.08311450e+00 -9.30915415e-01
7.17433989e-01 1.13126421e+00 5.89572012e-01 3.55062872e-01
-1.00608301e+00 -3.03218544e-01 -3.53335291e-02 2.02462152e-01
-1.03606057e+00 -5.18801391e-01 4.92467731e-01 -5.47092974e-01
1.61381519e+00 2.67964303e-01 8.42705965e-01 7.88320422e-01
-1.31742224e-01 8.52562189e-01 8.00338328e-01 -4.14659411e-01
4.54149872e-01 -1.85127988e-01 -1.96857020e-01 7.47823656e-01
1.20015554e-01 2.17300847e-01 -5.43121636e-01 -1.74485162e-01
9.31221783e-01 1.43924430e-01 -3.67249525e-03 -7.94392347e-01
-1.35981262e+00 5.65867245e-01 9.06342149e-01 1.19095422e-01
-6.59349442e-01 3.31562817e-01 2.42984027e-01 -2.59696901e-01
4.68437016e-01 5.86506784e-01 -6.55807257e-01 -1.84626758e-01
-1.04629767e+00 5.99756896e-01 4.98739004e-01 7.16348529e-01
6.94970727e-01 -1.69104218e-01 -2.00796947e-01 3.23704392e-01
2.86049414e-02 4.52933729e-01 1.60947859e-01 -1.11143041e+00
2.82239199e-01 7.63873875e-01 5.16807258e-01 -1.03189099e+00
-6.73434913e-01 -5.16423360e-02 -1.75717279e-01 8.20718229e-01
3.48376960e-01 -2.14812532e-03 -8.73473883e-01 1.30151451e+00
3.89103562e-01 2.26853609e-01 -2.60765076e-01 1.29959297e+00
7.39997804e-01 4.03256804e-01 1.84901133e-01 3.33046973e-01
1.32304466e+00 -9.63561952e-01 3.62902880e-02 -3.75709563e-01
3.59815985e-01 -7.35812962e-01 1.18324697e+00 2.68516809e-01
-9.58048940e-01 -5.38824558e-01 -1.02813375e+00 -3.70795488e-01
-3.82783920e-01 3.37637328e-02 9.58861411e-01 2.79167295e-01
-1.34181750e+00 6.88735366e-01 -8.20231676e-01 -5.00814319e-01
5.65212846e-01 4.82245862e-01 -3.23142231e-01 1.06736921e-01
-5.01516044e-01 1.11831188e+00 1.78304836e-01 -5.05815089e-01
-9.09987986e-01 -6.98868990e-01 -8.86685371e-01 2.03932628e-01
3.79237473e-01 -6.08078897e-01 1.34518301e+00 -9.58744168e-01
-1.19648004e+00 9.86908376e-01 -3.51281166e-01 -5.45981944e-01
6.06669128e-01 -4.49259430e-01 1.08807966e-01 4.13527936e-01
4.07481521e-01 1.12526309e+00 6.54706120e-01 -1.11117566e+00
-7.77276456e-01 -2.17122182e-01 1.99781686e-01 3.55522037e-01
3.37278359e-02 1.43224895e-01 -3.02090377e-01 -2.44755670e-01
-1.54901192e-01 -9.39024687e-01 -2.59490877e-01 4.59863901e-01
-1.38241649e-01 -3.22678238e-01 7.68681288e-01 -3.76299053e-01
3.61014545e-01 -2.22235346e+00 -3.14836353e-02 -2.10667640e-01
3.31400543e-01 5.00290632e-01 -4.72195223e-02 6.53035104e-01
3.73183414e-02 -2.48083457e-01 3.98351299e-03 -4.55713421e-01
-3.94916274e-02 -1.06749460e-01 -4.66240615e-01 5.34911036e-01
2.39194363e-01 8.72845948e-01 -1.06549966e+00 -2.90570080e-01
8.57060313e-01 6.47289157e-01 -4.08995718e-01 2.50958681e-01
-3.10010642e-01 3.57375801e-01 -1.90245703e-01 5.96019149e-01
4.96782392e-01 -2.39945039e-01 -1.82997942e-01 4.97629903e-02
-4.97911394e-01 3.40476125e-01 -1.04011428e+00 1.75904250e+00
-2.96170592e-01 9.11784410e-01 -2.92931288e-01 -5.38663030e-01
8.02368402e-01 2.68478215e-01 4.50742185e-01 -5.14647484e-01
1.49089783e-01 2.31945455e-01 -1.77525342e-01 -5.43007791e-01
4.45061237e-01 9.33651701e-02 2.10250556e-01 3.42813879e-01
1.11933656e-01 -1.00948133e-01 -5.03418893e-02 4.46333326e-02
1.32111084e+00 5.71456432e-01 3.64111185e-01 -2.63498604e-01
4.41755168e-02 1.62588462e-01 1.96210548e-01 6.67683125e-01
-2.67388672e-01 9.11686420e-01 2.82535255e-01 -9.49909687e-01
-1.15848434e+00 -9.00120497e-01 2.68866837e-01 1.28934538e+00
5.19514680e-01 -1.78180143e-01 -6.02190137e-01 -3.51088792e-01
-2.40244064e-02 6.90933168e-01 -6.96058989e-01 2.41295755e-01
-3.49259645e-01 -2.62935907e-01 5.84374554e-02 8.68549168e-01
3.86996955e-01 -1.25623953e+00 -1.65553272e+00 3.52659598e-02
2.91507412e-02 -1.15835333e+00 7.20536476e-03 1.37708768e-01
-7.71918595e-01 -1.38321912e+00 -6.38542771e-01 -6.58166766e-01
5.41781068e-01 6.51732981e-01 1.20833659e+00 1.70647442e-01
-5.77152669e-01 2.45234668e-01 -2.44355813e-01 -7.47845292e-01
2.99004585e-01 -1.12167567e-01 6.83692247e-02 -1.84013873e-01
6.25008881e-01 -5.18357754e-01 -8.54987741e-01 9.80611816e-02
-4.93265569e-01 2.34330460e-01 5.25149405e-01 3.20292830e-01
3.95160049e-01 -3.00301135e-01 -1.12038087e-02 -4.13715243e-01
1.93532631e-01 -2.40121678e-01 -5.74504614e-01 -5.21319686e-03
-4.74117786e-01 -2.31157281e-02 1.39958039e-01 -1.35462552e-01
-8.56507957e-01 5.50180435e-01 3.09963878e-02 -4.61405694e-01
-6.30717993e-01 -2.87643760e-01 3.26678276e-01 -2.82007664e-01
7.83948779e-01 -2.54617929e-01 -1.98065802e-01 -3.80218506e-01
2.74341494e-01 2.40153864e-01 4.00701612e-01 -2.63184875e-01
5.79546094e-01 8.87731552e-01 8.15792829e-02 -8.26173782e-01
-6.72191441e-01 -7.90450156e-01 -1.00410831e+00 -3.66559923e-01
9.75337148e-01 -9.61371422e-01 -8.40430140e-01 3.56500030e-01
-1.38248897e+00 -3.21292341e-01 -2.48271883e-01 4.37637478e-01
-6.37583077e-01 -6.98477402e-02 -3.90713722e-01 -1.04263127e+00
-2.89658487e-01 -1.04323661e+00 1.66740072e+00 2.43723020e-01
-4.20275658e-01 -6.06458008e-01 4.88729496e-03 2.23790601e-01
4.15161639e-01 6.40083075e-01 6.01952851e-01 -6.02925241e-01
-8.29737067e-01 -3.66209984e-01 -5.56726873e-01 -2.51679093e-01
-2.74843067e-01 7.42424950e-02 -1.24392045e+00 -1.51720852e-01
-3.07222664e-01 -3.42473358e-01 6.83250129e-01 4.22788858e-01
8.03892314e-01 1.05085254e-01 -5.21495759e-01 4.17502820e-01
1.61184609e+00 -2.68973708e-01 6.19208634e-01 5.90812385e-01
7.24207938e-01 6.70006156e-01 5.35904646e-01 4.01338845e-01
4.99673694e-01 8.00215483e-01 1.01122999e+00 -2.57120430e-01
-1.56684116e-01 -1.74076498e-01 -6.62553608e-02 -5.80119295e-03
-4.71421510e-01 -4.82803285e-02 -1.19352150e+00 8.09270680e-01
-2.11112738e+00 -8.78834903e-01 -3.35871011e-01 2.25694299e+00
1.85214639e-01 1.30608365e-01 4.24727350e-01 1.42232582e-01
4.53319699e-01 2.52553642e-01 -5.57095766e-01 -3.24863642e-01
1.08643450e-01 3.32702935e-01 5.67499697e-01 2.95333087e-01
-1.14754534e+00 9.97778177e-01 6.78162432e+00 3.25349599e-01
-1.14405620e+00 9.58620161e-02 3.12962681e-01 -4.71294582e-01
1.65489092e-01 -4.76737022e-02 -5.04161417e-01 1.81132823e-01
6.21769071e-01 3.94754618e-01 4.46098149e-01 1.06846428e+00
6.57059178e-02 -4.31815088e-01 -1.08032274e+00 1.00511515e+00
2.49762565e-01 -1.41497862e+00 -4.35578912e-01 1.26382068e-01
6.22744620e-01 6.62949026e-01 -1.93063214e-01 -8.01701471e-03
2.16743171e-01 -9.42353189e-01 9.07566130e-01 4.83771771e-01
5.35555542e-01 -7.29186118e-01 6.19019747e-01 5.49786091e-01
-9.74547327e-01 -9.78582278e-02 -4.32854414e-01 -7.48929977e-01
-2.05134861e-02 2.25755021e-01 -9.81995642e-01 1.28692642e-01
9.22004342e-01 4.07982737e-01 -7.24723041e-01 1.43814743e+00
-3.71690780e-01 1.45723015e-01 -4.45571452e-01 -3.32395196e-01
3.30720872e-01 4.54101384e-01 5.89652956e-01 1.13687706e+00
1.96670294e-01 -6.07120804e-02 2.75532901e-01 5.99979222e-01
2.19742119e-01 1.58719584e-01 -8.43498766e-01 4.96557981e-01
4.24737245e-01 1.19623625e+00 -1.01048088e+00 -3.29713881e-01
-2.52461463e-01 9.46016431e-01 5.84448278e-01 1.20674782e-01
-6.00266099e-01 -3.64897370e-01 7.24559724e-01 4.01149362e-01
7.39311516e-01 -3.48467261e-01 -4.03001428e-01 -8.62479091e-01
2.08162621e-01 -4.44152892e-01 -6.43231049e-02 -1.22523952e+00
-8.47702801e-01 7.43925691e-01 -1.99166182e-02 -1.22648227e+00
-3.10958087e-01 -7.76001930e-01 -7.33453929e-01 6.64419651e-01
-1.50183749e+00 -1.35528076e+00 -6.04673505e-01 4.31240439e-01
4.62670833e-01 -1.27435001e-02 1.07771385e+00 -1.06106207e-01
-4.02082540e-02 -4.85161059e-02 -3.02729905e-01 -2.77315117e-02
4.99344021e-01 -1.27803612e+00 6.96111143e-01 8.88248503e-01
3.18437129e-01 2.43665546e-01 8.17390501e-01 -4.28843975e-01
-1.16097677e+00 -7.48740375e-01 9.05210972e-01 -8.34676385e-01
4.65830863e-01 -4.18820411e-01 -6.67730153e-01 3.94706100e-01
2.16660634e-01 1.98543280e-01 3.28757882e-01 1.50294736e-01
-4.14510787e-01 1.03247397e-01 -1.17988539e+00 5.68845034e-01
9.63714004e-01 -4.41888571e-01 -6.79112971e-01 5.12542009e-01
6.04201019e-01 -6.61850691e-01 -1.42218396e-01 4.22955245e-01
5.01802027e-01 -1.50532401e+00 1.04961920e+00 -4.24606472e-01
3.69498819e-01 -3.48401815e-01 -1.38399348e-01 -8.61441135e-01
-4.40718234e-01 -4.19201940e-01 -2.45723426e-01 6.18711650e-01
2.00211823e-01 -2.82536030e-01 7.52194583e-01 6.93404853e-01
-1.09646702e-02 -8.09629321e-01 -9.17122424e-01 -5.11102498e-01
-6.21267498e-01 -4.81558383e-01 7.06563771e-01 5.68754494e-01
-2.76222050e-01 3.70347947e-01 -2.87600517e-01 2.50112653e-01
5.27460814e-01 1.49286881e-01 8.60338271e-01 -1.49195623e+00
-1.14817202e-01 -5.19140065e-01 -9.16817069e-01 -8.13198030e-01
-5.56540936e-02 -3.60122859e-01 1.92781880e-01 -1.80009711e+00
-4.17274050e-02 -2.48742178e-01 -2.60010958e-01 6.45262897e-01
2.27778852e-02 3.70370179e-01 1.96456730e-01 3.35816205e-01
-8.12229753e-01 1.82487488e-01 7.73114324e-01 1.41384706e-01
-8.02864730e-02 -1.41343892e-01 -3.95492703e-01 8.81449401e-01
6.70816481e-01 -2.27193594e-01 -2.36359835e-01 -7.16370702e-01
1.67074919e-01 -2.99008101e-01 8.17890406e-01 -1.50004828e+00
2.35719919e-01 1.87983736e-03 6.03561938e-01 -5.32608449e-01
6.97183967e-01 -6.43599093e-01 -1.66104510e-01 4.10593718e-01
-1.72174767e-01 9.57670528e-03 2.94732660e-01 5.83397269e-01
1.82937115e-01 -1.17805243e-01 6.13974750e-01 -3.68795425e-01
-1.18666887e+00 3.05074930e-01 -1.11417137e-01 -2.44048044e-01
1.18604815e+00 -3.14992607e-01 -1.40762106e-01 -3.26530814e-01
-2.97491044e-01 -7.99120665e-02 8.36721420e-01 6.34575307e-01
6.49809182e-01 -1.00232267e+00 -5.46956539e-01 8.93532187e-02
1.76885262e-01 1.66144013e-01 -1.55659914e-02 7.01648951e-01
-8.22200537e-01 3.74886900e-01 -3.67984504e-01 -7.74405360e-01
-1.24590516e+00 6.52170062e-01 2.14877069e-01 2.21852839e-01
-9.72707093e-01 1.06406176e+00 1.39013929e-02 -2.17070296e-01
3.35020155e-01 -1.67943150e-01 -1.55417010e-01 -1.78822398e-01
7.84955978e-01 4.05731022e-01 1.27059266e-01 -7.38024294e-01
-5.97525775e-01 5.48511982e-01 3.01276207e-01 -1.71160161e-01
1.61022568e+00 1.78528816e-01 -6.66687563e-02 5.36102891e-01
9.28183734e-01 -1.46702647e-01 -1.81639814e+00 -3.52906361e-02
-4.67917472e-02 -7.72900403e-01 2.54218519e-01 -8.23929548e-01
-6.86189950e-01 1.04587638e+00 6.37650251e-01 2.05610007e-01
8.20614755e-01 1.46485612e-01 7.00448990e-01 4.55337852e-01
6.54521346e-01 -1.05784798e+00 1.94009721e-01 4.04138476e-01
8.14114511e-01 -1.45886207e+00 1.52701691e-01 -2.14348003e-01
-7.02202559e-01 1.05317044e+00 6.22335553e-01 -6.79092288e-01
3.38768125e-01 5.88379279e-02 -2.56966054e-02 -6.53030396e-01
-7.02958763e-01 -6.35112286e-01 4.25470054e-01 8.26763868e-01
3.27541798e-01 1.17614634e-01 3.59761447e-01 -1.70297489e-01
-1.52272612e-01 -1.17535628e-01 9.60558578e-02 8.88666272e-01
-7.28657305e-01 -5.31344950e-01 -1.17551647e-01 2.83556938e-01
-1.66048303e-01 -8.06760490e-02 -3.80916953e-01 7.52238333e-01
3.55343491e-01 9.29895282e-01 1.89598233e-01 -2.93005884e-01
4.71920729e-01 -1.07018545e-01 5.80647409e-01 -8.09909105e-01
-5.43949306e-01 -1.26301259e-01 3.70082259e-02 -9.38004434e-01
-5.02193868e-01 -8.27342153e-01 -1.15578473e+00 -9.27908570e-02
-2.36397952e-01 -4.19258952e-01 9.75453556e-01 1.25352490e+00
6.51779830e-01 3.96099418e-01 3.77077460e-01 -1.64893305e+00
-6.53751045e-02 -7.05908954e-01 -1.59438282e-01 3.46886128e-01
4.65008169e-01 -9.40836132e-01 -1.75074592e-01 -2.45667085e-01] | [7.771450042724609, -1.8089286088943481] |
63cdfea1-a39a-445e-b8b7-f5b066f12cae | lemma-a-multi-view-dataset-for-learning-multi | 2007.15781 | null | https://arxiv.org/abs/2007.15781v1 | https://arxiv.org/pdf/2007.15781v1.pdf | LEMMA: A Multi-view Dataset for Learning Multi-agent Multi-task Activities | Understanding and interpreting human actions is a long-standing challenge and a critical indicator of perception in artificial intelligence. However, a few imperative components of daily human activities are largely missed in prior literature, including the goal-directed actions, concurrent multi-tasks, and collaborations among multi-agents. We introduce the LEMMA dataset to provide a single home to address these missing dimensions with meticulously designed settings, wherein the number of tasks and agents varies to highlight different learning objectives. We densely annotate the atomic-actions with human-object interactions to provide ground-truths of the compositionality, scheduling, and assignment of daily activities. We further devise challenging compositional action recognition and action/task anticipation benchmarks with baseline models to measure the capability of compositional action understanding and temporal reasoning. We hope this effort would drive the machine vision community to examine goal-directed human activities and further study the task scheduling and assignment in the real world. | ['Song-Chun Zhu', 'Yixin Zhu', 'Baoxiong Jia', 'Siyuan Huang', 'Yixin Chen'] | 2020-07-31 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5581_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123710766.pdf | eccv-2020-8 | ['action-understanding'] | ['computer-vision'] | [ 6.95227623e-01 5.99978529e-02 -3.06454837e-01 -3.99184614e-01
-4.61537153e-01 -6.27210140e-01 1.10812473e+00 -8.86627883e-02
-1.99473709e-01 5.84507346e-01 9.40223217e-01 -7.04811560e-03
-1.95835501e-01 -2.86609888e-01 -6.91888571e-01 -4.18575138e-01
-4.84415501e-01 6.78993165e-01 1.62600726e-01 -1.86746180e-01
4.60054278e-02 5.79858050e-02 -1.52143800e+00 6.24375701e-01
3.68150383e-01 6.55420244e-01 7.18323737e-02 7.67801702e-01
3.40820551e-01 1.61137915e+00 -3.20982426e-01 -8.97036269e-02
4.22934622e-01 -5.64153969e-01 -1.13198376e+00 6.40408874e-01
3.05869788e-01 -4.43856627e-01 -4.60285008e-01 5.70125461e-01
1.20851994e-01 4.17524606e-01 5.61296165e-01 -1.78109419e+00
-5.53901792e-01 8.29594910e-01 -4.40901816e-01 3.62090260e-01
6.27143681e-01 6.67452574e-01 1.25940442e+00 -2.52162963e-02
4.71545368e-01 1.43178380e+00 3.84759188e-01 3.80833536e-01
-1.01432371e+00 -4.27692384e-01 5.41915834e-01 4.37440813e-01
-7.28611171e-01 -5.71739256e-01 4.57002610e-01 -8.32239628e-01
1.16567433e+00 1.26922727e-01 5.19198895e-01 1.51372671e+00
5.59830479e-02 1.23895562e+00 9.32412207e-01 -1.65693149e-01
1.58034906e-01 -6.88836098e-01 2.99086988e-01 7.69559562e-01
-9.61060524e-02 1.67735502e-01 -9.10082340e-01 -1.82311516e-02
7.63607919e-01 3.43523115e-01 -2.16749366e-02 -3.97140920e-01
-1.99093580e+00 3.99901032e-01 4.16332483e-02 1.33870870e-01
-7.59092867e-01 6.22198105e-01 4.87712264e-01 1.46775797e-01
6.62795380e-02 7.11640656e-01 -4.27570045e-01 -6.64859533e-01
-2.95550793e-01 6.09313309e-01 8.04685891e-01 1.00257838e+00
5.39510608e-01 -1.71637669e-01 -5.95643282e-01 3.36974233e-01
-1.48032025e-01 4.49253172e-01 1.69071540e-01 -1.64137864e+00
7.12326348e-01 8.73525798e-01 5.67938983e-01 -7.53346384e-01
-6.54332221e-01 2.12022681e-02 -4.83262122e-01 -7.36051276e-02
7.77638674e-01 -1.22581452e-01 -6.13832831e-01 1.92518365e+00
2.00098738e-01 2.29234993e-01 1.98577177e-02 8.76976550e-01
1.66070908e-01 2.47297376e-01 4.64831769e-01 6.09994680e-02
1.56831968e+00 -1.49378753e+00 -6.66890383e-01 -9.14960265e-01
7.61075616e-01 -2.26401180e-01 1.05872703e+00 1.82765469e-01
-1.08840203e+00 -5.57985008e-01 -5.61510444e-01 1.31144067e-02
4.41197902e-02 -2.26815008e-02 1.19609988e+00 8.60458016e-02
-5.60941100e-01 2.32961610e-01 -1.38703942e+00 -4.00122255e-01
5.17019451e-01 1.03123754e-03 -3.29327375e-01 -6.55956045e-02
-8.86153698e-01 1.01636791e+00 2.52487391e-01 -2.65042037e-01
-1.49973071e+00 -4.85077053e-01 -9.64842618e-01 -4.68100235e-02
1.02754438e+00 -6.76770687e-01 1.84669089e+00 -9.77127492e-01
-9.82687712e-01 7.07185447e-01 -2.90380567e-01 -6.10636473e-01
5.30623615e-01 -3.06887060e-01 -1.56684861e-01 -1.00429222e-01
4.06232655e-01 6.34327054e-01 4.77909684e-01 -9.01289940e-01
-9.44346726e-01 -5.03454745e-01 5.77058613e-01 5.49405456e-01
5.57734035e-02 1.97829586e-02 -1.63816988e-01 -3.98506284e-01
-1.82762355e-01 -1.21985221e+00 -1.45885244e-01 -2.16158718e-01
-1.95084468e-01 -4.90931273e-01 3.69330943e-01 -5.09745240e-01
9.22912896e-01 -2.16789770e+00 3.24199438e-01 -5.69465160e-01
3.43118280e-01 -4.19634730e-01 -3.26725096e-01 6.53142035e-01
4.01562229e-02 -2.20909327e-01 -9.35454220e-02 -4.67708051e-01
4.36353087e-01 3.51947427e-01 -3.26406598e-01 3.48627687e-01
1.16806276e-01 1.40541637e+00 -1.25556183e+00 -3.21375221e-01
1.64268866e-01 -1.07629284e-01 -4.18301344e-01 2.45242298e-01
-7.32748926e-01 6.00203156e-01 -6.05900228e-01 8.95748377e-01
-1.20384000e-01 -6.31267250e-01 4.33877617e-01 -1.45594217e-02
1.74955353e-01 4.77522582e-01 -1.00193632e+00 1.86234152e+00
-2.28992477e-01 5.91155529e-01 -9.02075414e-03 -7.59137154e-01
9.16171372e-02 3.56360048e-01 7.95024931e-01 -9.68253374e-01
-1.93933725e-01 -3.84550422e-01 3.87792617e-01 -7.78243423e-01
4.50507492e-01 1.47889495e-01 -3.28005224e-01 9.09439683e-01
-3.63649666e-01 1.90794185e-01 4.59071368e-01 3.22598904e-01
1.62683582e+00 4.82344657e-01 5.09718537e-01 1.19866349e-01
-5.88964373e-02 4.60295081e-01 4.79585916e-01 9.56167102e-01
-7.94886768e-01 2.01842219e-01 6.63886011e-01 -9.48591888e-01
-9.41553831e-01 -8.42457831e-01 6.67671680e-01 1.93508208e+00
1.72796667e-01 -2.31456384e-01 -3.73029590e-01 -7.15487778e-01
-6.05711155e-03 7.37954378e-01 -7.83952057e-01 -6.05490468e-02
-7.79726148e-01 -5.94010890e-01 4.82112139e-01 9.70630646e-01
4.11165893e-01 -1.54533970e+00 -1.21658170e+00 1.65113449e-01
-7.07434416e-01 -1.60098314e+00 -6.79420948e-01 1.18334904e-01
-3.25172901e-01 -1.47427845e+00 -2.18771547e-01 -4.39633131e-01
4.52698201e-01 3.65047753e-01 1.56100249e+00 -1.45737126e-01
-1.66397557e-01 8.25789332e-01 -4.22371238e-01 -4.27475989e-01
-2.93824911e-01 -2.02096954e-01 1.48106232e-01 2.42085792e-02
4.93528634e-01 -6.56720281e-01 -5.82001567e-01 6.41071439e-01
-6.19479001e-01 4.55511153e-01 7.48086393e-01 3.33556116e-01
1.14971407e-01 -7.88889974e-02 2.62798935e-01 -6.00690544e-01
5.09745121e-01 -5.22147119e-01 -2.87961334e-01 3.77008766e-01
-1.55528039e-01 1.55991092e-01 2.44368359e-01 -7.18280733e-01
-9.44250941e-01 1.78483695e-01 7.48619914e-01 -1.16110958e-01
-4.29594874e-01 2.70393282e-01 -9.08117667e-02 5.29491961e-01
6.24531686e-01 3.75621706e-01 -7.86519572e-02 -6.61050230e-02
3.63918602e-01 2.65976071e-01 6.73043609e-01 -1.12548828e+00
3.59499276e-01 6.78652465e-01 -2.66309474e-02 -3.90733778e-01
-1.00558746e+00 -3.93593699e-01 -4.96926367e-01 -2.53796637e-01
1.04326451e+00 -1.19574988e+00 -1.26650620e+00 4.65383887e-01
-1.18244481e+00 -1.02696955e+00 -6.15397468e-02 3.13395917e-01
-9.98331487e-01 3.41227919e-01 -4.76971865e-01 -9.13215041e-01
1.33275777e-01 -1.27812004e+00 1.37991738e+00 -7.56602809e-02
-8.11075628e-01 -7.93618441e-01 5.08957766e-02 9.38630939e-01
1.14959002e-01 2.81520605e-01 7.48144448e-01 -6.55878425e-01
-9.09191668e-01 1.61087707e-01 -1.48789093e-01 -4.62440252e-02
3.11017781e-01 -5.42505264e-01 -5.82135022e-01 -6.55550733e-02
-2.54242182e-01 -8.37345004e-01 5.78985512e-01 3.55837256e-01
9.95269179e-01 -2.62198001e-01 -3.67425770e-01 1.05725639e-01
7.50654578e-01 3.84356320e-01 4.39990819e-01 2.70081252e-01
5.61209857e-01 7.54649460e-01 7.17649877e-01 6.05869949e-01
9.80917454e-01 6.86128736e-01 5.26626587e-01 2.08239257e-01
-1.14124604e-02 -2.15035215e-01 4.99895722e-01 -3.32164578e-02
-2.85967827e-01 -3.87396991e-01 -1.25534117e+00 7.13383555e-01
-2.33608508e+00 -1.42189097e+00 1.42066255e-01 1.63701153e+00
6.77154541e-01 1.00423589e-01 4.10157114e-01 -1.84691563e-01
3.37550104e-01 5.80146611e-01 -6.07584357e-01 1.28238067e-01
2.89727449e-01 -3.43317121e-01 3.25110972e-01 5.34698188e-01
-1.38222563e+00 8.54375899e-01 6.90574884e+00 1.85850337e-01
-4.17128861e-01 6.18285723e-02 3.57666373e-01 -3.76793027e-01
1.36354312e-01 -1.01046391e-01 -6.47124767e-01 5.09131789e-01
6.73267722e-01 4.09993045e-02 8.32458913e-01 6.86792850e-01
4.43955600e-01 -2.80797094e-01 -1.92498243e+00 7.88617134e-01
8.88451859e-02 -9.16553497e-01 -1.79785624e-01 1.85554042e-01
7.33715177e-01 -8.92462302e-03 -2.42057636e-01 6.47637010e-01
1.01699388e+00 -1.15691209e+00 7.96722591e-01 6.20635808e-01
1.82791069e-01 -4.32826094e-02 1.79091632e-01 6.11453116e-01
-1.19451821e+00 -5.31851888e-01 4.30615842e-01 -8.65587592e-01
5.02834678e-01 -7.70111978e-02 -6.04655325e-01 1.81814045e-01
5.16355515e-01 8.77345264e-01 -4.06798899e-01 3.41424972e-01
-1.32955983e-01 2.97392964e-01 -1.33177593e-01 -3.24273040e-03
4.13498014e-01 -1.36421546e-01 3.70721817e-01 8.20594251e-01
-2.41154179e-01 2.62751043e-01 8.62383008e-01 7.79412568e-01
1.27339035e-01 -7.23356009e-01 -5.16960979e-01 -7.62303293e-01
3.97550493e-01 7.99323916e-01 -4.60682273e-01 -4.20728952e-01
-5.76366484e-01 9.88110721e-01 3.32237184e-01 5.07812858e-01
-1.07778132e+00 4.15561646e-01 1.23274469e+00 9.12208557e-02
-3.35108899e-02 -6.69763744e-01 -3.14085126e-01 -1.07625461e+00
2.61309952e-01 -1.37826228e+00 4.43043828e-01 -8.96443307e-01
-1.11347139e+00 2.09434498e-02 2.76095808e-01 -8.74425411e-01
-7.51221955e-01 -3.77140403e-01 -3.88252795e-01 4.85026002e-01
-1.15762091e+00 -1.42592359e+00 -7.04512298e-01 4.95567620e-01
9.75035906e-01 -9.56101343e-02 6.57909811e-01 -4.54278104e-02
-5.22207558e-01 4.71884124e-02 -6.09968722e-01 2.66958475e-01
5.17029107e-01 -1.11455071e+00 5.86912394e-01 8.44585538e-01
2.29257047e-01 3.94854963e-01 7.50991404e-01 -7.08428919e-01
-1.54744232e+00 -8.92343223e-01 7.51134217e-01 -1.25477874e+00
8.78145933e-01 -3.98425907e-01 -5.03890455e-01 1.56064689e+00
2.20270976e-01 -2.42369056e-01 4.54243392e-01 3.27286243e-01
-4.21538860e-01 2.53137439e-01 -4.72468585e-01 8.41553688e-01
1.77300417e+00 -6.11781001e-01 -8.90447259e-01 6.65261865e-01
9.31307852e-01 -4.02426749e-01 -4.57631975e-01 2.16741711e-01
6.79615617e-01 -1.07433426e+00 1.14686418e+00 -1.04994011e+00
8.07047248e-01 -2.47102588e-01 -2.71386266e-01 -1.02724230e+00
-4.16275769e-01 -5.59271336e-01 -3.79137158e-01 7.98909426e-01
1.47846356e-01 -3.15755576e-01 7.65261889e-01 9.87084389e-01
-2.26635590e-01 -4.17845994e-01 -4.35023218e-01 -7.62115419e-01
-5.91680527e-01 -4.58905458e-01 6.31796360e-01 8.35668981e-01
1.16632216e-01 3.27467263e-01 -5.45710683e-01 1.08456865e-01
6.36560559e-01 2.88433760e-01 1.04431522e+00 -9.48791683e-01
-6.55467093e-01 -4.76088345e-01 -1.62599325e-01 -1.22353411e+00
3.11589688e-01 -6.52679384e-01 1.48201913e-01 -1.64009190e+00
4.86717850e-01 -4.40417118e-02 -3.11715547e-02 9.38835442e-01
-6.97760135e-02 -1.55592561e-01 2.17279688e-01 3.36282849e-01
-1.36206424e+00 3.95091742e-01 1.38348973e+00 -2.31265903e-01
-1.57739520e-01 4.97652702e-02 -8.56804430e-01 7.06474781e-01
4.68756735e-01 -1.96786240e-01 -6.21789336e-01 -8.27165723e-01
2.41894394e-01 1.65545374e-01 7.23144114e-01 -1.08058679e+00
2.97850221e-01 -9.83296752e-01 6.31813053e-03 -1.61948845e-01
5.41720390e-01 -8.23048472e-01 9.99648795e-02 4.47628289e-01
-7.69713223e-01 3.20633620e-01 -2.43196636e-01 7.43214488e-01
2.88663786e-02 5.42539120e-01 2.49462962e-01 -5.85129380e-01
-1.19192851e+00 3.04386824e-01 -2.77773440e-01 2.94429094e-01
1.41548038e+00 -1.19686313e-01 -5.91139376e-01 -5.76713502e-01
-6.90371990e-01 7.12760985e-01 3.78995508e-01 5.58353186e-01
5.91728911e-02 -1.13500130e+00 -6.67160034e-01 -1.31052807e-01
3.04934680e-01 -1.56554952e-02 2.42582887e-01 9.15531397e-01
-1.57502159e-01 4.68906015e-01 -4.27180648e-01 -4.27538097e-01
-1.01188171e+00 5.33243895e-01 3.04221451e-01 -4.66107786e-01
-5.59909165e-01 5.65641701e-01 3.66680533e-01 -3.00920725e-01
5.15364885e-01 -3.77114773e-01 1.02352202e-01 -7.25385919e-02
5.79201281e-01 5.56879699e-01 -5.41036010e-01 -3.50017935e-01
-4.68542546e-01 -8.42929482e-02 1.79419711e-01 -7.35679343e-02
1.25944257e+00 -1.07265897e-01 -1.18181957e-02 4.76763755e-01
5.86629748e-01 -5.83226264e-01 -1.71784008e+00 -3.00678849e-01
2.50253171e-01 -2.52403021e-01 -5.08911133e-01 -1.13920689e+00
-3.82265627e-01 5.91657162e-01 3.27582035e-04 2.13559195e-01
7.52562642e-01 3.35299850e-01 6.91228390e-01 5.85567772e-01
6.33242548e-01 -9.73942578e-01 7.07606673e-01 5.49624085e-01
8.62444639e-01 -1.46288431e+00 -6.09433390e-02 -1.42519653e-01
-1.05571258e+00 5.18083096e-01 1.04919922e+00 2.20446587e-01
6.94439262e-02 3.37459803e-01 -7.44404122e-02 -3.84824783e-01
-1.20746660e+00 -4.21838373e-01 3.52920294e-02 5.26676059e-01
3.29462707e-01 3.41302216e-01 8.48117992e-02 4.50751245e-01
2.59990226e-02 2.47307837e-01 3.34357470e-01 1.14044404e+00
-1.63972870e-01 -7.58198798e-01 -3.71202305e-02 4.19240952e-01
-8.71331245e-02 2.41298497e-01 -4.56726372e-01 5.72201967e-01
-3.54733430e-02 1.00634861e+00 3.04724902e-01 -2.38320872e-01
3.30887169e-01 4.23261702e-01 6.19125843e-01 -6.63268209e-01
-3.54828954e-01 -4.81215209e-01 2.83064604e-01 -1.12326288e+00
-7.08796680e-01 -9.32998061e-01 -1.32312357e+00 -2.19011515e-01
5.03029227e-01 -4.61046934e-01 2.74027765e-01 1.40355611e+00
4.70413238e-01 7.35391140e-01 7.67420828e-02 -9.32629287e-01
-7.75687456e-01 -9.64277625e-01 -2.38657460e-01 1.01374042e+00
3.95758808e-01 -7.02025890e-01 -3.62511575e-01 3.75128925e-01] | [8.250495910644531, 0.5404900312423706] |
47ca57a2-0446-4b19-b8f2-830ec1bd37d1 | speak-foreign-languages-with-your-own-voice | 2303.03926 | null | https://arxiv.org/abs/2303.03926v1 | https://arxiv.org/pdf/2303.03926v1.pdf | Speak Foreign Languages with Your Own Voice: Cross-Lingual Neural Codec Language Modeling | We propose a cross-lingual neural codec language model, VALL-E X, for cross-lingual speech synthesis. Specifically, we extend VALL-E and train a multi-lingual conditional codec language model to predict the acoustic token sequences of the target language speech by using both the source language speech and the target language text as prompts. VALL-E X inherits strong in-context learning capabilities and can be applied for zero-shot cross-lingual text-to-speech synthesis and zero-shot speech-to-speech translation tasks. Experimental results show that it can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker's voice, emotion, and acoustic environment. Moreover, VALL-E X effectively alleviates the foreign accent problems, which can be controlled by a language ID. Audio samples are available at \url{https://aka.ms/vallex}. | ['Furu Wei', 'Sheng Zhao', 'Lei He', 'Jinyu Li', 'Huaming Wang', 'Yanqing Liu', 'Zhuo Chen', 'Shujie Liu', 'Yu Wu', 'Sanyuan Chen', 'Chengyi Wang', 'Long Zhou', 'Ziqiang Zhang'] | 2023-03-07 | null | null | null | null | ['text-to-speech-synthesis', 'speech-to-speech-translation', 'speech-synthesis'] | ['speech', 'speech', 'speech'] | [-1.77136019e-01 1.47136897e-01 -2.49855787e-01 -5.57257533e-01
-1.47482252e+00 -4.46497560e-01 3.33619416e-01 -3.76147270e-01
-4.86535877e-02 5.07429004e-01 3.97092402e-01 -6.33629620e-01
6.34902775e-01 -3.70266706e-01 -8.09728265e-01 -5.60114086e-01
4.25296098e-01 1.81558922e-01 -1.07126452e-01 -2.53680199e-01
-4.35650855e-01 -8.95477384e-02 -1.22417474e+00 4.56744373e-01
7.94469833e-01 7.59222269e-01 6.68031991e-01 7.78469145e-01
-2.72589564e-01 7.96014428e-01 -5.34790814e-01 -4.30813938e-01
-1.40403628e-01 -6.24670029e-01 -5.41075230e-01 -7.54964128e-02
3.61424196e-03 -1.99725807e-01 -1.79666713e-01 1.05538023e+00
6.70548737e-01 1.42751068e-01 3.84697974e-01 -1.06252635e+00
-8.48179579e-01 8.80328119e-01 2.34937435e-03 -1.45825893e-01
3.97390068e-01 1.76428363e-01 1.04688704e+00 -1.23484695e+00
4.21360523e-01 1.42153740e+00 2.60236651e-01 9.59090233e-01
-1.14946258e+00 -8.82429600e-01 4.91940267e-02 -1.25174224e-01
-1.49953938e+00 -1.18067038e+00 8.88811529e-01 -1.50367722e-01
1.09900737e+00 2.17084423e-01 1.37387067e-01 1.69515646e+00
9.49305668e-02 9.86070216e-01 7.93469906e-01 -7.58968532e-01
3.84836495e-01 1.18317984e-01 -4.27866846e-01 5.65230310e-01
-8.24768364e-01 3.03369224e-01 -8.70999038e-01 6.47893995e-02
2.40606442e-01 -4.69855517e-01 -3.71835768e-01 2.73788631e-01
-1.22560108e+00 8.28576028e-01 -9.81594920e-02 4.24742669e-01
-1.51161745e-01 7.48712867e-02 6.99263632e-01 3.59324962e-01
6.02477014e-01 1.14085473e-01 -5.99180698e-01 -3.32971811e-01
-6.76611364e-01 -1.45120934e-01 5.53524673e-01 1.39247155e+00
4.67182338e-01 8.99958253e-01 -1.61772832e-01 1.27006328e+00
2.69850373e-01 1.12596440e+00 7.51354218e-01 -8.19265723e-01
5.96745968e-01 -2.42421761e-01 -2.14620024e-01 -1.62959620e-01
2.25070104e-01 -1.28367782e-01 -5.26988089e-01 -2.01229215e-01
-2.36461192e-01 -5.52373350e-01 -7.60271430e-01 1.99990809e+00
7.47210383e-02 2.52172470e-01 7.07550049e-01 5.41752338e-01
7.49675930e-01 1.29727995e+00 -7.56039936e-03 -4.47701603e-01
1.17136037e+00 -1.23870850e+00 -1.06153941e+00 -4.96657729e-01
6.34018183e-01 -1.03920233e+00 1.57779181e+00 -3.68580967e-02
-1.18803155e+00 -7.24515259e-01 -8.90018880e-01 -1.09292656e-01
-2.64644146e-01 4.70858991e-01 -4.78293374e-02 5.31638265e-01
-1.07524788e+00 -1.89430282e-01 -7.01786757e-01 -8.11336860e-02
-1.33278772e-01 -1.39049798e-01 -5.44592142e-02 -7.40619898e-02
-1.53798020e+00 4.36925203e-01 2.83989787e-01 -2.80719072e-01
-1.14500153e+00 -5.01379907e-01 -1.30600071e+00 1.18817233e-01
2.63859093e-01 9.43641886e-02 1.77930081e+00 -1.00825262e+00
-2.06252527e+00 5.42309284e-01 -6.00406945e-01 -1.33462816e-01
-5.11966981e-02 -7.12514594e-02 -1.11769915e+00 -6.48669377e-02
3.53466541e-01 6.75485253e-01 9.38838542e-01 -1.10476696e+00
-6.90613866e-01 4.89243269e-02 -6.06080592e-01 1.70899302e-01
-1.94324404e-01 4.90882605e-01 -5.32646894e-01 -9.87193227e-01
-3.69124979e-01 -9.99720156e-01 1.28662303e-01 -4.32349443e-01
-4.15485173e-01 -3.77263188e-01 9.01940763e-01 -7.58577526e-01
1.12596548e+00 -2.48154306e+00 -9.54960436e-02 -2.09732905e-01
-6.74965560e-01 4.31676626e-01 -3.68880332e-01 4.09297317e-01
1.58372615e-02 -8.07160810e-02 -1.46536320e-01 -7.47725308e-01
1.42039880e-01 3.40153784e-01 -8.03231359e-01 -6.37670606e-02
2.89485395e-01 9.91085589e-01 -8.19556534e-01 -2.75790513e-01
1.67059913e-01 7.36123800e-01 -3.74899983e-01 6.41620457e-01
-4.08826619e-01 4.87401962e-01 -2.21544951e-01 4.70507681e-01
9.54003185e-02 2.84533203e-01 1.53723687e-01 2.73398966e-01
-2.67654240e-01 9.17181551e-01 -5.34128964e-01 1.91698408e+00
-9.61767554e-01 4.57092255e-01 2.47375786e-01 -6.37958705e-01
1.22617126e+00 1.05174756e+00 2.14454811e-02 -9.68138754e-01
1.41872212e-01 5.97595394e-01 -1.74209327e-01 -3.18453640e-01
3.32028419e-01 -2.86750704e-01 -4.23366994e-01 4.72887546e-01
4.18926835e-01 -4.28513467e-01 -2.53326833e-01 -3.12794112e-02
4.22411174e-01 6.30771071e-02 1.91718251e-01 -2.09556207e-01
4.96046841e-01 -5.05239010e-01 5.99346340e-01 4.19125199e-01
-1.28522336e-01 4.83531773e-01 4.22426611e-02 1.80906475e-01
-1.05108583e+00 -1.24043190e+00 7.17885271e-02 1.50185561e+00
-3.11845452e-01 -5.32127917e-01 -1.00911093e+00 -3.56253833e-01
-5.17081738e-01 1.28035128e+00 3.00537143e-02 -2.90155262e-01
-6.29804194e-01 8.60290229e-02 9.48762715e-01 2.62518287e-01
-3.34538817e-02 -1.21577823e+00 1.22134700e-01 4.76663798e-01
-6.71367645e-01 -1.36728311e+00 -1.21666515e+00 2.77445018e-01
-1.77959383e-01 -2.02761680e-01 -7.38869965e-01 -1.28021920e+00
1.41551748e-01 4.44379114e-02 7.65582085e-01 -4.44222212e-01
1.94445252e-01 2.18960434e-01 -5.26602805e-01 -3.41616184e-01
-1.16536260e+00 1.34743191e-02 4.90310520e-01 2.73899525e-01
4.41179246e-01 -4.04320866e-01 1.34934425e-01 1.25084504e-01
-5.02368748e-01 2.01427221e-01 1.18942305e-01 7.35465348e-01
6.95335209e-01 -2.26646841e-01 1.08483124e+00 -3.88669312e-01
6.86167300e-01 -3.78134966e-01 -5.39023697e-01 2.16295585e-01
-3.21442395e-01 2.07673945e-02 1.12917149e+00 -4.95730549e-01
-1.30838239e+00 -2.78305840e-02 -6.18655384e-01 -6.51885688e-01
-2.40551546e-01 4.25073564e-01 -6.19368076e-01 5.36559880e-01
3.63794029e-01 6.09835982e-01 -2.35351682e-01 -4.79497224e-01
5.24086654e-01 1.33811736e+00 8.86940658e-01 -8.29083979e-01
4.34439659e-01 -2.10101604e-01 -8.51034403e-01 -9.93533552e-01
-7.39105403e-01 -1.97106421e-01 -3.46884280e-01 -2.37947240e-01
1.05279028e+00 -1.32572067e+00 -4.15201724e-01 5.54957569e-01
-1.47564447e+00 -6.15347266e-01 -1.32555485e-01 6.28287077e-01
-7.76022851e-01 4.30074101e-03 -8.01317930e-01 -8.80373776e-01
-3.23615789e-01 -1.47182381e+00 1.28063345e+00 -1.27375975e-01
-6.90621808e-02 -9.26146328e-01 -3.19933482e-02 2.41789639e-01
4.38413352e-01 -5.59249699e-01 1.06473076e+00 -6.24305487e-01
-4.29791987e-01 2.07784429e-01 3.52920681e-01 6.86072409e-01
3.44372243e-01 -1.43496826e-01 -1.19580925e+00 -3.74445736e-01
1.24383688e-01 -5.44741154e-01 1.08248107e-01 1.59599572e-01
8.54270875e-01 -5.63706577e-01 1.21382274e-01 5.96584916e-01
1.04754293e+00 6.11497521e-01 2.41439193e-01 -4.85206991e-01
7.16227353e-01 5.54801166e-01 4.42174613e-01 2.90290833e-01
4.41300571e-01 9.22879100e-01 -1.37249365e-01 -5.48402928e-02
-4.67434198e-01 -6.48177803e-01 9.90554750e-01 1.96946859e+00
7.02245414e-01 -3.90887469e-01 -8.51112425e-01 7.57755280e-01
-1.43821740e+00 -6.87088966e-01 1.25365034e-01 1.99855709e+00
1.30545056e+00 -5.73672391e-02 -2.85808474e-01 -2.52985567e-01
8.21275890e-01 3.05043966e-01 -4.10664380e-01 -6.92557335e-01
-1.37339801e-01 2.20526487e-01 7.59070739e-02 1.00842118e+00
-8.49009097e-01 1.47780430e+00 5.93910503e+00 1.21597672e+00
-1.60753703e+00 4.93387014e-01 4.66675103e-01 -1.94058314e-01
-5.41409552e-01 -1.03994951e-01 -9.05436873e-01 6.82609141e-01
1.65512943e+00 -4.43079501e-01 7.01036274e-01 8.99162471e-01
4.75594968e-01 5.08580804e-01 -1.11572194e+00 1.03205979e+00
1.77528009e-01 -1.13916719e+00 -8.58054981e-02 -3.28470170e-01
6.18720114e-01 3.62193972e-01 5.02679199e-02 7.17471719e-01
4.92788315e-01 -8.20373893e-01 1.20732641e+00 3.50481011e-02
1.53665268e+00 -8.67950499e-01 2.30304763e-01 5.82668185e-01
-1.41314042e+00 2.15960935e-01 2.33396851e-02 2.43939817e-01
5.28429210e-01 5.87180555e-02 -8.05708408e-01 4.33348566e-01
5.02904475e-01 4.88142312e-01 -5.50925396e-02 1.02298625e-01
-4.86864060e-01 1.22621107e+00 2.67842393e-02 -2.40824353e-02
3.45070064e-01 -1.97939575e-03 6.97986424e-01 1.46560574e+00
6.39791250e-01 -8.44932795e-02 5.63335121e-01 7.31757998e-01
-2.06696570e-01 4.37294871e-01 -6.14371240e-01 -1.15240425e-01
9.06917453e-01 5.69432139e-01 4.87767532e-03 -4.24568623e-01
-5.88026285e-01 1.16748905e+00 3.28862607e-01 7.43862212e-01
-7.41014659e-01 -6.29664898e-01 8.63815844e-01 -3.25499505e-01
4.43874180e-01 -2.99511880e-01 6.38661534e-02 -1.15422142e+00
1.06810883e-01 -1.21595061e+00 -1.98200732e-01 -1.17690849e+00
-1.19416916e+00 1.10347784e+00 -3.47729921e-01 -1.19616854e+00
-8.35109293e-01 -3.88828933e-01 -4.85717177e-01 1.17124259e+00
-1.45654535e+00 -1.33633828e+00 5.38797736e-01 7.65497983e-01
1.14062822e+00 -7.43052542e-01 1.40196836e+00 3.75028700e-01
-5.00366747e-01 8.18932235e-01 2.54190415e-01 3.94972324e-01
8.21370840e-01 -9.05675828e-01 6.99484229e-01 1.13281989e+00
1.55517340e-01 2.84306616e-01 4.91647631e-01 -5.25359631e-01
-1.17827642e+00 -1.48747993e+00 1.45211613e+00 -2.49139480e-02
8.09517980e-01 -8.79176557e-01 -9.82038200e-01 8.25424850e-01
4.97673601e-01 3.26674171e-02 8.06347847e-01 -1.65385172e-01
-4.86309081e-01 -1.89917997e-01 -6.74184084e-01 8.30734551e-01
5.85936725e-01 -1.45002306e+00 -4.17878240e-01 1.29906625e-01
1.66633844e+00 -3.58203620e-01 -6.40080035e-01 -1.62248896e-03
1.02650553e-01 -4.70765352e-01 6.48831248e-01 -3.80268842e-01
2.00082645e-01 -1.90608546e-01 -7.17751861e-01 -1.55596471e+00
7.71875307e-02 -9.98763382e-01 1.56601101e-01 1.61384428e+00
6.39141738e-01 -5.22841096e-01 -6.85915947e-02 2.04482391e-01
-7.34645426e-01 -4.13649827e-01 -1.23418438e+00 -1.03451312e+00
3.03296059e-01 -8.75275254e-01 7.81825244e-01 9.19403553e-01
9.47795734e-02 7.56733716e-01 -7.40404010e-01 3.72447282e-01
3.72774631e-01 -3.24988998e-02 4.79153931e-01 -4.85449731e-01
-4.08863038e-01 -1.18741959e-01 2.61121511e-01 -1.23779964e+00
8.21970344e-01 -1.27753472e+00 6.84324980e-01 -9.54998612e-01
-5.54110646e-01 -4.46695417e-01 -1.34080648e-01 7.31613696e-01
1.30113019e-02 -1.20362885e-01 2.73354203e-01 -3.75826992e-02
-2.11350694e-01 1.05879796e+00 1.02632093e+00 -9.68583599e-02
-1.82392105e-01 3.22823748e-02 -4.04369950e-01 3.84778261e-01
8.14211071e-01 -5.92434585e-01 -5.18577516e-01 -6.47370458e-01
-4.60087061e-01 6.21046007e-01 -1.64675713e-01 -7.60391891e-01
7.33519047e-02 -3.04081410e-01 -4.02041078e-01 -3.70338023e-01
6.24239087e-01 -4.73566979e-01 -9.53937098e-02 8.17400515e-02
-6.00501776e-01 -1.23518631e-01 2.98918307e-01 2.00214326e-01
-5.36721528e-01 -2.08212301e-01 9.73945677e-01 9.68813226e-02
-5.45711696e-01 2.75879800e-01 -7.26669669e-01 1.99362889e-01
6.82273686e-01 3.90341163e-01 -2.18249097e-01 -5.93302846e-01
-5.91649652e-01 1.95443481e-02 2.03278229e-01 1.00166285e+00
5.67652106e-01 -1.63284361e+00 -9.07682955e-01 7.44536519e-01
3.43886644e-01 -3.13116640e-01 9.67355520e-02 3.29235524e-01
1.12876989e-01 6.73868895e-01 2.59777963e-01 -5.38184106e-01
-1.16423571e+00 5.50874293e-01 3.64613980e-01 3.84681970e-01
-4.53252494e-01 8.50842834e-01 2.90323853e-01 -8.12585413e-01
3.76109272e-01 -2.02154532e-01 4.01719868e-01 -4.93414432e-01
6.87510550e-01 -2.42726371e-01 -7.05179339e-03 -1.07319772e+00
-3.20896745e-01 1.69028640e-01 -2.66670287e-02 -6.26018405e-01
1.00828695e+00 -5.55402935e-01 1.08225666e-01 9.24437106e-01
1.40749419e+00 5.01586318e-01 -1.19695008e+00 -5.00445366e-01
-1.06452323e-01 -1.13185056e-01 1.24065906e-01 -7.47851729e-01
-8.01977575e-01 1.11163902e+00 2.63067663e-01 -1.34315655e-01
8.65820944e-01 2.77414978e-01 1.15932095e+00 3.43090117e-01
3.31039906e-01 -1.28816879e+00 5.79589680e-02 1.01608741e+00
1.02682745e+00 -1.15126932e+00 -9.67645824e-01 -7.84129556e-03
-1.05178273e+00 8.21301818e-01 4.76245344e-01 3.86879921e-01
7.32293785e-01 6.41662061e-01 7.84535825e-01 3.48822296e-01
-1.10356808e+00 -7.39688203e-02 4.27294374e-02 6.71595633e-01
6.06148660e-01 5.52208900e-01 3.39834809e-01 6.67254210e-01
-6.56341732e-01 -3.66174191e-01 3.71372461e-01 5.82746983e-01
-3.87473494e-01 -1.30728996e+00 -3.29518855e-01 -2.93081164e-01
-4.82995242e-01 -5.55645704e-01 -9.20789018e-02 8.21191445e-02
-4.35769670e-02 1.30438197e+00 5.99628128e-02 -3.79684865e-01
2.24588856e-01 7.19874740e-01 -1.00216009e-01 -7.38748133e-01
-2.44674891e-01 6.16338670e-01 1.72071040e-01 -4.14669693e-01
5.38380668e-02 -6.16236210e-01 -1.44907892e+00 -4.48898077e-02
-1.69644639e-01 2.54547954e-01 8.82555366e-01 9.59269702e-01
4.24767822e-01 6.14602208e-01 1.06262267e+00 -6.14713430e-01
-4.81053770e-01 -9.84519422e-01 -5.63228071e-01 -9.70806368e-03
7.52636969e-01 -2.03033742e-02 -3.44387084e-01 4.86373305e-01] | [14.685746192932129, 6.883449077606201] |
4324f686-f9c6-4c63-835f-08c57956c2f3 | metaiqa-deep-meta-learning-for-no-reference | 2004.05508 | null | https://arxiv.org/abs/2004.05508v1 | https://arxiv.org/pdf/2004.05508v1.pdf | MetaIQA: Deep Meta-learning for No-Reference Image Quality Assessment | Recently, increasing interest has been drawn in exploiting deep convolutional neural networks (DCNNs) for no-reference image quality assessment (NR-IQA). Despite of the notable success achieved, there is a broad consensus that training DCNNs heavily relies on massive annotated data. Unfortunately, IQA is a typical small sample problem. Therefore, most of the existing DCNN-based IQA metrics operate based on pre-trained networks. However, these pre-trained networks are not designed for IQA task, leading to generalization problem when evaluating different types of distortions. With this motivation, this paper presents a no-reference IQA metric based on deep meta-learning. The underlying idea is to learn the meta-knowledge shared by human when evaluating the quality of images with various distortions, which can then be adapted to unknown distortions easily. Specifically, we first collect a number of NR-IQA tasks for different distortions. Then meta-learning is adopted to learn the prior knowledge shared by diversified distortions. Finally, the quality prior model is fine-tuned on a target NR-IQA task for quickly obtaining the quality model. Extensive experiments demonstrate that the proposed metric outperforms the state-of-the-arts by a large margin. Furthermore, the meta-model learned from synthetic distortions can also be easily generalized to authentic distortions, which is highly desired in real-world applications of IQA metrics. | ['Jinjian Wu', 'Hancheng Zhu', 'Leida Li', 'Guangming Shi', 'Weisheng Dong'] | 2020-04-11 | metaiqa-deep-meta-learning-for-no-reference-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Zhu_MetaIQA_Deep_Meta-Learning_for_No-Reference_Image_Quality_Assessment_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zhu_MetaIQA_Deep_Meta-Learning_for_No-Reference_Image_Quality_Assessment_CVPR_2020_paper.pdf | cvpr-2020-6 | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 2.22472981e-01 -4.16835546e-01 -1.46785393e-01 -4.61857110e-01
-1.02822590e+00 -2.51962483e-01 4.46334183e-01 -2.32477352e-01
-2.21288949e-01 4.14652795e-01 1.75231352e-01 -5.94065785e-02
-3.28053087e-01 -7.91762710e-01 -6.49403870e-01 -6.16873264e-01
-7.29916198e-03 -4.90227118e-02 -1.62921533e-01 -3.00165445e-01
2.63695329e-01 4.27834898e-01 -1.41149473e+00 1.91162467e-01
1.21149480e+00 1.23822212e+00 1.70055762e-01 5.40777445e-01
3.10305655e-01 5.46743095e-01 -8.69497478e-01 -7.07920015e-01
4.77798671e-01 -5.43095648e-01 -7.08051085e-01 1.08492665e-01
5.21627724e-01 -5.51836312e-01 -6.14182293e-01 1.29344630e+00
7.65847504e-01 3.03182900e-01 6.33862257e-01 -1.32516873e+00
-9.42180634e-01 1.52431011e-01 -3.20772797e-01 3.33155274e-01
2.21096903e-01 5.44413090e-01 1.04060006e+00 -8.63242626e-01
1.91122264e-01 1.30993390e+00 4.64408159e-01 4.69840825e-01
-7.90574431e-01 -6.97056115e-01 -2.20893454e-02 6.01669848e-01
-1.32434809e+00 -2.85557151e-01 7.65059531e-01 -2.32010886e-01
6.60897970e-01 4.51577157e-02 3.74713331e-01 1.07630301e+00
1.46338433e-01 7.49910295e-01 1.23707092e+00 -1.49490193e-01
1.84550986e-01 -8.04458335e-02 -3.50252658e-01 3.86269927e-01
6.85372874e-02 3.82436275e-01 -3.28566879e-01 2.76234031e-01
7.79701889e-01 -1.03398465e-01 -4.48127717e-01 -1.96789071e-01
-1.26449203e+00 4.25686896e-01 7.25645602e-01 1.48696497e-01
-2.47962087e-01 -2.36257352e-02 5.93997836e-01 5.61404586e-01
2.79843599e-01 5.61806083e-01 -3.14353794e-01 -1.89823925e-01
-9.72778559e-01 6.97988793e-02 3.31254601e-01 8.26155722e-01
7.35946536e-01 1.83168262e-01 -3.60656142e-01 1.08351636e+00
1.32312194e-01 6.38079762e-01 5.51987410e-01 -1.13345373e+00
6.27195299e-01 5.30076265e-01 2.03339115e-01 -1.27045667e+00
-1.92315251e-01 -7.97398746e-01 -1.29266155e+00 5.42547822e-01
4.36630100e-01 1.12336852e-01 -7.60505915e-01 1.62663543e+00
6.08405583e-02 1.10944420e-01 7.37151727e-02 1.09824789e+00
6.22853577e-01 7.15280116e-01 -1.49075404e-01 -1.79353669e-01
1.03846598e+00 -1.04911733e+00 -5.66578507e-01 6.91144019e-02
3.12047750e-01 -7.25154638e-01 1.33745563e+00 7.72390723e-01
-1.13577414e+00 -1.25102770e+00 -1.34545755e+00 1.79123189e-02
-2.29709312e-01 5.53363673e-02 2.48634309e-01 7.15111911e-01
-1.00785887e+00 7.67336369e-01 -3.83384973e-01 -7.34195206e-03
6.23952150e-01 2.00948000e-01 -1.23934664e-01 -4.61192489e-01
-1.35511506e+00 8.15881968e-01 2.40537614e-01 2.40655228e-01
-1.34445918e+00 -6.45591497e-01 -5.54015934e-01 -5.43623138e-03
4.12166089e-01 -7.25622952e-01 1.07986844e+00 -1.33253646e+00
-1.67061436e+00 6.58213496e-01 2.22215846e-01 -1.95293769e-01
6.27294302e-01 -1.89517528e-01 -9.07010257e-01 1.98405400e-01
-6.17566667e-02 5.11593640e-01 9.77714598e-01 -1.34866130e+00
-5.60660839e-01 -8.31492692e-02 5.48283279e-01 2.54579544e-01
-4.00834382e-01 -7.55833387e-02 -6.26394868e-01 -8.70756149e-01
-2.13621423e-01 -7.61923552e-01 -3.85889038e-02 1.96239129e-01
-3.13202560e-01 -9.78097171e-02 4.82306063e-01 -6.24526024e-01
1.25595403e+00 -2.08517146e+00 7.96934441e-02 1.82142124e-01
2.14331850e-01 5.68834364e-01 -6.20689452e-01 2.54568577e-01
-9.56316590e-02 1.23641521e-01 -1.92342862e-01 2.20213421e-02
1.02842882e-01 -5.93736991e-02 -1.10512309e-01 3.11440527e-01
3.19787830e-01 8.32473278e-01 -1.07005417e+00 -3.11139315e-01
3.27817559e-01 4.56548035e-01 -4.22242075e-01 5.49452484e-01
1.26451515e-02 6.02285147e-01 -2.09023684e-01 7.12705910e-01
8.60103190e-01 -3.14473242e-01 -2.82447308e-01 -7.09670186e-01
1.23809516e-01 -5.88001497e-02 -1.15092480e+00 1.84932041e+00
-7.60104179e-01 4.81520146e-01 -3.31700563e-01 -1.15901291e+00
8.70803416e-01 3.03430259e-01 4.59323049e-01 -1.22970486e+00
2.02513024e-01 3.47570777e-01 2.68710613e-01 -5.52586317e-01
3.05164248e-01 -5.35044558e-02 7.08851069e-02 3.39127630e-01
2.39951625e-01 -1.88057199e-02 2.84177840e-01 -4.71775234e-02
9.15987074e-01 4.99495976e-02 2.31592372e-01 8.91287401e-02
8.72069299e-01 -4.63458717e-01 6.26125813e-01 5.88612258e-01
-5.53837419e-01 8.99586380e-01 5.34230657e-02 -5.46380520e-01
-1.25762963e+00 -1.20400774e+00 -7.88544416e-02 9.03725803e-01
4.95856524e-01 -5.34981787e-02 -7.96447992e-01 -6.42362475e-01
-4.33762282e-01 3.97972375e-01 -5.23860633e-01 -3.68274510e-01
-4.16164815e-01 -7.13400781e-01 3.62990409e-01 4.54563200e-01
1.04542768e+00 -1.09719276e+00 -3.55776340e-01 1.85885891e-01
-2.60256827e-01 -1.07559526e+00 -5.16891420e-01 -3.93395394e-01
-7.16535747e-01 -1.18120992e+00 -1.08836997e+00 -4.83713001e-01
5.20511925e-01 4.93437737e-01 1.23587739e+00 2.80680716e-01
-5.50832786e-02 2.92146325e-01 -4.38232303e-01 -1.23183407e-01
-4.82665092e-01 -1.05481923e-01 1.01646110e-02 2.83666432e-01
1.39751688e-01 -6.80782437e-01 -1.09934866e+00 6.08440280e-01
-1.05841804e+00 2.53448393e-02 9.90874171e-01 8.10818672e-01
5.24650872e-01 3.95665050e-01 8.63014221e-01 -4.71480936e-01
6.07865393e-01 -3.97103071e-01 -3.93454224e-01 4.11354840e-01
-6.69156969e-01 -1.17061146e-01 8.36174786e-01 -5.07855356e-01
-1.11770749e+00 -5.94307184e-01 -2.21222252e-01 -5.02866030e-01
-1.65429652e-01 4.93527472e-01 -7.38990426e-01 -1.96043327e-01
6.72476292e-01 1.44730851e-01 -2.98350215e-01 -1.98811784e-01
4.49177623e-01 6.97512805e-01 7.82342851e-01 -6.46700919e-01
1.08784878e+00 3.18580866e-01 6.87474087e-02 -2.60528147e-01
-1.12890589e+00 -2.94642657e-01 -4.69475567e-01 -4.39443737e-01
6.52696252e-01 -9.28292990e-01 -4.52889591e-01 8.27646613e-01
-9.36272383e-01 -3.26058537e-01 -2.32503235e-01 4.29446250e-01
-6.86898291e-01 3.68452787e-01 -3.50810409e-01 -4.30429846e-01
-4.20170903e-01 -1.35370064e+00 7.68549740e-01 3.60059232e-01
2.31648967e-01 -8.78612161e-01 6.68366477e-02 4.98467326e-01
6.50170028e-01 1.35503292e-01 8.61524522e-01 -2.69954741e-01
-6.40294135e-01 -1.45454332e-01 -6.16813123e-01 8.84296715e-01
2.90679008e-01 -1.50319070e-01 -9.74587142e-01 -4.73647952e-01
-8.11463222e-02 -6.23252213e-01 4.96924132e-01 2.33163550e-01
1.53552377e+00 -2.23406404e-01 3.51521283e-01 8.21840644e-01
1.39763522e+00 2.13058680e-01 9.64538872e-01 5.55042803e-01
5.96366346e-01 2.22554803e-01 7.69946873e-01 2.74748087e-01
3.55195612e-01 7.73584127e-01 6.41049147e-01 -1.86274558e-01
-3.31388324e-01 -5.66322803e-02 2.91772693e-01 1.00808942e+00
-2.42730990e-01 -3.93827081e-01 -7.22415984e-01 6.42404735e-01
-1.49457908e+00 -9.29277420e-01 4.12695259e-01 2.13301182e+00
8.55296791e-01 1.82229832e-01 7.39218518e-02 5.20729184e-01
7.38207519e-01 1.99017793e-01 -9.03136015e-01 -2.49529511e-01
-1.58025146e-01 1.84922293e-01 1.18022017e-01 -3.45835090e-03
-1.08365655e+00 4.49046791e-01 5.62954712e+00 1.08794796e+00
-1.11379743e+00 2.11436778e-01 8.20265770e-01 2.22749077e-03
-1.78471133e-01 -2.55661339e-01 -2.09897667e-01 5.28855264e-01
9.77061331e-01 -4.16493028e-01 5.76448798e-01 6.32820368e-01
2.38298342e-01 2.56923079e-01 -1.05785942e+00 1.17383945e+00
5.64815141e-02 -8.97883773e-01 3.98023635e-01 -8.60304236e-02
1.22142041e+00 -1.46874368e-01 5.78544259e-01 4.76167053e-01
6.41592592e-02 -1.10763621e+00 5.84279418e-01 6.66971505e-01
1.16751683e+00 -9.75378692e-01 9.25910354e-01 1.51224062e-01
-9.60331023e-01 -2.11112276e-01 -7.18691051e-01 1.32768661e-01
8.55921060e-02 8.21814001e-01 -2.53518611e-01 8.91686916e-01
6.80403948e-01 7.30802119e-01 -7.32956231e-01 1.31489015e+00
-1.91745773e-01 4.58175719e-01 2.81905860e-01 4.16479707e-01
2.38543779e-01 -6.71836510e-02 2.56006688e-01 9.79218364e-01
6.54842198e-01 1.20296113e-01 -7.22235665e-02 7.35384583e-01
-4.67241585e-01 1.14351012e-01 -2.40678445e-01 1.65722847e-01
2.94617206e-01 1.25793743e+00 -2.04031199e-01 -3.46783370e-01
-6.50496900e-01 1.04106390e+00 8.59518200e-02 4.74879086e-01
-8.65921319e-01 -5.62559068e-01 5.66534281e-01 -6.49030060e-02
1.24629021e-01 3.03489319e-03 -5.62327318e-02 -1.15342629e+00
1.34531921e-02 -1.34063518e+00 1.52232572e-01 -9.85412598e-01
-1.56432068e+00 6.82490826e-01 -6.75478727e-02 -1.78550541e+00
-1.71265066e-01 -5.20854175e-01 -7.08904386e-01 8.85281384e-01
-1.93628883e+00 -1.07196629e+00 -7.86611736e-01 6.55362248e-01
6.53230429e-01 -2.35190049e-01 5.23737907e-01 5.09013295e-01
-5.13429165e-01 1.00670183e+00 8.27748179e-02 3.04792523e-01
8.65003049e-01 -1.07983792e+00 4.16505963e-01 1.06035423e+00
3.50890718e-02 4.40518707e-01 4.92576897e-01 -2.77451575e-01
-1.22636914e+00 -1.40158880e+00 2.62799710e-01 -1.65176153e-01
4.76740599e-01 1.72416002e-01 -1.14146936e+00 4.54749167e-02
3.30755681e-01 4.36213344e-01 5.48880339e-01 -1.38104245e-01
-6.57262921e-01 -5.85643351e-01 -1.17518854e+00 4.76346493e-01
1.13555145e+00 -6.65820658e-01 -3.22806895e-01 1.13347687e-01
5.86492658e-01 -3.03779602e-01 -1.17401719e+00 6.51241899e-01
5.20035923e-01 -1.23314810e+00 1.10415161e+00 -2.90641487e-01
7.47851253e-01 -4.31030154e-01 -2.38062665e-01 -1.64985406e+00
-3.13387930e-01 -3.21843326e-01 -2.11834848e-01 1.13228035e+00
6.89119771e-02 -2.81586617e-01 3.76284271e-01 3.23915690e-01
-3.06835383e-01 -7.29498029e-01 -8.55892122e-01 -9.95498717e-01
1.53013140e-01 -4.74404097e-01 9.18991446e-01 8.74736369e-01
-4.45184231e-01 -3.82405496e-03 -5.37658036e-01 3.43075305e-01
8.43697786e-01 -1.61341280e-02 8.95294487e-01 -9.92100060e-01
-3.51675481e-01 -5.80920696e-01 -6.82909727e-01 -1.04870963e+00
-3.89806069e-02 -6.82150781e-01 -1.50076766e-02 -1.35208917e+00
2.97729224e-01 -4.73617166e-01 -8.12854767e-01 1.00543790e-01
-4.74150211e-01 3.96235138e-01 1.79537579e-01 3.18357259e-01
-9.16429043e-01 8.12064171e-01 1.71966660e+00 -4.64410186e-01
2.57856175e-02 -5.15776388e-02 -7.37368584e-01 5.29255927e-01
8.68101597e-01 -2.57419884e-01 -6.49999619e-01 -5.78773499e-01
3.99217099e-01 -2.34940741e-02 4.56323534e-01 -1.43789732e+00
1.20638087e-01 -2.86676794e-01 4.25950378e-01 -4.88042444e-01
9.21159387e-02 -7.10497200e-01 2.45919582e-02 1.48026362e-01
-3.88884544e-01 -8.96442533e-02 -9.85502265e-03 5.86014926e-01
-5.45513988e-01 -1.27224818e-01 9.98832583e-01 -1.76351994e-01
-7.82389402e-01 6.87071919e-01 2.55576074e-01 2.93724954e-01
6.25991464e-01 -2.76131183e-01 -2.67639995e-01 -5.18796921e-01
-3.71165782e-01 -1.75869241e-02 4.85283256e-01 4.88126993e-01
8.19337428e-01 -1.70324445e+00 -8.44994605e-01 -3.94738885e-03
4.23034251e-01 5.13696633e-02 6.17270470e-01 5.66235006e-01
-3.09243023e-01 1.09594338e-01 -5.33754349e-01 -5.52998900e-01
-7.46008694e-01 7.96442270e-01 5.13337612e-01 -2.80484617e-01
-3.57996047e-01 5.59388101e-01 2.41773650e-01 -2.05211028e-01
2.18562812e-01 -1.33350953e-01 -1.19291067e-01 -3.15604597e-01
8.45413029e-01 5.12058079e-01 1.75652415e-01 -7.06308305e-01
-2.97458097e-02 6.85980916e-01 8.50986466e-02 1.09492894e-02
1.24730003e+00 -3.72971237e-01 2.29923993e-01 3.02405834e-01
1.34260952e+00 -2.75026411e-01 -1.45361400e+00 -5.09739459e-01
-2.03245148e-01 -8.16743195e-01 1.11210331e-01 -1.07706976e+00
-1.54591334e+00 1.21693814e+00 9.76217747e-01 -1.04774579e-01
1.71527207e+00 -4.67614055e-01 8.79719794e-01 4.26385969e-01
5.19716740e-01 -1.09137321e+00 6.47605658e-01 1.58338964e-01
1.15659988e+00 -1.44771624e+00 2.59675886e-02 1.13620095e-01
-5.50265491e-01 1.08108008e+00 7.40971565e-01 -1.22202478e-01
4.57142204e-01 -3.01400006e-01 2.41480365e-01 1.39198661e-01
-4.74372685e-01 1.45439599e-02 6.08928144e-01 8.75959694e-01
2.35884592e-01 -1.40284494e-01 5.38307708e-03 5.10759413e-01
-1.56757608e-02 1.53705686e-01 5.69200933e-01 4.05121744e-01
-2.82333165e-01 -1.00534689e+00 -2.17836097e-01 4.14493144e-01
-4.24041957e-01 7.35719455e-03 1.73904851e-01 6.20525658e-01
3.51090133e-01 1.11185718e+00 -3.43501866e-01 -5.40651143e-01
5.60088873e-01 -5.15024245e-01 4.29173648e-01 -3.23348671e-01
-4.72645640e-01 -2.65207469e-01 -3.25560868e-01 -8.54324698e-01
-7.66882777e-01 -3.58774513e-01 -7.52074420e-01 -3.05049151e-01
-2.84346730e-01 -1.81810036e-01 4.67890710e-01 8.99221659e-01
1.35499537e-01 6.59670353e-01 1.09706461e+00 -8.31489146e-01
-5.95516264e-01 -9.98961627e-01 -4.23282266e-01 8.12977016e-01
3.34419489e-01 -6.67021990e-01 -4.37533349e-01 -6.07326673e-03] | [11.886719703674316, -1.8961541652679443] |
cdbbf1ad-56e7-468f-9849-a3b94c5e69b2 | composition-of-relational-features-with-an | 2206.00738 | null | https://arxiv.org/abs/2206.00738v2 | https://arxiv.org/pdf/2206.00738v2.pdf | Composition of Relational Features with an Application to Explaining Black-Box Predictors | Relational machine learning programs like those developed in Inductive Logic Programming (ILP) offer several advantages: (1) The ability to model complex relationships amongst data instances; (2) The use of domain-specific relations during model construction; and (3) The models constructed are human-readable, which is often one step closer to being human-understandable. However, these ILP-like methods have not been able to capitalise fully on the rapid hardware, software and algorithmic developments fuelling current developments in deep neural networks. In this paper, we treat relational features as functions and use the notion of generalised composition of functions to derive complex functions from simpler ones. We formulate the notion of a set of $\text{M}$-simple features in a mode language $\text{M}$ and identify two composition operators ($\rho_1$ and $\rho_2$) from which all possible complex features can be derived. We use these results to implement a form of "explainable neural network" called Compositional Relational Machines, or CRMs, which are labelled directed-acyclic graphs. The vertex-label for any vertex $j$ in the CRM contains a feature-function $f_j$ and a continuous activation function $g_j$. If $j$ is a "non-input" vertex, then $f_j$ is the composition of features associated with vertices in the direct predecessors of $j$. Our focus is on CRMs in which input vertices (those without any direct predecessors) all have $\text{M}$-simple features in their vertex-labels. We provide a randomised procedure for constructing and learning such CRMs. Using a notion of explanations based on the compositional structure of features in a CRM, we provide empirical evidence on synthetic data of the ability to identify appropriate explanations; and demonstrate the use of CRMs as 'explanation machines' for black-box models that do not provide explanations for their predictions. | ['Devanshu Shah', 'Tirtharaj Dash', 'A Baskar', 'Ashwin Srinivasan'] | 2022-06-01 | null | null | null | null | ['inductive-logic-programming'] | ['methodology'] | [ 3.96235466e-01 8.05075228e-01 -1.08438417e-01 -6.49057150e-01
-1.19224399e-01 -4.12091345e-01 6.62984788e-01 2.59211659e-01
6.73752874e-02 7.02712297e-01 -8.14620927e-02 -9.27449346e-01
-6.19757652e-01 -1.36452270e+00 -1.15891612e+00 -4.40256774e-01
-6.85923815e-01 5.88614643e-01 2.16130942e-01 -4.89861488e-01
7.91232288e-02 6.25871241e-01 -2.08581781e+00 7.10674345e-01
4.26649213e-01 9.02695894e-01 8.14739466e-02 6.02999866e-01
-4.26205367e-01 1.05392325e+00 -3.62370044e-01 -5.42112947e-01
1.67758539e-01 -5.09065688e-01 -1.13551748e+00 -2.31522471e-01
3.77599662e-03 1.71682954e-01 -2.11679354e-01 7.61644363e-01
-3.26785862e-01 6.22225106e-02 6.83890820e-01 -1.69784892e+00
-8.52592349e-01 1.07181489e+00 6.41431734e-02 -1.53989270e-01
3.20757151e-01 -1.18323296e-01 1.49412847e+00 -6.11673594e-01
5.89234412e-01 1.19320190e+00 7.12861121e-01 7.27436662e-01
-1.62928772e+00 -4.54176307e-01 2.05557808e-01 2.49007605e-02
-1.28575969e+00 -2.06199318e-01 5.22720933e-01 -4.28319693e-01
1.55884075e+00 6.84613049e-01 6.49159908e-01 3.44794929e-01
2.93141603e-01 4.28107768e-01 7.72996902e-01 -8.21269453e-01
1.08756289e-01 2.30801225e-01 3.73265952e-01 1.21800303e+00
1.01485506e-01 3.82576317e-01 -3.86500478e-01 -5.79173937e-02
8.76161814e-01 -6.96930885e-02 -2.62581371e-02 -3.66647869e-01
-8.57169390e-01 1.14092028e+00 7.92462051e-01 3.80013227e-01
-3.29107419e-02 6.68723822e-01 2.22663790e-01 6.16574585e-01
-6.04860894e-02 5.92239082e-01 -7.69145370e-01 3.53190899e-01
-9.54119191e-02 1.73836038e-01 9.55690920e-01 1.12313235e+00
1.18643475e+00 1.22829564e-02 3.89784366e-01 4.51328099e-01
2.72107065e-01 -9.44783259e-03 1.63017750e-01 -1.00868869e+00
1.44811049e-01 1.17078412e+00 -2.79985666e-01 -1.02598763e+00
-4.91531193e-01 -1.49519920e-01 -8.21352243e-01 2.89651722e-01
2.69470036e-01 2.75725991e-01 -8.23199272e-01 2.05658674e+00
1.89317372e-02 -2.40186349e-01 2.59012341e-01 3.68043512e-01
9.31107759e-01 7.45589256e-01 4.31042723e-02 -2.11817399e-01
1.12272203e+00 -5.57422698e-01 -7.66077712e-02 -2.08267093e-01
8.77913475e-01 -1.51824653e-01 1.03394306e+00 3.18547964e-01
-1.13444662e+00 -6.11535609e-01 -1.07087302e+00 4.08756174e-02
-7.32031643e-01 -4.41456676e-01 1.35923398e+00 4.49261695e-01
-1.33819127e+00 8.20283711e-01 -5.20569086e-01 -6.65287450e-02
9.63065699e-02 1.09233057e+00 -6.48889661e-01 1.22558653e-01
-1.23073852e+00 9.54001784e-01 4.93901402e-01 7.11004585e-02
-6.95143104e-01 -2.34550148e-01 -1.12207329e+00 2.39149943e-01
3.53053004e-01 -7.44125247e-01 1.00623119e+00 -1.50181699e+00
-7.45716512e-01 9.64549303e-01 -2.33766735e-01 -5.06638229e-01
-7.15109110e-02 5.26444674e-01 -5.35526454e-01 -1.34950608e-01
2.56765895e-02 9.06234384e-01 5.39698541e-01 -1.43855774e+00
-7.47322798e-01 -3.18208486e-01 6.03604376e-01 -3.95065658e-02
1.95163846e-01 2.23733075e-02 -2.30060667e-01 -1.32935748e-01
5.41098416e-01 -9.54262555e-01 -2.53552824e-01 -1.58209771e-01
-7.57274568e-01 -3.84187520e-01 3.60181093e-01 7.27863535e-02
1.14348698e+00 -1.97069979e+00 7.60091245e-02 7.57168293e-01
6.22946799e-01 -3.74541618e-02 9.91910100e-02 4.43726867e-01
-5.79315126e-01 5.10533988e-01 -3.09576780e-01 2.49324888e-01
1.68021008e-01 7.86273956e-01 -1.58888817e-01 9.01668444e-02
4.99092817e-01 8.43945384e-01 -6.75457656e-01 -2.65027344e-01
1.14457615e-01 3.98583382e-01 -7.37014949e-01 1.39801465e-02
-6.54825926e-01 -4.91892099e-02 -2.97441304e-01 2.37089038e-01
4.99700792e-02 -3.38337034e-01 3.65427732e-01 -7.08543957e-05
-1.63430378e-01 5.25816977e-01 -1.34725523e+00 8.17398489e-01
-2.83368558e-01 5.02916217e-01 -2.82521755e-01 -1.27091670e+00
1.13853288e+00 1.43064260e-01 1.03302315e-01 -4.90808874e-01
1.26584321e-01 2.55187958e-01 2.80311704e-01 -3.22714835e-01
2.43557230e-01 -5.26439905e-01 -1.91425726e-01 4.31280315e-01
-9.33595300e-02 -9.45610851e-02 2.23231956e-01 3.40403378e-01
1.33710790e+00 5.36764860e-02 2.81464905e-01 -4.27281827e-01
7.60106087e-01 5.79202501e-03 5.29080808e-01 8.38499904e-01
3.58838230e-01 1.97981760e-01 1.05787528e+00 -1.01008117e+00
-1.00458658e+00 -1.09378695e+00 -8.30373839e-02 1.37827981e+00
-8.35000724e-02 -5.65807402e-01 -5.44364750e-01 -4.86631334e-01
1.20927364e-01 8.62705529e-01 -8.45549047e-01 -1.73806787e-01
-5.92751086e-01 -3.30751181e-01 5.68688571e-01 6.70305789e-01
-2.09695902e-02 -1.64788246e+00 -6.15353346e-01 2.53311455e-01
2.24500939e-01 -4.41951334e-01 8.31153467e-02 1.08098054e+00
-9.00718689e-01 -1.24481916e+00 4.04738963e-01 -1.08036160e+00
1.15262127e+00 -2.01124415e-01 1.31428015e+00 7.49236703e-01
-1.72513992e-01 8.61644894e-02 -7.48090819e-02 -3.53701919e-01
-6.42209411e-01 -8.56075808e-02 -8.90250355e-02 -2.63812989e-01
3.63870233e-01 -6.64298892e-01 -5.61398640e-02 2.50891477e-01
-1.17292249e+00 3.18951547e-01 4.41755563e-01 9.83970106e-01
7.09323347e-01 3.62866282e-01 5.68898141e-01 -1.39069426e+00
4.03848350e-01 -3.82303178e-01 -5.57959735e-01 3.06606919e-01
-7.91315377e-01 7.55967796e-01 9.91541386e-01 -1.78195149e-01
-6.05228007e-01 2.12155238e-01 -4.74095419e-02 7.68471435e-02
-1.53357610e-01 9.07378793e-01 -3.33985656e-01 1.50313288e-01
1.03815150e+00 1.31720677e-01 5.81563339e-02 -9.96742174e-02
5.22128820e-01 2.44539544e-01 6.21039212e-01 -9.13687646e-01
6.46618962e-01 8.13916326e-02 3.74017537e-01 -4.49889898e-01
-2.19635144e-01 3.17385346e-01 -5.46458423e-01 3.83117311e-02
4.92519259e-01 -3.51711422e-01 -1.15039277e+00 -1.29567251e-01
-1.11008561e+00 -3.20997983e-01 -5.07153451e-01 4.05250527e-02
-6.74265921e-01 -1.07928507e-01 -5.44193149e-01 -7.79546559e-01
-1.05558105e-01 -1.24249029e+00 3.64877135e-01 8.50050077e-02
-5.31814098e-01 -8.71612012e-01 -4.52956825e-01 -9.68281031e-02
2.05083787e-01 1.46404982e-01 1.91042650e+00 -9.84994113e-01
-7.39885867e-01 -2.67710924e-01 -3.15016329e-01 4.27872539e-01
-2.71211296e-01 3.47575098e-01 -7.58589923e-01 2.03851536e-01
-3.57067674e-01 -7.91249350e-02 4.29719269e-01 3.26428145e-01
1.04181957e+00 -6.08194411e-01 -5.09826362e-01 3.45785141e-01
1.44140959e+00 5.30972779e-01 7.38539398e-01 2.98900217e-01
6.06808126e-01 8.63906145e-01 -1.06082007e-01 -3.60999629e-02
4.49669480e-01 2.29841202e-01 6.56724513e-01 6.72524795e-02
3.17013890e-01 -2.71870047e-01 1.13166809e-01 4.31806058e-01
-2.80553102e-01 4.96309847e-02 -1.02201617e+00 5.96643388e-01
-1.75993752e+00 -8.71231258e-01 -5.88164508e-01 2.21114039e+00
8.27412784e-01 5.82826197e-01 -1.35282904e-01 3.42721790e-01
6.99953914e-01 -2.83951938e-01 -3.14421296e-01 -1.03825009e+00
-5.86511679e-02 7.85081863e-01 2.67350227e-01 7.34068632e-01
-7.67724872e-01 8.63734424e-01 5.78954220e+00 2.51491100e-01
-6.95006311e-01 -3.71374816e-01 6.72096729e-01 2.39624560e-01
-8.31266999e-01 2.83378094e-01 -7.00858235e-01 1.43984959e-01
1.03530920e+00 -1.34063167e-02 7.96944082e-01 9.69088972e-01
-3.25484365e-01 6.64989511e-03 -1.88357222e+00 4.38830972e-01
-2.82699198e-01 -1.62976587e+00 5.84522076e-02 -5.45192510e-02
4.35763419e-01 -2.65440524e-01 -1.05102494e-01 3.95606905e-01
6.84594393e-01 -1.56237781e+00 7.65520513e-01 2.62896806e-01
7.83332527e-01 -1.00906861e+00 4.72494334e-01 3.12152147e-01
-1.38358629e+00 -3.23064923e-01 -3.53457689e-01 -4.19606388e-01
-4.07566875e-01 2.50821501e-01 -8.09341609e-01 3.03785592e-01
5.98618805e-01 2.79556543e-01 -3.26582015e-01 2.77599663e-01
-5.03632784e-01 2.22104788e-01 -3.81231874e-01 -1.72333017e-01
2.72248924e-01 3.99835855e-02 9.76303890e-02 1.21514726e+00
1.02486648e-01 3.19113672e-01 -1.54796496e-01 1.07969177e+00
-9.78000909e-02 -1.20048985e-01 -9.84497845e-01 1.04362354e-01
4.35558110e-01 9.89098191e-01 -8.34864020e-01 -3.48008156e-01
-4.19211030e-01 3.04634005e-01 4.24912900e-01 2.28669047e-02
-6.86547339e-01 -4.39679384e-01 6.31835461e-01 2.65758246e-01
1.36209801e-01 1.95135564e-01 -3.08672577e-01 -7.52539754e-01
-1.26623750e-01 -1.01172757e+00 5.28663039e-01 -9.50724244e-01
-1.15325868e+00 8.77464354e-01 1.28540367e-01 -6.53406918e-01
-4.64409113e-01 -8.09514582e-01 -4.32293832e-01 1.09686327e+00
-8.09658051e-01 -1.14556372e+00 1.58056542e-01 7.83711374e-01
-3.20626386e-02 -1.26778021e-01 1.28594482e+00 -1.74004644e-01
-2.26245224e-01 3.58031780e-01 -2.74444759e-01 2.41537809e-01
-3.42611760e-01 -1.31823909e+00 4.76421684e-01 5.39084792e-01
3.33388269e-01 1.03095114e+00 6.85109794e-01 -5.08888364e-01
-1.42318296e+00 -8.03854942e-01 1.42548275e+00 -4.30554956e-01
5.66521466e-01 -3.44463408e-01 -9.44291174e-01 1.42490005e+00
-1.50696903e-01 1.76514518e-02 5.97178400e-01 5.00142932e-01
-5.85249603e-01 -2.83140481e-01 -1.05405080e+00 6.53361380e-01
1.26806521e+00 -5.41148007e-01 -6.49814606e-01 2.20849738e-01
7.66548097e-01 -1.23106129e-01 -8.18606079e-01 4.69691753e-01
6.37405217e-01 -1.16180634e+00 1.02037144e+00 -1.00878775e+00
6.73183262e-01 -2.77386338e-01 -3.19953471e-01 -9.89972353e-01
-4.68963653e-01 -3.55781764e-01 8.18161666e-02 9.05732930e-01
1.00304544e+00 -9.27786767e-01 8.75560462e-01 1.08534718e+00
-3.34324837e-01 -9.66049790e-01 -8.63661945e-01 -4.32833016e-01
2.98738554e-02 -8.38241100e-01 8.15789402e-01 9.11676466e-01
3.62551451e-01 5.34292996e-01 1.14146322e-01 4.15913835e-02
2.81512022e-01 1.90467015e-01 3.97235185e-01 -1.57744443e+00
-6.55588925e-01 -5.75810015e-01 -5.47810793e-01 -6.19741321e-01
2.89382845e-01 -1.27787232e+00 1.21566355e-01 -1.41457558e+00
1.09507091e-01 -9.22561407e-01 -1.92147434e-01 1.07170129e+00
3.75049680e-01 -6.15802035e-02 -8.75920290e-04 1.53861225e-01
-1.27395421e-01 -1.65417597e-01 9.12179291e-01 -7.67938048e-02
-1.96892828e-01 -1.63125128e-01 -1.05085254e+00 9.95122015e-01
6.10861182e-01 -5.97421288e-01 -5.74165523e-01 -3.39063466e-01
8.42899084e-01 3.47361833e-01 3.94907534e-01 -6.72425389e-01
1.51260778e-01 -2.36598611e-01 3.75635564e-01 -1.13860704e-01
1.94700897e-01 -9.21417475e-01 8.02299261e-01 5.51607966e-01
-6.76669240e-01 3.56395811e-01 3.23439762e-02 2.73213089e-01
3.98753621e-02 -3.82421553e-01 4.97729659e-01 -4.31703061e-01
-6.62459850e-01 -1.69570476e-01 -3.26470464e-01 -3.49674284e-01
9.40878510e-01 -3.62291366e-01 -4.07828987e-01 -2.93921888e-01
-9.83543098e-01 -3.90161425e-02 3.62461150e-01 1.12171248e-01
6.93896413e-01 -1.06001365e+00 -2.42822498e-01 6.20389819e-01
1.68310419e-01 3.92218322e-01 -2.71619409e-01 3.67041260e-01
-7.69644558e-01 4.43404287e-01 -1.98260441e-01 -3.61789137e-01
-1.03668904e+00 7.52663910e-01 5.53265333e-01 -1.84465513e-01
-4.28758353e-01 1.14648449e+00 2.52885669e-01 -8.41333270e-01
2.00127751e-01 -5.91927886e-01 -1.16847672e-01 -2.62931854e-01
2.75838256e-01 -8.68116021e-02 1.03264237e-02 -6.92865193e-01
-3.33815396e-01 9.21264365e-02 3.22854258e-02 -5.04974276e-02
1.47262812e+00 4.11076635e-01 -6.18637979e-01 3.72279912e-01
1.07612085e+00 -3.84270668e-01 -7.64086246e-01 -9.92339253e-02
1.41320020e-01 -1.03619255e-01 -4.41470325e-01 -7.60396659e-01
-1.11685610e+00 7.23649025e-01 -1.75703261e-02 8.60745668e-01
9.86606956e-01 4.94717240e-01 1.09385192e-01 6.02290750e-01
5.34702301e-01 -6.58920527e-01 -2.75172949e-01 6.27440810e-01
7.71871507e-01 -6.46915436e-01 -1.59201875e-01 -5.85078597e-01
-3.90501231e-01 1.30083311e+00 5.07836521e-01 -5.93926683e-02
4.31307584e-01 3.86220008e-01 -2.42801979e-01 -4.78439659e-01
-1.00203550e+00 -1.60941303e-01 1.30122870e-01 5.65204263e-01
5.54117560e-01 3.00601065e-01 -2.31222644e-01 8.22308660e-01
-4.33684319e-01 -1.46049201e-01 4.04768497e-01 9.52951908e-01
-5.79977214e-01 -1.37231112e+00 -2.52487093e-01 8.16317976e-01
-1.92503273e-01 -3.37908655e-01 -4.26390171e-01 1.13640392e+00
6.10347688e-01 7.79214799e-01 2.03637019e-01 -6.91899657e-01
3.54864389e-01 4.61255223e-01 5.71041524e-01 -7.63216972e-01
-7.81215310e-01 -4.35669661e-01 3.50219578e-01 -2.49555528e-01
-3.47011745e-01 -5.82790852e-01 -1.99913788e+00 -8.52598429e-01
-2.23144889e-01 2.43041858e-01 5.89240491e-01 1.01950753e+00
1.39480546e-01 3.99979025e-01 3.99632990e-01 -4.93975341e-01
-1.85587481e-01 -4.21462506e-01 -6.13282979e-01 4.45950866e-01
1.69279668e-02 -5.09921849e-01 -4.32062536e-01 2.20474824e-01] | [8.882976531982422, 6.9422783851623535] |
f36c5fc9-821c-41f5-9900-7b848cf1137e | policyclustergcn-identifying-efficient | 2306.14357 | null | https://arxiv.org/abs/2306.14357v1 | https://arxiv.org/pdf/2306.14357v1.pdf | PolicyClusterGCN: Identifying Efficient Clusters for Training Graph Convolutional Networks | Graph convolutional networks (GCNs) have achieved huge success in several machine learning (ML) tasks on graph-structured data. Recently, several sampling techniques have been proposed for the efficient training of GCNs and to improve the performance of GCNs on ML tasks. Specifically, the subgraph-based sampling approaches such as ClusterGCN and GraphSAINT have achieved state-of-the-art performance on the node classification tasks. These subgraph-based sampling approaches rely on heuristics -- such as graph partitioning via edge cuts -- to identify clusters that are then treated as minibatches during GCN training. In this work, we hypothesize that rather than relying on such heuristics, one can learn a reinforcement learning (RL) policy to compute efficient clusters that lead to effective GCN performance. To that end, we propose PolicyClusterGCN, an online RL framework that can identify good clusters for GCN training. We develop a novel Markov Decision Process (MDP) formulation that allows the policy network to predict ``importance" weights on the edges which are then utilized by a clustering algorithm (Graclus) to compute the clusters. We train the policy network using a standard policy gradient algorithm where the rewards are computed from the classification accuracies while training GCN using clusters given by the policy. Experiments on six real-world datasets and several synthetic datasets show that PolicyClusterGCN outperforms existing state-of-the-art models on node classification task. | ['Srinivasan Parthasarathy', 'Balaraman Ravindran', 'Shaileshh Bojja Venkatakrishnan', 'Saket Gurukar'] | 2023-06-25 | null | null | null | null | ['node-classification', 'graph-partitioning'] | ['graphs', 'graphs'] | [-1.21336348e-01 5.42918205e-01 -7.97796726e-01 -3.90798271e-01
-5.29768050e-01 -1.34482384e-01 5.34376919e-01 3.17577362e-01
-3.34292293e-01 6.02060854e-01 -2.23076701e-01 -5.51659644e-01
-1.95152223e-01 -1.09330118e+00 -8.87139559e-01 -8.49696815e-01
-4.12424505e-01 9.23217356e-01 -1.00886106e-01 2.57707626e-01
1.37122333e-01 3.13141406e-01 -1.36774957e+00 2.78598607e-01
1.29159403e+00 9.38531697e-01 3.48460317e-01 6.24053657e-01
-1.85988516e-01 9.72323298e-01 -2.49013603e-01 -1.74394146e-01
3.36612463e-01 -4.33752537e-01 -8.29719841e-01 2.59999543e-01
-1.39832914e-01 -7.36317970e-03 -2.44557723e-01 1.02534115e+00
2.73679316e-01 4.99300301e-01 6.43941402e-01 -1.46925664e+00
-3.58616650e-01 1.19036281e+00 -5.55228591e-01 -1.33468121e-01
-1.14919648e-01 2.18970329e-01 1.31115746e+00 -4.25020099e-01
6.98114932e-01 1.42317653e+00 4.44790751e-01 7.11730421e-01
-1.34459889e+00 -6.14480972e-01 6.12142563e-01 2.81115592e-01
-1.14424682e+00 2.78847292e-02 8.26545596e-01 -2.09409595e-01
9.55533683e-01 -5.88510334e-02 1.02914608e+00 1.02312374e+00
4.78941835e-02 1.26746869e+00 9.31914032e-01 -5.40699780e-01
7.64092147e-01 -1.28047034e-01 -1.50023580e-01 8.88261795e-01
1.88651115e-01 1.87756315e-01 -3.01913410e-01 -1.12460695e-01
6.14736259e-01 1.04967788e-01 1.64596662e-02 -6.84066713e-01
-5.77090859e-01 1.35281003e+00 8.76004219e-01 1.18811585e-01
-5.99540353e-01 5.32532871e-01 4.59169120e-01 1.75752878e-01
6.93147600e-01 4.26247776e-01 -3.47116172e-01 1.37706324e-01
-8.09489608e-01 -2.06224271e-03 9.42539215e-01 8.55407476e-01
1.03218949e+00 -2.84732953e-02 -2.78820366e-01 7.30014861e-01
4.30140465e-01 1.98847651e-01 3.06211978e-01 -8.58710587e-01
5.03301263e-01 9.82247412e-01 -3.23621660e-01 -8.93649578e-01
-3.19953531e-01 -5.13599575e-01 -1.08300996e+00 -7.18281092e-03
6.10721111e-02 -3.34936142e-01 -1.18870997e+00 1.69003129e+00
4.49317038e-01 5.15117586e-01 4.85104546e-02 6.83771551e-01
4.05205667e-01 6.62137687e-01 3.66042078e-01 -8.99822712e-02
6.24850929e-01 -9.71561849e-01 -4.43554848e-01 -1.34187013e-01
9.35728312e-01 6.12063967e-02 9.99433756e-01 5.21705925e-01
-5.88122487e-01 -4.90591109e-01 -8.60247076e-01 5.95047534e-01
-3.84184837e-01 1.75092161e-01 8.59572232e-01 6.44724250e-01
-1.34678304e+00 1.04284728e+00 -1.24131846e+00 -4.60253179e-01
7.05703318e-01 5.62574029e-01 2.48778507e-01 -3.45590800e-01
-8.60126793e-01 5.02031744e-01 9.22535956e-01 9.58666652e-02
-1.52779877e+00 -3.89980286e-01 -7.49433756e-01 1.62266359e-01
8.20622861e-01 -4.41069990e-01 1.18552041e+00 -1.27334535e+00
-1.45703769e+00 4.43080246e-01 2.54991621e-01 -9.44557250e-01
3.43240201e-01 7.76076876e-03 -1.11497842e-01 2.98815906e-01
-1.31708933e-02 1.05214798e+00 8.48082900e-01 -1.31859493e+00
-8.35168540e-01 -1.39922932e-01 -1.26239702e-01 3.56948614e-01
-3.76868725e-01 -3.48708838e-01 -6.11120284e-01 -1.91403911e-01
-1.77434638e-01 -1.09511375e+00 -7.69631624e-01 -5.71947038e-01
-8.16601872e-01 -9.42121208e-01 8.11886191e-01 -2.48448923e-01
1.27301109e+00 -1.74306583e+00 2.53742874e-01 7.62053609e-01
4.26233500e-01 2.97867358e-01 -1.80186614e-01 4.18045878e-01
3.04387242e-01 3.00180167e-01 5.01833111e-02 -4.69331503e-01
-7.29251355e-02 5.21573126e-01 6.72916174e-02 3.88064563e-01
6.96193278e-02 8.94472420e-01 -1.09214091e+00 -4.91429716e-01
2.47627050e-01 6.87953979e-02 -7.64702082e-01 4.19735700e-01
-9.22302485e-01 2.86334783e-01 -6.25735044e-01 5.19631624e-01
3.19418192e-01 -6.06182873e-01 7.65312135e-01 3.34994167e-01
3.29685330e-01 6.50819317e-02 -1.00113463e+00 1.30178881e+00
-3.19482207e-01 3.69597077e-01 1.34414911e-01 -1.53203523e+00
9.46662664e-01 7.37718195e-02 5.96388996e-01 -2.48928547e-01
1.35775015e-01 -2.40196791e-02 1.12149730e-01 -2.76403874e-01
1.74098834e-01 2.88687140e-01 1.32422859e-03 4.69051927e-01
1.43094331e-01 6.07316904e-02 3.70252520e-01 5.23066044e-01
1.24872267e+00 4.12974767e-02 1.72077253e-01 -3.58057588e-01
2.20390722e-01 4.25287783e-02 6.91631317e-01 9.78051364e-01
-1.03307232e-01 -6.81380332e-02 8.02527606e-01 -4.88835305e-01
-8.15541267e-01 -6.52218223e-01 5.39096892e-01 1.24743664e+00
-1.18000269e-01 -5.58866322e-01 -9.94528949e-01 -1.08502614e+00
-3.99975711e-03 7.99329937e-01 -8.40262353e-01 -4.19412196e-01
-3.66436809e-01 -7.10362613e-01 1.64074913e-01 5.70376873e-01
4.66753215e-01 -1.56665349e+00 -3.93311203e-01 4.56989527e-01
9.45829749e-02 -9.76977766e-01 -1.98128372e-01 4.90986943e-01
-1.00658131e+00 -1.28701508e+00 -1.34281993e-01 -8.30430865e-01
9.69194531e-01 1.23245828e-01 1.29961777e+00 3.75045002e-01
-1.14995576e-01 4.61173743e-01 -6.00396574e-01 -4.06045914e-01
-4.10073549e-01 6.16358876e-01 -7.53963664e-02 1.39180407e-01
3.21971118e-01 -4.94022340e-01 -6.66239142e-01 1.00979775e-01
-6.80743337e-01 1.96086809e-01 7.79429913e-01 6.34549797e-01
8.24183524e-01 1.94982365e-01 6.66308701e-01 -1.44995522e+00
8.89985621e-01 -6.47027612e-01 -7.97107816e-01 3.05673301e-01
-1.07389235e+00 3.24729294e-01 1.14915514e+00 -4.75395828e-01
-6.48565412e-01 6.72879294e-02 2.00382501e-01 -1.03966308e+00
-9.17413533e-02 7.35494256e-01 4.19556350e-02 1.12190552e-01
6.23518586e-01 1.04195163e-01 -5.00258915e-02 -1.22982182e-01
5.14921725e-01 4.39972192e-01 6.54883534e-02 -9.44362044e-01
5.59738398e-01 2.69629985e-01 4.78100181e-02 -4.18980241e-01
-1.04396641e+00 -5.33849180e-01 -4.88176912e-01 -4.94613975e-01
9.75247741e-01 -8.17129016e-01 -9.22409177e-01 1.10460930e-01
-6.15073025e-01 -1.06562042e+00 -2.03085855e-01 3.08800131e-01
-6.87377691e-01 -5.02912737e-02 -6.47005379e-01 -1.08082688e+00
-4.60716993e-01 -1.06711400e+00 7.95638740e-01 4.84617978e-01
2.74424285e-01 -1.17043173e+00 -1.12658612e-01 2.46262420e-02
1.86371267e-01 3.70039880e-01 1.06334543e+00 -8.31112385e-01
-7.74002910e-01 1.29099593e-01 -6.40774965e-02 2.45632067e-01
-1.39745651e-02 1.67517751e-01 -6.20642483e-01 -7.64039636e-01
-6.70303702e-01 -6.90059841e-01 1.01048315e+00 7.58011997e-01
1.61660695e+00 -5.24013400e-01 -9.19731200e-01 6.69404507e-01
1.61616695e+00 3.46507996e-01 3.59405518e-01 9.12143290e-02
9.37942624e-01 3.64987463e-01 6.75913930e-01 4.64428872e-01
4.33451146e-01 2.12118775e-01 8.38152230e-01 -1.48286074e-01
2.41843224e-01 -6.93488479e-01 2.39779711e-01 5.73870301e-01
-2.18979567e-02 -4.79197860e-01 -9.76143241e-01 4.54299808e-01
-2.34083581e+00 -6.79168046e-01 2.89799452e-01 1.98850667e+00
5.87672353e-01 2.27136120e-01 3.47524136e-01 -1.81412324e-01
9.37917233e-01 1.54091775e-01 -8.94407511e-01 -3.85952741e-01
4.20544535e-01 1.12031370e-01 5.46528459e-01 3.25297832e-01
-1.17215943e+00 1.46520329e+00 5.38439322e+00 9.11162078e-01
-9.95028794e-01 -3.00312549e-01 1.16765416e+00 1.58906743e-01
-1.21452160e-01 1.60426781e-01 -1.00429249e+00 3.42319727e-01
1.08135355e+00 1.84335317e-02 9.50337172e-01 1.26090658e+00
3.86905074e-01 -1.45167150e-02 -1.12554801e+00 6.23584807e-01
-2.01779619e-01 -1.39101851e+00 6.11573048e-02 3.22259307e-01
1.07399035e+00 2.90801734e-01 -8.98422748e-02 7.81168580e-01
1.33317339e+00 -1.06899667e+00 2.87873536e-01 2.49411926e-01
5.58890104e-01 -1.21875942e+00 5.36981821e-01 6.30311191e-01
-1.29030800e+00 -4.07939106e-01 -7.02366769e-01 3.05957079e-01
-3.90826225e-01 6.59514368e-01 -1.54044580e+00 6.02544427e-01
6.33596897e-01 8.65496755e-01 -6.23426318e-01 1.03340554e+00
-4.48225617e-01 1.12036347e+00 -2.08991036e-01 -5.17834365e-01
6.64933383e-01 -5.13388574e-01 9.82801467e-02 9.08681512e-01
7.41515979e-02 -4.95981835e-02 9.53591406e-01 9.01203036e-01
-3.51163149e-01 1.97483540e-01 -5.21716416e-01 -2.76456654e-01
7.01972306e-01 1.42020166e+00 -1.25580323e+00 -3.46218437e-01
-1.07445888e-01 5.50032079e-01 1.06983578e+00 4.03865665e-01
-6.09121382e-01 -1.27209976e-01 2.59673715e-01 -2.14586798e-02
5.60847938e-01 2.42774002e-03 5.07642291e-02 -7.95479953e-01
-5.08259237e-01 -8.36429656e-01 8.33369076e-01 -3.11647952e-01
-1.27879262e+00 5.31172752e-01 5.76181076e-02 -9.66892123e-01
-4.75015044e-01 -3.87070358e-01 -8.64690721e-01 2.44041175e-01
-1.44210184e+00 -8.88958335e-01 -1.32973224e-01 6.88677967e-01
3.73585612e-01 -2.84756094e-01 5.21701753e-01 -2.74224579e-01
-9.06093419e-01 4.63895589e-01 2.22938374e-01 4.35586840e-01
2.10180670e-01 -1.60046279e+00 1.67989120e-01 7.48680055e-01
2.34590918e-01 2.00640604e-01 4.22398210e-01 -8.73901427e-01
-1.30088401e+00 -1.64329600e+00 2.23523483e-01 2.69389808e-01
2.90819138e-01 -5.48500061e-01 -6.30463183e-01 7.98059821e-01
2.36990094e-01 2.01945648e-01 4.44538891e-01 3.61754537e-01
-4.52142581e-02 -2.42984608e-01 -1.09144473e+00 5.22913337e-01
9.45742667e-01 -8.71501267e-02 2.15819433e-01 6.27479196e-01
8.07074249e-01 -1.89162880e-01 -7.05164492e-01 2.05365390e-01
-2.12852538e-01 -7.91448355e-01 5.37086189e-01 -8.78129065e-01
2.73995548e-01 1.41708376e-02 2.10046381e-01 -1.90362692e+00
-2.54597157e-01 -7.65321732e-01 -3.74828666e-01 9.42427218e-01
4.92389113e-01 -5.84354281e-01 1.38048625e+00 3.92269313e-01
-2.02834576e-01 -1.19106543e+00 -7.03174770e-01 -5.71662962e-01
-4.64230292e-02 -2.47787818e-01 5.26635110e-01 8.40168476e-01
-1.11744009e-01 3.25895548e-01 -1.25598490e-01 -2.68451534e-02
8.65737677e-01 3.14208001e-01 7.90091276e-01 -1.47306979e+00
-4.00460750e-01 -4.25617933e-01 -6.69802278e-02 -8.21413159e-01
6.24645174e-01 -1.23128498e+00 2.10781038e-01 -1.68476248e+00
2.46914074e-01 -7.89732337e-01 -2.91286141e-01 7.34511554e-01
-3.33326936e-01 -5.82711697e-01 2.34675407e-01 8.34954083e-02
-1.33636367e+00 6.35805666e-01 1.20626247e+00 -7.00712949e-02
-4.92538661e-01 3.31695080e-01 -6.56393468e-01 3.90197963e-01
1.05118299e+00 -6.71722829e-01 -7.10812688e-01 1.10297330e-01
1.91889450e-01 2.75552213e-01 1.01225413e-02 -1.09718716e+00
3.17223847e-01 -3.36108238e-01 4.10873264e-01 -6.81030154e-01
-8.01982507e-02 -5.94832361e-01 -1.89939380e-01 7.69457281e-01
-6.32716656e-01 -8.45906660e-02 -2.49527663e-01 1.05304718e+00
7.33014494e-02 -6.48231879e-02 7.97724307e-01 -3.38490933e-01
-5.51729560e-01 7.76569366e-01 -4.18108761e-01 2.15410843e-01
1.18906820e+00 1.75323352e-01 -3.35186766e-03 -6.68450058e-01
-6.93495333e-01 8.25067639e-01 1.38875797e-01 2.13438332e-01
5.68499744e-01 -1.23232388e+00 -5.50186574e-01 1.82445589e-02
-1.74891308e-01 4.18626547e-01 -1.38777167e-01 5.75366497e-01
-2.97028959e-01 3.79804313e-01 9.28021222e-02 -6.51088893e-01
-8.68007183e-01 8.14559698e-01 5.14417529e-01 -1.00147712e+00
-5.31034350e-01 7.20795214e-01 -1.49922147e-01 -6.82241321e-01
5.47355175e-01 -4.70793426e-01 -3.04025322e-01 -1.53698921e-01
-1.23784505e-02 2.98141807e-01 -8.24460164e-02 9.47182849e-02
3.41222435e-02 -1.07977055e-01 -2.62761444e-01 1.83997512e-01
1.47846448e+00 1.48267627e-01 1.71515584e-01 7.12262914e-02
1.05711687e+00 -7.01093078e-01 -1.62310028e+00 -3.06001663e-01
3.24202955e-01 -1.51814610e-01 5.06923608e-02 -6.06904566e-01
-1.46178758e+00 7.19940305e-01 3.32644612e-01 5.89223385e-01
1.03336835e+00 9.09529328e-02 4.84376907e-01 5.09813070e-01
4.37285542e-01 -1.49950135e+00 3.11618894e-01 2.76411951e-01
3.70232105e-01 -1.19836974e+00 -1.35029122e-01 -3.23152661e-01
-8.11547101e-01 9.94684279e-01 1.11382437e+00 -5.53013444e-01
7.48899043e-01 -3.22156362e-02 -2.85024703e-01 -2.80627549e-01
-1.21319163e+00 -3.81154716e-01 7.46946856e-02 7.26372957e-01
1.28858820e-01 6.22099161e-01 2.32541896e-02 2.34658346e-01
3.50141227e-02 -1.12970099e-01 2.47625783e-01 6.75707698e-01
-5.89071751e-01 -1.14381921e+00 -6.24100119e-02 1.03290343e+00
-2.05145925e-01 1.25398606e-01 -5.15015900e-01 7.86377788e-01
-8.96826759e-02 1.00783980e+00 -1.21575696e-02 -6.11460149e-01
-1.58052161e-01 -1.79826275e-01 2.10817724e-01 -7.52748013e-01
-6.05792701e-01 1.75188735e-01 1.18978292e-01 -7.60601521e-01
-3.62170935e-01 -5.85656345e-01 -1.40838897e+00 -2.35261187e-01
-3.97178799e-01 4.84301865e-01 4.93184149e-01 9.09210443e-01
4.38948005e-01 5.70584416e-01 9.21451867e-01 -9.06397939e-01
-6.06481910e-01 -1.00975239e+00 -6.74937665e-01 3.04243356e-01
-2.93598950e-01 -5.81101000e-01 -2.82139003e-01 -5.35443544e-01] | [7.331878185272217, 6.215417385101318] |
455db635-bbca-4fa7-9d8c-e6b531e89465 | hierarchical-federated-learning | 2307.00233 | null | https://arxiv.org/abs/2307.00233v1 | https://arxiv.org/pdf/2307.00233v1.pdf | Hierarchical Federated Learning Incentivization for Gas Usage Estimation | Accurately estimating gas usage is essential for the efficient functioning of gas distribution networks and saving operational costs. Traditional methods rely on centralized data processing, which poses privacy risks. Federated learning (FL) offers a solution to this problem by enabling local data processing on each participant, such as gas companies and heating stations. However, local training and communication overhead may discourage gas companies and heating stations from actively participating in the FL training process. To address this challenge, we propose a Hierarchical FL Incentive Mechanism for Gas Usage Estimation (HI-GAS), which has been testbedded in the ENN Group, one of the leading players in the natural gas and green energy industry. It is designed to support horizontal FL among gas companies, and vertical FL among each gas company and heating station within a hierarchical FL ecosystem, rewarding participants based on their contributions to FL. In addition, a hierarchical FL model aggregation approach is also proposed to improve the gas usage estimation performance by aggregating models at different levels of the hierarchy. The incentive scheme employs a multi-dimensional contribution-aware reward distribution function that combines the evaluation of data quality and model contribution to incentivize both gas companies and heating stations within their jurisdiction while maintaining fairness. Results of extensive experiments validate the effectiveness of the proposed mechanism. | ['Han Yu', 'Zengxiang Li', 'Qijie Ding', 'Xiuli Wang', 'Zhenpeng Yu', 'Chengyi Yang', 'Xiaoli Tang', 'Has Sun'] | 2023-07-01 | null | null | null | null | ['fairness', 'fairness'] | ['computer-vision', 'miscellaneous'] | [-4.00276452e-01 3.75248939e-01 -3.44742537e-01 -2.91848123e-01
-3.11834604e-01 -6.09546125e-01 2.05709562e-01 3.30821782e-01
-4.24914688e-01 8.45903337e-01 -1.39990747e-01 -4.40698355e-01
-4.02946204e-01 -1.51578176e+00 -1.36872873e-01 -1.04276145e+00
-1.97241873e-01 3.55959177e-01 -3.54524374e-01 2.67955512e-01
-7.11526647e-02 5.12255967e-01 -1.02635264e+00 -1.77468911e-01
1.02126992e+00 1.30389237e+00 4.06210870e-02 3.09224218e-01
-1.69716746e-01 7.34187782e-01 -6.05670989e-01 -1.85402870e-01
7.99899936e-01 -5.97584136e-02 -7.85600126e-01 -2.99835265e-01
-5.41802049e-01 -5.92344701e-01 4.00687493e-02 1.20921850e+00
3.28628421e-01 3.59014571e-01 7.93550089e-02 -1.87257588e+00
-1.32161662e-01 1.06098211e+00 -4.62084115e-01 -1.92001700e-01
-2.02227205e-01 2.81588554e-01 1.12941432e+00 -1.22019295e-02
2.22615302e-02 7.91783929e-01 4.30396199e-01 5.45044482e-01
-9.67965662e-01 -1.00027955e+00 1.77456185e-01 -8.69946629e-02
-1.26899505e+00 -9.66930911e-02 6.82171226e-01 -2.35409468e-01
6.46367610e-01 9.33610499e-01 6.99726939e-01 1.37327105e-01
-4.44253534e-01 5.33600509e-01 1.01121974e+00 -3.16410810e-01
6.95549488e-01 3.76673728e-01 -2.78745294e-01 3.02427728e-02
5.77085674e-01 1.01421446e-01 -1.93532109e-02 -5.59084475e-01
6.29813492e-01 1.10530689e-01 -2.98329204e-01 -3.94051403e-01
-7.97482669e-01 8.48302603e-01 6.69294000e-01 4.44761574e-01
-8.27279866e-01 1.60468861e-01 4.68517452e-01 1.74581371e-02
4.95200783e-01 2.99689174e-01 -4.02174652e-01 2.30158344e-01
-9.64315593e-01 4.06647380e-03 9.36680198e-01 9.03485715e-01
9.08557475e-01 8.43631849e-03 -3.39884669e-01 1.40153527e-01
4.87944365e-01 3.01095814e-01 -4.82025295e-02 -1.34506190e+00
5.09550810e-01 8.72196317e-01 3.99597108e-01 -7.37464786e-01
-2.37231821e-01 -3.10132265e-01 -1.19144666e+00 4.42045450e-01
9.81065333e-02 -5.47827125e-01 -3.36550996e-02 1.65748298e+00
6.44364834e-01 -1.51047289e-01 -8.83340240e-02 8.63084614e-01
1.12763211e-01 4.85972017e-01 4.46175307e-01 -4.78870720e-01
1.10431445e+00 -9.19148326e-01 -6.93699837e-01 6.61120594e-01
5.71403801e-01 -6.21029064e-02 4.08679068e-01 2.76936352e-01
-1.14870059e+00 -1.32878795e-01 -6.75278366e-01 4.05240893e-01
-5.29982150e-01 -1.58541098e-01 7.02855587e-01 9.98192668e-01
-8.25840831e-01 6.51426136e-01 -6.70058429e-01 -2.38701597e-01
5.76109052e-01 7.64655530e-01 -3.04163508e-02 1.23624615e-01
-1.35065293e+00 4.60980207e-01 5.57586551e-01 1.81036890e-01
-7.70361185e-01 -8.85329664e-01 -5.36620617e-01 4.86663043e-01
3.41589808e-01 -7.62182355e-01 1.18132484e+00 -3.55300754e-01
-9.81420159e-01 2.24246055e-01 4.39481825e-01 -7.84548402e-01
9.85924184e-01 2.88199097e-01 -4.82020110e-01 -3.47554199e-02
-9.92707238e-02 3.01724792e-01 2.74583250e-01 -1.29425240e+00
-1.09641707e+00 -4.54833239e-01 3.66317451e-01 1.19070269e-01
-6.47998393e-01 -7.20719993e-02 1.82222471e-01 -2.17670545e-01
-3.84011418e-01 -4.34159398e-01 -6.85065031e-01 -2.13518545e-01
-3.12631965e-01 -4.64209884e-01 1.14821160e+00 -5.80836475e-01
1.57490838e+00 -1.88467181e+00 -5.10068059e-01 8.93026769e-01
4.13957715e-01 2.57064015e-01 5.24390221e-01 4.58172709e-01
3.32619578e-01 4.86393303e-01 -8.50042626e-02 -2.37583324e-01
4.70323205e-01 1.64854378e-01 9.44143608e-02 3.35469574e-01
-4.09988701e-01 6.01946771e-01 -9.75989938e-01 -2.79343635e-01
5.80935597e-01 1.41074181e-01 -3.11747044e-01 4.69704777e-01
-3.91895711e-01 6.23505116e-01 -8.10719728e-01 7.01780319e-01
1.01155639e+00 -2.23845869e-01 5.63438535e-01 -2.92666852e-01
-4.64246064e-01 -1.80243373e-01 -1.27225316e+00 1.21600509e+00
-6.81326926e-01 -3.68266255e-01 8.26660812e-01 -9.77248609e-01
9.62719619e-01 6.56851947e-01 1.11503768e+00 -3.75031531e-01
3.36296558e-01 3.68716240e-01 -5.03953159e-01 -2.46176273e-01
4.86298531e-01 -7.96059668e-02 -1.73496798e-01 8.01146150e-01
-3.40708613e-01 1.29848078e-01 1.01130620e-01 9.40372050e-02
9.17785525e-01 -3.21076900e-01 3.60327959e-01 -2.90390402e-01
8.10765386e-01 -3.09112161e-01 6.62671447e-01 5.95126390e-01
-5.86414874e-01 -4.25032616e-01 2.28149414e-01 -4.29604977e-01
-9.23032165e-01 -5.40587306e-01 2.83844292e-01 1.07081437e+00
1.45340443e-01 -2.82556891e-01 -8.56594384e-01 -1.00951421e+00
2.89980888e-01 9.35583174e-01 -3.18001717e-01 7.19453692e-02
-1.54233247e-01 -5.36849439e-01 5.12048721e-01 2.65989095e-01
1.02753496e+00 -7.95729816e-01 -8.33971620e-01 3.38715851e-01
-3.60546499e-01 -7.49035120e-01 -3.75731945e-01 2.14687660e-01
-3.77647430e-01 -9.94315565e-01 -3.42934728e-01 -9.92182717e-02
6.74323082e-01 1.81981474e-01 8.18494618e-01 1.12535022e-01
-7.54106268e-02 2.32958600e-01 -1.88544244e-02 -5.56866169e-01
-5.09792149e-01 2.83544987e-01 -5.90252466e-02 2.82370839e-02
4.85117108e-01 -5.16184270e-01 -1.03136623e+00 5.26569188e-01
-9.29281533e-01 -2.28489384e-01 1.28877193e-01 4.02853996e-01
2.48932660e-01 7.94748366e-01 9.17437255e-01 -1.09176815e+00
5.91380775e-01 -9.06815112e-01 -9.65549171e-01 4.07017171e-01
-1.15212405e+00 -2.76387960e-01 9.36730146e-01 3.25795740e-01
-1.14356077e+00 -1.00858966e-02 2.92395115e-01 -3.88764679e-01
-8.18965510e-02 2.02912077e-01 -5.05473673e-01 -2.68790692e-01
2.24347115e-01 -3.53430420e-01 7.25458488e-02 -4.78691012e-01
1.83334440e-01 9.75392580e-01 3.66262078e-01 -5.62880099e-01
9.24959064e-01 3.82537067e-01 1.85193554e-01 -5.09224460e-02
-4.59854901e-01 -5.62944949e-01 -1.37736341e-02 -3.78302902e-01
6.94262326e-01 -1.07468259e+00 -1.58151186e+00 2.10982591e-01
-8.54592860e-01 -4.01130691e-02 -5.59382081e-01 4.60083723e-01
-2.95556992e-01 1.04594633e-01 -3.39955777e-01 -1.42185771e+00
-8.91214669e-01 -6.88988090e-01 2.79953599e-01 5.35255075e-01
2.01163422e-02 -9.49850738e-01 -2.54606754e-01 4.73624021e-01
1.12771833e+00 5.53788543e-01 7.03074634e-01 -4.60480899e-01
-9.39812958e-01 -1.96438506e-01 -2.30596468e-01 6.11577630e-01
4.47713584e-01 -1.71249956e-01 -1.05804718e+00 -7.20114589e-01
4.40026373e-02 -2.09429160e-01 3.26253772e-01 1.51795775e-01
1.53903067e+00 -7.62375176e-01 -1.52279690e-01 3.35954309e-01
1.71712494e+00 2.87132829e-01 2.93600440e-01 2.71370560e-01
5.11182666e-01 9.40851808e-01 7.33991802e-01 1.05362964e+00
5.51120400e-01 1.51619285e-01 1.15179420e+00 -3.17045271e-01
6.75120413e-01 -2.24284157e-01 1.91203563e-03 3.63931626e-01
-2.39225045e-01 -3.39628816e-01 -4.14144605e-01 4.35492396e-01
-1.91259801e+00 -9.06844318e-01 1.28864333e-01 2.42909598e+00
4.43845332e-01 -1.95410490e-01 4.54048485e-01 2.16946468e-01
7.85141826e-01 2.53849328e-01 -6.55086517e-01 -4.52401936e-01
2.12772042e-01 -1.45476356e-01 1.12348104e+00 2.66252104e-02
-8.25132668e-01 2.73095042e-01 5.37333393e+00 8.25049102e-01
-8.75507295e-01 1.25625491e-01 8.22845638e-01 7.94872642e-02
-5.43262959e-01 1.65984288e-01 -3.64121556e-01 5.19713223e-01
9.74146307e-01 -7.75697708e-01 8.31735671e-01 9.37250972e-01
7.42064714e-01 -4.04756293e-02 -7.48771608e-01 5.40021241e-01
-8.66450071e-01 -1.38468122e+00 -3.80790621e-01 3.66073906e-01
8.54169428e-01 1.82661172e-02 -4.24356848e-01 3.14376969e-03
8.05988491e-01 -8.84205461e-01 4.98763382e-01 4.34716314e-01
6.58397496e-01 -1.22800481e+00 6.99184775e-01 5.42488754e-01
-1.49702859e+00 -5.70400059e-01 -9.81917009e-02 1.94278479e-01
2.25054488e-01 5.89653373e-01 -6.61372423e-01 1.32453966e+00
8.62395406e-01 2.84878705e-02 3.77918035e-02 8.39598179e-01
9.89120528e-02 4.12615061e-01 -3.99306655e-01 1.49355263e-01
2.20839053e-01 -6.13353908e-01 1.17083296e-01 7.03254461e-01
3.87635529e-01 3.08209863e-02 2.77681470e-01 1.18387699e+00
-3.66947830e-01 1.52621374e-01 -3.84968728e-01 -1.13612846e-01
9.26904738e-01 1.74864841e+00 -1.68481991e-01 -1.71782494e-01
-3.02213371e-01 2.36975700e-01 -2.13089250e-02 1.96842805e-01
-5.84204793e-01 -6.16752565e-01 5.90427041e-01 3.77025157e-02
9.88321006e-02 1.63838807e-02 -2.77101547e-01 -4.70730782e-01
-2.20393881e-01 -4.58882391e-01 8.30039620e-01 -1.80719331e-01
-1.30257452e+00 3.55817467e-01 -1.52775958e-01 -1.10304511e+00
-2.27946147e-01 4.55841795e-02 -9.01912332e-01 1.12038553e+00
-1.80489171e+00 -1.20925593e+00 -4.64546651e-01 8.75773787e-01
-2.33835265e-01 -8.98853391e-02 9.06366825e-01 4.85981613e-01
-7.20269918e-01 4.57541853e-01 1.50104210e-01 -1.53231502e-01
-2.97773313e-02 -1.38331974e+00 -1.02558136e-01 8.46629202e-01
-4.29229856e-01 4.79016364e-01 3.41629446e-01 -3.32192540e-01
-1.05029273e+00 -1.59070230e+00 7.42365479e-01 3.33185226e-01
4.15640503e-01 -2.84191936e-01 -6.85779631e-01 3.29506785e-01
2.86509156e-01 2.67890334e-01 8.89965117e-01 -2.53245890e-01
1.35121956e-01 -5.18808424e-01 -2.15598774e+00 6.20463490e-02
5.86282432e-01 -3.77692372e-01 4.28387642e-01 3.00900578e-01
7.53029287e-01 6.99728727e-02 -1.28896666e+00 3.25389087e-01
1.78691760e-01 -9.23940003e-01 4.54656273e-01 -4.32786345e-01
-3.37793201e-01 -5.25601864e-01 -1.86401725e-01 -1.18065071e+00
-2.85982728e-01 -1.14820838e+00 -4.28405941e-01 1.70101154e+00
-8.69352072e-02 -9.87842858e-01 1.17882264e+00 1.37684965e+00
2.48823494e-01 -4.99587029e-01 -9.82510626e-01 -8.93394530e-01
-3.97146419e-02 -2.81467021e-01 1.70256245e+00 1.09160566e+00
7.81263039e-02 -4.23999190e-01 -3.75615090e-01 4.30684447e-01
1.04542184e+00 8.34389776e-02 6.88445866e-01 -1.06115198e+00
-2.62290776e-01 -2.60319382e-01 1.35168567e-01 -3.86698663e-01
-6.16879873e-02 -8.02475274e-01 -2.44270623e-01 -1.69671309e+00
-2.24696800e-01 -1.20006919e+00 -8.45314801e-01 6.70308769e-01
3.85524511e-01 -2.30085403e-01 3.80102187e-01 1.27912730e-01
-5.71103930e-01 4.21047330e-01 9.77511585e-01 -1.09573126e-01
-4.05677892e-02 5.43546140e-01 -8.23923707e-01 3.72149289e-01
1.22672021e+00 -3.81471843e-01 -5.47197282e-01 -1.58090085e-01
-9.73256677e-02 2.24861056e-01 3.21159363e-01 -8.65077853e-01
4.16082293e-01 -5.77895284e-01 -8.50762054e-02 -6.58183873e-01
-2.53614277e-01 -1.63391066e+00 9.11961377e-01 7.44393766e-01
-1.96961403e-01 -3.27090055e-01 -4.05332148e-01 4.92338002e-01
-7.51255676e-02 4.35041077e-03 6.99749351e-01 -3.36059541e-01
-2.75951922e-01 6.55699968e-01 -6.08677976e-02 -4.64903772e-01
1.58417118e+00 8.60312358e-02 -4.33323830e-01 -2.63642609e-01
-4.18854475e-01 1.16657817e+00 3.71581852e-01 3.44933905e-02
-1.43168196e-01 -1.25392699e+00 -4.55546349e-01 2.29712963e-01
-7.81938359e-02 4.45600569e-01 4.78861183e-01 8.10934126e-01
-2.14536622e-01 5.36010861e-01 1.46680437e-02 -3.68909761e-02
-9.02765870e-01 5.44771552e-01 4.97474045e-01 -7.53799677e-01
4.42808382e-02 5.69233000e-01 -1.21345371e-01 -6.77562296e-01
4.89919573e-01 -1.61689281e-01 -1.14848770e-01 3.19626592e-02
2.42318988e-01 1.11316490e+00 1.20338023e-01 -2.40670726e-01
-1.36456177e-01 -2.46494800e-01 2.54725814e-01 3.74737978e-01
1.21373892e+00 -5.83078265e-01 -4.05552834e-01 -7.18925670e-02
9.70783293e-01 -1.35428607e-01 -9.80961323e-01 -2.77310014e-01
5.15097901e-02 -7.38080919e-01 2.00677484e-01 -8.78365159e-01
-1.52665198e+00 2.68391043e-01 3.76623571e-01 8.39667201e-01
1.41521835e+00 -4.68932092e-01 9.32866871e-01 -1.35385737e-01
9.92435098e-01 -1.26974428e+00 -6.90697014e-01 -3.57077926e-01
2.57565588e-01 -1.08049071e+00 -5.97044118e-02 -2.80940175e-01
-4.55329865e-01 8.99512887e-01 9.26243782e-01 2.98060030e-01
5.87732375e-01 1.36203617e-01 -2.08218589e-01 -1.08807951e-01
-6.60047591e-01 2.18902320e-01 -3.70507628e-01 4.99883354e-01
-1.09943271e-01 6.96844816e-01 -4.52703744e-01 7.94160604e-01
2.19698474e-01 2.50641882e-01 2.56323189e-01 1.00789332e+00
-1.70311764e-01 -1.36780870e+00 -2.17830881e-01 6.42985284e-01
-5.35107136e-01 4.36249077e-01 7.81182200e-02 3.65863770e-01
5.99565804e-01 1.17720377e+00 -9.62913632e-02 -1.58280879e-01
3.70847166e-01 -3.50308031e-01 -3.22651893e-01 -1.80668503e-01
-1.07846773e+00 -1.78967789e-02 7.24476799e-02 -5.86843193e-01
-5.16910493e-01 -3.89163852e-01 -1.35418642e+00 -8.32139909e-01
-5.15173137e-01 9.77711618e-01 8.59489918e-01 3.98210943e-01
2.33501539e-01 3.82787406e-01 1.78372467e+00 -4.64296579e-01
-9.88345027e-01 -5.50547838e-01 -1.33065748e+00 1.30472675e-01
2.87936032e-01 -1.41408518e-01 -7.58609772e-01 -6.68832839e-01] | [5.834280490875244, 6.044119834899902] |
265871dd-7437-4fb6-a839-896e77beb2ba | crowdsensing-based-road-damage-detection | 2211.11362 | null | https://arxiv.org/abs/2211.11362v1 | https://arxiv.org/pdf/2211.11362v1.pdf | Crowdsensing-based Road Damage Detection Challenge (CRDDC-2022) | This paper summarizes the Crowdsensing-based Road Damage Detection Challenge (CRDDC), a Big Data Cup organized as a part of the IEEE International Conference on Big Data'2022. The Big Data Cup challenges involve a released dataset and a well-defined problem with clear evaluation metrics. The challenges run on a data competition platform that maintains a real-time online evaluation system for the participants. In the presented case, the data constitute 47,420 road images collected from India, Japan, the Czech Republic, Norway, the United States, and China to propose methods for automatically detecting road damages in these countries. More than 60 teams from 19 countries registered for this competition. The submitted solutions were evaluated using five leaderboards based on performance for unseen test images from the aforementioned six countries. This paper encapsulates the top 11 solutions proposed by these teams. The best-performing model utilizes ensemble learning based on YOLO and Faster-RCNN series models to yield an F1 score of 76% for test data combined from all 6 countries. The paper concludes with a comparison of current and past challenges and provides direction for the future. | ['Yoshihide Sekimoto', 'Takehiro Kashiyama', 'Hiroshi Omata', 'Durga Toshniwal', 'Sanjay Kumar Ghosh', 'Hiroya Maeda', 'Deeksha Arya'] | 2022-11-21 | null | null | null | null | ['road-damage-detection'] | ['computer-vision'] | [-3.34783852e-01 -6.97065890e-02 2.67467380e-01 1.29740849e-01
-1.26780915e+00 -4.98856097e-01 7.15109646e-01 1.16638236e-01
-5.28819621e-01 7.32187867e-01 6.34328783e-01 2.20591232e-01
1.25678435e-01 -7.96018362e-01 -5.25934279e-01 -6.55724645e-01
-5.52785695e-02 3.22908282e-01 5.00587285e-01 -4.45928395e-01
2.57141054e-01 3.50345731e-01 -1.87854540e+00 5.45007885e-01
1.24465191e+00 8.29009652e-01 -2.66913567e-02 7.61893034e-01
3.55271041e-01 1.00257206e+00 -9.66274381e-01 -4.18990344e-01
3.10723275e-01 2.40256503e-01 -6.62888527e-01 -3.09957683e-01
5.69639325e-01 -2.55183633e-02 -2.27734163e-01 5.77389717e-01
1.20029354e+00 2.57216040e-02 3.81626070e-01 -1.21434331e+00
-9.15004849e-01 2.33234644e-01 -3.40745836e-01 5.17848015e-01
4.48936790e-01 4.48460169e-02 6.88033879e-01 -1.09952044e+00
7.29713976e-01 1.00462806e+00 8.96424055e-01 3.29187423e-01
-4.79000717e-01 -1.97334126e-01 -2.63604205e-02 6.76422536e-01
-1.47718000e+00 -2.76205420e-01 2.61689156e-01 -7.18798995e-01
1.29503489e+00 3.36957306e-01 1.47604316e-01 1.02589321e+00
-7.61493817e-02 6.93381369e-01 1.13807595e+00 -3.09789330e-01
6.30890846e-01 1.64254252e-02 2.24644214e-01 3.36305261e-01
4.19536382e-01 1.35074675e-01 -4.74431336e-01 -1.90159410e-01
-2.96426415e-01 -2.86940932e-01 5.22330217e-02 2.28061914e-01
-9.65236783e-01 7.42286682e-01 4.96979028e-01 9.91088524e-02
-5.39544642e-01 -1.60898775e-01 4.49251831e-01 2.11686388e-01
9.71544564e-01 -8.84092897e-02 -1.66569337e-01 -1.74152046e-01
-3.81829113e-01 6.47785187e-01 7.31955588e-01 5.84881425e-01
5.07930696e-01 -1.02991005e-02 -2.92702973e-01 1.12373018e+00
1.33985281e-01 8.13466430e-01 5.63239716e-02 -1.02600849e+00
1.17331922e+00 6.68416858e-01 5.44101179e-01 -1.06114972e+00
-7.01230109e-01 5.99344261e-02 -6.38915956e-01 1.64989144e-01
3.74300092e-01 -7.48826504e-01 -7.45356917e-01 9.38934982e-01
4.98245031e-01 2.98190325e-01 1.35992974e-01 1.06127727e+00
1.17109573e+00 5.74799597e-01 1.16006210e-02 3.00051242e-01
1.27446413e+00 -1.07578027e+00 -5.81286371e-01 -4.09264080e-02
8.14881027e-01 -5.46780407e-01 8.02295744e-01 5.85255146e-01
-7.77469397e-01 -6.66451454e-01 -8.15911055e-01 -1.21807031e-01
-1.08653069e+00 2.23300368e-01 -2.54148878e-02 8.76088738e-01
-1.07467282e+00 3.61954600e-01 -2.30840012e-01 -6.27631903e-01
5.77666163e-01 -5.18090501e-02 -3.38271290e-01 -3.22962582e-01
-1.33956695e+00 1.07219505e+00 -1.28650397e-01 4.63828444e-01
-9.98052239e-01 -5.98174870e-01 -5.31606615e-01 -6.47422075e-01
-7.74871930e-02 -1.66373834e-01 8.29042852e-01 -1.62463915e-02
-9.32168722e-01 1.16001976e+00 1.14506662e-01 -8.10674608e-01
6.65139318e-01 -6.49913192e-01 -9.80892837e-01 1.98468082e-02
5.53743005e-01 4.37537223e-01 3.64493132e-01 -1.13620532e+00
-1.25506485e+00 -5.07611036e-01 -8.23342651e-02 8.62819627e-02
-5.61536588e-02 5.03128111e-01 -3.23984712e-01 -4.53001589e-01
-5.53865910e-01 -1.02033424e+00 -3.60255986e-01 -8.52856934e-01
-4.49512571e-01 -4.67932463e-01 9.58531141e-01 -7.67803729e-01
1.16095090e+00 -2.12875557e+00 -3.82635504e-01 1.10875525e-01
2.17898428e-01 4.41511601e-01 -2.27623492e-01 8.93842041e-01
1.87774330e-01 2.04066813e-01 -1.92708269e-01 -3.86581093e-01
-1.54465005e-01 1.03561372e-01 -4.99993980e-01 7.15968072e-01
5.02428859e-02 7.33022273e-01 -6.98004723e-01 -1.13057017e-01
2.98754305e-01 3.90968651e-01 1.37858883e-01 -1.39217209e-02
3.31259161e-01 3.43247473e-01 -1.38970017e-01 6.47684455e-01
8.21790993e-01 3.06290835e-01 -5.36786675e-01 2.58708745e-01
-4.49130774e-01 -1.26275569e-01 -1.26813269e+00 1.28952503e+00
-2.03789666e-01 7.26955056e-01 -8.28691050e-02 -7.20437348e-01
1.13518667e+00 4.15399343e-01 4.53463376e-01 -9.41358209e-01
-2.40731090e-01 4.68045741e-01 -5.36209881e-01 -9.82924700e-01
7.34023333e-01 3.72255415e-01 -3.39124143e-01 3.00068170e-01
-5.17008483e-01 -1.29812807e-01 4.31879014e-01 1.03857599e-01
1.32350183e+00 -1.46601230e-01 -1.40282273e-01 -9.26442295e-02
5.13944685e-01 2.18466774e-01 6.49193585e-01 8.33445847e-01
-8.21727514e-01 1.13696730e+00 1.90684810e-01 -1.02627134e+00
-9.18694079e-01 -7.45574176e-01 5.49698025e-02 1.14053369e+00
1.22963399e-01 -2.91818261e-01 -9.83026743e-01 -5.10658801e-01
2.03380715e-02 4.04532313e-01 -8.58652472e-01 3.42217088e-01
-5.23117185e-01 -1.17380512e+00 1.04647374e+00 4.04715776e-01
8.48012447e-01 -1.19366610e+00 -5.22646606e-01 4.65724580e-02
-6.39102757e-01 -1.38401902e+00 2.03898400e-02 -3.73650670e-01
-4.02321607e-01 -1.22218084e+00 -7.79759049e-01 -7.20617175e-01
2.04232052e-01 6.26863718e-01 9.43332076e-01 -1.12098187e-01
-4.79981810e-01 2.86111236e-01 -6.00674450e-01 -1.07221484e+00
1.25055864e-01 1.60963878e-01 -5.96840940e-02 1.49601519e-01
6.91367209e-01 2.41644704e-03 -5.85724533e-01 7.67549098e-01
-3.48825842e-01 -6.81335628e-01 8.12339336e-02 1.50740579e-01
4.33311939e-01 -5.73425665e-02 1.06250715e+00 -5.39806962e-01
8.25959444e-01 -7.65729964e-01 -6.18882298e-01 9.76546630e-02
-3.89771700e-01 -5.96171200e-01 2.44713426e-01 2.99283743e-01
-1.16441190e+00 1.18376613e-01 -1.98327955e-02 2.60474533e-01
-3.81363451e-01 2.06349060e-01 -7.98172280e-02 1.03634715e-01
1.35071766e+00 -3.08744520e-01 -3.91036063e-01 -4.99887645e-01
3.43481034e-01 1.29128551e+00 7.64173806e-01 -4.25811440e-01
6.89545333e-01 8.12980771e-01 -5.21338582e-01 -1.08413017e+00
-8.07787597e-01 -9.93444324e-01 -5.45864582e-01 -4.50830758e-01
1.12740672e+00 -1.53530908e+00 -3.76237422e-01 1.08755147e+00
-1.35238636e+00 -2.81610012e-01 -4.92646992e-01 1.97165981e-01
-9.81261656e-02 1.85112327e-01 -3.43722820e-01 -9.15076137e-01
-5.71894884e-01 -8.06510687e-01 1.55558467e+00 2.90880591e-01
8.88682753e-02 -9.46701825e-01 7.54760087e-01 1.00117397e+00
5.57367504e-01 7.22209811e-01 -4.33087051e-02 -4.61850226e-01
-2.47646496e-01 -9.26757097e-01 -4.13363636e-01 4.29555833e-01
-2.49452084e-01 1.16562456e-01 -1.71734595e+00 -3.79365049e-02
-4.86260742e-01 -4.37869757e-01 1.17268050e+00 3.74036402e-01
8.33594501e-01 9.22436193e-02 -2.45654956e-01 -2.50783890e-01
1.23927653e+00 -1.36031777e-01 1.22490788e+00 6.76553547e-01
8.32352877e-01 8.94090176e-01 5.76147258e-01 5.03910542e-01
1.13630009e+00 5.07073700e-01 6.79524064e-01 5.76803554e-03
-1.59983948e-01 3.05554066e-02 5.25843561e-01 5.07371664e-01
-7.51005232e-01 -2.15163469e-01 -1.38889813e+00 1.03945315e+00
-1.85315859e+00 -1.28710294e+00 -1.03718984e+00 2.26823258e+00
2.43256286e-01 -1.71064109e-01 6.27270103e-01 3.07279587e-01
8.84182155e-01 3.65661502e-01 1.38564594e-02 -4.68205303e-01
-7.91745961e-01 -9.76354927e-02 7.42445290e-01 4.26058650e-01
-1.48718476e+00 8.76543224e-01 6.42089415e+00 7.28666186e-01
-6.88634932e-01 5.62190413e-01 3.37358445e-01 -1.43852502e-01
-2.09831074e-02 -2.08024248e-01 -1.04558647e+00 5.04914999e-01
1.38925803e+00 -2.19523031e-02 2.31352895e-01 5.88451445e-01
6.24024510e-01 -3.81321490e-01 -1.13958389e-01 7.19129860e-01
1.37568742e-01 -1.34538376e+00 -6.13632619e-01 3.22706044e-01
1.35067821e+00 1.19830859e+00 -2.04199776e-01 4.20398206e-01
4.54976141e-01 -1.04091990e+00 5.62844992e-01 6.01864457e-01
3.56555283e-01 -5.41024804e-01 1.07483160e+00 4.47376519e-01
-9.30231452e-01 -6.62247658e-01 -7.86837399e-01 -4.45153028e-01
2.09169954e-01 8.83989573e-01 -6.27555788e-01 8.57994854e-01
1.22153163e+00 9.17788506e-01 -8.77919912e-01 1.33734131e+00
-3.81677955e-01 8.32554221e-01 -3.18563670e-01 1.18736133e-01
1.79789394e-01 1.73591554e-01 4.50604618e-01 1.28883314e+00
3.18225741e-01 -1.36471596e-02 1.95990443e-01 -1.39849350e-01
-1.04503855e-01 7.94600248e-02 -1.03917909e+00 7.83306301e-01
5.11620641e-01 1.49380088e+00 -1.39928848e-01 -1.20524675e-01
-4.06876266e-01 5.85906029e-01 1.89103112e-01 3.36793780e-01
-7.24301815e-01 -4.44973052e-01 4.63787854e-01 3.74796763e-02
4.82071005e-02 7.70145059e-02 -5.94708562e-01 -7.99254835e-01
2.12737098e-01 -6.93340242e-01 6.46571100e-01 -8.71168733e-01
-1.28199637e+00 6.10192299e-01 -1.97884530e-01 -1.59312594e+00
1.73527300e-01 -4.39717144e-01 -9.03854132e-01 9.62330222e-01
-1.67133689e+00 -1.28160667e+00 -9.43391502e-01 5.67423344e-01
4.99994189e-01 -3.95626336e-01 7.74865687e-01 6.43957734e-01
-9.18380260e-01 1.38273090e-01 2.59113312e-01 2.55061034e-02
8.17389667e-01 -1.36726010e+00 9.92337525e-01 1.03244650e+00
-5.47127984e-02 -3.05452943e-01 3.71607751e-01 -6.57244384e-01
-5.95049024e-01 -1.57124662e+00 1.50085151e+00 -1.33008873e+00
4.20620441e-01 -1.98233992e-01 -7.14868367e-01 2.44438961e-01
3.13981831e-01 5.08256137e-01 5.78669369e-01 -2.15544980e-02
-2.89048016e-01 -4.22446191e-01 -1.19186270e+00 3.80972214e-02
7.98787713e-01 -5.69527686e-01 -2.75725961e-01 7.29954898e-01
3.52720827e-01 -4.67981905e-01 -7.97723413e-01 1.27050862e-01
2.92622209e-01 -1.08438766e+00 8.80381107e-01 -5.89255869e-01
3.79193336e-01 -4.57339585e-01 -4.21204269e-01 -1.39817631e+00
7.47166807e-03 -5.42720318e-01 1.61428079e-01 1.18472028e+00
4.79091406e-01 -5.60783565e-01 5.51950276e-01 5.82631350e-01
-3.83246124e-01 -2.69614369e-01 -1.29523170e+00 -9.28160667e-01
3.03827792e-01 -1.05066860e+00 4.99842077e-01 4.83986497e-01
-3.38083565e-01 2.91528136e-01 -5.52454114e-01 5.83592355e-01
7.37196386e-01 -4.68756735e-01 1.08202004e+00 -1.58009040e+00
7.36512005e-01 -6.66746357e-03 -6.42544925e-01 -1.85478300e-01
-2.77201772e-01 -5.38612723e-01 1.03912443e-01 -2.04292917e+00
-2.59678632e-01 -3.13765407e-01 -1.52175128e-01 5.77927232e-01
-1.96258366e-01 5.89056730e-01 6.13829307e-02 3.79686326e-01
-9.03852522e-01 2.68040687e-01 7.46475339e-01 -2.90870488e-01
-1.38590574e-01 1.17230214e-01 -7.97654212e-01 4.57038760e-01
1.06757367e+00 -4.28997338e-01 -1.06417909e-01 -7.14865685e-01
4.37597483e-01 -6.20085776e-01 6.81418419e-01 -1.36746836e+00
3.38857472e-01 9.56848189e-02 4.74040210e-02 -8.55515659e-01
3.51021700e-02 -4.65717345e-01 -6.11378662e-02 2.59834617e-01
4.59989969e-04 7.23899379e-02 1.76065132e-01 5.51174104e-01
-1.98307335e-01 7.08908364e-02 4.99162376e-01 2.40251213e-01
-1.15414584e+00 -1.63651854e-01 -2.81365573e-01 4.72418278e-01
1.40706134e+00 -2.03674942e-01 -1.09057713e+00 -2.32162476e-01
-5.98047614e-01 6.36072934e-01 -1.50435060e-01 9.94531989e-01
5.65329909e-01 -1.29792440e+00 -1.55842781e+00 -1.34982079e-01
5.88509798e-01 -2.38440838e-02 6.66436434e-01 1.04313099e+00
-4.67106104e-01 3.80979866e-01 -1.91297587e-02 -3.93041581e-01
-1.19012260e+00 2.68697441e-01 2.50657856e-01 -3.86570632e-01
-4.61445391e-01 4.70379710e-01 -5.17325163e-01 -1.03898263e+00
1.26817832e-02 -8.94259810e-02 -7.02010095e-01 4.75772560e-01
1.08868361e+00 1.35810125e+00 7.13508606e-01 -8.44743550e-01
-5.82091749e-01 6.76348746e-01 3.86699140e-01 -7.89633468e-02
1.48419166e+00 -3.78385186e-01 1.50990471e-01 2.75066257e-01
8.54588270e-01 1.20596401e-01 -9.24318135e-01 -8.22643712e-02
2.73160338e-01 -1.39213443e-01 -1.40019655e-01 -1.21036792e+00
-9.26997066e-01 9.76172745e-01 1.08715928e+00 3.97084415e-01
8.09214890e-01 -7.72772282e-02 8.49149823e-01 4.44487661e-01
6.01429343e-01 -1.82632148e+00 -2.46821582e-01 6.24016583e-01
1.43591893e+00 -1.47724354e+00 -2.34167561e-01 -2.79416051e-02
-1.04779530e+00 6.95321441e-01 3.67420137e-01 -2.74720073e-01
6.17366374e-01 -1.08209923e-01 4.01999235e-01 -3.40007514e-01
-6.00191474e-01 -6.50993466e-01 2.15173349e-01 1.19650388e+00
-4.71333452e-02 2.30514690e-01 -2.49304652e-01 5.12493074e-01
3.04896794e-02 1.87296078e-01 7.95522332e-01 6.90743685e-01
-8.30674767e-01 -6.83959007e-01 -8.16280425e-01 2.60555297e-01
-4.16956961e-01 2.39977479e-01 -6.62692308e-01 5.63457131e-01
4.82241422e-01 1.77616966e+00 -1.15270264e-01 -6.23403609e-01
1.09529006e+00 -4.23729271e-01 -3.10276896e-01 -4.79116887e-01
-7.87545860e-01 -3.71871769e-01 6.78889811e-01 -6.53480113e-01
-5.22107840e-01 -8.16436768e-01 -1.19415784e+00 -4.77997899e-01
7.41455182e-02 -9.37106321e-04 7.93603182e-01 8.04515123e-01
8.05694163e-01 -4.51633614e-03 1.02458644e+00 -8.42444599e-01
-2.81748801e-01 -1.10267782e+00 -5.41819751e-01 2.91891873e-01
2.42659211e-01 -6.11365855e-01 -3.02519619e-01 -4.84458841e-02] | [7.403700351715088, 1.0739742517471313] |
d0c0b60e-a52d-4b9e-93ba-b3e04c9c8e13 | gcdf1-a-goal-and-context-driven-f-score-for | null | null | https://aclanthology.org/2021.eancs-1.2 | https://aclanthology.org/2021.eancs-1.2.pdf | GCDF1: A Goal- and Context- Driven F-Score for Evaluating User Models | The evaluation of dialogue systems in interaction with simulated users has been proposed to improve turn-level, corpus-based metrics which can only evaluate test cases encountered in a corpus and cannot measure system’s ability to sustain multi-turn interactions. Recently, little emphasis was put on automatically assessing the quality of the user model itself, so unless correlations with human studies are measured, the reliability of user model based evaluation is unknown. We propose GCDF1, a simple but effective measure of the quality of semantic-level conversations between a goal-driven user agent and a system agent. In contrast with previous approaches we measure the F-score at dialogue level and consider user and system behaviours to improve recall and precision estimation. We facilitate scores interpretation by providing a rich hierarchical structure with information about conversational patterns present in the test data and tools to efficiently query the conversations generated. We apply our framework to assess the performance and weaknesses of a Convlab2 user model. | ['Bill Byrne', 'Bo-Hsiang Tseng', 'Alexandru Coca'] | null | null | null | null | eancs-2021-11 | ['dialogue-evaluation'] | ['natural-language-processing'] | [-9.36880037e-02 5.58593333e-01 3.88275236e-01 -5.49492955e-01
-9.61170793e-01 -9.12172914e-01 1.19374096e+00 2.73572862e-01
-3.51365268e-01 8.09600532e-01 5.27436435e-01 -4.28182274e-01
-7.94267282e-02 -4.75354016e-01 -8.16474259e-02 -1.34767279e-01
-3.12845744e-02 8.97832394e-01 3.62064183e-01 -5.64672887e-01
3.24782610e-01 -2.20499821e-02 -1.55251086e+00 7.49106824e-01
9.47359085e-01 5.05483031e-01 3.26143354e-01 1.33277965e+00
-3.44685465e-02 1.10913372e+00 -1.26237011e+00 -2.81181961e-01
-2.56368995e-01 -8.91122282e-01 -1.36506212e+00 2.07122117e-01
1.10093564e-01 -4.12246019e-01 5.46270348e-02 6.45410538e-01
5.36707461e-01 9.13178921e-02 5.19048631e-01 -1.45751441e+00
2.50001758e-01 7.80087113e-01 4.70352054e-01 2.23102510e-01
1.10836768e+00 6.63139641e-01 1.04012120e+00 -4.01930362e-01
7.43539631e-01 1.48025882e+00 3.16933215e-01 4.78801906e-01
-1.09043801e+00 -1.21341228e-01 -2.95552194e-01 3.46418507e-02
-6.42139554e-01 -7.25264966e-01 1.06843216e-02 -5.99279881e-01
1.23825681e+00 5.11012316e-01 6.16814315e-01 1.17447770e+00
-2.41538569e-01 8.62956285e-01 9.87626970e-01 -4.65298384e-01
2.14099944e-01 6.17377877e-01 3.23406547e-01 7.04730392e-01
-4.48728621e-01 -1.73625685e-02 -6.71260476e-01 -4.00979847e-01
4.20981765e-01 -8.90740991e-01 -3.72208446e-01 -1.20642960e-01
-1.16284788e+00 6.96097076e-01 -6.46617115e-02 7.51329899e-01
-3.09437066e-01 -3.91234487e-01 7.34896183e-01 7.98434973e-01
1.78484976e-01 1.00114071e+00 -6.38356805e-01 -1.11970782e+00
-5.57915986e-01 5.78823268e-01 1.58663929e+00 9.55173850e-01
3.31874579e-01 -3.64275038e-01 -5.85539043e-01 9.68992710e-01
2.69733012e-01 1.72326773e-01 6.18238747e-01 -1.26191258e+00
4.09661889e-01 9.54322398e-01 3.85166734e-01 -4.83182728e-01
-5.59702873e-01 2.88651492e-02 -3.36561650e-02 2.32442290e-01
9.03038859e-01 -3.55752081e-01 -3.08596473e-02 1.68438864e+00
2.00799391e-01 -4.78690356e-01 4.00445551e-01 5.85637748e-01
1.00346673e+00 3.29038143e-01 -8.28191563e-02 -3.82991374e-01
1.23233652e+00 -9.94857728e-01 -7.49624670e-01 1.82040066e-01
1.18398750e+00 -8.99957001e-01 1.53629136e+00 3.48981053e-01
-1.29900241e+00 -4.77554888e-01 -9.24407780e-01 3.05391848e-01
-1.80030137e-01 -2.15987697e-01 2.96540707e-01 8.33581388e-01
-1.31013894e+00 4.26389843e-01 -4.57912505e-01 -6.05653167e-01
-4.07846868e-01 1.42734572e-01 -9.37024727e-02 4.11767304e-01
-1.17845750e+00 1.28717208e+00 3.40657718e-02 -3.94225448e-01
-8.11326683e-01 -3.37450027e-01 -7.88249850e-01 7.46018291e-02
1.20075285e-01 -3.24018329e-01 2.18992567e+00 -6.74153805e-01
-1.96516097e+00 6.73044264e-01 9.63140652e-02 -3.78299028e-01
8.27453375e-01 2.18121558e-01 -1.43864021e-01 -3.40268351e-02
-7.08644539e-02 3.60559106e-01 -1.32426023e-01 -1.03675568e+00
-9.10936713e-01 -1.00604564e-01 6.57043636e-01 5.51546574e-01
4.55602184e-02 8.01208317e-02 -2.11180523e-01 2.67264992e-01
-3.68414104e-01 -8.23004425e-01 9.88911316e-02 -6.83727384e-01
-2.13063478e-01 -6.06329441e-01 5.59837639e-01 -5.52085698e-01
1.23788309e+00 -1.80261683e+00 -1.23085961e-01 2.59527028e-01
5.75481728e-02 3.53362322e-01 -2.96982139e-01 7.59247482e-01
3.63735706e-01 2.70037707e-02 -4.75927815e-02 -1.95126086e-01
3.98548782e-01 -1.14904717e-01 1.73034623e-01 -9.89006981e-02
-1.03090897e-01 6.50613904e-01 -1.12072432e+00 -4.36389863e-01
4.81903881e-01 1.53184012e-01 -6.19988799e-01 8.47113669e-01
-4.70830142e-01 4.99789298e-01 -3.18265826e-01 2.31020935e-02
-4.61143963e-02 -1.25882804e-01 3.16702932e-01 1.87275440e-01
-2.22120255e-01 9.84367967e-01 -7.86888719e-01 1.46023989e+00
-8.57714117e-01 7.23631680e-01 5.81311621e-03 -4.55332160e-01
7.15852737e-01 7.47984171e-01 9.14828330e-02 -1.00284553e+00
3.20872068e-01 -1.52451359e-02 5.12667894e-01 -8.16217244e-01
6.37655556e-01 3.10040116e-01 -1.29272401e-01 8.00078094e-01
2.65384942e-01 -4.10250515e-01 6.34977341e-01 4.33320612e-01
1.43644977e+00 -1.53009310e-01 3.24958682e-01 -3.62257034e-01
6.26013100e-01 6.02591336e-02 -2.42853850e-01 8.54302764e-01
-3.43660682e-01 2.58270472e-01 9.20048594e-01 -2.63597872e-02
-7.98680961e-01 -5.95834553e-01 -4.91824485e-02 1.35570264e+00
-2.03611165e-01 -7.08199143e-01 -1.36063695e+00 -8.50157678e-01
-5.23079216e-01 1.02374911e+00 -3.11083764e-01 6.11353070e-02
-2.31712103e-01 -5.38226068e-01 7.02396393e-01 -9.78939608e-02
4.40636843e-01 -1.40348041e+00 -8.55565965e-01 4.24404830e-01
-6.23888195e-01 -1.14275551e+00 -2.38128096e-01 -2.49418229e-01
-5.25808871e-01 -1.54210913e+00 -1.55351669e-01 -4.47555631e-01
1.63199499e-01 -7.56044611e-02 1.61678326e+00 5.19056201e-01
7.05095101e-03 6.92823470e-01 -6.81926489e-01 -1.78843170e-01
-1.32282138e+00 9.94236320e-02 -1.38877437e-01 -5.07990241e-01
2.80811995e-01 -3.28039318e-01 -4.78705347e-01 8.09241831e-01
-5.12316942e-01 9.43358988e-02 3.89370739e-01 8.08712006e-01
-5.76258540e-01 -3.16085190e-01 6.43636227e-01 -9.45202649e-01
1.52612782e+00 -1.44953877e-01 -2.87970930e-01 3.98406833e-01
-6.43196464e-01 8.91019702e-02 3.10535491e-01 -1.13447979e-01
-9.83896613e-01 -4.14601684e-01 -3.36406142e-01 4.73525047e-01
-3.77568483e-01 4.28597748e-01 -1.87655002e-01 2.98028111e-01
9.87754464e-01 -6.23409450e-02 2.25896463e-01 -5.34990914e-02
2.08573386e-01 1.10260820e+00 2.61627167e-01 -7.24243224e-01
-2.20793486e-02 -3.92251432e-01 -8.62696528e-01 -1.05979347e+00
-4.44393069e-01 -6.14324450e-01 -4.29943144e-01 -6.90811396e-01
6.41730249e-01 -5.40575624e-01 -1.31580520e+00 1.72909126e-01
-1.30307961e+00 -9.01230812e-01 4.69528465e-03 3.25864255e-01
-9.95366096e-01 2.44455889e-01 -5.13585746e-01 -1.31653988e+00
-2.11069003e-01 -1.38317871e+00 9.14956987e-01 4.80364002e-02
-1.04921675e+00 -1.04182231e+00 3.33903402e-01 5.58779061e-01
4.35005099e-01 -1.22123100e-01 7.67821550e-01 -1.10545027e+00
-1.69198126e-01 -1.16285503e-01 -1.26488348e-02 3.10493827e-01
1.28243733e-02 -1.26494560e-02 -1.09900546e+00 -1.99060395e-01
1.32647958e-02 -7.57718563e-01 -1.34055063e-01 -5.18083125e-02
4.82016474e-01 -5.00321090e-01 2.83220634e-02 -5.07722676e-01
6.16981864e-01 3.80094200e-01 5.67928553e-01 2.93997467e-01
3.03991009e-02 1.01682580e+00 6.55865729e-01 4.18601960e-01
4.32841957e-01 8.20577562e-01 3.80425416e-02 2.04338849e-01
9.84579176e-02 -6.70426190e-02 4.74987179e-01 8.50329041e-01
9.82738733e-02 -6.09408259e-01 -9.39911664e-01 4.48692203e-01
-1.82267368e+00 -9.47992206e-01 -3.08966219e-01 2.18381929e+00
9.22593057e-01 5.97005546e-01 5.82987487e-01 1.65255517e-01
5.10298252e-01 -1.76833361e-01 -8.10415857e-03 -5.63999712e-01
4.47458357e-01 -1.88819423e-01 -2.48477802e-01 1.05962300e+00
-5.10320842e-01 7.16234446e-01 7.09339857e+00 4.72990692e-01
-4.39101309e-01 6.79162666e-02 5.04060030e-01 1.02039196e-01
3.81463468e-02 -1.62794307e-01 -4.92532492e-01 4.49181050e-01
1.51252913e+00 -3.48893404e-01 3.87058675e-01 6.30070150e-01
5.55265188e-01 -3.53200734e-01 -1.65287995e+00 6.17073119e-01
-8.16802159e-02 -9.03123438e-01 -3.07382047e-01 1.40366197e-01
1.91017389e-01 2.08424982e-02 -3.59976411e-01 8.85774851e-01
6.72047079e-01 -7.92025685e-01 4.32365894e-01 3.57472837e-01
3.39921713e-01 -5.84480047e-01 8.92029226e-01 7.82002807e-01
-5.72824299e-01 3.33444804e-01 1.25874728e-01 -2.99640536e-01
9.88714956e-03 -6.94433898e-02 -1.82554781e+00 2.46476606e-02
2.71216720e-01 -1.79981515e-01 -4.93599296e-01 8.29730928e-01
-8.41957107e-02 5.03516018e-01 -2.97441542e-01 -5.96947253e-01
3.29920292e-01 -8.94113630e-03 5.31679869e-01 1.47354209e+00
-5.89898899e-02 1.02718636e-01 3.29617977e-01 5.39590836e-01
1.90157562e-01 2.39587411e-01 -6.36924088e-01 -8.90375897e-02
5.74169040e-01 1.32450902e+00 -4.34392959e-01 -4.82293427e-01
-9.85188931e-02 7.31668293e-01 1.90160185e-01 -3.77064683e-02
-5.42355061e-01 -1.61598548e-01 6.09109461e-01 4.76241782e-02
-4.59337354e-01 -1.53506786e-01 5.49152717e-02 -5.83019316e-01
-6.28429055e-02 -1.39208317e+00 2.13114232e-01 -6.33481324e-01
-8.14313293e-01 9.72648203e-01 -4.75557446e-02 -1.08483267e+00
-1.09249973e+00 -4.82646316e-01 -6.22766614e-01 9.03557718e-01
-6.21922493e-01 -4.87245172e-01 -3.74735892e-01 2.29679883e-01
1.02385652e+00 -4.08729047e-01 1.21360958e+00 -8.64417776e-02
-2.47171089e-01 6.21093988e-01 -2.69797713e-01 -5.72332703e-02
7.07389951e-01 -1.58591485e+00 4.39872593e-01 1.46490633e-01
-2.07147133e-02 4.95491058e-01 1.27567458e+00 -3.80167305e-01
-9.12041008e-01 -5.58303475e-01 8.30053210e-01 -9.49287117e-01
6.60649896e-01 -2.22984120e-01 -8.42830658e-01 3.00437719e-01
6.26371562e-01 -8.89344990e-01 7.30934680e-01 5.14547646e-01
-8.89300089e-03 5.03176570e-01 -1.19749475e+00 5.67060292e-01
8.72593284e-01 -6.53714657e-01 -7.57599115e-01 5.03535509e-01
5.25430501e-01 -5.39983749e-01 -8.51652861e-01 1.41514450e-01
5.30071318e-01 -1.38374650e+00 3.89292121e-01 -5.35503864e-01
3.17384839e-01 1.20716907e-01 2.97694802e-02 -1.58458197e+00
2.63600558e-01 -8.12118590e-01 2.10937545e-01 1.06812453e+00
8.52008939e-01 -2.20604137e-01 5.16766429e-01 9.44557726e-01
-8.60226601e-02 -3.45203906e-01 -6.28324986e-01 -3.98670495e-01
-2.90354878e-01 -3.76379281e-01 4.65767026e-01 7.88534284e-01
9.04523373e-01 9.26620245e-01 1.99204832e-02 -1.49807021e-01
2.29354784e-01 -4.15622413e-01 1.20430875e+00 -1.17372477e+00
-4.18241560e-01 -7.98452020e-01 -2.72702515e-01 -8.61563146e-01
6.18558899e-02 -5.57141662e-01 4.16108698e-01 -1.42020857e+00
-1.04962371e-01 -1.73502296e-01 2.49128059e-01 3.52065600e-02
4.68640253e-02 -1.44570321e-01 1.02021150e-01 5.97655252e-02
-1.00782788e+00 3.56074393e-01 1.13484228e+00 1.48083732e-01
-4.86921370e-01 1.43167764e-01 -4.38801795e-01 6.48277640e-01
7.30653465e-01 -5.41441254e-02 -6.20570302e-01 -5.52472565e-03
1.68307684e-02 5.41909933e-01 3.22379582e-02 -1.18461251e+00
2.52774507e-01 6.58576041e-02 -1.90188289e-01 -2.28624150e-01
9.01591405e-02 -4.90564942e-01 -3.02676141e-01 2.69823283e-01
-8.39613259e-01 1.34405866e-01 5.11194654e-02 1.45523190e-01
-2.09849253e-01 -3.26923549e-01 5.23123145e-01 -2.90518671e-01
-3.44236761e-01 -3.61320585e-01 -8.83058548e-01 4.45002951e-02
7.23597646e-01 7.10853636e-02 -3.76089126e-01 -8.26938570e-01
-7.89032876e-01 6.22516870e-01 2.88274795e-01 4.82373476e-01
1.87475011e-01 -7.61572242e-01 -7.36826301e-01 -4.65224823e-03
4.24200892e-01 -4.93192971e-01 4.11678525e-03 6.52669132e-01
-5.64488292e-01 5.72017252e-01 -1.51252031e-01 -7.45193720e-01
-1.60633361e+00 -1.50478095e-01 5.89684665e-01 -5.04544079e-01
-2.52682239e-01 5.93335271e-01 5.82154654e-02 -9.11281109e-01
5.28334916e-01 -2.33524442e-01 -5.28700531e-01 1.59452915e-01
8.56764495e-01 3.14187109e-01 4.78585541e-01 -3.99353266e-01
5.14894612e-02 -4.76320446e-01 -5.98457456e-02 -7.25968301e-01
9.00074720e-01 -1.67267218e-01 1.99328035e-01 7.67257512e-01
9.06818509e-01 -3.04484218e-01 -8.67767990e-01 -5.94223179e-02
2.37419561e-01 -1.84183717e-01 -1.68465585e-01 -1.28934920e+00
-1.05188660e-01 6.61784232e-01 5.86498916e-01 1.06826270e+00
3.32945466e-01 3.67504335e-03 3.38166654e-01 8.00104976e-01
4.61084485e-01 -1.36396348e+00 3.17629814e-01 9.12275553e-01
1.19398761e+00 -1.32525206e+00 -5.07395148e-01 -2.24495009e-01
-8.37744057e-01 9.23080921e-01 9.01210606e-01 3.80321592e-01
1.44599423e-01 2.71363258e-01 4.20958668e-01 -3.54936540e-01
-1.26775634e+00 -3.64768505e-01 3.63104641e-02 4.62873220e-01
1.07428932e+00 1.16472803e-01 -5.21382928e-01 3.53090167e-01
-8.31738710e-01 -1.21156164e-01 6.69966221e-01 8.51079464e-01
-6.71250045e-01 -1.11475515e+00 -8.80457461e-02 4.15796697e-01
-9.01779830e-02 1.71301231e-01 -8.71161819e-01 7.83624291e-01
-5.65754175e-01 1.56898654e+00 -3.24157663e-02 -5.00750840e-01
8.89496386e-01 5.47927022e-01 4.04517293e-01 -7.87535906e-01
-1.03081024e+00 -2.67610431e-01 9.33502376e-01 -5.53613722e-01
-3.44431758e-01 -7.11649001e-01 -9.22850430e-01 -3.15148979e-01
-3.63282621e-01 8.18063140e-01 7.42961466e-01 1.11754251e+00
7.09705725e-02 6.74089789e-01 6.13695860e-01 -6.65201902e-01
-6.74312413e-01 -1.66254044e+00 -1.08738765e-01 5.27917206e-01
9.79499817e-02 -5.10066569e-01 -4.66896802e-01 -1.42752185e-01] | [12.871386528015137, 8.05367660522461] |
ea522497-631a-4f02-bebf-fad3739a5912 | domain-mismatch-doesn-t-always-prevent-cross | 2211.16671 | null | https://arxiv.org/abs/2211.16671v1 | https://arxiv.org/pdf/2211.16671v1.pdf | Domain Mismatch Doesn't Always Prevent Cross-Lingual Transfer Learning | Cross-lingual transfer learning without labeled target language data or parallel text has been surprisingly effective in zero-shot cross-lingual classification, question answering, unsupervised machine translation, etc. However, some recent publications have claimed that domain mismatch prevents cross-lingual transfer, and their results show that unsupervised bilingual lexicon induction (UBLI) and unsupervised neural machine translation (UNMT) do not work well when the underlying monolingual corpora come from different domains (e.g., French text from Wikipedia but English text from UN proceedings). In this work, we show that a simple initialization regimen can overcome much of the effect of domain mismatch in cross-lingual transfer. We pre-train word and contextual embeddings on the concatenated domain-mismatched corpora, and use these as initializations for three tasks: MUSE UBLI, UN Parallel UNMT, and the SemEval 2017 cross-lingual word similarity task. In all cases, our results challenge the conclusions of prior work by showing that proper initialization can recover a large portion of the losses incurred by domain mismatch. | ['Noah A. Smith', 'Phillip Keung', 'Daniel Edmiston'] | 2022-11-30 | null | null | null | null | ['word-similarity', 'unsupervised-machine-translation', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-4.39894572e-03 -1.92300856e-01 -5.02766609e-01 -3.50107700e-01
-1.44455945e+00 -9.51753974e-01 8.50764990e-01 3.10228765e-02
-8.04554760e-01 1.15442002e+00 4.21739012e-01 -5.82368910e-01
3.17756832e-01 -4.05535907e-01 -9.43936229e-01 -4.04510647e-01
4.77357775e-01 1.04275453e+00 7.28224590e-02 -4.85010028e-01
-8.02982897e-02 -9.77663845e-02 -1.11012781e+00 3.87422919e-01
1.18367934e+00 3.48827034e-01 1.02119409e-01 1.76143065e-01
-3.60149890e-01 4.20436680e-01 -2.59651631e-01 -8.09723556e-01
3.86902004e-01 -6.87299728e-01 -1.14719045e+00 -4.26473439e-01
6.68566227e-01 -1.02483472e-02 -1.80924267e-01 1.11234176e+00
7.84608245e-01 -2.06199437e-02 8.99264693e-01 -1.02988625e+00
-8.73964787e-01 7.99941242e-01 -4.32697266e-01 1.35355309e-01
2.30759278e-01 -1.12440065e-01 1.05173731e+00 -1.18240845e+00
1.01275098e+00 1.28131759e+00 7.97162235e-01 4.73552644e-01
-1.42068434e+00 -8.50746334e-01 -3.34934801e-01 2.48242766e-01
-1.27559352e+00 -7.18876719e-01 5.03836691e-01 -3.71981382e-01
1.20402455e+00 -1.24760367e-01 -3.69900614e-02 1.55036831e+00
1.34978652e-01 7.83572555e-01 1.37016642e+00 -1.01120961e+00
-2.28865311e-01 5.76233387e-01 3.62062901e-02 3.18991840e-01
2.25191429e-01 1.30663782e-01 -5.53028524e-01 -9.55241844e-02
2.42478386e-01 -6.04574978e-01 -1.68663308e-01 -4.63936239e-01
-1.25930917e+00 9.63533998e-01 -3.04270606e-03 6.66591883e-01
1.24854065e-01 -3.75276238e-01 8.45919251e-01 9.72085059e-01
8.36758494e-01 5.83177507e-01 -8.70953560e-01 9.46895499e-03
-6.49704516e-01 -1.88711613e-01 8.18836212e-01 1.16838884e+00
1.09225130e+00 -2.61479467e-01 2.41015732e-01 1.32405126e+00
-1.82541713e-01 6.57801449e-01 9.63529408e-01 -4.37408209e-01
8.26567352e-01 2.42883638e-01 -3.45015861e-02 -4.32372540e-01
-3.61738726e-02 -4.15873565e-02 -4.97733653e-01 -3.07641774e-01
6.98071718e-01 -4.12727028e-01 -6.27432287e-01 2.17298198e+00
7.57092759e-02 -1.34319469e-01 5.09007752e-01 5.33107221e-01
5.50001860e-01 5.41380167e-01 1.86045840e-01 -9.91936028e-02
1.24455202e+00 -9.97453272e-01 -5.98391056e-01 -4.99535054e-01
1.20718372e+00 -1.22231090e+00 1.30399919e+00 -2.56494302e-02
-8.59493017e-01 -5.79013944e-01 -9.94020104e-01 -3.74267370e-01
-8.20423305e-01 4.41220999e-02 4.25811946e-01 4.81945455e-01
-8.57836902e-01 4.52323377e-01 -4.57134068e-01 -1.00181210e+00
4.04787809e-02 1.54755726e-01 -7.18986750e-01 -6.83042049e-01
-1.67476964e+00 1.53387558e+00 4.49105620e-01 -4.49981451e-01
-5.95453560e-01 -6.84964180e-01 -8.56151402e-01 -4.54687029e-01
8.06074217e-02 -3.62354517e-01 1.11551368e+00 -1.39089477e+00
-1.18897295e+00 1.64815116e+00 1.83103867e-02 -4.13033843e-01
4.23176378e-01 -3.68195087e-01 -5.19767225e-01 -3.14333230e-01
6.16148770e-01 7.23546445e-01 4.05761659e-01 -8.29897344e-01
-4.31753844e-01 -4.14358318e-01 -3.06988329e-01 4.31188196e-01
-3.59920382e-01 3.45239460e-01 -3.60784143e-01 -5.21610081e-01
-4.26629543e-01 -9.88965690e-01 1.06134087e-01 -6.81309998e-01
-3.70758697e-02 -4.71776128e-01 4.08899754e-01 -8.81296635e-01
7.18509793e-01 -2.05643678e+00 2.31867969e-01 -2.29503021e-01
-5.80923915e-01 4.45553482e-01 -5.92960119e-01 7.74098873e-01
-2.36077160e-01 6.61162511e-02 -2.57814586e-01 -2.34990105e-01
8.36831182e-02 5.55146933e-01 -3.88515711e-01 5.13763130e-01
2.95000434e-01 1.05633712e+00 -9.36364174e-01 -6.32102311e-01
-9.45305526e-02 3.43211502e-01 -3.89486700e-01 9.13271755e-02
-4.41527553e-02 5.46293437e-01 7.68011212e-02 3.06554914e-01
5.33751428e-01 1.88715234e-01 5.16510725e-01 -1.04927786e-01
-6.49889559e-02 8.86174023e-01 -4.87876952e-01 2.18149829e+00
-6.55876875e-01 6.61780417e-01 -1.78803086e-01 -1.18612432e+00
7.49392867e-01 4.77594882e-01 2.05844283e-01 -1.17316389e+00
1.14644378e-01 7.60487080e-01 1.22893974e-01 -3.95316631e-01
3.26639444e-01 -6.34215176e-01 -4.00076598e-01 7.74774253e-01
6.21104181e-01 -1.76685616e-01 1.24583319e-01 -4.38554399e-02
6.90919399e-01 3.29857975e-01 2.60950178e-01 -5.37201285e-01
3.86111557e-01 3.86249661e-01 4.66784418e-01 4.21824753e-01
-2.20044926e-01 3.94107282e-01 1.52421519e-01 -2.04221979e-01
-1.32875478e+00 -1.17885363e+00 -3.09024870e-01 1.45396471e+00
-2.63961032e-02 -1.89877450e-01 -8.35451126e-01 -9.04812813e-01
-9.82428715e-02 9.55929637e-01 -4.60251629e-01 -2.63573974e-01
-8.55069637e-01 -8.44139397e-01 9.05855536e-01 2.19498664e-01
1.21705130e-01 -1.07269907e+00 1.25848219e-01 2.56053984e-01
-5.74047089e-01 -1.35166407e+00 -7.44673908e-01 4.54491109e-01
-7.68891394e-01 -8.87756228e-01 -7.08649337e-01 -1.44632137e+00
4.06041682e-01 1.16844088e-01 1.63606870e+00 -4.50737596e-01
-9.68713462e-02 3.17791134e-01 -3.46973956e-01 -1.80227980e-01
-6.67003930e-01 4.89446193e-01 4.66662854e-01 -3.34546149e-01
1.17095423e+00 -5.97462893e-01 1.34495735e-01 4.20185030e-01
-7.06845343e-01 -8.19488391e-02 5.36294103e-01 1.14725733e+00
5.08712888e-01 -5.49000204e-01 8.02536905e-01 -1.10977972e+00
7.59972751e-01 -5.39721966e-01 -3.63104552e-01 5.83279073e-01
-4.47655499e-01 3.43933344e-01 6.14413619e-01 -6.37210131e-01
-9.69037890e-01 -2.25357577e-01 -1.07222334e-01 -2.52830267e-01
-2.31646001e-01 4.16846693e-01 -2.68642187e-01 1.78474680e-01
8.93498838e-01 1.86000526e-01 -1.81548387e-01 -7.12079287e-01
6.47457719e-01 8.54687095e-01 5.00980794e-01 -8.99147272e-01
5.93897104e-01 -1.35227535e-02 -7.70127654e-01 -5.38654029e-01
-9.49177265e-01 -4.67370808e-01 -1.10209441e+00 3.30711126e-01
9.99659538e-01 -1.20230091e+00 2.61654556e-01 2.14878976e-01
-1.38392162e+00 -5.17991185e-01 -2.24282935e-01 7.76046693e-01
-5.68243206e-01 9.91822332e-02 -6.11124933e-01 -1.07261352e-01
-2.99949199e-01 -9.78469551e-01 8.08222055e-01 -2.16709197e-01
-4.09672678e-01 -1.40886915e+00 6.83761358e-01 2.54311562e-01
3.31588030e-01 -1.78437129e-01 1.18944740e+00 -1.13439715e+00
-6.92680404e-02 -1.82217974e-02 -1.26744375e-01 6.05327845e-01
2.73961335e-01 -6.00328743e-01 -8.66716683e-01 -4.85642135e-01
5.43115381e-03 -8.89777780e-01 5.42864323e-01 -9.51547921e-02
5.47985211e-02 -1.11276843e-01 -2.21128106e-01 5.35544395e-01
1.57089949e+00 -7.90279508e-02 4.07792389e-01 3.66369992e-01
6.57363057e-01 8.43987823e-01 4.95658040e-01 -5.02016485e-01
4.66675848e-01 6.94328189e-01 -3.73566717e-01 -1.72749221e-01
-3.89963180e-01 -3.44193250e-01 8.04300487e-01 1.61912155e+00
2.62977630e-01 5.27335666e-02 -1.07594037e+00 1.09536064e+00
-1.60823750e+00 -6.61376894e-01 -1.20862490e-02 2.32833314e+00
1.35512209e+00 -1.13693587e-01 -7.92249441e-02 -5.42332053e-01
6.39762640e-01 -2.71140248e-01 -4.15406615e-01 -6.86028421e-01
-5.56625426e-01 6.04200959e-01 6.99106753e-01 6.93842471e-01
-1.05005872e+00 1.53507304e+00 6.19357967e+00 9.62433279e-01
-1.01937723e+00 8.46943080e-01 2.65648663e-01 1.42172992e-01
-3.60140681e-01 -3.69114541e-02 -7.77672946e-01 2.34269634e-01
1.18491220e+00 -4.41529512e-01 4.62660819e-01 6.66311681e-01
-5.33328831e-01 1.60666928e-01 -1.42946136e+00 9.42246258e-01
4.52621222e-01 -7.53635228e-01 -1.16193093e-01 -1.15389705e-01
1.04657352e+00 8.13312650e-01 -7.87916780e-02 6.96538806e-01
8.19738030e-01 -9.36066210e-01 3.15012366e-01 -1.90697670e-01
1.20830655e+00 -7.30142951e-01 7.93906629e-01 4.56302106e-01
-7.49637425e-01 5.75446248e-01 -6.60951793e-01 1.24812409e-01
3.88412774e-02 2.10133910e-01 -7.70172000e-01 6.26482129e-01
4.15471226e-01 6.55569315e-01 -6.02169812e-01 3.62393349e-01
-3.80203664e-01 6.71876132e-01 -2.40362242e-01 2.90474743e-01
4.15377647e-01 -3.40398192e-01 3.22832704e-01 1.47480714e+00
3.16387624e-01 -5.23759365e-01 9.16357487e-02 4.24940884e-01
-3.92892748e-01 6.92093611e-01 -1.13786244e+00 -1.38722494e-01
1.96155772e-01 7.43706763e-01 -1.93874463e-01 -4.42710638e-01
-7.73732007e-01 1.38228285e+00 8.49141359e-01 3.29545051e-01
-5.97687185e-01 -4.28368419e-01 8.11317682e-01 -1.45232022e-01
1.62235364e-01 -2.09468529e-01 -1.80999264e-01 -1.64934492e+00
-4.62141586e-03 -1.08199322e+00 3.36299211e-01 -4.75247979e-01
-1.83480501e+00 7.18274474e-01 -2.69591093e-01 -1.22635269e+00
-5.13057709e-01 -6.58342183e-01 -2.76258379e-01 1.14760971e+00
-1.55064964e+00 -1.28290474e+00 5.21195889e-01 7.18251705e-01
5.65235376e-01 -4.16266531e-01 1.16089392e+00 7.64959753e-01
-2.93237656e-01 9.61128473e-01 5.63611507e-01 4.89237905e-01
1.64623189e+00 -1.03144813e+00 5.44835687e-01 7.62221813e-01
5.61788380e-01 6.16022468e-01 5.53534925e-01 -5.52490652e-01
-1.21462941e+00 -1.17734587e+00 1.61725271e+00 -8.86390626e-01
1.01015019e+00 -7.06217527e-01 -1.00296891e+00 1.03251219e+00
8.40405166e-01 -2.90188938e-01 9.56976235e-01 2.99032420e-01
-8.06783795e-01 1.15767360e-01 -9.05901909e-01 6.97012901e-01
9.50951576e-01 -1.08340430e+00 -1.09105361e+00 6.08183682e-01
8.02893579e-01 -6.76354691e-02 -8.32219303e-01 2.20753804e-01
4.65446532e-01 -6.19543731e-01 8.37949812e-01 -1.18731427e+00
4.51460749e-01 1.20121352e-01 -3.97107840e-01 -1.72158659e+00
3.00537832e-02 -3.46016943e-01 6.96716964e-01 1.42895842e+00
7.28033543e-01 -7.72399783e-01 8.28505754e-02 9.54423398e-02
-1.33672073e-01 -7.49826282e-02 -1.05626261e+00 -1.14109290e+00
1.08798265e+00 -4.16634917e-01 2.33687311e-01 1.57611680e+00
1.93713546e-01 1.06374586e+00 -4.25091088e-01 -3.19056183e-01
6.71642482e-01 3.43752988e-02 8.19092333e-01 -1.00527966e+00
-1.61648601e-01 -2.62533307e-01 7.42836073e-02 -8.77808034e-01
6.96559966e-01 -1.46045172e+00 2.28739336e-01 -1.15934455e+00
1.95415333e-01 -5.30662656e-01 -2.99058020e-01 4.50405777e-01
-1.46759093e-01 4.55250859e-01 -5.82550168e-02 3.11354399e-01
-4.28546548e-01 4.63997841e-01 9.54098582e-01 -1.54742479e-01
8.28480050e-02 -6.00416064e-01 -6.03081286e-01 5.81532121e-01
7.12934017e-01 -8.60090435e-01 -1.32664010e-01 -1.01943517e+00
7.17288628e-02 -2.45471522e-01 -2.97279149e-01 -5.65246046e-01
5.98894730e-02 5.28663732e-02 2.05074921e-02 -1.36937454e-01
-1.14568938e-02 -6.79858625e-01 -2.21303493e-01 2.09777743e-01
-3.28030974e-01 1.26882955e-01 4.91699785e-01 1.13150962e-01
-4.14729387e-01 -2.41439521e-01 8.56665254e-01 -1.11343808e-01
-5.75664580e-01 3.06068808e-02 -3.15030575e-01 8.52197349e-01
6.55953884e-01 2.41078004e-01 -3.30526531e-01 -8.13684240e-03
-5.01241982e-01 1.20577239e-01 7.33267128e-01 8.30900252e-01
-2.17605218e-01 -1.74728847e+00 -1.06187451e+00 3.00155401e-01
4.69748497e-01 -5.26971579e-01 -1.43442661e-01 9.84711051e-01
-1.76645264e-01 6.74851954e-01 -4.01855618e-01 -4.28760797e-01
-9.48114276e-01 6.30573332e-01 1.07747972e-01 -5.36518157e-01
-1.77317113e-01 8.42594385e-01 3.37701827e-01 -1.31737804e+00
-2.83195786e-02 1.71210930e-01 1.50717214e-01 2.79569447e-01
2.91155040e-01 -8.46311972e-02 2.26483405e-01 -9.54384804e-01
-4.13122207e-01 8.49984467e-01 -3.01660419e-01 -4.43987697e-01
1.31497097e+00 -2.29967862e-01 -1.29065558e-01 7.30660141e-01
1.58499861e+00 2.79864103e-01 -4.86751914e-01 -7.95537174e-01
2.92113900e-01 -5.58539154e-03 -4.46729064e-01 -8.17252755e-01
-4.73463356e-01 1.23114741e+00 4.08217192e-01 -2.70944715e-01
8.39783967e-01 2.42737338e-01 1.02527738e+00 3.79269749e-01
6.25691772e-01 -1.37353432e+00 -3.43907356e-01 1.01980531e+00
6.60914838e-01 -1.44110179e+00 -4.56844896e-01 -2.14054696e-02
-7.26086736e-01 6.81993783e-01 5.74858308e-01 -1.66670740e-01
4.07611609e-01 1.30650938e-01 4.39633846e-01 2.13775262e-01
-7.41609633e-01 -4.32781070e-01 2.20644310e-01 7.16215193e-01
6.99230552e-01 1.03988878e-01 -5.26626348e-01 5.30793130e-01
-3.75989527e-01 -2.22940013e-01 4.40293662e-02 6.99789226e-01
-1.02178141e-01 -1.72335708e+00 -2.15920061e-01 -7.66717419e-02
-5.65393567e-01 -7.03589678e-01 -7.27258980e-01 8.96804094e-01
2.65459538e-01 7.12020099e-01 1.96587235e-01 -2.65238047e-01
2.48689368e-01 8.15766692e-01 7.64841020e-01 -7.54697859e-01
-6.08769357e-01 4.42784317e-02 3.14258695e-01 -2.57191509e-01
-3.94518793e-01 -6.14751160e-01 -7.74740458e-01 -1.45292148e-01
-2.29981422e-01 4.13592190e-01 5.65147638e-01 1.17580605e+00
1.97682247e-01 1.77661199e-02 3.23551446e-01 -3.85181308e-01
-4.76894796e-01 -1.36459446e+00 -2.47692868e-01 7.20362484e-01
-6.12960954e-04 -4.87962127e-01 -4.96192753e-01 2.68103093e-01] | [11.032645225524902, 9.944483757019043] |
7ad46b73-bde8-4c11-b234-d31fb1429572 | a-neural-method-for-goal-oriented-dialog | null | null | https://openreview.net/forum?id=ByhthReRb | https://openreview.net/pdf?id=ByhthReRb | A Neural Method for Goal-Oriented Dialog Systems to interact with Named Entities | Many goal-oriented dialog tasks, especially ones in which the dialog system has to interact with external knowledge sources such as databases, have to handle a large number of Named Entities (NEs). There are at least two challenges in handling NEs using neural methods in such settings: individual NEs may occur only rarely making it hard to learn good representations of them, and many of the Out Of Vocabulary words that occur during test time may be NEs. Thus, the need to interact well with these NEs has emerged as a serious challenge to building neural methods for goal-oriented dialog tasks. In this paper, we propose a new neural method for this problem, and present empirical evaluations on a structured Question answering task and three related goal-oriented dialog tasks that show that our proposed method can be effective in interacting with NEs in these settings. | ['Satinder Singh', 'Xiaoxiao Guo', 'Mo Yu', 'Jatin Ganhotra', 'Janarthanan Rajendran'] | 2018-01-01 | null | null | null | iclr-2018-1 | ['goal-oriented-dialog'] | ['natural-language-processing'] | [-2.50690728e-01 3.76645714e-01 1.54319704e-01 -5.94513059e-01
-5.02810717e-01 -8.39122951e-01 4.83823836e-01 2.63967842e-01
-7.55967140e-01 1.12438703e+00 3.20898414e-01 -3.71099204e-01
-9.39725339e-02 -9.65631127e-01 -2.94782072e-01 -2.86444817e-02
-7.71950856e-02 1.20685244e+00 3.79399389e-01 -9.39067304e-01
-1.38539607e-02 2.22652406e-01 -1.03338766e+00 3.59615088e-02
8.54140282e-01 7.64319122e-01 1.83820665e-01 5.84127784e-01
-7.05087423e-01 1.03459394e+00 -1.17443228e+00 -6.01205885e-01
4.21110801e-02 -2.91181117e-01 -1.38439631e+00 -7.92491734e-02
3.57303143e-01 -6.40988410e-01 -1.59274995e-01 1.09897268e+00
5.27390063e-01 1.08919442e+00 6.52272940e-01 -1.27075255e+00
-5.82016647e-01 8.89603674e-01 3.64550412e-01 2.58620113e-01
3.21949571e-01 1.12361930e-01 1.31788230e+00 -7.58485734e-01
6.45621121e-01 1.55332148e+00 4.13062274e-01 1.05765402e+00
-1.07267666e+00 -2.41555512e-01 2.29868785e-01 2.88571846e-02
-9.35250938e-01 -5.20273507e-01 5.58251381e-01 -2.16874972e-01
1.31524372e+00 1.35024535e-02 3.00400183e-02 1.07105172e+00
-2.58115143e-01 6.84012115e-01 6.31365895e-01 -3.72034907e-01
4.29638773e-01 4.30427492e-01 7.40525305e-01 4.64320779e-01
-1.13465808e-01 -2.22717002e-01 -4.04154539e-01 -3.89385164e-01
5.92219770e-01 -4.27916050e-01 -3.21446538e-01 -1.19705372e-01
-8.32178175e-01 1.34767604e+00 2.72713214e-01 4.63376969e-01
-2.42513001e-01 -1.75500438e-01 4.29951608e-01 4.16999042e-01
1.46746367e-01 9.38905656e-01 -7.75045276e-01 -3.04967016e-01
-1.23284660e-01 3.66787285e-01 1.64160955e+00 1.09838092e+00
5.22965491e-01 1.90340877e-02 -1.05248488e-01 1.29427981e+00
9.06163156e-02 2.27946863e-01 6.22714460e-01 -8.84037256e-01
7.22154677e-01 8.41446042e-01 4.01419848e-01 -7.36584723e-01
-7.21240699e-01 2.29403362e-01 -7.41094351e-01 -1.59044102e-01
8.66719723e-01 -6.28688276e-01 -5.85907817e-01 2.23778629e+00
3.68811309e-01 -3.81120056e-01 4.41258550e-01 6.89036310e-01
1.26832247e+00 5.53051591e-01 3.89688402e-01 -1.57514602e-01
1.56151640e+00 -8.56493890e-01 -1.10839570e+00 -6.67546749e-01
6.24612033e-01 -3.32015574e-01 1.20187473e+00 -1.28626272e-01
-1.17673075e+00 -3.59000087e-01 -7.45430708e-01 -2.51029193e-01
-8.33096564e-01 -2.62339264e-01 6.43301189e-01 4.44485188e-01
-9.08176780e-01 3.08941424e-01 -4.10801470e-01 -5.52048147e-01
-1.13902494e-01 5.59048295e-01 -2.71011621e-01 2.05191761e-01
-1.70927536e+00 1.35364008e+00 7.55974770e-01 5.53855971e-02
-4.89732772e-01 -3.59919339e-01 -1.10561740e+00 4.88377273e-01
8.18270981e-01 -4.70849693e-01 1.69758928e+00 -4.24563497e-01
-1.34803009e+00 6.49038553e-01 -1.09385587e-01 -3.94610137e-01
2.67650157e-01 -1.92167401e-01 -3.13079447e-01 -1.85809955e-01
-2.14706242e-01 6.22294307e-01 2.90646583e-01 -9.95070219e-01
-5.24377644e-01 -2.55205065e-01 8.33290160e-01 4.50639307e-01
-5.77698171e-01 8.83322433e-02 -1.98348954e-01 -1.62450254e-01
-2.53259391e-01 -7.03571737e-01 -4.03334171e-01 -3.14771146e-01
-4.89751220e-01 -8.83448124e-01 5.55448890e-01 -4.11474437e-01
1.06642509e+00 -1.71513951e+00 1.46292835e-01 -2.53183935e-02
3.84337246e-01 4.05529976e-01 -6.07879870e-02 5.26841879e-01
1.41695291e-01 1.00200534e-01 -2.97583919e-02 -1.30825460e-01
2.18043312e-01 3.72663647e-01 -2.83486515e-01 -4.11555856e-01
1.27341986e-01 8.98194075e-01 -8.65783334e-01 -3.52871150e-01
2.60172468e-02 -1.08613260e-01 -4.12064165e-01 8.69640350e-01
-8.01229954e-01 3.31274003e-01 -4.39700693e-01 2.09687993e-01
1.69848308e-01 -5.04128635e-01 3.58925045e-01 4.45730425e-02
3.77392173e-01 6.12853050e-01 -1.22874129e+00 1.24658346e+00
-5.38562477e-01 3.95260811e-01 2.06938118e-01 -8.53966296e-01
8.39489460e-01 4.73398715e-01 -3.54002342e-02 -4.60850745e-01
3.35411757e-01 -2.20799029e-01 2.19706446e-01 -6.28522396e-01
6.04334116e-01 -4.90197659e-01 -1.72181919e-01 5.83184779e-01
3.94670188e-01 -2.66338706e-01 5.52339792e-01 2.76685923e-01
1.13070047e+00 -7.01450706e-01 5.52338183e-01 -1.17560230e-01
7.14859962e-01 3.01587641e-01 6.29285038e-01 9.55250502e-01
-2.89959073e-01 4.05774042e-02 5.92243910e-01 -4.54554945e-01
-7.89713621e-01 -8.07180524e-01 9.05943960e-02 1.42330861e+00
7.76692033e-02 -3.49402606e-01 -6.34774387e-01 -1.00115764e+00
-2.60544628e-01 1.11677170e+00 -1.44804984e-01 -2.88774163e-01
-6.51459813e-01 -4.54520106e-01 5.44398248e-01 7.32467234e-01
7.30928361e-01 -1.35273397e+00 -3.76870185e-01 5.72093368e-01
-3.75722289e-01 -1.41300344e+00 -3.57267767e-01 4.95328695e-01
-7.17561185e-01 -1.20002890e+00 -3.72032583e-01 -9.67004001e-01
3.78330469e-01 -1.23618059e-01 1.60461366e+00 1.84789687e-01
8.66702572e-02 3.60335022e-01 -2.35227689e-01 -3.66525173e-01
-6.99786544e-01 1.73826516e-01 6.29627854e-02 -5.24247169e-01
9.18636084e-01 -3.59722137e-01 -1.55335749e-02 5.91804087e-01
-9.50097442e-01 -3.33940804e-01 -9.14387181e-02 1.22624958e+00
-1.51831880e-01 2.45235190e-01 1.00433528e+00 -1.25650537e+00
1.35115218e+00 -6.18196130e-01 -7.50020146e-01 4.67742354e-01
-1.14525907e-01 7.42550343e-02 9.08522844e-01 -6.88744187e-01
-1.22478533e+00 -1.80659652e-01 -4.82549548e-01 7.41224959e-02
-5.90604663e-01 8.63784075e-01 -3.72970641e-01 2.29892552e-01
9.00650859e-01 -1.79455042e-01 -5.85995130e-02 -4.42932487e-01
2.61900008e-01 7.38595486e-01 1.01257414e-01 -8.03559363e-01
5.87307274e-01 -1.59272745e-01 -3.82007957e-01 -1.17103529e+00
-1.06016326e+00 -4.98073012e-01 -3.37500662e-01 -4.84799594e-02
9.82469618e-01 -6.77288890e-01 -9.88883674e-01 2.25746751e-01
-1.25379574e+00 -6.94598854e-01 -2.57266164e-01 2.32365966e-01
-4.70196128e-01 2.60338813e-01 -7.85630465e-01 -1.04307497e+00
-2.03603104e-01 -1.08608294e+00 5.49184382e-01 6.24222934e-01
-5.14939308e-01 -1.35123110e+00 8.46618861e-02 4.70944524e-01
7.37282395e-01 -4.92132187e-01 1.49761462e+00 -1.77363098e+00
-2.94053942e-01 -4.79821563e-02 -1.03672490e-01 3.42298418e-01
2.05769971e-01 -7.39315987e-01 -7.90657580e-01 -1.30252421e-01
2.71845996e-01 -1.01326871e+00 2.35662401e-01 7.56022856e-02
7.10771859e-01 -2.84690499e-01 9.34228525e-02 -1.98841691e-01
9.27524090e-01 5.01004398e-01 1.57246634e-01 -4.24008258e-02
3.70255888e-01 9.96287704e-01 6.30421579e-01 2.72011489e-01
5.63560545e-01 7.34420061e-01 1.85714543e-01 1.65165365e-01
3.20704848e-01 -8.88943151e-02 -1.48950834e-02 5.42445779e-01
2.97950566e-01 -6.45117581e-01 -1.11278880e+00 6.41393185e-01
-1.99067581e+00 -7.64804065e-01 8.03783163e-02 1.94635177e+00
1.25215411e+00 -1.73396096e-02 1.43093228e-01 -4.37408924e-01
7.64230847e-01 6.38061315e-02 -7.78166771e-01 -1.16916746e-01
8.74239057e-02 2.53759593e-01 -1.08073391e-01 6.63653135e-01
-1.27575958e+00 1.37018919e+00 6.48173523e+00 2.92652786e-01
-6.64389849e-01 2.18137689e-02 5.07722497e-01 3.82072508e-01
1.35563359e-01 -1.71898007e-01 -1.13143086e+00 7.50749409e-02
1.11935115e+00 -2.66931444e-01 3.32003742e-01 1.15932310e+00
-2.51409203e-01 -4.19009402e-02 -1.44850743e+00 7.26603806e-01
-8.44767988e-02 -1.06702459e+00 -4.19273600e-03 -2.47732863e-01
3.17701429e-01 -2.83227533e-01 -4.63921756e-01 9.64000165e-01
1.03276718e+00 -1.16252887e+00 -1.47129014e-01 3.21368538e-02
3.67796659e-01 -4.70362157e-01 9.97206450e-01 6.69391215e-01
-8.64651203e-01 -1.90079629e-01 -6.86252832e-01 4.40240353e-02
3.87232423e-01 1.69722810e-01 -1.17511952e+00 6.09897226e-02
4.86752182e-01 -4.44131717e-02 -2.33808994e-01 9.11244154e-01
-2.86678284e-01 4.60636020e-01 -3.93768579e-01 -5.84919751e-01
2.08554268e-01 -6.55774400e-02 4.79249626e-01 8.05194855e-01
-1.10883698e-01 5.35850823e-01 3.82760406e-01 9.14629400e-01
-4.58964348e-01 2.48318464e-01 -8.71600270e-01 -2.93117344e-01
4.39893425e-01 1.12114871e+00 -4.12367761e-01 -4.00993288e-01
-5.28783739e-01 6.96995080e-01 6.99937463e-01 4.05135095e-01
-5.90023994e-01 -2.82468885e-01 6.97396636e-01 -2.97870159e-01
-1.32575296e-02 -3.97886008e-01 1.77537203e-01 -1.37791598e+00
-3.95814478e-02 -1.21571839e+00 7.98681617e-01 -6.23374999e-01
-1.79808605e+00 8.79908919e-01 1.54869750e-01 -6.80639863e-01
-6.41872704e-01 -8.59083533e-01 -6.66828334e-01 8.28594029e-01
-1.19410813e+00 -8.11100245e-01 -2.77951688e-01 8.71940672e-01
9.80930209e-01 -3.39551389e-01 1.20808101e+00 2.22797051e-01
-3.94151479e-01 4.50517029e-01 -3.10662717e-01 5.70659876e-01
7.72073805e-01 -1.52357137e+00 2.52939492e-01 3.78706276e-01
3.99703160e-02 1.10093629e+00 6.70436263e-01 -7.26407409e-01
-1.25092864e+00 -7.53879786e-01 1.08000779e+00 -6.16756439e-01
8.00041735e-01 -5.75995684e-01 -1.18304098e+00 9.83053863e-01
3.22599500e-01 -4.72586900e-01 8.46322894e-01 6.39562130e-01
-2.56734014e-01 4.16352987e-01 -1.17835116e+00 8.01585317e-01
9.24624562e-01 -4.80788678e-01 -1.31363988e+00 8.08067918e-01
6.85239971e-01 -5.60183167e-01 -7.23501682e-01 1.99700609e-01
-2.19931558e-01 -5.20638406e-01 9.96942222e-01 -1.47604537e+00
1.68610454e-01 1.76497236e-01 -1.33576080e-01 -1.63900626e+00
1.77070498e-01 -4.25998032e-01 -1.96457252e-01 1.36320651e+00
5.79253316e-01 -7.24357486e-01 7.97597826e-01 1.54084921e+00
1.20081432e-01 -2.55522907e-01 -9.94277775e-01 -6.92176223e-01
4.09993291e-01 -1.16881086e-02 4.04030770e-01 1.08600271e+00
3.63292277e-01 1.13841915e+00 -3.26712191e-01 9.33946297e-02
1.78752527e-01 -7.28999749e-02 7.45847046e-01 -1.59809399e+00
-2.14401066e-01 -2.14822486e-01 -6.58746511e-02 -1.41860902e+00
6.26853168e-01 -5.03972232e-01 2.66016632e-01 -1.33678603e+00
-9.00657102e-02 -5.39380968e-01 2.47675687e-01 5.54104030e-01
-4.68651503e-01 -4.87078398e-01 1.09071448e-03 -1.46004304e-01
-5.39846361e-01 8.31029713e-01 8.78320038e-01 -2.95942008e-01
-2.96597570e-01 2.49818981e-01 -6.09291971e-01 8.94690692e-01
7.28571534e-01 -2.40134433e-01 -5.21533668e-01 -2.53496379e-01
3.16890866e-01 4.52782422e-01 -5.27627990e-02 -7.03760982e-01
5.84845126e-01 -3.39234054e-01 -6.16894923e-02 -3.00263852e-01
8.19556475e-01 -1.08122063e+00 -4.12677735e-01 9.50569194e-03
-6.90507948e-01 4.16737422e-02 1.89902112e-01 4.28115666e-01
-5.18951952e-01 -8.00173104e-01 7.79351234e-01 -4.87032384e-01
-8.61066580e-01 1.12140149e-01 -5.51398516e-01 7.84111142e-01
8.49109530e-01 3.53555262e-01 -6.43386304e-01 -9.53119338e-01
-9.58128512e-01 8.45696449e-01 -1.02374451e-02 5.94181299e-01
4.89885628e-01 -8.96072268e-01 -2.88203865e-01 -1.46608815e-01
1.00490220e-01 6.18538149e-02 2.39207983e-01 2.91990250e-01
-2.09263638e-01 4.86421794e-01 -8.10793489e-02 -1.63556993e-01
-1.25112116e+00 3.89125466e-01 5.82282662e-01 -6.21448100e-01
-3.28051895e-01 1.05859709e+00 3.46919417e-01 -1.05329311e+00
7.40229964e-01 -5.18709831e-02 -8.14154267e-01 2.31626958e-01
7.24204421e-01 -1.33558214e-01 4.65867221e-02 -3.45369577e-01
-4.24108952e-02 -1.63293973e-01 -3.21718007e-01 -1.45156473e-01
1.30858529e+00 7.75033310e-02 2.70610861e-02 4.86667007e-01
7.32001722e-01 -2.54935443e-01 -6.01702809e-01 -5.25193691e-01
3.27871978e-01 -1.04888313e-01 -3.94955665e-01 -8.70545566e-01
-7.28897393e-01 8.46774101e-01 2.68984020e-01 6.88621581e-01
6.48651004e-01 1.10806830e-01 8.47195685e-01 1.46493995e+00
5.31794846e-01 -1.15464675e+00 9.37536508e-02 1.33129299e+00
8.31322014e-01 -1.65126681e+00 -5.85759342e-01 -4.60081518e-01
-8.09138358e-01 1.06149733e+00 1.27388442e+00 2.39627495e-01
6.53701127e-01 2.86335852e-02 2.68438160e-01 -4.83488649e-01
-9.28192675e-01 -4.13594693e-01 -5.32494858e-02 5.40868521e-01
4.15626377e-01 -3.14594716e-01 -2.26216376e-01 6.65258527e-01
-1.58148542e-01 -4.82790112e-01 5.86860061e-01 1.09990191e+00
-3.45823914e-01 -1.30389237e+00 -1.30317405e-01 5.69332540e-01
-3.75565678e-01 -2.68411756e-01 -8.13980758e-01 8.80459309e-01
-6.10546708e-01 1.43727028e+00 -1.40806213e-01 -1.61735460e-01
6.87837601e-01 8.34748268e-01 -1.63112566e-01 -1.07009780e+00
-1.05944121e+00 -4.88578528e-01 7.97947764e-01 -5.69894791e-01
-6.75266385e-02 -6.93551302e-02 -1.34945583e+00 -1.50256038e-01
-6.97611511e-01 5.93077898e-01 4.08067077e-01 1.10078788e+00
9.59248319e-02 2.95769393e-01 1.42221436e-01 -1.65871263e-01
-1.03359389e+00 -1.22233045e+00 -5.67463219e-01 7.38088012e-01
1.21147409e-01 -8.63110721e-01 -4.03689891e-01 -2.68007278e-01] | [12.67417049407959, 7.979498386383057] |
5fc13877-66b0-4f94-9a49-075a4dedfd5e | knowledge-graph-augmented-abstractive | 2005.01159 | null | https://arxiv.org/abs/2005.01159v1 | https://arxiv.org/pdf/2005.01159v1.pdf | Knowledge Graph-Augmented Abstractive Summarization with Semantic-Driven Cloze Reward | Sequence-to-sequence models for abstractive summarization have been studied extensively, yet the generated summaries commonly suffer from fabricated content, and are often found to be near-extractive. We argue that, to address these issues, the summarizer should acquire semantic interpretation over input, e.g., via structured representation, to allow the generation of more informative summaries. In this paper, we present ASGARD, a novel framework for Abstractive Summarization with Graph-Augmentation and semantic-driven RewarD. We propose the use of dual encoders---a sequential document encoder and a graph-structured encoder---to maintain the global context and local characteristics of entities, complementing each other. We further design a reward based on a multiple choice cloze test to drive the model to better capture entity interactions. Results show that our models produce significantly higher ROUGE scores than a variant without knowledge graph as input on both New York Times and CNN/Daily Mail datasets. We also obtain better or comparable performance compared to systems that are fine-tuned from large pretrained language models. Human judges further rate our model outputs as more informative and containing fewer unfaithful errors. | ['Lu Wang', 'Lingfei Wu', 'Luyang Huang'] | 2020-05-03 | knowledge-graph-augmented-abstractive-1 | https://aclanthology.org/2020.acl-main.457 | https://aclanthology.org/2020.acl-main.457.pdf | acl-2020-6 | ['cloze-test'] | ['natural-language-processing'] | [ 3.36139441e-01 6.81549191e-01 -1.65167570e-01 -4.41738307e-01
-1.00972450e+00 -6.46944344e-01 5.41534424e-01 5.36307752e-01
-3.96741629e-01 8.83308291e-01 1.09304118e+00 -1.35712117e-01
7.08843768e-02 -7.11592197e-01 -1.00893605e+00 -1.09762579e-01
1.83377713e-01 6.05773270e-01 -4.62810602e-03 -3.68562311e-01
5.19660175e-01 -9.52624008e-02 -1.28666890e+00 5.07115901e-01
1.40869117e+00 4.56639022e-01 3.61569881e-01 1.05695534e+00
-1.07183106e-01 1.12753558e+00 -9.87862945e-01 -8.81006837e-01
-1.79202780e-01 -8.08103383e-01 -1.01268065e+00 4.62579243e-02
4.93456095e-01 -4.42885399e-01 -5.64419508e-01 9.46388483e-01
6.47514403e-01 3.79023165e-01 8.20278525e-01 -7.66285121e-01
-1.29614127e+00 1.28997791e+00 -2.17805207e-01 2.26199016e-01
5.23435593e-01 8.46918896e-02 1.67496669e+00 -7.85321653e-01
8.29780281e-01 1.26839387e+00 3.77013505e-01 6.05879188e-01
-1.12186897e+00 -2.84296513e-01 3.08400393e-01 1.60176963e-01
-8.79499912e-01 -6.66420460e-01 7.09456205e-01 1.04495011e-01
1.37576067e+00 4.21016693e-01 4.94730145e-01 1.45211840e+00
9.66464430e-02 1.08143520e+00 2.46480256e-01 -2.68920720e-01
1.36530533e-01 2.20583603e-02 1.76702991e-01 8.39781940e-01
6.81041241e-01 -3.62908781e-01 -6.40651226e-01 -1.18507884e-01
4.09470111e-01 -2.20900550e-01 -5.04006982e-01 -2.93571204e-02
-1.10702074e+00 8.59092832e-01 6.20579958e-01 1.59431305e-02
-7.56028891e-01 3.18430215e-01 6.40400171e-01 1.43910691e-01
6.70457959e-01 1.09563291e+00 -1.40299410e-01 -2.07279116e-01
-9.71160889e-01 4.66066897e-01 1.09680808e+00 1.27201498e+00
4.91591752e-01 1.29049063e-01 -7.90552735e-01 8.22977126e-01
1.23639412e-01 3.61222565e-01 5.82464576e-01 -8.31334352e-01
8.49522114e-01 6.30601108e-01 4.44486290e-02 -1.00052726e+00
-1.18631445e-01 -7.73794711e-01 -9.17900145e-01 -5.83313286e-01
-1.80700049e-01 -1.51301220e-01 -8.88859451e-01 1.68106830e+00
-3.12301993e-01 1.87752265e-02 3.59424323e-01 6.99963510e-01
1.10373950e+00 8.37527215e-01 5.44268414e-02 -1.36926159e-01
1.14838207e+00 -1.38257277e+00 -9.53047395e-01 -4.89487737e-01
8.80685210e-01 -3.58200341e-01 1.20413315e+00 5.49089834e-02
-1.39238429e+00 -3.43985558e-01 -1.11050105e+00 -4.21325862e-01
-8.72192532e-02 3.03878784e-01 2.66440958e-01 2.37223879e-01
-1.10581291e+00 8.52556109e-01 -6.68452978e-01 -4.67741519e-01
4.46484029e-01 3.60824056e-02 -7.21252412e-02 -2.65926048e-02
-1.13528836e+00 9.42430913e-01 6.32923186e-01 -2.05546647e-01
-7.23686039e-01 -4.31904703e-01 -1.01593709e+00 5.02086341e-01
4.85206038e-01 -1.22788775e+00 1.48207307e+00 -7.83547938e-01
-1.35769343e+00 4.41321403e-01 -3.11485231e-01 -7.17125773e-01
3.62060010e-01 -4.56115842e-01 -1.44035116e-01 2.99230129e-01
3.83592904e-01 5.87193191e-01 5.38402736e-01 -1.26237094e+00
-2.88059711e-01 -1.34302229e-01 1.01832077e-01 5.86192548e-01
-4.23575342e-01 -1.33970082e-01 -4.61152017e-01 -9.11355019e-01
-4.31004584e-01 -6.10282302e-01 -3.60831618e-01 -6.65002286e-01
-1.02229619e+00 -3.82951051e-01 3.64469796e-01 -1.08925831e+00
1.54036582e+00 -1.75719976e+00 4.20393944e-01 -1.16239227e-01
4.72166568e-01 3.23148847e-01 -4.60224509e-01 7.39074409e-01
1.80440620e-01 4.39676076e-01 -3.28872085e-01 -6.16055965e-01
-4.54909988e-02 1.17808454e-01 -4.38603818e-01 -8.67649391e-02
4.96624589e-01 1.29386330e+00 -1.25070238e+00 -3.45778912e-01
-3.69144201e-01 9.78463739e-02 -7.52317905e-01 4.42341536e-01
-5.91923416e-01 -3.36228311e-02 -6.31576538e-01 2.85728395e-01
1.62815273e-01 -7.35872328e-01 1.00865059e-01 -6.45720074e-03
3.49209130e-01 7.70892084e-01 -5.95060349e-01 1.79382801e+00
-4.09217864e-01 5.23925662e-01 -2.65510201e-01 -8.29746902e-01
9.01479661e-01 1.96516380e-01 -1.57871589e-01 -7.23978400e-01
1.72444526e-02 2.19964191e-01 -1.97513819e-01 -4.26159769e-01
1.44626462e+00 2.96806365e-01 -6.26713932e-02 5.84806979e-01
1.64051339e-01 -3.63269858e-02 4.05275166e-01 9.12563801e-01
1.43211174e+00 -4.98549528e-02 3.38858783e-01 8.94921049e-02
1.86129391e-01 1.06733508e-01 3.80333006e-01 1.11856759e+00
3.04880083e-01 7.50033975e-01 7.25876570e-01 1.38540670e-01
-1.17649090e+00 -7.29290605e-01 6.47268713e-01 1.02575612e+00
1.35557443e-01 -8.97318244e-01 -8.69437099e-01 -9.17604983e-01
-3.92712988e-02 1.46679330e+00 -3.45122427e-01 -5.82041740e-01
-7.14760900e-01 -4.08310413e-01 7.70812631e-01 7.58294642e-01
3.76322180e-01 -1.13558984e+00 -4.32740778e-01 3.52647573e-01
-4.91025209e-01 -1.12163258e+00 -7.96950758e-01 3.85587104e-02
-9.00173724e-01 -7.26040244e-01 -7.89922535e-01 -6.30168378e-01
6.10395968e-01 1.81461722e-01 1.41572618e+00 1.97809815e-01
1.47143677e-01 3.54885072e-01 -7.55071819e-01 -3.19114178e-01
-7.98357964e-01 5.21818697e-01 -2.53679305e-01 -2.47800112e-01
4.47656624e-02 -4.36952323e-01 -6.35683596e-01 -1.84389159e-01
-9.82642949e-01 2.78371513e-01 8.83178234e-01 9.52578783e-01
2.01616749e-01 -3.70850652e-01 9.38922167e-01 -1.15888119e+00
1.34062862e+00 -5.01160979e-01 1.10306591e-01 5.01226008e-01
-7.75757134e-01 4.07213300e-01 9.29261565e-01 -9.92037132e-02
-1.05177796e+00 -4.08246487e-01 -5.35506010e-02 -4.18889970e-01
1.83673516e-01 7.73693025e-01 -1.06748240e-02 6.62486494e-01
1.04035115e+00 4.95270222e-01 -6.41281754e-02 -3.68381739e-01
7.05383182e-01 8.21928918e-01 6.83956027e-01 -4.74123329e-01
5.46202302e-01 -2.42132265e-02 -5.47024369e-01 -7.20416784e-01
-1.06117749e+00 -3.76944393e-01 -1.46449193e-01 4.45310511e-02
7.25526690e-01 -1.00899005e+00 -2.59516001e-01 2.77813990e-02
-1.57760751e+00 -2.08547935e-01 -4.30352271e-01 1.84695497e-01
-5.16492784e-01 5.78256309e-01 -7.68014967e-01 -5.78056514e-01
-8.47419143e-01 -8.33007574e-01 1.18830228e+00 2.71792948e-01
-4.97485340e-01 -9.71936703e-01 1.86887314e-03 3.40080112e-01
4.18817550e-01 1.01831798e-02 1.01023209e+00 -1.24544644e+00
-5.87584794e-01 -1.53798655e-01 -3.24603558e-01 4.74448889e-01
4.59561050e-02 -3.10331546e-02 -6.93467319e-01 -2.23042756e-01
-5.18165469e-01 -4.19242918e-01 1.20757604e+00 2.45529950e-01
1.25709200e+00 -1.02029645e+00 -3.39731991e-01 3.24835718e-01
1.06926143e+00 -1.01949088e-01 6.90335512e-01 5.05710505e-02
8.57349515e-01 5.45501888e-01 3.92738223e-01 5.32669485e-01
6.53784156e-01 3.12219560e-01 3.31429332e-01 2.46257722e-01
-3.55784714e-01 -9.04852808e-01 4.11316305e-01 1.04193509e+00
1.89510994e-02 -1.08936965e+00 -5.80820441e-01 7.37188280e-01
-2.04344249e+00 -1.01966989e+00 -1.89015847e-02 1.85640562e+00
9.49730217e-01 1.23146474e-01 7.35896379e-02 -3.44982296e-01
7.18105435e-01 4.46407259e-01 -5.17290473e-01 -4.63005185e-01
-2.31356353e-01 1.05325535e-01 5.55439532e-01 5.20344198e-01
-5.84364176e-01 1.12678945e+00 5.94411230e+00 6.49142981e-01
-7.78295577e-01 -2.91534126e-01 6.07862115e-01 -7.67028704e-02
-9.41816509e-01 2.17563007e-02 -6.57460213e-01 4.66456622e-01
1.08110321e+00 -6.78583324e-01 3.14237297e-01 7.93186784e-01
1.06966801e-01 1.98050126e-01 -1.23899651e+00 6.86366916e-01
4.49216515e-01 -1.74658597e+00 6.33200407e-01 -1.13905594e-01
6.81682408e-01 -4.49737646e-02 -2.08492637e-01 5.23116827e-01
6.53625786e-01 -1.04999006e+00 8.16033781e-01 5.51315188e-01
6.28619790e-01 -6.71365321e-01 6.37614906e-01 4.76484209e-01
-7.12096155e-01 -8.89008772e-03 -4.04020011e-01 1.05190985e-01
2.49703512e-01 5.44077337e-01 -1.18270075e+00 8.77584755e-01
2.23909885e-01 1.10720873e+00 -7.17408121e-01 7.97936976e-01
-5.16961992e-01 7.07467735e-01 2.42948487e-01 -6.61221027e-01
2.07927987e-01 2.55447067e-02 8.72687757e-01 1.45029199e+00
3.32109809e-01 4.15132940e-02 1.80614054e-01 1.01447511e+00
-6.49752438e-01 1.87614501e-01 -7.13851094e-01 -5.66916406e-01
6.43600345e-01 9.59123254e-01 -3.64057243e-01 -7.80077755e-01
-1.47542790e-01 1.38444817e+00 6.13502979e-01 5.14285028e-01
-6.66701376e-01 -7.75702417e-01 2.23826885e-01 -4.56655361e-02
3.19823325e-01 -7.42836297e-02 -3.07085395e-01 -1.44691074e+00
1.48191944e-01 -9.35607374e-01 3.72296304e-01 -9.38219607e-01
-1.18829870e+00 8.40019941e-01 -9.99862999e-02 -7.98688531e-01
-5.63436985e-01 -1.44710153e-01 -7.49155581e-01 7.44518340e-01
-1.53468633e+00 -9.00264204e-01 -1.92208752e-01 3.89831662e-02
9.01804566e-01 -1.86705649e-01 6.41143262e-01 -4.72516790e-02
-5.91599107e-01 6.26607180e-01 -4.12427597e-02 1.98407024e-01
6.77698851e-01 -1.44715035e+00 7.87604570e-01 9.97779548e-01
1.76990360e-01 7.31684744e-01 8.98939133e-01 -1.00970948e+00
-1.27573383e+00 -1.36251211e+00 1.11533594e+00 -4.48946446e-01
4.24669206e-01 -2.21682504e-01 -1.05605745e+00 9.17220414e-01
6.34104252e-01 -5.67667723e-01 5.77438116e-01 2.39772856e-01
-3.35331649e-01 2.86621243e-01 -7.95143425e-01 8.45770657e-01
1.31916487e+00 -4.38759357e-01 -1.01031005e+00 4.03731823e-01
1.26116323e+00 -4.02952492e-01 -5.58624804e-01 1.88367039e-01
1.09686278e-01 -7.22855210e-01 6.48342669e-01 -9.36342239e-01
1.09753191e+00 9.07005519e-02 -1.28442980e-02 -1.94124854e+00
-5.15539706e-01 -7.21153557e-01 -4.94703531e-01 1.31367910e+00
6.11571729e-01 -4.63365793e-01 7.05593288e-01 3.02935123e-01
-6.69435918e-01 -5.73479652e-01 -3.90012532e-01 -8.42607379e-01
-1.64688826e-01 5.68151288e-02 5.72071135e-01 7.15554655e-01
1.53869033e-01 1.09469795e+00 -2.30279878e-01 6.32674694e-02
4.06955868e-01 -6.65132143e-03 7.11790681e-01 -1.04560220e+00
-2.72921652e-01 -6.01190031e-01 3.89518104e-02 -1.41828704e+00
4.10349905e-01 -1.21833110e+00 1.58844933e-01 -2.20508933e+00
3.33537489e-01 2.30285246e-02 1.16393715e-02 3.11814517e-01
-6.38472617e-01 -3.44706893e-01 -6.73591942e-02 4.97706756e-02
-1.01248074e+00 1.02679420e+00 1.33165789e+00 -1.74634501e-01
-2.19930649e-01 -2.22108677e-01 -1.33675659e+00 2.63452441e-01
6.44271612e-01 -3.77772540e-01 -5.94721317e-01 -6.84142053e-01
4.49211150e-01 3.12460840e-01 2.65566230e-01 -6.55592680e-01
4.26741630e-01 7.10260272e-02 1.13204151e-01 -6.11859739e-01
1.64869353e-01 -1.17876835e-01 -3.30056190e-01 2.26049230e-01
-1.01946545e+00 2.05610231e-01 -1.28888175e-01 9.49918926e-01
-2.39800051e-01 -3.27675819e-01 1.74627960e-01 -3.60503584e-01
-4.21952963e-01 2.12892875e-01 -2.11942300e-01 4.22629982e-01
4.91327733e-01 -1.16161749e-01 -6.81859910e-01 -8.93710375e-01
-3.54223162e-01 4.59321409e-01 4.32014376e-01 4.88490909e-01
8.56024861e-01 -1.17769849e+00 -1.11627841e+00 -2.47685701e-01
3.22062045e-01 1.88686237e-01 7.48688206e-02 4.38757658e-01
-4.65346456e-01 4.55014825e-01 7.51171932e-02 -1.57798886e-01
-1.05972660e+00 3.63361418e-01 1.00884892e-01 -5.02591610e-01
-8.33653152e-01 8.10022473e-01 2.72012632e-02 -3.67878079e-01
2.09499195e-01 -4.65319782e-01 -2.17437834e-01 9.97210294e-02
4.82598960e-01 4.98195499e-01 1.47492483e-01 -1.68487355e-01
-2.57753115e-02 -1.42269745e-01 -4.66091454e-01 -3.52572091e-02
1.46309578e+00 -9.36456099e-02 -4.43167165e-02 8.07648078e-02
1.08405745e+00 1.78021237e-01 -9.70884621e-01 -4.12540019e-01
2.52623290e-01 -2.52515703e-01 -1.55023396e-01 -8.48501384e-01
-5.97933710e-01 6.13080978e-01 -4.18305546e-01 5.53377807e-01
7.87935257e-01 1.07754067e-01 9.44364071e-01 7.59024203e-01
-1.43666238e-01 -1.11607420e+00 5.12846470e-01 6.91457927e-01
1.13397264e+00 -9.57238734e-01 -6.13802411e-02 -3.77203643e-01
-1.13315606e+00 9.67674196e-01 6.10147059e-01 -1.54885441e-01
-1.52916506e-01 -1.72891989e-01 -4.25114900e-01 -2.99163997e-01
-1.18595004e+00 -1.26297086e-01 3.75510812e-01 4.11508799e-01
5.55656672e-01 5.88262752e-02 -1.40127614e-01 7.54486203e-01
-4.91478622e-01 -1.86270580e-01 8.65043461e-01 6.60391271e-01
-6.15203738e-01 -7.48326540e-01 1.65287733e-01 9.13831413e-01
-4.08752561e-01 -3.02269340e-01 -7.00015962e-01 2.76190072e-01
-8.16585481e-01 1.03156912e+00 6.23650365e-02 -4.04216230e-01
5.01830935e-01 1.24229044e-01 2.79742271e-01 -1.02523303e+00
-7.26597726e-01 -3.02098632e-01 7.25903153e-01 -3.84960115e-01
7.21550407e-03 -4.24900204e-01 -1.21034002e+00 -2.83794433e-01
-2.95995653e-01 4.41857994e-01 4.51155335e-01 6.76520646e-01
8.13746631e-01 9.75808382e-01 6.14092708e-01 -5.45599937e-01
-9.77950096e-01 -1.20747185e+00 -3.92519921e-01 4.69357789e-01
3.55508566e-01 -6.32723048e-02 -2.44631529e-01 -1.06141828e-01] | [12.279693603515625, 9.250367164611816] |
f9f6089b-b05e-4590-af40-aed93d8aff1f | lilgym-natural-language-visual-reasoning-with | 2211.01994 | null | https://arxiv.org/abs/2211.01994v3 | https://arxiv.org/pdf/2211.01994v3.pdf | lilGym: Natural Language Visual Reasoning with Reinforcement Learning | We present lilGym, a new benchmark for language-conditioned reinforcement learning in visual environments. lilGym is based on 2,661 highly-compositional human-written natural language statements grounded in an interactive visual environment. We introduce a new approach for exact reward computation in every possible world state by annotating all statements with executable Python programs. Each statement is paired with multiple start states and reward functions to form thousands of distinct Markov Decision Processes of varying difficulty. We experiment with lilGym with different models and learning regimes. Our results and analysis show that while existing methods are able to achieve non-trivial performance, lilGym forms a challenging open problem. lilGym is available at https://lil.nlp.cornell.edu/lilgym/. | ['Yoav Artzi', 'Noriyuki Kojima', 'Kianté Brantley', 'Anne Wu'] | 2022-11-03 | null | null | null | null | ['visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning'] | [-1.62096903e-01 -1.41366795e-01 -2.84314662e-01 -1.13436371e-01
-8.21660638e-01 -1.02565491e+00 5.41985393e-01 2.08802834e-01
-4.98166174e-01 1.06068814e+00 6.55999482e-02 -9.47982132e-01
3.65788341e-01 -5.07807910e-01 -9.06447709e-01 -4.84130710e-01
-4.49619979e-01 6.15691483e-01 3.79481912e-01 -7.41443187e-02
1.82700828e-01 2.71649957e-01 -1.15513015e+00 4.58523542e-01
7.47448444e-01 5.91578782e-01 2.39284545e-01 1.27981377e+00
-1.25217363e-01 1.44561350e+00 -4.55421388e-01 -2.44933844e-01
8.00695121e-02 -4.99154687e-01 -1.00068557e+00 -4.16123122e-01
1.56753391e-01 -3.04581463e-01 -4.31014560e-02 1.16054046e+00
3.55732024e-01 1.32516190e-01 4.97169822e-01 -1.75187600e+00
-6.02415979e-01 8.90139997e-01 -4.47647959e-01 2.57567495e-01
7.65355229e-01 8.28860700e-01 1.33978891e+00 -4.87053931e-01
6.63162827e-01 1.40199518e+00 2.14792892e-01 7.32997000e-01
-1.84011781e+00 -7.35406101e-01 2.08154857e-01 1.89179033e-01
-9.47997272e-01 -1.99808404e-01 4.02914941e-01 -6.38706148e-01
1.47164476e+00 2.41711363e-01 7.74743855e-01 1.27500772e+00
5.51609576e-01 9.58797932e-01 1.74343240e+00 -2.68346608e-01
6.88040078e-01 -2.50209153e-01 7.25193098e-02 9.66407418e-01
-1.48328558e-01 6.25608623e-01 -6.13140583e-01 -5.46514213e-01
8.93842995e-01 -3.90824974e-01 2.87678868e-01 -5.64011097e-01
-1.45608580e+00 7.91605473e-01 2.66490430e-01 -9.34792608e-02
6.86925948e-02 8.25127184e-01 5.26164532e-01 4.33404297e-01
-1.93142563e-01 5.32707155e-01 -2.24724054e-01 -4.64493334e-01
-5.67476392e-01 6.92416608e-01 1.05200016e+00 1.02005136e+00
7.65628278e-01 2.42345288e-01 -4.48735684e-01 4.73145843e-02
3.07883531e-01 7.28359699e-01 1.77093491e-01 -1.32723773e+00
1.84707120e-01 1.93224162e-01 5.39524555e-01 -5.27804971e-01
-4.89096195e-01 6.41187578e-02 -4.16154623e-01 8.28487754e-01
5.90381384e-01 -4.16282177e-01 -7.51891434e-01 1.63819087e+00
6.39384016e-02 7.77725875e-02 1.33071199e-01 6.77622259e-01
5.69671035e-01 1.01101267e+00 6.10525846e-01 -6.60964921e-02
1.03422773e+00 -1.09586763e+00 -2.62925357e-01 -3.47085714e-01
4.42234784e-01 -3.75607103e-01 1.54605258e+00 5.46034157e-01
-1.08971322e+00 -1.76180080e-01 -8.10489953e-01 1.67281121e-01
2.10632831e-02 -1.51065961e-01 9.44446206e-01 2.91490942e-01
-1.37245977e+00 2.65431285e-01 -1.11764133e+00 2.11541757e-01
2.40033180e-01 5.61390817e-02 -1.03545122e-01 1.74607277e-01
-1.10477233e+00 9.75116253e-01 6.40164137e-01 -2.72778928e-01
-1.79914558e+00 -1.99803665e-01 -9.36605871e-01 6.61594495e-02
5.94848037e-01 -3.73761863e-01 1.89661992e+00 -1.07157171e+00
-1.67361045e+00 8.67387950e-01 -5.83816022e-02 -5.37678182e-01
7.28412628e-01 -1.14370897e-01 -2.26870358e-01 1.31033912e-01
1.67499885e-01 6.10973120e-01 8.10557246e-01 -1.16617870e+00
-3.60468209e-01 1.22710496e-01 2.21746698e-01 2.19915092e-01
4.81573820e-01 5.04257083e-01 -3.51513959e-02 -4.75027740e-01
-7.00497687e-01 -1.05138850e+00 -6.34919167e-01 4.06692922e-02
-5.16129673e-01 -4.33005035e-01 2.49180391e-01 -4.65603471e-01
9.87971485e-01 -2.05988932e+00 2.91617930e-01 6.58824146e-02
3.42967808e-01 -1.71955258e-01 -1.25657260e-01 7.91943073e-01
3.20299454e-02 9.06279311e-02 -3.54385316e-01 -7.35614672e-02
3.50754499e-01 2.75389075e-01 -6.06726348e-01 3.56510341e-01
2.27607414e-01 1.14550722e+00 -1.35601306e+00 -8.01021695e-01
1.51523054e-01 -2.56374061e-01 -8.18990469e-01 4.51880634e-01
-9.96701241e-01 5.01122952e-01 -3.75803351e-01 6.40993893e-01
9.41616669e-02 -5.01841009e-01 4.78660822e-01 7.89157808e-01
-1.19706832e-01 3.63001108e-01 -1.14610362e+00 1.72074807e+00
-2.65552580e-01 5.05171061e-01 -2.07746789e-01 -6.79803491e-01
6.83321476e-01 1.04709506e-01 5.99133186e-02 -4.93248850e-01
-5.60358018e-02 5.55245876e-02 -6.76737577e-02 -4.21195656e-01
5.39768457e-01 -3.98283452e-01 -6.45406246e-01 6.53254747e-01
1.51492566e-01 -3.78998846e-01 5.74163020e-01 4.84119117e-01
1.18904769e+00 6.19467735e-01 7.89806843e-01 -3.22248727e-01
-1.07918521e-02 2.44188607e-01 4.04636651e-01 1.21138299e+00
-5.27833462e-01 -1.91873640e-01 1.08916092e+00 -2.70488143e-01
-1.05171227e+00 -1.76632273e+00 2.50836700e-01 1.45925629e+00
1.09270662e-01 -5.70013821e-01 -3.76779914e-01 -5.20210922e-01
5.61729297e-02 1.09189343e+00 -4.65717286e-01 5.08589372e-02
-3.82026881e-01 -2.55035460e-01 8.08887243e-01 6.48470581e-01
3.05018693e-01 -1.81851768e+00 -1.12502038e+00 2.75815368e-01
1.57558993e-01 -8.44595373e-01 -3.21908236e-01 6.83113933e-01
-5.74188113e-01 -8.83421481e-01 -3.50400865e-01 -6.57664776e-01
4.26844090e-01 -4.56066579e-01 1.64544106e+00 7.90906511e-03
-1.90577909e-01 2.21424222e-01 -1.29669875e-01 -2.47959852e-01
-8.46401393e-01 -1.80418864e-01 -1.75086617e-01 -7.74948835e-01
5.65935038e-02 -3.37898433e-01 -1.35168284e-01 -1.22388387e-02
-5.78001678e-01 5.15524447e-01 8.55906606e-02 1.00471818e+00
4.28277642e-01 -2.65293092e-01 1.14883758e-01 -8.14399004e-01
8.03014100e-01 -2.58543015e-01 -1.30383003e+00 2.53571481e-01
-2.18420520e-01 4.82404083e-01 6.91558003e-01 -4.61583495e-01
-1.00156546e+00 2.74272919e-01 -6.09888397e-02 -1.63042530e-01
-3.78852606e-01 6.07381761e-01 3.63810748e-01 3.05861562e-01
9.92955029e-01 3.58927280e-01 -2.67873853e-01 1.34307429e-01
6.55800700e-01 3.30108106e-02 6.51758552e-01 -1.24976897e+00
6.49049640e-01 -9.49449092e-03 -1.60922989e-01 -2.71705151e-01
-5.91217875e-01 3.09125371e-02 -1.54220551e-01 -3.74887973e-01
8.85726929e-01 -7.69164443e-01 -1.46680939e+00 4.61483657e-01
-9.24905479e-01 -1.39463449e+00 -4.10311729e-01 6.14998415e-02
-1.01136100e+00 3.02305341e-01 -9.89549756e-01 -1.08584559e+00
-2.04704050e-02 -1.26287615e+00 7.71048784e-01 1.13574430e-01
-3.74989599e-01 -8.31960320e-01 4.73745197e-01 -2.24436700e-01
1.05448790e-01 2.88825423e-01 9.64315832e-01 -3.74410897e-01
-6.16329432e-01 3.62603039e-01 1.08146355e-01 1.28260151e-01
-3.57411504e-01 2.89916784e-01 -6.65923774e-01 -4.29533362e-01
-4.87543344e-01 -1.09182274e+00 5.30702889e-01 9.04956758e-02
1.23180175e+00 -4.13764656e-01 5.99804223e-02 2.24840701e-01
1.48331511e+00 3.07728052e-01 4.47810292e-01 4.95190084e-01
4.32742566e-01 1.50403023e-01 5.82331181e-01 7.19786227e-01
3.34130645e-01 3.40349197e-01 6.47714853e-01 1.44261271e-01
4.38462883e-01 -6.39823854e-01 9.57692921e-01 2.19675869e-01
-5.03535643e-02 -8.27311128e-02 -1.47338676e+00 2.72995532e-01
-1.94546390e+00 -1.29206967e+00 1.78422645e-01 1.79081583e+00
1.31086075e+00 3.87102276e-01 4.55494732e-01 -2.51512527e-01
3.90056849e-01 2.30787650e-01 -6.77359521e-01 -7.70721018e-01
1.01373658e-01 4.99130398e-01 5.04266798e-01 7.21732378e-01
-1.01648021e+00 1.31212211e+00 7.27648544e+00 6.86783612e-01
-1.02891421e+00 -4.61168252e-02 4.21378165e-01 -1.32820100e-01
-4.64017421e-01 3.08326006e-01 -6.70473218e-01 2.99330473e-01
1.03161108e+00 -6.46251619e-01 8.82295430e-01 9.42893863e-01
-8.48304015e-03 -4.68419909e-01 -1.14043283e+00 7.44038820e-01
-4.41036403e-01 -1.40645349e+00 -2.71308541e-01 -5.00284076e-01
8.14039111e-01 3.52674603e-01 1.69818047e-02 6.02862298e-01
1.62259388e+00 -1.33660328e+00 1.20515835e+00 3.05825561e-01
9.41355169e-01 -8.30011129e-01 -5.88738769e-02 5.82402527e-01
-9.72448885e-01 -9.71004516e-02 -1.02424860e-01 -5.58402658e-01
1.03169017e-01 4.36580628e-02 -8.33068252e-01 2.38619864e-01
9.95602608e-01 7.72828996e-01 -4.88602221e-01 8.11552048e-01
-9.27417874e-01 9.14414525e-01 -5.10108396e-02 -6.04072988e-01
4.58691388e-01 6.20783493e-02 3.74495655e-01 1.27362382e+00
-6.62232041e-02 -2.26246893e-01 6.89574063e-01 1.18593073e+00
1.20694757e-01 -8.22240934e-02 -7.54167497e-01 -4.35782760e-01
2.06812978e-01 1.03827071e+00 -6.80861890e-01 -6.18819118e-01
-2.43287757e-01 1.02715528e+00 8.43738675e-01 6.31152093e-01
-1.21207488e+00 -1.79082248e-02 6.38521969e-01 -3.05390120e-01
5.33896759e-02 -7.58694828e-01 -1.90892220e-02 -1.35028636e+00
-2.95224756e-01 -1.28604007e+00 5.23887813e-01 -1.18867564e+00
-1.13327682e+00 5.28483331e-01 2.30634972e-01 -8.98386240e-01
-5.52629173e-01 -8.60444844e-01 -4.80923831e-01 7.46751308e-01
-1.20166039e+00 -7.65511453e-01 5.46847284e-02 7.57660568e-01
3.29217166e-01 -3.39288563e-02 1.06230187e+00 -2.81331390e-01
-3.55718255e-01 2.15714216e-01 -5.91834076e-03 -2.45888978e-02
5.54579258e-01 -1.72237682e+00 7.48950481e-01 7.71871865e-01
1.99152812e-01 5.02991974e-01 9.00252819e-01 -6.08891547e-01
-1.25489688e+00 -8.27814460e-01 6.06596529e-01 -5.32044291e-01
1.15845072e+00 -6.12739682e-01 -6.46327734e-01 1.23553681e+00
5.04134715e-01 6.74824566e-02 4.39509600e-01 -1.39465585e-01
-5.12145519e-01 3.90711755e-01 -8.06696117e-01 1.08763111e+00
9.63301480e-01 -9.28995550e-01 -5.19803524e-01 3.12680334e-01
7.40849614e-01 -8.91773343e-01 -5.11784673e-01 -2.37663582e-01
2.19864532e-01 -9.37880516e-01 7.45790005e-01 -1.02231491e+00
8.27205837e-01 -5.69517314e-01 -6.62638918e-02 -1.27022505e+00
-2.63473451e-01 -8.39414477e-01 -9.79806334e-02 5.52586496e-01
3.93979549e-01 -7.36170650e-01 5.99869072e-01 4.56071585e-01
1.63188025e-01 -5.14630973e-01 -9.61537123e-01 -8.60667765e-01
3.29269618e-01 -7.44010985e-01 2.55089134e-01 6.67332888e-01
3.94388437e-01 1.96103543e-01 -7.07615972e-01 -5.68732321e-02
5.21949410e-01 5.11345863e-01 7.29421258e-01 -6.63279414e-01
-8.82465780e-01 -3.57327163e-01 -3.84615734e-02 -1.03071642e+00
5.84997416e-01 -1.30189943e+00 3.72925580e-01 -1.36062574e+00
4.05482739e-01 -4.14035767e-01 -2.30691507e-01 1.08145356e+00
-9.06666294e-02 -1.32814333e-01 4.19874758e-01 -8.02127942e-02
-9.79590654e-01 4.71429706e-01 1.26530159e+00 -1.98598728e-01
1.45586386e-01 -1.19595654e-01 -3.31350714e-01 6.67378843e-01
1.08988726e+00 -6.35963082e-01 -3.27541292e-01 -1.36424571e-01
6.48515642e-01 6.18704140e-01 9.08371091e-01 -9.32205379e-01
1.24910556e-01 -8.94231558e-01 2.02749282e-01 -3.60731483e-01
-8.31907168e-02 -5.50062478e-01 2.22726054e-02 9.32052135e-01
-7.80623198e-01 6.01681352e-01 4.53112036e-01 4.36612785e-01
1.28818199e-01 -1.40780807e-01 6.83863699e-01 -2.88987607e-01
-1.11873472e+00 -9.38560814e-02 -9.56821680e-01 5.41406453e-01
1.28560364e+00 4.77452546e-01 -3.96486521e-01 -4.29068357e-01
-1.04382253e+00 4.79518384e-01 6.10627830e-01 1.34996831e-01
6.11751556e-01 -1.03312075e+00 -4.78998184e-01 -1.10382654e-01
1.51753858e-01 -3.71965230e-01 -4.52411175e-02 3.87675315e-01
-7.81194508e-01 4.40328903e-02 -6.72110379e-01 -4.79491621e-01
-1.23276472e+00 8.06009471e-01 3.89856666e-01 -6.62707031e-01
-4.66669321e-01 6.71798110e-01 1.12118296e-01 -5.23734450e-01
1.44153476e-01 -5.52530706e-01 1.98527709e-01 -5.78229308e-01
3.11121494e-01 4.62296623e-04 -5.97023606e-01 -1.18530013e-01
-3.23238552e-01 -1.45030618e-01 2.09048524e-01 -5.36023676e-01
1.05545759e+00 3.80793840e-01 -1.83953106e-01 8.65547538e-01
4.89428818e-01 -1.17313400e-01 -1.50081837e+00 -2.22952887e-02
2.17426762e-01 -1.60695255e-01 -4.86440122e-01 -8.91510010e-01
-4.93452638e-01 8.56570184e-01 3.40365916e-01 -3.37440744e-02
8.08322728e-01 1.06115282e-01 1.35521770e-01 7.83715665e-01
9.07509446e-01 -8.19687366e-01 4.88258600e-01 8.02364290e-01
8.75640750e-01 -1.26014853e+00 -1.85126543e-01 1.39945596e-01
-1.08637774e+00 9.34054375e-01 8.55700016e-01 -3.12889367e-01
1.29825681e-01 8.35595071e-01 5.74838817e-02 -3.91609147e-02
-1.13599956e+00 -3.56297754e-02 -2.98296005e-01 6.13464355e-01
4.57512915e-01 5.31866431e-01 5.47308289e-02 3.30590725e-01
-2.58555859e-01 1.48377359e-01 5.56533337e-01 1.34297860e+00
-3.50572497e-01 -1.01037633e+00 -3.61534536e-01 2.81997979e-01
-2.88981289e-01 -3.71222347e-01 -2.02423222e-02 5.97600758e-01
-5.09033799e-01 5.85063934e-01 -1.77275032e-01 -1.42090708e-01
-1.27543271e-01 4.68974747e-02 5.93531013e-01 -7.42941082e-01
-7.89676726e-01 -2.66143411e-01 2.32720554e-01 -8.02108169e-01
-1.66399688e-01 -7.15456784e-01 -1.79828513e+00 -3.87691200e-01
4.28643614e-01 6.60885647e-02 -4.96614277e-02 5.13661146e-01
-1.00594804e-01 6.06903374e-01 1.87312186e-01 -8.12097907e-01
-9.68950868e-01 -3.20536405e-01 -5.58676243e-01 1.59920186e-01
3.96599203e-01 -4.50779796e-01 -1.44300491e-01 1.20464675e-01] | [4.029333591461182, 1.5427740812301636] |
42ffb2ec-2d75-47f5-a143-85ecf125e0cc | diffload-uncertainty-quantification-in-load | 2306.01001 | null | https://arxiv.org/abs/2306.01001v1 | https://arxiv.org/pdf/2306.01001v1.pdf | DiffLoad: Uncertainty Quantification in Load Forecasting with Diffusion Model | Electrical load forecasting is of great significance for the decision makings in power systems, such as unit commitment and energy management. In recent years, various self-supervised neural network-based methods have been applied to electrical load forecasting to improve forecasting accuracy and capture uncertainties. However, most current methods are based on Gaussian likelihood methods, which aim to accurately estimate the distribution expectation under a given covariate. This kind of approach is difficult to adapt to situations where temporal data has a distribution shift and outliers. In this paper, we propose a diffusion-based Seq2seq structure to estimate epistemic uncertainty and use the robust additive Cauchy distribution to estimate aleatoric uncertainty. Rather than accurately forecasting conditional expectations, we demonstrate our method's ability in separating two types of uncertainties and dealing with the mutant scenarios. | ['Yi Wang', 'Liang Sun', 'Chaoli Zhang', 'Qingsong Wen', 'Zhixian Wang'] | 2023-05-31 | null | null | null | null | ['load-forecasting', 'energy-management'] | ['miscellaneous', 'time-series'] | [-1.08632166e-02 -1.79492176e-01 -6.85152085e-03 -5.25449276e-01
-6.51665390e-01 -4.99081671e-01 5.71781695e-01 2.58715898e-01
-1.52859524e-01 1.19975662e+00 2.07196966e-01 -3.84892672e-01
-4.62576509e-01 -9.81354892e-01 -5.88110685e-01 -8.23703110e-01
-1.24262288e-01 4.65511143e-01 -4.05475870e-02 2.15195283e-01
2.87099510e-01 3.80611509e-01 -1.12057090e+00 -2.03543946e-01
1.34677744e+00 1.07247782e+00 -5.10394424e-02 3.78275245e-01
-2.51149327e-01 6.01355374e-01 -9.16994393e-01 -2.51116991e-01
8.36368278e-02 -3.38018537e-01 -1.36889741e-01 -3.78021240e-01
-6.96390629e-01 -4.21228081e-01 -9.76704583e-02 1.27173436e+00
6.36791348e-01 1.72258303e-01 1.07215667e+00 -1.30896485e+00
-3.69121373e-01 1.09798408e+00 -5.22817194e-01 3.09417367e-01
7.36366725e-03 1.58364192e-01 3.70779097e-01 -5.26420236e-01
-1.66783154e-01 1.05883670e+00 8.87107372e-01 5.66940196e-02
-1.02187884e+00 -5.96022785e-01 1.06217757e-01 2.37400204e-01
-1.47659457e+00 -2.26993218e-01 7.75270879e-01 -5.42020321e-01
7.34534264e-01 1.20198922e-02 3.34214717e-01 1.03751528e+00
5.57344973e-01 7.80991256e-01 1.13096344e+00 -1.57073393e-01
8.88672769e-01 -2.82030925e-02 -2.00783208e-01 -3.47075433e-01
2.88483441e-01 3.21423799e-01 -2.09493354e-01 -2.41117984e-01
2.82861769e-01 -2.49800533e-02 -1.67228431e-01 1.39300883e-01
-9.42042649e-01 8.47710609e-01 -4.76609543e-02 2.87589282e-01
-4.90219593e-01 3.55025083e-01 3.82552654e-01 -1.28408298e-01
9.91851985e-01 9.89968143e-03 -5.10757565e-01 -5.74441373e-01
-1.12917387e+00 5.31134605e-02 9.59937692e-01 8.42269719e-01
2.42781818e-01 6.21966600e-01 -2.95838892e-01 5.26662827e-01
4.36316192e-01 7.75443792e-01 5.43795109e-01 -6.33172631e-01
3.39032769e-01 9.19325128e-02 3.37065011e-01 -7.53782630e-01
-6.21141851e-01 -4.86994207e-01 -1.16504347e+00 3.03439468e-01
5.06066680e-01 -6.45915926e-01 -7.52299130e-01 1.47804761e+00
2.06859499e-01 4.43170965e-01 2.64492273e-01 4.29951340e-01
1.64919630e-01 8.22130620e-01 4.35697958e-02 -3.41807097e-01
8.23230207e-01 -3.48523974e-01 -1.06854427e+00 1.33215681e-01
5.15307665e-01 -5.08705139e-01 3.35166097e-01 5.95630288e-01
-9.21951115e-01 -2.33541131e-01 -9.69230831e-01 7.43617177e-01
-4.85695034e-01 -1.73162282e-01 1.57571316e-01 9.07003105e-01
-4.43198860e-01 1.00528264e+00 -9.06291306e-01 7.21808011e-03
2.04908296e-01 -6.39367923e-02 2.88832575e-01 2.37286076e-01
-1.52559090e+00 1.08610320e+00 9.26837027e-01 7.17876077e-01
-7.42854834e-01 -6.00115716e-01 -5.98958194e-01 2.10814491e-01
5.43292761e-01 -2.83136845e-01 1.16877115e+00 -7.19548643e-01
-1.77646220e+00 -3.58308136e-01 1.54012203e-01 -6.06281817e-01
6.67177796e-01 4.48134951e-02 -7.51209676e-01 -4.36256230e-01
-4.35385883e-01 -2.28237495e-01 7.78717935e-01 -8.76839578e-01
-5.30838311e-01 -1.92538202e-01 -4.08325493e-01 -2.47091055e-02
-7.71290585e-02 -8.99501666e-02 2.28873670e-01 -8.93964171e-01
-2.36645658e-02 -6.81452572e-01 -1.67624697e-01 -4.79662061e-01
-4.77994323e-01 -3.39708060e-01 3.27041388e-01 -8.95100772e-01
1.37508500e+00 -2.02221274e+00 -2.16236845e-01 5.73882818e-01
-3.90314430e-01 1.10544130e-01 3.66764575e-01 8.18137765e-01
1.52365521e-01 7.23172352e-02 -6.05742633e-01 -2.34136268e-01
4.89586264e-01 3.60989720e-01 -3.58404368e-01 4.02339607e-01
2.20608383e-01 6.71674490e-01 -8.38915646e-01 1.88193265e-02
1.88675269e-01 2.40669683e-01 6.01646025e-03 1.50701061e-01
-4.68618840e-01 4.50251907e-01 -3.34224552e-01 2.97693044e-01
9.47614849e-01 2.18832552e-01 1.32868126e-01 -1.33465044e-02
-1.86315820e-01 -1.27892688e-01 -1.64891648e+00 1.36130822e+00
-4.07268435e-01 2.82256931e-01 -1.77024275e-01 -1.11389649e+00
9.23265219e-01 2.76230961e-01 3.83472413e-01 -2.88802952e-01
2.13319004e-01 2.39172369e-01 -3.54966037e-02 -4.28954542e-01
4.53326792e-01 -2.63735384e-01 -8.12726840e-02 5.98512948e-01
-1.73155695e-01 -2.31259316e-01 1.88674778e-01 -1.52679637e-01
7.83548415e-01 3.16899419e-01 1.28347471e-01 -2.67230719e-01
3.05271924e-01 -3.30252022e-01 9.46088672e-01 6.08787656e-01
1.17236726e-01 5.12355089e-01 6.51127517e-01 -1.13658011e-01
-8.26706231e-01 -1.13445294e+00 -3.62594634e-01 3.79502296e-01
-2.65491366e-01 1.45084783e-01 -5.34172237e-01 -3.94036770e-01
3.46945465e-01 1.62521434e+00 -3.26072931e-01 -3.53795320e-01
-5.20841628e-02 -1.28507054e+00 5.62592328e-01 7.58315027e-01
3.07292759e-01 -6.98385239e-01 -2.79753357e-01 6.08663559e-01
1.05415173e-01 -5.66197693e-01 -4.02653158e-01 3.88246924e-01
-6.15902841e-01 -7.10411847e-01 -8.36466134e-01 1.55940071e-01
5.61583698e-01 -5.13260484e-01 8.32372367e-01 -4.63901609e-01
7.61222690e-02 2.45984420e-01 -3.00395489e-01 -8.46541405e-01
-5.60716927e-01 -1.44073829e-01 2.19395921e-01 9.38631967e-02
2.34685048e-01 -5.12685537e-01 -3.11469167e-01 3.34257662e-01
-1.30183935e+00 -4.67736304e-01 3.77960742e-01 7.61057794e-01
2.82701850e-01 8.26044798e-01 1.28545606e+00 -7.63007045e-01
1.03257227e+00 -9.89846349e-01 -9.80548143e-01 4.75419164e-01
-1.04669476e+00 2.01243937e-01 5.81854939e-01 -4.67063427e-01
-1.48206401e+00 -9.38092992e-02 -1.79845780e-01 -2.00932145e-01
-1.99896142e-01 1.18403089e+00 -5.38433641e-02 3.82708967e-01
5.52311651e-02 1.59637675e-01 2.60008164e-02 -4.34700191e-01
1.56034857e-01 9.00961876e-01 3.85001391e-01 -5.70400715e-01
6.19339943e-01 -1.13680894e-02 2.21184641e-01 -4.13307875e-01
-5.28408408e-01 -1.20838813e-01 -5.07521689e-01 -3.55078071e-01
5.78918576e-01 -8.27314258e-01 -8.76724422e-01 9.97889459e-01
-8.99241567e-01 -2.53317237e-01 -4.72370297e-01 8.97024810e-01
-5.18223524e-01 5.31019747e-01 -3.43470722e-01 -1.47390580e+00
-1.88921556e-01 -1.08995152e+00 5.19874096e-01 3.23288023e-01
7.85198659e-02 -1.37014019e+00 2.64877528e-01 -2.59914875e-01
8.54842961e-01 3.92328590e-01 7.83584118e-01 -8.59712064e-01
-1.15979463e-01 -3.25767517e-01 1.00877270e-01 6.65152490e-01
8.85721147e-02 1.25278383e-01 -8.20802152e-01 -2.55942076e-01
-3.00937388e-02 -6.05552644e-02 3.15963686e-01 5.68435311e-01
1.10556340e+00 -1.13853082e-01 -3.37713510e-02 2.47805893e-01
1.27013052e+00 5.69906235e-01 5.98197758e-01 1.87770799e-02
1.35984039e-02 5.70702434e-01 4.92732286e-01 9.62885976e-01
3.88939142e-01 3.29005092e-01 4.04032826e-01 4.10737395e-01
5.89871287e-01 -3.25387061e-01 3.85378838e-01 8.73266876e-01
3.84367049e-01 -6.43332779e-01 -9.91929948e-01 5.31793356e-01
-1.98784363e+00 -1.09062505e+00 -1.07926287e-01 2.38054180e+00
9.38359678e-01 3.10479283e-01 -8.48189518e-02 -2.49427482e-02
7.21468747e-01 -2.02660158e-01 -9.71999168e-01 -3.58876735e-01
-2.94119000e-01 -2.68883437e-01 6.78939700e-01 4.00394112e-01
-5.73654592e-01 1.55886143e-01 6.62873793e+00 1.17504978e+00
-7.24765599e-01 -1.55483291e-01 7.67742515e-01 1.21820018e-01
-6.28166199e-01 -1.05515659e-01 -6.16993368e-01 1.09345531e+00
1.22482479e+00 -4.14291710e-01 4.71692234e-01 3.63040775e-01
4.71159011e-01 -6.76293612e-01 -9.38236773e-01 6.30186737e-01
-1.64189354e-01 -8.46446872e-01 -5.25773764e-01 -8.49419758e-02
9.46030021e-01 -4.83611934e-02 -8.96794125e-02 2.84671277e-01
6.92765176e-01 -1.06983912e+00 6.95613384e-01 1.37274110e+00
3.01887572e-01 -1.08256602e+00 1.27740729e+00 8.59239042e-01
-6.36595488e-01 -2.14349210e-01 -2.92627633e-01 -1.69524308e-02
7.30524480e-01 1.36699951e+00 -6.87778115e-01 8.82119000e-01
7.13555336e-01 4.14239228e-01 -1.37227485e-02 1.27394950e+00
-2.45713249e-01 9.11847055e-01 -7.65784264e-01 -1.75536692e-01
-9.29402038e-02 -5.75901031e-01 4.06664789e-01 8.37069333e-01
8.35564435e-01 -8.07638019e-02 -1.05552964e-01 1.04444873e+00
1.62798747e-01 -1.48611099e-01 -3.77050459e-01 -1.26006424e-01
7.43260920e-01 8.50889564e-01 -6.68084741e-01 -2.38007993e-01
-2.85808891e-01 7.33577788e-01 -2.39575669e-01 4.87818182e-01
-8.33982646e-01 -5.98364413e-01 1.98447406e-01 -4.41105723e-01
2.22005814e-01 -3.84792835e-01 -3.21320683e-01 -1.01333153e+00
-5.94560839e-02 -4.70744073e-01 4.02394772e-01 -9.82655585e-01
-1.78009558e+00 2.83064842e-01 3.92343879e-01 -1.00363743e+00
-7.43274331e-01 -4.22393978e-01 -1.04827619e+00 1.19540370e+00
-1.52920949e+00 -6.74955130e-01 1.55630574e-01 2.50278920e-01
2.96956480e-01 -7.44994953e-02 5.33376753e-01 6.37760609e-02
-7.52521336e-01 2.12759480e-01 1.20099211e+00 -2.25084141e-01
6.39624834e-01 -1.47013855e+00 1.30915195e-01 8.01064909e-01
-4.78976369e-01 1.97868705e-01 8.53095114e-01 -9.66401696e-01
-1.24663305e+00 -9.23879087e-01 5.50657511e-01 -1.48127759e-02
1.00044215e+00 -1.09456651e-01 -9.74894106e-01 4.33073014e-01
4.26587015e-01 -2.40030214e-01 6.20061636e-01 -1.05715536e-01
2.07957372e-01 -2.79581189e-01 -1.44044363e+00 2.29292810e-01
3.43088031e-01 -1.60266802e-01 -4.80451405e-01 2.56821603e-01
2.96992034e-01 -2.43347391e-01 -1.36019337e+00 3.71860027e-01
2.06501931e-01 -6.10118628e-01 5.12604415e-01 -3.15800488e-01
-1.34776175e-01 -5.42628586e-01 -2.89079212e-02 -2.10150504e+00
-1.50963962e-01 -6.31340444e-01 -9.48853940e-02 1.81807292e+00
3.70636761e-01 -1.12837625e+00 3.42059195e-01 8.81614566e-01
-9.29690078e-02 -3.61140698e-01 -1.29569483e+00 -9.68386114e-01
3.65620613e-01 -7.86159039e-01 1.11451817e+00 7.35346138e-01
1.44712508e-01 -2.37660855e-01 -4.68767762e-01 3.92552257e-01
8.05515409e-01 -1.72004297e-01 3.69706541e-01 -1.20544386e+00
-1.46880314e-01 -4.49722767e-01 -1.01565227e-01 -6.95480883e-01
2.54117846e-01 -5.11882782e-01 3.66042405e-01 -1.25784636e+00
-1.22280456e-01 -4.49776649e-01 -3.71199310e-01 9.51939002e-02
-1.07571930e-01 -2.09847614e-01 -1.26365632e-01 -1.40847176e-01
-1.78337961e-01 9.73762035e-01 4.23099846e-01 -6.19798973e-02
1.23018049e-01 4.49086547e-01 -2.79149026e-01 7.28868365e-01
1.18300259e+00 -5.25578678e-01 -4.25634563e-01 -3.35075289e-01
6.55800164e-01 2.37313554e-01 -3.13271396e-02 -8.31679523e-01
4.89389777e-01 -4.31078166e-01 4.12907124e-01 -1.04371727e+00
-1.45560071e-01 -1.06887436e+00 8.84293199e-01 3.72234017e-01
-2.92331100e-01 2.38351747e-01 3.11832249e-01 9.31007266e-01
-3.26122731e-01 -7.90473640e-01 2.74166048e-01 1.63991779e-01
-4.08607334e-01 7.61847794e-02 -8.18771005e-01 -9.97457951e-02
1.15652490e+00 1.45881772e-01 -2.54551619e-01 -7.03553438e-01
-6.88627481e-01 6.53462350e-01 5.76314703e-02 7.42015094e-02
3.75751317e-01 -1.29690301e+00 -9.21275735e-01 -4.41253260e-02
-1.89584509e-01 1.63228605e-02 4.49082583e-01 8.38962615e-01
-1.55618936e-01 2.12857977e-01 3.09023291e-01 -4.71301943e-01
-3.47139686e-01 5.45149922e-01 3.69681478e-01 -2.62975872e-01
-6.57080486e-02 2.76172668e-01 -4.07243460e-01 -5.51250398e-01
2.48266175e-01 -4.65282500e-01 -1.93159372e-01 2.24263296e-01
4.01528060e-01 8.38169813e-01 1.03849843e-01 -2.64814138e-01
-7.41724744e-02 3.11647415e-01 5.97859919e-01 -1.32042050e-01
1.24013722e+00 -4.69728291e-01 -1.08455755e-01 9.86186087e-01
7.22234607e-01 -1.66092351e-01 -1.50821912e+00 -2.95227438e-01
4.26293135e-01 -2.22022474e-01 3.19849283e-01 -1.22848940e+00
-9.20078039e-01 8.68980765e-01 5.02675891e-01 5.51208675e-01
1.11049473e+00 -5.63954532e-01 4.02435601e-01 5.73410429e-02
3.15197229e-01 -1.53080094e+00 -6.03891134e-01 4.40404058e-01
7.52163649e-01 -1.11761785e+00 2.79528983e-02 1.59691185e-01
-6.76517129e-01 1.37233555e+00 1.28778160e-01 3.11852574e-01
1.02320123e+00 6.77266598e-01 -1.89091936e-01 4.71515894e-01
-6.17726266e-01 8.32887646e-03 2.37053469e-01 5.48767567e-01
3.80193144e-02 1.77368701e-01 -3.33411336e-01 8.51547539e-01
1.98159426e-01 1.76660761e-01 6.55972362e-01 9.26050901e-01
-1.69884235e-01 -8.93036067e-01 -5.28445482e-01 5.90529025e-01
-5.34690201e-01 -1.61716551e-01 3.47257316e-01 3.44025135e-01
-4.09019664e-02 1.05673349e+00 2.29710683e-01 1.77886374e-02
3.51435065e-01 2.45402321e-01 -1.06249653e-01 -5.81427813e-02
-2.72524953e-01 1.54771924e-01 3.96926468e-03 3.02595906e-02
-2.73277462e-01 -9.25569057e-01 -1.12605023e+00 -4.70430106e-01
-6.28378689e-01 3.80692542e-01 9.82683599e-01 1.03394854e+00
2.80076891e-01 6.33545935e-01 8.21359456e-01 -7.30547488e-01
-1.36790299e+00 -1.25844705e+00 -1.07802951e+00 4.99600507e-02
-2.10993420e-02 -5.21976113e-01 -7.63446569e-01 -3.14536601e-01] | [6.1752142906188965, 2.9737255573272705] |
6bf38a7e-4dc5-453e-b584-6cc672bdf19c | long-range-constraints-for-neural-texture | 2211.11137 | null | https://arxiv.org/abs/2211.11137v1 | https://arxiv.org/pdf/2211.11137v1.pdf | Long Range Constraints for Neural Texture Synthesis Using Sliced Wasserstein Loss | In the past decade, exemplar-based texture synthesis algorithms have seen strong gains in performance by matching statistics of deep convolutional neural networks. However, these algorithms require regularization terms or user-added spatial tags to capture long range constraints in images. Having access to a user-added spatial tag for all situations is not always feasible, and regularization terms can be difficult to tune. It would be ideal to create an algorithm that does not have any of the aforementioned drawbacks. Thus, we propose a new set of statistics for exemplar based texture synthesis based on Sliced Wasserstein Loss and create a multi-scale algorithm to synthesize textures without a user-added spatial tag. Lastly, we study the ability of our proposed algorithm to capture long range constraints in images and compare our results to other exemplar-based neural texture synthesis algorithms. | ['Albert Chua', 'Liping Yin'] | 2022-11-21 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 2.06261098e-01 -2.50252128e-01 2.60002941e-01 -4.60561723e-01
-6.01178586e-01 -2.46749595e-01 6.29899979e-01 5.22420555e-02
-3.83757293e-01 8.70049536e-01 -2.12802216e-01 -7.32515529e-02
-5.91512382e-01 -9.40317929e-01 -8.95861328e-01 -8.23094368e-01
2.64731497e-02 3.90864998e-01 3.73843282e-01 -1.94006130e-01
3.76625806e-01 8.57895374e-01 -1.65577793e+00 1.98717877e-01
7.45224118e-01 1.31822228e+00 3.70081395e-01 3.66672575e-01
-1.96518958e-01 3.10563415e-01 -2.25749791e-01 -2.87328750e-01
5.06993830e-01 -4.17835563e-01 -3.07560027e-01 1.20019175e-01
8.84642899e-01 -7.50860572e-02 -1.87109888e-01 9.19903517e-01
5.35626888e-01 3.05120021e-01 8.01891804e-01 -9.72132146e-01
-7.32171595e-01 3.00341457e-01 -6.64267600e-01 -1.05190895e-01
-1.59378886e-01 -4.36784886e-02 8.10635328e-01 -7.54069149e-01
6.54933333e-01 1.15756249e+00 7.56577313e-01 5.15317321e-01
-1.31127310e+00 -4.84764427e-01 -3.58481333e-02 -6.31470978e-02
-1.33064342e+00 -4.24174935e-01 1.10406983e+00 -1.58289760e-01
5.66983223e-01 3.13602507e-01 5.31733871e-01 1.09768498e+00
3.31188619e-01 5.38991034e-01 1.45577991e+00 -5.45665264e-01
3.85784477e-01 1.77394506e-02 -4.85013574e-01 1.00019836e+00
1.32954121e-01 1.20824099e-01 -5.26483297e-01 6.53431639e-02
1.39458644e+00 -8.95929486e-02 9.96447653e-02 -9.53831494e-01
-1.21747291e+00 7.58234620e-01 5.11371732e-01 2.49832705e-01
-1.57337219e-01 5.41424215e-01 2.47239485e-01 2.69010901e-01
5.87070346e-01 7.86024511e-01 -3.39136660e-01 1.02526769e-01
-1.31935120e+00 2.72418380e-01 4.51144993e-01 9.35679078e-01
8.13723087e-01 1.95535034e-01 -3.06906998e-01 1.13307011e+00
-1.16020903e-01 4.18210536e-01 1.67113066e-01 -1.07472253e+00
-4.77715954e-02 3.40406299e-02 2.46029422e-01 -1.30647624e+00
-2.95296639e-01 -5.48094928e-01 -1.09484720e+00 5.12453675e-01
7.59038985e-01 1.58434734e-01 -1.02985454e+00 1.74003839e+00
1.22558363e-01 2.62822658e-01 -2.99008340e-01 1.00812316e+00
3.94819230e-01 3.39623153e-01 -1.56076178e-01 -4.53211516e-02
8.94206285e-01 -6.86782718e-01 -5.70899069e-01 3.57173711e-01
4.29330289e-01 -9.89405811e-01 1.38640010e+00 3.60331118e-01
-1.05840158e+00 -5.34698844e-01 -1.14658201e+00 2.06951573e-01
-3.46951395e-01 5.40150218e-02 7.93753564e-01 5.84775805e-01
-1.08756006e+00 9.54294980e-01 -7.36480236e-01 -3.98690522e-01
3.65254313e-01 3.05275559e-01 -3.12488168e-01 8.92590508e-02
-9.75535929e-01 8.17875624e-01 1.32370755e-01 1.40258178e-01
-5.77690423e-01 -4.74243402e-01 -8.06336164e-01 -5.54941520e-02
4.17050093e-01 -7.13681221e-01 7.10682034e-01 -9.88242626e-01
-1.50053144e+00 6.73774362e-01 -4.06702273e-02 -2.89417535e-01
7.44414628e-01 5.79412803e-02 -1.27766296e-01 -1.34092376e-01
-7.65052959e-02 8.76978338e-01 9.78027225e-01 -1.39345050e+00
-4.35120314e-01 4.96337563e-02 6.65730536e-02 -9.40868910e-03
-3.09252411e-01 -2.77664781e-01 -2.64109224e-01 -1.00348079e+00
2.73817062e-01 -9.24434245e-01 -2.32081950e-01 4.08665538e-01
-2.03094572e-01 1.43170163e-01 9.27277148e-01 -1.84094355e-01
7.91947722e-01 -2.13051033e+00 1.37031712e-02 3.29278052e-01
4.47533391e-02 5.32883145e-02 -3.44531506e-01 2.27693766e-01
1.66387141e-01 1.23503424e-01 -3.61870080e-01 -2.38984674e-01
1.79725021e-01 4.12446678e-01 -1.11875348e-01 3.09407026e-01
2.67192096e-01 6.50830626e-01 -8.15433741e-01 -5.91080248e-01
5.10875225e-01 5.03604770e-01 -8.60979915e-01 -1.64505616e-01
-5.19454658e-01 4.42133486e-01 -4.63162422e-01 5.56306064e-01
8.42876375e-01 -1.01089559e-01 -9.67932120e-02 -3.88914913e-01
-7.11632445e-02 -1.94478154e-01 -1.33135056e+00 1.63691044e+00
-7.01949656e-01 5.95450163e-01 -1.97315589e-01 -1.06925464e+00
1.13805163e+00 2.08019435e-01 7.30868816e-01 -9.46631491e-01
-9.16381087e-03 4.97078627e-01 -2.89804339e-02 -3.64089727e-01
5.63511968e-01 -2.69786477e-01 -1.19870909e-01 3.69078845e-01
-9.05435160e-03 -2.96208471e-01 1.56820878e-01 -3.37410569e-01
1.13089597e+00 4.69970524e-01 -8.62113088e-02 -5.11641204e-01
3.15391630e-01 -1.59606770e-01 6.09162867e-01 1.00860465e+00
-5.47379628e-02 1.01407766e+00 3.72025490e-01 -8.12619746e-01
-1.64552820e+00 -1.23098516e+00 -4.77551252e-01 1.01555336e+00
2.08187057e-03 1.05437368e-01 -6.50195658e-01 -4.45345312e-01
7.37481117e-02 2.15459645e-01 -7.68440783e-01 1.25601575e-01
-6.32229805e-01 -7.24319458e-01 5.00850379e-01 3.15543264e-01
7.22284794e-01 -1.12680554e+00 -5.71399033e-01 3.21987480e-01
7.33769033e-03 -9.37791765e-01 -3.97374541e-01 3.97620320e-01
-7.56688595e-01 -5.90581119e-01 -9.62166667e-01 -7.13149667e-01
7.84169853e-01 9.54981893e-02 1.02912998e+00 5.56318387e-02
-3.73751670e-01 5.74509166e-02 -4.18344051e-01 -2.11640731e-01
-3.73679549e-02 1.16841950e-01 8.57321247e-02 2.17062205e-01
5.28832041e-02 -7.09070027e-01 -7.42491007e-01 5.32571733e-01
-1.01235497e+00 1.62712976e-01 5.66666067e-01 1.09096098e+00
8.66916180e-01 2.06332490e-01 5.52579403e-01 -7.92985678e-01
5.02387464e-01 -8.10364038e-02 -5.74564159e-01 2.17332587e-01
-2.81183898e-01 3.86154860e-01 7.55696833e-01 -6.01426363e-01
-1.03546274e+00 2.02353805e-01 -6.11533783e-02 -4.84753072e-01
-1.83273435e-01 4.30772334e-01 3.36026959e-02 -4.75531429e-01
6.42765462e-01 1.33971095e-01 -1.19462430e-01 -4.02812511e-01
3.76996696e-01 3.35090697e-01 3.08591604e-01 -1.05569696e+00
6.05010867e-01 5.46418250e-01 3.68468881e-01 -8.41317594e-01
-9.31250691e-01 3.09708039e-03 -8.02740157e-01 -1.07850686e-01
6.70909107e-01 -3.94200414e-01 -4.47106838e-01 4.91292387e-01
-8.26316655e-01 -5.57242632e-01 -3.47788572e-01 3.22222054e-01
-1.11188662e+00 1.96701601e-01 -3.40397626e-01 -5.57615280e-01
1.32537335e-02 -1.30107796e+00 1.05509853e+00 6.51088506e-02
-1.13507606e-01 -1.01414263e+00 -2.81391382e-01 -3.38758156e-02
7.99650967e-01 5.76163411e-01 1.03361094e+00 -1.87580481e-01
-8.47412825e-01 5.81254400e-02 -5.20167351e-01 1.41732678e-01
2.93820262e-01 4.20717821e-02 -8.23456287e-01 -2.90389270e-01
-2.11854905e-01 -3.35313946e-01 9.64125693e-01 6.47427499e-01
1.58309674e+00 -4.43619490e-02 1.15520628e-02 9.71627653e-01
1.46289706e+00 1.35689914e-01 6.95637822e-01 5.12783051e-01
7.41980970e-01 4.83743876e-01 5.50380468e-01 3.71619433e-01
1.30456891e-02 8.60838354e-01 1.63986832e-01 -2.02965766e-01
-3.31032813e-01 -2.08054688e-02 -2.35661879e-01 7.02301383e-01
-2.28972420e-01 -2.46442422e-01 -6.48552120e-01 4.18791443e-01
-1.90620017e+00 -8.95639598e-01 2.93387957e-02 2.32377267e+00
7.78587520e-01 2.78266907e-01 -1.18876494e-01 1.16147533e-01
4.61500794e-01 1.93347260e-01 -4.85505253e-01 -5.31436026e-01
-3.18620145e-01 4.73707169e-01 5.47562957e-01 3.66230786e-01
-1.24078298e+00 1.10722780e+00 6.50089550e+00 1.20244014e+00
-1.40291452e+00 -9.91897583e-02 6.07319772e-01 -1.15423940e-01
-4.89356220e-01 -1.24598689e-01 -4.76303458e-01 3.63041639e-01
4.32584524e-01 2.27507114e-01 4.82798964e-01 5.22086084e-01
3.12820792e-01 -2.69591928e-01 -1.03554678e+00 9.34724808e-01
3.03584952e-02 -1.34831107e+00 2.11620286e-01 3.36914621e-02
9.22051370e-01 -3.27186644e-01 4.47186112e-01 -1.62864059e-01
2.86294788e-01 -1.15584314e+00 6.69748664e-01 8.70097816e-01
1.02967596e+00 -8.43471289e-01 5.77883959e-01 -2.73369011e-02
-9.76059496e-01 2.32201457e-01 -7.18780160e-01 6.40124381e-02
-7.35287964e-02 7.15508521e-01 -5.16661763e-01 2.67884970e-01
6.23108804e-01 4.18615699e-01 -4.77392823e-01 1.26585531e+00
2.72133231e-01 2.50127822e-01 -5.10506988e-01 -1.87661201e-01
4.87509489e-01 -2.93060809e-01 4.46380168e-01 1.03686666e+00
7.63841450e-01 -2.69841403e-01 -4.42709401e-02 9.30433154e-01
1.16681438e-02 6.80006072e-02 -7.68387675e-01 2.52825677e-01
3.70414943e-01 1.14682508e+00 -1.09721243e+00 -1.15014397e-01
-2.79757589e-01 9.93929088e-01 4.10094202e-01 3.85280758e-01
-7.20588207e-01 -4.19130892e-01 4.82285589e-01 3.71891886e-01
3.23578566e-01 -4.92997497e-01 -5.08685827e-01 -9.12703335e-01
-5.89947961e-03 -8.01033676e-01 -1.38176724e-01 -8.85257661e-01
-1.31920767e+00 4.42597538e-01 -9.09779817e-02 -1.23853242e+00
4.85595912e-02 -6.58074260e-01 -3.23745370e-01 8.38802814e-01
-1.27806878e+00 -1.41453564e+00 -2.69004494e-01 5.14130473e-01
3.56532633e-01 -1.83249354e-01 9.29823935e-01 4.13648963e-01
-2.69007891e-01 8.69210005e-01 2.82411337e-01 -1.26657054e-01
9.23970878e-01 -1.24217355e+00 1.86420918e-01 4.45384413e-01
1.89232290e-01 5.50517321e-01 9.30679202e-01 -5.82428038e-01
-1.16358840e+00 -1.02451372e+00 6.16091669e-01 -2.42990240e-01
4.84304816e-01 -4.32719320e-01 -9.45688426e-01 4.26988363e-01
-3.85194458e-02 1.54002458e-01 3.63213450e-01 2.00253546e-01
-2.73289055e-01 -4.49991345e-01 -1.14256561e+00 8.76194060e-01
1.15300262e+00 -4.38177913e-01 -1.44547507e-01 2.14244291e-01
1.64130554e-01 -2.95551151e-01 -8.13758969e-01 7.16085553e-01
8.43223751e-01 -9.85238492e-01 9.95437622e-01 -3.20642710e-01
5.72267890e-01 -2.49757469e-01 -3.55469376e-01 -1.45467591e+00
-5.01546919e-01 -4.50223207e-01 4.59821045e-01 1.03569186e+00
3.86458695e-01 -5.33295393e-01 9.62383568e-01 3.77554059e-01
-2.16076672e-01 -8.69115055e-01 -9.13202107e-01 -9.74592447e-01
2.80637175e-01 -2.23313570e-01 6.45976543e-01 1.05128574e+00
-2.70007432e-01 -3.52511108e-01 -7.49027133e-01 -7.85525665e-02
7.76961386e-01 1.33342311e-01 7.11536705e-01 -1.22089565e+00
-1.74852550e-01 -6.62538826e-01 -2.60462075e-01 -8.09530079e-01
-2.50577573e-02 -7.36498475e-01 1.20447882e-01 -1.43530536e+00
1.89432576e-01 -1.31019092e+00 -6.67777717e-01 3.57235014e-01
2.13547394e-01 7.02137291e-01 -8.95547047e-02 2.08350010e-02
-4.26409721e-01 6.12105012e-01 1.76660490e+00 -2.82312557e-02
6.16272166e-02 -3.08746696e-01 -3.76405001e-01 6.84347808e-01
8.00927401e-01 -3.79245579e-01 -4.62081671e-01 -5.76585591e-01
3.82441550e-01 4.09523733e-02 3.37906599e-01 -1.17587948e+00
1.02294900e-01 -4.81215686e-01 6.01785362e-01 -2.79158801e-01
4.95475829e-01 -8.18918824e-01 3.22772175e-01 -1.16880141e-01
-3.90227705e-01 -1.93470880e-01 1.66931793e-01 3.63120675e-01
-2.09771082e-01 -1.25906765e-01 1.05891001e+00 -3.54757905e-01
-5.70286274e-01 3.76066118e-01 -2.54129857e-01 -2.00696260e-01
8.11033487e-01 -5.68075180e-01 -4.20733206e-02 -1.83631822e-01
-8.03501248e-01 -2.68921584e-01 8.38744581e-01 3.37338686e-01
6.01215661e-01 -1.62039733e+00 -3.91249239e-01 3.29576075e-01
7.55610988e-02 -2.02789679e-01 2.21278131e-01 6.17445886e-01
-8.72140706e-01 2.22345263e-01 -7.14949667e-01 -5.23124397e-01
-8.58319461e-01 5.05851030e-01 4.65786159e-01 -2.10253701e-01
-5.93836963e-01 9.16020751e-01 2.87619293e-01 -5.11543751e-01
1.80338576e-01 -2.52684057e-01 1.93934247e-01 -1.57094225e-01
7.09496345e-03 -7.04996586e-02 5.10838404e-02 -3.54792327e-01
-9.89479423e-02 8.51544738e-01 1.27458777e-02 -1.24711782e-01
1.37294376e+00 -6.21404648e-02 8.55217129e-02 4.53577936e-01
9.43017781e-01 -5.81649691e-02 -1.51521564e+00 -2.80428588e-01
-1.87805653e-01 -7.30929196e-01 1.95309654e-01 -7.54860878e-01
-1.18530369e+00 9.02501106e-01 6.16874456e-01 6.62479028e-02
1.14573050e+00 -4.38413858e-01 9.07836616e-01 4.30748492e-01
6.87300920e-01 -1.36630785e+00 3.72890562e-01 6.00435078e-01
8.41985703e-01 -1.28122962e+00 -1.30630299e-01 -3.13446760e-01
-2.57263422e-01 9.90651608e-01 5.66290081e-01 -5.41536868e-01
7.67026484e-01 4.10195410e-01 1.34293184e-01 -1.71835124e-01
-4.86169368e-01 -1.67611554e-01 3.75273347e-01 6.46431565e-01
5.69578230e-01 8.80997926e-02 -3.32078636e-01 -1.70768291e-01
-1.74107626e-01 -1.24039434e-01 3.28264207e-01 8.23997617e-01
-4.09039319e-01 -1.27968097e+00 -3.20999712e-01 6.89664423e-01
-4.44971979e-01 -1.59105137e-01 4.33894880e-02 7.46598542e-01
1.65878817e-01 4.74660397e-01 2.28556946e-01 -3.55795771e-01
2.90156066e-01 -2.03259811e-02 8.47670972e-01 -2.58049279e-01
-5.42963222e-02 2.66497552e-01 1.35225235e-02 -4.26651955e-01
-3.97133797e-01 -4.89893079e-01 -7.88002014e-01 -5.48478484e-01
-1.31345898e-01 -1.59237340e-01 6.48021817e-01 8.11455309e-01
1.61946461e-01 5.15417695e-01 5.14226258e-01 -9.43955839e-01
-1.84758887e-01 -8.63280594e-01 -8.46238315e-01 5.83546102e-01
1.87821835e-01 -1.17225873e+00 -1.53304085e-01 4.02606875e-02] | [11.468525886535645, -0.5854077935218811] |
f1e001e4-c3f8-46eb-a6c2-618cf2fead60 | can-learning-from-natural-image-denoising-be | 1902.10379 | null | https://arxiv.org/abs/1902.10379v3 | https://arxiv.org/pdf/1902.10379v3.pdf | Can learning from natural image denoising be used for seismic data interpolation? | We propose a convolutional neural network (CNN) denoising based method for seismic data interpolation. It provides a simple and efficient way to break though the lack problem of geophysical training labels that are often required by deep learning methods. The new method consists of two steps: (1) Train a set of CNN denoisers from natural image clean-noisy pairs to learn denoising; (2) Integrate the trained CNN denoisers into project onto convex set (POCS) framework to perform seismic data interpolation. The method alleviates the demanding of seismic big data with similar features as applications of end-to-end deep learning on seismic data interpolation. Additionally, the proposed method is flexible for many cases of traces missing because missing cases are not involved in the training step, and thus it is of plug-and-play nature. These indicate the high generalizability of our approach and the reduction of the need of the problem-specific training. Primary results on synthetic and field data show promising interpolation performances of the presented CNN-POCS method in terms of signal-to-noise ratio, de-aliasing and weak-feature reconstruction, in comparison with traditional $f$-$x$ prediction filtering and curvelet transform based POCS methods. | ['Xiuyan Yang', 'Hao Zhang', 'Jianwei Ma'] | 2019-02-27 | null | null | null | null | ['de-aliasing'] | ['computer-vision'] | [-4.79556955e-02 4.27120514e-02 7.94171572e-01 -3.59000295e-01
-9.14434195e-01 -7.10892230e-02 4.68427896e-01 -1.24199502e-01
-6.61634982e-01 8.50449204e-01 1.53367847e-01 -1.61694556e-01
-4.04771119e-01 -9.91720378e-01 -1.03800690e+00 -9.06762719e-01
-2.12277129e-01 1.18494324e-01 6.26815036e-02 -5.90723932e-01
-2.52212323e-02 6.49600267e-01 -1.56013203e+00 1.85416549e-01
9.38265860e-01 1.25338829e+00 2.45202348e-01 5.32899320e-01
4.04004427e-03 8.22158277e-01 -3.47203344e-01 -1.97761655e-01
5.86326361e-01 -3.03383976e-01 -4.57544327e-01 -2.83391297e-01
1.82978183e-01 -3.77617300e-01 -3.24430108e-01 9.00892198e-01
7.14829147e-01 3.59795183e-01 5.24643660e-01 -5.51471353e-01
-4.17227626e-01 4.52889979e-01 -5.87176740e-01 2.36874651e-02
3.06417868e-02 1.89087316e-01 4.46971238e-01 -1.27154517e+00
4.00228709e-01 1.07937622e+00 1.55148757e+00 2.21940696e-01
-1.39706588e+00 -5.50313532e-01 -5.44690669e-01 7.82912523e-02
-1.19782007e+00 -5.90157390e-01 8.87521923e-01 -5.33425212e-01
7.18967557e-01 4.03100818e-01 5.79578638e-01 8.09103251e-01
-1.72606573e-01 2.59002537e-01 1.10075104e+00 -3.66596580e-01
2.38823727e-01 -3.65380049e-01 -6.08792678e-02 3.49393904e-01
-1.02957105e-02 2.76739776e-01 -5.20111561e-01 -2.23850198e-02
9.83749509e-01 -8.32464620e-02 -2.01205313e-01 1.02657288e-01
-1.13438666e+00 7.32375383e-01 7.35573649e-01 3.64049792e-01
-6.86362863e-01 5.38195312e-01 5.85384309e-01 4.84432608e-01
8.31092119e-01 2.80303866e-01 -1.78538635e-01 1.06050797e-01
-1.49361897e+00 5.26948452e-01 5.12894273e-01 4.85220581e-01
9.00899410e-01 8.32675576e-01 9.80244949e-02 9.86425400e-01
2.14561686e-01 4.77129608e-01 3.47503334e-01 -1.20683753e+00
2.57316798e-01 1.20695299e-02 2.41542116e-01 -1.03180218e+00
-3.70890766e-01 -5.93804240e-01 -1.45039976e+00 6.87704027e-01
3.98952663e-01 -1.75173908e-01 -7.64970779e-01 1.41764486e+00
2.43878797e-01 4.04753774e-01 1.21507257e-01 9.10162985e-01
8.70463789e-01 7.62808383e-01 -6.35250360e-02 2.55622785e-03
1.06370032e+00 -5.74584484e-01 -7.58167744e-01 -1.25761032e-01
1.91009328e-01 -8.05972815e-01 8.40070784e-01 5.18498898e-01
-1.27455378e+00 -1.04089379e+00 -8.80626798e-01 -2.89633367e-02
-1.65595651e-01 -4.50269766e-02 4.43659902e-01 4.72750574e-01
-1.30665600e+00 1.23408437e+00 -8.79118145e-01 1.78565770e-01
4.89775479e-01 3.71584296e-01 -4.32320625e-01 1.40099376e-01
-1.18910837e+00 7.26006031e-01 4.19320077e-01 8.41423035e-01
-9.16385055e-01 -9.32162106e-01 -7.40947187e-01 1.44820794e-01
-1.15541421e-01 -5.26023030e-01 7.82469153e-01 -1.23236084e+00
-1.35347426e+00 5.52944660e-01 7.35008195e-02 -6.33880675e-01
9.65047300e-01 -4.12188172e-01 -2.99905419e-01 -3.95570733e-02
1.60243407e-01 4.59378511e-01 1.10113871e+00 -1.37342238e+00
-3.80151898e-01 9.99334008e-02 -4.45191354e-01 -2.88168460e-01
-1.35237440e-01 -1.93622217e-01 2.13237647e-02 -1.08150065e+00
4.39246982e-01 -2.54968673e-01 -4.69769984e-01 1.95179820e-01
-1.77201524e-01 1.10820800e-01 6.44110620e-01 -1.26097226e+00
7.11749673e-01 -2.33049560e+00 -5.03557250e-02 4.08290058e-01
6.22849911e-02 4.53141034e-01 -8.02397504e-02 6.78397417e-01
-6.13465726e-01 -9.49700251e-02 -7.50092030e-01 -5.18455386e-01
-2.04067782e-01 1.76419377e-01 -3.45859766e-01 5.73098361e-01
3.22370350e-01 4.46991056e-01 -6.69146240e-01 -2.41741940e-01
4.48618501e-01 7.59626985e-01 -5.39812803e-01 2.94411808e-01
3.78590189e-02 1.00778306e+00 -1.19775180e-02 3.72324616e-01
1.14821899e+00 4.96535242e-01 -4.45009470e-01 -3.62176955e-01
-4.26199198e-01 -1.70173526e-01 -1.68523669e+00 1.67022920e+00
-5.31826138e-01 6.26033425e-01 4.03663307e-01 -1.46595931e+00
1.25238705e+00 4.71579194e-01 4.57846612e-01 -6.63235724e-01
-1.82950541e-01 7.88217247e-01 -3.81847143e-01 -8.59774530e-01
3.22058648e-01 -5.00621378e-01 2.44491935e-01 -9.99550223e-02
1.42521605e-01 -2.27054775e-01 -4.19020876e-02 -3.94144952e-01
8.03131819e-01 5.46081722e-01 -9.14811045e-02 -5.65907001e-01
6.92186356e-01 -1.81059092e-01 5.90591431e-01 8.10743153e-01
3.20825726e-01 1.05407596e+00 2.86043972e-01 -7.00220108e-01
-1.42675936e+00 -7.91342139e-01 -2.47694999e-01 7.18070328e-01
-4.31771487e-01 3.19016367e-01 -9.04571235e-01 2.39812672e-01
-1.53623641e-01 5.54304421e-01 -5.74582160e-01 2.66832978e-01
-1.09473205e+00 -6.41932964e-01 7.76853442e-01 3.98724407e-01
6.77086473e-01 -1.28386021e+00 -4.71230775e-01 5.43776989e-01
-3.56285810e-01 -7.59865582e-01 2.41167694e-02 4.22582477e-01
-1.22656560e+00 -6.57629669e-01 -9.15159762e-01 -9.45754528e-01
4.62604582e-01 -1.47051498e-01 9.76622164e-01 1.08847745e-01
4.87561934e-02 -1.01736955e-01 -2.40347743e-01 -1.39588729e-01
-5.08338809e-01 -2.08819970e-01 -2.98496366e-01 3.60718101e-01
-2.17618316e-01 -1.13950586e+00 -9.30154145e-01 -4.27313475e-03
-9.18934762e-01 -8.93114284e-02 3.48708421e-01 1.19058990e+00
4.95407999e-01 5.11248335e-02 7.27897525e-01 -5.82733750e-01
6.30521953e-01 -5.22357047e-01 -7.08387554e-01 -3.10224771e-01
-3.22619140e-01 5.54417679e-03 8.65478754e-01 -8.62922613e-03
-1.08217001e+00 3.08324903e-01 -9.46866095e-01 -2.96190381e-01
-2.72145659e-01 7.46899247e-01 2.48542637e-01 -3.04687530e-01
1.01476038e+00 2.84964919e-01 1.18901394e-01 -9.23738301e-01
6.21038489e-02 4.76782531e-01 1.06101358e+00 -4.84875083e-01
9.25319016e-01 6.47165120e-01 2.25209415e-01 -1.14630437e+00
-3.64001691e-01 -3.88141990e-01 -6.07172132e-01 -3.04331064e-01
8.49168539e-01 -1.08918142e+00 -6.60830081e-01 7.42322147e-01
-1.37992191e+00 -3.57738227e-01 -7.31449604e-01 3.62173140e-01
-5.49547732e-01 6.12883806e-01 -7.51605630e-01 -8.01772952e-01
-5.52306771e-01 -1.09903884e+00 8.41047406e-01 4.68133129e-02
2.14090452e-01 -8.60914528e-01 1.18322473e-03 3.51547673e-02
7.14364707e-01 8.28661084e-01 5.14953256e-01 -2.51998633e-01
-3.68029267e-01 -3.46515983e-01 -1.48561761e-01 9.46264088e-01
-1.96796745e-01 -1.60235360e-01 -1.36393273e+00 -2.35184208e-01
5.38878918e-01 -1.26714870e-01 1.06248367e+00 6.28814340e-01
1.19238460e+00 -2.92863846e-01 3.95880580e-01 1.05372024e+00
1.76696467e+00 -1.37919381e-01 1.18370378e+00 4.62419808e-01
5.97769558e-01 6.77278817e-01 6.31227940e-02 4.44769144e-01
-8.44698548e-02 4.22849476e-01 8.31523001e-01 -5.50547361e-01
-2.36637294e-01 5.53945191e-02 1.30551979e-01 8.14922631e-01
-6.78269029e-01 1.62214160e-01 -9.35117006e-01 7.05055237e-01
-1.80189669e+00 -1.10221136e+00 -7.58656681e-01 2.12736821e+00
8.26475680e-01 1.69881503e-03 -6.69901147e-02 7.08945513e-01
5.41895390e-01 1.10401720e-01 -3.11901537e-03 -2.92939216e-01
-2.14856669e-01 8.63494396e-01 5.87171197e-01 5.96520841e-01
-1.21343696e+00 4.84757990e-01 5.27929878e+00 1.01230490e+00
-1.15280592e+00 3.70300680e-01 5.15812933e-01 5.68338573e-01
-2.87085116e-01 -1.46391556e-01 -1.60026833e-01 4.18931007e-01
8.83409441e-01 5.10388553e-01 4.79149222e-01 6.21751666e-01
5.41416883e-01 2.19780579e-02 -7.50721097e-01 1.04234266e+00
-3.57714206e-01 -1.65237343e+00 -2.92273432e-01 -3.74795020e-01
5.59829175e-01 4.96638343e-02 -1.65059432e-01 2.89266240e-02
3.00533567e-02 -1.16083896e+00 1.08181286e+00 1.00414205e+00
7.01003313e-01 -7.57017732e-01 8.68668854e-01 4.23232228e-01
-1.05755496e+00 -1.33976396e-02 -4.16483670e-01 -2.41311178e-01
3.98317158e-01 9.34325814e-01 -3.58635128e-01 7.96459079e-01
1.02703965e+00 5.08560121e-01 -1.35849208e-01 1.33656013e+00
-1.28420204e-01 6.97444022e-01 -3.91995519e-01 4.88049954e-01
3.86864483e-01 -4.77952570e-01 5.08695543e-01 1.52117717e+00
7.04896390e-01 2.73611210e-02 -2.39821985e-01 8.65197241e-01
1.77526623e-01 -1.40587777e-01 -4.53648537e-01 6.26083910e-01
8.48876387e-02 1.02510750e+00 -5.51420152e-01 -1.68709353e-01
-1.69188783e-01 6.79630935e-01 -1.50214836e-01 4.32998300e-01
-6.10734344e-01 -4.68451291e-01 2.98042715e-01 3.75300109e-01
5.52282810e-01 -2.56974012e-01 -6.73009038e-01 -8.08573544e-01
1.59418210e-01 -7.39966750e-01 5.54402322e-02 -8.11238050e-01
-1.29346979e+00 8.11990082e-01 -2.76044101e-01 -1.53279901e+00
-1.99717999e-01 -2.36271754e-01 -7.38021612e-01 1.25962234e+00
-1.72233558e+00 -1.22170544e+00 -6.61859095e-01 7.24237442e-01
5.64663708e-01 8.35607499e-02 8.21755707e-01 8.57201219e-01
-5.43715293e-03 2.47067079e-01 5.42214632e-01 1.82715997e-01
4.53411102e-01 -1.02758718e+00 4.81813431e-01 1.19434452e+00
-4.04421061e-01 2.08155930e-01 1.03314316e+00 -4.70851451e-01
-9.81772542e-01 -1.03591514e+00 7.46962368e-01 2.17313886e-01
4.61158812e-01 -2.61047304e-01 -1.30631971e+00 3.57571453e-01
2.35857904e-01 3.13095272e-01 1.07442349e-01 -3.82672220e-01
-6.60386831e-02 -5.63881576e-01 -1.36826956e+00 3.02746385e-01
5.76524675e-01 -2.69300044e-01 -5.27912021e-01 3.14332247e-01
4.22389060e-01 -5.34795344e-01 -9.18026090e-01 4.70176131e-01
3.80729973e-01 -1.35251534e+00 1.44155085e+00 -8.39097276e-02
6.71595216e-01 -4.38907266e-01 -1.40460297e-01 -1.30070865e+00
-2.64460713e-01 -7.27522135e-01 3.44216615e-01 1.16919482e+00
2.20570013e-01 -6.80396020e-01 5.28780162e-01 2.90826291e-01
-7.37981915e-01 -3.09481770e-01 -1.45308518e+00 -6.18554354e-01
1.61503240e-01 -5.85251331e-01 5.20069957e-01 9.35690463e-01
-6.99797869e-01 -3.68563503e-01 -6.58542633e-01 2.94133782e-01
1.03240037e+00 -4.36634451e-01 7.13253558e-01 -1.31791055e+00
-3.87660116e-01 -1.52354702e-01 -1.06252454e-01 -7.25978494e-01
-2.58097976e-01 -5.78765512e-01 4.06557798e-01 -1.54720187e+00
-6.26876593e-01 -7.06679940e-01 -1.13773473e-01 3.67442310e-01
1.17730528e-01 5.36904156e-01 -1.97316483e-02 3.70391935e-01
2.82998770e-01 5.23624420e-01 1.18547153e+00 -1.30285369e-03
-6.82789758e-02 1.25765309e-01 -1.87765241e-01 8.54579806e-01
5.54772079e-01 -6.70798659e-01 5.60865477e-02 -7.82165408e-01
2.51112133e-01 4.41617548e-01 7.69228220e-01 -1.33250082e+00
2.02273801e-01 1.51501402e-01 3.81446213e-01 -4.40674186e-01
2.73530841e-01 -1.07269204e+00 7.78476834e-01 6.67712450e-01
4.27680314e-02 -1.58790514e-01 1.57365918e-01 4.02405292e-01
-5.70131361e-01 -5.24457991e-01 1.13302052e+00 -2.18045846e-01
-7.39176452e-01 -6.31305724e-02 -2.31403351e-01 -2.35537097e-01
4.74752843e-01 -6.72755659e-01 2.07342729e-01 -2.97700197e-01
-9.94753540e-01 -3.32568109e-01 -3.38786617e-02 -2.42131874e-01
4.61220384e-01 -1.21347916e+00 -1.09866345e+00 3.71624261e-01
-4.12809402e-01 5.16791940e-01 5.53749800e-01 9.08882260e-01
-1.28378570e+00 -3.58753264e-01 -3.40967447e-01 -6.49232566e-01
-5.84036231e-01 2.07267150e-01 5.92366755e-01 -5.68753220e-02
-1.00018227e+00 9.21988606e-01 -1.63845867e-01 -4.16365802e-01
1.23965465e-01 -4.15940732e-01 -6.15434647e-02 1.41676873e-01
2.90991813e-01 6.31273568e-01 4.35150951e-01 -5.42680681e-01
-7.43650123e-02 4.94405925e-01 6.83776021e-01 -2.18638629e-01
2.21025491e+00 1.47460297e-01 -4.04339284e-01 1.78953677e-01
1.11130404e+00 -1.19757429e-01 -1.60076869e+00 -4.73503582e-02
-3.51491459e-02 -5.06485224e-01 1.03316166e-01 -4.19267952e-01
-1.39384174e+00 1.00792730e+00 8.88881624e-01 2.90954351e-01
1.27813876e+00 -5.03254652e-01 8.09179902e-01 1.64051160e-01
1.41733428e-02 -1.08746278e+00 -1.93568379e-01 4.65272069e-01
1.34587824e+00 -9.30907011e-01 -1.71263024e-01 -1.71434820e-01
-1.08643331e-01 1.47359729e+00 -5.15234172e-02 -7.58981168e-01
8.38000238e-01 2.93436766e-01 1.14918472e-02 -1.95852622e-01
1.72046512e-01 9.33560804e-02 -6.35157451e-02 7.01767325e-01
4.15088356e-01 -2.66353548e-01 -3.85248393e-01 4.83333766e-01
-2.28137210e-01 1.95373252e-01 4.94373679e-01 7.33842790e-01
-3.95946950e-01 -8.31120253e-01 -7.82933652e-01 2.11978018e-01
-4.54765946e-01 -3.12630057e-01 6.36431217e-01 6.97819471e-01
4.71939772e-01 6.87671423e-01 -8.93397406e-02 -9.52962413e-02
6.65005326e-01 -9.73190367e-02 -4.51457873e-02 -1.51060015e-01
-1.01631522e+00 4.04870987e-01 9.12134051e-02 -2.92125404e-01
-5.17032146e-01 -6.16379261e-01 -9.64953363e-01 -3.79594117e-01
-2.73026705e-01 1.04673773e-01 7.34994471e-01 8.90446961e-01
2.04691663e-02 6.97883129e-01 5.90808749e-01 -1.49998057e+00
-5.77431858e-01 -1.01064229e+00 -6.29650235e-01 5.68605125e-01
6.80606127e-01 -4.04290766e-01 -4.44559038e-01 4.60981488e-01] | [11.53306770324707, -2.305431842803955] |
7551251a-e704-4c38-a6c6-0fd7281771a1 | save-spectral-shift-aware-adaptation-of-image | 2305.18670 | null | https://arxiv.org/abs/2305.18670v1 | https://arxiv.org/pdf/2305.18670v1.pdf | SAVE: Spectral-Shift-Aware Adaptation of Image Diffusion Models for Text-guided Video Editing | Text-to-Image (T2I) diffusion models have achieved remarkable success in synthesizing high-quality images conditioned on text prompts. Recent methods have tried to replicate the success by either training text-to-video (T2V) models on a very large number of text-video pairs or adapting T2I models on text-video pairs independently. Although the latter is computationally less expensive, it still takes a significant amount of time for per-video adaption. To address this issue, we propose SAVE, a novel spectral-shift-aware adaptation framework, in which we fine-tune the spectral shift of the parameter space instead of the parameters themselves. Specifically, we take the spectral decomposition of the pre-trained T2I weights and only control the change in the corresponding singular values, i.e. spectral shift, while freezing the corresponding singular vectors. To avoid drastic drift from the original T2I weights, we introduce a spectral shift regularizer that confines the spectral shift to be more restricted for large singular values and more relaxed for small singular values. Since we are only dealing with spectral shifts, the proposed method reduces the adaptation time significantly (approx. 10 times) and has fewer resource constrains for training. Such attributes posit SAVE to be more suitable for real-world applications, e.g. editing undesirable content during video streaming. We validate the effectiveness of SAVE with an extensive experimental evaluation under different settings, e.g. style transfer, object replacement, privacy preservation, etc. | ['Nazanin Rahnavard', 'Chen Chen', 'Mohsen Joneidi', 'Umar Khalid', 'Nazmul Karim'] | 2023-05-30 | null | null | null | null | ['style-transfer'] | ['computer-vision'] | [ 6.14692152e-01 -2.56964326e-01 -2.04585288e-02 -2.86230862e-01
-5.19410491e-01 -5.93847334e-01 4.99189794e-01 -1.79698035e-01
-6.50600076e-01 5.52225292e-01 7.27254748e-02 -1.67597368e-01
-4.44274880e-02 -4.18579787e-01 -7.99309254e-01 -8.85675073e-01
1.87838599e-01 1.50308490e-01 3.38600844e-01 -1.67647954e-02
3.54934990e-01 2.25463122e-01 -1.23282123e+00 1.28597230e-01
9.73421693e-01 9.76559997e-01 2.67743260e-01 5.95275283e-01
-1.29997477e-01 4.24442530e-01 -5.64174891e-01 -5.46513200e-01
5.21806717e-01 -5.44290304e-01 -4.56459105e-01 4.24291611e-01
4.41588104e-01 -2.25249320e-01 -3.66019368e-01 1.26786506e+00
5.11686623e-01 3.97306055e-01 5.00991642e-01 -1.08122587e+00
-6.10238492e-01 4.29013819e-01 -7.47046649e-01 3.03660721e-01
1.83159679e-01 8.37170780e-02 6.24216914e-01 -9.06535625e-01
6.48900747e-01 1.02444315e+00 5.13277948e-01 7.05262125e-01
-1.33499479e+00 -7.02045441e-01 3.03859085e-01 2.92781800e-01
-1.51415777e+00 -7.08623350e-01 7.76486039e-01 -3.27448338e-01
7.15589762e-01 4.46538180e-01 4.77405310e-01 1.18161452e+00
1.98819805e-02 7.01423168e-01 8.14752936e-01 -4.58834291e-01
2.07192898e-01 3.09443414e-01 -3.90853941e-01 3.63816440e-01
5.74003905e-02 -4.14843172e-01 -5.45416117e-01 -1.88384339e-01
9.31826651e-01 -1.00847064e-02 -7.41124272e-01 -5.38018525e-01
-1.40423059e+00 7.28483021e-01 7.46209780e-03 3.09762299e-01
-2.60634691e-01 3.23614776e-02 5.13144910e-01 5.79516470e-01
5.07300377e-01 2.37026840e-01 -3.68121624e-01 -1.01819642e-01
-9.26274538e-01 7.86365345e-02 4.70715791e-01 1.14288116e+00
7.00399876e-01 1.12327300e-01 -2.57144660e-01 1.05925095e+00
3.49517986e-02 4.66916859e-01 8.20149064e-01 -1.00071120e+00
7.98178971e-01 1.13409214e-01 1.60264835e-01 -9.58042026e-01
-1.58505500e-04 -2.06053972e-01 -1.03785193e+00 -1.02270149e-01
3.33661348e-01 -2.75161594e-01 -8.25831294e-01 1.79916644e+00
3.37320179e-01 3.10212046e-01 -1.20736666e-01 9.66432333e-01
1.80934161e-01 7.67696381e-01 -2.35295713e-01 -5.34298658e-01
1.02005339e+00 -8.67312193e-01 -8.14728081e-01 -1.37877569e-01
5.16325831e-01 -9.02996778e-01 1.24644685e+00 3.18330675e-01
-1.08531296e+00 -4.26276028e-01 -9.90489304e-01 2.34339043e-01
-7.38351494e-02 2.11990356e-01 6.12699538e-02 7.30895281e-01
-9.88013148e-01 5.99104047e-01 -5.61004460e-01 -3.58381867e-01
1.89887792e-01 4.60806191e-01 -4.77947950e-01 4.70510013e-02
-1.15960240e+00 3.71698231e-01 3.99205416e-01 -5.71870543e-02
-4.17362452e-01 -6.32600904e-01 -5.63488305e-01 1.32467613e-01
7.83548892e-01 -5.96427143e-01 8.28224361e-01 -1.42948949e+00
-1.81566310e+00 6.98173881e-01 -1.39946088e-01 -4.29815859e-01
8.27588320e-01 -7.28433654e-02 -4.30590302e-01 1.67610466e-01
-1.45417154e-01 5.95829964e-01 1.67637205e+00 -1.02585173e+00
-5.85809827e-01 -1.92251325e-01 -1.38301551e-01 3.70239705e-01
-9.79233503e-01 7.29176775e-02 -9.22640204e-01 -1.09270751e+00
-3.03574298e-02 -1.16824305e+00 -1.15427911e-01 1.85095295e-01
-1.21855266e-01 1.17721081e-01 8.01692903e-01 -5.71989238e-01
1.40708947e+00 -2.59307456e+00 3.86599004e-01 2.98743904e-01
1.29833743e-01 5.79813898e-01 -3.69741529e-01 3.18333507e-01
-4.58561406e-02 8.34242851e-02 -3.40671629e-01 -4.28935528e-01
-3.12759817e-01 2.13687792e-01 -9.73119065e-02 3.94753754e-01
-4.19165241e-03 4.04462934e-01 -6.34075999e-01 -4.92369771e-01
6.91716000e-02 4.54505533e-01 -8.99554670e-01 9.48197320e-02
-8.55087638e-02 4.87359941e-01 -4.03614253e-01 1.28581554e-01
7.07141757e-01 -1.42713860e-01 1.92341119e-01 -2.89295435e-01
4.54049520e-02 -1.33482531e-01 -1.39642918e+00 1.62838531e+00
-4.58809584e-01 6.27700746e-01 2.58895218e-01 -1.07639921e+00
7.60755420e-01 4.47583854e-01 6.12631321e-01 -3.74257594e-01
7.04717729e-03 1.08350553e-01 -1.18441753e-01 -5.97135246e-01
3.93643469e-01 -6.88760281e-02 3.45326245e-01 1.54477865e-01
-3.77256833e-02 1.88781898e-02 2.50360668e-01 7.62539133e-02
7.10173190e-01 -2.24551242e-02 9.99607667e-02 -1.53479055e-01
7.57044494e-01 -6.25208080e-01 7.50352621e-01 5.22891998e-01
-1.68873847e-01 9.42151666e-01 3.64207953e-01 -1.69041991e-01
-1.17075229e+00 -6.36858284e-01 1.54349685e-01 1.06143248e+00
8.57845470e-02 -3.94422054e-01 -1.08103931e+00 -7.65740931e-01
-1.64473504e-01 7.39497542e-01 -5.30281484e-01 -1.95355713e-01
-7.12970912e-01 -7.85891771e-01 3.54858607e-01 2.74685234e-01
5.81670105e-01 -7.41005182e-01 -3.95598322e-01 3.12752336e-01
-3.67041886e-01 -1.20697975e+00 -1.25760388e+00 -1.31360054e-01
-8.79679978e-01 -5.55034339e-01 -1.04190040e+00 -6.41476512e-01
8.23462307e-01 4.69264656e-01 6.05198801e-01 -1.66721586e-02
4.38243300e-02 4.85140622e-01 -5.17292321e-01 -3.71195413e-02
-3.57950091e-01 6.26103655e-02 9.82017666e-02 7.83168793e-01
-1.30826786e-01 -6.34581149e-01 -5.65808535e-01 5.86315930e-01
-1.42327654e+00 1.84093136e-02 3.59163672e-01 9.69474971e-01
5.74196041e-01 1.12227619e-01 4.23433304e-01 -8.47406328e-01
7.10260749e-01 -1.88561156e-01 -4.72285450e-01 3.78345758e-01
-6.45474374e-01 3.38127427e-02 8.44217718e-01 -1.02462149e+00
-1.20341098e+00 3.02658360e-02 -5.28023839e-02 -8.48875761e-01
1.26052901e-01 2.70231962e-01 -2.01845407e-01 -3.17100912e-01
4.99619365e-01 4.07704741e-01 -5.02345748e-02 -4.24344480e-01
2.59503245e-01 7.27991521e-01 4.03780341e-01 -4.81876463e-01
9.02044237e-01 3.22448432e-01 -2.34231263e-01 -9.79118824e-01
-2.66997904e-01 -3.65058750e-01 -5.08868396e-01 -9.79444310e-02
5.94446898e-01 -7.44248807e-01 -3.37455213e-01 4.00955677e-01
-9.25258279e-01 -3.85272413e-01 -2.98203200e-01 5.82938313e-01
-6.47849858e-01 8.28073859e-01 -3.07858407e-01 -4.57755387e-01
-3.48679960e-01 -1.08221591e+00 6.49839997e-01 6.94987103e-02
-6.97290748e-02 -8.49746287e-01 -6.74930513e-02 2.51489937e-01
4.30591613e-01 -2.88575143e-01 8.72119784e-01 -5.05932987e-01
-3.66243571e-01 -1.93326488e-01 -1.43721685e-01 5.82496583e-01
1.41150132e-01 -3.77716497e-02 -6.08608484e-01 -5.87820709e-01
2.78483719e-01 1.00752927e-01 7.81535447e-01 2.29273722e-01
1.27710199e+00 -5.28887868e-01 -1.79300845e-01 9.02486145e-01
1.23606122e+00 4.01119798e-01 7.09401667e-01 3.33026350e-01
7.59275377e-01 3.49284738e-01 5.64058006e-01 6.99433565e-01
-3.72041240e-02 9.83799934e-01 6.55509830e-02 1.50822401e-01
-1.28412366e-01 -1.08482473e-01 4.70865935e-01 9.39707816e-01
-2.21710861e-01 -4.84451801e-01 -4.83182967e-01 4.96585935e-01
-1.86312366e+00 -8.09647441e-01 3.13195854e-01 2.64372325e+00
1.05182326e+00 1.20693505e-01 1.64109524e-02 1.04351021e-01
8.09094310e-01 1.41520783e-01 -5.70672154e-01 -2.02401131e-01
-1.02677993e-01 -1.60332099e-01 6.81688190e-01 5.06023645e-01
-9.38435793e-01 9.27905440e-01 5.44446754e+00 1.20631623e+00
-1.29907358e+00 1.63029432e-01 5.23930490e-01 -4.29472566e-01
-3.64741743e-01 -2.66772658e-02 -4.56144631e-01 6.56158090e-01
7.27831602e-01 -1.27528772e-01 8.54473948e-01 4.80347991e-01
2.43220061e-01 2.13008404e-01 -8.51413190e-01 1.20582366e+00
1.33194715e-01 -1.18969822e+00 3.26856464e-01 -1.43400177e-01
6.67114854e-01 -2.20035002e-01 1.60437956e-01 9.96956136e-03
-3.24050426e-01 -5.53672671e-01 7.34866321e-01 3.11382383e-01
9.28105652e-01 -6.48911715e-01 4.30912584e-01 2.44742647e-01
-9.25313532e-01 -1.92617923e-01 -4.42606747e-01 4.33003992e-01
2.42697839e-02 4.65816617e-01 -4.76292491e-01 4.10444558e-01
7.19984472e-01 5.58011115e-01 -4.35781330e-01 8.96093309e-01
2.08592460e-01 6.03741705e-01 -5.15612006e-01 1.45560324e-01
1.02897950e-01 -5.85155189e-01 7.35568464e-01 1.22392237e+00
5.86147845e-01 1.59423709e-01 6.61688074e-02 3.92840773e-01
-1.21094920e-01 4.35761809e-01 -5.44903934e-01 -5.06989360e-02
3.28513563e-01 9.96324539e-01 -6.22374773e-01 -4.06757206e-01
-5.76199770e-01 1.44336498e+00 9.07603130e-02 6.98521256e-01
-8.47890437e-01 -4.77581859e-01 6.13058925e-01 2.56539583e-01
6.66495562e-01 -2.11907402e-01 -3.72154750e-02 -1.50161827e+00
2.65913606e-01 -1.05683470e+00 2.09358409e-01 -5.84700644e-01
-1.02355421e+00 6.84987605e-01 6.43166825e-02 -1.37637472e+00
-2.94347942e-01 -2.63869703e-01 -3.11111987e-01 5.59393525e-01
-1.28409851e+00 -7.53284812e-01 -6.79235384e-02 9.60420907e-01
7.29441404e-01 -2.40084231e-01 5.16640544e-01 5.14974833e-01
-6.69336379e-01 1.09803283e+00 5.47876656e-01 -9.71522033e-02
1.14072287e+00 -7.08914936e-01 2.54154801e-01 8.15462410e-01
1.88066766e-01 5.94343185e-01 7.55755067e-01 -5.27968884e-01
-1.26183462e+00 -9.54117537e-01 6.27094686e-01 9.00349244e-02
5.03269196e-01 -3.53199989e-01 -1.06305695e+00 6.09865546e-01
1.82986617e-01 -3.45366485e-02 4.77967769e-01 -4.36821640e-01
-2.56111681e-01 -4.40240592e-01 -1.04193377e+00 9.50933933e-01
1.07109475e+00 -3.66671681e-01 8.05286784e-03 3.18876714e-01
6.08282149e-01 -3.20582807e-01 -9.21022713e-01 2.83985972e-01
3.77069980e-01 -9.43337560e-01 9.82343197e-01 -1.95152849e-01
-1.41166272e-02 -4.13339168e-01 -1.59042597e-01 -1.28940594e+00
-3.66443992e-01 -1.15837824e+00 1.03768483e-01 1.33559120e+00
2.74123609e-01 -7.40802944e-01 6.08483315e-01 6.97806120e-01
9.18305740e-02 -5.58802187e-01 -1.15637159e+00 -8.21707964e-01
-2.05937803e-01 -1.29429057e-01 4.34835881e-01 9.39161897e-01
-1.55470327e-01 1.85577363e-01 -9.06878293e-01 8.35682359e-03
4.68454808e-01 -1.43302575e-01 6.65218890e-01 -8.28147590e-01
-4.79608178e-01 -4.20635045e-01 -2.61728942e-01 -1.23687160e+00
1.59197673e-02 -5.30280352e-01 7.28166336e-03 -7.88141370e-01
5.15193827e-02 -3.45886171e-01 -2.56638914e-01 2.99020469e-01
-4.11500275e-01 2.11373404e-01 4.36375290e-01 4.45413530e-01
-3.37633163e-01 7.51990736e-01 1.22169828e+00 -1.71622545e-01
-4.17393863e-01 1.88571036e-01 -3.64581585e-01 6.40273988e-01
8.74643147e-01 -4.37845558e-01 -7.01854527e-01 -6.33918524e-01
1.60264969e-01 4.20828350e-02 -1.28599852e-02 -1.00648415e+00
1.88389972e-01 -2.77362049e-01 1.97674364e-01 2.17712689e-02
2.90575564e-01 -1.01154101e+00 2.00499535e-01 1.03200786e-01
-4.62557226e-01 -1.19205400e-01 1.50825173e-01 7.93147147e-01
-1.55631557e-01 -6.04383111e-01 8.48006725e-01 1.03762209e-01
-4.68032330e-01 4.35275346e-01 -5.14144421e-01 -6.15735278e-02
9.59652066e-01 -4.03445512e-01 2.79579967e-01 -6.24649584e-01
-6.40347242e-01 -3.16267051e-02 7.08157182e-01 4.14111912e-01
6.35841846e-01 -1.24166417e+00 -6.90356016e-01 5.78863382e-01
3.02533228e-02 -2.95032829e-01 4.18567836e-01 8.83452773e-01
-1.80482954e-01 1.06926002e-01 -1.17378905e-01 -4.96412188e-01
-1.45904505e+00 8.20234597e-01 1.52024776e-01 -1.21919319e-01
-8.74094963e-01 7.29036808e-01 4.44733262e-01 -7.51328841e-02
4.22591954e-01 -2.11070448e-01 -1.37824446e-01 5.82730360e-02
4.19423431e-01 2.47817621e-01 -5.79186529e-02 -5.55934012e-01
-2.13239163e-01 9.29193437e-01 -3.07804018e-01 -2.96901435e-01
1.04319239e+00 -6.07056379e-01 1.18252501e-01 1.86283156e-01
1.20374060e+00 1.23133220e-01 -1.59601927e+00 -4.53786701e-01
-3.07256371e-01 -7.72771060e-01 7.37315938e-02 -3.35791439e-01
-1.34413576e+00 7.30234146e-01 5.75230420e-01 1.14066854e-01
1.46133840e+00 -4.96146560e-01 9.39943612e-01 4.50032562e-01
1.69614777e-01 -1.26446927e+00 9.32004750e-02 2.32554048e-01
8.51511955e-01 -9.76021945e-01 -2.06116196e-02 -3.38347852e-01
-8.50988925e-01 1.08878481e+00 3.43644947e-01 1.23880424e-01
5.17022550e-01 7.57237375e-02 6.40220791e-02 3.96809429e-01
-6.04428828e-01 3.09402823e-01 2.61321574e-01 5.37630439e-01
1.64375037e-01 -1.01232857e-01 -3.30945939e-01 1.60388276e-01
2.86876559e-01 5.93298934e-02 5.34427106e-01 7.25000858e-01
-1.43511176e-01 -1.20800781e+00 -5.45087755e-01 2.99099058e-01
-4.49687421e-01 -1.15930155e-01 -6.19738474e-02 3.64759564e-01
4.32694890e-02 8.79982352e-01 -8.77202675e-02 -4.05201048e-01
3.36163044e-01 1.55342981e-01 4.72355425e-01 -1.39761165e-01
-4.91796941e-01 4.99773324e-01 -1.32487148e-01 -4.25637752e-01
-3.92054051e-01 -7.33852684e-01 -9.72190619e-01 -3.79702836e-01
-2.94641137e-01 1.10517759e-02 4.56845701e-01 7.64049828e-01
6.17019057e-01 2.76394188e-01 7.43472040e-01 -9.04838204e-01
-7.19441056e-01 -8.41096282e-01 -5.50748110e-01 7.30386674e-01
2.07691520e-01 -4.18911397e-01 -3.02563995e-01 3.88193578e-01] | [11.114380836486816, -0.8145363926887512] |
4e49634c-35db-4418-a373-4dd8d6fc7fe6 | retrogan-a-cyclic-post-specialization-system | 2108.12941 | null | https://arxiv.org/abs/2108.12941v1 | https://arxiv.org/pdf/2108.12941v1.pdf | RetroGAN: A Cyclic Post-Specialization System for Improving Out-of-Knowledge and Rare Word Representations | Retrofitting is a technique used to move word vectors closer together or further apart in their space to reflect their relationships in a Knowledge Base (KB). However, retrofitting only works on concepts that are present in that KB. RetroGAN uses a pair of Generative Adversarial Networks (GANs) to learn a one-to-one mapping between concepts and their retrofitted counterparts. It applies that mapping (post-specializes) to handle concepts that do not appear in the original KB in a manner similar to how some natural language systems handle out-of-vocabulary entries. We test our system on three word-similarity benchmarks and a downstream sentence simplification task and achieve the state of the art (CARD-660). Altogether, our results demonstrate our system's effectiveness for out-of-knowledge and rare word generalization. | ['Peter Chin', 'Cynthia Breazeal', 'Catherine Havasi', 'Henry Lieberman', 'Yida Xin', 'Pedro Colon-Hernandez'] | 2021-08-30 | null | https://aclanthology.org/2021.findings-acl.183 | https://aclanthology.org/2021.findings-acl.183.pdf | findings-acl-2021-8 | ['word-similarity'] | ['natural-language-processing'] | [ 3.06299597e-01 7.50068128e-02 -1.63762663e-02 -2.19241232e-01
-7.51854062e-01 -7.92810857e-01 4.50773507e-01 9.17730927e-02
-7.63342977e-01 9.34580624e-01 5.73477745e-01 -3.23254168e-01
6.14974126e-02 -1.07816625e+00 -8.26302588e-01 -5.66911161e-01
4.24349099e-01 5.11024952e-01 -1.23653397e-01 -1.07791126e+00
-8.06700718e-03 4.10330325e-01 -1.15218723e+00 3.18709254e-01
9.96128261e-01 6.21454492e-02 8.54034200e-02 6.03323817e-01
-4.45041239e-01 5.73441386e-01 -1.10029793e+00 -1.00637913e+00
4.10713047e-01 -2.67661601e-01 -7.82494903e-01 -7.48683512e-01
7.71617293e-01 6.40496388e-02 -7.75080740e-01 1.06025815e+00
4.82588649e-01 4.73446816e-01 8.68909419e-01 -9.47301447e-01
-1.51426005e+00 8.31980348e-01 -1.26781970e-01 3.63852650e-01
2.91046500e-01 2.61359453e-01 1.06071198e+00 -8.06580842e-01
7.35139549e-01 1.23374259e+00 9.48927402e-01 7.94882119e-01
-1.35403216e+00 -7.99556017e-01 6.64073750e-02 1.03139289e-01
-1.63334298e+00 -1.97309047e-01 3.83905262e-01 -1.41319811e-01
1.61992431e+00 4.75894362e-01 5.66364646e-01 1.33256423e+00
2.88462043e-01 3.71574402e-01 6.00496352e-01 -4.68960822e-01
1.01748019e-01 -1.29048601e-01 2.16687813e-01 3.41156542e-01
3.24455112e-01 -2.33993933e-01 -4.30220842e-01 -2.24411845e-01
4.95739788e-01 -2.98733916e-02 -3.04986387e-01 -1.43875167e-01
-9.63382542e-01 1.05060256e+00 5.50709009e-01 3.15862477e-01
-1.79897636e-01 4.19169128e-01 4.11412358e-01 5.52798569e-01
3.58167291e-01 1.19578946e+00 -5.92161477e-01 -1.17221564e-01
-6.19599640e-01 5.88564098e-01 7.44748533e-01 1.12236810e+00
6.91314757e-01 2.53066540e-01 -7.86747515e-01 8.89186323e-01
-3.22003603e-01 6.33365691e-01 8.99154782e-01 -4.81492162e-01
6.12981677e-01 4.61173594e-01 -3.08699876e-01 -8.07956338e-01
-1.17626013e-02 -5.61794817e-01 -9.02352393e-01 -1.47549137e-01
-6.84931576e-02 -8.96651484e-03 -1.34700572e+00 2.04668260e+00
-2.43988082e-01 3.37507427e-01 5.46568871e-01 4.63397771e-01
9.89001870e-01 7.29309738e-01 4.53496456e-01 2.92638421e-01
1.20240605e+00 -8.58334124e-01 -5.97149789e-01 -5.15501261e-01
8.99279237e-01 -6.01725578e-01 1.61364675e+00 1.11302294e-01
-9.08886671e-01 -4.89398479e-01 -1.01785386e+00 -5.23816466e-01
-9.80301440e-01 -4.97807443e-01 6.17477715e-01 8.19371700e-01
-1.04917538e+00 6.21140182e-01 -2.45716959e-01 -1.56701460e-01
5.48586428e-01 2.01027498e-01 -6.35803342e-01 -2.82067239e-01
-1.79438722e+00 1.38948333e+00 6.60029650e-01 -2.98103392e-01
-5.46408296e-01 -1.51243353e+00 -1.02869606e+00 2.44678065e-01
1.03498980e-01 -1.29617667e+00 1.05706882e+00 -7.54830897e-01
-1.07787645e+00 8.09107661e-01 -5.92713319e-02 -7.34914899e-01
2.69770503e-01 -4.90062296e-01 -7.29174376e-01 -3.55036795e-01
1.20319836e-02 6.48811936e-01 7.52306700e-01 -1.01414692e+00
-2.70997733e-01 -4.12025377e-02 2.35145748e-01 3.77447993e-01
-5.08250773e-01 -3.80070299e-01 -5.70524871e-01 -1.37986004e+00
-5.02811968e-01 -7.50153124e-01 -2.06989586e-01 -4.05939639e-01
-7.42973745e-01 -1.09119155e-01 4.93011564e-01 -8.35182607e-01
1.47404575e+00 -1.96108377e+00 5.23848653e-01 3.08229685e-01
2.47135863e-01 7.55573511e-01 -5.08605301e-01 6.30986810e-01
-3.65418643e-01 3.71342629e-01 -5.38810134e-01 -4.15053785e-01
3.56068537e-02 5.52051187e-01 -7.16032743e-01 1.28200622e-02
1.91308200e-01 1.34928215e+00 -1.13116705e+00 1.11182993e-02
-9.12688896e-02 5.11848986e-01 -8.87696564e-01 -4.99860197e-02
-2.07808912e-01 -9.55907032e-02 -8.28772783e-02 1.68342188e-01
5.27222753e-01 2.90362298e-01 3.20939766e-03 -1.76085562e-01
6.52134717e-01 4.51664269e-01 -6.72551811e-01 1.95045245e+00
-6.93239570e-01 5.54311275e-01 -8.22849989e-01 -7.62748301e-01
9.57161009e-01 3.95429204e-04 -9.30970237e-02 -6.45153403e-01
-2.87938356e-01 -9.29718390e-02 -1.75276771e-02 -3.10838282e-01
1.04370821e+00 -5.75921953e-01 -4.41420615e-01 1.21476486e-01
3.48824352e-01 -4.30670887e-01 2.10467100e-01 4.82645601e-01
1.41145170e+00 -9.38689858e-02 3.94820720e-01 -1.96006641e-01
3.27263862e-01 -6.72444031e-02 4.14720267e-01 1.15340328e+00
3.36917430e-01 6.21113420e-01 3.28878552e-01 -2.97667384e-01
-1.11293852e+00 -1.47553754e+00 2.07723454e-01 1.16718340e+00
-1.23500019e-01 -6.82602286e-01 -6.34061873e-01 -8.39276373e-01
5.25612831e-01 1.55030239e+00 -1.00213385e+00 -1.06498778e+00
-5.95781744e-01 -2.93209404e-01 1.24416649e+00 6.76787674e-01
5.96831515e-02 -1.16953599e+00 1.34291455e-01 1.61201969e-01
2.68976931e-02 -6.52112842e-01 -8.12392712e-01 2.54453570e-01
-5.48300803e-01 -8.63224208e-01 -6.20668292e-01 -7.51742184e-01
6.13665938e-01 7.42732137e-02 1.70932364e+00 -7.81353861e-02
-2.06488848e-01 1.94163531e-01 -2.90819257e-01 -4.94177282e-01
-6.90123320e-01 3.96611542e-01 2.27191314e-01 -5.31520545e-01
7.82959163e-01 -5.32858729e-01 -1.99047387e-01 -1.94645360e-01
-1.28030419e+00 -3.17285836e-01 3.20091784e-01 1.08056819e+00
5.26975811e-01 -2.87714563e-02 5.74951291e-01 -1.31680238e+00
1.11764991e+00 -5.26082993e-01 -8.93563852e-02 4.73843545e-01
-3.97338599e-01 3.63415569e-01 9.10864115e-01 -6.50387824e-01
-8.84126365e-01 -4.03044403e-01 -3.80265176e-01 -5.85823894e-01
5.45026213e-02 4.36382353e-01 -2.71733493e-01 3.63053441e-01
9.14843798e-01 4.36083853e-01 -2.79588521e-01 -5.05057096e-01
1.06818688e+00 4.34901267e-01 9.47826505e-01 -4.75302786e-01
1.11974609e+00 1.40686154e-01 -3.99966985e-01 -5.19925237e-01
-1.06328833e+00 -3.15439850e-01 -3.80504191e-01 6.93881333e-01
6.45776272e-01 -9.72923696e-01 -2.16377079e-01 1.81794345e-01
-1.15863657e+00 -2.15139553e-01 -1.00034606e+00 3.01315095e-02
-2.53964126e-01 1.96903367e-02 -2.95689881e-01 -6.74150884e-02
-7.92627454e-01 -7.11849749e-01 7.32763529e-01 1.94787070e-01
-6.24743879e-01 -1.07093871e+00 3.13127518e-01 1.87725071e-02
6.57264292e-01 -2.73695648e-01 1.39450419e+00 -1.04426026e+00
6.88331127e-02 -2.13108033e-01 3.14323008e-01 7.48451948e-01
2.29606017e-01 -3.42432499e-01 -8.13555658e-01 -4.63696182e-01
-3.65494132e-01 -1.00834347e-01 1.25634301e+00 1.65083318e-03
1.19828379e+00 -4.71156925e-01 -3.23225379e-01 1.02526534e+00
1.30562186e+00 3.06605082e-03 1.25121582e+00 3.67548823e-01
9.67213213e-01 1.20361350e-01 1.10194035e-01 -2.55579576e-02
1.50380179e-01 6.77163064e-01 1.63538754e-01 -1.29834175e-01
-4.15217876e-01 -7.78724909e-01 1.18689969e-01 8.66123736e-01
2.79301375e-01 -5.15385807e-01 -8.43251407e-01 7.73116410e-01
-1.33193016e+00 -1.12549508e+00 1.82298467e-01 2.04892635e+00
1.31946278e+00 2.10157141e-01 -3.62496316e-01 -1.67563468e-01
5.47748446e-01 6.16758913e-02 -6.49686038e-01 -7.93472946e-01
-5.65070450e-01 9.67924118e-01 7.02215791e-01 5.16760767e-01
-8.57275069e-01 1.46046829e+00 6.80788231e+00 1.26558971e+00
-8.83080542e-01 2.35440865e-01 3.48731637e-01 -3.09201539e-01
-1.07932687e+00 -1.62132770e-01 -8.56196821e-01 4.83831137e-01
7.35212088e-01 -3.59516233e-01 6.00552380e-01 5.97453654e-01
-5.16483128e-01 4.47306603e-01 -1.10553408e+00 8.89153004e-01
4.16343629e-01 -1.60894561e+00 7.74345994e-01 -4.13729340e-01
9.34991300e-01 -5.28031476e-02 1.75164565e-01 1.06646085e+00
7.07195520e-01 -1.61589408e+00 3.84907186e-01 8.64109933e-01
1.02352476e+00 -1.11144757e+00 7.12472498e-01 6.56065717e-03
-6.94648087e-01 5.54609932e-02 -6.74234509e-01 4.71601337e-02
1.47904698e-02 4.92597669e-01 -9.42485988e-01 3.79849285e-01
5.63414633e-01 4.21109438e-01 -7.94397831e-01 6.77319586e-01
-5.50049484e-01 4.90470260e-01 -5.29555930e-03 -7.30615482e-02
1.27587602e-01 -6.33062199e-02 7.10971713e-01 1.54432702e+00
2.15455711e-01 -6.18422404e-02 -3.14287305e-01 8.77103984e-01
-7.22137153e-01 -3.54723223e-02 -7.56822586e-01 3.31483595e-02
9.03505325e-01 8.38474452e-01 6.98765740e-02 -6.19547665e-01
-2.28846088e-01 1.35823119e+00 6.77795708e-01 5.07847488e-01
-6.86674595e-01 -9.58675385e-01 1.43372500e+00 -9.56600234e-02
5.14379799e-01 1.35713458e-01 -4.00239944e-01 -1.14998233e+00
8.49669948e-02 -1.15206766e+00 4.16425765e-01 -9.36689019e-01
-1.66156840e+00 4.72949654e-01 9.08432435e-03 -8.31311643e-01
-7.30759129e-02 -3.16223949e-01 -5.63577771e-01 1.30830204e+00
-1.30210936e+00 -1.07854497e+00 -5.47293723e-02 5.77456117e-01
4.35522944e-01 -5.24559081e-01 1.32506859e+00 2.01366588e-01
-2.45468706e-01 1.15652251e+00 4.26907182e-01 3.31932127e-01
8.46118212e-01 -1.43599844e+00 1.25397384e+00 7.60338783e-01
2.72257119e-01 1.17302978e+00 6.51839018e-01 -7.46056318e-01
-1.22246742e+00 -1.34870446e+00 1.15766478e+00 -7.20713973e-01
6.98328555e-01 -4.54413861e-01 -1.28322256e+00 9.09264028e-01
2.46310979e-01 -2.75869463e-02 7.84651518e-01 1.80405602e-01
-1.02875853e+00 -6.34901822e-02 -1.06671214e+00 1.02539420e+00
1.36613631e+00 -7.78080583e-01 -1.37169576e+00 2.32586995e-01
1.17817366e+00 -6.37527227e-01 -6.75157428e-01 2.77037084e-01
3.48554254e-01 -3.55490148e-01 1.42798281e+00 -1.56753254e+00
5.71860433e-01 -1.55359998e-01 -2.49231145e-01 -2.24672318e+00
-5.79124629e-01 -6.30201101e-01 -1.97263867e-01 1.26984203e+00
4.59993035e-01 -6.41368389e-01 6.87306404e-01 2.12145582e-01
-3.65937442e-01 -6.33404493e-01 -8.53880763e-01 -1.05215108e+00
6.95660651e-01 -3.16233426e-01 9.87749279e-01 1.25489223e+00
-2.74771392e-01 5.04622638e-01 -1.82617322e-01 -1.27477618e-02
1.35539293e-01 -3.19409341e-01 6.37234092e-01 -8.85391355e-01
-3.51722181e-01 -4.88965541e-01 -6.39604628e-01 -9.06986952e-01
5.44647694e-01 -1.43176997e+00 -1.70132026e-01 -1.49715841e+00
1.32070959e-01 -4.94732499e-01 -4.97490257e-01 8.34303260e-01
-6.22048616e-01 4.63519394e-01 2.94624388e-01 -5.27894832e-02
-1.54886037e-01 7.61935711e-01 1.06478417e+00 -5.81800282e-01
-1.06845506e-01 -5.35383403e-01 -1.27337158e+00 4.53107208e-01
8.82474661e-01 -7.38879621e-01 -5.04167557e-01 -7.93515086e-01
3.57552439e-01 -5.78352273e-01 2.88468927e-01 -7.68246412e-01
1.00010663e-01 -1.88167244e-02 4.17708814e-01 -2.86261111e-01
5.22878408e-01 -5.23058832e-01 2.29404971e-01 4.34106678e-01
-6.45913601e-01 3.12374353e-01 5.26162565e-01 6.27279162e-01
-4.70159017e-02 -4.91657019e-01 4.98118848e-01 -1.07208237e-01
-8.82951558e-01 -6.74890280e-02 -1.69434518e-01 7.52256215e-01
8.70796084e-01 -8.57326239e-02 -6.89516664e-01 -3.23315173e-01
-5.58179140e-01 -1.73129588e-02 4.21114951e-01 8.21032822e-01
6.42534435e-01 -1.52632821e+00 -9.25366104e-01 3.14282715e-01
4.71568406e-01 -7.27828965e-02 2.11203709e-01 1.34093277e-02
-5.89649856e-01 4.21788096e-01 -1.67334646e-01 2.00061481e-02
-9.57415164e-01 6.32757902e-01 4.93089199e-01 -4.03928131e-01
-6.72233105e-01 1.23970592e+00 1.49914622e-01 -9.05595958e-01
1.48568392e-01 -2.47986540e-01 -7.63146579e-02 -1.49516433e-01
5.22013545e-01 1.51772171e-01 3.67221892e-01 -4.28099036e-01
-3.20909619e-01 2.64788240e-01 -4.42177236e-01 1.95973635e-01
1.30420303e+00 2.68577754e-01 -6.39487952e-02 2.00707570e-01
1.41771090e+00 3.26828212e-01 -7.64802396e-01 -4.39030170e-01
-1.41767099e-01 -3.66194636e-01 -1.62018642e-01 -8.35499585e-01
-9.37054336e-01 4.71457005e-01 2.45881051e-01 -2.14925364e-01
9.50644255e-01 -1.04264781e-01 9.03298616e-01 8.50149035e-01
3.45602036e-02 -1.01067197e+00 -1.08072788e-01 9.04855669e-01
1.10771537e+00 -6.55775547e-01 -2.77457237e-02 -3.56715083e-01
-8.58505130e-01 7.53650367e-01 5.20948052e-01 -2.92166084e-01
3.53621542e-01 1.45564213e-01 -2.34035496e-03 6.01157919e-02
-6.08842611e-01 -1.54518038e-01 4.84298080e-01 1.06864691e+00
2.86239952e-01 7.84714520e-02 -9.49800909e-02 6.85544550e-01
-6.02389812e-01 -2.40900278e-01 4.21950221e-01 7.76123762e-01
-8.18372294e-02 -1.14219415e+00 -1.71179950e-01 7.62767255e-01
-2.09572539e-01 -9.38568592e-01 -4.54654753e-01 8.78116608e-01
9.97361392e-02 5.96387088e-01 2.79326946e-01 -4.67250943e-01
8.38129997e-01 3.47524464e-01 3.76844674e-01 -9.80220437e-01
-9.32763457e-01 -8.44773948e-01 1.19929932e-01 -4.87774581e-01
2.88324654e-01 -1.55637905e-01 -1.26081729e+00 -4.73103821e-01
-5.33628371e-03 6.57268614e-02 3.54929239e-01 8.32328677e-01
6.13731265e-01 9.40805316e-01 3.14528844e-03 -2.69844413e-01
-8.68481696e-01 -9.51184332e-01 -5.17062902e-01 1.04703403e+00
1.40767932e-01 -4.06157196e-01 -3.94412763e-02 1.22208018e-02] | [10.605355262756348, 8.704318046569824] |
0a2a36d8-4b64-45e5-aa55-3806d83f80fd | discovering-object-masks-with-transformers-1 | 2206.06363 | null | https://arxiv.org/abs/2206.06363v1 | https://arxiv.org/pdf/2206.06363v1.pdf | Discovering Object Masks with Transformers for Unsupervised Semantic Segmentation | The task of unsupervised semantic segmentation aims to cluster pixels into semantically meaningful groups. Specifically, pixels assigned to the same cluster should share high-level semantic properties like their object or part category. This paper presents MaskDistill: a novel framework for unsupervised semantic segmentation based on three key ideas. First, we advocate a data-driven strategy to generate object masks that serve as a pixel grouping prior for semantic segmentation. This approach omits handcrafted priors, which are often designed for specific scene compositions and limit the applicability of competing frameworks. Second, MaskDistill clusters the object masks to obtain pseudo-ground-truth for training an initial object segmentation model. Third, we leverage this model to filter out low-quality object masks. This strategy mitigates the noise in our pixel grouping prior and results in a clean collection of masks which we use to train a final segmentation model. By combining these components, we can considerably outperform previous works for unsupervised semantic segmentation on PASCAL (+11% mIoU) and COCO (+4% mask AP50). Interestingly, as opposed to existing approaches, our framework does not latch onto low-level image cues and is not limited to object-centric datasets. The code and models will be made available. | ['Luc van Gool', 'Simon Vandenhende', 'Wouter Van Gansbeke'] | 2022-06-13 | discovering-object-masks-with-transformers | https://arxiv.org/abs/2206.06363 | https://arxiv.org/pdf/2206.06363.pdf | null | ['unsupervised-semantic-segmentation'] | ['computer-vision'] | [ 6.49699867e-01 3.43808889e-01 -1.19552299e-01 -4.87985551e-01
-6.90746307e-01 -6.49610400e-01 5.82478642e-01 3.96170586e-01
-5.31939268e-01 3.60375911e-01 -1.19664706e-02 -5.31763136e-02
2.38349184e-01 -8.37633669e-01 -7.48066187e-01 -6.18843317e-01
3.95156413e-01 3.92668277e-01 7.69155502e-01 1.30156651e-01
3.48610610e-01 3.79525393e-01 -1.74407661e+00 3.05372357e-01
1.17309570e+00 9.33954418e-01 5.68950832e-01 4.48517799e-01
-3.30297768e-01 5.85096359e-01 -5.49804389e-01 -1.77833855e-01
2.80428469e-01 -4.01979744e-01 -1.08598554e+00 6.72391295e-01
3.87851477e-01 -2.55872756e-02 2.69484282e-01 1.20680523e+00
2.82541402e-02 2.03507140e-01 6.94765866e-01 -1.06318033e+00
-5.83804213e-02 5.13953567e-01 -4.96831059e-01 -1.92081854e-01
-6.55888245e-02 1.74956247e-01 1.19663429e+00 -8.26314211e-01
5.44936955e-01 1.10255885e+00 6.11108363e-01 3.92642111e-01
-1.44749701e+00 -1.78144857e-01 3.90326291e-01 -2.45246217e-01
-1.32372200e+00 -4.38676387e-01 7.69647062e-01 -4.96077448e-01
4.95839894e-01 3.45404357e-01 5.60166895e-01 6.86183035e-01
-4.02051717e-01 1.13888121e+00 1.10421526e+00 -4.38438743e-01
4.04428780e-01 1.61547035e-01 3.07348639e-01 5.13506830e-01
2.18104973e-01 -2.59826720e-01 -2.97479779e-01 1.58727899e-01
5.93174219e-01 -7.56623223e-02 -1.21396556e-01 -6.57529831e-01
-1.14459467e+00 6.72782660e-01 5.49555004e-01 2.92790294e-01
-4.56766665e-01 1.06527597e-01 9.52897370e-02 -3.00809592e-01
4.59032029e-01 5.91865182e-01 -4.28070664e-01 2.08317488e-01
-1.36509132e+00 2.10021049e-01 6.09200597e-01 1.00267363e+00
1.31546009e+00 -1.35653228e-01 -3.15174460e-01 9.95307267e-01
4.97541815e-01 1.24885030e-01 1.89708814e-01 -1.16667366e+00
1.59786880e-01 7.81499565e-01 1.22259952e-01 -8.13374877e-01
-3.18348408e-01 -5.09156466e-01 -2.36256450e-01 1.79531023e-01
4.95971203e-01 3.59407067e-02 -1.50579166e+00 1.42930043e+00
3.84468764e-01 8.80277008e-02 -1.44483149e-02 9.28586721e-01
8.66010964e-01 3.58817637e-01 3.89075309e-01 2.93096066e-01
1.30305469e+00 -1.20734406e+00 -3.92965257e-01 -5.46979010e-01
5.00006318e-01 -7.54190564e-01 1.12787414e+00 3.05134952e-01
-1.00649416e+00 -5.40180743e-01 -9.51549828e-01 -1.21800033e-02
-4.96847957e-01 4.50421087e-02 7.82143831e-01 7.59620905e-01
-1.09987986e+00 4.85261470e-01 -9.81352866e-01 -4.71657842e-01
7.92469263e-01 2.03837782e-01 3.50230709e-02 3.91093232e-02
-6.55902088e-01 5.10148108e-01 9.52950478e-01 -4.73917909e-02
-9.14022982e-01 -4.95411277e-01 -9.25584972e-01 -5.88732436e-02
6.86398029e-01 -5.51155567e-01 1.16739428e+00 -1.30714834e+00
-1.28318822e+00 1.09415841e+00 -1.88464656e-01 -5.40978968e-01
4.70813692e-01 -2.57020265e-01 7.65191168e-02 3.89273494e-01
3.73528600e-01 1.35881615e+00 7.91025758e-01 -1.79485154e+00
-8.18978071e-01 -1.64849460e-02 6.03618100e-02 2.71838546e-01
-1.72491446e-01 -4.43650857e-02 -1.18252122e+00 -7.89316773e-01
4.17700052e-01 -7.95670629e-01 -5.64631104e-01 -2.26314545e-01
-7.70320952e-01 4.54830118e-02 7.98178196e-01 -3.40342760e-01
9.74524915e-01 -2.09173751e+00 -3.26359197e-02 3.38709712e-01
1.01319559e-01 7.11076036e-02 -9.75433737e-02 1.41719386e-01
1.24209084e-01 2.20988721e-01 -8.31001341e-01 -6.36388242e-01
2.86405161e-02 3.46743137e-01 3.20418272e-03 2.06760928e-01
5.14352560e-01 9.38245595e-01 -9.35764015e-01 -7.24681199e-01
4.74903524e-01 1.88474715e-01 -8.02139342e-01 1.40245423e-01
-6.34212911e-01 5.88015854e-01 -4.30924416e-01 7.60944605e-01
7.19173849e-01 -2.78569937e-01 6.01729825e-02 -1.45616680e-01
-5.19652627e-02 3.66540968e-01 -1.21804023e+00 1.77595079e+00
-8.26332420e-02 3.67091089e-01 1.79129407e-01 -1.11163652e+00
8.06715965e-01 -3.32431123e-02 7.36330986e-01 -4.28255111e-01
2.93375939e-01 2.06012726e-01 -2.49569342e-01 -1.68693662e-01
6.56862557e-01 -4.18806449e-02 -1.21449754e-01 2.22460747e-01
2.04912975e-01 -6.23632967e-01 3.94041747e-01 1.62151143e-01
7.66682088e-01 3.53145808e-01 -9.97712240e-02 -5.95724106e-01
4.02437598e-01 2.54454434e-01 7.12105811e-01 9.72780228e-01
-1.40181199e-01 1.25740314e+00 3.67753744e-01 8.45725015e-02
-7.88948774e-01 -1.12260675e+00 -2.12437481e-01 9.37242866e-01
4.67437565e-01 -2.98969060e-01 -1.21113014e+00 -8.32655489e-01
-1.12757340e-01 7.12579191e-01 -4.08554614e-01 1.18432932e-01
-3.55976969e-01 -8.79460871e-01 2.42744148e-01 5.26942432e-01
4.85624164e-01 -1.27586126e+00 -5.79873800e-01 1.18298039e-01
-1.48493618e-01 -1.31070185e+00 -3.42440248e-01 3.92930627e-01
-8.72522473e-01 -1.23824286e+00 -5.43644249e-01 -8.40014219e-01
1.08029950e+00 2.88511306e-01 1.30085385e+00 2.67503917e-01
-2.46648222e-01 4.30347323e-01 -5.79142213e-01 -4.14398551e-01
-3.11239064e-01 1.59534693e-01 -5.21497190e-01 3.20091069e-01
4.03836966e-01 -2.72011608e-01 -6.76012337e-01 2.68374950e-01
-1.08157933e+00 3.42380881e-01 5.98415077e-01 4.30537075e-01
8.42058718e-01 5.71939796e-02 3.24331582e-01 -1.24085689e+00
2.11292982e-01 -3.00378054e-01 -6.03304327e-01 -5.89664727e-02
-3.70755285e-01 -1.35617584e-01 4.11511987e-01 -8.88675228e-02
-9.31749046e-01 5.10222614e-01 -8.59687254e-02 -2.38756835e-01
-5.87956429e-01 3.49156618e-01 -3.97301227e-01 1.35149524e-01
4.65176642e-01 8.11248422e-02 -1.29950717e-01 -6.70126736e-01
5.91651559e-01 5.77400744e-01 7.03534544e-01 -6.78860843e-01
7.44035423e-01 6.44916773e-01 -4.98271525e-01 -8.33518267e-01
-1.03294802e+00 -7.78708458e-01 -9.03979361e-01 -4.03978266e-02
1.23454165e+00 -8.85703087e-01 -1.80023581e-01 5.54072022e-01
-7.81778812e-01 -7.28882909e-01 -5.06080449e-01 3.81361157e-01
-6.78943038e-01 4.67142373e-01 -4.56911236e-01 -6.55694306e-01
-4.94493730e-02 -1.13308036e+00 1.30833364e+00 3.84341627e-01
-3.19347948e-01 -8.60458553e-01 -4.49644506e-01 5.72540820e-01
1.08529506e-02 3.40560049e-01 5.42655706e-01 -6.21330440e-01
-7.68388212e-01 5.01362123e-02 -4.18151647e-01 5.90854108e-01
2.40236431e-01 -5.26313263e-04 -1.06006992e+00 -1.16826899e-01
-1.28683537e-01 -2.74430513e-01 1.14350164e+00 4.22628105e-01
1.43850029e+00 -1.38750812e-02 -3.73471469e-01 6.37445748e-01
1.41725409e+00 1.99938379e-02 5.36345422e-01 3.65479350e-01
1.11345160e+00 8.11566710e-01 7.75015473e-01 2.36698866e-01
4.11029607e-01 5.11825144e-01 4.14923936e-01 -5.10611176e-01
-2.57436693e-01 -2.28358880e-01 9.10089538e-03 4.29333657e-01
2.50735253e-01 -2.69051015e-01 -9.56783712e-01 9.59190905e-01
-1.84906709e+00 -4.03702319e-01 -2.89666981e-01 2.07458901e+00
8.89531791e-01 3.92253906e-01 2.16562167e-01 5.51480688e-02
7.28384674e-01 1.20782509e-01 -4.47403789e-01 -3.00080031e-02
-1.38435692e-01 3.93376917e-01 6.63175464e-01 5.42303801e-01
-1.41453588e+00 1.50443649e+00 6.15527678e+00 7.71716714e-01
-8.08136702e-01 3.55174243e-02 7.83445776e-01 2.28431642e-01
-3.73044103e-01 2.97291577e-01 -7.67206013e-01 5.03365040e-01
4.09495503e-01 4.39936310e-01 8.98014084e-02 7.24108279e-01
3.36956948e-01 -4.64835346e-01 -9.66927409e-01 5.91494739e-01
-1.01784326e-01 -1.01745057e+00 3.06860059e-02 -9.16301832e-02
9.96008277e-01 1.10504344e-01 -1.73878491e-01 -3.39900777e-02
6.80436134e-01 -9.33777928e-01 1.05164731e+00 3.02525938e-01
3.15981060e-01 -6.08983457e-01 4.12662655e-01 2.11482748e-01
-1.11992908e+00 1.85897887e-01 -1.80211693e-01 1.91715032e-01
1.35035545e-01 7.45985210e-01 -7.02415347e-01 5.86786866e-01
7.18746006e-01 7.19402194e-01 -6.32597625e-01 1.12023175e+00
-3.14229965e-01 8.04766536e-01 -4.44713622e-01 3.67620975e-01
5.15883744e-01 -3.30552608e-01 3.06042165e-01 1.27160668e+00
-1.36043087e-01 -1.25591040e-01 7.04129338e-01 1.12158060e+00
5.55865690e-02 9.49338228e-02 -2.93126069e-02 -7.48878866e-02
3.61759245e-01 1.27710390e+00 -1.56356692e+00 -5.11735141e-01
-3.92407507e-01 1.10469711e+00 1.46445081e-01 5.20650923e-01
-7.55931854e-01 -1.59372941e-01 4.68530387e-01 1.61958858e-01
5.62609434e-01 -2.10724697e-01 -7.76418626e-01 -9.91169810e-01
-3.55653577e-02 -7.99557805e-01 2.93940306e-01 -4.30794150e-01
-1.06653559e+00 3.75597030e-01 3.93011011e-02 -9.86630499e-01
1.63137048e-01 -4.87772554e-01 -5.33118606e-01 7.07917452e-01
-1.68430495e+00 -1.20970976e+00 -4.94199306e-01 3.20420206e-01
7.24097669e-01 3.48354608e-01 5.00986397e-01 2.21905857e-01
-5.52846789e-01 1.82075143e-01 -2.03906849e-01 2.50495225e-01
4.70856488e-01 -1.49687350e+00 4.03732806e-01 1.18729305e+00
2.51581460e-01 4.87376750e-01 7.09107399e-01 -6.97197914e-01
-7.63125837e-01 -1.32265866e+00 5.05545855e-01 -5.33725977e-01
3.72649968e-01 -5.22786379e-01 -9.10349727e-01 4.10639703e-01
-9.72279906e-02 4.07662317e-02 4.51158494e-01 -3.80445719e-02
-5.81947640e-02 1.41850397e-01 -9.63351786e-01 6.59199715e-01
9.85258460e-01 -3.13087583e-01 -4.34378207e-01 3.07223499e-01
6.31917596e-01 -2.49204457e-01 -6.29838407e-01 5.31504929e-01
1.05662353e-01 -1.07678080e+00 9.87716138e-01 -2.19750360e-01
3.06996763e-01 -6.53852046e-01 2.45193355e-02 -9.59424555e-01
-4.69867513e-02 -5.44340611e-01 2.70450026e-01 1.41138518e+00
3.99639636e-01 -4.49400336e-01 1.12153184e+00 5.84894359e-01
-4.22815353e-01 -5.27949154e-01 -3.70976329e-01 -7.49363482e-01
-9.90007296e-02 -6.28574193e-01 4.82263684e-01 9.08175528e-01
-5.41265428e-01 -6.96361437e-02 1.23350807e-01 2.17668429e-01
5.97642183e-01 2.19482973e-01 8.99172843e-01 -1.15292847e+00
-2.12733328e-01 -6.16960585e-01 -2.60606974e-01 -1.30374277e+00
8.17857608e-02 -9.14081812e-01 6.14138126e-01 -1.77427125e+00
1.31902710e-01 -9.12834346e-01 -4.39806968e-01 6.97607160e-01
-5.63911617e-01 6.65238678e-01 2.53918529e-01 2.41035581e-01
-7.68990338e-01 4.48207736e-01 1.01460373e+00 -5.97860105e-02
-4.62572277e-01 -1.37620404e-01 -8.85465503e-01 8.91497076e-01
8.00842524e-01 -4.65135217e-01 -4.58818346e-01 -3.19595903e-01
-3.54108244e-01 -6.44465268e-01 5.20673573e-01 -1.10133576e+00
-2.67418176e-02 -2.84837961e-01 1.14039071e-01 -6.68763399e-01
1.63963735e-01 -8.11352253e-01 -3.30513082e-02 2.42949963e-01
-1.41249329e-01 -6.84435308e-01 1.18919358e-01 3.68502945e-01
-3.00320923e-01 -4.40339684e-01 9.22029376e-01 -3.30762446e-01
-1.02843869e+00 1.27010584e-01 -4.01780844e-01 1.69080988e-01
1.00387013e+00 -6.40934646e-01 1.38658613e-01 -4.77517694e-02
-6.67073548e-01 4.57139760e-01 9.09577370e-01 5.34254730e-01
2.18114585e-01 -8.44253659e-01 -4.79999155e-01 1.36384577e-01
2.51603723e-01 6.83232367e-01 -1.35169821e-02 6.59297049e-01
-6.38369203e-01 1.89527541e-01 7.55043253e-02 -9.02761102e-01
-9.85805333e-01 4.18118268e-01 1.34584382e-01 8.72384757e-03
-6.92822576e-01 8.95276546e-01 6.57493234e-01 -4.41706449e-01
1.23947985e-01 -4.43545431e-01 -1.54609010e-01 6.77257404e-02
5.38040064e-02 -8.55420902e-02 -8.68275203e-03 -7.28208303e-01
-2.93091714e-01 5.28625190e-01 -5.87079041e-02 -1.04404502e-01
1.24058902e+00 -1.36920244e-01 -6.89287484e-02 3.20878208e-01
1.01886237e+00 9.72320959e-02 -1.71153510e+00 -1.53837427e-01
4.08537120e-01 -3.88611615e-01 6.57551736e-02 -6.85387909e-01
-1.10180819e+00 6.53704107e-01 2.41377726e-01 1.37158573e-01
1.17642975e+00 1.83251277e-01 7.44189203e-01 -1.09766319e-01
1.31298572e-01 -1.38630402e+00 7.43392706e-02 4.19804782e-01
3.74504894e-01 -1.28224850e+00 -4.07051034e-02 -1.02553797e+00
-5.09369969e-01 7.04602420e-01 6.21277392e-01 -1.59825414e-01
3.20278853e-01 1.61051631e-01 2.01314330e-01 -2.93626249e-01
-7.16082901e-02 -9.19507682e-01 5.11715531e-01 5.64112008e-01
3.28945249e-01 2.35711206e-02 -3.36803585e-01 4.23199385e-01
-2.00889349e-01 -2.42247596e-01 2.64675766e-01 1.04197490e+00
-7.23860919e-01 -1.23028636e+00 -3.47424001e-01 3.71064425e-01
-4.62582976e-01 -2.23240361e-01 -5.78539729e-01 5.38580477e-01
3.26036066e-01 9.99765158e-01 1.56632885e-01 -2.82913237e-03
1.56043500e-01 1.75275728e-02 1.89488292e-01 -1.08324790e+00
-5.24514258e-01 4.12936330e-01 -2.20915787e-02 -5.95366538e-01
-6.18516922e-01 -6.15683377e-01 -1.50949860e+00 4.32116091e-01
-4.03493851e-01 3.80156301e-02 5.59141874e-01 9.74963307e-01
2.95953959e-01 6.01122439e-01 3.16557050e-01 -1.01437378e+00
2.42509708e-01 -7.36908734e-01 -3.45822930e-01 7.70709276e-01
1.77422818e-02 -5.87994754e-01 -1.89732760e-01 4.72514957e-01] | [9.527420043945312, 0.5319823026657104] |
d6de6e86-4d33-4981-98a1-d5321d868220 | segmentation-mask-guided-end-to-end-person | 1908.10179 | null | https://arxiv.org/abs/1908.10179v1 | https://arxiv.org/pdf/1908.10179v1.pdf | Segmentation Mask Guided End-to-End Person Search | Person search aims to search for a target person among multiple images recorded by multiple surveillance cameras, which faces various challenges from both pedestrian detection and person re-identification. Besides the large intra-class variations owing to various illumination conditions, occlusions and varying poses, background clutters in the detected pedestrian bounding boxes further deteriorate the extracted features for each person, making them less discriminative. To tackle these problems, we develop a novel approach which guides the network with segmentation masks so that discriminative features can be learned invariant to the background clutters. We demonstrate that joint optimization of pedestrian detection, person re-identification and pedestrian segmentation enables to produce more discriminative features for pedestrian, and consequently leads to better person search performance. Extensive experiments on benchmark dataset CUHK-SYSU, show that our proposed model achieves the state-of-the-art performance with 86.3% mAP and 86.5 top-1 accuracy respectively. | ['Kai-Zhu Huang', 'Yao Zhao', 'Jimin Xiao', 'Dingyuan Zheng'] | 2019-08-27 | null | null | null | null | ['person-search'] | ['computer-vision'] | [-1.48834661e-01 -9.21964228e-01 1.19022802e-01 -2.10189924e-01
-5.40196240e-01 -6.02506399e-01 3.82358938e-01 -3.06924451e-02
-8.66092622e-01 6.66492641e-01 8.54365006e-02 2.43294135e-01
3.24643165e-01 -5.17447591e-01 -4.15170848e-01 -8.75135362e-01
6.42626286e-02 2.53572017e-01 5.57758987e-01 3.07463378e-01
1.74145028e-02 3.28885913e-01 -1.51609838e+00 -1.79313973e-01
6.79013789e-01 8.89130950e-01 2.18717530e-01 6.87971056e-01
4.09406155e-01 -9.76243094e-02 -8.06347728e-01 -7.19241083e-01
4.49612111e-01 1.52999520e-01 -5.18188775e-01 5.01194239e-01
7.09291518e-01 -4.70626324e-01 -5.76266885e-01 1.38176322e+00
7.67603815e-01 3.38132650e-01 6.15113437e-01 -1.11067271e+00
-5.78057230e-01 -1.34756804e-01 -1.06510806e+00 5.60763478e-01
5.41076660e-01 5.37959874e-01 5.30770183e-01 -9.09991741e-01
-6.57550665e-03 1.55821097e+00 6.81340098e-01 7.74132192e-01
-1.30090559e+00 -8.55804265e-01 3.45783770e-01 5.19019961e-01
-1.82113218e+00 -4.86703157e-01 4.04128045e-01 -3.06677878e-01
5.29050589e-01 5.61553419e-01 7.76301503e-01 1.09684825e+00
-3.97833705e-01 9.98208404e-01 1.03859210e+00 -2.40492716e-01
-1.97697997e-01 4.85398591e-01 5.57714105e-01 7.05938756e-01
6.33100331e-01 2.06827998e-01 -1.48011252e-01 -1.33119851e-01
6.96065843e-01 1.37894541e-01 -2.97025889e-01 -2.05386534e-01
-1.04715943e+00 4.12065119e-01 5.24807870e-01 -1.05282336e-01
-2.55295902e-01 -1.39685363e-01 5.78638434e-01 -2.76711166e-01
-1.87606812e-01 -2.59155780e-01 -2.65091062e-01 3.20053190e-01
-6.56315148e-01 4.25608426e-01 3.90721887e-01 1.28069103e+00
5.04295826e-01 -1.70095339e-01 -6.00832045e-01 1.03758442e+00
4.99805689e-01 9.09747422e-01 2.87412673e-01 -4.25563186e-01
5.28047383e-01 5.70129097e-01 5.21293759e-01 -1.24014068e+00
-2.68411070e-01 -7.09220231e-01 -1.04384148e+00 -3.71067613e-01
5.36236584e-01 -1.19055435e-01 -9.59279418e-01 1.41863644e+00
6.11635923e-01 3.73665988e-01 -8.94024894e-02 1.23100960e+00
1.00006092e+00 5.62000453e-01 4.96132731e-01 -9.29084048e-02
1.91781449e+00 -1.16327214e+00 -2.48771816e-01 -5.25754869e-01
-5.82247563e-02 -8.87656391e-01 6.04538560e-01 1.42763659e-01
-7.39351690e-01 -1.10390425e+00 -8.30178618e-01 4.18458194e-01
-2.70441979e-01 4.88292247e-01 2.10997373e-01 1.09000194e+00
-7.96264887e-01 -1.21513426e-01 -3.87142390e-01 -7.08983600e-01
5.69741428e-01 5.91329336e-01 -3.04550409e-01 -1.90189645e-01
-8.68317425e-01 5.29326141e-01 4.35792685e-01 4.70054060e-01
-6.84401333e-01 -2.34350309e-01 -5.83181441e-01 -1.91123933e-02
3.54481936e-01 -8.69742692e-01 9.14801300e-01 -4.51575786e-01
-9.29938018e-01 1.00795662e+00 -5.21614850e-01 -3.42991829e-01
7.45849013e-01 -3.07613045e-01 -5.16885817e-01 -1.13734484e-01
5.33637822e-01 5.90471864e-01 7.43227899e-01 -1.32224607e+00
-1.20382333e+00 -7.40389466e-01 -3.66705477e-01 2.73278892e-01
-3.55745554e-01 4.95421112e-01 -1.18347383e+00 -4.91652697e-01
-1.53596953e-01 -9.50006545e-01 -5.10967851e-01 -3.86279196e-01
-5.65296113e-01 -4.41312701e-01 8.23961735e-01 -6.89503074e-01
1.03005004e+00 -1.73001182e+00 -2.49153767e-02 1.52795181e-01
2.23957915e-02 5.20451903e-01 1.66289389e-01 -2.82802463e-01
3.04530442e-01 -2.08639130e-01 3.43297452e-01 -4.94330585e-01
3.22263129e-02 -1.05348960e-01 2.55259573e-01 6.36097729e-01
3.16446088e-02 7.82913864e-01 -6.98950052e-01 -8.31279397e-01
5.73115408e-01 3.67106110e-01 -1.46687269e-01 2.72714943e-01
5.85152328e-01 6.64616227e-01 -6.10523283e-01 1.12360990e+00
1.20082140e+00 -9.01262313e-02 -2.65393078e-01 -3.82314652e-01
-8.93867984e-02 -5.66673875e-01 -1.60944927e+00 1.07506061e+00
1.09137036e-01 2.59951890e-01 1.54269859e-01 -8.18447292e-01
7.24068105e-01 4.50619161e-02 1.27654552e-01 -6.65839732e-01
2.18696535e-01 4.08790112e-02 -2.55103827e-01 -3.87617022e-01
3.73026282e-01 4.78386253e-01 -2.95247644e-01 -2.28015795e-01
-1.23197369e-01 8.65071476e-01 3.45309973e-01 -2.19459057e-01
4.65283364e-01 -2.63595343e-01 3.15738529e-01 -3.49582911e-01
1.31618130e+00 -2.84680814e-01 7.75144219e-01 1.04120433e+00
-8.91448915e-01 3.94141972e-01 -4.72236961e-01 -7.71675289e-01
-9.33820724e-01 -1.19620287e+00 -3.01834643e-01 1.01060033e+00
6.85101151e-01 2.39019655e-02 -1.00798929e+00 -5.54834723e-01
-3.73968519e-02 7.75117874e-02 -2.20092550e-01 4.76544201e-02
-8.64522040e-01 -1.31715298e+00 5.64959645e-01 6.47871196e-01
1.13527608e+00 -6.80866420e-01 -5.31616390e-01 1.72455117e-01
-2.91682959e-01 -1.49715686e+00 -9.23180521e-01 -4.49346632e-01
-3.68282557e-01 -1.15655935e+00 -1.26905143e+00 -1.21230912e+00
7.99990296e-01 7.37216294e-01 8.85130107e-01 1.88904509e-01
-6.62931681e-01 4.67351288e-01 5.50228208e-02 -8.80207792e-02
1.54761732e-01 -1.42829791e-01 3.53707850e-01 3.41180682e-01
7.94281662e-01 5.64842708e-02 -1.16295660e+00 7.88468063e-01
-2.75545120e-01 -2.36217752e-01 3.83356631e-01 8.46502542e-01
4.82559234e-01 3.23484272e-01 1.82330549e-01 -6.68858513e-02
3.36859465e-01 -1.26869187e-01 -7.30107248e-01 5.19948602e-01
-3.30329955e-01 -6.79566562e-01 5.17291963e-01 -4.17905986e-01
-1.01026821e+00 4.08281237e-01 1.20717049e-01 -9.04101580e-02
-6.86941504e-01 -5.07507920e-01 -5.39760232e-01 -4.30720210e-01
3.69697869e-01 6.07242405e-01 -6.83913589e-01 -5.42586029e-01
9.69984606e-02 8.03422034e-01 9.03254807e-01 -6.33120835e-01
1.07016492e+00 3.21993530e-01 -1.08531289e-01 -9.11095083e-01
-5.51871598e-01 -1.17960322e+00 -9.31348205e-01 -4.36731666e-01
1.10532022e+00 -1.39942896e+00 -1.06001616e+00 8.29181254e-01
-1.32162309e+00 5.56773782e-01 4.51768637e-01 2.90579587e-01
6.26786798e-02 7.71347702e-01 -4.59940940e-01 -1.24449551e+00
-6.56206310e-01 -1.35643029e+00 1.04366028e+00 1.12312031e+00
1.11827403e-01 -5.12984872e-01 -4.83437330e-01 5.99335492e-01
7.02371523e-02 1.81698993e-01 2.78352559e-01 -4.78145808e-01
-9.67277050e-01 -5.74883163e-01 -6.92338228e-01 1.59693584e-01
1.80122629e-01 -3.40835333e-01 -9.79782939e-01 -5.89090168e-01
-3.76729339e-01 8.77499655e-02 9.33472335e-01 3.37760299e-01
9.68943894e-01 -3.22284579e-01 -8.25121164e-01 5.99145412e-01
1.32454515e+00 2.59490907e-01 2.92030036e-01 5.60365856e-01
8.62107813e-01 4.32433575e-01 5.62012613e-01 5.46745658e-01
4.96730357e-01 9.66230571e-01 -7.69788818e-03 -1.94847569e-01
-9.57846120e-02 -1.40335169e-02 -6.89987913e-02 1.67496741e-01
-2.68161118e-01 -4.87629548e-02 -7.71330774e-01 6.84462368e-01
-2.05855131e+00 -1.02569664e+00 -2.79300272e-01 2.31463814e+00
5.54257870e-01 -1.32663874e-03 7.72588551e-01 -3.10829747e-02
1.29623806e+00 -5.32021932e-02 -4.45571572e-01 4.94293660e-01
-2.58808017e-01 -2.18545377e-01 9.04109538e-01 2.99942255e-01
-1.80667794e+00 1.00241733e+00 5.90644169e+00 8.98692429e-01
-4.39974666e-01 6.04417548e-02 8.64237666e-01 1.23421565e-01
7.27020264e-01 -3.99764389e-01 -1.64641857e+00 7.92815328e-01
2.39991084e-01 -9.18218791e-02 2.13543087e-01 7.73023367e-01
1.74385905e-01 -1.47313684e-01 -8.34941030e-01 1.55955172e+00
8.68321285e-02 -7.75494277e-01 -4.98589277e-02 -1.87797293e-01
6.94969237e-01 -5.15049875e-01 2.32049033e-01 3.20755005e-01
1.03544638e-01 -8.97145748e-01 6.17478848e-01 4.17938232e-01
3.02925617e-01 -9.25990045e-01 8.54771972e-01 3.50189447e-01
-1.77290261e+00 -2.66750902e-01 -5.74432611e-01 2.62880653e-01
3.16613972e-01 4.82133850e-02 -7.95083344e-01 3.63921821e-01
1.12189388e+00 4.14714068e-01 -1.02377605e+00 1.69403827e+00
2.10330069e-01 1.56998679e-01 -3.40377480e-01 -1.09460890e-01
7.11030662e-02 -6.54797852e-02 7.20392883e-01 1.71139324e+00
-2.39283778e-02 -5.07721901e-02 8.44091177e-01 7.46105194e-01
2.82801539e-01 1.92150116e-01 -6.47245571e-02 6.61139011e-01
3.79546553e-01 1.27210593e+00 -7.38329351e-01 -4.18834656e-01
-4.95539278e-01 1.30611825e+00 3.99502665e-02 6.01964056e-01
-1.05843556e+00 1.26337498e-01 7.54385412e-01 -1.47399843e-01
3.29321086e-01 -1.81424737e-01 -9.22443252e-03 -1.25614452e+00
3.18787396e-01 -8.44349146e-01 6.46366596e-01 -1.07560106e-01
-1.53406262e+00 5.14602065e-01 7.54527794e-03 -9.68096137e-01
1.66243166e-01 -5.98724127e-01 -4.76811171e-01 1.17094219e+00
-1.49470139e+00 -1.46428680e+00 -6.89611316e-01 7.80454695e-01
8.52124572e-01 -4.40953255e-01 5.67140162e-01 7.53695309e-01
-1.21467221e+00 9.90288556e-01 -8.04697052e-02 6.12625301e-01
6.24957561e-01 -9.73292053e-01 5.37960887e-01 1.18986738e+00
-2.68282592e-01 6.48362875e-01 6.34516239e-01 -6.59587860e-01
-1.20106196e+00 -1.27412915e+00 5.10249615e-01 -4.65649039e-01
1.63205966e-01 -2.41467208e-01 -5.26747108e-01 3.62828463e-01
-5.22562116e-02 1.24260180e-01 6.10506237e-01 -3.47049870e-02
-1.51613668e-01 -2.73815393e-01 -1.11726320e+00 7.28407800e-01
1.08091593e+00 -1.13822915e-01 -1.96835399e-01 3.44124287e-01
3.16597074e-01 -3.61058563e-01 -4.99101132e-01 3.14050168e-01
6.33231103e-01 -8.94934952e-01 1.75940013e+00 -5.33286214e-01
-6.68150544e-01 -7.14374065e-01 -2.87125289e-01 -6.41264141e-01
-7.63516128e-01 -3.99835885e-01 1.10643178e-01 1.48570180e+00
-3.20537686e-02 -5.78085840e-01 9.15086746e-01 1.02030873e+00
4.68386948e-01 -3.04630697e-01 -1.03401756e+00 -9.26599920e-01
-3.95106196e-01 1.63627133e-01 6.19238019e-01 2.52937227e-01
-5.80882967e-01 2.59200871e-01 -6.67854488e-01 7.68473327e-01
1.34717298e+00 -4.96997721e-02 1.10010076e+00 -1.21663105e+00
-1.38505369e-01 -4.63941127e-01 -7.07862377e-01 -1.56044507e+00
-6.51979372e-02 -2.98019230e-01 8.91777128e-02 -1.09865010e+00
8.25811446e-01 -3.97129774e-01 -3.05071443e-01 9.98101942e-03
-7.68119454e-01 3.61363202e-01 4.11697626e-01 3.39161277e-01
-1.04648936e+00 3.55834395e-01 9.67798591e-01 -4.25421894e-01
-1.24049909e-01 4.65945959e-01 -6.24266803e-01 7.08797276e-01
4.67584044e-01 -1.12880759e-01 2.74210870e-01 -4.45465595e-01
-8.20580006e-01 -3.27220708e-01 1.00767314e+00 -1.43711305e+00
5.77361584e-01 4.52255160e-02 1.23838258e+00 -9.64608967e-01
4.86859292e-01 -8.30440938e-01 8.83476809e-02 5.65944672e-01
2.40746215e-01 5.39913997e-02 3.10964435e-01 8.73617589e-01
8.09315406e-03 -1.55265242e-01 1.00926197e+00 -3.53926450e-01
-1.11194026e+00 6.02651417e-01 -8.30139313e-03 -2.62383312e-01
1.13419843e+00 -5.86402476e-01 -1.93366095e-01 5.10283671e-02
-5.37487149e-01 5.17243981e-01 2.42729455e-01 3.29985261e-01
5.24307251e-01 -1.45020628e+00 -7.94752717e-01 2.57807702e-01
1.40991909e-02 -1.18299454e-01 4.02562350e-01 6.39909029e-01
-2.16033518e-01 6.63918316e-01 -7.92609900e-02 -9.21462178e-01
-1.90261745e+00 6.08140290e-01 4.74153012e-01 1.94570571e-02
-3.27820182e-01 1.08306956e+00 4.21524465e-01 -1.18371226e-01
5.12046993e-01 3.23482454e-01 -4.56590146e-01 -9.65704098e-02
7.53647268e-01 6.86292946e-01 -4.65024620e-01 -1.35614395e+00
-6.72926307e-01 9.74730968e-01 -3.18135500e-01 3.31500828e-01
6.44791543e-01 -5.94011247e-01 2.57055819e-01 -3.31229657e-01
1.04701793e+00 -1.99342027e-01 -1.24239516e+00 -4.87176031e-01
-7.17567876e-02 -1.01528454e+00 -3.07976753e-01 -5.71201801e-01
-9.22196209e-01 4.93600100e-01 1.14014554e+00 7.43798837e-02
9.71953690e-01 -2.11929917e-01 1.10872948e+00 3.65025967e-01
5.38959026e-01 -1.18046582e+00 1.84224062e-02 2.54885495e-01
4.99002337e-01 -1.59451461e+00 9.39021632e-02 -6.84912741e-01
-3.98820251e-01 9.00816798e-01 9.93097961e-01 -1.54509936e-02
3.39954168e-01 -2.83824623e-01 -8.36261585e-02 2.88482696e-01
8.49203318e-02 -5.87487817e-01 5.10704637e-01 9.40454245e-01
-1.67512640e-01 2.15317845e-01 1.40574664e-01 8.14344168e-01
5.71251940e-03 -3.67517382e-01 -2.29128003e-01 4.23113495e-01
-5.92704952e-01 -8.85647655e-01 -1.13462818e+00 1.83106691e-01
-4.05631453e-01 1.48160428e-01 -1.83243752e-01 5.96417427e-01
4.74419594e-01 1.24850535e+00 -1.97914377e-01 -2.63350606e-01
6.18344486e-01 -2.29702100e-01 4.74149525e-01 -2.34785274e-01
-4.60603803e-01 3.42228323e-01 -5.69397993e-02 -2.08564952e-01
-4.40679252e-01 -9.50156271e-01 -7.09140301e-01 -3.73750865e-01
-5.08244157e-01 -3.64274681e-02 1.27886578e-01 7.73330688e-01
4.20205668e-02 1.92635641e-01 5.88087082e-01 -8.21812153e-01
-6.11834466e-01 -7.40793705e-01 -3.28866124e-01 6.14812195e-01
2.74011046e-01 -7.14386642e-01 -3.30527984e-02 1.64853185e-01] | [14.781792640686035, 0.8371515274047852] |
52b991e1-ff4e-4a2a-9b07-3c0acacba52a | deep-unsupervised-anomaly-detection | null | null | https://openaccess.thecvf.com/content/WACV2021/papers/Li_Deep_Unsupervised_Anomaly_Detection_WACV_2021_paper.pdf | https://openaccess.thecvf.com/content/WACV2021/papers/Li_Deep_Unsupervised_Anomaly_Detection_WACV_2021_paper.pdf | Deep unsupervised anomaly detection | This paper proposes a novel method to detect anomalies in large datasets under a fully unsupervised setting. The key idea behind our algorithm is to learn the representation underlying normal data. To this end, we leverage the latest clustering technique suitable for handling high dimensional data. This hypothesis provides a reliable starting point for normal data selection. We train an autoencoder from the normal data subset, and iterate between hypothesizing nor-mal candidate subset based on clustering and representation learning. The reconstruction error from the learned autoencoder serves as a scoring function to assess the normality of the data. Experimental results on several public benchmark datasets show that the proposed method outperforms state-of-the-art unsupervised techniques and is comparable to semi-supervised techniques in most cases | ['Wen-Yan Lin', 'Siying Liu', 'Zheng Wang', 'Tangqing Li'] | 2021-01-05 | null | null | null | winter-conference-on-applications-of-computer-4 | ['unsupervised-anomaly-detection-with-specified-5', 'unsupervised-anomaly-detection-with-specified-4', 'unsupervised-anomaly-detection-with-specified-7', 'unsupervised-anomaly-detection-with-specified-6', 'unsupervised-anomaly-detection-with-specified'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-1.42714083e-01 1.35070896e-02 -4.56761196e-03 -4.86177683e-01
-4.73477632e-01 -2.68348247e-01 5.47404706e-01 4.18631941e-01
-2.06640705e-01 1.51888788e-01 4.48043011e-02 -7.18855113e-02
-3.46468091e-01 -9.57970977e-01 -5.09034395e-01 -1.04965520e+00
-3.03822219e-01 9.03880298e-01 2.33021099e-02 6.19083419e-02
3.01856905e-01 4.75138038e-01 -1.71455061e+00 2.73665309e-01
6.74138308e-01 1.12862158e+00 -2.72725821e-01 4.56293851e-01
-1.00894861e-01 7.54762769e-01 -5.85936904e-01 -3.30189317e-02
3.21607471e-01 -3.95776778e-01 -6.03963256e-01 4.38160956e-01
3.63266058e-02 -2.80309081e-01 -2.85216302e-01 1.01729667e+00
2.56890148e-01 2.98243225e-01 8.19477260e-01 -1.18529475e+00
-7.40442649e-02 6.52161539e-01 -3.62526149e-01 3.45486641e-01
2.59810776e-01 -1.43537372e-01 1.24893320e+00 -8.73193681e-01
2.95969963e-01 7.96994507e-01 6.00890279e-01 2.95413464e-01
-1.23677170e+00 -3.07987720e-01 -3.79523896e-02 2.45908707e-01
-1.47233784e+00 -3.99869561e-01 8.40325952e-01 -5.78637779e-01
6.89330935e-01 3.01978942e-02 5.18606246e-01 1.08807611e+00
-6.76240250e-02 7.25929081e-01 7.41089880e-01 -4.47420150e-01
6.28442466e-01 1.09502718e-01 3.57865155e-01 4.63467449e-01
4.97179568e-01 -1.22115342e-02 -1.99380025e-01 -6.32021487e-01
2.84783751e-01 3.92663717e-01 -9.76757780e-02 -4.85452980e-01
-7.60207713e-01 1.07236123e+00 -2.38727480e-02 4.27811325e-01
-7.90228784e-01 -4.07305986e-01 6.24662757e-01 5.35855234e-01
4.96924192e-01 4.36620384e-01 -4.06976998e-01 -1.78624745e-02
-1.04140007e+00 7.40427896e-02 7.96717405e-01 2.53872365e-01
6.01157129e-01 2.12497070e-01 8.07534009e-02 7.97807395e-01
3.33986729e-01 1.69249341e-01 9.12386596e-01 -4.94778037e-01
1.38463616e-01 9.29066181e-01 -2.81210423e-01 -8.78395796e-01
-5.01273692e-01 -4.20139164e-01 -1.13735878e+00 2.19918162e-01
1.91773534e-01 -1.64990991e-01 -7.89418459e-01 1.15503335e+00
4.93031830e-01 4.92052615e-01 2.39021882e-01 3.91541541e-01
3.01837862e-01 5.92143774e-01 -3.62082571e-01 -2.98858136e-01
6.57468259e-01 -3.27614278e-01 -5.93992591e-01 1.87871963e-01
6.38834476e-01 -3.05560619e-01 6.31791711e-01 9.76944268e-01
-5.54712117e-01 -4.83393848e-01 -9.67726767e-01 8.15544605e-01
-3.90595853e-01 2.47502640e-01 3.62612247e-01 6.34281754e-01
-7.91414499e-01 7.66043425e-01 -1.16555619e+00 -4.72211003e-01
2.72332698e-01 3.18816483e-01 -4.33190048e-01 1.66514575e-01
-9.09020305e-01 2.75917828e-01 1.01694512e+00 -1.20023422e-01
-9.36121404e-01 -3.94611925e-01 -6.62838578e-01 1.29967064e-01
2.60662735e-01 -1.45395935e-01 8.86520088e-01 -6.44501805e-01
-1.25387454e+00 7.39542902e-01 7.49285296e-02 -1.13796711e+00
3.32806557e-01 -3.02695423e-01 -8.05557668e-01 3.10872197e-01
-4.84703332e-02 -8.08088854e-02 1.34820330e+00 -1.37340987e+00
-7.05624104e-01 -3.60500276e-01 -7.45452464e-01 -3.14026028e-01
-8.19841266e-01 -1.36577249e-01 -3.84710431e-01 -6.81588769e-01
3.49708408e-01 -7.18116879e-01 -2.35871911e-01 -7.07107604e-01
-7.27913499e-01 -4.12370414e-01 1.23852968e+00 -4.96109605e-01
1.55080831e+00 -2.31715703e+00 -1.55445486e-01 1.13944149e+00
5.27059913e-01 1.50871649e-01 3.59620601e-01 5.17987251e-01
-5.75843930e-01 -2.81550914e-01 -5.15792727e-01 -4.36456084e-01
-1.11816727e-01 2.31978253e-01 -6.26981914e-01 7.36941159e-01
1.81578547e-01 2.85458684e-01 -6.44602001e-01 -2.37443358e-01
2.69367129e-01 2.40730539e-01 -5.50331831e-01 8.09967995e-01
3.63222673e-03 2.02550620e-01 -3.33049268e-01 5.71239293e-01
4.22425598e-01 -2.12527931e-01 3.55037674e-02 1.50114238e-01
2.64654696e-01 6.24040551e-02 -1.38017237e+00 1.12025678e+00
2.16688607e-02 3.81338656e-01 -3.93902898e-01 -1.67652082e+00
1.22322273e+00 3.25798988e-01 8.37511480e-01 -1.79569215e-01
2.02905208e-01 1.78414341e-02 6.17210707e-03 -4.97383058e-01
-1.19591504e-02 1.05314754e-01 -7.75568411e-02 6.63470447e-01
2.50775695e-01 4.17322814e-01 3.69999915e-01 2.49549001e-01
1.48617327e+00 -2.87587851e-01 4.40266520e-01 -1.66373596e-01
6.76961958e-01 -2.13547826e-01 6.42366588e-01 8.14115524e-01
6.15261458e-02 6.77751482e-01 7.93796837e-01 -6.46081746e-01
-9.06765640e-01 -1.14494455e+00 -1.75767854e-01 8.50827456e-01
-4.34965044e-01 -6.66181743e-01 -7.29410648e-01 -1.14542723e+00
3.81211899e-02 7.88005531e-01 -1.07005477e+00 -4.99583751e-01
-3.81799668e-01 -1.06331325e+00 5.44016004e-01 6.11845374e-01
5.70152774e-02 -9.75451171e-01 -5.51163435e-01 6.24726601e-02
1.64703548e-01 -9.68901694e-01 3.37150007e-01 6.17580056e-01
-1.03684592e+00 -1.29767144e+00 -5.77976555e-02 -4.46630448e-01
9.56637979e-01 -1.24496914e-01 1.04734778e+00 1.86650842e-01
-2.31467292e-01 2.12039053e-01 -6.71605349e-01 -2.11267576e-01
-5.86652994e-01 -8.44051912e-02 5.34997940e-01 6.49242759e-01
7.84296155e-01 -7.27107108e-01 -3.67198139e-01 7.89524242e-02
-1.10242450e+00 -9.25324321e-01 7.66952455e-01 9.32758093e-01
7.89320052e-01 9.55465496e-01 5.49242914e-01 -1.20655429e+00
6.90207899e-01 -9.65804458e-01 -5.92988610e-01 -8.27249065e-02
-9.52259421e-01 2.72160679e-01 1.15657985e+00 -8.51950422e-02
-7.84110487e-01 2.11503431e-01 -2.19539717e-01 -7.77909517e-01
-8.65471184e-01 4.45158541e-01 -2.07517728e-01 5.09207666e-01
7.54484355e-01 5.05917370e-01 -2.29046550e-02 -6.88566566e-01
7.14889681e-03 7.16023564e-01 6.82034910e-01 -3.77799809e-01
1.02566147e+00 7.29141653e-01 -1.96112797e-01 -7.56412566e-01
-6.41501784e-01 -9.62190747e-01 -9.25710261e-01 5.04865870e-02
6.01626873e-01 -6.45054400e-01 -4.76948559e-01 3.34886521e-01
-5.84221780e-01 -1.05783433e-01 -4.73371595e-01 3.66268426e-01
-3.89153153e-01 5.31060398e-01 -1.98787406e-01 -9.01312351e-01
-3.91838700e-01 -7.09533453e-01 8.90267551e-01 -2.10082382e-01
-3.17591041e-01 -1.09744501e+00 5.12034297e-01 -4.36380021e-02
7.15454072e-02 4.14462119e-01 7.91360855e-01 -1.77815855e+00
9.26485285e-02 -9.74047422e-01 2.01636866e-01 6.42987549e-01
1.15634717e-01 1.50085583e-01 -1.02459121e+00 -4.33674634e-01
1.97605863e-01 -2.35898212e-01 1.17160523e+00 1.20741844e-01
1.65755415e+00 -2.58574247e-01 -2.82261431e-01 5.18716455e-01
1.37761796e+00 3.26534733e-02 5.59297442e-01 2.87238538e-01
6.08991504e-01 4.27067876e-01 5.18528998e-01 9.32600737e-01
7.25158453e-02 1.81616113e-01 4.58384812e-01 1.41307279e-01
5.49327195e-01 2.24157199e-02 5.85856199e-01 7.55049467e-01
1.79889426e-01 -1.20638199e-01 -1.21009302e+00 5.49111962e-01
-1.83767617e+00 -1.07311773e+00 2.51820218e-02 2.40634894e+00
4.32437539e-01 5.92949986e-01 5.45703769e-01 9.16147649e-01
6.83766723e-01 -1.26251787e-01 -2.90967613e-01 -1.70070887e-01
6.38376847e-02 1.31070688e-01 1.48477390e-01 -4.66328673e-02
-1.50602686e+00 5.69858611e-01 6.23287249e+00 6.06883168e-01
-1.02337301e+00 -2.47796118e-01 4.58379209e-01 1.32991076e-01
3.19936424e-01 -1.99437872e-01 -6.73983872e-01 5.29949248e-01
1.39610469e+00 8.45568478e-02 1.74568906e-01 1.18666947e+00
1.59368485e-01 1.03132471e-01 -1.10018408e+00 7.92026162e-01
3.39458615e-01 -9.64672029e-01 -8.96206591e-03 9.97541398e-02
6.74156904e-01 1.90400258e-01 2.44677756e-02 3.84910226e-01
5.02910614e-01 -8.59599411e-01 1.58006459e-01 5.79794824e-01
3.16522866e-01 -1.09214890e+00 1.01985192e+00 3.99270713e-01
-8.60000491e-01 -4.06336129e-01 -2.97300965e-01 1.31446302e-01
-3.00480723e-01 1.16855335e+00 -1.11597705e+00 4.04513180e-01
9.29682791e-01 8.69711518e-01 -7.60178149e-01 1.09133446e+00
-4.32092920e-02 1.08613765e+00 -4.17961627e-01 4.80616510e-01
3.87935750e-02 -1.65632680e-01 4.95789289e-01 1.27546513e+00
2.85151541e-01 -3.03302914e-01 4.19420421e-01 5.08287847e-01
-2.93015633e-02 2.98647672e-01 -8.09833467e-01 -2.41748884e-01
4.08898413e-01 1.28600752e+00 -9.47294712e-01 -4.77844834e-01
-3.84406269e-01 8.89770687e-01 2.88511366e-01 1.92025706e-01
-3.72208774e-01 -6.37812495e-01 3.56599361e-01 1.75812095e-01
6.25697672e-01 6.55725226e-02 -2.46111780e-01 -1.32974923e+00
3.49855646e-02 -1.00345707e+00 9.93492544e-01 2.29398124e-02
-1.57410467e+00 7.21857190e-01 -1.45434827e-01 -1.63896203e+00
-7.63840735e-01 -5.48646033e-01 -1.14955175e+00 1.28193542e-01
-1.04241395e+00 -7.07703829e-01 -3.11132610e-01 8.31356227e-01
3.70468795e-01 -7.64266670e-01 1.07461846e+00 1.61255114e-02
-9.68038023e-01 5.70324421e-01 6.25754416e-01 5.79841018e-01
6.65848196e-01 -1.68723607e+00 3.48658055e-01 1.23925865e+00
4.48113799e-01 8.19570303e-01 7.69260943e-01 -6.72486484e-01
-1.12412155e+00 -1.34996581e+00 3.30610245e-01 -5.37655413e-01
7.35089719e-01 -2.97412485e-01 -1.20779312e+00 6.08337224e-01
-1.52351186e-02 2.46443078e-01 1.16288424e+00 2.87959367e-01
-4.35295522e-01 -2.95128167e-01 -1.11534691e+00 6.08230717e-02
2.78308511e-01 -3.11684519e-01 -7.95278847e-01 2.75754333e-01
1.85520932e-01 2.32434366e-02 -9.79439914e-01 3.76724839e-01
9.75112393e-02 -1.19875479e+00 7.34215379e-01 -6.64623857e-01
3.14774424e-01 -2.97984123e-01 -1.23481177e-01 -1.44614840e+00
-2.20173746e-01 -4.75937814e-01 -8.79104912e-01 1.08898687e+00
1.61695004e-01 -4.56487477e-01 8.95356953e-01 9.93889123e-02
2.94265710e-02 -7.31697619e-01 -6.26354039e-01 -7.39451051e-01
-1.94730967e-01 -7.14882493e-01 5.32877624e-01 1.01061201e+00
2.71503218e-02 2.31939815e-02 -3.79957527e-01 4.22047436e-01
9.96646523e-01 1.30040839e-01 9.07833040e-01 -1.79618728e+00
-4.91788119e-01 -4.14414167e-01 -7.95813203e-01 -4.21110243e-01
4.06432092e-01 -7.18337178e-01 -1.67886764e-02 -9.50059831e-01
-1.03041448e-01 -3.45355682e-02 -7.93854415e-01 3.66206199e-01
-1.23867430e-01 2.69897014e-01 -5.23356855e-01 2.91519076e-01
-9.38580334e-01 5.72289586e-01 -1.36747792e-01 1.33263797e-01
-4.14462149e-01 3.29802871e-01 -6.65589809e-01 9.46704924e-01
9.86990154e-01 -5.83242178e-01 -2.94331610e-01 2.19502330e-01
7.15828836e-02 -5.51139235e-01 1.45918936e-01 -1.32890224e+00
2.70611674e-01 5.74819073e-02 7.76676357e-01 -8.61234963e-01
-1.86821342e-01 -9.96208906e-01 -5.68047881e-01 4.01995808e-01
-3.42792422e-01 -5.17769270e-02 -2.71728605e-01 8.96301806e-01
-5.32611132e-01 -4.74955924e-02 7.98415303e-01 1.46783531e-01
-6.40741885e-01 3.24333340e-01 -4.99110520e-01 8.25143158e-02
1.06584227e+00 -9.72139183e-03 2.88034767e-01 -4.39529687e-01
-8.15662205e-01 2.12531030e-01 3.31127584e-01 4.34063822e-01
8.83505166e-01 -1.28826964e+00 -6.89140141e-01 7.75611043e-01
4.78173822e-01 7.57444045e-03 -2.54042000e-02 8.04427922e-01
-3.92932802e-01 3.49862218e-01 -5.10614775e-02 -7.68405855e-01
-9.48921919e-01 6.72831416e-01 1.58934370e-01 -2.88099170e-01
-8.61454785e-01 4.19918478e-01 -2.11264580e-01 -3.70909870e-01
4.67267543e-01 5.46284541e-02 -5.36746979e-01 1.49138644e-01
8.80222976e-01 4.26583827e-01 5.96021533e-01 -5.86536169e-01
-2.65091419e-01 1.56064257e-01 -3.63611430e-01 1.79584473e-01
1.70686805e+00 -3.38918827e-02 -1.81282550e-01 6.75613701e-01
1.24471354e+00 -2.01054037e-01 -8.37744951e-01 -3.87937874e-01
5.06806016e-01 -3.23592514e-01 1.93576679e-01 -2.33683929e-01
-1.08661115e+00 7.45222867e-01 6.65783286e-01 6.65992022e-01
1.36113226e+00 2.10559312e-02 6.28823876e-01 8.85288417e-01
-2.22026974e-01 -1.30341327e+00 2.06825033e-01 3.48544478e-01
4.60574597e-01 -1.37756872e+00 -7.60427713e-02 2.38390252e-01
-7.48709917e-01 1.42351723e+00 6.87014222e-01 -5.54648161e-01
1.00219548e+00 1.80913612e-01 9.67689753e-02 -4.50537324e-01
-8.48572433e-01 -2.03050509e-01 4.24557388e-01 5.22909939e-01
1.97243392e-01 -1.30607441e-01 1.58730417e-01 6.71108782e-01
-1.90809637e-01 -5.40660143e-01 4.93417114e-01 7.57835567e-01
-6.46603703e-01 -1.05790186e+00 -6.11466825e-01 8.83170366e-01
-7.51798630e-01 2.78673053e-01 -4.65724230e-01 3.56844038e-01
4.90043014e-02 8.97718191e-01 1.69254407e-01 -6.19216859e-01
1.37552127e-01 5.73477507e-01 -4.02911156e-01 -5.94332874e-01
-3.64488721e-01 3.03919822e-01 -4.06382740e-01 -6.71330154e-01
-4.08381447e-02 -9.68664765e-01 -1.35812092e+00 2.00152040e-01
-1.92099139e-01 4.83619899e-01 3.96055281e-01 1.02797592e+00
3.75210166e-01 3.71708333e-01 1.24842107e+00 -4.91759241e-01
-6.93890095e-01 -8.94412637e-01 -7.41752505e-01 8.80800784e-01
5.26079059e-01 -6.11711025e-01 -6.43755496e-01 1.35062054e-01] | [7.590263366699219, 2.428600549697876] |
72b08e62-a539-4a42-8afe-36046fd08ace | generalized-source-free-domain-adaptation | 2108.01614 | null | https://arxiv.org/abs/2108.01614v2 | https://arxiv.org/pdf/2108.01614v2.pdf | Generalized Source-free Domain Adaptation | Domain adaptation (DA) aims to transfer the knowledge learned from a source domain to an unlabeled target domain. Some recent works tackle source-free domain adaptation (SFDA) where only a source pre-trained model is available for adaptation to the target domain. However, those methods do not consider keeping source performance which is of high practical value in real world applications. In this paper, we propose a new domain adaptation paradigm called Generalized Source-free Domain Adaptation (G-SFDA), where the learned model needs to perform well on both the target and source domains, with only access to current unlabeled target data during adaptation. First, we propose local structure clustering (LSC), aiming to cluster the target features with its semantically similar neighbors, which successfully adapts the model to the target domain in the absence of source data. Second, we propose sparse domain attention (SDA), it produces a binary domain specific attention to activate different feature channels for different domains, meanwhile the domain attention will be utilized to regularize the gradient during adaptation to keep source information. In the experiments, for target performance our method is on par with or better than existing DA and SFDA methods, specifically it achieves state-of-the-art performance (85.4%) on VisDA, and our method works well for all domains after adapting to single or multiple target domains. Code is available in https://github.com/Albert0147/G-SFDA. | ['Shangling Jui', 'Luis Herranz', 'Joost Van de Weijer', 'Yaxing Wang', 'Shiqi Yang'] | 2021-08-03 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Yang_Generalized_Source-Free_Domain_Adaptation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Yang_Generalized_Source-Free_Domain_Adaptation_ICCV_2021_paper.pdf | iccv-2021-1 | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 7.52646700e-02 -2.18749255e-01 -4.65155900e-01 -4.02665406e-01
-7.42079973e-01 -6.40850723e-01 6.00667477e-01 -8.59805495e-02
-2.61522144e-01 8.23118985e-01 1.99580610e-01 1.19976841e-01
1.38009742e-01 -6.43760741e-01 -7.11567402e-01 -7.44783878e-01
3.13936949e-01 7.39569545e-01 4.23710465e-01 -1.83094412e-01
-5.43180481e-03 4.32471365e-01 -1.21336973e+00 1.77952677e-01
1.05466735e+00 9.17333841e-01 5.92074275e-01 1.82281792e-01
-3.87050182e-01 3.99442673e-01 -5.67602873e-01 -6.89033195e-02
1.06887676e-01 -6.78357601e-01 -8.53518009e-01 2.21471190e-01
2.53682047e-01 -1.39723688e-01 -1.19442061e-01 9.92960155e-01
5.44918299e-01 4.05049264e-01 1.01997471e+00 -9.94343877e-01
-1.14031672e+00 2.14034632e-01 -5.41718781e-01 3.74141783e-01
1.04401194e-01 -3.24064121e-02 6.24132633e-01 -1.07009089e+00
7.91285098e-01 1.10998273e+00 1.79800168e-01 9.47636127e-01
-1.10514045e+00 -7.53884137e-01 5.78387856e-01 3.06845367e-01
-1.28912473e+00 -6.08819902e-01 9.17828143e-01 -2.99833745e-01
8.02204669e-01 -4.07166451e-01 1.59811556e-01 1.31309032e+00
-2.41239786e-01 8.51179838e-01 7.96786427e-01 -4.88791227e-01
5.25180101e-01 3.90679181e-01 5.25337122e-02 9.85512733e-02
6.78444579e-02 1.70620922e-02 -3.43256027e-01 -7.20022991e-02
8.91967356e-01 3.54793742e-02 -2.65538454e-01 -6.32128298e-01
-1.21256423e+00 8.27346385e-01 5.97584486e-01 5.31160176e-01
-5.24938703e-01 -5.85384250e-01 3.32466930e-01 4.17992830e-01
5.52551568e-01 2.48525575e-01 -6.85263634e-01 2.29767695e-01
-4.43633825e-01 -9.77247301e-03 4.94607002e-01 1.17325819e+00
9.40175533e-01 1.28350809e-01 -1.42642871e-01 1.19842398e+00
2.28912458e-01 6.67845190e-01 9.12866354e-01 -6.09858513e-01
5.18246114e-01 7.06367970e-01 1.81459039e-01 -4.91829336e-01
-1.38195083e-01 -5.16420066e-01 -7.72985756e-01 4.22476120e-02
3.50746512e-01 -1.78187013e-01 -1.18060017e+00 1.93862367e+00
5.23773193e-01 4.05520022e-01 4.86436009e-01 9.35986638e-01
8.27640712e-01 7.34927595e-01 3.44625950e-01 -1.56493172e-01
1.00203228e+00 -1.11138296e+00 -4.81480330e-01 -6.92021787e-01
5.15625536e-01 -6.03265941e-01 1.29274774e+00 1.96631536e-01
-7.83586442e-01 -8.38589489e-01 -9.02209997e-01 5.44010662e-02
-5.01482248e-01 3.01245391e-01 9.14758295e-02 2.29153708e-01
-7.83479750e-01 1.12332366e-01 -6.00313365e-01 -7.97507823e-01
5.42881668e-01 4.48609054e-01 -5.21311581e-01 -5.36843896e-01
-1.25850320e+00 7.82257915e-01 5.34982443e-01 -5.01352608e-01
-1.15926063e+00 -6.48429155e-01 -8.44120800e-01 2.27284655e-02
3.08511555e-01 -6.27864003e-01 1.16964483e+00 -1.55560374e+00
-1.59473836e+00 9.13235784e-01 -5.13671339e-01 -2.96493262e-01
-7.38696102e-03 -2.19566554e-01 -7.59171784e-01 1.52915329e-01
3.51239622e-01 7.76505768e-01 9.80836391e-01 -1.36270964e+00
-6.81010187e-01 -3.99711102e-01 -1.74727350e-01 4.79836226e-01
-6.17813468e-01 -1.90671742e-01 -6.30716622e-01 -6.92555726e-01
-1.11504577e-01 -7.41276264e-01 -2.58346815e-02 -2.32810155e-01
1.38052423e-02 -3.55864853e-01 9.87719059e-01 -5.79485059e-01
1.18417835e+00 -2.44430470e+00 3.01101953e-01 8.85369256e-02
-1.73774794e-01 6.58464491e-01 -4.84841675e-01 2.44747236e-01
-1.92453012e-01 -2.63027370e-01 -5.80570519e-01 -2.53875732e-01
-2.46217817e-01 2.61553675e-01 -2.83707291e-01 2.96905071e-01
3.27208877e-01 7.34198332e-01 -8.44990909e-01 -2.65708923e-01
1.34673655e-01 3.79247874e-01 -5.10850430e-01 4.06993061e-01
-2.87291765e-01 8.06604862e-01 -6.75905406e-01 5.76013684e-01
8.64141047e-01 -4.21577901e-01 2.09676877e-01 1.14835627e-01
2.50863761e-01 1.13057874e-01 -1.07315612e+00 2.00957799e+00
-5.18923342e-01 8.84085670e-02 1.09259546e-01 -1.18321455e+00
1.19199121e+00 2.63759702e-01 3.09557170e-01 -8.82802486e-01
-5.67916781e-02 3.11405510e-01 -6.77987710e-02 -1.68628842e-01
-7.57979304e-02 -1.93801478e-01 -8.75285342e-02 2.25551605e-01
4.02356207e-01 3.09595883e-01 -6.25645295e-02 1.30300641e-01
8.98317695e-01 2.27808222e-01 5.53042531e-01 -2.48946533e-01
8.22261393e-01 2.81164438e-01 6.59659803e-01 3.60341877e-01
-3.03993553e-01 6.20958865e-01 2.48103458e-02 -1.69247001e-01
-1.03807259e+00 -1.20541775e+00 3.76738869e-02 1.42966020e+00
3.67228985e-01 2.67215252e-01 -7.20789969e-01 -1.09206259e+00
-7.58414045e-02 8.26550484e-01 -5.51276982e-01 -4.43667531e-01
-5.51254392e-01 -3.69039893e-01 1.83208004e-01 4.92659718e-01
7.28348672e-01 -1.18236852e+00 1.35381464e-02 3.07312429e-01
-1.67253554e-01 -9.60955501e-01 -7.29484439e-01 2.11584225e-01
-9.81899381e-01 -9.00096178e-01 -1.21528852e+00 -1.18221200e+00
8.50182116e-01 3.21296155e-01 1.04708087e+00 -4.02944267e-01
4.39279586e-01 3.79428208e-01 -6.16784334e-01 -1.62469253e-01
-3.07746589e-01 4.55482334e-01 1.50638744e-01 2.27300242e-01
7.76002884e-01 -6.17249668e-01 -4.45648998e-01 5.60526729e-01
-9.77419436e-01 -3.79224598e-01 8.26642931e-01 9.30935442e-01
7.82402456e-01 -1.85258135e-01 1.05888700e+00 -1.06856847e+00
5.39307833e-01 -1.06477404e+00 -3.26188862e-01 1.77235097e-01
-5.24766803e-01 -3.76483463e-02 1.03238070e+00 -7.87274361e-01
-1.31711817e+00 2.92613566e-01 -2.65633538e-02 -7.94431567e-01
-6.77156925e-01 3.12775493e-01 -6.35744154e-01 1.92798987e-01
1.03420770e+00 4.96836573e-01 -1.32279366e-01 -7.94548213e-01
2.12985128e-01 8.01481068e-01 5.17565966e-01 -6.14473641e-01
7.50592947e-01 4.19850081e-01 -5.52759647e-01 -5.26358187e-01
-8.10162187e-01 -7.70763397e-01 -8.53362381e-01 2.95311689e-01
6.37924790e-01 -1.12424374e+00 1.77656680e-01 4.45310056e-01
-8.57189894e-01 -5.70811868e-01 -2.96749204e-01 4.81023908e-01
-3.88319492e-01 2.94755995e-01 -7.10523576e-02 -3.77104253e-01
-2.12755799e-01 -8.39225531e-01 9.01340544e-01 4.00850654e-01
3.93249746e-03 -1.30910313e+00 1.92298606e-01 -4.99805296e-03
4.01993990e-01 1.42525360e-02 8.24576676e-01 -1.21290183e+00
-1.43386081e-01 1.73985526e-01 -1.91331476e-01 5.84658861e-01
5.26205778e-01 -6.73205078e-01 -9.16611969e-01 -4.93605912e-01
-2.71550342e-02 -2.40882561e-01 8.00559461e-01 4.82462674e-01
9.21577275e-01 -2.22769007e-01 -6.09893262e-01 4.78148282e-01
1.41535699e+00 4.48519409e-01 5.87323844e-01 3.93531471e-01
6.03819191e-01 3.77516985e-01 9.34693933e-01 3.67293745e-01
3.62834841e-01 6.99589968e-01 1.85205907e-01 -1.55565217e-01
-4.60220903e-01 -3.31402063e-01 6.15930676e-01 5.81562996e-01
3.32181394e-01 -3.64479870e-01 -9.42441285e-01 1.10776472e+00
-1.79498410e+00 -5.95892787e-01 2.08379194e-01 2.36780262e+00
8.16830754e-01 -3.35753076e-02 3.24075222e-01 -3.72895628e-01
9.77653146e-01 -1.82622328e-01 -1.15258968e+00 -1.86958000e-01
-1.81535557e-01 3.04170251e-01 3.67888361e-01 4.05206501e-01
-1.20113873e+00 1.24710572e+00 5.59206581e+00 1.08261883e+00
-1.21447158e+00 4.00818139e-01 3.69361311e-01 -1.33640002e-02
-2.07337186e-01 -2.05989406e-01 -8.78846109e-01 5.82754672e-01
9.41387832e-01 -1.50738999e-01 3.95286113e-01 1.15374672e+00
-3.47783454e-02 1.88873887e-01 -9.68339741e-01 7.63287306e-01
6.72754087e-03 -9.76709843e-01 2.24807337e-01 -6.21542446e-02
8.59060585e-01 1.08850263e-01 1.12612255e-01 6.16831481e-01
4.25088614e-01 -6.62401676e-01 2.49160901e-01 1.49971157e-01
1.03670514e+00 -7.73201942e-01 6.79023802e-01 4.69695240e-01
-1.14507508e+00 -2.22798139e-01 -6.13469601e-01 1.87727958e-01
-9.86240506e-02 4.23277825e-01 -1.01493728e+00 5.95409572e-01
6.53842986e-01 1.04519796e+00 -4.31154072e-01 8.98517787e-01
-2.94334739e-01 8.33148837e-01 -2.73946047e-01 3.52905214e-01
2.02083021e-01 2.34345067e-03 5.56808174e-01 1.04287875e+00
5.12316942e-01 1.22922294e-01 2.54194945e-01 6.83012784e-01
-1.72418401e-01 3.17043632e-01 -6.94283605e-01 1.49831712e-01
7.37889290e-01 8.23572338e-01 -3.31456780e-01 -5.81258655e-01
-4.86784160e-01 1.35190666e+00 4.74665403e-01 7.20121622e-01
-6.93760991e-01 -3.40118647e-01 8.20704103e-01 1.61344737e-01
5.85503697e-01 -4.50323569e-03 -2.09747270e-01 -1.32272732e+00
-2.14365795e-02 -8.06008875e-01 8.59416187e-01 -6.06968939e-01
-1.73844898e+00 5.85814893e-01 -6.05134517e-02 -1.49249232e+00
-1.84786946e-01 -3.65335405e-01 -5.35975635e-01 1.02144825e+00
-1.90680563e+00 -1.22119844e+00 -2.16699332e-01 1.26488698e+00
8.88341725e-01 -5.46295822e-01 8.45094740e-01 5.09637177e-01
-4.09931302e-01 8.11609209e-01 5.38019478e-01 5.92457578e-02
1.32260346e+00 -1.01207304e+00 3.21421385e-01 7.98043787e-01
-1.03344232e-01 5.71460009e-01 2.68194050e-01 -7.19951510e-01
-9.31908131e-01 -1.41802609e+00 7.02394187e-01 -3.37587595e-01
2.93580472e-01 -1.58775955e-01 -1.31944740e+00 8.03236365e-01
1.13596179e-01 1.10063948e-01 6.93416178e-01 -1.83151141e-02
-4.81081367e-01 -1.78499877e-01 -1.44211495e+00 2.17046231e-01
1.07554662e+00 -2.09723562e-01 -7.15324938e-01 2.32848510e-01
7.97663033e-01 -2.78314203e-01 -7.76261091e-01 2.78270781e-01
-3.39545570e-02 -6.26673043e-01 1.01050687e+00 -5.90940535e-01
2.39576176e-01 -6.14401877e-01 -1.65935799e-01 -1.60409415e+00
-7.56653666e-01 -4.39347848e-02 -1.80552617e-01 1.37025070e+00
3.33937973e-01 -7.86305368e-01 5.48993826e-01 2.11506277e-01
-3.35653573e-01 -2.99318850e-01 -9.17997003e-01 -1.04680824e+00
4.66044903e-01 5.56817800e-02 7.07202494e-01 1.32370973e+00
-2.79741973e-01 5.55193305e-01 -2.28561267e-01 4.44662213e-01
4.12738562e-01 1.81498259e-01 6.47516429e-01 -1.21649063e+00
-1.93795875e-01 -2.81064391e-01 -6.88249618e-02 -1.37507629e+00
2.99578309e-01 -9.61311877e-01 -1.82082325e-01 -1.56260502e+00
6.13630265e-02 -4.99252737e-01 -6.78930938e-01 7.00985134e-01
-2.75390416e-01 4.09717001e-02 4.03944328e-02 6.29320443e-01
-6.59522831e-01 7.03767478e-01 1.28167474e+00 -2.04733297e-01
-5.65400541e-01 8.38426948e-02 -1.01806784e+00 3.61185372e-01
9.10031438e-01 -5.02270818e-01 -6.33226752e-01 -5.42442858e-01
-5.64178288e-01 -3.14755350e-01 1.28121629e-01 -1.07554495e+00
1.39112458e-01 -3.33680898e-01 5.60908556e-01 -1.93614244e-01
1.71603873e-01 -1.02976429e+00 -6.24987185e-02 1.82276443e-01
-1.94876462e-01 -4.16081548e-01 3.52726370e-01 6.64233506e-01
-4.84548777e-01 -2.08668068e-01 1.22540104e+00 -8.77625793e-02
-1.40548670e+00 3.51358950e-01 -6.39250055e-02 1.60049364e-01
1.23207939e+00 -1.28596097e-01 -1.63549125e-01 -2.36906558e-01
-9.66947854e-01 3.03235471e-01 6.88989222e-01 4.56656545e-01
5.68506420e-01 -1.61174083e+00 -7.81600833e-01 2.78739154e-01
5.04229069e-01 2.23226309e-01 4.38108206e-01 3.70520055e-01
4.75321031e-05 3.54281485e-01 -3.75514865e-01 -6.35425270e-01
-8.88208866e-01 9.46591973e-01 2.36407697e-01 -2.14967504e-01
-3.94736379e-01 9.01514888e-01 7.15263844e-01 -7.08893478e-01
6.36143163e-02 -2.67256405e-02 -4.17283624e-01 -2.39217337e-02
5.44649541e-01 1.08098909e-01 4.65610400e-02 -6.41309261e-01
-5.11493385e-01 7.01415896e-01 -2.46129110e-01 1.24462925e-01
1.20046246e+00 -4.16199952e-01 3.25947613e-01 2.56300509e-01
1.18957508e+00 -1.23263068e-01 -1.49102962e+00 -8.01631212e-01
-1.35174885e-01 -3.33762348e-01 -1.76715955e-01 -1.15144277e+00
-1.01881349e+00 8.94974470e-01 7.73579121e-01 -3.28926712e-01
1.58155119e+00 3.00485522e-01 8.03657770e-01 1.93284780e-01
2.49835745e-01 -1.06905878e+00 1.54386938e-01 6.45938039e-01
8.29781413e-01 -1.39136827e+00 -2.94735759e-01 -1.67178452e-01
-1.20752609e+00 7.98670232e-01 9.81889427e-01 -2.52945840e-01
5.98342359e-01 -3.11619580e-01 1.29193112e-01 1.16813272e-01
-5.77420950e-01 -4.43638027e-01 3.01462710e-01 1.11134386e+00
1.63844392e-01 -7.25441659e-03 9.37141553e-02 8.50756109e-01
3.71044874e-01 -1.59622039e-04 6.06477521e-02 6.97814584e-01
-6.19351804e-01 -1.43053949e+00 -3.90465230e-01 1.78094357e-01
-1.27970293e-01 -1.69374440e-02 -5.06565511e-01 6.72554612e-01
2.26355150e-01 8.95802319e-01 -7.52895325e-03 -8.55884776e-02
5.25421679e-01 3.56662184e-01 8.98823813e-02 -1.03670752e+00
-3.12109351e-01 2.05453709e-01 -2.23751843e-01 -3.07413310e-01
-4.52714503e-01 -6.64385796e-01 -1.36390889e+00 7.60127455e-02
2.29935665e-02 1.88962698e-01 3.37169915e-01 7.73720741e-01
8.59245777e-01 3.33660036e-01 6.52880192e-01 -5.41771293e-01
-2.44487986e-01 -1.07406521e+00 -6.07186973e-01 5.96911430e-01
3.27866822e-01 -9.06787336e-01 -1.81194082e-01 3.42623144e-01] | [10.358477592468262, 3.0133755207061768] |
bf4eaa65-cb04-4e47-8a64-5e1bbc946e4e | deep-learning-and-traffic-classification | 2104.03182 | null | https://arxiv.org/abs/2104.03182v2 | https://arxiv.org/pdf/2104.03182v2.pdf | Deep Learning and Traffic Classification: Lessons learned from a commercial-grade dataset with hundreds of encrypted and zero-day applications | The increasing success of Machine Learning (ML) and Deep Learning (DL) has recently re-sparked interest towards traffic classification. While classification of known traffic is a well investigated subject with supervised classification tools (such as ML and DL models) are known to provide satisfactory performance, detection of unknown (or zero-day) traffic is more challenging and typically handled by unsupervised techniques (such as clustering algorithms). In this paper, we share our experience on a commercial-grade DL traffic classification engine that is able to (i) identify known applications from encrypted traffic, as well as (ii) handle unknown zero-day applications. In particular, our contribution for (i) is to perform a thorough assessment of state of the art traffic classifiers in commercial-grade settings comprising few thousands of very fine grained application labels, as opposite to the few tens of classes generally targeted in academic evaluations. Additionally, we contribute to the problem of (ii) detection of zero-day applications by proposing a novel technique, tailored for DL models, that is significantly more accurate and light-weight than the state of the art. Summarizing our main findings, we gather that (i) while ML and DL models are both equally able to provide satisfactory solution for classification of known traffic, however (ii) the non-linear feature extraction process of the DL backbone provides sizeable advantages for the detection of unknown classes. | ['Dario Rossi', 'Feng Jun', 'Alessandro Finamore', 'Lixuan Yang'] | 2021-04-07 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [ 4.32214364e-02 -5.77495635e-01 -3.95211309e-01 -2.09718138e-01
-7.40822852e-01 -5.78794420e-01 7.24332154e-01 1.81747258e-01
-1.35226622e-01 8.43860507e-01 -4.67018366e-01 -9.94231164e-01
-3.73444796e-01 -1.00390100e+00 -4.43306983e-01 -7.32121944e-01
-1.26223490e-01 7.38422215e-01 5.51211357e-01 -1.14237629e-01
1.85438797e-01 1.03490603e+00 -2.07013702e+00 4.99691755e-01
6.34188414e-01 1.62747347e+00 -2.52455890e-01 6.81754410e-01
-6.02856934e-01 1.14580214e+00 -8.42530727e-01 -6.72357738e-01
2.31556669e-01 1.35852277e-01 -9.00583267e-01 2.56728768e-01
9.69307423e-01 -5.84847219e-02 -3.94153506e-01 5.32489717e-01
3.00822645e-01 -1.74466446e-01 7.89636672e-01 -1.80892944e+00
-3.61929893e-01 1.66273490e-01 -3.96120071e-01 5.78865290e-01
-9.84401256e-02 2.63600886e-01 9.93847907e-01 -5.25901794e-01
3.52913350e-01 1.04088235e+00 7.51727402e-01 2.53334194e-01
-1.28753066e+00 -9.00903821e-01 5.45855016e-02 5.36003411e-01
-1.36284792e+00 -5.47969639e-01 5.99635065e-01 -7.10124850e-01
8.60612750e-01 2.18794197e-01 1.95990913e-02 1.21546245e+00
-9.79446620e-02 7.81380177e-01 1.15873981e+00 -1.86195284e-01
2.02040061e-01 6.65934682e-01 5.51445961e-01 4.16019887e-01
2.81404138e-01 1.90064684e-01 -2.89569497e-02 -2.25594252e-01
1.83452498e-02 2.47165002e-02 2.39453748e-01 -5.02835572e-01
-8.45921338e-01 8.40112031e-01 -9.99654923e-03 3.43973577e-01
-4.29015994e-01 3.60988602e-02 9.40972447e-01 5.19605100e-01
7.21351624e-01 1.38437822e-01 -6.80014789e-01 -4.19207662e-01
-1.05847645e+00 2.28559315e-01 1.23327661e+00 9.59717035e-01
1.07527387e+00 3.99266571e-01 -2.40075499e-01 7.59923458e-01
-1.92438275e-01 5.50782025e-01 -3.81662436e-02 -6.38007283e-01
4.76181418e-01 5.93024850e-01 -2.13521808e-01 -1.10091150e+00
-1.65496573e-01 -6.36819899e-01 -7.46490598e-01 3.56052071e-01
7.94544697e-01 -4.60877791e-02 -4.43392336e-01 1.37294006e+00
-4.22771983e-02 4.29402918e-01 -1.96656525e-01 2.12797150e-01
6.17251396e-01 4.77267146e-01 2.63558745e-01 -8.39595124e-02
1.30029142e+00 -5.46447098e-01 -1.78786889e-01 2.18095124e-01
6.67437375e-01 -7.22654045e-01 8.34500670e-01 5.10180056e-01
-6.02956355e-01 -8.07102799e-01 -6.21098280e-01 4.90628481e-01
-9.45397258e-01 -2.25507431e-02 7.78569400e-01 1.20272577e+00
-1.02067876e+00 2.03768745e-01 -1.88399211e-01 -6.06627107e-01
8.85489285e-01 5.27057707e-01 -2.35318169e-01 9.40793939e-03
-9.52800333e-01 6.72436655e-01 -7.27437511e-02 -5.80932975e-01
-1.09094274e+00 -9.46024120e-01 -4.93792683e-01 -4.51301858e-02
7.81822145e-01 -2.68746257e-01 1.04039609e+00 -8.40905845e-01
-1.08290267e+00 1.19086301e+00 -1.13416798e-01 -6.93194866e-01
4.02669936e-01 1.37205720e-01 -8.63581836e-01 1.06443301e-01
1.45017862e-01 2.26794615e-01 1.03920782e+00 -1.36432266e+00
-1.16094005e+00 -1.02666795e-01 3.10070574e-01 -7.84289420e-01
-5.75631797e-01 2.03176841e-01 3.02023627e-02 -4.88368601e-01
-7.15421915e-01 -6.87832415e-01 1.98439106e-01 -1.10927306e-01
-2.81208754e-01 -6.65732086e-01 1.49639297e+00 -2.53784537e-01
1.47159636e+00 -1.95738745e+00 -7.56708264e-01 3.08522254e-01
5.76407790e-01 6.64432347e-01 -8.15741252e-03 5.09883285e-01
-1.31890357e-01 8.81073549e-02 -9.75833386e-02 -1.65130705e-01
3.27218980e-01 2.68387347e-01 -6.86744213e-01 3.87179941e-01
3.41290057e-01 8.40588212e-01 -6.86333001e-01 -4.08428729e-01
8.13600123e-01 2.92233706e-01 -3.28331739e-01 9.69561189e-02
-1.64271370e-01 9.02409852e-02 -2.17737079e-01 1.01010358e+00
7.95367897e-01 -3.78320903e-01 -2.13557214e-01 -4.43537295e-01
-2.08394647e-01 3.74176025e-01 -8.72566164e-01 5.26775837e-01
-8.62457573e-01 1.07998979e+00 6.40871748e-02 -1.56479609e+00
1.06508780e+00 1.67555287e-01 9.01734173e-01 -6.72085226e-01
3.92087325e-02 3.42126340e-01 4.08842042e-03 -4.01144534e-01
9.87503380e-02 6.44324794e-02 -4.13976386e-02 6.07629836e-01
1.02373578e-01 4.32665110e-01 5.85946798e-01 1.77603602e-01
1.34573853e+00 -2.98822343e-01 3.30912392e-03 -3.15097898e-01
1.09560096e+00 -1.04876697e-01 5.69465943e-02 9.29973781e-01
-5.25528550e-01 1.08270302e-01 6.87475502e-01 -6.30366147e-01
-1.01548326e+00 -1.03541040e+00 -2.06935629e-01 1.39834321e+00
-2.08081797e-01 -1.38474613e-01 -6.06638491e-01 -1.09866846e+00
3.35330278e-01 4.67687577e-01 -2.42310777e-01 1.69597536e-01
-7.78973162e-01 -8.72518241e-01 7.47906566e-01 3.80230188e-01
5.31344414e-01 -9.81842816e-01 -1.01345792e-01 2.75166005e-01
1.04231410e-01 -1.65944397e+00 6.62054941e-02 4.98885930e-01
-3.85070086e-01 -1.45074713e+00 -4.32885081e-01 -6.47098422e-01
2.04041153e-01 4.91187960e-01 1.64881217e+00 1.51897639e-01
-4.80200708e-01 5.54765999e-01 -3.78898561e-01 -4.64073032e-01
-6.83787465e-01 4.84980047e-01 6.86018616e-02 5.72828174e-01
1.10248935e+00 -8.19759786e-01 -2.00507149e-01 5.53515792e-01
-6.71735406e-01 -8.62266243e-01 6.87942743e-01 3.27061415e-01
7.82380700e-02 4.24726576e-01 7.87064493e-01 -1.11945438e+00
6.49345934e-01 -8.23573530e-01 -5.24607062e-01 1.28659187e-02
-7.99831390e-01 -3.86393785e-01 7.90587008e-01 -4.39409226e-01
-6.85176849e-01 -3.21075946e-01 -3.29860359e-01 -3.88701349e-01
-8.70062172e-01 -1.61245242e-01 -6.98991716e-02 -4.51767087e-01
7.15309799e-01 3.69282991e-01 -1.09726742e-01 -4.36083108e-01
1.38809770e-01 1.15098119e+00 3.28101873e-01 -8.73506010e-01
9.93923068e-01 5.60994029e-01 2.48209223e-01 -1.06507730e+00
-7.76280880e-01 -6.83427751e-01 -8.47166955e-01 -3.48474324e-01
5.63336313e-01 -5.82726717e-01 -1.03429568e+00 4.25464153e-01
-8.28291535e-01 -1.37534991e-01 -3.89430016e-01 1.09118976e-01
-5.15208066e-01 6.64839447e-01 -4.70444232e-01 -1.09095728e+00
1.79755203e-02 -1.15533638e+00 1.12100661e+00 -3.77128780e-01
-2.54999042e-01 -1.21358860e+00 -3.29763979e-01 5.31015515e-01
6.73625469e-01 2.42163867e-01 1.20797849e+00 -1.10105586e+00
-6.29938185e-01 -4.92849588e-01 -6.60627544e-01 7.28241801e-01
2.39040136e-01 2.71890551e-01 -1.29035544e+00 -8.75750110e-02
-3.37481111e-01 -2.03847557e-01 7.85092533e-01 1.80238396e-01
1.33312571e+00 -1.86375994e-02 -3.28619242e-01 2.43463933e-01
1.19543111e+00 2.57830709e-01 6.29944503e-01 4.61834937e-01
4.55531150e-01 7.00242400e-01 3.75582457e-01 3.54620397e-01
3.13241035e-01 8.33680034e-01 4.49687302e-01 -2.10524440e-01
-4.64136809e-01 2.03751609e-01 3.28698665e-01 4.12936985e-01
4.40427437e-02 -4.47060704e-01 -9.79821324e-01 5.13651669e-01
-1.60911226e+00 -1.28334439e+00 -5.08471489e-01 2.12917495e+00
2.94962466e-01 6.08224094e-01 8.66862118e-01 7.87062347e-01
7.92421460e-01 8.53218809e-02 -1.28396094e-01 -6.01629972e-01
-8.88425633e-02 6.28222048e-01 6.93756938e-01 1.40850931e-01
-1.33805013e+00 6.75691605e-01 6.66791964e+00 1.49237692e+00
-1.08431804e+00 1.85173765e-01 6.60312474e-01 3.58331501e-01
2.39417836e-01 -1.81599066e-01 -1.16115940e+00 9.44514871e-01
1.34829664e+00 9.82405692e-02 1.68546364e-01 1.01482236e+00
3.24616469e-02 3.86271477e-02 -1.01882017e+00 1.11439097e+00
-3.56388241e-02 -1.43992305e+00 3.66135746e-01 4.65420127e-01
4.59709913e-01 -7.60614797e-02 1.54356822e-01 8.65062892e-01
1.11861914e-01 -7.05141664e-01 2.81061321e-01 3.85813415e-01
7.40918338e-01 -9.18832421e-01 9.55816031e-01 4.37197208e-01
-1.27177954e+00 -4.91699010e-01 -1.86635092e-01 -1.53293028e-01
-1.74885280e-02 1.01982200e+00 -4.89978701e-01 4.94630754e-01
5.73065996e-01 7.58563280e-01 -7.17959046e-01 1.05102909e+00
4.13954049e-01 9.07850862e-01 -6.26739040e-02 1.01527676e-01
5.08352458e-01 -1.25703111e-01 3.78389686e-01 1.47910690e+00
1.45447731e-01 -4.80287164e-01 4.87975955e-01 6.51734173e-01
-4.16781846e-03 -1.56840205e-03 -8.00543725e-01 -9.58142430e-02
3.31783861e-01 1.40106118e+00 -7.13566482e-01 -5.13255239e-01
-7.57502794e-01 3.39966506e-01 8.53278264e-02 1.47901371e-01
-8.18102419e-01 -4.58212823e-01 1.04163277e+00 8.32385063e-01
3.33547235e-01 -2.18415499e-01 -2.47902334e-01 -7.02028334e-01
-3.16581465e-02 -8.57450068e-01 4.24335927e-01 -2.33143851e-01
-1.86303949e+00 6.21826053e-01 -4.52611782e-02 -1.35966492e+00
-3.29657868e-02 -1.17572618e+00 -5.77830732e-01 5.13350725e-01
-1.74733305e+00 -1.21675599e+00 -3.25916857e-01 6.33376479e-01
5.47520161e-01 -8.54481161e-01 5.26990175e-01 1.16781294e+00
-5.32923996e-01 7.07479239e-01 1.46808010e-02 8.30328986e-02
5.96883178e-01 -1.12525582e+00 2.96811372e-01 4.00396585e-01
1.25694469e-01 2.57245243e-01 3.55049878e-01 -2.40453109e-01
-1.02446318e+00 -1.41179633e+00 9.59324419e-01 -9.97519672e-01
1.00755453e+00 -4.44057077e-01 -7.28604853e-01 2.17209086e-01
-3.02521020e-01 1.91038311e-01 7.82662034e-01 -2.24467181e-02
-4.96154994e-01 -6.76601708e-01 -1.18004966e+00 1.05065338e-01
8.55031013e-01 -7.53075302e-01 -1.23199541e-02 5.45817077e-01
1.87091246e-01 3.67455184e-01 -8.18363011e-01 4.90864366e-01
3.71692717e-01 -1.24684823e+00 1.08714616e+00 -7.58970618e-01
-1.04794219e-01 -1.78813815e-01 -3.43572944e-01 -5.31717718e-01
-2.16617510e-01 -6.56624615e-01 -6.54909551e-01 1.45593953e+00
1.22432627e-01 -6.87574923e-01 1.10676539e+00 2.66008973e-01
-1.06208049e-01 -7.28524566e-01 -8.93838823e-01 -1.26061809e+00
-2.85985861e-02 -9.07692611e-01 4.96318102e-01 7.96102524e-01
-5.68460166e-01 2.61992902e-01 -3.56352568e-01 -1.88128918e-01
6.80011213e-01 -7.10997684e-03 1.25706840e+00 -1.80352736e+00
7.99833164e-02 -6.83465421e-01 -1.00066221e+00 -9.81938601e-01
5.55170715e-01 -8.18710625e-01 -5.61047733e-01 -9.63361621e-01
7.44472444e-02 -6.77750349e-01 -3.89965355e-01 2.58694649e-01
4.28922892e-01 6.58035576e-01 -1.49706334e-01 2.68140197e-01
-7.80694246e-01 -1.28680244e-01 5.84250629e-01 -1.89715400e-01
3.40463728e-01 7.31637001e-01 -7.46518075e-01 4.99836296e-01
7.36440599e-01 -3.61476004e-01 -2.64172912e-01 1.76872998e-01
7.34832883e-02 -5.03658950e-01 6.95511222e-01 -1.32100618e+00
3.71427909e-02 -6.67123795e-02 1.00007281e-01 -8.24332833e-01
1.39373988e-02 -1.08243680e+00 -3.97732109e-01 2.89954513e-01
1.64916478e-02 -1.47573382e-01 4.32147235e-02 6.29260838e-01
-2.54012078e-01 -6.44962192e-02 8.64250362e-01 -9.70620513e-02
-1.03301179e+00 6.57120466e-01 -7.85440326e-01 2.24134028e-01
1.25655317e+00 -5.61375797e-01 -3.83383572e-01 -5.32183588e-01
-5.91504455e-01 -1.04888014e-01 5.79258986e-02 5.02071559e-01
-1.45377927e-02 -1.10536957e+00 -6.22760534e-01 9.57154557e-02
2.45966703e-01 -7.28355408e-01 -1.43549338e-01 1.04171908e+00
-4.44418371e-01 8.66376817e-01 -3.37800495e-02 -7.77892947e-01
-1.27849424e+00 9.23253238e-01 1.68760970e-01 -3.33394736e-01
-3.04881066e-01 3.74088258e-01 4.53905016e-02 -6.42878830e-01
5.61070919e-01 1.55882046e-01 -1.32899478e-01 3.46128762e-01
3.83822650e-01 1.06184590e+00 4.89145309e-01 -8.59439075e-01
-5.79308212e-01 4.48206991e-01 -6.85066879e-02 6.60405159e-01
1.06805921e+00 -1.11428201e-01 -8.57708603e-02 2.94361562e-01
1.39425135e+00 -4.86187525e-02 -7.11319089e-01 -2.39842951e-01
3.18403572e-01 -4.16952342e-01 -1.20092764e-01 -6.18012369e-01
-1.07973909e+00 1.38910425e+00 7.25195348e-01 1.07007885e+00
9.42895949e-01 -9.73286778e-02 1.17831743e+00 5.45180440e-01
4.73344803e-01 -9.21929240e-01 7.25344196e-02 5.34759045e-01
5.52073168e-03 -1.36504519e+00 -1.37172759e-01 -6.23489022e-01
-2.28784055e-01 1.29965031e+00 6.37879968e-01 -8.84465352e-02
9.00564551e-01 5.27129591e-01 -8.86901170e-02 -3.38662207e-01
-7.62762845e-01 -7.42817461e-01 9.21664611e-02 1.05738974e+00
3.60657275e-01 -1.64678156e-01 1.17505148e-01 9.78388861e-02
1.10802643e-01 -4.16150205e-02 2.66682386e-01 6.58869386e-01
-5.98244190e-01 -1.20282924e+00 -2.83295035e-01 8.69779110e-01
-6.56512618e-01 7.60060921e-02 -4.91575986e-01 9.11018670e-01
4.38852996e-01 1.30469847e+00 1.95925146e-01 -4.26251322e-01
4.93113965e-01 1.13535292e-01 4.87471260e-02 -5.07229269e-01
-7.82832623e-01 -5.81830263e-01 3.08178872e-01 -5.24956942e-01
-3.79884660e-01 -2.39863738e-01 -4.48907286e-01 -8.08146358e-01
-2.77051181e-01 1.74751014e-01 5.20360172e-01 1.04764795e+00
3.18268269e-01 5.03769755e-01 1.04262125e+00 -6.70855403e-01
-4.42033231e-01 -4.97307897e-01 -6.72506571e-01 5.95363617e-01
3.33281606e-01 -9.24058914e-01 -6.81995332e-01 -3.12347524e-02] | [5.064985752105713, 7.233597755432129] |
1cfaa8dd-91f8-4e38-b769-a028c4f4aed6 | decoupled-diffusion-models-with-explicit | 2306.13720 | null | https://arxiv.org/abs/2306.13720v1 | https://arxiv.org/pdf/2306.13720v1.pdf | Decoupled Diffusion Models with Explicit Transition Probability | Recent diffusion probabilistic models (DPMs) have shown remarkable abilities of generated content, however, they often suffer from complex forward processes, resulting in inefficient solutions for the reversed process and prolonged sampling times. In this paper, we aim to address the aforementioned challenges by focusing on the diffusion process itself that we propose to decouple the intricate diffusion process into two comparatively simpler process to improve the generative efficacy and speed. In particular, we present a novel diffusion paradigm named DDM (\textbf{D}ecoupled \textbf{D}iffusion \textbf{M}odels) based on the It\^{o} diffusion process, in which the image distribution is approximated by an explicit transition probability while the noise path is controlled by the standard Wiener process. We find that decoupling the diffusion process reduces the learning difficulty and the explicit transition probability improves the generative speed significantly. We prove a new training objective for DPM, which enables the model to learn to predict the noise and image components separately. Moreover, given the novel forward diffusion equation, we derive the reverse denoising formula of DDM that naturally supports fewer steps of generation without ordinary differential equation (ODE) based accelerators. Our experiments demonstrate that DDM outperforms previous DPMs by a large margin in fewer function evaluations setting and gets comparable performances in long function evaluations setting. We also show that our framework can be applied to image-conditioned generation and high-resolution image synthesis, and that it can generate high-quality images with only 10 function evaluations. | ['Kai Xu', 'Xinwang Liu', 'Zheng Qin', 'Yuhang Huang'] | 2023-06-23 | null | null | null | null | ['image-generation'] | ['computer-vision'] | [ 3.96598399e-01 -6.52953163e-02 3.37528169e-01 1.10087715e-01
-7.03583777e-01 -4.75945622e-01 8.07816148e-01 -3.37505072e-01
-3.24191779e-01 6.49329901e-01 -2.88638640e-02 -3.38003218e-01
-3.54471982e-01 -8.96403074e-01 -6.37396693e-01 -1.41050529e+00
1.89348578e-01 4.28483665e-01 2.74165243e-01 -1.86162010e-01
8.38003606e-02 4.26276505e-01 -1.18004727e+00 -1.27696261e-01
1.06471944e+00 1.06837928e+00 5.39909005e-01 8.96501839e-01
1.51369313e-04 9.42320764e-01 -6.71656787e-01 -4.20015514e-01
2.46188641e-01 -7.69337654e-01 -3.83496344e-01 1.39917331e-02
-7.02819228e-02 -3.78288031e-01 -6.94309533e-01 1.13143182e+00
9.74669635e-01 1.23765014e-01 9.04891491e-01 -9.92052436e-01
-8.59722078e-01 6.36086583e-01 -8.18047643e-01 2.02852175e-01
-1.22866116e-01 4.48408633e-01 4.08394665e-01 -7.31102765e-01
5.91001451e-01 1.16944361e+00 6.26668155e-01 7.47079015e-01
-1.45475924e+00 -6.75910234e-01 -5.67346588e-02 -4.00806144e-02
-1.36857009e+00 -3.35563004e-01 7.83519149e-01 -3.77350956e-01
7.55495489e-01 7.23181665e-02 3.63353491e-01 1.29336333e+00
5.72775364e-01 6.95131063e-01 1.29438329e+00 -2.65929401e-01
4.48532313e-01 -2.08889574e-01 6.36542812e-02 7.36731827e-01
6.77736178e-02 1.88015148e-01 -4.57538575e-01 -1.61160395e-01
1.12822807e+00 -2.29995430e-01 -5.25669217e-01 -8.91386792e-02
-1.21073818e+00 8.83774102e-01 1.91655606e-01 1.42256737e-01
-3.92951280e-01 4.54770952e-01 9.29361507e-02 1.23727687e-01
3.32373232e-01 1.36207029e-01 -7.44257942e-02 -1.38261139e-01
-1.16344571e+00 3.71511519e-01 8.23048592e-01 9.44149613e-01
5.02925456e-01 3.85463029e-01 -4.33737576e-01 6.14265859e-01
3.22882950e-01 9.44575787e-01 4.23432708e-01 -1.18401825e+00
1.99401483e-01 -1.96091667e-01 1.01313345e-01 -7.48025954e-01
-2.99933106e-01 -6.33193195e-01 -1.65884006e+00 4.53094333e-01
3.81875575e-01 -5.28460920e-01 -9.96471822e-01 1.76346481e+00
3.50208282e-01 3.38341892e-01 1.59779266e-01 7.06512094e-01
3.72832566e-01 1.20989871e+00 -1.54287830e-01 -3.57015669e-01
1.17272794e+00 -1.07456720e+00 -7.23205686e-01 1.60757661e-01
3.62161070e-01 -8.36971581e-01 7.34287620e-01 8.25273335e-01
-1.49086106e+00 -4.61067557e-01 -1.13280797e+00 2.19731382e-03
2.90152192e-01 9.60085541e-02 5.72776914e-01 7.26077676e-01
-1.29931748e+00 1.03049231e+00 -8.51104856e-01 2.11475268e-01
4.25323129e-01 3.16598833e-01 1.82426378e-01 -1.27850488e-01
-1.01443028e+00 6.37312174e-01 -1.79528575e-02 2.85065711e-01
-1.23599577e+00 -7.64102221e-01 -5.62990427e-01 1.16905883e-01
4.10473794e-02 -1.19725645e+00 1.03295064e+00 -6.60392046e-01
-2.23109627e+00 2.58853763e-01 -4.96667400e-02 -6.71826839e-01
7.61553526e-01 1.15678579e-01 -2.59610355e-01 2.06040323e-01
-1.59609228e-01 4.80428845e-01 1.43622184e+00 -1.13966620e+00
-3.65190178e-01 -1.22900411e-01 -2.54370451e-01 1.78761519e-02
-1.94437504e-01 -3.04644525e-01 -2.91980803e-01 -9.47098970e-01
-4.37083654e-02 -8.04879785e-01 -5.55647910e-01 2.11829264e-02
-2.58069992e-01 1.93308905e-01 6.44163191e-01 -5.30280888e-01
1.22607005e+00 -2.02683091e+00 5.15897274e-01 2.07962230e-01
4.93371516e-01 2.75840938e-01 -2.47249752e-02 3.05397034e-01
1.55657083e-01 -6.66503906e-02 -5.22501051e-01 -4.80941772e-01
4.29471321e-02 2.83970386e-01 -5.17108560e-01 4.30441618e-01
1.01576746e-01 6.99079692e-01 -7.73551464e-01 -3.18023145e-01
-9.17033851e-02 8.23341012e-01 -6.02904379e-01 -3.90518904e-02
-1.95157096e-01 7.03632653e-01 -2.75833726e-01 1.68659538e-01
1.14463818e+00 -2.95798630e-01 -3.14488448e-02 -2.62974352e-01
1.20810112e-02 -1.83242679e-01 -1.34956992e+00 1.67404556e+00
-6.57392800e-01 3.20867449e-01 4.29603726e-01 -8.52621436e-01
8.20564628e-01 1.81372449e-01 2.37443805e-01 -5.75816929e-01
7.58989453e-02 2.34217554e-01 3.93996835e-02 -9.69409198e-02
3.91507030e-01 -4.00178134e-01 1.01478957e-01 4.83547926e-01
2.40238518e-01 -4.37153250e-01 1.07569113e-01 2.10757583e-01
1.33580804e+00 4.32076305e-02 -2.24214971e-01 -3.68786693e-01
5.10653019e-01 -2.90799171e-01 3.73673141e-01 9.99378860e-01
9.40626040e-02 6.22783840e-01 4.65115428e-01 7.50044063e-02
-1.07217526e+00 -1.29287088e+00 -6.48733005e-02 4.15308893e-01
2.70341367e-01 -4.12115544e-01 -8.73059213e-01 -3.80140781e-01
-4.25552219e-01 7.63784111e-01 -3.61535877e-01 -1.51589736e-01
-6.46555424e-01 -1.43998623e+00 6.54705226e-01 2.77202994e-01
7.01405466e-01 -6.10459507e-01 -2.41491124e-01 3.16657424e-01
2.39348918e-01 -8.26377511e-01 -5.32040060e-01 2.16745496e-01
-8.75716269e-01 -3.67379934e-01 -1.36549282e+00 -7.07270920e-01
5.62955141e-01 2.60674935e-02 8.07385087e-01 -3.79324913e-01
-4.89783287e-02 2.42985100e-01 1.23471290e-01 -5.66598177e-02
-6.43264115e-01 -5.51182292e-02 -9.50301960e-02 6.75084814e-02
-4.33150262e-01 -1.00588346e+00 -9.65341926e-01 2.48212233e-01
-1.07393134e+00 3.62450331e-01 7.03716040e-01 1.00133860e+00
7.44601667e-01 6.43557668e-01 4.01192337e-01 -6.60230577e-01
7.02085614e-01 -3.88851434e-01 -8.15286994e-01 -1.80437230e-02
-7.76431620e-01 4.45136368e-01 7.29820251e-01 -7.27488518e-01
-1.43197238e+00 -4.67210300e-02 -4.70689952e-01 -3.88681799e-01
3.56354654e-01 1.58679232e-01 -1.27002448e-01 -1.18734486e-01
5.47807992e-01 5.86085975e-01 -1.12193609e-02 -5.20854890e-01
4.91700023e-01 3.59005004e-01 5.53199589e-01 -7.78764784e-01
9.90186512e-01 7.17748880e-01 4.00199950e-01 -8.42443347e-01
-3.82519096e-01 1.21364370e-01 -9.04362947e-02 -1.22048669e-01
8.82918835e-01 -8.55656862e-01 -8.16297710e-01 8.54435742e-01
-1.27107322e+00 -5.26189983e-01 -4.07023549e-01 4.03673500e-01
-6.36437118e-01 5.87574482e-01 -1.01549160e+00 -9.12939012e-01
-5.89768052e-01 -1.38279676e+00 1.02631617e+00 1.75762445e-01
1.13510169e-01 -8.87834251e-01 -3.78176980e-02 -1.14056453e-01
6.86200619e-01 1.75657943e-02 1.05584991e+00 9.39956680e-02
-8.26825082e-01 1.15361236e-01 -1.03375368e-01 6.22102618e-01
-2.22938061e-01 -8.06959346e-02 -7.08997786e-01 -2.06933156e-01
7.89710820e-01 1.92422584e-01 1.11215150e+00 5.09486914e-01
9.59193289e-01 -2.41401702e-01 -3.05651307e-01 8.19439948e-01
1.42726672e+00 1.68648124e-01 8.97949815e-01 -7.03473836e-02
6.27719998e-01 2.45266289e-01 -9.32392199e-03 6.19497478e-01
3.04247700e-02 4.46117014e-01 1.22593977e-01 5.49986847e-02
-4.71200794e-01 -7.21966773e-02 6.67810678e-01 1.11847830e+00
-6.26231879e-02 -5.63188553e-01 -5.67867935e-01 9.35233906e-02
-1.60679615e+00 -8.82397532e-01 -4.13053066e-01 2.17752266e+00
8.99885178e-01 1.68764353e-01 -2.97152609e-01 1.33275121e-01
6.89143777e-01 8.73036534e-02 -5.93113601e-01 -1.23936273e-01
-2.96607971e-01 6.68929636e-01 5.65681815e-01 6.67396307e-01
-7.52944946e-01 5.42465448e-01 6.19008255e+00 1.44374144e+00
-1.01399660e+00 4.37863410e-01 7.36037910e-01 -1.95950150e-01
-4.95224833e-01 -5.54035529e-02 -9.45567012e-01 6.95673108e-01
9.93974686e-01 -1.68870509e-01 5.17324626e-01 6.17925882e-01
1.75663412e-01 -1.40812263e-01 -7.38038063e-01 1.16331732e+00
-9.34855491e-02 -1.35786664e+00 1.92524612e-01 1.14153743e-01
8.36877346e-01 -2.59158701e-01 4.19131458e-01 8.97933841e-02
4.00760233e-01 -9.37165201e-01 7.36705899e-01 7.81233072e-01
5.74824989e-01 -7.34155893e-01 3.34122568e-01 5.33392251e-01
-8.44280660e-01 1.05478086e-01 -3.34670544e-01 1.22976787e-01
5.22842824e-01 1.15006721e+00 -2.16196641e-01 5.94525278e-01
4.52047706e-01 3.24781209e-01 -1.09114438e-01 8.48956168e-01
-2.17053786e-01 6.14397109e-01 -4.40028578e-01 8.21507871e-02
1.26245409e-01 -6.06373668e-01 6.96005464e-01 9.79212463e-01
8.98388207e-01 5.63482521e-03 -2.21533686e-01 1.04447424e+00
4.46955413e-02 -2.99084038e-01 -1.84919700e-01 1.55383870e-01
1.14938654e-02 1.20413268e+00 -7.31076717e-01 -3.59535903e-01
-8.38351250e-02 1.48580599e+00 -5.57354055e-02 5.53898871e-01
-1.23928058e+00 -2.39672616e-01 5.58640838e-01 5.54977804e-02
6.08006418e-01 -4.72827077e-01 -2.22779170e-01 -1.21699572e+00
-1.26177862e-01 -7.31332600e-01 -8.65738913e-02 -7.11850941e-01
-1.41585267e+00 7.40394831e-01 -2.61131227e-01 -9.33341444e-01
-1.16627254e-01 -4.81496930e-01 -3.79424751e-01 1.18336296e+00
-1.52688456e+00 -8.04721653e-01 -1.63875028e-01 5.49622953e-01
3.76230121e-01 1.54534742e-01 5.85155189e-01 4.60387200e-01
-6.98473692e-01 3.96974951e-01 6.71244681e-01 -4.20973092e-01
5.22675335e-01 -1.23892331e+00 4.18808073e-01 8.44676614e-01
-2.24068686e-01 3.80678743e-01 7.62464523e-01 -7.64126122e-01
-1.62385356e+00 -9.52022374e-01 3.13076884e-01 -2.04079941e-01
7.63378084e-01 -4.06914443e-01 -6.59506023e-01 1.74697176e-01
3.16405058e-01 -2.12831214e-01 3.02720279e-01 -6.16138935e-01
2.58683618e-02 -1.12500124e-01 -1.11115575e+00 7.06101537e-01
1.08616483e+00 -2.33155027e-01 -1.12825250e-02 3.76558900e-01
6.32081866e-01 -5.32578409e-01 -8.98048341e-01 2.06499115e-01
2.33692899e-01 -8.76410067e-01 1.08679628e+00 1.63100302e-01
6.55450761e-01 -4.23171103e-01 -4.70778085e-02 -1.13999724e+00
-2.83527285e-01 -1.30238986e+00 -5.98859429e-01 1.40010774e+00
2.01135859e-01 -6.68340743e-01 5.40341258e-01 2.81046331e-01
-1.02197610e-01 -8.38337004e-01 -9.92167354e-01 -8.47022772e-01
2.75296479e-01 -4.71456379e-01 3.51217002e-01 3.12540442e-01
-6.01652145e-01 4.19183880e-01 -6.11248493e-01 3.84498000e-01
8.02527547e-01 4.05361764e-02 4.16509986e-01 -8.13013017e-01
-1.04503441e+00 -5.81486166e-01 -3.40304412e-02 -1.72325099e+00
-2.39703387e-01 -7.61934459e-01 9.51366276e-02 -1.40351629e+00
5.18415086e-02 -5.98652482e-01 7.68621713e-02 -3.35755378e-01
-1.29447743e-01 1.26520276e-01 7.42885172e-02 3.22149247e-01
-2.05710962e-01 9.58605409e-01 1.53237033e+00 -1.90631032e-01
-4.50836867e-02 1.64487377e-01 -4.78854716e-01 6.18652105e-01
4.88269269e-01 -4.77847666e-01 -6.55080974e-01 -6.36197150e-01
1.97272614e-01 1.97228104e-01 3.80812764e-01 -1.28837276e+00
4.63135213e-01 3.03723902e-01 4.17711556e-01 -3.87913018e-01
5.39760947e-01 -4.57312226e-01 4.19233918e-01 4.70271796e-01
-6.01437762e-02 -1.31724393e-02 5.51301241e-02 7.54573107e-01
1.64249558e-02 -4.99602973e-01 9.13155973e-01 -1.53942946e-02
-2.56786555e-01 4.32256550e-01 -7.08948851e-01 -1.36221170e-01
8.65577340e-01 2.43521541e-01 -2.32022807e-01 -5.36575615e-01
-8.62710476e-01 -1.69246867e-01 3.77788067e-01 -1.91172302e-01
4.53195512e-01 -1.26709187e+00 -5.94208658e-01 1.16877787e-01
-7.47815013e-01 1.39149755e-01 7.66217828e-01 9.85016763e-01
-6.44303977e-01 -8.80685225e-02 2.57683933e-01 -6.04364276e-01
-7.17258155e-01 7.74083436e-01 3.51479113e-01 -5.70593476e-01
-8.00820470e-01 9.60787535e-01 5.03615558e-01 4.98585105e-02
8.71226117e-02 -2.50607997e-01 3.91719401e-01 -6.56242967e-02
6.99368417e-01 5.33186495e-01 -9.29926708e-02 -3.63810867e-01
8.15797225e-02 6.32611036e-01 -6.65626377e-02 -3.83023143e-01
1.30505800e+00 -2.52412885e-01 -1.52701005e-01 2.94950586e-02
1.10153067e+00 5.70353046e-02 -1.64786410e+00 -4.15827632e-02
-4.35529917e-01 -7.57896602e-02 4.53993499e-01 -6.55013323e-01
-1.24675214e+00 9.77184236e-01 8.01666081e-01 1.08921647e-01
1.40512502e+00 -2.56085515e-01 1.30368626e+00 6.00992851e-02
3.37983638e-01 -8.25916290e-01 4.26252969e-02 3.65393788e-01
7.02715635e-01 -5.98451257e-01 -9.70071256e-02 -5.32993793e-01
-4.00370270e-01 1.13433981e+00 6.82191402e-02 -1.97054520e-01
8.03795397e-01 7.63812125e-01 -3.99834335e-01 -3.90210561e-02
-5.66003084e-01 5.49440198e-02 -9.80524272e-02 6.28458381e-01
-1.72676146e-02 -2.62443244e-01 -2.47880369e-01 7.66582429e-01
1.45845152e-02 6.26581758e-02 4.60169286e-01 8.64917040e-01
-1.94881111e-01 -1.33054388e+00 -3.63734215e-01 1.34788632e-01
-3.56660843e-01 -4.65934902e-01 3.12734753e-01 4.33167309e-01
1.45234570e-01 6.17246568e-01 -1.11693211e-01 -1.15862526e-01
1.43355438e-02 -4.07543667e-02 7.89127648e-01 -3.40544060e-02
-3.69228274e-01 3.20412248e-01 -3.04664105e-01 -2.87389219e-01
-1.53205454e-01 -5.22151887e-01 -1.16601300e+00 -5.78499556e-01
-2.66622990e-01 6.94785863e-02 5.59590340e-01 6.86757267e-01
6.20001554e-01 7.55458236e-01 4.96534169e-01 -9.73793805e-01
-8.25518370e-01 -7.64980972e-01 -7.02793002e-01 -1.18278705e-01
1.95925876e-01 -4.66953754e-01 -3.94069523e-01 -4.36814241e-02] | [11.230767250061035, -0.548655092716217] |
c3915dab-e9b1-4a30-8662-3f4f6ddec31c | deep-snake-for-real-time-instance | 2001.01629 | null | https://arxiv.org/abs/2001.01629v3 | https://arxiv.org/pdf/2001.01629v3.pdf | Deep Snake for Real-Time Instance Segmentation | This paper introduces a novel contour-based approach named deep snake for real-time instance segmentation. Unlike some recent methods that directly regress the coordinates of the object boundary points from an image, deep snake uses a neural network to iteratively deform an initial contour to match the object boundary, which implements the classic idea of snake algorithms with a learning-based approach. For structured feature learning on the contour, we propose to use circular convolution in deep snake, which better exploits the cycle-graph structure of a contour compared against generic graph convolution. Based on deep snake, we develop a two-stage pipeline for instance segmentation: initial contour proposal and contour deformation, which can handle errors in object localization. Experiments show that the proposed approach achieves competitive performances on the Cityscapes, KINS, SBD and COCO datasets while being efficient for real-time applications with a speed of 32.3 fps for 512$\times$512 images on a 1080Ti GPU. The code is available at https://github.com/zju3dv/snake/. | ['Xiuli Li', 'Sida Peng', 'Hujun Bao', 'Wen Jiang', 'Xiaowei Zhou', 'Huaijin Pi'] | 2020-01-06 | deep-snake-for-real-time-instance-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Peng_Deep_Snake_for_Real-Time_Instance_Segmentation_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Peng_Deep_Snake_for_Real-Time_Instance_Segmentation_CVPR_2020_paper.pdf | cvpr-2020-6 | ['real-time-instance-segmentation'] | ['computer-vision'] | [-2.38834694e-01 6.76656738e-02 6.50042593e-02 -3.49325359e-01
-5.54119408e-01 -6.32514656e-01 2.58900076e-01 4.39153790e-01
-6.18228495e-01 9.56104398e-02 -4.64784384e-01 -2.94438094e-01
4.58752900e-01 -9.23013449e-01 -5.27212620e-01 -5.54984152e-01
-3.18870038e-01 4.51705098e-01 7.24408448e-01 -1.33435220e-01
4.88221705e-01 8.97565186e-01 -9.84774828e-01 -4.23484109e-02
6.56345487e-01 9.62888956e-01 -4.53616641e-02 7.84114838e-01
-3.06309760e-01 1.75237745e-01 -2.06990033e-01 -5.30673862e-01
3.96255940e-01 -1.03132635e-01 -7.40575373e-01 2.89482623e-01
3.91587287e-01 -9.87002477e-02 -9.41770971e-02 1.20205259e+00
4.63552266e-01 -8.89912024e-02 5.18576443e-01 -1.05417120e+00
-3.07135403e-01 1.37882695e-01 -8.99529755e-01 1.17558070e-01
6.61815926e-02 1.88720241e-01 6.76955462e-01 -9.09285486e-01
5.85082531e-01 1.23457193e+00 1.13841379e+00 2.61918336e-01
-1.10837305e+00 -4.38408166e-01 -1.12082846e-01 -2.07066163e-01
-1.37325680e+00 -2.98893563e-02 9.95955288e-01 -3.17140788e-01
4.80009079e-01 4.91407663e-02 8.86810005e-01 3.78950387e-01
3.14270437e-01 8.05673480e-01 9.96946394e-01 -2.93381423e-01
2.85856009e-01 -2.96839952e-01 3.29913586e-01 1.21625006e+00
2.80387282e-01 -5.46339601e-02 -1.14796888e-02 -2.32345194e-01
1.24083233e+00 -1.53977498e-01 -6.70049638e-02 -5.66741288e-01
-7.45036364e-01 1.06185627e+00 9.48029459e-01 8.77593011e-02
-3.00286502e-01 4.00882989e-01 3.60556602e-01 -1.12204775e-01
4.76536930e-01 -1.41493648e-01 -3.14180583e-01 2.60946840e-01
-1.04683232e+00 1.89354584e-01 8.04438829e-01 8.55487108e-01
9.05485809e-01 6.43935129e-02 1.64820567e-01 4.79390740e-01
6.16447747e-01 3.69980425e-01 3.95721555e-01 -1.19014764e+00
9.34922881e-03 6.94915235e-01 -5.16523086e-02 -1.15412724e+00
-7.09714532e-01 -2.79360503e-01 -8.50104928e-01 7.30578363e-01
6.96688771e-01 -1.40010178e-01 -1.41296101e+00 1.03004587e+00
8.77604723e-01 4.14698362e-01 -2.30981633e-01 8.77245009e-01
8.92029643e-01 5.34318388e-01 1.60451755e-01 1.57051966e-01
1.58710730e+00 -1.08685541e+00 -3.89166683e-01 -2.36958906e-01
5.54086626e-01 -7.71921396e-01 9.13874686e-01 1.25515819e-01
-1.28607273e+00 -5.62706351e-01 -8.14872205e-01 -2.53302842e-01
-3.38563472e-01 1.60147175e-02 8.35891426e-01 7.01823652e-01
-1.21975529e+00 8.35498929e-01 -1.40110469e+00 -2.76541740e-01
8.48079681e-01 4.21329558e-01 -1.01855420e-01 3.04641932e-01
-5.25992095e-01 4.47110862e-01 2.77677298e-01 2.06406295e-01
-5.00388861e-01 -5.80121219e-01 -9.06191230e-01 4.12860848e-02
-2.48373784e-02 -4.85096395e-01 9.60017741e-01 -9.90023136e-01
-1.58896887e+00 1.21137035e+00 1.42026991e-01 -6.59104049e-01
7.29247749e-01 -5.41101359e-02 1.77531034e-01 3.12035233e-01
-1.85827408e-02 9.31803524e-01 8.44814539e-01 -1.17039633e+00
-5.40677488e-01 -4.72643733e-01 -3.64348054e-01 -7.85988867e-02
2.23822311e-01 1.46035859e-02 -8.78818452e-01 -7.02828825e-01
4.23365980e-01 -9.31049347e-01 -6.19258106e-01 5.97640753e-01
-2.79434413e-01 -3.94371182e-01 9.78956640e-01 -5.94957709e-01
7.27759898e-01 -2.18076205e+00 -3.46691400e-01 2.90064454e-01
4.14782614e-02 5.80141008e-01 1.93225127e-02 1.85627848e-01
1.41798317e-01 -2.47156098e-02 -8.20564985e-01 -5.63674688e-01
-3.89991522e-01 1.38191819e-01 1.06279582e-01 8.05208564e-01
2.82527298e-01 1.03764093e+00 -7.53179073e-01 -7.80388653e-01
1.25065058e-01 7.38610804e-01 -3.61870170e-01 2.25620136e-01
-8.44490975e-02 3.85948360e-01 -5.57082295e-01 8.13960671e-01
1.16532540e+00 -2.45943069e-01 -4.84827198e-02 -1.46833167e-03
-3.43949080e-01 -2.98176646e-01 -1.28453827e+00 2.11786675e+00
2.09891461e-02 4.53596264e-01 5.04198670e-01 -9.52550828e-01
1.18277383e+00 3.98884527e-02 3.89162332e-01 -3.19863677e-01
3.38365108e-01 2.66224682e-01 -2.90472418e-01 -2.03032210e-01
2.32186049e-01 -4.59860153e-02 1.39639184e-01 1.57821193e-01
-3.47751164e-04 -4.74138200e-01 1.03926644e-01 1.40840532e-02
8.88584554e-01 5.98409891e-01 1.09454423e-01 -5.58499396e-01
5.24335027e-01 3.96494448e-01 5.01646757e-01 4.21008080e-01
-4.45228577e-01 8.31290364e-01 4.16081876e-01 -8.63269925e-01
-7.85114586e-01 -1.03294837e+00 -2.69541413e-01 7.38046885e-01
3.54621530e-01 1.67041253e-02 -1.11383092e+00 -5.21904767e-01
-4.64349538e-02 3.45312625e-01 -5.87744474e-01 3.00936818e-01
-1.04700661e+00 -6.96213186e-01 5.03305256e-01 5.52387178e-01
7.83911109e-01 -1.27506936e+00 -1.18732071e+00 4.20467257e-01
2.51846641e-01 -8.14537466e-01 -6.10657454e-01 1.98814660e-01
-1.29207790e+00 -1.02233636e+00 -8.77958596e-01 -1.17068017e+00
6.98608398e-01 4.60383818e-02 1.00661480e+00 4.12781864e-01
-6.04667783e-01 9.96953994e-02 -5.18704355e-02 -2.06087306e-01
-1.63952798e-01 -7.59175271e-02 -5.89667559e-01 -5.96548170e-02
-7.57667199e-02 -5.09370148e-01 -9.13055062e-01 2.48927977e-02
-7.25829363e-01 4.65875566e-02 4.12046403e-01 5.37829399e-01
7.62949347e-01 -3.96420509e-01 2.32224479e-01 -7.90673077e-01
3.68918151e-01 -2.24537030e-01 -9.32573199e-01 7.63940886e-02
-4.51130122e-01 -4.76177260e-02 1.60336897e-01 -3.08369637e-01
-8.73707712e-01 7.27529049e-01 -5.64875007e-01 -4.67797875e-01
-2.59500355e-01 2.26912573e-01 2.63120264e-01 -4.15969133e-01
4.05541450e-01 1.35692567e-01 1.76788911e-01 -6.91170990e-01
4.51840162e-01 3.19948465e-01 7.73456037e-01 -4.32762533e-01
6.42209709e-01 9.15239990e-01 1.30526438e-01 -7.63122439e-01
-4.30681020e-01 -5.79070926e-01 -1.03238630e+00 -1.63519502e-01
1.02996457e+00 -6.40985906e-01 -7.88342178e-01 6.68707609e-01
-1.36168349e+00 -6.53691709e-01 -2.03999057e-01 4.96673845e-02
-6.59499049e-01 5.42615891e-01 -8.54930937e-01 -5.78445971e-01
-7.37429857e-01 -9.78240252e-01 1.13702726e+00 6.86385810e-01
2.19797745e-01 -1.23262787e+00 2.05667138e-01 -1.30866274e-01
2.91045636e-01 6.84125423e-01 3.94194961e-01 -4.22721177e-01
-4.87814456e-01 -3.51184517e-01 -3.09621871e-01 3.64959612e-02
-3.98059189e-02 2.52656221e-01 -6.91402555e-01 -3.28649014e-01
-5.75723834e-02 -9.43123326e-02 7.85534620e-01 5.81142664e-01
9.60131943e-01 5.89848671e-04 -4.83248532e-01 9.85428214e-01
1.66664290e+00 1.65093169e-01 6.62500679e-01 2.74336904e-01
7.56856978e-01 3.94123912e-01 4.94201809e-01 3.03815216e-01
3.80261958e-01 4.00622159e-01 5.57080746e-01 -5.91605663e-01
-2.02777788e-01 5.09110019e-02 5.68775013e-02 5.10391653e-01
-1.20703220e-01 2.95220353e-02 -1.16234195e+00 6.17341101e-01
-1.63233173e+00 -4.69270706e-01 -7.24704862e-01 1.94797409e+00
7.80417383e-01 3.07770044e-01 2.12786034e-01 -1.61387786e-01
7.69893527e-01 -6.69589490e-02 -3.97389174e-01 -5.90923190e-01
1.78299263e-01 6.06899261e-01 6.02166831e-01 6.60129547e-01
-1.36995721e+00 1.21086490e+00 5.47042608e+00 7.63351858e-01
-1.32073772e+00 1.56282321e-01 7.95410454e-01 6.13305807e-01
2.56287247e-01 5.59916571e-02 -5.43770552e-01 2.68681645e-01
6.99803591e-01 1.37277201e-01 -1.53442696e-01 8.73960853e-01
1.14448056e-01 -7.90325329e-02 -5.10715246e-01 6.87104404e-01
-3.14827859e-01 -1.52476704e+00 -3.51808190e-01 -6.38813451e-02
4.49488968e-01 3.26767862e-01 -2.45862678e-01 -1.69418901e-01
3.67340833e-01 -8.80984485e-01 7.50660002e-01 2.91930228e-01
4.37461019e-01 -7.61465549e-01 5.45748770e-01 2.57154167e-01
-1.48051524e+00 4.08740222e-01 -4.32510793e-01 1.31522790e-01
2.09847525e-01 2.49280035e-01 -7.24486828e-01 3.88535112e-01
8.80820692e-01 4.43168491e-01 -6.84095442e-01 1.39043450e+00
-3.12336683e-01 5.80526054e-01 -6.08579755e-01 1.75064489e-01
5.58651865e-01 -4.57441777e-01 3.34979504e-01 1.58342528e+00
1.44828215e-01 4.10986781e-01 2.99437195e-01 7.64961660e-01
1.34100514e-02 1.98765025e-01 -3.69873554e-01 4.04531807e-01
4.03797552e-02 1.72952247e+00 -1.70052814e+00 -4.05136138e-01
-2.75974035e-01 1.13942575e+00 3.58964086e-01 2.15060543e-02
-8.86325300e-01 -4.49314743e-01 2.69975126e-01 1.00384377e-01
7.46705532e-01 -5.63448966e-01 -5.09587526e-01 -7.33556032e-01
-2.81778187e-01 -2.38871679e-01 4.49630111e-01 -6.76588058e-01
-8.47682655e-01 6.98885441e-01 -3.49672765e-01 -9.94563937e-01
4.27392900e-01 -5.77740729e-01 -1.16109407e+00 5.72708070e-01
-1.45335174e+00 -1.40124893e+00 -5.00379145e-01 5.22258282e-01
5.89023590e-01 5.75083077e-01 7.49764740e-01 3.38751264e-02
-3.62515837e-01 1.84782848e-01 1.00393137e-02 6.48321807e-01
2.70471543e-01 -1.36777258e+00 9.32787061e-01 7.68360615e-01
2.44400725e-01 3.77835453e-01 4.78758097e-01 -7.82552958e-01
-1.15194809e+00 -1.12660897e+00 5.07312238e-01 -1.47242069e-01
5.57926059e-01 -2.92392910e-01 -1.17843473e+00 6.13821089e-01
5.14039516e-01 7.14922369e-01 2.13690057e-01 -6.80217445e-01
-1.57845225e-02 2.66686678e-01 -1.38993502e+00 4.50697571e-01
8.71181607e-01 2.57691532e-01 -4.43616509e-01 2.70451486e-01
5.07988095e-01 -7.17188537e-01 -8.87055576e-01 2.41694838e-01
3.79178792e-01 -9.93647635e-01 1.09174585e+00 -3.81190896e-01
1.26982806e-02 -4.66509938e-01 3.49010140e-01 -9.39582765e-01
-1.98945612e-01 -7.21673608e-01 1.72613844e-01 1.01972055e+00
1.23139918e-01 -4.42767113e-01 1.13470089e+00 3.63319904e-01
-2.20234796e-01 -9.11138594e-01 -1.07444167e+00 -6.14482343e-01
3.27450305e-01 -2.24239141e-01 2.64485151e-01 7.88782358e-01
-4.32941288e-01 -5.29536158e-02 2.58073032e-01 3.63657802e-01
9.94963169e-01 3.43145519e-01 5.89667916e-01 -1.26863337e+00
1.48405775e-01 -4.77276593e-01 -5.30704796e-01 -9.37706530e-01
-7.76140671e-03 -8.41886997e-01 3.92738171e-03 -1.65329742e+00
-2.34350279e-01 -7.24991322e-01 5.93326353e-02 6.36239648e-01
-8.68441761e-02 6.12453699e-01 1.90131024e-01 1.35320544e-01
-3.29856873e-01 2.85943806e-01 1.39451444e+00 -8.36084038e-03
-3.02130014e-01 2.22745687e-01 -2.10462317e-01 1.15318882e+00
9.59404111e-01 -7.65049517e-01 1.29325911e-01 -3.01632822e-01
-3.75177115e-01 2.13546261e-01 5.66289365e-01 -1.03784275e+00
4.81811464e-01 3.97848636e-02 3.29569429e-01 -7.24471688e-01
2.63584554e-01 -7.95223355e-01 3.22749428e-02 1.12414265e+00
8.34710002e-02 2.22274706e-01 4.14761931e-01 4.82990980e-01
2.37191170e-02 -3.92260224e-01 1.20465469e+00 -3.59835118e-01
-7.23896563e-01 3.59000981e-01 -3.11576694e-01 5.65518625e-02
1.23413455e+00 -3.33661497e-01 4.44398373e-02 6.47721253e-03
-8.96939337e-01 2.03939557e-01 5.42151690e-01 1.90492123e-02
6.87465012e-01 -1.09173095e+00 -7.41784215e-01 2.10049435e-01
-3.84171933e-01 4.71932083e-01 -1.37902945e-01 8.25237155e-01
-1.27836287e+00 -8.09369907e-02 -3.09801310e-01 -8.97204340e-01
-1.21979582e+00 5.20896971e-01 5.77579021e-01 -4.26714011e-02
-1.01526535e+00 1.08944070e+00 2.60957163e-02 -2.27735788e-01
2.38384735e-02 -5.07001519e-01 -7.78739303e-02 -3.17786001e-02
1.40553162e-01 2.45539278e-01 1.61908135e-01 -6.69286072e-01
-6.28190577e-01 9.49923038e-01 2.03074709e-01 -6.94420040e-02
1.37619841e+00 1.18927509e-01 -1.24157339e-01 5.77846095e-02
1.19622123e+00 -9.30888951e-02 -1.61575317e+00 -8.55632499e-02
4.22393858e-01 -3.33005697e-01 1.44325286e-01 -4.01412278e-01
-1.48662317e+00 9.59594429e-01 1.01544547e+00 1.34663507e-01
9.46049750e-01 2.04094425e-02 8.30018342e-01 -1.23133939e-02
2.55959690e-01 -8.30284059e-01 5.25519326e-02 2.89625168e-01
7.76383698e-01 -1.13723624e+00 6.52343556e-02 -6.89034998e-01
-4.94835436e-01 1.50292861e+00 2.91061759e-01 -7.91449070e-01
9.28293347e-01 5.17681658e-01 4.52439994e-01 -4.38566804e-01
3.66667919e-02 -2.54891038e-01 9.73945335e-02 4.57575560e-01
3.10281605e-01 6.27378225e-02 -4.68932003e-01 1.77861243e-01
1.62334248e-01 -1.61479443e-01 4.86050189e-01 1.19327295e+00
-6.75941050e-01 -1.12026703e+00 -4.96785700e-01 -6.59527183e-02
-5.06718218e-01 6.58364873e-03 -1.58648655e-01 9.66121674e-01
-7.42641883e-03 6.09292209e-01 4.14044052e-01 2.31130764e-01
3.22259545e-01 9.91171040e-03 3.73624831e-01 -4.56822932e-01
-8.78064632e-01 4.68012750e-01 -3.26658368e-01 -6.96166694e-01
-4.34679180e-01 -7.74185181e-01 -1.91048336e+00 -6.50000200e-02
-1.86558336e-01 9.61100981e-02 8.76877248e-01 5.04691541e-01
3.66161495e-01 1.04535952e-01 4.31317329e-01 -1.30997074e+00
-2.62361258e-01 -7.26296842e-01 -3.82004768e-01 3.24322253e-01
1.66444585e-01 -4.88138616e-01 -1.17189936e-01 3.99621457e-01] | [9.514043807983398, 0.14285610616207123] |
47f318a5-b24f-4cfc-8b9c-7ec17e17cac7 | dynamic-causal-graph-convolutional-network | 2306.07019 | null | https://arxiv.org/abs/2306.07019v1 | https://arxiv.org/pdf/2306.07019v1.pdf | Dynamic Causal Graph Convolutional Network for Traffic Prediction | Modeling complex spatiotemporal dependencies in correlated traffic series is essential for traffic prediction. While recent works have shown improved prediction performance by using neural networks to extract spatiotemporal correlations, their effectiveness depends on the quality of the graph structures used to represent the spatial topology of the traffic network. In this work, we propose a novel approach for traffic prediction that embeds time-varying dynamic Bayesian network to capture the fine spatiotemporal topology of traffic data. We then use graph convolutional networks to generate traffic forecasts. To enable our method to efficiently model nonlinear traffic propagation patterns, we develop a deep learning-based module as a hyper-network to generate stepwise dynamic causal graphs. Our experimental results on a real traffic dataset demonstrate the superior prediction performance of the proposed method. | ['Chen Zhang', 'Rui Zhao', 'Lei Bai', 'Zhishuai Li', 'Ziyue Li', 'Junpeng Lin'] | 2023-06-12 | null | null | null | null | ['traffic-prediction'] | ['time-series'] | [-1.92936331e-01 -1.87145397e-01 -2.53443211e-01 -3.74705702e-01
-1.86332483e-02 -9.51969549e-02 5.97025335e-01 -4.17296141e-01
3.58376026e-01 8.34430635e-01 1.95082158e-01 -8.52767289e-01
-3.93068552e-01 -1.16000426e+00 -7.26977289e-01 -1.91635087e-01
-4.28307593e-01 3.88964653e-01 8.39480758e-01 -3.63946348e-01
1.25515357e-01 6.80707097e-01 -1.19336653e+00 2.53595650e-01
1.07319474e+00 8.92064154e-01 3.67844999e-02 8.46721053e-01
-2.11433277e-01 1.38992441e+00 -4.96355653e-01 -4.90323782e-01
1.46473765e-01 -2.19436541e-01 -3.32981288e-01 -7.48685673e-02
7.31234923e-02 -1.29908189e-01 -1.22166622e+00 5.64806700e-01
-4.48439009e-02 9.84097868e-02 6.30188823e-01 -1.45491076e+00
-4.86141354e-01 8.03676903e-01 -3.67194802e-01 8.85767460e-01
-1.37715489e-01 5.52140772e-01 8.98870587e-01 -1.81241512e-01
4.45704997e-01 1.28337824e+00 7.33678043e-01 1.44292548e-01
-1.21362221e+00 -9.49094594e-01 3.90815169e-01 7.35466421e-01
-1.25373292e+00 -3.67248684e-01 1.33130634e+00 -5.44126451e-01
9.04726624e-01 4.78155650e-02 7.83421159e-01 1.17373824e+00
5.57379127e-01 4.40065771e-01 8.87566149e-01 2.92571068e-01
-5.05061522e-02 -2.64540464e-01 2.72292465e-01 7.29665756e-01
3.29578221e-02 4.04709756e-01 -3.74985039e-01 4.11087051e-02
7.12930679e-01 3.05080805e-02 1.79480880e-01 5.30686826e-02
-7.50920355e-01 5.09413481e-01 8.39512706e-01 3.43903564e-02
-5.60452700e-01 6.12730980e-01 1.93260133e-01 1.12556957e-01
5.87153375e-01 -9.08434242e-02 -1.20419465e-01 -1.42496139e-01
-8.52516413e-01 1.87815711e-01 7.34591365e-01 9.09041286e-01
7.04301298e-01 4.84346479e-01 -3.14488828e-01 5.29212713e-01
2.73723751e-01 6.24985993e-01 -3.06382954e-01 -7.61139154e-01
7.21185803e-01 5.67035317e-01 -1.51194200e-01 -1.65717328e+00
-5.00618935e-01 -5.39662004e-01 -9.30638850e-01 -2.51334161e-01
4.51158762e-01 -8.69647115e-02 -7.95124292e-01 1.54479182e+00
8.65081623e-02 1.16806960e+00 -2.78496534e-01 7.75785029e-01
4.22150552e-01 8.77038717e-01 3.32317770e-01 8.72739702e-02
7.64770925e-01 -8.15960884e-01 -6.84880495e-01 2.72900999e-01
4.55431163e-01 -2.39018977e-01 6.52880609e-01 -1.14389695e-01
-7.98192322e-01 -6.37887359e-01 -6.88773692e-01 4.01100844e-01
-4.67662930e-01 -2.88279444e-01 5.75277269e-01 5.31225741e-01
-8.87105644e-01 5.02280891e-01 -8.64714265e-01 -6.83421269e-02
8.12924623e-01 2.27027029e-01 2.41374627e-01 1.06922790e-01
-1.51339686e+00 6.51771903e-01 4.84786302e-01 4.30989116e-01
-7.83657312e-01 -9.08870935e-01 -5.28296053e-01 3.45910251e-01
3.62341195e-01 -7.19635665e-01 7.08542287e-01 -4.14010674e-01
-1.47685587e+00 -1.28308550e-01 -2.22290978e-01 -7.46435106e-01
4.80250955e-01 3.20157439e-01 -1.03644824e+00 9.45553333e-02
-5.65832071e-02 1.53116539e-01 6.58260703e-01 -9.84468937e-01
-7.88430393e-01 1.48858562e-01 5.76295182e-02 -4.34238493e-01
-1.91603228e-01 -1.43653661e-01 -4.83368665e-01 -7.94367313e-01
-4.17038172e-01 -9.61439371e-01 -4.23364937e-01 -4.66772586e-01
-5.49969077e-01 -4.90021586e-01 1.21237564e+00 -6.95994735e-01
1.72754908e+00 -1.76161134e+00 -3.70340168e-01 6.62784636e-01
4.78442401e-01 2.52004027e-01 -1.44618720e-01 4.13267821e-01
8.97983536e-02 2.31068388e-01 -9.50107276e-02 3.88768800e-02
1.41343186e-02 2.89869487e-01 -6.91811860e-01 -3.55720185e-02
4.40878779e-01 1.35514128e+00 -8.86302352e-01 -5.39989710e-01
3.32238615e-01 6.56480789e-01 -5.21435559e-01 1.49411470e-01
-4.34793800e-01 6.44748032e-01 -6.23691499e-01 1.78472281e-01
7.12967634e-01 -4.87318367e-01 2.22051904e-01 -2.90108770e-01
5.19572832e-02 4.19196784e-01 -8.36810172e-01 9.08048630e-01
-6.28668487e-01 8.34383309e-01 -4.93625790e-01 -1.10734677e+00
1.03369844e+00 5.67316562e-02 5.42632341e-01 -1.06686282e+00
-1.33769467e-01 -2.38741413e-01 1.41353890e-01 -4.56240922e-01
3.60556185e-01 1.54575661e-01 9.13775712e-02 3.58457536e-01
-3.58012706e-01 4.49252874e-01 4.10055369e-01 2.79058754e-01
1.38555658e+00 -7.38179833e-02 -5.34619451e-01 6.80791622e-04
7.00736284e-01 9.13166702e-02 7.34955490e-01 6.32776260e-01
-1.11860171e-01 4.74388972e-02 9.22591865e-01 -8.03340733e-01
-9.80929852e-01 -1.23936760e+00 9.02187899e-02 7.88619339e-01
6.41197488e-02 -4.47615117e-01 -3.06863278e-01 -9.22084212e-01
4.07354627e-03 7.51990736e-01 -5.35301745e-01 -1.64353773e-01
-1.22748685e+00 -8.15697730e-01 5.42356849e-01 6.47623062e-01
5.12928605e-01 -1.01019895e+00 8.75222459e-02 4.40738410e-01
-1.98217884e-01 -1.55953324e+00 -3.18039626e-01 -4.54180926e-01
-6.87696457e-01 -1.15148079e+00 5.49569540e-02 -4.15301174e-01
4.39506412e-01 2.20238954e-01 1.08013284e+00 2.03934699e-01
1.14330426e-02 8.24203044e-02 1.03097875e-02 -7.59200752e-02
-5.40120661e-01 2.91783124e-01 -2.31320843e-01 4.09796298e-01
1.88933253e-01 -1.08998990e+00 -6.97968185e-01 5.57583094e-01
-6.69510961e-01 1.64180487e-01 4.44602132e-01 4.91769403e-01
2.68118471e-01 5.05578697e-01 6.42400324e-01 -8.57796788e-01
7.10588098e-01 -8.96317363e-01 -7.74900556e-01 2.11343408e-01
-5.81240475e-01 1.78853825e-01 9.19543743e-01 -3.32170308e-01
-1.19350517e+00 -2.41269931e-01 9.29914229e-03 -6.11965418e-01
-1.36269465e-01 7.93214023e-01 1.38206080e-01 3.07192415e-01
3.00919890e-01 2.45872870e-01 -2.74493396e-01 -2.23154202e-01
3.36086005e-01 2.84270078e-01 3.56283277e-01 -5.65649450e-01
1.08589101e+00 5.35941541e-01 5.20362616e-01 -7.18044698e-01
-5.94612658e-01 1.68331694e-02 -7.37213314e-01 -6.71503127e-01
7.99890399e-01 -6.59297168e-01 -1.02920055e+00 1.67810574e-01
-1.17690301e+00 -5.53155899e-01 1.41417146e-01 3.82874489e-01
-4.34766024e-01 1.69079542e-01 -5.86995721e-01 -6.76972151e-01
6.23538950e-03 -8.98588717e-01 6.72352314e-01 -1.01163043e-02
1.45731360e-01 -1.38229799e+00 1.99017569e-01 3.19900155e-01
6.43251479e-01 3.54664207e-01 1.16911697e+00 -2.71615714e-01
-1.31330836e+00 -1.68408439e-01 -6.01508379e-01 -1.40915498e-01
1.31143302e-01 4.69091237e-01 -4.50227559e-01 3.67567509e-01
-7.10627317e-01 4.67517704e-01 9.75620627e-01 4.63330418e-01
1.42458129e+00 -3.89648288e-01 -7.00104833e-01 6.70586884e-01
9.96814787e-01 1.71434239e-01 7.72809088e-01 -4.42410121e-03
1.16207969e+00 7.27587640e-01 1.58775464e-01 1.42487481e-01
9.60686028e-01 7.55203664e-01 3.02751631e-01 9.09528434e-02
-4.60938454e-01 -6.58248782e-01 2.50907660e-01 1.11212265e+00
-1.80831730e-01 -4.04586643e-01 -1.37327111e+00 3.88755441e-01
-2.14814782e+00 -1.43991280e+00 -7.57991850e-01 1.59742105e+00
1.18809529e-01 6.18222356e-01 4.55662608e-01 -1.70470253e-01
8.10568273e-01 2.91936874e-01 -4.33795989e-01 -2.40289405e-01
-1.00117601e-01 8.84163566e-03 5.29232621e-01 5.25618851e-01
-8.76698732e-01 1.23626029e+00 6.81482220e+00 8.04516435e-01
-1.21043432e+00 -3.59607190e-02 6.61125600e-01 9.77274850e-02
-3.77108723e-01 5.80472685e-02 -6.87040329e-01 8.67852509e-01
1.46946502e+00 -1.94995776e-01 4.25085664e-01 4.01182890e-01
7.60152221e-01 3.85435164e-01 -6.30376399e-01 6.32509947e-01
-5.59093118e-01 -1.86561918e+00 3.24938744e-01 1.77350864e-01
7.29015708e-01 1.69916779e-01 2.21457437e-01 4.09450591e-01
7.68228173e-01 -9.04478192e-01 2.52199471e-01 1.14968741e+00
3.11087847e-01 -9.09948468e-01 3.24957550e-01 2.35200956e-01
-1.70376492e+00 -2.18177065e-01 -1.88642576e-01 -2.32370161e-02
5.58077037e-01 7.23780513e-01 -1.02040911e+00 5.53850591e-01
5.38722336e-01 1.22406387e+00 -7.79409647e-01 8.66239071e-01
-2.32840285e-01 1.25420368e+00 -2.67610431e-01 -5.11387885e-02
1.52563289e-01 -3.33827227e-01 5.85650384e-01 1.39731884e+00
1.17471881e-01 -2.54467856e-02 1.20784417e-01 1.21005964e+00
1.50216706e-02 -2.73854047e-01 -8.41299713e-01 -3.63726430e-02
4.81254905e-01 1.01010239e+00 -6.91424847e-01 -3.33107650e-01
-5.03476679e-01 2.50860691e-01 5.24016678e-01 7.26878107e-01
-1.41475689e+00 -2.52760053e-01 6.98394537e-01 5.94635963e-01
5.25794744e-01 -7.68820047e-01 -3.25916976e-01 -1.05580413e+00
7.14546442e-03 -2.64541000e-01 4.42067802e-01 -7.19140649e-01
-1.50438166e+00 6.57126009e-01 7.11905211e-02 -1.22153318e+00
-1.27547666e-01 -3.27589750e-01 -1.13654876e+00 7.80912936e-01
-1.53449929e+00 -1.43332851e+00 -1.94536895e-01 6.81568086e-01
2.53806025e-01 -2.87781030e-01 2.65437309e-02 3.07191521e-01
-1.05703342e+00 2.35819042e-01 -3.33010256e-01 2.23361865e-01
1.00738592e-01 -8.66157949e-01 8.24761271e-01 9.19684649e-01
1.84327178e-02 4.55609560e-01 4.32241410e-01 -1.11396337e+00
-1.23498893e+00 -1.59933305e+00 8.03728163e-01 -6.82686150e-01
1.39053559e+00 -3.42325509e-01 -9.14055288e-01 6.99292362e-01
1.65835824e-02 2.11535737e-01 4.18205678e-01 1.61762074e-01
-3.84308815e-01 -6.44551337e-01 -5.40313482e-01 7.34606862e-01
1.34184611e+00 -5.24323046e-01 -1.54081032e-01 6.34571090e-02
7.30273008e-01 2.17807978e-01 -9.19089139e-01 5.38014352e-01
5.05151212e-01 -7.73021638e-01 9.07724440e-01 -7.43671715e-01
3.68397713e-01 -5.74022293e-01 1.26798257e-01 -1.28420663e+00
-6.41871214e-01 -6.31655276e-01 -6.64778411e-01 1.12332273e+00
5.35790145e-01 -7.00174868e-01 9.93098617e-01 5.58281481e-01
-2.77290214e-02 -5.31177282e-01 -8.23202908e-01 -6.37630820e-01
-1.13110304e-01 -9.86684084e-01 1.05794144e+00 7.90723741e-01
-2.60405749e-01 4.34303015e-01 -4.98937428e-01 4.27716881e-01
7.11481929e-01 6.56690449e-02 9.52432215e-01 -1.28353775e+00
2.65487339e-02 -7.61915624e-01 -6.29649222e-01 -1.09297121e+00
6.18699253e-01 -8.40753376e-01 -4.27337766e-01 -1.32369316e+00
-1.03728056e-01 -7.14041114e-01 -5.41024864e-01 -6.39002174e-02
-2.68570840e-01 1.01741694e-01 6.14517741e-02 1.15872003e-01
-7.72819519e-01 7.12319911e-01 1.20014906e+00 -2.10355207e-01
-2.09599316e-01 2.19127208e-01 -2.15796426e-01 3.67125630e-01
1.11679614e+00 -5.38721085e-01 -6.59660518e-01 -3.20777476e-01
4.23377663e-01 2.73328036e-01 6.05712712e-01 -1.06731868e+00
5.29186368e-01 -3.82508874e-01 2.06198052e-01 -1.09205675e+00
9.20462832e-02 -9.08711553e-01 3.37252736e-01 1.71849266e-01
-2.88476229e-01 2.19637722e-01 2.04100177e-01 1.04677415e+00
-1.53436735e-01 7.62409806e-01 2.71394610e-01 3.05777222e-01
-6.47412837e-01 7.17062652e-01 -7.75755703e-01 -2.15388685e-01
9.51893210e-01 -8.37831050e-02 -4.38723624e-01 -5.24649501e-01
-8.15237701e-01 4.53736871e-01 -5.39195314e-02 6.48430645e-01
5.78602731e-01 -1.56696570e+00 -5.13263822e-01 1.39765367e-01
-6.23540208e-02 -5.20784378e-01 3.46167296e-01 1.00896132e+00
-4.81010973e-01 6.19436562e-01 -1.38493478e-01 -6.86090648e-01
-6.39880598e-01 6.93547487e-01 3.58401984e-01 -5.68674624e-01
-7.07103014e-01 2.06888855e-01 1.84708666e-02 -3.41759086e-01
-1.17428169e-01 -3.70456904e-01 -2.74183929e-01 -3.06794226e-01
2.09921405e-01 8.01631331e-01 -1.77661344e-01 -6.87768817e-01
-2.16805831e-01 4.45913643e-01 2.84657985e-01 2.07779288e-01
1.31838083e+00 -3.46348614e-01 -6.23652618e-03 4.02507186e-01
9.63180661e-01 -2.24198774e-01 -1.54925394e+00 -3.00489962e-01
2.06587747e-01 -3.09150338e-01 3.45558859e-02 -5.17123759e-01
-1.43736124e+00 9.42068100e-01 3.14691037e-01 5.36347687e-01
9.53378141e-01 -2.46398166e-01 1.10658431e+00 4.46947187e-01
2.29048222e-01 -8.02213371e-01 -4.20683064e-03 7.76098609e-01
5.54137826e-01 -1.03868222e+00 -3.16777647e-01 -6.84844494e-01
-5.32134473e-01 1.33671486e+00 7.25589931e-01 -3.51267099e-01
1.38444996e+00 2.03514844e-01 -2.08598629e-01 -4.80451494e-01
-1.17781508e+00 -4.75493848e-01 4.88425165e-01 6.28089070e-01
4.32341211e-02 1.70535177e-01 1.66158706e-01 3.59514058e-01
-2.51489341e-01 1.20878376e-01 3.07629257e-01 3.50301951e-01
-2.02314541e-01 -7.68672645e-01 -3.89251811e-03 5.61003149e-01
-3.05070058e-02 1.05561778e-01 -2.11620912e-01 5.87380707e-01
-1.77983195e-01 1.14127898e+00 3.94389361e-01 -9.09019709e-01
3.93380404e-01 -1.49708554e-01 7.59315044e-02 -4.17554602e-02
-3.13410670e-01 -3.30925107e-01 2.36273050e-01 -7.35674322e-01
-2.28554294e-01 -3.38489711e-01 -1.02858436e+00 -8.85672569e-01
1.39980540e-01 4.61621135e-02 3.34723681e-01 8.68760288e-01
6.96357965e-01 8.71469975e-01 9.05994594e-01 -5.70245922e-01
3.77538472e-01 -7.11721420e-01 -2.61822313e-01 4.13752198e-01
2.45944813e-01 -8.46123040e-01 -3.03542018e-02 -3.01935337e-03] | [6.459692001342773, 2.056347608566284] |
cbb8dd7f-dcc7-4f13-94bb-62a213e18d3c | what-makes-a-top-performing-precision | 2006.02785 | null | https://arxiv.org/abs/2006.02785v2 | https://arxiv.org/pdf/2006.02785v2.pdf | What Makes a Top-Performing Precision Medicine Search Engine? Tracing Main System Features in a Systematic Way | From 2017 to 2019 the Text REtrieval Conference (TREC) held a challenge task on precision medicine using documents from medical publications (PubMed) and clinical trials. Despite lots of performance measurements carried out in these evaluation campaigns, the scientific community is still pretty unsure about the impact individual system features and their weights have on the overall system performance. In order to overcome this explanatory gap, we first determined optimal feature configurations using the Sequential Model-based Algorithm Configuration (SMAC) program and applied its output to a BM25-based search engine. We then ran an ablation study to systematically assess the individual contributions of relevant system features: BM25 parameters, query type and weighting schema, query expansion, stop word filtering, and keyword boosting. For evaluation, we employed the gold standard data from the three TREC-PM installments to evaluate the effectiveness of different features using the commonly shared infNDCG metric. | ['Udo Hahn', 'Michel Oleynik', 'Erik Faessler'] | 2020-06-04 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [ 3.97970736e-01 -3.34010988e-01 -3.92837554e-01 -1.02436557e-01
-1.32670736e+00 -7.55328894e-01 1.01817942e+00 1.01327538e+00
-1.17789578e+00 5.31007767e-01 5.84356368e-01 -7.24528849e-01
-7.01060951e-01 -2.78287202e-01 -2.84983993e-01 -2.74287790e-01
-2.52631098e-01 6.26251161e-01 3.23958844e-01 -1.34454072e-01
7.64256060e-01 2.73293167e-01 -1.30104744e+00 6.65115893e-01
7.41639197e-01 5.87846696e-01 -6.93787774e-03 7.22855210e-01
-9.70529616e-02 3.25487405e-01 -6.44207418e-01 -2.88437605e-01
-5.35214543e-02 -9.83023122e-02 -8.87247205e-01 -7.13261485e-01
-8.35196003e-02 1.80502340e-01 8.91236812e-02 6.66898251e-01
9.92548466e-01 -3.20018083e-01 7.36984313e-01 -6.17198348e-01
-1.85156927e-01 5.21002293e-01 -3.46180081e-01 5.52476287e-01
5.20203650e-01 2.88222004e-02 9.30714786e-01 -7.74250567e-01
8.90664697e-01 1.02882028e+00 4.47252154e-01 8.64064619e-02
-1.17219687e+00 -2.72035956e-01 -1.77694336e-01 2.12541800e-02
-1.41614401e+00 -4.29142267e-01 -4.89683412e-02 -6.04666352e-01
1.19447923e+00 6.02225006e-01 4.00988877e-01 7.23428607e-01
5.82581758e-01 4.11841005e-01 8.45343113e-01 -9.43210304e-01
3.17593426e-01 2.11092755e-01 5.09825885e-01 2.94431418e-01
4.29869682e-01 1.05283916e-01 -2.42988497e-01 -9.38092530e-01
-1.83946460e-01 -1.85349852e-01 -3.78385335e-01 1.02683708e-01
-1.16407728e+00 8.09782028e-01 1.77521840e-01 6.63007379e-01
-4.43786323e-01 -6.21373812e-03 5.17801404e-01 4.56355035e-01
2.96054631e-01 9.23229337e-01 -5.75283587e-01 -7.71332011e-02
-9.48980689e-01 5.23937225e-01 7.23592639e-01 6.08672202e-01
1.20196164e-01 -9.99430478e-01 -8.24103594e-01 8.08852315e-01
4.24294740e-01 3.98746520e-01 6.84129655e-01 -6.01245403e-01
3.47303510e-01 6.55660331e-01 2.39001393e-01 -8.14772487e-01
-7.41311014e-01 -5.43039680e-01 -3.46076667e-01 -4.58478540e-01
1.15373895e-01 -1.74924150e-01 -1.19881403e+00 1.17712069e+00
3.92524377e-02 -6.40921593e-01 -2.52305955e-01 5.47013640e-01
8.18314433e-01 1.97462857e-01 6.58457160e-01 -2.32827321e-01
1.63575983e+00 -3.03481430e-01 -6.11016929e-01 1.45620003e-01
1.31303477e+00 -1.27614176e+00 6.86351597e-01 1.11120954e-01
-9.31872666e-01 -6.29064441e-02 -1.03214598e+00 1.31843954e-01
-6.53394520e-01 -1.58379003e-01 4.58753437e-01 8.26254666e-01
-9.32632208e-01 5.94268024e-01 -7.70070672e-01 -4.45347518e-01
-4.46070172e-02 2.58832306e-01 2.55209971e-02 -3.09432358e-01
-1.38500237e+00 1.27415788e+00 3.47746879e-01 -2.73676813e-01
-2.80727237e-01 -9.77135420e-01 -2.18806848e-01 1.17878400e-01
1.69110909e-01 -9.34286594e-01 1.06025314e+00 -1.07026167e-01
-6.96651697e-01 6.69020653e-01 -2.82673333e-02 -3.94134432e-01
2.79469281e-01 -6.82775751e-02 -4.93989050e-01 6.02723248e-02
7.70839006e-02 3.67123544e-01 4.10947762e-02 -6.57127321e-01
-7.57835209e-01 -6.42580211e-01 -1.70354322e-01 1.55320823e-01
2.84294342e-03 6.70483053e-01 -7.75736153e-01 -4.63169038e-01
-7.76943490e-02 -1.03184330e+00 -4.41156596e-01 -6.26270831e-01
-1.58837676e-01 -3.47642750e-01 -1.24570146e-01 -5.50797939e-01
1.66824722e+00 -1.90506053e+00 -2.55723268e-01 7.02085376e-01
-6.75508901e-02 3.67927760e-01 -2.83072054e-01 7.31934249e-01
4.28569168e-02 6.58091724e-01 3.62044066e-01 2.09837362e-01
-2.31762305e-01 -3.56958807e-01 9.12017226e-02 3.04154962e-01
-9.75595042e-02 9.60295200e-01 -6.90923572e-01 -7.78181493e-01
9.44761336e-02 5.66539347e-01 -5.39286375e-01 -5.71548283e-01
-1.11558959e-01 -1.15756214e-01 -7.15538800e-01 6.94920659e-01
2.32767820e-01 -5.64891517e-01 3.80315155e-01 -2.09302790e-02
-2.73698904e-02 3.09574336e-01 -5.83433032e-01 1.55683935e+00
-1.30845383e-01 1.94417104e-01 -5.00670493e-01 -4.36444968e-01
3.41529042e-01 5.11240780e-01 7.69780755e-01 -1.03714085e+00
1.94048434e-01 3.83349478e-01 2.04369634e-01 -4.09181267e-01
4.22036588e-01 2.23215118e-01 -6.39013052e-02 3.47772747e-01
-1.67366847e-01 1.57586545e-01 2.79172182e-01 4.03739333e-01
1.48222744e+00 -5.04907429e-01 3.02132368e-01 -5.75072348e-01
3.02402854e-01 4.15284246e-01 2.33007729e-01 1.03708398e+00
-3.16923708e-02 6.90524101e-01 2.68346280e-01 -2.68965244e-01
-6.74454689e-01 -6.00832939e-01 -6.07258260e-01 8.76684070e-01
-3.21008593e-01 -1.03585482e+00 -4.36616957e-01 -4.37603444e-01
1.31442040e-01 7.30383873e-01 -7.53529847e-01 -2.93524742e-01
-2.57346183e-01 -1.19302857e+00 5.99852741e-01 -7.86335245e-02
-2.06013352e-01 -8.35775435e-01 -8.26703072e-01 2.16801926e-01
-3.45813334e-02 -8.11577201e-01 -6.08402848e-01 4.01538461e-02
-7.81866550e-01 -1.31984603e+00 -9.72064435e-01 -3.56712580e-01
2.59942055e-01 3.86050455e-02 1.24667096e+00 2.82496721e-01
-6.61135435e-01 7.08873153e-01 -5.11577189e-01 -8.00922155e-01
-3.43747258e-01 4.53251749e-01 -4.63561475e-01 -7.13174284e-01
5.98681390e-01 8.54426622e-02 -1.11653149e+00 3.87401208e-02
-1.01105905e+00 -4.32204664e-01 8.79951656e-01 5.87489307e-01
4.90864158e-01 -2.19676390e-01 4.38872457e-01 -1.12467396e+00
1.18684542e+00 -4.07306701e-01 -4.50185627e-01 8.81317794e-01
-1.36374056e+00 2.02870399e-01 -2.59393543e-01 -8.09817240e-02
-6.37856007e-01 -1.19787976e-01 -6.88896701e-02 3.93752605e-01
3.85376006e-01 1.18913066e+00 3.48475903e-01 3.05746961e-02
1.00527537e+00 -7.81133547e-02 -1.07576035e-01 -5.97644031e-01
1.05641037e-01 5.64340115e-01 -6.00518361e-02 -4.72817689e-01
2.78105259e-01 -5.13208322e-02 -1.09320302e-02 -6.80835068e-01
-4.36073124e-01 -8.06222022e-01 2.42988840e-02 1.27397537e-01
6.64894402e-01 -8.23081255e-01 -7.01289594e-01 -1.16396107e-01
-1.02432704e+00 1.61226928e-01 9.53481719e-02 5.91974497e-01
5.50956950e-02 1.44576699e-01 -4.65149164e-01 -5.12480438e-01
-8.44813943e-01 -1.12751758e+00 1.03061748e+00 -5.93350548e-03
-5.12138665e-01 -8.15389156e-01 4.29058969e-01 3.11614484e-01
7.36187696e-01 4.92520630e-02 1.43425107e+00 -8.86022747e-01
-8.13125167e-03 -8.37654710e-01 -1.05253242e-01 -2.51314551e-01
8.18731338e-02 3.15768123e-02 -5.88034272e-01 -3.20739478e-01
-4.35750157e-01 -5.03944196e-02 9.15095329e-01 6.55139446e-01
1.04435623e+00 1.29867658e-01 -7.24705756e-01 2.63511296e-02
1.52034891e+00 3.89218688e-01 5.69612384e-01 6.68114841e-01
-7.93142766e-02 4.41896409e-01 4.87058282e-01 1.37374535e-01
9.65219960e-02 7.93227077e-01 -2.53236234e-01 1.46244183e-01
1.12883992e-01 -1.53779447e-01 -2.27313057e-01 1.44003093e-01
-9.21413824e-02 -3.72397035e-01 -1.26733279e+00 3.10133338e-01
-1.33689499e+00 -4.86315370e-01 2.44360209e-01 2.44979811e+00
9.08480763e-01 5.54087043e-01 8.21118709e-03 -4.24287915e-02
3.16355228e-01 -3.02898258e-01 -1.30897731e-01 -2.20832095e-01
-1.61948223e-02 3.88977706e-01 8.19342554e-01 2.58890033e-01
-7.76115060e-01 3.76303583e-01 7.40985727e+00 9.61486340e-01
-9.42072213e-01 -2.54971117e-01 6.16955221e-01 -2.64935881e-01
-4.35049832e-01 9.19295102e-02 -7.59417832e-01 4.90201443e-01
1.38670337e+00 -6.25525713e-01 -2.29397863e-02 3.32310796e-01
2.08407208e-01 -4.41183776e-01 -9.83014643e-01 5.24282336e-01
-1.53237283e-01 -1.54527569e+00 8.05333629e-02 1.60507604e-01
4.40429777e-01 4.28000152e-01 9.28050876e-02 4.07657653e-01
3.29778880e-01 -9.63378429e-01 1.22892126e-01 5.77871799e-01
6.93887949e-01 -2.81048536e-01 9.60799754e-01 1.23247221e-01
-2.86524087e-01 1.12618797e-01 -7.19703734e-02 5.79998016e-01
-1.35605469e-01 6.97879553e-01 -1.20455170e+00 6.93520486e-01
5.55201352e-01 -3.16366404e-02 -1.08470523e+00 1.37936997e+00
5.36875486e-01 7.13526487e-01 -2.82706797e-01 -3.29857707e-01
1.76804796e-01 5.54942310e-01 3.59765530e-01 1.40896571e+00
1.60595819e-01 3.80965918e-02 -3.10337484e-01 2.37717345e-01
1.04761198e-02 5.24813771e-01 -1.84592262e-01 -3.90677869e-01
6.72389567e-01 1.05731153e+00 -9.28833187e-01 -2.43588731e-01
-2.55494207e-01 2.71060228e-01 -2.86482751e-01 4.32014495e-01
-2.92990595e-01 -3.48984569e-01 7.03017041e-02 2.80295581e-01
-6.30045682e-02 2.82781184e-01 -1.58129454e-01 -6.96395814e-01
-4.73614037e-02 -1.13522398e+00 7.88119793e-01 -3.96086574e-01
-9.33872581e-01 5.23317933e-01 1.74487621e-01 -8.37205052e-01
-3.22937310e-01 -3.35053682e-01 -2.22538933e-01 1.08818924e+00
-1.17901945e+00 -4.67914611e-01 3.88602197e-01 1.55657113e-01
1.68607812e-02 -2.51093775e-01 1.03592074e+00 4.58743572e-01
-3.23347121e-01 7.78855145e-01 6.27913773e-01 -1.02450311e-01
9.75517392e-01 -1.02588987e+00 1.74542308e-01 1.25627592e-01
-1.66943118e-01 1.26492059e+00 8.68177235e-01 -6.81832969e-01
-1.34672940e+00 -6.80467606e-01 1.28077602e+00 -7.37959564e-01
4.41984415e-01 1.10286139e-01 -4.93269503e-01 1.68521076e-01
7.88626224e-02 -5.34047067e-01 9.27001119e-01 5.32410324e-01
-1.60039648e-01 7.45432870e-03 -8.96373391e-01 4.71833438e-01
4.19904858e-01 -2.56407350e-01 -4.02732760e-01 3.63291293e-01
7.61368573e-01 -7.14255795e-02 -1.13280582e+00 8.34777653e-01
9.31317270e-01 -3.76523674e-01 1.15438163e+00 -9.93425071e-01
2.56951243e-01 3.84834334e-02 -2.38346890e-01 -1.08072078e+00
-3.26420486e-01 -2.72852778e-01 4.90954757e-01 8.42834115e-01
1.07890475e+00 -5.43882906e-01 5.33231974e-01 1.05563176e+00
1.13755815e-01 -9.44459021e-01 -6.77884221e-01 -4.27498817e-01
4.76549149e-01 -9.75016057e-02 4.58992809e-01 4.63064313e-01
1.39320597e-01 4.74139661e-01 2.14318678e-01 -2.80365795e-01
2.58316427e-01 -2.70696394e-02 3.01404774e-01 -1.07578695e+00
-2.30346411e-01 -5.58459699e-01 -2.50549138e-01 -1.93319649e-01
-6.68810546e-01 -9.92407978e-01 -1.50388911e-01 -1.51444471e+00
6.16929948e-01 -4.41580355e-01 -9.16540623e-01 2.20117956e-01
-5.09858072e-01 -1.88378215e-01 8.69955122e-02 2.68102676e-01
-7.74429381e-01 -1.11632384e-01 8.84180427e-01 -2.61206239e-01
-1.52602375e-01 1.50354162e-01 -9.96851325e-01 9.04054493e-02
4.17826682e-01 -6.95414722e-01 -3.54065895e-01 -3.88948381e-01
6.24471188e-01 2.98414230e-01 4.75444794e-02 -6.01304114e-01
4.35789764e-01 3.46035883e-02 4.68945652e-01 -6.83457673e-01
-1.03421032e-01 -5.89467645e-01 2.82796800e-01 8.68834794e-01
-9.04242456e-01 3.68647516e-01 3.13959897e-01 4.50364172e-01
-1.06128216e-01 -3.71517897e-01 2.40844399e-01 2.65212916e-02
-1.49211988e-01 -9.75778922e-02 -4.03579533e-01 1.08930767e-01
4.97442037e-01 2.04582945e-01 -3.96050930e-01 -1.46936983e-01
-5.20067453e-01 6.00215137e-01 2.34122559e-01 3.62744659e-01
2.35137209e-01 -9.08161104e-01 -9.46127832e-01 -2.34782487e-01
4.44218874e-01 -6.41126990e-01 2.69092143e-01 9.00824726e-01
-3.83181006e-01 1.18891394e+00 1.95927188e-01 -4.69808847e-01
-1.43433881e+00 4.60338682e-01 8.38483199e-02 -7.87986398e-01
-2.51784384e-01 4.62199301e-01 -3.60025704e-01 -1.70476153e-01
3.38992685e-01 -2.71741915e-02 -1.67562574e-01 5.02134003e-02
6.99078441e-01 2.57420242e-01 8.68176997e-01 1.64723888e-01
-5.67640841e-01 3.31254423e-01 -4.55822557e-01 -4.10682350e-01
1.16664612e+00 6.56677485e-02 1.90247625e-01 9.86811742e-02
1.29672277e+00 -2.30752081e-01 1.34076953e-01 -3.74034107e-01
7.27030933e-01 -1.35053769e-01 2.43001387e-01 -1.44062817e+00
-4.40452576e-01 3.12744707e-01 9.48803604e-01 4.70497161e-02
1.04966700e+00 -1.48495823e-01 2.23907381e-01 4.68371242e-01
2.80804098e-01 -1.11213160e+00 -4.88398910e-01 2.95385331e-01
7.01179326e-01 -9.94887650e-01 4.92173404e-01 6.28639981e-02
-3.66001159e-01 7.22115040e-01 -2.34778851e-01 3.98276538e-01
1.01797831e+00 -1.07277177e-01 1.25970587e-01 -6.36526585e-01
-1.04680431e+00 1.06134444e-01 6.53817594e-01 1.49748977e-02
9.89066780e-01 -1.63910523e-01 -1.20820034e+00 4.46511954e-01
-9.65804234e-02 4.39161032e-01 -5.19301407e-02 1.03499544e+00
-3.08910012e-01 -1.40216899e+00 -6.97061792e-02 9.51504230e-01
-1.09518921e+00 -2.57383913e-01 -4.45664048e-01 8.75146389e-01
-2.48200625e-01 1.09601736e+00 -4.64765966e-01 -3.26227456e-01
4.44696158e-01 2.26885647e-01 4.39728647e-01 -4.81716216e-01
-8.44072461e-01 2.94830650e-01 3.48156869e-01 -4.67333764e-01
-3.49260122e-01 -8.61527503e-01 -8.13179195e-01 -1.74622741e-02
-5.24385393e-01 7.44487226e-01 1.00788617e+00 8.45795572e-01
6.54027820e-01 4.72227931e-01 2.72826850e-01 6.17256165e-02
-9.30484891e-01 -1.25264740e+00 -1.65095508e-01 2.77654678e-01
-6.56063296e-03 -4.22878504e-01 -2.47204438e-01 -3.66891086e-01] | [8.806992530822754, 8.687763214111328] |
aa2650f3-175d-4d41-a84b-7cdc77dcf858 | soft-anchor-point-object-detection | 1911.12448 | null | https://arxiv.org/abs/1911.12448v2 | https://arxiv.org/pdf/1911.12448v2.pdf | Soft Anchor-Point Object Detection | Recently, anchor-free detection methods have been through great progress. The major two families, anchor-point detection and key-point detection, are at opposite edges of the speed-accuracy trade-off, with anchor-point detectors having the speed advantage. In this work, we boost the performance of the anchor-point detector over the key-point counterparts while maintaining the speed advantage. To achieve this, we formulate the detection problem from the anchor point's perspective and identify ineffective training as the main problem. Our key insight is that anchor points should be optimized jointly as a group both within and across feature pyramid levels. We propose a simple yet effective training strategy with soft-weighted anchor points and soft-selected pyramid levels to address the false attention issue within each pyramid level and the feature selection issue across all the pyramid levels, respectively. To evaluate the effectiveness, we train a single-stage anchor-free detector called Soft Anchor-Point Detector (SAPD). Experiments show that our concise SAPD pushes the envelope of speed/accuracy trade-off to a new level, outperforming recent state-of-the-art anchor-free and anchor-based detectors. Without bells and whistles, our best model can achieve a single-model single-scale AP of 47.4% on COCO. | ['Zhiqiang Shen', 'Fangyi Chen', 'Marios Savvides', 'Chenchen Zhu'] | 2019-11-27 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/721_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123540086.pdf | eccv-2020-8 | ['dense-object-detection'] | ['computer-vision'] | [-1.05809897e-01 -2.13766754e-01 -2.44201973e-01 -5.50616905e-02
-1.40737712e+00 -4.06000286e-01 2.73049682e-01 4.61789995e-01
-3.63352001e-01 -8.49722326e-02 -2.54479665e-02 -3.01334430e-02
8.24417844e-02 -4.13845062e-01 -7.08959937e-01 -7.05872893e-01
-3.76162171e-01 -1.02323875e-01 1.05161297e+00 -2.19619662e-01
3.38030636e-01 5.37379980e-01 -1.20831203e+00 4.57690619e-02
7.29044557e-01 1.29848731e+00 2.18145475e-01 6.85741186e-01
2.39612505e-01 1.81182057e-01 -6.78379834e-01 -3.19332957e-01
5.35003424e-01 1.63766444e-01 3.92750725e-02 -1.96652621e-01
6.76756561e-01 -4.08228904e-01 -3.40306163e-01 1.08492923e+00
7.79950440e-01 -2.83884525e-01 5.31583190e-01 -1.36227310e+00
-5.96810222e-01 2.58344024e-01 -1.30542064e+00 4.15318072e-01
1.91013470e-01 -4.78946045e-02 1.42676246e+00 -1.67256987e+00
6.61767200e-02 1.07451713e+00 1.10672450e+00 -1.27079617e-02
-9.78844583e-01 -7.25520670e-01 3.24355841e-01 2.60514647e-01
-1.75359547e+00 -5.18162966e-01 5.90850830e-01 -1.91147313e-01
7.66401768e-01 3.71441580e-02 4.15637106e-01 6.38994396e-01
2.75776893e-01 1.13070893e+00 5.49145341e-01 -4.92608130e-01
2.52514780e-01 -1.05764121e-01 1.96516946e-01 7.61320472e-01
5.24531007e-01 4.72790934e-02 -8.51417363e-01 -3.80281448e-01
7.43093431e-01 1.08127467e-01 -2.95870125e-01 -6.63071930e-01
-1.15043104e+00 8.65980506e-01 8.22426617e-01 1.27769664e-01
-1.92596748e-01 1.86160833e-01 2.46357918e-01 7.25382492e-02
2.18454331e-01 3.23976636e-01 -3.98182899e-01 1.60040379e-01
-1.13672018e+00 9.50247124e-02 4.59253252e-01 1.12907612e+00
4.96736228e-01 -2.63874352e-01 -3.24034691e-01 6.66315436e-01
5.22825956e-01 5.01173854e-01 2.35170901e-01 -5.60652614e-01
6.57491982e-01 7.77401567e-01 3.47746551e-01 -1.31370652e+00
-4.21041995e-01 -7.43836820e-01 -6.08114421e-01 1.87742800e-01
4.52876449e-01 -5.90590127e-02 -9.16684926e-01 1.45853376e+00
5.89670479e-01 2.87406772e-01 -1.65495545e-01 8.30575228e-01
3.02092940e-01 6.78646028e-01 -1.02101892e-01 -6.67154193e-02
1.76204872e+00 -1.09020674e+00 -3.70504588e-01 -4.96881098e-01
7.54694998e-01 -7.61666954e-01 9.34354186e-01 2.30295822e-01
-9.32087958e-01 -5.17974138e-01 -1.63410008e+00 -4.78969254e-02
-1.28507763e-01 4.12296057e-01 3.85337055e-01 5.27658522e-01
-9.02038276e-01 3.05476636e-01 -9.74250913e-01 -1.92478850e-01
3.31211567e-01 4.27344173e-01 -2.77639955e-01 1.00498900e-01
-9.57632899e-01 8.18777204e-01 1.15466602e-01 -3.53855863e-02
-4.73458439e-01 -7.17010379e-01 -6.91135764e-01 3.98062468e-01
6.11488461e-01 -4.05909956e-01 1.29038954e+00 -3.66407782e-01
-1.03919971e+00 5.65601707e-01 -2.32701316e-01 -6.79139853e-01
3.49157363e-01 -5.68706036e-01 -3.49846482e-01 2.40379989e-01
4.50803429e-01 5.47492564e-01 1.02889740e+00 -1.18143928e+00
-1.36647689e+00 -3.87453914e-01 -1.95179150e-01 1.83565468e-01
-3.62362921e-01 -4.90540564e-02 -9.06817794e-01 -8.01020563e-01
5.46368003e-01 -8.78055394e-01 -1.05515294e-01 4.51850861e-01
-3.74577045e-01 -4.53503132e-01 7.81954408e-01 -3.39753687e-01
1.44260228e+00 -2.48156190e+00 -2.61223614e-01 2.16450408e-01
3.96841615e-01 3.89463454e-01 -1.22831821e-01 3.55393022e-01
5.38948067e-02 -7.27415755e-02 1.34274229e-01 -6.14973903e-01
7.74857476e-02 -1.11059256e-01 -1.83259368e-01 7.48037696e-01
3.82635713e-01 7.36947536e-01 -9.27543044e-01 -6.02315366e-01
-1.75140426e-02 3.73204678e-01 -6.75729334e-01 -1.27947494e-01
4.14757609e-01 -1.10373043e-01 -5.89590609e-01 1.06235158e+00
7.79611647e-01 -3.72358233e-01 -3.47784787e-01 -5.56599677e-01
-2.41055429e-01 2.87466124e-02 -1.56230927e+00 1.54696274e+00
-7.75358528e-02 4.76028323e-01 3.31660032e-01 -7.01528251e-01
9.06611264e-01 7.73186535e-02 3.49915802e-01 -6.82867289e-01
-1.54939502e-01 5.61189234e-01 -7.11209327e-02 5.60383908e-02
6.95071757e-01 1.11496374e-01 -1.58816189e-01 -2.06878573e-01
-2.83712521e-02 2.99936652e-01 3.07333302e-02 2.01562479e-01
1.28518414e+00 -2.05281496e-01 5.98485470e-01 -2.45426416e-01
3.44990730e-01 -2.00188339e-01 6.73853934e-01 1.12597513e+00
-4.68457371e-01 7.24613965e-01 3.20078671e-01 -1.13309726e-01
-8.04519475e-01 -9.87166703e-01 -1.99524894e-01 1.32924330e+00
4.83500212e-01 -6.89525664e-01 -4.60452050e-01 -8.17706048e-01
2.77734637e-01 2.32294798e-01 -5.60105026e-01 -1.33263335e-01
-7.44124711e-01 -5.94876826e-01 4.02475417e-01 7.69526660e-01
4.44343805e-01 -2.16416344e-01 -9.26751375e-01 1.80601358e-01
6.47800565e-02 -1.17779911e+00 -7.68558323e-01 4.29525644e-01
-5.08580267e-01 -9.56817269e-01 -9.38231587e-01 -8.04761708e-01
5.94409704e-01 8.14737558e-01 9.56514835e-01 6.82554618e-02
-3.27793434e-02 1.65497094e-01 -6.18174076e-01 -5.91560006e-01
8.81280825e-02 1.71730563e-01 2.16144338e-01 7.52096847e-02
2.76344001e-01 -4.17595118e-01 -1.05203509e+00 6.75617516e-01
-5.32376766e-01 -2.01509058e-01 1.08230364e+00 8.61103117e-01
7.79820323e-01 -2.58002818e-01 4.14929271e-01 -1.53277591e-01
1.80621147e-01 -3.13043654e-01 -6.12411916e-01 1.13110110e-01
-5.16476035e-01 -5.50521873e-02 3.59110683e-01 -5.06241500e-01
-2.92255908e-01 3.41228247e-01 -2.88770348e-01 -4.15009737e-01
4.64007139e-01 2.65813708e-01 -2.01386958e-01 -4.89111513e-01
6.86115682e-01 2.36076012e-01 -4.12590802e-01 -4.88153875e-01
1.88181445e-01 6.21700227e-01 6.15859210e-01 -3.16029996e-01
1.16620946e+00 6.27981246e-01 5.42941615e-02 -7.05617666e-01
-1.12846708e+00 -9.80796576e-01 -3.90717536e-01 8.43420625e-02
4.14691895e-01 -1.21671999e+00 -5.34024596e-01 3.32268655e-01
-1.10241592e+00 2.51298040e-01 -2.89648324e-01 1.64937645e-01
-2.03544870e-01 5.69494367e-01 -3.54033560e-01 -7.80160546e-01
-4.97206092e-01 -1.19073808e+00 1.62096357e+00 3.23049635e-01
6.43581301e-02 -3.01762283e-01 -1.23873293e-01 -1.94053173e-01
2.49937996e-01 1.40125886e-01 2.37705469e-01 -9.32509303e-01
-5.82017899e-01 -8.09136808e-01 -5.81602037e-01 6.73179924e-02
-4.95207198e-02 -2.67939925e-01 -7.62465119e-01 -5.66750944e-01
1.09239951e-01 2.49526165e-02 9.40088689e-01 2.54307032e-01
6.90655053e-01 -1.22764632e-01 -7.88191259e-01 6.42023921e-01
1.37619603e+00 4.74838447e-03 3.02784443e-01 4.92983252e-01
5.50890625e-01 -6.57623485e-02 9.03175056e-01 5.17294228e-01
4.97243822e-01 1.14805233e+00 5.24317026e-01 -2.54624158e-01
-4.44349423e-02 -3.42073083e-01 3.29193175e-01 4.28086519e-01
3.93065661e-01 -8.52140412e-02 -9.88232195e-01 5.37787974e-01
-1.98301852e+00 -5.91035903e-01 -7.21348524e-02 2.24716139e+00
5.12843072e-01 6.10540628e-01 2.54798114e-01 3.75687927e-01
6.96153641e-01 1.48255676e-01 -5.39195538e-01 2.55797535e-01
2.83802748e-02 -3.09451491e-01 8.65072787e-01 3.88130993e-01
-1.39522803e+00 6.32001996e-01 6.16925287e+00 1.08861959e+00
-9.36519325e-01 1.34069666e-01 2.20273554e-01 2.35975091e-03
5.31250060e-01 1.21594258e-02 -1.52845252e+00 5.16478002e-01
6.09232068e-01 2.08148435e-02 -3.95177424e-01 1.17097497e+00
-7.16231903e-03 -1.02024637e-01 -1.24667406e+00 1.03304470e+00
-3.66999395e-02 -1.18042517e+00 -7.33109415e-02 2.22090185e-02
2.15702683e-01 1.84233755e-01 -4.07647155e-02 4.73629147e-01
-1.25154912e-01 -4.25095677e-01 9.22425449e-01 -5.19929305e-02
4.66622353e-01 -5.78110695e-01 6.25034988e-01 3.74207050e-01
-1.75923002e+00 -2.61709124e-01 -2.85683244e-01 1.68561295e-01
1.41293094e-01 5.54701626e-01 -5.44550002e-01 4.72391158e-01
7.96725869e-01 4.09964979e-01 -7.76029885e-01 1.49961817e+00
-1.23305090e-01 4.02192265e-01 -7.71638691e-01 -9.10079405e-02
3.89961213e-01 3.35584819e-01 8.10803950e-01 1.45679033e+00
4.77009624e-01 -1.40319224e-02 3.38149458e-01 3.31757307e-01
8.41275081e-02 1.67640653e-02 -9.55396593e-02 4.13885146e-01
9.22317088e-01 1.26307964e+00 -8.43370199e-01 -2.41037026e-01
-6.26793265e-01 8.99135232e-01 2.75642276e-01 -2.50727646e-02
-1.07586098e+00 -6.15191042e-01 5.92906594e-01 2.99986035e-01
8.45340192e-01 -3.60864639e-01 -2.59424537e-01 -1.08998668e+00
3.36811244e-01 -6.50632381e-01 3.06954712e-01 -4.42451239e-01
-1.05357921e+00 4.56953049e-02 -1.93776473e-01 -1.55828357e+00
3.11465532e-01 -6.41603827e-01 -6.31987870e-01 5.58882713e-01
-1.75911117e+00 -1.17754102e+00 -2.04199255e-01 4.15008634e-01
6.08924687e-01 2.37744108e-01 5.12642741e-01 5.84109902e-01
-4.94738638e-01 1.18467021e+00 1.84586227e-01 3.69206458e-01
7.77501523e-01 -1.15307927e+00 7.74499357e-01 1.17500091e+00
1.94129944e-01 6.46937788e-01 7.02270150e-01 -4.16566402e-01
-1.66465962e+00 -8.88399422e-01 6.37117684e-01 -4.29009080e-01
8.84519637e-01 -5.62357008e-01 -9.51934636e-01 3.64964128e-01
-4.19867933e-01 4.02898788e-01 3.75083178e-01 1.23804584e-01
-5.70501029e-01 -2.33345136e-01 -9.01937068e-01 6.03133500e-01
8.92183483e-01 -1.98623061e-01 -6.39477015e-01 3.52836579e-01
7.83131003e-01 -5.58015704e-01 -6.91468656e-01 6.64185941e-01
5.18775880e-01 -1.03357387e+00 1.38843369e+00 1.34965479e-01
-9.89329070e-02 -6.03386521e-01 -3.47692192e-01 -9.68900621e-01
-5.68652630e-01 -7.72319198e-01 -6.83437645e-01 1.16900969e+00
3.03150475e-01 -5.15960097e-01 9.23328757e-01 1.10511534e-01
-2.98626721e-01 -9.61395860e-01 -1.35254633e+00 -8.73254895e-01
-2.98588842e-01 -2.44947746e-01 3.37740451e-01 4.73222017e-01
-1.00963237e-02 6.05183125e-01 -3.57567608e-01 6.93585515e-01
6.64856136e-01 -1.08277783e-01 8.00667763e-01 -1.14746702e+00
-4.50890630e-01 -5.05761981e-01 -5.99298835e-01 -1.68981647e+00
-6.70571744e-01 -3.34419459e-01 1.46751508e-01 -1.38005626e+00
2.21043583e-02 -4.49864000e-01 -4.26677078e-01 5.79554856e-01
-5.26540339e-01 3.59767377e-01 5.02526999e-01 5.98086596e-01
-9.42961454e-01 3.95300329e-01 7.89514065e-01 1.01048211e-02
-2.49299690e-01 4.09588888e-02 -9.36011255e-01 9.82152581e-01
4.01154816e-01 -5.76104939e-01 2.91159377e-02 -3.29567492e-01
1.53602362e-01 -9.07561257e-02 3.32551926e-01 -1.40323436e+00
7.54649699e-01 3.28828037e-01 4.19219345e-01 -8.08752120e-01
5.13673306e-01 -8.42218161e-01 -5.21497071e-01 6.39641345e-01
5.17585911e-02 2.03160077e-01 1.29989162e-01 9.38872039e-01
-1.30763277e-01 -4.19412553e-03 9.90156412e-01 3.55544627e-01
-9.26336288e-01 2.21799448e-01 1.03413880e-01 -1.04733117e-01
1.21409464e+00 -5.24957001e-01 -3.29224885e-01 1.02377096e-02
-4.22551483e-01 5.42321265e-01 2.92421788e-01 4.47023243e-01
5.95206499e-01 -1.22413683e+00 -7.40019262e-01 1.90882370e-01
4.05172974e-01 -1.72588825e-02 -3.70806386e-03 1.06301320e+00
-2.15454444e-01 3.00450891e-01 3.06352556e-01 -8.93575013e-01
-1.24495971e+00 5.22961020e-01 3.49450529e-01 -3.79853904e-01
-7.61650860e-01 1.12962222e+00 2.62578726e-01 1.82786360e-01
6.06535137e-01 -4.37162578e-01 -4.27744575e-02 2.66386867e-01
7.53640831e-01 2.20717028e-01 2.39724606e-01 -6.04647577e-01
-6.44738257e-01 8.86198819e-01 -2.28900433e-01 2.53394008e-01
1.03617024e+00 -1.83405891e-01 5.84632874e-01 2.06778795e-01
1.08134007e+00 3.02855462e-01 -1.34243488e+00 -3.85553062e-01
1.81974411e-01 -5.42747676e-01 1.40711948e-01 -5.19142509e-01
-6.90563381e-01 6.26694798e-01 8.85031641e-01 4.12967473e-01
1.00306451e+00 7.97587708e-02 8.47106040e-01 2.70702243e-01
3.48085672e-01 -9.12722945e-01 2.17979908e-01 3.55805010e-01
8.18411827e-01 -1.36948895e+00 2.53471434e-01 -5.78002214e-01
-1.91558868e-01 7.89833903e-01 5.13745010e-01 -4.59686667e-01
6.19745672e-01 4.66279626e-01 -2.49389648e-01 -8.74171928e-02
-4.72744137e-01 -9.47458073e-02 4.81173277e-01 4.53150481e-01
1.13502070e-01 -1.09324224e-01 -9.80096757e-02 7.17536628e-01
1.62375078e-01 -2.90550113e-01 5.32039180e-02 1.10699141e+00
-9.28599238e-01 -7.52669096e-01 -6.46015525e-01 3.49680185e-01
-6.09974265e-01 -2.09481806e-01 -1.11615397e-01 1.03939807e+00
4.94881533e-02 8.06358397e-01 -6.58609197e-02 -5.14189243e-01
8.34858775e-01 -2.46530712e-01 -5.30304154e-03 -3.77372980e-01
-4.12512690e-01 2.17818320e-01 -1.77356690e-01 -6.72676444e-01
-1.01018049e-01 -4.80878621e-01 -1.18899918e+00 -3.95312570e-02
-9.79684651e-01 1.74248010e-01 5.58524966e-01 7.75990665e-01
7.54535973e-01 2.99796820e-01 5.81285894e-01 -1.30321240e+00
-1.08078074e+00 -7.50532508e-01 -4.27806884e-01 -1.26069084e-01
7.07823932e-01 -7.47119009e-01 -5.34227788e-01 -4.63617653e-01] | [8.710086822509766, -0.39567244052886963] |
a97dcfcf-9868-4e27-abb9-0ba8378f2a41 | learning-with-a-mole-transferable-latent | 2306.03857 | null | https://arxiv.org/abs/2306.03857v1 | https://arxiv.org/pdf/2306.03857v1.pdf | Learning with a Mole: Transferable latent spatial representations for navigation without reconstruction | Agents navigating in 3D environments require some form of memory, which should hold a compact and actionable representation of the history of observations useful for decision taking and planning. In most end-to-end learning approaches the representation is latent and usually does not have a clearly defined interpretation, whereas classical robotics addresses this with scene reconstruction resulting in some form of map, usually estimated with geometry and sensor models and/or learning. In this work we propose to learn an actionable representation of the scene independently of the targeted downstream task and without explicitly optimizing reconstruction. The learned representation is optimized by a blind auxiliary agent trained to navigate with it on multiple short sub episodes branching out from a waypoint and, most importantly, without any direct visual observation. We argue and show that the blindness property is important and forces the (trained) latent representation to be the only means for planning. With probing experiments we show that the learned representation optimizes navigability and not reconstruction. On downstream tasks we show that it is robust to changes in distribution, in particular the sim2real gap, which we evaluate with a real physical robot in a real office building, significantly improving performance. | ['Christian Wolf', 'Gianluca Monaci', 'Assem Sadek', 'Leonid Antsfeld', 'Guillaume Bono'] | 2023-06-06 | null | null | null | null | ['navigate'] | ['reasoning'] | [ 1.94454387e-01 8.01003635e-01 -2.99720801e-02 -3.70783120e-01
-6.73514426e-01 -7.61019111e-01 8.56256366e-01 2.30359778e-01
-5.69107652e-01 6.95798159e-01 3.44859034e-01 -2.67502785e-01
-3.84529829e-01 -8.87972295e-01 -1.12540889e+00 -8.49568188e-01
-3.76097590e-01 1.13880169e+00 1.06466062e-01 -2.50930041e-01
4.47823584e-01 4.56655681e-01 -1.49070764e+00 -6.63976371e-02
5.54437459e-01 5.74535370e-01 8.33332419e-01 9.90496695e-01
9.94482860e-02 1.14074862e+00 -2.74491608e-01 3.12022239e-01
5.48033953e-01 -3.38150978e-01 -1.10268855e+00 4.41620618e-01
2.11547032e-01 -3.65208805e-01 -4.55103964e-01 8.70786130e-01
2.53092587e-01 4.37637895e-01 7.04802632e-01 -9.61911142e-01
-2.94840187e-01 3.62590015e-01 1.31720155e-01 -1.72057465e-01
6.05883539e-01 3.33826989e-01 8.67391467e-01 -5.25559068e-01
8.50767732e-01 1.31996179e+00 3.97256017e-01 2.54739493e-01
-1.28197742e+00 2.74761647e-01 5.20819783e-01 1.92098618e-01
-1.16079414e+00 -5.64797938e-01 4.31948632e-01 -3.93601775e-01
1.17032480e+00 8.18190798e-02 6.20545030e-01 1.01597500e+00
1.94226101e-01 6.51195586e-01 8.96165133e-01 -3.82671237e-01
7.36848474e-01 3.76090109e-02 -1.04663089e-01 1.01891613e+00
4.16583605e-02 4.05198991e-01 -5.85513473e-01 9.95869339e-02
9.98494267e-01 -9.67527181e-03 -4.46642160e-01 -1.17727160e+00
-1.43448102e+00 7.36070633e-01 7.35908270e-01 -1.48996130e-01
-5.84943354e-01 3.01471740e-01 -7.11511821e-02 5.63463032e-01
-1.02204293e-01 5.54849803e-01 -4.71823722e-01 -2.86597997e-01
-2.91947156e-01 2.38437817e-01 1.01933444e+00 1.17797589e+00
8.82249236e-01 -2.85596967e-01 1.82842150e-01 3.46073270e-01
3.37899387e-01 4.97727394e-01 3.69040996e-01 -1.58298123e+00
6.62801266e-01 3.05001646e-01 7.11896837e-01 -8.13882232e-01
-7.50087619e-01 -3.30948532e-01 -3.66496533e-01 8.38487446e-01
6.63713813e-01 1.02893032e-01 -1.19094002e+00 1.93141174e+00
1.69903293e-01 -1.70234561e-01 3.04274976e-01 9.37157094e-01
2.12764636e-01 5.69056273e-01 -2.32633159e-01 5.29573448e-02
8.18724930e-01 -1.20369112e+00 -3.01271677e-01 -7.10328043e-01
7.85418272e-01 -1.97152346e-01 1.07800841e+00 5.70231080e-01
-9.35741425e-01 -3.20768595e-01 -1.15949392e+00 -3.32306623e-01
-3.33897293e-01 -1.69715837e-01 5.38685262e-01 1.46454155e-01
-1.30921078e+00 7.07849979e-01 -1.19350767e+00 -7.41179347e-01
3.65966260e-02 2.99320221e-01 -8.54288816e-01 -3.37444633e-01
-5.62622309e-01 1.29167521e+00 3.61233890e-01 5.58271110e-02
-1.44515991e+00 1.16710983e-01 -1.08800018e+00 -1.90100566e-01
3.97709936e-01 -9.80558634e-01 1.36687231e+00 -4.16589409e-01
-1.77404153e+00 7.13945925e-01 -3.07401389e-01 -6.01651251e-01
7.32703567e-01 -4.44300920e-01 3.60875756e-01 2.31754333e-01
4.14378464e-01 7.05119967e-01 8.51401627e-01 -1.47155285e+00
-6.11208439e-01 -6.54941559e-01 5.48415244e-01 5.64624786e-01
4.95136917e-01 -9.15647089e-01 -3.04315329e-01 3.24196517e-02
7.82468617e-01 -1.13487899e+00 -7.44411588e-01 1.40332967e-01
-3.60273778e-01 1.69963643e-01 4.19868529e-01 -5.51207662e-01
3.21097642e-01 -1.86638796e+00 6.56895161e-01 2.10212976e-01
3.26791145e-02 -5.15313864e-01 -2.99288511e-01 5.51172435e-01
1.80509925e-01 -2.01190248e-01 -3.78529102e-01 -7.01952875e-01
2.33473018e-01 6.91358447e-01 -5.39476335e-01 9.74641740e-01
-1.10844679e-01 8.47737670e-01 -1.17855811e+00 1.07276011e-02
5.50666809e-01 2.03166425e-01 -6.88446462e-01 3.32004011e-01
-4.92771715e-01 9.60788190e-01 -6.58315182e-01 2.28500351e-01
2.41065204e-01 -2.50983387e-01 1.95273712e-01 5.19678175e-01
-2.44165629e-01 7.45627820e-01 -1.21061587e+00 2.41858196e+00
-6.40247583e-01 4.70946521e-01 1.52010590e-01 -1.18540037e+00
7.67866492e-01 1.68314148e-02 1.10774249e-01 -7.94803381e-01
-1.43511772e-01 1.83978230e-01 -3.20154458e-01 -6.07245386e-01
4.08091724e-01 1.27140377e-02 -2.00548068e-01 5.09297431e-01
2.38071848e-03 -4.22254831e-01 -1.91144422e-01 2.25600183e-01
1.56136799e+00 5.75347662e-01 3.00150812e-01 -2.18019620e-01
8.47201347e-02 4.34562802e-01 2.04346329e-01 1.25761938e+00
6.90150559e-02 7.24675655e-01 2.91020334e-01 -4.87835616e-01
-1.06401980e+00 -1.31264961e+00 2.05751225e-01 8.75734091e-01
6.57728970e-01 -2.78782964e-01 -4.31266904e-01 -5.93751132e-01
-3.07817847e-01 1.09243238e+00 -7.32593119e-01 -2.05418020e-01
-6.72687471e-01 -1.83831543e-01 -5.11131622e-02 3.10706437e-01
2.83160925e-01 -1.25652337e+00 -1.32557082e+00 3.59070987e-01
-1.94312260e-01 -8.74444246e-01 2.56730355e-02 7.37899244e-01
-9.88159776e-01 -1.13689840e+00 -4.11261916e-01 -6.30239904e-01
1.07665491e+00 3.65734667e-01 1.09921908e+00 -2.59994328e-01
1.27635226e-01 7.70406902e-01 -3.44651014e-01 -9.60708596e-03
-4.19019639e-01 8.10224190e-03 2.73705255e-02 -4.05958593e-01
-1.42169625e-01 -9.07329321e-01 -3.85295659e-01 1.19150639e-01
-4.07201201e-01 2.89780945e-01 7.01288939e-01 6.94672167e-01
5.34167349e-01 2.66043972e-02 -8.21971893e-02 -5.94338775e-01
2.59230316e-01 -3.67282629e-01 -7.17434764e-01 -7.58831650e-02
-3.52957100e-01 6.80820227e-01 3.90072823e-01 -4.65985760e-02
-8.53474855e-01 5.10020018e-01 8.57985765e-02 -1.23860978e-01
-4.91733313e-01 4.80598450e-01 -2.96200931e-01 1.68983355e-01
8.17465425e-01 2.45122001e-01 -1.72143523e-02 -5.03695667e-01
5.99481404e-01 1.45349845e-01 5.77724218e-01 -5.15711963e-01
6.55416727e-01 6.96291745e-01 1.16237953e-01 -7.31992483e-01
-6.06206179e-01 -2.99576730e-01 -9.34635699e-01 9.66451317e-02
7.77358592e-01 -8.63430142e-01 -7.34700978e-01 1.10731177e-01
-1.23014188e+00 -1.03707337e+00 -5.60133815e-01 5.89847803e-01
-1.33638155e+00 3.57226259e-03 -3.08075905e-01 -9.09394681e-01
5.39774001e-01 -1.14838254e+00 1.19266248e+00 -2.20389962e-01
-2.59456694e-01 -8.67353618e-01 2.54236341e-01 -1.12747755e-02
1.59057945e-01 2.21347317e-01 7.42212355e-01 -1.89510435e-01
-1.20929694e+00 -7.40482882e-02 6.24075085e-02 -2.26873308e-01
1.30395874e-01 -8.41016352e-01 -9.02871072e-01 -1.96048439e-01
2.66781598e-01 -1.99142307e-01 1.02626598e+00 3.54287118e-01
6.05446756e-01 -5.19702435e-01 -4.87018645e-01 5.03698111e-01
1.44056404e+00 3.03402413e-02 7.03581452e-01 6.81571066e-01
5.22890925e-01 9.04694200e-01 7.32786179e-01 4.12289202e-01
6.08880997e-01 5.96505642e-01 1.12834299e+00 4.38223600e-01
1.18380912e-01 -5.05655348e-01 5.75409591e-01 3.29078436e-01
-6.43378273e-02 -3.04930717e-01 -9.14317667e-01 6.97970152e-01
-2.17324805e+00 -8.69785070e-01 1.64672956e-01 2.34824944e+00
4.04029608e-01 1.78587765e-01 -2.30987385e-01 9.84480232e-03
2.80295789e-01 5.46178818e-02 -7.47404277e-01 -1.48476496e-01
5.38181029e-02 -2.42018119e-01 8.23225200e-01 1.21994507e+00
-9.20491576e-01 9.05988812e-01 6.46180916e+00 -1.88503433e-02
-7.08318591e-01 6.32834956e-02 -8.18627514e-03 4.01783921e-02
-3.21423829e-01 2.42494136e-01 -4.07617569e-01 -3.34900841e-02
7.46059179e-01 3.45100403e-01 8.45293283e-01 1.06719100e+00
2.37972260e-01 -4.63413179e-01 -1.62902689e+00 8.28060031e-01
-1.91368848e-01 -1.24459720e+00 -2.51475662e-01 4.35203880e-01
3.17683309e-01 3.34910691e-01 -6.03265539e-02 3.48763853e-01
8.55294466e-01 -1.20805204e+00 1.26658916e+00 6.77417099e-01
2.38707021e-01 -5.05499542e-01 4.33327973e-01 9.81283307e-01
-8.13037097e-01 -2.72951365e-01 -3.28744233e-01 -4.63108569e-01
3.90333951e-01 9.29136351e-02 -9.69170213e-01 2.68487036e-01
4.58916038e-01 6.80270731e-01 -2.49175176e-01 9.78629947e-01
-5.93405128e-01 6.41532466e-02 -5.89280248e-01 6.06067516e-02
4.24841464e-01 -1.98016956e-01 9.14510906e-01 7.11297512e-01
4.52912688e-01 7.33165070e-02 4.62450296e-01 7.45765448e-01
5.34723818e-01 -5.66346705e-01 -9.87507284e-01 1.23508543e-01
2.90620029e-01 6.18934393e-01 -7.35246778e-01 -1.02659851e-01
2.23537069e-02 1.23426795e+00 6.30591691e-01 5.94553173e-01
-5.12385309e-01 -2.92309048e-03 6.75031006e-01 -1.83567125e-02
5.19461036e-01 -8.34683001e-01 -3.04288477e-01 -9.52463984e-01
1.13912866e-01 -3.75847489e-01 1.58542637e-02 -8.83285761e-01
-7.01756716e-01 4.60766375e-01 -1.04288317e-01 -1.10318744e+00
-6.68638706e-01 -6.08622015e-01 -2.41724640e-01 7.82632351e-01
-1.55178976e+00 -9.13913071e-01 -4.68760580e-01 5.13950765e-01
5.62185824e-01 -4.38013673e-03 1.08209682e+00 -3.01688582e-01
2.15640813e-02 -1.64150596e-01 -5.83694838e-02 -3.23990077e-01
3.46971959e-01 -1.47529042e+00 5.01294971e-01 7.31509566e-01
3.23244423e-01 6.78098619e-01 1.21211374e+00 -5.53631008e-01
-1.61648536e+00 -9.25692976e-01 5.76052308e-01 -8.39269578e-01
3.15851808e-01 -4.21075284e-01 -5.83976626e-01 1.19597995e+00
-5.89988343e-02 -1.29441679e-01 1.70525964e-02 4.71769683e-02
-1.60782978e-01 4.62382972e-01 -1.11397815e+00 6.67453408e-01
1.48899889e+00 -4.99005228e-01 -7.95532107e-01 4.35265124e-01
8.76930594e-01 -6.26279652e-01 -1.20214567e-01 1.20859332e-01
2.98588514e-01 -1.15916812e+00 9.75665390e-01 -5.45246184e-01
1.71302795e-01 -4.62239951e-01 -5.72809160e-01 -1.29225743e+00
-3.88790518e-01 -5.99160552e-01 -3.14389050e-01 5.29008150e-01
3.97184759e-01 -7.03290343e-01 9.92087007e-01 5.51203430e-01
-2.50069588e-01 -2.26599842e-01 -9.63450372e-01 -7.03382730e-01
-2.59792536e-01 -6.84512198e-01 3.61167312e-01 4.79386151e-01
-9.18573514e-02 4.84492600e-01 -3.12240720e-01 8.09924543e-01
7.10764110e-01 2.31742829e-01 1.03370965e+00 -9.94696796e-01
-4.57104504e-01 -1.73858806e-01 -4.85022783e-01 -1.78114188e+00
5.15122153e-02 -6.92361891e-01 6.85052037e-01 -1.96009421e+00
-2.77959287e-01 -5.83999097e-01 9.19147357e-02 3.72555137e-01
4.39963311e-01 -3.78485441e-01 5.35935536e-02 2.53406942e-01
-8.41313243e-01 7.14441538e-01 1.08467185e+00 -1.43339649e-01
-4.16472465e-01 4.16480787e-02 -4.62607652e-01 8.21633458e-01
7.20069289e-01 -5.00608146e-01 -6.60614848e-01 -7.76739717e-01
3.19622755e-01 2.02230230e-01 7.18657792e-01 -9.51597214e-01
4.11647141e-01 -3.08890045e-01 2.39964128e-01 -4.78362322e-01
8.30813944e-01 -9.48968649e-01 -8.08373839e-03 5.47679901e-01
-5.23254871e-01 -2.29696296e-02 -1.75490856e-01 9.59310889e-01
2.29871288e-01 -4.04714197e-01 2.99076051e-01 -5.65240324e-01
-1.08982933e+00 1.91791400e-01 -4.38553184e-01 -2.58330643e-01
7.99659193e-01 -2.78841764e-01 -6.12837374e-02 -7.96973646e-01
-1.21418023e+00 3.71138126e-01 8.62992108e-01 2.72957742e-01
6.63766980e-01 -1.08249485e+00 -4.83818978e-01 2.64585823e-01
1.35157049e-01 5.11234879e-01 -2.56809611e-02 6.08021677e-01
-5.91464758e-01 5.85553050e-01 -2.19325960e-01 -7.34628499e-01
-6.54652417e-01 6.83893204e-01 2.87668288e-01 -1.34833530e-01
-1.08305752e+00 7.71631062e-01 4.31818128e-01 -9.47501421e-01
4.60936546e-01 -4.56684381e-01 -6.70418292e-02 -3.22634459e-01
3.64924997e-01 7.47164786e-02 -2.26316154e-01 -5.75383008e-01
-2.42903903e-01 4.67840075e-01 4.15972024e-01 -6.59712017e-01
1.48203909e+00 -7.37361908e-01 4.00449447e-02 6.94863558e-01
9.18535471e-01 -8.42580050e-02 -1.75493145e+00 -2.09154755e-01
1.90091923e-01 -4.44006473e-01 -1.57846987e-01 -6.13092601e-01
-4.52301741e-01 9.03579712e-01 4.00551468e-01 2.99242884e-01
4.97191250e-01 2.77102262e-01 3.04035187e-01 1.05319428e+00
1.15447283e+00 -8.92152548e-01 1.83385193e-01 8.44835341e-01
1.24895716e+00 -1.33172309e+00 -1.71162739e-01 -9.86550748e-02
-4.42625910e-01 9.80190456e-01 2.87548065e-01 -3.05734605e-01
3.84112716e-01 3.35605294e-02 1.74858980e-02 -2.78333724e-01
-7.02995181e-01 -4.86613244e-01 -1.74121127e-01 9.36912179e-01
-3.56703073e-01 -4.45298180e-02 5.99116623e-01 1.55050978e-01
-5.35280704e-01 -3.53145659e-01 4.73772407e-01 1.21804416e+00
-9.28021133e-01 -9.38238263e-01 -2.88202316e-01 1.52920680e-02
4.41742241e-01 2.73115456e-01 -2.47932434e-01 6.83555365e-01
-2.47588381e-02 8.59554410e-01 2.04816107e-02 -1.18659392e-01
2.24119246e-01 -6.21467307e-02 8.74954104e-01 -8.91698480e-01
2.60046154e-01 -2.60522932e-01 2.64463872e-01 -1.12056601e+00
-2.79349357e-01 -7.96242297e-01 -1.58637536e+00 1.98472571e-02
1.14042647e-01 7.41348416e-02 7.90035963e-01 1.17052567e+00
3.30567300e-01 4.57244635e-01 3.12006086e-01 -1.48509896e+00
-4.45417941e-01 -7.58444369e-01 -4.17401671e-01 1.06104933e-01
1.02545559e+00 -8.43526065e-01 -3.98095429e-01 1.99041776e-02] | [4.665308475494385, 0.7159745693206787] |
3fdda359-bdae-424f-9306-2ba411c0b816 | attribute-recognition-by-joint-recurrent | 1709.08553 | null | http://arxiv.org/abs/1709.08553v1 | http://arxiv.org/pdf/1709.08553v1.pdf | Attribute Recognition by Joint Recurrent Learning of Context and Correlation | Recognising semantic pedestrian attributes in surveillance images is a
challenging task for computer vision, particularly when the imaging quality is
poor with complex background clutter and uncontrolled viewing conditions, and
the number of labelled training data is small. In this work, we formulate a
Joint Recurrent Learning (JRL) model for exploring attribute context and
correlation in order to improve attribute recognition given small sized
training data with poor quality images. The JRL model learns jointly pedestrian
attribute correlations in a pedestrian image and in particular their sequential
ordering dependencies (latent high-order correlation) in an end-to-end
encoder/decoder recurrent network. We demonstrate the performance advantage and
robustness of the JRL model over a wide range of state-of-the-art deep models
for pedestrian attribute recognition, multi-label image classification, and
multi-person image annotation on two largest pedestrian attribute benchmarks
PETA and RAP. | ['Jingya Wang', 'Shaogang Gong', 'Xiatian Zhu', 'Wei Li'] | 2017-09-25 | attribute-recognition-by-joint-recurrent-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Wang_Attribute_Recognition_by_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Wang_Attribute_Recognition_by_ICCV_2017_paper.pdf | iccv-2017-10 | ['pedestrian-attribute-recognition', 'multi-label-image-classification'] | ['computer-vision', 'computer-vision'] | [ 4.42722172e-01 -3.29455256e-01 -1.39220757e-02 -1.01612282e+00
-8.29188168e-01 -3.95100266e-01 6.41740561e-01 1.65586531e-01
-4.51528490e-01 5.52808166e-01 4.17860597e-01 5.88971563e-03
1.09474510e-01 -5.14015138e-01 -8.82851541e-01 -7.66846418e-01
-6.48104772e-02 6.42579019e-01 -1.32320225e-01 1.92009628e-01
-4.37048912e-01 2.06223801e-01 -1.60947120e+00 6.48818433e-01
3.75411123e-01 1.25150251e+00 -1.27098739e-01 1.02759361e+00
4.53424066e-01 9.96751904e-01 -4.91120875e-01 -9.69092488e-01
1.52979568e-01 -1.55210719e-01 -7.85882592e-01 6.32021904e-01
1.06470072e+00 -4.02249664e-01 -5.51091075e-01 9.38540816e-01
5.92687309e-01 -2.52864975e-02 6.89594865e-01 -1.44933283e+00
-8.06211829e-01 2.52095252e-01 -5.09621263e-01 3.53093088e-01
3.29483539e-01 5.12241364e-01 1.02726460e+00 -7.55635262e-01
2.61149049e-01 1.59412575e+00 8.35848033e-01 5.08812726e-01
-1.46787715e+00 -5.78157067e-01 5.78763187e-01 5.73113978e-01
-1.49138939e+00 -4.90491122e-01 5.34579098e-01 -5.38952470e-01
1.19630611e+00 8.91616642e-02 7.73377538e-01 1.74976873e+00
1.12987585e-01 1.03522134e+00 1.04583633e+00 -2.00313963e-02
-1.16145916e-01 -1.24316670e-01 2.59765714e-01 7.74239302e-01
7.73325562e-02 1.41964048e-01 -6.56558216e-01 -1.68835431e-01
4.37992454e-01 -1.25954017e-01 3.64702970e-01 -2.57609129e-01
-1.25194979e+00 6.51716590e-01 2.26632431e-01 -5.54877758e-01
-4.85487580e-01 1.84356555e-01 9.10447657e-01 2.03776240e-01
6.89345062e-01 -5.65851144e-02 -4.50736642e-01 7.47448653e-02
-4.97917563e-01 1.01544835e-01 5.86913943e-01 1.25417614e+00
3.72370392e-01 4.48537990e-02 -4.37730312e-01 1.09622562e+00
4.06715363e-01 7.60863721e-01 -1.92914858e-01 -9.83637154e-01
5.21283686e-01 2.49047115e-01 -1.40496433e-01 -5.52310765e-01
-4.00939286e-01 -4.63187605e-01 -1.19339001e+00 1.70915753e-01
5.10120749e-01 6.32004365e-02 -8.44923973e-01 1.79274523e+00
1.41564205e-01 1.95525125e-01 1.63288370e-01 9.95586574e-01
1.00195158e+00 4.92561370e-01 9.56858635e-01 1.30769983e-01
1.68599606e+00 -1.03254271e+00 -5.01097441e-01 -5.62570393e-01
3.82141978e-01 -7.10762084e-01 6.96859956e-01 6.07152879e-02
-7.29510248e-01 -9.20742512e-01 -7.99560189e-01 -1.50935605e-01
-2.34706968e-01 5.29689193e-01 6.09755278e-01 6.32499993e-01
-9.12826478e-01 9.39920694e-02 -4.95357752e-01 -5.50186098e-01
1.06248426e+00 3.97320807e-01 -5.20340681e-01 -3.71671051e-01
-1.12311733e+00 7.85787523e-01 1.21022403e-01 1.44161135e-01
-1.28780305e+00 -5.36080003e-01 -9.63814974e-01 -8.24517906e-02
3.10899138e-01 -1.17418694e+00 9.58434284e-01 -9.92339313e-01
-1.17091167e+00 1.46931970e+00 -2.55654067e-01 -6.89728677e-01
5.96070111e-01 -6.87911630e-01 -4.78889108e-01 -1.32336497e-01
3.76594514e-01 8.92440856e-01 8.70554030e-01 -1.02464151e+00
-7.05409288e-01 -6.05684161e-01 -2.91043431e-01 2.39143550e-01
2.46527698e-02 3.97319257e-01 -1.91273615e-01 -6.44217849e-01
-4.19975400e-01 -1.05479145e+00 -4.68917280e-01 5.68909124e-02
-4.68045086e-01 -4.35111612e-01 7.12967336e-01 -7.09843338e-01
4.00723487e-01 -2.31200266e+00 5.12810834e-02 -5.92268817e-02
1.45176277e-01 2.96003342e-01 -2.87219077e-01 -3.23894769e-01
-2.67164737e-01 -1.86727211e-01 -7.64191225e-02 -5.47186136e-01
2.24422052e-04 4.57387567e-01 -1.42395929e-01 6.34809196e-01
5.17359912e-01 9.64417696e-01 -8.11755896e-01 -7.21362829e-01
1.82728007e-01 7.28768826e-01 -2.70467132e-01 4.98835742e-01
-1.14414431e-01 6.78976417e-01 -5.04072785e-01 1.01613939e+00
4.41361964e-01 -2.69751191e-01 -1.83583617e-01 -4.78713691e-01
1.09948404e-01 -8.21521953e-02 -9.52838659e-01 1.51629877e+00
-2.91886389e-01 7.83204675e-01 -2.19537914e-01 -7.28804469e-01
9.38044190e-01 4.08648580e-01 5.08707643e-01 -7.13500023e-01
-6.89857230e-02 -6.26350492e-02 -3.46549779e-01 -5.07520735e-01
3.20241004e-01 3.13812286e-01 -3.32644314e-01 8.43095034e-02
2.70258874e-01 6.22089148e-01 2.95881093e-01 1.00395985e-01
7.91538954e-01 2.33362734e-01 3.35010469e-01 -2.03173563e-01
7.05353320e-01 -3.57282221e-01 7.70052314e-01 9.72650826e-01
-4.48315531e-01 5.71989119e-01 5.19590080e-01 -1.14148247e+00
-1.68398583e+00 -1.03122067e+00 -2.04551399e-01 1.61754024e+00
-1.29766241e-01 -2.14688286e-01 -5.09878457e-01 -7.51878142e-01
-1.59104109e-01 4.73233461e-01 -5.82596719e-01 -1.89453349e-01
-6.90004885e-01 -1.07333934e+00 8.52463543e-01 9.73927498e-01
1.04918718e+00 -8.98842573e-01 -5.95786989e-01 2.44353250e-01
-3.48561227e-01 -1.91077363e+00 -4.76240844e-01 7.43300244e-02
-2.36415282e-01 -9.82968211e-01 -5.77511966e-01 -6.65332377e-01
4.75067317e-01 -6.31363764e-02 1.62832797e+00 -3.59414697e-01
-5.98583043e-01 7.11856186e-01 -2.16043010e-01 -3.68896842e-01
-3.55261832e-01 -1.17935091e-01 1.32739440e-01 4.27730352e-01
6.94231153e-01 -1.67214110e-01 -5.59968233e-01 6.60332620e-01
-1.89404473e-01 1.09784417e-01 8.84624541e-01 1.19907248e+00
7.01913059e-01 -3.68139416e-01 3.68544310e-01 -6.41620278e-01
8.76809061e-02 -1.80918604e-01 -5.95612288e-01 4.91105080e-01
-9.06169936e-02 2.03354284e-03 4.94937271e-01 -2.41903365e-01
-1.24425673e+00 4.68711823e-01 -3.19301456e-01 -2.41051018e-01
-8.75138462e-01 -1.60224736e-01 -5.92800140e-01 2.43383244e-01
4.89812374e-01 1.45872474e-01 -1.40881941e-01 -2.33912706e-01
3.55815977e-01 4.36951071e-01 9.73200142e-01 -7.44682908e-01
9.35796052e-02 2.45548829e-01 2.90795952e-01 -7.01913893e-01
-1.42739213e+00 -5.92603564e-01 -1.19791484e+00 -4.97077942e-01
1.48247302e+00 -1.38740253e+00 -9.63254750e-01 7.92096496e-01
-1.27411008e+00 -2.88162827e-01 -2.76042223e-01 4.37447548e-01
-7.97573745e-01 2.18173832e-01 -6.95559978e-01 -8.07453096e-01
-3.22910756e-01 -1.32057571e+00 1.39359963e+00 1.66353863e-02
-1.80130556e-01 -6.26736879e-01 -4.39193040e-01 7.11799800e-01
1.49624944e-01 3.79689366e-01 8.50041866e-01 -6.83993340e-01
-6.99916124e-01 -1.93944484e-01 -5.17198324e-01 3.41123104e-01
-2.50111997e-01 -2.27423340e-01 -1.19567072e+00 -2.33594671e-01
-5.57011366e-01 -6.65743649e-01 1.17026293e+00 5.07179439e-01
1.33763242e+00 -3.34430754e-01 -1.83186337e-01 7.20441937e-01
9.02015567e-01 -2.53903806e-01 4.92811769e-01 3.88546705e-01
1.09305704e+00 6.71893477e-01 6.86543167e-01 6.68087065e-01
5.08828819e-01 9.35112298e-01 2.54442126e-01 -2.40421310e-01
-4.53027219e-01 2.03694582e-01 3.72036397e-01 2.65229702e-01
-2.67154664e-01 -1.68452084e-01 -1.02034760e+00 3.44261557e-01
-2.05742073e+00 -1.25007927e+00 -3.32549095e-01 2.12585449e+00
5.97969294e-01 2.28400510e-02 4.31354940e-01 -1.91594154e-01
7.93540061e-01 2.21848413e-01 -5.18833816e-01 -2.60076851e-01
-5.87819934e-01 -4.61664289e-01 7.51008928e-01 4.92000133e-02
-2.00348139e+00 8.95980299e-01 6.28905201e+00 4.25356627e-01
-3.74012113e-01 3.22444364e-03 1.42184591e+00 7.11843446e-02
4.85370964e-01 -4.35626924e-01 -1.33771586e+00 4.35570598e-01
1.12452245e+00 5.66684902e-01 3.03295497e-02 1.02916586e+00
-3.24752361e-01 7.89444894e-02 -1.42855263e+00 1.14034772e+00
1.76313460e-01 -8.94477367e-01 2.00708225e-01 -2.13102460e-01
4.06054735e-01 5.15095070e-02 5.67591488e-01 2.17991024e-01
5.89472294e-01 -1.22287428e+00 7.18737602e-01 6.30473197e-01
8.99901628e-01 -8.11111391e-01 9.98938143e-01 -1.31016418e-01
-1.41331911e+00 -2.09587395e-01 -4.41206664e-01 4.21651185e-01
2.45714575e-01 3.87899995e-01 -5.05872726e-01 2.73850352e-01
1.17248344e+00 1.11480367e+00 -1.02418983e+00 9.83945131e-01
-6.00507036e-02 5.74023008e-01 -9.87195671e-02 3.09017807e-01
3.74684006e-01 -9.51809585e-02 5.98074973e-01 1.80707014e+00
-7.96806142e-02 2.47019738e-01 4.80710983e-01 5.52749038e-01
-8.49807188e-02 -4.70951945e-01 -5.61819971e-01 3.91670823e-01
1.26748517e-01 1.13291073e+00 -4.77761835e-01 -4.28904474e-01
-7.09327877e-01 1.07516241e+00 1.64342195e-01 4.76279378e-01
-8.13021064e-01 5.21439373e-01 1.02888310e+00 -4.01753902e-01
4.50485617e-01 -2.48666093e-01 -4.30991471e-01 -8.56252611e-01
-1.91265494e-02 -1.03786373e+00 8.48716795e-01 -4.76733297e-01
-1.81975067e+00 7.33206451e-01 8.70466158e-02 -1.14900482e+00
-3.26569915e-01 -7.39154935e-01 -1.89374387e-01 7.45934188e-01
-1.39654362e+00 -1.94439280e+00 -6.21943891e-01 7.83843398e-01
9.07014489e-01 -7.91526914e-01 1.08024323e+00 4.87654507e-01
-6.99915111e-01 8.33163559e-01 -2.16771767e-01 4.66497540e-01
8.26696515e-01 -1.30631316e+00 7.23463416e-01 7.85086870e-01
1.26623377e-01 -1.11933932e-01 8.33957493e-01 -5.23189545e-01
-1.09011245e+00 -1.49281251e+00 7.38665223e-01 -8.15366745e-01
3.20010126e-01 -5.78059614e-01 -8.90186787e-01 7.15066552e-01
3.24348003e-01 6.19500339e-01 1.06986892e+00 3.48831862e-01
-8.27514112e-01 -2.45201617e-01 -7.49570668e-01 2.98366666e-01
1.07822347e+00 -6.89885080e-01 -6.93335310e-02 5.08001506e-01
4.30810899e-01 -1.96108505e-01 -9.64925170e-01 5.14743447e-01
6.63772404e-01 -6.37067199e-01 1.59147656e+00 -7.88180113e-01
2.10249871e-01 -1.28123477e-01 -4.80321407e-01 -1.10079324e+00
-6.54319465e-01 -2.98293382e-01 1.46424726e-01 1.38574934e+00
3.89147371e-01 -1.64437264e-01 6.20333493e-01 9.26208675e-01
-1.66584209e-01 -5.46396673e-01 -1.00102890e+00 -6.88630283e-01
-3.19067687e-01 -1.53404325e-01 4.33139652e-01 6.09875202e-01
-9.85775650e-01 5.67289770e-01 -8.28493834e-01 4.07645136e-01
1.41299164e+00 -1.79743811e-01 7.62743413e-01 -1.24597967e+00
-1.90301970e-01 -3.27750057e-01 -9.00229692e-01 -8.34273875e-01
4.71191019e-01 -5.22788286e-01 -1.04202859e-01 -1.23720491e+00
6.51387572e-01 -4.11697000e-01 -2.58257538e-01 4.36879188e-01
-3.57918233e-01 2.02262595e-01 1.37990937e-01 2.31597945e-01
-1.38682795e+00 7.24327981e-01 6.52625918e-01 -6.36987507e-01
4.19479817e-01 2.33016580e-01 -1.08162910e-01 6.79954588e-01
5.16669631e-01 -4.36056525e-01 -6.79665655e-02 -4.43625569e-01
6.01317696e-02 -1.22430347e-01 1.07839024e+00 -1.02132118e+00
3.37534994e-01 2.58471727e-01 1.09876835e+00 -5.92749536e-01
5.57345569e-01 -7.12936699e-01 -8.81680474e-02 5.37664741e-02
-8.14426422e-01 3.58604163e-01 1.71296597e-01 9.59252954e-01
-2.04509839e-01 3.17252010e-01 8.92926216e-01 -9.56896394e-02
-1.16014993e+00 5.87883651e-01 -4.46566939e-01 1.83839589e-01
1.00889647e+00 -3.62735577e-02 -3.27203691e-01 -3.05581301e-01
-9.45000708e-01 3.41434956e-01 -6.42107218e-04 4.69085693e-01
9.15591538e-01 -1.60369277e+00 -1.30892766e+00 1.59419969e-01
3.83983642e-01 -1.75799698e-01 3.00890774e-01 6.60439432e-01
-8.41638744e-02 3.28662574e-01 -7.32257962e-01 -1.02118802e+00
-1.77338135e+00 5.00908315e-01 3.04431558e-01 -1.71309501e-01
-8.30968738e-01 1.09483731e+00 3.68550837e-01 -1.79505959e-01
4.10548568e-01 3.17786992e-01 -4.67642725e-01 3.21203023e-02
4.95547384e-01 2.11580500e-01 -2.28278726e-01 -1.20955420e+00
-2.90432692e-01 5.52749276e-01 -4.24071103e-01 2.14074031e-01
1.15268183e+00 -2.34334767e-01 3.66039835e-02 2.91166693e-01
1.21230209e+00 -9.36504424e-01 -1.82745600e+00 -6.02134585e-01
4.89273109e-02 -4.58688885e-01 -2.63823539e-01 -6.56978071e-01
-8.35167944e-01 8.89344394e-01 9.30927217e-01 -3.40278745e-01
1.00483763e+00 2.92615831e-01 6.74485922e-01 6.25281811e-01
1.59298912e-01 -1.27522230e+00 3.34749639e-01 4.76469457e-01
6.92469597e-01 -1.76890099e+00 9.17294398e-02 -5.64133763e-01
-1.28289688e+00 1.04712677e+00 8.15845609e-01 2.09649265e-01
4.17124212e-01 1.48940220e-01 -1.68072544e-02 -1.17158040e-01
-8.75254631e-01 -2.93554217e-01 4.51326936e-01 7.70749152e-01
4.59177196e-01 2.25283414e-01 6.68214917e-01 4.35815483e-01
1.25770494e-01 -5.76662242e-01 1.12495184e-01 3.71747345e-01
-8.15503150e-02 -8.91075850e-01 -2.86069036e-01 4.39256519e-01
-4.88515198e-01 -3.99031490e-02 -2.61506498e-01 2.89034456e-01
1.79551229e-01 8.97742391e-01 2.44220600e-01 -2.80291259e-01
3.03739786e-01 1.20034061e-01 2.97321349e-01 -8.23130980e-02
-5.58376968e-01 -4.77112159e-02 7.01757669e-01 -7.52567470e-01
-6.42201126e-01 -1.11036813e+00 -5.51453412e-01 -8.89922231e-02
2.82388359e-01 -4.30098951e-01 3.59001279e-01 1.06575525e+00
2.78566748e-01 7.59624898e-01 4.28761899e-01 -5.36031485e-01
-4.41787809e-01 -8.91186357e-01 -1.18496552e-01 7.21000433e-01
4.68813837e-01 -6.79170728e-01 3.55650127e-01 5.10240853e-01] | [14.464241981506348, 0.9811956882476807] |
05efa87c-5316-468a-8f76-bdf646b2635f | co-search-covid-19-information-retrieval-with | 2006.09595 | null | https://arxiv.org/abs/2006.09595v1 | https://arxiv.org/pdf/2006.09595v1.pdf | CO-Search: COVID-19 Information Retrieval with Semantic Search, Question Answering, and Abstractive Summarization | The COVID-19 global pandemic has resulted in international efforts to understand, track, and mitigate the disease, yielding a significant corpus of COVID-19 and SARS-CoV-2-related publications across scientific disciplines. As of May 2020, 128,000 coronavirus-related publications have been collected through the COVID-19 Open Research Dataset Challenge. Here we present CO-Search, a retriever-ranker semantic search engine designed to handle complex queries over the COVID-19 literature, potentially aiding overburdened health workers in finding scientific answers during a time of crisis. The retriever is built from a Siamese-BERT encoder that is linearly composed with a TF-IDF vectorizer, and reciprocal-rank fused with a BM25 vectorizer. The ranker is composed of a multi-hop question-answering module, that together with a multi-paragraph abstractive summarizer adjust retriever scores. To account for the domain-specific and relatively limited dataset, we generate a bipartite graph of document paragraphs and citations, creating 1.3 million (citation title, paragraph) tuples for training the encoder. We evaluate our system on the data of the TREC-COVID information retrieval challenge. CO-Search obtains top performance on the datasets of the first and second rounds, across several key metrics: normalized discounted cumulative gain, precision, mean average precision, and binary preference. | ['Anuprit Kale', 'Richard Socher', 'Kazuma Hashimoto', 'Andre Esteva', 'Dragomir Radev', 'Wenpeng Yin', 'Romain Paulus'] | 2020-06-17 | null | null | null | null | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 4.00673412e-02 -1.35122895e-01 -5.71596384e-01 -1.17035367e-01
-1.36708260e+00 -7.67154038e-01 5.22929788e-01 8.09493303e-01
-5.94049811e-01 7.99690664e-01 8.07327330e-01 -3.86032581e-01
-4.38302130e-01 -6.60789192e-01 -8.17023039e-01 -1.01660296e-01
4.63369042e-02 1.07936776e+00 -8.46488178e-02 -2.26029187e-01
3.80087286e-01 3.08983207e-01 -1.06952405e+00 5.00522316e-01
1.12164736e+00 9.25813615e-01 3.54772836e-01 8.41644049e-01
-1.84143111e-01 5.23587167e-01 -8.15330386e-01 -6.60439193e-01
-2.29055569e-01 -2.02070307e-02 -8.86313319e-01 -7.98245609e-01
1.19816564e-01 -7.22197518e-02 -4.64989096e-01 7.26933300e-01
6.39368117e-01 5.98095544e-02 8.52338314e-01 -8.85600448e-01
-9.66872692e-01 4.68067646e-01 -4.49590862e-01 8.12361896e-01
4.14806753e-01 -1.27513632e-01 1.52927697e+00 -6.69407666e-01
1.13143122e+00 1.30834794e+00 4.83913928e-01 2.46470407e-01
-1.00871301e+00 -4.65509355e-01 -3.20130169e-01 1.35803893e-01
-1.25164366e+00 -1.59235865e-01 2.43832827e-01 -5.10296643e-01
1.47490239e+00 2.43644476e-01 2.42351949e-01 1.24215150e+00
6.62517428e-01 3.96794051e-01 2.61245459e-01 2.89829820e-01
2.11899355e-01 -1.17182322e-01 2.13907406e-01 5.69118917e-01
6.40729964e-01 -2.17800930e-01 -4.35826391e-01 -7.12541640e-01
1.50021181e-01 2.65749454e-01 -3.21912110e-01 3.15508664e-01
-1.15579450e+00 9.95840073e-01 5.26281476e-01 3.01611632e-01
-8.06817830e-01 5.18982718e-03 5.87787509e-01 2.25847080e-01
5.88125408e-01 1.01922691e+00 -5.87683141e-01 -6.44752085e-02
-1.06401110e+00 6.84190154e-01 9.74949241e-01 7.43765056e-01
3.05476099e-01 -5.64211547e-01 -8.51948440e-01 8.69691789e-01
3.57475162e-01 9.78266120e-01 3.44553560e-01 -8.37111533e-01
5.83049893e-01 6.17928505e-01 -3.53766419e-03 -1.25520217e+00
-4.69584882e-01 -7.20388234e-01 -7.04718530e-01 -9.87103045e-01
-3.85644704e-01 -1.71097741e-01 -1.01011872e+00 1.68910837e+00
-6.54194057e-02 1.21493533e-01 2.57126480e-01 8.06080043e-01
1.18900037e+00 1.03655040e+00 2.38195032e-01 -2.96671122e-01
1.85602129e+00 -7.80939996e-01 -9.83566105e-01 -5.32592088e-02
4.88118798e-01 -6.23774171e-01 2.99434662e-01 4.09890898e-02
-9.86045241e-01 -1.84824362e-01 -1.01656818e+00 -3.94085199e-01
-8.28341722e-01 -2.32961953e-01 2.91702330e-01 1.59490053e-02
-1.09820712e+00 3.53211105e-01 -3.43994528e-01 -3.91576022e-01
4.24477488e-01 -5.13129905e-02 -1.33884355e-01 -3.62093896e-01
-1.69247878e+00 1.17687178e+00 2.18780369e-01 -2.36676723e-01
-8.89950097e-01 -1.16036201e+00 -5.55844307e-01 3.71602148e-01
1.60190046e-01 -1.29144049e+00 8.36327910e-01 3.10047030e-01
-6.93769455e-01 6.83297038e-01 -3.42349261e-01 -7.25723267e-01
-1.50024652e-01 -9.24228057e-02 -8.29930305e-01 5.20997047e-01
6.29488766e-01 6.16418123e-01 3.64936054e-01 -7.64830232e-01
-3.84007543e-01 -5.44558644e-01 -2.33390197e-01 2.45913029e-01
6.70069605e-02 2.10054234e-01 -6.55675173e-01 -8.02788436e-01
-5.57345748e-01 -7.02577472e-01 -1.78215489e-01 -6.60193563e-01
-3.95962328e-01 -4.60277408e-01 5.27604580e-01 -1.10931611e+00
1.51881933e+00 -1.86036360e+00 3.50282043e-01 9.93311852e-02
3.54047328e-01 3.38230699e-01 -4.59071487e-01 6.97238982e-01
2.35821903e-01 5.37662566e-01 -1.61340356e-01 -8.45417604e-02
-1.85154915e-01 1.38232589e-01 -6.41969681e-01 1.12142079e-01
3.41717839e-01 1.25440478e+00 -1.15173388e+00 -4.53260034e-01
-5.84047258e-01 5.65474927e-01 -5.90036392e-01 3.44777197e-01
-7.25939095e-01 -1.25637800e-02 -7.82500684e-01 6.81372046e-01
3.53212416e-01 -8.43107283e-01 4.49621864e-02 -1.09509245e-01
2.27313012e-01 4.95113373e-01 -1.55373186e-01 1.74475038e+00
-2.58211017e-01 3.78795207e-01 -7.08731562e-02 -7.25205839e-01
6.81146264e-01 4.09295589e-01 6.48707628e-01 -8.25927794e-01
5.57763726e-02 4.00897026e-01 -5.63320041e-01 -5.38411558e-01
7.57063687e-01 2.65737981e-01 -3.51488024e-01 2.88843513e-01
2.61761636e-01 4.29867022e-02 5.09351134e-01 6.69219255e-01
1.49640977e+00 -5.79467893e-01 -1.48726869e-02 -2.03026056e-01
4.06998903e-01 2.66199589e-01 3.56402010e-01 8.61580491e-01
2.03832835e-01 4.08146739e-01 4.81516361e-01 -2.50578523e-01
-8.71894121e-01 -1.16620696e+00 -2.19867781e-01 8.63326788e-01
-1.72522798e-01 -5.18275440e-01 -4.57860351e-01 -5.11242151e-01
4.01134282e-01 7.93729782e-01 -5.52441835e-01 -2.76500821e-01
-4.09776449e-01 -5.08492291e-01 9.19533074e-01 2.06562385e-01
-6.94699958e-03 -9.34496224e-01 -3.23206633e-01 3.70929480e-01
-6.29899442e-01 -1.06765664e+00 -7.56411910e-01 7.14296922e-02
-5.90619802e-01 -1.09341383e+00 -1.00167596e+00 -7.32950568e-01
2.12931558e-01 -1.76508315e-02 1.38469303e+00 -1.85611784e-01
-3.44685644e-01 3.49223405e-01 -2.83260792e-01 -4.83237833e-01
-2.24381939e-01 3.59672338e-01 -2.44532041e-02 -4.75278020e-01
5.19473374e-01 -2.32709162e-02 -7.83233643e-01 -1.90677300e-01
-9.39982116e-01 -3.84331137e-01 5.43898642e-01 8.80507708e-01
6.79945111e-01 -4.42435235e-01 1.09473658e+00 -6.89423919e-01
1.15220356e+00 -1.04977393e+00 -5.48723876e-01 5.04427969e-01
-9.56538677e-01 3.29446167e-01 2.99734920e-01 1.31670862e-01
-7.10612714e-01 -7.22712040e-01 -1.53145105e-01 -4.55567002e-01
4.40025866e-01 1.15914440e+00 4.18827593e-01 6.72722101e-01
6.37233675e-01 1.64260477e-01 4.93848436e-02 -5.82178712e-01
5.41376531e-01 1.05407441e+00 6.67920470e-01 -2.85680026e-01
4.05298859e-01 3.72158065e-02 -7.44525716e-02 -6.40775144e-01
-1.11100054e+00 -8.61095548e-01 5.03726482e-01 9.48046073e-02
1.28513360e+00 -1.18735480e+00 -8.58517230e-01 -1.51046842e-01
-1.64903581e+00 4.39674318e-01 -2.01842301e-02 2.85537571e-01
-2.35971659e-01 -6.27500415e-02 -7.85680056e-01 -1.83982745e-01
-1.09671283e+00 -1.16300774e+00 1.34651279e+00 1.79742113e-01
-3.10786009e-01 -8.56779695e-01 5.21541297e-01 6.94975197e-01
6.23334289e-01 8.10920075e-02 1.31049120e+00 -1.09252453e+00
-4.29013491e-01 -6.09680638e-02 -4.80690509e-01 1.72943488e-01
-4.67822105e-02 -5.14586926e-01 -4.97149825e-01 -2.92272836e-01
-4.05466497e-01 -3.37676257e-01 1.28109384e+00 4.42176819e-01
1.11684918e+00 -6.54922724e-01 -5.79008877e-01 4.19816554e-01
1.28124106e+00 3.81634921e-01 3.04904163e-01 6.92402348e-02
4.07165706e-01 3.61308604e-01 2.26918563e-01 3.12582135e-01
7.30331898e-01 4.40584034e-01 1.76147312e-01 3.67591619e-01
-1.58113185e-02 -3.57119322e-01 -8.28096494e-02 1.11903477e+00
2.73362935e-01 -7.94875503e-01 -1.00137115e+00 6.74357295e-01
-1.41984141e+00 -9.86606956e-01 3.20792764e-01 1.78621244e+00
1.09742785e+00 -1.16915144e-01 -3.74657661e-01 -5.73253989e-01
4.52790171e-01 3.17962736e-01 -6.88299179e-01 -6.47611380e-01
-2.00995252e-01 3.05785120e-01 5.73141634e-01 5.16958237e-01
-7.98990548e-01 5.75062990e-01 6.47550058e+00 7.35070407e-01
-9.17653918e-01 7.00433999e-02 5.92030883e-01 -5.99337928e-02
-8.18547964e-01 -3.37611705e-01 -7.79650748e-01 6.69779062e-01
1.36553693e+00 -7.37902939e-01 4.63634491e-01 4.15380687e-01
-1.55315250e-01 3.26599717e-01 -8.87949347e-01 8.34440291e-01
2.74495274e-01 -2.07176113e+00 3.86525571e-01 -6.60750270e-02
7.74875104e-01 6.82176650e-01 6.47784844e-02 5.27483225e-01
1.68250099e-01 -1.11914766e+00 1.04574502e-01 1.00963569e+00
1.00013840e+00 -7.87271976e-01 9.18070495e-01 1.19265370e-01
-8.11859012e-01 -4.84071672e-02 -1.86791107e-01 7.62982070e-01
2.30443045e-01 8.13598633e-01 -1.01149559e+00 8.46417904e-01
6.40697002e-01 7.26406753e-01 -2.93784648e-01 8.94963443e-01
1.72871754e-01 5.77809930e-01 -7.60198683e-02 -3.34141344e-01
2.95299321e-01 1.79862171e-01 8.40789258e-01 1.43198156e+00
3.22516292e-01 2.21489176e-01 -9.72577631e-02 8.58012080e-01
-7.23908663e-01 -3.24512273e-03 -5.58311760e-01 -7.51925528e-01
7.15865493e-01 7.73404837e-01 -2.20911235e-01 -6.09210074e-01
-5.97166829e-03 7.57711470e-01 1.64814010e-01 4.95367765e-01
-6.30242825e-01 -6.58644140e-01 7.03595877e-01 -6.95115775e-02
4.50015575e-01 1.64033309e-01 -9.23941731e-02 -1.03565061e+00
-2.68351167e-01 -1.01150465e+00 8.16397071e-01 -8.72125149e-01
-1.50629985e+00 7.02023506e-01 -8.41085911e-02 -5.95448315e-01
-4.42427993e-01 -2.41770059e-01 -1.90710142e-01 1.15251040e+00
-1.42790723e+00 -6.89252436e-01 3.10465723e-01 8.93932134e-02
4.73302573e-01 -3.60597372e-01 9.12443876e-01 5.28221488e-01
-2.97132462e-01 3.02459180e-01 3.22256356e-01 -1.09820403e-01
7.06734776e-01 -9.30540383e-01 4.05981004e-01 2.80960560e-01
-3.42173547e-01 1.05973041e+00 7.17477918e-01 -1.01961601e+00
-1.66161931e+00 -1.33190894e+00 1.64673007e+00 -7.87966490e-01
7.03197837e-01 -1.51048332e-01 -9.22313690e-01 3.25976342e-01
3.43211651e-01 -4.44990098e-01 6.35407269e-01 -1.95884295e-02
-5.13178527e-01 -1.21200949e-01 -1.09743738e+00 3.77649307e-01
6.98234558e-01 -7.26155937e-01 -1.04867792e+00 6.26478136e-01
1.54531097e+00 -2.84801930e-01 -1.08278620e+00 5.34207821e-01
3.90711844e-01 -5.78380637e-02 1.17763698e+00 -1.01917064e+00
7.71079719e-01 -3.32862362e-02 -2.41217390e-01 -1.42134047e+00
-4.82538432e-01 -3.82086188e-01 -3.76253396e-01 7.04588234e-01
7.15873420e-01 -5.43461442e-01 2.43519560e-01 -4.59395684e-02
-1.21632881e-01 -1.01666033e+00 -8.79156053e-01 -5.61318517e-01
3.29637378e-02 -2.72748042e-02 6.75846159e-01 7.14884698e-01
-1.37764355e-02 7.34635115e-01 -1.02774397e-01 2.39144769e-02
4.67375308e-01 1.73893586e-01 6.04670085e-02 -1.39991593e+00
-1.71547979e-01 -6.06659353e-01 -7.31670037e-02 -9.71611023e-01
5.43151656e-03 -1.28816879e+00 -9.29434448e-02 -2.09894633e+00
4.48861122e-01 -1.09058194e-01 -6.36342347e-01 3.22544366e-01
-3.20011616e-01 -1.91116378e-01 -1.86953381e-01 1.07272126e-01
-8.10507417e-01 3.30933809e-01 1.09770131e+00 -6.02403045e-01
7.62071162e-02 -4.64187413e-01 -1.03679180e+00 2.97333556e-03
4.25982982e-01 -6.72991872e-01 -4.44635808e-01 -6.32570386e-01
8.19840491e-01 5.18974125e-01 3.28994058e-02 -5.22710741e-01
4.57779408e-01 -6.60194904e-02 2.60251671e-01 -1.02371526e+00
1.02060266e-01 -3.34460765e-01 1.15672357e-01 5.01557231e-01
-7.88273513e-01 5.28568089e-01 2.52448201e-01 7.89640367e-01
-2.22178265e-01 2.49065775e-02 2.56519943e-01 1.00903645e-01
-4.35691103e-02 4.51477170e-01 -4.54374194e-01 7.56908000e-01
3.87239516e-01 6.62413776e-01 -9.72407818e-01 -2.88251549e-01
-2.42306516e-01 7.54418969e-01 -1.08241595e-01 8.91296387e-01
6.81462944e-01 -1.07304037e+00 -1.13670337e+00 -2.14634091e-02
2.99041063e-01 -1.74228340e-01 4.14088935e-01 5.94436884e-01
-4.83525097e-01 1.22349441e+00 4.16479856e-02 -2.66475111e-01
-5.79866886e-01 7.69570172e-01 -1.72447532e-01 -8.35736215e-01
-2.57971674e-01 7.28899717e-01 -1.08240455e-01 -1.95199907e-01
2.26612583e-01 -2.35154107e-01 -3.92082542e-01 3.15250576e-01
7.57853866e-01 5.28147042e-01 2.96816051e-01 -4.07518178e-01
-6.65458739e-01 1.73492163e-01 -1.73150674e-01 -1.03433944e-01
1.44997811e+00 1.75455093e-01 -3.58750850e-01 6.38660267e-02
1.76351857e+00 -1.23560935e-01 -9.70598459e-02 -2.90121019e-01
1.65450320e-01 2.09193215e-01 1.68547451e-01 -1.27422953e+00
-7.53022134e-01 4.61332828e-01 2.31941566e-01 7.84019157e-02
8.90441597e-01 2.43982971e-01 1.14551187e+00 5.57665467e-01
-1.99981183e-02 -8.62609208e-01 -2.82170046e-02 7.65530229e-01
1.16355562e+00 -9.14277494e-01 -2.26804167e-02 3.37569088e-01
-5.22566319e-01 6.07870817e-01 -5.47726965e-03 2.96892107e-01
6.19455457e-01 4.11156751e-03 -2.00307801e-01 -7.84315884e-01
-1.25962532e+00 1.12338267e-01 6.67634666e-01 1.80798396e-01
2.41008803e-01 1.72410116e-01 -6.21303260e-01 8.12900186e-01
-1.96198635e-02 4.63994257e-02 8.71702656e-02 6.19933546e-01
-4.53995287e-01 -7.26832092e-01 4.46578637e-02 1.02523601e+00
-7.62772620e-01 -4.51302409e-01 -1.09049574e-01 -6.72370195e-02
-1.30994827e-01 1.14598608e+00 2.08740965e-01 -3.17601025e-01
2.96056956e-01 1.32817402e-01 -1.02866001e-01 -4.93820310e-01
-6.22576475e-01 -3.14823985e-01 3.54287297e-01 -4.69992757e-01
1.12354279e-01 -4.28923339e-01 -1.19590807e+00 -1.30549133e-01
8.75914097e-02 7.18353927e-01 9.31480408e-01 5.60107708e-01
1.07388890e+00 6.81335986e-01 3.97302389e-01 -3.34991515e-02
-7.04511166e-01 -9.24412131e-01 -1.71702281e-01 2.09075764e-01
5.85066319e-01 -3.99491996e-01 -2.20839068e-01 -2.26899177e-01] | [8.691701889038086, 8.709358215332031] |
93a7cd5c-bec7-4092-8f5c-08ad6705f2df | deriving-explanation-of-deep-visual-saliency | 2109.03575 | null | https://arxiv.org/abs/2109.03575v1 | https://arxiv.org/pdf/2109.03575v1.pdf | Deriving Explanation of Deep Visual Saliency Models | Deep neural networks have shown their profound impact on achieving human level performance in visual saliency prediction. However, it is still unclear how they learn the task and what it means in terms of understanding human visual system. In this work, we develop a technique to derive explainable saliency models from their corresponding deep neural architecture based saliency models by applying human perception theories and the conventional concepts of saliency. This technique helps us understand the learning pattern of the deep network at its intermediate layers through their activation maps. Initially, we consider two state-of-the-art deep saliency models, namely UNISAL and MSI-Net for our interpretation. We use a set of biologically plausible log-gabor filters for identifying and reconstructing the activation maps of them using our explainable saliency model. The final saliency map is generated using these reconstructed activation maps. We also build our own deep saliency model named cross-concatenated multi-scale residual block based network (CMRNet) for saliency prediction. Then, we evaluate and compare the performance of the explainable models derived from UNISAL, MSI-Net and CMRNet on three benchmark datasets with other state-of-the-art methods. Hence, we propose that this approach of explainability can be applied to any deep visual saliency model for interpretation which makes it a generic one. | ['Santanu Chaudhury', 'Chaker Larabi', 'Jayanta Mukhopadhyay', 'Sai Phani Kumar Malladi'] | 2021-09-08 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [ 4.05270785e-01 5.57166040e-01 -5.34130037e-02 -2.42074832e-01
1.23354107e-01 -4.67640981e-02 6.56201303e-01 -2.74777338e-02
1.26046941e-01 7.23753154e-01 3.24392706e-01 -2.42086798e-01
-1.10678010e-01 -5.63888073e-01 -1.07810235e+00 -4.24816132e-01
4.24673930e-02 9.83769223e-02 8.32619607e-01 -5.16089916e-01
6.05332434e-01 1.69685215e-01 -2.10203505e+00 6.00606322e-01
1.04122066e+00 9.75686491e-01 7.19760716e-01 5.66510260e-01
1.64057408e-02 9.82324600e-01 -3.81989568e-01 -1.38005331e-01
-1.38231844e-01 -6.57417119e-01 -1.06512499e+00 -1.89063892e-01
3.61271560e-01 -1.07626822e-02 -1.61052495e-01 1.00470388e+00
1.31471798e-01 2.99225952e-02 7.41299689e-01 -1.47145414e+00
-1.17728996e+00 6.18746221e-01 -5.14708042e-01 5.24526060e-01
-7.95283820e-03 -2.28613392e-02 1.03785062e+00 -1.04168510e+00
4.78733033e-01 1.34610200e+00 4.87298518e-01 6.82749391e-01
-1.07826519e+00 -1.88600287e-01 2.84260809e-01 5.95826447e-01
-9.60625589e-01 -1.41657203e-01 8.87930393e-01 -2.57585764e-01
9.30951655e-01 4.54767674e-01 8.77504885e-01 9.47572708e-01
5.07291019e-01 1.19885063e+00 1.36219442e+00 -3.99996161e-01
2.21460238e-01 1.55150026e-01 1.67509302e-01 7.62091756e-01
1.82514191e-01 1.48935914e-01 -8.87564123e-01 3.76026809e-01
1.09727955e+00 1.28201202e-01 -2.42897406e-01 -4.69698310e-01
-1.32048464e+00 7.32188582e-01 1.36608982e+00 3.70691419e-01
-5.65209568e-01 4.29419369e-01 -8.33039209e-02 -2.72813708e-01
3.86986673e-01 5.62833071e-01 -3.75977397e-01 6.72974706e-01
-1.09523106e+00 3.44837844e-01 1.21900357e-01 6.65569365e-01
1.03339267e+00 4.68588799e-01 -2.39472032e-01 4.93551016e-01
4.97979313e-01 2.49862239e-01 7.20483661e-01 -6.37111187e-01
-1.81764081e-01 7.73371935e-01 3.86674851e-02 -1.25505328e+00
-5.93532860e-01 -6.96969330e-01 -7.45086789e-01 4.47143495e-01
3.40235308e-02 3.48815769e-01 -1.36036646e+00 1.70595098e+00
-2.42513698e-02 6.32448971e-01 2.25921944e-01 1.42678297e+00
1.28094995e+00 4.85073864e-01 3.48124027e-01 4.20827597e-01
1.29907811e+00 -1.37539816e+00 -3.75858724e-01 -6.21282279e-01
6.68953508e-02 -5.98923206e-01 1.08621609e+00 -9.41376574e-03
-1.06976676e+00 -9.15160954e-01 -1.14185369e+00 -3.72070968e-01
-6.10111296e-01 1.72175780e-01 9.19092119e-01 5.77573664e-02
-1.56378376e+00 5.91480196e-01 -6.07918561e-01 -3.45108151e-01
6.39497161e-01 4.35231209e-01 -3.93091738e-02 5.90393305e-01
-1.23542523e+00 1.23422658e+00 6.00188911e-01 1.54508695e-01
-1.14636362e+00 -6.37466252e-01 -8.38575304e-01 4.57242668e-01
-4.07008529e-02 -9.16363060e-01 1.06621957e+00 -1.57189858e+00
-8.72318685e-01 9.11093771e-01 -5.42678177e-01 -8.14677417e-01
-1.43869847e-01 -5.03498549e-03 -8.25782940e-02 1.31971195e-01
2.13071018e-01 1.45006001e+00 1.02200854e+00 -1.56914961e+00
-7.67438829e-01 -4.53059152e-02 1.78027675e-01 1.89373970e-01
1.89016536e-01 -2.17998937e-01 5.01443027e-03 -6.84092760e-01
3.81055474e-01 -8.54706407e-01 -1.32861361e-01 -4.76393066e-02
-5.34457445e-01 -1.89597622e-01 8.11705828e-01 -6.72326624e-01
6.77608490e-01 -1.90281880e+00 3.23419541e-01 -7.23944977e-02
6.65294647e-01 4.14534897e-01 -1.62271634e-01 -1.92466795e-01
-3.58481497e-01 1.96926609e-01 -3.92650157e-01 -2.15102673e-01
-1.63206905e-01 9.58838314e-02 -6.50271177e-01 6.13933317e-02
6.27583146e-01 1.39529371e+00 -9.81251717e-01 -4.42485809e-01
4.80690926e-01 5.76133072e-01 -5.40295899e-01 7.09218904e-02
-2.78487533e-01 5.32992542e-01 -3.98671657e-01 6.18097246e-01
5.57373166e-01 -3.22967798e-01 -2.21082076e-01 -2.01506197e-01
-1.18568651e-01 2.23250166e-01 -6.52450085e-01 1.67403007e+00
-1.77364677e-01 1.04324996e+00 -6.43595457e-01 -1.09212017e+00
1.13418698e+00 2.26014983e-02 -7.01879710e-02 -8.22826684e-01
5.64563163e-02 3.46658558e-01 2.09963888e-01 -5.58981858e-02
6.86171412e-01 -1.12514086e-01 2.80133188e-01 1.93956196e-01
3.71576786e-01 1.48166537e-01 -2.60178566e-01 1.82347074e-01
4.86877680e-01 2.52459526e-01 3.02002758e-01 -7.00646877e-01
6.02868021e-01 1.28317162e-01 2.93995261e-01 7.13617623e-01
-4.18083191e-01 1.06683338e+00 4.95395809e-01 -8.24040413e-01
-1.14597499e+00 -1.24821079e+00 3.92671935e-02 1.15687299e+00
7.12134182e-01 1.34614348e-01 -9.87924457e-01 -4.71332461e-01
-1.84008449e-01 8.10083389e-01 -1.14152873e+00 -4.42108214e-01
-3.11048001e-01 -5.70074201e-01 1.97015777e-01 5.67780495e-01
7.55141139e-01 -1.84970069e+00 -1.09448922e+00 1.13057122e-02
-2.36148283e-01 -6.11955941e-01 -2.88880076e-02 1.12294175e-01
-8.29813898e-01 -1.11579883e+00 -6.59723997e-01 -1.12512732e+00
7.18428850e-01 7.50346184e-01 1.32614362e+00 5.62865019e-01
-2.75783867e-01 -9.42121446e-02 -1.94675937e-01 -8.33801031e-01
-1.10471256e-01 1.26054168e-01 -1.34925038e-01 3.63169648e-02
3.33377987e-01 -4.08518136e-01 -9.57236826e-01 3.23778301e-01
-1.02760422e+00 7.51684427e-01 6.23594642e-01 6.69911504e-01
7.99595594e-01 -4.47674215e-01 9.33781624e-01 -5.73996723e-01
3.72208267e-01 -6.37158930e-01 -1.37049079e-01 3.49073976e-01
-3.96672875e-01 4.39865559e-01 3.00297916e-01 -5.46919107e-02
-1.08674288e+00 5.12882285e-02 7.39232898e-02 -3.75645757e-01
-2.50430316e-01 3.00588995e-01 1.91388950e-01 -2.12526917e-01
7.14657485e-01 5.48825502e-01 -2.70788699e-01 -2.93963730e-01
4.16163653e-01 3.96606416e-01 8.20877731e-01 -2.47519985e-02
5.57302713e-01 4.60220039e-01 6.45036111e-03 -5.33409536e-01
-1.26851916e+00 -1.84402049e-01 -7.00649202e-01 -2.96987772e-01
1.06915891e+00 -7.81810164e-01 -5.00706315e-01 4.50880319e-01
-1.37975240e+00 -1.78063795e-01 -2.02389404e-01 6.77515268e-02
-7.76663840e-01 -8.61897394e-02 -1.07053556e-01 -5.04208565e-01
-4.37410384e-01 -1.26735079e+00 1.22247636e+00 7.34174967e-01
-2.48551086e-01 -1.21107852e+00 -1.96201429e-01 1.77438170e-01
6.52462482e-01 3.81464005e-01 8.20516229e-01 -5.23460209e-01
-8.86089742e-01 6.13875389e-01 -6.90429389e-01 -1.78982317e-02
5.66507280e-02 -1.76308587e-01 -1.19941103e+00 1.00225158e-01
-1.16994649e-01 -2.41940260e-01 1.35737562e+00 8.57832670e-01
1.36545455e+00 -2.16200620e-01 -3.56941670e-01 6.45129204e-01
1.50261509e+00 -1.93910792e-01 9.68557894e-01 6.40485346e-01
8.16404819e-01 6.94561064e-01 5.89461207e-01 -1.85504287e-01
4.32983637e-01 4.88370985e-01 1.07418871e+00 -6.11455917e-01
-4.80420530e-01 -4.90249872e-01 -3.16711664e-02 4.30687755e-01
-3.06304455e-01 8.28710869e-02 -8.54911745e-01 1.00953293e+00
-2.19535971e+00 -1.00226665e+00 -4.25011098e-01 1.73515868e+00
5.62690020e-01 1.95438087e-01 1.57583468e-02 -6.58190949e-03
7.94458628e-01 2.53798902e-01 -6.73348010e-01 -8.20614398e-01
-4.21419740e-01 3.56899887e-01 2.17827186e-01 4.92712706e-01
-1.10679901e+00 1.45282793e+00 6.46810198e+00 6.22460842e-01
-1.37809622e+00 9.15022865e-02 9.04754698e-01 3.24062258e-01
-5.68215668e-01 6.24917001e-02 -6.54108405e-01 3.62996459e-01
9.10696447e-01 -3.33694629e-02 4.32482570e-01 1.02260244e+00
8.47471282e-02 -2.93158501e-01 -9.01440382e-01 7.04720557e-01
2.30968311e-01 -1.78021789e+00 4.08007085e-01 -3.04859966e-01
9.86035883e-01 1.00135848e-01 4.30252373e-01 6.30487055e-02
-7.40238559e-03 -1.52758157e+00 1.09340394e+00 7.73972154e-01
3.38086307e-01 -5.48031509e-01 6.90324187e-01 1.90621376e-01
-1.01448512e+00 1.66658498e-03 -8.14111590e-01 -1.84250638e-01
-2.05932185e-01 2.34259635e-01 -9.83569682e-01 1.62441835e-01
7.57773817e-01 9.14475501e-01 -1.08328569e+00 1.20264089e+00
-5.66567004e-01 4.72570866e-01 1.68672487e-01 -1.34223327e-01
4.58248585e-01 3.26701492e-01 4.95719850e-01 1.01961660e+00
2.01505974e-01 -2.43151188e-01 -4.01923925e-01 1.55943847e+00
2.02759326e-01 -3.19969535e-01 -5.17969847e-01 5.07149041e-01
2.68655479e-01 1.18080032e+00 -1.02470684e+00 -4.28776175e-01
-7.07895681e-02 9.09707665e-01 3.87839347e-01 3.46944898e-01
-1.02463973e+00 -1.41287163e-01 7.73860335e-01 8.14972520e-02
5.10859489e-01 2.10375845e-01 -8.02671373e-01 -1.08056176e+00
-3.61734927e-01 -6.54729128e-01 -2.19103023e-01 -1.44278121e+00
-7.31491983e-01 8.96268308e-01 -5.83466627e-02 -9.21422899e-01
-1.37869745e-01 -5.92626095e-01 -8.60859394e-01 1.07433426e+00
-2.01898694e+00 -1.55718172e+00 -4.61710215e-01 6.46899045e-01
8.30641031e-01 -1.85696349e-01 5.65974236e-01 -3.58522654e-01
-9.20519009e-02 1.52318746e-01 -3.50991249e-01 -1.93043157e-01
1.17730364e-01 -1.37220848e+00 8.82685542e-01 8.97817612e-01
4.48791265e-01 5.70777893e-01 1.04330051e+00 -4.81067866e-01
-7.86822736e-01 -1.16626787e+00 9.10119355e-01 -6.28570974e-01
4.60366458e-01 -2.53749609e-01 -1.25517976e+00 6.70103848e-01
4.01536494e-01 -7.81251490e-02 1.25663280e-01 -1.02985322e-01
-1.03301788e-02 1.34579659e-01 -9.36105371e-01 6.98334575e-01
1.06197393e+00 -3.39901060e-01 -1.11771858e+00 1.26157448e-01
9.63515341e-01 -2.26061955e-01 -5.25599308e-02 3.92380565e-01
3.77321213e-01 -1.38912475e+00 1.21398580e+00 -7.00426042e-01
8.92409801e-01 -5.74842930e-01 1.26441136e-01 -1.46882248e+00
-6.60068750e-01 -6.76457509e-02 4.16142903e-02 7.78426230e-01
4.61329520e-01 -5.58387160e-01 7.30217755e-01 2.33423308e-01
-4.96456802e-01 -9.11945820e-01 -7.72798955e-01 -2.01081127e-01
-2.08414495e-01 -3.69970351e-02 6.83609486e-01 5.87880373e-01
-2.55136520e-01 3.12234998e-01 -4.39867258e-01 2.78197318e-01
7.01727152e-01 3.33224356e-01 3.03434104e-01 -1.27026117e+00
3.20936516e-02 -3.87852103e-01 -7.20832765e-01 -7.43824840e-01
2.97287792e-01 -8.11941564e-01 1.04523122e-01 -1.72190738e+00
4.55594361e-01 -2.01789275e-01 -7.07905769e-01 7.44213104e-01
-4.03745949e-01 5.39561629e-01 4.73731637e-01 3.45170885e-01
-6.86355472e-01 6.11318827e-01 1.46418595e+00 -8.22118968e-02
-1.37437165e-01 -2.33762383e-01 -1.11446750e+00 9.53456521e-01
1.00254619e+00 -3.27960074e-01 -4.88030761e-01 -4.40181106e-01
4.82173935e-02 -3.95401984e-01 1.11174083e+00 -1.24666011e+00
2.09618192e-02 -2.86284238e-01 5.49778104e-01 -6.08432591e-01
6.54339045e-02 -3.93497616e-01 -2.18494702e-02 6.27040505e-01
-4.60530221e-01 -1.65093958e-01 3.73712420e-01 3.55946273e-01
-3.93480361e-01 -1.34873986e-01 8.17251682e-01 -2.67280638e-01
-1.34210253e+00 -3.74804474e-02 -2.89608270e-01 -1.95333511e-01
8.86585593e-01 -4.59806591e-01 -5.51186144e-01 -2.70236343e-01
-8.46358776e-01 -1.15297593e-01 4.18307483e-01 7.69203126e-01
1.20620847e+00 -1.36652243e+00 -5.65034270e-01 2.97423184e-01
-7.43447095e-02 -1.39954194e-01 2.45596856e-01 6.62590802e-01
-5.68431139e-01 7.73892581e-01 -9.61021245e-01 -8.37871909e-01
-1.05872166e+00 6.73750579e-01 4.54687208e-01 1.59510031e-01
-2.25692406e-01 9.16764259e-01 9.71653044e-01 -2.79476613e-01
-2.51561522e-01 -5.64699054e-01 -6.47039890e-01 -4.70676035e-01
4.17223334e-01 -3.60633172e-02 -2.36154169e-01 -9.62671757e-01
-4.36676145e-01 5.16427875e-01 2.31354102e-01 3.03659797e-01
1.37746096e+00 -1.65458426e-01 -2.20281482e-01 3.51894826e-01
8.41469884e-01 -5.29286742e-01 -1.48044062e+00 7.38749504e-02
-1.08772121e-01 -3.27285022e-01 8.30929279e-02 -8.09241116e-01
-1.18744147e+00 1.24383044e+00 7.53946900e-01 1.26904100e-01
1.23631155e+00 2.49171257e-01 6.22301996e-01 -7.47825280e-02
2.22643808e-01 -6.52920544e-01 2.83871084e-01 4.63812500e-01
1.29608548e+00 -1.46224511e+00 -2.20024377e-01 -3.76378626e-01
-9.57491100e-01 8.07786644e-01 7.52094388e-01 -6.13486826e-01
4.95766103e-01 -5.86306572e-01 -1.68220893e-01 -3.27399671e-01
-6.96712554e-01 -5.70365548e-01 9.01839375e-01 7.86401689e-01
3.04835647e-01 2.64597535e-01 -1.12558007e-01 6.70355856e-01
-3.50286871e-01 -3.82798016e-02 6.40055239e-01 5.36541283e-01
-1.05607605e+00 -5.42114854e-01 -2.72591293e-01 2.33255938e-01
-1.94659427e-01 -4.58252013e-01 -6.25143170e-01 5.95856547e-01
3.67657661e-01 6.72167242e-01 1.76152249e-03 -3.93271595e-01
-3.80473100e-02 -1.74747482e-01 1.17269933e-01 -7.10227907e-01
-4.63175237e-01 -3.54147702e-01 -3.55832875e-01 -4.26559925e-01
-7.19937861e-01 -3.61908197e-01 -1.54939723e+00 -1.02361023e-01
-5.96415326e-02 -1.36840194e-01 6.46848977e-01 1.13603294e+00
4.85220820e-01 7.71017075e-01 2.88026571e-01 -1.21892786e+00
7.95883164e-02 -8.04166615e-01 -4.46863413e-01 4.90853697e-01
5.67715108e-01 -1.03233135e+00 -1.69353366e-01 2.60924578e-01] | [10.055135726928711, 1.61247718334198] |
3279b072-074d-4e8d-bf3e-5d0932074586 | chatgpt-is-not-enough-enhancing-large | 2306.11489 | null | https://arxiv.org/abs/2306.11489v1 | https://arxiv.org/pdf/2306.11489v1.pdf | ChatGPT is not Enough: Enhancing Large Language Models with Knowledge Graphs for Fact-aware Language Modeling | Recently, ChatGPT, a representative large language model (LLM), has gained considerable attention due to its powerful emergent abilities. Some researchers suggest that LLMs could potentially replace structured knowledge bases like knowledge graphs (KGs) and function as parameterized knowledge bases. However, while LLMs are proficient at learning probabilistic language patterns based on large corpus and engaging in conversations with humans, they, like previous smaller pre-trained language models (PLMs), still have difficulty in recalling facts while generating knowledge-grounded contents. To overcome these limitations, researchers have proposed enhancing data-driven PLMs with knowledge-based KGs to incorporate explicit factual knowledge into PLMs, thus improving their performance to generate texts requiring factual knowledge and providing more informed responses to user queries. This paper reviews the studies on enhancing PLMs with KGs, detailing existing knowledge graph enhanced pre-trained language models (KGPLMs) as well as their applications. Inspired by existing studies on KGPLM, this paper proposes to enhance LLMs with KGs by developing knowledge graph-enhanced large language models (KGLLMs). KGLLM provides a solution to enhance LLMs' factual reasoning ability, opening up new avenues for LLM research. | ['Xindong Wu', 'Xiao Ding', 'Zhao Li', 'Hongyang Chen', 'Linyao Yang'] | 2023-06-20 | null | null | null | null | ['knowledge-graphs'] | ['knowledge-base'] | [-3.25029224e-01 9.95120108e-01 -4.30540770e-01 5.40869758e-02
-4.91832912e-01 -5.28752089e-01 7.70806253e-01 2.76614904e-01
-1.42904043e-01 1.04360747e+00 5.28052866e-01 -5.34411371e-01
-1.97053030e-01 -1.44392335e+00 -7.58756220e-01 -3.58220190e-02
-7.44709745e-02 7.45434642e-01 4.98081565e-01 -2.67842174e-01
3.69828552e-01 1.78057238e-01 -1.02533793e+00 6.83686852e-01
1.32014275e+00 2.53579468e-01 6.28526330e-01 4.43268567e-01
-9.43528712e-01 1.64856076e+00 -7.60852277e-01 -1.04218030e+00
-4.96384472e-01 -3.57173055e-01 -1.40687370e+00 -3.58674556e-01
9.74544734e-02 -6.33259267e-02 -5.00992179e-01 7.21064210e-01
2.85409629e-01 4.47005421e-01 5.73209941e-01 -1.35897613e+00
-1.44990933e+00 1.64091277e+00 5.07377125e-02 -9.64170024e-02
6.75036252e-01 5.85871562e-02 9.37628150e-01 -8.29588175e-01
7.93405890e-01 1.69169033e+00 9.02772963e-01 6.23465657e-01
-8.28940034e-01 -4.08315092e-01 3.63951027e-01 5.50759614e-01
-1.56699717e+00 2.93294201e-03 6.46188915e-01 -1.21786229e-01
1.77689707e+00 -5.70382066e-02 8.97253335e-01 1.11578310e+00
1.23524420e-01 1.28651476e+00 9.80307937e-01 -7.63631403e-01
1.36783287e-01 5.74763477e-01 1.10902265e-01 1.01519239e+00
5.82627356e-01 -2.96917737e-01 -9.85297441e-01 -3.32023919e-01
9.41651225e-01 -4.20965463e-01 -1.40949920e-01 7.99110066e-03
-1.26861644e+00 8.13468516e-01 1.05692908e-01 5.89482009e-01
-4.94954646e-01 -7.40299793e-03 3.21147144e-02 1.13236308e-01
3.21021378e-01 7.05029070e-01 -5.04834294e-01 -2.00100094e-01
-7.02799559e-01 3.07615429e-01 1.23381793e+00 1.31339669e+00
7.83894598e-01 2.31184721e-01 -3.67938906e-01 9.77295816e-01
5.09358883e-01 6.81354284e-01 6.78585887e-01 -8.80848885e-01
5.19792020e-01 1.09730291e+00 -2.95420326e-02 -1.20164919e+00
-3.62055898e-01 -2.53655046e-01 -6.70007586e-01 -8.42735171e-01
2.95948461e-02 -8.14897269e-02 -5.64198136e-01 1.79639971e+00
1.88549049e-02 3.66860420e-01 7.20501184e-01 2.36952871e-01
1.26899874e+00 1.15101337e+00 6.20903611e-01 -2.11732015e-01
1.14165092e+00 -8.53038013e-01 -7.89778590e-01 -3.85002136e-01
9.81163979e-01 -2.92899191e-01 1.34215045e+00 3.91165704e-01
-1.06134832e+00 -3.95196587e-01 -5.28447449e-01 7.60867540e-03
-8.12216520e-01 -1.82243496e-01 1.02235711e+00 6.62616611e-01
-1.39515615e+00 1.48540616e-01 -4.12566781e-01 -6.39584243e-01
2.23088041e-01 -1.44932957e-04 2.37158891e-02 -2.99604744e-01
-2.11686373e+00 1.29279649e+00 1.28093493e+00 -1.15405917e-01
-8.86163771e-01 -6.89907551e-01 -1.08908522e+00 5.94066922e-03
7.49498129e-01 -1.10265112e+00 1.30790615e+00 -3.27308238e-01
-1.59540284e+00 5.30937195e-01 -1.77710116e-01 -5.54738700e-01
-1.13068067e-01 -2.06096731e-02 -6.12741530e-01 1.60448775e-01
2.27422506e-01 6.20776355e-01 4.02430713e-01 -1.43451703e+00
-4.18594658e-01 2.94208407e-01 5.86703002e-01 4.37766254e-01
-4.50431943e-01 -1.92879081e-01 -3.71727884e-01 -5.98894954e-01
-3.37594837e-01 -4.22950655e-01 -2.13125497e-01 -9.20325220e-01
-3.50577325e-01 -7.49064982e-01 5.70672333e-01 -4.89425242e-01
1.72146547e+00 -1.33462548e+00 -1.99297011e-01 2.36042663e-01
1.66697115e-01 7.74958789e-01 -3.74521196e-01 1.18072534e+00
6.07968092e-01 3.82712215e-01 9.42195952e-02 -1.30504416e-02
6.69995397e-02 6.79989755e-01 -6.01751685e-01 -7.44248986e-01
6.37243465e-02 1.67253160e+00 -1.34938121e+00 -8.99770319e-01
2.46974885e-01 5.22308350e-02 -4.57706928e-01 2.13304222e-01
-7.94764102e-01 -7.28316829e-02 -6.02160811e-01 5.42209446e-01
3.02271098e-01 -6.51586115e-01 4.01967674e-01 2.45230228e-01
1.89029872e-01 4.12450641e-01 -1.12001908e+00 1.54561818e+00
-8.52182806e-01 3.29074711e-01 -4.99604255e-01 -8.09739292e-01
7.77092695e-01 5.75719416e-01 -1.15059786e-01 -2.30692893e-01
-2.17261925e-01 -1.24354642e-02 -2.55067617e-01 -7.72917509e-01
7.55327880e-01 -4.14764553e-01 -8.52518599e-04 4.56255674e-01
3.77252907e-01 -5.27762651e-01 3.54195356e-01 1.00307095e+00
7.92332232e-01 -3.44074331e-02 6.62660062e-01 -1.07342303e-01
5.99123895e-01 2.03148708e-01 1.90513775e-01 1.18499684e+00
2.34248295e-01 -3.06281328e-01 2.75304288e-01 -5.24301641e-03
-2.75298089e-01 -1.18637621e+00 4.44683731e-01 8.53654265e-01
5.46715595e-02 -1.02289212e+00 -6.07297480e-01 -6.46298349e-01
-1.74866572e-01 1.55247629e+00 -1.31798059e-01 -2.48324797e-01
-3.36418092e-01 -6.80436492e-01 1.08158481e+00 6.10305846e-01
8.58758390e-01 -1.61280215e+00 4.15295847e-02 4.34233099e-01
-7.06675410e-01 -1.17450631e+00 4.46540974e-02 -3.53117079e-01
-7.92769194e-01 -9.03133810e-01 -2.99586356e-01 -8.78486276e-01
6.64797962e-01 3.11236888e-01 1.40573061e+00 2.38983016e-02
2.36747161e-01 9.62038100e-01 -6.65870368e-01 -8.02289426e-01
-6.80145204e-01 1.33565977e-01 1.00003153e-01 -5.34671724e-01
5.43019593e-01 -6.24653339e-01 -3.79837267e-02 5.52641228e-02
-1.04336131e+00 4.52315390e-01 7.02353477e-01 5.33327401e-01
1.84161514e-01 5.18367469e-01 1.08957040e+00 -1.11113441e+00
1.28801572e+00 -6.45993292e-01 -1.89215869e-01 8.11326861e-01
-5.84233224e-01 1.34003535e-01 7.15491474e-01 -5.47898650e-01
-1.73998082e+00 -8.07575047e-01 -1.10485338e-01 9.01643485e-02
5.99007756e-02 1.27179706e+00 -5.62218167e-02 -1.25636414e-01
8.02621603e-01 6.39006615e-01 -4.19203818e-01 6.19549677e-02
9.42585111e-01 5.82781553e-01 1.37522191e-01 -1.32840943e+00
6.51882827e-01 7.58394301e-02 -4.63861912e-01 -1.14379084e+00
-1.14298713e+00 -3.72580022e-01 -3.35966706e-01 -3.09888691e-01
4.10769999e-01 -9.21923339e-01 -8.18752110e-01 4.43798125e-01
-1.23303175e+00 -6.66718066e-01 -3.80222440e-01 4.77484316e-01
-5.94394684e-01 7.15549588e-01 -8.25767219e-01 -1.10070074e+00
-3.03869605e-01 -3.61528248e-01 6.23457432e-01 2.52309650e-01
-4.67270702e-01 -1.75316143e+00 7.63066262e-02 5.58511972e-01
5.98371863e-01 -2.36833885e-01 1.42703462e+00 -7.26399541e-01
-7.03275502e-01 -8.83538369e-03 3.64008173e-02 2.17254370e-01
9.04656723e-02 -2.42315054e-01 -6.73304021e-01 2.30381072e-01
-1.52954161e-01 -7.54867792e-01 5.13223708e-01 1.56919450e-01
8.73687327e-01 -9.25365806e-01 -4.95025486e-01 -9.09580365e-02
1.28788936e+00 1.32626340e-01 6.23513937e-01 2.51779020e-01
6.80317819e-01 5.16650140e-01 4.12641853e-01 2.69949347e-01
1.09368289e+00 2.24645380e-02 -1.78904846e-01 5.81585824e-01
-2.26757377e-01 -1.11812615e+00 6.37266099e-01 1.45776761e+00
-2.08364308e-01 -5.66853344e-01 -1.08307433e+00 8.42235625e-01
-2.01706648e+00 -1.05200398e+00 -2.32718270e-02 1.57012498e+00
1.33859229e+00 -1.52701531e-02 -3.79932702e-01 -3.12239498e-01
4.85118181e-01 1.98422775e-01 -2.54469365e-01 -3.22575599e-01
-4.68639374e-01 1.34597674e-01 -7.93297514e-02 8.78874540e-01
-2.54082948e-01 1.67860174e+00 6.38922024e+00 1.30481100e+00
-5.69023609e-01 -8.93244892e-02 5.67532703e-02 4.39303458e-01
-7.87708461e-01 2.89465398e-01 -1.05773854e+00 1.93929821e-02
1.17678475e+00 -6.04368627e-01 3.93679351e-01 8.90963376e-01
3.95517722e-02 -3.82915527e-01 -6.02666080e-01 8.21000218e-01
2.51581937e-01 -1.75237215e+00 1.22965240e+00 -4.36593384e-01
9.86912847e-01 -2.93846965e-01 -3.72481763e-01 1.03218269e+00
1.05331421e+00 -1.02124918e+00 4.78481710e-01 6.95912063e-01
3.06209445e-01 -5.16051292e-01 7.07692206e-01 1.04575944e+00
-1.20253742e+00 1.64703399e-01 -5.85671127e-01 -3.36261630e-01
4.09776509e-01 4.26721156e-01 -1.35072470e+00 1.04000294e+00
3.08456421e-01 5.04150391e-01 -5.17112017e-01 5.22951663e-01
-7.70689726e-01 9.45597410e-01 -1.95774570e-01 -5.46103656e-01
3.11184436e-01 4.25996184e-02 2.91619569e-01 1.38999891e+00
2.56381959e-01 4.80582654e-01 3.73216659e-01 1.02829576e+00
-4.18230472e-03 2.38515407e-01 -8.01717341e-01 -5.17702281e-01
8.01536500e-01 9.37107146e-01 -5.65162480e-01 -8.17678988e-01
-7.12003946e-01 6.37959838e-01 6.97431505e-01 5.99763811e-01
-3.90203744e-01 -3.67300749e-01 4.32663932e-02 1.10754922e-01
-1.80477634e-01 -3.43721449e-01 -1.43056823e-04 -1.37291074e+00
-3.23202521e-01 -7.59230137e-01 6.22084141e-01 -1.23767984e+00
-1.68870306e+00 3.72084647e-01 5.33438444e-01 -5.07822573e-01
-3.43594253e-01 -6.10576451e-01 -3.88502836e-01 7.92174637e-01
-1.59354174e+00 -1.70513451e+00 6.22081198e-02 8.53618979e-01
5.62182426e-01 -1.98594466e-01 9.94004965e-01 -1.90840468e-01
-1.09324969e-01 3.17438632e-01 -4.84052390e-01 1.56401798e-01
4.29361224e-01 -1.02440226e+00 2.92374492e-01 6.40385747e-01
4.09167171e-01 1.42061937e+00 3.46392304e-01 -1.27770710e+00
-1.32197213e+00 -1.24802840e+00 1.50460517e+00 -9.56177354e-01
8.12775731e-01 -1.64003193e-01 -1.15779006e+00 1.10568964e+00
3.65454555e-01 -6.74100935e-01 9.91128564e-01 2.20087722e-01
-2.14200184e-01 3.91240835e-01 -9.59616303e-01 9.77823317e-01
1.32302320e+00 -6.40011847e-01 -1.37957811e+00 4.73456353e-01
8.79593849e-01 -2.61192381e-01 -8.72006476e-01 2.37391651e-01
-4.75155413e-02 -5.46249151e-01 8.93817902e-01 -7.49290943e-01
1.43646017e-01 -3.94332767e-01 6.87528104e-02 -1.54521596e+00
-4.57741112e-01 -7.94533610e-01 -7.19115138e-01 1.40464580e+00
5.34093201e-01 -7.51437604e-01 4.92882699e-01 6.98274374e-01
-1.50653869e-01 -6.24464512e-01 -5.16101182e-01 -1.05008149e+00
8.87043178e-02 -9.67365682e-01 4.99191523e-01 1.09454215e+00
8.32412422e-01 5.48008144e-01 -3.47256303e-01 3.24499577e-01
1.12445556e-01 5.39378300e-02 5.75695932e-01 -1.07661879e+00
-1.84547558e-01 -2.43587911e-01 9.29541141e-02 -1.21601164e+00
5.28631568e-01 -1.29015577e+00 -2.44737387e-01 -2.26808023e+00
2.81304240e-01 -3.30243200e-01 -4.40562554e-02 7.52846837e-01
-4.29650635e-01 -3.30079526e-01 -4.63496670e-02 -6.30039796e-02
-9.60411608e-01 6.35478795e-01 1.30838740e+00 6.33689165e-02
-2.38779768e-01 -2.46038333e-01 -9.93961036e-01 9.24984276e-01
7.45099604e-01 -2.01687738e-01 -1.17566168e+00 -2.84255981e-01
7.57620513e-01 8.80137086e-02 2.43786201e-01 -7.01943398e-01
5.34659564e-01 -6.08790874e-01 -1.72630383e-03 -3.89692098e-01
5.27244359e-02 -4.45575953e-01 2.06491984e-02 2.18666166e-01
-3.70467931e-01 -1.74989760e-01 5.50900042e-01 5.13655961e-01
-4.70088542e-01 -4.21986997e-01 1.23560645e-01 -7.77192473e-01
-1.29694474e+00 1.39417559e-01 -1.02696812e+00 4.86815065e-01
7.91099906e-01 -1.83801770e-01 -4.96420950e-01 -7.06370533e-01
-4.00801837e-01 4.64445502e-01 -3.94813754e-02 5.00687897e-01
9.99417603e-01 -1.27162325e+00 -4.72585052e-01 -1.88333571e-01
2.77288795e-01 9.31205973e-03 4.83020723e-01 3.58141333e-01
-3.56000751e-01 9.58452761e-01 1.86790049e-01 1.05827399e-01
-6.67426109e-01 7.36812890e-01 -3.50737087e-02 -6.11881316e-01
-3.30434173e-01 1.00356793e+00 9.77622271e-02 -8.03239644e-01
1.16552919e-01 -4.87460196e-01 -4.04211730e-01 -1.46681771e-01
4.82342422e-01 2.91905433e-01 -2.84100056e-01 -7.83040971e-02
-1.50988802e-01 2.88509816e-01 7.05607459e-02 -9.57155898e-02
1.00343490e+00 -2.59462714e-01 -3.79400998e-01 6.03949249e-01
3.17434579e-01 3.82824481e-01 -3.94824147e-01 -7.61332929e-01
2.47333616e-01 2.24210904e-03 -4.44314271e-01 -1.17256689e+00
-3.51575583e-01 5.49057126e-01 -6.63857460e-01 1.23060063e-01
7.60401666e-01 2.50160456e-01 8.05142105e-01 1.06212091e+00
1.01815820e+00 -1.06879175e+00 4.62134570e-01 1.14523280e+00
8.98162246e-01 -9.11986947e-01 -1.93837732e-01 -4.84360307e-01
-1.02458775e+00 9.88892555e-01 9.10449445e-01 3.69390577e-01
6.13586307e-01 2.28102639e-01 9.49065313e-02 -3.38435590e-01
-1.02319157e+00 -1.79160863e-01 1.64062172e-01 8.56164694e-01
4.23803240e-01 9.05609131e-02 -2.83766896e-01 1.01272595e+00
-5.56560934e-01 2.56843805e-01 3.55948001e-01 8.91372263e-01
-6.22934818e-01 -1.17362571e+00 -1.26958445e-01 4.56679761e-01
4.50204089e-02 -6.96140826e-01 -5.95397353e-01 8.53128314e-01
5.95203824e-02 1.15606010e+00 -7.11568058e-01 -2.63398379e-01
2.40492269e-01 5.63075662e-01 4.64667052e-01 -1.06506515e+00
-4.72464681e-01 -8.35657477e-01 4.76874143e-01 -1.71568185e-01
-6.12527907e-01 -9.21584517e-02 -1.41998780e+00 -4.21207666e-01
-3.34570557e-01 6.56256318e-01 -2.21302822e-01 1.09110582e+00
3.45649540e-01 1.82679981e-01 -2.80674994e-01 -4.43743020e-02
-1.07219249e-01 -1.15351307e+00 -5.79912484e-01 1.14476338e-01
-4.97507364e-01 -4.24739033e-01 1.27977535e-01 7.10589662e-02] | [10.256875038146973, 8.024232864379883] |
f0b77972-912b-452d-bc82-5697c2f06334 | aspect-extraction-with-automated-prior | null | null | https://aclanthology.org/P14-1033 | https://aclanthology.org/P14-1033.pdf | Aspect Extraction with Automated Prior Knowledge Learning | null | ['Zhiyuan Chen', 'Bing Liu', 'Arjun Mukherjee'] | 2014-06-01 | null | null | null | acl-2014-6 | ['aspect-extraction'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.269987106323242, 3.704164743423462] |
f2d5d1a4-30df-4fe5-89c9-f1ebc198f7ae | micro-expression-spotting-a-new-benchmark | 2007.12421 | null | https://arxiv.org/abs/2007.12421v2 | https://arxiv.org/pdf/2007.12421v2.pdf | Micro-expression spotting: A new benchmark | Micro-expressions (MEs) are brief and involuntary facial expressions that occur when people are trying to hide their true feelings or conceal their emotions. Based on psychology research, MEs play an important role in understanding genuine emotions, which leads to many potential applications. Therefore, ME analysis has become an attractive topic for various research areas, such as psychology, law enforcement, and psychotherapy. In the computer vision field, the study of MEs can be divided into two main tasks, spotting and recognition, which are used to identify positions of MEs in videos and determine the emotion category of the detected MEs, respectively. Recently, although much research has been done, no fully automatic system for analyzing MEs has yet been constructed on a practical level for two main reasons: most of the research on MEs only focuses on the recognition part, while abandoning the spotting task; current public datasets for ME spotting are not challenging enough to support developing a robust spotting algorithm. The contributions of this paper are threefold: (1) we introduce an extension of the SMIC-E database, namely the SMIC-E-Long database, which is a new challenging benchmark for ME spotting; (2) we suggest a new evaluation protocol that standardizes the comparison of various ME spotting techniques; (3) extensive experiments with handcrafted and deep learning-based approaches on the SMIC-E-Long database are performed for baseline evaluation. | ['Quang-Nhat Vo', 'Thuong-Khanh Tran', 'Xiaopeng Hong', 'Xiaobai Li', 'Guoying Zhao'] | 2020-07-24 | null | null | null | null | ['micro-expression-spotting'] | ['computer-vision'] | [ 2.21234411e-01 -3.16710591e-01 -3.45246106e-01 -4.76144850e-01
-4.83317047e-01 -1.19198114e-01 4.03998256e-01 -3.21909547e-01
-3.35067332e-01 3.80165368e-01 2.65683550e-02 3.27132702e-01
1.10008404e-01 -3.56563300e-01 -1.40190855e-01 -9.85242724e-01
-1.10553652e-01 -1.02652349e-01 -2.47192442e-01 -3.01162213e-01
1.04258806e-01 5.27920425e-01 -1.68408799e+00 4.12940353e-01
3.16165775e-01 1.49712145e+00 -1.89751685e-01 1.33779943e-01
2.33237967e-02 7.34322786e-01 -6.93107009e-01 -5.56661367e-01
1.58357680e-01 -7.83514857e-01 -4.57523555e-01 3.37323457e-01
1.45928323e-01 -2.86344886e-01 -2.04145104e-01 1.28081715e+00
4.80555534e-01 1.29295051e-01 3.88707727e-01 -1.54109716e+00
-1.99804738e-01 1.29095525e-01 -8.94009590e-01 2.85981297e-02
2.94875085e-01 -3.63966152e-02 8.96781802e-01 -9.78353500e-01
7.07604706e-01 1.17654967e+00 3.27816546e-01 8.76165926e-01
-7.95490503e-01 -1.04342520e+00 1.50597030e-02 4.29566056e-01
-1.49777496e+00 -6.08218908e-01 1.10050726e+00 -2.92195350e-01
5.08819342e-01 4.98899877e-01 7.52118766e-01 1.24016356e+00
2.13749390e-02 1.29504299e+00 1.31578445e+00 -2.68575668e-01
1.28258392e-01 3.88888478e-01 -1.81704052e-02 5.30423880e-01
-4.09371227e-01 -2.05511808e-01 -5.46341181e-01 2.77933683e-02
5.58279872e-01 -8.23714212e-02 -4.85277802e-01 -1.08904466e-01
-9.18436468e-01 7.12746739e-01 3.01335543e-01 7.05937088e-01
-5.12153864e-01 -2.82121569e-01 7.71002769e-01 3.72206360e-01
6.16396189e-01 1.19047418e-01 -1.47255078e-01 -4.83665586e-01
-1.00397873e+00 1.87322065e-01 4.90094453e-01 4.96683598e-01
5.62456787e-01 -5.59482761e-02 -1.36048906e-02 1.10028005e+00
-1.85397774e-01 2.61023223e-01 6.43284678e-01 -6.74833953e-01
2.72187978e-01 6.94561124e-01 2.06176877e-01 -1.83838105e+00
-4.77629036e-01 -1.37996033e-01 -1.17806709e+00 3.00111901e-02
1.59990788e-01 -3.09938431e-01 -3.56449485e-01 1.99497771e+00
2.46261865e-01 3.77001762e-01 -1.57903522e-01 1.15662622e+00
1.14370322e+00 7.46126890e-01 -1.95623845e-01 -4.06135201e-01
1.38420320e+00 -8.09293747e-01 -9.52637911e-01 -2.02594519e-01
4.59474146e-01 -6.33941829e-01 8.63042116e-01 7.47215152e-01
-8.36209893e-01 -3.70415270e-01 -8.22755933e-01 2.42675170e-01
-2.75233388e-01 6.90558672e-01 8.20855141e-01 5.32872915e-01
-8.25456142e-01 2.39370555e-01 -6.30069017e-01 -3.04082870e-01
4.15093631e-01 3.29441249e-01 -6.39843702e-01 2.67120749e-01
-1.17663765e+00 7.13941574e-01 1.71464816e-01 6.29506350e-01
-4.51237559e-01 -4.32265140e-02 -9.71907735e-01 1.20611992e-02
3.99285853e-01 4.43939120e-02 9.37605739e-01 -1.85112357e+00
-1.49561369e+00 1.44205332e+00 -3.80851299e-01 -1.06729113e-01
3.93491983e-01 -1.29968092e-01 -8.72824550e-01 1.99499100e-01
-4.73194346e-02 7.91964829e-01 1.12531495e+00 -9.92712021e-01
-4.74896759e-01 -5.26122510e-01 -1.84697047e-01 4.57919315e-02
-5.10064662e-01 5.57895005e-01 -6.40859663e-01 -7.99924493e-01
-2.61000600e-02 -1.04997039e+00 8.64266306e-02 3.17751579e-02
-5.07601142e-01 -3.97701144e-01 9.84067082e-01 -6.35812044e-01
1.41905129e+00 -2.55734897e+00 2.62214452e-01 9.97568220e-02
2.93931365e-01 5.83188057e-01 -2.07872167e-01 1.70007437e-01
-5.05001783e-01 -2.12683752e-01 3.94206978e-02 -3.91309381e-01
-6.29768670e-02 -7.35608712e-02 -4.59175736e-01 5.70182800e-01
1.55202687e-01 9.24894571e-01 -7.09212601e-01 -3.99069548e-01
2.10215390e-01 2.49485657e-01 -1.89188093e-01 3.69462430e-01
1.17858447e-01 3.48201662e-01 -3.31957072e-01 7.75089324e-01
7.80396700e-01 -8.49804580e-02 1.19912334e-01 -2.86711693e-01
1.27063572e-01 -2.13840917e-01 -1.18438399e+00 1.22470880e+00
-2.23248065e-01 7.93912351e-01 4.87062126e-01 -1.21339524e+00
9.64047968e-01 3.97617668e-01 7.13414490e-01 -5.40671051e-01
5.09005308e-01 1.27658814e-01 -2.47668982e-01 -8.78093719e-01
4.15479660e-01 -3.78579736e-01 -2.64770806e-01 3.20794851e-01
-2.05722541e-01 2.89953887e-01 7.38370046e-02 -1.30040348e-01
7.34085321e-01 -3.05129495e-02 4.10439402e-01 2.42014468e-01
6.20044053e-01 -2.96983480e-01 9.63849783e-01 5.43706492e-02
-6.68933749e-01 2.78867543e-01 5.94215214e-01 -6.11388028e-01
-3.23835224e-01 -5.50824404e-01 -1.25202507e-01 9.67816114e-01
4.25684303e-01 -4.50051785e-01 -8.60460520e-01 -7.14697540e-01
-3.53573054e-01 3.88870299e-01 -8.49960625e-01 -1.23031586e-01
-2.35175774e-01 -9.29299653e-01 5.45352876e-01 3.48409802e-01
8.07789564e-01 -1.51015329e+00 -7.97173023e-01 2.34413520e-02
-5.85640848e-01 -1.23437905e+00 -3.51145416e-01 -2.73257762e-01
-5.35023212e-01 -1.11429274e+00 -8.12014580e-01 -6.60504580e-01
5.76444805e-01 4.46267664e-01 9.58846390e-01 1.46661133e-01
-3.41337472e-01 2.55936295e-01 -3.63832951e-01 -4.61154819e-01
-8.99465829e-02 -3.13235581e-01 7.81877413e-02 6.18639290e-01
9.74803269e-01 -4.24576730e-01 -5.61982810e-01 6.42155826e-01
-9.51388240e-01 3.96941900e-02 6.81977808e-01 8.21937323e-01
4.06327844e-01 2.99599245e-02 5.88914633e-01 -6.60776854e-01
6.27649665e-01 -4.21803445e-01 -2.87313342e-01 1.68292373e-01
1.98030211e-02 -4.76680547e-01 3.46350610e-01 -4.80574936e-01
-1.08197057e+00 1.85866505e-02 -4.03492779e-01 -5.77385485e-01
-3.99385840e-01 4.30335164e-01 -2.83785075e-01 -1.59324229e-01
2.74702489e-01 4.21840221e-01 5.74355982e-02 -2.44080916e-01
5.02574781e-04 1.03709686e+00 5.30774653e-01 -4.10617918e-01
3.02882701e-01 5.14252245e-01 -1.17482632e-01 -1.14861131e+00
-9.29191351e-01 -5.63585520e-01 -4.03365105e-01 -3.26526105e-01
8.71534765e-01 -7.80412138e-01 -7.78961360e-01 7.60696828e-01
-1.06344008e+00 3.81712727e-02 3.28797013e-01 2.70510733e-01
-5.25331438e-01 5.09926856e-01 -7.47992635e-01 -8.54587018e-01
-4.22826886e-01 -1.26995409e+00 1.19329882e+00 3.03728372e-01
-4.75787014e-01 -6.60400510e-01 2.63044909e-02 4.79726315e-01
1.85975447e-01 4.12121832e-01 6.08921170e-01 -3.23616982e-01
-1.50813803e-01 -3.49852353e-01 -2.59714663e-01 5.02984405e-01
2.81986356e-01 5.38820364e-02 -1.11506987e+00 -1.57800987e-01
1.61453500e-01 -7.56322086e-01 7.43472755e-01 1.64599583e-01
1.42021155e+00 -1.16468310e-01 -2.82247454e-01 6.38242245e-01
9.03338194e-01 3.30052793e-01 9.17466462e-01 2.80691177e-01
2.84897745e-01 8.40057909e-01 9.04207468e-01 6.47046745e-01
7.95699880e-02 9.94809806e-01 5.61980963e-01 -2.50603378e-01
3.90452117e-01 1.63852721e-01 5.61148584e-01 6.28413439e-01
-1.58863232e-01 -4.63372655e-02 -3.41584206e-01 4.60730612e-01
-1.91238451e+00 -1.27649498e+00 2.54135936e-01 1.92829466e+00
7.55615592e-01 -2.23191321e-01 1.98805913e-01 1.85275301e-01
8.32463086e-01 4.60863858e-01 -5.04879534e-01 -4.47611690e-01
-9.48818177e-02 1.07228570e-01 -3.46750110e-01 -1.49283051e-01
-1.45442581e+00 9.15712774e-01 5.26267481e+00 1.03788841e+00
-1.79534364e+00 -1.69494927e-01 9.74840581e-01 -8.25120360e-02
3.84272188e-01 -5.68705201e-01 -7.16020465e-01 5.23964047e-01
4.02725160e-01 1.32209986e-01 2.96883553e-01 1.05430746e+00
4.08317268e-01 -3.06844078e-02 -1.03714097e+00 1.83193290e+00
2.78209001e-01 -9.47257698e-01 -2.79348910e-01 -1.81580260e-01
4.61264729e-01 -4.23652858e-01 1.68631375e-01 3.33592772e-01
-5.39643288e-01 -1.10336745e+00 4.47127134e-01 2.54705012e-01
7.82307088e-01 -9.06748652e-01 8.86817515e-01 3.70472193e-01
-1.02023613e+00 6.84902966e-02 -3.19276482e-01 -1.18065387e-01
2.87881102e-02 6.18506014e-01 -4.55153733e-01 3.83896291e-01
7.67713845e-01 9.60443079e-01 -2.78690398e-01 7.53879607e-01
-3.66797268e-01 5.54686964e-01 -1.79494038e-01 -1.15682229e-01
2.85474539e-01 -3.44298840e-01 5.71453631e-01 1.25840425e+00
2.73303241e-01 1.10329919e-01 -6.18351772e-02 8.70926857e-01
-1.79676235e-01 4.00473416e-01 -4.00129169e-01 -2.12015048e-01
7.29582980e-02 1.76477802e+00 -6.83114529e-01 -7.80280530e-02
-3.33831042e-01 1.28973472e+00 1.74942702e-01 1.87552050e-01
-8.54859173e-01 -4.53337252e-01 9.31694865e-01 -1.73891872e-01
2.81428188e-01 1.23991836e-02 1.97670475e-01 -1.53950071e+00
9.46110338e-02 -1.20043421e+00 4.28094774e-01 -7.29625225e-01
-1.16409862e+00 7.22271025e-01 -1.57086059e-01 -1.28855956e+00
-4.52138066e-01 -6.56378686e-01 -7.97587276e-01 5.03244162e-01
-1.21603394e+00 -9.07673120e-01 -5.95995188e-01 6.15520775e-01
4.83560711e-01 -2.67767459e-02 9.77040589e-01 4.58309293e-01
-9.91724253e-01 8.03303719e-01 -1.89405337e-01 4.60944504e-01
8.72762263e-01 -6.75974786e-01 -2.34901696e-01 6.74579859e-01
3.14532518e-01 5.77001870e-01 6.28084004e-01 -2.60728002e-01
-1.27193987e+00 -8.36805046e-01 7.47989833e-01 -1.87809959e-01
5.02864957e-01 -4.47599411e-01 -7.98614621e-01 6.26468658e-01
5.90438992e-02 1.28949746e-01 7.12933898e-01 -8.15432072e-02
-1.80921052e-02 -1.94529086e-01 -1.15600407e+00 6.87574804e-01
7.39074588e-01 -3.20299119e-01 -3.67812216e-01 1.77029110e-02
-1.41270876e-01 -3.26089859e-01 -6.60024941e-01 3.63967210e-01
6.65042639e-01 -1.32781470e+00 6.78680122e-01 -3.95511925e-01
5.39231539e-01 -8.97108694e-04 7.23217055e-02 -1.23384225e+00
6.59194514e-02 -5.97908556e-01 1.32173346e-02 1.37374270e+00
-1.62308738e-01 -4.59746689e-01 9.63204622e-01 5.07953882e-01
3.32181185e-01 -1.12906396e+00 -9.32325363e-01 -4.75747943e-01
-5.07363021e-01 -5.64198017e-01 4.93534446e-01 1.16283274e+00
9.13979858e-02 3.03526938e-01 -8.36647570e-01 -2.82373846e-01
2.57943511e-01 5.79184532e-01 1.01573098e+00 -9.11632955e-01
-2.02334806e-01 -7.04564631e-01 -7.18999326e-01 -1.19086468e+00
4.07890171e-01 -5.15462756e-01 2.87060700e-02 -8.49509954e-01
3.55126113e-01 -1.95030794e-01 -3.81192863e-01 6.07573271e-01
-1.51069090e-01 4.04403090e-01 1.42905429e-01 2.00039729e-01
-6.62793338e-01 9.07653093e-01 1.08932745e+00 -7.67275020e-02
2.08527409e-02 1.22213513e-01 -6.85508966e-01 1.05144310e+00
5.52442133e-01 -1.98379576e-01 -2.85452038e-01 1.15408339e-01
2.68094063e-01 1.72755301e-01 3.34499955e-01 -6.68582797e-01
-1.37186930e-01 -3.33953649e-01 2.85069823e-01 -5.95118582e-01
5.58298647e-01 -7.92605698e-01 -1.06718421e-01 1.53078176e-02
-7.49722198e-02 -7.39701465e-02 2.22581655e-01 2.67624885e-01
-7.05426514e-01 -1.85004771e-01 9.42602038e-01 -1.08470410e-01
-1.28723133e+00 3.89751792e-01 -3.40972364e-01 -1.60605542e-03
1.32578361e+00 -8.35877657e-02 8.56296495e-02 -8.66340518e-01
-6.26097977e-01 2.85046734e-02 2.66745806e-01 5.99255443e-01
7.66514659e-01 -1.39517164e+00 -4.80815351e-01 1.85667321e-01
1.70467451e-01 -3.32876921e-01 3.20437074e-01 1.02545285e+00
-1.66953400e-01 3.09859097e-01 -3.85790884e-01 -5.13594866e-01
-1.68967462e+00 5.15129745e-01 3.40877950e-01 -8.68084952e-02
-5.02768219e-01 9.49017763e-01 5.38157105e-01 -5.94834834e-02
3.42027098e-01 2.06312016e-02 -4.71818298e-01 2.30165645e-01
9.90186989e-01 2.50246853e-01 -1.25125954e-02 -1.11458838e+00
-5.38922429e-01 4.81471747e-01 1.18398100e-01 2.82603264e-01
1.45707703e+00 -1.67761803e-01 -4.71855283e-01 3.34179610e-01
1.24321234e+00 -3.64396051e-02 -7.69623339e-01 -1.31867439e-01
-3.58090848e-02 -6.16757870e-01 1.12635665e-01 -5.60608327e-01
-1.39247036e+00 1.17176735e+00 5.71400642e-01 -4.29824926e-02
1.55963922e+00 -1.12140208e-01 1.10629368e+00 3.26189041e-01
5.05693853e-01 -1.11195624e+00 2.94271111e-01 1.44020453e-01
9.94339526e-01 -1.32403994e+00 -3.14849913e-01 -3.74232501e-01
-8.53971720e-01 1.17631459e+00 6.83772922e-01 1.98576555e-01
4.36155498e-01 1.31021708e-01 2.68518686e-01 -3.90126377e-01
-4.91202682e-01 7.09912181e-02 4.09676433e-01 1.80631727e-01
4.19716954e-01 -7.63968378e-02 -3.51832271e-01 1.11043024e+00
8.28972235e-02 3.26935649e-01 2.22576410e-01 6.60834432e-01
-3.53175312e-01 -9.72978354e-01 -3.52954596e-01 3.30116749e-01
-7.70203829e-01 2.35602409e-01 -6.14604473e-01 6.73660159e-01
2.69605905e-01 8.37475359e-01 -1.16085805e-01 -6.43298209e-01
8.70150104e-02 -4.23169788e-03 3.99900585e-01 -3.95561904e-01
-3.19115460e-01 -1.35876136e-02 1.35368288e-01 -8.06047559e-01
-7.40965545e-01 -5.65780997e-01 -1.02725673e+00 -2.77046621e-01
-2.40025863e-01 1.72334015e-01 4.32243705e-01 1.01573968e+00
3.57917994e-01 -2.96112988e-03 8.00141931e-01 -1.04314888e+00
-4.29925323e-01 -7.34834433e-01 -8.42480540e-01 8.90169799e-01
5.09613752e-02 -1.01350129e+00 -2.86662817e-01 -2.08958685e-01] | [13.603320121765137, 1.8062536716461182] |
893720dd-5d0a-47ce-899e-4c66c3d6480f | safe-deep-rl-for-intraoperative-planning-of | 2305.05354 | null | https://arxiv.org/abs/2305.05354v2 | https://arxiv.org/pdf/2305.05354v2.pdf | Safe Deep RL for Intraoperative Planning of Pedicle Screw Placement | Spinal fusion surgery requires highly accurate implantation of pedicle screw implants, which must be conducted in critical proximity to vital structures with a limited view of anatomy. Robotic surgery systems have been proposed to improve placement accuracy, however, state-of-the-art systems suffer from the limitations of open-loop approaches, as they follow traditional concepts of preoperative planning and intraoperative registration, without real-time recalculation of the surgical plan. In this paper, we propose an intraoperative planning approach for robotic spine surgery that leverages real-time observation for drill path planning based on Safe Deep Reinforcement Learning (DRL). The main contributions of our method are (1) the capability to guarantee safe actions by introducing an uncertainty-aware distance-based safety filter; and (2) the ability to compensate for incomplete intraoperative anatomical information, by encoding a-priori knowledge about anatomical structures with a network pre-trained on high-fidelity anatomical models. Planning quality was assessed by quantitative comparison with the gold standard (GS) drill planning. In experiments with 5 models derived from real magnetic resonance imaging (MRI) data, our approach was capable of achieving 90% bone penetration with respect to the GS while satisfying safety requirements, even under observation and motion uncertainty. To the best of our knowledge, our approach is the first safe DRL approach focusing on orthopedic surgeries. | ['Philipp Fuernstahl', 'Andreas Krause', 'Benjamin F. Grewe', 'Mazda Farshad', 'Yarden As', 'Fabio Carrillo', 'Hooman Esfandiari', 'Yunke Ao'] | 2023-05-09 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [-8.77450109e-02 7.66875148e-01 -3.18274319e-01 5.66032454e-02
-8.35570097e-01 -4.27683860e-01 2.60310173e-01 3.10572416e-01
-5.37019908e-01 7.83303916e-01 3.05978566e-01 -6.69693351e-01
-5.32019079e-01 -6.05009735e-01 -1.03821576e+00 -3.18781883e-01
-3.82180154e-01 9.28687096e-01 3.09604526e-01 -3.36551040e-01
3.12869042e-01 6.34752691e-01 -1.04308057e+00 -1.71190932e-01
9.91348326e-01 1.09286022e+00 2.70796567e-01 2.76693523e-01
4.85150844e-01 7.98126101e-01 4.64718938e-02 6.19265102e-02
4.24943060e-01 -4.31150794e-02 -9.22328293e-01 -3.58900785e-01
-1.10212257e-02 -5.26471198e-01 -3.37910384e-01 8.21711302e-01
5.13794065e-01 2.77409889e-02 3.86703849e-01 -5.83905876e-01
-2.27705315e-01 6.32240295e-01 -1.62095800e-01 -3.54623161e-02
2.32720286e-01 3.77620786e-01 4.02976334e-01 -6.69824958e-01
8.30370963e-01 4.12937522e-01 8.41767073e-01 6.77007318e-01
-9.08409238e-01 -4.80692267e-01 -1.92963913e-01 -8.31866190e-02
-1.05371916e+00 -1.41571224e-01 4.19239640e-01 -6.85926437e-01
1.04491210e+00 3.47925685e-02 1.24409735e+00 8.21654081e-01
1.39354527e+00 2.70699233e-01 1.13157213e+00 -4.60515529e-01
5.35878599e-01 -4.35364902e-01 -4.72383231e-01 8.22638094e-01
2.32450947e-01 9.82558787e-01 -1.43990651e-01 1.45463675e-01
1.14827490e+00 -8.08352325e-03 -5.70352793e-01 -1.03000748e+00
-1.25844920e+00 6.88740969e-01 1.09782016e+00 3.13376784e-01
-6.72956467e-01 2.96386361e-01 3.37764829e-01 -1.62406951e-01
-2.43796706e-01 7.94061780e-01 -2.61516631e-01 -2.33367860e-01
-8.47600937e-01 -1.99718364e-02 8.21893156e-01 8.11543465e-01
1.45860510e-02 -2.40656674e-01 3.94531935e-02 2.97816455e-01
4.13736880e-01 2.53987849e-01 7.47855186e-01 -1.00570726e+00
1.98877677e-01 2.54310250e-01 1.64971232e-01 -5.88829279e-01
-9.72985983e-01 -6.20764196e-01 -4.77754831e-01 4.55215454e-01
2.03490883e-01 -4.83035892e-02 -1.19557929e+00 1.39976978e+00
1.18437730e-01 6.19662553e-02 -7.50604123e-02 9.23584640e-01
3.92794192e-01 -2.60089021e-02 -1.79560900e-01 -4.24352437e-01
1.20532703e+00 -8.79210174e-01 -5.17515004e-01 -5.39515674e-01
6.23477340e-01 -4.70782965e-01 8.39394987e-01 7.37586677e-01
-1.18555963e+00 -1.18373841e-01 -1.13035107e+00 1.62864476e-01
1.44295260e-01 6.96865022e-02 6.70742214e-01 5.80327690e-01
-1.17096817e+00 6.55376554e-01 -1.55279493e+00 -2.48195142e-01
4.30111617e-01 7.59949446e-01 -5.94222426e-01 1.28485814e-01
-1.09185493e+00 1.61058939e+00 3.53961915e-01 1.43159747e-01
-1.14512873e+00 -7.63498962e-01 -9.02355552e-01 -1.45750567e-01
6.77430212e-01 -8.43494117e-01 1.63463140e+00 -4.20920223e-01
-2.07006693e+00 5.67379177e-01 5.19892156e-01 -7.88957715e-01
5.60866594e-01 -4.94706452e-01 -2.57923678e-02 5.14507890e-01
-5.21981940e-02 5.45355737e-01 3.21208984e-01 -1.19967842e+00
-4.48737666e-02 -4.87973213e-01 1.36920750e-01 2.50829339e-01
2.09770858e-01 -6.27258360e-01 -1.86183199e-01 -4.93186921e-01
4.02267247e-01 -1.32915556e+00 -9.89063501e-01 3.86882544e-01
-6.86153993e-02 5.72758079e-01 1.13555372e-01 -7.07645595e-01
1.06126714e+00 -1.72125089e+00 2.29606569e-01 3.64150882e-01
4.80358414e-02 9.08273682e-02 4.13949460e-01 2.51395911e-01
4.09242511e-02 -2.97292084e-01 -2.19280198e-01 8.37677792e-02
-5.67545831e-01 4.91770416e-01 -1.49963230e-01 6.95622683e-01
-3.08851331e-01 9.65333641e-01 -1.08969796e+00 -5.00354171e-01
5.67531168e-01 2.15880990e-01 -9.70034480e-01 4.84836623e-02
-1.73426792e-01 9.51929033e-01 -3.26809824e-01 5.20574629e-01
2.27832288e-01 -3.48531343e-02 3.04885924e-01 -2.82963604e-01
-8.50837901e-02 3.37991834e-01 -7.04230487e-01 2.44160652e+00
-9.65704858e-01 -2.96154767e-01 3.20661575e-01 -6.43970907e-01
7.09278405e-01 4.13713545e-01 9.76691306e-01 -8.21044505e-01
6.11024618e-01 8.46510768e-01 2.65735835e-01 -2.23563001e-01
1.55195013e-01 -5.71668148e-01 -3.12552154e-02 1.22075602e-01
3.02054286e-01 -8.94438803e-01 -5.49049675e-01 1.14568640e-02
1.12686467e+00 1.65464595e-01 5.02095520e-01 -6.30671501e-01
2.22897053e-01 1.68794960e-01 3.03320974e-01 6.22441530e-01
-1.29755318e-01 6.17479324e-01 2.07680047e-01 -4.69517678e-01
-6.32923305e-01 -1.46460521e+00 -2.39172712e-01 -8.37921947e-02
4.45183247e-01 -2.19255984e-02 -7.04352617e-01 -5.32469273e-01
-5.49694113e-02 8.55360627e-01 -7.53744721e-01 -5.40491760e-01
-6.75443053e-01 -5.29095717e-02 2.67129868e-01 6.63026035e-01
-2.70614624e-01 -8.73242617e-01 -1.31083739e+00 5.98511100e-01
1.13642715e-01 -9.18487966e-01 -1.50046036e-01 7.30633020e-01
-1.36602485e+00 -1.26261210e+00 -5.39863288e-01 -4.60546076e-01
7.10557342e-01 -2.05109820e-01 8.22670281e-01 -8.34765881e-02
-3.90711963e-01 5.18002033e-01 -1.87466532e-01 -2.56955326e-01
-5.21144509e-01 2.40155309e-02 1.09215081e-01 -7.71402657e-01
-5.13731539e-01 -4.17291313e-01 -9.56022859e-01 3.44928414e-01
-6.72672093e-01 9.96669382e-02 9.28143322e-01 9.85969782e-01
8.78616929e-01 -2.80265301e-01 2.54979610e-01 -6.62099779e-01
4.74176824e-01 -3.36674839e-01 -7.02527106e-01 -5.50729595e-02
-9.36777353e-01 1.21851631e-01 5.46182334e-01 -6.05620183e-02
-1.03263676e+00 2.33431816e-01 -3.99185985e-01 -4.47658688e-01
1.73975199e-01 7.31271565e-01 2.88987875e-01 -4.46822912e-01
8.53481114e-01 -3.17087799e-01 5.94128430e-01 -5.00648692e-02
3.98431242e-01 1.96423650e-01 9.00367975e-01 -8.33325565e-01
1.19312830e-01 6.12582982e-01 3.49803120e-01 1.96909276e-03
-7.08889604e-01 -2.65569627e-01 -7.18387842e-01 -4.53808308e-01
7.19308734e-01 -6.47590995e-01 -8.07326674e-01 1.28667466e-02
-6.67490542e-01 -5.44511497e-01 -6.03256941e-01 1.03118491e+00
-1.21820080e+00 3.20801228e-01 -7.95437336e-01 -3.55566233e-01
-5.19082725e-01 -1.74158788e+00 1.06748366e+00 4.20583151e-02
-3.94961983e-01 -9.29000556e-01 1.64295927e-01 4.60471250e-02
5.59529126e-01 5.52220285e-01 6.47558987e-01 -3.23873967e-01
-6.16950512e-01 -3.99878114e-01 4.37606454e-01 1.21290073e-01
4.96386886e-02 -3.86998951e-01 -5.02191186e-01 -3.30351412e-01
3.64610821e-01 -2.95851380e-01 2.70279646e-01 8.89659166e-01
1.19773328e+00 -1.25548571e-01 -5.36962688e-01 6.62607849e-01
1.60864925e+00 2.30150312e-01 7.22258270e-01 5.62176168e-01
4.10273075e-01 3.14328820e-01 8.73110235e-01 2.91422367e-01
2.72748291e-01 7.84590304e-01 9.56393063e-01 3.11216682e-01
-2.86221728e-02 -2.65936822e-01 1.64367519e-02 7.46026456e-01
-8.71883929e-02 1.67108342e-01 -1.30078602e+00 4.29471582e-01
-1.63117492e+00 -4.47881073e-01 1.02981023e-01 2.54311180e+00
8.50852013e-01 4.10025150e-01 -5.56941569e-01 -2.10490733e-01
-3.89515162e-02 -2.69838661e-01 -5.13335407e-01 -3.38332951e-01
6.97517991e-01 6.10458314e-01 1.00787044e+00 7.54848599e-01
-7.34666407e-01 6.56548202e-01 6.10299063e+00 4.02468234e-01
-1.48971665e+00 2.54581988e-01 3.94566683e-03 -3.14092517e-01
-2.22750142e-01 1.02618054e-01 -2.34409317e-01 2.91595101e-01
8.40250969e-01 1.11846521e-01 1.87385723e-01 8.17305863e-01
1.33784205e-01 -6.28766358e-01 -1.15054262e+00 4.52279419e-01
-1.61893144e-02 -1.64278710e+00 -3.80740672e-01 -5.23622334e-02
5.56120694e-01 2.58512229e-01 -1.17518097e-01 1.90757155e-01
2.21788138e-01 -1.00749290e+00 9.03355479e-01 8.58280599e-01
8.81440699e-01 -6.07069612e-01 7.57388473e-01 4.09104586e-01
-7.73614883e-01 -2.58333951e-01 5.84139340e-02 1.02292277e-01
5.58774352e-01 4.07720178e-01 -1.01035857e+00 9.56240177e-01
5.52124798e-01 5.35355926e-01 -1.50261922e-02 1.13806450e+00
-4.15497124e-01 1.79393992e-01 -5.12841940e-01 3.00727993e-01
2.21777037e-01 7.76139423e-02 3.93607736e-01 3.63242269e-01
5.36376834e-01 3.46451879e-01 1.37417197e-01 4.32270974e-01
2.92862266e-01 -4.30628508e-02 -5.51406026e-01 4.76249427e-01
2.33817086e-01 7.42579877e-01 -8.18201125e-01 2.79608965e-01
-1.19772283e-02 5.38877785e-01 2.66401827e-01 -2.96433479e-01
-7.11678803e-01 1.24748945e-01 1.15496285e-01 3.47768277e-01
-8.51413310e-02 -2.97883600e-01 -8.92136574e-01 -7.19908893e-01
3.86603773e-02 -5.90076923e-01 2.35086873e-01 -9.95445669e-01
-5.98062336e-01 6.86092675e-01 -5.83659708e-02 -1.36130202e+00
-4.52562332e-01 -7.67657638e-01 -3.68387312e-01 6.56601250e-01
-1.22527635e+00 -9.38549161e-01 -2.00644150e-01 2.99887955e-01
8.57574791e-02 2.18821689e-01 8.88052583e-01 -1.21067435e-01
1.69128746e-01 9.62974355e-02 -1.78658172e-01 -4.17010099e-01
6.99890852e-01 -1.06624007e+00 -1.74530149e-01 5.93798041e-01
-4.38076854e-01 8.02733362e-01 7.12529063e-01 -8.64839792e-01
-1.42641127e+00 -7.49050081e-01 2.97186106e-01 -4.92918789e-01
6.55212879e-01 1.78192884e-01 -6.20224655e-01 9.58182752e-01
-6.95720837e-02 2.30897173e-01 5.10220170e-01 -4.88610789e-02
2.78862745e-01 -3.63923758e-02 -1.22886002e+00 5.72872043e-01
1.03362429e+00 -1.88908815e-01 -8.95323455e-01 2.41481140e-01
3.83024633e-01 -1.34639072e+00 -1.27177262e+00 1.12041938e+00
7.53831387e-01 -8.29649627e-01 1.03248799e+00 -9.18854922e-02
6.52951479e-01 -2.14646488e-01 1.87343702e-01 -1.38790751e+00
7.33882710e-02 -4.03125703e-01 -1.30362719e-01 1.38379782e-01
1.20092355e-01 -5.99514365e-01 7.50142276e-01 5.32535076e-01
-7.83014417e-01 -1.45202315e+00 -1.48457229e+00 -7.11016357e-01
2.68905252e-01 -5.04132628e-01 2.05528513e-01 4.67981398e-01
5.28689444e-01 -3.32421869e-01 -4.06611636e-02 5.20618260e-01
1.78117335e-01 -7.45805129e-02 3.41240585e-01 -1.13453650e+00
-4.55872238e-01 -5.25041580e-01 -3.75145108e-01 -6.63287163e-01
-5.15126847e-02 -9.52344656e-01 5.88920116e-01 -1.97508717e+00
-2.28981614e-01 -8.47283423e-01 -1.80396765e-01 3.61383915e-01
2.96751797e-01 -1.27131939e-01 -3.35618109e-02 3.60060394e-01
-6.03159517e-02 7.14057505e-01 1.59872854e+00 3.66235137e-01
-2.19394341e-01 6.20721690e-02 -3.15160483e-01 8.23752582e-01
7.40405798e-01 -3.83436799e-01 -6.07614100e-01 -3.91071588e-01
1.64420620e-01 6.54920995e-01 3.44731539e-01 -1.25656343e+00
5.18153787e-01 6.25309423e-02 1.39582619e-01 -4.14412886e-01
2.12263986e-01 -1.38766122e+00 3.80362898e-01 1.35662544e+00
-4.97976877e-02 -2.44139303e-02 4.68459010e-01 5.85642517e-01
-1.88278154e-01 -1.45763636e-01 8.75725627e-01 -1.94086090e-01
-4.64575768e-01 2.19561890e-01 -4.10121560e-01 -1.32183403e-01
1.34193099e+00 -1.85183004e-01 1.27135683e-02 5.13558909e-02
-9.96242464e-01 1.86365552e-03 6.74554467e-01 3.66108835e-01
6.93348765e-01 -9.13885593e-01 -7.53171369e-02 2.34067604e-01
1.26554504e-01 4.57321972e-01 4.11016256e-01 1.36483204e+00
-1.24672115e+00 6.56158209e-01 -6.43507719e-01 -7.94222713e-01
-3.90633941e-01 8.01885068e-01 1.05633128e+00 -5.02917647e-01
-7.87031889e-01 7.88690090e-01 1.99750617e-01 -7.08838046e-01
2.27821693e-01 -7.76433229e-01 9.81016010e-02 -6.12710297e-01
5.10610752e-02 -8.33257437e-02 5.54873824e-01 -1.89168662e-01
-5.65021515e-01 5.21324396e-01 -5.34776337e-02 -1.90420151e-01
1.13427567e+00 4.68485057e-02 1.11931160e-01 2.02203467e-01
4.43473727e-01 9.20332745e-02 -1.28384030e+00 1.41146317e-01
-1.23219065e-01 -4.00725693e-01 3.10295194e-01 -1.28304887e+00
-1.13763809e+00 6.53144240e-01 6.21674538e-01 -8.21242630e-01
1.05072176e+00 -4.80530038e-02 7.14169025e-01 -1.40267033e-02
1.30704141e+00 -9.69014943e-01 1.44128539e-02 3.34416330e-01
1.12189579e+00 -8.93457174e-01 -4.11207601e-02 -6.25141144e-01
-5.91360390e-01 1.17337525e+00 7.01890469e-01 -2.51228541e-01
1.09009159e+00 6.21888220e-01 1.81457460e-01 -4.32084858e-01
-3.25570107e-01 2.66254097e-01 2.09476367e-01 3.67904097e-01
2.72366583e-01 2.45446742e-01 -6.43447042e-01 5.25151730e-01
-3.17962021e-01 6.34899974e-01 5.19979239e-01 1.35791683e+00
-3.50708276e-01 -7.88973570e-01 -5.92168681e-02 4.54808950e-01
-4.94179547e-01 2.58138310e-02 4.34644222e-01 9.03187394e-01
-1.78542256e-01 5.15370905e-01 -1.12820126e-01 -1.79434605e-02
5.80832362e-01 -3.91377717e-01 9.30500746e-01 -7.90619135e-01
-7.66871095e-01 -6.08470216e-02 -1.22010380e-01 -1.00307810e+00
8.46020132e-02 -5.64222455e-01 -1.71696353e+00 4.18171227e-01
-3.64652872e-01 1.25228986e-01 7.24577725e-01 9.37556326e-01
2.98857599e-01 1.00298059e+00 9.42864493e-02 -1.02889597e+00
-8.09246600e-01 -5.94267488e-01 -4.53980386e-01 -1.56003550e-01
3.02334487e-01 -1.15668702e+00 5.72872534e-03 -4.19182152e-01] | [13.824329376220703, -3.0089073181152344] |
30699e4f-099d-48fb-bf9e-b875342780b2 | data-efficient-paraphrase-generation-to | null | null | https://aclanthology.org/2020.coling-industry.2 | https://aclanthology.org/2020.coling-industry.2.pdf | Data-Efficient Paraphrase Generation to Bootstrap Intent Classification and Slot Labeling for New Features in Task-Oriented Dialog Systems | Recent progress through advanced neural models pushed the performance of task-oriented dialog systems to almost perfect accuracy on existing benchmark datasets for intent classification and slot labeling. However, in evolving real-world dialog systems, where new functionality is regularly added, a major additional challenge is the lack of annotated training data for such new functionality, as the necessary data collection efforts are laborious and time-consuming. A potential solution to reduce the effort is to augment initial seed data by paraphrasing existing utterances automatically. In this paper, we propose a new, data-efficient approach following this idea. Using an interpretation-to-text model for paraphrase generation, we are able to rely on existing dialog system training data, and, in combination with shuffling-based sampling techniques, we can obtain diverse and novel paraphrases from small amounts of seed data. In experiments on a public dataset and with a real-world dialog system, we observe improvements for both intent classification and slot labeling, demonstrating the usefulness of our approach. | ['Daniil Sorokin', 'Caglar Tirkaz', 'Tobias Falke', 'Shailza Jolly'] | 2020-12-01 | null | null | null | coling-2020-8 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 2.72411942e-01 3.47348779e-01 -1.88818462e-02 -6.83166146e-01
-8.36296916e-01 -8.21276963e-01 7.26768672e-01 3.64259899e-01
-5.88369191e-01 1.01238143e+00 5.98252296e-01 -3.26720893e-01
3.48023891e-01 -5.55180848e-01 -2.23995924e-01 -7.40606114e-02
3.85197431e-01 1.21140385e+00 3.28633815e-01 -7.44208336e-01
3.41392338e-01 8.61269236e-02 -1.32847202e+00 4.24144417e-01
8.93809736e-01 6.68856084e-01 2.49129131e-01 6.20714545e-01
-5.79518795e-01 7.54344702e-01 -8.99860919e-01 -3.96968931e-01
1.42287999e-01 -6.80545270e-01 -1.34385860e+00 3.01681697e-01
1.17796488e-01 -4.91007715e-01 3.10062878e-02 6.66448712e-01
4.51714456e-01 3.57330650e-01 4.71861511e-01 -1.05091441e+00
-2.41645679e-01 9.16980565e-01 -2.26562489e-02 -2.44529158e-01
6.72934294e-01 1.50781050e-01 1.17624450e+00 -7.55888343e-01
7.41703570e-01 1.15551019e+00 5.49000919e-01 8.42388391e-01
-1.61047935e+00 -2.90471196e-01 -3.55634652e-02 -4.29401807e-02
-8.28570545e-01 -6.66031659e-01 6.48269415e-01 -8.46491158e-02
1.12362099e+00 3.50345880e-01 3.35830837e-01 1.23151243e+00
-2.90865093e-01 8.79589379e-01 9.00229931e-01 -7.28440225e-01
5.05940378e-01 6.57931149e-01 4.22605336e-01 5.96201837e-01
-2.51732707e-01 -3.45211864e-01 -4.26904261e-01 -4.78373498e-01
2.37510845e-01 -6.22180961e-02 -1.82058558e-01 -4.41820294e-01
-9.10853207e-01 1.11064136e+00 2.57911950e-01 3.83334190e-01
-7.75878429e-02 -4.32796359e-01 6.61853552e-01 5.55936456e-01
5.18015742e-01 1.02741146e+00 -5.41414499e-01 -4.22519922e-01
-9.20473278e-01 4.12658244e-01 1.50498962e+00 9.25699294e-01
8.22462797e-01 -3.38693738e-01 -2.53477216e-01 1.34770203e+00
-9.22900811e-03 8.71419609e-02 9.67159629e-01 -1.00419557e+00
5.15374362e-01 8.94701660e-01 2.71015197e-01 -3.51562500e-01
-4.45533991e-01 9.37954187e-02 -4.15522039e-01 -2.82383170e-02
7.74693489e-01 -3.42543215e-01 -8.52692604e-01 1.74990547e+00
1.39043689e-01 -4.11865532e-01 3.24272454e-01 5.83076119e-01
6.19081736e-01 4.98401642e-01 1.23199150e-02 -1.44398108e-01
1.40044749e+00 -9.73427653e-01 -4.16415125e-01 -3.99374187e-01
7.77501643e-01 -7.67379522e-01 1.49062693e+00 1.14803746e-01
-1.02262664e+00 -4.27252084e-01 -8.92438412e-01 -1.27548277e-01
-2.98102468e-01 2.75830664e-02 7.08000183e-01 6.78304315e-01
-1.00364792e+00 4.99201894e-01 -5.39223254e-01 -7.76156902e-01
-3.46510410e-02 4.84922647e-01 -2.97834069e-01 4.55868542e-02
-1.11362684e+00 8.07286084e-01 3.70834261e-01 -3.62243503e-01
-4.61284876e-01 -5.15998065e-01 -9.04003501e-01 3.08310419e-01
5.86062849e-01 -5.45411527e-01 1.90710413e+00 -7.25464940e-01
-1.89382768e+00 6.32438123e-01 -2.64092267e-01 -7.33257055e-01
3.97483557e-01 4.04825620e-02 1.84509397e-01 -1.15343265e-01
5.29070571e-03 9.64590013e-01 4.81764078e-01 -1.21772432e+00
-6.50822222e-01 -1.09833673e-01 3.97020936e-01 3.28463316e-01
-6.43289983e-01 -1.50714263e-01 -1.89843997e-01 -1.43603384e-01
-6.17661364e-02 -1.21737778e+00 -4.28208590e-01 -3.89018238e-01
-3.01871836e-01 -4.52509046e-01 6.46768332e-01 -5.86845994e-01
9.41781282e-01 -2.01865196e+00 1.53661072e-01 -1.13854297e-01
1.45123884e-01 3.80802900e-01 -8.66483431e-03 7.32773423e-01
2.98993200e-01 -5.87026663e-02 -3.02177846e-01 -8.28693449e-01
1.08831838e-01 3.17176521e-01 -5.11477232e-01 -3.42427313e-01
1.21363342e-01 8.07535350e-01 -9.20827091e-01 -4.69364405e-01
2.40626246e-01 -3.29583794e-01 -7.72182941e-01 5.97635984e-01
-7.97568619e-01 3.59523654e-01 -1.24647923e-01 2.89774418e-01
1.96838513e-01 -2.42377132e-01 3.81649166e-01 1.92335084e-01
1.85553417e-01 7.42864847e-01 -8.08876216e-01 1.93026698e+00
-1.05626798e+00 5.30599475e-01 7.97438174e-02 -9.20887351e-01
9.37933505e-01 2.83819288e-01 2.44754449e-01 -4.76439834e-01
1.85873225e-01 3.54991823e-01 -1.17011890e-02 -2.80910194e-01
1.00607204e+00 -1.49347767e-01 -4.64115292e-01 8.91616046e-01
2.35939756e-01 -3.69495988e-01 4.40679252e-01 3.18148702e-01
1.16133380e+00 -2.38581151e-01 3.60935509e-01 -2.08258722e-02
5.45217574e-01 5.31996012e-01 4.64454629e-02 7.34460354e-01
-1.11863278e-01 4.59059656e-01 4.61386740e-01 -3.34478259e-01
-1.13272965e+00 -7.57774949e-01 1.20046876e-01 1.36278772e+00
-2.04251513e-01 -3.87467623e-01 -8.70643437e-01 -9.40165579e-01
-6.81223869e-02 8.67271602e-01 -3.25666994e-01 -1.62875459e-01
-7.80919790e-01 -5.42463958e-01 5.30591190e-01 4.42140341e-01
5.82865298e-01 -1.33675957e+00 -4.97023702e-01 5.96401453e-01
-3.74796599e-01 -1.13732898e+00 -3.63461822e-01 3.77384335e-01
-8.13842595e-01 -5.34995258e-01 -7.23387420e-01 -7.80816972e-01
6.05506718e-01 2.75461614e-01 1.36656952e+00 1.90991908e-01
-1.53258771e-01 1.15386307e-01 -4.64208454e-01 -1.16946585e-01
-1.03079534e+00 5.56805432e-01 -1.38634756e-01 -3.99637699e-01
4.94343668e-01 -3.73256534e-01 -2.91661471e-01 3.03314507e-01
-7.83463001e-01 -1.85570065e-02 4.42870677e-01 1.41225028e+00
-2.19177246e-01 -3.36387783e-01 9.48112667e-01 -1.01768255e+00
1.15780163e+00 -3.76190931e-01 -4.60015327e-01 2.02549547e-01
-6.05181694e-01 3.67379278e-01 8.19351971e-01 -4.05538708e-01
-1.32310700e+00 2.41758838e-01 -4.43734676e-01 -2.59805974e-02
-4.51502413e-01 4.68919784e-01 8.15435648e-02 1.44034982e-01
9.54052031e-01 3.22979420e-01 3.33932042e-01 -5.45631111e-01
5.78231573e-01 1.14176238e+00 3.26260149e-01 -6.61950529e-01
3.77575368e-01 1.87133476e-01 -5.14528990e-01 -1.00507462e+00
-7.93855846e-01 -5.81243455e-01 -4.82977778e-01 1.92514226e-01
5.89775920e-01 -6.30231321e-01 -6.38116658e-01 2.00022757e-01
-1.16079044e+00 -6.29897416e-01 -4.06540751e-01 8.85759369e-02
-4.87654269e-01 4.88418132e-01 -6.18257284e-01 -6.76236749e-01
-4.54322368e-01 -1.24590421e+00 1.14833689e+00 2.65782595e-01
-8.90372872e-01 -1.09674180e+00 2.62923747e-01 6.77921951e-01
6.35810554e-01 -5.78331053e-01 1.06058538e+00 -1.38212943e+00
-4.15624797e-01 -3.87189269e-01 -1.15904264e-01 2.98876286e-01
2.71982670e-01 -6.04165494e-01 -8.60085726e-01 -2.28669703e-01
1.69620708e-01 -1.07610190e+00 4.88424033e-01 -1.37082666e-01
7.74153411e-01 -2.48284832e-01 -1.91005915e-01 1.33333560e-02
8.35004568e-01 1.79466173e-01 8.60494152e-02 1.62284464e-01
1.80953860e-01 8.81600857e-01 6.82898104e-01 3.77089411e-01
4.94818509e-01 1.07640111e+00 -1.42153352e-01 5.86303249e-02
-3.96515392e-02 -3.78854722e-01 1.87254623e-01 4.85140175e-01
4.96012151e-01 -2.89425045e-01 -7.92325437e-01 6.39456511e-01
-1.94274867e+00 -7.26822197e-01 3.22841257e-01 2.17229438e+00
1.29858935e+00 2.24457204e-01 3.70591193e-01 -5.30927889e-02
3.77327204e-01 -9.78750810e-02 -5.01260936e-01 -4.18235362e-01
3.64091396e-01 3.48756790e-01 1.29090548e-01 7.24961996e-01
-8.48642111e-01 1.20357680e+00 6.08940554e+00 5.41431367e-01
-1.03011978e+00 -4.96442094e-02 6.13661408e-01 4.82130721e-02
-2.50464678e-01 8.12658444e-02 -9.71278846e-01 4.20535803e-01
1.08277488e+00 -2.67806888e-01 5.24608433e-01 1.14526856e+00
-7.72335604e-02 -2.86201477e-01 -1.30029833e+00 5.34462214e-01
1.05151989e-01 -1.35524344e+00 -1.79687500e-01 -1.32631436e-01
3.92522186e-01 -6.19729273e-02 -3.63510281e-01 7.86760211e-01
7.65312970e-01 -6.57106161e-01 1.39369100e-01 -3.05472821e-01
4.77731526e-01 -3.04657966e-01 8.40000987e-01 5.06691039e-01
-7.52409339e-01 -5.90368919e-02 -4.81626689e-01 -1.82940841e-01
3.66605192e-01 1.60441518e-01 -1.73903406e+00 2.02961490e-01
1.99827120e-01 1.38703272e-01 -5.62051654e-01 6.75236404e-01
-1.12338886e-01 4.10502791e-01 -4.17563558e-01 -6.16558015e-01
2.36498624e-01 1.14326142e-01 2.91796178e-01 1.12826836e+00
-7.62366652e-02 -1.79181010e-01 3.77710611e-01 7.51641810e-01
-1.88308522e-01 1.17776014e-01 -7.12025642e-01 3.01747136e-02
5.53203225e-01 1.44007540e+00 -7.61180043e-01 -4.75013882e-01
-2.76866138e-01 1.20963061e+00 5.23980141e-01 1.75058725e-03
-5.25404453e-01 -3.25038791e-01 4.24707890e-01 -2.87424982e-01
-2.06238148e-03 -2.50353694e-01 -3.26078743e-01 -1.20048892e+00
-3.12629305e-02 -1.17567730e+00 4.38445807e-01 -3.15897137e-01
-1.23301470e+00 8.39543045e-01 1.67982966e-01 -8.48808587e-01
-1.04077148e+00 -4.60058928e-01 -6.66032434e-01 8.42223942e-01
-9.08582747e-01 -9.37275171e-01 -3.52589428e-01 3.70333880e-01
1.23835850e+00 -4.13089186e-01 1.11841357e+00 6.74900189e-02
-1.99675515e-01 6.83988452e-01 -3.61610316e-02 5.91503866e-02
8.51854503e-01 -1.51736140e+00 8.27272892e-01 3.65080059e-01
3.94548297e-01 8.21071982e-01 9.58487689e-01 -4.38700616e-01
-1.08985388e+00 -5.88751435e-01 9.50301886e-01 -5.23994327e-01
8.76518548e-01 -8.63403380e-01 -1.02208316e+00 5.02177894e-01
5.16608953e-01 -7.38613963e-01 7.78625965e-01 5.98870575e-01
-1.06628321e-01 2.41612390e-01 -1.13928628e+00 8.91711235e-01
8.87752056e-01 -4.86913413e-01 -9.68932450e-01 6.37006342e-01
8.20004463e-01 -5.17149091e-01 -5.29566526e-01 8.19240585e-02
3.85541648e-01 -8.52687061e-01 7.56298602e-01 -7.46171713e-01
3.42235565e-01 1.57863081e-01 1.59382876e-02 -1.51594818e+00
3.04263473e-01 -5.66447496e-01 2.65278965e-01 1.34037876e+00
7.65183091e-01 -5.06412268e-01 1.27310777e+00 1.01145124e+00
-8.93662274e-02 -5.42812705e-01 -9.14045155e-01 -6.28396213e-01
-4.34676297e-02 -1.64318234e-02 5.50870657e-01 6.35452271e-01
7.00437188e-01 1.09855902e+00 -3.82060766e-01 -4.49441880e-01
1.14406377e-01 3.67348105e-01 1.25888371e+00 -1.10072684e+00
-6.90399468e-01 -2.55431861e-01 -1.29267192e-02 -1.36350894e+00
4.19890612e-01 -7.58123755e-01 3.58871311e-01 -1.37157273e+00
4.36504185e-02 -6.52407944e-01 3.53855848e-01 5.84146440e-01
-3.34815741e-01 -2.35472452e-02 2.07804888e-01 2.75411844e-01
-4.41912115e-01 5.83859086e-01 5.75023115e-01 -1.07063510e-01
-6.16024613e-01 3.64886761e-01 -5.94064951e-01 5.48892379e-01
8.18009973e-01 -3.02385718e-01 -7.77842045e-01 -1.21512935e-01
-2.38380313e-01 3.59340668e-01 -4.31218520e-02 -8.96366119e-01
2.43732244e-01 2.54242700e-02 -1.15440652e-01 -3.13013345e-01
6.66972756e-01 -8.32985222e-01 -3.09358567e-01 5.05195856e-01
-7.49271154e-01 -7.74179995e-02 2.24958301e-01 2.76847124e-01
-2.18474329e-01 -8.97842467e-01 6.77248716e-01 -3.40278387e-01
-5.86565435e-01 -3.10664058e-01 -4.60194170e-01 8.55273455e-02
8.42055976e-01 5.59013635e-02 -4.94613111e-01 -7.73905754e-01
-6.03784144e-01 2.36212060e-01 6.21096075e-01 5.42460322e-01
2.00453043e-01 -8.23016465e-01 -4.10514921e-01 2.60153294e-01
2.99535990e-01 -1.96855307e-01 -9.17382166e-02 3.11722666e-01
-3.21444392e-01 7.12565720e-01 -1.88369215e-01 -5.81945479e-01
-1.35907638e+00 2.75015116e-01 1.28905296e-01 -6.72377169e-01
-4.35067087e-01 7.81472743e-01 -7.78697059e-02 -7.71930039e-01
4.27443415e-01 -3.55010390e-01 -1.17643796e-01 1.33149877e-01
4.16038811e-01 -8.80892128e-02 1.63907066e-01 1.20941643e-02
-4.53150049e-02 -2.66481489e-01 -6.10413134e-01 -6.15436733e-01
1.09636426e+00 -6.51048794e-02 1.66330457e-01 3.94281864e-01
9.42336440e-01 -3.76227796e-02 -9.57433820e-01 -3.24082881e-01
3.70144486e-01 -2.74923235e-01 -5.77270746e-01 -7.92427242e-01
-3.60096514e-01 7.45634139e-01 1.82621270e-01 7.81831324e-01
8.28200102e-01 7.17378631e-02 9.92777050e-01 1.00018156e+00
5.27809501e-01 -9.08292890e-01 3.93319190e-01 7.85004437e-01
7.64744341e-01 -1.53485227e+00 -3.09882492e-01 -2.55790442e-01
-9.35344756e-01 9.39171433e-01 6.95091784e-01 1.94291860e-01
2.17273593e-01 1.03809588e-01 3.28506231e-01 -2.70075817e-02
-9.15190697e-01 -4.49166298e-02 -1.87845275e-01 4.24442410e-01
6.56720936e-01 -2.34895363e-01 -4.04205948e-01 3.02110642e-01
-4.80542511e-01 -1.35546297e-01 7.81439185e-01 1.03348482e+00
-6.67089999e-01 -1.76469421e+00 2.03296319e-02 5.18962026e-01
-1.14542276e-01 -2.06766427e-01 -7.91752040e-01 6.02327824e-01
-6.57174528e-01 1.13795412e+00 -2.35541221e-02 -3.05684566e-01
3.81542087e-01 5.90916157e-01 1.38965249e-01 -1.11834371e+00
-7.82049000e-01 -2.80720890e-01 7.00012028e-01 -2.75821149e-01
-1.86164137e-02 -5.83465040e-01 -1.11398733e+00 -2.00871006e-01
-4.66327190e-01 5.56231678e-01 7.40869403e-01 1.07562661e+00
4.48475271e-01 1.50674477e-01 6.16467297e-01 -8.37032139e-01
-1.16171253e+00 -1.31403482e+00 -1.37920395e-01 6.08181357e-01
2.71186400e-02 -5.35289824e-01 -3.24437529e-01 3.44876572e-02] | [12.762572288513184, 7.9821672439575195] |
ac96e2f9-0641-405c-9fad-f867df4a2027 | tabbie-pretrained-representations-of-tabular | 2105.02584 | null | https://arxiv.org/abs/2105.02584v1 | https://arxiv.org/pdf/2105.02584v1.pdf | TABBIE: Pretrained Representations of Tabular Data | Existing work on tabular representation learning jointly models tables and associated text using self-supervised objective functions derived from pretrained language models such as BERT. While this joint pretraining improves tasks involving paired tables and text (e.g., answering questions about tables), we show that it underperforms on tasks that operate over tables without any associated text (e.g., populating missing cells). We devise a simple pretraining objective (corrupt cell detection) that learns exclusively from tabular data and reaches the state-of-the-art on a suite of table based prediction tasks. Unlike competing approaches, our model (TABBIE) provides embeddings of all table substructures (cells, rows, and columns), and it also requires far less compute to train. A qualitative analysis of our model's learned cell, column, and row representations shows that it understands complex table semantics and numerical trends. | ['Mohit Iyyer', 'Varun Manjunatha', 'Dung Thai', 'Hiroshi Iida'] | 2021-05-06 | null | https://aclanthology.org/2021.naacl-main.270 | https://aclanthology.org/2021.naacl-main.270.pdf | naacl-2021-4 | ['cell-detection', 'table-annotation', 'table-annotation', 'column-type-annotation'] | ['computer-vision', 'knowledge-base', 'natural-language-processing', 'natural-language-processing'] | [ 3.79602350e-02 4.99540716e-01 -4.65515286e-01 -3.27665478e-01
-1.05576110e+00 -8.92326415e-01 6.27962470e-01 1.14999247e+00
4.78105694e-02 8.99132490e-01 6.09195352e-01 -6.61391914e-01
2.55524218e-01 -1.16732657e+00 -1.24679959e+00 -1.32979617e-01
-1.74318492e-01 1.05282295e+00 -1.31456837e-01 -3.91392946e-01
3.40398520e-01 9.55180749e-02 -1.41907799e+00 9.32025075e-01
7.20194757e-01 1.13865697e+00 -4.96155679e-01 4.72207695e-01
-7.68276036e-01 1.46611714e+00 -5.05187035e-01 -7.01833963e-01
4.55796048e-02 -1.65367514e-01 -7.56488621e-01 -2.25383282e-01
7.41420448e-01 -2.70117790e-01 -4.01205331e-01 6.41785741e-01
-6.15367703e-02 -7.94442445e-02 7.53740013e-01 -1.24567163e+00
-1.00729132e+00 1.29554534e+00 -4.10425037e-01 -7.14327991e-02
3.72202247e-01 -1.47098303e-01 1.68641448e+00 -9.22257602e-01
7.42516518e-01 1.43095255e+00 9.79522765e-01 2.13791236e-01
-1.71137989e+00 -3.85610908e-01 4.06496376e-01 -3.98201883e-01
-1.13372147e+00 -3.85100842e-01 2.55912066e-01 -4.29364264e-01
1.17015505e+00 3.62414062e-01 1.08251311e-01 8.30575109e-01
3.37666214e-01 1.04105067e+00 8.06985140e-01 -1.28593788e-01
1.90485299e-01 6.46748006e-01 2.37094238e-01 7.39004016e-01
9.62958574e-01 -5.53296685e-01 -6.67240977e-01 -1.28383234e-01
6.07329667e-01 -5.30001558e-02 1.02859728e-01 -8.23134184e-01
-1.47995865e+00 8.06470692e-01 2.79774427e-01 1.40823931e-01
1.67132333e-01 4.05936539e-01 6.96444273e-01 3.14580768e-01
3.38233143e-01 6.70969784e-01 -9.64097500e-01 -1.57358989e-01
-7.79230833e-01 4.50987130e-01 1.23004317e+00 1.44465590e+00
8.20454836e-01 -1.21454239e-01 -2.47241750e-01 6.80074573e-01
-2.68845563e-03 3.56384188e-01 5.78739680e-02 -5.69460273e-01
1.02268600e+00 9.65304732e-01 9.82365981e-02 -8.68257701e-01
-4.26808357e-01 -3.50639731e-01 -7.53948808e-01 -2.56280482e-01
6.57828510e-01 -4.84072790e-02 -9.12769377e-01 1.44594693e+00
-2.19519705e-01 -4.76783842e-01 1.73572183e-01 2.36660868e-01
1.09022069e+00 7.81596780e-01 -9.14239231e-03 1.02908015e-01
1.27170515e+00 -9.02443469e-01 -8.11181426e-01 -5.05449295e-01
1.21509445e+00 -4.78936285e-01 1.23637080e+00 3.86789620e-01
-1.26591527e+00 -4.16514695e-01 -1.02715290e+00 -7.56179571e-01
-1.13700020e+00 5.72990216e-02 1.14117467e+00 6.35700583e-01
-1.06244445e+00 5.12688875e-01 -7.18863726e-01 -3.85227114e-01
3.87544185e-01 2.91203082e-01 -4.99098033e-01 -1.00767791e-01
-1.06023920e+00 9.90049005e-01 4.10489142e-01 -1.55935064e-01
-4.23351884e-01 -9.94182050e-01 -1.37445617e+00 3.24124306e-01
6.71188056e-01 -4.68219489e-01 1.07240701e+00 -3.18178922e-01
-8.89053166e-01 9.91464734e-01 -1.79741934e-01 -7.16352761e-01
3.17117721e-01 -3.38988543e-01 -1.86127439e-01 -3.73599321e-01
4.79520746e-02 8.00637484e-01 1.46617904e-01 -1.60626769e+00
-4.23922807e-01 -2.07694441e-01 -2.19712351e-02 2.66344682e-03
-3.53183359e-01 -3.56611967e-01 -4.62184370e-01 -6.77108586e-01
2.15223327e-01 -3.61070663e-01 -2.83895358e-02 1.35699585e-01
-7.39395618e-01 -1.53106853e-01 2.95204490e-01 -7.06224144e-01
1.60386026e+00 -1.90196919e+00 -6.08833618e-02 2.51878262e-01
3.25047046e-01 -5.25671244e-01 1.45886645e-01 8.46805394e-01
-1.53068393e-01 6.29906774e-01 -3.39121372e-01 -3.91653001e-01
4.91188765e-01 2.53241718e-01 -5.93809724e-01 5.40230684e-02
4.25285220e-01 1.06555259e+00 -6.51997685e-01 -6.38361156e-01
-6.37789592e-02 7.55162016e-02 -1.06696451e+00 -8.77941325e-02
-7.23359346e-01 -5.29072881e-01 4.53833975e-02 1.08708799e+00
4.79136139e-01 -5.52107930e-01 5.88184416e-01 -2.28449665e-02
1.13369962e-02 8.97634745e-01 -1.34611058e+00 1.45603192e+00
-2.72047490e-01 4.78826851e-01 -1.25700161e-01 -7.52829731e-01
9.80919957e-01 -2.18265146e-01 3.73773128e-01 -5.06894231e-01
-2.85362769e-02 3.49056005e-01 -1.86061099e-01 -1.00501873e-01
8.15904498e-01 1.75424397e-01 -5.45025408e-01 4.16221529e-01
2.35475034e-01 -3.17359149e-01 4.72480386e-01 6.08682513e-01
1.05454218e+00 2.81814367e-01 3.83379191e-01 -5.68081975e-01
3.82759809e-01 3.86532098e-01 1.93381295e-01 8.75535190e-01
3.16779584e-01 5.23905754e-01 1.24247217e+00 -5.87120116e-01
-1.11095881e+00 -1.17078114e+00 -2.69730568e-01 1.22842371e+00
2.10138364e-03 -1.11701214e+00 -3.85858953e-01 -7.32984364e-01
7.91999757e-01 8.17851901e-01 -9.90367770e-01 1.22069567e-01
-6.07100427e-01 -4.79254961e-01 2.96205908e-01 1.02997279e+00
-6.03553466e-02 -8.90532255e-01 7.18775615e-02 3.12224686e-01
-8.59867707e-02 -8.82107317e-01 -3.28190207e-01 1.02662718e+00
-1.01194131e+00 -9.32309270e-01 1.96537763e-01 -9.50503051e-01
6.70373440e-01 -3.04982811e-01 1.75699270e+00 -6.49924390e-03
-9.49573293e-02 1.42268300e-01 6.27707019e-02 -4.55772072e-01
-3.40310186e-01 3.64164174e-01 -3.32619667e-01 -4.33500826e-01
4.02948529e-01 -2.48799428e-01 2.06712663e-01 -7.04256147e-02
-8.86606932e-01 1.09690316e-01 4.21173096e-01 1.00467229e+00
5.76808989e-01 -4.60657328e-02 2.44528279e-01 -1.55538905e+00
4.27361071e-01 -6.16972804e-01 -6.30068064e-01 5.43400347e-01
-6.46923423e-01 6.46587491e-01 7.96155453e-01 -5.83116040e-02
-5.59156179e-01 -2.78629839e-01 1.59462079e-01 1.89665554e-03
4.26470414e-02 6.78539336e-01 -2.41832897e-01 5.73314011e-01
5.34032702e-01 1.14329189e-01 2.88872924e-02 -5.31330645e-01
5.17749727e-01 2.63456076e-01 6.57218516e-01 -8.03790927e-01
1.12286735e+00 3.16856205e-01 -1.49850518e-01 -1.81365952e-01
-7.86159098e-01 -1.71890825e-01 -6.99245870e-01 3.81645203e-01
7.34399915e-01 -1.21242034e+00 -7.90977776e-01 -9.41731408e-02
-8.71053457e-01 -4.97431070e-01 -3.87927204e-01 -5.53030632e-02
-5.95628917e-01 -1.10217720e-01 -9.17597175e-01 -6.56401515e-01
-1.63957905e-02 -6.63644254e-01 1.23220789e+00 -5.03626764e-01
-6.31976306e-01 -1.24799681e+00 -2.07890034e-01 1.26196176e-01
2.36038402e-01 2.88316131e-01 1.75760686e+00 -9.78178322e-01
-9.62458551e-01 -1.15468517e-01 -3.32777739e-01 -1.34497315e-01
1.08858630e-01 9.77746025e-02 -6.70265496e-01 9.20758769e-03
-5.26780784e-01 -7.81620324e-01 1.08302975e+00 1.08723631e-02
1.47633517e+00 -5.75326502e-01 -4.63232368e-01 6.88361168e-01
1.49225402e+00 1.01293631e-01 7.00422168e-01 5.61224103e-01
6.57873631e-01 6.80844903e-01 4.88559246e-01 4.56534177e-01
8.41205537e-01 9.28646177e-02 3.20944011e-01 -6.92616627e-02
1.37321457e-01 -8.30897272e-01 2.43508130e-01 4.43225771e-01
6.59881294e-01 -5.70549428e-01 -1.05474412e+00 5.64760804e-01
-1.88780642e+00 -6.80060923e-01 -1.19323038e-01 2.06776929e+00
1.11101449e+00 7.98109889e-01 2.39429027e-02 2.70887554e-01
2.53107101e-01 2.42830053e-01 -6.02757990e-01 -6.14658952e-01
-3.47944230e-01 3.33667099e-01 7.59388626e-01 5.53255558e-01
-1.25694740e+00 1.10745740e+00 7.29450178e+00 3.77572000e-01
-5.41674256e-01 -5.01987576e-01 1.18724096e+00 1.01191457e-02
-9.56077814e-01 -6.11851104e-02 -9.70274806e-01 4.65583615e-02
8.68117988e-01 2.19600368e-02 5.93583763e-01 9.60628927e-01
-3.74879628e-01 -1.82219774e-01 -1.70055866e+00 9.19358611e-01
3.05706114e-01 -1.74917543e+00 4.89778310e-01 -6.33192658e-02
5.58675349e-01 -5.10145783e-01 1.57286808e-01 8.37950528e-01
7.16509044e-01 -1.51685882e+00 9.31569636e-01 2.58896708e-01
9.86911476e-01 -6.16221189e-01 7.80146360e-01 -5.70999011e-02
-1.08291936e+00 -2.02486619e-01 -2.15054199e-01 -1.84374675e-01
-4.95467991e-01 3.66169810e-01 -5.99013567e-01 3.03190082e-01
6.07610464e-01 1.02787125e+00 -1.08459365e+00 3.74986678e-01
1.49751887e-01 4.60587084e-01 -2.22491980e-01 -1.49721488e-01
1.86568111e-01 -3.73577811e-02 -1.41864091e-01 1.34895504e+00
6.20497614e-02 -2.05324739e-01 7.61922225e-02 1.08329809e+00
-5.80612719e-01 1.68487489e-01 -6.64474189e-01 -4.77818370e-01
5.48681796e-01 9.48791921e-01 -8.86695385e-01 -6.19068205e-01
-5.00836015e-01 3.21264803e-01 5.47993243e-01 1.35294467e-01
-6.79705024e-01 -3.37181687e-01 7.03759134e-01 4.01373923e-01
5.41402638e-01 -1.24147184e-01 -1.43092942e+00 -1.30560684e+00
5.63967414e-02 -1.12688887e+00 6.04755759e-01 -8.16392541e-01
-1.18431425e+00 3.27967823e-01 -3.73075195e-02 -8.35165501e-01
-3.25796664e-01 -9.89577770e-01 -1.63964942e-01 6.99711382e-01
-1.14387774e+00 -8.15766513e-01 -1.13772452e-02 3.22469056e-01
2.30455697e-01 -2.25824006e-02 8.91591609e-01 -6.43325672e-02
-5.42446315e-01 9.54181731e-01 1.73311859e-01 4.93161976e-01
9.14917469e-01 -1.82059741e+00 7.78150618e-01 5.38238347e-01
7.34407306e-02 1.02164721e+00 7.26959229e-01 -7.16993868e-01
-1.92907751e+00 -9.83431697e-01 1.38676596e+00 -1.11355805e+00
1.02469206e+00 -1.29181385e+00 -9.59116340e-01 1.20865715e+00
2.06323564e-01 7.02358857e-02 5.47564030e-01 6.67203784e-01
-8.61638427e-01 -4.10497695e-01 -9.20660853e-01 7.51963556e-01
1.06780934e+00 -5.83972335e-01 -5.64437807e-01 3.33539665e-01
7.96441495e-01 -5.91386735e-01 -9.28677797e-01 1.98089838e-01
5.96167266e-01 -7.29959846e-01 9.53817964e-01 -8.29484224e-01
9.33119535e-01 -1.46110862e-01 -4.70057994e-01 -1.01125669e+00
-2.73790926e-01 -4.87918884e-01 -5.74155092e-01 1.19248140e+00
1.00833714e+00 -5.11974335e-01 1.05803549e+00 6.49718523e-01
-4.29927818e-02 -8.27734530e-01 -1.54761225e-01 -5.78139067e-01
5.27501404e-01 -1.02788627e-01 7.24172950e-01 1.02396011e+00
3.60771328e-01 3.29770476e-01 -1.28684826e-02 -1.96022511e-01
3.07270497e-01 3.42912465e-01 1.00467193e+00 -1.30254197e+00
-9.08933952e-03 -5.33030629e-01 1.40909459e-02 -9.82522249e-01
-9.57403705e-02 -1.14337265e+00 -1.16319330e-02 -1.91235483e+00
9.80659053e-02 -5.85775554e-01 -2.94441670e-01 7.80123472e-01
-2.43726835e-01 -2.99361087e-02 2.05453128e-01 -7.51010552e-02
-6.60722852e-01 1.33705899e-01 9.25819337e-01 -3.57450306e-01
-1.79305747e-02 -6.42612040e-01 -1.32895136e+00 3.32910508e-01
6.86749578e-01 -4.96058583e-01 -2.19846427e-01 -3.09684038e-01
7.28646398e-01 1.76458478e-01 -9.35484841e-02 -1.00253403e+00
2.08962172e-01 -1.30490690e-01 8.48578513e-01 -1.07061756e+00
4.86029014e-02 -6.66537285e-01 -3.03930104e-01 2.35140294e-01
-9.58500087e-01 5.01028836e-01 6.87309027e-01 3.54811490e-01
-3.62978727e-01 2.44727969e-01 2.82577664e-01 -2.77920485e-01
-2.12161675e-01 3.35095860e-02 -3.77721578e-01 4.06619430e-01
5.11278987e-01 -2.47074053e-01 -6.29811823e-01 -3.83497059e-01
-5.02498209e-01 5.55279315e-01 5.73603094e-01 4.63419467e-01
2.78525591e-01 -1.21022224e+00 -4.89837766e-01 3.74253601e-01
4.83214080e-01 1.68805882e-01 -4.99571055e-01 3.54890049e-01
-7.24679828e-01 7.29922473e-01 3.66434753e-02 -2.83288002e-01
-6.90906703e-01 7.92676330e-01 2.59153340e-02 -9.13768947e-01
-2.69789815e-01 8.57198477e-01 1.95626393e-01 -9.91134346e-01
6.68644130e-01 -1.28497684e+00 1.06017152e-02 2.60773391e-01
2.66324610e-01 -1.28220588e-01 3.21804285e-01 1.42351627e-01
-4.34039682e-01 -2.12035179e-02 -3.49112302e-01 1.81523293e-01
1.31507742e+00 1.74236104e-01 -3.64756823e-01 9.10401106e-01
1.09253585e+00 3.44946057e-01 -1.01324522e+00 -1.03836253e-01
3.19050163e-01 -5.69147058e-02 -3.69641662e-01 -1.18634069e+00
-4.34522957e-01 8.33810091e-01 -2.51801342e-01 1.52680486e-01
6.29370451e-01 -1.59568831e-01 6.16694868e-01 8.57190967e-01
3.93562764e-01 -9.14330900e-01 4.79274569e-03 7.56705701e-01
7.11329758e-01 -1.33738339e+00 1.86275959e-01 -5.52362800e-01
-6.23207211e-01 1.29068089e+00 8.68871748e-01 -1.23048043e-02
5.99274576e-01 9.79739726e-01 -8.46606791e-02 -8.18243995e-02
-1.35616004e+00 -8.24282691e-02 1.74014300e-01 3.76934022e-01
8.40246499e-01 9.34668351e-03 1.01117432e-01 6.19474530e-01
-6.64586484e-01 -3.08743447e-01 6.80966496e-01 1.22515130e+00
-4.06909496e-01 -1.02200973e+00 -5.04793227e-01 1.08570862e+00
-2.82433867e-01 -6.25724852e-01 -6.86660945e-01 1.26014555e+00
-4.88534744e-04 7.21807122e-01 5.20953000e-01 -3.97806197e-01
1.35457993e-01 4.25562978e-01 3.40290070e-01 -9.74450529e-01
-8.88663590e-01 -6.57806322e-02 3.20778638e-01 -4.35663730e-01
2.21170768e-01 -7.20415056e-01 -1.21418774e+00 -7.28442073e-01
1.10984489e-01 2.32669383e-01 2.36893460e-01 4.69632953e-01
2.43111446e-01 6.85544789e-01 4.12378199e-02 -1.15372576e-01
-4.20210868e-01 -8.62358630e-01 -5.32707572e-01 3.73296946e-01
4.69260037e-01 -4.83371973e-01 -1.83050722e-01 -5.17982878e-02] | [9.659161567687988, 7.80434513092041] |
a1af1e56-fc82-4d38-a34b-fa166344bbbf | let-s-not-quote-out-of-context-unified-vision | 2306.00931 | null | https://arxiv.org/abs/2306.00931v1 | https://arxiv.org/pdf/2306.00931v1.pdf | "Let's not Quote out of Context": Unified Vision-Language Pretraining for Context Assisted Image Captioning | Well-formed context aware image captions and tags in enterprise content such as marketing material are critical to ensure their brand presence and content recall. Manual creation and updates to ensure the same is non trivial given the scale and the tedium towards this task. We propose a new unified Vision-Language (VL) model based on the One For All (OFA) model, with a focus on context-assisted image captioning where the caption is generated based on both the image and its context. Our approach aims to overcome the context-independent (image and text are treated independently) nature of the existing approaches. We exploit context by pretraining our model with datasets of three tasks: news image captioning where the news article is the context, contextual visual entailment, and keyword extraction from the context. The second pretraining task is a new VL task, and we construct and release two datasets for the task with 1.1M and 2.2K data instances. Our system achieves state-of-the-art results with an improvement of up to 8.34 CIDEr score on the benchmark news image captioning datasets. To the best of our knowledge, ours is the first effort at incorporating contextual information in pretraining the models for the VL tasks. | ['Sumit Shekhar', 'Niyati Chhaya', 'Pushpak Bhattacharyya', 'Abisek Rajakumar Kalarani'] | 2023-06-01 | null | null | null | null | ['image-captioning', 'marketing', 'keyword-extraction', 'visual-entailment'] | ['computer-vision', 'miscellaneous', 'natural-language-processing', 'reasoning'] | [ 6.61486566e-01 -2.04090811e-02 -3.26876819e-01 -4.77792591e-01
-1.26492584e+00 -8.03486884e-01 1.09477174e+00 1.03046671e-01
-5.47445357e-01 4.25866634e-01 4.82504219e-01 -4.99290377e-01
4.21016634e-01 -1.55618131e-01 -1.22726059e+00 -4.07686710e-01
3.94935399e-01 4.88802075e-01 1.08792372e-01 -3.13871145e-01
2.38072067e-01 -2.69259270e-02 -1.56796873e+00 1.12339783e+00
4.23670560e-01 1.20609295e+00 6.07109129e-01 8.21161091e-01
-3.47988129e-01 8.15208912e-01 -3.25779706e-01 -8.56496692e-01
2.35408962e-01 -3.73828113e-01 -1.01074195e+00 4.02162075e-01
1.00170732e+00 -2.52899140e-01 -8.41432065e-02 8.34084988e-01
2.69665778e-01 -2.50154048e-01 4.76286054e-01 -1.33290517e+00
-1.31865716e+00 8.43267739e-01 -7.92884171e-01 1.22599900e-01
2.59153247e-01 3.19407135e-02 1.27688074e+00 -1.12686658e+00
9.54073787e-01 1.07221532e+00 3.61269861e-01 5.55721700e-01
-9.66015160e-01 -4.31775004e-01 5.96210897e-01 2.76253939e-01
-1.22593832e+00 -4.11516905e-01 6.14850819e-01 -5.14124632e-01
7.19950795e-01 2.31036201e-01 2.36208305e-01 1.50375855e+00
-7.45224133e-02 9.44003403e-01 1.06771171e+00 -6.87279522e-01
-7.01519698e-02 5.06631911e-01 4.16335603e-03 2.25979954e-01
-1.13501735e-02 -3.13861459e-01 -1.44249603e-01 1.03566773e-01
5.65488100e-01 -1.70394674e-01 -3.56752202e-02 -3.55554640e-01
-1.27805364e+00 9.58151758e-01 3.81857365e-01 2.45436579e-01
-3.53452563e-01 3.91114116e-01 4.45453644e-01 5.86057343e-02
3.34677994e-01 4.09413010e-01 -4.11897182e-01 1.48886472e-01
-9.70358610e-01 2.82026350e-01 6.76715612e-01 1.26522887e+00
4.17069048e-01 -4.03965682e-01 -3.76611203e-01 8.56416166e-01
3.26940835e-01 6.25818372e-01 4.44681495e-01 -7.27600217e-01
5.77564418e-01 3.55433345e-01 3.87365788e-01 -9.69524801e-01
-5.58616593e-04 -5.86815238e-01 -4.21680540e-01 -3.33028346e-01
2.16855407e-01 4.14480157e-02 -1.35182583e+00 1.73211288e+00
1.73397988e-01 4.22679096e-01 2.44416088e-01 8.92862678e-01
1.13805413e+00 9.78333175e-01 5.42826951e-01 -8.84999055e-03
1.93189991e+00 -1.34450305e+00 -8.13391566e-01 -8.52106214e-01
5.34496009e-01 -1.24661505e+00 1.07813120e+00 9.84072685e-02
-8.58362317e-01 -6.44430399e-01 -1.02564418e+00 -1.67076170e-01
-4.93121922e-01 2.99459428e-01 5.01777589e-01 3.58360350e-01
-1.14842522e+00 -2.03863829e-01 -3.92371453e-02 -4.40050900e-01
2.56542444e-01 -5.68978935e-02 -4.68376815e-01 -5.19428432e-01
-1.31677258e+00 8.59375596e-01 3.04995060e-01 -4.18331511e-02
-9.02885437e-01 -4.92443919e-01 -9.70969141e-01 -2.40804553e-01
5.00342250e-01 -5.30284941e-01 1.55369115e+00 -1.60274959e+00
-7.44873941e-01 1.31993651e+00 -2.14394438e-03 -6.74395561e-01
3.32535774e-01 -2.87667304e-01 -6.38853788e-01 1.83002114e-01
4.26693678e-01 1.10182381e+00 9.49135005e-01 -1.95423841e+00
-9.44165885e-01 -5.66501729e-02 2.68706679e-01 2.66332865e-01
-2.45137364e-02 3.39557230e-01 -1.12169719e+00 -7.39969671e-01
-2.05409870e-01 -1.06728494e+00 -1.47187859e-01 -2.28480682e-01
-1.43254921e-01 1.03948191e-01 8.22548866e-01 -9.69213128e-01
1.02896059e+00 -2.14646053e+00 -2.44916186e-01 -1.90174326e-01
-1.51612982e-01 1.96621105e-01 -5.99354506e-01 4.53192472e-01
-1.21758528e-01 1.73716784e-01 -1.07957907e-01 -6.29632175e-01
6.63163513e-02 3.62443686e-01 -5.89116693e-01 5.48742246e-03
1.59258217e-01 1.08780086e+00 -9.20618892e-01 -6.72615707e-01
-7.83073455e-02 5.63514650e-01 -3.86948228e-01 1.72495380e-01
-6.11951828e-01 2.33941168e-01 -1.19459659e-01 5.94352365e-01
6.07894957e-01 -4.76496488e-01 3.69061410e-01 -5.09238183e-01
2.00286694e-02 -1.92071646e-02 -1.09370267e+00 1.78950477e+00
-6.89643383e-01 6.80436373e-01 1.53003037e-01 -7.94465899e-01
6.51766896e-01 4.18265134e-01 3.39453608e-01 -1.02499449e+00
5.51578961e-02 1.07803434e-01 -3.71964693e-01 -6.82475865e-01
9.18653429e-01 -3.87368910e-02 -5.15053689e-01 2.39894792e-01
7.67665654e-02 1.81060970e-01 2.36944064e-01 5.47444165e-01
5.03007114e-01 3.68141204e-01 8.16447213e-02 8.44694078e-02
4.08110350e-01 2.72268444e-01 1.73623189e-01 9.36420858e-01
-1.89415008e-01 8.02106798e-01 2.60865390e-01 -3.34765524e-01
-1.33724248e+00 -6.57006621e-01 5.50503731e-02 1.47359192e+00
3.54626954e-01 -3.46474797e-01 -6.59429312e-01 -8.12531650e-01
-2.26893708e-01 9.48539734e-01 -8.00268352e-01 1.75250724e-01
-5.43486714e-01 -4.06896561e-01 2.20141292e-01 5.53512633e-01
4.44423288e-01 -1.01025391e+00 -1.11286402e-01 2.48329997e-01
-6.22556448e-01 -1.75215924e+00 -8.71211112e-01 -2.01704562e-01
-2.35976577e-01 -8.36978018e-01 -7.37332761e-01 -1.29931641e+00
4.25336182e-01 3.54061246e-01 1.34914100e+00 -1.41857758e-01
-1.40784204e-01 6.86241031e-01 -5.14959216e-01 -4.37131405e-01
-6.00443244e-01 -2.53376812e-01 -4.85613227e-01 3.51777554e-01
3.23030174e-01 -3.37771773e-02 -7.68997133e-01 3.24512482e-01
-1.19880319e+00 5.11114120e-01 8.96686375e-01 7.71097362e-01
8.63731503e-01 -4.87313330e-01 5.46045303e-01 -1.01435137e+00
4.44373637e-01 -6.72143459e-01 -1.74787223e-01 5.71029067e-01
-5.58448553e-01 -1.12017773e-01 1.69617057e-01 -5.94898522e-01
-1.10728860e+00 4.67247277e-01 -6.19213395e-02 -2.99908966e-01
-2.47180954e-01 4.73354518e-01 -1.04535379e-01 2.89283872e-01
5.06915152e-01 2.44133919e-01 -3.27536881e-01 -4.15134281e-01
1.00807416e+00 9.49709535e-01 1.00269628e+00 -4.81211841e-01
4.42965299e-01 4.90532964e-01 -4.31346089e-01 -4.39473420e-01
-1.27596629e+00 -8.11113477e-01 -4.36896831e-01 -4.29127961e-01
1.00449955e+00 -1.36077285e+00 -8.55413154e-02 1.18437901e-01
-1.28804600e+00 -3.24834175e-02 -1.31839126e-01 2.65259206e-01
-5.14122963e-01 4.68874156e-01 -4.96877909e-01 -7.01664448e-01
-4.79804248e-01 -1.27555144e+00 1.21544385e+00 -7.25386515e-02
1.53848767e-01 -6.94795012e-01 -1.54940143e-01 8.37308228e-01
3.80337715e-01 3.11514378e-01 4.51161176e-01 -7.68303096e-01
-5.00166476e-01 -3.02140504e-01 -6.48220778e-01 3.70981127e-01
-1.60030946e-01 -3.06197613e-01 -1.09293866e+00 -1.81985766e-01
-2.41010755e-01 -4.04658198e-01 9.03516948e-01 2.69641697e-01
9.07884538e-01 -5.06269753e-01 -3.33383292e-01 1.83187217e-01
1.66125059e+00 1.91064224e-01 7.05306768e-01 6.15663767e-01
8.07479322e-01 7.05727279e-01 8.82573962e-01 1.99770659e-01
6.98170364e-01 8.86317670e-01 7.26568937e-01 -3.18176031e-01
-4.03725743e-01 -3.45552713e-01 2.34236121e-01 5.43005347e-01
3.60702634e-01 -5.69417953e-01 -8.55581343e-01 1.06077254e+00
-2.07385993e+00 -9.64334369e-01 -2.83778876e-01 1.82618451e+00
9.81238902e-01 9.97743011e-02 -9.75882262e-02 -5.19217432e-01
9.69392121e-01 1.19641066e-01 -2.94536412e-01 -5.44875801e-01
-1.30578250e-01 -3.00658941e-01 6.55429065e-01 4.81628597e-01
-1.54656672e+00 1.04937482e+00 5.79159117e+00 7.75527537e-01
-9.06246841e-01 4.21934187e-01 8.38068604e-01 3.65091972e-02
-3.55928034e-01 1.75981522e-01 -9.22284901e-01 5.30611455e-01
1.15817010e+00 3.06492776e-01 4.37293462e-02 1.11910641e+00
-1.11255810e-01 1.20775394e-01 -1.00850689e+00 1.21464050e+00
6.54962182e-01 -1.44257128e+00 4.45974439e-01 -7.54879192e-02
8.79703283e-01 -1.21049378e-02 3.10981393e-01 4.78039086e-01
1.45824745e-01 -8.77388537e-01 1.10164344e+00 3.60556632e-01
9.61072564e-01 -6.18526518e-01 9.72411513e-01 -5.18299900e-02
-1.06515825e+00 1.11020952e-02 7.24104419e-02 2.47870430e-01
4.82568890e-01 3.45757216e-01 -9.24369931e-01 3.58340353e-01
5.47682106e-01 5.00071764e-01 -7.31563747e-01 8.40224326e-01
-2.60320324e-02 4.49712723e-01 -5.74613288e-02 1.48004815e-01
7.01489747e-01 2.09419861e-01 3.06633681e-01 1.67575324e+00
6.28635958e-02 -2.13439822e-01 3.76858532e-01 4.03379649e-01
-4.08765137e-01 3.31782043e-01 -3.74721169e-01 -8.60518217e-02
2.69089401e-01 1.37579024e+00 -7.44162261e-01 -5.31837106e-01
-7.24555850e-01 1.20608079e+00 5.95789403e-03 2.64592499e-01
-1.11498892e+00 -8.72314274e-02 3.96237165e-01 2.62759507e-01
6.74161077e-01 1.83594953e-02 -1.29543692e-01 -9.59044039e-01
1.07961752e-01 -9.11630452e-01 6.51071966e-01 -1.17061245e+00
-1.36695135e+00 8.69060993e-01 9.51298773e-02 -1.08115101e+00
-2.11184725e-01 -6.20039046e-01 -2.31273964e-01 6.17833018e-01
-1.90394020e+00 -1.89781165e+00 -3.32397103e-01 4.46251720e-01
8.12407494e-01 2.23062292e-01 4.71866310e-01 5.72309613e-01
-1.21704526e-01 5.87796390e-01 7.63961002e-02 8.56343582e-02
1.10352087e+00 -1.05340528e+00 5.23836613e-01 9.41622376e-01
3.14735800e-01 3.44839841e-01 9.01673019e-01 -6.66955829e-01
-1.20466697e+00 -1.25992572e+00 1.18141425e+00 -6.97553158e-01
8.09562802e-01 -6.35668635e-01 -5.56722701e-01 8.32881927e-01
6.43161178e-01 -1.08370021e-01 7.00394094e-01 -2.26795360e-01
-6.27722979e-01 8.04116800e-02 -8.84595752e-01 5.87992847e-01
8.35667491e-01 -5.24390221e-01 -6.07331514e-01 6.04589224e-01
1.19237566e+00 -3.77900183e-01 -6.86893344e-01 1.77694961e-01
4.55865532e-01 -3.55300248e-01 1.14242899e+00 -5.53038955e-01
7.04240501e-01 -3.76016825e-01 -4.80368584e-01 -6.74494982e-01
-2.90779322e-01 -3.23867679e-01 5.17489910e-02 1.59632635e+00
6.80153489e-01 -1.02660567e-01 3.73729080e-01 4.74982649e-01
-1.97655022e-01 -4.20071006e-01 -7.37995207e-01 -5.38223803e-01
-2.35149026e-01 -5.74851394e-01 4.20857817e-01 1.08590615e+00
-4.20667380e-01 6.60309494e-01 -6.49863422e-01 2.22473830e-01
4.23670143e-01 3.84721398e-01 6.15221918e-01 -8.18375766e-01
-2.29607880e-01 -1.46875933e-01 -2.02320293e-01 -1.00771260e+00
-1.85349330e-01 -8.81596506e-01 1.72432244e-01 -1.90136528e+00
5.74241221e-01 -1.81466237e-01 -2.20078528e-01 5.80224156e-01
-1.34785950e-01 5.66979527e-01 3.25456083e-01 2.77645886e-01
-1.06705606e+00 9.28454399e-02 1.17472458e+00 -2.40345314e-01
1.47585362e-01 -4.49884892e-01 -1.14128149e+00 3.96168351e-01
6.97344124e-01 -2.80034631e-01 -4.28381056e-01 -5.12551785e-01
1.96499124e-01 -2.65998036e-01 5.50119102e-01 -5.79554856e-01
3.03211641e-02 -1.69077203e-01 2.24626839e-01 -6.50086284e-01
4.31015432e-01 -9.50738728e-01 1.55041412e-01 -1.64880827e-02
-6.26723766e-01 2.56843656e-01 3.88978004e-01 7.50113964e-01
-2.50770956e-01 -2.83918023e-01 6.48963392e-01 -2.99055666e-01
-1.26401997e+00 1.30251512e-01 -4.95377854e-02 1.84234217e-01
1.10251617e+00 -2.28120144e-02 -5.30714571e-01 -6.54527307e-01
-7.49212742e-01 3.32139432e-01 2.84434289e-01 8.50942612e-01
4.91671681e-01 -1.40460300e+00 -1.02954865e+00 -2.03599364e-01
6.36786222e-01 -3.33380967e-01 4.21663523e-01 4.92015898e-01
-3.28655213e-01 6.20330691e-01 -5.77323288e-02 -6.30649686e-01
-1.37976360e+00 1.08350384e+00 -2.60077920e-02 -3.80231202e-01
-3.69833410e-01 7.32983828e-01 4.89061177e-01 -5.71312830e-02
2.68536270e-01 -1.24788277e-01 -4.61052090e-01 1.06004260e-01
7.70194650e-01 -4.76533562e-01 7.32108951e-02 -1.10491526e+00
-2.58072227e-01 3.97469550e-01 -5.40835440e-01 -2.72477984e-01
1.21467590e+00 -5.17123759e-01 8.91114026e-02 1.34276822e-01
1.28823483e+00 -2.25501776e-01 -1.18564045e+00 -4.75038290e-01
6.96420763e-03 -3.18050057e-01 2.23828912e-01 -1.25211632e+00
-7.97899306e-01 5.00976682e-01 7.40360141e-01 2.43635420e-02
1.07121444e+00 3.92538935e-01 9.53622043e-01 -7.54197733e-03
-7.80679733e-02 -1.25276625e+00 2.08407000e-01 3.26485813e-01
1.24138761e+00 -1.58107579e+00 -9.01469290e-02 -2.54543602e-01
-1.27135181e+00 6.45722270e-01 6.78481996e-01 2.60430992e-01
3.37304831e-01 3.79727781e-02 2.86874086e-01 -2.62252510e-01
-7.21893549e-01 -4.86113191e-01 6.09727144e-01 6.03908718e-01
2.97630548e-01 -1.58117022e-02 -2.50817001e-01 7.79874504e-01
1.95160210e-01 -3.22185189e-01 3.64792079e-01 1.11171317e+00
-3.71972412e-01 -1.02402508e+00 -2.70305008e-01 3.15973982e-02
-7.50524163e-01 -4.49813306e-01 -4.82630461e-01 7.01064646e-01
1.22586131e-01 1.08142865e+00 1.04662776e-01 -2.00761408e-01
4.00484115e-01 1.70320898e-01 8.91450122e-02 -5.73817194e-01
-5.04941702e-01 1.99480966e-01 3.81281197e-01 -4.00764436e-01
-6.62499070e-01 -5.47035396e-01 -9.26984429e-01 2.84544617e-01
-2.64514506e-01 5.60802743e-02 1.31238222e+00 7.84466147e-01
6.00131691e-01 2.56167233e-01 4.80072021e-01 -6.85567439e-01
-1.47634313e-01 -8.75930965e-01 -1.62490487e-01 9.44599688e-01
2.37073213e-01 -3.32924426e-01 -1.46100536e-01 6.23600364e-01] | [10.976485252380371, 1.1512154340744019] |
ff43cf0f-3512-4ea7-9a01-b4394b19f488 | piks-a-technique-to-identify-actionable | 2304.02208 | null | https://arxiv.org/abs/2304.02208v1 | https://arxiv.org/pdf/2304.02208v1.pdf | PIKS: A Technique to Identify Actionable Trends for Policy-Makers Through Open Healthcare Data | With calls for increasing transparency, governments are releasing greater amounts of data in multiple domains including finance, education and healthcare. The efficient exploratory analysis of healthcare data constitutes a significant challenge. Key concerns in public health include the quick identification and analysis of trends, and the detection of outliers. This allows policies to be rapidly adapted to changing circumstances. We present an efficient outlier detection technique, termed PIKS (Pruned iterative-k means searchlight), which combines an iterative k-means algorithm with a pruned searchlight based scan. We apply this technique to identify outliers in two publicly available healthcare datasets from the New York Statewide Planning and Research Cooperative System, and California's Office of Statewide Health Planning and Development. We provide a comparison of our technique with three other existing outlier detection techniques, consisting of auto-encoders, isolation forests and feature bagging. We identified outliers in conditions including suicide rates, immunity disorders, social admissions, cardiomyopathies, and pregnancy in the third trimester. We demonstrate that the PIKS technique produces results consistent with other techniques such as the auto-encoder. However, the auto-encoder needs to be trained, which requires several parameters to be tuned. In comparison, the PIKS technique has far fewer parameters to tune. This makes it advantageous for fast, "out-of-the-box" data exploration. The PIKS technique is scalable and can readily ingest new datasets. Hence, it can provide valuable, up-to-date insights to citizens, patients and policy-makers. We have made our code open source, and with the availability of open data, other researchers can easily reproduce and extend our work. This will help promote a deeper understanding of healthcare policies and public health issues. | ['Hang Peng', 'Soumyabrata Dey', 'Subrata Garai', 'A. Ravishankar Rao'] | 2023-04-05 | null | null | null | null | ['outlier-detection'] | ['methodology'] | [-1.25840560e-01 -5.41261137e-02 -3.47833097e-01 -6.32979929e-01
-8.94408226e-01 -1.76210463e-01 1.55368581e-01 7.87587464e-01
-3.38580012e-01 6.26507342e-01 6.87373221e-01 -8.22325706e-01
-2.77739853e-01 -7.78037429e-01 -6.38634086e-01 -5.43379009e-01
-1.83600426e-01 4.92930084e-01 -2.37120315e-01 1.19930945e-01
1.74677417e-01 5.08972943e-01 -1.11765409e+00 3.83785814e-01
1.20949888e+00 6.44710004e-01 -2.58849680e-01 4.75222707e-01
-6.41766889e-03 6.98786557e-01 -3.22365373e-01 -4.18368936e-01
1.33855820e-01 -4.43573624e-01 -4.94362056e-01 -3.37448657e-01
9.94021669e-02 -5.58303595e-01 2.42524538e-02 7.16194749e-01
6.58891797e-01 -1.20705277e-01 3.31081003e-01 -1.14401019e+00
-6.66369617e-01 4.93197381e-01 -5.16389012e-01 3.41985792e-01
2.80760139e-01 3.54398280e-01 6.68233514e-01 -7.76523829e-01
5.82548261e-01 9.32684600e-01 1.12845027e+00 4.06972080e-01
-1.33519602e+00 -9.50484872e-01 8.63767192e-02 -3.55183557e-02
-1.10378695e+00 -6.15329623e-01 3.48595887e-01 -7.11158156e-01
1.05502307e+00 3.46107692e-01 8.21403086e-01 9.36576247e-01
3.52936327e-01 4.90213901e-01 8.00466776e-01 -1.69929460e-01
5.10831118e-01 -5.84609322e-02 -7.54626244e-02 7.50890017e-01
7.02466190e-01 1.38145268e-01 -5.07278800e-01 -1.03108609e+00
5.17744780e-01 7.29515493e-01 -1.15955779e-02 -1.52552962e-01
-1.00215399e+00 9.81034577e-01 -2.51671094e-02 1.19534381e-01
-6.27804399e-01 6.09580278e-02 4.08494443e-01 2.16679350e-01
6.34868920e-01 3.84854287e-01 -6.28260195e-01 -4.91016775e-01
-1.12284422e+00 2.65101105e-01 7.46126711e-01 5.06610334e-01
5.17960906e-01 -7.39817247e-02 -1.03252292e-01 5.09379566e-01
4.23853546e-01 4.57967311e-01 5.25847495e-01 -9.77625191e-01
3.65465701e-01 8.95546734e-01 2.84690678e-01 -1.01149869e+00
-6.53516114e-01 -2.15432048e-01 -8.58452260e-01 4.29261625e-02
3.12621623e-01 -3.30749124e-01 -1.00294149e+00 1.49909258e+00
4.68428850e-01 1.44005045e-01 1.16439588e-01 4.35914874e-01
6.23577774e-01 2.31794775e-01 2.00013705e-02 -2.93987751e-01
1.30093241e+00 -4.05749440e-01 -7.54453897e-01 1.62870184e-01
7.31782556e-01 -4.87511218e-01 9.91212010e-01 4.10077423e-01
-9.25742209e-01 1.47551671e-01 -4.05219316e-01 2.73065865e-01
-4.06324267e-01 -2.42212564e-01 7.47190654e-01 6.74689889e-01
-7.89894342e-01 5.70102930e-01 -1.50582957e+00 -5.68731189e-01
8.88162017e-01 2.67296463e-01 -3.91433686e-01 -1.22437820e-01
-8.99013281e-01 5.87658823e-01 -1.58161931e-02 -1.59036070e-01
-5.94294488e-01 -7.72938073e-01 -9.57773507e-01 2.31693000e-01
2.48710260e-01 -7.04663396e-01 1.18635941e+00 -5.02933741e-01
-8.65951538e-01 4.55219060e-01 -4.49707806e-01 -3.52660775e-01
4.02835459e-01 -2.76873261e-01 -6.06499970e-01 -2.06099823e-01
3.81298631e-01 -5.00425622e-02 2.73593366e-01 -5.48334956e-01
-4.69781429e-01 -5.23793459e-01 -8.12318921e-01 -1.69271007e-01
-2.30462655e-01 4.97646451e-01 -8.19086060e-02 -5.77964485e-01
1.85556456e-01 -6.57261133e-01 -7.73075700e-01 -3.24250072e-01
-4.13091451e-01 1.45596027e-01 4.07342017e-01 -7.58730292e-01
1.57917202e+00 -2.06022191e+00 -5.73131919e-01 4.42123443e-01
2.88560271e-01 1.25534907e-01 3.46458405e-01 6.09108925e-01
-1.37733325e-01 3.81636709e-01 -2.16261104e-01 -9.04018506e-02
-3.15928310e-01 8.73027965e-02 -4.15868983e-02 5.73029637e-01
2.56683290e-01 9.01171267e-01 -9.60183620e-01 -2.63783157e-01
-3.43547948e-02 3.11368763e-01 -9.92827833e-01 1.38872832e-01
2.27311581e-01 5.07817686e-01 -3.20641488e-01 1.13030565e+00
3.11995804e-01 -5.39772630e-01 1.77339520e-02 4.12590474e-01
-3.65318954e-01 2.55324602e-01 -1.04505444e+00 1.17537212e+00
6.17233105e-02 2.54745454e-01 -2.55201012e-01 -7.27664232e-01
8.72768641e-01 2.35623300e-01 7.35866129e-01 -5.24267614e-01
-1.28927454e-01 3.91500860e-01 4.83770743e-02 -7.77254999e-01
1.94816917e-01 -8.36196095e-02 -7.03548118e-02 6.83087170e-01
-4.43391472e-01 3.93488437e-01 4.11633626e-02 1.88440263e-01
1.50292587e+00 -2.74675757e-01 7.03713715e-01 -9.33061074e-03
-1.50552392e-01 1.54324800e-01 9.67920601e-01 7.88728237e-01
-2.43544787e-01 5.00948489e-01 7.66477764e-01 -9.59435463e-01
-8.86621058e-01 -9.93642271e-01 -4.30690974e-01 8.67621422e-01
-8.21333170e-01 -5.84402680e-01 -3.50330204e-01 -6.77414715e-01
5.15796781e-01 6.04228199e-01 -7.76329458e-01 -9.29068774e-02
-4.99311209e-01 -1.08278167e+00 6.20333195e-01 6.11712635e-01
1.18170790e-01 -1.02883685e+00 -9.47060883e-01 3.88766199e-01
-1.15658730e-01 -5.27252913e-01 -3.71984363e-01 2.32190669e-01
-9.52839017e-01 -1.31440127e+00 -5.09608924e-01 -2.13500053e-01
7.65061200e-01 -3.76964360e-01 9.20011699e-01 -1.74951144e-02
-6.50283873e-01 1.91782176e-01 -1.76200598e-01 -8.76173675e-01
-4.05068278e-01 -1.30775094e-01 3.37398052e-01 -2.95810670e-01
8.07890117e-01 -3.20725232e-01 -7.74337351e-01 -8.08379799e-03
-8.11792910e-01 -3.29386532e-01 5.04756987e-01 9.51179743e-01
5.76000154e-01 -2.05819914e-03 7.89585710e-01 -1.37784457e+00
5.75462401e-01 -9.44497287e-01 -5.97205400e-01 2.20541470e-02
-1.09228694e+00 -9.88110974e-02 3.28051478e-01 -2.31657904e-02
-7.78869271e-01 1.63509727e-01 -2.02718601e-01 -1.92373917e-01
-1.99616507e-01 6.37372136e-01 2.22140342e-01 3.09708685e-01
7.78358698e-01 -6.06995113e-02 2.11178064e-01 -6.42413199e-01
-5.81529401e-02 9.98179734e-01 3.55125785e-01 -1.50658220e-01
3.85560483e-01 5.25258303e-01 -3.87879997e-01 -5.80566168e-01
-4.18230176e-01 -5.81071854e-01 -1.86613023e-01 1.33887142e-01
6.26123726e-01 -9.35205042e-01 -6.81327105e-01 3.08329880e-01
-6.62237644e-01 -2.99264938e-01 -4.45592999e-01 5.78775167e-01
-3.54619861e-01 1.67910129e-01 -4.84376758e-01 -9.28125918e-01
-4.76706594e-01 -1.04548442e+00 8.67566764e-01 1.42958537e-01
-6.11494482e-01 -7.55539298e-01 4.71756935e-01 1.32417291e-01
5.37526131e-01 6.63746238e-01 1.01661766e+00 -1.08276141e+00
-2.66670614e-01 -1.73245922e-01 -2.31439434e-02 3.50860618e-02
4.55462098e-01 3.00430834e-01 -7.23795295e-01 -3.88164133e-01
-1.70281336e-01 2.10818406e-02 8.39257538e-01 6.76611781e-01
1.30195999e+00 -8.96223247e-01 -5.27906895e-01 8.41800451e-01
1.14540553e+00 4.30263132e-01 4.39843804e-01 4.29837942e-01
4.66402471e-01 4.00608093e-01 4.81529683e-01 9.59375203e-01
7.23386765e-01 2.73438334e-01 2.41522759e-01 -3.86409640e-01
5.41194081e-01 -1.71125412e-01 1.01993345e-01 5.10950387e-01
-9.16563123e-02 1.55103311e-01 -1.50969493e+00 6.63587213e-01
-1.96454310e+00 -1.13030994e+00 6.13362528e-02 2.32225800e+00
9.11157966e-01 -1.00856692e-01 2.84021437e-01 -9.46937650e-02
3.12958330e-01 -2.63322651e-01 -8.11837912e-01 -7.64087379e-01
2.96802014e-01 1.50976315e-01 5.49889565e-01 1.38901159e-01
-8.34022760e-01 5.50827503e-01 7.17928219e+00 -2.16550529e-02
-9.29780483e-01 2.29432397e-02 1.05444181e+00 -3.15279871e-01
-4.85480398e-01 -1.06308684e-01 -5.64057946e-01 7.94077814e-01
1.22680748e+00 -1.85824081e-01 2.23846957e-01 9.60997045e-01
5.71365833e-01 -1.08252056e-01 -9.02071178e-01 6.73541605e-01
-2.16749042e-01 -1.50773668e+00 -5.08542418e-01 2.58501500e-01
7.33393669e-01 3.44622642e-01 4.09272388e-02 2.28918456e-02
7.90709019e-01 -1.22175968e+00 1.80115718e-02 7.79713511e-01
8.88099313e-01 -9.91330206e-01 8.05599511e-01 1.44810706e-01
-6.46355450e-01 -4.99951422e-01 -2.08362117e-01 -7.36594945e-02
4.83573265e-02 1.07567489e+00 -1.14074564e+00 2.75878131e-01
1.22990537e+00 4.59144145e-01 -3.13719571e-01 1.19178724e+00
2.95847952e-01 8.67388308e-01 -2.38698632e-01 2.67070144e-01
-6.07192442e-02 9.26449448e-02 2.43047401e-01 1.17321658e+00
5.73258817e-01 1.31922126e-01 1.23850562e-01 6.53305233e-01
1.59761071e-01 3.58383179e-01 -7.30553985e-01 -2.56550252e-01
6.07477009e-01 8.11123729e-01 -4.95581865e-01 -3.55201811e-01
-6.24112904e-01 4.36630338e-01 2.80922413e-01 3.55739653e-01
-4.81060207e-01 -4.26021695e-01 9.37955678e-01 3.29124063e-01
2.52621263e-01 1.49237320e-01 -3.64297748e-01 -1.05235207e+00
-2.51609087e-01 -1.18243420e+00 8.01662207e-01 -1.76601306e-01
-1.16734231e+00 2.06633300e-01 -7.83375949e-02 -9.19346929e-01
-5.92585206e-01 -2.41478756e-01 -5.41897118e-01 7.62373626e-01
-1.26156449e+00 -6.52115107e-01 -2.13028789e-01 4.61021870e-01
7.55410269e-02 -2.59193599e-01 9.82923925e-01 1.90000474e-01
-8.52080882e-01 5.72064996e-01 5.40525019e-01 1.40417114e-01
7.77713299e-01 -1.13993108e+00 5.71025908e-01 7.84961045e-01
-2.10519925e-01 9.54715312e-01 4.60557044e-01 -1.14652967e+00
-9.31840479e-01 -1.24184966e+00 1.19516909e+00 -4.76137012e-01
3.76927137e-01 -1.13716573e-01 -8.95672143e-01 1.06761384e+00
-1.84108227e-01 -7.83272088e-02 1.41412532e+00 4.10338044e-01
-1.92999676e-01 -1.62344769e-01 -1.44393182e+00 3.54553759e-01
6.43912852e-01 -1.66574612e-01 -3.35394412e-01 4.69319493e-01
5.48955083e-01 -4.98860687e-01 -9.92154002e-01 3.59367013e-01
7.85443723e-01 -1.20992804e+00 6.61794662e-01 -1.08127379e+00
3.63858461e-01 5.73542416e-02 -1.97124965e-02 -1.17613637e+00
-5.24162829e-01 -7.96780109e-01 -2.90925205e-01 8.81044924e-01
4.03593034e-01 -1.08660042e+00 7.72383749e-01 1.10813928e+00
-1.14941254e-01 -1.16382718e+00 -7.77481616e-01 -5.53680480e-01
-1.15347050e-01 -3.88856381e-01 1.23327947e+00 1.17300475e+00
2.14384064e-01 -4.54340696e-01 -3.01383615e-01 2.92130977e-01
4.31116998e-01 9.72595289e-02 7.98523664e-01 -1.30520213e+00
-2.73590267e-01 -2.09174305e-01 -3.51573974e-01 -1.36414051e-01
-4.76108760e-01 -5.48935890e-01 -2.90544182e-01 -1.56441140e+00
5.10152638e-01 -5.71327150e-01 -5.88769317e-01 1.05943584e+00
-3.26157898e-01 -3.04600727e-02 -3.80856663e-01 2.35023931e-01
-3.23458493e-01 7.81432390e-02 6.47148013e-01 2.09190056e-01
-5.80188572e-01 1.88239157e-01 -1.23420501e+00 7.66193271e-01
9.53282773e-01 -7.91407287e-01 -1.25090415e-02 -1.58608764e-01
3.66108805e-01 -7.55702332e-02 1.29825220e-01 -7.16969192e-01
2.38408476e-01 -5.68892241e-01 7.41981745e-01 -6.45668209e-01
-2.51442164e-01 -6.58768117e-01 3.90218347e-01 9.51952338e-01
-1.86137453e-01 6.23442709e-01 1.86469153e-01 4.47459698e-01
8.10294673e-02 1.10539861e-01 4.84507233e-01 -3.28401357e-01
-9.32580829e-02 2.36480013e-01 -5.27453780e-01 1.33589849e-01
1.08641982e+00 -1.29693672e-01 -5.24687827e-01 -5.33912480e-01
-6.21816337e-01 3.76164317e-01 6.53694391e-01 9.42036957e-02
6.41519189e-01 -1.17212141e+00 -8.53140056e-01 6.59180939e-01
1.41992375e-01 2.47908890e-01 1.91706136e-01 1.15968657e+00
-4.59165722e-01 4.32832301e-01 4.11413386e-02 -4.18545067e-01
-1.17926800e+00 5.01918733e-01 9.95232239e-02 -1.77989319e-01
-7.16013193e-01 5.63141346e-01 -1.77798823e-01 -5.12261033e-01
2.38676012e-01 -5.90171337e-01 4.21336815e-02 9.63003859e-02
8.05007994e-01 7.40553737e-01 6.13236912e-02 -3.43291312e-01
-6.83997393e-01 -6.28122538e-02 -2.04669297e-01 1.94187254e-01
1.65373898e+00 8.67810026e-02 -1.02721564e-01 5.17575383e-01
9.20237601e-01 3.27481031e-01 -8.72363031e-01 1.58794466e-02
8.35715234e-02 -6.14583313e-01 -2.24608988e-01 -8.20281088e-01
-8.99329782e-01 4.85853165e-01 5.22524774e-01 2.81412989e-01
1.25114834e+00 -6.64380416e-02 6.83147490e-01 2.58258373e-01
2.07037386e-02 -9.05537486e-01 -8.13963637e-02 1.26324981e-01
4.57756817e-01 -1.29464781e+00 2.05494240e-01 1.69253975e-01
-7.13236570e-01 8.03409815e-01 3.35791767e-01 3.45425040e-01
5.49180925e-01 5.18683255e-01 3.24878156e-01 -4.09441769e-01
-9.79119420e-01 1.26450375e-01 4.99175005e-02 4.47379023e-01
5.65758646e-01 2.07798794e-01 -9.53627378e-02 7.51073897e-01
-2.34083280e-01 1.69057533e-01 5.97825766e-01 8.06848109e-01
-4.21530724e-01 -1.09299636e+00 -4.56678182e-01 1.36344099e+00
-8.56871128e-01 -4.00640815e-01 -1.75917149e-01 6.03268921e-01
9.90125835e-02 6.55895591e-01 2.43044451e-01 -8.44702050e-02
3.29532444e-01 4.48199660e-01 -2.75269181e-01 -6.77398086e-01
-5.51499844e-01 4.67153080e-02 1.14678316e-01 -8.86254132e-01
1.20230876e-02 -1.26215160e+00 -1.26368272e+00 -3.74315560e-01
-2.24367499e-01 2.32611641e-01 6.89006448e-01 5.46289265e-01
9.32280064e-01 3.96632135e-01 5.67633510e-01 -7.89774135e-02
-5.06775022e-01 -6.49754345e-01 -4.04747099e-01 2.91016638e-01
6.29344821e-01 -3.75510901e-01 -2.48755097e-01 -2.87960619e-01] | [7.861851692199707, 5.9829301834106445] |
bb45b692-ec37-4dde-a812-efc712879222 | learning-pixel-adaptive-weights-for-portrait | 2112.03536 | null | https://arxiv.org/abs/2112.03536v1 | https://arxiv.org/pdf/2112.03536v1.pdf | Learning Pixel-Adaptive Weights for Portrait Photo Retouching | Portrait photo retouching is a photo retouching task that emphasizes human-region priority and group-level consistency. The lookup table-based method achieves promising retouching performance by learning image-adaptive weights to combine 3-dimensional lookup tables (3D LUTs) and conducting pixel-to-pixel color transformation. However, this paradigm ignores the local context cues and applies the same transformation to portrait pixels and background pixels when they exhibit the same raw RGB values. In contrast, an expert usually conducts different operations to adjust the color temperatures and tones of portrait regions and background regions. This inspires us to model local context cues to improve the retouching quality explicitly. Firstly, we consider an image patch and predict pixel-adaptive lookup table weights to precisely retouch the center pixel. Secondly, as neighboring pixels exhibit different affinities to the center pixel, we estimate a local attention mask to modulate the influence of neighboring pixels. Thirdly, the quality of the local attention mask can be further improved by applying supervision, which is based on the affinity map calculated by the groundtruth portrait mask. As for group-level consistency, we propose to directly constrain the variance of mean color components in the Lab space. Extensive experiments on PPR10K dataset verify the effectiveness of our method, e.g. on high-resolution photos, the PSNR metric receives over 0.5 gains while the group-level consistency metric obtains at least 2.1 decreases. | ['Yongqiang Zhao', 'Dawei Yan', 'Chengzhe Lu', 'Binglu Wang'] | 2021-12-07 | null | null | null | null | ['photo-retouching'] | ['computer-vision'] | [ 3.93218398e-01 -2.07116634e-01 -2.39541888e-01 -1.92075819e-01
-5.16566992e-01 -5.34293115e-01 1.87816739e-01 -1.33116513e-01
-2.93551266e-01 6.56822383e-01 -3.65870558e-02 -1.64435955e-03
1.80646658e-01 -6.80054605e-01 -9.43870842e-01 -1.03788435e+00
6.10153615e-01 -3.57823074e-01 3.96784574e-01 -2.08670333e-01
3.22801709e-01 4.09122169e-01 -1.39388740e+00 2.93287694e-01
8.97138834e-01 1.25247478e+00 1.35212258e-01 5.32219470e-01
-8.87420683e-05 3.95220160e-01 -5.37907064e-01 -3.84449214e-01
5.64181209e-01 -5.47342539e-01 -2.96950400e-01 3.50688279e-01
9.41619992e-01 -4.06601012e-01 -3.95362109e-01 1.30743575e+00
4.45213854e-01 -5.84752373e-02 2.69356251e-01 -1.03105104e+00
-1.24704051e+00 4.13752258e-01 -1.17591953e+00 3.03594589e-01
2.03393966e-01 3.87475967e-01 9.11683440e-01 -7.60491729e-01
4.81162786e-01 1.07444155e+00 4.13596392e-01 4.61345524e-01
-1.60916948e+00 -8.96605730e-01 4.25236732e-01 3.71116132e-01
-1.73852301e+00 -2.66315609e-01 9.05941904e-01 -4.85101081e-02
4.64561075e-01 6.79808080e-01 7.50286341e-01 7.81377912e-01
4.58792806e-01 6.15642369e-01 1.30910277e+00 -2.03559890e-01
2.23993734e-01 3.75575900e-01 -2.19516307e-01 4.78278428e-01
2.35269144e-01 -1.66202992e-01 -7.00869560e-01 1.94321290e-01
1.18952811e+00 1.09941982e-01 -6.04381621e-01 -1.45988166e-01
-1.29778564e+00 6.29498884e-02 1.01291966e+00 2.83310264e-01
-9.18694288e-02 3.78996760e-01 -8.01156685e-02 1.55733168e-01
7.13004544e-02 4.09961432e-01 -2.75115877e-01 3.35620642e-01
-8.11252177e-01 -3.07356536e-01 1.00066066e-01 9.47145998e-01
1.11928976e+00 -9.40030217e-02 -5.27214646e-01 7.95874953e-01
-2.84959353e-03 6.96521044e-01 2.86781728e-01 -8.31531346e-01
2.59232670e-01 6.42205358e-01 2.73927808e-01 -1.33418036e+00
-5.50783351e-02 -1.97477356e-01 -1.07393336e+00 3.50437135e-01
5.00012100e-01 1.48453295e-01 -1.10799551e+00 1.50246096e+00
4.96776909e-01 2.67892957e-01 -2.98607498e-01 1.38784719e+00
3.25077385e-01 6.93099439e-01 2.08380837e-02 -1.48434296e-01
1.50883830e+00 -9.45026875e-01 -8.82752180e-01 -2.51236171e-01
-3.65482382e-02 -7.83977747e-01 1.53031576e+00 2.49068707e-01
-9.95111346e-01 -8.36403310e-01 -1.31640184e+00 -2.47083858e-01
-2.94264972e-01 2.00953588e-01 2.50568300e-01 6.61709309e-01
-1.01743388e+00 4.12527293e-01 -5.35332620e-01 -2.07000941e-01
2.22564474e-01 3.01277786e-01 -4.19985689e-02 -7.30888098e-02
-1.05255651e+00 4.92310315e-01 8.90287012e-02 2.89773285e-01
-3.78373444e-01 -8.59494269e-01 -3.89275163e-01 9.05722082e-02
3.91785294e-01 -4.69045013e-01 6.40214503e-01 -1.39144480e+00
-1.72617197e+00 7.65310049e-01 3.62599492e-02 -6.62024990e-02
4.71008092e-01 1.71221271e-01 -7.07132757e-01 8.43051970e-02
-1.46583632e-01 9.12508428e-01 1.23479855e+00 -1.42828262e+00
-7.39461541e-01 -2.91160971e-01 -7.62547180e-02 2.38395005e-01
-6.22506440e-01 -5.73024273e-01 -1.03614283e+00 -7.11083472e-01
3.17341179e-01 -8.40627313e-01 -1.25604913e-01 5.38878083e-01
-5.13494492e-01 4.75089401e-01 6.49335444e-01 -5.13971567e-01
1.41510260e+00 -2.53134274e+00 -2.46543482e-01 3.63369852e-01
2.06709981e-01 -3.17750834e-02 -2.10252792e-01 -1.73047453e-01
8.15922860e-03 1.30205974e-01 -1.26603112e-01 2.03301460e-01
-1.64191827e-01 6.19836263e-02 -2.12531313e-01 4.53542262e-01
2.92225361e-01 1.01439142e+00 -7.24452138e-01 -3.81015241e-01
2.32018352e-01 6.21800542e-01 -4.41829056e-01 -2.08800435e-01
-6.23859875e-02 2.61483490e-01 -3.60092372e-01 7.14536130e-01
1.16527116e+00 -3.64908189e-01 1.87147200e-01 -8.19126070e-01
-1.16511352e-01 -1.09155357e-01 -1.19373727e+00 1.44902170e+00
-2.82548547e-01 5.45491219e-01 -8.22990313e-02 -1.69909135e-01
1.02865994e+00 -2.39851370e-01 3.18601400e-01 -1.26204443e+00
1.26185656e-01 -5.57110086e-02 -2.19859809e-01 -6.94400668e-02
8.02405655e-01 1.62370116e-01 -7.95511082e-02 2.56830245e-01
-5.73954701e-01 4.90472056e-02 -4.24741685e-01 -2.43272651e-02
7.74037600e-01 -1.43575937e-01 2.67527495e-02 -4.12859917e-01
5.34391820e-01 -1.75878316e-01 6.07346475e-01 4.77995932e-01
-2.92657167e-01 8.25900614e-01 3.06393772e-01 -5.38273156e-01
-1.06318581e+00 -1.05914116e+00 -2.62462258e-01 9.30022895e-01
9.93777335e-01 -7.11698011e-02 -7.74873555e-01 -3.71239781e-01
-2.01837458e-02 5.36134541e-01 -9.51515496e-01 -2.45753244e-01
-4.27120388e-01 -7.09798694e-01 2.03230783e-01 3.99120986e-01
8.94986093e-01 -5.56654215e-01 -5.13930082e-01 3.13772149e-02
1.33268302e-02 -9.70699072e-01 -1.10744238e+00 -6.79309219e-02
-6.14694834e-01 -7.50987947e-01 -7.68348396e-01 -6.31153643e-01
1.03470206e+00 6.24718308e-01 7.88747072e-01 1.89567178e-01
-2.68280894e-01 2.70748794e-01 -1.37790397e-01 2.72765011e-01
2.52729673e-02 -1.55153545e-02 -1.08911373e-01 4.33968216e-01
3.71240415e-02 -5.23028493e-01 -1.08482742e+00 8.26482296e-01
-1.04517353e+00 3.47687572e-01 7.11064935e-01 7.81306326e-01
1.09163773e+00 1.98861107e-01 1.50493458e-01 -7.06974864e-01
2.86244094e-01 1.62882835e-01 -8.01062226e-01 5.87944508e-01
-6.59890831e-01 -2.78946143e-02 6.04841053e-01 -8.69250596e-01
-9.97426271e-01 9.63088423e-02 4.43072796e-01 -6.87205672e-01
1.85487017e-01 -2.58222163e-01 -4.61780518e-01 -3.70115638e-01
5.67974925e-01 2.14920998e-01 -4.64261711e-01 -5.78830503e-02
6.13300979e-01 3.21211100e-01 7.58762300e-01 -3.26951385e-01
9.86717403e-01 6.39656603e-01 -2.84046948e-01 -3.86627406e-01
-4.10235256e-01 4.25149053e-02 -4.47466344e-01 -3.94497007e-01
9.41743255e-01 -8.88693810e-01 -9.51550186e-01 3.70353401e-01
-6.86572909e-01 -5.75703919e-01 -3.29294831e-01 2.72869989e-02
-5.40896580e-02 1.35481358e-01 -5.53007722e-01 -5.61477721e-01
-2.40447789e-01 -1.06649792e+00 1.10195172e+00 7.02709079e-01
1.21224687e-01 -5.68316340e-01 -3.98185521e-01 1.73799276e-01
4.14024323e-01 1.38895646e-01 9.23861742e-01 3.13751280e-01
-1.04856932e+00 4.56589507e-03 -6.41896844e-01 5.74847963e-03
3.70316207e-01 1.17692247e-01 -1.04419971e+00 -1.15999393e-01
-3.18910748e-01 2.14984000e-01 6.92433894e-01 2.54685193e-01
1.28036976e+00 -2.58085191e-01 -2.21254423e-01 6.36821866e-01
1.51302063e+00 2.27221444e-01 1.03987050e+00 3.40403944e-01
1.02475381e+00 1.89223692e-01 6.01420760e-01 3.60446602e-01
3.78506839e-01 7.53196895e-01 2.19820336e-01 -4.59324867e-01
-4.13530648e-01 -5.12428641e-01 3.45633715e-01 5.83102167e-01
2.70507574e-01 -2.26786599e-01 -4.38476562e-01 2.92954057e-01
-1.66127884e+00 -6.12082720e-01 -6.42950982e-02 2.40681553e+00
1.05265677e+00 8.31620321e-02 -2.32993573e-01 -9.67427045e-02
1.06393969e+00 1.82492092e-01 -9.97147262e-01 -1.52403519e-01
-4.46639001e-01 -1.86250404e-01 1.00043285e+00 4.70712095e-01
-6.02729082e-01 9.82229292e-01 5.87324905e+00 1.08507454e+00
-1.48114502e+00 -1.94242179e-01 1.32186592e+00 -1.58655822e-01
-7.66076803e-01 -8.66274908e-02 -7.50211775e-01 6.93300784e-01
2.43257329e-01 6.80480152e-02 9.21194017e-01 4.51626807e-01
1.41429484e-01 -4.23045456e-01 -9.11417842e-01 1.18255401e+00
1.02338165e-01 -1.18349767e+00 1.32932097e-01 3.86881754e-02
1.17122757e+00 -5.67802370e-01 6.36997521e-01 -1.74205422e-01
-8.50692466e-02 -6.88759267e-01 8.43947887e-01 7.23427117e-01
9.93002295e-01 -6.29382133e-01 2.16132239e-01 -3.52990240e-01
-1.24805415e+00 6.74338043e-02 -5.84253371e-01 3.71890187e-01
-1.12486079e-01 6.14373028e-01 -4.85090137e-01 2.23552361e-01
9.99749422e-01 5.00590503e-01 -8.05790007e-01 7.82981634e-01
-2.90061861e-01 2.63810456e-01 -2.32689962e-01 8.73868167e-02
4.14503999e-02 -4.33978379e-01 1.51056245e-01 7.19013453e-01
2.56107330e-01 2.92314231e-01 -2.27516502e-01 1.10028434e+00
-3.88711601e-01 7.48330578e-02 1.37365341e-01 2.52352029e-01
5.39812446e-01 1.40310764e+00 -8.50103199e-01 -3.16313654e-01
-1.32173255e-01 1.57110560e+00 -3.43085043e-02 7.17481971e-01
-1.11896396e+00 -3.14893663e-01 8.75903249e-01 3.19955260e-01
5.37094295e-01 -3.18213739e-03 -7.15986311e-01 -1.00253105e+00
9.13797095e-02 -7.34079182e-01 1.32360056e-01 -1.14608133e+00
-1.30098343e+00 5.54191828e-01 -4.60673928e-01 -1.36820722e+00
6.50703609e-01 -3.38860542e-01 -4.96516973e-01 7.82812655e-01
-1.61920547e+00 -9.40845251e-01 -6.78310990e-01 7.07261384e-01
2.06210330e-01 4.57283139e-01 2.47292727e-01 3.83735895e-01
-9.38376069e-01 1.05125487e+00 9.50414240e-02 -1.42532988e-02
1.16822135e+00 -9.49363768e-01 3.81406873e-01 8.93216252e-01
3.92631739e-02 3.96606535e-01 5.45780420e-01 -5.06398320e-01
-1.64822352e+00 -1.04744291e+00 3.09587449e-01 -3.75884712e-01
6.23533368e-01 -3.55544627e-01 -1.17490387e+00 2.13031888e-01
2.82616258e-01 2.41262779e-01 3.67716402e-01 -4.07920212e-01
-5.59908092e-01 -8.43035400e-01 -1.24466932e+00 1.01770103e+00
9.18959200e-01 -6.21796489e-01 2.06652954e-01 -3.54480580e-03
7.17859149e-01 -5.12113333e-01 -8.30121934e-01 3.61500531e-02
8.12105775e-01 -8.48762333e-01 8.97388041e-01 8.88356268e-02
2.15999469e-01 -9.44030464e-01 -9.00156498e-02 -9.88714218e-01
-5.91901302e-01 -6.34278774e-01 3.04159850e-01 1.42019737e+00
3.42795104e-01 -6.44682825e-01 6.83146834e-01 1.20356882e+00
2.79243141e-01 -6.05668902e-01 -9.37805057e-01 -4.26803023e-01
-3.77740175e-01 -4.24766401e-03 9.01586533e-01 8.93118382e-01
-2.55964816e-01 1.38567150e-01 -5.65445125e-01 5.22188365e-01
4.51808721e-01 2.46946990e-01 6.61579370e-01 -5.14011085e-01
-2.30892792e-01 -6.32663667e-01 -2.97173232e-01 -1.18930948e+00
-3.26101512e-01 -3.94088119e-01 1.21075846e-01 -1.04935694e+00
1.18673466e-01 -5.96873462e-01 -7.07553208e-01 6.22495651e-01
-6.31704211e-01 9.02943432e-01 2.99312115e-01 2.22110555e-01
-5.70628643e-01 4.89611477e-01 1.72964180e+00 -4.78457212e-01
-4.89107996e-01 -4.82584476e-01 -8.87244880e-01 1.76047429e-01
7.07344413e-01 -2.45743483e-01 -3.33292961e-01 -5.93660533e-01
3.03980649e-01 -3.04201484e-01 3.73982489e-01 -1.04670775e+00
4.19753969e-01 -4.09720004e-01 9.08270836e-01 -3.29019338e-01
2.36563295e-01 -1.09664750e+00 4.46785301e-01 1.54244706e-01
-2.28338242e-01 -7.12425038e-02 3.43141019e-01 7.25998163e-01
1.77238867e-01 3.91224772e-01 1.00408530e+00 2.81309783e-01
-9.02519584e-01 2.47626886e-01 -2.15575948e-01 -3.02998304e-01
1.07503903e+00 -6.79890871e-01 -3.53041381e-01 -2.63562083e-01
-2.28250340e-01 2.55083777e-02 8.88831615e-01 2.94412136e-01
5.10061800e-01 -1.58958256e+00 -3.18922549e-01 5.20323098e-01
2.38457710e-01 -2.89035201e-01 6.86680198e-01 9.15785372e-01
-3.46302688e-01 -1.27954796e-01 -4.15773183e-01 -6.50923193e-01
-1.07795715e+00 7.24845946e-01 3.18454504e-01 2.67390072e-01
-4.80127126e-01 8.77881825e-01 4.58260804e-01 3.53699744e-01
-8.20200983e-03 -7.17593670e-01 1.60158083e-01 -9.23958421e-02
7.59470522e-01 1.76873356e-01 -5.49086258e-02 -5.15294969e-01
-3.61273825e-01 1.01471937e+00 -1.46130756e-01 4.79687117e-02
8.87671828e-01 -5.61980605e-01 1.01840068e-02 3.16169709e-01
1.03077435e+00 2.41158083e-01 -1.80778742e+00 -2.25972563e-01
-5.07088006e-01 -1.00571167e+00 2.68458962e-01 -9.65613246e-01
-1.52018237e+00 5.38038194e-01 9.20116484e-01 -5.45707066e-04
1.75823033e+00 -2.44838402e-01 8.36026609e-01 1.20992467e-01
2.91857868e-01 -1.26321983e+00 1.96029231e-01 -9.63693038e-02
7.54672289e-01 -9.62774217e-01 1.45489767e-01 -4.29785103e-01
-8.71630251e-01 9.57615614e-01 8.33647192e-01 -7.68672377e-02
2.16567695e-01 2.15452939e-01 2.72697628e-01 1.80217564e-01
-5.01960933e-01 -8.03693533e-02 5.12079120e-01 2.66511232e-01
1.60845350e-02 6.89601749e-02 9.94407982e-02 2.61760145e-01
1.27646834e-01 -3.01104516e-01 2.97217011e-01 3.76048237e-01
-2.53085285e-01 -8.06210160e-01 -6.94954932e-01 1.74811870e-01
6.45115748e-02 -1.61819816e-01 -3.72476846e-01 6.00570619e-01
2.52072513e-01 7.85094738e-01 2.47427911e-01 -6.98029876e-01
4.09933269e-01 -4.41879570e-01 4.36872214e-01 9.58411470e-02
-6.12225831e-01 3.86614680e-01 -5.20686388e-01 -6.44872785e-01
-3.13765347e-01 -1.69206217e-01 -1.18620312e+00 -6.31362379e-01
-3.15247148e-01 -2.21858710e-01 3.95179302e-01 3.47190291e-01
4.96605247e-01 6.83184326e-01 8.00755739e-01 -7.05040932e-01
2.45543078e-01 -4.13048655e-01 -7.48892903e-01 3.44045550e-01
2.82156706e-01 -2.98559010e-01 -2.44056255e-01 2.39685029e-01] | [11.331398963928223, -1.1540988683700562] |
6e8b0ade-27b0-4658-b29d-9cff6a414ff5 | a-novel-approach-to-vehicle-pose-estimation | 2107.09607 | null | https://arxiv.org/abs/2107.09607v1 | https://arxiv.org/pdf/2107.09607v1.pdf | A Novel Approach to Vehicle Pose Estimation using Automotive Radar | This paper presents a set of novel scan-matching techniques for vehicle pose estimation using automotive radar measurements. The proposed approach modifies the Normal Distributions Transform (NDT) -- a state-of-the-art scan-matching SLAM technique, widely used in lidar-based localization -- to account for particular aspects of radar environment perception. First, the polar NDT (PNDT) is introduced by solving the NDT problem in the polar coordinate system, natural for radar measurements. A better agreement between the measurement uncertainties and their representation in the scan-matching algorithm is achieved. Second, the extension of PNDT to take into account the Doppler measurements -- Doppler polar NDT (DPNDT) -- is proposed. Third, the SNR of detected targets is added to the optimization procedure to minimize the impact of RCS fluctuation. The improvement over the conventional NDT is demonstrated in numerical simulations and real data processing, showing the ability to decrease the localization error by a factor of 3 to 5, depending on the scenario, with a negligible increase in computational complexity. Finally, a DPNDT extension with the capability to compensate angular bias in array beam-forming is presented. Simulation results and real data processing show the possibility to correct it with the accuracy of $0.1^\circ$ almost in real-time. | ['Alexander Yarovoy', 'Nikita Petrov', 'Martijn Heller'] | 2021-07-20 | null | null | null | null | ['vehicle-pose-estimation'] | ['computer-vision'] | [ 2.67833173e-01 -5.47467209e-02 3.67669851e-01 -6.62621558e-01
-6.05221391e-01 -1.28193691e-01 6.93608880e-01 1.77442078e-02
-8.20763528e-01 9.73270833e-01 -4.06906188e-01 -3.72169018e-01
-8.66295159e-01 -1.04912460e+00 -5.34701407e-01 -7.82854140e-01
-1.99142590e-01 1.01536191e+00 1.84375286e-01 -3.17647010e-01
2.76451558e-01 1.25476074e+00 -1.64071929e+00 -5.84178448e-01
9.86574352e-01 9.37187493e-01 2.49259293e-01 1.86550662e-01
1.15479063e-03 -1.69988573e-01 -6.22933745e-01 -2.24636063e-01
4.02308136e-01 -1.11362122e-01 4.09011871e-01 -3.66808236e-01
4.29138303e-01 -5.57035245e-02 -9.10311714e-02 1.21014082e+00
4.58268195e-01 -6.19547926e-02 9.55811560e-01 -1.03503096e+00
2.90526360e-01 2.68246531e-01 -3.19825232e-01 -2.29005199e-02
1.16111241e-01 -4.70595621e-02 3.53424698e-01 -1.04280758e+00
8.11535239e-01 9.91709948e-01 4.43245769e-01 -5.74535467e-02
-1.13191688e+00 -7.52739668e-01 -7.52604306e-01 5.78754246e-01
-1.76335824e+00 -1.18590027e-01 7.75463343e-01 -5.60396552e-01
4.70772713e-01 3.65404874e-01 7.69163728e-01 4.59150612e-01
7.19559669e-01 -3.31791162e-01 1.38981926e+00 -4.02917981e-01
2.45989218e-01 1.98470697e-01 1.95239440e-01 1.52309865e-01
1.10412323e+00 6.49452925e-01 -1.77537814e-01 8.79019722e-02
4.94272649e-01 -3.63255352e-01 -1.82423443e-01 -8.36035430e-01
-9.25500393e-01 7.04483211e-01 2.27207512e-01 5.47013283e-01
-5.90528846e-01 -4.06597741e-02 1.40019596e-01 2.62689769e-01
3.67478102e-01 3.72848183e-01 -1.94195341e-02 -1.84059799e-01
-1.14281976e+00 4.89504308e-01 6.93822384e-01 7.86386132e-01
8.98699820e-01 5.99041104e-01 1.03131466e-01 9.92219076e-02
4.33986664e-01 1.38968253e+00 2.08380688e-02 -3.43572319e-01
2.57379204e-01 3.06988597e-01 2.46124923e-01 -8.64851236e-01
-6.19492769e-01 -1.01571631e+00 -6.14987671e-01 7.91832089e-01
4.28439379e-01 -1.80518508e-01 -6.62163854e-01 1.42796171e+00
5.28636038e-01 -3.14175412e-02 3.38953227e-01 7.48041987e-01
1.80315882e-01 4.61471438e-01 -5.15750706e-01 -6.05933130e-01
1.26811302e+00 3.87737416e-02 -9.95620012e-01 -1.81550398e-01
3.11456323e-01 -1.03553867e+00 1.38525084e-01 6.35815144e-01
-6.99488759e-01 -5.22771418e-01 -1.28159606e+00 8.97454679e-01
-2.31973499e-01 1.99062109e-01 -1.34323642e-01 9.29172695e-01
-6.48487210e-01 4.03234094e-01 -2.69119948e-01 -4.42988545e-01
-1.09411925e-01 5.40896021e-02 -3.96787018e-01 -1.73993364e-01
-1.04855549e+00 1.53728843e+00 1.34697616e-01 4.22538221e-01
-2.71346688e-01 -8.84912133e-01 -6.79268837e-01 -5.48056848e-02
1.70750752e-01 -2.94223279e-01 8.31059694e-01 -2.95850605e-01
-1.37920535e+00 3.62294108e-01 -4.75895584e-01 -7.91725159e-01
7.59926081e-01 -2.68063664e-01 -8.58366191e-01 1.65271103e-01
-2.72964090e-02 1.59262970e-01 8.82587016e-01 -1.20132363e+00
-2.24715412e-01 -4.13583010e-01 -6.64553761e-01 -3.48456651e-02
3.93527538e-01 -4.05154914e-01 1.89417720e-01 -2.76680499e-01
5.90479493e-01 -8.77901912e-01 -3.66783202e-01 -2.23065689e-01
8.30552354e-02 2.41464853e-01 9.22272861e-01 -3.00161898e-01
6.56572223e-01 -2.06004667e+00 -1.32034287e-01 8.66035223e-01
-1.01249509e-01 2.86952555e-01 1.06284209e-01 6.63847268e-01
-1.47319317e-01 -7.11854517e-01 -7.10273087e-02 -3.77935693e-02
-1.02874853e-01 2.31098041e-01 -2.00542957e-01 8.65402937e-01
-3.26505229e-02 4.84577715e-01 -3.67608130e-01 -1.63855284e-01
6.35436773e-01 5.70123911e-01 -1.57436486e-02 -2.27782637e-01
-2.15726588e-02 4.33066905e-01 -2.67379135e-01 2.73730278e-01
1.40208435e+00 7.74958789e-01 -1.16318762e-02 -3.22213233e-01
-7.38941073e-01 -1.09631188e-01 -1.58472872e+00 9.02183056e-01
-3.79006892e-01 9.17533696e-01 3.05439651e-01 -9.32278991e-01
1.71852160e+00 -2.02332754e-02 6.17374241e-01 -1.11086214e+00
2.73013651e-01 6.56866848e-01 4.65956181e-01 -2.44676307e-01
5.34550905e-01 -3.32277268e-01 6.91055283e-02 -8.55208561e-02
-2.70771533e-01 -5.33120275e-01 1.93497270e-01 -3.51660490e-01
6.00031972e-01 -1.73412457e-01 5.42288303e-01 -7.70086050e-01
8.95629168e-01 1.15393531e-02 4.52063024e-01 3.81840557e-01
2.40274265e-01 2.00706884e-01 2.87531614e-01 -1.35898525e-02
-8.59280109e-01 -9.57906008e-01 -6.47791803e-01 -1.61715776e-01
2.47150719e-01 1.85980916e-01 -4.88295764e-01 1.95149016e-02
4.18334097e-01 1.11876798e+00 -2.81791896e-01 -1.99483125e-03
-7.74371743e-01 -5.86702824e-01 3.18326980e-01 -3.61546315e-02
3.07779104e-01 -5.40316820e-01 -1.02460682e+00 1.91610485e-01
2.65733272e-01 -1.06534243e+00 4.77885544e-01 2.79571295e-01
-9.82793212e-01 -9.72970128e-01 -4.10643190e-01 -9.97699052e-02
4.84052330e-01 2.71557570e-01 6.83477163e-01 -3.46972764e-01
-4.34492886e-01 4.45592463e-01 -2.99592108e-01 -6.58517003e-01
-4.64197457e-01 -4.91278648e-01 2.69077837e-01 -2.29458767e-03
6.05458200e-01 -6.08322978e-01 -1.34851083e-01 4.62654531e-01
-4.95896131e-01 -6.26705527e-01 9.64471579e-01 4.69322354e-01
3.25163126e-01 -1.00625237e-03 3.68430048e-01 -6.27918601e-01
2.80122966e-01 1.28912941e-01 -1.40756881e+00 -3.39374810e-01
-9.86666560e-01 8.30004960e-02 2.75595099e-01 -9.48889256e-02
-9.79087472e-01 -1.52356714e-01 -4.32576150e-01 -2.36515686e-01
-2.65366167e-01 4.38934416e-01 -4.46749300e-01 -7.50764430e-01
7.39860535e-01 3.30445945e-01 3.37101668e-01 -3.05557549e-01
2.00501651e-01 3.24467391e-01 4.85457182e-01 -1.90002635e-01
1.56286228e+00 4.52341288e-01 9.23233569e-01 -1.25053382e+00
-1.48892924e-01 -6.17373288e-01 -4.45422530e-01 -5.12364328e-01
6.07606590e-01 -7.41993546e-01 -5.42095840e-01 2.96529204e-01
-1.11688495e+00 2.80209541e-01 -5.57904541e-01 1.28834879e+00
-4.08243805e-01 5.36733091e-01 2.56913841e-01 -1.09093189e+00
-1.46210000e-01 -1.11186790e+00 4.27797318e-01 1.28766194e-01
5.35703525e-02 -6.32661164e-01 1.12463191e-01 -3.00290938e-02
6.77331030e-01 2.35904023e-01 6.13151670e-01 -6.13592863e-01
-8.26166630e-01 -5.98282218e-01 -1.67861909e-01 3.10917586e-01
-3.22169602e-01 -3.84433478e-01 -6.09066308e-01 -3.03588480e-01
4.06835526e-01 5.33891082e-01 5.78204691e-01 5.71088135e-01
1.12469293e-01 3.88065010e-01 -4.51809973e-01 4.91518855e-01
1.86000741e+00 4.53138858e-01 8.36831510e-01 4.51281786e-01
-8.73853937e-02 4.05237556e-01 1.60581338e+00 4.52938706e-01
-3.47187012e-01 1.05158138e+00 8.29307556e-01 1.92561179e-01
-5.94790205e-02 3.01908612e-01 1.57711878e-01 1.77923545e-01
-1.12355441e-01 2.65592542e-02 -5.10701954e-01 3.08431119e-01
-1.46359360e+00 -9.67864990e-01 -7.89716423e-01 2.77868199e+00
-2.82820873e-02 5.97422719e-01 -3.99246782e-01 4.06658053e-01
5.68832994e-01 1.51189566e-01 1.78009316e-01 -4.42999899e-01
-1.36886075e-01 5.71135044e-01 1.09161115e+00 1.07502091e+00
-4.96753991e-01 2.05736473e-01 4.63580608e+00 8.02163482e-01
-1.30667460e+00 1.21233754e-01 -6.42048776e-01 3.18564653e-01
-3.61867785e-01 1.99843988e-01 -1.18437052e+00 3.63359839e-01
9.23339128e-01 -2.42247328e-01 -1.56872213e-01 6.19398773e-01
4.42227781e-01 -6.68002069e-01 -5.31858981e-01 8.95376027e-01
3.56149286e-01 -8.58569324e-01 -1.32261813e-01 4.96824563e-01
1.38159424e-01 -2.16502726e-01 -9.69556868e-02 7.08325803e-02
-6.97992325e-01 -5.30633211e-01 9.05006528e-01 7.98652947e-01
6.82600141e-01 -8.56451690e-01 1.12701094e+00 3.86576027e-01
-1.26933014e+00 -4.66864835e-03 -4.93654519e-01 -6.10628314e-02
8.08610797e-01 1.29499733e+00 -1.27978039e+00 1.21095848e+00
1.84021413e-01 9.70136225e-02 -2.45470345e-01 1.39005721e+00
-1.99755311e-01 2.92917997e-01 -6.44289076e-01 -3.39147508e-01
1.19705029e-01 -8.43546152e-01 1.28549254e+00 1.15935159e+00
1.16596615e+00 -2.44485974e-01 -3.50540847e-01 7.21926451e-01
7.85286367e-01 1.07628778e-02 -7.07963705e-01 3.90659988e-01
6.82744026e-01 1.25793839e+00 -2.40043953e-01 3.28137283e-03
1.01696789e-01 1.93353489e-01 -7.13737488e-01 1.94599509e-01
-9.75006878e-01 -5.60461521e-01 7.15561628e-01 7.82942772e-01
5.45841634e-01 -6.08360291e-01 -2.22775161e-01 -2.99070388e-01
1.42367452e-01 -3.32152277e-01 -2.66262263e-01 -6.83255970e-01
-6.89722061e-01 6.66224897e-01 5.45121074e-01 -1.82002127e+00
-6.83207273e-01 -9.28125024e-01 -3.37768316e-01 1.22670615e+00
-1.63658369e+00 -9.65124667e-01 -3.30613434e-01 2.43400261e-01
1.18649259e-01 -3.36143076e-01 4.74179715e-01 4.65928525e-01
2.61295736e-01 1.76421925e-01 3.75154138e-01 -3.79958302e-01
6.86329603e-01 -8.40528190e-01 -4.53474000e-02 1.01168907e+00
-2.39392534e-01 3.86248589e-01 1.60540175e+00 -8.12634647e-01
-1.25105441e+00 -7.62374997e-01 1.06565297e+00 1.83551937e-01
6.05220377e-01 -4.03982371e-01 -6.07513070e-01 2.68633544e-01
1.02280058e-01 -1.71778202e-01 1.73302323e-01 -1.65070757e-01
-1.24990031e-01 -6.50614619e-01 -1.30105078e+00 1.74834713e-01
5.29814839e-01 -6.95949495e-02 -8.38734627e-01 -5.20951999e-03
2.29926080e-01 -4.27616090e-01 -7.43864238e-01 8.55718374e-01
5.16323388e-01 -1.26142883e+00 9.35103297e-01 3.92795533e-01
-5.93551040e-01 -6.08792901e-01 -2.07586408e-01 -1.24801278e+00
-1.56446174e-01 -4.05159503e-01 5.21704733e-01 1.03268838e+00
2.64442801e-01 -1.27930462e+00 8.44560146e-01 -3.31818551e-01
-1.73659787e-01 -3.29341888e-01 -1.75830746e+00 -1.08651865e+00
-3.20729524e-01 -5.31980455e-01 3.42505932e-01 3.60196114e-01
-3.07141989e-01 1.25528574e-01 -8.33759978e-02 6.63108468e-01
1.11454582e+00 1.09477371e-01 9.96299028e-01 -1.88613582e+00
-1.55679375e-01 -6.20596893e-02 -8.92963052e-01 -7.62639642e-01
-4.80576716e-02 -3.76982927e-01 5.67975044e-02 -1.33906758e+00
-6.81858122e-01 -4.43055630e-01 2.48996437e-01 -5.75131655e-01
6.64947510e-01 1.04624957e-01 2.94461727e-01 -1.21613294e-01
3.11092377e-01 5.26474535e-01 8.72621596e-01 1.67878404e-01
3.08533143e-02 4.89443094e-01 2.22406596e-01 5.95617056e-01
5.31916559e-01 -5.99161506e-01 -3.97219896e-01 2.84297504e-02
1.88929558e-01 1.34210587e-01 5.01824021e-01 -1.74993587e+00
2.40515441e-01 9.40509140e-02 2.23042145e-01 -1.37093580e+00
8.98451686e-01 -1.41991615e+00 6.51528239e-01 9.02189732e-01
6.98085666e-01 1.33134618e-01 2.47366890e-01 4.94951397e-01
-3.30739141e-01 -7.64527559e-01 9.21698034e-01 3.38865250e-01
-6.15382731e-01 -3.25067341e-01 -6.51996732e-01 -6.10777795e-01
1.17942321e+00 -6.15291059e-01 -5.16272962e-01 -4.57951188e-01
-5.06932259e-01 -2.49206513e-01 2.53627032e-01 -2.61773914e-01
5.67298651e-01 -1.11691368e+00 -7.82829285e-01 5.03800631e-01
1.84511602e-01 -4.50579852e-01 5.06590188e-01 1.28661513e+00
-6.68702900e-01 7.60674894e-01 -3.12337458e-01 -7.55143046e-01
-1.45015788e+00 2.61502296e-01 4.54992861e-01 -1.85754672e-01
-3.67859066e-01 2.35769257e-01 -3.16145778e-01 -2.69997865e-01
-1.66415349e-01 -2.21210241e-01 -2.60446906e-01 2.06173122e-01
2.04735100e-01 4.12248880e-01 4.43960279e-01 -7.70303309e-01
-5.10315299e-01 1.02882349e+00 5.83126366e-01 -3.71910304e-01
1.11876750e+00 4.06131186e-02 -1.79901555e-01 2.82134265e-01
7.08418906e-01 7.91012824e-01 -8.54943514e-01 8.50551501e-02
7.34773837e-03 -4.33139324e-01 3.68158659e-03 -5.82516074e-01
-5.44436872e-01 7.99544871e-01 8.15896928e-01 1.73408985e-02
7.95888722e-01 -4.82976794e-01 1.51156396e-01 4.84616011e-01
6.82913005e-01 -7.96734691e-01 -6.20335519e-01 6.00473285e-01
7.71676302e-01 -4.87067819e-01 4.30028796e-01 -8.15246463e-01
-1.13798782e-01 1.34818959e+00 2.61186630e-01 -2.52511024e-01
7.64418304e-01 6.33472025e-01 7.27296099e-02 4.87666316e-02
1.04333684e-02 -5.06695986e-01 1.65655747e-01 7.01160312e-01
-1.11969905e-02 7.48155713e-02 -1.13323212e+00 1.66966781e-01
-1.58159614e-01 -7.85502344e-02 5.89004278e-01 8.18394363e-01
-9.23702717e-01 -1.36476481e+00 -1.06726718e+00 2.30444565e-01
3.22881073e-01 1.85065567e-01 1.01832189e-01 1.46941066e+00
3.87496471e-01 6.81646585e-01 2.36889839e-01 -1.62601084e-01
7.77863085e-01 1.54690623e-01 6.02405190e-01 -1.14030674e-01
-1.36000156e-01 2.70636499e-01 1.36280581e-01 -4.11166519e-01
-2.81914592e-01 -8.78873229e-01 -1.03927648e+00 4.84625436e-02
-3.64963233e-01 6.87414587e-01 1.24739087e+00 8.89182985e-01
-8.85664299e-03 4.20272142e-01 4.82356280e-01 -8.77660036e-01
-7.63596654e-01 -1.07032704e+00 -1.01268935e+00 -4.06614959e-01
4.42565829e-02 -9.74005401e-01 -5.89958131e-01 -7.42794752e-01] | [6.728953838348389, 0.9547088146209717] |
22325d73-dbd8-4ca5-81b8-480822e713cf | document-level-relation-extraction-with | 2010.11304 | null | https://arxiv.org/abs/2010.11304v3 | https://arxiv.org/pdf/2010.11304v3.pdf | Document-Level Relation Extraction with Adaptive Thresholding and Localized Context Pooling | Document-level relation extraction (RE) poses new challenges compared to its sentence-level counterpart. One document commonly contains multiple entity pairs, and one entity pair occurs multiple times in the document associated with multiple possible relations. In this paper, we propose two novel techniques, adaptive thresholding and localized context pooling, to solve the multi-label and multi-entity problems. The adaptive thresholding replaces the global threshold for multi-label classification in the prior work with a learnable entities-dependent threshold. The localized context pooling directly transfers attention from pre-trained language models to locate relevant context that is useful to decide the relation. We experiment on three document-level RE benchmark datasets: DocRED, a recently released large-scale RE dataset, and two datasets CDRand GDA in the biomedical domain. Our ATLOP (Adaptive Thresholding and Localized cOntext Pooling) model achieves an F1 score of 63.4, and also significantly outperforms existing models on both CDR and GDA. | ['Jing Huang', 'Tengyu Ma', 'Kevin Huang', 'Wenxuan Zhou'] | 2020-10-21 | null | null | null | null | ['document-level-relation-extraction'] | ['natural-language-processing'] | [ 4.04220909e-01 1.12404227e-01 -4.20172989e-01 -4.20963436e-01
-1.35095525e+00 -5.96575558e-01 3.98560524e-01 8.16676974e-01
-7.59456217e-01 9.23429132e-01 1.97314978e-01 -1.08725771e-01
-1.15338355e-01 -6.15515947e-01 -3.96621794e-01 -5.90151548e-01
8.24180618e-02 5.33055544e-01 1.67701244e-01 1.23285010e-01
2.15960462e-02 3.70708257e-01 -1.06735408e+00 7.95487881e-01
7.26005614e-01 1.03779733e+00 2.65199631e-01 5.00896394e-01
-2.28880182e-01 8.27330649e-01 -8.45727921e-01 -4.74347770e-01
-1.88614398e-01 -2.37128794e-01 -1.31121349e+00 -3.18125904e-01
4.37789559e-01 2.65574902e-01 -2.86449064e-02 9.17308211e-01
5.75524449e-01 4.23386171e-02 7.53497481e-01 -9.32729661e-01
-8.60206664e-01 6.79102242e-01 -8.44087124e-01 6.53911829e-01
3.36297065e-01 -7.67889738e-01 1.51649475e+00 -1.04907775e+00
1.12341070e+00 1.11401153e+00 6.07494771e-01 2.78745055e-01
-1.02305174e+00 -7.44081616e-01 2.42400736e-01 2.18005896e-01
-1.77974141e+00 -2.25375086e-01 1.71152845e-01 -2.99240857e-01
1.68365479e+00 2.07661062e-01 1.52327912e-02 9.27965701e-01
4.89192426e-01 5.87296188e-01 1.08474469e+00 -4.94724393e-01
-6.28113151e-02 -1.71453521e-01 5.77120543e-01 5.63703299e-01
3.19958478e-01 -6.42649353e-01 -3.68986726e-01 -5.27591765e-01
1.54324025e-01 -5.44850938e-02 -2.00797647e-01 4.41945016e-01
-1.31920958e+00 4.28065032e-01 1.14302389e-01 6.05413556e-01
-3.55795860e-01 -9.94101539e-02 4.50201571e-01 2.23456040e-01
5.46985805e-01 5.85831881e-01 -9.93223369e-01 3.58694166e-01
-7.65286148e-01 1.00431486e-03 6.58262491e-01 1.27298999e+00
6.04544044e-01 -9.28457737e-01 -7.93479979e-01 1.10228252e+00
1.77149981e-01 3.53283077e-01 5.56932926e-01 -4.71855640e-01
6.89850926e-01 7.12510943e-01 -1.76450863e-01 -1.14786744e+00
-8.37462604e-01 -4.47131455e-01 -9.48114038e-01 -7.55889237e-01
5.88922091e-02 -1.95232779e-01 -9.86138940e-01 1.64250135e+00
2.99584270e-01 1.94317743e-01 2.71187901e-01 4.62563992e-01
1.44555974e+00 6.01312280e-01 4.89884764e-01 -3.77793580e-01
2.00169706e+00 -1.05190194e+00 -1.18234611e+00 -4.27083641e-01
9.19593096e-01 -8.97585213e-01 4.92424160e-01 2.42496118e-01
-7.34247744e-01 -2.71806657e-01 -1.05803859e+00 -6.33893251e-01
-6.82712018e-01 5.46311319e-01 4.21957731e-01 8.74043182e-02
-7.99257040e-01 2.81459719e-01 -2.69449204e-01 -3.74490470e-01
3.36808592e-01 4.11864519e-01 -6.71675861e-01 -1.12082779e-01
-1.59813905e+00 1.08593142e+00 5.75886846e-01 2.15784833e-01
-3.17945778e-01 -3.58845830e-01 -8.76134872e-01 2.95408696e-01
5.87057233e-01 -5.61002314e-01 1.15278137e+00 -1.04985349e-01
-8.55192065e-01 1.30128825e+00 -5.00502229e-01 -1.14519060e-01
-1.14960261e-01 -1.59033671e-01 -5.62952399e-01 1.36124060e-01
3.56128246e-01 3.67681652e-01 2.45434046e-01 -9.85309720e-01
-8.63696396e-01 -2.65061200e-01 -1.99067414e-01 1.14159338e-01
-9.04674903e-02 5.68893850e-01 -6.07110739e-01 -7.94343293e-01
1.25541180e-01 -7.14487195e-01 -1.71989605e-01 -6.95315003e-01
-7.37398624e-01 -7.59313941e-01 5.39421618e-01 -7.17960536e-01
1.52951336e+00 -1.95633614e+00 1.80123318e-02 1.67816117e-01
4.08025831e-01 1.29080817e-01 -2.72473067e-01 3.57029319e-01
-4.09461945e-01 4.85084623e-01 -1.22386523e-01 -4.51929301e-01
-3.66527289e-01 1.12516403e-01 4.33805808e-02 1.55419052e-01
4.41241950e-01 1.01256347e+00 -1.04291892e+00 -9.11485255e-01
-7.66513944e-01 2.97497511e-01 -4.23047459e-03 5.96479550e-02
-2.25944426e-02 2.99972240e-02 -5.04887402e-01 8.20901275e-01
4.72782850e-01 -7.30102777e-01 7.58665502e-01 -4.04477477e-01
2.26686612e-01 5.22547543e-01 -9.54694510e-01 1.47381496e+00
-5.03947079e-01 4.86951739e-01 -2.33173802e-01 -7.31989145e-01
9.75792110e-01 5.56912959e-01 4.02732760e-01 -5.23319364e-01
2.42829338e-01 1.39698014e-01 -2.73777843e-01 -6.79332614e-01
7.29367733e-01 5.70004992e-02 -3.63722414e-01 2.11464956e-01
2.80785918e-01 4.77276683e-01 2.86977112e-01 3.07611614e-01
1.57506430e+00 -2.29747608e-01 7.88932502e-01 -2.11761519e-01
6.47796154e-01 -1.68794274e-01 1.04893625e+00 6.16318762e-01
-3.62392366e-02 6.41841531e-01 7.62415528e-01 -1.13758668e-01
-4.55047697e-01 -5.22228062e-01 -2.90588915e-01 1.10734749e+00
1.16926566e-01 -9.26174462e-01 -3.26907724e-01 -1.18317688e+00
-6.16258532e-02 4.47658747e-01 -7.42964923e-01 8.76274481e-02
-7.04553843e-01 -1.12869883e+00 6.27616704e-01 5.65729856e-01
4.62045878e-01 -9.65908468e-01 1.25231324e-02 4.97780859e-01
-7.56940126e-01 -1.63078749e+00 -7.07869411e-01 6.48588181e-01
-4.84568536e-01 -1.15679955e+00 -5.69614530e-01 -1.11438990e+00
6.54191732e-01 3.07122171e-02 1.31215489e+00 6.23017258e-04
-2.85827428e-01 -1.10150836e-01 -4.23622400e-01 -2.61217922e-01
-1.22939222e-01 4.60714281e-01 -2.97515035e-01 -3.03415209e-01
8.51328433e-01 -1.79665029e-01 -2.82316267e-01 3.18333209e-01
-6.10689223e-01 -1.01114817e-01 8.61249804e-01 1.07980192e+00
1.18842983e+00 -1.25055790e-01 1.04289889e+00 -1.28446794e+00
7.81650007e-01 -7.24950790e-01 -1.29387394e-01 1.02688241e+00
-7.56984532e-01 7.29799867e-02 3.36089224e-01 -3.14463317e-01
-9.58436966e-01 9.07004904e-03 -1.05689518e-01 2.14946553e-01
-2.33188942e-01 8.80202949e-01 -2.76396692e-01 3.65690321e-01
3.75656664e-01 -2.89290845e-01 -7.92330086e-01 -5.43561339e-01
3.45741004e-01 1.07760286e+00 5.56166291e-01 -4.99928772e-01
6.83973506e-02 5.72391134e-03 1.16164245e-01 -2.79264838e-01
-1.40164471e+00 -7.90912092e-01 -6.99477196e-01 3.27697664e-01
1.23954463e+00 -9.67831612e-01 -5.00711143e-01 2.20290080e-01
-1.47236741e+00 -4.16044369e-02 3.17127220e-02 3.21472496e-01
-2.08694041e-02 1.71069831e-01 -1.06256151e+00 -2.69450247e-01
-6.76473856e-01 -9.42642450e-01 1.23302889e+00 2.64973223e-01
-2.60874808e-01 -6.69077277e-01 1.51747212e-01 3.51435304e-01
-1.52072102e-01 1.83611646e-01 1.18783116e+00 -1.02329445e+00
-1.56839982e-01 -3.34054977e-01 -6.17470920e-01 -6.79069683e-02
2.95939744e-01 -2.17618123e-01 -9.93868232e-01 -5.83463013e-02
-4.40369040e-01 -3.75990927e-01 1.09602511e+00 -8.28325469e-03
1.11619210e+00 1.79696288e-02 -8.99479330e-01 8.13218728e-02
1.16837275e+00 3.64384234e-01 3.29556525e-01 2.10140899e-01
6.55903816e-01 4.46634382e-01 8.86157870e-01 4.80203629e-02
7.73658216e-01 6.69402659e-01 -2.74524868e-01 -4.36929613e-02
-1.51966140e-01 1.00681819e-01 -7.50237480e-02 1.02606058e+00
2.35894863e-02 -6.51641071e-01 -9.99011934e-01 5.44642627e-01
-1.96844006e+00 -7.44386971e-01 -1.34819269e-01 1.70361733e+00
1.25539029e+00 3.92885767e-02 -3.81304204e-01 -2.23922178e-01
8.84142518e-01 -1.46993518e-01 -1.73239022e-01 -4.03841615e-01
-3.91441435e-01 2.38942161e-01 5.00213861e-01 3.52690786e-01
-1.48724937e+00 9.38900530e-01 6.34542370e+00 9.93560493e-01
-5.54809928e-01 3.51052195e-01 8.49451065e-01 1.35781646e-01
-5.51225543e-02 -1.34831950e-01 -1.41916430e+00 5.67425117e-02
1.15253854e+00 -1.32394046e-01 -2.86002427e-01 5.80302119e-01
-2.42033258e-01 -2.46687099e-01 -1.21195805e+00 7.96915889e-01
2.68334925e-01 -1.22882390e+00 -2.62924194e-01 9.47563797e-02
6.61032319e-01 6.72259778e-02 -2.74356574e-01 3.91287327e-01
2.23839432e-01 -9.37957287e-01 2.07268193e-01 4.26367730e-01
1.01042497e+00 -6.80237293e-01 1.21290851e+00 4.35289517e-02
-1.41044962e+00 1.10864028e-01 -2.76038408e-01 5.23849189e-01
4.30174060e-02 8.17539036e-01 -8.86711359e-01 8.97862375e-01
7.37482131e-01 7.18865216e-01 -6.79957688e-01 7.66994119e-01
-5.12534261e-01 4.04658616e-01 -2.75572300e-01 -5.11217602e-02
4.80193906e-02 2.12249681e-01 2.08473429e-01 1.81642461e+00
8.40519667e-02 4.61384654e-01 2.12956935e-01 4.84706879e-01
-6.89031839e-01 4.63454545e-01 -1.43944502e-01 1.30216986e-01
5.89613080e-01 1.73467517e+00 -9.04983282e-01 -4.46984708e-01
-5.68154991e-01 9.51722622e-01 8.22470188e-01 3.90114605e-01
-6.05988324e-01 -8.31731021e-01 3.15445036e-01 -5.36239862e-01
3.94634128e-01 -8.42058361e-02 -1.21408544e-01 -1.12524498e+00
-4.29602079e-02 -6.03345454e-01 9.08967972e-01 -4.95289683e-01
-1.46452880e+00 9.49100077e-01 -5.13831861e-02 -8.10088158e-01
6.84551336e-03 -4.88882482e-01 -1.85006529e-01 9.44551468e-01
-1.73470581e+00 -1.17516017e+00 -1.63698588e-02 9.38053802e-02
2.80845731e-01 6.54319376e-02 1.13662624e+00 7.64845848e-01
-9.27704513e-01 9.12775397e-01 5.97494841e-02 4.05999452e-01
1.21772385e+00 -1.44597316e+00 2.86218435e-01 6.41740084e-01
9.23487395e-02 8.41382146e-01 9.09256116e-02 -9.29475307e-01
-7.02467978e-01 -1.13014126e+00 1.98869932e+00 -4.08706516e-01
5.25534153e-01 -5.42710185e-01 -1.03147733e+00 8.00397575e-01
2.63441116e-01 2.20688850e-01 1.14184940e+00 4.67436790e-01
-4.99976397e-01 1.88267067e-01 -1.38617194e+00 2.17420802e-01
1.03010154e+00 -6.19247019e-01 -7.44067729e-01 5.78603566e-01
7.75433779e-01 -4.81448412e-01 -1.41395307e+00 4.38331574e-01
3.98689985e-01 -6.44127131e-02 8.37449849e-01 -7.64196634e-01
3.89409751e-01 -2.82121211e-01 -1.58569351e-01 -1.03500426e+00
-5.62715173e-01 -3.56699765e-01 -3.90491843e-01 1.61719978e+00
9.96184647e-01 -5.02314687e-01 2.80854642e-01 5.37630320e-01
-1.47759527e-01 -1.15879118e+00 -9.78230298e-01 -6.48342490e-01
1.15835465e-01 -1.75661109e-02 5.21118641e-01 1.22427082e+00
2.14466572e-01 6.95781410e-01 -4.87714857e-02 3.74402225e-01
1.71752334e-01 3.14734042e-01 -6.94034919e-02 -1.05125999e+00
-1.17804065e-01 -1.82303324e-01 -2.15844929e-01 -8.97125065e-01
3.46235305e-01 -1.33887947e+00 1.91653550e-01 -1.74352372e+00
5.70527256e-01 -5.34545481e-01 -7.47444570e-01 1.02412951e+00
-5.51814795e-01 2.68680423e-01 -1.83542028e-01 6.58536032e-02
-1.01316381e+00 1.13136664e-01 9.81823564e-01 -2.03562498e-01
-5.45630641e-02 -3.88641506e-01 -7.85306454e-01 5.28808534e-01
6.14874840e-01 -9.36578333e-01 9.90418810e-03 -1.65226892e-01
6.63884103e-01 -1.36659071e-01 -1.04664490e-01 -3.98132414e-01
4.52157319e-01 -9.84686017e-02 4.20332223e-01 -7.69600689e-01
-1.44212931e-01 -5.71798027e-01 -2.79438674e-01 -1.01926982e-01
-7.39617705e-01 1.56009465e-01 1.21458188e-01 5.94355047e-01
-2.78245181e-01 -3.37533861e-01 5.03667355e-01 7.09004924e-02
-5.78802764e-01 2.04434782e-01 -4.54632074e-01 3.80026340e-01
7.56419599e-01 5.37010491e-01 -7.82826245e-01 1.47807643e-01
-9.88825977e-01 6.48121715e-01 -3.30111623e-01 5.80903590e-01
5.39188743e-01 -1.28949451e+00 -6.77931428e-01 -2.53071785e-01
4.45299864e-01 8.98623988e-02 1.67279378e-01 7.75343418e-01
-3.88237722e-02 5.33931673e-01 3.65412474e-01 -1.37340873e-01
-1.73598886e+00 5.79421699e-01 2.64777303e-01 -8.91332150e-01
-4.95913684e-01 1.06849313e+00 -1.27126396e-01 -4.22733575e-01
8.74564201e-02 -7.03448415e-01 -5.40622413e-01 3.34619731e-01
7.54239321e-01 1.09646335e-01 4.40824777e-01 -7.07174778e-01
-7.12566674e-01 6.17774904e-01 -6.44049048e-01 2.85550624e-01
1.16468382e+00 -1.98489323e-01 -5.01593053e-01 4.39836115e-01
1.24566448e+00 6.27600253e-02 -3.25210541e-01 -6.49884045e-01
6.45133495e-01 4.32899222e-02 1.08522929e-01 -1.26707244e+00
-8.07136297e-01 4.47527021e-01 3.09057534e-01 1.52280569e-01
1.02221000e+00 5.42635024e-01 9.03377712e-01 6.89406872e-01
1.44244820e-01 -1.12817347e+00 -2.74057865e-01 6.95640326e-01
8.95873845e-01 -1.28580034e+00 3.09429407e-01 -8.43906522e-01
-5.69386601e-01 9.67868447e-01 6.60055578e-01 4.04851496e-01
7.74786413e-01 5.48056483e-01 7.76224062e-02 -3.95492315e-01
-8.47112417e-01 -3.21259439e-01 6.72338188e-01 2.64712572e-01
8.31057847e-01 -5.78197837e-02 -7.59266436e-01 1.22696316e+00
2.87088215e-01 -5.00674732e-02 2.49344051e-01 1.18476522e+00
-1.93216860e-01 -1.39969373e+00 -2.20535714e-02 7.06597269e-01
-9.80564713e-01 -3.63236815e-01 -4.58089232e-01 3.76931548e-01
4.41044927e-01 1.12560511e+00 -5.98504320e-02 -3.94312501e-01
4.04066026e-01 3.46213311e-01 2.88963795e-01 -9.10913110e-01
-9.63078976e-01 1.95896015e-01 5.68123341e-01 -3.40032101e-01
-6.94285274e-01 -7.86107004e-01 -1.71483159e+00 2.99201727e-01
-6.74306035e-01 9.76953506e-02 2.36836150e-01 1.10080814e+00
6.44141912e-01 7.22762406e-01 2.78526902e-01 -8.11859295e-02
-3.65696289e-02 -1.09112465e+00 -5.03873527e-01 2.03047529e-01
2.85832584e-01 -7.03591585e-01 7.95903429e-03 -1.15446620e-01] | [9.151293754577637, 8.726655006408691] |
5096a77e-1d3e-409f-ac6c-92f55a5ec322 | real-time-3d-head-pose-and-facial-landmark | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Papazov_Real-Time_3D_Head_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Papazov_Real-Time_3D_Head_2015_CVPR_paper.pdf | Real-Time 3D Head Pose and Facial Landmark Estimation From Depth Images Using Triangular Surface Patch Features | We present a real-time system for 3D head pose estimation and facial landmark localization using a commodity depth sensor. We introduce a novel triangular surface patch (TSP) descriptor, which encodes the shape of the 3D surface of the face within a triangular area. The proposed descriptor is viewpoint invariant, and it is robust to noise and to variations in the data resolution. Using a fast nearest neighbor lookup, TSP descriptors from an input depth map are matched to the most similar ones that were computed from synthetic head models in a training phase. The matched triangular surface patches in the training set are used to compute estimates of the 3D head pose and facial landmark positions in the input depth map. By sampling many TSP descriptors, many votes for pose and landmark positions are generated which together yield robust final estimates. We evaluate our approach on the publicly available Biwi Kinect Head Pose Database to compare it against state-of-the-art methods. Our results show a significant improvement in the accuracy of both pose and landmark location estimates while maintaining real-time speed. | ['Tim K. Marks', 'Chavdar Papazov', 'Michael Jones'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['head-pose-estimation'] | ['computer-vision'] | [-1.56843720e-03 5.79160489e-02 7.89598823e-02 -7.45043337e-01
-1.11894655e+00 -3.04646343e-01 4.94014442e-01 1.56236216e-01
-4.34880167e-01 1.48582697e-01 2.25896001e-01 6.42300546e-01
2.14720994e-01 -5.89732230e-01 -5.99124610e-01 -5.67805886e-01
-6.82358295e-02 1.04301262e+00 3.53800684e-01 -1.65445656e-01
3.55114847e-01 1.04942226e+00 -2.04468441e+00 -2.01308340e-01
9.17363837e-02 1.18955231e+00 -1.44496292e-01 1.56063646e-01
1.99692219e-01 -2.87155837e-01 -4.81897086e-01 -1.16226412e-01
2.31565252e-01 7.07244650e-02 -3.13584507e-01 2.34483942e-01
9.64793980e-01 -3.51110041e-01 3.12228911e-02 7.50717103e-01
9.86425579e-01 1.73468105e-02 6.45054042e-01 -1.10452199e+00
2.64568657e-01 -4.94797468e-01 -4.99670953e-01 -3.32905114e-01
1.27468741e+00 -5.19166827e-01 5.08007884e-01 -1.52177477e+00
1.02762866e+00 1.33718455e+00 8.92225862e-01 7.15062737e-01
-1.10328388e+00 -8.26653123e-01 -1.11825638e-01 5.36708869e-02
-2.04142547e+00 -1.04945540e+00 8.20644855e-01 -2.22494051e-01
7.85686791e-01 3.80100787e-01 9.35595155e-01 5.19524455e-01
1.01180471e-01 3.20660174e-01 8.97622645e-01 -5.88494480e-01
5.69887042e-01 -1.99235886e-01 -2.84433991e-01 1.05120838e+00
2.46181991e-03 1.14239536e-01 -1.08812058e+00 -5.99587619e-01
9.28475201e-01 2.92702261e-02 -1.89132169e-01 -8.52967441e-01
-6.72177553e-01 5.62370181e-01 5.00102818e-01 1.00008868e-01
-5.84927380e-01 -1.57972902e-01 -2.72135716e-02 -1.84127972e-01
6.03563547e-01 -2.63490736e-01 -1.71521425e-01 -1.23652898e-01
-1.00239980e+00 4.52426344e-01 8.92690122e-01 1.12758470e+00
9.59383368e-01 -3.85578096e-01 2.40458712e-01 6.79658115e-01
7.02065647e-01 9.52084303e-01 3.01901609e-01 -1.06954634e+00
7.54166842e-02 6.57532752e-01 -4.17247117e-02 -9.90638852e-01
-6.48799777e-01 4.30959702e-01 -2.09098786e-01 2.91577518e-01
2.89455086e-01 2.79092789e-01 -1.00226212e+00 1.34791005e+00
1.05712771e+00 2.36194730e-02 -5.51730633e-01 9.26612914e-01
9.12831843e-01 9.34099108e-02 -3.31872910e-01 -1.46899045e-01
1.47720754e+00 -1.36455968e-01 -4.75394726e-01 -3.65199894e-01
1.54749483e-01 -7.64143288e-01 4.33483362e-01 3.20543468e-01
-1.01697874e+00 -2.30778873e-01 -7.79412150e-01 -2.13133559e-01
-5.13522699e-02 -9.49240327e-02 1.59237444e-01 7.91886508e-01
-1.40175605e+00 3.72640371e-01 -1.08780360e+00 -6.16303384e-01
2.06527203e-01 9.07387257e-01 -9.24334049e-01 -1.39192730e-01
-3.69687766e-01 8.27066422e-01 -3.04765016e-01 1.38928201e-02
-5.03123343e-01 -4.29315299e-01 -1.18075788e+00 -4.77754086e-01
-1.44904941e-01 -3.22889358e-01 1.27239430e+00 -3.24627578e-01
-1.76864231e+00 1.54396641e+00 -9.29618359e-01 2.81822473e-01
3.78803790e-01 1.55025767e-02 6.58949465e-02 1.66528463e-01
3.33790123e-01 7.43936360e-01 9.29535687e-01 -1.10813165e+00
-3.56922269e-01 -1.34438062e+00 -5.71399033e-01 3.39605153e-01
2.15722501e-01 1.62510768e-01 -8.32455337e-01 -7.84458891e-02
1.09022272e+00 -1.00898528e+00 7.91158453e-02 5.85808456e-01
-3.37559357e-02 -3.57858241e-01 5.32896221e-01 -6.24614179e-01
6.39434099e-01 -2.10089731e+00 -1.27461288e-04 5.59279323e-01
1.61038004e-02 -2.85212606e-01 1.57396883e-01 1.18926376e-01
1.36795148e-01 -4.13392842e-01 1.46129802e-01 -1.03804874e+00
6.42404482e-02 2.44652241e-01 1.35482892e-01 1.18344343e+00
-3.76295418e-01 5.63907087e-01 -7.10416198e-01 -5.94425201e-01
3.13480139e-01 9.29380417e-01 -3.92909944e-01 2.76460469e-01
1.73456639e-01 3.52109402e-01 -3.94581437e-01 1.21772170e+00
9.37886477e-01 4.68430251e-01 3.76340970e-02 -1.37327909e-01
-1.14524670e-01 4.30043757e-01 -1.33734155e+00 1.96590972e+00
-1.66627660e-01 2.97697425e-01 3.26435953e-01 -1.57343790e-01
1.19475925e+00 4.18731213e-01 5.18957198e-01 -7.64947057e-01
2.74708241e-01 4.92164075e-01 -7.73970902e-01 -1.36703670e-01
3.28051865e-01 -1.75490499e-01 6.65379390e-02 5.10811269e-01
-3.64195965e-02 -5.17652094e-01 -5.66888630e-01 -5.02826869e-01
6.71988964e-01 2.31958434e-01 5.11503339e-01 -1.26245022e-01
3.24742913e-01 -2.90985048e-01 3.81774038e-01 2.65545025e-02
-2.79876977e-01 1.02213562e+00 7.94623643e-02 -6.71111465e-01
-8.44637156e-01 -1.15047920e+00 -4.92468774e-01 9.91019070e-01
5.50690712e-03 -3.41234475e-01 -1.06139231e+00 -2.99324006e-01
2.57852882e-01 6.02716915e-02 -7.38262594e-01 3.74738574e-01
-6.78308308e-01 -2.03147307e-01 1.70156866e-01 4.03653741e-01
1.04745075e-01 -8.65913987e-01 -9.41852272e-01 1.35576827e-02
1.32154226e-01 -9.67686176e-01 -5.19024611e-01 -2.30028480e-02
-9.71650124e-01 -1.03199804e+00 -7.40582228e-01 -9.25652504e-01
1.12162614e+00 5.05927019e-02 8.57783258e-01 -9.33594704e-02
-1.94811136e-01 5.06448686e-01 -1.26212299e-01 -4.63646591e-01
1.77170217e-01 -2.07142293e-01 5.49599528e-01 -5.89242252e-03
7.34876573e-01 -5.02901793e-01 -6.91240430e-01 5.26016772e-01
-2.64054239e-01 -5.13668239e-01 -7.41609409e-02 3.10681671e-01
1.03439748e+00 -7.24204421e-01 -1.63309813e-01 -2.56900251e-01
2.16806203e-01 -7.70724863e-02 -8.12522352e-01 -5.89937568e-02
-2.98202127e-01 6.49761036e-02 -2.47244596e-01 -3.59317631e-01
-4.07479495e-01 7.44963706e-01 -3.36259276e-01 -3.68761480e-01
-2.51266778e-01 -5.83102508e-03 -2.58083761e-01 -5.19745648e-01
7.02421606e-01 7.73979723e-02 3.76920909e-01 -5.94678700e-01
1.54880613e-01 7.79446721e-01 5.57581306e-01 -4.09027278e-01
6.68886721e-01 8.51095557e-01 2.17932045e-01 -9.15821970e-01
-2.54860044e-01 -6.97189391e-01 -1.23886657e+00 -2.20556244e-01
5.24174511e-01 -8.93927395e-01 -9.76882577e-01 5.53121865e-01
-1.37324071e+00 8.67061540e-02 -1.14609130e-01 2.76895285e-01
-6.51639819e-01 7.30044395e-02 -3.85184884e-02 -1.00509989e+00
-4.34502125e-01 -1.12842143e+00 2.11836648e+00 1.36315629e-01
-3.39861900e-01 -6.67991042e-01 2.98705250e-01 -9.08192806e-03
1.50841335e-02 5.61637759e-01 2.30550736e-01 -2.65725374e-01
-4.29250091e-01 -8.70229304e-01 2.97259361e-01 -4.32098687e-01
1.22339495e-01 -1.73908770e-01 -1.38487327e+00 -4.22287434e-01
6.47774339e-02 5.78143671e-02 9.31075588e-02 4.93671358e-01
4.26915824e-01 -1.10928267e-01 -5.92017114e-01 6.99364245e-01
1.09934497e+00 1.37537956e-01 3.83560985e-01 3.15351725e-01
3.56691658e-01 7.41614759e-01 6.72605097e-01 6.96438372e-01
7.26392567e-01 1.29537773e+00 3.91662985e-01 1.80998623e-01
2.40242016e-02 -4.01312709e-01 3.55407834e-01 5.66681325e-01
-2.05367595e-01 5.27655661e-01 -1.01274550e+00 3.98077935e-01
-1.28605258e+00 -4.64769125e-01 2.77154177e-01 2.68845987e+00
6.53508067e-01 -3.15203041e-01 4.16739225e-01 2.66144454e-01
6.23549819e-01 -1.10286511e-01 -4.62457895e-01 -2.78961688e-01
3.18169117e-01 6.61851585e-01 3.45540911e-01 7.32226491e-01
-9.63636994e-01 9.06561196e-01 6.82165241e+00 2.66045868e-01
-1.12899113e+00 1.85465336e-01 9.26166922e-02 -1.79098770e-01
7.23904893e-02 -4.52240169e-01 -1.09148538e+00 3.00564594e-03
7.28942752e-01 -9.90631953e-02 2.43115947e-01 1.05076289e+00
1.18996285e-01 -4.01068836e-01 -1.25172603e+00 1.36467123e+00
5.45011401e-01 -8.24862599e-01 -3.92202228e-01 4.50884134e-01
3.65527898e-01 1.13735884e-01 -1.09873116e-01 -1.97620749e-01
-3.93858284e-01 -9.42483306e-01 9.30141568e-01 4.31269258e-01
1.18062127e+00 -7.33131886e-01 5.80351412e-01 2.88820207e-01
-1.50410414e+00 3.99346441e-01 -4.98026311e-01 2.66040377e-02
-1.71491094e-02 1.68273926e-01 -1.30389953e+00 -6.28421456e-02
6.07817829e-01 2.31573537e-01 -4.91286665e-01 1.08164060e+00
-2.48730123e-01 -2.08002895e-01 -8.67884696e-01 -2.47378461e-02
-3.90903234e-01 -3.70660797e-02 4.59504306e-01 7.41880417e-01
4.95220214e-01 2.29435280e-01 2.22800449e-02 5.57172477e-01
6.27183169e-02 3.08505088e-01 -8.42303514e-01 7.58751452e-01
8.72702003e-01 1.02446163e+00 -5.91450214e-01 1.12541288e-01
-5.45816384e-02 1.08189547e+00 2.34527707e-01 -8.85926038e-02
-2.58411556e-01 -9.14845467e-02 8.62305582e-01 5.69323421e-01
1.07560724e-01 -3.75545353e-01 -1.48503691e-01 -6.87592149e-01
3.65193367e-01 -5.09138942e-01 1.46016538e-01 -6.70928836e-01
-7.06958771e-01 7.58710086e-01 2.23408490e-02 -1.26848149e+00
-7.09685922e-01 -5.95171809e-01 -2.33187556e-01 8.93278241e-01
-1.19192326e+00 -1.12720835e+00 -5.28509796e-01 6.69988334e-01
1.63193122e-01 1.72713250e-01 1.41336882e+00 8.75657946e-02
-1.07033029e-01 8.12135816e-01 -1.56202331e-01 1.13442671e-02
5.30578852e-01 -8.76596987e-01 7.65987158e-01 1.46506950e-01
1.92552999e-01 6.90532386e-01 4.73685086e-01 -6.47047579e-01
-1.66800821e+00 -6.22040451e-01 1.29503810e+00 -8.19915712e-01
-8.20847377e-02 -7.41272092e-01 -5.28568447e-01 6.84364259e-01
-6.21968329e-01 3.69040430e-01 4.49847102e-01 -1.79077350e-02
-4.61581767e-01 -2.33042061e-01 -1.72646940e+00 1.18201815e-01
1.09661567e+00 -7.34423459e-01 -5.69679379e-01 2.88553685e-01
-3.57674062e-02 -1.05151224e+00 -8.40946496e-01 2.34554470e-01
1.14555776e+00 -9.82981920e-01 1.01360428e+00 1.65809095e-01
-5.77441275e-01 -2.48813018e-01 -3.24618816e-01 -1.01381588e+00
1.33497613e-02 -5.46564817e-01 1.93122387e-01 7.66672194e-01
-4.29070927e-02 -4.97527778e-01 1.18234205e+00 9.84946012e-01
1.55827016e-01 -6.54666603e-01 -1.89190006e+00 -5.16002536e-01
-4.13763940e-01 -3.82107258e-01 7.82450438e-01 2.69257337e-01
1.26263231e-01 -1.24165244e-01 2.74835855e-01 3.83128524e-01
9.99488235e-01 3.91812213e-02 8.34583998e-01 -1.55679917e+00
7.11518407e-01 -4.19381373e-02 -1.25209296e+00 -9.23661947e-01
2.13194594e-01 -6.20582342e-01 2.40079612e-01 -1.28643596e+00
-1.28599837e-01 -3.44527006e-01 3.80800188e-01 6.37315929e-01
3.94784063e-01 6.55928791e-01 -6.06375039e-02 2.13274330e-01
-3.01618725e-01 3.82655650e-01 7.46852756e-01 2.97150373e-01
-3.41656655e-01 1.35934681e-01 -1.59265604e-02 8.93923342e-01
2.83090353e-01 -6.23856843e-01 2.95745969e-01 -3.75689387e-01
-3.63017261e-01 1.53546855e-01 1.89268321e-01 -1.12933302e+00
4.54060763e-01 1.64531872e-01 6.20387375e-01 -8.16216469e-01
1.08510768e+00 -9.06071246e-01 2.68985450e-01 4.06056106e-01
3.12041223e-01 2.42580608e-01 7.69534940e-03 1.07535772e-01
1.41640067e-01 -9.51444451e-03 8.61241341e-01 -4.21160944e-02
-4.82167184e-01 5.70472479e-01 1.78903520e-01 -3.91877055e-01
9.87741470e-01 -6.44959807e-01 2.54286468e-01 -4.94254172e-01
-7.96381414e-01 -3.44117373e-01 9.95282650e-01 3.31264049e-01
1.07896233e+00 -1.59754908e+00 -6.12786949e-01 1.04576564e+00
3.26413304e-01 1.40231147e-01 -3.19224924e-01 5.81864178e-01
-6.72736526e-01 4.95221943e-01 -1.19004160e-01 -1.00690961e+00
-1.55176616e+00 -1.09969065e-01 6.16570234e-01 5.33164978e-01
-4.43894386e-01 9.59377050e-01 -2.53111213e-01 -7.42374957e-01
4.83909070e-01 -3.65390658e-01 9.10112709e-02 2.29063496e-01
8.01890433e-01 3.95107627e-01 5.71826518e-01 -1.37121904e+00
-9.73048568e-01 1.41763353e+00 4.90197927e-01 -4.17515099e-01
1.37737608e+00 -1.83128849e-01 -1.90172926e-01 3.08989614e-01
1.41391444e+00 2.34063074e-01 -9.85866427e-01 -3.15732181e-01
3.20304669e-02 -7.90748179e-01 2.08460316e-02 -3.61002445e-01
-9.50273752e-01 6.75179601e-01 1.15210688e+00 -3.66728514e-01
8.93457055e-01 2.48984113e-01 5.04004836e-01 9.71290097e-02
1.26140690e+00 -8.12098324e-01 -2.52317488e-01 4.41412687e-01
1.10235727e+00 -1.00606644e+00 2.35265404e-01 -5.21396995e-01
-2.37030789e-01 1.10667777e+00 1.61324650e-01 -1.45093799e-01
7.61148155e-01 3.74675184e-01 2.73375958e-01 -5.64043939e-01
-2.82184910e-02 -2.53447115e-01 5.62064409e-01 1.01433015e+00
3.77542555e-01 1.12897694e-01 9.63594541e-02 -9.41687599e-02
-6.27988696e-01 -7.10086599e-02 -1.54304534e-01 1.04011381e+00
-6.02647722e-01 -1.10757375e+00 -9.31537330e-01 -3.79925780e-02
-8.09980333e-02 2.96173781e-01 -3.79938930e-01 7.03930378e-01
3.12503465e-02 7.59597838e-01 3.50851476e-01 -3.32951725e-01
8.30828249e-01 2.33254433e-01 9.85222161e-01 -7.89104998e-01
-2.70976275e-01 3.50180835e-01 -2.01302052e-01 -1.02680588e+00
-2.85208642e-01 -1.05446506e+00 -1.32513654e+00 -2.49354914e-01
-2.70758241e-01 -1.00147069e-01 1.23886561e+00 6.50945008e-01
4.50247049e-01 -5.70503950e-01 7.00652957e-01 -1.83090627e+00
-3.81494671e-01 -8.69316995e-01 -8.77861738e-01 2.17305109e-01
4.80935544e-01 -1.00067937e+00 -3.21534276e-01 -9.90720540e-02] | [13.500707626342773, 0.16287916898727417] |
ce9bdcdd-456e-4b6b-b4df-fadb33e35da6 | mo-gym-a-library-of-multi-objective | null | null | https://bnaic2022.uantwerpen.be/wp-content/uploads/BNAICBeNeLearn_2022_submission_6485.pdf | https://bnaic2022.uantwerpen.be/wp-content/uploads/BNAICBeNeLearn_2022_submission_6485.pdf | MO-Gym: A Library of Multi-Objective Reinforcement Learning Environments | We introduce MO-Gym, an extensible library containing a diverse set of multi-objective reinforcement learning environments. It introduces a standardized API that facilitates conducting experiments and performance analyses of algorithms designed to interact with multi-objective Markov decision processes. Importantly, it extends the widely used OpenAI Gym API, allowing the reuse of algorithms and features that are well-established in the reinforcement learning community. MO-Gym is available at: https://github.com/LucasAlegre/mo-gym. | ['Bruno C. da Silva', 'Ana L. C. Bazzan', 'Ann Nowé', 'Grégoire Danoy', 'El-Ghazali Talbi', 'Florian Felten', 'Lucas N. Alegre'] | 2022-11-30 | null | null | null | benelux-conference-on-artificial-intelligence-1 | ['multi-objective-reinforcement-learning'] | ['methodology'] | [-5.71453750e-01 -1.83274373e-01 -2.14958966e-01 -1.00634418e-01
-8.15867722e-01 -5.00457168e-01 3.40689510e-01 7.47734010e-02
-4.59056169e-01 1.00747657e+00 -1.38341516e-01 -3.98776740e-01
-4.50953454e-01 -8.37148428e-01 -5.43468595e-01 -9.18943167e-01
-3.55290383e-01 7.09804296e-01 2.62809634e-01 -4.06469494e-01
2.81977981e-01 1.87074184e-01 -1.93192971e+00 -5.09837233e-02
6.70871139e-01 4.52918470e-01 2.15528250e-01 1.31756210e+00
4.40760553e-01 9.70656931e-01 -3.73430878e-01 -4.86369729e-02
8.91673714e-02 -1.96094722e-01 -1.10671532e+00 -4.48569268e-01
-2.49682084e-01 -2.60081559e-01 -4.95042443e-01 5.75813353e-01
7.81588197e-01 5.83254457e-01 3.16389412e-01 -1.79014885e+00
-4.21489209e-01 5.07488608e-01 2.36368254e-02 5.51161230e-01
7.35041559e-01 9.36939359e-01 1.03741848e+00 -1.51942134e-01
4.95369017e-01 1.46395230e+00 2.08770916e-01 7.58178473e-01
-1.18523228e+00 -4.69010383e-01 -6.16339967e-02 4.23917294e-01
-1.13632369e+00 -3.05724472e-01 3.94638866e-01 -1.29652387e-02
1.39584756e+00 4.15164173e-01 6.77321970e-01 1.33712971e+00
6.18367612e-01 1.12890089e+00 1.46591973e+00 -1.93807870e-01
6.41694009e-01 -5.42511523e-01 -1.29566416e-01 8.62003446e-01
-2.71838874e-01 6.32855833e-01 -5.81525505e-01 -1.79024681e-01
6.91074371e-01 -3.25703293e-01 2.14360788e-01 -5.23187220e-01
-1.32321870e+00 7.61276186e-01 5.68374135e-02 1.26981050e-01
-4.64357525e-01 7.73231089e-01 2.39196748e-01 3.99452865e-01
-3.50244269e-02 5.44393301e-01 -5.97321749e-01 -1.02350569e+00
-3.71967942e-01 8.24207902e-01 1.06297362e+00 7.96012580e-01
8.50559175e-01 2.33947918e-01 -2.81723648e-01 4.84242141e-01
6.77015007e-01 3.94461095e-01 4.30116445e-01 -1.68417335e+00
3.45480181e-02 2.42463768e-01 3.21754664e-01 -1.31263703e-01
-5.80476999e-01 -1.46478653e-01 -1.64101094e-01 8.04465890e-01
3.66554141e-01 -5.42004347e-01 -5.16354680e-01 1.49508846e+00
6.23778641e-01 1.23364143e-01 3.93807262e-01 7.55547643e-01
5.21649659e-01 4.99430329e-01 5.41139603e-01 1.07383013e-01
1.04927516e+00 -1.28077054e+00 -5.39995849e-01 -2.32383937e-01
3.96619618e-01 -6.73883021e-01 1.28812790e+00 5.21724164e-01
-1.25919902e+00 -1.94119379e-01 -1.09804273e+00 2.57435054e-01
-6.41949534e-01 -5.52094698e-01 9.88677144e-01 6.28481627e-01
-1.24189460e+00 1.00366592e+00 -1.28727198e+00 -2.03054562e-01
2.43408263e-01 4.80490535e-01 5.46947345e-02 2.02576786e-01
-1.24241436e+00 1.10177672e+00 6.94865644e-01 -4.65212822e-01
-1.54461849e+00 -5.27676523e-01 -7.70007193e-01 4.42298912e-02
7.39431798e-01 -8.22883010e-01 2.08776569e+00 -6.39107466e-01
-2.37396598e+00 3.97000432e-01 3.32365483e-01 -1.57573715e-01
5.87906480e-01 -2.87067175e-01 -4.02699620e-01 1.28420785e-01
-1.19370408e-01 7.09006608e-01 5.60733795e-01 -9.33715224e-01
-7.60666013e-01 -1.14916436e-01 4.83712941e-01 6.69094086e-01
2.73221821e-01 7.13289157e-02 -1.36747062e-01 -4.89827126e-01
-7.72831798e-01 -8.19869578e-01 -5.49579322e-01 -3.66725177e-01
-1.47627160e-01 -4.43189621e-01 5.62959731e-01 -2.73132980e-01
1.13528955e+00 -1.80317891e+00 3.22242260e-01 -6.62322491e-02
-1.96439158e-02 -1.29221767e-01 -3.25601697e-01 1.00452721e+00
6.05661422e-02 1.38533384e-01 -2.44400755e-01 -2.10191220e-01
7.24222541e-01 5.97753525e-01 4.14802343e-01 2.41723269e-01
2.09902540e-01 8.15242648e-01 -1.31270909e+00 -5.06568015e-01
5.52171350e-01 -6.03613704e-02 -6.71790183e-01 4.38686579e-01
-6.74633622e-01 6.34980738e-01 -7.25060165e-01 9.65530515e-01
2.63506502e-01 -2.79058278e-01 3.63121480e-01 8.65403712e-01
-3.29334319e-01 4.55299526e-01 -1.55000758e+00 2.13647580e+00
-3.20583761e-01 -7.87708238e-02 -5.80450445e-02 -7.04386294e-01
2.85125285e-01 4.55606133e-01 7.64322937e-01 -4.00912344e-01
1.89774767e-01 -4.44502234e-02 -6.31633550e-02 -4.27600384e-01
4.81850833e-01 2.02711180e-01 -1.59145147e-01 6.82274580e-01
4.89229858e-01 -4.78379428e-01 1.05601704e+00 -1.31617159e-01
1.38747776e+00 9.16584134e-01 4.69851643e-01 -4.25971150e-01
4.75893766e-01 1.86189562e-01 5.66574872e-01 8.68434787e-01
-5.90408683e-01 -1.61206484e-01 4.84933376e-01 -2.46682167e-01
-6.27732694e-01 -1.61419356e+00 6.26668567e-03 1.69414115e+00
4.02870812e-02 -6.73195899e-01 -6.53857470e-01 -6.21925414e-01
9.84547138e-02 8.54990423e-01 -5.20884752e-01 3.40171531e-02
-2.65565049e-02 -8.31349015e-01 6.00731730e-01 2.93309063e-01
4.88946259e-01 -1.59285355e+00 -7.75538981e-01 3.65835786e-01
1.27486158e-02 -5.68117499e-01 -9.64685977e-02 3.43400776e-01
-7.25346327e-01 -1.35639977e+00 -1.77231640e-01 -5.37857413e-01
-5.56042120e-02 -1.57505333e-01 1.48700631e+00 4.75845598e-02
-3.49710405e-01 9.90860581e-01 -3.69104981e-01 -3.81842673e-01
-7.20303297e-01 2.37053975e-01 -1.32292643e-01 -7.12104380e-01
2.71530569e-01 -6.40271544e-01 -5.70797384e-01 3.53266060e-01
-8.99693847e-01 -5.18073700e-02 1.15480676e-01 5.36530137e-01
4.98094559e-01 -1.61077663e-01 4.80879635e-01 -1.23153798e-01
9.06113088e-01 -5.09104133e-01 -9.67849076e-01 1.93576202e-01
-7.81463265e-01 3.22606295e-01 3.29724610e-01 -1.90166265e-01
-8.64264429e-01 -1.89538822e-01 -4.45962936e-01 4.78496701e-02
-6.96033716e-01 4.29084271e-01 1.04815044e-01 4.29679416e-02
3.62095386e-01 1.11349873e-01 4.52249981e-02 -2.39048883e-01
3.69140953e-01 3.35071862e-01 -2.95730270e-02 -9.95095015e-01
4.81809556e-01 -2.84770839e-02 1.23430248e-02 -4.20150965e-01
-2.29021952e-01 -3.07427108e-01 -5.72796986e-02 -5.82910597e-01
8.66253197e-01 -5.46044290e-01 -1.45570230e+00 6.55417085e-01
-4.31392610e-01 -1.25793016e+00 -5.34834862e-01 4.94718343e-01
-1.24239802e+00 1.01780377e-01 -7.49709487e-01 -6.26520157e-01
-3.38970840e-01 -1.47314203e+00 8.00195754e-01 7.25249171e-01
-1.81739658e-01 -1.15081835e+00 8.01931024e-01 2.13629872e-01
4.21995759e-01 9.68681276e-02 5.72882295e-01 -5.17258048e-01
-4.57407147e-01 3.92168343e-01 6.63044572e-01 1.86848775e-01
-1.17677569e-01 5.46800196e-01 -8.42714667e-01 -3.30370754e-01
-6.68517709e-01 -8.90951574e-01 2.20058978e-01 4.29718524e-01
1.06099176e+00 -4.56886962e-02 2.31968164e-02 2.94099748e-01
1.46815288e+00 2.89998144e-01 5.57723165e-01 9.26982045e-01
-1.33439198e-01 1.17496401e-01 7.89163291e-01 8.66579771e-01
8.32805037e-01 3.30586910e-01 8.24478745e-01 1.66159809e-01
5.61123133e-01 2.12445688e-02 6.69526637e-01 7.46921420e-01
-3.04006487e-01 -3.04013431e-01 -1.28523338e+00 1.70777112e-01
-2.15016365e+00 -1.18563509e+00 1.74197420e-01 1.85654712e+00
9.64891732e-01 -8.52520317e-02 4.53813106e-01 1.69419274e-02
4.01275307e-01 1.85945019e-01 -7.06389427e-01 -9.13163066e-01
1.57726213e-01 7.10770965e-01 2.85814673e-01 6.93972647e-01
-1.20258760e+00 1.02721786e+00 7.61552477e+00 9.05295908e-01
-6.22323155e-01 1.87703580e-01 2.43559390e-01 -2.19771832e-01
1.32773697e-01 1.85896941e-02 -6.04353189e-01 1.73166513e-01
1.41998482e+00 -7.55820513e-01 9.88083243e-01 9.68889952e-01
5.00929356e-01 -5.25750458e-01 -9.09345210e-01 2.03873634e-01
-6.86841369e-01 -1.28898692e+00 -3.32570732e-01 -2.64999662e-02
5.39791167e-01 4.24229562e-01 9.61282998e-02 8.78801882e-01
1.30661917e+00 -9.22070622e-01 6.62589312e-01 4.37175810e-01
1.74788371e-01 -1.03604650e+00 3.36309761e-01 2.23236412e-01
-1.00506032e+00 -1.41357884e-01 -5.80137596e-02 -3.39335859e-01
-1.41324937e-01 7.95362110e-04 -4.89019215e-01 8.17177117e-01
1.06617081e+00 7.11535215e-01 -5.11590898e-01 1.24461663e+00
-5.33192337e-01 5.97361863e-01 -2.69231290e-01 -1.45303339e-01
4.91847694e-01 -4.50063378e-01 5.38444102e-01 1.02165866e+00
-1.15734646e-02 -2.14527518e-01 4.57230061e-01 6.82049811e-01
3.57309163e-01 1.19601503e-01 -4.45408970e-01 -1.67195961e-01
2.26513207e-01 1.48170161e+00 -5.55051029e-01 -2.18473285e-01
-3.39816093e-01 7.10864604e-01 3.32670867e-01 3.54882538e-01
-1.35660052e+00 -2.59962052e-01 1.26663888e+00 -5.82293749e-01
1.88772589e-01 -5.54037392e-01 1.86223030e-01 -8.24004710e-01
-4.64562356e-01 -1.33932066e+00 7.65533030e-01 -8.94671559e-01
-1.09539104e+00 1.44462243e-01 1.06232077e-01 -1.01818216e+00
-3.50656629e-01 -6.02412343e-01 -5.76329350e-01 5.10997832e-01
-1.60635459e+00 -7.04751849e-01 -1.17010795e-01 8.07697058e-01
6.11156344e-01 9.69760120e-02 1.22703910e+00 6.76951781e-02
-1.05898821e+00 6.15262315e-02 3.37549567e-01 -3.51473838e-01
5.12056708e-01 -1.84132850e+00 3.83093804e-01 5.41928649e-01
-2.53771633e-01 1.30699396e-01 7.81907737e-01 -2.86499172e-01
-1.77533078e+00 -6.58894598e-01 -1.51228989e-02 -3.19969624e-01
1.08306456e+00 4.00127769e-02 -5.60114980e-01 7.77684748e-01
7.63476670e-01 -2.62417167e-01 7.15303838e-01 -1.30734414e-01
2.42427677e-01 3.26409712e-02 -9.81164515e-01 9.59460795e-01
8.12797248e-01 -1.82920933e-01 -5.69238961e-01 4.17015582e-01
2.39757091e-01 -6.65950835e-01 -1.48230827e+00 2.34643608e-01
2.85932124e-01 -9.28358316e-01 9.51111197e-01 -6.36938155e-01
5.06102681e-01 -2.80787259e-01 -1.27388472e-02 -1.74596524e+00
-3.30342740e-01 -1.17843664e+00 -5.83315253e-01 8.05193841e-01
5.23731411e-01 -7.84237206e-01 5.32838702e-01 3.63215148e-01
-1.44882202e-01 -6.60732567e-01 -1.11322832e+00 -9.23515081e-01
3.93192649e-01 -3.62125456e-01 6.85833216e-01 6.02082670e-01
2.64408141e-01 -7.54788071e-02 3.79972439e-03 2.30997562e-01
4.85424072e-01 -9.18854550e-02 6.72098875e-01 -7.62058675e-01
-6.76924050e-01 -6.02494776e-01 -6.40526265e-02 -6.93804979e-01
-5.19127324e-02 -9.44679379e-01 1.44126490e-01 -1.60385919e+00
-2.47038543e-01 -3.90800595e-01 -6.26120627e-01 7.40438342e-01
-2.05673322e-01 -7.92547986e-02 2.08423465e-01 -9.89719555e-02
-1.24990797e+00 1.01056719e+00 1.45331562e+00 7.97633007e-02
-2.21892804e-01 8.80284309e-02 -2.62498051e-01 6.57834589e-01
1.74840748e+00 -7.43991435e-01 -2.32045472e-01 -2.22693607e-01
1.44672215e-01 9.42444652e-02 2.26611987e-01 -1.46338451e+00
-1.59409747e-01 -8.74171317e-01 1.07303962e-01 -2.13313475e-01
3.15968573e-01 -5.14144480e-01 1.02587029e-01 8.17009270e-01
-3.45370173e-01 8.02132368e-01 5.63703656e-01 3.41128200e-01
8.66080672e-02 -2.63358027e-01 6.93605781e-01 -3.74997586e-01
-9.72743750e-01 2.57188767e-01 -1.24133658e+00 4.39951599e-01
1.58133960e+00 1.86624989e-01 -4.66903329e-01 -2.13661268e-01
-8.66353989e-01 7.67665923e-01 4.79618728e-01 5.84491014e-01
3.69035840e-01 -1.02667415e+00 -4.91658360e-01 -1.59479707e-01
1.39347211e-01 -5.31363487e-01 1.25612751e-01 7.96322286e-01
-8.22439611e-01 8.50005150e-02 -7.09994376e-01 -3.33921015e-01
-9.40469444e-01 5.07666051e-01 7.13419259e-01 -7.45378256e-01
-4.88185287e-01 2.30487898e-01 -6.79676294e-01 -7.69152582e-01
4.49110903e-02 -3.25830936e-01 7.40974471e-02 -4.57376987e-01
3.53428990e-01 8.33912909e-01 -9.41070393e-02 1.90456718e-01
-3.66970390e-01 -7.36436844e-02 4.37831819e-01 -4.36700523e-01
1.48250449e+00 1.60570275e-02 1.06817953e-01 7.08415091e-01
1.50970995e-01 -5.65453410e-01 -1.61270976e+00 3.11641186e-01
2.51677245e-01 2.61709727e-02 1.37264386e-01 -1.05027378e+00
-7.61529326e-01 6.04839087e-01 7.03178823e-01 1.21439934e-01
8.03156257e-01 -4.88400400e-01 1.97534382e-01 5.93320906e-01
6.24425590e-01 -1.49659240e+00 2.38204584e-01 7.46631742e-01
8.10577810e-01 -9.90917146e-01 -1.29905026e-02 2.14883909e-01
-8.09721589e-01 1.09110236e+00 9.05653596e-01 -2.07913801e-01
6.50206923e-01 4.70873088e-01 1.51190460e-01 -1.52463377e-01
-1.25266147e+00 -6.57431006e-01 -3.28559697e-01 7.83281267e-01
3.11154157e-01 2.36157894e-01 -2.42529422e-01 1.77838340e-01
-3.30489017e-02 3.82635444e-01 6.43629134e-01 1.63240075e+00
-2.33908281e-01 -1.47513580e+00 -3.74100268e-01 9.98444017e-03
-5.55791557e-01 4.83457185e-02 5.61947823e-02 9.05750334e-01
-2.62275130e-01 1.25472140e+00 -1.18659109e-01 -5.98717108e-02
1.46588191e-01 2.65304476e-01 6.38190567e-01 -3.91101062e-01
-1.04549205e+00 -2.92516887e-01 3.05223733e-01 -1.02124298e+00
-4.34145421e-01 -6.61680341e-01 -1.37818968e+00 -6.67845130e-01
3.02938104e-01 2.58143723e-01 6.56772673e-01 6.05857432e-01
4.04332012e-01 8.28662336e-01 4.41697538e-01 -9.80397105e-01
-6.91191733e-01 -7.58409619e-01 -6.13585830e-01 -7.68658966e-02
-1.30812392e-01 -8.40582728e-01 -1.01612344e-01 -3.62094432e-01] | [3.9273064136505127, 1.4718669652938843] |
2cafd96c-188a-4c03-b026-0d0b27d54930 | edge-but-not-least-cross-view-graph-pooling | 2109.11796 | null | https://arxiv.org/abs/2109.11796v1 | https://arxiv.org/pdf/2109.11796v1.pdf | Edge but not Least: Cross-View Graph Pooling | Graph neural networks have emerged as a powerful model for graph representation learning to undertake graph-level prediction tasks. Various graph pooling methods have been developed to coarsen an input graph into a succinct graph-level representation through aggregating node embeddings obtained via graph convolution. However, most graph pooling methods are heavily node-centric and are unable to fully leverage the crucial information contained in global graph structure. This paper presents a cross-view graph pooling (Co-Pooling) method to better exploit crucial graph structure information. The proposed Co-Pooling fuses pooled representations learnt from both node view and edge view. Through cross-view interaction, edge-view pooling and node-view pooling seamlessly reinforce each other to learn more informative graph-level representations. Co-Pooling has the advantage of handling various graphs with different types of node attributes. Extensive experiments on a total of 15 graph benchmark datasets validate the effectiveness of our proposed method, demonstrating its superior performance over state-of-the-art pooling methods on both graph classification and graph regression tasks. | ['Ivor W. Tsang', 'Jie Yin', 'Xiaowei Zhou'] | 2021-09-24 | null | null | null | null | ['graph-regression'] | ['graphs'] | [-1.07998222e-01 3.47147614e-01 -4.49182421e-01 -3.15193206e-01
-1.43869847e-01 -4.29840982e-01 6.65912032e-01 5.86633027e-01
1.10206723e-01 3.64812702e-01 4.52579856e-01 -1.31410509e-01
-1.11847915e-01 -1.28171957e+00 -5.84201515e-01 -6.17491126e-01
-4.01446760e-01 -1.90640256e-01 7.76625797e-02 -1.51432052e-01
-1.29072601e-02 5.60220301e-01 -9.71445739e-01 2.41348594e-01
6.07149005e-01 6.23034179e-01 1.30619958e-01 6.26511037e-01
-2.60641873e-01 8.62076223e-01 -2.64406651e-01 -5.02910435e-01
2.12261513e-01 5.39796650e-02 -5.49020886e-01 4.17106450e-02
3.59277427e-01 -7.49356579e-03 -9.06454682e-01 1.13924718e+00
4.21975374e-01 7.37833306e-02 3.31920087e-01 -1.36208832e+00
-1.24244797e+00 7.91687131e-01 -7.72868395e-01 2.11354479e-01
3.17702293e-01 -2.34656055e-02 1.63741958e+00 -8.00153017e-01
7.44569957e-01 1.28433371e+00 7.34220207e-01 2.82705188e-01
-1.26421773e+00 -4.69532371e-01 6.99585855e-01 -4.78357822e-03
-1.13801658e+00 2.78756410e-01 1.34013557e+00 -2.87498891e-01
1.39705765e+00 1.27727583e-01 8.28025579e-01 9.06147063e-01
3.63862514e-01 7.15661466e-01 8.93905222e-01 -7.28105828e-02
-2.90379703e-01 -7.08956867e-02 4.54734445e-01 1.15808725e+00
6.17791474e-01 -2.77899742e-01 -4.99615341e-01 -1.03455916e-01
7.80866921e-01 4.12941456e-01 -4.59177732e-01 -5.46415448e-01
-1.13641608e+00 9.23422694e-01 1.39912927e+00 3.66620094e-01
-5.48315883e-01 5.23210287e-01 5.75940609e-01 2.38190785e-01
7.72670865e-01 4.49524432e-01 -1.69609144e-01 5.31047165e-01
-4.55313802e-01 -9.15317684e-02 6.11447275e-01 1.00320709e+00
1.04148853e+00 1.57452136e-01 -4.95196193e-01 5.90438902e-01
3.49928856e-01 -7.82030001e-02 2.30393142e-01 3.88574600e-02
8.21626961e-01 1.53944600e+00 -5.84126234e-01 -1.52986383e+00
-5.61765552e-01 -6.09258473e-01 -1.10760570e+00 -1.95129234e-02
2.58544553e-02 8.02105144e-02 -1.17384517e+00 1.75164413e+00
-2.66191438e-02 3.47061872e-01 -5.10520712e-02 5.16230464e-01
1.47229147e+00 5.36347151e-01 1.63790971e-01 3.95048380e-01
1.34478354e+00 -1.24184167e+00 -5.35140216e-01 -2.95259565e-01
6.50709629e-01 -1.61885053e-01 9.52953517e-01 -1.74141198e-01
-8.70125055e-01 -4.91640121e-01 -1.30847371e+00 -1.27398998e-01
-8.83403301e-01 -1.75203592e-01 1.01493776e+00 5.30673563e-01
-1.25489223e+00 7.24427581e-01 -7.92460442e-01 -4.55573350e-01
7.31236994e-01 3.05168152e-01 -1.00018549e+00 -3.14939648e-01
-1.06433046e+00 4.97999936e-01 5.87135375e-01 1.76718920e-01
-5.31510055e-01 -7.96283126e-01 -1.25187349e+00 6.30259216e-01
4.14274842e-01 -8.41017365e-01 4.08667088e-01 -4.57142651e-01
-9.91601586e-01 8.02079856e-01 3.04313432e-02 -3.74487698e-01
-5.43994345e-02 7.06635863e-02 -3.76932114e-01 2.02770412e-01
-1.83133677e-01 4.98074919e-01 6.32161319e-01 -1.15683401e+00
4.63276589e-03 -5.12500525e-01 2.98538834e-01 2.77660519e-01
-5.64830184e-01 -3.73837858e-01 -6.30189896e-01 -7.63436735e-01
2.25597814e-01 -5.57902217e-01 -2.29126856e-01 -2.00376540e-01
-6.27219379e-01 -3.67309272e-01 9.18712795e-01 -5.73554456e-01
1.39352393e+00 -1.89116752e+00 3.11038285e-01 1.54618919e-01
1.02614701e+00 2.54309237e-01 -3.54907811e-01 1.05642080e+00
-2.88284004e-01 4.99093115e-01 9.22155380e-02 -3.97426873e-01
1.19643593e-02 2.55359620e-01 5.24378829e-02 3.87919337e-01
5.65758288e-01 1.44939208e+00 -1.04017258e+00 -2.58020282e-01
2.33417928e-01 6.76205635e-01 -7.00678647e-01 3.41539353e-01
1.57749420e-03 9.97423008e-02 -6.01823688e-01 6.24393880e-01
7.88212240e-01 -7.28681862e-01 4.10227746e-01 -3.23804289e-01
5.02723634e-01 4.19642366e-02 -8.49576592e-01 1.81069422e+00
-3.67638290e-01 3.30153048e-01 3.47239636e-02 -1.11238110e+00
1.10948157e+00 6.95808753e-02 2.45495647e-01 -2.88509727e-01
8.18143785e-02 -2.19948038e-01 -2.22281646e-02 -2.44090095e-01
4.61522132e-01 9.39139202e-02 -1.31314144e-01 3.98365557e-01
7.05485463e-01 2.49276608e-01 1.50978208e-01 6.05005383e-01
1.47190869e+00 -5.46572246e-02 5.69771528e-01 -3.28505516e-01
5.91140270e-01 -6.77917838e-01 3.52938473e-01 6.10997736e-01
-1.81277618e-01 6.30505502e-01 1.01206708e+00 -7.20593929e-01
-4.98888075e-01 -1.04878592e+00 5.86339533e-01 1.01081407e+00
1.16444737e-01 -9.36369240e-01 -4.04442608e-01 -1.08020127e+00
1.49984583e-01 1.38031170e-01 -1.06749189e+00 -4.50819969e-01
-5.47267675e-01 -8.26083660e-01 3.29573691e-01 7.97427893e-01
4.87299532e-01 -1.02905941e+00 -3.95032354e-02 1.78515002e-01
2.38688737e-01 -1.31683338e+00 -5.90957999e-01 7.77786449e-02
-9.34285939e-01 -1.26054490e+00 -4.84100133e-01 -9.25274730e-01
9.28039432e-01 6.40803337e-01 1.34132171e+00 6.01875901e-01
-3.29014540e-01 4.47344631e-01 -3.60821098e-01 3.09757907e-02
-4.51361239e-02 2.93230265e-01 -4.37101156e-01 1.82425454e-01
2.01010257e-01 -9.96038556e-01 -4.89860147e-01 -1.74003586e-01
-7.34207273e-01 1.43608779e-01 6.49382651e-01 9.68747318e-01
5.05835772e-01 1.56209450e-02 5.60563564e-01 -1.26025939e+00
1.15810490e+00 -5.94381988e-01 -3.91117215e-01 5.32699525e-01
-6.75263524e-01 2.84864724e-01 7.63296485e-01 9.35306773e-03
-5.97774565e-01 -3.14740568e-01 2.63472110e-01 -5.56160510e-01
1.44025773e-01 9.75632906e-01 -3.63485217e-01 -2.15545326e-01
2.29749545e-01 2.03066319e-01 5.12681603e-02 -3.87051523e-01
7.25452542e-01 -9.44383722e-03 2.77661115e-01 -3.09641123e-01
7.86039591e-01 2.43046969e-01 2.82566011e-01 -6.05329275e-01
-4.86218274e-01 -5.15071809e-01 -6.75221860e-01 -1.79578394e-01
9.87218142e-01 -9.58013654e-01 -6.64302111e-01 5.20022035e-01
-9.29714918e-01 -6.89509735e-02 8.79075751e-02 1.39894381e-01
-1.30354017e-01 5.22842705e-01 -6.49762392e-01 -3.84707421e-01
-5.35304189e-01 -1.07599485e+00 1.02379417e+00 3.40963453e-01
4.20212597e-01 -1.49993289e+00 -6.11867420e-02 -2.19736826e-02
4.18026686e-01 6.93880558e-01 1.07737255e+00 -8.22502673e-01
-7.17935205e-01 -6.40614331e-01 -8.03887963e-01 1.00370795e-01
3.66094232e-01 -1.38655737e-01 -8.33536506e-01 -5.52169383e-01
-8.04983735e-01 -1.26169756e-01 1.15192568e+00 1.78246602e-01
1.16243958e+00 -2.61547834e-01 -4.47560400e-01 7.03060210e-01
1.72154939e+00 -5.10369897e-01 4.74317908e-01 -9.42374617e-02
1.38030982e+00 3.13357085e-01 -1.10184744e-01 8.72661769e-02
5.69378376e-01 3.78659964e-01 7.93290436e-01 -2.71810949e-01
-3.52086544e-01 -7.28764534e-01 1.24268159e-01 9.41860080e-01
-1.34000912e-01 -5.66739500e-01 -7.24078834e-01 4.53255564e-01
-1.88187861e+00 -8.45792949e-01 -1.54837832e-01 1.73231578e+00
8.99747089e-02 2.67822981e-01 1.34686574e-01 -1.85431495e-01
8.18642855e-01 8.81721735e-01 -3.53420734e-01 -4.70494121e-01
2.01388774e-03 2.23617584e-01 6.52471066e-01 3.36109489e-01
-1.15624177e+00 1.03828990e+00 5.62852383e+00 3.92524451e-01
-9.45396006e-01 -9.36922878e-02 3.31067473e-01 4.82903630e-01
-5.88466406e-01 -4.86607514e-02 -2.44090930e-01 5.72652444e-02
5.48908770e-01 -2.94630468e-01 4.23427612e-01 8.76614988e-01
-4.03010696e-01 3.58036906e-01 -1.06472230e+00 9.27832365e-01
4.99352366e-02 -1.63580966e+00 3.80101144e-01 2.74909735e-01
6.74025059e-01 1.70261353e-01 -1.09163947e-01 5.11089921e-01
5.18024206e-01 -1.38004291e+00 -2.12873872e-02 5.38018227e-01
3.80457073e-01 -5.57413757e-01 8.60570848e-01 -3.20970118e-02
-1.85644054e+00 5.95174916e-02 -3.14860851e-01 -2.03966171e-01
2.66850237e-02 3.59516472e-01 -9.16125715e-01 1.36279762e+00
6.58370554e-01 1.31170213e+00 -9.70074475e-01 6.09471023e-01
-3.71541709e-01 3.59613538e-01 5.17380759e-02 -9.30780694e-02
5.77840805e-01 -2.63907343e-01 4.93668050e-01 1.14662528e+00
2.37615071e-02 -1.56918585e-01 3.56944323e-01 1.13384795e+00
-5.22528052e-01 1.40006334e-01 -9.83982921e-01 -6.47183716e-01
3.37620676e-01 1.66038632e+00 -8.89823854e-01 -1.97211087e-01
-8.81886661e-01 7.81476557e-01 9.94368792e-01 4.41862077e-01
-5.46946704e-01 -6.13075614e-01 7.43849099e-01 -7.75826797e-02
5.75258434e-01 -3.35953206e-01 -7.66092539e-02 -1.27213526e+00
-5.91872185e-02 -3.97466749e-01 7.12829888e-01 -6.22456193e-01
-1.57754636e+00 7.99607217e-01 3.30255218e-02 -7.82934666e-01
2.11364418e-01 -9.76050794e-01 -1.12466228e+00 9.97622252e-01
-1.58469570e+00 -1.82766593e+00 -6.84840858e-01 6.23292565e-01
5.23278676e-02 -2.55598307e-01 8.57030153e-01 7.59111866e-02
-6.46180451e-01 7.73522377e-01 -4.68921810e-01 3.64639580e-01
1.58083379e-01 -1.59036577e+00 7.89396286e-01 9.57684636e-01
3.63600522e-01 8.15420032e-01 1.65604874e-01 -7.08523989e-01
-1.74275470e+00 -1.25926161e+00 4.62581456e-01 -2.24214077e-01
7.49375343e-01 -7.45793104e-01 -1.07504547e+00 8.81486535e-01
3.40425402e-01 7.20614910e-01 6.30415320e-01 3.15063655e-01
-7.70331383e-01 -1.20291054e-01 -1.03237295e+00 5.77687204e-01
1.16146910e+00 -7.56684363e-01 -3.13868016e-01 8.45170543e-02
1.00623727e+00 -1.98000759e-01 -1.24494207e+00 3.87732178e-01
3.74002188e-01 -8.53408396e-01 1.02699864e+00 -9.41782415e-01
2.92017996e-01 -1.58099577e-01 -8.58837664e-02 -1.52239132e+00
-6.64736629e-01 -6.07723117e-01 -1.76976383e-01 1.22083914e+00
3.40729713e-01 -1.03346181e+00 8.37812006e-01 2.64729820e-02
-2.35345915e-01 -1.02941847e+00 -5.60515165e-01 -5.10479987e-01
-9.35242400e-02 2.74407342e-02 9.18822527e-01 1.05283415e+00
2.55513877e-01 4.61117268e-01 -2.75284976e-01 3.84571850e-01
4.56148267e-01 2.82772690e-01 9.33673143e-01 -1.22893000e+00
-9.51844379e-02 -7.47662306e-01 -1.10405588e+00 -5.54521978e-01
2.87764072e-01 -1.66231847e+00 -5.04900992e-01 -2.10776782e+00
3.22274208e-01 4.64917235e-02 -7.64332414e-01 6.02813959e-01
-6.07168972e-01 1.47743568e-01 3.21985930e-01 -3.60833585e-01
-3.79170895e-01 5.83884776e-01 1.51221168e+00 -2.79499352e-01
-9.99558121e-02 -4.11732882e-01 -1.00760615e+00 3.71399939e-01
7.75378764e-01 -1.55897483e-01 -6.00298047e-01 -3.60733539e-01
1.89413235e-01 -1.40280738e-01 4.42775339e-01 -8.12390745e-01
1.68554217e-01 1.56198189e-01 1.83586627e-01 -3.58807057e-01
2.17533797e-01 -6.64527118e-01 3.40979025e-02 4.80442524e-01
-1.64549887e-01 3.46165061e-01 2.05241829e-01 1.09324491e+00
-3.08140129e-01 2.75545448e-01 4.05274123e-01 -3.90839130e-01
-8.12454462e-01 7.73477435e-01 3.01043153e-01 -2.61588469e-02
9.06775057e-01 -3.45825046e-01 -5.29322207e-01 -4.13556576e-01
-8.86122763e-01 3.06312710e-01 3.35239530e-01 7.00900435e-01
6.86677039e-01 -1.56803429e+00 -6.45445585e-01 3.85757655e-01
5.03784478e-01 -2.73759782e-01 2.83636957e-01 8.27418327e-01
-4.03829455e-01 2.12737441e-01 -3.16818178e-01 -3.79867762e-01
-1.16376340e+00 8.28971267e-01 2.78512269e-01 -9.29064929e-01
-9.78476405e-01 1.12071550e+00 4.26487088e-01 -5.75017631e-01
3.23310383e-02 -6.79183960e-01 -2.51213044e-01 -6.52611256e-02
2.12981343e-01 1.32792294e-01 1.79874506e-02 -5.84372938e-01
-3.62766743e-01 4.32168007e-01 -2.18607724e-01 6.16210341e-01
1.65253401e+00 1.81060567e-01 -3.40414464e-01 2.45832995e-01
1.39400756e+00 -1.24767840e-01 -1.06752777e+00 -3.45066845e-01
-7.90409520e-02 -4.27288681e-01 9.49634910e-02 -3.70220155e-01
-1.43695426e+00 1.15677261e+00 -1.84303410e-02 5.27949095e-01
1.03964984e+00 5.08586355e-02 5.75006187e-01 1.63057134e-01
1.92186281e-01 -5.32666624e-01 3.34225476e-01 1.11769959e-01
1.12466431e+00 -1.25847471e+00 2.32655734e-01 -7.95229793e-01
-4.07384634e-01 1.27353418e+00 7.93385625e-01 -7.07420051e-01
1.01951909e+00 -2.17772767e-01 -5.16948462e-01 -7.58665264e-01
-7.49728620e-01 -2.64269799e-01 6.66624188e-01 5.96221864e-01
4.15413857e-01 3.70784372e-01 8.85865837e-03 7.17775643e-01
2.16741055e-01 -4.60909694e-01 4.39684480e-01 7.95150161e-01
-1.41358942e-01 -1.03041875e+00 3.83705497e-01 6.26450479e-01
-2.41823629e-01 -1.46783754e-01 -6.77266598e-01 1.11191332e+00
-3.27115953e-01 6.46608651e-01 -8.13276544e-02 -7.19108939e-01
2.14824453e-01 -6.67245463e-02 4.80881900e-01 -8.30799639e-01
-7.18037188e-01 -3.03890228e-01 5.97388949e-03 -7.50735521e-01
-2.81120598e-01 -6.01202175e-02 -1.01199758e+00 -3.74033839e-01
-3.21144611e-01 -2.25208730e-01 2.16992497e-01 5.59719265e-01
6.58018947e-01 1.01099718e+00 4.59252924e-01 -9.82957900e-01
-1.95431143e-01 -9.83626425e-01 -7.56620705e-01 4.49577183e-01
2.74786592e-01 -7.56160676e-01 -1.98212847e-01 -5.10113060e-01] | [7.091124057769775, 6.290483474731445] |
0c400778-60d3-4751-80eb-615ccbf8531b | chinese-paragraph-level-discourse-parsing | null | null | https://aclanthology.org/2020.coling-main.506 | https://aclanthology.org/2020.coling-main.506.pdf | Chinese Paragraph-level Discourse Parsing with Global Backward and Local Reverse Reading | Discourse structure tree construction is the fundamental task of discourse parsing and most previous work focused on English. Due to the cultural and linguistic differences, existing successful methods on English discourse parsing cannot be transformed into Chinese directly, especially in paragraph level suffering from longer discourse units and fewer explicit connectives. To alleviate the above issues, we propose two reading modes, i.e., the global backward reading and the local reverse reading, to construct Chinese paragraph level discourse trees. The former processes discourse units from the end to the beginning in a document to utilize the left-branching bias of discourse structure in Chinese, while the latter reverses the position of paragraphs in a discourse unit to enhance the differentiation of coherence between adjacent discourse units. The experimental results on Chinese MCDTB demonstrate that our model outperforms all strong baselines. | ['Qiaoming Zhu', 'Fang Kong', 'Peifeng Li', 'Xiaomin Chu', 'Feng Jiang'] | 2020-12-01 | null | null | null | coling-2020-8 | ['discourse-parsing'] | ['natural-language-processing'] | [ 3.54003042e-01 5.84500372e-01 -4.42660689e-01 -2.91345954e-01
-5.87266982e-01 -6.56810045e-01 5.19031703e-01 4.86023575e-02
-2.26413846e-01 9.30584848e-01 1.10737884e+00 -8.22777808e-01
6.69373512e-01 -9.56973076e-01 -2.58694172e-01 -6.84812307e-01
3.33703518e-01 6.18357696e-02 5.39187551e-01 -4.85598654e-01
4.13338691e-01 -3.77039522e-01 -7.94133425e-01 9.30439711e-01
1.15294480e+00 3.76151055e-01 5.92698157e-01 4.91213202e-01
-8.84912789e-01 1.45131779e+00 -9.99812603e-01 -1.25929847e-01
-3.05589139e-01 -9.93750930e-01 -1.28401852e+00 1.84154510e-01
-2.66638756e-01 -4.55030262e-01 -8.64250772e-03 1.02964973e+00
1.74990520e-01 -2.32345775e-01 2.81626821e-01 -5.23271620e-01
-7.39658833e-01 1.28291917e+00 -7.00205803e-01 3.46194118e-01
5.38138270e-01 -2.61535704e-01 1.08690894e+00 -3.92534822e-01
8.32311988e-01 1.48083925e+00 1.69717461e-01 7.48010218e-01
-7.04874694e-01 -3.99706155e-01 7.84299254e-01 2.99376100e-02
-6.54906571e-01 -3.32964540e-01 7.04834461e-01 -4.93059635e-01
9.86729205e-01 3.70501459e-01 3.79962623e-01 9.47521925e-01
-7.55400658e-02 8.89893174e-01 1.00711918e+00 -7.19331801e-01
-6.07567653e-02 -2.31946960e-01 4.88350630e-01 2.63598949e-01
-6.41435757e-02 -4.13429022e-01 -1.88809440e-01 2.78198332e-01
4.17967707e-01 -6.54102087e-01 -2.00248107e-01 6.60304666e-01
-1.17151976e+00 9.36065316e-01 3.34107488e-01 6.70189917e-01
5.48758432e-02 -2.18072742e-01 7.37172723e-01 1.30785212e-01
5.33435524e-01 3.45981717e-01 -1.52099907e-01 -2.44887069e-01
-5.79822779e-01 1.82217658e-01 8.76218438e-01 1.17324901e+00
3.04964960e-01 -2.34945506e-01 -4.04653788e-01 5.69846332e-01
3.79232287e-01 1.20299354e-01 4.78270859e-01 -9.30232823e-01
1.27287149e+00 7.89040685e-01 -1.91333398e-01 -8.52372646e-01
-1.12546340e-01 1.15395272e-02 -7.01432049e-01 -3.68308395e-01
4.46259975e-01 -7.65306294e-01 -3.55535239e-01 1.58325791e+00
4.14741457e-01 -4.15532470e-01 2.70311147e-01 7.36256719e-01
1.18286490e+00 1.15906155e+00 2.39479437e-01 -6.03076220e-01
1.69156790e+00 -1.37062228e+00 -1.17424214e+00 -4.05587345e-01
8.83420944e-01 -8.79377782e-01 9.93927002e-01 -1.59631863e-01
-1.33502638e+00 -3.59896600e-01 -1.13158631e+00 -5.61481714e-01
2.61954721e-02 8.61161724e-02 1.85046285e-01 4.47894454e-01
-5.28654993e-01 -4.92696390e-02 -8.89921188e-01 1.82248041e-01
3.34557712e-01 -3.03128004e-01 1.48197159e-01 1.58728957e-01
-1.37675273e+00 7.75870502e-01 7.68704772e-01 2.69621521e-01
-9.86886695e-02 -3.59730810e-01 -9.11332607e-01 6.78495020e-02
5.63328683e-01 -3.62763345e-01 1.62523997e+00 -1.02830863e+00
-1.74551070e+00 8.89466286e-01 -6.56376779e-01 -3.63240451e-01
6.79031730e-01 -4.93776798e-01 1.00489333e-02 -1.96230924e-03
1.85107470e-01 4.40762341e-01 3.25820297e-01 -1.15424907e+00
-1.04735267e+00 -1.61556333e-01 3.40903997e-01 5.89956641e-01
7.91948661e-03 4.60414350e-01 -5.35346866e-01 -6.05984628e-01
2.35081404e-01 -5.31224251e-01 -9.16398689e-02 -7.05390573e-01
-7.12962925e-01 -5.45180440e-01 1.04612851e+00 -1.07963645e+00
2.05525208e+00 -1.97642124e+00 3.51918697e-01 -3.44452769e-01
8.03058594e-02 1.63855940e-01 1.79500237e-01 6.34134352e-01
2.72890143e-02 4.02887464e-01 -2.17285633e-01 -4.96149696e-02
-1.81412846e-01 3.53952408e-01 -4.59944963e-01 -7.42049217e-02
3.75389606e-01 9.02822733e-01 -1.07999551e+00 -7.28566945e-01
-1.39678419e-01 -1.20630145e-01 -4.18233067e-01 2.50263870e-01
-6.53611541e-01 7.93178976e-01 -8.06117296e-01 3.75238776e-01
5.95641494e-01 -2.43160561e-01 8.16276908e-01 1.28693432e-01
-6.20735288e-01 1.37975037e+00 -7.91520655e-01 1.37819719e+00
-2.20627710e-01 7.20644832e-01 1.55307367e-01 -8.76378417e-01
5.47345281e-01 4.00709450e-01 -1.35208592e-01 -7.63018668e-01
2.55823165e-01 1.69268006e-03 4.78355974e-01 -7.17582822e-01
8.46173286e-01 -1.15818731e-01 -3.78841162e-01 3.87082160e-01
-6.72635078e-01 9.32195261e-02 5.27672172e-01 2.36509353e-01
7.30262876e-01 1.81728661e-01 5.24175048e-01 -3.00565004e-01
7.68467903e-01 3.27792257e-01 8.73253703e-01 1.48749992e-01
-1.45437270e-02 5.29053211e-01 1.16175449e+00 5.42016588e-02
-8.13221991e-01 -6.93331718e-01 2.53036246e-02 1.42012286e+00
2.80080199e-01 -6.56858563e-01 -1.26194882e+00 -7.03042924e-01
-5.68249464e-01 9.09733951e-01 -4.39495206e-01 4.49842274e-01
-1.77411985e+00 -6.62110984e-01 6.58667564e-01 6.65723860e-01
9.04280007e-01 -1.22303689e+00 -6.51994109e-01 4.45878446e-01
-8.00513327e-01 -1.19085491e+00 -4.69239473e-01 -2.47281045e-01
-6.41764402e-01 -1.07098961e+00 -4.50165689e-01 -1.35367596e+00
6.53397858e-01 2.16880471e-01 9.11280930e-01 5.73290408e-01
6.27605855e-01 -5.36934555e-01 -6.92402303e-01 -4.03633624e-01
-9.43677127e-01 3.98003250e-01 -7.64365196e-01 -6.97698057e-01
2.65359432e-01 8.43297616e-02 -4.42747504e-01 1.17910773e-01
-6.36259198e-01 6.47664309e-01 1.42106175e-01 8.39237392e-01
1.38010219e-01 -1.39674023e-01 6.28836215e-01 -1.40817833e+00
9.30980921e-01 -3.87385637e-01 -4.24558282e-01 4.22062933e-01
-4.63035665e-02 -7.42028654e-02 5.07046282e-01 -1.91609636e-01
-1.65427637e+00 -6.16603017e-01 -5.13330996e-01 7.71906555e-01
1.09345831e-01 8.91040802e-01 -5.68874061e-01 9.28939521e-01
2.50450879e-01 -1.65441688e-02 -2.03099430e-01 -3.37268382e-01
3.04544330e-01 7.02573717e-01 4.58927333e-01 -8.22281003e-01
3.92358780e-01 6.08446635e-02 -4.95294899e-01 -9.38434660e-01
-9.72602427e-01 -1.49317712e-01 -8.04923177e-01 -4.70261723e-02
1.49100268e+00 -9.58641291e-01 -4.23605680e-01 5.14937818e-01
-1.61549580e+00 -4.31285113e-01 -3.47057916e-02 5.79909906e-02
5.59516214e-02 7.20253110e-01 -1.31136870e+00 -6.01534545e-01
-3.30899626e-01 -9.85984862e-01 8.53928924e-01 5.62286258e-01
-4.77879971e-01 -9.92898226e-01 -3.51750553e-02 4.20509666e-01
1.09435871e-01 3.95577013e-01 1.39164376e+00 -3.50422323e-01
-5.47657013e-01 4.80590880e-01 -3.05904180e-01 2.52722472e-01
4.69437361e-01 2.91498840e-01 -6.63469553e-01 3.18068154e-02
2.38300681e-01 -6.89430684e-02 5.52712023e-01 2.37119451e-01
6.69277966e-01 -6.07835472e-01 -3.53850037e-01 4.01094109e-02
9.47556496e-01 4.95207608e-01 8.25143754e-01 6.66703284e-01
6.63375974e-01 8.27630639e-01 8.35261822e-01 -4.34745848e-02
8.04972231e-01 1.60737917e-01 1.42863141e-02 -4.64467481e-02
-2.50318527e-01 -3.96761268e-01 4.15442616e-01 1.63216913e+00
-1.05618648e-01 -3.84806246e-01 -9.97894645e-01 5.21406174e-01
-1.82352328e+00 -8.43857944e-01 -5.87206841e-01 1.64274740e+00
1.29019642e+00 4.52190906e-01 -9.66935530e-02 -6.59957901e-02
9.79900777e-01 6.12120509e-01 -1.46494940e-01 -5.13557076e-01
-3.62561017e-01 -1.78799942e-01 8.07690844e-02 6.83386505e-01
-8.00546944e-01 1.16908574e+00 6.09651518e+00 5.17074823e-01
-1.11338997e+00 1.23576348e-04 6.28744662e-01 2.82585055e-01
-3.58925670e-01 3.42414469e-01 -9.88726199e-01 6.32900715e-01
4.99194562e-01 -2.51695633e-01 -1.77453179e-02 4.87152010e-01
2.67984897e-01 -3.28113049e-01 -9.99704123e-01 2.46265054e-01
-3.40107471e-01 -1.68006468e+00 -8.86266977e-02 -8.98382962e-02
6.69353902e-01 -2.29478300e-01 -3.51043522e-01 3.22257966e-01
3.02502304e-01 -9.21641946e-01 8.57098341e-01 -2.35558450e-01
5.38027346e-01 -3.98322165e-01 5.68401396e-01 8.87694657e-01
-1.36904705e+00 1.78044438e-01 -2.44851723e-01 -5.35819888e-01
5.90842783e-01 7.66912475e-02 -6.12207174e-01 7.17222512e-01
3.12453181e-01 4.75967586e-01 -1.75243840e-01 1.41288117e-01
-8.35093975e-01 9.83638704e-01 -9.35278716e-04 -3.12412381e-01
5.97462952e-01 -4.67576027e-01 6.44339383e-01 1.48375082e+00
-2.39036158e-02 4.37168300e-01 2.86591798e-01 5.71006238e-01
-1.73451543e-01 6.26939014e-02 -5.46853095e-02 1.09589651e-01
8.41747761e-01 8.80783200e-01 -6.79137051e-01 -4.65897739e-01
-5.55040240e-01 5.21330774e-01 4.23986316e-01 2.59048402e-01
-7.65471339e-01 -2.19515830e-01 2.71711826e-01 2.06490561e-01
2.60523230e-01 -4.01556402e-01 -5.86534321e-01 -1.03833866e+00
2.61287659e-01 -1.09092438e+00 4.95783538e-01 -3.24101716e-01
-1.01385784e+00 6.02347374e-01 1.27315387e-01 -1.05452776e+00
-2.78265953e-01 -3.74223351e-01 -1.26936591e+00 1.06337655e+00
-1.58906484e+00 -8.75612319e-01 3.10059432e-02 -4.21286859e-02
1.14629579e+00 3.94126475e-01 6.49380744e-01 9.79204848e-02
-7.77824640e-01 3.16683114e-01 -3.58881871e-03 7.91749060e-01
4.45071101e-01 -1.33521080e+00 5.43717086e-01 1.14239109e+00
-5.53582191e-01 8.71604204e-01 5.94366908e-01 -7.81374216e-01
-8.73224676e-01 -6.39835596e-01 1.34761655e+00 -1.77240893e-01
7.03816950e-01 -4.61954445e-01 -1.33763564e+00 7.35393941e-01
9.73840356e-01 -9.66506124e-01 7.54357636e-01 2.40182817e-01
-6.50573149e-02 3.63770515e-01 -6.78089142e-01 8.57388914e-01
8.44053566e-01 -3.47922295e-01 -1.42838705e+00 1.12871811e-01
1.21967638e+00 -1.02935588e+00 -6.96336091e-01 1.79577604e-01
3.04855883e-01 -7.48586297e-01 4.04309750e-01 -5.73089480e-01
1.05690730e+00 -4.62441921e-01 7.73867518e-02 -9.91086304e-01
-7.51178786e-02 -6.74699664e-01 6.85177818e-02 1.70012426e+00
6.44822538e-01 -5.34821689e-01 6.15168512e-01 5.80266654e-01
-3.79854292e-01 -6.56558037e-01 -8.45393598e-01 -2.58118361e-01
8.01612794e-01 4.75354455e-02 6.84305489e-01 1.00104535e+00
6.68054223e-01 1.00146341e+00 -8.36905930e-03 -1.14891678e-01
-5.62930740e-02 4.09937978e-01 5.39862573e-01 -6.90562904e-01
-1.91861197e-01 -6.06885135e-01 6.26577377e-01 -1.89248085e+00
1.15061574e-01 -5.57147980e-01 2.78835982e-01 -1.79469705e+00
-6.82699829e-02 -4.64616716e-01 4.31140214e-01 5.24579324e-02
-6.45555615e-01 -7.39279628e-01 3.79297614e-01 2.70848006e-01
-4.55990613e-01 4.83686924e-01 1.70900583e+00 -6.05164804e-02
-4.87535954e-01 -7.47348294e-02 -9.52129126e-01 1.01494110e+00
7.81812012e-01 -2.55089760e-01 -3.92044187e-01 -9.46374059e-01
1.89780042e-01 3.34327847e-01 -4.97765005e-01 -2.01115340e-01
3.25186014e-01 -4.36113715e-01 1.84876304e-02 -9.66663897e-01
-2.90803045e-01 -1.98048174e-01 -4.64431584e-01 5.58068931e-01
-5.20274520e-01 4.00798112e-01 -1.18603799e-02 1.31925613e-01
-6.31009758e-01 -4.46483493e-01 5.44714570e-01 -3.65714133e-01
-5.43169677e-01 -4.18614298e-01 -7.30524004e-01 5.42952240e-01
9.20082211e-01 -2.19384924e-01 -9.95118141e-01 -5.06409965e-02
-4.73406196e-01 7.12082088e-01 2.39304662e-01 5.47517896e-01
1.79292321e-01 -1.08122241e+00 -8.76275122e-01 -2.58199602e-01
-4.32416081e-01 8.60686123e-01 8.30275789e-02 6.39499187e-01
-9.06803966e-01 4.21069086e-01 5.15700094e-02 -5.31914949e-01
-1.52780724e+00 2.61561632e-01 1.33332297e-01 -8.35764229e-01
-7.23396599e-01 7.23469615e-01 5.43851078e-01 -2.04017192e-01
2.42377043e-01 -7.11982191e-01 -6.19003952e-01 1.15362190e-01
7.58651733e-01 2.60632426e-01 -2.42494732e-01 -6.75310969e-01
-2.12307990e-01 3.27463329e-01 -3.01917851e-01 -8.46045688e-02
8.76823843e-01 -6.14742279e-01 -6.48201883e-01 5.07972300e-01
8.19149792e-01 5.24739683e-01 -1.04705751e+00 -1.67545930e-01
5.33753276e-01 -2.31482163e-01 -4.51089174e-01 -5.41226923e-01
-3.19852322e-01 8.40444505e-01 -4.74029720e-01 6.71893597e-01
9.78388429e-01 -6.37778593e-03 1.09892929e+00 3.25931847e-01
3.26859355e-02 -1.47734094e+00 -1.60438105e-01 1.21168268e+00
8.46376300e-01 -9.17206407e-01 6.39671534e-02 -1.01313758e+00
-8.14125776e-01 1.14541209e+00 8.65185678e-01 2.49562357e-02
2.94399768e-01 4.48360324e-01 3.01633328e-01 5.82183823e-02
-8.09746683e-01 4.59753536e-02 -6.25956357e-02 4.37796563e-01
1.04315174e+00 3.50453407e-02 -9.59791005e-01 7.03522384e-01
-4.09112304e-01 -4.30411011e-01 5.74239671e-01 1.22787416e+00
-6.23207450e-01 -1.47626185e+00 -3.81419539e-01 -2.85203997e-02
-6.55007899e-01 -1.83211610e-01 -5.76756895e-01 1.12730455e+00
7.57439211e-02 1.27941871e+00 2.55016416e-01 4.93864492e-02
3.20448726e-01 -7.41598830e-02 3.98785800e-01 -8.80151689e-01
-8.76794755e-01 4.09703285e-01 7.81931341e-01 8.01792964e-02
-6.58217788e-01 -8.95883679e-01 -1.90557218e+00 -4.41363871e-01
-4.42182064e-01 2.81452388e-01 9.50848013e-02 1.23181927e+00
-1.11398943e-01 9.26464379e-01 5.16758859e-01 -2.38634139e-01
-3.53876591e-01 -1.17975652e+00 -8.52255523e-03 1.13258526e-01
3.77144009e-01 -9.48372036e-02 5.25979809e-02 1.90158665e-01] | [10.832134246826172, 9.504203796386719] |
41b56495-89ea-47c2-9f74-b2c9a9724edf | diffusionct-latent-diffusion-model-for-ct | 2301.08815 | null | https://arxiv.org/abs/2301.08815v2 | https://arxiv.org/pdf/2301.08815v2.pdf | DiffusionCT: Latent Diffusion Model for CT Image Standardization | Computed tomography (CT) is one of the modalities for effective lung cancer screening, diagnosis, treatment, and prognosis. The features extracted from CT images are now used to quantify spatial and temporal variations in tumors. However, CT images obtained from various scanners with customized acquisition protocols may introduce considerable variations in texture features, even for the same patient. This presents a fundamental challenge to downstream studies that require consistent and reliable feature analysis. Existing CT image harmonization models rely on GAN-based supervised or semi-supervised learning, with limited performance. This work addresses the issue of CT image harmonization using a new diffusion-based model, named DiffusionCT, to standardize CT images acquired from different vendors and protocols. DiffusionCT operates in the latent space by mapping a latent non-standard distribution into a standard one. DiffusionCT incorporates an Unet-based encoder-decoder, augmented by a diffusion model integrated into the bottleneck part. The model is designed in two training phases. The encoder-decoder is first trained, without embedding the diffusion model, to learn the latent representation of the input data. The latent diffusion model is then trained in the next training phase while fixing the encoder-decoder. Finally, the decoder synthesizes a standardized image with the transformed latent representation. The experimental results demonstrate a significant improvement in the performance of the standardization task using DiffusionCT. | ['Jin Chen', 'Ge Wang', 'Michael A. Brooks', 'Jie Zhang', 'Md Selim'] | 2023-01-20 | null | null | null | null | ['image-harmonization', 'lung-cancer-diagnosis'] | ['computer-vision', 'medical'] | [ 4.54904169e-01 5.46095893e-02 -4.60574448e-01 -3.01026791e-01
-1.20248008e+00 -1.87504604e-01 3.44632775e-01 -9.21239778e-02
-3.00736874e-01 3.88034850e-01 3.93415123e-01 -1.96778789e-01
1.16290651e-01 -8.64238918e-01 -4.32899624e-01 -1.17639005e+00
2.70937324e-01 6.42981529e-01 3.47020149e-01 2.39593193e-01
-3.92635673e-01 2.23298386e-01 -6.89150691e-01 6.68481231e-01
6.36919379e-01 8.79872382e-01 5.31902373e-01 8.05368721e-01
-2.13373870e-01 9.06198382e-01 -3.47577363e-01 -1.10821612e-01
2.17832431e-01 -9.68544602e-01 -6.06848121e-01 2.49865323e-01
4.22155187e-02 -2.27767542e-01 -3.55504453e-01 8.39425325e-01
6.05441511e-01 -3.84357482e-01 6.38704717e-01 -8.43084753e-01
-4.19896364e-01 6.18365705e-01 -4.15696502e-01 1.60923481e-01
-1.10829972e-01 2.93104202e-01 4.99437392e-01 -5.07024527e-01
7.40430713e-01 7.84551084e-01 4.87156957e-01 8.38804543e-01
-1.19000316e+00 -5.06590664e-01 -3.01471293e-01 -4.55787070e-02
-1.14183939e+00 -8.30915123e-02 5.84780633e-01 -4.85349655e-01
6.50437295e-01 2.66173661e-01 8.78723741e-01 1.15643847e+00
8.03166032e-01 3.36375535e-01 1.22367013e+00 -3.70756209e-01
1.45318523e-01 1.41451031e-01 -3.55346262e-01 8.58797669e-01
8.71194974e-02 1.96155116e-01 -3.79142374e-01 -1.28294677e-02
9.18859124e-01 1.40406430e-01 -4.32590455e-01 -5.39871395e-01
-1.21295714e+00 9.21694815e-01 5.19017279e-01 6.36654973e-01
-5.35298347e-01 2.63091624e-01 4.50162977e-01 1.94343880e-01
3.93001765e-01 -6.66229874e-02 -2.96116043e-02 -3.72482873e-02
-1.07205462e+00 -3.95056933e-01 4.88893747e-01 7.21972644e-01
4.11162823e-01 -1.09640755e-01 -4.59995091e-01 6.81197107e-01
4.44915831e-01 5.46168566e-01 1.06228709e+00 -6.36296451e-01
4.88087744e-01 4.15677130e-01 -6.26232207e-01 -4.42788005e-01
-2.80756295e-01 -5.11392653e-01 -1.15328777e+00 4.17823792e-02
3.42519730e-02 7.84294531e-02 -1.25515497e+00 1.51254702e+00
2.88482279e-01 2.93861151e-01 -3.50560285e-02 6.63364768e-01
7.38538146e-01 3.49913418e-01 2.54185945e-01 -3.38507295e-01
1.18781197e+00 -1.16503358e+00 -6.70597792e-01 2.70423386e-02
7.78644562e-01 -9.94011104e-01 9.70313311e-01 4.56199497e-02
-1.03088748e+00 -4.17378902e-01 -1.01542556e+00 8.79626796e-02
-9.62074287e-03 5.17543592e-02 1.39965624e-01 9.38401580e-01
-9.77893651e-01 3.30110133e-01 -1.34428644e+00 -3.33040923e-01
3.48789424e-01 4.12076980e-01 -4.13615495e-01 -2.60629833e-01
-8.74742627e-01 7.72135913e-01 1.77116275e-01 -9.09490809e-02
-1.01689470e+00 -8.38705063e-01 -7.45870173e-01 -1.48967654e-01
1.28657594e-01 -1.14093792e+00 1.33067536e+00 -1.05074179e+00
-1.97172356e+00 8.42549384e-01 -2.17681244e-01 -3.35472375e-01
7.76437819e-01 3.77582908e-01 -2.57726610e-01 2.31500462e-01
3.49891514e-01 4.65975583e-01 1.12546313e+00 -1.11737895e+00
-3.32078695e-01 -1.66021556e-01 -6.51954472e-01 2.43906036e-01
-5.57430042e-03 -2.37182200e-01 -8.23022544e-01 -7.64388919e-01
3.13544899e-01 -1.07914197e+00 -3.78273070e-01 -5.94811887e-02
-3.29401195e-01 4.62549716e-01 9.05021846e-01 -6.18959665e-01
1.06883371e+00 -2.07815003e+00 2.16463774e-01 5.20860553e-01
4.24729645e-01 -1.17959224e-01 6.17215820e-02 6.79795668e-02
-3.65785599e-01 -2.64998935e-02 -5.11288106e-01 -3.75691652e-01
-3.80909234e-01 4.65042949e-01 1.67705595e-01 4.75714505e-01
-8.44330117e-02 1.20752823e+00 -7.84627080e-01 -7.30435431e-01
3.89458030e-01 4.37778890e-01 -4.95056897e-01 3.99036765e-01
-1.03333313e-02 1.13580728e+00 -5.50653160e-01 3.83710563e-01
5.80625594e-01 -5.72277725e-01 5.78257024e-01 -3.28473926e-01
1.46227013e-02 1.99376941e-01 -6.06787980e-01 1.88832009e+00
-6.65825188e-01 3.37669373e-01 -1.43384457e-01 -7.87519097e-01
4.81394559e-01 5.79907358e-01 1.09624636e+00 -8.30596566e-01
2.42743149e-01 3.45060617e-01 1.89330041e-01 -5.78493595e-01
1.10199861e-01 -5.33948898e-01 -5.33629023e-02 6.25330627e-01
-6.85368292e-03 -5.66241205e-01 -2.13484347e-01 1.34050667e-01
1.23800254e+00 -3.12612474e-01 1.79761812e-01 2.11093426e-02
4.76086378e-01 2.87352572e-03 3.15155864e-01 4.71705854e-01
-7.12173581e-02 8.75330687e-01 4.23048913e-01 -1.30959138e-01
-1.22391260e+00 -1.25943887e+00 -1.26850829e-01 3.95435989e-01
-1.57912791e-01 -4.30754066e-01 -6.61131442e-01 -9.27879274e-01
-3.46704870e-01 3.41438651e-01 -8.57293129e-01 -3.34097356e-01
-5.90716183e-01 -8.62493694e-01 3.24596256e-01 5.27004242e-01
7.20114231e-01 -7.26907849e-01 -6.29103005e-01 2.38753542e-01
-2.67348498e-01 -1.03964865e+00 -6.84897840e-01 4.62179095e-01
-1.16228485e+00 -8.52891743e-01 -9.62187588e-01 -6.21401429e-01
9.34036553e-01 8.32807869e-02 1.00425804e+00 1.31665349e-01
-2.98403889e-01 4.15902764e-01 -2.93195605e-01 -2.51118373e-02
-9.26033080e-01 1.88602105e-01 -5.11561096e-01 -2.23629382e-02
-1.28380284e-01 -3.17775130e-01 -7.54742682e-01 3.88427615e-01
-1.19289279e+00 3.54192495e-01 8.67080033e-01 1.18599021e+00
9.50077534e-01 1.93868116e-01 -9.71518010e-02 -1.22110748e+00
4.11137342e-01 -4.83464718e-01 -3.47893864e-01 2.67023265e-01
-4.76148635e-01 2.38386869e-01 4.99078035e-01 -1.98239163e-01
-1.10902572e+00 2.76949972e-01 -3.08423311e-01 -4.94771093e-01
2.27307051e-01 6.36592031e-01 -1.23994261e-01 -2.16081515e-01
6.28728330e-01 3.83022338e-01 -1.39998039e-03 -1.39115557e-01
1.24765195e-01 4.49012130e-01 4.16239202e-01 -3.16598624e-01
8.46296668e-01 6.06112540e-01 -5.64616024e-02 -5.20401001e-01
-5.64235449e-01 -4.12662983e-01 -7.46913731e-01 -2.55499721e-01
1.11629391e+00 -6.73100650e-01 1.19536616e-01 5.56011379e-01
-7.63282716e-01 -7.48002350e-01 -7.37260044e-01 7.79534996e-01
-4.99135345e-01 3.13199699e-01 -6.47859097e-01 4.92307283e-02
-4.27074552e-01 -1.57722092e+00 1.06435978e+00 5.49743213e-02
-9.45684463e-02 -1.20838892e+00 4.32823449e-01 2.76431531e-01
6.52378201e-01 1.40279546e-01 1.16966915e+00 -2.07493454e-01
-6.73483253e-01 -4.52376902e-02 -8.40874091e-02 3.87265235e-01
6.63119018e-01 -2.47103885e-01 -8.23965132e-01 -2.92745620e-01
4.33618456e-01 -9.04857665e-02 8.41213882e-01 7.24402845e-01
1.28010190e+00 1.30091116e-01 -3.99142653e-01 9.66969371e-01
1.43351352e+00 2.19221994e-01 6.82250440e-01 1.89933449e-01
6.98696375e-01 1.39182836e-01 9.92089137e-02 1.43124714e-01
2.31755495e-01 6.43100739e-01 1.89641446e-01 -4.31655824e-01
-5.57554662e-01 -2.02627733e-01 2.51081228e-01 1.28773081e+00
4.04799655e-02 -1.57576561e-01 -1.02249074e+00 1.24179654e-01
-1.49590802e+00 -7.56920218e-01 -2.49673799e-02 2.14861846e+00
8.21128011e-01 5.66219054e-02 -3.19242924e-01 3.20151933e-02
4.35544789e-01 4.91688843e-04 -5.16494870e-01 -5.39762303e-02
5.75422160e-02 6.51664853e-01 7.67367601e-01 4.14620191e-01
-8.14347684e-01 6.61033094e-01 6.84140730e+00 8.03152740e-01
-1.78448331e+00 6.74497545e-01 6.90075219e-01 6.97466284e-02
-4.62411135e-01 -1.12681106e-01 -2.17420071e-01 4.37934935e-01
9.38054919e-01 -3.10108252e-02 1.00207580e-02 5.46160996e-01
3.37070912e-01 -3.42547596e-01 -9.88009334e-01 8.66663098e-01
2.22764872e-02 -1.29028249e+00 3.80757563e-02 1.82238594e-01
9.03398097e-01 7.90339038e-02 4.31693375e-01 -2.85076052e-02
1.63184479e-01 -1.04863739e+00 5.09444296e-01 4.78908449e-01
1.19177043e+00 -3.30089152e-01 8.83074701e-01 1.27884492e-01
-1.30130816e+00 4.26384419e-01 -1.72378227e-01 8.01580071e-01
1.81431547e-01 5.28647542e-01 -1.14857769e+00 7.94741571e-01
4.69750077e-01 7.23608196e-01 -5.16215920e-01 7.79543579e-01
-2.37303838e-01 6.93637431e-01 -6.45513982e-02 5.26373327e-01
4.35688309e-02 -1.21341817e-01 2.16834709e-01 1.05172992e+00
5.24076104e-01 -7.21213669e-02 1.25261471e-01 5.85513055e-01
1.34932874e-02 1.51454523e-01 -4.33380604e-01 8.05218890e-02
5.65214790e-02 1.16771066e+00 -1.00112355e+00 -4.10325885e-01
-5.05883634e-01 1.30623317e+00 -5.84949851e-02 1.00699641e-01
-1.16781688e+00 3.57755661e-01 -1.05692163e-01 2.99492687e-01
2.03287452e-01 -2.05169171e-01 -2.77627736e-01 -1.10102046e+00
-2.96168745e-01 -8.87171626e-01 6.13384902e-01 -6.64871573e-01
-1.29212487e+00 8.75198364e-01 9.63419601e-02 -1.33352423e+00
-3.70327353e-01 -3.93217921e-01 -6.85628235e-01 9.50057864e-01
-1.44527340e+00 -1.36739135e+00 -6.62660778e-01 8.61373186e-01
5.07252991e-01 -1.31372765e-01 9.28026199e-01 3.87713671e-01
-6.21964216e-01 6.24778390e-01 2.27689877e-01 5.59479520e-02
7.91379690e-01 -1.21520317e+00 -5.05692996e-02 5.64748228e-01
-2.84054726e-01 2.50976115e-01 2.10483149e-01 -8.79765928e-01
-1.19515097e+00 -1.36394656e+00 5.69159865e-01 -2.12177441e-01
3.71966064e-01 -1.97257977e-02 -7.58365214e-01 7.45295227e-01
3.10952872e-01 2.14350969e-01 9.39781785e-01 -6.18532598e-01
4.14397344e-02 3.93762859e-03 -1.09568489e+00 3.18106771e-01
6.02925599e-01 -6.64549530e-01 -2.30750456e-01 3.40825170e-01
4.40231085e-01 -8.44435692e-01 -1.01938653e+00 2.12108165e-01
5.48777997e-01 -8.21506083e-01 8.74940813e-01 -5.83937392e-02
5.19200861e-01 -1.11298658e-01 -5.65552898e-02 -1.36419094e+00
-4.19352144e-01 -2.08176345e-01 3.29185098e-01 6.98978245e-01
4.37214762e-01 -5.62668622e-01 1.16353869e+00 4.58266109e-01
-1.52952895e-01 -6.77735746e-01 -1.10423040e+00 -6.29089892e-01
1.97275341e-01 -4.68161643e-01 4.00361776e-01 7.66839743e-01
-3.91234696e-01 5.16614839e-02 -8.64248723e-02 -5.04099466e-02
4.83110905e-01 -7.87431225e-02 5.08649468e-01 -5.77202559e-01
-5.80089211e-01 -3.34294766e-01 -2.53495216e-01 -8.25969458e-01
-1.46418795e-01 -1.19427872e+00 7.67013431e-03 -1.37519860e+00
5.09818852e-01 -5.15269458e-01 -2.76919156e-01 2.90719301e-01
-3.58723514e-02 2.21338287e-01 -4.80076857e-02 4.65455145e-01
-1.59318715e-01 5.69861829e-01 1.74184799e+00 -3.60484272e-01
-1.37965053e-01 4.24912982e-02 -2.21186772e-01 5.93787968e-01
8.62549067e-01 -9.32902575e-01 -5.45773447e-01 -4.60133612e-01
-2.20066234e-01 2.40380839e-01 2.90323228e-01 -1.08249772e+00
3.26994509e-01 -9.79273617e-02 5.12591243e-01 -4.42532480e-01
1.72441170e-01 -1.09466600e+00 5.64443767e-01 9.81731057e-01
-1.76556081e-01 2.59294003e-01 5.48410267e-02 3.83471072e-01
-4.57478642e-01 -2.52868384e-01 1.06852639e+00 -2.43837491e-01
-2.49236330e-01 5.92220485e-01 -5.22927046e-01 -7.36874267e-02
1.09191465e+00 -2.09118903e-01 -3.58912512e-03 -2.97918946e-01
-8.37363899e-01 -2.29875803e-01 4.43320245e-01 -5.24427332e-02
7.73934364e-01 -1.35702193e+00 -7.34774530e-01 4.60703701e-01
2.45308708e-02 2.69528508e-01 5.01704335e-01 1.25908899e+00
-7.51118064e-01 3.22268307e-01 -2.48695463e-01 -1.04121590e+00
-1.12211668e+00 2.29866862e-01 7.36703873e-01 -1.09931600e+00
-6.59604311e-01 6.16013944e-01 3.65845323e-01 -2.22799867e-01
-1.84206679e-01 -5.44837892e-01 2.27750376e-01 -3.09884489e-01
-4.54563312e-02 -8.55981782e-02 3.06014150e-01 -6.45354807e-01
-1.32963941e-01 7.43824959e-01 -9.76860300e-02 -4.18427318e-01
1.12822783e+00 -1.00479908e-02 -3.00934575e-02 2.41538599e-01
1.33917308e+00 1.67728052e-01 -1.00069618e+00 -2.46054128e-01
-2.30044499e-01 -4.13230807e-01 2.93379009e-01 -6.54397190e-01
-1.50441504e+00 8.34124565e-01 9.60578978e-01 -1.91947967e-01
1.45013928e+00 2.79895682e-03 8.74064326e-01 -2.18744501e-01
2.11477831e-01 -6.62016809e-01 4.93930727e-01 3.35110009e-01
7.41854012e-01 -1.17086864e+00 -2.84934286e-02 -5.58903933e-01
-8.16353798e-01 1.19124210e+00 4.55943406e-01 1.91440135e-01
8.17762434e-01 5.91085136e-01 4.14936960e-01 -3.21853191e-01
-7.31183410e-01 4.67647575e-02 3.24938387e-01 4.47425663e-01
5.78201771e-01 1.22773938e-01 -7.34962747e-02 4.28411633e-01
-1.27582043e-01 2.22072139e-01 1.15749866e-01 1.11476731e+00
-4.63037565e-02 -1.68095505e+00 -4.46793407e-01 5.31676948e-01
-3.29962283e-01 -1.70712117e-02 -5.81202246e-02 7.43084133e-01
3.33157033e-01 3.90865803e-01 -2.49051452e-02 -4.61804688e-01
1.88228577e-01 -6.30456582e-02 6.31462634e-01 -9.27314937e-01
-7.72065341e-01 3.40665340e-01 -4.28688645e-01 -4.69651639e-01
-5.81901848e-01 -6.56114459e-01 -1.28684449e+00 -1.35705546e-01
-2.13770956e-01 1.07659176e-01 6.35318577e-01 7.15769112e-01
-1.02326632e-01 1.05498207e+00 7.85233974e-01 -4.18341160e-01
-2.19840065e-01 -7.68390894e-01 -4.34878588e-01 3.36828321e-01
6.79651797e-02 -4.06214952e-01 -5.56450756e-03 3.56623203e-01] | [14.022889137268066, -2.30000901222229] |
05aa9e99-e336-402a-840f-ab64c3b8d16b | sublabel-accurate-convex-relaxation-of | 1604.01980 | null | http://arxiv.org/abs/1604.01980v2 | http://arxiv.org/pdf/1604.01980v2.pdf | Sublabel-Accurate Convex Relaxation of Vectorial Multilabel Energies | Convex relaxations of nonconvex multilabel problems have been demonstrated to
produce superior (provably optimal or near-optimal) solutions to a variety of
classical computer vision problems. Yet, they are of limited practical use as
they require a fine discretization of the label space, entailing a huge demand
in memory and runtime. In this work, we propose the first sublabel accurate
convex relaxation for vectorial multilabel problems. The key idea is that we
approximate the dataterm of the vectorial labeling problem in a piecewise
convex (rather than piecewise linear) manner. As a result we have a more
faithful approximation of the original cost function that provides a meaningful
interpretation for the fractional solutions of the relaxed convex problem. In
numerous experiments on large-displacement optical flow estimation and on color
image denoising we demonstrate that the computed solutions have superior
quality while requiring much lower memory and runtime. | ['Daniel Cremers', 'Thomas Möllenhoff', 'Michael Moeller', 'Jan Lellmann', 'Emanuel Laude'] | 2016-04-07 | null | null | null | null | ['color-image-denoising'] | ['computer-vision'] | [-3.67082097e-02 -2.44308282e-02 -3.67350459e-01 -4.43202078e-01
-1.14295864e+00 -6.95032895e-01 -1.86340306e-02 -1.34555167e-02
-4.26617384e-01 1.04435158e+00 9.17542130e-02 -2.38798440e-01
-9.06379670e-02 -3.19702029e-01 -7.79505610e-01 -9.03122187e-01
1.63013726e-01 5.47668934e-01 -3.16258341e-01 -1.08618604e-03
2.43188918e-01 5.17839849e-01 -1.22670424e+00 -1.23642646e-01
1.11379135e+00 8.92077625e-01 -1.84317932e-01 4.82506901e-01
-7.18824938e-02 5.91276407e-01 -2.39648849e-01 -2.32746512e-01
6.05098009e-01 -2.48260230e-01 -1.01094460e+00 5.10286808e-01
1.20140517e+00 -4.20616984e-01 -1.84936345e-01 1.25673997e+00
2.10317552e-01 4.59651411e-01 5.41969717e-01 -1.23143578e+00
-5.91109812e-01 2.60039449e-01 -9.93900836e-01 1.46429604e-02
1.41365677e-01 -2.57703304e-01 1.21580172e+00 -8.23096633e-01
6.91674471e-01 1.29964054e+00 8.38549256e-01 3.45018893e-01
-1.67835724e+00 -1.26658171e-01 3.07193369e-01 1.80239265e-03
-1.29451835e+00 -4.84382242e-01 5.28347433e-01 -4.99491096e-01
4.09659088e-01 4.28153872e-01 5.93131125e-01 3.22304338e-01
5.01777939e-02 7.10940003e-01 1.18010306e+00 -2.29007617e-01
1.33371934e-01 1.10837519e-01 5.05223513e-01 1.13581872e+00
3.11429381e-01 -2.19408825e-01 -1.09675072e-01 -4.26398695e-01
9.17267382e-01 -9.44719911e-02 -5.82306981e-01 -4.30797279e-01
-1.15947270e+00 1.04059029e+00 4.08858299e-01 5.29385097e-02
-2.68265694e-01 5.72268784e-01 3.91249031e-01 9.38469023e-02
6.69379354e-01 3.29968363e-01 -1.58738509e-01 -7.02883536e-03
-1.06484449e+00 4.23791856e-01 7.09413707e-01 9.58525836e-01
1.06681132e+00 1.95849612e-01 -6.64412156e-02 6.79309189e-01
1.80219635e-01 5.30816913e-01 -1.96325600e-01 -1.70339322e+00
3.62471044e-01 1.30699679e-01 4.90538061e-01 -1.26868045e+00
-3.81992638e-01 -3.56778055e-01 -7.42540896e-01 2.69886225e-01
8.37506354e-01 -8.50965902e-02 -7.46701479e-01 1.73659384e+00
4.28668916e-01 4.04695004e-01 -1.82400942e-01 1.11637843e+00
3.36693287e-01 9.86832261e-01 -6.99814707e-02 -6.85764611e-01
1.04378116e+00 -1.10422349e+00 -8.99941742e-01 -2.45416090e-02
7.02745855e-01 -8.76561999e-01 1.00497282e+00 4.45813149e-01
-1.35426724e+00 -3.84947151e-01 -8.62254441e-01 -5.19219995e-01
2.40130916e-01 -1.50417343e-01 8.30731571e-01 5.39032221e-01
-1.17173910e+00 5.37046969e-01 -7.43405342e-01 8.68693069e-02
2.78455079e-01 2.99382836e-01 -3.01004946e-01 -4.58508313e-01
-4.96197939e-01 6.11277103e-01 -4.77612317e-02 3.92356664e-01
-5.42867780e-01 -9.83586431e-01 -8.83575976e-01 -5.24930507e-02
4.17528212e-01 -5.15167415e-01 9.41551328e-01 -8.62599850e-01
-1.11521554e+00 9.47324038e-01 -5.37151456e-01 -1.51337564e-01
6.78221762e-01 2.44814381e-02 3.28695536e-01 2.72314578e-01
1.70981362e-01 6.35648549e-01 8.54195535e-01 -1.59558249e+00
-5.85643053e-01 -3.30960661e-01 2.91419387e-01 1.29837826e-01
-2.11708903e-01 -3.28852713e-01 -4.04960752e-01 -7.32121944e-01
1.46063924e-01 -1.21935165e+00 -5.05272388e-01 1.77313074e-01
-2.27532789e-01 -1.39277697e-01 6.15941584e-01 -7.87787676e-01
1.11710858e+00 -1.96295965e+00 5.33634722e-01 1.35385945e-01
4.55963820e-01 1.98606998e-01 -1.30400866e-01 6.02922291e-02
1.15755484e-01 1.49334386e-01 -5.59331954e-01 -5.83818376e-01
8.19297880e-02 7.19488859e-01 -3.71287227e-01 1.10489798e+00
-3.32050174e-01 8.67005944e-01 -1.06265271e+00 -6.11715198e-01
1.43827125e-01 4.36001241e-01 -9.53892291e-01 -2.94635091e-02
-1.02172635e-01 4.87392306e-01 -4.25169021e-01 4.53244179e-01
9.20735061e-01 -1.62122518e-01 4.64381129e-02 -3.93024594e-01
-1.42261878e-01 -4.08699065e-01 -1.35813630e+00 1.84961951e+00
-4.57528591e-01 6.80530846e-01 6.41155303e-01 -1.30987644e+00
4.16587919e-01 1.51234433e-01 9.63044345e-01 -2.56177664e-01
5.19711971e-02 1.40533194e-01 -5.23518264e-01 -3.82280439e-01
7.46940732e-01 -4.84637350e-01 -3.54990587e-02 4.62066650e-01
-3.29648077e-01 -1.78802013e-01 4.32752460e-01 1.01753823e-01
5.71066916e-01 3.33553962e-02 3.00070960e-02 -7.13767886e-01
6.92633033e-01 2.45633289e-01 9.36442494e-01 5.11214435e-01
-2.17471048e-01 6.51184261e-01 5.50261140e-01 -5.39309263e-01
-9.86223996e-01 -8.08014989e-01 -4.57037717e-01 8.99793684e-01
2.84809321e-01 -1.61062747e-01 -9.32144523e-01 -4.67556536e-01
9.51784626e-02 2.62669295e-01 -3.40888113e-01 2.69274950e-01
-1.05655086e+00 -8.32381785e-01 2.54502624e-01 4.51508582e-01
1.77215084e-01 -3.97036165e-01 -4.06161964e-01 2.85430700e-01
-4.82056409e-01 -1.43459320e+00 -1.04094589e+00 -1.52010113e-01
-9.79590178e-01 -9.87123847e-01 -7.94027984e-01 -1.01369536e+00
1.10930812e+00 4.90013689e-01 1.04840684e+00 2.47413054e-01
-2.37595767e-01 5.25310576e-01 6.29070178e-02 2.62100667e-01
-9.20239910e-02 -3.67453873e-01 2.08215471e-02 3.25939804e-01
-4.09388870e-01 -1.95580229e-01 -4.27822709e-01 2.06470594e-01
-9.16109800e-01 -2.38145337e-01 -7.34160468e-02 1.02442265e+00
1.10314143e+00 -2.90538073e-02 2.35783100e-01 -1.03008199e+00
6.64557576e-01 -1.40546918e-01 -1.07533360e+00 2.52172887e-01
-6.83971286e-01 2.67911762e-01 8.06891143e-01 -2.96784788e-01
-1.07318521e+00 2.38773867e-01 4.90925126e-02 -5.65111756e-01
2.19432548e-01 5.45024097e-01 1.47511035e-01 -5.52255929e-01
3.74488711e-01 -6.12592213e-02 -8.03603530e-02 -3.90392929e-01
6.93589389e-01 2.38217250e-01 7.57714689e-01 -1.11858082e+00
6.81346416e-01 9.55860674e-01 4.14140701e-01 -7.93406725e-01
-1.07276249e+00 -5.62744677e-01 -4.86529619e-01 -1.43024832e-01
9.43149388e-01 -5.80965757e-01 -8.62917066e-01 3.17024916e-01
-1.17573762e+00 -3.30345690e-01 -4.45943207e-01 3.92693788e-01
-8.02493215e-01 7.38254309e-01 -7.27069974e-01 -6.06665909e-01
-3.76471598e-03 -1.46578789e+00 1.03558302e+00 1.68027915e-02
1.57513246e-01 -1.33514857e+00 4.72396314e-02 6.70840561e-01
9.49084461e-02 5.96226394e-01 8.40683937e-01 3.13388675e-01
-7.16370523e-01 -1.05803788e-01 -3.98234427e-01 5.43738186e-01
-7.73054361e-02 -4.65934910e-02 -7.04342067e-01 -5.86880863e-01
1.69021472e-01 -2.67428339e-01 8.07154536e-01 8.49962413e-01
1.13244474e+00 -4.10443693e-01 -9.18294713e-02 1.16026068e+00
1.66389096e+00 -1.19474128e-01 2.78388947e-01 -8.64085853e-02
1.08164883e+00 6.15522206e-01 8.19789886e-01 3.55762184e-01
4.35026675e-01 6.47140563e-01 3.42047393e-01 -2.95368701e-01
-1.05058156e-01 2.03293979e-01 1.83368355e-01 9.16759193e-01
-2.21473604e-01 -1.64753899e-01 -7.50299573e-01 5.84500611e-01
-2.22160959e+00 -8.14306021e-01 -5.10667384e-01 2.23069549e+00
7.88444459e-01 -4.68833238e-01 3.57699133e-02 6.00599684e-02
6.98232055e-01 2.03799307e-01 -3.67610514e-01 -6.20218754e-01
-2.87224315e-02 -8.69622454e-02 8.46588731e-01 1.18728876e+00
-1.05611706e+00 8.26595783e-01 7.21267700e+00 6.54378355e-01
-1.08072078e+00 1.66448891e-01 6.59458995e-01 -7.73214102e-02
-4.71606702e-01 2.43242402e-02 -6.67966545e-01 2.70960003e-01
5.85223794e-01 -2.19556093e-01 8.29277754e-01 4.84747916e-01
3.33584070e-01 -1.07864045e-01 -1.11212957e+00 1.25615239e+00
2.16498189e-02 -1.54263651e+00 -9.50764492e-02 4.08615649e-01
1.20398676e+00 -2.60948688e-01 1.28956199e-01 -2.96437740e-01
1.16519809e-01 -9.74171817e-01 8.37774992e-01 4.07841921e-01
7.85197556e-01 -8.37436080e-01 5.52413523e-01 2.63276279e-01
-1.07328200e+00 7.93082714e-02 -6.91217661e-01 -3.13466378e-02
5.56330144e-01 7.07625210e-01 -1.32374421e-01 2.87444890e-01
3.91647249e-01 7.96058178e-01 2.04029307e-01 9.39629018e-01
2.10542548e-02 4.57286507e-01 -2.38063812e-01 5.28315604e-01
6.14337325e-01 -8.10647666e-01 6.20778620e-01 1.17834449e+00
3.55967209e-02 3.48459393e-01 6.49818659e-01 7.62914181e-01
-2.33468756e-01 2.51556754e-01 -3.18584293e-01 9.77667645e-02
4.25312743e-02 1.26773787e+00 -6.58314347e-01 -2.74575591e-01
-5.28350472e-01 7.62623489e-01 4.89841163e-01 7.91141272e-01
-9.88995194e-01 1.05939731e-02 8.77085149e-01 2.63740811e-02
1.08302720e-01 -4.64265734e-01 -4.19925481e-01 -1.33161414e+00
1.01420864e-01 -6.35355592e-01 3.64732414e-01 -3.23296398e-01
-1.19482601e+00 3.48806471e-01 -1.55256558e-02 -8.70540261e-01
-1.15500398e-01 -5.80588639e-01 -1.06882900e-01 8.72552812e-01
-1.70308173e+00 -9.74757433e-01 -3.38932008e-01 6.87361479e-01
2.59274662e-01 4.17897254e-01 5.32105207e-01 5.93227625e-01
-5.87307632e-01 4.61035341e-01 2.69358307e-01 -1.04001835e-01
5.92578053e-01 -1.38816059e+00 -1.21111646e-01 9.76034641e-01
7.48491427e-03 5.34969509e-01 8.44244123e-01 -2.90151983e-01
-1.69307756e+00 -1.07339394e+00 6.43619597e-01 -1.55732885e-01
6.53442919e-01 5.82335405e-02 -9.70211327e-01 9.20049012e-01
1.46684125e-02 5.09592712e-01 4.17765856e-01 -1.54067785e-01
-1.86799705e-01 -1.17533393e-01 -1.38032281e+00 3.29503238e-01
9.25465286e-01 -4.05628711e-01 -1.83926627e-01 5.86101830e-01
4.27091986e-01 -5.11872888e-01 -1.05381179e+00 7.13229775e-02
3.37848276e-01 -7.42082894e-01 1.17631471e+00 -5.56983531e-01
1.71521142e-01 -4.55292881e-01 -3.14590454e-01 -1.20875669e+00
-4.16898012e-01 -8.14205766e-01 -1.04376942e-01 9.50863123e-01
5.07304333e-02 -7.74289906e-01 7.68050015e-01 7.60266483e-01
-3.26956809e-01 -8.52221668e-01 -1.05953646e+00 -8.48588526e-01
2.82925874e-01 -2.38031611e-01 2.27777120e-02 1.17048466e+00
-2.13726044e-01 -3.28866728e-02 -6.55323863e-01 1.23530902e-01
1.26957202e+00 3.89778823e-01 4.92300451e-01 -1.03490043e+00
-1.92500114e-01 -3.10247540e-01 -2.71458924e-01 -1.43057811e+00
6.27292812e-01 -1.11577082e+00 2.75069714e-01 -1.50683641e+00
2.98350215e-01 -8.63857925e-01 -1.13320075e-01 3.37484717e-01
-2.81654835e-01 6.08475924e-01 1.43429607e-01 2.01281264e-01
-4.12653327e-01 2.64677912e-01 1.55984867e+00 -3.83594453e-01
-9.87972617e-02 -2.93740451e-01 -7.13518918e-01 7.48456538e-01
4.66402501e-01 -4.54628378e-01 -5.62974215e-01 -8.77404273e-01
1.71140447e-01 2.73188770e-01 2.33037114e-01 -3.71074468e-01
5.07645644e-02 -4.08843905e-01 -3.36716384e-01 -3.20062876e-01
3.59195828e-01 -7.13241816e-01 -2.38244720e-02 2.67639458e-01
-2.87749439e-01 -7.34315580e-03 4.01534811e-02 5.80302119e-01
-2.72876531e-01 -3.57851952e-01 1.10716081e+00 -7.28469715e-03
-5.82908392e-01 5.82498789e-01 -1.65515035e-01 4.46195900e-01
1.04266274e+00 -1.62957221e-01 -2.29057789e-01 -3.87580872e-01
-7.23111570e-01 2.23346472e-01 6.49349451e-01 -5.59667870e-03
4.23541963e-01 -1.35596812e+00 -7.27542818e-01 -1.53500959e-01
-3.88805717e-01 1.78843871e-01 1.60592481e-01 9.86598969e-01
-9.79339421e-01 4.62961912e-01 -1.34802490e-01 -7.12562203e-01
-1.15325928e+00 6.83835626e-01 4.25288796e-01 -1.68985933e-01
-8.34081352e-01 9.10846710e-01 3.49445343e-01 -2.61211008e-01
2.86189884e-01 -5.59560895e-01 2.74967521e-01 1.31990045e-01
4.29815859e-01 8.97561371e-01 -1.78644404e-01 -9.61365342e-01
-2.19682395e-01 1.05410218e+00 2.74486512e-01 -1.41061619e-02
1.35674572e+00 -3.08768868e-01 -5.77861607e-01 6.86671361e-02
1.52511144e+00 1.82508215e-01 -1.57785571e+00 -1.09699912e-01
-2.05909744e-01 -7.74395049e-01 4.78487462e-01 -7.57723451e-02
-1.49348235e+00 7.20300555e-01 2.68925875e-01 2.23495573e-01
9.99150395e-01 -9.02748331e-02 1.26015449e+00 2.67640173e-01
4.88537192e-01 -1.03814781e+00 -3.48856628e-01 2.04639152e-01
8.15528154e-01 -1.20213604e+00 3.63389969e-01 -7.20564723e-01
-2.59240597e-01 1.02466607e+00 1.50206849e-01 -4.01668191e-01
5.17543793e-01 3.70569646e-01 2.45082662e-01 1.84995290e-02
-2.46607929e-01 5.26289567e-02 3.74221683e-01 3.94872189e-01
2.83874482e-01 2.01275453e-01 -6.51540637e-01 6.07768865e-03
7.43765458e-02 -2.93465406e-01 7.38174081e-01 7.37807631e-01
-3.53957206e-01 -1.00602150e+00 -5.70731699e-01 5.91269694e-02
-4.54492271e-01 6.64235428e-02 1.48679376e-01 4.25397813e-01
5.94407693e-02 9.90464330e-01 -1.58545256e-01 2.34381601e-01
2.31899038e-01 -3.14536750e-01 7.91216314e-01 -3.06891233e-01
-2.87397206e-01 3.18341672e-01 -9.15190130e-02 -8.86893034e-01
-5.47053337e-01 -7.25491047e-01 -1.58131039e+00 -6.19987488e-01
-9.11446139e-02 2.04188630e-01 4.91917998e-01 8.03502738e-01
8.17655306e-03 3.44126999e-01 5.18021703e-01 -1.17121351e+00
-6.92718327e-01 -2.27265149e-01 -6.19169950e-01 6.07757568e-01
6.50287449e-01 -5.95087111e-01 -3.62591028e-01 2.61552066e-01] | [6.978598594665527, 4.350953102111816] |
a4291695-520a-4835-9df6-e898f96a2c8f | glodyne-global-topology-preserving-dynamic | 2008.01935 | null | https://arxiv.org/abs/2008.01935v4 | https://arxiv.org/pdf/2008.01935v4.pdf | GloDyNE: Global Topology Preserving Dynamic Network Embedding | Learning low-dimensional topological representation of a network in dynamic environments is attracting much attention due to the time-evolving nature of many real-world networks. The main and common objective of Dynamic Network Embedding (DNE) is to efficiently update node embeddings while preserving network topology at each time step. The idea of most existing DNE methods is to capture the topological changes at or around the most affected nodes (instead of all nodes) and accordingly update node embeddings. Unfortunately, this kind of approximation, although can improve efficiency, cannot effectively preserve the global topology of a dynamic network at each time step, due to not considering the inactive sub-networks that receive accumulated topological changes propagated via the high-order proximity. To tackle this challenge, we propose a novel node selecting strategy to diversely select the representative nodes over a network, which is coordinated with a new incremental learning paradigm of Skip-Gram based embedding approach. The extensive experiments show GloDyNE, with a small fraction of nodes being selected, can already achieve the superior or comparable performance w.r.t. the state-of-the-art DNE methods in three typical downstream tasks. Particularly, GloDyNE significantly outperforms other methods in the graph reconstruction task, which demonstrates its ability of global topology preservation. The source code is available at https://github.com/houchengbin/GloDyNE | ['Ke Tang', 'Han Zhang', 'Chengbin Hou', 'Shan He'] | 2020-08-05 | null | null | null | null | ['graph-reconstruction'] | ['graphs'] | [-2.86779225e-01 3.26242633e-02 -1.75179020e-01 5.91115877e-02
6.86621852e-03 -6.16979361e-01 6.28383696e-01 4.05393600e-01
-1.86635450e-01 4.53130007e-01 3.29594880e-01 -1.80668890e-01
-5.07537961e-01 -1.11711609e+00 -4.41393882e-01 -8.74276400e-01
-5.72269499e-01 5.77215612e-01 5.51694274e-01 -3.44246447e-01
8.25886950e-02 8.09159994e-01 -9.13503945e-01 -3.46375823e-01
6.35260820e-01 5.98026752e-01 1.73230886e-01 5.67049980e-01
-2.03568012e-01 3.31268430e-01 -1.23680428e-01 -4.71355855e-01
2.48116642e-01 8.44003931e-02 -8.35498869e-01 -2.36543044e-01
3.13513339e-01 -2.03117803e-01 -1.26166260e+00 9.78722870e-01
5.35045028e-01 3.33883286e-01 4.05313611e-01 -1.33784342e+00
-5.36664903e-01 7.01134324e-01 -5.07314861e-01 5.43474436e-01
6.45972714e-02 5.90017028e-02 1.34835613e+00 -8.93740773e-01
8.60056043e-01 1.22922146e+00 7.66212821e-01 2.58534610e-01
-1.29504108e+00 -6.14082158e-01 5.81887722e-01 2.95928121e-01
-1.52505565e+00 -2.54555970e-01 1.18665266e+00 -1.98490396e-01
5.28554261e-01 2.66007960e-01 7.46572435e-01 1.08325648e+00
2.32614070e-01 4.40404743e-01 4.89890456e-01 1.58210099e-01
-5.68211861e-02 -2.86188781e-01 4.89690714e-02 8.52716565e-01
3.31925303e-01 -6.14302456e-02 -3.79150778e-01 -2.46350288e-01
6.98875606e-01 4.62563723e-01 -3.71168554e-01 -5.92117012e-01
-1.42089379e+00 6.37024879e-01 1.03768480e+00 6.09295964e-01
-3.56772125e-01 4.50754225e-01 7.78464019e-01 4.50392455e-01
5.28686881e-01 2.21493214e-01 -4.87558812e-01 -2.52334148e-01
-5.44937253e-01 -4.08896208e-02 6.65425479e-01 6.80072904e-01
7.36629546e-01 -1.34311229e-01 -8.36050063e-02 6.76123083e-01
2.50933379e-01 -4.13788995e-03 2.60379046e-01 -4.66409266e-01
5.75932801e-01 9.63717222e-01 -2.75293022e-01 -1.85582900e+00
-3.27076137e-01 -7.09345400e-01 -1.40232921e+00 -1.90658227e-01
1.70948684e-01 7.87300393e-02 -5.83012104e-01 1.88280916e+00
7.01528966e-01 7.26552188e-01 -3.03228378e-01 4.31156844e-01
7.19577134e-01 8.43040109e-01 -1.86397374e-01 -1.42041191e-01
1.03527236e+00 -9.44275975e-01 -4.80758160e-01 2.84148734e-02
6.94165349e-01 -3.45961481e-01 8.34681809e-01 -8.91559348e-02
-6.06300712e-01 -3.36825460e-01 -8.85589838e-01 -6.33240044e-02
-6.11764371e-01 -2.02155709e-01 6.92539394e-01 2.34147251e-01
-1.22076631e+00 8.14045846e-01 -9.95560110e-01 -4.80150521e-01
4.64002550e-01 3.49566519e-01 -6.52591169e-01 -2.07606450e-01
-1.28023541e+00 4.40499783e-01 3.98558587e-01 4.47264403e-01
-7.24424064e-01 -9.45784509e-01 -8.50572407e-01 3.06260228e-01
5.35106063e-01 -4.50285167e-01 5.66654980e-01 -3.63128275e-01
-1.21235824e+00 3.92580390e-01 -1.68081269e-01 -2.15970308e-01
5.51277757e-01 1.86744943e-01 -5.06698251e-01 2.85195619e-01
-8.19522217e-02 3.20075542e-01 5.96357584e-01 -1.18249130e+00
-2.32944518e-01 -3.16048890e-01 7.39819556e-02 1.76213130e-01
-1.03786814e+00 -4.84296948e-01 -8.91377091e-01 -8.02415371e-01
3.03909183e-01 -9.84468758e-01 -4.12491173e-01 4.32989329e-01
-3.83734405e-01 -3.85564774e-01 1.26397622e+00 -5.50556958e-01
1.70986676e+00 -1.96522141e+00 5.78200281e-01 4.56031203e-01
8.56546521e-01 4.25375700e-01 -3.92223179e-01 8.25274169e-01
-1.65643409e-01 3.94173056e-01 -1.45205110e-01 -2.90709525e-01
-8.52814689e-02 7.77394995e-02 -1.34757474e-01 5.67621052e-01
-1.87951863e-01 1.03805280e+00 -1.24332249e+00 -3.98646683e-01
4.20506835e-01 6.46414161e-01 -3.19848537e-01 -1.82193205e-01
1.16252869e-01 2.56980419e-01 -5.36570847e-01 3.71770859e-01
7.93704629e-01 -3.83101046e-01 5.00176609e-01 -5.72787941e-01
1.37452021e-01 3.11139822e-02 -1.26465976e+00 1.65486336e+00
-4.02170241e-01 6.99455917e-01 1.44902647e-01 -1.13409424e+00
8.63701105e-01 2.02183455e-01 8.82883072e-01 -5.76050460e-01
-3.07714827e-02 1.00356191e-01 -8.77390429e-03 -1.51953578e-01
2.47896045e-01 2.30740890e-01 -2.33109016e-02 4.25717771e-01
-8.12248588e-02 3.82785499e-01 2.54523516e-01 7.16871858e-01
1.59725368e+00 -5.33222198e-01 1.22693039e-01 -1.67624563e-01
6.17479146e-01 -5.37665308e-01 7.43329883e-01 3.90796363e-01
-3.59517276e-01 9.77946743e-02 8.25227559e-01 -6.39934957e-01
-9.53508914e-01 -1.00895000e+00 8.65245014e-02 8.88632357e-01
3.74700367e-01 -6.81443155e-01 -2.14518815e-01 -9.52639341e-01
3.72724146e-01 8.92025530e-02 -7.63604701e-01 -4.88556445e-01
-7.96258032e-01 -5.91935992e-01 4.49538440e-01 2.75484830e-01
3.60625893e-01 -7.46154249e-01 3.56059298e-02 3.66024882e-01
1.60070900e-02 -9.05541897e-01 -7.92543471e-01 -2.84286439e-01
-1.06738710e+00 -1.26968956e+00 -5.58121741e-01 -8.76950860e-01
8.94693077e-01 5.31531394e-01 9.11093235e-01 5.90035439e-01
-2.71407872e-01 2.51754493e-01 -4.81556952e-01 5.07216632e-01
-1.08187638e-01 4.83367264e-01 2.28946745e-01 1.31544068e-01
-1.54879287e-01 -1.07917345e+00 -8.05772483e-01 3.52176756e-01
-8.69225442e-01 -8.72348100e-02 6.33093059e-01 8.30280304e-01
6.03117347e-01 6.51385069e-01 4.96345252e-01 -1.00706208e+00
5.35522103e-01 -6.64730132e-01 -2.87943661e-01 2.63958395e-01
-6.39517844e-01 -1.24465339e-01 9.78258669e-01 -5.00333369e-01
-5.11648297e-01 -3.17974776e-01 8.66505317e-03 -5.98037720e-01
5.01138747e-01 6.53545141e-01 -2.00963959e-01 -3.57458264e-01
2.07850784e-01 2.53497213e-01 -1.65607035e-02 -6.01343989e-01
3.43118250e-01 2.43942454e-01 2.56275773e-01 -3.42935562e-01
1.29886830e+00 5.78986526e-01 3.33181798e-01 -7.06584930e-01
-3.72749388e-01 -5.05360246e-01 -7.54019082e-01 -2.11280033e-01
1.01036608e-01 -6.06046975e-01 -7.41040826e-01 4.79498178e-01
-8.51593435e-01 -2.82797217e-01 -1.41313478e-01 1.49526000e-01
8.43692385e-03 6.84934080e-01 -5.94423831e-01 -3.03235650e-01
-4.68118250e-01 -6.83743834e-01 6.29534662e-01 -9.84756555e-03
1.82576060e-01 -1.49193490e+00 2.35284045e-01 -1.41287193e-01
4.10481572e-01 6.04249477e-01 1.01947260e+00 -4.87064540e-01
-6.11087799e-01 -4.56720054e-01 -2.94675916e-01 3.13627198e-02
4.77736592e-01 2.37668753e-01 -2.40364537e-01 -8.50284100e-01
-5.91225266e-01 2.20900476e-01 9.24869001e-01 4.35117148e-02
1.25662708e+00 -5.15480518e-01 -7.84940660e-01 7.61490345e-01
1.54254806e+00 -3.75318289e-01 4.43145812e-01 8.57785344e-02
1.06548786e+00 4.40978825e-01 4.30916309e-01 3.77972275e-01
5.15208483e-01 7.20755219e-01 7.05027163e-01 3.59938778e-02
-2.81063497e-01 -6.36721551e-01 2.86734104e-01 1.27988577e+00
2.51463473e-01 -5.06835282e-01 -1.04439497e+00 8.08587015e-01
-1.96094501e+00 -9.41423237e-01 -1.11256100e-01 1.94673920e+00
5.85899830e-01 2.93755829e-01 4.52515781e-02 1.18640102e-01
9.90723968e-01 7.47919321e-01 -6.90319121e-01 5.12499064e-02
2.36880612e-02 -1.40897110e-01 5.02117515e-01 4.80049521e-01
-9.47979391e-01 9.05963659e-01 5.11231804e+00 8.98077250e-01
-1.04615772e+00 1.81767777e-01 4.41622585e-01 7.23696500e-02
-4.90832061e-01 3.24788466e-02 -4.50278014e-01 7.38847911e-01
8.05513322e-01 -4.43522185e-01 4.37659711e-01 5.82550406e-01
3.38578135e-01 4.87164170e-01 -8.54709446e-01 8.02540898e-01
-2.13802949e-01 -1.66524768e+00 3.12334627e-01 1.38264999e-01
7.36776590e-01 2.03868702e-01 -4.33373749e-02 1.87618211e-01
2.72712171e-01 -8.27208400e-01 2.02598676e-01 5.05500972e-01
9.04976368e-01 -7.94568896e-01 6.10189974e-01 1.94329828e-01
-1.97820759e+00 -2.27446914e-01 -4.42880780e-01 3.04533303e-01
2.36741245e-01 8.99248540e-01 -6.16873503e-01 7.55442798e-01
6.78213775e-01 1.24586082e+00 -6.96615100e-01 1.19193578e+00
-9.37334821e-02 5.37030578e-01 -4.89392847e-01 -7.41967559e-02
3.58120829e-01 -2.77488917e-01 1.03031468e+00 9.48887885e-01
3.23990971e-01 -2.11687982e-01 1.42379373e-01 5.22699475e-01
-4.51188445e-01 4.02021408e-03 -6.64630830e-01 -2.04363987e-01
1.05102992e+00 1.49117422e+00 -7.45864630e-01 1.61184184e-02
-1.70484796e-01 9.94667768e-01 6.24471843e-01 3.16805989e-01
-7.38145590e-01 -7.04164267e-01 8.66218746e-01 3.59717250e-01
3.71735156e-01 -4.94678319e-01 1.65351316e-01 -1.01212883e+00
3.39591444e-01 -5.04516304e-01 4.46513116e-01 -1.92800507e-01
-1.24033689e+00 5.36090434e-01 -3.73116404e-01 -1.22787559e+00
4.02094066e-01 -3.53408784e-01 -1.10184598e+00 3.44226629e-01
-1.43995905e+00 -1.25816309e+00 -6.06354773e-01 5.19177675e-01
3.06355029e-01 -1.38639500e-02 6.09222770e-01 6.66584730e-01
-9.02270019e-01 7.68395185e-01 5.12131095e-01 3.03058296e-01
4.46507126e-01 -1.21201611e+00 5.59342742e-01 9.29916799e-01
9.74852890e-02 7.52480149e-01 4.42500383e-01 -7.32965589e-01
-1.74808645e+00 -1.39625895e+00 8.20672572e-01 -1.89672813e-01
9.78415072e-01 -5.93339562e-01 -9.93977368e-01 6.86137617e-01
-3.14375728e-01 5.33205092e-01 1.94956601e-01 1.66713610e-01
-3.56909394e-01 -6.03535056e-01 -9.35410559e-01 8.20874393e-01
1.48171926e+00 -5.32780707e-01 -7.36966059e-02 4.42927897e-01
9.64654267e-01 -1.93524301e-01 -1.14865863e+00 4.68387634e-01
2.87821233e-01 -5.77948809e-01 1.02121806e+00 -5.60689926e-01
-6.06806763e-03 -5.44703186e-01 2.91171432e-01 -1.50533330e+00
-6.37282431e-01 -1.01558018e+00 -6.98836982e-01 1.41584074e+00
1.65800527e-01 -9.13347960e-01 9.74513888e-01 9.78495032e-02
6.91863447e-02 -1.07520509e+00 -1.25213146e+00 -7.25928664e-01
-8.73344913e-02 3.37418504e-02 7.04982638e-01 1.10858226e+00
-2.48031706e-01 1.47917002e-01 -2.49586746e-01 2.50007778e-01
5.41416705e-01 -2.15575844e-02 7.84912765e-01 -1.45835137e+00
3.23872492e-02 -5.12344539e-01 -8.66753578e-01 -1.10377216e+00
2.83999860e-01 -1.21194589e+00 -4.14936900e-01 -1.72289741e+00
7.86708891e-02 -7.91008949e-01 -5.85324943e-01 3.76001567e-01
-1.56259254e-01 -1.39892912e-02 2.25647707e-02 2.83334732e-01
-6.97773695e-01 9.66970563e-01 1.31802821e+00 -2.22744673e-01
-1.99246749e-01 -5.50949946e-02 -7.22472906e-01 3.29350650e-01
7.42819607e-01 -4.39469784e-01 -5.61405122e-01 -3.77044469e-01
4.37391967e-01 -2.80511826e-01 4.17790025e-01 -8.67223918e-01
5.35699308e-01 9.37598795e-02 5.68463355e-02 -5.17283261e-01
1.76431939e-01 -1.01511431e+00 3.51961821e-01 6.29259646e-01
-1.21798702e-01 3.98747236e-01 1.96599327e-02 1.27823544e+00
-1.42803699e-01 1.77184358e-01 7.30112731e-01 2.15557262e-01
-7.91877508e-01 1.12358081e+00 6.41256943e-02 1.10700196e-02
1.15419078e+00 -1.62723869e-01 -3.66438061e-01 -1.97058678e-01
-5.49528003e-01 6.40805602e-01 5.55789173e-01 5.33502936e-01
7.74579406e-01 -1.64408052e+00 -5.12702346e-01 -9.06062648e-02
6.78769648e-02 1.40491992e-01 5.11882842e-01 1.01120102e+00
-5.54028273e-01 1.18965164e-01 -3.13523971e-02 -2.09777206e-01
-1.21299207e+00 4.64284986e-01 3.66043746e-01 -6.54669702e-01
-1.04608035e+00 9.10670757e-01 -1.76428005e-01 -6.21363282e-01
2.63688236e-01 7.44455606e-02 -1.24351032e-01 1.50055632e-01
2.00124621e-01 7.00930655e-01 -5.93807437e-02 -6.13534451e-01
-4.87339169e-01 6.03384137e-01 -2.82655507e-01 4.08916920e-01
1.58498883e+00 -2.98563510e-01 -4.77887601e-01 3.55051219e-01
1.59784913e+00 1.95340291e-02 -1.09841657e+00 -6.05665922e-01
-1.28604203e-01 -6.38097107e-01 2.06948787e-01 -2.09978804e-01
-1.54563034e+00 6.36957943e-01 4.01648283e-01 2.84634113e-01
9.41066146e-01 -1.09224468e-01 1.09323835e+00 5.15026987e-01
2.63506651e-01 -9.57298219e-01 3.57034385e-01 4.95121896e-01
8.15004826e-01 -8.56891632e-01 1.71676815e-01 -5.20438850e-01
-2.19870195e-01 1.15617073e+00 5.89169204e-01 -1.94388911e-01
1.10139704e+00 -1.57428041e-01 -5.54492533e-01 -4.69946593e-01
-8.50744963e-01 1.26911774e-01 -1.61965825e-02 3.66806537e-01
-3.09501193e-04 5.43595329e-02 -1.29902706e-01 1.55838579e-02
1.01184718e-01 -6.33609056e-01 3.44393939e-01 5.59377551e-01
-2.55697995e-01 -1.22668898e+00 2.33392268e-01 7.05031037e-01
1.09414719e-02 4.84724268e-02 -5.27956605e-01 7.87171960e-01
-1.82821423e-01 5.47339559e-01 -2.90200450e-02 -6.60324216e-01
2.81677872e-01 -3.03745508e-01 -3.59442160e-02 -5.28455913e-01
-4.32419926e-01 -3.89874995e-01 -1.39502734e-01 -6.34928644e-01
-5.66744767e-02 -7.49846399e-01 -1.14736760e+00 -9.49279666e-01
-3.05916756e-01 9.61384773e-02 1.93205282e-01 3.68245393e-01
7.33205795e-01 5.85174978e-01 1.17171013e+00 -6.94985151e-01
-4.15760905e-01 -7.98535526e-01 -6.08205199e-01 5.07790864e-01
4.00251597e-01 -8.30500782e-01 -7.22204387e-01 -5.79447031e-01] | [7.172929286956787, 6.1025309562683105] |
9274aee1-f861-430a-9053-2f5bc1916603 | towards-explainable-music-emotion-recognition | 1907.03572 | null | https://arxiv.org/abs/1907.03572v1 | https://arxiv.org/pdf/1907.03572v1.pdf | Towards Explainable Music Emotion Recognition: The Route via Mid-level Features | Emotional aspects play an important part in our interaction with music. However, modelling these aspects in MIR systems have been notoriously challenging since emotion is an inherently abstract and subjective experience, thus making it difficult to quantify or predict in the first place, and to make sense of the predictions in the next. In an attempt to create a model that can give a musically meaningful and intuitive explanation for its predictions, we propose a VGG-style deep neural network that learns to predict emotional characteristics of a musical piece together with (and based on) human-interpretable, mid-level perceptual features. We compare this to predicting emotion directly with an identical network that does not take into account the mid-level features and observe that the loss in predictive performance of going through the mid-level features is surprisingly low, on average. The design of our network allows us to visualize the effects of perceptual features on individual emotion predictions, and we argue that the small loss in performance in going through the mid-level features is justified by the gain in explainability of the predictions. | ['Shreyan Chowdhury', 'Verena Haunschmid', 'Gerhard Widmer', 'Andreu Vall'] | 2019-07-08 | null | null | null | null | ['music-emotion-recognition'] | ['music'] | [ 8.83230940e-02 3.67229998e-01 3.26582432e-01 -3.31359148e-01
-4.11066152e-02 -6.14244103e-01 5.29788077e-01 3.19012582e-01
-5.91550693e-02 3.60693187e-01 5.39334714e-01 4.92360555e-02
-4.30572659e-01 -6.34794712e-01 -4.52686638e-01 -3.01139444e-01
-2.42046028e-01 2.09053040e-01 -1.62504748e-01 -4.89333749e-01
5.82107417e-02 4.13083106e-01 -1.85829675e+00 6.06613040e-01
4.79782224e-01 1.27441537e+00 -1.23526402e-01 7.54478931e-01
-3.61214876e-02 1.08953512e+00 -5.18209219e-01 -4.26606327e-01
1.86803058e-01 -7.63838172e-01 -8.02277625e-01 -1.29679054e-01
1.65478632e-01 1.37786537e-01 2.48229489e-01 7.60783315e-01
2.78561354e-01 2.60857910e-01 8.58133495e-01 -1.19460821e+00
-4.12081212e-01 7.32949853e-01 -2.45119184e-01 -3.98075670e-01
4.60370511e-01 1.54416338e-01 1.61024272e+00 -5.59696972e-01
5.70532262e-01 1.08246219e+00 9.54555213e-01 5.46140432e-01
-1.47735429e+00 -2.74687082e-01 1.07105337e-01 7.61839449e-02
-1.06633139e+00 -1.45805523e-01 9.52330709e-01 -6.53844535e-01
5.69201767e-01 6.32127404e-01 1.08804023e+00 8.41688335e-01
-6.02653529e-03 5.17856002e-01 1.00525987e+00 -3.55522454e-01
2.49752298e-01 1.35145038e-01 -7.43942186e-02 4.00569588e-01
-2.77137429e-01 2.17232719e-01 -6.55761421e-01 8.53660405e-02
7.24258840e-01 -1.23372614e-01 -2.63259858e-01 -3.24034154e-01
-1.19707835e+00 7.03734338e-01 8.19015563e-01 4.59836870e-01
-4.78000522e-01 3.93007755e-01 4.09204513e-01 4.64443445e-01
2.22826749e-01 1.23379576e+00 -6.58476412e-01 -5.21391690e-01
-1.02405536e+00 1.45737723e-01 7.99740076e-01 2.16571778e-01
6.17754221e-01 5.71686402e-02 -7.91160837e-02 7.23306954e-01
-4.84161377e-02 -2.40248665e-01 4.89875466e-01 -1.26539314e+00
-3.43575358e-01 5.66730678e-01 1.36054456e-01 -1.22693610e+00
-6.81157053e-01 -5.62838018e-01 -7.94425726e-01 8.56915891e-01
6.08283758e-01 -3.09696436e-01 -4.65135515e-01 2.10691261e+00
-2.01083392e-01 -1.10444061e-01 -9.84918997e-02 1.07051134e+00
7.80503154e-01 4.28837657e-01 1.49580553e-01 -9.43537951e-02
1.22292316e+00 -7.58170843e-01 -3.78883332e-01 -1.97720036e-01
3.93829912e-01 -6.80551469e-01 1.55405223e+00 6.95735693e-01
-1.03683746e+00 -8.98231745e-01 -1.09119582e+00 -2.22525761e-01
-2.78833628e-01 -3.07142753e-02 9.63408291e-01 2.34118566e-01
-9.73833025e-01 1.29475188e+00 -5.49629807e-01 -2.84251571e-01
9.16615948e-02 5.86009145e-01 -3.23620051e-01 7.88248062e-01
-1.09062421e+00 9.23760235e-01 2.36055106e-01 1.54972856e-03
-1.33675963e-01 -5.91096997e-01 -4.90565151e-01 5.43965995e-01
1.78722322e-01 -6.91910803e-01 1.14110148e+00 -1.54363143e+00
-1.52071536e+00 6.11135960e-01 1.23618312e-01 -7.37016946e-02
4.39080626e-01 -1.27947628e-01 -1.17712505e-01 -2.01853916e-01
-4.73120213e-01 1.00832021e+00 5.01609862e-01 -1.29051113e+00
-2.28193626e-01 -1.44963682e-01 3.09484094e-01 2.36939207e-01
-3.27166557e-01 -3.17311853e-01 3.34834307e-02 -7.35189974e-01
1.94321573e-01 -1.05066156e+00 -2.93111145e-01 1.70191631e-01
-3.87326688e-01 1.12078190e-02 1.59568295e-01 -4.93460685e-01
1.00086308e+00 -2.33383346e+00 3.76501292e-01 1.94138333e-01
2.28838623e-01 -2.67550170e-01 -1.26184344e-01 4.27069128e-01
-3.51106048e-01 2.39004195e-01 3.99152003e-02 -1.61417350e-01
3.59503776e-01 -5.37677482e-02 -3.98336112e-01 -2.31119633e-01
1.62442312e-01 8.87552619e-01 -6.63687646e-01 -9.73180234e-02
1.59651265e-01 6.11918986e-01 -9.24478173e-01 1.33885592e-01
-3.68360072e-01 5.68984687e-01 -3.03864032e-01 8.22055340e-02
2.04136353e-02 -5.05084023e-02 1.48036942e-01 -4.24011886e-01
3.47866379e-02 3.79750013e-01 -1.07863188e+00 1.59386814e+00
-6.05006635e-01 9.59649146e-01 -1.85045511e-01 -6.99919224e-01
1.00630987e+00 4.37514037e-01 4.56107616e-01 -4.22288269e-01
1.81712180e-01 6.56846017e-02 4.10587639e-01 -3.92221898e-01
7.26150930e-01 -9.41716254e-01 -2.93540776e-01 6.13276184e-01
-1.32538915e-01 -3.57209325e-01 -2.56850600e-01 -5.62033355e-02
7.61623859e-01 2.76870757e-01 2.79852629e-01 -1.06775068e-01
1.92258403e-01 -1.45975605e-01 4.49233264e-01 6.04460478e-01
3.88166867e-02 7.45034635e-01 8.16736400e-01 -6.69312716e-01
-9.06814635e-01 -9.26669121e-01 2.68273145e-01 1.16373575e+00
1.35505786e-02 -6.30132794e-01 -6.00135744e-01 -3.18973005e-01
-1.21277114e-02 7.37712681e-01 -9.45643425e-01 -3.97391409e-01
-3.22018117e-02 -3.55606228e-01 3.52187067e-01 5.81210792e-01
-5.39610237e-02 -1.52017307e+00 -8.15753102e-01 2.30509192e-01
-1.00206971e-01 -6.82757914e-01 1.37026444e-01 6.89016879e-01
-8.36430371e-01 -6.22260809e-01 -1.85105547e-01 -3.99043471e-01
2.24326149e-01 -4.23138052e-01 1.38774014e+00 3.16686481e-01
-3.07463378e-01 -6.48572147e-02 -3.23594362e-01 -5.99102259e-01
-2.93720633e-01 -2.55818330e-02 -2.16800794e-02 -2.32848734e-01
3.04184496e-01 -1.05632007e+00 -5.78076005e-01 6.23237453e-02
-8.33862364e-01 2.18017429e-01 3.61894220e-01 8.33636403e-01
3.55135113e-01 -5.07066201e-04 4.71375763e-01 -6.69748902e-01
7.16855466e-01 -6.35247156e-02 1.56483620e-01 -1.63227901e-01
-3.87898803e-01 2.51280606e-01 6.90733433e-01 -5.59986353e-01
-6.31407857e-01 7.09014386e-02 -3.54438990e-01 -1.85571879e-01
-2.56545454e-01 6.91228688e-01 3.95637415e-02 1.50286451e-01
7.81025708e-01 -3.04246694e-01 -6.80140033e-02 -6.04548514e-01
6.68545723e-01 3.37718427e-01 7.05689549e-01 -5.57082713e-01
6.01955175e-01 1.26261234e-01 1.42960832e-01 -4.42283601e-01
-1.14399207e+00 9.81331468e-02 -6.98128402e-01 -3.19113046e-01
9.27877009e-01 -6.41271234e-01 -1.17359769e+00 -1.04586512e-01
-8.72075319e-01 -5.05061865e-01 -8.64698589e-01 4.11565125e-01
-1.07973313e+00 2.78504994e-02 -7.22448409e-01 -8.60382676e-01
-3.82817723e-02 -9.43538249e-01 8.21867347e-01 1.06712468e-01
-1.24617648e+00 -8.57040882e-01 -1.89585462e-02 1.65324494e-01
3.43145579e-01 6.22214854e-01 1.20832849e+00 -5.68891466e-01
-2.09532961e-01 -1.18210480e-01 -1.00546166e-01 2.75413543e-01
7.98693746e-02 1.32569969e-01 -1.26211774e+00 2.23565534e-01
3.05576883e-02 -5.71290433e-01 8.86680901e-01 2.47836411e-01
1.33162570e+00 -1.56171545e-01 1.40751749e-01 5.80780327e-01
1.18280220e+00 -1.85725078e-01 5.95523655e-01 3.11277360e-01
6.14844143e-01 8.05087447e-01 4.47091222e-01 4.19136554e-01
4.40550968e-02 9.42911744e-01 5.26280820e-01 -3.99714440e-01
-2.35976558e-02 -3.27341318e-01 1.62461385e-01 5.98216116e-01
-4.45247442e-01 8.82743075e-02 -4.99362081e-01 1.75907671e-01
-1.89931250e+00 -1.17446995e+00 -6.15999810e-02 2.08463931e+00
8.07605088e-01 2.62866497e-01 2.48768672e-01 5.39065719e-01
1.99359104e-01 1.06282793e-01 -4.66609776e-01 -1.00219774e+00
-5.70859537e-02 2.85442322e-01 -2.73221135e-01 4.31742907e-01
-8.56964827e-01 8.18871737e-01 6.79762602e+00 5.54625273e-01
-1.40335763e+00 -4.21627283e-01 6.92417562e-01 -2.98488557e-01
-4.61221099e-01 -1.25104025e-01 2.83509523e-01 1.98734179e-02
6.57187045e-01 -9.34565719e-03 7.36808717e-01 7.91292012e-01
3.42736781e-01 5.59872501e-02 -1.54463899e+00 7.59990633e-01
-3.15700859e-01 -1.09396672e+00 9.42017958e-02 4.65051606e-02
4.45722699e-01 -4.32831675e-01 1.74956739e-01 2.84391671e-01
5.74456789e-02 -1.53593087e+00 1.12788868e+00 8.31820905e-01
4.74220574e-01 -9.36603606e-01 5.10242224e-01 4.16441530e-01
-8.13162625e-01 -1.58102021e-01 -2.26978227e-01 -9.38389301e-01
-1.18482769e-01 4.06996906e-01 -6.28933489e-01 1.37564674e-01
5.58838129e-01 5.10767043e-01 -4.19942826e-01 7.68628359e-01
-3.03993255e-01 4.89270866e-01 -1.83164507e-01 -1.16896123e-01
8.58766809e-02 -1.67238995e-01 3.61523032e-01 1.07983196e+00
5.22754312e-01 8.41358379e-02 -4.45438102e-02 1.18060565e+00
-1.44127151e-02 3.36665273e-01 -3.35088223e-01 -2.01937169e-01
2.40625627e-02 1.41337585e+00 -7.78687119e-01 -1.91087633e-01
-8.13665465e-02 1.08993828e+00 4.16486800e-01 1.61636963e-01
-5.12123048e-01 -3.09793502e-01 9.49697912e-01 1.06468461e-01
3.40170532e-01 -6.63372800e-02 -7.95697510e-01 -1.04624772e+00
-1.24872327e-01 -8.63074183e-01 6.32375628e-02 -1.30606508e+00
-1.21804523e+00 7.38645911e-01 -7.10437536e-01 -1.19556916e+00
-5.85571766e-01 -5.29596090e-01 -8.08351457e-01 1.04015541e+00
-9.13206100e-01 -1.00369430e+00 -2.99951136e-01 1.71867475e-01
1.63446873e-01 2.65676975e-01 1.20263600e+00 -2.29439408e-01
3.16679068e-02 5.90417206e-01 -2.10174844e-01 1.16173297e-01
6.37605906e-01 -1.56545138e+00 8.95083919e-02 1.71014696e-01
6.84055209e-01 6.30909204e-01 1.35088670e+00 -7.00272024e-02
-7.23779917e-01 -5.32466292e-01 9.48020339e-01 -6.19743824e-01
6.14654124e-01 -3.18861574e-01 -8.15119863e-01 4.87923324e-01
1.07291780e-01 -2.67310917e-01 1.00426376e+00 7.62436509e-01
-4.74845111e-01 9.90934297e-02 -9.10819769e-01 7.32296228e-01
8.89093697e-01 -6.30460441e-01 -7.45049894e-01 -1.93862662e-01
5.49065411e-01 -4.99509834e-02 -1.11349833e+00 4.14235562e-01
1.12721157e+00 -1.44067717e+00 1.01916194e+00 -7.45327115e-01
1.06739640e+00 -3.28618556e-01 -8.35756399e-03 -1.51765871e+00
-5.84194601e-01 -3.36924970e-01 3.10108155e-01 1.09757495e+00
7.01946080e-01 -4.30962164e-03 8.81947517e-01 6.73101664e-01
1.09211177e-01 -9.09537613e-01 -4.81896311e-01 -4.55915540e-01
2.12615848e-01 -7.79820979e-01 6.72608078e-01 9.88119364e-01
6.65127277e-01 5.57465374e-01 -5.93601763e-01 -2.31064484e-01
8.38265643e-02 6.44292653e-01 7.18774557e-01 -1.66597176e+00
-9.20837402e-01 -7.81688929e-01 -7.14373410e-01 -6.81255281e-01
-7.22475126e-02 -8.63771498e-01 1.33629665e-01 -1.43230319e+00
2.84011304e-01 -3.57864588e-01 -5.47113657e-01 5.84359050e-01
-2.43910193e-01 7.90535033e-01 6.49819613e-01 5.37466258e-02
-4.02970850e-01 5.67916393e-01 1.28633130e+00 -3.35976370e-02
-2.81135648e-01 -5.21843359e-02 -1.03209960e+00 1.12681842e+00
6.72097921e-01 -1.43350720e-01 -2.70480931e-01 -8.99993852e-02
7.10644245e-01 1.27232745e-01 5.46281874e-01 -1.27509487e+00
-1.83631316e-01 -1.11618407e-01 8.73385131e-01 -9.72065777e-02
7.04638481e-01 -7.22264588e-01 3.84129733e-01 3.25556129e-01
-9.13866401e-01 -3.12232107e-01 1.77754149e-01 2.66265869e-01
-3.87325794e-01 -1.13684379e-01 4.95171010e-01 -9.65367630e-02
-7.29808331e-01 -1.76463470e-01 -3.20160687e-01 -2.26371929e-01
3.73162448e-01 -4.52687919e-01 2.49663979e-01 -9.14512575e-01
-1.45321918e+00 -2.57340223e-01 7.12023795e-01 5.29079497e-01
9.85230953e-02 -1.40317619e+00 -3.70764196e-01 -9.11083259e-03
2.35438481e-01 -4.44855392e-01 9.20811519e-02 5.95737934e-01
-2.06036538e-01 -4.95876325e-03 -6.23294055e-01 -3.26338768e-01
-1.33640742e+00 5.11047781e-01 3.80340517e-01 -2.00046360e-01
-5.44755518e-01 9.04419541e-01 3.59829545e-01 -4.04035896e-01
1.43272385e-01 -4.06682134e-01 -4.23889548e-01 1.80280313e-01
2.41140783e-01 -1.19497418e-01 -2.50662506e-01 -6.11560643e-01
-1.01179264e-01 7.08185792e-01 5.11162281e-01 -3.25759888e-01
1.66888630e+00 3.29082310e-02 -5.44795096e-02 1.01035523e+00
8.00779581e-01 2.58700222e-01 -1.28341246e+00 2.87799060e-01
1.16017744e-01 -1.32041588e-01 -4.43725139e-02 -1.25935483e+00
-9.10331845e-01 1.17563736e+00 2.76791304e-01 5.16317129e-01
1.23353863e+00 -1.18151354e-02 5.98783553e-01 2.96443909e-01
2.42516384e-01 -9.11989033e-01 1.33593127e-01 4.16357875e-01
9.63466823e-01 -9.22547817e-01 -9.56535935e-02 -3.92582834e-01
-7.44327843e-01 1.32105315e+00 4.42038506e-01 -1.41845137e-01
3.90515208e-01 1.19488798e-01 3.50116313e-01 -3.19929093e-01
-9.42572474e-01 -1.75055042e-01 6.71362758e-01 3.27643991e-01
1.04336131e+00 2.88344592e-01 -1.31733477e-01 1.01616263e+00
-9.90290403e-01 5.16307987e-02 5.29476762e-01 3.50941032e-01
-5.44200361e-01 -1.10846770e+00 -8.77929404e-02 4.25917923e-01
-5.66108823e-01 -8.94012302e-02 -8.44376743e-01 7.02744126e-01
5.87104678e-01 8.06124091e-01 1.16399087e-01 -8.63275528e-01
2.87108868e-01 2.65298218e-01 3.83207589e-01 -4.68113184e-01
-7.70752430e-01 -1.05911093e-02 2.18773097e-01 -6.00085557e-01
-2.60349780e-01 -4.88360465e-01 -1.39584422e+00 -1.59456834e-01
8.51055235e-02 3.23550582e-01 6.51976943e-01 8.74838710e-01
1.57562047e-01 8.17331612e-01 5.00231862e-01 -1.12554002e+00
-2.76865840e-01 -9.09056604e-01 -7.62405157e-01 8.76120567e-01
2.80825078e-01 -5.59554100e-01 -4.23034102e-01 1.06367558e-01] | [15.868280410766602, 5.394870281219482] |
97130c44-60f6-4f8a-a05d-4eccae96d5a6 | a-dynamic-graph-interactive-framework-with | 2211.04023 | null | https://arxiv.org/abs/2211.04023v1 | https://arxiv.org/pdf/2211.04023v1.pdf | A Dynamic Graph Interactive Framework with Label-Semantic Injection for Spoken Language Understanding | Multi-intent detection and slot filling joint models are gaining increasing traction since they are closer to complicated real-world scenarios. However, existing approaches (1) focus on identifying implicit correlations between utterances and one-hot encoded labels in both tasks while ignoring explicit label characteristics; (2) directly incorporate multi-intent information for each token, which could lead to incorrect slot prediction due to the introduction of irrelevant intent. In this paper, we propose a framework termed DGIF, which first leverages the semantic information of labels to give the model additional signals and enriched priors. Then, a multi-grain interactive graph is constructed to model correlations between intents and slots. Specifically, we propose a novel approach to construct the interactive graph based on the injection of label semantics, which can automatically update the graph to better alleviate error propagation. Experimental results show that our framework significantly outperforms existing approaches, obtaining a relative improvement of 13.7% over the previous best model on the MixATIS dataset in overall accuracy. | ['Yuexian Zou', 'Tengtao Song', 'Xuxin Cheng', 'Weiyuan Xu', 'Zhihong Zhu'] | 2022-11-08 | null | null | null | null | ['spoken-language-understanding', 'intent-detection', 'slot-filling', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'speech'] | [ 4.07580495e-01 6.55656517e-01 -5.41088998e-01 -6.60538852e-01
-7.97776282e-01 -1.74599186e-01 4.48783785e-01 2.90128767e-01
-2.83138216e-01 5.76498747e-01 3.99053782e-01 -1.99946299e-01
1.52426928e-01 -4.94259745e-01 -4.78740126e-01 -3.17453504e-01
1.12719357e-01 7.04278469e-01 4.09584224e-01 5.03816828e-02
2.28171758e-02 -3.94560069e-01 -1.24843681e+00 2.40178823e-01
9.99084830e-01 9.42901909e-01 3.44623476e-01 5.43729328e-02
-4.18020368e-01 9.36294615e-01 -3.48585784e-01 -3.81092787e-01
-1.38620362e-01 -2.97642380e-01 -8.43555987e-01 2.62282342e-01
-1.13760926e-01 -2.45189816e-01 -1.54011130e-01 1.10915005e+00
1.94261178e-01 1.86042473e-01 3.93186420e-01 -1.17823458e+00
-2.05283433e-01 1.14389431e+00 -7.33802974e-01 -1.30085170e-01
2.03535900e-01 -2.88733661e-01 1.48060465e+00 -8.13428342e-01
4.52270657e-01 1.47151315e+00 6.84162557e-01 3.57836187e-01
-1.28677166e+00 -7.18527615e-01 6.00674510e-01 2.71661967e-01
-1.34007657e+00 -4.75975156e-01 9.27669346e-01 -1.18620194e-01
9.12162721e-01 1.82431698e-01 2.80194223e-01 9.18289125e-01
-3.84219110e-01 1.17416859e+00 1.02520525e+00 -6.52522087e-01
2.18793795e-01 8.29844251e-02 4.53839213e-01 6.93197429e-01
-9.57133546e-02 -2.88567066e-01 -6.37905121e-01 -3.27851176e-01
3.23422760e-01 2.34486796e-02 9.11757126e-02 -3.15209180e-01
-1.04664207e+00 8.76696587e-01 2.29038179e-01 3.20154816e-01
-4.83211935e-01 6.02472089e-02 3.51345956e-01 -1.18524216e-01
7.96220005e-01 2.93004662e-01 -6.15195096e-01 -1.30853415e-01
-7.62750447e-01 2.47491449e-01 6.18725657e-01 9.39525306e-01
1.16494799e+00 -4.11814779e-01 -4.80280191e-01 1.26167965e+00
8.06117058e-01 1.16193347e-01 3.47184837e-01 -7.79807568e-01
5.79227269e-01 7.19569743e-01 1.34212270e-01 -9.44875717e-01
-7.65163124e-01 -5.09901702e-01 -3.90007108e-01 -7.96598017e-01
1.57306895e-01 -4.82597910e-02 -1.03443861e+00 2.01325512e+00
6.00694716e-01 5.08549154e-01 -1.12259559e-01 7.77661800e-01
5.94282925e-01 5.13553083e-01 5.75625420e-01 -5.35379648e-02
1.64784956e+00 -1.20308876e+00 -1.20598722e+00 -7.13024735e-01
1.01135576e+00 -6.97641015e-01 9.60989714e-01 -6.96245581e-02
-6.09494746e-01 -3.91243368e-01 -8.06017697e-01 -6.31916076e-02
-2.17059359e-01 1.91491634e-01 9.35096264e-01 5.79026103e-01
-7.81910181e-01 2.22249866e-01 -7.95912087e-01 -3.02231342e-01
1.15455352e-01 2.03565225e-01 1.08445277e-02 -2.18825802e-01
-1.51068759e+00 5.97090244e-01 6.61062419e-01 -6.90624192e-02
-4.81087297e-01 -6.68385327e-01 -1.14309657e+00 8.62585455e-02
9.89319026e-01 -3.82580847e-01 1.43277466e+00 -5.69688141e-01
-1.38804424e+00 5.27978897e-01 -5.16213834e-01 -3.23202461e-01
1.40715055e-02 -2.61624157e-01 -3.20377946e-01 -2.92039394e-01
4.72263902e-01 8.35758388e-01 5.09454608e-01 -1.34451461e+00
-8.45796347e-01 -3.15440148e-01 1.66906312e-01 4.28892344e-01
-3.19796711e-01 -1.34430856e-01 -1.00451875e+00 -6.56927347e-01
6.22304022e-01 -9.98805463e-01 -2.58640528e-01 -5.60554445e-01
-8.22099984e-01 -5.07716238e-01 6.65850759e-01 -7.19185054e-01
1.47737682e+00 -2.16843534e+00 -1.33899674e-01 1.96800977e-01
2.66111791e-01 3.36132497e-02 -7.17471391e-02 3.70130748e-01
2.04793319e-01 2.42035184e-02 -2.06992075e-01 -8.58466685e-01
3.76248688e-01 4.47313696e-01 -2.71508485e-01 5.29696196e-02
1.12603202e-01 9.18879449e-01 -9.41832721e-01 -6.26128256e-01
7.21809044e-02 1.77831635e-01 -7.84385026e-01 2.10176706e-01
-5.07213056e-01 3.84934932e-01 -6.30308986e-01 6.25109017e-01
7.51636267e-01 -5.59753776e-01 7.62404263e-01 -3.06575298e-01
2.38554373e-01 8.72168481e-01 -1.08355224e+00 1.79906797e+00
-4.31269288e-01 7.14042857e-02 -5.42569906e-02 -1.03085744e+00
8.43572497e-01 3.17072839e-01 5.21425843e-01 -7.17900574e-01
4.68544923e-02 6.82235956e-02 -3.58847529e-01 -1.73215985e-01
6.85404122e-01 -6.77686259e-02 -3.45805258e-01 4.65414077e-01
3.90788764e-02 3.28858614e-01 6.45859912e-02 2.95664966e-01
1.09207571e+00 5.85252680e-02 2.18610346e-01 3.31857082e-05
3.38580728e-01 -2.75482267e-01 9.95323479e-01 9.25755262e-01
-2.74179786e-01 2.20411777e-01 5.34997940e-01 -8.73942152e-02
-6.01819992e-01 -7.65813470e-01 5.13311513e-02 1.40261292e+00
3.60005587e-01 -7.37617016e-01 -5.51050544e-01 -1.02048814e+00
-8.92773718e-02 8.55667412e-01 -5.04642069e-01 -8.42011794e-02
-4.06133235e-01 -9.12680030e-01 4.37832743e-01 4.60142821e-01
5.69058836e-01 -8.05623412e-01 -6.05546534e-02 4.02170062e-01
-8.10306847e-01 -1.45028305e+00 -3.01595420e-01 4.65591043e-01
-7.08509743e-01 -7.71140099e-01 -2.09091261e-01 -5.34016430e-01
5.80505669e-01 2.26810470e-01 1.06531572e+00 2.20411584e-01
1.53114051e-01 9.55051854e-02 -5.90331018e-01 -8.64906237e-02
-2.18736395e-01 2.43691787e-01 -1.71224356e-01 1.12325087e-01
6.64477229e-01 -3.67426038e-01 -3.19503009e-01 3.97889137e-01
-8.40642750e-01 3.71026188e-01 5.86994469e-01 1.06298947e+00
5.21287680e-01 1.51155621e-01 7.33261049e-01 -1.41593313e+00
3.55670422e-01 -7.30080664e-01 -3.08842957e-01 2.96329081e-01
-8.38240862e-01 2.43304253e-01 2.20348179e-01 -1.43003762e-01
-1.40335393e+00 4.99225706e-02 -2.46325552e-01 -1.77176818e-02
-1.32672802e-01 7.25582182e-01 -2.40624696e-01 3.55324835e-01
6.10035062e-02 3.35632004e-02 -5.02990007e-01 -7.06153452e-01
4.60692674e-01 6.77322388e-01 2.25168854e-01 -6.84000015e-01
4.24533069e-01 3.37912053e-01 -4.11024779e-01 -5.33207178e-01
-1.37419581e+00 -8.62722635e-01 -4.22334254e-01 -9.92055144e-03
6.14053667e-01 -1.09566891e+00 -2.23698482e-01 3.46896440e-01
-1.09240675e+00 -2.31344834e-01 -1.79310013e-02 5.60429394e-01
-3.29189986e-01 5.75277448e-01 -8.50806296e-01 -9.52223599e-01
6.80023730e-02 -1.06797612e+00 1.28557396e+00 2.49799088e-01
-2.63320565e-01 -1.01671946e+00 -1.83906294e-02 6.80269718e-01
2.40423918e-01 -4.33447897e-01 1.10071993e+00 -9.64814663e-01
-5.67292154e-01 -7.04306923e-03 -4.81923312e-01 -1.33302972e-01
2.53575206e-01 -5.46423972e-01 -1.18160248e+00 -3.55008394e-02
-2.47387066e-01 -4.52715546e-01 9.36811447e-01 2.05279842e-01
8.92901182e-01 -1.62803799e-01 -8.19961786e-01 9.41067711e-02
1.03404939e+00 1.27174631e-01 3.04042518e-01 1.67007759e-01
6.43192530e-01 6.87203348e-01 1.02010190e+00 6.78316891e-01
9.91286397e-01 1.09812784e+00 2.93359488e-01 -2.23598734e-01
-4.18084934e-02 -6.65657878e-01 4.04310785e-02 1.02960789e+00
4.94282544e-01 -2.73612380e-01 -6.81783915e-01 5.98147571e-01
-2.24280977e+00 -3.84548783e-01 1.39649257e-01 1.79381132e+00
9.87591624e-01 2.69961536e-01 -9.20382440e-02 -2.39155404e-02
8.67600918e-01 3.99754047e-01 -4.56662595e-01 1.37916714e-01
1.48177877e-01 -9.90176052e-02 3.55366111e-01 4.83576447e-01
-1.30119896e+00 1.38718843e+00 6.14885902e+00 8.86038184e-01
-6.10389888e-01 3.53040695e-01 6.82299912e-01 5.22192121e-01
-5.23240566e-01 2.98199922e-01 -1.05689609e+00 5.00197113e-01
6.84760034e-01 2.67830998e-01 1.04359411e-01 8.55968595e-01
-1.40283396e-02 -2.40763560e-01 -8.47600222e-01 5.86527646e-01
-1.24221072e-01 -8.39351177e-01 -2.02796400e-01 -5.37273735e-02
6.19477212e-01 -1.99495137e-01 -1.63254276e-01 7.72398651e-01
6.15800917e-01 -3.78039151e-01 6.17867529e-01 3.71150792e-01
2.56516337e-01 -5.93702555e-01 7.92290986e-01 4.42219943e-01
-1.24811018e+00 8.27479139e-02 -2.92340100e-01 -7.68580958e-02
2.79509127e-01 9.24490988e-01 -1.19820178e+00 6.73023939e-01
4.54815149e-01 6.12441182e-01 -4.23653185e-01 8.55424583e-01
-3.23812276e-01 8.82332206e-01 -3.77741992e-01 2.66493801e-02
2.87131041e-01 1.32523105e-01 3.10222298e-01 1.21755707e+00
1.34057831e-02 -9.47391167e-02 5.55153072e-01 8.21397483e-01
-1.16704628e-01 3.26437652e-01 -2.75285304e-01 -4.87873964e-02
8.23450625e-01 1.26081073e+00 -9.51663554e-01 -4.09170181e-01
-6.62894070e-01 9.53246891e-01 6.01882458e-01 3.58037204e-01
-9.64481294e-01 -2.72647683e-02 6.27887726e-01 -3.99028957e-01
4.46202606e-01 -1.07801091e-02 -2.72650361e-01 -1.08866942e+00
-4.17548791e-02 -4.24690217e-01 4.96454507e-01 -4.50170577e-01
-1.08640027e+00 2.80440241e-01 3.35298106e-02 -6.03669822e-01
-3.83876771e-01 -1.27402917e-01 -2.57002741e-01 6.02913857e-01
-1.73379242e+00 -1.14702880e+00 -1.47474427e-02 7.46013820e-02
6.15467846e-01 2.23305091e-01 8.62627864e-01 5.13109505e-01
-6.84869528e-01 7.81140506e-01 -1.17162943e-01 9.53462813e-03
6.86637998e-01 -1.05992699e+00 3.53468955e-01 7.66645730e-01
3.75806153e-01 5.25518537e-01 6.52085960e-01 -8.32870841e-01
-8.78751278e-01 -1.13369894e+00 1.30372107e+00 -8.71488452e-02
5.29800236e-01 -4.44197536e-01 -8.83509278e-01 8.59907925e-01
-1.26961708e-01 -2.17164978e-01 8.38462412e-01 6.91703916e-01
-5.43086529e-01 1.98611930e-01 -8.87859166e-01 4.80897695e-01
1.04528379e+00 -5.66814542e-01 -5.31509995e-01 3.43666643e-01
1.15463042e+00 -3.98100495e-01 -6.01531088e-01 5.65309048e-01
2.55891234e-01 -8.03347349e-01 7.92666733e-01 -3.33826810e-01
-2.71100327e-02 -3.57762426e-01 -2.93753773e-01 -1.14540970e+00
-3.28268439e-01 -3.38385910e-01 -2.16938496e-01 1.45545542e+00
6.42983019e-01 -5.05560279e-01 8.10716927e-01 6.84233248e-01
-2.36930698e-01 -6.81704104e-01 -8.43770981e-01 -4.65107858e-01
-7.44089305e-01 -7.42947042e-01 7.20787644e-01 1.05463672e+00
2.52926558e-01 5.46051204e-01 -6.90790713e-01 2.92246252e-01
5.15596807e-01 1.60721749e-01 4.92582858e-01 -1.14195526e+00
-4.79320735e-01 -7.69063225e-03 -1.21614851e-01 -1.55522180e+00
5.30145884e-01 -7.66436398e-01 5.06045699e-01 -1.43098593e+00
3.32167804e-01 -1.03353441e+00 -5.69687486e-01 8.73139977e-01
-6.33980930e-01 2.48304680e-01 5.20396940e-02 1.50120094e-01
-1.29256022e+00 6.68736994e-01 8.17558706e-01 -5.01132868e-02
-1.62230924e-01 2.69228816e-02 -7.94693708e-01 6.74185693e-01
6.12600982e-01 -6.46996558e-01 -4.93370444e-01 -2.74412870e-01
1.38396740e-01 3.15865427e-02 7.52340555e-02 -7.51172304e-01
1.08552925e-01 -4.01984490e-02 -1.10626355e-01 -6.55350626e-01
5.29728115e-01 -7.86901951e-01 7.78861791e-02 2.76379474e-02
-5.15149176e-01 -4.72214133e-01 4.82842624e-02 8.73867929e-01
-2.36144155e-01 -3.89512062e-01 3.72268975e-01 6.30394295e-02
-9.19129670e-01 1.31995436e-02 -2.29927257e-01 7.46046230e-02
7.30595887e-01 3.65905225e-01 -2.90687412e-01 -3.87381315e-01
-7.76074886e-01 3.89317423e-01 1.99512973e-01 5.64315915e-01
1.79168850e-01 -1.40043807e+00 -1.65555477e-01 2.31587067e-01
4.00637269e-01 4.12082439e-03 3.89264822e-01 7.49616027e-01
3.07046950e-01 7.71275878e-01 4.06823546e-01 -5.65375686e-01
-9.68196929e-01 2.24785835e-01 -8.52477476e-02 -7.58598506e-01
-5.31009018e-01 8.80891800e-01 3.92807364e-01 -7.85405934e-01
3.61638099e-01 -2.66362518e-01 -2.46915072e-01 4.26747650e-02
2.81580329e-01 1.29371155e-02 3.83476578e-02 -6.58158362e-01
-3.64893377e-01 1.14612952e-01 -4.69681770e-01 -7.49475583e-02
1.02258766e+00 -5.17558217e-01 -1.34679973e-02 5.34105062e-01
8.93697977e-01 -1.19857848e-01 -1.21466637e+00 -8.60004365e-01
5.89639425e-01 -4.10461009e-01 8.41043368e-02 -8.96797001e-01
-8.71086836e-01 4.88917977e-01 2.78779536e-01 3.03019971e-01
9.09906745e-01 3.51806551e-01 9.45868433e-01 2.56987721e-01
3.96615714e-01 -1.16825151e+00 8.44157264e-02 6.54453576e-01
1.25749797e-01 -1.22875726e+00 -2.53683150e-01 -8.68205905e-01
-7.31299877e-01 6.12764478e-01 6.83617592e-01 5.05786419e-01
5.51547825e-01 1.00580849e-01 1.61188379e-01 -2.48400569e-01
-8.37221086e-01 -4.17720079e-01 1.45294845e-01 4.06742185e-01
6.10634804e-01 2.53778398e-01 -4.39510852e-01 8.69521379e-01
1.23364344e-01 -1.48035914e-01 -5.39241694e-02 9.49671626e-01
-5.75073838e-01 -1.45816755e+00 -9.04884562e-02 5.18252730e-01
-3.12975526e-01 -2.69905120e-01 -8.93559307e-02 4.24699485e-01
8.46582502e-02 1.01759124e+00 -2.74185576e-02 -4.96994317e-01
1.69070169e-01 5.26270568e-01 4.62477989e-02 -8.27552080e-01
-1.21416524e-01 3.57914835e-01 4.18060601e-01 -6.33821785e-01
-5.51023960e-01 -6.69026613e-01 -1.38024867e+00 1.72035947e-01
-7.01297343e-01 3.48645538e-01 6.90841854e-01 1.28635716e+00
5.23555279e-01 6.93588555e-01 4.40387279e-01 -5.54226756e-01
-3.87671262e-01 -1.17858422e+00 -6.95747554e-01 3.23738873e-01
-5.58595471e-02 -1.11131060e+00 -2.83415705e-01 -1.76915839e-01] | [12.514104843139648, 7.2993388175964355] |
dbe03d0d-efb2-4e3a-a70f-eec78fd611cd | an-experimental-study-in-real-time-facial-1 | 2304.03064 | null | https://arxiv.org/abs/2304.03064v1 | https://arxiv.org/pdf/2304.03064v1.pdf | An experimental study in Real-time Facial Emotion Recognition on new 3RL dataset | Although real-time facial emotion recognition is a hot topic research domain in the field of human-computer interaction, state-of the-art available datasets still suffer from various problems, such as some unrelated photos such as document photos, unbalanced numbers of photos in each class, and misleading images that can negatively affect correct classification. The 3RL dataset was created, which contains approximately 24K images and will be publicly available, to overcome previously available dataset problems. The 3RL dataset is labelled with five basic emotions: happiness, fear, sadness, disgust, and anger. Moreover, we compared the 3RL dataset with other famous state-of-the-art datasets (FER dataset, CK+ dataset), and we applied the most commonly used algorithms in previous works, SVM and CNN. The results show a noticeable improvement in generalization on the 3RL dataset. Experiments have shown an accuracy of up to 91.4% on 3RL dataset using CNN where results on FER2013, CK+ are, respectively (approximately from 60% to 85%). | ['Khloud Al Jallad', 'Tarek Barhoum', 'Rasha Albezreh', 'Rouaa Alaraj', 'Lana Ahmad Abdullah', 'Rahmeh Abou Zafra'] | 2023-04-06 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [-1.25081643e-01 -5.76415919e-02 -2.22707298e-02 -6.30801260e-01
-9.00254026e-02 -1.65232435e-01 3.91830772e-01 -8.00367072e-02
-4.74719524e-01 9.19588327e-01 -1.02697715e-01 4.16484773e-01
2.25387573e-01 -3.91165227e-01 -2.87374854e-01 -7.90450275e-01
-2.66261697e-02 -1.50741696e-01 -2.13091969e-01 -4.73584682e-01
2.24954158e-01 6.05214655e-01 -1.97069120e+00 5.52934170e-01
4.45488781e-01 1.78002417e+00 -7.79682755e-01 3.40851337e-01
1.00211322e-01 1.05429780e+00 -8.47482741e-01 -9.04201329e-01
2.88976319e-02 -1.28233179e-01 -7.38063037e-01 2.21554711e-01
3.18053216e-01 -8.62126723e-02 -1.62521362e-01 1.06377518e+00
6.28164589e-01 4.57748957e-02 6.58242047e-01 -2.09564114e+00
-8.48720372e-01 7.57525489e-02 -9.14660215e-01 -2.09825844e-01
4.02231395e-01 -2.00113803e-01 4.89427745e-01 -9.18472052e-01
6.38883293e-01 1.37258327e+00 7.03893602e-01 7.61135936e-01
-5.81068456e-01 -1.25081027e+00 -9.12526324e-02 5.46634674e-01
-1.58281076e+00 -4.46587145e-01 9.27243233e-01 -4.16851670e-01
7.44581044e-01 2.85883665e-01 6.71225965e-01 1.31952751e+00
1.34232923e-01 8.07443619e-01 1.64629829e+00 -2.55906701e-01
7.44420961e-02 5.92914879e-01 1.12519927e-01 4.78717297e-01
-2.91075230e-01 -1.44762322e-01 -5.45108855e-01 -8.36865082e-02
2.28311032e-01 -6.02618903e-02 -3.02606285e-01 -7.00253621e-02
-7.17136323e-01 6.76220417e-01 6.23091578e-01 3.39939207e-01
-2.82703936e-01 -4.44095373e-01 8.06423306e-01 4.60334092e-01
6.27126932e-01 5.22175506e-02 -5.00871718e-01 -2.01660275e-01
-5.82936049e-01 7.57607222e-02 5.86576581e-01 8.44013870e-01
8.35145950e-01 1.51992845e-03 1.42887875e-01 1.04361999e+00
4.22340706e-02 4.05581921e-01 7.74225056e-01 -5.33756554e-01
1.69686209e-02 7.87612259e-01 1.08742118e-01 -1.60417557e+00
-4.57470417e-01 -9.50827822e-02 -1.27812767e+00 2.70355195e-01
-1.58313755e-02 -1.74850151e-01 -8.23665857e-01 1.67613328e+00
1.76885128e-01 2.39323657e-02 4.57410365e-01 1.00765061e+00
1.48231375e+00 6.64050579e-01 2.22402692e-01 -4.04401660e-01
1.03435087e+00 -1.24857616e+00 -1.10862482e+00 -3.31648737e-02
4.21003729e-01 -9.00394380e-01 1.00460124e+00 8.11123610e-01
-7.17508435e-01 -6.88716531e-01 -9.19569016e-01 1.46660626e-01
-7.53918946e-01 5.83555222e-01 7.77931809e-01 8.01719487e-01
-1.03140509e+00 4.75019366e-01 -2.09867172e-02 -6.66450322e-01
7.29712248e-01 3.40410233e-01 -1.08680546e+00 -7.14018382e-03
-1.14464533e+00 9.50254798e-01 3.29103559e-01 5.53159237e-01
-5.19028485e-01 -2.39340097e-01 -6.98063195e-01 -2.03306735e-01
1.30980775e-01 3.31317365e-01 9.76338744e-01 -1.85965657e+00
-1.45918226e+00 1.38710344e+00 1.25610158e-01 -5.33510335e-02
4.02167231e-01 -2.94323504e-01 -1.06154025e+00 1.26189530e-01
-2.87566185e-01 8.53871644e-01 7.87469208e-01 -1.08299685e+00
-2.19191223e-01 -5.61333001e-01 -1.44076079e-01 -2.01706998e-02
-5.91079593e-01 1.95145294e-01 -1.27990276e-01 -4.39802408e-01
-2.14525804e-01 -9.64053154e-01 2.19437703e-01 -2.97175236e-02
-3.36376250e-01 -3.48408699e-01 1.16314542e+00 -6.25709832e-01
9.40591991e-01 -2.31769586e+00 -2.31522352e-01 -9.03743133e-02
-6.27235994e-02 5.86210012e-01 2.81409249e-02 2.77072966e-01
-7.14300394e-01 2.00478494e-01 1.10585969e-02 -1.69257298e-01
-1.55184299e-01 5.45753837e-02 -2.76195668e-02 4.44158643e-01
5.39141595e-01 6.88987076e-01 -5.36212027e-01 -6.06292129e-01
1.84957668e-01 6.17155075e-01 -6.41321465e-02 3.20678264e-01
3.82881969e-01 1.48579061e-01 -1.67375892e-01 1.08146918e+00
1.08524537e+00 1.53064460e-01 -1.63914174e-01 -6.38508558e-01
1.68956593e-01 -7.38940597e-01 -9.80044901e-01 1.17481601e+00
-1.72209859e-01 7.87294567e-01 -1.47118017e-01 -1.11553872e+00
1.40089655e+00 5.19638479e-01 3.72982055e-01 -6.62565589e-01
7.70106375e-01 -5.32204285e-02 -2.28370354e-01 -9.88174379e-01
5.40500402e-01 -2.00005636e-01 -6.45990521e-02 1.19242473e-02
2.45055512e-01 2.73198754e-01 5.71177714e-02 -2.51771491e-02
6.27210021e-01 -5.30680679e-02 3.36071342e-01 -4.22074795e-02
7.06068695e-01 -8.23327675e-02 5.79486668e-01 4.71199257e-03
-8.39103222e-01 4.95197803e-01 6.85381949e-01 -6.83090270e-01
-7.24237621e-01 -5.45846462e-01 -1.54511705e-01 8.73411655e-01
9.79921743e-02 -2.61955768e-01 -7.22847462e-01 -7.37895727e-01
-1.84316441e-01 2.82106370e-01 -9.64635670e-01 -3.36224407e-01
-9.19948146e-03 -9.24717724e-01 6.91180468e-01 4.18227375e-01
1.14139080e+00 -1.49342239e+00 -4.73862827e-01 -2.25591213e-01
-2.71747500e-01 -1.38801479e+00 3.37812036e-01 -8.45789164e-02
-4.21209693e-01 -1.27805603e+00 -8.42241466e-01 -7.42250502e-01
8.27951252e-01 -8.68843049e-02 1.15893376e+00 1.64748996e-01
-6.44318759e-01 4.59270403e-02 -6.10739470e-01 -5.62097490e-01
1.09186575e-01 -3.97663981e-01 -7.76110683e-03 4.35339004e-01
7.24038899e-01 -2.57159382e-01 -3.39822024e-01 4.37445819e-01
-7.22184837e-01 -1.21039368e-01 5.83407938e-01 8.87180090e-01
1.20589323e-01 1.54854849e-01 8.48030388e-01 -8.04924667e-01
4.15583283e-01 -3.83284539e-01 -1.17098875e-01 2.52722055e-01
-3.23267430e-01 -6.46565020e-01 4.38668638e-01 -6.01903677e-01
-1.21943569e+00 -1.43239927e-02 -2.37349376e-01 -5.31704366e-01
-6.08019650e-01 1.39212936e-01 -1.83882236e-01 -3.27309757e-01
5.20458043e-01 9.52040218e-03 -2.55876593e-02 -1.20841287e-01
-8.91895499e-03 1.23525953e+00 4.66152966e-01 -1.75938696e-01
2.61412591e-01 2.63057470e-01 -1.15778297e-01 -8.41864884e-01
-8.96948695e-01 -3.43814403e-01 -4.66802478e-01 -6.85049355e-01
6.75287426e-01 -1.17086661e+00 -1.03004038e+00 1.14324772e+00
-8.59531581e-01 1.90676883e-01 1.86105311e-01 2.43620053e-01
-2.49958664e-01 6.89520612e-02 -6.11572981e-01 -1.13864315e+00
-4.69872475e-01 -1.05561399e+00 9.40376699e-01 6.51675045e-01
-1.68005213e-01 -6.02307260e-01 -3.93905997e-01 5.23568213e-01
3.99297595e-01 8.48552167e-01 6.86819375e-01 -4.88509417e-01
2.65949011e-01 -4.88807231e-01 -4.96478647e-01 7.44616985e-01
-4.33437526e-02 4.75075871e-01 -1.41593742e+00 -1.32896125e-01
-6.63679764e-02 -1.24222386e+00 6.19849622e-01 -1.85722917e-01
1.39876735e+00 -2.11773351e-01 -1.70997515e-01 1.70708477e-01
1.40021551e+00 4.39853460e-01 1.13241804e+00 2.06702620e-01
2.25746647e-01 7.37134755e-01 1.00005054e+00 5.97168386e-01
2.06290141e-01 3.82779419e-01 6.13863051e-01 -4.44023818e-01
3.38769138e-01 1.03495546e-01 2.04869226e-01 5.87056756e-01
-1.73652738e-01 -2.17973724e-01 -6.78127646e-01 3.46059501e-01
-1.80351818e+00 -9.39877093e-01 -1.90334916e-01 1.81530523e+00
7.34166920e-01 -1.22493595e-01 1.31581575e-01 5.93062341e-01
8.11684310e-01 1.23043530e-01 -3.31908762e-01 -7.80212522e-01
-4.84748244e-01 2.87519366e-01 9.11847726e-02 -1.72171623e-01
-1.36811519e+00 1.13475060e+00 6.01835346e+00 7.84308732e-01
-1.72103190e+00 -1.40691355e-01 1.31376386e+00 -5.88311702e-02
6.08402193e-01 -5.78713536e-01 -6.18048310e-01 3.49094272e-01
7.19268560e-01 7.81915325e-04 1.33091703e-01 1.28375173e+00
-2.13091254e-01 -3.31178784e-01 -7.48673379e-01 1.70675385e+00
6.45633459e-01 -7.11069524e-01 -9.20367762e-02 -4.57413346e-01
7.44952381e-01 -4.66985881e-01 -2.21189447e-02 7.44027793e-01
-2.45364338e-01 -1.54454911e+00 4.66214627e-01 4.84588206e-01
7.91801512e-01 -1.16470516e+00 1.30576968e+00 6.85736164e-02
-8.02945435e-01 -4.01721336e-02 -5.97117424e-01 -2.99237132e-01
-3.33765119e-01 6.36554241e-01 -3.49015146e-01 5.29686391e-01
1.40431464e+00 1.08665287e+00 -8.17284405e-01 7.74077535e-01
-3.66857499e-02 2.20019817e-01 2.99357064e-02 -3.95102203e-01
6.41967654e-02 6.94444031e-02 -1.63789362e-01 1.14697909e+00
7.51653388e-02 2.60010928e-01 -1.09916873e-01 3.65746349e-01
-1.97497204e-01 5.40591419e-01 -5.27317762e-01 -1.60565823e-01
-1.37003645e-01 1.89947259e+00 -4.80211884e-01 -3.15733582e-01
-2.18837187e-01 1.06498075e+00 2.75049537e-01 3.10705990e-01
-9.64325547e-01 -5.88939309e-01 5.47453284e-01 -4.79579985e-01
-5.59028536e-02 4.16259885e-01 1.83816135e-01 -1.09359431e+00
1.31590620e-01 -1.06071019e+00 3.08867604e-01 -1.40129173e+00
-1.42074049e+00 1.06395388e+00 -3.01484525e-01 -1.07314622e+00
1.74932793e-01 -9.36864674e-01 -4.07977521e-01 5.21331489e-01
-1.44332933e+00 -1.15724003e+00 -9.18125451e-01 8.14932883e-01
2.42863938e-01 -3.87263775e-01 9.91117239e-01 5.43919027e-01
-8.34625661e-01 6.09799981e-01 -2.58640021e-01 3.52062076e-01
1.08016396e+00 -7.81067491e-01 -4.47002888e-01 1.36573076e-01
-3.30860078e-01 1.71162024e-01 4.44520473e-01 -1.80969059e-01
-1.16303933e+00 -9.36360478e-01 7.61951685e-01 -5.15252762e-02
2.52431899e-01 -2.47807026e-01 -6.59656882e-01 5.55116355e-01
4.92675215e-01 3.40502173e-01 9.06202257e-01 -3.23195569e-02
-3.90616506e-01 -2.91493356e-01 -1.64864385e+00 5.24051070e-01
6.74138665e-01 -1.94889918e-01 -2.43997231e-01 1.73730597e-01
9.56164524e-02 -3.69636148e-01 -8.43813717e-01 7.76228309e-01
7.51008987e-01 -1.49860525e+00 5.82433403e-01 -7.58157194e-01
7.71112144e-01 -1.85711011e-02 -2.74735689e-01 -1.38519812e+00
-4.63146567e-02 -1.75101325e-01 2.28567749e-01 1.63720679e+00
4.50160056e-02 -4.18396145e-01 7.47591019e-01 5.93440950e-01
3.21385384e-01 -1.09491217e+00 -6.49802864e-01 -5.08988321e-01
-2.75358379e-01 -2.23400027e-01 4.84569579e-01 1.24683273e+00
5.91787435e-02 3.78887981e-01 -6.66958809e-01 -2.71124154e-01
1.78685576e-01 -6.55431598e-02 9.74455655e-01 -1.28421402e+00
5.45097053e-01 -2.99167901e-01 -9.02166307e-01 -1.32028893e-01
3.46196115e-01 -4.72610116e-01 -2.32400969e-01 -1.12851393e+00
2.48724654e-01 -2.06893906e-01 -4.56706971e-01 8.79248798e-01
7.31996745e-02 8.22480381e-01 1.38583526e-01 -3.29220951e-01
-5.95509827e-01 7.69173145e-01 1.15308821e+00 -8.72450098e-02
2.20037445e-01 -4.73344892e-01 -7.36389875e-01 9.63188708e-01
7.34439135e-01 -1.12011254e-01 -1.35232955e-01 1.22988917e-01
1.19295999e-01 -1.01460941e-01 3.92006963e-01 -1.22050309e+00
-2.24812672e-01 -1.12804100e-01 1.00768757e+00 -6.24147892e-01
7.33326733e-01 -9.81414795e-01 3.14338893e-01 1.44381791e-01
-2.83176124e-01 1.37517512e-01 3.96684974e-01 3.14247981e-02
-7.92103112e-01 8.15151352e-03 1.17289925e+00 -9.83251706e-02
-9.97663081e-01 2.93360382e-01 3.16512436e-02 3.46252918e-02
1.47495842e+00 -1.30765170e-01 -4.35916990e-01 -6.48729622e-01
-7.21153796e-01 -4.31440026e-02 1.97466686e-01 7.40307331e-01
6.73788965e-01 -1.62957168e+00 -5.31368852e-01 1.33948550e-01
4.10244256e-01 -5.06740570e-01 5.08946180e-01 8.74544919e-01
-3.68643105e-01 1.15590788e-01 -9.20442402e-01 -4.67097461e-01
-1.72731674e+00 5.26058614e-01 2.34122038e-01 1.53149039e-01
-8.45610425e-02 9.05499995e-01 -4.55808677e-02 -3.83733839e-01
3.16664457e-01 1.41209960e-01 -5.01925707e-01 3.59928548e-01
6.25417233e-01 3.34783673e-01 9.71154347e-02 -1.14539862e+00
-6.46146357e-01 7.53212094e-01 -2.00243536e-02 3.88025939e-01
1.26244354e+00 1.29987329e-01 -2.91911095e-01 5.01041949e-01
1.48567963e+00 -3.42495412e-01 -6.15402162e-01 6.98126629e-02
-3.22589040e-01 -5.67317665e-01 -2.43158359e-02 -1.18285453e+00
-1.50549161e+00 1.02690244e+00 1.10761380e+00 1.57910250e-02
1.61214566e+00 -4.07835066e-01 7.66757131e-01 4.63930279e-01
2.50403225e-01 -1.24189591e+00 3.08559179e-01 3.37558061e-01
1.15859175e+00 -1.64607418e+00 -4.98298034e-02 -4.08551425e-01
-1.09132373e+00 1.27303159e+00 1.06742442e+00 -5.28535806e-02
7.51284778e-01 3.88014317e-02 4.85848635e-01 -1.48165701e-02
-6.94345713e-01 1.09555341e-01 8.76733810e-02 4.92600828e-01
5.73109984e-01 -1.01476982e-01 -2.55638361e-01 8.96842480e-01
-3.91952008e-01 4.41319227e-01 4.26511467e-01 8.29634368e-01
-1.33368328e-01 -7.01687276e-01 -4.11787003e-01 2.32054487e-01
-8.32392216e-01 2.40177915e-01 -7.50315964e-01 1.12201250e+00
4.40601856e-01 1.08475697e+00 2.80487407e-02 -6.05703175e-01
3.45081270e-01 1.35057151e-01 3.05326432e-01 1.05395295e-01
-6.34862423e-01 -2.55126953e-01 1.68409944e-01 -5.21643400e-01
-7.83848405e-01 -1.78611070e-01 -9.05890465e-01 -3.97270232e-01
-2.31537640e-01 3.47719155e-02 6.41821921e-01 7.42220998e-01
2.04024166e-01 2.01189056e-01 7.93243647e-01 -8.65503967e-01
1.79003607e-02 -1.06961453e+00 -6.68147326e-01 9.43292320e-01
5.62344678e-02 -9.13475692e-01 -4.40487206e-01 2.36284107e-01] | [13.569962501525879, 1.8432459831237793] |
8585a96c-c578-4a38-a50d-a87e57f55ec1 | beyond-contrastive-learning-a-variational | 2212.10726 | null | https://arxiv.org/abs/2212.10726v2 | https://arxiv.org/pdf/2212.10726v2.pdf | Beyond Contrastive Learning: A Variational Generative Model for Multilingual Retrieval | Contrastive learning has been successfully used for retrieval of semantically aligned sentences, but it often requires large batch sizes or careful engineering to work well. In this paper, we instead propose a generative model for learning multilingual text embeddings which can be used to retrieve or score sentence pairs. Our model operates on parallel data in $N$ languages and, through an approximation we introduce, efficiently encourages source separation in this multilingual setting, separating semantic information that is shared between translations from stylistic or language-specific variation. We show careful large-scale comparisons between contrastive and generation-based approaches for learning multilingual text embeddings, a comparison that has not been done to the best of our knowledge despite the popularity of these approaches. We evaluate this method on a suite of tasks including semantic similarity, bitext mining, and cross-lingual question retrieval -- the last of which we introduce in this paper. Overall, our Variational Multilingual Source-Separation Transformer (VMSST) model outperforms both a strong contrastive and generative baseline on these tasks. | ['Taylor Berg-Kirkpatrick', 'Graham Neubig', 'William W. Cohen', 'Jonathan H. Clark', 'John Wieting'] | 2022-12-21 | null | null | null | null | ['open-domain-question-answering'] | ['natural-language-processing'] | [-2.24084668e-02 -2.65235603e-01 -1.99473932e-01 -4.77659792e-01
-1.75327551e+00 -9.06494737e-01 9.20608461e-01 2.70385087e-01
-7.01330900e-01 5.57834804e-01 6.33925140e-01 -5.56279004e-01
2.59002368e-03 -3.20377737e-01 -7.64944851e-01 -4.71619248e-01
2.42246598e-01 8.93921793e-01 -1.01698495e-01 -4.96503562e-01
1.61873117e-01 -1.24938518e-01 -1.18987346e+00 2.74863034e-01
9.63103831e-01 4.97235686e-01 2.34296486e-01 6.46864772e-01
-4.57214147e-01 5.69100440e-01 -4.33833718e-01 -7.84066796e-01
1.65096119e-01 -5.25649190e-01 -1.01940203e+00 -2.86182880e-01
9.72927988e-01 -2.51031835e-02 -1.43664867e-01 9.28571999e-01
7.55933166e-01 -9.97269992e-03 8.82788897e-01 -9.17713344e-01
-1.13100350e+00 1.00814223e+00 -4.24742013e-01 4.72320527e-01
2.95783967e-01 -2.22595766e-01 1.68570101e+00 -1.17352831e+00
8.59685361e-01 1.31982338e+00 5.84386349e-01 4.56490219e-01
-1.33291256e+00 -4.83323157e-01 7.77543560e-02 1.38858497e-01
-1.27176881e+00 -7.34670222e-01 7.72666514e-01 -3.27250272e-01
1.21855795e+00 1.45895287e-01 3.06337774e-01 1.24338627e+00
8.18948597e-02 1.14391363e+00 1.12515461e+00 -7.84662843e-01
-6.51930794e-02 4.24867183e-01 4.57396917e-03 6.33926332e-01
-1.15925804e-01 -1.69347122e-01 -6.48146510e-01 -2.38460958e-01
1.32046461e-01 -2.61360884e-01 -2.66410321e-01 -4.89542902e-01
-1.16638768e+00 1.16002584e+00 1.74273327e-01 5.90127051e-01
-2.37067882e-02 1.13171726e-01 5.53329587e-01 7.64744580e-01
8.77721190e-01 5.53376615e-01 -6.41609609e-01 -1.23200193e-01
-1.17071307e+00 2.80895799e-01 6.83164001e-01 1.00236309e+00
6.48857296e-01 -3.86359394e-02 -8.68076310e-02 1.25153315e+00
2.70113081e-01 7.92424679e-01 7.96870649e-01 -7.22891927e-01
6.60883844e-01 8.55557993e-02 -1.97719589e-01 -6.96398377e-01
-2.43760850e-02 -2.79578149e-01 -3.17778021e-01 -4.62638855e-01
3.14880520e-01 7.88956694e-03 -5.04195452e-01 1.93081653e+00
1.93354994e-01 -1.45463720e-01 3.26967984e-01 6.17338777e-01
6.59104168e-01 7.01233983e-01 -6.90865666e-02 -2.72307191e-02
1.42437720e+00 -1.02203763e+00 -7.04221666e-01 -2.42123634e-01
8.67623925e-01 -1.37135863e+00 1.36907697e+00 -5.64195886e-02
-1.25545979e+00 -3.03945273e-01 -9.02162552e-01 -5.27601480e-01
-4.06800926e-01 9.35773328e-02 5.38990498e-01 6.33347511e-01
-1.24017715e+00 4.29342061e-01 -6.79189444e-01 -6.56517625e-01
1.21702559e-01 -7.65381604e-02 -1.01085655e-01 -2.82036126e-01
-1.39098716e+00 1.01316142e+00 -1.62273068e-02 -3.25280070e-01
-6.26129568e-01 -8.15518618e-01 -1.01455998e+00 -1.37075320e-01
1.15713917e-01 -8.36106420e-01 1.50241017e+00 -8.83997440e-01
-1.27669036e+00 1.08694756e+00 -4.73753154e-01 -3.87980878e-01
2.12063804e-01 -4.06014472e-01 -2.50664830e-01 9.69632789e-02
3.25081319e-01 8.11054707e-01 7.91146219e-01 -8.71279478e-01
-3.65724921e-01 -4.82060492e-01 -5.23274280e-02 4.91176039e-01
-5.98620117e-01 5.08193552e-01 -4.50038165e-01 -9.24377680e-01
-1.18981019e-01 -9.91513968e-01 1.17672384e-01 -3.14063430e-01
-2.38288313e-01 -5.96392930e-01 4.45244521e-01 -8.71136487e-01
9.62712348e-01 -1.86786675e+00 3.35495055e-01 -8.66821334e-02
-1.48975253e-01 8.30667838e-02 -4.81871098e-01 7.79476047e-01
6.54766858e-02 1.88458815e-01 -2.49119729e-01 -8.06701064e-01
3.91035527e-01 8.42957571e-02 -6.48474753e-01 3.12062651e-01
2.26304054e-01 1.15150738e+00 -9.24110830e-01 -6.69150710e-01
-7.77485520e-02 5.86741269e-01 -5.73490739e-01 2.10213348e-01
-1.91411778e-01 -1.73251797e-02 -2.51595706e-01 4.88963574e-01
2.76202440e-01 4.71824296e-02 4.60519284e-01 -9.30509493e-02
1.81654170e-01 9.40692008e-01 -6.14534378e-01 2.20140862e+00
-9.24288690e-01 8.00878227e-01 -1.13943629e-01 -8.74273002e-01
6.83612466e-01 4.23359275e-01 3.54478538e-01 -8.82833660e-01
-3.25958170e-02 5.60814738e-01 -2.69999206e-01 -4.80247527e-01
8.84631932e-01 -3.41473311e-01 -3.16333562e-01 9.45323586e-01
3.32528621e-01 -3.58377129e-01 2.94267833e-01 5.42195499e-01
8.08712840e-01 2.03698173e-01 5.84421493e-02 -5.68531334e-01
3.32528353e-01 -1.00635476e-02 1.33433253e-01 6.23248935e-01
-3.97998579e-02 4.31408614e-01 2.12557375e-01 1.02565987e-02
-1.12471366e+00 -1.24626374e+00 -2.52673447e-01 1.37415731e+00
-1.38497472e-01 -5.61686397e-01 -7.21645355e-01 -8.74041378e-01
-4.44105379e-02 7.87305415e-01 -3.02736610e-01 -6.93316236e-02
-8.68177176e-01 -8.03843975e-01 6.09731615e-01 5.43148160e-01
-1.16265126e-01 -8.55903685e-01 -7.77795538e-03 1.82725921e-01
-4.66722548e-01 -1.05047607e+00 -1.00375283e+00 1.02832176e-01
-8.07615161e-01 -9.22167063e-01 -6.76752269e-01 -1.20220017e+00
1.82866693e-01 2.25521624e-01 1.59541285e+00 -3.16916734e-01
-2.06706002e-01 6.78685725e-01 -3.83992970e-01 -3.78061026e-01
-5.32974124e-01 4.16473746e-01 -2.82677598e-02 -3.39301735e-01
7.09662139e-01 -5.09349167e-01 -3.36267680e-01 -2.30881393e-01
-1.02699792e+00 -4.73865688e-01 4.26320404e-01 1.02231920e+00
5.54049909e-01 -5.97287059e-01 6.26773000e-01 -8.88132095e-01
1.05705833e+00 -5.92760384e-01 -3.77838731e-01 5.53533912e-01
-7.06030607e-01 6.07501626e-01 5.54437876e-01 -2.32232228e-01
-6.70210481e-01 -5.88894486e-01 -2.49404147e-01 -2.85765737e-01
1.82120293e-01 5.99417806e-01 -3.09704598e-02 3.35593402e-01
5.55145264e-01 2.50225604e-01 -6.98436275e-02 -7.58707702e-01
7.63026655e-01 8.24784577e-01 2.40181044e-01 -8.23444664e-01
7.91860640e-01 1.77928254e-01 -4.97503072e-01 -6.74044967e-01
-9.46680486e-01 -5.28586984e-01 -5.00747740e-01 2.60430515e-01
8.36206555e-01 -1.05858982e+00 1.30418867e-01 8.71140882e-02
-1.36363876e+00 -1.64663494e-01 -4.26945448e-01 4.63567793e-01
-3.27107996e-01 3.84473532e-01 -8.90548408e-01 -4.07551080e-01
-5.43086588e-01 -1.14373875e+00 1.30108178e+00 -3.70500416e-01
-2.82292187e-01 -1.56515503e+00 6.27429307e-01 5.04153311e-01
6.57244146e-01 -3.30243558e-01 1.02750266e+00 -9.43082154e-01
-4.18878585e-01 2.16586739e-02 5.93009219e-02 4.71987933e-01
1.19373612e-01 -1.48492485e-01 -8.80427659e-01 -4.63860095e-01
5.48004434e-02 -8.05834055e-01 9.94151711e-01 2.49768585e-01
6.76710308e-01 -2.37453401e-01 -7.95999244e-02 3.37616503e-01
1.34853697e+00 -3.08440179e-01 2.71207720e-01 2.84572303e-01
6.93250418e-01 7.30397403e-01 3.51651430e-01 -1.92062601e-01
8.06561649e-01 8.61991465e-01 -1.75466731e-01 -9.15569514e-02
-3.53991568e-01 -2.50490874e-01 7.15265214e-01 1.66315782e+00
4.82385874e-01 -2.57092148e-01 -7.63500333e-01 9.27322447e-01
-1.53535998e+00 -9.86519575e-01 7.81838819e-02 2.12132382e+00
1.21905267e+00 -2.69800842e-01 1.63617823e-02 -3.45155120e-01
4.39603448e-01 4.29209888e-01 -2.69628853e-01 -7.19992578e-01
-2.98956305e-01 8.16121638e-01 2.49272123e-01 8.53055179e-01
-1.06154132e+00 1.16492605e+00 6.40926504e+00 9.82391953e-01
-9.88394320e-01 5.99232197e-01 4.09812063e-01 -3.06652904e-01
-1.02059460e+00 7.60515928e-02 -7.37357974e-01 2.55236059e-01
1.10756612e+00 -2.18789756e-01 3.74363661e-01 6.52456939e-01
-2.39595696e-01 3.22118580e-01 -1.31285954e+00 9.32450056e-01
6.78190947e-01 -1.09626007e+00 1.28388152e-01 -1.64687917e-01
9.15419877e-01 4.97473359e-01 3.01997930e-01 4.77640629e-01
6.56325698e-01 -9.27719355e-01 7.16533840e-01 7.91368410e-02
5.70663035e-01 -6.72882497e-01 4.57975537e-01 -4.00506146e-02
-1.07428634e+00 3.59491169e-01 -3.06900948e-01 5.70171893e-01
3.32182765e-01 5.35015881e-01 -5.50083697e-01 6.19452357e-01
5.92038393e-01 7.29667068e-01 -5.70492089e-01 4.81026322e-01
-3.19866359e-01 7.47116745e-01 -1.41365483e-01 -1.02786399e-01
4.41777170e-01 -1.89532354e-01 5.95687568e-01 1.36480498e+00
3.01448643e-01 -6.76994264e-01 1.70065880e-01 7.82885551e-01
-3.53976190e-01 6.37850523e-01 -6.35914028e-01 -1.87509418e-01
5.42136610e-01 1.07928395e+00 -9.06217769e-02 -3.97216976e-01
-4.82608557e-01 1.00871515e+00 7.13826299e-01 3.44154537e-01
-6.24997199e-01 -3.31554592e-01 7.46772289e-01 -1.15973532e-01
4.51113552e-01 -2.98197210e-01 1.52959719e-01 -1.59763300e+00
3.57493639e-01 -1.16558397e+00 5.05500317e-01 -5.80103457e-01
-1.74323654e+00 7.98707008e-01 -1.95234660e-02 -8.47069144e-01
-7.22025037e-01 -4.22322482e-01 -2.98094153e-01 1.09008574e+00
-1.94326127e+00 -1.20846272e+00 4.72129166e-01 5.98666430e-01
8.03458631e-01 -3.16059798e-01 9.55873132e-01 4.13232386e-01
-3.59514624e-01 8.64367902e-01 5.23317099e-01 8.56156349e-02
1.22565842e+00 -1.51127839e+00 4.90151435e-01 8.33522677e-01
9.02908087e-01 8.50117385e-01 5.54010332e-01 -2.73122132e-01
-1.35237300e+00 -1.16377568e+00 1.73609626e+00 -7.60293245e-01
1.02420366e+00 -5.51841021e-01 -8.12661588e-01 8.09092522e-01
8.56463730e-01 -3.54398310e-01 9.84253645e-01 5.04325628e-01
-9.14183617e-01 -1.42653123e-01 -6.26037717e-01 6.75536215e-01
7.69450665e-01 -1.14241815e+00 -8.56457651e-01 6.22810185e-01
8.50361526e-01 -4.69740853e-02 -8.69465947e-01 1.16226852e-01
3.64704162e-01 -7.22142935e-01 1.00479972e+00 -7.95361161e-01
5.34742415e-01 8.50264430e-02 -3.97997588e-01 -1.56661510e+00
4.23377268e-02 -5.85291147e-01 2.53845990e-01 1.48785639e+00
6.18323088e-01 -6.26232147e-01 2.10825920e-01 8.31098482e-02
-2.56580055e-01 -6.49768174e-01 -8.60185683e-01 -9.42111611e-01
1.00343060e+00 -4.62945104e-01 5.87008893e-01 1.15012574e+00
4.26729880e-02 7.78438568e-01 -1.13604128e-01 -3.50701779e-01
5.64143777e-01 2.69581199e-01 5.20506799e-01 -7.74571180e-01
-5.55670917e-01 -6.08919263e-01 -2.95519494e-02 -1.15957189e+00
6.27286315e-01 -1.52635872e+00 -2.33007967e-03 -1.37036157e+00
3.87838930e-01 -4.75475192e-01 -3.00335020e-01 2.03022823e-01
-4.28968012e-01 3.00210744e-01 -3.61536182e-02 3.02070022e-01
-6.51425421e-01 6.90749824e-01 7.61964321e-01 -3.47682029e-01
8.06081221e-02 -5.79990923e-01 -8.62322927e-01 2.07689732e-01
8.78325820e-01 -6.04080200e-01 -5.75396657e-01 -1.03927398e+00
3.25149536e-01 -1.76223099e-01 1.74736202e-01 -4.76079911e-01
-1.15753487e-01 1.46490440e-01 -1.95455432e-01 -2.76192874e-01
2.27454588e-01 -3.60926658e-01 -3.76255661e-01 4.69086617e-02
-7.74454236e-01 5.21972775e-01 5.99533245e-02 3.64485919e-01
-5.10967910e-01 -2.87205279e-01 3.90673369e-01 -5.57116903e-02
-3.03523719e-01 2.47415245e-01 -2.56415576e-01 7.83711433e-01
4.41871673e-01 3.83342922e-01 -3.40384454e-01 -3.97852182e-01
-2.34823644e-01 1.13936365e-01 4.45966870e-01 7.48095632e-01
4.14001435e-01 -1.55533493e+00 -1.15852988e+00 -7.16853738e-02
3.95157009e-01 -4.99090433e-01 -2.12694239e-02 9.29702342e-01
-1.93095431e-01 6.65807188e-01 4.00094986e-01 -4.94243056e-01
-1.28109694e+00 5.05413592e-01 2.50438124e-01 -6.20477557e-01
-2.91943133e-01 9.10193324e-01 5.97553924e-02 -8.24565530e-01
-1.89460218e-02 -9.54508930e-02 -6.35337681e-02 2.26324454e-01
2.82771319e-01 -1.01681110e-02 3.18568170e-01 -7.88966298e-01
-3.15220594e-01 5.93085587e-01 -4.46341217e-01 -5.97031355e-01
1.24121881e+00 -3.54793340e-01 -3.53964984e-01 7.10656345e-01
1.76400232e+00 3.56147945e-01 -5.90307772e-01 -5.39133966e-01
1.41989022e-01 -2.42910579e-01 -6.47811145e-02 -4.80191201e-01
-8.35014880e-01 1.22233653e+00 3.46697330e-01 5.49734849e-03
7.70155668e-01 3.42403352e-01 1.25190449e+00 4.62370574e-01
1.12912819e-01 -1.25322509e+00 1.96026802e-01 5.92031956e-01
7.08027363e-01 -1.29259825e+00 -3.57605278e-01 5.12450263e-02
-5.58700442e-01 7.89156020e-01 1.82923257e-01 -1.18530512e-01
5.98148823e-01 1.76333874e-01 3.04625034e-01 -2.19355866e-01
-8.48338902e-01 -1.86232299e-01 5.62033832e-01 3.56052607e-01
9.28816199e-01 -9.59751010e-03 -4.58479077e-01 7.81502575e-02
-4.70366806e-01 -6.39599621e-01 9.16844383e-02 9.76021588e-01
-1.92894310e-01 -1.67683816e+00 -1.29254580e-01 2.62951404e-01
-6.47086680e-01 -8.58994484e-01 -4.70243424e-01 6.44081354e-01
-2.13787138e-01 1.01591301e+00 2.16196567e-01 -8.85716379e-02
5.94157428e-02 5.33523917e-01 7.43200541e-01 -5.45591533e-01
-6.91347539e-01 1.80340052e-01 2.00762555e-01 -4.35298473e-01
-6.53572261e-01 -1.08417308e+00 -6.69814289e-01 -1.70022547e-01
-2.14188278e-01 4.75111157e-01 8.40018988e-01 1.00230896e+00
3.83727044e-01 1.65069222e-01 6.32713437e-01 -2.78164715e-01
-8.66371691e-01 -1.02345037e+00 -3.93354028e-01 5.94702780e-01
1.40795231e-01 -2.03712881e-01 -4.51901674e-01 -4.99907918e-02] | [11.141907691955566, 9.849842071533203] |
dc202d4b-3a11-44f2-98e7-09b3cd5677d6 | a-brief-survey-of-recent-edge-preserving | 1503.07297 | null | http://arxiv.org/abs/1503.07297v1 | http://arxiv.org/pdf/1503.07297v1.pdf | A Brief Survey of Recent Edge-Preserving Smoothing Algorithms on Digital Images | Edge preserving filters preserve the edges and its information while blurring
an image. In other words they are used to smooth an image, while reducing the
edge blurring effects across the edge like halos, phantom etc. They are
nonlinear in nature. Examples are bilateral filter, anisotropic diffusion
filter, guided filter, trilateral filter etc. Hence these family of filters are
very useful in reducing the noise in an image making it very demanding in
computer vision and computational photography applications like denoising,
video abstraction, demosaicing, optical-flow estimation, stereo matching, tone
mapping, style transfer, relighting etc. This paper provides a concrete
introduction to edge preserving filters starting from the heat diffusion
equation in olden to recent eras, an overview of its numerous applications, as
well as mathematical analysis, various efficient and optimized ways of
implementation and their interrelationships, keeping focus on preserving the
boundaries, spikes and canyons in presence of noise. Furthermore it provides a
realistic notion for efficient implementation with a research scope for
hardware realization for further acceleration. | ['Amlan Chakrabarti', 'Chandrajit Pal', 'Ranjan Ghosh'] | 2015-03-25 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 2.28660986e-01 -4.49095249e-01 6.14635944e-01 -1.53307423e-01
2.19896823e-01 -6.62662745e-01 4.84930843e-01 -6.63412139e-02
-5.02671063e-01 9.02932525e-01 6.02252126e-01 -3.90032679e-02
-1.74772754e-01 -7.65113115e-01 -4.37484682e-01 -7.23555505e-01
-1.11401305e-02 -1.25091344e-01 5.21820784e-01 -3.72127593e-01
6.67322755e-01 7.61894643e-01 -1.72224510e+00 1.69097751e-01
8.17264855e-01 5.59983313e-01 1.66843131e-01 9.76612747e-01
-6.76723942e-02 7.07874358e-01 -2.15709612e-01 -2.96712220e-01
3.95831496e-01 -4.81616914e-01 -5.48240900e-01 2.67532200e-01
5.31653225e-01 -3.87236387e-01 -4.85197365e-01 1.25769818e+00
3.72287720e-01 4.06048387e-01 7.35190034e-01 -9.64082122e-01
-5.90550482e-01 -7.46184662e-02 -7.32408047e-01 5.63505650e-01
2.01293513e-01 4.76688892e-02 2.14773282e-01 -8.08266699e-01
9.25207078e-01 1.20357299e+00 8.44494581e-01 3.33113760e-01
-1.10717607e+00 -2.30399996e-01 -2.79066652e-01 5.13922930e-01
-1.04803145e+00 -3.84126544e-01 6.20687187e-01 -3.41239929e-01
7.43669629e-01 8.39423001e-01 7.72906840e-01 2.70973295e-01
6.24598503e-01 5.88850439e-01 1.49191129e+00 -4.79138792e-01
-1.28198816e-02 2.71641433e-01 9.67722684e-02 7.57976949e-01
1.80417255e-01 1.33470386e-01 -4.55404580e-01 6.21597935e-03
1.04376638e+00 6.01318777e-02 -8.04321170e-01 -1.40209585e-01
-1.07917011e+00 4.67415214e-01 3.42061609e-01 4.82634872e-01
-3.21721643e-01 -4.14556935e-02 2.62825608e-01 4.95930284e-01
5.94686210e-01 -2.50171393e-01 -1.27721041e-01 4.33559343e-03
-8.71185660e-01 2.40690306e-01 6.69207215e-01 7.76723266e-01
8.51741314e-01 3.14588919e-02 -6.31839186e-02 5.52369297e-01
5.33788241e-02 2.01994896e-01 3.96242023e-01 -1.09072745e+00
-1.08928248e-01 2.60389000e-01 2.50383526e-01 -1.11703122e+00
-3.68252724e-01 -2.37059951e-01 -1.04871130e+00 1.00317252e+00
2.91931540e-01 -7.01192096e-02 -9.30443227e-01 1.15872526e+00
3.67888689e-01 6.41820312e-01 -1.77165344e-01 8.96373510e-01
6.74160659e-01 8.22794020e-01 -2.68000454e-01 -3.59780282e-01
1.43890178e+00 -7.55817473e-01 -9.72233117e-01 1.85886007e-02
1.18989259e-01 -1.52201867e+00 5.23545742e-01 6.27008259e-01
-1.45263267e+00 -5.13448119e-01 -7.58057356e-01 -5.50383747e-01
-4.32130247e-01 -1.06287532e-01 4.82973039e-01 6.38900340e-01
-1.23626387e+00 9.19203401e-01 -6.04523361e-01 -5.78424156e-01
1.45157129e-01 1.62439078e-01 -5.53983808e-01 -1.02484167e-01
-7.44359434e-01 1.11117089e+00 1.86601117e-01 1.94255233e-01
-2.05767289e-01 -8.51938665e-01 -6.30861819e-01 3.37438136e-02
-2.49123484e-01 -8.69777620e-01 3.95041764e-01 -1.00851154e+00
-1.33803511e+00 1.19205594e+00 -3.58784080e-01 -6.07124746e-01
9.33033824e-01 -1.22737676e-01 -3.23875129e-01 -1.37669638e-01
-3.60053509e-01 5.00317097e-01 1.10852945e+00 -8.52518499e-01
-8.17067564e-01 -3.11334759e-01 -4.38754886e-01 4.86183822e-01
6.37826398e-02 -1.44828394e-01 -1.59422770e-01 -9.04976130e-01
1.68281376e-01 -5.53728640e-01 -1.80716403e-02 3.37017745e-01
-6.74295500e-02 3.07526678e-01 1.38078988e+00 -1.13787150e+00
1.02456307e+00 -2.44486856e+00 2.86935389e-01 1.29316688e-01
1.19391136e-01 2.84798473e-01 1.12694621e-01 5.16658902e-01
-1.19755752e-01 -3.57908219e-01 -5.32592595e-01 -2.69852746e-02
-3.98233831e-01 1.03282154e-01 -8.41567144e-02 1.08575273e+00
-2.14883596e-01 4.35004473e-01 -6.65848851e-01 -4.88316387e-01
7.78611660e-01 1.04348743e+00 -3.83901983e-01 -3.56324948e-02
4.35432374e-01 5.74864388e-01 1.59826815e-01 2.49252781e-01
1.22311568e+00 6.05830908e-01 -6.06760681e-01 -4.24169064e-01
-7.99856246e-01 -4.58952010e-01 -1.63036263e+00 1.38060427e+00
-4.74289089e-01 1.37214053e+00 6.53338313e-01 -9.84201372e-01
1.00368452e+00 2.66686708e-01 3.12864363e-01 -3.87207687e-01
1.09754615e-01 3.78697544e-01 -3.79481167e-01 -6.65188611e-01
6.51837170e-01 -2.66974390e-01 8.03334773e-01 8.25908855e-02
-3.25437009e-01 -2.27358103e-01 3.68023276e-01 -2.78703105e-02
8.84022117e-01 -1.63899660e-01 3.65307778e-01 -7.59307563e-01
1.11019397e+00 -2.93050110e-01 4.79143143e-01 3.46436262e-01
-2.59818465e-01 8.70601058e-01 7.00581521e-02 -6.57019734e-01
-1.11760283e+00 -8.97758543e-01 -4.02491480e-01 3.67779613e-01
4.18982297e-01 2.07716018e-01 -8.58092248e-01 1.23748146e-01
-1.63089074e-02 5.40434241e-01 -6.15162134e-01 -7.78375491e-02
-8.07244837e-01 -4.81981963e-01 -3.21359336e-02 -1.20493054e-01
1.21027088e+00 -9.71517622e-01 -1.00074553e+00 2.45781049e-01
3.56312618e-02 -6.99833512e-01 -6.37575686e-01 -2.76317745e-01
-1.31217861e+00 -9.48276579e-01 -1.19143462e+00 -1.01240635e+00
5.25690138e-01 6.04792476e-01 1.08334064e+00 2.05956623e-01
-7.40345359e-01 2.26093471e-01 4.73978370e-03 -1.38822183e-01
-2.25387678e-01 -7.76477039e-01 -2.67632842e-01 1.69114187e-01
3.09394479e-01 -8.60793948e-01 -9.93531883e-01 2.13164523e-01
-1.12912524e+00 -4.15893123e-02 4.92789179e-01 7.97706664e-01
2.11009875e-01 5.04109919e-01 -2.82950938e-01 -9.49025214e-01
8.96359205e-01 3.37753482e-02 -7.80983269e-01 9.29352120e-02
-3.71725827e-01 -4.72063310e-02 4.11216319e-01 -2.69385651e-02
-1.53204346e+00 -2.93616056e-02 6.07027598e-02 1.30163580e-01
-1.63707837e-01 -2.47673616e-01 1.69968069e-01 -5.55225849e-01
7.64278889e-01 3.50918114e-01 3.75049263e-02 -7.79839158e-01
3.44959676e-01 5.82603395e-01 8.91612351e-01 2.93874592e-02
7.02583611e-01 1.28011692e+00 3.24407250e-01 -1.29703951e+00
-2.57393755e-02 -7.82797337e-01 -5.18424809e-01 -3.84516448e-01
1.00253165e+00 -4.58810985e-01 -6.06882334e-01 6.29667461e-01
-1.40578032e+00 1.41312048e-01 -2.34550834e-01 4.39662784e-01
-5.73147476e-01 5.80425680e-01 -8.38914037e-01 -7.90791273e-01
-3.64846200e-01 -1.03371632e+00 6.26875043e-01 7.61587441e-01
6.64136782e-02 -1.30800819e+00 -1.45675261e-02 8.60685483e-02
6.69656992e-01 3.11077625e-01 6.80345476e-01 5.05112767e-01
-6.05863333e-01 -7.97318891e-02 -3.91323745e-01 5.17750263e-01
2.92063296e-01 -1.31048402e-02 -8.87317955e-01 -2.75563985e-01
4.09413815e-01 5.52764893e-01 1.02841032e+00 8.12945068e-01
6.17793381e-01 -1.61557361e-01 -2.08900556e-01 9.41920757e-01
1.76395047e+00 1.34142891e-01 1.17744184e+00 3.04618508e-01
2.82392442e-01 6.14861488e-01 3.16209942e-01 2.62304932e-01
-4.43907678e-01 4.34671789e-01 2.20975786e-01 -3.74317199e-01
-5.91911852e-01 4.49196160e-01 1.02282844e-01 4.04013306e-01
-4.61496294e-01 -7.87118524e-02 -2.82765329e-01 5.27484715e-01
-1.58141541e+00 -9.97304797e-01 -9.24498677e-01 2.55361819e+00
4.15699065e-01 -1.05211571e-01 -1.98257253e-01 4.00753379e-01
1.12418413e+00 -1.71629544e-02 -1.00206539e-01 -8.28758657e-01
-3.90761524e-01 2.56787658e-01 1.00037944e+00 1.16069317e+00
-1.01594412e+00 7.96179116e-01 5.74787426e+00 7.48972356e-01
-1.11774409e+00 2.02623948e-01 5.47679067e-01 4.08751905e-01
-1.81855142e-01 7.46631920e-02 -4.08177108e-01 5.20138621e-01
3.55197757e-01 -1.44982561e-01 5.92007339e-01 2.36242697e-01
5.38849354e-01 -7.31788158e-01 -4.64003295e-01 1.37787259e+00
-9.52490345e-02 -1.47468686e+00 -2.77029783e-01 -4.69654836e-02
9.51569021e-01 -2.66499251e-01 1.64411608e-02 -6.19244099e-01
-1.84440285e-01 -8.61206353e-01 4.02067870e-01 7.87533760e-01
3.59314471e-01 -8.46532047e-01 7.27084696e-01 -2.20431690e-03
-9.61624444e-01 1.98166400e-01 -5.28934717e-01 -2.16150686e-01
6.63219213e-01 8.74516606e-01 -3.92602496e-02 2.51815319e-01
7.29300439e-01 6.67502046e-01 -1.40723974e-01 1.66944134e+00
2.29326457e-01 1.10110268e-01 -3.87429625e-01 2.86433935e-01
3.72746736e-01 -9.75523651e-01 9.81058538e-01 1.50875390e+00
3.33330482e-01 3.00123096e-01 -5.86325765e-01 6.21736765e-01
1.86619669e-01 1.28229961e-01 -5.33437669e-01 6.18607342e-01
1.05134495e-01 1.17339408e+00 -1.18707824e+00 -3.52331221e-01
-3.71799886e-01 1.49152851e+00 -5.24085045e-01 5.20394802e-01
-4.72122729e-01 -8.79177868e-01 8.54822814e-01 2.95522958e-01
-2.53822953e-02 -2.67940670e-01 -4.34762061e-01 -9.63023305e-01
-1.13876812e-01 -5.41552722e-01 2.56057858e-01 -7.86781371e-01
-7.56414711e-01 3.88532996e-01 -2.02857569e-01 -1.22866988e+00
4.45845515e-01 -7.22248018e-01 -7.61631966e-01 1.28335989e+00
-1.47370362e+00 -6.04469597e-01 -6.66718543e-01 7.16560066e-01
7.99832761e-01 7.02304170e-02 1.83878005e-01 5.70216596e-01
-7.52638876e-02 -1.57421291e-01 4.75791007e-01 -3.04911345e-01
8.13022375e-01 -1.21784997e+00 1.33485988e-01 1.23183799e+00
-7.95473903e-03 4.18681294e-01 1.30849993e+00 -5.47019601e-01
-1.10340214e+00 -5.14707029e-01 8.85219157e-01 2.55624890e-01
2.90468544e-01 1.75096709e-02 -1.03465199e+00 3.94410431e-01
7.21653581e-01 7.28239194e-02 -6.71027005e-02 -6.86653793e-01
3.83843035e-01 -1.02690272e-01 -1.39144385e+00 5.87553322e-01
8.90673339e-01 -1.49644226e-01 -2.96124607e-01 5.75749934e-01
-1.14667574e-02 -5.87890983e-01 -4.35392052e-01 -1.65875256e-01
4.32231277e-01 -1.98351729e+00 1.07984877e+00 -1.96768314e-01
7.68931583e-02 -4.74000096e-01 3.98567706e-01 -1.26488411e+00
-6.36189163e-01 -1.20147860e+00 4.48476464e-01 1.06765079e+00
-2.20672134e-02 -7.08092034e-01 8.54710698e-01 3.02115530e-01
-2.30492160e-01 -3.65283079e-02 -9.15132403e-01 -4.19071764e-01
-3.49143595e-01 1.76145136e-01 -7.72425486e-03 9.17548537e-01
-4.02961731e-01 -6.06409200e-02 -5.57671249e-01 7.62930661e-02
9.53332782e-01 -2.67144352e-01 4.41697538e-01 -1.19097090e+00
-6.97572157e-02 -6.58045053e-01 -9.22460854e-01 -8.92867386e-01
-3.86499643e-01 -4.36941296e-01 -2.44395420e-01 -1.57776511e+00
-2.29531080e-01 -1.33798018e-01 2.31942803e-01 -4.52916682e-01
-1.94221493e-02 7.02637374e-01 -2.01649545e-03 2.63791047e-02
3.17298204e-01 -3.00670750e-02 1.36408365e+00 8.83432776e-02
-2.47915059e-01 6.55498430e-02 -1.53038636e-01 8.75262201e-01
7.46460080e-01 -1.62825778e-01 -4.16613728e-01 -6.50862992e-01
5.08159511e-02 -2.24162012e-01 3.55521232e-01 -9.74551380e-01
5.06559372e-01 7.00998753e-02 3.38207930e-01 -2.78160989e-01
4.24199373e-01 -1.09725690e+00 4.00640249e-01 7.55630195e-01
1.12138629e-01 3.32623750e-01 1.46672234e-01 6.28040791e-01
-4.33362126e-01 -6.41408265e-01 1.40068889e+00 -3.93207192e-01
-1.13954496e+00 -6.47827536e-02 -5.84612191e-01 -8.90658945e-02
1.16992009e+00 -7.54467249e-01 -9.20857638e-02 -5.53660393e-01
-8.66804540e-01 -2.50364929e-01 6.78087473e-01 -1.18052036e-01
7.11084962e-01 -9.50841784e-01 -8.68387878e-01 4.05336440e-01
-5.30669630e-01 -2.84837246e-01 6.29355133e-01 1.07290173e+00
-1.59001851e+00 1.40869632e-01 -8.49197507e-01 -2.58129239e-01
-1.84110594e+00 4.77582604e-01 3.72897059e-01 3.13212842e-01
-1.09242165e+00 1.07280409e+00 3.36300403e-01 5.79189897e-01
1.23332851e-01 -3.28874409e-01 -1.54197752e-01 -2.69863248e-01
7.59738564e-01 1.07068896e+00 1.44313127e-01 -6.27787828e-01
-1.32544264e-01 1.04565692e+00 2.43012220e-01 -1.74817264e-01
1.19247174e+00 -5.14551938e-01 -7.08246469e-01 1.80024788e-01
1.15250063e+00 3.01120698e-01 -1.31713319e+00 8.60115960e-02
-7.25027323e-02 -9.99430835e-01 5.60686171e-01 -3.81599069e-01
-1.08045495e+00 9.78959262e-01 8.77578080e-01 4.92352396e-01
1.46110392e+00 -6.40520751e-01 7.40124106e-01 -1.46742120e-01
-6.11002147e-02 -1.17377698e+00 -4.77267712e-01 1.44818708e-01
9.07980561e-01 -8.37589741e-01 1.96405694e-01 -7.44751930e-01
-2.83266097e-01 1.56561053e+00 -6.91889152e-02 -5.78882158e-01
8.42786193e-01 4.61963505e-01 -1.67923812e-02 1.50594220e-01
-2.61539519e-01 -2.41067708e-01 1.73184901e-01 8.84842336e-01
6.90907538e-01 -2.54917860e-01 -1.01757371e+00 -8.31548989e-01
4.01249938e-02 1.35407642e-01 8.74597847e-01 8.54889750e-01
-7.57098198e-01 -8.98356736e-01 -8.74734044e-01 3.08561444e-01
-3.60887557e-01 -1.44647151e-01 -1.26398593e-01 7.17358470e-01
3.17753851e-01 7.90888786e-01 2.15522230e-01 1.96303636e-01
3.92530769e-01 -1.60221845e-01 7.11825371e-01 1.46998158e-02
-6.43989205e-01 1.07045598e-01 4.49885940e-03 -4.35688764e-01
-4.40174252e-01 -6.04580164e-01 -7.56186724e-01 -7.76893973e-01
-2.01203860e-02 1.99097008e-01 7.47728765e-01 5.14034748e-01
8.46745819e-02 2.61024952e-01 2.48719200e-01 -9.10065830e-01
1.02432691e-01 -8.31459761e-01 -9.70764101e-01 6.94976330e-01
5.90162635e-01 -5.30182779e-01 -5.18504024e-01 5.15818298e-01] | [11.12946891784668, -2.5562186241149902] |
2a720965-bf39-4e7b-ba73-9cb6a6d296d7 | rare-and-zero-shot-word-sense-disambiguation | null | null | https://aclanthology.org/2022.acl-long.323 | https://aclanthology.org/2022.acl-long.323.pdf | Rare and Zero-shot Word Sense Disambiguation using Z-Reweighting | Word sense disambiguation (WSD) is a crucial problem in the natural language processing (NLP) community. Current methods achieve decent performance by utilizing supervised learning and large pre-trained language models. However, the imbalanced training dataset leads to poor performance on rare senses and zero-shot senses. There are more training instances and senses for words with top frequency ranks than those with low frequency ranks in the training dataset. We investigate the statistical relation between word frequency rank and word sense number distribution. Based on the relation, we propose a Z-reweighting method on the word level to adjust the training on the imbalanced dataset. The experiments show that the Z-reweighting strategy achieves performance gain on the standard English all words WSD benchmark. Moreover, the strategy can help models generalize better on rare and zero-shot senses. | ['Tong Zhang', 'Yangqiu Song', 'Hongming Zhang', 'Ying Su'] | null | null | null | null | acl-2022-5 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 6.55100942e-02 -2.06670657e-01 -6.57368422e-01 -2.93052882e-01
-4.29103613e-01 -2.57407576e-01 3.83988529e-01 7.46339262e-01
-9.50131536e-01 6.61032617e-01 4.25364316e-01 -3.02460313e-01
-2.27283284e-01 -9.89102900e-01 7.37068430e-02 -4.51428771e-01
-9.19328183e-02 3.43580335e-01 2.55124897e-01 -7.82783568e-01
3.38286698e-01 -2.28054337e-02 -1.49678886e+00 1.38519689e-01
1.08742142e+00 7.68553615e-01 3.91000301e-01 1.34558193e-02
-8.44119787e-01 3.75511386e-02 -7.43588030e-01 -2.46547803e-01
2.21894354e-01 6.30946159e-02 -7.46324420e-01 -3.85231763e-01
-1.26406690e-02 4.38373715e-01 -9.05235335e-02 1.33696067e+00
7.75750399e-01 4.35739785e-01 4.78561580e-01 -1.17299211e+00
-7.84424722e-01 7.74997473e-01 -1.00084448e+00 7.56701469e-01
4.10141528e-01 -2.80798793e-01 1.33123970e+00 -8.94642055e-01
3.98851663e-01 1.40796983e+00 5.66144645e-01 4.19140071e-01
-8.59041154e-01 -9.78365362e-01 5.55932701e-01 3.56972039e-01
-1.62582326e+00 5.47180623e-02 5.73222458e-01 -2.27336526e-01
1.56965864e+00 1.09063558e-01 5.39385796e-01 9.28436458e-01
1.85820371e-01 6.91523433e-01 6.97365999e-01 -6.83466852e-01
2.14437023e-01 -2.17797935e-01 6.74447954e-01 2.05270961e-01
8.12496006e-01 -2.50769824e-01 -6.49432540e-01 -2.95927495e-01
2.81496137e-01 1.68323845e-01 -1.12357758e-01 6.09403616e-03
-1.03951526e+00 9.23550606e-01 1.48217827e-01 7.05515325e-01
-2.24235043e-01 -4.67554450e-01 6.10404551e-01 4.46082592e-01
9.68837380e-01 9.44876850e-01 -1.03487599e+00 -7.28378221e-02
-6.80858970e-01 3.25887769e-01 5.22784233e-01 6.82501614e-01
7.42860138e-01 -1.81633219e-01 -4.70711052e-01 1.36307406e+00
2.19948530e-01 4.70265538e-01 1.35875428e+00 -1.75317243e-01
5.71814001e-01 9.68972266e-01 -4.57528755e-02 -1.38898039e+00
-5.41096032e-01 -5.99058032e-01 -8.60936463e-01 -3.51111680e-01
-6.15712022e-03 6.57348931e-02 -1.03152812e+00 1.81963968e+00
3.55900973e-01 3.36223751e-01 8.79763216e-02 7.62025356e-01
6.90680444e-01 5.45295179e-01 5.11876702e-01 -5.42458773e-01
1.70812941e+00 -4.19717044e-01 -9.51651394e-01 -8.00143063e-01
7.49238193e-01 -8.98270786e-01 1.69004488e+00 2.25903735e-01
-3.91138345e-01 -3.98302495e-01 -1.24937069e+00 -1.13941431e-02
-7.85983622e-01 -3.23127449e-01 5.86193264e-01 5.88590562e-01
-4.89177465e-01 3.56958002e-01 -4.15309340e-01 -3.96700978e-01
2.72932827e-01 4.15256359e-02 -1.87267303e-01 -2.23161548e-01
-2.12574553e+00 9.91007209e-01 8.78305078e-01 -7.26901412e-01
1.16866574e-01 -9.07643914e-01 -1.06722450e+00 1.02741979e-01
3.42733979e-01 -4.82700914e-01 9.98662949e-01 -5.60197711e-01
-7.33338475e-01 1.15139616e+00 -3.24962139e-01 -4.12673861e-01
1.38991280e-02 -2.95808941e-01 -9.44630682e-01 -3.38889897e-01
3.77562582e-01 1.72531545e-01 4.91402030e-01 -8.52298856e-01
-8.00870776e-01 -4.39917892e-01 -7.88254514e-02 3.85467887e-01
-7.57428885e-01 -3.12388182e-01 -2.03251392e-01 -1.04416263e+00
1.99852362e-01 -3.99940878e-01 -3.62383336e-01 -6.69293702e-01
-3.29091817e-01 -8.05890918e-01 4.28182423e-01 -7.62927234e-02
1.86509085e+00 -2.17078590e+00 -4.60223556e-01 3.36507797e-01
3.18053901e-01 6.42546475e-01 -4.33807403e-01 2.73506820e-01
-4.48170722e-01 2.91824549e-01 1.98964216e-02 -1.78346764e-02
3.15886140e-02 3.40741605e-01 -4.85408902e-01 -6.54178672e-03
7.97737241e-02 5.85378706e-01 -1.38392377e+00 -4.05265927e-01
-9.05506127e-03 1.19555637e-01 -4.81904358e-01 -5.12742884e-02
4.09800448e-02 -3.99337500e-01 -6.20327711e-01 3.84975284e-01
6.36290193e-01 -2.73115098e-01 2.16762722e-01 -1.22251444e-01
2.11800203e-01 7.22516179e-01 -1.23263001e+00 1.56105053e+00
-5.90491235e-01 1.40891120e-01 -6.15307450e-01 -9.86841142e-01
9.82650220e-01 4.46676835e-02 4.38981712e-01 -9.83256400e-01
1.08394355e-01 1.53350383e-01 7.89747853e-03 -6.17667019e-01
8.27200353e-01 -5.68550050e-01 -3.10379684e-01 2.76157349e-01
7.12931007e-02 -1.06843181e-01 5.74184477e-01 3.22369367e-01
9.61820722e-01 -6.71262980e-01 9.11909580e-01 -5.48086345e-01
2.97717631e-01 -2.80930903e-02 9.14191008e-01 6.46710932e-01
-2.12938815e-01 2.40536764e-01 2.94635445e-01 -4.31294501e-01
-5.94740391e-01 -9.30496812e-01 -4.02524889e-01 1.46309447e+00
5.36576807e-01 -9.21456695e-01 -3.43314111e-01 -5.77142298e-01
2.08560362e-01 9.66251194e-01 -4.92669970e-01 -6.62607074e-01
-7.81739205e-02 -1.19636583e+00 3.89582992e-01 2.75279850e-01
2.54964560e-01 -8.74114931e-01 -1.65512592e-01 2.66642332e-01
-3.43868047e-01 -9.80734944e-01 -6.15048945e-01 2.28264719e-01
-5.32289982e-01 -1.01887846e+00 -4.18254018e-01 -8.10342431e-01
3.92618865e-01 5.48184514e-01 1.53970289e+00 9.75507796e-02
-3.17292094e-01 -6.19562939e-02 -8.31066728e-01 -8.75825405e-01
1.56184390e-01 4.27726358e-01 7.55874872e-01 -4.33539331e-01
1.34158242e+00 -4.97815818e-01 -3.83755565e-01 1.10203415e-01
-1.15701628e+00 -4.33365136e-01 1.60280377e-01 8.10815990e-01
5.51790476e-01 3.77013803e-01 8.41012001e-01 -9.50343251e-01
1.20972323e+00 -6.84361577e-01 -2.44991407e-01 2.06658363e-01
-1.14560366e+00 2.59207606e-01 2.76011378e-01 -7.37455606e-01
-9.26040232e-01 -4.00143534e-01 -2.11819246e-01 1.34366482e-01
1.04786858e-01 7.60803223e-01 -2.47223884e-01 4.08953846e-01
8.49566460e-01 -1.19900353e-01 -5.60612202e-01 -5.18225312e-01
3.25794071e-01 7.57755399e-01 6.64396882e-02 -5.62960923e-01
6.96361065e-01 9.46557149e-02 -4.72368300e-01 -9.23794866e-01
-1.40034080e+00 -9.75134969e-01 -3.17347407e-01 2.88581192e-01
5.08621275e-01 -9.34099495e-01 -1.98002845e-01 2.89823651e-01
-9.62778270e-01 2.06474721e-01 -2.27625132e-01 5.62423348e-01
1.05949514e-01 4.63927329e-01 -2.10663974e-01 -7.37766445e-01
-4.37035710e-01 -5.96024573e-01 9.62792814e-01 2.48974785e-01
-7.03294873e-01 -9.98108685e-01 3.86242509e-01 -2.17930190e-02
3.54360580e-01 -1.76312312e-01 1.16592586e+00 -1.14447320e+00
3.58001143e-01 -3.08891922e-01 -1.19889088e-01 1.22868143e-01
4.96619344e-01 -4.20238972e-01 -8.04174602e-01 -1.08476587e-01
-8.92262608e-02 -2.03261077e-01 1.18692482e+00 2.58561641e-01
1.14210057e+00 -2.64758527e-01 -4.97804254e-01 1.50413051e-01
1.43104959e+00 4.00188230e-02 4.22212124e-01 3.92220974e-01
5.47465265e-01 4.24778283e-01 8.99637341e-01 5.36346018e-01
4.44470674e-01 3.00964713e-01 7.39325732e-02 -6.37976304e-02
2.55771309e-01 -3.67514551e-01 -4.13484313e-02 9.95022058e-01
2.72346228e-01 -3.69193226e-01 -1.25650525e+00 8.51434886e-01
-1.61694372e+00 -8.16513836e-01 1.90177262e-02 2.22695422e+00
1.41362309e+00 4.75967199e-01 -1.51965305e-01 5.38071632e-01
9.28253114e-01 4.59351033e-01 -3.91888350e-01 -2.88741082e-01
-3.33578795e-01 3.98704231e-01 4.05472696e-01 6.23678088e-01
-9.94644165e-01 1.26855886e+00 6.31484699e+00 1.38681185e+00
-9.37628746e-01 9.53283235e-02 4.98058110e-01 -1.68370202e-01
-6.38284981e-01 -4.32449341e-01 -8.97520006e-01 5.02868652e-01
4.37377483e-01 -7.64523625e-01 6.45729005e-02 7.63227522e-01
3.18406522e-02 5.84909655e-02 -6.01518035e-01 1.25862920e+00
1.63874567e-01 -8.55701983e-01 3.78393173e-01 -2.41937712e-01
7.78280437e-01 2.83658534e-01 -2.25106061e-01 5.91169834e-01
5.37966549e-01 -9.46512163e-01 2.12490097e-01 1.35309502e-01
8.50758612e-01 -1.02425313e+00 9.09120262e-01 3.68954539e-01
-1.24917245e+00 -8.02490637e-02 -8.15941691e-01 -4.13625538e-01
6.36707321e-02 1.50069213e+00 -6.89899027e-01 3.42846394e-01
6.07340157e-01 6.61007345e-01 -3.21781456e-01 7.68354297e-01
-2.16788843e-01 5.99128306e-01 -2.25993380e-01 -2.79884219e-01
-2.45291721e-02 4.91343364e-02 6.46365881e-01 1.29587591e+00
2.84384936e-01 3.60572219e-01 3.64825547e-01 1.62219226e-01
-7.31241480e-02 4.36833173e-01 -4.90295142e-01 -1.80785790e-01
7.60906398e-01 1.00293076e+00 -5.91070414e-01 -4.79746312e-01
-2.64461488e-01 6.11459851e-01 2.37021148e-01 3.56869698e-01
-2.94104010e-01 -6.58448935e-01 1.37536883e+00 1.23686172e-01
-6.64163232e-02 -1.57632619e-01 -5.42945802e-01 -1.38722956e+00
-7.74468407e-02 -6.07658446e-01 8.87177050e-01 -4.13427472e-01
-2.02724242e+00 5.83314061e-01 -2.73631886e-02 -1.25384820e+00
-7.69612417e-02 -7.06550598e-01 -4.03433114e-01 7.91394711e-01
-1.93688810e+00 -2.89178401e-01 4.89851758e-02 4.22704637e-01
5.96191704e-01 -3.07749361e-01 1.05468619e+00 3.77707332e-01
-3.42608988e-01 7.38719761e-01 -1.05580762e-01 1.10775709e-01
8.75416279e-01 -1.30067503e+00 5.67006648e-01 5.36641896e-01
1.89012945e-01 9.01227355e-01 8.91008258e-01 -7.04748392e-01
-5.93086004e-01 -9.63127732e-01 1.57964623e+00 -2.79115826e-01
7.41569519e-01 -7.79724866e-02 -1.02181470e+00 9.04200226e-02
1.90195963e-02 -6.71623945e-02 1.21363020e+00 5.86526990e-01
-6.76098585e-01 -1.42629236e-01 -1.11312842e+00 5.89795053e-01
1.21341133e+00 -5.92990875e-01 -1.32402849e+00 2.13373065e-01
1.03834140e+00 -1.74853623e-01 -3.89091015e-01 4.72195625e-01
2.99072146e-01 -3.27689022e-01 1.23674810e+00 -9.97588336e-01
4.24689978e-01 -3.23940873e-01 -1.85441554e-01 -1.89871264e+00
-3.92309725e-01 -1.91950664e-01 9.78101790e-02 1.27724373e+00
5.42491019e-01 -6.52038991e-01 2.79964685e-01 2.96622664e-01
3.14636141e-01 -7.02560067e-01 -9.11604881e-01 -8.07694614e-01
2.32944354e-01 -8.00421834e-01 9.27613199e-01 1.51680970e+00
4.85783547e-01 7.15812147e-01 -9.87951010e-02 -1.77275501e-02
2.68823594e-01 5.75111946e-03 6.67201588e-03 -1.53578377e+00
4.90702093e-02 -4.21496034e-01 -5.92010558e-01 -9.07372355e-01
3.55168879e-01 -1.02870739e+00 -6.41568704e-03 -1.36859214e+00
1.70670733e-01 -3.13206464e-01 -9.90550637e-01 7.02965796e-01
-1.03361285e+00 1.40033305e-01 -1.03463814e-01 -1.38873652e-01
-7.07362771e-01 5.31817436e-01 1.10395658e+00 -3.22370797e-01
-2.01116234e-01 -2.16118217e-01 -1.07317460e+00 8.12102914e-01
9.34057117e-01 -7.12467909e-01 -6.11436427e-01 -4.43783224e-01
9.27590013e-01 -6.91531360e-01 -2.68527776e-01 -5.07091761e-01
-1.50970405e-03 -4.56281424e-01 2.27437600e-01 -3.23856473e-01
-7.30389357e-02 -5.43836713e-01 -7.24950671e-01 2.98884779e-01
-4.80540663e-01 2.37825319e-01 2.67319530e-01 5.28726399e-01
-2.36821398e-01 -8.37648660e-02 6.40718043e-01 -1.65534005e-01
-9.29695368e-01 2.52459586e-01 -2.94083804e-01 7.87881076e-01
5.08638859e-01 -5.93844838e-02 -1.10704489e-01 -3.08828354e-02
-2.90401965e-01 4.90744352e-01 1.77306056e-01 1.09005010e+00
5.30370414e-01 -1.63341355e+00 -7.36427248e-01 2.71104276e-01
7.53490090e-01 -4.22830209e-02 1.47390991e-01 1.34352863e-01
3.12029775e-02 1.38068259e-01 1.07040681e-01 -1.73164651e-01
-1.10769260e+00 5.24433672e-01 5.82662560e-02 -5.33695042e-01
-1.23694278e-01 1.08559132e+00 1.12023316e-01 -6.53335154e-01
1.02833733e-01 -3.69209886e-01 -6.12997115e-01 5.68049848e-01
1.09514773e+00 2.18314320e-01 4.16154563e-01 -4.60104436e-01
-6.09074235e-01 4.19802755e-01 -1.88498095e-01 1.33428514e-01
1.30730152e+00 -2.73712546e-01 -1.59216985e-01 6.07913315e-01
1.10157764e+00 1.43742114e-01 -3.83334786e-01 -6.56987488e-01
4.34411734e-01 -6.54835582e-01 7.03966469e-02 -6.48303568e-01
-8.04488778e-01 5.47279358e-01 5.72182059e-01 4.68498200e-01
9.80526507e-01 -5.39490953e-02 1.06627905e+00 5.56575358e-01
3.48954231e-01 -1.42964303e+00 -8.57431740e-02 1.05709684e+00
4.61501539e-01 -1.11301816e+00 -1.38400301e-01 -3.98194879e-01
-5.90754271e-01 7.09475517e-01 6.23275816e-01 -1.00820027e-02
8.88145745e-01 3.34468067e-01 3.76905203e-01 -1.36806816e-01
-5.77762783e-01 -5.00286639e-01 4.59728837e-01 6.09049499e-01
5.63180566e-01 2.30192214e-01 -9.79067445e-01 1.16726136e+00
-5.23043156e-01 -2.03697667e-01 4.15211283e-02 4.84540522e-01
-7.94165790e-01 -1.25777268e+00 -1.58168122e-01 7.29673624e-01
-5.50793231e-01 -6.92370057e-01 -4.14576471e-01 4.10723865e-01
3.76291871e-01 1.16614139e+00 2.01971963e-01 -4.90062326e-01
4.24898207e-01 2.73800462e-01 4.06417996e-02 -8.54772806e-01
-2.43248522e-01 -3.80593807e-01 -4.36879806e-02 -4.83866543e-01
-2.09454179e-01 -1.77225962e-01 -1.45260227e+00 -1.87442407e-01
-4.59920824e-01 3.12873423e-01 1.75413415e-01 1.29968202e+00
3.77193332e-01 5.21486223e-01 5.32600105e-01 -1.83475211e-01
-5.99478364e-01 -1.11477733e+00 -9.73782897e-01 8.74214053e-01
2.43441939e-01 -7.94331670e-01 -4.21742231e-01 -3.55154961e-01] | [10.265185356140137, 8.96411418914795] |
2cd8efe0-ec58-4c68-8031-22917928a158 | financial-trading-model-with-stock-bar-chart | 1903.04610 | null | http://arxiv.org/abs/1903.04610v1 | http://arxiv.org/pdf/1903.04610v1.pdf | Financial Trading Model with Stock Bar Chart Image Time Series with Deep Convolutional Neural Networks | Even though computational intelligence techniques have been extensively
utilized in financial trading systems, almost all developed models use the time
series data for price prediction or identifying buy-sell points. However, in
this study we decided to use 2-D stock bar chart images directly without
introducing any additional time series associated with the underlying stock. We
propose a novel algorithmic trading model CNN-BI (Convolutional Neural Network
with Bar Images) using a 2-D Convolutional Neural Network. We generated 2-D
images of sliding windows of 30-day bar charts for Dow 30 stocks and trained a
deep Convolutional Neural Network (CNN) model for our algorithmic trading
model. We tested our model separately between 2007-2012 and 2012-2017 for
representing different market conditions. The results indicate that the model
was able to outperform Buy and Hold strategy, especially in trendless or bear
markets. Since this is a preliminary study and probably one of the first
attempts using such an unconventional approach, there is always potential for
improvement. Overall, the results are promising and the model might be
integrated as part of an ensemble trading model combined with different
strategies. | ['Omer Berat Sezer', 'Ahmet Murat Ozbayoglu'] | 2019-03-11 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-7.25487351e-01 -1.63339153e-01 2.29812726e-01 -2.16636121e-01
-3.81128639e-01 -8.64008009e-01 8.81888986e-01 -1.12564139e-01
-4.29436237e-01 4.80323136e-01 -1.29983261e-01 -7.16283321e-01
-1.84422076e-01 -1.14456820e+00 -5.94842672e-01 -2.64694512e-01
-5.43736696e-01 5.54269433e-01 2.66003937e-01 -6.90100133e-01
5.08190513e-01 8.47029805e-01 -1.37857211e+00 3.28092068e-01
3.04425269e-01 1.76541400e+00 -4.61990356e-01 6.45845771e-01
-5.13316691e-01 9.27551210e-01 -8.09457123e-01 -6.50470376e-01
1.19898415e+00 -3.60057741e-01 -9.08062458e-02 -1.32074192e-01
2.97246277e-01 -4.90726113e-01 2.24018306e-03 8.21564674e-01
2.23812982e-01 -1.18167512e-01 3.75856817e-01 -8.30785215e-01
-9.93054032e-01 8.07377875e-01 -6.29579067e-01 5.89042842e-01
-3.28618467e-01 2.13708282e-02 1.16177964e+00 -8.03426087e-01
4.50705618e-01 7.68999994e-01 9.36882734e-01 -1.17343612e-01
-1.20400500e+00 -7.06976354e-01 -4.68407832e-02 5.87500725e-03
-5.82447112e-01 1.70075297e-01 1.00236106e+00 -5.84454954e-01
1.06373274e+00 1.42705768e-01 1.20430994e+00 5.40214360e-01
7.18857825e-01 7.21346796e-01 1.35313892e+00 -1.62044987e-01
1.68338776e-01 3.90396948e-04 1.14348635e-01 1.33041188e-01
3.77743989e-01 5.45546710e-01 -1.87899053e-01 7.38653122e-04
1.03981733e+00 2.13406175e-01 2.02010468e-01 -2.31155694e-01
-1.17591989e+00 1.28725266e+00 6.90893114e-01 7.11332619e-01
-5.41847289e-01 1.09067939e-01 4.27053750e-01 9.69928980e-01
9.80219483e-01 7.73285031e-01 -8.11569214e-01 -1.69929400e-01
-1.47256672e+00 7.70131826e-01 9.66370463e-01 1.93853989e-01
4.01088953e-01 5.49053073e-01 2.30710745e-01 5.07234931e-01
-2.40184683e-02 2.66997278e-01 8.40856969e-01 -4.78642106e-01
2.43784964e-01 7.52748191e-01 1.61641017e-01 -1.04350066e+00
-6.92970395e-01 -5.14623046e-01 -6.27139390e-01 9.80785966e-01
7.19037890e-01 -4.14681584e-01 -9.35991824e-01 9.06426728e-01
-3.26976240e-01 9.26445499e-02 1.01894833e-01 7.76973069e-01
3.75847101e-01 5.97406566e-01 -4.75656480e-01 -6.87759817e-02
1.29355538e+00 -7.52930999e-01 -6.23762012e-01 2.77630329e-01
2.22573102e-01 -8.99200857e-01 4.11005735e-01 6.34334624e-01
-1.16491449e+00 -6.92292154e-01 -1.36787724e+00 3.30515802e-01
-1.02003407e+00 -5.69779193e-03 6.51450038e-01 5.80823958e-01
-1.09664190e+00 1.00470150e+00 -6.20919287e-01 2.63532460e-01
1.87483847e-01 2.61173338e-01 -1.55983388e-01 8.85218561e-01
-1.39909530e+00 1.05909801e+00 3.35391879e-01 5.32623768e-01
-1.70800701e-01 -6.01073503e-01 -5.26925981e-01 6.53097555e-02
-3.65867279e-02 -8.97474959e-02 1.41554201e+00 -1.23244298e+00
-1.57144582e+00 5.39089382e-01 7.80534267e-01 -1.55707836e+00
1.15024996e+00 -3.29455763e-01 -6.18166089e-01 -2.60387033e-01
-4.24059927e-01 4.14425015e-01 8.33299816e-01 -7.56017208e-01
-6.92290425e-01 -2.75155604e-01 -3.09123285e-02 -3.18754703e-01
1.24207705e-01 2.09895689e-02 2.78424650e-01 -1.31850767e+00
8.19407851e-02 -8.22857738e-01 -3.46505642e-02 -2.62313038e-01
1.03552550e-01 9.65201780e-02 8.34992409e-01 -7.80514777e-01
1.00543749e+00 -1.85104656e+00 -4.66327041e-01 3.69005919e-01
3.22742690e-03 3.22593778e-01 1.55026093e-01 4.65559989e-01
-4.54760998e-01 2.04014346e-01 -2.53615260e-01 1.31921411e-01
1.30904943e-01 -1.62022591e-01 -7.77343869e-01 3.11960906e-01
5.00416815e-01 1.27521777e+00 -2.87242144e-01 2.07994908e-01
3.56989026e-01 2.65665919e-01 -7.73498639e-02 -1.22935250e-01
-5.20645738e-01 4.80247438e-02 -9.59883779e-02 4.61291313e-01
9.89112020e-01 9.84368324e-02 -2.88123935e-01 -4.09117229e-02
-7.37670958e-01 7.24967644e-02 -1.09049177e+00 9.18484330e-01
-1.90554112e-01 9.08706844e-01 -4.27651525e-01 -9.37336802e-01
1.51474059e+00 2.59955525e-01 6.27607763e-01 -1.18618178e+00
1.60724446e-01 6.99774206e-01 4.38240141e-01 2.22103447e-02
5.69431365e-01 -3.73864025e-01 1.12748004e-01 6.01309121e-01
-1.65380076e-01 -2.61027068e-01 1.80326775e-01 -6.73632979e-01
6.86108649e-01 2.30476066e-01 4.21648920e-02 -2.40074530e-01
3.85754675e-01 2.96565115e-01 1.74896315e-01 4.50092703e-01
-1.96005881e-01 7.57927954e-01 7.75258660e-01 -1.33917987e+00
-1.16625094e+00 -8.19556057e-01 -2.98824251e-01 5.59678495e-01
-4.88680094e-01 2.90060073e-01 -3.96761119e-01 -4.08336580e-01
4.54898804e-01 7.52067029e-01 -1.02643657e+00 4.27879095e-01
-7.29301989e-01 -9.25101399e-01 4.45440084e-01 5.25925696e-01
8.32521677e-01 -1.41816330e+00 -9.63433564e-01 6.72311127e-01
9.52377737e-01 -4.28793371e-01 -7.67222866e-02 6.89668119e-01
-1.06432843e+00 -1.06680298e+00 -1.23788643e+00 -3.74415070e-01
-1.07882485e-01 -2.55350411e-01 1.16752040e+00 -2.76886612e-01
3.00354809e-02 -1.50976986e-01 -3.07907611e-01 -1.17329788e+00
3.98314121e-04 -4.28725891e-02 -3.96396667e-01 1.21588781e-01
5.02151191e-01 -2.87641168e-01 -7.51066625e-01 2.63879389e-01
-9.24066126e-01 -3.66810530e-01 4.99506325e-01 7.75855839e-01
2.57573277e-01 5.81797631e-03 8.15937877e-01 -7.39869833e-01
1.01257324e+00 -2.66603798e-01 -1.31500804e+00 3.10700620e-03
-7.64531493e-01 7.70722479e-02 4.52643931e-01 -2.97148168e-01
-8.24707389e-01 -2.59614468e-01 -9.02535766e-02 -5.59756637e-01
3.24319214e-01 7.42655337e-01 7.60285199e-01 1.47788048e-01
3.13276082e-01 1.04209527e-01 3.42715681e-01 -5.88587165e-01
1.16603538e-01 1.16984598e-01 2.46463224e-01 -5.55158185e-04
8.05231035e-01 3.81857932e-01 -1.10604152e-01 -3.64218652e-01
-4.07143891e-01 -4.85666804e-02 -7.99404323e-01 -2.95004249e-01
9.09192145e-01 -7.24563777e-01 -5.56128323e-01 9.03828382e-01
-8.04482698e-01 -4.26332563e-01 -3.37905407e-01 4.68711346e-01
-4.99246836e-01 -2.77895212e-01 -7.02258229e-01 -1.03862381e+00
-4.24548060e-01 -9.08325553e-01 5.92860699e-01 2.13052437e-01
7.83339422e-03 -1.28352046e+00 6.13406062e-01 -2.49902129e-01
9.46697533e-01 7.03598201e-01 6.78949714e-01 -1.46854973e+00
-4.99044836e-01 -6.86486959e-01 -2.91190714e-01 5.76494634e-01
1.59321144e-01 3.43270957e-01 -9.59949315e-01 8.22500512e-02
2.92949378e-01 -6.07898980e-02 1.06199515e+00 5.45622110e-01
6.16419435e-01 -1.04622304e-01 5.34871221e-01 3.90833318e-01
1.42196977e+00 8.84007394e-01 6.72797203e-01 1.11734629e+00
1.40060512e-02 4.75495636e-01 4.37850773e-01 3.95238668e-01
-2.90485807e-02 3.22231978e-01 4.14134026e-01 -1.79566368e-01
3.90773654e-01 1.64688855e-01 3.82508934e-01 5.50539672e-01
-3.26320589e-01 1.62714779e-01 -9.95324552e-01 3.31977218e-01
-1.71578681e+00 -1.20414710e+00 1.53832221e-02 1.97347188e+00
2.71936208e-01 8.50037575e-01 4.59162086e-01 1.53108001e-01
3.91279817e-01 4.29357111e-01 -6.19902670e-01 -8.48844647e-01
-4.12584633e-01 5.90444088e-01 8.09178531e-01 -6.89424202e-02
-1.38704479e+00 3.75652552e-01 6.45182610e+00 3.06962311e-01
-1.61363280e+00 -4.24828380e-01 1.08149397e+00 -9.61570367e-02
-1.08165599e-01 -3.83961052e-01 -4.81208503e-01 4.11867023e-01
1.19220412e+00 -3.47033828e-01 1.00521795e-01 8.62714946e-01
1.19803436e-01 1.76601112e-01 -7.93965220e-01 5.77265203e-01
-2.49663308e-01 -1.85651696e+00 -3.40737440e-02 2.89103866e-01
7.30601311e-01 1.56181619e-01 4.26565528e-01 4.41229761e-01
2.11832657e-01 -9.41317439e-01 1.05714178e+00 9.37622666e-01
6.86335936e-02 -1.01831770e+00 1.27326608e+00 -6.20684735e-02
-1.14928138e+00 -7.72678852e-02 -2.59252101e-01 -1.58312410e-01
-5.15703112e-02 2.68438548e-01 -5.23115695e-01 5.59729934e-01
7.50097930e-01 7.92800009e-01 -7.92532325e-01 9.37034011e-01
5.11217117e-01 2.62756824e-01 -4.14917827e-01 -1.36435717e-01
8.47453535e-01 -5.27920187e-01 8.05148110e-02 8.94004464e-01
8.18909585e-01 -1.67030916e-01 -3.24481040e-01 8.02221179e-01
6.49962127e-02 1.71662167e-01 -7.09023178e-01 -2.82914162e-01
-1.88011616e-01 9.86162841e-01 -1.07471073e+00 -3.11720341e-01
-8.78093719e-01 4.86215353e-01 -9.40992609e-02 1.09538533e-01
-7.43197203e-01 -4.39209163e-01 6.12214744e-01 2.46834606e-01
8.29578042e-01 -2.38969341e-01 -7.14080811e-01 -1.02033901e+00
-2.62141787e-02 -5.94514847e-01 4.78588730e-01 -8.29805136e-01
-1.51439834e+00 8.55226934e-01 -3.65325771e-02 -1.68585408e+00
-6.11695051e-01 -1.22677052e+00 -1.02136433e+00 8.98064196e-01
-1.62180102e+00 -6.30799472e-01 2.79412448e-01 2.64298946e-01
5.41171849e-01 -8.19936574e-01 5.35954952e-01 -1.50237381e-01
-2.23271146e-01 1.05349734e-01 5.52535415e-01 5.21272898e-01
5.10324657e-01 -1.62557197e+00 8.64448786e-01 6.34574533e-01
4.19706047e-01 3.57739210e-01 5.74273467e-01 -6.57772779e-01
-8.64609241e-01 -8.44114006e-01 5.30944049e-01 -1.48121223e-01
1.45401239e+00 -2.88703702e-02 -1.13181233e+00 7.29334354e-01
8.19792092e-01 -2.07598984e-01 5.33096373e-01 -4.10888046e-01
-2.95477599e-01 -3.36612374e-01 -1.04682004e+00 3.20355445e-01
7.12851882e-02 -2.63229962e-02 -1.05630875e+00 -1.82691976e-01
4.20964569e-01 -2.67980009e-01 -1.05041683e+00 2.33010188e-01
7.03416526e-01 -1.50562215e+00 7.59167552e-01 -2.11913079e-01
4.39635813e-01 -9.34541896e-02 1.39824659e-01 -1.38692772e+00
9.91129279e-02 -5.29695392e-01 -1.57848105e-01 7.81011462e-01
8.59392524e-01 -1.03787029e+00 7.64001369e-01 5.95145226e-01
1.23256527e-01 -6.18850529e-01 -1.05163610e+00 -7.77303755e-01
6.32203519e-01 -5.77196956e-01 9.91014600e-01 7.27380514e-01
-2.56621867e-01 -2.75886983e-01 -2.64467835e-01 -3.87802064e-01
2.98600614e-01 4.60665256e-01 5.09685040e-01 -1.54394150e+00
-6.89010993e-02 -1.24880838e+00 -3.50019664e-01 -3.24086189e-01
-1.33780882e-01 -4.81451601e-01 -4.89218801e-01 -1.13086522e+00
-5.68590999e-01 -1.32867187e-01 -7.64863253e-01 1.29010469e-01
3.19825888e-01 2.39597231e-01 4.72624928e-01 3.57766420e-01
2.61586070e-01 1.76481843e-01 1.23346937e+00 -1.20168127e-01
-2.30320886e-01 3.41627866e-01 -2.70150363e-01 5.75382769e-01
9.87832725e-01 1.81828663e-01 -8.75880569e-02 -8.05089101e-02
4.49912161e-01 1.04005396e-01 2.63295650e-01 -8.94024193e-01
3.49169970e-02 2.76871443e-01 8.56185377e-01 -1.23377657e+00
9.22775939e-02 -9.56990123e-01 4.78458703e-01 6.05916977e-01
-2.70125240e-01 9.40271556e-01 5.26521981e-01 2.70742327e-01
-8.31288636e-01 -1.37695178e-01 6.60088360e-01 -4.24705446e-01
-6.52702510e-01 1.40129432e-01 -3.26895267e-01 -5.05984187e-01
1.19112754e+00 -4.14514542e-01 -2.12149233e-01 -3.71771663e-01
-8.48609865e-01 7.30488449e-02 3.23204637e-01 5.60793161e-01
4.55483198e-01 -1.36871040e+00 -7.81006277e-01 3.28468084e-01
-1.43462598e-01 -2.88234562e-01 -1.21632189e-01 4.81105119e-01
-1.30668652e+00 7.38324225e-01 -6.90987110e-01 -1.70710191e-01
-5.68897307e-01 7.12221742e-01 7.02828526e-01 -5.41950464e-01
-7.42658794e-01 5.14074743e-01 -2.39677981e-01 -9.28609893e-02
-1.80118848e-02 -9.37382340e-01 -5.41718423e-01 9.66687918e-01
5.58285058e-01 1.60212457e-01 3.77188802e-01 -3.71024221e-01
1.97338848e-03 5.72124839e-01 -1.56081945e-01 -4.19560432e-01
2.00879049e+00 5.12062430e-01 -9.00551863e-03 9.50356185e-01
9.93873656e-01 -1.86785564e-01 -1.42978930e+00 1.82760268e-01
7.24559426e-01 -2.26063445e-01 -3.01173851e-02 -6.77973747e-01
-1.58537984e+00 7.66662061e-01 6.41686201e-01 1.02356744e+00
9.98925090e-01 -5.30376077e-01 7.24676907e-01 2.26269439e-01
2.17386439e-01 -9.79165316e-01 -1.64350703e-01 6.11018181e-01
1.19232082e+00 -1.21883309e+00 -1.25417620e-01 5.40273726e-01
-6.46089852e-01 1.78909397e+00 2.04412296e-01 -9.78569686e-01
1.16988873e+00 4.12729234e-01 5.55481911e-01 -3.20598185e-01
-9.00141954e-01 -1.83953419e-01 4.17007595e-01 1.55627400e-01
7.04995573e-01 -8.88312161e-02 -2.77386934e-01 5.11267126e-01
-6.40653789e-01 9.00379717e-02 5.29067576e-01 9.10642326e-01
-2.88166404e-01 -9.99291241e-01 -4.95419800e-01 5.81880867e-01
-7.95378745e-01 -5.17220981e-02 -4.84280109e-01 1.54009283e+00
-1.26217961e-01 2.73003429e-01 8.04485381e-01 -2.24575892e-01
6.24513328e-01 4.78272110e-01 7.04657063e-02 -1.40242890e-01
-1.27022386e+00 4.17986929e-01 -1.07388198e-01 -1.91848546e-01
-4.25554931e-01 -9.42709386e-01 -8.35670948e-01 -4.84200805e-01
-1.28552586e-01 1.00923397e-01 6.87441409e-01 4.86835539e-01
-4.56565320e-02 5.73705316e-01 7.60051906e-01 -1.09698391e+00
-5.86047411e-01 -1.11190295e+00 -9.52590346e-01 1.50939390e-01
7.01897919e-01 -6.05208755e-01 -3.03448558e-01 -3.56415063e-02] | [4.366585731506348, 4.283901214599609] |
7b124d0f-f9df-48d4-b1d2-adcc5d162785 | toward-reliable-human-pose-forecasting-with | 2304.06707 | null | https://arxiv.org/abs/2304.06707v1 | https://arxiv.org/pdf/2304.06707v1.pdf | Toward Reliable Human Pose Forecasting with Uncertainty | Recently, there has been an arms race of pose forecasting methods aimed at solving the spatio-temporal task of predicting a sequence of future 3D poses of a person given a sequence of past observed ones. However, the lack of unified benchmarks and limited uncertainty analysis have hindered progress in the field. To address this, we first develop an open-source library for human pose forecasting, featuring multiple models, datasets, and standardized evaluation metrics, with the aim of promoting research and moving toward a unified and fair evaluation. Second, we devise two types of uncertainty in the problem to increase performance and convey better trust: 1) we propose a method for modeling aleatoric uncertainty by using uncertainty priors to inject knowledge about the behavior of uncertainty. This focuses the capacity of the model in the direction of more meaningful supervision while reducing the number of learned parameters and improving stability; 2) we introduce a novel approach for quantifying the epistemic uncertainty of any model through clustering and measuring the entropy of its assignments. Our experiments demonstrate up to $25\%$ improvements in accuracy and better performance in uncertainty estimation. | ['Alexandre Alahi', 'Taylor Mordan', 'Zahra Tehraninasab', 'Amirhossein Alimohammadi', 'Yashar Zoroofchi Benisi', 'Parham Saremi', 'Matin Daghyani', 'Mehrshad Mirmohammadi', 'Saeed Saadatnejad'] | 2023-04-13 | null | null | null | null | ['human-pose-forecasting'] | ['computer-vision'] | [-4.97412644e-02 3.05132717e-01 7.13155642e-02 -6.40175045e-01
-7.89001405e-01 -4.37445939e-01 6.91616952e-01 1.48050785e-01
-3.07246923e-01 7.10160911e-01 4.91424769e-01 1.05851352e-01
-3.02322209e-01 -5.33841193e-01 -7.93983519e-01 -5.42783618e-01
-1.62676245e-01 7.84035623e-01 2.60979354e-01 1.35024518e-01
5.21523394e-02 3.28943193e-01 -1.49162149e+00 -1.45432100e-01
6.89694762e-01 1.09920597e+00 -2.49774992e-01 3.24392587e-01
4.62816626e-01 6.09577477e-01 -5.06776452e-01 -7.63167620e-01
1.60966739e-01 -2.34644830e-01 -9.46812749e-01 -4.95141335e-02
7.91127607e-02 -1.00388773e-01 -1.83737308e-01 7.54812181e-01
4.17470634e-01 3.45843762e-01 8.51426482e-01 -1.26258147e+00
-1.68204919e-01 7.53412604e-01 -2.25100100e-01 1.01463661e-01
4.27746296e-01 1.48638949e-01 8.95204842e-01 -3.31343144e-01
5.63215435e-01 1.21499848e+00 7.29631841e-01 6.14947736e-01
-9.88200545e-01 -2.52141565e-01 3.25839758e-01 2.06417814e-01
-1.42968559e+00 -2.38465533e-01 6.04677498e-01 -6.74181223e-01
5.76992571e-01 2.88429886e-01 4.67243999e-01 1.35836411e+00
8.94281045e-02 6.60507262e-01 8.36671114e-01 -1.44002721e-01
5.24618626e-01 1.68204829e-01 7.91755393e-02 5.26034713e-01
1.44372806e-01 1.05083920e-01 -7.62706876e-01 -2.07359374e-01
3.72477800e-01 -1.29847482e-01 -8.87896493e-02 -5.33780217e-01
-1.09260035e+00 4.58393961e-01 4.41150904e-01 1.36142284e-01
-1.87297806e-01 2.75179684e-01 4.50472124e-02 -2.49449253e-01
6.73297346e-01 5.03383517e-01 -3.41436595e-01 -3.37374479e-01
-8.23025942e-01 5.67898273e-01 7.60321498e-01 5.88376284e-01
4.84023482e-01 -4.52972770e-01 -2.10855290e-01 3.57772231e-01
5.51954865e-01 4.22313809e-01 1.08576126e-01 -1.13162303e+00
4.86455828e-01 4.03620660e-01 3.91253620e-01 -1.07413733e+00
-5.33574164e-01 -4.94179428e-01 -5.00646949e-01 8.60640332e-02
5.71634412e-01 -2.53028452e-01 -6.45943463e-01 1.93453515e+00
3.45279783e-01 2.95095950e-01 -2.84883291e-01 9.60485041e-01
1.41550809e-01 4.10790026e-01 -5.45759611e-02 1.16714314e-01
9.36331809e-01 -6.81526005e-01 -2.74471104e-01 -3.11293125e-01
4.19350863e-01 -3.55595559e-01 7.53494263e-01 4.29135859e-01
-1.00077975e+00 -3.82414252e-01 -9.05802608e-01 2.08263516e-01
-2.19362512e-01 -4.41395231e-02 5.43738127e-01 7.90869594e-01
-6.46380305e-01 8.98365378e-01 -1.20979786e+00 -1.94362879e-01
1.34117424e-01 2.07196444e-01 -1.77701905e-01 1.27611041e-01
-1.24729633e+00 1.17709816e+00 2.20859826e-01 1.94763899e-01
-7.59985387e-01 -6.88139498e-01 -8.23185444e-01 -1.05399683e-01
5.84604621e-01 -7.24981070e-01 9.52963233e-01 -5.76995075e-01
-1.30626774e+00 4.02253509e-01 1.48590952e-01 -5.69917560e-01
9.01417792e-01 -5.46878994e-01 -1.56262681e-01 -1.82433948e-01
6.22188002e-02 7.54475892e-01 7.62588561e-01 -1.24359322e+00
-4.09300297e-01 -6.39978051e-01 -4.10267562e-02 3.14028323e-01
-2.44815171e-01 -2.63037473e-01 -6.12913787e-01 -4.58604842e-01
7.79230222e-02 -1.14997804e+00 -3.13441575e-01 -1.13259092e-01
-3.11569661e-01 -1.23430490e-01 1.32476479e-01 -7.40852892e-01
1.22618508e+00 -1.88516927e+00 5.02297759e-01 4.70025301e-01
8.54568481e-02 -2.54555583e-01 2.86546111e-01 2.11997330e-01
3.57707590e-01 5.84601909e-02 -3.94386321e-01 -8.04755390e-01
2.76789129e-01 2.42377803e-01 -4.18189198e-01 4.19659466e-01
1.70671269e-01 6.67914212e-01 -8.58044982e-01 -4.36021119e-01
2.78929114e-01 5.67356110e-01 -5.62624156e-01 1.36894137e-02
-5.17296314e-01 7.87680447e-01 -3.65654856e-01 4.74758118e-01
3.00610751e-01 -3.37377578e-01 1.78199589e-01 -5.18054627e-02
1.89553440e-01 2.11608887e-01 -1.56640768e+00 1.50812805e+00
-1.73708111e-01 2.60759354e-01 -3.16845834e-01 -4.80228186e-01
7.41457164e-01 8.85144621e-02 7.87128508e-01 -5.55523075e-02
3.48263800e-01 -1.17788009e-01 -3.94645095e-01 -2.62275577e-01
5.15669227e-01 2.22068027e-01 -3.88983399e-01 3.42566013e-01
-1.16906621e-01 -1.06773846e-01 9.52864289e-02 7.85054043e-02
9.48156536e-01 4.35088426e-01 -2.24394143e-01 -1.13658749e-01
2.71400541e-01 -2.40806609e-01 4.25536215e-01 6.25179350e-01
-2.81713486e-01 6.41312420e-01 3.81880850e-01 -3.79264295e-01
-7.47792125e-01 -1.18835115e+00 -2.08721280e-01 8.74209762e-01
2.32529655e-01 -4.41636086e-01 -7.60366559e-01 -6.11189067e-01
1.61082953e-01 8.06375027e-01 -8.55010033e-01 -1.31672442e-01
-2.73732841e-01 -9.13885832e-01 4.74226952e-01 6.65625334e-01
3.46130788e-01 -5.22219718e-01 -6.92670703e-01 -1.70103312e-01
-5.77382088e-01 -1.00889599e+00 -3.90264988e-01 -1.31147161e-01
-8.22271764e-01 -1.01476061e+00 -5.38473725e-01 -6.48276582e-02
5.37369132e-01 -4.13478822e-01 1.08995867e+00 -3.05305034e-01
-8.91361311e-02 6.97186291e-01 -1.90917566e-01 -4.78418797e-01
-6.57090694e-02 1.56719312e-01 3.26814473e-01 1.48152933e-02
1.43732727e-01 -4.35359031e-01 -5.34944713e-01 4.91961539e-01
-7.34731436e-01 -2.09072232e-01 3.89451087e-01 4.51298684e-01
3.72275561e-01 2.16642790e-03 3.18168253e-01 -4.45431769e-01
6.63056076e-01 -4.76376593e-01 -5.52146614e-01 3.59956920e-01
-8.86639595e-01 5.46523213e-01 5.94958663e-03 -1.75561115e-01
-1.20911622e+00 1.55821845e-01 -4.28123574e-04 -2.76526362e-01
7.78022632e-02 4.62486058e-01 7.59067610e-02 1.63427025e-01
8.86833012e-01 -1.46602854e-01 7.64348283e-02 -3.93257320e-01
4.16785389e-01 3.21397275e-01 6.91095054e-01 -9.01313424e-01
4.57310259e-01 5.51494956e-01 1.61055937e-01 -3.81782502e-01
-1.05539560e+00 -3.36761445e-01 -7.19558001e-01 -5.57737589e-01
6.98688686e-01 -8.50176692e-01 -8.62128139e-01 2.87694424e-01
-8.38663042e-01 -2.03078911e-01 -2.36500368e-01 5.97116590e-01
-5.66455603e-01 5.87662399e-01 -4.33974624e-01 -9.67745662e-01
1.39367068e-02 -1.07683361e+00 1.22509897e+00 5.51415011e-02
-4.03068006e-01 -8.89249384e-01 4.63172704e-01 7.04520583e-01
1.32399037e-01 3.25509012e-01 4.01505411e-01 -3.96959275e-01
-6.05906427e-01 -2.41537184e-01 1.23440608e-01 2.56036848e-01
-2.58858591e-01 -6.38916865e-02 -9.31087255e-01 -1.13958783e-01
-7.11199343e-02 -1.57643482e-01 7.86535740e-01 4.11907583e-01
1.05408943e+00 -1.42275775e-02 -5.69764972e-01 2.76222676e-01
8.77490163e-01 -2.86781073e-01 4.67010856e-01 2.45132819e-01
2.78551638e-01 8.21303546e-01 6.11995637e-01 6.74393475e-01
7.25134611e-01 7.77985573e-01 5.07315516e-01 5.76152086e-01
2.22065240e-01 -3.06160390e-01 2.79712200e-01 2.83270150e-01
-3.70234668e-01 -1.65305540e-01 -1.11492658e+00 4.10172552e-01
-2.16247869e+00 -6.83423162e-01 3.08053285e-01 2.38487816e+00
6.37768805e-01 2.70841449e-01 2.68321097e-01 3.80487926e-02
5.48005700e-01 -1.54358204e-02 -5.48614204e-01 2.30202794e-01
3.10000151e-01 -3.93500268e-01 3.75117660e-01 6.16593480e-01
-1.09569812e+00 5.44272721e-01 6.75458622e+00 4.83908355e-01
-8.11206400e-01 -1.70592815e-01 7.87006557e-01 -3.86300564e-01
-2.70761400e-01 -1.41194314e-01 -8.98398221e-01 6.68282747e-01
9.12711203e-01 -1.86816268e-02 2.72749454e-01 8.89690101e-01
-5.26069142e-02 -2.98439890e-01 -1.19298458e+00 7.64569283e-01
2.39644513e-01 -1.16435313e+00 -1.97740301e-01 2.84328703e-02
6.23673260e-01 -8.26795772e-02 1.22377560e-01 1.69422135e-01
3.83225054e-01 -8.00005138e-01 9.34516191e-01 1.10519481e+00
1.24279551e-01 -8.77694964e-01 7.87078738e-01 5.40855169e-01
-8.27018559e-01 -2.54615117e-02 2.62084231e-02 -1.29509240e-01
2.53196865e-01 7.12476909e-01 -7.77177036e-01 6.37804449e-01
7.96291053e-01 4.56347883e-01 -6.29514813e-01 8.40690434e-01
-1.99410617e-01 3.78929734e-01 -6.67578697e-01 -1.22980289e-01
-1.29416406e-01 -3.49643007e-02 5.38178027e-01 8.36486757e-01
2.07962200e-01 -1.26054585e-01 1.92170337e-01 8.74027669e-01
1.74566850e-01 -2.61910528e-01 -2.73765981e-01 1.35576263e-01
4.88329619e-01 8.87778759e-01 -6.65621459e-01 1.42636016e-01
1.34591162e-01 8.18150401e-01 3.36210072e-01 -6.25384375e-02
-1.06124818e+00 1.87573791e-01 4.98738199e-01 1.14287950e-01
-3.86969186e-02 -4.62828785e-01 -5.52460134e-01 -1.14243591e+00
2.36689448e-01 -4.95017260e-01 6.36794984e-01 -6.79289818e-01
-1.28882957e+00 6.10741615e-01 4.64972377e-01 -9.26123798e-01
-6.19765341e-01 -5.40331066e-01 -1.63339734e-01 5.65692484e-01
-9.95149195e-01 -9.78816211e-01 -2.55745053e-01 3.46344233e-01
6.23339936e-02 3.85942534e-02 5.59813321e-01 7.96831474e-02
-5.31976283e-01 4.11284655e-01 -1.34408459e-01 -2.17138454e-01
8.52392316e-01 -1.20824075e+00 2.14682609e-01 8.78734291e-01
6.10473230e-02 4.44201678e-01 1.07157338e+00 -8.30020547e-01
-9.54602182e-01 -8.60575616e-01 7.87665188e-01 -1.23335719e+00
7.04751074e-01 -3.81805003e-01 -4.09877747e-01 7.53376186e-01
-4.52019542e-01 -3.43376279e-01 5.41676342e-01 5.39047658e-01
-3.08609962e-01 -3.52351032e-02 -1.21246982e+00 5.08187652e-01
1.02752125e+00 -3.28057498e-01 -5.58180451e-01 1.89751819e-01
5.83174467e-01 -4.80365157e-01 -1.02759910e+00 6.02485836e-01
6.58660233e-01 -1.11883330e+00 9.37943995e-01 -2.80841053e-01
2.79864460e-01 -2.81430721e-01 -2.13617891e-01 -1.14625382e+00
-1.49014935e-01 -4.04613614e-01 -3.69965702e-01 1.04053414e+00
3.85181308e-01 -3.67656738e-01 1.05904055e+00 1.18452072e+00
8.68135393e-02 -8.25555801e-01 -9.46183741e-01 -7.02728808e-01
-1.23662367e-01 -8.78970981e-01 5.90752423e-01 4.00887281e-01
5.85832745e-02 1.64673895e-01 -6.20089531e-01 3.75052243e-01
8.62355292e-01 -1.45129710e-01 6.10701323e-01 -1.30405784e+00
-3.95740598e-01 -2.29685947e-01 -4.84305859e-01 -8.98559153e-01
2.33008321e-02 -4.39699858e-01 1.73142374e-01 -1.18167675e+00
1.99353263e-01 -3.43969136e-01 -6.16545863e-02 2.51806736e-01
-1.02777272e-01 8.41766521e-02 1.29658625e-01 3.06095779e-01
-9.60389018e-01 5.57479501e-01 7.35228419e-01 3.32599580e-02
-3.10071502e-02 4.02683824e-01 -5.45748711e-01 8.04265678e-01
5.57362437e-01 -3.26920599e-01 -4.18843120e-01 -3.56506497e-01
5.68737090e-01 -4.16517816e-02 6.31446600e-01 -1.36566770e+00
2.64120191e-01 -8.15459117e-02 4.48136628e-01 -5.95363796e-01
6.12510383e-01 -7.86616623e-01 2.95782775e-01 2.62279540e-01
-4.77504283e-01 -2.39659280e-01 5.07376119e-02 1.01806259e+00
-3.42123024e-02 -1.52436763e-01 5.13410509e-01 -1.97942983e-02
-6.03681564e-01 1.88244298e-01 1.04450518e-02 -7.42783397e-02
1.05793059e+00 1.11076303e-01 -5.68753332e-02 -6.40885592e-01
-1.07568276e+00 4.82632965e-01 4.13488477e-01 5.89471757e-01
2.24613726e-01 -1.02530992e+00 -3.93228501e-01 4.78888042e-02
1.35688916e-01 -3.91077958e-02 3.41663301e-01 5.92770934e-01
-1.71405494e-01 4.21012223e-01 -1.53191581e-01 -7.60574579e-01
-1.03725481e+00 3.12292010e-01 4.84054118e-01 -4.24299121e-01
-1.46563902e-01 8.39493454e-01 -4.19629246e-01 -5.67141652e-01
6.53503716e-01 -2.70668358e-01 4.23629507e-02 -7.88058862e-02
3.33107233e-01 8.19032013e-01 1.04385838e-01 -4.52482551e-01
-5.76071978e-01 2.62882531e-01 2.02838495e-01 -4.06071424e-01
1.29967856e+00 -3.28537434e-01 5.33979945e-02 6.71864629e-01
5.43164074e-01 -4.19572629e-02 -1.69287300e+00 -1.27057835e-01
2.41850480e-01 -3.52482855e-01 -1.04440652e-01 -1.00087047e+00
-4.68302697e-01 4.64956403e-01 6.50700092e-01 9.31598172e-02
7.54352152e-01 2.84705818e-01 4.44953799e-01 5.07068992e-01
6.67832732e-01 -1.29007590e+00 1.38509154e-01 4.55214977e-01
7.47806370e-01 -1.39078689e+00 1.62731305e-01 -3.52479130e-01
-8.35159659e-01 8.26692343e-01 4.82413620e-01 2.29070172e-01
8.45387101e-01 1.44126073e-01 -1.14273444e-01 -1.49012551e-01
-5.67362905e-01 -1.58348799e-01 7.37566590e-01 4.73550349e-01
1.67532817e-01 7.12183565e-02 8.03899318e-02 8.27276111e-01
-2.89528430e-01 1.01973459e-01 9.98335034e-02 7.18444467e-01
-4.46363837e-01 -1.02641428e+00 -4.55209345e-01 2.56135702e-01
-2.67208844e-01 3.53849292e-01 -3.79082590e-01 3.78481090e-01
1.77681148e-01 9.31545496e-01 -7.15722218e-02 -5.34433186e-01
2.20489994e-01 1.52614012e-01 6.55773699e-01 -4.07138854e-01
-1.99044064e-01 -3.33828002e-01 8.32153410e-02 -5.58574021e-01
-3.55166793e-01 -9.16472673e-01 -8.41613770e-01 -2.05353275e-01
-6.89061508e-02 2.08237395e-01 8.12947989e-01 1.17809403e+00
5.29648662e-01 3.76142740e-01 3.89843047e-01 -8.90811324e-01
-7.95277238e-01 -8.43054533e-01 -4.09485847e-01 4.38375324e-01
-1.82211965e-01 -9.46634531e-01 -3.34019572e-01 -7.82022849e-02] | [7.14342737197876, -0.861072301864624] |
60c5fafa-b996-41b1-9732-44cde740a7d1 | beyond-lexical-a-semantic-retrieval-framework-1 | null | null | https://openreview.net/forum?id=H1xvMrHqAE | https://openreview.net/pdf?id=H1xvMrHqAE | Beyond Lexical: A Semantic Retrieval Framework for Textual Search Engine | Search engine has become a fundamental component in various web and mobile applications. Retrieving relevant documents from the massive datasets is challenging for a search engine system, especially when faced with verbose or tail queries. In this paper, we explore a vector space search framework for document retrieval. Specifically, we trained a deep semantic matching model so that each query and document can be encoded as a low dimensional embedding. Our model was trained based on BERT architecture. We deployed a fast k-nearest-neighbor index service for online serving. Both offline and online metrics demonstrate that our method improved retrieval performance and search quality considerably, particularly for tail queries | ['Anonymous'] | 2019-06-09 | null | null | null | null | ['semantic-retrieval'] | ['natural-language-processing'] | [-2.05720738e-01 -6.40506864e-01 -6.10803902e-01 -2.93957412e-01
-1.12623370e+00 -7.32839227e-01 8.37122977e-01 2.30325237e-02
-5.07676959e-01 6.58356324e-02 4.38854784e-01 -2.11069360e-01
-8.47342551e-01 -8.72195482e-01 -2.82701463e-01 -1.69086277e-01
6.77759945e-03 8.33116829e-01 3.66885900e-01 -4.48640674e-01
4.80413347e-01 4.62246805e-01 -1.50873935e+00 2.53744751e-01
5.08742690e-01 1.57407665e+00 3.95741194e-01 5.79714417e-01
-4.32371706e-01 3.63419592e-01 -2.80450016e-01 -3.86471868e-01
3.06637019e-01 2.08716571e-01 -1.13235152e+00 -6.38811111e-01
2.76149869e-01 -5.65863669e-01 -1.10697627e+00 8.97309840e-01
7.81277239e-01 3.19130927e-01 6.49265528e-01 -1.09417963e+00
-1.17523777e+00 3.47309232e-01 -2.94733755e-02 5.06873786e-01
4.82454807e-01 -3.54855537e-01 1.41124189e+00 -1.20450521e+00
4.01829779e-01 1.09532344e+00 2.44341671e-01 3.20661664e-01
-7.46255457e-01 -4.89500523e-01 -1.66445643e-01 6.81943953e-01
-1.57917631e+00 -4.83029604e-01 6.59643948e-01 -8.01979974e-02
1.16913414e+00 5.34523904e-01 4.19040233e-01 1.09771752e+00
9.11604986e-02 1.19170797e+00 1.78700402e-01 -1.66232914e-01
1.96887910e-01 3.99756245e-02 3.14677477e-01 3.06344450e-01
8.14056993e-02 1.67443991e-01 -4.38310236e-01 -5.02588749e-01
3.93806517e-01 5.77569485e-01 2.29781941e-02 -2.67861784e-01
-9.06504512e-01 1.11020923e+00 7.60598540e-01 4.30225164e-01
-4.89614308e-01 4.06200081e-01 5.07507205e-01 5.08738935e-01
1.82666361e-01 4.59951818e-01 -1.62096128e-01 -1.06283054e-01
-9.57578003e-01 3.02451670e-01 9.45984483e-01 1.14581442e+00
5.55430770e-01 -6.79066062e-01 -6.76268339e-01 1.10333645e+00
4.23436403e-01 7.11062670e-01 9.70495939e-01 -9.78885293e-01
3.61366212e-01 5.84271133e-01 8.76930356e-02 -1.47438252e+00
-1.03528462e-01 -4.80498523e-01 -6.82972133e-01 -7.83200920e-01
-4.51398790e-01 5.56843400e-01 -6.99847877e-01 1.09937918e+00
4.01736684e-02 -1.00135334e-01 -3.55625600e-02 1.15893018e+00
7.80948043e-01 8.22464108e-01 -5.94636537e-02 1.78367376e-01
1.33722806e+00 -1.29578865e+00 -6.37542605e-01 -3.20801198e-01
6.84271991e-01 -9.01948512e-01 1.14955330e+00 -6.05756156e-02
-1.06274581e+00 -3.30854237e-01 -7.34896958e-01 -4.97748643e-01
-6.02306902e-01 -2.98202753e-01 6.72601104e-01 1.45184159e-01
-1.15684271e+00 4.05518323e-01 -5.04913867e-01 -4.01648790e-01
2.87208974e-01 2.35858843e-01 -1.82833105e-01 -4.65346873e-01
-1.48310184e+00 5.08794010e-01 2.31244534e-01 -8.53481740e-02
-5.24478853e-01 -3.64492923e-01 -4.43946153e-01 6.59228742e-01
2.61997432e-01 -7.62304366e-01 1.49020278e+00 -1.12158887e-01
-1.04346824e+00 7.84339011e-01 -2.92669058e-01 -1.91034600e-01
2.52395980e-02 -2.53503621e-01 -7.71946728e-01 2.54255563e-01
2.40084156e-01 2.70796806e-01 6.90317571e-01 -6.82594836e-01
-6.16162956e-01 -6.38446689e-01 -9.16326568e-02 3.25659662e-01
-8.64536107e-01 4.01724517e-01 -1.27996659e+00 -5.74502230e-01
2.63192028e-01 -6.58887088e-01 5.77222072e-02 2.01633215e-01
-1.05944630e-02 -8.51262748e-01 9.07729745e-01 -7.17405498e-01
1.68412602e+00 -2.15845275e+00 -3.54654528e-02 5.36752701e-01
2.95695752e-01 3.26001704e-01 -2.95157731e-01 8.49680543e-01
4.61337864e-01 6.80835545e-02 4.63869750e-01 2.79159229e-02
6.62861824e-01 -1.01379501e-02 -7.28649557e-01 -4.46721539e-02
-3.93223703e-01 1.54973042e+00 -8.20951998e-01 -6.79903328e-01
-1.98925495e-01 5.55354118e-01 -5.02145290e-01 4.19296980e-01
-1.53251708e-01 -5.91781974e-01 -1.02019835e+00 9.29650724e-01
3.03140879e-01 -8.12160552e-01 -3.64805043e-01 -1.40010804e-01
5.55388391e-01 3.04972470e-01 -6.46028697e-01 2.17664194e+00
-7.62842417e-01 6.20318115e-01 -6.92505464e-02 -7.73523092e-01
8.74949694e-01 2.07987890e-01 3.91061604e-01 -1.64941895e+00
-1.95396155e-01 6.23667002e-01 -7.60595679e-01 -4.59460139e-01
8.86441469e-01 6.15732908e-01 -1.84394181e-01 6.62517309e-01
-3.35310668e-01 1.92162707e-01 -4.62862775e-02 4.32805747e-01
1.28872550e+00 -4.72176820e-01 -2.70828784e-01 -3.70071940e-02
4.17235523e-01 -4.12992463e-02 -2.61713892e-01 1.02297115e+00
-3.29742655e-02 2.70474255e-01 -1.66648984e-01 -5.52446187e-01
-8.06173444e-01 -9.21920180e-01 -1.23042107e-01 1.63643825e+00
5.06024837e-01 -5.18181860e-01 -3.59796494e-01 -5.36795497e-01
4.17142719e-01 2.64014721e-01 -9.98135060e-02 -7.57593989e-01
-4.67022896e-01 -1.54480055e-01 3.06308538e-01 5.63963771e-01
3.03700089e-01 -1.11603415e+00 -2.38998845e-01 3.45374525e-01
-1.97127178e-01 -9.26852703e-01 -1.20569587e+00 -9.03326739e-03
-6.56456470e-01 -7.99520493e-01 -9.55806375e-01 -1.20690310e+00
1.23500057e-01 8.32410574e-01 1.09676194e+00 3.78587008e-01
-4.14893627e-01 6.50379419e-01 -5.75658441e-01 1.29893720e-01
2.21755922e-01 4.15906727e-01 -7.62849152e-02 -3.10971946e-01
1.11242604e+00 -1.74881220e-01 -1.45506442e+00 5.23137987e-01
-1.23951852e+00 -7.45220959e-01 4.96235818e-01 7.98887789e-01
5.65070808e-01 -1.09459214e-01 4.52700734e-01 -1.90833986e-01
1.33603859e+00 -7.92319775e-01 -5.63670635e-01 5.25783420e-01
-1.18680286e+00 2.94107229e-01 4.54368591e-01 -1.53725937e-01
-3.45963776e-01 -5.18572628e-01 -8.04401264e-02 -5.10528445e-01
3.98737252e-01 7.03799903e-01 1.82001755e-01 -7.32062161e-02
6.78430438e-01 5.98152637e-01 -2.40550086e-01 -9.17721570e-01
3.67233068e-01 1.52693152e+00 3.68321449e-01 -4.11726147e-01
6.49206102e-01 2.71598101e-01 -3.43872607e-01 -4.18558449e-01
-7.14823842e-01 -1.18410861e+00 -1.22749664e-01 3.25938135e-01
3.66867959e-01 -8.96624982e-01 -7.47331023e-01 -2.84164120e-02
-1.10973108e+00 1.40719935e-01 -5.16321994e-02 3.45686078e-01
-4.54377860e-01 3.36865962e-01 -4.91187662e-01 -4.06984448e-01
-9.89258349e-01 -1.04837704e+00 1.40323019e+00 1.10019170e-01
3.04264249e-03 -1.04132891e+00 3.46922845e-01 5.09080529e-01
1.00957751e+00 -9.28771436e-01 8.44342887e-01 -1.05062401e+00
-6.59116745e-01 -8.43699813e-01 -9.02382791e-01 -1.13952376e-01
7.44421333e-02 -7.00185537e-01 -7.83133566e-01 -5.67302883e-01
-2.71860927e-01 -5.09350598e-01 9.24489915e-01 -1.41571715e-01
1.74914765e+00 -5.42312384e-01 -6.78513527e-01 8.72727156e-01
1.37425959e+00 1.69807419e-01 3.17355275e-01 6.61721051e-01
4.38663691e-01 3.17717463e-01 6.14354193e-01 3.92009258e-01
4.07013237e-01 1.12950516e+00 2.44612128e-01 3.08931559e-01
-1.13517478e-01 -6.39077663e-01 -1.83714718e-01 7.38846898e-01
6.67609036e-01 -5.28187513e-01 -9.74201620e-01 6.65915132e-01
-1.98227561e+00 -8.86063993e-01 6.56122446e-01 1.96090209e+00
8.85700345e-01 -2.45531306e-01 -1.68959469e-01 -1.50494322e-01
4.03535157e-01 2.06974238e-01 -8.23435724e-01 -1.50634632e-01
2.00988483e-02 4.00156915e-01 3.34147274e-01 3.51929575e-01
-8.55684221e-01 9.95971620e-01 6.65468693e+00 1.30718803e+00
-8.92500341e-01 -9.28814560e-02 1.10190295e-01 -2.91219980e-01
-6.34107590e-01 -1.57656372e-01 -8.94379735e-01 7.29948223e-01
9.37489986e-01 -7.03489304e-01 6.41671598e-01 1.08098817e+00
-1.92627952e-01 5.77478766e-01 -1.14010918e+00 1.47643185e+00
1.37024119e-01 -1.58513081e+00 4.78042573e-01 1.96234405e-01
3.89387071e-01 2.13173583e-01 3.43843281e-01 5.83545089e-01
1.47930101e-01 -1.26337159e+00 1.82233796e-01 5.97437978e-01
6.94579840e-01 -6.15999162e-01 4.95336235e-01 2.40698025e-01
-1.05469787e+00 -3.38975221e-01 -5.54684460e-01 4.76157457e-01
-7.09037110e-02 2.78127491e-01 -5.24734855e-01 2.09436089e-01
9.15559471e-01 5.25277019e-01 -5.38010001e-01 1.21687496e+00
5.21382034e-01 1.63843602e-01 -4.59391892e-01 -3.84354234e-01
4.83730912e-01 -6.08192608e-02 2.30126917e-01 1.14271677e+00
6.23109639e-01 -1.76474765e-01 4.18912061e-02 7.04196274e-01
-4.30113375e-01 3.04229140e-01 -7.29311526e-01 -4.44195241e-01
8.97529542e-01 1.13804638e+00 -1.69939652e-01 -2.31936857e-01
-3.85657221e-01 1.50793469e+00 3.74909401e-01 5.60137510e-01
-4.43221807e-01 -9.66789067e-01 7.91475832e-01 -2.23063696e-02
3.21368873e-01 3.49958055e-02 3.79923105e-01 -1.11978877e+00
5.98465800e-01 -8.28383029e-01 7.39484966e-01 -7.87184715e-01
-1.42931795e+00 6.43344223e-01 -2.80114323e-01 -9.17716861e-01
-6.77181184e-01 -3.64661247e-01 -2.39879757e-01 1.04670608e+00
-1.81258881e+00 -9.08867061e-01 -4.00033861e-01 8.59890163e-01
4.73221630e-01 -5.93455791e-01 1.03939605e+00 6.68572247e-01
-1.89637125e-01 9.28519845e-01 8.63513529e-01 1.16060846e-01
5.92931092e-01 -8.00066650e-01 5.81284404e-01 1.30079597e-01
5.78623176e-01 9.25147712e-01 5.50487898e-02 -2.73625493e-01
-1.90437543e+00 -8.50407243e-01 1.20596540e+00 -5.25654316e-01
9.17909205e-01 -1.58939794e-01 -9.10480678e-01 3.56012374e-01
4.06446727e-03 1.23687178e-01 7.41877079e-01 2.39779279e-02
-5.91247976e-01 -3.40946764e-01 -8.28226924e-01 6.28046513e-01
1.09635329e+00 -1.16539264e+00 -5.05616069e-01 8.54190707e-01
1.30056059e+00 3.19402665e-02 -9.28905666e-01 1.25479817e-01
7.19095945e-01 -5.00231624e-01 1.52273858e+00 -8.65245700e-01
1.36777596e-03 5.52813709e-02 -6.39074266e-01 -9.30546403e-01
-4.61075425e-01 -5.30029178e-01 -8.69148672e-01 6.31539762e-01
2.90817976e-01 -4.28122699e-01 9.09057856e-01 8.48061144e-01
2.13052183e-01 -7.64662266e-01 -7.50229478e-01 -9.99677718e-01
-8.16791225e-03 -2.81526893e-01 9.95167851e-01 4.77237344e-01
-1.54449465e-02 2.58155942e-01 -7.92642906e-02 -1.27897382e-01
4.78840381e-01 6.44033074e-01 3.43234122e-01 -1.24088168e+00
-2.16128662e-01 -7.41998255e-01 -4.33427066e-01 -1.77610385e+00
2.58838207e-01 -1.37041008e+00 -4.53959316e-01 -1.58952618e+00
5.33308327e-01 -5.35655379e-01 -7.20001400e-01 1.75734639e-01
3.56097030e-03 1.63922235e-01 -1.67894393e-01 7.89451540e-01
-1.11089790e+00 7.31845677e-01 8.92038047e-01 -5.00707865e-01
1.05751671e-01 6.73207864e-02 -8.64505529e-01 -6.60651103e-02
4.50897962e-01 -3.58689576e-01 -5.25231481e-01 -1.15413380e+00
7.22154737e-01 7.97828510e-02 4.21646923e-01 -4.93675232e-01
9.62250054e-01 3.16972355e-03 2.04808936e-01 -5.72021484e-01
4.39275533e-01 -1.09272540e+00 -2.36887455e-01 2.46925935e-01
-7.50135899e-01 7.77785704e-02 -1.64747417e-01 7.56693006e-01
-6.43049657e-01 -2.80310482e-01 1.30776176e-02 2.08292693e-01
-8.93174231e-01 8.06882679e-01 2.37645343e-01 1.75369114e-01
6.43548250e-01 -5.14236130e-02 -2.52195626e-01 -6.37913406e-01
-3.82167101e-01 6.51444852e-01 1.94589362e-01 8.08568239e-01
9.70272183e-01 -1.85385597e+00 -4.04289544e-01 2.71800607e-01
6.99360013e-01 -2.17185736e-01 7.77359772e-03 1.62353456e-01
-4.34376240e-01 1.27278328e+00 3.18288743e-01 -2.88268715e-01
-7.95444548e-01 9.31183100e-01 1.43929804e-02 -1.33798748e-01
-4.46748495e-01 8.70800436e-01 -1.80402070e-01 -2.98640072e-01
7.38980353e-01 1.52273595e-01 -1.28840581e-01 -3.82214412e-02
8.38342190e-01 3.63143444e-01 2.04193354e-01 -2.39762202e-01
-3.09266210e-01 7.09620118e-01 -4.21283692e-01 2.60778759e-02
1.11456740e+00 -5.00938296e-01 -9.81889740e-02 -6.88290745e-02
2.12187529e+00 -4.75877613e-01 -3.93613160e-01 -8.50486815e-01
3.92902166e-01 -6.93159997e-01 6.78474307e-01 -6.27186656e-01
-9.90136325e-01 6.44345343e-01 7.48707771e-01 3.05335343e-01
9.89780366e-01 3.09573174e-01 1.62620521e+00 1.14923227e+00
4.15625393e-01 -1.32892478e+00 1.31590113e-01 5.37151873e-01
8.58913183e-01 -1.54942751e+00 -4.10806715e-01 3.73462886e-01
-1.23383969e-01 9.55883920e-01 8.96858796e-02 8.93884748e-02
9.11579728e-01 -3.71404409e-01 8.30866694e-02 -6.87035620e-01
-9.33555126e-01 -1.81302696e-01 8.73248875e-01 2.66577572e-01
2.58007556e-01 -2.77985305e-01 -1.38300762e-01 5.06999075e-01
-8.61877725e-02 4.66613173e-02 -5.92044890e-01 1.04000270e+00
-6.35015130e-01 -1.11383724e+00 9.81572792e-02 6.97904289e-01
-3.15169334e-01 -4.52194393e-01 -3.34428817e-01 1.09336413e-01
-9.68740106e-01 9.63595569e-01 2.84832954e-01 -5.48049808e-01
3.02098989e-01 8.10057670e-02 8.33450723e-03 -3.95803034e-01
-2.64186949e-01 -2.31674209e-01 -3.84879947e-01 -1.16989148e+00
2.88390607e-01 -4.55995910e-02 -9.93150234e-01 -3.46602827e-01
-3.01472306e-01 5.36616623e-01 9.28826034e-01 5.07343709e-01
8.80396187e-01 -1.17599472e-01 1.01199377e+00 -2.59001911e-01
-1.21670938e+00 -6.39381945e-01 -6.99116588e-01 5.56996346e-01
2.24432960e-01 -3.09314519e-01 -4.19240981e-01 -5.10010660e-01] | [11.356941223144531, 7.467426776885986] |
ccad82a1-daa8-4958-bba0-68b3eed741ae | invero-making-semantic-role-labeling | null | null | https://aclanthology.org/2020.emnlp-demos.11 | https://aclanthology.org/2020.emnlp-demos.11.pdf | InVeRo: Making Semantic Role Labeling Accessible with Intelligible Verbs and Roles | Semantic Role Labeling (SRL) is deeply dependent on complex linguistic resources and sophisticated neural models, which makes the task difficult to approach for non-experts. To address this issue we present a new platform named Intelligible Verbs and Roles (InVeRo). This platform provides access to a new verb resource, VerbAtlas, and a state-of-the-art pretrained implementation of a neural, span-based architecture for SRL. Both the resource and the system provide human-readable verb sense and semantic role information, with an easy to use Web interface and RESTful APIs available at http://nlp.uniroma1.it/invero. | ['Roberto Navigli', 'Davide Zanfardino', 'Fabrizio Brignone', 'Simone Conia'] | 2020-10-01 | null | null | null | emnlp-2020-11 | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 1.55627295e-01 4.30795550e-01 -4.93178487e-01 -4.17098552e-01
-4.36986595e-01 -1.03398681e+00 5.87618768e-01 4.02484417e-01
-6.38301969e-01 8.64910722e-01 7.06473053e-01 -1.66133001e-01
-1.91627413e-01 -7.02519476e-01 -3.36314768e-01 -2.63855457e-01
1.15316279e-01 8.53235781e-01 2.45255679e-01 -8.34128201e-01
-1.52220577e-01 3.49710733e-01 -1.57146025e+00 7.26002574e-01
6.61185801e-01 8.40587676e-01 6.47970021e-01 2.70125836e-01
-3.80595237e-01 1.16973829e+00 -3.78196836e-01 -1.91125423e-01
-1.46675184e-01 -1.14098921e-01 -1.30258870e+00 -7.33606517e-01
-1.56703204e-01 7.73265436e-02 -7.23059922e-02 9.38574970e-01
3.42443347e-01 4.74316359e-01 4.45917100e-02 -1.21425557e+00
-8.71647239e-01 1.20063388e+00 4.55130309e-01 2.65185714e-01
8.77700269e-01 3.32352668e-02 1.43085611e+00 -7.20150471e-01
1.11463332e+00 1.46091914e+00 5.55348814e-01 1.07342482e+00
-9.22959626e-01 -2.49392048e-01 3.00294273e-02 5.85069954e-01
-7.26758361e-01 -5.92266619e-01 9.01261687e-01 -2.20439151e-01
1.37525785e+00 2.94755727e-01 7.69969344e-01 1.49283755e+00
-3.19671631e-01 7.37767816e-01 1.05846155e+00 -6.12401187e-01
9.23454240e-02 -1.81770101e-01 6.13278151e-01 7.18258739e-01
3.13613534e-01 -1.65177628e-01 -8.61991763e-01 -5.05311675e-02
7.59219289e-01 -5.29017806e-01 -2.56753892e-01 -1.55146196e-01
-1.27932358e+00 5.59383333e-01 4.44530994e-01 9.15933311e-01
-5.11974871e-01 3.44239771e-01 6.87127352e-01 4.64201927e-01
3.27372134e-01 7.19490826e-01 -8.21375430e-01 -4.24966633e-01
-1.71974584e-01 2.22528204e-01 9.29909885e-01 5.60077965e-01
1.85524240e-01 -1.57364875e-01 9.08221230e-02 9.93393958e-01
4.64678615e-01 -1.10146411e-01 7.64826238e-01 -1.13977206e+00
3.68814558e-01 9.47358489e-01 2.72972107e-01 -5.07087171e-01
-8.66371989e-01 9.52600390e-02 -3.69953424e-01 -1.25089571e-01
7.27360964e-01 3.50998491e-01 -6.01512730e-01 1.99455190e+00
3.68848056e-01 -4.90683734e-01 2.46443897e-01 9.05575573e-01
1.27928388e+00 4.19855148e-01 6.85062706e-01 -6.62326068e-02
1.75874805e+00 -7.46154249e-01 -1.05271769e+00 -6.77083254e-01
6.66904807e-01 -3.71544480e-01 1.77916574e+00 7.89295658e-02
-1.18284869e+00 -3.55417043e-01 -9.78498161e-01 -6.96294665e-01
-9.16825712e-01 -4.33974154e-03 1.05051303e+00 2.02059567e-01
-1.19002986e+00 5.33279538e-01 -8.03937197e-01 -8.17792594e-01
4.84488070e-01 1.96483046e-01 -7.36117601e-01 4.42594960e-02
-1.93458533e+00 1.30700123e+00 1.07485485e+00 3.06487195e-02
-6.18524849e-01 -5.51337421e-01 -1.23421514e+00 1.13017209e-01
7.07032382e-01 -6.59521997e-01 1.46368480e+00 -8.10651839e-01
-1.38202536e+00 1.47071755e+00 -1.93777248e-01 -4.50033575e-01
2.90310800e-01 -2.25292072e-01 -2.82210737e-01 3.40100288e-01
3.64007503e-01 5.42841911e-01 4.15614516e-01 -9.92795646e-01
-3.32111299e-01 -5.32267690e-01 5.96601963e-01 4.77484822e-01
-2.03439325e-01 5.66627860e-01 -1.30571708e-01 -8.80841672e-01
-6.69557080e-02 -4.72870052e-01 -2.62910463e-02 8.45067874e-02
1.19378075e-01 -9.97684836e-01 3.11975718e-01 -7.96055675e-01
1.10076940e+00 -2.07914925e+00 1.56130582e-01 -4.29880828e-01
3.59274715e-01 1.90287441e-01 -1.59770772e-01 4.04295444e-01
-3.70346844e-01 2.67521173e-01 -3.16943109e-01 -4.40836996e-02
3.04880142e-01 5.99667966e-01 6.50590509e-02 2.33894825e-01
1.17313322e-02 9.90281820e-01 -1.43772137e+00 -4.91239727e-01
4.26096171e-01 1.90419808e-01 -2.25373119e-01 1.64316922e-01
-6.04313254e-01 4.81221974e-01 -3.40771943e-01 5.72434902e-01
2.30444640e-01 -2.74916530e-01 6.90645993e-01 4.71095964e-02
-7.54317790e-02 1.00714993e+00 -1.04703152e+00 2.03675747e+00
-7.59489894e-01 3.10727328e-01 1.31003559e-01 -8.54779661e-01
6.12984240e-01 7.00240254e-01 3.01165581e-01 -4.35120225e-01
2.03161970e-01 3.96852374e-01 -1.88605785e-02 -5.14544487e-01
2.54647523e-01 -1.88549340e-01 -3.18625808e-01 3.72293949e-01
3.75992805e-01 -4.89248261e-02 6.18884206e-01 -5.86649179e-02
1.04999757e+00 6.71188951e-01 1.03718853e+00 -5.31616628e-01
7.36081302e-01 1.78461179e-01 5.86269557e-01 4.47704881e-01
-2.62613028e-01 7.86113087e-03 6.38826728e-01 -6.83472991e-01
-7.15027153e-01 -1.02550781e+00 -4.84701991e-02 1.69697750e+00
7.26973414e-02 -4.99881417e-01 -7.55220234e-01 -6.47985518e-01
-2.31064677e-01 1.06607437e+00 -6.62389994e-01 6.69454485e-02
-8.05100918e-01 -4.14605916e-01 7.39396811e-01 6.41010642e-01
5.46782196e-01 -1.98264539e+00 -8.94148886e-01 4.13598448e-01
-5.74468791e-01 -1.29058695e+00 -1.53420074e-02 4.37616050e-01
-5.23472190e-01 -1.24157691e+00 -1.39351664e-02 -8.77087772e-01
1.80539206e-01 -3.97989839e-01 1.55958021e+00 5.69995791e-02
-7.54043693e-03 6.62699994e-03 -5.81597149e-01 -3.24455321e-01
-5.62052548e-01 1.20152332e-01 -2.47165989e-02 -3.89569551e-01
4.34123814e-01 -6.05136812e-01 -3.51869278e-02 -1.77654937e-01
-7.71125197e-01 2.50693783e-02 -3.07300258e-02 6.75933659e-01
5.97700179e-01 -2.21437752e-01 5.95938623e-01 -1.02809525e+00
9.80507672e-01 -5.01246989e-01 -6.75604761e-01 2.29434580e-01
-3.83541048e-01 1.28852665e-01 7.89171994e-01 2.95578577e-02
-1.34677744e+00 5.97711504e-02 -6.70197070e-01 1.85140133e-01
-4.33229476e-01 5.53556442e-01 -7.56448925e-01 4.17852968e-01
7.60342717e-01 -4.40451130e-02 -2.52551377e-01 -7.65676320e-01
6.91608250e-01 5.73017836e-01 6.27083063e-01 -7.29202509e-01
2.19941676e-01 4.20019299e-01 -3.46061081e-01 -6.26303792e-01
-1.18291545e+00 -3.90404135e-01 -6.81548715e-01 -1.08239837e-01
1.08918214e+00 -7.00469851e-01 -6.77219868e-01 3.97532076e-01
-1.49532592e+00 -7.67448783e-01 -6.56893253e-01 5.61486818e-02
-7.99402118e-01 5.64229600e-02 -6.31542504e-01 -3.17841947e-01
-5.23905516e-01 -6.06224716e-01 7.53038526e-01 8.84280577e-02
-7.96047926e-01 -1.23292279e+00 -3.04605812e-01 6.11445963e-01
2.86627948e-01 3.85676503e-01 1.06964028e+00 -1.17543888e+00
-1.06002353e-02 2.21987851e-02 -2.24888489e-01 2.79594004e-01
1.58948787e-02 -7.46772051e-01 -8.02509904e-01 2.81716436e-01
6.44852743e-02 -6.08899295e-01 4.83179778e-01 7.06399158e-02
1.18467724e+00 -3.09912533e-01 -1.88160166e-01 2.69849122e-01
1.20912743e+00 8.45051408e-02 4.69690353e-01 5.68857312e-01
6.44444883e-01 6.81642294e-01 7.08191454e-01 2.55079091e-01
6.22156799e-01 7.90085614e-01 3.95823210e-01 4.05665517e-01
-3.06636631e-01 -3.39304149e-01 3.10702294e-01 3.49305928e-01
-2.46295482e-01 -3.02602887e-01 -1.19625831e+00 6.55467033e-01
-2.07367921e+00 -9.44322646e-01 -2.78498083e-01 1.67383289e+00
1.10823393e+00 1.28826752e-01 -9.79253352e-02 2.10116357e-01
6.00852311e-01 4.98711288e-01 -4.16767120e-01 -5.09988248e-01
-1.73366323e-01 4.32208300e-01 -3.28170992e-02 7.86058545e-01
-1.11679471e+00 1.63241041e+00 5.83545828e+00 1.01358604e+00
-6.25404656e-01 8.81244600e-01 3.86275426e-02 1.23076007e-01
-3.50494146e-01 -1.83958307e-01 -6.32198155e-01 3.42267960e-01
1.09877348e+00 -3.97773147e-01 7.51116633e-01 9.62199330e-01
4.57772106e-01 -1.73679739e-02 -1.09130132e+00 1.00141883e+00
-7.98856243e-02 -1.51286471e+00 8.05912837e-02 -4.28808182e-01
-1.55818388e-01 2.09173322e-01 -6.21431410e-01 3.06696862e-01
4.87685382e-01 -1.19317806e+00 1.12196481e+00 1.56096295e-01
1.06810308e+00 -2.98427939e-01 7.18220830e-01 2.31679514e-01
-1.32685876e+00 -2.51371235e-01 -1.14457034e-01 -4.30136293e-01
2.38802537e-01 3.30773830e-01 -3.28641713e-01 3.43021840e-01
8.01755130e-01 7.81472743e-01 -4.21983033e-01 3.30020189e-01
-1.17912996e+00 5.20553350e-01 -3.96632731e-01 -2.75212377e-01
1.48350805e-01 2.08727360e-01 7.16193140e-01 9.77982342e-01
-2.69633859e-01 1.52950764e-01 3.02499712e-01 7.39712119e-01
-4.09270585e-01 2.37299800e-01 -5.67197978e-01 -3.61564368e-01
7.55201042e-01 1.16483760e+00 -8.86685312e-01 -2.47370899e-01
-1.10908851e-01 1.00232351e+00 5.96161664e-01 -1.97956413e-01
-6.47164524e-01 -3.78310263e-01 7.71579146e-01 2.31311873e-01
-4.83848631e-01 -1.72714755e-01 -4.20648187e-01 -1.15597939e+00
5.41512594e-02 -8.39790761e-01 9.24669921e-01 -1.07349873e+00
-1.13966310e+00 7.11605191e-01 2.74332881e-01 -5.55740476e-01
-4.37262326e-01 -1.00509262e+00 -1.70052126e-01 7.51324773e-01
-1.37560952e+00 -1.45989823e+00 -2.34839782e-01 6.66479170e-01
8.01282823e-01 8.95496309e-02 1.21451068e+00 9.16491374e-02
-1.26641005e-01 1.44772813e-01 -5.85762560e-01 1.67751119e-01
2.35650629e-01 -1.58616638e+00 6.16663516e-01 5.80573678e-01
9.27771553e-02 6.48374557e-01 5.97298801e-01 -4.99457210e-01
-9.86546636e-01 -7.62213707e-01 1.56564569e+00 -8.30763996e-01
1.07174170e+00 -4.99651700e-01 -8.36274266e-01 1.01864541e+00
3.06053936e-01 2.06821874e-01 7.12030232e-01 1.15805961e-01
-3.32655370e-01 3.28842461e-01 -1.33591390e+00 5.15954971e-01
1.78262508e+00 -6.48468256e-01 -1.50719082e+00 6.50512695e-01
9.30873513e-01 -4.46194232e-01 -7.55202830e-01 4.53683324e-02
2.69260615e-01 -7.27173269e-01 8.59774828e-01 -6.58124506e-01
3.05687517e-01 -2.84653306e-01 -1.62675723e-01 -1.19307005e+00
-2.32718542e-01 -7.42221773e-01 -3.59959193e-02 1.00946200e+00
5.83264172e-01 -8.93418193e-01 3.53004664e-01 4.27065700e-01
-4.56580251e-01 -4.79214281e-01 -1.24206591e+00 -7.52061546e-01
-1.11382529e-01 -7.47879267e-01 6.54106319e-01 1.05062628e+00
4.00977820e-01 8.43633235e-01 1.04880139e-01 -1.01206921e-01
2.38503277e-01 -2.55766381e-02 -7.59121925e-02 -1.61458564e+00
-2.17033938e-01 -3.91825289e-01 5.93532771e-02 -4.93956298e-01
8.19616914e-01 -1.46188867e+00 -1.10496685e-01 -2.18982482e+00
-3.24644655e-01 -3.10784817e-01 -9.22774822e-02 1.36227453e+00
-4.71992344e-02 2.09332451e-01 2.14911968e-01 1.65583923e-01
-7.35952377e-01 4.45056975e-01 9.62350130e-01 2.32374489e-01
-1.76919803e-01 -4.96402353e-01 -8.73235524e-01 1.27915418e+00
1.25467050e+00 -7.05609977e-01 -2.95803249e-01 -6.59140766e-01
6.66677952e-01 -2.04141080e-01 3.99131536e-01 -8.30122411e-01
-2.01137900e-01 -1.39855310e-01 6.01035841e-02 -2.23890319e-02
4.17554498e-01 -6.55664086e-01 -1.47498310e-01 5.06714880e-01
-4.73459810e-01 3.69514003e-02 2.74463147e-01 3.76641490e-02
-1.79094225e-01 -7.57483304e-01 7.43856788e-01 -7.94115067e-01
-1.18379879e+00 -2.42826238e-01 -4.35163558e-01 4.09839064e-01
8.41527939e-01 5.44126518e-02 -4.04313087e-01 -7.71737620e-02
-1.13066959e+00 2.28858694e-01 3.76633674e-01 6.67640567e-01
4.03897047e-01 -1.21201181e+00 -5.80175519e-01 -1.61809012e-01
4.23858464e-01 -1.29518703e-01 7.57815242e-02 2.18726888e-01
-7.83793271e-01 4.72712189e-01 -3.24041545e-01 3.09736639e-01
-1.07717025e+00 3.88161600e-01 4.99211580e-01 -7.14847624e-01
-5.17449558e-01 9.16823804e-01 -3.72811645e-01 -9.25686598e-01
-1.69674084e-01 -2.51542062e-01 -8.15126777e-01 4.14381057e-01
5.70316255e-01 9.39087868e-02 1.41812861e-02 -5.84815979e-01
-6.38512313e-01 2.94265505e-02 4.54832166e-01 -1.10196039e-01
1.46392846e+00 1.32983923e-01 -6.80775285e-01 7.00483322e-01
5.74893773e-01 -6.08560853e-02 -5.79519987e-01 1.64559960e-01
5.73697507e-01 -6.10078052e-02 -1.10353254e-01 -1.15609848e+00
-4.84028131e-01 5.32210946e-01 2.32794687e-01 4.23890024e-01
8.11981738e-01 5.15006840e-01 7.10090280e-01 5.65668106e-01
5.56384563e-01 -1.12286973e+00 -1.69132397e-01 9.97715592e-01
1.35059071e+00 -9.51268613e-01 -4.25534278e-01 -8.23881030e-01
-5.25568187e-01 8.80764484e-01 5.52306473e-01 -1.68845523e-02
5.46380520e-01 7.70304203e-02 1.78102106e-01 -3.45635235e-01
-8.66472244e-01 -6.28860235e-01 -8.07616711e-02 8.83256912e-01
7.09835947e-01 1.91070586e-01 -1.02596736e+00 1.14279819e+00
-5.27870595e-01 2.88719237e-01 3.83806646e-01 8.53390217e-01
-1.61233425e-01 -1.36160970e+00 -2.36885130e-01 2.09486395e-01
-6.98456764e-01 -4.82597142e-01 -4.81248111e-01 8.21126640e-01
2.80220300e-01 9.58227396e-01 4.85551124e-03 2.57024407e-01
4.76211607e-01 6.26051545e-01 4.53033626e-01 -1.17769241e+00
-8.58220637e-01 -4.20727551e-01 8.60685050e-01 -9.59748983e-01
-4.75865006e-01 -5.72072983e-01 -1.96198928e+00 1.34864807e-01
4.44596291e-01 7.93279037e-02 4.96350825e-01 9.20875132e-01
3.04317296e-01 5.61042368e-01 -2.94431448e-01 -7.70908415e-01
-3.16006154e-01 -1.07688820e+00 -2.85586357e-01 6.31111562e-01
-3.18969786e-01 -7.72881031e-01 -1.42057940e-01 3.03875539e-03] | [10.275971412658691, 9.42556381225586] |
19b62514-dd6a-4b63-8f0f-4f66de71d50b | improving-statistical-fidelity-for-neural | 2301.11189 | null | https://arxiv.org/abs/2301.11189v2 | https://arxiv.org/pdf/2301.11189v2.pdf | Improving Statistical Fidelity for Neural Image Compression with Implicit Local Likelihood Models | Lossy image compression aims to represent images in as few bits as possible while maintaining fidelity to the original. Theoretical results indicate that optimizing distortion metrics such as PSNR or MS-SSIM necessarily leads to a discrepancy in the statistics of original images from those of reconstructions, in particular at low bitrates, often manifested by the blurring of the compressed images. Previous work has leveraged adversarial discriminators to improve statistical fidelity. Yet these binary discriminators adopted from generative modeling tasks may not be ideal for image compression. In this paper, we introduce a non-binary discriminator that is conditioned on quantized local image representations obtained via VQ-VAE autoencoders. Our evaluations on the CLIC2020, DIV2K and Kodak datasets show that our discriminator is more effective for jointly optimizing distortion (e.g., PSNR) and statistical fidelity (e.g., FID) than the state-of-the-art HiFiC model. On the CLIC2020 test set, we obtain the same FID as HiFiC with 30-40% fewer bits. | ['Jakob Verbeek', 'Hervé Jégou', 'Karen Ullrich', 'Alaaeldin El-Nouby', 'Matthew J. Muckley'] | 2023-01-26 | null | null | null | null | ['ms-ssim'] | ['computer-vision'] | [ 3.81505787e-01 -1.94034472e-01 8.51461291e-02 -3.31121653e-01
-9.41189170e-01 -4.97303486e-01 4.42931533e-01 -3.44890982e-01
-1.89973757e-01 7.54446447e-01 4.35153186e-01 -2.11408928e-01
9.65105668e-02 -7.61051118e-01 -9.53438997e-01 -7.44281590e-01
-1.50381342e-01 -1.27267197e-01 -3.58520359e-01 -2.32741330e-02
7.99919218e-02 3.23776811e-01 -1.26850939e+00 4.23287898e-01
8.04112315e-01 1.20150411e+00 2.08319381e-01 9.84362185e-01
5.05160213e-01 1.06938136e+00 -8.93942714e-01 -6.73730493e-01
5.51966906e-01 -7.29448199e-01 -5.43920398e-01 5.72612733e-02
7.24006772e-01 -1.04824579e+00 -1.10321689e+00 1.40468955e+00
4.54694510e-01 -1.62523091e-01 7.86395967e-01 -1.06456923e+00
-1.13491225e+00 4.05728638e-01 -1.64757341e-01 3.49897951e-01
8.31098333e-02 4.87012446e-01 9.18773830e-01 -5.06724834e-01
6.59599364e-01 1.34550834e+00 6.05506122e-01 3.15518767e-01
-1.64939523e+00 -5.45896530e-01 -5.96350253e-01 2.96034664e-01
-1.71179390e+00 -7.47700572e-01 5.63137233e-01 -2.44426921e-01
6.81834579e-01 4.68948424e-01 2.74827957e-01 1.17639101e+00
7.47666240e-01 4.92476344e-01 9.73201811e-01 -1.40884325e-01
3.21048111e-01 -7.34284800e-03 -7.35588789e-01 3.04019600e-01
3.05537462e-01 3.65036219e-01 -4.73927796e-01 5.33030666e-02
9.68034089e-01 -1.34009227e-01 -6.33446276e-01 -1.07313938e-01
-9.47861433e-01 8.51370394e-01 5.57452500e-01 1.44278944e-01
-5.28772593e-01 7.32470632e-01 9.48821902e-02 4.78184223e-01
3.58756602e-01 4.37870860e-01 7.90222064e-02 -2.71034241e-01
-1.34299850e+00 2.73765862e-01 4.76535290e-01 9.97699440e-01
6.32599175e-01 3.74863625e-01 -3.23641032e-01 7.58533955e-01
1.86398402e-01 6.45314991e-01 5.47730982e-01 -1.62194216e+00
4.81390566e-01 -2.42616892e-01 1.28079891e-01 -1.34860063e+00
4.97946829e-01 -5.96162558e-01 -1.00505006e+00 3.87138307e-01
-3.37430695e-03 1.83457389e-01 -1.03577423e+00 1.82763970e+00
-5.02610445e-01 1.23412006e-01 2.38717988e-01 1.03743207e+00
4.55405533e-01 9.71315563e-01 -1.39425620e-01 -1.50467142e-01
7.54219890e-01 -6.33937418e-01 -8.92190516e-01 -2.62670010e-01
1.73498407e-01 -8.69051754e-01 9.31160510e-01 4.22791153e-01
-1.45922601e+00 -6.88509166e-01 -1.39290082e+00 -1.42710358e-01
1.71634167e-01 -1.23075992e-01 1.44069687e-01 8.19126785e-01
-1.20209253e+00 8.55353832e-01 -8.62793982e-01 1.94147393e-01
5.96664310e-01 -5.23437560e-02 -2.42366463e-01 -4.50754464e-01
-1.12445307e+00 6.67003810e-01 2.10212186e-01 -3.82723868e-01
-1.48582792e+00 -7.97193706e-01 -8.49848270e-01 3.06713939e-01
-2.24904284e-01 -5.84443510e-01 1.09883082e+00 -9.28476334e-01
-1.09385705e+00 5.77092886e-01 1.18789256e-01 -9.02800798e-01
6.35271966e-01 -1.77724168e-01 -3.74619097e-01 5.60201645e-01
-1.17800817e-01 9.69287872e-01 1.06222332e+00 -1.51156616e+00
-1.69765607e-01 -2.39202883e-02 3.67148630e-02 1.13171004e-01
-3.61739844e-01 -4.02580470e-01 -3.58217001e-01 -1.01436365e+00
8.78989026e-02 -7.91567147e-01 3.30642462e-02 3.01309705e-01
-3.93997908e-01 6.35617197e-01 8.70505095e-01 -1.12718832e+00
1.16854692e+00 -2.31954384e+00 5.53342849e-02 -7.29774013e-02
2.75632918e-01 2.17310861e-01 -2.67190099e-01 3.20717007e-01
-2.50860769e-02 4.48239088e-01 -4.14760351e-01 -4.56115156e-01
-7.25376084e-02 4.88199145e-01 -3.76483023e-01 7.79472053e-01
2.07822770e-01 9.12785470e-01 -7.08554149e-01 -4.01193917e-01
2.82130569e-01 9.06447113e-01 -8.90882671e-01 3.09473664e-01
-1.41974427e-02 3.12720656e-01 5.04416041e-02 6.19748831e-01
9.38745856e-01 -3.05936694e-01 -1.28659964e-01 -3.64297450e-01
4.26630914e-01 2.04649583e-01 -7.47032106e-01 1.65643656e+00
-5.64455330e-01 1.08644092e+00 1.91378372e-03 -7.36174405e-01
7.03254402e-01 2.99430937e-01 3.30453902e-01 -9.26389277e-01
1.03580328e-02 1.13062009e-01 -9.87104848e-02 -2.45064646e-01
6.79771185e-01 -1.68748364e-01 1.87381759e-01 1.67049959e-01
3.06248695e-01 -4.36509043e-01 -2.95669697e-02 2.51145273e-01
1.12212014e+00 -2.92909920e-01 5.61212040e-02 4.94439490e-02
1.53688630e-02 -3.85236233e-01 4.20728385e-01 8.66426587e-01
-3.00958008e-01 1.20024168e+00 3.15544099e-01 4.84064408e-02
-1.85427356e+00 -1.24961984e+00 -3.18333149e-01 2.38072515e-01
2.60618776e-01 -4.61552262e-01 -9.27846313e-01 -1.47219241e-01
-9.40413922e-02 1.06676626e+00 -3.90105933e-01 -5.79596877e-01
-3.02169651e-01 -3.92707735e-01 9.11304891e-01 2.94652820e-01
7.17813373e-01 -6.15197837e-01 -5.54388940e-01 9.92358774e-02
-3.70541960e-01 -1.35476506e+00 -6.47266507e-01 -6.88156635e-02
-9.36399996e-01 -5.46249330e-01 -9.39128876e-01 -2.77039051e-01
5.74913323e-01 1.80895641e-01 1.25957811e+00 -1.52765319e-01
-2.58105785e-01 2.55393565e-01 -3.74851346e-01 -3.76808969e-03
-9.73161399e-01 -4.44443107e-01 -2.01733544e-01 -1.00028224e-01
-2.30583444e-01 -8.15718353e-01 -8.77013624e-01 1.92135274e-01
-1.39777374e+00 1.65515225e-02 6.24098599e-01 9.92365658e-01
7.75025368e-01 3.97211552e-01 1.06882103e-01 -5.00213563e-01
5.31531632e-01 -5.20214021e-01 -3.19214761e-01 -3.08975894e-02
-9.77096379e-01 2.05242768e-01 9.08145130e-01 -3.28226268e-01
-8.07646453e-01 -4.65935767e-01 -2.44017422e-01 -9.52860236e-01
2.03798190e-02 4.26537365e-01 -1.69938505e-01 -1.64886504e-01
8.70262921e-01 5.72460234e-01 1.64690278e-02 -3.26102048e-01
2.92571187e-01 7.75982261e-01 1.07543397e+00 -2.97198534e-01
7.68738985e-01 4.25584882e-01 1.44758550e-02 -4.89584386e-01
-2.26657435e-01 1.83152109e-02 7.81248808e-02 -2.92678867e-02
5.97346544e-01 -1.05759597e+00 -2.48469457e-01 3.97875190e-01
-1.04621041e+00 -1.31040230e-01 -4.65643138e-01 4.47180092e-01
-1.02293384e+00 5.00226796e-01 -8.27584088e-01 -5.70365191e-01
-2.51524061e-01 -1.35274220e+00 9.76284087e-01 -4.92247008e-02
2.63903942e-02 -8.61760437e-01 -1.77876264e-01 3.21887672e-01
7.03862309e-01 4.34878021e-01 7.79665172e-01 -3.21949762e-03
-7.30785251e-01 -2.90765554e-01 -2.23249838e-01 1.07135606e+00
1.49961889e-01 -2.33630255e-01 -9.70991790e-01 -6.17422938e-01
2.46050492e-01 -3.73265147e-01 9.32600439e-01 4.50782746e-01
1.43386161e+00 -7.78571427e-01 3.92788231e-01 1.20832205e+00
1.74529672e+00 2.80443698e-01 1.34617710e+00 1.08704582e-01
3.27328235e-01 -5.62522523e-02 2.34963924e-01 5.83486676e-01
-3.91916744e-02 6.73779011e-01 7.66455889e-01 1.01900734e-01
-6.73414111e-01 -5.72651863e-01 4.92593676e-01 6.16896391e-01
1.67849004e-01 -7.01076806e-01 -6.12393975e-01 5.31512499e-01
-1.24992597e+00 -1.04688632e+00 4.48902190e-01 2.14257693e+00
1.13872170e+00 1.48740234e-02 -5.12056470e-01 3.45172107e-01
5.47139466e-01 4.36862648e-01 -5.69963396e-01 -4.10802722e-01
-4.65224177e-01 1.31547466e-01 8.68268192e-01 5.96172750e-01
-8.13868582e-01 5.10889113e-01 7.03349543e+00 9.33154464e-01
-1.15555596e+00 1.49459466e-01 1.12820971e+00 -1.15648225e-01
-5.15525937e-01 -5.93055971e-02 -1.80071995e-01 8.60126019e-01
1.36510515e+00 -3.53259116e-01 9.22142744e-01 7.06483841e-01
1.96304828e-01 8.38591084e-02 -1.08757854e+00 1.36104894e+00
2.19457000e-01 -1.46991575e+00 3.62788618e-01 3.29468161e-01
1.00859809e+00 -1.49126694e-01 6.54052317e-01 -1.02166973e-01
1.34035885e-01 -1.57045805e+00 9.77880478e-01 5.43990254e-01
1.30757082e+00 -6.82735264e-01 8.13113511e-01 1.72785342e-01
-5.12332559e-01 1.04461834e-01 -6.81752384e-01 4.14532959e-01
2.31581017e-01 6.73668206e-01 -6.44001424e-01 3.72225672e-01
7.21772194e-01 7.00814247e-01 -4.77906585e-01 8.72434378e-01
-5.03688492e-02 8.35379779e-01 -1.25075459e-01 6.35245919e-01
1.88343510e-01 -2.27456652e-02 6.99302554e-01 1.14522338e+00
7.24040985e-01 -2.42719613e-03 -4.48678762e-01 1.16701305e+00
-4.47783738e-01 -3.15366030e-01 -6.52916431e-01 -2.90864766e-01
4.09332514e-01 5.41289866e-01 -4.38574981e-03 -2.79402137e-01
-1.30607307e-01 1.31017160e+00 -2.87415504e-01 5.37647426e-01
-9.11704898e-01 -2.08810329e-01 9.25020993e-01 1.99641779e-01
5.55612445e-01 -3.09654415e-01 -4.06527549e-01 -1.12029445e+00
-4.53840978e-02 -1.32323802e+00 -2.12175399e-01 -1.12724924e+00
-9.16575134e-01 5.36660850e-01 -7.40654469e-02 -1.40561748e+00
-5.24590492e-01 -3.08084965e-01 -1.92509174e-01 1.08418834e+00
-1.49148929e+00 -7.55150974e-01 -4.25711513e-01 5.59212685e-01
4.28053916e-01 -2.18211621e-01 7.95750260e-01 5.65432250e-01
-2.04271879e-02 9.46780920e-01 5.92144012e-01 5.80274425e-02
5.73974252e-01 -9.20879126e-01 4.32873696e-01 1.12299168e+00
1.97726458e-01 3.66067648e-01 1.22010660e+00 -3.59640658e-01
-1.41327405e+00 -1.24797440e+00 4.53407884e-01 3.24960463e-02
1.40541583e-01 1.50146708e-01 -8.69119883e-01 5.63742042e-01
2.80144274e-01 3.62012655e-01 3.75567168e-01 -7.68027663e-01
-5.23276746e-01 -1.83061555e-01 -1.54693544e+00 4.91145343e-01
8.41473997e-01 -8.43267798e-01 -9.05972794e-02 1.28438264e-01
8.96623433e-01 -4.33130175e-01 -1.04149199e+00 2.80674756e-01
3.69959980e-01 -1.16285551e+00 1.25695300e+00 -3.70125204e-01
1.16923928e+00 -1.69202164e-01 -9.65990424e-01 -1.53821933e+00
-3.85142595e-01 -5.24884582e-01 -3.89148414e-01 1.03412926e+00
-4.60355058e-02 -1.62830487e-01 6.29847944e-01 4.00351405e-01
1.14032184e-03 -4.82636094e-01 -1.02467728e+00 -1.12252665e+00
1.38152570e-01 -4.73037094e-01 5.40120721e-01 5.44907212e-01
-6.25880063e-01 -2.63542950e-01 -8.29085350e-01 1.51882574e-01
9.02051389e-01 -3.58376473e-01 5.56702912e-01 -3.59400481e-01
-6.31833315e-01 -2.46704653e-01 -9.78032887e-01 -1.33490741e+00
-1.20317139e-01 -7.98994184e-01 9.30764526e-02 -1.17774045e+00
2.87387758e-01 -2.42352635e-01 -3.55309159e-01 1.20867446e-01
7.64891580e-02 5.20030558e-01 3.35444421e-01 3.98821652e-01
-1.89402491e-01 7.83992410e-01 1.43476641e+00 -5.20892739e-01
3.26129109e-01 -4.87106800e-01 -8.16744924e-01 1.82923377e-01
5.94880104e-01 -5.78123808e-01 -5.54636896e-01 -8.21087301e-01
6.52257577e-02 3.15697521e-01 6.31648779e-01 -1.36202312e+00
1.26591906e-01 -1.06470823e-01 5.32269120e-01 -1.43599063e-01
5.48793018e-01 -6.89324617e-01 5.90317667e-01 4.77115750e-01
-5.66832542e-01 -7.40063787e-02 4.92718294e-02 6.39796674e-01
-6.14367783e-01 -2.53997862e-01 1.03942633e+00 2.12108772e-02
-4.00272071e-01 2.10446402e-01 -4.34841141e-02 5.67364283e-02
6.26426160e-01 -1.38055772e-01 -4.38668400e-01 -1.11119044e+00
-3.60273868e-01 -4.22983289e-01 8.30110729e-01 1.81254178e-01
9.99459326e-01 -1.50144160e+00 -1.09532893e+00 2.44037598e-01
-5.54180779e-02 -2.91262776e-01 2.50424623e-01 3.40818375e-01
-9.21638668e-01 3.49991560e-01 -4.14106339e-01 -5.34975290e-01
-1.14490807e+00 4.79917675e-01 3.90848845e-01 -1.41169503e-01
-5.20802438e-01 7.63248384e-01 1.76894203e-01 3.53481412e-01
2.21302316e-01 -1.30317107e-01 5.03470242e-01 -5.78219533e-01
6.12602174e-01 2.53765047e-01 1.22455240e-03 -7.41163552e-01
-2.47231033e-02 1.60924196e-01 1.45946950e-01 -3.16437662e-01
1.29651034e+00 -3.05984020e-01 1.98719800e-01 -4.85808402e-03
1.79727137e+00 -1.80580720e-01 -1.66288626e+00 -6.23766109e-02
-7.14832306e-01 -1.19352400e+00 5.81608713e-01 -7.12626159e-01
-1.36767828e+00 8.82661700e-01 9.53010261e-01 1.65533736e-01
1.41392553e+00 -2.58118093e-01 1.01999807e+00 2.46876422e-02
3.93188059e-01 -7.13976264e-01 1.89405411e-01 9.96470917e-03
1.25568569e+00 -1.05517387e+00 9.65310186e-02 -8.06306452e-02
-5.80668151e-01 8.03505421e-01 -3.56183834e-02 -1.79419741e-01
3.65660012e-01 3.53533477e-01 -1.43973351e-01 3.49433243e-01
-7.79081762e-01 4.35018539e-01 1.59224465e-01 6.25680983e-01
1.97653785e-01 1.21803008e-01 -6.33333698e-02 1.59262612e-01
-4.71194685e-01 -2.19011158e-02 7.55442202e-01 7.44990408e-01
-1.64088428e-01 -7.81030297e-01 -6.10035777e-01 3.70858252e-01
-7.74362445e-01 -3.73654872e-01 1.10109933e-01 2.78214425e-01
1.27817526e-01 1.09148395e+00 2.49061242e-01 -6.82122350e-01
-6.54815510e-02 -4.35418993e-01 5.26576817e-01 -9.44705456e-02
-1.24870941e-01 -4.52076420e-02 -1.77388191e-01 -7.43612647e-01
-1.69141367e-01 -5.37446141e-01 -7.70213842e-01 -9.49754655e-01
7.90486038e-02 -1.56032801e-01 8.01740646e-01 4.64553475e-01
3.91907424e-01 6.28838837e-01 9.17232990e-01 -6.82762027e-01
-9.48557436e-01 -8.53720129e-01 -6.36890411e-01 8.13746870e-01
6.77404463e-01 -1.83192074e-01 -6.48424089e-01 5.38210928e-01] | [11.445757865905762, -1.5955421924591064] |
bfeffae5-5cd4-4919-9776-1c0d17a73e90 | indoor-localization-with-robust-global | 2210.06294 | null | https://arxiv.org/abs/2210.06294v1 | https://arxiv.org/pdf/2210.06294v1.pdf | Indoor Localization with Robust Global Channel Charting: A Time-Distance-Based Approach | Fingerprinting-based positioning significantly improves the indoor localization performance in non-line-of-sight-dominated areas. However, its deployment and maintenance is cost-intensive as it needs ground-truth reference systems for both the initial training and the adaption to environmental changes. In contrast, channel charting (CC) works without explicit reference information and only requires the spatial correlations of channel state information (CSI). While CC has shown promising results in modelling the geometry of the radio environment, a deeper insight into CC for localization using multi-anchor large-bandwidth measurements is still pending. We contribute a novel distance metric for time-synchronized single-input/single-output CSIs that approaches a linear correlation to the Euclidean distance. This allows to learn the environment's global geometry without annotations. To efficiently optimize the global channel chart we approximate the metric with a Siamese neural network. This enables full CC-assisted fingerprinting and positioning only using a linear transformation from the chart to the real-world coordinates. We compare our approach to the state-of-the-art of CC on two different real-world data sets recorded with a 5G and UWB radio setup. Our approach outperforms others with localization accuracies of 0.69m for the UWB and 1.4m for the 5G setup. We show that CC-assisted fingerprinting enables highly accurate localization and reduces (or eliminates) the need for annotated training data. | ['Christopher Mutschler', 'Bjoern M. Eskofier', 'Tobias Feigl', 'George Yammine', 'Maximilian Stahlke'] | 2022-10-07 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [ 5.32424897e-02 -1.35769501e-01 5.51806912e-02 -4.32105094e-01
-1.16340959e+00 -8.09433341e-01 9.83389989e-02 2.34920651e-01
-6.11808419e-01 8.93562138e-01 -2.25114986e-01 -5.84685981e-01
-3.95957470e-01 -6.39183342e-01 -9.76989508e-01 -8.12564552e-01
-6.30415082e-01 3.39994341e-01 -1.19385138e-01 7.85254985e-02
8.70464444e-02 5.91090858e-01 -8.53564858e-01 -5.78284442e-01
7.21895933e-01 1.42416501e+00 2.34192371e-01 9.83721077e-01
1.62291557e-01 5.19051440e-02 -9.96102452e-01 1.45568335e-02
3.27375740e-01 -2.49434575e-01 -1.71993345e-01 -4.39571977e-01
5.45935750e-01 -1.58035591e-01 -4.23883855e-01 5.77683449e-01
1.04624820e+00 -1.89867958e-01 8.72720033e-02 -1.10065103e+00
-7.78911039e-02 4.51483101e-01 -1.76303431e-01 -1.48418784e-01
5.38117468e-01 -5.35394847e-01 4.79372889e-01 -3.69570434e-01
3.60022008e-01 5.80098450e-01 1.45781636e+00 -3.07584386e-02
-1.14044487e+00 -6.99562430e-01 -2.92936116e-01 -1.48589006e-02
-1.86034942e+00 -5.28468311e-01 4.64082301e-01 -3.88378128e-02
4.65608180e-01 3.60992134e-01 4.22670096e-01 1.01340365e+00
6.83884025e-02 -1.64152938e-03 6.01049662e-01 -6.64112270e-01
4.35361862e-01 -7.20209032e-02 -4.29520339e-01 5.55660903e-01
3.96144837e-01 1.26023188e-01 -5.76605499e-01 -3.02646011e-01
8.52194428e-01 -1.86716542e-01 -5.15449822e-01 -7.28287756e-01
-1.44062066e+00 2.25789249e-01 6.30484521e-01 4.98764634e-01
-3.40657622e-01 9.14757609e-01 -1.92638457e-01 3.90544772e-01
-1.47918254e-01 7.37417281e-01 -5.38396180e-01 -5.44468999e-01
-1.33117843e+00 -1.00358397e-01 9.75810111e-01 1.35487890e+00
8.92718673e-01 1.14143966e-02 -4.41752821e-02 5.34451187e-01
4.24750775e-01 1.19287717e+00 1.82101019e-02 -1.10568488e+00
8.05657268e-01 -2.43145168e-01 6.29145384e-01 -1.23309672e+00
-9.11887765e-01 -1.36748624e+00 -6.92577481e-01 -4.99176353e-01
1.04449821e+00 -6.53687596e-01 -5.73344707e-01 1.70793247e+00
5.99316172e-02 6.18030846e-01 -1.53442010e-01 4.33516175e-01
1.25253163e-02 2.29558215e-01 -6.00636363e-01 5.60326986e-02
7.68704951e-01 -6.94564104e-01 -5.44684649e-01 -1.79246455e-01
7.65865445e-01 -6.78127885e-01 4.37075436e-01 2.61829019e-01
-5.28915167e-01 -4.26788688e-01 -1.27613533e+00 6.25782073e-01
-3.82341772e-01 3.21784407e-01 3.26682657e-01 1.44990969e+00
-1.24443519e+00 3.88639688e-01 -7.45972931e-01 -5.32656133e-01
1.53628081e-01 7.52843559e-01 -3.45757633e-01 -2.42162198e-01
-1.29591072e+00 6.26993299e-01 -9.04476084e-03 3.83710116e-01
-1.47722617e-01 -8.31743956e-01 -7.14413226e-01 1.25767171e-01
2.22944543e-01 -3.66933405e-01 1.20629692e+00 -3.95802200e-01
-1.70675778e+00 -9.58561450e-02 -3.07651043e-01 -7.05941737e-01
5.49242795e-01 -9.33357254e-02 -8.03749442e-01 -3.83826718e-02
1.54948458e-01 -1.67783182e-02 4.52855080e-01 -1.20601809e+00
-6.02984726e-01 -2.23659635e-01 -1.75744653e-01 -2.98999071e-01
-1.09863058e-02 -7.36388445e-01 -7.61412382e-01 -5.07230341e-01
7.56250381e-01 -1.32977688e+00 -4.23758894e-01 -5.48423305e-02
-4.26582307e-01 1.00112176e+00 4.62123126e-01 -5.94081759e-01
1.32074678e+00 -1.88481820e+00 -6.21448398e-01 1.00842500e+00
-1.97232902e-01 -2.29263142e-01 8.20276588e-02 7.36151397e-01
4.10094887e-01 9.15296748e-03 6.48855418e-02 -6.25543296e-01
-7.94258788e-02 1.21675022e-01 -2.23305169e-03 7.70621598e-01
-7.49661446e-01 6.33585036e-01 -1.11299992e+00 -5.64999990e-02
2.18647361e-01 5.47652364e-01 -4.89815116e-01 -1.80296406e-01
5.17173529e-01 9.02273357e-01 -3.57302397e-01 6.41219616e-01
9.59182978e-01 -1.48002744e-01 4.18238342e-01 -2.89255977e-01
-3.03156823e-01 -1.98106728e-02 -1.38013268e+00 2.07926607e+00
-1.13626885e+00 8.53146970e-01 1.63970634e-01 -8.40477586e-01
1.03582036e+00 3.20738375e-01 6.29175365e-01 -1.03883219e+00
8.76285508e-03 6.83234453e-01 -1.80720150e-01 -1.38230845e-01
3.24810266e-01 3.20632190e-01 -4.59532738e-01 3.84935379e-01
2.37038569e-03 1.01277143e-01 -2.99855232e-01 -5.16371876e-02
1.32399201e+00 2.27062795e-02 2.71334767e-01 -2.16006309e-01
6.60282135e-01 -4.72555220e-01 4.26707566e-01 1.37802815e+00
1.27345815e-01 6.19928062e-01 6.72126263e-02 -9.09972116e-02
-5.80653727e-01 -1.02598703e+00 -2.03932896e-01 5.46369791e-01
4.80377674e-01 -3.87172431e-01 -5.50911725e-01 -2.43142158e-01
4.04965401e-01 5.81272006e-01 -5.68809569e-01 1.73575670e-01
-5.06324291e-01 -4.14956301e-01 1.03089857e+00 2.15754420e-01
7.59225368e-01 7.30083957e-02 -2.83693522e-01 6.51575565e-01
-2.08122864e-01 -1.40606773e+00 -4.33452845e-01 3.67395103e-01
-3.58400077e-01 -6.88751757e-01 -7.81134248e-01 -2.75619864e-01
7.04590380e-01 3.52339447e-01 5.57889283e-01 -1.28353700e-01
1.04394086e-01 6.68204248e-01 -1.74232662e-01 -1.72710896e-01
1.29223922e-02 4.73013997e-01 2.48288468e-01 2.58842111e-01
-3.14306378e-01 -8.56983781e-01 -6.79927051e-01 8.23913455e-01
4.95182723e-02 -4.90777194e-01 6.17137074e-01 6.56667590e-01
4.38757926e-01 -9.08402205e-02 4.37184602e-01 -5.26314676e-01
3.02536756e-01 -4.69664991e-01 -9.90137935e-01 2.45109111e-01
-7.81157196e-01 1.87585726e-01 6.74364448e-01 -7.64307305e-02
-6.85760260e-01 3.09417844e-01 -1.33398503e-01 1.47899479e-01
1.65756837e-01 5.10620713e-01 -2.87493497e-01 -9.61174965e-01
6.35057390e-01 9.87905711e-02 -5.13218880e-01 -5.79361498e-01
2.85177678e-01 8.02186310e-01 8.70371222e-01 -7.01650679e-01
1.14933026e+00 4.64842588e-01 4.84404653e-01 -9.38332617e-01
-5.02001405e-01 -6.49714231e-01 -7.32979476e-01 1.91344842e-02
1.05872922e-01 -7.85291374e-01 -8.05460155e-01 2.89879203e-01
-1.07566798e+00 -3.19360167e-01 1.20265760e-01 6.21720672e-01
-4.40191180e-01 2.59886414e-01 2.47170478e-02 -8.56941462e-01
-4.71871756e-02 -7.94844329e-01 1.00451231e+00 1.35580629e-01
-8.68107975e-02 -1.04748034e+00 2.18652084e-01 -3.09226178e-02
1.08383107e+00 4.15057153e-01 2.34650090e-01 -1.28654420e-01
-7.40776896e-01 -7.68929303e-01 -6.58737943e-02 -9.90256369e-02
1.30736828e-01 -5.45621514e-01 -1.06642282e+00 -3.15824509e-01
-4.57371205e-01 6.33484125e-01 2.09532499e-01 6.04720116e-01
1.13038707e+00 -1.87620148e-01 -7.11407840e-01 1.27497804e+00
1.48156631e+00 2.30618730e-01 6.52286947e-01 5.12412667e-01
5.25335133e-01 -1.70899153e-01 6.31171465e-01 5.59629798e-01
5.65963507e-01 1.23823559e+00 4.55820918e-01 -2.67810792e-01
2.02766240e-01 -4.30053890e-01 -1.64064646e-01 2.37863734e-01
-1.32581592e-01 -4.14019644e-01 -8.29044878e-01 1.68751016e-01
-1.68943059e+00 -9.21312094e-01 -2.86936313e-01 2.73113513e+00
2.20971167e-01 -1.50888059e-02 -2.29027674e-01 2.60202765e-01
5.13326287e-01 -4.06074859e-02 -2.99108416e-01 -7.97671378e-02
4.41570617e-02 2.02459589e-01 1.65993083e+00 8.60090196e-01
-1.11789918e+00 3.73555571e-01 5.51356697e+00 3.18944842e-01
-1.40413165e+00 8.92893448e-02 1.52729094e-01 1.67568356e-01
-1.83660220e-02 4.24405262e-02 -5.06479323e-01 6.23465002e-01
1.23711836e+00 2.17013359e-01 5.22727370e-01 7.16389298e-01
2.46697009e-01 -3.42429399e-01 -8.50715160e-01 1.42469728e+00
-2.79067397e-01 -1.55752909e+00 -8.72722447e-01 3.59765649e-01
6.77079499e-01 3.04591715e-01 -2.73548067e-03 3.43584716e-02
-3.63113135e-01 -7.12159157e-01 9.15499389e-01 7.76123881e-01
1.08292830e+00 -5.82624376e-01 1.07148993e+00 1.58772290e-01
-1.59740126e+00 -2.04672411e-01 1.43972943e-02 7.66312853e-02
2.69880772e-01 7.95556664e-01 -1.26922679e+00 9.65757132e-01
4.64001924e-01 2.12966144e-01 -6.57246470e-01 1.65494943e+00
-1.17766380e-01 5.49616992e-01 -7.27370203e-01 1.25640705e-02
1.73148975e-01 1.05400600e-01 2.40845889e-01 1.19593394e+00
1.08753836e+00 -3.62892777e-01 2.43119430e-02 2.15815067e-01
1.57372430e-01 -1.82383850e-01 -1.43438324e-01 5.08338571e-01
1.18716097e+00 9.72152650e-01 -5.06403565e-01 2.20594376e-01
-5.49751930e-02 8.78281474e-01 -4.41619098e-01 7.18954027e-01
-6.99809551e-01 -1.03604984e+00 4.30760324e-01 2.02509001e-01
4.50247526e-01 -8.32172573e-01 -2.40358725e-01 -7.17543125e-01
4.40291828e-03 -1.93215713e-01 -2.66169161e-01 -1.71170831e-01
-3.92883718e-01 4.60508019e-01 -4.45745528e-01 -1.38153327e+00
-5.37074625e-01 -1.64472073e-01 -3.26796055e-01 1.00956845e+00
-1.46745658e+00 -9.68239903e-01 -7.53490746e-01 4.34108466e-01
-5.24380982e-01 6.18488230e-02 1.21615946e+00 8.84357810e-01
-2.20197305e-01 9.61421967e-01 9.42736447e-01 1.36699080e-01
7.84525990e-01 -1.30515277e+00 7.02739596e-01 7.73594201e-01
1.36263162e-01 7.00901389e-01 8.93361807e-01 -2.71847069e-01
-1.60671711e+00 -8.09251308e-01 9.13562119e-01 -2.79919654e-01
5.12832165e-01 -7.10758984e-01 -2.29939684e-01 3.85731012e-01
-4.94022906e-01 4.82784361e-01 6.86390102e-01 2.51947850e-01
-9.44730714e-02 -7.12251544e-01 -1.24244595e+00 2.04515338e-01
1.07434130e+00 -5.24476767e-01 5.21711588e-01 1.28881522e-02
1.48737088e-01 -5.93218982e-01 -9.66167688e-01 -9.94626582e-02
1.20521057e+00 -8.61188293e-01 1.13779593e+00 6.02365553e-01
-9.54193413e-01 -6.88460588e-01 -5.14482498e-01 -1.19976079e+00
-7.76451454e-02 -1.10961783e+00 -7.78824091e-02 1.13788462e+00
5.32433033e-01 -1.02704179e+00 1.01310575e+00 1.91354871e-01
1.04052082e-01 -4.75432754e-01 -1.33200717e+00 -9.79998946e-01
-6.94730759e-01 -8.02325666e-01 1.06828701e+00 6.98545337e-01
-3.40158552e-01 -1.93078876e-01 -5.04768908e-01 8.89998794e-01
9.19595540e-01 -2.20885724e-01 1.22249186e+00 -1.38404393e+00
-4.18452621e-01 -1.46610318e-02 -6.33715987e-01 -1.37961853e+00
-3.59351605e-01 -4.66404140e-01 1.23341367e-01 -1.48560381e+00
-1.01865196e+00 -1.36913621e+00 -3.08497250e-01 1.55699328e-01
5.11761308e-01 3.08355093e-01 -1.16136901e-01 -2.01348037e-01
-6.40374601e-01 2.22797036e-01 5.68486512e-01 7.92204291e-02
-3.89657468e-01 8.02496314e-01 -4.32502240e-01 2.77372986e-01
9.05936301e-01 -4.85981792e-01 -2.84674942e-01 -5.10134280e-01
3.03644329e-01 3.49999785e-01 2.00385541e-01 -1.84194803e+00
4.95900869e-01 1.70657158e-01 5.73527932e-01 -4.16880399e-01
4.35698122e-01 -1.25457442e+00 4.58149672e-01 4.48980093e-01
9.76696238e-02 -2.79530346e-01 7.82140791e-02 7.02920377e-01
7.99786896e-02 -2.94935144e-02 3.71249229e-01 4.16257381e-01
-5.15614927e-01 3.16411674e-01 -5.15710711e-02 -2.62791425e-01
5.44134557e-01 -5.17661572e-01 -1.33570015e-01 -1.02032471e+00
-4.40371513e-01 1.19813673e-01 5.54351568e-01 -3.22168246e-02
3.62551473e-02 -1.28134441e+00 -1.43297330e-01 2.68717319e-01
1.25929385e-01 -3.53545189e-01 1.96203947e-01 9.60267723e-01
-7.93449998e-01 9.09473777e-01 5.31971976e-02 -7.08322525e-01
-9.31753099e-01 -1.15332708e-01 8.26777875e-01 -3.93898115e-02
-8.41587409e-03 6.95819676e-01 -5.99146962e-01 -6.02293670e-01
5.57504714e-01 -5.26400983e-01 3.69408935e-01 -2.90549845e-01
4.29051846e-01 5.37090898e-01 4.10265714e-01 -5.46825349e-01
-6.59538209e-01 1.01740122e+00 6.22787178e-01 -3.52153778e-01
1.04960418e+00 -6.12870753e-01 2.79842794e-01 1.67093068e-01
1.42204261e+00 6.32760525e-01 -1.20490527e+00 -2.72931516e-01
3.81809980e-01 -5.78736007e-01 5.86665496e-02 -8.48756135e-01
-8.68479252e-01 6.22361004e-01 1.05910087e+00 2.31483564e-01
7.82982707e-01 -3.24060798e-01 5.91288686e-01 8.52989495e-01
1.27258992e+00 -8.86312544e-01 -5.33273518e-01 4.17044073e-01
3.90066892e-01 -9.56326485e-01 4.26573073e-03 -2.51935005e-01
2.14808866e-01 1.26340544e+00 -1.79925069e-01 2.65010595e-01
7.90964365e-01 3.10312897e-01 3.04266572e-01 2.93216795e-01
2.87767500e-01 5.13899215e-02 9.44487825e-02 1.09986258e+00
2.66500026e-01 1.62643522e-01 2.52013296e-01 4.82370794e-01
-5.73117971e-01 -2.11874664e-01 3.19089592e-01 8.54702115e-01
-3.31969351e-01 -1.38879097e+00 -7.09927142e-01 2.98274338e-01
-2.64915109e-01 1.22102529e-01 2.61032879e-01 7.88608670e-01
2.62428731e-01 1.05929518e+00 -2.14019045e-01 -3.61817807e-01
3.32027048e-01 -3.50284696e-01 5.63008964e-01 -1.22492187e-01
-9.69188735e-02 -8.47890899e-02 2.98365414e-01 -9.72081602e-01
-4.51037921e-02 -6.68195784e-01 -1.19284916e+00 -8.73461217e-02
-3.90118182e-01 5.06424665e-01 1.44756353e+00 7.66922772e-01
5.98215103e-01 5.29469967e-01 8.44113946e-01 -1.01527309e+00
-1.70876071e-01 -4.83915895e-01 -7.24768341e-01 -5.83137512e-01
7.09321380e-01 -5.64051092e-01 -1.96422532e-01 -4.92323399e-01] | [6.34637975692749, 0.955189049243927] |
e2be40e1-3f4b-4de7-931a-c2daf274512a | extending-adversarial-attacks-and-defenses-to | 1901.03006 | null | https://arxiv.org/abs/1901.03006v4 | https://arxiv.org/pdf/1901.03006v4.pdf | Extending Adversarial Attacks and Defenses to Deep 3D Point Cloud Classifiers | 3D object classification and segmentation using deep neural networks has been extremely successful. As the problem of identifying 3D objects has many safety-critical applications, the neural networks have to be robust against adversarial changes to the input data set. There is a growing body of research on generating human-imperceptible adversarial attacks and defenses against them in the 2D image classification domain. However, 3D objects have various differences with 2D images, and this specific domain has not been rigorously studied so far. We present a preliminary evaluation of adversarial attacks on deep 3D point cloud classifiers, namely PointNet and PointNet++, by evaluating both white-box and black-box adversarial attacks that were proposed for 2D images and extending those attacks to reduce the perceptibility of the perturbations in 3D space. We also show the high effectiveness of simple defenses against those attacks by proposing new defenses that exploit the unique structure of 3D point clouds. Finally, we attempt to explain the effectiveness of the defenses through the intrinsic structures of both the point clouds and the neural network architectures. Overall, we find that networks that process 3D point cloud data are weak to adversarial attacks, but they are also more easily defensible compared to 2D image classifiers. Our investigation will provide the groundwork for future studies on improving the robustness of deep neural networks that handle 3D data. | ['Ronald Yu', 'Hao Su', 'Daniel Liu'] | 2019-01-10 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [ 3.58441211e-02 3.30601223e-02 4.36097711e-01 -3.82241398e-01
-2.52342790e-01 -1.14466131e+00 7.25031137e-01 -2.13475332e-01
-3.01037282e-01 1.01301089e-01 -6.10835969e-01 -8.13993275e-01
1.50560737e-01 -9.66130793e-01 -1.16341257e+00 -7.23103762e-01
-4.24716443e-01 3.12732935e-01 4.96840030e-01 -4.10212725e-01
3.12550038e-01 1.71752298e+00 -1.33565807e+00 8.52750391e-02
5.08607566e-01 1.05748498e+00 -5.89879274e-01 8.88135791e-01
1.97659410e-03 2.38014981e-01 -1.22760963e+00 -4.53540385e-01
9.81300652e-01 2.27587223e-01 -6.85805738e-01 -1.61351472e-01
8.91556144e-01 -4.78153855e-01 -5.37589133e-01 1.52237022e+00
6.40544295e-01 -8.04356411e-02 6.12033725e-01 -1.86651695e+00
-9.37839389e-01 8.40730742e-02 -3.70381504e-01 2.31454894e-01
8.28792035e-05 4.95273501e-01 2.12183908e-01 -2.44034305e-01
2.57842660e-01 1.73728037e+00 7.43149877e-01 1.08195567e+00
-9.53743339e-01 -1.13051522e+00 1.58117399e-01 -4.21796553e-02
-1.11278319e+00 2.41889328e-01 1.01427102e+00 -3.00926298e-01
1.05063951e+00 4.74429935e-01 3.90581876e-01 1.59903014e+00
4.19944912e-01 6.71793401e-01 9.46032345e-01 -7.80021623e-02
2.19618425e-01 -1.41883343e-01 2.09432572e-01 3.64573359e-01
3.15417260e-01 7.72396088e-01 2.38806322e-01 -4.86636162e-01
8.48118663e-01 -9.88852158e-02 2.94077042e-02 -4.79490459e-01
-7.49798119e-01 9.94924545e-01 1.04004776e+00 6.71752393e-02
-1.12761423e-01 3.13581586e-01 4.91611987e-01 3.74700069e-01
4.73977000e-01 6.10071480e-01 -4.56508666e-01 5.71438730e-01
-3.39948565e-01 5.99858105e-01 6.21443808e-01 1.10854232e+00
3.51437569e-01 3.86686653e-01 4.05482203e-01 2.96674907e-01
2.71526635e-01 8.81914616e-01 2.11711936e-02 -9.36343372e-01
2.19966054e-01 4.01057929e-01 -2.01600015e-01 -1.37964165e+00
-3.90042096e-01 -1.42444968e-01 -9.28436399e-01 1.16834497e+00
2.70970762e-01 -1.72112405e-01 -1.23617232e+00 1.64472020e+00
4.89740670e-01 4.67020988e-01 1.91177711e-01 8.32497001e-01
6.80144966e-01 5.56404948e-01 3.46596688e-02 6.43472970e-01
1.02577972e+00 -2.40445673e-01 -1.14794955e-01 -4.43532206e-02
3.33984315e-01 -5.72503686e-01 5.16448081e-01 1.30424559e-01
-8.57436717e-01 -7.63948202e-01 -1.45153463e+00 1.66588977e-01
-7.72937179e-01 -8.76088440e-01 5.36029816e-01 1.33807373e+00
-8.31142306e-01 7.11361706e-01 -1.05443573e+00 -1.38919726e-01
8.67107987e-01 7.70003200e-01 -4.82963443e-01 3.32169920e-01
-1.44026589e+00 1.15887153e+00 3.87065709e-01 -3.39247175e-02
-9.95130479e-01 -6.95857823e-01 -7.46652603e-01 -3.10449839e-01
-1.34924605e-01 -6.38489783e-01 1.08911169e+00 -6.74706757e-01
-1.10042787e+00 1.17126703e+00 5.88743746e-01 -9.53402460e-01
6.15202665e-01 -1.90548733e-01 -3.76958281e-01 3.53938192e-01
-1.76233202e-01 1.03100860e+00 1.08638191e+00 -1.46887064e+00
-4.65577096e-01 -7.15148628e-01 3.11131686e-01 -5.19572943e-02
-5.78358918e-02 1.45124063e-01 3.98956239e-02 -7.78058052e-01
1.35698795e-01 -1.40526962e+00 -3.84496391e-01 4.25731510e-01
-7.14318991e-01 -1.51526377e-01 1.49393678e+00 1.52780816e-01
-6.41792268e-02 -2.40911984e+00 -3.87273699e-01 4.07614231e-01
2.28463143e-01 8.50231111e-01 -2.65702248e-01 1.37318755e-02
-5.34842849e-01 7.52290666e-01 -2.29835615e-01 8.44162330e-02
1.13081202e-01 4.00269866e-01 -9.96280313e-01 7.82517493e-01
6.11229002e-01 8.96253943e-01 -6.91294253e-01 -3.70747782e-02
5.02616286e-01 4.69445407e-01 -4.77836549e-01 6.36446923e-02
-8.05428177e-02 2.76146084e-01 -7.08772779e-01 5.23514867e-01
1.28837895e+00 3.09907377e-01 -7.73418009e-01 1.18308470e-01
3.86389554e-01 -3.54933739e-02 -8.40015829e-01 1.07455480e+00
6.39985502e-02 6.01238191e-01 8.80709738e-02 -8.86069536e-01
1.19664812e+00 2.35961109e-01 4.20731604e-01 -9.12126750e-02
4.05320227e-01 1.64491013e-01 1.50183186e-01 -1.74406230e-01
2.28088915e-01 -1.97011188e-01 -2.83465862e-01 3.23590845e-01
-1.00396924e-01 -9.66734946e-01 -7.12300122e-01 9.13596302e-02
1.15199935e+00 -1.98210910e-01 -3.13082218e-01 -2.12208088e-02
4.37063813e-01 1.60129115e-01 2.82896608e-01 1.16246688e+00
-6.26212537e-01 7.20690370e-01 3.33949357e-01 -6.96404994e-01
-1.09591985e+00 -1.43302953e+00 -2.44502589e-01 1.55958548e-01
5.25901437e-01 3.13179165e-01 -8.49349558e-01 -1.20429528e+00
5.81456840e-01 7.39779651e-01 -7.00537086e-01 -8.39743435e-01
-5.14574111e-01 -4.36893880e-01 1.60875773e+00 5.94442964e-01
7.79653966e-01 -9.16635096e-01 -5.13604224e-01 -2.28061184e-01
4.41475838e-01 -1.26269996e+00 -5.91508811e-03 3.01442623e-01
-1.03392208e+00 -1.12502956e+00 -3.08615357e-01 -7.53367960e-01
5.77284575e-01 4.26444471e-01 9.45000410e-01 2.52295434e-01
-1.46562994e-01 4.35024798e-01 -3.13664556e-01 -1.32432353e+00
-9.03475523e-01 -1.88295454e-01 5.41120052e-01 -3.98796827e-01
8.39374900e-01 -6.83730066e-01 -1.29519343e-01 6.31820917e-01
-1.17169380e+00 -6.75335288e-01 1.67098135e-01 4.37650084e-01
3.14731985e-01 4.31168944e-01 1.26674235e-01 -6.52372837e-01
5.64183831e-01 -2.45265603e-01 -7.06210494e-01 -4.10490036e-01
1.06552854e-01 -2.98299491e-01 5.96396029e-01 -6.97689354e-01
-3.73147368e-01 1.41605213e-01 -7.15812325e-01 -1.13584638e+00
-8.44456255e-01 -4.57739592e-01 -2.75484771e-01 -8.74027371e-01
1.00652635e+00 -2.31555283e-01 1.68197334e-01 -2.20654085e-01
3.82752270e-01 4.81555492e-01 6.68879271e-01 -5.91791749e-01
1.69052231e+00 8.48294854e-01 3.93611670e-01 -9.71222699e-01
-4.17082876e-01 4.01339233e-02 -6.83309972e-01 4.63976711e-02
9.87040281e-01 -5.40401042e-01 -9.53787744e-01 1.01169956e+00
-1.67947125e+00 9.54702944e-02 -2.99051404e-01 1.57421187e-01
-6.21254802e-01 5.28132796e-01 -6.50241673e-01 -5.71138501e-01
-1.65644929e-01 -1.45547771e+00 9.24277127e-01 -2.13963211e-01
-1.00366823e-01 -7.57090688e-01 -2.47573525e-01 1.60543118e-02
1.53514162e-01 9.42273855e-01 1.11073124e+00 -1.27901387e+00
-6.87092423e-01 -6.77685440e-01 -4.23165783e-02 8.70612741e-01
-5.03906645e-02 2.43487805e-01 -1.22961867e+00 -2.39599943e-01
6.85035229e-01 -2.82329977e-01 5.03144383e-01 1.49121478e-01
1.41874325e+00 -1.84790269e-01 -1.96716160e-01 8.92758250e-01
1.00884557e+00 4.06776696e-01 7.62227297e-01 4.86694008e-01
7.41878986e-01 4.59221363e-01 2.57304132e-01 -2.53214955e-01
-4.82132107e-01 3.93067390e-01 1.29133296e+00 -2.31127262e-01
2.29023144e-01 -7.78810456e-02 8.19324255e-02 -4.47175652e-02
6.29290417e-02 -3.09859306e-01 -9.82690036e-01 1.76314950e-01
-1.30025625e+00 -1.05300581e+00 -2.12045744e-01 1.74131727e+00
3.18416029e-01 7.39163518e-01 -7.91742280e-02 4.43004608e-01
9.99890268e-01 2.37995386e-01 -8.11861098e-01 -6.55953526e-01
-2.24051774e-01 3.93470913e-01 9.12356555e-01 2.33814240e-01
-1.62439358e+00 1.02181244e+00 6.71217203e+00 5.90391159e-01
-1.21884048e+00 -3.34519118e-01 3.92364055e-01 2.25064661e-02
7.06206262e-02 -3.84731382e-01 -6.32977545e-01 3.24317068e-01
6.71417236e-01 1.70046359e-01 -8.58204216e-02 1.29064238e+00
-2.46953592e-01 7.46844351e-01 -1.17628574e+00 7.49085248e-01
-1.41678452e-01 -1.39003348e+00 3.96762133e-01 3.64614397e-01
5.77566445e-01 4.39635217e-01 4.33599174e-01 1.37837559e-01
7.99469113e-01 -1.08242476e+00 7.29191124e-01 -2.20044017e-01
3.03636551e-01 -1.11458206e+00 7.15201437e-01 4.88412201e-01
-5.28246403e-01 6.08664975e-02 -6.94586873e-01 7.55055845e-02
-1.42336324e-01 1.96714655e-01 -7.92232513e-01 3.10142428e-01
1.11663473e+00 4.16567594e-01 -5.01227736e-01 8.22420597e-01
-5.03211379e-01 3.65685433e-01 -7.11902618e-01 -4.80986610e-02
7.46734858e-01 1.67977512e-01 1.27435791e+00 8.10380220e-01
8.85164179e-03 1.85389817e-01 -3.92356627e-02 1.12467504e+00
-6.04591221e-02 -6.26500368e-01 -1.41897643e+00 2.19043866e-01
5.27279496e-01 6.82912290e-01 -7.65585899e-01 1.79398462e-01
-9.46315527e-02 8.12937617e-01 -1.40936330e-01 1.95614100e-01
-8.66897047e-01 -5.66033065e-01 1.50932944e+00 -2.51939923e-01
3.42630982e-01 -6.41192853e-01 -6.45743310e-01 -6.49199903e-01
-2.07235977e-01 -1.10571337e+00 1.28333986e-01 -7.71901250e-01
-1.82050908e+00 7.72912145e-01 8.07362869e-02 -1.35280824e+00
1.53917586e-02 -1.17887628e+00 -8.60099912e-01 7.74680316e-01
-1.07162631e+00 -1.04841268e+00 3.39883231e-02 8.61936450e-01
5.85499816e-02 -4.02051419e-01 8.00182283e-01 -5.67751266e-02
-1.07445449e-01 7.41176844e-01 -2.66287446e-01 7.03363597e-01
2.59676784e-01 -1.08267701e+00 1.45141566e+00 1.08079171e+00
1.74297377e-01 7.15244353e-01 6.43529773e-01 -6.60954773e-01
-1.21473014e+00 -1.27056777e+00 1.71823889e-01 -1.06378865e+00
5.32925308e-01 -5.44730425e-01 -1.17530549e+00 6.59325361e-01
-9.56519619e-02 9.12299007e-02 4.81941938e-01 -2.81548589e-01
-7.08893538e-01 2.02652484e-01 -1.74331844e+00 8.23320448e-01
9.91722703e-01 -6.45168960e-01 -9.06258643e-01 3.17316025e-01
1.32974815e+00 -7.20163465e-01 -6.25745714e-01 6.38197660e-01
-7.95818791e-02 -9.30212140e-01 1.70145261e+00 -9.97215331e-01
1.61513627e-01 -4.26222563e-01 -2.46026650e-01 -1.18143582e+00
-3.00912559e-01 -5.73846996e-01 1.86982632e-01 8.39104831e-01
6.13819547e-02 -8.99246216e-01 1.11071968e+00 6.86680377e-01
-3.96358639e-01 -2.63465077e-01 -1.21982586e+00 -1.22436249e+00
7.69367993e-01 -8.55153322e-01 1.02299559e+00 1.00865138e+00
-8.53268027e-01 -1.55669346e-01 2.87294015e-02 1.16487288e+00
8.12996924e-01 -3.01072419e-01 1.00855720e+00 -1.20398164e+00
2.30703726e-01 -5.20354867e-01 -1.32140231e+00 -7.22823441e-01
4.60531026e-01 -9.94938314e-01 -1.70182511e-01 -7.68406451e-01
-8.09480250e-01 -6.27212226e-01 -1.71953678e-01 5.17737329e-01
-9.98059381e-03 5.63832700e-01 5.94710410e-01 9.04285237e-02
2.29993373e-01 3.06965709e-01 1.13412285e+00 -6.22281492e-01
7.49889854e-03 5.05212963e-01 -5.63761115e-01 1.03037584e+00
8.27434659e-01 -7.82206357e-01 -3.55648309e-01 -7.54675806e-01
-1.93721265e-01 -6.85862541e-01 9.52226579e-01 -1.17513716e+00
1.13215923e-01 -4.94306162e-02 5.95466673e-01 -9.59953129e-01
5.52683413e-01 -1.42010176e+00 -8.01144987e-02 7.00661242e-01
-1.38630480e-01 3.01994337e-03 7.35341191e-01 6.05160534e-01
2.78534926e-02 -2.23553821e-01 1.23460209e+00 -2.70766199e-01
-6.15751684e-01 6.48090422e-01 -3.80632937e-01 -8.65770951e-02
1.39574897e+00 -5.08386791e-01 -3.92531633e-01 -1.07421644e-01
-5.93017459e-01 1.18763223e-01 6.41767919e-01 8.42779398e-01
8.49726379e-01 -1.38177967e+00 -5.94793558e-01 5.41036129e-01
-2.32798532e-01 3.76513124e-01 1.36918891e-02 -2.62583107e-01
-8.92103612e-01 4.48913649e-02 -7.02713311e-01 -1.04097235e+00
-1.31033003e+00 1.14118826e+00 7.19341755e-01 4.26802009e-01
-6.36066735e-01 1.13793862e+00 2.43400842e-01 -7.97950685e-01
4.74683821e-01 -2.36659169e-01 2.49803171e-01 -5.42168379e-01
3.37477654e-01 1.46041140e-01 1.66379824e-01 -5.97402692e-01
-6.25566721e-01 5.36374331e-01 -9.18012187e-02 2.84232825e-01
1.32662201e+00 4.35210288e-01 -1.30567774e-02 -1.54822007e-01
1.38643467e+00 -3.68558198e-01 -1.12440431e+00 1.77564174e-01
-2.61273384e-01 -5.63058376e-01 -1.41682342e-01 -3.92721534e-01
-1.14373529e+00 1.19519067e+00 7.66333163e-01 7.03032792e-01
8.53085160e-01 -1.53779253e-01 8.45716476e-01 6.18216097e-01
4.13396925e-01 -3.77378762e-01 -4.56879884e-02 8.39581370e-01
9.05250490e-01 -9.53682303e-01 -1.75962031e-01 -5.21363080e-01
-3.14732701e-01 9.05014098e-01 7.39328146e-01 -9.23518181e-01
1.00338101e+00 2.99794644e-01 5.03741026e-01 -2.01527014e-01
-2.33403638e-01 1.74190983e-01 1.87496781e-01 1.30685127e+00
-5.87591171e-01 -2.71224290e-01 6.45357728e-01 1.88220143e-01
-5.23413599e-01 -6.61389351e-01 3.84621114e-01 9.86718357e-01
-2.11405784e-01 -1.00882602e+00 -1.00311935e+00 -7.15316534e-02
-5.48025787e-01 2.67188311e-01 -9.88832593e-01 1.11969531e+00
1.78622290e-01 8.23289037e-01 1.58927038e-01 -7.21156299e-01
5.73659360e-01 -4.81960811e-02 2.69436747e-01 -4.35620189e-01
-8.38328123e-01 -5.82518160e-01 -3.68163735e-01 -5.69372714e-01
-2.31948480e-01 -4.18135107e-01 -1.09906518e+00 -7.21335530e-01
-2.09691703e-01 -1.34603232e-01 9.86683726e-01 5.79134762e-01
4.13331479e-01 3.62078875e-01 9.42332029e-01 -1.35387015e+00
-1.02144670e+00 -3.61369550e-01 -3.89753997e-01 6.91776693e-01
5.87037385e-01 -6.58473194e-01 -7.27342308e-01 -4.42914426e-01] | [7.704028129577637, -4.45789909362793] |
f9a0d9f8-a381-4d0c-b741-15dd50e84acd | overview-of-the-ninth-dialog-system | 2011.06486 | null | https://arxiv.org/abs/2011.06486v1 | https://arxiv.org/pdf/2011.06486v1.pdf | Overview of the Ninth Dialog System Technology Challenge: DSTC9 | This paper introduces the Ninth Dialog System Technology Challenge (DSTC-9). This edition of the DSTC focuses on applying end-to-end dialog technologies for four distinct tasks in dialog systems, namely, 1. Task-oriented dialog Modeling with unstructured knowledge access, 2. Multi-domain task-oriented dialog, 3. Interactive evaluation of dialog, and 4. Situated interactive multi-modal dialog. This paper describes the task definition, provided datasets, baselines and evaluation set-up for each track. We also summarize the results of the submitted systems to highlight the overall trends of the state-of-the-art technologies for the tasks. | ['Rajen Subba', 'Shivani Poddar', 'Seungwhan Moon', 'Satwik Kottur', 'Alborz Geramifard', 'Ankita De', 'Paul A. Crook', 'Cho', 'Eunjoon', 'Ahmad Beirami', 'Maxine Eskenazi', 'David Traum', 'Seyed Hossein Alavi', 'Carla Gordon', 'Yulan Feng', 'Shikib Mehri', 'Jianfeng Gao', 'Minlie Huang', 'Swadheen Shukla', 'Zheng Zhang', 'Baolin Peng', 'Runze Liang', 'Shahin Shayandeh', 'Kaili Huang', 'Lars Liden', 'Lingxiao Luo', 'Qi Zhu', 'Jinchao Li', 'Dilek Hakkani-Tür', 'Chao-Wei Huang', 'Yang Liu', 'Karthik Gopalakrishnan', 'Behnam Hedayatnia', 'Mihail Eric', 'Yun-Nung Chen', 'Abhinav Rastogi', "Luis Fernando D'Haro", 'Seokhwan Kim', 'Chulaka Gunasekara'] | 2020-11-12 | null | null | null | null | ['interactive-evaluation-of-dialog'] | ['natural-language-processing'] | [-5.35132051e-01 5.99785388e-01 1.77348182e-01 -6.83765829e-01
-7.07590222e-01 -1.13311911e+00 1.09261668e+00 -2.04409152e-01
-4.72179800e-01 9.11120713e-01 8.16445410e-01 -3.01201612e-01
1.35458902e-01 8.67742673e-02 3.18319231e-01 -1.46083027e-01
2.38792270e-01 1.70119572e+00 4.91070002e-01 -1.08424592e+00
7.30614066e-02 7.39957672e-03 -7.34444737e-01 9.34623897e-01
4.85414833e-01 7.58181691e-01 2.31143519e-01 1.06488967e+00
-6.42577946e-01 1.21228039e+00 -9.52515781e-01 -5.34670472e-01
-2.44458586e-01 -1.85126707e-01 -1.65738499e+00 -3.92035358e-02
2.72564948e-01 -8.66633892e-01 -2.18337327e-01 5.17366707e-01
8.04740131e-01 6.25717163e-01 7.99116194e-01 -1.37015128e+00
-4.49550897e-01 9.84648049e-01 4.21621919e-01 -2.18932062e-01
8.45461965e-01 1.73438594e-01 8.20649922e-01 -9.41450119e-01
8.27208042e-01 1.98845196e+00 4.28468972e-01 1.11949587e+00
-1.05445051e+00 -1.20177373e-01 8.59583244e-02 -2.74275541e-01
-7.91611016e-01 -8.15210402e-01 3.27806115e-01 -7.54162312e-01
1.44268954e+00 2.04843819e-01 -2.53316969e-01 1.57635105e+00
-4.66087684e-02 8.39242995e-01 1.01815248e+00 -4.57518727e-01
4.12845984e-03 6.86782062e-01 6.65679216e-01 2.38983214e-01
-8.11385751e-01 -4.57645506e-02 -6.08051479e-01 -4.50251818e-01
3.66895586e-01 -6.93239212e-01 2.36341983e-01 8.73340666e-02
-1.29563093e+00 1.00783432e+00 -3.29883359e-02 4.37883288e-01
-1.27382139e-02 -4.53944653e-01 9.63433981e-01 5.78328788e-01
5.54779470e-01 5.13083816e-01 -7.03167975e-01 -3.36141139e-01
-2.21183430e-02 5.92132568e-01 1.72118926e+00 1.34025335e+00
8.32453445e-02 -1.94758505e-01 -7.55375564e-01 1.37210238e+00
5.07200658e-01 3.21262211e-01 2.69227356e-01 -1.11076379e+00
8.21108580e-01 5.66486239e-01 6.18719757e-01 1.65223181e-02
-7.45427966e-01 9.29223001e-01 -2.96105742e-01 1.52871817e-01
6.95665717e-01 -1.00036108e+00 -4.08584893e-01 1.52474689e+00
1.72570705e-01 -8.04564536e-01 8.32967281e-01 5.38507879e-01
1.90012813e+00 6.37645721e-01 5.37604868e-01 -3.89895104e-02
1.92781425e+00 -1.29799700e+00 -1.32499981e+00 -3.91327798e-01
5.95362306e-01 -9.64559793e-01 1.20462871e+00 1.58615075e-02
-1.24355948e+00 -5.61023712e-01 -5.86043119e-01 -4.82339859e-01
-8.96226704e-01 1.33508295e-01 2.03093991e-01 6.49187624e-01
-1.35595798e+00 -2.94421643e-01 -1.84748858e-01 -7.56515563e-01
-6.61985576e-01 4.45579410e-01 -1.06534570e-01 6.64094031e-01
-1.71665001e+00 1.52692521e+00 5.09298801e-01 -4.90186751e-01
-8.00173402e-01 -4.22181159e-01 -7.40310431e-01 -1.25095189e-01
3.61786395e-01 -2.77352124e-01 2.51324654e+00 -4.32357304e-02
-2.19516134e+00 1.08442903e+00 -1.36092469e-01 -4.40071285e-01
5.79129338e-01 -5.19741058e-01 -6.03112519e-01 -7.14954808e-02
-4.95884530e-02 1.12873781e+00 4.25638780e-02 -1.27216518e+00
-7.95262694e-01 7.58061260e-02 4.47617322e-01 7.73479044e-01
-9.75163952e-02 6.07593596e-01 -1.79571137e-01 -1.71454236e-01
-5.45556486e-01 -9.91018772e-01 -3.28633487e-01 -6.08214736e-01
-6.82124555e-01 -7.47675359e-01 1.30260170e+00 -4.76030350e-01
9.45647597e-01 -2.03354645e+00 5.36982305e-02 -8.46729457e-01
2.65225232e-01 1.94620460e-01 3.27674389e-01 1.18234515e+00
6.06008649e-01 -1.58670053e-01 1.54288784e-01 -9.03448224e-01
6.54870272e-01 1.16702542e-01 -5.89124620e-01 -6.01958454e-01
-3.08732003e-01 1.00414443e+00 -7.12514222e-01 -6.73004329e-01
7.61333168e-01 -1.86957598e-01 2.77068876e-02 9.52506483e-01
-1.13574636e+00 1.02100551e+00 -4.60790813e-01 2.47930363e-01
3.39723885e-01 -7.92459100e-02 3.54205698e-01 -1.12618230e-01
-3.66270244e-01 6.42643273e-01 -6.34708285e-01 2.05505419e+00
-6.42191708e-01 7.06116736e-01 5.93475282e-01 1.04734071e-01
9.27827120e-01 1.20860648e+00 1.63808465e-01 -2.54268378e-01
1.44368500e-01 4.44308557e-02 -6.98239133e-02 -5.97108305e-01
1.04350293e+00 3.88527587e-02 -8.37954223e-01 6.00656867e-01
4.88695025e-01 -2.45242670e-01 1.23502761e-02 5.68848550e-01
6.45987868e-01 -4.46460009e-01 4.21920836e-01 -4.81214136e-01
7.39903331e-01 3.06252629e-01 -2.53248990e-01 8.06234241e-01
-4.18205649e-01 4.38156612e-02 3.44702363e-01 -2.53066391e-01
-4.02886391e-01 -1.24203789e+00 -1.48852281e-02 1.91593623e+00
1.60146073e-01 -3.37595820e-01 -7.94106960e-01 -1.14876330e+00
-7.65619203e-02 9.90952671e-01 -3.85271102e-01 3.24192047e-01
-4.26955611e-01 -1.52665764e-01 1.14521778e+00 5.33673882e-01
1.08171070e+00 -1.49248958e+00 -2.91469306e-01 3.12690049e-01
-5.18711209e-01 -1.65797424e+00 -5.17103493e-01 1.39061928e-01
-4.56211746e-01 -6.93979621e-01 -6.00100994e-01 -1.04914665e+00
-2.58884221e-01 -1.19859409e-02 1.44982958e+00 -5.87515235e-01
4.12176043e-01 8.21871519e-01 -3.64661545e-01 -4.01115775e-01
-1.01964462e+00 2.74426341e-01 -1.01339519e-01 -5.14447927e-01
6.74413204e-01 9.75549296e-02 -2.77101517e-01 9.22318876e-01
-2.91289240e-01 4.48239297e-02 1.75596282e-01 8.59017432e-01
-5.19941807e-01 -6.65844858e-01 9.62235868e-01 -1.05539489e+00
1.54206884e+00 -3.02890986e-01 -3.72996956e-01 6.75442815e-01
-7.01392477e-04 -1.19898818e-01 2.33821258e-01 -3.97042245e-01
-2.22174454e+00 1.53432593e-01 -4.66714859e-01 1.60858244e-01
-6.43808484e-01 2.16423839e-01 -3.74783099e-01 2.18489408e-01
9.26984727e-01 -1.28710061e-01 -1.02293514e-01 -7.68987298e-01
9.34753180e-01 1.34265447e+00 7.17468262e-01 -7.83661544e-01
2.30526701e-02 7.56181255e-02 -7.95942485e-01 -9.15982962e-01
-5.20571709e-01 -7.44100809e-01 -9.69664812e-01 -5.39753973e-01
1.56367803e+00 -8.35175037e-01 -1.13230419e+00 5.37103653e-01
-1.56963992e+00 -1.12539542e+00 -5.65831810e-02 1.17008008e-01
-4.99166906e-01 -9.08384006e-03 -1.12885737e+00 -1.18365836e+00
-7.19067335e-01 -1.40803039e+00 1.29429269e+00 1.89062923e-01
-7.14132786e-01 -1.72084677e+00 5.62776089e-01 6.33476198e-01
4.92349386e-01 -1.47958860e-01 6.72797620e-01 -1.55010509e+00
7.14576095e-02 5.25067374e-03 -3.67475003e-02 2.32360568e-02
1.04167134e-01 -5.27354181e-01 -1.44141817e+00 4.03862000e-02
-4.89075556e-02 -1.30345201e+00 6.18228354e-02 1.59148663e-01
1.04684435e-01 -1.88741609e-01 -4.46149588e-01 -4.55789238e-01
4.81633753e-01 7.74799824e-01 4.86372225e-02 -1.16275936e-01
3.47726166e-01 1.10573637e+00 9.97724950e-01 3.70528370e-01
7.66498566e-01 1.16047978e+00 -1.47973299e-01 1.81204170e-01
-4.89633083e-02 -9.70561132e-02 4.64621633e-01 6.32336855e-01
6.07506216e-01 -6.35727823e-01 -1.10767043e+00 3.76800627e-01
-2.24534202e+00 -6.10093474e-01 -3.26959759e-01 1.52333570e+00
1.00445366e+00 -2.13048863e-03 6.60334587e-01 -7.65116870e-01
6.30271852e-01 4.17427838e-01 -4.02384132e-01 -7.60989428e-01
-6.65254518e-02 -2.45129719e-01 -1.59829259e-01 1.10797405e+00
-1.26054907e+00 1.59837151e+00 7.05924988e+00 4.69337195e-01
-4.77415204e-01 5.53647459e-01 4.26933348e-01 4.39865172e-01
4.03700471e-01 7.11385608e-02 -1.29878390e+00 9.04768333e-02
1.24673557e+00 -2.82348514e-01 1.52090907e-01 1.09531701e+00
-4.64941263e-02 -1.89837500e-01 -1.45092106e+00 4.19357836e-01
-2.51755863e-01 -1.10982335e+00 -1.79982901e-01 -1.70761719e-01
2.78043777e-01 4.62839827e-02 -1.72989804e-03 9.26205695e-01
1.21773529e+00 -8.37199509e-01 3.79872382e-01 -6.43783389e-03
8.64071727e-01 6.61438555e-02 7.16554463e-01 3.21197271e-01
-1.19356847e+00 1.21428959e-01 -3.70422006e-02 1.02672204e-01
8.55300128e-01 -3.91483963e-01 -1.48613131e+00 4.16777015e-01
3.78296047e-01 2.40164548e-01 -1.16258152e-01 2.49653056e-01
-3.79178002e-02 1.13481693e-02 -3.92882377e-02 -3.21827412e-01
3.30882370e-01 2.32492462e-02 5.96002221e-01 1.69488776e+00
-3.93953323e-01 3.18585396e-01 6.94180727e-01 4.64132816e-01
-8.09519216e-02 -1.08126014e-01 -7.05291033e-01 1.80070281e-01
9.50317442e-01 1.28397071e+00 -2.67002672e-01 -5.97003937e-01
-3.43676805e-01 9.47786331e-01 9.85292569e-02 1.73361406e-01
-3.23296726e-01 -1.92687929e-01 5.42411029e-01 -6.43340588e-01
-4.43592340e-01 -2.56401360e-01 -3.70274752e-01 -7.15784431e-01
-6.14741325e-01 -9.70847666e-01 1.00159276e+00 -6.35773897e-01
-1.72547913e+00 1.09062409e+00 6.28108561e-01 -8.10666084e-01
-9.64427292e-01 -7.21064150e-01 -8.86201322e-01 9.18961167e-01
-5.81411839e-01 -1.23642504e+00 -4.39427197e-01 8.14627647e-01
1.44314122e+00 -4.47032839e-01 1.38610446e+00 1.02581166e-01
-1.21800356e-01 5.99292397e-01 6.16445541e-02 3.48436117e-01
1.43410611e+00 -1.81310856e+00 8.07580233e-01 -2.43716940e-01
-4.56234068e-01 3.39374959e-01 7.23515630e-01 -8.18172455e-01
-9.89180744e-01 -6.86370134e-01 8.62400889e-01 -1.16371834e+00
8.17704141e-01 -9.11949158e-01 -6.27576530e-01 1.09585702e+00
1.25165617e+00 -1.07908201e+00 5.75409353e-01 4.19626564e-01
-1.50900528e-01 4.56523150e-01 -1.41448474e+00 8.81167412e-01
7.26207554e-01 -7.27313876e-01 -8.72718036e-01 9.05777574e-01
1.19129789e+00 -9.47482109e-01 -1.12892342e+00 2.53983408e-01
5.20515680e-01 -6.52798772e-01 1.01454866e+00 -6.67336524e-01
2.92485170e-02 3.72932881e-01 -2.55965054e-01 -1.18947601e+00
3.26849461e-01 -1.06812596e+00 -1.56343699e-01 1.40301633e+00
8.24748695e-01 -4.55412686e-01 5.20299911e-01 1.19901943e+00
-5.58143556e-01 6.14876710e-02 -9.08213317e-01 -6.27887726e-01
5.04210711e-01 1.82823313e-03 1.97583422e-01 8.17679882e-01
1.01616871e+00 1.37741148e+00 -6.92706585e-01 -2.53098160e-01
8.85853469e-02 -2.57371962e-01 1.20646453e+00 -1.34175301e+00
-1.87659651e-01 -3.26938689e-01 4.03392285e-01 -1.67328835e+00
2.56187767e-01 -2.69886613e-01 3.12985331e-01 -1.55577505e+00
-3.07905674e-01 -2.42625296e-01 5.76636016e-01 2.59319693e-01
-4.57412079e-02 -6.31497979e-01 3.01268429e-01 2.57080287e-01
-1.00182164e+00 7.14992762e-01 8.93980920e-01 -3.68107930e-02
-8.46031964e-01 3.87319803e-01 -3.98666680e-01 5.80253482e-01
9.72322822e-01 4.74953093e-02 -8.07705998e-01 -2.99592406e-01
-6.99254036e-01 7.03830957e-01 -1.17242083e-01 -6.62621558e-01
7.16937304e-01 -4.10788096e-02 -3.44702363e-01 -9.25815761e-01
1.27915573e+00 -6.31334126e-01 -2.85835385e-01 2.73406506e-01
-1.01938212e+00 6.19150251e-02 5.66778064e-01 1.68583870e-01
-1.31730616e-01 -3.11409503e-01 7.25377142e-01 -2.73879647e-01
-8.80522311e-01 -4.24922824e-01 -9.69018459e-01 3.52864861e-01
8.70437026e-01 3.00551414e-01 -1.16799319e+00 -9.78648245e-01
-1.15822637e+00 8.68165791e-01 -1.64318427e-01 7.02896833e-01
2.96573490e-01 -9.59796727e-01 -6.96250498e-01 -6.94397092e-01
3.03824455e-01 -3.73571306e-01 3.75037402e-01 5.30519903e-01
-6.94959387e-02 1.08545280e+00 -2.41314337e-01 -4.69958812e-01
-1.59861612e+00 1.17862128e-01 4.10836250e-01 -7.39782810e-01
-3.67904991e-01 7.21551895e-01 4.18720216e-01 -1.00688899e+00
8.60183716e-01 4.47791480e-02 -7.40983009e-01 1.97844192e-01
6.05705142e-01 2.80530304e-01 -8.69436469e-03 -3.66553336e-01
-2.48730257e-01 -3.68099421e-01 -6.45319164e-01 -1.22926867e+00
3.52037460e-01 -4.28546399e-01 1.73982263e-01 9.77243960e-01
6.44603431e-01 -4.38321859e-01 -8.97395790e-01 -5.34594357e-01
6.92844331e-01 2.33464926e-01 -4.41740155e-01 -1.70052588e+00
-8.40703472e-02 8.23231757e-01 5.66115379e-01 6.94405258e-01
4.95629668e-01 2.92968959e-01 9.03031826e-01 1.10570347e+00
2.67121673e-01 -1.28128266e+00 3.71653050e-01 1.40747821e+00
1.27305913e+00 -1.56328273e+00 -5.92689872e-01 -4.46307391e-01
-1.71051598e+00 9.38314795e-01 1.30379581e+00 6.89839423e-01
8.17523479e-01 2.99657345e-01 8.26145470e-01 -5.51851630e-01
-1.09799254e+00 -3.08887005e-01 1.76451519e-01 7.27030337e-01
6.54949725e-01 -1.57034531e-01 1.05025716e-01 8.19111168e-01
-1.97071865e-01 -5.14628947e-01 8.88564736e-02 1.21849632e+00
-5.37818849e-01 -1.23544419e+00 -2.18667746e-01 -2.33026370e-02
-1.84505746e-01 -1.81148812e-01 -1.77240539e+00 1.08020318e+00
-7.78986871e-01 1.81916678e+00 -2.40598291e-01 -7.51282990e-01
7.86813319e-01 7.05125391e-01 -1.99432045e-01 -9.32131290e-01
-1.45730436e+00 6.14823811e-02 1.19955492e+00 -2.88186185e-02
-2.12922573e-01 -3.47817570e-01 -1.16445601e+00 -3.06825846e-01
-2.89966911e-01 6.41594589e-01 4.33275193e-01 7.78738141e-01
3.80373567e-01 1.07198812e-01 3.67599845e-01 -8.40082824e-01
-6.25017881e-01 -1.80363584e+00 -1.59897059e-01 6.16155565e-02
5.69990724e-02 -6.72132850e-01 -2.26697810e-02 1.17354497e-01] | [12.82015609741211, 7.997310638427734] |
063f2e4c-1290-4957-abdd-775eac2c1d71 | prompt-engineering-for-healthcare | 2304.14670 | null | https://arxiv.org/abs/2304.14670v1 | https://arxiv.org/pdf/2304.14670v1.pdf | Prompt Engineering for Healthcare: Methodologies and Applications | This review will introduce the latest advances in prompt engineering in the field of natural language processing (NLP) for the medical domain. First, we will provide a brief overview of the development of prompt engineering and emphasize its significant contributions to healthcare NLP applications such as question-answering systems, text summarization, and machine translation. With the continuous improvement of general large language models, the importance of prompt engineering in the healthcare domain is becoming increasingly prominent. The aim of this article is to provide useful resources and bridges for healthcare NLP researchers to better explore the application of prompt engineering in this field. We hope that this review can provide new ideas and inspire ample possibilities for research and application in medical NLP. | ['Shu Zhang', 'Tianming Liu', 'Dinggang Shen', 'Yixuan Yuan', 'Dajiang Zhu', 'Bao Ge', 'Xiang Li', 'Yiheng Liu', 'Haiyang Zhang', 'Chenxi Yue', 'Huawen Hu', 'Jinru Wu', 'Yanqing Kang', 'Qiushi Yang', 'Haixing Dai', 'Chong Ma', 'Zihao Wu', 'Sigang Yu', 'Enze Shi', 'Jiaqi Wang'] | 2023-04-28 | null | null | null | null | ['prompt-engineering', 'text-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.07062006e-01 6.87376559e-01 -5.37641227e-01 -3.62281591e-01
-1.08412218e+00 -2.49683216e-01 3.49394143e-01 1.16384649e+00
-4.33765292e-01 9.45435762e-01 9.86489236e-01 -3.30256492e-01
-4.70883399e-01 -5.31572580e-01 -1.16880432e-01 -4.08036888e-01
5.34117483e-02 6.26767218e-01 -2.25127473e-01 -2.95652121e-01
3.84141892e-01 4.42398399e-01 -9.11898851e-01 5.83048403e-01
1.11223054e+00 2.54690051e-01 2.11984411e-01 8.94611359e-01
-5.44980049e-01 1.15969074e+00 -7.45208502e-01 -2.39476234e-01
-5.54111063e-01 -7.20873535e-01 -1.35885739e+00 -2.86908090e-01
4.45580408e-02 -9.56151038e-02 -1.32046178e-01 7.81245828e-01
7.94938207e-01 1.08920380e-01 5.06557226e-01 -6.85457468e-01
-5.11255741e-01 4.09942418e-01 8.96384474e-03 4.45776910e-01
1.21349883e+00 -1.71564415e-01 7.83460617e-01 -5.01430213e-01
7.96113789e-01 1.02816153e+00 4.94745761e-01 9.17340457e-01
-5.57701886e-01 1.34608880e-01 1.70650065e-01 3.89314115e-01
-8.73801053e-01 -5.04034162e-01 6.22243106e-01 -9.02877897e-02
1.35175729e+00 4.72904503e-01 7.37986445e-01 9.41446781e-01
7.36261547e-01 1.21130538e+00 5.60934007e-01 -7.80006230e-01
2.10456997e-01 2.76110489e-02 7.07033157e-01 6.15283310e-01
2.44440138e-01 -1.26745477e-01 -8.27370524e-01 -5.31785011e-01
3.53664368e-01 -1.86950877e-01 -3.47901344e-01 4.12462026e-01
-1.26812768e+00 9.52458739e-01 -2.52564460e-01 5.51159203e-01
-9.02544379e-01 -3.63452137e-01 5.24821460e-01 1.05684705e-01
3.64181399e-01 8.87437522e-01 -6.14906311e-01 -5.94304860e-01
-9.62607980e-01 5.16125023e-01 1.29601908e+00 8.59088600e-01
-9.96836126e-02 -2.51646549e-01 -4.50508446e-01 9.50051665e-01
1.79451391e-01 6.44870222e-01 4.85551625e-01 -1.07855678e+00
1.49715737e-01 4.23995376e-01 1.15356140e-01 -9.16256249e-01
-7.76362538e-01 -8.80392641e-02 -7.62852371e-01 -8.95624220e-01
-4.82389852e-02 -4.74445373e-01 -5.81559360e-01 1.23486030e+00
2.44647041e-01 -1.85019702e-01 4.60642129e-01 4.44212079e-01
1.48152816e+00 8.41579139e-01 4.82133240e-01 -1.06927085e+00
1.89442706e+00 -8.97615671e-01 -1.40480602e+00 -3.57414484e-01
7.01586723e-01 -1.04443741e+00 5.53449333e-01 2.57878244e-01
-1.22131467e+00 2.65530236e-02 -5.11429191e-01 -3.40061814e-01
-1.20893441e-01 -4.45001125e-02 7.20618069e-01 2.37432420e-01
-7.93066859e-01 2.20771819e-01 -9.29325461e-01 -1.09171605e+00
2.51521498e-01 1.29469857e-01 -3.22008096e-02 -1.64894566e-01
-1.51436973e+00 1.23894191e+00 2.61421591e-01 -1.57752469e-01
-7.30233192e-02 -1.03667653e+00 -9.34628665e-01 -1.74683034e-01
3.95491034e-01 -1.37060177e+00 1.77950871e+00 -2.64928162e-01
-1.49300015e+00 8.77026856e-01 -5.65823138e-01 -7.03044772e-01
-1.25378398e-02 -5.33222735e-01 -5.41603148e-01 7.10079014e-01
9.33139995e-02 4.04318750e-01 2.84481049e-02 -5.90663731e-01
-7.53085136e-01 -1.71827406e-01 -3.36274743e-01 4.03841019e-01
-1.03180604e-02 6.16545856e-01 -1.34441227e-01 -4.80680138e-01
-3.75614911e-01 -3.72903019e-01 -5.76519191e-01 -3.64908665e-01
-4.48044449e-01 -6.35974705e-01 2.16035377e-02 -7.29295969e-01
1.60087168e+00 -1.71753061e+00 -3.94342512e-01 -2.94289380e-01
1.58252776e-01 4.73325133e-01 -1.54254138e-01 1.27227426e+00
-4.36634310e-02 -1.18153661e-01 -2.79933959e-01 2.53725052e-01
-4.20180321e-01 4.86388624e-01 -6.48072958e-01 9.35174599e-02
2.41957709e-01 1.34404433e+00 -1.34319019e+00 -8.92770171e-01
2.88882107e-01 4.15803879e-01 -2.30073646e-01 1.75112411e-01
-3.07135701e-01 3.60606760e-01 -9.41247046e-01 4.68578041e-01
-1.19826660e-01 -4.09248620e-01 9.01652947e-02 2.19573915e-01
-2.09358379e-01 7.25180745e-01 -3.22280884e-01 1.58197343e+00
-7.44449794e-02 2.97715813e-01 -1.05480134e-01 -9.11581635e-01
6.79004133e-01 8.59475613e-01 8.92613530e-01 -6.30545139e-01
1.81209389e-02 1.03217825e-01 -2.37684473e-01 -1.16036749e+00
6.52670443e-01 -4.63422805e-01 -2.52161305e-02 4.47315246e-01
-2.87739605e-01 -2.32699633e-01 4.08002526e-01 5.76377153e-01
1.07812035e+00 -5.37175596e-01 1.29794490e+00 -2.62796789e-01
7.93185174e-01 5.11566818e-01 3.48748893e-01 9.17576611e-01
-2.41128191e-01 2.16295153e-01 4.12635863e-01 -5.41484416e-01
-4.63594407e-01 -7.31350660e-01 -2.91512281e-01 6.97540641e-01
-2.55964577e-01 -7.81059504e-01 -6.44276321e-01 -6.40718162e-01
-2.62604147e-01 1.04747272e+00 -3.27066153e-01 -2.79056467e-02
-8.73040557e-01 -1.04531908e+00 6.33841038e-01 4.50305432e-01
-5.59881106e-02 -1.48312759e+00 -8.52294743e-01 5.06703198e-01
-8.35328698e-01 -1.09261680e+00 -2.58336961e-01 -1.60109758e-01
-1.32163310e+00 -1.30170918e+00 -9.59437430e-01 -1.09210265e+00
6.67682111e-01 -4.21762392e-02 1.10962152e+00 -1.16928359e-02
-3.38684708e-01 1.18880105e+00 -6.05449438e-01 -8.88946652e-01
-7.45562971e-01 1.75244406e-01 5.15515395e-02 -6.55191481e-01
8.42740119e-01 -1.44973099e-01 -6.85811639e-01 -5.23642302e-01
-9.98136699e-01 -5.32489680e-02 4.14564133e-01 7.85359442e-01
6.14759982e-01 -3.02692294e-01 1.05053461e+00 -1.35828924e+00
1.53004789e+00 -4.51693505e-01 1.27790183e-01 5.95173895e-01
-8.53151321e-01 6.31988719e-02 3.35826635e-01 -5.93850203e-02
-1.11378360e+00 -1.96379125e-01 -7.80172944e-01 8.41908514e-01
-4.04169858e-01 1.01290238e+00 2.85988778e-01 3.58582079e-01
9.58821774e-01 4.13015932e-01 8.12107697e-02 -5.84184051e-01
6.21796846e-01 7.09089816e-01 5.18481672e-01 -3.54895771e-01
-4.69195843e-02 1.63335919e-01 -1.16762012e-01 -1.22835922e+00
-1.00400329e+00 -8.59700263e-01 -4.52814549e-02 -1.42257381e-02
8.93853486e-01 -3.26062769e-01 -5.93976200e-01 1.25880297e-02
-1.52236843e+00 -4.49376144e-02 -8.62076461e-01 4.73767221e-01
-6.14481926e-01 9.96065915e-01 -9.56878245e-01 -6.60739362e-01
-1.14543700e+00 -6.88432217e-01 1.12270367e+00 4.86490756e-01
-9.39877808e-01 -1.27451730e+00 4.96665090e-01 3.69459897e-01
1.86065838e-01 1.25586595e-02 1.42169869e+00 -1.19894552e+00
6.19492419e-02 -4.76711392e-01 1.80836767e-01 -9.30283684e-03
4.59257632e-01 -2.99800992e-01 -4.05205548e-01 2.86772281e-01
5.76949239e-01 -1.60635456e-01 4.76678550e-01 6.48073018e-01
6.10913873e-01 -6.86264277e-01 -6.59104228e-01 -2.29263812e-01
1.11734211e+00 3.69737118e-01 5.24330020e-01 6.51850924e-02
1.60936639e-01 1.12364256e+00 8.72132182e-01 3.68692935e-01
5.54066658e-01 2.42226988e-01 -4.30921167e-01 -3.39912660e-02
-1.17639527e-01 -2.33219922e-01 -9.24234651e-03 1.42979419e+00
2.43461072e-01 -1.51209563e-01 -1.01798117e+00 4.27412808e-01
-1.96146011e+00 -6.42841578e-01 -2.08734453e-01 1.68297219e+00
1.28623879e+00 -4.53201920e-01 -2.38499597e-01 -1.43817157e-01
3.32153380e-01 1.13994129e-01 -2.27505162e-01 -7.13831961e-01
9.30954888e-03 2.51641870e-01 -4.45958935e-02 8.55940163e-01
-7.47945726e-01 8.44204605e-01 7.69405270e+00 5.04921973e-01
-7.43358076e-01 -1.80842713e-01 3.83693755e-01 8.07695910e-02
-2.13216364e-01 -3.84470135e-01 -6.84919357e-01 1.24493390e-01
1.17726135e+00 -7.53237128e-01 1.46441646e-02 5.53140938e-01
7.72118092e-01 -3.77097607e-01 -1.12478125e+00 1.09081948e+00
2.79833049e-01 -1.57733130e+00 2.27641657e-01 -3.21524173e-01
5.38770854e-01 -1.22517139e-01 -4.63012487e-01 6.65857792e-02
4.30757031e-02 -6.90432310e-01 -1.44823730e-01 7.85588503e-01
4.70971704e-01 -3.32262754e-01 9.48838472e-01 7.11509764e-01
-6.18568897e-01 1.53126568e-01 -2.80436069e-01 -1.04594260e-01
9.76814985e-01 9.96917725e-01 -1.03850210e+00 1.03895724e+00
1.19165570e-01 6.75218463e-01 4.90382053e-02 1.24118018e+00
-4.69213635e-01 4.50587302e-01 -1.09611906e-01 -3.30692679e-01
-6.06170185e-02 1.06236786e-01 7.42442250e-01 1.47292471e+00
-4.73254062e-02 8.45888555e-01 2.43317083e-01 1.55595407e-01
1.66592047e-01 7.66988575e-01 -5.41127443e-01 -4.91892666e-01
2.62063384e-01 7.67695308e-01 -5.63344657e-01 -7.22397685e-01
-3.79359692e-01 7.68976390e-01 -8.87770057e-02 2.35043794e-01
-2.93907046e-01 -3.04724932e-01 2.12938294e-01 1.54465791e-02
-4.16725487e-01 2.64356378e-02 -4.91327554e-01 -9.40679669e-01
-2.61342198e-01 -1.18979144e+00 8.57819438e-01 -5.88591874e-01
-1.26966524e+00 5.62395453e-01 2.41524011e-01 -6.46397829e-01
-6.37588382e-01 -3.38123143e-01 -2.78851360e-01 6.51533008e-01
-1.49242401e+00 -4.92469996e-01 2.55589157e-01 2.23240793e-01
7.53494143e-01 1.20469637e-01 1.31899059e+00 -7.46922893e-03
-2.53831178e-01 1.65510431e-01 -1.65249050e-01 -2.40055144e-01
8.52258861e-01 -1.03121078e+00 1.97345808e-01 5.49865484e-01
4.79752384e-02 1.06642163e+00 9.36578631e-01 -9.84126806e-01
-1.35790372e+00 -8.84943306e-01 2.04270458e+00 -7.74346411e-01
3.49017888e-01 4.36334968e-01 -7.62750149e-01 3.86268288e-01
4.54730898e-01 -8.57349277e-01 1.22751939e+00 1.47177093e-02
3.70022058e-01 4.16911989e-02 -1.18611467e+00 6.15670264e-01
4.89998132e-01 -3.79050672e-01 -1.46563959e+00 9.26270366e-01
8.41740727e-01 -4.10098404e-01 -9.53252435e-01 5.91487110e-01
3.46405536e-01 -2.26515263e-01 8.19734991e-01 -8.85428786e-01
2.98058569e-01 6.21836856e-02 4.86470670e-01 -9.42706883e-01
-2.09316403e-01 -8.58135939e-01 -2.70946383e-01 7.39263177e-01
3.34141791e-01 -7.11560011e-01 7.00326204e-01 9.51852143e-01
-1.36513695e-01 -1.31527960e+00 -7.45648444e-01 -1.81239977e-01
1.66868150e-01 -4.63031322e-01 1.57739192e-01 7.55449712e-01
9.79799628e-01 7.73216426e-01 -2.19337329e-01 -3.58588487e-01
1.52364612e-01 -9.16768089e-02 2.66326010e-01 -1.21929884e+00
1.88461579e-02 -2.58245468e-01 1.99995220e-01 -1.43489337e+00
-2.15406284e-01 -7.80933082e-01 2.33310461e-01 -2.46407223e+00
2.90640861e-01 4.76252198e-01 -3.12516093e-02 6.30535364e-01
-3.67868721e-01 -2.40184650e-01 -1.09961346e-01 1.47618815e-01
-6.69804811e-01 3.42260063e-01 1.32456017e+00 -1.05677910e-01
-4.44540381e-01 2.64441311e-01 -1.23750448e+00 6.25079572e-01
8.49539459e-01 -6.14545465e-01 -4.79488671e-01 -2.45143652e-01
5.12894630e-01 4.70477611e-01 -3.04739088e-01 -2.07623705e-01
6.32578790e-01 -2.07666308e-01 -1.02655768e-01 -6.74293280e-01
-1.05218209e-01 -5.67031384e-01 -4.67283875e-01 7.68792093e-01
-6.29383028e-01 2.91133255e-01 2.92157590e-01 4.70752954e-01
-3.39058310e-01 -4.27427292e-01 3.98264676e-01 -4.47022438e-01
-4.37063158e-01 3.65250092e-03 -1.00980234e+00 2.90411234e-01
7.64485538e-01 7.56455213e-02 -3.13093841e-01 -4.96778369e-01
-6.93643212e-01 6.60725176e-01 7.77579425e-03 5.16110599e-01
8.85849535e-01 -7.30433285e-01 -8.27682793e-01 -3.12864393e-01
1.85626373e-01 -1.12633362e-01 4.84693825e-01 1.05165148e+00
-6.75304234e-01 1.26017153e+00 3.22551906e-01 -1.77230433e-01
-1.36574352e+00 7.36790538e-01 6.40543476e-02 -6.51650071e-01
-1.11665249e+00 5.76698422e-01 1.35688722e-01 -9.03404132e-03
3.75992030e-01 -3.50413024e-01 -6.23891354e-01 -5.96666746e-02
1.16717291e+00 4.25887972e-01 7.31381997e-02 -9.02353674e-02
-6.75905764e-01 2.13992938e-01 -3.85168403e-01 -8.33745450e-02
1.21043622e+00 -1.73844367e-01 -6.82422161e-01 3.53780091e-01
6.62044823e-01 7.43969977e-02 -1.14806928e-01 -1.97624192e-01
4.35558140e-01 3.61689597e-01 -3.56127381e-01 -1.23521197e+00
-1.57711759e-01 6.51221156e-01 8.41922238e-02 1.67084306e-01
1.15479434e+00 3.29330653e-01 1.12746024e+00 5.71677923e-01
1.66796014e-01 -1.19670033e+00 -8.09339061e-02 6.88774943e-01
1.05747175e+00 -7.56683648e-01 1.67394847e-01 -4.38262731e-01
-7.20657885e-01 1.23234296e+00 -1.71877176e-01 2.61671215e-01
6.54516697e-01 2.82271415e-01 4.32200134e-01 -5.87585986e-01
-8.39450657e-01 9.33998451e-02 3.52986574e-01 6.60082459e-01
8.39549482e-01 -6.84493929e-02 -1.25390828e+00 7.52915144e-01
-2.36241728e-01 5.01417458e-01 5.49640357e-01 1.03560507e+00
-6.38496339e-01 -1.62460303e+00 -3.54006827e-01 7.05315351e-01
-7.41900682e-01 -4.50877517e-01 -6.51375055e-01 3.22216809e-01
-3.81754756e-01 1.36367786e+00 -4.94053334e-01 4.26778018e-01
5.48112452e-01 4.89114732e-01 5.51732302e-01 -1.07845056e+00
-6.40857160e-01 -2.63360441e-02 5.82265556e-01 -5.02214789e-01
-6.33488834e-01 -7.59079576e-01 -1.43790221e+00 -1.35750808e-02
3.09932940e-02 7.70259738e-01 3.37593615e-01 1.11009908e+00
4.16481137e-01 5.55571139e-01 -2.28768468e-01 2.49260858e-01
-4.62931663e-01 -9.95680511e-01 -2.66354859e-01 -1.75300881e-01
4.00253922e-01 1.16548553e-01 1.51986718e-01 2.25809857e-01] | [12.234389305114746, 9.436200141906738] |
c312eb98-2e09-4e30-8cd9-dc05f784f975 | fame-vil-multi-tasking-vision-language-model | 2303.02483 | null | https://arxiv.org/abs/2303.02483v1 | https://arxiv.org/pdf/2303.02483v1.pdf | FAME-ViL: Multi-Tasking Vision-Language Model for Heterogeneous Fashion Tasks | In the fashion domain, there exists a variety of vision-and-language (V+L) tasks, including cross-modal retrieval, text-guided image retrieval, multi-modal classification, and image captioning. They differ drastically in each individual input/output format and dataset size. It has been common to design a task-specific model and fine-tune it independently from a pre-trained V+L model (e.g., CLIP). This results in parameter inefficiency and inability to exploit inter-task relatedness. To address such issues, we propose a novel FAshion-focused Multi-task Efficient learning method for Vision-and-Language tasks (FAME-ViL) in this work. Compared with existing approaches, FAME-ViL applies a single model for multiple heterogeneous fashion tasks, therefore being much more parameter-efficient. It is enabled by two novel components: (1) a task-versatile architecture with cross-attention adapters and task-specific adapters integrated into a unified V+L model, and (2) a stable and effective multi-task training strategy that supports learning from heterogeneous data and prevents negative transfer. Extensive experiments on four fashion tasks show that our FAME-ViL can save 61.5% of parameters over alternatives, while significantly outperforming the conventional independently trained single-task models. Code is available at https://github.com/BrandonHanx/FAME-ViL. | ['Tao Xiang', 'Yi-Zhe Song', 'Li Zhang', 'Licheng Yu', 'Xiatian Zhu', 'Xiao Han'] | 2023-03-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Han_FAME-ViL_Multi-Tasking_Vision-Language_Model_for_Heterogeneous_Fashion_Tasks_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Han_FAME-ViL_Multi-Tasking_Vision-Language_Model_for_Heterogeneous_Fashion_Tasks_CVPR_2023_paper.pdf | cvpr-2023-1 | ['multi-modal-classification'] | ['miscellaneous'] | [ 8.63936618e-02 -4.41314071e-01 -3.36513668e-01 -2.57137299e-01
-1.22252488e+00 -7.44448721e-01 6.09788477e-01 -2.17242897e-01
-4.30234104e-01 4.37897563e-01 2.03411326e-01 -1.14805900e-01
4.15728800e-02 -3.48120570e-01 -8.25554311e-01 -5.14086187e-01
4.38515037e-01 5.49468279e-01 1.63897842e-01 -2.26442859e-01
5.92193939e-02 1.09258644e-01 -1.57311201e+00 6.42926216e-01
7.38717020e-01 9.96779621e-01 6.52623653e-01 6.62348986e-01
-1.85344547e-01 6.33484602e-01 -3.79781514e-01 -4.73904043e-01
1.72322243e-01 -5.61374724e-02 -7.53328562e-01 -8.02835391e-04
6.53522789e-01 -9.25813690e-02 -1.81871355e-01 5.37906170e-01
8.69438887e-01 9.01833326e-02 7.34543860e-01 -1.35820103e+00
-1.14101613e+00 1.49612457e-01 -6.11061454e-01 -1.89821333e-01
1.83553517e-01 3.39762062e-01 7.82428324e-01 -1.09127450e+00
5.09183943e-01 1.26361454e+00 5.62234938e-01 7.75984645e-01
-1.34989488e+00 -6.78636134e-01 2.79749423e-01 -8.88608098e-02
-1.46699131e+00 -5.68022251e-01 7.77409017e-01 -4.78907794e-01
9.22157228e-01 2.95685112e-01 3.70153099e-01 1.44594026e+00
2.90313333e-01 1.21798491e+00 1.07141149e+00 -3.84232044e-01
-1.84851959e-01 3.99185956e-01 3.66575904e-02 7.04373956e-01
-1.39470771e-01 -1.45174861e-01 -5.89141726e-01 -8.84171054e-02
8.52408588e-01 6.27173707e-02 -1.42865732e-01 -5.66051185e-01
-1.36550999e+00 8.04323852e-01 3.72056454e-01 2.70060450e-01
-2.49227747e-01 2.61165351e-01 8.47251236e-01 5.81537127e-01
6.02858484e-01 3.87420863e-01 -6.62926495e-01 1.53668135e-01
-8.09338570e-01 2.73217529e-01 6.42824888e-01 1.23526597e+00
6.67892337e-01 1.64149478e-02 -6.81033611e-01 1.18162608e+00
1.42707005e-01 6.14044964e-01 6.35063469e-01 -7.42863536e-01
4.58136708e-01 4.28001314e-01 -6.29043728e-02 -6.72039449e-01
-2.36602828e-01 -4.81773257e-01 -9.05221522e-01 -1.36829510e-01
2.44235232e-01 1.15464561e-01 -1.13287270e+00 1.99640036e+00
7.54100233e-02 -1.64988682e-01 -4.19104099e-02 8.94156575e-01
1.10097992e+00 6.15796030e-01 4.00616616e-01 1.87795058e-01
1.53000224e+00 -1.49825895e+00 -6.07307971e-01 -5.94677329e-01
5.98805606e-01 -1.09954441e+00 1.64419496e+00 1.87544152e-01
-1.11362183e+00 -8.79449368e-01 -6.80028856e-01 -4.90260392e-01
-7.42137313e-01 4.17494178e-01 7.10665286e-01 2.54826367e-01
-1.32322276e+00 -1.82580873e-01 -1.31693467e-01 -4.94548321e-01
2.70290881e-01 3.90061975e-01 -4.98705328e-01 -1.89594418e-01
-1.08828163e+00 7.85247684e-01 3.42470735e-01 -6.64742440e-02
-8.49418521e-01 -7.28175640e-01 -8.83298099e-01 -3.61837931e-02
4.04024035e-01 -1.30797803e+00 1.21764529e+00 -1.17294157e+00
-1.40064991e+00 1.22626817e+00 -1.28684297e-01 -3.73782367e-02
4.06456739e-01 -3.84610862e-01 -2.02802956e-01 7.55114928e-02
1.54806152e-01 1.16120708e+00 1.27344263e+00 -1.48012233e+00
-2.21887201e-01 -2.19424099e-01 5.50661944e-02 3.38138789e-01
-4.16864485e-01 1.20970190e-01 -1.24590671e+00 -8.44338536e-01
-4.89693642e-01 -1.12277329e+00 -2.22188830e-02 1.03201628e-01
-2.70759910e-01 -3.71925801e-01 8.58119726e-01 -5.38122892e-01
1.15753543e+00 -2.26466131e+00 2.88367689e-01 -2.81098098e-01
2.59840697e-01 4.81909335e-01 -7.40485549e-01 4.83202547e-01
3.49187180e-02 1.33126732e-02 3.37805226e-02 -6.39327109e-01
-4.14514355e-02 2.12655038e-01 -1.08252928e-01 2.09817305e-01
2.24416137e-01 1.41184914e+00 -6.74287617e-01 -6.89636290e-01
3.37455332e-01 5.79667509e-01 -4.49886203e-01 3.51895362e-01
-3.88532370e-01 4.48891133e-01 -4.95687932e-01 9.03260708e-01
4.85067278e-01 -5.56939900e-01 -1.15255281e-01 -5.42418718e-01
3.62879559e-02 -2.68815249e-01 -7.78523088e-01 2.02826309e+00
-1.03280795e+00 4.54842776e-01 2.97861516e-01 -8.49666476e-01
8.09310734e-01 3.57876599e-01 3.41636062e-01 -9.82593179e-01
1.69312641e-01 1.12684913e-01 -6.22622430e-01 -7.32475758e-01
6.49194658e-01 1.39795810e-01 -4.14614260e-01 2.74660319e-01
3.03460330e-01 -9.00384188e-02 -7.07205459e-02 2.46982902e-01
8.69226098e-01 2.64628828e-01 4.24419679e-02 -1.54725119e-01
4.85086173e-01 -9.51165408e-02 1.57445103e-01 8.27174366e-01
-1.66751280e-01 7.08308756e-01 -4.84186932e-02 -3.30241382e-01
-1.08902800e+00 -1.05399704e+00 2.34860312e-02 1.64377248e+00
1.66121244e-01 -4.73132521e-01 -2.68032312e-01 -5.85212648e-01
1.91071227e-01 5.28565109e-01 -5.85581422e-01 -3.20569247e-01
-4.33781564e-01 -4.63451743e-01 3.78727585e-01 4.37544584e-01
4.65456486e-01 -1.06105506e+00 -2.32679263e-01 7.18929023e-02
-3.57388884e-01 -1.22131085e+00 -9.56359267e-01 1.18542686e-01
-6.86962903e-01 -6.71876252e-01 -1.00769854e+00 -1.07642663e+00
4.79960412e-01 6.33346260e-01 1.42676246e+00 -2.19273299e-01
-2.18678355e-01 8.82418692e-01 -2.50241816e-01 -3.47912371e-01
-1.42958671e-01 2.87880540e-01 -2.17657581e-01 7.05149472e-02
3.79439890e-01 -2.95857668e-01 -6.76483750e-01 3.49172056e-01
-1.04390621e+00 3.50197524e-01 9.67081368e-01 1.09177732e+00
6.49289548e-01 -6.23093724e-01 8.14037144e-01 -7.52476394e-01
8.04292440e-01 -5.44904113e-01 -2.32889205e-01 5.67445993e-01
-5.08317530e-01 -8.11628923e-02 6.13679647e-01 -8.89027297e-01
-9.62221503e-01 1.77821875e-01 2.87030395e-02 -9.06661928e-01
-7.43187889e-02 3.51281643e-01 -1.00355662e-01 -1.79658905e-01
4.59466666e-01 4.17098701e-01 1.07841842e-01 -6.20016634e-01
7.43899047e-01 9.21090007e-01 5.38857698e-01 -6.97200477e-01
5.29413998e-01 2.07553923e-01 -3.52545589e-01 -7.40120828e-01
-8.69897902e-01 -6.86632752e-01 -4.37258512e-01 -3.43199402e-01
9.51936245e-01 -1.40973914e+00 -6.73716903e-01 5.25600374e-01
-1.17820108e+00 -5.35057724e-01 -8.75933692e-02 2.48131707e-01
-5.97132802e-01 1.20925881e-01 -5.80028713e-01 -5.74031055e-01
-6.73456609e-01 -1.19223845e+00 1.56094456e+00 4.48192656e-02
-1.60219684e-01 -1.03682470e+00 -1.00817032e-01 8.34046543e-01
7.48171866e-01 -1.36540160e-01 8.41216445e-01 -5.10105669e-01
-6.30978346e-01 -1.11393325e-01 -5.11235058e-01 4.08585876e-01
-1.12956017e-01 -3.92353237e-01 -1.07587647e+00 -5.36069214e-01
-2.24977598e-01 -8.53732347e-01 1.01567090e+00 4.44600254e-01
1.37881660e+00 -1.25249654e-01 -4.08928782e-01 7.58498192e-01
1.53155601e+00 -8.84315521e-02 5.06695151e-01 3.66501540e-01
8.25914085e-01 4.85162884e-01 5.34132421e-01 7.33985230e-02
6.67925000e-01 8.49833250e-01 2.47370914e-01 -4.26896065e-01
-3.90268564e-01 -2.11930692e-01 4.50268924e-01 8.81261170e-01
1.56080008e-01 -3.91797692e-01 -6.27931654e-01 8.44209135e-01
-2.02837777e+00 -7.24969327e-01 1.00783268e-02 2.04307461e+00
8.83474290e-01 -2.54612803e-01 1.15165688e-01 -5.18961668e-01
5.45953691e-01 2.45269820e-01 -6.58617496e-01 -4.48041916e-01
-1.47977203e-01 -5.40484637e-02 4.61254686e-01 2.86611408e-01
-1.27546287e+00 1.15413678e+00 6.00033951e+00 1.20544016e+00
-1.09965277e+00 4.43031967e-01 4.08635676e-01 -3.70453984e-01
-3.44978809e-01 -4.41993952e-01 -8.00941706e-01 3.60639125e-01
7.08997846e-01 -1.65122952e-02 2.65275955e-01 8.28225672e-01
-1.02161422e-01 1.22736335e-01 -1.03808403e+00 1.30595946e+00
4.26566213e-01 -1.04144704e+00 3.12519342e-01 -2.35465467e-01
5.26928723e-01 1.96476787e-01 3.40949237e-01 6.56955123e-01
1.49994820e-01 -8.45721424e-01 7.56305039e-01 3.56538117e-01
1.22370362e+00 -3.16110998e-01 4.55449134e-01 1.51542976e-01
-1.18739080e+00 -1.51546925e-01 -7.90831223e-02 3.89677107e-01
9.33858156e-02 3.73128027e-01 -2.79568642e-01 6.44179106e-01
7.60715008e-01 6.52146816e-01 -6.33512259e-01 8.55749667e-01
2.07655594e-01 3.71108621e-01 -1.34288698e-01 1.95025742e-01
3.43824208e-01 -7.28947297e-02 4.52040732e-01 1.52163398e+00
1.44846529e-01 -4.09034967e-01 3.71931523e-01 6.49046481e-01
-9.80764106e-02 2.70048022e-01 -7.93618739e-01 8.48616809e-02
2.24360093e-01 1.40073526e+00 -2.00380161e-01 -3.27125579e-01
-7.95369387e-01 1.33850729e+00 4.81264383e-01 5.52516282e-01
-8.60675454e-01 -1.14795089e-01 5.60782194e-01 -6.61761463e-02
3.88157874e-01 -1.66016236e-01 -6.64643422e-02 -1.33178079e+00
1.16822638e-01 -9.22638535e-01 4.49482173e-01 -9.88300264e-01
-1.67184711e+00 7.58182526e-01 2.28983536e-02 -1.13423193e+00
-2.40171235e-02 -4.69256282e-01 -2.42446274e-01 8.62996399e-01
-1.88069224e+00 -1.96322751e+00 -2.77820379e-01 9.84599710e-01
8.67483139e-01 -2.88400203e-01 9.08672035e-01 6.09901845e-01
-5.23876667e-01 8.44758511e-01 1.28835201e-01 -1.16578571e-01
1.25642884e+00 -1.03252804e+00 1.06953770e-01 3.40875030e-01
-1.76901579e-01 4.43436801e-01 3.61442000e-01 -4.05219376e-01
-1.70958292e+00 -1.37032187e+00 9.17528450e-01 -6.02165282e-01
6.48703694e-01 -6.67957366e-01 -7.69856393e-01 8.40706706e-01
4.35894608e-01 -6.78784996e-02 7.91152179e-01 7.30724111e-02
-6.28548026e-01 -2.98547685e-01 -8.77057016e-01 6.80535495e-01
1.03751302e+00 -7.77387619e-01 -4.03079867e-01 6.81055427e-01
8.47672462e-01 -2.44710445e-01 -8.66583765e-01 2.37241849e-01
6.94333911e-01 -6.41084731e-01 1.32202029e+00 -5.77824712e-01
4.28882360e-01 -6.88407868e-02 -2.64795542e-01 -1.21539009e+00
-5.93931913e-01 -5.48490822e-01 -9.33647826e-02 1.30912018e+00
3.28043729e-01 -5.92238724e-01 2.76589036e-01 4.95494932e-01
-2.74179786e-01 -8.42734218e-01 -7.02052295e-01 -8.63055289e-01
7.42935985e-02 -3.11911255e-01 3.01505744e-01 9.37074900e-01
-5.46119690e-01 7.51346529e-01 -7.61363924e-01 -7.63298646e-02
5.82876742e-01 2.61314005e-01 9.48037267e-01 -9.64420319e-01
-3.60334396e-01 -4.04252291e-01 4.07900028e-02 -1.28091288e+00
2.02221557e-01 -1.08837748e+00 -9.12036151e-02 -1.62977314e+00
4.65411782e-01 -5.67524254e-01 -4.10942644e-01 7.31480300e-01
-2.05169663e-01 4.49016601e-01 4.69625503e-01 4.50751036e-01
-9.72096384e-01 5.69150090e-01 1.43724692e+00 -2.22651646e-01
-2.66900957e-01 -1.54429212e-01 -8.96390021e-01 3.72877002e-01
8.38220656e-01 -2.61344880e-01 -4.93466228e-01 -8.40628803e-01
8.75146091e-02 2.71172412e-02 5.60736656e-01 -6.38050437e-01
2.53931552e-01 1.57486256e-02 3.01209122e-01 -5.55647492e-01
6.03590846e-01 -8.61396372e-01 1.25760958e-01 1.53190330e-01
-3.09594691e-01 1.39140636e-01 4.04046476e-01 4.85804260e-01
-2.46951193e-01 3.28585617e-02 6.69616044e-01 -2.89147586e-01
-9.72295642e-01 3.15698534e-01 -3.11934292e-01 -4.42719162e-02
1.12781441e+00 1.89840831e-02 -6.09727740e-01 -2.72452921e-01
-5.77706695e-01 4.52179939e-01 3.77807468e-01 8.87851536e-01
5.12707293e-01 -1.41480541e+00 -7.81616747e-01 2.68556535e-01
6.02640688e-01 -1.58114210e-01 5.84308922e-01 8.51287544e-01
1.39538171e-02 6.58648193e-01 -9.46051404e-02 -7.06977963e-01
-1.40545070e+00 8.30308378e-01 9.98563468e-02 -4.74144071e-01
-3.21268827e-01 7.93937147e-01 4.95502234e-01 -6.14729345e-01
2.10375413e-01 5.70055172e-02 -3.25810984e-02 8.40540007e-02
3.29900086e-01 -6.74275309e-02 -1.60249695e-01 -6.31952703e-01
-2.09691346e-01 8.73426318e-01 -3.44974130e-01 1.88577861e-01
1.15116894e+00 -3.57530475e-01 -6.67955279e-02 5.16173601e-01
1.32488275e+00 -2.72832155e-01 -9.32692289e-01 -4.69918370e-01
-3.19869488e-01 -4.38818872e-01 2.64762223e-01 -1.10335362e+00
-1.11228395e+00 7.09545314e-01 5.78735352e-01 -3.02926893e-03
1.47967196e+00 3.27824056e-01 9.40253615e-01 2.29528114e-01
2.67058849e-01 -9.96978641e-01 3.59751940e-01 3.50893587e-01
1.22816968e+00 -1.49101090e+00 -2.23194733e-01 -3.20716828e-01
-9.07290936e-01 6.30528450e-01 8.51771772e-01 1.71784773e-01
5.26693165e-01 2.09291026e-01 2.42448062e-01 -2.77515054e-01
-1.03928447e+00 -3.00436825e-01 5.47743618e-01 6.33767128e-01
5.13717294e-01 -1.04752727e-01 -1.00116014e-01 5.53626597e-01
2.83275992e-01 3.10744196e-01 -1.41681239e-01 9.12404954e-01
-2.50942223e-02 -1.15816259e+00 -2.84184188e-01 5.04677474e-01
-3.09789896e-01 -1.97497860e-01 -1.94259048e-01 7.39721239e-01
4.28954996e-02 8.08393598e-01 -2.01251749e-02 -3.29174250e-01
4.58320528e-01 1.88318714e-01 5.45094967e-01 -5.41155577e-01
-1.00650656e+00 1.89436778e-01 6.32871613e-02 -6.64871395e-01
-4.18223411e-01 -2.49474019e-01 -4.77327973e-01 -1.19385064e-01
-2.36888975e-01 -2.29184061e-01 6.82198763e-01 5.55409133e-01
8.94568622e-01 5.17028987e-01 4.64302659e-01 -9.22890365e-01
-4.42824960e-01 -8.66139948e-01 -3.80295694e-01 4.96733993e-01
2.51525134e-01 -5.87468147e-01 -1.54455811e-01 8.36303830e-02] | [10.688091278076172, 1.4691765308380127] |
cd9b6434-1876-47ab-843c-35a79f1c04ee | beamlearning-an-end-to-end-deep-learning | 2104.13347 | null | https://arxiv.org/abs/2104.13347v1 | https://arxiv.org/pdf/2104.13347v1.pdf | BeamLearning: an end-to-end Deep Learning approach for the angular localization of sound sources using raw multichannel acoustic pressure data | Sound sources localization using multichannel signal processing has been a subject of active research for decades. In recent years, the use of deep learning in audio signal processing has allowed to drastically improve performances for machine hearing. This has motivated the scientific community to also develop machine learning strategies for source localization applications. In this paper, we present BeamLearning, a multi-resolution deep learning approach that allows to encode relevant information contained in unprocessed time domain acoustic signals captured by microphone arrays. The use of raw data aims at avoiding simplifying hypothesis that most traditional model-based localization methods rely on. Benefits of its use are shown for realtime sound source 2D-localization tasks in reverberating and noisy environments. Since supervised machine learning approaches require large-sized, physically realistic, precisely labelled datasets, we also developed a fast GPU-based computation of room impulse responses using fractional delays for image source models. A thorough analysis of the network representation and extensive performance tests are carried out using the BeamLearning network with synthetic and experimental datasets. Obtained results demonstrate that the BeamLearning approach significantly outperforms the wideband MUSIC and SRP-PHAT methods in terms of localization accuracy and computational efficiency in presence of heavy measurement noise and reverberation. | ['Alexandre Garcia', 'Éric Bavu', 'Hadrien Pujol'] | 2021-04-27 | null | null | null | null | ['audio-signal-processing'] | ['audio'] | [ 2.79262692e-01 -8.07155252e-01 9.22504783e-01 -2.77604401e-01
-1.31219769e+00 -5.28849304e-01 2.87012368e-01 2.50244558e-01
-4.94580746e-01 5.94209015e-01 2.88855106e-01 -2.08438650e-01
-4.59910363e-01 -6.36846721e-01 -5.49221516e-01 -9.64392483e-01
-6.28733695e-01 7.66410828e-02 8.39577839e-02 1.93958320e-02
1.30834743e-01 5.11809707e-01 -1.76482975e+00 3.24674994e-01
3.22812349e-01 1.11509740e+00 4.75529373e-01 1.13685215e+00
1.69708699e-01 7.62658596e-01 -8.95175159e-01 3.20711464e-01
1.59898251e-01 -3.83430004e-01 -2.10110053e-01 -6.36555135e-01
4.88485605e-01 -9.04972106e-02 -3.18510234e-02 8.67299616e-01
1.49245286e+00 5.42834699e-01 3.81057024e-01 -7.55878270e-01
-5.36682159e-02 6.99610949e-01 -2.17499986e-01 5.11494339e-01
6.20284259e-01 -4.11447346e-01 6.86262667e-01 -1.17262554e+00
-1.39862597e-01 1.13734245e+00 1.25574434e+00 -2.85699051e-02
-1.01043868e+00 -6.91423416e-01 -6.00440681e-01 3.79516244e-01
-1.43676555e+00 -6.50088131e-01 1.04811335e+00 -1.56503469e-01
9.99721944e-01 4.87006575e-01 3.55078727e-01 9.54510629e-01
1.94245934e-01 4.56049383e-01 1.43097293e+00 -1.00646329e+00
5.85577071e-01 -1.36293173e-01 -8.89505446e-02 5.06125331e-01
-1.27685964e-01 4.08947140e-01 -7.92864978e-01 -3.53642017e-01
7.11405337e-01 -4.52961832e-01 -4.50851440e-01 3.25171906e-03
-1.21036494e+00 6.65107667e-01 4.13716793e-01 7.37220764e-01
-2.74645209e-01 4.41413879e-01 3.81471425e-01 3.35211515e-01
7.17567682e-01 4.08398598e-01 -4.09505606e-01 -1.83243930e-01
-1.26493311e+00 2.90724784e-01 9.79561448e-01 3.79086554e-01
2.62315929e-01 5.54894090e-01 1.88561231e-01 1.15166509e+00
4.39252377e-01 9.04613793e-01 5.61853886e-01 -8.96671593e-01
-3.17164287e-02 -4.21207279e-01 1.90580413e-01 -1.33056772e+00
-8.07250321e-01 -1.06175947e+00 -9.84086573e-01 2.30752274e-01
4.14868414e-01 -2.64931053e-01 -6.25074267e-01 1.45111978e+00
3.19901705e-01 7.27499187e-01 3.68881300e-02 1.02413654e+00
7.39248633e-01 9.62569773e-01 -5.33045769e-01 -2.49146506e-01
8.93033326e-01 -8.39857519e-01 -8.28400731e-01 2.01977760e-01
1.53447628e-01 -1.33122289e+00 8.77349734e-01 8.23945642e-01
-9.61580336e-01 -9.19260323e-01 -9.81856108e-01 3.61298531e-01
-4.01568145e-01 -3.04806381e-02 5.12379646e-01 1.03813708e+00
-1.20686126e+00 5.35716057e-01 -7.92677045e-01 2.62250826e-02
5.08655719e-02 3.38643521e-01 -7.01485723e-02 -1.46869436e-01
-1.06819057e+00 5.09378672e-01 -2.20969051e-01 4.10655886e-01
-1.18261015e+00 -8.86069000e-01 -7.42680430e-01 -1.77585501e-02
-2.47424036e-01 -3.34425926e-01 1.54991400e+00 -7.31962860e-01
-1.50989425e+00 2.48369500e-01 -2.36081108e-01 -6.25405073e-01
3.01964521e-01 -5.46016097e-01 -8.43528390e-01 1.82959467e-01
7.20207840e-02 -9.98938456e-02 1.15653574e+00 -1.27159286e+00
-4.53096747e-01 -2.33449358e-02 -3.03025037e-01 -6.44103438e-02
-1.41804188e-01 1.15079455e-01 2.99503982e-01 -8.99102211e-01
3.81126910e-01 -5.16205430e-01 -2.61301965e-01 -3.33680153e-01
-8.21395293e-02 1.32223517e-01 5.77165008e-01 -6.60356522e-01
7.59883583e-01 -2.07344103e+00 -2.48174742e-01 4.12380427e-01
-1.60288975e-01 1.77381560e-01 -1.33863419e-01 6.49522901e-01
-2.17082411e-01 -5.97149074e-01 -2.15222195e-01 -6.38173759e-01
-1.49489388e-01 -2.01509714e-01 -6.36680603e-01 7.36338973e-01
-6.00515902e-01 1.94145679e-01 -8.27117622e-01 -2.40910083e-01
4.00436521e-01 9.97757316e-01 -4.74222362e-01 2.96493202e-01
2.47210309e-01 8.22035670e-01 -6.23092465e-02 5.80970049e-01
7.14980364e-01 1.95518658e-01 -4.51103270e-01 -2.17800111e-01
-4.77358580e-01 2.02578023e-01 -1.56283176e+00 1.97275662e+00
-1.09100997e+00 7.96875656e-01 6.33621335e-01 -9.08318460e-01
8.87838542e-01 7.16688991e-01 4.66989249e-01 -8.35355520e-01
1.32542893e-01 8.05444896e-01 -7.28278905e-02 -6.13785982e-01
1.01562679e-01 -3.15639526e-01 2.13497579e-01 5.42274356e-01
9.70371664e-02 -2.65225559e-01 -3.71889234e-01 -4.25533712e-01
1.05829561e+00 -1.20793164e-01 -3.56603116e-02 -4.27332073e-01
8.54619741e-01 -3.56618404e-01 5.41486293e-02 9.79830980e-01
-5.11507876e-02 6.95322812e-01 -5.80798984e-01 -4.40853029e-01
-5.45511007e-01 -1.08356738e+00 -2.38842428e-01 1.41688812e+00
-3.37978154e-01 7.07463697e-02 -8.31680119e-01 2.63300478e-01
-3.72063518e-01 5.92109203e-01 -2.46976435e-01 1.58673182e-01
-9.74835694e-01 -7.44555771e-01 8.13132346e-01 3.76611412e-01
4.64403003e-01 -1.06512380e+00 -7.55715191e-01 5.38017154e-01
-2.04972982e-01 -8.16120684e-01 2.17795633e-02 5.23316979e-01
-6.49644971e-01 -5.86671412e-01 -1.05014455e+00 -1.15653026e+00
1.40014157e-01 3.50548685e-01 8.59620988e-01 -2.41323039e-01
-6.24784887e-01 9.03487802e-01 -4.22814995e-01 -7.47752190e-01
-2.35507339e-01 -2.63826162e-01 4.84143704e-01 1.33460164e-01
-1.14786118e-01 -1.13340497e+00 -6.92830384e-01 1.49879619e-01
-7.39058316e-01 -5.38024127e-01 3.32143933e-01 6.50701106e-01
5.22165656e-01 5.34806669e-01 9.50667202e-01 -1.28757358e-01
9.81837094e-01 -2.22713739e-01 -6.69808269e-01 -9.32637323e-03
-2.81755701e-02 -3.01588953e-01 7.18727112e-01 -3.35855037e-01
-1.15066493e+00 -4.04621959e-02 -6.81613266e-01 -1.71790317e-01
-4.25720274e-01 5.08117020e-01 1.08761452e-01 -4.50136304e-01
9.07689095e-01 4.26324904e-01 -4.47079092e-01 -7.33967960e-01
9.82541293e-02 7.01523483e-01 6.49481893e-01 -6.05133057e-01
6.17694318e-01 4.69336927e-01 1.79932028e-01 -1.31910694e+00
-6.76737726e-01 -7.20353305e-01 -5.08311450e-01 -3.01994234e-01
6.50284529e-01 -8.85070562e-01 -6.42473817e-01 7.15406418e-01
-1.32321608e+00 -3.07899863e-01 -3.57223004e-02 1.05455983e+00
-5.61605632e-01 8.64973441e-02 -4.92842942e-01 -1.22542095e+00
-4.31260914e-01 -9.12636757e-01 9.65461075e-01 -3.09664626e-02
-6.44334853e-02 -1.28099883e+00 6.71976507e-01 4.14745100e-02
9.19979155e-01 2.05404252e-01 5.90138972e-01 -4.65416014e-01
-2.87615806e-01 -3.97625834e-01 2.19425857e-01 4.54362601e-01
2.00953096e-01 -7.27286696e-01 -1.72275865e+00 -4.69845742e-01
5.55455089e-01 -2.07199663e-01 6.99897289e-01 8.50546360e-01
1.09967649e+00 -1.02462051e-02 -5.49696684e-02 6.47852540e-01
1.53974319e+00 2.73977101e-01 1.47280768e-01 1.19016960e-01
3.77516419e-01 5.48815012e-01 3.05192441e-01 4.84348267e-01
-2.04878196e-01 3.63653988e-01 5.08196831e-01 -1.81013256e-01
-3.22849691e-01 -3.33680883e-02 1.29358873e-01 1.16823554e+00
-6.17676452e-02 -2.87372082e-01 -9.01860535e-01 5.88429034e-01
-1.31760502e+00 -8.94343376e-01 -2.54316300e-01 2.18498516e+00
5.23416340e-01 -2.04304725e-01 -1.27580822e-01 8.18787992e-01
3.93685520e-01 1.46106541e-01 -1.64087549e-01 -2.28098691e-01
-1.34849727e-01 9.87685919e-01 2.91087598e-01 8.34187925e-01
-1.11632037e+00 4.65360522e-01 6.24429512e+00 7.73276985e-01
-1.40701115e+00 4.18898284e-01 1.68494537e-01 -7.84493536e-02
-2.54562180e-02 -6.98802352e-01 -3.23937595e-01 1.11025512e-01
1.20153308e+00 2.85779983e-01 6.05625987e-01 6.67784870e-01
5.00498056e-01 -1.02196403e-01 -9.84139800e-01 1.39134896e+00
1.91645592e-01 -1.16088390e+00 -5.17599702e-01 -5.13478279e-01
5.10256231e-01 2.09283486e-01 4.48713541e-01 -5.84378764e-02
-1.91058755e-01 -1.26048517e+00 8.89201641e-01 4.02504385e-01
5.20808995e-01 -8.45894277e-01 7.11061656e-01 4.12554920e-01
-1.33411741e+00 -1.82653755e-01 -5.25243700e-01 -4.24440444e-01
3.70275319e-01 1.00920749e+00 -1.06059408e+00 3.95751655e-01
9.97357726e-01 2.66290158e-01 -1.59297690e-01 1.52375758e+00
-6.02073148e-02 1.18164515e+00 -6.05492473e-01 -1.16836932e-02
1.79972664e-01 1.76856607e-01 6.22855246e-01 1.54107940e+00
8.08500826e-01 -1.75650208e-03 -7.70336315e-02 5.00600398e-01
3.02116930e-01 1.47942603e-01 -5.60393631e-01 4.35979992e-01
3.63899261e-01 1.26187968e+00 -7.93537319e-01 1.99512362e-01
-6.73267767e-02 6.64974868e-01 -4.16719049e-01 4.42699552e-01
-7.33482182e-01 -6.19088113e-01 3.13494503e-01 1.08264431e-01
2.77942300e-01 -7.12178290e-01 -1.53362900e-01 -4.37775970e-01
-2.60605872e-01 -6.51867330e-01 2.93499064e-02 -8.33619058e-01
-1.03479707e+00 8.67105544e-01 -1.22069620e-01 -1.15861678e+00
-2.54628569e-01 -6.30155563e-01 -7.89708793e-01 1.01452982e+00
-1.58510196e+00 -9.27894354e-01 -2.72042662e-01 9.48074341e-01
5.83947897e-01 -2.13994369e-01 1.34202230e+00 6.83247447e-01
4.83289026e-02 4.33643281e-01 5.33745170e-01 1.86420400e-02
7.07014740e-01 -1.09663785e+00 8.00894201e-02 6.61379516e-01
5.98354757e-01 5.84464073e-01 1.13890707e+00 -6.13823757e-02
-1.22816253e+00 -1.01138270e+00 6.56240821e-01 -1.08985901e-01
5.88187873e-01 -5.93366385e-01 -7.80422270e-01 2.29274869e-01
5.00476539e-01 1.75565645e-01 1.03661764e+00 -2.34853402e-01
-1.47103205e-01 -4.56807792e-01 -1.12834942e+00 -5.52027822e-02
6.40107095e-01 -6.04341865e-01 -5.86784065e-01 5.59730172e-01
5.83970249e-01 -3.11836243e-01 -5.29304862e-01 2.06593707e-01
4.51494426e-01 -1.11375713e+00 1.37486053e+00 7.88620859e-02
-3.49325538e-02 -3.18642974e-01 -5.52891254e-01 -1.57456434e+00
-7.56122395e-02 -7.45880008e-01 -4.46058996e-02 1.05459881e+00
2.58894205e-01 -6.30875409e-01 3.23676497e-01 -4.14288908e-01
-3.57710302e-01 -3.76577467e-01 -1.21571004e+00 -7.07091630e-01
5.38705550e-02 -1.12494504e+00 5.12701809e-01 6.34744942e-01
-4.48621690e-01 1.60085574e-01 -4.62722838e-01 7.51569986e-01
1.07217479e+00 2.73305774e-01 4.02088076e-01 -1.15624821e+00
-5.75144112e-01 1.47555536e-02 -1.27299696e-01 -9.94259298e-01
1.23508461e-01 -6.27801895e-01 3.94051015e-01 -1.58039701e+00
-8.09239030e-01 -6.91337526e-01 -6.03879452e-01 6.61604479e-02
4.13469672e-01 5.79870820e-01 -3.01334798e-01 -4.29103635e-02
-2.97573656e-01 3.11635613e-01 9.17363703e-01 4.12021652e-02
-1.27553403e-01 4.83509839e-01 5.15513457e-02 9.49874103e-01
9.73665357e-01 -5.39985716e-01 -5.27467608e-01 -8.16422343e-01
2.98035681e-01 1.34582564e-01 5.61529100e-01 -1.78057611e+00
6.54015541e-01 4.89190966e-01 5.92518985e-01 -6.24732912e-01
7.23291516e-01 -1.09959412e+00 -4.06479537e-02 5.06322443e-01
-4.41602349e-01 -7.08931834e-02 4.07616496e-01 6.23607695e-01
-5.08395612e-01 -5.81660643e-02 7.29441881e-01 -2.06998378e-01
-5.62033296e-01 -2.59392381e-01 -5.86139023e-01 -3.72525573e-01
4.44269150e-01 -2.72076931e-02 1.86598539e-01 -6.66096210e-01
-7.84472406e-01 -6.36073709e-01 -4.04864907e-01 2.49647155e-01
8.38174343e-01 -1.17056894e+00 -7.24088192e-01 4.74998444e-01
-3.79236788e-01 -3.80292498e-02 3.75504106e-01 7.14967728e-01
-7.48781323e-01 6.25526190e-01 -1.01090193e-01 -7.78307438e-01
-1.29029775e+00 2.98496336e-01 6.23421550e-01 1.22230858e-01
-4.12435681e-01 1.44504845e+00 4.48074751e-02 -5.86879194e-01
5.33048868e-01 -5.63023925e-01 -1.65561154e-01 -1.18332878e-01
8.25829387e-01 7.92026401e-01 5.03122509e-01 -5.30020118e-01
-3.64135623e-01 7.48379707e-01 5.16986966e-01 -6.61600173e-01
1.58542824e+00 -3.95260751e-02 -2.85053968e-01 7.14376509e-01
1.42064118e+00 5.15646398e-01 -7.65478313e-01 -2.16626838e-01
-1.26354739e-01 -4.22839910e-01 4.16167885e-01 -1.12142098e+00
-7.28462934e-01 1.31098509e+00 1.25397611e+00 3.60138774e-01
1.41305470e+00 -3.11427027e-01 8.05051684e-01 4.91492927e-01
8.73602152e-01 -7.10994005e-01 2.54931062e-01 5.36672711e-01
1.02101874e+00 -9.42619801e-01 -4.65478182e-01 -5.62720820e-02
3.58410239e-01 1.23113894e+00 3.41330469e-02 -1.21647410e-01
1.08797133e+00 6.69681907e-01 5.41399658e-01 4.96986732e-02
-8.14448819e-02 1.95209831e-01 2.39068344e-01 8.11630309e-01
6.90161347e-01 -2.04370376e-02 3.70222956e-01 5.19080520e-01
-6.28171444e-01 -3.19979966e-01 8.11011121e-02 9.56237555e-01
-7.87099481e-01 -9.79415417e-01 -1.14284194e+00 -4.15047668e-02
-7.02162325e-01 -2.59525359e-01 4.23536934e-02 2.70373702e-01
1.99842006e-01 1.30762720e+00 -1.73930913e-01 -5.08210585e-02
3.42112452e-01 -6.18991368e-02 5.72546601e-01 -3.74364555e-01
-7.18836784e-01 4.80551094e-01 -2.32859910e-01 -2.34394163e-01
-6.19601309e-01 -4.46502745e-01 -1.30205643e+00 8.08184221e-02
-3.11613351e-01 5.34468293e-01 1.10459864e+00 6.39728844e-01
4.79599200e-02 9.80058551e-01 6.35436893e-01 -1.28933132e+00
-5.55466831e-01 -1.04455364e+00 -7.14126647e-01 -3.43389720e-01
9.52447653e-01 -4.74213153e-01 -6.20332956e-01 -5.71847446e-02] | [15.153164863586426, 5.7441534996032715] |
394961fa-1b63-43d9-9d2c-0de347c95e18 | vice-self-supervised-visual-concept | 2111.12460 | null | https://arxiv.org/abs/2111.12460v3 | https://arxiv.org/pdf/2111.12460v3.pdf | ViCE: Improving Dense Representation Learning by Superpixelization and Contrasting Cluster Assignment | Recent self-supervised models have demonstrated equal or better performance than supervised methods, opening for AI systems to learn visual representations from practically unlimited data. However, these methods are typically classification-based and thus ineffective for learning high-resolution feature maps that preserve precise spatial information. This work introduces superpixels to improve self-supervised learning of dense semantically rich visual concept embeddings. Decomposing images into a small set of visually coherent regions reduces the computational complexity by $\mathcal{O}(1000)$ while preserving detail. We experimentally show that contrasting over regions improves the effectiveness of contrastive learning methods, extends their applicability to high-resolution images, improves overclustering performance, superpixels are better than grids, and regional masking improves performance. The expressiveness of our dense embeddings is demonstrated by improving the SOTA unsupervised semantic segmentation benchmark on Cityscapes, and for convolutional models on COCO. | ['Kazuya Takeda', 'Kento Ohtani', 'Alexander Carballo', 'Keisuke Fujii', 'Tomoki Hayashi', 'Robin Karlsson'] | 2021-11-24 | null | null | null | null | ['unsupervised-semantic-segmentation', 'learning-word-embeddings'] | ['computer-vision', 'methodology'] | [ 1.88524768e-01 3.02740782e-01 -4.29607809e-01 -4.86425847e-01
-6.99659467e-01 -6.16386473e-01 6.77113652e-01 3.98421198e-01
-8.40405047e-01 6.08157039e-01 2.36206606e-01 1.32974666e-02
1.00787049e-02 -8.01056564e-01 -6.50991976e-01 -6.49245441e-01
-3.26698750e-01 3.74235034e-01 4.95130926e-01 1.24566495e-01
4.38921489e-02 5.19442379e-01 -1.78330636e+00 2.37172097e-01
8.43750775e-01 8.30982149e-01 3.03545147e-01 4.87599581e-01
-2.93503642e-01 6.28596187e-01 -3.51348430e-01 1.10326745e-01
1.42986432e-01 -2.15748504e-01 -8.07561696e-01 3.31475288e-01
9.21497703e-01 -2.45066255e-01 -1.08359642e-01 1.00130510e+00
1.69652253e-01 1.99627221e-01 9.84865725e-01 -1.03718293e+00
-9.96819317e-01 3.95316958e-01 -8.10531616e-01 4.36407268e-01
-1.52546793e-01 -3.42624709e-02 1.28051758e+00 -8.21291685e-01
8.55352044e-01 1.15500367e+00 7.43144035e-01 3.12007815e-01
-1.63209176e+00 -4.49174434e-01 2.85957098e-01 -3.45428735e-02
-1.44099832e+00 -2.98751384e-01 6.79922640e-01 -3.80001277e-01
1.14933813e+00 9.09600183e-02 7.09641874e-01 8.08219016e-01
-2.15528756e-01 8.25703025e-01 1.16131580e+00 -4.11599964e-01
4.28028494e-01 2.25086376e-01 2.34096482e-01 8.11143994e-01
4.79347527e-01 2.60001332e-01 -3.89567882e-01 2.74645835e-01
1.14488804e+00 1.50224075e-01 -1.34537354e-01 -7.79585302e-01
-1.10362399e+00 1.00725460e+00 9.96175885e-01 6.21309996e-01
2.36200523e-02 2.42023736e-01 1.81689963e-01 5.58993109e-02
7.09043741e-01 7.20150948e-01 -2.77177542e-01 1.37296870e-01
-1.34656262e+00 -5.97756132e-02 3.16793054e-01 8.63422930e-01
1.10474849e+00 3.88414323e-01 1.46636635e-01 9.72777963e-01
-8.45406428e-02 3.67961437e-01 4.84434098e-01 -1.14860499e+00
2.57704761e-02 7.46682048e-01 1.81541853e-02 -1.09317720e+00
-6.60740614e-01 -5.42578459e-01 -7.90560722e-01 6.14947438e-01
5.81685483e-01 1.24443851e-01 -1.29328561e+00 1.33954680e+00
-3.58921625e-02 -1.80314660e-01 3.65302041e-02 9.58822548e-01
7.62492597e-01 6.24839664e-01 2.84913570e-01 1.31558284e-01
1.01623130e+00 -1.16951072e+00 -5.29913008e-01 -4.91466016e-01
7.63738632e-01 -4.18853074e-01 1.27951097e+00 2.82712787e-01
-9.19794619e-01 -7.76108801e-01 -1.20670640e+00 -4.25624877e-01
-8.02990317e-01 1.29265279e-01 9.36109006e-01 7.54498720e-01
-1.11275184e+00 6.68488026e-01 -9.13260639e-01 -4.28246439e-01
1.18235862e+00 2.57579267e-01 -4.87556636e-01 -1.49745107e-01
-7.24680305e-01 6.52801335e-01 5.94816685e-01 -5.00209570e-01
-7.46683836e-01 -9.35597539e-01 -1.21518695e+00 1.28594920e-01
-8.51194337e-02 -1.33861139e-01 5.68082035e-01 -1.26753235e+00
-9.12156641e-01 1.16622758e+00 -8.44211578e-02 -7.87251711e-01
2.65982836e-01 -1.27918333e-01 -3.22337776e-01 6.10294759e-01
3.15896869e-01 1.49647355e+00 8.19417596e-01 -1.53319538e+00
-6.88083768e-01 -3.27103049e-01 5.17665148e-02 2.50958383e-01
-7.56846666e-01 -4.46353018e-01 -4.12730247e-01 -6.63484752e-01
1.14401817e-01 -5.16803324e-01 -3.28406990e-01 3.44940782e-01
-8.58836249e-02 -2.50161514e-02 9.25310075e-01 -3.95521313e-01
8.08636248e-01 -2.38753057e+00 -1.04023375e-01 1.01159640e-01
3.61189216e-01 1.09249599e-01 -2.07783669e-01 -3.94907966e-03
-1.31890848e-01 3.95903915e-01 -5.21799862e-01 -2.60383785e-01
-5.77730760e-02 5.58382869e-01 -1.68119401e-01 5.44247985e-01
4.75343466e-01 1.27788734e+00 -8.64309609e-01 -7.62031078e-01
5.79052329e-01 5.79363704e-01 -4.92227226e-01 -1.20918781e-01
-1.19512834e-01 1.37414843e-01 -9.77646559e-02 4.66498584e-01
5.52922845e-01 -6.00030541e-01 -3.91683094e-02 4.60754000e-02
-1.08700648e-01 -1.07975192e-01 -1.00332928e+00 1.91944432e+00
-3.87700677e-01 9.75442648e-01 -9.31663215e-02 -1.18847907e+00
9.28691745e-01 -2.57105231e-01 4.20929641e-01 -1.15090001e+00
-1.13047272e-01 -2.30358735e-01 -3.13220531e-01 -1.09904952e-01
4.29584473e-01 -1.17072769e-01 8.81500468e-02 2.87858725e-01
3.71067077e-01 -2.60445565e-01 2.36618295e-01 2.18895718e-01
7.56436884e-01 2.91437119e-01 2.36591324e-01 -5.96556723e-01
-1.13149077e-01 4.90764916e-01 3.33405882e-01 6.83105230e-01
-2.08630532e-01 1.09760332e+00 4.32071775e-01 -4.51753676e-01
-1.21717417e+00 -1.42989278e+00 -5.21329880e-01 1.28318071e+00
4.11320627e-01 -2.67542064e-01 -7.43681073e-01 -6.92965925e-01
1.35281324e-01 5.87905049e-01 -1.04035127e+00 7.14553371e-02
-3.22228849e-01 -7.73276091e-01 4.73533958e-01 1.05450177e+00
6.11555517e-01 -8.13179910e-01 -8.22787106e-01 -2.38968562e-02
1.33555576e-01 -1.10826766e+00 -2.26155207e-01 5.79898536e-01
-1.07483280e+00 -9.77151453e-01 -9.19202685e-01 -1.15209723e+00
1.01427841e+00 4.12048161e-01 1.17769098e+00 -7.87265077e-02
-7.78652549e-01 3.12436730e-01 -4.13100302e-01 -1.15740240e-01
7.38736168e-02 6.29191846e-02 -1.66123405e-01 -3.77746761e-01
3.51884604e-01 -5.11740148e-01 -4.84930217e-01 1.55817553e-01
-9.36389685e-01 -2.75187436e-02 5.20636439e-01 9.90124404e-01
9.09044981e-01 1.66864991e-01 3.26074779e-01 -1.12370396e+00
7.66175091e-02 -1.71761245e-01 -5.02232492e-01 -3.55622545e-02
-6.87876761e-01 2.92540491e-01 6.04039729e-01 -3.62558633e-01
-9.66047823e-01 3.03337276e-01 1.57544211e-01 -3.86403859e-01
-6.09560549e-01 5.68646528e-02 2.14691177e-01 -7.30692148e-02
9.92777646e-01 1.54736787e-01 4.15946636e-03 -4.43316579e-01
8.23473096e-01 3.13460261e-01 6.40798807e-01 -4.06012952e-01
9.02671218e-01 9.30385053e-01 -4.07561004e-01 -1.10279751e+00
-7.74359226e-01 -5.59690356e-01 -1.07636178e+00 2.38235265e-01
1.14378309e+00 -1.09758532e+00 -1.91319641e-02 -1.44313812e-01
-5.17853260e-01 -4.98595595e-01 -7.23542452e-01 2.69575149e-01
-6.15339577e-01 3.73040885e-01 -3.60967577e-01 -4.64250684e-01
9.66314822e-02 -6.98434055e-01 9.83730972e-01 3.97163391e-01
-2.90327460e-01 -1.31791019e+00 -4.13302854e-02 2.62229294e-01
2.45222643e-01 3.86476189e-01 8.74144137e-01 -4.26555425e-01
-4.68309015e-01 9.87763256e-02 -6.96648300e-01 4.98765767e-01
1.76835746e-01 -1.52309403e-01 -1.28262937e+00 -2.13349074e-01
-7.16477096e-01 -4.74058092e-01 1.49054074e+00 5.10056853e-01
1.20686960e+00 -1.11329362e-01 -4.55720097e-01 7.81144857e-01
1.68363965e+00 -1.70250863e-01 6.80903256e-01 4.21232492e-01
8.23707342e-01 5.90969026e-01 4.03242350e-01 8.19345489e-02
9.63056609e-02 2.29428515e-01 4.20102119e-01 -7.98841059e-01
-5.18196881e-01 -2.08072826e-01 7.50265792e-02 2.64767706e-01
1.01562003e-02 1.47501886e-01 -9.04545546e-01 1.09047508e+00
-1.55342185e+00 -8.24101567e-01 5.47510087e-02 1.88201356e+00
8.21646750e-01 2.17527732e-01 1.49042025e-01 2.77560294e-01
5.25639534e-01 3.64317834e-01 -4.30304557e-01 -3.43791574e-01
-4.17654514e-01 6.15676820e-01 7.36983776e-01 5.16006529e-01
-1.39418745e+00 1.34677052e+00 6.74074554e+00 8.71792614e-01
-9.39499557e-01 1.23667076e-01 7.98626721e-01 -1.03312045e-01
-2.67469943e-01 -2.61892289e-01 -5.35674393e-01 1.21537432e-01
5.68367660e-01 3.52608263e-01 1.24287039e-01 1.05310762e+00
-8.25564861e-02 -3.64845365e-01 -1.09694767e+00 1.03363442e+00
2.46194825e-01 -1.72944725e+00 9.09819156e-02 -4.70135175e-02
1.10468554e+00 8.51844698e-02 4.43099052e-01 1.37910292e-01
5.76899111e-01 -1.60339558e+00 4.85817790e-01 2.94270981e-02
8.99760962e-01 -9.34686661e-01 4.19737428e-01 -2.11870987e-02
-1.14647436e+00 -1.82437062e-01 -6.30463123e-01 -6.36004657e-02
-2.84125030e-01 4.70638812e-01 -6.44465387e-01 2.78515816e-01
1.01885545e+00 1.10734832e+00 -1.15094614e+00 8.08660328e-01
-1.68364376e-01 4.87720400e-01 -3.04962754e-01 1.14920311e-01
6.23914421e-01 -1.79143697e-02 -5.34716919e-02 1.50855160e+00
-3.60168636e-01 -4.76119034e-02 1.57937348e-01 1.05190516e+00
9.33949091e-03 5.22000082e-02 -6.97002411e-01 -2.06333175e-01
1.83993399e-01 1.22076035e+00 -1.27756071e+00 -4.18578267e-01
-3.46676558e-01 1.13118923e+00 4.78142500e-01 6.34287059e-01
-6.04519010e-01 -4.83959585e-01 6.36403501e-01 3.90991718e-01
7.09612668e-01 -4.39987510e-01 -8.04644644e-01 -1.01998997e+00
-2.83853114e-01 -4.28975284e-01 3.33595246e-01 -7.99462259e-01
-1.08689189e+00 4.44005936e-01 -5.57708205e-04 -1.14667404e+00
-6.18013814e-02 -7.12564647e-01 -2.46321276e-01 4.90572512e-01
-1.78442478e+00 -1.21474445e+00 -4.09433126e-01 4.62520599e-01
4.67946231e-01 -1.41458148e-02 1.00549138e+00 3.48204048e-03
-2.57024080e-01 5.01961946e-01 4.34886485e-01 5.27015090e-01
5.06489754e-01 -1.59869933e+00 1.85771435e-01 7.60495961e-01
7.48094320e-01 3.41502845e-01 4.00170386e-01 -2.83475161e-01
-7.17517912e-01 -1.23126101e+00 5.24729192e-01 -4.51984018e-01
5.25889218e-01 -4.62081492e-01 -1.02428699e+00 4.74410325e-01
2.78248191e-01 4.84695017e-01 6.04490995e-01 8.49735215e-02
-8.86318564e-01 -3.89680602e-02 -1.23824453e+00 5.44004858e-01
1.05974090e+00 -5.91977000e-01 -7.56080627e-01 2.03717500e-01
5.80078602e-01 -1.03032626e-02 -1.01217127e+00 8.50411952e-02
1.99865028e-01 -9.85588968e-01 1.26425242e+00 -4.94869947e-01
5.24673820e-01 -2.27419019e-01 -1.32868648e-01 -1.24501455e+00
-3.97552490e-01 6.36740029e-02 2.56780624e-01 9.35852051e-01
3.76865417e-01 -4.28373456e-01 9.70988154e-01 9.34768319e-02
1.27037913e-01 -4.24722016e-01 -7.72532523e-01 -8.79924059e-01
4.16033983e-01 -2.22201437e-01 1.32673159e-01 1.24578190e+00
-5.78712635e-02 2.29946688e-01 1.89260796e-01 1.21057712e-01
8.87740076e-01 2.70872742e-01 4.11276400e-01 -1.23998272e+00
-1.00964345e-01 -5.94188631e-01 -7.45290875e-01 -9.54157054e-01
8.27300549e-02 -1.01012766e+00 -1.79590032e-01 -1.79768491e+00
1.59687549e-01 -4.80627805e-01 -3.03298891e-01 8.59752297e-01
3.75562832e-02 8.90695095e-01 3.37166823e-02 5.80268577e-02
-7.16057062e-01 3.51886362e-01 1.11474121e+00 -4.63175178e-01
-2.07478434e-01 -8.21497321e-01 -7.22093523e-01 7.74173975e-01
8.14829409e-01 -3.47037524e-01 -5.17297864e-01 -6.46467447e-01
-1.89116925e-01 -6.00308478e-01 6.23579741e-01 -1.22660911e+00
-6.84738234e-02 9.56051052e-02 1.06324613e+00 -3.95830899e-01
1.77338973e-01 -7.08284557e-01 -4.57343012e-01 2.56474853e-01
-5.66166580e-01 -4.06009376e-01 5.64732909e-01 6.68016374e-01
-3.08183134e-01 2.14184113e-02 1.16266084e+00 -1.51992068e-01
-1.22510993e+00 1.16735205e-01 -2.81008512e-01 4.44785178e-01
9.51199889e-01 -6.76666856e-01 -2.42208958e-01 -1.03261225e-01
-9.01194632e-01 9.86666828e-02 8.10006678e-01 4.98187065e-01
6.31374121e-01 -1.15288842e+00 -3.38117331e-01 2.95879543e-01
4.01185334e-01 1.46582410e-01 1.69259012e-01 3.15608591e-01
-8.40321958e-01 3.29680055e-01 -6.51386976e-01 -9.15895641e-01
-1.20569098e+00 5.40256798e-01 1.95625558e-01 1.14913583e-01
-9.71798897e-01 1.01618624e+00 5.01313210e-01 -1.93998784e-01
3.31793934e-01 -5.49444377e-01 -3.93708825e-01 1.90134808e-01
4.97669250e-01 2.03534067e-01 -2.23073304e-01 -6.56179667e-01
-3.67207199e-01 7.48601079e-01 -7.73666129e-02 -1.31102622e-01
1.51039398e+00 -5.62470779e-02 2.32182175e-01 4.52799797e-01
1.55347133e+00 -2.97203034e-01 -1.81109464e+00 -2.31576562e-01
1.18642524e-01 -4.64589804e-01 4.36797142e-01 -7.08009660e-01
-1.11940813e+00 1.21930623e+00 8.95671368e-01 1.55617431e-01
9.03009534e-01 1.85973331e-01 4.48655903e-01 2.29465783e-01
1.91321298e-01 -1.28579271e+00 3.69872212e-01 2.25774661e-01
4.82039660e-01 -1.49784565e+00 2.38640204e-01 -3.40247929e-01
-8.00201595e-01 1.11648691e+00 4.42438394e-01 -5.29588699e-01
6.25288904e-01 3.38701129e-01 1.09692544e-01 -1.30023539e-01
-1.87404379e-01 -7.24017620e-01 3.83190185e-01 1.25420868e+00
1.87619030e-01 9.47646499e-02 1.96566284e-01 2.84618527e-01
-1.12515002e-01 -3.54019493e-01 4.26403254e-01 8.03924739e-01
-7.72274673e-01 -6.46795154e-01 -1.58431843e-01 3.50775540e-01
-1.49171263e-01 -2.36342266e-01 -3.23357880e-01 1.15953302e+00
2.81478137e-01 7.37681806e-01 7.79220104e-01 -1.62822334e-03
-3.68750095e-02 1.94405466e-02 5.75544775e-01 -7.80247092e-01
-1.96138486e-01 5.40643521e-02 -8.77629593e-02 -5.39222956e-01
-5.19149184e-01 -5.27118325e-01 -1.69445515e+00 1.08731780e-02
4.22273800e-02 -4.53679375e-02 4.54429716e-01 7.78313041e-01
2.37265423e-01 5.09336770e-01 4.79451686e-01 -9.07039404e-01
6.80137798e-02 -6.50428772e-01 -7.61908889e-01 5.29341280e-01
3.42093796e-01 -7.48658419e-01 -4.01403159e-01 5.07224679e-01] | [9.70186996459961, 1.1673661470413208] |
bd32add4-36d1-4683-a2db-6d3b94627eed | abstractive-summarization-as-augmentation-for | 2305.18023 | null | https://arxiv.org/abs/2305.18023v1 | https://arxiv.org/pdf/2305.18023v1.pdf | Abstractive Summarization as Augmentation for Document-Level Event Detection | Transformer-based models have consistently produced substantial performance gains across a variety of NLP tasks, compared to shallow models. However, deep models are orders of magnitude more computationally expensive than shallow models, especially on tasks with large sequence lengths, such as document-level event detection. In this work, we attempt to bridge the performance gap between shallow and deep models on document-level event detection by using abstractive text summarization as an augmentation method. We augment the DocEE dataset by generating abstractive summaries of examples from low-resource classes. For classification, we use linear SVM with TF-IDF representations and RoBERTa-base. We use BART for zero-shot abstractive summarization, making our augmentation setup less resource-intensive compared to supervised fine-tuning. We experiment with four decoding methods for text generation, namely beam search, top-k sampling, top-p sampling, and contrastive search. Furthermore, we investigate the impact of using document titles as additional input for classification. Our results show that using the document title offers 2.04% and 3.19% absolute improvement in macro F1-score for linear SVM and RoBERTa, respectively. Augmentation via summarization further improves the performance of linear SVM by about 0.5%, varying slightly across decoding methods. Overall, our augmentation setup yields insufficient improvements for linear SVM compared to RoBERTa. | ['Domagoj Pluščec', 'Filip Karlo Došilović', 'Janko Vidaković'] | 2023-05-29 | null | null | null | null | ['abstractive-text-summarization', 'text-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 5.49342394e-01 9.48730484e-02 -3.40758473e-01 -2.21762791e-01
-1.52764606e+00 -5.56372464e-01 9.65431929e-01 5.97382724e-01
-5.13730466e-01 1.02110779e+00 7.10427463e-01 -3.27875018e-01
3.02917928e-01 -6.32286191e-01 -6.72068655e-01 -5.66665351e-01
9.68156978e-02 4.14947957e-01 8.46704766e-02 -1.19942501e-01
4.32480514e-01 9.39729288e-02 -1.37311745e+00 8.06702614e-01
9.44087446e-01 8.08367014e-01 -2.51863375e-02 9.89827335e-01
-2.45240614e-01 6.83976650e-01 -1.13321292e+00 -2.89409786e-01
-1.02757618e-01 -7.11667895e-01 -8.21164131e-01 -1.86772317e-01
6.48323953e-01 -3.87645364e-01 -5.08404136e-01 4.02897477e-01
8.11779439e-01 1.63152844e-01 9.04049456e-01 -9.55302060e-01
-3.61738265e-01 9.68657434e-01 -4.65644866e-01 5.33983588e-01
6.73503101e-01 1.44248143e-01 1.25986421e+00 -1.00087559e+00
4.10377532e-01 1.09191549e+00 8.18849146e-01 4.28154141e-01
-1.29189229e+00 -5.05908549e-01 -4.76709604e-02 1.03943221e-01
-9.11383510e-01 -8.03701401e-01 2.57057399e-01 -2.84071833e-01
1.60703635e+00 5.48175156e-01 4.17906225e-01 1.36777341e+00
1.79337740e-01 1.16062081e+00 8.61216366e-01 -4.87376094e-01
2.40117848e-01 1.64049529e-02 4.71561730e-01 4.99252051e-01
2.58833170e-01 -2.98512906e-01 -8.42782676e-01 -5.59367657e-01
1.91346586e-01 -3.56827468e-01 -1.94815144e-01 4.84540135e-01
-1.25492704e+00 9.54314828e-01 -1.54081862e-02 1.87640842e-02
-5.03951550e-01 1.91740796e-01 7.73407698e-01 5.96097894e-02
7.81823456e-01 7.45881379e-01 -4.78868693e-01 -3.94205540e-01
-1.51216114e+00 3.93945545e-01 9.60979462e-01 7.51781344e-01
1.56857967e-01 3.85966569e-01 -9.81825709e-01 1.21309125e+00
-2.70862222e-01 5.05220950e-01 8.73313785e-01 -6.01493537e-01
8.82491350e-01 3.56661767e-01 9.59398896e-02 -5.35152674e-01
-4.02783543e-01 -6.12713754e-01 -8.66346300e-01 -4.55373019e-01
2.92486191e-01 -3.24777454e-01 -1.15281868e+00 1.46857333e+00
-1.14228360e-01 6.56993187e-04 3.73521626e-01 3.94330055e-01
1.05859351e+00 1.10876226e+00 -9.07337945e-03 -3.98259163e-01
1.31236196e+00 -1.32342243e+00 -6.00178659e-01 -4.19695586e-01
8.54095340e-01 -6.65517032e-01 1.19804859e+00 2.53243029e-01
-1.40455127e+00 -2.68994778e-01 -9.25180554e-01 -2.16202080e-01
-2.02316828e-02 4.76967990e-01 4.97379482e-01 4.36230212e-01
-8.14570487e-01 5.95341325e-01 -7.75981069e-01 -4.36484426e-01
5.37364960e-01 6.37739152e-02 -9.03827846e-02 1.16023585e-01
-1.27719593e+00 8.41084003e-01 5.23992896e-01 -4.46290821e-01
-6.78120136e-01 -8.94872963e-01 -8.74785602e-01 4.42216635e-01
1.15559891e-01 -6.81998849e-01 1.45190895e+00 -3.54106545e-01
-1.28137314e+00 6.55948341e-01 -5.87934554e-01 -9.83704925e-01
3.81349564e-01 -3.41862470e-01 -1.08282082e-01 1.49171636e-01
1.84679121e-01 5.48185349e-01 5.23236096e-01 -7.44511843e-01
-6.86294138e-01 -6.52831346e-02 -1.72368661e-01 4.32943970e-01
-4.55262005e-01 1.57955930e-01 -3.39348257e-01 -9.27253067e-01
-2.45358720e-01 -6.80281878e-01 -1.51445225e-01 -7.25231349e-01
-7.69604445e-01 -3.53994697e-01 3.98047328e-01 -1.02572584e+00
1.50840127e+00 -1.79583931e+00 -7.77715594e-02 -1.99531555e-01
-2.17374191e-01 3.57650965e-01 -2.39540577e-01 6.36102378e-01
1.06709071e-01 1.36685118e-01 -4.20952737e-01 -4.72784758e-01
-1.94458172e-01 -9.05804709e-02 -5.69946945e-01 -9.58626042e-04
4.25490052e-01 9.34381127e-01 -7.82508433e-01 -3.64668787e-01
-2.29231626e-01 1.10615261e-01 -5.29048562e-01 1.12836538e-02
-3.48901540e-01 -3.60544175e-02 -1.56919971e-01 5.69014907e-01
1.38823956e-01 -4.60134119e-01 -1.18110739e-01 5.70217706e-02
1.55397609e-01 1.02568960e+00 -6.78605378e-01 1.62671828e+00
-6.17726982e-01 8.65604162e-01 -4.81811315e-01 -1.13232160e+00
6.73624814e-01 3.97368997e-01 6.08806275e-02 -6.05922520e-01
-1.94835216e-02 1.63052112e-01 -9.38483775e-02 -1.20115966e-01
9.14494753e-01 2.47872397e-02 -2.76786059e-01 4.68646020e-01
6.36019334e-02 8.80855247e-02 6.07537389e-01 4.73714918e-01
1.46438062e+00 -1.59225017e-01 4.65239882e-01 6.52106926e-02
2.69541293e-01 2.86608964e-01 2.67570883e-01 1.11669743e+00
3.68830711e-01 8.41404140e-01 6.57101691e-01 -1.08854212e-02
-9.74171460e-01 -8.12305212e-01 -1.87360764e-01 1.27731001e+00
-3.04531187e-01 -6.87397718e-01 -7.74265707e-01 -8.19305301e-01
-2.24571000e-03 1.28643727e+00 -3.56299937e-01 -3.25737923e-01
-8.01977575e-01 -1.24490535e+00 1.00416172e+00 7.20304906e-01
3.77438277e-01 -1.03847504e+00 -4.15657014e-01 3.03512067e-01
-5.48034787e-01 -1.17624986e+00 -5.43309867e-01 4.08359349e-01
-9.90976334e-01 -5.38600385e-01 -1.02417874e+00 -5.25617182e-01
4.03290063e-01 6.43593147e-02 1.16101694e+00 -2.23042965e-01
-1.42218545e-01 7.14672580e-02 -4.61562395e-01 -3.99180770e-01
-7.14673758e-01 6.83763742e-01 -2.38724612e-02 -4.79743421e-01
4.26328257e-02 -3.06732893e-01 -4.58275378e-01 -4.98337066e-03
-5.52784204e-01 2.54184961e-01 7.46054828e-01 1.07328105e+00
3.14211130e-01 -3.78110349e-01 9.62651789e-01 -8.66362751e-01
8.45826209e-01 -6.05756521e-01 -1.11630835e-01 1.21883959e-01
-6.10998094e-01 2.24575698e-01 8.65684450e-01 -3.90831739e-01
-9.87273276e-01 -4.51165706e-01 -3.02478790e-01 -4.25264379e-03
1.51153980e-02 6.79843724e-01 2.67515808e-01 6.19486928e-01
1.02457213e+00 6.93481743e-01 -1.57668099e-01 -4.35083300e-01
1.21762075e-01 9.62307334e-01 4.32755858e-01 -3.04645479e-01
4.41344827e-01 3.19042206e-02 -3.06843996e-01 -1.01546788e+00
-1.12772048e+00 -4.30275649e-01 -6.13738336e-02 3.11313689e-01
4.82109487e-01 -1.03270459e+00 -1.10656753e-01 4.88094598e-01
-1.12326813e+00 -4.95877862e-01 -4.02019262e-01 3.52614015e-01
-4.96646076e-01 2.89854914e-01 -8.05193543e-01 -6.08729720e-01
-9.98392880e-01 -8.54109883e-01 1.31877553e+00 7.53053576e-02
-6.11921251e-01 -6.58396423e-01 -2.08907388e-02 5.68957508e-01
3.87100875e-01 2.12912466e-02 8.61890793e-01 -1.21373451e+00
-2.32590422e-01 -3.33438843e-01 -1.32069424e-01 3.60675454e-01
-5.52127557e-03 -3.25954974e-01 -9.70719993e-01 -3.04800481e-01
-3.88770461e-01 -4.18394655e-01 1.36216152e+00 4.16182071e-01
1.28853273e+00 -6.39298260e-01 -4.60439473e-01 3.95579934e-01
8.80770147e-01 7.58011043e-02 4.96851027e-01 3.17917377e-01
5.11338532e-01 2.24637434e-01 7.65250266e-01 5.98785937e-01
2.53294826e-01 7.77839899e-01 -1.44515380e-01 1.20950356e-01
-3.47523779e-01 -2.51982450e-01 6.27546728e-01 6.95268691e-01
1.05210677e-01 -7.19647646e-01 -9.93440032e-01 7.07666457e-01
-1.66623092e+00 -1.11242020e+00 -6.94524422e-02 2.17244101e+00
1.31599903e+00 3.31222326e-01 1.65109619e-01 2.88866699e-01
4.94746417e-01 2.51212656e-01 -4.07073915e-01 -6.03933573e-01
-1.23408452e-01 3.58877659e-01 5.20779967e-01 3.29629451e-01
-1.08467007e+00 9.34361696e-01 6.28568792e+00 1.14103293e+00
-1.15566504e+00 1.29252151e-01 6.42260313e-01 -6.25262976e-01
-1.40865654e-01 -2.71677434e-01 -1.28792405e+00 8.27207446e-01
1.49808705e+00 -3.47601533e-01 6.23718277e-02 5.84863961e-01
2.57576525e-01 -2.79065162e-01 -1.31246018e+00 7.07489967e-01
2.68067300e-01 -1.77240920e+00 3.20618957e-01 -5.18247485e-02
8.98954630e-01 2.03496531e-01 -2.16958374e-01 7.20691502e-01
1.56539857e-01 -9.36876357e-01 6.73553467e-01 1.25657350e-01
8.51363838e-01 -5.72470725e-01 7.58543730e-01 5.19219995e-01
-7.27540135e-01 -9.36064695e-04 -1.84432268e-01 4.96263392e-02
2.72901833e-01 7.97432899e-01 -1.13575590e+00 4.90631878e-01
4.56172794e-01 5.25122821e-01 -4.19758409e-01 9.88872528e-01
-9.63424984e-03 1.17230392e+00 -4.11949813e-01 -2.85687685e-01
2.91839242e-01 2.92320788e-01 7.60822773e-01 1.68370569e+00
4.29403782e-01 3.38830985e-02 -2.00810167e-03 3.96757632e-01
-4.50358123e-01 1.22416407e-01 -2.90394604e-01 -1.24816924e-01
6.71141386e-01 9.05457437e-01 -5.73047996e-01 -8.30958188e-01
-2.00606167e-01 1.17860913e+00 1.47040606e-01 1.68656871e-01
-9.96186733e-01 -6.77457631e-01 3.53155047e-01 7.44923577e-02
3.72093618e-01 4.91920710e-02 -4.77522016e-01 -1.21938312e+00
1.34524405e-01 -9.30670917e-01 4.05856699e-01 -7.48703301e-01
-8.84615839e-01 3.81754369e-01 7.46246055e-02 -9.11097825e-01
-6.14207745e-01 -2.84183566e-02 -8.22797716e-01 9.32370484e-01
-1.29113483e+00 -7.95082867e-01 -1.40253782e-01 5.54765835e-02
1.20795774e+00 -2.24704146e-01 8.41222107e-01 1.83494180e-01
-5.98003983e-01 9.14791465e-01 4.19187069e-01 3.81848938e-03
7.64203548e-01 -1.36754215e+00 8.32055449e-01 7.80104876e-01
2.01444030e-01 4.54398096e-01 7.83760548e-01 -6.55974865e-01
-1.25104558e+00 -1.30276787e+00 1.14818764e+00 -4.29388642e-01
3.21434230e-01 -4.03260171e-01 -9.27309215e-01 5.74725389e-01
2.47973800e-01 -4.26379889e-01 6.64522648e-01 2.52597451e-01
-1.56503424e-01 1.22884132e-01 -9.99268472e-01 6.36883020e-01
8.48632038e-01 -3.23681593e-01 -7.95171320e-01 6.15158379e-01
8.07662785e-01 -5.07969856e-01 -7.55149245e-01 3.53313923e-01
4.45084453e-01 -5.67351878e-01 9.58211005e-01 -5.92696667e-01
9.04645920e-01 2.18384027e-01 -9.41174477e-03 -1.64019191e+00
-4.85906042e-02 -3.72651756e-01 -4.70266432e-01 1.27370632e+00
8.07714880e-01 -5.97594023e-01 8.45664203e-01 4.14181389e-02
-3.62244248e-01 -9.86817062e-01 -6.13571584e-01 -8.85125399e-01
1.98082671e-01 -3.07296544e-01 2.57012218e-01 7.40625560e-01
1.31228492e-01 7.34923482e-01 -2.84795910e-01 -7.31021613e-02
4.07747567e-01 9.55673084e-02 7.15620339e-01 -1.01401687e+00
-3.92370492e-01 -5.53743541e-01 4.15052548e-02 -1.11457455e+00
1.62763000e-01 -1.26979554e+00 1.23165481e-01 -1.95354819e+00
3.90791416e-01 -1.37681216e-01 -9.95971784e-02 7.86917865e-01
-4.21224415e-01 2.18966454e-01 -3.18832062e-02 2.02661201e-01
-2.81257927e-01 6.11496806e-01 6.67532623e-01 -2.28786722e-01
-3.54085445e-01 9.67276171e-02 -8.42434525e-01 5.72371364e-01
1.01442850e+00 -4.75793451e-01 -2.27251276e-01 -3.91634136e-01
2.37897271e-03 3.73780906e-01 8.47906247e-02 -9.27753329e-01
1.58130601e-02 1.32821277e-01 4.08815593e-01 -8.71900976e-01
2.85068870e-01 1.19982645e-01 -4.33847547e-01 4.76853818e-01
-8.91386569e-01 -1.20218672e-01 5.07318974e-01 5.28536737e-01
-1.54635295e-01 -3.46165895e-01 6.30874038e-01 3.43300663e-02
-1.47254616e-01 -8.54768530e-02 -6.82810187e-01 4.68792945e-01
5.69447637e-01 -3.24953884e-01 -4.58773553e-01 -3.82358402e-01
-5.59706151e-01 2.45114133e-01 9.61741284e-02 3.12269598e-01
3.25863659e-01 -1.01265132e+00 -1.09087563e+00 -7.15914369e-02
1.02212571e-01 -2.07309276e-01 -4.32680640e-03 9.20013666e-01
-2.85981953e-01 8.01886678e-01 1.90045595e-01 -5.29817820e-01
-1.59311175e+00 -3.57834399e-02 -1.73171610e-01 -6.27397060e-01
-6.63342774e-01 8.12536895e-01 6.30367994e-02 -1.85766548e-01
3.50786120e-01 -5.19054830e-01 -7.12694675e-02 3.03379357e-01
5.92763126e-01 7.14673340e-01 4.50014442e-01 -1.62943885e-01
-3.12677503e-01 6.16664812e-03 -3.81948739e-01 -2.38122731e-01
1.19626427e+00 2.74893522e-01 1.64313391e-01 3.34199041e-01
1.14750862e+00 1.31919431e-02 -6.88260138e-01 -1.15211338e-01
-5.06896712e-03 -2.87797544e-02 2.02621266e-01 -1.20153010e+00
-4.74057168e-01 8.12306464e-01 1.11453660e-01 1.86792031e-01
8.49543929e-01 2.69806781e-03 1.15639949e+00 6.37205660e-01
2.30059307e-02 -7.61497974e-01 3.05725262e-03 7.50874996e-01
9.34460878e-01 -1.11463332e+00 8.39431733e-02 -8.23968500e-02
-6.38121903e-01 7.96771884e-01 3.51533771e-01 3.48092392e-02
-9.29774344e-02 2.22279832e-01 -4.47578669e-01 7.44297057e-02
-1.06883407e+00 1.81443989e-02 3.82316232e-01 1.11014143e-01
5.49357712e-01 -6.47624880e-02 -4.86518443e-01 5.86505890e-01
-4.90608454e-01 -5.57621568e-02 7.76530862e-01 7.99741924e-01
-6.32571578e-01 -7.80760944e-01 -2.75118709e-01 1.15252447e+00
-8.55127394e-01 -6.24554336e-01 -2.11919934e-01 5.83116889e-01
-5.89100540e-01 9.58062828e-01 2.80259371e-01 -9.52166021e-02
2.96613544e-01 5.40128827e-01 4.25267875e-01 -9.33814824e-01
-7.60417581e-01 -1.08405553e-01 7.64078140e-01 -2.87498444e-01
1.06637865e-01 -9.36519027e-01 -1.32682025e+00 -1.52687460e-01
-4.43591982e-01 2.52195060e-01 6.48366153e-01 9.04590189e-01
4.45711046e-01 6.65330589e-01 3.80326658e-01 -8.27187359e-01
-9.48536396e-01 -1.23622465e+00 -4.00448255e-02 -3.59188616e-02
3.35049927e-01 -2.89092928e-01 -5.26250541e-01 8.68893135e-03] | [12.251676559448242, 9.259209632873535] |
0f7b2787-0a8f-432d-a874-5e525b363980 | joint-non-parametric-point-process-model-for | 2209.04142 | null | https://arxiv.org/abs/2209.04142v6 | https://arxiv.org/pdf/2209.04142v6.pdf | Causal Modeling of Policy Interventions From Sequences of Treatments and Outcomes | A treatment policy defines when and what treatments are applied to affect some outcome of interest. Data-driven decision-making requires the ability to predict what happens if a policy is changed. Existing methods that predict how the outcome evolves under different scenarios assume that the tentative sequences of future treatments are fixed in advance, while in practice the treatments are determined stochastically by a policy and may depend, for example, on the efficiency of previous treatments. Therefore, the current methods are not applicable if the treatment policy is unknown or a counterfactual analysis is needed. To handle these limitations, we model the treatments and outcomes jointly in continuous time, by combining Gaussian processes and point processes. Our model enables the estimation of a treatment policy from observational sequences of treatments and outcomes, and it can predict the interventional and counterfactual progression of the outcome after an intervention on the treatment policy (in contrast with the causal effect of a single treatment). We show with real-world and semi-synthetic data on blood glucose progression that our method can answer causal queries more accurately than existing alternatives. | ['Pekka Marttinen', 'Kirsi Pietiläinen', 'Tuure Saarinen', 'Anne Juuti', 'ST John', 'Çağlar Hızlı'] | 2022-09-09 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 5.15422165e-01 7.26304427e-02 -5.98672569e-01 -3.32147211e-01
-3.55690897e-01 -6.05499387e-01 8.10890019e-01 4.95828718e-01
-4.64260340e-01 1.29321492e+00 6.97140276e-01 -8.06919396e-01
-3.83320570e-01 -1.03097188e+00 -8.36975157e-01 -6.84468806e-01
-1.87222794e-01 9.61567700e-01 -1.09788924e-01 4.00088727e-01
1.41653538e-01 2.34097570e-01 -1.28888845e+00 1.43954426e-01
9.79128242e-01 4.13342983e-01 1.92517117e-02 6.65408969e-01
-1.56653821e-01 3.85607898e-01 -3.76521111e-01 -4.03146774e-01
9.58444849e-02 -5.97210050e-01 -5.13992190e-01 6.13356829e-02
-2.69399375e-01 -3.53263676e-01 -1.86077982e-01 8.26047242e-01
5.06179869e-01 -1.68715101e-02 1.07584095e+00 -1.13368464e+00
-4.63062882e-01 8.02439988e-01 -4.81755257e-01 -6.96664825e-02
1.70838788e-01 4.85246986e-01 5.24539828e-01 1.23268366e-01
3.19778800e-01 1.55865848e+00 4.17110205e-01 4.82148588e-01
-1.80278075e+00 -3.31936419e-01 3.96327138e-01 2.69041061e-02
-7.49747336e-01 -4.78937775e-01 2.79288173e-01 -7.22102523e-01
1.75510854e-01 1.22218959e-01 6.49360716e-01 1.42036974e+00
7.13536859e-01 3.62299293e-01 1.33620012e+00 -2.92706072e-01
6.85863972e-01 -1.72771618e-01 -5.59346303e-02 6.04645871e-02
5.56655645e-01 6.60727084e-01 7.77941756e-03 -7.83596873e-01
5.88472724e-01 3.82289082e-01 -3.64684999e-01 -3.90907735e-01
-1.10237014e+00 8.72039676e-01 -7.73037672e-02 -2.33720466e-01
-1.22531772e+00 3.23384047e-01 4.58063871e-01 1.12454899e-01
4.50661957e-01 2.01046348e-01 -8.29661191e-01 -1.09490275e-01
-7.29631603e-01 5.64392388e-01 7.93737233e-01 4.27299500e-01
2.68758744e-01 -3.04889411e-01 -7.41880417e-01 2.24452421e-01
-2.23446786e-02 8.88969302e-01 3.63849580e-01 -1.06729066e+00
2.24600837e-01 6.20231731e-03 1.04088676e+00 -2.93020219e-01
-3.28624636e-01 -6.91961721e-02 -6.92200303e-01 3.50849591e-02
9.03672636e-01 -6.65936112e-01 -1.15097773e+00 1.91200101e+00
2.66445071e-01 3.89686137e-01 -1.41925454e-01 5.69199204e-01
-7.93442279e-02 4.47091669e-01 6.19030118e-01 -9.16873932e-01
1.21953917e+00 1.66784242e-01 -8.63953471e-01 1.41057730e-01
6.88134193e-01 -5.23370087e-01 7.12608159e-01 1.86703727e-01
-1.09408069e+00 -7.80908614e-02 -5.26382625e-02 8.07946205e-01
6.67120442e-02 -3.22485805e-01 6.21245921e-01 5.73004663e-01
-6.57318771e-01 7.42888629e-01 -9.30757046e-01 -4.44885820e-01
3.94059867e-01 1.57365203e-01 2.84287035e-01 -8.93681049e-02
-1.16689003e+00 7.93661296e-01 2.38252610e-01 -1.45048782e-01
-1.08442223e+00 -1.01403987e+00 -2.94604003e-01 3.32911491e-01
6.77044868e-01 -1.22388268e+00 1.45740497e+00 -8.48842859e-01
-1.41296101e+00 3.24741334e-01 -3.17416102e-01 -8.82072568e-01
8.74775589e-01 2.46378288e-01 -2.57363856e-01 -2.89056301e-01
2.38888443e-01 1.92309603e-01 6.15314841e-01 -7.62756467e-01
-8.56935740e-01 -5.96744061e-01 3.11596990e-02 2.54062861e-01
3.23398858e-01 -1.22380592e-01 1.09285347e-01 -5.29485881e-01
-4.27868217e-01 -1.09565616e+00 -6.30054891e-01 -1.35289893e-01
-4.99815792e-01 -4.00335751e-02 2.35123113e-01 -3.59151363e-01
9.50753272e-01 -1.81515157e+00 -5.80039173e-02 -9.68511477e-02
-1.98015198e-01 -2.29945228e-01 1.56713620e-01 5.31983078e-01
-1.63168401e-01 3.23744416e-01 -3.23694915e-01 3.32476020e-01
-7.19539374e-02 3.77444893e-01 -3.99946481e-01 7.21142590e-01
-2.09045038e-01 5.62571347e-01 -8.69641840e-01 -1.06323719e-01
2.66848087e-01 -2.27877460e-02 -5.20007730e-01 3.06947948e-03
-6.02466762e-01 6.98890746e-01 -7.62253642e-01 -5.65677807e-02
5.44679463e-01 -2.31638372e-01 5.88818312e-01 3.01043004e-01
-2.48713598e-01 2.10639119e-01 -1.19600105e+00 9.62986410e-01
-3.84167105e-01 -5.04681617e-02 -3.58303279e-01 -1.29464257e+00
3.07152957e-01 5.99660456e-01 7.62444139e-01 -5.12338877e-01
2.35997755e-02 -2.17875503e-02 3.14357311e-01 -6.21214449e-01
-3.67665201e-01 -6.44320130e-01 7.63014751e-03 7.59246409e-01
-4.46244657e-01 1.64593145e-01 8.61737207e-02 -1.78305745e-01
1.13863671e+00 6.53436556e-02 6.04473889e-01 -3.86849016e-01
-1.54325711e-02 9.30305123e-02 1.01988006e+00 1.14694822e+00
-3.11143603e-02 1.53748453e-01 8.60510528e-01 -4.70547557e-01
-1.01189804e+00 -1.19045413e+00 -3.46570462e-01 4.88027751e-01
-2.10967645e-01 4.06314582e-01 -3.19850475e-01 -6.31523073e-01
4.26357478e-01 1.25923419e+00 -8.70960593e-01 -1.86633199e-01
-2.22030893e-01 -1.40142655e+00 -2.20320791e-01 2.08433673e-01
3.59606706e-02 -8.91948640e-01 -4.61411297e-01 6.50359213e-01
-1.08040702e-02 -5.56549370e-01 -3.68865997e-01 -3.01225841e-01
-1.04863369e+00 -1.39632082e+00 -8.31180215e-01 1.18077986e-01
3.97930831e-01 -1.10570610e-01 9.85819101e-01 -5.97023308e-01
1.53763726e-01 2.83699632e-01 1.48769081e-01 -9.25950706e-01
-4.94233847e-01 -4.38113898e-01 3.00238878e-01 2.14541584e-01
2.30796382e-01 -5.35385966e-01 -1.17632592e+00 -6.25038892e-02
-7.73035407e-01 1.68703701e-02 2.59202123e-01 6.87820196e-01
4.64738131e-01 1.39963448e-01 7.65116870e-01 -1.30883098e+00
7.33972251e-01 -7.58104682e-01 -7.90085316e-01 5.14817357e-01
-8.49127173e-01 3.42790812e-01 5.82331359e-01 -7.65768409e-01
-1.43626428e+00 -6.67361245e-02 5.44911563e-01 -7.38560110e-02
-3.86047930e-01 6.39175415e-01 -1.94746837e-01 9.49843466e-01
5.11592329e-01 -2.29218658e-02 -8.25354978e-02 -6.09204829e-01
5.12416422e-01 4.58607256e-01 3.53697747e-01 -8.98438931e-01
1.11112922e-01 6.52794003e-01 2.43586063e-01 -1.70816004e-01
-6.26820266e-01 1.13783017e-01 -2.69505769e-01 6.29461482e-02
7.22476006e-01 -8.55673850e-01 -1.22969520e+00 3.37754726e-01
-9.68232632e-01 -4.53065664e-01 -5.66582203e-01 9.18303311e-01
-8.87407362e-01 -1.38475493e-01 -1.33810684e-01 -1.09459734e+00
9.15440992e-02 -1.09617746e+00 5.35819232e-01 1.10814564e-01
-2.72442460e-01 -1.30634344e+00 1.73417181e-01 -2.10935339e-01
2.61889428e-01 5.06877840e-01 1.32429314e+00 -4.62537199e-01
-4.30039316e-01 -1.40117154e-01 1.61222994e-01 -2.46762037e-01
5.98937273e-01 -1.20290443e-02 -3.36910278e-01 -2.28232443e-01
5.30242026e-02 3.89677256e-01 4.87352818e-01 1.44189763e+00
1.24100792e+00 -7.73766041e-01 -6.63622856e-01 1.89563379e-01
1.21440446e+00 7.13851452e-01 3.83306533e-01 2.63776742e-02
1.05743498e-01 7.38066256e-01 3.92698765e-01 6.86420977e-01
3.66584629e-01 7.13435054e-01 2.15017945e-01 5.50914779e-02
4.30280805e-01 -3.95457208e-01 1.11550629e-01 -3.78013492e-01
5.66140898e-02 -4.15001661e-01 -8.97589743e-01 7.01063871e-01
-1.95359147e+00 -1.15438187e+00 -4.77297783e-01 2.99374056e+00
9.10630882e-01 1.17168278e-01 5.67160904e-01 -3.80543709e-01
8.61169279e-01 -3.64347458e-01 -9.90064204e-01 -5.24316132e-01
8.79520327e-02 -8.75944570e-02 1.01802802e+00 4.86143678e-01
-7.41294265e-01 3.28032672e-01 7.10951471e+00 2.70984620e-01
-1.02050304e+00 7.79489242e-03 1.07729948e+00 -8.76920819e-02
-3.29659283e-01 2.44931757e-01 -4.89025652e-01 8.32951009e-01
1.49330688e+00 -8.74764085e-01 1.98430941e-01 5.20695336e-02
1.22445691e+00 -2.95611382e-01 -1.32438874e+00 3.08978230e-01
-8.11123371e-01 -1.26739156e+00 -5.50113060e-02 3.55471909e-01
7.77812243e-01 -3.03357303e-01 2.74398588e-02 3.17703411e-02
1.39802980e+00 -6.64242029e-01 4.12156284e-01 1.13474250e+00
7.13199139e-01 -7.43559241e-01 6.68343186e-01 5.51106274e-01
-3.04137737e-01 -2.93752283e-01 -6.08534329e-02 -2.78482407e-01
5.07825613e-01 9.47363436e-01 -9.91313398e-01 3.99888009e-01
2.22322717e-01 3.38690400e-01 6.99719563e-02 1.28343844e+00
-1.97397247e-01 1.12304342e+00 -1.81239337e-01 2.43588403e-01
-5.84058184e-03 -4.06788886e-01 4.50132400e-01 6.60502017e-01
5.94134748e-01 2.54847795e-01 2.47499049e-01 7.55048156e-01
2.67054856e-01 -4.01388928e-02 -6.29036486e-01 -9.21307430e-02
4.75439847e-01 4.04182762e-01 -6.18789434e-01 -4.20833558e-01
-5.22940755e-01 5.35922647e-01 -3.10874045e-01 7.76894391e-01
-6.10234439e-01 4.74940598e-01 8.52128446e-01 3.65206093e-01
2.43210886e-02 3.06240350e-01 -4.68076199e-01 -9.69869196e-01
-3.01020235e-01 -7.76766479e-01 6.87255740e-01 -6.01790190e-01
-1.58036268e+00 -3.37293684e-01 3.32420230e-01 -8.38038802e-01
-5.11491656e-01 -1.53468281e-01 -6.51051819e-01 1.35526979e+00
-1.05663693e+00 -3.89790982e-01 6.32828236e-01 1.68143943e-01
4.03165370e-01 3.79032433e-01 7.65051544e-01 -3.14969510e-01
-5.57139218e-01 -2.55435467e-01 5.48905313e-01 -2.50320792e-01
6.49066865e-01 -1.26740980e+00 2.01410159e-01 4.44329232e-01
-4.27241743e-01 2.85585761e-01 1.20081544e+00 -9.66608405e-01
-8.86184812e-01 -1.03457153e+00 6.74900055e-01 -3.12375486e-01
8.52530420e-01 -4.09404784e-02 -8.66660297e-01 6.82188809e-01
-9.53593254e-02 -2.88401306e-01 4.26066756e-01 2.66112149e-01
2.74578243e-01 -7.74137080e-02 -1.23102260e+00 8.20471823e-01
8.08357120e-01 -1.36899054e-01 -5.09300709e-01 6.18357241e-01
7.08778441e-01 5.97103164e-02 -6.57647908e-01 4.05159950e-01
4.96729404e-01 -5.71284533e-01 8.11101913e-01 -1.32989872e+00
2.90880024e-01 3.56528386e-02 1.09713919e-01 -1.77772784e+00
-2.30484262e-01 -6.47095025e-01 1.15577221e-01 1.02775669e+00
3.47334087e-01 -9.77150977e-01 5.10904789e-01 1.04689443e+00
6.08080149e-01 -3.77643734e-01 -1.08679068e+00 -6.68681562e-01
3.40541452e-01 -2.20900521e-01 1.09994745e+00 9.00551021e-01
-3.01663905e-01 1.76898941e-01 -3.87863308e-01 3.18305939e-01
7.58955836e-01 4.12225068e-01 4.57394838e-01 -1.32027555e+00
-3.39376062e-01 -3.36053669e-01 -3.33442278e-02 -5.38615882e-01
1.76955700e-01 -3.43060017e-01 -1.85754731e-01 -1.47862923e+00
6.76239848e-01 -4.76173222e-01 -2.85048187e-01 3.02944094e-01
-5.65269411e-01 -8.16898525e-01 1.97960082e-02 2.90380083e-02
1.35361314e-01 6.18488848e-01 1.13464200e+00 5.80376387e-02
-4.48431671e-01 7.86086559e-01 -7.37585425e-01 6.60017908e-01
6.72144890e-01 -7.54539311e-01 -6.47263646e-01 4.49855141e-02
-8.95467177e-02 8.92463267e-01 3.61965656e-01 -3.04274946e-01
-1.05294824e-01 -1.05893934e+00 1.85413003e-01 -1.52630076e-01
-3.59602273e-01 -7.00228810e-01 7.94899583e-01 8.65580022e-01
-6.58721268e-01 2.32758597e-02 -4.93421103e-04 1.18447793e+00
3.72243375e-01 -6.29585981e-02 4.91685659e-01 -1.34827062e-01
1.17887894e-03 5.71139097e-01 -6.88206613e-01 4.19777958e-03
8.99343431e-01 2.02845126e-01 -2.30447233e-01 -8.74794900e-01
-1.17549193e+00 4.57312763e-01 4.29238826e-01 2.76219696e-01
2.02488348e-01 -1.13962102e+00 -1.01928627e+00 -3.05437237e-01
-2.29593396e-01 -3.95798355e-01 3.35763395e-01 7.54224181e-01
9.46990103e-02 5.07498860e-01 1.03330232e-01 -3.43289703e-01
-8.80900621e-01 9.99839902e-01 6.18308187e-01 -3.49542528e-01
-5.52109361e-01 -1.45037457e-01 7.82474697e-01 -1.25090748e-01
-1.76900610e-01 -5.68099357e-02 -7.48538375e-02 5.03889322e-02
5.02903819e-01 3.49901706e-01 -3.39296281e-01 -4.25893292e-02
8.58062059e-02 -1.01863153e-01 8.54746252e-02 -4.65719461e-01
1.44165468e+00 -4.81067389e-01 6.27139658e-02 8.46639931e-01
5.17301083e-01 -4.09299195e-01 -1.51407814e+00 -4.09443021e-01
7.58294538e-02 -4.81201023e-01 -4.45045792e-02 -1.07376409e+00
-6.78681016e-01 5.99418700e-01 7.77611434e-01 2.98655868e-01
1.10297036e+00 -2.04426676e-01 1.11920878e-01 -3.06677878e-01
3.38260978e-01 -6.96514726e-01 -7.38211572e-01 -1.52982846e-01
7.50746608e-01 -9.31686103e-01 -5.05111068e-02 -2.07931355e-01
-4.39411193e-01 6.62258983e-01 -1.09107867e-02 2.32439891e-01
8.32886875e-01 1.02282271e-01 -1.96395293e-01 -3.06183230e-02
-1.16028333e+00 -2.08659217e-01 5.68181910e-02 5.49596548e-01
3.81138861e-01 8.25526178e-01 -8.70316446e-01 2.54566073e-01
3.98798101e-03 5.43295979e-01 6.74334884e-01 6.48782730e-01
6.94668517e-02 -1.15329969e+00 -5.76583862e-01 9.61901307e-01
-6.90955281e-01 -3.79846478e-03 7.31177554e-02 6.84928596e-01
4.69373651e-02 7.56753802e-01 4.22846854e-01 6.09401643e-01
6.31453633e-01 3.33890647e-01 2.94260770e-01 -6.34507298e-01
6.99643865e-02 1.52088746e-01 7.45462850e-02 -3.77863586e-01
-3.40124726e-01 -1.24042428e+00 -9.50541079e-01 -5.43784499e-01
1.05416164e-01 8.01133588e-02 3.81463587e-01 1.15254486e+00
4.51668680e-01 5.11828244e-01 7.88122296e-01 -3.78696829e-01
-1.04717422e+00 -7.51216948e-01 -6.65369928e-01 5.65898597e-01
2.22290754e-01 -7.11975694e-01 -3.00346375e-01 1.34155795e-01] | [8.037847518920898, 5.340810298919678] |
b7fc8ea6-562d-42a7-9af4-9f197c90ae13 | knowledge-driven-slot-constraints-for-goal | null | null | https://aclanthology.org/2021.naacl-main.266 | https://aclanthology.org/2021.naacl-main.266.pdf | Knowledge-Driven Slot Constraints for Goal-Oriented Dialogue Systems | In goal-oriented dialogue systems, users provide information through slot values to achieve specific goals. Practically, some combinations of slot values can be invalid according to external knowledge. For example, a combination of {``}cheese pizza{''} (a menu item) and {``}oreo cookies{''} (a topping) from an input utterance {``}Can I order a cheese pizza with oreo cookies on top?{''} exemplifies such invalid combinations according to the menu of a restaurant business. Traditional dialogue systems allow execution of validation rules as a post-processing step after slots have been filled which can lead to error accumulation. In this paper, we formalize knowledge-driven slot constraints and present a new task of constraint violation detection accompanied with benchmarking data. Then, we propose methods to integrate the external knowledge into the system and model constraint violation detection as an end-to-end classification task and compare it to the traditional rule-based pipeline approach. Experiments on two domains of the MultiDoGO dataset reveal challenges of constraint violation detection and sets the stage for future work and improvements. | ['Saab Mansour', 'Daniele Bonadiman', 'Piyawat Lertvittayakumjorn'] | 2021-06-01 | null | null | null | naacl-2021-4 | ['goal-oriented-dialogue-systems'] | ['natural-language-processing'] | [ 2.06308559e-01 5.07235110e-01 4.33366224e-02 -7.32437372e-01
-4.94410098e-01 -7.70118833e-01 3.98890615e-01 6.18869841e-01
-3.42749774e-01 7.52830923e-01 5.43021113e-02 -6.01818383e-01
-3.48369092e-01 -7.88044274e-01 -4.42434847e-01 -7.29762316e-02
1.14001758e-01 1.08876657e+00 6.84544981e-01 -8.99309635e-01
2.67595619e-01 -1.83403119e-01 -1.43891704e+00 9.19085741e-01
5.23163617e-01 8.14150691e-01 1.78255141e-01 4.82820153e-01
-4.76886034e-01 5.73067367e-01 -9.26460147e-01 -4.01438028e-01
1.42749414e-01 -4.23217505e-01 -1.28803110e+00 8.12550113e-02
-1.03290461e-01 -2.62307435e-01 5.05448222e-01 1.06052864e+00
1.74849868e-01 4.23970371e-01 3.38908434e-01 -1.32694590e+00
-9.22204256e-02 9.40725267e-01 1.59751087e-01 -1.14887871e-01
8.43209863e-01 3.15933347e-01 1.29941297e+00 -4.90037143e-01
8.27501118e-01 1.17808485e+00 4.88509834e-01 6.47967100e-01
-9.82990444e-01 -1.48300916e-01 4.28623348e-01 2.31387734e-01
-9.35848117e-01 -4.85771656e-01 5.48605323e-01 -4.79673237e-01
1.39858603e+00 6.83725774e-01 4.95572597e-01 7.47497320e-01
-3.52535963e-01 8.95094633e-01 7.49895751e-01 -5.78428149e-01
4.52393800e-01 1.65507585e-01 2.66785264e-01 7.03799129e-01
-1.67779505e-01 -1.61835983e-01 -6.73218310e-01 -2.52747774e-01
2.79934645e-01 -6.15669012e-01 1.32871987e-02 -2.30961293e-01
-9.56347823e-01 9.54038203e-01 -2.23629609e-01 1.82984620e-01
-1.50963992e-01 -4.69441444e-01 8.20795476e-01 5.45657694e-01
1.32563058e-03 6.77815735e-01 -1.00112903e+00 -3.19032043e-01
-6.33960545e-01 7.17521250e-01 1.47148132e+00 1.23664284e+00
5.21355391e-01 -4.49469626e-01 -3.48320931e-01 1.06467760e+00
4.24862981e-01 -1.60614669e-01 2.37331256e-01 -9.57623065e-01
4.88092422e-01 8.76823425e-01 7.47498572e-01 -5.21593750e-01
-8.91454220e-01 1.25362203e-01 -1.85570642e-01 1.37131736e-01
1.27387333e+00 -2.24699914e-01 -5.63414633e-01 1.39376581e+00
5.88042200e-01 -2.81959683e-01 1.85963452e-01 1.23600900e+00
9.28343236e-01 3.05175543e-01 1.45045817e-01 -3.00128460e-01
1.93063366e+00 -9.46502984e-01 -9.37895417e-01 -2.09284127e-01
8.62380326e-01 -9.24890459e-01 1.28139865e+00 7.47624397e-01
-9.90050495e-01 -2.93455511e-01 -7.45169282e-01 1.25819463e-02
-4.99409586e-01 -1.87594846e-01 7.93991625e-01 5.80800176e-01
-5.05372822e-01 5.52339137e-01 -5.74187338e-01 -4.63068426e-01
-3.68483335e-01 2.73993969e-01 -2.55526274e-01 1.58907637e-01
-1.69619083e+00 9.41878438e-01 7.38196671e-01 2.73859411e-01
-3.67260307e-01 -4.91603583e-01 -1.06635928e+00 6.50211722e-02
1.13836360e+00 -2.52566367e-01 1.86019075e+00 -6.07928753e-01
-1.82143462e+00 8.17538738e-01 5.21930680e-02 -3.45563978e-01
7.07461119e-01 -1.45653009e-01 -4.53325957e-01 -1.87265053e-01
2.71339923e-01 1.35337114e-01 3.22282761e-01 -8.27697635e-01
-1.09793377e+00 -2.49555126e-01 6.30733550e-01 2.12588668e-01
4.75329161e-01 5.30773461e-01 -5.55806935e-01 -9.66138020e-02
4.17144358e-01 -7.62985229e-01 -1.74776077e-01 -4.33327705e-01
-7.57607400e-01 -6.08040869e-01 2.22209528e-01 -6.16257727e-01
1.62987995e+00 -1.78295302e+00 -9.75732051e-04 3.89741093e-01
-2.14561969e-01 4.18015987e-01 2.84564167e-01 6.54938161e-01
-2.49831267e-02 7.73095489e-02 -2.89788783e-01 1.64946973e-01
5.93208134e-01 4.50271755e-01 1.65056989e-01 -1.29037142e-01
2.72169441e-01 5.07368505e-01 -1.08826399e+00 -3.51494342e-01
3.43315870e-01 -1.70759559e-01 -7.54452050e-01 2.07855403e-01
-1.10001123e+00 3.46462607e-01 -4.59230959e-01 6.23992682e-01
4.29196566e-01 6.82429969e-02 9.03577268e-01 -1.49454966e-01
-3.77794415e-01 7.59647965e-01 -1.42910457e+00 1.77611804e+00
-1.93231583e-01 7.07674176e-02 3.28164041e-01 -9.84025240e-01
9.52612817e-01 3.66195470e-01 2.11518496e-01 -6.49698555e-01
1.49489775e-01 3.04217398e-01 3.37717146e-01 -9.06503737e-01
5.65679789e-01 -1.18894793e-01 -5.65127730e-01 2.42642179e-01
-3.32834929e-01 -2.63241351e-01 6.21254086e-01 -8.85332003e-02
1.21037471e+00 6.71836317e-01 4.26993042e-01 -1.76390052e-01
8.53311181e-01 5.42314768e-01 7.13945031e-01 7.89936543e-01
-3.77490580e-01 2.56728083e-01 9.30905223e-01 -5.97027898e-01
-9.08585131e-01 -5.35663128e-01 -6.81563392e-02 1.69495571e+00
-1.12658106e-01 -7.50317037e-01 -5.56471467e-01 -8.04644704e-01
-1.80352051e-02 9.15443242e-01 -4.20550913e-01 2.65746593e-01
-5.93767047e-01 -5.03759563e-01 5.58344305e-01 1.03941765e-02
4.39754158e-01 -1.20774734e+00 -1.04962695e+00 5.67829728e-01
-6.80048943e-01 -1.05922115e+00 -2.34695390e-01 5.88338256e-01
-2.31514394e-01 -1.32668340e+00 1.29724160e-01 -5.80147743e-01
1.88520506e-01 -4.11271781e-01 1.39717150e+00 5.08319378e-01
-5.70227802e-02 -2.73972526e-02 -7.58105278e-01 -4.17966694e-01
-5.49396038e-01 -1.26996562e-02 -1.05147608e-01 -1.78157166e-01
9.16995823e-01 -4.35529612e-02 -3.43037307e-01 5.97119629e-01
-7.93040395e-01 1.19037077e-01 -6.24793507e-02 9.70890284e-01
4.05701548e-01 4.98359278e-02 3.60307068e-01 -1.40317011e+00
8.65375400e-01 -5.38671255e-01 -7.23007977e-01 2.87710160e-01
-3.93305808e-01 2.72386938e-01 5.75787544e-01 -3.62199135e-02
-1.05330980e+00 2.41688520e-01 -5.03795147e-01 2.41329327e-01
-5.37774444e-01 8.83704662e-01 -3.91951680e-01 8.07147920e-01
7.70941496e-01 -2.24422261e-01 -7.93923065e-02 -6.60446703e-01
1.87113687e-01 5.80906570e-01 5.99055171e-01 -9.83429134e-01
2.08407834e-01 -5.90645038e-02 -3.29594344e-01 -5.25109708e-01
-9.61981714e-01 -5.78331292e-01 -6.56774700e-01 -1.12545572e-01
7.23458767e-01 -4.75534171e-01 -1.31714332e+00 6.05191924e-02
-1.11386013e+00 -5.57365954e-01 -1.75778821e-01 5.21824807e-02
-7.05530524e-01 4.64645982e-01 -3.82839441e-01 -1.05717635e+00
-1.43319875e-01 -1.06719053e+00 8.14250767e-01 2.21359178e-01
-6.18883550e-01 -7.67007947e-01 -2.52737075e-01 4.90214407e-01
2.51407206e-01 1.81158811e-01 1.00430667e+00 -1.34273565e+00
-4.33201268e-02 2.77768821e-02 1.92689061e-01 -2.28668358e-02
-2.50398903e-03 -1.61328856e-02 -5.52865267e-01 9.09670591e-02
-3.57237548e-01 -3.08331043e-01 2.03012004e-01 -1.46529805e-02
4.26302403e-01 -3.76526713e-01 -5.55002503e-02 -3.13658156e-02
1.06269753e+00 4.32060242e-01 4.15623754e-01 7.45841384e-01
1.45023956e-03 1.05533338e+00 1.26811194e+00 8.41334879e-01
5.46075761e-01 1.20108366e+00 3.32914323e-01 3.78734499e-01
2.88527280e-01 -4.94768284e-02 1.87881052e-01 7.55223073e-03
1.70064718e-01 -1.19592667e-01 -1.00133729e+00 4.84662384e-01
-2.30215955e+00 -8.93133581e-01 -4.86540377e-01 2.04296303e+00
1.14069545e+00 4.03534204e-01 3.14078003e-01 3.38242948e-01
5.16150594e-01 -2.06156462e-01 -4.65161473e-01 -9.02851582e-01
2.16287360e-01 1.56942025e-01 -2.33107544e-02 8.83900642e-01
-1.11740053e+00 1.36632586e+00 4.97137022e+00 3.87060970e-01
-6.42879188e-01 3.77916358e-02 1.82230212e-02 1.34769514e-01
-1.38539076e-01 2.21337244e-01 -9.47483242e-01 4.68589991e-01
7.71146297e-01 1.68790042e-01 3.58740687e-01 7.27055430e-01
3.61591935e-01 -5.40185273e-01 -1.45444310e+00 5.47922730e-01
-3.50652605e-01 -1.05295205e+00 -3.62508476e-01 -4.22456414e-01
-1.04884796e-01 -4.29009140e-01 -5.35226405e-01 7.58320391e-01
6.89398825e-01 -7.14270353e-01 8.11809361e-01 3.44553977e-01
3.39914292e-01 -5.24476767e-01 7.52189875e-01 6.30880833e-01
-6.97425425e-01 -2.52746884e-02 -3.48166913e-01 -2.92709619e-01
2.44277120e-01 2.95368612e-01 -1.34176898e+00 7.98545599e-01
7.09361732e-01 -1.32180870e-01 -1.49503320e-01 8.00746381e-01
-5.50702870e-01 3.45386088e-01 -7.21446157e-01 -2.11623043e-01
2.68736362e-01 -1.77718803e-01 6.74869001e-01 1.36664176e+00
-2.26695046e-01 4.12029773e-01 3.90920550e-01 7.89382041e-01
3.90544176e-01 2.99316555e-01 -1.93098530e-01 2.64849663e-01
5.23470581e-01 1.11639583e+00 -6.96819544e-01 -3.65883559e-01
-4.10986632e-01 8.96523416e-01 2.19182760e-01 8.55183005e-02
-6.85538650e-01 -2.48766825e-01 6.78011239e-01 6.17530197e-03
2.20581502e-01 9.49862525e-02 -8.21480229e-02 -9.44280028e-01
4.03855368e-02 -1.13935602e+00 7.57614136e-01 -4.47759390e-01
-9.19913650e-01 4.05511111e-01 1.07854195e-01 -1.03418016e+00
-4.99712110e-01 -5.96861005e-01 -2.70910770e-01 8.97458613e-01
-1.38891482e+00 -9.38868880e-01 -7.90157616e-02 5.05815089e-01
8.06049049e-01 1.17714673e-01 1.30072618e+00 2.17647746e-01
-4.14392382e-01 5.02619386e-01 -4.60604191e-01 1.74233392e-01
6.83248758e-01 -1.34295928e+00 2.52088666e-01 6.11203313e-01
-1.78886443e-01 5.75331330e-01 1.25053895e+00 -7.09780753e-01
-1.22510898e+00 -6.00023448e-01 1.25366628e+00 -6.55719280e-01
4.68052685e-01 -2.59012759e-01 -1.15623438e+00 7.01541007e-01
1.01195641e-01 -3.97477627e-01 9.89360034e-01 5.24779022e-01
-3.64604563e-01 5.03076434e-01 -1.20031774e+00 3.61011833e-01
9.00306165e-01 -1.31827742e-01 -9.81424391e-01 5.90952098e-01
4.30414051e-01 -1.00675452e+00 -9.11399901e-01 2.50595212e-01
7.49796391e-01 -1.18010747e+00 6.19394720e-01 -1.19474590e+00
2.39997283e-01 -5.27359724e-01 -1.23438135e-01 -1.16352844e+00
-2.36902982e-02 -7.43738353e-01 1.19669996e-01 1.09084439e+00
8.26546609e-01 -7.64689520e-02 6.56336129e-01 1.43383503e+00
-5.29247224e-01 -3.51890117e-01 -1.06283987e+00 -5.40055037e-01
-1.91113740e-01 -7.79395819e-01 8.28833103e-01 1.12942362e+00
9.77573812e-01 3.64746630e-01 -4.55442071e-01 1.94872811e-01
1.08209297e-01 5.71598485e-02 8.16501021e-01 -1.25905967e+00
-4.43488777e-01 -3.63284588e-01 1.90667421e-01 -1.00952029e+00
-7.90369604e-03 -8.73699129e-01 3.46015215e-01 -1.61795187e+00
-4.31176901e-01 -8.50244403e-01 -1.40549336e-02 1.03124416e+00
-4.00309712e-02 -6.50933459e-02 3.50480795e-01 -1.72912344e-01
-8.40665162e-01 -2.69789398e-01 9.15504873e-01 -1.19796433e-01
-6.82286203e-01 1.85196757e-01 -5.94445527e-01 7.38916755e-01
6.95103943e-01 -1.87275901e-01 -1.56283423e-01 -3.06367636e-01
8.30992758e-01 4.80218023e-01 -1.08887576e-01 -4.27203983e-01
4.22191381e-01 -5.22411644e-01 -1.52310997e-01 -3.85967970e-01
3.09105277e-01 -9.95901823e-01 1.56940490e-01 2.50884533e-01
-4.50297147e-01 -2.94374973e-01 2.95476109e-01 1.20142788e-01
-1.42687857e-01 -6.67310596e-01 4.72103000e-01 -4.74678785e-01
-8.76295328e-01 -5.01218498e-01 -6.27499878e-01 -3.70652974e-02
8.47640574e-01 -1.23311721e-01 -4.28504854e-01 -1.22417420e-01
-1.41840851e+00 7.93673754e-01 1.44369379e-02 6.55794680e-01
3.61717820e-01 -8.09516370e-01 -5.80585897e-01 3.56256783e-01
3.40968043e-01 3.75318944e-01 9.87270847e-03 5.68094492e-01
-4.91768509e-01 4.52295512e-01 -3.78844708e-01 -4.26191688e-01
-1.20681679e+00 3.85015756e-01 3.99279565e-01 -5.40720820e-01
-4.89442706e-01 8.79163861e-01 -4.37735796e-01 -8.07919323e-01
3.79041433e-01 -6.31934345e-01 -5.87682068e-01 2.84086019e-01
4.04603869e-01 -6.59413114e-02 3.41860682e-01 -3.98835361e-01
-5.64773023e-01 -1.60126865e-01 4.11639782e-03 -9.98413488e-02
1.01125193e+00 -1.06300436e-01 2.25461461e-02 3.91538918e-01
2.84595549e-01 -1.27005383e-01 -7.98741341e-01 -4.11808431e-01
6.83861077e-01 -4.97897327e-01 -6.43374205e-01 -1.55082214e+00
-4.49030846e-01 4.11710143e-01 2.43808642e-01 7.09348023e-01
7.32433200e-01 -2.19760761e-02 5.01968265e-01 6.85184538e-01
4.85666960e-01 -1.70950270e+00 -3.00721407e-01 9.87254024e-01
9.77564931e-01 -1.29340625e+00 -2.23949879e-01 -6.59762740e-01
-8.76270413e-01 1.14901924e+00 7.50236332e-01 4.03290033e-01
4.03520823e-01 1.41549230e-01 2.11229876e-01 -3.27387601e-01
-1.01238012e+00 -4.28362131e-01 -1.51213989e-01 5.21525323e-01
5.85754991e-01 2.77711064e-01 -1.05112863e+00 1.12845671e+00
-4.07181114e-01 -1.58533491e-02 4.66732889e-01 1.10186017e+00
-6.50234401e-01 -1.58526969e+00 -4.40579683e-01 3.80759686e-01
-4.38059360e-01 -8.65926594e-02 -5.16876519e-01 7.03182518e-01
2.66255170e-01 1.36809385e+00 -8.19699466e-02 -1.94130138e-01
1.02197921e+00 7.83025980e-01 2.01516092e-01 -1.01861453e+00
-1.32264447e+00 3.19023162e-01 8.55304122e-01 -4.61913317e-01
-3.51396590e-01 -7.93276012e-01 -1.81597233e+00 1.06774247e-03
-4.12918985e-01 5.11393189e-01 5.29619873e-01 1.14223278e+00
7.02167079e-02 4.26096350e-01 1.54901147e-01 -2.94457465e-01
-6.41899228e-01 -1.07174945e+00 -3.23769450e-01 8.60227644e-01
-1.84698790e-01 -7.00307190e-01 -1.81488395e-01 1.59074560e-01] | [12.777647018432617, 7.7884721755981445] |
0ff71c0e-fc21-4d57-be2f-764829b07bd3 | knowledge-bridged-causal-interaction-network | 2212.02995 | null | https://arxiv.org/abs/2212.02995v1 | https://arxiv.org/pdf/2212.02995v1.pdf | Knowledge-Bridged Causal Interaction Network for Causal Emotion Entailment | Causal Emotion Entailment aims to identify causal utterances that are responsible for the target utterance with a non-neutral emotion in conversations. Previous works are limited in thorough understanding of the conversational context and accurate reasoning of the emotion cause. To this end, we propose Knowledge-Bridged Causal Interaction Network (KBCIN) with commonsense knowledge (CSK) leveraged as three bridges. Specifically, we construct a conversational graph for each conversation and leverage the event-centered CSK as the semantics-level bridge (S-bridge) to capture the deep inter-utterance dependencies in the conversational context via the CSK-Enhanced Graph Attention module. Moreover, social-interaction CSK serves as emotion-level bridge (E-bridge) and action-level bridge (A-bridge) to connect candidate utterances with the target one, which provides explicit causal clues for the Emotional Interaction module and Actional Interaction module to reason the target emotion. Experimental results show that our model achieves better performance over most baseline models. Our source code is publicly available at https://github.com/circle-hit/KBCIN. | ['Bing Qin', 'Zhuojun Li', 'Yanyan Zhao', 'Weixiang Zhao'] | 2022-12-06 | null | null | null | null | ['causal-emotion-entailment'] | ['natural-language-processing'] | [ 8.74680430e-02 8.85260761e-01 -2.51023322e-01 -7.66069055e-01
-5.61258614e-01 -3.75187427e-01 6.65771425e-01 1.60617679e-01
4.00917590e-01 6.69854879e-01 1.14304161e+00 -6.44336343e-02
-2.61604115e-02 -5.16647160e-01 -6.12842262e-01 -3.39783996e-01
-1.32830173e-01 1.58296302e-01 -2.24070743e-01 -5.01238406e-01
3.10992934e-02 -2.43423179e-01 -9.98364806e-01 7.51451790e-01
7.55991638e-01 6.91307843e-01 -2.17761680e-01 7.75255919e-01
-1.97997659e-01 1.63660610e+00 -4.40133393e-01 -7.63025284e-01
-5.64282596e-01 -1.07698810e+00 -1.34331524e+00 -1.67564303e-01
-3.36110026e-01 -3.75731885e-02 -4.06966895e-01 9.39470649e-01
2.15508446e-01 3.63461792e-01 4.32418615e-01 -1.82105863e+00
-5.55980802e-01 1.30537283e+00 -2.09763408e-01 3.58135626e-02
8.55088472e-01 2.13452756e-01 1.42322552e+00 -7.33346820e-01
7.09411740e-01 1.72667599e+00 3.83407146e-01 8.07997406e-01
-8.50092709e-01 -6.22094393e-01 4.98224556e-01 6.41245008e-01
-8.97342086e-01 -4.01179075e-01 1.26797879e+00 -1.47823825e-01
1.24759459e+00 4.24284816e-01 8.99360955e-01 1.45542634e+00
-7.25920871e-02 1.08230317e+00 6.82032764e-01 -2.83113718e-01
-1.72340497e-02 -2.04843894e-01 5.11902392e-01 7.08931744e-01
-5.73519945e-01 -2.94404298e-01 -8.54309440e-01 -3.56904685e-01
3.51015002e-01 -2.08249986e-01 -4.76187080e-01 2.17644274e-01
-9.63963389e-01 1.06057096e+00 6.63991034e-01 3.41231704e-01
-6.75304234e-01 5.04365623e-01 5.88390470e-01 1.69066072e-01
4.72211152e-01 3.38430375e-01 -6.32147312e-01 -3.76953334e-01
-1.72189474e-02 1.03021294e-01 1.24556446e+00 8.65731120e-01
6.60444498e-01 -3.71795744e-01 -3.46413821e-01 8.17309976e-01
3.38025868e-01 2.79438905e-02 2.84362789e-02 -1.08720183e+00
3.13667119e-01 9.87926781e-01 -1.79664582e-01 -1.21228993e+00
-5.00233710e-01 4.36317585e-02 -5.75226367e-01 -8.19733024e-01
2.65807398e-02 -4.90633786e-01 -1.85897425e-01 2.14611268e+00
6.69813633e-01 9.48978066e-01 4.17588681e-01 9.77017283e-01
1.25254226e+00 7.28091538e-01 3.95305485e-01 -4.74389851e-01
1.58316731e+00 -1.08489382e+00 -1.10154998e+00 -2.35065505e-01
7.51646221e-01 -5.17422676e-01 1.30752850e+00 -3.77741419e-02
-7.58598924e-01 -1.28124496e-02 -5.74669123e-01 -2.05362126e-01
-5.91751263e-02 -3.30780745e-01 8.43014240e-01 -1.62394404e-01
-6.65953934e-01 1.19515359e-01 -2.22252518e-01 -4.53424871e-01
6.97855577e-02 -4.25968505e-02 -6.95152059e-02 6.05664402e-03
-2.24713111e+00 7.70751119e-01 2.88634032e-01 3.25074792e-01
-9.16054070e-01 -7.64452398e-01 -1.06521714e+00 1.11134939e-01
7.13289857e-01 -6.90328598e-01 1.43300736e+00 -1.16816044e+00
-1.73528862e+00 6.02946877e-01 -3.58757854e-01 -2.42541432e-01
-8.83257017e-02 -2.03976035e-01 -5.22422194e-01 3.07370692e-01
-2.54622214e-02 4.35161561e-01 6.01363242e-01 -1.24535537e+00
-2.82565415e-01 -8.68595317e-02 4.20020729e-01 6.07606173e-01
-3.04763429e-02 1.68090671e-01 -5.46170592e-01 -3.24379027e-01
-2.74344683e-01 -8.43833804e-01 1.17179165e-02 -8.81439090e-01
-8.59534860e-01 -8.95250499e-01 6.60217285e-01 -5.88503003e-01
1.41679215e+00 -2.08586192e+00 4.59223062e-01 6.82107210e-02
3.36183667e-01 -3.70930701e-01 -1.46485403e-01 8.97487879e-01
-4.26176548e-01 2.17878535e-01 9.65587273e-02 -2.37110093e-01
1.19552098e-01 3.51177454e-01 -4.93895680e-01 4.98410985e-02
4.88795400e-01 1.20362723e+00 -1.26170373e+00 -4.84752029e-01
8.22588727e-02 6.05975747e-01 -5.62692165e-01 6.50441170e-01
-6.27054095e-01 4.42501277e-01 -7.23311186e-01 1.54159397e-01
1.32328197e-02 -3.15309703e-01 5.85614383e-01 -5.63267112e-01
3.33423495e-01 6.90729678e-01 -5.54359436e-01 1.52435994e+00
-4.37022954e-01 4.56404239e-01 1.98864982e-01 -7.46711075e-01
7.52282083e-01 6.26451790e-01 1.21589877e-01 -3.15769255e-01
3.53441089e-01 -4.19045389e-01 5.86444549e-02 -9.87599194e-01
1.12258293e-01 -3.20529133e-01 -5.52260697e-01 4.99469280e-01
6.28334805e-02 -1.43616140e-01 -2.26091728e-01 1.02012253e+00
1.30451846e+00 -1.32899582e-01 3.79581124e-01 3.62122245e-02
5.99369586e-01 -2.18156889e-01 7.41950512e-01 2.09084705e-01
-3.95455182e-01 -5.07075675e-02 1.39721406e+00 -9.21948478e-02
-2.73051113e-01 -4.79894340e-01 6.19813025e-01 1.25155663e+00
3.76429081e-01 -6.46798372e-01 -8.57539594e-01 -7.39360571e-01
-2.05060050e-01 1.29965842e+00 -1.04153633e+00 -4.17865694e-01
-5.26083231e-01 -2.75503904e-01 6.41967654e-01 3.66120666e-01
3.68742377e-01 -1.49195731e+00 -2.12363396e-02 6.25626370e-02
-1.03552461e+00 -1.18002653e+00 -6.01617336e-01 5.54963611e-02
-1.18944816e-01 -1.39992845e+00 1.99724674e-01 -4.25439507e-01
3.91332120e-01 -7.92007074e-02 1.21805978e+00 3.70024770e-01
3.05140410e-02 6.36231303e-01 -5.41455090e-01 -2.75187105e-01
-3.30143064e-01 -3.74010801e-01 -2.66176373e-01 3.38959754e-01
5.53177476e-01 -5.48429370e-01 -6.11751795e-01 2.91518327e-02
-4.19863701e-01 3.37447315e-01 2.33162325e-02 7.41414666e-01
4.00462262e-02 -3.68785225e-02 8.85743022e-01 -9.07862842e-01
1.13101017e+00 -1.02961111e+00 2.15652987e-01 1.47160575e-01
-3.27585697e-01 -1.55985981e-01 3.95696551e-01 -1.76744133e-01
-1.50925207e+00 -2.84581393e-01 -4.31032444e-04 -3.74392778e-01
-2.27235883e-01 8.38325799e-01 -4.16215479e-01 9.13454056e-01
2.78882176e-01 -7.54286498e-02 -2.39705458e-01 1.39139533e-01
7.46609032e-01 3.83634955e-01 4.90994424e-01 -7.31033623e-01
1.68066740e-01 8.24050233e-02 -3.99729311e-01 -4.79303092e-01
-1.37569702e+00 -4.32455987e-01 -2.45967641e-01 -7.85866082e-01
1.22355700e+00 -8.89613748e-01 -1.36972392e+00 1.49974465e-01
-1.56227040e+00 -5.22220671e-01 7.63313770e-02 3.22644144e-01
-3.91261905e-01 5.05081005e-02 -9.97940958e-01 -1.09966743e+00
-4.76114452e-01 -7.79842973e-01 9.43341851e-01 2.85087407e-01
-6.60995483e-01 -1.17649221e+00 -1.19297601e-01 6.87263072e-01
-4.87333760e-02 4.50959921e-01 1.04803753e+00 -7.99525738e-01
-1.25285879e-01 -1.39279306e-01 -2.37788990e-01 -2.46543121e-02
-9.56259086e-04 1.09649472e-01 -9.62013304e-01 3.91976446e-01
-1.81009620e-01 -6.89807713e-01 6.11529291e-01 2.88025945e-01
8.37433875e-01 -4.99234229e-01 -4.27738279e-01 -1.52438238e-01
7.17413425e-01 1.83573574e-01 5.41958213e-01 -4.09803480e-01
8.42599213e-01 9.89989400e-01 6.21889532e-01 3.76576096e-01
1.00343919e+00 4.59883124e-01 5.06394446e-01 -8.07412341e-02
-6.59369444e-03 -5.73115170e-01 5.11643171e-01 1.04370892e+00
1.12957181e-02 -4.26145881e-01 -8.83858442e-01 5.43366551e-01
-2.39385629e+00 -1.22084343e+00 -5.87923825e-01 1.22495139e+00
1.21357334e+00 3.43753062e-02 -1.70818716e-01 -2.37652853e-01
8.97210658e-01 2.53357232e-01 -4.98327583e-01 -5.45667946e-01
-4.57825512e-03 -3.37665647e-01 -5.10972261e-01 1.17400455e+00
-7.33504117e-01 1.18449903e+00 4.63435030e+00 6.46569014e-01
-5.91565788e-01 3.31005305e-01 7.07398891e-01 -1.54200848e-02
-6.02528036e-01 2.27880374e-01 -3.63260210e-01 1.92568585e-01
9.38959718e-01 -2.05685973e-01 4.33189154e-01 7.56428301e-01
5.28371871e-01 -3.09544262e-02 -1.19169772e+00 7.40841210e-01
6.10708073e-02 -1.36785555e+00 -1.87360361e-01 -4.00632977e-01
4.37972546e-01 -3.42173845e-01 -6.05056822e-01 6.52954280e-01
5.79099536e-01 -8.35771441e-01 4.17400271e-01 6.40298367e-01
4.27510321e-01 -9.12183702e-01 7.78977334e-01 1.31884262e-01
-1.33914411e+00 1.12831116e-01 1.80429772e-01 -3.32600027e-01
3.45735580e-01 6.48164988e-01 -9.51606512e-01 5.37275553e-01
4.80751067e-01 8.25526059e-01 1.67206734e-01 -1.58238530e-01
-1.04767227e+00 1.11612475e+00 1.17024243e-01 -4.30771470e-01
2.76937991e-01 -7.35383257e-02 6.13406837e-01 1.49042499e+00
-4.10269201e-01 9.53564942e-01 1.62550434e-01 8.19423139e-01
-4.00189042e-01 4.07939367e-02 -4.24493313e-01 -2.00998649e-01
6.86853707e-01 1.41656005e+00 -4.71949607e-01 -4.72820878e-01
-2.36203730e-01 1.16985142e+00 6.36426866e-01 4.27482933e-01
-1.13379514e+00 -2.64955461e-01 9.47839558e-01 -4.54410046e-01
-1.51715010e-01 3.27371597e-01 6.71899915e-02 -9.95763421e-01
-3.90454859e-01 -6.93140328e-01 7.50295341e-01 -1.17304623e+00
-1.64666545e+00 4.72579479e-01 -7.63874948e-02 -3.57219666e-01
-2.06239164e-01 5.38424850e-02 -1.18849218e+00 7.30872869e-01
-1.20344520e+00 -1.33875501e+00 -4.18572158e-01 8.50295484e-01
6.25101924e-01 3.96464735e-01 9.20248985e-01 -1.33824304e-01
-8.84961665e-01 3.28639746e-01 -9.67684984e-01 3.24609905e-01
6.79060459e-01 -1.17470956e+00 -5.46509661e-02 5.78123271e-01
-3.43397319e-01 7.68823683e-01 8.00453842e-01 -9.89254475e-01
-1.45727682e+00 -1.09827924e+00 1.29792607e+00 -5.88000000e-01
1.05051661e+00 -4.50826228e-01 -9.91768301e-01 8.58610928e-01
7.71241426e-01 -4.35229123e-01 1.03226459e+00 6.44699275e-01
-5.72437644e-01 2.70528167e-01 -6.83354318e-01 7.59137154e-01
1.16048288e+00 -7.35659957e-01 -9.49773252e-01 3.81801456e-01
1.55138016e+00 -2.16420785e-01 -8.03105831e-01 1.33348361e-01
1.62390456e-01 -7.58143604e-01 5.82647264e-01 -8.90331388e-01
1.00560594e+00 -7.61536583e-02 -4.19529947e-03 -1.34405029e+00
-1.77690104e-01 -9.71135616e-01 -2.43308276e-01 1.58745360e+00
5.69321692e-01 -2.51032531e-01 2.06414551e-01 1.00133669e+00
-4.58996654e-01 -8.11877072e-01 -6.50779605e-01 -3.29872631e-02
-4.72597480e-01 -8.70635509e-01 6.00074112e-01 1.58812952e+00
1.01903367e+00 1.19997430e+00 -5.93887925e-01 3.13901007e-01
3.34126413e-01 3.32784027e-01 5.58380723e-01 -9.08864975e-01
-2.21006572e-01 -4.36132193e-01 2.80261099e-01 -6.51414037e-01
9.04480338e-01 -9.57888782e-01 2.40560994e-01 -1.67845726e+00
4.41968173e-01 -4.90631275e-02 -1.80966869e-01 7.80358374e-01
-7.11788237e-01 -2.51625478e-01 -1.33162336e-02 -1.07634321e-01
-8.47586095e-01 8.29442859e-01 1.29170310e+00 6.92670494e-02
-3.44608963e-01 -5.32002449e-01 -9.43239272e-01 8.02684724e-01
8.13120544e-01 -3.64128321e-01 -7.13253379e-01 6.42509460e-02
5.08026838e-01 5.36615133e-01 4.61482048e-01 1.14232965e-01
2.73136050e-01 -5.68892539e-01 -2.03104869e-01 -2.44078130e-01
4.41755593e-01 -4.01478618e-01 1.04972459e-01 2.46377617e-01
-1.00018394e+00 -4.18441951e-01 -5.89232929e-02 8.10365140e-01
-3.33528340e-01 2.10647583e-01 3.47934753e-01 8.55662227e-02
-7.43807495e-01 6.78537264e-02 -3.33424568e-01 2.32848018e-01
9.71147180e-01 3.87511939e-01 -4.93960112e-01 -1.00724995e+00
-8.26072931e-01 6.63019955e-01 -2.77440280e-01 6.89562798e-01
7.38826275e-01 -1.34826100e+00 -7.66185284e-01 -4.22795683e-01
2.92609721e-01 -1.36171296e-01 5.42573631e-01 1.05148256e+00
2.71493763e-01 2.48163447e-01 4.53270197e-01 -1.42618278e-02
-1.30798328e+00 5.16403079e-01 3.01091343e-01 -1.96101308e-01
-5.00250638e-01 1.43914342e+00 2.97491193e-01 -4.16705161e-01
2.05096677e-01 -2.70371556e-01 -3.28749061e-01 1.67773843e-01
4.44633663e-01 -3.09796422e-03 -5.40834069e-01 -5.29428720e-01
-7.28037655e-01 -4.76970933e-02 2.56786197e-01 -4.48215194e-02
1.14506578e+00 -3.32824558e-01 -7.27576733e-01 6.68223202e-01
1.09498835e+00 -1.07420407e-01 -9.85911250e-01 -1.56923264e-01
-1.07888319e-02 1.28731862e-01 9.49730456e-04 -1.06326866e+00
-8.77104998e-01 5.79829931e-01 -4.03428137e-01 3.44898045e-01
1.03105021e+00 5.77132583e-01 7.26304114e-01 1.17545441e-01
-8.22427869e-02 -8.67333829e-01 4.84393269e-01 7.61883736e-01
1.57758164e+00 -9.97632384e-01 -4.75525856e-01 -9.30029511e-01
-1.25513721e+00 9.63446975e-01 9.69164133e-01 1.42835692e-01
6.97927892e-01 2.68029273e-01 1.63889647e-01 -8.03343892e-01
-1.54874742e+00 -2.61725366e-01 2.24450260e-01 1.48709610e-01
7.86723137e-01 4.57996398e-01 -3.17245811e-01 1.25274110e+00
-1.12630986e-01 -3.25916052e-01 4.02939051e-01 3.94123673e-01
-6.44393265e-02 -6.11271501e-01 1.82043780e-02 8.34772810e-02
-7.73264021e-02 -4.50731128e-01 -1.23280358e+00 5.28083622e-01
-2.92265303e-02 1.62989950e+00 -1.23142965e-01 -7.52102911e-01
3.23968619e-01 5.34233809e-01 5.91339320e-02 -5.10168970e-01
-9.05916750e-01 1.67782888e-01 8.96121919e-01 -9.11443233e-01
-6.17093503e-01 -5.21443486e-01 -2.06186700e+00 -3.88336092e-01
-3.56319577e-01 4.65588033e-01 2.23571852e-01 1.08447587e+00
4.88500923e-01 7.68566191e-01 6.32671356e-01 -3.65236849e-01
2.27804467e-01 -1.08074665e+00 -1.26741186e-01 7.81933665e-01
1.07682236e-01 -5.01265526e-01 -4.86404479e-01 1.14099979e-01] | [12.838027000427246, 6.392460823059082] |
8b85c77a-add3-4089-9341-7119af555b18 | graph-masked-autoencoder-for-sequential | 2305.04619 | null | https://arxiv.org/abs/2305.04619v3 | https://arxiv.org/pdf/2305.04619v3.pdf | Graph Masked Autoencoder for Sequential Recommendation | While some powerful neural network architectures (e.g., Transformer, Graph Neural Networks) have achieved improved performance in sequential recommendation with high-order item dependency modeling, they may suffer from poor representation capability in label scarcity scenarios. To address the issue of insufficient labels, Contrastive Learning (CL) has attracted much attention in recent methods to perform data augmentation through embedding contrasting for self-supervision. However, due to the hand-crafted property of their contrastive view generation strategies, existing CL-enhanced models i) can hardly yield consistent performance on diverse sequential recommendation tasks; ii) may not be immune to user behavior data noise. In light of this, we propose a simple yet effective Graph Masked AutoEncoder-enhanced sequential Recommender system (MAERec) that adaptively and dynamically distills global item transitional information for self-supervised augmentation. It naturally avoids the above issue of heavy reliance on constructing high-quality embedding contrastive views. Instead, an adaptive data reconstruction paradigm is designed to be integrated with the long-range item dependency modeling, for informative augmentation in sequential recommendation. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art baseline models and can learn more accurate representations against data noise and sparsity. Our implemented model code is available at https://github.com/HKUDS/MAERec. | ['Chao Huang', 'Lianghao Xia', 'Yaowen Ye'] | 2023-05-08 | null | null | null | null | ['sequential-recommendation'] | ['miscellaneous'] | [ 2.43623003e-01 -1.02384597e-01 -5.65470397e-01 -3.60330343e-01
-3.11594278e-01 -4.85722214e-01 5.25371313e-01 -4.54522111e-02
-7.54439011e-02 4.25222337e-01 5.74321210e-01 -3.86423290e-01
-1.52982593e-01 -7.11159885e-01 -6.29506528e-01 -5.78737974e-01
2.13828668e-01 4.01431233e-01 -3.79700035e-01 -4.94098336e-01
-8.91528726e-02 1.66668802e-01 -1.38325787e+00 2.15726838e-01
9.33655798e-01 9.08683777e-01 2.04889894e-01 2.23395213e-01
-1.97808042e-01 7.86176860e-01 -1.23884887e-01 -4.35688972e-01
3.71442735e-01 -4.95828956e-01 -5.12228847e-01 2.45113462e-01
4.74547714e-01 -4.78910059e-01 -6.44318640e-01 9.41029429e-01
4.33545113e-01 4.96144235e-01 5.83733261e-01 -1.20630586e+00
-1.44283164e+00 9.21383917e-01 -5.48584700e-01 1.72972739e-01
1.26724258e-01 9.11180954e-03 1.33683431e+00 -1.24377310e+00
3.76330435e-01 8.64601374e-01 6.76020682e-01 7.14507878e-01
-1.36043048e+00 -6.91141784e-01 7.25338519e-01 -4.16128524e-03
-1.05685532e+00 -3.43077421e-01 1.16239345e+00 -1.33839622e-01
7.08044291e-01 2.22852737e-01 6.15385592e-01 1.59283328e+00
-4.20315892e-01 8.61669838e-01 9.10336018e-01 -2.49174088e-01
-1.63333900e-02 1.23778664e-01 4.06360060e-01 7.81635821e-01
2.35959619e-01 1.26942694e-01 -5.28697550e-01 -1.48908362e-01
9.76103961e-01 5.64941525e-01 -3.57123941e-01 -3.11421007e-01
-1.11307418e+00 1.02435374e+00 6.55513585e-01 3.37743551e-01
-4.82569188e-01 4.36955281e-02 4.48029786e-01 5.73654413e-01
7.06004262e-01 5.61996937e-01 -5.12430072e-01 3.74917150e-01
-4.79605854e-01 -1.12116861e-03 5.26025474e-01 1.02599347e+00
5.24293900e-01 6.82130158e-01 -1.49182200e-01 1.08361161e+00
3.31459612e-01 3.13330114e-01 6.60648763e-01 -6.69477344e-01
4.75032181e-01 7.03970909e-01 -1.37604982e-01 -1.28069174e+00
-3.75979245e-01 -1.15141654e+00 -1.23735964e+00 -3.21685046e-01
9.97725353e-02 3.20780985e-02 -7.87541449e-01 1.75772250e+00
2.73890316e-01 3.09155166e-01 -1.43930182e-01 9.82314050e-01
1.11392903e+00 6.24591589e-01 1.05252802e-01 -1.87854871e-01
1.26133943e+00 -1.25162637e+00 -8.26167285e-01 -2.29420736e-01
5.20786643e-01 -4.30860847e-01 1.41328180e+00 2.94915289e-01
-8.41714203e-01 -8.19676578e-01 -1.12384975e+00 -6.17334135e-02
-2.38283381e-01 2.27754980e-01 9.45553958e-01 4.28106725e-01
-1.02078974e+00 4.90156293e-01 -6.27331138e-01 -1.48437142e-01
2.93189138e-01 3.25948954e-01 -3.47555429e-01 -2.27753162e-01
-1.27212477e+00 4.18439209e-01 2.49148622e-01 1.53956771e-01
-6.25400662e-01 -6.03062153e-01 -8.76343966e-01 1.30425766e-01
7.87468791e-01 -9.03426588e-01 9.85901892e-01 -1.17735589e+00
-1.56648278e+00 3.00383568e-01 2.19727814e-01 -3.15663218e-01
-8.82584676e-02 -3.36149871e-01 -6.74753726e-01 -1.50273219e-01
-1.33652747e-01 5.08859515e-01 1.02034676e+00 -1.45956635e+00
-2.62912691e-01 -4.41135168e-01 2.85026103e-01 4.85652357e-01
-1.02014482e+00 -3.34364384e-01 -5.02416372e-01 -1.32976413e+00
1.72330141e-01 -1.02014816e+00 -4.46383417e-01 -2.62891710e-01
-2.46959180e-01 -2.21921816e-01 6.31632447e-01 -5.42659581e-01
1.43536663e+00 -1.99408448e+00 2.23751992e-01 6.66659698e-02
3.65726113e-01 3.33223164e-01 -5.79306304e-01 4.81433302e-01
-1.08724184e-01 1.03668105e-02 2.10946500e-02 -5.96332550e-01
-1.98954225e-01 1.82173327e-01 -2.47353211e-01 3.15161437e-01
1.15959123e-02 9.85598087e-01 -9.62106943e-01 -3.46318670e-02
1.32148832e-01 5.97333133e-01 -7.99973190e-01 4.03535277e-01
-3.24693829e-01 4.80019182e-01 -4.14522886e-01 7.00952411e-01
3.90472740e-01 -8.81042838e-01 4.49143678e-01 -4.72510964e-01
3.30064267e-01 3.95292044e-01 -1.14873135e+00 1.73429298e+00
-5.55381358e-01 1.51871964e-02 -1.23667337e-01 -1.06065679e+00
8.86698961e-01 2.83542812e-01 4.67057049e-01 -7.04458833e-01
4.85481508e-02 6.28701877e-04 -1.40595526e-01 -1.42473474e-01
8.31722021e-01 1.10811377e-02 1.50361016e-01 6.11215830e-01
2.24875897e-01 5.85603714e-01 -1.56985343e-01 4.40797091e-01
9.21682835e-01 1.65497288e-01 3.19517523e-01 -4.04049493e-02
4.30800676e-01 -3.09258133e-01 5.92308283e-01 5.80264390e-01
1.44855946e-01 6.25862598e-01 -8.02979469e-02 -3.33121628e-01
-8.31720233e-01 -6.60648584e-01 2.69006729e-01 1.60580862e+00
-4.28899676e-02 -5.90134561e-01 -2.61304915e-01 -1.03609097e+00
-2.94044092e-02 6.09918058e-01 -6.01210177e-01 -3.70017231e-01
-6.44223809e-01 -7.50187755e-01 1.38709933e-01 7.90615797e-01
2.16880977e-01 -1.07259095e+00 4.29669648e-01 4.36554223e-01
-2.42544338e-01 -9.54639792e-01 -8.45942557e-01 7.92336762e-02
-1.03507030e+00 -8.13688397e-01 -6.18104517e-01 -8.57865274e-01
8.50041032e-01 1.04194129e+00 1.28254986e+00 2.84371287e-01
4.02095884e-01 5.61896980e-01 -8.28324199e-01 4.07186672e-02
-3.04857373e-01 1.48355171e-01 3.81279081e-01 1.60818934e-01
2.34344453e-01 -9.47300315e-01 -8.70247602e-01 3.99046659e-01
-9.62552011e-01 1.45128816e-01 5.79626441e-01 1.17309499e+00
8.36099267e-01 -2.31865987e-01 1.01971543e+00 -1.32809603e+00
7.05368817e-01 -8.07142913e-01 -3.96365881e-01 2.20456123e-01
-1.13795936e+00 -3.68007347e-02 1.11855364e+00 -7.18142748e-01
-9.47630703e-01 -1.56122133e-01 -1.56798884e-01 -7.35495210e-01
1.42685771e-02 8.39712679e-01 -6.22150004e-02 1.11858733e-01
6.91715240e-01 3.47660929e-01 1.19473442e-01 -7.67174721e-01
6.65732384e-01 5.25559127e-01 3.55116397e-01 -4.14095610e-01
8.13417196e-01 3.04304957e-01 -3.00676018e-01 -4.23774481e-01
-1.20551443e+00 -4.72516328e-01 -3.12550098e-01 -1.32461295e-01
3.61882806e-01 -1.19868541e+00 -2.67976493e-01 2.35937238e-01
-5.70704162e-01 -2.35443562e-01 -3.57803285e-01 4.81326669e-01
-2.41423696e-01 5.50616145e-01 -7.79906213e-01 -5.86985171e-01
-7.01578677e-01 -8.17760229e-01 5.91846168e-01 3.96032706e-02
8.81028622e-02 -1.06060076e+00 -4.74859364e-02 6.47128165e-01
6.65742338e-01 -8.25126842e-02 8.89827847e-01 -9.70881522e-01
-3.97977769e-01 -1.21711448e-01 -1.26082838e-01 5.55087030e-01
3.55948865e-01 -3.75070632e-01 -7.65629292e-01 -5.87652326e-01
-1.00874230e-01 -4.50823396e-01 8.52607071e-01 2.73310184e-01
1.45168507e+00 -6.05506539e-01 -4.31448109e-02 5.55303514e-01
1.32934451e+00 -5.46973608e-02 2.73111969e-01 6.96621537e-02
1.20727062e+00 3.68149191e-01 4.91132796e-01 4.53235775e-01
6.34644270e-01 7.06041694e-01 4.99509573e-01 -1.14190243e-01
-4.01418746e-01 -6.23754025e-01 3.79141361e-01 1.49116421e+00
-2.03999683e-01 -6.31725430e-01 -4.09824818e-01 3.04552346e-01
-1.95518363e+00 -7.79966712e-01 -9.26571488e-02 1.97683227e+00
6.46118045e-01 5.45078032e-02 2.99746782e-01 -3.43925017e-03
5.17284930e-01 3.20960641e-01 -6.37617469e-01 7.55093321e-02
-1.06893949e-01 -1.22615293e-01 2.64867693e-01 1.44979417e-01
-1.01042068e+00 8.73845279e-01 5.11408901e+00 7.58785903e-01
-8.20349216e-01 3.06252241e-01 6.26447201e-01 3.74628082e-02
-7.49908388e-01 -2.19070822e-01 -7.23578274e-01 4.49592799e-01
8.34390283e-01 3.95889848e-01 6.26370907e-01 8.27415407e-01
-6.84796274e-02 6.59872591e-01 -9.19695020e-01 9.73859012e-01
2.82267034e-01 -1.40552711e+00 2.92713135e-01 1.77080363e-01
8.96296322e-01 3.25770341e-02 4.15348172e-01 7.25032568e-01
4.64621156e-01 -7.71629333e-01 3.64405751e-01 3.32066149e-01
8.25213134e-01 -6.46192789e-01 6.93084359e-01 3.38959187e-01
-1.17512059e+00 -2.56141514e-01 -3.81823093e-01 -6.49743974e-02
7.55990520e-02 5.52238703e-01 -4.34015691e-01 6.94328666e-01
4.18475062e-01 1.06468630e+00 -5.69939554e-01 4.86190021e-01
-1.57472521e-01 9.62661386e-01 7.49445986e-04 3.17785852e-02
2.09085733e-01 -4.65229124e-01 4.17975038e-01 8.39595795e-01
2.23500863e-01 2.37319306e-01 4.38396573e-01 5.24846315e-01
-4.42854762e-01 3.01670343e-01 -7.15857565e-01 -1.70654386e-01
6.17048502e-01 1.36812115e+00 -5.13223171e-01 -2.28400767e-01
-8.66620660e-01 9.33090389e-01 7.15916216e-01 5.02374649e-01
-7.39914060e-01 2.69443959e-01 3.74231577e-01 1.83279678e-01
2.97312349e-01 -1.90529957e-01 -1.60950497e-01 -1.46552265e+00
-1.27532035e-01 -1.27060187e+00 5.61723709e-01 -4.52088714e-01
-1.86685205e+00 6.94702625e-01 -3.06942940e-01 -1.53278506e+00
-8.11035186e-02 -4.57709879e-01 -3.28833342e-01 4.31695879e-01
-1.44567204e+00 -1.46481025e+00 -1.63792267e-01 7.69250393e-01
7.27651417e-01 -5.57806611e-01 1.01619828e+00 5.85706532e-01
-7.32589900e-01 1.01011789e+00 1.53243482e-01 -5.27784266e-02
6.49026632e-01 -1.19265902e+00 4.05994594e-01 6.90872550e-01
6.06532574e-01 8.08481574e-01 3.65534961e-01 -6.33599699e-01
-1.65008879e+00 -1.36536980e+00 3.19579244e-01 -3.33638519e-01
7.07124054e-01 -3.55197906e-01 -1.11038184e+00 8.50102067e-01
7.73640797e-02 2.78113127e-01 1.02172899e+00 6.22124851e-01
-6.39757335e-01 -1.59151420e-01 -6.56580925e-01 6.49210513e-01
1.36797738e+00 -5.22809505e-01 -2.67913908e-01 3.86864603e-01
8.94914865e-01 -8.34647566e-02 -9.81014490e-01 4.78100657e-01
3.62709910e-01 -6.59258544e-01 1.11181784e+00 -7.98535645e-01
4.00571644e-01 -1.25874549e-01 -2.66374201e-01 -1.37412453e+00
-9.18097496e-01 -4.31035757e-01 -8.03935528e-01 1.37583852e+00
4.89018381e-01 -5.57685614e-01 7.58125603e-01 2.79855818e-01
-3.14528048e-01 -1.04371035e+00 -4.31632817e-01 -6.70745313e-01
-3.24417233e-01 -1.96328387e-01 5.40820122e-01 1.37021196e+00
-2.90586025e-01 8.85694981e-01 -9.49820578e-01 9.70461890e-02
4.26200688e-01 2.62330830e-01 6.66758120e-01 -1.23040259e+00
-7.26026773e-01 -2.03276217e-01 1.78853050e-01 -1.31280398e+00
1.86377168e-01 -1.21401882e+00 -3.20743293e-01 -1.53611684e+00
2.61130780e-01 -7.11949110e-01 -8.76362503e-01 5.80187201e-01
-4.51769501e-01 3.78812462e-01 5.18218465e-02 4.23532933e-01
-7.57458866e-01 8.64602387e-01 1.30690062e+00 1.11279730e-02
-2.51349688e-01 1.82359725e-01 -1.25193393e+00 5.27120352e-01
9.93238807e-01 -3.81225377e-01 -7.86390841e-01 -4.68669385e-01
5.67534804e-01 8.53622034e-02 1.13941737e-01 -5.72699070e-01
3.35172378e-02 -6.13345131e-02 2.57729292e-01 -4.71782416e-01
2.03745797e-01 -9.09783781e-01 1.49267256e-01 1.57445371e-01
-5.17457128e-01 2.62637824e-01 -2.19119310e-01 1.11190307e+00
-6.65089488e-02 -6.86613843e-02 4.17322069e-01 -2.11207822e-01
-5.80774128e-01 7.53869534e-01 -1.64028689e-01 -1.26144275e-01
3.62264395e-01 3.40548232e-02 -3.35179001e-01 -5.53327799e-01
-6.08799994e-01 2.07360953e-01 2.98194766e-01 7.06444681e-01
7.19748557e-01 -1.71619165e+00 -6.37725592e-01 1.62383527e-01
2.48703286e-01 -2.17443824e-01 5.01064956e-01 6.83048010e-01
1.47645622e-01 7.56827220e-02 1.13729164e-02 -8.25370923e-02
-9.52239990e-01 9.21228826e-01 -2.96708866e-04 -5.96638978e-01
-8.38317394e-01 9.04052377e-01 1.27286524e-01 -6.68843865e-01
2.41899967e-01 6.59040501e-03 -6.05058670e-01 1.47838116e-01
3.85972291e-01 2.21281230e-01 6.55252114e-02 -4.99261528e-01
4.59726825e-02 8.16954598e-02 -4.49002326e-01 3.95609468e-01
1.56474447e+00 -3.08501005e-01 1.77339673e-01 3.60028595e-01
1.03673828e+00 6.51452020e-02 -1.14364660e+00 -7.58275628e-01
-3.04804087e-01 -2.78543353e-01 3.07682604e-01 -7.75419235e-01
-1.40139544e+00 5.19319057e-01 4.27309245e-01 3.77500981e-01
1.14852798e+00 -2.45581016e-01 9.66221511e-01 4.67223734e-01
2.46979937e-01 -9.52849686e-01 5.08228242e-01 4.00680780e-01
8.60676169e-01 -1.52969265e+00 1.99969232e-01 -3.58054191e-01
-8.47565591e-01 7.99286306e-01 8.62437785e-01 -2.26895452e-01
7.61878729e-01 -7.24353790e-02 -8.93705040e-02 -2.76191950e-01
-8.16725910e-01 -1.65304065e-01 6.71942711e-01 4.73816603e-01
5.99669278e-01 1.03011064e-01 -1.31216854e-01 8.88686240e-01
9.35426727e-02 -3.09832573e-01 3.42652112e-01 5.41359901e-01
-1.44702777e-01 -1.10638654e+00 1.09592721e-01 8.05602133e-01
-4.18619722e-01 -3.27273130e-01 -1.36409551e-01 7.14199722e-01
-1.58768430e-01 9.50213253e-01 -1.54821336e-01 -6.61962152e-01
2.14793965e-01 -5.74428365e-02 2.08935052e-01 -7.14081705e-01
-7.27445662e-01 2.31987521e-01 1.68792501e-01 -3.58011484e-01
-5.76806605e-01 -3.67936701e-01 -8.16345215e-01 -1.93769440e-01
-6.86321616e-01 6.82769120e-02 2.99184293e-01 6.61424518e-01
6.01629019e-01 6.76837265e-01 8.67761970e-01 -7.34944046e-01
-8.13655674e-01 -1.09128594e+00 -5.88976681e-01 5.90099037e-01
1.79701418e-01 -7.26036131e-01 -2.36780956e-01 -2.17660680e-01] | [10.195484161376953, 5.585234642028809] |
98a7e15d-293e-4ae0-990c-57daec0150e8 | domain-adaptation-in-reinforcement-learning | 2102.05714 | null | https://arxiv.org/abs/2102.05714v2 | https://arxiv.org/pdf/2102.05714v2.pdf | Domain Adaptation In Reinforcement Learning Via Latent Unified State Representation | Despite the recent success of deep reinforcement learning (RL), domain adaptation remains an open problem. Although the generalization ability of RL agents is critical for the real-world applicability of Deep RL, zero-shot policy transfer is still a challenging problem since even minor visual changes could make the trained agent completely fail in the new task. To address this issue, we propose a two-stage RL agent that first learns a latent unified state representation (LUSR) which is consistent across multiple domains in the first stage, and then do RL training in one source domain based on LUSR in the second stage. The cross-domain consistency of LUSR allows the policy acquired from the source domain to generalize to other target domains without extra training. We first demonstrate our approach in variants of CarRacing games with customized manipulations, and then verify it in CARLA, an autonomous driving simulator with more complex and realistic visual observations. Our results show that this approach can achieve state-of-the-art domain adaptation performance in related RL tasks and outperforms prior approaches based on latent-representation based RL and image-to-image translation. | ['Jeffrey L. Krichmar', 'Emre Neftci', 'Xinyun Zou', 'Kexin Chen', 'Takashi Nagata', 'Jinwei Xing'] | 2021-02-10 | null | null | null | null | ['transfer-reinforcement-learning'] | ['methodology'] | [ 1.66842192e-01 1.05687365e-01 -2.82397598e-01 -1.76440358e-01
-7.62053967e-01 -7.35376954e-01 8.43468368e-01 -3.24285865e-01
-6.81603253e-01 1.08910263e+00 -4.67687882e-02 -6.63247555e-02
2.91211516e-01 -4.69188958e-01 -9.81915116e-01 -7.09029317e-01
9.32188481e-02 6.87822402e-01 5.47302127e-01 -5.03292322e-01
-3.44250235e-04 5.64384699e-01 -1.60070837e+00 3.93843539e-02
5.93483508e-01 5.43829143e-01 5.76797545e-01 6.53706014e-01
1.82454228e-01 1.11663413e+00 -5.54529071e-01 3.27250689e-01
5.47354043e-01 -6.27008259e-01 -7.18297720e-01 6.04560971e-02
2.70061195e-01 -7.58800030e-01 -3.53069484e-01 8.43142688e-01
5.35576761e-01 5.26927769e-01 5.61382771e-01 -1.43287623e+00
-5.87963402e-01 5.85840605e-02 -3.74930084e-01 -3.39594148e-02
2.53836572e-01 6.60662472e-01 5.68190932e-01 -3.89420986e-01
1.10718179e+00 1.26908374e+00 1.55384943e-01 1.04854178e+00
-1.30368578e+00 -8.75115156e-01 3.30469728e-01 -1.27728637e-02
-9.08775210e-01 -4.32989776e-01 5.96959114e-01 -3.58322978e-01
9.36537504e-01 -5.87142766e-01 5.38785934e-01 1.54028726e+00
2.53473490e-01 8.27967644e-01 1.52216053e+00 -2.09713548e-01
6.17218435e-01 1.43592700e-01 -5.96225619e-01 6.00309014e-01
4.99793440e-02 5.47290802e-01 -5.03435016e-01 1.56907722e-01
1.20588398e+00 -3.10551465e-01 1.25093469e-02 -9.39413965e-01
-1.35019267e+00 9.05669630e-01 4.17760700e-01 1.28958151e-01
-5.00294507e-01 4.54246134e-01 4.57431912e-01 5.98396957e-01
1.97427213e-01 6.61249816e-01 -4.79979306e-01 -3.22789490e-01
-7.19406009e-01 5.69747627e-01 6.24790907e-01 9.31356549e-01
8.23494077e-01 4.26482409e-01 -1.32907569e-01 6.65911317e-01
-7.45800361e-02 5.65935135e-01 6.46018147e-01 -1.41342628e+00
1.99016094e-01 2.16716990e-01 3.73699367e-01 -3.17935199e-01
-1.79090247e-01 -1.97585925e-01 -3.26087773e-01 9.75196958e-01
4.30687726e-01 -3.85619104e-01 -9.78576660e-01 2.14024138e+00
3.54159981e-01 3.11734378e-01 5.44471443e-01 1.03918970e+00
3.04735214e-01 6.51836574e-01 2.58664280e-01 -6.72339424e-02
8.41982305e-01 -1.00857782e+00 -4.76258725e-01 -7.09838808e-01
4.19083983e-01 -4.12097752e-01 1.22388363e+00 1.37816504e-01
-1.04767370e+00 -7.52927244e-01 -1.16627026e+00 -4.90766317e-02
-2.34301373e-01 -5.54678962e-02 4.47483331e-01 1.46018360e-02
-1.06265640e+00 4.51551288e-01 -9.13949728e-01 -8.60670507e-01
2.62924612e-01 4.29099590e-01 -5.75376511e-01 -1.15558416e-01
-1.27971947e+00 1.30823386e+00 4.73045051e-01 -5.21804035e-01
-1.56128287e+00 -5.09422839e-01 -8.80047321e-01 -2.06477642e-01
5.36725283e-01 -7.02531219e-01 1.66946948e+00 -1.21057618e+00
-2.30104065e+00 8.05366099e-01 5.32994643e-02 -5.88211000e-01
5.66771686e-01 -2.17027143e-01 -2.17610672e-02 2.29976699e-01
2.79238999e-01 1.17357087e+00 1.05080366e+00 -1.43230402e+00
-6.36582077e-01 -9.41916928e-02 4.26857263e-01 6.27634108e-01
2.16125488e-01 -1.93125024e-01 -1.75465927e-01 -3.62492710e-01
-3.66215348e-01 -1.23439515e+00 -2.26866156e-01 1.42634213e-01
4.16791707e-01 -1.22676797e-01 8.90674293e-01 -4.66087431e-01
3.44733059e-01 -2.17918825e+00 5.01048326e-01 -3.08541030e-01
-8.31271857e-02 2.48150095e-01 -4.48250562e-01 3.48013431e-01
2.82545816e-02 -4.83972579e-01 -2.13107184e-01 -2.45843440e-01
1.03269070e-02 6.71143055e-01 -4.96310532e-01 4.93675530e-01
3.30106258e-01 1.01524734e+00 -1.19933891e+00 -2.60292470e-01
2.19660670e-01 1.12264827e-01 -5.34050643e-01 5.10043144e-01
-5.81220567e-01 1.10060370e+00 -4.16537225e-01 7.59484097e-02
2.34658301e-01 -1.39207780e-01 1.91230685e-01 2.32555851e-01
-1.08583488e-01 8.83121118e-02 -8.90447080e-01 2.16868496e+00
-5.99622130e-01 7.56794810e-01 8.76103565e-02 -1.02990222e+00
9.00376916e-01 2.08893880e-01 5.09150743e-01 -9.23863828e-01
4.46159542e-02 1.04737923e-01 2.01794971e-02 -4.47849333e-01
5.82072496e-01 -3.88561100e-01 -3.65710616e-01 5.49688458e-01
2.86691874e-01 -6.32499635e-01 5.43224104e-02 1.64929375e-01
1.02629721e+00 9.78672743e-01 4.53744441e-01 2.82974960e-03
1.60286307e-01 3.70768547e-01 5.41157901e-01 7.64108539e-01
-5.22863567e-01 3.57897341e-01 4.77097750e-01 -2.71319538e-01
-1.30953968e+00 -1.20015478e+00 3.96528959e-01 1.34664786e+00
2.43305102e-01 2.30543882e-01 -5.70982575e-01 -8.56074333e-01
7.65897240e-03 1.02329016e+00 -7.65427470e-01 -4.32462215e-01
-5.21891952e-01 7.03450814e-02 5.40168822e-01 5.55896223e-01
7.63232589e-01 -1.38135660e+00 -1.16247070e+00 3.23725134e-01
-9.81015489e-02 -1.36623836e+00 -1.43515453e-01 2.65741080e-01
-6.66173279e-01 -7.73393154e-01 -8.83451343e-01 -7.20447361e-01
4.93107826e-01 1.52495340e-01 1.05490363e+00 -3.74794722e-01
-4.23042402e-02 6.44865930e-01 -3.43071371e-01 -2.25066423e-01
-8.58871639e-01 -6.79318309e-02 1.93261892e-01 -3.41994166e-01
7.36098439e-02 -4.51343089e-01 -3.97858381e-01 2.83339083e-01
-8.89396489e-01 1.64354742e-01 7.63515711e-01 9.36658382e-01
4.55581635e-01 -3.67793143e-01 7.62683928e-01 -5.14865875e-01
7.63167560e-01 -3.85958552e-01 -8.69032919e-01 2.36168969e-02
-2.47324184e-01 4.46406692e-01 6.13135397e-01 -8.40057671e-01
-1.17546940e+00 3.44888657e-01 1.80803210e-01 -6.17781222e-01
-3.64201099e-01 1.38348073e-01 6.52854741e-02 2.53274525e-03
8.74578893e-01 3.81748408e-01 3.44356924e-01 2.53321137e-03
6.94049895e-01 3.25259358e-01 6.17018223e-01 -7.19411373e-01
8.65532577e-01 5.47401965e-01 -7.51073509e-02 -7.35952854e-01
-6.22873843e-01 -2.53449947e-01 -6.61186397e-01 -3.17849278e-01
1.08683157e+00 -1.11279202e+00 -6.68311834e-01 2.74805188e-01
-1.02629948e+00 -1.14456856e+00 -6.78259313e-01 4.83603239e-01
-1.18614674e+00 2.15272699e-02 -3.01797032e-01 -4.22683060e-01
1.86485976e-01 -1.42791355e+00 1.08686757e+00 1.79813549e-01
-1.88016683e-01 -9.31190133e-01 4.62606251e-01 4.74334434e-02
4.11811322e-01 2.10372582e-01 7.17994153e-01 -4.80272502e-01
-4.45849568e-01 1.88095763e-01 -1.15972355e-01 4.36312556e-01
8.78628641e-02 -4.96775478e-01 -7.91083574e-01 -6.74992263e-01
-1.35194466e-01 -1.02205551e+00 5.76810420e-01 1.51422858e-01
6.60352707e-01 1.13768205e-01 -1.72337845e-01 4.12463963e-01
1.35012364e+00 2.07103342e-01 6.69421673e-01 5.59998155e-01
3.19858104e-01 4.28191036e-01 8.67665231e-01 2.82835662e-01
3.99657607e-01 6.85578227e-01 4.84579802e-01 -1.77047178e-01
-3.26462865e-01 -4.45755064e-01 9.00268078e-01 3.08577180e-01
-3.11905649e-02 1.12623945e-01 -6.57041967e-01 5.81939280e-01
-1.93087494e+00 -9.96909797e-01 6.91622913e-01 2.01450658e+00
7.64949262e-01 8.60367492e-02 1.99934542e-01 -5.62937021e-01
3.81755322e-01 5.27029634e-02 -1.09744895e+00 -5.60128987e-01
7.64832199e-02 2.74035513e-01 5.78260839e-01 3.67223501e-01
-8.37152541e-01 1.62047815e+00 6.57111406e+00 5.49308062e-01
-1.36166716e+00 2.58623600e-01 1.19354725e-01 -2.89774220e-02
1.18306480e-01 1.38210401e-01 -5.23670495e-01 1.13001660e-01
8.18840444e-01 -1.54044420e-01 7.95256495e-01 1.02580309e+00
2.23050460e-01 -3.68465394e-01 -1.15122235e+00 6.66283190e-01
1.15559753e-02 -1.11316502e+00 -9.98475775e-02 7.86497295e-02
9.27401304e-01 2.75411010e-01 2.26898327e-01 9.17728603e-01
9.99363184e-01 -8.19745302e-01 6.34875596e-01 2.75551289e-01
9.76944447e-01 -6.35957301e-01 3.40782583e-01 5.88413715e-01
-7.98302352e-01 -8.75614956e-02 -5.21076739e-01 -1.59338668e-01
-2.65687276e-02 -4.25874978e-01 -1.10421479e+00 3.09682578e-01
4.26402926e-01 7.70025015e-01 -2.79076695e-01 4.97369885e-01
-5.00832021e-01 3.36347103e-01 -1.42790720e-01 9.76871103e-02
4.46414769e-01 -7.30931833e-02 5.80578029e-01 7.38191068e-01
1.94792017e-01 -2.13713497e-02 3.81434381e-01 8.73702347e-01
1.10030705e-02 -1.89216822e-01 -1.06021178e+00 -2.55825277e-03
1.21093392e-01 9.64832246e-01 -5.02660990e-01 -4.11584139e-01
-3.95111233e-01 1.34828424e+00 6.90800428e-01 7.41347730e-01
-1.04736626e+00 9.80060082e-03 9.07612681e-01 -7.94370100e-02
3.93243998e-01 -5.84711432e-01 2.51187354e-01 -1.22602904e+00
-3.95780534e-01 -1.04063845e+00 3.97165157e-02 -1.00361609e+00
-1.01497388e+00 4.99861747e-01 2.21847638e-01 -1.52985859e+00
-7.96871841e-01 -4.50077325e-01 -1.40976280e-01 7.50923336e-01
-1.76990032e+00 -1.23084223e+00 -2.99252987e-01 8.29015851e-01
8.62754524e-01 -4.77477998e-01 9.21627045e-01 -1.81297854e-01
-1.96000978e-01 3.71007353e-01 2.09720463e-01 -8.87073800e-02
1.14324558e+00 -1.10538399e+00 3.79473776e-01 6.82727158e-01
-1.84546426e-01 4.01795775e-01 8.67117405e-01 -6.19271457e-01
-1.28967404e+00 -1.07436895e+00 -1.88505426e-02 -3.84647101e-01
6.05500340e-01 -3.30160618e-01 -8.35874259e-01 9.29763198e-01
4.71035451e-01 6.03273511e-02 1.27736494e-01 -3.35944891e-01
-3.49655151e-01 -1.63041111e-02 -1.07199407e+00 8.98930430e-01
8.21421802e-01 -4.81673330e-01 -7.23864555e-01 1.93450511e-01
7.28643537e-01 -5.68735957e-01 -5.72381079e-01 2.10499242e-01
5.66225410e-01 -7.84446836e-01 8.56274068e-01 -7.67425835e-01
4.27999854e-01 -3.83628547e-01 -9.19738337e-02 -1.75502789e+00
-2.56473362e-01 -5.77089727e-01 1.96002185e-01 6.65592194e-01
1.35205284e-01 -7.14914441e-01 5.61336517e-01 2.58168459e-01
4.41806018e-02 -1.06767520e-01 -8.30978215e-01 -1.12656796e+00
5.07700741e-01 -2.92010009e-01 3.02893877e-01 8.22750628e-01
-8.73944908e-02 4.59351420e-01 -4.99637306e-01 1.52113184e-01
4.80522335e-01 1.12553924e-01 1.33838534e+00 -9.23002303e-01
-5.65483391e-01 -1.09857172e-01 -2.14071333e-01 -1.38967729e+00
7.14484394e-01 -7.27595925e-01 5.37297070e-01 -1.53434789e+00
-1.95669699e-02 -2.61249632e-01 -8.43826458e-02 7.53624856e-01
1.34691671e-01 -6.28229976e-02 4.74738747e-01 3.72294039e-01
-7.97281384e-01 7.81082094e-01 1.62193739e+00 -8.19638744e-02
-3.36424887e-01 -3.22771639e-01 -2.92829871e-01 5.90732038e-01
9.02997673e-01 -4.41621989e-01 -8.04006875e-01 -4.15536731e-01
-1.05908282e-01 2.10954636e-01 4.22653228e-01 -1.20481992e+00
9.35212523e-02 -5.28962433e-01 2.72921920e-01 -3.55678461e-02
5.97647488e-01 -8.17423940e-01 -2.00197712e-01 5.62870026e-01
-4.68377501e-01 7.73236156e-03 5.99293828e-01 6.10383570e-01
-7.60175381e-03 5.98589256e-02 1.02417266e+00 -2.56479144e-01
-1.15616190e+00 8.26322287e-02 -6.01626277e-01 1.25867024e-01
1.40195262e+00 -1.93063930e-01 -2.88093388e-01 -6.67657375e-01
-6.85662031e-01 2.11866707e-01 9.58229780e-01 6.13990486e-01
5.74058890e-01 -1.13487542e+00 -7.00095952e-01 2.57002890e-01
3.32326651e-01 -1.42662138e-01 1.88194290e-02 4.20095325e-01
-3.75541925e-01 1.19203441e-01 -8.61950517e-01 -7.07777739e-01
-7.99169838e-01 6.53227389e-01 3.26541573e-01 -4.35026050e-01
-6.89354539e-01 4.62895483e-01 4.76409733e-01 -5.12650311e-01
-1.03636406e-01 -1.83745578e-01 -8.53362679e-02 -3.37852031e-01
1.92120165e-01 -5.24160974e-02 -4.10660416e-01 -6.71691120e-01
-1.26382276e-01 4.66953486e-01 -1.18947938e-01 -7.42243171e-01
1.14784479e+00 -1.42300338e-01 4.00109053e-01 5.73986709e-01
8.54651690e-01 -4.37369704e-01 -2.13421798e+00 -2.41758704e-01
-3.49598318e-01 -2.91756898e-01 -6.22830428e-02 -8.71474028e-01
-7.26077914e-01 8.35055530e-01 6.84047818e-01 -3.92951041e-01
9.86994982e-01 -6.02005124e-02 5.60424805e-01 5.00995159e-01
7.45045066e-01 -1.18145216e+00 6.02318048e-01 7.04860270e-01
1.06809521e+00 -1.38482213e+00 -3.17799486e-02 3.47943783e-01
-1.27976203e+00 8.57833803e-01 8.02643597e-01 -4.36579466e-01
5.72828725e-02 7.55978376e-02 2.02310130e-01 1.23023644e-01
-9.35182571e-01 -4.84544069e-01 -1.67953596e-01 8.83320093e-01
-2.31443276e-03 -1.35809034e-01 1.72556862e-01 -6.84455559e-02
7.31685534e-02 3.29441279e-01 6.05850995e-01 1.21723032e+00
-4.16474283e-01 -1.16163111e+00 -1.62137970e-01 -1.96131021e-01
2.76664998e-02 3.25843573e-01 -8.43675807e-02 1.19491732e+00
-4.44634110e-02 6.59159720e-01 2.79647298e-03 -1.10645942e-01
3.66058767e-01 9.62459221e-02 8.43218982e-01 -8.06066811e-01
-2.78535306e-01 -4.62235585e-02 -1.44615963e-01 -8.04432392e-01
-4.54359204e-01 -8.13696682e-01 -1.54371870e+00 2.57945596e-03
2.12747067e-01 -1.28704056e-01 7.60579586e-01 9.48981404e-01
3.98833215e-01 6.43041968e-01 3.27764988e-01 -1.13441765e+00
-6.78914547e-01 -8.56433570e-01 -4.27852780e-01 6.67579949e-01
5.14197409e-01 -1.03182161e+00 -6.79341927e-02 3.19806963e-01] | [4.467667102813721, 1.0221922397613525] |
c837b439-bd90-4465-94da-5aaaa3e82b09 | exploring-the-role-of-audio-in-video | 2306.12559 | null | https://arxiv.org/abs/2306.12559v1 | https://arxiv.org/pdf/2306.12559v1.pdf | Exploring the Role of Audio in Video Captioning | Recent focus in video captioning has been on designing architectures that can consume both video and text modalities, and using large-scale video datasets with text transcripts for pre-training, such as HowTo100M. Though these approaches have achieved significant improvement, the audio modality is often ignored in video captioning. In this work, we present an audio-visual framework, which aims to fully exploit the potential of the audio modality for captioning. Instead of relying on text transcripts extracted via automatic speech recognition (ASR), we argue that learning with raw audio signals can be more beneficial, as audio has additional information including acoustic events, speaker identity, etc. Our contributions are twofold. First, we observed that the model overspecializes to the audio modality when pre-training with both video and audio modality, since the ground truth (i.e., text transcripts) can be solely predicted using audio. We proposed a Modality Balanced Pre-training (MBP) loss to mitigate this issue and significantly improve the performance on downstream tasks. Second, we slice and dice different design choices of the cross-modal module, which may become an information bottleneck and generate inferior results. We proposed new local-global fusion mechanisms to improve information exchange across audio and video. We demonstrate significant improvements by leveraging the audio modality on four datasets, and even outperform the state of the art on some metrics without relying on the text modality as the input. | ['Heng Wang', 'Ehsan Elhamifar', 'Haichao Yu', 'Longyin Wen', 'Linjie Yang', 'YuHan Shen'] | 2023-06-21 | null | null | null | null | ['video-captioning', 'automatic-speech-recognition'] | ['computer-vision', 'speech'] | [ 5.59709549e-01 1.80259228e-01 5.14239296e-02 -3.91495228e-01
-1.51006055e+00 -7.00273812e-01 6.69267595e-01 -2.50897646e-01
-2.60641932e-01 5.60552895e-01 6.38741255e-01 -2.69920230e-01
3.90698612e-01 -1.88679919e-01 -1.05018663e+00 -5.07577658e-01
9.97501165e-02 5.94160929e-02 1.06673770e-01 2.89201252e-02
-2.34581888e-01 -1.39866590e-01 -1.68813288e+00 8.44946384e-01
5.35833597e-01 1.31684077e+00 1.73891827e-01 9.08893049e-01
-1.40348196e-01 8.88811171e-01 -5.76550186e-01 -4.99326944e-01
2.93307692e-01 -5.94076872e-01 -6.77562416e-01 9.83752962e-03
5.74545324e-01 -6.10765219e-01 -5.98137796e-01 7.67765105e-01
6.03337824e-01 -4.36672270e-02 3.89154851e-01 -1.44052529e+00
-3.46318722e-01 9.08257544e-01 -5.22250354e-01 -1.37849107e-01
5.12875378e-01 2.58142442e-01 1.24641347e+00 -9.63713229e-01
5.08784950e-01 1.18734002e+00 5.71320593e-01 6.63856864e-01
-1.03781199e+00 -7.86116242e-01 3.06584001e-01 4.79855925e-01
-1.31154478e+00 -9.99858022e-01 7.42029011e-01 -4.69371468e-01
7.22096682e-01 3.95807862e-01 3.10704380e-01 1.59683132e+00
-3.24337214e-01 1.10287130e+00 7.40670741e-01 -2.92874157e-01
5.99473566e-02 1.18077293e-01 -3.06963205e-01 1.66357249e-01
-4.44107205e-01 1.03857584e-01 -9.09351647e-01 -4.96771894e-02
4.73918468e-01 -3.55266750e-01 -6.31235957e-01 -1.03055365e-01
-1.29324377e+00 5.96423149e-01 1.92731202e-01 1.43380463e-01
-3.65348011e-01 2.93256193e-01 6.03474975e-01 3.02184254e-01
3.02316159e-01 2.49738589e-01 -3.60588521e-01 -4.39814717e-01
-1.31855989e+00 -2.34008893e-01 5.69577515e-01 8.64673197e-01
6.77912593e-01 1.83161050e-02 -4.33038682e-01 8.62708569e-01
3.70009601e-01 5.98241031e-01 3.30510676e-01 -1.24342561e+00
9.71088946e-01 1.47712631e-02 5.01933768e-02 -7.16731787e-01
-1.27385482e-01 -2.41005868e-01 -6.15075648e-01 -1.41251117e-01
4.73052025e-01 -4.77524072e-01 -1.09114075e+00 1.93101037e+00
1.77386478e-02 6.00793660e-01 4.12016846e-02 1.23532367e+00
8.51850033e-01 8.50778997e-01 9.94548127e-02 -7.32318610e-02
1.23707664e+00 -1.04838490e+00 -8.59591186e-01 -2.41389483e-01
4.65935349e-01 -8.45674932e-01 8.71605814e-01 3.93381536e-01
-1.10050440e+00 -5.80491900e-01 -9.89533186e-01 -4.96354476e-02
3.28051783e-02 1.27509430e-01 3.66261333e-01 5.45171916e-01
-1.25117183e+00 3.50523353e-01 -8.71092558e-01 -2.62141585e-01
3.03775549e-01 1.85525760e-01 -4.55723405e-01 -1.56345725e-01
-1.30550230e+00 4.93187338e-01 2.83379376e-01 8.63854066e-02
-1.08452427e+00 -7.11844981e-01 -9.71519053e-01 2.48759955e-01
5.65789700e-01 -6.25531852e-01 1.38875520e+00 -1.45902967e+00
-1.58037329e+00 3.23532641e-01 -2.83451408e-01 -6.24753177e-01
6.44380331e-01 -2.78789520e-01 -3.85702193e-01 6.25258088e-01
-2.75388628e-01 1.16197050e+00 1.09362411e+00 -1.31950474e+00
-5.38125873e-01 2.48386972e-02 2.40785480e-01 3.44236344e-01
-5.19859314e-01 3.20542865e-02 -8.09340060e-01 -6.57136261e-01
-1.55994922e-01 -1.08163440e+00 1.42570928e-01 5.47618195e-02
-3.19897652e-01 2.38652393e-01 8.01448941e-01 -1.02260458e+00
1.12637794e+00 -2.40842676e+00 -2.92302519e-02 1.66137218e-02
-4.81902324e-02 2.81784475e-01 -4.41356212e-01 5.16391695e-01
-1.26804739e-01 1.31902978e-01 -1.74581155e-01 -7.03113437e-01
1.04842568e-03 1.07906505e-01 -6.19773626e-01 1.54395595e-01
4.76576656e-01 7.29513586e-01 -7.11004257e-01 -6.82319641e-01
1.72120064e-01 8.81138146e-01 -7.22408772e-01 1.81012183e-01
-1.93042219e-01 5.03138363e-01 -5.81248812e-02 5.24054706e-01
5.66401958e-01 -4.32471901e-01 1.65973052e-01 -3.32629532e-01
2.69854397e-01 5.90104580e-01 -1.18889689e+00 1.92238760e+00
-5.82050741e-01 8.38129282e-01 5.68924665e-01 -8.09069276e-01
4.80706453e-01 9.92860436e-01 5.67590535e-01 -6.23349011e-01
4.84205671e-02 1.54634088e-01 -2.04117060e-01 -4.64661211e-01
4.24335271e-01 -3.30042318e-02 8.25173110e-02 2.39736125e-01
3.08267117e-01 2.18526825e-01 -3.17306668e-02 4.49766904e-01
1.16273570e+00 3.20834249e-01 -2.39152148e-01 4.11990404e-01
3.92122924e-01 -1.19574986e-01 3.20423782e-01 6.88203633e-01
-6.91352785e-02 1.21381283e+00 3.98737490e-01 3.11820656e-01
-9.16370511e-01 -9.06044006e-01 1.60402179e-01 1.30725718e+00
9.11001787e-02 -6.72459602e-01 -7.37794340e-01 -6.70345843e-01
-3.72172445e-01 6.22942984e-01 -3.19996983e-01 7.01026758e-03
-5.33277392e-01 -3.40264678e-01 7.94003725e-01 6.73116088e-01
2.25756228e-01 -6.40596807e-01 -2.42484078e-01 2.36385807e-01
-7.73537457e-01 -1.46159685e+00 -7.26869643e-01 4.53755483e-02
-5.87156951e-01 -7.21071720e-01 -9.34571564e-01 -3.10647696e-01
1.89881414e-01 2.71350771e-01 8.52531016e-01 -2.87254333e-01
1.53719798e-01 6.73409641e-01 -6.56864345e-01 -2.04008728e-01
-4.83986557e-01 7.44595975e-02 -1.97399706e-01 4.93836552e-01
4.49161604e-02 -6.34519517e-01 -5.85042417e-01 3.39066833e-01
-9.93710637e-01 3.90932947e-01 6.43221915e-01 8.30800593e-01
2.24532977e-01 -4.30824727e-01 6.56070650e-01 -4.68601823e-01
1.20920636e-01 -4.98593360e-01 -2.27222905e-01 6.14408068e-02
-1.45678729e-01 4.55407538e-02 6.88130736e-01 -5.09608626e-01
-1.12060332e+00 2.51227558e-01 -3.13954383e-01 -9.40167487e-01
-9.33178663e-02 4.73160893e-01 -3.21187586e-01 1.87969208e-01
3.70728225e-01 4.58310433e-02 6.30480722e-02 -6.20420814e-01
3.13357681e-01 1.08436620e+00 6.27566755e-01 -5.12720168e-01
5.88895380e-01 5.90683222e-01 -2.16367438e-01 -6.63930357e-01
-6.43908024e-01 -6.25832081e-01 -9.32762474e-02 -2.64936298e-01
7.82012641e-01 -1.44695461e+00 -5.35608113e-01 1.48357064e-01
-1.29192400e+00 -6.41730875e-02 -2.62920316e-02 7.13090837e-01
-5.23589432e-01 4.78675753e-01 -7.13538706e-01 -8.89158189e-01
-2.55545139e-01 -1.23508680e+00 1.35974407e+00 -1.41639635e-01
-1.85541853e-01 -5.81160963e-01 -2.48303130e-01 6.21027231e-01
4.18874115e-01 -1.28298834e-01 3.94106209e-01 -8.58768225e-01
-6.84356689e-01 -8.59768167e-02 -4.22531337e-01 4.04596120e-01
-1.74560562e-01 -1.16384134e-01 -1.61846244e+00 -3.06203097e-01
-2.57202059e-01 -4.39008355e-01 1.04227507e+00 1.17787994e-01
1.17727709e+00 -3.10471475e-01 -1.76030800e-01 4.83365506e-01
1.13420892e+00 4.93974537e-02 7.16412544e-01 -2.17032321e-02
7.82772005e-01 7.23240614e-01 4.73402947e-01 3.75936955e-01
4.69576389e-01 9.40049827e-01 4.99508470e-01 -2.85432432e-02
-5.29488504e-01 -4.75551218e-01 8.60886574e-01 8.77688050e-01
1.64442644e-01 -5.12994528e-01 -8.19080830e-01 7.43520439e-01
-1.87823510e+00 -1.00261927e+00 1.11734696e-01 2.15783787e+00
8.87469232e-01 -1.11289218e-01 1.23768494e-01 1.96021438e-01
7.79012501e-01 3.76895480e-02 -3.73268902e-01 2.02289689e-02
-1.44230381e-01 6.35018200e-02 4.39613104e-01 4.59211320e-01
-1.10196567e+00 4.50422764e-01 5.95403767e+00 9.26647067e-01
-1.41235864e+00 2.58797497e-01 4.36844647e-01 -5.39924622e-01
-3.89115423e-01 -4.91825193e-02 -5.39910674e-01 6.20791137e-01
1.44747567e+00 1.63397118e-01 4.18442398e-01 4.17078704e-01
2.93101043e-01 1.27230421e-01 -1.39556861e+00 1.22667253e+00
1.38150886e-01 -1.07468081e+00 1.20157771e-01 -4.21848968e-02
4.25887465e-01 1.81574628e-01 2.86569111e-02 3.92703950e-01
-1.03230894e-01 -9.59614635e-01 1.05240691e+00 1.43229768e-01
9.33789670e-01 -4.03807819e-01 7.80206740e-01 9.55817550e-02
-1.24490595e+00 -6.97248131e-02 6.67091981e-02 6.04344010e-02
4.39522684e-01 4.73270476e-01 -9.70110178e-01 8.34712386e-01
7.34501600e-01 5.24276376e-01 -4.83060867e-01 1.16935861e+00
1.57531106e-03 9.62922633e-01 -6.78595066e-01 4.11909759e-01
2.71095514e-01 2.99308211e-01 6.06308281e-01 1.35286367e+00
4.80440319e-01 -2.10095957e-01 2.08813131e-01 4.61997062e-01
-1.76445380e-01 5.87694570e-02 -3.85913402e-01 -2.44322792e-01
4.08016831e-01 1.09706175e+00 -3.35579246e-01 -3.88910681e-01
-7.61235714e-01 8.78526926e-01 -2.14362338e-01 5.45031726e-01
-1.13042724e+00 -3.46411049e-01 5.64549506e-01 1.69219151e-01
5.09059608e-01 7.19166100e-02 -1.77826896e-01 -1.33908474e+00
2.62457550e-01 -9.01757002e-01 3.78017217e-01 -9.49369788e-01
-1.08127439e+00 6.09357536e-01 -5.15778922e-02 -1.44740498e+00
-6.43460751e-01 -4.76135880e-01 -2.59502351e-01 7.53218532e-01
-1.66376042e+00 -1.17554438e+00 -3.91030051e-02 6.58051252e-01
4.66375858e-01 1.99738428e-01 4.32334304e-01 8.34486365e-01
-2.85143286e-01 9.68171418e-01 -2.03565732e-02 3.10350180e-01
1.13388133e+00 -8.38144422e-01 7.10319653e-02 8.16989362e-01
4.15501922e-01 2.61929482e-01 8.40463877e-01 -3.17505717e-01
-1.43521881e+00 -1.01864302e+00 6.27826810e-01 -2.93060929e-01
6.36397302e-01 -6.05780125e-01 -9.08989370e-01 6.34059548e-01
3.55091155e-01 -1.00148082e-01 7.28541613e-01 -3.31384875e-03
-5.84031403e-01 -5.09325713e-02 -7.21967638e-01 5.98805666e-01
9.82346594e-01 -9.08598781e-01 -3.20209831e-01 -5.96471764e-02
1.07542098e+00 -3.85962754e-01 -6.99885011e-01 3.63776982e-01
5.03743529e-01 -7.56762505e-01 8.26880932e-01 -3.60190630e-01
5.85063815e-01 -4.17224795e-01 -3.81553739e-01 -1.08211279e+00
2.36280486e-01 -8.64425480e-01 -2.46498555e-01 1.54190052e+00
6.43374979e-01 -2.70341069e-01 6.84624612e-01 6.13729060e-01
-2.67662168e-01 -3.47595096e-01 -1.08975995e+00 -6.89572155e-01
-3.02426428e-01 -7.74206340e-01 4.99107480e-01 7.38351107e-01
1.56743377e-01 5.41102767e-01 -8.35716009e-01 2.61723369e-01
1.91396832e-01 1.50400111e-02 6.49076521e-01 -9.54910100e-01
-5.24509728e-01 -6.00129962e-02 -2.35008061e-01 -1.29114342e+00
1.27154604e-01 -8.12698007e-01 3.21379453e-01 -1.37331581e+00
6.70998469e-02 7.39762885e-03 -3.82529557e-01 6.93712890e-01
-8.99102464e-02 4.94683534e-01 6.42069161e-01 2.14566678e-01
-7.87412643e-01 6.22974694e-01 1.04982984e+00 -1.39159441e-01
-8.77352655e-02 -2.98368990e-01 -7.16386020e-01 5.02007842e-01
4.62383837e-01 -3.63186002e-01 -4.16964442e-01 -7.16824949e-01
1.22572295e-01 5.23959756e-01 5.44560432e-01 -1.07511997e+00
3.67595911e-01 1.94583029e-01 2.25656331e-01 -4.37735051e-01
9.17750299e-01 -1.03920603e+00 2.31416136e-01 -1.88121751e-01
-5.43261647e-01 -3.34415823e-01 4.26240325e-01 6.95755124e-01
-5.14852822e-01 2.38446873e-02 4.38902825e-01 2.02314973e-01
-4.28913504e-01 4.23635356e-02 -2.88846612e-01 2.16971003e-02
6.76632881e-01 -6.62176460e-02 -3.78572702e-01 -9.70364392e-01
-6.47893310e-01 3.30098897e-01 4.36165392e-01 5.94847083e-01
4.33119416e-01 -1.28383064e+00 -8.13429058e-01 1.92955621e-02
9.41174775e-02 -2.79444963e-01 4.35233593e-01 9.90229130e-01
-4.07329062e-03 5.37410378e-01 7.51937553e-02 -8.32451105e-01
-1.24046493e+00 5.53626716e-01 5.05448170e-02 -1.74678043e-02
-3.91172469e-01 6.17135525e-01 4.18421686e-01 -4.29029763e-02
5.10351777e-01 -3.63563420e-03 -8.74794871e-02 3.07959974e-01
6.78151071e-01 9.19175148e-02 2.10653171e-01 -7.83367276e-01
-3.34019601e-01 3.59116852e-01 2.03667935e-02 -5.96401095e-01
1.11015022e+00 -4.58277017e-01 3.72495919e-01 4.25997645e-01
1.40120041e+00 3.48463170e-02 -1.48452795e+00 -2.40873992e-01
-2.67600477e-01 -3.66427541e-01 1.62558794e-01 -9.36811745e-01
-1.16474807e+00 1.20687556e+00 4.30890888e-01 8.78267586e-02
1.31210947e+00 -2.16236692e-02 1.07062519e+00 1.99662507e-01
3.24894130e-01 -8.99142146e-01 -3.05792466e-02 3.70792031e-01
7.29415298e-01 -1.29845130e+00 -4.07008410e-01 -3.45607638e-01
-9.19371724e-01 9.92951274e-01 3.87641400e-01 4.08440530e-01
3.91577363e-01 2.81563580e-01 2.38656163e-01 3.32064599e-01
-1.06604803e+00 -3.52641106e-01 4.71288085e-01 4.34216946e-01
4.26637679e-01 -2.72661984e-01 2.32661501e-01 7.37362921e-01
-6.61704317e-02 6.69238567e-02 4.95291263e-01 7.48319924e-01
-1.31842315e-01 -1.06329274e+00 -6.33621395e-01 3.17946821e-01
-7.33564615e-01 -3.58501673e-01 -3.06782871e-01 3.68423998e-01
-1.14220962e-01 1.32910299e+00 4.20926437e-02 -5.67512512e-01
3.11510265e-03 3.16027552e-01 3.15850377e-01 -3.78500879e-01
-5.62104166e-01 4.12391275e-01 2.71753311e-01 -7.19972908e-01
-4.88744676e-01 -5.03585160e-01 -1.02756202e+00 -3.60517614e-02
-3.71230721e-01 2.74163514e-01 7.35808551e-01 8.36362958e-01
7.50862122e-01 5.58739305e-01 4.74165678e-01 -1.11509621e+00
-4.73075628e-01 -9.11602020e-01 -1.37781546e-01 3.79017919e-01
6.77726090e-01 -4.74574298e-01 -5.52968442e-01 4.41145062e-01] | [15.129231452941895, 4.866677761077881] |
890427d8-c87a-4e4b-bfdd-dd8760cc55ae | measuring-intelligence-through-games | 1109.1314 | null | https://arxiv.org/abs/1109.1314v1 | https://arxiv.org/pdf/1109.1314v1.pdf | Measuring Intelligence through Games | Artificial general intelligence (AGI) refers to research aimed at tackling the full problem of artificial intelligence, that is, create truly intelligent agents. This sets it apart from most AI research which aims at solving relatively narrow domains, such as character recognition, motion planning, or increasing player satisfaction in games. But how do we know when an agent is truly intelligent? A common point of reference in the AGI community is Legg and Hutter's formal definition of universal intelligence, which has the appeal of simplicity and generality but is unfortunately incomputable. Games of various kinds are commonly used as benchmarks for "narrow" AI research, as they are considered to have many important properties. We argue that many of these properties carry over to the testing of general intelligence as well. We then sketch how such testing could practically be carried out. The central part of this sketch is an extension of universal intelligence to deal with finite time, and the use of sampling of the space of games expressed in a suitably biased game description language. | ['Jürgen Schmidhuber', 'Julian Togelius', 'Tom Schaul'] | 2011-09-06 | null | null | null | null | ['motion-planning'] | ['robots'] | [ 3.16265523e-01 4.68207270e-01 8.69191624e-03 2.36090776e-02
-1.43605247e-01 -6.47176504e-01 8.08832169e-01 -2.02755153e-01
-5.44143915e-01 9.15777922e-01 -2.34952867e-01 -6.46439016e-01
-5.97330749e-01 -1.27347124e+00 -2.18647107e-01 -6.58444285e-01
-2.23059151e-02 1.01121724e+00 5.50965965e-01 -8.16601276e-01
5.00505686e-01 3.66722107e-01 -1.83629155e+00 -3.85705531e-01
7.39787042e-01 6.57932937e-01 4.30536605e-02 1.02259851e+00
2.39942037e-02 1.04517102e+00 -7.66112983e-01 -5.04486084e-01
1.28966123e-01 -5.81712425e-01 -1.29489052e+00 1.42084941e-01
-5.15274882e-01 2.75878590e-02 -1.57739878e-01 1.22894931e+00
1.74442068e-01 3.52998525e-01 6.67987943e-01 -1.49386656e+00
-2.67416805e-01 4.65636849e-01 1.83567122e-01 1.07477695e-01
7.82463729e-01 3.47356468e-01 1.19714165e+00 1.08269371e-01
5.92309177e-01 1.13442874e+00 3.87603819e-01 9.66463089e-01
-6.14228904e-01 -3.62279452e-02 -3.58297713e-02 2.47607738e-01
-1.14655471e+00 -1.40268072e-01 3.69666100e-01 -2.45675266e-01
8.88787508e-01 8.17939043e-01 7.22224951e-01 4.85588491e-01
5.37598394e-02 1.03504622e+00 9.38327909e-01 -6.23222113e-01
5.15578330e-01 -8.86703804e-02 1.38088226e-01 5.35583079e-01
4.53891426e-01 4.34988081e-01 -5.86229265e-02 9.25240666e-03
1.10625505e+00 -4.80818808e-01 3.91524509e-02 -4.37122315e-01
-1.27854109e+00 1.23820424e+00 -1.82423383e-01 4.79623079e-01
-2.20727146e-01 6.24900311e-03 4.96597797e-01 6.89700663e-01
-4.17637043e-02 1.14757466e+00 -3.50941926e-01 -6.91381872e-01
-8.49302262e-02 8.54020655e-01 1.36883354e+00 7.48923600e-01
5.21288514e-01 5.85536510e-02 3.87443751e-01 4.00996417e-01
1.88497946e-01 3.64278592e-02 7.24822283e-01 -1.33748710e+00
-9.94499922e-02 6.58628464e-01 2.46396378e-01 -7.61770129e-01
-6.00676179e-01 -7.71337971e-02 -4.74492460e-01 6.63483977e-01
4.75839198e-01 -3.83475661e-01 -3.65160614e-01 1.86149645e+00
2.05493644e-01 1.00720674e-01 5.31399250e-01 6.81953788e-01
6.00736916e-01 4.12205189e-01 -1.90765545e-01 -2.65138835e-01
1.19065678e+00 -8.08986127e-01 -2.78541237e-01 -4.97885466e-01
8.17454219e-01 -1.02656737e-01 8.01972687e-01 4.82361853e-01
-1.30720329e+00 -1.03470258e-01 -1.23461509e+00 2.48814330e-01
-5.13851285e-01 -9.33197021e-01 1.48092449e+00 1.09490967e+00
-1.20318878e+00 5.33359408e-01 -5.16373575e-01 -5.12826741e-01
-1.70035347e-01 7.49512076e-01 -3.29607040e-01 6.83174357e-02
-1.38818920e+00 1.14327109e+00 8.46679986e-01 -3.84010017e-01
-5.61906934e-01 3.60687561e-02 -1.09290421e+00 -2.64874667e-01
9.11977112e-01 -8.27132285e-01 1.64855540e+00 -1.19335926e+00
-1.74042559e+00 9.63779628e-01 1.49592593e-01 -7.00018406e-01
5.32596886e-01 3.58994305e-01 -1.81611583e-01 -1.60064802e-01
1.97849736e-01 3.41489285e-01 2.27732688e-01 -7.77449071e-01
-1.02854633e+00 -5.21974385e-01 1.03900313e+00 5.65866947e-01
2.64948457e-01 2.72098124e-01 -3.57241809e-01 -3.93144041e-01
9.92207974e-02 -1.06946456e+00 -7.71946728e-01 -5.44840276e-01
-6.79672584e-02 -5.81164300e-01 3.36304665e-01 1.58571824e-01
1.02812898e+00 -1.76834536e+00 3.55986834e-01 2.20861509e-01
2.24656522e-01 2.02529758e-01 -1.75312325e-01 3.71614248e-01
5.65342866e-02 1.91287309e-01 -9.97916907e-02 6.01682663e-01
5.07565320e-01 7.08251059e-01 3.08577001e-01 1.30405724e-01
3.72882821e-02 1.29563963e+00 -1.10068786e+00 -2.50043064e-01
3.95857364e-01 -2.31034368e-01 -8.04569483e-01 -1.36167537e-02
-4.67076927e-01 1.16851248e-01 -8.28849435e-01 3.41026604e-01
1.21081341e-03 -1.01062350e-01 2.17963472e-01 8.09197426e-01
-9.09830332e-02 3.56233984e-01 -1.43034732e+00 1.30798423e+00
1.19734108e-01 5.09113908e-01 -1.10549726e-01 -1.30209291e+00
7.50278056e-01 4.25175935e-01 3.97281677e-01 -6.71179712e-01
3.03206533e-01 1.62767544e-01 5.99018514e-01 -3.24897647e-01
5.59530437e-01 -5.19262433e-01 -4.93662775e-01 5.58904350e-01
-2.93767571e-01 -7.53860474e-01 5.55431008e-01 -1.37798607e-01
1.23593068e+00 -1.00885443e-01 7.67361104e-01 -4.51280355e-01
7.79493630e-01 3.49076122e-01 4.24555212e-01 1.21456182e+00
-3.32362801e-01 1.73503101e-01 6.58450961e-01 -7.25633144e-01
-9.91236985e-01 -6.38205826e-01 2.03352958e-01 1.06670940e+00
4.30486113e-01 -3.90711695e-01 -8.97288144e-01 -1.14795275e-01
-4.99496013e-01 7.34990895e-01 -5.60727537e-01 -1.87213838e-01
-2.88893461e-01 -7.28899896e-01 5.71598470e-01 2.20464766e-01
4.93052065e-01 -1.65536237e+00 -1.13415468e+00 4.17923331e-01
-1.50703089e-02 -8.21589172e-01 2.78580755e-01 3.47931981e-01
-7.04011619e-01 -1.22322309e+00 -3.23931724e-01 -7.11880684e-01
6.63304999e-02 1.39273629e-01 1.39695442e+00 4.84295696e-01
-1.36909276e-01 6.33536160e-01 -4.73686427e-01 -7.37339973e-01
-6.89217269e-01 1.07967421e-01 1.34184927e-01 -5.29617906e-01
5.30082345e-01 -3.60442370e-01 -1.16140209e-01 4.11383003e-01
-1.08412063e+00 1.65499717e-01 1.95989400e-01 8.73770773e-01
-3.55045274e-02 6.48753762e-01 3.52069587e-01 -9.51630175e-01
1.00598073e+00 -3.47120941e-01 -3.40206861e-01 1.93979129e-01
-1.27803177e-01 7.99921826e-02 3.13806027e-01 -2.60199070e-01
-5.02768338e-01 -3.09670985e-01 -3.62613529e-01 4.79918927e-01
-4.34734613e-01 6.38061941e-01 -5.31184196e-01 -2.28530958e-01
6.18265510e-01 2.64797777e-01 2.32506648e-01 3.18452150e-01
1.79341510e-01 4.12117600e-01 5.61709821e-01 -9.66307104e-01
6.65938973e-01 5.76627478e-02 2.26973325e-01 -9.92822289e-01
-4.83450949e-01 -2.55286902e-01 -3.34074527e-01 -2.20183656e-02
6.42439544e-01 -1.61744535e-01 -1.25063825e+00 4.26369488e-01
-7.35038459e-01 -5.84433496e-01 -6.61361098e-01 1.81102619e-01
-1.35472178e+00 2.79542208e-01 -3.50849032e-01 -1.11269164e+00
1.30761340e-01 -1.27205026e+00 5.93141317e-01 3.69961143e-01
-5.08657217e-01 -1.11034012e+00 1.60378128e-01 2.26271585e-01
1.35377884e-01 1.06823005e-01 1.05456054e+00 -1.10478806e+00
-1.50866732e-01 -4.38313603e-01 1.86960727e-01 1.72384351e-01
-3.12231630e-02 -1.44006521e-01 -6.26709998e-01 -9.82039720e-02
1.78109571e-01 -3.41631562e-01 6.60736114e-02 3.66813779e-01
6.41817868e-01 -2.77587324e-01 6.77510723e-02 1.57201946e-01
1.46649086e+00 1.04418743e+00 1.12018228e+00 9.18008149e-01
1.33546889e-01 6.69092298e-01 8.31579268e-01 5.54910481e-01
3.80888313e-01 6.93968296e-01 5.98750830e-01 2.27393210e-01
6.52321994e-01 3.35138708e-01 -4.98580467e-03 4.16490197e-01
-7.76963770e-01 -2.92125344e-01 -1.04960346e+00 4.55946326e-01
-2.16126347e+00 -1.39351010e+00 7.13791549e-02 2.14747715e+00
5.92836559e-01 4.35607404e-01 4.80174899e-01 6.62636936e-01
5.58271527e-01 -4.48636450e-02 -4.49782759e-01 -9.23777819e-01
-1.86789826e-01 3.31936091e-01 2.83225268e-01 7.83724010e-01
-1.09271896e+00 1.10231912e+00 7.15178204e+00 8.92922699e-01
-6.20358944e-01 -2.03301251e-01 5.64979970e-01 4.27708387e-01
-1.16405778e-01 6.77870065e-02 -2.49251381e-01 2.25995183e-01
8.90725136e-01 -5.57912707e-01 6.58360481e-01 8.11619282e-01
-1.02831692e-01 -2.49044627e-01 -9.74715352e-01 7.46957600e-01
-2.09685445e-01 -1.22138143e+00 -2.97023263e-02 2.87669629e-01
4.67658341e-01 -3.59394729e-01 -2.06728339e-01 5.05960882e-01
8.76627088e-01 -1.39129639e+00 6.18024349e-01 1.72736004e-01
3.06639403e-01 -1.17639029e+00 8.91573370e-01 7.71001875e-01
-9.22779322e-01 -1.03614241e-01 -4.57898855e-01 -9.73575175e-01
-2.65050381e-01 -2.54193246e-01 -4.95930642e-01 5.89003861e-01
1.10551514e-01 2.22710893e-01 -1.94163084e-01 1.19107056e+00
-5.74912131e-02 8.31067711e-02 -4.08556491e-01 -6.28963590e-01
7.57148027e-01 -4.66763318e-01 6.05401576e-01 7.73841798e-01
8.17518681e-02 6.70216918e-01 1.75585032e-01 4.28057581e-01
5.89538455e-01 -1.98957458e-01 -9.99042988e-01 -2.13024959e-01
2.11067900e-01 6.97372019e-01 -7.64609277e-01 -2.61037350e-01
-4.97673303e-01 8.34913909e-01 -3.84461544e-02 -1.25959218e-01
-7.13043630e-01 -4.37249601e-01 9.24585879e-01 -1.11297823e-01
-6.09547272e-02 -2.32589960e-01 -3.26192826e-01 -1.04714274e+00
-3.75041008e-01 -1.33162355e+00 5.12620509e-01 -6.77893758e-01
-8.17974627e-01 5.59852481e-01 -1.09862484e-01 -8.35260332e-01
-1.06956959e+00 -9.93292212e-01 -7.94786870e-01 5.99065185e-01
-8.64461064e-01 -1.00407612e+00 -1.92963071e-02 7.19425380e-01
6.11363113e-01 -3.18257421e-01 1.01171899e+00 -3.46024126e-01
-3.72910798e-01 3.10895383e-01 -2.57283866e-01 1.26312062e-01
-2.96149611e-01 -1.42456806e+00 5.76345801e-01 8.24165404e-01
5.70085533e-02 5.19957721e-01 1.07859516e+00 -2.18772575e-01
-1.67213798e+00 -5.51209331e-01 6.84714317e-01 -6.10176504e-01
9.09433901e-01 2.10166529e-01 -5.02719641e-01 7.68943787e-01
-7.49396235e-02 -4.54062164e-01 2.23540589e-01 2.17707723e-01
9.35183242e-02 5.44784009e-01 -8.67807925e-01 1.02216959e+00
1.09543133e+00 -3.01403522e-01 -7.12209105e-01 2.31916130e-01
5.41545153e-01 -3.94145787e-01 -7.84440517e-01 4.28299546e-01
4.97145116e-01 -1.17143083e+00 9.94720817e-01 -1.07586229e+00
2.54561931e-01 -3.23022038e-01 -1.44506976e-01 -1.02794230e+00
-5.30099571e-01 -1.02340746e+00 3.13764960e-01 6.35327339e-01
2.86609679e-01 -9.23210382e-01 1.18299782e+00 1.00988078e+00
3.53812836e-02 -7.15274990e-01 -9.44955707e-01 -1.05092704e+00
3.57576400e-01 -9.00016427e-01 7.84968674e-01 8.56858313e-01
6.37668967e-01 4.20785099e-01 -9.39656794e-02 -1.87334150e-01
3.34461838e-01 -8.26314911e-02 8.59413445e-01 -1.70182693e+00
-5.56188464e-01 -9.61277187e-01 -1.09104395e+00 -8.81465852e-01
9.04444754e-02 -4.14697230e-01 1.30145043e-01 -1.42838824e+00
6.55971467e-02 -5.54984629e-01 2.25004461e-02 2.88170099e-01
2.34395694e-02 2.11927876e-01 2.01679841e-01 2.14789659e-02
-9.98049259e-01 -9.52559337e-02 1.37855542e+00 -8.95219892e-02
-2.80636959e-02 3.54124427e-01 -1.19124103e+00 1.07100964e+00
9.51683283e-01 2.13189393e-01 -6.82689309e-01 -4.10637259e-02
6.58156633e-01 3.12350005e-01 2.38928974e-01 -1.21314287e+00
4.99936163e-01 -8.96092415e-01 -4.63713825e-01 2.40850151e-01
2.92260319e-01 -6.00456417e-01 3.98138851e-01 6.62072718e-01
-1.63090438e-01 8.85482877e-02 -6.20306795e-03 2.79259477e-02
-2.99346983e-01 -9.55113530e-01 7.13821828e-01 -6.59558475e-01
-1.23058820e+00 2.61707425e-01 -7.41469920e-01 3.57261151e-01
1.30681276e+00 -7.99353719e-01 -1.27318040e-01 -7.37347364e-01
-5.70466995e-01 4.99999598e-02 7.62903333e-01 9.30627063e-02
4.22573090e-01 -9.35207307e-01 -5.95710635e-01 5.02930060e-02
1.74045354e-01 -4.44701947e-02 -1.10593982e-01 4.13454145e-01
-8.01392794e-01 6.44651234e-01 -4.18391615e-01 -5.36427833e-02
-1.15468740e+00 6.41701102e-01 6.28183246e-01 -4.40530121e-01
-5.40563464e-01 7.82434404e-01 4.73130494e-01 -3.39480698e-01
1.10937804e-01 -2.31134035e-02 -4.13634896e-01 -5.10912001e-01
8.04057419e-01 8.15547854e-02 -9.41966102e-02 -5.07808626e-01
-3.93191576e-01 2.29514480e-01 1.93907619e-01 -4.32492584e-01
1.26701903e+00 9.18558165e-02 -3.75564992e-01 2.77589053e-01
5.05644798e-01 -4.50984538e-01 -6.12103820e-01 7.90479332e-02
2.89969414e-01 -9.32395160e-02 -4.47270185e-01 -5.19483984e-01
-5.53774893e-01 5.00562131e-01 1.38479657e-02 9.98616934e-01
1.07675958e+00 -2.15500742e-02 2.64387131e-01 6.48368299e-01
1.09967875e+00 -1.01015246e+00 -3.88293505e-01 1.19881654e+00
5.45301259e-01 -1.10896540e+00 -2.18753353e-01 -2.85454392e-01
-8.12874198e-01 1.03733444e+00 5.41538179e-01 -2.59292901e-01
2.01589763e-01 4.05037493e-01 -1.87917799e-01 -3.33619714e-01
-9.31283355e-01 -6.61728799e-01 -1.29184872e-01 1.04149270e+00
2.46895388e-01 2.88984984e-01 -4.23230141e-01 6.37867332e-01
-6.48270547e-01 -4.34915861e-03 6.82261169e-01 1.13229668e+00
-8.62765551e-01 -1.15861320e+00 -3.24994355e-01 3.78683358e-01
-3.44278276e-01 8.90693665e-02 -6.22650445e-01 1.11555076e+00
7.81309903e-02 1.10866964e+00 -1.14526249e-01 -5.03174245e-01
5.50396927e-02 -1.99278593e-01 6.60055637e-01 -6.63361192e-01
-4.81332064e-01 -5.87473094e-01 4.00064856e-01 -4.43334401e-01
-5.40898204e-01 -5.81080019e-01 -1.11300981e+00 -7.93209136e-01
-7.13260695e-02 7.37178028e-01 3.42283309e-01 1.25909281e+00
-4.00030404e-01 2.95076966e-01 2.29306012e-01 -8.09858084e-01
-2.75045842e-01 -4.63506460e-01 -8.22282970e-01 2.85161257e-01
1.20442698e-03 -4.44544792e-01 -2.61758178e-01 -3.89207780e-01] | [3.510091781616211, 1.4715774059295654] |
d614a1fd-3e32-479a-b347-54c6fe7bd6ed | ynu-hpcc-at-semeval-2022-task-6-transformer | null | null | https://aclanthology.org/2022.semeval-1.134 | https://aclanthology.org/2022.semeval-1.134.pdf | YNU-HPCC at SemEval-2022 Task 6: Transformer-based Model for Intended Sarcasm Detection in English and Arabic | In this paper, we (a YNU-HPCC team) describe the system we built in the SemEval-2022 competition. As participants in Task 6 (titled “iSarcasmEval: Intended Sarcasm Detection In English and Arabic”), we implement the sentiment system for all three subtasks in English and Arabic. All subtasks involve the detection of sarcasm (binary and multilabel classification) and the determination of the sarcastic text location (sentence pair classification). Our system primarily applies the sequence classification model of a bidirectional encoder representation from a transformer (BERT). The BERT is used to extract sentence information from both directions for downstream classification tasks. A single basic model is used for single-sentence and sentence-pair binary classification tasks. For the multilabel task, the Label-Powerset method and binary cross-entropy loss function with weights are used. Our system exhibits competitive performance, obtaining 12/43 (21/32), 11/22, and 3/16 (8/13) rankings in the three official rankings for English (Arabic). | ['Xuejie Zhang', 'Jin Wang', 'Guangmin Zheng'] | null | null | null | null | semeval-naacl-2022-7 | ['sentence-pair-classification'] | ['natural-language-processing'] | [ 1.82231650e-01 2.16479808e-01 -1.39649466e-01 -7.10156083e-01
-1.25496602e+00 -7.08971024e-01 4.95125562e-01 3.88076097e-01
-7.94583142e-01 5.85617185e-01 4.59385216e-01 -5.01578748e-01
4.41492289e-01 -2.29813471e-01 -4.54402834e-01 -5.45700312e-01
3.90536398e-01 5.78268170e-01 2.63260137e-02 -6.63679183e-01
4.04381514e-01 -1.15497656e-01 -8.94876719e-01 1.06329501e+00
7.13866234e-01 8.91755581e-01 4.32114154e-02 1.00318658e+00
2.28789911e-01 1.53721547e+00 -6.97202682e-01 -1.08209097e+00
-1.32379904e-01 -6.02432966e-01 -1.35131741e+00 -3.22373182e-01
2.28997931e-01 -8.48292336e-02 4.70956415e-02 1.01946914e+00
7.30907023e-01 -1.62768006e-01 7.69447029e-01 -1.05974889e+00
-4.48149741e-01 1.10555232e+00 -6.34140074e-01 1.41729742e-01
4.23751324e-01 -2.60720670e-01 1.36469769e+00 -1.20215499e+00
5.36849558e-01 1.13908291e+00 8.42666447e-01 4.46384490e-01
-9.55478251e-01 -3.26383233e-01 -3.25825572e-01 3.65342975e-01
-9.44634497e-01 -5.64949453e-01 5.47658920e-01 -6.07085109e-01
1.13750410e+00 4.00132537e-01 4.05222654e-01 1.11356878e+00
2.29201972e-01 1.25902951e+00 1.14450586e+00 -7.47050166e-01
2.04524476e-05 3.99878502e-01 5.16066074e-01 5.90205908e-01
-4.53324080e-01 -3.28691334e-01 -7.54126132e-01 -3.02033946e-02
-3.02127659e-01 -5.62559783e-01 -7.75411353e-02 1.35770038e-01
-1.16118300e+00 1.20592880e+00 3.15600008e-01 2.37684891e-01
8.17016736e-02 -2.45086268e-01 1.05964887e+00 5.18083930e-01
6.39162958e-01 3.50970626e-01 -5.76207638e-01 -3.81481677e-01
-1.09271073e+00 1.70857497e-02 8.70345891e-01 7.54066348e-01
-7.47096613e-02 -3.40875894e-01 -3.44230473e-01 1.24839902e+00
1.53423682e-01 5.86344540e-01 7.90601313e-01 -6.70513213e-01
8.70225012e-01 3.05332154e-01 -1.14707068e-01 -8.57871234e-01
-6.41718626e-01 -2.07142681e-01 -7.16746449e-01 -2.66036272e-01
4.42059815e-01 -1.94889590e-01 -4.15538549e-01 1.58735442e+00
7.23479316e-02 -5.96891820e-01 4.88615006e-01 8.82109642e-01
1.12617302e+00 7.05702066e-01 1.15484402e-01 -1.30095780e-01
1.70281982e+00 -1.40992534e+00 -4.48464602e-01 -4.12064850e-01
1.34669507e+00 -1.22614706e+00 1.05067647e+00 2.06371233e-01
-1.19722331e+00 -4.11779433e-01 -1.02332020e+00 -5.04625857e-01
-1.75146207e-01 6.94500566e-01 -1.11890517e-01 2.95484185e-01
-8.40415895e-01 1.84371591e-01 -4.78966922e-01 -2.97069043e-01
-2.46172801e-01 1.84769347e-01 -3.07682306e-01 1.23088874e-01
-1.47147620e+00 1.26973116e+00 1.47828251e-01 1.61231577e-01
-4.30426031e-01 -2.56585866e-01 -9.81606245e-01 -1.59538135e-01
-3.76844943e-01 -8.48144740e-02 1.60753596e+00 -1.35530698e+00
-1.32017910e+00 1.53927517e+00 -4.87640090e-02 -6.51808739e-01
4.42402750e-01 -2.51632035e-01 -3.45631510e-01 1.24511600e-01
3.97177786e-01 6.08029962e-01 4.89754707e-01 -6.10139430e-01
-4.79746163e-01 -3.38049293e-01 7.85044059e-02 5.90926588e-01
-3.40684235e-01 6.85745239e-01 1.34178683e-01 -7.61872113e-01
-8.51450637e-02 -1.05063033e+00 2.48966143e-01 -7.40203977e-01
-7.05264807e-01 -4.08354610e-01 4.87781376e-01 -1.34084451e+00
1.14955437e+00 -2.18915677e+00 1.93521798e-01 -2.92672336e-01
-8.80203303e-03 5.16087078e-02 -8.91926363e-02 5.09665370e-01
-2.99403429e-01 -1.96693331e-01 -2.07990527e-01 -6.32539511e-01
-3.43586393e-02 -2.93094873e-01 -3.79094929e-01 4.58083749e-01
1.13822713e-01 8.59880805e-01 -9.30348694e-01 -5.13579130e-01
-3.07501644e-01 2.11733505e-01 -2.56716132e-01 7.67214298e-02
2.45486051e-01 1.03583597e-01 -2.84794648e-03 3.55377257e-01
4.83834267e-01 -1.74247399e-01 5.06141186e-01 -2.36300513e-01
-4.40571532e-02 8.72395992e-01 -3.84334266e-01 1.46930814e+00
-5.95264196e-01 6.90555155e-01 1.45116657e-01 -1.03508496e+00
8.72052312e-01 4.25795525e-01 2.17328727e-01 -7.58087277e-01
1.39914826e-01 4.56482768e-01 -1.92130674e-02 -4.59968746e-01
7.04896390e-01 -2.68892556e-01 -7.61486828e-01 5.94422162e-01
3.93449217e-02 -2.11147264e-01 4.62116092e-01 5.22273064e-01
1.06229329e+00 -5.34146614e-02 2.50905901e-01 -4.47735012e-01
9.17440116e-01 1.06367573e-01 5.27658761e-01 3.81838799e-01
-2.85019785e-01 6.47698522e-01 9.25464094e-01 -3.58630747e-01
-1.13192916e+00 -7.54690468e-01 -2.68895440e-02 1.47593868e+00
-2.10469648e-01 -6.75543129e-01 -7.93741882e-01 -1.05575323e+00
-3.17844778e-01 9.14643347e-01 -4.47927415e-01 -2.73136050e-01
-7.32993603e-01 -8.70319963e-01 8.21131229e-01 5.31726480e-01
1.57475546e-01 -1.24847353e+00 -5.41012764e-01 -1.61881801e-02
-8.52821171e-01 -9.77153718e-01 -8.25522065e-01 5.79129279e-01
-4.89669323e-01 -9.80724394e-01 -5.28453529e-01 -1.24530005e+00
3.91041428e-01 -1.84247777e-01 1.34123349e+00 -1.45004392e-01
-9.96796116e-02 -2.61808485e-01 -6.70816243e-01 -1.23861328e-01
-8.39697719e-01 1.66200072e-01 -2.32357550e-02 -2.60289669e-01
6.40557230e-01 2.48791441e-01 -3.30987185e-01 1.30595639e-01
-4.48775500e-01 3.53288591e-01 2.67880023e-01 1.21432686e+00
4.74436246e-02 -5.38706124e-01 6.77362144e-01 -8.75641823e-01
7.61475146e-01 -3.16740751e-01 -2.71664917e-01 2.90403843e-01
-5.18589139e-01 -1.10613391e-01 7.42247581e-01 3.99891622e-02
-9.39500511e-01 9.06352997e-02 -5.15974760e-01 3.70175272e-01
3.56439531e-01 7.21048057e-01 7.59517550e-02 6.35182440e-01
7.58350253e-01 1.72879696e-01 2.35541482e-02 -4.44052279e-01
2.90382922e-01 1.41302705e+00 5.32987475e-01 -2.00328693e-01
2.26763566e-03 -1.12195194e-01 -5.50431490e-01 -5.23205757e-01
-1.47822487e+00 -5.91573358e-01 -7.16824889e-01 -1.53686255e-01
1.10218370e+00 -1.16494286e+00 -5.60230196e-01 7.29937553e-01
-1.32216859e+00 -2.26194620e-01 -1.61807701e-01 2.70695806e-01
-4.96296555e-01 3.73716861e-01 -1.28843844e+00 -7.48101473e-01
-7.48830736e-01 -1.34168077e+00 1.29069686e+00 -2.05788925e-01
-5.56203783e-01 -8.46698999e-01 2.21081004e-01 9.59644437e-01
2.40726098e-02 -4.18914109e-01 1.01934934e+00 -8.92413735e-01
6.20549679e-01 -2.43705228e-01 -2.69592941e-01 7.20488429e-01
-3.27342898e-01 -2.67891407e-01 -9.37827587e-01 -3.87400240e-01
1.20140359e-01 -1.07809889e+00 9.36640680e-01 2.73278981e-01
7.81476974e-01 -2.93587625e-01 1.27912521e-01 2.60412917e-02
9.01033401e-01 -4.90134731e-02 4.21032131e-01 2.16443926e-01
3.69668454e-01 8.40381801e-01 7.24143207e-01 2.97515422e-01
6.81723654e-01 8.46239865e-01 9.04394984e-02 -2.82740057e-03
-2.00455859e-01 -2.10419104e-01 1.16382372e+00 1.43893850e+00
5.46211720e-01 -2.53741801e-01 -8.55473161e-01 5.37449360e-01
-1.90982676e+00 -9.06716406e-01 -5.93957961e-01 1.97903073e+00
1.25166130e+00 1.52136728e-01 2.37686366e-01 2.63297647e-01
7.08662033e-01 5.69785163e-02 -6.15983456e-02 -1.01327431e+00
-4.27899212e-01 -9.02985781e-02 3.50599110e-01 6.69285834e-01
-1.34930253e+00 9.93211985e-01 6.10143185e+00 8.75559092e-01
-8.61467540e-01 3.62945139e-01 1.01647735e+00 -2.58014202e-02
2.26186384e-02 -3.46969627e-02 -1.03192973e+00 7.32187927e-01
1.34343517e+00 2.21900597e-01 8.59748572e-02 6.58560812e-01
7.65167102e-02 -2.96540350e-01 -8.12954724e-01 6.63336873e-01
5.57852447e-01 -1.04174829e+00 -5.15297294e-01 -5.04784822e-01
5.78920424e-01 3.73510212e-01 -1.01636305e-01 5.36841810e-01
3.35468203e-01 -9.19807494e-01 1.11375749e+00 2.47260526e-01
8.20106030e-01 -7.41869271e-01 1.18988121e+00 4.25073564e-01
-7.72008657e-01 -1.41973704e-01 -2.06141755e-01 1.53896688e-02
3.19486260e-01 8.44437182e-01 -8.86832714e-01 2.27908298e-01
7.20526755e-01 9.80748713e-01 -6.36246145e-01 3.85471642e-01
-6.24300420e-01 8.91154051e-01 -5.14965132e-02 -2.74233937e-01
9.49093401e-02 -1.27485871e-01 5.22107780e-01 1.48924005e+00
-1.00622093e-03 -4.33578134e-01 6.80766106e-02 1.94830611e-01
-2.91444719e-01 5.29069364e-01 -1.17222846e-01 3.53113897e-02
1.02922760e-01 1.46421254e+00 -6.28689766e-01 -3.92680198e-01
-2.75382102e-01 1.23107302e+00 5.77230930e-01 -1.48365453e-01
-8.03757012e-01 -6.29966140e-01 -2.74337456e-02 -3.56068075e-01
1.46098614e-01 1.97637826e-01 -4.81285095e-01 -1.16741645e+00
1.35892425e-02 -1.07038033e+00 5.56871593e-01 -9.17694569e-01
-1.41547132e+00 7.88068116e-01 -4.38586682e-01 -8.61019790e-01
-4.39907253e-01 -7.93361783e-01 -4.37417060e-01 9.25740719e-01
-1.03685057e+00 -1.27009404e+00 3.53061020e-01 1.73198730e-01
7.20233679e-01 -3.82825971e-01 8.16782713e-01 3.20360005e-01
-4.88277614e-01 7.66984940e-01 2.20111653e-01 5.54049432e-01
1.16096377e+00 -1.36637533e+00 2.27978960e-01 6.27498507e-01
1.25633702e-02 1.29986435e-01 6.15261853e-01 -4.99603212e-01
-7.08924174e-01 -9.40188408e-01 1.99492669e+00 -6.15740478e-01
9.68106389e-01 -5.61428130e-01 -3.27885181e-01 4.87555563e-01
3.15703869e-01 -4.62012142e-01 7.25765884e-01 1.42996967e-01
-3.80898356e-01 8.76213610e-02 -8.33511949e-01 2.57144123e-01
4.37380046e-01 -7.84257889e-01 -6.71710014e-01 8.05950582e-01
7.04653740e-01 -3.70058209e-01 -7.72858500e-01 3.17310750e-01
5.22228181e-01 -7.92768598e-01 6.05941832e-01 -9.02593374e-01
1.30270815e+00 2.99093686e-02 -3.60766232e-01 -1.23150361e+00
-1.09981194e-01 -1.89421609e-01 5.75186275e-02 1.08322656e+00
8.60851943e-01 -1.15433052e-01 5.06312788e-01 -1.29848523e-02
-1.99173525e-01 -8.25207710e-01 -9.54209208e-01 -3.49251419e-01
4.21553522e-01 -2.94264168e-01 -6.58569559e-02 8.30119491e-01
6.42135203e-01 1.13338351e+00 -4.24224764e-01 -5.41242361e-01
1.25797763e-01 3.25530469e-01 1.86989248e-01 -7.16979146e-01
-4.41295534e-01 -4.71607804e-01 -1.51237115e-01 -9.77823317e-01
4.35214549e-01 -1.43833148e+00 3.18037093e-01 -1.17523801e+00
6.41115367e-01 -1.65082186e-01 -1.98424429e-01 6.32471263e-01
-9.48513299e-02 5.63321173e-01 2.50013113e-01 2.86264688e-01
-6.89198494e-01 4.88995761e-01 5.46473265e-01 -3.41975272e-01
3.13051313e-01 9.91206616e-02 -7.22322643e-01 6.06463730e-01
8.83635223e-01 -5.73243916e-01 4.95623425e-02 -5.00486016e-01
7.45131493e-01 1.12377428e-01 2.67290860e-01 -5.36823273e-01
2.09698058e-03 3.25554490e-01 8.64421576e-02 -7.00863242e-01
2.94040948e-01 -2.17364565e-01 -4.84400839e-01 6.33895159e-01
-9.53674614e-01 4.49639678e-01 -2.00429201e-01 1.75405126e-02
-3.90897453e-01 -6.81372523e-01 9.42936242e-01 1.12897426e-01
-1.90974846e-01 -5.48146725e-01 -5.80150247e-01 4.80652988e-01
8.02447319e-01 4.39629585e-01 -5.50749600e-01 -4.17698622e-01
-7.98051596e-01 2.82975197e-01 1.68294594e-01 5.05781591e-01
4.04778033e-01 -1.21570218e+00 -1.15233231e+00 7.34242052e-02
2.36114830e-01 -4.84865308e-01 -1.19717522e-02 1.42354238e+00
-6.00568473e-01 4.47848558e-01 2.06940383e-01 -4.28401321e-01
-1.56721377e+00 1.34940177e-01 2.64318645e-01 -7.03039646e-01
-1.87023580e-01 1.27874756e+00 -1.35112479e-01 -1.00987756e+00
-2.75087450e-02 2.75926709e-01 -3.42899024e-01 3.37948829e-01
4.90967900e-01 3.85113627e-01 4.89188999e-01 -9.96242344e-01
-3.85316283e-01 2.55111922e-02 -2.95123249e-01 -3.29731196e-01
1.22454226e+00 -1.37707755e-01 -6.08577609e-01 5.85857570e-01
1.35604167e+00 6.06399253e-02 -5.29297411e-01 -3.00717019e-02
2.13416725e-01 2.72897124e-01 4.04404476e-02 -1.21269405e+00
-6.50785267e-01 1.06301224e+00 2.30910942e-01 2.40381941e-01
8.75414193e-01 3.58002782e-02 9.90804255e-01 9.87317264e-02
-1.47420242e-01 -1.60725617e+00 -1.00031577e-01 1.22054780e+00
9.61783886e-01 -1.23795235e+00 -3.01359296e-01 -5.80771752e-02
-1.16937637e+00 1.16059256e+00 4.94425535e-01 2.33106464e-01
5.70108891e-01 4.32430685e-01 2.85931706e-01 -1.41950890e-01
-7.66351521e-01 3.33806157e-01 3.52289915e-01 -8.32636899e-04
1.10132623e+00 2.15841889e-01 -8.70313346e-01 8.80285025e-01
-5.62225461e-01 -4.87711161e-01 4.82884556e-01 7.68842697e-01
-3.25498641e-01 -1.12900198e+00 -3.29340100e-02 4.58125472e-01
-7.04249382e-01 -5.61522841e-01 -6.33930624e-01 1.48924410e-01
-2.19607070e-01 1.22322786e+00 -5.80245536e-03 -5.40761530e-01
1.22596450e-01 2.43629694e-01 1.75912932e-01 -5.27636707e-01
-1.02718878e+00 -3.70573774e-02 6.08093977e-01 -2.86559969e-01
-3.54155958e-01 -9.85286117e-01 -1.20949566e+00 -1.65681377e-01
-3.93468559e-01 4.77635056e-01 8.02074969e-01 9.42070067e-01
1.25442982e-01 2.60021657e-01 7.61162162e-01 -3.48199695e-01
-1.08054018e+00 -1.31496489e+00 -5.34212589e-01 5.19907176e-01
4.48606163e-02 -1.46522641e-01 -3.66085261e-01 2.71449059e-01] | [9.185791969299316, 10.595486640930176] |
1f94f2d3-19d7-44d7-a887-e48dcd60cf68 | crosslingual-embeddings-are-essential-in-unmt | 2106.04995 | null | https://arxiv.org/abs/2106.04995v1 | https://arxiv.org/pdf/2106.04995v1.pdf | Crosslingual Embeddings are Essential in UNMT for Distant Languages: An English to IndoAryan Case Study | Recent advances in Unsupervised Neural Machine Translation (UNMT) have minimized the gap between supervised and unsupervised machine translation performance for closely related language pairs. However, the situation is very different for distant language pairs. Lack of lexical overlap and low syntactic similarities such as between English and Indo-Aryan languages leads to poor translation quality in existing UNMT systems. In this paper, we show that initializing the embedding layer of UNMT models with cross-lingual embeddings shows significant improvements in BLEU score over existing approaches with embeddings randomly initialized. Further, static embeddings (freezing the embedding layer weights) lead to better gains compared to updating the embedding layer weights during training (non-static). We experimented using Masked Sequence to Sequence (MASS) and Denoising Autoencoder (DAE) UNMT approaches for three distant language pairs. The proposed cross-lingual embedding initialization yields BLEU score improvement of as much as ten times over the baseline for English-Hindi, English-Bengali, and English-Gujarati. Our analysis shows the importance of cross-lingual embedding, comparisons between approaches, and the scope of improvements in these systems. | ['Pushpak Bhattacharyya', 'Rudra Murthy V', 'Tamali Banerjee'] | 2021-06-09 | null | https://aclanthology.org/2021.mtsummit-research.3 | https://aclanthology.org/2021.mtsummit-research.3.pdf | mtsummit-2021-8 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [-2.53705811e-02 4.92914282e-02 -3.52513731e-01 -4.07237858e-01
-1.18538046e+00 -8.26670527e-01 7.73557186e-01 1.57942958e-02
-7.42984116e-01 8.76624584e-01 6.71270430e-01 -7.48945117e-01
4.75988865e-01 -4.53628182e-01 -7.13301778e-01 -6.21698678e-01
3.26063156e-01 6.43557847e-01 -3.86482805e-01 -4.29683149e-01
-1.29345447e-01 2.41679344e-02 -7.54054487e-01 2.75747925e-01
9.88127887e-01 2.08798304e-01 9.25806090e-02 6.11130595e-01
-1.69052750e-01 3.64409357e-01 -4.23213869e-01 -8.87590706e-01
4.05640692e-01 -7.53043354e-01 -8.06582451e-01 -3.54712099e-01
4.48901832e-01 -2.65978515e-01 -2.71873891e-01 9.26220655e-01
8.74215841e-01 -2.06601545e-01 7.95041442e-01 -9.30530846e-01
-1.42311954e+00 1.03359807e+00 -3.57178628e-01 1.34883031e-01
-1.30836204e-01 -8.30358937e-02 1.26810205e+00 -1.13861322e+00
7.00552523e-01 1.14667010e+00 9.55286264e-01 6.40295088e-01
-1.37053204e+00 -3.34250331e-01 -4.52246428e-01 6.46962523e-02
-1.21164262e+00 -6.15357101e-01 3.54533434e-01 -2.42809772e-01
1.51636398e+00 1.05939314e-01 1.22204408e-01 1.35326493e+00
4.46842700e-01 6.78946316e-01 1.20899582e+00 -9.42970455e-01
-1.72239747e-02 4.26501483e-01 -1.43519357e-01 4.52339768e-01
2.72679061e-01 1.75736204e-01 -5.45947433e-01 -1.10816173e-01
5.65744936e-01 -3.27134609e-01 1.71297342e-01 -9.85165387e-02
-1.73120153e+00 9.99716818e-01 1.83238536e-01 7.18656063e-01
-3.66840810e-01 1.27497196e-01 7.20796883e-01 8.27830553e-01
5.56236327e-01 5.56817293e-01 -6.85217857e-01 -3.40479434e-01
-9.48791206e-01 -2.44423360e-01 6.64666295e-01 9.34477985e-01
7.94845521e-01 3.27899516e-01 -8.15711468e-02 1.09959269e+00
5.33255972e-02 6.21379316e-01 9.72625852e-01 -5.08883953e-01
8.89411688e-01 1.17549241e-01 -1.31371751e-01 -6.49022341e-01
1.47504196e-01 -3.61516804e-01 -7.59408534e-01 -1.83918446e-01
2.16816649e-01 -4.65342522e-01 -8.33372831e-01 1.76570511e+00
-9.96940210e-02 -3.82410139e-01 6.47412121e-01 7.45765626e-01
4.27179843e-01 9.85149622e-01 -1.56993821e-01 -1.30552530e-01
1.16690719e+00 -1.32496345e+00 -9.73242223e-01 -3.75627905e-01
8.50845277e-01 -1.25610352e+00 1.20607173e+00 -2.78272361e-01
-1.00797188e+00 -5.79930186e-01 -1.11565685e+00 -3.87645334e-01
-5.85755289e-01 3.48011285e-01 3.69672418e-01 6.51662827e-01
-1.21076596e+00 5.84589958e-01 -8.26865911e-01 -8.06625128e-01
-1.39534846e-01 5.60490310e-01 -5.13441563e-01 -3.30627039e-02
-1.60344517e+00 1.43071067e+00 4.97151434e-01 -2.58042831e-02
-5.46662211e-01 -3.66471767e-01 -8.94891560e-01 -1.94766000e-01
-5.16084909e-01 -3.21216673e-01 1.10473013e+00 -1.18913686e+00
-1.57725060e+00 8.50509346e-01 -2.83635885e-01 -5.32641053e-01
3.89662415e-01 -2.46099159e-01 -4.36798573e-01 -2.17924133e-01
1.30059302e-01 9.22681928e-01 5.60253620e-01 -7.73370564e-01
-2.99756408e-01 -5.59429079e-02 -3.10792714e-01 4.77231085e-01
-7.46519268e-01 4.96718019e-01 -1.82077631e-01 -9.20993388e-01
-5.21302894e-02 -1.03912508e+00 -1.04644276e-01 -4.78221714e-01
8.87652785e-02 -8.18373859e-02 5.83862543e-01 -1.21156621e+00
1.29910326e+00 -2.09154701e+00 3.85994524e-01 -3.52723688e-01
-3.98426414e-01 2.31637821e-01 -5.08436501e-01 8.93087864e-01
-6.33190125e-02 2.12233394e-01 -2.92021811e-01 -3.13930869e-01
2.34079331e-01 6.98087037e-01 -4.67849523e-02 2.94360310e-01
3.92448008e-01 1.32341039e+00 -7.57952690e-01 -4.83745784e-01
2.69007180e-02 5.37816167e-01 -3.23334962e-01 1.47729829e-01
1.08173028e-01 1.78699270e-01 1.29348278e-01 6.63576305e-01
3.73552591e-01 2.14886263e-01 3.43481213e-01 -1.96158676e-03
-1.14595011e-01 8.37558925e-01 -6.04303181e-01 1.80159807e+00
-6.05685532e-01 9.50809181e-01 -2.74647057e-01 -8.01522493e-01
9.51886117e-01 7.33680546e-01 2.53641993e-01 -6.00648940e-01
3.28256004e-02 6.68074727e-01 3.58051032e-01 -1.78414315e-01
6.58058941e-01 -3.95398051e-01 -1.57070488e-01 7.86894798e-01
5.11120141e-01 8.82420093e-02 7.18639791e-02 -1.79873034e-01
8.62306654e-01 2.20818877e-01 2.71588176e-01 -4.81993049e-01
2.64222234e-01 2.81213913e-02 4.91219282e-01 3.49720716e-01
-3.09976459e-01 6.67786300e-01 1.84616819e-01 -4.45355028e-01
-1.72974455e+00 -1.20191848e+00 -1.09120101e-01 1.24533832e+00
-3.06272715e-01 -4.74569350e-01 -9.15523648e-01 -6.55690014e-01
-2.33219475e-01 6.82361782e-01 -5.46977401e-01 -1.38774529e-01
-9.34842765e-01 -1.18839478e+00 1.00312710e+00 4.89634693e-01
3.43944639e-01 -1.00381875e+00 1.45057067e-01 5.47694862e-01
-5.26168764e-01 -1.17332911e+00 -7.21997619e-01 6.09725475e-01
-1.09584820e+00 -1.14524014e-01 -9.36712027e-01 -1.34602296e+00
5.16861379e-01 -2.09737167e-01 1.17859530e+00 -4.86607611e-01
2.74766058e-01 -1.39720410e-01 -3.38826269e-01 -2.62281716e-01
-8.53229225e-01 5.31342030e-01 5.15751064e-01 -3.00310165e-01
9.98310804e-01 -4.90975231e-01 -1.59843341e-01 1.69344068e-01
-7.58044958e-01 -1.44423753e-01 8.94627213e-01 1.05855989e+00
3.97908956e-01 -5.38049459e-01 5.01178563e-01 -5.67704737e-01
8.79598439e-01 -3.75280172e-01 -2.74280515e-02 3.76329839e-01
-8.70561719e-01 4.62974995e-01 6.33924246e-01 -6.28071785e-01
-7.06594825e-01 -1.51505083e-01 -1.82932943e-01 -2.04439573e-02
2.05754295e-01 4.40699041e-01 -1.59262270e-02 2.57711321e-01
8.39140654e-01 3.60233843e-01 -6.62941486e-02 -5.59234202e-01
4.83547837e-01 1.19840825e+00 3.18404794e-01 -6.36751771e-01
8.42577517e-01 -2.53118813e-01 -7.83467948e-01 -5.19528985e-01
-2.16246232e-01 -2.24182233e-01 -1.17420089e+00 3.42983842e-01
9.41902995e-01 -1.07747900e+00 3.39422435e-01 2.28061557e-01
-1.27757907e+00 -2.17579424e-01 -1.74554870e-01 8.41029167e-01
-4.02237058e-01 2.13530734e-01 -1.08763444e+00 -2.66210914e-01
-6.20583534e-01 -1.26698136e+00 9.38399434e-01 -3.48592788e-01
-5.27166665e-01 -1.28458762e+00 4.91849244e-01 2.87494868e-01
5.92095912e-01 -9.63055417e-02 1.20064533e+00 -7.81039834e-01
-5.83445355e-02 -8.13420713e-02 -7.03404993e-02 8.20607603e-01
4.94911820e-01 -2.12156594e-01 -8.21511090e-01 -3.92393768e-01
-2.53050148e-01 -3.02232593e-01 5.66913605e-01 1.06757276e-01
1.45670064e-02 -5.04906714e-01 1.03817627e-01 3.37777644e-01
1.50773859e+00 -8.67528170e-02 6.62619591e-01 4.79232907e-01
6.44786656e-01 2.97784865e-01 1.88656151e-01 -1.60348341e-01
3.20314199e-01 5.67939341e-01 -2.74388224e-01 -2.99327642e-01
-1.93137616e-01 -3.77364039e-01 1.23334515e+00 1.97512877e+00
2.80405283e-02 -8.23805183e-02 -8.56830001e-01 9.73967850e-01
-1.68465471e+00 -7.50712037e-01 7.80735910e-02 2.02025723e+00
1.37332869e+00 -1.55403102e-02 -4.65149693e-02 -1.93847731e-01
6.89143240e-01 8.47944841e-02 -1.43301710e-01 -1.38984442e+00
-3.59540761e-01 4.56240147e-01 7.71313965e-01 7.83151984e-01
-9.05107617e-01 1.32595205e+00 6.82467985e+00 7.41662800e-01
-1.19846225e+00 7.04535186e-01 3.76703560e-01 1.36350080e-01
-3.76722157e-01 8.74543935e-02 -7.63534844e-01 3.65617752e-01
1.59263170e+00 -1.83579519e-01 5.58227658e-01 5.70354819e-01
-3.31484042e-02 5.44888258e-01 -1.25504422e+00 7.65944719e-01
1.86105445e-01 -1.14729774e+00 7.95181170e-02 1.65842012e-01
1.11481905e+00 7.07636356e-01 -1.00911045e-02 4.61100698e-01
5.07282853e-01 -1.01100481e+00 5.42820871e-01 -2.15493292e-01
9.25307393e-01 -8.54309678e-01 1.10478044e+00 2.18289852e-01
-8.40894818e-01 4.00426656e-01 -7.21780121e-01 -4.93898951e-02
-1.01703681e-01 3.41494709e-01 -1.01879847e+00 4.87023354e-01
4.23326135e-01 5.69672406e-01 -5.01243353e-01 1.39265195e-01
-2.70935148e-01 8.95004869e-01 -2.68049687e-01 -1.36606410e-01
6.27704024e-01 -3.58079642e-01 3.39466780e-01 1.68630826e+00
3.86316359e-01 -5.37574530e-01 -2.44161844e-01 4.39356714e-01
-3.02408934e-01 4.92599964e-01 -8.19156528e-01 -4.71279621e-01
2.23536476e-01 7.78958738e-01 -2.73443937e-01 -4.28009748e-01
-5.93645275e-01 1.61285746e+00 4.90320325e-01 4.06228215e-01
-8.31849873e-01 -3.77710283e-01 9.54969406e-01 -1.34513691e-01
2.88633347e-01 -6.16858840e-01 -3.44607443e-01 -1.40463173e+00
2.13447258e-01 -1.18059826e+00 1.58195645e-02 -3.98066461e-01
-1.30786622e+00 8.74059439e-01 -3.47286373e-01 -1.20064616e+00
-5.25883913e-01 -7.00107515e-01 -2.72732377e-01 1.20102644e+00
-1.35615349e+00 -1.30932236e+00 7.70530403e-01 8.09067786e-02
7.00287461e-01 -4.47475374e-01 1.32161295e+00 5.80264270e-01
-3.24893266e-01 1.06227028e+00 8.98222029e-01 5.70774436e-01
1.07580149e+00 -1.29352045e+00 9.47071016e-01 1.03490520e+00
6.36863530e-01 8.38486612e-01 4.81570989e-01 -4.26689923e-01
-1.34089649e+00 -1.10375810e+00 1.86499536e+00 -7.33703077e-01
7.74810672e-01 -7.39070654e-01 -6.53414071e-01 8.89056802e-01
9.97722208e-01 -2.75165975e-01 8.39480877e-01 1.73965231e-01
-4.08299118e-01 9.97772217e-02 -8.66500258e-01 8.73497963e-01
6.70009315e-01 -9.39267218e-01 -8.40186715e-01 3.64317447e-01
8.42881441e-01 3.96234170e-02 -1.06047428e+00 1.88475579e-01
6.80850685e-01 -4.89691943e-01 7.32165575e-01 -9.14248765e-01
6.31018221e-01 -1.33300737e-01 -6.06028616e-01 -1.58500838e+00
-4.12512332e-01 -7.33361125e-01 1.04463719e-01 1.30083263e+00
1.00510550e+00 -5.55598319e-01 4.28529471e-01 1.25132576e-01
-2.56599355e-02 -5.48389077e-01 -9.97671187e-01 -9.45234299e-01
8.15777719e-01 -1.27153963e-01 4.12829190e-01 1.21150124e+00
1.25268579e-01 7.36084640e-01 -5.24131179e-01 1.02672335e-02
3.29315275e-01 -2.09976435e-01 5.21535516e-01 -6.22585297e-01
-3.58842134e-01 -2.21035048e-01 -4.16810125e-01 -8.27418864e-01
2.12823123e-01 -1.35790098e+00 2.94905971e-03 -1.57415533e+00
1.31080166e-01 -1.36932179e-01 -5.79868317e-01 6.38088048e-01
-2.31791228e-01 6.08538985e-01 2.09722221e-02 2.47115940e-01
-7.05085099e-02 4.55807537e-01 8.76177371e-01 -2.03651458e-01
4.04469017e-03 -6.35735273e-01 -4.74058002e-01 2.86136061e-01
9.54213798e-01 -7.44928181e-01 -1.32342786e-01 -1.24438536e+00
1.71930611e-01 -3.02832603e-01 -3.20131600e-01 -7.83765137e-01
-1.98235020e-01 2.87047345e-02 3.63658190e-01 -3.35030854e-01
5.53382561e-02 -7.76743948e-01 -1.41277546e-02 4.27123755e-01
-4.86568332e-01 7.76046455e-01 2.41922781e-01 2.17540357e-02
-4.24135685e-01 -2.47099265e-01 6.54131770e-01 -1.68657854e-01
-4.67067808e-01 -5.03155105e-02 -6.02452338e-01 -2.97621451e-02
4.83958632e-01 -2.29086339e-01 1.26837879e-01 -1.84378296e-01
-5.45960724e-01 -1.96461171e-01 5.51745832e-01 8.36962521e-01
2.46418208e-01 -1.87067056e+00 -1.19995391e+00 3.47953111e-01
2.57643849e-01 -7.14778841e-01 -5.65044463e-01 7.96311140e-01
-6.84432030e-01 5.26676118e-01 -3.94756913e-01 -4.08530235e-01
-1.07742214e+00 2.60547698e-01 1.59744129e-01 -4.12439644e-01
-1.88016921e-01 6.59323394e-01 -2.16674909e-01 -1.24735522e+00
1.05839990e-01 -2.92997837e-01 3.43153149e-01 -1.65046751e-02
1.43244371e-01 1.51659459e-01 2.82084137e-01 -9.68631744e-01
-3.02110136e-01 4.60450858e-01 -3.20618868e-01 -5.54424942e-01
1.37384737e+00 -2.59910315e-01 -3.33938450e-01 6.06567323e-01
1.63386106e+00 2.59161681e-01 -5.47049463e-01 -4.02435362e-01
1.17256612e-01 8.32308270e-03 -1.10360518e-01 -9.19706643e-01
-5.47589481e-01 1.16897523e+00 6.88229978e-01 -4.57330406e-01
7.86594391e-01 -2.70546407e-01 1.22250557e+00 6.09129965e-01
2.76813716e-01 -1.23698974e+00 -3.42476368e-01 8.13169777e-01
5.56032360e-01 -1.34188747e+00 -2.70542294e-01 3.64868760e-01
-7.41546750e-01 1.13147318e+00 2.94489741e-01 -8.79750699e-02
3.30945104e-01 3.29612017e-01 6.67823017e-01 4.17634100e-01
-7.31141567e-01 -2.43459027e-02 2.61451155e-01 4.84480828e-01
9.74726439e-01 2.88692415e-01 -6.26768291e-01 1.80466428e-01
-4.08431619e-01 -3.24572712e-01 2.97431529e-01 8.78522277e-01
-2.05002218e-01 -1.66689110e+00 -3.49230796e-01 9.01165903e-02
-7.68625975e-01 -7.53766477e-01 -6.18725717e-01 7.79040277e-01
9.93878469e-02 7.65464067e-01 3.45972963e-02 -4.50568289e-01
-5.29386364e-02 6.27011538e-01 5.00188649e-01 -5.43261945e-01
-8.85804832e-01 2.87937433e-01 2.02702180e-01 -2.69493796e-02
-3.15635145e-01 -5.64407885e-01 -1.00915658e+00 -3.78123939e-01
-3.38306397e-01 3.45579982e-01 9.14521813e-01 7.73897469e-01
3.52535367e-01 6.80977032e-02 5.91039121e-01 -4.07427102e-01
-6.79660857e-01 -1.47102726e+00 -8.89849737e-02 3.08117718e-01
3.29560965e-01 2.67028809e-03 -2.81353146e-01 4.53260899e-01] | [11.437277793884277, 10.24424934387207] |
7730adfd-868d-4e2c-82e9-9ebe7746a3ce | wekws-a-production-first-small-footprint-end | 2210.16743 | null | https://arxiv.org/abs/2210.16743v1 | https://arxiv.org/pdf/2210.16743v1.pdf | WeKws: A production first small-footprint end-to-end Keyword Spotting Toolkit | Keyword spotting (KWS) enables speech-based user interaction and gradually becomes an indispensable component of smart devices. Recently, end-to-end (E2E) methods have become the most popular approach for on-device KWS tasks. However, there is still a gap between the research and deployment of E2E KWS methods. In this paper, we introduce WeKws, a production-quality, easy-to-build, and convenient-to-be-applied E2E KWS toolkit. WeKws contains the implementations of several state-of-the-art backbone networks, making it achieve highly competitive results on three publicly available datasets. To make WeKws a pure E2E toolkit, we utilize a refined max-pooling loss to make the model learn the ending position of the keyword by itself, which significantly simplifies the training pipeline and makes WeKws very efficient to be applied in real-world scenarios. The toolkit is publicly available at https://github.com/wenet-e2e/wekws. | ['Fuping Pan', 'Lei Xie', 'Xiao-Lei Zhang', 'BinBin Zhang', 'Jingyong Hou', 'Menglong Xu', 'Jie Wang'] | 2022-10-30 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [-1.43419713e-01 -1.02252439e-01 -3.89626473e-01 -3.96683425e-01
-1.08974147e+00 -3.51173967e-01 1.71908885e-01 -5.29375732e-01
-4.34593111e-01 3.56016397e-01 3.49894524e-01 -5.99459827e-01
3.41030210e-01 -3.43689173e-01 -5.86975276e-01 -3.96596223e-01
3.81025136e-01 7.18923435e-02 2.41819128e-01 -1.02364466e-01
-3.98852259e-01 2.30345428e-02 -1.22298968e+00 4.18482929e-01
4.25208747e-01 1.17094624e+00 5.85111201e-01 8.36000443e-01
5.59050143e-02 4.90689009e-01 -4.01140034e-01 -4.49161112e-01
3.95511277e-02 -5.18415719e-02 -8.26305568e-01 -4.97103602e-01
8.18999037e-02 -5.56276262e-01 -6.89281702e-01 5.62700808e-01
1.15995634e+00 1.33709177e-01 1.36103883e-01 -1.38260376e+00
-8.56248558e-01 7.19760478e-01 -1.34003595e-01 6.29147608e-03
3.50265771e-01 3.30063730e-01 1.12881303e+00 -1.16434586e+00
2.34354421e-01 8.46741140e-01 7.03344822e-01 7.27909505e-01
-8.07626724e-01 -8.53941202e-01 6.36458173e-02 4.25462991e-01
-1.65071356e+00 -1.07291019e+00 7.55394638e-01 1.39394715e-01
1.48536181e+00 4.59212810e-01 6.52708709e-01 1.54565454e+00
-4.71653759e-01 1.51001143e+00 7.90917277e-01 -2.06348673e-01
1.25789285e-01 1.17081985e-01 -4.24941927e-02 5.60580134e-01
-4.14654493e-01 -5.11075974e-01 -9.28678632e-01 -1.89086720e-01
3.73537689e-01 -7.42665008e-02 -5.68518460e-01 2.00370178e-01
-1.16465616e+00 4.12869692e-01 1.95240691e-01 4.09790456e-01
-3.62444282e-01 3.52755964e-01 5.74728489e-01 1.70156941e-01
6.88889027e-01 1.67957008e-01 -1.00968122e+00 -1.04147446e+00
-8.69321048e-01 2.97898293e-01 7.26190627e-01 1.01956904e+00
1.57554969e-01 -1.89268842e-01 -3.29974622e-01 1.15035057e+00
1.86868906e-01 6.07852459e-01 4.99598384e-01 -5.58473289e-01
7.92570531e-01 1.24795482e-01 1.16608478e-01 -2.84683466e-01
-2.56133318e-01 -9.75292549e-02 -7.18253016e-01 -5.77176690e-01
1.52462944e-02 -2.41883263e-01 -6.75486922e-01 1.67440033e+00
2.12195143e-01 4.48471546e-01 -9.08647478e-02 7.75348067e-01
1.20019424e+00 9.52889979e-01 5.21293026e-04 -2.31972504e-02
1.28815889e+00 -1.38785577e+00 -9.63982642e-01 -1.30581796e-01
8.15361857e-01 -6.30386710e-01 1.80174983e+00 3.60191256e-01
-1.06739783e+00 -3.50569874e-01 -8.52774441e-01 -4.63409990e-01
-5.70991755e-01 3.97600502e-01 5.30311227e-01 5.62646627e-01
-1.22838497e+00 5.00160277e-01 -8.83964300e-01 -2.18511879e-01
3.90253425e-01 3.41510296e-01 -2.63241470e-01 2.26724371e-01
-1.59401274e+00 5.95717788e-01 3.93966407e-01 2.23851293e-01
-3.16030085e-01 -9.76337492e-01 -6.71360970e-01 1.43357560e-01
6.03398919e-01 -5.64085305e-01 1.93232310e+00 -4.75751847e-01
-2.14650106e+00 6.19730592e-01 -3.07744712e-01 -3.24351490e-02
1.59123898e-01 -3.67208153e-01 -7.23411322e-01 -1.74354270e-01
-2.72163659e-01 5.31462193e-01 6.48292243e-01 -6.29943550e-01
-5.00994205e-01 -4.83820140e-02 -6.42909110e-02 2.86911368e-01
-8.38652611e-01 2.93278337e-01 -1.14733589e+00 -7.12947786e-01
-5.78689754e-01 -7.94585764e-01 2.75871038e-01 3.91301885e-02
-6.67695343e-01 -8.17201614e-01 8.63106787e-01 -1.11422968e+00
1.76565063e+00 -2.42895675e+00 -2.16775939e-01 -1.66584089e-01
4.00232226e-01 7.37769067e-01 -1.09912731e-01 7.19242990e-01
1.77581906e-01 1.53360903e-01 1.98812872e-01 -9.83382463e-01
4.14765149e-01 -2.59683300e-02 -7.60928541e-02 2.15366587e-01
-1.74353972e-01 1.37590182e+00 -8.71488571e-01 -7.46764615e-02
3.34256470e-01 6.62166476e-01 -1.91147432e-01 5.69643676e-01
-9.71348956e-02 1.10213667e-01 -4.41628069e-01 4.88992453e-01
3.48813087e-01 -3.79470676e-01 -1.33211434e-01 -2.16373205e-01
-3.95001546e-02 9.42803919e-01 -7.87005723e-01 2.07991171e+00
-9.45081592e-01 7.07696140e-01 2.06758492e-02 -5.94374418e-01
4.91350621e-01 6.13929093e-01 2.33547136e-01 -6.68524027e-01
1.17482506e-01 4.37069237e-01 -4.28521693e-01 -4.59037751e-01
6.74655557e-01 3.92433763e-01 -1.34189889e-01 2.22937599e-01
1.59909740e-01 -1.22200087e-01 -2.52180338e-01 9.82415751e-02
1.06801844e+00 -8.64676535e-02 3.33062530e-01 8.43763947e-02
7.63772503e-02 -8.83184075e-01 4.09340382e-01 5.90507686e-01
-1.68186650e-01 5.88852406e-01 -4.32079621e-02 -2.11197034e-01
-9.50539291e-01 -9.41449821e-01 2.15849087e-01 1.12972009e+00
-9.36438069e-02 -9.21748757e-01 -9.06220496e-01 -5.75236142e-01
-2.26058125e-01 8.50942254e-01 -1.80012703e-01 -2.76296407e-01
-2.03932971e-01 -3.21291208e-01 8.67589056e-01 6.54657960e-01
7.41896689e-01 -1.10246754e+00 -4.05732483e-01 2.31481254e-01
-5.72985351e-01 -1.44526005e+00 -1.14705431e+00 9.36016142e-02
-2.87088513e-01 -4.67569232e-01 -9.07818973e-01 -6.01978421e-01
1.08776219e-01 5.14040828e-01 9.76279616e-01 -9.07064527e-02
-3.07900272e-02 3.19648772e-01 -7.02222407e-01 -4.57747459e-01
4.16924693e-02 5.52224398e-01 2.41957158e-01 -1.33772686e-01
6.21800601e-01 -6.57360792e-01 -7.74454653e-01 2.48994231e-01
-4.26174134e-01 3.07958812e-01 3.68612856e-01 4.78913665e-01
2.99062043e-01 -6.78413585e-02 5.03312409e-01 -5.53509593e-01
9.33662772e-01 -3.72816205e-01 -2.51098692e-01 3.12905639e-01
-6.02616549e-01 -2.02932060e-01 6.53980434e-01 -6.69015169e-01
-7.19516277e-01 -2.96737105e-01 -1.01656902e+00 -5.19976497e-01
7.32170567e-02 4.38989103e-01 -5.37590563e-01 1.46293014e-01
6.08470440e-02 4.73272085e-01 -5.44384599e-01 -8.83258522e-01
5.99557042e-01 1.63520157e+00 4.56877619e-01 -2.86021262e-01
5.87699831e-01 -5.19288816e-02 -9.05307055e-01 -1.10664356e+00
-7.46051192e-01 -8.01114738e-01 -8.35374892e-02 2.21151728e-02
7.20413923e-01 -1.02417529e+00 -8.71836901e-01 9.39543486e-01
-1.40598404e+00 -8.90529990e-01 -2.21458480e-01 1.87271595e-01
-3.24403733e-01 2.73729742e-01 -6.33832753e-01 -8.89256239e-01
-1.02666044e+00 -1.06745195e+00 1.37784326e+00 2.52822846e-01
-2.74908066e-01 -8.22455645e-01 -2.27159292e-01 4.63544816e-01
6.87585592e-01 -4.65240836e-01 3.58806491e-01 -6.18641436e-01
-1.17792249e-01 -1.67356297e-01 -1.31418079e-01 5.97931623e-01
2.00067729e-01 -2.44570836e-01 -1.28411710e+00 -1.66118935e-01
-1.83770865e-01 -5.18065929e-01 5.54891586e-01 1.56094298e-01
1.90079665e+00 -5.16014695e-01 -2.77993262e-01 7.41683960e-01
8.18090737e-01 1.13952838e-01 6.28758073e-01 7.55147496e-03
7.69966781e-01 -3.10670286e-02 2.38087252e-01 4.30573970e-01
9.06830370e-01 1.22302711e+00 -1.61402434e-01 -3.41547430e-01
-1.77859604e-01 -5.77741802e-01 4.38683659e-01 1.54749727e+00
1.86241984e-01 -5.80196559e-01 -5.97174764e-01 6.17128968e-01
-1.93140590e+00 -6.57994688e-01 1.94428071e-01 2.03572321e+00
1.19283211e+00 1.50340199e-01 1.79190248e-01 -4.52737175e-02
4.35024619e-01 3.78954709e-01 -7.11334765e-01 -3.27414453e-01
8.89572203e-02 3.07794809e-01 4.17734444e-01 3.08708221e-01
-1.13337779e+00 1.16416776e+00 5.42377377e+00 1.55982602e+00
-1.04012811e+00 6.06999516e-01 7.97763586e-01 -2.14895710e-01
-3.19202006e-01 -5.19807160e-01 -9.25474048e-01 8.92878115e-01
1.38073170e+00 1.92267951e-02 7.64640927e-01 1.03575099e+00
5.32121003e-01 1.57263920e-01 -9.66093481e-01 1.57639909e+00
-1.69661015e-01 -1.29719818e+00 -5.47879398e-01 -1.31299585e-01
2.61789769e-01 3.86503607e-01 8.56379345e-02 4.95798379e-01
1.15291581e-01 -7.95405626e-01 7.11668253e-01 2.35160142e-02
1.47065115e+00 -6.00928128e-01 5.94828188e-01 4.39426064e-01
-1.55249774e+00 1.49713919e-01 -3.51950340e-02 1.67786106e-02
4.98935431e-01 7.80176759e-01 -6.82844341e-01 3.56213182e-01
9.18636620e-01 4.12814945e-01 -7.58755356e-02 6.85422063e-01
-4.81128871e-01 9.81507599e-01 -5.70140898e-01 -3.77979666e-01
2.53986955e-01 1.78411901e-01 1.57625288e-01 1.36001599e+00
3.49378347e-01 8.83095786e-02 1.24645114e-01 5.14799535e-01
-6.48819864e-01 1.64414257e-01 -4.94496375e-01 -3.26746315e-01
8.49319637e-01 1.40392232e+00 -4.65098947e-01 -3.70366424e-01
-6.72384143e-01 1.73643255e+00 3.76469046e-01 3.19584489e-01
-9.38586056e-01 -6.94135725e-01 1.09445465e+00 -1.49414524e-01
4.50570852e-01 -1.88288450e-01 3.12633246e-01 -1.28793311e+00
5.47808647e-01 -7.98419237e-01 -1.94281191e-01 -9.08031404e-01
-1.22374082e+00 5.85231125e-01 -8.22292082e-03 -8.57705116e-01
-6.35927320e-02 -6.18565500e-01 -6.66847050e-01 9.89247262e-01
-1.49803472e+00 -1.31933594e+00 -1.66191280e-01 5.37495792e-01
8.96689594e-01 8.73095468e-02 9.57192063e-01 7.18098164e-01
-6.63568139e-01 1.13802600e+00 1.67329907e-01 1.23540558e-01
6.22670054e-01 -1.09080946e+00 1.28978014e+00 6.57130182e-01
2.08093569e-01 6.68491840e-01 4.08969074e-01 -4.25154537e-01
-1.51080263e+00 -9.97975647e-01 1.18714917e+00 -3.85418236e-01
8.76668513e-01 -1.17997587e+00 -5.77174664e-01 4.05061901e-01
3.09296966e-01 1.67828515e-01 8.55988681e-01 1.95978999e-01
-4.12026852e-01 -1.26026943e-01 -7.24247575e-01 8.84033501e-01
1.44341671e+00 -1.12804508e+00 -1.36198383e-02 5.25499105e-01
1.35414279e+00 -4.92363185e-01 -6.83651030e-01 2.01010913e-01
7.25271881e-01 -5.96288323e-01 7.64839292e-01 -2.38383800e-01
-1.50517508e-01 -1.81583909e-03 -6.36470467e-02 -1.50233519e+00
1.16728703e-02 -1.30945730e+00 -6.66536748e-01 1.39172542e+00
6.95109725e-01 -7.01728046e-01 7.57557929e-01 7.88533151e-01
-3.13879222e-01 -1.23044705e+00 -1.12249124e+00 -7.46357083e-01
-3.08018893e-01 -8.78310800e-01 7.25831747e-01 6.45693541e-01
2.08902746e-01 3.10301006e-01 -8.72865915e-01 -1.84082717e-01
1.22728841e-02 -2.77783513e-01 7.06663132e-01 -6.33145273e-01
-6.29974902e-01 -3.40947628e-01 -7.81484544e-02 -1.82266998e+00
2.47204736e-01 -6.49413347e-01 2.42032766e-01 -1.43294728e+00
1.32023677e-01 -1.97567970e-01 -1.52918130e-01 9.22756851e-01
-3.60585600e-01 3.05210382e-01 1.26701608e-01 -1.37919001e-03
-7.72879064e-01 9.11716640e-01 7.61357307e-01 -5.44575900e-02
-3.24373364e-01 2.51746863e-01 -7.28242338e-01 4.20554936e-01
1.15443742e+00 -1.28882274e-01 -6.25880480e-01 -4.63550806e-01
1.30326778e-01 -4.18987036e-01 -1.05494887e-01 -6.63971782e-01
3.02837729e-01 2.32050434e-01 -1.70139343e-01 -3.65484118e-01
6.81243896e-01 -6.82006896e-01 -2.52447575e-02 -5.14293201e-02
-1.45635411e-01 4.81383130e-02 5.41315451e-02 1.29364774e-01
1.16723940e-01 5.56775648e-03 3.34371567e-01 1.33006588e-01
-8.24073136e-01 6.02120101e-01 -2.97238827e-01 2.96606310e-02
6.03178978e-01 5.98151088e-02 -7.46660829e-02 -6.98776960e-01
-1.79097682e-01 2.58294374e-01 6.16037808e-02 6.03229463e-01
6.91296041e-01 -1.31784594e+00 -2.77691066e-01 5.77319367e-03
3.68211389e-01 -4.96745966e-02 3.28412443e-01 7.61253953e-01
-4.36301194e-02 5.20767152e-01 5.77332914e-01 -1.00138597e-01
-1.41977298e+00 3.45550388e-01 8.89185965e-02 -2.50734031e-01
-8.44497859e-01 1.25597155e+00 -1.81761935e-01 -7.56094813e-01
6.38881624e-01 -5.70256472e-01 2.42840484e-01 -4.27694321e-01
9.01521564e-01 3.22975129e-01 2.95003772e-01 -3.35093886e-01
-4.18096632e-01 9.67216343e-02 -8.85410756e-02 -1.17522329e-01
1.17053604e+00 -1.80201828e-01 2.39833683e-01 7.18934238e-01
1.42341244e+00 -9.12582874e-02 -1.09291029e+00 -4.49725330e-01
-9.23063383e-02 -2.21425369e-01 3.46109450e-01 -8.39961827e-01
-1.03456366e+00 7.94413686e-01 3.60900193e-01 1.29475817e-02
1.24305212e+00 1.62285894e-01 1.54669106e+00 5.96391499e-01
3.45074683e-01 -1.37709153e+00 -2.07729533e-01 2.93956161e-01
7.17148304e-01 -1.06217265e+00 -4.72518086e-01 -4.85802174e-01
-6.60660684e-01 7.70438075e-01 3.76326829e-01 5.56655049e-01
8.06711853e-01 5.53321302e-01 3.70230824e-02 5.80787100e-02
-6.82350814e-01 -1.24201052e-01 4.02276307e-01 4.56459731e-01
3.81285787e-01 2.19006851e-01 -1.45709902e-01 9.92647946e-01
-1.93099633e-01 3.27367514e-01 -3.75804789e-02 7.19109833e-01
1.37312755e-01 -1.25423610e+00 4.23609525e-01 5.09269834e-01
-4.29160237e-01 -6.90325677e-01 -3.08633566e-01 3.87266278e-01
-2.14786783e-01 1.05948067e+00 -5.26237428e-01 -8.93290460e-01
5.25018871e-01 2.31649056e-01 2.09663108e-01 -5.63279688e-01
-5.57532132e-01 -8.23424831e-02 3.56240213e-01 -8.90383184e-01
-7.54796565e-02 -1.95702344e-01 -1.00114596e+00 -4.84618723e-01
-4.85852122e-01 1.89800173e-01 9.95752513e-01 1.01235163e+00
8.63246977e-01 4.97238576e-01 6.25645220e-01 -1.01557600e+00
-5.80627322e-01 -1.17303765e+00 -6.09475255e-01 1.65728211e-01
1.98061809e-01 -4.28682834e-01 -1.59646705e-01 -2.38456026e-01] | [14.228819847106934, 6.408493995666504] |
80ca41ac-b28a-4a55-8176-86c5230ebe11 | machine-learning-based-prototyping-of | 1802.02312 | null | https://arxiv.org/abs/1802.02312v2 | https://arxiv.org/pdf/1802.02312v2.pdf | Machine Learning-Based Prototyping of Graphical User Interfaces for Mobile Apps | It is common practice for developers of user-facing software to transform a mock-up of a graphical user interface (GUI) into code. This process takes place both at an application's inception and in an evolutionary context as GUI changes keep pace with evolving features. Unfortunately, this practice is challenging and time-consuming. In this paper, we present an approach that automates this process by enabling accurate prototyping of GUIs via three tasks: detection, classification, and assembly. First, logical components of a GUI are detected from a mock-up artifact using either computer vision techniques or mock-up metadata. Then, software repository mining, automated dynamic analysis, and deep convolutional neural networks are utilized to accurately classify GUI-components into domain-specific types (e.g., toggle-button). Finally, a data-driven, K-nearest-neighbors algorithm generates a suitable hierarchical GUI structure from which a prototype application can be automatically assembled. We implemented this approach for Android in a system called ReDraw. Our evaluation illustrates that ReDraw achieves an average GUI-component classification accuracy of 91% and assembles prototype applications that closely mirror target mock-ups in terms of visual affinity while exhibiting reasonable code structure. Interviews with industrial practitioners illustrate ReDraw's potential to improve real development workflows. | ['Denys Poshyvanyk', 'Richard Bonett', 'Michael Curcio', 'Carlos Bernal-Cárdenas', 'Kevin Moran'] | 2018-02-07 | null | null | null | null | ['component-classification'] | ['natural-language-processing'] | [ 2.14807764e-02 -1.82469815e-01 -8.98561329e-02 -2.29450718e-01
-6.44007385e-01 -9.39003229e-01 2.17951745e-01 3.00255358e-01
4.46089596e-01 -1.27126100e-02 -1.87765226e-01 -6.61184132e-01
-3.08685184e-01 -6.62378490e-01 -5.20816684e-01 1.61326349e-01
-3.88276391e-02 2.09379762e-01 2.25846678e-01 -2.99095139e-02
5.52466750e-01 5.19223690e-01 -1.99457717e+00 6.96484149e-01
9.16432858e-01 8.73843551e-01 3.39625835e-01 8.18954408e-01
-3.58981401e-01 4.63443905e-01 -7.58986235e-01 -3.16968501e-01
2.16254950e-01 -2.48839736e-01 -6.39850020e-01 2.00925678e-01
2.18598515e-01 -1.25364885e-01 6.89061701e-01 7.54713416e-01
2.22703561e-01 -4.43774581e-01 3.98369223e-01 -1.42137194e+00
-5.58836699e-01 5.99065304e-01 -5.92110336e-01 -5.30119658e-01
8.14930737e-01 -2.35853661e-02 8.17266166e-01 -1.08474779e+00
5.30692518e-01 6.54864550e-01 1.13346946e+00 3.10576886e-01
-1.27983940e+00 -5.39737642e-01 -1.03438407e-01 -2.64565945e-02
-1.55594277e+00 -2.68052012e-01 9.36232448e-01 -9.29456353e-01
9.66011584e-01 5.49516618e-01 1.01360798e+00 6.83212221e-01
1.52000532e-01 3.77858311e-01 6.21498525e-01 -7.47820318e-01
6.21620893e-01 4.81336862e-01 1.69180349e-01 5.62515318e-01
3.40144873e-01 -5.11712611e-01 -3.11869115e-01 -5.99601269e-01
4.39283103e-01 1.42899007e-01 -7.12480471e-02 -6.01143181e-01
-7.18986750e-01 3.66775393e-01 -9.79301799e-03 5.21268845e-01
-5.62874913e-01 -7.15504214e-02 2.58906335e-01 3.40583444e-01
2.61250615e-01 8.12345266e-01 -5.35110772e-01 -6.61810398e-01
-1.34218502e+00 -1.79981008e-01 1.14064562e+00 1.21075451e+00
8.67740691e-01 -1.48037761e-01 2.26659313e-01 8.52079690e-01
2.82668769e-01 -2.61047006e-01 6.01126313e-01 -7.61373281e-01
-3.29871476e-02 1.47309124e+00 1.09358363e-01 -9.19707000e-01
-1.55484498e-01 -2.02288747e-01 -2.81194806e-01 7.98595071e-01
7.68906297e-03 2.96216514e-02 -4.53895688e-01 9.98686612e-01
3.42772335e-01 -1.77725196e-01 -2.15343788e-01 1.60120606e-01
4.27616686e-01 2.32003272e-01 -2.59035140e-01 6.35731891e-02
1.36009324e+00 -7.57021964e-01 -3.47512275e-01 1.32700019e-02
4.88928854e-01 -9.02732849e-01 1.52936792e+00 8.11463594e-01
-9.49899733e-01 -8.04218531e-01 -1.37261891e+00 2.77720898e-01
-3.99498880e-01 4.95351225e-01 5.72538674e-01 1.10725737e+00
-1.08530974e+00 6.54697776e-01 -8.28936815e-01 -6.33938551e-01
4.84730899e-01 1.60145134e-01 -1.68862239e-01 2.61497766e-01
-1.50230840e-01 5.40227592e-01 -1.20365648e-02 -2.99326897e-01
-6.82449043e-01 -9.95590925e-01 -5.13148546e-01 1.56314537e-01
3.32200348e-01 -4.73026931e-01 1.45064116e+00 -1.23957384e+00
-1.38757062e+00 4.00022149e-01 1.92078859e-01 -9.61786776e-04
3.52722168e-01 -2.16600180e-01 -5.04871368e-01 -3.04966539e-01
-7.73708373e-02 -7.00858906e-02 1.03962958e+00 -1.50015724e+00
-9.12229061e-01 -3.25264335e-01 1.35665508e-02 -2.40874395e-01
-6.04771614e-01 1.68145858e-02 -6.84111059e-01 -1.31669879e-01
-2.31077105e-01 -7.47457504e-01 1.32132962e-01 -5.86218499e-02
-2.26607531e-01 -1.72759756e-01 8.77524674e-01 -7.71673501e-01
1.92471528e+00 -2.20183659e+00 -2.89305687e-01 4.89851832e-01
3.59680921e-01 1.61649659e-01 1.04976622e-02 7.57601321e-01
-3.91208053e-01 2.52794236e-01 1.17634811e-01 2.84118466e-02
-2.48704143e-02 -4.53817427e-01 -1.00804158e-02 -1.89342536e-02
2.33548135e-02 5.19223630e-01 -7.64000952e-01 -1.42445758e-01
1.35689721e-01 4.37143028e-01 -5.48367083e-01 3.01166624e-01
-3.07039171e-01 -8.02051425e-02 -1.46328866e-01 9.13025618e-01
3.89988542e-01 -3.76569033e-01 5.45265377e-01 -2.46741056e-01
-4.59546506e-01 1.99212059e-01 -1.23027432e+00 1.64624774e+00
-8.68516505e-01 6.92263901e-01 -9.16577205e-02 -6.14365339e-01
1.12666786e+00 1.50588825e-01 2.61807233e-01 -2.58049369e-01
-1.82076544e-01 1.54184163e-01 -3.70505564e-02 -6.45551324e-01
4.27774668e-01 7.45419979e-01 2.81487219e-02 7.03686774e-01
-3.11223835e-01 -3.88935208e-02 4.23949063e-02 5.05805910e-02
1.29280686e+00 6.30468309e-01 7.57645309e-01 -7.23541006e-02
2.54382461e-01 2.75283813e-01 1.47014931e-01 5.63356102e-01
5.91978192e-01 6.95245445e-01 6.12161219e-01 -4.20550108e-01
-1.05040908e+00 -7.17346847e-01 2.37171710e-01 1.22078967e+00
-4.42071974e-01 -1.01178026e+00 -1.13379979e+00 -6.94557071e-01
-1.46921262e-01 9.05289114e-01 -6.47576272e-01 -2.25082889e-01
-2.50805199e-01 -1.30465955e-01 2.89679796e-01 5.22135496e-01
2.39932742e-02 -9.51698840e-01 -1.34047508e+00 2.35470265e-01
3.83503079e-01 -3.59858394e-01 -3.97387564e-01 -1.38315847e-02
-5.50101340e-01 -1.42763579e+00 -3.13287437e-01 -6.32464468e-01
8.00858080e-01 2.54685074e-01 1.41944790e+00 4.48590130e-01
-8.02705705e-01 6.41549528e-01 -6.07457995e-01 -3.46690685e-01
-7.72105873e-01 9.50896647e-03 -3.10723454e-01 -8.78196061e-02
4.27761197e-01 -8.16248953e-01 -4.90308106e-01 2.99890935e-01
-7.31829584e-01 4.22352403e-02 7.71884143e-01 4.22902852e-01
4.51052040e-01 2.35820949e-01 2.64824212e-01 -8.07853699e-01
9.64697599e-01 -5.52156806e-01 -9.12080646e-01 6.83260798e-01
-8.50881338e-01 -3.14965665e-01 7.60621846e-01 -7.62835562e-01
-7.55210996e-01 3.40372384e-01 -4.21186946e-02 -3.51409853e-01
-3.59542906e-01 8.00841630e-01 -3.03978890e-01 7.77830631e-02
1.01102829e+00 1.19306527e-01 -7.05731362e-02 -5.98370075e-01
3.48933220e-01 1.24054158e+00 2.55351931e-01 -4.65828240e-01
8.04656088e-01 9.73060131e-02 -6.58351898e-01 -8.07800174e-01
-1.08704194e-02 -3.26097012e-01 -5.43387830e-01 -2.80488640e-01
4.39787954e-01 -5.72426856e-01 -6.89639568e-01 2.49392688e-01
-9.91795778e-01 -3.44645679e-01 -2.68514365e-01 9.46134049e-03
-1.85819089e-01 1.63312450e-01 2.34601244e-01 -9.11143601e-01
-4.19504315e-01 -1.15362203e+00 9.79662955e-01 1.25290439e-01
-9.03757095e-01 -5.72836041e-01 2.47316118e-02 1.41869664e-01
4.75901484e-01 3.47369164e-01 1.20183039e+00 -7.04652786e-01
-3.26546758e-01 -5.48380673e-01 4.42167521e-02 3.34723085e-01
6.13120735e-01 7.21697152e-01 -8.24565589e-01 -1.43012375e-01
-1.34906471e-01 4.15485978e-01 -2.49280736e-01 -6.14382252e-02
1.29168546e+00 -2.77060539e-01 -4.81863022e-01 5.75774550e-01
1.35413849e+00 8.15723479e-01 4.70172018e-01 4.06549960e-01
5.86429060e-01 5.03915250e-01 6.29190326e-01 6.69385850e-01
3.76358777e-01 6.74025536e-01 4.59196806e-01 3.92534845e-02
-9.49166268e-02 -2.04312101e-01 3.62320632e-01 6.31170332e-01
7.32034892e-02 2.39523277e-01 -1.28585255e+00 4.89942521e-01
-1.64031732e+00 -6.24131739e-01 -2.44024307e-01 2.49341917e+00
6.80931687e-01 1.45600915e-01 3.70009303e-01 1.90598592e-01
6.24459863e-01 -7.19851732e-01 -4.90606040e-01 -5.18883169e-01
7.24615037e-01 1.07274644e-01 -9.82604325e-02 -5.38724437e-02
-6.41703427e-01 3.21353912e-01 5.73358345e+00 5.32504916e-01
-1.22526169e+00 4.58560139e-02 4.05552000e-01 6.39081597e-02
-1.36400759e-01 1.64382488e-01 -5.04154027e-01 5.73561609e-01
7.94339657e-01 -3.16521764e-01 7.54137933e-01 1.58391726e+00
1.46284029e-01 -1.87216431e-01 -1.50400209e+00 1.08244443e+00
2.04752430e-01 -1.63603938e+00 -1.84833378e-01 2.74215192e-02
4.62628692e-01 -3.13412815e-01 -1.79167658e-01 1.37316763e-01
2.66047984e-01 -8.72800767e-01 8.46062124e-01 4.55117941e-01
1.17608976e+00 -6.32620633e-01 3.33302587e-01 2.49921888e-01
-1.28974664e+00 -2.21939191e-01 1.48188099e-01 -6.47769347e-02
-4.21012849e-01 4.12456542e-01 -1.20716536e+00 4.15744245e-01
1.12941289e+00 3.43691647e-01 -9.82538819e-01 1.13853502e+00
5.88281639e-02 4.05022591e-01 -6.40942752e-02 -2.39762664e-01
-4.43341494e-01 -8.17219764e-02 4.03911807e-02 1.25305986e+00
6.85492218e-01 -4.66968536e-01 -4.74969335e-02 1.06420004e+00
1.44807503e-01 1.29068434e-01 -5.50887406e-01 -1.42801538e-01
7.98076928e-01 1.66191864e+00 -9.94365036e-01 -2.58652031e-01
-4.25273240e-01 8.54214251e-01 1.34006944e-02 1.65688157e-01
-7.37560034e-01 -8.10080349e-01 6.70629203e-01 5.08794725e-01
4.18309718e-01 -3.63968983e-02 -4.28958774e-01 -7.07633197e-01
2.75061518e-01 -1.18225014e+00 -1.01134628e-01 -9.11159337e-01
-9.84523416e-01 8.32512856e-01 -2.02172667e-01 -1.57015395e+00
-2.64829993e-01 -2.51974016e-01 -9.83734250e-01 6.45458102e-01
-5.37909269e-01 -1.17326343e+00 -7.74304867e-01 2.40608558e-01
5.88586450e-01 -2.21280903e-01 8.96906972e-01 2.28421167e-01
-3.73772711e-01 5.20480156e-01 1.02135144e-01 -1.04282863e-01
3.56957942e-01 -1.40398002e+00 6.90449774e-01 8.09205115e-01
2.44678050e-01 1.19377840e+00 4.29444730e-01 -7.89293647e-01
-1.69131947e+00 -1.03082824e+00 2.76382118e-01 -3.80916327e-01
6.62687838e-01 -7.54041255e-01 -7.98468828e-01 4.67669457e-01
2.34952122e-01 -2.71173030e-01 9.68587339e-01 2.72729993e-01
-5.45027852e-01 -3.16053838e-01 -1.03266430e+00 6.63210392e-01
7.79529333e-01 -5.36863744e-01 -4.10708338e-01 9.31401476e-02
2.60819346e-01 -1.56028345e-01 -1.00396490e+00 -2.39178881e-01
9.88365412e-01 -1.22914207e+00 7.16833711e-01 -2.12109461e-01
1.50929108e-01 -6.59938872e-01 1.23009123e-01 -1.17798257e+00
-1.38069615e-01 -1.25862658e+00 -7.96400383e-02 1.48988926e+00
4.62484956e-01 -1.97394609e-01 4.93155777e-01 6.32249415e-01
-1.31094098e-01 -7.22475648e-01 -4.02361751e-01 -5.81083894e-01
-6.20306313e-01 -7.45206594e-01 1.02260923e+00 1.06114805e+00
2.47561827e-01 2.78877527e-01 1.47544593e-01 9.89194289e-02
1.42016038e-01 2.83036113e-01 1.23030901e+00 -1.55764723e+00
-4.43421572e-01 -5.85664630e-01 -3.60948801e-01 -5.67403376e-01
-3.45959306e-01 -5.55046558e-01 -1.28137007e-01 -1.57890308e+00
5.25562167e-02 -4.63479698e-01 7.25647248e-03 6.36369705e-01
2.38958851e-01 1.47972265e-02 -1.22587040e-01 7.34619498e-02
-5.21903276e-01 -3.81119937e-01 1.90066308e-01 -1.95279960e-02
-7.61610806e-01 2.47816324e-01 -1.04344881e+00 7.54014850e-01
7.96347678e-01 -3.39432389e-01 -5.28845549e-01 -2.34736338e-01
6.12729728e-01 -1.61952734e-01 3.69439833e-02 -1.28171420e+00
1.83933154e-01 5.48002310e-02 3.61294419e-01 -5.98756552e-01
-1.07705474e-01 -1.11817718e+00 8.23594093e-01 4.06715661e-01
-1.61514029e-01 3.97855401e-01 1.04202472e-01 7.27346763e-02
2.00991750e-01 -5.76095521e-01 4.80649590e-01 -8.61576013e-03
-5.91553986e-01 -2.35751867e-01 -5.40544808e-01 -5.42777002e-01
1.29456592e+00 -7.85442531e-01 -1.73603490e-01 1.69505440e-02
-4.47969437e-01 -3.04220170e-01 9.39092577e-01 7.34927237e-01
6.70636892e-01 -9.53855157e-01 -1.25710264e-01 5.11421025e-01
4.20300156e-01 -2.50148624e-01 -1.44460663e-01 6.16811216e-01
-7.78269708e-01 -2.28799120e-01 -1.33503482e-01 -5.88842809e-01
-1.73595965e+00 7.10340321e-01 1.08382255e-01 2.75100797e-01
-6.27971530e-01 4.28793401e-01 -2.24875599e-01 -1.84742361e-01
2.80874252e-01 -5.80832899e-01 -1.80761293e-01 1.52705938e-01
8.29806089e-01 4.40332383e-01 5.06905615e-01 -1.07778266e-01
-4.64504868e-01 4.51944739e-01 1.11681648e-01 2.60044366e-01
1.50235105e+00 1.07824355e-01 -2.85828263e-01 5.90797722e-01
9.61655855e-01 2.94633806e-01 -1.14208400e+00 4.44925100e-01
4.51522976e-01 -4.86304522e-01 -2.84791976e-01 -1.05649281e+00
-6.61782622e-01 6.73670650e-01 6.90979421e-01 8.73609543e-01
1.16845024e+00 1.42334759e-01 1.29777327e-01 3.85670930e-01
4.75286037e-01 -9.43062007e-01 4.81839664e-02 -9.71557721e-02
8.90341163e-01 -7.32799232e-01 1.12879813e-01 -1.54766232e-01
-3.50954980e-01 1.42982316e+00 6.27644897e-01 8.21222886e-02
5.31938314e-01 6.92591786e-01 -1.85484305e-01 -2.56560981e-01
-8.43534231e-01 2.87144542e-01 2.35818520e-01 8.97893727e-01
5.78173339e-01 -3.23944911e-02 -7.18465522e-02 8.18427503e-01
-1.47633836e-01 2.50167340e-01 7.20284700e-01 1.14864850e+00
-3.28798294e-01 -1.09776151e+00 -4.58249062e-01 6.64617121e-01
-3.61066103e-01 -1.44987404e-01 -5.97005069e-01 8.08488131e-01
3.22493225e-01 9.61763859e-01 -3.86019871e-02 -6.83811069e-01
3.79559755e-01 1.06241778e-01 3.00653487e-01 -8.67727637e-01
-8.99828732e-01 1.77360624e-01 -5.54880612e-02 -6.33147717e-01
1.09466739e-01 -5.67257941e-01 -8.17494988e-01 8.49054307e-02
-4.62771505e-01 3.44071746e-01 1.18693364e+00 5.22729933e-01
9.16396022e-01 7.72818387e-01 6.44063532e-01 -7.71354914e-01
-2.75083125e-01 -6.33965850e-01 -1.45761281e-01 1.88594237e-01
3.15678529e-02 -4.41392422e-01 5.59406653e-02 4.94757444e-01] | [8.223063468933105, 7.336215019226074] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.