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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
4178f08e-5438-4717-a4d3-dc1e4f419672 | a-metal-artifact-reduction-scheme-for | 2202.00116 | null | https://arxiv.org/abs/2202.00116v1 | https://arxiv.org/pdf/2202.00116v1.pdf | A Metal Artifact Reduction Scheme For Accurate Iterative Dual-Energy CT Algorithms | CT images have been used to generate radiation therapy treatment plans for more than two decades. Dual-energy CT (DECT) has shown high accuracy in estimating electronic density or proton stopping-power maps used in treatment planning. However, the presence of metal implants introduces severe streaking artifacts in the reconstructed images, affecting the diagnostic accuracy and treatment performance. In order to reduce the metal artifacts in DECT, we introduce a metal-artifact reduction scheme for iterative DECT algorithms. An estimate is substituted for the corrupt data in each iteration. We utilize normalized metal-artifact reduction (NMAR) composed with image-domain decomposition to initialize the algorithm and speed up the convergence. A fully 3D joint statistical DECT algorithm, dual-energy alternating minimization (DEAM), with the proposed scheme is tested on experimental and clinical helical data acquired on a Philips Brilliance Big Bore scanner. We compared DEAM with the proposed method to the original DEAM and vendor reconstructions with and without metal-artifact reduction for orthopedic implants (O-MAR). The visualization and quantitative analysis show that DEAM with the proposed method has the best performance in reducing streaking artifacts caused by metallic objects. | ["Joseph A. O'Sullivan", 'Bruce R. Whiting', 'David G. Politte', 'Jeffrey F. Williamson', 'Rui Liao', 'Maria Medrano', 'Tao Ge'] | 2022-01-31 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 2.97139406e-01 -1.58644065e-01 3.50240320e-01 -2.96016838e-02
-1.00260806e+00 5.19423038e-02 1.70149416e-01 -3.62062119e-02
-3.70905399e-01 6.96557581e-01 4.36577171e-01 -2.48544231e-01
-5.77878833e-01 -4.44685668e-01 -3.21661502e-01 -1.03122485e+00
2.43409768e-01 8.55788946e-01 4.65396285e-01 2.31373474e-01
2.56710231e-01 3.82593960e-01 -7.69300878e-01 5.84969997e-01
9.33632314e-01 8.34275365e-01 7.11646438e-01 4.50543255e-01
2.47847661e-01 8.78702700e-01 -5.97154796e-02 -9.45691839e-02
1.13624416e-01 -7.81214237e-01 -5.44170678e-01 1.93226606e-01
-4.21176106e-01 -2.70685166e-01 3.01137511e-02 8.95637810e-01
8.25894415e-01 -1.94044188e-01 1.04408491e+00 -6.78801715e-01
-1.44529089e-01 6.36058748e-01 -1.07874644e+00 3.14740211e-01
9.51018557e-02 8.77014101e-02 -1.78743601e-01 -9.07600582e-01
5.35592437e-01 6.96487606e-01 1.07189989e+00 5.57948768e-01
-1.07703543e+00 -4.63802308e-01 -8.41189802e-01 3.21581990e-01
-1.12045336e+00 2.03434285e-02 6.19545162e-01 -6.07451856e-01
7.69933701e-01 6.52982831e-01 6.31946266e-01 6.47299349e-01
1.01561260e+00 1.70500964e-01 1.37835896e+00 -5.59953034e-01
2.90531814e-01 2.74313074e-02 9.71820056e-02 6.60066843e-01
7.90810287e-01 1.14171449e-02 3.50977592e-02 -4.65442568e-01
9.13897276e-01 -2.03089625e-01 -5.94251335e-01 -2.10227951e-01
-1.03100026e+00 5.01410902e-01 1.97693482e-01 7.74500489e-01
-9.19717491e-01 1.34843066e-01 5.31447649e-01 -3.86769742e-01
3.38520825e-01 2.12824330e-01 -2.91509349e-02 4.41240221e-02
-7.67878532e-01 5.90779260e-02 2.98958391e-01 4.76335078e-01
-1.52207643e-01 2.07528789e-02 -4.41354245e-01 9.43735182e-01
5.73595524e-01 4.04424042e-01 1.03995085e+00 -4.19852287e-01
8.58934149e-02 2.83442348e-01 4.24295291e-02 -5.32988012e-01
-4.45047796e-01 -5.02695262e-01 -8.71198058e-01 3.73378873e-01
2.27551684e-01 -1.27564862e-01 -1.28879130e+00 1.09968412e+00
6.78131819e-01 1.25872642e-01 -4.11433764e-02 1.15410376e+00
6.72291458e-01 6.55067384e-01 2.57113159e-01 -8.88722658e-01
1.54325783e+00 -6.15721524e-01 -1.26860189e+00 -1.23574115e-01
7.77785182e-01 -1.22847974e+00 5.73535144e-01 5.65583169e-01
-1.53490603e+00 -1.65422007e-01 -1.08388174e+00 3.43292743e-01
7.53821850e-01 7.98311755e-02 2.93169379e-01 7.46073723e-01
-6.57884240e-01 6.69477463e-01 -1.25711894e+00 -8.98850034e-04
4.44099069e-01 2.99536794e-01 -1.94699034e-01 -1.49365813e-01
-7.98829257e-01 1.35727012e+00 -1.75983403e-02 1.78805232e-01
-6.36595011e-01 -1.03180575e+00 -5.25892019e-01 -4.75815147e-01
1.53663903e-01 -9.46947634e-01 1.43972456e+00 -5.72847188e-01
-1.62365901e+00 4.73479360e-01 -6.50169188e-03 -3.19233209e-01
9.02350426e-01 7.58849419e-05 -2.08622694e-01 7.40102679e-02
3.01258475e-01 -3.56098443e-01 5.00059009e-01 -1.33940518e+00
5.06288521e-02 -6.10806525e-01 -1.01346028e+00 4.86121207e-01
1.04801401e-01 2.69512255e-02 -2.17751473e-01 -6.26283765e-01
9.06713307e-01 -1.02356076e+00 -6.88409984e-01 -1.71880335e-01
-3.98612082e-01 2.65846282e-01 6.03189409e-01 -9.97461855e-01
1.24848890e+00 -1.81750238e+00 -9.98276100e-02 2.60053188e-01
3.11868340e-02 -2.45215312e-01 5.82286000e-01 6.30066618e-02
-4.50499535e-01 -3.65414023e-01 -8.00088882e-01 -1.36421904e-01
-6.90429628e-01 7.91389197e-02 5.06631553e-01 9.15578246e-01
-8.29287171e-01 5.41565061e-01 -5.82961679e-01 -6.27828419e-01
3.91839832e-01 6.53829873e-01 -4.15247589e-01 -4.56327312e-02
1.86425313e-01 7.78943300e-01 -5.26764929e-01 2.78263569e-01
1.30958223e+00 -2.36409888e-01 2.30048478e-01 -7.14956224e-01
-2.83757985e-01 -1.64504722e-01 -1.06451225e+00 1.58104765e+00
-6.24635398e-01 5.00753410e-02 2.18896165e-01 -4.77824181e-01
3.31200123e-01 6.20799541e-01 1.11956358e+00 -8.34999740e-01
5.68850219e-01 6.01307273e-01 3.96292865e-01 -8.81253660e-01
1.32282004e-01 -9.10338342e-01 4.87307042e-01 3.05074453e-01
-4.94682342e-01 -4.98115003e-01 -4.14731354e-01 2.19363011e-02
1.06561565e+00 -9.52138305e-02 2.33111575e-01 -6.38162017e-01
4.32015657e-01 3.24066490e-01 4.29609925e-01 2.63786256e-01
2.96505820e-02 7.83614576e-01 -1.25631601e-01 -1.87370881e-01
-1.23943675e+00 -9.72270012e-01 -7.80872881e-01 -2.10128948e-01
1.42496124e-01 8.11601728e-02 -9.61721122e-01 -1.88656598e-01
-3.01286161e-01 8.73909175e-01 -4.30096000e-01 -1.34925442e-02
-8.12474906e-01 -1.61900342e+00 -4.00248095e-02 4.73516941e-01
4.22246546e-01 -8.60997021e-01 -8.86374354e-01 5.48145890e-01
-3.13394904e-01 -9.45804298e-01 -4.05604929e-01 3.05610478e-01
-1.33664846e+00 -8.95695329e-01 -9.18331146e-01 -3.96778584e-01
7.23541498e-01 -8.42215642e-02 8.43736053e-01 -7.47732893e-02
-6.99205816e-01 2.82529056e-01 -2.67831385e-01 -2.35062927e-01
-9.16888773e-01 -8.24191093e-01 1.38181672e-02 -1.90032333e-01
-1.40423223e-01 -4.81823146e-01 -9.01208460e-01 4.50684845e-01
-8.92307818e-01 3.50407898e-01 6.33194029e-01 8.53463769e-01
9.82772052e-01 2.95464993e-01 1.87719211e-01 -1.08332324e+00
6.41357422e-01 -4.36395615e-01 -4.76187110e-01 -2.04631388e-02
-7.48642981e-01 1.51427492e-01 2.51315415e-01 -2.19287157e-01
-1.58334112e+00 -1.05747044e-01 -5.26748776e-01 -1.45745769e-01
1.91918120e-01 2.90353924e-01 1.41393587e-01 -4.15187448e-01
9.24983323e-01 3.28497231e-01 1.46939373e-02 -3.82742733e-01
-3.00410092e-01 5.85208952e-01 4.07055587e-01 -2.26324737e-01
3.51624787e-01 5.47744989e-01 3.24710339e-01 -6.61104500e-01
-3.24388444e-01 -3.10277104e-01 -2.24960238e-01 -4.33944523e-01
1.07815361e+00 -5.17128527e-01 -5.82413554e-01 7.06321836e-01
-1.04654396e+00 -1.00379899e-01 -2.56976306e-01 1.09032881e+00
-6.12044215e-01 7.70419836e-01 -8.15924764e-01 -7.56220996e-01
-6.28148317e-01 -1.62658250e+00 5.72015524e-01 1.09976456e-02
-2.10750371e-01 -8.68891656e-01 2.88070410e-01 3.57189327e-01
5.31223774e-01 5.58845162e-01 1.01073778e+00 5.32704853e-02
-1.83638096e-01 -5.37156910e-02 1.68188795e-01 1.47865430e-01
2.64446139e-01 -5.41337907e-01 -5.34129441e-01 -6.94005787e-02
1.20508099e+00 1.94292665e-01 3.13422352e-01 1.29266381e+00
1.27803254e+00 9.12439749e-02 -5.27843654e-01 6.29670084e-01
2.03066540e+00 7.85773456e-01 9.49965894e-01 3.22129309e-01
5.30249715e-01 3.58862132e-02 5.32264888e-01 7.18252659e-01
-1.51559502e-01 7.55639255e-01 3.29468280e-01 -6.59588054e-02
-2.82046795e-01 3.04284126e-01 -1.89920843e-01 9.32683468e-01
-4.72554833e-01 -1.16776496e-01 -9.63363707e-01 4.06486601e-01
-1.49531674e+00 -6.53710902e-01 -1.12542033e+00 2.10643864e+00
8.38167608e-01 -1.03555351e-01 -4.54857171e-01 3.30467016e-01
7.15814590e-01 -5.21838605e-01 -3.56069863e-01 -2.88541824e-01
2.39898294e-01 5.58867514e-01 8.64177108e-01 5.98812938e-01
-4.22172576e-01 -1.26677498e-01 6.14629650e+00 9.99603868e-01
-9.07914698e-01 8.39708626e-01 4.87159640e-01 1.52880773e-01
-5.12744725e-01 -1.90571234e-01 -2.49187991e-01 6.46443367e-01
9.18922901e-01 6.98067620e-03 -1.36295363e-01 6.09599948e-01
6.13937020e-01 -9.55916047e-01 -5.32646060e-01 1.06988108e+00
-1.60977498e-01 -1.01204610e+00 -2.00634822e-01 1.65109485e-01
7.25323558e-01 -3.97384644e-01 -4.22905013e-02 -3.20450604e-01
4.87658940e-02 -7.32694745e-01 5.88120818e-01 3.71104389e-01
6.31568193e-01 -6.11876428e-01 9.18605626e-01 1.55999824e-01
-8.52926970e-01 1.94128558e-01 -2.61281818e-01 4.56198573e-01
7.12273121e-01 1.21141422e+00 -1.02990055e+00 9.22749281e-01
4.57491726e-01 1.49788737e-01 -1.48260057e-01 1.30395842e+00
2.08331659e-01 3.92117739e-01 -3.99548084e-01 3.62118095e-01
-4.13207673e-02 -4.67602313e-01 6.76944852e-01 9.04206455e-01
7.26923466e-01 7.22788155e-01 -4.07942981e-01 4.61657792e-01
4.25034463e-01 3.09274942e-01 -2.33688325e-01 6.31122053e-01
1.01519087e-02 9.32474256e-01 -1.01374018e+00 -2.86864609e-01
-1.98850468e-01 1.06880593e+00 -4.75745559e-01 -3.29013802e-02
-1.27299416e+00 4.13618326e-01 -7.06404671e-02 7.93524861e-01
-2.94139147e-01 -3.78725976e-02 -9.71778750e-01 -4.96294469e-01
-1.31051376e-01 -3.60848248e-01 4.93755817e-01 -9.89924848e-01
-1.13389039e+00 9.51495290e-01 2.02708885e-01 -1.33907926e+00
-2.63329577e-02 -3.82700115e-01 -4.82356191e-01 9.70725358e-01
-9.78526771e-01 -6.30479157e-01 -6.42668784e-01 3.31055015e-01
5.33596098e-01 3.64415944e-01 4.94024694e-01 6.36241436e-01
-2.28351071e-01 1.68466806e-01 2.39866659e-01 -5.09860396e-01
4.95994061e-01 -9.99260783e-01 -4.45363402e-01 5.48093021e-01
-6.26215339e-01 1.01040125e-01 1.05846322e+00 -1.12465167e+00
-1.48885739e+00 -7.27011085e-01 2.24926379e-02 -2.85450667e-01
2.46931121e-01 2.41559505e-01 -8.03722858e-01 5.53363740e-01
4.15220261e-01 5.70098348e-02 6.50388300e-01 -8.00901651e-01
6.31559670e-01 1.86317757e-01 -1.69871461e+00 2.51973987e-01
5.28487504e-01 2.46094808e-01 -4.45565224e-01 4.43443000e-01
1.85231894e-01 -9.81196821e-01 -9.79061484e-01 9.13430631e-01
3.49229306e-01 -1.06220043e+00 8.84268463e-01 1.71138525e-01
5.27045131e-01 -2.44463488e-01 1.19684242e-01 -1.20428932e+00
-5.83126664e-01 -1.06232233e-01 4.08493966e-01 5.26503444e-01
3.28129411e-01 -4.90382165e-01 8.17348719e-01 7.69415200e-01
-7.49053061e-01 -7.66148686e-01 -1.12869561e+00 -3.78323406e-01
7.03110248e-02 -4.84893173e-01 -2.07866341e-01 8.51995647e-01
-1.49827480e-01 -2.08670139e-01 -2.17809111e-01 3.01889420e-01
1.03194225e+00 -3.68677378e-01 4.26311716e-02 -9.01886761e-01
-4.58274633e-01 2.05004588e-01 -1.61371693e-01 -4.03836757e-01
-8.20449114e-01 -7.80222356e-01 2.34965920e-01 -1.97943246e+00
6.62542820e-01 -6.73259139e-01 -5.67878187e-02 -4.20634225e-02
-9.69398692e-02 4.97018516e-01 -3.17599177e-01 2.81065106e-01
4.53718752e-01 4.34863538e-01 1.86635685e+00 1.79435223e-01
-2.22647786e-01 -3.24849016e-03 -1.63122967e-01 7.80682743e-01
3.51755410e-01 -8.62140119e-01 -4.27299857e-01 -2.96649069e-01
1.91672165e-02 6.58076048e-01 2.85011262e-01 -1.37212300e+00
8.76919478e-02 1.06843673e-01 6.64235711e-01 -9.00672555e-01
1.72630623e-01 -1.28159499e+00 1.22326648e+00 1.24883723e+00
2.39188984e-01 1.33702368e-01 4.36971575e-01 1.74203798e-01
-7.36632124e-02 -7.13030159e-01 1.41069722e+00 -3.99794638e-01
-7.43509308e-02 4.48436402e-02 -5.93211055e-01 -4.66998369e-01
1.16947091e+00 -4.48848754e-01 1.69810757e-01 1.55262694e-01
-1.06079233e+00 -3.44050795e-01 2.61157691e-01 -2.79803574e-01
7.60028660e-01 -1.37613118e+00 -8.45474303e-01 -1.05069838e-01
-3.52842867e-01 3.08768731e-03 1.13068807e+00 1.52921224e+00
-1.09816802e+00 7.10048750e-02 -9.81408358e-02 -8.23368490e-01
-1.24427676e+00 4.20594007e-01 7.14487374e-01 -6.86241925e-01
-9.14213955e-01 7.77668357e-01 3.64796668e-01 2.85092145e-01
-5.70549726e-01 -3.26862395e-01 -6.61177710e-02 -3.66305053e-01
2.71958411e-01 3.85712981e-01 7.28452802e-01 -5.55998623e-01
-3.31909269e-01 8.63031507e-01 -1.29558831e-01 -3.08181137e-01
1.34528530e+00 -1.04962073e-01 -8.49230438e-02 1.63123354e-01
8.16467583e-01 -3.71840745e-02 -6.90651000e-01 3.48942190e-01
-2.65792310e-01 -5.54316461e-01 4.31208998e-01 -9.79000986e-01
-9.28027689e-01 1.96002647e-01 1.23764646e+00 -2.53879249e-01
1.28468001e+00 -2.95370638e-01 7.73781359e-01 -6.17270887e-01
5.03205001e-01 -8.32511783e-01 6.12941049e-02 -1.29736871e-01
8.42271626e-01 -8.81982207e-01 4.38670367e-01 -9.83568370e-01
-7.98298717e-01 1.05207968e+00 5.55545926e-01 -9.19822752e-02
7.84483790e-01 8.19703877e-01 -1.21784322e-01 -2.45288014e-01
-2.72217840e-01 4.92468029e-01 -4.47397977e-02 3.16602558e-01
6.88308537e-01 -7.73850530e-02 -1.14853621e+00 4.76201773e-01
2.28893057e-01 2.56948650e-01 8.85078549e-01 1.12514293e+00
-4.22908127e-01 -9.38396096e-01 -9.81285572e-01 4.39998686e-01
-6.89792514e-01 -3.19729410e-02 4.03066009e-01 5.36943197e-01
-1.98214706e-02 5.50024033e-01 -3.59569579e-01 6.53755292e-02
5.03554583e-01 -8.99172500e-02 9.64345753e-01 -4.11957860e-01
-8.08786333e-01 5.38836539e-01 -7.54892230e-02 -4.05179530e-01
-4.67436224e-01 -5.86167574e-01 -1.50773132e+00 7.37702250e-02
-5.58374643e-01 4.56087887e-01 1.16873121e+00 8.90093684e-01
-2.04754025e-01 1.21343398e+00 4.76623923e-01 -5.29152989e-01
-2.68477887e-01 -1.31192851e+00 -6.74095094e-01 3.39119107e-01
-1.31981120e-01 -7.34562695e-01 -3.84411931e-01 -1.08933635e-01] | [13.23569107055664, -2.635951042175293] |
6ec62e25-07b9-41a2-8141-48e9ec32188b | coloc-conditioned-localizer-and-classifier | 2210.13932 | null | https://arxiv.org/abs/2210.13932v1 | https://arxiv.org/pdf/2210.13932v1.pdf | CoLoC: Conditioned Localizer and Classifier for Sound Event Localization and Detection | In this article, we describe Conditioned Localizer and Classifier (CoLoC) which is a novel solution for Sound Event Localization and Detection (SELD). The solution constitutes of two stages: the localization is done first and is followed by classification conditioned by the output of the localizer. In order to resolve the problem of the unknown number of sources we incorporate the idea borrowed from Sequential Set Generation (SSG). Models from both stages are SELDnet-like CRNNs, but with single outputs. Conducted reasoning shows that such two single-output models are fit for SELD task. We show that our solution improves on the baseline system in most metrics on the STARSS22 Dataset. | ['Jakub Tkaczuk', 'Sławomir Kapka'] | 2022-10-25 | null | null | null | null | ['sound-event-localization-and-detection'] | ['audio'] | [ 1.44973293e-01 1.50684521e-01 3.15330505e-01 -2.47075588e-01
-1.32081461e+00 -7.36385286e-01 4.95235860e-01 -9.13304836e-02
-4.23760891e-01 5.68173766e-01 2.14204729e-01 -3.26708883e-01
1.51247576e-01 -3.46119851e-01 -8.16141129e-01 -5.95069587e-01
-5.75677045e-02 2.64078323e-02 6.29913867e-01 -9.36129987e-02
1.66160196e-01 5.34452617e-01 -1.51605332e+00 6.67892814e-01
3.10028195e-01 1.04401910e+00 5.01531243e-01 9.20208395e-01
4.02179845e-02 1.25369847e+00 -7.19352841e-01 7.29449540e-02
4.92644385e-02 -6.29884243e-01 -7.04935908e-01 -5.10066211e-01
2.15067416e-01 5.46676554e-02 -1.54599860e-01 7.17049122e-01
7.50958920e-01 1.77328572e-01 6.87484443e-01 -1.13616800e+00
-3.13793659e-01 1.16199899e+00 -3.35178941e-01 3.61530691e-01
3.61865819e-01 -2.00676218e-01 1.18794501e+00 -1.27191758e+00
3.98193538e-01 1.06144547e+00 8.60469699e-01 5.04380047e-01
-1.13105834e+00 -6.73526824e-01 1.09093778e-01 2.45456114e-01
-1.56621766e+00 -7.92209983e-01 6.59768820e-01 -2.43065700e-01
1.13123035e+00 2.05198467e-01 2.88712472e-01 1.41062498e+00
-1.51074722e-01 7.15211153e-01 1.06606483e+00 -8.66598010e-01
6.01828516e-01 2.91090757e-01 2.01886371e-01 4.53779399e-01
-5.74950278e-02 2.16985062e-01 -8.86810541e-01 1.44561590e-03
4.58109826e-01 -3.07744920e-01 -3.64991099e-01 2.01100200e-01
-1.04575658e+00 6.33152783e-01 4.41350520e-01 5.87350488e-01
-2.76043326e-01 4.90706414e-01 1.06679097e-01 1.90553218e-01
4.95426327e-01 6.17313921e-01 -6.49315357e-01 9.35687870e-02
-1.26480126e+00 1.66804925e-01 8.74132454e-01 1.11166430e+00
4.22877312e-01 1.25714168e-01 -3.51274610e-01 7.40319610e-01
4.78128433e-01 4.07088399e-01 6.03952289e-01 -7.93345332e-01
3.70061427e-01 4.04210314e-02 1.37458548e-01 -6.87485397e-01
-6.05977476e-01 -6.37815297e-01 -2.52397776e-01 -1.43925578e-03
3.85354191e-01 -3.46290737e-01 -9.38124955e-01 1.83321929e+00
-3.70358229e-02 7.56376207e-01 1.20500855e-01 7.71105111e-01
7.50095189e-01 7.35675752e-01 2.19559461e-01 4.28242143e-03
1.26129305e+00 -8.95265400e-01 -7.94441581e-01 -3.19442213e-01
5.28437793e-01 -4.84395683e-01 6.85369313e-01 4.52985734e-01
-9.03788388e-01 -6.23724818e-01 -1.23951280e+00 -2.20392644e-02
-5.16672790e-01 4.90071863e-01 5.31708896e-01 6.26000047e-01
-1.33855164e+00 5.45000017e-01 -8.05978656e-01 -4.31092143e-01
1.23780064e-01 1.53731123e-01 4.40420001e-04 5.22051811e-01
-1.11816168e+00 8.16274941e-01 2.65776664e-01 1.85721830e-01
-1.17986703e+00 -2.73241907e-01 -6.41363740e-01 2.13624209e-01
2.78026104e-01 -2.68615901e-01 1.64797437e+00 -4.85603601e-01
-1.54708254e+00 5.62273383e-01 -2.39465177e-01 -8.46040368e-01
9.42315459e-02 -2.52225995e-01 -4.20207649e-01 3.70905519e-01
7.77309164e-02 5.41749895e-01 9.66757238e-01 -1.32683039e+00
-8.83884907e-01 -2.51050871e-02 -1.93332881e-01 -6.60831854e-02
6.15988933e-02 4.14286733e-01 -1.20606199e-01 -5.84277570e-01
3.93210709e-01 -5.17412007e-01 -2.68198907e-01 -3.70429367e-01
-5.87037981e-01 -2.65772462e-01 3.93159181e-01 -5.78942657e-01
1.05082583e+00 -2.38375449e+00 -1.90159976e-01 4.89574336e-02
-2.44308747e-02 7.82454666e-03 -3.32358062e-01 6.41891301e-01
-4.53725994e-01 2.37469792e-01 -7.60575607e-02 -8.40849757e-01
8.89004394e-02 -3.64256948e-02 -9.26218152e-01 3.08156908e-01
3.56239259e-01 6.61557496e-01 -8.19462001e-01 -4.42253053e-01
-1.55734479e-01 1.05189621e-01 -5.25474250e-01 3.72094363e-01
-3.32467794e-01 3.46501768e-01 -2.37631842e-01 5.49674749e-01
4.30572331e-01 -3.53187963e-04 6.83884919e-02 -1.40937105e-01
-3.16828728e-01 7.22985268e-01 -1.45915568e+00 1.73457253e+00
-5.09177566e-01 6.03927433e-01 9.71454754e-02 -8.79716814e-01
8.59104037e-01 7.13014424e-01 9.19934809e-02 -2.67054021e-01
1.21360175e-01 3.14981103e-01 -9.12207216e-02 -4.24226761e-01
3.70547235e-01 -5.61069250e-02 -1.86821237e-01 5.34667552e-01
7.24988759e-01 -1.25178799e-01 -6.73489785e-03 4.77395475e-01
1.36278462e+00 3.87256980e-01 3.22750241e-01 -5.36663122e-02
2.32776865e-01 -1.96022525e-01 4.80541974e-01 1.36315668e+00
-9.88672301e-03 8.87798727e-01 3.90013903e-01 1.42574742e-01
-5.01147985e-01 -1.10733914e+00 2.34268587e-02 1.10130227e+00
-9.13574994e-02 -5.11422276e-01 -7.62871146e-01 -6.64815009e-01
-2.74206668e-01 8.64534140e-01 -3.96128803e-01 -6.54179975e-02
-4.89901870e-01 -6.18011832e-01 9.99966383e-01 7.80561566e-01
-2.39155791e-03 -1.31627762e+00 -7.16426492e-01 3.39212000e-01
-2.34906614e-01 -1.38125861e+00 -6.86883703e-02 1.01060760e+00
-4.60821092e-01 -8.13572466e-01 -3.12441915e-01 -8.61795783e-01
2.38273889e-01 1.79576904e-01 9.55425143e-01 -7.71386325e-02
-1.23370737e-01 2.04674318e-01 -7.15399742e-01 -7.11132348e-01
-3.16351801e-01 3.71302478e-02 1.47124007e-01 2.25801468e-01
2.01173440e-01 -8.53847265e-01 -1.86058909e-01 -1.52226761e-01
-5.39593995e-01 -1.93586126e-01 6.18790209e-01 6.40741229e-01
4.63181406e-01 -5.02599590e-02 9.51094031e-01 -6.20887101e-01
4.14774030e-01 -6.31776571e-01 -5.62511146e-01 1.36976734e-01
-3.19241405e-01 -4.85583432e-02 6.33712351e-01 -5.37727177e-01
-9.14376736e-01 4.04061913e-01 -4.26661223e-01 -3.53523016e-01
-4.63697135e-01 1.49872184e-01 -7.71508068e-02 1.41309664e-01
8.16566944e-01 1.90395981e-01 -5.50884306e-01 -8.24392617e-01
4.91016090e-01 9.01526868e-01 6.56184793e-01 -4.90672052e-01
6.36207104e-01 3.12326580e-01 -3.13231558e-01 -5.40153325e-01
-1.15866148e+00 -4.45373118e-01 -5.34213185e-01 -1.43088192e-01
7.69932926e-01 -1.15524101e+00 -6.10859036e-01 2.99724609e-01
-1.60611010e+00 -2.64615327e-01 -5.57892025e-01 6.40014946e-01
-4.37230438e-01 -1.07543848e-01 -7.79275179e-01 -1.19310343e+00
-7.65296295e-02 -8.26711118e-01 1.08905756e+00 1.79152697e-01
-4.86928746e-02 -5.90799689e-01 1.41533554e-01 -2.61283338e-01
3.05231065e-01 -1.15015976e-01 4.70798969e-01 -1.30916917e+00
-4.79639053e-01 -1.94031876e-02 -7.54685402e-02 4.28161114e-01
-4.09838110e-01 -3.03143144e-01 -1.67106247e+00 1.56813145e-01
4.71748888e-01 -3.49914610e-01 1.09105897e+00 2.69473016e-01
1.09599733e+00 -1.00190729e-01 -4.23072308e-01 5.93975067e-01
1.46570468e+00 3.49859834e-01 4.04295206e-01 -9.04480740e-02
3.79601479e-01 5.08883536e-01 2.85910577e-01 1.72063053e-01
2.47546002e-01 4.52063292e-01 3.33652467e-01 -1.43157735e-01
-4.07874584e-01 -5.73091686e-01 5.89361191e-01 9.60065007e-01
3.34622830e-01 -5.17414868e-01 -7.23233521e-01 7.78635383e-01
-1.84921527e+00 -9.98759508e-01 -4.00097758e-01 1.76014817e+00
7.72616506e-01 1.03361681e-01 -6.66476712e-02 3.70250940e-01
6.38387561e-01 -2.48966203e-03 -1.99642759e-02 -3.32883924e-01
-1.66547000e-01 6.90521240e-01 4.04899329e-01 4.78475600e-01
-8.48587990e-01 1.01473987e+00 7.44261360e+00 9.36041951e-01
-8.54935110e-01 5.88757992e-01 1.59310803e-01 -2.98554134e-02
4.35826778e-02 2.84197092e-01 -1.06329215e+00 5.27933061e-01
1.29408264e+00 5.32660007e-01 4.57746327e-01 7.16093957e-01
5.77006936e-02 -3.83786112e-01 -1.36139381e+00 8.17135572e-01
1.81604385e-01 -1.09541547e+00 -3.71476203e-01 -3.40083897e-01
4.23184961e-01 2.46146664e-01 -2.72422224e-01 5.23193479e-01
3.70747089e-01 -6.98627591e-01 1.32043445e+00 5.41996181e-01
6.39872909e-01 -4.86389041e-01 5.84122658e-01 5.52271426e-01
-1.11107540e+00 -2.43790165e-01 -2.84945816e-01 -8.68337750e-02
1.81073904e-01 6.14562094e-01 -1.07754457e+00 4.55132842e-01
8.07240248e-01 4.64918405e-01 -5.52546859e-01 1.06214964e+00
-7.52331495e-01 1.36508358e+00 -5.25659919e-01 -1.70143977e-01
-8.68473295e-03 3.09413552e-01 7.35156953e-01 1.60254943e+00
3.20435047e-01 -6.67292550e-02 9.21780840e-02 1.24979031e+00
-2.65670102e-02 -1.36418045e-01 -5.99115014e-01 2.28176862e-01
7.99915552e-01 1.33780456e+00 -8.73708189e-01 -2.91766256e-01
-1.22842938e-01 8.46402228e-01 4.66831446e-01 4.45479602e-01
-8.54893208e-01 -6.09480798e-01 -1.18357502e-01 -2.03405946e-01
6.01672828e-01 1.09732915e-02 -2.82379329e-01 -9.10525620e-01
-1.79031789e-01 -4.38022137e-01 1.81100011e-01 -1.05021870e+00
-1.07895553e+00 6.60916865e-01 -1.27864420e-01 -1.13608158e+00
-3.07823569e-01 -6.46561980e-01 -7.36152351e-01 8.10374439e-01
-1.51449871e+00 -1.10736084e+00 1.84850454e-01 4.63713735e-01
4.08766598e-01 -8.21213424e-02 1.03393483e+00 3.60958606e-01
-5.33115447e-01 4.13682729e-01 -2.43287116e-01 2.10097194e-01
6.52645826e-01 -1.49415648e+00 4.46784079e-01 1.22643292e+00
6.70484006e-01 5.47399819e-01 5.27343631e-01 -3.98019135e-01
-9.79280293e-01 -1.01525795e+00 1.35402083e+00 -5.13799489e-01
6.75211847e-01 -9.51617360e-01 -4.64137018e-01 8.07323456e-01
1.90305635e-01 2.18614284e-02 4.19384837e-01 2.49449998e-01
-4.30657804e-01 -1.59254462e-01 -9.73806381e-01 3.39328736e-01
1.00667989e+00 -7.87599683e-01 -7.83332348e-01 3.63091290e-01
9.21317220e-01 -2.99609125e-01 -4.29311603e-01 1.87953711e-01
9.80405957e-02 -7.86076188e-01 7.81782925e-01 -3.65403086e-01
7.13703781e-02 -5.08088708e-01 -4.86736983e-01 -1.24885285e+00
-2.17344999e-01 -8.03663373e-01 -2.39362866e-01 1.37609637e+00
6.45636082e-01 -5.89792788e-01 4.34195660e-02 -1.28179982e-01
-5.19989371e-01 -4.59722787e-01 -9.82786775e-01 -7.43739843e-01
-3.61394823e-01 -1.04536855e+00 4.39331740e-01 5.61578333e-01
7.19305053e-02 7.31446147e-01 -2.72117823e-01 5.33137381e-01
3.60274404e-01 -3.46185081e-02 2.74263382e-01 -9.90136683e-01
-6.23358130e-01 -1.11747809e-01 -2.87820455e-02 -1.09538782e+00
-3.66925634e-02 -1.00300181e+00 6.89320266e-01 -1.20699024e+00
-7.13280216e-02 -5.17907798e-01 -6.16199195e-01 7.90094078e-01
1.05605863e-01 3.75466764e-01 2.17741191e-01 9.96129811e-02
-7.95874178e-01 2.00440928e-01 3.91174138e-01 2.23167166e-01
-1.93381891e-01 1.05185531e-01 -9.08159435e-01 8.08193922e-01
8.11622143e-01 -9.68598485e-01 -1.48040682e-01 -2.44233131e-01
3.51393104e-01 2.91712821e-01 7.13179767e-01 -1.36501527e+00
6.56620502e-01 4.25970107e-01 3.07162672e-01 -8.45476568e-01
4.65808123e-01 -7.44455993e-01 -1.71709731e-01 9.10397023e-02
-5.12950778e-01 -1.80387467e-01 3.85571755e-02 5.11181176e-01
-1.41869560e-01 -5.56597710e-01 6.65066183e-01 -8.45321491e-02
-4.88077462e-01 -3.03863108e-01 -6.11630678e-01 -5.53873815e-02
6.36089325e-01 2.26646990e-01 -1.11473233e-01 -2.43597239e-01
-1.04879916e+00 -6.17184602e-02 -3.55218023e-01 2.51383215e-01
5.63577473e-01 -1.02549708e+00 -7.57506132e-01 1.27751142e-01
-4.95182537e-02 -7.60732964e-02 -2.43743405e-01 8.32415700e-01
1.15141114e-02 3.19716543e-01 4.95627761e-01 -4.75106508e-01
-5.19514680e-01 4.87771899e-01 5.85099220e-01 -1.05229244e-01
-5.46089888e-01 1.19494295e+00 -3.29514034e-02 -3.57369244e-01
5.12475669e-01 -6.91225350e-01 -3.34151417e-01 1.42969176e-01
5.91761529e-01 1.31283164e-01 2.13680834e-01 -2.99356461e-01
-4.21765119e-01 1.45450905e-01 6.90339208e-02 -6.71605825e-01
1.45000339e+00 -1.12467207e-01 6.95955679e-02 8.17750335e-01
8.91148806e-01 4.38099921e-01 -8.59491467e-01 -6.11525327e-02
3.30893606e-01 2.57467210e-01 4.59134988e-02 -1.17107534e+00
-6.94093525e-01 8.62236619e-01 5.78978837e-01 3.06615144e-01
9.90068376e-01 2.65847355e-01 4.83389705e-01 3.35465372e-01
5.18018186e-01 -1.08751762e+00 -1.00892216e-01 5.36837041e-01
8.62238824e-01 -6.56284332e-01 -3.84906799e-01 -1.48969844e-01
-4.49378133e-01 9.01901782e-01 4.77328300e-01 -3.77474934e-01
7.80164182e-01 7.19614148e-01 -2.07196753e-02 -2.18635961e-01
-8.92481506e-01 -3.94023925e-01 1.20239012e-01 5.46257794e-01
2.30053559e-01 -1.26856580e-01 -2.31226087e-01 1.24571896e+00
-1.55272052e-01 -2.48289675e-01 3.27512532e-01 8.95915926e-01
-6.22410238e-01 -8.19079161e-01 -4.50131357e-01 1.75549507e-01
-6.41843617e-01 -4.41942126e-01 -3.86455774e-01 5.32531619e-01
5.42274833e-01 1.48372185e+00 -1.58051953e-01 -6.07009470e-01
2.97743201e-01 2.34633550e-01 2.71432132e-01 -9.63504136e-01
-8.13423395e-01 3.69880348e-01 3.58372286e-04 -6.19925499e-01
-3.60917687e-01 -6.38029218e-01 -1.45090032e+00 6.79102361e-01
-7.62122214e-01 3.51028830e-01 8.27813923e-01 9.39532816e-01
3.23741883e-01 8.61016989e-01 8.05980265e-01 -9.58551645e-01
-7.56515503e-01 -1.36170197e+00 -7.50103295e-01 -1.29161969e-01
5.71388423e-01 -3.58943462e-01 -7.13909984e-01 1.04801729e-01] | [15.20635986328125, 5.2278265953063965] |
95042d31-09d6-44aa-8b3f-0eae2eb7fda6 | do-transformers-parse-while-predicting-the | 2303.08117 | null | https://arxiv.org/abs/2303.08117v1 | https://arxiv.org/pdf/2303.08117v1.pdf | Do Transformers Parse while Predicting the Masked Word? | Pre-trained language models have been shown to encode linguistic structures, e.g. dependency and constituency parse trees, in their embeddings while being trained on unsupervised loss functions like masked language modeling. Some doubts have been raised whether the models actually are doing parsing or only some computation weakly correlated with it. We study questions: (a) Is it possible to explicitly describe transformers with realistic embedding dimension, number of heads, etc. that are capable of doing parsing -- or even approximate parsing? (b) Why do pre-trained models capture parsing structure? This paper takes a step toward answering these questions in the context of generative modeling with PCFGs. We show that masked language models like BERT or RoBERTa of moderate sizes can approximately execute the Inside-Outside algorithm for the English PCFG [Marcus et al, 1993]. We also show that the Inside-Outside algorithm is optimal for masked language modeling loss on the PCFG-generated data. We also give a construction of transformers with $50$ layers, $15$ attention heads, and $1275$ dimensional embeddings in average such that using its embeddings it is possible to do constituency parsing with $>70\%$ F1 score on PTB dataset. We conduct probing experiments on models pre-trained on PCFG-generated data to show that this not only allows recovery of approximate parse tree, but also recovers marginal span probabilities computed by the Inside-Outside algorithm, which suggests an implicit bias of masked language modeling towards this algorithm. | ['Sanjeev Arora', 'Rong Ge', 'Abhishek Panigrahi', 'Haoyu Zhao'] | 2023-03-14 | null | null | null | null | ['constituency-parsing'] | ['natural-language-processing'] | [ 2.94096302e-02 1.11218226e+00 1.50743112e-01 -6.41443908e-01
-1.06930339e+00 -6.28759444e-01 2.12341458e-01 1.65140957e-01
-4.31445062e-01 6.17013514e-01 4.91351247e-01 -7.72547543e-01
3.93089503e-01 -1.19868398e+00 -1.06322908e+00 -5.48938155e-01
-3.79312515e-01 6.63965702e-01 8.37895423e-02 -6.41945824e-02
-2.56429315e-01 3.94704074e-01 -1.11952972e+00 4.02787805e-01
6.43862784e-01 5.66118777e-01 3.56619745e-01 1.04642069e+00
-5.79928935e-01 9.63320076e-01 -3.95980120e-01 -9.60242748e-01
2.98575193e-01 -2.68127441e-01 -9.52102184e-01 -1.04766369e-01
3.62802982e-01 -4.15711224e-01 -2.08731934e-01 9.53635097e-01
2.15350807e-01 -2.36045286e-01 6.17774904e-01 -6.59895480e-01
-8.49584222e-01 1.29615152e+00 -1.60106435e-01 3.01696450e-01
-4.24245372e-02 9.59804282e-02 1.58415282e+00 -7.12854445e-01
4.40998852e-01 1.52172327e+00 5.77449620e-01 8.12750697e-01
-1.60671258e+00 -5.46361387e-01 2.05413103e-01 -5.52416444e-01
-9.16795135e-01 -2.96715379e-01 5.71552336e-01 -2.40484595e-01
1.58511770e+00 8.30805376e-02 2.47564554e-01 8.27484608e-01
2.47205049e-01 7.51699388e-01 9.77447510e-01 -6.88431442e-01
1.36495426e-01 2.65832543e-01 4.69264150e-01 1.14946496e+00
4.20860052e-01 2.51294702e-01 -3.18590999e-01 -1.06527217e-01
5.22912383e-01 -5.99666595e-01 1.73623428e-01 8.79900008e-02
-6.73381925e-01 1.34642315e+00 3.09842408e-01 4.22554225e-01
-5.52909030e-03 3.53402495e-01 3.23018759e-01 2.44036406e-01
5.76208949e-01 3.91367733e-01 -6.65575087e-01 9.57093239e-02
-6.78379238e-01 2.34023347e-01 8.51464391e-01 1.10887063e+00
1.00766706e+00 2.16915175e-01 2.90990829e-01 6.22243643e-01
1.69181868e-01 3.55271727e-01 3.23996514e-01 -9.94625509e-01
8.29468310e-01 5.36858737e-01 -8.54057595e-02 -3.15099090e-01
-4.41577882e-01 -1.01634771e-01 -4.49333042e-01 1.45254850e-01
7.75641024e-01 -3.27588379e-01 -7.28593946e-01 2.34954691e+00
-5.09702181e-03 -4.28953081e-01 1.27058014e-01 3.77192765e-01
3.90326917e-01 9.46806788e-01 4.06984419e-01 -8.32606703e-02
1.51329660e+00 -5.87403357e-01 -5.43992162e-01 -6.74575984e-01
1.10765600e+00 -6.32948458e-01 1.39690006e+00 7.79158697e-02
-1.42454433e+00 -6.66902304e-01 -1.01227379e+00 -5.53948224e-01
-2.35400110e-01 1.32861868e-01 1.06438112e+00 1.23996115e+00
-1.26816916e+00 6.94088399e-01 -1.12983918e+00 -9.30911824e-02
2.34475285e-01 5.36684453e-01 -5.27202010e-01 -1.33163286e-02
-1.02940941e+00 9.28106487e-01 2.66855329e-01 -1.15818299e-01
-9.05724227e-01 -3.25542301e-01 -1.13331282e+00 3.25577468e-01
-1.84254080e-01 -3.54343921e-01 1.09569275e+00 -7.24070370e-01
-1.08501601e+00 1.08968818e+00 -4.64973003e-01 -7.11000800e-01
-1.78066827e-02 -8.22701231e-02 -2.36848712e-01 -7.28950575e-02
9.92674306e-02 7.45307207e-01 4.93510216e-01 -1.02588785e+00
-5.17514110e-01 -5.11141598e-01 3.90306622e-01 -1.65239915e-01
-2.64913559e-01 2.81330585e-01 2.68920243e-01 -3.82607847e-01
2.99246579e-01 -7.14730144e-01 -2.24240094e-01 -3.60384732e-01
-3.91771346e-01 -4.56505865e-01 1.42994806e-01 -8.97539496e-01
1.08717263e+00 -2.03022003e+00 -2.33664513e-01 -2.21969262e-01
1.79051906e-01 1.27716064e-01 -2.27567241e-01 4.28371817e-01
-3.48280489e-01 6.58727586e-01 -3.70544881e-01 -7.52490997e-01
4.39814657e-01 5.96423984e-01 -6.30982697e-01 2.62089491e-01
5.14425218e-01 8.23933005e-01 -3.53351057e-01 -4.76330638e-01
-1.13784477e-01 2.20137596e-01 -9.94718790e-01 4.48799044e-01
-1.82399571e-01 -1.45375147e-01 -1.08787112e-01 4.85359728e-01
7.50535727e-01 1.27238780e-01 4.85621274e-01 -3.48497764e-03
-3.18190865e-02 9.11073565e-01 -8.63528967e-01 1.28032196e+00
-7.77096033e-01 5.10908365e-01 3.55798334e-01 -1.09180081e+00
7.51190782e-01 2.97989249e-01 -2.28274971e-01 -3.16147536e-01
1.52977109e-01 2.37986401e-01 3.55540574e-01 -8.04598480e-02
3.91443253e-01 -7.45650172e-01 -5.20519078e-01 5.69939375e-01
4.66214597e-01 -6.84635714e-02 1.91427067e-01 3.03621918e-01
1.43668556e+00 -1.00147672e-01 -1.11856923e-01 -4.15669560e-01
7.52515495e-02 3.22735705e-03 7.34672546e-01 5.88998020e-01
-2.10465670e-01 3.81311655e-01 8.60935867e-01 -6.21691644e-01
-1.25978994e+00 -1.50675833e+00 -1.83953092e-01 1.36152565e+00
-4.05756384e-01 -4.47045028e-01 -1.10862637e+00 -6.18703067e-01
-3.49652231e-01 1.22588491e+00 -6.58269048e-01 2.36432366e-02
-1.10658395e+00 -1.14529896e+00 7.48129129e-01 9.49940383e-01
1.63992748e-01 -1.18746114e+00 -3.45376551e-01 4.95415390e-01
3.56407054e-02 -1.21994162e+00 -2.71064401e-01 9.41720605e-01
-1.23841631e+00 -7.70582855e-01 -7.63501525e-02 -1.20855772e+00
8.87181103e-01 -5.54530740e-01 1.30761182e+00 -3.85620967e-02
-5.28308824e-02 -3.32684927e-02 -1.10776789e-01 -1.10042334e-01
-7.89111078e-01 4.78496142e-02 -1.45425811e-01 -5.10888577e-01
6.96544468e-01 -7.07945406e-01 -1.95101634e-01 -1.31886467e-01
-8.19523036e-01 -1.54640600e-01 5.73646367e-01 8.62764835e-01
3.89176369e-01 -9.96137876e-03 2.56963879e-01 -1.20295465e+00
5.04990995e-01 -7.11988285e-02 -7.10171282e-01 1.38849571e-01
-4.35407460e-01 5.24093509e-01 1.06217098e+00 -1.16447799e-01
-1.04450297e+00 3.33114386e-01 -6.05746329e-01 1.71434969e-01
-1.17428750e-01 -8.54525045e-02 -7.13343799e-01 3.84300083e-01
7.62508512e-01 1.62257850e-01 -4.23288196e-01 -5.72138965e-01
5.91082752e-01 3.77883554e-01 2.48647630e-01 -1.05894506e+00
9.25426602e-01 3.03004891e-01 -2.59595066e-01 -6.33287489e-01
-6.96528018e-01 4.25653197e-02 -4.33782130e-01 6.84251189e-01
1.15735197e+00 -9.34357405e-01 -4.43754345e-01 -1.15426086e-01
-1.49606621e+00 -4.81036216e-01 -4.79985356e-01 3.32182109e-01
-6.83536112e-01 3.16810519e-01 -1.23863435e+00 -1.02886283e+00
-3.26545298e-01 -1.12421966e+00 8.36765885e-01 -1.84985161e-01
-3.02534550e-01 -1.03938186e+00 -7.75546348e-03 3.41678083e-01
1.99219748e-01 -8.17102194e-02 1.67743981e+00 -8.64788651e-01
-6.81074739e-01 -1.00168779e-01 -1.80709869e-01 7.90323436e-01
-4.08909693e-02 -2.38386035e-01 -1.23335540e+00 -5.36693931e-02
1.61531359e-01 -2.96313822e-01 9.04610634e-01 4.17173743e-01
8.25219274e-01 -5.34563959e-01 -3.29941399e-02 5.68334639e-01
1.45506012e+00 4.88681048e-02 5.78434587e-01 -1.57159790e-01
4.30251539e-01 7.84809887e-01 3.03048473e-02 -2.00189441e-01
4.03635293e-01 9.84768290e-03 2.50425220e-01 1.97964311e-01
-1.74725465e-02 -6.84976280e-01 7.27122426e-01 1.05001771e+00
1.96852610e-01 -1.78367808e-01 -6.67098224e-01 8.80383551e-01
-1.13345432e+00 -6.89607501e-01 -1.37758568e-01 1.99661791e+00
9.43595529e-01 6.81896389e-01 -1.77540243e-01 1.11887187e-01
7.00451016e-01 1.68632925e-01 -9.85837430e-02 -1.36398363e+00
-1.65636107e-01 1.10179901e+00 6.29912078e-01 1.06101942e+00
-8.89925718e-01 1.17183852e+00 6.37881041e+00 6.60316885e-01
-6.20333493e-01 4.86023158e-01 8.90338361e-01 1.30096346e-01
-6.97942019e-01 3.89308125e-01 -1.16133940e+00 2.53834724e-01
1.43043625e+00 5.07561192e-02 4.89820093e-01 9.58123028e-01
-1.58750743e-01 2.52844244e-02 -1.35261929e+00 3.04984510e-01
-3.96242201e-01 -1.10551023e+00 -5.93943894e-02 2.77078807e-01
2.19358459e-01 -6.96570799e-02 -1.06885709e-01 7.21542656e-01
7.74990380e-01 -1.26597536e+00 6.50498152e-01 -3.87075722e-01
6.53957665e-01 -8.34503472e-01 4.96004850e-01 6.55100644e-01
-1.31422734e+00 3.59158367e-02 -9.38260019e-01 -2.51058877e-01
2.34507769e-01 6.11048758e-01 -9.42079902e-01 -3.21133062e-02
3.36771846e-01 -3.56527627e-01 -2.87861258e-01 -9.01996344e-02
-6.13112032e-01 1.10004187e+00 -6.29255712e-01 -1.49203435e-01
2.74601430e-01 -8.59169569e-03 1.36979401e-01 1.34394002e+00
2.70996958e-01 2.01224044e-01 -2.87883371e-01 1.21274507e+00
-1.54104039e-01 8.28429610e-02 -5.87377906e-01 -3.15627635e-01
3.60380322e-01 9.78274882e-01 -6.89226925e-01 -2.84250408e-01
-4.39878881e-01 7.11406589e-01 7.92958498e-01 1.63670499e-02
-7.42412508e-01 -3.46712917e-01 5.13414979e-01 3.89908850e-01
4.59927768e-01 -3.45286697e-01 -5.22558749e-01 -1.13528776e+00
5.14482483e-02 -4.37905550e-01 2.60234326e-01 -5.30683875e-01
-1.13427031e+00 9.32403564e-01 -1.42514482e-01 -3.10257107e-01
-3.62738103e-01 -1.05527580e+00 -7.99760461e-01 1.24742472e+00
-1.28317785e+00 -1.01919317e+00 5.35154343e-01 1.87031165e-01
4.90097940e-01 3.45979184e-02 1.23095679e+00 4.69286256e-02
-4.51728761e-01 7.74001658e-01 -2.25292757e-01 5.12225688e-01
5.82092963e-02 -1.68314302e+00 8.17457438e-01 1.11284161e+00
4.78982776e-01 9.13947403e-01 7.25572646e-01 -3.30288321e-01
-1.15968132e+00 -1.07409692e+00 1.64117229e+00 -7.51819074e-01
4.86887306e-01 -8.94015849e-01 -8.50092888e-01 1.21361673e+00
3.57576370e-01 3.94473225e-02 7.51274467e-01 1.98686644e-01
-4.82661515e-01 -7.87822157e-02 -1.41991067e+00 3.21385533e-01
1.18933094e+00 -4.89461422e-01 -9.77799237e-01 2.90233195e-01
9.37319100e-01 1.24805234e-02 -6.09152198e-01 4.40036729e-02
2.17512146e-01 -1.15681493e+00 7.65150905e-01 -7.35221684e-01
6.20780408e-01 2.44269773e-01 -6.29041612e-01 -1.04766572e+00
-2.74599105e-01 -5.03543854e-01 1.69354156e-01 1.39768946e+00
7.88207948e-01 -7.71568537e-01 9.83379483e-01 8.76802921e-01
-1.76807970e-01 -6.22321606e-01 -1.14996064e+00 -7.32816339e-01
9.82119918e-01 -7.39155352e-01 5.00209808e-01 3.90208006e-01
2.25045849e-02 3.73586297e-01 1.00753017e-01 4.32824731e-01
6.15071893e-01 1.79774061e-01 4.15842652e-01 -8.50297511e-01
-5.84376395e-01 -9.77551267e-02 -1.98571056e-01 -1.11251485e+00
5.24634063e-01 -1.12930143e+00 9.55463350e-02 -1.33054042e+00
-9.42017585e-02 -4.83261108e-01 8.59658346e-02 4.30905163e-01
2.61609200e-02 -1.68153182e-01 6.47411868e-02 -3.05281639e-01
4.08061296e-02 1.20447323e-01 7.93767989e-01 1.18578143e-01
3.21312636e-01 -1.93415850e-01 -9.67619061e-01 9.35909271e-01
8.03477585e-01 -8.25951874e-01 -2.62437254e-01 -7.21498251e-01
2.51240492e-01 2.64617026e-01 1.28059104e-01 -7.80933440e-01
-2.35172391e-01 2.78518319e-01 2.19896451e-01 -2.56415635e-01
3.95310134e-01 -5.70824802e-01 -2.67251462e-01 5.04903615e-01
-3.95544052e-01 3.41844261e-01 2.80316889e-01 2.57086277e-01
7.49191269e-03 -6.89589083e-01 8.22419286e-01 -5.92855036e-01
-2.46550173e-01 -6.40593295e-04 -5.54507017e-01 5.11645079e-01
4.33085710e-01 3.24485116e-02 -8.71738568e-02 -1.15410738e-01
-1.09728205e+00 -2.75024265e-01 4.05573487e-01 -2.40558207e-01
2.49760449e-01 -9.53239501e-01 -6.25619411e-01 3.58137846e-01
-3.62157375e-01 1.16680868e-01 -6.69195205e-02 9.15455297e-02
-8.50357771e-01 3.05739671e-01 2.01850995e-01 4.08629514e-03
-9.07804608e-01 5.50122499e-01 2.36695781e-01 -7.75551200e-01
-4.41962987e-01 1.33911610e+00 5.96726656e-01 -7.51577258e-01
-3.25742476e-02 -9.87017989e-01 5.36744177e-01 -2.36509785e-01
4.01214063e-01 -7.05776066e-02 8.54303762e-02 -5.09271622e-01
-3.68076861e-01 3.88496339e-01 -5.98921925e-02 -2.07767427e-01
1.37063968e+00 1.12900130e-01 -2.38988940e-02 2.90441737e-02
1.44647336e+00 5.04364491e-01 -1.13556004e+00 -9.89592448e-03
9.58500654e-02 4.07692641e-02 -2.50943691e-01 -6.86167598e-01
-8.96809220e-01 1.61372471e+00 3.45232457e-01 3.39772671e-01
9.00666475e-01 3.61465156e-01 1.01227391e+00 3.26246917e-01
5.39956629e-01 -8.92586887e-01 -3.41342777e-01 4.23268795e-01
4.89605963e-01 -9.40763772e-01 -3.81567568e-01 -4.80685472e-01
-4.18785185e-01 9.60180759e-01 4.56884146e-01 -6.76754236e-01
5.06536365e-01 7.64572799e-01 -8.34116116e-02 1.93283856e-02
-1.05241740e+00 -2.14480653e-01 -3.80210131e-01 5.53399801e-01
5.81800878e-01 4.52771246e-01 -3.19400877e-01 1.12217629e+00
-7.53850818e-01 -5.57769001e-01 5.07975161e-01 7.42879331e-01
-7.75192082e-01 -1.38846111e+00 -1.64162830e-01 2.88299322e-01
-8.16522241e-01 -4.15278524e-01 -2.59879440e-01 9.24329400e-01
3.11336756e-01 9.50590074e-01 1.81156784e-01 -2.58133680e-01
2.16714740e-01 6.06312215e-01 7.36154020e-01 -1.05431211e+00
-6.15115881e-01 -1.71296913e-02 4.32963312e-01 -1.90396652e-01
2.57247776e-01 -3.84819835e-01 -1.48249328e+00 -4.19384837e-01
-3.36498171e-01 2.66549349e-01 5.72570503e-01 6.72466874e-01
1.14883734e-02 2.00281963e-01 4.09226507e-01 -4.55347151e-01
-8.85363936e-01 -1.06767547e+00 -8.23758602e-01 2.51486838e-01
9.58056247e-04 -8.32254961e-02 -8.17129314e-01 -9.33811185e-04] | [10.395525932312012, 9.565067291259766] |
fbb35cdf-80b6-4691-b972-536e90d76f3c | icdar-2021-competition-on-historical-map | 2105.13265 | null | https://arxiv.org/abs/2105.13265v1 | https://arxiv.org/pdf/2105.13265v1.pdf | ICDAR 2021 Competition on Historical Map Segmentation | This paper presents the final results of the ICDAR 2021 Competition on Historical Map Segmentation (MapSeg), encouraging research on a series of historical atlases of Paris, France, drawn at 1/5000 scale between 1894 and 1937. The competition featured three tasks, awarded separately. Task~1 consists in detecting building blocks and was won by the L3IRIS team using a DenseNet-121 network trained in a weakly supervised fashion. This task is evaluated on 3 large images containing hundreds of shapes to detect. Task~2 consists in segmenting map content from the larger map sheet, and was won by the UWB team using a U-Net-like FCN combined with a binarization method to increase detection edge accuracy. Task~3 consists in locating intersection points of geo-referencing lines, and was also won by the UWB team who used a dedicated pipeline combining binarization, line detection with Hough transform, candidate filtering, and template matching for intersection refinement. Tasks~2 and~3 are evaluated on 95 map sheets with complex content. Dataset, evaluation tools and results are available under permissive licensing at \url{https://icdar21-mapseg.github.io/}. | ['Pavel Král', 'Ladislav Lenc', 'Josef Baloun', 'Nam Nguyen', 'Vincent Nguyen', 'Thierry Géraud', 'Clément Mallet', 'Bertrand Duménieu', 'Julien Perret', 'Yizi Chen', 'Edwin Carlinet', 'Joseph Chazalon'] | 2021-05-27 | null | null | null | null | ['document-layout-analysis', 'line-segment-detection', 'line-detection', 'contour-detection'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-1.39759168e-01 2.48675108e-01 1.87040702e-01 -3.95691186e-01
-1.04572761e+00 -9.46132123e-01 9.24104691e-01 2.47100681e-01
-7.17555642e-01 4.76728827e-01 2.52030790e-01 -2.07560837e-01
8.00578296e-02 -1.12190199e+00 -8.69921863e-01 -1.87094286e-01
-8.09047520e-01 7.73641646e-01 7.09435999e-01 -5.17421663e-01
3.59904468e-01 8.00322235e-01 -1.25638044e+00 1.97398707e-01
6.19068325e-01 1.11079860e+00 3.85907441e-02 8.42660010e-01
-2.04914603e-02 4.00320739e-01 -4.85182166e-01 -5.44963241e-01
7.68490493e-01 1.29481852e-01 -8.47432852e-01 -4.38469976e-01
9.68978345e-01 -3.15050989e-01 -5.70913911e-01 8.54420960e-01
5.42218626e-01 1.55596659e-01 6.87936008e-01 -1.10052824e+00
-8.96621794e-02 8.35881591e-01 -6.97353959e-01 7.65745759e-01
1.05912223e-01 -2.51486868e-01 1.11941803e+00 -8.97945821e-01
7.15841532e-01 8.76579106e-01 1.31244290e+00 -1.43581508e-02
-9.33933973e-01 -7.16404021e-01 -1.81634016e-02 3.66515040e-01
-1.72207987e+00 -1.77176371e-01 7.11751282e-01 -6.12889826e-01
9.24019516e-01 5.30309856e-01 6.78799450e-01 3.29189956e-01
-2.37841398e-01 9.76988673e-01 8.70123506e-01 -2.46085614e-01
2.06098229e-01 -3.36595267e-01 1.31751031e-01 6.06351852e-01
-2.09302053e-01 -8.74592736e-02 -1.57430887e-01 -9.56649929e-02
8.93859982e-01 -4.22336340e-01 -1.22536816e-01 3.51097658e-02
-1.02910137e+00 9.24507678e-01 9.37773108e-01 3.82836640e-01
-2.97876626e-01 2.67510146e-01 3.61836374e-01 -4.40916121e-02
7.52261579e-01 5.16618371e-01 -1.10811017e-01 -3.64710321e-03
-1.49694693e+00 7.35034645e-01 5.10294497e-01 9.69451904e-01
7.44024634e-01 -2.06056222e-01 8.41965079e-02 9.74081516e-01
-1.28856763e-01 4.60174829e-01 -2.86967397e-01 -1.05933988e+00
5.52833676e-01 5.22989035e-01 -2.29899809e-02 -1.00844419e+00
-1.06560147e+00 -4.55215514e-01 -6.74697399e-01 2.64174342e-01
5.46266198e-01 -2.98798591e-01 -1.42281687e+00 1.18126476e+00
1.20733373e-01 1.33966774e-01 -6.36150181e-01 9.10218835e-01
1.02046537e+00 7.65577495e-01 -1.56794831e-01 5.57016134e-01
1.38294375e+00 -6.25541925e-01 -8.34297091e-02 -1.66769043e-01
4.22573924e-01 -7.08291113e-01 6.11110628e-01 1.59980148e-01
-1.22327232e+00 -4.05704916e-01 -1.11562347e+00 -2.55810350e-01
-7.16881990e-01 -7.16283871e-03 5.42847991e-01 2.93667048e-01
-1.38343167e+00 4.91771221e-01 -7.39430845e-01 -5.53414881e-01
7.94650733e-01 2.61377931e-01 -1.59170821e-01 1.75259471e-01
-1.24423230e+00 9.48830247e-01 3.71410131e-01 3.20929438e-01
-5.01748502e-01 -8.79651010e-01 -9.76662159e-01 -2.03909889e-01
-3.68432738e-02 -2.06563503e-01 1.01835907e+00 -4.14100319e-01
-5.65896273e-01 1.24768054e+00 4.24806535e-01 -9.05770123e-01
9.48950350e-01 -2.70542502e-01 -6.09449089e-01 1.54383361e-01
6.70276403e-01 1.14619207e+00 -1.75069347e-02 -1.14006746e+00
-1.28205097e+00 -3.33628297e-01 -2.27803513e-02 1.09990403e-01
2.93372840e-01 1.00481838e-01 -8.51406693e-01 -8.48868430e-01
3.74898404e-01 -6.53490484e-01 -2.70281523e-01 -3.35701287e-01
-4.73206669e-01 -9.97438356e-02 6.96886063e-01 -1.16410732e+00
1.13437998e+00 -1.76356423e+00 -3.32146585e-01 8.41508329e-01
1.46850869e-01 -1.17312573e-01 1.68719232e-01 3.72670889e-01
1.17652901e-01 1.32619545e-01 -7.00117290e-01 -1.58723891e-01
-6.19791448e-02 -3.09612423e-01 -4.99589324e-01 6.90581322e-01
-3.21334712e-02 9.49895203e-01 -7.60172784e-01 -5.12187362e-01
4.51943666e-01 1.91038907e-01 -2.46504825e-02 -3.35220993e-01
-8.93915966e-02 -1.30253360e-02 1.04822973e-02 7.97116876e-01
1.11755407e+00 2.90980726e-01 -4.78591263e-01 -2.94703484e-01
-7.02474535e-01 -4.17665206e-02 -1.14828026e+00 1.63428605e+00
-3.92738357e-02 1.06319571e+00 1.28666818e-01 -6.28363311e-01
1.11596990e+00 -2.03261480e-01 6.18385673e-01 -1.10478532e+00
-5.79674132e-02 2.41794094e-01 -3.78249377e-01 -1.53331667e-01
7.84750342e-01 2.06559300e-01 -3.70124578e-01 1.80161223e-02
4.95476797e-02 -6.01458669e-01 5.97807288e-01 3.26687306e-01
1.28402317e+00 2.92492330e-01 -2.25386277e-01 -5.48930407e-01
3.25214684e-01 6.62681997e-01 5.33427477e-01 8.75167072e-01
-1.21520482e-01 1.07738531e+00 3.08471501e-01 -1.04158831e+00
-1.43114471e+00 -1.31215537e+00 -5.97639024e-01 8.46409738e-01
2.85476804e-01 -2.99501270e-01 -6.96949720e-01 -4.85338420e-01
-5.94405122e-02 8.17231417e-01 -7.95402825e-01 5.81652105e-01
-9.94230509e-01 -9.72570002e-01 1.12327993e+00 6.46469712e-01
9.69420433e-01 -1.03472364e+00 -5.17253041e-01 2.95672894e-01
-1.46529555e-01 -9.80497420e-01 -2.96807289e-01 2.59591639e-01
-4.33275074e-01 -9.75630581e-01 -1.05951679e+00 -8.92037153e-01
5.29923975e-01 -5.79173677e-02 1.17713213e+00 -1.67294201e-02
-3.91042322e-01 2.83392016e-02 -1.99027568e-01 -5.60309708e-01
9.97126698e-02 5.67639053e-01 -4.17203039e-01 -4.38740432e-01
1.57521412e-01 -4.71893489e-01 -6.68007731e-01 4.95071352e-01
-5.04276156e-01 2.87279606e-01 3.75360221e-01 2.45468825e-01
5.38648665e-01 -6.53654709e-02 3.13143164e-01 -6.58942282e-01
5.33502959e-02 -3.80562723e-01 -8.72309864e-01 2.57694840e-01
1.19483974e-02 -4.32623655e-01 2.30803713e-01 3.40113848e-01
-7.72833109e-01 2.71169662e-01 -7.11609125e-01 2.89425343e-01
-2.60202050e-01 1.21418267e-01 -5.53054959e-02 -1.57977924e-01
8.41010213e-01 9.80530456e-02 -6.91762388e-01 -3.75179172e-01
3.96975875e-01 5.14239132e-01 1.45636106e+00 -4.26886171e-01
1.27817583e+00 8.42561007e-01 -2.20021188e-01 -9.13038671e-01
-7.46658623e-01 -6.47961438e-01 -8.79578173e-01 -5.50358713e-01
1.07893384e+00 -9.51100469e-01 -5.16036712e-02 6.41896725e-01
-9.36477602e-01 -5.45162857e-01 -1.54780611e-01 2.72973865e-01
-4.26891357e-01 -1.72392204e-01 -6.88864231e-01 -4.66032326e-01
-5.39600730e-01 -6.20164812e-01 1.03161430e+00 5.35485566e-01
-8.25458840e-02 -7.60053933e-01 4.15298969e-01 2.85254270e-01
3.00409615e-01 7.63075650e-01 4.97358173e-01 -5.98259449e-01
-4.56567258e-01 -4.29295450e-01 -4.98194814e-01 -2.19237790e-01
-2.60415286e-01 -5.24917711e-03 -1.20957637e+00 8.55257660e-02
-7.19643831e-01 9.18196216e-02 1.23820531e+00 6.71466231e-01
1.00884366e+00 1.66649997e-01 -3.40393782e-01 8.88555408e-01
1.21059108e+00 2.44725466e-01 9.26021457e-01 9.14918303e-01
7.18896091e-01 5.80005407e-01 3.88881862e-01 1.18464619e-01
7.03536808e-01 6.72119856e-01 4.12966192e-01 -5.56038260e-01
-1.63870916e-01 -7.68124610e-02 -2.22204968e-01 6.14366643e-02
-4.64983016e-01 -3.39979609e-03 -1.58026826e+00 8.87306273e-01
-1.70810497e+00 -9.67165530e-01 -5.59832394e-01 2.02961493e+00
4.04616356e-01 5.73452711e-01 5.44912100e-01 -5.40693216e-02
8.67723286e-01 3.18013012e-01 -2.89493114e-01 -1.57579053e-02
-5.15045881e-01 1.39936700e-01 1.24562764e+00 5.77240348e-01
-1.69635785e+00 1.11908603e+00 5.75049591e+00 1.07744503e+00
-8.94285917e-01 2.33586263e-02 1.00831938e+00 1.50737225e-03
-1.72690377e-01 -1.13376997e-01 -8.13951790e-01 4.90931898e-01
6.26826346e-01 1.01348117e-01 1.78744234e-02 6.59966528e-01
-4.25811708e-02 -4.16347384e-01 -4.49558765e-01 7.54694939e-01
-1.30254729e-02 -2.05739141e+00 -4.40035909e-01 -1.16906330e-01
6.97808683e-01 8.83319616e-01 -4.96181369e-01 2.89254218e-01
3.16650152e-01 -1.00527000e+00 1.19405925e+00 6.14195347e-01
7.91675508e-01 -1.13245118e+00 9.33381259e-01 7.00377598e-02
-1.57847929e+00 1.08262360e-01 -1.84720501e-01 1.16228610e-01
5.28304517e-01 3.95221949e-01 -8.32257688e-01 7.27726758e-01
1.14572847e+00 3.93242210e-01 -9.79355395e-01 1.69140005e+00
-2.10807696e-01 5.48172295e-01 -7.98256278e-01 4.56293285e-01
5.06920099e-01 -1.81354652e-03 6.00669026e-01 1.71732974e+00
9.54537764e-02 6.90376684e-02 3.13984044e-02 7.63199568e-01
-1.74890861e-01 -2.10189208e-01 -3.65784764e-01 8.00462186e-01
5.45571506e-01 1.49566388e+00 -1.56995833e+00 -2.08298385e-01
2.81600356e-02 6.66460335e-01 2.03360822e-02 5.99868596e-02
-1.00181854e+00 -7.35481858e-01 1.35114774e-01 5.81857145e-01
1.29020810e-01 -1.76969439e-01 -7.38008857e-01 -5.35918772e-01
-3.77713963e-02 -9.70379189e-02 6.09441698e-01 -7.51408875e-01
-9.67150927e-01 6.88060045e-01 3.50875556e-01 -1.02662432e+00
1.73485577e-01 -9.89674255e-02 -1.11558402e+00 7.17823625e-01
-1.08594191e+00 -1.52076638e+00 -5.21261692e-01 2.19742939e-01
5.41565537e-01 4.65791598e-02 3.09053451e-01 6.23973906e-01
-4.59561229e-01 3.63176823e-01 4.81301844e-02 6.42102778e-01
3.95358473e-01 -1.63038146e+00 1.25951850e+00 9.41507459e-01
5.11834286e-02 -5.63028753e-02 6.30027950e-01 -9.12715018e-01
-4.88460690e-01 -1.30816925e+00 7.32053757e-01 -4.24808025e-01
5.97586989e-01 -8.08817089e-01 -8.69870842e-01 7.00663805e-01
1.39113562e-03 -1.47405416e-01 1.46069946e-02 -2.93192446e-01
4.60593477e-02 -1.90872103e-02 -1.25189257e+00 3.05719852e-01
9.60231841e-01 -4.06811029e-01 -5.38216770e-01 3.48109215e-01
2.88711548e-01 -8.25828314e-01 -8.75479758e-01 5.08809268e-01
4.61928219e-01 -6.43168807e-01 1.09940565e+00 3.05503219e-01
3.60271662e-01 -5.93108773e-01 -3.76378521e-02 -1.12716711e+00
-2.30792925e-01 -4.21867728e-01 4.98934597e-01 1.23786819e+00
6.15065217e-01 -2.20428199e-01 9.86855268e-01 3.40959042e-01
-6.25891626e-01 -4.80628043e-01 -1.14025688e+00 -7.16531634e-01
2.07028329e-01 -7.84205616e-01 7.11322546e-01 6.91345692e-01
-4.87487346e-01 -3.49122062e-02 3.18770260e-02 4.14810121e-01
7.93274403e-01 -2.35944182e-01 7.23253727e-01 -1.25577056e+00
4.66263711e-01 -8.18617105e-01 -6.36568010e-01 -7.74292707e-01
-4.29226488e-01 -9.29533780e-01 9.32453796e-02 -1.89990914e+00
-4.26600963e-01 -8.18222463e-01 1.75628394e-01 4.49019521e-01
7.82808810e-02 8.71131480e-01 1.12784319e-01 5.01609683e-01
-4.65164840e-01 1.93352863e-01 4.49540645e-01 -2.24612981e-01
-2.36407980e-01 1.41141564e-01 -1.49642229e-01 9.55457151e-01
8.15999746e-01 -3.82721245e-01 2.10045323e-01 -3.48607123e-01
3.73268634e-01 -3.72213960e-01 6.53454542e-01 -1.47644067e+00
4.16602284e-01 3.82209122e-01 8.36238205e-01 -1.57282388e+00
2.08723903e-01 -5.23479879e-01 1.85374603e-01 3.18241924e-01
4.01731692e-02 1.75531626e-01 5.15095174e-01 9.58306715e-02
-7.52913728e-02 -2.03734085e-01 7.21531212e-01 -1.76879406e-01
-1.33929205e+00 1.98854044e-01 -1.57047331e-01 8.14000294e-02
1.06726849e+00 -4.49851125e-01 -5.79538643e-01 -1.92914575e-01
-7.98040092e-01 8.96644413e-01 4.28677261e-01 3.02620858e-01
4.06092703e-01 -1.16147256e+00 -1.02645957e+00 2.04783268e-02
3.03685647e-02 4.40423191e-01 3.43411267e-01 6.25089049e-01
-1.38496268e+00 9.79935899e-02 -3.50196868e-01 -6.75574481e-01
-9.13087606e-01 -2.11477354e-01 5.27608931e-01 -9.04611722e-02
-1.05512357e+00 1.12780178e+00 -5.55825308e-02 -6.80887222e-01
2.26541087e-01 -1.10467650e-01 -4.81413931e-01 5.21736979e-01
6.46891475e-01 7.42286742e-01 3.58247906e-01 -7.50195742e-01
-5.96427977e-01 7.50366569e-01 2.64037997e-02 -5.29474080e-01
1.65388536e+00 4.74984534e-02 1.26023725e-01 1.23837188e-01
8.14835668e-01 -1.87904894e-01 -1.31606531e+00 1.10117765e-03
4.49604690e-01 -9.25626978e-02 2.44276688e-01 -9.70092833e-01
-1.15898848e+00 4.22911197e-01 7.97730446e-01 2.25798771e-01
8.06936324e-01 2.23313749e-01 8.85816753e-01 1.13293201e-01
3.83829057e-01 -1.37986565e+00 -7.95550942e-01 5.66158116e-01
1.17201149e+00 -1.10631621e+00 1.82108462e-01 -1.57823920e-01
-4.42934811e-01 1.28216290e+00 6.95887625e-01 -3.76266897e-01
6.21822655e-01 4.92235988e-01 -4.53136191e-02 -6.63017690e-01
2.65397966e-01 -2.99028426e-01 4.55788195e-01 4.23846632e-01
3.90098983e-04 1.79519624e-01 -6.15921915e-02 2.64480650e-01
-8.39402616e-01 -2.08410949e-01 3.39081168e-01 1.02123642e+00
-6.48232937e-01 -4.67618018e-01 -7.37029076e-01 6.58047259e-01
-1.77182704e-01 -6.80066720e-02 -4.42419857e-01 1.04991961e+00
6.15059972e-01 5.54714024e-01 7.42477119e-01 -4.06479120e-01
5.22612691e-01 -4.76490736e-01 1.78056121e-01 -2.65530229e-01
-8.17905545e-01 3.28530036e-02 4.14510936e-01 -4.38907236e-01
-5.46582155e-02 -6.58246040e-01 -1.57233822e+00 -4.42851663e-01
1.01909034e-01 1.87612578e-01 7.72581935e-01 7.89225936e-01
-1.33535445e-01 2.49241471e-01 3.91140908e-01 -1.34132969e+00
4.48025703e-01 -9.84336436e-01 -4.87268507e-01 -6.84085488e-02
1.61922738e-01 -3.89163405e-01 -2.71541536e-01 -1.46477103e-01] | [8.707534790039062, -1.6334044933319092] |
5fca6b5f-4170-4339-ba6c-cde9956c8c0f | constructing-large-proposition-databases | null | null | https://aclanthology.org/L12-1238 | https://aclanthology.org/L12-1238.pdf | Constructing Large Proposition Databases | With the advent of massive online encyclopedic corpora such as Wikipedia, it has become possible to apply a systematic analysis to a wide range of documents covering a significant part of human knowledge. Using semantic parsers, it has become possible to extract such knowledge in the form of propositions (predicate―argument structures) and build large proposition databases from these documents. This paper describes the creation of multilingual proposition databases using generic semantic dependency parsing. Using Wikipedia, we extracted, processed, clustered, and evaluated a large number of propositions. We built an architecture to provide a complete pipeline dealing with the input of text, extraction of knowledge, storage, and presentation of the resulting propositions. | ['Peter Exner', 'Pierre Nugues'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [-1.91155672e-01 6.17427528e-01 1.73385032e-02 -1.60533935e-01
-8.57131183e-01 -9.08958554e-01 9.61347938e-01 9.95939374e-01
-6.08915031e-01 1.06237459e+00 4.38460529e-01 -2.30578154e-01
-2.36707389e-01 -1.10283959e+00 -6.98022485e-01 1.23850904e-01
2.23240882e-01 5.82279265e-01 8.16742539e-01 -3.21029216e-01
3.49719226e-01 2.79419243e-01 -1.65250564e+00 5.25382280e-01
6.05270386e-01 8.79918218e-01 4.45885479e-01 1.83179855e-01
-1.09496260e+00 8.42085719e-01 -6.78207397e-01 -9.88996446e-01
-2.42826492e-01 -1.94488510e-01 -1.19576383e+00 -1.51695669e-01
3.54448184e-02 2.65567541e-01 1.93835989e-01 1.11360097e+00
1.07725054e-01 -1.89426750e-01 1.59414396e-01 -8.62784564e-01
-4.52778995e-01 1.25027013e+00 2.10658833e-01 -1.56388164e-01
8.50843668e-01 -6.48015320e-01 9.92511153e-01 -5.94649076e-01
1.45968890e+00 1.20065689e+00 6.01871550e-01 4.26688744e-03
-5.68283200e-01 -9.09957737e-02 -3.03716213e-01 2.71960199e-01
-1.00088120e+00 -2.04780355e-01 5.72255552e-01 -3.09097260e-01
1.40110707e+00 3.18054259e-02 8.23415935e-01 7.27637768e-01
5.55426404e-02 5.64593554e-01 9.61649597e-01 -1.08286202e+00
8.61079693e-02 5.10908842e-01 5.54223657e-01 5.72311640e-01
4.65526760e-01 -5.03106296e-01 -5.93974769e-01 -1.44833505e-01
2.79614329e-01 -7.06205726e-01 8.94040763e-02 -1.45579994e-01
-9.40100610e-01 4.26576048e-01 -1.25728756e-01 5.54324150e-01
-6.40024066e-01 -8.05469379e-02 9.67953026e-01 -9.26784053e-02
1.21324591e-01 4.63356316e-01 -5.38445950e-01 -3.75286996e-01
-4.24290746e-01 2.71586299e-01 1.05666852e+00 1.00496268e+00
7.25757420e-01 -6.02732778e-01 3.05600554e-01 6.65174186e-01
1.74465522e-01 3.03425401e-01 5.54793656e-01 -1.05819154e+00
8.66030931e-01 9.72884893e-01 3.41654390e-01 -1.08005714e+00
-4.29230541e-01 5.61752319e-02 5.65545447e-02 -4.09344584e-01
2.64989495e-01 8.86678919e-02 -6.64545596e-01 1.42037237e+00
6.00235522e-01 -6.52485251e-01 5.91167450e-01 3.59615713e-01
1.04327786e+00 4.33894128e-01 3.73126775e-01 -1.62615463e-01
1.57554173e+00 -3.81381273e-01 -9.12510872e-01 -1.30838426e-02
5.62101841e-01 -7.92478323e-01 6.49778903e-01 1.59319952e-01
-1.09047294e+00 -3.15630943e-01 -8.17982852e-01 -3.12488317e-01
-1.22003722e+00 -1.64366350e-01 8.77089143e-01 5.81745744e-01
-8.60775948e-01 2.82675445e-01 -6.37881517e-01 -6.12721682e-01
2.95130938e-01 -5.97089268e-02 -7.35981941e-01 -2.07466617e-01
-1.54676938e+00 1.21730089e+00 1.35550308e+00 -1.24037713e-01
-1.32421777e-01 -3.18624526e-01 -9.33912635e-01 -9.45314486e-03
5.30677855e-01 -6.89361811e-01 6.64209247e-01 -7.34740973e-01
-1.12282133e+00 9.20200408e-01 6.53335229e-02 -5.70006251e-01
1.62273422e-02 -2.38185108e-01 -6.82292998e-01 4.85624403e-01
5.33283412e-01 3.70756119e-01 3.65667522e-01 -8.84045005e-01
-9.03629899e-01 -4.55937922e-01 2.07900092e-01 2.10274786e-01
-2.94555932e-01 5.91587305e-01 -7.68951476e-01 -8.62744674e-02
-2.52414551e-02 -3.79711419e-01 1.48350999e-01 -6.35131240e-01
-3.05617243e-01 -2.62291312e-01 3.05514008e-01 -1.07572770e+00
1.21193182e+00 -2.06189442e+00 2.66137198e-02 2.45297909e-01
-1.06551582e-02 1.51998237e-01 2.25889236e-01 8.84509683e-01
2.91196644e-01 2.88807124e-01 -7.27341697e-02 1.78480297e-01
2.65083760e-01 5.28372407e-01 -2.46554255e-01 -3.40396881e-01
-9.68425721e-02 7.15246916e-01 -8.93933892e-01 -9.93908226e-01
1.40599295e-01 2.49355644e-01 -1.48190066e-01 -2.40451261e-01
-6.16812646e-01 -3.77602667e-01 -7.43764222e-01 4.76116568e-01
2.12610498e-01 3.64677645e-02 7.80286610e-01 -1.19926505e-01
-3.63560110e-01 6.37064397e-01 -1.01335299e+00 1.70135522e+00
-3.38948786e-01 5.23201466e-01 -1.80351838e-01 -8.13411832e-01
8.68770838e-01 4.73220825e-01 3.46381426e-01 -5.51568985e-01
6.33145347e-02 2.68459886e-01 -4.24662948e-01 -6.02913022e-01
1.02198589e+00 -4.53252010e-02 -6.13173187e-01 9.36623365e-02
3.29349488e-01 -1.39501467e-01 8.84721518e-01 5.00982046e-01
1.11971819e+00 5.03402352e-01 6.94861472e-01 -1.59895688e-01
6.97273791e-01 8.54616463e-01 3.10693443e-01 -4.24896479e-02
2.57219613e-01 -1.48849353e-01 7.67305911e-01 -3.73129845e-01
-1.16315675e+00 -1.03758192e+00 -2.90094346e-01 4.89721864e-01
-1.54420167e-01 -8.71453762e-01 -7.08424270e-01 -3.95131588e-01
-4.94030975e-02 7.75685489e-01 -3.14239919e-01 4.33937669e-01
-4.92226869e-01 -4.75162119e-01 6.23194814e-01 3.65921617e-01
6.93283498e-01 -1.16618609e+00 -6.70595229e-01 4.44083065e-01
-2.19735354e-01 -1.51475859e+00 7.38270938e-01 1.65362991e-02
-4.25295889e-01 -1.29018569e+00 1.79706007e-01 -7.53618538e-01
3.45667630e-01 -5.10225296e-01 1.10983121e+00 -3.47390473e-01
-1.87033579e-01 3.89515728e-01 -6.96077824e-01 -6.38668776e-01
-6.78425312e-01 -1.04206003e-01 -3.09072077e-01 -7.00086594e-01
3.89160395e-01 -1.09111510e-01 3.39245468e-01 -3.67197335e-01
-8.87462974e-01 -4.49522324e-02 3.99448425e-01 3.65113556e-01
7.53970921e-01 5.65143764e-01 6.82022035e-01 -9.96933401e-01
9.28234398e-01 -5.38602293e-01 -6.39102459e-01 7.98908412e-01
-2.78618962e-01 1.78505436e-01 7.84295201e-01 3.79412055e-01
-1.38812041e+00 -1.44765764e-01 -3.42898518e-01 3.68839711e-01
-1.77861810e-01 1.15977657e+00 -4.41757649e-01 5.38342655e-01
3.77904594e-01 1.33855358e-01 -9.79752764e-02 -4.00188625e-01
7.40153015e-01 7.09927201e-01 7.15357602e-01 -8.24086547e-01
2.81864792e-01 2.39642262e-01 1.06203511e-01 -9.55297947e-01
-7.20020652e-01 -4.50093925e-01 -7.19009161e-01 -2.42119282e-01
8.69569302e-01 -8.54587913e-01 -2.07178950e-01 -2.96993125e-02
-1.03149271e+00 7.03599229e-02 -4.97028798e-01 3.83076936e-01
-3.43897045e-01 5.75439453e-01 -4.08452630e-01 -3.76537383e-01
-3.46829265e-01 -4.07614321e-01 3.72319609e-01 3.16884816e-02
-3.73599917e-01 -1.14755857e+00 7.14523494e-02 4.53362614e-01
-1.13112547e-01 3.53648752e-01 1.12639558e+00 -8.36886227e-01
-4.80764300e-01 -4.59826350e-01 -1.83396995e-01 3.33699554e-01
-7.71097094e-02 2.00422764e-01 -2.25361645e-01 5.24364948e-01
-3.21479291e-01 -2.43778199e-01 3.64332050e-01 -1.59887865e-01
4.39065307e-01 -5.63459620e-02 -4.19032216e-01 -5.29563166e-02
1.53258038e+00 3.89142744e-02 8.51926923e-01 8.35444570e-01
2.73109555e-01 6.77081108e-01 8.58808339e-01 8.03282261e-02
7.83856690e-01 3.86816204e-01 -2.49920581e-02 9.34415936e-01
-2.53729910e-01 -3.28285217e-01 -1.01440037e-02 8.92406762e-01
-1.14381231e-01 7.68071413e-02 -1.23238957e+00 8.44966292e-01
-1.66583538e+00 -9.62623179e-01 -8.66280049e-02 1.66637611e+00
8.27909887e-01 3.38118076e-01 -5.49152270e-02 9.90924388e-02
5.59551775e-01 -2.21358284e-01 2.16572911e-01 -5.21811247e-01
-1.46406502e-01 2.25313619e-01 3.21718246e-01 3.87674749e-01
-7.68249989e-01 1.26476049e+00 6.50646496e+00 5.79592288e-01
-6.38848841e-01 1.77489191e-01 -3.29078823e-01 3.93661290e-01
-4.42157984e-01 1.62628263e-01 -1.06175435e+00 4.13344145e-01
1.23912549e+00 -4.55116451e-01 1.15471957e-02 8.04111958e-01
-1.89894065e-01 -6.87081397e-01 -5.24460733e-01 5.35164654e-01
1.70836374e-01 -1.69063413e+00 1.06613778e-01 -3.43631595e-01
3.71229738e-01 -4.09063743e-03 -6.46082878e-01 2.33692974e-01
5.13876140e-01 -4.29173172e-01 9.44116652e-01 5.72152138e-01
5.02828240e-01 -6.58210993e-01 8.79902840e-01 1.37630269e-01
-1.03630936e+00 3.00500076e-02 -4.52082038e-01 2.85379719e-02
2.90396303e-01 5.93449295e-01 -8.25683355e-01 1.14966929e+00
5.75832546e-01 5.56211412e-01 -5.54137766e-01 7.29840159e-01
-7.07203507e-01 1.09866589e-01 -4.77055550e-01 -3.42903018e-01
5.66017590e-02 -1.25822783e-01 1.69971526e-01 1.23820055e+00
3.18541259e-01 1.49510652e-01 -6.75109029e-02 3.31441909e-01
-1.77942246e-01 6.14259481e-01 -3.30832660e-01 -4.44939613e-01
6.93022072e-01 1.03193557e+00 -9.05517519e-01 -6.83632672e-01
-6.38543546e-01 3.92570347e-01 5.94068170e-01 1.15525238e-02
-4.49333638e-01 -6.28108382e-01 2.10138619e-01 -5.07251211e-02
3.65618765e-01 -2.39335045e-01 -1.21385947e-01 -9.53860104e-01
4.37561750e-01 -5.36819875e-01 6.44506752e-01 -9.77300465e-01
-7.80274808e-01 6.69512630e-01 3.91523540e-01 -4.54198420e-01
-4.65530396e-01 -6.25486612e-01 -6.32092431e-02 7.93053687e-01
-1.21124136e+00 -1.20694387e+00 9.49446186e-02 4.25235301e-01
-1.09309711e-01 -1.50905117e-01 1.04238510e+00 1.90719962e-01
-3.01731229e-01 -4.31626797e-01 -1.15946867e-01 3.35181803e-01
4.24594462e-01 -1.07621813e+00 4.04087417e-02 9.17975962e-01
2.31796935e-01 7.70066321e-01 6.43220782e-01 -1.00173616e+00
-1.27014804e+00 -6.84281945e-01 1.43458796e+00 -6.12450421e-01
1.06201112e+00 1.21839300e-01 -6.76586211e-01 8.98396194e-01
4.04180706e-01 -4.28522795e-01 7.24665880e-01 2.46534541e-01
-3.77071291e-01 -1.26197025e-01 -9.73554552e-01 1.75154701e-01
6.91203594e-01 -7.65318930e-01 -1.44524550e+00 2.33056769e-01
5.58615208e-01 -3.43656480e-01 -1.24305332e+00 -8.69208127e-02
3.57540876e-01 -6.10284388e-01 9.05810893e-01 -3.59729826e-01
5.27069092e-01 -2.96527416e-01 -1.60754293e-01 -9.99390543e-01
2.56381661e-01 -1.23258121e-01 -7.76809230e-02 1.35842443e+00
6.87331498e-01 -5.84706783e-01 4.05407131e-01 7.99081147e-01
-4.99476433e-01 -1.48798630e-01 -8.12065125e-01 -5.88836253e-01
-1.24213964e-01 -7.35143185e-01 6.14302337e-01 8.93385291e-01
9.50684369e-01 3.32755476e-01 3.20494950e-01 1.25827089e-01
4.12983030e-01 3.58993083e-01 5.13015687e-01 -1.19921231e+00
1.60729557e-01 -1.52771220e-01 -6.61170065e-01 -2.00661346e-01
3.73845398e-01 -1.14204478e+00 -2.49296159e-01 -2.36544657e+00
8.10765196e-03 -2.77971536e-01 9.13097560e-02 5.57356119e-01
8.98141712e-02 -2.88278937e-01 -2.70134956e-02 7.49999285e-02
-7.58533418e-01 2.46690083e-02 7.71396279e-01 4.48433518e-01
-4.51573282e-02 -6.92868769e-01 -6.31642342e-01 7.92499185e-01
7.62703478e-01 -3.58170599e-01 -1.23029582e-01 -3.79642159e-01
9.70317543e-01 -5.88583685e-02 1.84584782e-01 -1.02366745e+00
2.67828405e-01 -6.51384797e-03 2.28998229e-01 -7.04721093e-01
2.70850033e-01 -7.94104338e-01 3.61126006e-01 1.81751743e-01
1.32903727e-02 -5.79870753e-02 3.36041063e-01 2.49897495e-01
-6.37837410e-01 -6.54922068e-01 1.27654284e-01 -4.84628141e-01
-1.37592304e+00 -4.81704384e-01 -2.46982902e-01 1.69654399e-01
1.17154467e+00 2.19715722e-02 -6.52178407e-01 3.44702244e-01
-8.26182365e-01 8.78323987e-02 5.09690940e-01 1.71272129e-01
4.11504626e-01 -9.68384445e-01 -1.56804025e-01 -2.68938750e-01
2.82484898e-03 -2.28898183e-01 -1.94228105e-02 3.31435412e-01
-9.54199076e-01 8.49307001e-01 -6.85704648e-01 1.70297399e-01
-1.11684716e+00 7.67635524e-01 -3.21939945e-01 -4.04502690e-01
-6.80230319e-01 3.44725072e-01 -7.96617508e-01 -2.92781889e-01
-2.41670474e-01 -1.40058279e-01 -5.13966322e-01 4.52778637e-01
4.11165178e-01 7.18330741e-02 -3.10141360e-03 -7.67374456e-01
-4.55792755e-01 3.90448809e-01 3.82138073e-01 -2.83220708e-01
1.43209136e+00 -1.84976086e-01 -7.00645566e-01 2.70864308e-01
7.09965706e-01 3.84347498e-01 -4.86558229e-01 7.41114989e-02
6.19401693e-01 -1.57615185e-01 -2.74084777e-01 -7.39187062e-01
-2.93354303e-01 3.22639227e-01 -3.70228976e-01 5.31987309e-01
9.04577017e-01 4.36540693e-01 5.77068806e-01 5.91219008e-01
6.65546119e-01 -1.61228383e+00 -4.04503196e-01 6.73826337e-01
6.12171531e-01 -8.12375784e-01 2.48904020e-01 -8.05250943e-01
-5.79198301e-01 1.28849268e+00 -4.39717695e-02 2.28939310e-01
5.81914604e-01 3.22465330e-01 -5.67910336e-02 -4.37017620e-01
-6.09032035e-01 -5.19243658e-01 -7.43368715e-02 6.73270106e-01
2.50279844e-01 2.31854513e-01 -8.52577448e-01 8.09567690e-01
-4.63320106e-01 2.34426141e-01 6.04958236e-01 1.10751820e+00
-6.77840471e-01 -1.46919584e+00 -3.56894612e-01 2.28695899e-01
-6.61836624e-01 -1.41834125e-01 -4.56120253e-01 1.24293923e+00
2.31425822e-01 6.97780430e-01 1.48678310e-02 1.47619680e-01
4.10346538e-01 6.75511777e-01 5.66367567e-01 -7.70881653e-01
-4.19952571e-01 -2.37780482e-01 1.16115189e+00 -2.95780748e-01
-9.57309365e-01 -8.19590807e-01 -1.75532985e+00 -5.72888814e-02
1.24605894e-01 5.82764149e-01 1.17639112e+00 1.24888098e+00
1.77491784e-01 5.28417408e-01 -3.28423381e-01 5.64894900e-02
-3.23636383e-02 -6.79415584e-01 -4.09800917e-01 3.20796967e-01
-7.14242697e-01 -4.03969526e-01 1.21373229e-01 3.31383139e-01] | [9.378386497497559, 8.464118003845215] |
1b36321f-17ad-4a9f-8838-02f9b7fa647c | deep-multimodal-learning-for-audio-visual | 1501.05396 | null | http://arxiv.org/abs/1501.05396v1 | http://arxiv.org/pdf/1501.05396v1.pdf | Deep Multimodal Learning for Audio-Visual Speech Recognition | In this paper, we present methods in deep multimodal learning for fusing
speech and visual modalities for Audio-Visual Automatic Speech Recognition
(AV-ASR). First, we study an approach where uni-modal deep networks are trained
separately and their final hidden layers fused to obtain a joint feature space
in which another deep network is built. While the audio network alone achieves
a phone error rate (PER) of $41\%$ under clean condition on the IBM large
vocabulary audio-visual studio dataset, this fusion model achieves a PER of
$35.83\%$ demonstrating the tremendous value of the visual channel in phone
classification even in audio with high signal to noise ratio. Second, we
present a new deep network architecture that uses a bilinear softmax layer to
account for class specific correlations between modalities. We show that
combining the posteriors from the bilinear networks with those from the fused
model mentioned above results in a further significant phone error rate
reduction, yielding a final PER of $34.03\%$. | ['Youssef Mroueh', 'Etienne Marcheret', 'Vaibhava Goel'] | 2015-01-22 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 1.91242591e-01 1.13619283e-01 3.18412930e-01 -5.21758378e-01
-1.69924104e+00 -3.12684387e-01 4.56787795e-01 -1.46626845e-01
-4.37026381e-01 6.00857258e-01 2.24002182e-01 -2.88777918e-01
4.72358942e-01 -3.27856839e-01 -9.22865093e-01 -7.65610278e-01
-1.26464535e-02 7.05028921e-02 -2.63613522e-01 9.34454501e-02
-4.62097079e-01 3.95013571e-01 -1.79367197e+00 7.44587183e-01
2.84420788e-01 1.58298552e+00 -1.06073134e-01 1.16163850e+00
7.85986781e-02 8.31038654e-01 -8.05567563e-01 -3.80185068e-01
1.07553035e-01 -4.46056016e-02 -6.24597013e-01 6.98463544e-02
5.62964559e-01 -4.59258556e-01 -6.00361705e-01 8.10760796e-01
8.40983391e-01 3.34968954e-01 6.02164388e-01 -1.29408813e+00
-4.51750427e-01 6.31016910e-01 -3.89244765e-01 -2.48645976e-01
3.72619212e-01 2.08095431e-01 1.13095498e+00 -1.10878074e+00
4.03299630e-02 1.33807254e+00 6.37892544e-01 3.88004333e-01
-1.33783853e+00 -7.63078272e-01 -6.68905303e-02 1.19233668e-01
-1.78550124e+00 -1.00894856e+00 6.75980389e-01 -3.99675548e-01
1.55943692e+00 2.33526006e-01 2.50737906e-01 1.24343014e+00
-2.35248446e-01 7.69932985e-01 8.75273168e-01 -4.06591684e-01
-1.00481279e-01 5.29129505e-01 -1.92602620e-01 5.65901875e-01
-5.31618237e-01 6.36351407e-02 -6.96898639e-01 -4.67781052e-02
4.14828122e-01 -2.94988871e-01 -1.92423075e-01 1.95199147e-01
-9.83729362e-01 7.85422504e-01 3.69527221e-01 2.65527070e-01
-2.88766593e-01 4.29041177e-01 3.67719680e-01 1.87603742e-01
2.91149974e-01 -9.89908129e-02 -3.19572121e-01 -1.27194420e-01
-1.20610797e+00 -1.64272010e-01 5.42853594e-01 7.00668931e-01
6.26403034e-01 5.87493241e-01 -1.58703193e-01 1.29312551e+00
7.81038105e-01 6.21809125e-01 2.33818129e-01 -1.01855600e+00
5.16123772e-01 1.70107130e-02 -1.18676782e-01 -5.84794641e-01
-1.86953634e-01 -3.67538005e-01 -7.85091460e-01 2.68024176e-01
2.12606549e-01 -2.82448351e-01 -1.21709669e+00 1.85806870e+00
-1.50691077e-01 -1.75454635e-02 3.63723367e-01 6.16233647e-01
1.16411102e+00 1.08088481e+00 1.38365909e-01 -9.10832584e-02
1.24280560e+00 -7.69393444e-01 -7.32048213e-01 -1.31674722e-01
2.56311268e-01 -9.07460392e-01 7.16104627e-01 4.73912716e-01
-1.41105652e+00 -6.95511341e-01 -1.27872539e+00 -2.64830083e-01
-3.45486403e-01 3.99066091e-01 1.14268675e-01 6.85182333e-01
-1.44982946e+00 1.48971289e-01 -6.42632484e-01 -7.27156317e-03
3.54963005e-01 6.32879198e-01 -4.23535526e-01 8.29108059e-02
-1.16922915e+00 7.97718883e-01 2.45103970e-01 1.04236014e-01
-1.30334020e+00 -4.44410950e-01 -1.07154286e+00 3.57389450e-01
3.36695500e-02 -4.15038675e-01 1.39219081e+00 -9.88098741e-01
-1.65010262e+00 6.49717987e-01 -4.74240690e-01 -5.18627524e-01
2.32686326e-01 9.85783041e-02 -6.67275965e-01 2.17477486e-01
-3.78884852e-01 1.23325717e+00 1.07781792e+00 -1.45988882e+00
-7.03443646e-01 -7.35630170e-02 -1.94463223e-01 1.64862797e-01
-3.22773725e-01 1.86139550e-02 -6.02349877e-01 -5.75570822e-01
-1.52291104e-01 -7.77286768e-01 1.78198427e-01 -2.43982509e-01
-3.40227902e-01 -9.35364142e-02 5.56642711e-01 -1.15981781e+00
1.01792347e+00 -2.36866021e+00 3.65582928e-02 3.51544559e-01
2.13108435e-01 2.49469548e-01 -2.97928035e-01 1.13529049e-01
-2.12248966e-01 7.83325583e-02 -1.59152642e-01 -9.61378634e-01
1.99498191e-01 9.98098552e-02 -1.47659510e-01 2.02974454e-01
2.61349618e-01 7.38657653e-01 -6.91875368e-02 -2.50475109e-01
3.58239740e-01 1.26035297e+00 -4.78605747e-01 2.35213697e-01
1.27827257e-01 9.62266326e-02 4.23713297e-01 8.17765057e-01
6.14910305e-01 -1.54176295e-01 3.48634236e-02 -4.39487338e-01
2.44871050e-01 3.74933660e-01 -1.13592684e+00 1.55140209e+00
-6.49108887e-01 1.00918686e+00 7.53209233e-01 -8.95539463e-01
7.71936059e-01 8.44439328e-01 3.30770522e-01 -7.00681567e-01
3.09220344e-01 2.89361656e-01 -1.98808864e-01 -1.26831517e-01
5.82880318e-01 -4.67204034e-01 -1.48057416e-01 8.89529586e-02
6.89918995e-01 -7.24687651e-02 -2.90914923e-01 1.18943021e-01
8.03069711e-01 -4.66401488e-01 -4.01405305e-01 3.58867645e-01
5.22297263e-01 -5.69866538e-01 2.12178871e-01 6.84241772e-01
-3.49008322e-01 7.80754387e-01 2.18790472e-01 1.77854255e-01
-1.12758815e+00 -1.30220520e+00 -1.04121819e-01 1.05678046e+00
-3.98153782e-01 -2.22799003e-01 -6.55960739e-01 -3.29857290e-01
-9.53134298e-02 6.72630310e-01 -2.56826371e-01 -5.62995709e-02
-2.47537628e-01 -4.25870717e-01 1.00337553e+00 7.48678565e-01
4.29034501e-01 -8.02685738e-01 9.36868265e-02 -6.05739933e-03
-4.51033652e-01 -1.45998371e+00 -3.11014503e-01 4.16207522e-01
-2.82254189e-01 -3.96778017e-01 -1.00902307e+00 -8.36213291e-01
4.88028005e-02 -2.11787522e-01 8.20473194e-01 -4.20896441e-01
-7.46904835e-02 6.15005553e-01 -1.15986601e-01 -2.92695999e-01
-6.20948732e-01 -2.22345963e-01 2.31370181e-01 2.25478187e-01
3.50234360e-01 -4.32105929e-01 -2.54645228e-01 -6.26718700e-02
-7.00158060e-01 -1.54884934e-01 3.83434176e-01 8.91012549e-01
4.10800815e-01 2.76277065e-02 6.14311278e-01 9.85668674e-02
4.86065954e-01 -1.92388967e-01 -5.13951719e-01 1.38219297e-01
-2.19829634e-01 -1.52293935e-01 1.34582564e-01 -3.73749912e-01
-1.02525508e+00 2.98165441e-01 -7.37323165e-01 -6.97790205e-01
-3.94719929e-01 4.15515810e-01 -3.65761250e-01 6.32554144e-02
3.14287961e-01 1.89300239e-01 1.30417511e-01 -5.16743064e-01
4.89169538e-01 1.22039831e+00 7.05594301e-01 -3.72622758e-02
4.21228886e-01 2.20836267e-01 -3.59405518e-01 -1.07655215e+00
-2.81577021e-01 -3.42191666e-01 -2.14930519e-01 -1.58475801e-01
1.17004824e+00 -1.45705640e+00 -1.05309224e+00 5.87603569e-01
-1.17037487e+00 -1.92181230e-01 -2.05108359e-01 9.02733803e-01
-5.36391854e-01 2.17833355e-01 -7.46295393e-01 -1.17376733e+00
-2.95117497e-01 -1.51515126e+00 1.21748781e+00 5.66827320e-03
-6.82954490e-02 -8.26323092e-01 -2.62210429e-01 7.07507610e-01
3.57135087e-01 -1.87143981e-01 6.00982666e-01 -6.41514421e-01
-3.49251628e-01 -3.15455407e-01 -5.14371693e-01 9.82786715e-01
-6.08890206e-02 -6.12706803e-02 -1.78668630e+00 -3.60193819e-01
-1.91198796e-01 -4.57779825e-01 1.08244216e+00 5.34319341e-01
1.01986408e+00 -2.40967616e-01 5.41634895e-02 4.47755188e-01
1.19983554e+00 5.44899225e-01 6.74398065e-01 -3.72523814e-01
8.24816406e-01 3.26062173e-01 -1.45387918e-01 4.46290493e-01
2.72102624e-01 8.10881495e-01 4.35151756e-01 -2.43576884e-01
-3.53278279e-01 -5.04572801e-02 7.25757241e-01 9.37804282e-01
3.35496873e-01 -4.40236598e-01 -8.75675857e-01 7.02587426e-01
-1.38283885e+00 -8.54759216e-01 2.85489142e-01 2.26049590e+00
7.51580179e-01 2.30645668e-03 2.02603474e-01 3.21338087e-01
6.51664972e-01 9.70290676e-02 -5.15217409e-02 -6.47985041e-01
-4.16481018e-01 4.21820223e-01 3.25564057e-01 1.01091218e+00
-1.22506285e+00 6.19164407e-01 6.69144535e+00 1.08995473e+00
-1.30003715e+00 2.79654980e-01 6.00410521e-01 -4.38542545e-01
-1.69908017e-01 -4.49222594e-01 -6.21484041e-01 2.81758636e-01
1.60777879e+00 4.72021520e-01 6.05934501e-01 4.85687375e-01
7.65234977e-02 -1.32280841e-01 -1.01335347e+00 1.48785520e+00
3.05961281e-01 -1.12354052e+00 -3.18392217e-02 1.44138485e-01
4.39754784e-01 3.16745132e-01 4.77933347e-01 3.75327080e-01
3.18882316e-01 -1.56738687e+00 7.68526733e-01 3.47587824e-01
1.13296199e+00 -1.04240847e+00 7.46518433e-01 8.91954601e-02
-1.14090800e+00 -1.72118202e-01 8.17245692e-02 2.09085613e-01
4.22545940e-01 3.87797862e-01 -1.12776220e+00 3.47693056e-01
8.01095545e-01 2.38637865e-01 -1.24445803e-01 6.75101936e-01
1.79009199e-01 6.45588815e-01 -5.83503842e-01 3.34647268e-01
1.64129391e-01 4.87472296e-01 5.62166929e-01 1.49642086e+00
5.06468952e-01 -2.14568138e-01 -7.69490227e-02 6.22809112e-01
-3.77707422e-01 -1.19694263e-01 -5.23180723e-01 -3.69630188e-01
3.11550051e-01 1.06004095e+00 -1.71880677e-01 -5.71084261e-01
-3.97405654e-01 9.62148011e-01 5.15710153e-02 5.83918035e-01
-9.48100269e-01 -4.31595147e-01 7.00664461e-01 -2.55239099e-01
7.41313934e-01 -2.42922351e-01 -3.75345647e-01 -1.14226437e+00
-1.24564648e-01 -8.80065143e-01 1.01163857e-01 -1.02758563e+00
-1.05595815e+00 6.63410485e-01 -2.20156506e-01 -9.77335036e-01
-3.14171284e-01 -6.41985059e-01 -1.10587753e-01 1.35886204e+00
-1.15702355e+00 -1.26633406e+00 1.98431656e-01 6.38913333e-01
4.29938823e-01 -4.10020679e-01 9.08501625e-01 7.65690207e-01
-4.46935415e-01 1.12039435e+00 5.24140187e-02 2.38660365e-01
6.07874870e-01 -9.51299608e-01 1.35282790e-02 6.30280375e-01
2.68340439e-01 3.74955148e-01 5.40632010e-01 -2.59214550e-01
-1.08454633e+00 -8.95924866e-01 1.07136631e+00 -4.10766691e-01
3.42540979e-01 -5.11813581e-01 -8.33483100e-01 6.36631548e-01
6.10694289e-01 -2.29453996e-01 8.94331455e-01 2.48760190e-02
-7.01263726e-01 -2.20970377e-01 -1.21821702e+00 2.21463248e-01
3.19161475e-01 -1.22121739e+00 -3.16966027e-01 -6.28174022e-02
8.58407497e-01 -2.19834596e-01 -1.06918025e+00 4.57886040e-01
6.95876598e-01 -7.14237452e-01 1.13746512e+00 -3.68231595e-01
1.99123189e-01 -2.60291666e-01 -9.67290103e-01 -1.21145213e+00
7.47457221e-02 -4.10123736e-01 -1.21387377e-01 1.33445024e+00
7.97762275e-01 -4.40852225e-01 4.46007401e-01 4.78758276e-01
-2.17797056e-01 -3.82584333e-01 -1.40191829e+00 -4.70928431e-01
-1.47244101e-02 -1.00707340e+00 2.86657155e-01 6.77534819e-01
-2.13644262e-02 5.23074150e-01 -6.62049711e-01 2.88693637e-01
4.09876585e-01 -6.60193801e-01 4.48378772e-01 -7.79614329e-01
-4.43769574e-01 -3.77439171e-01 -3.57528657e-01 -1.03726745e+00
1.73028275e-01 -7.20900714e-01 2.59579957e-01 -1.47910082e+00
-3.86004411e-02 -2.45871898e-02 -4.51817840e-01 6.89698935e-01
3.17519814e-01 7.06540108e-01 2.96101719e-01 -1.69099495e-01
-3.58444124e-01 5.74293971e-01 5.93411505e-01 -4.55826104e-01
-2.83403933e-01 -1.79710120e-01 -5.54083049e-01 6.21946573e-01
5.24974048e-01 -1.14300393e-01 -9.79055166e-02 -4.47480261e-01
-6.56843111e-02 1.95460632e-01 5.32457769e-01 -9.60202515e-01
1.03839763e-01 4.14078891e-01 5.55091143e-01 -6.78791702e-01
1.24381554e+00 -9.45143282e-01 1.33582905e-01 4.14266959e-02
-4.54517543e-01 -3.12407315e-01 4.71352816e-01 4.51186001e-01
-5.99434853e-01 2.83182681e-01 8.99801075e-01 1.38024569e-01
-3.41327071e-01 -1.77543491e-01 -6.68097675e-01 -4.78047609e-01
5.51661372e-01 -1.62880287e-01 -1.46644801e-01 -9.54487622e-01
-1.41691923e+00 2.16094069e-02 -1.00485511e-01 4.31306720e-01
7.07760930e-01 -1.49402189e+00 -5.95648766e-01 3.10311407e-01
-5.77033348e-02 -4.29457486e-01 5.77680469e-01 9.77566898e-01
-1.24044679e-01 5.38201511e-01 1.50135770e-01 -7.90743232e-01
-1.54690325e+00 2.95483619e-01 5.20749390e-01 1.90001026e-01
5.87492716e-03 1.26672900e+00 6.02943413e-02 -4.16920453e-01
7.98855305e-01 -2.36548364e-01 -7.28027374e-02 3.81487489e-01
4.58237886e-01 2.90050477e-01 3.71851981e-01 -1.23609471e+00
-5.29947996e-01 2.15704381e-01 2.52917469e-01 -8.91012430e-01
1.16293299e+00 -2.64101446e-01 1.84302047e-01 6.03949070e-01
1.79095912e+00 4.48047556e-02 -1.15228426e+00 -2.66970277e-01
-7.76981354e-01 -6.01269770e-04 4.17115778e-01 -1.03810441e+00
-1.03862059e+00 1.38166904e+00 1.09937978e+00 1.79746285e-01
1.22176349e+00 1.11543827e-01 7.07927465e-01 1.61781743e-01
-3.34621102e-01 -1.17618978e+00 -1.17077030e-01 5.84390581e-01
8.37946296e-01 -1.32527232e+00 -4.65800971e-01 1.05406329e-01
-7.83216298e-01 9.09464061e-01 1.91187292e-01 2.99849421e-01
8.12079608e-01 4.56515968e-01 1.94384485e-01 1.25271887e-01
-7.01246381e-01 -3.69011670e-01 5.59630394e-01 7.12340117e-01
3.99301052e-01 1.53573275e-01 4.56914127e-01 6.73587620e-01
-1.09801777e-01 -2.15500712e-01 1.08091503e-01 6.82139397e-01
-4.23455566e-01 -7.93322682e-01 -5.91614008e-01 1.81407839e-01
-6.59815907e-01 -3.41638327e-01 -3.53824377e-01 3.54417026e-01
1.17649883e-01 1.35336494e+00 2.89700270e-01 -7.88845658e-01
1.26220897e-01 5.51194429e-01 4.00844842e-01 -2.33835265e-01
-5.85530519e-01 6.93153203e-01 3.96947622e-01 -3.80831331e-01
-2.80266732e-01 -3.82708728e-01 -1.18352878e+00 -1.89069659e-01
-1.52553245e-01 -6.90064654e-02 1.04309332e+00 8.78773510e-01
3.54676932e-01 7.07684815e-01 4.76895809e-01 -1.04829645e+00
-2.56637663e-01 -1.00150990e+00 -4.99764174e-01 7.09937140e-03
8.46169889e-01 -4.32333827e-01 -5.95662653e-01 2.11123750e-01] | [14.369234085083008, 5.167525768280029] |
10277e3c-ec93-4077-8e8e-5d09ae4f40ae | group-activity-recognition-via-dynamic | 2305.05583 | null | https://arxiv.org/abs/2305.05583v1 | https://arxiv.org/pdf/2305.05583v1.pdf | Group Activity Recognition via Dynamic Composition and Interaction | Previous group activity recognition approaches were limited to reasoning using human relations or finding important subgroups and tended to ignore indispensable group composition and human-object interactions. This absence makes a partial interpretation of the scene and increases the interference of irrelevant actions on the results. Therefore, we propose our DynamicFormer with Dynamic composition Module (DcM) and Dynamic interaction Module (DiM) to model relations and locations of persons and discriminate the contribution of participants, respectively. Our findings on group composition and human-object interaction inspire our core idea. Group composition tells us the location of people and their relations inside the group, while interaction reflects the relation between humans and objects outside the group. We utilize spatial and temporal encoders in DcM to model our dynamic composition and build DiM to explore interaction with a novel GCN, which has a transformer inside to consider the temporal neighbors of human/object. Also, a Multi-level Dynamic Integration is employed to integrate features from different levels. We conduct extensive experiments on two public datasets and show that our method achieves state-of-the-art. | ['Zheng Wang', 'Danni Xu', 'Wenxuan Liu', 'Zhuo Zhou', 'Youliang Zhang'] | 2023-05-09 | null | null | null | null | ['human-object-interaction-detection', 'group-activity-recognition'] | ['computer-vision', 'computer-vision'] | [-6.29810318e-02 -1.28235772e-01 1.92441046e-01 -5.01352191e-01
2.45227531e-01 -3.74303699e-01 1.04569066e+00 -3.02035771e-02
-6.23532176e-01 2.53271312e-01 8.48086536e-01 1.96087763e-01
-3.21549028e-01 -8.46355796e-01 -5.06444752e-01 -6.06663048e-01
-2.78829634e-01 5.64073384e-01 4.18961763e-01 -1.08891847e-02
-3.41583230e-02 4.09097880e-01 -1.79819155e+00 8.17415297e-01
5.61733007e-01 8.81771147e-01 1.22215413e-01 3.99722070e-01
2.27306351e-01 1.19908881e+00 -7.75693238e-01 -2.18150765e-01
3.88984382e-01 -6.98682010e-01 -6.96603358e-01 2.19104916e-01
4.07536209e-01 -4.45880473e-01 -5.86789489e-01 5.41337311e-01
5.66576421e-01 6.68774724e-01 6.72123611e-01 -1.72234476e+00
-7.61376441e-01 8.04176629e-01 -5.92600107e-01 3.91241580e-01
6.87586665e-01 4.72941339e-01 1.00733447e+00 -7.59265244e-01
8.03828835e-01 1.62589669e+00 3.39260429e-01 4.95645404e-01
-1.03121662e+00 -6.20159924e-01 8.40986490e-01 7.09769785e-01
-1.46578348e+00 -4.32900399e-01 5.42710364e-01 -5.53482652e-01
1.12344611e+00 5.31414747e-01 9.11256731e-01 1.36702883e+00
-2.27691591e-01 9.86356199e-01 7.02646673e-01 -1.44332692e-01
2.19567820e-01 -2.14019567e-01 4.81037289e-01 7.36024559e-01
1.18030585e-01 -2.82809436e-02 -8.97201598e-01 -3.51235159e-02
8.04927886e-01 3.75656128e-01 -2.28461787e-01 -4.02853131e-01
-1.66304481e+00 3.07962030e-01 6.30606294e-01 4.30924147e-01
-4.51767772e-01 1.50941148e-01 1.39296174e-01 -6.17351979e-02
7.59389773e-02 2.43974477e-01 -7.93993101e-02 -7.36593902e-02
-4.39797163e-01 5.00044167e-01 8.65454018e-01 8.68985713e-01
4.65474278e-01 -7.12397575e-01 -8.30890357e-01 6.00403607e-01
1.63279191e-01 1.00294434e-01 1.56280667e-01 -1.09903765e+00
4.65951920e-01 1.24107528e+00 -1.27732633e-02 -1.09253597e+00
-6.63848281e-01 -2.49393523e-01 -7.97372162e-01 -1.93492863e-02
4.87699032e-01 1.33275181e-01 -6.80219591e-01 1.99864268e+00
4.78193074e-01 2.94444144e-01 -4.81094927e-01 1.15106833e+00
7.92274773e-01 1.96633160e-01 2.86775112e-01 4.38544787e-02
1.58055198e+00 -1.42571354e+00 -7.20141888e-01 -3.66966456e-01
6.66617930e-01 5.22188358e-02 9.51882482e-01 1.31623074e-01
-9.60251987e-01 -8.74220490e-01 -6.60245836e-01 -3.34909588e-01
-5.23082376e-01 4.41505313e-02 8.93258512e-01 2.54743129e-01
-8.27593505e-01 4.68867600e-01 -8.79774213e-01 -6.29789829e-01
5.16326964e-01 3.54398102e-01 -4.72564757e-01 1.80990566e-02
-9.16969240e-01 7.79994667e-01 1.51255727e-01 3.67673844e-01
-8.29091609e-01 -4.36192572e-01 -7.97725320e-01 1.92101926e-01
7.89281905e-01 -1.09894168e+00 7.97823548e-01 -9.26694274e-01
-7.99149215e-01 5.93798101e-01 -4.39628184e-01 -2.37007976e-01
6.97103024e-01 -3.64294797e-01 -4.43715513e-01 -5.43568507e-02
1.64065525e-01 7.45881259e-01 2.47456953e-01 -1.10054231e+00
-7.95156956e-01 -7.15085685e-01 2.81494081e-01 6.48931742e-01
1.65961999e-02 2.63278723e-01 -6.89245701e-01 -4.41817045e-01
2.14605838e-01 -7.48200595e-01 2.60794032e-02 2.86556929e-01
-2.80629605e-01 -5.31986296e-01 5.96671939e-01 -6.92313135e-01
1.41667259e+00 -2.22319055e+00 3.44720513e-01 2.99988855e-02
5.79512656e-01 -2.10119337e-01 -2.98461765e-01 5.15916407e-01
-1.07912228e-01 5.10375798e-02 1.37581110e-01 -5.39590418e-01
1.82954818e-01 2.71340460e-01 8.47243518e-02 6.02982998e-01
-1.18819542e-01 1.34809363e+00 -8.69695425e-01 -4.13491309e-01
1.96431935e-01 4.59336728e-01 -7.71779537e-01 2.27788344e-01
-1.10936701e-01 4.16246593e-01 -3.10550034e-01 6.52113736e-01
5.96801162e-01 -3.53371531e-01 3.52962255e-01 -1.72920197e-01
-1.10256016e-01 3.72979224e-01 -1.36425662e+00 1.49276245e+00
2.04504371e-01 3.03483784e-01 3.60208973e-02 -6.26816511e-01
2.50964284e-01 -7.12311640e-02 5.17919123e-01 -8.48904669e-01
7.94869885e-02 -3.44011694e-01 5.28065443e-01 -6.27152085e-01
1.80553436e-01 5.88791370e-01 6.39856607e-02 7.38486230e-01
-1.05079852e-01 9.25381362e-01 2.50174403e-01 4.36082304e-01
1.47052610e+00 3.48293513e-01 3.11176986e-01 -1.23860918e-01
3.63950849e-01 -5.43160737e-01 5.47787309e-01 1.07221913e+00
-4.65782583e-01 3.85579050e-01 6.05401397e-01 -6.47439063e-01
-4.36425179e-01 -1.10983527e+00 5.96617699e-01 1.53578198e+00
5.00954330e-01 -6.70567214e-01 -5.79246938e-01 -8.97622883e-01
-4.44910899e-02 5.80488443e-01 -1.11976671e+00 -1.66134313e-01
-9.08690274e-01 -6.29377604e-01 2.15975136e-01 8.44186008e-01
8.26265931e-01 -1.34984803e+00 -7.78569937e-01 -1.39271379e-01
-5.77429116e-01 -1.04359639e+00 -6.26997173e-01 -1.65838152e-01
-3.16642940e-01 -1.54547024e+00 -9.93771106e-02 -5.87412477e-01
8.14852178e-01 4.19230044e-01 9.63993371e-01 1.71650723e-01
-2.91808605e-01 4.04499441e-01 -2.94744074e-01 -1.74545363e-01
3.94772053e-01 -1.47047251e-01 1.25713542e-01 2.96009302e-01
6.19136751e-01 -6.23968303e-01 -9.08261120e-01 7.73921728e-01
-4.53390151e-01 4.03163373e-01 3.79080445e-01 4.40840423e-01
2.38044537e-03 5.95005229e-02 -5.62888347e-02 -6.85023665e-01
3.35958511e-01 -4.83594775e-01 7.77576026e-03 4.65926975e-01
-8.56062621e-02 8.79318453e-03 1.31869912e-01 -8.32837403e-01
-1.06719708e+00 -1.50200650e-01 5.38283229e-01 -2.49249175e-01
-2.31803998e-01 1.51115999e-01 -6.81949854e-01 4.58521008e-01
5.68867266e-01 7.79897645e-02 -2.94561356e-01 -4.22190517e-01
2.68507659e-01 1.87212422e-01 5.48931122e-01 -6.50179863e-01
3.94865721e-01 7.80231237e-01 -1.91430256e-01 -3.52397352e-01
-6.69290423e-01 -5.57417989e-01 -9.20233667e-01 -3.93146574e-01
9.68382537e-01 -9.81803417e-01 -1.36114633e+00 5.31232059e-01
-1.23557353e+00 -5.68290770e-01 -3.06798190e-01 3.66492718e-01
-3.00606251e-01 2.44460896e-01 -3.83149713e-01 -1.00953484e+00
2.06320062e-01 -8.11950982e-01 1.17855096e+00 3.88706074e-04
-5.28970599e-01 -6.16985977e-01 -2.67665565e-01 4.93057251e-01
1.39105663e-01 1.79801106e-01 5.12937844e-01 -6.95263028e-01
-1.03523684e+00 1.17265955e-01 -3.58007818e-01 -3.38094026e-01
1.52312547e-01 -4.54353392e-01 -8.04731548e-01 -2.12408379e-02
-1.09594665e-01 2.52015114e-01 1.04534018e+00 2.87112743e-01
1.26725769e+00 -4.34901893e-01 -9.04012859e-01 5.01589000e-01
7.66650498e-01 2.81327307e-01 7.66308188e-01 8.19190368e-02
9.24634814e-01 9.43105102e-01 5.17148256e-01 5.90236425e-01
7.16362834e-01 8.60121667e-01 2.47899547e-01 4.73015681e-02
-2.04873279e-01 -3.31726223e-01 4.33006078e-01 1.38098508e-01
-5.46059728e-01 -7.17248619e-01 -7.74557173e-01 4.33717370e-01
-2.39611411e+00 -1.37640107e+00 -7.22998306e-02 1.89572549e+00
3.70462447e-01 -1.60831362e-01 3.98191571e-01 -1.36468366e-01
6.22411072e-01 2.91541904e-01 -3.44301045e-01 3.61074775e-01
-1.22502342e-01 -2.55498588e-01 6.92841262e-02 3.15947205e-01
-1.03937483e+00 8.22396398e-01 6.23106766e+00 5.14927268e-01
-3.99705768e-01 1.91392526e-01 4.55974936e-01 -5.97513676e-01
-1.61828902e-02 -9.55245197e-02 -7.15752959e-01 5.51900029e-01
4.67803895e-01 1.24682330e-01 7.18401790e-01 3.01927686e-01
2.86790341e-01 -5.16080379e-01 -1.69999504e+00 1.06974185e+00
3.21793020e-01 -7.88478911e-01 5.57630956e-02 3.29634756e-01
2.85451412e-01 -3.48869532e-01 -2.86407083e-01 4.12767231e-01
4.21608508e-01 -9.99966085e-01 9.35851455e-01 9.37587440e-01
1.57937378e-01 -2.95836687e-01 4.76711333e-01 6.27230167e-01
-1.48074746e+00 -5.18305302e-01 7.45377764e-02 -6.02933109e-01
2.55297989e-01 3.16502014e-03 -5.51039577e-01 3.70661169e-01
8.89468789e-01 7.64480472e-01 -8.90056849e-01 7.79857278e-01
-2.49508590e-01 1.05582505e-01 -3.36677849e-01 6.03164099e-02
-1.29162848e-01 -2.50443190e-01 4.28484350e-01 9.53641117e-01
1.25143409e-01 4.42759454e-01 3.97807717e-01 1.08371747e+00
1.31576434e-01 -2.66879559e-01 -4.32906896e-01 4.37171236e-02
5.31613231e-01 9.16166961e-01 -7.69766331e-01 -4.28368479e-01
-4.78834987e-01 1.19535458e+00 5.11572421e-01 6.33344531e-01
-1.13519371e+00 1.93999723e-01 8.74976099e-01 5.21150053e-01
1.65305763e-01 -3.36814493e-01 -1.35643318e-01 -1.15851736e+00
3.71358544e-01 -8.03222775e-01 8.33371878e-01 -8.03748846e-01
-1.17537951e+00 2.88474232e-01 2.44718015e-01 -9.91083622e-01
-5.74975871e-02 -4.46334362e-01 -3.97079051e-01 7.45054722e-01
-4.94646579e-01 -1.56649280e+00 -7.28903592e-01 6.14621997e-01
5.19553483e-01 1.99420862e-02 5.38143814e-01 3.19450378e-01
-7.32676506e-01 5.11162460e-01 -7.08422303e-01 4.22014475e-01
5.54556310e-01 -8.83117795e-01 3.61734152e-01 9.37181175e-01
2.42399544e-01 9.65439260e-01 3.41906220e-01 -9.23574209e-01
-1.10946453e+00 -7.82254219e-01 1.20481884e+00 -1.10331440e+00
2.96764314e-01 -1.02952421e+00 -6.40103519e-01 1.00669646e+00
1.86741725e-02 3.78086939e-02 7.11517751e-01 4.92893964e-01
-6.46331191e-01 1.90974072e-01 -9.14827049e-01 1.12213576e+00
2.29344416e+00 -5.50070107e-01 -5.32083571e-01 8.64660144e-02
7.15636373e-01 -1.24936879e-01 -2.73395538e-01 3.41497779e-01
8.27248037e-01 -1.33810544e+00 1.15405655e+00 -5.87568641e-01
2.20750734e-01 -5.43835461e-01 -8.43261331e-02 -9.49619293e-01
-8.34286034e-01 -1.42065033e-01 -5.80131412e-01 1.11141026e+00
-1.73021723e-02 -5.89832902e-01 5.84201336e-01 8.92431378e-01
5.12329042e-02 -4.95606571e-01 -5.77851295e-01 -6.44839764e-01
-6.99009597e-01 -3.32844079e-01 8.24583530e-01 9.47201014e-01
1.28760442e-01 3.17611217e-01 -2.50940442e-01 2.43946850e-01
2.59114385e-01 8.52000415e-02 6.79360509e-01 -1.32901955e+00
-4.85756099e-01 -5.33523858e-01 -4.45670009e-01 -1.09489548e+00
1.16463214e-01 -6.50023580e-01 -2.01281726e-01 -1.79298425e+00
5.78053415e-01 -1.16961986e-01 -3.51335913e-01 7.82989085e-01
-2.18571723e-01 -8.48127753e-02 1.80561215e-01 1.96080446e-01
-1.18239701e+00 4.08934116e-01 1.24070168e+00 -2.33287528e-01
-3.67500931e-01 -1.43388510e-01 -9.22780812e-01 7.31855929e-01
2.86981165e-01 -1.36466771e-01 -4.49390531e-01 -6.54001713e-01
2.31289372e-01 -4.10609990e-01 8.51309001e-01 -1.22733307e+00
5.45337439e-01 -3.02800745e-01 7.48933733e-01 -5.58552384e-01
5.36733508e-01 -8.24344039e-01 4.74623471e-01 2.71631867e-01
-4.22247589e-01 -8.41568708e-02 -1.63000315e-01 6.38247073e-01
1.66548453e-02 4.62397873e-01 1.76230446e-01 -3.59750360e-01
-8.55897307e-01 4.37027037e-01 -2.05513939e-01 -3.21208984e-01
1.04803634e+00 -4.42701012e-01 -4.90350842e-01 -3.83172095e-01
-9.46394503e-01 4.96745527e-01 4.61973995e-01 6.58729553e-01
3.94921362e-01 -1.43935347e+00 -4.61657912e-01 3.60983968e-01
2.50467628e-01 -7.56799653e-02 5.71346283e-01 9.76467073e-01
-1.15527764e-01 1.74269557e-01 -1.73603460e-01 -3.27414989e-01
-1.42636776e+00 7.61656463e-01 4.81273681e-01 -1.39989525e-01
-6.40249133e-01 9.58755136e-01 9.01229918e-01 -3.97145987e-01
5.53302884e-01 -3.93851161e-01 -3.73610646e-01 3.49455386e-01
8.01572740e-01 6.13288760e-01 -4.13475960e-01 -8.58251154e-01
-7.18498647e-01 1.96932107e-01 1.76356956e-01 -3.02948624e-01
1.23331702e+00 -1.60888165e-01 -3.05447638e-01 4.06890541e-01
7.60956168e-01 -1.73823237e-01 -1.27051330e+00 -3.45717579e-01
-2.21624244e-02 -6.44903362e-01 -4.23240095e-01 -1.03637016e+00
-8.13892305e-01 6.92159116e-01 4.23388004e-01 -6.16876930e-02
1.09754777e+00 3.29276383e-01 2.62029320e-01 3.61401826e-01
6.43447876e-01 -1.05610943e+00 2.21366331e-01 4.53687340e-01
1.04539394e+00 -9.15100515e-01 3.35879028e-02 -7.33593464e-01
-7.03299344e-01 5.16572654e-01 1.13910186e+00 2.38424554e-01
5.58341920e-01 5.00399657e-02 -4.07329977e-01 -4.03102666e-01
-1.09969008e+00 -4.25008029e-01 4.43672240e-01 6.32846475e-01
3.47356051e-01 1.54712573e-01 -2.85366595e-01 9.46558833e-01
-7.39649832e-02 -4.92697023e-02 -1.13882296e-01 9.87638652e-01
-1.53970718e-01 -8.44360352e-01 -1.76269427e-01 4.45868224e-01
1.21918023e-01 3.91718484e-02 -6.90867245e-01 7.46024609e-01
7.97654927e-01 8.41813862e-01 6.50067866e-01 -4.74432409e-01
4.15940553e-01 1.74073055e-01 6.91709459e-01 -5.93157947e-01
-8.17267418e-01 -1.16250210e-01 1.18153036e-01 -1.09434390e+00
-5.75100541e-01 -9.37894702e-01 -1.23448408e+00 -3.29202116e-01
1.07506357e-01 -2.80710638e-01 1.05109811e-01 8.90045881e-01
5.98307252e-01 7.47724950e-01 1.37643248e-01 -7.69163609e-01
2.26266291e-02 -1.12740886e+00 -5.84930658e-01 7.13389993e-01
9.20472965e-02 -8.91415000e-01 -2.64147371e-01 2.93509439e-02] | [8.214689254760742, 0.5870494246482849] |
4beb1297-4b7d-474f-b18b-910aff577c16 | simplifying-model-based-rl-learning | 2209.08466 | null | https://arxiv.org/abs/2209.08466v3 | https://arxiv.org/pdf/2209.08466v3.pdf | Simplifying Model-based RL: Learning Representations, Latent-space Models, and Policies with One Objective | While reinforcement learning (RL) methods that learn an internal model of the environment have the potential to be more sample efficient than their model-free counterparts, learning to model raw observations from high dimensional sensors can be challenging. Prior work has addressed this challenge by learning low-dimensional representation of observations through auxiliary objectives, such as reconstruction or value prediction. However, the alignment between these auxiliary objectives and the RL objective is often unclear. In this work, we propose a single objective which jointly optimizes a latent-space model and policy to achieve high returns while remaining self-consistent. This objective is a lower bound on expected returns. Unlike prior bounds for model-based RL on policy exploration or model guarantees, our bound is directly on the overall RL objective. We demonstrate that the resulting algorithm matches or improves the sample-efficiency of the best prior model-based and model-free RL methods. While sample efficient methods typically are computationally demanding, our method attains the performance of SAC in about 50% less wall-clock time. | ['Ruslan Salakhutdinov', 'Sergey Levine', 'Benjamin Eysenbach', 'Homanga Bharadhwaj', 'Raj Ghugare'] | 2022-09-18 | null | null | null | null | ['value-prediction'] | ['computer-code'] | [ 1.50317222e-01 4.38982189e-01 -8.26806068e-01 -2.46670172e-01
-1.23237181e+00 -6.08277500e-01 6.77676916e-01 1.48623526e-01
-7.32856810e-01 9.98382688e-01 2.23144591e-01 -3.78887206e-01
-2.34769478e-01 -6.72124028e-01 -1.01681900e+00 -6.33801341e-01
-2.17428967e-01 7.83234358e-01 -1.97948471e-01 3.43886673e-01
3.58057231e-01 3.84561241e-01 -1.29556525e+00 -4.04180080e-01
3.04898441e-01 1.24108875e+00 3.56823951e-01 8.60481977e-01
-2.24247202e-02 9.92810607e-01 -2.17753828e-01 2.59332716e-01
5.19112229e-01 -3.90315592e-01 -5.61068714e-01 1.47896543e-01
5.59013058e-03 -5.93616426e-01 -9.51431617e-02 9.22805429e-01
4.25184965e-01 2.53938854e-01 5.06071746e-01 -1.24332535e+00
-1.87638998e-01 4.89209890e-01 -4.78199184e-01 -8.76781344e-02
1.58569083e-01 1.36624977e-01 9.76788580e-01 -3.79884750e-01
4.44009334e-01 1.54779875e+00 2.82569915e-01 4.62989151e-01
-1.65796816e+00 -5.90789557e-01 5.69357395e-01 -3.99410278e-01
-7.39693105e-01 -6.93243384e-01 4.05882299e-01 -2.45510623e-01
7.58721113e-01 -5.04331514e-02 5.74252367e-01 1.08111918e+00
-5.79235293e-02 1.25501037e+00 1.29116940e+00 -2.37721026e-01
7.78792799e-01 1.96146280e-01 -1.47929519e-01 6.25929356e-01
5.30646622e-01 6.43457651e-01 -6.70225561e-01 -3.36878628e-01
1.14293528e+00 1.47419393e-01 6.30035102e-02 -1.01597309e+00
-1.05786085e+00 9.84918356e-01 4.63638827e-02 -5.40890455e-01
-5.42474031e-01 9.04186130e-01 5.52818589e-02 3.62596512e-01
3.01391482e-01 6.17590904e-01 -7.58230209e-01 -6.66780353e-01
-8.66205513e-01 5.95972061e-01 1.04154217e+00 1.00160730e+00
8.33923757e-01 1.58745214e-01 -1.37232602e-01 2.20388681e-01
6.52822375e-01 8.08722198e-01 1.82608262e-01 -1.52815938e+00
5.54282784e-01 -6.78114407e-03 8.97406459e-01 -3.71782482e-01
-2.18006134e-01 -6.69084668e-01 -2.27154508e-01 4.35855478e-01
4.21360075e-01 -4.25996274e-01 -8.03540945e-01 1.93643546e+00
4.53278929e-01 1.97744280e-01 2.16760501e-01 6.39186382e-01
-2.32310042e-01 5.67205191e-01 -1.95952699e-01 -5.01339257e-01
7.83874393e-01 -9.34130251e-01 -6.96202874e-01 -6.48906589e-01
4.60456967e-01 -4.02463973e-01 1.17971575e+00 4.32349086e-01
-1.48862338e+00 -1.45394474e-01 -1.01430929e+00 3.18324268e-01
3.18211555e-01 -1.62367582e-01 8.58225584e-01 5.99823356e-01
-8.56221199e-01 6.32179976e-01 -1.55388069e+00 3.15068811e-02
3.67340535e-01 4.81143862e-01 8.78013521e-02 1.42808110e-01
-4.37034786e-01 9.02958512e-01 2.58181006e-01 -4.21332717e-01
-1.50333369e+00 -7.41128325e-01 -7.93065012e-01 1.52626216e-01
9.56507564e-01 -4.86250877e-01 1.99497795e+00 -5.51608205e-01
-2.02323246e+00 1.54500127e-01 -2.65386581e-01 -9.42352235e-01
7.04490840e-01 -5.88909149e-01 3.03953022e-01 -1.21098012e-01
-8.47349316e-02 5.11057377e-01 9.66128528e-01 -1.41929114e+00
-8.38138759e-01 -1.81618467e-01 1.70534313e-01 2.97508985e-01
-1.65479243e-01 -5.14119804e-01 -2.79458284e-01 -2.95459956e-01
2.76643932e-02 -1.01362193e+00 -6.28130198e-01 1.18488401e-01
1.20157458e-01 4.02355269e-02 5.44975817e-01 -2.21497402e-01
1.00047374e+00 -1.84286559e+00 -1.02760985e-01 2.58345038e-01
-4.61294949e-02 -1.17745556e-01 -1.86426058e-01 5.04596651e-01
6.27577126e-01 -5.24459593e-02 -1.08445369e-01 -7.99652219e-01
3.90126288e-01 6.62389696e-01 -7.19851851e-01 7.22574532e-01
-3.04513797e-03 9.40556228e-01 -1.16935599e+00 -2.52367228e-01
2.65734792e-01 -6.45302385e-02 -8.87862563e-01 4.13504452e-01
-7.39417672e-01 4.51086879e-01 -5.52991450e-01 4.54513103e-01
3.00477028e-01 -4.73128051e-01 6.00922346e-01 4.79305297e-01
5.35436645e-02 5.41430354e-01 -1.39650881e+00 1.76313674e+00
-7.25490808e-01 2.25094602e-01 5.36886990e-01 -1.06168246e+00
7.54503131e-01 1.22782841e-01 5.98066747e-01 -7.75091469e-01
-1.47291660e-01 4.55847196e-02 -4.58602756e-01 -1.81980118e-01
5.03166199e-01 -2.14213416e-01 -2.12409511e-01 8.13919783e-01
-1.07099913e-01 -1.95826054e-01 -8.99854302e-03 -1.31002590e-02
1.04846478e+00 7.71209419e-01 4.33296770e-01 -1.60298347e-01
-3.15652102e-01 -2.16112271e-01 6.52161658e-01 1.32212627e+00
6.72002556e-04 -2.88516674e-02 5.64376771e-01 -1.69946700e-01
-9.38878477e-01 -1.17883277e+00 6.12734929e-02 1.00970101e+00
1.65508851e-01 -4.29914474e-01 -4.05366689e-01 -5.45514524e-01
4.06252295e-01 7.82824695e-01 -5.41352570e-01 8.35388899e-02
-4.01203394e-01 -5.04765332e-01 1.93263650e-01 5.13089538e-01
2.03894794e-01 -5.48995316e-01 -1.02607369e+00 5.82974970e-01
1.59239024e-01 -9.32110250e-01 -3.09048504e-01 4.36455488e-01
-1.14899242e+00 -8.34262967e-01 -4.08339947e-01 -2.56420695e-03
6.62743688e-01 7.07050860e-02 1.16926718e+00 -4.28849816e-01
1.33287176e-01 8.57844234e-01 7.13080317e-02 -8.14936936e-01
-2.06012517e-01 -6.01300225e-02 3.13181370e-01 -1.68580815e-01
4.15781373e-03 -6.78485870e-01 -4.91137534e-01 -8.74192789e-02
-7.24119544e-01 -4.85468507e-02 7.87377775e-01 9.07530785e-01
9.16224241e-01 -2.46375456e-01 4.35931325e-01 -7.40918517e-01
7.63345540e-01 -4.85370159e-01 -1.35436237e+00 2.63984073e-02
-1.12774348e+00 9.38805401e-01 4.12025124e-01 -6.59276664e-01
-9.38972890e-01 3.95602942e-01 3.71709079e-01 -5.40625870e-01
4.53549549e-02 4.11715597e-01 1.53133690e-01 2.02823997e-01
4.79308516e-01 2.18816057e-01 2.50061750e-01 -5.67251682e-01
3.82036239e-01 1.67842090e-01 4.62648124e-01 -1.09571803e+00
8.05624485e-01 4.68787134e-01 4.00444180e-01 -3.79601181e-01
-1.16530657e+00 -2.24394858e-01 -8.52879584e-02 2.16987301e-02
3.64853948e-01 -1.03131366e+00 -9.43734646e-01 -3.21287960e-01
-5.97767770e-01 -8.78010809e-01 -8.20140481e-01 8.67753506e-01
-1.28257883e+00 2.04936728e-01 -2.38072574e-01 -1.59664679e+00
1.31504178e-01 -9.67120826e-01 1.14724565e+00 -4.26762179e-02
-1.19580284e-01 -9.57340121e-01 3.81214380e-01 -7.79613480e-02
3.70427638e-01 2.86965579e-01 5.53097367e-01 -2.91661859e-01
-1.12147963e+00 -7.49530792e-02 1.48139298e-01 9.04320031e-02
-8.30561966e-02 -6.76898539e-01 -8.04502189e-01 -7.44345009e-01
1.27057105e-01 -7.53511310e-01 9.55443919e-01 6.74406111e-01
1.14822412e+00 -6.88278735e-01 -2.80999333e-01 4.82437581e-01
1.55147731e+00 1.42244533e-01 1.52428597e-01 4.75385189e-01
2.42907014e-02 2.61058599e-01 1.05665696e+00 1.09848881e+00
2.94006348e-01 5.85879207e-01 8.39084089e-01 4.54344600e-02
3.88535172e-01 -6.94324911e-01 7.01684892e-01 1.52219772e-01
1.73320189e-01 -3.10068317e-02 -5.55952966e-01 6.24312043e-01
-2.36552501e+00 -9.35205460e-01 6.42113805e-01 2.68993020e+00
8.63223255e-01 3.89445454e-01 3.29647839e-01 -2.23308235e-01
1.34687558e-01 1.11828968e-01 -1.21148670e+00 -2.37474039e-01
3.71154606e-01 1.50092199e-01 1.05908716e+00 8.19199443e-01
-8.55477214e-01 7.64816046e-01 7.26036406e+00 4.81329411e-01
-7.36601591e-01 1.78957716e-01 2.97301412e-01 -7.92594969e-01
-4.34345961e-01 3.75537395e-01 -1.11540639e+00 2.74769932e-01
1.23205376e+00 -2.54573226e-01 1.03447771e+00 1.16557598e+00
3.02359909e-01 -3.41336548e-01 -1.50005686e+00 1.06037617e+00
-2.89766252e-01 -1.30561757e+00 -3.11102957e-01 7.23204732e-01
8.18897665e-01 1.54097036e-01 1.89627409e-01 5.90992153e-01
1.20265543e+00 -1.08163738e+00 9.27216709e-01 6.42229974e-01
6.47492051e-01 -9.40833986e-01 2.26544052e-01 7.89245009e-01
-9.02294159e-01 -4.68117088e-01 -4.37618136e-01 -4.09204125e-01
1.70962811e-01 3.14041555e-01 -9.19357955e-01 6.17857166e-02
3.48078966e-01 5.13382494e-01 8.77606496e-02 7.80718982e-01
-2.54695356e-01 7.57656872e-01 -6.06979132e-01 -1.44158661e-01
6.13252640e-01 -1.47280425e-01 5.26265085e-01 8.42389762e-01
3.18631172e-01 -2.21920073e-01 7.03914642e-01 8.29087496e-01
-1.94925517e-02 -3.37906092e-01 -6.90142691e-01 -3.30327690e-01
6.39120400e-01 8.15560400e-01 -2.71739721e-01 -3.64993215e-01
-2.29184747e-01 4.61910486e-01 5.28300643e-01 3.72134686e-01
-7.93433666e-01 1.55411318e-01 9.40160453e-01 -7.50042796e-02
5.29407263e-01 -6.04929984e-01 -1.09750152e-01 -1.09614813e+00
1.39034197e-01 -7.13660538e-01 1.31222367e-01 -2.51684785e-01
-8.61928523e-01 -2.31408283e-01 7.50171915e-02 -1.16381073e+00
-9.05391216e-01 -4.01006550e-01 4.50687222e-02 6.34015203e-01
-1.80233967e+00 -6.40269876e-01 2.67867118e-01 4.01132077e-01
6.35010064e-01 -1.16278477e-01 8.76425683e-01 -3.23062837e-01
-1.14689775e-01 3.54933321e-01 6.12924099e-01 -3.50336760e-01
4.32084471e-01 -1.56989837e+00 1.43646672e-01 7.63581157e-01
1.38798445e-01 4.87456083e-01 8.18693757e-01 -6.11235261e-01
-1.90075564e+00 -9.82538462e-01 1.21534623e-01 -5.30525029e-01
6.32142603e-01 -3.95927250e-01 -2.02295467e-01 1.07926691e+00
-1.17153198e-01 1.26418918e-02 3.39644372e-01 3.81187081e-01
-1.61643580e-01 -9.32697207e-02 -9.45810616e-01 5.56592762e-01
9.38369751e-01 -3.51899385e-01 -3.04964066e-01 2.76375443e-01
7.25540876e-01 -5.45090735e-01 -8.67966235e-01 1.59143269e-01
6.29016876e-01 -6.61606789e-01 1.06387949e+00 -9.24862504e-01
-5.29988809e-03 -1.27059057e-01 -4.17596519e-01 -1.15771043e+00
-1.52063310e-01 -1.07618749e+00 -1.01867878e+00 7.44838059e-01
2.66134650e-01 -6.01709664e-01 1.02241588e+00 8.29083264e-01
1.13183536e-01 -7.83790052e-01 -7.96724916e-01 -1.26710415e+00
-5.36908209e-02 -5.23822129e-01 6.11618698e-01 4.67543811e-01
-3.02115202e-01 2.00237513e-01 -8.08202028e-01 2.71998107e-01
1.19980061e+00 3.78433883e-01 1.11243093e+00 -1.13653767e+00
-8.59331667e-01 -3.13301504e-01 2.14385241e-01 -1.74757195e+00
2.79635817e-01 -5.80145419e-01 1.07565179e-01 -1.50308585e+00
1.25416234e-01 -6.63653433e-01 -4.80632871e-01 4.26880687e-01
1.94660991e-01 -5.86761892e-01 2.66051173e-01 1.20498218e-01
-9.78091657e-01 8.42113376e-01 1.25761962e+00 9.50035527e-02
-5.46597898e-01 2.34736487e-01 -7.04067051e-01 6.47624075e-01
8.73175025e-01 -7.96830714e-01 -8.77276182e-01 -3.57541353e-01
3.85174155e-01 6.26039147e-01 1.76632822e-01 -8.43893647e-01
1.86668605e-01 -7.79235482e-01 3.57726604e-01 -6.39961421e-01
6.72249675e-01 -8.34382951e-01 -6.33421168e-02 6.63868666e-01
-8.17170382e-01 -1.19341336e-01 1.07082509e-04 1.18699503e+00
2.15032443e-01 -2.03840941e-01 6.88469410e-01 -4.74814296e-01
-3.05523634e-01 5.31117797e-01 -3.85631204e-01 1.89572856e-01
7.79803991e-01 8.08885992e-02 2.20856559e-03 -8.44039619e-01
-5.65187573e-01 4.91770446e-01 4.54955012e-01 1.56057268e-01
5.09140253e-01 -1.24558342e+00 -4.00675178e-01 -1.05358856e-02
-2.34464094e-01 1.96810052e-01 -3.54254723e-01 6.20161355e-01
1.86564669e-01 6.13740325e-01 8.00948143e-02 -4.41404790e-01
-7.81396210e-01 6.65588319e-01 2.18917444e-01 -8.24698865e-01
-4.57640976e-01 5.56023598e-01 8.64187256e-02 -5.75970531e-01
7.03453541e-01 -5.03876388e-01 3.71042252e-01 -3.95694554e-01
5.74874580e-01 4.12129998e-01 -5.53769231e-01 2.51833349e-01
8.80821124e-02 1.55254498e-01 7.80576840e-02 -7.68531919e-01
1.57227385e+00 -3.23331445e-01 4.16335553e-01 6.08500481e-01
8.25794339e-01 -5.19296192e-02 -2.25159860e+00 -6.09913588e-01
2.63573945e-01 -6.38760924e-01 3.16741228e-01 -7.29481220e-01
-6.27712965e-01 8.17662954e-01 6.37420297e-01 7.78345540e-02
7.24814117e-01 -2.30907753e-01 6.71243191e-01 9.44697380e-01
7.40006328e-01 -1.32520604e+00 3.67418915e-01 4.23378915e-01
5.74002326e-01 -1.22535264e+00 1.99701190e-01 3.51067513e-01
-4.10067886e-01 7.22951889e-01 4.14423943e-01 -2.18359858e-01
4.02621269e-01 4.79264319e-01 -3.83054346e-01 1.87044576e-01
-1.22155058e+00 -2.87211686e-01 -2.66067147e-01 6.11375809e-01
-7.43435621e-02 7.36303329e-02 1.76841065e-01 3.57055187e-01
-1.51592404e-01 1.96967483e-01 3.52727443e-01 1.38676488e+00
-7.97068954e-01 -1.35757160e+00 -4.14323658e-01 4.88126606e-01
-5.44453025e-01 1.46724626e-01 1.53606176e-01 6.70989871e-01
-6.91500843e-01 1.07049263e+00 9.00721177e-02 8.98617432e-02
6.88463748e-02 -1.01881482e-01 7.79649079e-01 -5.60176492e-01
2.92573106e-02 3.13612401e-01 1.91268295e-01 -1.04306912e+00
-4.40740436e-01 -8.52391541e-01 -1.23040450e+00 -2.59702981e-01
-3.43131870e-02 2.15506554e-01 7.98651636e-01 9.07507598e-01
4.49179441e-01 1.63677230e-01 8.27063262e-01 -9.21441197e-01
-1.51837695e+00 -6.01774693e-01 -6.68692529e-01 2.33044494e-02
8.17842722e-01 -8.23808610e-01 -1.10359028e-01 -3.00985247e-01] | [4.1476664543151855, 2.2184669971466064] |
b96d3bb6-fd67-4578-a5c9-547cb737868b | exploring-novel-prognostic-biomarkers-and | 2306.16184 | null | https://arxiv.org/abs/2306.16184v1 | https://arxiv.org/pdf/2306.16184v1.pdf | Exploring novel prognostic biomarkers and biologic processes involved in NASH, cirrhosis and HCC based on survival analysis using systems biology approach | There is an unmet need to develop medications or drug combinations which can stop advancement of NASH to liver cirrhosis and HCC. Therefore, identifying key biomarkers based on overall survival and the exploring biological processes involved in the pathogenesis and progression of NASH toward cirrhosis and HCC to improve therapeutic interventions is necessary. The microarray dataset from the GPL13667 platform was downloaded from the Gene Expression Omnibus (GEO). The inclusion criteria for the DEGs included an adjusted p-value <0.05 and a log(2) fold change >1. In the first step, protein-protein interaction (PPI) network of differentially expressed genes (DEGs) in every three groups (NASH, Cirrhosis and HCC) was constructed using STRING online database. In the second step, the DEGs of each group were imported to Cytoscape software separately, and the genes with Degree >3 were selected. In the third step, the genes with Degree >3 were imported to Gephi software (0.9.2 version) and the genes with betweenness centrality >0 were selected. Venn diagram of these genes was depicted for the NASH, cirrhotic and HCC groups. According to venn diagram, 96 (NASH), 30 (cirrhotic) and 213 (HCC) genes were specifically upregulated (based on inclusion criteria). Among specific genes, 22 (NASH), 5 (cirrhotic) and 82 (HCC) genes were with poor overall survival. The overlap among the 3 groups (NASH, Cirrhosis and HCC) contained 4 upregulated genes HLA-F, HLA-DPA1, TPM1 and YWHAZ. From 4 upregulated genes, only YWHAZ gene was with poor overall survival. The present study detected new candidate genes and key biological processes in NASH, cirrhosis and HCC based on overall survival and using in silico analysis. Therefore, performing in vitro and in vivo researches to verify the results is necessary. | ['Sedigheh Behrouzifar'] | 2023-06-28 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [-5.52480519e-01 -5.72426856e-01 -1.11604847e-01 -2.94953566e-02
-1.03614651e-01 -4.80723441e-01 4.27528750e-03 2.96447605e-01
2.08320439e-01 7.85647929e-01 4.21707243e-01 -5.02992451e-01
-2.71499753e-01 -9.04510438e-01 -5.87069942e-03 -1.25526392e+00
-8.96072507e-01 2.29787767e-01 -3.11768919e-01 -4.59150523e-02
5.97417029e-03 5.47619581e-01 -1.10103202e+00 1.31776422e-01
9.98207092e-01 4.13436770e-01 -1.43599793e-01 4.40710068e-01
-2.59612709e-01 -9.50360075e-02 7.20212832e-02 3.41634929e-01
3.62362087e-01 -9.83483911e-01 -2.96773404e-01 -3.87716621e-01
-4.81558830e-01 9.41553712e-02 -8.27704072e-02 1.11704683e+00
8.00886571e-01 -2.95711547e-01 5.90719163e-01 -1.36658859e+00
-2.68194258e-01 5.89932621e-01 -2.73621291e-01 -1.34849951e-01
1.91961527e-01 4.88320410e-01 4.10258472e-01 -9.82206762e-01
5.89481294e-01 1.25900924e+00 5.98652363e-01 2.49237880e-01
-9.75306153e-01 -7.36497104e-01 -5.47905803e-01 2.76586086e-01
-1.58768976e+00 -1.34979546e-01 2.42518678e-01 -6.69437528e-01
6.50252521e-01 6.27901852e-01 1.38602662e+00 1.73588648e-01
5.41846633e-01 -1.26429290e-01 1.38547146e+00 -1.95634335e-01
8.67135376e-02 -2.22352266e-01 -1.08368518e-02 6.88816845e-01
6.02995813e-01 1.84841886e-01 4.59806435e-02 -3.95008445e-01
4.71719801e-01 9.94343460e-02 -5.79442024e-01 -6.16297834e-02
-1.48623812e+00 6.04008973e-01 3.44810754e-01 7.49646306e-01
-3.59970689e-01 -1.66386068e-02 6.80712759e-01 3.86752397e-01
-4.44570370e-02 2.92226315e-01 -6.64608657e-01 1.79569289e-01
-2.57971585e-01 -4.51069742e-01 9.92502511e-01 5.23606718e-01
6.32756829e-01 -6.07622303e-02 -1.29729584e-01 4.54859376e-01
6.41057312e-01 2.24415377e-01 3.56301248e-01 -3.56291980e-01
-6.80764496e-01 1.00964499e+00 -1.34906486e-01 -1.17812312e+00
-6.11153245e-01 -4.28659797e-01 -1.30813277e+00 -1.05156749e-01
3.67856443e-01 -3.04569721e-01 -4.72633660e-01 1.44872415e+00
3.33189011e-01 1.64011091e-01 1.03159577e-01 1.11951637e+00
8.72676075e-01 3.94990951e-01 5.12541294e-01 -8.19536150e-01
1.82494104e+00 -4.19214100e-01 -6.77534282e-01 7.38281429e-01
8.24589133e-01 -9.02843773e-01 5.84746182e-01 7.17425719e-02
-4.11267370e-01 9.14391782e-03 -7.73347378e-01 4.47586775e-01
-4.82722223e-01 5.43576153e-03 6.07952178e-01 4.45867211e-01
-1.16411567e+00 5.80453515e-01 -7.12661088e-01 -1.21543813e+00
1.28341056e-02 3.47486079e-01 -5.04159331e-01 -1.62154719e-01
-1.52566493e+00 8.93484831e-01 1.80448622e-01 1.73743591e-01
-7.13383675e-01 -7.28100419e-01 -4.80928272e-01 2.24071071e-01
-3.34797174e-01 -1.34648216e+00 1.98195815e-01 -5.17127991e-01
-1.32801151e+00 7.13620186e-01 -1.26786754e-01 2.37662420e-01
8.13399404e-02 4.92209166e-01 -3.80263925e-01 9.15753767e-02
-2.14716285e-01 7.90863112e-02 -3.10532928e-01 -7.44555652e-01
-3.11176300e-01 -5.61944604e-01 -5.25573790e-01 2.64115125e-01
7.64579475e-02 3.54872733e-01 2.00573392e-02 -2.83522934e-01
3.70860219e-01 -9.83844876e-01 -4.57960010e-01 8.20280686e-02
-5.32986522e-01 -3.41771424e-01 6.78606808e-01 -7.57028162e-01
8.86360168e-01 -1.90998340e+00 -1.56753995e-02 4.18724239e-01
1.83430284e-01 5.66350184e-02 1.14760259e-02 6.72307134e-01
-4.12620366e-01 9.72506464e-01 3.78442228e-01 9.37472880e-01
-2.76790947e-01 -1.90350890e-01 7.66788840e-01 1.05130184e+00
-1.61688745e-01 7.55085230e-01 -1.11975920e+00 -3.13938916e-01
2.37618014e-01 5.30009866e-01 -2.07652852e-01 1.93078116e-01
8.32838267e-02 6.08965695e-01 -5.90937734e-01 1.04162002e+00
8.50413144e-01 -4.93269190e-02 6.86853766e-01 -5.95844865e-01
-5.50427198e-01 -2.22285315e-01 -6.97124839e-01 1.02383912e+00
1.73907325e-01 3.56437773e-01 2.99551755e-01 -6.43465757e-01
1.00680721e+00 6.96029961e-01 7.27674127e-01 1.04120448e-02
3.53434324e-01 2.75505155e-01 5.76122820e-01 -9.06481922e-01
-8.29577088e-01 -3.43823314e-01 2.21895888e-01 -7.56338090e-02
-2.74094820e-01 3.84073466e-01 2.26895183e-01 -1.73262451e-02
1.00984526e+00 -4.00408387e-01 6.74206913e-01 -7.99772799e-01
6.96590960e-01 2.02258855e-01 6.90035284e-01 1.93072036e-01
-3.36226344e-01 -2.99160033e-02 1.26467943e+00 -1.82033196e-01
-1.04130554e+00 -5.79183578e-01 -2.57430226e-01 2.44140446e-01
7.54570439e-02 -1.94298774e-01 -1.57373562e-01 -2.27931365e-01
2.14666370e-02 3.49600822e-01 -2.50677496e-01 1.20966688e-01
2.00734045e-02 -1.20033813e+00 5.75311422e-01 -2.95771211e-01
6.86579823e-01 -2.16685936e-01 2.76775062e-01 2.26785362e-01
8.72164071e-02 -4.79249051e-03 -1.17634147e-01 -2.12795824e-01
-8.88891816e-01 -1.64558244e+00 -3.44842851e-01 -9.41475511e-01
1.01451635e+00 1.56664670e-01 5.22399306e-01 4.74567294e-01
-4.80694890e-01 -1.98558614e-01 -2.48898894e-01 -2.47916207e-01
-3.28891337e-01 -4.46076483e-01 2.53902674e-01 -5.34942687e-01
3.43820482e-01 -7.61815429e-01 -1.01097190e+00 4.94065642e-01
-3.15460533e-01 7.19438046e-02 8.40789080e-01 8.62475932e-01
4.54981208e-01 1.37610838e-01 7.48239040e-01 -3.81082594e-01
3.44040722e-01 -1.15468025e+00 -5.95361650e-01 -1.26467403e-02
-8.26409996e-01 -6.38917506e-01 3.78335148e-01 1.06861264e-01
-4.30782020e-01 9.48760584e-02 -7.00177997e-02 -4.00168449e-02
-1.89052746e-01 1.07497776e+00 -6.50672495e-01 -2.30810363e-02
1.82336837e-01 1.88960835e-01 7.91935205e-01 -3.67454737e-01
-2.10534051e-01 5.19817948e-01 -2.72138983e-01 5.53653091e-02
4.48413521e-01 -3.42029221e-02 7.18610585e-01 -7.17658997e-01
2.42012009e-01 -6.38888061e-01 -2.16415212e-01 -2.61347771e-01
7.36837506e-01 -9.85793531e-01 -1.16615176e+00 5.01762092e-01
-6.25506580e-01 1.28146425e-01 6.23466969e-01 1.06175756e+00
4.27711084e-02 4.38654125e-01 -8.88185501e-01 -4.02608335e-01
-8.29502225e-01 -9.83218968e-01 -1.45922855e-01 2.75419086e-01
-1.21767610e-01 -6.88062012e-01 1.75902545e-01 -2.00903386e-01
6.87640607e-01 7.38115430e-01 1.62853098e+00 -9.46568012e-01
-5.12004733e-01 -2.12599441e-01 -3.33018631e-01 1.49744339e-02
3.61069232e-01 5.91419756e-01 -5.12500964e-02 -1.82108372e-01
-2.40237311e-01 4.74162102e-01 2.87597686e-01 3.27896237e-01
7.31889665e-01 -6.66616917e-01 -7.41687477e-01 6.86146855e-01
1.84639823e+00 7.24981189e-01 9.04437363e-01 9.77742374e-02
2.53251523e-01 4.16671723e-01 4.44605708e-01 1.52729407e-01
6.15936480e-02 2.58397937e-01 4.86912847e-01 -2.30147481e-01
-3.84852923e-02 1.50791168e-01 2.48887613e-01 7.54482448e-01
-1.83000579e-01 -1.39407590e-01 -1.04371023e+00 5.46489239e-01
-1.41417074e+00 -9.18972433e-01 -1.22233450e+00 2.11284423e+00
6.75925732e-01 -6.85644209e-01 -1.33064225e-01 -4.80941981e-01
9.13503349e-01 -7.28721082e-01 -3.79367501e-01 -1.65563464e-01
-3.60548109e-01 -1.88581705e-01 2.02476323e-01 2.54689187e-01
-6.19112551e-01 2.33238563e-01 5.90033340e+00 1.78308547e-01
-1.23433161e+00 -2.10510299e-01 7.46568978e-01 1.48065567e-01
-5.98495863e-02 6.18786991e-01 -7.06057474e-02 4.26682740e-01
7.79121935e-01 -7.67140567e-01 3.45642000e-01 5.76387107e-01
9.42632973e-01 -2.70769984e-01 -7.10263729e-01 5.45653641e-01
-4.39380586e-01 -1.31239152e+00 -3.39366555e-01 1.83811799e-01
6.25678778e-01 1.78734720e-01 -7.44026482e-01 3.42489034e-01
1.00101024e-01 -9.08916056e-01 -5.50726235e-01 9.04545605e-01
9.87272799e-01 -5.11736691e-01 1.53351033e+00 -6.00848570e-02
-9.52399015e-01 2.75403708e-01 -2.10750297e-01 1.51891646e-03
-2.06587479e-01 1.20276523e+00 -1.26171589e+00 7.40669370e-01
4.81551141e-01 5.94625115e-01 -2.55076826e-01 1.52135158e+00
-3.29799861e-01 6.82142556e-01 -1.45551667e-01 -3.20143133e-01
-2.60065377e-01 -4.88175660e-01 8.02459002e-01 1.09370661e+00
9.13958490e-01 6.10845089e-01 1.31522790e-01 6.36234820e-01
2.17733011e-01 6.30647659e-01 -2.70721883e-01 -1.34656549e-01
6.88548565e-01 1.65908551e+00 -6.76207721e-01 -3.81030023e-01
-3.02630842e-01 4.02313948e-01 -5.76307297e-01 3.10515970e-01
-3.88438582e-01 -4.84269738e-01 1.12026966e+00 2.35631093e-01
-4.33040470e-01 1.17154092e-01 -2.80540735e-01 -9.77885306e-01
-7.65630543e-01 -6.28224790e-01 4.29606736e-01 -5.27439237e-01
-1.32541430e+00 1.49335504e-01 -2.70295650e-01 -1.23152471e+00
1.82385892e-01 -3.26858461e-01 -8.97374749e-01 1.44773471e+00
-1.04369414e+00 -9.07070041e-01 -5.12628675e-01 -3.11689880e-02
-1.68593585e-01 1.82073772e-01 1.16761231e+00 2.15016037e-01
-9.29122329e-01 -2.25213617e-01 4.68560755e-01 9.86798555e-02
7.91642785e-01 -8.40579987e-01 -7.76804447e-01 2.35877752e-01
-1.36777318e+00 1.12197065e+00 7.70398438e-01 -8.62158835e-01
-1.77664995e+00 -1.16766703e+00 1.25334740e+00 4.31506902e-01
4.99474227e-01 -5.45707298e-04 -6.13258839e-01 3.00020754e-01
3.36248159e-01 -9.64498520e-02 1.21993411e+00 4.43861494e-03
1.87174797e-01 8.26886017e-03 -1.44256127e+00 7.64057875e-01
5.35152614e-01 9.45831910e-02 8.63525867e-02 6.13450468e-01
1.98527545e-01 -8.99965987e-02 -1.61835396e+00 5.18011749e-01
8.11342180e-01 -8.33940864e-01 6.32621944e-01 -6.78561509e-01
8.41863453e-02 -9.73113358e-01 1.37235284e-01 -1.50417387e+00
-7.12760687e-01 -2.87466437e-01 5.48664927e-01 1.10619795e+00
1.07402168e-01 -7.67377198e-01 3.04912150e-01 3.17981184e-01
-3.32749575e-01 -9.82873678e-01 -8.97477746e-01 -3.95520419e-01
-3.27150188e-02 6.45817161e-01 7.03398168e-01 1.30096304e+00
7.34524190e-01 -2.58021913e-02 1.69754326e-01 7.61113390e-02
3.24887276e-01 -4.09808531e-02 5.01572907e-01 -1.20292449e+00
4.69075561e-01 -5.55443287e-01 -5.93238354e-01 1.35819241e-01
-2.45062947e-01 -1.24765718e+00 -4.59028244e-01 -2.07404590e+00
5.41431546e-01 -5.97868621e-01 -3.76819074e-01 7.04438448e-01
-1.79718435e-02 -2.01474220e-01 -1.84414953e-01 2.35392839e-01
2.50640959e-01 1.20498382e-01 1.02261519e+00 1.72203198e-01
-2.21034735e-01 -2.51913637e-01 -8.92071605e-01 4.77950782e-01
1.17031419e+00 -3.04626673e-01 1.61480621e-01 4.71963465e-01
1.07936598e-01 4.38752621e-01 3.91864359e-01 -5.04167497e-01
1.64875045e-01 -5.76192617e-01 6.91892624e-01 -5.91802895e-01
-3.82778049e-01 -8.47980142e-01 1.13606477e+00 1.48011684e+00
1.67611971e-01 -1.54499918e-01 6.22963943e-02 1.16788000e-01
7.83899501e-02 1.20609097e-01 7.09591568e-01 -2.22265899e-01
-3.96763921e-01 1.96422949e-01 -9.65662122e-01 -7.46830583e-01
1.31971288e+00 -1.18267529e-01 -4.50587809e-01 -1.53193429e-01
-1.07931674e+00 4.34201717e-01 5.91904402e-01 -1.29350096e-01
3.71436268e-01 -1.41717470e+00 -1.10783505e+00 5.57195991e-02
-6.61915988e-02 -4.62785751e-01 5.37515342e-01 1.81109571e+00
-7.41378903e-01 5.74941456e-01 -5.20418942e-01 -4.30178076e-01
-1.47382402e+00 4.19582218e-01 6.04358315e-01 -7.75189474e-02
-7.12900311e-02 3.69266927e-01 -1.38108969e-01 -1.87810242e-01
-2.62516499e-01 5.09455055e-02 -4.41722393e-01 1.40888736e-01
2.13479728e-01 8.12180400e-01 -8.76112431e-02 -2.45437309e-01
-6.43282950e-01 2.34822378e-01 5.37784278e-01 6.23458326e-01
1.24194133e+00 -2.21079320e-01 -1.43271613e+00 1.33925661e-01
1.20849466e+00 3.20285439e-01 -2.45328760e-03 4.25737560e-01
-7.39360973e-02 -5.40317595e-01 -2.09741548e-01 -1.10055697e+00
-9.74363983e-01 8.57353508e-02 8.31403852e-01 -6.61917105e-02
1.26952529e+00 -1.45991206e-01 1.86566398e-01 -3.09575319e-01
3.91494006e-01 -1.33640587e-01 -8.13017786e-01 5.24364769e-01
9.00236905e-01 -7.58999109e-01 -1.42378360e-01 -7.41659105e-01
-1.47916386e-02 1.51928365e+00 5.68374813e-01 1.01156719e-02
8.27926755e-01 1.32348482e-02 -1.26584386e-02 -1.77405700e-01
-9.88839447e-01 2.01483056e-01 -4.13811654e-02 3.29660952e-01
1.00032341e+00 3.13632369e-01 -1.52358615e+00 7.85756588e-01
2.22192734e-01 2.79110730e-01 4.93018538e-01 4.83626634e-01
-5.11576653e-01 -1.02313077e+00 -4.69621629e-01 7.88501441e-01
-3.29445243e-01 -1.93555444e-01 -3.56478989e-01 1.11250520e+00
4.15914327e-01 8.69297862e-01 1.00510061e-01 -3.33079457e-01
8.59161392e-02 2.65122443e-01 -1.16544917e-01 -3.55616331e-01
-7.70272374e-01 5.92301726e-01 4.52005118e-01 -3.34126592e-01
-2.39441290e-01 -7.80133128e-01 -1.51201773e+00 -4.93818998e-01
-2.88631737e-01 5.77464044e-01 9.18136299e-01 4.03206676e-01
7.25359142e-01 2.86664426e-01 7.69247293e-01 -7.64227659e-02
-1.52345255e-01 -1.27995634e+00 -7.84973323e-01 3.85269895e-02
-1.37815505e-01 -1.70333669e-01 -8.83008361e-01 -1.53676435e-01] | [5.9919915199279785, 5.675660133361816] |
09e80ae1-c978-4958-b85a-8a5dde0198d2 | a-continuous-relaxation-of-beam-search-for | 1708.00111 | null | http://arxiv.org/abs/1708.00111v2 | http://arxiv.org/pdf/1708.00111v2.pdf | A Continuous Relaxation of Beam Search for End-to-end Training of Neural Sequence Models | Beam search is a desirable choice of test-time decoding algorithm for neural
sequence models because it potentially avoids search errors made by simpler
greedy methods. However, typical cross entropy training procedures for these
models do not directly consider the behaviour of the final decoding method. As
a result, for cross-entropy trained models, beam decoding can sometimes yield
reduced test performance when compared with greedy decoding. In order to train
models that can more effectively make use of beam search, we propose a new
training procedure that focuses on the final loss metric (e.g. Hamming loss)
evaluated on the output of beam search. While well-defined, this "direct loss"
objective is itself discontinuous and thus difficult to optimize. Hence, in our
approach, we form a sub-differentiable surrogate objective by introducing a
novel continuous approximation of the beam search decoding procedure. In
experiments, we show that optimizing this new training objective yields
substantially better results on two sequence tasks (Named Entity Recognition
and CCG Supertagging) when compared with both cross entropy trained greedy
decoding and cross entropy trained beam decoding baselines. | ['Taylor Berg-Kirkpatrick', 'Chris Dyer', 'Kartik Goyal', 'Graham Neubig'] | 2017-08-01 | null | null | null | null | ['ccg-supertagging'] | ['natural-language-processing'] | [ 3.83485734e-01 4.55112725e-01 -4.16447148e-02 -5.39633811e-01
-1.27833688e+00 -5.91809988e-01 6.23632550e-01 2.82115936e-01
-6.95014834e-01 9.98378217e-01 2.58023548e-03 -6.67211652e-01
1.40669465e-01 -5.74440718e-01 -8.48000705e-01 -7.44925916e-01
5.85213676e-02 6.85595453e-01 -3.70446295e-02 9.77666825e-02
8.60598087e-02 7.49906451e-02 -1.27151823e+00 -4.18098308e-02
1.02337754e+00 8.15795600e-01 1.86599255e-01 7.48589575e-01
-3.25787105e-02 4.95153368e-01 -3.75343919e-01 -8.89179528e-01
-9.19905677e-02 -8.75375807e-01 -9.31437314e-01 -3.34616244e-01
2.34776750e-01 4.40917872e-02 8.13611671e-02 1.17825603e+00
6.34833515e-01 2.69237578e-01 6.72774017e-01 -6.44393861e-01
-9.33886245e-02 1.17333114e+00 -9.15813968e-02 2.30369747e-01
3.18014115e-01 -2.23674238e-01 1.33803427e+00 -6.74231768e-01
6.15804613e-01 9.48961735e-01 8.74531329e-01 4.94205475e-01
-1.54468453e+00 -4.68634427e-01 -9.23016593e-02 3.38448919e-02
-1.31520772e+00 -6.11613095e-01 3.21205765e-01 -1.35753244e-01
1.14440966e+00 6.25979900e-01 4.35562968e-01 1.15801215e+00
3.09708137e-02 1.18622112e+00 9.21328664e-01 -5.72544932e-01
3.25921595e-01 3.05488795e-01 1.84626609e-01 7.90028155e-01
1.10105321e-01 2.21765101e-01 -2.93620795e-01 -2.19621211e-01
-6.51799934e-03 -7.09689021e-01 -5.90860784e-01 -4.14451808e-01
-1.05158842e+00 9.63668525e-01 7.92028233e-02 5.47874153e-01
-8.30982327e-02 2.30400965e-01 4.00220513e-01 4.95267481e-01
6.52582228e-01 4.90502030e-01 -5.95461667e-01 -6.80105686e-01
-1.34073842e+00 1.44505709e-01 1.09147203e+00 6.93962991e-01
6.78218126e-01 -1.35315403e-01 -4.37408984e-01 1.03225398e+00
4.43436444e-01 3.61188889e-01 5.38736045e-01 -4.54359561e-01
5.75218618e-01 1.74071401e-01 -2.06692219e-01 -4.13793564e-01
-4.23225760e-01 -9.39892232e-01 -7.03222632e-01 -3.11997652e-01
4.86995190e-01 -1.89451009e-01 -8.63189816e-01 2.02098942e+00
2.11404204e-01 9.13799647e-03 -9.55244005e-02 4.79664236e-01
2.89201140e-01 6.03430033e-01 1.30713567e-01 -6.31943882e-01
9.24992085e-01 -7.31545508e-01 -7.06788898e-01 -2.00530082e-01
1.45977163e+00 -5.15939653e-01 9.69316423e-01 1.91244200e-01
-1.52091300e+00 -1.36638567e-01 -9.06153023e-01 6.73895404e-02
-2.74030000e-01 -1.47985831e-01 2.29207218e-01 9.24705207e-01
-1.15981793e+00 8.88769507e-01 -9.54564512e-01 -2.42648378e-01
5.16915740e-03 3.51651460e-01 -1.47657946e-01 1.44091457e-01
-1.32038677e+00 1.26572967e+00 8.22318256e-01 7.72449151e-02
-4.54541653e-01 -5.34377992e-01 -8.77124548e-01 2.59011835e-01
1.84245363e-01 -6.97705567e-01 1.51883364e+00 -7.34830618e-01
-1.53640163e+00 9.60827947e-01 -1.76562384e-01 -6.91911340e-01
7.08847582e-01 -5.43853678e-02 -1.60804331e-01 -4.37328756e-01
-2.70544022e-01 3.90636176e-01 4.41412598e-01 -9.07330990e-01
-4.00617570e-01 -2.64001161e-01 -2.90982604e-01 2.44690493e-01
-1.29526764e-01 2.09338982e-02 -4.55149680e-01 -6.82493091e-01
-1.39304891e-01 -1.03138864e+00 -6.79392293e-02 -6.33981228e-01
-6.70255899e-01 -2.20407650e-01 2.60397941e-01 -8.19073975e-01
1.75322783e+00 -2.09778976e+00 4.03606623e-01 4.66142982e-01
-2.28807330e-01 3.97496760e-01 6.92425817e-02 2.86193132e-01
-2.96012033e-02 4.01789039e-01 -6.80407763e-01 -6.55626297e-01
2.69199550e-01 1.33502185e-01 4.64188345e-02 5.08361101e-01
-4.32426706e-02 9.71960008e-01 -7.91000783e-01 -4.27152097e-01
-8.23957622e-02 3.36674929e-01 -8.64852071e-01 1.69495985e-01
-2.71594852e-01 2.25804999e-01 -2.00259745e-01 3.80371273e-01
4.79815155e-01 -3.07545334e-01 4.15318549e-01 2.76692331e-01
-1.85398366e-02 6.84604645e-01 -7.24178195e-01 1.68486845e+00
-7.19732821e-01 6.30366147e-01 -2.40766108e-01 -1.07942343e+00
6.61050916e-01 3.71208072e-01 8.07196274e-02 -7.32117891e-01
1.94447905e-01 4.19757575e-01 -3.44930366e-02 -2.58783728e-01
4.79291409e-01 -2.31723011e-01 -3.99272107e-02 5.23854613e-01
9.03303698e-02 1.44104749e-01 5.19655086e-02 -1.80329606e-01
1.17983258e+00 1.51387155e-01 5.38019896e-01 -2.29421079e-01
4.96013075e-01 -4.03490216e-01 5.53633988e-01 1.03511727e+00
4.54966687e-02 5.32084942e-01 4.73638296e-01 1.01335719e-01
-1.10244918e+00 -8.36884618e-01 -4.37817901e-01 1.08854353e+00
-2.41894081e-01 -5.08032382e-01 -1.00713360e+00 -9.04337287e-01
-2.32357085e-01 1.16822875e+00 -5.81551433e-01 -4.22006756e-01
-8.33871543e-01 -1.07740700e+00 7.86209404e-01 2.76682884e-01
4.21575010e-01 -8.16559196e-01 -3.99032325e-01 3.28612924e-01
-3.89361918e-01 -7.45074153e-01 -6.03693902e-01 6.57721758e-01
-9.65937018e-01 -6.20227695e-01 -8.18759143e-01 -6.19944096e-01
2.52259940e-01 -4.73598093e-01 1.41516709e+00 1.26655325e-01
1.78367317e-01 -4.10419255e-02 -5.19571424e-01 -2.67314427e-02
-7.84295261e-01 6.48208320e-01 -3.24923337e-01 -1.30272359e-01
2.94556767e-01 -4.58216906e-01 -3.63593131e-01 1.45493314e-01
-7.23888397e-01 9.93976742e-03 6.94238305e-01 1.17812276e+00
5.01733065e-01 -4.57810432e-01 3.35418850e-01 -1.03827715e+00
5.95406353e-01 -5.37257910e-01 -6.58938229e-01 7.32572913e-01
-8.95026028e-01 7.68547952e-01 4.06771064e-01 -2.35344067e-01
-8.08773935e-01 -1.71340197e-01 -6.84839010e-01 6.68664789e-03
1.02544114e-01 6.99733734e-01 -1.45740211e-01 -6.42703399e-02
4.96167004e-01 4.91749018e-01 -1.49485931e-01 -7.50866771e-01
9.34549049e-02 6.50330961e-01 1.51994050e-01 -2.52817601e-01
4.11889464e-01 -3.35459471e-01 -1.72754854e-01 -5.33495843e-01
-6.71072066e-01 -3.23477685e-01 -2.24673375e-01 -5.40903881e-02
9.22507107e-01 -4.51486558e-01 -6.55434132e-01 2.70221949e-01
-9.72054541e-01 -4.57465291e-01 -5.99690713e-02 5.41842580e-01
-8.80360126e-01 4.40803021e-01 -4.08954203e-01 -1.11140335e+00
-3.14264089e-01 -1.18405974e+00 1.18377113e+00 -2.37067431e-01
-2.82185048e-01 -1.25485408e+00 3.24524522e-01 -5.85772060e-02
3.30945194e-01 -3.80523726e-02 1.05380964e+00 -1.06769085e+00
-3.69539142e-01 -2.98537612e-01 1.29384443e-01 4.16009068e-01
-4.13376242e-01 -2.44142368e-01 -1.08591890e+00 -5.29712975e-01
1.08358450e-01 -2.27234900e-01 1.12626910e+00 2.65690535e-01
1.17478502e+00 -3.51624846e-01 -4.03412700e-01 9.47598577e-01
1.43419325e+00 2.24244803e-01 6.44600928e-01 4.23167050e-01
3.16663474e-01 2.68473834e-01 4.81438071e-01 2.47970790e-01
2.01913372e-01 1.10155964e+00 1.15117535e-01 3.46561998e-01
1.81255743e-01 -3.85093629e-01 4.46414441e-01 9.44197595e-01
4.53022569e-02 -6.96168005e-01 -1.03415823e+00 2.76631236e-01
-1.67799199e+00 -9.83293891e-01 3.67498845e-02 2.58714795e+00
1.12165320e+00 2.23347202e-01 7.42398202e-02 2.61906862e-01
5.81382573e-01 3.04847211e-02 -3.49674284e-01 -6.39002740e-01
-2.11620823e-01 2.79197782e-01 6.68143928e-01 7.78510273e-01
-1.11324978e+00 6.11781061e-01 6.66604042e+00 9.36351895e-01
-8.21558058e-01 2.94815183e-01 6.06522739e-01 -7.49875903e-02
-6.14907682e-01 -1.12333875e-02 -9.74687457e-01 7.73788869e-01
1.56092286e+00 -5.87775707e-02 3.65553677e-01 6.60248816e-01
-1.64335847e-01 2.36227643e-02 -1.43267775e+00 9.63777483e-01
-1.24233827e-01 -1.08489943e+00 -2.79067844e-01 2.91806757e-01
4.51638550e-01 1.91579401e-01 4.16340344e-02 4.60545003e-01
3.09418678e-01 -9.19507980e-01 8.74103546e-01 3.40602636e-01
7.22755969e-01 -7.39605725e-01 7.38434196e-01 6.22021437e-01
-6.90985918e-01 -6.26225071e-03 -1.12930886e-01 4.38176036e-01
3.50424767e-01 8.04379642e-01 -9.89390969e-01 4.87854689e-01
1.98795930e-01 2.79386103e-01 -3.61665159e-01 1.39530098e+00
-4.10592230e-03 9.44305718e-01 -4.32819605e-01 -3.43874007e-01
3.64231527e-01 -2.88469214e-02 7.59291708e-01 1.62024236e+00
5.57360351e-01 -2.92287976e-01 -2.92499274e-01 5.97333848e-01
-2.86529928e-01 2.70379215e-01 -7.36850321e-01 -1.02364577e-01
3.08102787e-01 6.22707486e-01 -4.16526854e-01 -1.58476934e-01
-2.33562365e-01 9.34055328e-01 6.54752493e-01 2.27324963e-01
-9.61524546e-01 -2.33887255e-01 5.36939502e-01 -1.91462725e-01
5.06358266e-01 -1.18065022e-01 -3.72798026e-01 -1.33359933e+00
1.30971342e-01 -7.94205248e-01 4.82184201e-01 -2.96729475e-01
-7.39848375e-01 8.32263529e-01 -6.80645034e-02 -9.88424063e-01
-9.08161044e-01 -2.67349601e-01 -1.94387540e-01 8.93113315e-01
-1.49837124e+00 -5.24649918e-01 3.18661809e-01 1.11814745e-01
3.46756905e-01 1.06006965e-01 9.86233890e-01 4.47110176e-01
-6.17727757e-01 1.35047150e+00 6.66473925e-01 -4.08302248e-02
2.99426764e-01 -1.45163500e+00 5.14744222e-01 8.58292162e-01
3.46986353e-01 4.40976083e-01 8.94380510e-01 -4.80384111e-01
-1.06436586e+00 -8.96343946e-01 1.46862853e+00 -3.86013359e-01
4.63361412e-01 -4.83938992e-01 -1.05816376e+00 7.45793581e-01
5.30224033e-02 -4.71412301e-01 7.30992556e-01 4.00023013e-01
-1.51166931e-01 2.28419587e-01 -9.29174066e-01 4.72474158e-01
1.06963623e+00 -5.85227966e-01 -4.99975353e-01 3.58304083e-01
5.65737486e-01 -3.28711867e-01 -7.67714024e-01 5.38774848e-01
4.80797738e-01 -9.97701287e-01 7.07788527e-01 -5.41701734e-01
-5.87137043e-02 2.41007641e-01 -1.98870957e-01 -1.45756924e+00
-1.08610027e-01 -8.56523931e-01 -2.48606786e-01 1.06599593e+00
9.40253377e-01 -6.97282910e-01 6.98090851e-01 5.43335676e-01
-1.97080433e-01 -9.56864536e-01 -1.14620328e+00 -9.93514597e-01
2.27441713e-01 -6.03492320e-01 4.07799572e-01 7.35208511e-01
2.84006238e-01 9.16924551e-02 -3.50609422e-01 -2.01756269e-01
5.89673519e-01 -1.27568156e-01 2.37400368e-01 -9.75160658e-01
-6.35341227e-01 -8.25655878e-01 -4.40489531e-01 -1.25179124e+00
3.32943976e-01 -1.24287009e+00 4.90590781e-01 -1.07222545e+00
1.04406729e-01 -5.82351744e-01 -3.14610928e-01 3.20797801e-01
-3.87089014e-01 7.58979917e-02 -1.46440743e-02 -4.18446586e-02
-4.26448852e-01 5.36123276e-01 7.85010099e-01 1.77079767e-01
-1.77962240e-02 2.66342849e-01 -3.30183804e-01 3.50892454e-01
6.59162879e-01 -6.27476037e-01 -1.94279224e-01 -4.78178024e-01
6.15407586e-01 3.38785321e-01 5.55352867e-02 -7.25290060e-01
2.64533609e-01 3.32955629e-01 -1.88914865e-01 -4.25452411e-01
2.29682639e-01 -4.87618506e-01 2.90966928e-01 3.98889869e-01
-7.50928402e-01 1.30487327e-02 -4.05703299e-02 4.57545519e-01
-3.17573935e-01 -5.85506201e-01 7.71077335e-01 -3.45743410e-02
-1.11852266e-01 -6.20018877e-02 -2.62287259e-01 1.94895178e-01
6.40572309e-01 -2.85623312e-01 1.53393611e-01 -3.16799372e-01
-1.01158094e+00 1.54283911e-01 3.95315975e-01 -3.22919227e-02
2.62662679e-01 -1.24647498e+00 -8.20072472e-01 3.18449587e-01
3.10976580e-02 -2.64489770e-01 -1.15608849e-01 1.19668448e+00
-3.48875880e-01 7.41460323e-01 4.08040941e-01 -4.44680154e-01
-1.26705348e+00 3.35023105e-01 5.04351914e-01 -6.87779367e-01
-3.65470052e-01 1.18583512e+00 8.10440555e-02 -4.66141760e-01
4.71943170e-01 -3.26360166e-01 7.67024457e-02 2.24676467e-02
2.83920884e-01 2.14861482e-01 5.10088563e-01 -4.32723075e-01
-3.84444684e-01 2.60301322e-01 -5.59952557e-02 -2.65623629e-01
1.15613282e+00 -2.55595893e-01 7.19721392e-02 5.83214641e-01
1.68502152e+00 -2.10944906e-01 -8.32371950e-01 -2.67414540e-01
4.50534135e-01 -3.20806921e-01 1.70453325e-01 -8.61540139e-01
-7.70118833e-01 7.70347953e-01 4.50385422e-01 4.55293924e-01
1.21640718e+00 -6.14253469e-02 7.10953534e-01 4.99930769e-01
2.93582290e-01 -1.00659072e+00 -7.32593954e-01 7.13981628e-01
4.79208529e-01 -1.18688285e+00 -5.66538930e-01 -1.85570553e-01
-2.21186429e-01 9.41160977e-01 3.67940478e-02 4.54136014e-01
5.77889502e-01 2.78464943e-01 -2.14490876e-01 7.82744214e-02
-8.62696946e-01 -2.05156893e-01 4.56031799e-01 1.38583854e-01
7.24286258e-01 1.45971985e-03 -6.87543392e-01 1.40669569e-01
-4.58867967e-01 -8.91046003e-02 7.72025660e-02 7.48509884e-01
-4.25202608e-01 -1.21355212e+00 7.82974884e-02 4.75864351e-01
-7.97867239e-01 -6.23056650e-01 -1.86272234e-01 5.92005193e-01
-3.13284129e-01 6.24897659e-01 -1.35435862e-02 -6.33981705e-01
5.93191832e-02 6.63579881e-01 4.55888927e-01 -3.57881576e-01
-6.22411847e-01 -1.89938262e-01 4.98061717e-01 -4.59491730e-01
-2.10646942e-01 -9.14371729e-01 -7.31953859e-01 -2.98215926e-01
-7.60840178e-01 5.98989248e-01 9.59923506e-01 1.16025412e+00
9.46860462e-02 2.51884311e-01 5.42567968e-01 -4.55015868e-01
-9.39671576e-01 -1.02728486e+00 -3.99545938e-01 2.40128294e-01
3.65825802e-01 -5.28367937e-01 -4.18626755e-01 -3.69338870e-01] | [11.674718856811523, 9.569070816040039] |
c0ad57d5-8a2c-434b-91b0-1cfee5a5f427 | myriad-a-multi-array-room-acoustic-database | 2301.13057 | null | https://arxiv.org/abs/2301.13057v2 | https://arxiv.org/pdf/2301.13057v2.pdf | MYRiAD: A Multi-Array Room Acoustic Database | In the development of acoustic signal processing algorithms, their evaluation in various acoustic environments is of utmost importance. In order to advance evaluation in realistic and reproducible scenarios, several high-quality acoustic databases have been developed over the years. In this paper, we present another complementary database of acoustic recordings, referred to as the Multi-arraY Room Acoustic Database (MYRiAD). The MYRiAD database is unique in its diversity of microphone configurations suiting a wide range of enhancement and reproduction applications (such as assistive hearing, teleconferencing, or sound zoning), the acoustics of the two recording spaces, and the variety of contained signals including 1214 room impulse responses (RIRs), reproduced speech, music, and stationary noise, as well as recordings of live cocktail parties held in both rooms. The microphone configurations comprise a dummy head (DH) with in-ear omnidirectional microphones, two behind-the-ear (BTE) pieces equipped with 2 omnidirectional microphones each, 5 external omnidirectional microphones (XMs), and two concentric circular microphone arrays (CMAs) consisting of 12 omnidirectional microphones in total. The two recording spaces, namely the SONORA Audio Laboratory (SAL) and the Alamire Interactive Laboratory (AIL), have reverberation times of 2.1s and 0.5s, respectively. Audio signals were reproduced using 10 movable loudspeakers in the SAL and a built-in array of 24 loudspeakers in the AIL. MATLAB and Python scripts are included for accessing the signals as well as microphone and loudspeaker coordinates. The database is publicly available at [1]. | ['Toon van Waterschoot', 'Maja Taseska', 'Randall Ali', 'Thomas Dietzen'] | 2023-01-30 | null | null | null | null | ['audio-signal-processing'] | ['audio'] | [ 7.10359514e-02 -5.81895888e-01 1.20279217e+00 -3.45826484e-02
-1.19035149e+00 -7.22674012e-01 1.40619963e-01 -2.55289793e-01
-3.16068441e-01 2.06437930e-01 4.15956944e-01 -3.95368785e-01
-1.84706658e-01 -2.47626781e-01 -3.53434771e-01 -9.77405250e-01
-3.18764597e-01 -1.77534923e-01 1.80717781e-01 -8.84175859e-03
-1.92411616e-01 4.50585127e-01 -1.82094347e+00 1.82124659e-01
2.41173431e-01 1.18303263e+00 5.53410053e-01 1.22788334e+00
5.95204830e-01 2.21088648e-01 -1.00671017e+00 -1.13036156e-01
1.84383973e-01 -1.85828254e-01 2.72970080e-01 -3.35221469e-01
1.19303666e-01 -1.42072260e-01 -1.66129559e-01 6.82132185e-01
1.37562025e+00 3.48557532e-01 3.80210936e-01 -9.23888505e-01
-1.23047054e-01 4.71791089e-01 -4.11572367e-01 1.15907915e-01
7.04037547e-01 -9.91281047e-02 8.37484717e-01 -1.10486674e+00
-1.24172963e-01 1.19120836e+00 7.03802586e-01 3.77080202e-01
-8.66759777e-01 -8.85759473e-01 -4.80745137e-01 -6.01809584e-02
-1.48912680e+00 -1.09618998e+00 9.49520528e-01 -2.85107970e-01
7.51556933e-01 8.36348057e-01 4.66010422e-01 1.09298170e+00
-2.01324686e-01 2.95967638e-01 1.04736400e+00 -7.90373385e-01
4.80797619e-01 3.01164806e-01 -1.40331134e-01 -3.65389697e-02
-9.87043232e-02 1.24161161e-01 -4.40295547e-01 -5.88895261e-01
6.92040563e-01 -2.91375130e-01 -7.09302008e-01 -1.65367946e-02
-1.04172122e+00 2.82826632e-01 -1.43268898e-01 6.24625206e-01
-1.62127703e-01 -1.09905846e-01 2.83304006e-01 2.19365731e-01
1.42945468e-01 2.80167550e-01 -2.45237619e-01 -1.03230692e-01
-3.76025707e-01 1.70225441e-01 1.00015426e+00 8.74080539e-01
8.01948234e-02 4.55107450e-01 2.50377566e-01 1.72588003e+00
7.36304998e-01 7.64070272e-01 5.25733769e-01 -7.81617045e-01
5.79597831e-01 -5.51449835e-01 1.82401508e-01 -1.14165163e+00
-4.24338400e-01 -6.02342546e-01 -8.26645017e-01 -6.97910488e-02
2.61198252e-01 -4.15307343e-01 -4.36499298e-01 1.50541806e+00
4.39106762e-01 2.43030772e-01 1.07518725e-01 7.77228713e-01
8.97205293e-01 7.84532785e-01 -4.75632757e-01 -4.89825398e-01
1.39578414e+00 -7.33687758e-01 -9.53872204e-01 -1.02787085e-01
4.39284457e-04 -1.38844168e+00 1.13855326e+00 5.93683243e-01
-1.15280247e+00 -7.75961280e-01 -9.15794194e-01 4.89598334e-01
-4.11157943e-02 -2.41184175e-01 -1.64196417e-01 1.13166606e+00
-1.07910991e+00 -1.67262405e-01 -7.25063205e-01 1.16345592e-01
-4.48924422e-01 1.05591364e-01 -2.65659899e-01 1.11921735e-01
-9.94622171e-01 3.04416567e-01 -6.29444540e-01 5.46950042e-01
-9.50404823e-01 -6.12736821e-01 -9.69802737e-01 -2.71286606e-03
-1.10573523e-01 -3.16062570e-01 1.53435278e+00 -3.15533876e-01
-1.74955273e+00 4.48313326e-01 -1.34563550e-01 1.16815619e-01
2.50302255e-01 -2.03194797e-01 -1.20784748e+00 -8.68946612e-02
-2.52326485e-02 -2.65635014e-01 7.72922993e-01 -1.51400447e+00
-3.91547948e-01 -3.62068564e-01 -4.75201607e-01 2.02979907e-01
-1.99440390e-01 6.34286165e-01 -2.44282648e-01 -8.57087553e-01
4.34324682e-01 -6.84328854e-01 -1.40311062e-01 -5.94861686e-01
-6.04997158e-01 4.71812785e-01 6.29020035e-01 -7.69636095e-01
1.39044285e+00 -2.79540277e+00 -2.55393922e-01 4.54182267e-01
-3.06286395e-01 1.93249404e-01 -1.87428266e-01 6.02521122e-01
-2.07874030e-01 -2.98709869e-01 -4.52938043e-02 -7.65017569e-01
7.00352490e-02 -4.90540676e-02 -3.11244637e-01 6.30208969e-01
-7.12181509e-01 5.00976555e-02 -5.64852774e-01 -1.49866611e-01
3.14187258e-01 6.01111650e-01 -5.44124305e-01 6.02988660e-01
7.05290854e-01 6.91581130e-01 -2.67398626e-01 8.08662772e-01
8.33491921e-01 7.86736429e-01 -2.89526880e-01 6.29445538e-02
-4.02444243e-01 3.35376173e-01 -1.90119886e+00 1.17457271e+00
-8.44213307e-01 4.30825979e-01 1.13580251e+00 -4.13022995e-01
1.13214493e+00 8.95926058e-01 6.04580455e-02 -4.52753186e-01
-1.30970806e-01 6.13654137e-01 1.56574756e-01 -7.10690439e-01
3.32200646e-01 -2.37483919e-01 3.80457565e-02 3.78487289e-01
-2.14952499e-01 -5.26246190e-01 -1.13015190e-01 -3.13171238e-01
1.04299903e+00 -6.61573291e-01 2.25071356e-01 -4.50715199e-02
6.51708484e-01 -1.00789976e+00 5.92924595e-01 5.97770393e-01
-3.28044713e-01 8.95615876e-01 -3.27635437e-01 2.32466608e-01
-7.49074996e-01 -1.40446973e+00 -2.33932227e-01 1.08019686e+00
-2.58536965e-01 -2.98456073e-01 -4.68600601e-01 6.53082192e-01
-2.90343583e-01 7.21710563e-01 3.40186767e-02 3.16398203e-01
-6.46500409e-01 -2.34092459e-01 7.54691720e-01 2.74688900e-01
3.10697734e-01 -1.07042944e+00 -3.12236309e-01 1.57617331e-01
-2.62753278e-01 -9.61689234e-01 -5.95201433e-01 2.90195018e-01
-3.27598482e-01 -6.05984449e-01 -8.34006608e-01 -9.61516857e-01
2.35978469e-01 3.70322645e-01 7.71925271e-01 -6.32547796e-01
-3.23035181e-01 8.44027817e-01 -3.27545524e-01 -6.38603389e-01
-6.82213962e-01 -7.19433546e-01 5.88488400e-01 2.63888538e-01
-2.29415432e-01 -9.96590793e-01 -6.34223342e-01 8.39936137e-01
-7.27841198e-01 -5.75570285e-01 1.54293612e-01 6.47384226e-01
3.45814526e-01 3.72026324e-01 6.35860980e-01 -5.47889471e-02
7.92710841e-01 -3.89540523e-01 -4.69775617e-01 -2.36362338e-01
2.68607080e-01 -9.68880534e-01 8.44290197e-01 -3.37946266e-01
-1.22978544e+00 -2.96545833e-01 -7.61460423e-01 -2.18438312e-01
-6.71363831e-01 1.45245403e-01 -7.89411604e-01 2.26792112e-01
5.42825520e-01 2.22327456e-01 -3.24056327e-01 -7.49584436e-01
-1.05425511e-02 1.41296518e+00 8.37427914e-01 -2.70922840e-01
7.41444051e-01 7.01609720e-03 -4.40043449e-01 -1.34477460e+00
-1.19087165e-02 -8.21670830e-01 -1.12428218e-01 -3.20173949e-01
5.12210667e-01 -8.99950743e-01 -6.16944015e-01 1.01810884e+00
-9.67929184e-01 -1.32450104e-01 -2.91172490e-02 1.05872846e+00
-3.86320025e-01 2.19126999e-01 -6.84912205e-01 -1.39374471e+00
-2.19739407e-01 -1.28101146e+00 8.78507972e-01 1.41391531e-01
-2.90604562e-01 -8.09007347e-01 2.13388607e-01 5.44749498e-01
6.58605576e-01 2.22876668e-03 6.42478824e-01 -4.29780960e-01
5.16271889e-02 -2.92907000e-01 6.37494326e-01 6.54429078e-01
4.54814941e-01 -9.97539610e-02 -1.66335881e+00 -3.80059034e-01
5.42933226e-01 -1.04232445e-01 2.13503033e-01 8.56467605e-01
8.53754044e-01 -1.23827666e-01 -1.16839737e-01 4.26025301e-01
9.86956418e-01 9.92922127e-01 5.60911536e-01 7.71158142e-03
1.54235825e-01 7.39150763e-01 3.09206575e-01 6.57265961e-01
-3.18788663e-02 6.10006273e-01 5.01285911e-01 -1.34657145e-01
8.04160442e-03 -4.16652299e-02 4.64392275e-01 1.50013649e+00
1.13048762e-01 -3.90718639e-01 -5.82832158e-01 4.64672565e-01
-7.72659361e-01 -8.57033253e-01 -2.38977820e-01 2.45564222e+00
6.98630989e-01 -3.34326595e-01 3.50670307e-03 7.25551546e-01
8.38763535e-01 2.36100033e-01 -1.97813079e-01 -6.71614349e-01
-5.36865629e-02 1.39253467e-01 -1.19314035e-02 8.27988148e-01
-8.95873487e-01 -1.32241219e-01 5.89361668e+00 5.68795562e-01
-8.57549667e-01 5.44977374e-02 2.84740597e-01 -7.21971020e-02
-3.58840346e-01 -5.60564578e-01 -6.25392795e-01 4.30513978e-01
1.00481224e+00 1.64768249e-01 5.89568734e-01 7.08443046e-01
5.12066245e-01 -1.16137899e-01 -9.78380263e-01 1.20663333e+00
-6.57711998e-02 -6.21216357e-01 -6.99440598e-01 -1.04886845e-01
2.74912149e-01 -1.43405691e-01 4.38522577e-01 2.97323801e-02
-9.33473036e-02 -7.74099350e-01 8.35103691e-01 1.72199801e-01
8.15244377e-01 -5.67139685e-01 5.10873079e-01 5.09226859e-01
-1.37818778e+00 -2.41200238e-01 -1.65075362e-01 -5.79351969e-02
3.06493789e-01 6.90704405e-01 -7.29874671e-01 5.24019897e-01
1.13524520e+00 -4.83863503e-02 1.65590718e-01 1.32206976e+00
6.18979260e-02 1.04931903e+00 -6.08548045e-01 9.62767154e-02
-2.47382179e-01 -2.10949451e-01 9.64945078e-01 1.25230074e+00
5.70107341e-01 1.78749561e-01 -3.56877029e-01 4.26360935e-01
1.61875024e-01 1.57679155e-01 -5.91427743e-01 4.80710715e-01
8.84943426e-01 1.29776359e+00 -2.42462024e-01 2.49801859e-01
-3.70660216e-01 3.68060827e-01 -7.68898129e-01 7.37190604e-01
-5.76957345e-01 -8.95878375e-01 7.77187586e-01 8.29834584e-03
1.39705211e-01 -3.41478914e-01 -3.44161429e-02 -3.86247605e-01
1.61917642e-01 -1.06854224e+00 6.86585754e-02 -9.45541501e-01
-1.12100995e+00 7.95545816e-01 -7.46012032e-02 -1.36529100e+00
-3.22087169e-01 -4.65679377e-01 -8.37514758e-01 1.22619665e+00
-8.92403781e-01 -6.25621259e-01 -1.97655022e-01 6.99755430e-01
5.01006007e-01 -1.98938414e-01 1.13713181e+00 9.24035847e-01
-5.29946268e-01 7.33185470e-01 6.72378540e-01 1.16096042e-01
8.31148267e-01 -8.75676394e-01 7.40784407e-02 5.40520191e-01
-7.34757036e-02 8.15338254e-01 1.11687934e+00 1.36238992e-01
-1.41915798e+00 -8.13491881e-01 7.62444675e-01 -3.52716446e-02
5.83182633e-01 -9.43691790e-01 -8.05694401e-01 4.54156667e-01
6.88625425e-02 -2.62065139e-02 1.26249886e+00 -1.31887496e-01
-2.02839732e-01 -4.84895617e-01 -9.80212569e-01 6.39681220e-01
6.78414404e-01 -3.87548000e-01 -3.74704391e-01 1.51121289e-01
5.30988216e-01 -5.75459301e-01 -5.49260199e-01 1.71600178e-01
6.76842511e-01 -1.14663601e+00 9.88622129e-01 3.38484973e-01
-1.31046683e-01 -4.42344189e-01 -8.64341795e-01 -1.42534494e+00
-7.44320527e-02 -9.37648416e-01 3.18573564e-01 1.72518897e+00
3.57727677e-01 -1.09536517e+00 6.64272383e-02 3.44429404e-01
-6.81903183e-01 -5.30156851e-01 -1.23873258e+00 -7.62955904e-01
-2.83663988e-01 -9.11845744e-01 4.74890858e-01 3.47843468e-01
-8.23995024e-02 2.32971877e-01 -3.28369439e-01 5.38088202e-01
5.03819168e-01 -3.09579372e-01 7.82668948e-01 -7.61495948e-01
-7.09652185e-01 -2.49257877e-01 -1.98546454e-01 -1.19129813e+00
-2.98037589e-01 -2.66448706e-01 2.18618408e-01 -1.21247411e+00
-4.63912129e-01 -6.60569847e-01 -2.55240023e-01 -3.79816554e-02
1.46350548e-01 -3.55756804e-02 -3.87629755e-02 -2.06173748e-01
1.32227018e-01 4.27090257e-01 9.47390437e-01 -6.12777807e-02
-3.51463765e-01 8.72634649e-01 -4.53920096e-01 1.02097631e+00
4.12982166e-01 -2.01249495e-01 -3.92792523e-01 -3.99548531e-01
-3.16577464e-01 4.40470487e-01 1.59945890e-01 -1.12416613e+00
5.09847999e-01 4.25440729e-01 4.31843437e-02 -6.29209578e-01
8.27740550e-01 -9.53439832e-01 5.39190531e-01 2.03969151e-01
-2.40227386e-01 5.81345847e-03 3.99686068e-01 4.37515169e-01
-4.91293877e-01 -8.01123753e-02 8.28144848e-01 5.37915640e-02
-9.79231670e-02 -2.51032084e-01 -6.70165837e-01 -2.51423120e-01
7.53358603e-01 -2.67490685e-01 1.47220194e-01 -8.63332272e-01
-7.15208173e-01 -1.04273848e-01 -8.83477032e-02 3.32487285e-01
7.26665258e-01 -1.16690505e+00 -7.24604905e-01 7.11195469e-01
-2.06395984e-01 -1.25568332e-02 6.11687779e-01 7.59525478e-01
-2.68204182e-01 4.61751819e-01 2.03049734e-01 -6.00782156e-01
-1.58739686e+00 6.69163838e-02 6.42570376e-01 1.76004022e-01
-2.76220441e-01 1.24285674e+00 4.58872586e-01 -4.80613738e-01
6.96530104e-01 -2.88938612e-01 -2.47820139e-01 -2.15051308e-01
7.15296090e-01 7.05290318e-01 2.82652408e-01 -6.98467672e-01
-4.25399244e-01 6.25863314e-01 4.44599479e-01 -4.43424284e-01
1.04186606e+00 -4.48327452e-01 -2.08578855e-01 8.57989669e-01
1.29554904e+00 1.04384077e+00 -5.52406490e-01 -6.47124499e-02
-3.77649963e-01 -7.11250901e-01 -3.62357855e-01 -7.15026438e-01
-6.68073237e-01 9.11725402e-01 6.18006170e-01 3.67659032e-01
1.58884692e+00 -5.48835136e-02 5.70596457e-01 1.04364343e-01
5.21491945e-01 -7.42554784e-01 5.33730052e-02 3.40067744e-01
1.21393538e+00 -5.70530951e-01 -5.95924854e-01 -3.88383657e-01
-3.72957230e-01 8.21695924e-01 2.69358784e-01 4.53533351e-01
8.93738151e-01 8.09805453e-01 6.01773024e-01 4.39202875e-01
-4.79032367e-01 3.65005612e-01 -1.06592365e-02 8.46086264e-01
7.83204675e-01 1.27798945e-01 2.90953130e-01 1.07007420e+00
-7.19659150e-01 -7.99889386e-01 3.29720408e-01 7.14269459e-01
-4.29078937e-01 -7.34768808e-01 -1.24236536e+00 1.29358321e-01
-7.52436638e-01 -5.25479838e-02 5.18303830e-03 2.40607366e-01
-4.56538759e-02 1.77224553e+00 5.93477748e-02 -4.47532177e-01
8.84110749e-01 -1.48445070e-01 -5.48867509e-02 -3.76003474e-01
-7.08467841e-01 7.95672178e-01 1.25070527e-01 -1.52799368e-01
2.07276437e-02 -7.79402554e-01 -1.07912254e+00 -9.08244997e-02
-4.36565787e-01 5.55238366e-01 9.51949418e-01 3.96762490e-01
1.04957558e-01 7.18404949e-01 1.08491814e+00 -9.56331730e-01
-6.78361297e-01 -1.28367317e+00 -1.42601979e+00 -1.46495208e-01
6.07520461e-01 -2.74240673e-01 -8.11994970e-01 -6.95348382e-02] | [15.112653732299805, 5.783239841461182] |
f63882c9-2ef0-42b1-8569-85e0803c0974 | deepscreen-high-performance-drug-target | null | null | https://pubs.rsc.org/en/content/articlelanding/2020/sc/c9sc03414e#!divAbstract | https://pubs.rsc.org/en/content/articlepdf/2020/sc/c9sc03414e | DEEPScreen: high performance drug–target interaction prediction with convolutional neural networks using 2-D structural compound representations | The identification of physical interactions between drug candidate compounds and target biomolecules is an important process in drug discovery. Since conventional screening procedures are expensive and time consuming, computational approaches are employed to provide aid by automatically predicting novel drug–target interactions (DTIs). In this study, we propose a large-scale DTI prediction system, DEEPScreen, for early stage drug discovery, using deep convolutional neural networks. One of the main advantages of DEEPScreen is employing readily available 2-D structural representations of compounds at the input level instead of conventional descriptors that display limited performance. DEEPScreen learns complex features inherently from the 2-D representations, thus producing highly accurate predictions. The DEEPScreen system was trained for 704 target proteins (using curated bioactivity data) and finalized with rigorous hyper-parameter optimization tests. We compared the performance of DEEPScreen against the state-of-the-art on multiple benchmark datasets to indicate the effectiveness of the proposed approach and verified selected novel predictions through molecular docking analysis and literature-based validation. Finally, JAK proteins that were predicted by DEEPScreen as new targets of a well-known drug cladribine were experimentally demonstrated in vitro on cancer cells through STAT3 phosphorylation, which is the downstream effector protein. The DEEPScreen system can be exploited in the fields of drug discovery and repurposing for in silico screening of the chemogenomic space, to provide novel DTIs which can be experimentally pursued. The source code, trained "ready-to-use" prediction models, all datasets and the results of this study are available at https://github.com/cansyl/DEEPscreen. | ['Rengul Cetin-Atalay', 'Esra Nalbat', 'Volkan Atalay', 'Tunca Dogan', 'Ahmet Sureyya Rifaioglu', 'Maria Jesus Martin'] | 2020-01-08 | null | null | null | chemical-science-2020-1 | ['molecular-docking'] | ['medical'] | [ 1.14129290e-01 -3.90601724e-01 -4.81406122e-01 -1.53059676e-01
-6.88558042e-01 -6.88479006e-01 3.23501408e-01 6.49638414e-01
-2.87712157e-01 1.34677458e+00 -2.22521409e-01 -6.30626142e-01
-2.13764265e-01 -6.19189620e-01 -8.43382239e-01 -9.38681066e-01
-2.21911848e-01 6.89389765e-01 8.47058594e-02 -1.63135171e-01
2.52044499e-01 1.10008454e+00 -9.47789133e-01 3.46290946e-01
9.80508327e-01 6.98892355e-01 3.73028591e-02 4.14360315e-01
2.05146492e-01 1.99817270e-01 -3.28255534e-01 -2.35647708e-01
-1.33499494e-02 -2.53419340e-01 -5.86679518e-01 -5.94260991e-01
-2.71704756e-02 -2.98568457e-02 -1.21054024e-01 5.35705209e-01
8.39866459e-01 -1.38836652e-01 7.52251744e-01 -4.75592941e-01
-5.66524148e-01 -2.01433808e-01 2.36280859e-02 1.57556698e-01
5.49445212e-01 3.93142015e-01 8.45855594e-01 -1.24179947e+00
5.82740784e-01 8.76061082e-01 5.19551814e-01 4.97756094e-01
-1.35305643e+00 -7.92884409e-01 -5.30953467e-01 2.74927169e-01
-1.39374328e+00 -1.66349620e-01 3.19230855e-01 -5.72166264e-01
1.44265723e+00 3.54820907e-01 5.72095990e-01 1.17411041e+00
6.90056384e-01 4.21149254e-01 9.09963071e-01 -9.51806679e-02
4.05952662e-01 9.45349783e-02 -1.85123738e-02 7.40053654e-01
2.70510912e-01 4.46639657e-01 -4.37386781e-01 -7.21573293e-01
4.73980576e-01 1.64216176e-01 -3.19057226e-01 -3.14862818e-01
-8.52048218e-01 9.37018335e-01 5.84721684e-01 2.73254544e-01
-4.41373527e-01 -2.37854049e-01 2.58347929e-01 -3.28552797e-02
3.36559743e-01 8.39554071e-01 -9.41159189e-01 1.75802931e-01
-3.95947307e-01 2.19157726e-01 6.84697270e-01 7.13883042e-02
6.21729851e-01 -2.65662163e-01 -1.20979035e-02 6.02676570e-01
4.06739473e-01 1.60552859e-01 5.46608210e-01 -3.73303071e-02
-5.47244251e-02 7.79867053e-01 1.27984762e-01 -6.69067442e-01
-8.40776086e-01 -5.44642925e-01 -6.22417331e-01 2.30494514e-01
3.06699872e-01 -2.36747116e-01 -7.87723422e-01 1.37743092e+00
4.83986676e-01 2.36187056e-01 3.33277345e-01 9.47598636e-01
9.18844998e-01 6.36081100e-01 5.78099489e-01 -3.66465926e-01
1.14155400e+00 -4.82655376e-01 -1.90528333e-01 6.13876045e-01
9.50796366e-01 -6.73194528e-01 6.46511555e-01 5.57476938e-01
-6.01875186e-01 -2.99033850e-01 -1.15514624e+00 2.72424966e-01
-7.74254382e-01 2.46526107e-01 9.09165144e-01 4.23992783e-01
-5.70351422e-01 1.13593161e+00 -7.77279973e-01 -1.42527699e-01
6.44253671e-01 1.07414436e+00 -7.58426487e-01 2.59483099e-01
-1.41654313e+00 1.00563598e+00 6.74965322e-01 -4.30854484e-02
-1.06008327e+00 -9.66205716e-01 -4.50958997e-01 6.27086461e-02
-1.19644562e-02 -5.80454886e-01 8.12073231e-01 -6.15629435e-01
-1.66983259e+00 4.81877863e-01 -2.45237793e-03 -4.41265285e-01
5.98990172e-02 -2.44804248e-01 -3.89295548e-01 7.95130357e-02
-1.20390140e-01 2.43096873e-01 7.52407312e-02 -6.44841969e-01
-3.85541692e-02 -5.10494173e-01 -2.69565523e-01 1.87109008e-01
-7.82855302e-02 -3.25555429e-02 -6.60745278e-02 -4.09756660e-01
-4.73991364e-01 -9.79026020e-01 -4.68241662e-01 -3.13462466e-01
-6.86069489e-01 -2.72774428e-01 5.36752403e-01 -4.00708169e-01
9.03534770e-01 -1.70484746e+00 1.80054739e-01 4.47186053e-01
3.04987699e-01 9.45664346e-01 -2.47903362e-01 9.54887450e-01
-6.19184852e-01 1.73455864e-01 1.57816976e-01 5.43666124e-01
-4.89372104e-01 -4.15931582e-01 -4.02000733e-02 7.10936666e-01
4.06387180e-01 1.01354671e+00 -7.75799632e-01 1.62001923e-02
2.81394809e-01 7.82970667e-01 -3.66610318e-01 1.77535012e-01
-6.59246802e-01 7.99994886e-01 -9.93392169e-01 7.59055436e-01
6.97451472e-01 -5.42066276e-01 5.34508169e-01 -1.83706582e-01
-1.01485327e-01 3.22170764e-01 -4.24301565e-01 1.44074678e+00
1.74760818e-04 1.71039879e-01 -8.82377028e-01 -7.99829483e-01
8.88920724e-01 5.07453799e-01 6.82070792e-01 -8.05333972e-01
1.39997154e-01 4.19258893e-01 1.74242690e-01 -2.17271835e-01
-4.05056179e-01 -3.14217173e-02 3.22591901e-01 -8.68569687e-02
1.75063666e-02 4.09941614e-01 7.66106742e-03 -1.21341810e-01
1.14247680e+00 1.60503447e-01 5.08191645e-01 -2.51231939e-01
7.41570115e-01 3.34295243e-01 3.10560524e-01 8.10392126e-02
1.99067309e-01 -5.54429367e-03 5.18855691e-01 -8.69394362e-01
-9.10010934e-01 -7.28698790e-01 -6.06457591e-01 7.58725703e-01
-2.04049349e-01 -5.42114079e-01 -4.56833035e-01 -6.67513192e-01
2.49222487e-01 2.93411076e-01 -6.91302598e-01 -9.09878910e-02
-1.48982644e-01 -1.03246725e+00 5.77702224e-01 5.85893244e-02
-7.13058636e-02 -8.95649433e-01 -4.90286201e-02 6.30486310e-01
4.87107188e-01 -5.12285590e-01 2.16322660e-01 6.43067300e-01
-7.16189206e-01 -1.68540788e+00 -7.23352969e-01 -4.61467355e-01
3.73403281e-01 -1.97857305e-01 5.92703581e-01 5.12490124e-02
-5.25246263e-01 -5.33038616e-01 -7.73006082e-02 -6.86047077e-01
-3.88530761e-01 -9.88409482e-03 3.53504747e-01 -3.43794316e-01
5.26785672e-01 -6.49831951e-01 -9.48589206e-01 2.95247257e-01
-6.07903421e-01 -5.90406395e-02 9.04057324e-01 9.87664819e-01
9.00262773e-01 -3.08876216e-01 8.43006134e-01 -1.05711830e+00
6.65211678e-01 -5.91911733e-01 -8.41383755e-01 1.67747155e-01
-5.70596218e-01 2.00706318e-01 8.79799902e-01 -4.08411205e-01
-6.36778891e-01 4.33134198e-01 -5.87224007e-01 -1.46870852e-01
-2.48255849e-01 9.73247230e-01 -2.38574117e-01 -4.28167194e-01
8.75429153e-01 1.12447560e-01 5.57008721e-02 -6.95219219e-01
-1.79218367e-01 4.01623487e-01 -1.73911959e-01 -3.85796338e-01
4.41746265e-01 -4.68922816e-02 5.07414222e-01 -5.83153605e-01
-4.61723626e-01 -3.87312591e-01 -4.48264688e-01 2.17449322e-01
6.54681683e-01 -8.96592438e-01 -1.33563054e+00 2.88602829e-01
-1.09058118e+00 -2.13586450e-01 4.21994120e-01 7.03086555e-01
-2.87941188e-01 3.34509134e-01 -5.15653253e-01 -2.53901094e-01
-7.27551401e-01 -1.33964181e+00 8.33913028e-01 2.93368012e-01
-2.03750566e-01 -9.82222438e-01 6.27076089e-01 2.70894438e-01
7.72014335e-02 7.16103256e-01 1.30893779e+00 -1.36989963e+00
-5.78774095e-01 -3.69015068e-01 -5.39281927e-02 4.75684553e-02
2.40744457e-01 1.85194820e-01 -9.11781490e-01 -1.97288349e-01
-6.14803612e-01 -5.40739000e-01 6.49531543e-01 4.64523077e-01
1.15359461e+00 -2.83031076e-01 -6.96621120e-01 6.38576806e-01
1.66475821e+00 7.19678998e-01 7.79684782e-01 3.69482636e-01
5.95799088e-01 1.23848490e-01 5.87267041e-01 5.72428167e-01
-2.72343189e-01 9.07421649e-01 5.35789967e-01 -4.76164669e-01
3.97163242e-01 -4.04779427e-02 6.85026944e-02 -9.30923522e-02
-3.57546538e-01 -3.06367129e-01 -1.04920042e+00 -4.97226715e-02
-1.61115444e+00 -6.80784047e-01 -3.36199999e-01 2.23408937e+00
1.27696633e+00 -7.31840730e-02 1.10354327e-01 -2.73547798e-01
2.05953956e-01 -5.66225827e-01 -9.07182395e-01 -6.24181509e-01
-2.51729280e-01 9.76696908e-01 4.02231693e-01 3.33813131e-01
-1.18535173e+00 9.58745420e-01 5.78055143e+00 1.01669574e+00
-1.46150529e+00 -3.98372322e-01 1.03682888e+00 9.93270278e-02
1.15790583e-01 -6.00203988e-04 -7.42366493e-01 2.01086357e-01
1.37117481e+00 -2.54016638e-01 8.28269050e-02 7.45010138e-01
5.97116411e-01 5.46787865e-03 -1.25945663e+00 6.00413084e-01
-7.76643574e-01 -2.14418817e+00 1.94757745e-01 3.75147194e-01
5.43367028e-01 1.57166421e-01 1.48528501e-01 -3.12395990e-02
1.08819887e-01 -1.45011079e+00 -2.68387109e-01 5.62548697e-01
9.40994024e-01 -8.45226645e-01 7.47422874e-01 1.16164930e-01
-7.62088180e-01 9.59089622e-02 -4.01488096e-01 1.55528888e-01
-5.03362656e-01 5.53314090e-01 -1.31729269e+00 6.79738998e-01
2.37001628e-01 7.25849509e-01 -6.27236366e-01 1.07360840e+00
6.70278743e-02 3.26811224e-01 -1.21609785e-01 -4.03025448e-01
1.88024268e-01 2.37778341e-03 6.53747618e-02 1.14618301e+00
-3.60001102e-02 1.58210427e-01 1.88453764e-01 6.13458037e-01
-1.35607868e-01 6.08377159e-01 -4.52734917e-01 -4.40574288e-01
1.35230079e-01 1.20975649e+00 -3.21044922e-01 -1.40383378e-01
-1.39711231e-01 6.83997810e-01 1.95049360e-01 3.70511740e-01
-8.58517170e-01 -2.28018448e-01 8.80772889e-01 1.32738754e-01
9.80205238e-02 1.40910774e-01 1.07563294e-01 -8.09029281e-01
-5.52558184e-01 -1.01105738e+00 3.48223418e-01 -4.33154941e-01
-1.20657551e+00 4.71612602e-01 -1.84706390e-01 -1.19630897e+00
1.36148229e-01 -1.12434995e+00 -5.84275544e-01 1.19007504e+00
-1.33955944e+00 -9.87079382e-01 2.03425840e-01 5.42868733e-01
2.46077895e-01 -2.96062738e-01 1.38117254e+00 2.41820782e-01
-7.93135822e-01 4.37730491e-01 7.72012115e-01 -3.16462576e-01
8.01132739e-01 -8.87255549e-01 6.19284399e-02 -2.55166017e-03
-2.50938624e-01 8.18523407e-01 5.54485202e-01 -8.34517002e-01
-1.44710588e+00 -1.10139084e+00 5.31241477e-01 -2.40235761e-01
7.77621627e-01 -2.12264881e-01 -8.24019790e-01 7.41422027e-02
-1.04631670e-01 -7.67965317e-02 1.49860311e+00 -1.02091536e-01
-1.59809262e-01 2.02603802e-01 -9.64967668e-01 4.00735408e-01
4.19091672e-01 -2.03861848e-01 5.97237758e-02 8.84303689e-01
3.15683544e-01 -4.59082276e-01 -1.25298083e+00 5.96639097e-01
7.12083161e-01 -7.10788786e-01 1.11394012e+00 -1.26014543e+00
2.87951976e-01 -3.36768061e-01 1.62564114e-01 -1.13271785e+00
-4.93873805e-01 -4.26461130e-01 -1.02717504e-01 4.32633698e-01
9.64391232e-01 -5.95656395e-01 1.05916524e+00 3.88809055e-01
-1.40007026e-03 -1.41306770e+00 -1.07767558e+00 -4.67535615e-01
2.85850137e-01 1.88960299e-01 4.00799811e-01 7.88886845e-01
2.57186204e-01 6.62384391e-01 -1.39456525e-01 6.87170029e-02
9.33261067e-02 -3.17052007e-02 6.99536443e-01 -1.48484886e+00
-3.90994817e-01 -2.53967732e-01 -5.19037545e-01 -3.24549496e-01
1.18464515e-01 -8.64341557e-01 -6.44992530e-01 -1.34440589e+00
3.33139598e-01 -2.50436038e-01 -6.68031275e-01 7.00303793e-01
6.17583878e-02 8.25361931e-04 -4.17149812e-01 3.39616448e-01
-2.74334878e-01 4.39263552e-01 1.22671628e+00 -3.33895832e-01
-4.80778813e-01 2.07206547e-01 -6.28701270e-01 2.69136965e-01
9.73032773e-01 -3.23018879e-01 -3.03905576e-01 5.03817856e-01
1.99821129e-01 5.95316775e-02 2.94113517e-01 -7.83634365e-01
-2.54494190e-01 -4.25015271e-01 1.06790423e+00 -4.64481086e-01
4.85875934e-01 -7.69697964e-01 7.32215881e-01 8.88373315e-01
-1.27361238e-01 -3.47053379e-01 6.34823799e-01 5.29103100e-01
-1.62144881e-02 2.17346475e-01 8.92961144e-01 1.64425354e-02
-4.94134814e-01 6.40707433e-01 -4.93825823e-01 -7.24481463e-01
1.24979937e+00 -3.44793260e-01 -3.44653726e-01 1.43212914e-01
-9.51691151e-01 -2.59059906e-01 3.17506284e-01 -4.04440910e-02
7.81917334e-01 -1.00331712e+00 -5.21292090e-01 4.97502908e-02
3.59897524e-01 -4.66559798e-01 1.45827815e-01 8.56767058e-01
-9.43476021e-01 9.95916545e-01 -3.05794686e-01 -3.69590312e-01
-1.42946720e+00 7.96583295e-01 6.85049713e-01 -4.54775572e-01
-1.91358566e-01 6.93287671e-01 4.86343235e-01 -1.01825699e-01
2.33667484e-03 -4.02341783e-02 -4.46949303e-01 -2.54091531e-01
3.82994354e-01 -1.96579695e-01 4.08833265e-01 -4.57928121e-01
-7.59803593e-01 2.96983749e-01 -4.12968159e-01 7.45624661e-01
1.72114086e+00 7.01221764e-01 5.12079941e-03 -2.53699690e-01
1.35749292e+00 -2.28700712e-01 -8.82526398e-01 -2.50560101e-02
1.46807730e-01 -1.67758107e-01 -3.92357223e-02 -1.32996321e+00
-5.18782139e-01 5.68435073e-01 1.02738285e+00 -4.97677863e-01
8.97046804e-01 9.38355643e-03 5.02391279e-01 8.10429335e-01
6.68817386e-02 -7.66282439e-01 7.89286941e-02 3.20877492e-01
8.93477798e-01 -1.14293671e+00 2.95656323e-01 -3.33107561e-01
-4.57975686e-01 1.34726858e+00 4.74954933e-01 1.02462076e-01
5.95843196e-01 -1.02941394e-01 -1.60935208e-01 -5.51251948e-01
-9.06111360e-01 1.84851602e-01 5.31551480e-01 5.29588580e-01
9.68593240e-01 7.06114396e-02 -6.29247487e-01 6.49212539e-01
4.75421906e-01 1.63306475e-01 1.51410401e-01 8.21309328e-01
-3.04306179e-01 -1.65415192e+00 -1.17170721e-01 2.77543932e-01
-8.15398633e-01 -3.45728934e-01 -1.05268586e+00 8.23766053e-01
-1.15762092e-02 6.25343442e-01 -7.37371266e-01 -4.71080482e-01
2.74994045e-01 3.99932303e-02 2.89446265e-01 -5.82618177e-01
-6.27649367e-01 2.44503424e-01 2.40704626e-01 -3.34448218e-01
-2.12582111e-01 -2.27061570e-01 -1.33648920e+00 -2.54188985e-01
-6.05763078e-01 5.11758268e-01 7.20631719e-01 8.03406477e-01
8.66445303e-01 2.68846869e-01 7.54079521e-01 -7.80427217e-01
-1.76961899e-01 -8.43715429e-01 -3.77447248e-01 5.96914142e-02
9.06229988e-02 -7.22005725e-01 1.91325501e-01 -2.02262886e-02] | [5.054746150970459, 5.662865161895752] |
818c0f0e-0256-4928-826b-5bbba7c70174 | kilm-knowledge-injection-into-encoder-decoder | 2302.09170 | null | https://arxiv.org/abs/2302.09170v1 | https://arxiv.org/pdf/2302.09170v1.pdf | KILM: Knowledge Injection into Encoder-Decoder Language Models | Large pre-trained language models (PLMs) have been shown to retain implicit knowledge within their parameters. To enhance this implicit knowledge, we propose Knowledge Injection into Language Models (KILM), a novel approach that injects entity-related knowledge into encoder-decoder PLMs, via a generative knowledge infilling objective through continued pre-training. This is done without architectural modifications to the PLMs or adding additional parameters. Experimental results over a suite of knowledge-intensive tasks spanning numerous datasets show that KILM enables models to retain more knowledge and hallucinate less, while preserving their original performance on general NLU and NLG tasks. KILM also demonstrates improved zero-shot performances on tasks such as entity disambiguation, outperforming state-of-the-art models having 30x more parameters. | ['Dilek Hakkani-Tür', 'Yang Liu', 'Aishwarya Padmakumar', 'Devamanyu Hazarika', 'Mahdi Namazifar', 'Yan Xu'] | 2023-02-17 | null | null | null | null | ['entity-disambiguation'] | ['natural-language-processing'] | [ 7.45166242e-02 8.00981343e-01 -4.88583863e-01 -4.91678342e-02
-9.99549448e-01 -5.25358737e-01 9.92137074e-01 1.98465288e-02
-5.93193769e-01 9.41888452e-01 6.03963733e-01 -1.71282113e-01
1.84647575e-01 -4.93482322e-01 -9.16703284e-01 -7.77261183e-02
-6.33997694e-02 6.91960931e-01 2.47810811e-01 -1.32757321e-01
-1.35427788e-01 -4.45422716e-02 -9.41632569e-01 4.90826517e-01
9.05300140e-01 3.73706967e-01 3.97550493e-01 3.60940218e-01
-3.08764786e-01 1.05192733e+00 -4.73999768e-01 -7.24042833e-01
-4.84717786e-02 -2.40747705e-01 -9.47486997e-01 -2.16272205e-01
3.65514755e-01 -2.95880526e-01 -5.39906681e-01 6.83199763e-01
4.93862927e-01 1.58446416e-01 6.70905948e-01 -8.15106869e-01
-1.36485362e+00 1.10889673e+00 -2.99130350e-01 2.06888810e-01
7.08786398e-02 1.45526692e-01 1.03038704e+00 -1.29834700e+00
6.52553439e-01 1.54965472e+00 6.63964987e-01 6.42043471e-01
-1.60330820e+00 -7.89344668e-01 1.44319847e-01 9.72422138e-02
-1.80092001e+00 -8.50765705e-01 1.26704693e-01 -1.34652659e-01
1.92986572e+00 -3.91703874e-01 2.05990851e-01 1.20504630e+00
1.19019479e-01 1.10249209e+00 6.88865066e-01 -5.46892047e-01
-1.49611598e-02 5.25604308e-01 3.06569263e-02 9.61556554e-01
5.66684484e-01 -4.30691391e-02 -8.67800534e-01 -1.66669548e-01
6.64564610e-01 -5.04034936e-01 -2.33644783e-01 -1.70694992e-01
-1.17665672e+00 6.56022251e-01 1.68716848e-01 2.25044161e-01
-4.73773897e-01 4.26751882e-01 1.88165575e-01 2.01200917e-01
5.28693914e-01 9.36041534e-01 -7.59023368e-01 -1.47987917e-01
-1.04634404e+00 7.57045969e-02 1.00824618e+00 1.36570919e+00
1.02113998e+00 3.18815529e-01 -4.50759083e-01 8.49906564e-01
2.87643462e-01 5.49448431e-01 7.98763692e-01 -8.85027766e-01
5.26836097e-01 5.48169255e-01 -1.77554460e-03 -5.33589780e-01
-2.08541438e-01 -8.75004470e-01 -4.65559065e-01 -4.05881643e-01
-2.78883576e-01 -1.74982324e-01 -1.32509232e+00 2.03758335e+00
-2.63157606e-01 5.27925074e-01 5.55448055e-01 3.54383677e-01
7.95378387e-01 8.22706342e-01 5.97202659e-01 1.51189432e-01
1.31842470e+00 -1.06428087e+00 -7.00331151e-01 -7.17381001e-01
9.43395317e-01 -4.42931086e-01 1.08095276e+00 1.11867726e-01
-1.25229871e+00 -4.21692133e-01 -9.49627042e-01 -3.52201849e-01
-5.18208504e-01 2.31009349e-01 7.06533611e-01 5.39669514e-01
-1.39992368e+00 3.43544006e-01 -9.43689406e-01 -2.99705982e-01
6.55159593e-01 3.50961745e-01 -3.54698420e-01 -4.73543257e-02
-1.40858674e+00 1.41122401e+00 1.05593228e+00 -3.34166110e-01
-1.17708218e+00 -1.15650928e+00 -1.28385532e+00 3.85347933e-01
4.47411060e-01 -1.17464304e+00 1.47726774e+00 -5.98638892e-01
-1.43148410e+00 7.03529000e-01 -3.93567562e-01 -7.92084038e-01
-4.18263003e-02 -6.97165072e-01 -5.56995153e-01 -7.53872991e-02
1.71400189e-01 1.27530074e+00 6.29049659e-01 -1.44786513e+00
-2.82221645e-01 2.78249294e-01 -1.53990770e-02 3.82226735e-01
-5.61889291e-01 -2.22990140e-01 -8.85382891e-01 -7.56078005e-01
-3.26455832e-01 -6.65538251e-01 -1.86778262e-01 -5.57462990e-01
-5.45794964e-01 -4.37748544e-02 6.50587857e-01 -9.67710495e-01
1.48450828e+00 -1.88240194e+00 3.03577129e-02 -4.13231328e-02
1.35628283e-01 8.55431437e-01 -6.50831819e-01 4.75419521e-01
2.00541154e-01 3.35365057e-01 -2.53597230e-01 -8.41319203e-01
2.78420120e-01 5.62870204e-01 -4.78384793e-01 -2.65877306e-01
5.13753176e-01 1.71491241e+00 -1.02761805e+00 -4.95091259e-01
-1.18215479e-01 6.95721149e-01 -6.31630540e-01 -6.58437237e-02
-7.00690985e-01 8.25081095e-02 -2.09877089e-01 5.22162676e-01
2.92463869e-01 -5.53456724e-01 2.59538144e-01 -5.88429607e-02
3.05122703e-01 6.82431340e-01 -8.54798317e-01 1.95046031e+00
-8.15343261e-01 5.99040031e-01 -3.11230808e-01 -4.63666826e-01
5.66237569e-01 5.46071470e-01 -1.29405841e-01 -8.91469836e-01
-8.60025361e-02 1.22667037e-01 -1.69873431e-01 -1.59465045e-01
8.58069539e-01 -1.49988487e-01 -3.43844965e-02 4.50062066e-01
6.49922192e-01 2.19867453e-01 1.83663711e-01 6.40537083e-01
1.19406044e+00 8.88743550e-02 4.26000178e-01 -7.40307942e-02
1.78676292e-01 -1.36583429e-02 3.80775034e-01 9.63837981e-01
3.01528692e-01 1.29674837e-01 6.31599277e-02 3.20084393e-01
-9.51311588e-01 -1.30260241e+00 1.38538972e-01 1.07801664e+00
-2.25428417e-01 -7.50982761e-01 -6.08008087e-01 -5.67400634e-01
2.17167139e-01 1.40663457e+00 -3.57606292e-01 -6.23819590e-01
-4.18768734e-01 -8.09084296e-01 1.13645828e+00 8.33245873e-01
6.22604728e-01 -1.00868618e+00 -1.44190207e-01 4.16966796e-01
-5.13707623e-02 -1.35186279e+00 -4.17034328e-01 2.12085694e-01
-7.82349586e-01 -4.40869868e-01 -6.17751002e-01 -8.49486470e-01
6.19879663e-01 1.31990155e-03 1.50228310e+00 -3.26678932e-01
-9.24843103e-02 5.65102041e-01 -1.67520776e-01 -2.91532815e-01
-6.35908544e-01 5.80588400e-01 -6.64827228e-03 -3.12346309e-01
5.62866032e-01 -6.17053032e-01 -1.28840685e-01 -1.36780635e-01
-9.52067316e-01 2.16482073e-01 1.16250753e+00 8.19848955e-01
6.07047021e-01 -1.62454754e-01 1.05016470e+00 -1.00929606e+00
7.19904482e-01 -5.82303643e-01 -2.91074127e-01 3.80373508e-01
-9.05631185e-01 5.94998777e-01 3.77983987e-01 -5.12240469e-01
-1.44280517e+00 -2.71652848e-01 1.34505972e-01 -4.89326060e-01
1.77252308e-01 8.38064611e-01 -2.70917982e-01 5.77706192e-03
7.37326980e-01 5.34556627e-01 -3.87842774e-01 -7.89543211e-01
1.05237079e+00 3.41176510e-01 7.12441504e-01 -5.54518223e-01
7.04850018e-01 4.77916189e-02 -4.93326873e-01 -6.86944306e-01
-1.00470734e+00 -5.28424144e-01 -6.08618677e-01 4.64341968e-01
6.83241546e-01 -1.50488818e+00 -1.94324419e-01 8.40106979e-02
-1.33784020e+00 -6.24905705e-01 -4.06252325e-01 4.92397666e-01
-3.26189876e-01 1.52482390e-01 -5.86305916e-01 -4.57635999e-01
-4.95767832e-01 -6.60823822e-01 1.07070184e+00 1.79202348e-01
-4.72074032e-01 -1.42449248e+00 1.53819144e-01 2.66495615e-01
6.71067595e-01 -4.44818258e-01 1.25334787e+00 -8.91795158e-01
-8.70397329e-01 4.24028002e-02 -3.79558206e-01 4.61176515e-01
4.55941400e-03 -7.84651697e-01 -1.03713930e+00 -3.59634250e-01
-5.68155408e-01 -5.31297207e-01 1.33641696e+00 -5.67262545e-02
8.01961601e-01 -6.54065251e-01 -7.87446856e-01 6.44492745e-01
1.43042552e+00 9.43718776e-02 6.19601250e-01 2.10282385e-01
8.00700307e-01 -9.15198773e-03 4.90511544e-02 1.46371275e-01
6.62731290e-01 5.47321796e-01 -1.67799369e-01 -5.20597287e-02
-7.41657555e-01 -8.13019753e-01 6.67946398e-01 8.96585226e-01
2.52196550e-01 -5.63055754e-01 -1.01456964e+00 1.06713486e+00
-1.81688654e+00 -7.25492001e-01 4.33310807e-01 1.86709785e+00
1.54935777e+00 9.56853405e-02 -5.66796124e-01 -5.55470467e-01
2.94646502e-01 2.49508068e-01 -6.90110683e-01 -2.67607093e-01
-3.60466570e-01 5.59150934e-01 6.82207882e-01 8.58589709e-01
-9.44830954e-01 1.74138188e+00 6.84657049e+00 1.06106722e+00
-8.17262053e-01 4.67866778e-01 2.29997337e-01 -1.38067305e-01
-5.65935433e-01 2.33182117e-01 -1.17304695e+00 1.54282853e-01
1.25233853e+00 -5.96747756e-01 3.97192240e-01 7.41063774e-01
-1.26202896e-01 -1.21872224e-01 -1.12220097e+00 7.19422877e-01
3.10253650e-01 -1.41190505e+00 7.09008753e-01 -1.77249178e-01
9.67327595e-01 4.06585991e-01 3.41847539e-02 9.60184038e-01
6.75272048e-01 -1.27124989e+00 3.92744750e-01 8.20229292e-01
8.75252306e-01 -5.95864356e-01 4.54080850e-01 4.52327281e-01
-8.79483879e-01 2.18912736e-01 -3.32273245e-01 7.27033764e-02
5.79682767e-01 5.62750518e-01 -1.71920288e+00 4.63500977e-01
9.30610299e-02 5.89081764e-01 -8.39021266e-01 7.25694656e-01
-5.13740182e-01 9.58181620e-01 -2.16498688e-01 2.57923007e-01
2.49447227e-01 5.13319671e-01 4.92283076e-01 1.84588051e+00
9.80824977e-02 -5.00145890e-02 -3.41063156e-03 1.24105489e+00
-5.92795968e-01 -1.01555206e-01 -5.73014498e-01 -3.94845605e-01
6.76918626e-01 8.67182314e-01 -3.25280167e-02 -6.83671355e-01
-5.20947397e-01 1.35186505e+00 5.57000220e-01 6.93938196e-01
-6.20704830e-01 -2.74790734e-01 6.28660858e-01 3.14708278e-02
4.11894619e-01 -4.98785079e-01 -1.71439201e-01 -1.23742223e+00
-2.83630669e-01 -5.90966642e-01 1.56308144e-01 -8.73053968e-01
-9.98166084e-01 4.96437490e-01 1.41447812e-01 -5.34628093e-01
-6.13790572e-01 -3.98381352e-01 -3.07956152e-02 1.10376906e+00
-2.00067258e+00 -1.42760587e+00 2.64308155e-01 3.68216097e-01
4.49602783e-01 -4.17134434e-01 1.18638110e+00 3.56295437e-01
-3.73897105e-01 9.38938200e-01 9.01925191e-02 1.89101800e-01
7.36122847e-01 -1.24473774e+00 6.94265246e-01 8.15755606e-01
4.00150418e-01 9.74406004e-01 4.03689831e-01 -9.81899679e-01
-1.26604426e+00 -1.57873917e+00 1.45723712e+00 -8.42657983e-01
5.56301773e-01 -4.26435590e-01 -1.13206553e+00 1.30420315e+00
6.09233975e-01 -3.44888538e-01 7.80077338e-01 2.92117596e-01
-5.95506728e-01 3.87653500e-01 -7.48598635e-01 6.27503037e-01
1.09072912e+00 -7.97204137e-01 -1.15836382e+00 1.73145518e-01
1.10254979e+00 -2.37285808e-01 -8.18331301e-01 3.58957142e-01
2.27539673e-01 -1.44605413e-01 1.16362119e+00 -7.93079257e-01
3.34219784e-01 -2.50728339e-01 -1.10391319e-01 -1.58693755e+00
-4.26461965e-01 -7.40463555e-01 -7.99756348e-01 1.32876945e+00
9.15502310e-01 -5.47850370e-01 5.86299062e-01 5.56866109e-01
-3.36820960e-01 -5.13803303e-01 -6.45949066e-01 -1.15742445e+00
8.67121890e-02 -3.73844653e-01 2.82957137e-01 1.05690598e+00
6.02149665e-02 9.07809913e-01 -3.98489773e-01 3.15259069e-01
3.08166176e-01 -4.20653999e-01 5.45788527e-01 -7.39944041e-01
-3.88585210e-01 -3.19578290e-01 -1.80434108e-01 -1.28952551e+00
6.00166261e-01 -1.35843670e+00 -1.92495957e-01 -1.90218818e+00
4.22870696e-01 -2.11061507e-01 -3.33226174e-01 1.22019458e+00
-3.66915435e-01 1.53293118e-01 1.79399595e-01 1.22109145e-01
-1.05716538e+00 9.15148616e-01 9.88498867e-01 -8.28908384e-02
-2.36353874e-01 -6.14064217e-01 -1.01365745e+00 5.76995969e-01
4.43261236e-01 -4.66821522e-01 -8.46963525e-01 -9.55693603e-01
2.85407871e-01 -2.89566189e-01 3.82445455e-01 -1.02274871e+00
4.51963156e-01 2.10881725e-01 2.78700978e-01 -4.58384365e-01
6.55603528e-01 -3.09120059e-01 -2.81596631e-02 2.41403326e-01
-4.15275425e-01 -9.00007944e-05 8.23002040e-01 5.44274986e-01
-4.19984460e-01 -7.51706809e-02 5.28476954e-01 -1.97695285e-01
-1.17997277e+00 5.63194156e-02 -2.78995246e-01 3.80412489e-01
6.58210516e-01 9.76536795e-02 -4.23272938e-01 -4.60592151e-01
-7.99580634e-01 4.02330756e-01 2.33048245e-01 5.66833973e-01
5.67646682e-01 -1.22054207e+00 -8.53356898e-01 4.44208533e-02
2.27516428e-01 -5.95380180e-02 3.04439459e-02 5.61701596e-01
-1.15731210e-01 9.98254418e-01 2.88777441e-01 -1.53270200e-01
-7.50726938e-01 5.08059740e-01 3.13610822e-01 -5.33705533e-01
-5.36392808e-01 9.17421341e-01 3.72306883e-01 -5.40390670e-01
1.84115514e-01 -3.20563793e-01 -2.86054965e-02 -6.82207420e-02
5.59119582e-01 4.70008142e-02 1.10779591e-01 -2.80793220e-01
-2.84814477e-01 -1.88944880e-02 -4.62928265e-01 -3.83770615e-01
1.12668967e+00 -2.83178389e-01 2.17579260e-01 2.07578033e-01
1.14262974e+00 8.74977559e-02 -1.09046960e+00 -8.92701149e-01
3.10023278e-01 8.12117103e-03 2.64553547e-01 -1.57139742e+00
-6.84753835e-01 7.36784041e-01 2.26793531e-02 -5.52961588e-01
9.27860260e-01 2.03538924e-01 1.05340970e+00 7.78333485e-01
3.95176113e-01 -9.76589799e-01 1.49473876e-01 1.12688398e+00
6.97588027e-01 -9.04378831e-01 -1.91661313e-01 -2.77300388e-01
-8.97570133e-01 4.38998103e-01 6.14635289e-01 2.84955144e-01
5.27562618e-01 3.58157426e-01 -2.56276876e-03 -4.06624638e-02
-1.22032666e+00 -4.62686718e-01 3.30832064e-01 6.46711051e-01
4.30001020e-01 2.36539301e-02 -5.42170480e-02 1.07562685e+00
-1.47857875e-01 1.72233641e-01 1.94901511e-01 8.98002863e-01
-4.98732388e-01 -1.08685696e+00 9.77679566e-02 2.79723376e-01
-2.76798546e-01 -1.00370121e+00 -3.60985667e-01 7.76734114e-01
1.25442624e-01 5.93699515e-01 -3.77356298e-02 -1.91750705e-01
1.99875638e-01 6.31056428e-01 3.90489817e-01 -1.03387523e+00
-4.34546053e-01 -1.50753051e-01 4.82329845e-01 -4.55649525e-01
-1.27979830e-01 -4.24061418e-01 -1.49717271e+00 -3.91220972e-02
-2.08850831e-01 5.48164286e-02 2.71397501e-01 9.71828699e-01
1.03017449e+00 7.37237275e-01 -1.07266016e-01 -3.84426653e-01
-5.07316232e-01 -1.08537364e+00 -3.81290227e-01 8.11151266e-02
7.63650909e-02 -6.20212793e-01 1.48077995e-01 2.88937747e-01] | [10.594982147216797, 8.26716423034668] |
0710e70c-8b5c-46ae-b3c4-e15f243764df | computing-expected-motif-counts-for | 2305.01089 | null | https://arxiv.org/abs/2305.01089v1 | https://arxiv.org/pdf/2305.01089v1.pdf | Computing Expected Motif Counts for Exchangeable Graph Generative Models | Estimating the expected value of a graph statistic is an important inference task for using and learning graph models. This note presents a scalable estimation procedure for expected motif counts, a widely used type of graph statistic. The procedure applies for generative mixture models of the type used in neural and Bayesian approaches to graph data. | ['Oliver Schulte'] | 2023-05-01 | null | null | null | null | ['type'] | ['speech'] | [ 3.24829996e-01 2.58464098e-01 -3.71444464e-01 -4.40681487e-01
-3.77305359e-01 -5.14828265e-01 4.84319866e-01 2.06362978e-01
-3.13074499e-01 8.26003671e-01 -1.18678704e-01 -5.54690778e-01
-1.45328194e-01 -9.24974740e-01 -7.12255120e-01 -7.28815258e-01
-5.56949079e-01 7.19046772e-01 2.27745190e-01 4.58399624e-01
4.70264107e-01 5.83387434e-01 -1.43663645e+00 -2.54867315e-01
4.49378669e-01 3.08427781e-01 7.97424987e-02 1.27508080e+00
-2.03849792e-01 5.65053582e-01 -8.05588365e-01 -4.38599437e-01
-3.19289893e-01 -5.54204345e-01 -2.54416168e-01 -4.63883346e-03
7.37872481e-01 -2.83347726e-01 -4.73245442e-01 1.29779220e+00
4.68068153e-01 1.79828018e-01 1.63680947e+00 -1.29901123e+00
-9.01812166e-02 1.03713274e+00 -5.41103482e-01 5.87521791e-01
2.58345425e-01 -3.33292097e-01 1.06199419e+00 -4.57848877e-01
7.35001862e-01 1.32725751e+00 6.98011458e-01 1.22295059e-01
-1.62763751e+00 -5.37855327e-01 -3.28082651e-01 3.77062596e-02
-1.63265216e+00 -2.00972423e-01 7.16907322e-01 -6.57847464e-01
1.16183984e+00 1.36249676e-01 9.53622997e-01 1.29787695e+00
6.59116566e-01 6.84171915e-01 7.52292275e-01 -3.20448101e-01
6.23302817e-01 -2.61586785e-01 3.63493592e-01 1.08348989e+00
7.18856692e-01 -5.23498580e-02 -5.40122747e-01 -6.27930045e-01
1.22729778e+00 -2.49121219e-01 6.08751327e-02 -7.50206858e-02
-5.65854609e-01 1.11938727e+00 -2.25952297e-01 9.44623277e-02
-2.92181730e-01 1.00351918e+00 3.19146603e-01 6.69140816e-02
6.80591524e-01 -2.01276317e-01 -3.76633294e-02 -3.62366021e-01
-1.10312521e+00 2.85265952e-01 1.30054212e+00 1.19550228e+00
6.57208622e-01 4.76662606e-01 -2.42762372e-01 7.53974319e-01
9.52008903e-01 4.48762625e-01 -2.83742398e-01 -7.81435907e-01
-1.61093339e-01 1.79905728e-01 -5.93309343e-01 -9.08870339e-01
-3.10963541e-01 -4.57664073e-01 -9.64684248e-01 5.57753108e-02
7.06710696e-01 -3.87708656e-02 -9.59193885e-01 1.73342979e+00
2.65769511e-01 3.76854062e-01 -4.85072792e-01 1.22256331e-01
1.13679564e+00 5.75733185e-01 -8.28455985e-02 -2.77048975e-01
1.26380336e+00 -2.83517957e-01 -7.30057895e-01 -2.28468955e-01
2.90283918e-01 -5.56339502e-01 4.85745102e-01 5.32251477e-01
-8.21779549e-01 -1.17227726e-01 -9.49858606e-01 2.89298743e-01
-4.13000137e-01 -9.11737755e-02 1.00133920e+00 1.15775239e+00
-1.21914017e+00 7.56224394e-01 -8.66674304e-01 -5.43212771e-01
4.67694968e-01 3.37218493e-01 -2.07424983e-01 -1.88727275e-01
-6.68078780e-01 6.30998015e-01 7.65871048e-01 -1.77481502e-01
-8.82584393e-01 -2.12041304e-01 -1.01971793e+00 1.08567081e-01
1.14048325e-01 -4.71660107e-01 8.78843248e-01 -2.72782624e-01
-1.42754829e+00 9.54885304e-01 -4.18920368e-02 -2.61646509e-01
-4.75685075e-02 3.56557310e-01 -3.57062183e-02 3.44495773e-02
-2.26133719e-01 4.35460985e-01 9.77048457e-01 -6.79486632e-01
1.58798844e-01 -3.30175817e-01 -5.93556881e-01 -3.14779907e-01
-9.57393087e-03 -2.04045385e-01 -3.69434983e-01 -7.05807745e-01
-1.53222005e-03 -6.74985945e-01 -1.91179737e-01 -2.42711291e-01
-6.05713248e-01 -5.60546279e-01 2.47873053e-01 -8.76580417e-01
1.21827424e+00 -1.68270230e+00 1.85909681e-02 8.00976157e-01
5.36137342e-01 -3.72701615e-01 -9.67540294e-02 8.21538925e-01
1.04003318e-01 -8.17532837e-02 -3.21484625e-01 7.99865276e-02
3.54188047e-02 2.92235225e-01 -1.80581920e-02 6.43478692e-01
1.38822675e-03 7.59112537e-01 -8.75192583e-01 -7.05979407e-01
6.69780523e-02 5.20761549e-01 -5.68296611e-01 2.64841884e-01
-4.31006700e-01 4.42047268e-02 -2.06857830e-01 1.99836612e-01
6.05966389e-01 -5.58973849e-01 6.96375966e-01 4.20343354e-02
3.13245267e-01 -1.02006637e-01 -1.33268487e+00 1.51228845e+00
2.50650018e-01 8.45446706e-01 -2.56183326e-01 -1.02926743e+00
1.15804827e+00 5.40468469e-02 1.27047867e-01 3.48034829e-01
4.00452763e-01 -7.27564618e-02 1.66551396e-01 -2.59984910e-01
3.80198173e-02 7.95129761e-02 -2.51349434e-02 7.79071629e-01
7.14291751e-01 -2.94158041e-01 6.21870100e-01 6.01866364e-01
1.67426169e+00 2.11859390e-01 8.36570740e-01 -7.13312566e-01
-1.04295455e-01 -3.56077701e-01 -3.30933854e-02 1.07356083e+00
2.16485217e-01 1.92132339e-01 1.09211087e+00 -2.71860421e-01
-9.80131984e-01 -1.49364495e+00 -8.03216398e-02 1.10553741e+00
-5.17148316e-01 -7.35967100e-01 -8.12251687e-01 -2.90092677e-01
-4.73172478e-02 5.96247375e-01 -4.31564182e-01 -2.68315762e-01
-1.39783129e-01 -9.84776318e-01 7.13385165e-01 4.60207492e-01
-5.47444113e-02 -9.69400823e-01 -1.08953953e-01 1.39037505e-01
2.13115782e-01 -8.53956342e-01 -4.08116549e-01 6.56232715e-01
-1.12318349e+00 -1.06007481e+00 -6.00454032e-01 -6.94891453e-01
6.14621043e-01 -2.44101778e-01 1.51266801e+00 1.65983662e-01
-6.11584425e-01 6.89447582e-01 1.92148492e-01 -2.14269519e-01
-7.65574694e-01 -3.70342890e-03 -2.11338580e-01 -3.49477351e-01
3.30520540e-01 -1.01900232e+00 -7.11034983e-02 -1.08063474e-01
-8.99325728e-01 -7.36061409e-02 6.87011480e-01 5.84575474e-01
4.21239644e-01 1.39053196e-01 2.68195927e-01 -1.13917840e+00
1.06368172e+00 -6.24827266e-01 -9.49923098e-01 1.53819710e-01
-6.05744302e-01 5.61789811e-01 3.04179400e-01 -7.44280934e-01
-4.36807960e-01 -3.58800925e-02 -5.42010628e-02 -2.35956103e-01
-3.72548252e-01 8.49395096e-01 -2.28107944e-01 -6.98204041e-02
4.41939801e-01 1.63280845e-01 1.17837556e-01 4.05790620e-02
2.41802648e-01 3.40464562e-01 2.85601288e-01 -2.59508997e-01
2.94135422e-01 8.00889209e-02 6.31891370e-01 -1.37548876e+00
-3.79721999e-01 -4.60524201e-01 -4.93995607e-01 -6.44545436e-01
8.15006793e-01 -5.18168867e-01 -9.47021484e-01 5.30098975e-01
-1.26288843e+00 -5.05757749e-01 9.68629196e-02 6.47049963e-01
-7.47412086e-01 4.85188335e-01 -6.96440518e-01 -1.22029042e+00
-4.49385419e-02 -5.77705026e-01 1.15731907e+00 2.36155123e-01
-3.08060735e-01 -1.64240348e+00 7.29870677e-01 -2.26402938e-01
2.73052335e-01 4.40264314e-01 9.94984090e-01 -8.00130188e-01
-4.99422193e-01 -3.67835999e-01 -1.44497976e-01 -2.47057244e-01
-2.01635256e-01 6.11535668e-01 -7.41097689e-01 -9.44255367e-02
-4.17519033e-01 -9.60704610e-02 9.55370605e-01 1.13670003e+00
8.86271834e-01 -2.51851588e-01 -5.57996809e-01 2.68244147e-01
1.40281606e+00 -2.32068121e-01 5.55865109e-01 -4.64967042e-01
6.26501441e-01 3.75215918e-01 -5.48516260e-03 4.75063413e-01
1.62794188e-01 3.73935580e-01 1.51930332e-01 2.61923373e-01
1.20669663e-01 -3.80172044e-01 4.55190688e-01 9.02217031e-01
6.67192116e-02 -7.33206570e-01 -1.01850748e+00 2.70687670e-01
-1.97128534e+00 -1.36705387e+00 -6.08902395e-01 2.12472486e+00
7.36391008e-01 3.00271094e-01 4.48056936e-01 4.65884209e-02
1.26610315e+00 1.90664694e-01 -5.10383964e-01 -4.77552176e-01
4.44949828e-02 4.43697333e-01 6.67401612e-01 5.91364384e-01
-7.70603001e-01 9.64842141e-01 8.85009193e+00 1.34541357e+00
-2.71414846e-01 -2.57291824e-01 3.25983733e-01 3.66467625e-01
-2.55206078e-01 1.97294563e-01 -7.79850543e-01 2.93139368e-01
1.46891999e+00 2.60319095e-02 7.98902929e-01 6.80253923e-01
-2.91193947e-02 -6.41051412e-01 -1.19183421e+00 1.18629134e+00
2.50451744e-01 -1.26308429e+00 -1.30828977e-01 6.30277455e-01
3.57596755e-01 1.31006399e-02 -1.35603696e-01 9.99635383e-02
8.42398047e-01 -1.13397658e+00 1.96411610e-01 8.17053258e-01
5.97536147e-01 -7.97330201e-01 4.82295245e-01 4.37011957e-01
-1.10821688e+00 7.32941031e-01 -6.98113859e-01 2.65067425e-02
1.35988578e-01 1.16609526e+00 -1.12817180e+00 -2.43455082e-01
2.91607648e-01 6.57225907e-01 -7.19126105e-01 1.16387820e+00
-4.29439723e-01 1.16944218e+00 -7.56845295e-01 -6.47853076e-01
-3.27881694e-01 -5.49686372e-01 7.10247636e-01 1.38083827e+00
2.46886745e-01 -5.61411619e-01 5.57749160e-02 1.14666378e+00
6.46051671e-03 -6.49067089e-02 -1.40189922e+00 -6.69941783e-01
5.83269954e-01 1.40442336e+00 -1.57013428e+00 -4.12785858e-01
1.12477187e-02 8.31298053e-01 5.97391784e-01 2.72824258e-01
-7.65439808e-01 -2.66575813e-01 -1.64813042e-01 -1.69565305e-01
5.16719341e-01 -6.30594611e-01 5.25039174e-02 -9.21336770e-01
-5.26015401e-01 -5.85337162e-01 6.63371325e-01 -6.43093288e-01
-1.48561072e+00 1.11727066e-01 5.69563270e-01 -5.56399226e-01
-6.31647229e-01 -8.99736226e-01 -9.86943901e-01 6.01079226e-01
-5.04720986e-01 -8.03201616e-01 -1.91797674e-01 4.39613253e-01
-3.19028012e-02 -2.28066042e-01 8.54852140e-01 -1.69146970e-01
-3.40292305e-01 3.21605057e-01 2.37270236e-01 1.09491147e-01
2.85361528e-01 -1.51846552e+00 7.50689089e-01 7.11758375e-01
7.62299776e-01 6.28968656e-01 1.18736649e+00 -1.15984786e+00
-1.61595476e+00 -6.65441871e-01 3.29928577e-01 -1.97525308e-01
7.39096940e-01 -7.10586011e-01 -6.35336339e-01 7.10208595e-01
1.86030179e-01 -3.47962141e-01 1.01890695e+00 3.49864185e-01
-1.92364022e-01 5.51908553e-01 -1.02732313e+00 5.51121473e-01
1.10185587e+00 -4.91424769e-01 -8.25548396e-02 3.90042514e-01
3.65372077e-02 1.02375355e-02 -1.12109256e+00 5.08197658e-02
7.15738475e-01 -8.76898885e-01 5.84765911e-01 -4.12453681e-01
-1.36211276e-01 -2.89530549e-02 -1.46226123e-01 -1.17344379e+00
-4.20924753e-01 -9.29884136e-01 -6.61639035e-01 1.11705041e+00
2.89463907e-01 -6.00050688e-01 1.21690214e+00 1.99152887e-01
4.81929988e-01 -2.36663043e-01 -1.15868223e+00 -7.51220465e-01
-3.77543032e-01 -4.78292286e-01 -1.90997589e-02 5.73777616e-01
-1.33492664e-01 7.99408257e-01 -6.02362417e-02 -2.89234072e-02
1.30378795e+00 -2.04517215e-01 9.54710484e-01 -1.55105233e+00
-6.19259775e-01 -7.03066945e-01 -9.21297729e-01 -9.36191738e-01
4.39420700e-01 -1.27778614e+00 6.11676695e-03 -1.62519324e+00
4.67612892e-01 4.24418449e-01 2.38545731e-01 -1.66505426e-01
-1.71325758e-01 1.92700103e-01 -5.93131900e-01 -3.45012873e-01
-5.98443568e-01 3.79055589e-01 6.06180489e-01 7.13832974e-02
2.24962413e-01 1.09725654e-01 -1.05217822e-01 7.74797261e-01
8.18422914e-01 -8.85403156e-01 -5.26548207e-01 3.42179537e-01
5.53250432e-01 2.14228481e-01 4.75907207e-01 -7.57647336e-01
2.38504469e-01 -6.86777616e-03 3.99352103e-01 -9.67598259e-01
2.85154670e-01 -4.33328748e-01 5.63602924e-01 6.20308816e-01
-1.52703315e-01 2.45729864e-01 -6.05003051e-02 1.09255517e+00
2.04466641e-01 -5.62859595e-01 6.76528752e-01 -2.12386340e-01
-3.24276626e-01 3.02228391e-01 -1.09304488e+00 2.28063941e-01
7.38764703e-01 -4.16559160e-01 -2.50025332e-01 -7.79929042e-01
-7.10622370e-01 -3.04021180e-01 1.37874603e-01 -2.55429298e-01
1.02738285e+00 -1.16118586e+00 -7.24367082e-01 3.79212461e-02
-5.57571463e-02 -4.83317971e-01 -2.59091556e-01 8.30099285e-01
-7.09143519e-01 -1.11941844e-01 -1.68453425e-01 -7.08534241e-01
-1.31885254e+00 5.12328684e-01 1.61014855e-01 -4.47630733e-01
-4.58210588e-01 8.25495183e-01 -2.19554290e-01 -2.54519790e-01
1.33543298e-01 -8.58221799e-02 -2.65468210e-01 -3.18543464e-02
3.14791977e-01 6.83795094e-01 -2.87639558e-01 -1.66856289e-01
-3.17667395e-01 1.91057533e-01 2.88651079e-01 -4.26599741e-01
1.33691168e+00 1.78619206e-01 -4.57214773e-01 1.13776731e+00
9.95136976e-01 -1.06225342e-01 -7.36090124e-01 1.01383522e-01
2.43835568e-01 5.12279533e-02 1.82781830e-01 -1.91631600e-01
-6.98913753e-01 8.14348340e-01 2.38575935e-01 5.56322575e-01
4.30414826e-01 3.24558884e-01 -1.86371803e-03 3.18604201e-01
3.89224946e-01 -8.44785631e-01 -2.38375589e-02 4.24600184e-01
4.99843955e-01 -7.47648537e-01 3.80937815e-01 -4.45118874e-01
-7.93532878e-02 1.26430285e+00 2.75219828e-01 -6.09109223e-01
1.05106997e+00 6.76858187e-01 -7.15204298e-01 -7.81798184e-01
-7.65258431e-01 -1.72531307e-01 6.58441067e-01 9.84444857e-01
8.25734317e-01 3.46017932e-03 -2.66095400e-01 5.12528047e-02
-1.71337515e-01 -5.04503667e-01 6.39967322e-01 7.71651626e-01
-6.06053889e-01 -7.56256402e-01 -1.19298354e-01 9.64970231e-01
-4.74569738e-01 -5.45245409e-01 -7.19655335e-01 5.65580070e-01
-5.53780079e-01 8.11192632e-01 2.19832972e-01 -1.97748125e-01
-5.97563624e-01 4.63330895e-02 1.09721124e+00 -6.42578602e-01
-1.44807696e-01 1.71040878e-01 2.85571694e-01 -1.52312323e-01
-3.78942460e-01 -5.37144721e-01 -1.00535762e+00 -5.74706316e-01
-7.34250426e-01 2.14191720e-01 6.21007740e-01 8.52711082e-01
-9.50064212e-02 4.60913420e-01 -3.79802734e-02 -8.40334952e-01
-1.31629407e-01 -1.27452230e+00 -1.20668483e+00 -7.28465021e-02
-1.85552910e-01 -6.21473491e-01 -6.45673513e-01 -5.31725725e-03] | [6.92313289642334, 4.487251281738281] |
3e7a7aed-515e-498b-8147-24a72bfa0e64 | responsible-task-automation-empowering-large | 2306.01242 | null | https://arxiv.org/abs/2306.01242v1 | https://arxiv.org/pdf/2306.01242v1.pdf | Responsible Task Automation: Empowering Large Language Models as Responsible Task Automators | The recent success of Large Language Models (LLMs) signifies an impressive stride towards artificial general intelligence. They have shown a promising prospect in automatically completing tasks upon user instructions, functioning as brain-like coordinators. The associated risks will be revealed as we delegate an increasing number of tasks to machines for automated completion. A big question emerges: how can we make machines behave responsibly when helping humans automate tasks as personal copilots? In this paper, we explore this question in depth from the perspectives of feasibility, completeness and security. In specific, we present Responsible Task Automation (ResponsibleTA) as a fundamental framework to facilitate responsible collaboration between LLM-based coordinators and executors for task automation with three empowered capabilities: 1) predicting the feasibility of the commands for executors; 2) verifying the completeness of executors; 3) enhancing the security (e.g., the protection of users' privacy). We further propose and compare two paradigms for implementing the first two capabilities. One is to leverage the generic knowledge of LLMs themselves via prompt engineering while the other is to adopt domain-specific learnable models. Moreover, we introduce a local memory mechanism for achieving the third capability. We evaluate our proposed ResponsibleTA on UI task automation and hope it could bring more attentions to ensuring LLMs more responsible in diverse scenarios. The research project homepage is at https://task-automation-research.github.io/responsible_task_automation. | ['Yan Lu', 'Wenxuan Xie', 'Xiaoyi Zhang', 'Zhizheng Zhang'] | 2023-06-02 | null | null | null | null | ['prompt-engineering'] | ['natural-language-processing'] | [ 2.39090443e-01 6.20903075e-01 1.06445588e-02 -4.12133306e-01
-4.69370008e-01 -7.49177575e-01 7.32467532e-01 -2.61696994e-01
-4.35286105e-01 4.68538642e-01 2.31664523e-01 -4.95076060e-01
-1.93106338e-01 -4.41364616e-01 -5.32113254e-01 -3.02171618e-01
1.00063078e-01 4.87256378e-01 -2.15627760e-01 -9.36334953e-02
2.11251393e-01 3.12554926e-01 -1.56938505e+00 3.57876033e-01
8.93054187e-01 8.90914023e-01 3.82893503e-01 5.77862680e-01
1.15746163e-01 9.47480083e-01 -5.10850906e-01 -2.64029533e-01
3.51566225e-01 2.23645955e-01 -1.18482018e+00 -2.10798904e-01
-1.26215696e-01 -4.40904230e-01 1.00377098e-01 9.00907278e-01
2.97608841e-02 8.15838668e-03 3.59415591e-01 -1.83864522e+00
-3.60882372e-01 7.47576177e-01 8.34770128e-02 -4.85358745e-01
5.97490728e-01 4.29419875e-01 8.20153117e-01 -3.84862483e-01
3.21253002e-01 8.80385041e-01 1.89135820e-01 8.43079686e-01
-8.67631495e-01 -4.79532808e-01 2.53632665e-01 1.79432124e-01
-1.31506026e+00 -6.85113192e-01 5.07023096e-01 -7.26558328e-01
1.03278172e+00 6.88960016e-01 -1.70434210e-02 1.30640280e+00
3.32105279e-01 8.97267222e-01 1.14370453e+00 -5.03508270e-01
3.38110328e-01 6.88437641e-01 5.81971407e-01 7.31579125e-01
3.03450376e-01 5.12531213e-02 -6.55985653e-01 -3.29690725e-01
5.28350234e-01 -1.98755208e-02 -3.71221930e-01 -3.46681207e-01
-1.40605056e+00 5.00341475e-01 -1.80474415e-01 5.23367703e-01
-4.33155894e-01 -5.33783343e-03 4.54629362e-01 3.43867689e-01
-4.38746698e-02 8.26948047e-01 -7.37252057e-01 -2.52504170e-01
-4.67860758e-01 1.81259349e-01 1.06751823e+00 1.26430082e+00
6.47580147e-01 -1.77034408e-01 -2.09320217e-01 1.25713557e-01
2.04981834e-01 2.27554575e-01 4.41697478e-01 -1.07706714e+00
4.17114317e-01 7.39424229e-01 6.57330573e-01 -7.01863229e-01
-6.17113829e-01 -2.59849012e-01 -7.30274022e-01 3.61512393e-01
4.72238451e-01 -2.45734364e-01 -2.48194888e-01 1.85878086e+00
4.01921161e-02 -2.72703141e-01 -1.26746409e-02 8.96595359e-01
6.23392388e-02 3.89904201e-01 2.21994117e-01 -7.02106357e-02
1.75014555e+00 -6.90837443e-01 -8.43079865e-01 -3.93331140e-01
8.08895290e-01 -4.39319432e-01 1.39540637e+00 7.37193823e-01
-8.84576678e-01 -7.15344489e-01 -9.53750014e-01 1.00271299e-01
-3.56791705e-01 5.02110779e-01 1.02371871e+00 8.63801897e-01
-9.05014455e-01 4.40789074e-01 -8.35671365e-01 -2.46443033e-01
2.40936682e-01 5.70180714e-01 -4.41111952e-01 3.25254679e-01
-1.19784319e+00 8.64285111e-01 4.08737123e-01 -5.51638268e-02
-1.04710293e+00 -3.53246689e-01 -7.62859046e-01 1.99905381e-01
7.31735706e-01 -8.50371480e-01 1.44985902e+00 -5.27410448e-01
-1.43366325e+00 5.88900626e-01 -2.71814074e-02 -4.92693275e-01
6.71863198e-01 -5.15043557e-01 -1.47087604e-01 -3.09233040e-01
7.17408881e-02 5.22281289e-01 8.37199926e-01 -1.16567791e+00
-6.91130400e-01 -5.37924945e-01 4.03781861e-01 7.99127370e-02
-7.08344519e-01 2.51296423e-02 7.79958442e-02 -2.30753258e-01
-4.34393317e-01 -9.55016792e-01 1.08694599e-03 -2.56218940e-01
-5.96633077e-01 -5.12499094e-01 6.46176398e-01 -6.00570977e-01
1.37842989e+00 -2.11073399e+00 1.18131310e-01 1.52296275e-01
4.65870231e-01 3.71807933e-01 -1.39873713e-01 4.01956588e-01
1.58061102e-01 2.72148341e-01 1.00103706e-01 -4.41886753e-01
4.64118510e-01 -4.41572033e-02 -4.66002464e-01 9.77704227e-02
-4.37020790e-03 9.26083982e-01 -5.93993008e-01 -1.06611125e-01
2.37778664e-01 1.26976550e-01 -3.55311096e-01 5.84112287e-01
-9.40461606e-02 4.97486353e-01 -6.82827592e-01 5.05694568e-01
4.05601501e-01 -1.15648814e-01 4.56611276e-01 6.60969242e-02
-1.42418846e-01 3.15217972e-01 -1.11104250e+00 1.47139204e+00
-6.29611611e-01 1.90655872e-01 4.90604728e-01 -5.84450901e-01
7.44336367e-01 5.39813578e-01 -7.58061558e-02 -3.57625961e-01
2.11173326e-01 -3.89882959e-02 -7.41799995e-02 -5.00805616e-01
3.83785009e-01 3.53159517e-01 -3.55976522e-01 9.32711542e-01
-2.88091868e-01 1.62206382e-01 -2.15574607e-01 2.40165055e-01
1.15528667e+00 2.02314898e-01 5.76972902e-01 -2.17825025e-01
7.48823941e-01 -9.33494568e-02 2.56809473e-01 8.48143518e-01
-5.12315989e-01 -6.79886192e-02 4.43905562e-01 -6.61297977e-01
-6.34144127e-01 -5.35616040e-01 4.30154651e-01 1.30434072e+00
-2.88919568e-01 -6.09267294e-01 -1.16980731e+00 -9.35149729e-01
-2.43163660e-01 1.30243266e+00 -3.81991118e-01 -2.88207918e-01
-3.37202579e-01 6.98013529e-02 7.06774056e-01 4.81992006e-01
5.97556531e-01 -1.25914884e+00 -1.29164219e+00 -1.76827088e-01
-4.97570515e-01 -1.19103038e+00 -5.47686338e-01 4.63748246e-01
-4.83462304e-01 -9.57517803e-01 -4.13374417e-02 -3.60591143e-01
6.72371864e-01 3.40028524e-01 7.54585326e-01 1.76629156e-01
-5.58776334e-02 5.97540796e-01 -1.47103310e-01 -6.81968510e-01
-5.56759357e-01 2.16950089e-01 6.35345638e-01 -3.74052748e-02
4.62511569e-01 -5.27558565e-01 -1.86653659e-01 5.79476774e-01
-6.81536376e-01 3.48328322e-01 6.50044501e-01 4.46014345e-01
-1.23703398e-03 1.41639993e-01 4.26144034e-01 -1.02947330e+00
1.03377068e+00 -1.67048141e-01 -6.13667190e-01 5.20308554e-01
-8.30690920e-01 2.40600184e-01 8.47658873e-01 -2.55682081e-01
-1.19250739e+00 3.22363615e-01 2.23247558e-01 -1.44340739e-01
-7.48762608e-01 1.62184373e-01 -4.96337086e-01 1.38024926e-01
6.60509050e-01 3.58719856e-01 -4.75824848e-02 -2.59831756e-01
2.61871636e-01 9.27159607e-01 4.40193832e-01 -9.64937627e-01
9.11602676e-01 1.63552016e-01 -3.44186276e-01 -5.00499487e-01
-4.53960806e-01 -3.53532195e-01 -4.70016479e-01 -1.94128007e-01
8.23156595e-01 -6.08580828e-01 -1.53096485e+00 3.74283642e-01
-1.41744876e+00 -4.80113059e-01 -4.14996296e-02 1.80386543e-01
-7.37809718e-01 4.50856298e-01 -3.16753477e-01 -1.30597687e+00
-6.16045296e-01 -1.21257460e+00 1.01752639e+00 5.48806489e-02
-8.47827733e-01 -6.12169504e-01 -3.69235367e-01 8.53528738e-01
5.31858623e-01 -2.59062767e-01 1.15266347e+00 -9.62312818e-01
-6.09291017e-01 -3.11378866e-01 5.54680228e-02 3.33267212e-01
9.91019160e-02 -6.28687561e-01 -1.22977495e+00 -2.72823721e-01
3.85296106e-01 -4.68457401e-01 4.32295576e-02 1.17381036e-01
1.11057687e+00 -6.38787150e-01 -2.76085049e-01 2.77392358e-01
7.09714234e-01 2.22967491e-01 5.56425154e-01 3.05255413e-01
3.51199448e-01 1.01759410e+00 8.41214836e-01 5.33684254e-01
3.79195094e-01 6.42285287e-01 4.07393634e-01 3.14767540e-01
3.04531783e-01 -2.71216482e-01 6.27544224e-01 4.02585864e-01
-3.97656977e-01 -8.86135474e-02 -9.50620353e-01 3.23035792e-02
-2.10394335e+00 -6.74042106e-01 -2.99810572e-03 2.33736062e+00
5.77524185e-01 2.32306197e-01 -9.82546508e-02 1.18044831e-01
4.76683319e-01 -1.80441946e-01 -4.21700984e-01 -3.83039474e-01
5.86721539e-01 -2.33321920e-01 1.50810704e-01 3.92191470e-01
-1.06145024e+00 8.28858435e-01 5.42113352e+00 5.84941506e-01
-7.01629400e-01 2.17167512e-01 5.62027037e-01 2.75558144e-01
-1.01554751e-01 5.69383092e-02 -7.87359297e-01 1.74511075e-01
9.47417736e-01 4.57637496e-02 8.12351644e-01 1.24800205e+00
1.76871583e-01 -1.62537992e-01 -1.43421257e+00 8.50293517e-01
-1.61389276e-01 -1.09887171e+00 1.02468506e-02 1.70928493e-01
-4.29650880e-02 -5.18493831e-01 -3.31362523e-02 4.55416054e-01
1.19677119e-01 -9.32992280e-01 9.59888577e-01 5.90201080e-01
6.80086732e-01 -5.98962545e-01 7.98172176e-01 1.15908289e+00
-9.42254245e-01 -4.06066537e-01 -1.58596739e-01 -5.57470143e-01
-4.18412350e-02 1.27177551e-01 -1.05983210e+00 5.36569536e-01
4.65560496e-01 -4.59790528e-02 -4.62055624e-01 2.14224175e-01
-4.48815912e-01 2.83140630e-01 -2.33305395e-02 -6.13007508e-02
-1.66492671e-01 -6.37872741e-02 4.91150826e-01 8.25925410e-01
4.48696017e-02 1.37399063e-02 1.42199203e-01 1.07437742e+00
1.85136780e-01 -2.81361878e-01 -7.83505917e-01 -3.55294257e-01
7.01406181e-01 1.53279984e+00 -4.22145009e-01 -1.62196323e-01
-1.90805167e-01 1.05750322e+00 4.62588280e-01 3.10726553e-01
-6.75616741e-01 -3.29246610e-01 6.82877898e-01 8.78517777e-02
-4.90826130e-01 -4.39188123e-01 -4.61114138e-01 -1.05539727e+00
1.77300677e-01 -1.19946587e+00 2.88253069e-01 -8.69170964e-01
-9.21357214e-01 7.53119767e-01 2.73294318e-02 -8.58633816e-01
-3.49962592e-01 -7.29112566e-01 -5.06996810e-01 7.96074629e-01
-1.13310599e+00 -1.40816236e+00 -4.35466141e-01 7.25610793e-01
5.52083313e-01 -2.61186957e-01 1.26764679e+00 -1.26619428e-01
-5.25570035e-01 5.89450300e-01 -7.79656231e-01 -4.65674475e-02
6.97874963e-01 -9.74835455e-01 4.19136286e-01 8.44272614e-01
9.02833641e-02 1.25841057e+00 6.12252772e-01 -6.57067358e-01
-1.86050010e+00 -8.86517763e-01 1.07068944e+00 -1.02254117e+00
4.07343179e-01 -8.93631458e-01 -6.10490203e-01 1.08118176e+00
7.40014538e-02 -6.23872817e-01 5.93350649e-01 1.90083697e-01
-3.03188056e-01 1.32254288e-01 -1.16710413e+00 5.21607339e-01
1.07880008e+00 -5.98222792e-01 -5.68527341e-01 4.34482098e-01
8.07849407e-01 -1.19436495e-02 -5.56366682e-01 1.43871335e-02
5.87889791e-01 -9.81459439e-01 6.48344338e-01 -5.86228251e-01
-4.06559557e-02 -3.84943992e-01 -9.64343622e-02 -9.34828997e-01
-3.55435342e-01 -1.13040924e+00 -1.67872697e-01 1.24362743e+00
1.96714416e-01 -8.98745835e-01 6.06234193e-01 1.47295630e+00
-1.99984968e-01 -4.37800109e-01 -7.35292852e-01 -7.54132628e-01
-4.86984968e-01 -8.40168953e-01 8.01246464e-01 8.26697171e-01
5.95250666e-01 3.20076764e-01 -6.55928433e-01 2.68466324e-01
3.00578952e-01 -2.63107538e-01 1.00320029e+00 -1.28286219e+00
-3.39810878e-01 -1.89217523e-01 2.72170715e-02 -8.15425754e-01
2.39981145e-01 -8.71313155e-01 9.52860992e-03 -1.09378171e+00
1.23792686e-01 -3.42807382e-01 -1.26552746e-01 1.00328624e+00
1.71056956e-01 -4.60956454e-01 3.30566496e-01 3.38885456e-01
-8.61085117e-01 5.74948899e-02 8.67455304e-01 -1.88986100e-02
-2.85462469e-01 4.29460704e-01 -1.00989389e+00 8.06103051e-01
1.00080824e+00 -2.04683572e-01 -5.26363075e-01 -3.14019501e-01
2.77677655e-01 7.78557211e-02 4.74982351e-01 -1.16656017e+00
6.35673881e-01 -2.83683419e-01 2.14059111e-02 7.70682693e-02
1.70292199e-01 -1.20925808e+00 2.32541263e-01 5.88763833e-01
-5.93807936e-01 -2.11093441e-01 1.35832429e-02 3.67702395e-01
2.78994024e-01 -2.15652138e-01 2.00827509e-01 -1.62995517e-01
-7.05671489e-01 -1.04437299e-01 -7.98947573e-01 -5.45231998e-01
1.15647566e+00 -7.36711621e-02 -4.03225213e-01 -4.95598435e-01
-6.58262849e-01 2.96971619e-01 3.16010922e-01 5.28056741e-01
4.84961540e-01 -7.94559717e-01 -1.95990190e-01 5.10790467e-01
2.41588816e-01 -3.36633474e-01 1.70157835e-01 8.68504584e-01
1.09903589e-01 8.58015835e-01 -3.17587346e-01 -1.14513218e-01
-1.31455624e+00 6.81749046e-01 4.09369776e-03 -4.95188624e-01
-2.19506174e-01 5.82852900e-01 4.55082476e-01 -5.88948607e-01
6.74314439e-01 -4.24268216e-01 -8.56811926e-02 -3.52223277e-01
8.16475153e-01 3.69635969e-01 6.26969784e-02 -7.41800247e-03
-5.74895263e-01 -3.36828195e-02 -2.05150634e-01 1.07891552e-01
1.17223060e+00 -1.55657768e-01 -1.57614231e-01 1.75161093e-01
3.30237120e-01 -4.87205572e-02 -1.04280961e+00 2.27310993e-02
2.52637774e-01 -2.10960209e-01 -1.52598426e-01 -1.27304804e+00
-5.02873182e-01 9.01075363e-01 3.31823111e-01 3.59394997e-01
9.30300653e-01 -1.25707686e-01 6.39394283e-01 8.29390168e-01
1.12569427e+00 -1.07271016e+00 -4.13796492e-02 5.10911465e-01
8.38914096e-01 -1.20497954e+00 -2.09457546e-01 -3.78140807e-01
-1.04544234e+00 1.20096600e+00 8.39434266e-01 4.83419925e-01
1.84236810e-01 4.76279467e-01 -3.53639796e-02 -1.62554488e-01
-9.08907235e-01 1.08385146e-01 1.80361807e-01 9.97785985e-01
3.60795975e-01 4.10420507e-01 -1.38994008e-01 1.49521804e+00
-1.07255459e-01 3.22810054e-01 4.75615352e-01 9.97344971e-01
-4.46534216e-01 -1.20385838e+00 -4.82612252e-01 3.02797377e-01
-7.07121566e-02 1.10268019e-01 -6.54703796e-01 5.58159530e-01
1.18075140e-01 1.27317619e+00 -5.75976014e-01 -7.69804418e-01
2.98932523e-01 5.97229004e-01 4.89839502e-02 -6.42433822e-01
-8.06955576e-01 -4.19679821e-01 3.23842198e-01 -9.46246028e-01
1.56374946e-01 -3.91476125e-01 -1.14381909e+00 -2.96548456e-01
-1.37313381e-01 1.30835786e-01 7.52532244e-01 9.45335865e-01
7.16276169e-01 3.32274407e-01 3.27575088e-01 -7.79030144e-01
-1.01975620e+00 -7.85452306e-01 -3.54742825e-01 1.18294910e-01
3.02105527e-02 -3.98821086e-01 -2.55693048e-01 -5.73488064e-02] | [10.499284744262695, 7.737028121948242] |
d7a4b79e-650a-4e4c-baaf-d8ba48b255ce | threshnet-segmentation-refinement-inspired-by | 2211.06560 | null | https://arxiv.org/abs/2211.06560v2 | https://arxiv.org/pdf/2211.06560v2.pdf | ThreshNet: Segmentation Refinement Inspired by Region-Specific Thresholding | We present ThreshNet, a post-processing method to refine the output of neural networks designed for binary segmentation tasks. ThreshNet uses the confidence map produced by a base network along with global and local patch information to significantly improve the performance of even state-of-the-art methods. Binary segmentation models typically convert confidence maps into predictions by thresholding the confidence scores at 0.5 (or some other fixed number). However, we observe that the best threshold is image-dependent and often even region-specific -- different parts of the image benefit from using different thresholds. Thus ThreshNet takes a trained segmentation model and learns to correct its predictions by using a memory-efficient post-processing architecture that incorporates region-specific thresholds as part of the training mechanism. Our experiments show that ThreshNet consistently improves over current the state-of-the-art methods in binary segmentation and saliency detection, typically by 3 to 5% in mIoU and mBA. | ['Daniel Kifer', 'Chaopeng Shen', 'Savinay Nagendra'] | 2022-11-12 | null | null | null | null | ['saliency-detection'] | ['computer-vision'] | [ 6.60681427e-01 3.31872046e-01 -3.09332341e-01 -6.45955861e-01
-7.55506217e-01 -4.55795497e-01 2.76390791e-01 3.73508453e-01
-5.74596941e-01 4.94023234e-01 -1.83289781e-01 -2.95749575e-01
3.66911918e-01 -6.83795393e-01 -9.53049004e-01 -3.71890306e-01
8.01131129e-02 4.57909942e-01 1.34057939e+00 -2.18586281e-01
5.61045647e-01 1.74192816e-01 -1.64002466e+00 6.68573201e-01
9.36422229e-01 1.29405224e+00 4.55524743e-01 8.76935542e-01
9.48660299e-02 4.67254609e-01 -6.43607795e-01 -1.91104993e-01
3.80496591e-01 -5.04291356e-01 -8.09822679e-01 -1.66686356e-01
8.28786731e-01 -6.73678666e-02 8.57613906e-02 1.18627334e+00
1.24739006e-01 3.71448137e-02 5.98988414e-01 -9.54329669e-01
-2.99714684e-01 9.50857341e-01 -6.96368158e-01 5.24339914e-01
-1.55956268e-01 2.00935408e-01 9.15654540e-01 -8.61631572e-01
5.27901530e-01 1.07756341e+00 9.26363111e-01 3.40335548e-01
-1.41117513e+00 -5.05679131e-01 1.20663039e-01 1.49197623e-01
-1.26972818e+00 -3.40800703e-01 2.81376809e-01 -1.33034378e-01
9.94115114e-01 2.37151444e-01 8.55999231e-01 3.40147913e-01
3.52943718e-01 8.55434418e-01 1.18865573e+00 -3.42503458e-01
3.69355857e-01 -9.68057662e-02 4.47995365e-02 6.97478950e-01
2.59583831e-01 -4.73427027e-02 -4.31244642e-01 3.29921752e-01
1.05199265e+00 -3.59808832e-01 3.84256952e-02 -3.58831435e-01
-1.17085028e+00 6.87823415e-01 1.08342040e+00 3.10988218e-01
-4.30531472e-01 3.16047668e-01 8.49929750e-02 -6.06908686e-02
4.08748239e-01 7.62416244e-01 -5.86133540e-01 9.56498086e-02
-1.94205177e+00 8.13015699e-02 4.02392477e-01 6.88074410e-01
1.00937188e+00 -1.64921712e-02 -4.34570789e-01 6.54273689e-01
8.87862146e-02 7.35727400e-02 4.59320217e-01 -1.26069355e+00
8.08574036e-02 6.14428520e-01 -1.07432738e-01 -7.44874001e-01
-3.47129434e-01 -5.63927531e-01 -3.71577740e-01 6.06848419e-01
6.13034308e-01 -1.77239686e-01 -1.79802561e+00 1.46578324e+00
2.29326740e-01 2.77560890e-01 -3.55977416e-01 1.00147355e+00
6.03175819e-01 5.02171755e-01 3.06596100e-01 3.09870511e-01
1.09163117e+00 -1.27545512e+00 -1.96156397e-01 -8.37162495e-01
2.21438840e-01 -8.08390617e-01 1.02211761e+00 3.50199640e-01
-1.42197883e+00 -6.74944043e-01 -1.37369442e+00 -7.72098601e-02
-4.46326584e-01 5.48321940e-02 5.56026399e-01 5.65280974e-01
-1.57897401e+00 8.52921963e-01 -9.36694622e-01 -2.85875231e-01
7.12524712e-01 5.95053792e-01 6.61793426e-02 2.21230805e-01
-9.07608986e-01 1.00142169e+00 7.23011732e-01 -1.20675154e-01
-7.85368025e-01 -7.50885546e-01 -9.38647985e-01 1.55386031e-01
2.69983053e-01 -5.26528418e-01 1.60553885e+00 -1.36499691e+00
-1.25629652e+00 9.53006208e-01 -1.98265180e-01 -8.10834408e-01
4.82714951e-01 -5.94561584e-02 -2.04087477e-02 4.97817993e-01
3.04279059e-01 1.62171972e+00 1.09168935e+00 -1.09571195e+00
-1.06205988e+00 -1.80488508e-02 -6.60234168e-02 2.12249801e-01
2.37419948e-01 2.63006538e-02 -8.33944678e-01 -7.60477364e-01
5.50472021e-01 -6.85044110e-01 -5.36787152e-01 2.89437063e-02
-2.98693568e-01 5.59010319e-02 8.87432814e-01 -6.31067812e-01
9.64706242e-01 -2.01157618e+00 -1.41158447e-01 3.28043967e-01
1.76728308e-01 2.75198221e-01 -7.34191462e-02 -4.26014215e-01
-1.40202604e-02 2.17556685e-01 -5.59950471e-01 -3.23372185e-01
-3.97431225e-01 4.75077843e-03 3.81084830e-02 2.59810239e-01
6.13336146e-01 1.03696227e+00 -8.74775469e-01 -7.54540980e-01
3.75845879e-01 2.43949816e-01 -4.64239210e-01 -1.65330932e-01
-5.08349299e-01 1.19955033e-01 -3.09444088e-02 6.99588716e-01
5.35800338e-01 -4.10256922e-01 -1.09519124e-01 -1.75864086e-01
-2.64803976e-01 4.68210429e-01 -9.96384621e-01 1.64557028e+00
-6.52311817e-02 9.61633205e-01 2.16443643e-01 -9.13754821e-01
6.64364398e-01 -5.28169349e-02 1.55894935e-01 -6.21368051e-01
3.57693255e-01 3.64301264e-01 1.83442309e-01 1.98678732e-01
8.74139309e-01 1.30846456e-01 7.53409341e-02 2.08877563e-01
2.33625770e-01 -3.36890548e-01 3.13634127e-01 2.37615004e-01
1.02879810e+00 9.09923762e-02 9.52662900e-02 -4.97146338e-01
1.85974702e-01 4.37616050e-01 5.81101537e-01 1.17594254e+00
-4.47338015e-01 1.25263536e+00 3.91883940e-01 -1.12593673e-01
-1.15349269e+00 -1.01640582e+00 -3.18161637e-01 1.45298266e+00
6.00392461e-01 -6.64840341e-02 -1.31357634e+00 -6.76682472e-01
-1.57798633e-01 6.56604230e-01 -8.70490253e-01 -7.97281638e-02
-6.36831999e-01 -5.53683043e-01 4.71617609e-01 9.16552484e-01
8.62760961e-01 -1.36854184e+00 -9.80424941e-01 4.02056843e-01
1.13082923e-01 -1.02656472e+00 -4.78597641e-01 7.02541292e-01
-1.11431420e+00 -8.29645038e-01 -8.24408054e-01 -1.07909739e+00
7.65134394e-01 2.11746812e-01 1.28669572e+00 3.95149261e-01
-2.08245397e-01 -3.66262943e-01 -2.17408121e-01 -1.99255735e-01
-3.02260101e-01 3.89364630e-01 -5.04604638e-01 -4.36883897e-01
5.45179211e-02 -3.22253495e-01 -8.62049401e-01 5.02211511e-01
-7.57383466e-01 4.19517308e-01 8.38511169e-01 6.49619460e-01
8.36705446e-01 -8.66021141e-02 3.36652249e-01 -8.91092420e-01
3.17214251e-01 -2.23769806e-02 -6.40267432e-01 2.15237916e-01
-6.31725669e-01 1.56282857e-01 1.48868695e-01 -4.17691380e-01
-7.81280518e-01 3.12925935e-01 -9.19318795e-02 -2.32631758e-01
-2.88122594e-01 1.96554258e-01 1.22773305e-01 -3.78312439e-01
7.78693736e-01 -6.33884519e-02 -3.33076060e-01 -1.35075927e-01
4.45016891e-01 2.53576428e-01 1.12462366e+00 -1.70861229e-01
4.08965081e-01 2.60372192e-01 -2.06983685e-01 -3.58576089e-01
-1.02238262e+00 -4.76887256e-01 -8.01819086e-01 -2.03003943e-01
1.09512699e+00 -6.46371663e-01 -1.16626576e-01 6.31524086e-01
-1.01203823e+00 -8.63202453e-01 -2.97296047e-01 5.00476025e-02
-4.54113930e-01 -1.91061005e-01 -7.27468848e-01 -4.24570441e-01
-3.57204705e-01 -1.23463738e+00 1.01914823e+00 7.90271223e-01
-2.61618614e-01 -6.66405320e-01 -2.55092084e-01 1.59307688e-01
5.07032990e-01 1.41327813e-01 5.84306359e-01 -6.20615959e-01
-5.71306884e-01 1.11318275e-01 -6.82835817e-01 1.88120827e-01
-1.83905110e-01 1.55875221e-01 -9.05292451e-01 6.70972094e-02
-3.81062955e-01 -2.51596540e-01 1.36815083e+00 9.91240025e-01
1.21640134e+00 -8.13513398e-02 -4.72369522e-01 5.91048002e-01
1.28658712e+00 -9.83885489e-03 7.11395800e-01 5.41151643e-01
4.77567941e-01 2.09925786e-01 7.94121087e-01 1.34403063e-02
2.35019282e-01 3.42375427e-01 5.69621325e-01 -5.38285911e-01
-4.51771826e-01 -1.67465076e-01 -7.96419755e-03 8.03185403e-02
1.64395928e-01 5.83344623e-02 -8.39248717e-01 7.93934286e-01
-1.92141688e+00 -8.39181304e-01 -7.75292292e-02 2.03380322e+00
1.34132493e+00 8.25410247e-01 2.49967948e-01 1.94293126e-01
1.10457146e+00 1.39732808e-01 -7.69276083e-01 -5.05625844e-01
-1.97680682e-01 6.14851594e-01 9.68445897e-01 6.16987288e-01
-1.41227472e+00 1.57358837e+00 7.37591887e+00 8.90897632e-01
-1.39488053e+00 -4.42359410e-02 1.12485695e+00 1.38222292e-01
5.14227413e-02 4.80590463e-02 -7.52344549e-01 2.94076771e-01
7.50351965e-01 3.38762313e-01 2.51762301e-01 9.66402650e-01
-1.10054724e-02 -7.55514205e-01 -9.17864144e-01 5.87071300e-01
2.12603752e-02 -1.57791078e+00 -3.64902258e-01 -3.08842629e-01
1.10237324e+00 4.07058597e-01 2.65942156e-01 8.60758796e-02
4.87638891e-01 -1.12101483e+00 1.25299001e+00 1.27514184e-01
6.35257900e-01 -5.83876371e-01 6.86886370e-01 1.21000916e-01
-1.04283988e+00 2.07031325e-01 -3.20394427e-01 -7.69656559e-04
1.26792610e-01 5.82481623e-01 -1.13988233e+00 -2.50976980e-01
1.00983739e+00 5.86384714e-01 -1.12479520e+00 1.54793251e+00
-4.50390846e-01 8.20919752e-01 -5.98215044e-01 3.19114402e-02
5.36686718e-01 2.66412795e-01 1.79527685e-01 1.48544812e+00
9.02065337e-02 7.14489296e-02 -6.86022826e-03 1.02094710e+00
-1.28136918e-01 -3.02227825e-01 1.10815138e-01 3.05100888e-01
4.71348971e-01 1.30991971e+00 -1.51532066e+00 -7.33997524e-01
4.37973142e-02 1.01730359e+00 1.05643779e-01 1.81609675e-01
-8.85206699e-01 -5.64790964e-01 3.15912098e-01 2.23517910e-01
9.92543101e-01 -9.31387916e-02 -1.19440377e+00 -6.19585514e-01
-3.61755282e-01 -5.87937891e-01 2.12638587e-01 -1.01691115e+00
-6.42903805e-01 6.63558900e-01 -4.56545986e-02 -7.00053155e-01
-2.10319474e-01 -5.33510029e-01 -7.36136317e-01 7.59593189e-01
-1.39538276e+00 -9.14892673e-01 -2.07613185e-01 2.53726602e-01
7.25427151e-01 3.23226213e-01 2.95885533e-01 -1.24671437e-01
-3.97857249e-01 4.49217826e-01 -3.08440387e-01 1.02351733e-01
5.53998232e-01 -1.52163446e+00 9.17990804e-01 1.14313602e+00
2.85494179e-01 3.53718907e-01 9.00781929e-01 -7.56620705e-01
-3.34568799e-01 -9.32332933e-01 7.10503161e-01 -1.78132117e-01
4.61652666e-01 -2.30173960e-01 -1.01549053e+00 4.79032040e-01
3.83762062e-01 1.98509723e-01 8.38384032e-02 -1.74472690e-01
-2.33475506e-01 1.60151154e-01 -1.44057477e+00 6.39796674e-01
8.73680592e-01 -2.61141628e-01 -6.77645683e-01 -1.42181247e-01
7.86082387e-01 -7.46326208e-01 -2.69760221e-01 5.88410914e-01
5.17769158e-01 -1.23687267e+00 8.78852010e-01 -1.13568522e-01
4.83871430e-01 -5.18981457e-01 1.82611287e-01 -1.08890915e+00
-3.84470493e-01 -2.97323316e-01 7.19200224e-02 8.07785869e-01
9.34174895e-01 -1.91748872e-01 1.18510032e+00 6.91119552e-01
-2.56390065e-01 -7.71578312e-01 -8.76142919e-01 -2.63580680e-01
-5.00465743e-02 -5.39169073e-01 4.76393551e-01 4.24289286e-01
-2.22826943e-01 7.38403052e-02 2.11363778e-01 1.88506231e-01
4.85294461e-01 6.82692602e-02 3.05747926e-01 -1.11110806e+00
-1.44830003e-01 -7.98037648e-01 -5.24221718e-01 -1.28913796e+00
-2.46198922e-01 -4.16183770e-01 6.65544987e-01 -1.52429414e+00
-6.85280859e-02 -5.34547329e-01 -5.00684261e-01 9.04328942e-01
-3.99455905e-01 9.83125985e-01 2.37975717e-01 -2.52390690e-02
-8.80711854e-01 -2.69379485e-02 1.19919980e+00 -2.53063500e-01
-4.00669634e-01 -9.20322537e-02 -7.16257215e-01 8.97464216e-01
9.24864471e-01 -5.62615991e-01 -1.17301688e-01 -3.59293997e-01
-3.52049172e-02 -3.23447883e-01 5.09638906e-01 -1.46049464e+00
3.96199077e-01 -2.08517849e-01 7.78787434e-01 -8.63065302e-01
2.19811395e-01 -4.20572430e-01 -3.46154898e-01 4.91460294e-01
-4.50501204e-01 -5.32433353e-02 4.53416824e-01 1.65819719e-01
-1.24590620e-01 -3.20483953e-01 1.17504275e+00 -1.86378434e-01
-1.19323432e+00 -1.18580081e-01 -4.47719306e-01 1.60535380e-01
7.95939803e-01 -7.26078451e-01 -2.60468662e-01 -1.96453914e-01
-6.94161236e-01 1.82916999e-01 6.88677430e-01 1.99831814e-01
6.76865041e-01 -8.29951406e-01 -4.18481559e-01 3.33383143e-01
-3.00754964e-01 3.52781832e-01 -2.69251615e-01 7.70998836e-01
-8.14020634e-01 1.03432789e-01 -4.26147997e-01 -1.01246715e+00
-1.14000618e+00 6.66946992e-02 4.27556634e-01 2.38661319e-02
-3.75601113e-01 1.26574039e+00 5.83498999e-02 4.54735160e-02
2.56567299e-01 -8.10455799e-01 -4.32741977e-02 -3.36683035e-01
3.25372487e-01 5.57248257e-02 1.89635798e-01 -5.88096142e-01
-4.61427033e-01 5.00918806e-01 -1.71538934e-01 -5.73949397e-01
1.11572325e+00 1.81186218e-02 6.41955584e-02 1.58533022e-01
7.34056771e-01 -3.87437046e-01 -1.80258083e+00 2.90033575e-02
4.53860313e-02 -3.15995753e-01 4.51566666e-01 -1.23603845e+00
-1.31911659e+00 7.22468376e-01 5.37323475e-01 6.08104542e-02
1.20450377e+00 1.75018191e-01 8.34966362e-01 4.80672047e-02
2.00389713e-01 -1.37483072e+00 3.40781994e-02 5.47241807e-01
6.03826761e-01 -1.24716890e+00 3.48473974e-02 -3.99797857e-01
-5.70630193e-01 8.56564760e-01 8.56256604e-01 -4.64536786e-01
5.84366977e-01 3.08405280e-01 2.65329272e-01 8.77327695e-02
-3.75131190e-01 -3.54561001e-01 5.58511257e-01 5.67392826e-01
3.18370432e-01 -8.04778747e-03 -2.11392611e-01 3.57372284e-01
-4.11888272e-01 -1.42129093e-01 2.51486987e-01 9.53877091e-01
-1.00627697e+00 -9.48131502e-01 -4.77648646e-01 5.76782227e-01
-4.87661779e-01 -3.96135420e-01 -4.72754896e-01 5.62547624e-01
4.22713995e-01 9.06221330e-01 3.39670211e-01 -4.07588154e-01
-7.97932073e-02 -2.12965906e-01 3.41912121e-01 -7.45119572e-01
-9.00910020e-01 2.45737538e-01 -2.37093866e-01 -7.54291117e-01
-3.91157955e-01 -6.62095368e-01 -1.66626227e+00 3.80883329e-02
-5.84201574e-01 -1.36568114e-01 6.64421678e-01 9.51920092e-01
2.70606846e-01 5.97392738e-01 5.01960330e-02 -1.36584258e+00
4.85156989e-03 -9.40815210e-01 -2.59882987e-01 4.95232269e-02
4.64919806e-01 -4.81675744e-01 -1.90794393e-01 1.88089058e-01] | [9.540233612060547, 0.2643972933292389] |
edb3ea14-d99f-4a1c-8488-9f68543d9e68 | abnormal-event-detection-in-videos-using-1 | 1708.09644 | null | http://arxiv.org/abs/1708.09644v1 | http://arxiv.org/pdf/1708.09644v1.pdf | Abnormal Event Detection in Videos using Generative Adversarial Nets | In this paper we address the abnormality detection problem in crowded scenes.
We propose to use Generative Adversarial Nets (GANs), which are trained using
normal frames and corresponding optical-flow images in order to learn an
internal representation of the scene normality. Since our GANs are trained with
only normal data, they are not able to generate abnormal events. At testing
time the real data are compared with both the appearance and the motion
representations reconstructed by our GANs and abnormal areas are detected by
computing local differences. Experimental results on challenging abnormality
detection datasets show the superiority of the proposed method compared to the
state of the art in both frame-level and pixel-level abnormality detection
tasks. | ['Moin Nabi', 'Mahdyar Ravanbakhsh', 'Lucio Marcenaro', 'Enver Sangineto', 'Carlo Regazzoni', 'Nicu Sebe'] | 2017-08-31 | null | null | null | null | ['abnormal-event-detection-in-video', 'semi-supervised-anomaly-detection', 'abnormal-event-detection-in-video'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 4.79061395e-01 7.33351111e-02 5.66484511e-01 2.44769175e-02
-1.89812377e-01 -2.98818409e-01 5.56318879e-01 -1.48769364e-01
-2.80009389e-01 7.86280513e-01 5.10582179e-02 2.01302722e-01
6.42170787e-01 -8.85801196e-01 -7.26210296e-01 -7.33700156e-01
5.38846441e-02 2.24975333e-01 4.58723426e-01 -7.05182552e-02
3.32956724e-02 3.79075080e-01 -1.61165249e+00 4.05710578e-01
9.22582567e-01 8.20021033e-01 -3.34148079e-01 1.14417267e+00
1.81972086e-02 1.08961010e+00 -9.59262133e-01 -1.98522806e-01
5.08449912e-01 -9.17947710e-01 -6.32428467e-01 6.69616818e-01
6.68595493e-01 -7.69256592e-01 -5.71897388e-01 1.19263327e+00
3.65740627e-01 3.91074419e-01 5.11293709e-01 -1.26024508e+00
-5.28112292e-01 -4.43495274e-01 -4.99357134e-01 8.16592634e-01
7.72137225e-01 5.67057967e-01 4.04790074e-01 -6.86778784e-01
7.96815217e-01 1.20501018e+00 2.96810329e-01 8.59706521e-01
-1.03310025e+00 -2.15218440e-01 1.41438082e-01 3.06268990e-01
-8.85568321e-01 -3.82139713e-01 9.44367766e-01 -5.57920277e-01
7.20791399e-01 1.90905392e-01 9.52072501e-01 1.27596736e+00
3.76877636e-01 8.39151084e-01 8.84956002e-01 -3.58579576e-01
3.94891560e-01 -5.03766418e-01 -5.61711609e-01 8.21350515e-01
3.63525271e-01 3.29459190e-01 -3.28839958e-01 2.84488443e-02
1.04625988e+00 2.50636488e-01 -4.83438671e-01 -1.96470380e-01
-1.12233484e+00 6.70463920e-01 3.31245929e-01 3.82275611e-01
-7.17688322e-01 2.87683189e-01 4.84060198e-01 3.54967773e-01
7.38869071e-01 1.17358059e-01 1.45896211e-01 -8.00449848e-02
-8.26295257e-01 2.84782410e-01 3.29672754e-01 3.11025053e-01
3.78394246e-01 7.10080326e-01 -4.32624370e-01 4.69005972e-01
5.88489370e-03 4.61663842e-01 8.57319176e-01 -9.84470069e-01
2.31788412e-01 7.02581942e-01 1.28673524e-01 -1.22922814e+00
-2.07775027e-01 -2.73833901e-01 -9.12447155e-01 6.42235458e-01
5.30097306e-01 -1.46171406e-01 -1.36182308e+00 1.41022277e+00
3.02738369e-01 7.68236697e-01 2.66806722e-01 9.04163837e-01
6.29526675e-01 6.01399422e-01 -8.58790874e-02 -8.18502977e-02
6.60300314e-01 -9.21552837e-01 -9.23249066e-01 -4.23865139e-01
2.98613995e-01 -7.24646509e-01 8.28254700e-01 3.67327154e-01
-1.56916583e+00 -8.21557999e-01 -8.55823874e-01 2.28079811e-01
-2.23697856e-01 -2.09431559e-01 1.36665896e-01 6.06133282e-01
-1.24419034e+00 3.51792663e-01 -1.17994761e+00 -3.32368940e-01
4.94315773e-01 2.14564912e-02 -4.11975205e-01 -3.02746862e-01
-8.24911714e-01 5.78298509e-01 3.73857528e-01 1.88833728e-01
-1.31405616e+00 -1.74847394e-01 -1.09408343e+00 -3.42185438e-01
-4.67439517e-02 -8.81097615e-01 8.09443414e-01 -1.32241774e+00
-1.35639143e+00 9.92010832e-01 -2.65296102e-01 -4.90186393e-01
9.63852584e-01 -2.13777676e-01 -3.67642730e-01 5.24074912e-01
-7.49755427e-02 4.91722763e-01 1.16118896e+00 -1.21507871e+00
-6.53364897e-01 -6.83763549e-02 1.96603127e-02 1.33565664e-01
-1.01636807e-02 -1.21159084e-01 -4.03943121e-01 -1.05374086e+00
2.34879643e-01 -6.38600945e-01 -2.95544893e-01 7.89911151e-02
-4.88810152e-01 3.83243382e-01 1.27914071e+00 -1.10022283e+00
6.96415663e-01 -1.98930871e+00 8.74006450e-02 2.38979921e-01
1.72742322e-01 6.11539960e-01 -4.02996004e-01 -5.28348200e-02
-1.89550534e-01 -2.30618238e-01 -4.61906552e-01 -4.41539049e-01
-5.71626663e-01 6.12999439e-01 -1.28840163e-01 8.80709350e-01
4.27368551e-01 1.05090201e+00 -1.08381736e+00 -3.53961855e-01
7.76478946e-01 4.56179559e-01 -6.13313735e-01 7.94170737e-01
-1.34838268e-01 1.20506024e+00 -1.88801885e-01 8.23343337e-01
6.71344042e-01 2.90953010e-01 -4.03798610e-01 2.39406392e-01
5.43878973e-01 -2.00353846e-01 -9.12670076e-01 1.58091044e+00
-2.44376943e-01 6.31886363e-01 -2.43984967e-01 -1.07267272e+00
7.60823071e-01 3.95616800e-01 5.43891549e-01 -7.26023257e-01
9.61127654e-02 -3.68850902e-02 1.98264852e-01 -7.91215479e-01
1.04388736e-01 -1.95391458e-02 4.79290724e-01 2.41620928e-01
-1.67085771e-02 -1.77389413e-01 3.03437561e-01 -5.71572557e-02
1.48550379e+00 1.76410228e-01 1.00575611e-01 2.99672335e-01
9.52906311e-01 -2.52047569e-01 6.87637210e-01 6.56164706e-01
-4.24057007e-01 1.05456090e+00 4.93389070e-01 -8.11224103e-01
-1.09824800e+00 -1.29629695e+00 2.06011295e-01 4.26087439e-01
2.26948202e-01 1.19291678e-01 -8.68089855e-01 -9.83007371e-01
-4.08784449e-01 6.55653000e-01 -9.10573959e-01 -3.07805479e-01
-7.33706713e-01 -6.61555529e-01 4.86152530e-01 6.09819651e-01
7.71628201e-01 -1.30647624e+00 -9.19483662e-01 2.04696819e-01
-3.82473588e-01 -1.53813577e+00 -3.31326544e-01 -4.94356006e-01
-9.81524050e-01 -1.31520689e+00 -8.74655843e-01 -6.83301866e-01
1.21272624e+00 -1.64301664e-01 1.16520405e+00 4.41785127e-01
-6.12194777e-01 5.29308617e-01 -5.68229318e-01 -2.54877418e-01
-7.89214849e-01 -8.28194439e-01 -2.14908168e-01 5.72915733e-01
-6.91870078e-02 -6.96810484e-01 -9.31450248e-01 2.02398196e-01
-1.41255414e+00 -1.33361474e-01 5.10975838e-01 1.04753995e+00
5.80543518e-01 2.32853383e-01 1.24244899e-01 -8.64433348e-01
3.24215889e-01 -2.00553581e-01 -4.32080209e-01 -2.60955319e-02
-8.31089690e-02 -3.59963477e-01 5.87774158e-01 -2.97735959e-01
-1.15359163e+00 8.66466761e-02 -2.42846310e-01 -8.76214206e-01
-7.49211490e-01 -1.70407727e-01 9.91755947e-02 -6.21227883e-02
4.83696193e-01 4.76748198e-01 -3.58237699e-02 -1.54810637e-01
1.24235816e-01 3.08874585e-02 1.06106663e+00 -2.16582283e-01
9.93965268e-01 9.84440982e-01 9.52946395e-02 -6.82826161e-01
-6.30425513e-01 -4.20158237e-01 -7.06153035e-01 -4.61648971e-01
1.26987636e+00 -6.61313355e-01 -8.31451863e-02 7.63580918e-01
-1.32960665e+00 -3.87384713e-01 -8.45464826e-01 2.59363413e-01
-8.12829852e-01 6.36495471e-01 -7.14386761e-01 -8.77457976e-01
-2.72331834e-01 -1.00675058e+00 1.22193205e+00 1.53575450e-01
8.81902874e-02 -1.30480266e+00 2.56100804e-01 2.42131516e-01
3.87446791e-01 1.16841638e+00 6.06104314e-01 -5.01526237e-01
-4.58534211e-01 -4.99553859e-01 6.52977377e-02 8.05121720e-01
2.64267564e-01 5.64296171e-03 -1.09700060e+00 -3.78735244e-01
2.04985306e-01 1.70676149e-02 7.83530474e-01 5.39559007e-01
1.28881967e+00 -2.95280904e-01 -4.77781706e-02 7.10638225e-01
1.29536223e+00 3.40226501e-01 1.26173556e+00 1.74904972e-01
8.16059411e-01 4.92292255e-01 4.41034585e-01 2.97082692e-01
-2.60418504e-01 5.52029848e-01 8.89080107e-01 -6.05391264e-01
-5.92826486e-01 -1.93951115e-01 5.95611453e-01 2.27826908e-01
-3.79687607e-01 -4.99394566e-01 -8.90075028e-01 7.31600642e-01
-1.79522133e+00 -1.09646928e+00 -2.81616062e-01 2.09902024e+00
1.11706026e-01 1.42627805e-01 4.48870808e-02 4.50684458e-01
7.73606777e-01 2.05565244e-01 -5.36658108e-01 -3.82275194e-01
-3.18735540e-01 3.23844641e-01 1.11400813e-01 4.20244455e-01
-1.23168802e+00 6.32936239e-01 6.55110502e+00 3.06361586e-01
-1.01614070e+00 3.55657727e-01 8.71884465e-01 2.94271596e-02
6.79869428e-02 -2.78827637e-01 2.08275124e-01 6.16567314e-01
5.36555946e-01 1.58492714e-01 1.95273712e-01 5.37656307e-01
3.75450134e-01 -3.41557741e-01 -9.65924680e-01 9.81782913e-01
5.35474598e-01 -9.53422308e-01 2.22337201e-01 -6.48184419e-02
1.20646942e+00 -1.36275813e-01 1.00728601e-01 -1.90492377e-01
1.16072781e-01 -1.02669597e+00 5.77216327e-01 8.63685250e-01
6.45787001e-01 -6.21230066e-01 1.25453269e+00 1.09688632e-01
-8.19374979e-01 5.56232221e-02 -3.72700691e-01 -8.40640441e-02
3.99831086e-01 6.40204489e-01 -7.16348529e-01 3.96864653e-01
7.17062354e-01 8.30856800e-01 -8.88659477e-01 1.02278316e+00
-4.15754616e-01 5.05197704e-01 -3.21266986e-02 6.91232204e-01
1.90210506e-01 -2.58949727e-01 8.85256290e-01 1.09083664e+00
3.08793187e-01 -2.04382613e-01 2.34525546e-01 9.31908786e-01
2.05470353e-01 -1.57713905e-01 -9.57854569e-01 2.72535026e-01
-5.37893116e-01 8.36116970e-01 -8.01070511e-01 -5.11897206e-01
-3.62518489e-01 1.62682033e+00 -2.17523426e-01 5.97335458e-01
-6.53926373e-01 -5.43321744e-02 6.65009856e-01 2.28917345e-01
1.82677925e-01 1.13047548e-02 -2.31214818e-02 -1.43676674e+00
1.72196344e-01 -6.90951824e-01 5.40743291e-01 -8.11371982e-01
-1.33409393e+00 6.61417305e-01 -2.91186899e-01 -1.46724069e+00
-7.04429984e-01 -4.58130866e-01 -1.24103475e+00 6.84237421e-01
-1.45814288e+00 -9.44680929e-01 -7.02982366e-01 1.08706212e+00
7.87370145e-01 -1.15603708e-01 5.88226318e-01 1.10097863e-01
-5.06131589e-01 2.62560964e-01 -1.43205851e-01 5.10444760e-01
4.14047122e-01 -1.34501886e+00 5.89016199e-01 1.70969892e+00
1.51541635e-01 -7.83165097e-02 7.71824539e-01 -7.99808979e-01
-7.10989892e-01 -1.45717454e+00 2.40120351e-01 -5.23655295e-01
1.25369772e-01 -4.88188583e-03 -9.87861872e-01 6.49462223e-01
3.19994152e-01 8.37770581e-01 2.10133076e-01 -8.74827147e-01
1.77756712e-01 2.88460672e-01 -1.62722600e+00 2.73942530e-01
9.08565044e-01 -1.96805209e-01 -6.37292385e-01 3.96595567e-01
2.83077508e-01 -7.36108840e-01 -3.98968130e-01 5.43663502e-01
8.61196890e-02 -1.46533692e+00 9.03401613e-01 -6.10302925e-01
5.36324680e-01 -4.34173822e-01 5.49476855e-02 -1.43771338e+00
2.04398200e-01 -3.40015918e-01 -4.19439733e-01 9.09696817e-01
-1.55240133e-01 -7.05831349e-01 8.51489186e-01 2.13007450e-01
-8.62618163e-02 -5.56180298e-01 -1.02431071e+00 -7.25215375e-01
-2.87596941e-01 -3.97485733e-01 2.60655344e-01 6.65373385e-01
-6.94993138e-01 -2.76915967e-01 -4.51321155e-01 1.95237279e-01
6.52486801e-01 -2.54791111e-01 8.15099299e-01 -7.96204567e-01
-3.58574688e-01 -2.07110241e-01 -1.29210484e+00 -3.07040513e-01
1.05076477e-01 -4.60954487e-01 -7.51235709e-03 -1.48664653e+00
-1.56619713e-01 2.88972676e-01 -2.84546345e-01 3.67556602e-01
-3.47313374e-01 8.37114334e-01 7.76175782e-02 1.17037178e-03
-4.47669357e-01 4.83811796e-01 1.46373844e+00 -2.57315695e-01
-1.84705392e-01 -1.13410363e-03 9.17524472e-02 1.03939867e+00
7.17620492e-01 -3.19104195e-01 -3.22979510e-01 -3.14419180e-01
-3.27968389e-01 -3.10361870e-02 8.10287714e-01 -1.49113047e+00
-1.77620232e-01 1.02499910e-02 9.94569004e-01 -3.94745529e-01
2.86834449e-01 -6.43395483e-01 -4.90659960e-02 9.45978820e-01
-2.08511442e-01 4.03801173e-01 1.40388504e-01 9.35145736e-01
-5.44680119e-01 -1.00234495e-02 1.06814408e+00 -2.11796254e-01
-7.54986882e-01 3.96953166e-01 -3.77990186e-01 1.72823310e-01
1.29023373e+00 -3.40163618e-01 -2.45831087e-01 -7.23797202e-01
-9.30145681e-01 -2.22795263e-01 7.54162431e-01 3.78428459e-01
1.10684443e+00 -1.13170850e+00 -8.53364706e-01 7.54497051e-01
9.58773866e-02 2.06649616e-01 3.87109131e-01 8.68410528e-01
-1.05090272e+00 -4.27340567e-02 -5.77326119e-01 -8.35009098e-01
-9.09692407e-01 6.45416856e-01 6.69058561e-01 -3.72731060e-01
-9.58941936e-01 5.49444914e-01 4.69593555e-01 1.38620093e-01
9.22675282e-02 -2.96555609e-01 -1.36057198e-01 -6.16988242e-01
7.06814170e-01 4.34780240e-01 1.59377381e-01 -8.87511551e-01
-1.19425043e-01 4.83495414e-01 3.47869009e-01 -1.28003955e-01
1.01092124e+00 -1.01939864e-01 -1.54916897e-01 7.83078894e-02
9.88522947e-01 -3.41765396e-02 -1.39521372e+00 7.28400797e-02
-4.44566041e-01 -1.10157120e+00 -6.70991081e-04 -5.35507083e-01
-1.69369924e+00 8.35700631e-01 1.18062079e+00 2.01967880e-01
1.50087881e+00 -1.73959315e-01 1.03040588e+00 -8.96343216e-02
1.49787236e-02 -6.88210785e-01 4.92867619e-01 3.19189541e-02
8.18094194e-01 -1.14476764e+00 -4.05458719e-01 -2.91776478e-01
-5.38394749e-01 1.05173266e+00 8.69548500e-01 -7.11043537e-01
2.77042270e-01 1.43008262e-01 3.06370586e-01 -1.04985245e-01
-5.03276944e-01 -4.88666594e-01 3.33611131e-01 1.18931353e+00
1.19594492e-01 -2.83417910e-01 -5.82234561e-02 -3.43600482e-01
1.54636294e-01 -2.35562369e-01 7.84797132e-01 8.15330803e-01
-9.23991948e-02 -1.00466716e+00 -8.47590804e-01 5.28896391e-01
-6.54492319e-01 4.06306058e-01 -3.26499522e-01 5.93677998e-01
5.31738043e-01 1.07455051e+00 3.22049081e-01 -1.60294905e-01
4.03885096e-01 2.29552388e-02 4.19237554e-01 -5.65653026e-01
-2.61426061e-01 1.13012781e-02 -2.79808104e-01 -1.10454726e+00
-6.89200938e-01 -7.25433350e-01 -1.01787412e+00 4.73011062e-02
1.31380573e-01 -2.80185640e-01 3.61480534e-01 8.44736159e-01
-3.14439535e-02 9.40028787e-01 6.03283226e-01 -1.00655770e+00
1.15741141e-01 -7.74092376e-01 -5.14424682e-01 1.15569687e+00
8.06453764e-01 -6.28574371e-01 -5.17998517e-01 4.56077576e-01] | [7.870909214019775, 1.5821926593780518] |
e255e841-ca3a-4047-8eb1-ea234277723f | robust-transformer-with-locality-inductive | 2301.11553 | null | https://arxiv.org/abs/2301.11553v1 | https://arxiv.org/pdf/2301.11553v1.pdf | Robust Transformer with Locality Inductive Bias and Feature Normalization | Vision transformers have been demonstrated to yield state-of-the-art results on a variety of computer vision tasks using attention-based networks. However, research works in transformers mostly do not investigate robustness/accuracy trade-off, and they still struggle to handle adversarial perturbations. In this paper, we explore the robustness of vision transformers against adversarial perturbations and try to enhance their robustness/accuracy trade-off in white box attack settings. To this end, we propose Locality iN Locality (LNL) transformer model. We prove that the locality introduction to LNL contributes to the robustness performance since it aggregates local information such as lines, edges, shapes, and even objects. In addition, to further improve the robustness performance, we encourage LNL to extract training signal from the moments (a.k.a., mean and standard deviation) and the normalized features. We validate the effectiveness and generality of LNL by achieving state-of-the-art results in terms of accuracy and robustness metrics on German Traffic Sign Recognition Benchmark (GTSRB) and Canadian Institute for Advanced Research (CIFAR-10). More specifically, for traffic sign classification, the proposed LNL yields gains of 1.1% and ~35% in terms of clean and robustness accuracy compared to the state-of-the-art studies. | ['Shahriar Baradaran Shokouhi', 'Hojat Asgarian Dehkordi', 'Hossein Kashiani', 'Omid Nejati Manzari'] | 2023-01-27 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 5.37329540e-02 -4.57942039e-01 6.13521188e-02 -9.21604857e-02
-5.71678817e-01 -5.89655161e-01 6.62363350e-01 -3.84425461e-01
-3.76845181e-01 4.74568516e-01 1.42116278e-01 -4.57672238e-01
-6.78777918e-02 -5.79485476e-01 -8.69644105e-01 -8.73688936e-01
1.32148966e-01 -4.07174945e-01 4.81221735e-01 -2.70409197e-01
2.52510548e-01 9.28305030e-01 -1.26434290e+00 1.09439969e-01
1.01841080e+00 1.18864322e+00 -4.97964352e-01 5.99230230e-01
2.26690009e-01 9.53618705e-01 -7.39631176e-01 -8.30551088e-01
5.78911722e-01 -1.48006052e-01 -2.55548805e-01 -1.40025660e-01
9.29731071e-01 -4.34313029e-01 -1.01708603e+00 1.48213613e+00
7.19285071e-01 3.67422663e-02 6.19919062e-01 -1.65744209e+00
-1.11152887e+00 1.37327507e-01 -6.83109224e-01 4.45719987e-01
-9.78554040e-02 8.33376765e-01 7.35071957e-01 -7.69069016e-01
3.39585096e-01 1.24246132e+00 6.97241127e-01 6.94093764e-01
-9.13911104e-01 -9.32552934e-01 2.93007851e-01 6.61576033e-01
-1.01757371e+00 -4.77397054e-01 8.14372480e-01 -1.98667660e-01
7.26760805e-01 2.07589820e-01 1.31448358e-01 1.45434904e+00
2.34752953e-01 8.44638050e-01 1.13340199e+00 -2.68731385e-01
5.82883582e-02 -8.18854496e-02 1.77904129e-01 6.49210751e-01
5.05203426e-01 4.53346521e-01 -3.51842254e-01 2.20909655e-01
6.11691594e-01 -1.27319857e-01 -4.50331241e-01 -2.66748935e-01
-9.66624260e-01 5.90134203e-01 8.54331553e-01 3.47498097e-02
-3.52862686e-01 3.83824825e-01 6.52433813e-01 2.17603192e-01
6.58552125e-02 2.82486677e-01 -8.83203968e-02 -2.84830481e-02
-5.59286177e-01 -7.78350979e-02 3.19958657e-01 8.90220821e-01
4.24368531e-02 5.96524477e-01 -6.18951678e-01 7.16554105e-01
2.90094852e-01 9.90968525e-01 4.30771232e-01 -6.48127139e-01
7.46565402e-01 3.47844720e-01 -4.63552326e-02 -1.13622320e+00
-1.81351885e-01 -6.15256369e-01 -1.09286392e+00 7.22596765e-01
5.66333354e-01 5.92719857e-03 -1.21172857e+00 1.78880692e+00
2.60427166e-02 4.37202185e-01 1.97719783e-01 9.79999006e-01
7.86324263e-01 4.32025820e-01 -4.52030003e-02 1.18679605e-01
1.15173972e+00 -1.10354257e+00 -5.40452898e-01 -1.60739034e-01
1.64368406e-01 -6.95602000e-01 1.20314991e+00 4.18486983e-01
-7.85009146e-01 -7.18751490e-01 -9.87719119e-01 1.59896046e-01
-4.50285316e-01 1.44648060e-01 2.59157002e-01 9.81223106e-01
-7.04050362e-01 4.74475712e-01 -7.23138154e-01 -3.52081865e-01
7.94456363e-01 2.29100809e-02 -4.47321802e-01 -1.25200242e-01
-1.00303519e+00 9.69113231e-01 -1.37530357e-01 2.16419294e-01
-9.73567426e-01 -7.91166782e-01 -6.18088365e-01 -4.91787083e-02
2.39704028e-01 -4.29392844e-01 9.73700464e-01 -6.00439310e-01
-1.57789898e+00 4.83972549e-01 8.18104818e-02 -7.73653388e-01
7.69945800e-01 -4.31426585e-01 -6.19192183e-01 2.42750019e-01
-1.45647049e-01 4.31377560e-01 1.06912494e+00 -1.06984973e+00
-5.74837565e-01 -2.74117559e-01 2.24040449e-01 -1.96368888e-01
-6.63728058e-01 2.05533028e-01 -4.47158933e-01 -1.07120335e+00
-2.28189945e-01 -9.04006064e-01 9.52207297e-02 3.78708512e-01
-3.88246357e-01 7.48790652e-02 1.23103607e+00 -7.39792705e-01
1.00091207e+00 -2.31880093e+00 -2.57641405e-01 1.72927529e-01
8.07794109e-02 1.00117922e+00 -3.43555570e-01 -8.85543763e-04
-9.63295773e-02 1.68409407e-01 -3.61245610e-02 -9.78302956e-02
2.27793947e-01 5.82515076e-02 -6.97130263e-01 7.57033765e-01
3.11376452e-01 1.16846406e+00 -6.29426241e-01 -1.28180116e-01
4.63566273e-01 5.19132316e-01 -3.67740810e-01 -1.26102164e-01
1.67643294e-01 3.18587631e-01 -3.16029280e-01 9.16393638e-01
7.93829858e-01 1.85447574e-01 -6.12459481e-01 -5.69023430e-01
1.22544192e-01 -1.93597794e-01 -7.32718229e-01 1.03663385e+00
-3.32612842e-01 1.08177483e+00 -2.14986935e-01 -8.38412583e-01
9.02950346e-01 1.94946185e-01 1.06930390e-01 -9.87979829e-01
2.26840541e-01 1.33106068e-01 2.69960731e-01 -4.76142049e-01
1.54896289e-01 1.25738829e-01 8.14467892e-02 3.27138640e-02
-1.13252990e-01 1.14342876e-01 1.02147171e-02 5.77584617e-02
1.14360607e+00 -2.07477957e-02 5.50856628e-02 7.01742694e-02
6.90377891e-01 -5.20695388e-01 4.74938989e-01 9.35767114e-01
-8.23788166e-01 4.26277786e-01 4.15417045e-01 -1.64730564e-01
-1.15161335e+00 -1.24701178e+00 -3.81302424e-02 8.46919358e-01
2.15431780e-01 1.15978174e-01 -7.74374843e-01 -8.64483774e-01
4.04454023e-02 8.20277750e-01 -5.67463934e-01 -6.39159083e-01
-6.64853454e-01 -5.60132861e-01 1.34883785e+00 9.70848143e-01
1.19662869e+00 -1.02081919e+00 -5.00522912e-01 -2.55200863e-01
2.12501567e-02 -1.55240941e+00 -6.27600372e-01 -3.17336440e-01
-6.26073599e-01 -9.72713828e-01 -7.78986931e-01 -4.68229026e-01
6.79019272e-01 2.63515949e-01 4.64996547e-01 -1.83103949e-01
-2.53638387e-01 2.57095516e-01 -5.10670185e-01 -4.59325463e-01
-2.18815282e-01 -4.14414167e-01 2.15999305e-01 3.31243008e-01
2.80655771e-02 -5.68962991e-01 -6.74724340e-01 6.26494765e-01
-9.49119091e-01 -4.96538490e-01 8.09637785e-01 9.08908010e-01
1.34179190e-01 -3.81995668e-03 4.91942227e-01 -2.07544163e-01
6.59083784e-01 4.71301712e-02 -6.53710663e-01 3.44332278e-01
-4.20439005e-01 4.46729623e-02 9.72186387e-01 -6.92067027e-01
-8.83851707e-01 -3.35233659e-01 -1.07940122e-01 -9.04332340e-01
-2.13879868e-01 5.85770234e-02 -3.77894193e-01 -6.64250076e-01
8.26246083e-01 3.54671091e-01 -1.31657630e-01 -1.78736627e-01
5.65499187e-01 6.47317290e-01 1.03799295e+00 -4.33585912e-01
1.20870590e+00 5.40348470e-01 1.99887142e-01 -7.45890677e-01
-6.44446075e-01 -2.29242265e-01 -2.85367578e-01 -3.39938343e-01
5.52476764e-01 -5.18066466e-01 -8.95833552e-01 1.01431537e+00
-9.28781569e-01 -1.22873187e-01 -2.48263493e-01 3.68745506e-01
-5.31756997e-01 7.41289258e-01 -4.88369882e-01 -7.64926434e-01
-5.75238943e-01 -1.23699176e+00 7.18081951e-01 4.60485518e-01
3.71479154e-01 -6.44586504e-01 -3.91997248e-01 3.42930913e-01
7.79686093e-01 4.50276971e-01 7.36416042e-01 -5.37562251e-01
-5.85233092e-01 -4.33416635e-01 -6.24166787e-01 7.27219403e-01
2.46631112e-02 6.47356138e-02 -1.15376222e+00 -4.06899780e-01
-1.65361121e-01 -1.86228231e-01 1.11420214e+00 2.39413634e-01
1.07925594e+00 -3.96564305e-01 -5.62348254e-02 9.10400510e-01
1.13993764e+00 2.46872261e-01 1.11465454e+00 4.98287022e-01
8.12115908e-01 1.48973569e-01 3.79355282e-01 1.79203898e-01
-7.51759708e-02 8.11280549e-01 6.69985652e-01 5.48298135e-02
-5.13645887e-01 -1.44613579e-01 7.05825269e-01 2.03132719e-01
-6.32987469e-02 -3.76248926e-01 -7.67163515e-01 5.44984102e-01
-1.79250014e+00 -1.10940218e+00 5.45599777e-03 2.18460941e+00
3.93714935e-01 4.54070538e-01 1.47617817e-01 3.80462945e-01
7.59281456e-01 2.52923250e-01 -8.55596662e-01 -1.85423508e-01
-3.59435558e-01 1.84357092e-01 8.18676353e-01 3.44942600e-01
-1.24524045e+00 1.16548693e+00 5.38419056e+00 1.01832736e+00
-1.43848681e+00 -5.94022162e-02 4.65377808e-01 -6.23294909e-04
2.08206981e-01 -3.88957381e-01 -6.81048095e-01 5.80562413e-01
5.67109406e-01 -8.32011625e-02 4.60147649e-01 6.51789963e-01
2.22247913e-01 3.83174956e-01 -6.68834090e-01 9.24526274e-01
2.24552527e-01 -1.03353167e+00 1.88782856e-01 2.38099862e-02
7.16178477e-01 9.22833160e-02 6.00416303e-01 3.12253982e-01
2.04932898e-01 -1.05654597e+00 7.67248690e-01 7.33557999e-01
8.59176815e-01 -7.53349304e-01 8.97056997e-01 9.33077559e-02
-1.12881279e+00 -4.12341863e-01 -2.78592706e-01 3.25267792e-01
1.31332010e-01 2.72704124e-01 -2.96254694e-01 5.79211652e-01
8.16502333e-01 5.53789854e-01 -7.64767110e-01 1.33560038e+00
-6.44725084e-01 7.54810691e-01 -3.12407225e-01 9.49772913e-03
3.98676455e-01 1.29969910e-01 8.31267715e-01 1.17386258e+00
1.36413261e-01 -1.74816921e-01 -1.54427722e-01 8.33975911e-01
-1.84408292e-01 -2.33111098e-01 -5.06407857e-01 7.32885525e-02
4.19436008e-01 9.53556538e-01 -4.49940711e-01 -1.30476698e-01
-2.15382949e-01 1.00340128e+00 -2.27427663e-04 6.17312908e-01
-1.22919917e+00 -7.95803726e-01 9.27229285e-01 -2.61869580e-01
6.71349704e-01 -2.09649101e-01 -3.57552737e-01 -1.09910023e+00
4.31845874e-01 -1.11178756e+00 1.12673163e-01 -7.94163764e-01
-1.58303523e+00 6.02693796e-01 -2.03156799e-01 -1.35120952e+00
1.54339522e-01 -9.48006749e-01 -8.40488076e-01 6.82056606e-01
-1.48600900e+00 -1.47372675e+00 -5.19740880e-01 7.19325483e-01
3.26967835e-01 -5.34334898e-01 3.27553153e-01 3.41583550e-01
-7.66022086e-01 1.40070152e+00 2.31088474e-01 6.34228528e-01
6.68061554e-01 -8.63818109e-01 8.41465890e-01 1.45351088e+00
1.78021669e-01 4.92976278e-01 5.46850681e-01 -4.05247569e-01
-1.33295703e+00 -1.27143204e+00 3.12548161e-01 -3.44523191e-01
8.88146698e-01 3.35591584e-02 -7.22302437e-01 4.30219114e-01
-4.88464534e-03 5.92382669e-01 6.22955635e-02 -5.17298639e-01
-9.93665516e-01 -4.11320567e-01 -1.35032117e+00 9.55420554e-01
1.06166017e+00 -6.68025196e-01 -4.40608412e-01 -1.09060369e-01
5.84276319e-01 -3.40402335e-01 -4.88274515e-01 5.80895662e-01
6.29961133e-01 -9.71266508e-01 1.24090564e+00 -7.11316347e-01
2.66976267e-01 -6.18712127e-01 -4.30077940e-01 -1.28433907e+00
-3.49089950e-01 -6.44203126e-01 -2.92532176e-01 1.33459175e+00
1.13306023e-01 -8.87567222e-01 5.53198934e-01 3.34997356e-01
-2.02097461e-01 -5.92211664e-01 -1.14863861e+00 -1.18357754e+00
1.90819278e-01 -6.69245601e-01 4.67913896e-01 5.38917780e-01
-3.97190064e-01 -1.26009122e-01 -5.89975536e-01 5.02054393e-01
8.20458591e-01 -2.41936639e-01 9.29802597e-01 -6.84778810e-01
-9.40930098e-02 -8.32359195e-01 -1.05585194e+00 -9.60521221e-01
1.43263459e-01 -5.64924121e-01 -1.13345914e-01 -1.29722023e+00
-1.16797030e-01 2.87677813e-02 -5.52405179e-01 6.55474842e-01
-3.05047423e-01 5.04531264e-01 7.61467874e-01 -1.35733373e-02
-4.89156097e-01 6.51509464e-01 9.97702599e-01 -5.15440404e-01
1.44072175e-01 4.90943268e-02 -5.97787797e-01 6.96449101e-01
9.56113338e-01 -2.20415294e-01 -1.87605903e-01 -3.67198795e-01
-5.79563200e-01 -5.06806374e-01 8.78204823e-01 -1.28108835e+00
4.18733150e-01 -1.02810986e-01 3.05631369e-01 -3.54327649e-01
3.25688750e-01 -7.91987836e-01 -4.47301507e-01 5.42228103e-01
-1.75836682e-01 -2.44639993e-01 5.25287807e-01 6.49875820e-01
-1.53882101e-01 1.36705488e-02 1.14034319e+00 4.03448999e-01
-7.54646301e-01 3.26361954e-01 -1.49549529e-01 7.45476112e-02
1.03644490e+00 -3.40674192e-01 -8.71510744e-01 -4.52172399e-01
-2.11959943e-01 -1.76519882e-02 2.38951176e-01 5.79156339e-01
8.04529190e-01 -1.46999180e+00 -7.33959377e-01 3.62578273e-01
2.20940948e-01 -6.09263539e-01 1.82367533e-01 9.18842971e-01
-5.04820526e-01 4.33952987e-01 -5.35560250e-01 -4.88102704e-01
-1.35650039e+00 7.36598790e-01 5.27540922e-01 -1.15983181e-01
-7.15453744e-01 7.54154742e-01 1.52402177e-01 -9.22592953e-02
5.87240577e-01 -4.69687521e-01 5.18667251e-02 -3.89230698e-01
5.12721300e-01 6.16875052e-01 -5.32644652e-02 -7.00045228e-01
-4.90516573e-01 8.27660143e-01 -7.14030713e-02 -6.27487302e-02
1.02300036e+00 2.34514743e-01 3.29338431e-01 -2.28059635e-01
1.04189599e+00 1.05821595e-01 -1.44655740e+00 -2.53311545e-01
-2.32422322e-01 -7.43630826e-01 5.45258597e-02 -9.13811386e-01
-1.44985545e+00 1.02959108e+00 8.14960718e-01 -4.85482290e-02
1.24538350e+00 -3.99673969e-01 9.05016243e-01 4.41782713e-01
1.11891225e-01 -7.46377826e-01 2.69445419e-01 5.86525798e-01
1.05846751e+00 -1.06285655e+00 -1.67130172e-01 -1.78652540e-01
-6.62753701e-01 9.45225835e-01 6.51758850e-01 -2.60846138e-01
4.98741806e-01 2.87765354e-01 2.54729927e-01 1.79575086e-01
-3.75434041e-01 -2.98209816e-01 6.00343049e-01 9.58099365e-01
-7.63852149e-02 -1.70587286e-01 6.74565360e-02 3.08939457e-01
-1.55957386e-01 -1.25371471e-01 2.36569762e-01 5.82691073e-01
-3.00926894e-01 -7.61312306e-01 -6.80991709e-01 1.94482744e-01
-4.39147413e-01 -1.61794171e-01 -4.84329581e-01 7.26086020e-01
2.43872777e-02 1.11907113e+00 -3.52144390e-01 -5.74681938e-01
6.79661274e-01 -2.21576393e-01 4.23548341e-01 2.96549022e-01
-4.48227733e-01 -2.34227732e-01 -1.27361551e-01 -6.31930888e-01
-1.24614447e-01 -3.43051493e-01 -8.33873630e-01 -4.41701472e-01
-4.11849707e-01 -3.40697944e-01 4.79686618e-01 8.32784593e-01
3.36279839e-01 6.89730644e-01 6.23409450e-01 -6.83629870e-01
-9.33572829e-01 -9.26852643e-01 -1.84170067e-01 6.39319003e-01
4.64489013e-01 -7.23135114e-01 -4.48932886e-01 -1.22825935e-01] | [5.4156928062438965, 7.915650844573975] |
7f9bcac8-cdbe-4f85-809e-c0dc61f5085f | linguistic-information-in-neural-semantic | null | null | https://aclanthology.org/W19-0504 | https://aclanthology.org/W19-0504.pdf | Linguistic Information in Neural Semantic Parsing with Multiple Encoders | Recently, sequence-to-sequence models have achieved impressive performance on a number of semantic parsing tasks. However, they often do not exploit available linguistic resources, while these, when employed correctly, are likely to increase performance even further. Research in neural machine translation has shown that employing this information has a lot of potential, especially when using a multi-encoder setup. We employ a range of semantic and syntactic resources to improve performance for the task of Discourse Representation Structure Parsing. We show that (i) linguistic features can be beneficial for neural semantic parsing and (ii) the best method of adding these features is by using multiple encoders. | ['Rik van Noord', 'Antonio Toral', 'Johan Bos'] | 2019-05-01 | null | null | null | ws-2019-5 | ['drs-parsing'] | ['natural-language-processing'] | [ 4.92980808e-01 3.54275674e-01 -2.58920550e-01 -5.41751862e-01
-1.07114899e+00 -5.81721187e-01 7.28104413e-01 2.12858200e-01
-5.20783305e-01 8.73407960e-01 7.21368074e-01 -6.25708818e-01
2.92753875e-01 -7.93465972e-01 -8.08808804e-01 -1.51728109e-01
2.93669283e-01 4.80526268e-01 2.56752312e-01 -4.52320337e-01
2.93993890e-01 5.80297895e-02 -1.40234470e+00 5.16575336e-01
7.22427964e-01 5.19465327e-01 4.29211676e-01 5.52442312e-01
-5.84956408e-01 1.10281086e+00 -6.02415025e-01 -6.84328198e-01
1.67587437e-02 -7.18129635e-01 -1.35409284e+00 -2.20115870e-01
1.43206701e-01 -2.66053915e-01 3.63447294e-02 8.17063570e-01
3.29820335e-01 1.09771073e-01 3.52969170e-01 -5.96817732e-01
-5.71976900e-01 6.55322433e-01 -2.25641817e-01 3.21431339e-01
4.14783746e-01 -2.52814479e-02 1.33398330e+00 -6.06117308e-01
8.89299452e-01 1.52556610e+00 6.19010806e-01 6.72051549e-01
-1.05298185e+00 -3.74612808e-01 -9.69983917e-03 -2.18425896e-02
-6.80118799e-01 -6.75953686e-01 6.39716923e-01 -1.01318672e-01
1.55988765e+00 -8.08787048e-02 1.95933074e-01 1.00335741e+00
-2.33480707e-02 9.25248265e-01 1.19062448e+00 -7.93963194e-01
-1.44920677e-01 6.34752885e-02 8.59149694e-02 8.13584030e-01
-1.67289376e-02 5.63945845e-02 -5.20057797e-01 -9.90392566e-02
5.50580204e-01 -6.49476528e-01 -9.51558724e-02 8.62998962e-02
-9.79085803e-01 1.18141246e+00 2.61339277e-01 6.60845339e-01
-8.09428841e-02 3.39300126e-01 8.37982953e-01 5.75965941e-01
6.88281119e-01 9.87681389e-01 -6.26814723e-01 -4.57440645e-01
-6.73491776e-01 1.38234586e-01 8.76720846e-01 6.01469755e-01
6.99032068e-01 -7.93108195e-02 1.54190361e-01 1.14223790e+00
3.43288183e-02 8.40277672e-02 6.09171152e-01 -1.25170565e+00
6.85465693e-01 3.73418331e-01 -4.71030138e-02 -8.55879068e-01
-4.66011226e-01 3.83446328e-02 -2.45107293e-01 -8.34222883e-05
6.48212552e-01 -2.35306337e-01 -7.03925967e-01 2.05135727e+00
-3.74010950e-02 -8.55554640e-02 3.25998306e-01 6.74992323e-01
5.00451446e-01 6.33611560e-01 4.57122296e-01 -8.25887248e-02
1.30817544e+00 -9.23769116e-01 -6.65405750e-01 -7.62234807e-01
1.14675736e+00 -8.95571053e-01 9.62784171e-01 -2.26570033e-02
-1.42255592e+00 -3.13186198e-01 -9.58611846e-01 -2.40139246e-01
-3.75863045e-01 -4.59835865e-02 9.54808593e-01 6.07639492e-01
-1.11859059e+00 8.16673338e-01 -9.09315944e-01 -4.91497129e-01
2.94999003e-01 2.02791303e-01 -3.94060940e-01 -2.63767004e-01
-1.32046878e+00 1.46048391e+00 5.05356371e-01 -2.21392363e-01
-2.92733550e-01 -2.41606414e-01 -1.15422380e+00 1.39402032e-01
4.01780158e-01 -8.59469175e-01 1.31043482e+00 -1.38821185e+00
-1.39434433e+00 9.25911486e-01 -4.67902243e-01 -5.55243015e-01
1.30610406e-01 -2.65921414e-01 6.54649884e-02 4.35990930e-01
2.13797167e-01 1.05012345e+00 4.92853522e-01 -8.67555857e-01
-5.96674442e-01 -4.42669094e-01 3.64010811e-01 4.27893162e-01
-4.66013849e-01 5.95858097e-01 -1.07161991e-01 -5.17892480e-01
-1.29590809e-01 -8.67406428e-01 -2.53240854e-01 -3.39875430e-01
-1.60103887e-02 -4.20322508e-01 3.90972525e-01 -9.80560780e-01
8.56386304e-01 -1.79987562e+00 2.37165898e-01 -2.46661469e-01
-3.24812293e-01 4.10707891e-01 -1.86927244e-01 4.71040398e-01
1.78715140e-01 4.62264180e-01 -3.04388642e-01 -3.82307321e-01
-2.24858195e-01 4.96826768e-01 -1.85472906e-01 -1.35833323e-01
6.45053923e-01 1.10513163e+00 -7.22057819e-01 -3.75544310e-01
1.55096084e-01 3.57461721e-01 -6.54424489e-01 -3.98038253e-02
-4.54329848e-01 4.22545195e-01 -4.98923361e-01 1.38624758e-01
-7.81035284e-03 -3.34424078e-01 4.59849238e-01 3.29010189e-01
1.09931514e-01 1.08567643e+00 -6.64941728e-01 1.78090930e+00
-7.86661327e-01 6.07765198e-01 1.35810256e-01 -1.17106891e+00
6.34983003e-01 4.70575482e-01 1.45225793e-01 -7.35815346e-01
9.52221006e-02 3.47425342e-01 2.94509888e-01 -4.96781051e-01
5.22323728e-01 -4.20362949e-01 -1.58345513e-02 6.58836007e-01
1.73375726e-01 6.94422424e-02 2.98336625e-01 1.04081064e-01
1.38044047e+00 4.44930077e-01 2.94393837e-01 -2.15688914e-01
3.29015374e-01 3.02686036e-01 5.59354603e-01 5.50792336e-01
1.13356616e-02 4.49801028e-01 6.28662825e-01 -2.73648947e-01
-1.31465447e+00 -4.25113440e-01 7.06328377e-02 1.35188150e+00
-2.24898919e-01 -2.66914815e-01 -8.72620225e-01 -6.21009767e-01
-3.22545826e-01 8.94305646e-01 -3.26208651e-01 -9.85449553e-02
-1.05550671e+00 -7.75964141e-01 7.45202601e-01 9.38445330e-01
2.44784474e-01 -1.13981295e+00 -6.87508106e-01 5.47849774e-01
-1.61180347e-01 -1.42104745e+00 -6.52143778e-03 4.08369541e-01
-9.93776679e-01 -9.06405866e-01 -5.12861371e-01 -7.83616066e-01
3.68585110e-01 3.15106630e-01 1.17132211e+00 3.73545289e-01
1.31264746e-01 2.89500147e-01 -4.77076769e-01 -3.85305285e-01
-9.37535167e-01 3.26258302e-01 -4.52995688e-01 -5.31885564e-01
4.54871118e-01 -4.10513818e-01 4.49873880e-02 -4.09590006e-02
-6.15223408e-01 1.98828056e-01 5.27904749e-01 1.07807660e+00
-1.39169125e-02 -4.18270707e-01 1.05787790e+00 -1.25107527e+00
8.00312281e-01 -3.82071674e-01 -2.44952217e-01 1.82385296e-01
-4.82922524e-01 3.68868470e-01 7.16951966e-01 2.96064466e-02
-1.22803330e+00 -1.64678067e-01 -5.77797413e-01 -7.24880993e-02
-3.90117526e-01 6.22264862e-01 -5.50441369e-02 1.45786986e-01
5.27688980e-01 -1.74011916e-01 2.87254095e-01 -6.39295101e-01
4.04141724e-01 5.35110891e-01 1.37952313e-01 -7.73562193e-01
2.89758235e-01 1.66598454e-01 -1.86690744e-02 -6.98569596e-01
-1.09498525e+00 -2.89261878e-01 -5.39393723e-01 3.95850062e-01
1.09331727e+00 -8.62551689e-01 -1.35072172e-01 6.72859848e-02
-1.39384556e+00 -3.42158616e-01 6.39833063e-02 3.90086770e-01
-6.39173746e-01 3.85016918e-01 -8.55408192e-01 -5.15045762e-01
-6.07687980e-02 -1.30543566e+00 1.08461332e+00 3.51972468e-02
-4.34923828e-01 -1.33176923e+00 -2.16062054e-01 5.73572040e-01
5.23396254e-01 -3.09628956e-02 1.35230851e+00 -9.55304027e-01
-4.68144417e-01 1.13065265e-01 -1.99388415e-01 4.29897636e-01
9.77506489e-02 -4.29497540e-01 -9.53722537e-01 2.49544382e-02
5.65064028e-02 -5.43688357e-01 9.63262558e-01 2.44525224e-01
8.06217372e-01 -2.42546022e-01 -3.41806322e-01 1.79153442e-01
1.31030965e+00 1.26069903e-01 4.81743455e-01 4.45451647e-01
6.26628518e-01 7.15784788e-01 3.22041154e-01 -2.78619587e-01
6.48498058e-01 7.78481245e-01 7.30265528e-02 1.62289329e-02
-3.42898250e-01 -3.07683229e-01 3.50184739e-01 8.55707705e-01
8.57708454e-02 -1.49834007e-01 -9.79386747e-01 6.78264618e-01
-1.74534297e+00 -8.55586052e-01 1.37410671e-01 1.77364969e+00
9.10390496e-01 2.84426033e-01 1.02054477e-01 -5.57466820e-02
5.01466274e-01 1.08262643e-01 -2.93554723e-01 -9.07345057e-01
-2.53347814e-01 5.52505076e-01 4.90707338e-01 6.30778313e-01
-8.38909268e-01 1.15653050e+00 7.23169422e+00 4.15176928e-01
-8.91082764e-01 3.49121541e-01 6.18653178e-01 3.35299522e-01
-4.40923572e-01 2.71120965e-01 -8.09734106e-01 3.38534534e-01
1.35745347e+00 2.38781199e-02 2.91434735e-01 6.48998976e-01
-1.88603446e-01 -2.64550596e-01 -1.09133506e+00 2.78084695e-01
2.27530837e-01 -1.49805653e+00 3.71402130e-03 -1.33027256e-01
5.21697283e-01 1.08272120e-01 -4.95458663e-01 3.33946884e-01
5.14354289e-01 -1.15240109e+00 3.71882439e-01 -1.10385880e-01
4.58291769e-01 -7.39441156e-01 8.02313685e-01 5.18826008e-01
-7.22191155e-01 -1.01226337e-01 -4.21981871e-01 -3.42047632e-01
4.45449322e-01 8.69493783e-02 -7.65703619e-01 5.35017133e-01
2.41745934e-01 6.20627940e-01 -5.73024750e-01 5.91944814e-01
-5.34205914e-01 7.34082997e-01 -3.04466546e-01 -1.61025852e-01
4.77368772e-01 5.79345003e-02 3.36289972e-01 1.31225634e+00
1.14708520e-01 7.13827312e-02 3.18872035e-02 5.12201786e-01
-1.08588636e-01 2.37452984e-01 -8.72954607e-01 -2.02009678e-01
2.95794249e-01 7.80889988e-01 -6.74042702e-01 -3.88307512e-01
-9.59809005e-01 8.14819098e-01 6.63355172e-01 -7.07519799e-02
-4.17184621e-01 -8.00341740e-02 6.70147955e-01 -1.30826026e-01
3.35143626e-01 -3.76430988e-01 -4.72876728e-01 -1.11852193e+00
-1.48522019e-01 -9.32074070e-01 3.33297729e-01 -7.97041595e-01
-1.03433585e+00 4.15835440e-01 -5.45740798e-02 -4.03612643e-01
-6.96656942e-01 -8.84256542e-01 -3.80528599e-01 8.29766214e-01
-1.65154648e+00 -1.27943373e+00 4.68633175e-01 9.77798253e-02
8.34816456e-01 -8.33987892e-02 1.20452297e+00 2.57342935e-01
-2.57011056e-01 3.39972138e-01 -1.76134735e-01 3.09936941e-01
5.90652823e-01 -1.19268894e+00 6.56054318e-01 7.45627284e-01
4.26435977e-01 4.73613560e-01 6.63537383e-01 -4.34407830e-01
-1.26083040e+00 -7.66830921e-01 1.16536951e+00 -6.17185593e-01
5.76216459e-01 -3.02894890e-01 -1.14074206e+00 9.66617107e-01
4.10599589e-01 -5.47435760e-01 6.16756022e-01 5.32502651e-01
-3.48489672e-01 5.97674370e-01 -8.48596931e-01 4.99325663e-01
9.00381684e-01 -6.73986852e-01 -1.10396767e+00 3.87610316e-01
9.41695690e-01 -3.95289838e-01 -8.45749259e-01 3.68791580e-01
2.50196338e-01 -7.14305878e-01 8.86585116e-01 -8.63961935e-01
8.89532745e-01 1.27633691e-01 -1.50345236e-01 -1.43330741e+00
-8.19750503e-02 -4.36609179e-01 5.44622913e-02 1.29397225e+00
7.05778956e-01 -6.25792861e-01 5.22198498e-01 7.34670162e-01
-3.30402732e-01 -7.89741516e-01 -9.53985810e-01 -7.45532870e-01
6.45278931e-01 -3.12084556e-01 4.57991868e-01 9.68989491e-01
1.96641654e-01 1.02068806e+00 -3.37015539e-01 -4.07486349e-01
9.01621282e-02 -4.72041294e-02 4.48001266e-01 -1.07433355e+00
-3.84451479e-01 -6.00260794e-01 -2.94514775e-01 -9.48529541e-01
7.49949217e-01 -1.14687979e+00 -6.88050091e-02 -1.68782759e+00
8.23751464e-02 -4.49858069e-01 -7.46096224e-02 6.28585696e-01
-4.49709713e-01 5.11007980e-02 3.70324552e-01 8.03234577e-02
-3.20041895e-01 1.71975702e-01 1.05545712e+00 3.02096188e-01
1.09428272e-01 -3.00596535e-01 -9.45905268e-01 8.09084296e-01
9.77051258e-01 -4.21686411e-01 -1.28435448e-01 -8.91016603e-01
2.33900100e-01 3.53235245e-01 6.24510087e-02 -6.38910651e-01
-8.75768960e-02 -4.88019548e-04 1.03137121e-01 1.07027516e-01
6.43891692e-01 -3.28883320e-01 -2.98534691e-01 3.34201932e-01
-6.48035347e-01 2.25031793e-01 4.06038284e-01 3.91236871e-01
-3.51866424e-01 -7.31705666e-01 6.76908851e-01 -6.59766674e-01
-7.11911559e-01 -3.95696104e-01 -4.60701913e-01 2.64345646e-01
6.18571341e-01 4.46603298e-02 -3.63745987e-01 -3.55420291e-01
-4.81664062e-01 1.91277824e-02 6.16413176e-01 5.22316813e-01
1.19974926e-01 -1.00092959e+00 -5.66706240e-01 -4.94779721e-02
-2.13512108e-01 -2.83020258e-01 -3.61067563e-01 5.74275494e-01
-2.11561278e-01 6.65198028e-01 -1.94614485e-01 -3.56105864e-01
-1.22966814e+00 3.34993273e-01 1.62408113e-01 -5.09781122e-01
-7.73053288e-01 7.72039413e-01 -1.89886950e-02 -3.15700978e-01
-1.27314717e-01 6.56619519e-02 -3.07725906e-01 1.64528102e-01
4.98339266e-01 -1.60300825e-02 8.66150707e-02 -6.93758249e-01
-2.53994316e-01 4.48900819e-01 -2.87370801e-01 -2.62538612e-01
1.60641789e+00 -1.26140475e-01 -4.39292528e-02 3.47052634e-01
1.11286712e+00 -2.26244807e-01 -8.92627120e-01 -2.81359136e-01
5.22967160e-01 -2.37650916e-01 -4.14259098e-02 -7.92729318e-01
-5.84676027e-01 1.24202943e+00 -1.25842229e-01 1.71373829e-01
1.02165008e+00 8.21156800e-02 1.08534336e+00 3.35115969e-01
5.21814823e-01 -1.12793899e+00 -8.68893713e-02 8.36159647e-01
3.55744243e-01 -1.28431177e+00 -1.32414818e-01 -3.95551652e-01
-8.32336783e-01 1.16825616e+00 3.49997878e-01 -1.51095122e-01
2.75072873e-01 4.22194004e-01 -4.74810712e-02 -1.76315576e-01
-1.08474255e+00 -4.45146501e-01 -1.36134177e-01 4.38609064e-01
9.14720774e-01 -1.10180646e-01 -5.36225617e-01 2.76349217e-01
-1.19158752e-01 -2.97264867e-02 5.23314297e-01 9.88029718e-01
-6.59529805e-01 -1.60519159e+00 -1.03886321e-01 5.16133070e-01
-1.03663659e+00 -1.89562410e-01 -3.90905678e-01 6.61694527e-01
-3.74500871e-01 1.06242096e+00 -6.39375225e-02 1.39995754e-01
5.37857413e-04 5.92490435e-01 6.88428760e-01 -9.74415660e-01
-7.30915785e-01 -2.67317370e-02 8.18635821e-01 -5.86746931e-01
-6.46562874e-01 -8.00780714e-01 -1.29506028e+00 -3.24876383e-02
-2.84424186e-01 3.45590971e-02 9.18368459e-01 1.38100100e+00
3.54672551e-01 5.63921988e-01 1.96036562e-01 -5.24933755e-01
-6.38055801e-01 -9.85458016e-01 1.66246787e-01 3.27695400e-01
4.01376523e-02 -5.75645804e-01 -9.41031575e-02 2.10284442e-02] | [10.631903648376465, 9.307860374450684] |
85acee1a-c489-48e7-9e8e-3524007f7618 | safe-exploration-in-markov-decision-processes-1 | 1205.4810 | null | https://arxiv.org/abs/1205.4810v3 | https://arxiv.org/pdf/1205.4810v3.pdf | Safe Exploration in Markov Decision Processes | In environments with uncertain dynamics exploration is necessary to learn how to perform well. Existing reinforcement learning algorithms provide strong exploration guarantees, but they tend to rely on an ergodicity assumption. The essence of ergodicity is that any state is eventually reachable from any other state by following a suitable policy. This assumption allows for exploration algorithms that operate by simply favoring states that have rarely been visited before. For most physical systems this assumption is impractical as the systems would break before any reasonable exploration has taken place, i.e., most physical systems don't satisfy the ergodicity assumption. In this paper we address the need for safe exploration methods in Markov decision processes. We first propose a general formulation of safety through ergodicity. We show that imposing safety by restricting attention to the resulting set of guaranteed safe policies is NP-hard. We then present an efficient algorithm for guaranteed safe, but potentially suboptimal, exploration. At the core is an optimization formulation in which the constraints restrict attention to a subset of the guaranteed safe policies and the objective favors exploration policies. Our framework is compatible with the majority of previously proposed exploration methods, which rely on an exploration bonus. Our experiments, which include a Martian terrain exploration problem, show that our method is able to explore better than classical exploration methods. | ['Pieter Abbeel', 'Teodor Mihai Moldovan'] | 2012-05-22 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-5.80218248e-02 3.46838325e-01 -4.85531509e-01 2.11284369e-01
-4.58545625e-01 -7.09388733e-01 6.85277283e-01 1.09741390e-02
-5.48764765e-01 1.40600085e+00 -1.34811178e-01 -7.22598195e-01
-5.39259970e-01 -1.13320374e+00 -7.70173371e-01 -1.07462037e+00
-6.38813734e-01 6.28777146e-01 3.39327335e-01 -4.40129429e-01
3.53766203e-01 5.66979051e-01 -1.51098323e+00 -5.64033747e-01
8.79612923e-01 6.75533056e-01 1.32120624e-01 5.95638633e-01
7.89846405e-02 5.75319111e-01 -2.84951419e-01 4.31493193e-01
3.33228111e-01 -4.01900768e-01 -1.19428921e+00 -9.11239684e-02
-4.76687908e-01 -3.93966556e-01 -2.14824468e-01 1.22164083e+00
7.58014340e-03 6.76297188e-01 4.79078412e-01 -1.52111173e+00
5.55497743e-02 9.07217503e-01 -2.95094728e-01 5.56284189e-02
1.47427350e-01 2.79047847e-01 8.68870199e-01 7.76460171e-02
6.74487114e-01 1.15792513e+00 1.13422945e-01 6.38133645e-01
-1.29205143e+00 -5.06855845e-01 6.87319458e-01 1.64451580e-02
-1.17765975e+00 -2.25166440e-01 4.05262023e-01 -1.01965733e-01
1.04852593e+00 5.31056821e-01 8.99794757e-01 1.05501556e+00
6.84439659e-01 6.39327645e-01 1.45819032e+00 -3.17253947e-01
1.01942873e+00 -1.12916604e-01 -9.50368643e-02 4.38106567e-01
5.49259067e-01 7.96164989e-01 -3.21311355e-01 -2.83971816e-01
6.76720500e-01 -3.43063831e-01 -2.22777292e-01 -9.47808206e-01
-1.04107785e+00 9.18274581e-01 9.37278271e-02 8.66639763e-02
-2.10832849e-01 4.94131088e-01 2.96298802e-01 5.33438623e-01
-5.33956140e-02 8.22029710e-01 -2.08456099e-01 -4.23606038e-01
-7.41303682e-01 5.55915713e-01 1.08799160e+00 6.95930064e-01
5.53609490e-01 1.85751438e-01 1.98013142e-01 -1.27876878e-01
2.86108732e-01 4.07200158e-01 8.28859359e-02 -1.17689323e+00
2.02752396e-01 1.20631587e-02 7.75766253e-01 -3.46843421e-01
-2.98069000e-01 -3.86676967e-01 -4.45298254e-01 9.22505319e-01
3.66774857e-01 -2.71008670e-01 -8.39791417e-01 2.06059742e+00
3.96707833e-01 1.52928643e-02 3.93038273e-01 6.57899380e-01
-4.29612279e-01 6.95061207e-01 -1.05649233e-01 -6.17257297e-01
7.02786922e-01 -5.06762505e-01 -7.38971353e-01 -3.34292315e-02
5.47469378e-01 -1.20921560e-01 1.07424855e+00 6.21081352e-01
-1.15930212e+00 3.63284536e-02 -1.34633851e+00 7.96619773e-01
-2.75036782e-01 -7.64666200e-01 7.82449007e-01 8.79892051e-01
-1.03764391e+00 1.00667036e+00 -1.38943768e+00 -3.37123215e-01
2.12607756e-02 4.26002681e-01 3.23126651e-02 5.17279863e-01
-1.21768236e+00 1.29601657e+00 8.06344390e-01 -3.15853059e-02
-1.43044412e+00 -5.30107021e-02 -7.38232017e-01 7.70172924e-02
9.86809909e-01 -3.15144598e-01 1.48191190e+00 -5.19069016e-01
-1.69029021e+00 -3.86616588e-02 8.92710090e-02 -8.16649199e-01
8.57200623e-01 -3.02634478e-01 -1.90622732e-01 6.92745019e-03
-1.64775699e-02 3.12033266e-01 4.66532230e-01 -1.38253582e+00
-6.93386078e-01 5.24483621e-02 4.68361378e-01 3.77571642e-01
-2.25828052e-01 -4.21691597e-01 -2.42889225e-02 -9.01633576e-02
1.09703600e-01 -1.32081723e+00 -8.76276672e-01 -3.46131444e-01
-5.66636741e-01 -5.59022725e-02 6.45508885e-01 6.74239323e-02
1.26816249e+00 -1.66657639e+00 2.21614942e-01 7.35847950e-01
-1.55285358e-01 -2.26309419e-01 1.55820131e-01 7.76867032e-01
2.19643444e-01 3.41100395e-01 -3.88764352e-01 -2.45016944e-02
4.49575335e-01 6.06872618e-01 -8.50934088e-01 6.18459582e-01
-3.29637170e-01 5.66656411e-01 -1.13066518e+00 -3.36544812e-01
3.36554229e-01 -1.27173275e-01 -6.03941917e-01 -4.85592484e-02
-5.91795385e-01 4.51706648e-01 -7.06170142e-01 4.35764611e-01
4.24113870e-01 7.72866458e-02 4.46572602e-01 7.31762290e-01
-6.69572055e-01 4.08657998e-01 -1.56859231e+00 1.36409903e+00
-2.77695984e-01 1.80505902e-01 1.49664015e-01 -9.93326664e-01
5.55184662e-01 1.90332115e-01 6.87896013e-01 -4.45851266e-01
1.15189508e-01 3.37257385e-01 1.26484688e-03 -1.48203105e-01
6.62929177e-01 -3.07370305e-01 -3.42283934e-01 7.92346716e-01
-4.60081995e-01 -3.35228086e-01 1.41544268e-01 1.24511234e-01
1.18537414e+00 5.54977536e-01 4.23865139e-01 -9.36418474e-01
2.15039417e-01 1.93376452e-01 7.74650812e-01 1.25638831e+00
-3.33063960e-01 -2.77153075e-01 8.53834331e-01 -3.44025880e-01
-9.10889983e-01 -1.21898246e+00 -1.94681346e-01 7.50987709e-01
7.03243196e-01 -3.68002206e-01 -7.00216889e-01 -6.69102430e-01
-9.19850841e-02 1.09108186e+00 -1.05585527e+00 -1.32378504e-01
-5.02552927e-01 -4.49602216e-01 3.01765978e-01 3.82542133e-01
4.99436617e-01 -1.14373744e+00 -1.42740035e+00 3.67252767e-01
1.16165079e-01 -5.40481806e-01 2.79059298e-02 7.31672883e-01
-9.30271506e-01 -9.63248730e-01 -1.46187246e-01 -1.07461803e-01
4.41538393e-01 -1.27056241e-01 7.38839984e-01 1.49955507e-02
1.51086688e-01 3.25946152e-01 -1.92432389e-01 -3.90321910e-01
-3.75702024e-01 2.03574389e-01 4.95259851e-01 -5.60704529e-01
-6.13796227e-02 -6.57135844e-01 -3.12869430e-01 2.66582012e-01
-9.28895950e-01 2.06523184e-02 3.28495651e-01 7.26147652e-01
7.37804055e-01 5.97940147e-01 4.65528399e-01 -5.17798662e-01
4.39908564e-01 -5.21977961e-01 -9.64030027e-01 2.24518865e-01
-8.16815138e-01 7.25658774e-01 4.99681532e-01 -3.88533145e-01
-1.03981280e+00 3.93989049e-02 6.06322922e-02 -1.48976773e-01
-8.70510340e-02 4.59411204e-01 -3.23941559e-01 -2.73857936e-02
4.95993346e-01 3.29891682e-01 2.11659595e-02 -6.84115142e-02
1.69716984e-01 2.24265829e-02 2.38667652e-01 -1.09069014e+00
8.90389442e-01 6.23129010e-01 4.34072077e-01 -7.49767423e-01
-3.84216726e-01 4.94878404e-02 -3.43560934e-01 -2.85978973e-01
4.65374559e-01 -2.56462693e-01 -1.12660849e+00 -1.46443680e-01
-4.62304085e-01 -8.54772925e-01 -7.02620089e-01 4.70806032e-01
-1.12740290e+00 4.07909960e-01 -2.56377935e-01 -1.53294134e+00
1.43923879e-01 -1.24567819e+00 4.70183372e-01 2.64171779e-01
-3.55000138e-01 -8.39693010e-01 4.01794434e-01 -4.63947177e-01
4.53264892e-01 5.48117995e-01 6.27255857e-01 -3.82016122e-01
-7.69167185e-01 4.96872723e-01 5.62343478e-01 -3.17889273e-01
7.90736675e-02 -1.31054381e-02 -4.77250934e-01 -6.14602149e-01
8.71594921e-02 -3.25453401e-01 9.24243689e-01 4.17027384e-01
1.08315873e+00 -5.99180460e-01 -6.91185117e-01 4.03561920e-01
1.47873044e+00 5.63450456e-01 6.29203200e-01 1.00826538e+00
7.27141872e-02 5.94416857e-01 1.06355548e+00 7.45832384e-01
9.58618522e-03 4.49325979e-01 9.13894534e-01 3.95890355e-01
8.24909806e-01 -2.22390771e-01 5.92013955e-01 4.85428460e-02
-1.41112342e-01 -1.73560128e-01 -9.10184741e-01 7.07635522e-01
-2.09961748e+00 -1.06245422e+00 3.04241836e-01 2.57219386e+00
8.90445590e-01 5.91685712e-01 1.62408605e-01 2.28700146e-01
2.44603470e-01 -6.02729470e-02 -8.78254116e-01 -9.02698278e-01
1.93399101e-01 2.29054540e-01 8.06531787e-01 1.04344583e+00
-1.10625911e+00 8.17017138e-01 6.71406460e+00 6.10324860e-01
-8.91137481e-01 -1.20714791e-01 2.90594935e-01 -4.23824728e-01
-6.60281181e-01 4.89139587e-01 -7.79965222e-01 3.57767344e-01
1.01021719e+00 -3.70053828e-01 5.00240564e-01 9.15911198e-01
2.80189395e-01 -5.38166285e-01 -1.03500545e+00 1.61083132e-01
-7.12806106e-01 -1.11462295e+00 -2.65445769e-01 5.07371724e-01
7.53724813e-01 -2.59225547e-01 1.20581813e-01 4.12663877e-01
9.43340719e-01 -1.20610952e+00 9.15742576e-01 3.73634428e-01
4.30217683e-01 -1.39910483e+00 3.19744080e-01 5.64187765e-01
-1.18601215e+00 -2.87577391e-01 -1.85308799e-01 -2.16159672e-01
2.98630238e-01 2.78101146e-01 -5.23088694e-01 6.01048410e-01
6.27132058e-01 1.24916390e-01 1.20564520e-01 9.29411113e-01
-3.35601896e-01 4.60476816e-01 -7.44266033e-01 -3.44658047e-01
8.52021337e-01 -3.31721365e-01 8.70966554e-01 7.52023935e-01
4.44171488e-01 5.85038811e-02 4.15068328e-01 8.41531157e-01
6.87291145e-01 -3.17456007e-01 -9.09662664e-01 7.37876147e-02
4.85395074e-01 7.46594191e-01 -1.06133318e+00 -1.82701543e-01
1.05312511e-01 6.07725859e-01 9.46664512e-02 3.22226435e-01
-9.87782359e-01 -3.31191719e-01 8.04619551e-01 -1.19948603e-01
2.32917249e-01 -4.77298081e-01 -3.20326149e-01 -8.82985830e-01
-9.35440063e-02 -8.12819600e-01 4.77194607e-01 -8.17395002e-02
-5.46597123e-01 4.99451190e-01 4.16322261e-01 -1.05709648e+00
-4.51448321e-01 -2.95227945e-01 -6.50469601e-01 6.57535732e-01
-1.55179751e+00 -6.17994130e-01 1.61268085e-01 3.76241356e-01
4.82910514e-01 1.14090994e-01 7.88253427e-01 -5.22692442e-01
-4.52249110e-01 1.06316827e-01 3.66066128e-01 -6.84633374e-01
1.76583290e-01 -1.53984833e+00 1.70344025e-01 1.14741683e+00
-2.30047524e-01 7.47436404e-01 1.26352358e+00 -1.03621364e+00
-1.47526741e+00 -8.94173265e-01 4.47951108e-01 -2.28578404e-01
7.51679301e-01 -1.32975042e-01 -7.96064377e-01 7.95528293e-01
2.03214288e-01 -3.64732087e-01 1.76751792e-01 2.99710810e-01
2.11446032e-01 2.81533331e-01 -9.00605083e-01 9.80096579e-01
1.00919271e+00 -9.22244042e-02 -4.69982684e-01 1.31556451e-01
4.91912246e-01 -3.52622151e-01 -5.90924144e-01 5.06436944e-01
5.36270738e-01 -7.92627633e-01 7.04838276e-01 -7.74531007e-01
4.11892831e-02 -5.04709601e-01 -2.05879807e-01 -1.15180135e+00
-1.41119510e-01 -1.23933721e+00 -6.19111717e-01 5.91844916e-01
2.44516820e-01 -7.41320252e-01 9.09122527e-01 6.68017745e-01
-6.04782626e-02 -1.00110519e+00 -1.26706529e+00 -1.53671646e+00
5.10995746e-01 -3.56304854e-01 7.62754321e-01 7.31802821e-01
5.43198109e-01 -3.96874666e-01 -3.89785707e-01 3.29696715e-01
9.22579765e-01 3.73662323e-01 3.67533863e-01 -9.25559402e-01
-4.97480720e-01 -5.94871640e-01 2.90189683e-01 -6.91841006e-01
2.20574230e-01 -3.56198758e-01 4.62847233e-01 -1.40126371e+00
-1.56708807e-02 -9.81501162e-01 -4.58910227e-01 4.47121710e-01
7.83576965e-02 -4.40778762e-01 3.91577370e-02 2.04223618e-01
-7.12391853e-01 8.40536654e-01 1.13162005e+00 1.54334232e-01
-6.61546409e-01 1.03055969e-01 -5.72087169e-01 6.84668660e-01
1.27401471e+00 -3.99978518e-01 -7.20224380e-01 2.18027964e-01
3.68488640e-01 2.94088781e-01 1.48303241e-01 -9.88761365e-01
8.91207680e-02 -9.81936693e-01 -6.80623427e-02 -5.48883080e-01
4.55685765e-01 -7.54541755e-01 4.67517704e-01 1.19135618e+00
-5.15796363e-01 -1.52297124e-01 2.34485507e-01 7.84702301e-01
2.54783154e-01 -3.78620923e-01 7.96686649e-01 -5.92546538e-04
-8.36363196e-01 1.97184414e-01 -8.38495672e-01 -1.08644716e-01
1.53200793e+00 -2.82916635e-01 -3.44748423e-02 -4.50024247e-01
-9.61560011e-01 6.23558283e-01 7.20412612e-01 1.20088011e-01
4.34329897e-01 -1.18701696e+00 -2.86140621e-01 -1.41638219e-01
-1.55307665e-01 -2.13098228e-01 -4.45335396e-02 7.12550402e-01
-2.56181002e-01 4.48609322e-01 -4.58367586e-01 -2.43894845e-01
-9.03686941e-01 8.10040593e-01 4.52247143e-01 -5.59182823e-01
-7.69548178e-01 4.50949878e-01 1.93311460e-02 -8.54212940e-02
2.02023387e-01 -4.44944143e-01 -6.05926067e-02 -2.56253570e-01
3.05626452e-01 4.07519877e-01 -3.45166206e-01 -2.37703659e-02
-4.29673225e-01 1.82435483e-01 6.57643974e-02 -9.09389615e-01
1.22262311e+00 -2.11619630e-01 1.57708943e-01 4.94199395e-01
4.00993109e-01 -1.01478785e-01 -1.51978135e+00 9.05977115e-02
7.87456036e-02 -4.77762938e-01 2.74709929e-02 -5.30522943e-01
-5.17985165e-01 5.96354425e-01 2.67895222e-01 4.60515261e-01
9.57564354e-01 -2.61595666e-01 4.35405314e-01 8.93497169e-01
9.32917595e-01 -1.52293849e+00 -3.04147899e-01 6.84403658e-01
6.90093935e-01 -7.75550067e-01 -7.07137818e-03 -1.39250159e-02
-6.16441369e-01 1.05764198e+00 6.55226231e-01 -1.75043270e-01
4.17873383e-01 4.79754090e-01 -5.46396136e-01 3.00798584e-02
-9.53904212e-01 -3.73598963e-01 -3.42114925e-01 3.92683268e-01
-2.51681745e-01 2.66586035e-01 -4.05680060e-01 2.54433692e-01
-3.34857583e-01 -2.16242090e-01 7.37263441e-01 1.44403148e+00
-9.86543775e-01 -1.26434731e+00 -5.65800965e-01 2.40884051e-01
-3.14909160e-01 3.34063828e-01 -4.49311510e-02 1.10512900e+00
-1.39448449e-01 8.77689123e-01 -1.28813908e-01 -7.91256130e-02
-8.97251517e-02 -6.40193522e-02 6.02940619e-01 -3.06966364e-01
-1.58368379e-01 -2.85962764e-02 2.01734334e-01 -9.95947897e-01
-9.46778804e-02 -1.03843951e+00 -1.48176241e+00 -5.51443279e-01
-3.14559162e-01 7.05922306e-01 3.57170701e-01 9.58088696e-01
-1.23109639e-01 5.43463290e-01 6.23580575e-01 -4.45852578e-01
-1.07324791e+00 -3.66782814e-01 -7.99600482e-01 -3.08975488e-01
5.45757115e-01 -1.01906049e+00 -3.47840697e-01 -3.82613480e-01] | [4.4851555824279785, 2.1935787200927734] |
8f13ee0b-d96c-4309-8586-f86d428dd2a6 | pslt-a-light-weight-vision-transformer-with | 2304.03481 | null | https://arxiv.org/abs/2304.03481v1 | https://arxiv.org/pdf/2304.03481v1.pdf | PSLT: A Light-weight Vision Transformer with Ladder Self-Attention and Progressive Shift | Vision Transformer (ViT) has shown great potential for various visual tasks due to its ability to model long-range dependency. However, ViT requires a large amount of computing resource to compute the global self-attention. In this work, we propose a ladder self-attention block with multiple branches and a progressive shift mechanism to develop a light-weight transformer backbone that requires less computing resources (e.g. a relatively small number of parameters and FLOPs), termed Progressive Shift Ladder Transformer (PSLT). First, the ladder self-attention block reduces the computational cost by modelling local self-attention in each branch. In the meanwhile, the progressive shift mechanism is proposed to enlarge the receptive field in the ladder self-attention block by modelling diverse local self-attention for each branch and interacting among these branches. Second, the input feature of the ladder self-attention block is split equally along the channel dimension for each branch, which considerably reduces the computational cost in the ladder self-attention block (with nearly 1/3 the amount of parameters and FLOPs), and the outputs of these branches are then collaborated by a pixel-adaptive fusion. Therefore, the ladder self-attention block with a relatively small number of parameters and FLOPs is capable of modelling long-range interactions. Based on the ladder self-attention block, PSLT performs well on several vision tasks, including image classification, objection detection and person re-identification. On the ImageNet-1k dataset, PSLT achieves a top-1 accuracy of 79.9% with 9.2M parameters and 1.9G FLOPs, which is comparable to several existing models with more than 20M parameters and 4G FLOPs. Code is available at https://isee-ai.cn/wugaojie/PSLT.html. | ['Qi Tian', 'Yutong Lu', 'Wei-Shi Zheng', 'Gaojie Wu'] | 2023-04-07 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [ 8.37274715e-02 -2.76953608e-01 1.99085057e-01 -4.14924920e-02
-2.67419368e-01 3.89431678e-02 3.41345847e-01 -8.77148414e-04
-5.03390491e-01 4.73519683e-01 -1.40751138e-01 -1.18845649e-01
7.21485987e-02 -9.42882299e-01 -7.40685880e-01 -9.40378904e-01
3.12028021e-01 1.14604905e-01 6.52913451e-01 -2.15896651e-01
1.18520543e-01 3.55802596e-01 -1.68584335e+00 9.80376154e-02
1.10549712e+00 1.28570879e+00 5.78418970e-01 4.95433509e-01
3.26258466e-02 4.67382908e-01 -4.30088669e-01 -4.34734493e-01
1.23017736e-01 -3.36933374e-01 -5.12757361e-01 -1.32033918e-02
5.51902711e-01 -2.44883150e-01 -5.23242176e-01 1.14980340e+00
8.47274065e-01 3.99663532e-03 2.93302387e-01 -1.20153010e+00
-7.21450448e-01 4.59436685e-01 -9.37524736e-01 5.21229327e-01
2.13132426e-02 2.82704115e-01 7.80344307e-01 -8.56650710e-01
2.79610306e-02 1.26555479e+00 6.87316298e-01 2.73325413e-01
-1.16995227e+00 -1.08130753e+00 2.47190192e-01 6.06823564e-01
-1.57013941e+00 -1.65603235e-01 6.47445142e-01 -2.57039756e-01
8.95440996e-01 1.47075400e-01 9.35710251e-01 6.10645413e-01
2.88294405e-01 6.95744216e-01 1.13784599e+00 -3.11623394e-01
2.59761196e-02 8.53348449e-02 2.32640266e-01 7.51327932e-01
2.23259822e-01 -7.82391950e-02 -5.71144819e-01 2.20127046e-01
1.05312037e+00 3.63777697e-01 -2.49184951e-01 3.69882621e-02
-9.05071855e-01 6.95990503e-01 1.02124107e+00 3.99049789e-01
-3.75054270e-01 1.44159153e-01 1.21728972e-01 1.35221794e-01
2.74844080e-01 2.24126708e-02 -2.95256197e-01 2.14994684e-01
-7.31079519e-01 -3.93138677e-02 1.27428308e-01 8.52302074e-01
9.47619200e-01 -1.44960992e-02 -4.35807586e-01 1.05475426e+00
2.05033541e-01 8.76167476e-01 4.93311822e-01 -6.18508101e-01
4.29965377e-01 9.11057174e-01 -2.31155023e-01 -1.02212405e+00
-4.45191264e-01 -7.56583452e-01 -1.36667550e+00 1.27650574e-01
2.44736195e-01 1.30962682e-04 -9.20659661e-01 1.67058671e+00
1.10428549e-01 1.85375661e-01 -1.76658630e-01 8.23537648e-01
8.13425899e-01 6.86134458e-01 1.30714357e-01 -1.79571658e-01
1.73995900e+00 -1.16916275e+00 -4.38469797e-01 -3.19528580e-01
2.15047210e-01 -5.87132037e-01 1.07758129e+00 1.78209618e-01
-1.17971635e+00 -1.02322626e+00 -8.83794427e-01 -1.17107794e-01
-1.36604175e-01 2.97761321e-01 4.63838398e-01 3.36736619e-01
-1.17190862e+00 2.69169182e-01 -5.07347524e-01 -4.60428357e-01
6.15472138e-01 5.96658468e-01 -8.10257867e-02 -1.66735977e-01
-1.20655525e+00 4.66784596e-01 1.83799833e-01 2.31726244e-01
-4.75138813e-01 -5.82038283e-01 -7.10144222e-01 3.13617945e-01
8.93425494e-02 -7.80405104e-01 7.92268574e-01 -9.20925140e-01
-1.29710388e+00 4.60467726e-01 -3.56209517e-01 -3.80548745e-01
2.92186946e-01 -1.14359315e-02 -3.38938802e-01 1.53766781e-01
1.88249394e-01 9.72642660e-01 1.13365233e+00 -7.66541779e-01
-8.31545293e-01 -6.11564279e-01 7.03974888e-02 3.51248592e-01
-7.94360161e-01 2.87254211e-02 -9.08957422e-01 -7.70918906e-01
9.74476710e-02 -1.05424941e+00 -1.42190412e-01 -5.42934313e-02
-8.73119161e-02 -2.02143282e-01 8.42695594e-01 -4.93304163e-01
1.45820582e+00 -2.21399641e+00 4.07989090e-03 -3.47758718e-02
4.42448050e-01 5.91753602e-01 -9.47174281e-02 -1.25093269e-03
-4.02080007e-02 -2.07545847e-01 -2.66067803e-01 -2.59886652e-01
-3.55418116e-01 -4.73909751e-02 1.54472040e-02 1.04531489e-01
2.81031653e-02 1.15131176e+00 -5.60825109e-01 -5.52039385e-01
3.61405522e-01 6.39111578e-01 -5.81795096e-01 -2.47134715e-02
3.95975500e-01 2.30451599e-01 -3.92849863e-01 6.01652324e-01
1.07960963e+00 -5.43637156e-01 -1.78913057e-01 -5.31853080e-01
-3.45059067e-01 -5.22451513e-02 -1.03268027e+00 1.26708472e+00
-5.08830607e-01 5.63005567e-01 -9.56836566e-02 -8.13793361e-01
1.01841331e+00 6.52979910e-02 3.76031250e-01 -1.16094029e+00
2.03023463e-01 3.37199979e-02 -3.60892564e-02 -2.40539014e-01
2.66744614e-01 3.03381179e-02 3.82717401e-02 2.53843486e-01
-4.82164090e-03 3.98749739e-01 7.02938214e-02 1.12488888e-01
9.14626658e-01 -2.59871721e-01 3.65253203e-02 -1.96464255e-01
8.52895617e-01 -3.74450773e-01 6.85558558e-01 5.80654085e-01
-2.04379544e-01 3.61116439e-01 -1.11585564e-03 -4.32721198e-01
-9.34165001e-01 -9.22339022e-01 -1.20922461e-01 1.10894477e+00
5.98758936e-01 -3.89473617e-01 -9.21243548e-01 -3.71226281e-01
2.33752672e-02 2.31460214e-01 -4.28841263e-01 -3.09323490e-01
-5.09283066e-01 -8.84160280e-01 4.88001972e-01 8.61538053e-01
1.32619953e+00 -1.24978864e+00 -7.16774523e-01 2.09620804e-01
-2.68665016e-01 -8.90740156e-01 -7.96229601e-01 -9.81227979e-02
-6.88458264e-01 -8.96411300e-01 -1.07233012e+00 -8.08305085e-01
7.80352890e-01 6.10515475e-01 7.17084467e-01 4.18422893e-02
-3.39570254e-01 1.85373425e-01 -1.30260348e-01 -4.31979507e-01
3.73259991e-01 2.20251102e-02 4.76907473e-03 3.44108254e-01
3.85208160e-01 -4.91188288e-01 -8.79456699e-01 5.38566411e-01
-6.00528181e-01 1.53752133e-01 8.30743730e-01 9.40396011e-01
6.27598047e-01 1.19284779e-01 5.03090382e-01 -3.26346844e-01
3.48116279e-01 -4.65485454e-02 -6.84905350e-01 2.23883763e-01
-3.95558417e-01 -1.92649424e-01 6.29396200e-01 -7.08662212e-01
-1.14376187e+00 -8.15735981e-02 -1.04097292e-01 -4.88381714e-01
2.52857506e-01 -8.22499171e-02 -2.31983468e-01 -2.88429141e-01
3.58284652e-01 4.34177101e-01 -2.08504677e-01 -5.26561975e-01
6.20427467e-02 7.56095111e-01 4.34630454e-01 -1.32481977e-01
6.38985097e-01 3.58489156e-01 -1.01918010e-02 -8.35336387e-01
-5.47016442e-01 -1.29271597e-01 -4.19142365e-01 -1.84587255e-01
8.50722671e-01 -1.02806461e+00 -1.21417522e+00 1.15927613e+00
-9.01730835e-01 -2.56428063e-01 -2.58896112e-01 4.72725898e-01
-1.62744552e-01 3.13602597e-01 -6.97259188e-01 -6.90426052e-01
-6.79441750e-01 -1.24320924e+00 8.23825836e-01 7.50846922e-01
1.17946513e-01 -6.16595447e-01 -6.13407373e-01 5.18692374e-01
5.83446503e-01 -3.66305947e-01 6.87794268e-01 -6.17143996e-02
-7.22211182e-01 1.23374961e-01 -7.67094791e-01 4.33266133e-01
1.98168922e-02 -5.34660041e-01 -9.54245269e-01 -5.28549910e-01
-2.57990882e-02 -7.87356645e-02 1.04400694e+00 5.55295706e-01
1.20599055e+00 -5.65843992e-02 -4.48733479e-01 7.57661343e-01
1.23188460e+00 4.29495573e-01 6.54595017e-01 2.85038799e-01
8.46075416e-01 3.05739492e-01 4.48454648e-01 2.71460444e-01
6.02851450e-01 7.86639094e-01 3.18759024e-01 -5.65810204e-01
-3.54396433e-01 -1.21722884e-01 2.14964360e-01 7.31885195e-01
-2.90354699e-01 -1.08428545e-01 -7.49414384e-01 3.63140792e-01
-1.78014362e+00 -8.87661517e-01 -1.78988263e-01 2.21145892e+00
6.18500650e-01 7.51523525e-02 3.69493738e-02 1.43640116e-01
9.04091477e-01 2.90891547e-02 -7.60041714e-01 -2.12902710e-01
-2.52489984e-01 2.06318617e-01 6.31737471e-01 4.53247398e-01
-8.89831126e-01 9.08927143e-01 5.22471619e+00 1.06008387e+00
-1.19661129e+00 7.13255480e-02 6.68022633e-01 -3.45622040e-02
1.73025712e-01 -1.29928365e-01 -1.22617328e+00 8.83000195e-01
4.84388292e-01 -2.86683217e-02 3.99164557e-01 5.20034611e-01
3.17504816e-02 -3.39188665e-01 -6.24217212e-01 1.35372066e+00
1.56977534e-01 -1.02039766e+00 1.84187010e-01 2.04543382e-01
4.74544615e-01 7.41211697e-02 1.49274543e-01 1.20844379e-01
1.87716670e-02 -8.05191755e-01 4.98248965e-01 5.38627803e-01
9.76894259e-01 -7.50272989e-01 8.23761463e-01 3.07637185e-01
-1.59153354e+00 -5.65401912e-01 -3.81450832e-01 -5.94235361e-02
1.97340585e-02 6.18178606e-01 -8.83706138e-02 3.54792655e-01
1.24806166e+00 5.93521833e-01 -6.81361079e-01 1.01716816e+00
-1.66604608e-01 3.97120267e-01 -4.12310928e-01 8.40038136e-02
6.18251748e-02 -2.66498059e-01 4.62997854e-01 1.13183880e+00
4.29844946e-01 2.03002498e-01 9.88541245e-02 5.64612329e-01
-1.29627883e-01 -9.36713368e-02 -1.61143150e-02 5.41566670e-01
5.02595544e-01 1.16901672e+00 -4.34616983e-01 -5.52559078e-01
-4.80185091e-01 1.13030660e+00 2.37114310e-01 3.48143578e-01
-9.38883960e-01 -7.77055562e-01 5.28914452e-01 2.84774423e-01
6.08143151e-01 1.01522289e-01 -1.69192195e-01 -1.02565968e+00
8.18454400e-02 -6.38173342e-01 3.92172992e-01 -1.00002956e+00
-1.06901467e+00 8.86087954e-01 -1.76859096e-01 -1.02243996e+00
3.48692209e-01 -4.64737505e-01 -5.11481762e-01 1.14130270e+00
-1.62623525e+00 -1.41922307e+00 -7.53008127e-01 8.82821381e-01
4.82653111e-01 -1.41804501e-01 4.77282912e-01 4.82851923e-01
-8.61229718e-01 8.57627332e-01 -1.42414853e-01 -1.01307675e-03
7.29448318e-01 -9.26126122e-01 2.36885935e-01 7.58094609e-01
-2.94099778e-01 6.47584558e-01 2.63262361e-01 -4.89523321e-01
-1.03918409e+00 -1.08930218e+00 9.68411267e-01 1.14970185e-01
4.08706963e-01 -3.18698406e-01 -1.05736315e+00 3.95756006e-01
1.15223505e-01 2.05346107e-01 3.38350832e-01 -8.79368782e-02
-3.03639799e-01 -7.38049150e-01 -1.04816544e+00 5.32922268e-01
1.16438103e+00 -3.13256681e-01 -3.58474404e-01 2.99088471e-02
3.94302130e-01 -1.62599176e-01 -7.37721384e-01 3.19057584e-01
6.33609354e-01 -9.62208331e-01 1.02123559e+00 3.33201021e-01
-6.10402739e-03 -5.66316843e-01 -1.03174858e-02 -1.06871998e+00
-9.89433646e-01 -4.03275847e-01 -5.49616618e-03 1.28768408e+00
1.71921551e-02 -9.64118063e-01 5.44429421e-01 2.35958755e-01
-1.38183385e-02 -9.11163092e-01 -9.62062716e-01 -6.68265164e-01
-2.92974144e-01 -1.14687175e-01 5.38474143e-01 4.17317033e-01
-2.85208255e-01 6.34075403e-01 -2.61458457e-01 9.65980515e-02
7.23290265e-01 9.36770365e-02 5.25203705e-01 -1.31047630e+00
-2.30028138e-01 -3.40844393e-01 -4.12304610e-01 -1.45376575e+00
-3.25617701e-01 -5.56648970e-01 -3.00512046e-01 -1.47619855e+00
4.43379819e-01 -5.34229755e-01 -5.27231514e-01 7.29489684e-01
-3.43183577e-01 7.10437655e-01 4.37863827e-01 3.22860241e-01
-4.56162244e-01 6.71319783e-01 1.32923222e+00 -2.76997179e-01
-3.02514881e-01 8.45486298e-02 -8.74177158e-01 8.79993677e-01
7.38499463e-01 -1.48005351e-01 -4.30529207e-01 -5.06957889e-01
-1.81638435e-01 -2.54347712e-01 6.95768595e-01 -1.26891744e+00
6.13447011e-01 1.65927976e-01 7.36271143e-01 -7.79905260e-01
5.26754856e-01 -7.26347566e-01 -2.09751446e-03 6.70004606e-01
1.09139189e-01 5.86684868e-02 4.07883257e-01 3.06928098e-01
-1.87331289e-01 8.86128843e-02 1.01507926e+00 8.15805863e-04
-6.94727302e-01 4.86803055e-01 -2.86205471e-01 -2.73456156e-01
1.18816876e+00 -5.46264768e-01 -2.07184494e-01 -7.27759376e-02
-6.14628553e-01 3.41716290e-01 2.77283818e-01 3.89427215e-01
6.54279292e-01 -1.38185680e+00 -5.41032314e-01 6.43019080e-01
-2.48558912e-02 -9.06483829e-02 8.23346019e-01 8.30235720e-01
-1.00058012e-01 4.75456268e-01 -4.66522664e-01 -7.70535290e-01
-1.58777893e+00 5.57831645e-01 3.09860855e-01 -3.27642381e-01
-6.70654833e-01 9.72402871e-01 6.59503341e-01 1.80224195e-01
-2.53261998e-02 -1.83705971e-01 -3.60670358e-01 1.10187657e-01
7.65384793e-01 6.10054076e-01 -1.01934031e-01 -8.85588288e-01
-4.70752269e-01 1.05642986e+00 -3.08489203e-01 1.58803120e-01
1.18885279e+00 -3.25732410e-01 -1.35386303e-01 2.14896370e-02
9.94781256e-01 -9.12150592e-02 -1.51408017e+00 -5.07760763e-01
-6.29493892e-01 -4.04677927e-01 3.63217711e-01 -6.83461845e-01
-1.31512594e+00 1.04479432e+00 9.52794850e-01 -9.28870738e-02
1.81338680e+00 -6.74692332e-04 9.32654142e-01 4.91744578e-02
3.75062644e-01 -9.57574785e-01 2.30145022e-01 4.89655316e-01
8.61338854e-01 -9.00869906e-01 -5.46551384e-02 -4.58613098e-01
-6.01661980e-01 6.43612862e-01 9.54194903e-01 -3.02796252e-02
5.78762233e-01 3.01175743e-01 -2.36071691e-01 -5.49545977e-03
-3.71246427e-01 -3.14438254e-01 2.62756079e-01 6.68884873e-01
1.44110486e-01 1.42392535e-02 -1.32385671e-01 4.98491615e-01
3.86359952e-02 7.15085194e-02 -3.93753611e-02 5.08449376e-01
-4.10076082e-01 -9.11448240e-01 -4.69114780e-01 3.79066974e-01
-2.06397623e-01 -3.40188980e-01 7.86369666e-02 3.65132272e-01
6.00400865e-01 1.01802778e+00 3.91307443e-01 -5.02731860e-01
4.14142549e-01 -4.22410429e-01 3.21689099e-01 -2.24496275e-01
-6.46686673e-01 2.69815087e-01 -5.46482563e-01 -4.31087941e-01
-3.53435129e-01 -4.55933064e-01 -1.08612585e+00 -4.05229479e-01
-3.87478471e-01 4.67413627e-02 3.36975455e-01 6.60584390e-01
6.41552567e-01 8.21440637e-01 6.21631682e-01 -7.32203960e-01
-1.51550144e-01 -1.11197686e+00 -5.00127137e-01 1.45279363e-01
8.98591727e-02 -6.92487419e-01 -5.11360180e-04 -3.39119099e-02] | [10.853123664855957, -1.6956642866134644] |
3f9f0f89-ac6e-40ce-87e0-fd16ba25ba6c | deep-intra-image-contrastive-learning-for | 2302.04607 | null | https://arxiv.org/abs/2302.04607v1 | https://arxiv.org/pdf/2302.04607v1.pdf | Deep Intra-Image Contrastive Learning for Weakly Supervised One-Step Person Search | Weakly supervised person search aims to perform joint pedestrian detection and re-identification (re-id) with only person bounding-box annotations. Recently, the idea of contrastive learning is initially applied to weakly supervised person search, where two common contrast strategies are memory-based contrast and intra-image contrast. We argue that current intra-image contrast is shallow, which suffers from spatial-level and occlusion-level variance. In this paper, we present a novel deep intra-image contrastive learning using a Siamese network. Two key modules are spatial-invariant contrast (SIC) and occlusion-invariant contrast (OIC). SIC performs many-to-one contrasts between two branches of Siamese network and dense prediction contrasts in one branch of Siamese network. With these many-to-one and dense contrasts, SIC tends to learn discriminative scale-invariant and location-invariant features to solve spatial-level variance. OIC enhances feature consistency with the masking strategy to learn occlusion-invariant features. Extensive experiments are performed on two person search datasets CUHK-SYSU and PRW, respectively. Our method achieves a state-of-the-art performance among weakly supervised one-step person search approaches. We hope that our simple intra-image contrastive learning can provide more paradigms on weakly supervised person search. The source code is available at \url{https://github.com/jiabeiwangTJU/DICL}. | ['Xuelong Li', 'Zhuang Shao', 'Hanqing Sun', 'Jiale Cao', 'Yanwei Pang', 'Jiabei Wang'] | 2023-02-09 | null | null | null | null | ['pedestrian-detection', 'person-search'] | ['computer-vision', 'computer-vision'] | [-1.53807998e-01 -4.84554917e-01 -2.65463412e-01 -3.84237051e-01
-6.63910568e-01 -2.86671251e-01 7.19984531e-01 -2.39491984e-01
-9.50592756e-01 7.17987716e-01 3.15939486e-01 2.54379958e-01
7.13979229e-02 -6.57718122e-01 -5.97486317e-01 -6.65721297e-01
5.57971671e-02 4.61151063e-01 5.45348108e-01 -2.67643183e-01
-2.84829326e-02 3.64187807e-01 -1.80439293e+00 2.42805719e-01
8.04584205e-01 8.60191643e-01 1.80222362e-01 6.01177096e-01
3.43307406e-01 4.51134324e-01 -4.57988858e-01 -6.10280097e-01
5.44413507e-01 -3.57370466e-01 -8.45580161e-01 -2.00038448e-01
8.92135024e-01 -2.74175018e-01 -7.62418985e-01 1.13053894e+00
9.56954896e-01 2.52282679e-01 4.93611634e-01 -1.29180598e+00
-6.85310245e-01 1.10083167e-02 -7.31835008e-01 6.58750713e-01
3.97049665e-01 7.76623428e-01 7.84677267e-01 -9.40033495e-01
3.17379326e-01 1.30702305e+00 7.36011624e-01 6.46819115e-01
-1.12490165e+00 -7.56778955e-01 2.35227585e-01 5.21041572e-01
-1.70550179e+00 -1.55487090e-01 8.00284386e-01 -2.94343740e-01
8.73873651e-01 4.82015133e-01 8.99058878e-01 1.07412505e+00
-2.27037504e-01 1.00104523e+00 1.40384245e+00 -3.96041811e-01
-1.25658229e-01 1.28879711e-01 2.77648717e-01 9.98673260e-01
3.23804379e-01 8.27074707e-01 -7.03842580e-01 -1.02650151e-01
9.26167250e-01 7.29790330e-02 -2.10458040e-01 -2.07334712e-01
-1.02237761e+00 5.83564878e-01 9.37025368e-01 2.42538810e-01
-1.03052240e-03 3.72028686e-02 3.81770045e-01 9.98395979e-02
2.59755939e-01 7.79513568e-02 -2.06264049e-01 3.40579674e-02
-9.39057350e-01 4.15801525e-01 4.00035173e-01 9.01136041e-01
7.88868487e-01 -1.95280582e-01 -6.69213891e-01 1.05084383e+00
1.91424787e-01 7.10032880e-01 4.27905351e-01 -7.40262449e-01
4.65627819e-01 4.36774075e-01 1.18218042e-01 -8.74595046e-01
-4.98029262e-01 -8.84107590e-01 -9.81597066e-01 2.72517502e-01
8.13451052e-01 1.47071272e-01 -8.72978032e-01 1.86331213e+00
2.30855227e-01 1.12922952e-01 -5.26278734e-01 1.42675447e+00
1.08638680e+00 1.76009566e-01 4.43429947e-01 2.80397952e-01
1.80459023e+00 -1.43917882e+00 -1.72495127e-01 -5.43344557e-01
1.01409949e-01 -6.00383461e-01 1.33564079e+00 -1.89714462e-01
-1.37652862e+00 -1.13958156e+00 -9.42322373e-01 -1.95720226e-01
-4.60529119e-01 3.06092322e-01 3.88633907e-01 6.61253214e-01
-1.11887860e+00 2.76977539e-01 -3.88437212e-01 -5.19132972e-01
5.66971898e-01 2.89614558e-01 -3.18276733e-01 -7.39106089e-02
-1.32334530e+00 7.75692523e-01 3.96995656e-02 7.95206875e-02
-5.22844553e-01 -6.30900204e-01 -7.17297316e-01 1.22660741e-01
3.05481613e-01 -1.10873151e+00 9.37515497e-01 -6.60814703e-01
-1.06653833e+00 1.45542884e+00 -4.89286870e-01 -3.27043444e-01
8.62941444e-01 -1.91012144e-01 -2.86454707e-01 2.57381111e-01
5.33207476e-01 8.97625268e-01 7.23941267e-01 -1.22621977e+00
-8.12101305e-01 -6.38165176e-01 -2.33418986e-01 3.03753704e-01
-2.28284448e-01 3.47527415e-01 -9.28135574e-01 -8.39345515e-01
-5.79042919e-02 -8.44934046e-01 -6.17160723e-02 4.25324023e-01
-1.82461470e-01 -4.54224467e-01 3.87047529e-01 -7.37441421e-01
1.13884425e+00 -1.83791566e+00 -1.04389444e-01 1.78333968e-01
2.27018461e-01 3.36704105e-01 -3.79506558e-01 -1.05191797e-01
3.17998715e-02 -2.68691659e-01 6.23922376e-03 -5.46279967e-01
2.63772719e-02 -2.84287900e-01 2.05614761e-01 6.60084188e-01
8.77492204e-02 1.43101192e+00 -8.76562238e-01 -7.26191163e-01
2.05475837e-01 4.05759066e-01 -3.99468154e-01 9.44138095e-02
3.70065570e-01 5.64986825e-01 -9.73161012e-02 1.08211660e+00
8.35421801e-01 -2.74497569e-01 -3.71823758e-01 -4.60264653e-01
-2.02009916e-01 -1.34880900e-01 -1.02138686e+00 1.49597311e+00
-3.71685252e-03 5.01374900e-01 -6.65426394e-03 -8.51323128e-01
5.98796427e-01 -3.04613322e-01 2.00717360e-01 -1.26671422e+00
7.63400719e-02 2.00741217e-01 -2.18994528e-01 -1.00269161e-01
3.19043040e-01 2.66890019e-01 -1.33326322e-01 7.03829676e-02
-4.38099168e-02 3.61892849e-01 1.80496976e-01 -4.24678344e-03
7.40334332e-01 -2.50147413e-02 1.94522873e-01 -5.41616499e-01
9.30075586e-01 -2.46640250e-01 5.31125665e-01 1.19917059e+00
-8.66427362e-01 7.66538680e-01 -1.18058279e-01 -7.38307297e-01
-1.16211677e+00 -1.30575597e+00 -2.10302383e-01 1.25268924e+00
5.64215422e-01 -2.82270879e-01 -6.94468200e-01 -6.03047311e-01
9.30612460e-02 7.90105015e-02 -4.97709513e-01 -1.06424600e-01
-8.46958756e-01 -6.60425365e-01 7.69165635e-01 7.95985460e-01
1.34257448e+00 -9.85006571e-01 -4.30735707e-01 -1.09752044e-01
-2.66058415e-01 -1.06697130e+00 -1.09668016e+00 -3.20473611e-02
-2.94270009e-01 -1.03267002e+00 -1.40345490e+00 -1.21951222e+00
6.96584105e-01 4.87822145e-01 1.34507632e+00 3.17553252e-01
-6.74821675e-01 4.76849139e-01 5.41699268e-02 2.79632676e-02
3.59937966e-01 4.31556674e-03 2.56198943e-01 -2.87317634e-01
7.22181380e-01 -3.22927982e-01 -1.24148905e+00 8.11451077e-01
-2.48290554e-01 -1.28119066e-01 5.03024280e-01 1.20748389e+00
6.92771733e-01 -1.09069929e-01 -4.85632904e-02 -2.75436249e-02
3.56535137e-01 1.01379275e-01 -5.90471208e-01 3.39000612e-01
-5.59252143e-01 -1.72855914e-01 2.17552960e-01 -6.01502478e-01
-1.08690870e+00 9.51307360e-03 -1.73179358e-01 -9.49283764e-02
-3.60771567e-01 -1.14188157e-01 -2.92922288e-01 -3.60692501e-01
7.49238968e-01 5.51500857e-01 -2.14033365e-01 -2.67977417e-01
1.59647524e-01 4.87770200e-01 9.62226033e-01 -6.97692275e-01
8.86189342e-01 6.82593226e-01 -7.82657936e-02 -6.38527811e-01
-8.39583457e-01 -7.75557399e-01 -7.79492080e-01 -2.99758404e-01
9.84959960e-01 -1.24161482e+00 -9.53520775e-01 7.97039151e-01
-8.72499347e-01 -5.24959981e-01 -3.85679394e-01 3.34630251e-01
-2.22823992e-01 5.51067054e-01 -7.46484935e-01 -7.01663554e-01
-5.14384270e-01 -1.03784895e+00 1.16401875e+00 6.21138930e-01
-4.91087846e-02 -6.60806239e-01 -4.04702201e-02 6.07724488e-01
4.24480200e-01 -1.94466665e-01 2.20935464e-01 -2.39049911e-01
-6.16029084e-01 -2.46294439e-01 -8.13095689e-01 1.52703553e-01
-2.08114833e-01 -8.02396476e-01 -9.33237255e-01 -6.00282729e-01
-4.14093673e-01 -2.96008795e-01 1.26854551e+00 5.20406663e-01
1.19985580e+00 -2.21556231e-01 -4.47695106e-01 8.47769439e-01
1.17452240e+00 -1.53755844e-01 5.75317502e-01 6.39196455e-01
7.06001997e-01 5.23060262e-01 4.50193018e-01 1.87896699e-01
6.22767389e-01 1.24509621e+00 -1.50296345e-01 -5.33764660e-01
-7.33604372e-01 -4.15266156e-01 5.62155843e-02 -5.28291762e-02
-4.44577068e-01 1.40574977e-01 -8.45236003e-01 4.02782559e-01
-1.94744170e+00 -1.43332648e+00 -3.29973698e-02 2.12331390e+00
8.16657782e-01 1.70573890e-01 6.59640789e-01 -1.78023934e-01
8.42773616e-01 1.04827926e-01 -5.85265160e-01 2.98371702e-01
-6.37805223e-01 9.54180881e-02 5.15496790e-01 5.89518070e-01
-1.46187246e+00 9.51965630e-01 5.25530863e+00 1.06292164e+00
-7.29123950e-01 4.16758716e-01 8.73157561e-01 -2.22945705e-01
1.29322827e-01 -2.46244416e-01 -1.17407668e+00 8.20767999e-01
2.04282388e-01 3.09847414e-01 3.02837580e-01 9.14726615e-01
-2.15475261e-02 -2.61754721e-01 -9.96364355e-01 1.60456669e+00
2.57826626e-01 -1.17486680e+00 -2.07336962e-01 -3.77520849e-03
6.56221986e-01 -6.70706853e-02 3.82405758e-01 4.07206059e-01
-1.02080867e-01 -9.22545791e-01 7.84113765e-01 4.94366318e-01
7.96295464e-01 -6.98438406e-01 7.42689967e-01 2.03038052e-01
-1.79952955e+00 -2.42188886e-01 -3.43725532e-01 -2.70577706e-02
1.89129531e-01 2.54204959e-01 4.26168703e-02 6.30102158e-02
1.38644648e+00 4.65636492e-01 -9.95322168e-01 1.24354637e+00
-1.74826741e-01 1.08768612e-01 -3.52230281e-01 -1.33955002e-01
7.80101418e-02 -1.41119108e-01 6.25424683e-01 1.60946977e+00
-1.89491868e-01 -1.79294944e-02 4.23109353e-01 1.00925446e+00
1.83213294e-01 -2.28852198e-01 -1.39997736e-01 8.56064916e-01
3.35920274e-01 9.25685823e-01 -6.09007835e-01 -5.52174032e-01
-5.21730304e-01 1.55957139e+00 3.52235883e-01 4.75379854e-01
-8.94910157e-01 -8.62056464e-02 6.18569195e-01 4.09724981e-01
2.49462396e-01 -1.44461393e-01 -4.24454480e-01 -1.28453588e+00
3.61001909e-01 -7.61281610e-01 7.55165875e-01 -5.44050932e-01
-1.62065232e+00 5.42642415e-01 1.48649141e-01 -1.07727659e+00
1.04272202e-01 -4.48245317e-01 -5.89724958e-01 1.10721171e+00
-1.71515346e+00 -1.59591138e+00 -7.17711806e-01 8.96866202e-01
7.03488946e-01 -5.32215714e-01 4.33257729e-01 5.37921250e-01
-5.08604348e-01 1.20437586e+00 -3.40276062e-01 4.57960218e-01
8.54338229e-01 -1.13705134e+00 4.32613283e-01 9.28824663e-01
-7.26091042e-02 7.35147774e-01 3.82170409e-01 -7.62497127e-01
-8.41885924e-01 -8.82403612e-01 9.69211519e-01 -5.57523966e-01
2.98355788e-01 -4.51741815e-01 -6.30403161e-01 1.37775391e-01
3.55246589e-02 4.27710801e-01 4.46575969e-01 4.54728007e-02
-5.37905872e-01 -2.23071203e-01 -1.17980254e+00 7.30092347e-01
1.69103885e+00 -7.15201676e-01 -3.37945074e-01 3.25107694e-01
3.12499851e-01 -2.68032402e-01 -4.17642981e-01 4.13150877e-01
7.94123292e-01 -1.21381235e+00 1.63410068e+00 -3.24074715e-01
-1.28565639e-01 -4.85408962e-01 2.03758348e-02 -7.59928465e-01
-8.42373490e-01 -2.93912739e-01 -1.29170746e-01 1.09713030e+00
1.31617054e-01 -7.15266585e-01 9.68771219e-01 7.47204065e-01
2.10862622e-01 -5.95972657e-01 -1.08731019e+00 -1.08866143e+00
1.19637422e-01 -2.39575356e-02 4.13905084e-01 4.60606694e-01
-4.04868156e-01 2.02251703e-01 -4.38255340e-01 1.42180845e-01
1.12754476e+00 2.71802559e-03 7.48407662e-01 -1.09973955e+00
-5.17409205e-01 -7.72933424e-01 -4.35611486e-01 -1.43661320e+00
7.27749169e-02 -7.92738676e-01 -1.31393149e-01 -1.13858020e+00
7.69733548e-01 -3.68728727e-01 -2.83042401e-01 3.48023385e-01
-4.88495648e-01 6.05671465e-01 3.26321810e-01 4.81859177e-01
-7.51787961e-01 5.66905975e-01 1.13536263e+00 -5.20016015e-01
-2.08041251e-01 9.42443311e-02 -5.28817952e-01 5.85008025e-01
7.26351559e-01 -4.31687534e-02 4.48585413e-02 -3.13544303e-01
-3.12144190e-01 -5.27707338e-01 1.21814179e+00 -1.18858302e+00
7.22030699e-01 1.71741694e-01 9.77859437e-01 -8.02696764e-01
4.84919578e-01 -4.04855013e-01 -1.82774752e-01 7.88929045e-01
-2.53696382e-01 6.15220517e-02 5.62944897e-02 3.90626371e-01
-2.44587213e-01 1.47039473e-01 1.04532480e+00 -4.40620184e-01
-1.01149881e+00 5.20528793e-01 1.46113083e-01 1.46613091e-01
6.34000003e-01 -4.33958739e-01 -4.12807435e-01 -3.14552426e-01
-5.53941667e-01 3.59269440e-01 3.69581193e-01 3.79107684e-01
7.14234233e-01 -1.39163530e+00 -8.67663503e-01 3.38463902e-01
2.06746936e-01 -3.84802073e-01 4.29502070e-01 8.25763881e-01
-2.24312708e-01 5.57338178e-01 -1.60154328e-01 -7.90902197e-01
-1.52341449e+00 5.45690835e-01 5.66779971e-01 -2.53817201e-01
-5.11767507e-01 1.31712306e+00 4.20297205e-01 -3.17469358e-01
5.79731524e-01 2.54686981e-01 6.18214011e-02 -2.64256239e-01
7.25636899e-01 5.27094841e-01 -4.32033062e-01 -9.36385214e-01
-6.29409432e-01 9.83951509e-01 -1.00232624e-02 -2.38961112e-02
9.40742552e-01 -2.71782666e-01 3.56167853e-02 -3.05483580e-01
1.23291802e+00 -2.33984366e-01 -1.43585002e+00 -4.49798375e-01
-2.08144248e-01 -6.39110684e-01 -1.50764182e-01 -8.81782413e-01
-9.29705024e-01 6.51830792e-01 1.35161161e+00 -3.38639200e-01
1.01313901e+00 3.57426465e-01 8.82410645e-01 2.42215514e-01
4.71212536e-01 -1.33169210e+00 4.91070837e-01 2.77808070e-01
9.29400861e-01 -1.71713138e+00 6.24933615e-02 -3.97549093e-01
-4.78408247e-01 5.77158570e-01 1.11054850e+00 -6.41113669e-02
6.80116355e-01 1.44212514e-01 -2.07881063e-01 -6.97001070e-02
-4.94736284e-02 -8.89544725e-01 8.13579679e-01 9.03199255e-01
7.74889961e-02 -3.59815778e-03 -2.63342559e-01 6.47820055e-01
-2.33684868e-01 -2.11348981e-01 -6.15525067e-01 6.54607832e-01
-2.73792893e-01 -1.02639282e+00 -6.10187411e-01 1.63248062e-01
-6.55187806e-03 -2.51992136e-01 -3.52756590e-01 7.50017107e-01
4.85376924e-01 9.19961989e-01 1.44506218e-02 -1.96975857e-01
4.83677894e-01 -2.11572438e-01 6.19786561e-01 -8.10486823e-02
-6.28235221e-01 1.25271175e-02 -8.13115090e-02 -7.32238352e-01
-4.06218529e-01 -7.02472091e-01 -8.48629892e-01 -3.06139678e-01
-7.66571909e-02 -1.78593040e-01 9.36818272e-02 6.40011966e-01
2.77242303e-01 4.21435013e-02 3.01049173e-01 -9.10985827e-01
-3.71940196e-01 -7.62100577e-01 -2.51904279e-01 6.42029762e-01
3.42483938e-01 -8.48444402e-01 -2.66431421e-01 -1.97967499e-01] | [14.851670265197754, 0.7884725332260132] |
f73cbf79-20ec-4bb6-ba01-97a4c1c45f28 | an-object-slam-framework-for-association | 2305.07299 | null | https://arxiv.org/abs/2305.07299v1 | https://arxiv.org/pdf/2305.07299v1.pdf | An Object SLAM Framework for Association, Mapping, and High-Level Tasks | Object SLAM is considered increasingly significant for robot high-level perception and decision-making. Existing studies fall short in terms of data association, object representation, and semantic mapping and frequently rely on additional assumptions, limiting their performance. In this paper, we present a comprehensive object SLAM framework that focuses on object-based perception and object-oriented robot tasks. First, we propose an ensemble data association approach for associating objects in complicated conditions by incorporating parametric and nonparametric statistic testing. In addition, we suggest an outlier-robust centroid and scale estimation algorithm for modeling objects based on the iForest and line alignment. Then a lightweight and object-oriented map is represented by estimated general object models. Taking into consideration the semantic invariance of objects, we convert the object map to a topological map to provide semantic descriptors to enable multi-map matching. Finally, we suggest an object-driven active exploration strategy to achieve autonomous mapping in the grasping scenario. A range of public datasets and real-world results in mapping, augmented reality, scene matching, relocalization, and robotic manipulation have been used to evaluate the proposed object SLAM framework for its efficient performance. | ['Jian Zhang', 'Xin Chen', 'Wenkai Sun', 'Zhiqiang Deng', 'Delong Zhu', 'Yunzhou Zhang', 'Yanmin Wu'] | 2023-05-12 | null | null | null | null | ['object-slam'] | ['computer-vision'] | [ 1.16598666e-01 -3.19128990e-01 -2.15241432e-01 -8.54074717e-01
-5.47905087e-01 -4.28077281e-01 5.16047597e-01 3.75896186e-01
-4.53473479e-01 5.66903412e-01 -1.70975521e-01 2.14089781e-01
-8.63652408e-01 -6.44004643e-01 -8.32163513e-01 -4.56309855e-01
-2.02421859e-01 1.04036164e+00 3.04325253e-01 3.03896740e-02
8.36266935e-01 8.07467639e-01 -1.92529380e+00 -2.00258031e-01
1.19737220e+00 1.13447785e+00 9.04196620e-01 2.43475154e-01
-2.17823043e-01 1.13544762e-01 -2.05835268e-01 1.18900217e-01
4.38985646e-01 2.02678338e-01 -5.38699687e-01 8.64248425e-02
4.22837228e-01 -2.08356917e-01 -7.28775486e-02 1.25198555e+00
5.12717426e-01 4.96158808e-01 6.19438231e-01 -1.87564945e+00
-6.72832906e-01 3.63333046e-01 -3.19627702e-01 -3.31032932e-01
4.81854200e-01 -1.60699159e-01 7.26638973e-01 -1.17721713e+00
5.63170731e-01 1.56598687e+00 5.11582732e-01 -8.79914984e-02
-9.94225860e-01 -6.20353401e-01 2.08377987e-01 6.71205223e-01
-1.60804236e+00 -2.57422537e-01 6.93538189e-01 -4.74908680e-01
7.41196156e-01 2.31011912e-01 5.15015662e-01 5.78703761e-01
2.49945924e-01 5.59257746e-01 9.72460747e-01 -2.99413264e-01
5.13771832e-01 2.48596355e-01 1.52459919e-01 5.82956135e-01
6.73369169e-01 -1.45091815e-02 -7.58136332e-01 -3.02004904e-01
5.60542405e-01 2.33595610e-01 1.48710236e-01 -1.32447588e+00
-1.63543212e+00 6.18896604e-01 6.36612117e-01 -8.77286419e-02
-5.27574062e-01 1.20236188e-01 2.74490297e-01 -3.43510062e-02
-3.18109728e-02 5.33403277e-01 -1.74122825e-01 7.53448009e-02
-3.48864853e-01 4.45070297e-01 5.84138393e-01 1.73569262e+00
1.20480549e+00 -3.80903661e-01 5.84767498e-02 7.88467348e-01
5.33591866e-01 8.58397126e-01 3.88651520e-01 -1.00258493e+00
1.74265653e-01 8.62013042e-01 3.44795197e-01 -1.43070078e+00
-5.65507770e-01 -2.42260899e-02 -4.32417333e-01 2.31402710e-01
-1.53669879e-01 8.13911617e-01 -8.63370299e-01 1.33229887e+00
4.14405257e-01 1.29424155e-01 3.08186840e-02 1.05986929e+00
5.93688309e-01 5.41986013e-03 -6.99426904e-02 -1.80641897e-02
1.27675271e+00 -8.74583066e-01 -7.07930446e-01 -1.57434270e-01
6.51538193e-01 -5.89367628e-01 9.13754404e-01 4.59342629e-01
-5.84864259e-01 -5.39641023e-01 -1.06740320e+00 -1.98885366e-01
-6.05475426e-01 1.58389777e-01 7.67428041e-01 2.09638506e-01
-7.34801769e-01 2.39292294e-01 -7.13675916e-01 -8.11341643e-01
2.71306187e-01 4.80953872e-01 -7.18180060e-01 -2.98015386e-01
-5.35994768e-01 1.28805435e+00 1.21924043e+00 2.30345026e-01
-8.31296265e-01 -2.82611579e-01 -9.37581837e-01 -4.24221873e-01
4.50373620e-01 -4.20026511e-01 7.49011040e-01 -1.22215385e-02
-1.21513700e+00 7.50800550e-01 -1.53181717e-01 -2.77576774e-01
3.99397284e-01 -4.42416936e-01 -2.94689626e-01 -1.20333552e-01
3.97981435e-01 6.62564754e-01 4.09025192e-01 -1.46929240e+00
-8.20400596e-01 -8.23468387e-01 -2.00016424e-01 6.88331485e-01
1.83619151e-03 -1.03871502e-01 -3.56459707e-01 -1.95688948e-01
1.23932266e+00 -1.02250004e+00 -1.49501130e-01 2.73363173e-01
-3.57660651e-01 -3.99289787e-01 1.06007755e+00 -4.78720278e-01
4.49378580e-01 -2.17637753e+00 2.02115804e-01 3.18012595e-01
-3.26548852e-02 -4.57515389e-01 -8.91206786e-02 3.43132585e-01
4.70248818e-01 -3.75032485e-01 -1.49058342e-01 -3.84551913e-01
2.91521937e-01 4.85501438e-01 -3.92780960e-01 8.30930650e-01
-1.64236590e-01 8.61562729e-01 -9.90571916e-01 -6.91470563e-01
5.87452233e-01 -5.76535836e-02 -4.06908035e-01 2.76159972e-01
-8.77890140e-02 4.37786669e-01 -4.53142017e-01 1.15583539e+00
1.00318038e+00 3.35496277e-01 -2.40200665e-02 -3.72332871e-01
-4.23266917e-01 -1.60670448e-02 -1.52729511e+00 2.31604743e+00
-2.94972599e-01 2.52330899e-01 -8.36187452e-02 -1.02649093e+00
1.63390744e+00 -3.05389345e-01 5.47895372e-01 -5.92270494e-01
4.59431000e-02 7.08364367e-01 -1.44833848e-01 -5.23286223e-01
9.45560873e-01 5.93480229e-01 -1.84528753e-01 3.01887020e-02
7.02382624e-02 -5.98722517e-01 -8.06188360e-02 -1.17025129e-01
6.97491527e-01 5.21688759e-01 4.90892380e-01 -5.68716228e-01
2.64671594e-01 3.68430674e-01 5.24646461e-01 8.94485116e-01
-2.69851565e-01 4.00937885e-01 -3.05674314e-01 -3.49167287e-01
-9.42601740e-01 -1.39351821e+00 -3.93460572e-01 1.01064837e+00
1.15511835e+00 -6.34753332e-02 1.92436308e-03 -2.57419616e-01
5.43093622e-01 8.10011327e-01 -4.27684039e-01 -2.57386118e-01
-5.13806343e-01 -4.73949194e-01 6.20932057e-02 3.83318394e-01
5.71766615e-01 -1.19741893e+00 -7.70850658e-01 1.75738692e-01
-1.92706883e-01 -9.72071767e-01 9.36238691e-02 2.14690745e-01
-6.22239172e-01 -1.05488908e+00 -4.68511730e-02 -8.60873759e-01
8.11667681e-01 6.36958718e-01 4.02982056e-01 -4.34395939e-01
-4.64251697e-01 4.28547353e-01 -6.07087195e-01 -7.67745733e-01
1.89186893e-02 -1.87061191e-01 8.29655588e-01 -9.13431421e-02
4.64687586e-01 -5.15688479e-01 -4.17080939e-01 7.83790767e-01
-3.86477649e-01 -1.17284387e-01 8.79318655e-01 5.51999569e-01
7.71550775e-01 -9.09066573e-02 5.24387062e-01 -1.40444532e-01
2.52356976e-01 -5.28021157e-01 -6.62903905e-01 4.66401070e-01
-6.46504998e-01 -1.00364141e-01 -1.77730709e-01 -5.83545268e-01
-7.42059350e-01 2.34629303e-01 6.82191074e-01 -4.20842081e-01
-1.30372748e-01 3.40931058e-01 -4.94020462e-01 -3.97247434e-01
5.98078847e-01 2.91532248e-01 1.39819726e-01 -5.39343357e-01
5.05663514e-01 8.31023693e-01 8.88926446e-01 -7.75177240e-01
7.49868751e-01 6.88190341e-01 8.91497880e-02 -5.55833697e-01
-3.86994004e-01 -7.34608650e-01 -9.71450806e-01 -2.46547431e-01
5.99752784e-01 -9.05836642e-01 -8.65571678e-01 3.02526414e-01
-1.14704299e+00 1.19783543e-01 -1.88664392e-01 1.15401149e+00
-1.12805498e+00 3.95611286e-01 2.88146406e-01 -9.72649395e-01
2.98621766e-02 -1.25393629e+00 1.14828503e+00 4.01500277e-02
-5.85485324e-02 -2.86155075e-01 -7.61892423e-02 1.63014531e-01
2.37198278e-01 2.01252908e-01 5.09979606e-01 -9.60204363e-01
-9.44095016e-01 -1.83007330e-01 -5.46680093e-01 -2.83617586e-01
2.61360258e-01 -3.67421061e-01 -5.59686720e-01 -4.86433893e-01
-1.46096528e-01 -2.81622112e-01 4.59546506e-01 9.58495662e-02
9.97062802e-01 5.38705848e-02 -6.87258661e-01 6.78278685e-01
1.23723471e+00 3.33666652e-01 2.77539730e-01 6.84887707e-01
6.48461461e-01 8.41050327e-01 1.59315884e+00 5.84685504e-01
5.11416495e-01 9.21995997e-01 9.66006219e-01 4.82894659e-01
3.50671202e-01 -2.39410684e-01 -2.87530310e-02 5.86206555e-01
1.92613363e-01 -6.42877966e-02 -1.00426006e+00 5.25366902e-01
-2.33879638e+00 -5.26641369e-01 -8.91890079e-02 2.26657891e+00
2.16321379e-01 -1.34416535e-01 -3.16644043e-01 -2.32189149e-01
9.42956567e-01 -3.60256046e-01 -8.26314569e-01 1.04689404e-01
-2.31267095e-01 -5.89026928e-01 6.94407701e-01 3.59196901e-01
-9.38439310e-01 9.54118609e-01 5.41488886e+00 5.98868489e-01
-8.34853113e-01 8.28337893e-02 -4.95801717e-01 3.04881364e-01
-9.26039070e-02 3.18960667e-01 -8.46771896e-01 3.20343196e-01
1.92173094e-01 -2.44279817e-01 3.49110395e-01 1.25624597e+00
-1.91699155e-02 -4.90627676e-01 -1.15232265e+00 1.28722894e+00
3.23857695e-01 -8.78816545e-01 2.13099476e-02 1.75091431e-01
4.54610705e-01 8.02498311e-02 -1.47113934e-01 1.73243165e-01
2.66073614e-01 -7.91542351e-01 1.18081737e+00 7.85138667e-01
4.90465432e-01 -4.03760642e-01 6.21616840e-01 4.96430606e-01
-1.10985112e+00 -3.24284405e-01 -7.62275994e-01 5.90474755e-02
1.64651021e-01 2.73554772e-01 -1.20121300e+00 9.78468060e-01
1.00402510e+00 6.17064059e-01 -5.95273256e-01 1.44305670e+00
2.17575505e-01 -3.43565613e-01 -6.30850494e-01 -1.16831139e-01
-2.02006400e-01 -3.67033899e-01 7.50534236e-01 6.83848619e-01
6.23205602e-01 -2.13432729e-01 7.12880611e-01 8.30854058e-01
5.08687973e-01 2.95372248e-01 -5.88336766e-01 2.97482789e-01
1.11370826e+00 1.09469128e+00 -9.69483376e-01 -2.46085599e-01
-5.18214889e-02 7.29339063e-01 3.60121667e-01 9.88296643e-02
-6.90908849e-01 -3.69085342e-01 4.80522841e-01 -1.79077134e-01
-8.47120881e-02 -7.91997135e-01 -3.74733865e-01 -8.57463896e-01
2.46587113e-01 -2.90916324e-01 1.34854332e-01 -9.13020611e-01
-1.01782358e+00 3.42537194e-01 4.63031322e-01 -1.29218841e+00
-5.26520386e-02 -6.51548982e-01 -1.41329378e-01 7.86658227e-01
-1.31392527e+00 -1.22275245e+00 -1.08545172e+00 4.20815408e-01
3.88251424e-01 -3.02985251e-01 7.32467175e-01 8.55784044e-02
-3.96912638e-03 1.00952148e-01 2.29153261e-01 -3.32020432e-01
9.52402234e-01 -9.21949267e-01 6.26604035e-02 6.18598700e-01
-1.46619007e-02 7.47627139e-01 8.60411525e-01 -1.08735979e+00
-1.78323507e+00 -1.07760274e+00 3.54742497e-01 -5.46462476e-01
4.82081652e-01 -6.43191040e-01 -8.57054949e-01 7.51243293e-01
-6.58147097e-01 1.48661911e-01 8.14858451e-02 7.47974589e-02
-2.83115178e-01 -1.59277946e-01 -1.27716148e+00 3.72704297e-01
1.40772355e+00 -3.42083037e-01 -7.08028793e-01 5.06605685e-01
7.63489425e-01 -4.62967813e-01 -7.53934503e-01 9.92487550e-01
7.53407001e-01 -8.17414820e-01 9.32997584e-01 -2.83531874e-01
-5.42312622e-01 -7.57814765e-01 -9.28807795e-01 -9.27453995e-01
-3.98051977e-01 -6.84974492e-02 6.86887130e-02 1.14216423e+00
-2.15824872e-01 -8.79870117e-01 7.19131649e-01 2.63159394e-01
-5.04043818e-01 -2.79745907e-01 -1.10769248e+00 -1.23636329e+00
-6.75877154e-01 -3.60505015e-01 6.68636441e-01 7.71873176e-01
-2.38933906e-01 -3.54446024e-01 -1.61439210e-01 7.12775171e-01
9.19819832e-01 3.86505008e-01 1.18730319e+00 -1.53356767e+00
3.29482704e-01 -2.50441462e-01 -1.11597252e+00 -8.07227015e-01
2.94123799e-01 -9.28617537e-01 6.14272475e-01 -1.56733060e+00
2.86107659e-01 -1.03670204e+00 -2.47305483e-01 2.15143427e-01
2.09766388e-01 2.69948363e-01 1.31121680e-01 7.37134814e-01
-8.27232420e-01 7.64439583e-01 7.42667079e-01 4.10615327e-03
-2.14636371e-01 -2.83448547e-01 -9.56763923e-02 6.21209741e-01
8.16832423e-01 -2.61138499e-01 -2.91910082e-01 -4.03685689e-01
-3.54919955e-03 -2.79839396e-01 7.65842855e-01 -1.13475871e+00
5.32760620e-01 -5.01416266e-01 7.85972849e-02 -1.11819971e+00
7.30578005e-01 -1.13335598e+00 3.35570395e-01 5.63632250e-01
-2.36821130e-01 2.19455324e-02 -1.42930731e-01 7.72991836e-01
-3.84123251e-02 -3.56616080e-01 5.80499291e-01 4.38975655e-02
-1.41510594e+00 4.62501287e-01 2.57937759e-01 -4.21467394e-01
1.41362262e+00 -5.90217948e-01 -2.49374717e-01 -3.41483317e-02
-5.73732793e-01 3.68088692e-01 7.47342169e-01 9.19082463e-01
8.35651934e-01 -1.47990155e+00 -6.66062772e-01 3.14505368e-01
8.34043503e-01 2.52558589e-01 1.88363954e-01 9.41237211e-01
-5.92397988e-01 3.81257862e-01 -6.11837327e-01 -1.20272064e+00
-1.00655818e+00 6.97340667e-01 -1.38391018e-01 7.15328217e-01
-5.12911141e-01 5.50743580e-01 1.91694319e-01 -9.80896175e-01
3.13677430e-01 -2.48023197e-01 1.30310446e-01 -1.65599048e-01
1.40834033e-01 5.89672267e-01 -9.37677324e-02 -8.40087235e-01
-8.01477671e-01 8.02649617e-01 2.14514181e-01 2.06193738e-02
9.74530399e-01 -5.87178826e-01 -4.60236073e-01 7.21630275e-01
8.92629325e-01 -1.49638355e-01 -1.02935135e+00 -5.20405054e-01
4.52948898e-01 -7.31986701e-01 -4.16699350e-01 -5.20584345e-01
1.52342869e-02 4.84441251e-01 8.09560120e-01 -1.84625506e-01
6.04844809e-01 2.78374374e-01 2.69963387e-02 9.03399467e-01
1.22974980e+00 -1.14805150e+00 7.98622891e-02 5.07482767e-01
1.23568070e+00 -1.55897713e+00 3.21930468e-01 -6.73101783e-01
-3.53539199e-01 9.41957235e-01 9.18995798e-01 -1.24967434e-01
4.14185762e-01 -4.23444621e-02 -1.38236806e-01 -6.45122454e-02
-1.26021624e-01 -1.42377585e-01 2.43114322e-01 8.93198431e-01
-4.43351448e-01 1.79174751e-01 -4.90465946e-02 1.00338809e-01
-4.85913366e-01 -2.78588474e-01 5.27516007e-02 1.11495352e+00
-1.11605155e+00 -4.79741335e-01 -5.14394224e-01 2.61095554e-01
4.99707162e-01 3.45170587e-01 1.52377989e-02 8.13441992e-01
1.53749242e-01 7.98546851e-01 3.32172185e-01 -4.06311780e-01
3.14850748e-01 -2.20575362e-01 6.47342324e-01 -5.37870049e-01
3.79223078e-01 -4.32995379e-01 -2.69096673e-01 -8.12361896e-01
-3.89263362e-01 -6.80216193e-01 -1.41877007e+00 4.72995527e-02
-6.51610255e-01 1.61757469e-01 1.40946269e+00 8.01757693e-01
5.25882125e-01 9.46970135e-02 3.48857075e-01 -1.14360833e+00
-7.30356693e-01 -1.05722570e+00 -7.19257653e-01 2.58998841e-01
-3.89245339e-02 -1.40007687e+00 -2.04619840e-01 -4.25441533e-01] | [7.230001449584961, -2.181178092956543] |
5906942b-1741-4d51-90d2-35318357ccdc | beyond-low-frequency-information-in-graph | 2101.00797 | null | https://arxiv.org/abs/2101.00797v1 | https://arxiv.org/pdf/2101.00797v1.pdf | Beyond Low-frequency Information in Graph Convolutional Networks | Graph neural networks (GNNs) have been proven to be effective in various network-related tasks. Most existing GNNs usually exploit the low-frequency signals of node features, which gives rise to one fundamental question: is the low-frequency information all we need in the real world applications? In this paper, we first present an experimental investigation assessing the roles of low-frequency and high-frequency signals, where the results clearly show that exploring low-frequency signal only is distant from learning an effective node representation in different scenarios. How can we adaptively learn more information beyond low-frequency information in GNNs? A well-informed answer can help GNNs enhance the adaptability. We tackle this challenge and propose a novel Frequency Adaptation Graph Convolutional Networks (FAGCN) with a self-gating mechanism, which can adaptively integrate different signals in the process of message passing. For a deeper understanding, we theoretically analyze the roles of low-frequency signals and high-frequency signals on learning node representations, which further explains why FAGCN can perform well on different types of networks. Extensive experiments on six real-world networks validate that FAGCN not only alleviates the over-smoothing problem, but also has advantages over the state-of-the-arts. | ['HuaWei Shen', 'Chuan Shi', 'Xiao Wang', 'Deyu Bo'] | 2021-01-04 | null | null | null | null | ['node-classification-on-non-homophilic'] | ['graphs'] | [-3.56303565e-02 1.37599781e-01 -3.25023770e-01 -1.16234370e-01
1.01945512e-01 -1.16103984e-01 4.33673143e-01 2.80658364e-01
-3.07538241e-01 4.98352319e-01 5.70904575e-02 -2.94120193e-01
-4.84795272e-01 -1.30835402e+00 -6.20309591e-01 -7.96331823e-01
-7.55913496e-01 -5.29121533e-02 6.32678628e-01 -7.43212998e-01
-1.06018983e-01 3.65122855e-01 -1.44200552e+00 1.20611690e-01
9.40247059e-01 8.27970445e-01 1.57787278e-01 4.01209503e-01
-4.38186854e-01 4.92985308e-01 -7.61846364e-01 4.81948629e-02
3.58725153e-02 -4.60489839e-01 -7.06263244e-01 -2.64249593e-01
9.09951106e-02 2.52704501e-01 -6.55293405e-01 1.16032827e+00
4.15710419e-01 1.49436519e-01 4.15813386e-01 -1.27960181e+00
-5.24405837e-01 1.09209549e+00 -4.60173041e-01 6.49485171e-01
4.60223258e-02 1.55074254e-01 9.65027452e-01 -3.20576876e-01
4.15309191e-01 1.36218238e+00 8.65117788e-01 4.46808577e-01
-9.82862592e-01 -7.77859569e-01 5.45569181e-01 2.55581796e-01
-1.33655775e+00 -6.33047596e-02 1.17458570e+00 -4.87411432e-02
6.78443193e-01 1.10448368e-01 6.95910215e-01 1.15779269e+00
3.72245729e-01 4.99347210e-01 5.04726112e-01 -3.06446046e-01
-2.85659209e-02 -2.66320676e-01 1.26295686e-01 6.69542849e-01
4.15921181e-01 2.59946615e-01 -6.85674250e-01 2.57260837e-02
1.00962973e+00 9.14922729e-02 -6.41296983e-01 7.97202718e-03
-1.10738242e+00 9.19042289e-01 1.24909890e+00 9.63589132e-01
-3.17048639e-01 6.26899064e-01 4.05125648e-01 6.32285357e-01
3.21287304e-01 2.02228308e-01 -3.93998295e-01 3.00363123e-01
-3.90221059e-01 -3.11089844e-01 6.54692233e-01 5.07559717e-01
9.37085390e-01 4.87523019e-01 -1.01663433e-01 7.25822031e-01
4.32950646e-01 2.33762294e-01 4.94182169e-01 -2.20432341e-01
1.87163472e-01 6.94212377e-01 -6.65047288e-01 -1.29708707e+00
-8.23429883e-01 -1.12262583e+00 -1.33940125e+00 -2.69088238e-01
4.48237479e-01 -3.64426941e-01 -8.19401026e-01 2.03014064e+00
1.30518124e-01 6.96200013e-01 -1.11494094e-01 7.19771028e-01
1.20295024e+00 6.33851051e-01 3.20476323e-01 -4.49045636e-02
1.29172146e+00 -6.72432125e-01 -7.09687114e-01 -2.90698469e-01
5.48427880e-01 -2.75574505e-01 9.22581971e-01 -1.08901784e-01
-4.19647157e-01 -8.34462583e-01 -1.20605075e+00 4.39608008e-01
-4.40608442e-01 -1.72464579e-01 1.13763845e+00 7.64589369e-01
-1.32247198e+00 8.93123329e-01 -6.59063041e-01 -5.73727906e-01
3.82675469e-01 4.42037612e-01 -2.44611979e-01 1.20400846e-01
-1.97429979e+00 4.27692205e-01 5.53774297e-01 1.24408886e-01
-6.51187837e-01 -7.65233517e-01 -8.15269232e-01 4.95216787e-01
5.11006534e-01 -4.56298143e-01 7.16899216e-01 -1.03266811e+00
-1.33928955e+00 2.60061055e-01 2.71692097e-01 -4.39152002e-01
1.40974447e-02 2.35635296e-01 -8.84041309e-01 3.76582265e-01
-2.72420764e-01 5.64079821e-01 9.22731340e-01 -9.49705064e-01
-3.33674222e-01 -7.47040957e-02 1.81036517e-01 -1.31762117e-01
-8.68340611e-01 -4.51498181e-01 -3.93943727e-01 -8.53137672e-01
2.46328473e-01 -5.95988214e-01 -2.27183998e-01 -7.54031688e-02
-1.25075340e-01 -4.87960756e-01 8.09308231e-01 -1.52688190e-01
1.41951060e+00 -2.19596171e+00 -2.04971388e-01 4.77065772e-01
5.26559651e-01 3.89384598e-01 -4.06849802e-01 5.95337987e-01
-1.86885875e-02 2.44876459e-01 1.17279842e-01 4.33863699e-01
-2.82647222e-01 1.82985097e-01 -7.35049918e-02 2.47462586e-01
4.05669242e-01 1.03677142e+00 -1.14646924e+00 -2.09394962e-01
1.28813416e-01 9.61196482e-01 -3.75245392e-01 -1.79231837e-01
-1.81329831e-01 3.53101403e-01 -6.42672539e-01 3.75488460e-01
6.23519540e-01 -7.25771487e-01 3.85652333e-01 -3.72942716e-01
3.80556673e-01 1.20274603e-01 -1.08189154e+00 1.42179358e+00
-1.83803767e-01 6.79969549e-01 1.88259333e-01 -1.28634477e+00
9.79797125e-01 1.37497187e-01 5.17396450e-01 -8.15035701e-01
1.64295197e-01 6.95741400e-02 4.40081149e-01 -2.40659580e-01
8.24087579e-03 -3.64081226e-02 1.19076937e-01 1.85793310e-01
4.28900898e-01 3.89552444e-01 2.08653465e-01 4.19068724e-01
1.25890136e+00 -5.88150442e-01 9.73240137e-02 -5.60912371e-01
6.04961812e-01 -5.53902388e-01 5.06859601e-01 8.95763695e-01
-3.13294739e-01 -1.51496138e-02 6.29953623e-01 -4.17504400e-01
-1.48241773e-01 -8.60403478e-01 2.54479479e-02 1.47287643e+00
5.36564171e-01 -6.84660196e-01 -5.27961314e-01 -7.94895828e-01
2.01942056e-01 2.03483626e-01 -7.15537786e-01 -6.91788435e-01
-5.11007249e-01 -1.11622930e+00 5.50619304e-01 5.13084292e-01
6.71109021e-01 -1.09819436e+00 -3.82462144e-01 3.68697882e-01
3.06226388e-02 -9.54083979e-01 -3.45976174e-01 3.04099768e-01
-9.07963336e-01 -9.54691589e-01 -5.31334877e-01 -7.22224414e-01
5.91742814e-01 6.56489193e-01 1.37659776e+00 8.03998470e-01
-1.18171670e-01 2.28923187e-01 -5.02724886e-01 -2.82880217e-01
-3.46997887e-01 5.75088561e-01 -1.55066445e-01 1.33172199e-01
3.52429122e-01 -1.04681957e+00 -5.88501811e-01 2.65175283e-01
-9.81066942e-01 -2.88737357e-01 9.54446971e-01 8.55261683e-01
9.80554521e-02 6.58737898e-01 1.14736736e+00 -1.18403625e+00
7.95107245e-01 -6.00567639e-01 -2.96403795e-01 1.42463520e-01
-3.86408567e-01 2.56928086e-01 9.04656589e-01 -5.92923224e-01
-6.66714549e-01 -2.92336315e-01 -2.82771081e-01 -1.59321278e-01
1.35847554e-01 8.41907680e-01 -1.07898697e-01 -5.23542225e-01
9.01763082e-01 2.29319576e-02 -3.16483676e-02 -2.07240537e-01
1.87887847e-01 1.38728665e-02 2.63422638e-01 -5.47576606e-01
1.02701592e+00 4.19285059e-01 1.88232869e-01 -1.15374339e+00
-6.34308219e-01 -3.16285461e-01 -2.82445461e-01 -4.82448995e-01
4.36266452e-01 -8.02550912e-01 -8.03141773e-01 3.88987452e-01
-7.33896792e-01 -4.08780038e-01 -1.83549270e-01 3.16408485e-01
9.93468538e-02 2.97243565e-01 -6.85013831e-01 -6.01192772e-01
-3.68954122e-01 -8.24133992e-01 5.42908907e-01 7.21189439e-01
2.60382622e-01 -1.38741481e+00 -3.81321192e-01 -4.16492045e-01
1.04987276e+00 2.21581534e-01 1.03857613e+00 -6.45087361e-01
-6.28618836e-01 1.19343050e-01 -2.61804372e-01 -1.05907969e-01
7.33378902e-02 -3.13407294e-02 -8.88169706e-01 -6.02283299e-01
-2.23885223e-01 -8.83233473e-02 1.33464658e+00 4.79144782e-01
1.27375495e+00 -4.41210978e-02 -5.56550264e-01 6.93216920e-01
1.52865911e+00 -1.59617960e-01 5.07500827e-01 -6.33011535e-02
7.10756063e-01 4.02250469e-01 2.28999462e-02 1.64464921e-01
2.71969318e-01 3.56378227e-01 7.14410305e-01 -2.96067268e-01
-2.65839845e-01 -2.21045181e-01 2.43040323e-01 9.82820034e-01
5.79961799e-02 -5.47529399e-01 -8.29184294e-01 4.55436707e-01
-1.81513107e+00 -8.02980304e-01 -8.24977458e-03 1.74218655e+00
6.75850034e-01 4.83212084e-01 2.23435294e-02 3.06885004e-01
8.16605210e-01 6.38221860e-01 -4.03455406e-01 1.21571280e-01
-1.45890757e-01 3.02795589e-01 4.96634662e-01 1.77202731e-01
-1.15153730e+00 7.24372387e-01 6.08117342e+00 1.14198375e+00
-1.46258235e+00 -1.34010226e-01 5.58660686e-01 3.84033322e-01
-5.00437140e-01 -2.22402498e-01 -5.99506676e-01 4.56879854e-01
1.17062116e+00 -1.54092520e-01 4.09969211e-01 6.78393006e-01
-8.08197334e-02 2.70686805e-01 -8.91501546e-01 7.02316523e-01
-2.12214500e-01 -1.52155674e+00 1.13559969e-01 -1.12753086e-01
5.64829230e-01 2.59230524e-01 -2.52656210e-02 6.04705453e-01
4.42200899e-01 -1.26898897e+00 1.92726195e-01 3.92522126e-01
6.02574646e-01 -7.17332363e-01 8.23985934e-01 3.07922900e-01
-1.83448088e+00 -7.52457455e-02 -5.18793404e-01 9.20200422e-02
-2.85962313e-01 9.78380740e-01 -6.61080122e-01 9.98261690e-01
6.87551558e-01 9.80938673e-01 -6.91327274e-01 9.89913762e-01
-2.87252426e-01 8.50004911e-01 -5.15725911e-01 -2.92157680e-01
2.90202111e-01 1.50967643e-01 3.83036375e-01 1.15157235e+00
2.40437344e-01 -3.94093012e-03 2.74343342e-01 7.95646250e-01
-3.31225961e-01 -7.00312406e-02 -3.40966165e-01 -1.72701195e-01
6.70563817e-01 1.32591796e+00 -1.16557896e+00 -5.83948977e-02
-3.68050963e-01 4.46762681e-01 4.32178855e-01 5.77880204e-01
-5.87820351e-01 -6.20434701e-01 4.54095691e-01 1.97772712e-01
2.29340419e-01 -1.25326768e-01 1.73080936e-01 -1.01022899e+00
-2.14895144e-01 -6.09641373e-01 6.52825356e-01 -1.95073724e-01
-1.46628797e+00 6.82621896e-01 -3.21936846e-01 -9.91734743e-01
8.70084111e-03 -4.18036312e-01 -9.92554605e-01 6.01983190e-01
-1.87604690e+00 -8.76124740e-01 -6.84298098e-01 5.18433034e-01
2.79148281e-01 -2.83570145e-03 6.16556585e-01 4.20461148e-01
-5.12513757e-01 7.88503766e-01 -1.34776220e-01 4.10580546e-01
4.73807991e-01 -1.04447865e+00 4.36817408e-01 7.95114636e-01
2.29662046e-01 7.01699972e-01 4.89838719e-01 -5.81001699e-01
-1.60501122e+00 -1.19745493e+00 4.31850344e-01 1.57614991e-01
5.68402112e-01 -3.33736002e-01 -1.24876904e+00 3.11452180e-01
6.29789531e-02 4.11609322e-01 4.30135339e-01 3.18841666e-01
-3.08890194e-01 -3.11510235e-01 -9.02784050e-01 4.55420107e-01
1.25203133e+00 -4.56517786e-01 -1.10767543e-01 2.22743586e-01
9.52716887e-01 -1.64229602e-01 -7.11440861e-01 6.84629977e-01
8.14560726e-02 -9.13504004e-01 1.11989689e+00 -4.72628325e-01
9.54274610e-02 -2.97558337e-01 3.19900602e-01 -1.57291329e+00
-5.06501079e-01 -6.83215320e-01 -2.42088467e-01 1.02179217e+00
1.61329776e-01 -9.00821626e-01 8.10049176e-01 -2.82411277e-01
1.09525681e-01 -8.20751607e-01 -9.25464869e-01 -7.11558759e-01
-8.13951939e-02 -3.20152521e-01 6.17094159e-01 1.06871378e+00
4.67886403e-02 4.99821991e-01 -1.60151795e-01 1.73418984e-01
4.53983217e-01 -9.48798135e-02 3.62304956e-01 -1.80510104e+00
-1.69266462e-01 -6.81060433e-01 -6.48337901e-01 -1.10856628e+00
2.21371397e-01 -1.06432939e+00 -2.54539132e-01 -1.44960487e+00
-1.77454934e-01 -6.02215052e-01 -7.21216261e-01 3.46522391e-01
-5.44736922e-01 5.81567138e-02 6.02592975e-02 -1.27796486e-01
-6.80697322e-01 6.48761213e-01 1.26394939e+00 -2.12857977e-01
-8.31048936e-02 -3.81775871e-02 -9.03179705e-01 5.24893701e-01
6.81742251e-01 -3.03638607e-01 -8.21805120e-01 -2.50636131e-01
3.79664063e-01 -1.02780126e-01 2.33431652e-01 -1.23157501e+00
3.46854866e-01 -1.25813978e-02 4.65478390e-01 -3.38584423e-01
-8.99843350e-02 -8.64329398e-01 8.51337165e-02 7.73115218e-01
-2.44006932e-01 -2.00613067e-01 2.83131987e-01 1.04101384e+00
-1.56167656e-01 4.43802401e-02 7.35555172e-01 -8.47751424e-02
-9.34833229e-01 4.71202284e-01 -4.35113937e-01 2.10763142e-01
7.09633589e-01 3.57111543e-02 -5.01943171e-01 -6.19652331e-01
-5.08895934e-01 4.31523383e-01 2.70500388e-02 5.27452767e-01
3.53625208e-01 -1.33333409e+00 -6.95873082e-01 6.35080218e-01
-3.34434323e-02 -2.20940083e-01 4.36757714e-01 6.01188362e-01
-2.29842618e-01 2.72837996e-01 -1.62974849e-01 -8.02241147e-01
-6.80500269e-01 4.59972084e-01 5.15605390e-01 -4.19122696e-01
-6.61457241e-01 1.11825848e+00 2.51408190e-01 -3.42598051e-01
2.79340446e-01 -3.84723783e-01 -5.87655783e-01 2.35352814e-02
3.40975136e-01 6.86736181e-02 1.45755842e-01 -4.07600641e-01
-4.13464308e-01 5.40915668e-01 -5.18406853e-02 5.18844843e-01
1.37314260e+00 8.01571161e-02 5.18121049e-02 2.16679290e-01
1.02056921e+00 -2.95481056e-01 -1.01229119e+00 -5.34853101e-01
1.49777979e-01 -1.63617581e-01 1.33191213e-01 -3.63102168e-01
-1.81822443e+00 8.31150532e-01 4.10551041e-01 8.21513712e-01
1.33249974e+00 1.39812153e-04 7.25660443e-01 3.22386116e-01
4.16231573e-01 -9.30835545e-01 4.12966609e-01 6.84010267e-01
6.13051653e-01 -1.05613005e+00 -1.19837232e-01 -8.34589660e-01
-9.93489176e-02 1.33366370e+00 7.47159004e-01 -2.24654734e-01
1.13945973e+00 9.07350853e-02 -2.77454350e-02 -4.76706177e-01
-8.90873849e-01 -3.13733250e-01 3.34846079e-01 6.59200907e-01
4.21796769e-01 -4.66989428e-02 -1.86131343e-01 5.87532222e-01
-5.56908883e-02 -3.72064054e-01 3.00570041e-01 6.56427562e-01
-5.91034412e-01 -9.67876315e-01 -4.60507944e-02 6.75062776e-01
-3.14898729e-01 3.48145701e-02 -2.75869220e-01 9.01954770e-01
-4.11101384e-03 9.45111811e-01 -1.07540153e-02 -6.62307739e-01
1.27693757e-01 -1.71276703e-01 1.63121387e-01 -4.99583125e-01
-8.11317205e-01 1.00803256e-01 -1.17762320e-01 -4.79561269e-01
-4.49211091e-01 1.10150822e-01 -1.48799217e+00 -4.88884866e-01
-5.26685655e-01 3.20471138e-01 2.48758584e-01 6.51682079e-01
4.61367279e-01 1.33664107e+00 5.57622313e-01 -7.09418178e-01
-4.22251821e-01 -9.78565991e-01 -4.96582687e-01 3.36387217e-01
4.46186066e-01 -7.44621873e-01 -8.35287452e-01 -5.52605987e-01] | [7.075004577636719, 6.184177875518799] |
fe0e89c4-1912-48d8-94c1-6413738c60a7 | compress-self-supervised-learning-by | 2010.14713 | null | https://arxiv.org/abs/2010.14713v1 | https://arxiv.org/pdf/2010.14713v1.pdf | CompRess: Self-Supervised Learning by Compressing Representations | Self-supervised learning aims to learn good representations with unlabeled data. Recent works have shown that larger models benefit more from self-supervised learning than smaller models. As a result, the gap between supervised and self-supervised learning has been greatly reduced for larger models. In this work, instead of designing a new pseudo task for self-supervised learning, we develop a model compression method to compress an already learned, deep self-supervised model (teacher) to a smaller one (student). We train the student model so that it mimics the relative similarity between the data points in the teacher's embedding space. For AlexNet, our method outperforms all previous methods including the fully supervised model on ImageNet linear evaluation (59.0% compared to 56.5%) and on nearest neighbor evaluation (50.7% compared to 41.4%). To the best of our knowledge, this is the first time a self-supervised AlexNet has outperformed supervised one on ImageNet classification. Our code is available here: https://github.com/UMBCvision/CompRess | ['Hamed Pirsiavash', 'Ajinkya Tejankar', 'Soroush Abbasi Koohpayegani'] | 2020-10-28 | null | http://proceedings.neurips.cc/paper/2020/hash/975a1c8b9aee1c48d32e13ec30be7905-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/975a1c8b9aee1c48d32e13ec30be7905-Paper.pdf | neurips-2020-12 | ['self-supervised-image-classification'] | ['computer-vision'] | [ 1.08444266e-01 4.66505349e-01 -5.95775306e-01 -7.85783350e-01
-4.48511422e-01 -2.32098550e-01 4.15388465e-01 2.48416692e-01
-6.23836339e-01 6.40217245e-01 2.34561577e-01 -1.15340114e-01
1.91709161e-01 -7.11462259e-01 -8.10372651e-01 -4.66126323e-01
8.75843838e-02 6.41138554e-01 -1.64766274e-02 5.58949746e-02
4.11833599e-02 3.00211996e-01 -1.59937727e+00 2.24861220e-01
8.13765407e-01 9.67941701e-01 2.64882565e-01 3.51244956e-01
-9.38380063e-02 1.00197327e+00 -3.06130826e-01 -2.08691120e-01
4.07378078e-01 -4.98968542e-01 -1.04454029e+00 1.25638962e-01
9.27165806e-01 -5.00462949e-01 -6.67856455e-01 1.03921449e+00
4.27299231e-01 9.29954872e-02 7.44385839e-01 -1.30635726e+00
-7.14452863e-01 9.12885308e-01 -4.02237654e-01 -4.58190255e-02
-1.83983296e-01 -7.11728632e-02 9.03867304e-01 -8.60735774e-01
5.36211848e-01 1.08833098e+00 6.32907689e-01 7.09529638e-01
-1.19118202e+00 -8.59320700e-01 -2.71001637e-01 2.32212514e-01
-1.34004426e+00 -5.53951621e-01 6.63119316e-01 -1.23882204e-01
8.12261343e-01 -2.19712943e-01 6.91643298e-01 8.74163508e-01
-3.10285360e-01 1.05626452e+00 1.28720975e+00 -6.19140327e-01
2.70577669e-01 4.60386872e-01 3.24735135e-01 8.77061844e-01
1.89930037e-01 2.62679547e-01 -4.32621896e-01 -1.64490622e-02
5.15682042e-01 3.20130944e-01 -7.78053794e-03 -6.32561088e-01
-9.61516857e-01 9.64309275e-01 7.56696403e-01 3.47787321e-01
-4.44365740e-02 2.28062138e-01 3.76921982e-01 5.78627169e-01
6.30428135e-01 3.71819079e-01 -5.35524189e-01 -1.30604342e-01
-1.06310928e+00 -2.60240763e-01 9.38491285e-01 1.08524656e+00
1.17763567e+00 2.10114688e-01 3.28980654e-01 1.02898979e+00
2.05907673e-01 3.99905503e-01 1.07759261e+00 -8.70473504e-01
2.55533040e-01 5.77074766e-01 -5.26860237e-01 -5.63739061e-01
-1.27676249e-01 -8.39155734e-01 -9.22627628e-01 1.03951961e-01
3.51539761e-01 -5.78036755e-02 -8.22919130e-01 1.48595750e+00
9.32906792e-02 4.43947941e-01 3.37851375e-01 4.72907841e-01
9.29545760e-01 6.17125213e-01 -2.53940164e-03 9.39896330e-03
7.62936950e-01 -1.53601813e+00 -4.17911589e-01 -3.69669616e-01
1.14122152e+00 -6.38727784e-01 1.10146308e+00 2.76753426e-01
-9.12409663e-01 -8.93671691e-01 -1.16237772e+00 -2.04207171e-02
-3.69708031e-01 3.14893007e-01 6.42235398e-01 5.57256281e-01
-1.17237413e+00 8.80020022e-01 -7.54231334e-01 -4.19052660e-01
9.05863583e-01 3.13153476e-01 -5.39300859e-01 -2.81456053e-01
-9.01933253e-01 9.22870696e-01 5.87256074e-01 -6.05634511e-01
-1.05179930e+00 -7.68588364e-01 -8.79541099e-01 1.27780303e-01
1.66566968e-01 -1.77841276e-01 1.39913404e+00 -1.31608558e+00
-1.31445014e+00 1.04108822e+00 -1.44506037e-01 -9.24614787e-01
2.69520074e-01 -1.75759941e-01 -6.90742359e-02 3.08808893e-01
-6.90550357e-02 1.25359833e+00 9.75165963e-01 -1.20219851e+00
-4.08267945e-01 -3.43628913e-01 -1.26317032e-02 2.54524350e-01
-9.06370819e-01 -3.51810247e-01 -2.11207345e-01 -5.93190491e-01
2.22630814e-01 -1.03672051e+00 -2.56143898e-01 2.41436437e-01
-1.45077094e-01 -4.59489673e-01 1.05681574e+00 -2.44810119e-01
8.74704301e-01 -2.33776259e+00 -2.12217063e-01 -9.06750280e-03
5.02248704e-01 5.84722579e-01 -3.19394797e-01 2.98808813e-01
-3.37490022e-01 -3.59491183e-04 -3.00414562e-01 -7.42961943e-01
-2.69120663e-01 5.15968025e-01 -2.27413192e-01 4.45822030e-01
-8.36744905e-02 9.24111009e-01 -1.05680573e+00 -6.98036373e-01
1.58325553e-01 4.94968116e-01 -5.10873079e-01 2.80651271e-01
5.94850257e-02 1.24897026e-01 2.20522680e-03 4.75674868e-01
6.41413152e-01 -4.09043998e-01 2.61351224e-02 -9.98995900e-02
2.79321998e-01 3.39188665e-01 -8.75378668e-01 1.81688464e+00
-5.70764422e-01 7.97518134e-01 -4.48430300e-01 -1.36396468e+00
1.26337409e+00 2.23329872e-01 4.11128163e-01 -5.54921508e-01
-2.27463413e-02 3.24634135e-01 -6.10625139e-04 -1.71830550e-01
2.26038679e-01 2.05553751e-02 5.41902065e-01 7.66237855e-01
6.80948675e-01 -1.69490501e-01 1.12466432e-01 4.61076945e-01
8.81656289e-01 3.55528072e-02 3.86152536e-01 -3.06284159e-01
2.31671348e-01 6.68660551e-02 2.48204768e-01 5.89020491e-01
-3.52292359e-01 6.51542783e-01 1.28826171e-01 -6.41960084e-01
-9.82169628e-01 -8.73495579e-01 -2.03169629e-01 9.94923353e-01
-4.24281359e-02 -7.50622869e-01 -8.39907587e-01 -1.01822829e+00
7.29481429e-02 6.21906996e-01 -5.13926148e-01 -4.56000239e-01
-4.74268526e-01 -1.82995901e-01 5.28451979e-01 6.64297998e-01
8.63568962e-01 -9.93966520e-01 -2.42046043e-01 -3.18193398e-02
1.43845350e-01 -1.05224168e+00 -4.17709708e-01 5.77622294e-01
-1.47728395e+00 -9.51056242e-01 -5.98329961e-01 -1.17553985e+00
1.14486873e+00 5.29821575e-01 1.08059454e+00 2.56668836e-01
-1.33374125e-01 4.26465183e-01 -5.19032776e-01 -4.44515914e-01
-5.33605576e-01 3.14248502e-01 1.95157647e-01 -1.40465885e-01
6.79385960e-01 -6.07249439e-01 -4.41977620e-01 2.28965580e-01
-1.01662147e+00 7.49175772e-02 7.40066588e-01 1.00597763e+00
7.40465343e-01 1.60152093e-01 3.71083170e-01 -1.05234957e+00
1.55855983e-01 -5.03508210e-01 -3.21386456e-01 8.57161507e-02
-1.14561784e+00 1.70643851e-01 9.06966865e-01 -6.38238192e-01
-7.39946127e-01 5.73564351e-01 -3.73473801e-02 -6.18659496e-01
-4.88442510e-01 3.11366647e-01 2.58457899e-01 -2.03216076e-01
9.40257609e-01 3.60134155e-01 3.05254817e-01 -6.18119776e-01
3.32383245e-01 9.17642057e-01 2.76148915e-01 -3.29751849e-01
9.29284155e-01 6.02882862e-01 -2.59364426e-01 -8.36728454e-01
-1.12914097e+00 -5.53843319e-01 -9.99246478e-01 3.97058353e-02
4.29899544e-01 -1.07866371e+00 -4.50430363e-02 3.10266167e-01
-7.95029998e-01 -6.53250694e-01 -8.33482265e-01 7.38131404e-01
-6.02388144e-01 3.77554327e-01 -5.70145011e-01 -3.54348660e-01
-2.64881879e-01 -9.77980077e-01 6.50273860e-01 7.57480189e-02
-1.51385859e-01 -1.18853128e+00 1.11701123e-01 4.23828870e-01
4.78053331e-01 -3.51626754e-01 3.30988914e-01 -9.95898068e-01
-2.29805335e-01 -2.61771679e-01 -1.92537576e-01 9.45460975e-01
1.48787871e-01 -3.34139884e-01 -1.09851277e+00 -6.44545674e-01
1.41817123e-01 -1.05546653e+00 1.21477306e+00 1.96939006e-01
1.31260204e+00 -4.42370564e-01 -1.99420303e-01 7.97463477e-01
1.33972526e+00 -1.11017376e-01 6.90702975e-01 3.61800104e-01
7.56852508e-01 4.49467868e-01 5.26735723e-01 1.34566545e-01
3.56309623e-01 3.68953764e-01 4.71987247e-01 -2.67936677e-01
-4.55578238e-01 -6.91664159e-01 4.04802978e-01 1.00722611e+00
1.76139757e-01 1.56321317e-01 -9.58409607e-01 4.55580533e-01
-1.56971860e+00 -7.83807218e-01 2.25456133e-01 2.10338569e+00
1.10939384e+00 1.01932317e-01 -1.94282923e-02 1.79471046e-01
5.20112455e-01 2.13732958e-01 -6.31026745e-01 -1.56943277e-01
-6.91360161e-02 4.41136271e-01 6.81364298e-01 4.61623251e-01
-1.33432305e+00 1.14972317e+00 5.85919666e+00 1.01075447e+00
-1.13045800e+00 2.08743393e-01 7.96762705e-01 -4.86085676e-02
1.21158376e-01 2.01363266e-01 -9.74797726e-01 3.74426067e-01
1.16804469e+00 -6.68694675e-02 2.11055860e-01 1.35044765e+00
-2.11227491e-01 1.04379341e-01 -1.21249008e+00 1.23458755e+00
2.43396670e-01 -1.25863886e+00 1.15784273e-01 7.06212893e-02
9.91969287e-01 3.87425274e-01 4.22987193e-02 4.10637587e-01
3.27911317e-01 -1.11606395e+00 3.37049276e-01 1.31979898e-01
8.08546484e-01 -5.79678118e-01 6.93908632e-01 5.14664769e-01
-9.63466942e-01 2.63067801e-02 -7.96998918e-01 -3.64344195e-02
-2.91593671e-01 5.79662681e-01 -9.86213863e-01 -3.53180692e-02
6.76553249e-01 1.28133726e+00 -8.42657983e-01 9.22200382e-01
-3.10792744e-01 1.12320316e+00 -3.26293856e-01 1.99487824e-02
3.16555679e-01 -7.76055157e-02 8.87680650e-02 1.18290615e+00
4.36503850e-02 -1.04982905e-01 1.94880337e-01 5.48757136e-01
-4.04687971e-01 2.55430609e-01 -7.81494856e-01 -1.22330844e-01
4.29574907e-01 1.05863476e+00 -4.92245436e-01 -6.49640083e-01
-5.17793298e-01 1.11876059e+00 7.20820546e-01 2.12811232e-01
-4.40325141e-01 -3.23187888e-01 1.85908586e-01 1.72549531e-01
1.72215641e-01 -7.47837573e-02 -1.93283528e-01 -1.19891560e+00
-1.46586239e-01 -8.89369845e-01 3.65051955e-01 -6.27487063e-01
-1.20618641e+00 7.33054638e-01 5.90846129e-03 -1.44369400e+00
-2.32633352e-01 -4.90856916e-01 -4.17959392e-01 3.70272547e-01
-1.61048174e+00 -9.76837039e-01 -3.61401111e-01 6.69137895e-01
6.20013654e-01 -6.78600013e-01 1.05278921e+00 1.45188347e-01
-2.34174699e-01 8.92077804e-01 3.38409841e-01 3.62430423e-01
9.10315871e-01 -1.33248436e+00 2.09480807e-01 4.42249209e-01
6.09375715e-01 4.86645669e-01 2.68940032e-01 -4.00714844e-01
-1.09370744e+00 -1.16699517e+00 1.11202538e+00 -5.63996360e-02
6.04542077e-01 -7.95606598e-02 -1.01425242e+00 7.90513813e-01
3.18493396e-01 2.88119644e-01 8.85428905e-01 -2.39682481e-01
-7.84584641e-01 -4.89132106e-01 -1.32708800e+00 4.14243966e-01
1.05128849e+00 -5.90216398e-01 -5.42011857e-01 5.78934312e-01
7.38432169e-01 -1.28510520e-01 -9.41574633e-01 2.25909993e-01
3.59155804e-01 -9.52849805e-01 9.26411927e-01 -5.12741208e-01
7.22475767e-01 -4.59833480e-02 -8.03043917e-02 -1.52640927e+00
-9.77399722e-02 -2.98029095e-01 -4.96696323e-01 9.09981668e-01
2.92370200e-01 -7.77010679e-01 1.22905433e+00 1.76532924e-01
-1.02254733e-01 -9.58662510e-01 -6.63268447e-01 -9.84598577e-01
2.51144379e-01 -1.47127062e-01 3.17628741e-01 1.26355577e+00
-1.98793709e-02 3.42792243e-01 -3.28063667e-01 -1.94214106e-01
8.83555591e-01 -2.25921571e-02 7.65611947e-01 -1.27339506e+00
-3.39989543e-01 -1.93242937e-01 -5.41773081e-01 -1.42225170e+00
4.54276830e-01 -1.35076916e+00 -3.72858137e-01 -1.31906962e+00
3.66503686e-01 -7.25645065e-01 -3.34889442e-01 8.64951193e-01
1.28119156e-01 4.72244084e-01 1.86908901e-01 5.85846364e-01
-6.00832820e-01 5.62650502e-01 1.16273975e+00 -3.06330562e-01
-8.10437053e-02 -1.91840008e-02 -7.12188423e-01 7.34424651e-01
1.66475534e+00 -7.66629696e-01 -6.47189319e-01 -5.51504791e-01
-3.27103138e-01 -4.10365790e-01 1.52335569e-01 -1.20620096e+00
5.58074415e-01 1.98146164e-01 3.43686581e-01 -4.93791193e-01
2.38320455e-01 -8.98421705e-01 -3.69501531e-01 8.33532929e-01
-7.44794965e-01 -1.76064923e-01 -1.20840846e-02 4.32839513e-01
-5.52978098e-01 -6.72569692e-01 9.67424572e-01 -2.09603414e-01
-7.54883409e-01 5.04604042e-01 -1.24114424e-01 1.66456657e-03
9.65240180e-01 -2.01697528e-01 -3.62188548e-01 -6.55845284e-01
-6.36000872e-01 1.11330330e-01 3.61342281e-01 2.39285067e-01
9.72279847e-01 -1.37939954e+00 -7.06097603e-01 3.19197625e-01
2.26668596e-01 5.60129471e-02 -5.12606315e-02 5.64755797e-01
-5.45385301e-01 3.53631645e-01 -1.60275176e-01 -6.86410308e-01
-1.29005826e+00 4.61356252e-01 2.32146338e-01 -2.06368044e-01
-4.95053619e-01 8.69581044e-01 -1.68447644e-01 -8.15522015e-01
6.93858802e-01 -7.43508711e-02 -1.42063871e-01 -1.29088491e-01
6.16128385e-01 4.68019992e-01 -8.75968859e-02 -6.51591241e-01
1.40840352e-01 4.50528473e-01 -4.39228296e-01 2.38547876e-01
1.52513659e+00 1.60299286e-01 -1.10602900e-01 4.09337848e-01
1.89441919e+00 -3.85193020e-01 -1.29298604e+00 -7.10661232e-01
-1.36975959e-01 -2.59846479e-01 2.71969110e-01 -4.61616993e-01
-1.32767630e+00 9.47391927e-01 7.76336372e-01 -1.55256003e-01
9.68259513e-01 1.67340577e-01 8.71245265e-01 9.36321676e-01
3.14606130e-01 -1.14607394e+00 4.48047727e-01 5.75867176e-01
6.66290700e-01 -1.65367770e+00 1.95068762e-01 -2.29600057e-01
-7.08326638e-01 1.16136062e+00 7.61199415e-01 -3.40783894e-01
1.08093059e+00 1.03185669e-01 1.96315095e-01 -4.87411337e-04
-6.64118469e-01 -3.00772250e-01 1.85447827e-01 6.15077615e-01
4.68869865e-01 -6.63662609e-03 3.74130644e-02 2.60839332e-02
-4.67447430e-01 1.13298729e-01 3.28215808e-01 9.84612107e-01
-6.44459307e-01 -1.29420888e+00 -4.40572249e-03 6.74288511e-01
-6.74709976e-02 -4.12872791e-01 -3.42290044e-01 5.98656118e-01
-8.27649236e-03 8.91615152e-01 2.56086320e-01 -4.98467416e-01
1.76549908e-02 1.02888115e-01 3.08074594e-01 -8.87649536e-01
-3.78732622e-01 -2.36931428e-01 -2.17164189e-01 -4.76309001e-01
-6.47133887e-01 -2.60340422e-01 -1.24302924e+00 -3.29876184e-01
-2.38003805e-01 2.61707515e-01 7.45364070e-01 5.64872503e-01
2.00532496e-01 3.59390490e-02 8.47205758e-01 -7.64550686e-01
-1.02320504e+00 -1.04451871e+00 -5.50826490e-01 4.83897865e-01
1.70640841e-01 -3.78827453e-01 -6.37728155e-01 1.39801979e-01] | [9.476737976074219, 2.702094078063965] |
e6f11218-0e25-4299-9840-74885360cd5d | convolutional-recurrent-neural-networks-for-4 | 1703.02317 | null | http://arxiv.org/abs/1703.02317v1 | http://arxiv.org/pdf/1703.02317v1.pdf | Convolutional Recurrent Neural Networks for Bird Audio Detection | Bird sounds possess distinctive spectral structure which may exhibit small
shifts in spectrum depending on the bird species and environmental conditions.
In this paper, we propose using convolutional recurrent neural networks on the
task of automated bird audio detection in real-life environments. In the
proposed method, convolutional layers extract high dimensional, local frequency
shift invariant features, while recurrent layers capture longer term
dependencies between the features extracted from short time frames. This method
achieves 88.5% Area Under ROC Curve (AUC) score on the unseen evaluation data
and obtains the second place in the Bird Audio Detection challenge. | ['EmreÇakır', 'Sharath Adavanne', 'Tuomas Virtanen', 'Konstantinos Drossos', 'Giambattista Parascandolo'] | 2017-03-07 | null | null | null | null | ['bird-audio-detection'] | ['audio'] | [ 7.12163448e-02 -8.36021781e-01 3.71180445e-01 -1.20346166e-01
-3.44338298e-01 -5.72909296e-01 -4.22183089e-02 1.66243896e-01
-6.23101711e-01 9.17093754e-02 4.22934026e-01 2.50197709e-01
-9.70287845e-02 -5.68332970e-01 -3.30007732e-01 -3.65757823e-01
-8.37305248e-01 -7.03310609e-01 3.20623189e-01 -3.08953494e-01
-1.53316095e-01 3.08899760e-01 -1.86791086e+00 5.51153958e-01
1.77281257e-02 9.87254560e-01 -1.48851052e-01 1.42862916e+00
7.87419677e-01 6.48818433e-01 -8.42297614e-01 4.65487652e-02
3.70108843e-01 -3.62059951e-01 -6.04268670e-01 -4.28742945e-01
3.11186820e-01 -1.84718207e-01 -5.48081815e-01 8.91274631e-01
9.35935020e-01 3.61195207e-01 3.56790155e-01 -9.11066473e-01
-2.73000926e-01 9.95164037e-01 -3.79402757e-01 9.59521234e-01
6.33768618e-01 1.70412108e-01 1.34363031e+00 -7.04006016e-01
-3.68584767e-02 7.05027819e-01 1.20244968e+00 3.13836038e-01
-1.07224309e+00 -7.90370643e-01 -4.08390611e-02 1.95821777e-01
-1.63124359e+00 -4.31388766e-01 9.01892483e-01 -5.09958029e-01
1.22321188e+00 1.98249817e-01 9.21462595e-01 8.23105872e-01
2.28634566e-01 5.26201725e-01 5.92949569e-01 -2.67074943e-01
-3.19901139e-01 -1.92359433e-01 1.18273899e-01 5.08130908e-01
-2.41008401e-01 7.12356150e-01 -7.99197912e-01 -2.12351829e-01
3.00517350e-01 -1.29758483e-02 -3.98550838e-01 2.72556871e-01
-1.24244571e+00 5.13050795e-01 9.02446449e-01 4.98923212e-01
-3.23838413e-01 3.61496508e-01 8.99898231e-01 6.78055882e-01
4.15638864e-01 6.01876140e-01 -4.75236923e-01 -1.66662410e-01
-8.51132393e-01 2.73176074e-01 6.84111238e-01 4.09243762e-01
2.37709507e-01 4.30397362e-01 -5.76461889e-02 1.03357565e+00
2.25739311e-02 2.95604527e-01 7.77988195e-01 -1.78859517e-01
-6.54494017e-02 7.41165131e-02 -3.31146568e-02 -1.11336100e+00
-8.99917185e-01 -1.05965865e+00 -6.86156571e-01 -4.13634688e-01
-1.49834543e-01 -5.63521266e-01 -5.85456252e-01 1.82151306e+00
1.25523463e-01 3.78859818e-01 -6.16326854e-02 1.09314156e+00
1.25910914e+00 9.41162944e-01 6.54829592e-02 -2.21874565e-01
1.21778190e+00 -6.66681588e-01 -4.00090307e-01 3.30294743e-02
1.13621920e-01 -9.09623504e-01 7.17580140e-01 3.51685494e-01
-5.80091417e-01 -9.34786975e-01 -1.39055181e+00 2.73523927e-01
-4.30316299e-01 4.86526698e-01 2.92409480e-01 5.99815428e-01
-1.15361273e+00 4.46926117e-01 -6.55990124e-01 -1.99800506e-01
-1.71387941e-01 4.60457444e-01 -4.78590913e-02 6.25500560e-01
-1.48223615e+00 2.72395551e-01 1.70651078e-01 5.98106325e-01
-1.27471817e+00 -7.43391156e-01 -7.66880453e-01 1.92208424e-01
-3.06180418e-02 -3.59067440e-01 1.50080681e+00 -7.93009758e-01
-1.39656758e+00 8.03449452e-01 4.84164357e-01 -1.05887687e+00
2.98947170e-02 -4.17160094e-01 -9.02395070e-01 -1.40547872e-01
-5.33732511e-02 5.64369619e-01 1.13510633e+00 -2.37658978e-01
-7.87396133e-01 -8.86622742e-02 3.39176208e-02 -3.09404843e-02
-2.44325951e-01 4.48563963e-01 4.57262546e-01 -1.07384861e+00
-1.48446620e-01 -9.97361541e-01 -1.90366030e-01 -4.65040833e-01
3.07437638e-03 -2.80178100e-01 6.67682350e-01 -5.20811319e-01
1.49408245e+00 -2.75347638e+00 1.27080092e-02 -1.67102784e-01
2.11030677e-01 3.32099855e-01 -2.35021010e-01 3.64910454e-01
-1.22325681e-01 -2.49789998e-01 1.21855415e-01 3.66798401e-01
-2.84601618e-02 -5.89317977e-01 -5.81950605e-01 5.14886856e-01
2.80461460e-01 4.32799250e-01 -9.78461742e-01 1.07494056e-01
-1.32472869e-02 5.94563603e-01 -5.83896816e-01 3.17715675e-01
2.67341197e-01 -1.91174179e-01 4.85650301e-02 3.84440511e-01
3.15391541e-01 5.00420749e-01 -2.61367083e-01 -2.46510386e-01
-4.04990435e-01 5.53460181e-01 -9.67442930e-01 1.58518136e+00
-6.15982592e-01 1.14665318e+00 -6.61069229e-02 -6.39934182e-01
9.92915630e-01 5.57502925e-01 3.64727795e-01 -2.82605678e-01
1.84477374e-01 1.21763408e-01 5.09649754e-01 -2.13061213e-01
5.94738662e-01 -1.63714007e-01 -4.83011723e-01 2.89567441e-01
4.80683029e-01 -7.83407092e-02 -1.93153191e-02 -3.52710813e-01
1.07306659e+00 -3.85120094e-01 4.42317665e-01 -2.20575184e-01
4.71155137e-01 -7.35370696e-01 5.79134524e-01 6.65815473e-01
-5.02818763e-01 3.78690302e-01 1.68774575e-01 -9.13732529e-01
-4.74053681e-01 -9.35863316e-01 -2.74752736e-01 1.64796138e+00
-1.97680235e-01 -8.47908735e-01 -5.89109540e-01 -4.69323903e-01
-1.32082611e-01 8.17686021e-02 -6.79403007e-01 -6.68351650e-01
-5.89824378e-01 -6.77191734e-01 1.17346632e+00 6.44716859e-01
3.98756206e-01 -1.35214424e+00 -1.27148545e+00 5.43231189e-01
6.59648180e-02 -9.00112271e-01 -6.95188582e-01 4.92929816e-01
-6.66375086e-02 -7.24452436e-01 -3.29757571e-01 -9.75881636e-01
-1.94684550e-01 4.01839703e-01 9.05688882e-01 -2.20452785e-01
-7.36295998e-01 3.35031927e-01 -5.31470478e-01 -6.69919193e-01
-5.61816292e-03 4.95263815e-01 4.52546090e-01 9.63819586e-03
3.53253812e-01 -6.53651118e-01 -6.38422072e-01 3.87282997e-01
-6.23907864e-01 -5.99955082e-01 1.61349967e-01 1.02028167e+00
3.99056941e-01 3.70716333e-01 9.26553488e-01 -7.22277984e-02
8.98647547e-01 -2.11651325e-01 -5.28125942e-01 3.40002328e-02
8.07502046e-02 -3.26986879e-01 8.33101928e-01 -6.24341965e-01
-4.23573315e-01 2.46189684e-01 -1.71810269e-01 -2.46246010e-01
-2.32772201e-01 4.62673783e-01 3.74597698e-01 -7.01399595e-02
1.04223406e+00 8.29811767e-02 -6.83112085e-01 -4.83671248e-01
-8.71472582e-02 7.76102602e-01 6.91142619e-01 6.02622144e-02
5.20738661e-01 -9.84808709e-03 -3.43016297e-01 -1.08493090e+00
-1.06003308e+00 -6.19718075e-01 -5.64669013e-01 -4.87229526e-01
7.10567594e-01 -1.05536222e+00 -6.54345155e-01 4.86748308e-01
-8.66468668e-01 1.48293510e-01 -2.56563157e-01 8.31701219e-01
-2.45482460e-01 -3.10886115e-01 -6.80242538e-01 -8.84608984e-01
-7.85801649e-01 -8.44503880e-01 8.10918570e-01 2.25607246e-01
-3.57247531e-01 -3.33754838e-01 5.32712936e-01 -3.95232022e-01
4.77645159e-01 2.02150330e-01 3.33073825e-01 -4.99886930e-01
-4.77056429e-02 -5.58890939e-01 2.30292052e-01 4.05657023e-01
3.73627365e-01 1.89160019e-01 -1.33126295e+00 -5.27806461e-01
-5.78982718e-02 -1.68865368e-01 1.29786265e+00 6.75220966e-01
1.02379775e+00 -2.18189135e-01 2.22298071e-01 8.32436204e-01
8.60761702e-01 2.37373888e-01 -1.74920604e-01 -1.87835097e-01
4.05734181e-01 2.94100553e-01 4.43628192e-01 7.01140404e-01
-1.89645022e-01 6.94554090e-01 1.17080614e-01 1.07514504e-02
-1.48424610e-01 -5.31155050e-01 7.55390227e-01 1.01474988e+00
5.21616787e-02 5.35697266e-02 -9.87917721e-01 7.07381129e-01
-1.48756552e+00 -1.25782073e+00 3.17602664e-01 2.12346387e+00
5.62109113e-01 1.58427626e-01 7.75252640e-01 5.28638482e-01
7.86147654e-01 2.96514273e-01 -3.61123055e-01 -6.31662607e-01
-2.23643091e-02 2.38671869e-01 1.17850512e-01 1.31657615e-01
-1.76178563e+00 8.50663662e-01 6.94183683e+00 3.44918489e-01
-1.46270680e+00 -5.55503592e-02 9.76479873e-02 -5.55842578e-01
5.61616540e-01 -6.00271285e-01 -4.98053908e-01 1.93698913e-01
1.32040513e+00 -5.17280102e-01 5.78103244e-01 7.21717834e-01
-7.84659535e-02 4.14561391e-01 -1.01948619e+00 1.15598583e+00
-4.98276167e-02 -8.29553604e-01 -2.90503293e-01 -2.98330545e-01
4.57065612e-01 3.78662288e-01 5.91325581e-01 4.74074930e-01
-5.92931956e-02 -1.08980906e+00 6.79752886e-01 3.24596763e-01
8.44726980e-01 -1.21751547e+00 8.36299002e-01 1.56932846e-01
-1.84129119e+00 -3.87120724e-01 -4.03214961e-01 -3.10159147e-01
-2.82350749e-01 2.42040977e-01 -1.35213280e+00 8.40849653e-02
1.20911562e+00 9.82045352e-01 -6.52603984e-01 1.43559313e+00
1.12943202e-01 9.56652343e-01 -3.82840931e-01 -2.59956867e-01
1.79023176e-01 4.33886290e-01 9.02713537e-01 1.73146260e+00
1.34591937e-01 -7.28968084e-02 1.18186191e-01 4.99164343e-01
-1.30441651e-01 9.91255045e-02 -7.67183900e-01 -1.20759591e-01
1.12546206e-01 1.18583453e+00 -6.81738079e-01 2.99270630e-01
-1.28401935e-01 4.00824159e-01 -9.51333623e-03 3.82982269e-02
-9.62457001e-01 -1.02763724e+00 7.36082554e-01 -1.64661303e-01
3.83961618e-01 7.16369366e-03 4.09838855e-01 -9.33429003e-01
-4.42674346e-02 -7.36758947e-01 7.34820902e-01 -3.22476804e-01
-1.36161423e+00 1.07170045e+00 -2.87072897e-01 -1.61485255e+00
-5.60559452e-01 -2.76478946e-01 -5.97636104e-01 4.14128482e-01
-1.02378821e+00 -9.29776132e-01 -6.60594627e-02 4.84037429e-01
6.12615585e-01 -3.78877133e-01 9.11193967e-01 3.79609436e-01
-4.58189875e-01 8.90507579e-01 -2.43242174e-01 4.32053238e-01
6.35961890e-01 -1.03758943e+00 7.23553061e-01 8.78811538e-01
8.05897713e-01 4.32736009e-01 7.45526016e-01 -1.37631193e-01
-9.06495869e-01 -1.16922629e+00 5.62572360e-01 1.44839436e-02
7.20046997e-01 -5.42875051e-01 -6.44745588e-01 2.10824385e-01
3.14445868e-02 2.67736584e-01 1.11265314e+00 2.73807377e-01
-5.50327778e-01 -5.89847744e-01 -7.94771969e-01 1.23313643e-01
9.11812246e-01 -1.10211730e+00 -5.46673834e-01 -4.41770628e-02
9.55702007e-01 -2.39219487e-01 -6.45798266e-01 6.04234219e-01
9.38747227e-01 -8.20315421e-01 9.63856578e-01 -6.77775621e-01
-4.21357229e-02 -5.62027454e-01 -2.40061119e-01 -1.46758366e+00
-5.72466433e-01 -9.55336750e-01 1.94360375e-01 8.67376566e-01
5.22979319e-01 -2.25860864e-01 7.79223070e-02 -4.02043223e-01
-1.10744707e-01 -3.64532948e-01 -9.75715518e-01 -5.38220763e-01
-4.40891683e-01 -5.06064057e-01 7.49804795e-01 5.50876737e-01
1.05051450e-01 4.66711193e-01 -5.50281644e-01 3.72811586e-01
1.55459285e-01 3.94794464e-01 4.28058326e-01 -1.27547562e+00
-3.22765082e-01 -3.85322541e-01 -1.06729257e+00 -6.77648902e-01
1.74512386e-01 -7.16414213e-01 2.71304905e-01 -5.97798407e-01
-1.16270334e-01 2.99502850e-01 -1.14990151e+00 2.83366442e-01
8.71731490e-02 4.90372807e-01 1.33619741e-01 -3.36147994e-01
-7.64500380e-01 5.31581521e-01 4.24840242e-01 -3.35377723e-01
-5.28733134e-01 5.97124696e-01 -2.28648394e-01 7.28317320e-01
7.62589574e-01 -4.84072357e-01 -2.49563992e-01 -3.72324347e-01
3.03295821e-01 1.30533367e-01 5.06900549e-01 -1.42142785e+00
1.86954603e-01 3.23554128e-01 5.67782700e-01 -5.73000014e-01
1.75925180e-01 -5.79862893e-01 9.55370907e-03 7.67806351e-01
-8.75296831e-01 1.73295245e-01 5.10202229e-01 6.33179307e-01
-5.30924618e-01 3.77179347e-02 7.56681621e-01 1.35743007e-01
-4.48377341e-01 2.33222514e-01 -6.55152559e-01 -1.40667051e-01
5.19243717e-01 1.83068082e-01 -2.76441360e-03 -2.75693893e-01
-8.39137316e-01 -2.88010746e-01 -3.51294100e-01 7.91987896e-01
7.33973742e-01 -1.29846799e+00 -9.14262593e-01 5.57367384e-01
1.09925739e-01 -6.11948311e-01 5.28078675e-01 3.96024555e-01
-3.58625203e-01 4.72570270e-01 -3.64912897e-01 -6.05728269e-01
-1.52474236e+00 3.38932425e-01 8.27529907e-01 3.03947359e-01
-5.33176601e-01 1.24939120e+00 1.56463355e-01 -2.24642843e-01
3.40532511e-01 -6.91478193e-01 -4.97208446e-01 2.44186640e-01
7.13992000e-01 3.07500094e-01 2.15517238e-01 -8.46258819e-01
-5.39359212e-01 3.86951685e-01 -8.52802843e-02 8.15545693e-02
1.49721873e+00 2.39792630e-01 4.82227579e-02 8.99238050e-01
1.27872431e+00 1.48514152e-01 -8.41062844e-01 -2.84439743e-01
-1.26829207e-01 -4.50680733e-01 4.19956297e-01 -6.97322130e-01
-1.01527965e+00 1.08844471e+00 1.13788176e+00 8.22818816e-01
1.43809092e+00 -1.02453664e-01 8.18809450e-01 4.27692711e-01
1.00724533e-01 -8.94942403e-01 -2.70435542e-01 5.51151037e-01
1.13738537e+00 -6.54099584e-01 -1.67185456e-01 1.94534421e-01
-4.23272908e-01 1.16833162e+00 5.81786990e-01 -4.30035055e-01
9.84917402e-01 2.75535136e-01 1.40524164e-01 -6.61695600e-02
-1.19235480e+00 -4.09406573e-01 6.03257596e-01 5.70394754e-01
7.54750252e-01 5.22293091e-01 1.56066284e-01 8.71805429e-01
-1.01908326e+00 -5.92681289e-01 2.76977241e-01 3.97994101e-01
-4.74189073e-01 -3.04480046e-01 -9.57508665e-03 3.39728922e-01
-7.14376748e-01 4.54841480e-02 -7.23706722e-01 1.37970254e-01
1.84124023e-01 1.01911807e+00 6.87947869e-02 -9.02395785e-01
5.60920477e-01 -4.50138077e-02 1.59047499e-01 -5.16516268e-01
-1.30675447e+00 4.49305087e-01 -1.60390019e-01 -2.27676958e-01
-3.98634940e-01 -6.56457365e-01 -7.70896435e-01 1.09653339e-01
-5.76626778e-01 2.93184042e-01 3.43320906e-01 3.37951899e-01
2.60020107e-01 7.66302168e-01 1.22121012e+00 -7.61461794e-01
-4.64552701e-01 -1.07153320e+00 -5.97353339e-01 1.32505065e-02
1.18680155e+00 -4.79702085e-01 -4.84809935e-01 -1.12945542e-01] | [15.220247268676758, 5.283101558685303] |
2df0b15b-c560-4e54-aa3d-9aafbd53cc35 | improving-weakly-supervised-sound-event-1 | 2303.05678 | null | https://arxiv.org/abs/2303.05678v1 | https://arxiv.org/pdf/2303.05678v1.pdf | Improving Weakly Supervised Sound Event Detection with Causal Intervention | Existing weakly supervised sound event detection (WSSED) work has not explored both types of co-occurrences simultaneously, i.e., some sound events often co-occur, and their occurrences are usually accompanied by specific background sounds, so they would be inevitably entangled, causing misclassification and biased localization results with only clip-level supervision. To tackle this issue, we first establish a structural causal model (SCM) to reveal that the context is the main cause of co-occurrence confounders that mislead the model to learn spurious correlations between frames and clip-level labels. Based on the causal analysis, we propose a causal intervention (CI) method for WSSED to remove the negative impact of co-occurrence confounders by iteratively accumulating every possible context of each class and then re-projecting the contexts to the frame-level features for making the event boundary clearer. Experiments show that our method effectively improves the performance on multiple datasets and can generalize to various baseline models. | ['Yuexian Zou', 'Yujun Wang', 'Fan Cui', 'Dongchao Yang', 'Yifei Xin'] | 2023-03-10 | null | null | null | null | ['sound-event-detection'] | ['audio'] | [ 4.92999762e-01 -4.09882665e-02 -1.67021021e-01 -3.43149483e-01
-8.41795564e-01 -3.77418756e-01 4.86478657e-01 2.26321489e-01
-7.48556107e-02 6.64701879e-01 6.72702014e-01 -7.93824568e-02
-1.29793555e-01 -5.88205874e-01 -1.00708830e+00 -7.36508071e-01
-4.70322043e-01 -3.55492502e-01 6.57086074e-01 3.70008558e-01
4.23248708e-02 -1.31145358e-01 -1.59495974e+00 6.77881777e-01
4.99987721e-01 6.09308243e-01 2.98032373e-01 4.18404341e-01
7.29649290e-02 1.17746973e+00 -8.53672981e-01 -1.42919213e-01
-9.68806967e-02 -8.64512026e-01 -4.13655162e-01 -2.32160985e-01
5.02827466e-01 -1.68139964e-01 -3.87974441e-01 1.01725161e+00
3.77259642e-01 -5.76662757e-02 4.71167564e-01 -1.42779529e+00
-2.91336030e-01 9.84896660e-01 -7.49441564e-01 6.49842501e-01
2.40806356e-01 1.40012046e-02 1.14149761e+00 -1.01845658e+00
3.74436885e-01 1.55123019e+00 8.55379701e-01 1.93996668e-01
-1.08188009e+00 -1.09266281e+00 7.05927908e-01 4.56558943e-01
-1.29104125e+00 -3.63996536e-01 9.60182607e-01 -2.57868946e-01
4.42375958e-01 4.11146879e-01 4.14216071e-01 1.50822794e+00
-2.01677099e-01 7.40598857e-01 8.08634520e-01 -4.11850601e-01
2.27282107e-01 -3.82510543e-01 -1.46697626e-01 5.21689057e-01
7.48242512e-02 8.11325610e-02 -9.63235915e-01 -2.71582425e-01
5.72277009e-01 2.71264445e-02 -1.76461130e-01 2.21086755e-01
-1.64308250e+00 6.07967675e-01 3.27475518e-01 3.53344798e-01
-2.76384473e-01 4.28347021e-01 4.09880400e-01 -2.90401578e-02
6.17955089e-01 2.71984845e-01 -5.16206622e-01 -6.21516369e-02
-8.67813289e-01 4.01651651e-01 2.92049855e-01 6.86997890e-01
5.42427719e-01 -2.46244788e-01 -5.51760912e-01 6.69256330e-01
2.21132785e-01 3.73601943e-01 1.00772329e-01 -7.96393931e-01
4.99433726e-01 3.32088828e-01 1.15787506e-01 -1.32132339e+00
-3.60130757e-01 -4.34095234e-01 -6.84849679e-01 -3.83576274e-01
5.05069554e-01 -2.95515209e-01 -6.45289421e-01 2.09384584e+00
4.63017732e-01 1.30440569e+00 -5.84202349e-01 1.00079775e+00
7.46768892e-01 5.26229143e-01 5.52780688e-01 -5.30203819e-01
1.23738205e+00 -7.00155616e-01 -8.97949874e-01 -1.98297545e-01
4.10597026e-01 -6.77684665e-01 1.26619399e+00 1.10573299e-01
-6.75821066e-01 -5.20117044e-01 -9.07068014e-01 3.03390533e-01
2.98811775e-02 2.96960697e-02 5.96536934e-01 1.27883837e-01
-1.40010312e-01 5.81915498e-01 -8.72055829e-01 8.55557546e-02
5.47450602e-01 -5.32192625e-02 2.83809472e-02 6.55503720e-02
-1.63169181e+00 4.76583272e-01 1.79251894e-01 2.06463486e-02
-1.15101027e+00 -1.05083299e+00 -4.52709943e-01 1.03216872e-01
7.54953146e-01 -3.48322064e-01 1.09374201e+00 -8.73064995e-01
-8.65953445e-01 3.06069523e-01 -5.66759348e-01 -2.50937611e-01
3.43347937e-01 -4.82950985e-01 -5.89304328e-01 1.05974665e-02
3.35293025e-01 2.56447971e-01 9.17220831e-01 -1.24133694e+00
-1.02566898e+00 2.16873898e-03 1.96633637e-02 4.15240154e-02
-3.37949038e-01 2.76135325e-01 -5.00774324e-01 -1.02886033e+00
2.32677773e-01 -6.71966493e-01 -1.05369478e-01 -2.20964685e-01
-5.35773277e-01 -2.67924190e-01 8.44012380e-01 -3.40330899e-01
1.75827801e+00 -2.50545049e+00 -1.66131526e-01 -1.61402300e-01
2.65304178e-01 -9.41802263e-02 -6.06626682e-02 1.59227565e-01
-2.89549679e-01 1.53859437e-01 -1.91023439e-01 -2.98218668e-01
-1.40478894e-01 2.44341061e-01 -8.15475166e-01 3.85297567e-01
3.19000036e-01 4.79048193e-01 -1.47125721e+00 -5.99546909e-01
1.49972811e-01 2.58841813e-01 -5.87590158e-01 2.18957961e-01
-2.69431233e-01 5.24125934e-01 -3.08033943e-01 3.14563245e-01
4.26513076e-01 -1.30624473e-01 -4.20174282e-03 -3.67752224e-01
-1.44112632e-01 6.92003846e-01 -1.47770703e+00 1.33907950e+00
-3.36171091e-01 5.43401599e-01 -2.56277680e-01 -9.42033410e-01
4.39229250e-01 5.58138907e-01 5.09910643e-01 -3.74168068e-01
-2.89881021e-01 -8.46774597e-03 -3.48708965e-02 -7.94360280e-01
-1.16579235e-01 -2.64226198e-01 -6.26675263e-02 3.09041142e-01
-9.78919938e-02 5.71668148e-01 -1.17561750e-01 1.61162332e-01
1.39767408e+00 -9.91814211e-02 5.76130934e-02 3.89080346e-02
2.72626787e-01 -2.71911353e-01 1.24961174e+00 1.11579418e+00
-2.91729122e-01 7.82226264e-01 7.61505425e-01 -2.45690852e-01
-6.31705999e-01 -1.15305722e+00 -3.02670915e-02 1.20360410e+00
3.81638765e-01 -6.97832227e-01 -6.79099143e-01 -9.54626501e-01
-1.84887469e-01 7.98978508e-01 -6.92320049e-01 -3.15246046e-01
-7.65512526e-01 -9.97921884e-01 7.11365700e-01 5.02918065e-01
2.38556772e-01 -1.05148232e+00 -4.53619838e-01 4.57164884e-01
-6.21583998e-01 -1.04173100e+00 -5.65118074e-01 2.07124442e-01
-4.70450461e-01 -1.17080879e+00 -1.13235220e-01 -3.89332801e-01
3.99628222e-01 2.64823437e-01 1.15049887e+00 1.41406894e-01
-1.88574806e-01 -2.31471628e-01 -4.40744847e-01 -5.53378105e-01
-1.98688135e-01 -3.07545662e-01 3.07889320e-02 2.41956607e-01
3.85248482e-01 -7.31206536e-01 -7.49547720e-01 3.86322975e-01
-7.47522652e-01 5.18737398e-02 2.60726959e-01 8.75821948e-01
5.18683016e-01 5.34857690e-01 9.90559697e-01 -9.43507493e-01
1.72356412e-01 -7.38498330e-01 -2.63382286e-01 1.09902127e-02
-2.42283612e-01 -3.84938478e-01 5.40837109e-01 -8.74279082e-01
-1.20337415e+00 3.62846255e-02 2.71400195e-02 -6.40513182e-01
-4.62602824e-01 3.57930869e-01 -5.54180562e-01 6.56186759e-01
5.58730364e-01 -5.46221286e-02 -7.13740945e-01 -4.64243889e-01
2.95556933e-01 3.67775351e-01 6.87266111e-01 -3.81011367e-01
6.67005956e-01 6.67104959e-01 -1.73219115e-01 -5.88961661e-01
-1.39284110e+00 -4.18685794e-01 -4.24389720e-01 -2.41057262e-01
8.47923994e-01 -1.04236078e+00 -2.68856436e-01 2.26807192e-01
-1.27789986e+00 -4.26792763e-02 -1.01398528e-01 6.95271075e-01
-3.32989395e-02 1.66463956e-01 -4.66999263e-01 -9.04039681e-01
4.01882142e-01 -9.01941061e-01 1.20763683e+00 -9.51096788e-03
-5.17313123e-01 -6.20542407e-01 1.09261632e-01 5.05313389e-02
-1.42374516e-01 2.21110880e-01 8.69522989e-01 -4.41357553e-01
-5.13239264e-01 -2.28005126e-02 -2.15688571e-01 -7.48894513e-02
3.72581095e-01 1.56663299e-01 -1.25760877e+00 3.91397804e-01
1.95007976e-02 2.03794897e-01 1.01534379e+00 4.77496356e-01
1.45859611e+00 -4.76502180e-01 -5.36126733e-01 3.36640745e-01
9.33848143e-01 1.31930649e-01 3.62626106e-01 -1.83708034e-02
1.02802920e+00 5.70451558e-01 6.37243509e-01 4.43459421e-01
3.69799107e-01 8.62449348e-01 5.26033461e-01 -3.05629820e-01
-3.53743881e-01 -7.04349577e-01 4.30150211e-01 5.31521499e-01
7.39475414e-02 -2.01648757e-01 -5.62391758e-01 7.87922025e-01
-2.03469276e+00 -1.28070939e+00 -6.39192462e-01 2.13291502e+00
1.07112896e+00 1.56721801e-01 4.22642641e-02 4.21513200e-01
1.09596825e+00 3.56752127e-01 -3.66910279e-01 3.23063105e-01
-2.42643297e-01 4.24296781e-02 3.89410794e-01 3.00229341e-01
-1.38248372e+00 6.51939869e-01 6.28061628e+00 9.73836064e-01
-8.94362092e-01 4.86726731e-01 5.56322455e-01 -5.18938482e-01
-3.30245644e-01 2.24206164e-01 -6.46253347e-01 9.70796406e-01
7.29754984e-01 8.00990984e-02 1.42252222e-01 4.61433172e-01
7.18467414e-01 -8.91847536e-02 -1.07766819e+00 6.80809557e-01
-2.11629182e-01 -1.11670792e+00 -1.30105376e-01 -2.29400679e-01
6.84538245e-01 -1.95637226e-01 -2.54829764e-01 2.23314539e-01
3.10026199e-01 -6.86786473e-01 1.07499170e+00 4.33637053e-01
3.86847734e-01 -5.95823288e-01 4.12716120e-01 4.08878446e-01
-1.19252968e+00 -1.62004828e-01 -2.57195592e-01 -3.43309551e-01
3.01788926e-01 1.21245730e+00 -4.20911491e-01 3.73968601e-01
9.00418460e-01 8.04164290e-01 -3.03728729e-01 9.35187638e-01
-6.68732643e-01 1.44506299e+00 -3.77631158e-01 2.07868621e-01
-1.78557977e-01 4.52246279e-01 8.53790462e-01 1.42071486e+00
2.13203907e-01 1.86222240e-01 1.33683428e-01 9.52116489e-01
1.94787960e-02 -2.46361092e-01 -4.05950338e-01 3.91744941e-01
7.21199751e-01 8.92438591e-01 -8.27512443e-01 -3.85255486e-01
-5.83907008e-01 7.44390607e-01 3.76852117e-02 4.28744286e-01
-1.36014843e+00 -3.90052795e-02 6.44408882e-01 5.11104167e-02
3.79056543e-01 1.92309663e-01 -4.85537857e-01 -1.35375535e+00
1.15112498e-01 -5.68846345e-01 6.58396780e-01 -5.62931955e-01
-1.44152820e+00 1.04809321e-01 2.25825105e-02 -1.29188275e+00
2.27994740e-01 2.10035577e-01 -9.69961107e-01 6.10598743e-01
-1.40587449e+00 -9.34396982e-01 -1.01471014e-01 5.79416394e-01
5.48559666e-01 3.82916480e-01 5.93232632e-01 6.69114113e-01
-8.48872244e-01 5.98272443e-01 -4.62028027e-01 8.77334774e-02
1.01643777e+00 -1.13847458e+00 2.01301724e-02 1.24498594e+00
5.19121706e-01 5.12974620e-01 8.54702175e-01 -8.87360692e-01
-8.93737793e-01 -1.38878000e+00 1.22891593e+00 -6.13301873e-01
8.47602189e-01 -5.70532441e-01 -1.04789555e+00 4.91611838e-01
-2.17183083e-01 2.33254895e-01 5.76504886e-01 4.58245635e-01
-4.38230127e-01 -2.02755958e-01 -5.97856045e-01 5.84101975e-01
1.49958277e+00 -5.80575049e-01 -8.63542616e-01 3.46530139e-01
9.47593629e-01 -2.22770214e-01 -2.23841131e-01 5.60653806e-01
3.54340196e-01 -9.93530929e-01 1.09612763e+00 -4.33913440e-01
6.94906831e-01 -7.20695019e-01 6.23570681e-02 -1.10458136e+00
-3.61929119e-01 -5.17101824e-01 -3.72101456e-01 1.70851493e+00
4.09067124e-01 2.81485706e-03 2.74989724e-01 2.84084797e-01
-1.08036615e-01 -3.28175604e-01 -1.13058114e+00 -6.52778804e-01
-4.50672746e-01 -9.29763794e-01 7.81764805e-01 1.21193254e+00
5.87179922e-02 4.38736439e-01 -6.87231064e-01 8.84031773e-01
6.01016641e-01 1.75730944e-01 4.46093440e-01 -9.47884321e-01
-3.59299690e-01 -3.41982573e-01 -9.44050252e-02 -7.45141268e-01
-1.85614489e-02 -5.08637249e-01 5.23356974e-01 -9.34781313e-01
2.42658913e-01 -6.16827369e-01 -7.09713042e-01 6.24750912e-01
-1.02622890e+00 1.57246500e-01 -7.00285062e-02 3.09512824e-01
-7.67869294e-01 4.02665049e-01 9.97860014e-01 6.00431561e-02
-2.02090904e-01 6.90820366e-02 -5.34147501e-01 1.06688666e+00
2.72553027e-01 -9.83251989e-01 -4.54648077e-01 -2.85627335e-01
2.65163213e-01 7.81407505e-02 8.41525137e-01 -8.96483123e-01
2.41539538e-01 -2.78943986e-01 1.67536050e-01 -4.60954726e-01
-1.99152350e-01 -7.62110591e-01 1.81469157e-01 9.04737934e-02
-6.73612118e-01 -3.73547614e-01 2.73810215e-02 9.93892133e-01
-4.37068313e-01 4.33817394e-02 5.15829682e-01 6.42331913e-02
-4.95345056e-01 1.47682384e-01 -3.86240840e-01 1.86321944e-01
7.52639711e-01 2.27111429e-01 -1.41021714e-01 -2.37739131e-01
-7.40392804e-01 5.83785102e-02 -1.49331748e-01 5.54184616e-01
2.12966874e-01 -1.43710732e+00 -6.80886507e-01 2.87327208e-02
6.01193644e-02 -2.07409337e-02 3.74695927e-01 8.21650863e-01
3.73915613e-01 3.68606001e-02 5.88161409e-01 -5.48515201e-01
-1.18875456e+00 6.30612493e-01 9.87775624e-02 -1.44777894e-01
-8.49544585e-01 1.19371223e+00 8.37001503e-01 1.81767762e-01
4.62381661e-01 -5.86192131e-01 -1.83559716e-01 1.69665664e-01
7.24246144e-01 3.45463753e-01 -8.92564729e-02 -3.59501988e-01
-5.67628741e-01 1.77961409e-01 3.89420539e-01 -4.74867187e-02
1.32949197e+00 -2.39502981e-01 -2.15746295e-02 8.52898717e-01
9.11036849e-01 2.39106655e-01 -1.67470086e+00 -2.76623785e-01
1.77786991e-01 -4.91824955e-01 2.19849929e-01 -8.25705230e-01
-1.11554098e+00 7.75745809e-01 3.58958185e-01 1.53546125e-01
1.29367936e+00 3.31526995e-01 7.70268917e-01 -2.21973225e-01
2.82996031e-03 -9.03590560e-01 3.22641730e-02 2.70503730e-01
6.58986211e-01 -9.64737594e-01 -1.85010359e-01 -7.40720928e-01
-4.13591236e-01 7.30318844e-01 4.86165792e-01 -8.82378966e-02
7.40530849e-01 5.33414662e-01 -7.92346895e-02 -2.70470977e-01
-9.65118468e-01 -2.27022395e-01 2.61488050e-01 3.13433796e-01
3.14872891e-01 1.33500531e-01 -3.21135670e-01 1.08408570e+00
8.34741369e-02 1.60065647e-02 1.41279876e-01 6.53672278e-01
-1.82965040e-01 -8.00802946e-01 -4.29228991e-01 3.74923736e-01
-6.85646296e-01 -3.59739959e-01 -2.03107759e-01 4.43923742e-01
8.23607922e-01 1.19993305e+00 1.26020610e-01 -4.38762903e-01
5.68387032e-01 9.30982605e-02 1.52805433e-01 -5.98971248e-01
-5.03857791e-01 4.31637168e-01 1.03655115e-01 -7.67275691e-01
-6.23244405e-01 -9.15652275e-01 -1.29674470e+00 2.07363442e-01
-5.03852665e-01 -4.30732518e-02 1.38783649e-01 1.25580990e+00
3.31354916e-01 9.08708870e-01 7.14900076e-01 -5.30914187e-01
3.78798763e-03 -9.09544647e-01 -4.18959737e-01 7.01720297e-01
6.25979602e-01 -1.06611788e+00 -9.35163021e-01 5.36204875e-01] | [8.894486427307129, 0.8135494589805603] |
3ac7a656-215b-4450-ae4c-efe071d7ba22 | sportsmot-a-large-multi-object-tracking | 2304.05170 | null | https://arxiv.org/abs/2304.05170v2 | https://arxiv.org/pdf/2304.05170v2.pdf | SportsMOT: A Large Multi-Object Tracking Dataset in Multiple Sports Scenes | Multi-object tracking in sports scenes plays a critical role in gathering players statistics, supporting further analysis, such as automatic tactical analysis. Yet existing MOT benchmarks cast little attention on the domain, limiting its development. In this work, we present a new large-scale multi-object tracking dataset in diverse sports scenes, coined as \emph{SportsMOT}, where all players on the court are supposed to be tracked. It consists of 240 video sequences, over 150K frames (almost 15\times MOT17) and over 1.6M bounding boxes (3\times MOT17) collected from 3 sports categories, including basketball, volleyball and football. Our dataset is characterized with two key properties: 1) fast and variable-speed motion and 2) similar yet distinguishable appearance. We expect SportsMOT to encourage the MOT trackers to promote in both motion-based association and appearance-based association. We benchmark several state-of-the-art trackers and reveal the key challenge of SportsMOT lies in object association. To alleviate the issue, we further propose a new multi-object tracking framework, termed as \emph{MixSort}, introducing a MixFormer-like structure as an auxiliary association model to prevailing tracking-by-detection trackers. By integrating the customized appearance-based association with the original motion-based association, MixSort achieves state-of-the-art performance on SportsMOT and MOT17. Based on MixSort, we give an in-depth analysis and provide some profound insights into SportsMOT. The dataset and code will be available at https://deeperaction.github.io/datasets/sportsmot.html. | ['LiMin Wang', 'Gangshan Wu', 'Yichun Yang', 'Xiaoyu Zhao', 'Chenkai Zeng', 'Yutao Cui'] | 2023-04-11 | null | null | null | null | ['multiple-object-tracking'] | ['computer-vision'] | [-2.80582309e-01 -6.74164176e-01 -4.18532073e-01 9.98418406e-02
-6.98847353e-01 -7.36525416e-01 2.69186884e-01 -3.76034491e-02
-4.87470239e-01 5.50996900e-01 1.32464439e-01 7.87186399e-02
-1.32572636e-01 -4.18069243e-01 -7.56540656e-01 -6.18343174e-01
-2.71201521e-01 5.86275041e-01 1.00249493e+00 -3.51652652e-01
-1.86304078e-02 1.27547637e-01 -1.77177823e+00 2.88116306e-01
4.46973234e-01 8.69309664e-01 1.80147544e-01 8.64151895e-01
3.01176280e-01 7.32041657e-01 -5.37939191e-01 -5.81731200e-01
5.70235729e-01 -3.37753564e-01 -3.51293981e-01 -1.57074243e-01
1.05688369e+00 -4.24515754e-02 -6.72109246e-01 8.38933587e-01
6.56704128e-01 2.91827589e-01 1.82781041e-01 -1.44966495e+00
-2.87625760e-01 5.80894649e-01 -9.38504279e-01 9.30384457e-01
4.65483576e-01 5.55196643e-01 9.66387033e-01 -7.44805038e-01
7.16696978e-01 1.14557123e+00 8.97741795e-01 5.02218664e-01
-1.02839303e+00 -1.03299546e+00 4.42758948e-01 4.21144843e-01
-1.45834172e+00 -2.51965582e-01 3.70325595e-01 -4.78835613e-01
1.12372667e-01 7.73377538e-01 1.28408957e+00 1.02510178e+00
2.23361269e-01 1.27109587e+00 1.09088659e+00 5.87467439e-05
-1.44354150e-01 -2.90344775e-01 2.87482291e-01 4.20353472e-01
4.21669036e-01 2.05024138e-01 -9.90806401e-01 1.11376770e-01
7.22428501e-01 1.97254214e-02 6.88891262e-02 -2.45784819e-01
-1.47612989e+00 4.80485708e-01 2.36301258e-01 1.27219230e-01
-1.13815993e-01 5.29454827e-01 4.63367224e-01 2.99451835e-02
3.68079543e-01 -1.57936722e-01 8.75730738e-02 -6.49350941e-01
-8.90771091e-01 7.62164235e-01 1.21559098e-01 1.03534019e+00
2.95355827e-01 7.99534544e-02 -5.80383778e-01 6.53932154e-01
1.86336294e-01 7.26191342e-01 1.98620394e-01 -9.35387492e-01
5.32508612e-01 4.54557329e-01 5.74105196e-02 -9.00835633e-01
-4.81298506e-01 -6.23345613e-01 -4.51693207e-01 2.34712780e-01
8.58972788e-01 7.50326142e-02 -5.29060483e-01 1.64098370e+00
7.73182154e-01 6.94061756e-01 -5.71299732e-01 1.31156099e+00
1.22300255e+00 1.53776467e-01 4.95121405e-02 4.07219231e-02
1.81139708e+00 -1.05269182e+00 -6.73412204e-01 -1.86352760e-01
5.48225522e-01 -7.62192607e-01 7.65256703e-01 3.77197266e-01
-1.12500691e+00 -7.94901907e-01 -7.41654813e-01 2.93930292e-01
4.39964533e-02 7.61272535e-02 6.73722208e-01 6.59548700e-01
-6.13338709e-01 2.60808855e-01 -8.23852718e-01 -2.33649582e-01
4.82164532e-01 1.83821872e-01 -2.34264642e-01 1.07346691e-01
-9.53976333e-01 5.78669071e-01 2.71835804e-01 9.29171592e-02
-9.75316942e-01 -7.28052318e-01 -6.54519558e-01 -4.15055454e-01
8.37324381e-01 -5.86455166e-01 1.17809093e+00 -4.27422076e-01
-1.06531549e+00 9.15200174e-01 -4.25589783e-03 -4.16019529e-01
8.80566895e-01 -6.47790790e-01 -5.55960953e-01 7.61650223e-03
5.21294951e-01 3.77669811e-01 6.07367516e-01 -1.05121398e+00
-1.20208645e+00 -1.13325417e-01 1.43340021e-01 1.98866010e-01
-2.44865641e-01 4.23851341e-01 -1.15006161e+00 -1.00582337e+00
-4.37495634e-02 -1.38711226e+00 -3.37759927e-02 7.80021474e-02
-5.05653262e-01 -2.02736363e-01 8.21459830e-01 -2.85528719e-01
1.58063805e+00 -2.00017047e+00 1.67640150e-01 -1.91412166e-01
5.47786415e-01 3.12211901e-01 -7.12943822e-02 1.90351173e-01
1.55283555e-01 -3.80894065e-01 3.48461568e-01 -3.24474722e-01
3.10469382e-02 1.29212901e-01 -6.94447160e-02 7.19960511e-01
-3.65837365e-01 1.11031258e+00 -1.13614500e+00 -8.87687147e-01
4.32306260e-01 8.32618624e-02 -4.07607287e-01 -1.55708686e-01
2.68132538e-02 7.95369446e-01 -4.68440652e-01 1.14443278e+00
6.68399215e-01 -1.34865101e-02 -1.48943260e-01 -1.56021789e-01
-5.04499674e-01 -1.66237727e-01 -1.66393960e+00 1.74818063e+00
2.70271510e-01 7.49198139e-01 1.58289418e-01 -6.86360896e-01
5.51279247e-01 4.03165780e-02 1.07228231e+00 -6.69486642e-01
1.53374240e-01 6.13071863e-03 2.33398482e-01 -2.71782547e-01
9.47348893e-01 9.66260433e-02 -2.61112392e-01 3.46020073e-01
-7.61496723e-02 6.36466026e-01 7.96312988e-01 4.20946032e-01
1.10401464e+00 4.78175402e-01 -8.90997425e-02 -2.06085637e-01
1.89541280e-01 1.99979335e-01 1.04233515e+00 1.20650005e+00
-6.24593735e-01 5.32486439e-01 -2.69727223e-02 -6.71126306e-01
-6.33627892e-01 -1.21795845e+00 -8.01860094e-02 1.44185674e+00
6.65872693e-01 -7.52452731e-01 -4.03034925e-01 -5.42081475e-01
2.77624488e-01 -1.99691907e-01 -7.72154689e-01 5.50896525e-02
-1.06472135e+00 -9.03156340e-01 8.15593839e-01 7.44690180e-01
4.76374269e-01 -9.09378469e-01 -6.75231516e-01 3.26875836e-01
-4.73435760e-01 -1.39454556e+00 -8.37022901e-01 -3.58846247e-01
-6.97472990e-01 -1.18066204e+00 -8.79669428e-01 -2.71732956e-01
-4.39033657e-02 6.92269325e-01 1.21030307e+00 2.88554311e-01
-5.49890697e-01 4.05105144e-01 -5.27256966e-01 -5.65559864e-01
1.70734897e-02 -6.26416951e-02 5.49248338e-01 2.30284840e-01
2.63708979e-01 -4.47204053e-01 -7.49120295e-01 9.07571673e-01
-5.66372931e-01 1.47998720e-01 5.83121777e-01 3.49403650e-01
6.83670998e-01 -3.16693217e-01 4.21334170e-02 -4.32872772e-01
6.01240396e-02 -3.84325325e-01 -5.48934460e-01 9.18249264e-02
-5.44597255e-03 -4.90223855e-01 2.60272250e-02 -1.06131291e+00
-5.02769947e-01 -5.48518961e-03 5.89906424e-02 -7.08981752e-01
2.22048536e-03 -1.17425155e-02 8.78429264e-02 -3.44745576e-01
5.36291838e-01 2.37951159e-01 -3.64923596e-01 -6.05519235e-01
2.04046875e-01 1.16240278e-01 9.24079239e-01 -8.03943753e-01
1.16122949e+00 8.34770977e-01 7.29146525e-02 -6.16235197e-01
-1.07191873e+00 -9.07214046e-01 -5.99502265e-01 -1.05265677e+00
9.50933278e-01 -1.09196901e+00 -1.31180000e+00 4.67560112e-01
-6.41309381e-01 -3.39525521e-01 -3.74895394e-01 8.39925468e-01
-3.66177022e-01 4.45415974e-01 -5.77745199e-01 -9.30553973e-01
-8.05043206e-02 -9.89033639e-01 1.19600022e+00 4.09299791e-01
-1.32725075e-01 -6.62423968e-01 3.59795868e-01 7.38882065e-01
1.86154507e-02 4.39490765e-01 -3.21776420e-01 -4.16532218e-01
-8.48729253e-01 -2.04088166e-01 -7.17190579e-02 -2.82815337e-01
-2.33636409e-01 -1.57424599e-01 -7.34516382e-01 -4.16163683e-01
-6.20533466e-01 -7.08061159e-02 9.58513677e-01 7.34534264e-01
9.73539948e-01 3.89560044e-01 -5.63242972e-01 7.65745521e-01
9.81546700e-01 -1.09150365e-01 3.99516046e-01 8.46522212e-01
9.67864573e-01 1.92302763e-01 1.22190881e+00 5.22789598e-01
5.50082922e-01 1.42967510e+00 5.65518916e-01 4.52231541e-02
-4.94672060e-01 -1.27409384e-01 5.84447622e-01 7.68362463e-01
-9.66098726e-01 -1.59182891e-01 -7.75780499e-01 5.82282603e-01
-2.27569294e+00 -1.51457679e+00 -7.05439329e-01 2.07295585e+00
4.70689565e-01 3.67091119e-01 1.03218043e+00 5.32890782e-02
1.01558852e+00 1.39674351e-01 -4.12523568e-01 3.72532845e-01
-2.62403607e-01 -8.47942233e-02 6.25596106e-01 8.30426663e-02
-1.45627201e+00 8.71592283e-01 5.53661776e+00 1.36336040e+00
-7.09109664e-01 3.38493466e-01 -1.30720004e-01 -5.61626017e-01
3.42269033e-01 9.98322815e-02 -1.43892407e+00 8.15531611e-01
4.90724176e-01 2.13867072e-02 -9.22795907e-02 6.44332826e-01
2.22447932e-01 -2.53588587e-01 -6.21890843e-01 1.11108208e+00
-1.34072334e-01 -1.46285999e+00 -2.38547474e-01 4.10807550e-01
4.06584918e-01 -4.75115664e-02 4.42065857e-02 5.71734428e-01
4.14148271e-01 -5.36470056e-01 1.36458373e+00 4.37673450e-01
6.19500577e-01 -2.47970015e-01 4.58490342e-01 2.99074262e-01
-2.23391986e+00 -1.78021699e-01 -9.43604186e-02 -1.93595171e-01
6.03616416e-01 3.11302155e-01 -1.62100643e-02 8.83257806e-01
1.26279521e+00 1.18383026e+00 -7.08776474e-01 1.55249441e+00
1.71352923e-01 6.16074264e-01 -4.44859445e-01 1.13408536e-01
1.69155389e-01 -1.53018087e-01 1.07486677e+00 1.33320677e+00
1.89289793e-01 5.76213002e-02 8.21804643e-01 3.80924493e-01
2.15528786e-01 -1.49472952e-01 -1.63829505e-01 3.52479577e-01
4.07058179e-01 1.50857186e+00 -1.02745473e+00 -4.90985066e-01
-4.76605237e-01 4.72525865e-01 -2.31570080e-02 -7.82024488e-02
-1.51360440e+00 2.07619816e-01 1.03543055e+00 5.34625649e-01
3.75235677e-01 -2.01978892e-01 1.31167332e-02 -1.32403970e+00
2.03964502e-01 -8.68343592e-01 8.30696821e-01 -5.71961761e-01
-1.03644025e+00 3.63877356e-01 3.43580037e-01 -1.92940855e+00
3.60106587e-01 -3.69248390e-01 -6.89492881e-01 2.90779561e-01
-1.10947526e+00 -1.43379962e+00 -6.24200284e-01 7.07525432e-01
6.98080063e-01 -1.23853698e-01 8.49718079e-02 9.49288070e-01
-9.21713650e-01 6.57284915e-01 -1.06794871e-01 4.39122051e-01
8.58548939e-01 -1.10797107e+00 2.39298806e-01 9.86343265e-01
4.41420645e-01 2.81076014e-01 8.88280988e-01 -9.03717935e-01
-1.55010331e+00 -1.05119872e+00 1.41594961e-01 -1.03521419e+00
8.19694817e-01 -4.77247179e-01 -5.79643846e-01 8.10346544e-01
-1.49328142e-01 4.83522862e-01 6.34404361e-01 2.27345616e-01
2.50503961e-02 -1.93611369e-01 -4.59154189e-01 6.35284185e-01
1.63704479e+00 1.39918774e-01 -4.30215418e-01 2.53524393e-01
2.95608252e-01 -9.69833493e-01 -1.00566518e+00 3.29510778e-01
9.37347472e-01 -9.69473004e-01 1.39894485e+00 -4.26861703e-01
-1.89449206e-01 -6.78438008e-01 -2.70436555e-02 -6.51602089e-01
-5.31232715e-01 -7.87105620e-01 -5.49098372e-01 1.18939209e+00
-2.03862488e-01 -2.68319517e-01 1.10694432e+00 9.12218019e-02
-5.09212911e-01 -6.00478470e-01 -1.24807274e+00 -1.33456016e+00
-2.58224308e-01 -7.40821064e-01 2.89893210e-01 7.40557551e-01
-2.32862145e-01 -8.38108286e-02 -8.45823705e-01 1.04577705e-01
1.11244750e+00 4.02536243e-01 1.38233066e+00 -1.32304895e+00
-4.06016380e-01 -4.84706700e-01 -7.14156091e-01 -1.34398949e+00
-4.73700017e-01 -6.15175784e-01 -1.28126115e-01 -1.07678223e+00
5.50277293e-01 -6.73847556e-01 -2.41507784e-01 4.86238211e-01
-3.39097977e-01 9.84871805e-01 6.76921368e-01 4.96562988e-01
-1.37875450e+00 2.98321337e-01 1.50827920e+00 -3.21590118e-02
1.69881079e-02 1.96942538e-01 -4.74838138e-01 7.82165706e-01
4.92248714e-01 -7.39859283e-01 2.52998322e-01 -2.46077836e-01
1.15785390e-01 3.47163193e-02 7.58647144e-01 -1.50673056e+00
3.35540593e-01 -3.35600197e-01 2.05738485e-01 -9.44677114e-01
7.09376156e-01 -3.93699020e-01 4.66839433e-01 6.66474700e-01
7.27661774e-02 2.74409711e-01 4.15559947e-01 5.66172719e-01
-1.47800758e-01 1.64231539e-01 4.89018291e-01 -5.81548214e-02
-1.07500529e+00 7.40999341e-01 -2.40735084e-01 4.33284968e-01
1.27716136e+00 -6.68389797e-01 -4.22371596e-01 -2.12795854e-01
-9.42127645e-01 4.95945632e-01 2.77227342e-01 7.00171947e-01
3.46482426e-01 -1.67663336e+00 -1.08917248e+00 -3.03440630e-01
2.87157059e-01 -3.80163252e-01 5.88921785e-01 1.55453813e+00
-2.51427591e-01 1.75125018e-01 -3.03929508e-01 -1.13818407e+00
-1.82353938e+00 3.36105227e-01 -6.85891416e-03 -4.19179469e-01
-1.07471311e+00 6.96194768e-01 1.99235857e-01 6.87081292e-02
7.19113275e-02 -2.63119549e-01 -1.71542421e-01 5.74374478e-03
7.56341815e-01 5.95950186e-01 -3.38024855e-01 -1.05357492e+00
-5.46161115e-01 9.36648488e-01 5.46527095e-02 4.53487039e-02
1.08145058e+00 -1.23517156e-01 4.30077136e-01 3.68024439e-01
2.82516599e-01 2.74513572e-01 -1.46570241e+00 -2.70095587e-01
-1.96396112e-01 -9.08421874e-01 -3.04795206e-01 -2.94401348e-01
-1.23529053e+00 5.42029321e-01 7.45012403e-01 2.86361337e-01
8.09424341e-01 2.85257488e-01 1.03569818e+00 -1.87475160e-01
5.86551964e-01 -1.03625095e+00 3.54933500e-01 3.87716442e-01
4.52647626e-01 -1.24400246e+00 3.39351803e-01 -4.91520315e-01
-6.40420437e-01 7.22105145e-01 9.22429562e-01 -1.59497917e-01
2.57142931e-01 4.07998979e-01 7.12886676e-02 -4.33056742e-01
-5.76030850e-01 -7.17602968e-01 5.58820963e-01 5.16273201e-01
2.27048218e-01 -2.51126364e-02 -2.24842429e-01 6.06071115e-01
-1.31307200e-01 -1.10447705e-01 2.04320922e-01 1.04741907e+00
-4.52105224e-01 -1.11764503e+00 -9.29711699e-01 2.80966610e-01
-4.53869164e-01 2.35975593e-01 -9.71275289e-03 1.05287409e+00
6.16252601e-01 7.22881794e-01 -6.45573139e-02 -6.03827596e-01
5.20786345e-01 -6.26956105e-01 5.74389577e-01 -4.08396542e-01
-1.00195193e+00 2.95130104e-01 1.45936489e-01 -8.74630749e-01
-7.22390890e-01 -1.09963930e+00 -9.06584144e-01 -6.39933109e-01
-4.14210916e-01 1.84882097e-02 1.85745612e-01 8.25130165e-01
1.63130425e-02 7.77730644e-01 -3.85919511e-02 -1.21546352e+00
1.33493379e-01 -8.66145909e-01 -5.79201937e-01 6.52094066e-01
1.65247217e-01 -1.34798110e+00 1.26793012e-01 -5.36327623e-02] | [6.36254358291626, -2.0535950660705566] |
47d481e6-5ea8-48ce-8ba5-098ce0b9b410 | a-multilingual-evaluation-of-ner-robustness | 2305.18933 | null | https://arxiv.org/abs/2305.18933v1 | https://arxiv.org/pdf/2305.18933v1.pdf | A Multilingual Evaluation of NER Robustness to Adversarial Inputs | Adversarial evaluations of language models typically focus on English alone. In this paper, we performed a multilingual evaluation of Named Entity Recognition (NER) in terms of its robustness to small perturbations in the input. Our results showed the NER models we explored across three languages (English, German and Hindi) are not very robust to such changes, as indicated by the fluctuations in the overall F1 score as well as in a more fine-grained evaluation. With that knowledge, we further explored whether it is possible to improve the existing NER models using a part of the generated adversarial data sets as augmented training data to train a new NER model or as fine-tuning data to adapt an existing NER model. Our results showed that both these approaches improve performance on the original as well as adversarial test sets. While there is no significant difference between the two approaches for English, re-training is significantly better than fine-tuning for German and Hindi. | ['Sowmya Vajjala', 'Akshay Srinivasan'] | 2023-05-30 | null | null | null | null | ['named-entity-recognition-ner'] | ['natural-language-processing'] | [-2.41490349e-01 1.72556981e-01 4.33629602e-01 -2.73798585e-01
-1.01627743e+00 -1.22992146e+00 8.00911605e-01 1.21504746e-01
-1.03371716e+00 1.08771837e+00 5.08692443e-01 -3.87117624e-01
2.34100863e-01 -7.84694016e-01 -7.12285519e-01 -2.57034868e-01
5.96911311e-02 3.95105213e-01 2.86855549e-01 -6.58818543e-01
-1.72260866e-01 7.58388340e-01 -8.61002922e-01 8.31594244e-02
6.91066921e-01 6.48867413e-02 -5.42095900e-01 7.72097826e-01
-4.03082296e-02 5.98253548e-01 -9.54436302e-01 -8.89651060e-01
3.41288298e-01 -2.42204741e-01 -9.01239455e-01 -5.98461092e-01
3.06166738e-01 -2.05985326e-02 -3.61713558e-01 9.96203959e-01
9.03186023e-01 2.73528248e-01 3.72575670e-01 -9.36950207e-01
-5.71137011e-01 8.32883835e-01 1.30307794e-01 2.55718708e-01
4.63967472e-01 2.01900259e-01 5.29657543e-01 -7.37492621e-01
8.45610440e-01 1.07166374e+00 1.18118942e+00 7.34736264e-01
-1.22701585e+00 -5.67910969e-01 -1.51881382e-01 -3.98538113e-01
-1.45473421e+00 -5.49476564e-01 2.85892814e-01 -1.23212077e-01
1.12345159e+00 2.20816657e-01 3.15662543e-03 1.44938362e+00
2.39854887e-01 2.04305932e-01 1.23585880e+00 -5.64165115e-01
2.22473219e-01 4.18093592e-01 -2.12563157e-01 2.84925312e-01
2.27414951e-01 4.76062328e-01 6.36839494e-02 -3.69816542e-01
3.55920702e-01 -5.90559185e-01 -5.30016758e-02 9.57929268e-02
-1.21448886e+00 7.74953544e-01 3.22207272e-01 7.87065506e-01
-3.87474209e-01 -1.27343759e-01 6.14022195e-01 5.21434486e-01
3.43421787e-01 1.06454420e+00 -8.23037386e-01 -3.62655073e-01
-1.03966248e+00 1.55937701e-01 1.17768681e+00 5.73147237e-01
4.58447695e-01 4.47111934e-01 -2.11050317e-01 7.89560854e-01
-1.63319454e-01 5.01823485e-01 7.22485363e-01 -5.67328632e-01
6.21778548e-01 1.95150301e-01 2.02819869e-01 -7.92187035e-01
-4.89794254e-01 -2.77891487e-01 -4.83448267e-01 3.53082657e-01
6.36133969e-01 -7.51225770e-01 -1.18508041e+00 2.02063537e+00
1.66788176e-01 -1.00917488e-01 4.90772814e-01 4.55674559e-01
7.71430433e-01 4.87501472e-01 6.68065131e-01 2.78395385e-01
1.09728086e+00 -6.51316643e-01 -4.94307905e-01 -3.94383222e-01
7.88680136e-01 -1.04107344e+00 9.44795787e-01 6.88936785e-02
-9.37701523e-01 -6.30050123e-01 -9.62367713e-01 1.36909112e-01
-9.58301842e-01 -2.80668885e-01 2.62499839e-01 1.05338669e+00
-1.10283172e+00 7.06681728e-01 -7.24539518e-01 -6.39599323e-01
-1.17320366e-01 3.09625089e-01 -9.48588610e-01 -4.18959856e-02
-1.76224458e+00 1.31087792e+00 6.45330608e-01 -2.77027935e-01
-6.73250973e-01 -5.94325602e-01 -1.02776992e+00 -7.81077966e-02
-9.24120173e-02 -3.15758377e-01 1.23115122e+00 -7.95701325e-01
-1.42707288e+00 8.28717411e-01 3.00225943e-01 -4.34641093e-01
7.17391193e-01 -1.33075014e-01 -9.49813366e-01 -3.17981809e-01
6.86351815e-03 6.01529360e-01 3.17977428e-01 -1.16113985e+00
-3.74224573e-01 -3.98219489e-02 2.07848608e-01 1.11194802e-02
-2.17028305e-01 4.80564505e-01 -8.80380347e-02 -9.93606269e-01
-5.68141282e-01 -1.01716602e+00 -4.64079469e-01 -7.77517378e-01
-5.18840611e-01 2.42221072e-01 4.25275892e-01 -9.61713016e-01
1.25163805e+00 -2.03298783e+00 -2.50932723e-01 2.64798880e-01
-2.11307585e-01 7.40580380e-01 -4.54070359e-01 6.74668372e-01
-4.85611796e-01 8.11280251e-01 -1.79039016e-01 -1.50754631e-01
1.15095399e-01 2.44764328e-01 -1.95533261e-01 2.83387661e-01
3.91956717e-01 7.87565470e-01 -9.53109205e-01 -2.02551916e-01
3.62154171e-02 6.95861340e-01 -3.39437872e-01 2.25284815e-01
2.27884293e-01 4.61620420e-01 -2.08875075e-01 2.19953403e-01
5.14527857e-01 5.36835611e-01 -5.02981842e-02 8.59828014e-03
-1.56366080e-01 3.28571409e-01 -1.23289979e+00 1.35498130e+00
-5.89901924e-01 4.97138232e-01 -1.95357367e-01 -5.61451018e-01
8.29499543e-01 7.05002606e-01 8.88995901e-02 -6.58837378e-01
-6.86907247e-02 3.60671848e-01 6.73172772e-02 -1.38433814e-01
6.02870822e-01 -5.14651597e-01 -6.75734341e-01 2.83348650e-01
2.91479468e-01 -1.00194156e-01 2.97336906e-01 -6.47017285e-02
1.36700737e+00 -3.85096520e-02 3.86435121e-01 -1.10946476e-01
5.33090591e-01 1.57871604e-01 5.49087346e-01 8.38461578e-01
-3.15191478e-01 6.52367771e-01 1.15103148e-01 -1.59121558e-01
-1.24963486e+00 -1.16471052e+00 -1.40168011e-01 9.34307814e-01
-3.98225039e-01 -5.36672533e-01 -1.08782089e+00 -1.05344760e+00
-3.07766110e-01 1.19999981e+00 -5.72435141e-01 -3.27441841e-01
-6.38833284e-01 -8.05178404e-01 1.60365582e+00 5.96157491e-01
4.51636881e-01 -1.34787524e+00 -2.20304534e-01 3.72899204e-01
-8.30205902e-02 -1.18777990e+00 -4.69650656e-01 4.40904230e-01
-5.80116272e-01 -6.47832155e-01 -6.99652374e-01 -7.15741932e-01
3.32675368e-01 -6.90322161e-01 1.29861820e+00 -2.83374399e-01
2.28103474e-02 3.86945218e-01 -4.16786939e-01 -5.41632056e-01
-1.12626421e+00 4.04448599e-01 2.38281488e-01 -5.34798145e-01
4.15788442e-01 -3.26318979e-01 -1.28790960e-01 3.15655828e-01
-1.17726171e+00 -7.87069619e-01 5.00899553e-01 7.27717340e-01
2.95787603e-01 2.18839467e-01 5.96874535e-01 -1.18548429e+00
8.63144815e-01 -4.77759570e-01 -2.98328727e-01 3.09391201e-01
-4.67354536e-01 3.70348603e-01 1.01008022e+00 -4.80461270e-01
-1.10095644e+00 -8.30491409e-02 -7.03972816e-01 -1.38405934e-02
-6.54110491e-01 5.26975572e-01 -3.69102925e-01 -1.53547168e-01
1.06483495e+00 -2.62078494e-01 -6.26060247e-01 -6.35154545e-01
4.14802223e-01 6.12104535e-01 6.03196561e-01 -5.79666138e-01
1.18600738e+00 -4.43375707e-02 -4.35272425e-01 -6.25559628e-01
-4.25752789e-01 -1.15253508e-01 -7.17128098e-01 2.99936503e-01
8.91283214e-01 -8.04522991e-01 -1.55690906e-03 5.22861719e-01
-1.11810887e+00 -4.34410721e-01 -5.07696331e-01 5.57331800e-01
-1.55776769e-01 2.43693113e-01 -8.29023659e-01 -2.66106039e-01
-3.27733487e-01 -1.04696405e+00 7.07017183e-01 1.30882069e-01
-4.82383102e-01 -1.54192221e+00 5.62383354e-01 -8.30792338e-02
7.22943008e-01 6.23480260e-01 6.68284416e-01 -1.44675672e+00
3.77238512e-01 -6.36628926e-01 3.23145598e-01 6.15711927e-01
1.70944571e-01 -1.62909683e-02 -8.40135992e-01 -4.36869770e-01
-1.02285288e-01 -3.32175106e-01 3.92509192e-01 -3.58413309e-01
2.99350411e-01 -1.80899486e-01 7.96767231e-03 5.07969141e-01
1.42441750e+00 9.36135426e-02 9.29013252e-01 7.73551285e-01
5.89634478e-01 3.46996635e-01 4.09997374e-01 -4.16230559e-02
-9.75248695e-04 6.28510356e-01 2.77905371e-02 -3.92483473e-01
-9.38836709e-02 -4.38468814e-01 6.08666062e-01 6.40337288e-01
2.40861028e-02 -4.07539457e-01 -1.28150558e+00 6.11746848e-01
-1.27560961e+00 -1.07560635e+00 2.10349143e-01 2.30870914e+00
9.84700263e-01 2.93224782e-01 1.12359552e-02 -1.42537341e-01
8.47811043e-01 -3.48412916e-02 -1.80324689e-01 -9.21463192e-01
-2.60191798e-01 7.28536129e-01 7.87597597e-01 5.58065176e-01
-1.27877355e+00 1.17289829e+00 7.20339394e+00 5.92535496e-01
-1.17236376e+00 1.44593298e-01 3.17283601e-01 2.47760072e-01
-2.14901611e-01 3.64815034e-02 -7.55579472e-01 3.30709368e-01
1.56839180e+00 -2.83454329e-01 4.48973328e-01 6.38603389e-01
-3.13980132e-02 4.58067805e-01 -8.69126379e-01 3.92062992e-01
-1.28735065e-01 -8.30120862e-01 3.89835946e-02 -2.02414095e-02
7.30401278e-01 3.78068179e-01 -3.50479811e-01 7.91241407e-01
9.62679923e-01 -1.29540527e+00 5.68434358e-01 4.42548662e-01
7.98135340e-01 -9.37557995e-01 1.15766668e+00 3.47398192e-01
-8.83265018e-01 4.05495167e-01 -2.53116578e-01 2.81305552e-01
2.50144094e-01 1.33957550e-01 -9.32366550e-01 6.49248719e-01
5.76415777e-01 -7.72199258e-02 -9.30508912e-01 8.10781300e-01
-4.20747846e-01 8.29091012e-01 -4.34521317e-01 3.94862890e-01
3.33395094e-01 3.85165185e-01 6.29153430e-01 1.63919342e+00
2.15582430e-01 -2.41874456e-01 1.47623077e-01 3.74782950e-01
-2.48883784e-01 2.35871270e-01 -6.52851641e-01 -2.98290670e-01
4.48539823e-01 1.17794597e+00 -3.91194195e-01 -1.63131192e-01
-4.43281174e-01 1.16088235e+00 3.75283748e-01 3.93241197e-01
-9.43394959e-01 -7.99280941e-01 5.39420307e-01 1.15454420e-02
3.60002726e-01 -8.77964869e-02 5.77697940e-02 -1.28834403e+00
-2.35261604e-01 -1.23787916e+00 5.51084280e-01 -6.45540476e-01
-1.33108568e+00 1.17467439e+00 -9.48857889e-02 -8.23125660e-01
-5.26963174e-01 -6.09234989e-01 -7.27179468e-01 1.22796643e+00
-1.08190370e+00 -1.03961074e+00 6.97391778e-02 6.80582821e-01
-4.77145845e-03 -2.51509637e-01 1.35370529e+00 3.88187021e-01
-4.27920967e-01 1.23407781e+00 2.20579505e-01 7.12338150e-01
1.07162082e+00 -1.41673315e+00 9.56772149e-01 1.30198991e+00
3.86610508e-01 7.15305626e-01 9.41692173e-01 -6.78767979e-01
-5.55419385e-01 -1.12640679e+00 1.16045403e+00 -7.72665441e-01
7.23894894e-01 -1.90015018e-01 -9.83316243e-01 8.79090607e-01
3.99048567e-01 5.05960248e-02 8.31779659e-01 6.25497475e-02
-5.14415085e-01 3.83092731e-01 -1.52601433e+00 6.18617117e-01
7.53585875e-01 -6.09467328e-01 -9.29514110e-01 -2.01383489e-03
7.09657371e-01 -3.89506787e-01 -1.46770227e+00 4.65138197e-01
2.63484180e-01 -7.56271839e-01 9.02493775e-01 -1.15807116e+00
1.40165135e-01 -2.01339677e-01 -3.45494539e-01 -1.75068343e+00
-3.11581343e-01 -3.29397172e-01 4.81235713e-01 1.80085182e+00
7.73382187e-01 -7.94346511e-01 4.43246692e-01 7.28254020e-01
-1.37210218e-02 -1.63411319e-01 -8.07954371e-01 -8.81739318e-01
7.37076938e-01 -4.54277635e-01 6.51055753e-01 1.18268871e+00
-1.22485332e-01 3.43917519e-01 -1.92610100e-01 3.54397774e-01
1.49228334e-01 -7.31348336e-01 7.57547677e-01 -9.00791228e-01
-2.90534288e-01 -1.34066164e-01 -7.23177731e-01 -5.65788448e-02
3.45633179e-01 -9.09087241e-01 9.66809466e-02 -1.43626225e+00
-2.37972304e-01 -4.17542219e-01 -4.59969372e-01 8.18417370e-01
-3.53532165e-01 3.78202587e-01 3.67833227e-01 -5.40185608e-02
-1.99922919e-01 1.04034938e-01 6.18530810e-01 8.97558108e-02
-9.95632783e-02 -7.75015727e-02 -8.55444193e-01 6.65934801e-01
8.75691891e-01 -7.94389248e-01 -5.74190430e-02 -4.35258865e-01
1.89228997e-01 -2.46760085e-01 1.59735277e-01 -1.25456369e+00
9.40135389e-04 1.56420425e-01 5.07180154e-01 1.65740922e-01
-1.33202508e-01 -6.73980296e-01 2.88855940e-01 2.98745036e-01
-3.13759774e-01 3.54731321e-01 8.05761516e-01 2.20701620e-01
-3.15952510e-01 -3.84180576e-01 8.85220349e-01 -1.00590169e-01
-8.26948941e-01 -2.24347711e-02 -4.22647327e-01 5.29009998e-01
8.34860981e-01 -1.30940065e-01 -1.72675446e-01 -3.59944344e-01
-9.20558214e-01 -8.79898295e-02 7.48117745e-01 6.35756075e-01
-9.73901749e-02 -1.23631239e+00 -1.03718412e+00 1.82506919e-01
7.80691132e-02 -4.96611357e-01 -2.49488149e-02 2.30975866e-01
-5.20722687e-01 3.16484541e-01 -3.38574558e-01 1.16914652e-01
-1.04309905e+00 7.35488594e-01 6.51252449e-01 -8.84821892e-01
-9.70511809e-02 5.40096819e-01 -1.49022222e-01 -1.24175501e+00
8.80943239e-02 9.76110995e-02 -2.34221622e-01 -1.51121989e-01
3.20252001e-01 1.96327984e-01 3.49924892e-01 -9.71391618e-01
-6.35327756e-01 2.85939008e-01 -1.12952232e-01 -3.54934573e-01
1.19419324e+00 1.77737340e-01 2.38714024e-01 3.35276097e-01
1.19651115e+00 7.95459926e-01 -7.90442288e-01 1.15319872e-02
1.54102981e-01 -1.36659602e-02 -1.77609324e-01 -1.21758950e+00
-8.07723701e-01 5.33101141e-01 6.31598055e-01 2.46383950e-01
9.54569340e-01 -2.90833145e-01 7.16553867e-01 5.11553288e-01
3.93847317e-01 -9.89420712e-01 -5.56527674e-01 8.61431301e-01
6.23345077e-01 -1.03292000e+00 -2.46480361e-01 1.55128352e-03
-7.19506443e-01 1.00650930e+00 4.88230854e-01 -8.72860774e-02
5.79351366e-01 5.28171599e-01 4.43521887e-01 2.11017311e-01
-2.85527021e-01 -1.03191696e-01 1.07480668e-01 7.28648663e-01
7.73826301e-01 -8.75528082e-02 -3.45365375e-01 5.57756782e-01
-5.74354172e-01 -1.53067544e-01 5.31874716e-01 8.91256630e-01
9.30933133e-02 -1.46037483e+00 -4.89680767e-01 2.71988753e-02
-1.04402149e+00 -3.68224531e-01 -5.52257121e-01 1.27260661e+00
7.94022381e-02 1.03232682e+00 -2.46947989e-01 -6.10303342e-01
8.57704461e-01 4.33983654e-01 8.22262540e-02 -6.52661741e-01
-1.16222131e+00 -5.20250142e-01 5.33099174e-01 -3.93631339e-01
-1.46469548e-01 -5.52957475e-01 -1.17741883e+00 -4.26336378e-01
-2.13571683e-01 5.39912999e-01 6.59764111e-01 9.99037027e-01
2.87926406e-01 4.93091851e-01 4.08537030e-01 -5.70380032e-01
-8.67619038e-01 -1.11294138e+00 -3.81868511e-01 7.52812207e-01
-2.27650702e-02 -2.03844503e-01 -5.65279543e-01 -2.20542271e-02] | [9.841837882995605, 9.695231437683105] |
2d878105-a086-44fc-880e-0c4a6e50fefc | skin-cancer-reorganization-and-classification | 1703.00534 | null | http://arxiv.org/abs/1703.00534v1 | http://arxiv.org/pdf/1703.00534v1.pdf | Skin cancer reorganization and classification with deep neural network | As one kind of skin cancer, melanoma is very dangerous. Dermoscopy based
early detection and recarbonization strategy is critical for melanoma therapy.
However, well-trained dermatologists dominant the diagnostic accuracy. In order
to solve this problem, many effort focus on developing automatic image analysis
systems. Here we report a novel strategy based on deep learning technique, and
achieve very high skin lesion segmentation and melanoma diagnosis accuracy: 1)
we build a segmentation neural network (skin_segnn), which achieved very high
lesion boundary detection accuracy; 2) We build another very deep neural
network based on Google inception v3 network (skin_recnn) and its well-trained
weight. The novel designed transfer learning based deep neural network
skin_inceptions_v3_nn helps to achieve a high prediction accuracy. | ['Hao Chang'] | 2017-03-01 | null | null | null | null | ['melanoma-diagnosis', 'skin-lesion-segmentation'] | ['computer-vision', 'medical'] | [ 2.53049701e-01 -1.00969456e-01 -2.40100771e-01 4.99956980e-02
-2.94930756e-01 -1.00065820e-01 2.06803337e-01 -1.90681927e-02
-4.09546643e-01 5.45413852e-01 -2.04810172e-01 -5.34919858e-01
1.05844662e-01 -1.10367560e+00 -1.65650845e-01 -8.14395964e-01
2.82354116e-01 3.36695239e-02 4.53449935e-01 -8.45978260e-02
3.46422642e-01 4.53828573e-01 -7.89841712e-01 1.86363012e-01
1.43383563e+00 9.67758358e-01 1.35007739e-01 1.09438455e+00
-3.44738841e-01 6.69530571e-01 -3.41407031e-01 -1.77978680e-01
2.09512532e-01 -3.28588188e-01 -9.09857035e-01 -2.24806983e-02
1.85745396e-02 -3.52040857e-01 1.72730163e-01 1.12990153e+00
5.00990510e-01 -3.76986384e-01 9.30155754e-01 -7.89705038e-01
-6.92051053e-01 1.30586177e-01 -8.73151660e-01 9.72923860e-02
4.39606048e-02 2.17222556e-01 5.70437536e-02 -1.61769316e-01
4.95353639e-01 5.69709778e-01 1.28879690e+00 9.39325333e-01
-4.88520056e-01 -4.24360901e-01 -1.77428082e-01 3.36598605e-01
-1.18025041e+00 1.54823467e-01 4.23859090e-01 -4.93985593e-01
5.75979769e-01 5.09645641e-01 9.92980421e-01 9.67730224e-01
6.07138157e-01 6.12486124e-01 1.44231236e+00 -4.13538843e-01
-2.55536977e-02 3.68604541e-01 2.40919694e-01 9.20313478e-01
1.54551193e-01 -5.32518290e-02 2.44457617e-01 3.48431051e-01
1.23766351e+00 2.40863338e-01 -1.61084607e-01 5.04869103e-01
-4.55875903e-01 5.10218978e-01 8.32594454e-01 4.58637387e-01
-3.77402574e-01 1.85812622e-01 4.48260397e-01 1.59138456e-01
4.04080361e-01 1.21338904e-01 -1.59578815e-01 5.97069366e-03
-5.82010865e-01 -4.07469928e-01 6.28546298e-01 6.06217571e-02
4.22756046e-01 -1.10633001e-02 -1.30175844e-01 1.02613807e+00
2.01663747e-01 2.53665745e-01 6.89849496e-01 -2.76194185e-01
-1.27784595e-01 9.30904686e-01 -2.50086576e-01 -6.20065153e-01
-6.09216809e-01 -2.54579663e-01 -1.12132382e+00 5.33604324e-01
3.38799864e-01 -5.17867327e-01 -1.50440216e+00 1.00725174e+00
4.34209079e-01 1.89056277e-01 1.17010690e-01 9.09484863e-01
8.64744425e-01 5.74293673e-01 5.59452534e-01 1.08951181e-01
1.24808991e+00 -1.05032194e+00 -5.05558431e-01 1.83893129e-01
6.30421579e-01 -6.66837335e-01 6.88198268e-01 7.03999281e-01
-7.40056574e-01 -5.00095487e-01 -7.92541981e-01 -3.75108384e-02
-6.75438225e-01 2.49321118e-01 9.50278938e-01 9.77112353e-01
-1.05671895e+00 6.86523914e-01 -8.29210103e-01 -9.77991283e-01
6.39587104e-01 4.82483745e-01 -3.58192354e-01 3.22719643e-05
-9.42465186e-01 9.82085884e-01 3.35122198e-01 2.68766761e-01
-7.04694927e-01 -4.16733742e-01 -5.57935596e-01 -3.89878780e-01
5.67576736e-02 -7.34661043e-01 9.86442685e-01 -1.35054147e+00
-1.57459414e+00 1.03043818e+00 1.48193717e-01 -5.28368950e-01
5.28689265e-01 6.19839877e-02 -5.89060247e-01 3.66334856e-01
-3.58530641e-01 5.57584226e-01 7.58192241e-01 -1.12514257e+00
-7.05745041e-01 -5.28996229e-01 -2.47332975e-01 3.72606367e-01
-7.00974107e-01 -7.42327422e-02 -2.81354457e-01 -2.02515826e-01
-2.20353246e-01 -6.05763257e-01 -5.10257840e-01 3.33378196e-01
-6.61987424e-01 -2.96409428e-01 6.66650534e-01 -1.23572898e+00
1.13961422e+00 -1.76450086e+00 -2.56049514e-01 3.13963234e-01
3.14376205e-01 9.37088370e-01 -6.93005230e-03 1.85276538e-01
-4.14857976e-02 4.42363888e-01 -2.21488222e-01 2.80453950e-01
-6.91025496e-01 -9.40404609e-02 5.21306813e-01 1.95602924e-01
2.33837754e-01 7.67305613e-01 -7.57731497e-01 -7.76123226e-01
5.60394585e-01 4.55641270e-01 -3.50136571e-02 9.36860889e-02
2.67665926e-03 1.33602276e-01 -4.64776367e-01 1.04234767e+00
9.91207719e-01 -1.26370966e-01 -3.95241529e-02 -3.04661840e-01
-3.46937403e-02 -7.56757498e-01 -6.31201863e-01 1.37991071e+00
-5.50085545e-01 4.48986739e-01 1.87497973e-01 -8.61924946e-01
9.27953780e-01 2.29990989e-01 5.86718380e-01 -5.89895964e-01
5.46333671e-01 5.47944084e-02 5.93849197e-02 -1.15242219e+00
-6.23757690e-02 -3.32392395e-01 4.47714299e-01 -1.82814226e-01
-2.88440645e-01 1.59369409e-02 -3.19612563e-01 -5.26419319e-02
1.08886778e+00 5.87593280e-02 2.64977038e-01 -4.24498357e-02
7.28546679e-01 3.93635660e-01 3.90963584e-01 1.71843603e-01
-4.99248743e-01 4.23424602e-01 5.69767714e-01 -5.65537453e-01
-8.03858221e-01 -1.00035000e+00 -3.77118319e-01 4.81200308e-01
3.12547296e-01 2.85292566e-01 -1.21634793e+00 -6.68697238e-01
-2.01732926e-02 2.27458075e-01 -1.03354251e+00 -1.79107357e-02
4.87929356e-04 -9.51565921e-01 6.57957017e-01 6.69942796e-01
1.04840672e+00 -9.44256425e-01 -3.04943830e-01 1.99488383e-02
4.23617035e-01 -4.54188287e-01 -1.06341198e-01 2.49116682e-02
-6.87750638e-01 -1.50981879e+00 -1.23364055e+00 -9.68248963e-01
8.99673462e-01 -1.37410492e-01 4.22989547e-01 3.80390257e-01
-9.83506382e-01 9.38740149e-02 -3.12261522e-01 -7.03235924e-01
-3.84202600e-01 1.46145314e-01 -2.42909953e-01 1.45133823e-01
7.19374716e-01 -2.05215544e-01 -8.34012687e-01 -1.96819648e-01
-7.99447834e-01 1.69727981e-01 9.41721559e-01 7.03243494e-01
5.33134997e-01 3.11201848e-02 6.31835461e-01 -1.35226297e+00
9.57764864e-01 -5.15090227e-01 -3.00848871e-01 5.20359755e-01
-6.78415120e-01 -6.04473054e-01 7.64558971e-01 -1.82179436e-01
-1.20338237e+00 1.14910617e-01 -7.84984529e-01 -2.26000860e-01
-2.98409253e-01 3.81387413e-01 4.48296428e-01 -5.51598966e-01
1.00142431e+00 2.34624907e-01 5.05202472e-01 -2.61312932e-01
-1.21272624e-01 1.08204770e+00 4.62515950e-01 1.86642528e-01
4.37563896e-01 2.60043055e-01 1.06623523e-01 -1.02069628e+00
-5.20134985e-01 -4.65021700e-01 -5.49975276e-01 -4.73487198e-01
1.39399147e+00 -7.35316932e-01 -7.99528182e-01 1.11852801e+00
-7.88062751e-01 -7.56412148e-01 1.11632809e-01 1.74494773e-01
-5.62091544e-03 5.92440665e-01 -8.89651179e-01 -9.40727592e-01
-8.86683285e-01 -7.46101081e-01 5.28176725e-01 9.66828942e-01
1.29137337e-01 -1.66442394e+00 1.79080784e-01 3.75474781e-01
5.58724940e-01 8.18770826e-01 5.61720908e-01 -3.29640001e-01
-1.05492502e-01 -5.16200125e-01 -7.68940687e-01 6.27222836e-01
3.64769220e-01 6.30111158e-01 -8.26415122e-01 9.20870081e-02
-3.70346516e-01 -2.36015961e-01 1.20398808e+00 7.38483012e-01
1.64490008e+00 5.67875020e-02 -8.03509653e-01 8.95144820e-01
1.92249560e+00 4.68692005e-01 9.44786906e-01 1.71944082e-01
8.36289644e-01 2.96702325e-01 3.40309203e-01 5.93275949e-02
2.54765451e-01 -2.06267029e-01 5.07690609e-01 -8.79758358e-01
-4.01551038e-01 -2.18326986e-01 -1.51628077e-01 6.47932827e-01
-6.21669292e-01 -9.32031795e-02 -9.81294572e-01 6.36536121e-01
-1.49584830e+00 -6.78808272e-01 -5.29651821e-01 2.05788112e+00
8.41839314e-01 8.65433738e-02 1.28594890e-01 3.49061415e-02
8.48470271e-01 -2.90710092e-01 -6.20373487e-01 -9.25272584e-01
1.90183595e-01 6.22290969e-01 6.23928845e-01 4.34315324e-01
-1.15544426e+00 9.01559591e-01 6.49747849e+00 1.00607681e+00
-1.74342811e+00 1.16225205e-01 7.89185047e-01 2.79043823e-01
-1.29758418e-01 -3.72841120e-01 -5.63367069e-01 5.69611430e-01
6.38393939e-01 1.65610805e-01 5.64720668e-02 7.01804280e-01
3.49345505e-02 -4.17202324e-01 -3.96008998e-01 8.77185225e-01
1.93298329e-02 -1.30777574e+00 6.79095536e-02 4.71160226e-02
5.94166875e-01 -3.66313696e-01 1.23414658e-01 2.26043195e-01
4.90905866e-02 -1.36674643e+00 -6.24883115e-01 9.71435010e-01
1.38350523e+00 -8.05776179e-01 9.59233820e-01 2.01296791e-01
-9.32016373e-01 -1.14434130e-01 -5.55248737e-01 3.49575095e-02
-2.17354760e-01 7.21382916e-01 -1.16369355e+00 5.23584008e-01
3.82737279e-01 7.26357400e-01 -6.49816394e-01 1.32452238e+00
-1.83196649e-01 6.93871439e-01 -9.14596617e-02 -5.67537904e-01
4.07148838e-01 -3.32113236e-01 7.06830770e-02 1.20900667e+00
1.29029348e-01 -1.93408970e-02 -3.82652953e-02 7.12705612e-01
1.17649838e-01 3.34253967e-01 -5.08269966e-01 4.99633141e-03
2.38371249e-02 1.77439094e+00 -7.69897938e-01 -1.89169556e-01
-4.65902090e-02 1.19677830e+00 3.12818550e-02 1.36837617e-01
-7.99323857e-01 -8.10301721e-01 2.86722034e-01 3.63847136e-01
-3.81536454e-01 3.11314821e-01 -6.04239106e-01 -8.34324002e-01
-5.22628605e-01 -2.54419595e-01 3.46600413e-01 -6.60642028e-01
-1.34630203e+00 5.21760941e-01 -7.69761860e-01 -9.32246566e-01
3.16533148e-01 -1.04348242e+00 -1.29173958e+00 1.04667115e+00
-1.77418983e+00 -1.51887405e+00 -9.57947075e-01 7.28838682e-01
4.68493253e-01 -1.23339802e-01 9.67519581e-01 5.39938286e-02
-1.07584119e+00 6.55138671e-01 -8.94143805e-03 3.86495143e-01
7.88590193e-01 -1.57960379e+00 -2.25270703e-01 4.93388265e-01
-6.98202550e-01 2.90252328e-01 -8.14511478e-02 -7.89281189e-01
-1.21417415e+00 -1.34099662e+00 2.06297383e-01 -1.09706923e-01
4.28728491e-01 3.39042157e-01 -6.63892746e-01 3.45781177e-01
4.86558288e-01 -1.72824696e-01 1.00314486e+00 1.76161498e-01
2.07615450e-01 -3.64981353e-01 -1.65037072e+00 6.27119660e-01
3.79336357e-01 -2.82607168e-01 -1.94091275e-01 8.61212671e-01
2.48724818e-01 -4.27734017e-01 -9.94397283e-01 4.12923872e-01
6.81284070e-01 -9.46321309e-01 6.66647851e-01 -4.08026129e-01
7.82614291e-01 9.08399820e-02 6.48539841e-01 -1.15145016e+00
-2.70150095e-01 -1.21781707e-01 2.66730398e-01 8.68314743e-01
3.29210967e-01 -7.71501303e-01 1.25723851e+00 5.22557437e-01
3.45067047e-02 -1.33760750e+00 -5.14541745e-01 -4.41118687e-01
3.53138566e-01 1.05370782e-01 5.91140650e-02 1.01585162e+00
-7.00900629e-02 5.51177636e-02 -1.14360772e-01 -4.51197401e-02
5.52118301e-01 -3.47401410e-01 4.54111397e-01 -1.15557218e+00
8.74598026e-02 -4.11796153e-01 -4.71243143e-01 -5.13815880e-01
-4.70926315e-01 -6.33203924e-01 -3.32004368e-01 -1.99923766e+00
2.67155111e-01 -5.81213772e-01 -5.13030827e-01 7.13770509e-01
-3.12493742e-01 4.80298787e-01 -3.87908012e-01 -1.88000515e-01
-1.72157109e-01 -1.88097015e-01 1.61751366e+00 -2.28395835e-01
-3.47168416e-01 2.28380919e-01 -6.43667400e-01 7.57485032e-01
9.34915185e-01 2.33017102e-01 -1.39163449e-01 -4.27149683e-01
-1.73222259e-01 1.63030609e-01 2.97102720e-01 -1.29439533e+00
5.40019810e-01 -4.24664408e-01 9.47278976e-01 -3.57353359e-01
2.00136900e-01 -6.08522534e-01 -2.74118632e-02 1.04362798e+00
1.09478042e-01 -7.92350650e-01 2.14478046e-01 3.02345395e-01
-1.28258184e-01 -4.55384314e-01 1.01533365e+00 -4.68143821e-01
-1.01129425e+00 5.79914808e-01 -4.96194392e-01 -6.13428891e-01
1.71224499e+00 -7.42546678e-01 -3.75065804e-01 1.20189458e-01
-8.77126276e-01 2.63202101e-01 4.38500464e-01 -9.03337598e-02
6.89995289e-01 -9.72908497e-01 -4.94950503e-01 -4.83037606e-02
-9.27864984e-02 1.92217678e-02 6.75173402e-01 9.20335591e-01
-1.45979881e+00 2.68449094e-02 -7.36114025e-01 -4.54274565e-01
-1.31594503e+00 3.93823057e-01 7.51348376e-01 -2.58632213e-01
-3.08428943e-01 1.35777378e+00 -3.66721362e-01 -3.79201829e-01
-7.62515217e-02 -2.12286204e-01 -5.94402730e-01 -1.74599543e-01
4.72538084e-01 4.20268267e-01 -2.83447087e-01 6.98318854e-02
-2.24657014e-01 8.91641617e-01 -2.14630902e-01 3.93985987e-01
1.10735500e+00 2.66814232e-01 -2.73893118e-01 -1.13582887e-01
9.65878665e-01 -1.36009172e-01 -1.03033793e+00 4.15886343e-01
-5.76051772e-01 -3.09924603e-01 2.96461016e-01 -1.30666602e+00
-1.15746462e+00 1.10876501e+00 1.06736529e+00 4.05125678e-01
1.51043093e+00 -7.89725602e-01 1.09914780e+00 1.41701832e-01
1.04010411e-01 -1.31562245e+00 -5.90858534e-02 -3.27389911e-02
4.61418241e-01 -1.48338950e+00 -6.52319640e-02 -7.71319866e-01
-5.88505268e-01 1.47976887e+00 1.04796898e+00 -5.25068700e-01
7.94600964e-01 2.87059844e-01 4.23676878e-01 -5.92010804e-02
-1.76592320e-01 -2.96444088e-01 -3.39341983e-02 1.00272954e+00
1.82640001e-01 2.89323926e-01 -5.20598173e-01 7.13478088e-01
2.71572262e-01 5.25081992e-01 4.28480774e-01 6.11712813e-01
-7.03991413e-01 -9.81942594e-01 -3.11515063e-01 1.02813339e+00
-4.76837337e-01 -9.44669824e-03 -6.26601338e-01 7.48833597e-01
5.34607947e-01 5.45513451e-01 3.01359706e-02 -7.05572426e-01
-1.20251015e-01 -1.99670583e-01 5.66969216e-01 -4.77801323e-01
-8.18406880e-01 -2.69081056e-01 -1.24571353e-01 -1.86908543e-01
-1.63476005e-01 3.19827721e-02 -1.31932616e+00 -2.75034547e-01
-3.78283530e-01 -1.93501785e-01 9.27480578e-01 7.34144568e-01
-8.85032862e-02 7.57060587e-01 6.95693314e-01 -1.99401274e-01
-2.70987809e-01 -1.07700932e+00 -9.75346565e-01 2.55597025e-01
-3.38452645e-02 -1.01401068e-01 -6.84896484e-02 6.15319498e-02] | [15.663595199584961, -3.011268377304077] |
089072da-ec05-4529-bf25-cf7901ef1d9f | satellite-dynamics-toolbox-library-a-tool-to | 2303.15872 | null | https://arxiv.org/abs/2303.15872v1 | https://arxiv.org/pdf/2303.15872v1.pdf | Satellite Dynamics Toolbox Library: a tool to model multi-body space systems for robust control synthesis and analysis | The level of maturity reached by robust control theory techniques nowadays contributes to a considerable minimization of the development time of an end-to-end control design of a spacecraft system. The advantage offered by this framework is twofold: all system uncertainties can be included from the very beginning of the design process; the validation and verification (V\&V) process is improved by fast detection of worst-case configurations that could escape to a classical sample-based Monte Carlo simulation campaign. Before proceeding to the control synthesis and analysis, a proper uncertain plant model has to be available in order to push these techniques to their limits of performance. In this spirit, the Satellite Dynamics Toolbox Library (SDTlib) offers many features to model a spacecraft system in a multi-body fashion on SIMULINK. Parametric models can be easily built in a Linear Fractional Transformation (LFT) form by including uncertainties and varying parameters with minimal number of repetitions. Uncertain Linear Time Invariant (LTI) and uncertain Linear Parameter-Varying (LPV) controllers can then be synthesized and analyzed in a straightforward way. The authors present in this article a tutorial, that can be downloaded at https://nextcloud.isae.fr/index.php/s/XDfRfHntejHTmmp, to show how to deal with an end-to-end robust design of a spacecraft mission and to provide to researchers a benchmark to test their own algorithms. | ['Franca Somers', 'Ervan Kassarian', 'Daniel Alazard', 'Francesco Sanfedino'] | 2023-03-28 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [-1.38761923e-01 1.57306984e-01 2.27082044e-01 4.98696417e-02
-2.47215673e-01 -1.03370476e+00 8.59078765e-01 -3.22374366e-02
-3.72702517e-02 1.17344570e+00 -6.17525876e-01 -4.26828206e-01
-7.28333473e-01 -6.31994188e-01 -6.48137629e-01 -8.24828565e-01
-2.81090647e-01 6.58265710e-01 4.01769318e-02 -5.68456411e-01
-1.58270746e-01 8.63704860e-01 -1.62120116e+00 -4.33463991e-01
8.30046952e-01 8.20438564e-01 2.28501424e-01 3.39739680e-01
3.77774030e-01 1.20480955e-01 -4.11483556e-01 3.16144019e-01
3.79820853e-01 -2.78977066e-01 -3.11552852e-01 3.28054465e-02
-5.62799215e-01 1.87483326e-01 1.20208889e-01 1.11769247e+00
4.88415062e-01 3.04475993e-01 7.58567870e-01 -1.26341045e+00
3.09718758e-01 2.33987898e-01 1.78879470e-01 -1.92857191e-01
3.24795932e-01 3.51788729e-01 1.88450649e-01 -6.92306042e-01
6.27951145e-01 9.52710330e-01 5.20422339e-01 -2.84256227e-02
-1.30760229e+00 -3.32884431e-01 -1.48794636e-01 -1.13434717e-01
-1.68141377e+00 -3.19240063e-01 4.23114210e-01 -7.23553538e-01
6.42103851e-01 8.00413847e-01 6.94838762e-01 6.36115074e-01
3.05777371e-01 7.53361210e-02 1.26099420e+00 -3.49319726e-01
4.94173259e-01 2.09822044e-01 -5.53019419e-02 2.50700653e-01
4.97781515e-01 5.14491796e-01 4.41920221e-01 -1.98150263e-03
7.25794017e-01 -4.92503852e-01 -5.06382406e-01 -5.16529500e-01
-1.11584914e+00 7.27417827e-01 3.46294433e-01 6.88109875e-01
-2.47693643e-01 -1.30278260e-01 5.43110728e-01 5.58889687e-01
1.29862484e-02 5.26298583e-01 -2.30179384e-01 5.20687252e-02
-7.15784371e-01 6.07937574e-01 1.00500631e+00 9.39064384e-01
2.96540558e-01 6.22043073e-01 1.68239430e-01 2.48797879e-01
4.38314170e-01 7.39355505e-01 1.80018708e-01 -8.27466249e-01
-2.06516787e-01 4.03898925e-01 6.95107400e-01 -6.68550611e-01
-6.48970962e-01 -7.69881845e-01 -7.75333941e-01 7.60117292e-01
3.01967770e-01 -3.58184457e-01 -3.96942586e-01 1.32798278e+00
6.00517690e-01 -2.42589757e-01 1.59509197e-01 9.76930201e-01
-7.01819137e-02 9.47001040e-01 -6.04201853e-01 -7.08032370e-01
1.19560719e+00 -2.15391845e-01 -8.75666261e-01 -1.00652888e-01
2.97333926e-01 -9.04151917e-01 6.09367132e-01 8.58320415e-01
-1.01596069e+00 -3.99235368e-01 -1.35424006e+00 6.50852501e-01
-3.86779368e-01 1.88147202e-01 -4.39174436e-02 6.79332495e-01
-8.40578377e-01 8.26065958e-01 -7.20297933e-01 -2.50294447e-01
-4.92453754e-01 4.70980763e-01 -9.49061885e-02 3.31129730e-01
-1.46687305e+00 1.35761225e+00 7.90683150e-01 4.83281463e-01
-8.46899867e-01 -6.38515890e-01 -6.36970520e-01 -2.00333789e-01
5.91791570e-01 -8.12655687e-01 1.29968011e+00 -9.16408718e-01
-1.78840542e+00 1.94577798e-01 1.63950741e-01 -4.79915082e-01
8.94710124e-01 3.06810975e-01 -3.49764943e-01 -8.36143196e-02
-2.54625738e-01 -8.92002732e-02 1.04804397e+00 -1.19885111e+00
-2.69856513e-01 9.28147435e-02 -7.30920583e-02 -1.85577869e-02
2.66466349e-01 7.88379610e-02 -6.61970153e-02 -6.44118369e-01
-2.68210739e-01 -1.22395968e+00 -2.87946939e-01 -2.35319212e-01
-2.38153070e-01 2.19889015e-01 6.66040003e-01 -5.67162037e-01
9.82257485e-01 -1.75577497e+00 6.76285446e-01 5.00084996e-01
-5.44259131e-01 3.53699088e-01 4.53539014e-01 9.60196137e-01
-4.59978789e-01 -2.17726097e-01 -3.56948972e-01 1.76404580e-01
3.99704516e-01 1.36916833e-02 -3.61392409e-01 7.74899185e-01
-2.27727499e-02 1.57613292e-01 -7.01529443e-01 -8.51207525e-02
7.04732656e-01 4.09359008e-01 -1.15251042e-01 -1.35502338e-01
-4.33131993e-01 7.33490229e-01 -6.51876628e-01 3.79228920e-01
5.40196836e-01 3.85592371e-01 2.20643848e-01 -7.16564804e-02
-8.21470141e-01 -2.35872224e-01 -1.73800886e+00 1.21856880e+00
-4.48701769e-01 2.74503231e-01 7.95623541e-01 -9.94083822e-01
1.09508002e+00 5.93438148e-01 4.14095461e-01 -2.63000369e-01
6.49629354e-01 6.25224531e-01 -5.21787442e-02 -1.81461796e-02
3.31968099e-01 -4.18151408e-01 -3.51941846e-02 -8.71725082e-02
-2.38286242e-01 -7.68465340e-01 6.03884816e-01 -2.55482763e-01
4.48543370e-01 1.78365201e-01 4.91660684e-01 -9.93744969e-01
1.24206626e+00 1.40933633e-01 4.96151388e-01 -8.01880509e-02
2.63350427e-01 -1.07701637e-01 6.19084477e-01 1.39592886e-01
-1.29885554e+00 -4.88214046e-01 -5.60146630e-01 1.49432048e-01
-2.41390273e-01 -2.74567846e-02 -6.49073064e-01 2.41124779e-01
4.49755155e-02 1.10475934e+00 -3.29418272e-01 -1.80644572e-01
-3.22515011e-01 -4.79530573e-01 1.11989692e-01 -9.69959050e-02
-6.85189813e-02 -3.91692400e-01 -6.17808819e-01 4.01097476e-01
3.86647344e-01 -5.49172878e-01 8.36639553e-02 2.42343992e-01
-9.11138654e-01 -9.79292154e-01 -5.85588694e-01 -4.65328962e-01
5.61100900e-01 -3.34007472e-01 6.21327877e-01 -1.73287079e-01
-2.73536116e-01 2.77976274e-01 -2.15445206e-01 -4.18569565e-01
-6.79690480e-01 -3.67100239e-01 5.36746442e-01 -2.73735702e-01
-7.61627316e-01 -3.56739789e-01 -2.71274894e-01 8.64531219e-01
-1.03890228e+00 -3.33174378e-01 4.30646017e-02 7.81903505e-01
5.32177508e-01 4.57440525e-01 5.73323727e-01 -1.04218394e-01
5.29875338e-01 -2.52042472e-01 -1.65435052e+00 6.89892098e-02
-4.76669848e-01 -8.80304649e-02 9.36471999e-01 -1.94968253e-01
-8.86579454e-01 3.15030724e-01 -2.20747650e-01 -4.38499987e-01
5.36451377e-02 7.45011270e-01 -4.22415823e-01 -3.69754046e-01
4.46854323e-01 -1.84668079e-02 5.80763757e-01 -3.31431627e-01
3.17217678e-01 4.33422238e-01 2.70369589e-01 -6.84940159e-01
1.15531743e+00 2.16224983e-01 5.36454022e-01 -8.91413331e-01
2.15041474e-01 -4.01087180e-02 -5.53734958e-01 -4.83836263e-01
2.20630467e-01 -8.53220761e-01 -8.99440765e-01 3.02429467e-01
-7.43187904e-01 -4.55431163e-01 -5.00411749e-01 4.62067455e-01
-7.60865986e-01 2.21301615e-01 -8.65701661e-02 -1.28262424e+00
4.70558666e-02 -1.40438044e+00 5.13771832e-01 1.46109343e-01
-2.31696948e-01 -1.00990736e+00 2.12210286e-02 -2.82661259e-01
3.95265430e-01 7.48879671e-01 3.87239039e-01 -3.07392716e-01
-5.12613893e-01 -5.36155164e-01 5.45228839e-01 3.88114154e-01
-2.31696784e-01 4.28643882e-01 -5.72719574e-01 -9.21032608e-01
4.67411280e-01 1.44711524e-01 1.44203082e-01 5.03622174e-01
1.56257868e-01 -2.04485923e-01 -3.57913345e-01 3.37939739e-01
1.81170380e+00 4.72055107e-01 4.76090550e-01 5.63293338e-01
-8.09858739e-02 7.91840017e-01 1.15781164e+00 5.22253513e-01
-3.32513839e-01 9.67466772e-01 7.45464802e-01 3.25727195e-01
4.89019692e-01 4.47224945e-01 6.01498306e-01 3.64758909e-01
-1.27395153e-01 -1.06813185e-01 -8.59491408e-01 4.22889829e-01
-1.54928124e+00 -7.62158573e-01 -7.92679846e-01 2.67485690e+00
4.35561389e-01 2.72145927e-01 1.52810737e-01 5.33652544e-01
8.74644518e-01 -1.57134831e-01 -1.75691292e-01 -5.44082046e-01
2.05198321e-02 -2.20997095e-01 6.58864617e-01 7.52341449e-01
-9.57454264e-01 1.26626045e-01 5.06224728e+00 8.89380932e-01
-1.16054535e+00 -2.00867176e-01 -1.01154475e-02 -1.03265688e-01
-1.20930821e-01 2.18799591e-01 -7.07352936e-01 4.42411959e-01
1.41866624e+00 -7.30159879e-01 5.22621930e-01 7.75708020e-01
8.72714996e-01 -2.21356511e-01 -6.78127527e-01 5.73183775e-01
-4.13784921e-01 -1.02218950e+00 -5.32426178e-01 1.45177305e-01
6.67343736e-01 -2.74796247e-01 -1.65260397e-02 3.16106498e-01
-1.01404771e-01 -6.93108022e-01 9.63301063e-01 8.85018289e-01
5.31322658e-01 -1.09693992e+00 6.12813711e-01 5.58304191e-01
-1.11237764e+00 -3.20306003e-01 -2.38859490e-01 -1.13625221e-01
7.59206116e-01 7.43991673e-01 -8.20183933e-01 1.50108349e+00
1.32935122e-01 2.97274232e-01 -2.10573122e-01 1.39988422e+00
-5.28244041e-02 2.68780082e-01 -6.67776942e-01 -2.36831129e-01
2.40161404e-01 -6.97855294e-01 1.12953532e+00 7.10823953e-01
8.91700268e-01 -3.76061834e-02 -8.11334401e-02 9.44971383e-01
8.25766325e-01 1.42337620e-01 -4.73355979e-01 -1.09082438e-01
2.58105159e-01 1.18337476e+00 -6.49531960e-01 -1.00598313e-01
1.44185171e-01 4.25927132e-01 -6.63981616e-01 8.63253400e-02
-1.07409346e+00 -6.12085342e-01 5.41054189e-01 1.76472172e-01
3.10339689e-01 -5.06671786e-01 -3.20051461e-02 -7.58135021e-01
-1.11579098e-01 -1.04362655e+00 9.65821519e-02 -8.89953613e-01
-7.71325111e-01 6.17307603e-01 2.79422224e-01 -1.83333683e+00
-9.02800858e-01 -7.74046004e-01 -5.02784431e-01 1.15615988e+00
-9.92102325e-01 -8.18981528e-01 1.61962524e-01 4.09234822e-01
2.03359485e-01 -6.06102496e-02 8.24535489e-01 3.50408584e-01
-6.59433722e-01 -2.55173415e-01 7.99473643e-01 -6.34880066e-01
5.87035835e-01 -1.14496589e+00 -8.39865282e-02 9.70815539e-01
-8.55550230e-01 6.13551259e-01 1.55891895e+00 -6.77883863e-01
-1.65413904e+00 -9.80166018e-01 5.52167356e-01 -2.94490773e-02
1.13522875e+00 4.71782358e-03 -7.69235492e-01 5.37966609e-01
3.28110069e-01 -2.69283414e-01 -1.02067687e-01 -3.49613070e-01
6.11246765e-01 -9.48794857e-02 -1.12460768e+00 5.92439234e-01
1.73145890e-01 -1.26064137e-01 -5.94928265e-01 4.01829988e-01
4.62666780e-01 -5.13385296e-01 -1.39426506e+00 5.08220315e-01
1.66910648e-01 -7.78307736e-01 9.91869450e-01 2.24485267e-02
-4.51526076e-01 -1.01489341e+00 2.70153433e-01 -1.53297341e+00
1.16572738e-01 -1.28711212e+00 2.24652812e-01 1.32739389e+00
1.66033924e-01 -1.00981092e+00 -1.49484172e-01 4.42083776e-01
-5.17553329e-01 -1.76759809e-01 -1.24429190e+00 -1.32628405e+00
1.91254944e-01 -3.73647124e-01 4.51840937e-01 7.63512969e-01
1.84758455e-01 -1.16818801e-01 -2.17329383e-01 4.98741060e-01
5.92911303e-01 5.84898889e-02 6.30101204e-01 -1.31668460e+00
-3.59897107e-01 -4.84558433e-01 -1.76857263e-01 -1.03725523e-01
-7.22664520e-02 -5.48692226e-01 -8.04693550e-02 -1.22993243e+00
-9.20990944e-01 -2.19785348e-01 9.66914147e-02 -6.65490851e-02
4.02271986e-01 -1.20370701e-01 2.68255889e-01 5.75581975e-02
2.03664228e-01 5.80509841e-01 1.04124320e+00 7.74973854e-02
-1.56699091e-01 4.40683335e-01 1.48819521e-01 5.83638489e-01
9.18505609e-01 -1.82897061e-01 -4.93937731e-01 3.44861954e-01
3.92447561e-01 5.87614059e-01 5.43646216e-01 -1.31084490e+00
8.92667100e-03 -3.49662192e-02 4.76022623e-03 -5.28288424e-01
3.54448259e-01 -1.17596841e+00 1.11934972e+00 8.87880564e-01
2.69230902e-01 -7.07127675e-02 5.10743499e-01 2.92398453e-01
-3.45638335e-01 -6.97301328e-01 9.47818458e-01 8.90876427e-02
-4.28313106e-01 -1.79249614e-01 -7.70271122e-01 -5.72173595e-01
1.39146471e+00 -1.06081098e-01 -7.71435723e-02 -2.58742422e-01
-1.06552756e+00 3.73153627e-01 8.06928694e-01 -9.44305360e-02
-1.74090579e-01 -1.05288339e+00 -5.24155378e-01 8.14802130e-04
-1.43240809e-01 -2.37715289e-01 7.24052668e-01 1.12939370e+00
-6.73467577e-01 9.12323415e-01 -4.23674434e-01 -4.75893348e-01
-1.05526781e+00 9.53854978e-01 7.76816070e-01 3.76582928e-02
-2.29116648e-01 2.50716031e-01 -4.68416810e-01 -3.14185798e-01
-1.41347021e-01 -4.80722308e-01 -9.91095304e-02 3.84145200e-01
4.45340693e-01 4.99609590e-01 4.13606405e-01 -5.96431851e-01
-3.33971143e-01 5.17381489e-01 7.55102098e-01 -3.66907686e-01
1.22938120e+00 -1.78244695e-01 -1.45211518e-01 3.85233015e-01
7.76019990e-01 9.77239534e-02 -1.35054505e+00 4.84440506e-01
-3.97069193e-02 -1.14313267e-01 2.85828859e-01 -7.73953617e-01
-6.57505989e-01 2.64239401e-01 6.05885923e-01 5.88322818e-01
1.04657638e+00 -7.40327954e-01 -8.96099880e-02 2.68145204e-01
6.54616714e-01 -1.08380127e+00 -8.24885309e-01 5.48069656e-01
1.53650939e+00 -5.65701485e-01 2.24567965e-01 -5.14893413e-01
-5.49070895e-01 1.39915168e+00 -1.02307312e-01 -4.03180033e-01
5.34894228e-01 6.45545900e-01 -3.11636537e-01 3.00525904e-01
-4.79495764e-01 -4.32824492e-01 3.01624298e-01 3.57102156e-01
2.71752238e-01 -1.03523828e-01 -9.79901135e-01 4.11814183e-01
6.71613067e-02 1.44932494e-01 7.46569276e-01 9.69596863e-01
-5.60279071e-01 -1.37359369e+00 -1.20535445e+00 -2.53616810e-01
-2.98121944e-02 4.12452519e-01 -5.88832535e-02 1.30874467e+00
-9.99746472e-02 8.27230096e-01 -2.87036777e-01 5.63274026e-02
8.26156020e-01 3.30221683e-01 3.45018566e-01 -9.34377760e-02
-7.75037348e-01 4.23695326e-01 5.03294826e-01 -4.20110196e-01
-2.47858658e-01 -9.26981747e-01 -1.20971584e+00 -2.96078026e-01
-3.14488918e-01 6.07733727e-01 1.05015540e+00 6.15341306e-01
5.26560657e-02 1.01670468e+00 6.00907147e-01 -1.15094972e+00
-9.40836430e-01 -8.13451946e-01 -9.19851243e-01 -4.25459653e-01
7.37038255e-02 -9.63695765e-01 -5.26748598e-01 -1.09351315e-01] | [5.370342254638672, 2.3436851501464844] |
b615da15-b7a8-4d32-be55-52bb030f124c | perfect-match-a-simple-method-for-learning | 1810.00656 | null | https://arxiv.org/abs/1810.00656v5 | https://arxiv.org/pdf/1810.00656v5.pdf | Perfect Match: A Simple Method for Learning Representations For Counterfactual Inference With Neural Networks | Learning representations for counterfactual inference from observational data is of high practical relevance for many domains, such as healthcare, public policy and economics. Counterfactual inference enables one to answer "What if...?" questions, such as "What would be the outcome if we gave this patient treatment $t_1$?". However, current methods for training neural networks for counterfactual inference on observational data are either overly complex, limited to settings with only two available treatments, or both. Here, we present Perfect Match (PM), a method for training neural networks for counterfactual inference that is easy to implement, compatible with any architecture, does not add computational complexity or hyperparameters, and extends to any number of treatments. PM is based on the idea of augmenting samples within a minibatch with their propensity-matched nearest neighbours. Our experiments demonstrate that PM outperforms a number of more complex state-of-the-art methods in inferring counterfactual outcomes across several benchmarks, particularly in settings with many treatments. | ['Walter Karlen', 'Patrick Schwab', 'Lorenz Linhardt'] | 2018-10-01 | perfect-match-a-simple-method-for-learning-1 | https://openreview.net/forum?id=rkxt8oC9FQ | https://openreview.net/pdf?id=rkxt8oC9FQ | iclr-2019-5 | ['counterfactual-inference'] | ['miscellaneous'] | [ 5.27015328e-01 4.13192421e-01 -8.35721493e-01 -5.56955218e-01
-6.89629614e-01 -2.64330775e-01 8.08098853e-01 1.94352135e-01
-6.27684116e-01 1.56233442e+00 8.51781785e-01 -1.08152962e+00
-6.06224775e-01 -8.78914475e-01 -1.00205958e+00 -6.61830664e-01
-3.75128627e-01 6.15030885e-01 -6.83496237e-01 -1.36618288e-02
1.13948211e-01 3.62225980e-01 -1.39815128e+00 3.33048344e-01
7.05923557e-01 6.63567662e-01 -3.67176771e-01 1.10900193e-01
2.10949838e-01 5.75994253e-01 -3.13846707e-01 -7.26726115e-01
2.63742447e-01 -2.25262225e-01 -8.42833459e-01 -4.96119797e-01
4.50804323e-01 -7.29903758e-01 -4.94818956e-01 7.17992127e-01
8.74038398e-01 3.60733867e-01 9.24295664e-01 -1.07616270e+00
-7.04974830e-01 1.08968580e+00 -4.37006742e-01 1.94538236e-01
2.17920020e-01 3.35265934e-01 6.86374664e-01 -2.17879638e-01
4.46852565e-01 1.48528969e+00 8.50549757e-01 7.44853497e-01
-1.53440702e+00 -8.64544034e-01 5.93473911e-02 1.54537529e-01
-7.23969817e-01 -6.56303167e-01 6.13557935e-01 -3.85921150e-01
9.17511463e-01 3.71460766e-01 5.44554591e-01 1.55007005e+00
3.67153317e-01 7.30018258e-01 1.26151872e+00 -3.25722247e-01
6.30283415e-01 -3.13915431e-01 -2.12589070e-01 -4.80182096e-02
5.18901110e-01 1.02557838e+00 -1.24509603e-01 -6.77189291e-01
7.92587399e-01 4.07118708e-01 -4.46598887e-01 -7.89596677e-01
-1.29697919e+00 1.35873067e+00 5.65272689e-01 -2.46226370e-01
-8.76084030e-01 3.02798182e-01 7.30960548e-01 2.16572210e-01
5.41722953e-01 7.76028931e-01 -6.87346756e-01 4.70557474e-02
-9.38219011e-01 9.25007105e-01 7.89465547e-01 2.28615060e-01
2.16851145e-01 -2.00723615e-02 -5.25297403e-01 7.61357725e-01
-2.11185127e-01 5.04375517e-01 7.08072722e-01 -1.33386397e+00
5.74048281e-01 1.74982220e-01 6.04253173e-01 -2.99896091e-01
-6.72829628e-01 -2.66637236e-01 -1.20833540e+00 2.84329027e-01
5.81026137e-01 -6.44832850e-01 -8.71900320e-01 2.01688480e+00
4.28781450e-01 3.51331532e-01 1.56865314e-01 8.00707400e-01
5.94414949e-01 1.15101904e-01 2.71011859e-01 -5.47497988e-01
1.17999625e+00 -7.18863383e-02 -6.60606325e-01 -1.85118556e-01
7.59344697e-01 -4.78181809e-01 7.76668429e-01 -1.33378999e-02
-1.18207991e+00 -8.48674774e-02 -7.32583821e-01 3.01418781e-01
-4.60525721e-01 -3.83967191e-01 1.19789028e+00 8.85666609e-01
-8.48755479e-01 1.11618364e+00 -4.24067587e-01 -4.89769876e-02
9.12823737e-01 5.40654361e-01 -2.29764387e-01 -1.47922575e-01
-1.59072900e+00 1.04590666e+00 5.39932370e-01 -8.10640827e-02
-8.45810175e-01 -1.24951684e+00 -1.11721122e+00 3.40384722e-01
5.20214915e-01 -1.39422441e+00 1.55781615e+00 -8.56798589e-01
-1.34652984e+00 6.17189825e-01 -1.13214672e-01 -1.25904655e+00
8.72421980e-01 -5.25839776e-02 -3.84256929e-01 -4.32201236e-01
2.86512375e-01 7.99419463e-01 3.42965961e-01 -7.79575646e-01
-4.36227262e-01 -5.54653764e-01 2.55705476e-01 3.54810596e-01
2.53180414e-01 -3.17930132e-02 7.14347541e-01 -6.98032856e-01
-4.40510869e-01 -7.46534228e-01 -7.59813964e-01 -2.15243831e-01
-4.76507902e-01 -3.11244100e-01 2.86976069e-01 -2.52381027e-01
8.88215423e-01 -1.69767606e+00 -5.14470518e-01 -7.51505271e-02
7.10825995e-02 1.83204323e-01 7.28812590e-02 2.06772640e-01
-7.53796697e-01 9.91447717e-02 -4.01779175e-01 1.43249840e-01
2.37595469e-01 5.64789511e-02 -6.20189011e-01 5.16091883e-01
-1.65080413e-01 1.04003656e+00 -9.96206820e-01 -2.59563744e-01
4.63111639e-01 -1.52002899e-02 -7.12049246e-01 1.09659277e-01
-1.59255728e-01 5.63963838e-02 -7.55433291e-02 -1.78456418e-02
8.08584154e-01 -2.02987254e-01 4.16476399e-01 2.44816259e-01
9.90356319e-03 8.69080663e-01 -1.17152214e+00 1.33459258e+00
-4.32826996e-01 3.50639492e-01 -4.66909379e-01 -1.46425843e+00
2.68277645e-01 6.97348833e-01 4.13348824e-01 -5.31739056e-01
2.66713023e-01 1.39609993e-01 3.29798371e-01 -4.20733213e-01
2.51098454e-01 -8.53654265e-01 -2.04077378e-01 7.24478602e-01
-2.66335338e-01 8.39597210e-02 6.51317984e-02 -2.23122537e-01
8.33761454e-01 -7.90291727e-02 9.78304863e-01 -2.52152622e-01
-1.01130649e-01 -2.10691631e-01 6.06941462e-01 1.25910509e+00
-1.99744970e-01 4.61723208e-01 4.99190778e-01 -8.53827596e-01
-1.04772854e+00 -1.41294646e+00 -3.42869818e-01 6.18368626e-01
-4.73196924e-01 5.92607081e-01 -5.32751501e-01 -6.94873452e-01
4.83294725e-01 1.35646749e+00 -1.00362098e+00 -4.18541253e-01
-4.87282634e-01 -1.18778288e+00 4.24921662e-01 7.27766097e-01
3.82566512e-01 -1.30492663e+00 -6.47315979e-01 2.98837692e-01
-2.22922608e-01 -3.83495212e-01 -3.21711957e-01 -6.18576631e-02
-1.19766390e+00 -1.17717469e+00 -9.22817290e-01 -7.01663569e-02
1.56818181e-01 -2.79531833e-02 1.36045027e+00 -5.85034907e-01
-6.06789580e-03 -3.47131163e-01 4.00659800e-01 -8.93829942e-01
-3.68101478e-01 -3.67387235e-01 3.76249343e-01 -4.10856366e-01
7.02426493e-01 -6.70608044e-01 -1.05964434e+00 -1.28120780e-01
-6.68823361e-01 -9.99500677e-02 5.88958502e-01 1.21152306e+00
1.39112875e-01 -3.18683773e-01 1.06627870e+00 -1.34239209e+00
9.43776548e-01 -7.51505136e-01 -5.34466803e-01 3.61655578e-02
-5.03811002e-01 1.08798034e-01 8.06014538e-01 -5.48193991e-01
-1.26252687e+00 -3.63486886e-01 -1.34663329e-01 -2.71385968e-01
-5.36635399e-01 4.78187203e-01 1.16552832e-02 7.88979709e-01
9.63609219e-01 -1.23875350e-01 8.25681761e-02 -3.48727018e-01
5.53167582e-01 6.74081564e-01 5.82619548e-01 -3.89742225e-01
7.73951737e-03 6.80450678e-01 3.51854078e-02 -1.27034381e-01
-9.50749815e-01 -7.82611966e-02 7.52632990e-02 4.10930634e-01
4.55635637e-01 -7.76598394e-01 -1.37053430e+00 1.46492541e-01
-9.76256907e-01 -5.52014232e-01 -7.17089355e-01 1.06455505e+00
-9.67683375e-01 5.31494915e-02 -1.99673399e-01 -8.37416828e-01
-3.77065420e-01 -8.54201138e-01 6.17909729e-01 8.03833753e-02
-5.39314270e-01 -1.14117432e+00 1.39695987e-01 1.19053222e-01
3.21266681e-01 4.83123928e-01 1.15365171e+00 -5.85801780e-01
-2.01408193e-02 -8.00019726e-02 -2.75925934e-01 -6.83656782e-02
3.06465656e-01 -6.01767182e-01 -8.65486979e-01 -3.48624498e-01
5.60179986e-02 -3.24166030e-01 9.20262158e-01 1.37467885e+00
1.72622073e+00 -9.74788189e-01 -6.00398958e-01 3.47648978e-01
1.25087607e+00 2.75274038e-01 6.08018935e-01 1.62151694e-01
5.41001372e-02 6.63630366e-01 2.44844317e-01 5.36505520e-01
4.42667037e-01 5.07563233e-01 5.20242453e-01 -2.06379771e-01
3.42432261e-01 -2.27038980e-01 -1.00854024e-01 -5.56671798e-01
-3.41284238e-02 -1.55438825e-01 -7.03262269e-01 1.06934178e+00
-1.88096619e+00 -1.43005490e+00 1.39258549e-01 2.58926153e+00
1.04594982e+00 8.25067908e-02 2.68539459e-01 1.89485438e-02
7.07474589e-01 1.65830269e-01 -9.12961185e-01 -8.51186514e-01
1.23520322e-01 5.18566549e-01 8.57048512e-01 1.62448630e-01
-1.23962355e+00 2.78830051e-01 6.97598553e+00 6.86496794e-01
-9.64224994e-01 9.86053795e-02 1.11115563e+00 -5.18911123e-01
-5.10209262e-01 -5.32821007e-02 -2.66686618e-01 7.20062852e-01
1.20644355e+00 -4.48846459e-01 2.64708400e-01 6.29932046e-01
6.97398365e-01 -4.30449992e-02 -1.62234616e+00 6.56206071e-01
-5.20654917e-01 -2.06425142e+00 1.98784590e-01 9.74552408e-02
1.01117992e+00 1.77811012e-01 2.55203873e-01 4.40366179e-01
1.27348053e+00 -1.48306823e+00 2.28686094e-01 4.17623878e-01
9.60356653e-01 -7.86933601e-01 1.07131386e+00 5.12243330e-01
-1.53160736e-01 -2.56873280e-01 -3.77869993e-01 -5.75611889e-01
3.34512383e-01 7.17387736e-01 -1.08671451e+00 5.59428275e-01
6.46750808e-01 2.12669820e-01 2.93906778e-01 1.14843595e+00
-2.36174222e-02 5.63706458e-01 -2.44639948e-01 -4.87241484e-02
2.89869338e-01 2.51948237e-01 7.31411949e-02 8.93852234e-01
3.72948200e-01 1.53090999e-01 -3.31193745e-01 8.37819755e-01
-3.34492534e-01 -1.54909998e-01 -1.21706748e+00 4.04383630e-01
6.80347085e-01 3.14713508e-01 -5.40613681e-02 -6.44767880e-01
-2.19363019e-01 3.18905979e-01 3.35422307e-01 3.83644640e-01
-6.26186848e-01 -9.81425270e-02 8.94884944e-01 1.45369470e-01
7.68504664e-02 7.01316118e-01 -5.07312715e-01 -9.57290232e-01
-1.88279346e-01 -1.06493163e+00 9.21534538e-01 -6.26099765e-01
-1.47790205e+00 -2.49221563e-01 3.91270310e-01 -1.10099983e+00
-9.13747907e-01 -6.20681643e-01 -7.94538736e-01 9.74800825e-01
-1.35946882e+00 -6.66868627e-01 4.92783129e-01 3.32831979e-01
2.43795723e-01 1.01325326e-01 1.02009368e+00 2.50802916e-02
-3.05920690e-01 5.32115221e-01 4.47379351e-01 3.36434245e-02
5.19500494e-01 -1.32482541e+00 4.68744338e-01 3.75700474e-01
-2.69471914e-01 7.81520188e-01 1.00564504e+00 -3.76253456e-01
-6.39008820e-01 -9.82726634e-01 1.10733843e+00 -2.97048628e-01
3.84107232e-01 -3.72200608e-02 -5.59063613e-01 8.05152178e-01
2.52033979e-01 -1.68228000e-01 7.58892119e-01 5.95433772e-01
-1.97898969e-01 -7.97847137e-02 -1.52118003e+00 1.07700932e+00
1.11054027e+00 -2.81081408e-01 -1.07336211e+00 4.62132365e-01
6.99208379e-01 -3.74660015e-01 -6.88337803e-01 6.41140342e-01
7.56434441e-01 -1.02745426e+00 1.33092666e+00 -1.37484753e+00
8.58670413e-01 2.39632189e-01 1.08989775e-01 -1.78589380e+00
-1.03796437e-01 -5.67555547e-01 -5.75549006e-02 5.51120698e-01
4.33986366e-01 -1.16587961e+00 9.23972726e-01 8.18176508e-01
2.40767285e-01 -8.63945723e-01 -1.30258727e+00 -5.70768476e-01
6.71588123e-01 -3.83455604e-01 1.45105648e+00 1.25691938e+00
1.62292123e-01 4.67669107e-02 -5.31331599e-01 -8.83698314e-02
6.40654385e-01 5.82323909e-01 6.16178572e-01 -1.13868511e+00
-3.49150121e-01 -6.35205746e-01 2.61156876e-02 -7.71947026e-01
3.44816506e-01 -7.04990447e-01 -2.88332969e-01 -1.43539000e+00
3.94585818e-01 -3.30655068e-01 -4.03856665e-01 3.22484106e-01
-4.07023013e-01 -1.22753739e-01 -2.57621467e-01 -3.52384895e-01
9.17429924e-02 6.45649970e-01 1.02790880e+00 -1.46070436e-01
4.47016470e-02 4.18107450e-01 -1.03094602e+00 7.47998297e-01
8.61312091e-01 -5.77171743e-01 -4.44593042e-01 -1.16450265e-01
-3.16823013e-02 6.15764320e-01 6.53347969e-01 -3.81569564e-01
-2.11369827e-01 -6.09130979e-01 6.01811886e-01 -3.67354333e-01
1.35287479e-01 -4.65548605e-01 2.80965090e-01 8.00335884e-01
-6.10688031e-01 -1.84443146e-02 2.04191238e-01 5.30061603e-01
1.66567534e-01 -1.75359726e-01 5.87245643e-01 -4.66991216e-01
-2.09078521e-01 1.91922307e-01 -1.85115442e-01 2.12398678e-01
7.63023078e-01 -7.98357204e-02 -5.44238567e-01 -5.05111814e-01
-6.06024444e-01 4.04431611e-01 1.59592971e-01 3.26345980e-01
4.34602499e-01 -1.45223331e+00 -9.86573279e-01 -9.71965790e-02
-4.66726646e-02 -4.21132520e-02 5.50526679e-01 6.40154004e-01
-4.53966074e-02 7.28941262e-01 -2.32350796e-01 -9.85156521e-02
-7.71899521e-01 9.63783503e-01 5.72317362e-01 -6.97680712e-01
-4.78340149e-01 3.91200453e-01 6.80525303e-01 -9.21940088e-01
-1.44674748e-01 -1.82744607e-01 -1.42407969e-01 -1.86873332e-01
5.84018946e-01 4.64753985e-01 -7.69579038e-02 -5.10311052e-02
-1.62034631e-01 -3.39215606e-01 -1.62613377e-01 -6.37261719e-02
1.39845967e+00 2.96792984e-01 1.41371891e-01 2.63310105e-01
9.87016499e-01 -6.64138138e-01 -1.17958236e+00 -1.27911881e-01
-1.76474392e-01 -5.57947099e-01 -1.43499613e-01 -1.00464487e+00
-5.88594019e-01 7.13220835e-01 6.20579243e-01 1.63215816e-01
9.00604665e-01 -2.29154065e-01 3.53910834e-01 3.56986344e-01
1.76796049e-01 -9.85543907e-01 -6.39258206e-01 -4.53340225e-02
9.42133605e-01 -1.35900044e+00 2.30786338e-01 1.54737204e-01
-2.78810054e-01 4.89386886e-01 2.22204149e-01 -2.32885286e-01
6.48220718e-01 -4.61043864e-02 -2.66820490e-02 -1.21604197e-01
-9.38509405e-01 5.39668426e-02 8.29607397e-02 6.08406186e-01
7.13013411e-01 7.40470946e-01 -6.17816687e-01 2.16486663e-01
-6.55840039e-01 5.15486598e-01 5.63151896e-01 3.99108082e-01
3.02692026e-01 -8.63709211e-01 -3.15710366e-01 1.30958652e+00
-7.99289405e-01 -3.30296844e-01 5.46384566e-02 1.00820601e+00
3.95699367e-02 7.11524844e-01 4.55243826e-01 3.97858858e-01
3.88178438e-01 7.16288313e-02 3.20300043e-01 -3.72692168e-01
-3.48358303e-01 -4.76303607e-01 3.60172570e-01 -5.52672923e-01
-5.44447362e-01 -9.45306897e-01 -6.98580921e-01 -6.98666096e-01
-3.80517580e-02 1.58927813e-01 2.84280241e-01 1.08718121e+00
3.27385157e-01 3.78557444e-01 5.27177989e-01 -7.48849094e-01
-1.26972938e+00 -1.16971958e+00 -3.44157606e-01 5.37899792e-01
5.88605940e-01 -7.88630664e-01 -1.57450885e-01 -4.77638900e-01] | [8.082780838012695, 5.413482189178467] |
59760a98-6bec-4abb-81c2-54af1c9d0c5e | deep-kernel-learning-via-random-fourier | 1910.02660 | null | https://arxiv.org/abs/1910.02660v1 | https://arxiv.org/pdf/1910.02660v1.pdf | Deep Kernel Learning via Random Fourier Features | Kernel learning methods are among the most effective learning methods and have been vigorously studied in the past decades. However, when tackling with complicated tasks, classical kernel methods are not flexible or "rich" enough to describe the data and hence could not yield satisfactory performance. In this paper, via Random Fourier Features (RFF), we successfully incorporate the deep architecture into kernel learning, which significantly boosts the flexibility and richness of kernel machines while keeps kernels' advantage of pairwise handling small data. With RFF, we could establish a deep structure and make every kernel in RFF layers could be trained end-to-end. Since RFF with different distributions could represent different kernels, our model has the capability of finding suitable kernels for each layer, which is much more flexible than traditional kernel-based methods where the kernel is pre-selected. This fact also helps yield a more sophisticated kernel cascade connection in the architecture. On small datasets (less than 1000 samples), for which deep learning is generally not suitable due to overfitting, our method achieves superior performance compared to advanced kernel methods. On large-scale datasets, including non-image and image classification tasks, our method also has competitive performance. | ['Jiaxuan Xie', 'Kaijie Wang', 'Fanghui Liu', 'Xiaolin Huang'] | 2019-10-07 | null | null | null | null | ['small-data'] | ['computer-vision'] | [-5.30970275e-01 -4.62462872e-01 -2.25562334e-01 -4.78179961e-01
-3.79567057e-01 -4.82225984e-01 2.67759085e-01 4.99522984e-02
-4.84893680e-01 3.73588502e-01 -9.61125642e-02 -4.08023447e-01
-2.66308159e-01 -8.62425983e-01 -5.49637437e-01 -7.90871620e-01
-1.63769379e-01 -3.69833894e-02 6.01892710e-01 -3.12933028e-02
7.41774067e-02 6.11295223e-01 -1.25422001e+00 1.73239276e-01
1.24459779e+00 1.01015294e+00 2.74626791e-01 3.36587727e-01
-6.06299080e-02 7.82083690e-01 -2.55821586e-01 -4.43365961e-01
-7.99816996e-02 -5.25479391e-02 -7.03502178e-01 -3.01634401e-01
2.30897456e-01 -3.06721568e-01 -4.79996175e-01 9.77278233e-01
4.54276830e-01 8.50863159e-02 8.98974478e-01 -9.51789737e-01
-9.03452516e-01 4.71298784e-01 -5.63720286e-01 7.86292776e-02
-6.42074272e-02 -8.35202783e-02 7.00865507e-01 -9.88215804e-01
-1.29776075e-01 9.96121168e-01 9.10856009e-01 3.53857696e-01
-1.33426094e+00 -6.81520045e-01 6.06353171e-02 3.68160576e-01
-1.48514974e+00 -7.65297636e-02 7.29299068e-01 -3.76631558e-01
7.88592339e-01 1.13313951e-01 6.05118990e-01 9.80411112e-01
1.18327983e-01 1.00041258e+00 1.06647742e+00 -2.09954351e-01
1.92862540e-01 3.81141752e-01 2.52238572e-01 8.08810055e-01
8.19514915e-02 -4.41899113e-02 -1.37193590e-01 -2.99820840e-01
1.10150397e+00 3.53839129e-01 -4.68346566e-01 -4.70849514e-01
-1.43198919e+00 1.10588443e+00 7.83022523e-01 2.37688944e-01
-2.45288685e-01 -4.19297168e-04 4.92669970e-01 4.21864688e-01
5.07249773e-01 3.39559555e-01 -5.69072902e-01 8.88311937e-02
-8.58030498e-01 1.25521049e-01 7.36526430e-01 4.16481882e-01
8.50122511e-01 -4.96117137e-02 -1.26298442e-01 1.14796889e+00
-1.09372591e-03 3.05385828e-01 5.86128056e-01 -3.09809834e-01
3.74382436e-02 7.55895257e-01 9.87963080e-02 -7.92834699e-01
-5.80199957e-01 -3.44344616e-01 -1.31088579e+00 3.56688559e-01
5.38721502e-01 -4.52245586e-02 -8.55982602e-01 1.48405063e+00
2.08229139e-01 2.66191185e-01 5.35000302e-02 9.22479451e-01
5.22219896e-01 8.60204756e-01 -8.38154107e-02 1.35473624e-01
1.44781017e+00 -1.06163621e+00 -3.96004438e-01 -8.68942291e-02
6.95786953e-01 -8.50911736e-01 1.43153226e+00 4.38454717e-01
-5.02345681e-01 -6.62758112e-01 -8.36822152e-01 9.11988094e-02
-5.50899625e-01 5.94443917e-01 1.27826989e+00 4.18136895e-01
-1.00505126e+00 5.36834776e-01 -7.55053580e-01 -1.65979877e-01
5.52292824e-01 3.29299420e-01 -5.27084410e-01 -6.54949024e-02
-1.33334470e+00 9.55214739e-01 5.60500681e-01 1.37547955e-01
-5.05607069e-01 -7.83619523e-01 -7.81054556e-01 2.81736910e-01
2.86402106e-01 -4.96078491e-01 9.36058521e-01 -8.46232235e-01
-1.62917256e+00 4.61579561e-01 6.71959147e-02 -2.98203707e-01
2.85631478e-01 -3.04891884e-01 -2.79865503e-01 6.28438517e-02
-1.41847029e-01 4.33523715e-01 1.30045462e+00 -5.77973902e-01
-3.43469828e-01 3.05083138e-03 1.72441043e-02 -1.54213533e-01
-6.69942260e-01 3.00018210e-02 -1.34746075e-01 -7.08775282e-01
-2.28540659e-01 -8.74089301e-01 -3.58860701e-01 -7.47590419e-03
-2.23160073e-01 -6.42867208e-01 7.43626595e-01 -3.10082406e-01
1.37432647e+00 -2.43388915e+00 -1.59068853e-01 2.16638044e-01
1.73165232e-01 7.18544126e-01 -7.25465044e-02 5.33011079e-01
-2.43331686e-01 -1.54926732e-01 -2.44933125e-02 1.44315407e-01
2.27443967e-02 -1.17375799e-01 -4.61738169e-01 4.17333692e-01
3.93904179e-01 1.11172545e+00 -6.77645087e-01 -2.80502021e-01
2.78793216e-01 6.72538638e-01 -3.40353698e-01 4.07460868e-01
1.09056585e-01 -9.33519751e-03 -5.60967386e-01 5.81065416e-01
8.10020208e-01 -7.38095641e-01 -3.49589884e-01 -4.70691472e-01
-7.35492930e-02 -1.90949321e-01 -1.14160407e+00 1.59774637e+00
-5.79359770e-01 4.85533774e-01 -2.64089555e-01 -1.21653759e+00
9.23888683e-01 9.34920907e-02 6.92301095e-02 -3.05402517e-01
-6.53272942e-02 2.77008623e-01 8.35774001e-03 -4.05637741e-01
1.23416791e-02 -4.31051515e-02 1.09673413e-02 4.02235299e-01
1.64294794e-01 1.81179672e-01 2.41787545e-02 5.20009287e-02
9.60153401e-01 -9.65662301e-02 2.14657143e-01 -4.47514266e-01
6.14260972e-01 -2.24551246e-01 1.99441761e-01 6.50065243e-01
1.01788640e-01 4.34232920e-01 5.38747907e-01 -8.47607732e-01
-8.31037104e-01 -1.25435019e+00 -4.43688750e-01 1.16236460e+00
-2.41170973e-02 -3.07802141e-01 -6.42420053e-01 -7.42113411e-01
2.12769881e-01 1.91231266e-01 -6.92403793e-01 -4.72069502e-01
-2.25614578e-01 -1.13785660e+00 6.92677975e-01 7.29690135e-01
7.76100218e-01 -8.73985171e-01 -2.60009468e-01 2.20517173e-01
9.69857201e-02 -9.30874586e-01 -5.63088775e-01 4.36589360e-01
-9.36602056e-01 -1.03982246e+00 -9.84095931e-01 -8.62633288e-01
6.06881678e-01 6.16562843e-01 6.37492597e-01 -6.07618056e-02
-5.07383406e-01 -1.76427871e-01 -4.51604933e-01 -2.35342726e-01
-6.45223334e-02 2.27204457e-01 1.23759560e-01 3.21288973e-01
2.96618015e-01 -5.49349904e-01 -7.75943398e-01 2.52665758e-01
-1.01442349e+00 -1.48852721e-01 9.78641629e-01 1.11002612e+00
1.43185228e-01 4.35087532e-01 8.12836707e-01 -7.14545131e-01
7.69948900e-01 -3.98114204e-01 -4.83921677e-01 3.96461368e-01
-6.51323617e-01 2.24807635e-01 1.15456402e+00 -1.00063300e+00
-8.81523550e-01 -3.98244560e-02 -3.36590856e-02 -4.78413165e-01
-1.90206572e-01 5.08173704e-01 6.48422092e-02 -3.65044624e-01
7.74876833e-01 2.99972028e-01 2.27757588e-01 -6.05404854e-01
2.83629477e-01 5.06331205e-01 1.31698176e-01 -5.14365792e-01
7.55018532e-01 3.98070693e-01 -2.49809504e-01 -6.01631284e-01
-9.04546678e-01 -4.64382827e-01 -6.53184652e-01 2.47224540e-01
6.30130470e-01 -9.44174767e-01 -6.73527002e-01 9.23216820e-01
-8.93742919e-01 -2.93223172e-01 3.24990437e-03 8.18663061e-01
-2.73449153e-01 3.17694992e-01 -9.60837364e-01 -6.34481847e-01
-2.81044215e-01 -9.94094431e-01 8.81486237e-01 5.62335134e-01
1.18303120e-01 -1.25603783e+00 -1.78477496e-01 -1.30847469e-01
8.45694125e-01 -1.57554261e-02 1.16954422e+00 -6.42542779e-01
-2.59516746e-01 -4.93964612e-01 -6.52352571e-01 6.42358243e-01
3.99936050e-01 -2.80329608e-03 -1.01797140e+00 -3.65526825e-01
-1.93693131e-01 -6.92903340e-01 1.16939831e+00 2.61299193e-01
1.35404456e+00 -5.33356592e-02 -3.02782655e-01 8.78928006e-01
1.36984313e+00 -1.96419194e-01 6.89143777e-01 3.33147615e-01
7.51002908e-01 4.15860474e-01 3.57272416e-01 2.66420811e-01
3.59932274e-01 3.81194949e-01 5.71750440e-02 -6.27348065e-01
2.99435765e-01 -4.68362123e-02 5.29642820e-01 5.84826827e-01
-1.33963808e-01 2.38118470e-01 -6.94898188e-01 1.80320293e-01
-2.04206157e+00 -9.44964707e-01 1.01094181e-02 2.13385177e+00
9.80172575e-01 6.60806745e-02 1.68445125e-01 -1.00995675e-01
5.46232760e-01 9.16448534e-02 -6.26633942e-01 -2.97088742e-01
-1.05111390e-01 4.38752681e-01 3.29267323e-01 -9.72065609e-03
-1.39965248e+00 1.10916114e+00 6.22146845e+00 1.25554049e+00
-1.44679892e+00 -1.90278843e-01 5.15377283e-01 2.69345790e-01
7.53569677e-02 -2.39213984e-02 -6.85460269e-01 6.04637206e-01
8.63754869e-01 1.93969056e-01 4.09305632e-01 1.01252937e+00
-2.02876702e-02 -1.83923393e-01 -9.33984280e-01 1.13498855e+00
-2.37028033e-01 -1.32339847e+00 2.23526582e-02 6.65676519e-02
3.59079778e-01 -1.41739370e-02 1.57022163e-01 5.67348957e-01
2.39556208e-01 -1.22873700e+00 2.69373357e-01 5.80642521e-01
6.14000142e-01 -1.03967118e+00 8.35198879e-01 6.16947293e-01
-1.07249105e+00 -1.46483898e-01 -1.00384438e+00 2.56350487e-02
-1.81897923e-01 1.10266566e+00 -5.40197670e-01 4.80565876e-01
7.99230039e-01 6.43090904e-01 -7.29805291e-01 1.14494812e+00
-1.53415695e-01 7.69414663e-01 -3.60455006e-01 -8.39533582e-02
4.40047383e-01 -3.53074074e-01 -2.25570485e-01 1.41686380e+00
2.27960005e-01 -2.04086155e-01 3.04772645e-01 7.81240523e-01
3.22236955e-01 3.28930318e-01 -6.20284438e-01 -5.19660711e-02
2.09105909e-01 1.61040425e+00 -8.37161064e-01 -3.48364472e-01
-6.70362175e-01 1.29087234e+00 8.76428008e-01 4.28133518e-01
-1.00801837e+00 -6.98596239e-01 8.07184100e-01 1.63252465e-02
4.20781612e-01 -2.15455383e-01 2.89796293e-01 -1.57575858e+00
-1.91789214e-02 -5.90789437e-01 2.71439701e-01 -4.21809822e-01
-1.80693436e+00 6.20220184e-01 -1.08771048e-01 -1.20574677e+00
5.06022349e-02 -1.04268706e+00 -9.71287847e-01 9.60433662e-01
-1.58052146e+00 -1.34277308e+00 -1.12075128e-01 8.27435136e-01
8.24699551e-02 -4.03737761e-02 1.02443695e+00 2.83414423e-01
-5.51621974e-01 7.16519058e-01 3.70438516e-01 4.43389684e-01
1.07494330e+00 -1.24084091e+00 2.95669317e-01 5.13055742e-01
8.72914717e-02 9.10141289e-01 2.26681810e-02 -2.58820415e-01
-1.50540936e+00 -9.83058751e-01 4.37123060e-01 -1.68195069e-01
9.78329718e-01 -5.12772083e-01 -1.35315645e+00 9.80855674e-02
-1.18503697e-01 4.53116834e-01 1.05873966e+00 2.93584913e-01
-7.94496298e-01 -3.06036800e-01 -7.23228931e-01 3.42893600e-01
6.02276504e-01 -7.30262876e-01 -3.38800460e-01 2.65143454e-01
4.93048519e-01 -7.87296146e-02 -9.32487667e-01 5.58071256e-01
5.18456042e-01 -1.09639096e+00 1.10539603e+00 -7.25735188e-01
2.42640302e-02 -3.40247780e-01 1.86265960e-01 -1.50770617e+00
-8.00184309e-01 -2.85813570e-01 -2.23391294e-01 1.02481163e+00
3.88286918e-01 -9.89435673e-01 5.66218019e-01 3.68107826e-01
1.11626178e-01 -1.16945887e+00 -5.71343541e-01 -9.03026640e-01
2.47307107e-01 -2.59910256e-01 6.04957283e-01 1.08267176e+00
-2.77917720e-02 2.68118620e-01 -5.46053886e-01 1.52544305e-01
5.22552371e-01 4.89360243e-01 6.59402192e-01 -1.34159124e+00
-3.19421619e-01 -6.48435771e-01 -4.67441410e-01 -1.02445555e+00
3.39243859e-01 -9.55738485e-01 -2.54320353e-01 -1.27376854e+00
2.82911867e-01 -6.11731350e-01 -5.66165924e-01 7.24318504e-01
-5.87676227e-01 2.08558083e-01 -1.49257213e-01 3.41863483e-01
-3.79929096e-01 4.89569515e-01 1.31719351e+00 2.68739890e-02
-1.52841121e-01 2.18575567e-01 -7.85660744e-01 7.93344975e-01
8.74975145e-01 -1.56922385e-01 -4.82216090e-01 -1.08187251e-01
1.03012323e-01 -3.06033641e-01 5.59752822e-01 -7.59940863e-01
2.95256764e-01 -6.04423508e-02 9.97101963e-01 -1.86072543e-01
1.18968494e-01 -6.74292207e-01 -1.88847259e-01 3.38671923e-01
-1.93512570e-02 -6.99878111e-02 4.96472232e-02 5.25260210e-01
-3.72156322e-01 -1.60043046e-01 8.71526539e-01 -6.86661759e-03
-8.56887758e-01 6.12047613e-01 -4.14499015e-01 -2.77668834e-01
9.92172599e-01 -2.06357434e-01 -4.75797132e-02 -8.84804130e-02
-5.32790303e-01 1.34275004e-01 5.06211042e-01 3.45919162e-01
7.20400393e-01 -1.44020319e+00 -5.39021730e-01 4.22491431e-01
1.54048324e-01 1.68741435e-01 3.36601615e-01 1.01224554e+00
-4.82615203e-01 1.79539487e-01 -1.10790409e-01 -5.04900098e-01
-6.91253662e-01 8.30759823e-01 2.00266257e-01 -3.06463927e-01
-7.67360806e-01 9.29661512e-01 4.78508145e-01 -4.79281574e-01
1.58583060e-01 -4.91572350e-01 -1.35690212e-01 1.26641676e-01
8.73949587e-01 9.91888121e-02 -5.49433865e-02 -1.10227451e-01
-2.73780674e-01 6.31826341e-01 -4.50636357e-01 4.68565047e-01
1.25827610e+00 3.50841641e-01 -2.36836195e-01 3.69522750e-01
1.34943569e+00 -9.62452069e-02 -1.34738481e+00 -2.86293924e-01
1.77742969e-02 -3.47211897e-01 4.06416841e-02 -6.41336262e-01
-1.04953349e+00 1.07332194e+00 3.88718069e-01 2.71189272e-01
1.14557910e+00 6.99668452e-02 8.85212600e-01 5.17909467e-01
3.21491539e-01 -9.40135300e-01 1.42257899e-01 6.71881437e-01
4.30680901e-01 -1.48835111e+00 -1.76729381e-01 -3.51213574e-01
-6.25513017e-01 1.64802277e+00 6.54377401e-01 -3.21211398e-01
9.15726900e-01 3.07809144e-01 9.46697667e-02 -8.30490084e-04
-6.67878270e-01 -1.83347553e-01 5.61483026e-01 4.19320464e-01
5.11264861e-01 1.66694045e-01 9.65900440e-03 6.53358817e-01
3.06093842e-02 -1.65766880e-01 1.19332112e-01 6.24025345e-01
-5.82897663e-01 -1.14102614e+00 -1.67634085e-01 6.40196741e-01
-3.20932657e-01 -2.45230436e-01 -7.64530599e-02 7.16342866e-01
-1.21624850e-01 4.75329399e-01 -1.77558273e-01 -4.74045038e-01
1.99985355e-01 1.10709161e-01 4.23160642e-01 -4.59092259e-01
-4.69634891e-01 1.71134481e-03 -3.57616365e-01 -5.62492013e-01
-2.36239195e-01 -1.71109691e-01 -1.27080762e+00 -3.70814145e-01
-6.09338522e-01 3.51839006e-01 4.33934033e-01 7.80786097e-01
3.30635071e-01 1.02321193e-01 8.25562775e-01 -8.43190312e-01
-9.80226338e-01 -8.82290781e-01 -7.79870272e-01 2.32762322e-01
2.33188346e-01 -7.68457055e-01 -3.86832297e-01 -1.60438567e-01] | [8.970918655395508, 2.439159631729126] |
bd8eb5fe-618c-4220-aabc-f4670850b386 | primitive3d-3d-object-dataset-synthesis-from | 2205.12627 | null | https://arxiv.org/abs/2205.12627v1 | https://arxiv.org/pdf/2205.12627v1.pdf | Primitive3D: 3D Object Dataset Synthesis from Randomly Assembled Primitives | Numerous advancements in deep learning can be attributed to the access to large-scale and well-annotated datasets. However, such a dataset is prohibitively expensive in 3D computer vision due to the substantial collection cost. To alleviate this issue, we propose a cost-effective method for automatically generating a large amount of 3D objects with annotations. In particular, we synthesize objects simply by assembling multiple random primitives. These objects are thus auto-annotated with part labels originating from primitives. This allows us to perform multi-task learning by combining the supervised segmentation with unsupervised reconstruction. Considering the large overhead of learning on the generated dataset, we further propose a dataset distillation strategy to remove redundant samples regarding a target dataset. We conduct extensive experiments for the downstream tasks of 3D object classification. The results indicate that our dataset, together with multi-task pretraining on its annotations, achieves the best performance compared to other commonly used datasets. Further study suggests that our strategy can improve the model performance by pretraining and fine-tuning scheme, especially for the dataset with a small scale. In addition, pretraining with the proposed dataset distillation method can save 86\% of the pretraining time with negligible performance degradation. We expect that our attempt provides a new data-centric perspective for training 3D deep models. | ['Yeow Meng Chee', 'Yuwei Wu', 'Zekun Tong', 'Henghui Ding', 'Xinke Li'] | 2022-05-25 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Primitive3D_3D_Object_Dataset_Synthesis_From_Randomly_Assembled_Primitives_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Primitive3D_3D_Object_Dataset_Synthesis_From_Randomly_Assembled_Primitives_CVPR_2022_paper.pdf | cvpr-2022-1 | ['3d-object-classification'] | ['computer-vision'] | [ 2.53962934e-01 1.90666512e-01 -2.15622131e-02 -4.86019045e-01
-8.92346203e-01 -4.13204640e-01 4.74029124e-01 -1.71502009e-02
-3.14168960e-01 4.43959087e-01 -1.24839291e-01 -1.46276161e-01
1.73451409e-01 -8.27006638e-01 -8.63330901e-01 -7.37255275e-01
3.17960858e-01 8.46581697e-01 3.53450835e-01 2.32758522e-01
1.35618553e-01 8.45609307e-01 -1.64684701e+00 4.21499014e-02
9.77565229e-01 1.08359492e+00 6.43178821e-01 2.66682297e-01
-3.97374183e-01 2.70362318e-01 -7.25583732e-01 -2.94605374e-01
6.48937047e-01 -1.68055758e-01 -7.61412859e-01 7.01745450e-01
4.75623727e-01 -6.09924972e-01 9.49609801e-02 7.60408342e-01
6.66155875e-01 -2.74200421e-02 7.14074552e-01 -1.07510293e+00
-1.38234571e-01 3.51486176e-01 -7.53977299e-01 -2.71700501e-01
-2.29985826e-02 1.74157605e-01 8.01647842e-01 -1.02551866e+00
6.19153321e-01 1.09118927e+00 5.11319697e-01 4.84564215e-01
-1.12592697e+00 -6.37264967e-01 1.38569549e-01 -2.25706026e-01
-1.37436390e+00 -3.34533989e-01 1.08577585e+00 -5.22195876e-01
7.29650736e-01 2.06900779e-02 5.32824337e-01 8.51546168e-01
-2.49474719e-01 9.66296613e-01 1.01620519e+00 -4.36552197e-01
1.73405752e-01 4.64074276e-02 -3.25293764e-02 7.96656728e-01
3.99966568e-01 -2.17667684e-01 -9.56733972e-02 1.15597904e-01
8.88716817e-01 5.70710301e-02 8.44103694e-02 -5.69004714e-01
-1.22489464e+00 6.41909063e-01 4.28765178e-01 3.63210328e-02
-3.37076664e-01 4.39418629e-02 3.26545238e-01 -1.51871964e-01
5.95200121e-01 4.77691382e-01 -6.58926547e-01 2.60103673e-01
-8.90999258e-01 2.13305965e-01 5.28212726e-01 1.23666346e+00
9.57123458e-01 -1.25790238e-02 -9.00944024e-02 1.19997013e+00
2.40209088e-01 3.56459081e-01 2.85918623e-01 -8.77379000e-01
5.50982594e-01 8.29284430e-01 -6.36620000e-02 -7.70946622e-01
-3.96448821e-01 -4.31831449e-01 -9.27404642e-01 1.75405249e-01
4.44316119e-01 -1.39428124e-01 -1.20373356e+00 1.45053041e+00
7.24935234e-01 2.41610944e-01 2.99973749e-02 7.46773660e-01
9.04391646e-01 4.03620929e-01 1.84365287e-02 -8.55226368e-02
1.27669060e+00 -1.04169106e+00 -3.55979264e-01 -1.40262678e-01
7.15168118e-01 -8.40949893e-01 1.15942550e+00 3.94623607e-01
-7.66550601e-01 -7.99768507e-01 -9.58977103e-01 -1.23164266e-01
-3.89860570e-02 4.86678213e-01 7.75688291e-01 4.54825968e-01
-5.69468498e-01 3.45116615e-01 -8.45502317e-01 -1.85529456e-01
9.50543582e-01 4.35768783e-01 -3.54240686e-01 -5.34025095e-02
-4.57378477e-01 5.71808875e-01 7.22004831e-01 4.27308530e-02
-1.14405560e+00 -5.45513511e-01 -7.83617795e-01 -1.98674649e-01
6.07318223e-01 -7.73459554e-01 1.26963866e+00 -5.91525793e-01
-1.56733620e+00 9.84387755e-01 6.91149905e-02 -2.74381965e-01
5.68763256e-01 -2.80759305e-01 2.04464495e-01 1.13245852e-01
2.35443577e-01 1.01072955e+00 8.44441414e-01 -1.60182035e+00
-6.86260879e-01 -4.15923089e-01 4.80221435e-02 2.53647476e-01
-2.71194100e-01 -3.04708123e-01 -8.55106711e-01 -7.29163527e-01
5.37122071e-01 -1.08952034e+00 -5.89725494e-01 1.99864302e-02
-6.20104194e-01 -2.47663215e-01 8.54845524e-01 -2.02527478e-01
6.87952340e-01 -2.11553621e+00 -2.25943904e-02 2.85506956e-02
2.48364091e-01 2.78275162e-01 -8.37458074e-02 6.45563602e-02
2.44128823e-01 4.56179231e-02 -4.30342346e-01 -6.82848334e-01
-7.04114661e-02 3.60730678e-01 -1.85034066e-01 2.33114704e-01
3.09862167e-01 7.83713698e-01 -6.76106751e-01 -7.43400276e-01
2.80469298e-01 5.94520047e-02 -7.08534002e-01 3.50254089e-01
-6.28309846e-01 6.16865575e-01 -6.57948256e-01 7.89359272e-01
7.98981845e-01 -3.97091687e-01 1.37756451e-03 -3.20163280e-01
1.21480621e-01 1.77918762e-01 -1.31475317e+00 2.01876831e+00
-4.60528374e-01 2.22685009e-01 -1.32575154e-01 -1.17935371e+00
1.21060550e+00 1.42209619e-01 6.30175114e-01 -3.26009899e-01
2.28612393e-01 3.86024803e-01 -1.31681263e-01 -5.48707664e-01
4.94014293e-01 -2.09923446e-01 -2.09932372e-01 4.90338832e-01
1.44447267e-01 -6.17337644e-01 7.61418939e-02 -1.99894123e-02
7.63851225e-01 4.67264473e-01 1.61329538e-01 -2.91694347e-02
2.84828722e-01 1.85240582e-01 5.82403421e-01 4.97796625e-01
-7.29817376e-02 5.94067574e-01 2.18246207e-01 -3.92683148e-01
-1.07257450e+00 -6.54429436e-01 -1.61334887e-01 7.03233004e-01
2.94042677e-01 -1.70527354e-01 -7.64226139e-01 -1.06784987e+00
4.15740684e-02 4.51842815e-01 -2.15181574e-01 4.03043926e-02
-6.12832069e-01 -9.39878941e-01 5.21199524e-01 6.15786910e-01
7.68686414e-01 -9.71544623e-01 -5.73872328e-01 1.36369556e-01
1.23812947e-02 -1.29520440e+00 -1.71050847e-01 2.83332556e-01
-1.32431090e+00 -9.21416342e-01 -7.15764642e-01 -1.00885189e+00
9.93202865e-01 5.25171518e-01 9.20618951e-01 1.80272572e-02
-1.50198892e-01 6.20698892e-02 -3.41798156e-01 -4.08941835e-01
-4.15124238e-01 2.82162786e-01 4.07934468e-03 -1.10343099e-01
1.63167998e-01 -5.29828370e-01 -4.75974232e-01 4.19242620e-01
-9.15244520e-01 3.88835967e-01 9.62042212e-01 6.59829080e-01
1.05044401e+00 2.75562853e-01 6.52494967e-01 -1.08059776e+00
1.89006299e-01 -2.55977362e-01 -7.21446693e-01 6.73435479e-02
-3.95483106e-01 2.79695392e-01 5.59350729e-01 -4.76241797e-01
-1.30332577e+00 6.05659664e-01 -3.04936916e-01 -5.91074347e-01
-5.73041320e-01 1.35094225e-01 -5.76960564e-01 4.57469374e-02
4.71049398e-01 1.09737568e-01 -1.31274000e-01 -8.47360432e-01
4.20902342e-01 6.75274253e-01 3.54968458e-01 -8.70974422e-01
8.82197559e-01 4.36963201e-01 1.83728799e-01 -7.12172508e-01
-9.98314142e-01 -2.92477012e-01 -9.63245630e-01 1.22682070e-02
8.33824515e-01 -9.97380078e-01 -3.61064821e-01 6.71701074e-01
-1.24129951e+00 -4.37250644e-01 -3.47854316e-01 5.19749105e-01
-5.13081670e-01 3.89856249e-01 -4.45593327e-01 -5.94424546e-01
-2.64885217e-01 -1.26399970e+00 1.38773096e+00 -2.46060565e-02
6.46756738e-02 -6.36777759e-01 -1.25922278e-01 5.86874545e-01
-1.73795372e-01 3.63081157e-01 9.15223002e-01 -8.11119080e-01
-9.51526523e-01 -1.36620551e-01 -3.55923831e-01 4.77412760e-01
3.56467515e-01 -1.39037475e-01 -1.01493061e+00 -3.46772093e-03
7.12743327e-02 -4.62325275e-01 6.61200106e-01 1.99449763e-01
1.57930863e+00 6.28886074e-02 -3.86191785e-01 5.81019223e-01
1.38376534e+00 1.97178915e-01 2.00848162e-01 2.10583299e-01
1.04287350e+00 5.72855413e-01 9.92370486e-01 5.15261889e-01
4.00581837e-01 5.93974054e-01 4.71129745e-01 -1.71392336e-01
-3.38796407e-01 -3.91761750e-01 -1.43882632e-01 8.16495538e-01
-8.64171684e-02 -2.35510603e-01 -1.06037462e+00 4.54285115e-01
-1.63650584e+00 -6.52556241e-01 -1.29143164e-01 2.14107609e+00
9.67783093e-01 3.76161367e-01 1.02690898e-01 2.66600817e-01
5.77664614e-01 2.33162520e-03 -5.92619836e-01 1.93287715e-01
1.06319778e-01 2.41373092e-01 4.36740696e-01 2.75483638e-01
-1.20219111e+00 1.10622859e+00 5.89475489e+00 9.69832718e-01
-1.11943495e+00 -8.68476927e-02 7.66364276e-01 9.47244465e-02
-2.24615127e-01 -1.51145622e-01 -1.22645330e+00 3.38930279e-01
4.72587287e-01 2.00971067e-01 -1.30593896e-01 1.18462646e+00
2.72474825e-01 -1.98750898e-01 -1.15832961e+00 1.02238846e+00
4.89132293e-02 -1.27234006e+00 3.31335723e-01 1.78013355e-01
9.65207875e-01 -1.97616920e-01 -3.68907988e-01 1.93029419e-01
1.95089698e-01 -5.83650470e-01 7.17657924e-01 2.58784741e-01
6.72702372e-01 -6.51976764e-01 6.28452182e-01 6.88980043e-01
-1.06380153e+00 1.80349454e-01 -4.42163020e-01 3.66194248e-02
1.62859216e-01 9.81509566e-01 -1.30587649e+00 6.71367228e-01
5.01401782e-01 6.83887243e-01 -5.38637221e-01 1.07588959e+00
-2.12243050e-01 4.26956683e-01 -4.81384158e-01 1.06734283e-01
6.56152293e-02 -1.87816828e-01 3.13583225e-01 9.14508402e-01
4.22558486e-01 1.32055864e-01 5.37171125e-01 6.71031535e-01
-3.59652072e-01 7.25936443e-02 -6.98699117e-01 1.70431778e-01
4.98832911e-01 1.25548744e+00 -1.11162710e+00 -5.19022524e-01
-3.03619564e-01 9.22420382e-01 2.95705169e-01 8.21653008e-03
-8.25151563e-01 -2.29136288e-01 2.96309143e-01 7.65950903e-02
4.44437981e-01 -4.91873562e-01 -6.14053488e-01 -1.01370800e+00
9.96265858e-02 -6.62758172e-01 1.83679730e-01 -5.30001998e-01
-1.18824887e+00 6.11259580e-01 1.30217299e-01 -1.43358397e+00
-9.38107297e-02 -4.49771345e-01 -2.02371359e-01 4.51667696e-01
-1.50088215e+00 -1.13348937e+00 -5.74418366e-01 4.72759247e-01
9.13690746e-01 -4.74159904e-02 5.62012076e-01 4.39776659e-01
-6.60198748e-01 4.52783704e-01 -3.07235181e-01 4.35060523e-02
6.07356250e-01 -1.25254893e+00 3.35460216e-01 7.56836712e-01
1.92113116e-01 2.70639539e-01 2.73629487e-01 -6.81751966e-01
-1.19011271e+00 -1.36352980e+00 5.29650211e-01 -3.21709275e-01
1.51223704e-01 -4.96928662e-01 -7.65021801e-01 5.66622972e-01
-3.34693134e-01 1.05045540e-02 6.87483072e-01 -8.77390876e-02
-4.92817461e-02 -1.68788508e-01 -1.19989765e+00 4.55352485e-01
1.29757607e+00 -2.16124758e-01 -6.31908476e-01 4.68239158e-01
9.17725027e-01 -6.68447495e-01 -9.36816037e-01 6.73044264e-01
2.94439137e-01 -7.61079133e-01 9.26393151e-01 -1.71936303e-01
5.62933028e-01 -4.43043262e-01 -2.81147569e-01 -1.01548767e+00
5.22302575e-02 -2.58750111e-01 -6.68159351e-02 1.50702834e+00
2.74891794e-01 -3.69019926e-01 1.08440590e+00 5.92414975e-01
-5.10380864e-01 -8.31563830e-01 -5.33412695e-01 -8.38340342e-01
-1.28001109e-01 -4.98685092e-01 8.33096623e-01 8.15443575e-01
-7.22478926e-01 4.60297793e-01 -3.38052779e-01 2.50018388e-01
6.27458930e-01 5.22190511e-01 1.32865667e+00 -1.34261179e+00
-2.87956297e-01 -1.95688009e-01 -1.46187037e-01 -1.55991554e+00
4.34277356e-02 -9.48439181e-01 3.76073092e-01 -1.54907465e+00
5.71272634e-02 -1.18968558e+00 1.38812169e-01 5.81470311e-01
-1.55289695e-01 3.70221317e-01 9.34167951e-02 4.71898496e-01
-4.67098057e-01 6.25190020e-01 1.80793440e+00 -4.22950322e-03
-4.70360488e-01 1.97682485e-01 -5.74850976e-01 8.64255905e-01
7.59076536e-01 -5.50851703e-01 -5.67587197e-01 -7.02833772e-01
-4.50783253e-01 -7.37136826e-02 2.99308926e-01 -9.52601135e-01
-1.24527380e-01 -2.14719057e-01 3.81321043e-01 -9.97701645e-01
4.85019714e-01 -9.69342172e-01 1.05204530e-01 2.57064253e-01
-4.66691218e-02 -4.12390083e-01 2.15243876e-01 4.72414136e-01
-1.90157011e-01 -3.35982800e-01 7.62365758e-01 -2.87847131e-01
-7.54500806e-01 5.15428543e-01 -1.09982295e-02 -9.35592875e-02
1.16900301e+00 -3.46971631e-01 9.73793194e-02 2.15875298e-01
-5.59111297e-01 8.98611993e-02 5.16366959e-01 2.05720663e-01
4.67138380e-01 -1.26065958e+00 -4.85563219e-01 3.77593279e-01
-7.76573224e-03 1.00205910e+00 9.13918391e-02 4.04337674e-01
-5.75653195e-01 3.24397713e-01 -2.40090013e-01 -9.22742844e-01
-1.03928053e+00 4.65503007e-01 -9.98021476e-03 -2.05208853e-01
-7.23685563e-01 8.28704178e-01 2.50541836e-01 -4.77981508e-01
2.95615554e-01 -5.45640588e-01 8.80442262e-02 1.31585464e-01
-1.47400675e-02 2.63224602e-01 1.50034055e-01 -3.91359448e-01
-2.66807526e-01 6.70626163e-01 -7.88884014e-02 1.44118503e-01
1.52436829e+00 9.45041776e-02 1.13415219e-01 2.26587281e-01
1.10411704e+00 -6.46639103e-03 -1.48374069e+00 -2.18634501e-01
8.16548057e-03 -5.74602187e-01 -1.00679629e-01 -4.50217605e-01
-1.19196463e+00 8.47675145e-01 2.84692645e-01 -2.69217864e-02
1.19507527e+00 2.48035014e-01 9.61589754e-01 5.73489308e-01
4.90376472e-01 -8.83179963e-01 4.13453370e-01 2.64184445e-01
6.75659537e-01 -1.41653252e+00 1.31418243e-01 -6.83672369e-01
-5.33011496e-01 8.85710657e-01 9.09098446e-01 -2.26470441e-01
6.28866494e-01 1.38281167e-01 1.39981583e-02 -2.66379178e-01
-3.27597708e-01 -2.87526995e-01 2.00593129e-01 4.66310918e-01
1.07334025e-01 -3.78284976e-02 -1.55929580e-01 5.27858734e-01
-1.83775395e-01 -5.16947061e-02 3.48071992e-01 8.84654582e-01
-5.30430496e-01 -1.38327873e+00 -2.84920156e-01 5.32877266e-01
-1.84539586e-01 2.11031288e-01 -2.48161957e-01 8.35616112e-01
4.11104441e-01 6.25290573e-01 -1.75716374e-02 -3.02680790e-01
4.13110346e-01 5.98363355e-02 5.05097866e-01 -9.66336668e-01
-1.95739701e-01 3.39259535e-01 6.65093735e-02 -2.90544629e-01
-6.82460010e-01 -5.88295996e-01 -1.38216162e+00 1.12224050e-01
-5.32401145e-01 -1.38155401e-01 6.35351121e-01 1.00682187e+00
4.85175610e-01 4.88234043e-01 7.62542129e-01 -1.14660132e+00
-4.04402703e-01 -8.67066920e-01 -3.69107008e-01 4.24595892e-01
-6.24300260e-03 -9.68401432e-01 -7.82952905e-02 2.48735398e-01] | [8.08565902709961, -3.0518105030059814] |
41d3249a-a0b6-415e-a5c4-e8aca8d91b97 | dimensionality-reduction-for-general-kde-mode | 2305.18755 | null | https://arxiv.org/abs/2305.18755v3 | https://arxiv.org/pdf/2305.18755v3.pdf | Dimensionality Reduction for General KDE Mode Finding | Finding the mode of a high dimensional probability distribution $D$ is a fundamental algorithmic problem in statistics and data analysis. There has been particular interest in efficient methods for solving the problem when $D$ is represented as a mixture model or kernel density estimate, although few algorithmic results with worst-case approximation and runtime guarantees are known. In this work, we significantly generalize a result of (LeeLiMusco:2021) on mode approximation for Gaussian mixture models. We develop randomized dimensionality reduction methods for mixtures involving a broader class of kernels, including the popular logistic, sigmoid, and generalized Gaussian kernels. As in Lee et al.'s work, our dimensionality reduction results yield quasi-polynomial algorithms for mode finding with multiplicative accuracy $(1-\epsilon)$ for any $\epsilon > 0$. Moreover, when combined with gradient descent, they yield efficient practical heuristics for the problem. In addition to our positive results, we prove a hardness result for box kernels, showing that there is no polynomial time algorithm for finding the mode of a kernel density estimate, unless $\mathit{P} = \mathit{NP}$. Obtaining similar hardness results for kernels used in practice (like Gaussian or logistic kernels) is an interesting future direction. | ['Cas Widdershoven', 'Christopher Musco', 'Xinyu Luo'] | 2023-05-30 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [-4.10953224e-01 -1.36404425e-01 -8.18282589e-02 -5.99056818e-02
-8.83727312e-01 -8.58557045e-01 -7.22089633e-02 2.54034609e-01
-4.69945997e-01 5.79735637e-01 -4.67483819e-01 -6.83440149e-01
-6.63562357e-01 -9.36426938e-01 -6.08881831e-01 -8.49772871e-01
-5.67611158e-01 5.46062946e-01 2.70241350e-01 1.49789602e-01
2.70101935e-01 7.17865348e-01 -1.33659410e+00 -3.50925773e-01
1.00432587e+00 1.06654167e+00 -6.37955815e-02 1.06439829e+00
-1.15716923e-02 3.14495355e-01 -2.25374833e-01 -7.09251761e-01
3.94454539e-01 -2.41271809e-01 -7.99973309e-01 -8.26949347e-03
2.22157508e-01 -5.60643375e-02 -3.02496523e-01 1.44041216e+00
3.79531324e-01 3.22724700e-01 1.11328912e+00 -1.57899463e+00
-5.98865211e-01 5.92158198e-01 -8.25110793e-01 2.10887060e-01
3.23464349e-02 -1.80759326e-01 6.80421948e-01 -5.29386342e-01
-5.45714051e-02 1.04038036e+00 8.17861021e-01 2.80339509e-01
-1.44161570e+00 -5.76404095e-01 -1.00206435e-01 1.49919614e-01
-1.79214621e+00 -1.19365439e-01 4.38558429e-01 -4.49150771e-01
1.08000898e+00 5.34850955e-01 1.94310352e-01 1.27017722e-01
-1.99702576e-01 5.66202819e-01 9.31863487e-01 -5.23096561e-01
4.98149246e-01 3.87509525e-01 4.84436691e-01 8.90613258e-01
6.71086907e-01 -3.32407594e-01 1.98946014e-01 -7.55036175e-01
8.27653289e-01 1.67021435e-02 -2.19776362e-01 -4.54642147e-01
-6.93664908e-01 9.67735648e-01 -1.40144587e-01 2.39367504e-02
-4.68439125e-02 3.97774398e-01 1.70865968e-01 2.54161954e-01
6.00309491e-01 2.16342419e-01 -5.18991768e-01 -4.84069347e-01
-7.21699655e-01 3.69075835e-01 1.26563346e+00 1.00082111e+00
8.18280578e-01 -1.15392573e-01 3.52125525e-01 8.93695176e-01
4.88294549e-02 8.15065503e-01 -5.10023497e-02 -9.20536876e-01
5.04492462e-01 2.73495615e-01 5.08956730e-01 -9.62318420e-01
-3.93989295e-01 9.66930687e-02 -8.38594556e-01 4.61487472e-03
9.90475416e-01 -2.58567631e-01 -3.73862773e-01 1.84710670e+00
3.64193797e-01 1.68782324e-01 1.46891559e-02 5.09919405e-01
-7.60101434e-03 7.17986822e-01 -2.50737995e-01 -3.69448334e-01
1.33739567e+00 -4.19732302e-01 -2.94134706e-01 1.77871093e-01
9.23136234e-01 -9.08389449e-01 7.96955526e-01 6.09780252e-01
-1.17034864e+00 -2.50248238e-02 -5.59238732e-01 2.07826287e-01
-3.57358426e-01 3.49553317e-01 1.12405527e+00 1.08300233e+00
-1.09849072e+00 4.18237448e-01 -9.16532874e-01 -3.02607983e-01
3.03533435e-01 5.37982166e-01 -1.07650414e-01 -1.18394621e-01
-7.02775300e-01 5.69343567e-01 2.34450966e-01 -1.09798096e-01
-1.87083274e-01 -8.66983771e-01 -7.69040227e-01 3.36612714e-03
2.05753863e-01 -2.64522374e-01 1.04834557e+00 -2.05308437e-01
-1.24956894e+00 6.48204923e-01 -2.99623698e-01 -5.25949180e-01
2.25382507e-01 1.88885145e-02 -1.58611909e-01 2.89489269e-01
-2.29882941e-01 8.88561532e-02 6.39647424e-01 -6.57759368e-01
-7.03273296e-01 -4.51734006e-01 2.43531197e-01 -1.46297023e-01
-3.82662714e-01 7.38714188e-02 -2.67286688e-01 -2.42294759e-01
6.79173023e-02 -9.77139115e-01 -5.33760905e-01 -1.82022259e-01
-4.03448790e-01 -5.31940520e-01 3.75558168e-01 -4.47503299e-01
1.53605592e+00 -2.23875880e+00 -9.58558079e-03 6.46057069e-01
3.30976844e-01 9.00835544e-02 2.96276987e-01 4.88100886e-01
-9.15428922e-02 1.80462319e-02 -3.73786956e-01 -1.68053225e-01
5.18075883e-01 -2.01890677e-01 -4.06341314e-01 9.51525509e-01
-1.31493270e-01 5.11652946e-01 -5.46053767e-01 -2.35247597e-01
-3.43703330e-02 3.25991422e-01 -5.90019703e-01 -1.03780724e-01
2.62997393e-03 -6.31753802e-01 -7.34381452e-02 5.46103120e-01
1.28658104e+00 -3.80448699e-01 -7.29564354e-02 -2.17698455e-01
-1.61534362e-02 -2.41141260e-01 -1.76848090e+00 1.03483295e+00
-3.73079121e-01 3.36671591e-01 3.45604241e-01 -1.34791362e+00
6.35753214e-01 -1.51834479e-02 5.18881023e-01 1.25731334e-01
6.33076206e-02 4.29584175e-01 -2.56208718e-01 -3.17861229e-01
3.75480890e-01 -3.42200667e-01 -3.08660656e-01 6.32916093e-01
-7.17274249e-02 -1.87578753e-01 5.04276037e-01 2.43912935e-01
1.19670880e+00 -4.21864361e-01 -3.85574847e-02 -4.16805059e-01
2.36781999e-01 9.85879600e-02 7.10748136e-02 7.74466336e-01
-1.25525117e-01 2.80760944e-01 1.05531204e+00 -1.69579254e-03
-1.01195812e+00 -1.34193182e+00 -3.95358413e-01 9.66123939e-01
-2.62019239e-05 -3.98120463e-01 -1.00292206e+00 -3.89819860e-01
1.72297716e-01 5.67852199e-01 -6.06261134e-01 -1.57380030e-01
-2.33653173e-01 -1.32017910e+00 7.02499151e-01 5.70813537e-01
4.39708978e-01 -2.50559688e-01 -3.09802711e-01 3.85481343e-02
2.82748818e-01 -9.77132082e-01 -5.92139363e-01 5.12808502e-01
-9.89347935e-01 -1.26743257e+00 -8.54042590e-01 -7.40244865e-01
7.42499411e-01 2.36655712e-01 6.04825139e-01 -4.72817510e-01
-4.10709649e-01 6.38904870e-01 2.01057401e-02 -2.30753049e-01
-1.30079225e-01 -9.00145546e-02 1.99442774e-01 -2.70538241e-01
7.68431783e-01 -4.96155858e-01 -5.43857217e-01 2.93772668e-01
-9.95868266e-01 -4.94639188e-01 3.00991923e-01 4.61169958e-01
5.41344166e-01 7.88920701e-01 2.97338873e-01 -7.32717216e-01
8.60732019e-01 -7.16951013e-01 -1.16496146e+00 7.69061968e-02
-4.98820633e-01 2.51597434e-01 8.82488847e-01 -7.32686698e-01
-5.41338742e-01 1.45893797e-01 2.06972305e-02 -5.47604501e-01
-1.96855903e-01 2.99315274e-01 -1.09094732e-01 -9.14854556e-02
6.44272983e-01 4.99664128e-01 5.92500754e-02 -5.27442813e-01
3.93208474e-01 5.40566385e-01 3.97253722e-01 -7.75708258e-01
7.80191839e-01 5.93442380e-01 2.89891034e-01 -1.04671013e+00
-6.69302940e-01 -7.24705219e-01 -1.99009851e-01 4.34864134e-01
5.98608017e-01 -5.08712709e-01 -1.18783891e+00 4.44829017e-01
-8.27507734e-01 -4.12065268e-01 -1.81607589e-01 7.05037534e-01
-9.03933465e-01 6.48737431e-01 -6.08803332e-01 -1.37703931e+00
-1.56570990e-02 -9.35487509e-01 7.45981634e-01 1.69328436e-01
8.16059336e-02 -1.12989736e+00 1.66246057e-01 3.47100385e-02
4.33250278e-01 -2.56355494e-01 1.23530853e+00 -7.22001195e-01
-2.35722333e-01 -4.43442404e-01 -5.43652236e-01 3.60164493e-01
-6.15036562e-02 -6.22959957e-02 -5.61240256e-01 -1.69081688e-01
-4.02622111e-03 -8.64369050e-02 7.99112856e-01 6.52752697e-01
1.42682552e+00 -4.86950547e-01 -3.72668445e-01 7.17789531e-01
1.40909314e+00 -5.45017608e-02 6.19319856e-01 -2.44474132e-03
4.17315036e-01 2.48181418e-01 3.08351219e-01 6.65824234e-01
3.51634532e-01 3.51051062e-01 -2.96043884e-02 2.40265548e-01
4.87696856e-01 2.29294300e-01 3.00958097e-01 2.15683401e-01
-1.21780038e-01 -7.80188888e-02 -7.26457953e-01 5.14549255e-01
-1.76546896e+00 -1.07778418e+00 -3.13461483e-01 2.71731186e+00
9.06817973e-01 2.75438316e-02 6.96658432e-01 2.07136869e-01
9.02579486e-01 -5.64644694e-01 -4.20061827e-01 -6.28327847e-01
-2.08499152e-02 5.71376026e-01 9.28727329e-01 5.76656878e-01
-1.23656058e+00 7.78541207e-01 6.64487934e+00 1.42877424e+00
-6.94529831e-01 -2.13293154e-02 4.54551131e-01 -2.03570649e-01
-1.68061346e-01 -4.93924543e-02 -9.98264790e-01 5.56483746e-01
1.04022241e+00 -2.42269441e-01 6.07529759e-01 1.20715261e+00
-8.68140608e-02 -5.29355645e-01 -8.85346472e-01 1.29314280e+00
-1.85780570e-01 -9.85325217e-01 -6.50275946e-01 6.29848182e-01
5.39753377e-01 -5.14101744e-01 3.96003991e-01 3.60635519e-01
4.87857342e-01 -9.85265374e-01 1.19237624e-01 2.98775941e-01
6.63015783e-01 -1.36725295e+00 5.28955519e-01 3.86238217e-01
-1.37282538e+00 2.06087697e-02 -9.94922519e-01 6.46539852e-02
-4.73621115e-03 1.05464590e+00 -6.44903958e-01 5.70378043e-02
5.32935619e-01 -9.96875204e-03 -3.44198108e-01 1.22207987e+00
4.00062621e-01 7.47892261e-01 -1.11474705e+00 -1.98174223e-01
2.24148203e-02 -4.04854000e-01 3.12714309e-01 1.36866605e+00
5.48339605e-01 2.66810447e-01 -3.33846584e-02 8.80916417e-01
1.38244823e-01 1.63860306e-01 -5.34555435e-01 -3.26173782e-01
5.03831983e-01 1.23236132e+00 -1.03027713e+00 -2.04594359e-01
-3.20922822e-01 1.07649970e+00 3.72587353e-01 3.89466047e-01
-9.70489979e-01 -8.80081356e-01 9.08826172e-01 1.40032813e-01
7.44719028e-01 -6.43458247e-01 -2.35580564e-01 -1.07000625e+00
5.41843772e-02 -2.56807476e-01 6.13069654e-01 -1.10309646e-01
-1.59536564e+00 1.20806217e-01 5.44969499e-01 -8.28141391e-01
-1.78702712e-01 -8.84245038e-01 -4.46454942e-01 9.52048063e-01
-8.70479405e-01 -6.26781046e-01 3.77065271e-01 5.59395134e-01
-1.72640737e-02 1.02962792e-01 8.23968291e-01 1.12070687e-01
-4.91489500e-01 8.03761840e-01 5.34747481e-01 -1.64510664e-02
4.22375113e-01 -1.57148898e+00 -2.52633113e-02 8.04389656e-01
-1.81077987e-01 6.64570630e-01 7.78753519e-01 -5.29750168e-01
-1.91629636e+00 -8.03197265e-01 5.59481621e-01 -2.19332427e-01
9.42819178e-01 -2.47976184e-01 -7.78396428e-01 2.89811164e-01
-2.67489731e-01 -5.38380556e-02 1.03815579e+00 1.97620258e-01
-4.06682551e-01 4.32642363e-02 -1.49241471e+00 5.04279017e-01
8.33156049e-01 -4.33455557e-01 -7.04866871e-02 3.53015602e-01
5.35330713e-01 -2.54871905e-01 -1.07532310e+00 1.71290010e-01
3.31863493e-01 -7.76057184e-01 1.06624901e+00 -6.52037501e-01
-5.56712821e-02 -2.40693212e-01 -4.66395348e-01 -9.56423163e-01
-1.73199907e-01 -7.21007347e-01 -7.66999424e-01 9.28222477e-01
5.23644984e-01 -7.90868104e-01 7.61599660e-01 1.02880931e+00
2.89474785e-01 -1.09277868e+00 -9.52661157e-01 -1.12669539e+00
4.12453026e-01 -9.83749330e-01 3.60149384e-01 6.44869089e-01
3.14983845e-01 -1.42667010e-01 -1.07799232e-01 4.72154915e-01
8.25917840e-01 1.18973821e-01 7.75623858e-01 -9.56491888e-01
-7.33879745e-01 -7.80601263e-01 -6.37629271e-01 -1.33796012e+00
2.11765066e-01 -7.76640058e-01 -2.32843518e-01 -1.20293856e+00
3.23010504e-01 -8.42403233e-01 -4.76380326e-02 2.12674066e-01
-5.18388785e-02 1.56929016e-01 -1.83639944e-01 -1.16774313e-01
-6.65810347e-01 1.49707928e-01 6.74512684e-01 2.46742219e-01
-3.22329462e-01 4.20389473e-01 -8.44125569e-01 8.75755250e-01
8.80651236e-01 -3.46943855e-01 -4.20248061e-01 1.00611657e-01
5.30824125e-01 3.85087729e-02 3.54729563e-01 -8.47729027e-01
2.13402793e-01 -3.20170492e-01 1.19290136e-01 -6.80678010e-01
4.04170007e-01 -6.33655250e-01 -1.07308269e-01 2.84491122e-01
-2.91602071e-02 2.11418852e-01 3.62350106e-01 6.94909573e-01
4.14455682e-01 -6.58726931e-01 7.99347162e-01 2.75119603e-01
-2.49024957e-01 4.13522631e-01 -6.08162582e-01 3.21853429e-01
1.12149251e+00 -2.22380176e-01 -3.87269944e-01 -3.79665643e-01
-7.81359851e-01 -8.38725492e-02 4.44929421e-01 -3.18556845e-01
5.56515932e-01 -1.13884473e+00 -4.37510639e-01 -2.77539697e-02
-1.96218222e-01 3.27984546e-03 3.74333888e-01 1.28942776e+00
-5.93868971e-01 4.16815847e-01 5.02718866e-01 -4.77261335e-01
-9.25352037e-01 6.84000194e-01 2.33444497e-01 -1.80854276e-01
-1.16841778e-01 1.06973612e+00 1.38117522e-02 -1.21838860e-02
3.09614986e-01 -6.02842569e-01 5.21124959e-01 -1.98537737e-01
7.47826815e-01 9.26836193e-01 -1.36602804e-01 -3.96334156e-02
-4.35702264e-01 3.82519066e-01 -9.14448500e-02 -5.01021564e-01
1.33182859e+00 -6.43071160e-02 -3.41425270e-01 3.42908144e-01
1.48814392e+00 2.70395905e-01 -9.17297482e-01 -2.20046174e-02
-1.94353014e-01 -3.29764038e-01 -2.79056996e-01 -2.82124788e-01
-7.95734227e-01 6.98336422e-01 5.55633783e-01 7.16219306e-01
1.24612427e+00 5.33332825e-01 7.04906166e-01 7.02544093e-01
3.53213608e-01 -1.07435703e+00 -2.90723622e-01 5.67513585e-01
2.60199219e-01 -9.16157305e-01 -1.57018170e-01 -4.90481853e-01
-3.69831085e-01 1.20852089e+00 4.24588799e-01 -3.25857699e-01
1.12329888e+00 4.31135595e-01 -6.74120486e-01 -6.93984255e-02
-4.31033939e-01 -4.31038052e-01 1.99076340e-01 6.44051015e-01
2.97272384e-01 2.14413539e-01 -3.46799701e-01 9.42202330e-01
-3.74109954e-01 -1.73671275e-01 4.09846932e-01 8.65077138e-01
-7.55977392e-01 -1.02428639e+00 -5.26803255e-01 7.82976091e-01
-4.47581440e-01 -2.76410580e-01 3.94266844e-02 5.22533000e-01
-8.76054764e-02 8.49258959e-01 -3.06950901e-02 -2.34234467e-01
5.54813119e-03 2.62924910e-01 8.07036221e-01 -3.78945559e-01
7.78155029e-02 1.16322346e-01 -2.56336719e-01 -2.33317569e-01
1.01279467e-01 -5.32005072e-01 -1.39996231e+00 -8.18314195e-01
-4.64754015e-01 5.10313988e-01 6.25340879e-01 8.79459083e-01
2.96075195e-01 -2.67491966e-01 5.72622359e-01 -3.52304697e-01
-8.09119821e-01 -7.58207917e-01 -1.10794747e+00 -1.67282104e-01
-2.16824654e-02 -5.35777271e-01 -7.79638171e-01 -3.88736986e-02] | [7.349287986755371, 4.181121349334717] |
dd509644-f755-4208-b17c-4dc7986a698c | homological-neural-networks-a-sparse | 2306.15337 | null | https://arxiv.org/abs/2306.15337v1 | https://arxiv.org/pdf/2306.15337v1.pdf | Homological Neural Networks: A Sparse Architecture for Multivariate Complexity | The rapid progress of Artificial Intelligence research came with the development of increasingly complex deep learning models, leading to growing challenges in terms of computational complexity, energy efficiency and interpretability. In this study, we apply advanced network-based information filtering techniques to design a novel deep neural network unit characterized by a sparse higher-order graphical architecture built over the homological structure of underlying data. We demonstrate its effectiveness in two application domains which are traditionally challenging for deep learning: tabular data and time series regression problems. Results demonstrate the advantages of this novel design which can tie or overcome the results of state-of-the-art machine learning and deep learning models using only a fraction of parameters. | ['Tomaso Aste', 'Antonio Briola', 'Yuanrong Wang'] | 2023-06-27 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [ 2.66365916e-01 3.66421528e-02 -2.85907179e-01 -1.93720460e-01
-1.28481448e-01 -5.15697896e-01 6.81711555e-01 4.04170066e-01
-2.99371123e-01 7.52672315e-01 -1.50816932e-01 -8.12357903e-01
-5.98034739e-01 -7.73878098e-01 -7.69964635e-01 -5.91526866e-01
-5.97509623e-01 4.35194433e-01 -6.35383949e-02 -2.94610143e-01
2.14153245e-01 8.83016884e-01 -1.34927952e+00 3.06550741e-01
8.75535309e-01 1.41911280e+00 -2.45191529e-01 3.81056964e-01
-1.03847601e-01 1.00850070e+00 -3.68145347e-01 -4.52557087e-01
2.82231718e-01 -2.09177434e-01 -5.17805874e-01 -4.50805843e-01
3.40197563e-01 -3.89178619e-02 -8.35353911e-01 7.84752011e-01
3.84725749e-01 1.89176545e-01 7.12740600e-01 -1.10709214e+00
-6.67488694e-01 5.86198866e-01 -4.27879721e-01 1.12120189e-01
-2.32978955e-01 -2.01136157e-01 1.33750069e+00 -7.97281444e-01
5.33946931e-01 1.03945971e+00 9.99072909e-01 2.69089580e-01
-1.46160865e+00 -6.46867692e-01 -2.62036454e-02 7.02051818e-01
-1.31795323e+00 -2.33632252e-01 1.02612221e+00 -3.65798622e-01
1.03185105e+00 1.86124146e-01 8.02366912e-01 8.63265574e-01
3.66133898e-01 6.36348784e-01 7.08077669e-01 -3.00868362e-01
4.88410443e-01 -2.30538368e-01 2.44007975e-01 1.13481140e+00
6.34655416e-01 5.67532852e-02 -6.08983338e-01 -8.54657516e-02
9.40145910e-01 2.05113590e-01 1.58804983e-01 -5.33814549e-01
-6.09166443e-01 1.18853009e+00 8.07736099e-01 3.16622227e-01
-1.83328316e-01 3.14917564e-01 3.49600077e-01 1.81218758e-01
4.98123348e-01 6.85398877e-01 -6.02684557e-01 7.69309998e-02
-8.97597015e-01 1.28886610e-01 1.00485635e+00 5.90498447e-01
7.06507683e-01 3.76394957e-01 2.78447986e-01 6.43422842e-01
1.45494908e-01 -1.12305537e-01 4.14941639e-01 -8.47031951e-01
1.51416853e-01 9.32675898e-01 -4.70748454e-01 -1.29370546e+00
-9.51359332e-01 -8.89021933e-01 -1.48520756e+00 1.93411976e-01
3.43122929e-01 -1.23839099e-02 -1.00773478e+00 1.60995221e+00
1.29789650e-01 -2.44783357e-01 -8.63506868e-02 4.46940392e-01
7.14078665e-01 7.72428751e-01 -3.07957470e-01 -3.45071964e-02
1.03634536e+00 -7.77022004e-01 -7.33560145e-01 3.06998212e-02
7.09336579e-01 -1.57792673e-01 6.10468745e-01 7.49765337e-01
-1.04100418e+00 -4.49225873e-01 -1.42908609e+00 -3.96730751e-01
-7.33487546e-01 8.76324773e-02 1.30574512e+00 4.73123252e-01
-1.00383401e+00 9.93252277e-01 -9.39680696e-01 5.98466545e-02
9.00425911e-01 6.80586815e-01 -2.41323844e-01 5.39123118e-02
-1.17111468e+00 8.67963850e-01 5.77189088e-01 2.74228632e-01
-5.52301466e-01 -9.33674395e-01 -6.71535373e-01 3.24405998e-01
3.40344965e-01 -7.22889781e-01 1.06318367e+00 -8.99111450e-01
-1.46745551e+00 4.03760910e-01 2.50445217e-01 -9.15793240e-01
3.56075257e-01 -1.23015881e-01 -2.27914885e-01 6.84291795e-02
-6.51031792e-01 3.96286845e-01 6.80459797e-01 -8.62652183e-01
-2.33777210e-01 -5.09014964e-01 -7.01998547e-02 -2.80335635e-01
-5.79884410e-01 -4.05521601e-01 -2.44228303e-01 -7.36218750e-01
2.51376480e-01 -8.78196657e-01 -5.54481864e-01 2.77831256e-01
-2.50365108e-01 -3.54061931e-01 7.72021234e-01 -5.39699972e-01
1.33713257e+00 -1.80917060e+00 3.51362646e-01 4.34484839e-01
6.81826293e-01 2.60906488e-01 5.77824861e-02 6.46897018e-01
-3.89985532e-01 2.67901510e-01 -1.68997556e-01 -4.42645960e-02
-1.44405086e-02 2.34612599e-01 -2.87565351e-01 4.49945986e-01
3.48246455e-01 8.71168554e-01 -5.51579833e-01 -1.71921298e-01
9.38597843e-02 5.08210778e-01 -5.67423761e-01 9.14136097e-02
-3.15834820e-01 9.67065468e-02 -2.94673681e-01 4.56069946e-01
3.25008154e-01 -5.45478523e-01 4.27207291e-01 -5.43257356e-01
-7.85537660e-02 3.61468703e-01 -8.76432002e-01 1.74220431e+00
-3.87984663e-01 9.27214324e-01 -2.16431826e-01 -1.66699612e+00
8.74858439e-01 -1.33656859e-01 5.94040990e-01 -9.45213735e-01
2.90197879e-01 1.05699569e-01 2.50492930e-01 -3.11777383e-01
2.60363251e-01 8.46657753e-02 -1.31406948e-01 2.39942670e-01
1.83545560e-01 1.14215584e-02 2.14872167e-01 2.10516095e-01
1.01746523e+00 -1.21957503e-01 5.35012245e-01 -5.62984049e-01
2.57097393e-01 -9.06632990e-02 3.52892756e-01 9.17985022e-01
3.42722356e-01 2.47418448e-01 7.66596973e-01 -1.11701381e+00
-1.21402109e+00 -8.63181293e-01 -2.44982779e-01 1.03034902e+00
-3.25830162e-01 -6.61958694e-01 -6.43836558e-01 -1.71486184e-01
-6.83779493e-02 4.14116591e-01 -8.11292768e-01 -2.98534811e-01
-6.72467351e-01 -7.65496016e-01 6.67251706e-01 5.62493324e-01
4.55783010e-01 -6.19195700e-01 -6.13002777e-01 2.36203268e-01
4.33874339e-01 -8.75723600e-01 1.74227655e-01 7.27514267e-01
-1.13783586e+00 -8.80585849e-01 -3.00017744e-01 -4.93594378e-01
5.10252357e-01 -2.40566641e-01 1.17373061e+00 1.27239659e-01
-6.66224599e-01 -3.34572583e-01 -2.19815113e-02 -5.57051599e-01
-2.67739624e-01 4.10437733e-01 8.56879205e-02 -9.74012911e-02
1.82965733e-02 -8.98310304e-01 -4.97119159e-01 -1.56093791e-01
-1.16395044e+00 4.01117414e-01 7.53914058e-01 1.05367506e+00
3.13582182e-01 3.05472672e-01 5.33480704e-01 -7.38230884e-01
5.18305540e-01 -4.68789756e-01 -1.02613604e+00 3.96562479e-02
-7.59290338e-01 7.29523838e-01 9.49455321e-01 -3.95026356e-01
-5.62523782e-01 1.03341803e-01 -4.40856367e-02 -2.31666535e-01
2.38306895e-01 1.15080798e+00 1.07357502e-01 -2.74123579e-01
6.14726186e-01 3.76565009e-02 5.38408756e-02 -5.98027110e-01
3.46108198e-01 3.03045899e-01 5.50762534e-01 -4.46133912e-01
6.33496284e-01 3.86557519e-01 8.70238006e-01 -7.83763885e-01
-9.40436125e-01 2.34067202e-01 -6.84913218e-01 -1.66610003e-01
5.50788999e-01 -7.04423666e-01 -9.88599360e-01 5.13387322e-01
-1.14022088e+00 -9.84174460e-02 2.18390059e-02 1.04323737e-01
-2.45978355e-01 1.72853649e-01 -4.81245041e-01 -7.63417661e-01
-4.53299075e-01 -8.96581829e-01 6.14276946e-01 6.75827116e-02
-1.78378478e-01 -1.23702145e+00 -1.51308952e-03 -8.63322318e-02
4.27867085e-01 5.55361092e-01 1.50829411e+00 -8.30281615e-01
-7.78204262e-01 -2.35020041e-01 -2.53097683e-01 1.56135693e-01
-1.70159221e-01 8.69943723e-02 -7.83694386e-01 -2.26986572e-01
-1.54543862e-01 -2.64142483e-01 1.21199489e+00 4.23854828e-01
1.82513273e+00 -5.67632496e-01 -2.65498042e-01 9.38976884e-01
1.50664353e+00 3.94568324e-01 3.95301163e-01 1.79409534e-01
9.45230186e-01 5.46718478e-01 -2.82359928e-01 4.58138227e-01
1.21325806e-01 6.32497013e-01 5.64583480e-01 -1.50940701e-01
2.24467620e-01 -2.89368749e-01 -1.01173699e-01 9.60719228e-01
-1.81323618e-01 -9.29147452e-02 -1.10004091e+00 3.20938557e-01
-2.06966472e+00 -7.69846201e-01 -8.78565013e-02 2.03233576e+00
8.95267129e-01 3.50238949e-01 -5.42529598e-02 4.43004340e-01
1.57690674e-01 6.59003854e-02 -9.32036996e-01 -5.25758266e-01
-9.85876173e-02 6.37415946e-01 5.83922923e-01 9.07737538e-02
-1.07571661e+00 5.47511458e-01 7.11574125e+00 1.01724350e+00
-1.18067169e+00 -2.20741391e-01 8.90491128e-01 -6.59888536e-02
-2.68852621e-01 -3.42974991e-01 -5.69556415e-01 7.89881274e-02
1.28378189e+00 -1.93391725e-01 5.80984771e-01 8.26000273e-01
-4.33506630e-02 1.83526427e-01 -1.42897117e+00 1.23430955e+00
-1.36929527e-01 -2.07163167e+00 2.61623532e-01 2.07734071e-02
6.34868562e-01 7.67248869e-02 2.99901813e-01 1.45974368e-01
3.41049641e-01 -1.46768904e+00 7.15509355e-01 4.58698124e-01
5.84480226e-01 -1.00148261e+00 3.06684345e-01 1.69472530e-01
-9.40229297e-01 -4.43081617e-01 -2.96376228e-01 -5.11309564e-01
-3.50906998e-01 7.21701443e-01 -5.42165518e-01 6.32278264e-01
8.19167972e-01 8.97343755e-01 -5.21185040e-01 9.89343107e-01
1.49983540e-01 6.43420935e-01 -5.68500519e-01 -3.75346750e-01
3.92534047e-01 -2.90881753e-01 2.37659588e-01 8.80727172e-01
2.95702696e-01 4.55132797e-02 -1.40593067e-01 9.44039345e-01
-4.58241373e-01 -2.69113362e-01 -8.60562146e-01 -2.64951438e-01
3.29055160e-01 1.20636833e+00 -6.65446579e-01 -2.36855939e-01
-3.25940102e-01 4.49527383e-01 6.08417511e-01 2.29272217e-01
-5.81764281e-01 -4.37722504e-01 5.80877185e-01 -4.60508503e-02
4.99207556e-01 -5.61176121e-01 -8.02907765e-01 -1.01255631e+00
-9.43800211e-02 -8.44319522e-01 5.01468599e-01 -4.60428953e-01
-1.16952193e+00 5.28000295e-01 -6.55294359e-02 -8.62329423e-01
-4.91216838e-01 -1.05653536e+00 -3.56361598e-01 6.18204832e-01
-1.15961981e+00 -8.43854249e-01 -8.28429163e-02 4.14469928e-01
1.24431886e-01 -3.97906154e-01 9.48990285e-01 2.95508921e-01
-7.12351918e-01 6.32508636e-01 7.41778195e-01 5.12885265e-02
-6.30582869e-02 -1.12678516e+00 4.69030350e-01 6.38051629e-01
3.55477899e-01 5.19911528e-01 6.60958171e-01 -1.61240533e-01
-1.96769726e+00 -8.72747123e-01 4.50073183e-01 -1.76659524e-01
1.03676069e+00 -8.57781827e-01 -8.52655470e-01 3.79955560e-01
1.59292325e-01 -1.85742423e-01 7.13997185e-01 3.59194666e-01
-5.14164150e-01 -4.88160580e-01 -8.81574750e-01 7.75029063e-01
1.01710761e+00 -5.45162559e-01 -4.10320401e-01 2.59478658e-01
5.82252026e-01 -1.96051329e-01 -9.00506914e-01 5.07870555e-01
8.00347209e-01 -7.00246632e-01 9.18883622e-01 -9.22553301e-01
5.90054333e-01 5.79099096e-02 1.26388110e-02 -1.31118870e+00
-4.55001950e-01 -9.50676680e-01 -6.12653911e-01 5.50577462e-01
4.75172937e-01 -4.73549455e-01 9.20999110e-01 4.97626424e-01
-1.35647267e-01 -1.06470096e+00 -9.38647389e-01 -5.44933319e-01
1.44302696e-01 -3.24497312e-01 2.99849749e-01 8.69315803e-01
-4.07143421e-02 6.35338902e-01 -3.55548441e-01 -2.02824384e-01
5.89604735e-01 8.60429183e-02 3.13115239e-01 -1.89161086e+00
-1.89081728e-01 -7.35846639e-01 -4.84040469e-01 -8.92432392e-01
7.03271329e-02 -8.43881130e-01 -2.81731725e-01 -1.32357395e+00
7.77532253e-03 -1.91717729e-01 -5.23388565e-01 6.07265055e-01
3.07823032e-01 1.64473116e-01 -1.25929667e-02 -2.75652073e-02
-3.84778321e-01 5.36058307e-01 8.48532140e-01 -2.39249408e-01
-4.78263050e-02 -2.54304260e-01 -7.22542882e-01 6.55398190e-01
6.99823558e-01 -3.63220870e-01 -4.22977984e-01 -5.73076129e-01
8.84285390e-01 5.48868440e-02 2.78152853e-01 -1.14891672e+00
4.11653340e-01 2.31372893e-01 7.37205029e-01 -7.98877656e-01
3.05752099e-01 -1.04136515e+00 3.34859967e-01 6.55346394e-01
-5.83595991e-01 4.03795689e-01 5.77102721e-01 4.80115235e-01
-1.20092489e-01 -2.22797934e-02 7.70356178e-01 9.79055539e-02
-5.16389430e-01 4.56760168e-01 -5.59128582e-01 -4.04321909e-01
5.97152293e-01 8.93470421e-02 -3.15193236e-01 -3.70030135e-01
-5.66082478e-01 -1.11618862e-01 -9.21168644e-03 3.30980301e-01
5.38842440e-01 -1.38193357e+00 -4.43820268e-01 2.60103315e-01
-2.34729066e-01 1.33989453e-01 -1.48980254e-02 6.00864947e-01
-7.00652242e-01 7.57657766e-01 -5.16136229e-01 -5.09958506e-01
-9.67464268e-01 5.13294399e-01 2.86322832e-01 -5.05728781e-01
-5.09120941e-01 5.88592649e-01 1.05565958e-01 -1.47056684e-01
3.95524800e-01 -4.46328044e-01 -1.51960608e-02 1.78996861e-01
3.52439463e-01 3.84295881e-01 3.08723420e-01 -4.91512427e-03
-1.91255599e-01 2.00155079e-01 -2.63091087e-01 2.80210525e-01
1.81225049e+00 5.29387653e-01 -3.04617524e-01 4.32081908e-01
1.42869222e+00 -5.74379146e-01 -1.14945197e+00 -1.65157706e-01
2.30994374e-01 1.36377022e-01 4.78086382e-01 -7.13462770e-01
-1.20373058e+00 9.79354501e-01 4.97268289e-01 6.50000334e-01
1.23584759e+00 -1.99895203e-01 4.09014165e-01 1.14454985e+00
8.90319869e-02 -9.15850818e-01 1.43337473e-01 6.95320070e-01
7.94013500e-01 -8.52837324e-01 2.04343826e-01 -3.95810045e-02
1.53233498e-01 1.49185729e+00 1.85898945e-01 -2.48821348e-01
9.08792734e-01 3.33806783e-01 -5.31804800e-01 -3.48498791e-01
-9.48659420e-01 7.41869137e-02 5.40715218e-01 2.72285342e-01
3.85835975e-01 -2.19868258e-01 -1.84562787e-01 6.84319377e-01
-2.21048042e-01 -4.72124815e-02 4.12624478e-01 6.93135977e-01
-1.61025658e-01 -8.94747853e-01 6.74660578e-02 7.32877195e-01
-5.99790633e-01 -4.07320201e-01 -5.72163403e-01 8.08601975e-01
-3.78450096e-01 3.95071030e-01 1.37814507e-01 -4.29927617e-01
2.55087726e-02 -8.85667950e-02 5.10166168e-01 -1.09995902e-01
-4.18583333e-01 -2.25851730e-01 1.79715842e-01 -4.90397215e-01
-2.98474729e-01 -4.03353542e-01 -1.04937303e+00 -5.38850009e-01
-1.15408055e-01 -7.60092959e-02 1.07930756e+00 9.27923501e-01
5.89849710e-01 6.64598107e-01 5.90242326e-01 -1.02804625e+00
-5.49307048e-01 -7.66698122e-01 -3.51295650e-01 3.82585786e-02
5.00721872e-01 -6.68947399e-01 -1.35303825e-01 -1.64517596e-01] | [6.601881980895996, 5.775589942932129] |
677246ba-c3eb-426f-a9ed-1d1a51189467 | bevstereo-accurate-depth-estimation-in-multi | 2304.04185 | null | https://arxiv.org/abs/2304.04185v1 | https://arxiv.org/pdf/2304.04185v1.pdf | BEVStereo++: Accurate Depth Estimation in Multi-view 3D Object Detection via Dynamic Temporal Stereo | Bounded by the inherent ambiguity of depth perception, contemporary multi-view 3D object detection methods fall into the performance bottleneck. Intuitively, leveraging temporal multi-view stereo (MVS) technology is the natural knowledge for tackling this ambiguity. However, traditional attempts of MVS has two limitations when applying to 3D object detection scenes: 1) The affinity measurement among all views suffers expensive computational cost; 2) It is difficult to deal with outdoor scenarios where objects are often mobile. To this end, we propose BEVStereo++: by introducing a dynamic temporal stereo strategy, BEVStereo++ is able to cut down the harm that is brought by introducing temporal stereo when dealing with those two scenarios. Going one step further, we apply Motion Compensation Module and long sequence Frame Fusion to BEVStereo++, which shows further performance boosting and error reduction. Without bells and whistles, BEVStereo++ achieves state-of-the-art(SOTA) on both Waymo and nuScenes dataset. | ['Li Xiao', 'Zheng Ge', 'Han Bao', 'Jianjian Sun', 'Jinrong Yang', 'Yinhao Li'] | 2023-04-09 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 1.87103704e-01 -1.82078317e-01 -2.01235134e-02 -7.17326859e-03
-4.14440483e-01 -5.91047704e-01 7.56807983e-01 -1.72233149e-01
-3.40796381e-01 3.02831203e-01 5.54822646e-02 -9.62713733e-02
2.45546624e-02 -7.97416449e-01 -5.68744302e-01 -5.83456039e-01
1.99993625e-01 2.25473985e-01 1.01569796e+00 -4.22989368e-01
3.90924901e-01 4.80019838e-01 -1.97955108e+00 3.94158542e-01
6.01554215e-01 9.74076748e-01 6.38783514e-01 7.67554283e-01
-1.85343511e-02 7.94447303e-01 3.51502299e-02 -4.56460267e-01
6.15108609e-01 -7.18962476e-02 -6.84101164e-01 2.25759700e-01
7.57145107e-01 -6.80672288e-01 -3.06213170e-01 9.56936300e-01
7.33702183e-01 1.41503266e-03 4.53532994e-01 -1.19700754e+00
4.75569582e-03 -1.34397447e-02 -9.83746111e-01 5.21820486e-01
6.94858730e-01 3.26288611e-01 8.59099329e-01 -1.22347975e+00
9.20126796e-01 1.39681733e+00 4.06419128e-01 5.39087713e-01
-1.09603250e+00 -2.49660492e-01 4.35060561e-01 3.63759130e-01
-1.11138761e+00 -4.43791002e-01 8.04147005e-01 -4.41423506e-01
1.01929343e+00 4.35234636e-01 7.95025706e-01 1.24459386e+00
5.01424037e-02 1.11968684e+00 1.27960765e+00 -2.37453729e-01
7.02945888e-02 6.81742132e-02 3.22798342e-02 4.96893197e-01
8.17429274e-02 4.66693133e-01 -7.18594730e-01 2.12812439e-01
5.68953335e-01 1.10040240e-01 -3.16901803e-01 -7.96293080e-01
-1.22227013e+00 4.61062640e-01 3.07240218e-01 1.52994260e-01
-1.13728479e-01 -7.58415461e-02 4.32959229e-01 2.12585464e-01
2.66992629e-01 -4.03554104e-02 -2.14795768e-01 -2.11022526e-01
-8.03229272e-01 3.43398094e-01 2.99206883e-01 1.10448039e+00
6.02932453e-01 -9.66868922e-02 1.18023723e-01 6.50774062e-01
2.58873701e-01 4.45901364e-01 1.70379564e-01 -1.05004942e+00
7.30474710e-01 7.41867661e-01 1.52204767e-01 -9.17169988e-01
-3.81621391e-01 -2.30370775e-01 -6.71810031e-01 6.52959108e-01
4.17484492e-01 2.08322972e-01 -8.32082212e-01 1.30399370e+00
7.73380458e-01 -6.71883598e-02 -5.34871854e-02 1.18741369e+00
8.45695734e-01 3.88219446e-01 -4.54510599e-01 -2.33419672e-01
1.27615523e+00 -1.16227734e+00 -4.11759347e-01 -5.62808037e-01
5.14134347e-01 -8.45680714e-01 9.11304533e-01 5.08638978e-01
-1.22313273e+00 -5.38270116e-01 -1.06064439e+00 -1.98060691e-01
-2.80073553e-01 -3.73216301e-01 7.88038492e-01 7.66957939e-01
-9.74346042e-01 2.37505868e-01 -7.89632738e-01 -6.00300431e-01
3.20811838e-01 2.57070303e-01 -5.22242188e-01 -3.48079652e-01
-7.79672682e-01 8.17247748e-01 3.43172163e-01 7.44656846e-02
-9.16788280e-01 -6.81736946e-01 -7.56747007e-01 -1.98232949e-01
7.37004578e-01 -1.16167510e+00 9.44617748e-01 -5.66862106e-01
-1.24401641e+00 1.10444188e+00 -1.67762876e-01 -3.81928653e-01
1.11199093e+00 -4.51046228e-01 -2.13351354e-01 1.18125342e-01
1.24086507e-01 6.99716985e-01 7.60951042e-01 -1.52529967e+00
-9.80203152e-01 -7.13135600e-01 3.77380073e-01 6.67400062e-01
-1.24747321e-01 -1.56534225e-01 -7.15963602e-01 -4.04127181e-01
5.26777327e-01 -8.71526718e-01 -2.15737596e-01 1.44383222e-01
-3.92402709e-01 7.70440046e-03 1.03288841e+00 -2.43495312e-02
1.08015132e+00 -2.28246212e+00 3.81278843e-01 -3.07150126e-01
3.52617115e-01 3.31533343e-01 3.69159691e-02 3.75748426e-01
2.31947854e-01 -2.42374167e-01 1.06690921e-01 -3.77891660e-01
-1.48444027e-01 6.81452006e-02 -1.66575909e-01 4.48717564e-01
-2.34373853e-01 6.64901435e-01 -9.05495048e-01 -5.52121580e-01
5.87807000e-01 1.63321555e-01 -7.21692383e-01 1.89578265e-01
-1.42888680e-01 4.03623253e-01 -2.55204350e-01 8.45925748e-01
9.17653024e-01 -9.84877795e-02 -1.31306767e-01 -2.31857195e-01
-4.16994065e-01 1.25944251e-02 -1.53916430e+00 1.90381145e+00
-2.90002197e-01 4.14782941e-01 1.51633546e-01 -5.57939649e-01
5.55965364e-01 -9.88644958e-02 5.06005645e-01 -6.64338648e-01
1.88625231e-01 2.09708646e-01 -1.82869032e-01 -6.29057169e-01
6.06007397e-01 -2.91023050e-02 2.68760711e-01 -2.79156696e-02
-1.36489898e-01 -2.83964217e-01 1.17290266e-01 2.61217564e-01
9.54937816e-01 3.93484712e-01 4.10980195e-01 -7.00690225e-02
6.50007844e-01 2.06848122e-02 4.48348224e-01 7.01606095e-01
-4.32704031e-01 8.18104982e-01 6.71065897e-02 -5.12897253e-01
-9.07021761e-01 -1.33752155e+00 -7.55328611e-02 9.14688170e-01
5.79139233e-01 -2.14064404e-01 -3.63019437e-01 -6.78045809e-01
-6.33240193e-02 3.14955682e-01 -4.59553510e-01 1.90744400e-01
-3.30544293e-01 -5.46777904e-01 2.29340032e-01 4.63999957e-01
8.31066728e-01 -3.70073915e-01 -1.21991920e+00 1.02590881e-01
-2.45032266e-01 -1.58478951e+00 -3.97047013e-01 2.65360661e-02
-1.03596365e+00 -1.02378213e+00 -8.23733985e-01 -3.13446105e-01
8.21552575e-02 1.10151911e+00 1.03906929e+00 -2.29032218e-01
-3.52806240e-01 5.36983907e-01 -4.40963984e-01 -4.24776644e-01
-9.42293182e-02 -1.78627193e-01 8.03914294e-03 3.34706046e-02
3.32766622e-01 -8.07557881e-01 -1.06850052e+00 5.43893933e-01
-8.22229981e-01 3.60216916e-01 3.69664997e-01 6.57970130e-01
4.59985763e-01 -2.52187610e-01 -3.27514164e-04 -6.04667664e-01
-2.83960193e-01 -1.30000412e-01 -6.64324462e-01 4.01722789e-02
-5.19520462e-01 -4.26618844e-01 1.32453591e-01 -1.61149919e-01
-1.17777419e+00 2.55835921e-01 -1.54277921e-01 -6.34190798e-01
-2.20238104e-01 -2.43789658e-01 -1.91949606e-01 -1.21400557e-01
6.43842161e-01 3.45259339e-01 -2.82912910e-01 -3.38788778e-01
2.81202555e-01 5.07120788e-01 4.74391431e-01 9.88958590e-03
7.47462928e-01 1.25844824e+00 2.55717546e-01 -8.74238193e-01
-8.83508861e-01 -7.01084197e-01 -8.45882237e-01 -6.03926539e-01
1.06544125e+00 -1.18833327e+00 -8.28872383e-01 5.48886418e-01
-1.11513031e+00 6.52893931e-02 -1.40473675e-02 4.30908591e-01
-6.41043842e-01 7.74745643e-01 -2.86363780e-01 -1.07147646e+00
-6.30108342e-02 -1.17103279e+00 1.32762849e+00 -5.58778383e-02
8.73137489e-02 -7.32841551e-01 -1.92896858e-01 7.99632788e-01
2.06855133e-01 3.44440103e-01 4.99359488e-01 -7.60165974e-02
-1.07249320e+00 7.81671256e-02 -3.16440344e-01 7.68643767e-02
-1.77603096e-01 -3.65144789e-01 -1.37217057e+00 -3.32915545e-01
3.32821071e-01 -9.72841457e-02 8.58090818e-01 2.46328413e-01
6.71361327e-01 1.92906320e-01 -2.98609555e-01 8.12944233e-01
1.52959609e+00 2.01205745e-01 5.63257992e-01 7.22086310e-01
8.14017653e-01 7.02195942e-01 9.53484952e-01 5.89080811e-01
6.27483785e-01 1.11096919e+00 9.29200888e-01 5.23154438e-03
-2.34842598e-01 -1.62976071e-01 4.86545533e-01 5.57351291e-01
-2.80544221e-01 -3.28009874e-01 -9.33101833e-01 5.24220109e-01
-1.84775782e+00 -1.06605256e+00 -5.31680167e-01 2.06949496e+00
1.79804549e-01 3.56648237e-01 2.31443733e-01 3.21239293e-01
4.82405961e-01 4.52280909e-01 -4.52756733e-01 -1.11501394e-02
-3.99213761e-01 -5.43122292e-01 5.23784518e-01 2.89951205e-01
-1.01592433e+00 7.26547480e-01 5.43782043e+00 8.37434649e-01
-1.10678339e+00 1.19347714e-01 3.20382863e-01 -4.00049508e-01
-1.13790371e-01 2.99364384e-02 -9.98878062e-01 3.05272788e-01
1.84744731e-01 2.72634298e-01 2.55089313e-01 7.68821061e-01
1.93181276e-01 -6.10245705e-01 -1.20058894e+00 1.41812718e+00
1.86508611e-01 -1.11726439e+00 3.67935710e-02 6.61609620e-02
6.24496162e-01 2.62691170e-01 1.27386749e-01 -1.57526389e-01
3.05731334e-02 -6.72465801e-01 9.28665578e-01 8.01638514e-02
4.12942469e-01 -5.02654910e-01 5.70319295e-01 7.46610284e-01
-1.17303753e+00 -1.71203673e-01 -2.16212660e-01 -1.64929971e-01
4.51261073e-01 6.09711051e-01 -6.46500409e-01 9.27707911e-01
9.59754229e-01 7.76261747e-01 -5.68379104e-01 9.67240036e-01
1.29207909e-01 -1.19265832e-01 -3.38185757e-01 2.56273836e-01
2.06917182e-01 -9.20011252e-02 9.54125285e-01 1.02803338e+00
3.80458623e-01 -5.89072406e-02 1.96467683e-01 4.06956404e-01
4.26346481e-01 -2.35266671e-01 -1.01229179e+00 4.60245609e-01
1.40183508e-01 9.63675737e-01 -8.15572321e-01 -2.70864695e-01
-6.57391429e-01 1.43918419e+00 6.79190531e-02 1.46716133e-01
-7.62582004e-01 1.25608191e-01 4.98503000e-01 3.96095127e-01
4.82221931e-01 -2.49312833e-01 -2.99345285e-01 -1.45502627e+00
1.76500648e-01 -7.28579521e-01 4.03910071e-01 -9.75006700e-01
-1.06476688e+00 4.97998804e-01 3.53781022e-02 -1.67956245e+00
5.77479824e-02 -4.95507181e-01 -6.74021021e-02 4.08413172e-01
-1.70802832e+00 -1.32736278e+00 -5.48066378e-01 7.34992325e-01
1.01244104e+00 1.53496608e-01 2.97560811e-01 5.34910798e-01
-1.81789637e-01 1.97622567e-01 -1.16638646e-01 -4.21026081e-01
6.67673707e-01 -1.11404049e+00 3.89072448e-01 1.00556087e+00
6.80705458e-02 2.09333405e-01 8.16559494e-01 -2.81410426e-01
-1.78084898e+00 -7.76236534e-01 6.81645393e-01 -7.99132526e-01
4.15070742e-01 -5.62849045e-01 -6.14906788e-01 4.65738773e-01
6.20283484e-02 2.36293003e-01 2.88060576e-01 -1.99661896e-01
-3.85667592e-01 -4.19367366e-02 -1.03252244e+00 6.46892667e-01
1.76602125e+00 -4.88028914e-01 -5.07847667e-01 1.16185158e-01
5.21324217e-01 -8.41595352e-01 -7.18243301e-01 5.95288038e-01
6.98278368e-01 -1.92896450e+00 1.33497548e+00 -1.86388105e-01
4.64311689e-01 -5.61630964e-01 -5.93429565e-01 -8.94739568e-01
1.90199148e-02 -5.13628542e-01 -1.80629939e-01 1.17576206e+00
-6.89517260e-02 -4.21208113e-01 8.32680583e-01 1.28797919e-01
-1.81030989e-01 -6.50820017e-01 -1.17719698e+00 -9.61879671e-01
-3.57199699e-01 -6.66284978e-01 3.17561060e-01 7.60464847e-01
-3.36909264e-01 4.15378183e-01 -5.35814166e-01 2.92655438e-01
8.28835607e-01 2.99927413e-01 1.16553450e+00 -1.11176634e+00
-5.43338895e-01 -2.69274473e-01 -6.95811391e-01 -1.50136447e+00
-5.84616303e-01 -3.96576911e-01 -1.26207009e-01 -1.33113003e+00
2.35770121e-01 -1.01683520e-01 8.26929063e-02 -2.38130420e-01
-1.31412894e-01 3.12790334e-01 6.86875105e-01 1.57119289e-01
-6.61439121e-01 4.36925858e-01 1.55205834e+00 4.07380015e-02
-1.12970702e-01 -9.44503695e-02 -2.82853842e-01 9.90721166e-01
1.97944015e-01 -2.10345864e-01 -5.13597071e-01 -5.57171106e-01
2.67081052e-01 4.86438334e-01 6.43964708e-01 -1.07193482e+00
3.15421999e-01 -1.53600872e-01 1.66689798e-01 -1.06996572e+00
7.70937681e-01 -1.03741908e+00 1.67510644e-01 5.73019862e-01
3.24325979e-01 1.99317813e-01 1.17018580e-01 8.61396015e-01
-2.86490619e-01 3.02704740e-02 1.01267350e+00 -3.75782728e-01
-1.21375656e+00 1.96631029e-01 -2.51697838e-01 4.40202169e-02
1.27267385e+00 -7.97882199e-01 -3.46426487e-01 -1.29511699e-01
-6.76031947e-01 4.26364303e-01 7.57883668e-01 5.68913102e-01
6.74210072e-01 -1.14403498e+00 -2.57342905e-01 2.79622465e-01
3.32786620e-01 2.21272677e-01 7.76503384e-01 1.16141379e+00
-3.74548227e-01 3.93282175e-01 -2.50823945e-01 -1.02398157e+00
-1.61682987e+00 8.14142227e-01 2.20258474e-01 -1.40681863e-01
-8.22203040e-01 9.31722879e-01 6.73371375e-01 -1.51428863e-01
1.55502290e-01 -1.81242406e-01 -7.30358213e-02 2.57014930e-01
3.41255248e-01 6.46752119e-01 1.57621011e-01 -5.79800487e-01
-5.50448060e-01 8.17523062e-01 -8.71414095e-02 -2.03423738e-01
1.22070730e+00 -7.40621150e-01 2.14586735e-01 4.60350871e-01
9.55521941e-01 1.15980648e-01 -1.47032559e+00 1.60952508e-02
-8.85422528e-02 -9.55165029e-01 2.99892630e-02 -4.09752101e-01
-7.60922313e-01 1.09493017e+00 9.06967461e-01 1.35645613e-01
1.21651196e+00 -2.76911333e-02 7.43504584e-01 1.43153638e-01
6.96661413e-01 -8.45664918e-01 2.69892544e-01 5.53166270e-01
5.89746714e-01 -1.54065061e+00 4.39455323e-02 -7.49856830e-01
-6.04282200e-01 9.79066670e-01 8.00233483e-01 9.78257582e-02
3.90550286e-01 9.22384113e-02 -3.84504558e-04 -3.66252691e-01
-7.21974313e-01 -5.07156312e-01 2.98341550e-02 6.62777901e-01
2.51905322e-02 -4.25379366e-01 -1.26324967e-01 -3.23503539e-02
1.01570323e-01 -1.57390490e-01 4.93479997e-01 1.04398715e+00
-4.53210056e-01 -8.72011960e-01 -4.76706237e-01 -1.76311154e-02
-4.29806590e-01 8.77851434e-03 -2.62616694e-01 9.33621585e-01
3.47676933e-01 8.43066990e-01 -2.57256806e-01 -4.92179513e-01
6.32354021e-01 -1.97504744e-01 5.93986452e-01 -4.67626899e-01
-4.21904206e-01 1.91150650e-01 -1.12947831e-02 -8.61232281e-01
-8.03753018e-01 -7.48749197e-01 -7.81341970e-01 -2.05986634e-01
-3.51332128e-01 -5.40388584e-01 4.26173329e-01 8.29529047e-01
3.30982447e-01 4.09963816e-01 6.45092249e-01 -1.09506929e+00
-5.01727879e-01 -5.34750700e-01 -4.50635493e-01 4.76932645e-01
6.10983253e-01 -8.47904503e-01 -3.01013291e-01 -2.13562220e-01] | [8.08616828918457, -2.229245185852051] |
13051aa7-e796-42ef-980a-47936e05f220 | confidence-aware-levenberg-marquardt | 1609.01524 | null | http://arxiv.org/abs/1609.01524v1 | http://arxiv.org/pdf/1609.01524v1.pdf | Confidence-aware Levenberg-Marquardt optimization for joint motion estimation and super-resolution | Motion estimation across low-resolution frames and the reconstruction of
high-resolution images are two coupled subproblems of multi-frame
super-resolution. This paper introduces a new joint optimization approach for
motion estimation and image reconstruction to address this interdependence. Our
method is formulated via non-linear least squares optimization and combines two
principles of robust super-resolution. First, to enhance the robustness of the
joint estimation, we propose a confidence-aware energy minimization framework
augmented with sparse regularization. Second, we develop a tailor-made
Levenberg-Marquardt iteration scheme to jointly estimate motion parameters and
the high-resolution image along with the corresponding model confidence
parameters. Our experiments on simulated and real images confirm that the
proposed approach outperforms decoupled motion estimation and image
reconstruction as well as related state-of-the-art joint estimation algorithms. | ['Thomas Köhler', 'Cosmin Bercea', 'Andreas Maier'] | 2016-09-06 | null | null | null | null | ['multi-frame-super-resolution'] | ['computer-vision'] | [ 1.63496569e-01 -4.00608063e-01 -1.22167729e-01 -1.51083902e-01
-1.29012275e+00 -1.03098422e-01 3.37044626e-01 -5.94273269e-01
-5.03525019e-01 8.88157308e-01 3.26935589e-01 4.27355260e-01
-1.29910991e-01 -2.70083934e-01 -5.12355685e-01 -9.05915618e-01
1.42111450e-01 -1.13730639e-01 4.56071615e-01 6.76160231e-02
3.31320167e-01 2.33277053e-01 -1.26023996e+00 2.03879201e-03
7.40558207e-01 6.23201728e-01 6.40365124e-01 7.87228882e-01
3.66387874e-01 8.44624996e-01 3.58809121e-02 1.34018809e-01
1.86203420e-02 -5.11186123e-01 -6.40221834e-01 3.84066790e-01
5.24570525e-01 -5.34090519e-01 -1.27479583e-01 1.16722548e+00
4.96680766e-01 2.97489911e-01 4.44863737e-01 -4.30600464e-01
-4.43518221e-01 -2.12951023e-02 -1.15690351e+00 7.05230117e-01
5.24634600e-01 -1.25526741e-01 5.31729758e-01 -1.33522236e+00
7.80280530e-01 1.21416855e+00 5.60426831e-01 2.64573008e-01
-1.37659860e+00 -2.64746785e-01 6.81350306e-02 7.96868727e-02
-1.67170620e+00 -7.08863676e-01 8.64737749e-01 -4.67821062e-01
5.00901997e-01 8.76845270e-02 2.36057654e-01 7.39896476e-01
3.16581219e-01 5.11514127e-01 1.10616517e+00 -4.13354725e-01
-3.29269916e-02 -2.76419043e-04 -1.93885520e-01 9.42211926e-01
2.80203909e-01 3.16473216e-01 -6.38928354e-01 -3.75284046e-01
1.71900260e+00 -4.81652468e-01 -5.10990620e-01 -3.95903170e-01
-1.52366757e+00 5.78594446e-01 -1.06428742e-01 4.74666536e-01
-7.40876138e-01 1.41548544e-01 -7.33028725e-02 -4.35713977e-01
6.17644787e-01 -2.57070437e-02 5.75050786e-02 1.12423815e-01
-1.18205023e+00 1.73989832e-01 2.06414104e-01 6.78401768e-01
6.73021138e-01 4.89137083e-01 7.10343421e-02 9.27132010e-01
6.48524523e-01 3.73343736e-01 3.00949544e-01 -1.44857526e+00
1.78330734e-01 -3.72712731e-01 6.73697531e-01 -1.17359817e+00
-1.61322787e-01 -4.33785319e-01 -9.82126057e-01 1.68997645e-01
2.52863497e-01 -9.71530974e-02 -4.12583947e-01 1.64530468e+00
6.86014354e-01 8.56304467e-01 -7.18841702e-02 1.38679087e+00
5.39083660e-01 7.71050453e-01 -1.53664827e-01 -9.67886209e-01
1.13120687e+00 -9.90096867e-01 -1.21819568e+00 -1.03736646e-01
-1.36197761e-01 -8.93471420e-01 4.40728158e-01 3.27134639e-01
-1.76525140e+00 -8.92801464e-01 -1.11360574e+00 -9.44032073e-02
5.84362805e-01 7.96895176e-02 2.52797961e-01 2.58308232e-01
-8.83524239e-01 6.53822124e-01 -9.42687869e-01 1.63585082e-01
5.10851778e-02 1.28121033e-01 -2.90412962e-01 -4.04962301e-02
-8.75888586e-01 9.92923141e-01 3.13817024e-01 1.54308856e-01
-7.45216668e-01 -6.47207141e-01 -9.28268790e-01 -3.41502309e-01
2.69537508e-01 -9.82633114e-01 9.97992814e-01 -6.93028152e-01
-1.72869062e+00 6.91303551e-01 -6.54198170e-01 -2.37595495e-02
3.84858400e-01 -4.03084069e-01 -4.60815698e-01 4.14282203e-01
1.30313858e-01 3.18032146e-01 1.30861688e+00 -1.61791325e+00
-7.47162700e-01 -1.41092345e-01 -6.33520544e-01 4.22611296e-01
1.17713146e-01 2.14227706e-01 -7.36950517e-01 -8.67535591e-01
4.03215706e-01 -6.85606003e-01 -5.66669285e-01 -1.34297952e-01
6.21912517e-02 5.26656210e-01 4.91509408e-01 -9.92118537e-01
1.48962104e+00 -2.07462645e+00 6.70418978e-01 -5.20539396e-02
1.64620906e-01 -3.89549658e-02 -5.65051101e-02 -2.53147870e-01
3.43584195e-02 -2.99455673e-01 -4.19197857e-01 -3.44367176e-01
-5.73112130e-01 -2.44253082e-03 1.31152868e-01 9.16729689e-01
1.82719424e-01 5.28436124e-01 -9.81620133e-01 -7.55945623e-01
5.33301175e-01 1.03446555e+00 -2.96283394e-01 3.11934561e-01
3.46549243e-01 1.05572474e+00 -5.44856906e-01 4.32708323e-01
8.86437654e-01 -3.33141416e-01 2.36554623e-01 -7.24350631e-01
-5.57340622e-01 -1.81413814e-01 -1.70454574e+00 1.79758167e+00
-3.09765726e-01 3.18253487e-01 6.78487778e-01 -6.90874517e-01
5.56210518e-01 4.59174365e-01 9.10820305e-01 -4.79823023e-01
-8.15204382e-02 1.16497122e-01 -5.71290135e-01 -4.88205224e-01
7.50482380e-01 -3.63699913e-01 2.78018951e-01 1.37233093e-01
-1.09751254e-01 9.79779754e-04 7.98179433e-02 -8.79990086e-02
5.12878895e-01 6.28493369e-01 6.77731335e-01 -3.39826643e-01
9.78785515e-01 -2.99051136e-01 9.06417131e-01 4.79122490e-01
-2.12010026e-01 8.91682804e-01 -3.67383391e-01 -2.94259876e-01
-1.35511875e+00 -1.10829914e+00 -4.29785728e-01 7.28019893e-01
4.00724113e-01 -1.06426321e-01 -7.06371665e-01 -6.22674339e-02
-4.06211257e-01 2.98875153e-01 -2.90403157e-01 4.64669734e-01
-1.02233982e+00 -1.09235418e+00 -1.11133717e-01 3.35928112e-01
5.47617912e-01 -4.42842543e-01 -6.49220586e-01 5.37223816e-01
-7.64447212e-01 -1.62689614e+00 -7.15601623e-01 -4.28670138e-01
-1.05211616e+00 -7.25756884e-01 -1.11539364e+00 -7.74113297e-01
6.29933655e-01 5.80861568e-01 1.04489875e+00 -1.20108925e-01
-1.96091682e-01 4.87866223e-01 1.26399947e-02 5.06106019e-01
-3.82196128e-01 -5.87114573e-01 2.53155529e-01 4.28457022e-01
-3.58124495e-01 -6.85071528e-01 -7.28605092e-01 5.75663626e-01
-8.46670032e-01 2.16836900e-01 5.88378668e-01 7.98431933e-01
1.30857575e+00 1.74138665e-01 2.13986829e-01 -4.58495826e-01
2.79603332e-01 -3.66235107e-01 -8.81116867e-01 1.01242483e-01
-4.67908114e-01 1.86706409e-01 1.91266730e-01 -5.63693464e-01
-1.55126214e+00 3.74317765e-01 6.58134520e-02 -6.01760507e-01
5.72428061e-03 1.91351831e-01 -4.26472463e-02 -4.94083911e-01
4.23430830e-01 4.88569051e-01 -2.17530057e-01 -4.76787150e-01
4.18708235e-01 1.66849315e-01 1.14145362e+00 -3.55974555e-01
9.11307573e-01 8.89539719e-01 1.42001122e-01 -1.09755242e+00
-7.23667085e-01 -7.74406612e-01 -8.80382776e-01 -3.65348965e-01
1.26885128e+00 -1.23763525e+00 -3.43724579e-01 3.43833447e-01
-1.10368145e+00 2.84963977e-02 1.14175186e-01 9.35543358e-01
-9.28270936e-01 9.63769555e-01 -6.23377264e-01 -9.51548457e-01
-1.33919358e-01 -1.17285728e+00 1.06095994e+00 1.02559321e-01
-3.55849639e-02 -1.06737161e+00 3.15148264e-01 4.89166677e-01
3.16907614e-01 4.31849480e-01 2.95028090e-01 5.04563630e-01
-9.22388256e-01 2.16261804e-01 -3.07358921e-01 1.76775053e-01
7.71605968e-02 -8.14293921e-02 -6.40083134e-01 -4.09513444e-01
5.12843788e-01 9.87949222e-02 6.31189466e-01 9.73702788e-01
6.01527393e-01 -2.81143427e-01 -2.27280095e-01 9.11447167e-01
1.82148325e+00 -9.77352113e-02 7.76809335e-01 4.20074373e-01
8.29702914e-01 3.88749361e-01 9.44780648e-01 7.25145578e-01
4.00657922e-01 1.10082221e+00 1.75610363e-01 3.16757634e-02
-1.31507546e-01 1.44179359e-01 3.83670717e-01 1.00037682e+00
-5.10084271e-01 2.94471949e-01 -4.41363931e-01 5.32150149e-01
-2.00168800e+00 -1.15798664e+00 -3.83601904e-01 2.34335542e+00
7.07926393e-01 -2.84914553e-01 7.90477768e-02 -2.09612697e-01
8.46704483e-01 3.82633984e-01 -4.28511053e-01 2.39603013e-01
-2.63739526e-01 -1.58463597e-01 4.25787449e-01 1.01714945e+00
-1.18404400e+00 7.34517515e-01 6.97450590e+00 7.88140237e-01
-6.32501543e-01 4.11534369e-01 3.60695779e-01 -7.92656839e-02
-1.90608978e-01 -1.29083425e-01 -8.14270139e-01 3.84680748e-01
9.13421452e-01 -7.62662292e-02 4.58218873e-01 3.73200059e-01
5.72893083e-01 -3.20574462e-01 -5.92061520e-01 1.19759202e+00
7.85341710e-02 -1.46461761e+00 -1.21208079e-01 9.26757976e-02
9.32727575e-01 -2.05977842e-01 1.21226199e-02 -4.12970543e-01
5.48710749e-02 -7.32093513e-01 6.67074263e-01 8.95605087e-01
6.96928203e-01 -7.80998588e-01 3.71878773e-01 2.89511472e-01
-1.51264369e+00 1.43731296e-01 -3.95628870e-01 3.23889434e-01
9.65083301e-01 6.71369553e-01 8.64920095e-02 7.97256112e-01
6.56518936e-01 7.53154635e-01 1.74813457e-02 7.85320580e-01
-9.03459787e-02 3.02829109e-02 -8.53395835e-02 8.43530953e-01
1.24567458e-02 -5.27874947e-01 9.76781428e-01 1.28778481e+00
3.23638827e-01 7.31039584e-01 1.95013002e-01 8.57345462e-01
4.00280684e-01 -8.43766928e-02 -1.42199278e-01 4.35880125e-01
3.01462740e-01 1.36501157e+00 -3.79270852e-01 -3.05731058e-01
-7.22359180e-01 1.16192329e+00 1.49121314e-01 5.57576120e-01
-8.91194999e-01 3.64324272e-01 4.64820445e-01 1.85367361e-01
5.32958806e-01 -5.69678962e-01 -6.90192729e-02 -1.55671465e+00
-1.20515659e-01 -6.80243552e-01 3.15080047e-01 -8.70354831e-01
-8.80058587e-01 4.77276713e-01 1.95834607e-01 -1.33416331e+00
-4.40212846e-01 -6.74967021e-02 -2.98907310e-01 9.60107625e-01
-1.90091610e+00 -8.88467252e-01 -2.50948399e-01 7.95140445e-01
6.99256599e-01 2.10802704e-01 4.13598776e-01 3.16765785e-01
-5.96252322e-01 1.51116371e-01 3.26107591e-01 -2.19789252e-01
5.78408182e-01 -7.58004010e-01 2.70088147e-02 1.20443130e+00
-1.61406338e-01 3.31926972e-01 9.25849199e-01 -6.76421642e-01
-1.46707857e+00 -8.08125496e-01 4.05504167e-01 -1.79904208e-01
3.81252319e-01 3.39599162e-01 -1.34407258e+00 4.78636920e-01
1.97729751e-01 3.15687269e-01 3.99848431e-01 -6.62981629e-01
1.13174163e-01 2.44067550e-01 -1.26488054e+00 3.54836524e-01
7.88151205e-01 -3.82017285e-01 -4.24438834e-01 6.13349676e-02
3.20994079e-01 -5.03183544e-01 -1.30777991e+00 5.38345039e-01
5.80926657e-01 -1.09728682e+00 1.52755356e+00 -2.41145622e-02
2.54975230e-01 -5.95267117e-01 -4.32859570e-01 -9.25157487e-01
-9.72679019e-01 -7.94759929e-01 -6.98419571e-01 9.43809986e-01
1.67787857e-02 -1.65279508e-01 5.89585781e-01 4.51726705e-01
2.29353160e-01 -4.80661362e-01 -9.36145484e-01 -6.00735664e-01
-2.94653088e-01 -7.35450536e-02 -2.15525344e-01 1.00632858e+00
-2.83616006e-01 2.67896503e-01 -9.19994652e-01 6.95580542e-01
1.64829350e+00 -1.51582365e-03 4.26686585e-01 -7.75359213e-01
-4.90368217e-01 -1.54491201e-01 -2.13068992e-01 -1.11587286e+00
1.06388494e-01 -2.34327152e-01 2.17729405e-01 -1.42623806e+00
3.47067088e-01 1.91690385e-01 -2.14968398e-01 -3.78541768e-01
-4.15756315e-01 3.25011879e-01 -7.55610541e-02 5.58183789e-01
-5.84949434e-01 4.69508529e-01 1.36401498e+00 3.65622550e-01
-2.32042134e-01 -1.79517895e-01 -2.58741379e-01 1.00021601e+00
2.89808810e-01 -3.50311697e-01 -2.37635851e-01 -5.03692985e-01
-1.08759180e-01 7.43509471e-01 3.39892745e-01 -8.63134921e-01
2.21669719e-01 -4.17514384e-01 5.65170586e-01 -6.06505334e-01
5.74298382e-01 -6.91005647e-01 3.78109604e-01 1.37519464e-01
-1.29948676e-01 -3.22558321e-02 -4.45558369e-04 7.46501088e-01
-3.15744638e-01 -1.14546292e-01 1.47244942e+00 -3.33317578e-01
-8.61019075e-01 2.79686302e-01 -5.18059611e-01 -6.44240230e-02
7.42062271e-01 -3.54892373e-01 2.16939170e-02 -4.94336039e-01
-8.65815103e-01 -2.34038718e-02 4.97918189e-01 9.62783843e-02
1.06736743e+00 -1.44315779e+00 -1.06994760e+00 4.14211787e-02
-3.25148374e-01 -1.66940153e-01 4.95914876e-01 1.22915483e+00
-2.54374802e-01 2.64911771e-01 -1.83108687e-01 -7.09169805e-01
-1.26128912e+00 6.29860222e-01 3.83925140e-01 -4.34120685e-01
-7.12431967e-01 5.75021803e-01 3.50237370e-01 1.79069772e-01
-1.94565758e-01 6.75565302e-02 -3.34562212e-01 -2.43144810e-01
9.76179123e-01 7.53444791e-01 -5.02784431e-01 -1.29229259e+00
-3.50877136e-01 1.24639201e+00 -7.33536994e-03 -5.23719609e-01
1.34193349e+00 -9.50478971e-01 -1.02965847e-01 2.95744181e-01
1.21835792e+00 -2.13598553e-02 -1.75621974e+00 -3.53253365e-01
-5.95485866e-02 -8.91248941e-01 6.10950351e-01 -2.05309123e-01
-9.70331848e-01 4.00965869e-01 5.98616660e-01 -2.63047367e-01
1.16127110e+00 -2.55127370e-01 7.23520517e-01 -2.61115968e-01
4.12112594e-01 -1.09409630e+00 1.93281174e-01 1.70799389e-01
8.30659807e-01 -1.28972244e+00 5.43397248e-01 -6.31620705e-01
-5.08294463e-01 8.79462361e-01 4.69311804e-01 -1.63321763e-01
4.28943545e-01 3.68519425e-01 -1.56066671e-01 2.13335589e-01
-6.73506498e-01 -8.97712260e-02 5.79809964e-01 5.64762533e-01
5.68479121e-01 -4.79691118e-01 -3.77641231e-01 1.67509601e-01
5.74986398e-01 2.94395834e-01 6.11513913e-01 7.44266152e-01
-6.07879043e-01 -7.68565714e-01 -8.49799812e-01 -2.51437783e-01
-7.65796781e-01 3.24872951e-03 4.64011371e-01 4.02187020e-01
-2.79856473e-01 8.58084083e-01 -1.35529533e-01 1.83600914e-02
1.70524180e-01 -3.62195104e-01 8.83674502e-01 -1.26297772e-01
3.08933947e-02 8.39676797e-01 6.36367276e-02 -9.76055861e-01
-9.98991072e-01 -8.14174175e-01 -1.16740119e+00 -4.30062041e-02
-4.51385349e-01 5.48937246e-02 5.88296473e-01 7.41557658e-01
2.07204089e-01 5.13349116e-01 7.46188045e-01 -1.46854115e+00
-2.73965806e-01 -4.98876274e-01 -6.50021195e-01 1.84912965e-01
7.40047216e-01 -5.60492575e-01 -4.02212620e-01 2.65413135e-01] | [11.109798431396484, -2.175032377243042] |
52964099-8913-46b9-901b-a59f07399be9 | speaker-verification-using-attentive-multi | 2306.00426 | null | https://arxiv.org/abs/2306.00426v1 | https://arxiv.org/pdf/2306.00426v1.pdf | Speaker verification using attentive multi-scale convolutional recurrent network | In this paper, we propose a speaker verification method by an Attentive Multi-scale Convolutional Recurrent Network (AMCRN). The proposed AMCRN can acquire both local spatial information and global sequential information from the input speech recordings. In the proposed method, logarithm Mel spectrum is extracted from each speech recording and then fed to the proposed AMCRN for learning speaker embedding. Afterwards, the learned speaker embedding is fed to the back-end classifier (such as cosine similarity metric) for scoring in the testing stage. The proposed method is compared with state-of-the-art methods for speaker verification. Experimental data are three public datasets that are selected from two large-scale speech corpora (VoxCeleb1 and VoxCeleb2). Experimental results show that our method exceeds baseline methods in terms of equal error rate and minimal detection cost function, and has advantages over most of baseline methods in terms of computational complexity and memory requirement. In addition, our method generalizes well across truncated speech segments with different durations, and the speaker embedding learned by the proposed AMCRN has stronger generalization ability across two back-end classifiers. | ['Qisheng Huang', 'Wenchang Cao', 'Zhongjie Jiang', 'Yanxiong Li'] | 2023-06-01 | null | null | null | null | ['speaker-verification'] | ['speech'] | [ 1.01313576e-01 -3.52129221e-01 -4.72691432e-02 -4.86625254e-01
-1.04324305e+00 -3.22805315e-01 3.08137119e-01 -1.41281083e-01
-4.94291484e-01 3.06306452e-01 3.51384491e-01 -3.10826987e-01
1.11992501e-01 -3.02256376e-01 -3.69369835e-01 -6.96799576e-01
-4.83034812e-02 -8.98651928e-02 1.95536733e-01 -1.82981506e-01
1.29025131e-01 4.70045567e-01 -1.41873372e+00 1.69334680e-01
5.50331950e-01 1.17479610e+00 3.29146385e-01 6.97562516e-01
3.62635136e-01 6.43956542e-01 -7.56816447e-01 -1.42577454e-01
-8.99665952e-02 -4.41449016e-01 -6.03005946e-01 -2.70979255e-01
3.03976953e-01 -8.67774040e-02 -5.97704172e-01 9.92600739e-01
9.74718511e-01 5.92263520e-01 3.75246733e-01 -1.00302732e+00
-9.07648087e-01 8.68030310e-01 -2.77791589e-01 6.67881012e-01
2.69016057e-01 -3.98765616e-02 1.17539287e+00 -1.22430885e+00
2.03188911e-01 1.39402795e+00 7.87173748e-01 8.33706975e-01
-7.98430324e-01 -1.03703809e+00 3.39053392e-01 5.53050280e-01
-1.57140994e+00 -9.20096099e-01 9.69424486e-01 -7.19567463e-02
9.74202752e-01 3.38773429e-01 1.93723083e-01 1.26223290e+00
-1.54103532e-01 9.53592420e-01 7.52905071e-01 -2.83404112e-01
2.42947146e-01 2.19330341e-01 2.44820476e-01 5.15747488e-01
-4.10875201e-01 2.90726215e-01 -9.04063880e-01 -1.20913886e-01
4.48659360e-01 -9.63961259e-02 -4.16234821e-01 2.67856926e-01
-1.29155457e+00 8.23971033e-01 4.98474628e-01 5.03771722e-01
-2.18444556e-01 -2.45602161e-01 7.15085089e-01 5.15890658e-01
4.01595384e-01 -1.32195726e-01 -2.55622536e-01 -1.51919678e-01
-1.03526747e+00 -9.44515318e-02 4.54440564e-01 7.13967860e-01
1.28669560e-01 4.37225342e-01 -1.04311496e-01 1.19354165e+00
7.45000899e-01 5.86830080e-01 1.24326611e+00 -2.13648945e-01
8.23015749e-01 1.93029225e-01 -3.47693086e-01 -8.51894438e-01
-3.01116973e-01 -4.56168443e-01 -9.37047005e-01 -2.11413819e-02
-1.67783588e-01 -1.44430995e-01 -6.98905826e-01 1.73779941e+00
2.43560538e-01 5.16660273e-01 5.04621148e-01 1.06921065e+00
1.25895178e+00 1.02892029e+00 -2.32863903e-01 -2.68772364e-01
1.34446359e+00 -1.21247602e+00 -9.73830223e-01 -1.09048471e-01
1.57711387e-01 -7.44403124e-01 1.06476521e+00 2.21341506e-01
-8.21405947e-01 -8.24496984e-01 -1.36118686e+00 9.95238051e-02
-4.50405359e-01 4.38063055e-01 -4.89953570e-02 8.25959444e-01
-1.07295632e+00 2.64071018e-01 -8.09737444e-01 -1.21500075e-01
1.56860247e-01 3.22192401e-01 -1.85709834e-01 3.02207440e-01
-1.50293398e+00 6.44703090e-01 2.26909116e-01 5.41642070e-01
-1.08117330e+00 -3.19581091e-01 -1.13785326e+00 2.08160192e-01
-2.63940960e-01 -5.14507806e-03 1.45258510e+00 -7.96547234e-01
-2.01012206e+00 3.72130185e-01 -4.12068844e-01 -4.70290273e-01
2.24086583e-01 9.63546634e-02 -1.04340780e+00 -8.93406644e-02
-1.92680493e-01 3.33000720e-01 1.01433885e+00 -5.68606853e-01
-3.92724603e-01 -3.43691498e-01 -4.05230254e-01 2.36755446e-01
-6.18316114e-01 5.17476380e-01 -1.38413636e-02 -9.35640454e-01
3.46860766e-01 -6.80961847e-01 1.60563126e-01 -3.18120331e-01
-3.77603471e-01 -5.63767195e-01 1.30500221e+00 -9.80702221e-01
1.27522397e+00 -2.52828860e+00 1.31465182e-01 8.14195201e-02
-2.45220751e-01 4.68423784e-01 -3.50603849e-01 4.14752722e-01
-2.22267553e-01 -1.56341761e-01 -1.06351390e-01 -6.50887012e-01
1.01917639e-01 -4.30072337e-01 -5.78036427e-01 7.30061829e-01
-2.47827433e-02 7.12296963e-01 -5.03920615e-01 -3.69670123e-01
1.59066901e-01 8.18624377e-01 -1.49919152e-01 2.43934870e-01
2.85340190e-01 2.64964670e-01 -6.07674941e-02 6.66542172e-01
6.57390714e-01 2.89313272e-02 -1.38165027e-01 1.47976175e-01
4.12964588e-03 7.74542749e-01 -1.22753489e+00 1.69323754e+00
-6.25286758e-01 8.65397990e-01 1.40089020e-02 -1.00756729e+00
1.20533741e+00 8.05946052e-01 -2.18513742e-01 -5.17157614e-01
1.22112609e-01 1.98242083e-01 2.17119157e-02 -3.17012429e-01
2.02538148e-01 -4.07057703e-02 -6.63113222e-02 5.21015525e-01
1.90640077e-01 1.67636603e-01 -4.15716350e-01 -6.23946339e-02
7.65743792e-01 -4.57248926e-01 1.81368992e-01 -4.24396880e-02
1.11206675e+00 -7.84188569e-01 6.88507140e-01 3.80754888e-01
-5.80142140e-01 3.05374086e-01 -1.41885474e-01 -2.52585649e-01
-7.93090940e-01 -9.52184439e-01 -2.12131813e-01 1.19664156e+00
9.29415226e-02 -9.27805454e-02 -8.18099499e-01 -5.63177168e-01
-7.79412165e-02 4.70255435e-01 -5.40092170e-01 -1.89872816e-01
-7.05218434e-01 -2.16754019e-01 1.09245491e+00 8.28445017e-01
6.85661793e-01 -1.38203084e+00 -1.12397246e-01 2.41256565e-01
-1.60974056e-01 -1.10651207e+00 -1.12690079e+00 -1.76838711e-01
-6.98068619e-01 -4.61592644e-01 -8.14631939e-01 -1.34067512e+00
2.20315024e-01 2.90863365e-01 3.40484172e-01 -1.72448963e-01
-8.00508112e-02 -1.43656403e-01 -3.95619601e-01 -2.67996907e-01
-3.37674707e-01 2.30987579e-01 5.34999073e-01 4.45408106e-01
5.83712816e-01 -4.80428845e-01 -4.11828518e-01 5.05506814e-01
-4.79759514e-01 -5.55719376e-01 5.60158372e-01 1.12031305e+00
2.85160035e-01 -1.59164563e-01 1.25841749e+00 -1.15434140e-01
9.58898485e-01 -1.59502700e-01 -6.83110952e-01 2.82588750e-01
-3.39513242e-01 4.52364143e-03 6.56187296e-01 -7.81039476e-01
-1.04150283e+00 -6.10008165e-02 -3.73768210e-01 -5.25249004e-01
-1.09135568e-01 4.98419613e-01 -2.34723628e-01 4.16942015e-02
4.09309894e-01 6.98040783e-01 4.20210650e-04 -6.42116725e-01
1.23551413e-01 1.39005387e+00 5.19986093e-01 2.87226867e-02
6.59911692e-01 3.51566263e-03 -7.45980322e-01 -1.10445929e+00
-4.26544905e-01 -5.62428415e-01 -4.70115900e-01 -3.06536946e-02
6.72779918e-01 -1.07398212e+00 -9.78474677e-01 6.77371800e-01
-1.30922294e+00 1.61078468e-01 -5.72206080e-03 9.24964607e-01
-3.12153071e-01 3.12611878e-01 -9.64582741e-01 -1.25928974e+00
-8.06584954e-01 -1.18916595e+00 1.14448416e+00 2.46660814e-01
7.44532794e-02 -9.76192832e-01 1.52473450e-01 4.35799628e-01
7.23395944e-01 -4.55671936e-01 4.20310289e-01 -1.04828346e+00
-2.80106843e-01 -4.81236845e-01 5.47836535e-02 6.04847133e-01
2.18025550e-01 -3.27626437e-01 -1.57267046e+00 -6.67144299e-01
2.69499362e-01 -3.56229633e-01 9.76459682e-01 2.55707085e-01
1.22354972e+00 -5.44459581e-01 -2.30586752e-01 4.55128878e-01
9.43058074e-01 4.48343366e-01 3.64926130e-01 1.90402716e-02
3.37133586e-01 2.74156988e-01 3.70709062e-01 3.65491025e-02
2.35811546e-01 9.07698154e-01 2.03665765e-03 6.75961003e-02
-2.04629168e-01 -2.50917703e-01 9.77008641e-01 1.71708930e+00
2.22172678e-01 -1.13119632e-01 -6.52519822e-01 6.22187138e-01
-1.70877266e+00 -1.33625841e+00 5.78632772e-01 2.03205681e+00
7.56780148e-01 2.26065107e-02 2.99759328e-01 5.10586321e-01
1.09415245e+00 4.04114783e-01 -7.78379858e-01 -4.67670590e-01
-1.73788697e-01 1.64590746e-01 -1.12633221e-01 6.39705718e-01
-1.18853116e+00 8.11106324e-01 6.00652075e+00 8.66035879e-01
-1.57460034e+00 4.27985847e-01 2.77268082e-01 -1.45805538e-01
2.22462460e-01 -5.62525928e-01 -1.11825287e+00 5.22220850e-01
1.33427870e+00 -8.83661509e-02 5.37244141e-01 1.01324868e+00
1.83188155e-01 7.45621860e-01 -1.06114364e+00 1.27315331e+00
6.49195492e-01 -1.07100534e+00 -2.06203222e-01 -1.96693629e-01
3.67588520e-01 1.92293197e-01 4.03803438e-01 4.90850836e-01
-5.47770262e-02 -1.17538047e+00 7.47617364e-01 7.87064135e-02
8.94324124e-01 -8.59054685e-01 9.97370183e-01 3.04076523e-01
-1.66576874e+00 -3.01831156e-01 -2.52302825e-01 1.59204647e-01
1.67015880e-01 1.82878509e-01 -9.70463932e-01 3.22507054e-01
7.02700555e-01 5.89341402e-01 -2.71100223e-01 8.27618778e-01
-2.27959812e-01 8.80921960e-01 -1.11874476e-01 -5.23921669e-01
1.51893705e-01 2.15120494e-01 5.51083446e-01 1.35796177e+00
2.39380255e-01 -1.29912093e-01 -1.19699031e-01 5.81357002e-01
-3.08977574e-01 3.23479772e-01 -4.57953423e-01 1.31631950e-02
8.43570173e-01 1.03457546e+00 -1.82199359e-01 -3.56912315e-01
-2.89082050e-01 9.50880527e-01 2.78169364e-01 4.84851867e-01
-9.13420677e-01 -1.05110323e+00 6.05599701e-01 -3.91698092e-01
6.19554520e-01 1.01660170e-01 -1.20931230e-01 -1.24564767e+00
4.58804518e-01 -9.78259921e-01 3.80030036e-01 -5.35746157e-01
-1.28391838e+00 1.16853774e+00 -5.42705357e-01 -1.35967720e+00
-3.37878197e-01 -5.16922951e-01 -8.13214839e-01 1.07735109e+00
-1.56621778e+00 -1.16461802e+00 -5.57015091e-02 8.17249537e-01
1.03763914e+00 -7.71481812e-01 1.00142825e+00 5.10145366e-01
-7.96512008e-01 1.37268281e+00 6.47759661e-02 5.36321342e-01
5.88488221e-01 -7.48430133e-01 7.25229800e-01 9.03185248e-01
2.70752639e-01 6.99048936e-01 1.66898817e-01 -2.36977190e-01
-1.09441316e+00 -1.11938751e+00 1.08007920e+00 -3.70863415e-02
5.24314463e-01 -7.09358811e-01 -9.78961587e-01 7.07568049e-01
3.79233330e-01 1.20211676e-01 7.44703591e-01 1.56189486e-01
-7.81843841e-01 -4.08801287e-01 -1.21438241e+00 2.85536170e-01
7.29654372e-01 -1.23084331e+00 -9.21583056e-01 1.62450358e-01
1.12987971e+00 -4.03055221e-01 -7.23937333e-01 2.80622214e-01
7.20173597e-01 -3.59785169e-01 8.54777813e-01 -5.17996013e-01
-1.73493907e-01 -3.91650230e-01 -4.69206780e-01 -1.47280490e+00
-1.97011754e-01 -5.54147005e-01 -8.16666037e-02 1.28968835e+00
8.27118397e-01 -9.10203636e-01 3.90700728e-01 -1.93046495e-01
-2.66869158e-01 -6.23556793e-01 -1.50994873e+00 -1.04005539e+00
-2.19473653e-02 -3.81756276e-01 8.64854336e-01 1.07945478e+00
3.45217496e-01 5.08643627e-01 -3.61036450e-01 6.10054255e-01
4.34734792e-01 1.39639350e-02 2.99318880e-01 -1.03077996e+00
-2.24093169e-01 -3.97245497e-01 -8.20615411e-01 -1.30746579e+00
3.73115987e-01 -9.36302662e-01 1.75896898e-01 -1.01440990e+00
6.73272535e-02 -1.04106650e-01 -6.51806533e-01 3.66123110e-01
1.82749622e-03 5.16172647e-02 -5.21909557e-02 1.36270504e-02
-3.36334705e-01 8.05843890e-01 7.10197210e-01 -4.56980497e-01
-3.47281992e-01 2.03219473e-01 -2.64318824e-01 4.24192846e-01
9.08382893e-01 -2.65942454e-01 -2.96618819e-01 -2.81864315e-01
-5.77283978e-01 2.42696881e-01 2.42353067e-01 -1.14317644e+00
6.22993350e-01 2.96784073e-01 2.87054718e-01 -8.11636209e-01
7.21743941e-01 -5.00713348e-01 -2.84251869e-01 4.43956226e-01
-6.91022277e-01 1.79988578e-01 2.87703186e-01 4.62209821e-01
-6.13729656e-01 -6.07030578e-02 9.73698258e-01 4.22937751e-01
-2.81312943e-01 2.27676421e-01 -2.69742727e-01 -2.81455398e-01
8.72623801e-01 -6.42220769e-03 -3.18306834e-01 -3.91021371e-01
-7.24605203e-01 -6.35687411e-02 -2.26644769e-01 9.03074622e-01
1.04069722e+00 -1.73269486e+00 -8.89572084e-01 7.13190675e-01
1.24975204e-01 -3.88808340e-01 2.89330840e-01 6.47902846e-01
-1.06961824e-01 6.89574659e-01 2.05108374e-01 -6.75165176e-01
-1.56677306e+00 4.66631949e-01 4.96550113e-01 1.83390677e-01
-5.49737155e-01 1.16515899e+00 -1.40248895e-01 -8.32855523e-01
7.94658124e-01 -4.73022938e-01 -8.95292461e-02 1.32474890e-02
9.58647192e-01 2.15072870e-01 3.16417307e-01 -1.02423298e+00
-7.16279030e-01 5.04557729e-01 -3.39976788e-01 -5.24278939e-01
1.33839285e+00 -6.37105033e-02 1.68161511e-01 6.84580505e-01
1.38338423e+00 1.30853161e-01 -7.93227434e-01 -4.89282995e-01
-1.77413389e-01 -1.74228936e-01 7.90283307e-02 -6.53707385e-01
-1.15696704e+00 1.23196197e+00 1.07185113e+00 1.74004257e-01
9.53334033e-01 1.69068463e-02 1.06004095e+00 4.73517150e-01
-4.54831049e-02 -1.18385339e+00 1.69114634e-01 4.69315976e-01
1.15195978e+00 -1.21754944e+00 -5.06741881e-01 8.08944702e-02
-7.50898182e-01 9.94024217e-01 5.55082440e-01 2.62415782e-03
8.83104622e-01 6.93697482e-02 4.47063506e-01 1.15345404e-01
-7.58115053e-01 2.88575530e-01 3.00739974e-01 4.17499751e-01
3.77233356e-01 1.08693346e-01 1.47336811e-01 6.68627322e-01
-5.21519005e-01 -4.90276843e-01 9.40262377e-02 5.42316914e-01
-4.22966897e-01 -6.85576558e-01 -4.99291569e-01 -5.93757033e-02
-3.39528948e-01 -2.48229340e-01 -3.61568600e-01 2.85676003e-01
-3.27929825e-01 1.19073904e+00 8.44756048e-03 -6.77548289e-01
4.61296052e-01 2.88107097e-01 -8.27589557e-02 -3.16874325e-01
-5.68788946e-01 -9.38665941e-02 -1.78838283e-01 -2.86344081e-01
-2.69868404e-01 -8.12263012e-01 -1.11100209e+00 -2.31932383e-03
-8.35298657e-01 4.44799066e-01 9.93847966e-01 7.54484713e-01
4.56872463e-01 5.73663771e-01 1.13635051e+00 -6.97815776e-01
-1.06659377e+00 -1.59528363e+00 -4.86093968e-01 9.13956761e-02
1.07130849e+00 -5.21950006e-01 -7.46529818e-01 -4.47248966e-02] | [14.343430519104004, 6.067890644073486] |
138ecf85-4a40-4c96-b99b-aeab28db503b | hicem-a-high-coverage-emotion-model-for | 2206.07593 | null | https://arxiv.org/abs/2206.07593v1 | https://arxiv.org/pdf/2206.07593v1.pdf | HICEM: A High-Coverage Emotion Model for Artificial Emotional Intelligence | As social robots and other intelligent machines enter the home, artificial emotional intelligence (AEI) is taking center stage to address users' desire for deeper, more meaningful human-machine interaction. To accomplish such efficacious interaction, the next-generation AEI need comprehensive human emotion models for training. Unlike theory of emotion, which has been the historical focus in psychology, emotion models are a descriptive tools. In practice, the strongest models need robust coverage, which means defining the smallest core set of emotions from which all others can be derived. To achieve the desired coverage, we turn to word embeddings from natural language processing. Using unsupervised clustering techniques, our experiments show that with as few as 15 discrete emotion categories, we can provide maximum coverage across six major languages--Arabic, Chinese, English, French, Spanish, and Russian. In support of our findings, we also examine annotations from two large-scale emotion recognition datasets to assess the validity of existing emotion models compared to human perception at scale. Because robust, comprehensive emotion models are foundational for developing real-world affective computing applications, this work has broad implications in social robotics, human-machine interaction, mental healthcare, and computational psychology. | ['James Z. Wang', 'Benjamin Wortman'] | 2022-06-15 | null | null | null | null | ['emotional-intelligence'] | ['natural-language-processing'] | [-8.04248750e-02 4.36652690e-01 -1.73048928e-01 -6.18441701e-01
4.15184572e-02 -2.87876815e-01 2.59923697e-01 4.43890035e-01
-5.73384047e-01 4.67158735e-01 3.90661180e-01 1.40393123e-01
-3.34063247e-02 -4.87996548e-01 1.61085173e-01 -9.60251689e-02
-1.07125789e-01 2.96926647e-01 -6.38286889e-01 -6.38787806e-01
6.67840987e-02 2.52841711e-01 -1.50111842e+00 5.10748774e-02
9.71214056e-01 1.04678416e+00 -2.27669686e-01 3.04709524e-01
1.15966894e-01 7.24073529e-01 -3.61651093e-01 -5.16256392e-01
-2.98275918e-01 -3.70045990e-01 -7.58780956e-01 -3.16124745e-02
-5.10985434e-01 -4.04132009e-02 1.41031250e-01 1.07180393e+00
5.85199773e-01 3.95588487e-01 8.17121267e-01 -1.72229278e+00
-1.07186115e+00 3.87215942e-01 -7.83393309e-02 -6.32920682e-01
8.09667349e-01 -1.61744490e-01 1.15164399e+00 -9.74879026e-01
7.49210536e-01 1.17217648e+00 6.74977303e-01 8.18890870e-01
-6.15221858e-01 -4.67621297e-01 9.77524444e-02 1.34733900e-01
-1.08520734e+00 -4.16950285e-01 8.14000487e-01 -2.50575423e-01
1.20754242e+00 1.12953089e-01 9.24466133e-01 1.20547879e+00
1.12517208e-01 7.81918168e-01 9.92045224e-01 -6.61719918e-01
4.50732976e-01 6.64517045e-01 2.88455874e-01 4.50255454e-01
1.40197545e-01 -3.62269670e-01 -4.63865012e-01 -1.00873083e-01
1.91115588e-01 -9.29594785e-02 -4.72944975e-02 -1.69320613e-01
-1.15889955e+00 1.02936125e+00 2.76341945e-01 4.94611531e-01
-6.79260790e-01 -1.78818434e-01 6.50114536e-01 2.54230231e-01
4.46312487e-01 8.45389187e-01 -4.79645818e-01 -8.62091541e-01
-2.96137873e-02 -1.22114800e-01 1.00721550e+00 8.60816121e-01
6.03096306e-01 -5.91121353e-02 4.52726841e-01 1.34667861e+00
4.82205600e-01 4.02456045e-01 6.74841404e-01 -1.21562076e+00
-3.12848389e-01 9.08890367e-01 1.72216848e-01 -1.24710739e+00
-7.52705395e-01 1.84719875e-01 -6.95740819e-01 2.62033986e-03
-1.16594069e-01 -5.28979897e-01 -3.59516829e-01 1.84221077e+00
3.45477164e-01 -4.75064784e-01 4.74869609e-01 8.54137063e-01
7.45907426e-01 5.51428258e-01 3.33365202e-01 -3.17638457e-01
1.43885410e+00 -6.48334384e-01 -9.70007658e-01 -5.55379987e-01
8.23821187e-01 -5.34278214e-01 1.10638380e+00 5.42596102e-01
-6.80783272e-01 -3.70625973e-01 -1.06096697e+00 -2.50956099e-02
-7.34764993e-01 -1.13498569e-02 1.14680421e+00 7.26031959e-01
-1.07802474e+00 5.05149215e-02 -5.28210163e-01 -1.06422532e+00
2.39150032e-01 3.12275052e-01 -5.76140761e-01 6.79153530e-03
-1.35511494e+00 1.42600763e+00 1.81302190e-01 9.67859924e-02
1.57426354e-02 -1.47933057e-02 -1.05939257e+00 -1.74464256e-01
-1.72574073e-01 -3.47029150e-01 1.10262954e+00 -1.48571110e+00
-1.56175363e+00 9.42670643e-01 -8.86254907e-02 -4.73934822e-02
-3.36934119e-01 -1.33718282e-01 -8.14359784e-01 2.29257807e-01
-6.46579126e-03 9.53670025e-01 3.79469573e-01 -1.14318407e+00
-5.26441991e-01 -5.65405488e-01 4.95984778e-02 5.75940013e-01
-9.49971199e-01 4.41826224e-01 -1.34986594e-01 -1.57119930e-01
-7.62683228e-02 -1.18739104e+00 -3.15231770e-01 -1.30968556e-01
2.50500709e-01 -5.55145144e-01 5.59286118e-01 -5.30036211e-01
1.04206777e+00 -2.43074203e+00 2.16137618e-01 2.56474674e-01
2.36583069e-01 -1.71338897e-02 -1.12706341e-01 3.62171352e-01
-9.34058055e-02 3.44060659e-02 1.27740234e-01 -2.86001086e-01
5.24609149e-01 3.34368348e-01 1.30597636e-01 2.17570111e-01
2.58092195e-01 9.43071842e-01 -1.04717886e+00 -4.72707421e-01
2.39621595e-01 4.43822592e-01 -7.71973670e-01 2.00487927e-01
1.68357790e-01 -2.03993805e-02 -4.07495350e-01 6.47258520e-01
4.36513312e-02 -1.86920837e-02 2.62755930e-01 -1.24513488e-02
1.55077875e-01 -1.55911013e-01 -8.71434391e-01 1.48097646e+00
-5.59967577e-01 7.22615421e-01 1.95952624e-01 -9.69144762e-01
1.11153436e+00 4.22174901e-01 1.01012337e+00 -6.14770174e-01
6.02356732e-01 5.40746860e-02 2.63906866e-01 -7.42870092e-01
8.17146540e-01 -3.57721597e-01 -6.13180757e-01 5.52645385e-01
-1.36190327e-03 -4.60193694e-01 -7.67432898e-02 1.43028453e-01
8.53382111e-01 -1.11135818e-01 5.52993953e-01 -1.51403531e-01
5.19816950e-02 -7.30196312e-02 5.99522591e-01 1.62588835e-01
-1.06284082e+00 6.52358755e-02 2.89005041e-01 -2.20484197e-01
-7.58703411e-01 -6.54607892e-01 -5.85106425e-02 1.42341053e+00
2.05233946e-01 -4.37623292e-01 -7.77331233e-01 -2.65767723e-01
-3.66781890e-01 8.75964701e-01 -5.58593392e-01 -4.99028653e-01
2.36338645e-01 -6.86925590e-01 4.50515479e-01 3.63743335e-01
2.25064665e-01 -1.75172675e+00 -1.02950501e+00 2.49453157e-01
-3.97048801e-01 -1.28020310e+00 1.18564442e-02 3.54657203e-01
-2.41209775e-01 -6.79801583e-01 -2.22122267e-01 -1.03375995e+00
4.75780129e-01 -2.17872672e-02 9.68680084e-01 -2.01497078e-01
-2.95980453e-01 1.17872417e+00 -8.97907376e-01 -8.60415936e-01
-2.45352238e-01 -2.73884565e-01 6.25880003e-01 -2.36090034e-01
1.08758676e+00 -5.43205142e-01 -3.38551700e-01 4.04468030e-02
-7.17312455e-01 -1.80504486e-01 2.47301027e-01 5.41262746e-01
-3.07890009e-02 1.71927765e-01 1.00526202e+00 -2.13729724e-01
1.17000151e+00 -7.99803257e-01 5.56824625e-01 2.93910533e-01
-5.00856936e-01 -5.23319542e-01 3.91655117e-01 -5.65036356e-01
-1.03181207e+00 -2.02886239e-01 -1.98282614e-01 1.80189908e-01
-4.35576886e-01 7.32625723e-01 -1.75310344e-01 3.40648711e-01
6.80665731e-01 -3.42404276e-01 1.54320940e-01 3.53850096e-01
6.54893577e-01 1.30852461e+00 2.59779155e-01 -7.08857596e-01
-1.42906271e-02 1.67421594e-01 -6.77482665e-01 -1.33666527e+00
-4.72896516e-01 -4.27576840e-01 -3.84737015e-01 -6.44476116e-01
1.13968074e+00 -7.75344253e-01 -9.20175135e-01 2.30006337e-01
-1.19028723e+00 -3.81910801e-01 -3.37882459e-01 6.99114978e-01
-5.72627962e-01 1.79235905e-01 -6.92950010e-01 -1.21642900e+00
-4.57436264e-01 -9.54270363e-01 8.87778938e-01 3.88675839e-01
-1.19720209e+00 -1.08275235e+00 4.11298126e-02 2.33637601e-01
2.25851417e-01 1.93746448e-01 9.66641426e-01 -8.31209481e-01
7.26345897e-01 -2.38162667e-01 -1.10972166e-01 5.65385222e-01
2.85855621e-01 7.30806366e-02 -7.71640658e-01 1.42537251e-01
1.93250299e-01 -1.00552857e+00 -6.43406762e-03 6.50379062e-02
5.68086267e-01 3.31683084e-02 -3.74204144e-02 7.44638965e-03
1.04628944e+00 7.28249609e-01 7.34588087e-01 2.69127339e-01
3.78673643e-01 1.01536596e+00 9.72090602e-01 7.43649364e-01
9.60751116e-01 1.16408288e-01 9.18879360e-02 9.76875331e-03
6.57930553e-01 1.49460763e-01 6.29443824e-01 1.18219960e+00
-3.83132957e-02 -1.48967924e-02 -9.45904434e-01 5.03761649e-01
-1.85790420e+00 -9.18633819e-01 3.19252700e-01 1.70507324e+00
7.58593559e-01 -1.38558194e-01 -4.99948636e-02 9.32603180e-02
3.85002166e-01 -1.88438222e-01 -5.21599770e-01 -1.19733536e+00
1.32112756e-01 1.76569715e-01 -3.01219970e-01 2.55141020e-01
-9.52338099e-01 1.25262344e+00 6.30269718e+00 3.20865065e-01
-1.01867855e+00 -1.42423371e-02 7.42031813e-01 8.65193978e-02
-2.22303137e-01 -4.28682357e-01 -8.93241465e-02 1.31771326e-01
1.05525458e+00 -1.21886902e-01 7.15034068e-01 1.27384436e+00
3.00132841e-01 -1.05296500e-01 -1.00552905e+00 1.36341226e+00
2.52523750e-01 -3.75881523e-01 -5.49778879e-01 -9.61403400e-02
5.66045403e-01 -1.15222640e-01 -8.42494592e-02 8.01141143e-01
2.89290398e-01 -1.05349982e+00 5.39861381e-01 3.97992283e-01
6.58832073e-01 -9.83058095e-01 6.94270194e-01 6.20937571e-02
-7.53024638e-01 -1.19049489e-01 -3.33819360e-01 -5.57053447e-01
5.29384278e-02 3.14385593e-01 -4.56630349e-01 1.56709924e-01
7.22590685e-01 8.15450132e-01 -2.03119591e-01 1.94367930e-01
5.35474680e-02 1.92732438e-01 -2.64726341e-01 -6.49716735e-01
3.17781344e-02 -5.09356678e-01 1.34982154e-01 1.23145127e+00
3.95167738e-01 4.82062668e-01 1.83378354e-01 3.19807619e-01
1.11823246e-01 4.28142726e-01 -9.62311804e-01 -7.19800174e-01
4.76463407e-01 1.58005989e+00 -9.21339512e-01 -1.41450137e-01
-5.68586886e-01 1.39427114e+00 4.67527717e-01 2.44802773e-01
-8.91656160e-01 -6.65564299e-01 1.05017912e+00 -7.54202008e-01
-4.76580143e-01 -4.31066006e-01 -5.00654519e-01 -1.11942220e+00
-3.30024958e-01 -1.25009465e+00 3.38037051e-02 -1.02305138e+00
-1.45874357e+00 7.11916149e-01 -3.55496943e-01 -9.38079953e-01
-4.48421508e-01 -9.32640970e-01 -3.03411841e-01 1.40062258e-01
-7.70136058e-01 -1.17584670e+00 -3.74094516e-01 4.62637872e-01
2.48640820e-01 -1.01291366e-01 1.42295051e+00 2.20355257e-01
-5.83644688e-01 4.38459277e-01 -1.60260811e-01 5.82982302e-02
9.12236810e-01 -9.67116058e-01 -2.06558287e-01 2.48863414e-01
-6.24450371e-02 7.19761193e-01 6.11612320e-01 -4.72193152e-01
-1.53175581e+00 -4.85829681e-01 8.76490355e-01 -4.48058605e-01
8.23028386e-01 -2.62506366e-01 -4.27070916e-01 6.81154490e-01
5.02760410e-01 -5.05701363e-01 1.31223702e+00 6.39209211e-01
-9.85761657e-02 1.88533291e-01 -1.22271955e+00 9.81851459e-01
9.92055893e-01 -6.00918949e-01 -6.55018628e-01 -4.95357290e-02
7.42017806e-01 1.27508894e-01 -1.23229456e+00 3.09055656e-01
8.65988135e-01 -8.25179219e-01 6.45255923e-01 -5.85863829e-01
6.86321199e-01 2.45110333e-01 -4.27738488e-01 -1.61437809e+00
-2.92306632e-01 -4.77223188e-01 2.22576305e-01 1.31757009e+00
3.43944430e-01 -6.98345721e-01 4.36753154e-01 1.44702244e+00
-2.39638135e-01 -8.82699192e-01 -5.75483382e-01 -3.22555542e-01
8.55807401e-03 -1.05853605e+00 3.93368602e-01 1.37533617e+00
1.23559380e+00 6.36930525e-01 -1.64555192e-01 -3.28808606e-01
2.65544541e-02 -3.53322923e-01 6.90670133e-01 -1.34869564e+00
1.05237998e-01 -5.97687364e-01 -5.91936707e-01 -4.64087456e-01
5.40293872e-01 -6.24968052e-01 3.51239234e-01 -1.80205047e+00
9.60962996e-02 -4.02446508e-01 -3.72307807e-01 5.49360335e-01
-1.37024939e-01 3.14385653e-01 1.78160682e-01 -3.80718112e-01
-8.74990225e-01 9.33351755e-01 8.80369961e-01 8.32323208e-02
-3.54819775e-01 -6.59985065e-01 -1.32064366e+00 1.18683374e+00
9.72105801e-01 1.15452982e-01 -4.91838992e-01 -2.46295929e-02
5.77883959e-01 -2.61014223e-01 -2.24852547e-01 -8.77129555e-01
6.98706210e-02 -4.15048122e-01 2.82257587e-01 1.99886963e-01
6.83707952e-01 -1.07416892e+00 -2.67115980e-01 5.28092533e-02
-1.98726371e-01 -3.10401414e-02 1.57257825e-01 8.54874104e-02
-1.84372559e-01 -2.40627810e-01 6.14523709e-01 7.09136799e-02
-9.79955614e-01 4.35490534e-02 -8.75828981e-01 2.71000583e-02
1.23845220e+00 -2.65374124e-01 6.12202771e-02 -7.55501390e-01
-7.96241343e-01 2.69805372e-01 5.47315776e-01 7.74178505e-01
8.86413932e-01 -1.39099717e+00 -2.82878399e-01 -1.00349031e-01
4.62624550e-01 -3.67383480e-01 6.27144575e-02 6.20756447e-01
-1.16473012e-01 1.35970622e-01 -3.97404999e-01 -4.77991067e-02
-1.01175630e+00 5.07529378e-01 -8.42899829e-03 2.11301431e-01
-1.37422547e-01 5.20133436e-01 -2.73597389e-01 -7.80857325e-01
2.87085861e-01 -2.50725448e-01 -3.48762929e-01 3.64480376e-01
2.92231113e-01 2.53184468e-01 -4.01511014e-01 -9.51158226e-01
-3.92823189e-01 1.28677115e-01 3.41226310e-01 -4.81612086e-01
1.44393921e+00 -3.28697354e-01 -5.48569858e-01 7.60171652e-01
1.25586057e+00 -1.02005318e-01 -4.14793879e-01 3.36672693e-01
2.57961214e-01 1.24046490e-01 -2.41389662e-01 -7.29483843e-01
-3.33913833e-01 6.88956916e-01 6.26946568e-01 2.74949640e-01
1.33907342e+00 -9.34313461e-02 7.20218718e-01 5.63577056e-01
6.28270745e-01 -1.73354495e+00 3.25731009e-01 7.56393731e-01
8.76161575e-01 -1.39206433e+00 -2.63996184e-01 -6.30920902e-02
-1.33760619e+00 9.47493911e-01 8.51491272e-01 1.33981317e-01
7.86700010e-01 1.03720881e-01 1.79923803e-01 -2.45144323e-01
-5.54889560e-01 -2.99670845e-01 4.15358432e-02 7.79540002e-01
8.71981621e-01 4.29543287e-01 -3.72873932e-01 1.21983194e+00
-3.77013862e-01 2.66656075e-02 4.00002986e-01 8.52597594e-01
-5.64716220e-01 -1.02666712e+00 -1.50180072e-01 4.74241555e-01
-2.42033899e-01 -3.84072214e-03 -7.13286698e-01 6.04095519e-01
1.96272239e-01 1.49475789e+00 5.97518496e-02 -6.29939973e-01
3.18256557e-01 3.51894021e-01 3.08442593e-01 -5.26585162e-01
-4.30536062e-01 -4.30754572e-01 4.42507058e-01 -4.37562972e-01
-6.23175085e-01 -4.95575666e-01 -1.68760848e+00 -1.00999035e-01
-3.40554863e-02 1.85165942e-01 8.49745750e-01 1.00813627e+00
5.33800066e-01 1.42834887e-01 5.14787376e-01 -8.57377470e-01
5.66332340e-02 -1.21062362e+00 -5.36812305e-01 5.62849224e-01
-3.64569753e-01 -4.42956060e-01 -2.27234706e-01 -6.57042190e-02] | [13.018081665039062, 5.744825839996338] |
825314e7-ea7b-4d8f-ae20-1c89ba2bce1d | computing-and-exploiting-document-structure | 2211.03229 | null | https://arxiv.org/abs/2211.03229v1 | https://arxiv.org/pdf/2211.03229v1.pdf | Computing and Exploiting Document Structure to Improve Unsupervised Extractive Summarization of Legal Case Decisions | Though many algorithms can be used to automatically summarize legal case decisions, most fail to incorporate domain knowledge about how important sentences in a legal decision relate to a representation of its document structure. For example, analysis of a legal case summarization dataset demonstrates that sentences serving different types of argumentative roles in the decision appear in different sections of the document. In this work, we propose an unsupervised graph-based ranking model that uses a reweighting algorithm to exploit properties of the document structure of legal case decisions. We also explore the impact of using different methods to compute the document structure. Results on the Canadian Legal Case Law dataset show that our proposed method outperforms several strong baselines. | ['Diane Litman', 'Yang Zhong'] | 2022-11-06 | null | null | null | null | ['unsupervised-extractive-summarization', 'extractive-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.74612057e-01 6.13486767e-01 -8.35546613e-01 -6.32784367e-01
-8.31592798e-01 -8.72368515e-01 9.63888645e-01 9.26763713e-01
-2.51791447e-01 7.44216740e-01 1.51873732e+00 -7.75420606e-01
-6.78879023e-01 -7.59031057e-01 -1.42270312e-01 -9.37232599e-02
1.94814935e-01 4.96779889e-01 3.04014981e-01 -5.67981958e-01
9.68445837e-01 3.67173791e-01 -1.05570436e+00 7.84475148e-01
1.16501904e+00 8.37675482e-03 -3.86051744e-01 4.08831060e-01
-3.78192723e-01 1.49978626e+00 -1.04180801e+00 -8.91568124e-01
1.37389377e-01 -5.52850187e-01 -1.13424253e+00 -1.38722956e-01
1.02174091e+00 -3.50766003e-01 -6.25812531e-01 8.73487055e-01
3.85000110e-01 1.28291160e-01 1.31976032e+00 -8.68777394e-01
-3.54393721e-01 1.10683727e+00 -6.27510965e-01 9.23797965e-01
7.49789655e-01 -4.49189425e-01 1.47723269e+00 8.10427517e-02
1.36070001e+00 1.37924933e+00 4.36198205e-01 3.40544343e-01
-1.18740332e+00 -1.82014853e-01 3.99465948e-01 3.91338795e-01
-5.41918337e-01 -7.72522926e-01 1.03305006e+00 -5.62458038e-01
1.20275319e+00 1.43264338e-01 5.20069897e-01 7.30712116e-01
6.10505521e-01 6.24030650e-01 7.97520638e-01 -5.04599690e-01
1.68551952e-01 -6.15897477e-01 1.10191238e+00 5.18314898e-01
1.06346202e+00 -4.69448984e-01 -2.65659779e-01 -9.75731254e-01
-2.75262713e-01 -1.90490440e-01 2.40679756e-02 2.33641006e-02
-5.31118631e-01 1.09483755e+00 2.78497990e-02 6.55579448e-01
-4.19396788e-01 2.12205157e-01 7.33425558e-01 3.49350601e-01
5.26226580e-01 6.16781890e-01 1.12597626e-02 -2.37000659e-01
-8.81915212e-01 7.24237800e-01 8.93032610e-01 5.97584963e-01
4.59257215e-01 -5.26235461e-01 -8.34997833e-01 8.51502240e-01
1.48313388e-01 8.37678239e-02 1.55904382e-01 -1.12400448e+00
1.16357601e+00 1.00135970e+00 -2.76068062e-01 -1.28994179e+00
-3.07574660e-01 -1.27742728e-02 -3.16727132e-01 -4.24390053e-03
2.68214732e-01 -1.98601022e-01 -7.90151477e-01 1.23717928e+00
1.39393836e-01 -5.57202101e-01 2.42768794e-01 1.54099107e-01
8.63902271e-01 3.90985638e-01 1.52342096e-01 -2.60370582e-01
9.87236142e-01 -4.49344695e-01 -8.35990012e-01 -2.01490819e-01
9.66338873e-01 -7.51569748e-01 4.76864040e-01 -1.72228739e-01
-9.48052347e-01 1.15627229e-01 -9.64483261e-01 -4.97225374e-01
1.10912547e-01 -2.85627712e-02 7.12057471e-01 4.97510076e-01
-7.36903191e-01 5.05473197e-01 -4.02503788e-01 -5.30632675e-01
7.99477756e-01 -2.42931485e-01 -1.21809848e-01 -2.46479064e-01
-9.80410397e-01 1.12968624e+00 4.16600674e-01 -4.88157600e-01
-3.00297309e-02 -4.97093648e-01 -8.66985083e-01 1.83593810e-01
6.65167630e-01 -7.38028526e-01 1.16670787e+00 -2.96939343e-01
-7.46545851e-01 8.93968880e-01 -5.35391390e-01 -6.76336408e-01
3.16651732e-01 -1.55614138e-01 -3.71425986e-01 5.42044938e-01
6.58738434e-01 3.94197106e-02 5.23134053e-01 -9.80327845e-01
-7.62511432e-01 -4.69676584e-01 7.62357593e-01 1.22649416e-01
-1.19974345e-01 5.05400002e-01 -9.49272737e-02 -7.39147782e-01
-4.72976193e-02 -6.04729533e-01 -4.27119255e-01 -6.31535113e-01
-6.98401928e-01 -6.25641465e-01 5.22750676e-01 -6.07888699e-01
1.88258135e+00 -1.73950219e+00 -1.18126869e-01 4.57875758e-01
5.29276729e-01 -7.33042136e-02 5.30304685e-02 1.16651130e+00
1.45207837e-01 6.96716309e-01 -3.93652141e-01 2.94159681e-01
1.62534222e-01 4.73352671e-01 -7.18044043e-01 3.25313777e-01
-1.12788297e-01 7.97678232e-01 -1.00176847e+00 -8.90782833e-01
-1.28213093e-01 -8.06571916e-02 -7.07786620e-01 -5.97193122e-01
1.13823805e-02 -2.91943640e-01 -8.83809924e-01 3.47867846e-01
2.40078613e-01 -4.54139411e-02 6.81051373e-01 -1.43993035e-01
1.40659912e-02 1.13444614e+00 -3.06402594e-01 1.35492802e+00
6.63699508e-02 7.88844407e-01 -2.67269909e-01 -9.73506153e-01
3.42863888e-01 9.89925861e-02 3.95382792e-01 -7.04439402e-01
2.58468185e-03 -9.51522961e-03 6.06172621e-01 -7.02392817e-01
6.73669279e-01 3.47593836e-02 -4.57921892e-01 9.46807623e-01
-3.91096413e-01 -1.99327633e-01 1.26219082e+00 1.14508927e+00
1.58631134e+00 -4.84687924e-01 8.50331783e-01 -3.01543683e-01
3.02662671e-01 5.45001209e-01 6.22320771e-01 1.14309943e+00
1.02997482e-01 3.07903647e-01 1.30217075e+00 -3.74356270e-01
-8.41917157e-01 -6.34038150e-01 -2.41755709e-01 8.22760403e-01
-8.13139081e-02 -9.67634737e-01 -5.83478093e-01 -1.17944348e+00
3.15232456e-01 1.13701546e+00 -8.68109524e-01 -1.62993342e-01
-9.94680405e-01 -6.11271441e-01 5.09857655e-01 4.26097840e-01
1.30274668e-01 -6.58231974e-01 -8.01792741e-01 2.83944249e-01
-1.53902113e-01 -9.81365681e-01 -4.78231251e-01 -3.39562327e-01
-8.26931477e-01 -1.64640629e+00 -8.46562162e-03 -3.45604479e-01
8.88489783e-01 2.69683719e-01 1.03147030e+00 3.91326696e-01
-9.67860222e-02 4.04513091e-01 -5.60942173e-01 -4.87411946e-01
-9.12427545e-01 4.18804795e-01 -3.70413095e-01 -5.45909107e-01
4.30720687e-01 -5.32153010e-01 -1.39918134e-01 -4.87349957e-01
-1.06449914e+00 -2.48320192e-01 2.59123355e-01 3.99203181e-01
-1.21365273e-02 -9.39354077e-02 5.41651011e-01 -1.68072701e+00
1.62960565e+00 -5.10116696e-01 -2.16257244e-01 6.35898232e-01
-7.88314760e-01 4.15339023e-01 3.05514783e-01 1.75505951e-01
-1.13358474e+00 -5.39203644e-01 -1.59238409e-02 5.14302611e-01
1.50788575e-01 9.83935952e-01 1.75732747e-01 5.78188837e-01
8.53531897e-01 -3.27769578e-01 -1.66479155e-01 -1.48725688e-01
5.25274873e-01 5.89089692e-01 2.56148219e-01 -6.40612721e-01
9.10544574e-01 5.81276476e-01 5.52008227e-02 -6.00015759e-01
-1.19490147e+00 -4.83430177e-01 -6.67792022e-01 -2.16994315e-01
7.17510164e-01 -3.08577567e-01 -5.86787947e-02 -3.00040424e-01
-1.45170224e+00 2.03970790e-01 -3.81236404e-01 1.45849571e-01
-1.52510986e-01 8.00621152e-01 -5.51267862e-01 -5.69920361e-01
-3.86487722e-01 -5.90151072e-01 7.85670698e-01 -1.12111077e-01
-7.39754438e-01 -8.63288999e-01 4.36153740e-01 8.00094485e-01
2.24322483e-01 5.32995701e-01 1.57219589e+00 -9.78769183e-01
8.10229108e-02 -3.70729297e-01 -1.58474967e-02 -1.44212291e-01
5.47825336e-01 3.64771575e-01 -3.88298362e-01 -2.79449895e-02
-2.51943201e-01 5.68989106e-03 1.16308868e+00 4.08495724e-01
4.92280960e-01 -9.60807920e-01 -6.26018345e-01 -1.59121364e-01
9.95871127e-01 3.60483229e-01 6.46812677e-01 3.08261424e-01
4.56050754e-01 8.27914596e-01 3.12518835e-01 2.47005805e-01
4.80074376e-01 4.34379995e-01 -1.30344152e-01 2.75176913e-01
-2.77752042e-01 -2.99692214e-01 2.97370583e-01 5.86304784e-01
-2.78023005e-01 -5.31376600e-01 -1.15730417e+00 5.91888666e-01
-1.95701504e+00 -1.61312032e+00 -4.27031785e-01 1.61191475e+00
5.46724975e-01 4.14362878e-01 1.23609878e-01 1.91918224e-01
7.46421456e-01 6.78037584e-01 -3.54771852e-01 -7.19802201e-01
-2.06423223e-01 5.88338673e-02 3.99772525e-01 5.46478391e-01
-9.00500119e-01 8.19312572e-01 7.25790834e+00 6.49354815e-01
-4.53655809e-01 -1.03243269e-01 3.04467797e-01 -2.66398430e-01
-7.89140999e-01 4.65133905e-01 -5.78715026e-01 2.48539001e-01
6.35370135e-01 -9.62448657e-01 -3.16902250e-01 6.01575971e-01
2.92471111e-01 -3.47676963e-01 -9.45110857e-01 3.23630273e-01
4.24274683e-01 -1.83196342e+00 5.99031687e-01 4.37545627e-01
8.61389458e-01 3.25290509e-03 -5.67584336e-01 8.33153054e-02
7.74566412e-01 -5.40809274e-01 5.73329031e-01 3.99623871e-01
2.54090011e-01 -6.83074892e-01 7.62187719e-01 1.62788868e-01
-6.90707743e-01 -3.29518020e-01 -2.89125144e-01 -2.91973621e-01
3.93040717e-01 4.87740576e-01 -9.01347995e-01 7.94056952e-01
2.06964239e-01 1.01582623e+00 -7.37409472e-01 8.00813794e-01
-6.41306460e-01 8.26900125e-01 1.10313363e-01 -1.88110113e-01
3.20941895e-01 -1.97362781e-01 8.52368653e-01 1.11161172e+00
-1.12257384e-01 5.09433210e-01 2.22656190e-01 2.40004987e-01
-3.86054546e-01 2.55716443e-01 -1.07827628e+00 -6.91552311e-02
3.63080442e-01 7.81928480e-01 -6.76901281e-01 -7.09333718e-01
-3.28960806e-01 3.30543846e-01 4.27935660e-01 3.66709083e-01
-2.81153858e-01 -4.97071385e-01 5.36053538e-01 5.00608802e-01
2.21468121e-01 -2.63982136e-02 -4.20751795e-02 -1.17234266e+00
2.65586525e-01 -7.64892340e-01 9.19542432e-01 -6.69657767e-01
-1.36154306e+00 4.27406847e-01 4.98694032e-01 -1.01324284e+00
-5.83756745e-01 -7.57205635e-02 -1.04962277e+00 1.91638753e-01
-1.36603224e+00 -5.25769472e-01 4.17647868e-01 6.68426156e-02
5.25500834e-01 -1.39280751e-01 3.52336884e-01 -2.72959210e-02
-4.50419933e-01 1.60403028e-01 -2.62459870e-02 3.87163401e-01
5.86843908e-01 -1.23679757e+00 4.35645610e-01 1.12088752e+00
3.88371378e-01 1.04528654e+00 7.78406382e-01 -1.07417071e+00
-9.50061560e-01 -7.92003036e-01 1.41713417e+00 -5.37354469e-01
8.24437857e-01 1.58644468e-01 -7.89930344e-01 7.62622952e-01
9.23665047e-01 -7.81279624e-01 9.72795129e-01 3.58670592e-01
-6.11180127e-01 1.99500676e-02 -9.99639869e-01 6.69534802e-01
1.33814394e+00 -5.28400064e-01 -1.65205693e+00 5.51812887e-01
3.33297849e-01 -1.06429167e-01 -4.96367693e-01 2.34559327e-01
4.40194756e-01 -7.02958941e-01 5.97909570e-01 -1.28398108e+00
8.40816975e-01 -1.54833719e-01 1.05084740e-01 -1.07546341e+00
-5.17379761e-01 -5.09470522e-01 -4.76613976e-02 1.21670175e+00
7.74191678e-01 -7.30285168e-01 4.22305822e-01 8.01130652e-01
-2.30236009e-01 -5.50337195e-01 -1.18197238e+00 -5.40801466e-01
9.84346718e-02 -2.45016664e-01 3.44257116e-01 1.06186044e+00
5.84443510e-01 7.87333548e-01 4.14139666e-02 -2.53803700e-01
5.84496617e-01 4.67649937e-01 6.81661129e-01 -1.52993000e+00
1.70995072e-01 -6.96014941e-01 -5.08211672e-01 -4.80806112e-01
4.83515918e-01 -1.38098633e+00 -5.81423461e-01 -2.64063525e+00
5.95637441e-01 1.13579690e-01 -7.05574900e-02 6.44441068e-01
-1.41749620e-01 -4.99584764e-01 1.74357146e-01 4.97454077e-01
-7.90761650e-01 1.41011879e-01 8.37831676e-01 -6.21664047e-01
-3.32159735e-02 -3.61131489e-01 -1.35001111e+00 9.27289486e-01
7.23946989e-01 -9.39989805e-01 -5.69800615e-01 -6.15564466e-01
4.19827372e-01 2.08797138e-02 -1.99037269e-01 -6.19633734e-01
5.55128634e-01 -2.74411440e-01 -1.19093500e-01 -7.14193225e-01
-2.87304401e-01 -2.32336372e-01 -1.69530481e-01 5.93778372e-01
-8.89458835e-01 2.01386258e-01 -3.11161913e-02 7.30521381e-01
-3.07612807e-01 -6.35516882e-01 1.75249409e-02 4.70467098e-02
-2.12495893e-01 -1.33558854e-01 -8.93830657e-01 5.70705891e-01
8.03198874e-01 -2.90516913e-01 -9.85402703e-01 -3.46217036e-01
-1.92572817e-01 4.49759960e-01 5.69886744e-01 4.38005775e-01
4.45247054e-01 -9.58391666e-01 -1.18278110e+00 -7.83807218e-01
2.92845130e-01 -4.21587706e-01 -8.05508904e-03 5.40283203e-01
-5.00634789e-01 4.76179123e-01 1.74202010e-01 3.59524265e-02
-1.54359257e+00 9.10905078e-02 -8.93070251e-02 -7.65517056e-01
-7.93848813e-01 -1.54938642e-03 -3.61928463e-01 2.37377957e-02
-2.84335285e-01 -4.98416334e-01 -8.15830886e-01 5.88782370e-01
6.22406006e-01 4.32908773e-01 -5.52572273e-02 -7.99045801e-01
-5.55873096e-01 6.01876318e-01 -6.76057518e-01 -1.23742484e-01
1.57274377e+00 1.81318223e-01 -5.70911288e-01 1.01576470e-01
9.03852344e-01 7.09426939e-01 -4.31549966e-01 -5.78336567e-02
6.53098106e-01 -4.39359546e-01 -5.66146262e-02 -7.26191580e-01
-7.34601498e-01 2.06308335e-01 -4.34724152e-01 5.42666435e-01
7.18691468e-01 3.66652876e-01 3.52821857e-01 9.95168328e-01
2.29600266e-01 -1.31576872e+00 -4.95049208e-01 6.64022446e-01
1.13521957e+00 -7.71329105e-01 7.74690330e-01 -7.13957667e-01
-7.63031900e-01 1.24090528e+00 5.13690710e-02 -2.11275160e-01
3.42227519e-01 1.63502023e-01 1.53502822e-01 -7.78271496e-01
-1.02548933e+00 -2.14428231e-01 4.22409415e-01 1.72089607e-01
5.43682575e-01 -3.20375144e-01 -1.39027107e+00 6.69398010e-02
-3.29732835e-01 -2.50062972e-01 1.09811473e+00 1.23343682e+00
-4.47408438e-01 -1.34261250e+00 -1.51561067e-01 1.18624556e+00
-8.87911737e-01 -3.76470953e-01 -1.15526581e+00 7.57849634e-01
-3.13334197e-01 1.39812779e+00 -1.61995098e-01 7.39957020e-02
5.16676843e-01 5.72220720e-02 3.91162306e-01 -1.01977694e+00
-9.83216465e-01 -1.41403943e-01 1.05429792e+00 -3.61898780e-01
-8.70913804e-01 -1.12201786e+00 -1.46825516e+00 -1.21693499e-01
-6.05429970e-02 6.10754073e-01 2.10490108e-01 1.18285751e+00
6.12642169e-01 5.25476515e-01 4.87735122e-02 1.44652903e-01
-3.49546045e-01 -9.62547839e-01 -3.41718197e-01 6.74662352e-01
1.58252016e-01 -4.50352609e-01 -3.27782243e-01 -9.02882740e-02] | [12.11978530883789, 9.598570823669434] |
5883ad4a-4b8d-4e07-a3b6-4269d0898846 | computer-aided-recognition-and-assessment-of | 2201.11987 | null | https://arxiv.org/abs/2201.11987v2 | https://arxiv.org/pdf/2201.11987v2.pdf | Computer-aided Recognition and Assessment of a Porous Bioelastomer on Ultrasound Images for Regenerative Medicine Applications | Biodegradable elastic scaffolds have attracted more and more attention in the field of soft tissue repair and tissue engineering. These scaffolds made of porous bioelastomers support tissue ingrowth along with their own degradation. It is necessary to develop a computer-aided analyzing method based on ultrasound images to identify the degradation performance of the scaffold, not only to obviate the need to do destructive testing, but also to monitor the scaffold's degradation and tissue ingrowth over time. It is difficult using a single traditional image processing algorithm to extract continuous and accurate contour of a porous bioelastomer. This paper proposes a joint algorithm for the bioelastomer's contour detection and a texture feature extraction method for monitoring the degradation behavior of the bioelastomer. Mean-shift clustering method is used to obtain the bioelastomer's and native tissue's clustering feature information. Then the OTSU image binarization method automatically selects the optimal threshold value to convert the grayscale ultrasound image into a binary image. The Canny edge detector is used to extract the complete bioelastomer's contour. The first-order and second-order statistical features of texture are extracted. The proposed joint algorithm not only achieves the ideal extraction of the bioelastomer's contours in ultrasound images, but also gives valuable feedback of the degradation behavior of the bioelastomer at the implant site based on the changes of texture characteristics and contour area. The preliminary results of this study suggest that the proposed computer-aided image processing techniques have values and potentials in the non-invasive analysis of tissue scaffolds in vivo based on ultrasound images and may help tissue engineers evaluate the tissue scaffold's degradation and cellular ingrowth progress and improve the scaffold designs. | ['Jiao Yu', 'Aliona Dreglea', 'Jia Sun', 'Yanying Zhu', 'Kaixuan Guo', 'Dun Wang'] | 2022-01-28 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 2.35383242e-01 -1.97317719e-01 -3.83700728e-02 3.21257085e-01
-2.80634344e-01 -2.92890131e-01 -1.28898248e-01 5.36489785e-01
-2.18844011e-01 3.80670488e-01 6.39767498e-02 -1.37201026e-01
-3.43019813e-01 -7.74776220e-01 -3.06423992e-01 -1.20113051e+00
-3.36946517e-01 3.75529468e-01 6.62807167e-01 1.58202481e-02
2.72938401e-01 7.29049742e-01 -1.46415806e+00 3.15765589e-01
8.10733259e-01 1.27231610e+00 6.03652477e-01 6.05523467e-01
-3.06868348e-02 5.15227556e-01 -1.63585991e-01 3.04421961e-01
-5.45307389e-03 -2.73463935e-01 -3.41743290e-01 1.58622742e-01
-3.55233997e-01 -3.07128727e-01 1.11825474e-01 1.07002473e+00
4.04392838e-01 -2.42165849e-01 1.02318692e+00 -5.09874284e-01
-2.89213598e-01 -2.39556022e-02 -3.47390175e-01 3.27884257e-01
8.83281305e-02 -4.38619778e-02 8.03755522e-02 -8.26111436e-01
8.41415942e-01 7.85250247e-01 5.17039418e-01 5.54975085e-02
-9.87089753e-01 -3.58959675e-01 -7.69532561e-01 4.88408297e-01
-1.11114383e+00 -1.83904499e-01 8.46537828e-01 -9.34837699e-01
3.90899777e-01 6.70391917e-01 9.44263160e-01 2.35430315e-01
1.11455429e+00 4.29547280e-01 1.45207632e+00 -6.50985062e-01
5.03238857e-01 9.37715471e-02 7.57099539e-02 8.95441949e-01
3.96407038e-01 3.02782327e-01 8.57409388e-02 -5.42392172e-02
1.00367785e+00 -1.11939512e-01 -4.90258455e-01 -2.98781067e-01
-6.55544102e-01 1.80933818e-01 -2.84610074e-02 8.13502669e-01
-7.10628688e-01 -2.27208510e-02 2.95896709e-01 6.12741162e-04
4.20200080e-01 -6.78103091e-03 -8.18746835e-02 -2.52287745e-01
-9.33260918e-01 -3.87938946e-01 8.14004421e-01 2.49608994e-01
3.87270272e-01 -1.46666244e-01 8.29646587e-02 7.73338318e-01
3.37907434e-01 6.22223139e-01 6.30843103e-01 -7.34115303e-01
-5.45865059e-01 5.19689262e-01 -4.53023240e-02 -1.24244010e+00
-1.22580685e-01 -3.01224768e-01 -6.22985065e-01 6.39955878e-01
4.56066281e-01 7.98133090e-02 -1.01019943e+00 7.86106169e-01
6.09940290e-01 -1.20656095e-01 -1.24516703e-01 9.66173470e-01
6.56684995e-01 6.77946031e-01 1.41993593e-02 -6.41955853e-01
1.69295287e+00 -3.75948727e-01 -1.02553964e+00 2.76254982e-01
3.33299637e-01 -8.10286760e-01 7.49630511e-01 2.58720040e-01
-9.07324016e-01 -1.30486384e-01 -1.20881271e+00 6.51740909e-01
-3.74256223e-02 4.11286861e-01 4.69903171e-01 5.94562113e-01
-7.69039571e-01 7.44573772e-01 -1.41278028e+00 -3.68010491e-01
1.60315752e-01 3.71286720e-01 -6.13720298e-01 -1.28304809e-01
-4.42568690e-01 7.89946914e-01 -7.11233355e-03 3.96716744e-01
-4.63292360e-01 -4.23628926e-01 -5.68536639e-01 4.24488857e-02
-1.55913547e-01 -3.70942205e-01 6.51598632e-01 -9.32287455e-01
-1.68252623e+00 7.40696013e-01 -2.25409050e-03 1.67905524e-01
3.79177570e-01 3.42035025e-01 -1.12754732e-01 6.29985929e-01
1.53171286e-01 -3.21181595e-01 3.67643446e-01 -1.26051068e+00
-2.19611272e-01 -5.49876988e-01 -8.86937499e-01 -1.85694486e-01
-3.63244683e-01 1.06196404e-01 -2.62325466e-01 -7.30699301e-01
6.56443715e-01 -8.55938196e-01 1.30617479e-02 1.80400550e-01
-8.58247131e-02 -7.49933571e-02 7.91519880e-01 -9.15663779e-01
9.81766462e-01 -2.25802445e+00 -2.44940937e-01 5.57012022e-01
5.48552051e-02 7.17370734e-02 1.27248853e-01 3.97741139e-01
-4.04841863e-02 -9.82076377e-02 -1.71542406e-01 4.62858379e-01
-6.12191439e-01 1.81968838e-01 6.91919208e-01 9.31044042e-01
-4.45527345e-01 3.92363459e-01 -7.47494817e-01 -8.99316192e-01
-8.70574787e-02 4.35223073e-01 -1.77286312e-01 1.02590825e-02
1.74370468e-01 1.97688028e-01 -5.34724534e-01 1.04384637e+00
6.53829575e-01 2.25474074e-01 1.45985261e-01 -6.99202836e-01
-2.64068902e-01 -5.34869850e-01 -8.36095035e-01 9.20540929e-01
-1.77006293e-02 6.25397742e-01 5.14608562e-01 -8.86699915e-01
1.22711790e+00 6.06152177e-01 1.16049731e+00 -5.87698519e-01
5.68092227e-01 7.59778738e-01 -2.79022623e-02 -1.23732066e+00
-6.11142814e-02 -8.02153349e-01 5.81990957e-01 1.16364965e-02
-2.04009011e-01 3.98406237e-02 1.27275407e-01 -2.82364875e-01
9.04170573e-01 -8.23135898e-02 -3.60723212e-02 -7.07288086e-01
3.88833225e-01 1.44460276e-01 4.91481453e-01 5.28127924e-02
-2.20908239e-01 3.51653785e-01 1.55012444e-01 -3.76082689e-01
-8.76836836e-01 -9.51487362e-01 -3.68807048e-01 4.24712360e-01
4.29837137e-01 4.13983405e-01 -6.28459632e-01 3.83301228e-02
2.10596919e-01 1.46888969e-02 -5.93145072e-01 -1.21406481e-01
-4.03373450e-01 -5.19967139e-01 -1.47052690e-01 1.67818904e-01
2.61869282e-01 -8.98891211e-01 -7.02969134e-01 7.00676143e-01
-6.57269731e-02 -7.67196536e-01 -2.17882991e-01 4.08360623e-02
-1.40803838e+00 -9.29205298e-01 -8.24821830e-01 -1.21773148e+00
7.75472581e-01 9.92131233e-03 2.61750787e-01 1.77375108e-01
-5.48245728e-01 3.90534669e-01 -5.79269946e-01 -2.96787262e-01
-6.85509145e-01 -3.45187604e-01 -1.95698813e-01 3.31652984e-02
-1.03834398e-01 -6.95379972e-01 -9.12602484e-01 6.13851666e-01
-8.92573774e-01 -1.50306940e-01 7.52898455e-01 5.84284067e-01
8.38377953e-01 4.27094012e-01 6.03676438e-01 -5.87032497e-01
5.51227689e-01 -3.54444504e-01 -2.44135261e-01 1.11176185e-01
-7.52314866e-01 -1.86302334e-01 3.95770259e-02 -5.69148004e-01
-9.84289765e-01 -1.92804009e-01 5.33829071e-03 -2.86484748e-01
-6.31839260e-02 1.19602513e+00 1.89251706e-01 -1.67893454e-01
6.52337968e-01 2.15715900e-01 8.24742138e-01 -4.61142659e-01
-4.55289811e-01 9.81199980e-01 7.41490841e-01 -5.51574945e-01
1.64794549e-01 6.11846149e-01 -1.34931296e-01 -1.12864530e+00
1.27439350e-01 -7.19062805e-01 8.73618871e-02 -9.57068622e-01
7.33038127e-01 -3.99498731e-01 -9.26396787e-01 7.03636467e-01
-9.13607240e-01 -3.18007588e-01 -6.81313053e-02 7.83056676e-01
-4.50316876e-01 6.68651879e-01 -1.05606902e+00 -9.54374552e-01
-6.43944323e-01 -1.10856414e+00 4.45942640e-01 2.82178372e-01
-1.62541550e-02 -7.63333678e-01 9.85395089e-02 5.02559245e-01
6.42561138e-01 6.76320791e-01 1.01006424e+00 8.90046805e-02
-3.54126453e-01 -7.24015951e-01 1.36834130e-01 4.95353550e-01
3.93494159e-01 1.83006242e-01 -4.23658669e-01 -1.26610264e-01
4.61733192e-01 3.87799829e-01 4.89854038e-01 9.10912335e-01
6.22546673e-01 -5.28914988e-01 -5.55109501e-01 2.54665822e-01
1.62127769e+00 7.72451758e-01 1.07826388e+00 4.98492479e-01
1.96165103e-03 6.89923763e-01 8.10721874e-01 2.66675383e-01
-1.59532309e-01 3.21611285e-01 5.89802265e-01 -3.79916310e-01
-4.19114113e-01 2.82874741e-02 2.55553752e-01 1.07200301e+00
-5.49262702e-01 -3.72748880e-04 -7.72350430e-01 8.04181576e-01
-1.16086650e+00 -7.13084877e-01 -2.44066834e-01 2.05116892e+00
8.95744503e-01 -1.71869665e-01 -1.87782615e-01 2.42588222e-01
9.33557868e-01 -6.99337184e-01 -1.49491012e-01 -3.26371759e-01
-3.09174191e-02 4.32347119e-01 6.45614684e-01 5.34740388e-01
-5.89404166e-01 4.04325843e-01 5.34037018e+00 8.29419255e-01
-1.47046590e+00 -1.95839927e-01 3.47438216e-01 4.36013103e-01
-3.28407794e-01 9.43948030e-02 1.53772846e-01 6.13126159e-01
3.50062042e-01 -6.38784170e-02 -9.06163305e-02 6.55353844e-01
4.74040091e-01 -9.38674986e-01 -5.41842341e-01 6.81295276e-01
-4.16714758e-01 -1.28058493e+00 -3.81556690e-01 1.60168603e-01
3.08157265e-01 -3.81466269e-01 -3.27370435e-01 -6.48768246e-01
-5.03668308e-01 -4.68484670e-01 8.40292096e-01 8.20589662e-01
8.26689363e-01 -2.00099394e-01 1.26918173e+00 -2.71680998e-04
-9.43996191e-01 3.56351025e-02 -6.51424751e-02 -9.80692729e-02
2.34381765e-01 9.15188730e-01 -1.22927904e+00 2.49971300e-01
6.97498679e-01 1.96155876e-01 -1.91397388e-02 1.66535687e+00
4.30519521e-01 6.75556540e-01 -3.56879115e-01 -2.52447516e-01
-2.78980315e-01 -6.19597495e-01 7.95661390e-01 9.77863252e-01
7.14817941e-01 5.09001553e-01 -1.23610899e-01 5.36040485e-01
6.17808342e-01 6.82906747e-01 -1.50935516e-01 -2.46111587e-01
5.77273905e-01 9.18572128e-01 -1.48405409e+00 -2.45045945e-01
1.35624319e-01 5.56733489e-01 -6.32675827e-01 3.24250311e-01
-2.77223796e-01 -3.03670138e-01 3.00890058e-01 7.71193564e-01
1.10544786e-02 -5.48070133e-01 -6.10450804e-01 -3.91005993e-01
5.56197129e-02 -3.79697293e-01 2.73210078e-01 -7.56986439e-01
-8.61494303e-01 2.92992532e-01 -1.59352005e-01 -1.12165225e+00
5.19479394e-01 -5.60000002e-01 -5.94479918e-01 6.95750177e-01
-8.12834501e-01 -9.13437426e-01 -3.86910766e-01 8.24102908e-02
2.60991335e-01 1.70331150e-01 6.18005753e-01 1.41178310e-01
-3.90812933e-01 3.02240066e-02 3.44512910e-01 -9.65942219e-02
3.79921079e-01 -7.19310820e-01 -1.01264012e+00 5.64231038e-01
-8.95501077e-01 3.89444232e-01 1.20877981e+00 -1.11605096e+00
-1.27067375e+00 -1.85739145e-01 4.46603477e-01 3.95751357e-01
4.28983003e-01 2.17977881e-01 -9.25704122e-01 4.39964719e-02
-6.40193820e-02 -2.05814809e-01 8.35661173e-01 -5.73545337e-01
5.41706026e-01 -3.22217941e-01 -1.31049633e+00 3.67044210e-01
2.63139457e-01 2.13750973e-02 -2.03189060e-01 2.80726194e-01
-1.19861029e-01 -3.96121711e-01 -1.32835746e+00 6.86944783e-01
1.25703645e+00 -6.75495088e-01 5.53565979e-01 -2.99231615e-02
3.89358401e-01 -4.28478628e-01 8.63674358e-02 -5.39751053e-01
-4.27756816e-01 -7.78463483e-02 6.08799934e-01 1.01904309e+00
2.41160557e-01 -7.93171406e-01 9.87148583e-01 4.83386666e-01
-5.60876727e-01 -9.86349821e-01 -1.15143597e+00 -7.06809223e-01
-2.60991782e-01 5.42937256e-02 -3.69187057e-01 7.10903585e-01
3.28556836e-01 -4.70779240e-01 2.85169929e-01 -2.25315318e-02
4.17860091e-01 -1.23970591e-01 1.58824310e-01 -1.26823294e+00
5.54803610e-02 -3.23463589e-01 -9.61923718e-01 -2.28829980e-01
-3.88651460e-01 -7.71492004e-01 4.09851164e-01 -1.91388893e+00
1.77185729e-01 -7.87271023e-01 -1.60706833e-01 2.63912737e-01
2.05731049e-01 3.96062583e-02 -2.89327234e-01 5.37927985e-01
2.48371810e-01 3.54074866e-01 1.70269895e+00 -2.95857161e-01
-3.68618250e-01 -2.55955779e-03 -1.89285845e-01 3.61283332e-01
5.36815286e-01 -3.32598537e-01 -1.35882929e-01 7.29845464e-02
-8.47193450e-02 5.50930858e-01 3.08162451e-01 -1.23306179e+00
3.98337215e-01 -1.97377086e-01 1.13944031e-01 -2.10739359e-01
-8.73292834e-02 -1.15870214e+00 7.85241604e-01 8.97696853e-01
3.15815747e-01 -5.00405490e-01 -1.19632687e-02 4.76270586e-01
-2.25969940e-01 -3.22244257e-01 9.78657067e-01 -3.60264280e-03
-2.61255682e-01 -1.25203416e-01 -1.20189989e+00 -9.17210042e-01
1.23160338e+00 -1.18001032e+00 -1.78240880e-01 -1.55243114e-01
-1.07515609e+00 -3.19633871e-01 7.17890501e-01 -2.31693596e-01
7.80001640e-01 -1.14588857e+00 -4.28282231e-01 1.08090773e-01
-3.07780206e-01 -2.72092909e-01 8.52339864e-01 1.43425369e+00
-1.40876698e+00 -3.22183967e-01 -6.96693718e-01 -8.88848662e-01
-1.36114085e+00 3.74312013e-01 4.41824764e-01 -1.58102930e-01
-4.20859069e-01 5.57820082e-01 -8.06736574e-02 5.57494581e-01
-2.03354686e-01 -4.23897982e-01 -2.55160272e-01 2.03839643e-03
-3.38195078e-02 7.49300957e-01 1.92479849e-01 -4.16329890e-01
-3.81276250e-01 7.92813420e-01 1.79134741e-01 1.42990261e-01
1.32843757e+00 -2.04002798e-01 -5.84831834e-01 3.47045153e-01
1.20922840e+00 2.57253051e-02 -1.01281476e+00 1.82665274e-01
4.63385768e-02 -4.61505741e-01 2.73716450e-01 -8.60080600e-01
-1.18952441e+00 4.70530659e-01 9.87975061e-01 3.77608955e-01
1.45634258e+00 1.74766839e-01 8.27366710e-01 -4.62797850e-01
3.65182668e-01 -1.24559844e+00 2.17332184e-01 -2.01920569e-01
1.10623407e+00 -5.94646573e-01 1.07264541e-01 -1.09741366e+00
-3.22931781e-02 1.53301311e+00 3.47805582e-02 -1.79908931e-01
9.55704808e-01 6.95295215e-01 2.83921808e-01 -4.04745728e-01
-1.40715182e-01 1.14086062e-01 9.71377194e-02 4.64060545e-01
4.04787391e-01 2.59872139e-01 -1.02640224e+00 7.79486060e-01
1.96198195e-01 4.83559489e-01 5.74478209e-01 1.19618618e+00
-1.10293269e+00 -7.40041912e-01 -6.73358738e-01 4.72143859e-01
-6.22165322e-01 5.04934013e-01 1.17264315e-01 5.35580575e-01
4.66214158e-02 7.63977230e-01 -2.33567923e-01 -2.58009315e-01
3.35024655e-01 -7.26344064e-02 5.15123606e-01 -1.50903270e-01
1.81804541e-02 8.15513551e-01 9.27444324e-02 -2.31337637e-01
-4.24709111e-01 -6.59967363e-01 -1.31002569e+00 1.49017558e-01
-5.10488153e-01 3.56612802e-01 1.25110626e+00 8.30125332e-01
9.52371135e-02 5.64235926e-01 4.80596691e-01 -7.56803095e-01
8.62499177e-02 -9.32032943e-01 -1.08431828e+00 1.47518605e-01
1.36513293e-01 -9.29108024e-01 -3.11416656e-01 4.26713496e-01] | [13.674081802368164, -2.7929961681365967] |
e13d9191-d8c9-4142-baf0-f6dd8df1976f | smiles-transformer-pre-trained-molecular | 1911.04738 | null | https://arxiv.org/abs/1911.04738v1 | https://arxiv.org/pdf/1911.04738v1.pdf | SMILES Transformer: Pre-trained Molecular Fingerprint for Low Data Drug Discovery | In drug-discovery-related tasks such as virtual screening, machine learning is emerging as a promising way to predict molecular properties. Conventionally, molecular fingerprints (numerical representations of molecules) are calculated through rule-based algorithms that map molecules to a sparse discrete space. However, these algorithms perform poorly for shallow prediction models or small datasets. To address this issue, we present SMILES Transformer. Inspired by Transformer and pre-trained language models from natural language processing, SMILES Transformer learns molecular fingerprints through unsupervised pre-training of the sequence-to-sequence language model using a huge corpus of SMILES, a text representation system for molecules. We performed benchmarks on 10 datasets against existing fingerprints and graph-based methods and demonstrated the superiority of the proposed algorithms in small-data settings where pre-training facilitated good generalization. Moreover, we define a novel metric to concurrently measure model accuracy and data efficiency. | ['Shoi Shi', 'Shion Honda', 'Hiroki R. Ueda'] | 2019-11-12 | null | null | null | null | ['small-data'] | ['computer-vision'] | [ 7.82507300e-01 -2.15213090e-01 -7.32119441e-01 -3.74683172e-01
-6.60643756e-01 -6.86985254e-01 7.52088547e-01 8.79484892e-01
-1.79362386e-01 9.01882708e-01 2.96716560e-02 -6.17492616e-01
-3.27179670e-01 -1.12890995e+00 -8.49855125e-01 -4.81142104e-01
-2.90226489e-01 4.67873335e-01 -2.33138911e-02 9.62665230e-02
7.48261929e-01 7.57773638e-01 -1.12434971e+00 5.22035599e-01
9.59228277e-01 8.21734309e-01 1.25353038e-02 4.80870962e-01
-4.24023569e-01 8.68394315e-01 -1.69910520e-01 -5.21433949e-01
9.62800086e-02 -3.75380069e-01 -6.60185277e-01 -4.81707662e-01
6.11519694e-01 3.31965417e-01 -4.21073556e-01 1.03236806e+00
2.53590912e-01 1.58715993e-01 9.05776441e-01 -7.89896905e-01
-6.24390960e-01 4.08061892e-01 -2.60215402e-01 -2.62035847e-01
8.11008215e-01 -1.26239911e-01 8.87911618e-01 -1.07601511e+00
9.22357082e-01 1.18048728e+00 7.99818218e-01 2.61504620e-01
-1.69304037e+00 -6.52138889e-01 -2.37667277e-01 7.24688247e-02
-1.41640532e+00 -3.57123137e-01 5.07577419e-01 -7.17517316e-01
9.90219653e-01 3.95933867e-01 2.49171823e-01 9.14502800e-01
4.45277750e-01 4.06945944e-01 9.55729961e-01 -2.34490037e-01
5.49549818e-01 9.40812379e-02 6.13810793e-02 8.44756722e-01
3.67285192e-01 1.76538706e-01 -7.44181156e-01 -9.44125473e-01
4.33893174e-01 5.03859043e-01 -1.46256030e-01 -5.54690003e-01
-1.09866261e+00 1.00385416e+00 3.84834558e-01 3.20606306e-02
-3.54821563e-01 -1.24303132e-01 4.08015221e-01 2.41532311e-01
3.66338581e-01 7.86181867e-01 -3.79530668e-01 5.62032759e-02
-1.06341326e+00 5.23192920e-02 9.52317834e-01 6.76844299e-01
7.32507527e-01 -2.78315485e-01 -1.69493362e-01 7.25229263e-01
2.45682582e-01 3.08098823e-01 4.39303488e-01 -1.96565181e-01
4.44724679e-01 7.88630605e-01 -5.30820154e-02 -1.21397519e+00
-3.19968015e-01 -3.84842306e-01 -9.93507326e-01 -1.75310537e-01
3.83620709e-01 2.72052407e-01 -7.91762948e-01 1.38793111e+00
2.54565090e-01 2.57988751e-01 1.08383477e-01 4.91656005e-01
5.95525920e-01 8.22734177e-01 4.06904250e-01 -2.84932643e-01
8.47581327e-01 -5.00869572e-01 -9.20543447e-02 3.13943624e-01
9.37048733e-01 -7.80447960e-01 9.69709575e-01 7.26606190e-01
-4.90087539e-01 -4.41584051e-01 -9.48927283e-01 1.82740733e-01
-7.46639132e-01 5.90559281e-02 1.02880955e+00 8.34319055e-01
-5.61571956e-01 1.31617868e+00 -5.42446792e-01 -1.66437700e-01
7.41056442e-01 5.95788836e-01 -6.33523226e-01 -2.93128341e-01
-8.95273089e-01 5.35916448e-01 5.55523336e-01 -3.16575885e-01
-6.96829796e-01 -7.76678085e-01 -7.58828342e-01 -2.77809985e-02
1.77624613e-01 -5.67218721e-01 6.82734787e-01 -6.17175043e-01
-1.75033295e+00 5.04025280e-01 -2.36082748e-01 -6.22806668e-01
6.08044863e-02 3.14872712e-02 -4.43260849e-01 -2.45598294e-02
-1.88724726e-01 1.84747681e-01 6.11952305e-01 -8.23535383e-01
-5.57578392e-02 -4.21289623e-01 -1.68348163e-01 -4.32665050e-01
-4.34292644e-01 -2.29155421e-01 5.74738458e-02 -5.23422003e-01
-6.88120499e-02 -7.71689415e-01 -6.45758867e-01 -1.52837604e-01
-7.49338329e-01 -1.66633606e-01 4.14414316e-01 -2.83640772e-01
1.15248573e+00 -1.79920304e+00 7.84437433e-02 8.25718403e-01
4.15989637e-01 3.25473100e-01 -4.72860873e-01 9.66952801e-01
-3.05080831e-01 1.62271053e-01 -3.25798988e-01 1.29679203e-01
-1.20369710e-01 -1.01096161e-01 -5.10935605e-01 4.94727314e-01
3.51117738e-02 7.90001690e-01 -1.20177412e+00 -3.73389482e-01
1.83425725e-01 5.07202864e-01 -5.87594211e-01 1.93150774e-01
-5.58897555e-01 2.69666553e-01 -7.16446757e-01 7.53262639e-01
6.59432530e-01 -5.48873365e-01 7.04086900e-01 -7.59241879e-02
-7.00001270e-02 4.03296024e-01 -7.61785626e-01 1.85322893e+00
-4.64955091e-01 2.55905747e-01 -9.05073881e-01 -1.19200301e+00
1.08671319e+00 1.95684587e-03 5.75084209e-01 -8.28082979e-01
-2.79420048e-01 2.78032243e-01 -1.99006826e-01 -2.62705982e-01
2.60971367e-01 -1.50079712e-01 1.59597412e-01 1.61155954e-01
2.09976863e-02 9.01478156e-03 2.89553195e-01 1.98850274e-01
1.17911148e+00 3.31442282e-02 7.27873564e-01 -1.55112192e-01
5.66412330e-01 1.29533753e-01 1.78402126e-01 7.70989239e-01
6.64025068e-01 8.90402421e-02 4.87838328e-01 -6.33580327e-01
-6.70012891e-01 -9.92579401e-01 -2.84849674e-01 1.14742970e+00
-7.69087970e-02 -9.94553983e-01 -3.44097048e-01 -6.56485915e-01
2.97869444e-01 5.19660234e-01 -5.59041679e-01 -1.75452441e-01
-4.00326699e-02 -8.17319632e-01 5.00148714e-01 8.18695575e-02
-2.02058852e-01 -7.84375966e-01 3.96586545e-02 4.61225927e-01
5.35468459e-01 -7.43567407e-01 -4.23636585e-01 1.97004512e-01
-1.07026315e+00 -1.37335682e+00 -5.80338478e-01 -5.40936470e-01
7.18068123e-01 -8.04310944e-03 9.37475026e-01 -1.16720989e-01
-4.71899688e-01 -1.85474977e-01 -1.57455206e-01 -3.25272113e-01
-5.63736379e-01 1.08956240e-01 1.17243089e-01 1.21270113e-01
2.77207583e-01 -8.26529860e-01 -6.63709641e-01 9.93103683e-02
-9.43227291e-01 6.32758886e-02 8.40600669e-01 9.70545888e-01
8.97537172e-01 -4.08238947e-01 5.59956610e-01 -1.37224412e+00
7.69778550e-01 -5.18045247e-01 -8.11389089e-01 5.25689006e-01
-7.51548529e-01 5.20504892e-01 1.11900127e+00 -5.35115004e-01
-4.25041705e-01 6.53931677e-01 -1.30936757e-01 -3.92146334e-02
-2.26065312e-02 1.05264044e+00 -1.28762305e-01 -6.15616083e-01
8.90867829e-01 5.18896699e-01 -2.56061733e-01 -5.93177319e-01
4.68246192e-01 6.83299780e-01 3.52428079e-01 -7.78609395e-01
5.79818368e-01 1.72272325e-01 5.42840183e-01 -8.44922781e-01
-5.90101063e-01 -5.83448946e-01 -4.90620822e-01 4.14720088e-01
2.64239222e-01 -6.33678675e-01 -1.12211323e+00 -1.21810280e-01
-1.11329818e+00 -6.02758378e-02 6.89331889e-02 5.21954656e-01
-4.37394381e-01 8.06359947e-01 -4.53893334e-01 -6.03845119e-01
-5.82238019e-01 -8.46707821e-01 9.60260570e-01 -9.71156135e-02
-2.60434389e-01 -1.02169490e+00 7.42009878e-01 -7.98025802e-02
2.75801778e-01 6.16132021e-01 1.40014696e+00 -9.73513484e-01
-4.32876229e-01 -4.73999441e-01 -1.97481543e-01 -4.82805967e-02
4.39914495e-01 -1.31123979e-02 -7.96396494e-01 -2.29442507e-01
-6.53919578e-01 -2.95229077e-01 9.50605214e-01 9.65611339e-02
1.66868389e+00 -5.56896746e-01 -7.41369605e-01 6.87806606e-01
1.53399897e+00 2.97524154e-01 6.64374590e-01 9.70612280e-03
6.55983925e-01 3.12794298e-01 3.37336302e-01 6.49891257e-01
-1.63464323e-01 7.13849723e-01 1.48426116e-01 -5.44860810e-02
3.71472985e-01 -9.00544822e-01 2.43175745e-01 5.00384331e-01
-1.26785621e-01 -2.82132357e-01 -8.79292190e-01 -2.58942023e-02
-1.66327333e+00 -1.09438527e+00 -1.14567123e-01 2.70291519e+00
1.10115528e+00 7.94362724e-02 2.55057395e-01 -2.67869681e-02
3.38749290e-01 -1.47306964e-01 -6.35582209e-01 -3.05551261e-01
2.44105533e-02 8.09231699e-01 5.42860031e-01 3.86253357e-01
-8.66512060e-01 9.05699313e-01 6.45031309e+00 1.09901035e+00
-1.47949147e+00 -3.05604905e-01 6.06692731e-01 1.80896312e-01
-4.41412002e-01 2.18376275e-02 -4.62561458e-01 5.48877060e-01
1.27267754e+00 -3.48350734e-01 3.20110053e-01 8.89845192e-01
3.91322933e-02 2.10335895e-01 -1.46998525e+00 1.24541259e+00
-2.30048209e-01 -2.18625021e+00 7.04705179e-01 1.90551415e-01
6.48290217e-01 2.65116692e-02 6.61407933e-02 -5.72729930e-02
1.67567238e-01 -1.59732568e+00 1.97960377e-01 6.40680730e-01
1.03078318e+00 -6.58770084e-01 3.44130695e-01 2.71871030e-01
-1.18676829e+00 1.18378371e-01 -7.76543796e-01 3.90985049e-03
-3.35019022e-01 7.78349280e-01 -1.20330071e+00 8.06578875e-01
-1.58823803e-01 1.05305791e+00 -6.00704432e-01 1.23321104e+00
1.59635350e-01 6.34302437e-01 -2.24697784e-01 -4.09206867e-01
4.49274629e-02 -5.19151390e-01 1.04166456e-02 1.47686028e+00
3.53897005e-01 -3.93361151e-02 2.41821542e-01 6.53543472e-01
-3.50875765e-01 5.29152274e-01 -8.81921828e-01 -6.82806849e-01
3.22726995e-01 1.03596175e+00 -5.17438114e-01 -4.00575578e-01
-3.57574433e-01 7.14692473e-01 2.95249194e-01 3.32690984e-01
-4.84225661e-01 -6.24844074e-01 5.56921661e-01 4.08494711e-01
-9.43393931e-02 -3.01076770e-01 -1.23050950e-01 -1.08499146e+00
-4.11562413e-01 -1.01638782e+00 3.37288260e-01 -2.25283459e-01
-1.32260907e+00 5.75955689e-01 -4.27419692e-01 -1.34143460e+00
-1.79810598e-01 -9.43906009e-01 -4.92158711e-01 7.09245086e-01
-1.39301312e+00 -7.13250935e-01 1.21210150e-01 5.19193709e-01
1.04371317e-01 -2.88366914e-01 1.32133722e+00 2.60691404e-01
-3.96545887e-01 6.41933322e-01 6.60769343e-01 -1.63006008e-01
6.60299242e-01 -1.01534867e+00 5.06124437e-01 1.79314718e-01
6.59827173e-01 1.13159919e+00 4.90644634e-01 -7.54923522e-01
-1.89683294e+00 -1.12233520e+00 7.82363892e-01 -3.91230136e-01
7.50710845e-01 -4.49251324e-01 -8.53488684e-01 1.19242989e-01
-3.95447075e-01 1.03128560e-01 1.28518331e+00 1.66556880e-01
-6.92134202e-01 7.81940520e-02 -9.80515659e-01 3.05093288e-01
1.17565656e+00 -9.37209845e-01 -1.57147929e-01 7.99142897e-01
3.17204297e-01 -2.17432171e-01 -1.16719258e+00 1.57975227e-01
7.88594961e-01 -5.62713206e-01 1.23829126e+00 -1.19316483e+00
4.22407717e-01 -2.10306451e-01 -2.54772782e-01 -1.17339671e+00
-9.20123085e-02 -7.64576018e-01 -1.45484760e-01 5.32619715e-01
6.69469297e-01 -7.72573948e-01 1.01321065e+00 2.04834804e-01
2.77172893e-01 -9.32393789e-01 -7.01176405e-01 -9.11317647e-01
-3.45047265e-02 -1.58999115e-01 6.75978363e-01 9.85091746e-01
5.85749686e-01 3.75780255e-01 -4.21693087e-01 -2.20569968e-02
6.14159346e-01 5.70278645e-01 8.23933721e-01 -1.55994010e+00
-4.67371345e-01 -2.69847095e-01 -7.18581676e-01 -1.08354723e+00
1.86764628e-01 -1.36140394e+00 -3.04183304e-01 -1.17015493e+00
4.83870685e-01 -3.81486923e-01 -5.68807185e-01 4.45442706e-01
1.47632241e-01 3.00344522e-03 -8.86039212e-02 1.64477140e-01
-6.62642419e-01 3.89154255e-01 1.06177735e+00 -5.39232731e-01
-3.27913612e-01 1.84344903e-01 -5.72812438e-01 2.85076320e-01
7.42269278e-01 -7.89699078e-01 -4.79752272e-01 2.37567514e-01
4.34935659e-01 3.18935812e-01 2.08742291e-01 -7.36629844e-01
3.33158761e-01 -6.62706316e-01 4.89466637e-01 -3.13591599e-01
2.97230840e-01 -5.51195860e-01 4.00860727e-01 7.23916888e-01
-5.27784705e-01 -3.61703366e-01 1.37981430e-01 8.11811745e-01
-2.49398142e-01 5.21242805e-02 4.51224595e-01 2.34271616e-01
-4.10968125e-01 7.06621051e-01 -1.01733558e-01 -2.56995350e-01
6.23984694e-01 -2.33798549e-01 -2.41638318e-01 6.28811307e-03
-4.44806129e-01 -3.34761709e-01 5.65775514e-01 1.14861146e-01
9.53420997e-01 -1.10533583e+00 -3.46264869e-01 1.49251938e-01
4.89571989e-01 -4.48039651e-01 -2.08997741e-01 6.27231359e-01
-6.51721478e-01 8.45771849e-01 -1.03928097e-01 -3.43074769e-01
-1.23065603e+00 7.57523179e-01 2.39874929e-01 -3.62568557e-01
-1.87825024e-01 5.64675689e-01 2.10920855e-01 -4.49886292e-01
8.29965323e-02 -4.92645085e-01 3.67931835e-03 -3.36940795e-01
5.75663805e-01 9.18009877e-02 1.61442742e-01 -1.98807999e-01
-5.61209261e-01 6.34301603e-01 -1.07810125e-01 5.15478969e-01
1.51938534e+00 6.74740136e-01 -4.86238077e-02 6.81520067e-03
1.35933590e+00 2.02430680e-01 -7.55129874e-01 -3.48665506e-01
5.90689659e-01 -5.29163897e-01 -2.64576942e-01 -6.28348291e-01
-3.91427726e-01 8.50270391e-01 4.28148150e-01 -5.71853220e-02
6.82607234e-01 -2.22571209e-01 6.55541778e-01 1.06977403e+00
5.16094625e-01 -5.88133752e-01 -6.15532249e-02 3.12063187e-01
6.91266954e-01 -1.08670938e+00 2.79422194e-01 -6.46350503e-01
-1.60975546e-01 1.47113621e+00 -8.15340281e-02 1.75649688e-01
5.78340054e-01 -7.11972117e-02 -4.98235941e-01 -3.57318074e-01
-6.01873159e-01 1.30190223e-01 6.48738325e-01 5.97222388e-01
7.44495332e-01 2.91128308e-01 -1.24174684e-01 5.03326178e-01
-8.58578980e-02 4.62282300e-01 4.15212400e-02 7.35831499e-01
-4.37370121e-01 -1.67006087e+00 -1.01540536e-01 7.26166487e-01
-2.91104496e-01 -4.21000540e-01 -8.36678326e-01 3.52571726e-01
-2.18779579e-01 8.31703126e-01 -5.81374049e-01 -6.38329208e-01
2.50441015e-01 -2.18421817e-02 6.60289109e-01 -7.99984038e-01
-2.70390868e-01 -5.45773029e-01 -2.96071731e-02 -6.28124535e-01
-2.91372418e-01 -1.44164026e-01 -1.19389892e+00 -4.25720483e-01
-1.78837851e-01 6.02218270e-01 6.08681858e-01 7.81034589e-01
6.79321826e-01 -7.51965679e-03 7.70627081e-01 -6.81589603e-01
-6.26064777e-01 -5.47883987e-01 -5.83148599e-01 3.57505143e-01
6.20462969e-02 -3.98532301e-01 1.16788715e-01 9.24671292e-02] | [5.134660720825195, 5.779274940490723] |
36364e1b-075c-43e9-98e5-6aec9bd8a050 | srn-side-output-residual-network-for-object-1 | 1703.02243 | null | http://arxiv.org/abs/1703.02243v2 | http://arxiv.org/pdf/1703.02243v2.pdf | SRN: Side-output Residual Network for Object Symmetry Detection in the Wild | In this paper, we establish a baseline for object symmetry detection in
complex backgrounds by presenting a new benchmark and an end-to-end deep
learning approach, opening up a promising direction for symmetry detection in
the wild. The new benchmark, named Sym-PASCAL, spans challenges including
object diversity, multi-objects, part-invisibility, and various complex
backgrounds that are far beyond those in existing datasets. The proposed
symmetry detection approach, named Side-output Residual Network (SRN),
leverages output Residual Units (RUs) to fit the errors between the object
symmetry groundtruth and the outputs of RUs. By stacking RUs in a
deep-to-shallow manner, SRN exploits the 'flow' of errors among multiple scales
to ease the problems of fitting complex outputs with limited layers,
suppressing the complex backgrounds, and effectively matching object symmetry
of different scales. Experimental results validate both the benchmark and its
challenging aspects related to realworld images, and the state-of-the-art
performance of our symmetry detection approach. The benchmark and the code for
SRN are publicly available at https://github.com/KevinKecc/SRN. | ['Jie Chen', 'Jianbin Jiao', 'Qixiang Ye', 'Wei Ke', 'Guoying Zhao'] | 2017-03-07 | srn-side-output-residual-network-for-object-2 | http://openaccess.thecvf.com/content_cvpr_2017/html/Ke_SRN_Side-output_Residual_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Ke_SRN_Side-output_Residual_CVPR_2017_paper.pdf | cvpr-2017-7 | ['symmetry-detection'] | ['computer-vision'] | [ 2.92424858e-01 -2.62093339e-02 1.81996480e-01 -5.12481153e-01
-7.03049362e-01 -5.30653477e-01 3.19314718e-01 -5.22383690e-01
1.64148256e-01 9.14949998e-02 1.76295027e-01 1.21333294e-01
-1.09495498e-01 -4.58071381e-01 -1.06651914e+00 -5.64769685e-01
2.74795502e-01 3.16940308e-01 7.28087485e-01 -3.57854962e-01
2.74992436e-01 7.79046476e-01 -1.62025154e+00 7.13412166e-01
6.08170748e-01 1.28855658e+00 -1.68970093e-01 4.00330126e-01
2.61679709e-01 6.41735017e-01 -3.96860152e-01 -7.91571379e-01
8.61753166e-01 -1.82078302e-01 -4.38779771e-01 1.23542905e-01
1.22858667e+00 -2.06369966e-01 -5.63150108e-01 1.09790528e+00
7.17980027e-01 -1.36777997e-01 3.91074032e-01 -1.52056897e+00
-8.31484199e-01 2.36558527e-01 -9.59728718e-01 -2.79763825e-02
-1.86297566e-01 2.61803955e-01 9.59105611e-01 -1.05220401e+00
7.67561734e-01 1.37176740e+00 1.01157832e+00 4.34785098e-01
-8.59794497e-01 -7.67688155e-01 2.15124175e-01 3.30143392e-01
-1.32156932e+00 -4.53278154e-01 7.82656372e-01 -4.45145756e-01
7.88668990e-01 2.16384679e-01 2.60813773e-01 1.15915465e+00
3.00029740e-02 1.09959674e+00 8.57037723e-01 7.23097520e-03
-1.87440112e-01 -1.02316573e-01 -3.37904766e-02 4.53189820e-01
3.39201212e-01 -4.96889949e-02 -5.98839402e-01 2.19620973e-01
8.13095450e-01 2.38748699e-01 -3.07809532e-01 -9.49636042e-01
-1.04971802e+00 4.22180474e-01 9.54658031e-01 -1.61400348e-01
5.70376776e-03 -8.61220807e-02 6.64444983e-01 7.36614168e-02
2.29741573e-01 4.08795178e-01 -5.97532272e-01 3.85381758e-01
-7.84375846e-01 5.30999720e-01 3.02614778e-01 1.23859096e+00
4.28717822e-01 -1.03952087e-01 -4.56013709e-01 9.62056279e-01
5.72305396e-02 4.21892315e-01 1.20157994e-01 -6.29806459e-01
8.94868791e-01 9.14064825e-01 -1.33369640e-01 -9.22908008e-01
-6.44321918e-01 -7.93067396e-01 -8.72589052e-01 3.76122445e-01
4.20213997e-01 3.78172994e-01 -9.28914964e-01 1.52112734e+00
5.46590328e-01 1.37315482e-01 -7.31931999e-02 9.55583453e-01
1.15917683e+00 2.62653857e-01 -6.43634200e-01 4.92616624e-01
1.27667391e+00 -1.45760095e+00 -1.23422094e-01 -2.47386634e-01
3.37210387e-01 -1.03940904e+00 1.08262575e+00 2.47711822e-01
-1.07473433e+00 -6.46062255e-01 -8.94327819e-01 -3.66911590e-01
-3.04918140e-01 5.15419900e-01 4.84017074e-01 1.52696013e-01
-1.02935445e+00 5.47785103e-01 -6.64997995e-01 -2.09000915e-01
6.91368878e-01 2.60426968e-01 -4.62718546e-01 2.67298557e-02
-6.54218733e-01 7.61620104e-01 2.06597373e-01 3.35275590e-01
-4.03452903e-01 -9.55562532e-01 -9.53071654e-01 1.26314536e-01
6.12440348e-01 -6.98220909e-01 1.34770954e+00 -8.56242061e-01
-1.18032169e+00 9.63088632e-01 1.86325699e-01 -2.26703703e-01
1.00702465e+00 -3.71474206e-01 -2.94407308e-01 -8.42876732e-02
1.63945258e-01 7.64481664e-01 7.95861602e-01 -1.07180536e+00
-4.21740800e-01 -6.31122708e-01 -1.68585286e-01 1.40495211e-01
1.45787513e-02 3.89827996e-01 -6.77793086e-01 -6.15328729e-01
5.13487697e-01 -8.50833535e-01 1.46661326e-01 4.99129146e-01
-5.52120686e-01 -1.73349559e-01 9.37060297e-01 -6.73269451e-01
7.51722574e-01 -2.33572102e+00 -6.52613938e-02 7.22977985e-03
2.26309270e-01 2.45273843e-01 -3.66271079e-01 1.75616711e-01
-4.47679967e-01 -2.76813090e-01 -1.52234852e-01 -3.49168837e-01
3.55512321e-01 -2.66446978e-01 -2.45906204e-01 5.47541440e-01
5.55872381e-01 1.11994183e+00 -3.77675295e-01 -2.34335512e-01
1.12247644e-02 3.71771544e-01 -7.20361114e-01 8.61418396e-02
4.49761525e-02 2.45830670e-01 1.77858341e-02 1.05792236e+00
1.26922143e+00 -3.72022778e-01 -3.69586170e-01 -4.56371397e-01
-1.00041613e-01 3.29377264e-01 -1.52082419e+00 1.32279003e+00
-4.66916338e-02 6.23049319e-01 8.09810162e-02 -5.98178804e-01
9.39184964e-01 -2.94886380e-01 4.17567044e-01 -1.02878463e+00
-1.21086240e-01 3.48986894e-01 2.20800921e-01 -3.70984018e-01
4.61670697e-01 4.81284380e-01 1.34589029e-02 5.29053994e-02
1.85940236e-01 -1.24116451e-01 3.20727915e-01 -6.63310438e-02
9.06517208e-01 3.83575112e-01 4.02638875e-02 -9.39475000e-02
4.32718992e-01 -5.90778887e-01 9.43778932e-01 6.14501595e-01
-2.66047508e-01 1.22495222e+00 6.62643135e-01 -6.60365164e-01
-1.37068999e+00 -1.31332803e+00 -3.21521670e-01 1.09688950e+00
2.85727888e-01 -3.03236574e-01 -5.60305178e-01 -6.97033048e-01
2.98819304e-01 7.59279132e-02 -7.16123164e-01 -6.37897030e-02
-7.86774516e-01 -4.78975028e-01 5.36311924e-01 8.52684498e-01
8.06305528e-01 -1.00143647e+00 -5.31645298e-01 -2.24888906e-01
-1.28819823e-01 -1.46899927e+00 -6.54475510e-01 -6.55398220e-02
-6.44051194e-01 -1.29396605e+00 -8.46545875e-01 -6.44205272e-01
6.54204786e-01 4.38910097e-01 1.20370138e+00 -1.77882701e-01
-5.95956862e-01 -2.51858663e-02 -8.19311440e-02 -5.47946513e-01
-9.27327573e-02 -2.09249347e-01 -9.76599976e-02 1.33766457e-01
3.02873775e-02 -5.27948081e-01 -7.26635218e-01 7.95703173e-01
-8.05159688e-01 2.06725076e-01 7.52471685e-01 6.95999801e-01
4.94415313e-01 -1.05151601e-01 1.89333931e-01 -4.46997672e-01
-1.93720043e-01 -2.16404926e-02 -1.09017360e+00 3.85618240e-01
-1.15240194e-01 -1.82625782e-02 3.16405863e-01 -3.61843973e-01
-8.95243824e-01 8.62220582e-03 -8.58366340e-02 -7.19179034e-01
1.95149779e-02 -4.04594183e-01 -3.93334329e-01 -4.24292386e-01
6.60970271e-01 1.87193424e-01 -1.46349967e-01 -5.42181134e-01
-1.02807796e-02 5.21019220e-01 7.39216030e-01 -3.23035359e-01
1.00050819e+00 5.41291952e-01 4.28738780e-02 -4.14008230e-01
-9.79850173e-01 -7.21258640e-01 -5.90433300e-01 -6.55044569e-03
5.48126340e-01 -1.09694099e+00 -5.44614255e-01 9.93014216e-01
-1.17509425e+00 -1.92614093e-01 -3.46778423e-01 2.75065273e-01
-5.57177484e-01 4.70532775e-01 -4.34014887e-01 -4.79181528e-01
-4.05668706e-01 -1.07851875e+00 1.53600287e+00 5.62783718e-01
1.32220492e-01 -1.83884665e-01 -2.72429824e-01 5.50787508e-01
4.65861529e-01 2.24516347e-01 5.58487177e-01 -7.20184267e-01
-9.44480181e-01 -1.46877736e-01 -8.52034867e-01 4.41163003e-01
-1.68783665e-01 2.27519527e-01 -1.00624049e+00 -2.24994332e-01
-1.24489248e-01 -5.33635497e-01 1.02007878e+00 1.41776770e-01
1.30643904e+00 -1.73701048e-01 1.57370538e-01 1.14202869e+00
1.18285847e+00 -4.84704494e-01 7.80912519e-01 6.63657963e-01
9.38120246e-01 6.10837221e-01 4.76770043e-01 3.12263876e-01
4.43268120e-01 8.71890008e-01 5.79495609e-01 -4.97332156e-01
-3.82403582e-01 -5.18557988e-02 3.35642755e-01 1.48019180e-01
-1.26687318e-01 -2.30771750e-02 -8.05087626e-01 1.93616316e-01
-1.92121482e+00 -9.73309159e-01 -2.36225605e-01 2.16714621e+00
4.33946371e-01 3.56584102e-01 2.78130502e-01 2.81048995e-02
8.24375331e-01 1.36598125e-01 -7.31190741e-01 -1.98670164e-01
-4.18763310e-01 -9.36863795e-02 6.51586115e-01 -9.74700004e-02
-1.46664977e+00 7.32976556e-01 5.38270140e+00 6.33632958e-01
-1.16012108e+00 -1.34368837e-01 4.14999694e-01 -1.14585854e-01
1.46398142e-01 -2.92108923e-01 -1.19505167e+00 4.32242781e-01
2.16289803e-01 2.88500994e-01 1.57657191e-01 1.19428194e+00
-2.04589460e-02 2.32144728e-01 -1.09988582e+00 1.15690720e+00
4.00732785e-01 -1.32088697e+00 -1.32896081e-01 -2.40564689e-01
9.54302728e-01 4.49980885e-01 2.02553347e-01 4.17020589e-01
-1.44477651e-01 -6.88020229e-01 1.05675375e+00 1.82332233e-01
6.07236445e-01 -4.14416224e-01 8.44238162e-01 1.27094045e-01
-1.27348185e+00 -4.07371908e-01 -4.92643803e-01 2.68048167e-01
-1.44557029e-01 3.81141394e-01 -7.73855388e-01 6.62492454e-01
1.22220731e+00 6.77732468e-01 -1.14158487e+00 1.48506713e+00
-1.87225848e-01 -6.67292625e-02 -5.05974531e-01 2.17297405e-01
7.19302669e-02 -4.74078115e-03 5.91696322e-01 1.38336742e+00
2.38575846e-01 -4.66850996e-01 3.87580469e-02 1.10605025e+00
-1.77952588e-01 3.44611406e-02 -3.51281792e-01 2.39348710e-01
1.05047211e-01 1.21926546e+00 -8.02006841e-01 -1.17429160e-01
-3.95732880e-01 9.75045085e-01 3.72709215e-01 2.94890821e-01
-1.05419374e+00 -3.12286437e-01 6.06232405e-01 2.64662266e-01
5.13370872e-01 1.71832979e-01 -4.07597959e-01 -1.41005814e+00
7.98976541e-01 -1.03140509e+00 4.44698364e-01 -8.63603592e-01
-1.44184971e+00 3.08477461e-01 8.09299387e-03 -1.50447071e+00
2.70705760e-01 -9.91472006e-01 -8.82139981e-01 7.03446269e-01
-1.49183214e+00 -1.64880204e+00 -5.98804057e-01 4.79287356e-01
5.57956100e-01 -2.38649085e-01 2.22830653e-01 4.25096989e-01
-5.85259199e-01 9.13236141e-01 1.42354332e-02 3.83384049e-01
9.43034708e-01 -1.14977646e+00 9.03807104e-01 1.09832871e+00
-1.36839822e-01 3.04414600e-01 4.58097368e-01 -3.80806059e-01
-1.13703001e+00 -1.12614071e+00 8.35442662e-01 -4.99466270e-01
5.78828931e-01 -8.71379435e-01 -8.76569092e-01 5.86893916e-01
-2.78673440e-01 4.96454448e-01 2.01723129e-01 -1.76557694e-02
-9.26906526e-01 -2.60498405e-01 -8.53526354e-01 6.51341558e-01
1.56273341e+00 -3.43372077e-01 -3.66488457e-01 5.86546659e-01
6.51729524e-01 -9.90677774e-01 -3.81178260e-01 1.22797167e+00
5.66704810e-01 -1.44202471e+00 1.04902864e+00 -6.37106717e-01
6.63379312e-01 -5.00203907e-01 -1.10403128e-01 -7.76670694e-01
-5.06588221e-01 -5.20808101e-01 9.31768641e-02 1.17266071e+00
3.36620599e-01 -6.02788985e-01 9.27589238e-01 3.19392234e-01
-5.24501979e-01 -8.88022840e-01 -8.73847902e-01 -1.07595968e+00
-1.36137277e-01 -9.97859165e-02 6.51309073e-01 7.42622733e-01
-9.29003179e-01 6.23848364e-02 -2.00595170e-01 4.79027182e-01
6.71740234e-01 4.83405739e-01 1.21828949e+00 -1.06198204e+00
-5.49001515e-01 -8.20989609e-01 -1.06792140e+00 -9.18786108e-01
-9.76480097e-02 -9.22639608e-01 -8.49551521e-04 -1.34662557e+00
3.57260108e-01 -5.24287708e-02 -1.83085337e-01 4.74397928e-01
-2.11076215e-01 5.12490809e-01 5.15232980e-01 3.86282988e-02
-8.80706131e-01 6.27185106e-01 1.41297507e+00 -2.65511721e-01
2.14903932e-02 4.51756746e-01 -7.47646809e-01 9.27203536e-01
6.71519816e-01 -3.41826588e-01 -1.08211614e-01 -3.92894268e-01
3.43167990e-01 -6.30841434e-01 7.27901340e-01 -1.08983469e+00
1.68097898e-01 1.70783773e-01 7.99017370e-01 -1.03996265e+00
2.66591251e-01 -6.34016037e-01 -8.59274641e-02 2.53123134e-01
-7.05744699e-02 1.62245885e-01 2.95887321e-01 1.95505291e-01
-1.41701862e-01 7.85410553e-02 1.08684897e+00 1.03039540e-01
-6.05331957e-01 4.29127008e-01 5.56771517e-01 3.29783857e-01
8.64191532e-01 -4.95979071e-01 -5.41820228e-01 -4.26694974e-02
-5.30391097e-01 5.27222395e-01 7.62097061e-01 7.37486482e-01
5.56842029e-01 -1.21827126e+00 -1.03218842e+00 6.06690466e-01
5.05140305e-01 3.43997151e-01 5.57478130e-01 9.63632762e-01
-6.84571803e-01 1.23183012e-01 -4.52626914e-01 -1.01841199e+00
-1.35802495e+00 4.58862811e-01 6.52314663e-01 -2.03025877e-01
-8.65703642e-01 1.23869693e+00 6.81903601e-01 -9.87005949e-01
6.91620767e-01 -5.11529684e-01 4.31904346e-02 -3.41898173e-01
4.67968911e-01 4.59967732e-01 4.62792724e-01 -5.09363353e-01
-4.90103960e-01 8.25785339e-01 -7.62105957e-02 3.65754217e-01
1.33607936e+00 2.44028434e-01 -6.06893562e-02 7.72986487e-02
1.26364338e+00 1.34963900e-01 -1.60781384e+00 -3.89926374e-01
-1.20673396e-01 -6.08001828e-01 -5.79013348e-01 -7.20457137e-01
-9.97637212e-01 7.53263712e-01 6.58112049e-01 -1.92144319e-01
1.11132860e+00 2.09595133e-02 5.56933463e-01 5.36791623e-01
-7.66154304e-02 -1.16840374e+00 3.05598289e-01 7.20610261e-01
1.42255616e+00 -1.50346768e+00 2.77446732e-02 -6.55263722e-01
-6.79146647e-01 1.23555505e+00 1.09466362e+00 -2.41000712e-01
3.67383182e-01 4.47511584e-01 1.69945180e-01 -3.68779413e-02
-5.15971363e-01 -1.18551165e-01 6.62156582e-01 3.49688798e-01
1.25872895e-01 -2.41854683e-01 2.00917318e-01 4.74227428e-01
-3.55028063e-01 -3.31663877e-01 2.72613913e-01 6.55640483e-01
-1.33262917e-01 -8.03811431e-01 -5.36956549e-01 2.44949147e-01
-3.49597633e-01 7.13827312e-02 -8.11891198e-01 7.74431586e-01
3.52931917e-01 3.09018612e-01 1.06879525e-01 1.85122609e-03
7.95374632e-01 -1.30480021e-01 5.81243038e-01 -4.74062949e-01
-5.57282567e-01 2.36975849e-02 -1.01805694e-01 -9.53843832e-01
-7.13270009e-02 -6.53540492e-01 -1.03999317e+00 5.46467490e-02
-2.97469586e-01 -4.77657408e-01 6.34366930e-01 5.10744095e-01
5.10976970e-01 6.06180608e-01 7.64338791e-01 -1.28257656e+00
-1.07347989e+00 -1.01500618e+00 -4.11536425e-01 6.67478442e-01
3.49990785e-01 -6.42701089e-01 -3.55939269e-01 -3.51100564e-01] | [8.556533813476562, -1.7436587810516357] |
300547c9-fe33-4a1c-84e8-82f11e7c9e8b | texture-generation-using-graph-generative | 2206.08547 | null | https://arxiv.org/abs/2206.08547v2 | https://arxiv.org/pdf/2206.08547v2.pdf | Texture Generation Using A Graph Generative Adversarial Network And Differentiable Rendering | Novel photo-realistic texture synthesis is an important task for generating novel scenes, including asset generation for 3D simulations. However, to date, these methods predominantly generate textured objects in 2D space. If we rely on 2D object generation, then we need to make a computationally expensive forward pass each time we change the camera viewpoint or lighting. Recent work that can generate textures in 3D requires 3D component segmentation that is expensive to acquire. In this work, we present a novel conditional generative architecture that we call a graph generative adversarial network (GGAN) that can generate textures in 3D by learning object component information in an unsupervised way. In this framework, we do not need an expensive forward pass whenever the camera viewpoint or lighting changes, and we do not need expensive 3D part information for training, yet the model can generalize to unseen 3D meshes and generate appropriate novel 3D textures. We compare this approach against state-of-the-art texture generation methods and demonstrate that the GGAN obtains significantly better texture generation quality (according to Frechet inception distance). We release our model source code as open source. | ['Bradley Walls', 'Clayton T. Morrison', 'Dharma KC'] | 2022-06-17 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 5.21312118e-01 2.33582139e-01 4.98142898e-01 6.72265813e-02
-6.50272310e-01 -7.04217315e-01 7.89266229e-01 -3.35809827e-01
3.63184176e-02 7.91591644e-01 -3.79377246e-01 -2.69899547e-01
2.55471528e-01 -1.36111271e+00 -1.18806350e+00 -8.63799453e-01
2.71490335e-01 8.37624252e-01 3.68367821e-01 -2.35361293e-01
5.52770775e-03 8.88062775e-01 -1.71961129e+00 2.85326988e-02
7.50106931e-01 8.11563551e-01 5.12983575e-02 9.06287372e-01
-7.46385977e-02 3.11675459e-01 -5.98085821e-01 -4.95275944e-01
4.44379598e-01 -8.03588748e-01 -5.35373092e-01 5.41334569e-01
4.73884791e-01 -2.42022738e-01 -4.11976911e-02 9.17830110e-01
7.13125110e-01 1.24892153e-01 9.74509060e-01 -1.11154842e+00
-6.91729426e-01 1.92711934e-01 -4.67545450e-01 -5.73338628e-01
1.86130062e-01 2.95132399e-01 2.58245230e-01 -9.24876332e-01
9.66512620e-01 1.30390322e+00 6.58045411e-01 8.10557425e-01
-1.60638547e+00 -2.98575640e-01 -7.66416192e-02 -2.77468145e-01
-1.27228069e+00 -1.99322894e-01 1.17838573e+00 -3.82252336e-01
4.25195634e-01 3.72279257e-01 8.83853674e-01 1.50379157e+00
3.33540887e-01 5.72062671e-01 1.38086116e+00 -5.67011118e-01
4.75160867e-01 -6.15406856e-02 -7.70791173e-01 5.91809809e-01
1.51036918e-01 1.75664723e-01 -2.19485015e-01 8.67882296e-02
1.33584833e+00 -1.92489490e-01 -1.12067915e-01 -6.53322041e-01
-1.20195270e+00 7.54211009e-01 1.05714619e-01 -9.97558162e-02
-3.49889308e-01 4.37833369e-01 -8.06924794e-03 3.02765161e-01
9.78469968e-01 3.75356078e-01 5.43248728e-02 -4.97034155e-02
-8.15405488e-01 5.72150528e-01 8.02714884e-01 1.10150445e+00
8.82394314e-01 3.66936237e-01 -1.40351236e-01 7.32794464e-01
8.38696063e-02 9.54273045e-01 3.37994844e-02 -8.04083645e-01
-3.78147326e-02 2.62201488e-01 1.22172244e-01 -9.97288525e-01
-2.20895279e-02 -1.41903877e-01 -1.01777565e+00 6.87915683e-01
3.24665636e-01 -2.53893942e-01 -1.32052565e+00 1.29960454e+00
7.66246378e-01 1.14398018e-01 -7.91599676e-02 7.83579230e-01
7.65536606e-01 5.95712662e-01 -3.90718490e-01 -1.44077130e-02
9.58984554e-01 -7.19269454e-01 -5.88986218e-01 8.70014131e-02
9.16485190e-02 -1.13372076e+00 1.16489863e+00 3.17902535e-01
-1.34380817e+00 -6.34030163e-01 -9.78467941e-01 1.67030990e-01
-2.44924322e-01 -2.94284374e-01 5.86570024e-01 7.05670953e-01
-9.77586985e-01 6.41873598e-01 -9.54058051e-01 -2.88798749e-01
4.69285369e-01 4.47089523e-02 -2.84301668e-01 -3.91122326e-02
-9.63047802e-01 8.23674917e-01 2.25625664e-01 -4.41314951e-02
-1.20149779e+00 -5.11070967e-01 -9.69512403e-01 -2.67222941e-01
4.37389374e-01 -1.10467386e+00 1.06923962e+00 -8.56793106e-01
-2.16369390e+00 9.38910663e-01 7.29558989e-02 -2.69762248e-01
9.68172967e-01 2.58997660e-02 1.49790803e-02 1.17130145e-01
-1.13503464e-01 7.46385753e-01 1.23673022e+00 -1.72845113e+00
-3.83135304e-02 7.42942989e-02 6.84855655e-02 1.08362116e-01
2.33376324e-01 -4.97684658e-01 -4.86285985e-01 -1.04529309e+00
1.35215417e-01 -1.15950274e+00 -3.56568843e-01 1.05353750e-01
-5.41666925e-01 1.97991878e-02 9.01291251e-01 -2.79006362e-01
4.85593796e-01 -1.90482640e+00 2.62005001e-01 3.05420220e-01
5.29087074e-02 -1.49059920e-02 -5.68424836e-02 5.23365319e-01
1.16155036e-01 3.80244970e-01 -4.75306332e-01 -5.03320575e-01
2.43931517e-01 1.97525293e-01 -2.98687637e-01 2.01865539e-01
5.00121474e-01 9.92893040e-01 -8.16534519e-01 -3.62513185e-01
4.76702511e-01 5.76417983e-01 -9.14351165e-01 2.75145411e-01
-6.70598447e-01 9.08810496e-01 -4.40476596e-01 6.19042337e-01
1.00870335e+00 -1.95251316e-01 -1.32732779e-01 -8.35522115e-02
8.33384097e-02 -1.73651069e-01 -1.12843215e+00 1.82272339e+00
-5.09953141e-01 3.20148379e-01 -3.84913087e-01 -8.28287601e-01
1.21608055e+00 1.81326643e-01 1.59258574e-01 -5.02209425e-01
1.14302896e-01 3.88640910e-01 -3.09129715e-01 -2.27585435e-01
4.41337645e-01 -3.69953930e-01 -1.67613924e-01 5.01995921e-01
-7.46222883e-02 -1.27135921e+00 2.48681027e-02 -1.52924344e-01
8.93331170e-01 6.23705626e-01 -2.47227103e-01 -4.61715907e-02
2.16485620e-01 4.82763797e-02 2.58969635e-01 8.44176054e-01
6.47384882e-01 1.17722321e+00 5.65758944e-01 -5.03695130e-01
-1.52136970e+00 -1.29922497e+00 1.63555099e-03 1.61681488e-01
1.59513265e-01 2.30583586e-02 -9.61517751e-01 -6.66401505e-01
-1.55149296e-01 6.18227661e-01 -7.65454888e-01 -4.12370376e-02
-6.29230142e-01 -6.55298173e-01 3.16523254e-01 2.26670519e-01
5.90759099e-01 -1.20660007e+00 -4.49404180e-01 2.69537538e-01
1.98689729e-01 -1.02775013e+00 -2.42379561e-01 -8.97371992e-02
-7.41477370e-01 -8.66980076e-01 -1.09318304e+00 -6.99466646e-01
9.26182449e-01 -1.67337075e-01 1.22074866e+00 -9.00517181e-02
-2.71552831e-01 4.65778142e-01 -4.66643840e-01 -5.47435701e-01
-7.91801214e-01 1.86832435e-02 -2.37208486e-01 1.96820393e-01
-3.50796193e-01 -7.88362503e-01 -5.83370328e-01 3.86931241e-01
-1.20710588e+00 6.55732095e-01 3.61729383e-01 9.04280841e-01
1.13909936e+00 3.03587884e-01 3.89655381e-01 -1.17459345e+00
2.99985617e-01 -1.75946102e-01 -7.99583673e-01 -4.86835465e-02
-2.33333930e-02 6.98179901e-02 6.39205098e-01 -5.79699397e-01
-1.12773681e+00 9.21920091e-02 -3.09384465e-01 -8.00436497e-01
-3.45290601e-01 2.30218366e-01 -1.68261558e-01 -1.98903307e-01
6.15063429e-01 3.50078791e-01 3.60796340e-02 -5.53139985e-01
4.05113339e-01 1.24147795e-01 2.28941068e-01 -7.85992324e-01
1.35140944e+00 7.32501507e-01 4.20868188e-01 -7.05476284e-01
-5.36288440e-01 5.14633060e-01 -6.63677752e-01 -3.57693762e-01
1.01756036e+00 -6.14689112e-01 -6.31502867e-01 8.62719834e-01
-1.07797492e+00 -8.73536468e-01 -6.04832113e-01 1.41127631e-01
-1.05263710e+00 1.63280547e-01 -2.73288697e-01 -6.21780634e-01
-2.17420250e-01 -1.23099649e+00 1.36083949e+00 1.59884132e-02
-1.31963134e-01 -1.03161252e+00 -6.79330081e-02 3.09295729e-02
5.31767488e-01 1.17256308e+00 1.02019453e+00 7.37992451e-02
-9.31083024e-01 -9.74073112e-02 1.36841629e-02 3.32521558e-01
3.22432727e-01 4.88882884e-02 -7.35935569e-01 -1.25922784e-01
-1.33518353e-01 -1.25199348e-01 5.47518909e-01 1.91999421e-01
1.36226022e+00 -2.25594789e-01 -1.08674467e-01 7.88628936e-01
1.36506152e+00 2.29609743e-01 8.30472946e-01 1.28187135e-01
9.55100894e-01 4.62061584e-01 4.05884504e-01 2.78309911e-01
1.78977281e-01 6.84722245e-01 4.15005773e-01 -4.45476323e-01
-5.24724722e-01 -4.86293375e-01 3.87647897e-02 6.62291229e-01
-4.37166929e-01 -5.79423726e-01 -7.81861782e-01 4.07822698e-01
-1.64824772e+00 -8.46851230e-01 -1.06307425e-01 2.12505436e+00
8.22179258e-01 1.95222721e-01 -1.98666796e-01 1.45476282e-01
7.08555937e-01 -5.08050807e-03 -5.06006837e-01 -4.31251913e-01
-2.61116117e-01 8.30953360e-01 3.03371161e-01 4.97634321e-01
-8.37455809e-01 1.19438434e+00 5.81747341e+00 1.00355434e+00
-1.17370856e+00 -1.21852733e-01 7.04558909e-01 -7.28772208e-02
-8.02418113e-01 3.52114141e-02 -4.42228228e-01 4.46910113e-01
4.39593881e-01 -4.37628962e-02 3.01337004e-01 8.07310462e-01
2.04036549e-01 -1.46571532e-01 -1.02582550e+00 1.06450856e+00
3.31035405e-01 -1.47804117e+00 5.17696798e-01 1.06825717e-01
1.08668697e+00 -5.45871496e-01 1.01737946e-01 -9.96268243e-02
3.88122022e-01 -1.09955585e+00 9.37468410e-01 8.54029715e-01
1.05320656e+00 -9.32385683e-01 4.31510597e-01 1.22484587e-01
-8.39032948e-01 7.68473744e-01 -2.91006088e-01 2.13326558e-01
4.39979762e-01 8.24454069e-01 -7.03294337e-01 6.55988514e-01
6.29165769e-01 5.76363385e-01 -4.27588820e-01 9.09151852e-01
-4.45941180e-01 4.34779942e-01 -2.96241105e-01 2.81682666e-02
5.50690852e-02 -2.75964737e-01 7.80740917e-01 7.18117654e-01
7.81509936e-01 1.77426822e-02 1.09678499e-01 1.14643645e+00
-1.40370071e-01 -1.70445740e-01 -8.26677382e-01 1.31740034e-01
1.31956786e-01 8.35467875e-01 -1.01714790e+00 -3.18766087e-01
7.26283044e-02 1.43697095e+00 -1.18372835e-01 4.11956519e-01
-9.73872542e-01 -4.63691831e-01 3.63550007e-01 3.91682595e-01
4.91908580e-01 -3.34343493e-01 -2.00387165e-01 -1.14701295e+00
1.42732384e-02 -8.04902613e-01 -3.63361120e-01 -1.13311660e+00
-1.44925916e+00 7.68844068e-01 -1.56909488e-02 -1.37972355e+00
-2.87173808e-01 -4.33627218e-01 -3.48448396e-01 8.70044947e-01
-1.35064542e+00 -1.42242789e+00 -4.19931114e-01 6.32494450e-01
5.30419052e-01 8.73297825e-02 8.87447774e-01 1.78608611e-01
-1.38028368e-01 4.76269215e-01 -7.33811110e-02 -8.54930431e-02
7.42630780e-01 -1.14257777e+00 8.10083210e-01 7.35079288e-01
-2.70606726e-02 1.18396312e-01 7.19492435e-01 -7.73386657e-01
-1.42637479e+00 -1.22371757e+00 5.69546759e-01 -5.68808734e-01
2.86206901e-01 -7.88635790e-01 -7.22611308e-01 5.43821454e-01
1.31392077e-01 1.13888392e-02 2.82697797e-01 -4.13936824e-01
1.31085694e-01 9.71350148e-02 -1.18041849e+00 9.29013193e-01
1.43656611e+00 -2.19414115e-01 -2.07215831e-01 3.29077929e-01
7.73869395e-01 -9.25044298e-01 -9.17202711e-01 6.09439611e-01
3.64920706e-01 -9.04878020e-01 7.95468152e-01 -2.26338193e-01
6.11170948e-01 -5.91650486e-01 4.11308073e-02 -1.45691061e+00
5.17258747e-03 -9.47233677e-01 8.44808221e-02 1.00967395e+00
3.63171220e-01 -7.09335446e-01 8.41502130e-01 2.60279417e-01
-3.53078604e-01 -6.86741531e-01 -7.01413631e-01 -8.73357356e-01
1.39862090e-01 -4.24909353e-01 9.43943739e-01 8.02072167e-01
-8.73161733e-01 -1.33058075e-02 -5.22769988e-01 8.82048607e-02
6.97970927e-01 4.60284859e-01 1.32330775e+00 -1.07877028e+00
-3.11954886e-01 -2.04539463e-01 -4.31946844e-01 -9.79764342e-01
1.96931695e-04 -7.32518435e-01 3.10022496e-02 -1.53727746e+00
-8.55122954e-02 -6.83862746e-01 3.61820966e-01 3.47506821e-01
-9.18574855e-02 6.59912527e-01 8.78753811e-02 -7.27021471e-02
-8.50756094e-02 9.09692705e-01 1.88445163e+00 -1.79570228e-01
-1.19613416e-01 -4.43392843e-02 -3.12339008e-01 6.80569887e-01
5.90985775e-01 -5.12252569e-01 -5.43459356e-01 -5.66463411e-01
2.50263870e-01 -6.82940707e-02 7.49513865e-01 -1.08937430e+00
-2.61514634e-01 -2.90209949e-01 6.64308071e-01 -5.80527365e-01
3.94829988e-01 -8.04461360e-01 7.46064782e-01 1.60118863e-01
1.94082469e-01 -1.15912020e-01 3.48660469e-01 3.93646330e-01
-8.46021473e-02 3.42473164e-02 8.90525639e-01 -3.06115538e-01
-2.07740188e-01 7.04620838e-01 -4.65305775e-01 -6.58814162e-02
1.08377242e+00 -3.69077027e-01 4.40045667e-04 -3.44406396e-01
-9.64908063e-01 -4.44201350e-01 1.04698873e+00 3.95448953e-01
7.13001668e-01 -1.70173049e+00 -7.57179320e-01 4.06406611e-01
-1.54611155e-01 4.60486591e-01 5.24895906e-01 3.19976360e-01
-1.01585913e+00 -1.90091193e-01 -2.48218641e-01 -7.81213582e-01
-7.66804159e-01 4.65409786e-01 3.09013754e-01 -1.66046187e-01
-7.01912642e-01 7.20451593e-01 4.28185642e-01 -4.74122763e-01
-3.55136484e-01 -9.73650962e-02 3.42951596e-01 -1.58427343e-01
5.52721508e-02 -2.05276329e-02 1.62938818e-01 -2.54269987e-01
9.40116867e-02 9.57179546e-01 2.77084470e-01 -2.46125907e-01
1.31122279e+00 1.27085194e-01 2.93729678e-02 3.46950144e-01
8.88648987e-01 1.57909527e-01 -1.46730387e+00 1.78842232e-01
-7.31373370e-01 -6.87553823e-01 -1.98114291e-01 -4.71582860e-01
-1.19815719e+00 8.49563599e-01 5.73619425e-01 2.24356070e-01
1.11823297e+00 -6.23608604e-02 8.22830081e-01 1.48697868e-02
6.21332884e-01 -8.35719347e-01 2.53034949e-01 5.65090418e-01
1.17995930e+00 -1.00993061e+00 -1.50917113e-01 -6.82250142e-01
-3.92142624e-01 9.71658289e-01 4.96198535e-01 -5.06246567e-01
7.53831506e-01 2.07622856e-01 1.04362005e-02 -2.16241673e-01
-4.98387903e-01 -1.05081156e-01 2.66648382e-01 8.41843545e-01
1.14534952e-01 -2.41552480e-02 -1.11449786e-01 -2.01608259e-02
-4.52548951e-01 -2.71584094e-01 5.83995283e-01 8.05315793e-01
1.73373997e-01 -1.43025362e+00 -3.49608153e-01 6.49472550e-02
-2.53023714e-01 -1.54897094e-01 -2.35028848e-01 9.00759220e-01
2.46837273e-01 5.58909595e-01 2.29484946e-01 -1.66880563e-01
5.05344331e-01 3.33973058e-02 9.06623244e-01 -6.78864658e-01
-1.38325736e-01 2.15996370e-01 9.57679190e-03 -3.87340426e-01
-5.17517328e-01 -3.82110715e-01 -9.23549056e-01 -5.44004679e-01
-2.16408208e-01 -8.48724023e-02 6.19452536e-01 5.35250843e-01
4.65620697e-01 6.41670227e-01 7.62956142e-01 -1.31234264e+00
2.75874496e-01 -7.39988208e-01 -4.08360183e-01 5.05352974e-01
5.84961809e-02 -8.79453480e-01 -3.47831398e-01 5.01432836e-01] | [9.148998260498047, -3.461620807647705] |
dd6be1d3-ec4c-4ba9-adc9-2a1afe9a8d8d | minimum-norm-method-for-linear-and-planar | 2106.03666 | null | https://arxiv.org/abs/2106.03666v1 | https://arxiv.org/pdf/2106.03666v1.pdf | Minimum Norm Method for Linear and Planar Sparse Arrays | Coprime and nested arrays are sparse arrays with enhanced degrees of freedom, which can be exploited in direction of arrival estimation using algorithms such as product processing, min processing, and MUSIC. This paper applies the minimum norm method for direction of arrival estimation. Comparison of the root mean squared errors and probabilities of resolution of the minimum norm method with MUSIC for a given linear coprime or nested array demonstrates the superiority of the minimum norm method. Specifically, minimum norm method exhibits lower mean squared error, narrower peaks at the locations of the true sources, and a lower noise floor in the spatial spectral estimate. This work also formulates two different minimum norm methods for planar sparse arrays: direct and linear. Comparison of the linear minimum norm method with the linear MUSIC for planar arrays also demonstrates higher accuracy of the minimum norm method. | ['Kaushallya Adhikari', 'Tyler M. Trosclair'] | 2021-06-07 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [ 4.51594472e-01 -4.65934277e-01 2.90430546e-01 1.25447229e-01
-8.60783815e-01 -8.09194803e-01 7.94130266e-02 3.02990172e-02
-3.15347314e-02 7.07698822e-01 4.75251079e-01 -3.92460674e-02
-9.94071960e-01 -4.74288523e-01 -3.52445513e-01 -1.02590942e+00
-5.54141521e-01 -1.47862479e-01 -2.35046998e-01 -1.40952803e-02
2.21049353e-01 6.51543677e-01 -1.24194968e+00 7.09428862e-02
5.57244658e-01 1.32122099e+00 -5.29476851e-02 7.11488843e-01
3.65504533e-01 2.05168232e-01 -7.69149899e-01 -3.66965570e-02
5.20565391e-01 -3.35934937e-01 6.65715635e-01 -1.80729315e-01
4.20620441e-01 6.17307760e-02 -1.70680761e-01 1.03405631e+00
9.08290505e-01 3.61339927e-01 1.00556207e+00 -1.15345848e+00
-2.69583136e-01 5.79412520e-01 -1.05315852e+00 3.23763877e-01
7.05658078e-01 -6.36120021e-01 5.29320002e-01 -1.36113214e+00
-2.32122883e-01 1.12703443e+00 1.15668678e+00 -7.00216830e-01
-1.02002478e+00 -9.10448372e-01 -6.85049355e-01 -3.55279118e-01
-2.00161076e+00 -8.80269766e-01 8.06663156e-01 -4.38290566e-01
6.01866722e-01 5.17377436e-01 1.47876084e-01 4.75298584e-01
9.81022939e-02 3.00896317e-01 8.80300403e-01 -7.28602648e-01
2.99067765e-01 -1.76443651e-01 -2.30899993e-02 4.36664581e-01
8.72939348e-01 4.14590776e-01 -6.95335507e-01 -5.03112972e-01
9.51019466e-01 -2.08030522e-01 -5.30584931e-01 -3.09333473e-01
-1.59622419e+00 6.66254282e-01 -1.07910924e-01 6.58081114e-01
-7.77075589e-01 6.84128478e-02 -3.69279571e-02 4.27259654e-02
1.65878624e-01 7.79168010e-01 -6.99795559e-02 1.27903387e-01
-1.46307302e+00 1.34846792e-01 8.56441677e-01 1.14049900e+00
2.83575892e-01 9.02251542e-01 -2.28071347e-01 9.23818290e-01
5.18545866e-01 1.55620050e+00 -1.53630972e-02 -9.32622373e-01
4.49981779e-01 -2.14824349e-01 3.72601151e-01 -1.79060960e+00
-4.85300571e-01 -1.38347816e+00 -1.22672200e+00 -3.06503363e-02
3.09451133e-01 -6.43430650e-01 -3.84815097e-01 1.33934712e+00
8.39351416e-02 4.55172271e-01 5.14228642e-01 8.57025445e-01
4.44398403e-01 1.10018289e+00 -6.90990388e-01 -7.70461321e-01
9.03281450e-01 -7.08698988e-01 -1.25261021e+00 -2.53847271e-01
1.43455803e-01 -1.36598420e+00 -1.26710430e-01 7.05461502e-01
-1.04883766e+00 -5.25987625e-01 -1.23304975e+00 7.81524539e-01
9.61166248e-02 3.83116305e-01 4.53087330e-01 1.08502436e+00
-6.56367302e-01 -1.19402468e-01 -3.61891598e-01 3.02219450e-01
-3.65905195e-01 -7.21415281e-02 -3.10107082e-01 -1.01613976e-01
-7.50628054e-01 2.56313473e-01 -3.12046949e-02 1.67424172e-01
-2.00415492e-01 -1.08981371e+00 -1.15154672e+00 1.83206856e-01
-1.26724681e-02 -3.57471973e-01 8.49961698e-01 -6.70015275e-01
-9.95730877e-01 -7.92815015e-02 -6.01725578e-01 -3.02653283e-01
-2.27124751e-01 -3.44437540e-01 -1.12811601e+00 1.28754526e-01
4.25661236e-01 -2.10702434e-01 1.28565252e+00 -1.13976181e+00
-5.69206536e-01 -1.91724107e-01 -8.10877740e-01 4.94377539e-02
7.93704167e-02 -3.18449438e-01 3.76765400e-01 -1.13921189e+00
9.77964461e-01 -4.71823096e-01 -4.78688955e-01 -3.59111369e-01
-9.18683410e-02 5.83173871e-01 4.08418387e-01 -6.90001726e-01
1.56060767e+00 -2.81485963e+00 -3.30982327e-01 7.98293531e-01
-9.86721665e-02 -2.04183280e-01 -3.45222831e-01 7.67382562e-01
-3.19099456e-01 -5.71159124e-01 -2.10151061e-01 2.01157734e-01
-7.89099783e-02 -2.32864797e-01 -3.04208398e-01 6.07358396e-01
-1.50961637e-01 1.57646045e-01 -7.10054696e-01 2.20837127e-02
4.23904546e-02 4.27843839e-01 -4.73159432e-01 -5.32026999e-02
7.17038929e-01 3.42146337e-01 -9.10587087e-02 8.06123495e-01
1.06537759e+00 1.53847441e-01 -2.29280502e-01 -7.26475716e-01
-6.95635736e-01 -3.63252133e-01 -2.15390944e+00 1.36809778e+00
-5.52921355e-01 7.54465401e-01 6.51089251e-01 -9.62536335e-01
1.30819571e+00 5.50828218e-01 5.72106898e-01 -6.02656603e-01
-1.20346874e-01 7.10622907e-01 1.81804523e-01 -3.67462307e-01
4.05865312e-01 -4.18897085e-02 1.12983428e-01 4.11512107e-02
-7.95592517e-02 -1.24705300e-01 3.29906344e-01 4.90921549e-02
8.65995944e-01 -5.96234262e-01 7.15791583e-01 -5.36921501e-01
4.86834556e-01 -3.73534828e-01 5.00543654e-01 8.88804078e-01
3.10245574e-01 6.01062238e-01 -1.15546539e-01 -3.36413011e-02
-8.53369474e-01 -1.20492435e+00 -3.33344519e-01 7.15565920e-01
1.35747939e-01 -1.72869399e-01 -6.41363189e-02 4.36720043e-01
1.31387338e-01 9.54078138e-01 -5.31593747e-02 2.68794328e-01
-5.02009749e-01 -6.04046524e-01 4.76153404e-01 3.95722449e-01
4.35330749e-01 1.24834687e-01 -5.32177806e-01 3.37425828e-01
-1.58356458e-01 -1.13216090e+00 -4.07551557e-01 1.87128931e-01
-7.54128516e-01 -5.71650088e-01 -1.00111747e+00 -6.99669361e-01
6.09361231e-01 7.88171649e-01 7.49456525e-01 -8.61745298e-01
3.65924202e-02 8.20068657e-01 -6.03950679e-01 -7.42137611e-01
2.56616443e-01 -8.23229730e-01 4.71538067e-01 5.34534454e-01
-1.86645880e-01 -8.78988147e-01 -6.93648875e-01 6.99614525e-01
-5.36929607e-01 -4.39536959e-01 9.11016881e-01 7.49169111e-01
6.25560522e-01 5.74272871e-01 6.52295351e-01 7.55423401e-03
8.04207146e-01 -7.70847201e-01 -7.70473480e-01 -7.85437450e-02
-4.00614947e-01 -1.77325621e-01 3.11941028e-01 -2.71993011e-01
-1.08924711e+00 1.36747807e-01 6.06670044e-02 -2.85890430e-01
1.47625953e-01 8.55676591e-01 1.88126847e-01 -3.80920589e-01
9.23429251e-01 3.01313818e-01 -2.87501633e-01 -2.67464578e-01
1.63397610e-01 5.22048295e-01 6.18468523e-01 -2.72351980e-01
9.34407651e-01 2.85706908e-01 7.65235066e-01 -1.48743665e+00
-6.95864379e-01 -1.08813918e+00 -3.13212663e-01 8.86479951e-03
2.63443232e-01 -9.59412634e-01 -1.45299986e-01 2.32098922e-01
-1.11940742e+00 2.70768136e-01 -2.00918406e-01 1.31290233e+00
-2.76594818e-01 3.59941751e-01 9.37315449e-03 -1.36621940e+00
-3.21499139e-01 -6.18296802e-01 9.56763268e-01 1.95833951e-01
-3.56532693e-01 -1.00119638e+00 2.49874026e-01 -1.35979652e-01
7.54115880e-01 3.13449800e-01 6.45744920e-01 -4.31946546e-01
-1.08660422e-01 -6.98493838e-01 -1.68374151e-01 1.82545334e-01
1.69789448e-01 -4.16340798e-01 -5.62066972e-01 -3.05584371e-01
4.11643654e-01 6.52162433e-01 1.39615417e-01 1.23715949e+00
3.25293154e-01 -5.77085435e-01 -4.20448363e-01 7.37822711e-01
1.65785956e+00 6.55434608e-01 5.55082440e-01 -5.27946539e-02
7.71123469e-02 2.95310169e-01 9.09879565e-01 8.16162825e-01
-5.71874559e-01 4.60811108e-01 2.08742276e-01 2.45863777e-02
9.15676057e-02 1.65016323e-01 2.72733960e-02 1.19606841e+00
-4.86633331e-02 -3.80797267e-01 -5.69887280e-01 6.18319213e-01
-1.58630311e+00 -1.18167949e+00 -6.94065213e-01 2.48764420e+00
-2.34136619e-02 -3.70303869e-01 -2.30706006e-01 5.74819982e-01
4.99913931e-01 1.66597798e-01 -2.66000610e-02 -3.78645569e-01
-4.26285803e-01 3.09242725e-01 9.42339480e-01 5.71973860e-01
-9.45111156e-01 -5.95408008e-02 7.05515814e+00 1.04797089e+00
-7.20057726e-01 -5.01604155e-02 -1.52875274e-01 2.52341628e-01
-1.53778404e-01 -2.70473331e-01 -6.11746669e-01 2.89562523e-01
7.90787101e-01 -2.78811097e-01 2.39801779e-01 6.20336831e-01
3.38913262e-01 -4.89733815e-01 -8.40944409e-01 1.70391405e+00
2.63361126e-01 -1.08332324e+00 -2.72076368e-01 -9.01673213e-02
1.25009727e+00 -5.12443006e-01 3.71516079e-01 -1.33774415e-01
-5.57403088e-01 -9.82345581e-01 6.69560015e-01 5.96254468e-01
6.22881353e-01 -8.53907347e-01 9.06147659e-01 4.21530664e-01
-1.50329328e+00 -2.41560727e-01 -4.04014111e-01 -3.68940771e-01
5.88298142e-01 1.36692119e+00 -7.01410711e-01 8.04160357e-01
1.94306657e-01 4.39150929e-01 2.36501187e-01 1.69693160e+00
-1.02462508e-02 7.40123749e-01 -7.46668577e-01 5.16194366e-02
3.38228434e-01 -4.98850226e-01 1.28275049e+00 1.50289655e+00
1.43203616e+00 4.39139307e-01 2.05751464e-01 4.15891588e-01
6.92961693e-01 3.23169291e-01 -8.71767640e-01 -3.18825990e-02
8.32157612e-01 1.03148353e+00 -4.32276368e-01 6.44671358e-03
-5.45221925e-01 3.93775463e-01 -1.13311553e+00 6.94093883e-01
-4.49779332e-01 -7.28237510e-01 3.05029839e-01 2.08628535e-01
3.68371636e-01 -7.83453226e-01 -6.05399370e-01 -3.62535149e-01
-1.44612074e-01 -9.87960100e-01 2.84975886e-01 -8.69221449e-01
-9.82889116e-01 4.11643416e-01 1.33814529e-01 -1.66851807e+00
-2.57804394e-01 -6.59033120e-01 -5.08424342e-01 9.55337048e-01
-8.58897030e-01 -5.40426791e-01 -4.20977101e-02 5.88956952e-01
4.42649007e-01 -6.10263526e-01 8.02990794e-01 5.98150551e-01
2.39896160e-02 5.36861181e-01 6.07423723e-01 -2.19740290e-02
4.52227622e-01 -6.63947523e-01 -4.36959803e-01 1.13401270e+00
9.92111564e-02 6.86176598e-01 1.22304189e+00 -5.08942366e-01
-1.55111718e+00 -7.74369776e-01 4.60962862e-01 8.60215425e-02
6.40844107e-01 1.48325965e-01 -3.38645488e-01 3.88689429e-01
2.25768998e-01 -2.56640285e-01 1.07979393e+00 -3.91908437e-02
-2.57074952e-01 -2.11251214e-01 -1.05469489e+00 2.57657319e-01
5.53959906e-01 9.08944756e-02 -5.02722204e-01 4.41520959e-01
4.25971925e-01 -1.34995639e-01 -8.09091508e-01 9.65318084e-01
8.02067518e-01 -8.45286310e-01 1.26385009e+00 1.35699198e-01
1.63336862e-02 -5.68424225e-01 -1.08809447e+00 -1.48935640e+00
-9.07778382e-01 -7.68665791e-01 -2.00577397e-02 1.02670431e+00
8.59803617e-01 -9.05221701e-01 3.05654198e-01 -2.81398684e-01
-4.33618605e-01 -3.72294545e-01 -8.09085548e-01 -1.17198753e+00
-7.28212893e-01 -5.47782540e-01 9.26853865e-02 1.01446593e+00
-8.21366981e-02 4.11727250e-01 -5.54229379e-01 1.10208070e+00
1.19299352e+00 1.80596754e-01 7.07859516e-01 -1.03759742e+00
-6.02429509e-01 -1.96165547e-01 -5.87290466e-01 -1.22313929e+00
-4.77237612e-01 -5.87482154e-01 5.15104085e-02 -1.44814301e+00
-5.47521293e-01 -4.58127737e-01 -2.06058919e-01 -2.91651547e-01
3.96634072e-01 2.71775782e-01 -2.04350770e-01 -1.95357412e-01
-1.99271873e-01 2.56493241e-01 9.07066703e-01 -1.01393789e-01
-2.57612377e-01 5.37129104e-01 -3.08539093e-01 9.24835324e-01
4.46957558e-01 -6.28775239e-01 -4.92239475e-01 -5.44967055e-01
7.54521132e-01 4.40176249e-01 -1.17447577e-01 -1.55926728e+00
5.45746744e-01 1.99959412e-01 5.33242643e-01 -7.25600719e-01
5.49120545e-01 -9.96258974e-01 6.54044271e-01 3.64319533e-01
7.03155696e-02 1.15239307e-01 2.06527889e-01 7.26800382e-01
-6.69908285e-01 -4.48490053e-01 6.98646903e-01 1.25890657e-01
-5.18291712e-01 -3.19291741e-01 -6.57975614e-01 -2.99111396e-01
7.63642311e-01 -6.42815351e-01 1.60866782e-01 -9.40833211e-01
-6.50418282e-01 -2.39036694e-01 -6.29327178e-01 -2.25587592e-01
7.74884105e-01 -1.48720908e+00 -1.10085368e+00 5.19759774e-01
-2.70161718e-01 -4.63168293e-01 3.36935073e-01 1.42208970e+00
-4.09970403e-01 7.09064007e-01 -6.10567890e-02 -7.43962169e-01
-1.08931351e+00 4.25813437e-01 1.28772691e-01 1.10656776e-01
2.50991464e-01 1.01527083e+00 4.37948048e-01 -2.03221202e-01
6.67829514e-02 -4.39649448e-02 -1.30444974e-01 6.54738471e-02
8.61985743e-01 9.90556419e-01 -7.31506720e-02 -8.65716875e-01
-4.31818336e-01 1.03408504e+00 7.06211686e-01 -5.00356376e-01
1.03224802e+00 -1.80894434e-01 -4.88047749e-01 3.98987085e-01
1.13208103e+00 1.20697355e+00 -3.35101455e-01 -8.03132504e-02
-2.43638113e-01 -8.64001513e-01 1.03712946e-01 -6.87880576e-01
-5.28380752e-01 5.82357168e-01 8.49411190e-01 3.17512929e-01
1.30583799e+00 -4.15319920e-01 2.78669953e-01 5.19433677e-01
6.25869632e-01 -4.70354825e-01 -3.34134735e-02 4.68253911e-01
1.14759219e+00 -7.44199812e-01 5.06636873e-02 -7.50836372e-01
6.52724579e-02 1.12307727e+00 -8.83463547e-02 -2.18246415e-01
8.67699206e-01 7.41251230e-01 1.00462042e-01 7.10415840e-02
-9.55962911e-02 4.92669083e-02 5.85880756e-01 7.38662422e-01
4.56703275e-01 2.82523870e-01 -6.45359099e-01 8.29016268e-01
-3.60714972e-01 -6.25168800e-01 4.11720276e-01 6.46272361e-01
-9.21698987e-01 -4.71742868e-01 -1.40944731e+00 3.93921167e-01
-3.89360696e-01 -3.76728177e-01 2.37976626e-01 2.87114829e-01
-9.86254364e-02 1.45637572e+00 2.86572240e-02 -2.04997063e-02
7.22127557e-01 -1.59903452e-01 4.09527749e-01 -2.33492449e-01
9.97308865e-02 7.43169606e-01 2.50079691e-01 -2.04259545e-01
-3.02730650e-01 -5.93358576e-01 -8.01741183e-01 1.40670359e-01
-5.26913583e-01 6.06609583e-01 8.47735345e-01 4.60368633e-01
2.14655995e-01 5.58158457e-01 7.30072081e-01 -7.90675759e-01
-4.22137856e-01 -6.81207418e-01 -9.39385176e-01 -1.89111084e-01
3.67129534e-01 -5.30059040e-01 -6.98557913e-01 -4.62852381e-02] | [6.496457099914551, 1.3351184129714966] |
55e243e9-151b-455f-bfa1-157b36bace30 | near-realtime-facial-animation-by-deep-3d | 2305.03216 | null | https://arxiv.org/abs/2305.03216v1 | https://arxiv.org/pdf/2305.03216v1.pdf | Near-realtime Facial Animation by Deep 3D Simulation Super-Resolution | We present a neural network-based simulation super-resolution framework that can efficiently and realistically enhance a facial performance produced by a low-cost, realtime physics-based simulation to a level of detail that closely approximates that of a reference-quality off-line simulator with much higher resolution (26x element count in our examples) and accurate physical modeling. Our approach is rooted in our ability to construct - via simulation - a training set of paired frames, from the low- and high-resolution simulators respectively, that are in semantic correspondence with each other. We use face animation as an exemplar of such a simulation domain, where creating this semantic congruence is achieved by simply dialing in the same muscle actuation controls and skeletal pose in the two simulators. Our proposed neural network super-resolution framework generalizes from this training set to unseen expressions, compensates for modeling discrepancies between the two simulations due to limited resolution or cost-cutting approximations in the real-time variant, and does not require any semantic descriptors or parameters to be provided as input, other than the result of the real-time simulation. We evaluate the efficacy of our pipeline on a variety of expressive performances and provide comparisons and ablation experiments for plausible variations and alternatives to our proposed scheme. | ['Eftychios Sifakis', 'Ken Museth', 'Jonathan Swartz', 'Byungsoo Kim', 'Doyub Kim', 'Matthew Cong', 'Sangeetha Grama Srinivasan', 'Hyojoon Park'] | 2023-05-05 | null | null | null | null | ['semantic-correspondence'] | ['computer-vision'] | [ 2.32778475e-01 3.99228454e-01 2.71043956e-01 -4.35586065e-01
-5.82407415e-01 -3.01775515e-01 7.99282491e-01 -2.51289070e-01
-2.05061182e-01 6.92770481e-01 2.89927218e-02 1.23801433e-01
-5.49191758e-02 -8.82404685e-01 -9.48669434e-01 -3.54527116e-01
-3.23309392e-01 7.28818476e-01 2.30242193e-01 -8.26931894e-01
-4.76544797e-01 1.07420170e+00 -2.09663439e+00 2.96135485e-01
4.26413566e-01 8.45461249e-01 -1.77949965e-01 7.06629097e-01
1.00211307e-01 3.79188478e-01 -7.15442002e-01 -5.06039798e-01
3.92281741e-01 -4.47112441e-01 -6.18830919e-01 -1.07275300e-01
8.40285540e-01 -5.96271455e-01 -1.73462287e-01 6.93710744e-01
5.93989253e-01 8.55193213e-02 3.79592299e-01 -1.15121186e+00
1.94258410e-02 4.25019860e-01 -4.30723339e-01 -3.52323622e-01
5.87479711e-01 3.66589814e-01 2.94936597e-01 -5.16960621e-01
1.19625711e+00 1.68680596e+00 1.11231601e+00 9.55328286e-01
-1.44444203e+00 -7.84922600e-01 -1.73018888e-01 -2.46896848e-01
-1.37974715e+00 -6.34648979e-01 8.49719703e-01 -3.08338031e-02
6.65255249e-01 2.47573510e-01 9.85296249e-01 1.51441503e+00
1.67368010e-01 -4.48181666e-02 9.88363504e-01 -6.99008480e-02
3.40883791e-01 -8.94055516e-02 -5.82474411e-01 7.18535900e-01
-1.83202177e-01 6.50106311e-01 -4.06713635e-01 -3.57936412e-01
1.32541239e+00 -5.76836646e-01 -2.22994938e-01 -3.74203682e-01
-8.80177379e-01 5.00089824e-01 3.67313236e-01 2.49365970e-01
-5.37429154e-01 6.26595318e-01 4.49743032e-01 5.98478317e-02
2.69520938e-01 4.24099267e-01 -4.58207577e-01 -1.72217295e-01
-1.27939117e+00 6.74723327e-01 8.28651607e-01 6.55423403e-01
4.33448225e-01 7.13452637e-01 8.41964111e-02 4.01519299e-01
-1.75111681e-01 2.56181866e-01 1.25540569e-01 -1.56325090e+00
-3.63077223e-01 1.11269288e-01 2.11671695e-01 -9.58660603e-01
-5.34870028e-01 -4.44261372e-01 -8.61418605e-01 9.66922700e-01
4.74344164e-01 -1.38758346e-01 -8.80535722e-01 2.16714644e+00
7.01200604e-01 8.52342308e-01 8.98822322e-02 1.10329866e+00
9.39840913e-01 5.57342052e-01 2.86156595e-01 -2.36091956e-01
1.26300752e+00 -4.08816636e-01 -4.78582203e-01 3.81287605e-01
3.50652307e-01 -5.49934030e-01 1.09422266e+00 3.45161796e-01
-1.65154243e+00 -8.77192020e-01 -1.15162694e+00 2.77987979e-02
-4.43529040e-02 -2.12647364e-01 5.27194321e-01 2.90956408e-01
-1.47728395e+00 1.52148128e+00 -8.17341268e-01 -2.49405742e-01
3.45069826e-01 6.30321205e-01 -5.68493903e-01 5.30971587e-01
-1.46051991e+00 9.10903573e-01 6.88820332e-02 -1.60206944e-01
-9.64960337e-01 -1.31929851e+00 -7.79957116e-01 1.17269836e-01
8.83470569e-03 -9.37199593e-01 1.19593167e+00 -1.48527396e+00
-1.98319888e+00 8.92403722e-01 1.52045861e-01 -3.76781374e-01
7.59950936e-01 1.04094632e-01 -4.92710620e-01 1.75346285e-01
-5.30921757e-01 1.03835166e+00 9.17234063e-01 -1.74570453e+00
-8.89184847e-02 -8.97680670e-02 5.12494802e-01 4.10161652e-02
1.90551266e-01 1.76721498e-01 -3.10797125e-01 -6.16854966e-01
-4.30340379e-01 -8.45683575e-01 -4.17314082e-01 6.71536744e-01
8.07975605e-02 3.08687270e-01 9.73186910e-01 -7.70447791e-01
4.81475413e-01 -1.83604074e+00 2.45304957e-01 1.13569975e-01
2.33729377e-01 3.66344064e-01 -3.41919333e-01 8.50671977e-02
-5.35215437e-01 -9.01483744e-02 -2.65032768e-01 -4.97427881e-01
-1.72285184e-01 2.73966432e-01 -1.85103863e-01 3.92061859e-01
3.51430744e-01 7.42070913e-01 -7.98572600e-01 -5.59145033e-01
3.16865057e-01 1.30327785e+00 -5.10953307e-01 1.70770362e-01
-3.89653444e-01 9.27551270e-01 -1.32750496e-01 1.59631163e-01
8.84182155e-01 8.10171962e-02 2.28641599e-01 -6.94578528e-01
8.37812945e-02 -6.74672574e-02 -1.37027860e+00 2.07908821e+00
-7.70848870e-01 2.23416179e-01 5.57572007e-01 -5.76737761e-01
8.16794455e-01 3.47777158e-01 5.04296482e-01 -7.80071676e-01
3.17784190e-01 -5.52144535e-02 -9.59896669e-02 -1.79340884e-01
4.22811836e-01 -5.46385527e-01 -1.29787043e-01 2.11698323e-01
1.65788248e-01 -5.88218391e-01 -2.62466252e-01 -1.03390522e-01
7.41854489e-01 1.06657100e+00 1.89415589e-01 -2.31622592e-01
5.58399916e-01 -1.31127402e-01 4.96432513e-01 2.41206139e-01
4.11380269e-02 6.34243190e-01 3.57168078e-01 -5.45296848e-01
-1.52945125e+00 -1.08543599e+00 -1.07728370e-01 7.71364868e-01
6.11598194e-02 -4.80444491e-01 -1.22399449e+00 -9.97916237e-02
-1.64472654e-01 6.94783032e-01 -8.78430188e-01 -2.03071505e-01
-1.07140172e+00 -2.41711199e-01 9.33897316e-01 5.58974624e-01
3.68699670e-01 -1.02501738e+00 -1.05495858e+00 1.70434594e-01
3.84720534e-01 -1.32679701e+00 8.23162124e-02 -1.89877778e-01
-8.25496078e-01 -6.95979595e-01 -5.49905598e-01 -4.88920242e-01
2.39791989e-01 -6.44208252e-01 1.61151159e+00 3.85648251e-01
-4.62587953e-01 1.30271450e-01 2.17217952e-01 -3.32963653e-02
-1.02847314e+00 -4.67211068e-01 3.34210366e-01 -2.99851686e-01
-6.74961448e-01 -1.25684857e+00 -5.79438329e-01 2.42892653e-02
-9.66302931e-01 5.22162855e-01 -3.04372306e-03 5.95743418e-01
5.64218640e-01 -1.51592255e-01 3.40437144e-01 -5.10072649e-01
4.60733771e-01 -1.01312548e-01 -7.72696674e-01 -1.26543671e-01
-2.37531792e-02 2.27980688e-01 1.16436350e+00 -6.75549090e-01
-1.15967798e+00 2.51323372e-01 -4.87947077e-01 -9.59083676e-01
-3.87759745e-01 1.83667932e-02 -3.00636947e-01 -3.47705066e-01
7.42696106e-01 -1.92042932e-01 1.84274197e-01 -3.73062342e-01
6.55790627e-01 -3.35824676e-02 1.04605460e+00 -8.57581973e-01
9.79781389e-01 6.52831614e-01 6.12570703e-01 -6.42623603e-01
-2.48026147e-01 6.44838333e-01 -7.76945949e-01 -3.62315893e-01
4.91455257e-01 -6.85181677e-01 -1.15379798e+00 5.22048295e-01
-1.27713048e+00 -4.92006212e-01 -7.65789986e-01 2.16280043e-01
-8.68018031e-01 2.55351156e-01 -6.76400900e-01 -6.62545800e-01
-2.00000361e-01 -1.19807541e+00 1.32669604e+00 2.66762227e-01
-4.77800637e-01 -8.41945291e-01 4.87671845e-04 -2.60276258e-01
5.61587751e-01 1.10829711e+00 8.19437861e-01 -1.17144912e-01
-2.10111871e-01 3.23927701e-02 2.65834089e-02 1.17714949e-01
-9.54153463e-02 5.20141244e-01 -1.16290379e+00 -1.77572072e-01
-1.24020562e-01 -3.38934779e-01 1.93217784e-01 2.62210518e-01
1.03892481e+00 -1.79816067e-01 -1.98574036e-01 8.54883194e-01
1.43556786e+00 -1.07883349e-01 6.52464449e-01 -1.29445851e-01
6.95218980e-01 8.94654274e-01 2.59081453e-01 4.33115274e-01
-1.30674586e-01 1.25090468e+00 5.88514030e-01 -4.21282262e-01
-5.24173439e-01 -2.47791544e-01 1.89951807e-01 3.46471131e-01
-4.15076524e-01 2.69094497e-01 -5.36027491e-01 2.12010965e-02
-1.45901084e+00 -1.06644559e+00 1.24485001e-01 2.11599994e+00
9.63745594e-01 4.15258594e-02 3.56478781e-01 -3.04367561e-02
6.57253027e-01 -2.26092245e-03 -5.88513911e-01 -7.87216246e-01
-6.14188015e-02 8.93045545e-01 2.51488723e-02 6.52242243e-01
-6.39228821e-01 9.83018816e-01 6.59480810e+00 1.11936176e+00
-1.36278784e+00 8.28619748e-02 5.67158997e-01 -3.28771770e-01
-2.37341404e-01 -1.20777205e-01 -3.05312544e-01 1.79271027e-01
1.50249696e+00 -1.89267904e-01 4.83412236e-01 8.61503243e-01
5.69239736e-01 2.32669279e-01 -1.33863914e+00 8.63612115e-01
-1.31463230e-01 -1.87286699e+00 3.14773738e-01 -1.31164834e-01
4.15532291e-01 -5.26076794e-01 2.23707575e-02 2.21401453e-01
3.47523838e-01 -1.47253418e+00 9.34267700e-01 5.51529348e-01
1.35029066e+00 -1.05277026e+00 3.06558251e-01 3.18677872e-01
-1.27376437e+00 4.70594317e-01 3.16230543e-02 5.73125407e-02
3.39469045e-01 1.95639543e-02 -4.48558420e-01 5.52552640e-01
5.59379339e-01 2.00078428e-01 -1.78047001e-01 4.21998620e-01
2.44169205e-01 2.21809849e-01 -6.44412816e-01 3.44109893e-01
3.42937224e-02 -1.09344937e-01 6.29004002e-01 1.19530296e+00
9.50687379e-02 1.61736637e-01 -5.55653274e-02 1.13147819e+00
-1.20958097e-01 -8.62033218e-02 -4.01119143e-01 4.04816449e-01
2.72541106e-01 1.38392949e+00 -5.55806637e-01 -4.19527322e-01
8.18345696e-02 9.24690723e-01 1.95254639e-01 1.35735288e-01
-1.42967236e+00 9.22934115e-02 1.02385151e+00 4.38166678e-01
8.32652301e-02 -1.07167758e-01 4.16430682e-02 -7.88022518e-01
-2.36795917e-01 -9.47807491e-01 -1.18173264e-01 -1.22639465e+00
-7.43460476e-01 1.18892431e+00 3.90300184e-01 -1.10372317e+00
-7.87579894e-01 -2.57443219e-01 -4.00777906e-01 1.13987231e+00
-1.18988454e+00 -1.46331882e+00 -5.12467325e-01 5.96718550e-01
2.83200920e-01 2.33477309e-01 1.23193431e+00 3.38202387e-01
-4.79814224e-02 7.14619577e-01 -4.21039253e-01 -1.80710241e-01
2.02019632e-01 -7.22802222e-01 6.99695170e-01 3.98944348e-01
-3.00390571e-01 1.87271729e-01 1.24930227e+00 -5.04081249e-01
-1.21180117e+00 -1.02732646e+00 1.58272028e-01 -2.29430497e-01
4.38267291e-01 -2.73044258e-01 -1.18354619e+00 3.03085357e-01
-6.60640076e-02 2.47393921e-01 1.61715783e-02 -4.52379346e-01
-2.66985029e-01 1.80383492e-02 -1.67324018e+00 8.43889654e-01
1.38727701e+00 -2.83492059e-01 -3.13737720e-01 8.38101376e-04
7.69113660e-01 -9.26707566e-01 -1.14352250e+00 6.49087906e-01
8.85468483e-01 -1.15536523e+00 1.26740253e+00 -7.44997740e-01
5.96944094e-01 -1.96720198e-01 4.83962484e-02 -1.16826904e+00
-2.58731954e-02 -8.92863750e-01 -1.70331746e-01 9.95523274e-01
-1.99572399e-01 -4.82795015e-02 9.43035245e-01 5.30345798e-01
8.06269944e-02 -1.00501299e+00 -1.19874811e+00 -7.36637235e-01
2.33747393e-01 -5.76097608e-01 1.11751521e+00 8.90996516e-01
-5.88412941e-01 6.03894237e-04 -3.06139350e-01 1.92622170e-01
6.53283656e-01 4.42117406e-03 8.49121094e-01 -1.25718212e+00
-4.46207911e-01 -3.87654930e-01 -5.94945669e-01 -4.74962831e-01
6.01780951e-01 -5.55082262e-01 -1.35176539e-01 -1.04919529e+00
-2.15555847e-01 -3.11567634e-01 3.71257752e-01 2.82838613e-01
5.23766339e-01 4.72820491e-01 2.54432470e-01 3.58183198e-02
7.20654950e-02 6.45244360e-01 1.43316376e+00 3.31733167e-01
-2.01665759e-01 -4.85775590e-01 -3.26857477e-01 1.29556489e+00
4.00210947e-01 -2.21984804e-01 -3.35539490e-01 -2.74283979e-02
-9.97577161e-02 6.85911715e-01 7.93763757e-01 -1.32213140e+00
-3.38326663e-01 -2.14817867e-01 5.70842028e-01 -1.00520097e-01
9.83939230e-01 -9.33490813e-01 9.43212032e-01 4.32269335e-01
-4.31165963e-01 4.55233008e-02 6.66558623e-01 8.81806668e-03
3.00564259e-01 1.69067264e-01 1.32060647e+00 -1.53330296e-01
-6.30916476e-01 3.06623220e-01 -4.70924452e-02 -2.35595256e-01
7.45881736e-01 -5.08736134e-01 1.77607182e-02 -5.98386228e-01
-1.09312785e+00 -4.54371333e-01 8.90624642e-01 3.20493490e-01
5.29587388e-01 -1.42872941e+00 -8.71207893e-01 2.15801343e-01
-2.79702961e-01 -4.01668549e-02 4.34848368e-01 1.94288433e-01
-7.27330685e-01 -1.63004935e-01 -7.23000586e-01 -4.47480649e-01
-1.31710005e+00 5.02714157e-01 9.11275327e-01 -2.42220759e-01
-9.26506221e-01 4.77442563e-01 1.82435706e-01 -4.61053401e-01
2.05991883e-02 -2.38590881e-01 -5.48156612e-02 -4.65414494e-01
4.07026350e-01 1.66725963e-01 8.37020576e-02 -1.07297790e+00
-3.57332528e-01 7.79500782e-01 5.07665455e-01 -3.74625146e-01
1.28677297e+00 3.25460017e-01 1.86155841e-01 6.11271299e-02
1.18395436e+00 -8.28243718e-02 -1.61945105e+00 2.76322514e-01
-5.97574234e-01 -1.78985640e-01 2.44391267e-03 -8.12476814e-01
-1.34837031e+00 5.28548002e-01 7.10362673e-01 -4.32343513e-01
1.14847994e+00 -7.80587569e-02 8.84415567e-01 1.32994400e-02
5.33710182e-01 -8.42743993e-01 -1.55914739e-01 1.69301420e-01
1.22320235e+00 -4.31288749e-01 1.19680554e-01 -5.58811009e-01
-2.69104183e-01 1.16397536e+00 7.61388719e-01 -5.76131821e-01
5.15732050e-01 1.05534852e+00 -9.43786930e-03 -1.33562282e-01
-7.08049655e-01 2.60714233e-01 1.47522939e-02 9.99619603e-01
2.93358892e-01 -2.67022699e-01 1.02060147e-01 4.14213330e-01
-6.07035577e-01 3.76622319e-01 5.16365469e-01 7.07912862e-01
1.29875422e-01 -9.48313475e-01 -3.95200938e-01 -1.49890989e-01
-1.99353322e-01 2.04657838e-01 -5.39758019e-02 1.25722659e+00
3.65553528e-01 2.76281565e-01 3.93986017e-01 -3.67214859e-01
5.18507302e-01 -1.28850952e-01 1.01718247e+00 -1.47513464e-01
-8.90980363e-01 -3.98780592e-02 2.73833543e-01 -1.11846292e+00
-5.63591182e-01 -3.50700557e-01 -1.56105149e+00 -6.97800517e-01
1.58626810e-01 -1.30498931e-01 7.50651598e-01 7.87266314e-01
5.08187175e-01 6.90871179e-01 3.91035110e-01 -1.73093557e+00
-4.04581964e-01 -4.54214543e-01 -4.08937424e-01 7.43256450e-01
2.37025112e-01 -8.15130830e-01 -1.35302946e-01 2.04597071e-01] | [12.607062339782715, -0.4288526773452759] |
66aa5e0c-cdb0-4134-81a3-f2e9940e398e | a-fully-automated-pipeline-for-detection-and | 1703.06418 | null | http://arxiv.org/abs/1703.06418v1 | http://arxiv.org/pdf/1703.06418v1.pdf | A Fully-Automated Pipeline for Detection and Segmentation of Liver Lesions and Pathological Lymph Nodes | We propose a fully-automated method for accurate and robust detection and
segmentation of potentially cancerous lesions found in the liver and in lymph
nodes. The process is performed in three steps, including organ detection,
lesion detection and lesion segmentation. Our method applies machine learning
techniques such as marginal space learning and convolutional neural networks,
as well as active contour models. The method proves to be robust in its
handling of extremely high lesion diversity. We tested our method on volumetric
computed tomography (CT) images, including 42 volumes containing liver lesions
and 86 volumes containing 595 pathological lymph nodes. Preliminary results
under 10-fold cross validation show that for both the liver lesions and the
lymph nodes, a total detection sensitivity of 0.53 and average Dice score of
$0.71 \pm 0.15$ for segmentation were obtained. | ['Daniel L. Rubin', 'Yefeng Zheng', 'John W. Lambert', 'Dorin Comaniciu', 'Assaf Hoogi'] | 2017-03-19 | null | null | null | null | ['organ-detection'] | ['medical'] | [-1.61283731e-01 1.16899282e-01 -2.36379296e-01 -1.33840129e-01
-8.67367983e-01 -6.70422792e-01 3.76458913e-01 5.61136782e-01
-5.13536274e-01 5.22872627e-01 9.70744416e-02 -5.84770083e-01
8.80689770e-02 -6.74011528e-01 -3.26678455e-02 -9.41597700e-01
-7.19431639e-01 7.24142492e-01 3.07961583e-01 4.69161421e-01
2.49244068e-02 1.04495382e+00 -2.01433793e-01 -1.00951344e-01
1.02692854e+00 9.28283215e-01 -7.47629404e-02 9.66124356e-01
-6.36461303e-02 6.87856853e-01 -1.46615744e-01 -2.05456689e-01
4.15278733e-01 -4.19574827e-01 -7.33975828e-01 4.19497609e-01
1.16235010e-01 -3.34157705e-01 -1.12655826e-01 8.55347157e-01
4.65840399e-01 -1.95912242e-01 1.14534569e+00 -6.43430889e-01
-2.83818424e-01 5.14428794e-01 -6.21861458e-01 4.45240289e-01
-1.66392297e-01 5.04315138e-01 2.57928967e-01 -1.08697212e+00
4.69680846e-01 7.63446510e-01 1.08527160e+00 3.76656324e-01
-1.15299964e+00 -3.90827924e-01 -7.87559211e-01 -4.07639444e-01
-1.43160844e+00 -1.59145355e-01 7.72853121e-02 -8.79394233e-01
7.16000617e-01 4.41543400e-01 1.08018005e+00 2.33868454e-02
9.64461446e-01 6.20784283e-01 9.34470832e-01 -5.23146689e-01
2.69771218e-01 1.60839587e-01 5.91708049e-02 1.26463819e+00
5.20264506e-01 6.07925057e-02 4.77893651e-01 -5.66103220e-01
1.17270768e+00 -6.60716444e-02 -9.23669711e-02 -5.32523036e-01
-1.41264451e+00 1.11117113e+00 8.46019030e-01 5.42701602e-01
-7.43507087e-01 2.05355063e-01 5.20868361e-01 -4.30439264e-01
4.90891248e-01 2.74198145e-01 -9.10230204e-02 6.58558726e-01
-1.10117865e+00 -2.74747103e-01 8.48913074e-01 5.59591591e-01
7.64269382e-02 -6.45893514e-02 -3.80981266e-01 3.80431175e-01
5.21845162e-01 3.88575733e-01 4.93319243e-01 -8.00969779e-01
-2.13380694e-01 6.97160244e-01 -7.68447518e-02 -6.99606240e-01
-8.27754021e-01 -3.17706823e-01 -1.12931287e+00 2.98878938e-01
6.01170599e-01 -3.77042413e-01 -1.33814991e+00 9.39866066e-01
3.76744181e-01 -3.56398314e-01 -3.34311306e-01 8.06847930e-01
9.38690007e-01 2.74093539e-01 5.78445733e-01 -4.73134845e-01
1.41577888e+00 -7.54557371e-01 -5.49162030e-01 2.14049175e-01
7.84836769e-01 -8.25790286e-01 3.41787457e-01 -5.13531752e-02
-1.37654185e+00 -4.65283766e-02 -6.45744860e-01 3.80065650e-01
-1.37057856e-01 2.72026002e-01 6.87245250e-01 8.35352361e-01
-1.18842125e+00 3.32840294e-01 -1.20849502e+00 -4.54609931e-01
7.32754469e-01 5.11147857e-01 -2.33289748e-01 6.35006577e-02
-5.70895195e-01 1.07408142e+00 4.54082578e-01 1.12478375e-01
-9.56707299e-01 -7.63154864e-01 -8.23127747e-01 -1.22590691e-01
9.61269066e-02 -7.87195504e-01 1.10506463e+00 -6.21622324e-01
-1.10488105e+00 1.12254906e+00 7.65755251e-02 -5.45458436e-01
9.13919032e-01 5.06880581e-01 2.38313135e-02 5.26332140e-01
-6.41595647e-02 6.23248219e-01 3.97973686e-01 -1.18033409e+00
-2.79008597e-01 -2.53686339e-01 -6.29422426e-01 9.25602913e-02
2.54806340e-01 3.58046800e-01 -1.78674415e-01 -4.94571775e-01
5.31275272e-01 -1.06737375e+00 -8.05915475e-01 6.66788101e-01
-4.98716295e-01 6.18764125e-02 4.93433982e-01 -1.06698847e+00
6.72390938e-01 -1.67550766e+00 -3.74072134e-01 6.03270769e-01
6.63168013e-01 1.31950632e-01 4.57933843e-01 -2.23253995e-01
6.02366589e-02 3.42614442e-01 -5.76053321e-01 9.17649493e-02
-2.34183893e-01 -1.39133353e-02 5.52718461e-01 9.91220057e-01
-5.57145849e-02 1.40635109e+00 -8.98372769e-01 -1.12348795e+00
4.32577133e-01 6.65460885e-01 -1.32269353e-01 1.99744217e-02
3.23509812e-01 5.77107430e-01 -2.43911073e-01 9.67265248e-01
6.49257481e-01 -4.84674960e-01 2.78302222e-01 5.90292402e-02
-1.27811819e-01 -3.10218126e-01 -7.74347544e-01 1.31683445e+00
-1.77977666e-01 4.90856737e-01 5.41791618e-01 -4.40600693e-01
6.30214810e-01 7.54419386e-01 1.13535631e+00 -3.02340865e-01
3.77682745e-01 4.27203059e-01 2.69692481e-01 -5.76877117e-01
4.04573604e-02 -4.86857295e-01 2.38225237e-01 5.98807812e-01
-6.03830293e-02 -3.39558452e-01 2.51174808e-01 1.85216159e-01
1.02224040e+00 -4.95825380e-01 7.12673128e-01 -7.88904309e-01
5.45508742e-01 4.49788541e-01 4.32928562e-01 5.08794844e-01
-8.86413753e-01 4.78884637e-01 3.64716053e-01 -4.73387331e-01
-9.25895393e-01 -1.30923533e+00 -6.44259095e-01 4.81124312e-01
-2.91203465e-02 2.82348752e-01 -7.40876794e-01 -9.88173723e-01
1.55502558e-02 3.43360722e-01 -7.96513796e-01 2.65303105e-01
-8.24676514e-01 -1.36494720e+00 3.54443610e-01 5.27542293e-01
4.26561177e-01 -9.79673922e-01 -7.62431145e-01 3.37833285e-01
7.83295408e-02 -6.35424614e-01 -3.51617545e-01 1.17360927e-01
-1.35757661e+00 -1.43290102e+00 -1.22385132e+00 -9.63878512e-01
1.08098018e+00 -8.17773491e-02 1.35731542e+00 4.04616654e-01
-1.20864046e+00 2.93883771e-01 1.24297990e-02 -1.87314287e-01
-7.48673379e-01 -2.69217223e-01 -2.12449074e-01 -4.72607076e-01
7.24035278e-02 6.20840825e-02 -1.01621616e+00 3.87241036e-01
-6.52395010e-01 -2.66405046e-01 8.54621708e-01 7.28548050e-01
5.42592764e-01 -1.94417149e-01 2.03537524e-01 -6.73675716e-01
4.81437504e-01 -5.26432335e-01 -5.21658301e-01 2.58203357e-01
-5.90732932e-01 -6.18960619e-01 5.32396836e-03 -2.19880313e-01
-7.70053267e-01 5.35944104e-01 5.53455241e-02 -4.37664464e-02
-2.95313239e-01 2.80810744e-01 6.67818129e-01 -6.14613354e-01
9.54746902e-01 2.15040028e-01 2.44349808e-01 -1.16916774e-02
1.55950010e-01 2.59840280e-01 2.81501919e-01 1.47920683e-01
5.87040603e-01 5.36367595e-01 2.35321462e-01 -6.10901535e-01
1.87150612e-02 -6.49521351e-01 -1.12407446e+00 -3.45666379e-01
1.06077981e+00 -5.70670426e-01 -1.94165617e-01 1.42598435e-01
-7.91454434e-01 -4.58491266e-01 -3.80788535e-01 9.80795443e-01
-4.54454511e-01 5.16288459e-01 -1.20271862e+00 -5.98718822e-01
-8.99229646e-01 -1.47742987e+00 4.88100648e-01 2.24620998e-01
-1.81513712e-01 -1.57086325e+00 9.51410681e-02 -1.40256107e-01
8.72607827e-01 5.04749537e-01 9.91932213e-01 -8.03990781e-01
-6.96537316e-01 -5.10924816e-01 -3.58857423e-01 -8.19814131e-02
1.94769427e-01 2.45993569e-01 -4.50302869e-01 -4.83022034e-01
-9.95031446e-02 -1.96755137e-02 8.89928043e-01 1.14594460e+00
7.94752181e-01 -6.89381883e-02 -5.81658602e-01 5.45561552e-01
1.59536815e+00 4.90474343e-01 4.51892793e-01 -1.30583629e-01
1.76416665e-01 4.04314727e-01 1.07678369e-01 1.41749635e-01
-1.47076786e-01 9.66322646e-02 4.83137965e-01 -7.02917814e-01
-3.35416228e-01 5.65585613e-01 -3.06345701e-01 5.93699396e-01
3.54995742e-03 4.01107259e-02 -1.46537161e+00 7.69561052e-01
-1.17864001e+00 -5.55907011e-01 -4.36850846e-01 1.89184833e+00
6.84471428e-01 -1.39588445e-01 2.31472939e-01 -1.26416028e-01
8.26640546e-01 -3.97406191e-01 -2.48513713e-01 -1.28964961e-01
3.64297301e-01 1.90985277e-01 6.49395525e-01 5.84707916e-01
-1.37083542e+00 2.87086487e-01 7.76972008e+00 4.38235551e-01
-1.08653080e+00 1.49857581e-01 9.63148832e-01 2.61449248e-01
2.40584627e-01 -2.84542263e-01 -2.78998315e-01 2.84661084e-01
4.32011873e-01 -1.65129662e-01 -2.90467799e-01 5.70260882e-01
7.67320618e-02 -6.28572702e-01 -6.92804039e-01 4.48160768e-01
-7.04240948e-02 -1.40906715e+00 -2.98061043e-01 2.04490989e-01
8.16829860e-01 2.07938269e-01 -1.73555925e-01 7.94462934e-02
2.28213385e-01 -1.16393518e+00 1.37260795e-01 4.71877784e-01
9.02923942e-01 -4.40347165e-01 1.04422617e+00 3.66866738e-01
-1.14359677e+00 3.17345381e-01 -1.39937684e-01 5.60786188e-01
2.41987407e-02 7.96242297e-01 -1.63920736e+00 1.97235554e-01
1.99014381e-01 1.81525260e-01 -5.86205661e-01 1.80897486e+00
1.41531602e-01 4.10112739e-01 -4.47338939e-01 6.04027091e-03
2.53411651e-01 -1.59715489e-01 5.31339228e-01 1.66156709e+00
2.86378413e-01 2.57481158e-01 2.66114026e-01 9.05073583e-01
2.00650290e-01 5.57064354e-01 -2.80997962e-01 3.26868385e-01
1.57448083e-01 1.76431406e+00 -1.66486955e+00 -5.77732861e-01
-4.02215794e-02 6.23893440e-01 -3.25115055e-01 9.28444043e-02
-7.00441062e-01 -1.11972161e-01 -3.09020281e-01 1.17879100e-01
-2.02266648e-01 -1.85951203e-01 -5.47423124e-01 -9.75679278e-01
-5.26437581e-01 -4.28229757e-02 5.65335393e-01 -1.74838617e-01
-1.21959245e+00 4.91522014e-01 -1.23833030e-01 -9.57869172e-01
-2.02830061e-01 -6.32759392e-01 -9.48316813e-01 9.40839469e-01
-1.17218018e+00 -9.36045229e-01 -4.82708782e-01 3.91484976e-01
3.20833653e-01 -5.34949712e-02 9.27113771e-01 6.04642704e-02
-3.48919719e-01 2.69161731e-01 9.82885361e-02 6.41402543e-01
2.11275145e-01 -1.60665667e+00 9.03054886e-03 7.17299700e-01
-4.69989836e-01 4.53155607e-01 2.19001934e-01 -7.88411975e-01
-1.09807396e+00 -1.10857666e+00 8.25187504e-01 -1.59743071e-01
4.46942866e-01 1.14524759e-01 -6.55950487e-01 7.06866562e-01
2.55857885e-01 5.92698753e-01 9.19643521e-01 -5.25799930e-01
2.95853585e-01 5.34428477e-01 -1.84758401e+00 2.57598609e-01
1.19288988e-01 4.54163030e-02 -3.94323498e-01 6.50442660e-01
-1.92175001e-01 -6.13205969e-01 -1.27740240e+00 6.40853763e-01
4.49602187e-01 -9.09609914e-01 1.17675138e+00 -1.47057518e-01
4.34179641e-02 1.34253800e-01 2.98693210e-01 -9.78207648e-01
-6.22646332e-01 -1.96256220e-01 1.30319610e-01 3.46931487e-01
4.99494612e-01 -5.05007923e-01 1.05986035e+00 6.93569720e-01
-1.87679961e-01 -9.20457304e-01 -9.77756023e-01 -2.68298030e-01
6.32003844e-01 1.73076615e-01 -1.69234991e-01 1.00075829e+00
-3.23783495e-02 -3.88388425e-01 5.51124513e-01 -1.58179611e-01
1.10698378e+00 6.03633374e-03 2.39834711e-02 -1.25062430e+00
3.10038418e-01 -7.82563269e-01 -3.02906513e-01 -3.36770743e-01
-3.77565473e-01 -1.23654437e+00 5.71344309e-02 -1.93574083e+00
6.97407067e-01 -8.49819064e-01 -1.73768163e-01 3.93600941e-01
-1.20782644e-01 6.01284981e-01 1.30193979e-02 3.96357805e-01
-2.96397477e-01 -1.14553943e-01 1.55670190e+00 -4.92842272e-02
-1.56871334e-01 1.26859784e-01 -1.17624886e-01 8.84608388e-01
1.08166468e+00 -3.53870451e-01 2.38546371e-01 1.65436327e-01
-4.63860989e-01 4.54663336e-01 5.71015537e-01 -9.21456337e-01
2.33773351e-01 -4.21346398e-03 1.12570584e+00 -8.39770317e-01
8.18366036e-02 -7.68586814e-01 1.35349423e-01 1.47096217e+00
-2.49546185e-01 -1.34912372e-01 1.48624465e-01 1.44169882e-01
8.91635492e-02 -3.59029502e-01 1.37512887e+00 -8.46998811e-01
-3.27186674e-01 4.66751337e-01 -7.59887576e-01 -2.20129997e-01
1.47973633e+00 -3.30003239e-02 2.04392429e-02 -4.98016216e-02
-1.24261057e+00 1.68223143e-01 2.57410645e-01 -4.98835117e-01
5.78740835e-01 -1.20099819e+00 -1.04283345e+00 2.39654958e-01
-2.40704492e-01 -7.31180515e-03 -2.85115503e-02 1.61283684e+00
-1.44998145e+00 6.43765509e-01 -1.08354092e-01 -1.04163253e+00
-1.35729265e+00 2.98429221e-01 1.09524548e+00 -5.31396925e-01
-7.73858607e-01 8.83686602e-01 -3.14467698e-02 -3.37491065e-01
1.42954990e-01 -3.40352565e-01 -2.64641345e-02 -1.19002322e-02
2.76427537e-01 5.34829021e-01 1.25974327e-01 -6.93581760e-01
-4.47120309e-01 4.49600130e-01 9.14906114e-02 1.59956008e-01
7.99825370e-01 -4.95811505e-03 -4.17025179e-01 -5.76859154e-02
1.19371426e+00 -1.49207979e-01 -8.11984658e-01 -3.01699728e-01
7.24947527e-02 -3.49195719e-01 4.42767262e-01 -1.16812778e+00
-1.18666804e+00 7.55909085e-01 9.78269875e-01 3.91505927e-01
8.46532106e-01 6.61019236e-02 5.66065729e-01 -1.03253216e-01
1.72254965e-01 -5.61493874e-01 -1.90470934e-01 2.77834207e-01
6.46478653e-01 -1.41689157e+00 4.39897209e-01 -5.66729009e-01
-3.69023293e-01 1.34238982e+00 2.65666038e-01 -3.67879391e-01
7.52326787e-01 4.57634062e-01 4.33146209e-01 -3.37420166e-01
-4.03895438e-01 3.29502043e-03 4.48510766e-01 3.29097271e-01
6.84229314e-01 4.47877735e-01 -4.04213220e-01 -7.93406889e-02
4.15093720e-01 -2.19224137e-03 3.04474026e-01 1.05875564e+00
-8.42389524e-01 -4.43079323e-01 -6.44286156e-01 6.36314273e-01
-8.82268369e-01 3.08175664e-02 -2.89430708e-01 1.18722105e+00
-1.34398760e-02 4.04800415e-01 1.45929426e-01 3.78101349e-01
-1.09663285e-01 1.05527274e-01 6.48163795e-01 -3.95958275e-01
-1.04301417e+00 5.04271626e-01 -2.77828097e-01 -3.17313164e-01
-2.91301191e-01 -6.92530215e-01 -1.57002294e+00 -2.19474658e-01
-4.22086954e-01 1.13070652e-01 8.22901070e-01 5.56029558e-01
-3.44610363e-01 3.94874305e-01 6.03126824e-01 -8.74693811e-01
-4.61257935e-01 -1.04535532e+00 -8.69601667e-01 8.57004076e-02
3.42020273e-01 -3.61583859e-01 -4.43315506e-01 3.29098515e-02] | [14.4744234085083, -2.6865508556365967] |
0b89b5e8-91f6-4603-894f-3d04a3fff943 | heterogeneous-information-network-based-1 | 2204.11849 | null | https://arxiv.org/abs/2204.11849v2 | https://arxiv.org/pdf/2204.11849v2.pdf | Heterogeneous Information Network based Default Analysis on Banking Micro and Small Enterprise Users | Risk assessment is a substantial problem for financial institutions that has been extensively studied both for its methodological richness and its various practical applications. With the expansion of inclusive finance, recent attentions are paid to micro and small-sized enterprises (MSEs). Compared with large companies, MSEs present a higher exposure rate to default owing to their insecure financial stability. Conventional efforts learn classifiers from historical data with elaborate feature engineering. However, the main obstacle for MSEs involves severe deficiency in credit-related information, which may degrade the performance of prediction. Besides, financial activities have diverse explicit and implicit relations, which have not been fully exploited for risk judgement in commercial banks. In particular, the observations on real data show that various relationships between company users have additional power in financial risk analysis. In this paper, we consider a graph of banking data, and propose a novel HIDAM model for the purpose. Specifically, we attempt to incorporate heterogeneous information network with rich attributes on multi-typed nodes and links for modeling the scenario of business banking service. To enhance feature representation of MSEs, we extract interactive information through meta-paths and fully exploit path information. Furthermore, we devise a hierarchical attention mechanism respectively to learn the importance of contents inside each meta-path and the importance of different metapahs. Experimental results verify that HIDAM outperforms state-of-the-art competitors on real-world banking data. | ['Guangwen Yang', 'Xi Zhang', 'Jiachen Shen', 'Yingsheng Ji', 'Zheng Zhang'] | 2022-04-24 | null | null | null | null | ['implicit-relations'] | ['natural-language-processing'] | [-4.67758924e-01 1.30507261e-01 -2.74163276e-01 -2.72186130e-01
1.61479954e-02 -2.06044152e-01 4.32717085e-01 3.57543796e-01
-2.16209680e-01 4.00727302e-01 2.93636292e-01 -4.80763972e-01
-5.36117792e-01 -1.32229435e+00 -2.25676924e-01 -4.65889037e-01
-4.38216686e-01 4.19545531e-01 2.46648446e-01 -4.04554367e-01
3.60973001e-01 6.44302428e-01 -7.00728297e-01 1.50551021e-01
6.87107563e-01 1.06208813e+00 -1.32196978e-01 -1.47929743e-01
-3.26001167e-01 9.36927199e-01 -5.67189097e-01 -1.07419991e+00
8.99018720e-02 -7.42710978e-02 -5.53199828e-01 -3.12405303e-02
-3.67717773e-01 -2.89380282e-01 -6.34618342e-01 1.00521684e+00
2.88702786e-01 -6.13677315e-02 7.45736122e-01 -1.19450414e+00
-5.95054448e-01 1.13278234e+00 -9.79224861e-01 6.91240251e-01
-2.68218610e-02 -1.13041326e-01 1.48997116e+00 -1.01010263e+00
1.62648350e-01 1.10640979e+00 6.42000914e-01 -6.25352710e-02
-7.78001308e-01 -7.46577024e-01 5.37821651e-01 4.69630361e-01
-1.12153447e+00 4.14568931e-02 9.93335366e-01 -3.64481509e-01
9.00173306e-01 3.73454876e-02 7.39913642e-01 9.62695479e-01
2.63817638e-01 5.73284864e-01 6.23452902e-01 -7.85680339e-02
-2.23991469e-01 4.47166771e-01 3.76499414e-01 6.28067315e-01
4.09920335e-01 -1.12952635e-01 -4.33156997e-01 -1.75062820e-01
7.27160215e-01 3.64456445e-01 -2.92561442e-01 1.15969228e-02
-9.26123440e-01 8.64144385e-01 7.25184679e-01 2.98762113e-01
-3.15820903e-01 -3.61256123e-01 3.57843220e-01 3.74163091e-01
3.35590184e-01 3.95609811e-02 -2.89943516e-01 1.33804590e-01
-3.14262539e-01 -1.39903486e-01 9.14684892e-01 6.48883224e-01
5.62994421e-01 -9.29257125e-02 1.36971176e-01 7.34668732e-01
5.14945388e-01 -2.77611375e-01 4.93711770e-01 -3.64565372e-01
1.15282083e+00 1.08477449e+00 -2.31785312e-01 -1.68530071e+00
-4.73623395e-01 -9.77655411e-01 -1.33736253e+00 1.47705451e-01
3.62512678e-01 -2.93959565e-02 -1.79867566e-01 1.17002320e+00
1.81316510e-01 1.75732151e-01 -8.55684504e-02 6.87236071e-01
4.92735296e-01 5.93515515e-01 7.91286156e-02 -3.60463172e-01
1.20926297e+00 -9.81209636e-01 -4.49531347e-01 -1.59518272e-01
4.16181952e-01 -2.77510852e-01 8.84603500e-01 6.52965903e-01
-7.29365766e-01 -2.95511663e-01 -9.58720326e-01 3.72654319e-01
-3.75355542e-01 -2.76097864e-01 7.27087677e-01 7.54463136e-01
-5.13298213e-01 8.36697757e-01 -4.25357848e-01 2.22294033e-01
5.14435053e-01 1.50895521e-01 -3.01400214e-01 3.05561647e-02
-1.81127214e+00 7.11444914e-01 3.29160631e-01 4.89099056e-01
-5.46656847e-01 -4.27159220e-01 -5.36534309e-01 4.97888982e-01
5.92580676e-01 -4.97260988e-01 6.01944387e-01 -7.33893514e-01
-8.98627400e-01 3.17132980e-01 5.56481719e-01 -4.00611907e-01
6.70992136e-01 -1.19125098e-01 -7.20084488e-01 4.16850112e-02
-1.60986945e-01 -4.73757386e-01 5.26920915e-01 -8.03320467e-01
-6.57618880e-01 -5.02830207e-01 1.78239599e-01 2.82861024e-01
-1.10000956e+00 2.49230564e-01 -4.40215647e-01 -1.04206777e+00
1.82783008e-01 -2.95273870e-01 -3.57124478e-01 -3.58278096e-01
-5.47580481e-01 -2.35713005e-01 5.32456934e-01 -6.96772039e-01
1.86769569e+00 -1.91463387e+00 1.01405241e-01 7.21726239e-01
4.25676346e-01 7.71263242e-02 3.11354876e-01 5.91356337e-01
-8.78920257e-02 4.29498285e-01 -8.11681971e-02 -1.62007660e-01
-1.45359844e-01 -1.95021126e-02 -1.92565441e-01 2.35825077e-01
1.73283920e-01 6.45689785e-01 -6.20161712e-01 -6.65381134e-01
-2.41700187e-01 2.98185647e-01 -3.10927212e-01 2.21590415e-01
3.10689628e-01 8.79790336e-02 -1.00767958e+00 9.74262416e-01
6.11249149e-01 -5.29397249e-01 3.04751039e-01 -2.22825289e-01
2.49522492e-01 5.02587035e-02 -1.29929709e+00 8.02845299e-01
-3.66674036e-01 -8.34608302e-02 -9.18777809e-02 -1.31952989e+00
1.05203295e+00 1.61885664e-01 4.48853940e-01 -3.76896977e-01
4.55580764e-02 9.18228775e-02 6.76866295e-03 -4.09770727e-01
1.81652293e-01 -1.69369534e-01 2.10114643e-02 4.41763222e-01
-4.00183231e-01 5.87561727e-01 1.02373175e-02 5.19827724e-01
1.04958344e+00 -4.02784556e-01 3.39523762e-01 6.34917170e-02
8.08937788e-01 -4.83627319e-01 7.44017065e-01 3.54089290e-01
-1.38224021e-01 2.62616813e-01 1.02876043e+00 -4.02481556e-01
-4.03267026e-01 -7.37425566e-01 -1.05242841e-01 7.67321289e-01
3.21870446e-02 -3.11918169e-01 -3.68022770e-01 -1.17871988e+00
3.30777258e-01 3.63314360e-01 -7.02941358e-01 -1.48930937e-01
-6.06881082e-01 -1.30871236e+00 2.72061974e-01 6.86628282e-01
5.97860634e-01 -1.22098219e+00 -3.16386782e-02 5.38766980e-01
3.38343307e-02 -8.14666331e-01 -2.37109706e-01 4.36661020e-02
-1.02638125e+00 -1.32109869e+00 -6.08965456e-01 -3.96893770e-01
4.92461771e-01 4.80281487e-02 1.23859835e+00 2.70263493e-01
-1.34148449e-01 -8.69368166e-02 -3.78729045e-01 -1.29309282e-01
-1.09284125e-01 2.84114152e-01 -2.29440346e-01 4.92115468e-01
3.59924465e-01 -7.20044553e-01 -7.56709218e-01 6.77807152e-01
-5.48334420e-01 -3.46351415e-01 8.45691860e-01 8.96662295e-01
3.45494449e-01 6.47779226e-01 1.05844784e+00 -1.36831963e+00
7.05471039e-01 -1.19072771e+00 -3.63508821e-01 3.29974025e-01
-1.32145572e+00 -3.67644399e-01 6.15374863e-01 -3.10687814e-03
-1.29686892e+00 -5.47348380e-01 9.51991081e-02 -1.13051228e-01
2.18717948e-01 1.23737133e+00 -5.44310033e-01 -1.64872706e-02
9.97626260e-02 -4.82152589e-02 -1.34886160e-01 -7.12883890e-01
-1.80150405e-01 6.22301698e-01 1.35744616e-01 -4.80947942e-01
6.80430770e-01 1.07747309e-01 2.24225789e-01 -2.83784956e-01
-8.17081511e-01 -3.21405023e-01 -4.05545324e-01 -2.49593318e-01
3.89612257e-01 -6.94413841e-01 -6.25880003e-01 8.47231627e-01
-7.30455339e-01 2.03260452e-01 2.07005128e-01 3.93672496e-01
4.82088923e-02 7.49411464e-01 -1.02496266e+00 -7.71784663e-01
-4.00154293e-01 -8.07760656e-01 1.95714101e-01 2.88130134e-01
3.01864564e-01 -1.30573249e+00 -1.94767714e-01 5.97774684e-01
1.25831306e-01 3.59730780e-01 1.16985834e+00 -9.04091299e-01
-7.52747178e-01 -3.35788697e-01 -5.18163145e-01 4.28314060e-01
2.15266779e-01 -7.19091147e-02 -5.37545919e-01 -2.37910002e-01
2.76173595e-02 -1.99920195e-03 9.92097974e-01 -4.91489246e-02
1.12637281e+00 -4.01325136e-01 -4.20491695e-01 5.58599710e-01
1.42548835e+00 6.53412715e-02 4.97744292e-01 7.99881399e-01
7.63087571e-01 1.06960189e+00 7.89923906e-01 8.29077125e-01
5.44288695e-01 4.20217097e-01 9.43356872e-01 1.04836740e-01
5.67066789e-01 -1.74308941e-01 3.58721197e-01 8.73751819e-01
-4.16029543e-01 -4.36548054e-01 -1.13054109e+00 4.57020968e-01
-1.79744518e+00 -9.25861359e-01 -3.39368731e-01 1.95903111e+00
6.68706775e-01 6.49633050e-01 1.46131366e-01 3.41921210e-01
7.69853473e-01 1.29577994e-01 -4.50681984e-01 1.73063993e-01
-1.94848359e-01 -2.63253897e-01 2.89682686e-01 8.59064609e-02
-1.04075587e+00 3.07770491e-01 5.03598642e+00 7.38005698e-01
-5.44763088e-01 -1.19240373e-01 1.10772753e+00 5.49001843e-02
-4.20359761e-01 -1.27470315e-01 -9.58560169e-01 7.28581667e-01
7.60783553e-01 -3.27032387e-01 -3.70723233e-02 7.45968401e-01
2.97925491e-02 3.88591826e-01 -8.05497706e-01 4.16259319e-01
-2.28030264e-01 -1.00918984e+00 1.38887718e-01 2.38797724e-01
3.29252243e-01 -2.72463083e-01 1.44946054e-01 2.71204680e-01
1.68366566e-01 -8.94846618e-01 3.89637679e-01 7.01699138e-01
2.33388528e-01 -1.10384011e+00 1.15111470e+00 4.35204476e-01
-1.35637760e+00 -6.59163654e-01 -4.91416454e-01 1.68382958e-01
2.74166614e-01 8.08326185e-01 -4.65029359e-01 1.04042995e+00
9.58054781e-01 1.20303822e+00 -8.02382648e-01 9.23709214e-01
-1.67131037e-01 6.14804208e-01 -1.50794089e-02 1.47790521e-01
2.46886879e-01 -5.16442239e-01 2.42438644e-01 1.03316379e+00
4.59972292e-01 1.68549389e-01 7.51798898e-02 5.05924761e-01
-1.58583403e-01 5.27908623e-01 -4.46729511e-01 -2.83915307e-02
3.07424277e-01 1.33355021e+00 -5.10423779e-01 -1.10760167e-01
-8.85856569e-01 4.15813297e-01 5.49001992e-01 2.12732539e-01
-6.52075231e-01 -6.15279615e-01 3.32671106e-01 3.56665969e-01
1.76538199e-01 1.10243544e-01 -2.43506819e-01 -1.13412058e+00
-5.53789400e-02 -7.71554053e-01 1.01967108e+00 -2.75031686e-01
-1.77415395e+00 7.02258348e-01 -2.29721233e-01 -1.25892103e+00
1.36635257e-02 -5.14066339e-01 -9.52215314e-01 1.00018060e+00
-1.83963335e+00 -1.14223659e+00 -3.42030883e-01 6.93691611e-01
2.85160452e-01 -6.79845572e-01 3.87761533e-01 5.33432126e-01
-1.04222691e+00 7.68189549e-01 7.72391213e-04 4.46193188e-01
5.06763339e-01 -1.20131290e+00 4.14533347e-01 7.03916728e-01
8.97236168e-02 6.02604687e-01 2.24117041e-01 -7.18242705e-01
-8.51548016e-01 -1.00443435e+00 8.58989358e-01 -2.63935089e-01
1.00835061e+00 1.28119782e-01 -1.37143111e+00 6.55610561e-01
3.63702588e-02 5.40329739e-02 6.98094368e-01 2.73956090e-01
-4.04547244e-01 -3.71334821e-01 -9.15120006e-01 2.82959968e-01
1.04745197e+00 -1.76831916e-01 -5.22874713e-01 1.67729095e-01
4.81410533e-01 3.24039310e-01 -1.19880843e+00 4.46477801e-01
2.88105369e-01 -1.40864587e+00 1.12451518e+00 -6.94853663e-01
5.34746408e-01 2.63131559e-01 1.79885924e-01 -1.03991568e+00
-6.12285018e-01 -2.98645139e-01 -4.63653356e-01 1.64806366e+00
4.97119755e-01 -9.29207563e-01 1.01240373e+00 4.78309691e-01
1.75285742e-01 -1.08421385e+00 -6.98652565e-01 -6.36915147e-01
7.64257982e-02 1.58507843e-02 7.44438887e-01 1.17433691e+00
3.46309036e-01 2.48884544e-01 -3.55614692e-01 1.90509737e-01
7.37594903e-01 2.65795380e-01 2.11771011e-01 -1.65284824e+00
-6.69814408e-01 -7.14192510e-01 -3.28590751e-01 -6.08489871e-01
1.51254401e-01 -7.18940139e-01 -8.61024916e-01 -1.29162443e+00
2.31705666e-01 -7.85604894e-01 -1.03580213e+00 3.04009825e-01
-3.95548135e-01 -1.41131714e-01 -1.65494513e-02 5.38828969e-01
-1.35300517e-01 4.48305726e-01 1.10127902e+00 -1.65511280e-01
2.67772079e-02 5.90857148e-01 -7.97380269e-01 7.46950626e-01
8.52709174e-01 -3.47317994e-01 -5.17551064e-01 -2.52689332e-01
8.31907690e-01 4.29053128e-01 2.25854471e-01 -5.85588276e-01
2.40905702e-01 -2.86657870e-01 2.21883893e-01 -4.74591523e-01
1.51335280e-02 -9.93936181e-01 1.68864101e-01 4.00481820e-01
-1.21729396e-01 3.51906329e-01 -4.40835506e-01 1.20005667e+00
-5.61190069e-01 -5.20341039e-01 3.25710833e-01 -3.93446863e-01
-5.12257218e-01 7.00774908e-01 -5.34488261e-02 -1.02314716e-02
9.19198990e-01 -5.44458739e-02 -2.81993568e-01 -2.79420227e-01
-8.15677047e-01 4.17758912e-01 2.39964388e-02 4.08169746e-01
7.39425182e-01 -1.13098323e+00 -7.89609194e-01 2.04054624e-01
4.88018915e-02 2.42968984e-02 3.32433313e-01 7.85418332e-01
-3.58858466e-01 -1.13740587e-03 -1.75207034e-01 1.62120894e-01
-9.96393919e-01 6.22247815e-01 3.26687634e-01 -8.30241203e-01
-5.86788654e-01 8.16450775e-01 1.79908395e-01 1.19618913e-02
3.39421630e-01 9.58456248e-02 -1.11364293e+00 7.26126909e-01
3.02506238e-01 5.29461265e-01 4.30089049e-02 -4.78673726e-01
-2.01328650e-01 3.35688233e-01 -5.13502896e-01 3.66802186e-01
1.52469277e+00 -3.13169122e-01 -1.86304718e-01 1.27881542e-01
8.54141772e-01 1.06010549e-01 -9.53434169e-01 -5.39930403e-01
4.89584297e-01 -6.31676555e-01 6.49190322e-02 -5.86446643e-01
-1.80420220e+00 1.03195989e+00 -5.28153814e-02 6.33112073e-01
1.04455149e+00 -1.57890096e-01 6.89036846e-01 3.03370923e-01
4.61689502e-01 -9.86347616e-01 3.23928177e-01 1.08004123e-01
8.80155325e-01 -1.17239368e+00 1.25104219e-01 -6.45528674e-01
-7.92691171e-01 1.20944440e+00 8.04507613e-01 1.87783092e-01
1.19425952e+00 2.39433944e-02 -4.84478585e-02 -3.14967096e-01
-6.93988204e-01 3.46981972e-01 -1.51749533e-02 4.24366683e-01
2.71274388e-01 -1.87610060e-01 -2.42211476e-01 1.37322295e+00
7.70285875e-02 -3.86882216e-01 5.72033882e-01 6.13183200e-01
-3.82604510e-01 -1.20309842e+00 -3.38260829e-02 6.88408792e-01
-9.16459560e-01 -2.08713129e-01 -1.70112506e-01 7.37849474e-01
-1.48327380e-01 6.93152368e-01 -1.37512997e-01 -4.29980814e-01
5.23546994e-01 -5.22125401e-02 -2.70335734e-01 -4.98025000e-01
-1.11995912e+00 -7.04336837e-02 2.15329066e-01 -3.04349303e-01
-6.37551472e-02 -8.66241813e-01 -9.89352822e-01 -3.20247620e-01
-4.46775973e-01 2.29845956e-01 5.96703589e-02 4.98234540e-01
2.08379358e-01 6.32625580e-01 1.21319342e+00 -1.31442860e-01
-1.05656648e+00 -8.49381208e-01 -1.09775150e+00 3.45275849e-01
-3.42719294e-02 -7.28370190e-01 -6.38830423e-01 -3.37788731e-01] | [7.229859352111816, 6.072728157043457] |
13a5eeb4-5273-45dc-954a-770acab0ded9 | deeply-supervised-density-regression-for | 2011.03683 | null | https://arxiv.org/abs/2011.03683v2 | https://arxiv.org/pdf/2011.03683v2.pdf | Deeply-Supervised Density Regression for Automatic Cell Counting in Microscopy Images | Accurately counting the number of cells in microscopy images is required in many medical diagnosis and biological studies. This task is tedious, time-consuming, and prone to subjective errors. However, designing automatic counting methods remains challenging due to low image contrast, complex background, large variance in cell shapes and counts, and significant cell occlusions in two-dimensional microscopy images. In this study, we proposed a new density regression-based method for automatically counting cells in microscopy images. The proposed method processes two innovations compared to other state-of-the-art density regression-based methods. First, the density regression model (DRM) is designed as a concatenated fully convolutional regression network (C-FCRN) to employ multi-scale image features for the estimation of cell density maps from given images. Second, auxiliary convolutional neural networks (AuxCNNs) are employed to assist in the training of intermediate layers of the designed C-FCRN to improve the DRM performance on unseen datasets. Experimental studies evaluated on four datasets demonstrate the superior performance of the proposed method. | ['Hua Li', 'Mark A. Anastasio', 'Lilianna Solnica-Krezel', 'Kyaw Thu Minn', 'Shenghua He'] | 2020-11-07 | null | null | null | null | ['automatic-cell-counting'] | ['miscellaneous'] | [ 3.51379216e-01 -4.29449886e-01 4.14600194e-01 -3.31137031e-01
-6.02553487e-01 -1.53503090e-01 4.38933223e-01 2.85054743e-01
-9.55320954e-01 1.00384736e+00 -4.23312604e-01 -8.60764310e-02
2.70303607e-01 -6.64090395e-01 -4.34288383e-01 -1.07378864e+00
7.59750828e-02 6.83077753e-01 3.83895814e-01 4.93288904e-01
4.62668598e-01 7.72230327e-01 -1.24793112e+00 -1.54276550e-01
7.18050659e-01 9.00232911e-01 5.27116597e-01 1.00340211e+00
-3.81059408e-01 1.09301448e+00 -6.02376223e-01 -4.46474664e-02
-4.02889043e-01 -3.67218792e-01 -6.05178714e-01 3.14747505e-02
-4.95719202e-02 -4.81636763e-01 -8.48150700e-02 9.13975656e-01
5.70053339e-01 -3.61541100e-02 1.18282628e+00 -8.18375230e-01
-5.68188190e-01 -8.57033674e-03 -9.75335896e-01 6.89008892e-01
-1.73727334e-01 -6.98681623e-02 1.31970868e-01 -8.30517232e-01
3.37777764e-01 1.09209871e+00 5.55667937e-01 6.88625157e-01
-1.23603439e+00 -6.31511033e-01 -3.10761720e-01 -6.09048679e-02
-1.59461188e+00 -2.85062104e-01 3.61888856e-01 -6.34830236e-01
8.40239763e-01 -1.62409574e-01 4.83308375e-01 5.57155967e-01
1.83714256e-01 5.03569901e-01 1.12797928e+00 -3.34749371e-01
3.20954353e-01 2.27431133e-01 -1.91884905e-01 9.32885766e-01
4.39387798e-01 -4.32221532e-01 -7.30335191e-02 -4.37049232e-02
1.33859706e+00 1.42052457e-01 -1.18421108e-01 -3.86662558e-02
-1.12400317e+00 6.55203104e-01 3.82610470e-01 5.27271569e-01
-1.96821854e-01 1.26469687e-01 2.90098220e-01 -4.42317069e-01
5.69088936e-01 1.75793961e-01 -2.66065985e-01 -1.00071594e-01
-9.99479294e-01 -1.24645571e-03 5.38567662e-01 6.01720870e-01
7.06435442e-01 -1.01950727e-01 -1.73754409e-01 9.51286077e-01
1.76439270e-01 4.13711697e-01 5.48081160e-01 -6.99370325e-01
9.61890593e-02 9.09489453e-01 8.84359479e-02 -7.19116628e-01
-6.28014266e-01 -1.64252654e-01 -1.25081146e+00 2.05072667e-03
5.82146287e-01 -6.96133599e-02 -9.16225970e-01 1.38322389e+00
4.72854823e-01 2.08840743e-01 -1.70471430e-01 7.82402515e-01
6.58927560e-01 7.46697843e-01 2.89441168e-01 -3.64651263e-01
1.11691844e+00 -5.53909242e-01 -5.13364792e-01 3.74553762e-02
7.02872455e-01 -5.75193226e-01 7.43489087e-01 7.06369132e-02
-8.71990025e-01 -3.83365661e-01 -9.75713968e-01 -4.09058601e-01
-4.31732684e-01 4.83563453e-01 4.69355851e-01 2.32381672e-01
-7.91116178e-01 3.23549926e-01 -1.01443779e+00 -4.82344508e-01
9.40205395e-01 6.98533833e-01 -3.05799454e-01 -3.83726284e-02
-4.94208723e-01 6.74003243e-01 2.87918955e-01 2.31229320e-01
-7.13106096e-01 -5.27105808e-01 -8.64472389e-01 5.17299445e-03
-2.19889671e-01 -5.74110627e-01 9.46223736e-01 -3.59206825e-01
-1.33454013e+00 1.03013182e+00 -4.59384590e-01 -2.22929060e-01
4.42306638e-01 1.89139828e-01 1.55279443e-01 2.83603519e-01
3.47148478e-02 8.83254707e-01 2.90457338e-01 -1.26036239e+00
-7.66618252e-01 -5.35280466e-01 -4.80541527e-01 3.71467583e-02
-1.72913983e-01 -1.50334746e-01 -5.04438162e-01 -2.92645574e-01
-3.85328978e-02 -5.10458767e-01 -2.81458825e-01 1.64546147e-01
-1.76639706e-01 -2.92187724e-02 1.03752458e+00 -4.55207944e-01
9.40277815e-01 -1.89383531e+00 1.68849230e-01 -2.53581941e-01
3.40782374e-01 3.88104379e-01 1.27710924e-01 -1.26820639e-01
5.63664794e-01 -1.46383211e-01 -2.82504827e-01 -6.86926126e-01
-5.11757135e-01 1.81720093e-01 3.60810399e-01 7.79669523e-01
6.00187659e-01 7.67405808e-01 -7.41783738e-01 -1.11177957e+00
6.27753139e-01 1.12038112e+00 -2.02458337e-01 4.42128360e-01
-4.15392593e-02 1.00230074e+00 -3.17009449e-01 9.42541182e-01
8.03753197e-01 -6.77161813e-01 7.62682036e-02 -2.40388244e-01
-2.03626484e-01 -3.81025434e-01 -7.10346818e-01 1.17975461e+00
-4.29150522e-01 3.53373319e-01 -7.58455414e-03 -1.05677807e+00
8.89146864e-01 -6.30701706e-02 5.16349792e-01 -4.39986199e-01
6.78278863e-01 3.39817762e-01 -3.66803966e-02 -5.11279941e-01
1.89307496e-01 -2.32490376e-01 1.67488813e-01 1.82863131e-01
2.81166136e-01 -2.11612523e-01 3.91535610e-01 -2.60972269e-02
9.19645727e-01 -1.38624087e-01 5.11033952e-01 -2.83894986e-01
1.01257491e+00 -2.77593523e-01 6.33212209e-01 2.40838423e-01
-3.91030222e-01 6.64128363e-01 5.09130538e-01 -6.05894327e-01
-1.24441755e+00 -6.80261612e-01 -4.38719422e-01 5.12193143e-01
1.74665257e-01 2.61595219e-01 -8.90093565e-01 -2.97939777e-01
-1.66937456e-01 8.85333400e-03 -7.50612020e-01 3.12779814e-01
-5.59841990e-01 -1.15192246e+00 4.61238027e-01 5.37990928e-01
6.49125993e-01 -9.48949575e-01 -5.90759158e-01 2.51262993e-01
-1.32220224e-01 -1.41403735e+00 -3.09981018e-01 2.54686236e-01
-8.49701405e-01 -1.12687409e+00 -1.11520958e+00 -1.03353035e+00
1.07686126e+00 2.23767936e-01 7.85508394e-01 2.80601233e-01
-7.77241766e-01 -2.28311881e-01 6.58046873e-03 -5.94061792e-01
-2.78514475e-01 2.38429874e-01 -9.57159773e-02 -1.37364239e-01
7.93746233e-01 -5.66632152e-01 -8.95740747e-01 2.46709168e-01
-8.22694421e-01 -6.96049333e-02 6.93588197e-01 9.89196062e-01
1.06429744e+00 1.61137190e-02 6.24559939e-01 -1.08253360e+00
4.10418987e-01 -3.76242280e-01 -9.20619130e-01 1.40567943e-01
-3.02840680e-01 -3.20701897e-02 8.52559447e-01 -4.65371281e-01
-8.64701152e-01 2.62672633e-01 -2.42896959e-01 -3.35155308e-01
-1.12088703e-01 9.33694988e-02 8.60002041e-02 -2.59550512e-01
3.29885900e-01 4.30328339e-01 2.67417848e-01 -1.38219357e-01
-2.76818871e-01 8.32037210e-01 5.65702796e-01 -2.57446259e-01
3.12500000e-01 6.39436483e-01 3.88004810e-01 -9.54172134e-01
-7.53572464e-01 -7.26266623e-01 -8.18413019e-01 -9.74665508e-02
1.04414940e+00 -9.19084311e-01 -1.01985478e+00 8.97481501e-01
-1.07078862e+00 -2.95319796e-01 3.15068454e-01 5.65208197e-01
-4.49589461e-01 3.74796718e-01 -9.21937227e-01 -1.13421249e+00
-6.19457364e-01 -1.24162066e+00 1.23997557e+00 7.00727105e-01
1.62597168e-02 -1.14316440e+00 1.13029607e-01 4.10469621e-01
5.25824189e-01 2.08391681e-01 1.09511459e+00 -3.09409320e-01
-4.69991237e-01 -3.39138567e-01 -7.01607883e-01 1.79976359e-01
3.31018955e-01 2.25315928e-01 -1.08706903e+00 -1.54773265e-01
-1.61292642e-01 -6.13117158e-01 8.85840356e-01 8.91477644e-01
1.52284873e+00 -2.88302340e-02 -5.49230039e-01 8.23722482e-01
1.67576015e+00 9.59010422e-02 7.03437150e-01 5.16113117e-02
7.61309743e-01 3.66417676e-01 3.62134844e-01 4.85044897e-01
3.47442776e-01 2.39126444e-01 2.34305575e-01 -4.23405439e-01
3.02541871e-02 1.41971940e-02 -3.10341626e-01 9.67933595e-01
-1.81471854e-01 -1.97597682e-01 -6.15125597e-01 5.55386901e-01
-1.60095096e+00 -7.08945513e-01 -2.44746394e-02 1.97154772e+00
9.20207918e-01 -4.24033962e-02 2.15305939e-01 1.25453278e-01
8.29223871e-01 -4.02789086e-01 -7.91884065e-01 -1.12726726e-01
-3.45932469e-02 1.26923025e-01 5.34559727e-01 3.31990242e-01
-1.21810234e+00 6.68361247e-01 6.13958693e+00 8.24489653e-01
-1.12724447e+00 -1.52812108e-01 1.21612036e+00 -8.67213160e-02
3.21947604e-01 -5.54582238e-01 -1.00379205e+00 6.54785037e-01
5.81363440e-01 2.43084922e-01 3.00866336e-01 6.37579918e-01
9.47118178e-02 -5.53199649e-01 -8.84620905e-01 1.16589963e+00
-6.47638366e-02 -1.35065496e+00 6.08335733e-02 1.95413858e-01
5.38542867e-01 -3.67366187e-02 -1.34082764e-01 1.66515470e-01
1.34012654e-01 -1.21405327e+00 8.00462738e-02 5.04241586e-01
1.14749837e+00 -9.20939803e-01 1.31139982e+00 4.74790275e-01
-1.12007332e+00 1.50046498e-01 -8.68260443e-01 1.23157769e-01
1.42287195e-01 7.86300004e-01 -7.64652133e-01 -1.65385827e-01
5.56095362e-01 4.70988274e-01 -4.46333408e-01 9.22174454e-01
3.77060086e-01 2.61928171e-01 -2.28315875e-01 -3.04239511e-01
5.93446046e-02 -2.80530423e-01 -1.80142462e-01 1.55000067e+00
3.78461212e-01 2.41036732e-02 -1.50976419e-01 9.98169065e-01
-3.98106545e-01 2.62322072e-02 -3.86129469e-01 -2.23729566e-01
4.71125454e-01 1.90208673e+00 -1.22274041e+00 -1.94466546e-01
-3.34182709e-01 6.66851997e-01 7.97613978e-01 1.79351438e-02
-8.75451565e-01 -3.28575999e-01 2.13801593e-01 3.15542132e-01
1.99765638e-01 -1.10504113e-01 -6.86124563e-02 -9.54281509e-01
-4.14237320e-01 -3.01954955e-01 9.84307155e-02 -3.85619253e-01
-1.44049418e+00 4.52652365e-01 -3.04797977e-01 -8.92976999e-01
1.09741636e-01 -8.15488279e-01 -5.18532038e-01 9.73254502e-01
-1.74125314e+00 -1.09659100e+00 -6.43035471e-01 4.44548398e-01
5.30589283e-01 -7.25493282e-02 9.16598141e-01 4.51722771e-01
-8.09323490e-01 3.32768440e-01 9.22677070e-02 3.16409379e-01
4.50512826e-01 -1.33955002e+00 1.52936045e-04 4.74108189e-01
-3.11565638e-01 4.65920359e-01 2.73258388e-01 -4.30949420e-01
-1.05170083e+00 -1.23916399e+00 6.20784700e-01 -5.03055193e-02
3.85105699e-01 -3.73255104e-01 -9.46386576e-01 2.18385816e-01
-3.25568616e-01 5.13144612e-01 9.60925639e-01 -3.58840227e-01
1.39905944e-01 1.26321167e-01 -1.43993115e+00 2.99999356e-01
3.61633807e-01 -3.46907347e-01 1.26211599e-01 2.25410879e-01
2.46546254e-01 -2.94465750e-01 -9.22334611e-01 3.08684826e-01
7.01265633e-01 -8.63813341e-01 7.99777865e-01 1.68037731e-02
5.46985745e-01 -4.76396143e-01 -1.96268875e-02 -7.94568837e-01
-3.40554357e-01 1.02798007e-02 -2.13569388e-01 1.18715990e+00
1.45432174e-01 -3.57444704e-01 9.65820909e-01 2.63968468e-01
2.16602609e-01 -1.05432868e+00 -9.33541536e-01 -3.41334760e-01
2.25483045e-01 4.62185323e-01 3.70407611e-01 6.90766394e-01
-1.08140767e-01 5.35894156e-01 -6.28204942e-02 -9.13988799e-02
7.25608945e-01 -3.41336876e-01 7.12476373e-01 -1.25717044e+00
-5.71140498e-02 -2.49674454e-01 -6.53274953e-01 -8.73221695e-01
-5.89093529e-02 -5.10438800e-01 6.62325099e-02 -1.57276535e+00
7.23381698e-01 -4.34014857e-01 -2.39966094e-01 -3.92745994e-02
-3.96389008e-01 5.07142723e-01 -2.73629516e-01 4.01189238e-01
-6.45707011e-01 3.45186114e-01 1.53033733e+00 -1.77598506e-01
-2.63539068e-02 -9.37651843e-02 -4.64837462e-01 6.63286924e-01
8.66585255e-01 -4.39705014e-01 -4.35006469e-01 -3.42822134e-01
-3.78288925e-02 5.00171781e-02 2.72470921e-01 -1.21741140e+00
3.70174915e-01 -4.63732965e-02 1.06122744e+00 -8.17004323e-01
4.37032670e-01 -4.69014734e-01 -2.13678345e-01 4.20922309e-01
-2.57980339e-02 -1.70755357e-01 2.02378973e-01 5.62281907e-01
-1.68096781e-01 -2.65447646e-01 1.26145983e+00 -4.51806366e-01
-1.16201825e-01 3.88164908e-01 -4.31182712e-01 -1.04712673e-01
9.84899580e-01 -4.08740312e-01 -4.90002453e-01 6.13935515e-02
-2.08757386e-01 6.71080947e-02 5.50898552e-01 -4.13160950e-01
6.61450386e-01 -1.04447472e+00 -5.18101156e-01 1.22069441e-01
1.33368373e-01 6.94481909e-01 2.77981550e-01 1.00038147e+00
-1.14927125e+00 3.40439826e-01 -3.10816675e-01 -8.50597799e-01
-1.13645577e+00 4.45568025e-01 3.96569371e-01 -5.71226597e-01
-2.99262792e-01 1.07808137e+00 2.58985460e-01 -3.09427500e-01
1.58493504e-01 -4.57514852e-01 -5.23869693e-01 -3.44469100e-01
7.78164327e-01 3.76636505e-01 -1.11356102e-01 -6.61926031e-01
-2.86805868e-01 7.47387469e-01 -4.32205409e-01 3.30902100e-01
1.37222242e+00 -2.47545183e-01 -2.72543162e-01 5.78128219e-01
1.31349170e+00 -5.24767280e-01 -1.42508054e+00 1.13673499e-02
-3.04694474e-01 -3.52540821e-01 1.36222452e-01 -4.43402290e-01
-1.04194117e+00 1.16466773e+00 5.24465919e-01 -1.35521233e-01
9.58248556e-01 -2.20376134e-01 7.15425432e-01 3.40432078e-01
2.65969574e-01 -1.10140312e+00 7.70778507e-02 3.23573679e-01
2.91368634e-01 -1.46973515e+00 2.49494791e-01 -3.83477956e-01
-4.04664248e-01 1.08870888e+00 8.70293617e-01 -2.19604932e-02
6.07192576e-01 6.15156114e-01 1.95967425e-02 -1.94488242e-01
-5.99876881e-01 3.05684027e-03 -1.20089255e-01 8.06782246e-01
8.22509170e-01 -2.69589007e-01 -9.96260121e-02 2.59833843e-01
2.77634263e-01 5.06011665e-01 5.32946408e-01 9.93265748e-01
-6.19844258e-01 -6.30523682e-01 -1.63646847e-01 7.89797366e-01
-7.21393764e-01 1.12566657e-01 -1.48655310e-01 6.82959318e-01
1.43382251e-01 7.29769349e-01 4.33117598e-01 -2.81610172e-02
-2.05700532e-01 -2.59353310e-01 7.83781826e-01 -5.07453322e-01
-2.38151923e-01 2.85253912e-01 -6.16477191e-01 -6.96807285e-04
-7.25472689e-01 -3.35312396e-01 -1.50775003e+00 -5.09824336e-01
-5.16739666e-01 -1.82264328e-01 6.13574326e-01 9.64561820e-01
5.51974922e-02 6.05678678e-01 3.83928329e-01 -9.20960605e-01
1.58565696e-02 -1.19677603e+00 -9.08930063e-01 2.24633336e-01
3.14015597e-01 -7.28949845e-01 -2.72808015e-01 3.94334227e-01] | [14.749423027038574, -3.1843934059143066] |
541276ea-e497-4266-ab14-391591b17136 | dna-steganalysis-using-deep-recurrent-neural | 1704.08443 | null | http://arxiv.org/abs/1704.08443v3 | http://arxiv.org/pdf/1704.08443v3.pdf | DNA Steganalysis Using Deep Recurrent Neural Networks | Recent advances in next-generation sequencing technologies have facilitated
the use of deoxyribonucleic acid (DNA) as a novel covert channels in
steganography. There are various methods that exist in other domains to detect
hidden messages in conventional covert channels. However, they have not been
applied to DNA steganography. The current most common detection approaches,
namely frequency analysis-based methods, often overlook important signals when
directly applied to DNA steganography because those methods depend on the
distribution of the number of sequence characters. To address this limitation,
we propose a general sequence learning-based DNA steganalysis framework. The
proposed approach learns the intrinsic distribution of coding and non-coding
sequences and detects hidden messages by exploiting distribution variations
after hiding these messages. Using deep recurrent neural networks (RNNs), our
framework identifies the distribution variations by using the classification
score to predict whether a sequence is to be a coding or non-coding sequence.
We compare our proposed method to various existing methods and biological
sequence analysis methods implemented on top of our framework. According to our
experimental results, our approach delivers a robust detection performance
compared to other tools. | ['Byunghan Lee', 'Sunyoung Kwon', 'Sungroh Yoon', 'Ho Bae'] | 2017-04-27 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 1.18914104e+00 -5.69415689e-02 -1.35749310e-01 1.79170564e-01
-4.89158273e-01 -6.12202525e-01 5.26566327e-01 -1.31020352e-01
-2.53348529e-01 7.79525936e-01 1.00602970e-01 -7.19029844e-01
4.94963735e-01 -9.62518573e-01 -6.33359790e-01 -1.20058429e+00
-8.75834003e-02 1.83285087e-01 4.65405375e-01 -2.46037573e-01
6.43950641e-01 1.60712853e-01 -1.57975531e+00 5.79149246e-01
6.62210524e-01 1.79223537e-01 3.16427499e-01 1.12638044e+00
-2.43089229e-01 4.89669710e-01 -8.99285972e-01 -2.67874729e-03
1.04915254e-01 -1.15384746e+00 -3.50013435e-01 -2.89680511e-01
-4.70416397e-01 -1.93839699e-01 -3.76994163e-01 1.29182935e+00
5.74517548e-01 -4.17055666e-01 7.54920900e-01 -7.44756162e-01
-5.77045381e-01 9.17708695e-01 -6.59236610e-01 2.92550236e-01
5.04574120e-01 3.29646915e-01 5.99339485e-01 -3.56126130e-01
6.85416281e-01 1.37517560e+00 7.29012847e-01 6.58755898e-01
-8.03964913e-01 -7.13053286e-01 -5.18875241e-01 3.93719882e-01
-1.05543554e+00 -2.17452675e-01 8.54407430e-01 -3.94713104e-01
8.51253152e-01 4.50879961e-01 7.44870901e-01 1.29943717e+00
7.08614349e-01 7.54078627e-01 1.14699030e+00 -7.24081159e-01
5.69894537e-02 -1.41119242e-01 -3.70095104e-01 7.72750497e-01
4.29155529e-01 1.10920206e-01 -3.62227671e-02 -2.19714269e-01
3.31997454e-01 1.09532386e-01 -2.74144530e-01 -4.77241315e-02
-1.36744761e+00 1.02879536e+00 -2.44123071e-01 7.26211786e-01
8.77705868e-03 2.56814033e-01 7.73267627e-01 4.68154103e-01
2.11355656e-01 2.22682267e-01 3.72110009e-02 -1.41035572e-01
-7.48276114e-01 -9.49753001e-02 8.94432664e-01 4.34254199e-01
7.19034016e-01 3.40630487e-02 1.16338789e-01 3.68225276e-01
5.64151406e-01 5.83784521e-01 7.81402588e-01 -1.47430584e-01
1.54578596e-01 5.55024803e-01 -2.51368403e-01 -1.66698325e+00
-3.79754007e-01 -9.74983573e-02 -9.32485938e-01 8.13973974e-03
3.27916294e-01 -1.92745164e-01 -9.85516667e-01 1.24036765e+00
2.47828364e-01 5.88476717e-01 4.19986635e-01 6.59501016e-01
5.04936278e-01 8.24210227e-01 -3.33950490e-01 -1.61181524e-01
1.45490646e+00 -4.01080340e-01 -7.67434120e-01 7.69588575e-02
8.49333465e-01 -8.25298369e-01 3.55980575e-01 2.79124439e-01
-2.98597991e-01 -1.08423300e-01 -1.39946473e+00 3.34328800e-01
-5.51239789e-01 -3.72691035e-01 2.97069490e-01 1.45297682e+00
-6.75291121e-01 3.65806490e-01 -6.43591940e-01 -4.24645692e-01
3.01073268e-02 1.97015077e-01 -4.84452322e-02 5.51579893e-02
-1.62667012e+00 6.29065871e-01 8.79182637e-01 9.89784598e-02
-1.17460763e+00 6.51529655e-02 -7.61799753e-01 -3.19319926e-02
1.29843146e-01 -9.46997106e-02 6.25608742e-01 -9.69097137e-01
-1.61242259e+00 8.38799596e-01 9.68082100e-02 -5.31903207e-01
3.15889657e-01 4.35143888e-01 -5.71177900e-01 3.71187121e-01
-4.05170500e-01 8.63372684e-02 9.54853356e-01 -8.65476906e-01
-5.33676028e-01 -4.49766219e-02 -4.84242767e-01 -2.93941081e-01
-5.18011749e-02 -2.21435457e-01 1.42762959e-01 -5.46875894e-01
3.37047353e-02 -9.47462082e-01 -6.75159544e-02 -3.94163340e-01
-4.73293960e-01 5.79274856e-02 1.21734083e+00 -8.63512158e-01
1.05145478e+00 -2.04709697e+00 -9.93367750e-03 5.25449932e-01
1.40973791e-01 8.69305730e-01 -2.27724373e-01 9.22710717e-01
1.19499274e-01 1.88070312e-01 -4.92008328e-01 4.56551671e-01
-2.46626005e-01 4.15907949e-02 -1.72407612e-01 9.19899344e-01
-1.10922918e-01 6.76598430e-01 -1.14215100e+00 -3.25053841e-01
2.41842315e-01 7.93017268e-01 -2.07415432e-01 2.80268416e-02
-4.45638299e-01 5.37011266e-01 -6.43002570e-01 5.01693785e-01
9.04345512e-01 -2.24467605e-01 9.42939162e-01 3.28989655e-01
1.20829329e-01 1.08812563e-01 -6.70963883e-01 1.36915159e+00
-8.31459761e-02 8.86668861e-01 -3.78857553e-01 -1.32876515e+00
1.25158954e+00 4.31860656e-01 2.28798315e-01 -2.52390504e-01
2.12976068e-01 6.03408933e-01 4.11572933e-01 -9.87806141e-01
1.99068621e-01 -7.49042705e-02 1.93250477e-02 5.79037547e-01
-1.33750826e-01 4.57949698e-01 -1.17830768e-01 -1.49561793e-01
1.45681906e+00 1.23303145e-01 4.95025963e-01 -4.03513759e-03
8.36520910e-01 -2.72515565e-01 4.66440320e-01 9.14706886e-01
-1.64834946e-01 3.26942801e-01 7.96607792e-01 -2.67700136e-01
-1.40389037e+00 -2.22537011e-01 2.45404840e-01 6.80516899e-01
7.15096667e-02 -1.69452950e-02 -8.82856190e-01 -5.70792377e-01
1.07157528e-02 2.75979519e-01 -5.58546782e-01 -3.17842185e-01
-6.85913265e-01 -1.10206354e+00 1.26448929e+00 -2.48652622e-01
5.52557051e-01 -1.15929878e+00 -7.73507237e-01 3.93604249e-01
-1.78698882e-01 -6.74411535e-01 -2.11324841e-01 -4.88730296e-02
-5.48338771e-01 -1.21467912e+00 -9.67418432e-01 -7.50122428e-01
4.01032150e-01 2.83036292e-01 3.18547338e-01 4.33236659e-01
-4.78370428e-01 -1.97057217e-01 -6.44215941e-01 -3.12309235e-01
-1.16117418e+00 3.17476749e-01 -2.78551877e-01 1.69101968e-01
7.49269724e-01 -5.29486597e-01 -6.96350634e-01 1.45168155e-01
-1.17024827e+00 -1.80697143e-01 8.99684191e-01 8.55736077e-01
-1.94349945e-01 2.12163135e-01 5.63861847e-01 -1.27601945e+00
4.36401337e-01 -9.36516225e-01 -5.95576406e-01 2.88017273e-01
-5.03531873e-01 4.20589060e-01 5.17333508e-01 -3.52204531e-01
-7.07014501e-01 -1.51114553e-01 -4.17434216e-01 1.49646372e-01
-8.93011615e-02 6.56771839e-01 -1.46593735e-01 -2.77587146e-01
4.79074478e-01 1.11050284e+00 3.70114982e-01 -2.22473264e-01
-1.02778964e-01 1.09628260e+00 1.37063384e-01 3.21649164e-02
6.67835832e-01 6.05729043e-01 9.54103917e-02 -9.85263228e-01
5.93600608e-03 -4.71329540e-01 -2.58209944e-01 -6.28941804e-02
6.97859466e-01 -5.15133202e-01 -1.16664171e+00 9.25068974e-01
-1.07090712e+00 2.21583635e-01 6.39603913e-01 3.90129715e-01
-3.53454530e-01 1.12843454e+00 -6.19251251e-01 -1.26166070e+00
-2.18452632e-01 -1.14217961e+00 9.21378136e-01 -1.96707547e-01
1.28609657e-01 -9.98320222e-01 5.24340153e-01 -1.12479869e-02
3.61962855e-01 8.19750249e-01 9.83908772e-01 -8.18240523e-01
-5.51765263e-01 -3.99894893e-01 1.63387120e-01 -1.13370776e-01
2.63461053e-01 -8.11310560e-02 -8.44689190e-01 -4.73081678e-01
7.20912516e-02 9.94993001e-03 1.24057031e+00 1.27798378e-01
7.89333701e-01 -4.86809820e-01 -6.65637493e-01 5.50807476e-01
1.54795170e+00 7.12007761e-01 1.21990550e+00 5.62617779e-01
5.44222593e-01 6.09641910e-01 1.35579258e-01 3.77437770e-01
-1.60960034e-01 3.73556674e-01 3.90593886e-01 2.24809900e-01
2.80325949e-01 -5.31684101e-01 6.55800641e-01 9.53293860e-01
5.03387563e-02 -8.71497929e-01 -8.95104408e-01 3.45885098e-01
-1.66936576e+00 -1.30245984e+00 -1.02558471e-01 1.97610545e+00
7.37099469e-01 9.47405994e-02 1.25935897e-01 4.89124686e-01
1.27063596e+00 3.67800683e-01 -5.41276276e-01 -4.89678770e-01
-2.67734438e-01 -7.97102749e-02 7.55213857e-01 4.23218638e-01
-9.77986157e-01 5.26688099e-01 6.69884443e+00 9.45260108e-01
-1.20603216e+00 -4.81158160e-02 2.92485148e-01 5.41329205e-01
-6.10356748e-01 -1.24648502e-02 -5.07651865e-01 8.03693891e-01
1.20479786e+00 1.60822511e-01 2.27188051e-01 4.56797093e-01
-6.19715303e-02 1.06530420e-01 -5.50369263e-01 7.89372742e-01
2.83890694e-01 -1.52808416e+00 8.29454139e-02 2.69099146e-01
5.86491227e-01 -2.99856603e-01 8.23391229e-02 -1.66524902e-01
1.43854246e-01 -1.06165123e+00 4.73714694e-02 4.07664597e-01
7.99056649e-01 -8.05021405e-01 1.21107924e+00 3.51572186e-01
-8.59668732e-01 1.04058780e-01 -6.22079670e-01 -6.64688507e-03
-1.38338447e-01 7.70887494e-01 -1.41517210e+00 4.14419413e-01
3.76043543e-02 7.67401934e-01 -5.13038188e-02 7.21497536e-01
-2.62698054e-01 8.68917048e-01 -6.04802649e-03 -7.08838701e-01
2.66981453e-01 5.48807420e-02 7.86700964e-01 1.60749686e+00
7.02434957e-01 -2.31820181e-01 -4.56178412e-02 4.56153691e-01
1.05173051e-01 -3.43006179e-02 -9.05146956e-01 -6.58198833e-01
4.29433554e-01 7.68990397e-01 -1.06248009e+00 -3.58185291e-01
-2.79095709e-01 9.01749611e-01 -3.77050489e-01 1.01832405e-01
-4.29172844e-01 -1.02963972e+00 4.98057991e-01 -8.04230496e-02
6.23782337e-01 -8.57807174e-02 1.51323527e-01 -1.10345757e+00
-4.28673238e-01 -1.31341445e+00 2.85565317e-01 -1.12845369e-01
-9.04325545e-01 2.62740612e-01 -3.97786915e-01 -1.18004227e+00
-3.93333137e-01 -5.24260581e-01 -1.88289896e-01 6.52436256e-01
-1.46783078e+00 -7.99220979e-01 1.18082866e-01 1.70564473e-01
3.13589633e-01 -5.63658416e-01 8.71078074e-01 1.39707088e-01
-4.05914634e-01 5.72186589e-01 7.08867729e-01 4.60037500e-01
3.45341355e-01 -8.30125093e-01 7.82804549e-01 1.04360652e+00
-3.28726530e-01 5.90897858e-01 1.08828294e+00 -1.02135301e+00
-1.64223158e+00 -6.76736891e-01 7.62326062e-01 -2.15480756e-02
3.73174787e-01 -7.25537121e-01 -8.54807556e-01 6.39275432e-01
2.03388035e-01 -4.63470489e-01 7.83771574e-01 -5.42374611e-01
-4.20910954e-01 4.30073172e-01 -1.36118555e+00 3.66457134e-01
7.31493294e-01 -5.24351537e-01 -3.99042666e-01 1.64839372e-01
6.29693091e-01 -1.44405514e-01 -4.46140736e-01 2.82602329e-02
1.08777833e+00 -9.78346109e-01 6.54232085e-01 -2.85122871e-01
3.90364558e-01 -5.86793721e-01 -6.62591830e-02 -1.10436320e+00
-2.00981274e-01 -1.02002025e+00 -4.15290818e-02 7.01171100e-01
1.96085796e-01 -9.65991735e-01 9.13118303e-01 -6.65090740e-01
2.64155984e-01 -1.79703787e-01 -9.91739511e-01 -5.53864539e-01
-2.61594832e-01 1.22574538e-01 7.37046659e-01 1.03724062e+00
2.71632433e-01 -7.89530575e-02 -1.13298202e+00 9.65081900e-02
6.91813767e-01 -9.98355299e-02 6.03961289e-01 -9.66939747e-01
-4.57389772e-01 -1.72141567e-01 -1.08513534e+00 -8.23898375e-01
-4.65589501e-02 -9.94034171e-01 2.54648238e-01 -9.05632675e-01
3.22512358e-01 -3.46519798e-02 -3.67711067e-01 5.74050546e-02
-1.46825731e-01 4.55636322e-01 -1.79846287e-01 1.15593590e-01
-1.83803305e-01 9.96715277e-02 9.45754290e-01 -2.44654790e-01
1.00324929e-01 -1.87766194e-01 -4.92416650e-01 4.28404868e-01
8.62789094e-01 -7.96801329e-01 -1.24746844e-01 1.22396294e-02
5.23200631e-01 1.12422504e-01 5.10128289e-02 -1.01277518e+00
-1.51865594e-02 -1.25950086e-03 3.22306603e-01 -4.46281940e-01
-4.91875596e-02 -6.20833457e-01 4.55584168e-01 1.44015872e+00
-5.06306767e-01 -1.58780232e-01 -1.62912220e-01 9.79332685e-01
1.36799544e-01 -3.85493815e-01 7.08056867e-01 -4.43372041e-01
-6.05009377e-01 -3.10671866e-01 -1.25306880e+00 -5.05214751e-01
1.08285069e+00 -5.95941544e-01 -4.23861384e-01 -3.40795666e-01
-1.42549038e-01 -3.42518836e-01 3.39714348e-01 2.02677384e-01
8.63600969e-01 -8.59065533e-01 -9.52257693e-01 4.64475542e-01
1.55360267e-01 -9.51773226e-01 2.90353060e-01 5.51830351e-01
-1.01798689e+00 7.65797019e-01 -4.88620698e-01 -6.61523879e-01
-1.44274783e+00 7.80704081e-01 1.72246888e-01 -2.57769436e-01
-7.37473249e-01 5.32801390e-01 -1.08329706e-01 -4.31329310e-01
-1.08992502e-01 1.73843522e-02 -4.83361393e-01 -1.62100613e-01
8.11352193e-01 4.11353528e-01 -2.19675228e-01 -5.58401942e-01
-3.31172228e-01 5.45344770e-01 -1.21405840e-01 1.15273915e-01
1.05194974e+00 -1.71762541e-01 -3.73831004e-01 5.97812951e-01
1.55975461e+00 -9.81460810e-02 -6.95434690e-01 -7.14251623e-02
1.99388295e-01 -6.57777727e-01 -5.07742524e-01 -4.31624681e-01
-8.15699637e-01 9.57103252e-01 8.86714876e-01 3.33722919e-01
8.46225083e-01 -4.58604932e-01 1.08859432e+00 4.39511567e-01
4.72647548e-01 -6.78590357e-01 -1.23350471e-01 2.92825699e-01
1.19088821e-01 -9.31502581e-01 -2.71789849e-01 -3.23169023e-01
-1.39130265e-01 1.48450434e+00 -6.60236925e-02 -1.70752108e-01
3.10363531e-01 2.44596794e-01 1.42576620e-01 -7.21406192e-02
-6.82480872e-01 7.28216246e-02 -2.74569303e-01 9.53680873e-01
4.32328165e-01 2.27617174e-02 -7.61060715e-01 -5.90220153e-01
6.69230223e-02 1.08982585e-01 8.33032548e-01 9.84070599e-01
-9.98044729e-01 -1.26135337e+00 -6.67679012e-01 2.73844212e-01
-9.12153244e-01 2.50008591e-02 -3.05747807e-01 4.43615347e-01
3.44055891e-02 9.82548714e-01 -2.10120246e-01 -7.90686846e-01
-4.63915259e-01 1.33274198e-01 1.86609492e-01 -4.28785086e-01
-3.14923674e-01 1.43184617e-01 -1.86333105e-01 -1.43672684e-02
-7.11308777e-01 -6.18735492e-01 -9.54151928e-01 -5.74070156e-01
-3.36943895e-01 2.99261421e-01 6.44644797e-01 9.51852083e-01
2.24102169e-01 5.33394039e-01 7.56838262e-01 -4.05318290e-01
-3.17960441e-01 -9.30827200e-01 -6.27961516e-01 1.02778092e-01
7.57989466e-01 -3.05383414e-01 -4.75847334e-01 -1.07909434e-01] | [4.305810928344727, 8.04938793182373] |
4e88dac4-e285-4095-808b-3efe127fa741 | groupnet-multiscale-hypergraph-neural | 2204.08770 | null | https://arxiv.org/abs/2204.08770v2 | https://arxiv.org/pdf/2204.08770v2.pdf | GroupNet: Multiscale Hypergraph Neural Networks for Trajectory Prediction with Relational Reasoning | Demystifying the interactions among multiple agents from their past trajectories is fundamental to precise and interpretable trajectory prediction. However, previous works only consider pair-wise interactions with limited relational reasoning. To promote more comprehensive interaction modeling for relational reasoning, we propose GroupNet, a multiscale hypergraph neural network, which is novel in terms of both interaction capturing and representation learning. From the aspect of interaction capturing, we propose a trainable multiscale hypergraph to capture both pair-wise and group-wise interactions at multiple group sizes. From the aspect of interaction representation learning, we propose a three-element format that can be learnt end-to-end and explicitly reason some relational factors including the interaction strength and category. We apply GroupNet into both CVAE-based prediction system and previous state-of-the-art prediction systems for predicting socially plausible trajectories with relational reasoning. To validate the ability of relational reasoning, we experiment with synthetic physics simulations to reflect the ability to capture group behaviors, reason interaction strength and interaction category. To validate the effectiveness of prediction, we conduct extensive experiments on three real-world trajectory prediction datasets, including NBA, SDD and ETH-UCY; and we show that with GroupNet, the CVAE-based prediction system outperforms state-of-the-art methods. We also show that adding GroupNet will further improve the performance of previous state-of-the-art prediction systems. | ['Siheng Chen', 'Ya zhang', 'Zhenyang Ni', 'Maosen Li', 'Chenxin Xu'] | 2022-04-19 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Xu_GroupNet_Multiscale_Hypergraph_Neural_Networks_for_Trajectory_Prediction_With_Relational_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Xu_GroupNet_Multiscale_Hypergraph_Neural_Networks_for_Trajectory_Prediction_With_Relational_CVPR_2022_paper.pdf | cvpr-2022-1 | ['relational-reasoning'] | ['natural-language-processing'] | [-4.36917245e-01 1.67674571e-01 -2.68199295e-01 -2.44105414e-01
2.78264266e-02 -3.27091932e-01 8.39285672e-01 1.19954780e-01
2.10911945e-01 6.91519141e-01 5.28262734e-01 -4.54868019e-01
-6.50563598e-01 -1.22722208e+00 -1.03417516e+00 -3.71585935e-01
-7.58557022e-01 9.99170899e-01 3.10184091e-01 -6.17754579e-01
-2.74003237e-01 4.48376626e-01 -1.50579727e+00 1.99117213e-01
1.17452991e+00 4.42961007e-01 -3.97219062e-01 6.60543919e-01
4.32016701e-01 1.02849483e+00 -3.25633019e-01 -5.36910474e-01
2.64638782e-01 -3.81437510e-01 -6.02455199e-01 -3.77416819e-01
3.48352380e-02 -5.04832685e-01 -1.02582681e+00 5.64054191e-01
1.33845612e-01 5.08536816e-01 8.97151470e-01 -1.86028004e+00
-8.57964814e-01 1.19721723e+00 -2.71397144e-01 1.07635856e-01
5.43766856e-01 5.65844715e-01 1.07857621e+00 -3.52356166e-01
6.25558019e-01 1.60598814e+00 8.70958149e-01 4.99168336e-01
-1.11889601e+00 -9.11747098e-01 5.97942352e-01 3.79620522e-01
-1.07195222e+00 1.44091457e-01 6.84868276e-01 -5.13375223e-01
1.23231411e+00 2.46524647e-01 1.06570899e+00 1.29282427e+00
3.21373999e-01 6.26094520e-01 5.71465313e-01 4.01777089e-01
1.37588009e-01 -4.25102085e-01 4.12220985e-01 8.37997675e-01
2.51722544e-01 5.17396748e-01 -3.07925999e-01 -3.42315286e-01
8.75092089e-01 4.08053666e-01 -1.34331763e-01 -1.83575451e-01
-1.30019224e+00 7.36947775e-01 9.28498507e-01 -1.48007572e-01
-3.40139389e-01 6.20153546e-01 1.24549598e-01 1.81067973e-01
5.13900161e-01 2.05104291e-01 -3.92058671e-01 -2.66460240e-01
-2.97355056e-01 1.00151730e+00 9.62309301e-01 8.27195823e-01
4.78568822e-01 -1.12788744e-01 -4.60346788e-01 4.13623691e-01
4.51946139e-01 6.04466856e-01 -2.53025591e-01 -1.03233695e+00
5.03295422e-01 1.22734189e+00 9.45288390e-02 -1.29907179e+00
-7.87996590e-01 -3.04851532e-01 -1.10594654e+00 -1.24794757e-02
4.43446934e-01 -3.67658973e-01 -7.06570685e-01 2.12764287e+00
3.94251376e-01 9.79288936e-01 1.71246096e-01 9.55374777e-01
1.25114548e+00 7.98122287e-01 -4.05258425e-02 1.25891408e-02
9.16992307e-01 -1.27561259e+00 -3.97571206e-01 1.93837360e-01
8.77414107e-01 1.73070908e-01 9.32763696e-01 -1.06201628e-02
-1.03643477e+00 -4.77212876e-01 -7.35010624e-01 1.19972870e-01
-4.92385328e-01 -2.73403317e-01 1.06461215e+00 2.74482518e-01
-9.87135053e-01 7.89068878e-01 -1.05125308e+00 -5.68602324e-01
2.46077910e-01 6.02072358e-01 -5.47978878e-02 3.60371262e-01
-1.44714999e+00 6.99701250e-01 4.13095616e-02 -2.11416095e-01
-9.55690920e-01 -1.20403767e+00 -6.00115180e-01 3.79527152e-01
3.41382056e-01 -1.05978739e+00 8.88160408e-01 -1.03214689e-01
-1.41650999e+00 1.11526825e-01 1.19360676e-02 -6.99724793e-01
6.33428276e-01 9.39352959e-02 -5.19052267e-01 -1.02854282e-01
-1.65628701e-01 6.79793119e-01 1.50913179e-01 -1.30298460e+00
-3.56251150e-01 -1.26587274e-02 5.71230888e-01 3.32627475e-01
-8.08987245e-02 -3.11814964e-01 -2.73652822e-01 -4.93943542e-01
-4.93860215e-01 -1.31068277e+00 -2.09026799e-01 -9.59740728e-02
-4.19530064e-01 -6.77009344e-01 8.63405526e-01 -3.87664169e-01
1.08781910e+00 -1.63504934e+00 5.55208921e-01 2.96927541e-01
7.59946227e-01 1.47634625e-01 -2.31664643e-01 7.33321607e-01
1.60393089e-01 1.91664994e-01 -1.74917504e-01 -5.15490353e-01
3.27343762e-01 3.40253413e-01 -5.29541135e-01 2.21611843e-01
-2.35113055e-01 1.23472750e+00 -1.08073437e+00 -1.70566320e-01
3.22845936e-01 5.24569631e-01 -9.10489082e-01 3.70756388e-01
-5.64970970e-01 6.69125617e-01 -5.65065682e-01 1.99078903e-01
6.29464626e-01 -5.86000443e-01 1.98365375e-01 7.21318796e-02
2.29603380e-01 1.37017295e-01 -8.24701190e-01 1.11075556e+00
-6.90019950e-02 4.93032277e-01 -3.38175803e-01 -6.96938455e-01
6.54583871e-01 6.65097907e-02 7.51947761e-01 -3.92802984e-01
3.01218573e-02 -2.73252964e-01 3.43150854e-01 -2.22644076e-01
4.10107851e-01 4.86204922e-01 -1.87277153e-01 6.58609986e-01
-3.67268831e-01 2.36460567e-01 2.23431706e-01 6.84695244e-01
1.35681725e+00 2.84367949e-02 -5.52288666e-02 -1.38534233e-01
3.22467327e-01 -3.03490777e-02 5.57137251e-01 8.28190625e-01
-1.11042380e-01 1.30896002e-01 5.40879905e-01 -7.78292775e-01
-7.47732937e-01 -1.20659268e+00 2.61937737e-01 1.12475646e+00
6.42570972e-01 -7.77638137e-01 -6.63145900e-01 -7.20852494e-01
3.80627245e-01 7.10343182e-01 -9.30320680e-01 -3.67920965e-01
-6.62515581e-01 -8.78527045e-01 6.58644855e-01 6.20595753e-01
6.13607466e-01 -1.07445097e+00 -2.09257398e-02 4.30001728e-02
-1.95623457e-01 -1.02737212e+00 -4.28145945e-01 -6.66555583e-01
-3.46667677e-01 -1.15761495e+00 -4.43574749e-02 -3.54981959e-01
4.09968287e-01 3.48104745e-01 1.10369277e+00 6.98854446e-01
2.34811872e-01 4.33224112e-01 -2.70204008e-01 -1.24843545e-01
-3.83699149e-01 7.81617165e-02 4.93955374e-01 -3.50608081e-01
1.33426204e-01 -1.06483829e+00 -7.71837890e-01 5.99527240e-01
-4.01633799e-01 4.01529640e-01 1.28693953e-01 3.90751660e-01
1.19916379e-01 2.48110980e-01 5.17004550e-01 -6.03103042e-01
8.82327735e-01 -7.94618070e-01 -3.97981584e-01 2.07705542e-01
-4.45829898e-01 -1.77685753e-01 8.66718411e-01 -6.20752633e-01
-9.31313157e-01 -5.55852711e-01 9.21509638e-02 -5.50691009e-01
-4.73038033e-02 5.72829127e-01 9.19254348e-02 -4.91893180e-02
5.43413699e-01 -8.62352829e-03 -3.85870486e-02 -8.94735008e-02
6.00067437e-01 1.33059844e-01 2.56856263e-01 -7.71090329e-01
7.67203271e-01 3.53520870e-01 2.85392076e-01 -5.37403345e-01
-4.24138039e-01 4.59015705e-02 -3.10376823e-01 -5.37039340e-01
9.19155657e-01 -9.65629935e-01 -1.80366015e+00 5.39851487e-01
-1.16470766e+00 -9.09796655e-01 8.63879919e-02 4.82195824e-01
-7.54048705e-01 8.04188028e-02 -9.22543168e-01 -6.34959102e-01
-2.00367674e-01 -1.23943591e+00 8.06126952e-01 1.71128407e-01
-2.83585489e-01 -1.23523414e+00 3.85273904e-01 3.94560903e-01
2.09713504e-01 5.13172269e-01 1.00917387e+00 -7.46026576e-01
-1.03134108e+00 2.93009669e-01 -2.77184308e-01 -6.31682098e-01
-1.11125372e-01 1.77240193e-01 -2.95721233e-01 -2.30031416e-01
-8.02756667e-01 -9.97515842e-02 8.47906888e-01 3.62576127e-01
1.16942680e+00 -4.85568076e-01 -9.13117409e-01 6.27880812e-01
7.67902970e-01 1.09137058e-01 4.81388658e-01 -2.83378601e-01
1.10906827e+00 5.10186493e-01 2.49275401e-01 3.87292653e-01
1.31969976e+00 9.00432050e-01 3.96133810e-01 1.13219559e-01
-2.01066777e-01 -5.80140114e-01 4.05369222e-01 8.72935057e-01
-5.50185919e-01 -7.79334188e-01 -1.00265956e+00 2.27670416e-01
-2.75885916e+00 -1.44519043e+00 -2.27849841e-01 1.68209088e+00
4.57922995e-01 8.97011906e-02 7.40602493e-01 -4.36110556e-01
5.41015208e-01 1.12307675e-01 -8.71461928e-01 -1.77053973e-01
2.69286364e-01 -5.82522333e-01 2.66180813e-01 6.78782225e-01
-7.58244216e-01 1.09575963e+00 6.24245644e+00 7.28320062e-01
-5.95375776e-01 -1.69312537e-01 5.08823037e-01 -1.35836497e-01
-5.27020097e-01 -1.84742019e-01 -7.26651728e-01 4.52173948e-01
9.97855902e-01 -2.41082460e-01 1.14379799e+00 5.07481515e-01
3.05167556e-01 4.30781275e-01 -1.48841703e+00 7.38845766e-01
-2.62399733e-01 -1.60985708e+00 2.63492227e-01 3.07868451e-01
8.58257174e-01 1.24936469e-01 1.91247314e-02 8.96701753e-01
1.22419047e+00 -1.15382123e+00 3.66445035e-01 9.56784844e-01
2.34872833e-01 -8.41823101e-01 4.13217932e-01 6.32069468e-01
-1.70036030e+00 -1.30740821e-01 -1.44505069e-01 -5.13545215e-01
3.43645692e-01 7.06956759e-02 -6.81648195e-01 7.77597308e-01
6.82502210e-01 1.50497210e+00 -4.23272014e-01 7.33642697e-01
-8.61209407e-02 7.37650275e-01 -4.67008591e-01 -3.12920243e-01
1.75768986e-01 -5.29039264e-01 7.83030808e-01 8.34556699e-01
3.21219504e-01 5.90816915e-01 5.35325825e-01 1.22347295e+00
-2.25833595e-01 -4.36638504e-01 -7.03864872e-01 -1.47340596e-01
6.89751267e-01 8.48877311e-01 -6.65534019e-01 -4.05104458e-01
-2.23019019e-01 3.69764835e-01 8.03861260e-01 5.55478692e-01
-1.40780985e+00 1.78361163e-01 1.23727012e+00 1.55311421e-01
1.67566687e-01 -2.52490014e-01 -1.11403212e-01 -1.24349415e+00
-4.02525514e-01 -7.15645611e-01 2.04339802e-01 -6.15142405e-01
-1.60351813e+00 4.89047110e-01 5.77412069e-01 -1.10047603e+00
-3.00496906e-01 -2.90941894e-01 -1.17732358e+00 3.43066275e-01
-7.86616266e-01 -1.64075959e+00 -2.78066009e-01 5.27915061e-01
2.13051274e-01 -3.35153461e-01 5.12771845e-01 9.33061987e-02
-7.84922898e-01 5.50709128e-01 -1.13576345e-01 6.89435229e-02
2.63515592e-01 -1.07584834e+00 8.46980512e-01 3.71320844e-01
-1.34777829e-01 7.20124781e-01 8.08383763e-01 -1.17389834e+00
-1.46152937e+00 -1.36052299e+00 3.61758679e-01 -8.75207007e-01
9.05261695e-01 -6.14804864e-01 -8.05273294e-01 1.16267765e+00
8.91940743e-02 -1.66215971e-01 5.84742725e-01 5.72179496e-01
-1.36282235e-01 4.48742434e-02 -9.88178492e-01 1.20610225e+00
1.92603731e+00 -3.70101511e-01 -5.28803885e-01 4.77480650e-01
1.31002784e+00 -4.73664284e-01 -1.07547021e+00 5.75654268e-01
6.02383971e-01 -8.86950672e-01 1.45020378e+00 -8.24664712e-01
5.58908761e-01 -2.34728515e-01 1.28374055e-01 -1.57992804e+00
-6.27368212e-01 -5.13309300e-01 -6.38222098e-01 1.02709472e+00
4.44230884e-01 -9.21862304e-01 7.97764122e-01 8.91796470e-01
-3.88219692e-02 -1.04961240e+00 -5.42298257e-01 -7.88918018e-01
2.57636547e-01 -4.10621822e-01 1.47727740e+00 7.80778825e-01
4.49696779e-01 3.63694787e-01 -7.23774791e-01 4.00034696e-01
7.89331079e-01 3.71741772e-01 1.23171699e+00 -1.44793248e+00
-5.20460188e-01 -6.65684402e-01 -2.99302787e-01 -1.20501685e+00
5.65003932e-01 -9.62418139e-01 -3.18351805e-01 -1.77941847e+00
2.24123493e-01 -4.81787413e-01 8.92477110e-03 2.36010119e-01
-1.67288154e-01 -2.09125072e-01 3.57108116e-01 2.45747492e-01
-9.41813409e-01 9.99168456e-01 1.55457866e+00 -1.93083584e-01
-4.64872777e-01 1.05575420e-01 -4.00500506e-01 7.91937649e-01
6.70627296e-01 -2.67419040e-01 -9.64568019e-01 -1.74415901e-01
3.95644188e-01 4.26788956e-01 6.69608176e-01 -1.10252988e+00
4.99792039e-01 -5.21789670e-01 7.51144290e-02 -8.31625283e-01
6.27706707e-01 -6.64958835e-01 5.05379975e-01 3.89013916e-01
-6.24909699e-01 5.01764454e-02 1.79117188e-01 9.51868117e-01
4.65344667e-01 7.19905198e-01 1.93973053e-02 1.81471527e-01
-2.58352429e-01 9.35032547e-01 -4.38324660e-01 3.85170765e-02
1.20511341e+00 1.63502529e-01 -8.32692742e-01 -7.79149592e-01
-8.37772489e-01 9.65035796e-01 4.90500778e-01 5.27420640e-01
6.51037574e-01 -1.53839219e+00 -6.01865172e-01 7.94793889e-02
-8.20542723e-02 -2.28937402e-01 4.16347593e-01 7.08524585e-01
-3.39455128e-01 3.27978700e-01 -2.46032819e-01 -4.92900193e-01
-1.00047719e+00 7.63388813e-01 3.92591685e-01 -4.66805935e-01
-8.29070866e-01 4.28033531e-01 3.78446847e-01 -1.12260509e+00
1.99435398e-01 -6.58207417e-01 -4.45078313e-01 -5.80496967e-01
4.05800611e-01 7.61493564e-01 -7.60321915e-01 -7.02786624e-01
-2.75213420e-01 4.91693020e-01 3.87119725e-02 1.06347583e-01
1.40022826e+00 2.93053351e-02 4.27438579e-02 2.38372430e-01
7.85334408e-01 -2.76067883e-01 -1.56057751e+00 6.20492138e-02
-5.70530593e-01 -2.24712133e-01 -3.74499977e-01 -7.89513767e-01
-8.73526394e-01 5.99241614e-01 2.87982472e-03 5.05847573e-01
5.60184896e-01 1.62835896e-01 1.03398311e+00 6.93833768e-01
5.49807012e-01 -5.18585443e-01 6.98245093e-02 6.54811680e-01
9.61169183e-01 -1.10468185e+00 -3.86143997e-02 -6.54385984e-01
-6.17399633e-01 6.15671873e-01 1.08466280e+00 -4.32635516e-01
1.17609131e+00 -5.94729632e-02 -6.14253581e-01 -5.52377045e-01
-1.23824704e+00 -2.77974475e-02 5.35321116e-01 5.81263840e-01
-1.31078973e-01 4.88473207e-01 6.99896598e-03 8.27566743e-01
-4.15278256e-01 4.82819369e-03 3.43816251e-01 2.19949558e-01
-1.32102802e-01 -8.65538597e-01 1.56170756e-01 5.93868852e-01
3.14209789e-01 2.43004933e-01 -5.96305966e-01 7.83993065e-01
1.11909159e-01 1.09699988e+00 1.86739996e-01 -9.31199849e-01
1.75450817e-01 -5.65315843e-01 3.16591799e-01 -2.66612351e-01
-8.72902036e-01 -6.53225958e-01 2.66059339e-01 -6.59090698e-01
-3.69553030e-01 -5.23518920e-01 -1.71940196e+00 -1.36275518e+00
2.44809538e-01 1.15437396e-01 -2.21727178e-01 1.05861425e+00
6.57005250e-01 6.38462961e-01 2.75073260e-01 -1.11040747e+00
-1.01628073e-01 -7.75466502e-01 -2.33015925e-01 6.90590084e-01
2.34554768e-01 -1.07741582e+00 -3.65903974e-01 -4.39107835e-01] | [5.8844122886657715, 0.8531733751296997] |
ab048ab7-44d5-4e60-8e11-1a6c1b1541ce | an-mil-derived-transformer-for-weakly | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Yang_An_MIL-Derived_Transformer_for_Weakly_Supervised_Point_Cloud_Segmentation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Yang_An_MIL-Derived_Transformer_for_Weakly_Supervised_Point_Cloud_Segmentation_CVPR_2022_paper.pdf | An MIL-Derived Transformer for Weakly Supervised Point Cloud Segmentation | We address weakly supervised point cloud segmentation by proposing a new model, MIL-derived transformer, to mine additional supervisory signals. First, the transformer model is derived based on multiple instance learning (MIL) to explore pair-wise cloud-level supervision, where two clouds of the same category yield a positive bag while two of different classes produce a negative bag. It leverages not only individual cloud annotations but also pair-wise cloud semantics for model optimization. Second, Adaptive global weighted pooling (AdaGWP) is integrated into our transformer model to replace max pooling and average pooling. It introduces learnable weights to re-scale logits in the class activation maps. It is more robust to noise while discovering more complete foreground points under weak supervision. Third, we perform point subsampling and enforce feature equivariance between the original and subsampled point clouds for regularization. The proposed method is end-to-end trainable and is general because it can work with different backbones with diverse types of weak supervision signals, including sparsely annotated points and cloud-level labels. The experiments show that it achieves state-of-the-art performance on the S3DIS and ScanNet benchmarks. | ['Yen-Yu Lin', 'Yung-Yu Chuang', 'Kai-Syun Chen', 'Ji-Jia Wu', 'Cheng-Kun Yang'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['point-cloud-segmentation'] | ['computer-vision'] | [ 2.76758641e-01 1.15969047e-01 -3.63885403e-01 -6.72833443e-01
-1.26173711e+00 -5.00248671e-01 4.82205212e-01 2.36140322e-02
-2.16563687e-01 2.94095874e-01 -2.08308682e-01 8.32673013e-02
-2.69238241e-02 -7.08515823e-01 -1.28218651e+00 -8.09925318e-01
-1.48057081e-02 6.40597880e-01 8.48532796e-01 3.23155560e-02
2.08956301e-01 6.40925050e-01 -1.39424264e+00 6.96806252e-01
9.15870905e-01 1.30013764e+00 3.85449737e-01 1.93991020e-01
-3.08670610e-01 6.48451686e-01 -3.42348546e-01 -4.03596103e-01
6.50275350e-01 3.47436458e-01 -7.13756442e-01 3.25934052e-01
9.14182484e-01 1.10890135e-01 6.20077457e-03 1.01159728e+00
1.94324538e-01 -8.48651603e-02 4.72010523e-01 -1.39670932e+00
-5.05304575e-01 4.57343251e-01 -1.10801017e+00 5.60254557e-03
-3.30155045e-01 3.06057423e-01 1.25224793e+00 -1.35493290e+00
3.31411093e-01 1.35127783e+00 6.51813149e-01 2.59827346e-01
-1.33988059e+00 -8.59046400e-01 9.38646138e-01 -9.58198681e-02
-1.28825760e+00 3.82339960e-04 9.87741113e-01 -4.35797542e-01
6.85241401e-01 2.85237968e-01 5.92373669e-01 7.59468198e-01
-3.35573286e-01 1.10764325e+00 1.03438973e+00 -1.75050851e-02
1.97114930e-01 3.55914950e-01 2.60545999e-01 6.93927467e-01
8.32464024e-02 -3.04009378e-01 -6.64750218e-01 -2.01555416e-01
8.38782668e-01 4.14412171e-01 -4.12787534e-02 -6.92976594e-01
-1.28106129e+00 7.36122906e-01 1.06814969e+00 -3.32464289e-04
-2.04975769e-01 2.64196336e-01 -1.75432267e-03 -4.50788550e-02
8.38389695e-01 4.04305637e-01 -7.51263440e-01 5.26886463e-01
-9.68445361e-01 3.82954478e-01 4.72912014e-01 1.36567485e+00
1.24585831e+00 -1.36584505e-01 -3.09965730e-01 8.38876963e-01
3.52846652e-01 5.16238451e-01 1.14890330e-01 -8.83412123e-01
8.35918844e-01 9.81216490e-01 -9.51251760e-02 -8.23299289e-01
-9.69496742e-02 -6.80306613e-01 -4.69531745e-01 2.43834242e-01
9.74191278e-02 1.85465917e-01 -1.36157382e+00 1.50230670e+00
3.89615953e-01 9.64177907e-01 -3.06661218e-01 9.85268772e-01
8.97571683e-01 4.34104025e-01 1.30445048e-01 3.76595318e-01
1.21756029e+00 -1.35986602e+00 -1.03681087e-01 -6.59175754e-01
2.07575709e-01 -2.62746811e-01 1.28993952e+00 1.41404018e-01
-1.06522298e+00 -5.28046191e-01 -1.02358842e+00 -2.02411208e-02
-5.48204958e-01 6.63090795e-02 6.97004259e-01 2.45622218e-01
-7.60299563e-01 5.87191761e-01 -1.07108057e+00 1.82806656e-01
1.01573229e+00 3.03382069e-01 -1.45258904e-01 -1.19364537e-01
-6.93538904e-01 4.61686313e-01 9.41962674e-02 1.22709617e-01
-9.12487686e-01 -1.18222094e+00 -8.47648680e-01 -6.50271121e-03
5.25570989e-01 -6.01588070e-01 8.33356857e-01 -8.88994873e-01
-1.02391827e+00 1.11785567e+00 -2.92005092e-01 -3.25998873e-01
4.46208000e-01 -2.02833280e-01 -2.44748797e-02 2.02003986e-01
4.40400690e-01 9.38760519e-01 1.04213798e+00 -1.74262226e+00
-8.37524116e-01 -7.02608168e-01 -9.36424360e-02 1.75998271e-01
-7.40573555e-02 -1.32451802e-01 -8.26715887e-01 -7.64950335e-01
5.89432001e-01 -9.17236447e-01 -3.88638765e-01 1.05622485e-01
-5.47752678e-01 -1.80112451e-01 1.08494318e+00 -2.35609487e-01
6.26497447e-01 -2.10591006e+00 1.46865875e-01 4.19158757e-01
3.22878927e-01 -2.00716719e-01 -1.76901236e-01 -1.72493234e-01
-3.29480879e-02 1.43547922e-01 -7.59763837e-01 -8.17689717e-01
-2.96203722e-03 3.65345329e-01 -5.82632303e-01 4.20957178e-01
9.85321105e-01 1.02821970e+00 -9.40534174e-01 -5.37399948e-01
2.06101164e-01 3.26980382e-01 -6.36803269e-01 2.08371073e-01
-3.89472693e-01 5.16894937e-01 -6.04195297e-01 1.23170471e+00
1.03412044e+00 -5.47142267e-01 -4.04464722e-01 -2.52328128e-01
5.09929508e-02 7.54893124e-02 -1.20982468e+00 1.87744498e+00
-2.54337788e-01 1.52605936e-01 2.61965662e-01 -8.94202650e-01
9.23948526e-01 -1.66330278e-01 4.76652980e-01 -1.63209811e-01
-1.44672930e-01 2.91635901e-01 -4.68343049e-01 -2.81110257e-01
2.29951948e-01 -1.52682558e-01 -9.39765051e-02 -5.36491759e-02
4.06908959e-01 -6.03052199e-01 -3.97022307e-01 1.61992997e-01
9.06800568e-01 4.23944086e-01 -3.56277078e-01 -2.98031688e-01
3.74782920e-01 -7.51818195e-02 8.49008679e-01 7.36027002e-01
-1.58921495e-01 9.43417132e-01 2.84743339e-01 -2.55732685e-01
-5.77062964e-01 -1.49952400e+00 -2.72340626e-01 1.23603249e+00
6.37263119e-01 -1.25781491e-01 -5.71794331e-01 -9.28263128e-01
3.80122125e-01 3.82810473e-01 -4.89698470e-01 -2.90041082e-02
-6.18948758e-01 -7.80002713e-01 2.99782425e-01 8.87675464e-01
3.35586190e-01 -1.02297974e+00 -2.62957901e-01 3.38557139e-02
-7.08349794e-02 -1.31594491e+00 -3.53264302e-01 6.12797558e-01
-1.06402266e+00 -1.13732219e+00 -5.60807645e-01 -8.70581925e-01
7.27055013e-01 4.31763172e-01 1.25698042e+00 -2.83580162e-02
-2.87888329e-02 3.50843102e-01 -1.67276338e-01 -7.16764390e-01
4.24673259e-01 1.78864181e-01 -9.89433080e-02 3.77047807e-01
5.21190703e-01 -6.37031794e-01 -3.76727462e-01 5.04873216e-01
-8.02438796e-01 -1.82381943e-01 6.97035074e-01 6.71370268e-01
1.23338830e+00 -2.49886662e-01 4.86071318e-01 -1.08185184e+00
5.83186233e-03 -5.88295758e-01 -4.94157404e-01 2.31792971e-01
-3.53872120e-01 -1.58567116e-01 2.48560086e-01 -3.60587031e-01
-9.45308447e-01 3.22919458e-01 2.11512551e-01 -1.07853520e+00
-1.81576312e-01 1.30341843e-01 -3.87092859e-01 -3.63682300e-01
3.79257977e-01 -8.63574967e-02 -3.34523648e-01 -6.50922120e-01
5.42729199e-01 3.70977461e-01 8.22460055e-01 -8.63848269e-01
1.22832263e+00 9.54774976e-01 -2.23250955e-01 -6.73733652e-01
-1.15071559e+00 -8.52232397e-01 -9.57288563e-01 -1.44030273e-01
1.01997662e+00 -1.24436462e+00 -1.77665010e-01 5.23217916e-01
-1.04564166e+00 -2.45225623e-01 -4.57283825e-01 2.08809510e-01
-4.17867094e-01 1.49715645e-03 -4.49402601e-01 -6.94912195e-01
-1.38811916e-01 -1.33068848e+00 1.75871205e+00 1.90309972e-01
4.08342898e-01 -6.47680461e-01 -3.97555351e-01 4.75788265e-01
-4.30230945e-02 4.45592105e-01 5.39462268e-01 -6.40875459e-01
-1.19701433e+00 -1.56160384e-01 -4.00820106e-01 5.44090867e-01
1.62967795e-03 -1.66363984e-01 -1.38834536e+00 -3.51539731e-01
3.53818797e-02 -4.29810852e-01 1.32688332e+00 2.02316791e-01
1.77632558e+00 -8.42520073e-02 -5.58043659e-01 1.02168572e+00
1.30579555e+00 -4.50099766e-01 4.40543652e-01 2.50773758e-01
1.29114342e+00 6.66187525e-01 6.93174720e-01 5.52143268e-02
5.90264916e-01 5.43127656e-01 8.92848313e-01 -4.98465002e-01
8.38315636e-02 -2.93613940e-01 1.15082584e-01 5.37489712e-01
1.12099834e-02 2.99272358e-01 -8.85550916e-01 7.27563798e-01
-2.00493312e+00 -5.84982574e-01 -2.02674255e-01 1.90456247e+00
6.87316477e-01 4.97847438e-01 7.19343498e-02 -1.26412913e-01
6.88650668e-01 3.22538435e-01 -8.63796473e-01 3.45511705e-01
-2.40841478e-01 5.13430715e-01 8.71295154e-01 3.99760455e-01
-1.34094226e+00 1.21091819e+00 5.33847570e+00 8.33598137e-01
-1.06610894e+00 3.42218131e-01 6.97681069e-01 -3.44671607e-01
-3.52642804e-01 8.99121985e-02 -8.00114930e-01 5.34188509e-01
2.61948049e-01 5.02517164e-01 8.76175165e-02 1.08953917e+00
-1.07838951e-01 2.47157574e-01 -1.19082534e+00 9.03763354e-01
-4.30845991e-02 -1.42910588e+00 1.09974602e-02 -1.45104513e-01
8.27637732e-01 6.35209262e-01 1.21347733e-01 3.66968781e-01
2.92528331e-01 -8.66859436e-01 1.09934461e+00 4.11841094e-01
5.70772529e-01 -6.84196293e-01 6.50805056e-01 5.00105083e-01
-1.47452855e+00 -1.54897019e-01 -4.27913219e-01 1.49154380e-01
3.71293798e-02 6.77900970e-01 -6.13021135e-01 5.81768453e-01
1.10320103e+00 8.79998922e-01 -8.03501487e-01 9.26035464e-01
-2.52255499e-01 6.13884985e-01 -6.61338806e-01 4.70831871e-01
4.39104080e-01 -3.25277716e-01 5.89783251e-01 1.29191506e+00
-6.82944357e-02 -4.31411490e-02 5.58578312e-01 1.30395973e+00
-1.49333864e-01 -3.23290437e-01 -3.06125760e-01 5.33410788e-01
5.60065508e-01 1.34387457e+00 -7.69400954e-01 -2.79024512e-01
-4.19668317e-01 7.11539388e-01 5.06102145e-01 4.31109309e-01
-8.81758392e-01 -1.25903055e-01 8.55086386e-01 2.36710101e-01
5.80782294e-01 -1.57195590e-02 -8.11206877e-01 -1.22873557e+00
3.46679896e-01 -4.06000406e-01 3.99539381e-01 -6.37003839e-01
-1.79358912e+00 4.60493624e-01 1.84226960e-01 -1.44024110e+00
4.23013508e-01 -6.69512689e-01 -8.74386847e-01 1.01256394e+00
-2.01822925e+00 -1.68341470e+00 -5.07291734e-01 5.43589175e-01
5.80672562e-01 -2.83576492e-02 3.44186842e-01 2.08491430e-01
-5.23302376e-01 4.75130081e-01 -5.23422480e-01 4.02821526e-02
5.84513724e-01 -1.58463740e+00 4.39634830e-01 7.25771725e-01
1.82473481e-01 4.85932440e-01 2.34868720e-01 -6.04584098e-01
-1.08101869e+00 -1.74327910e+00 3.23206127e-01 -9.80053842e-01
5.41412234e-01 -7.71776140e-01 -1.18307173e+00 7.37074316e-01
-3.34505379e-01 7.27529168e-01 4.30893332e-01 -4.10630479e-02
-5.48886180e-01 -2.89799869e-01 -1.28491974e+00 1.04605593e-01
1.06942737e+00 -6.13085687e-01 -5.87390840e-01 4.87140656e-01
1.28946292e+00 -5.84691644e-01 -7.08194911e-01 7.19127536e-01
-6.77135587e-03 -7.44033277e-01 1.31012976e+00 -6.11137152e-01
3.49393457e-01 -6.54147446e-01 -3.01835567e-01 -1.10068381e+00
-3.90049934e-01 -1.50782272e-01 -2.18481213e-01 1.24107409e+00
6.81295276e-01 -5.09712636e-01 1.13615954e+00 4.24685389e-01
-5.76741636e-01 -1.22221577e+00 -9.27764475e-01 -8.44501734e-01
1.13942400e-01 -6.64704144e-01 1.03825462e+00 1.24929893e+00
-4.83680546e-01 1.44584164e-01 -1.84273683e-02 7.98376441e-01
1.00925767e+00 2.93600082e-01 7.73392379e-01 -1.49322402e+00
-2.94522673e-01 -3.46735120e-01 -3.72829795e-01 -1.03878760e+00
1.98579624e-01 -1.26436663e+00 6.18326515e-02 -1.30087519e+00
9.59862396e-02 -6.87616885e-01 -6.84414566e-01 6.62653029e-01
-3.74174118e-01 4.00280625e-01 3.64448190e-01 3.57802093e-01
-6.56587243e-01 5.69876432e-01 1.10542119e+00 -4.30769235e-01
-3.06192070e-01 2.78481811e-01 -7.56637216e-01 1.03137016e+00
4.66372013e-01 -6.46376908e-01 -2.11255133e-01 -6.36564076e-01
-9.29929167e-02 -4.85243529e-01 7.86294937e-01 -9.54187036e-01
1.30906895e-01 -2.64294058e-01 4.28766519e-01 -1.01188624e+00
4.45531338e-01 -1.01976204e+00 -3.40278417e-01 -1.18386373e-01
-2.10290343e-01 -2.55338073e-01 1.42414570e-01 9.04900551e-01
-2.44541720e-01 1.12551726e-01 9.62060392e-01 -3.68358493e-01
-7.59299457e-01 7.57078528e-01 6.17286861e-01 9.85812098e-02
9.83150244e-01 -3.84308934e-01 -9.98071805e-02 2.38051370e-01
-6.73491895e-01 8.55802238e-01 4.97643918e-01 6.07113540e-01
6.84246063e-01 -1.32150686e+00 -6.15270615e-01 2.23623306e-01
5.09739161e-01 1.05973673e+00 1.92507207e-02 7.36803174e-01
-1.14575095e-01 -3.27863172e-02 2.89712578e-01 -1.32399344e+00
-8.49875331e-01 2.33699888e-01 3.75944585e-01 -9.75314900e-02
-7.13794470e-01 1.34675109e+00 5.42024732e-01 -9.05840814e-01
4.11081463e-01 -6.96638882e-01 1.17017195e-01 -7.91728050e-02
2.95041770e-01 6.60724938e-02 2.94932097e-01 -4.97053027e-01
-6.10464156e-01 6.88376844e-01 -1.93166628e-03 1.44802198e-01
1.68931699e+00 2.52160668e-01 -1.72523350e-01 7.37970650e-01
1.09092820e+00 -7.82243758e-02 -1.75684738e+00 -5.53468049e-01
1.54195592e-01 -7.33495533e-01 1.56601295e-01 -7.44084656e-01
-1.35665405e+00 8.11766267e-01 4.26396310e-01 -1.26160666e-01
8.56870413e-01 5.35119355e-01 6.08968318e-01 3.47715706e-01
5.50410926e-01 -1.00764275e+00 2.09978834e-01 3.72564822e-01
9.20370638e-01 -1.42920482e+00 -6.93747699e-02 -7.37810552e-01
-7.45049834e-01 6.55478299e-01 1.04084778e+00 -5.24049222e-01
7.06789792e-01 3.28563213e-01 -1.02799542e-01 -5.25293052e-01
-5.25830567e-01 -2.71017581e-01 5.29567897e-01 6.70821786e-01
-1.44142836e-01 1.71125919e-01 5.34798324e-01 9.01547372e-01
-2.83962972e-02 -2.89910436e-01 2.13420857e-02 9.33478177e-01
-4.28873330e-01 -6.93417609e-01 -4.97635007e-01 7.92756796e-01
8.21981113e-03 1.43949231e-02 -6.03193402e-01 6.18027568e-01
4.72368509e-01 6.28655136e-01 3.29833686e-01 -3.59836280e-01
5.51225305e-01 -3.71950641e-02 3.46653819e-01 -9.76154208e-01
-6.93467081e-01 1.02303028e-01 -3.98391992e-01 -7.11429358e-01
-4.92584050e-01 -6.62700891e-01 -1.48947001e+00 2.46742040e-01
-5.36002219e-01 -9.89877880e-02 5.42540729e-01 8.77819598e-01
4.07202214e-01 6.30931497e-01 6.79121912e-01 -1.43482924e+00
-4.07559782e-01 -9.44741547e-01 -6.22348785e-01 4.59816575e-01
5.82289934e-01 -9.20539021e-01 -4.96068984e-01 -1.30906984e-01] | [8.027639389038086, -3.323402166366577] |
98b4dfd9-b80f-4ec1-9266-cd018f9b8fda | creating-training-corpora-for-nlg-micro | null | null | https://aclanthology.org/P17-1017 | https://aclanthology.org/P17-1017.pdf | Creating Training Corpora for NLG Micro-Planners | In this paper, we present a novel framework for semi-automatically creating linguistically challenging micro-planning data-to-text corpora from existing Knowledge Bases. Because our method pairs data of varying size and shape with texts ranging from simple clauses to short texts, a dataset created using this framework provides a challenging benchmark for microplanning. Another feature of this framework is that it can be applied to any large scale knowledge base and can therefore be used to train and learn KB verbalisers. We apply our framework to DBpedia data and compare the resulting dataset with Wen et al. 2016{'}s. We show that while Wen et al.{'}s dataset is more than twice larger than ours, it is less diverse both in terms of input and in terms of text. We thus propose our corpus generation framework as a novel method for creating challenging data sets from which NLG models can be learned which are capable of handling the complex interactions occurring during in micro-planning between lexicalisation, aggregation, surface realisation, referring expression generation and sentence segmentation. To encourage researchers to take up this challenge, we made available a dataset of 21,855 data/text pairs created using this framework in the context of the WebNLG shared task. | ['Laura Perez-Beltrachini', 'Claire Gardent', 'Shashi Narayan', 'Anastasia Shimorina'] | 2017-07-01 | null | null | null | acl-2017-7 | ['referring-expression-generation'] | ['computer-vision'] | [ 3.77866507e-01 7.95170724e-01 1.26747102e-01 -4.53660160e-01
-1.12504387e+00 -8.83631766e-01 1.01315463e+00 3.73508930e-01
-5.92282832e-01 1.15338850e+00 5.51167071e-01 -2.83566028e-01
-1.68461934e-01 -1.02359939e+00 -8.92128289e-01 -1.65061191e-01
1.88150823e-01 1.37068832e+00 3.37963164e-01 -6.70753121e-01
6.97768629e-02 5.11712655e-02 -1.36964762e+00 7.17364073e-01
1.04939365e+00 6.98367059e-01 3.97636175e-01 4.68396693e-01
-3.42759162e-01 8.59811306e-01 -7.38472998e-01 -6.80149138e-01
1.38108939e-01 -4.09365833e-01 -1.46021235e+00 -1.64050534e-01
4.71077949e-01 1.79369431e-02 1.59999415e-01 6.45064235e-01
7.82943428e-01 3.27809900e-01 6.67718172e-01 -7.21030951e-01
-3.44507694e-01 1.48022652e+00 2.11403340e-01 1.92088872e-01
8.75106812e-01 4.40507494e-02 1.16272748e+00 -5.66628277e-01
1.36582637e+00 1.37718356e+00 5.26875615e-01 9.49418008e-01
-1.35117257e+00 -1.19405746e-01 1.33239068e-02 2.77411491e-01
-1.12433302e+00 -7.28558004e-01 5.21512091e-01 -3.56107891e-01
1.50870287e+00 3.08787644e-01 6.12509131e-01 1.44535017e+00
-4.75949794e-01 9.65172112e-01 1.14863193e+00 -9.77324128e-01
2.76140627e-02 6.12789169e-02 -1.14717133e-01 3.51963133e-01
-5.84524535e-02 -9.76674333e-02 -5.11542141e-01 9.70856398e-02
3.56423706e-01 -8.88502359e-01 -1.60242110e-01 1.20942537e-02
-1.43282104e+00 8.09376359e-01 8.29096287e-02 6.12848401e-01
-1.41479611e-01 -1.16044939e-01 5.98665416e-01 2.16166541e-01
3.42727005e-01 7.96601057e-01 -6.73555553e-01 -5.59803426e-01
-7.07936108e-01 9.03984964e-01 1.31601548e+00 1.34799445e+00
5.15254438e-01 -3.43139172e-01 -3.51445824e-01 9.79211092e-01
1.82068184e-01 3.55395257e-01 6.17363155e-01 -8.60171437e-01
1.13225698e+00 6.74827754e-01 1.58996377e-02 -6.82392061e-01
-5.88553905e-01 1.89782396e-01 -3.75334471e-01 -5.03169477e-01
7.56398022e-01 -2.98894763e-01 -6.55259073e-01 1.81513810e+00
4.86292392e-01 -4.19687063e-01 5.50321937e-01 6.83226824e-01
1.23378384e+00 6.45392656e-01 1.78813245e-02 -2.28755385e-01
1.49293232e+00 -7.51631320e-01 -5.88869512e-01 -3.59334528e-01
9.76005435e-01 -6.11976087e-01 1.25581312e+00 3.98564100e-01
-1.49184215e+00 -2.65752405e-01 -7.22458601e-01 -4.22887295e-01
-5.52944183e-01 -2.91641176e-01 6.75657392e-01 2.25764498e-01
-1.04853237e+00 4.75520402e-01 -6.04148984e-01 -6.11643493e-01
1.88471720e-01 1.74898177e-01 -2.08329126e-01 -1.69032291e-01
-1.47912359e+00 1.32600093e+00 1.08317959e+00 -4.70640846e-02
-5.24287403e-01 -5.07401228e-01 -1.01744640e+00 -3.99348676e-01
8.50123703e-01 -7.67037809e-01 1.60643315e+00 -6.37433350e-01
-1.46711266e+00 1.06197357e+00 -4.92372038e-03 -6.59427702e-01
7.03561187e-01 -1.61782671e-02 -3.04499984e-01 -1.24486692e-01
3.01650256e-01 8.50618541e-01 1.69607222e-01 -8.96332741e-01
-7.12895334e-01 -3.92331839e-01 2.76600569e-01 3.79710436e-01
1.68943286e-01 1.46604896e-01 -3.47961843e-01 -5.65973341e-01
-2.49904543e-01 -7.42599726e-01 -1.08138792e-01 -1.12270677e+00
-7.50381172e-01 -5.46498477e-01 2.14265734e-01 -6.19088292e-01
1.22302508e+00 -1.56360447e+00 6.49945915e-01 5.47412671e-02
-3.93604934e-02 1.58355266e-01 -9.28925499e-02 8.44440758e-01
3.06995153e-01 3.85252267e-01 -4.12351102e-01 -2.45969355e-01
4.13455337e-01 5.62375188e-01 -1.14388466e-01 -2.13760525e-01
3.70313495e-01 1.19097412e+00 -8.71309161e-01 -8.13707530e-01
2.68417737e-03 1.11307524e-01 -6.26242995e-01 1.88740537e-01
-8.63608897e-01 4.91469115e-01 -3.36594909e-01 4.03623581e-01
9.70901772e-02 2.89897084e-01 3.70447010e-01 6.97908923e-02
-1.62514150e-01 8.10391426e-01 -1.05437493e+00 1.97802150e+00
-7.98287511e-01 3.66832376e-01 -2.56492019e-01 -8.52937102e-01
6.95022166e-01 4.47679073e-01 -7.30079412e-02 -7.92923570e-01
2.69920439e-01 4.36920851e-01 1.89405844e-01 -7.26150513e-01
5.88455856e-01 -4.59302217e-01 -5.12243390e-01 4.24930274e-01
3.15268427e-01 -8.37586582e-01 1.07858276e+00 2.22450316e-01
1.15358436e+00 1.19514510e-01 3.76331031e-01 -2.92770773e-01
5.74895144e-01 3.07423264e-01 3.70084554e-01 5.67525804e-01
3.05482805e-01 3.72869641e-01 7.18028367e-01 -3.68630677e-01
-9.68968987e-01 -7.30531871e-01 -2.91206241e-01 1.06249905e+00
-4.31269854e-01 -6.51855350e-01 -9.82855141e-01 -7.89294600e-01
-2.25449950e-01 1.10486662e+00 -6.20952964e-01 3.84113133e-01
-9.14855957e-01 -3.75308394e-01 9.31192458e-01 3.04646999e-01
2.59416789e-01 -1.78531563e+00 -5.74940264e-01 6.13514721e-01
-5.95710337e-01 -1.48858094e+00 3.10094957e-03 1.67389780e-01
-5.45943260e-01 -1.12762463e+00 -2.10632473e-01 -8.94377232e-01
2.55720258e-01 -7.33804882e-01 1.73552799e+00 -1.01216428e-01
-7.47929215e-02 2.23992005e-01 -7.19198167e-01 -8.13341796e-01
-9.74041164e-01 4.40099865e-01 -3.31410140e-01 -6.12295985e-01
3.25505823e-01 -3.99539351e-01 -5.20788506e-03 -2.04205215e-01
-1.18165255e+00 3.63913208e-01 2.61633784e-01 7.13240981e-01
6.34981632e-01 -1.60382837e-01 5.16221941e-01 -1.50431263e+00
8.72242033e-01 -5.46387672e-01 -4.94134456e-01 2.85395712e-01
-2.66195983e-01 1.28293559e-01 6.27347469e-01 -1.05180845e-01
-1.11606371e+00 -1.23592742e-01 -3.77955943e-01 5.15246451e-01
-4.60410327e-01 8.70553851e-01 -4.44647342e-01 5.16920447e-01
7.42812872e-01 -2.07029916e-02 -2.68042773e-01 -4.77341175e-01
8.66662145e-01 6.21999502e-01 5.41541696e-01 -1.16885841e+00
4.67458397e-01 -6.19701333e-02 -1.77679494e-01 -7.49666631e-01
-8.94279718e-01 -1.36485562e-01 -8.66839826e-01 1.19876862e-01
8.63386393e-01 -7.02490866e-01 -3.36170197e-01 2.74143785e-01
-1.39063609e+00 -7.99122989e-01 -5.35300016e-01 3.96423154e-02
-9.15925622e-01 1.40351683e-01 -7.51011074e-01 -4.00674731e-01
-3.18924546e-01 -1.02870536e+00 1.08947682e+00 -1.35731980e-01
-7.41227031e-01 -1.37960649e+00 1.95804536e-01 7.15848386e-01
2.57313192e-01 6.72645926e-01 1.23255587e+00 -9.75845098e-01
-1.92959741e-01 1.51408225e-01 5.02250381e-02 2.00364709e-01
-1.07234389e-01 7.64691830e-03 -6.74952686e-01 8.84982646e-02
-2.64222056e-01 -9.36135650e-01 4.85550672e-01 -2.10465435e-02
7.55329609e-01 -7.94671774e-01 2.32311655e-02 2.92392343e-01
1.21088827e+00 9.30860080e-03 7.15820134e-01 6.65425360e-01
7.08537221e-01 1.08222568e+00 7.21468508e-01 2.29973540e-01
1.05755329e+00 9.68455434e-01 1.77503765e-01 2.47448713e-01
-7.89031386e-02 -1.44596383e-01 3.44567716e-01 8.53308856e-01
-2.30476111e-01 -3.23368222e-01 -1.50560129e+00 7.52620220e-01
-1.94794405e+00 -9.34409201e-01 -2.82067984e-01 1.62485921e+00
1.56111073e+00 1.84811994e-01 2.39149392e-01 1.40477847e-02
3.49084556e-01 -6.38960749e-02 2.29327418e-02 -7.76151180e-01
-3.04654807e-01 5.97107708e-01 -6.26988336e-02 6.64063871e-01
-9.50695455e-01 1.42477632e+00 5.21772099e+00 9.58427012e-01
-7.21184850e-01 1.04827881e-01 1.93024293e-01 -2.06642285e-01
-5.20309091e-01 -1.66141808e-01 -8.87519538e-01 3.42337519e-01
1.25854588e+00 -2.07409665e-01 5.50614893e-01 5.49268901e-01
2.26492584e-02 -1.48837775e-01 -1.57192659e+00 6.41858160e-01
1.37996092e-01 -1.48804903e+00 1.56811759e-01 -2.01664656e-01
6.26311064e-01 1.82065666e-01 -4.72973704e-01 5.72042942e-01
5.99542022e-01 -1.28611076e+00 1.08250272e+00 2.77737349e-01
7.14973032e-01 -6.79783881e-01 8.53382587e-01 5.82910240e-01
-7.24893510e-01 2.61926562e-01 -6.27035797e-02 -1.84620023e-01
4.79570836e-01 5.18815577e-01 -1.24424982e+00 7.54470050e-01
3.61841053e-01 4.32992190e-01 -6.78915977e-01 5.01106739e-01
-4.89873976e-01 4.64216173e-01 -3.62361789e-01 -2.59395331e-01
3.82951349e-01 -6.79287836e-02 6.01522744e-01 1.57107449e+00
8.62058103e-02 8.11788440e-02 2.67607272e-01 9.16103363e-01
-9.59982350e-02 4.11718041e-01 -8.62499475e-01 -1.55673727e-01
4.85819459e-01 1.22418940e+00 -7.10393488e-01 -4.57102686e-01
-9.32684317e-02 6.72736526e-01 8.08601022e-01 2.51870956e-02
-5.11062026e-01 -4.36702698e-01 3.30346711e-02 6.87202439e-02
2.76548862e-01 -1.03256404e-01 -3.32531333e-02 -9.99190331e-01
2.59664506e-01 -1.21356976e+00 3.80644172e-01 -5.50354183e-01
-1.36430311e+00 8.56887579e-01 4.29261595e-01 -4.91929919e-01
-8.92507493e-01 -6.01435363e-01 -2.15675443e-01 7.01311350e-01
-1.31680608e+00 -1.44170880e+00 -3.76826003e-02 6.17413461e-01
5.93262970e-01 -5.38097173e-02 1.04525125e+00 9.59472135e-02
-4.50279117e-01 2.20710933e-01 -3.39511484e-01 1.08893335e-01
5.55446625e-01 -1.67402637e+00 5.91680229e-01 6.11129403e-01
4.33540940e-01 4.23796564e-01 6.46677911e-01 -6.15573585e-01
-1.27498484e+00 -1.09816706e+00 1.33894634e+00 -9.72879648e-01
7.84910560e-01 -7.03160942e-01 -7.57997692e-01 8.59271169e-01
2.87162840e-01 -3.30866486e-01 6.74205601e-01 2.83494323e-01
-2.16910765e-02 1.93988532e-01 -1.16963375e+00 5.23387432e-01
1.26243424e+00 -3.18753004e-01 -1.13789463e+00 5.80021799e-01
7.34314382e-01 -1.00412023e+00 -1.03823900e+00 3.91564518e-01
1.16946980e-01 -9.22221005e-01 4.73790377e-01 -6.59784853e-01
8.08400154e-01 4.02888246e-02 -2.64096946e-01 -1.53569150e+00
1.98126718e-01 -7.63234019e-01 -1.10902213e-01 1.79896331e+00
9.77698743e-01 -4.05082107e-01 3.57115090e-01 5.77714324e-01
-3.56340796e-01 -8.45262289e-01 -1.10143149e+00 -6.40442729e-01
5.13339877e-01 -9.24677551e-01 7.66795933e-01 8.13163459e-01
4.58407640e-01 6.87980652e-01 2.07345754e-01 -4.22273874e-01
1.56653449e-01 -6.15782179e-02 8.47920477e-01 -1.09987199e+00
-3.77218395e-01 -4.14073586e-01 -9.54774246e-02 -5.98007083e-01
3.93042117e-01 -1.32236981e+00 1.97656393e-01 -1.81005108e+00
-8.82104561e-02 -7.14291871e-01 3.48153353e-01 5.78306019e-01
1.63202807e-01 5.65457903e-02 3.09521496e-01 -5.10642044e-02
-5.81608236e-01 2.88623393e-01 1.36244822e+00 7.06599355e-02
-2.62247890e-01 -3.44707072e-01 -7.82332778e-01 7.55167663e-01
6.95170939e-01 -2.52567232e-01 -3.06179106e-01 -7.75279224e-01
6.06428385e-01 1.39340773e-01 9.00004134e-02 -6.77444398e-01
-9.56167467e-03 -2.04698831e-01 -9.07893851e-02 -1.91610828e-01
1.04106084e-01 -4.35171962e-01 -4.85841334e-02 -4.39805426e-02
-3.56926590e-01 -1.77154765e-02 4.09095526e-01 7.95285180e-02
-2.64708608e-01 -3.98090154e-01 3.02604228e-01 -5.06161630e-01
-6.57715797e-01 -1.14041030e-01 -3.34262133e-01 8.97521257e-01
1.09530950e+00 -7.42262825e-02 -4.27358657e-01 -2.04939112e-01
-7.16695249e-01 3.55087876e-01 3.66274983e-01 3.93746138e-01
3.19586426e-01 -1.16410029e+00 -9.44156706e-01 4.66808118e-02
3.06370288e-01 7.92988837e-01 -2.00205799e-02 6.68858886e-01
-6.96385324e-01 6.58188939e-01 -1.18770145e-01 -2.46511236e-01
-9.12772298e-01 1.89852148e-01 -3.83217297e-02 -6.95793509e-01
-5.48834682e-01 9.03632939e-01 -3.72502834e-01 -8.66100192e-01
7.29162320e-02 -6.64912581e-01 -4.61525500e-01 3.45434278e-01
3.42779666e-01 1.73173577e-01 3.80969077e-01 -7.80408800e-01
-2.79056042e-01 2.04997480e-01 -6.40633628e-02 -4.91134822e-01
1.56900799e+00 -7.92944506e-02 -2.91595936e-01 6.03180528e-01
7.91162729e-01 2.41338998e-01 -6.88011050e-01 -1.49193034e-01
5.61804473e-01 3.08005642e-02 -4.40702409e-01 -1.20468462e+00
-4.39991266e-01 6.61359966e-01 -2.84578234e-01 4.52600598e-01
7.77156949e-01 5.58118105e-01 8.12223017e-01 6.42734945e-01
4.08294231e-01 -1.44760108e+00 -1.08187452e-01 1.26615107e+00
1.29459012e+00 -1.17416775e+00 -3.35536510e-01 -3.84116143e-01
-8.38683486e-01 1.07804871e+00 6.28325284e-01 8.71438757e-02
3.33162844e-02 3.52264762e-01 7.90076181e-02 -4.36562121e-01
-9.89445090e-01 -4.54245001e-01 2.38964245e-01 9.67328489e-01
6.99449837e-01 1.36997640e-01 -5.99360526e-01 8.51313591e-01
-1.08482945e+00 -1.65038884e-01 6.44491017e-01 8.91865194e-01
-8.70645493e-02 -1.66693699e+00 -4.00624573e-02 4.21341151e-01
-4.09052342e-01 -3.15553337e-01 -7.40992010e-01 1.18291748e+00
3.46844971e-01 8.59499156e-01 -1.37587517e-01 4.66032960e-02
6.84509456e-01 2.81820685e-01 9.15117145e-01 -1.34111619e+00
-6.85540676e-01 -2.70123690e-01 9.92395937e-01 -4.26029801e-01
-5.84650218e-01 -8.71167243e-01 -1.40949094e+00 -1.35554567e-01
-1.10346250e-01 1.62482098e-01 5.31891465e-01 1.33396447e+00
5.81170805e-02 5.06774008e-01 -1.31915376e-01 -9.46517229e-01
-3.87472093e-01 -1.24774218e+00 -3.37875336e-01 6.71244383e-01
-1.80292547e-01 -4.47654009e-01 -5.41960038e-02 7.28411600e-02] | [11.196165084838867, 9.100493431091309] |
e36f032b-f48c-46f5-a36e-5b171dc267e0 | playing-lottery-tickets-with-vision-and | 2104.11832 | null | https://arxiv.org/abs/2104.11832v2 | https://arxiv.org/pdf/2104.11832v2.pdf | Playing Lottery Tickets with Vision and Language | Large-scale pre-training has recently revolutionized vision-and-language (VL) research. Models such as LXMERT and UNITER have significantly lifted the state of the art over a wide range of VL tasks. However, the large number of parameters in such models hinders their application in practice. In parallel, work on the lottery ticket hypothesis (LTH) has shown that deep neural networks contain small matching subnetworks that can achieve on par or even better performance than the dense networks when trained in isolation. In this work, we perform the first empirical study to assess whether such trainable subnetworks also exist in pre-trained VL models. We use UNITER as the main testbed (also test on LXMERT and ViLT), and consolidate 7 representative VL tasks for experiments, including visual question answering, visual commonsense reasoning, visual entailment, referring expression comprehension, image-text retrieval, GQA, and NLVR$^2$. Through comprehensive analysis, we summarize our main findings as follows. ($i$) It is difficult to find subnetworks that strictly match the performance of the full model. However, we can find "relaxed" winning tickets at 50%-70% sparsity that maintain 99% of the full accuracy. ($ii$) Subnetworks found by task-specific pruning transfer reasonably well to the other tasks, while those found on the pre-training tasks at 60%/70% sparsity transfer universally, matching 98%/96% of the full accuracy on average over all the tasks. ($iii$) Besides UNITER, other models such as LXMERT and ViLT can also play lottery tickets. However, the highest sparsity we can achieve for ViLT is far lower than LXMERT and UNITER (30% vs. 70%). ($iv$) LTH also remains relevant when using other training methods (e.g., adversarial training). | ['Zicheng Liu', 'Lijuan Wang', 'Jingjing Liu', 'Shuohang Wang', 'Yu Cheng', 'Tianlong Chen', 'Linjie Li', 'Yen-Chun Chen', 'Zhe Gan'] | 2021-04-23 | null | null | null | null | ['visual-commonsense-reasoning', 'visual-entailment'] | ['reasoning', 'reasoning'] | [ 1.92224473e-01 3.24172318e-01 -1.49984390e-01 -1.77805081e-01
-5.59398532e-01 -5.56476831e-01 5.71451068e-01 -1.13128521e-01
-5.10985315e-01 5.65698922e-01 8.17092657e-02 -6.28246129e-01
-2.95030866e-02 -6.52794719e-01 -9.48670805e-01 -3.99199814e-01
5.44516779e-02 2.54599780e-01 1.93158351e-02 -2.72558153e-01
6.68135062e-02 3.12172621e-01 -1.50322187e+00 6.98025644e-01
7.27468669e-01 9.85038221e-01 1.87913582e-01 3.87040228e-01
-1.66767552e-01 1.16027939e+00 -7.90374815e-01 -5.79129696e-01
3.34562361e-01 -3.22750896e-01 -8.87616098e-01 -2.09448218e-01
1.12864482e+00 -4.29079354e-01 -4.95221436e-01 9.27890837e-01
5.03077507e-01 1.21720098e-01 6.21693075e-01 -1.36540222e+00
-9.08433020e-01 7.27148831e-01 -4.79095310e-01 1.93974674e-01
3.89184207e-01 6.87573433e-01 1.35370076e+00 -7.04998136e-01
8.48520219e-01 1.42653584e+00 6.83708787e-01 6.24452353e-01
-1.23941541e+00 -1.06085300e+00 3.06376249e-01 3.16246748e-01
-1.38569629e+00 -4.80854839e-01 5.42420805e-01 -2.59249419e-01
1.50752401e+00 1.79147646e-01 7.03678131e-01 1.29156327e+00
8.94113258e-02 1.15282750e+00 1.26774216e+00 -3.03257495e-01
-1.86174572e-01 2.03749880e-01 7.09063336e-02 9.87789214e-01
1.98119149e-01 -9.55764428e-02 -5.54078162e-01 1.70582652e-01
7.59085834e-01 -4.05864239e-01 -5.23504257e-01 -1.80229440e-01
-1.23339403e+00 9.34364080e-01 7.42953718e-01 4.99856889e-01
-2.22979859e-02 3.63575578e-01 5.66064060e-01 4.94936615e-01
2.95999438e-01 6.60781264e-01 -2.92364776e-01 1.04036428e-01
-1.08608317e+00 3.23885798e-01 7.95258462e-01 9.63487208e-01
7.29493618e-01 4.66807693e-01 -2.93862224e-01 8.23502719e-01
-2.65118070e-02 3.37642342e-01 3.48171264e-01 -1.06417513e+00
7.01225638e-01 4.96024728e-01 -3.44948947e-01 -1.14973104e+00
-3.23621333e-01 -6.28802240e-01 -9.19679642e-01 2.23904684e-01
7.05102205e-01 -1.74664035e-02 -9.61933792e-01 2.11323285e+00
-3.09165299e-01 -5.92691600e-02 -2.63688564e-02 7.42477059e-01
1.21617031e+00 6.59891605e-01 3.38809669e-01 5.87627739e-02
1.45793808e+00 -7.91632712e-01 -4.23524052e-01 -5.52370012e-01
7.96404600e-01 -6.15504265e-01 1.60284638e+00 3.81095320e-01
-1.16010940e+00 -6.22634768e-01 -1.11721253e+00 -3.84988189e-01
-3.82452130e-01 -4.35113460e-02 9.18998420e-01 5.30508876e-01
-1.46844649e+00 5.11070073e-01 -3.27361435e-01 -4.80577171e-01
8.88867915e-01 3.17728579e-01 -2.98545778e-01 -5.76392710e-01
-1.29980075e+00 1.21465218e+00 2.11619362e-01 -6.63586631e-02
-1.10559380e+00 -9.81876373e-01 -1.02180374e+00 1.98808908e-01
3.59300584e-01 -9.64103401e-01 9.91203368e-01 -1.20792031e+00
-9.17533636e-01 1.45766735e+00 -2.46497139e-01 -7.88599133e-01
3.96533936e-01 -8.02600905e-02 -2.06639022e-01 3.12498748e-01
1.92094058e-01 1.25824678e+00 8.26044142e-01 -1.22818458e+00
-2.28047580e-01 -1.53741941e-01 4.81703758e-01 2.31526464e-01
-1.38612300e-01 1.16861435e-02 -3.60146880e-01 -5.22710562e-01
-2.51116991e-01 -5.44973433e-01 2.87476897e-01 1.37225717e-01
-4.76558745e-01 -4.32791203e-01 6.99528039e-01 -6.69073403e-01
8.46395314e-01 -2.08873105e+00 -6.53017138e-04 -7.35324994e-02
7.29927361e-01 2.71793097e-01 -3.92995507e-01 3.53839755e-01
-3.63368899e-01 2.93029696e-01 -8.84859189e-02 -5.02634168e-01
2.43042126e-01 3.98803473e-01 -4.23997492e-01 3.32537353e-01
2.26498827e-01 1.34347582e+00 -5.91795087e-01 -5.84865868e-01
2.02829346e-01 2.64065415e-01 -5.78327358e-01 -1.77382886e-01
-3.92515033e-01 9.34974253e-02 5.61980065e-04 6.65033400e-01
3.82315725e-01 -4.75598335e-01 3.49070504e-03 -4.50959563e-01
7.12209120e-02 3.09207022e-01 -6.82098567e-01 1.53004158e+00
-3.68889570e-01 1.24492395e+00 2.96453387e-01 -1.36967087e+00
5.81264496e-01 2.03767613e-01 4.32171151e-02 -1.20039189e+00
1.16082057e-01 3.77274211e-03 2.97909886e-01 -5.12279689e-01
3.53485495e-01 -4.72112089e-01 -4.72478159e-02 3.60257804e-01
2.53796875e-01 -3.02899867e-01 1.71598703e-01 4.88782138e-01
1.08882737e+00 -7.97763467e-02 2.00628176e-01 -3.11776966e-01
2.61833936e-01 1.69675753e-01 2.26269946e-01 1.06108570e+00
-3.27771246e-01 4.21906829e-01 8.25948477e-01 -2.68920094e-01
-1.02770793e+00 -8.67790103e-01 -6.16980046e-02 1.23526418e+00
-2.03400273e-02 -3.53833616e-01 -7.37110317e-01 -5.19340634e-01
-5.83615787e-02 1.10259187e+00 -6.08593941e-01 -2.05407083e-01
-5.53291500e-01 -5.99683404e-01 1.15732431e+00 5.42358100e-01
7.96343863e-01 -1.49971581e+00 -5.05687594e-01 -1.85955241e-01
-2.78214544e-01 -1.40598261e+00 3.94638143e-02 9.26114991e-02
-6.13019586e-01 -9.32962060e-01 -6.34992957e-01 -8.09461236e-01
4.67708915e-01 2.74376422e-01 1.53376877e+00 3.37059379e-01
-4.24577057e-01 4.83163744e-01 -6.92560747e-02 -2.74474919e-01
-2.44380027e-01 -1.80479735e-01 -3.48843515e-01 -5.00214040e-01
2.22780436e-01 -5.51027119e-01 -5.89034140e-01 1.46772042e-01
-8.10816407e-01 2.29457289e-01 6.40957773e-01 8.38064790e-01
4.38984573e-01 6.77533746e-02 4.60322917e-01 -7.31839120e-01
6.98466003e-01 -3.36785406e-01 -2.34210134e-01 6.81787431e-02
-4.98206735e-01 -1.38106033e-01 6.49075687e-01 -6.37726545e-01
-7.71791935e-01 -6.02741897e-01 -2.24144846e-01 -7.37222135e-01
-2.29489088e-01 4.40897316e-01 -9.83668715e-02 -1.38217479e-01
6.43029094e-01 1.98360354e-01 4.05175058e-04 -1.60381958e-01
5.22736132e-01 -9.68924630e-03 4.57330585e-01 -6.00343466e-01
7.14114249e-01 5.48677266e-01 -1.91364780e-01 -7.93590248e-01
-1.02506125e+00 -1.83195800e-01 -5.87091334e-02 3.39599587e-02
9.76124465e-01 -8.93237054e-01 -9.96245503e-01 2.12355003e-01
-1.17848182e+00 -7.13087022e-01 -5.16335011e-01 2.55929410e-01
-5.29836416e-01 3.44520211e-01 -6.11603498e-01 -5.28613985e-01
-4.19249952e-01 -1.09156656e+00 9.28145170e-01 -1.43823503e-02
-4.30576116e-01 -1.15615082e+00 -4.41106290e-01 5.79314411e-01
4.51828778e-01 2.70343184e-01 1.45261800e+00 -6.28277302e-01
-6.08061552e-01 1.15226336e-01 -7.85286427e-01 5.01266718e-01
-4.45819438e-01 -2.21642196e-01 -1.03148341e+00 -4.22821492e-01
7.06721395e-02 -8.07651758e-01 1.30955350e+00 4.78406101e-01
1.15129364e+00 -1.95135549e-01 -2.05763623e-01 6.62770867e-01
1.39037716e+00 3.75313275e-02 8.42913330e-01 3.38310689e-01
7.54673064e-01 6.02712750e-01 1.79382533e-01 -8.40639025e-02
4.33411092e-01 4.26732033e-01 5.66820860e-01 -2.86762238e-01
-5.66143036e-01 -2.04674020e-01 5.62056482e-01 2.77011245e-01
-3.94880399e-02 -4.17862356e-01 -9.66116250e-01 5.87575674e-01
-1.59540784e+00 -1.26455808e+00 -1.89932268e-02 1.90280461e+00
6.65173233e-01 3.79920781e-01 -1.20749451e-01 6.51997179e-02
4.12414312e-01 4.58681852e-01 -5.96497059e-01 -5.84837019e-01
-4.39150393e-01 6.85544074e-01 2.52573013e-01 4.36342180e-01
-8.44741106e-01 1.38150227e+00 6.69318914e+00 1.11253202e+00
-1.13632870e+00 1.72395438e-01 7.54253328e-01 -2.59054810e-01
-3.66544724e-01 -1.28673747e-01 -7.36642957e-01 -1.17761254e-01
6.75107598e-01 9.38288346e-02 5.06285906e-01 7.50893652e-01
5.10417745e-02 -2.60595232e-01 -1.32901692e+00 1.16497791e+00
4.84359860e-01 -1.42210448e+00 2.82294303e-01 -2.02134885e-02
6.81055069e-01 2.59750515e-01 3.19883466e-01 8.80154788e-01
3.13470453e-01 -1.64418793e+00 7.99337745e-01 7.54222348e-02
9.72564459e-01 -6.05170310e-01 5.48431158e-01 4.52398121e-01
-8.30569685e-01 3.36926691e-02 -4.78685886e-01 -1.96555808e-01
-3.53820957e-02 3.61435026e-01 -7.61874318e-01 2.57832646e-01
6.60743594e-01 7.01169908e-01 -7.48598814e-01 6.29210889e-01
-3.85956168e-01 6.50248230e-01 -3.12440366e-01 3.23429927e-02
4.77161735e-01 1.48319438e-01 5.42243838e-01 1.35120380e+00
5.94830029e-02 -6.74747750e-02 -7.21128061e-02 1.30043399e+00
-4.82286870e-01 -1.39292777e-02 -8.08456779e-01 -1.04098253e-01
4.90215234e-02 9.98522758e-01 -5.47545671e-01 -5.11297226e-01
-4.21436787e-01 8.57776940e-01 5.07752359e-01 6.84393466e-01
-9.70905185e-01 -2.64661014e-01 7.08394527e-01 1.36829212e-01
4.60384369e-01 -1.33363634e-01 -2.97267735e-01 -1.31414688e+00
-4.37776186e-02 -1.16674042e+00 3.12243611e-01 -1.20882237e+00
-1.35634577e+00 5.93089104e-01 1.40713051e-01 -7.27769971e-01
-2.36007005e-01 -8.88290823e-01 -7.23028302e-01 8.56725395e-01
-1.53854179e+00 -1.17350042e+00 -3.43955427e-01 1.04585660e+00
5.36134660e-01 1.65938120e-02 7.25921631e-01 1.37439549e-01
-6.27258658e-01 7.61111438e-01 -2.94497818e-01 3.52958888e-01
5.92580616e-01 -1.04202461e+00 2.16855362e-01 6.94800138e-01
3.68255585e-01 9.10095751e-01 6.20746613e-01 -3.74312788e-01
-1.15641332e+00 -7.84843743e-01 7.69153595e-01 -3.94367963e-01
7.36130178e-01 -4.45567667e-01 -8.26826453e-01 9.75229383e-01
4.98497069e-01 -1.50504038e-01 5.14496326e-01 2.93146223e-01
-8.30546975e-01 1.02784388e-01 -1.17584336e+00 7.99571872e-01
1.21257722e+00 -7.88694441e-01 -7.87430465e-01 4.94475782e-01
8.30810249e-01 -3.13034177e-01 -6.53509259e-01 3.28503609e-01
4.46209580e-01 -1.19918489e+00 1.20527554e+00 -7.13561416e-01
7.40620375e-01 1.65761888e-01 -2.66743839e-01 -1.03750753e+00
-1.75482914e-01 -4.00827140e-01 1.63196504e-01 1.02366555e+00
3.14621627e-01 -6.34631157e-01 6.89889312e-01 6.22526407e-01
-1.93385288e-01 -8.48832607e-01 -9.93992031e-01 -7.48715103e-01
3.97593647e-01 -9.07389641e-01 1.22406401e-01 1.09885812e+00
-1.46016002e-01 5.08583009e-01 -1.68791994e-01 -2.79811382e-01
6.03099346e-01 1.57353446e-01 5.90407133e-01 -8.34022939e-01
-4.95019436e-01 -8.20002496e-01 -2.05569848e-01 -1.24156904e+00
5.82262576e-01 -1.24907887e+00 -2.99011081e-01 -1.80476916e+00
3.09560508e-01 -3.06635261e-01 -1.14721380e-01 8.41908693e-01
9.19317156e-02 4.54085052e-01 4.09623921e-01 2.14510769e-01
-5.43524325e-01 2.84760416e-01 1.59417331e+00 -2.69342244e-01
1.98597997e-01 -4.06810969e-01 -1.07676423e+00 8.38596404e-01
7.66528726e-01 -8.35154280e-02 -4.82626468e-01 -6.22425377e-01
2.83693314e-01 -1.57661408e-01 8.53926480e-01 -7.19088078e-01
-8.43482614e-02 8.58975723e-02 5.16908288e-01 -2.85535604e-01
4.42718059e-01 -4.76734966e-01 -3.27153593e-01 3.48909348e-01
-3.20903629e-01 -1.28618091e-01 6.46786213e-01 2.67321199e-01
-1.30916104e-01 -1.42247781e-01 7.88588464e-01 -3.77358764e-01
-8.72358680e-01 6.65402263e-02 -3.60346824e-01 4.30715263e-01
5.57738900e-01 -4.62761521e-01 -7.55954504e-01 -6.39757872e-01
-6.91817880e-01 1.38635099e-01 2.50446677e-01 2.55254954e-01
7.61038542e-01 -9.34823334e-01 -5.94790399e-01 -2.04368606e-02
9.52710584e-02 1.05998460e-02 3.02148074e-01 9.33420181e-01
-3.45091134e-01 7.10787535e-01 -2.55751938e-01 -5.09851456e-01
-1.15295410e+00 5.42954981e-01 3.44128907e-01 -4.68853176e-01
-6.64875209e-01 9.80973840e-01 7.27227867e-01 -1.63509324e-01
2.89578378e-01 -5.44006646e-01 -3.17166075e-02 -3.91807668e-02
3.05462301e-01 4.71733361e-02 -1.67213991e-01 -5.67641854e-01
-2.97043771e-01 5.26824236e-01 -4.66642454e-02 3.40881199e-02
1.17946112e+00 9.28427354e-02 -1.46251127e-01 2.19834402e-01
1.36446345e+00 -1.53099850e-01 -8.81586134e-01 -4.97776195e-02
-4.21971589e-01 -1.77358851e-01 9.85632278e-03 -9.88674700e-01
-1.13596511e+00 1.17033041e+00 1.25333741e-01 9.80956666e-03
1.13200665e+00 2.54635185e-01 6.59023046e-01 5.11885166e-01
2.44194239e-01 -6.60920620e-01 1.74141556e-01 4.87381756e-01
1.12964320e+00 -1.12411082e+00 -4.20114920e-02 -2.47834355e-01
-6.19448483e-01 6.81497753e-01 7.57549703e-01 -3.09529126e-01
2.22273096e-01 6.71213046e-02 -2.79457510e-01 -5.15241742e-01
-8.99841309e-01 -3.34157079e-01 2.61345446e-01 6.70131505e-01
3.69676590e-01 -1.80161402e-01 -3.84778995e-03 3.73396575e-01
-4.33402270e-01 -2.18178764e-01 1.13616265e-01 5.29194653e-01
-3.59494090e-01 -6.19906187e-01 -2.35877782e-01 5.03242493e-01
-4.06392872e-01 -6.57944322e-01 -4.50000435e-01 1.43372488e+00
2.91257232e-01 9.02498960e-01 1.89996019e-01 -2.47498780e-01
3.61757398e-01 2.05164477e-01 7.67284274e-01 -5.91384351e-01
-8.43780279e-01 1.37968855e-02 3.72500986e-01 -7.67096698e-01
-3.47847551e-01 -3.20947558e-01 -1.23360264e+00 -6.88949943e-01
-7.04985633e-02 -2.88033187e-01 2.73385525e-01 9.86136317e-01
1.15587004e-01 5.24083316e-01 -2.85966545e-01 -6.23683929e-01
-4.06189948e-01 -9.28427279e-01 -4.39416379e-01 5.91687441e-01
8.56164321e-02 -5.81127822e-01 -6.39260352e-01 -6.50030747e-02] | [10.564112663269043, 1.8268144130706787] |
fd430378-bdaf-4f0c-ae7f-37ab79f1aed6 | da-gan-instance-level-image-translation-by-1 | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Ma_DA-GAN_Instance-Level_Image_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Ma_DA-GAN_Instance-Level_Image_CVPR_2018_paper.pdf | DA-GAN: Instance-Level Image Translation by Deep Attention Generative Adversarial Networks | Unsupervised image translation, which aims in translating two independent sets of images, is challenging in discovering the correct correspondences without paired data. Existing works build upon Generative Adversarial Networks (GANs) such that the distribution of the translated images are indistinguishable from the distribution of the target set. However, such set-level constraints cannot learn the instance-level correspondences (e.g. aligned semantic parts in object transfiguration task). This limitation often results in false positives (e.g. geometric or semantic artifacts), and further leads to mode collapse problem. To address the above issues, we propose a novel framework for instance-level image translation by Deep Attention GAN (DA-GAN). Such a design enables DA-GAN to decompose the task of translating samples from two sets into translating instances in a highly-structured latent space. Specifically, we jointly learn a deep attention encoder, and the instance-level correspondences could be consequently discovered through attending on the learned instances. Therefore, the constraints could be exploited on both set-level and instance-level. Comparisons against several state-of-the- arts demonstrate the superiority of our approach, and the broad application capability, e.g, pose morphing, data augmentation, etc., pushes the margin of domain translation problem. | ['Jianlong Fu', 'Chang Wen Chen', 'Shuang Ma', 'Tao Mei'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['deep-attention', 'image-animation', 'deep-attention'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 5.96877456e-01 3.57273012e-01 -7.70022720e-02 -2.85709918e-01
-1.00298882e+00 -7.40886152e-01 7.79106915e-01 -6.30951762e-01
1.01536950e-02 8.53924274e-01 -2.05937736e-02 -1.20935636e-02
4.79262434e-02 -9.84143496e-01 -1.25724971e+00 -9.31284606e-01
5.69988132e-01 6.87696397e-01 -2.29938492e-01 -1.38440937e-01
-5.86443804e-02 2.98163384e-01 -1.44721353e+00 3.24239433e-01
1.08956492e+00 8.66821051e-01 1.60280243e-01 1.75019607e-01
-2.67593205e-01 4.74258840e-01 -7.28112161e-01 -4.34708774e-01
4.66227174e-01 -9.23377097e-01 -6.08003020e-01 5.22247195e-01
6.35213375e-01 -3.63330275e-01 -3.03012818e-01 1.41937351e+00
1.79124579e-01 -4.68330160e-02 7.52039373e-01 -1.54596961e+00
-1.34106493e+00 3.90744299e-01 -6.41926348e-01 -4.05417204e-01
1.70474663e-01 3.78706485e-01 8.60338211e-01 -9.28406894e-01
6.65419400e-01 1.32136071e+00 2.50654876e-01 7.56351411e-01
-1.38955748e+00 -7.92734385e-01 1.76197022e-01 -9.57771912e-02
-1.14586604e+00 -3.09994906e-01 1.10111451e+00 -5.51240146e-01
3.05156738e-01 3.08294475e-01 5.32319188e-01 1.56882346e+00
5.09545356e-02 7.59421289e-01 1.47465920e+00 -2.42793098e-01
2.22695932e-01 1.75005376e-01 -5.67269087e-01 4.72324610e-01
1.88557610e-01 2.28070050e-01 -4.06084210e-01 1.73244640e-01
1.14219642e+00 1.97197974e-01 -2.52369761e-01 -4.61732328e-01
-1.52747309e+00 7.09176660e-01 7.05265820e-01 1.39966249e-01
-2.73712695e-01 2.99480073e-02 7.53714219e-02 3.85876745e-01
3.92444402e-01 6.02767229e-01 -4.02652442e-01 4.07583237e-01
-7.23428607e-01 2.50343889e-01 2.58543670e-01 1.43391788e+00
9.52666640e-01 1.60100013e-01 -3.46617818e-01 4.59508449e-01
1.84485182e-01 7.68919468e-01 4.23498273e-01 -7.33616173e-01
6.19398117e-01 5.35141528e-01 1.93859525e-02 -9.38957989e-01
2.01616213e-01 -5.12179196e-01 -1.13407016e+00 2.64713824e-01
3.86164427e-01 -4.82339635e-02 -1.22708762e+00 2.22592521e+00
2.89096892e-01 2.68268257e-01 1.03850506e-01 9.89916682e-01
6.70446217e-01 5.92297018e-01 -1.11995153e-01 -1.76861212e-01
1.27458823e+00 -9.19844806e-01 -7.72866845e-01 -3.70382935e-01
9.64929983e-02 -6.99523509e-01 1.40111804e+00 5.03733940e-02
-1.11893737e+00 -7.87919462e-01 -9.06949162e-01 -2.84918677e-02
-2.77726412e-01 9.17273313e-02 4.21935201e-01 2.49618590e-01
-8.46341074e-01 3.23248506e-01 -7.21074462e-01 -1.82622969e-01
6.94375694e-01 3.31196904e-01 -4.81465608e-01 -1.05390020e-01
-1.12895930e+00 5.56964159e-01 2.96947986e-01 1.18625537e-01
-1.03616893e+00 -6.65696204e-01 -8.98633301e-01 2.88692582e-02
3.55847001e-01 -1.00462067e+00 8.29812407e-01 -1.56777287e+00
-1.50874329e+00 1.02745807e+00 4.58989739e-02 -6.16996661e-02
7.40016460e-01 -7.65788928e-02 -2.19318360e-01 -1.22611590e-01
4.16187942e-01 7.75928259e-01 1.29492664e+00 -1.65611708e+00
-3.69990408e-01 -5.18670559e-01 1.51533201e-01 1.53962046e-01
-3.63534838e-01 -2.99458086e-01 -3.85946691e-01 -9.00154054e-01
2.29608953e-01 -9.79519725e-01 -1.18864238e-01 4.32616435e-02
-6.69991016e-01 2.83070296e-01 8.16774964e-01 -6.26697481e-01
5.91888309e-01 -2.22596550e+00 5.97112000e-01 -4.98142391e-02
2.00473145e-01 -6.89234436e-02 -3.12716663e-01 1.65484577e-01
-2.43243843e-01 1.31723434e-01 -4.69650984e-01 -3.28874439e-01
4.96002333e-03 3.84746075e-01 -6.05501354e-01 3.47783118e-01
5.90023577e-01 1.31068659e+00 -9.87481594e-01 -2.20755562e-01
5.55026978e-02 4.03271407e-01 -6.12964094e-01 5.75937092e-01
-4.78051066e-01 1.21358061e+00 -5.58596015e-01 5.04291654e-01
7.89335370e-01 -2.84564257e-01 -4.10211645e-02 -4.63730931e-01
2.80576587e-01 1.12940498e-01 -7.19392061e-01 1.99971151e+00
-5.13760030e-01 4.24706221e-01 -2.44057789e-01 -1.16179609e+00
9.36583400e-01 2.41251200e-01 3.44025284e-01 -6.72451079e-01
2.81401090e-02 2.21015200e-01 8.32903385e-02 -3.71487379e-01
2.38111734e-01 -4.13271755e-01 -1.91694260e-01 3.90593290e-01
1.69325009e-01 -1.84858337e-01 -3.46466273e-01 -7.41641372e-02
6.71285570e-01 4.48511660e-01 -2.00668350e-03 -1.65597141e-01
3.36733013e-01 -6.26403913e-02 6.60766840e-01 5.39251745e-01
8.97715986e-02 9.19809103e-01 5.34081519e-01 -2.77089477e-01
-1.18597341e+00 -1.18199909e+00 1.10429175e-01 6.65206611e-01
4.33252990e-01 1.27124250e-01 -1.01253247e+00 -9.44059014e-01
-9.43767428e-02 6.73699200e-01 -9.10251796e-01 -5.42912304e-01
-5.82519472e-01 -5.00107467e-01 3.36302012e-01 4.73941803e-01
6.72724783e-01 -1.15827227e+00 -1.81997940e-01 -2.95217838e-02
-3.25618297e-01 -1.25990438e+00 -6.80790722e-01 -8.18149596e-02
-8.24303091e-01 -8.35353911e-01 -6.89712882e-01 -8.84185255e-01
1.16193807e+00 7.36806095e-02 1.13310730e+00 -1.56962454e-01
4.92516253e-03 7.94073865e-02 -2.67539322e-01 -2.47057304e-01
-6.17365181e-01 -5.04497290e-02 1.04559295e-01 3.66304606e-01
1.48439869e-01 -8.95494342e-01 -5.87688208e-01 4.85172570e-01
-1.20499516e+00 3.76556456e-01 8.64466608e-01 1.17927945e+00
8.82322311e-01 1.10219695e-01 7.54945874e-01 -1.15032172e+00
3.50519121e-01 -3.59680533e-01 -5.39732456e-01 3.57730061e-01
-3.26660454e-01 1.12131052e-01 8.51369798e-01 -7.05568552e-01
-9.61414397e-01 7.82399103e-02 1.24552868e-01 -9.80417609e-01
-2.90507823e-01 2.31861070e-01 -9.59486187e-01 1.64157063e-01
6.20534539e-01 4.91144836e-01 7.57857636e-02 -3.10797542e-01
6.07223928e-01 3.26474905e-01 7.93786824e-01 -8.29632342e-01
1.36285472e+00 5.52527726e-01 -1.01206668e-01 -2.44625703e-01
-9.02248502e-01 1.50413483e-01 -7.53197610e-01 1.40793890e-01
9.64269519e-01 -9.32077408e-01 -9.90957320e-02 4.80003297e-01
-1.28026354e+00 -2.09105670e-01 -6.50498092e-01 1.41005680e-01
-8.80612433e-01 -5.50298803e-02 -3.22410911e-01 -3.03308994e-01
-1.23766372e-02 -1.40093589e+00 1.39799941e+00 1.91837236e-01
4.19235090e-03 -9.03354883e-01 -2.11942628e-01 4.05753493e-01
2.27050453e-01 6.45804524e-01 9.55806792e-01 -5.13745368e-01
-9.64500785e-01 -1.14347950e-01 -1.82091355e-01 3.97742778e-01
5.57922184e-01 -1.52998507e-01 -9.84174669e-01 -4.14114982e-01
2.13526323e-01 -3.41352671e-01 5.91157556e-01 1.83294222e-01
1.40189934e+00 -6.56105042e-01 -2.07529545e-01 9.29584265e-01
1.29864192e+00 7.04442635e-02 9.11693215e-01 1.48882166e-01
1.12448609e+00 5.56376576e-01 6.10348761e-01 6.20967560e-02
8.16299766e-02 8.06550145e-01 5.57715535e-01 -5.07950485e-01
-2.98977673e-01 -6.72931075e-01 2.63803065e-01 6.34268880e-01
6.21329509e-02 -3.69712293e-01 -5.45482874e-01 4.27214950e-01
-1.73015046e+00 -7.44662702e-01 3.57958853e-01 2.14124608e+00
9.36905086e-01 -3.94006670e-02 -5.20908758e-02 -1.86207741e-01
9.14972842e-01 6.04307838e-02 -8.74350727e-01 6.73013031e-02
-1.19944960e-01 2.21651703e-01 3.71977180e-01 3.10483038e-01
-9.70335901e-01 9.69876230e-01 5.08640766e+00 8.63460720e-01
-1.17761338e+00 2.72259265e-01 7.55012989e-01 6.07782975e-02
-8.17810893e-01 4.74512652e-02 -4.95216340e-01 6.52055383e-01
3.09404880e-01 -9.64095891e-02 5.59511304e-01 6.40993536e-01
-9.98614207e-02 5.43928742e-01 -1.43370056e+00 9.53338921e-01
1.55570507e-01 -1.24579966e+00 5.54634333e-01 3.04943979e-01
1.12035620e+00 -4.09338236e-01 5.74305952e-01 3.57766986e-01
2.58762389e-01 -1.15800166e+00 9.26307380e-01 4.07352746e-01
1.35711896e+00 -6.93230093e-01 5.75660944e-01 3.66906583e-01
-8.05063725e-01 2.45254576e-01 -2.77352452e-01 2.24478632e-01
1.47005558e-01 4.92142975e-01 -4.17900980e-01 7.55751312e-01
4.42435771e-01 7.78124034e-01 -2.08386019e-01 3.30084413e-01
-5.00398159e-01 3.21641535e-01 -7.64520932e-03 4.80991244e-01
1.37438148e-01 -4.55345780e-01 5.85979760e-01 6.29650652e-01
4.37120259e-01 -4.83644148e-03 1.54883042e-01 1.65627229e+00
-5.35606265e-01 -2.95531154e-01 -7.24871397e-01 -1.15682580e-01
3.58096987e-01 1.09649634e+00 -5.45000732e-01 -3.12399209e-01
-3.09496164e-01 1.37737536e+00 3.22218329e-01 5.34120381e-01
-1.13454843e+00 -2.21986160e-01 8.13654363e-01 2.09877282e-01
2.25076526e-01 7.66417906e-02 -5.31960666e-01 -1.41803145e+00
3.51688415e-01 -1.17041492e+00 -3.92659083e-02 -6.88369572e-01
-1.56479168e+00 7.68423140e-01 -1.92728698e-01 -1.67927885e+00
-9.65226069e-02 -3.93608302e-01 -6.73549175e-01 9.47346687e-01
-1.37038803e+00 -1.58574045e+00 -4.33009684e-01 6.48947179e-01
5.50800800e-01 -1.81161433e-01 7.54094779e-01 3.40821624e-01
-4.84671444e-01 9.55053270e-01 -4.26158793e-02 2.73875356e-01
7.19644070e-01 -1.11413288e+00 5.13495982e-01 8.84706378e-01
3.15467507e-01 5.96143782e-01 5.49473584e-01 -5.97979546e-01
-1.40926707e+00 -1.47335422e+00 3.51820320e-01 -8.04606676e-01
4.87186015e-01 -5.64704478e-01 -1.00958562e+00 8.67101967e-01
1.04211748e-01 2.09969178e-01 3.96072626e-01 -3.01244676e-01
-5.40324092e-01 -5.32357581e-03 -1.29780793e+00 6.80945337e-01
1.25436795e+00 -6.96783483e-01 -5.42219102e-01 4.46675062e-01
9.92778063e-01 -5.88405550e-01 -7.44115710e-01 4.73638088e-01
2.81796366e-01 -8.89310837e-01 1.05092037e+00 -9.54436421e-01
9.43503499e-01 -4.53034997e-01 -9.92614403e-02 -1.42265999e+00
-2.41742074e-01 -6.86075449e-01 3.35772298e-02 1.42720795e+00
3.17301095e-01 -7.33388543e-01 6.60204649e-01 5.23563325e-01
-2.00526863e-01 -6.94914937e-01 -8.80682588e-01 -9.03692484e-01
3.71204972e-01 8.39686617e-02 1.03165078e+00 1.25371301e+00
-5.25920153e-01 3.83208990e-01 -5.67803681e-01 3.50906581e-01
6.50542855e-01 4.19239938e-01 9.53855515e-01 -7.98371434e-01
-5.90078831e-01 -4.02944833e-01 -4.52362716e-01 -1.12109804e+00
4.01736110e-01 -9.06301498e-01 1.14662834e-01 -1.29958248e+00
2.57252812e-01 -6.10850394e-01 -1.77227929e-01 5.39403260e-01
-4.12972957e-01 3.74186099e-01 1.20819904e-01 4.80119139e-01
-1.51254967e-01 9.16246593e-01 1.79793763e+00 -3.70192826e-01
1.95986647e-02 -1.24246411e-01 -8.95056427e-01 5.25437474e-01
6.77759826e-01 -5.23995161e-01 -6.05794191e-01 -6.78689957e-01
3.10804751e-02 3.53563391e-02 5.79923868e-01 -7.12671459e-01
-6.05960377e-02 -4.03028399e-01 3.93293530e-01 -2.33108252e-01
1.98487535e-01 -9.90284324e-01 4.42715198e-01 2.32309818e-01
-3.96059662e-01 -2.40235895e-01 2.93322019e-02 6.80434227e-01
-4.70884889e-01 2.73851782e-01 7.66382098e-01 -1.74730256e-01
-3.17638695e-01 6.99695230e-01 3.34376007e-01 2.25819901e-01
1.09400392e+00 -3.06683391e-01 -2.99707949e-01 -2.47056648e-01
-5.70781171e-01 -3.12027112e-02 8.15637231e-01 5.74333727e-01
4.54738140e-01 -1.79816663e+00 -7.90481985e-01 6.16340637e-01
3.02790165e-01 5.98905623e-01 1.54572234e-01 6.82017207e-01
-1.22590452e-01 2.63228838e-04 -4.60154176e-01 -7.56913722e-01
-7.33507276e-01 7.78970003e-01 3.62113535e-01 -2.04951510e-01
-5.07490098e-01 8.66875947e-01 9.79732037e-01 -6.39345288e-01
-1.69985473e-01 -2.11775467e-01 3.09358656e-01 -2.32839793e-01
2.65091568e-01 -2.18464792e-01 -9.63959768e-02 -6.20912552e-01
-1.39884308e-01 6.85952902e-01 -2.46454086e-02 -3.24348696e-02
1.21568024e+00 3.12997960e-02 -1.87537834e-01 2.00148031e-01
1.29997253e+00 6.73055463e-03 -1.54671097e+00 -3.14943790e-01
-5.01234114e-01 -7.71799803e-01 -3.46002847e-01 -6.26474321e-01
-1.37638938e+00 9.14145291e-01 3.03496838e-01 -1.76849246e-01
1.25381386e+00 2.09435895e-01 8.89254391e-01 1.64145231e-02
4.58258122e-01 -6.02248967e-01 2.55079180e-01 1.35406703e-01
1.09712315e+00 -1.33930004e+00 -3.26349258e-01 -5.30657530e-01
-5.58945179e-01 8.16923499e-01 9.25203800e-01 -2.94591010e-01
2.84348726e-01 8.94915387e-02 -3.07546426e-02 -2.07561463e-01
-3.86808068e-01 -4.38179411e-02 4.42920834e-01 8.53191733e-01
1.48142457e-01 2.36882240e-01 1.07132912e-01 5.64061582e-01
-2.86344349e-01 -1.90758631e-01 2.27378637e-01 6.52905345e-01
3.78906429e-02 -1.19155562e+00 -3.79081517e-01 2.37612545e-01
-1.19696170e-01 -4.83460575e-02 -4.77473021e-01 8.03734720e-01
3.03354323e-01 4.47165430e-01 1.56967491e-01 -4.38841015e-01
3.77049655e-01 -2.33987346e-01 6.85250282e-01 -7.30850995e-01
-2.28358060e-01 5.37329465e-02 -4.07664239e-01 -4.51018870e-01
-4.23311740e-01 -5.39358258e-01 -9.81830955e-01 -1.28029227e-01
-3.43674034e-01 -1.06520131e-01 3.09534669e-01 9.03370023e-01
4.96559948e-01 7.67529011e-01 8.28045249e-01 -7.44902790e-01
-6.37271047e-01 -8.75307441e-01 -4.83387500e-01 9.26846981e-01
3.55438948e-01 -7.09223449e-01 -4.55417842e-01 4.04466987e-01] | [11.66914176940918, -0.37902504205703735] |
3d79bf4c-ab1d-48fe-b2f4-87623bf49181 | de-risking-geological-carbon-storage-from | 2211.03527 | null | https://arxiv.org/abs/2211.03527v1 | https://arxiv.org/pdf/2211.03527v1.pdf | De-risking geological carbon storage from high resolution time-lapse seismic to explainable leakage detection | Geological carbon storage represents one of the few truly scalable technologies capable of reducing the CO2 concentration in the atmosphere. While this technology has the potential to scale, its success hinges on our ability to mitigate its risks. An important aspect of risk mitigation concerns assurances that the injected CO2 remains within the storage complex. Amongst the different monitoring modalities, seismic imaging stands out with its ability to attain high resolution and high fidelity images. However, these superior features come, unfortunately, at prohibitive costs and time-intensive efforts potentially rendering extensive seismic monitoring undesirable. To overcome this shortcoming, we present a methodology where time-lapse images are created by inverting non-replicated time-lapse monitoring data jointly. By no longer insisting on replication of the surveys to obtain high fidelity time-lapse images and differences, extreme costs and time-consuming labor are averted. To demonstrate our approach, hundreds of noisy time-lapse seismic datasets are simulated that contain imprints of regular CO2 plumes and irregular plumes that leak. These time-lapse datasets are subsequently inverted to produce time-lapse difference images used to train a deep neural classifier. The testing results show that the classifier is capable of detecting CO2 leakage automatically on unseen data and with a reasonable accuracy. | ['Felix J. Herrmann', 'Mathias Louboutin', 'Abhinav Prakash Gahlot', 'Huseyin Tuna Erdinc', 'Ziyi Yin'] | 2022-10-07 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 3.78206968e-01 -2.02982113e-01 4.78903234e-01 2.60678399e-02
-9.58082736e-01 -7.03734815e-01 8.16435218e-01 1.60023451e-01
-5.14803469e-01 8.63258541e-01 -1.42229080e-01 -4.77396727e-01
-2.86907494e-01 -1.05142760e+00 -6.09236658e-01 -1.05062103e+00
-5.14683962e-01 1.06886238e-01 4.83417839e-01 8.63344688e-03
4.06382978e-01 9.30465758e-01 -1.73367691e+00 -1.32547751e-01
7.48666465e-01 1.12415528e+00 2.96893090e-01 4.87259448e-01
1.78117633e-01 4.45338160e-01 -7.27147400e-01 2.29558319e-01
6.21773422e-01 -2.95496374e-01 -3.72882873e-01 -2.95332581e-01
1.71290711e-01 -4.33979273e-01 -1.01838060e-01 8.79879475e-01
4.34679449e-01 7.42998794e-02 3.71725440e-01 -9.36507046e-01
-1.03395265e-02 1.58203170e-02 -6.27795696e-01 3.81545782e-01
5.07876724e-02 3.83283317e-01 2.80018330e-01 -7.10440099e-01
3.62051666e-01 5.78005731e-01 6.26685619e-01 1.19005516e-01
-1.17355096e+00 -5.73651254e-01 -2.71619022e-01 -1.84735149e-01
-1.16964650e+00 -7.27010310e-01 7.30065525e-01 -4.39046323e-01
1.13442135e+00 5.35564899e-01 7.20529854e-01 7.28255332e-01
2.20990211e-01 1.67479124e-02 1.49017417e+00 -2.82033563e-01
4.22850907e-01 1.34998271e-02 -6.14752412e-01 3.74875516e-01
4.02181357e-01 5.37295520e-01 -4.39980626e-01 -7.19229281e-02
5.48255086e-01 1.16365165e-01 -4.45578396e-01 1.06978133e-01
-9.74035740e-01 4.49389756e-01 4.77644354e-01 5.42220175e-01
-4.14998263e-01 2.73083568e-01 1.99484631e-01 2.58408636e-01
6.89672530e-01 4.66172844e-01 -1.14512436e-01 -2.70276785e-01
-1.22963679e+00 1.89992413e-01 3.89367759e-01 3.85328203e-01
6.83222890e-01 2.26111442e-01 3.93565863e-01 2.91656077e-01
2.23559991e-01 8.57822418e-01 4.39888388e-01 -9.03196812e-01
3.41691822e-01 3.87419611e-01 3.18605453e-01 -1.27354085e+00
-1.12501480e-01 -2.28405073e-01 -8.00968230e-01 6.35854542e-01
3.52353632e-01 -8.79597440e-02 -8.92744839e-01 1.36612582e+00
1.40451804e-01 1.03499785e-01 4.93127517e-02 7.41636574e-01
-4.48002815e-02 8.42709780e-01 -4.67628166e-02 -3.96691352e-01
9.51462805e-01 -3.34596902e-01 -6.60829425e-01 -3.73712003e-01
2.16400817e-01 -2.41236567e-01 7.74766266e-01 2.74965078e-01
-8.86483908e-01 -3.56387626e-03 -1.34644353e+00 3.81418318e-01
-8.73668253e-01 -4.41085339e-01 5.42640626e-01 7.55934656e-01
-6.85355544e-01 8.78142536e-01 -1.15015948e+00 4.57965471e-02
4.08823490e-01 5.69697358e-02 -3.52205545e-01 3.06822032e-01
-1.15756297e+00 8.90794396e-01 1.05918944e-01 4.44036692e-01
-1.15031803e+00 -8.79202068e-01 -6.08780086e-01 5.39944544e-02
2.17607141e-01 -2.33964156e-02 8.14962447e-01 -2.04012856e-01
-1.14612997e+00 5.22304356e-01 -2.17455835e-03 -5.13857245e-01
8.46935749e-01 1.74772572e-02 -4.45808470e-01 4.88558888e-01
1.71384126e-01 1.98264927e-01 7.48417795e-01 -1.27731824e+00
-6.57539666e-01 -4.57890540e-01 -4.39404517e-01 4.41868864e-02
-3.12471658e-01 9.59970206e-02 1.53311864e-01 -2.45201349e-01
3.18040103e-01 -8.65806520e-01 -1.90139934e-01 1.37800366e-01
-1.36519685e-01 6.23309135e-01 1.19042361e+00 -5.97144306e-01
8.07422876e-01 -2.07804251e+00 -3.95776808e-01 1.01684518e-01
-8.43576435e-03 2.66487092e-01 4.00474459e-01 4.78116721e-01
1.84867874e-01 5.87175369e-01 -9.36946094e-01 -1.81758404e-01
-3.01741213e-01 1.06413275e-01 -3.93476576e-01 6.78067029e-01
5.11069357e-01 5.72766066e-01 -9.96101439e-01 -1.22842856e-01
3.29104781e-01 4.90998775e-01 1.37804031e-01 1.51303113e-01
-3.31902914e-02 7.08998144e-01 -1.83180213e-01 7.57224143e-01
7.77535439e-01 1.68372154e-01 6.23157918e-02 1.27652019e-01
-7.39181757e-01 2.36584947e-01 -1.03485394e+00 1.20396709e+00
-3.01290125e-01 5.89390874e-01 2.76128381e-01 -9.63776112e-01
1.02788198e+00 3.50871027e-01 4.41466928e-01 -1.21503842e+00
-3.79115373e-01 7.05112636e-01 -4.90004748e-01 -7.29885757e-01
5.01176775e-01 -7.79150784e-01 8.39751679e-03 4.63972658e-01
-5.32290995e-01 -6.37507796e-01 -2.23710969e-01 -1.14976145e-01
1.09023619e+00 -1.27105787e-03 -3.75784039e-01 -3.05409700e-01
3.45237285e-01 6.84027523e-02 3.50473762e-01 5.74767232e-01
-2.83149898e-01 6.79253936e-01 1.01462893e-01 -5.25464296e-01
-1.31814349e+00 -9.28998768e-01 -1.93508804e-01 2.63166845e-01
-5.63070476e-02 2.69223571e-01 -2.25426316e-01 -3.09109509e-01
-5.64886928e-02 5.43891549e-01 -5.60365677e-01 4.97458316e-02
-5.60591817e-01 -1.03603959e+00 8.13982725e-01 4.76622730e-01
7.89635420e-01 -8.53426218e-01 -1.41571093e+00 2.88916081e-01
-7.28060827e-02 -5.25191247e-01 3.24373662e-01 5.57938993e-01
-9.50389147e-01 -9.44819510e-01 -6.75969839e-01 -1.65484801e-01
5.74346900e-01 4.18971211e-01 7.42879212e-01 4.01192158e-02
-5.13124764e-01 -8.21602996e-04 -9.04309526e-02 -6.29843295e-01
-3.39492500e-01 -4.01079714e-01 -9.51157883e-03 -2.59828538e-01
1.16484962e-01 -8.45087409e-01 -7.86952913e-01 8.04776922e-02
-1.23000669e+00 -2.25373223e-01 3.05861294e-01 6.47288322e-01
6.55093014e-01 4.34569329e-01 6.24434769e-01 -4.41589892e-01
4.01491880e-01 -5.30929029e-01 -1.00032687e+00 8.76948312e-02
-6.77169740e-01 -1.30831078e-01 4.11372453e-01 -1.01164103e-01
-1.20732784e+00 4.42582257e-02 1.21957317e-01 -1.96495444e-01
-2.18409345e-01 8.31413150e-01 1.05349600e-01 -8.01986307e-02
6.50035381e-01 3.97223324e-01 1.58433050e-01 -4.23893631e-01
-8.85711312e-02 6.77645862e-01 8.37308347e-01 -3.86127353e-01
9.99460042e-01 9.97339368e-01 1.59237683e-01 -1.14463067e+00
-3.94855469e-01 -1.31761953e-01 -5.16472518e-01 -6.67194128e-01
7.30094433e-01 -7.87087262e-01 -2.59596020e-01 6.56744838e-01
-7.07061529e-01 -1.99557364e-01 -2.80872345e-01 2.71314055e-01
-1.12242922e-01 2.03305423e-01 -1.29111603e-01 -1.38792574e+00
-1.73421696e-01 -9.49508488e-01 8.31459403e-01 3.05490047e-01
3.37290734e-01 -7.92421818e-01 5.77406399e-02 1.92252591e-01
9.05161083e-01 9.89473283e-01 5.49705327e-01 -2.50665426e-01
-9.01395917e-01 -5.36778927e-01 -4.55989093e-02 2.61587441e-01
4.73505795e-01 5.20348884e-02 -1.42609644e+00 -3.31410438e-01
5.30211985e-01 -2.53669918e-01 8.26799750e-01 1.95871834e-02
1.06993806e+00 -1.77102804e-01 -2.70230353e-01 6.47248745e-01
1.72004569e+00 6.91419959e-01 8.29677343e-01 7.44473040e-01
1.69384450e-01 8.18251789e-01 5.67240834e-01 2.97474086e-01
-2.62483716e-01 1.71284020e-01 8.94392371e-01 -3.38961221e-02
2.39032373e-01 -1.20789558e-01 3.09843309e-02 4.49212909e-01
-2.25911289e-01 -1.19059689e-01 -1.21539915e+00 1.04429805e+00
-1.35095894e+00 -1.16283858e+00 -2.35887945e-01 2.56482553e+00
6.14349604e-01 5.26024401e-02 -1.94236144e-01 4.51968551e-01
4.60428476e-01 4.45859671e-01 -5.94988823e-01 -3.58508015e-03
-1.72406912e-01 3.51083092e-02 8.06769490e-01 4.12793338e-01
-1.07630777e+00 3.13994557e-01 5.90431261e+00 7.26165175e-02
-1.57257557e+00 1.04817329e-02 6.95720851e-01 -8.67035836e-02
-5.50645947e-01 1.41392630e-02 -3.20280820e-01 6.39419377e-01
1.42158568e+00 -2.21802399e-01 2.78670311e-01 3.93160582e-01
5.02139866e-01 -6.58228278e-01 -6.05702400e-01 5.44646263e-01
-2.90015250e-01 -1.44380736e+00 -4.75148231e-01 1.50813729e-01
4.21526194e-01 2.78792351e-01 1.22416876e-02 -2.79269338e-01
-3.25248763e-02 -9.51720119e-01 8.16193819e-01 6.86247468e-01
9.66817558e-01 -7.15090096e-01 6.65406883e-01 4.57126319e-01
-8.86496961e-01 -2.27334395e-01 -2.70870000e-01 -2.18457103e-01
3.50188404e-01 8.85930657e-01 -6.05383396e-01 6.00459397e-01
9.93725300e-01 1.96834907e-01 -5.04054606e-01 9.56101120e-01
-6.23088703e-02 4.85234290e-01 -9.40621555e-01 2.06472948e-01
3.04960400e-01 -2.58484811e-01 3.64697874e-01 1.04780996e+00
7.90716290e-01 2.44270906e-01 -4.05045986e-01 1.09632611e+00
2.86205113e-01 -6.52382314e-01 -1.17062807e+00 -2.27724373e-01
7.93489337e-01 8.38515878e-01 -7.72967219e-01 -1.87552646e-01
-2.04947963e-01 6.58829868e-01 -2.45456278e-01 5.26619963e-02
-6.12348020e-01 -5.08610308e-01 3.47506404e-01 3.44577432e-01
2.35185459e-01 -3.96360457e-01 -3.86691988e-01 -8.29030275e-01
3.19395751e-01 -5.32039464e-01 7.48532936e-02 -5.65149665e-01
-1.00152004e+00 4.66397971e-01 7.07784370e-02 -1.21635354e+00
1.60660669e-02 -3.49104196e-01 -6.56314313e-01 1.02053213e+00
-1.78105938e+00 -8.20726395e-01 -7.25123584e-01 1.85615242e-01
5.00751957e-02 1.84536815e-01 8.43195558e-01 4.09249634e-01
-3.82147700e-01 1.33877695e-02 3.57463211e-01 -2.55269349e-01
3.95323992e-01 -1.09276104e+00 3.05171937e-01 1.07399631e+00
-4.06108975e-01 5.05276978e-01 7.24982023e-01 -9.68154371e-01
-1.37096190e+00 -1.19414485e+00 7.77665079e-01 -1.85451344e-01
7.74001539e-01 -3.69857758e-01 -1.23521948e+00 4.16817993e-01
-1.38237298e-01 3.38086784e-01 3.49183470e-01 -6.71522260e-01
-7.46500716e-02 -2.84768879e-01 -1.47217357e+00 2.38410771e-01
5.91569841e-01 -8.83020043e-01 -5.52734137e-01 3.40937018e-01
4.50859040e-01 -2.39276379e-01 -9.44884777e-01 4.89129543e-01
5.01568556e-01 -1.26300454e+00 7.98145294e-01 -1.84519246e-01
5.40034950e-01 -5.40410399e-01 -2.29482383e-01 -1.00287485e+00
2.56202847e-01 -5.14692307e-01 1.23564854e-01 1.25981593e+00
5.84253192e-01 -7.86676764e-01 8.40151608e-01 8.52366328e-01
-1.37180924e-01 -3.43543857e-01 -1.32331502e+00 -1.20063615e+00
2.59616345e-01 -2.89887190e-01 6.36679113e-01 9.33403432e-01
-1.31520510e-01 -4.58087295e-01 -1.98921606e-01 7.01432288e-01
8.36003125e-01 8.44484847e-03 3.67964000e-01 -1.17610908e+00
-3.95474136e-02 1.05140135e-01 -6.96779937e-02 -2.10041448e-01
-5.89925647e-01 -3.31855237e-01 4.08164203e-01 -1.20624292e+00
-9.81881171e-02 -5.81063032e-01 -3.01252246e-01 4.07619476e-01
1.98095944e-02 3.92814428e-01 6.44603595e-02 3.81487340e-01
1.78430557e-01 3.94352645e-01 6.82640851e-01 -2.87090868e-01
-1.95718929e-01 -3.72491717e-01 -5.17993048e-02 4.45455909e-01
8.80041897e-01 -7.43920445e-01 -2.61138380e-01 -7.28319347e-01
3.06681991e-01 3.23020905e-01 5.85579455e-01 -1.44890165e+00
3.08220446e-01 -2.05275461e-01 2.96064615e-01 -7.59620547e-01
3.59191597e-01 -1.02907979e+00 7.93804944e-01 6.46923900e-01
-1.71260759e-02 3.05631552e-02 3.05807263e-01 7.12053120e-01
-4.04799938e-01 -3.43907923e-01 1.01543880e+00 -6.48871541e-01
-3.12099606e-01 1.26363471e-01 -6.18614316e-01 -4.46183592e-01
1.18162096e+00 -3.43604892e-01 -5.92730999e-01 -8.13486874e-02
-4.30722177e-01 -6.38175234e-02 7.43928194e-01 1.45652935e-01
6.14058435e-01 -8.65410566e-01 -3.02103400e-01 3.40687722e-01
-6.76375329e-02 4.19194043e-01 3.12375277e-01 6.47719860e-01
-8.48671973e-01 2.69112080e-01 -1.09976776e-01 -6.75275087e-01
-6.86630070e-01 4.27727550e-01 6.30897284e-01 -7.05764443e-02
-8.57965291e-01 8.23498666e-01 -3.81024063e-01 -2.00812906e-01
-3.95821072e-02 -1.46674260e-01 1.69511974e-01 2.57177293e-01
7.87392437e-01 3.92177343e-01 4.71995413e-01 -2.03121766e-01
-2.89031029e-01 2.15262175e-01 3.12947720e-01 -4.60761696e-01
1.90700865e+00 -1.23883232e-01 -2.16324046e-01 6.09773040e-01
1.01833093e+00 3.09153218e-02 -1.55401886e+00 2.91943192e-01
3.48474950e-01 -8.98940563e-01 2.66086578e-01 -9.60181832e-01
-1.20659053e+00 1.04124248e+00 8.94627750e-01 5.00285149e-01
1.16900313e+00 -3.31387818e-01 6.60099208e-01 3.49375546e-01
4.40819383e-01 -1.04348624e+00 -3.60959508e-02 1.61037713e-01
8.73717904e-01 -1.10084784e+00 8.96813124e-02 1.19794704e-01
-4.82188165e-02 9.71668124e-01 2.62366235e-01 7.94705600e-02
4.67372149e-01 6.81557834e-01 2.10049331e-01 -6.18271887e-01
-4.60075200e-01 2.34932303e-01 -8.19233418e-01 7.74942875e-01
2.69763991e-02 -4.73958217e-02 -6.75691888e-02 -1.06473885e-01
3.38202380e-02 -1.41594470e-01 7.46285677e-01 1.34021366e+00
-6.12948716e-01 -6.54565632e-01 -6.32341623e-01 3.94806564e-01
-5.35692155e-01 1.01716489e-01 -2.29323849e-01 6.40985310e-01
-4.32932042e-02 8.82093966e-01 1.10676922e-01 -9.87445563e-02
2.81730264e-01 1.59282252e-01 -1.27487332e-01 -1.61185607e-01
-5.13346255e-01 3.80358994e-02 2.23278180e-01 -5.31419396e-01
-4.75274771e-01 -8.14577878e-01 -1.13331485e+00 -1.81276232e-01
-3.79581898e-01 2.84604013e-01 1.15320480e+00 6.80346251e-01
2.40142763e-01 4.01346385e-01 8.42024267e-01 -1.20531988e+00
-6.67587221e-01 -7.66907215e-01 -6.95241690e-01 3.47819179e-01
8.25267076e-01 -5.57917237e-01 -9.03657377e-01 -7.31344521e-02] | [6.843639373779297, 2.646791696548462] |
415e4e8e-77f5-4a93-bfb7-e2a2a23ed8a5 | faster-unsupervised-semantic-inpainting-a-gan | 1908.04968 | null | https://arxiv.org/abs/1908.04968v1 | https://arxiv.org/pdf/1908.04968v1.pdf | Faster Unsupervised Semantic Inpainting: A GAN Based Approach | In this paper, we propose to improve the inference speed and visual quality of contemporary baseline of Generative Adversarial Networks (GAN) based unsupervised semantic inpainting. This is made possible with better initialization of the core iterative optimization involved in the framework. To our best knowledge, this is also the first attempt of GAN based video inpainting with consideration to temporal cues. On single image inpainting, we achieve about 4.5-5$\times$ speedup and 80$\times$ on videos compared to baseline. Simultaneously, our method has better spatial and temporal reconstruction qualities as found on three image and one video dataset. | ['Arnav Kumar Jain', 'Avisek Lahiri', 'Prabir Kumar Biswas', 'Divyasri Nadendla'] | 2019-08-14 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [ 3.54850501e-01 2.79246449e-01 3.71969678e-02 -1.29649997e-01
-8.90508711e-01 -3.28659713e-01 5.78225851e-01 -5.97047806e-01
-3.45990479e-01 1.05253458e+00 1.36608273e-01 -1.19084731e-01
1.95259154e-01 -7.69828737e-01 -1.15197754e+00 -5.23976088e-01
4.62312102e-02 2.61124104e-01 -4.42478321e-02 -1.86102837e-01
-3.54788564e-02 3.30383718e-01 -1.19917917e+00 3.45232695e-01
7.08503366e-01 1.06571364e+00 1.32239744e-01 7.93062925e-01
6.98041469e-02 1.23155630e+00 -5.18808186e-01 -5.18506944e-01
4.63193476e-01 -7.91198134e-01 -8.13137770e-01 2.44158119e-01
6.96860433e-01 -8.14336181e-01 -8.86419058e-01 1.00170493e+00
2.49118209e-01 2.87092149e-01 4.21034843e-01 -1.07965207e+00
-7.15723395e-01 6.70073152e-01 -6.48936868e-01 2.88360506e-01
3.06002140e-01 3.15580606e-01 7.11924911e-01 -9.03994799e-01
1.00819659e+00 1.16555345e+00 5.82080126e-01 8.66424084e-01
-1.16624534e+00 -7.44138896e-01 5.35845309e-02 2.02321470e-01
-1.33786201e+00 -6.51589096e-01 1.02069640e+00 -1.06842861e-01
9.73924398e-01 9.63078216e-02 6.54270232e-01 1.31463206e+00
1.38331890e-01 8.21406245e-01 1.16655421e+00 -4.25808430e-01
1.93369046e-01 -3.32095802e-01 -6.50955915e-01 7.29210913e-01
-1.83168650e-01 5.77360332e-01 -6.98375940e-01 3.01559746e-01
1.18040383e+00 -5.96777499e-02 8.35004449e-02 4.03564051e-02
-1.00195980e+00 8.86480570e-01 5.84669113e-01 1.03435010e-01
-4.67050195e-01 9.67278600e-01 2.84753591e-01 4.77212071e-01
6.01546943e-01 2.84268886e-01 -7.44992681e-03 -3.93206656e-01
-1.45746136e+00 2.66724795e-01 3.26556087e-01 1.15382707e+00
6.22735381e-01 8.47184718e-01 -1.38982944e-02 5.42890072e-01
1.75931782e-01 6.06383681e-01 2.39514455e-01 -1.51691556e+00
3.05507779e-01 -1.14490360e-01 1.36962980e-01 -1.03506565e+00
2.44249448e-01 -1.53774261e-01 -9.08309817e-01 4.66429830e-01
2.12870747e-01 -1.82857797e-01 -1.18257833e+00 1.67674303e+00
-8.41236711e-02 6.20059371e-01 -1.67406216e-01 7.51062751e-01
6.38219893e-01 6.80512547e-01 7.52381384e-02 -3.43404949e-01
9.00942504e-01 -1.20591378e+00 -1.04588497e+00 -3.37414414e-01
-9.57706571e-02 -9.55079079e-01 6.95670009e-01 4.06934530e-01
-1.53716683e+00 -7.18773961e-01 -1.20864713e+00 8.59287381e-03
5.43263555e-02 -2.99475938e-01 8.61256242e-01 5.78850150e-01
-1.31858134e+00 8.40392351e-01 -1.08882868e+00 -2.19864964e-01
6.77694142e-01 3.24079990e-01 -2.71423429e-01 -2.60158539e-01
-1.01429999e+00 6.86859608e-01 1.93839461e-01 -3.63144018e-02
-1.39570403e+00 -6.81206286e-01 -7.45188951e-01 -4.34287071e-01
3.75568122e-01 -8.26306105e-01 1.23860276e+00 -1.27609372e+00
-1.74323845e+00 6.51777625e-01 -2.13499278e-01 -1.01305676e+00
7.32319713e-01 -5.11972785e-01 -3.97324443e-01 4.90703702e-01
7.37822652e-02 1.27727830e+00 1.27008402e+00 -1.33044636e+00
-1.47492960e-01 2.54702568e-02 6.99291527e-02 2.17728913e-02
1.53394312e-01 3.33002135e-02 -6.79653406e-01 -1.33054423e+00
-2.21366942e-01 -9.12006259e-01 -3.75182718e-01 1.97829440e-01
-1.81369886e-01 4.45512742e-01 9.90158498e-01 -1.15896535e+00
8.47006977e-01 -1.97104633e+00 3.86229217e-01 -1.76029235e-01
1.54437006e-01 2.50218034e-01 -7.15035871e-02 5.04543185e-01
-4.40521911e-02 1.43947393e-01 -1.98947623e-01 -7.52640963e-01
-1.62074447e-01 3.91443074e-01 -5.14941037e-01 3.77693474e-01
2.06310511e-01 1.29230893e+00 -7.11754084e-01 -4.26868856e-01
4.97911632e-01 5.29062331e-01 -7.41920054e-01 2.07743436e-01
-3.43452185e-01 7.40702152e-01 -1.57295823e-01 7.96987534e-01
7.13885307e-01 1.73673965e-02 -2.35753767e-02 -5.25608324e-02
3.63418400e-01 -9.05983970e-02 -6.52320445e-01 2.43398881e+00
-5.14213026e-01 9.83938098e-01 8.33712965e-02 -1.10524094e+00
6.38446867e-01 4.76898611e-01 5.40125608e-01 -8.02133799e-01
1.68796495e-01 -4.04567420e-02 -2.72628486e-01 -1.53445840e-01
5.15152931e-01 -4.52942073e-01 1.08913064e-01 2.06254125e-01
3.56847405e-01 -3.43258709e-01 2.60579195e-02 3.23975593e-01
1.14377117e+00 4.83938277e-01 -5.78331649e-02 -1.81995437e-01
1.97903037e-01 -1.12584054e-01 2.96809375e-01 7.78721094e-01
-1.54132634e-01 9.20728803e-01 1.54027820e-01 -4.21961069e-01
-1.47247624e+00 -1.29113102e+00 1.58779308e-01 5.15589476e-01
9.62287635e-02 -4.01109189e-01 -9.49181318e-01 -6.60615325e-01
-3.10294330e-01 7.37959743e-01 -7.29957104e-01 -2.69884057e-03
-9.02601063e-01 -3.36755455e-01 6.90502763e-01 7.66997576e-01
7.39622653e-01 -1.32966924e+00 -2.82631874e-01 3.65884125e-01
-1.88662827e-01 -1.44193542e+00 -5.96377492e-01 -1.94092005e-01
-1.22614110e+00 -7.64092326e-01 -6.36967063e-01 -6.91938102e-01
4.98475224e-01 -3.40446942e-02 1.30705249e+00 2.69181877e-02
-3.73988271e-01 2.33111307e-01 -4.35624212e-01 -1.16650984e-01
-4.55041647e-01 -2.87275612e-01 -1.78457737e-01 -2.90410638e-01
-2.81452328e-01 -1.12375045e+00 -9.01317835e-01 6.56437129e-02
-1.07979393e+00 2.51384825e-01 5.56698918e-01 1.00255239e+00
6.21284604e-01 1.32809594e-01 3.23661953e-01 -7.00079799e-01
2.44641587e-01 -3.53106171e-01 -3.82260323e-01 -2.42265388e-01
-6.36204779e-01 5.85119873e-02 8.61807048e-01 -2.87500441e-01
-1.03338659e+00 -2.73011494e-02 -3.09253633e-01 -9.69416738e-01
-1.22890756e-01 8.87035280e-02 1.07030660e-01 -2.26317346e-01
4.58646894e-01 4.13265973e-01 -4.61062267e-02 -4.43305641e-01
5.60274839e-01 -3.14942226e-02 7.21000910e-01 -5.24402916e-01
9.97296453e-01 8.35268617e-01 -3.84652391e-02 -5.49321115e-01
-4.27395940e-01 1.31315261e-01 -3.82416308e-01 -2.82301366e-01
9.21267271e-01 -1.09081209e+00 -3.89401376e-01 5.43632507e-01
-1.01235962e+00 -4.79893565e-01 -4.88054395e-01 2.42111579e-01
-9.37440991e-01 4.28814352e-01 -1.01889586e+00 -5.27082741e-01
-4.67396080e-01 -1.18551588e+00 1.00477862e+00 -2.39438266e-02
-3.18332575e-04 -9.62102890e-01 -8.71612579e-02 5.14766335e-01
5.63016653e-01 6.25395060e-01 2.18274891e-01 2.91487634e-01
-9.12042558e-01 1.45243362e-01 -1.36167094e-01 5.56563139e-01
6.51979446e-02 -1.65344566e-01 -8.31132352e-01 -5.52401543e-01
1.22847915e-01 -2.74604261e-01 1.02666080e+00 5.39467871e-01
1.11311352e+00 -4.14073706e-01 -1.52490705e-01 9.25287247e-01
1.69261730e+00 5.76255560e-01 1.27811718e+00 2.39593953e-01
7.74298549e-01 -2.81552412e-03 6.57976389e-01 4.04390931e-01
-1.09232917e-01 8.17530513e-01 6.55956507e-01 -1.82281747e-01
-5.97460330e-01 -5.33823013e-01 5.49955428e-01 6.88481092e-01
-3.72210681e-01 -3.00157249e-01 -4.86906052e-01 5.55338860e-01
-1.69958627e+00 -1.15751493e+00 4.83621478e-01 1.69892764e+00
7.71694779e-01 1.96108874e-02 3.56736779e-02 4.10052091e-02
5.72565615e-01 4.37913269e-01 -5.42567551e-01 -3.94714803e-01
9.08831879e-03 9.44543839e-01 5.57224333e-01 6.13842785e-01
-9.64589477e-01 1.21895266e+00 6.98783207e+00 9.95446146e-01
-9.81141984e-01 3.76829267e-01 7.85506248e-01 -3.82182956e-01
-3.14725399e-01 8.56014043e-02 -2.65928924e-01 6.88372195e-01
9.29204226e-01 1.53936222e-01 8.96705389e-01 7.47677386e-01
1.14874616e-01 3.41070816e-02 -8.81738365e-01 1.07088041e+00
2.04947829e-01 -1.67867208e+00 1.65097728e-01 8.93745199e-02
9.92875457e-01 -1.66812837e-01 2.38154605e-01 6.37505800e-02
3.21571857e-01 -1.57070065e+00 9.49631691e-01 4.77953494e-01
9.84029651e-01 -1.21024096e+00 5.71377993e-01 -6.74912035e-02
-1.06838202e+00 4.11627442e-01 -2.30181534e-02 -5.16177528e-02
5.33727527e-01 2.40364671e-01 -5.49048305e-01 5.89476943e-01
7.47337699e-01 7.60673881e-01 -3.19316566e-01 5.41113138e-01
-3.46035123e-01 7.57971466e-01 -3.23554724e-02 4.56984133e-01
3.11106622e-01 -1.34703979e-01 4.91116405e-01 8.70883048e-01
4.81274426e-01 1.77854717e-01 -5.73671982e-02 7.78611481e-01
-2.33722806e-01 -2.81005770e-01 -7.94385076e-01 -1.55620664e-01
2.15791106e-01 7.12720990e-01 -6.30608082e-01 -3.55267763e-01
-2.66166806e-01 1.66557693e+00 1.77194193e-01 3.51993799e-01
-1.33555913e+00 -1.56717777e-01 5.71632981e-01 9.36125964e-02
7.37720430e-01 -4.08506453e-01 -2.53758669e-01 -1.20358074e+00
3.81322205e-02 -9.61979151e-01 1.45025268e-01 -8.24741900e-01
-1.02287662e+00 6.26998425e-01 -9.45350826e-02 -1.13306808e+00
-4.69402134e-01 -3.79211187e-01 -3.97875428e-01 5.95533311e-01
-1.12189615e+00 -1.31099916e+00 -2.45429784e-01 7.41132796e-01
9.14518654e-01 -1.81887373e-01 7.73506761e-01 3.36350441e-01
-3.28418016e-01 7.65470386e-01 5.49724624e-02 1.20624769e-02
5.79563320e-01 -9.96162713e-01 6.65614784e-01 1.38218498e+00
3.19667846e-01 3.38452399e-01 1.01397264e+00 -5.04844487e-01
-1.61526215e+00 -1.07582259e+00 2.74955869e-01 -1.79058909e-01
5.54039121e-01 -1.23837322e-01 -4.51176137e-01 9.61499631e-01
9.59771276e-01 -1.29605178e-02 2.64173210e-01 -4.90439385e-01
-1.97330758e-01 -6.47315979e-02 -1.47068393e+00 6.45721734e-01
1.31386507e+00 -6.25028968e-01 -2.42884427e-01 2.41451666e-01
8.48604381e-01 -5.96851110e-01 -9.81665850e-01 3.62991452e-01
2.93016821e-01 -1.08352733e+00 1.31138766e+00 -2.82534480e-01
8.79068971e-01 -3.25508237e-01 -3.39220822e-01 -1.02342582e+00
-1.00890025e-01 -1.09025395e+00 -3.19908887e-01 1.07015777e+00
4.54933047e-02 -2.10502118e-01 1.05852747e+00 5.44059165e-02
-1.70439765e-01 -5.43451786e-01 -1.01503241e+00 -6.93068385e-01
1.15039401e-01 -7.32756555e-01 2.69779027e-01 7.37451196e-01
-5.25253654e-01 -5.96813969e-02 -1.21162212e+00 -7.99313188e-02
9.12595928e-01 1.18104033e-02 6.44216537e-01 -3.94311070e-01
-7.34995544e-01 -1.75706506e-01 -5.03751993e-01 -1.11950850e+00
1.89811677e-01 -4.71734941e-01 -3.61009873e-02 -1.26079297e+00
7.69569129e-02 -7.60822371e-02 -2.59191632e-01 3.09990406e-01
1.31531402e-01 9.22156751e-01 4.48530585e-01 3.32553014e-02
-4.92255300e-01 7.22899675e-01 1.49759936e+00 -2.69741416e-01
2.37725288e-01 -5.72430611e-01 -5.34227490e-01 4.70392168e-01
9.00084615e-01 -3.99485260e-01 -5.40914655e-01 -7.06422925e-01
-2.25686550e-01 6.24681950e-01 5.18674552e-01 -1.14400244e+00
-3.13608646e-02 -2.05838591e-01 5.33335447e-01 -4.59509343e-01
7.43613899e-01 -6.29436493e-01 7.32943475e-01 5.31496644e-01
-2.05557272e-02 2.72735715e-01 5.63171506e-01 6.63344622e-01
-4.67270911e-01 9.74790603e-02 9.60069776e-01 -3.23534429e-01
-9.11148310e-01 3.42529595e-01 -2.06214845e-01 5.52603751e-02
1.00813675e+00 -1.73360616e-01 -2.70697437e-02 -7.25847483e-01
-7.94446468e-01 -2.99057394e-01 7.83322573e-01 3.34986925e-01
8.56845737e-01 -1.56752801e+00 -5.96101582e-01 7.16556311e-02
-4.78176504e-01 -2.17699930e-02 4.54870045e-01 7.76563644e-01
-1.00749731e+00 3.01105499e-01 -5.39694786e-01 -4.47862923e-01
-1.02448201e+00 6.66718423e-01 1.60099477e-01 -2.94970930e-01
-7.42930830e-01 9.41742301e-01 8.13487172e-02 3.16867590e-01
2.71569658e-02 7.57356035e-03 4.52335864e-01 -4.22510743e-01
3.71069044e-01 3.26484472e-01 -1.76149487e-01 -5.55465639e-01
-2.51490653e-01 4.87211049e-01 -2.23039865e-01 -5.75151324e-01
1.31738853e+00 1.97620299e-02 -6.49601314e-03 -2.38326490e-02
1.10475349e+00 4.35725711e-02 -1.71211958e+00 1.46436065e-01
-6.12489164e-01 -6.73174560e-01 3.25768627e-02 -6.97750747e-01
-1.59271169e+00 5.69123864e-01 5.88509917e-01 -2.80197561e-01
1.42213535e+00 -8.52532312e-02 1.16505075e+00 -1.09959617e-01
5.74586749e-01 -9.13819075e-01 2.98445165e-01 1.28097638e-01
9.49647784e-01 -1.05266500e+00 7.18681514e-02 -3.43687415e-01
-4.52793658e-01 7.55039692e-01 5.17068624e-01 -7.60591984e-01
3.86347711e-01 5.07710695e-01 -1.74571108e-02 -1.93741433e-02
-7.47450709e-01 2.31647521e-01 1.64185882e-01 6.39131904e-01
1.44207746e-01 -1.68964006e-02 -3.83563995e-01 -3.83909531e-02
-1.84688851e-01 3.83448303e-02 2.36174926e-01 9.05177951e-01
2.88213827e-02 -1.32873464e+00 -1.03254795e-01 3.52063738e-02
-7.61871696e-01 -3.82957101e-01 -8.56572110e-03 1.00181866e+00
6.14938252e-02 9.17838454e-01 -8.43595043e-02 -5.33843994e-01
-5.26925400e-02 -2.01690465e-01 9.29704905e-01 -6.47907704e-02
-4.52786744e-01 1.50922909e-01 7.88061023e-02 -1.00520349e+00
-6.77705050e-01 -5.27705193e-01 -1.10580659e+00 -5.99484444e-01
1.08083382e-01 -1.27143055e-01 4.36452925e-01 7.78547227e-01
4.26445723e-01 6.75203502e-01 5.35709620e-01 -1.05482066e+00
-8.22860003e-02 -8.77892256e-01 -4.44119483e-01 5.72919011e-01
4.93496098e-03 -5.35163343e-01 -3.56202841e-01 4.33147222e-01] | [11.29134750366211, -0.6649293303489685] |
67d70ada-74b3-424f-a334-19eb928b425a | crysgnn-distilling-pre-trained-knowledge-to | 2301.05852 | null | https://arxiv.org/abs/2301.05852v1 | https://arxiv.org/pdf/2301.05852v1.pdf | CrysGNN : Distilling pre-trained knowledge to enhance property prediction for crystalline materials | In recent years, graph neural network (GNN) based approaches have emerged as a powerful technique to encode complex topological structure of crystal materials in an enriched representation space. These models are often supervised in nature and using the property-specific training data, learn relationship between crystal structure and different properties like formation energy, bandgap, bulk modulus, etc. Most of these methods require a huge amount of property-tagged data to train the system which may not be available for different properties. However, there is an availability of a huge amount of crystal data with its chemical composition and structural bonds. To leverage these untapped data, this paper presents CrysGNN, a new pre-trained GNN framework for crystalline materials, which captures both node and graph level structural information of crystal graphs using a huge amount of unlabelled material data. Further, we extract distilled knowledge from CrysGNN and inject into different state of the art property predictors to enhance their property prediction accuracy. We conduct extensive experiments to show that with distilled knowledge from the pre-trained model, all the SOTA algorithms are able to outperform their own vanilla version with good margins. We also observe that the distillation process provides a significant improvement over the conventional approach of finetuning the pre-trained model. We have released the pre-trained model along with the large dataset of 800K crystal graph which we carefully curated; so that the pretrained model can be plugged into any existing and upcoming models to enhance their prediction accuracy. | ['Niloy Ganguly', 'Satadeep Bhattacharjee', 'Seung-Cheol Lee', 'Pawan Goyal', 'Bidisha Samanta', 'Kishalay Das'] | 2023-01-14 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 2.44392291e-01 2.88470984e-01 -4.41386998e-01 -2.38326624e-01
-4.72597957e-01 -2.94937879e-01 2.97479123e-01 2.65752077e-01
-8.58408585e-03 1.08102322e+00 1.51425511e-01 -2.24030092e-01
-2.19534591e-01 -1.30056274e+00 -9.91554081e-01 -9.89463985e-01
-2.81911969e-01 6.31522477e-01 2.13336214e-01 -4.98588920e-01
3.49007159e-01 4.61713880e-01 -1.54696083e+00 1.51202574e-01
9.98275876e-01 1.08753896e+00 4.66039807e-01 5.34528494e-01
2.00803708e-02 8.45208704e-01 -7.94785917e-02 1.55905476e-02
2.59140909e-01 -8.66995007e-02 -8.90032828e-01 -3.26454222e-01
3.24967265e-01 1.13538176e-01 -5.05119026e-01 7.79333174e-01
4.92657989e-01 -2.93854941e-02 8.34332108e-01 -7.83089161e-01
-1.01614594e+00 9.69793260e-01 -2.86411792e-01 -4.35129143e-02
3.64113808e-01 2.46925652e-01 1.26255429e+00 -6.07160747e-01
9.71626103e-01 6.67299032e-01 5.26263833e-01 4.63794172e-01
-1.24602163e+00 -6.97494924e-01 -2.19202563e-01 3.99857670e-01
-1.03230321e+00 -3.77126895e-02 1.16373646e+00 -3.11733067e-01
1.32970917e+00 -3.01865023e-02 9.40038204e-01 1.15380049e+00
1.63749620e-01 4.18912709e-01 1.15681374e+00 -4.32612479e-01
1.68186188e-01 -1.33359089e-01 2.84907699e-01 8.91630650e-01
4.63728279e-01 3.17243725e-01 -5.64017832e-01 -3.74305882e-02
6.79909229e-01 -1.89232901e-01 -4.28166568e-01 -7.71870792e-01
-1.08662581e+00 8.21926296e-01 9.62568581e-01 2.71754265e-01
-2.28227019e-01 1.30820215e-01 2.17174679e-01 4.39318120e-01
3.89648587e-01 9.13973987e-01 -8.13985586e-01 -1.50467893e-02
-6.47263527e-01 1.87625453e-01 9.12182510e-01 8.92060220e-01
1.21049833e+00 -2.20786338e-03 2.51052052e-01 6.75988019e-01
1.48056075e-01 2.80647695e-01 5.04942715e-01 -2.78690815e-01
4.82392937e-01 1.05956185e+00 -5.34805059e-01 -8.35138261e-01
-5.50109804e-01 -5.10196507e-01 -9.21659172e-01 1.18326060e-01
1.68911308e-01 1.80494472e-01 -1.27118790e+00 1.62782943e+00
1.71204403e-01 5.66373356e-02 1.26532719e-01 5.26046157e-01
1.07643223e+00 8.23121011e-01 -1.42643332e-01 -2.65684128e-02
8.28337193e-01 -1.05791652e+00 -1.34253934e-01 7.98568502e-02
9.20766056e-01 -2.15208590e-01 1.03549886e+00 4.10275042e-01
-8.11679840e-01 -4.11178082e-01 -1.55408943e+00 -1.57665052e-02
-8.31929266e-01 -2.55932480e-01 1.28127682e+00 4.62258965e-01
-9.20672178e-01 1.29883301e+00 -6.58853531e-01 -1.58324406e-01
4.31877136e-01 8.41809332e-01 -6.15265191e-01 -2.09528536e-01
-1.29544902e+00 7.84535110e-01 8.24702799e-01 -9.23064798e-02
-6.86215639e-01 -8.37330878e-01 -8.88089240e-01 -1.96297973e-01
6.00792289e-01 -5.81344128e-01 6.36184216e-01 -3.58280301e-01
-1.56211317e+00 5.11974394e-01 2.83994675e-01 -2.19624758e-01
8.17564800e-02 2.99041182e-01 -4.11136717e-01 -4.03638482e-02
-6.17465712e-02 3.59300226e-01 5.34712851e-01 -1.10141289e+00
-1.00776151e-01 -1.90479562e-01 -5.78219481e-02 -2.03318581e-01
-3.64542395e-01 -5.70569456e-01 -2.10187778e-01 -4.78821814e-01
1.79569945e-01 -6.84832633e-01 -2.65252411e-01 -5.22867441e-01
-6.71571255e-01 -3.38485301e-01 6.58745170e-01 -4.05056983e-01
1.03561020e+00 -1.49474335e+00 2.70940721e-01 6.78168058e-01
4.70953822e-01 9.45420414e-02 -3.30862552e-01 6.57356381e-01
-3.64167839e-01 1.00218713e-01 -4.43220437e-01 3.27051073e-01
-1.02831706e-01 2.07942277e-01 2.94531286e-01 2.69212067e-01
2.11793080e-01 1.19943428e+00 -8.62052858e-01 -2.73991674e-01
1.75525635e-01 3.67612123e-01 -6.74867809e-01 8.44872743e-02
-7.34042108e-01 5.48624635e-01 -5.98495305e-01 9.16697204e-01
7.24103034e-01 -7.16540933e-01 3.61019462e-01 -2.99398243e-01
1.50413588e-01 4.89638567e-01 -9.17119384e-01 1.77224410e+00
-2.05736578e-01 1.93422914e-01 -4.09440905e-01 -1.26806021e+00
1.09489727e+00 5.00597525e-03 7.13035285e-01 -9.59798038e-01
1.23785064e-02 3.77190441e-01 2.86098629e-01 -5.29866278e-01
4.73188221e-01 -3.78943123e-02 6.19763173e-02 5.10700047e-01
2.37155944e-01 -1.13035947e-01 3.94957691e-01 1.05870545e-01
1.37570250e+00 3.94311577e-01 9.44410041e-02 -2.99073219e-01
6.47938490e-01 6.04218058e-02 3.22597921e-01 5.05346417e-01
4.55776602e-01 3.45209450e-01 3.29762906e-01 -6.04441285e-01
-1.53202879e+00 -7.88811982e-01 -2.55367547e-01 8.32855344e-01
-1.09977491e-01 -7.66054869e-01 -4.76454288e-01 -5.80325782e-01
8.49992484e-02 1.63433507e-01 -7.31439292e-01 -3.05980891e-01
-7.68542588e-01 -1.05402446e+00 1.79284647e-01 6.44492626e-01
4.06722248e-01 -1.27266693e+00 1.03723019e-01 2.79378444e-01
4.81091589e-01 -9.37804699e-01 4.55982573e-02 6.98036969e-01
-7.59686291e-01 -1.24479461e+00 -3.43460500e-01 -8.99039149e-01
7.78414905e-01 -3.42261106e-01 1.31955540e+00 4.27975178e-01
-3.02204609e-01 -2.88320214e-01 -5.32096207e-01 -6.12094179e-02
-5.98486006e-01 7.46686101e-01 6.48548976e-02 -5.09082139e-01
1.84931293e-01 -1.09057474e+00 -4.54894811e-01 -1.24721549e-01
-6.94038212e-01 2.80213267e-01 7.13340521e-01 8.73741388e-01
5.79652250e-01 3.63344371e-01 4.94660497e-01 -1.18903923e+00
5.21777928e-01 -3.34767401e-01 -5.11658311e-01 2.74189770e-01
-1.17209029e+00 7.94722319e-01 8.46357524e-01 -2.68693030e-01
-5.49182713e-01 9.67673659e-02 -1.50282353e-01 6.28398582e-02
1.59941688e-01 9.70285773e-01 -4.35567081e-01 -4.78521496e-01
5.75576782e-01 1.38794586e-01 -1.24251679e-01 -5.44920743e-01
3.81575197e-01 5.76246858e-01 5.62293768e-01 -1.03142750e+00
1.01680112e+00 -9.28201079e-02 4.88199413e-01 -4.07959074e-01
-6.89658344e-01 -1.69557437e-01 -9.32087660e-01 9.24310014e-02
4.65852231e-01 -6.83654904e-01 -8.64272773e-01 3.98971558e-01
-6.32830620e-01 -4.46613431e-01 -7.10018277e-02 3.55125010e-01
-5.80238581e-01 4.02528286e-01 -7.20271468e-01 -2.91874200e-01
-5.70750713e-01 -1.10975635e+00 6.96203768e-01 6.73011765e-02
1.22059122e-01 -1.09076941e+00 2.09392115e-01 2.65518099e-01
4.75574762e-01 5.86544394e-01 1.55437303e+00 -7.36178160e-01
-9.01656806e-01 -1.74589530e-01 -2.64177501e-01 2.41409630e-01
4.32919174e-01 4.27732728e-02 -7.19323695e-01 -3.08935344e-01
-3.34347308e-01 -6.55043602e-01 1.16561544e+00 2.14473352e-01
1.32380688e+00 -7.42680877e-02 -4.13746119e-01 7.71987677e-01
1.66076982e+00 -4.14529145e-02 6.83246434e-01 6.24554157e-01
1.20417690e+00 2.62294829e-01 3.43695357e-02 6.67692125e-02
3.15109134e-01 4.62951958e-01 5.71755469e-01 -1.92303192e-02
1.17327861e-01 -3.61429363e-01 2.67244041e-01 1.15017700e+00
-7.23582566e-01 -2.06897154e-01 -9.76593733e-01 9.91851091e-02
-1.65345073e+00 -7.25306153e-01 -1.04004897e-01 2.02190304e+00
1.05568981e+00 3.78398269e-01 -1.27970632e-02 1.59883305e-01
4.37294751e-01 2.69587934e-01 -8.42117250e-01 -1.68262199e-01
-2.87484527e-01 7.84145355e-01 8.66692245e-01 2.26380348e-01
-8.18275392e-01 1.01227987e+00 6.48531675e+00 9.33437705e-01
-1.32688689e+00 -4.24447924e-01 2.82929212e-01 1.83027223e-01
-5.48448682e-01 1.16861783e-01 -6.24981582e-01 4.23088789e-01
1.26646650e+00 -7.33188242e-02 5.86739659e-01 7.95593679e-01
-1.35538816e-01 8.30094740e-02 -1.34480476e+00 8.33109140e-01
-1.69360667e-01 -1.87819302e+00 7.78499246e-02 4.00177807e-01
7.16268957e-01 2.37357482e-01 -2.26480603e-01 4.07685250e-01
5.62946856e-01 -1.34675038e+00 1.56730369e-01 5.74045420e-01
8.62352788e-01 -8.42354953e-01 7.04719901e-01 1.25907779e-01
-1.27526402e+00 2.17686787e-01 -6.13047659e-01 -1.11479573e-01
-7.68589228e-02 7.04476476e-01 -9.31386352e-01 1.22831368e+00
5.18692613e-01 1.18649793e+00 -7.28295326e-01 7.26960123e-01
-2.32254490e-01 5.99357426e-01 -4.30418164e-01 -2.75213540e-01
1.51070595e-01 -4.55254048e-01 1.65871516e-01 5.15852749e-01
2.05780104e-01 -1.13108404e-01 1.13594472e-01 9.79503632e-01
-5.06634533e-01 2.92163771e-02 -6.27697647e-01 -4.59139735e-01
1.46985665e-01 1.24502790e+00 -5.98144531e-01 -3.24172050e-01
-3.66098762e-01 5.41394413e-01 6.20466948e-01 1.24632522e-01
-6.78500652e-01 -1.51711717e-01 2.58170336e-01 2.61173338e-01
4.27339464e-01 -1.45813793e-01 -7.92965069e-02 -1.06816065e+00
7.91066736e-02 -7.02280402e-01 3.52834985e-02 -7.97606349e-01
-1.48295701e+00 7.27213085e-01 -1.64185762e-01 -9.03963387e-01
-2.78800368e-01 -1.21550369e+00 -4.78134096e-01 6.73042655e-01
-1.64032173e+00 -1.29506516e+00 -2.38288507e-01 2.11128935e-01
-1.97587386e-01 -3.88711780e-01 7.43784428e-01 2.17455000e-01
-6.34859979e-01 4.62037951e-01 3.86588305e-01 9.96811315e-02
4.49021041e-01 -1.38822401e+00 4.15764093e-01 3.15152556e-01
2.82753073e-02 4.26226676e-01 6.19843304e-01 -8.77545893e-01
-1.73329735e+00 -1.07022607e+00 4.17169273e-01 -4.58721429e-01
9.39876914e-01 -6.47456288e-01 -1.06944311e+00 5.24094641e-01
-7.44553208e-02 1.19176447e-01 5.92814744e-01 2.44302869e-01
-3.31502467e-01 1.09824210e-01 -9.51073050e-01 3.05173039e-01
1.39559829e+00 -4.45682436e-01 -3.86586368e-01 5.00046909e-01
1.01816118e+00 -3.77953380e-01 -1.27825761e+00 8.11334193e-01
3.66655022e-01 -8.00347507e-01 9.57267284e-01 -7.33750105e-01
7.75540054e-01 -3.39211598e-02 -1.73029408e-01 -1.20856559e+00
-4.68206912e-01 -1.87620282e-01 -1.65184855e-01 1.08152473e+00
9.24406469e-01 -7.11848259e-01 1.38231277e+00 5.88468969e-01
-2.18190029e-01 -1.29263186e+00 -5.57894289e-01 -8.54289353e-01
2.83176303e-01 -3.03912133e-01 1.07492292e+00 9.69572544e-01
3.73984016e-02 6.32445872e-01 -4.44253981e-01 -1.03806391e-01
4.14087474e-01 3.24677587e-01 6.46781266e-01 -1.67260551e+00
-3.31511497e-01 -1.09189622e-01 -6.25352442e-01 -7.18690395e-01
2.95716166e-01 -1.51281786e+00 -3.96262199e-01 -1.61438048e+00
3.99073631e-01 -7.06552684e-01 -4.54676807e-01 6.27021551e-01
6.16398603e-02 1.00717224e-01 -2.45235071e-01 2.13505656e-01
-3.49166781e-01 7.65319765e-01 1.62696612e+00 -3.53518248e-01
-2.17162386e-01 -4.21289951e-01 -6.12433612e-01 2.30675101e-01
7.86926627e-01 -4.09971982e-01 -3.83397192e-01 1.60815656e-01
6.78614438e-01 -1.62294656e-01 1.96565956e-01 -1.22956741e+00
9.10649374e-02 -9.19061378e-02 5.36730587e-01 -6.71594679e-01
2.39359692e-01 -7.21843719e-01 2.32923791e-01 3.76363695e-01
-1.18667625e-01 -1.91296741e-01 -8.62543061e-02 5.99705756e-01
-1.14278816e-01 -2.11626172e-01 4.52981114e-01 -3.32084090e-01
-6.76194668e-01 7.77756214e-01 1.72411919e-01 -3.52844149e-01
6.98846400e-01 -5.01152575e-01 -4.39041227e-01 2.94367187e-02
-7.55104661e-01 1.51145861e-01 8.90857518e-01 3.40256095e-02
3.95806372e-01 -1.28132820e+00 -4.36103761e-01 4.75115657e-01
2.14949951e-01 3.71951282e-01 -1.89420469e-02 3.78283441e-01
-6.20275140e-01 4.75535780e-01 -4.89407331e-01 -4.11279351e-01
-8.22724044e-01 9.69459355e-01 9.15456936e-02 -6.86055183e-01
-7.44670331e-01 6.03545845e-01 -2.87467033e-01 -6.36196911e-01
-3.72790664e-01 -5.63219905e-01 -1.11373976e-01 -3.22062403e-01
-1.00634716e-01 9.00169276e-03 3.86152774e-01 -3.45707536e-01
-1.63834527e-01 7.01857507e-01 -3.78837228e-01 5.35236835e-01
2.10109711e+00 3.58620346e-01 -4.05537218e-01 1.59363806e-01
1.31783855e+00 1.05935246e-01 -9.35322940e-01 -3.03366572e-01
6.49464056e-02 7.15816915e-02 -7.19165280e-02 -7.36461401e-01
-1.35098469e+00 4.19019997e-01 3.22326899e-01 8.66302624e-02
1.00150704e+00 3.25702637e-01 8.16953242e-01 7.55333960e-01
6.37706459e-01 -1.10957038e+00 5.99873066e-02 5.05725265e-01
5.78635693e-01 -1.31236398e+00 3.51356119e-01 -7.35466838e-01
-4.41478975e-02 1.27815461e+00 7.64629006e-01 -2.65716702e-01
7.26870894e-01 2.24935472e-01 -3.26874316e-01 -7.03808308e-01
-8.09220850e-01 -5.35780676e-02 2.38569543e-01 7.16353953e-01
4.87583905e-01 4.81981896e-02 -3.16325552e-03 3.63695502e-01
-8.19788873e-01 -9.63025540e-02 4.39731717e-01 9.23775375e-01
-4.15388614e-01 -1.75102484e+00 -1.01027582e-02 9.71536577e-01
5.77656068e-02 -3.16046596e-01 -4.45860505e-01 8.98848891e-01
3.14974715e-03 3.12413901e-01 -2.60309726e-01 -8.27239931e-01
6.52030408e-02 1.05085388e-01 7.04751730e-01 -6.43742383e-01
-2.23467797e-01 -4.61423308e-01 1.64444596e-01 -4.09530193e-01
-5.70255935e-01 -1.08264670e-01 -1.46396792e+00 -5.00982106e-01
-4.95000005e-01 2.35209554e-01 3.94638658e-01 9.06732798e-01
2.13560164e-01 7.15928257e-01 5.36981165e-01 -9.39545274e-01
-1.18629850e-01 -1.06334817e+00 -8.30358088e-01 3.90768975e-01
2.25585282e-01 -8.23682010e-01 -1.19053721e-01 -3.21310729e-01] | [5.22313117980957, 5.533278465270996] |
6dc43e81-201a-4b86-aee5-6608edf25deb | progressive-one-shot-human-parsing | 2012.11810 | null | https://arxiv.org/abs/2012.11810v3 | https://arxiv.org/pdf/2012.11810v3.pdf | Progressive One-shot Human Parsing | Prior human parsing models are limited to parsing humans into classes pre-defined in the training data, which is not flexible to generalize to unseen classes, e.g., new clothing in fashion analysis. In this paper, we propose a new problem named one-shot human parsing (OSHP) that requires to parse human into an open set of reference classes defined by any single reference example. During training, only base classes defined in the training set are exposed, which can overlap with part of reference classes. In this paper, we devise a novel Progressive One-shot Parsing network (POPNet) to address two critical challenges , i.e., testing bias and small sizes. POPNet consists of two collaborative metric learning modules named Attention Guidance Module and Nearest Centroid Module, which can learn representative prototypes for base classes and quickly transfer the ability to unseen classes during testing, thereby reducing testing bias. Moreover, POPNet adopts a progressive human parsing framework that can incorporate the learned knowledge of parent classes at the coarse granularity to help recognize the descendant classes at the fine granularity, thereby handling the small sizes issue. Experiments on the ATR-OS benchmark tailored for OSHP demonstrate POPNet outperforms other representative one-shot segmentation models by large margins and establishes a strong baseline. Source code can be found at https://github.com/Charleshhy/One-shot-Human-Parsing. | ['DaCheng Tao', 'Bhavani Thuraisingham', 'Jing Zhang', 'Haoyu He'] | 2020-12-22 | null | null | null | null | ['one-shot-segmentation', 'human-parsing'] | ['computer-vision', 'computer-vision'] | [ 2.90077925e-01 3.83617848e-01 -2.13763133e-01 -6.43265605e-01
-9.22134638e-01 -4.72722054e-01 -6.87911827e-03 2.21276488e-02
-4.46374208e-01 3.74250501e-01 -8.24697316e-02 1.36397839e-01
1.69343278e-01 -1.04681230e+00 -7.20705450e-01 -5.38161159e-01
3.33322376e-01 6.55233681e-01 6.68150604e-01 -1.76888496e-01
-1.90539330e-01 8.30899552e-03 -1.42349792e+00 1.88404396e-01
9.99389946e-01 8.60905111e-01 3.87943447e-01 5.46557009e-01
-1.66002303e-01 3.97066116e-01 -6.37599051e-01 -6.38035178e-01
1.60596237e-01 -3.99896562e-01 -7.36361027e-01 3.96767184e-02
3.23078960e-01 -4.68279511e-01 -1.28588885e-01 1.17471051e+00
6.50150657e-01 4.15785521e-01 2.65756100e-01 -1.05323708e+00
-9.68846738e-01 6.62528038e-01 -6.00126207e-01 1.93113670e-01
3.01948607e-01 2.41549477e-01 1.08084679e+00 -7.14968383e-01
5.89931726e-01 1.21865833e+00 7.06748128e-01 1.04202080e+00
-8.85150552e-01 -7.13957548e-01 5.53255856e-01 2.69053895e-02
-1.02478194e+00 -1.45829156e-01 7.52106488e-01 -3.36324841e-01
5.13729692e-01 2.31951535e-01 3.79707754e-01 1.10648942e+00
-2.06241980e-01 9.93822455e-01 7.41293252e-01 -2.37915307e-01
3.26410115e-01 -3.05060204e-02 9.16561961e-01 7.04667449e-01
2.28524923e-01 1.22032631e-02 -1.50775209e-01 1.15620412e-01
5.43162286e-01 1.55505568e-01 -5.19304127e-02 -3.79292220e-01
-8.72594714e-01 9.49272692e-01 7.38592446e-01 1.49802223e-01
-3.77332643e-02 -2.12385178e-01 3.75726610e-01 -9.94049832e-02
2.89845616e-01 1.31491661e-01 -6.43427372e-01 1.01053439e-01
-6.46868527e-01 9.31579396e-02 6.47781312e-01 1.42749596e+00
7.79756010e-01 -3.85479689e-01 -4.32685256e-01 1.03797889e+00
1.22642450e-01 3.72775257e-01 5.54704070e-01 -8.76605332e-01
7.47988582e-01 7.56455719e-01 -1.75056666e-01 -5.94477594e-01
-4.63147551e-01 -3.48436922e-01 -7.01784372e-01 -1.60780221e-01
4.69350576e-01 -3.87405664e-01 -1.32877922e+00 1.88833857e+00
8.07968795e-01 7.45825395e-02 -2.71311942e-02 1.01938450e+00
1.24851406e+00 6.43141985e-01 3.87835860e-01 5.64294010e-02
1.66618967e+00 -1.28776813e+00 -3.11740488e-01 -5.55150449e-01
5.34672618e-01 -3.46216559e-01 1.34532297e+00 2.01551318e-01
-7.55457163e-01 -9.82106924e-01 -1.08390427e+00 -3.15109193e-01
-4.73492146e-01 8.40410497e-03 6.31438851e-01 7.62397170e-01
-5.35737336e-01 5.63297391e-01 -8.89097154e-01 -4.67076480e-01
7.54285038e-01 3.31437349e-01 -8.13460574e-02 -3.76538277e-01
-1.17680478e+00 2.29642540e-01 5.96360087e-01 5.98393977e-02
-7.83934534e-01 -5.16058922e-01 -1.12473309e+00 6.28591999e-02
7.63341665e-01 -5.49961388e-01 1.39359868e+00 -8.60573113e-01
-1.32882583e+00 9.03973520e-01 5.13543412e-02 -8.37925524e-02
5.90504527e-01 -3.11536103e-01 -4.36033875e-01 -1.05717918e-02
4.78173912e-01 8.15607667e-01 6.64333045e-01 -1.12238407e+00
-6.83034360e-01 -4.86941159e-01 3.49594653e-01 1.30170628e-01
-1.81406125e-01 -9.23641492e-03 -8.88954937e-01 -6.34619296e-01
1.50991678e-01 -9.72464979e-01 -3.95377189e-01 -3.12352180e-01
-6.71138763e-01 -3.54419291e-01 4.62213755e-01 -6.66551411e-01
1.19164777e+00 -2.25852394e+00 1.03476554e-01 -1.62042320e-01
2.04963848e-01 2.79477209e-01 -1.86844036e-01 -1.46633983e-01
6.75974414e-03 2.43781433e-02 -5.19521713e-01 -1.31420061e-01
1.43841535e-01 3.52231383e-01 -1.16818072e-02 5.01065515e-02
3.35646868e-01 9.63782310e-01 -1.05093408e+00 -6.20473385e-01
-8.09513628e-02 2.08564833e-01 -5.44544041e-01 3.19447160e-01
-2.00568691e-01 4.31139290e-01 -6.37352943e-01 9.71069574e-01
7.04216063e-01 -3.06578368e-01 9.52845253e-03 -8.53002816e-02
2.67340958e-01 -2.19328120e-01 -1.18395329e+00 1.85607696e+00
-1.89509273e-01 -4.01247703e-02 -1.71757162e-01 -7.06542969e-01
7.99417078e-01 -1.03502974e-01 6.83620051e-02 -6.99707210e-01
3.73444438e-01 -5.96039593e-02 5.31969331e-02 -3.62667710e-01
3.15890521e-01 -9.78239812e-03 -5.45602500e-01 1.82073891e-01
2.50691205e-01 4.17108059e-01 4.61768448e-01 1.17601797e-01
1.02651632e+00 2.50868440e-01 2.14846686e-01 -5.42867295e-02
3.60450864e-01 9.20170248e-02 1.30320632e+00 8.21419835e-01
-6.28448188e-01 9.63646173e-01 5.15636742e-01 -4.38954979e-01
-7.96877325e-01 -1.27728105e+00 -9.23843011e-02 1.66420424e+00
6.45020366e-01 -3.26580852e-01 -1.14894331e+00 -1.11914515e+00
-2.05097169e-01 6.34997964e-01 -8.02651525e-01 -1.70151323e-01
-7.25432694e-01 -9.20150220e-01 1.97914630e-01 9.35358822e-01
7.01814353e-01 -1.19911468e+00 -7.35841215e-01 1.80111185e-01
-5.07462583e-02 -1.11698377e+00 -6.53078020e-01 1.37357652e-01
-7.39232838e-01 -1.32495308e+00 -8.94106030e-01 -1.20346344e+00
7.21125901e-01 6.26064911e-02 1.09539688e+00 8.64207968e-02
-4.90424573e-01 2.67050356e-01 -6.16714716e-01 -3.74238014e-01
-1.33033961e-01 1.50200367e-01 -3.23716700e-01 -8.49120021e-02
7.58186102e-01 -1.20858043e-01 -6.77836597e-01 5.71273565e-01
-6.76932871e-01 7.05389902e-02 5.91796875e-01 7.72235274e-01
9.67767000e-01 -2.14868579e-02 5.22612751e-01 -1.48145497e+00
2.93861717e-01 -6.40889406e-01 -3.69587868e-01 4.43060786e-01
-2.34905377e-01 -3.37344438e-01 6.32424474e-01 -6.01993024e-01
-1.08568287e+00 2.34146684e-01 -1.39241442e-01 -2.18752846e-01
-3.96554470e-01 6.23938739e-02 -8.08136404e-01 3.17327708e-01
4.41254348e-01 -6.80154637e-02 -4.22308266e-01 -6.68805540e-01
4.45205718e-01 5.22704422e-01 6.74413919e-01 -7.95072913e-01
6.82942748e-01 1.04261868e-01 -5.25732636e-01 -4.55254614e-01
-1.10495269e+00 -4.86907631e-01 -8.67260635e-01 8.54833350e-02
1.32283258e+00 -9.16269243e-01 -1.40741885e-01 4.33653921e-01
-9.16259766e-01 -5.06712794e-01 -5.14724195e-01 1.74911141e-01
-3.83991361e-01 2.66114712e-01 -9.47693944e-01 -3.84918004e-01
-5.37435889e-01 -1.08497787e+00 1.19785368e+00 7.34479964e-01
-1.13008365e-01 -5.49707890e-01 -9.57596973e-02 5.34015954e-01
-1.11523092e-01 3.60547841e-01 8.05661380e-01 -1.05708206e+00
-4.75427777e-01 -2.60039479e-01 -1.47483051e-01 3.31931263e-01
-1.21019311e-01 -2.54867703e-01 -1.02577782e+00 -3.19216311e-01
-9.61677805e-02 -3.12515378e-01 8.98510039e-01 2.32665777e-01
1.29646027e+00 1.08122155e-01 -3.07107955e-01 7.96460032e-01
1.15078521e+00 3.09555650e-01 5.58244586e-01 2.35507101e-01
9.14102495e-01 5.19861221e-01 1.01257455e+00 1.54250666e-01
4.13833231e-01 3.53789955e-01 9.41204131e-02 -1.02457799e-01
-1.44799516e-01 -3.20902914e-01 4.96320501e-02 7.13450432e-01
1.32752359e-01 -1.94074333e-01 -8.76920521e-01 4.28525686e-01
-1.77119696e+00 -5.97995102e-01 1.53325558e-01 2.00727248e+00
6.52154326e-01 5.48358142e-01 1.90407351e-01 -1.47666588e-01
1.29084218e+00 9.74633694e-02 -8.81708801e-01 -2.48229250e-01
2.34508038e-01 3.64444315e-01 2.34633222e-01 1.77646264e-01
-1.43623435e+00 1.04710197e+00 4.41344595e+00 7.21827984e-01
-4.97461021e-01 4.45808738e-01 7.76101887e-01 -7.33718872e-02
1.41902521e-01 -1.39629543e-02 -1.16468608e+00 6.79253697e-01
6.91182613e-01 2.84792203e-02 5.52689470e-02 1.11811829e+00
-3.77064109e-01 1.63554832e-01 -1.18719113e+00 8.91821444e-01
3.81499343e-02 -6.93510711e-01 -5.94478995e-02 -2.51744241e-01
5.14738321e-01 -7.38360360e-02 -2.49028131e-01 9.48695779e-01
2.58410633e-01 -6.68216467e-01 5.33938468e-01 1.95204258e-01
8.73695672e-01 -7.53593981e-01 7.13221312e-01 3.80583704e-01
-1.37640321e+00 -3.31805676e-01 -7.44238973e-01 2.64938205e-01
1.95430204e-01 3.59215647e-01 -2.33845890e-01 5.21148264e-01
8.97654116e-01 4.84279215e-01 -7.44016290e-01 9.37611103e-01
-4.34280545e-01 5.06224036e-01 -8.93840641e-02 1.38002094e-02
1.70561939e-01 -3.05133536e-02 3.03362578e-01 1.04328358e+00
2.26888791e-01 5.13388753e-01 5.74829638e-01 7.94072092e-01
-1.44530937e-01 1.05596602e-01 -8.07956532e-02 1.67644843e-01
4.82334465e-01 1.38411880e+00 -1.04182792e+00 -5.08279800e-01
-6.40028417e-01 1.12840736e+00 4.20541346e-01 2.85348833e-01
-1.09688210e+00 -8.12873721e-01 4.94448930e-01 3.97028774e-02
5.33779740e-01 2.14434430e-01 -1.68518871e-01 -1.15720069e+00
1.17558107e-01 -7.65920043e-01 9.36615646e-01 -5.17951429e-01
-1.51152968e+00 7.65071869e-01 -5.86681291e-02 -1.16829693e+00
1.92302078e-01 -6.60015583e-01 -8.34541023e-01 7.25391209e-01
-1.29965365e+00 -1.33169520e+00 -6.02252722e-01 4.41750258e-01
9.17792857e-01 1.38551081e-02 6.31788135e-01 3.22027564e-01
-1.01905179e+00 8.48684430e-01 -2.60655671e-01 6.90014005e-01
6.18113101e-01 -1.38472223e+00 7.15355814e-01 9.12264526e-01
-7.67892376e-02 5.70475101e-01 3.23066384e-01 -8.55796516e-01
-9.43164468e-01 -1.35813367e+00 5.40203035e-01 -5.57829857e-01
3.90138924e-01 -5.83507061e-01 -1.10108221e+00 8.21505606e-01
-2.85910964e-01 3.22760582e-01 8.38127732e-01 2.13937327e-01
-4.05184895e-01 -1.53647408e-01 -1.13841188e+00 3.62958997e-01
1.29875934e+00 -2.45355636e-01 -8.37954640e-01 3.25726092e-01
1.07698226e+00 -6.55740499e-01 -8.75230193e-01 5.30240357e-01
3.61914009e-01 -7.57991552e-01 8.09042037e-01 -6.78177893e-01
4.33444291e-01 -1.67436361e-01 -9.05599073e-02 -9.83558238e-01
-6.25889122e-01 -3.48450929e-01 -1.72193825e-01 1.59597123e+00
3.78545880e-01 -5.10531485e-01 1.02794576e+00 9.48437989e-01
-3.43994051e-01 -7.63393700e-01 -8.27838182e-01 -7.34256506e-01
2.05610886e-01 -2.73831636e-01 8.89596403e-01 7.40843952e-01
-3.33137125e-01 5.21469116e-01 -7.05144107e-02 4.17745769e-01
7.39691317e-01 3.37862790e-01 7.79523015e-01 -1.29290497e+00
-6.06797457e-01 -1.30436629e-01 -4.58755434e-01 -1.03394651e+00
2.96463501e-02 -9.63544190e-01 2.37135768e-01 -1.47647059e+00
5.01441717e-01 -5.24666965e-01 -4.73300457e-01 5.03576994e-01
-7.07169771e-01 2.90835574e-02 2.50555098e-01 -2.58465880e-03
-1.00465024e+00 3.49719107e-01 1.31793404e+00 -1.62203282e-01
-2.73129135e-01 1.31534651e-01 -9.79862809e-01 7.74848163e-01
8.35470438e-01 -5.79314947e-01 -4.27505791e-01 -5.87265432e-01
-2.20371515e-01 -1.06250599e-01 2.30194956e-01 -1.20356417e+00
2.16023520e-01 8.08614641e-02 6.66089535e-01 -4.62988585e-01
3.13850306e-02 -5.96034169e-01 -1.10450581e-01 3.22639376e-01
-3.04672778e-01 -1.44313172e-01 -9.78885405e-03 8.43282402e-01
1.24875471e-01 -4.62802827e-01 8.44738662e-01 -3.90763104e-01
-1.31695819e+00 6.51886225e-01 5.30813992e-01 6.13819182e-01
1.19453490e+00 -3.65931839e-01 -4.92708176e-01 4.06528324e-01
-1.13877606e+00 7.00391233e-01 4.13608491e-01 6.17763042e-01
4.09449279e-01 -1.18836284e+00 -5.04393637e-01 2.11258665e-01
3.25525314e-01 4.32650059e-01 5.76049805e-01 4.22831088e-01
-3.83410007e-01 -8.46689492e-02 -3.78113128e-02 -5.83735168e-01
-1.15538669e+00 8.87911022e-01 2.30623141e-01 -7.30007589e-02
-8.52130830e-01 1.19722605e+00 5.81992090e-01 -8.03876460e-01
3.55056882e-01 -3.43469173e-01 -3.12909037e-01 8.50011408e-03
6.73312545e-01 3.16878408e-01 -1.83243513e-01 -5.81324399e-01
-2.45340511e-01 7.54227817e-01 -2.22716823e-01 5.10170460e-01
1.24244285e+00 -9.43877026e-02 2.63792485e-01 4.91596371e-01
1.10857999e+00 -2.23094776e-01 -1.52060497e+00 -3.35446149e-01
2.53683746e-01 -3.32575440e-01 -3.58139515e-01 -8.63439023e-01
-1.23008990e+00 8.89966965e-01 8.53142619e-01 -2.53997128e-02
1.11911500e+00 2.47798681e-01 1.30601585e+00 1.79077357e-01
3.60295296e-01 -1.32315648e+00 -7.75927678e-02 3.48947942e-01
3.39441150e-01 -1.48005891e+00 -3.11425209e-01 -7.69647598e-01
-7.78853834e-01 9.35250282e-01 1.13519287e+00 -7.94613957e-02
5.20226061e-01 4.17445181e-03 -3.25224772e-02 -1.67178348e-01
-3.50420862e-01 -4.16709214e-01 2.49839589e-01 7.01535821e-01
2.97234386e-01 2.65233696e-01 -2.50984341e-01 1.53531969e+00
-9.05918702e-02 -1.88699067e-01 1.28232494e-01 1.05318427e+00
-6.07191682e-01 -1.02105737e+00 -1.65405571e-01 5.58333397e-01
-3.47983181e-01 2.97994673e-04 -2.63779879e-01 8.86521995e-01
5.84792852e-01 8.59937668e-01 1.71737641e-01 -3.35196376e-01
6.98463440e-01 9.59904864e-02 2.55728453e-01 -1.11839688e+00
-4.94863421e-01 -4.92808968e-02 9.02277082e-02 -6.00569606e-01
-4.49416116e-02 -5.32310843e-01 -1.43197632e+00 1.63631260e-01
-3.70637178e-01 1.34213686e-01 1.40596628e-01 7.27569759e-01
1.30319282e-01 7.58267879e-01 3.46989691e-01 -6.17347181e-01
-6.52220726e-01 -7.44798958e-01 -6.26968324e-01 6.03371859e-01
-7.01487139e-02 -5.92628360e-01 -1.21989720e-01 -8.93196687e-02] | [9.023524284362793, 0.39980393648147583] |
8e00996b-d9cd-4b13-9728-26c66370e988 | glasses-detection-using-convolutional-neural | null | null | https://doi.org/10.1007/978-3-319-46654-5_78 | https://link.springer.com/chapter/10.1007/978-3-319-46654-5_78#chapter-info | Glasses Detection Using Convolutional Neural Networks | Glasses detection plays an important role in face recognition and soft biometrices for person identification. However, automatic glasses detection is still a challenging problem under real application scenarios, because face variations, light conditions, and self-occlusion, have significant influence on its performance. Inspired by the success of Deep Convolutional Neural Networks (DCNN) on face recognition, object detection and image classification, we propose a glasses detection method based on DCNN. Specifically, we devise a Glasses Network (GNet), and pre-train it as a face identification network with a large number of face images. The pre-trained GNet is finally fine-tuned as a glasses detection network by using another set of facial images wearing and not wearing glasses. Evaluation experiments have been done on two public databases, Multi-PIE and LFW. The results demonstrate the superior performance of the proposed method over competing methods. | ['Qijun Zhao', 'Ronghang Zhu', 'Li Shao'] | 2016-09-21 | null | null | null | conference-paper-2016-9 | ['person-identification', 'face-identification'] | ['computer-vision', 'computer-vision'] | [ 1.86332747e-01 -2.14078650e-01 1.48634404e-01 -5.70394993e-01
-1.73504110e-02 -2.89043456e-01 3.34542841e-01 -7.99612820e-01
-8.28667358e-02 4.95256841e-01 4.78455611e-02 2.78695077e-02
2.29848668e-01 -7.28387892e-01 -6.04132771e-01 -9.15078223e-01
7.75997192e-02 2.03455463e-02 1.02277905e-01 -5.11202291e-02
-1.42965151e-03 5.70717692e-01 -1.87497771e+00 5.73217310e-02
7.22246170e-01 1.39646900e+00 -1.60832942e-01 1.24412462e-01
4.03603196e-01 1.55341506e-01 -4.93410230e-01 -7.72785664e-01
3.90264452e-01 -2.00276986e-01 -1.75574422e-01 6.33626878e-01
6.42246127e-01 -5.11489034e-01 -3.52090031e-01 1.27925777e+00
7.21986949e-01 -1.28418639e-01 6.34812593e-01 -1.03706586e+00
-7.07534254e-01 2.14541450e-01 -4.18290138e-01 2.59327348e-02
2.80619502e-01 5.05808413e-01 3.38294417e-01 -9.72256243e-01
2.61652112e-01 1.38233781e+00 7.88275063e-01 9.14301991e-01
-7.59633422e-01 -9.90523100e-01 -1.11621253e-01 2.22716659e-01
-1.57840168e+00 -9.87066388e-01 4.82585311e-01 -4.25323009e-01
3.70425016e-01 2.11264983e-01 5.89585066e-01 9.40374732e-01
4.59772758e-02 6.24289215e-01 1.11564517e+00 -4.10163373e-01
1.59804560e-02 4.02669758e-01 -4.34283875e-02 7.67882943e-01
4.67173338e-01 1.89858153e-01 -2.94554025e-01 1.44311920e-01
7.43151963e-01 1.67178214e-01 -4.70473439e-01 1.82112575e-01
-4.39918101e-01 5.04821718e-01 6.14634573e-01 3.16137075e-01
-3.99068505e-01 -2.04160646e-01 1.38358455e-02 1.90659449e-01
4.12871927e-01 -2.42815554e-01 1.36257201e-01 6.07991576e-01
-8.31254005e-01 -3.05057261e-02 5.66620827e-01 6.75540805e-01
5.78723371e-01 1.67821020e-01 -4.82854337e-01 8.96730721e-01
6.46869361e-01 5.86742640e-01 5.31410038e-01 4.56414372e-02
1.68384537e-01 8.91508102e-01 9.12531093e-02 -1.04280627e+00
-2.18201965e-01 -4.93064612e-01 -9.02232468e-01 2.17977390e-01
2.74489731e-01 -3.83471288e-02 -1.16771221e+00 1.42825234e+00
4.77923691e-01 5.71139753e-01 -2.59521436e-02 1.10154688e+00
1.15221477e+00 1.70534089e-01 7.85393044e-02 -2.38979012e-01
1.60541987e+00 -5.47613621e-01 -7.36832976e-01 -2.09052220e-01
-2.60100613e-04 -7.76325583e-01 9.26850379e-01 4.53020811e-01
-8.15473855e-01 -7.53422856e-01 -1.03720093e+00 2.95681745e-01
-2.96581954e-01 7.31311560e-01 4.49480683e-01 1.51929688e+00
-1.12844825e+00 1.91497669e-01 -3.53976399e-01 -5.48914254e-01
8.68916869e-01 1.06452370e+00 -2.38394916e-01 -1.76724315e-01
-9.74039972e-01 5.22135198e-01 7.04271495e-02 7.51547158e-01
-1.07100129e+00 -7.37482756e-02 -6.84060931e-01 2.02200972e-02
4.23066378e-01 -3.39171469e-01 8.75344276e-01 -1.10087514e+00
-1.56191361e+00 1.16906357e+00 -1.02759771e-01 -2.16191232e-01
4.30657685e-01 -1.22112453e-01 -8.41195762e-01 -1.51692808e-01
-2.56463349e-01 1.49318740e-01 1.19693458e+00 -9.80578661e-01
-1.93368077e-01 -7.06982017e-01 -1.78502742e-02 -5.07627539e-02
-7.35352099e-01 3.64389300e-01 -5.32332361e-01 -3.76416236e-01
1.25423566e-01 -1.01049578e+00 4.93355751e-01 -2.21025914e-01
-6.34570777e-01 -3.71067852e-01 9.59232390e-01 -6.42472446e-01
1.02911270e+00 -2.11442351e+00 -2.78962731e-01 3.25446397e-01
2.12065071e-01 8.51915359e-01 1.13683864e-01 -1.05222039e-01
1.20451944e-02 -2.48823613e-01 7.19777644e-02 -4.71056551e-01
-2.08539680e-01 -1.45329863e-01 1.97606161e-01 5.79270601e-01
1.77284643e-01 6.54073536e-01 -2.86631495e-01 -3.03530186e-01
1.94818899e-01 6.03909552e-01 -3.60389829e-01 4.90726292e-01
1.99166499e-02 3.30111802e-01 -3.70920330e-01 1.32462573e+00
1.06253374e+00 -1.21476816e-03 1.86440408e-01 -5.04568219e-01
1.06488794e-01 -3.11134517e-01 -1.27379036e+00 8.35882664e-01
-1.47502393e-01 3.14844102e-01 3.29039176e-03 -5.35268068e-01
1.02961874e+00 5.32229364e-01 -2.26429068e-02 -3.26233119e-01
8.51521671e-01 1.77127719e-01 3.91024649e-02 -9.87315297e-01
7.95859173e-02 -1.44043639e-01 5.33608317e-01 -7.43079633e-02
1.54957846e-01 3.92135352e-01 1.23672143e-01 -3.02447855e-01
6.86751783e-01 -3.74680907e-02 5.21884337e-02 -1.81698233e-01
9.02332366e-01 -6.83932543e-01 6.32277012e-01 3.04833233e-01
-4.11887079e-01 5.71003020e-01 1.84894755e-01 -5.66627324e-01
-5.23498118e-01 -6.75967634e-01 -3.25801164e-01 8.02288592e-01
1.86332110e-02 2.89086886e-02 -1.20120502e+00 -7.18218148e-01
1.97342515e-01 -1.27604336e-01 -8.07029963e-01 -2.90448576e-01
-2.77309030e-01 -1.18281376e+00 5.08331060e-01 3.36262256e-01
9.45134103e-01 -1.20411205e+00 -1.55682430e-01 -2.31691189e-02
3.20535719e-01 -1.27421153e+00 -6.92167222e-01 -4.64152247e-01
-3.03944051e-01 -1.28952682e+00 -7.55230308e-01 -9.13010240e-01
6.99711919e-01 1.98737547e-01 6.40380442e-01 4.64345187e-01
-4.69063580e-01 6.28237128e-02 1.22051910e-02 -4.85011846e-01
-2.06271589e-01 -9.88217667e-02 5.64635873e-01 1.10444880e+00
7.79781997e-01 -3.98235708e-01 -8.57651949e-01 5.48170269e-01
-6.65752947e-01 -6.20675921e-01 6.63871109e-01 6.23104811e-01
9.28674117e-02 2.36584973e-02 6.43171012e-01 -1.12328839e+00
5.38779020e-01 -1.62277803e-01 -8.67826045e-01 3.43021125e-01
-4.64864314e-01 -4.79414165e-01 1.34650305e-01 -3.47425610e-01
-1.10285509e+00 3.50412101e-01 -2.13564247e-01 -6.28437340e-01
-3.46621066e-01 1.20114937e-01 -8.54914486e-01 -6.52474761e-01
5.73163390e-01 6.35597706e-02 1.52348533e-01 -4.27290112e-01
-2.82650560e-01 1.16907930e+00 4.10658926e-01 8.92718062e-02
8.29886019e-01 3.26702237e-01 -1.64451763e-01 -9.59946036e-01
-7.77126551e-01 -3.09438258e-01 -7.02036083e-01 -5.26657939e-01
9.98556256e-01 -1.07394946e+00 -1.40378129e+00 1.03883195e+00
-7.38833487e-01 1.32475525e-01 5.96021831e-01 4.14509386e-01
1.95588738e-01 1.37299910e-01 -3.84266943e-01 -1.23935485e+00
-5.50738156e-01 -1.26036417e+00 1.03548336e+00 6.43370271e-01
2.81902313e-01 -6.81440353e-01 -3.15890163e-01 4.69008207e-01
5.84202409e-01 2.62828410e-01 2.65105426e-01 -3.84033501e-01
-7.62543321e-01 -4.95247871e-01 -2.97504246e-01 4.66643929e-01
5.83962858e-01 -3.73836942e-02 -1.67020679e+00 -4.14459914e-01
-2.10937625e-03 -2.78606385e-01 8.39163959e-01 3.52220416e-01
9.57391977e-01 -2.35566646e-01 -3.86078924e-01 7.69825637e-01
1.46082127e+00 2.61737406e-01 8.33606839e-01 1.10362202e-01
7.87060618e-01 6.17159784e-01 -1.02983918e-02 2.78902173e-01
1.85941979e-01 4.63060915e-01 6.33293688e-01 -7.28396922e-02
-3.48040432e-01 1.75602168e-01 2.59168506e-01 3.73306364e-01
-4.28604901e-01 -2.23519057e-01 -6.74145401e-01 2.47372448e-01
-1.50174975e+00 -7.51844108e-01 7.67876133e-02 2.23736262e+00
4.95263785e-01 1.53405026e-01 3.47405165e-01 2.97660172e-01
9.60462153e-01 -1.39866024e-01 -6.08127892e-01 1.86150536e-01
-2.52199143e-01 2.00050220e-01 2.93340504e-01 -8.61481111e-03
-1.34362674e+00 1.01439214e+00 5.80710554e+00 4.14925486e-01
-1.65902197e+00 1.89208046e-01 7.91329265e-01 -3.54772918e-02
3.74329001e-01 -5.70984066e-01 -1.16338181e+00 7.14448631e-01
6.26596689e-01 4.79458570e-01 4.04352725e-01 7.53310978e-01
1.96914315e-01 1.42138869e-01 -7.64911294e-01 1.20810163e+00
4.98955816e-01 -9.64800656e-01 -1.04179554e-01 6.13959059e-02
6.46594584e-01 -4.16951686e-01 4.19160306e-01 1.32829010e-01
-1.36477977e-01 -1.12629282e+00 3.86186272e-01 5.60472786e-01
9.80057120e-01 -4.81974572e-01 9.88385558e-01 -2.01202542e-01
-1.26281726e+00 -2.59658009e-01 -5.04403710e-01 2.74814963e-01
-3.35222960e-01 4.35413003e-01 -8.49557400e-01 2.64337212e-01
8.32552612e-01 5.74115157e-01 -7.15564311e-01 1.25737453e+00
-1.03073739e-01 6.62282228e-01 -1.54856622e-01 -8.38044211e-02
-1.65910423e-01 -1.77863836e-01 1.81665227e-01 8.85857940e-01
2.78842658e-01 2.63790488e-01 2.40175221e-02 7.73535907e-01
-5.37876487e-01 2.07537822e-02 -4.53868300e-01 7.67172426e-02
3.15489113e-01 1.45193589e+00 -5.78406513e-01 -2.48191059e-01
-5.07322967e-01 7.64035702e-01 -7.98824951e-02 2.79522091e-01
-5.21111965e-01 -2.32246101e-01 6.85274303e-01 4.30662096e-01
2.46809632e-01 3.29346359e-01 -7.81534612e-03 -1.10825348e+00
1.74192563e-01 -1.05575430e+00 2.91340142e-01 -5.36882043e-01
-1.21102166e+00 1.07135332e+00 -4.22836989e-01 -1.24430084e+00
2.18146041e-01 -9.00161922e-01 -8.50191772e-01 9.92064118e-01
-1.56586945e+00 -1.44698894e+00 -9.03090298e-01 8.35145831e-01
2.94902861e-01 -5.45180440e-01 4.62490678e-01 7.59106159e-01
-1.10220313e+00 1.08895040e+00 -1.87626362e-01 2.92959183e-01
6.19408786e-01 -8.67173076e-01 3.57871279e-02 9.51568604e-01
-3.29301953e-01 5.60172081e-01 4.95981157e-01 -6.16357446e-01
-1.58232009e+00 -1.35583878e+00 3.71128559e-01 -3.34745556e-01
9.45004672e-02 -4.65901703e-01 -8.71116817e-01 6.86049581e-01
1.06821038e-01 2.37138465e-01 5.53612292e-01 -1.80042654e-01
-7.15127885e-02 -5.50866961e-01 -1.44642568e+00 2.56769210e-01
1.01905262e+00 -3.63596529e-01 -1.45307109e-01 5.14653027e-01
7.99968466e-02 -4.27804828e-01 -5.55919349e-01 5.71009338e-01
7.96877563e-01 -1.11209929e+00 7.58401573e-01 -2.72196442e-01
-1.41729623e-01 -2.14315817e-01 -1.95406321e-02 -8.92613947e-01
-1.52466178e-01 -7.28243053e-01 1.89690918e-01 1.53763282e+00
3.70878121e-03 -7.76878715e-01 1.15008128e+00 5.51712751e-01
-1.98610243e-03 -5.78469753e-01 -6.06511831e-01 -7.56683350e-01
-6.49928093e-01 1.00164138e-01 1.04919934e+00 7.64130652e-01
-4.70178664e-01 9.36285108e-02 -6.14366531e-01 5.41157603e-01
8.92789364e-01 -4.20304053e-02 8.27910185e-01 -1.41110587e+00
-9.45770647e-03 -1.29645959e-01 -7.05499291e-01 -5.82256913e-01
1.10732466e-01 -5.13544083e-01 -6.36072503e-03 -9.36291993e-01
1.27873212e-01 -3.12731832e-01 -2.95410722e-01 5.98192453e-01
-2.83185899e-01 9.09772098e-01 -1.56258330e-01 2.82054305e-01
-1.63907483e-01 4.78249311e-01 1.04970574e+00 -1.66318536e-01
-2.19895229e-01 4.99588609e-01 -4.81319338e-01 6.26903296e-01
4.76336747e-01 -1.35662138e-01 -7.56419748e-02 -1.45757884e-01
-5.01740932e-01 -3.10696363e-01 2.22381890e-01 -1.34089518e+00
2.01666802e-01 2.09240839e-01 7.89603055e-01 -2.28282064e-01
5.76262116e-01 -6.47174716e-01 5.55138066e-02 5.67828715e-01
2.44608164e-01 -1.56184912e-01 2.66080886e-01 2.44091898e-01
-1.34008452e-01 -2.11754233e-01 9.22181368e-01 -1.79301575e-01
-5.30529261e-01 6.37005806e-01 -1.04064569e-01 -5.49997330e-01
8.89861703e-01 -5.48512518e-01 -1.47026554e-01 -5.56746982e-02
-9.30027127e-01 -1.60885975e-01 1.23153172e-01 5.02924562e-01
7.51765847e-01 -1.25856364e+00 -6.05328262e-01 8.74597609e-01
9.55367237e-02 -4.82370436e-01 7.37450719e-02 7.63057590e-01
-2.42609411e-01 1.22876368e-01 -3.04162234e-01 -6.02671444e-01
-1.86893797e+00 4.57289189e-01 5.85666120e-01 4.94708449e-01
-1.96463212e-01 9.79968727e-01 3.07954520e-01 -6.29158467e-02
5.38825393e-01 -2.63263017e-01 -6.62127793e-01 -7.97378421e-02
6.40854359e-01 5.94633929e-02 4.89829838e-01 -8.54467392e-01
-3.39801043e-01 8.41453373e-01 -7.76720122e-02 4.76350904e-01
1.19358087e+00 -2.95603685e-02 -2.93064475e-01 -6.72124922e-02
9.31615174e-01 -1.27389133e-01 -9.79584932e-01 -4.03984785e-01
-2.39024341e-01 -7.38589108e-01 3.50002535e-02 -5.17043710e-01
-1.41591787e+00 8.49504530e-01 1.38511670e+00 3.28523695e-01
1.16711962e+00 -2.16199517e-01 6.05870843e-01 2.79149950e-01
1.90495804e-01 -6.44653559e-01 2.31853664e-01 -1.05073545e-02
8.83501112e-01 -1.53773308e+00 -2.20369145e-01 -7.55651414e-01
-2.80211449e-01 1.08431852e+00 9.24711704e-01 -7.77733847e-02
9.71134901e-01 -2.12094054e-01 2.20897034e-01 -3.71628791e-01
-1.13360234e-01 -3.60454440e-01 4.89454210e-01 4.35713530e-01
2.95812875e-01 2.51050498e-02 1.30310208e-01 6.61068678e-01
-2.65440017e-01 3.58320445e-01 1.90063909e-01 6.37748003e-01
-5.53072751e-01 -8.80140841e-01 -6.97616518e-01 6.33174539e-01
-5.16682267e-01 1.26699135e-01 -4.02070969e-01 6.15734279e-01
4.95041370e-01 1.03356516e+00 -1.65706128e-01 -5.97545326e-01
3.95580411e-01 2.85341442e-02 5.18029869e-01 -6.38593256e-01
-8.43159199e-01 3.63786481e-02 -2.64370829e-01 -1.78157941e-01
-4.07993436e-01 -5.15838683e-01 -4.95864511e-01 -3.14027667e-01
-6.27316236e-01 -1.54756531e-01 5.18185198e-01 1.11896193e+00
2.30401635e-01 4.56839055e-01 9.51997459e-01 -8.53925645e-01
-5.42855322e-01 -1.37926531e+00 -6.67243123e-01 4.24785435e-01
3.56175482e-01 -9.78852510e-01 -2.88733333e-01 1.75193667e-01] | [13.295698165893555, 0.6522663831710815] |
c44becba-bf4b-4d37-9289-4a4a393ed954 | lightweight-convolution-transformer-for-cross | 2305.04325 | null | https://arxiv.org/abs/2305.04325v1 | https://arxiv.org/pdf/2305.04325v1.pdf | Lightweight Convolution Transformer for Cross-patient Seizure Detection in Multi-channel EEG Signals | Background: Epilepsy is a neurological illness affecting the brain that makes people more likely to experience frequent, spontaneous seizures. There has to be an accurate automated method for measuring seizure frequency and severity in order to assess the efficacy of pharmacological therapy for epilepsy. The drug quantities are often derived from patient reports which may cause significant issues owing to inadequate or inaccurate descriptions of seizures and their frequencies. Methods and materials: This study proposes a novel deep learning architecture based lightweight convolution transformer (LCT). The transformer is able to learn spatial and temporal correlated information simultaneously from the multi-channel electroencephalogram (EEG) signal to detect seizures at smaller segment lengths. In the proposed model, the lack of translation equivariance and localization of ViT is reduced using convolution tokenization, and rich information from the transformer encoder is extracted by sequence pooling instead of the learnable class token. Results: Extensive experimental results demonstrate that the proposed model of cross-patient learning can effectively detect seizures from the raw EEG signals. The accuracy and F1-score of seizure detection in the cross-patient case on the CHB-MIT dataset are shown to be 96.31% and 96.32%, respectively, at 0.5 sec segment length. In addition, the performance metrics show that the inclusion of inductive biases and attention-based pooling in the model enhances the performance and reduces the number of transformer encoder layers, which significantly reduces the computational complexity. In this research work, we provided a novel approach to enhance efficiency and simplify the architecture for multi-channel automated seizure detection. | ['Anil K. Tiwari', 'Salim Rukhsar'] | 2023-05-07 | null | null | null | null | ['seizure-detection', 'eeg', 'eeg'] | ['medical', 'methodology', 'time-series'] | [ 1.61410883e-01 -4.85213697e-01 2.88078517e-01 -2.50096321e-01
-1.05065107e+00 -3.36256683e-01 1.82436630e-01 1.88880399e-01
-6.46170080e-01 9.01791632e-01 7.93886259e-02 -2.65229642e-02
-3.29493791e-01 -3.76130491e-01 -4.35484350e-01 -8.17087829e-01
-5.50670505e-01 -2.56626666e-01 -1.11037344e-01 1.62957251e-01
1.78732544e-01 4.88698930e-01 -1.16466665e+00 6.24823213e-01
9.01192188e-01 1.35944712e+00 6.16036773e-01 3.56086701e-01
1.74172580e-01 7.34573662e-01 -9.63438153e-01 1.34948760e-01
-4.55325991e-02 -3.95855486e-01 -5.29491961e-01 -1.79915875e-01
-2.44663507e-01 -3.97702873e-01 -3.16447914e-01 1.15849018e+00
1.08853590e+00 -4.62100834e-01 5.39163053e-01 -9.43718374e-01
-9.48409587e-02 7.30975270e-01 -4.41666186e-01 5.98938346e-01
2.36445457e-01 7.33466968e-02 5.03688455e-01 -7.23940492e-01
-1.11136176e-02 6.08280063e-01 7.30681658e-01 3.11436832e-01
-1.00224638e+00 -1.23842108e+00 -1.74653977e-01 5.63401639e-01
-1.82013381e+00 -2.49392375e-01 5.72179496e-01 -3.94036382e-01
1.24802792e+00 1.80989921e-01 9.24270153e-01 1.19183290e+00
6.88606083e-01 6.56463504e-01 1.38910472e+00 -7.21242428e-02
9.59171057e-02 -8.21923371e-03 -7.38728121e-02 3.05258155e-01
1.03627987e-01 2.85374001e-02 -7.78632641e-01 -3.47529389e-02
7.44340777e-01 3.27969879e-01 -7.23251045e-01 3.04576039e-01
-1.24946773e+00 5.45484722e-01 5.16746998e-01 7.13714123e-01
-7.52678692e-01 4.70796749e-02 6.47379577e-01 1.94327489e-01
3.40927482e-01 6.66110873e-01 -4.86051083e-01 -4.45511073e-01
-1.20140314e+00 1.14297867e-03 4.10612792e-01 7.71827519e-01
3.11737746e-01 1.36752307e-01 -3.42091262e-01 5.69413602e-01
-2.58704066e-01 5.03355563e-01 1.05601680e+00 5.32016333e-04
4.58106101e-01 6.32693648e-01 -2.53011048e-01 -6.53828621e-01
-7.29405224e-01 -8.13840270e-01 -1.05117011e+00 -2.33821228e-01
-5.52046159e-03 -2.05233827e-01 -9.24114287e-01 1.56207156e+00
-5.22231281e-01 4.33640361e-01 1.25710353e-01 7.04261005e-01
6.87252700e-01 3.36067200e-01 7.25629777e-02 -2.31842831e-01
1.59700274e+00 -2.90384710e-01 -9.45093036e-01 -1.31164476e-01
5.44152737e-01 -6.34290576e-01 7.19150484e-01 4.61608529e-01
-1.00741041e+00 -2.50158101e-01 -1.11634755e+00 3.71321231e-01
-3.32062721e-01 5.01847267e-01 4.07652169e-01 4.22313154e-01
-8.45779121e-01 5.61867476e-01 -1.03435290e+00 7.17815803e-03
7.31605053e-01 8.35364819e-01 -3.27948421e-01 1.44983321e-01
-1.37329054e+00 9.08092737e-01 4.99648899e-01 1.68083653e-01
-9.84165728e-01 -7.57139206e-01 -5.07846653e-01 3.89424622e-01
-3.03279370e-01 -9.77712944e-02 9.56092536e-01 -8.07731390e-01
-1.02849054e+00 2.96785414e-01 -9.75302011e-02 -7.86831498e-01
1.93394333e-01 -1.11937672e-01 -4.83211070e-01 3.17650944e-01
-2.90944539e-02 4.97876197e-01 7.37129927e-01 -2.58129090e-01
-7.38325119e-01 -5.40468395e-01 -3.51234794e-01 2.33842582e-02
-4.63598758e-01 1.93435490e-01 -1.05609195e-02 -9.29194987e-01
-2.34818049e-02 -6.21931970e-01 1.06243305e-01 -6.56980813e-01
-1.69627011e-01 -1.20840691e-01 7.17168987e-01 -9.69516873e-01
1.35705471e+00 -2.28650665e+00 -2.10515410e-01 8.85358751e-02
2.40300238e-01 1.88690543e-01 7.33385608e-02 3.07601362e-01
-3.38059962e-01 -3.71477939e-02 -4.00055766e-01 9.23105180e-02
-2.23337203e-01 -3.41631293e-01 -3.00877064e-01 5.29208720e-01
3.11095417e-01 9.12112534e-01 -5.99137962e-01 -1.55584961e-01
1.52352989e-01 9.23475981e-01 -2.76789039e-01 3.01905483e-01
3.67266148e-01 5.41030288e-01 -4.04737502e-01 6.22497439e-01
4.98826623e-01 -1.49343148e-01 -1.17841646e-01 -3.99298608e-01
-1.68552458e-01 6.68201387e-01 -8.28961849e-01 1.85686028e+00
-4.06251252e-01 7.69015253e-01 -3.28909308e-01 -1.01211536e+00
7.90629864e-01 8.86262357e-01 5.86220801e-01 -1.08198905e+00
3.34209800e-01 4.77313608e-01 3.69014323e-01 -6.51001990e-01
-2.62243986e-01 7.62005597e-02 -5.27217798e-02 2.27768868e-01
1.02972522e-01 5.42288460e-02 -8.63476321e-02 -2.71723002e-01
1.30372572e+00 -2.22882807e-01 4.80164081e-01 -3.00732911e-01
4.28384274e-01 -3.55714470e-01 6.09657764e-01 4.59578723e-01
3.28217400e-03 1.64536655e-01 1.92770928e-01 -3.88181806e-01
-6.37505472e-01 -9.68649924e-01 -5.88305414e-01 3.03513438e-01
-1.18234068e-01 -2.62172133e-01 -9.15658534e-01 -1.61263406e-01
-2.83300459e-01 2.52447635e-01 -3.66282463e-01 -2.16628686e-01
-4.18353081e-01 -1.14029586e+00 8.02109241e-01 7.31525183e-01
8.60204518e-01 -1.22487843e+00 -1.15207589e+00 5.72135806e-01
-3.67667258e-01 -1.09760427e+00 -3.72379512e-01 5.70210338e-01
-8.38225007e-01 -9.62175667e-01 -8.80828381e-01 -9.28252101e-01
4.87124592e-01 -3.50943089e-01 4.77603167e-01 -3.97088200e-01
-7.90508866e-01 -1.02635652e-01 -3.67420197e-01 -5.58437765e-01
4.17069227e-01 2.38378644e-02 1.52069598e-01 3.84695046e-02
8.49908829e-01 -7.93874502e-01 -8.66447031e-01 -3.68853174e-02
-6.93443775e-01 -2.85566133e-02 9.98724580e-01 9.14644539e-01
5.07808149e-01 1.90446287e-01 9.51708138e-01 -2.86645263e-01
8.99292588e-01 -2.78149635e-01 -4.86431658e-01 6.53384030e-02
-3.60913336e-01 5.28141744e-02 7.42609978e-01 -6.12658858e-01
-5.06562710e-01 1.31177798e-01 3.04840207e-02 -2.39123508e-01
-5.24365976e-02 3.82904738e-01 -6.84575140e-02 -2.57244408e-01
2.48794511e-01 6.40209258e-01 -4.11876291e-01 -3.92405868e-01
-4.69102532e-01 1.06075752e+00 4.28858608e-01 1.83119420e-02
1.29584029e-01 1.75107539e-01 -2.44145066e-01 -7.98143446e-01
-1.80279911e-01 -3.73389691e-01 -3.13854694e-01 1.28205627e-01
9.49852884e-01 -1.26365471e+00 -9.45319355e-01 6.36216104e-01
-1.20366383e+00 1.60794444e-02 1.61449760e-01 1.11891031e+00
-3.90436888e-01 -7.91311935e-02 -7.09429860e-01 -7.77683437e-01
-9.24538732e-01 -1.52533889e+00 8.00531268e-01 4.42244299e-02
-3.20240080e-01 -5.75151205e-01 -3.39121968e-01 -1.76895395e-01
6.48105860e-01 1.73500761e-01 9.18909132e-01 -8.82511854e-01
-5.89561880e-01 -2.51195401e-01 -2.03925520e-01 3.84044051e-01
4.50085342e-01 -8.62779498e-01 -1.34891975e+00 -5.35608292e-01
3.94613922e-01 -1.17579550e-01 6.32124066e-01 5.48006713e-01
1.46930492e+00 -2.09631905e-01 -2.78607965e-01 7.09933281e-01
1.56280696e+00 7.84872949e-01 6.43809438e-01 1.10506736e-01
2.73847163e-01 1.66486040e-01 -4.15583327e-02 6.16634250e-01
6.64240718e-02 3.84486496e-01 1.79810271e-01 -1.42111957e-01
-4.62424308e-02 1.52871519e-01 3.50972623e-01 1.08144343e+00
1.34796441e-01 7.96340480e-02 -8.74250531e-01 8.34722459e-01
-1.31701338e+00 -8.78422499e-01 2.33844012e-01 2.29896140e+00
9.58844066e-01 1.09302504e-02 -2.69357264e-01 5.45493841e-01
5.69587827e-01 -4.11784321e-01 -5.09032726e-01 -1.62502572e-01
-1.64032996e-01 9.26726878e-01 5.48842609e-01 4.72326204e-02
-9.95301306e-01 3.96513790e-01 5.71988344e+00 8.74657571e-01
-1.60896671e+00 1.99925020e-01 2.93953598e-01 -4.08531934e-01
3.10149044e-01 -5.87541759e-01 -5.53762436e-01 7.96467841e-01
1.15815699e+00 -2.12840766e-01 5.81237137e-01 2.30501771e-01
3.83054018e-01 -5.08062281e-02 -1.19208252e+00 1.48463428e+00
5.34529388e-02 -1.07862675e+00 -1.30779296e-01 -6.51271343e-02
4.93775547e-01 3.09070081e-01 9.79208425e-02 7.59290382e-02
-4.81061965e-01 -1.30646861e+00 4.76224244e-01 6.07337832e-01
1.05654979e+00 -1.01126516e+00 1.02270758e+00 2.21288309e-01
-1.46592176e+00 -3.68748456e-01 -5.20326644e-02 2.26541124e-02
-2.05327809e-01 4.07136947e-01 -1.12809396e+00 3.06640834e-01
1.03286481e+00 6.64790332e-01 -4.51112479e-01 1.35119820e+00
-1.28143057e-01 6.39413297e-01 -2.06239522e-01 -2.36645684e-01
1.94583878e-01 2.14670658e-01 3.38910252e-01 1.21175015e+00
7.42580295e-01 1.39891878e-01 -8.75516087e-02 7.95183778e-01
-1.13290489e-01 5.82694262e-02 -4.55637306e-01 -1.57914251e-01
4.58474755e-01 1.02592885e+00 -4.83640164e-01 -4.78631444e-02
-4.06186908e-01 9.27863777e-01 -2.47215107e-02 3.38179350e-01
-7.51543224e-01 -9.62382913e-01 4.26327735e-01 -2.14289688e-02
2.64547706e-01 1.43642828e-01 -3.71230870e-01 -8.81548107e-01
3.33763629e-01 -9.00645673e-01 3.85199152e-02 -6.22622907e-01
-8.62662673e-01 1.09496808e+00 -3.14836293e-01 -1.38362086e+00
-3.03914219e-01 -4.70909655e-01 -4.54746753e-01 1.20808256e+00
-1.58630514e+00 -8.57023239e-01 -1.30892679e-01 9.52272296e-01
4.65199798e-01 -2.48786896e-01 1.14681351e+00 6.72187090e-01
-4.66755450e-01 7.01056182e-01 6.44319803e-02 2.80189425e-01
5.25482893e-01 -9.62222219e-01 6.37488216e-02 8.07303071e-01
-1.57188535e-01 5.81784606e-01 3.19194049e-01 -4.87609565e-01
-1.08701944e+00 -1.17790449e+00 9.42722738e-01 1.64858043e-01
4.93125826e-01 -5.05010545e-01 -8.95988882e-01 3.74830633e-01
2.68114805e-01 -1.74190372e-01 6.74304724e-01 -5.33706963e-01
3.22273076e-02 -4.28520560e-01 -1.22938120e+00 2.73009419e-01
5.00286818e-01 -8.96951973e-01 -6.91284955e-01 1.36636734e-01
3.29182953e-01 -1.38644248e-01 -9.96972919e-01 4.80444163e-01
5.92997551e-01 -6.27873123e-01 5.66542983e-01 -7.29482844e-02
3.76597680e-02 -1.17779940e-01 -7.24483877e-02 -1.40674186e+00
-7.36246556e-02 -5.81335127e-01 1.56955272e-01 7.19674230e-01
4.37707484e-01 -8.75359774e-01 3.76099586e-01 1.07877597e-01
-2.04773158e-01 -1.12450433e+00 -1.11151350e+00 -7.50113666e-01
-2.09997848e-01 -3.86208087e-01 8.62235904e-01 5.33732414e-01
4.46584344e-01 2.70419735e-02 -1.70147464e-01 3.72468024e-01
4.33273673e-01 -1.03822276e-01 -3.15963954e-01 -1.11330068e+00
-2.30045281e-02 -2.71664262e-01 -7.59047031e-01 -5.60938537e-01
-1.67805567e-01 -9.01513338e-01 -4.83337082e-02 -1.30291212e+00
2.75495321e-01 -1.20085038e-01 -9.26292300e-01 6.26534522e-01
2.08056509e-01 2.92700887e-01 -1.83163747e-01 -1.75839644e-02
-8.25196654e-02 3.32330078e-01 8.90776098e-01 -2.71533012e-01
-1.89736009e-01 -1.14272505e-01 -3.47632617e-01 5.99964499e-01
1.01796114e+00 -6.47492766e-01 -4.07779247e-01 -4.08883959e-01
-1.43628731e-01 1.28763273e-01 3.29892933e-01 -1.35206854e+00
4.55001891e-01 5.17041922e-01 8.10426414e-01 -5.97736835e-01
4.22726214e-01 -7.79854834e-01 2.09987666e-02 8.66946340e-01
-1.14609078e-01 3.44922960e-01 5.70429444e-01 1.99239805e-01
-4.02662635e-01 1.34933278e-01 6.56506419e-01 -2.72314027e-02
-4.82508034e-01 3.64907354e-01 -4.88644451e-01 -1.41355842e-01
1.03137505e+00 -1.20295115e-01 6.28528446e-02 -1.52393319e-02
-5.81501782e-01 -3.78311649e-02 -2.20698968e-01 3.25990498e-01
9.06199276e-01 -1.29053009e+00 -7.05007851e-01 6.57384157e-01
1.58179611e-01 -4.11224246e-01 3.47398400e-01 1.18326974e+00
-3.33296239e-01 9.36599553e-01 -4.89394039e-01 -5.17260373e-01
-1.20001042e+00 1.83761373e-01 5.67733884e-01 -1.66952938e-01
-7.41504550e-01 8.78928423e-01 2.04095557e-01 3.41139317e-01
4.45795834e-01 -8.15173686e-01 -2.92130053e-01 8.62538349e-03
8.96544218e-01 2.06161916e-01 6.38606906e-01 -3.79633904e-01
-6.61441743e-01 4.11318481e-01 -2.02312216e-01 -9.56876017e-03
1.48548877e+00 3.13689768e-01 -1.01492286e-01 2.07743630e-01
1.23981333e+00 -3.83853197e-01 -9.07321513e-01 7.15250671e-02
-1.09060749e-01 -4.93082553e-02 3.54389548e-01 -1.16142309e+00
-1.10375583e+00 1.21308959e+00 1.21139145e+00 -4.78017628e-02
1.66830504e+00 -3.58638257e-01 8.95831048e-01 3.06143016e-01
5.44829607e-01 -9.20661628e-01 -3.20101172e-01 1.51304916e-01
9.25525248e-01 -6.70936823e-01 -2.82610863e-01 1.78393468e-01
-3.52992266e-01 1.08826768e+00 1.84915841e-01 -1.36932328e-01
5.90815187e-01 7.20525146e-01 -1.41404957e-01 -2.23855272e-01
-5.63702166e-01 4.24449816e-02 1.79295570e-01 4.17018741e-01
4.88717616e-01 2.89716154e-01 -4.20404911e-01 1.04398799e+00
-1.88761100e-01 1.44063503e-01 2.73284137e-01 8.19603980e-01
-2.38253802e-01 -7.91983664e-01 -3.10659498e-01 8.13803077e-01
-9.13504243e-01 -6.08990967e-01 -2.14879420e-02 5.58827877e-01
2.72192419e-01 9.07731712e-01 1.83725476e-01 -4.38549191e-01
3.24762166e-01 1.74552917e-01 6.01429880e-01 -3.30821157e-01
-8.87339592e-01 2.72453368e-01 -4.50938493e-01 -4.91269588e-01
-2.60869414e-01 -3.73098433e-01 -1.39229298e+00 2.42919073e-01
-4.22217041e-01 2.59719700e-01 7.64227331e-01 1.00944018e+00
5.80548882e-01 9.98895407e-01 5.95357597e-01 -4.66169775e-01
-4.65509295e-01 -1.28486145e+00 -8.29177976e-01 8.89607221e-02
6.16100192e-01 -5.91754377e-01 -3.09923381e-01 -2.19574012e-02] | [13.223191261291504, 3.5023458003997803] |
fe1e4e2c-3676-470f-933d-ae7ed7d45087 | feature-robustness-in-non-stationary-health | 1908.00690 | null | https://arxiv.org/abs/1908.00690v1 | https://arxiv.org/pdf/1908.00690v1.pdf | Feature Robustness in Non-stationary Health Records: Caveats to Deployable Model Performance in Common Clinical Machine Learning Tasks | When training clinical prediction models from electronic health records (EHRs), a key concern should be a model's ability to sustain performance over time when deployed, even as care practices, database systems, and population demographics evolve. Due to de-identification requirements, however, current experimental practices for public EHR benchmarks (such as the MIMIC-III critical care dataset) are time agnostic, assigning care records to train or test sets without regard for the actual dates of care. As a result, current benchmarks cannot assess how well models trained on one year generalise to another. In this work, we obtain a Limited Data Use Agreement to access year of care for each record in MIMIC and show that all tested state-of-the-art models decay in prediction quality when trained on historical data and tested on future data, particularly in response to a system-wide record-keeping change in 2008 (0.29 drop in AUROC for mortality prediction, 0.10 drop in AUROC for length-of-stay prediction with a random forest classifier). We further develop a simple yet effective mitigation strategy: by aggregating raw features into expert-defined clinical concepts, we see only a 0.06 drop in AUROC for mortality prediction and a 0.03 drop in AUROC for length-of-stay prediction. We demonstrate that this aggregation strategy outperforms other automatic feature preprocessing techniques aimed at increasing robustness to data drift. We release our aggregated representations and code to encourage more deployable clinical prediction models. | ['Willie Boag', 'Michael C. Hughes', 'Tristan Naumann', 'Matthew B. A. McDermott', 'Bret Nestor', 'Marzyeh Ghassemi', 'Gabriela Berner', 'Anna Goldenberg'] | 2019-08-02 | null | null | null | null | ['length-of-stay-prediction'] | ['medical'] | [-5.90685159e-02 4.53344360e-02 -2.78625757e-01 -5.54736972e-01
-9.15275216e-01 -6.73639536e-01 2.33378589e-01 9.26439881e-01
-4.03326005e-01 7.65486598e-01 4.77915913e-01 -8.02661955e-01
-5.25433838e-01 -8.27627301e-01 -6.55086279e-01 -3.19975555e-01
-4.05841857e-01 5.09311914e-01 -1.59186006e-01 1.71707347e-01
-2.28196084e-01 3.70607764e-01 -9.41639841e-01 5.52299559e-01
7.15432227e-01 8.25556159e-01 -5.29915214e-01 6.35187209e-01
3.84465486e-01 8.43141973e-01 -6.09901965e-01 -2.53237814e-01
3.29381645e-01 -2.85667956e-01 -4.84438002e-01 -5.18871784e-01
1.07525393e-01 -3.15318972e-01 -3.39766651e-01 2.53740013e-01
6.03454411e-01 -4.24484581e-01 7.32205629e-01 -1.10497808e+00
-4.88953590e-01 7.54860997e-01 6.98297098e-02 2.59281695e-01
4.24693286e-01 4.39879447e-01 7.68889785e-01 -1.93105370e-01
6.52280390e-01 5.42607546e-01 1.40175653e+00 4.51423377e-01
-1.47065938e+00 -6.00444674e-01 1.49181023e-01 -2.24855646e-01
-1.26937187e+00 -4.06105995e-01 -1.56968445e-01 -7.65204728e-01
1.10070312e+00 6.06290102e-01 7.32726336e-01 8.78646314e-01
6.29805565e-01 1.64041415e-01 7.03550935e-01 -1.26614243e-01
1.74779728e-01 1.71194389e-01 3.41549248e-01 2.35081777e-01
6.40483677e-01 2.38943964e-01 -8.49921927e-02 -8.18781257e-01
2.83083469e-01 5.42162001e-01 -3.32489848e-01 -1.65750727e-01
-1.16355968e+00 4.97260928e-01 1.59471631e-01 1.30715743e-01
-6.04670823e-01 -1.29793033e-01 4.33476150e-01 4.13763672e-01
2.23502472e-01 5.98889172e-01 -1.09487522e+00 -3.39829296e-01
-1.09884346e+00 3.81100684e-01 9.09239292e-01 8.86067510e-01
1.33217737e-01 -3.06872487e-01 -5.81230164e-01 6.09597206e-01
-4.40437645e-02 5.24484873e-01 5.41174173e-01 -5.67133486e-01
3.02372307e-01 7.38703549e-01 2.62352526e-01 -6.04886055e-01
-9.23894346e-01 -5.98569930e-01 -7.23661304e-01 -3.65799904e-01
4.36881542e-01 -4.82307404e-01 -9.33423400e-01 1.69069922e+00
-4.07438651e-02 -7.19927698e-02 2.61295438e-01 1.85632527e-01
5.42803168e-01 2.73982465e-01 5.21982670e-01 -4.92502570e-01
1.33480108e+00 -1.02715656e-01 -4.55312341e-01 1.50343925e-01
1.25865340e+00 -3.34093302e-01 6.21213675e-01 3.71406794e-01
-1.00184560e+00 -1.45267278e-01 -7.23952770e-01 5.82400322e-01
-2.66606510e-01 -1.60214394e-01 4.73347902e-01 7.53894925e-01
-5.99690497e-01 7.96077669e-01 -1.15662992e+00 -6.22612715e-01
5.22806287e-01 4.33044016e-01 -4.06286985e-01 -1.64049909e-01
-1.11467469e+00 9.17544842e-01 1.59287721e-01 -3.48771423e-01
-4.65124488e-01 -1.47134328e+00 -5.45046091e-01 1.79254800e-01
-1.91437826e-01 -1.03310478e+00 1.13369393e+00 -5.17592192e-01
-5.48249841e-01 5.66616893e-01 6.27938733e-02 -8.95458341e-01
6.34554565e-01 -2.83977777e-01 -9.85159159e-01 -2.39569172e-01
1.11318836e-02 8.54641497e-02 3.38057391e-02 -1.02448404e+00
-8.25374663e-01 -4.95685369e-01 -3.82365495e-01 -2.05826566e-01
-2.95377493e-01 5.79729909e-03 6.63762726e-03 -6.33259594e-01
-1.70820996e-01 -9.40796733e-01 -5.20054281e-01 -4.85226959e-01
-1.14199132e-01 2.69849807e-01 3.47052753e-01 -8.30542147e-01
2.02425337e+00 -1.95623600e+00 -5.23748696e-01 2.25713387e-01
2.11602032e-01 3.71892937e-02 1.11110456e-01 6.14029765e-01
-2.77954012e-01 3.87566894e-01 -2.83250213e-01 8.64586458e-02
-5.62147915e-01 6.12127818e-02 -1.26674011e-01 4.11752492e-01
3.24381143e-01 7.11079657e-01 -7.86051631e-01 -2.94731915e-01
6.50970936e-02 3.54020238e-01 -9.93631363e-01 2.57185161e-01
1.47975370e-01 4.01145905e-01 -2.37971917e-01 7.90682137e-01
4.24317688e-01 -5.39892733e-01 3.79494518e-01 2.88241506e-02
-9.12980214e-02 4.06447172e-01 -7.47514248e-01 1.17634594e+00
-3.58060420e-01 3.39407176e-02 -5.43800473e-01 -7.39111841e-01
7.91608870e-01 5.31471848e-01 1.19058895e+00 -3.99855971e-01
-2.52841473e-01 6.42485172e-02 4.60066885e-01 -2.89829463e-01
-2.99151912e-02 -9.41670313e-02 -1.55007809e-01 3.32662910e-01
-2.91709930e-01 3.26262772e-01 -1.68751761e-01 1.10472078e-02
1.76273060e+00 -2.55854160e-01 5.05578816e-01 -3.62928599e-01
-4.38017584e-03 3.82762909e-01 9.29963589e-01 8.04742575e-01
-1.06148079e-01 8.71903241e-01 4.15116608e-01 -7.61254966e-01
-8.94465864e-01 -1.14900815e+00 -8.65074992e-01 6.29914165e-01
-5.91824174e-01 -5.95309317e-01 -3.90877813e-01 -7.83983707e-01
5.52740574e-01 9.09258544e-01 -6.48776710e-01 -5.25428832e-01
-4.47510868e-01 -1.27951288e+00 6.41268194e-01 6.08515024e-01
-2.28450462e-01 -7.76674211e-01 -1.04899037e+00 7.58409858e-01
2.10378960e-01 -5.50307989e-01 -1.94401667e-01 3.10040593e-01
-1.05679393e+00 -1.32784462e+00 -5.24186671e-01 -1.64709449e-01
4.10435915e-01 -3.97806555e-01 1.36747217e+00 1.25156194e-01
-2.51315296e-01 5.22338867e-01 -1.85038298e-01 -8.30211401e-01
-5.29128611e-01 2.52155572e-01 1.30602747e-01 -4.05061275e-01
6.91930652e-01 -4.45755154e-01 -9.84911919e-01 2.87749290e-01
-6.97525620e-01 -3.39968622e-01 3.94309431e-01 8.39997768e-01
4.65093404e-01 -4.80284542e-01 1.07608795e+00 -1.28580356e+00
5.13800740e-01 -1.01914334e+00 -3.22389066e-01 3.86221349e-01
-1.54487264e+00 -1.45936251e-01 6.04499698e-01 -2.54981905e-01
-6.07033372e-01 1.32064492e-01 9.46767554e-02 -1.76757976e-01
-1.93419442e-01 7.73362875e-01 3.53059173e-01 6.49195254e-01
9.22355354e-01 -4.16908860e-02 5.65270334e-02 -4.43561554e-01
-2.09466502e-01 8.98704350e-01 3.56681049e-01 -3.76687139e-01
4.34441000e-01 2.30890721e-01 -1.39700949e-01 -1.50531456e-01
-6.73502684e-01 -6.48075640e-01 -4.51483041e-01 2.19484493e-01
5.73374391e-01 -1.24752641e+00 -8.54419768e-01 7.52804875e-02
-6.07618570e-01 -3.55451077e-01 -4.24035758e-01 5.95649302e-01
-4.66751814e-01 -1.39295176e-01 -5.04119337e-01 -7.13350713e-01
-6.59351766e-01 -8.14024329e-01 7.08069205e-01 -4.62760180e-02
-8.96838009e-01 -1.03377473e+00 4.44323510e-01 -1.48442229e-02
5.36410689e-01 6.67179704e-01 1.18417013e+00 -1.07233322e+00
-6.73104376e-02 -5.33440590e-01 -3.22505869e-02 1.69345081e-01
6.33475482e-01 1.61121890e-01 -7.27765739e-01 -5.94331682e-01
-3.07551533e-01 2.59330302e-01 6.16648555e-01 6.80594027e-01
1.15424216e+00 -5.70104182e-01 -7.08367109e-01 6.75448477e-01
1.40796173e+00 5.92361748e-01 4.99644130e-01 4.56448376e-01
4.03761029e-01 3.74766141e-01 5.30284822e-01 8.41159463e-01
4.38336700e-01 4.73312169e-01 2.61478741e-02 -1.50477141e-01
3.33967030e-01 -1.37638405e-01 1.88576318e-02 3.27379644e-01
-1.22294962e-01 2.06511207e-02 -1.45936465e+00 8.27169538e-01
-1.84168124e+00 -9.59828317e-01 -2.38736689e-01 2.72265792e+00
8.63861680e-01 2.27773830e-01 4.42228198e-01 -3.87638137e-02
2.40437523e-01 -4.87249136e-01 -7.85274208e-01 -4.15757656e-01
3.28791179e-02 1.17450729e-01 9.33885932e-01 -7.43681416e-02
-7.63696253e-01 2.31726408e-01 7.05742264e+00 -2.83195406e-01
-1.07772267e+00 -1.08733550e-01 1.01915717e+00 -3.79843771e-01
-2.32738331e-02 -5.93215451e-02 -5.32502115e-01 6.41054988e-01
1.62550414e+00 -4.05891359e-01 -7.25473911e-02 8.80809486e-01
4.33508426e-01 1.76231235e-01 -1.43196678e+00 7.86065400e-01
-4.10399139e-01 -1.63469279e+00 -1.69242799e-01 1.39861375e-01
7.65235126e-01 2.11144641e-01 -2.50655442e-01 4.45724756e-01
5.16103506e-01 -1.21579874e+00 4.47934389e-01 8.63568187e-01
1.13218379e+00 -6.24239087e-01 9.99813676e-01 7.36989677e-02
-8.11876774e-01 -3.88320118e-01 4.57722321e-03 -6.28183708e-02
1.55028805e-01 6.43237948e-01 -1.04935074e+00 5.99164486e-01
9.77472961e-01 7.32601225e-01 -5.92319846e-01 1.17695177e+00
7.41102457e-01 1.14746737e+00 -2.75131315e-01 5.65486431e-01
-1.92580208e-01 3.63322258e-01 5.11756204e-02 1.32425928e+00
4.68048513e-01 3.33475500e-01 3.14724632e-02 3.25240314e-01
6.93046972e-02 7.64550716e-02 -6.57326937e-01 1.95405364e-01
5.62745631e-01 9.02700722e-01 -5.32999560e-02 -4.61698920e-01
-5.36843777e-01 3.37418556e-01 -1.04161642e-01 2.87497938e-01
-6.87457621e-01 -9.86316204e-02 9.82891679e-01 8.74664187e-01
-2.42359247e-02 3.68656337e-01 -6.97112322e-01 -1.00061941e+00
-1.95752710e-01 -1.04944158e+00 9.49581087e-01 -2.56419897e-01
-1.39672995e+00 4.83148068e-01 -9.48064495e-03 -1.37815738e+00
-5.80152333e-01 -2.39883959e-01 -3.97926807e-01 9.73459661e-01
-1.06309533e+00 -7.25377142e-01 -1.15896687e-01 3.38475645e-01
1.41923819e-02 -9.16633978e-02 1.26045454e+00 3.29732090e-01
-3.90864193e-01 8.89663637e-01 3.78070474e-01 2.63557374e-01
1.05896258e+00 -9.83757675e-01 3.03498536e-01 3.90075892e-01
-3.18720400e-01 9.84629273e-01 6.05428696e-01 -8.15421999e-01
-1.16178656e+00 -1.39642572e+00 8.39631617e-01 -1.10981750e+00
4.20802087e-01 9.56275910e-02 -1.11731577e+00 9.59167898e-01
-2.73476899e-01 6.35611564e-02 1.10094202e+00 4.92676705e-01
-2.80373871e-01 -3.48047972e-01 -1.47317970e+00 1.92810267e-01
9.10109103e-01 -1.50576398e-01 -5.57599545e-01 1.98529199e-01
6.45288825e-01 -2.54732877e-01 -1.77284169e+00 9.76625144e-01
9.52556908e-01 -7.80257821e-01 7.59023607e-01 -1.33074570e+00
3.66937757e-01 -7.15748891e-02 -1.36862144e-01 -1.20246887e+00
-6.22082531e-01 -5.07631958e-01 -7.98444003e-02 1.04360807e+00
7.81841576e-01 -8.38060975e-01 6.45073414e-01 1.14263368e+00
3.70960124e-02 -1.00044954e+00 -6.72478676e-01 -6.73137724e-01
2.89107323e-01 -2.68183112e-01 8.46191645e-01 1.26675308e+00
1.49513245e-01 -4.78364676e-02 -1.42290115e-01 3.19143146e-01
2.33706757e-01 -8.78243968e-02 6.48855925e-01 -1.41053128e+00
-3.83158624e-01 -2.23453507e-01 -4.29476202e-01 -4.17938381e-02
-5.80692947e-01 -6.36393428e-01 -4.47154909e-01 -1.72804725e+00
4.46955979e-01 -9.07011032e-01 -1.06108594e+00 7.48828948e-01
-3.67281616e-01 -2.26892844e-01 1.22320831e-01 4.02489632e-01
-1.44524381e-01 -1.35271668e-01 4.14340973e-01 3.40518877e-02
-5.86311638e-01 3.35230976e-01 -8.08602273e-01 2.01187044e-01
8.67837906e-01 -8.62998605e-01 -2.53264070e-01 -1.96950063e-01
-1.61557477e-02 3.82083029e-01 1.24988921e-01 -1.24708652e+00
-7.96460882e-02 -2.12081626e-01 6.78835869e-01 -3.74473929e-01
-3.01308721e-01 -9.31119621e-01 7.89688945e-01 1.01097512e+00
-4.33083177e-01 5.95029771e-01 4.92144912e-01 4.92588490e-01
-1.37852132e-02 4.06571239e-01 5.32158554e-01 8.15208256e-02
1.51656523e-01 3.54376465e-01 -4.40862566e-01 8.06041509e-02
1.12284696e+00 -1.94306001e-01 -2.71022260e-01 -1.33691072e-01
-8.00332129e-01 3.48477513e-01 5.73040545e-01 4.21121150e-01
2.43513938e-02 -1.05709410e+00 -9.52108264e-01 1.24681719e-01
5.38573325e-01 -4.42051291e-02 2.83545703e-01 8.43839467e-01
-3.95758957e-01 5.91174781e-01 -2.89955195e-02 -6.32795990e-01
-1.21581447e+00 7.55008221e-01 4.29034829e-01 -4.62353438e-01
-5.84895551e-01 2.04053730e-01 1.29725877e-02 -2.95364201e-01
6.13654628e-02 -6.63406909e-01 1.50585487e-01 -1.43038004e-03
5.36196291e-01 2.75472403e-01 3.93958807e-01 -1.53209344e-01
-6.84655547e-01 1.30095273e-01 -2.34662816e-01 3.29630136e-01
1.65440488e+00 2.21293762e-01 2.59053648e-01 5.73918760e-01
9.70651507e-01 -2.96714120e-02 -1.11490929e+00 1.85147688e-01
2.29435563e-01 -4.41834033e-02 -3.45234454e-01 -1.25330865e+00
-7.50461221e-01 3.07884127e-01 9.88460481e-01 2.34201223e-01
1.21127665e+00 -1.67028829e-01 3.92845869e-01 1.83113083e-01
1.97757155e-01 -6.97071791e-01 -8.67772281e-01 3.62571701e-02
5.65419972e-01 -1.09986079e+00 1.69265479e-01 1.13165684e-01
-6.34339988e-01 7.23776460e-01 3.54026943e-01 1.35976061e-01
1.03701115e+00 2.48052806e-01 2.04538614e-01 -4.10733521e-02
-1.32546282e+00 4.27032530e-01 -4.32415456e-02 6.60995364e-01
7.80582488e-01 5.38230598e-01 -3.58390361e-01 9.31503832e-01
-1.73120961e-01 3.85665268e-01 5.28844476e-01 8.72889400e-01
2.27295533e-02 -1.08425486e+00 -3.07380587e-01 1.27808309e+00
-8.29643250e-01 -3.21683496e-01 4.09753434e-03 8.38398278e-01
6.34676814e-02 8.46016407e-01 2.25042447e-01 -2.97916085e-01
8.57639253e-01 5.15200794e-01 -1.70592852e-02 -7.83616185e-01
-1.17341793e+00 -2.87689894e-01 2.58659154e-01 -5.11455357e-01
-5.61070479e-02 -1.06709063e+00 -1.22346902e+00 -1.75149515e-01
-4.18428667e-02 1.50390089e-01 3.80500317e-01 5.48932016e-01
9.77248311e-01 8.06742311e-01 4.03019398e-01 2.18564719e-01
-8.38779449e-01 -9.06954348e-01 -2.56835163e-01 6.74751043e-01
3.34094167e-01 -3.05945873e-01 -7.63899237e-02 2.32823446e-01] | [7.8495869636535645, 6.145815849304199] |
7e11f526-1f81-4aa5-a8e3-2a24cb99fe14 | meta-mimicking-embedding-via-others | 2112.08684 | null | https://arxiv.org/abs/2112.08684v3 | https://arxiv.org/pdf/2112.08684v3.pdf | Mimic Embedding via Adaptive Aggregation: Learning Generalizable Person Re-identification | Domain generalizable (DG) person re-identification (ReID) aims to test across unseen domains without access to the target domain data at training time, which is a realistic but challenging problem. In contrast to methods assuming an identical model for different domains, Mixture of Experts (MoE) exploits multiple domain-specific networks for leveraging complementary information between domains, obtaining impressive results. However, prior MoE-based DG ReID methods suffer from a large model size with the increase of the number of source domains, and most of them overlook the exploitation of domain-invariant characteristics. To handle the two issues above, this paper presents a new approach called Mimic Embedding via adapTive Aggregation (META) for DG person ReID. To avoid the large model size, experts in META do not adopt a branch network for each source domain but share all the parameters except for the batch normalization layers. Besides multiple experts, META leverages Instance Normalization (IN) and introduces it into a global branch to pursue invariant features across domains. Meanwhile, META considers the relevance of an unseen target sample and source domains via normalization statistics and develops an aggregation module to adaptively integrate multiple experts for mimicking unseen target domain. Benefiting from a proposed consistency loss and an episodic training algorithm, META is expected to mimic embedding for a truly unseen target domain. Extensive experiments verify that META surpasses state-of-the-art DG person ReID methods by a large margin. Our code is available at https://github.com/xbq1994/META. | ['Zhenan Sun', 'Lingxiao He', 'Jian Liang', 'Boqiang Xu'] | 2021-12-16 | null | null | null | null | ['generalizable-person-re-identification'] | ['computer-vision'] | [-1.93869293e-01 -1.37115926e-01 -2.31663793e-01 -4.24101263e-01
-6.47308528e-01 -6.54360473e-01 7.52197802e-01 -1.69486195e-01
-5.71481764e-01 7.59207368e-01 1.29277885e-01 2.53984511e-01
-1.09489709e-01 -6.68071628e-01 -5.23756564e-01 -5.59653878e-01
1.55442610e-01 6.84262931e-01 9.82285738e-02 -2.47126177e-01
-3.41776669e-01 3.11297834e-01 -1.29910600e+00 -1.40015662e-01
1.12199461e+00 7.14591682e-01 -1.55835897e-01 4.39485237e-02
-2.68549509e-02 7.25163668e-02 -7.02342808e-01 -9.77036655e-01
5.53461492e-01 -3.20220172e-01 -3.92016351e-01 -1.23684309e-01
6.87834680e-01 -5.44879794e-01 -7.24102080e-01 1.14760399e+00
8.88138652e-01 2.61869431e-01 7.26398766e-01 -1.56079316e+00
-1.12301040e+00 3.48146796e-01 -4.01484400e-01 2.40729034e-01
2.78184235e-01 2.16871187e-01 6.02491558e-01 -9.47881460e-01
4.55455959e-01 1.16759932e+00 6.84965432e-01 1.03776312e+00
-1.16720688e+00 -1.08257759e+00 5.48556626e-01 9.84155573e-03
-1.53332829e+00 -4.95685846e-01 8.70294869e-01 -2.49746114e-01
5.22619426e-01 4.37914319e-02 3.33374500e-01 1.77027118e+00
-4.23138916e-01 7.90188432e-01 9.20134008e-01 -5.38584106e-02
3.05030167e-01 5.45422614e-01 3.75846326e-01 2.16589883e-01
6.20148778e-01 2.01574877e-01 -4.86632526e-01 -3.15586358e-01
6.79683924e-01 7.28284121e-02 -3.00499707e-01 -5.27005672e-01
-1.10195339e+00 5.84305525e-01 4.41301435e-01 1.67960301e-01
-3.03227305e-01 -3.29312235e-01 4.02100444e-01 3.97251427e-01
4.88504827e-01 2.18619049e-01 -3.95700783e-01 2.10120156e-01
-8.38361442e-01 4.89231110e-01 6.91505909e-01 1.15518570e+00
7.98029006e-01 5.62429503e-02 -5.44092618e-02 9.13590014e-01
7.69721791e-02 6.40716434e-01 7.77203500e-01 -5.92295229e-01
7.10963428e-01 8.32248628e-01 1.96953967e-01 -7.30287194e-01
-5.63054904e-02 -8.56008530e-01 -1.05815160e+00 -3.29545466e-03
4.47881281e-01 -2.74823070e-01 -8.42837214e-01 2.27288222e+00
6.24159992e-01 4.45573896e-01 1.54868022e-01 8.85647416e-01
7.72169411e-01 1.54991925e-01 2.65276313e-01 3.34526122e-01
1.46213627e+00 -8.60528290e-01 -3.43437672e-01 -5.87486923e-01
3.51654887e-01 -2.40738213e-01 8.07644844e-01 9.45868343e-02
-8.86985660e-01 -7.85537601e-01 -1.19286215e+00 -1.05216861e-01
-6.33689702e-01 1.80674881e-01 4.06679392e-01 8.03321064e-01
-9.72263157e-01 3.45625758e-01 -3.23731393e-01 -6.92516267e-01
4.24154311e-01 5.77771902e-01 -7.22623467e-01 -3.18297356e-01
-1.43669283e+00 7.19790995e-01 4.35509294e-01 -1.78731438e-02
-7.80875921e-01 -8.74019325e-01 -9.89061952e-01 -8.03687796e-03
1.88529268e-01 -1.10855293e+00 8.85277271e-01 -1.06597912e+00
-1.37463605e+00 8.69097054e-01 -2.61200964e-01 -3.59798580e-01
9.51429725e-01 -3.88451517e-01 -7.59316862e-01 -4.39011753e-02
3.80139291e-01 6.93787873e-01 1.07991445e+00 -1.08604801e+00
-4.72807556e-01 -4.56240505e-01 1.16422437e-01 1.91620946e-01
-7.90892780e-01 -1.83686376e-01 -5.82082272e-01 -7.89055526e-01
-2.39433363e-01 -8.41415942e-01 2.73898765e-02 1.16294492e-02
-2.22985148e-01 -3.35672259e-01 5.82348883e-01 -5.59880793e-01
1.00598991e+00 -2.33677030e+00 7.08159059e-02 1.46564156e-01
4.91890162e-01 4.14831430e-01 -4.54321474e-01 2.04756275e-01
-1.94437861e-01 -1.38186337e-02 -1.63587317e-01 -7.10439563e-01
3.33863825e-01 1.39116734e-01 -2.02660412e-01 5.25908232e-01
1.89679265e-01 9.58017230e-01 -9.38149512e-01 -2.80483961e-01
9.31698009e-02 5.81333160e-01 -4.15037185e-01 1.53488144e-01
1.75244465e-01 5.70869267e-01 -5.96364379e-01 6.38954043e-01
1.13939834e+00 -2.51388282e-01 1.28496766e-01 -5.84367998e-02
2.65062004e-01 8.92821401e-02 -1.32999539e+00 1.60822332e+00
-9.45555419e-02 1.24705434e-01 4.45112251e-02 -1.04140615e+00
9.43300188e-01 3.55057687e-01 2.89452910e-01 -7.60388434e-01
9.25963521e-02 3.68019879e-01 -2.14078501e-01 -7.93841407e-02
4.22494382e-01 -4.76831011e-02 -3.75268966e-01 2.84342736e-01
2.64188230e-01 7.00104237e-01 1.58063080e-02 1.55358732e-01
1.00068891e+00 -2.46584844e-02 1.81040570e-01 -1.28531337e-01
6.70961022e-01 -2.47393206e-01 9.38467324e-01 8.74744892e-01
-5.22263527e-01 6.19671583e-01 5.36230952e-02 -3.43593150e-01
-1.05601442e+00 -1.42384577e+00 -1.41987860e-01 9.88496065e-01
4.71550047e-01 -1.80299971e-02 -6.91548228e-01 -9.56324518e-01
5.02266228e-01 5.04873872e-01 -8.00007522e-01 -3.23657930e-01
-6.24605060e-01 -7.87054241e-01 7.51898587e-01 5.74557781e-01
9.37127292e-01 -5.27791202e-01 1.75099626e-01 1.04977019e-01
-1.18941978e-01 -1.14594948e+00 -6.92449510e-01 -3.34747016e-01
-6.12951636e-01 -9.51289237e-01 -1.25635982e+00 -7.08212197e-01
7.46598721e-01 5.06205738e-01 8.82133543e-01 -3.19204062e-01
8.53265449e-03 7.49491155e-01 -3.21881145e-01 -3.10509861e-01
-2.12194115e-01 3.12255621e-01 6.19962692e-01 1.88361406e-01
9.12446558e-01 -7.77037859e-01 -7.52290130e-01 6.51038766e-01
-8.24037015e-01 -3.52409303e-01 5.76826990e-01 9.23325121e-01
2.28582546e-01 -5.27115986e-02 9.67860997e-01 -5.33194721e-01
7.23301053e-01 -9.24159408e-01 -4.32875067e-01 3.20685983e-01
-5.42785466e-01 -6.74155429e-02 7.70178676e-01 -9.60119128e-01
-1.07736075e+00 -3.81656229e-01 2.35013738e-01 -5.45126498e-01
-2.81204998e-01 5.56164570e-02 -6.33071363e-01 6.59389794e-02
6.87958956e-01 3.11252862e-01 3.66469286e-02 -6.88822269e-01
2.14767411e-01 6.08221233e-01 7.02302873e-01 -6.68408155e-01
1.39112318e+00 5.28780878e-01 -4.70054209e-01 -3.68733972e-01
-5.22314847e-01 -5.73452711e-01 -6.79240763e-01 1.38943329e-01
6.35543287e-01 -1.38283145e+00 -4.84517664e-01 6.23599350e-01
-9.84459996e-01 -9.20243785e-02 -3.40191513e-01 5.78612030e-01
4.59825667e-03 5.78525901e-01 -3.23717356e-01 -5.09036779e-01
-2.34439492e-01 -7.39882052e-01 8.05999815e-01 4.25714821e-01
-2.80387819e-01 -1.03710485e+00 -5.64532466e-02 4.29944187e-01
5.13701677e-01 5.95401675e-02 5.85740328e-01 -1.17913783e+00
-4.74626929e-01 -5.25781155e-01 -3.23851496e-01 5.40625453e-01
2.64012396e-01 -7.90955424e-01 -1.19504809e+00 -6.94602191e-01
-1.30107729e-02 -8.89186934e-02 7.71979928e-01 -1.50362318e-02
7.35788047e-01 -2.54319549e-01 -4.64656621e-01 6.85205162e-01
1.14545262e+00 -1.51419804e-01 3.25784415e-01 5.94276249e-01
6.35498047e-01 6.28180385e-01 4.03775632e-01 3.98016304e-01
6.61696911e-01 6.58528030e-01 5.64716235e-02 1.02497106e-02
-8.48762989e-02 -4.74066585e-01 6.09716594e-01 3.30948651e-01
5.08845374e-02 -4.12422210e-01 -6.14601970e-01 7.42899895e-01
-1.73563266e+00 -9.54608321e-01 4.21543419e-01 2.46380782e+00
5.53817987e-01 -6.29002154e-02 4.23053890e-01 -3.03409427e-01
1.02624083e+00 -8.36679488e-02 -1.14267552e+00 1.48541987e-01
-3.79180133e-01 -1.42634595e-02 5.05025506e-01 6.13964126e-02
-1.13575113e+00 8.08761179e-01 5.09250021e+00 8.62772584e-01
-8.11131716e-01 3.60794902e-01 2.11313620e-01 -1.97794095e-01
-2.18049958e-01 -2.16784462e-01 -1.17799592e+00 7.44702756e-01
7.86204159e-01 -5.54879665e-01 3.66056383e-01 1.03750467e+00
-3.30178201e-01 3.51696283e-01 -1.26729536e+00 9.84337509e-01
5.70069402e-02 -8.46180797e-01 2.37882748e-01 2.52108544e-01
6.42126262e-01 6.22728355e-02 3.84475827e-01 6.45400167e-01
3.77763867e-01 -6.92524135e-01 5.93686223e-01 3.00959736e-01
6.69363797e-01 -6.24495566e-01 7.96973944e-01 3.11025202e-01
-1.16921878e+00 -1.55311197e-01 -5.82113564e-01 2.04318747e-01
3.58601101e-04 4.18397576e-01 -6.43227041e-01 8.62436593e-01
6.93168402e-01 6.84357882e-01 -6.87783241e-01 9.50729430e-01
-1.20020017e-01 2.42880866e-01 -3.44293952e-01 4.33424294e-01
-7.15885311e-02 1.55098597e-02 7.97869623e-01 1.00054741e+00
3.71618748e-01 -2.34174326e-01 1.52732223e-01 9.84735727e-01
-3.22455436e-01 -1.37185127e-01 -5.25701523e-01 1.28721744e-01
8.52616370e-01 9.46100473e-01 3.31920572e-02 -4.98038888e-01
-4.80591029e-01 1.50920606e+00 3.38285089e-01 7.86343098e-01
-9.43160355e-01 -2.39069507e-01 1.09040058e+00 1.81245312e-01
2.19863579e-01 -5.63755110e-02 2.10642926e-02 -1.48114038e+00
4.91339982e-01 -8.63762438e-01 5.37970126e-01 -1.85970351e-01
-2.11911440e+00 5.59976578e-01 1.67905182e-01 -1.63508904e+00
-2.18001068e-01 -5.08071482e-01 -5.81213534e-01 1.27879190e+00
-1.82724690e+00 -1.42735386e+00 -3.21773469e-01 8.90342414e-01
2.62420267e-01 -5.07185996e-01 6.88074648e-01 5.31863689e-01
-9.43238080e-01 1.44379008e+00 2.76302814e-01 3.89325291e-01
1.15701234e+00 -1.15054297e+00 7.51032412e-01 8.84891212e-01
-4.81473833e-01 1.12360036e+00 4.12323773e-01 -7.41543293e-01
-1.17190289e+00 -1.27015615e+00 8.63749325e-01 -6.35895729e-01
5.50177276e-01 -5.60607553e-01 -1.14448059e+00 8.04839432e-01
-6.46853894e-02 1.66451767e-01 8.79753470e-01 8.78804177e-02
-7.55707681e-01 -3.23138207e-01 -1.44536960e+00 5.31481862e-01
1.41603637e+00 -6.38152599e-01 -6.60951078e-01 6.18707947e-02
5.21714091e-01 -1.85110137e-01 -8.86315823e-01 3.29664618e-01
5.11981905e-01 -8.20401490e-01 1.25933373e+00 -5.74829817e-01
-1.33438617e-01 -3.92835766e-01 5.49000204e-02 -1.24798000e+00
-5.35541594e-01 -4.73159641e-01 -1.16828486e-01 1.61316276e+00
8.55587423e-02 -1.36250138e+00 7.49839842e-01 1.06283724e+00
7.67174736e-02 -2.00977087e-01 -1.12673199e+00 -1.35684037e+00
2.95680642e-01 -4.59075831e-02 1.08740628e+00 1.12445474e+00
-3.02762479e-01 1.49430245e-01 -3.73380721e-01 5.77598631e-01
1.04517782e+00 -2.02403009e-01 1.06924188e+00 -1.50538445e+00
-2.79607266e-01 -3.28891128e-01 -5.07823467e-01 -1.11458135e+00
4.49894458e-01 -9.97299731e-01 -4.60449904e-01 -1.15223849e+00
2.62922466e-01 -5.06240606e-01 -6.41772091e-01 4.00968462e-01
-3.86372715e-01 1.86432377e-01 1.65898293e-01 3.96397948e-01
-6.07470274e-01 6.98036849e-01 9.84489620e-01 -3.21638525e-01
-1.81470498e-01 -7.89837390e-02 -1.15975952e+00 5.21353304e-01
8.76270056e-01 -3.87282878e-01 -5.65135598e-01 -5.26396871e-01
-1.97381631e-01 -5.45539320e-01 9.02121961e-01 -1.16972399e+00
4.43916351e-01 2.28338212e-01 8.20300281e-01 -1.55858889e-01
4.26113516e-01 -7.02555835e-01 2.12352097e-01 8.33627884e-04
5.10895178e-02 2.60374304e-02 3.43340784e-01 7.50559866e-01
-1.70324668e-01 2.03381926e-02 7.32969224e-01 3.14660706e-02
-8.73254836e-01 6.11312151e-01 1.09565206e-01 8.53091329e-02
9.36693907e-01 -5.24479389e-01 -5.74879885e-01 -1.19407184e-01
-5.50758123e-01 5.32811284e-01 7.85837293e-01 6.13182127e-01
5.00898063e-01 -1.37096584e+00 -8.19627047e-01 4.39299077e-01
3.43758017e-01 -6.13927059e-02 7.15039134e-01 6.04968309e-01
2.46075496e-01 1.82162911e-01 -1.70678258e-01 -2.75460631e-01
-9.71198559e-01 5.97384214e-01 4.02030528e-01 -3.47224057e-01
-5.14111102e-01 9.02052104e-01 6.77773476e-01 -7.34283030e-01
1.49547547e-01 3.41352910e-01 1.57174133e-02 1.24782234e-01
7.53040254e-01 4.10658270e-01 -2.16072649e-01 -6.21359289e-01
-4.32641238e-01 5.11753440e-01 -5.72261870e-01 5.33244386e-02
1.06842792e+00 -2.37303749e-01 1.82788178e-01 -6.38707504e-02
1.20329595e+00 2.20107976e-02 -1.35267806e+00 -6.08106613e-01
-1.57697916e-01 -4.40879405e-01 -4.15991277e-01 -8.77944231e-01
-7.53420413e-01 6.02523863e-01 7.09101021e-01 -3.00528318e-01
1.12834013e+00 -3.41392048e-02 9.78363335e-01 2.40398571e-01
4.91244584e-01 -1.04970694e+00 5.25707379e-03 2.96835303e-01
6.69756114e-01 -1.36092746e+00 -1.34763524e-01 -2.53761947e-01
-6.23909950e-01 7.57172704e-01 8.72247815e-01 -1.52265891e-01
4.37395781e-01 -2.13396668e-01 -3.53081524e-02 1.80673599e-01
-2.58744419e-01 -1.12375416e-01 3.33561480e-01 9.49691236e-01
-1.10756181e-01 1.30753191e-02 2.46459749e-02 1.02973533e+00
-9.79560912e-02 4.05774005e-02 6.75268322e-02 6.61346555e-01
-1.40315266e-02 -1.55179918e+00 -3.81391555e-01 1.75549611e-01
-1.95687622e-01 -3.10686324e-02 -3.50172609e-01 9.83419359e-01
3.91106725e-01 7.13385701e-01 -1.14893600e-01 -4.83776927e-01
4.54531997e-01 1.20438859e-01 2.07324460e-01 -4.22377408e-01
-6.18943334e-01 -4.27037179e-01 -2.49664694e-01 -2.92786956e-01
-2.49406233e-01 -6.74189746e-01 -7.88115442e-01 -5.52011549e-01
-1.55771533e-02 1.79284275e-01 2.41094708e-01 6.74070477e-01
7.84214914e-01 1.98196173e-01 5.88108361e-01 -6.17092550e-01
-7.71677077e-01 -9.03712451e-01 -6.73030138e-01 6.30652547e-01
4.31444377e-01 -8.78254473e-01 -5.69770634e-01 -2.72376269e-01] | [14.727566719055176, 1.1051923036575317] |
24060922-a809-4e73-94a3-93becfd02c19 | cross-modality-attention-with-semantic-graph | 1912.07872 | null | https://arxiv.org/abs/1912.07872v2 | https://arxiv.org/pdf/1912.07872v2.pdf | Cross-Modality Attention with Semantic Graph Embedding for Multi-Label Classification | Multi-label image and video classification are fundamental yet challenging tasks in computer vision. The main challenges lie in capturing spatial or temporal dependencies between labels and discovering the locations of discriminative features for each class. In order to overcome these challenges, we propose to use cross-modality attention with semantic graph embedding for multi label classification. Based on the constructed label graph, we propose an adjacency-based similarity graph embedding method to learn semantic label embeddings, which explicitly exploit label relationships. Then our novel cross-modality attention maps are generated with the guidance of learned label embeddings. Experiments on two multi-label image classification datasets (MS-COCO and NUS-WIDE) show our method outperforms other existing state-of-the-arts. In addition, we validate our method on a large multi-label video classification dataset (YouTube-8M Segments) and the evaluation results demonstrate the generalization capability of our method. | ['Shilei Wen', 'Yingze Bao', 'Zhiyao Guo', 'Xiang Long', 'Lei Cui', 'Renchun You'] | 2019-12-17 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 2.69918680e-01 -4.39545125e-01 -5.40445268e-01 -4.43788797e-01
-7.40335584e-01 -6.13384724e-01 4.29702669e-01 2.04915658e-01
-3.20650846e-01 2.61982203e-01 9.32566896e-02 8.50869864e-02
-3.14793587e-01 -2.64535725e-01 -5.11382997e-01 -6.08135998e-01
8.64401087e-02 1.19721934e-01 1.56199917e-01 2.55111098e-01
4.22330290e-01 2.03261264e-02 -1.44475543e+00 4.06806439e-01
4.39739019e-01 1.23642719e+00 1.50759041e-01 2.53527045e-01
1.08387833e-03 9.89947379e-01 2.21905768e-01 -5.30018285e-02
7.08233938e-02 -2.98175603e-01 -9.65171993e-01 3.89098495e-01
7.77933657e-01 -2.90113054e-02 -4.44933623e-01 1.16361237e+00
2.35108703e-01 2.28899717e-01 8.32653344e-01 -1.56222975e+00
-9.83892262e-01 -1.44359199e-02 -8.80750239e-01 2.38376826e-01
3.61432493e-01 -2.75347859e-01 1.37310219e+00 -8.63389432e-01
6.29035294e-01 1.22026563e+00 7.44033456e-01 6.38902307e-01
-1.09564030e+00 -8.35877955e-01 4.65710700e-01 4.44084615e-01
-1.45742226e+00 -8.18746090e-02 8.23442400e-01 -6.36913538e-01
5.23441553e-01 5.02769873e-02 3.04296464e-01 1.14801693e+00
9.90196466e-02 8.60402465e-01 1.22991991e+00 -4.09720480e-01
-1.72742069e-01 -3.57077084e-02 2.66047448e-01 1.21794808e+00
-2.52686173e-01 -1.46713376e-01 -6.26643777e-01 -2.44471073e-01
6.75816953e-01 4.62560833e-01 -2.33047113e-01 -6.73670530e-01
-1.40276062e+00 1.06415713e+00 4.95855689e-01 2.58645982e-01
-1.42245308e-01 4.92446065e-01 8.88446689e-01 1.23997927e-01
8.31303000e-01 1.10445246e-01 -2.41052687e-01 1.89880997e-01
-5.77569246e-01 -1.81908131e-01 2.22799525e-01 9.85562086e-01
7.83681333e-01 -4.96738225e-01 -1.84373558e-01 1.23474061e+00
7.54006207e-01 2.47348383e-01 5.60718298e-01 -8.22074413e-01
4.34883535e-01 8.28986704e-01 -4.11106437e-01 -1.21161520e+00
-5.09812474e-01 -1.13850035e-01 -5.94263971e-01 -2.27803499e-01
6.83552697e-02 2.51618952e-01 -8.57444525e-01 1.66895962e+00
3.50319684e-01 9.21498477e-01 -1.79399222e-01 8.70751202e-01
8.72757435e-01 4.68681753e-01 5.33636332e-01 -1.51189789e-01
1.41233909e+00 -1.49738955e+00 -7.23205209e-01 -1.42811224e-01
1.05250728e+00 -3.97178888e-01 8.92061830e-01 -2.10370496e-01
-4.76376384e-01 -6.06869698e-01 -7.21614122e-01 -4.03781608e-02
-3.83051097e-01 1.35545164e-01 6.75843894e-01 1.80320695e-01
-8.31487954e-01 4.16635454e-01 -3.54920954e-01 -5.57399631e-01
7.04117417e-01 1.98724240e-01 -6.82002425e-01 -5.15610397e-01
-1.08748603e+00 4.20706272e-01 4.84511822e-01 -1.56948820e-01
-8.93568277e-01 -5.98637104e-01 -1.19456065e+00 -1.54061422e-01
3.17166209e-01 -4.07350898e-01 8.04363310e-01 -1.09345090e+00
-9.23304617e-01 1.28945708e+00 -1.17203116e-01 6.75590187e-02
4.40399684e-02 1.24187283e-01 -5.11293292e-01 5.49869418e-01
6.76999629e-01 9.68882143e-01 6.22425020e-01 -1.33025348e+00
-8.88638198e-01 -4.73899096e-01 1.76392481e-01 2.95574814e-01
-7.06015170e-01 7.38506615e-02 -4.70608979e-01 -6.04285538e-01
2.50441760e-01 -1.20810378e+00 -1.06107891e-02 -8.42954293e-02
-2.52052069e-01 -6.13821805e-01 9.96588051e-01 -3.65880936e-01
1.16273856e+00 -2.31395316e+00 2.67147541e-01 1.14917643e-01
2.70297915e-01 -1.48946494e-01 -4.85836565e-01 3.78979951e-01
-2.80101597e-01 1.40672058e-01 1.82730868e-01 -4.27857548e-01
-6.90819249e-02 1.80704936e-01 -4.58414964e-02 8.37256372e-01
6.67986423e-02 1.10187852e+00 -1.10373580e+00 -8.34581137e-01
1.58963040e-01 4.18855608e-01 -3.34530205e-01 1.51229843e-01
-9.07316953e-02 6.58624530e-01 -6.68615162e-01 9.16967392e-01
2.16410965e-01 -8.40762138e-01 3.78236115e-01 -5.84731340e-01
2.72806108e-01 -2.77609408e-01 -8.08689058e-01 2.19148731e+00
-5.11305273e-01 4.93695766e-01 -3.92916411e-01 -1.26683795e+00
4.65705454e-01 4.15321022e-01 7.83808947e-01 -6.95994079e-01
2.97559172e-01 3.68709862e-02 -6.67912424e-01 -7.63423562e-01
1.26920193e-01 3.36528569e-02 -1.65371597e-01 5.14037311e-01
3.72220546e-01 6.35872781e-01 2.52009392e-01 3.73292327e-01
8.59986186e-01 1.10700682e-01 1.92584440e-01 -3.57347816e-01
5.91294646e-01 -2.77099013e-01 7.57621825e-01 3.34715694e-01
-5.03825545e-01 5.03372967e-01 4.61456478e-01 -5.08178234e-01
-5.87152421e-01 -6.72458231e-01 -3.01825311e-02 1.48852587e+00
5.20184696e-01 -5.43845236e-01 -4.17264163e-01 -1.52172351e+00
-1.91555209e-02 1.86720528e-02 -9.93391275e-01 3.00055444e-02
-3.27105105e-01 -4.65764463e-01 2.78309911e-01 6.19719863e-01
4.04270619e-01 -8.18844795e-01 1.53791383e-02 1.25197917e-02
-2.86518425e-01 -1.35941529e+00 -9.35461164e-01 -1.56112522e-01
-6.48482919e-01 -1.45368040e+00 -4.89073038e-01 -1.39252210e+00
8.03380430e-01 7.05946624e-01 7.97623873e-01 2.03678340e-01
-3.74475241e-01 9.04349387e-01 -4.42800134e-01 3.10154498e-01
-1.67343188e-02 6.71127625e-03 -9.23154652e-02 4.75886673e-01
4.64706331e-01 -4.95626271e-01 -5.78476310e-01 5.08934796e-01
-8.46635401e-01 1.98166102e-01 3.28212082e-01 1.07745755e+00
8.99619818e-01 -3.01531672e-01 9.14699495e-01 -1.10687947e+00
2.63977677e-01 -8.78856599e-01 -3.94178212e-01 6.65390432e-01
-6.67632818e-01 -1.32761508e-01 2.10520104e-01 -5.31310916e-01
-5.07386744e-01 3.18777859e-01 2.99671590e-01 -8.73648405e-01
-2.29376659e-01 5.40611565e-01 -3.52260959e-03 -3.96992147e-01
4.57812510e-02 5.96552305e-02 -3.64846922e-02 -3.65369916e-01
6.02801383e-01 6.80614471e-01 1.43055454e-01 -4.82340723e-01
3.29268515e-01 6.46804452e-01 1.35601789e-01 -3.08630586e-01
-1.30387497e+00 -1.07878983e+00 -7.42691040e-01 -5.53325236e-01
1.38459909e+00 -1.17433906e+00 -7.94537365e-01 3.40362668e-01
-9.30131376e-01 -1.58632144e-01 2.77414769e-01 4.16534126e-01
-6.41384244e-01 5.84324360e-01 -8.85299325e-01 -2.25681812e-01
-3.14772427e-02 -1.35486984e+00 1.31997311e+00 2.42505550e-01
7.48042762e-02 -1.58550322e+00 2.93029964e-01 7.55139291e-01
-1.01959795e-01 2.02995300e-01 9.65256929e-01 -6.24348223e-01
-4.75353122e-01 2.11716089e-02 -7.59868264e-01 1.25083372e-01
2.08201110e-01 -4.33037311e-01 -8.70520055e-01 -5.03746033e-01
-5.25232732e-01 -7.47434020e-01 1.01756954e+00 1.81652695e-01
1.27500892e+00 -5.67002147e-02 -7.67843306e-01 4.86758858e-01
1.74832928e+00 -1.14603721e-01 9.60328206e-02 2.99090117e-01
1.31697261e+00 5.21288931e-01 8.22697043e-01 1.44376740e-01
7.13558495e-01 7.77540267e-01 6.13743603e-01 -4.61882278e-02
-1.13043353e-01 -2.20340848e-01 8.49442035e-02 1.06569421e+00
2.18275428e-01 -1.29012242e-01 -8.09789658e-01 6.74494147e-01
-2.26351213e+00 -8.36859584e-01 -1.12654403e-01 1.63271570e+00
5.35837293e-01 -3.33120108e-01 2.21740492e-02 -3.45710784e-01
8.42040658e-01 3.94748151e-01 -3.82068396e-01 -6.73278328e-03
1.72272339e-01 -2.68737525e-01 5.43113530e-01 1.61175221e-01
-1.80038118e+00 9.05403078e-01 5.64990950e+00 9.25149620e-01
-9.52749550e-01 5.46753526e-01 5.70221603e-01 -2.80068209e-03
-1.03130214e-01 -1.49956226e-01 -7.01525688e-01 5.67660093e-01
7.93192744e-01 3.53325844e-01 2.60761082e-01 7.61457741e-01
-2.63913453e-01 1.83017164e-01 -1.17912078e+00 1.24833405e+00
5.98513007e-01 -1.34434736e+00 -7.99866170e-02 1.17733292e-01
9.91497695e-01 9.75860581e-02 9.92386192e-02 3.92525882e-01
1.66001827e-01 -8.36918771e-01 4.88219261e-01 3.41968983e-01
1.00218797e+00 -6.32755518e-01 5.26256084e-01 -1.06572375e-01
-1.66542375e+00 -2.29306281e-01 -1.83144018e-01 2.45974526e-01
1.75307795e-01 1.98121235e-01 -3.39412957e-01 6.24113798e-01
5.52707195e-01 1.53164804e+00 -8.68740201e-01 7.15342760e-01
-1.03848740e-01 4.42008942e-01 3.36010933e-01 2.63595164e-01
5.11657119e-01 -2.12681994e-01 -5.82031831e-02 1.11779702e+00
7.61358067e-02 -7.52478689e-02 7.35481203e-01 3.93865138e-01
-3.22848380e-01 3.87232929e-01 -7.33818829e-01 -1.24126531e-01
2.85312861e-01 1.53229535e+00 -8.11825037e-01 -1.34690568e-01
-1.06747377e+00 1.23008764e+00 7.32149124e-01 5.25144696e-01
-1.08538687e+00 -7.66826123e-02 6.80404723e-01 -3.84862095e-01
2.70656615e-01 1.69064086e-02 1.49016544e-01 -1.19454837e+00
-2.33405769e-01 -5.52022517e-01 7.41075933e-01 -5.63862860e-01
-1.57722020e+00 2.50313044e-01 -3.06340992e-01 -1.29474580e+00
8.83793235e-02 -5.71110964e-01 -1.70298398e-01 3.65376532e-01
-1.63280618e+00 -1.86127377e+00 -3.13164175e-01 5.87216020e-01
6.11410797e-01 -2.43134663e-01 9.42092299e-01 8.04447651e-01
-6.25450134e-01 6.42272234e-01 2.64585227e-01 3.94395083e-01
8.33205819e-01 -9.41911578e-01 -2.51675963e-01 2.61604995e-01
5.48432589e-01 1.89631373e-01 -1.42120093e-01 -4.93970841e-01
-1.09100795e+00 -1.43738663e+00 7.24755704e-01 -4.43308324e-01
8.33381712e-01 -3.31836700e-01 -7.32553661e-01 9.93408263e-01
3.07230562e-01 5.37062585e-01 1.22339320e+00 1.07418001e-01
-9.20930803e-01 6.67714328e-02 -9.13836956e-01 2.50229388e-01
1.27031231e+00 -8.22167516e-01 -1.88390508e-01 7.16544092e-01
8.81826043e-01 -5.50807789e-02 -1.14323592e+00 4.80079055e-01
5.79180062e-01 -5.59687853e-01 9.57804978e-01 -8.31366479e-01
4.35350031e-01 -2.22565278e-01 -3.91158223e-01 -1.28347528e+00
-5.31319380e-01 2.66583651e-01 2.03946307e-01 1.36670935e+00
2.03991085e-01 -5.85080683e-01 4.69552010e-01 1.50069475e-01
-3.70557755e-02 -9.99853015e-01 -8.48960638e-01 -5.74784398e-01
-2.11480737e-01 -1.02608442e-01 6.36989474e-02 1.65854692e+00
-1.07279986e-01 4.92441982e-01 -4.45238471e-01 2.15155795e-01
7.97615409e-01 3.09769839e-01 2.02177972e-01 -1.27671540e+00
-8.04511830e-02 -2.08350047e-01 -8.21885765e-01 -8.03936303e-01
1.00420785e+00 -1.27193725e+00 -1.86262906e-01 -1.51971233e+00
7.32132316e-01 -6.17200136e-01 -9.53473926e-01 6.53951228e-01
-2.16191217e-01 7.54089534e-01 1.64463624e-01 3.34861755e-01
-1.56117618e+00 4.84728664e-01 1.02980804e+00 -3.92176688e-01
4.22736108e-01 -5.35617352e-01 -4.75012690e-01 5.40495038e-01
5.38704276e-01 -5.81281662e-01 -5.85561395e-01 -4.26400185e-01
2.08526418e-01 -1.24813773e-01 4.13484722e-01 -8.17142963e-01
7.62768611e-02 -2.87613183e-01 -1.28915096e-02 -3.64004701e-01
2.14877516e-01 -9.14128900e-01 -6.68131113e-02 1.17171109e-01
-6.47467792e-01 2.09323708e-02 -3.59952040e-02 1.07889938e+00
-4.07587200e-01 -1.99049652e-01 7.26769388e-01 4.47412170e-02
-1.33165193e+00 7.17863977e-01 -6.09278679e-02 2.26450443e-01
1.47786164e+00 1.72291175e-01 -3.69383752e-01 7.11849891e-03
-7.22615778e-01 6.62137449e-01 4.33335871e-01 8.02650392e-01
6.36872888e-01 -2.06660843e+00 -3.82163554e-01 1.00446321e-01
8.51902246e-01 -5.14260709e-01 6.01871610e-01 9.44937229e-01
-2.22469494e-01 4.48835135e-01 -7.90724009e-02 -7.49105096e-01
-1.57292354e+00 8.03527594e-01 1.32390082e-01 -2.55955368e-01
-3.67116570e-01 9.09316182e-01 5.73890686e-01 -4.52322364e-01
1.22688070e-01 1.60423428e-01 -4.92735386e-01 3.04498792e-01
4.77120548e-01 1.06653295e-01 -3.82713020e-01 -1.00877333e+00
-5.74346840e-01 1.20285928e+00 -1.23229720e-01 2.25056171e-01
1.18637955e+00 -4.56440091e-01 -2.94091910e-01 7.24205792e-01
1.99759293e+00 -4.14520562e-01 -1.18636227e+00 -4.51556116e-01
1.09780319e-01 -6.32434607e-01 -1.83102991e-02 -3.03090155e-01
-1.42263460e+00 9.08108234e-01 9.04597342e-01 7.26792663e-02
9.86518800e-01 3.43608111e-01 7.45640755e-01 -1.51053630e-02
3.57759416e-01 -1.09074187e+00 6.36054933e-01 2.60381669e-01
4.70504850e-01 -1.75451696e+00 -3.06501627e-01 -5.52712917e-01
-6.52964234e-01 9.19550657e-01 7.24969566e-01 -1.80127118e-02
7.65612900e-01 -5.40061176e-01 3.58885050e-01 -5.40178001e-01
-5.68108082e-01 -2.28720680e-01 4.06463027e-01 3.07452530e-01
4.88288134e-01 -2.62176394e-02 -3.79192054e-01 1.63916469e-01
9.90484536e-01 -1.66623384e-01 3.32506560e-02 7.76268244e-01
-2.71025747e-01 -1.21258593e+00 6.12145737e-02 4.33293402e-01
-5.53250730e-01 3.62645760e-02 8.94562453e-02 5.29095352e-01
1.65263548e-01 9.56490755e-01 -4.91157584e-02 -5.79259157e-01
-2.58357916e-02 2.12650970e-01 4.50592548e-01 -7.17135906e-01
-2.45151356e-01 -4.65772338e-02 -2.17853617e-02 -8.27864707e-01
-8.85665476e-01 -6.19019210e-01 -1.22244847e+00 1.44189581e-01
-4.19441968e-01 -4.35849018e-02 6.25122845e-01 8.33302438e-01
5.53813577e-01 6.42628193e-01 7.65442729e-01 -6.80277586e-01
-2.21229658e-01 -7.83552706e-01 -7.66438484e-01 1.11963296e+00
6.97257929e-03 -1.04746199e+00 -2.94771224e-01 8.10102448e-02] | [9.737200736999512, 3.963042974472046] |
fe4b795b-bece-4175-8df5-27df8b1ac15b | vehicle-trajectory-prediction-works-but-not | 2112.03909 | null | https://arxiv.org/abs/2112.03909v2 | https://arxiv.org/pdf/2112.03909v2.pdf | Vehicle trajectory prediction works, but not everywhere | Vehicle trajectory prediction is nowadays a fundamental pillar of self-driving cars. Both the industry and research communities have acknowledged the need for such a pillar by providing public benchmarks. While state-of-the-art methods are impressive, i.e., they have no off-road prediction, their generalization to cities outside of the benchmark remains unexplored. In this work, we show that those methods do not generalize to new scenes. We present a method that automatically generates realistic scenes causing state-of-the-art models to go off-road. We frame the problem through the lens of adversarial scene generation. The method is a simple yet effective generative model based on atomic scene generation functions along with physical constraints. Our experiments show that more than 60% of existing scenes from the current benchmarks can be modified in a way to make prediction methods fail (i.e., predicting off-road). We further show that the generated scenes (i) are realistic since they do exist in the real world, and (ii) can be used to make existing models more robust, yielding 30-40 reductions in the off-road rate. The code is available online: https://s-attack.github.io/. | ['Amir-Hossein Shahidzadeh', 'Alexandre Alahi', 'Seyed-Mohsen Moosavi-Dezfooli', 'Mohammad Shaverdikondori', 'Ahmad Rahimi', 'Saeed Saadatnejad', 'Mohammadhossein Bahari'] | 2021-12-07 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Bahari_Vehicle_Trajectory_Prediction_Works_but_Not_Everywhere_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Bahari_Vehicle_Trajectory_Prediction_Works_but_Not_Everywhere_CVPR_2022_paper.pdf | cvpr-2022-1 | ['scene-generation'] | ['computer-vision'] | [ 2.06948504e-01 2.89836287e-01 -1.23300612e-01 -2.47797564e-01
-7.89732993e-01 -5.88491261e-01 9.12920415e-01 -7.69398570e-01
5.03765307e-02 6.71696067e-01 6.34667575e-02 -7.43470550e-01
3.63181055e-01 -1.17155027e+00 -1.06880665e+00 -5.24908364e-01
1.75534785e-01 2.94018835e-01 6.77349925e-01 -7.04185367e-01
-2.33599730e-02 5.40050149e-01 -1.68464971e+00 2.01628000e-01
9.12330389e-01 5.08736849e-01 -1.33166891e-02 7.63462484e-01
3.22777539e-01 7.29548573e-01 -4.18185264e-01 -8.17673504e-01
5.60599566e-01 -3.07198793e-01 -5.47073305e-01 3.65227670e-03
5.60843408e-01 -4.53148991e-01 -9.78315830e-01 8.49853337e-01
2.61785388e-01 1.57344744e-01 4.45797533e-01 -1.75688112e+00
-6.03438437e-01 4.02626470e-02 -1.51309982e-01 -1.03788041e-01
1.94741055e-01 6.67574942e-01 6.47011161e-01 -7.41380990e-01
7.06073761e-01 1.15711641e+00 6.71084821e-01 6.95517302e-01
-1.13089883e+00 -9.25070643e-01 1.33907974e-01 3.56033921e-01
-1.51024473e+00 -6.91507399e-01 8.28520417e-01 -4.38927144e-01
7.59591460e-01 5.31269908e-01 6.09117568e-01 1.37643039e+00
5.22192061e-01 5.93863010e-01 7.13740706e-01 6.62260279e-02
1.24915972e-01 2.62745798e-01 -2.95414865e-01 4.07236367e-01
2.43346125e-01 7.90655196e-01 -1.61926538e-01 1.24808513e-01
3.02220553e-01 -1.57528758e-01 1.71125546e-01 -4.01012242e-01
-1.14420462e+00 8.96911919e-01 3.49946588e-01 -2.22305581e-01
-4.37262245e-02 4.52890515e-01 1.59649789e-01 -9.19756368e-02
4.02083308e-01 2.75210321e-01 -1.16208959e-02 -1.17853314e-01
-8.32241952e-01 7.94968069e-01 5.59004128e-01 1.28829730e+00
9.24748302e-01 2.76316851e-01 -8.75662267e-02 4.77329791e-01
2.41940543e-02 8.36778462e-01 -1.80442892e-02 -1.25672805e+00
6.05376542e-01 4.06443560e-03 3.38507205e-01 -1.11027515e+00
-1.80700764e-01 -9.38055739e-02 -5.65447807e-01 3.54816914e-01
2.17941716e-01 -3.70155871e-01 -8.96660149e-01 1.66458464e+00
2.64824957e-01 7.74697781e-01 1.39153570e-01 5.39535105e-01
7.51601577e-01 9.86894488e-01 -7.10371882e-02 2.77262151e-01
7.77886450e-01 -1.27576244e+00 -4.99039769e-01 -4.62084979e-01
7.17331111e-01 -8.61156762e-01 1.03708899e+00 1.98104739e-01
-9.60502326e-01 -7.45737731e-01 -9.58638608e-01 3.19196612e-01
-5.62338412e-01 -3.61849159e-01 7.11215734e-01 1.02215457e+00
-1.12303078e+00 3.62719506e-01 -5.82750380e-01 -1.90774813e-01
4.58754778e-01 -6.75648591e-03 -2.07471281e-01 -2.99585819e-01
-1.46343291e+00 9.81558919e-01 1.26281023e-01 -1.30175635e-01
-1.22269142e+00 -9.45463657e-01 -9.75717425e-01 -4.01577294e-01
2.78206915e-01 -6.68249190e-01 1.38864446e+00 -6.58454359e-01
-1.38415289e+00 4.78203386e-01 -3.08600426e-01 -5.02631068e-01
8.81066084e-01 -2.03830943e-01 -8.57609689e-01 -6.59709498e-02
2.29325965e-01 9.82323349e-01 7.15540051e-01 -1.69620633e+00
-8.16088140e-01 2.88910449e-01 1.98614612e-01 -7.36543685e-02
1.13529027e-01 -1.92055926e-01 -5.63455939e-01 -5.77529967e-01
-4.05229300e-01 -1.48685634e+00 -5.37924469e-01 -1.66976035e-01
-7.95964301e-01 1.85877442e-01 1.24603152e+00 -4.62269604e-01
9.54659104e-01 -2.02909589e+00 -7.10656285e-01 1.34018123e-01
-1.40882237e-02 3.19236010e-01 -2.62440056e-01 5.25081754e-01
-7.74456710e-02 5.00723124e-01 -1.42052978e-01 -3.42651844e-01
4.31507640e-02 2.34902784e-01 -7.80925155e-01 4.09098268e-01
2.77501404e-01 1.12180877e+00 -9.56671894e-01 -1.96187228e-01
6.29314780e-01 4.57777739e-01 -7.33489990e-01 -8.01106915e-02
-2.19500527e-01 6.53045535e-01 -3.61342281e-01 3.62981796e-01
9.37669694e-01 2.58509099e-01 -3.48668218e-01 5.51843550e-03
-2.02796683e-01 3.40154111e-01 -9.94364083e-01 1.21487546e+00
-5.06220639e-01 9.95068312e-01 -5.48195481e-01 -6.49961531e-01
8.00463319e-01 9.66531634e-02 3.80578309e-01 -7.00220287e-01
-2.18435571e-01 -1.29436776e-02 1.74440593e-02 -4.20331836e-01
7.90656805e-01 -6.06305189e-02 -1.46052092e-01 1.79437295e-01
-6.34532094e-01 -7.16645837e-01 2.04420298e-01 2.37123415e-01
1.04923797e+00 1.98750004e-01 -3.08263391e-01 -1.47328293e-02
2.03886926e-01 2.24052817e-01 4.81078357e-01 8.56748223e-01
-1.99374959e-01 7.85651505e-01 2.07048833e-01 -4.86373901e-01
-1.43287075e+00 -1.31820655e+00 -1.12717286e-01 4.88563299e-01
5.30200005e-01 -2.69304693e-01 -8.52382123e-01 -5.81313789e-01
5.59326075e-02 1.48013151e+00 -5.03128111e-01 -4.91098017e-01
-5.02640426e-01 -6.18248880e-01 8.24419558e-01 4.18945432e-01
5.15901923e-01 -8.24888468e-01 -3.88462543e-01 1.06709115e-01
-2.36362204e-01 -1.50878692e+00 -3.30098808e-01 -5.73931932e-01
-3.51415277e-01 -8.40446889e-01 -3.62341195e-01 -2.58560359e-01
5.02081513e-01 7.40263879e-01 1.20102549e+00 5.60091436e-03
-1.58920467e-01 2.55953521e-01 -1.74513832e-01 -5.00220597e-01
-9.97901738e-01 -7.90109262e-02 -5.78873791e-02 -6.59808293e-02
1.98236823e-01 -5.88657200e-01 -6.65669739e-01 7.66500592e-01
-7.19093144e-01 5.12394726e-01 3.20266575e-01 4.28828925e-01
5.85115910e-01 4.11545873e-01 6.84475780e-01 -1.06002474e+00
7.44435787e-02 -6.69296086e-01 -6.39741600e-01 -1.98560908e-01
-5.31006098e-01 -2.13813871e-01 7.00076401e-01 -4.00705785e-01
-1.11557937e+00 2.01037362e-01 -5.32049716e-01 -4.19826478e-01
-5.12164235e-01 -1.13773629e-01 -2.92484522e-01 -9.94390026e-02
7.08131075e-01 2.48756453e-01 -3.25668782e-01 7.36165568e-02
6.21742725e-01 3.51519823e-01 5.51923871e-01 -3.65698665e-01
1.49302912e+00 8.77020538e-01 1.98761284e-01 -5.95595896e-01
-7.19727576e-01 -1.22650504e-01 -2.78617054e-01 -5.18316031e-01
7.02501714e-01 -1.05004787e+00 -4.38373059e-01 4.23851758e-01
-9.73542511e-01 -7.37650990e-01 -2.88975686e-01 2.43754670e-01
-8.45182776e-01 2.33765975e-01 -2.22475320e-01 -7.19265401e-01
2.09802240e-01 -1.44885695e+00 1.07764959e+00 6.92323176e-03
-7.89604634e-02 -8.20138931e-01 -4.04922925e-02 5.43557882e-01
5.17961204e-01 4.53202248e-01 5.11272430e-01 -1.74962476e-01
-8.76092613e-01 -4.57756251e-01 -5.21289110e-02 2.78138727e-01
3.95171084e-02 4.90268469e-01 -1.10280514e+00 2.21753784e-04
-4.37029988e-01 1.43419102e-01 8.30361009e-01 1.82477966e-01
1.29043651e+00 -3.33930910e-01 -5.31648636e-01 6.44312203e-01
1.12683511e+00 3.33958268e-01 1.40950847e+00 2.84779847e-01
8.62716138e-01 5.26535690e-01 7.54012823e-01 1.75426081e-01
6.48627758e-01 1.08591628e+00 6.74486220e-01 -1.19285539e-01
-5.41755915e-01 -6.63256764e-01 4.71539795e-01 4.17782456e-01
4.50951941e-02 -4.03866172e-01 -1.11267960e+00 7.28380561e-01
-1.71624064e+00 -1.39497638e+00 -6.02903545e-01 2.25273180e+00
4.86503363e-01 5.64829588e-01 1.59764543e-01 2.35946309e-02
6.94714665e-01 1.87275395e-01 -5.76029718e-01 -4.30944502e-01
-6.03548810e-02 -5.21111153e-02 8.51006091e-01 5.87513566e-01
-1.12571990e+00 1.22065330e+00 6.63749933e+00 1.03231633e+00
-1.10827661e+00 1.87285364e-01 9.06323254e-01 2.76007410e-02
-6.93327308e-01 1.60932034e-01 -1.12808168e+00 6.54885769e-01
1.22760940e+00 -2.54346907e-01 4.44847018e-01 1.00727618e+00
5.81743300e-01 3.64603698e-02 -8.67439747e-01 6.78092182e-01
-3.36764902e-02 -1.57956994e+00 1.11419514e-01 2.97516406e-01
1.10133052e+00 1.66737854e-01 3.34810942e-01 5.80606699e-01
5.57125270e-01 -1.13139868e+00 9.82871950e-01 5.53648412e-01
9.31700706e-01 -9.57920372e-01 4.29084897e-01 4.74733740e-01
-1.20897651e+00 1.13648430e-01 -3.11914563e-01 8.43506381e-02
6.13322437e-01 5.25539696e-01 -8.03513944e-01 4.84242499e-01
4.77225542e-01 6.94530725e-01 -6.87887132e-01 1.02445853e+00
-2.70278126e-01 9.33906019e-01 -3.44883502e-01 3.02431285e-01
2.89716631e-01 -8.11036304e-02 7.03614712e-01 1.14329970e+00
5.78136802e-01 -2.71294862e-01 2.28088330e-02 1.02595973e+00
3.98135334e-02 -3.34013820e-01 -1.19595683e+00 3.68302107e-01
4.43367660e-01 8.98242474e-01 -3.32118094e-01 -1.65076002e-01
-5.65992773e-01 6.93891346e-01 -1.71017066e-01 5.09958386e-01
-1.53325832e+00 -1.19971372e-01 1.13654006e+00 5.68778634e-01
3.08355212e-01 -1.05956167e-01 -3.43815327e-01 -9.02666807e-01
-1.55292645e-01 -6.97848439e-01 -3.18324506e-01 -9.44725633e-01
-1.00180030e+00 4.84112948e-01 8.30858946e-02 -1.47424316e+00
-4.08178806e-01 -4.03302997e-01 -8.14683497e-01 5.67542672e-01
-1.62049305e+00 -1.30180144e+00 -4.62356508e-01 5.83567500e-01
6.37935936e-01 -1.02978103e-01 5.70542932e-01 5.14322817e-01
-6.27462387e-01 7.31076062e-01 7.75406808e-02 -8.29889774e-02
5.78650355e-01 -7.46993363e-01 1.25550032e+00 1.20109534e+00
1.10018581e-01 6.93043619e-02 9.76867557e-01 -6.41045749e-01
-1.30131030e+00 -1.85580456e+00 7.17048764e-01 -8.02907348e-01
6.81077898e-01 -4.68049854e-01 -4.85788852e-01 7.38194823e-01
-1.45235538e-01 1.14674129e-01 3.03067595e-01 -3.25324953e-01
-3.07489216e-01 -1.93015128e-01 -1.00974166e+00 1.16887009e+00
1.25804245e+00 -3.14849615e-01 -4.36512800e-03 4.48593318e-01
9.46096897e-01 -4.76188630e-01 -3.81957680e-01 5.51557004e-01
3.62355262e-01 -1.07378983e+00 1.19855237e+00 -5.09501159e-01
6.51498795e-01 -4.44137812e-01 -2.54572690e-01 -1.24597406e+00
-2.95116037e-01 -9.14941907e-01 -1.78653225e-01 1.00061548e+00
5.76861322e-01 -8.75525415e-01 7.85270870e-01 6.83811605e-01
-5.51057935e-01 -6.96881473e-01 -9.30771530e-01 -1.15996420e+00
3.76135975e-01 -1.04608881e+00 8.85255337e-01 6.01261914e-01
-5.77939391e-01 8.41591284e-02 -6.30813062e-01 2.65671194e-01
5.69048882e-01 -1.35804713e-01 1.41207051e+00 -8.22094798e-01
-1.23165801e-01 -4.70096886e-01 -6.05885744e-01 -1.10379958e+00
3.63828450e-01 -6.52629733e-01 1.53013900e-01 -1.45591712e+00
2.96790842e-02 -7.59958982e-01 1.50943205e-01 2.66900152e-01
-2.39948377e-01 5.48508406e-01 3.85151595e-01 5.89457937e-02
-4.05406654e-01 5.08533359e-01 1.32502985e+00 -1.63580254e-01
9.25653800e-02 4.10728663e-01 -7.72282183e-01 6.67566121e-01
1.31005669e+00 -4.50564861e-01 -7.15327024e-01 -1.43321440e-01
1.98599324e-01 -2.20343336e-01 8.43999445e-01 -1.34386921e+00
-1.40388146e-01 -4.33074951e-01 -5.55175431e-02 -6.90764546e-01
6.12962544e-01 -6.95850492e-01 6.30752504e-01 4.32016015e-01
-4.98140370e-03 -1.36423588e-01 4.92935210e-01 5.35801232e-01
1.09351531e-01 -8.51536356e-03 7.98334718e-01 2.30357602e-01
-8.48197401e-01 5.17656863e-01 -4.43753481e-01 2.08058327e-01
1.32354665e+00 -3.08805078e-01 -6.57502949e-01 -8.90402555e-01
-3.20368201e-01 1.62820034e-02 6.60377026e-01 6.93994641e-01
4.82974678e-01 -1.43096149e+00 -9.29925442e-01 4.43805046e-02
2.28376657e-01 -1.28445655e-01 4.23608482e-01 4.64646012e-01
-5.61892092e-01 4.37073588e-01 1.22837022e-01 -4.72780108e-01
-8.71311724e-01 7.62039602e-01 3.94059420e-01 -1.11068435e-01
-8.13367605e-01 4.68872845e-01 6.39195859e-01 -4.89551842e-01
-3.12729478e-01 9.42966491e-02 3.18182558e-01 -5.79133749e-01
4.99624938e-01 2.34418124e-01 6.46133423e-02 -1.14453709e+00
-1.76979333e-01 2.96090424e-01 2.20567867e-01 -4.94911000e-02
1.13570845e+00 -1.13360606e-01 5.41983008e-01 1.77756265e-01
1.02975810e+00 3.64313096e-01 -1.57574487e+00 2.11477950e-01
-6.60716832e-01 -7.11398542e-01 -1.09350726e-01 -5.62915444e-01
-1.10992479e+00 7.96303689e-01 4.44137990e-01 3.08859438e-01
7.64171362e-01 2.16115247e-02 1.29282653e+00 2.04814821e-01
5.20051360e-01 -8.88333321e-01 -1.85313985e-01 5.19493043e-01
8.64691556e-01 -1.13559628e+00 -2.36613363e-01 -8.92023981e-01
-8.17234457e-01 5.66144526e-01 6.54282570e-01 -2.20141202e-01
7.06762493e-01 4.37143743e-01 -1.19156912e-01 1.19169675e-01
-7.99104452e-01 -2.19571069e-01 6.75211251e-02 9.84521329e-01
-4.23886999e-02 3.80518168e-01 3.14787239e-01 4.26284879e-01
-6.65838182e-01 -1.36698067e-01 7.52453148e-01 4.79144305e-01
-1.93279192e-01 -9.69648540e-01 -3.55479836e-01 3.77886385e-01
-2.91280206e-02 -1.39888853e-01 3.76234129e-02 8.85843337e-01
2.80388176e-01 1.29035747e+00 -6.92300349e-02 -7.53636360e-01
4.91197258e-01 -2.27701634e-01 3.65356915e-02 -4.40897048e-01
-9.86089557e-02 -5.43035090e-01 4.70444739e-01 -7.63586342e-01
4.09361310e-02 -7.19493091e-01 -9.15189981e-01 -9.35212612e-01
-6.98751211e-02 -3.58339101e-01 5.10858476e-01 7.60051787e-01
5.34401536e-01 5.43278158e-01 9.39322948e-01 -1.17527390e+00
-2.03770563e-01 -4.31218356e-01 -1.13597669e-01 5.30136824e-01
2.19654843e-01 -7.47575104e-01 -3.94400537e-01 3.06594700e-01] | [6.358780860900879, 0.5161541700363159] |
c6fe0493-e6bb-465a-bda9-b4c90fbdebfe | computational-protein-design-with-deep | 1801.07130 | null | http://arxiv.org/abs/1801.07130v2 | http://arxiv.org/pdf/1801.07130v2.pdf | Computational Protein Design with Deep Learning Neural Networks | Computational protein design has a wide variety of applications. Despite its
remarkable success, designing a protein for a given structure and function is
still a challenging task. On the other hand, the number of solved protein
structures is rapidly increasing while the number of unique protein folds has
reached a steady number, suggesting more structural information is being
accumulated on each fold. Deep learning neural network is a powerful method to
learn such big data set and has shown superior performance in many machine
learning fields. In this study, we applied the deep learning neural network
approach to computational protein design for predicting the probability of 20
natural amino acids on each residue in a protein. A large set of protein
structures was collected and a multi-layer neural network was constructed. A
number of structural properties were extracted as input features and the best
network achieved an accuracy of 38.3%. Using the network output as residue type
restraints was able to improve the average sequence identity in designing three
natural proteins using Rosetta. Moreover, the predictions from our network show
~3% higher sequence identity than a previous method. Results from this study
may benefit further development of computational protein design methods. | [] | 2018-02-24 | null | null | null | null | ['protein-design'] | ['medical'] | [ 1.07027486e-01 -3.43137011e-02 -3.28535065e-02 -5.58048725e-01
-2.71281183e-01 -3.07056487e-01 -8.40420350e-02 9.36130583e-02
-5.22233129e-01 1.23738563e+00 8.99201911e-03 -4.91187423e-01
1.57093868e-01 -7.44357586e-01 -8.92197847e-01 -1.06627715e+00
1.09388568e-02 4.16382849e-01 9.69989449e-02 -3.47359329e-01
1.52086869e-01 7.07068443e-01 -1.30087352e+00 5.05429804e-01
8.75428200e-01 8.05101931e-01 7.51653373e-01 1.87897503e-01
-1.34781778e-01 5.11555374e-01 -4.55793798e-01 -1.41066208e-01
1.41717970e-01 -4.80754703e-01 -8.08644176e-01 -2.01607287e-01
-8.47356170e-02 5.25575615e-02 1.19548142e-01 7.00621188e-01
8.64469647e-01 -6.64947331e-02 5.67394972e-01 -2.54473209e-01
-8.69870245e-01 2.32988015e-01 -3.35083276e-01 -1.91017553e-01
1.43883616e-01 4.26380545e-01 1.10074317e+00 -9.76137161e-01
6.79415345e-01 8.44044030e-01 5.24719417e-01 5.23709357e-01
-1.33977079e+00 -5.64096510e-01 -3.02457273e-01 5.45089722e-01
-1.21622896e+00 1.04198404e-01 5.53740263e-01 -5.59055507e-01
1.50080633e+00 -1.54154509e-01 7.87528217e-01 7.56160796e-01
4.91564572e-01 3.62572163e-01 8.39074492e-01 -2.57149488e-01
4.05697942e-01 -1.94730952e-01 2.24750992e-02 6.43305182e-01
2.89017092e-02 -1.77593883e-02 -1.50815219e-01 -3.71014833e-01
3.66353452e-01 2.17073023e-01 -2.18954101e-01 -4.34865266e-01
-9.66384530e-01 1.02691150e+00 8.11816692e-01 3.33483458e-01
-7.13940144e-01 -2.65295655e-01 3.15808952e-01 1.38829246e-01
5.17169014e-02 8.78087759e-01 -1.09758699e+00 7.22876415e-02
-3.13950121e-01 4.04285759e-01 7.41692781e-01 2.52572656e-01
8.24225068e-01 -4.52021025e-02 3.57582361e-01 8.62722099e-01
-3.36883999e-02 2.30441824e-01 5.07759690e-01 -6.27588511e-01
-2.24089354e-01 8.36441517e-01 1.35575876e-01 -8.48805547e-01
-7.62279749e-01 -2.64238894e-01 -8.65585446e-01 5.02204776e-01
6.37724996e-01 -1.50340036e-01 -7.35594630e-01 1.73590934e+00
2.56472617e-01 -6.09942734e-01 1.58749342e-01 1.02527046e+00
6.00389719e-01 8.54689717e-01 1.71772897e-01 -3.42489481e-01
1.37481964e+00 -6.11454785e-01 -4.46148306e-01 1.33780256e-01
5.99580765e-01 -6.50127351e-01 7.80416369e-01 5.37053525e-01
-6.90591574e-01 -4.26773608e-01 -1.15081406e+00 -9.32006463e-02
-4.17572379e-01 1.48839444e-01 8.51795614e-01 3.85680020e-01
-5.96981227e-01 8.42923164e-01 -5.69693923e-01 -3.64427298e-01
3.62569630e-01 7.31112301e-01 -6.72749400e-01 7.73809925e-02
-1.19032133e+00 1.15796018e+00 8.86355519e-01 -7.73181692e-02
-4.19685543e-01 -5.46421707e-01 -3.62270325e-01 1.67151332e-01
1.48530185e-01 -5.96376002e-01 9.57427621e-01 -9.02522445e-01
-1.58152497e+00 7.37411678e-01 -2.63876796e-01 -4.56972629e-01
-1.66932881e-01 4.45313519e-03 -1.80449918e-01 -1.44154117e-01
-2.74613023e-01 7.47398496e-01 2.53344834e-01 -6.97993994e-01
-4.39939857e-01 -4.72226024e-01 -1.48745447e-01 -5.95825873e-02
-3.26732583e-02 2.11945385e-01 2.71342963e-01 -3.86652440e-01
6.08453378e-02 -9.26210284e-01 -5.72418332e-01 -1.12488912e-02
7.93111417e-03 -4.58156228e-01 3.43436450e-01 -5.86256146e-01
8.88675153e-01 -1.66741240e+00 6.26882851e-01 1.24581955e-01
2.59322584e-01 6.01267278e-01 3.37950699e-02 6.81821108e-01
-3.92654270e-01 -1.12362690e-01 -1.89152598e-01 7.16069818e-01
-4.78057712e-01 9.53298211e-02 9.85410586e-02 3.41188192e-01
3.27822238e-01 8.66961896e-01 -5.91372430e-01 1.62956506e-01
1.06038615e-01 6.72792435e-01 -6.25430882e-01 3.94980073e-01
-6.79705918e-01 5.80828130e-01 -4.07929599e-01 3.91421765e-01
6.97674155e-01 -7.02939093e-01 5.78088462e-01 -2.50664026e-01
-1.06413685e-01 2.47511283e-01 -4.87899065e-01 1.31113684e+00
-1.19166439e-02 2.85650343e-01 -4.28281516e-01 -1.06741536e+00
1.23932314e+00 3.14859807e-01 7.08566785e-01 -6.69189751e-01
1.84448272e-01 2.17551783e-01 8.58403563e-01 -6.40437245e-01
7.96143636e-02 -4.02202308e-01 1.79454759e-01 4.86825019e-01
-1.19634576e-01 4.02789265e-01 9.09801573e-02 -2.92342931e-01
1.04387033e+00 2.29447559e-01 5.66989243e-01 -2.45298728e-01
6.67487323e-01 3.20463270e-01 9.12526548e-01 2.07818151e-02
-1.48027483e-02 5.13061106e-01 6.29795551e-01 -1.39087510e+00
-1.58232939e+00 -3.97791713e-01 -1.85971871e-01 1.22316766e+00
-2.96301484e-01 -4.21578974e-01 -8.51260364e-01 -2.92239726e-01
-1.91097930e-02 3.37900043e-01 -4.54164952e-01 -1.88727289e-01
-7.04681814e-01 -1.15888548e+00 2.12385237e-01 2.22949535e-01
1.31333932e-01 -1.65850377e+00 -5.58123529e-01 6.00313962e-01
5.59515841e-02 -7.10306704e-01 -9.63284895e-02 5.75454533e-01
-6.46007299e-01 -1.30498242e+00 -7.81084895e-01 -1.05986512e+00
5.48597157e-01 1.51121810e-01 8.77009988e-01 -8.43078718e-02
-3.83982956e-01 -9.08549309e-01 -1.65777251e-01 -5.21190047e-01
-5.54310024e-01 2.24038213e-01 2.68661261e-01 -2.64878750e-01
9.79525387e-01 -6.91947877e-01 -6.17326081e-01 4.96908665e-01
-6.49403036e-01 3.17028850e-01 8.33066165e-01 1.09689474e+00
6.74446166e-01 -1.13037124e-01 1.00190163e+00 -6.70246422e-01
6.66455805e-01 -4.06771183e-01 -5.43394208e-01 1.98036611e-01
-5.60721755e-01 5.96735775e-01 8.86888742e-01 -3.71499538e-01
-7.75834799e-01 5.02219200e-01 -6.85493231e-01 2.10657284e-01
-2.60585308e-01 6.34249389e-01 -2.79237896e-01 2.70079612e-03
9.03019965e-01 3.26380998e-01 3.99386704e-01 -6.84622705e-01
3.52564082e-02 5.38349628e-01 -3.13455239e-02 -5.13870656e-01
1.68094546e-01 -1.05760917e-01 3.25496569e-02 -9.52291727e-01
-6.27992094e-01 -1.39110461e-01 -7.14089215e-01 1.10377200e-01
9.51302171e-01 -6.20983660e-01 -1.41259384e+00 5.05575716e-01
-1.28979087e+00 -6.33838028e-02 2.96857119e-01 4.54812467e-01
-4.00070786e-01 4.40519422e-01 -4.27808017e-01 -3.38343114e-01
-5.94437838e-01 -1.55897141e+00 6.07720435e-01 8.61997679e-02
-3.61353666e-01 -5.74206114e-01 1.45516306e-01 3.73331994e-01
4.22996342e-01 3.42960685e-01 1.35578656e+00 -7.01470315e-01
-4.95110422e-01 2.45866328e-02 -2.19901636e-01 2.38254294e-01
3.32212061e-01 -1.44998118e-01 -6.64583087e-01 -2.12538302e-01
-1.60451531e-01 -4.75448132e-01 7.30240166e-01 4.41036046e-01
1.11578918e+00 -1.48258835e-01 -1.52612522e-01 4.14874911e-01
1.35923409e+00 7.37327695e-01 6.32970870e-01 5.46964586e-01
6.53401494e-01 6.20347321e-01 3.61429453e-01 3.99034202e-01
-1.44803092e-01 7.88585901e-01 4.55942959e-01 -9.30874348e-02
4.79370534e-01 6.81417659e-02 1.29048899e-01 3.36549401e-01
-3.69695485e-01 -3.48703153e-02 -1.00662255e+00 -1.28890909e-02
-1.81041908e+00 -9.25344050e-01 -9.97335985e-02 1.98734748e+00
1.22893977e+00 -2.48990078e-02 2.31315851e-01 -1.03598073e-01
5.91795266e-01 -1.88236147e-01 -1.03874934e+00 -5.44341147e-01
-3.90386701e-01 3.35271686e-01 4.25399542e-01 1.28034696e-01
-7.58011460e-01 9.09983337e-01 6.69810677e+00 5.10821879e-01
-1.45543551e+00 -4.27258611e-01 6.97965384e-01 -1.88909415e-02
7.07973838e-02 -1.77792236e-01 -7.27826834e-01 4.81556624e-01
1.03631318e+00 -1.45053893e-01 3.70904952e-01 9.68578398e-01
4.40804511e-01 8.26471746e-02 -7.31632471e-01 7.14168608e-01
-4.77007836e-01 -1.81741524e+00 -4.85347658e-02 3.62466335e-01
4.99289215e-01 1.21612445e-01 -2.33954281e-01 -1.58037525e-02
2.48139009e-01 -1.31919158e+00 4.81778979e-02 4.06931013e-01
5.71314871e-01 -1.04060638e+00 7.63695121e-01 5.65863848e-01
-6.47553682e-01 5.42087853e-02 -7.24038541e-01 -4.04229522e-01
5.47223166e-02 6.28477812e-01 -9.99950111e-01 8.20501298e-02
5.93232691e-01 5.50707042e-01 -2.53625304e-01 8.03366423e-01
-4.40114737e-02 4.16559339e-01 -2.18491241e-01 -5.70252001e-01
1.24523841e-01 -5.54138064e-01 2.16634013e-02 8.58857810e-01
-9.23441947e-02 3.43627989e-01 2.04825953e-01 7.55187035e-01
-1.55897498e-01 3.99776638e-01 -3.18899065e-01 -2.29625627e-01
-2.21988745e-02 1.08168471e+00 -4.08403248e-01 -1.18721779e-02
-4.25804377e-01 6.41218543e-01 6.89517975e-01 1.40685275e-01
-7.33728230e-01 -4.42361236e-01 1.11212361e+00 3.07295006e-03
4.04663086e-01 -1.63027585e-01 -3.52932923e-02 -8.07841718e-01
-1.01525970e-02 -1.12686908e+00 5.31933904e-02 -6.75278068e-01
-1.15643394e+00 5.76140940e-01 -6.03602111e-01 -7.88961828e-01
-6.37393892e-02 -1.06096888e+00 -2.74764776e-01 1.17372501e+00
-1.03153765e+00 -7.49383509e-01 1.76951602e-01 -1.08073987e-01
4.99328434e-01 -5.67153096e-01 1.12702036e+00 1.77987486e-01
-6.39172196e-01 3.96971375e-01 6.17037117e-01 -2.50780076e-01
7.18854725e-01 -1.10423517e+00 4.10197884e-01 2.47012600e-01
-4.81897980e-01 8.79995048e-01 8.50389302e-01 -6.61505044e-01
-1.45516038e+00 -9.04719710e-01 1.02379513e+00 6.32492360e-03
4.66745138e-01 -2.57148266e-01 -1.39991415e+00 2.90001988e-01
2.26144612e-01 -3.41372848e-01 1.09774196e+00 5.11158928e-02
-1.39114872e-01 1.50065988e-01 -1.17011058e+00 4.43290949e-01
7.09959686e-01 -1.19654223e-01 -3.75661224e-01 4.34702158e-01
8.04554522e-01 -2.15921640e-01 -1.12664413e+00 4.57327843e-01
8.06182742e-01 -1.08106399e+00 9.46869016e-01 -1.08278954e+00
5.02997816e-01 -4.04578954e-01 -6.11047680e-03 -1.12607419e+00
-7.97940016e-01 -2.55702555e-01 1.00645199e-02 4.43729341e-01
7.49288261e-01 -6.17665410e-01 1.11444545e+00 6.57396376e-01
-7.77412280e-02 -1.31868482e+00 -5.93524992e-01 -4.57044959e-01
3.68334472e-01 7.76084885e-02 6.26020014e-01 6.13148153e-01
1.74045056e-01 7.45641887e-01 -6.27964199e-01 -2.65046030e-01
2.16065213e-01 2.20908508e-01 5.86219132e-01 -1.33667552e+00
-3.13305050e-01 -2.40537032e-01 -3.81533444e-01 -9.38715637e-01
1.81246281e-01 -9.73640859e-01 -3.44879627e-01 -1.42546201e+00
3.30103934e-01 -1.52046993e-01 -3.63983572e-01 6.78318977e-01
-1.22783713e-01 -1.93400860e-01 -3.04562718e-01 -4.24796231e-02
-1.80280134e-01 5.76107860e-01 1.17245495e+00 -2.05017075e-01
-2.96743304e-01 8.44717175e-02 -7.59751916e-01 4.39957291e-01
1.15797460e+00 -2.54914701e-01 4.04393114e-02 6.18431158e-02
4.72929239e-01 -2.29439422e-01 -2.89418519e-01 -7.97828913e-01
-9.31693390e-02 -3.14036101e-01 7.11658120e-01 -6.35319591e-01
2.04721197e-01 -9.55324709e-01 5.27832270e-01 8.52841794e-01
-2.34620646e-01 -7.97320455e-02 1.90351516e-01 2.83612549e-01
8.77662897e-02 -4.45160083e-02 1.08017826e+00 -1.94523260e-01
-7.69297183e-01 1.76500067e-01 -6.66711748e-01 -5.18934071e-01
9.86818016e-01 -2.80377984e-01 6.04956672e-02 1.53239444e-01
-8.07976782e-01 -6.47594780e-02 6.42325997e-01 4.15504009e-01
7.24487782e-01 -1.09494781e+00 -5.27016699e-01 2.61641055e-01
2.65519738e-01 -1.50252387e-01 1.24966860e-01 3.25546950e-01
-9.50521708e-01 6.72329605e-01 -5.93707085e-01 -5.02023399e-01
-1.48889649e+00 8.36454928e-01 4.44240123e-01 -8.86337906e-02
-5.71233690e-01 5.05430579e-01 1.47740990e-01 -5.68180919e-01
-7.84603059e-02 2.53687613e-02 -3.14919233e-01 -2.99286872e-01
6.60397172e-01 9.66757983e-02 7.84505904e-02 -7.40221620e-01
-2.24719658e-01 5.70876658e-01 -5.04128873e-01 7.58276284e-01
1.93967414e+00 4.00124907e-01 -3.81178856e-01 5.65759316e-02
1.32623160e+00 -6.31829500e-01 -1.11743689e+00 -2.85857737e-01
2.66596615e-01 -9.79171246e-02 -3.65581930e-01 -9.54834640e-01
-8.03129733e-01 6.96353555e-01 8.12023520e-01 -2.07415640e-01
8.87716293e-01 2.61270218e-02 9.51797962e-01 1.00728059e+00
5.82424521e-01 -8.40755403e-01 -1.97186828e-01 7.63463020e-01
8.15724254e-01 -1.52103984e+00 2.75211241e-02 2.50520371e-03
-5.64513147e-01 1.27922213e+00 4.80421722e-01 -3.18003967e-02
2.99489796e-01 8.21469799e-02 8.59057251e-03 -3.28161418e-01
-7.90962577e-01 4.43741828e-02 -1.14659527e-02 5.23666084e-01
9.67224956e-01 6.57130852e-02 -6.70106411e-01 5.15733123e-01
-1.08976781e-01 1.47370592e-01 2.40038216e-01 8.39958549e-01
-9.38835084e-01 -1.54774714e+00 -2.28056282e-01 2.89688140e-01
-7.61026919e-01 -3.42254043e-02 -6.10403061e-01 3.00425649e-01
-8.06001844e-06 6.60674036e-01 -4.50447530e-01 -2.45813623e-01
2.80628413e-01 2.38518938e-01 4.17007506e-01 -4.34363335e-01
-6.57805562e-01 -1.65297285e-01 -7.01970607e-02 -1.58291414e-01
-2.85210907e-01 -3.22783977e-01 -1.61667264e+00 -3.61903071e-01
-3.84090215e-01 3.90673518e-01 7.98246145e-01 7.49548137e-01
7.68571913e-01 3.89596671e-01 5.52727759e-01 -8.00570309e-01
-3.66418928e-01 -9.49221015e-01 -4.95562702e-01 3.55424225e-01
1.04715284e-02 -4.75000113e-01 1.49617270e-01 1.74047232e-01] | [4.714326858520508, 5.665234565734863] |
b2b9a45f-585e-43cd-b208-bf662afbf0f6 | efficient-multi-task-rgb-d-scene-analysis-for | 2207.04526 | null | https://arxiv.org/abs/2207.04526v1 | https://arxiv.org/pdf/2207.04526v1.pdf | Efficient Multi-Task RGB-D Scene Analysis for Indoor Environments | Semantic scene understanding is essential for mobile agents acting in various environments. Although semantic segmentation already provides a lot of information, details about individual objects as well as the general scene are missing but required for many real-world applications. However, solving multiple tasks separately is expensive and cannot be accomplished in real time given limited computing and battery capabilities on a mobile platform. In this paper, we propose an efficient multi-task approach for RGB-D scene analysis~(EMSANet) that simultaneously performs semantic and instance segmentation~(panoptic segmentation), instance orientation estimation, and scene classification. We show that all tasks can be accomplished using a single neural network in real time on a mobile platform without diminishing performance - by contrast, the individual tasks are able to benefit from each other. In order to evaluate our multi-task approach, we extend the annotations of the common RGB-D indoor datasets NYUv2 and SUNRGB-D for instance segmentation and orientation estimation. To the best of our knowledge, we are the first to provide results in such a comprehensive multi-task setting for indoor scene analysis on NYUv2 and SUNRGB-D. | ['Horst-Michael Groß', 'Mona Köhler', 'Söhnke Benedikt Fischedick', 'Daniel Seichter'] | 2022-07-10 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 4.10763770e-01 -2.87196726e-01 3.67833138e-01 -6.12286210e-01
-6.12422824e-01 -7.71465063e-01 4.26196814e-01 1.67929575e-01
-6.35666251e-01 5.12651205e-01 -6.22866273e-01 -4.53769058e-01
-1.81490421e-01 -8.75652969e-01 -8.37467730e-01 -6.53436184e-01
1.44645989e-01 7.02413261e-01 6.03606999e-01 -2.64690638e-01
-1.69285442e-04 5.54549336e-01 -1.78633308e+00 1.20712683e-01
8.06914926e-01 1.22944701e+00 8.94520104e-01 1.08242774e+00
-5.42948246e-02 6.81019366e-01 -4.52876508e-01 1.66083783e-01
4.02976036e-01 1.49137229e-02 -8.57650697e-01 3.79808098e-01
5.22666574e-01 -3.79675090e-01 -7.48300701e-02 1.03832805e+00
4.70821261e-01 3.68811101e-01 3.56219918e-01 -1.31637609e+00
1.98055506e-01 -1.28563643e-01 -3.54388744e-01 5.17654009e-02
4.68246520e-01 2.21896067e-01 5.11222899e-01 -4.48370874e-01
4.78813082e-01 8.96909118e-01 6.90302491e-01 1.59243450e-01
-8.59021962e-01 -1.22161791e-01 5.50112903e-01 2.53378063e-01
-1.25439692e+00 -3.92709583e-01 8.61291349e-01 -2.23017082e-01
1.14168847e+00 3.49450916e-01 8.00853670e-01 7.52207279e-01
-2.47375026e-01 9.17524815e-01 1.37704873e+00 -1.61935985e-01
4.45749998e-01 1.38889760e-01 2.39704981e-01 7.44548440e-01
7.41202235e-02 -3.63734692e-01 -2.92276651e-01 4.47407454e-01
5.27172744e-01 2.21041217e-01 -1.88956454e-01 -4.75949675e-01
-1.29367423e+00 4.45011765e-01 6.91799045e-01 1.78390667e-01
-4.72945899e-01 3.64390194e-01 9.48303193e-02 2.14185655e-01
6.22893095e-01 6.56854361e-02 -7.04226017e-01 -3.11901778e-01
-7.19302475e-01 3.72874588e-01 8.50871146e-01 1.08094871e+00
9.99150455e-01 -3.17447931e-01 5.21951497e-01 5.62490463e-01
3.90090138e-01 8.90992999e-01 1.28080949e-01 -1.16402984e+00
5.57492852e-01 4.82574761e-01 3.29696596e-01 -6.56292319e-01
-9.10730124e-01 -1.09592453e-01 -5.34760356e-01 4.22196150e-01
6.57458663e-01 -2.23504961e-01 -1.14542329e+00 1.37252676e+00
5.12557626e-01 1.59008488e-01 -7.94668309e-03 9.21875179e-01
1.01674402e+00 6.22448146e-01 -4.93255779e-02 1.76633045e-01
1.32749605e+00 -1.07900167e+00 -5.07320106e-01 -7.83419073e-01
6.28410697e-01 -6.72494054e-01 1.16208708e+00 3.60185176e-01
-7.35820413e-01 -4.80766982e-01 -1.07822859e+00 -3.34394783e-01
-7.20974982e-01 2.76603159e-02 1.02094519e+00 7.92977989e-01
-1.17404044e+00 3.00172061e-01 -1.13542604e+00 -6.63234413e-01
4.37714875e-01 4.97036427e-01 -2.46353939e-01 -2.65184581e-01
-7.38414288e-01 8.41146886e-01 3.74921322e-01 2.16457322e-01
-8.78547370e-01 -3.04736644e-01 -8.22732627e-01 -3.30862403e-01
8.65143478e-01 -7.02568412e-01 1.54214966e+00 -7.06512451e-01
-1.65040696e+00 7.83657968e-01 -2.31009483e-01 -2.48329163e-01
5.93118250e-01 -2.88696021e-01 1.22356459e-01 1.20565422e-01
1.00849941e-01 7.51450598e-01 5.68206549e-01 -1.51500762e+00
-8.97886992e-01 -6.43280268e-01 5.65436542e-01 7.00006902e-01
1.07305191e-01 -3.71677548e-01 -8.11678469e-01 -6.15493543e-02
4.71787691e-01 -1.12615573e+00 -6.16624057e-01 -1.28433511e-01
-4.83640045e-01 1.31953537e-01 1.07294643e+00 -7.52842188e-01
5.71952999e-01 -2.00878382e+00 1.03355013e-01 1.98477075e-01
1.07587226e-01 9.40068066e-02 1.69292301e-01 -6.42133355e-02
3.67800385e-01 -1.83520660e-01 -3.95581186e-01 -7.61194110e-01
-2.39110105e-02 4.64110613e-01 2.05039650e-01 5.21121681e-01
-4.29707110e-01 8.50661993e-01 -8.45479250e-01 -3.82449061e-01
6.70442760e-01 4.43524718e-01 -4.71847355e-01 -6.16741590e-02
-5.79398930e-01 5.84358275e-01 -5.76495826e-01 7.81037629e-01
6.97754025e-01 -3.05184692e-01 2.26727471e-01 -2.09793597e-02
-1.82814777e-01 2.97738075e-01 -1.38813424e+00 2.19707727e+00
-7.02962101e-01 7.49776483e-01 3.09828073e-01 -1.16515219e+00
4.21410561e-01 7.17192739e-02 5.93819797e-01 -1.01044524e+00
1.17313869e-01 3.47682208e-01 -4.01394844e-01 -6.27084792e-01
9.07270610e-01 1.09919049e-01 -2.98089117e-01 4.04179186e-01
-2.60339618e-01 -7.16656148e-01 1.28196463e-01 -6.45883009e-02
8.99575889e-01 3.95166367e-01 2.00787246e-01 -2.31140200e-02
1.80527553e-01 4.40225005e-01 1.72739461e-01 8.93409014e-01
-2.85060227e-01 4.86341476e-01 7.56266564e-02 -4.11585718e-01
-9.38035488e-01 -9.71328378e-01 1.93621621e-01 1.11958420e+00
6.16304755e-01 -1.88150018e-01 -8.95712376e-01 -6.86451435e-01
-1.86561689e-01 5.66835523e-01 -2.72349864e-01 5.91443062e-01
-6.67232752e-01 -9.46919024e-01 1.54776707e-01 4.47890252e-01
8.40534270e-01 -9.28782225e-01 -1.20415497e+00 2.42509514e-01
-3.18424851e-01 -1.65494752e+00 1.91595763e-01 4.88143027e-01
-7.63315022e-01 -1.13933539e+00 -5.03407061e-01 -5.85486591e-01
5.10807931e-01 8.68724763e-01 1.14958847e+00 -1.40593827e-01
-2.39004090e-01 5.85207701e-01 -2.98283964e-01 -6.96086466e-01
7.35171735e-02 2.79515654e-01 -2.00039789e-01 -4.15841043e-01
-5.84029332e-02 -5.10161698e-01 -6.79222763e-01 3.74271482e-01
-6.25489533e-01 4.27832037e-01 3.14563572e-01 7.92435929e-02
6.92082286e-01 3.49470347e-01 -6.53200829e-03 -8.37231576e-01
1.04059584e-01 -2.99338281e-01 -7.44160831e-01 1.02784820e-01
-8.65337476e-02 -3.82930130e-01 2.91032344e-01 7.56593794e-02
-1.19599450e+00 2.60688543e-01 -3.19713354e-01 1.86393410e-02
-6.63219154e-01 3.82237613e-01 -3.90364468e-01 -2.18975782e-01
3.87103707e-01 1.00457132e-01 -3.05793434e-01 -5.63214719e-01
3.85643035e-01 7.72032499e-01 3.85229409e-01 -4.53509629e-01
5.16981840e-01 8.05133939e-01 9.43230912e-02 -1.14630949e+00
-9.03705418e-01 -6.49708092e-01 -7.84819603e-01 -3.60187978e-01
1.12313724e+00 -1.05463541e+00 -7.22708762e-01 9.62616324e-01
-1.00062132e+00 -1.01674414e+00 -8.13443959e-02 2.26388946e-01
-8.03292155e-01 2.65143633e-01 -1.67752638e-01 -8.24952602e-01
1.44166157e-01 -1.49886179e+00 1.38112998e+00 1.61246896e-01
2.58807689e-01 -9.72371995e-01 -2.60155350e-01 7.63292193e-01
3.39973271e-01 2.78332859e-01 4.16702747e-01 -3.05625677e-01
-1.07793009e+00 -4.99400347e-02 -4.26159620e-01 -2.32244711e-02
1.64836466e-01 -3.76157045e-01 -1.23468006e+00 -4.08190936e-02
1.26027033e-01 -2.74094015e-01 7.32909083e-01 5.71784198e-01
1.20313430e+00 2.23926142e-01 -3.73163760e-01 6.86250269e-01
1.47776735e+00 5.18689692e-01 3.52900803e-01 6.57942474e-01
1.09505987e+00 5.12874186e-01 8.34944606e-01 2.34685451e-01
8.78712595e-01 1.02441394e+00 7.24802911e-01 -3.21478903e-01
-1.89372346e-01 4.34644282e-01 5.89250252e-02 3.81291866e-01
-3.49630564e-01 -5.15155554e-01 -1.19824612e+00 3.52283806e-01
-1.99637222e+00 -6.33672833e-01 -3.71957928e-01 1.95637918e+00
2.73747623e-01 6.61743283e-02 9.92952064e-02 2.26017565e-01
2.81640261e-01 3.10705066e-01 -5.57713509e-01 1.42096858e-02
-4.76515740e-02 -7.88829178e-02 7.87953556e-01 6.71497226e-01
-1.35997689e+00 9.99579132e-01 5.84310436e+00 5.80045521e-01
-1.04228961e+00 3.73563707e-01 5.61483681e-01 -2.44935006e-01
-7.84616992e-02 -2.53489614e-01 -6.17155612e-01 2.67659307e-01
7.71326721e-01 4.64466721e-01 6.62572801e-01 1.03994179e+00
1.75218165e-01 -6.84215069e-01 -8.65782619e-01 1.13874686e+00
-6.02187738e-02 -1.01049948e+00 -4.46494550e-01 4.23566066e-02
6.09162390e-01 5.29416621e-01 -1.66898623e-01 3.91065814e-02
1.96396112e-01 -6.94619834e-01 8.25632751e-01 3.46945703e-01
4.69916433e-01 -6.20047152e-01 5.81638813e-01 5.67794025e-01
-1.26028562e+00 -9.45212599e-03 -3.54881138e-02 -1.50605872e-01
3.64388943e-01 6.36378944e-01 -7.87277281e-01 7.21986294e-01
9.40801919e-01 7.07460344e-01 -6.27886891e-01 9.26889539e-01
-1.99976131e-01 -5.52708507e-02 -7.56400526e-01 -1.60705209e-01
4.28903311e-01 -3.10214192e-01 4.15132195e-01 8.99354637e-01
4.36615974e-01 9.49079022e-02 5.36009371e-01 2.22467735e-01
2.20543623e-01 -2.68117964e-01 -5.81287324e-01 2.26681232e-01
1.74668804e-02 1.05301630e+00 -1.29588032e+00 -4.42686796e-01
-3.60554457e-01 1.21084082e+00 1.73781529e-01 5.18672585e-01
-8.68668914e-01 -1.52407855e-01 7.02563584e-01 5.49084158e-04
2.66072690e-01 -8.31787646e-01 -5.18097043e-01 -1.19159961e+00
5.76309301e-02 -5.21805763e-01 6.01912402e-02 -1.05986845e+00
-6.93569601e-01 4.83921796e-01 1.32676363e-01 -9.49828684e-01
-1.43842220e-01 -8.86073649e-01 -1.05070665e-01 5.48691750e-01
-1.71445608e+00 -1.29183197e+00 -8.06002259e-01 8.39129031e-01
8.21167946e-01 3.17950368e-01 7.25643933e-01 3.26342165e-01
-6.06073558e-01 -2.84245282e-01 8.19025934e-02 -2.98179597e-01
1.43448308e-01 -1.47141469e+00 3.93961787e-01 9.62241590e-01
6.06194288e-02 4.41881828e-02 6.49236798e-01 -5.43358922e-01
-1.53771341e+00 -9.06769156e-01 5.49345851e-01 -4.59871501e-01
3.22315782e-01 -5.45963407e-01 -5.20535350e-01 7.98338830e-01
-1.20461300e-01 -4.01764289e-02 3.81646335e-01 1.32345691e-01
2.13258877e-01 -1.58079401e-01 -9.96507227e-01 5.61609447e-01
1.35845625e+00 -5.02594650e-01 -1.39844328e-01 7.14355350e-01
6.75667644e-01 -9.25487161e-01 -5.40999234e-01 3.88188988e-01
3.15223992e-01 -1.25933540e+00 1.36342168e+00 5.05155511e-02
-2.70616859e-02 -5.91426909e-01 -5.61520934e-01 -1.11720097e+00
3.75285476e-01 -3.02712977e-01 2.43711267e-02 6.48586869e-01
2.69049257e-01 -6.00456357e-01 8.85323107e-01 6.64612472e-01
-4.53579128e-01 -3.28448147e-01 -1.06386113e+00 -5.93216479e-01
-6.47627354e-01 -1.27789414e+00 6.08014345e-01 6.15178227e-01
-5.85750461e-01 8.71058628e-02 -1.66505530e-01 5.02428770e-01
7.48432100e-01 5.12417853e-01 1.08676112e+00 -1.07996607e+00
-2.06759796e-01 -2.77703166e-01 -2.46761650e-01 -1.19092906e+00
4.05726247e-02 -4.65320677e-01 3.31702620e-01 -2.06751871e+00
-7.82225356e-02 -9.30280387e-01 3.33537124e-02 4.62991625e-01
-4.90058400e-02 4.21358913e-01 2.61644661e-01 1.87134221e-01
-1.02374768e+00 2.14631110e-01 1.15329826e+00 -6.60874322e-02
-4.13060009e-01 2.43843049e-01 -3.05676937e-01 1.08570373e+00
8.57391298e-01 -1.47319883e-01 -7.34945834e-01 -7.87836730e-01
3.31219763e-01 8.60885009e-02 6.45293057e-01 -1.23088038e+00
2.79273301e-01 -1.83928892e-01 2.28853598e-01 -7.63304591e-01
8.26968491e-01 -1.04353595e+00 1.72334418e-01 3.16877306e-01
4.41378772e-01 -1.68596506e-01 2.74729818e-01 4.24224526e-01
3.21881589e-03 -8.40422884e-02 4.24372554e-01 -5.02939343e-01
-1.38818228e+00 2.66708434e-01 -4.06492829e-01 -4.70159948e-01
1.12496328e+00 -5.96741498e-01 -3.53966713e-01 -2.65821755e-01
-7.75607407e-01 1.79042563e-01 6.77246094e-01 1.62980780e-01
5.87089360e-01 -7.85681784e-01 -1.60821423e-01 2.24482149e-01
3.31978016e-02 7.22221136e-01 4.85122144e-01 8.51957142e-01
-8.29030216e-01 4.88589764e-01 -1.54584780e-01 -9.10705447e-01
-1.31680024e+00 3.19190234e-01 4.28477228e-01 -2.01549977e-01
-5.81269741e-01 7.86495209e-01 1.67447150e-01 -7.00788617e-01
8.34372193e-02 -5.61378121e-01 -8.90391991e-02 1.32569447e-01
1.15575157e-01 3.00070822e-01 4.12598372e-01 -5.32777488e-01
-5.00874758e-01 6.50145948e-01 3.63707960e-01 -2.52999425e-01
1.53959703e+00 -6.68613732e-01 -1.30417198e-01 4.64968383e-01
9.46644723e-01 -3.78566504e-01 -1.48248577e+00 1.07677346e-02
-2.94119149e-01 -4.54001963e-01 2.47110009e-01 -7.98351884e-01
-9.58888710e-01 8.52302432e-01 6.89517081e-01 3.95344406e-01
1.34199131e+00 3.16800214e-02 7.28676379e-01 7.60770202e-01
8.82028282e-01 -1.14884388e+00 8.02977756e-02 5.89155376e-01
4.78880823e-01 -1.39738035e+00 1.15967639e-01 -7.85636723e-01
-4.83011395e-01 8.76031041e-01 4.80006754e-01 3.15321833e-01
5.58709025e-01 1.86814830e-01 2.09534287e-01 -4.25808519e-01
-1.09368652e-01 -6.53661966e-01 9.09195766e-02 7.28628039e-01
-3.61482613e-02 1.57861158e-01 5.21439314e-01 1.42462924e-02
-3.42338920e-01 -9.77944359e-02 3.96135598e-01 1.43141282e+00
-5.66495001e-01 -9.12608743e-01 -4.77388114e-01 2.14986473e-01
-9.72549096e-02 8.35403502e-02 9.69867036e-03 8.12822104e-01
2.79902995e-01 1.15422022e+00 1.14989027e-01 -1.95146054e-01
3.96131307e-01 -1.86793506e-01 6.01658523e-01 -4.82324898e-01
-3.16604286e-01 1.33717090e-01 4.34841007e-01 -8.36356580e-01
-8.08611572e-01 -7.76602626e-01 -1.27257049e+00 -3.38613153e-01
7.85302091e-03 -3.60207647e-01 1.41386461e+00 1.17818010e+00
1.15540013e-01 8.55164707e-01 4.03321058e-01 -1.51697767e+00
3.68097365e-01 -6.56005085e-01 -4.96864945e-01 2.56474942e-01
3.32196057e-01 -5.53508639e-01 2.17472874e-02 3.64216194e-02] | [8.396195411682129, -2.354335069656372] |
fb356699-c860-4ecd-9ca7-7456f7c99cc7 | learning-pixel-level-distinctions-for-video | 2204.04615 | null | https://arxiv.org/abs/2204.04615v1 | https://arxiv.org/pdf/2204.04615v1.pdf | Learning Pixel-Level Distinctions for Video Highlight Detection | The goal of video highlight detection is to select the most attractive segments from a long video to depict the most interesting parts of the video. Existing methods typically focus on modeling relationship between different video segments in order to learning a model that can assign highlight scores to these segments; however, these approaches do not explicitly consider the contextual dependency within individual segments. To this end, we propose to learn pixel-level distinctions to improve the video highlight detection. This pixel-level distinction indicates whether or not each pixel in one video belongs to an interesting section. The advantages of modeling such fine-level distinctions are two-fold. First, it allows us to exploit the temporal and spatial relations of the content in one video, since the distinction of a pixel in one frame is highly dependent on both the content before this frame and the content around this pixel in this frame. Second, learning the pixel-level distinction also gives a good explanation to the video highlight task regarding what contents in a highlight segment will be attractive to people. We design an encoder-decoder network to estimate the pixel-level distinction, in which we leverage the 3D convolutional neural networks to exploit the temporal context information, and further take advantage of the visual saliency to model the spatial distinction. State-of-the-art performance on three public benchmarks clearly validates the effectiveness of our framework for video highlight detection. | ['Lixin Duan', 'Wen Li', 'Yuning Jiang', 'Tiezheng Ge', 'Biao Wang', 'Fanyue Wei'] | 2022-04-10 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wei_Learning_Pixel-Level_Distinctions_for_Video_Highlight_Detection_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wei_Learning_Pixel-Level_Distinctions_for_Video_Highlight_Detection_CVPR_2022_paper.pdf | cvpr-2022-1 | ['highlight-detection'] | ['computer-vision'] | [ 7.57866725e-02 -3.13543051e-01 -5.14594913e-01 -2.95286626e-01
-1.97683126e-01 -2.66977876e-01 4.28198904e-01 3.92881900e-01
-2.28948966e-01 3.47890764e-01 5.75584888e-01 1.08057491e-01
1.62952155e-01 -7.30698287e-01 -8.47714484e-01 -5.73799908e-01
-3.93611014e-01 -5.76294482e-01 5.85881293e-01 -1.90671712e-01
3.81437838e-01 3.20407242e-01 -1.62750340e+00 6.08501732e-01
8.32071185e-01 1.28145194e+00 3.73952180e-01 3.66177380e-01
-1.25649333e-01 1.08091295e+00 -5.40072262e-01 -7.18208179e-02
2.08555147e-01 -6.10566318e-01 -4.82581556e-01 1.97248295e-01
6.87726259e-01 -6.65025651e-01 -5.85317671e-01 1.09040642e+00
1.77850097e-01 2.96604604e-01 4.44508314e-01 -1.20073771e+00
-3.75619739e-01 7.79256046e-01 -8.78524184e-01 8.27478766e-01
4.02603596e-01 1.55392483e-01 1.07979238e+00 -7.66494036e-01
6.77656949e-01 1.11640203e+00 3.97132188e-01 1.83121517e-01
-6.58996582e-01 -6.14731133e-01 6.45594060e-01 4.96349186e-01
-1.18429542e+00 -8.68027583e-02 1.24905193e+00 -5.61145008e-01
2.48950839e-01 1.76818341e-01 1.00628555e+00 9.15958881e-01
1.52699411e-01 1.27787983e+00 8.99158955e-01 -1.62042528e-01
8.51190761e-02 -3.17455977e-02 1.45861253e-01 5.76701999e-01
-1.36965603e-01 7.91632384e-02 -9.35154736e-01 3.01133811e-01
7.01479614e-01 4.89050269e-01 -4.09657657e-01 -3.09509665e-01
-1.18016517e+00 5.13699472e-01 8.79151523e-01 3.54889184e-01
-5.38188875e-01 2.62005866e-01 5.63568771e-01 -7.64369890e-02
4.97358620e-01 1.83476835e-01 -5.99251091e-02 -2.05617875e-01
-1.33694565e+00 2.46548042e-01 2.06274167e-01 6.94648504e-01
9.84829724e-01 -1.56233281e-01 -6.46002352e-01 4.96644199e-01
9.48146954e-02 2.18746305e-01 2.06294805e-01 -7.17750430e-01
4.93825823e-01 7.16058433e-01 -6.58569559e-02 -1.28286803e+00
-2.41669506e-01 -5.27243972e-01 -5.30880988e-01 3.25493991e-01
2.97820240e-01 -1.71199709e-01 -7.51726866e-01 1.69274592e+00
1.61920786e-01 5.37592530e-01 -3.02675366e-01 1.25061107e+00
9.92654562e-01 6.70353055e-01 2.87189573e-01 -8.93063247e-02
1.45237172e+00 -1.03384745e+00 -6.09314799e-01 -1.52924731e-01
3.91213894e-01 -6.08359516e-01 1.00956357e+00 -2.15144455e-02
-1.10214067e+00 -8.85924160e-01 -1.05415952e+00 -8.75412673e-02
-2.32967585e-01 1.61270678e-01 4.76139158e-01 1.40384570e-01
-7.23752320e-01 4.42425579e-01 -5.01215696e-01 -2.95591563e-01
4.75058645e-01 -8.96793157e-02 -1.04047000e-01 1.61641072e-02
-1.34210157e+00 5.08169889e-01 3.98474723e-01 4.17997725e-02
-9.41027761e-01 -8.21842790e-01 -8.12057972e-01 3.64758074e-01
5.94541430e-01 -4.57478017e-01 8.26961279e-01 -1.61147285e+00
-8.38378310e-01 6.82193160e-01 -4.18803334e-01 -5.91591060e-01
6.17775083e-01 -3.55113894e-01 -3.18040013e-01 5.03896534e-01
1.43787801e-01 9.96582568e-01 8.60009849e-01 -1.15967464e+00
-1.21669209e+00 4.13223654e-02 6.59860313e-01 3.22925836e-01
-4.33680594e-01 2.41845042e-01 -9.62513447e-01 -9.76976693e-01
-1.35950640e-01 -6.65515602e-01 4.40624841e-02 2.01836720e-01
-5.80631375e-01 -1.07670650e-01 1.07652414e+00 -7.17954516e-01
1.59274375e+00 -2.34489679e+00 -1.32319167e-01 5.56423217e-02
3.69979590e-01 2.20952243e-01 -4.67636762e-03 2.21789196e-01
5.74938655e-02 8.56020749e-02 9.65487137e-02 9.89220664e-02
-2.62596905e-01 -3.31847191e-01 -3.19756895e-01 1.70248032e-01
4.97374713e-01 8.37835312e-01 -1.07155216e+00 -5.91925681e-01
2.75688618e-01 6.33309722e-01 -5.95149994e-01 1.89904183e-01
-2.39963189e-01 4.18148965e-01 -6.29894793e-01 5.50149739e-01
6.58376038e-01 -7.45639130e-02 -1.00060008e-01 -5.19250214e-01
-4.23806220e-01 2.47336641e-01 -1.12165105e+00 1.40383554e+00
-4.15914841e-02 1.07487869e+00 -3.55548233e-01 -7.36500800e-01
6.99506462e-01 -1.14926904e-01 6.18382633e-01 -8.99074435e-01
-2.28855424e-02 -1.16779998e-01 -6.01283498e-02 -7.00294077e-01
7.20808685e-01 3.13653588e-01 1.12602569e-01 1.52315363e-01
-4.72033143e-01 5.33499002e-01 4.00004059e-01 2.89053619e-01
8.23769987e-01 3.51681501e-01 1.72083423e-01 -2.35509485e-01
5.65925479e-01 -3.35032851e-01 8.82561207e-01 6.06157780e-01
-5.31534016e-01 6.77339673e-01 8.44619870e-01 -6.48148954e-01
-6.55118823e-01 -8.62055182e-01 2.61837989e-01 1.37658310e+00
7.73096800e-01 -6.22627318e-01 -7.35441148e-01 -7.70841241e-01
-1.45499706e-02 4.48183328e-01 -1.03940904e+00 -2.01367915e-01
-8.24605525e-01 -1.55798435e-01 1.73095800e-02 5.54962695e-01
7.13578343e-01 -1.05796695e+00 -1.18512309e+00 -5.74412271e-02
-3.28471154e-01 -1.06025231e+00 -9.03183937e-01 -1.84907799e-03
-5.97406268e-01 -1.26753187e+00 -9.14020956e-01 -9.05296922e-01
5.69593012e-01 8.11143041e-01 1.09919977e+00 2.31705666e-01
-7.39254355e-02 1.60536885e-01 -4.90041614e-01 -2.32763901e-01
9.62786824e-02 -9.58020836e-02 -3.09753358e-01 3.07544708e-01
4.02795345e-01 -2.12996140e-01 -1.00392282e+00 3.31462711e-01
-9.33323026e-01 4.39685762e-01 5.18328071e-01 5.55003047e-01
6.58516109e-01 3.05047214e-01 1.10051863e-01 -7.85148144e-01
2.53270686e-01 -4.01464552e-01 -1.49457440e-01 2.89750248e-01
4.23155688e-02 -1.38890177e-01 3.97062600e-01 -4.32118058e-01
-9.16855574e-01 -2.61872145e-03 2.59064168e-01 -6.42159760e-01
-4.51914556e-02 5.05077600e-01 -2.22361639e-01 2.34409675e-01
2.46008903e-01 1.53867722e-01 -2.68373102e-01 -2.06355810e-01
3.56307298e-01 2.22947568e-01 4.70612347e-01 -3.83724779e-01
4.67330039e-01 5.55909157e-01 -4.44189757e-02 -6.51052177e-01
-9.77067113e-01 -6.01705015e-01 -6.77773058e-01 -6.27225637e-01
9.26098406e-01 -1.12746215e+00 -5.86801171e-01 2.09445164e-01
-9.22027588e-01 -3.35841745e-01 -1.43149242e-01 3.94020110e-01
-2.72812039e-01 4.51633334e-01 -3.60186636e-01 -6.28316939e-01
-1.21741831e-01 -1.32435000e+00 1.12056613e+00 6.79209292e-01
-1.59214422e-01 -8.74736726e-01 -4.36351180e-01 -1.25617862e-01
2.20832735e-01 4.42777961e-01 8.26607823e-01 -2.02247947e-01
-8.24507773e-01 2.01111790e-02 -5.75730860e-01 1.22919023e-01
1.26806051e-01 4.59618926e-01 -8.26087534e-01 -6.58860803e-02
-3.57591689e-01 8.19767937e-02 1.22261941e+00 8.32805455e-01
1.31719565e+00 -1.61065698e-01 -4.91232276e-01 5.87370038e-01
1.21267569e+00 1.18075415e-01 6.97207868e-01 5.74087918e-01
8.89083803e-01 7.72631645e-01 1.02716684e+00 5.13278723e-01
4.19410497e-01 9.35865581e-01 5.55902481e-01 -4.19459075e-01
-3.78940403e-01 -3.11600834e-01 6.51845753e-01 2.04876781e-01
-9.11573619e-02 -5.49601018e-02 -6.69771373e-01 5.59271991e-01
-1.91211879e+00 -1.34834290e+00 -1.27417132e-01 2.18343282e+00
5.85924089e-01 3.62028956e-01 4.84389246e-01 -1.64305523e-01
1.00152791e+00 5.90853274e-01 -5.15683949e-01 -8.79003759e-03
-3.17699939e-01 -3.22621375e-01 2.68236309e-01 1.50105521e-01
-1.41222310e+00 9.34027433e-01 5.55749130e+00 7.65173256e-01
-1.37543368e+00 -2.15479136e-01 1.15325320e+00 -3.21741939e-01
-2.47615039e-01 -5.37544396e-03 -7.15608478e-01 9.47531104e-01
4.73553449e-01 -1.27885908e-01 3.05230003e-02 7.12700188e-01
6.48148656e-01 -4.94920850e-01 -1.16186786e+00 9.97670293e-01
2.37803951e-01 -1.35950494e+00 1.13446094e-01 -1.63364127e-01
8.90961528e-01 -2.95370907e-01 3.30852985e-01 1.16316862e-01
-3.88281018e-01 -6.90106273e-01 1.12434232e+00 6.50703967e-01
3.92579257e-01 -9.16292965e-01 4.70609993e-01 -4.89677675e-02
-1.55401945e+00 -1.98860437e-01 -2.90930420e-01 -3.51791009e-02
4.28449884e-02 7.24145830e-01 -4.93792415e-01 2.17548147e-01
9.56551254e-01 1.16043901e+00 -6.58828318e-01 1.47723234e+00
-4.34793949e-01 5.08322120e-01 1.29992276e-01 1.28343880e-01
3.90830666e-01 -1.57360405e-01 5.30262053e-01 1.57130265e+00
2.66915739e-01 -1.46889791e-01 3.41653794e-01 9.17793274e-01
-1.31979644e-01 1.11594060e-02 -2.73991525e-01 2.03144774e-01
3.54727507e-01 1.09422970e+00 -8.38964105e-01 -4.19112921e-01
-5.50303459e-01 9.71303403e-01 2.42844552e-01 4.97671336e-01
-1.12804031e+00 -3.48857164e-01 7.61579812e-01 3.74018073e-01
6.10471487e-01 -1.11551531e-01 -2.07048044e-01 -9.66577768e-01
2.51644284e-01 -7.30856299e-01 3.69463682e-01 -7.74635732e-01
-8.54030073e-01 4.79157269e-01 -1.51158392e-01 -1.62688816e+00
1.95027106e-02 -2.27637231e-01 -1.01465249e+00 7.58998990e-01
-1.90938532e+00 -1.03053689e+00 -7.43455291e-01 3.13282460e-01
7.58717060e-01 2.39552140e-01 -2.60456949e-02 2.21708596e-01
-5.73124349e-01 5.06016076e-01 -1.96828038e-01 3.83995354e-01
8.41286242e-01 -1.09280646e+00 1.73213810e-01 1.21530247e+00
1.05410695e-01 5.51799655e-01 6.95097029e-01 -7.13293731e-01
-9.78937864e-01 -1.13797581e+00 7.83322871e-01 -1.85574144e-01
4.66331184e-01 -1.85710534e-01 -8.64981592e-01 2.79241830e-01
2.25387871e-01 9.01911408e-02 5.62622070e-01 1.06963716e-01
-4.14547712e-01 -1.71675369e-01 -6.65807068e-01 7.05920100e-01
1.00921810e+00 -7.09267080e-01 -4.96427536e-01 -1.84827689e-02
5.89109004e-01 -2.95799762e-01 -3.73031169e-01 3.19825917e-01
6.44584656e-01 -1.33209765e+00 1.04497886e+00 -4.07431334e-01
1.00557363e+00 -5.05719423e-01 2.17960253e-01 -1.12680984e+00
-3.38586003e-01 -5.31469882e-01 -2.76824623e-01 1.41872942e+00
-2.22572871e-02 1.12524197e-01 7.64800787e-01 6.16425931e-01
-9.12908614e-02 -8.76195133e-01 -6.06507897e-01 -4.07000840e-01
-3.99270922e-01 -3.00984263e-01 5.54064572e-01 9.10624087e-01
-2.51192898e-02 -4.52055261e-02 -6.10307097e-01 1.93648800e-01
2.77037799e-01 3.53395849e-01 7.34835505e-01 -1.03754914e+00
-1.75066981e-02 -8.09630752e-01 -5.93340397e-01 -1.45658481e+00
-1.29132560e-02 -5.03224909e-01 -2.05790173e-04 -1.50788581e+00
6.25531554e-01 -3.43481213e-01 -8.22310567e-01 3.15300733e-01
-8.43983829e-01 2.47905239e-01 5.54443896e-01 3.77148360e-01
-9.90653098e-01 4.24646020e-01 1.27204239e+00 -3.01648200e-01
-3.27659696e-01 -1.59331650e-01 -7.08197415e-01 7.18158543e-01
4.15433735e-01 -1.28289729e-01 -3.40778440e-01 -3.63884985e-01
4.39943075e-02 -2.46406019e-01 5.16597450e-01 -1.23565805e+00
2.00571969e-01 -3.31652254e-01 6.45984352e-01 -7.63279378e-01
9.13198143e-02 -7.76450038e-01 -1.97831050e-01 4.37549710e-01
-5.59743583e-01 2.27080658e-02 1.41433805e-01 6.14412367e-01
-5.19938171e-01 -6.20562099e-02 6.96922541e-01 7.03398064e-02
-1.45510805e+00 4.03772891e-01 -2.27334425e-01 5.01631387e-02
1.20981300e+00 -4.44600582e-01 -2.91061640e-01 -4.37612861e-01
-3.05118471e-01 2.80554980e-01 6.45826280e-01 6.35616422e-01
6.52451396e-01 -1.31295669e+00 -4.97967392e-01 8.89003426e-02
2.62316257e-01 -3.56432229e-01 6.19009376e-01 9.01167333e-01
-4.84023362e-01 1.69182867e-01 -5.56935847e-01 -5.99311650e-01
-1.44135976e+00 8.30242097e-01 2.08623260e-01 -3.91337015e-02
-7.31729925e-01 9.22991991e-01 9.06072140e-01 7.31479108e-01
3.95785093e-01 -6.63238287e-01 -6.58299923e-01 2.84492642e-01
8.87732208e-01 2.78198361e-01 -5.29160142e-01 -8.71573567e-01
-3.11868489e-01 6.44064426e-01 -1.06936485e-01 3.76095504e-01
1.00050259e+00 -3.45752954e-01 1.59214526e-01 4.19590175e-01
1.25424087e+00 2.14415640e-01 -1.94957066e+00 -4.11236823e-01
-8.46739560e-02 -8.41153681e-01 1.68749467e-01 -5.02974868e-01
-1.36014163e+00 9.90824521e-01 5.26243269e-01 1.07702881e-01
1.38724351e+00 -1.52354643e-01 8.33324015e-01 -1.42522156e-01
7.17283636e-02 -1.21083844e+00 3.37950230e-01 3.60657692e-01
7.24136472e-01 -1.33485711e+00 8.89542773e-02 -6.56988978e-01
-9.92696106e-01 1.36793470e+00 6.59343064e-01 -7.90250525e-02
4.14023250e-01 -1.44471526e-01 6.82251230e-02 -1.80621356e-01
-4.11011130e-01 -5.65898299e-01 8.09226692e-01 4.03216988e-01
6.68701470e-01 -1.25955315e-02 -1.35843650e-01 2.96177447e-01
2.38214120e-01 -9.99111831e-02 3.96819174e-01 6.18870378e-01
-6.96655989e-01 -6.71778798e-01 -1.21871218e-01 3.21906358e-01
-4.97976393e-01 -9.17984545e-02 -3.90251070e-01 5.63970864e-01
4.17226583e-01 8.39028239e-01 2.45297417e-01 -3.91649991e-01
2.54994661e-01 -3.23644698e-01 3.37899476e-02 -3.42619509e-01
-6.32905841e-01 3.16166103e-01 -7.90699795e-02 -6.65412307e-01
-7.00830340e-01 -6.73753023e-01 -1.35395122e+00 -2.33647406e-01
7.95328096e-02 7.06597641e-02 3.33365411e-01 7.16692746e-01
3.86803359e-01 8.66591811e-01 6.57336116e-01 -9.45370734e-01
2.35563844e-01 -5.63802302e-01 -5.82976818e-01 8.76989484e-01
5.06984234e-01 -8.50474000e-01 -7.12227896e-02 2.04200014e-01] | [10.042572021484375, 0.3743288516998291] |
39461cb6-fb18-4997-a607-cb7715a70889 | image-animation-with-keypoint-mask | 2112.10457 | null | https://arxiv.org/abs/2112.10457v1 | https://arxiv.org/pdf/2112.10457v1.pdf | Image Animation with Keypoint Mask | Motion transfer is the task of synthesizing future video frames of a single source image according to the motion from a given driving video. This task is challenging due to the complexity of motion representation and the unknown relations between the driving video and the source image. Despite this difficulty, this problem attracted great interests from researches at the recent years, with gradual improvements. The problem can be thought as decoupling of motion and appearance, which is often solved by extracting the motion from keypoint movement. We chose to tackle the generic, unsupervised setting, where we need to apply animation to any arbitrary object, without any domain specific model for the structure of the input. In this work, we extract the structure from a keypoint heatmap, without an explicit motion representation. Then, the structures from the image and the video are extracted to warp the image according to the video, by a deep generator. | ['Dov Gertz', 'Yanir Marmor', 'Or Toledano'] | 2021-12-20 | null | null | null | null | ['image-animation'] | ['computer-vision'] | [ 4.73451346e-01 1.24444909e-01 -5.97694470e-03 -3.07429638e-02
-2.99668521e-01 -6.26048684e-01 9.24591243e-01 -3.46788913e-01
-3.76770109e-01 5.72931111e-01 -4.62924509e-04 8.04812834e-02
5.48739098e-02 -5.73139548e-01 -9.07637954e-01 -1.03678739e+00
3.66750389e-01 2.06300601e-01 2.51572520e-01 -2.25558072e-01
2.27080524e-01 2.94457346e-01 -1.47687459e+00 -7.18673766e-02
6.17130280e-01 5.90472400e-01 5.59902489e-01 7.62627065e-01
-3.46612670e-02 7.73380518e-01 -4.71026123e-01 -1.88023552e-01
1.73907638e-01 -1.00058174e+00 -6.46093726e-01 3.83385748e-01
3.18354845e-01 -2.19907910e-01 -5.72936416e-01 1.16213179e+00
5.55075286e-03 3.75106990e-01 6.29186153e-01 -1.43163073e+00
-3.48652780e-01 2.57488370e-01 -5.12974083e-01 2.07047109e-02
2.40477726e-01 3.80580246e-01 7.66384065e-01 -4.62868840e-01
1.15446365e+00 9.85606313e-01 -4.80486453e-02 8.17764640e-01
-1.19566178e+00 -2.49290511e-01 2.24032924e-01 5.46083868e-01
-1.26165509e+00 -3.60257387e-01 1.12528467e+00 -7.38750398e-01
2.19261378e-01 7.91237801e-02 7.89388657e-01 1.32569623e+00
1.35127917e-01 5.32664359e-01 6.80673778e-01 -3.46142679e-01
3.60429436e-01 8.40477645e-02 -3.02472144e-01 4.48545963e-01
-7.94644356e-02 -2.28273179e-02 -2.72757828e-01 2.10742176e-01
8.66587579e-01 2.99518695e-04 -5.44788003e-01 -5.63438058e-01
-1.21115577e+00 5.81945777e-01 2.80503094e-01 5.20513475e-01
-4.36034948e-01 1.35726467e-01 -1.15090981e-01 1.53942853e-01
1.63676664e-01 1.96711898e-01 -2.03313097e-01 -9.51372609e-02
-1.14931965e+00 3.96827012e-01 7.43776381e-01 8.71813297e-01
8.94233465e-01 1.95351735e-01 7.21350908e-02 1.59675583e-01
2.51645327e-01 5.00198483e-01 5.46813726e-01 -9.40090656e-01
4.54144329e-01 3.15676957e-01 2.06215695e-01 -1.44242311e+00
-1.39274120e-01 -1.33982792e-01 -9.49432552e-01 3.27288538e-01
8.15492809e-01 -2.18095645e-01 -8.45588088e-01 1.97057009e+00
5.13071775e-01 5.44035912e-01 8.38336051e-02 1.03599596e+00
4.36157614e-01 9.67429340e-01 -1.21788800e-01 -3.33163112e-01
1.14828229e+00 -9.61091995e-01 -6.81082189e-01 -4.29621696e-01
4.08322841e-01 -6.31811321e-01 7.30344296e-01 2.97094166e-01
-1.00932574e+00 -7.79872954e-01 -1.03214681e+00 -1.41236633e-01
4.19550464e-02 1.42410979e-01 -8.66843462e-02 1.89312119e-02
-8.61291587e-01 4.63558674e-01 -7.66792476e-01 -3.16446066e-01
-1.08025424e-01 8.32585022e-02 -4.04766172e-01 -1.73656587e-02
-1.22546399e+00 6.98228061e-01 5.03732085e-01 3.55086744e-01
-9.53201354e-01 -3.36513162e-01 -9.05232489e-01 -2.45515816e-02
5.00894845e-01 -9.48463500e-01 9.24795508e-01 -1.67101765e+00
-1.60078514e+00 5.47511816e-01 -2.12614149e-01 -3.02875966e-01
8.65761220e-01 -1.01819232e-01 -9.62162465e-02 1.65724665e-01
-3.19673792e-02 5.69177985e-01 1.55971348e+00 -1.18653619e+00
-7.73421049e-01 -1.38999894e-01 2.07529604e-01 2.31082186e-01
-2.11694650e-02 -3.14026922e-01 -7.88291335e-01 -6.90915585e-01
-9.82002094e-02 -1.27599013e+00 -2.87062883e-01 -1.29284874e-01
-3.01214486e-01 1.43478617e-01 1.07097280e+00 -8.76312792e-01
1.02986622e+00 -2.29310131e+00 8.80911350e-01 -1.09662808e-01
2.20521297e-02 1.27206996e-01 -2.05545962e-01 1.65874466e-01
-1.74714029e-01 -1.81512743e-01 -5.66458642e-01 -9.67515111e-02
-3.13290507e-01 5.88351898e-02 -4.26486194e-01 5.50208151e-01
4.47760046e-01 9.02617037e-01 -1.13733387e+00 -4.88120466e-01
1.88465714e-01 5.17736912e-01 -5.72285056e-01 6.41929388e-01
-5.81186473e-01 1.01094103e+00 -5.79235554e-01 -8.50986242e-02
5.64866722e-01 -3.26915793e-02 1.71193227e-01 -2.70059615e-01
-1.09152101e-01 -2.74841607e-01 -1.29415274e+00 1.97748518e+00
-3.00692409e-01 6.85048163e-01 1.22703105e-01 -1.31017458e+00
7.64029741e-01 3.03178817e-01 6.91841125e-01 -3.78470868e-01
1.03053749e-01 1.04992874e-01 2.45768681e-01 -8.49136591e-01
3.33740264e-01 -1.88143164e-01 -4.95758951e-02 3.15860569e-01
5.15867118e-03 -3.06904644e-01 2.14711651e-01 4.74302145e-03
8.38099122e-01 6.16489649e-01 7.72453099e-02 1.67440087e-01
7.90451050e-01 5.05987443e-02 4.22429711e-01 1.45339623e-01
3.26268137e-01 9.41895664e-01 5.10735631e-01 -4.78717923e-01
-1.20747793e+00 -6.19042575e-01 4.49941128e-01 3.61990005e-01
2.13305563e-01 -3.40953052e-01 -1.02953374e+00 -6.13980293e-01
-4.06595975e-01 3.37074637e-01 -6.16176248e-01 -3.19421947e-01
-9.99834359e-01 -5.34508824e-01 4.09334116e-02 1.34027541e-01
4.65092957e-01 -1.10680068e+00 -8.18117321e-01 3.58462900e-01
-5.21866858e-01 -1.44892108e+00 -7.01229692e-01 -3.38349223e-01
-6.47784531e-01 -1.06924129e+00 -1.13595164e+00 -7.63238192e-01
6.73479557e-01 2.56610692e-01 6.77429616e-01 -1.21156015e-02
-3.04037601e-01 1.72413826e-01 -2.24494204e-01 1.07332893e-01
-5.80475211e-01 8.40968192e-02 -1.22577429e-01 7.29807913e-01
-2.49620110e-01 -5.92777312e-01 -6.60064340e-01 2.25833982e-01
-1.47047985e+00 4.50666279e-01 6.31000400e-01 7.20176101e-01
3.34330708e-01 2.37291187e-01 1.81240603e-01 -5.66925585e-01
-3.02554015e-02 -5.39932311e-01 -6.78113997e-01 -2.53972411e-02
7.02032149e-02 3.85785550e-01 8.43086839e-01 -4.91808176e-01
-1.02251756e+00 6.28947854e-01 1.39882006e-02 -6.76917195e-01
-2.06664294e-01 2.24498153e-01 -6.07505798e-01 1.87341049e-01
2.67170131e-01 4.30946857e-01 7.86543563e-02 -3.08375567e-01
6.03238463e-01 3.76601219e-01 7.18036115e-01 -3.66154581e-01
1.13746333e+00 6.36699557e-01 2.51350135e-01 -9.82292116e-01
-3.69319648e-01 -8.54847357e-02 -1.00664783e+00 -2.14785546e-01
1.31424761e+00 -5.73959589e-01 -3.62408549e-01 5.18692970e-01
-1.52745795e+00 -4.16909814e-01 -1.71466544e-01 4.21310157e-01
-7.51456976e-01 5.03857315e-01 -2.39649013e-01 -5.76051772e-01
1.51573746e-02 -1.36709094e+00 8.98887634e-01 2.99335510e-01
-1.75777256e-01 -9.88958657e-01 2.15718895e-01 4.25726920e-02
7.10595399e-02 5.59810400e-01 9.19779718e-01 -2.88636982e-02
-9.45084393e-01 -4.26897928e-02 1.14267401e-01 3.62628430e-01
2.94029117e-01 3.84384573e-01 -7.21385062e-01 -1.39765874e-01
3.69408339e-01 1.56487793e-01 6.21357203e-01 3.41561198e-01
7.30996549e-01 -3.92146707e-01 -2.11181894e-01 7.65303552e-01
1.39529753e+00 4.39438969e-01 6.42923474e-01 3.27034742e-01
1.00171292e+00 1.04404080e+00 5.68298757e-01 2.02833042e-01
1.66171238e-01 1.09070170e+00 4.48564351e-01 7.21729547e-02
-1.20792687e-01 -3.51767480e-01 4.85155344e-01 8.41656446e-01
-2.77865648e-01 -1.44864976e-01 -8.38402808e-01 4.72281039e-01
-2.07929945e+00 -1.03073061e+00 -5.41861244e-02 2.16058016e+00
4.63202268e-01 6.07542880e-02 2.37048026e-02 -9.37854219e-03
8.32104385e-01 3.62976760e-01 -5.82065344e-01 1.72297686e-01
-1.07607305e-01 -1.75114334e-01 3.74568552e-02 6.20282292e-01
-1.01150405e+00 9.41570997e-01 5.01417780e+00 6.89078271e-01
-1.51992083e+00 -1.50688127e-01 6.17394686e-01 1.24829687e-01
-1.86168805e-01 3.37322026e-01 -6.03805721e-01 7.70858943e-01
9.25401688e-01 -3.47180367e-01 3.73481065e-01 4.70501363e-01
3.64232957e-01 6.02170732e-03 -1.26495624e+00 8.94362390e-01
2.17057317e-01 -1.02129233e+00 2.73824871e-01 2.22138762e-02
7.09819853e-01 -5.43686688e-01 3.24340761e-02 -5.17283566e-02
-2.71492958e-01 -7.84957886e-01 9.00128126e-01 8.23844671e-01
7.10673273e-01 -5.39533734e-01 3.04062575e-01 6.21140420e-01
-1.15282607e+00 8.74827951e-02 -2.05346227e-01 -4.17049862e-02
4.19311315e-01 1.90258712e-01 -4.08542931e-01 7.56372929e-01
3.22172046e-01 9.97563303e-01 -3.20130020e-01 8.54126990e-01
-4.05347675e-01 4.10128206e-01 -4.05916087e-02 3.03929657e-01
3.17061603e-01 -6.61665559e-01 7.45317757e-01 8.94847393e-01
6.49666429e-01 3.23962718e-02 -5.48886172e-02 9.26528513e-01
1.15077041e-01 2.04058439e-02 -8.13098669e-01 9.61143598e-02
-1.09772481e-01 1.43421555e+00 -7.66008437e-01 -3.54314983e-01
-3.00398529e-01 1.16584027e+00 -2.30841078e-02 5.35258710e-01
-8.99586856e-01 -3.19321811e-01 5.74939787e-01 2.71018803e-01
3.10591400e-01 -3.68916541e-01 3.47528666e-01 -1.34394693e+00
1.52886808e-01 -8.26266706e-01 1.89484644e-03 -8.02916646e-01
-6.22375488e-01 8.29168439e-01 1.46997809e-01 -1.71283793e+00
-6.46878004e-01 -3.75949740e-01 -6.84388340e-01 7.70397902e-01
-1.36958778e+00 -8.64466608e-01 -4.84168619e-01 5.58421135e-01
7.13784099e-01 7.11351112e-02 3.03541452e-01 2.06884772e-01
-4.95301396e-01 6.51986301e-02 2.52156481e-02 1.49994612e-01
7.12500095e-01 -1.06713974e+00 5.44092476e-01 9.64233220e-01
1.28970981e-01 1.99505135e-01 8.19357991e-01 -4.65352833e-01
-1.57022202e+00 -1.06658161e+00 6.97181582e-01 -4.86289561e-01
7.36012042e-01 -4.34282124e-01 -1.05527294e+00 4.82448190e-01
2.93272257e-01 1.14738114e-01 -1.22384593e-01 -9.50344741e-01
1.89117819e-01 -1.17890283e-01 -6.01010621e-01 7.67334819e-01
9.79906261e-01 -3.45908940e-01 -4.99767870e-01 5.58099635e-02
6.42234623e-01 -3.54599059e-01 -4.39746976e-01 1.10663017e-02
3.86302173e-01 -5.36524415e-01 8.16540241e-01 -5.83156884e-01
8.10387969e-01 -6.38603747e-01 1.83186904e-01 -1.45899379e+00
-2.32864872e-01 -8.82021427e-01 5.71447648e-02 1.30222344e+00
1.31111532e-01 -1.58161461e-01 8.75350118e-01 6.06497526e-01
1.48521602e-01 -5.78947403e-02 -9.12207425e-01 -3.91087919e-01
-2.95158364e-02 -7.16180801e-02 4.14943308e-01 8.09627056e-01
-3.04334551e-01 6.82436466e-01 -8.05892587e-01 1.21494859e-01
5.09027660e-01 2.76784480e-01 1.05765164e+00 -9.85104084e-01
-5.65837026e-01 -2.95131981e-01 -6.07919991e-01 -1.24432707e+00
3.70209336e-01 -6.44871771e-01 2.63792992e-01 -1.29960310e+00
-4.91683409e-02 2.02571135e-02 6.96271285e-02 7.96287227e-03
-4.44538474e-01 1.96710546e-02 5.37123442e-01 3.64435256e-01
-2.17953220e-01 7.60363817e-01 1.60780847e+00 -3.05401951e-01
-2.22243294e-01 -8.71672761e-03 -2.22486883e-01 7.49332190e-01
5.42479455e-01 -4.95699793e-01 -6.69211209e-01 -5.45003295e-01
1.35457531e-01 5.88840544e-01 4.11221206e-01 -1.08606398e+00
3.31885785e-01 -2.93094993e-01 2.18017057e-01 -1.67899311e-01
4.33959663e-01 -1.10813296e+00 6.19252563e-01 5.23483574e-01
-1.35217130e-01 2.25005463e-01 -4.60501686e-02 7.02554882e-01
-4.05101001e-01 -3.78664374e-01 6.74532592e-01 -2.15154842e-01
-8.60899091e-01 4.19552565e-01 -2.03578770e-01 6.91904128e-03
1.18713522e+00 -1.87775210e-01 -3.79321277e-02 -4.45359260e-01
-8.98114145e-01 -1.36079252e-01 6.88904166e-01 5.55064619e-01
4.96083021e-01 -1.34356582e+00 -6.89062893e-01 4.02426481e-01
-1.13115244e-01 3.18510711e-01 4.81243163e-01 8.56774747e-01
-4.98204529e-01 2.32761085e-01 -4.09865081e-01 -7.12144375e-01
-1.00275040e+00 8.56991708e-01 3.30141693e-01 -1.36321317e-02
-8.27373385e-01 1.46008238e-01 6.98503673e-01 2.37756774e-01
-6.51363805e-02 -4.57706422e-01 -4.13069516e-01 1.25836313e-01
5.95536172e-01 1.40652806e-01 -2.07555711e-01 -1.12209916e+00
-4.17602360e-02 1.04576504e+00 2.17867181e-01 -3.55539531e-01
1.13415599e+00 -1.45539209e-01 6.08125106e-02 5.48885286e-01
1.48252869e+00 -1.26515642e-01 -1.58537579e+00 -1.43160701e-01
3.73556688e-02 -4.66757834e-01 -2.98208982e-01 2.31704768e-02
-1.27780163e+00 9.35718477e-01 2.22093374e-01 5.72671667e-02
1.04502833e+00 -1.86996028e-01 7.60083258e-01 -1.29163787e-02
3.68033856e-01 -7.73507953e-01 1.03551149e-01 3.76155376e-01
8.22222650e-01 -9.98474717e-01 -3.14554185e-01 -1.69413984e-01
-5.47056198e-01 1.31205893e+00 4.17348623e-01 -2.29179427e-01
6.05511248e-01 -7.40357954e-03 4.19815704e-02 2.81648278e-01
-6.37255907e-01 -3.84262279e-02 4.39006537e-01 6.26909256e-01
-1.28338002e-02 -3.63019109e-01 -3.08214009e-01 3.66942495e-01
-2.07354240e-02 1.18078411e-01 6.97534382e-01 6.95843518e-01
-1.39936551e-01 -1.21329904e+00 -3.18373829e-01 -2.57538199e-01
-1.87355682e-01 2.37191767e-01 -3.56435061e-01 7.50497341e-01
2.82740891e-01 5.83312929e-01 7.20255151e-02 -3.30283642e-01
3.45019042e-01 -3.67844254e-02 3.49929392e-01 -4.56081599e-01
-1.20396025e-01 3.25186580e-01 -5.08109808e-01 -5.08266151e-01
-6.65527344e-01 -7.10841596e-01 -1.16005707e+00 9.79810432e-02
1.48671359e-01 4.21579093e-01 5.94599962e-01 8.03172410e-01
1.56355500e-01 4.96852994e-01 7.12654650e-01 -1.22838366e+00
-2.26086736e-01 -7.17337012e-01 -5.19363761e-01 8.66883039e-01
6.14935338e-01 -4.77393299e-01 -5.09803057e-01 6.68226123e-01] | [10.819844245910645, -0.8381256461143494] |
7e5ae592-2d29-4e3a-b81f-d3ff34dfe6d6 | a-class-wise-non-salient-region-generalized | 2212.14154 | null | https://arxiv.org/abs/2212.14154v1 | https://arxiv.org/pdf/2212.14154v1.pdf | A Class-wise Non-salient Region Generalized Framework for Video Semantic Segmentation | Video semantic segmentation (VSS) is beneficial for dealing with dynamic scenes due to the continuous property of the real-world environment. On the one hand, some methods alleviate the predicted inconsistent problem between continuous frames. On the other hand, other methods employ the previous frame as the prior information to assist in segmenting the current frame. Although the previous methods achieve superior performances on the independent and identically distributed (i.i.d) data, they can not generalize well on other unseen domains. Thus, we explore a new task, the video generalizable semantic segmentation (VGSS) task that considers both continuous frames and domain generalization. In this paper, we propose a class-wise non-salient region generalized (CNSG) framework for the VGSS task. Concretely, we first define the class-wise non-salient feature, which describes features of the class-wise non-salient region that carry more generalizable information. Then, we propose a class-wise non-salient feature reasoning strategy to select and enhance the most generalized channels adaptively. Finally, we propose an inter-frame non-salient centroid alignment loss to alleviate the predicted inconsistent problem in the VGSS task. We also extend our video-based framework to the image-based generalizable semantic segmentation (IGSS) task. Experiments demonstrate that our CNSG framework yields significant improvement in the VGSS and IGSS tasks. | ['Chen Xu', 'Wenbin Zou', 'Zhengyu Zhang', 'Muxin Liao', 'Shishun Tian', 'Yuhang Zhang'] | 2022-12-29 | null | null | null | null | ['video-semantic-segmentation'] | ['computer-vision'] | [ 3.81740272e-01 -2.09455028e-01 -2.24372432e-01 -5.64512610e-01
-5.94871044e-01 -2.54269540e-01 1.73159316e-01 -1.62780598e-01
-3.00525665e-01 4.84233052e-01 -2.25694049e-02 1.40576273e-01
-1.08374611e-01 -5.96639812e-01 -7.80602038e-01 -7.79028118e-01
2.37088948e-01 1.24630801e-01 9.16175604e-01 -3.18136960e-01
2.56458014e-01 2.29786590e-01 -1.57108879e+00 2.95852572e-01
1.06343806e+00 1.20539367e+00 7.63595104e-01 1.54632643e-01
-2.18644105e-02 4.37075615e-01 -4.74291384e-01 1.15486868e-01
3.41608882e-01 -5.72349548e-01 -9.92311120e-01 5.51265001e-01
4.00376886e-01 -3.16442460e-01 -3.25799324e-02 1.34944403e+00
2.24729910e-01 6.17646694e-01 4.03169185e-01 -1.67837298e+00
-5.00101209e-01 1.60666123e-01 -7.71481752e-01 5.48329830e-01
3.49732339e-01 5.33499084e-02 8.12432647e-01 -7.45591342e-01
7.61166513e-01 1.48045111e+00 3.93719971e-01 4.89474267e-01
-6.82549596e-01 -5.88002920e-01 7.71082222e-01 6.28677666e-01
-1.29584527e+00 -1.61781207e-01 9.24942374e-01 -2.59379208e-01
5.94139218e-01 3.35834622e-01 5.92813313e-01 8.15410435e-01
-1.03594474e-01 1.20425248e+00 8.72695565e-01 -1.69014663e-01
4.02932972e-01 -2.60155737e-01 1.35939931e-02 5.14029324e-01
5.87184280e-02 -4.00638461e-01 -4.23487991e-01 1.27874985e-01
8.80506277e-01 2.50593185e-01 -4.16821688e-01 -3.70953858e-01
-1.26159346e+00 6.14162624e-01 3.28611046e-01 3.65704417e-01
-3.85041475e-01 6.85648099e-02 4.15166944e-01 6.75485749e-03
4.38737005e-01 9.04274881e-02 -6.68911755e-01 -1.65145062e-02
-9.64633822e-01 2.95846850e-01 3.29189718e-01 1.28925323e+00
9.83099461e-01 -7.16002360e-02 -2.57440627e-01 8.47728610e-01
2.54369110e-01 2.94367254e-01 4.54888195e-01 -1.28096545e+00
5.48645318e-01 5.08401871e-01 5.41843697e-02 -1.19214404e+00
-4.34247524e-01 -2.80998170e-01 -7.23334014e-01 -2.87105054e-01
3.29338819e-01 4.76410203e-02 -1.00770128e+00 1.85388708e+00
5.41561425e-01 5.77718735e-01 3.84144038e-02 1.29064202e+00
7.57659972e-01 7.02987134e-01 3.17756861e-01 -4.79179472e-01
1.27078068e+00 -1.23781991e+00 -6.47126794e-01 -3.46728086e-01
3.15920383e-01 -5.54909289e-01 1.04325199e+00 1.09673359e-01
-1.03262997e+00 -7.14418590e-01 -8.25684547e-01 -7.65069202e-02
-1.52743444e-01 -1.42563030e-01 4.78430212e-01 2.72137195e-01
-9.88989234e-01 3.51102769e-01 -8.82829428e-01 -5.05800843e-01
4.25440639e-01 2.03177497e-01 -2.65303820e-01 -1.80971220e-01
-1.24789464e+00 3.63987327e-01 7.80148506e-01 2.93514524e-02
-6.30786479e-01 -4.07712847e-01 -8.33004057e-01 4.26791720e-02
9.28499818e-01 -6.94437325e-01 1.05808723e+00 -1.40190077e+00
-1.24171877e+00 6.54554129e-01 -3.52109820e-01 -2.19058603e-01
4.35338616e-01 8.63298122e-03 -4.22511309e-01 5.89826345e-01
4.37424511e-01 8.20923924e-01 9.31899190e-01 -1.32285464e+00
-1.02068293e+00 -3.19182903e-01 1.31990433e-01 6.07812166e-01
-2.92972594e-01 -4.89605255e-02 -9.64851975e-01 -1.03798127e+00
6.01830721e-01 -9.82031763e-01 -2.24463284e-01 1.52371195e-03
-4.24643427e-01 -3.82091701e-01 1.25503111e+00 -6.46452129e-01
1.07836604e+00 -2.43913746e+00 2.51920164e-01 5.05267121e-02
-3.38750072e-02 2.69177526e-01 -3.08466345e-01 4.13533747e-02
5.97688071e-02 -4.03667754e-03 -3.29092264e-01 -2.06762820e-01
-2.42917880e-01 5.17864406e-01 -2.55314237e-03 2.41774425e-01
1.64625466e-01 7.74680495e-01 -1.07878339e+00 -7.45489895e-01
2.18057632e-01 4.20035496e-02 -6.79081202e-01 4.28319722e-02
-3.24146628e-01 7.19376504e-01 -8.31502199e-01 6.54261708e-01
8.33353162e-01 -3.17208081e-01 -1.55972242e-01 -5.10676682e-01
6.19402453e-02 -3.96757334e-01 -1.31766796e+00 1.92005396e+00
3.81427482e-02 3.27885896e-01 -3.82830156e-03 -1.33324981e+00
7.56451786e-01 -3.27000245e-02 7.15570986e-01 -6.11494184e-01
-7.66573623e-02 1.61666483e-01 -2.71150798e-01 -7.44030297e-01
6.55613780e-01 -6.67880476e-02 -4.46403697e-02 -5.18285558e-02
5.05588911e-02 -1.23042157e-02 1.83767468e-01 3.63863528e-01
7.03050852e-01 2.10325003e-01 3.57810676e-01 -3.79730374e-01
8.22698116e-01 -6.35393187e-02 1.21511161e+00 6.22882962e-01
-6.49585545e-01 9.02472138e-01 3.89785111e-01 -2.51712054e-01
-6.49114728e-01 -1.01236713e+00 1.00339018e-01 1.24447823e+00
1.23947346e+00 -2.94611067e-01 -8.78502131e-01 -8.54160786e-01
-1.35774478e-01 5.75499594e-01 -2.93297052e-01 -1.69447467e-01
-5.19529283e-01 -5.21841526e-01 1.36978939e-01 7.34744906e-01
1.00264013e+00 -9.08239663e-01 -5.90560794e-01 1.64873093e-01
-5.75290084e-01 -1.44915986e+00 -7.55754769e-01 -2.69317746e-01
-6.84186101e-01 -1.04175305e+00 -9.13587153e-01 -9.39997554e-01
6.30071163e-01 7.92952478e-01 6.20719612e-01 1.60291478e-01
8.34321529e-02 5.88409185e-01 -6.72962248e-01 -1.23185836e-01
3.59789059e-02 -1.88161358e-01 -7.00494926e-03 1.71426073e-01
2.78874665e-01 -1.71849638e-01 -8.09456468e-01 6.81865156e-01
-1.12002397e+00 3.01190346e-01 2.04893708e-01 7.50814438e-01
9.17755663e-01 2.51581550e-01 8.22471380e-01 -6.89446628e-01
2.18998328e-01 -4.58983183e-01 -2.17808619e-01 3.96095932e-01
-2.98934788e-01 -2.69530594e-01 6.01352513e-01 -4.30752009e-01
-1.29670227e+00 1.03424005e-01 2.01742463e-02 -5.32552361e-01
-1.85924634e-01 3.16633075e-01 -5.42437911e-01 -9.76757891e-03
1.45869032e-01 3.08326960e-01 -2.29554638e-01 -2.75005400e-01
1.83514982e-01 6.20705724e-01 6.93086147e-01 -6.40625894e-01
5.74101031e-01 7.36479223e-01 -2.51242518e-02 -8.03608954e-01
-9.12083328e-01 -8.31707835e-01 -5.84831774e-01 -2.90644526e-01
1.11438537e+00 -1.05684757e+00 -3.12182009e-01 5.92846692e-01
-1.04367530e+00 -2.22466037e-01 -1.40782371e-01 5.01555979e-01
-8.09973359e-01 9.02727902e-01 -3.68597329e-01 -5.55431128e-01
-1.00892715e-01 -1.41469812e+00 1.36206818e+00 4.83164728e-01
4.23814580e-02 -7.69774437e-01 -5.77040553e-01 4.83140498e-01
9.42723751e-02 2.78603226e-01 6.54675484e-01 -7.01315343e-01
-7.47858763e-01 2.96003908e-01 -5.03535986e-01 5.01844108e-01
6.72791675e-02 -2.20514134e-01 -5.25106311e-01 -2.77163208e-01
9.58656445e-02 1.42460773e-02 8.44009757e-01 5.77078760e-01
1.49505424e+00 -6.60005137e-02 -3.33167106e-01 7.09645569e-01
1.35370111e+00 5.59753537e-01 6.18849814e-01 5.26600778e-01
8.78880322e-01 4.93650109e-01 1.28458822e+00 4.44538116e-01
6.00356340e-01 7.48136640e-01 4.34955537e-01 -2.21247181e-01
-1.04495734e-01 -9.83211175e-02 1.70112476e-01 5.45498133e-01
-2.01366246e-01 -3.84645641e-01 -6.47652388e-01 5.35242021e-01
-2.21229005e+00 -9.89593804e-01 -1.72194466e-01 1.74394500e+00
5.40103257e-01 -9.93188620e-02 4.87089679e-02 -2.02443134e-02
1.19136715e+00 1.71943903e-01 -7.06488431e-01 3.55969630e-02
-2.69414216e-01 -2.90682971e-01 3.84205610e-01 5.91135547e-02
-1.25497770e+00 1.09476233e+00 4.82352304e+00 1.23193336e+00
-9.27209377e-01 2.96014965e-01 6.86860383e-01 2.71436304e-01
-2.57908046e-01 7.34481439e-02 -6.83110058e-01 8.19604635e-01
1.71784326e-01 -2.56098986e-01 3.33998501e-01 8.72696757e-01
3.16753894e-01 -4.08174962e-01 -9.46558416e-01 1.23255157e+00
2.19783634e-01 -9.72243428e-01 2.13180527e-01 -4.07164961e-01
1.00913537e+00 -3.67575765e-01 6.33283630e-02 1.14896670e-01
-3.41826767e-01 -4.26239222e-01 1.04556894e+00 3.37397784e-01
6.30118608e-01 -6.66853607e-01 6.44318938e-01 3.09914827e-01
-1.36898243e+00 -1.54161945e-01 -4.88283485e-01 2.35203892e-01
4.51667368e-01 3.73568773e-01 -3.21422905e-01 8.12387466e-01
9.44800973e-01 9.97839212e-01 -5.15324056e-01 1.06781316e+00
-6.52700141e-02 3.23667526e-01 -2.45615467e-01 3.29452634e-01
4.75622267e-01 -2.93049842e-01 6.66479766e-01 1.03016710e+00
4.73631740e-01 4.69294757e-01 4.87838149e-01 6.02472603e-01
1.93319574e-01 6.61502704e-02 -9.58560184e-02 4.36045140e-01
5.02012610e-01 9.79203761e-01 -1.23357296e+00 -6.29261792e-01
-5.34468532e-01 1.22978044e+00 -1.63075700e-01 7.03078687e-01
-9.93779004e-01 -2.42704943e-01 4.18964982e-01 -5.85296750e-02
4.42780584e-01 -1.12899564e-01 -1.23835362e-01 -1.33200097e+00
1.58576176e-01 -6.87999845e-01 6.68276429e-01 -1.02545547e+00
-1.21355140e+00 4.06040102e-01 3.08416098e-01 -1.41540670e+00
-1.38997272e-01 -1.87695950e-01 -4.24750924e-01 4.53991532e-01
-1.66241407e+00 -1.16139007e+00 -6.47704840e-01 9.69152272e-01
1.01524770e+00 7.18136085e-03 1.56885907e-01 3.45193893e-01
-5.61922908e-01 3.50967288e-01 -5.30951694e-02 -4.81682830e-02
6.70704544e-01 -9.03353333e-01 1.62486240e-01 1.08762395e+00
-3.39119971e-01 4.37507570e-01 6.63996816e-01 -7.69635856e-01
-1.03311682e+00 -1.42781174e+00 3.65611017e-01 1.74643874e-01
4.00874346e-01 6.30358607e-02 -1.12942898e+00 7.18005240e-01
-2.61808842e-01 1.17689930e-01 1.90512925e-01 -3.24022025e-01
-1.10350937e-01 -1.37383416e-02 -1.23866594e+00 5.14169991e-01
1.43065643e+00 -2.86519319e-01 -6.07161641e-01 3.48645866e-01
8.71097028e-01 -4.76800352e-01 -6.92552090e-01 6.37142897e-01
1.65965393e-01 -9.98614669e-01 1.14077449e+00 -2.37362191e-01
2.99190819e-01 -6.45318389e-01 -3.12979549e-01 -1.05141068e+00
-2.20544234e-01 -4.30482149e-01 2.14658871e-01 1.19789970e+00
-1.00655623e-01 -5.76303840e-01 7.20129013e-01 6.84154332e-01
-4.99996901e-01 -5.47874451e-01 -1.05661047e+00 -1.11914277e+00
-2.79796720e-01 -3.61104012e-01 6.43220603e-01 9.53434110e-01
-1.36780992e-01 -1.18191615e-02 -2.99849361e-01 2.32001930e-01
5.49948335e-01 2.63960272e-01 5.77237368e-01 -9.57536697e-01
-1.72466174e-01 -1.84541062e-01 -7.86997139e-01 -1.39738929e+00
2.50182420e-01 -6.54366970e-01 2.71401674e-01 -1.54292536e+00
4.23097104e-01 -4.50476974e-01 -3.52057129e-01 4.45880264e-01
-5.36463439e-01 1.96307331e-01 4.47538376e-01 2.70670384e-01
-1.25089037e+00 6.15993977e-01 1.52471161e+00 -9.28640831e-03
-3.10797691e-01 -5.28303608e-02 -6.20707095e-01 8.73127401e-01
6.96814299e-01 -1.53281480e-01 -7.24930763e-01 -4.71870601e-01
-2.18181208e-01 1.16786346e-01 5.54242671e-01 -9.82923210e-01
2.55162239e-01 -5.31756818e-01 2.56166339e-01 -7.01465964e-01
2.81749874e-01 -7.79922187e-01 7.89783075e-02 2.11148217e-01
-1.35677293e-01 -1.97237045e-01 1.15676234e-02 9.08990204e-01
-5.03904521e-01 -1.64997011e-01 8.62486064e-01 -1.38851747e-01
-1.62138975e+00 4.76800889e-01 -1.78034097e-01 2.96794087e-01
1.41428804e+00 -5.58637321e-01 -3.21738482e-01 -4.10353750e-01
-7.31361687e-01 4.55043077e-01 6.02968693e-01 5.72022080e-01
7.54189491e-01 -1.18370819e+00 -3.30528110e-01 2.21933663e-01
2.01430187e-01 2.79870152e-01 7.98401177e-01 9.37763035e-01
-3.93523395e-01 1.05176374e-01 -2.88740188e-01 -8.75274360e-01
-1.23635113e+00 7.40670979e-01 7.14665651e-02 2.75724202e-01
-6.83921158e-01 8.80568802e-01 8.24060500e-01 1.82795733e-01
-6.18770812e-03 -4.18014944e-01 -2.02388942e-01 -5.09550646e-02
4.42297667e-01 5.01393735e-01 -1.28898606e-01 -9.77077484e-01
-4.55092251e-01 7.18229771e-01 1.05280047e-02 2.01865345e-01
1.17897511e+00 -6.80202842e-01 -8.48021582e-02 2.04800233e-01
1.30251169e+00 -4.60284203e-01 -1.63295913e+00 -2.32543454e-01
-6.85617775e-02 -6.25575185e-01 -8.00739378e-02 -4.65117753e-01
-1.29526269e+00 4.74015087e-01 4.71647322e-01 2.25863885e-02
1.61132586e+00 1.87346458e-01 1.12088156e+00 -3.30499746e-02
4.85439956e-01 -1.37929094e+00 1.64408222e-01 4.13603842e-01
6.37135386e-01 -1.17319643e+00 -6.31205216e-02 -9.56952631e-01
-1.05199599e+00 9.73841131e-01 8.17823112e-01 6.13968857e-02
4.31539744e-01 -2.18513563e-01 -1.63047150e-01 -1.16899729e-01
-2.69633442e-01 -4.25421298e-01 3.13345671e-01 7.05649734e-01
-1.73289999e-01 -2.92839613e-02 -5.20911574e-01 6.82366610e-01
3.60106230e-01 1.52314961e-01 5.14838219e-01 8.91937494e-01
-5.80403388e-01 -7.22154856e-01 -4.09732401e-01 3.02579135e-01
-2.15798929e-01 2.05535918e-01 6.50979653e-02 6.77324951e-01
4.46912318e-01 1.13164091e+00 7.40025416e-02 -3.20631593e-01
3.38999093e-01 -2.70409822e-01 3.67983550e-01 -5.25064051e-01
-8.16117078e-02 3.07691365e-01 -3.23333353e-01 -7.83692598e-01
-9.04006660e-01 -8.14471424e-01 -1.68272316e+00 2.36993227e-02
-2.36057788e-01 -3.34490016e-02 2.93738693e-01 1.07566142e+00
3.50003392e-01 4.43813771e-01 4.67360318e-01 -9.05498028e-01
-2.24691242e-01 -4.41875696e-01 -8.35180342e-01 9.29277658e-01
1.36940107e-01 -9.41485524e-01 -2.35356227e-01 3.49016517e-01] | [9.363456726074219, -0.12327134609222412] |
fdafd2a3-453d-49f4-a1e8-77cc1f77ccca | iconqa-a-new-benchmark-for-abstract-diagram | 2110.13214 | null | https://arxiv.org/abs/2110.13214v4 | https://arxiv.org/pdf/2110.13214v4.pdf | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | Current visual question answering (VQA) tasks mainly consider answering human-annotated questions for natural images. However, aside from natural images, abstract diagrams with semantic richness are still understudied in visual understanding and reasoning research. In this work, we introduce a new challenge of Icon Question Answering (IconQA) with the goal of answering a question in an icon image context. We release IconQA, a large-scale dataset that consists of 107,439 questions and three sub-tasks: multi-image-choice, multi-text-choice, and filling-in-the-blank. The IconQA dataset is inspired by real-world diagram word problems that highlight the importance of abstract diagram understanding and comprehensive cognitive reasoning. Thus, IconQA requires not only perception skills like object recognition and text understanding, but also diverse cognitive reasoning skills, such as geometric reasoning, commonsense reasoning, and arithmetic reasoning. To facilitate potential IconQA models to learn semantic representations for icon images, we further release an icon dataset Icon645 which contains 645,687 colored icons on 377 classes. We conduct extensive user studies and blind experiments and reproduce a wide range of advanced VQA methods to benchmark the IconQA task. Also, we develop a strong IconQA baseline Patch-TRM that applies a pyramid cross-modal Transformer with input diagram embeddings pre-trained on the icon dataset. IconQA and Icon645 are available at https://iconqa.github.io. | ['Song-Chun Zhu', 'Xiaodan Liang', 'Zhou Yu', 'Wei zhang', 'Yizhou Zhao', 'Tony Xia', 'Jiaqi Chen', 'Liang Qiu', 'Pan Lu'] | 2021-10-25 | null | null | null | null | ['math-word-problem-solving', 'mathematical-question-answering', 'mathematical-reasoning', 'arithmetic-reasoning', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'natural-language-processing', 'natural-language-processing', 'reasoning', 'reasoning', 'time-series'] | [ 1.03029490e-01 6.22882368e-03 1.37559131e-01 -2.99039185e-01
-6.96856797e-01 -9.19317305e-01 5.31783938e-01 4.73243743e-02
-9.82001796e-02 3.11896831e-01 3.59395176e-01 -7.09746063e-01
-6.35041073e-02 -9.27308500e-01 -9.98486876e-01 -5.35766743e-02
6.83254004e-01 5.48727036e-01 2.37090915e-01 -4.57430869e-01
4.44532186e-01 1.95941910e-01 -1.49193037e+00 1.26666081e+00
1.13588262e+00 1.25718439e+00 1.98474705e-01 9.66403425e-01
-4.33519781e-01 1.66561604e+00 -6.78889990e-01 -9.02525306e-01
3.80314328e-02 -5.15449405e-01 -1.25637329e+00 -7.24574476e-02
1.08128107e+00 -4.62433726e-01 -5.97677469e-01 9.53012586e-01
2.50439376e-01 2.46608973e-01 6.75911605e-01 -1.65139830e+00
-1.82137847e+00 2.70798951e-01 -2.72349626e-01 3.82585227e-01
7.84247756e-01 5.98956883e-01 1.41758478e+00 -1.20765841e+00
6.49797201e-01 1.62237239e+00 3.05524528e-01 5.03478050e-01
-9.84351993e-01 -6.53980017e-01 1.67166039e-01 9.87316072e-01
-9.70252216e-01 5.85467666e-02 7.66174853e-01 -3.77938688e-01
8.69761586e-01 5.60436010e-01 9.08555210e-01 1.05801547e+00
-7.20709041e-02 1.36946678e+00 1.11471808e+00 -2.33624905e-01
1.38004854e-01 -2.33129427e-01 1.71178758e-01 7.95986831e-01
8.91375616e-02 -2.29370773e-01 -4.89877254e-01 2.28078231e-01
9.33930397e-01 9.15409848e-02 -4.07289326e-01 -4.03710932e-01
-1.41090453e+00 9.33748841e-01 1.13105571e+00 5.30014448e-02
-1.71526283e-01 5.21150351e-01 4.73552883e-01 2.88637757e-01
-1.96955845e-01 7.61165261e-01 -2.51140803e-01 4.24898751e-02
-2.97367811e-01 3.55814785e-01 5.22220731e-01 1.02848399e+00
7.95392990e-01 -1.51193142e-01 -6.29482985e-01 7.34492600e-01
1.49536222e-01 8.70302975e-01 -1.39254391e-01 -1.29460740e+00
8.02983701e-01 1.21601868e+00 -1.59348205e-01 -1.06534290e+00
-1.83279067e-01 9.91335884e-02 -6.94623172e-01 1.47821814e-01
5.82704842e-01 4.57337588e-01 -1.26866901e+00 1.40918458e+00
9.95562747e-02 -2.96817958e-01 -1.47642875e-02 1.27750409e+00
1.83397627e+00 7.90725887e-01 4.67059642e-01 6.17519796e-01
1.98471570e+00 -1.32625854e+00 -6.65102541e-01 -5.35924852e-01
2.21853480e-01 -7.41453409e-01 1.94831622e+00 3.72301430e-01
-1.09955263e+00 -6.51411533e-01 -8.03369761e-01 -1.09792435e+00
-7.98333347e-01 2.39171505e-01 8.64529967e-01 3.95888329e-01
-1.00447953e+00 -3.61124247e-01 8.34102035e-02 -2.77757943e-01
9.32299435e-01 -2.54025578e-01 -2.88352460e-01 -8.33462954e-01
-1.35017753e+00 9.33339953e-01 1.31763309e-01 1.28417656e-01
-1.16806936e+00 -1.20886385e+00 -1.00989711e+00 9.49735045e-02
6.02583647e-01 -9.77284312e-01 1.15729904e+00 -1.07207429e+00
-9.14445996e-01 1.24307549e+00 -2.41254698e-02 1.69304907e-02
3.75037193e-01 -2.56301820e-01 -4.45163488e-01 8.21756780e-01
3.72369945e-01 1.14462769e+00 7.40972281e-01 -1.41575015e+00
-5.21153361e-02 -1.55273438e-01 9.74896312e-01 2.21494451e-01
2.82201082e-01 -2.79444575e-01 -5.87398052e-01 -7.02500045e-01
-1.48478448e-01 -6.00495636e-01 2.92340666e-01 5.98781824e-01
-3.12953383e-01 -5.52718997e-01 6.65298998e-01 -9.35762227e-01
7.98215628e-01 -2.01009989e+00 1.04956806e-01 -1.85586974e-01
4.76029217e-01 1.03875101e-01 -5.48139095e-01 5.62224567e-01
-9.74782780e-02 9.49366167e-02 -2.20337540e-01 3.12999159e-01
3.23650956e-01 2.25707963e-01 -6.53540492e-01 -7.57353231e-02
3.63125414e-01 1.90779245e+00 -9.90720212e-01 -6.36883438e-01
4.35111076e-01 1.69739559e-01 -7.09160209e-01 2.07060322e-01
-7.31944680e-01 2.56752342e-01 -2.63952285e-01 1.08281136e+00
8.24892342e-01 -7.17117250e-01 -1.94684416e-01 -5.50313652e-01
4.33738530e-01 -2.55496446e-02 -6.00407064e-01 1.92801178e+00
-6.74228966e-01 1.04436982e+00 -4.23422217e-01 -7.18418479e-01
6.50586724e-01 -1.79189712e-01 -1.68157488e-01 -1.24055934e+00
1.43186852e-01 -9.00571495e-02 -8.70894790e-02 -7.76967466e-01
4.30587590e-01 7.86251202e-03 -2.31342360e-01 4.35523689e-01
9.60931703e-02 -7.21565723e-01 3.06661338e-01 8.94242287e-01
9.51766372e-01 -3.73732634e-02 2.62142748e-01 -1.68000549e-01
6.84244931e-01 2.49148458e-01 -2.63349801e-01 8.08458626e-01
-4.42487121e-01 7.41280317e-01 8.60811055e-01 -6.08447492e-01
-9.31530416e-01 -1.59476292e+00 9.43958461e-02 1.34610474e+00
5.87745309e-01 -3.96122277e-01 -4.76441205e-01 -6.97847903e-01
1.44924313e-01 9.14214730e-01 -1.13629866e+00 2.49244496e-02
-3.05433720e-01 -1.34239361e-01 5.57255089e-01 7.60917068e-01
9.11381900e-01 -1.69680667e+00 -5.48172116e-01 -3.72040272e-01
-5.69506109e-01 -1.27921343e+00 -5.77947974e-01 -4.17179644e-01
-2.42056876e-01 -1.64612782e+00 -5.77221632e-01 -1.05646801e+00
7.44337440e-01 4.51951146e-01 1.81948972e+00 4.13629085e-01
-5.97742319e-01 1.03288901e+00 -4.97425020e-01 -3.29041034e-01
-6.64504841e-02 -4.78195816e-01 -8.09057236e-01 -3.24121058e-01
5.29751718e-01 1.28963552e-02 -9.26481128e-01 2.36661285e-01
-9.77234244e-01 3.47084135e-01 5.84136844e-01 6.13818228e-01
3.08611006e-01 -5.53255975e-01 4.11832452e-01 -5.72330832e-01
5.49795508e-01 -2.56969154e-01 -2.64552355e-01 8.17929029e-01
2.06943955e-02 -7.66648129e-02 6.19549990e-01 -3.12083334e-01
-1.22610223e+00 -3.96762133e-01 -1.01355925e-01 -3.79808962e-01
3.92466746e-02 9.10025761e-02 -4.32163179e-01 -5.40019646e-02
6.66264534e-01 2.74072260e-01 -2.18518272e-01 1.12980023e-01
1.18787920e+00 2.28175327e-01 7.10386813e-01 -8.52012455e-01
6.25623703e-01 5.58655322e-01 -3.47361058e-01 -7.27667212e-01
-1.18412435e+00 -3.84602278e-01 -2.89261043e-01 -5.29543936e-01
1.35252225e+00 -9.65892136e-01 -1.42339027e+00 3.01066667e-01
-1.53696585e+00 -4.90782470e-01 -2.53292918e-01 -1.38651326e-01
-5.49208462e-01 4.31807786e-01 -4.79705811e-01 -3.21036249e-01
-4.15240228e-01 -1.01535094e+00 1.07971776e+00 2.69675434e-01
-1.01922341e-01 -7.84812629e-01 -1.97243169e-01 1.11185467e+00
1.29251570e-01 -3.82830831e-03 1.37392962e+00 -8.01490843e-02
-1.15154815e+00 2.87089437e-01 -1.33771002e+00 2.30818808e-01
-3.93273741e-01 -3.39211226e-01 -9.13994968e-01 1.52578518e-01
-5.78205943e-01 -1.06580496e+00 1.00393033e+00 1.51744813e-01
1.64423358e+00 -9.37742069e-02 1.24487065e-01 5.73111534e-01
1.49955630e+00 1.33281082e-01 1.17782414e+00 2.51766264e-01
1.02503645e+00 5.06692529e-01 4.50325638e-01 8.69096294e-02
9.95273471e-01 1.12414166e-01 7.97028661e-01 -1.16256990e-01
-6.28383994e-01 -4.90530372e-01 -2.28420366e-02 2.98958361e-01
5.61674871e-02 -3.11700284e-01 -1.24966788e+00 7.24970400e-01
-1.50239134e+00 -1.06702352e+00 -3.85225624e-01 1.21109772e+00
7.76637852e-01 -3.73511225e-01 -1.54590800e-01 -1.56785578e-01
4.85080332e-01 3.51576418e-01 -8.04637074e-01 -4.99417841e-01
-2.08608404e-01 3.76165956e-01 -9.58670452e-02 3.21933925e-01
-9.60897923e-01 1.08469343e+00 5.48746681e+00 7.15229750e-01
-4.67510164e-01 7.85372332e-02 4.91605073e-01 2.27356046e-01
-7.92249858e-01 7.43702874e-02 -1.18771471e-01 1.52613774e-01
-6.45799562e-03 2.74528086e-01 6.65998280e-01 7.59060740e-01
-4.96581078e-01 -3.31269443e-01 -1.00152469e+00 1.52427900e+00
5.95088303e-01 -1.53578925e+00 8.87160182e-01 -5.68485141e-01
6.51961505e-01 -5.28327048e-01 3.07191074e-01 7.32443392e-01
3.56409252e-01 -1.58189237e+00 7.36353815e-01 3.88169080e-01
1.05133545e+00 -4.81904417e-01 3.61647487e-01 -2.54170179e-01
-1.28534126e+00 -2.26800069e-01 -4.32890505e-01 -1.63903534e-01
7.65246078e-02 2.56273985e-01 -2.37135634e-01 4.43761319e-01
1.01065600e+00 7.84890234e-01 -1.16435671e+00 7.13652015e-01
-7.29750514e-01 1.75395936e-01 2.80454129e-01 -4.19160038e-01
3.36867452e-01 -2.74374913e-02 -7.83724934e-02 8.12751472e-01
-1.79730192e-01 5.50887465e-01 -2.65392631e-01 1.19279289e+00
-2.76863486e-01 -4.86958437e-02 -2.72256583e-01 -7.98511058e-02
3.11856896e-01 1.10054791e+00 -5.97758889e-01 -4.29388106e-01
-6.56176448e-01 1.17253363e+00 5.00251472e-01 7.14793265e-01
-1.17863059e+00 -6.01863444e-01 4.70698208e-01 -1.17009729e-01
3.29530120e-01 -2.27562394e-02 -4.68670338e-01 -1.29111981e+00
-8.92686620e-02 -1.26990402e+00 6.95408225e-01 -1.77940464e+00
-1.53453088e+00 4.89302963e-01 -6.15791902e-02 -8.26115549e-01
4.29677278e-01 -1.26812136e+00 -5.81521451e-01 4.55340028e-01
-1.31584120e+00 -1.68173909e+00 -1.08390963e+00 8.33834708e-01
7.13980079e-01 4.69457395e-02 4.24629122e-01 1.31322309e-01
-1.48867583e-02 2.73219585e-01 -6.23369992e-01 4.32523519e-01
6.02545321e-01 -1.43810964e+00 5.33123374e-01 6.01732314e-01
2.81256735e-01 5.06725907e-01 2.67275542e-01 -4.70199674e-01
-1.40075612e+00 -8.54253531e-01 3.94768625e-01 -1.20930326e+00
8.43451202e-01 -6.12738967e-01 -8.96306396e-01 7.31621981e-01
5.97638965e-01 1.33490458e-01 4.76204097e-01 -4.55770716e-02
-1.02195764e+00 7.68446475e-02 -8.16298068e-01 8.30003858e-01
1.40576804e+00 -1.02572072e+00 -1.16787601e+00 3.83837879e-01
8.96283567e-01 -3.57720762e-01 -5.70788205e-01 2.09003910e-01
6.49473429e-01 -9.70380366e-01 1.54220819e+00 -8.37745845e-01
1.03382754e+00 -3.05719972e-01 -3.42721879e-01 -1.00297427e+00
-2.66084135e-01 1.51172196e-02 8.88405815e-02 8.36724579e-01
1.22572780e-01 -1.29393756e-01 5.47982156e-01 4.00251836e-01
9.26358029e-02 -5.84007859e-01 -6.56957805e-01 -3.04736167e-01
2.35807180e-01 -6.03218853e-01 5.33303916e-01 8.01951170e-01
-2.41516456e-01 5.89889348e-01 -1.34229690e-01 -1.57753397e-02
5.28970838e-01 5.55153191e-01 9.15005445e-01 -8.79612505e-01
-1.32275082e-03 -3.74335080e-01 -3.49960774e-01 -1.32173944e+00
2.63815135e-01 -1.04979408e+00 -1.79094866e-01 -2.19747782e+00
4.43749636e-01 2.58949734e-02 6.47224560e-02 3.78540605e-01
-4.14600313e-01 5.58739185e-01 4.73911434e-01 -1.30227476e-01
-1.18566263e+00 6.78205907e-01 2.09285378e+00 -6.05885684e-01
3.93187761e-01 -8.24661732e-01 -7.34855413e-01 5.56034207e-01
5.14630198e-01 2.16643274e-01 -8.34147036e-01 -7.69576848e-01
6.75185084e-01 -8.18115920e-02 1.26040161e+00 -7.92288363e-01
1.70972720e-02 -1.93309814e-01 8.49498689e-01 -9.44340169e-01
3.06648880e-01 -6.44838572e-01 -4.05525595e-01 3.93579990e-01
-3.20942730e-01 3.19987088e-01 3.09066653e-01 6.09467447e-01
-2.67991245e-01 -1.18673323e-02 6.33952618e-01 -3.80528063e-01
-1.58558285e+00 2.19640225e-01 -1.47109836e-01 8.96646142e-01
1.06349456e+00 -7.56916255e-02 -1.19939089e+00 -6.67189479e-01
-4.79049534e-01 5.93030512e-01 3.71111512e-01 6.96635664e-01
1.20093966e+00 -1.44398558e+00 -6.35376573e-01 -1.47102043e-01
8.05306435e-01 -5.88037027e-03 8.60477865e-01 4.20786649e-01
-1.08712077e+00 5.61753690e-01 -6.08932972e-01 -4.64010507e-01
-9.69371378e-01 1.04004693e+00 2.12273568e-01 1.48008764e-01
-5.66366017e-01 1.08259523e+00 9.13208008e-01 -7.19602466e-01
-4.51640300e-02 -5.56238949e-01 -7.92151913e-02 -5.75329997e-02
6.30461752e-01 2.47838557e-01 -4.67818439e-01 -2.82451093e-01
-6.21697128e-01 8.10178518e-01 2.11012304e-01 1.26516253e-01
6.70554757e-01 6.95482716e-02 -4.16559279e-01 1.73688278e-01
1.06058383e+00 -2.60879844e-01 -9.82357383e-01 9.49567333e-02
-2.86752015e-01 -5.39603055e-01 -3.82974327e-01 -1.34955466e+00
-7.77600348e-01 1.26056457e+00 1.59427449e-01 -2.08440170e-01
1.18919766e+00 5.67430139e-01 7.41520226e-01 5.42054296e-01
1.06660284e-01 -7.38910377e-01 1.12507594e+00 5.54402709e-01
1.73416233e+00 -1.44284773e+00 1.34528633e-02 -5.14143467e-01
-1.09124637e+00 1.02389550e+00 1.23152757e+00 -1.85684487e-01
3.45316112e-01 -5.22056341e-01 2.78311729e-01 -5.44553220e-01
-6.90615237e-01 -5.40224075e-01 7.50192881e-01 7.99674034e-01
1.31503001e-01 7.42979581e-03 1.53216839e-01 6.21870816e-01
-2.04044923e-01 -2.42645547e-01 4.02993053e-01 4.72238660e-01
-1.91458598e-01 -3.84182960e-01 -4.59081292e-01 4.37737525e-01
2.26949409e-01 -5.01107037e-01 -6.74012482e-01 1.05958879e+00
1.30057245e-01 9.23329890e-01 2.17994496e-01 1.57743972e-02
4.11344320e-01 -1.06039301e-01 8.83430600e-01 -3.24801534e-01
-3.22554857e-01 -9.81630147e-01 5.17008826e-02 -7.28390455e-01
-3.00458044e-01 -1.34141609e-01 -1.28472722e+00 -3.64778191e-01
2.87628651e-01 -2.34169245e-01 2.45713457e-01 7.44773209e-01
1.15178637e-01 6.94083810e-01 -2.79488295e-01 -3.34243536e-01
-6.85998844e-03 -7.29471385e-01 -2.01900721e-01 8.81337523e-01
2.99410969e-01 -8.04421842e-01 -2.72497356e-01 -8.17846432e-02] | [10.777029991149902, 1.797850489616394] |
a5b0b7d2-659b-4f9d-9907-b6610a4c6534 | reducing-quantum-annealing-biases-for-solving | 2103.04963 | null | https://arxiv.org/abs/2103.04963v1 | https://arxiv.org/pdf/2103.04963v1.pdf | Reducing quantum annealing biases for solving the graph partitioning problem | Quantum annealers offer an efficient way to compute high quality solutions of NP-hard problems when expressed in a QUBO (quadratic unconstrained binary optimization) or an Ising form. This is done by mapping a problem onto the physical qubits and couplers of the quantum chip, from which a solution is read after a process called quantum annealing. However, this process is subject to multiple sources of biases, including poor calibration, leakage between adjacent qubits, control biases, etc., which might negatively influence the quality of the annealing results. In this work, we aim at mitigating the effect of such biases for solving constrained optimization problems, by offering a two-step method, and apply it to Graph Partitioning. In the first step, we measure and reduce any biases that result from implementing the constraints of the problem. In the second, we add the objective function to the resulting bias-corrected implementation of the constraints, and send the problem to the quantum annealer. We apply this concept to Graph Partitioning, an important NP-hard problem, which asks to find a partition of the vertices of a graph that is balanced (the constraint) and minimizes the cut size (the objective). We first quantify the bias of the implementation of the constraint on the quantum annealer, that is, we require, in an unbiased implementation, that any two vertices have the same likelihood of being assigned to the same or to different parts of the partition. We then propose an iterative method to correct any such biases. We demonstrate that, after adding the objective, solving the resulting bias-corrected Ising problem on the quantum annealer results in a higher solution accuracy. | ['Hristo N. Djidjev', 'Georg Hahn', 'Elijah Pelofske'] | 2021-03-08 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [ 4.91142601e-01 3.54368597e-01 2.09856719e-01 -1.34314150e-01
-5.43253362e-01 -7.51845419e-01 1.22934006e-01 3.67345661e-01
-4.23644662e-01 7.46668041e-01 -3.65626782e-01 -3.19500834e-01
-3.96351814e-01 -1.38307977e+00 -9.45715249e-01 -9.48118746e-01
3.40923965e-01 7.78403223e-01 8.82493854e-02 -3.17828655e-01
4.70919490e-01 2.93722481e-01 -1.18267643e+00 -1.48431823e-01
9.92606819e-01 7.68388152e-01 -1.07180856e-01 5.89925766e-01
2.11344346e-01 2.63659865e-01 -5.28058112e-01 -4.22732949e-01
4.57451314e-01 -7.63858438e-01 -1.01668859e+00 -1.76355049e-01
-7.16247112e-02 1.66482389e-01 -2.66954333e-01 1.66044593e+00
2.81400800e-01 9.86783803e-02 5.81197798e-01 -1.09677219e+00
-3.24420482e-01 6.33195102e-01 -2.93155998e-01 -1.94593132e-01
1.80717736e-01 1.49252012e-01 1.16782224e+00 4.87952605e-02
6.80758238e-01 8.98515642e-01 1.17807187e-01 4.46440279e-01
-1.42884827e+00 -3.91084671e-01 -5.40887654e-01 3.76015455e-02
-1.51714396e+00 -3.24769944e-01 5.46891272e-01 -2.74391115e-01
7.25199163e-01 5.04854500e-01 5.22437811e-01 2.75061786e-01
4.82354522e-01 -2.23618060e-01 1.39338672e+00 -6.97773576e-01
7.75550425e-01 1.98285371e-01 4.20186698e-01 5.90511262e-01
5.12702703e-01 2.09970951e-01 -2.82355249e-01 -1.62559763e-01
2.89102588e-02 -4.53563541e-01 -2.80991465e-01 -3.04207563e-01
-6.74456537e-01 7.85213292e-01 8.03898335e-01 2.90888995e-01
-2.23336890e-01 2.80775845e-01 -1.72196999e-01 3.26639414e-01
-3.20567153e-02 8.36185277e-01 6.51780441e-02 9.30920020e-02
-6.70631766e-01 6.82794675e-02 9.62452531e-01 4.79413480e-01
1.27111018e+00 -7.82788515e-01 -1.75388753e-01 2.23773941e-01
2.11300984e-01 5.28485537e-01 -1.79531977e-01 -6.58704698e-01
3.73527050e-01 5.85962534e-01 1.93223342e-01 -9.43580985e-01
-4.08414394e-01 -2.76062846e-01 -7.08798826e-01 2.81023920e-01
4.84329939e-01 -1.53719634e-01 -1.03119624e+00 1.80237186e+00
3.74727368e-01 -2.76875615e-01 2.48669274e-02 1.10892630e+00
3.03230494e-01 9.01884317e-01 -3.19915116e-01 -3.43224674e-01
1.29403508e+00 -5.86867034e-01 -6.05876327e-01 -2.46356472e-01
7.63126314e-01 -4.80830580e-01 5.95117390e-01 3.94085616e-01
-9.64873254e-01 -1.00552030e-02 -1.48272920e+00 -1.75012667e-02
-2.93168664e-01 -3.79233599e-01 3.23237181e-01 1.10545158e+00
-1.00581169e+00 9.63760972e-01 -7.56217301e-01 -1.15694730e-02
-7.40738288e-02 8.70625138e-01 -1.91224128e-01 -1.63025156e-01
-1.27791393e+00 1.02689660e+00 3.38140666e-01 3.88894111e-01
-3.06615353e-01 -6.35570362e-02 -5.69788098e-01 3.68057787e-01
5.72117090e-01 -6.33439422e-01 5.26862383e-01 -7.07490146e-01
-1.64686430e+00 8.78565192e-01 -1.05068296e-01 -2.34964713e-01
2.07788587e-01 4.95045483e-01 2.05842108e-01 -1.60380781e-01
-1.21960379e-01 2.55610347e-01 5.62950969e-01 -9.61629272e-01
-1.92628633e-02 -6.56110227e-01 3.57530087e-01 -5.41914068e-02
9.42768436e-03 -2.12110698e-01 -4.76498723e-01 3.17403853e-01
5.12204051e-01 -1.42533541e+00 -4.57010478e-01 -7.10730433e-01
-7.73288667e-01 1.63363330e-02 1.11674808e-01 -3.50854129e-01
1.09335518e+00 -2.15575457e+00 7.31367826e-01 9.53955829e-01
2.69435227e-01 3.68944858e-03 -1.21308481e-02 3.60579163e-01
1.49807259e-01 3.27058017e-01 -6.19093716e-01 -1.63717613e-01
1.92884296e-01 2.74826199e-01 1.25669479e-01 6.49008512e-01
1.00316018e-01 8.24743092e-01 -7.43803024e-01 1.30634019e-02
1.15453536e-02 -3.86629514e-02 -7.43639708e-01 1.11473566e-02
-2.38734484e-01 3.68590385e-01 -3.72175336e-01 -3.32713276e-02
9.70651090e-01 -8.87616277e-02 5.91027677e-01 -2.24276960e-01
-7.40830526e-02 5.44690609e-01 -1.54600060e+00 1.20823717e+00
-1.77984372e-01 2.98911780e-01 1.28913030e-01 -1.04478836e+00
8.08877349e-01 -9.86566767e-02 1.97399259e-01 -1.04694784e+00
4.05392170e-01 4.94500548e-01 3.49034101e-01 -9.40681025e-02
7.37018049e-01 -3.95878792e-01 -4.07347918e-01 6.05291367e-01
-1.03368454e-01 -7.35862553e-01 2.72457063e-01 2.87570417e-01
1.32891798e+00 -4.85650897e-01 -1.95113197e-02 -3.87497365e-01
4.55602378e-01 -1.22563355e-01 5.23118138e-01 1.01063573e+00
1.15175964e-03 4.75273669e-01 1.05475855e+00 -8.78143162e-02
-1.06988001e+00 -6.07376158e-01 -1.33396819e-01 1.93851620e-01
5.95387936e-01 -4.17159706e-01 -1.05834997e+00 -3.95805985e-01
-1.18762709e-01 6.07099771e-01 -5.22789061e-01 -6.10782444e-01
-4.19438422e-01 -1.26299489e+00 1.74556404e-01 -2.09453523e-01
2.12803051e-01 -7.12149203e-01 -5.19813299e-01 1.69912085e-01
-2.29108870e-01 -9.42317963e-01 -4.28438298e-02 6.81011200e-01
-5.44730544e-01 -8.66576254e-01 1.94742857e-03 -1.99372113e-01
8.66336823e-01 -2.42701977e-01 8.06894779e-01 3.60695034e-01
-1.91626415e-01 -2.14972287e-01 -2.14804307e-01 6.89302459e-02
-5.62520802e-01 1.43696457e-01 -2.98268106e-02 2.25447789e-02
1.04709283e-01 -1.93287224e-01 -3.48742098e-01 1.22509778e-01
-8.70199382e-01 -1.42455816e-01 2.78443009e-01 6.39564693e-01
5.57173789e-01 3.90702248e-01 -3.36507335e-02 -8.91585410e-01
5.58682442e-01 -2.75580436e-01 -1.14807761e+00 2.18304381e-01
-5.51186204e-01 6.55620456e-01 6.25298440e-01 1.57925129e-01
-3.29812080e-01 1.53090447e-01 -7.80119449e-02 2.20638394e-01
2.59994954e-01 4.52762842e-01 -3.03232074e-01 -6.08881176e-01
5.65590620e-01 -1.33581832e-01 -2.72341222e-01 7.83624575e-02
2.27725953e-01 5.63952088e-01 1.89105049e-01 -6.67613268e-01
9.25897479e-01 4.06017780e-01 6.09540164e-01 -4.82645333e-01
-5.66734254e-01 -1.07169546e-01 -3.08872312e-01 -1.30083665e-01
9.75531816e-01 -8.43655244e-02 -7.99694896e-01 3.84848982e-01
-9.91059363e-01 -2.77789265e-01 -1.42288670e-01 3.92821044e-01
-3.47117186e-01 2.79543102e-01 -4.58963871e-01 -9.83874679e-01
-6.90064654e-02 -1.51958597e+00 6.92365050e-01 5.19437671e-01
9.69609618e-02 -3.90503168e-01 2.44599029e-01 4.60874557e-01
1.49477541e-01 1.28486589e-01 1.12286830e+00 -3.20252746e-01
-8.99842024e-01 -1.97931185e-01 -2.22700253e-01 2.91184336e-01
-2.02813089e-01 -3.77899292e-03 -7.41099238e-01 -3.89214039e-01
1.93733513e-01 -2.08529338e-01 7.47527897e-01 2.54355937e-01
8.53838086e-01 -8.40351060e-02 -2.33351380e-01 5.50123990e-01
1.66434777e+00 1.84388742e-01 8.29473972e-01 2.00661123e-01
4.30174142e-01 5.53652346e-01 3.25067759e-01 -7.02908784e-02
1.76953688e-01 9.64896381e-01 5.14684439e-01 7.39108548e-02
4.11647826e-01 1.77633777e-01 1.69245139e-01 7.96069682e-01
-7.35918656e-02 -5.39288938e-01 -8.59712780e-01 2.29390770e-01
-1.75548184e+00 -4.95667338e-01 -5.26940763e-01 2.63298225e+00
7.16843724e-01 2.99791753e-01 -2.23527983e-01 2.57096440e-01
8.35855901e-01 -1.17215514e-01 -2.02656716e-01 -9.13164735e-01
-6.75639417e-03 6.11129105e-01 8.56106818e-01 8.90988946e-01
-6.94409728e-01 7.24293470e-01 5.45635223e+00 4.93707359e-01
-1.08105528e+00 3.60903628e-02 4.28225219e-01 2.93687843e-02
-4.78783637e-01 6.10748768e-01 -4.41212535e-01 5.39048016e-01
1.02194369e+00 8.90194923e-02 1.15746164e+00 1.02417767e-01
-1.05813555e-01 -5.14553428e-01 -1.26230955e+00 6.31205499e-01
-2.58732498e-01 -1.09712183e+00 -3.82186055e-01 4.41007525e-01
8.57325137e-01 -1.09802701e-01 -6.07941747e-02 -4.11237217e-02
1.52133659e-01 -9.90123868e-01 7.25313544e-01 3.42034370e-01
4.35464919e-01 -7.93254256e-01 8.97908568e-01 3.07509661e-01
-5.72393894e-01 6.74778968e-02 -4.46993232e-01 -3.45449597e-01
1.58101082e-01 9.29636657e-01 -4.54420418e-01 7.49104798e-01
3.48239183e-01 -2.93414712e-01 -1.26603693e-01 9.21967030e-01
-4.00705099e-01 3.66710514e-01 -6.57146990e-01 -3.90858561e-01
2.95081884e-01 -9.21447575e-01 6.12276733e-01 6.49224699e-01
1.38880357e-01 4.26583052e-01 -1.69917345e-01 1.27642584e+00
-3.12588632e-01 -3.62249874e-02 -3.61712456e-01 -1.73101470e-01
3.49502385e-01 1.22767270e+00 -9.66770530e-01 -1.19047500e-01
8.23405758e-02 9.70042467e-01 4.58363235e-01 2.89615870e-01
-6.94440424e-01 -5.71291566e-01 2.34919488e-01 -2.02953786e-01
7.30992556e-02 -2.56679188e-02 -6.46794140e-01 -9.51444626e-01
1.44431800e-01 -7.22908318e-01 1.12933077e-01 -5.03475606e-01
-8.48551929e-01 1.65699616e-01 -4.44355071e-01 -4.05015171e-01
1.36496753e-01 -4.99442250e-01 -4.17584777e-01 1.04733324e+00
-1.33079302e+00 -2.66667873e-01 -1.06493458e-01 4.38188314e-02
-6.90164506e-01 7.41288126e-01 6.40083313e-01 4.24266160e-01
-6.49175048e-01 3.52896035e-01 2.41276398e-01 -1.20964482e-01
3.97627592e-01 -1.26346004e+00 1.44220144e-01 1.03720796e+00
5.01096360e-02 7.31634915e-01 8.69868338e-01 -5.17443001e-01
-1.70310318e+00 -6.59701943e-01 1.01000357e+00 -3.23910147e-01
4.89489347e-01 -7.15735793e-01 -6.34351969e-01 3.50556642e-01
-8.93580317e-02 -1.83561131e-01 2.86825895e-01 1.55080035e-01
2.84304917e-02 -6.39974251e-02 -1.37061238e+00 2.23520100e-01
6.93789124e-01 -6.69526696e-01 -3.28444660e-01 4.84494895e-01
4.09544438e-01 -4.91282552e-01 -6.13193452e-01 3.31905454e-01
3.48861098e-01 -8.40198934e-01 5.70200384e-01 -4.04267937e-01
4.38395619e-01 -5.37783921e-01 -8.48857388e-02 -1.37032712e+00
-2.72381872e-01 -7.74273396e-01 3.96268278e-01 8.17514181e-01
5.39630473e-01 -8.28584135e-01 7.43275404e-01 7.22562134e-01
1.53659731e-02 -5.02328515e-01 -1.21401024e+00 -5.75010955e-01
1.95369989e-01 -4.12707664e-02 6.36564910e-01 8.48935783e-01
2.32144713e-01 3.91765207e-01 -2.21493259e-01 5.57958186e-01
5.89578211e-01 2.47080743e-01 6.69428229e-01 -9.87626612e-01
-5.41815341e-01 -3.61252367e-01 -4.86152321e-01 -7.00711727e-01
-1.00197986e-01 -8.68351340e-01 4.66776699e-01 -1.41897213e+00
3.59819710e-01 -8.63961279e-01 -1.61482677e-01 2.01439068e-01
-2.26808503e-01 2.57989943e-01 3.27735096e-01 -6.72916099e-02
-5.03124177e-01 2.56288260e-01 9.74134207e-01 -2.09376633e-01
-2.86809117e-01 -1.07500330e-01 -3.51030171e-01 1.89982519e-01
5.07376134e-01 -1.03220701e+00 2.39501864e-01 -3.51958990e-01
8.95256221e-01 1.66842341e-01 2.53732890e-01 -9.61322129e-01
4.02377874e-01 4.79481593e-02 -3.36456537e-01 -1.43854544e-01
2.86943793e-01 -6.29388273e-01 4.78019267e-01 8.21124792e-01
5.20001352e-02 -4.35280830e-01 -9.45180431e-02 2.92432100e-01
1.01058245e-01 -8.85208905e-01 9.50672686e-01 7.15110898e-02
-3.27457587e-04 -1.71752974e-01 -2.26626933e-01 -1.22787319e-01
9.21711385e-01 7.21534640e-02 -4.72451776e-01 -1.90670669e-01
-6.03862882e-01 2.61343688e-01 7.87820220e-01 -1.43678129e-01
-4.26638313e-02 -7.88167536e-01 -4.35108155e-01 1.38646781e-01
-1.18762456e-01 -6.47275969e-02 1.91589892e-02 9.56020534e-01
-6.37928605e-01 4.02778298e-01 -8.60130787e-02 -5.09136617e-01
-7.68938303e-01 3.77877533e-01 5.75773776e-01 -3.89481276e-01
-3.11068911e-03 7.39227653e-01 -2.49361411e-01 -6.57563269e-01
-3.17902446e-01 -3.58790398e-01 2.05323309e-01 -1.21370621e-01
-5.48320599e-02 3.06106657e-01 5.81864476e-01 -4.43925351e-01
-5.32281280e-01 6.85152769e-01 2.17182875e-01 -3.62920195e-01
1.15992701e+00 1.43730976e-02 -6.86842620e-01 -4.23035957e-03
9.72403824e-01 3.26921701e-01 -4.97009814e-01 1.80710956e-01
-2.05559596e-01 -2.01580495e-01 2.67081410e-01 -5.55465519e-01
-8.50795209e-01 7.07568347e-01 4.01270211e-01 7.19663441e-01
8.97876859e-01 -4.67330180e-02 7.20919073e-01 4.70427632e-01
6.89043581e-01 -1.17340386e+00 -4.61841732e-01 5.24946868e-01
5.25563471e-02 -9.47109580e-01 7.99978375e-02 -6.09500468e-01
-1.46646388e-02 1.00237346e+00 1.65077239e-01 -1.52766913e-01
1.67535633e-01 9.72106457e-02 -3.13338578e-01 -4.61010247e-01
-1.98739737e-01 -1.13985293e-01 7.88786784e-02 1.98818371e-02
-1.40728310e-01 4.82349306e-01 -8.30502570e-01 2.90461779e-01
-3.45827699e-01 -2.02826291e-01 7.92722046e-01 7.91120410e-01
-3.69028181e-01 -1.30299783e+00 -5.30471563e-01 3.37354690e-01
-1.23914301e-01 -2.68105865e-02 -6.70099676e-01 2.21843749e-01
2.65047073e-01 1.14265919e+00 3.88712287e-02 -4.49738383e-01
2.64870852e-01 -2.21570544e-02 7.02100456e-01 -5.69710672e-01
-6.44708395e-01 -3.82362843e-01 1.55478820e-01 -5.22144020e-01
-9.84966680e-02 -3.62184614e-01 -1.39715028e+00 -4.22081441e-01
-8.51876855e-01 5.73004723e-01 9.72521245e-01 1.18221784e+00
2.33709306e-01 5.67084730e-01 6.65059090e-01 -5.47320127e-01
-7.80719995e-01 -4.18119878e-01 -5.84560931e-01 1.90188751e-01
9.36306491e-02 -4.53673393e-01 -7.58136272e-01 -7.06146479e-01] | [5.631623268127441, 4.904772758483887] |
7f4d64e8-0497-4a78-8b25-a4e842a507f6 | deformation-aware-3d-model-embedding-and | 2004.01228 | null | https://arxiv.org/abs/2004.01228v3 | https://arxiv.org/pdf/2004.01228v3.pdf | Deformation-Aware 3D Model Embedding and Retrieval | We introduce a new problem of retrieving 3D models that are deformable to a given query shape and present a novel deep deformation-aware embedding to solve this retrieval task. 3D model retrieval is a fundamental operation for recovering a clean and complete 3D model from a noisy and partial 3D scan. However, given a finite collection of 3D shapes, even the closest model to a query may not be satisfactory. This motivates us to apply 3D model deformation techniques to adapt the retrieved model so as to better fit the query. Yet, certain restrictions are enforced in most 3D deformation techniques to preserve important features of the original model that prevent a perfect fitting of the deformed model to the query. This gap between the deformed model and the query induces asymmetric relationships among the models, which cannot be handled by typical metric learning techniques. Thus, to retrieve the best models for fitting, we propose a novel deep embedding approach that learns the asymmetric relationships by leveraging location-dependent egocentric distance fields. We also propose two strategies for training the embedding network. We demonstrate that both of these approaches outperform other baselines in our experiments with both synthetic and real data. Our project page can be found at https://deformscan2cad.github.io/. | ['Jingwei Huang', 'Tolga Birdal', 'Minhyuk Sung', 'Mikaela Angelina Uy', 'Leonidas Guibas'] | 2020-04-02 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/264_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123520392.pdf | eccv-2020-8 | ['3d-object-reconstruction'] | ['computer-vision'] | [-2.84259498e-01 -6.23987839e-02 -4.34004664e-02 -4.37466413e-01
-1.06650496e+00 -8.12070370e-01 4.93605614e-01 1.72330253e-02
-9.20760334e-02 8.59572440e-02 3.12305868e-01 1.96246594e-01
-4.04426903e-01 -8.22497010e-01 -9.22259033e-01 -5.76403797e-01
2.03110158e-01 9.12681997e-01 3.03239167e-01 -1.75285071e-01
4.00657505e-01 1.08210099e+00 -1.25605583e+00 -1.17168911e-01
5.67434132e-01 8.71336579e-01 3.46163630e-01 3.31569195e-01
-4.68503591e-03 -3.48825067e-01 -3.39297801e-01 -2.48129889e-01
6.62045956e-01 6.60835803e-02 -8.35660338e-01 3.13475914e-02
5.62049031e-01 -6.94195867e-01 -6.58905268e-01 9.39779282e-01
7.83670187e-01 1.71830028e-01 9.14048076e-01 -8.98737192e-01
-8.55556369e-01 -1.31248817e-01 -3.96771401e-01 -1.25330299e-01
6.12673521e-01 -2.13204667e-01 9.99556065e-01 -1.32136476e+00
1.03328633e+00 1.12543583e+00 3.81586581e-01 6.67734623e-01
-1.14279735e+00 -2.85104185e-01 -7.15156347e-02 1.42633274e-01
-1.67596877e+00 -4.55362439e-01 1.27883482e+00 -3.02320480e-01
6.39574289e-01 3.66562992e-01 6.82628930e-01 8.92951608e-01
4.35286611e-02 5.99842072e-01 6.01032794e-01 -3.45549941e-01
8.35667551e-02 1.23443753e-02 -8.85098130e-02 6.46838486e-01
1.98262319e-01 -3.89421247e-02 -5.62304854e-01 -3.05571288e-01
9.34953451e-01 2.43229777e-01 -3.61194968e-01 -1.12506783e+00
-1.15732896e+00 6.79969192e-01 7.57795215e-01 2.64353812e-01
-3.87233585e-01 2.30127811e-01 -1.49679244e-01 1.28412724e-01
5.67306161e-01 5.41061044e-01 -2.25884974e-01 9.49824825e-02
-7.52921700e-01 6.12073958e-01 5.92897117e-01 1.05828393e+00
9.60307598e-01 -2.99421459e-01 2.25895762e-01 6.59329295e-01
5.66700518e-01 6.25913382e-01 1.53699279e-01 -1.04921389e+00
3.07605565e-01 8.35177183e-01 3.65161747e-01 -1.31620479e+00
-8.96693766e-02 -1.91556528e-01 -5.33925891e-01 1.53099701e-01
2.52243996e-01 5.64540565e-01 -8.32729101e-01 1.55482268e+00
7.04598546e-01 1.70415901e-02 -1.30156815e-01 1.19232583e+00
8.25102210e-01 4.39979315e-01 -6.23196542e-01 3.21691126e-01
9.29629862e-01 -6.54774070e-01 -3.77458632e-01 -1.75538063e-02
6.68288469e-01 -9.41272378e-01 8.86437893e-01 -1.69378906e-01
-1.26491404e+00 -2.47589469e-01 -1.11375546e+00 -4.12552476e-01
-2.42015630e-01 -2.11376965e-01 1.85980260e-01 1.93437189e-01
-1.03947425e+00 7.90088177e-01 -1.00593746e+00 -3.70261550e-01
4.17687774e-01 2.82245994e-01 -6.65033877e-01 -4.41529810e-01
-9.39241648e-01 9.18289065e-01 2.82322317e-02 2.52037197e-01
-7.87742674e-01 -6.85786545e-01 -8.29545677e-01 -1.12097144e-01
3.61697465e-01 -8.17303061e-01 9.36454952e-01 -1.78169072e-01
-1.08740139e+00 1.14765525e+00 -3.28596085e-01 9.64459032e-02
5.96492350e-01 -2.23162502e-01 7.93727674e-03 2.48821750e-01
1.69709940e-02 5.59248328e-01 7.89289713e-01 -1.57508636e+00
2.35841155e-01 -7.11447835e-01 2.45679677e-01 4.15562481e-01
-9.25415605e-02 -2.54738718e-01 -7.79800892e-01 -5.73027134e-01
9.01860237e-01 -9.86900330e-01 -1.46469042e-01 6.64723337e-01
-3.34902823e-01 -2.01779217e-01 8.66764307e-01 -5.85829020e-01
8.15362215e-01 -2.09473085e+00 4.09734279e-01 2.88451076e-01
2.71645039e-01 4.65050265e-02 -2.34340191e-01 7.36826837e-01
6.48079515e-02 3.06726545e-01 -2.10018545e-01 -4.87464845e-01
2.83434778e-01 2.92165309e-01 -4.28900838e-01 6.84299350e-01
9.19144154e-02 1.09834170e+00 -7.19911337e-01 -4.22221005e-01
1.47167623e-01 7.77866125e-01 -6.30551577e-01 3.38818371e-01
-2.54951399e-02 3.45820665e-01 -8.95364404e-01 8.39873552e-01
1.01500642e+00 -7.91867748e-02 -3.44557226e-01 -5.70957601e-01
2.34913632e-01 2.81825244e-01 -1.32928050e+00 2.23604345e+00
-1.82619125e-01 1.56608194e-01 -4.93952930e-02 -1.01578164e+00
1.22621691e+00 1.34367242e-01 8.55056643e-01 -5.56244373e-01
-1.03235632e-01 5.08814454e-01 -4.49900270e-01 -4.88063842e-01
5.02244234e-01 7.78452605e-02 -1.95720326e-02 8.25737596e-01
-2.29369134e-01 -6.96375132e-01 -3.64057153e-01 4.31596041e-02
8.01239967e-01 3.28351736e-01 -1.14365101e-01 -1.40861511e-01
2.40103871e-01 -1.32077128e-01 4.67832088e-01 5.15603065e-01
-4.98985425e-02 1.03657341e+00 1.01592712e-01 -7.77794778e-01
-1.14921939e+00 -1.32681346e+00 -1.81231201e-01 3.53769869e-01
6.82464361e-01 -2.63181925e-01 -5.71052015e-01 -6.16388202e-01
2.24490106e-01 2.84387857e-01 -5.58053851e-01 -3.50316107e-01
-7.61008978e-01 -1.84515566e-01 7.91032761e-02 4.94478106e-01
1.70065358e-01 -6.51667476e-01 -4.82930064e-01 4.58651455e-03
-1.21669598e-01 -8.62236500e-01 -8.78605962e-01 -2.26506144e-01
-9.79698300e-01 -1.18316221e+00 -7.35490918e-01 -9.74369049e-01
9.21032250e-01 4.78762329e-01 9.77355301e-01 4.05921459e-01
-3.15863267e-02 6.07126713e-01 -4.09185797e-01 -1.18540779e-01
-3.67440522e-01 1.27762347e-01 1.87404335e-01 -1.40251592e-01
2.31630757e-01 -6.39206648e-01 -8.45234752e-01 5.00960231e-01
-1.14289594e+00 -7.84617662e-02 2.91311622e-01 5.90347946e-01
1.11922932e+00 -2.70061255e-01 1.86533093e-01 -2.38368049e-01
4.47775722e-01 -1.95618749e-01 -3.96975577e-01 3.30217451e-01
-3.58748794e-01 2.81879634e-01 1.13172323e-01 -4.78514373e-01
-4.59465981e-01 1.00633770e-01 -1.14447489e-01 -1.00975204e+00
4.77274172e-02 4.34386611e-01 -3.60299289e-01 -1.90856963e-01
4.90211278e-01 2.51021177e-01 2.06949249e-01 -9.09961581e-01
2.67003298e-01 4.41491842e-01 2.71399498e-01 -6.37778997e-01
1.20495021e+00 7.19088137e-01 1.04531713e-01 -5.42956948e-01
-7.25327253e-01 -3.21317852e-01 -1.12131119e+00 -4.09054682e-02
5.77515483e-01 -5.82155108e-01 -3.50752413e-01 3.33173811e-01
-1.35256374e+00 -1.08056858e-01 -3.57933879e-01 5.16899705e-01
-6.89701557e-01 4.81126368e-01 -2.24948272e-01 -2.94468760e-01
-4.35584694e-01 -1.32154036e+00 1.54005814e+00 -7.79301077e-02
-1.85193662e-02 -8.22684884e-01 2.33482540e-01 1.28583997e-01
3.69632959e-01 4.17814016e-01 9.27228689e-01 -4.66623545e-01
-8.94434094e-01 -7.09294438e-01 2.13200543e-02 7.94738680e-02
2.77644515e-01 -4.04569581e-02 -7.68865108e-01 -4.76995766e-01
1.22102700e-01 -1.59939587e-01 6.83060348e-01 1.46062642e-01
1.10216665e+00 -7.57043213e-02 -4.85095918e-01 8.78999829e-01
1.29696667e+00 -8.69099125e-02 5.34157038e-01 2.66801655e-01
7.71430492e-01 5.18470407e-01 5.36342084e-01 3.10451984e-01
3.49439383e-01 9.84986067e-01 6.92011237e-01 2.53117293e-01
-8.18432346e-02 -4.74728912e-01 -4.34690714e-03 1.09841824e+00
2.88642608e-02 -1.97427139e-01 -8.86476696e-01 4.71572995e-01
-1.64136982e+00 -5.90462685e-01 2.89121926e-01 2.45249796e+00
6.06123090e-01 -5.71706668e-02 -3.82898927e-01 -8.67062807e-02
5.01548171e-01 2.86720335e-01 -8.21099520e-01 -2.05962807e-01
2.37774309e-02 2.11802721e-01 9.29856151e-02 5.68900585e-01
-9.26400185e-01 8.20815325e-01 5.72654629e+00 5.84245086e-01
-1.10678756e+00 -7.45361447e-02 3.12006027e-02 -1.22249424e-01
-8.77794862e-01 8.02074373e-02 -6.32946789e-01 1.37939632e-01
3.46074641e-01 -3.74315917e-01 3.30453336e-01 6.49498701e-01
1.81201756e-01 2.33150020e-01 -1.39219093e+00 1.06548548e+00
2.25398913e-01 -1.40515900e+00 4.54988778e-01 1.90235004e-01
5.42615175e-01 -1.32945269e-01 1.22382678e-02 -6.51209652e-02
-3.29948485e-01 -8.87837827e-01 7.24370897e-01 1.01711345e+00
7.06911325e-01 -5.87945580e-01 3.88353944e-01 3.56854856e-01
-1.19457757e+00 4.35068309e-01 -6.13533199e-01 5.04857838e-01
2.91948736e-01 2.77370811e-01 -6.78935468e-01 7.09269822e-01
8.15111041e-01 6.81544781e-01 -5.55434167e-01 1.09809554e+00
5.62484302e-02 -4.06775996e-02 -5.12441814e-01 2.30503708e-01
3.12767178e-03 -2.82020748e-01 9.55215514e-01 4.41112876e-01
6.24125063e-01 -3.68765928e-02 -3.22980247e-02 1.15417695e+00
-2.23436177e-01 1.17402911e-01 -9.12763059e-01 1.32439524e-01
6.53443456e-01 9.15936530e-01 -4.83490050e-01 1.19856536e-01
-1.96133614e-01 1.08495402e+00 3.99483383e-01 3.85440975e-01
-6.09850347e-01 -2.61222899e-01 7.83973694e-01 4.02323961e-01
3.45580161e-01 -5.72331429e-01 1.90990455e-02 -1.14879179e+00
6.39797807e-01 -5.05577326e-01 9.81151760e-02 -9.23666596e-01
-1.19771981e+00 5.06786048e-01 1.78620487e-01 -1.55679011e+00
-2.20795453e-01 -3.91528547e-01 -5.11515379e-01 9.08715963e-01
-1.44798279e+00 -1.26135302e+00 -4.06791806e-01 4.77077037e-01
2.99858093e-01 1.80456117e-01 9.21749353e-01 2.24842340e-01
-7.93145373e-02 4.56936777e-01 1.37062222e-01 1.17076682e-02
8.97585630e-01 -1.10384834e+00 4.31266874e-01 5.14895797e-01
3.23625058e-01 6.15340829e-01 4.15160596e-01 -5.04245162e-01
-1.85275078e+00 -9.12539124e-01 9.63125587e-01 -6.64942265e-01
2.04710960e-01 -4.21195477e-01 -1.29472971e+00 6.61449909e-01
-2.56425202e-01 3.32704037e-01 2.75827706e-01 -4.03550953e-01
-4.30686206e-01 -1.62557691e-01 -1.14114249e+00 5.92876554e-01
1.27516937e+00 -6.79579139e-01 -7.38009453e-01 3.74144435e-01
7.92664289e-01 -7.26408243e-01 -1.11009824e+00 4.86669123e-01
6.29636168e-01 -7.07635999e-01 1.29874134e+00 -5.88950455e-01
1.58175409e-01 -3.45335633e-01 -5.49836099e-01 -1.27205777e+00
-4.97077070e-02 -6.28199279e-01 -3.85507882e-01 7.70593762e-01
6.87938705e-02 -4.96048212e-01 8.99150848e-01 8.78788948e-01
-2.84942716e-01 -1.13624835e+00 -1.17623317e+00 -8.26198697e-01
4.99677241e-01 -3.02420944e-01 9.87215042e-01 8.31156373e-01
-6.24360204e-01 -2.30237216e-01 -9.62723345e-02 4.02170062e-01
6.11141801e-01 5.70578098e-01 9.49004829e-01 -1.21322417e+00
6.94528744e-02 -3.72620583e-01 -7.06421971e-01 -1.42350411e+00
2.30890483e-01 -1.17308939e+00 -2.30431527e-01 -1.71284652e+00
-8.58759508e-03 -6.30920947e-01 -8.79281834e-02 1.92249298e-01
2.85310894e-01 2.37526089e-01 1.47756664e-02 4.27411914e-01
-2.82206655e-01 7.75538623e-01 1.71743274e+00 -2.11671233e-01
-2.99250305e-01 -8.37110728e-02 -5.14192700e-01 4.76069182e-01
6.67486846e-01 -5.77116847e-01 -2.66756117e-01 -9.76875365e-01
2.67127186e-01 -4.12293486e-02 5.57698309e-01 -5.65171838e-01
2.74971008e-01 -1.19483799e-01 3.26604366e-01 -9.21779096e-01
8.00602198e-01 -9.51604664e-01 3.03140044e-01 1.85127616e-01
-1.89495087e-01 2.31383458e-01 -2.44195182e-02 6.18034601e-01
-9.13351327e-02 -3.83493006e-01 5.28935552e-01 -1.17707133e-01
-4.02644813e-01 8.47257257e-01 3.40138882e-01 5.46777397e-02
9.09637332e-01 -4.58740354e-01 1.03939913e-01 -3.62317026e-01
-9.06137347e-01 1.52492300e-01 9.97632265e-01 6.35109186e-01
1.16582584e+00 -1.80757368e+00 -7.02687621e-01 3.94103140e-01
3.28122079e-01 5.82274139e-01 1.64814055e-01 6.38737381e-01
-6.10311925e-01 2.32624620e-01 1.33709729e-01 -6.69966400e-01
-9.71661508e-01 4.72178161e-01 6.71149492e-01 1.08833298e-01
-7.02569783e-01 5.20383239e-01 8.30356404e-02 -7.95922637e-01
1.25791982e-01 -3.39869738e-01 1.55180067e-01 -1.53928161e-01
2.34521195e-01 2.08368599e-01 2.40535200e-01 -8.84020329e-01
-4.73274589e-01 1.14666975e+00 -3.69288623e-02 1.00447759e-01
1.68521404e+00 -1.98605448e-01 -5.37281483e-02 2.60062039e-01
1.66792011e+00 6.03892282e-02 -1.16993177e+00 -5.20835876e-01
-1.78998858e-01 -8.47780466e-01 -1.92756504e-02 -1.61698878e-01
-1.08397973e+00 8.46490979e-01 5.60398102e-01 -1.70687642e-02
7.68956423e-01 4.47401822e-01 9.56060529e-01 6.19995892e-01
3.93878222e-01 -8.30182910e-01 3.27957630e-01 2.78321177e-01
1.47711301e+00 -1.13200378e+00 1.00663722e-01 -2.02315867e-01
-1.12875901e-01 1.24803329e+00 4.10131574e-01 -3.77651930e-01
1.00838459e+00 -1.97166786e-01 -1.83811262e-02 -5.66469193e-01
-3.74929070e-01 4.10005413e-02 6.33926272e-01 5.24541378e-01
-8.41738135e-02 -8.85702670e-02 7.15886503e-02 2.60545313e-01
-7.48960152e-02 -4.28650767e-01 2.56849408e-01 1.04120159e+00
-2.72262216e-01 -1.26766539e+00 -3.07556033e-01 2.49723941e-01
-5.82018774e-03 2.90576279e-01 -5.44502199e-01 8.45950305e-01
-3.39806199e-01 3.63324165e-01 1.67370588e-01 -3.38948727e-01
6.89486384e-01 1.08453982e-01 6.79690540e-01 -5.23647070e-01
1.11245155e-01 1.64506316e-01 -5.43786049e-01 -7.30157971e-01
-5.06691217e-01 -6.15887046e-01 -1.14210582e+00 -2.53996640e-01
-3.83936644e-01 -1.77673608e-01 5.75428545e-01 5.79914808e-01
6.47068143e-01 -2.88920224e-01 9.04302537e-01 -1.25596428e+00
-7.85325706e-01 -6.05239630e-01 -5.99261463e-01 7.81512618e-01
2.98620582e-01 -1.03410864e+00 -4.91313487e-01 -2.02566177e-01] | [8.270700454711914, -3.411470890045166] |
83eff9f6-4714-4d36-a7fc-39d41287fbba | self-supervised-text-independent-speaker | 2012.07178 | null | https://arxiv.org/abs/2012.07178v2 | https://arxiv.org/pdf/2012.07178v2.pdf | Self-supervised Text-independent Speaker Verification using Prototypical Momentum Contrastive Learning | In this study, we investigate self-supervised representation learning for speaker verification (SV). First, we examine a simple contrastive learning approach (SimCLR) with a momentum contrastive (MoCo) learning framework, where the MoCo speaker embedding system utilizes a queue to maintain a large set of negative examples. We show that better speaker embeddings can be learned by momentum contrastive learning. Next, alternative augmentation strategies are explored to normalize extrinsic speaker variabilities of two random segments from the same speech utterance. Specifically, augmentation in the waveform largely improves the speaker representations for SV tasks. The proposed MoCo speaker embedding is further improved when a prototypical memory bank is introduced, which encourages the speaker embeddings to be closer to their assigned prototypes with an intermediate clustering step. In addition, we generalize the self-supervised framework to a semi-supervised scenario where only a small portion of the data is labeled. Comprehensive experiments on the Voxceleb dataset demonstrate that our proposed self-supervised approach achieves competitive performance compared with existing techniques, and can approach fully supervised results with partially labeled data. | ['Dong Yu', 'Meng Yu', 'Chao Weng', 'Chunlei Zhang', 'Wei Xia'] | 2020-12-13 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 8.03013891e-02 2.72943497e-01 -4.16816771e-01 -7.94504106e-01
-1.00436580e+00 -3.49544495e-01 6.87484801e-01 1.20092565e-02
-2.97238261e-01 4.29354757e-01 4.31250334e-01 -2.30035990e-01
1.22832991e-01 -1.89447820e-01 -5.42053640e-01 -9.18633699e-01
-4.10350300e-02 5.16323566e-01 -1.26254216e-01 -2.59820849e-01
-1.73174031e-02 3.84129792e-01 -1.50037241e+00 1.31865188e-01
7.87546933e-01 7.95628667e-01 -6.32549776e-03 7.16461182e-01
-3.08629811e-01 7.24715233e-01 -5.28477550e-01 -2.54141390e-01
1.07385918e-01 -5.51933408e-01 -7.29337096e-01 2.52057314e-01
4.47127640e-01 1.21281758e-01 -1.57285288e-01 1.08614111e+00
6.36443675e-01 3.44157159e-01 8.11050832e-01 -1.50203598e+00
-7.03593969e-01 1.16681862e+00 -6.32576048e-01 2.97719091e-01
2.73868181e-02 -2.46688407e-02 1.25267124e+00 -1.27766740e+00
1.02648042e-01 1.37019217e+00 5.02771735e-01 8.64069223e-01
-1.23230147e+00 -9.21292603e-01 5.18284976e-01 3.45536053e-01
-1.44776595e+00 -9.40901995e-01 1.27902055e+00 -2.77411491e-01
6.47139192e-01 2.75781870e-01 3.70730430e-01 1.00142097e+00
-6.34760439e-01 1.06295252e+00 8.18190217e-01 -6.60421252e-01
4.40971434e-01 7.92889118e-01 6.31597340e-01 5.68794727e-01
-3.24549615e-01 2.10544258e-01 -8.42662692e-01 -4.36816812e-01
2.38380395e-02 -1.35590985e-01 -3.13349187e-01 -5.76417923e-01
-9.64428663e-01 1.05145216e+00 2.45366678e-01 3.30357581e-01
-2.95911819e-01 -1.65350482e-01 4.60876405e-01 3.56861979e-01
5.66520929e-01 4.57650656e-03 -4.30819094e-01 -6.13221042e-02
-1.18858933e+00 -2.58181602e-01 6.94805503e-01 7.06706464e-01
8.23835433e-01 5.43740571e-01 -1.60611838e-01 1.10354483e+00
6.12390101e-01 4.08739746e-01 9.53369200e-01 -5.15053034e-01
4.41175282e-01 2.69684374e-01 -2.40159929e-01 -4.20043558e-01
-4.14518826e-02 -5.43011189e-01 -6.89964235e-01 4.10922617e-02
2.57337205e-02 -1.01557277e-01 -8.21903586e-01 1.88236678e+00
3.82036775e-01 6.38116121e-01 5.06295383e-01 7.41831064e-01
9.84344184e-01 7.65945017e-01 6.80549219e-02 -5.00915825e-01
9.63271022e-01 -1.32456005e+00 -6.84444606e-01 -1.95695356e-01
3.56184185e-01 -5.23946345e-01 1.14090431e+00 2.24193707e-01
-9.35811341e-01 -6.24227703e-01 -1.10757053e+00 3.76807034e-01
-1.41822442e-01 2.21431553e-01 2.39472270e-01 1.02905619e+00
-1.14375496e+00 3.72275531e-01 -8.65670443e-01 -1.72242776e-01
4.31427211e-01 3.31884205e-01 -7.43011534e-02 2.90664554e-01
-9.99987543e-01 5.09281337e-01 1.80225655e-01 1.33021519e-01
-9.65869069e-01 -6.97846770e-01 -1.01645362e+00 1.76931083e-01
-4.26543541e-02 -4.16199826e-02 1.22036052e+00 -1.18330395e+00
-1.89018595e+00 7.75119483e-01 -5.75278938e-01 -6.49375439e-01
2.82884151e-01 6.40131757e-02 -5.34287691e-01 1.03478394e-01
-3.12020004e-01 6.97322965e-01 1.36214626e+00 -1.34603429e+00
-3.91192585e-01 -2.09585175e-01 -3.60018671e-01 2.51423717e-01
-8.84015441e-01 1.83292001e-01 -5.61732240e-02 -5.72274327e-01
1.57907352e-01 -8.07201862e-01 -4.93690409e-02 -3.01342666e-01
-4.90541309e-01 -5.22429049e-01 7.61785805e-01 -5.05309165e-01
1.07747114e+00 -2.54028368e+00 1.37778252e-01 4.62363243e-01
1.11240871e-01 3.01461428e-01 -3.16561490e-01 9.48327929e-02
-4.56992239e-01 -3.32258418e-02 -3.28806072e-01 -8.48297536e-01
6.57700896e-02 8.90884325e-02 -4.94012326e-01 5.50892353e-01
3.16836029e-01 5.56579649e-01 -6.84354305e-01 -5.61442494e-01
-3.05823013e-02 5.42498410e-01 -6.33003712e-01 3.35771590e-01
2.04132989e-01 2.85692424e-01 4.41730395e-02 6.03174686e-01
8.01144361e-01 -1.82404160e-01 1.80405557e-01 -7.06027374e-02
1.29061610e-01 5.40544271e-01 -1.22120965e+00 1.28360796e+00
-3.38044554e-01 5.47954738e-01 1.84603065e-01 -1.22842717e+00
1.23023880e+00 4.47382808e-01 2.32981414e-01 -5.91707304e-02
7.28354752e-02 1.15239825e-02 3.32693234e-02 -1.25283003e-01
4.57593441e-01 -3.50641221e-01 3.03045511e-01 6.30392492e-01
2.93075442e-01 2.15494245e-01 -1.88915148e-01 2.02406898e-01
5.18915355e-01 -2.99013287e-01 2.76852936e-01 -1.84478551e-01
8.97507131e-01 -3.82041752e-01 7.76525140e-01 5.48884809e-01
-5.79522491e-01 5.51288962e-01 2.25507319e-01 1.17931627e-01
-5.96111894e-01 -1.17760420e+00 -2.55403787e-01 1.49007380e+00
-4.16731127e-02 -2.08125144e-01 -6.37292445e-01 -9.12583888e-01
-6.84401467e-02 8.54234815e-01 -5.82827151e-01 -2.74889618e-01
-5.99783957e-01 -8.54465306e-01 5.37524343e-01 7.48276949e-01
3.34201604e-02 -9.35398459e-01 7.25624487e-02 4.53433655e-02
4.11169678e-02 -6.73929453e-01 -7.25452781e-01 4.39759761e-01
-8.82048666e-01 -5.69534063e-01 -7.30838299e-01 -1.23049426e+00
8.04916024e-01 2.15310991e-01 6.02990806e-01 -1.65785164e-01
2.29081944e-01 4.05631602e-01 -2.95485109e-01 -2.21560523e-01
-7.35180080e-01 -1.06418272e-02 5.99989116e-01 4.71166968e-01
6.12312317e-01 -5.09235442e-01 -1.76449597e-01 2.25435272e-01
-4.51895714e-01 -3.39299411e-01 3.02492619e-01 1.26173007e+00
4.00764912e-01 -3.32105875e-01 1.07800221e+00 -6.63836360e-01
5.45639396e-01 -5.23757517e-01 -3.16910952e-01 2.45300487e-01
-7.69926369e-01 1.65320769e-01 5.64731956e-01 -9.19119477e-01
-1.09931374e+00 1.56799585e-01 -2.40758494e-01 -7.58781493e-01
-5.57694025e-02 3.62441659e-01 -1.58436254e-01 2.22515054e-02
7.12295830e-01 4.88712281e-01 4.06066716e-01 -3.72599125e-01
4.59640443e-01 1.05648589e+00 3.06118935e-01 -3.97218764e-01
1.03612661e+00 2.86054194e-01 -8.07444274e-01 -9.89752412e-01
-6.34049177e-01 -6.97706580e-01 -4.60826248e-01 -1.45031765e-01
3.88225615e-01 -1.03599298e+00 -6.48522615e-01 1.10082380e-01
-8.73064637e-01 -2.31431335e-01 -6.00482762e-01 8.19481373e-01
-3.52554858e-01 4.14164633e-01 -5.86378872e-01 -1.26993835e+00
-4.14642274e-01 -9.28745389e-01 7.56652713e-01 2.90650249e-01
-1.83611616e-01 -9.21910286e-01 3.53073180e-01 2.81072080e-01
3.18847716e-01 -6.83457196e-01 6.33363068e-01 -1.45306480e+00
1.68465953e-02 7.63662672e-03 2.27247015e-01 8.66384387e-01
3.08505476e-01 -1.10816337e-01 -1.58403146e+00 -5.63108206e-01
1.04158968e-01 -5.05808473e-01 9.77034569e-01 2.38656089e-01
9.03107107e-01 -3.28300685e-01 -1.80073559e-01 4.83455420e-01
7.26756215e-01 1.32677212e-01 9.72164422e-02 8.69946927e-02
5.92397749e-01 7.11819172e-01 4.59208816e-01 2.65004367e-01
3.92608494e-01 3.82675588e-01 7.13791996e-02 -2.77537823e-01
-5.67348301e-02 -2.51023501e-01 7.51175582e-01 1.46094739e+00
2.59920180e-01 4.65077497e-02 -7.33370006e-01 5.84061444e-01
-1.62577724e+00 -1.05484676e+00 4.25315410e-01 2.24600029e+00
8.93731773e-01 1.16027391e-03 3.44842404e-01 4.02824879e-01
1.15483427e+00 4.31800693e-01 -6.22537315e-01 -3.11522245e-01
-2.39076987e-01 1.07565679e-01 -3.50137427e-02 6.86021864e-01
-1.09595358e+00 8.96088719e-01 6.62555361e+00 5.95996261e-01
-1.42602682e+00 3.48359197e-01 5.86839080e-01 -2.24468261e-01
-3.01395655e-01 -2.41182223e-01 -9.86643076e-01 3.60832483e-01
1.17542982e+00 -3.42562973e-01 3.05122524e-01 1.14454198e+00
1.23311415e-01 5.95028937e-01 -1.30704999e+00 1.21833587e+00
6.11520469e-01 -1.10514462e+00 -1.61013395e-01 -1.11179896e-01
6.27363741e-01 -1.50276104e-03 4.18554783e-01 6.31962061e-01
2.43526086e-01 -6.81158662e-01 6.79436624e-01 1.66919246e-01
4.49300230e-01 -8.35159719e-01 6.29308045e-01 3.33240926e-01
-1.10659206e+00 -2.48524934e-01 -4.21332836e-01 3.96358520e-01
2.29402229e-01 2.62074649e-01 -1.20583165e+00 2.67214060e-01
5.92588723e-01 4.97780770e-01 -5.83145678e-01 8.30829501e-01
-1.10118128e-01 1.23775744e+00 -2.08260715e-01 -1.36443287e-01
-6.87613934e-02 -3.55743282e-02 6.02513611e-01 1.21889031e+00
6.42612278e-02 -2.58778632e-01 1.21125177e-01 7.12518811e-01
-2.54891932e-01 4.64050353e-01 -1.79687396e-01 -3.36322226e-02
6.40750587e-01 1.17177296e+00 -2.15613693e-01 -6.47868097e-01
-3.06610465e-01 8.55965137e-01 5.51699221e-01 5.44316053e-01
-6.81493223e-01 -1.40709504e-01 6.00412011e-01 -1.60604164e-01
5.11739254e-01 2.08095536e-01 -5.97598962e-02 -1.22746849e+00
-1.21693552e-01 -7.96553016e-01 4.61057484e-01 -4.22337174e-01
-1.55668247e+00 8.57886612e-01 -2.49671698e-01 -1.37738657e+00
-5.93689203e-01 -1.80688977e-01 -1.00429296e+00 8.66005599e-01
-1.92133260e+00 -1.02542365e+00 7.46959224e-02 6.37173712e-01
6.08321846e-01 -7.16922283e-01 8.47099543e-01 1.18338756e-01
-7.19169259e-01 1.11416364e+00 2.06132978e-01 2.22231343e-01
8.68589818e-01 -1.26348853e+00 2.23662391e-01 7.83067644e-01
4.59951341e-01 7.79275239e-01 5.60304284e-01 -3.38184685e-01
-1.08077586e+00 -1.06337476e+00 8.92854273e-01 -7.76241198e-02
6.48928881e-01 -4.27758813e-01 -1.23122036e+00 7.61408508e-01
2.67310411e-01 2.88853645e-01 1.22601533e+00 4.48364735e-01
-5.55141389e-01 -3.48460346e-01 -1.14806402e+00 3.50176752e-01
5.17649114e-01 -8.49283874e-01 -1.00914800e+00 2.86506563e-01
6.42359853e-01 -4.80373316e-02 -4.52109307e-01 1.58286840e-01
3.00116569e-01 -5.71109414e-01 8.25730205e-01 -8.01444113e-01
-1.73114166e-01 -3.60658973e-01 -1.05857894e-01 -1.45720780e+00
-1.56862244e-01 -6.83246434e-01 -4.35327649e-01 1.67114484e+00
6.53087795e-01 -6.89033031e-01 9.95505393e-01 1.88949361e-01
-1.60710186e-01 -5.20642340e-01 -1.07544589e+00 -9.80356753e-01
1.14651278e-01 -4.82385308e-01 6.03149712e-01 1.25859284e+00
3.30274522e-01 4.38534528e-01 -3.91413897e-01 4.29041207e-01
7.52887130e-01 1.11057349e-01 6.43839657e-01 -1.14218974e+00
-5.00700295e-01 -2.65337020e-01 -4.03059870e-01 -1.11633730e+00
6.74840152e-01 -1.28495800e+00 3.15316230e-01 -8.57947826e-01
3.09328228e-01 -4.40593749e-01 -7.53087342e-01 3.42400938e-01
-4.30161864e-01 1.93603590e-01 5.12771420e-02 2.20924899e-01
-4.24564838e-01 9.47105348e-01 5.07742763e-01 -4.80431616e-01
-6.81329966e-01 2.35556364e-01 -6.92929745e-01 6.76880538e-01
8.52505744e-01 -4.83338356e-01 -3.56728256e-01 3.51732299e-02
-6.52613938e-01 -8.60876292e-02 1.07664406e-01 -6.84043586e-01
4.74786669e-01 2.12538972e-01 1.48101032e-01 -6.53496206e-01
5.43072343e-01 -3.92457724e-01 -3.99588972e-01 2.43335024e-01
-7.45334148e-01 -2.96750337e-01 2.12000702e-02 5.22161067e-01
-4.30279255e-01 -3.52382064e-01 1.01962101e+00 4.02262628e-01
-2.72658229e-01 2.05670625e-01 -3.89996439e-01 -1.02040693e-01
8.06699395e-01 -5.67049272e-02 3.58513966e-02 -4.34291959e-01
-1.02760422e+00 2.53852427e-01 4.99554910e-02 4.38200474e-01
7.14418173e-01 -1.52936137e+00 -8.38952482e-01 3.52929771e-01
2.35724062e-01 -3.54517460e-01 2.98589647e-01 8.55018556e-01
1.86484098e-01 1.62830189e-01 2.76249617e-01 -9.57492471e-01
-1.52560568e+00 5.24921358e-01 1.79253727e-01 1.27444729e-01
-4.00204450e-01 1.08854306e+00 2.69646168e-01 -8.49202454e-01
6.53421938e-01 -1.08753853e-01 -4.06909466e-01 2.51787752e-01
8.24704885e-01 2.90462494e-01 1.03603804e-03 -9.57687080e-01
-5.02779186e-01 1.54570654e-01 -5.32685697e-01 -3.15982938e-01
1.41802609e+00 -1.18390135e-01 2.42394865e-01 7.88599193e-01
1.24891841e+00 3.97314042e-01 -1.30904174e+00 -5.66349387e-01
-4.09990512e-02 -1.12963729e-01 8.92347097e-02 -3.43347281e-01
-1.15928960e+00 9.57607985e-01 8.79776001e-01 1.04927406e-01
8.71245682e-01 2.52449334e-01 3.95352006e-01 4.31954712e-01
-2.04857111e-01 -1.14021635e+00 2.56276488e-01 2.81345218e-01
6.87210500e-01 -1.34363222e+00 -2.86765009e-01 -2.86184043e-01
-1.03942156e+00 9.98892367e-01 5.44873834e-01 1.39161736e-01
8.39028060e-01 1.97412223e-01 3.69339257e-01 2.17465192e-01
-7.23055482e-01 -1.21650293e-01 1.88798666e-01 6.01204872e-01
3.92452002e-01 9.14805532e-02 1.34492546e-01 7.99194038e-01
-3.16952407e-01 -6.01250410e-01 3.64013612e-01 7.85524905e-01
-4.74146038e-01 -1.00765014e+00 -5.72851360e-01 3.10464501e-01
-1.11432612e-01 -2.80466974e-01 -2.08980769e-01 3.38215679e-01
-4.17316824e-01 1.04681706e+00 2.25003377e-01 -3.46651345e-01
8.14463049e-02 5.90740085e-01 1.51819319e-01 -7.65911400e-01
-6.20383739e-01 3.03522676e-01 -1.95072860e-01 -3.27093676e-02
-4.58339512e-01 -9.24292028e-01 -1.26788986e+00 1.07124463e-01
-6.25310004e-01 5.22085190e-01 7.95414090e-01 8.54365706e-01
1.33238807e-01 4.86000836e-01 1.24679172e+00 -7.63877571e-01
-1.01885736e+00 -1.17539835e+00 -6.36107028e-01 1.26854718e-01
6.96582913e-01 -6.16942406e-01 -1.02359414e+00 2.17318371e-01] | [14.341720581054688, 6.178313732147217] |
d93a6ed3-b4a8-4153-9025-8f56cfc1d424 | emailsum-abstractive-email-thread | 2107.14691 | null | https://arxiv.org/abs/2107.14691v1 | https://arxiv.org/pdf/2107.14691v1.pdf | EmailSum: Abstractive Email Thread Summarization | Recent years have brought about an interest in the challenging task of summarizing conversation threads (meetings, online discussions, etc.). Such summaries help analysis of the long text to quickly catch up with the decisions made and thus improve our work or communication efficiency. To spur research in thread summarization, we have developed an abstractive Email Thread Summarization (EmailSum) dataset, which contains human-annotated short (<30 words) and long (<100 words) summaries of 2549 email threads (each containing 3 to 10 emails) over a wide variety of topics. We perform a comprehensive empirical study to explore different summarization techniques (including extractive and abstractive methods, single-document and hierarchical models, as well as transfer and semisupervised learning) and conduct human evaluations on both short and long summary generation tasks. Our results reveal the key challenges of current abstractive summarization models in this task, such as understanding the sender's intent and identifying the roles of sender and receiver. Furthermore, we find that widely used automatic evaluation metrics (ROUGE, BERTScore) are weakly correlated with human judgments on this email thread summarization task. Hence, we emphasize the importance of human evaluation and the development of better metrics by the community. Our code and summary data have been made available at: https://github.com/ZhangShiyue/EmailSum | ['Mohit Bansal', 'Jianfeng Gao', 'Asli Celikyilmaz', 'Shiyue Zhang'] | 2021-07-30 | null | https://aclanthology.org/2021.acl-long.537 | https://aclanthology.org/2021.acl-long.537.pdf | acl-2021-5 | ['email-thread-summarization'] | ['natural-language-processing'] | [ 4.62667108e-01 2.92658538e-01 -3.03234190e-01 -1.49519667e-01
-1.23027420e+00 -6.97436869e-01 9.54662740e-01 6.49410069e-01
-2.48634607e-01 1.05072522e+00 1.04656589e+00 -3.09698462e-01
2.11714476e-01 -2.23629773e-01 -1.50120929e-01 -2.68169552e-01
1.17537491e-01 4.86047804e-01 4.09073122e-02 -2.28401572e-01
8.63992870e-01 -2.92566046e-02 -1.16357040e+00 5.48130572e-01
1.10078931e+00 5.36605358e-01 1.79972589e-01 1.23256958e+00
-4.41113800e-01 8.52943540e-01 -1.29968011e+00 -7.98152328e-01
-4.80195642e-01 -7.45109081e-01 -1.22627258e+00 1.66846216e-01
4.56547290e-01 -1.81472927e-01 -1.37657270e-01 7.00913489e-01
7.55027235e-01 1.14393599e-01 7.07292318e-01 -1.19184840e+00
-3.74946594e-01 1.12435675e+00 -6.38182580e-01 5.54488599e-01
8.44798326e-01 6.35555433e-03 1.39581096e+00 -5.96671999e-01
5.89562595e-01 1.14816356e+00 6.86133921e-01 5.00491440e-01
-9.57627356e-01 -5.28480947e-01 5.86277582e-02 -4.15804498e-02
-5.86236060e-01 -6.74495816e-01 6.27345026e-01 -4.02235836e-01
9.34987545e-01 6.34351909e-01 3.61506552e-01 1.43791366e+00
3.20143878e-01 1.11320841e+00 8.30300868e-01 -4.33034778e-01
-1.13347790e-03 1.31626025e-01 7.81028569e-01 3.46837431e-01
3.69205236e-01 -7.70885766e-01 -8.00562024e-01 -7.40977108e-01
1.12404950e-01 -2.30211377e-01 -3.35908920e-01 4.93205339e-01
-1.45097029e+00 9.34780836e-01 -6.65269569e-02 2.86003262e-01
-5.15790343e-01 -3.98538960e-03 9.19487000e-01 2.82539904e-01
8.44578087e-01 7.03489006e-01 -2.85572946e-01 -6.40715659e-01
-9.21380341e-01 5.84794044e-01 1.40930581e+00 9.12258565e-01
6.38278127e-01 -3.36700439e-01 -6.62322640e-01 1.00447464e+00
-1.07532471e-01 2.66321242e-01 5.56732237e-01 -1.09107172e+00
8.22778463e-01 5.58884084e-01 2.13628784e-01 -9.19760823e-01
-3.76188606e-01 -3.34432632e-01 -9.71481562e-01 -6.35215461e-01
1.04319371e-01 -4.44402397e-01 -1.72542170e-01 1.33861840e+00
-5.98905124e-02 -1.70662180e-01 2.75565311e-02 2.52469689e-01
1.16518974e+00 9.12379861e-01 -1.71670020e-02 -6.53033257e-01
1.67988563e+00 -1.18370378e+00 -9.83534098e-01 -3.27620298e-01
6.67569041e-01 -1.23939717e+00 1.01027262e+00 1.74142599e-01
-1.22392547e+00 -4.31068778e-01 -8.40342224e-01 -5.44200391e-02
4.46900725e-02 1.79832727e-01 5.86651266e-01 4.72882599e-01
-9.33179677e-01 6.27725959e-01 -5.77498138e-01 -5.69195986e-01
3.15305978e-01 -1.61199756e-02 -1.07321506e-02 2.69817650e-01
-1.10060263e+00 6.94146454e-01 7.34770000e-02 -1.43122524e-01
-2.08896965e-01 -5.80086231e-01 -5.92893600e-01 2.31372640e-02
2.85114050e-01 -9.60729361e-01 1.93540895e+00 -6.09446526e-01
-1.42424548e+00 6.44229352e-01 -5.40454209e-01 -4.16639924e-01
4.94513422e-01 -5.08458555e-01 9.92161706e-02 2.43174896e-01
5.02248287e-01 2.58565038e-01 5.84765375e-01 -9.63782907e-01
-7.37019002e-01 -1.74132884e-01 -2.10994199e-01 3.27640414e-01
-3.79254669e-01 2.62339681e-01 -2.23489255e-02 -7.59104729e-01
-3.38474512e-01 -8.74659419e-01 -3.82375158e-02 -9.35616493e-01
-7.72095263e-01 -7.87149191e-01 6.69442117e-01 -9.71373439e-01
1.53150356e+00 -1.71810007e+00 7.56034553e-02 -3.30290884e-01
4.19921488e-01 1.14065997e-01 -9.86988172e-02 1.11226058e+00
2.80775815e-01 3.68591487e-01 -2.06570551e-01 -7.71060169e-01
-4.13308106e-02 -1.62943140e-01 -5.62360942e-01 -1.33391633e-03
-2.33882606e-01 9.85638738e-01 -9.40436304e-01 -6.35019422e-01
-2.90432304e-01 -9.89226326e-02 -2.17968270e-01 3.90749484e-01
-2.12477505e-01 3.23710382e-01 -6.41980708e-01 1.97754338e-01
2.26369932e-01 -5.03727317e-01 -2.45024860e-02 8.84309188e-02
-1.04229808e-01 8.96692097e-01 -5.96089303e-01 1.41772115e+00
-4.91438568e-01 1.08756769e+00 -5.93206696e-02 -5.36641419e-01
6.33810461e-01 4.57668334e-01 3.22755456e-01 -2.15519726e-01
1.00655958e-01 6.92298785e-02 -2.03731105e-01 -3.62251610e-01
1.22776377e+00 1.61455303e-01 -4.27190423e-01 1.27897060e+00
-1.93239912e-01 -1.86630100e-01 5.87819278e-01 9.13268149e-01
1.24465287e+00 -4.35803503e-01 5.87288201e-01 -1.75145585e-02
4.57213879e-01 1.52007351e-02 1.19725578e-01 1.18229353e+00
-1.59432068e-01 6.12165928e-01 9.35980141e-01 -2.35464554e-02
-1.00126719e+00 -6.45926714e-01 2.71987617e-01 1.38540077e+00
-1.37771755e-01 -9.47827101e-01 -9.05826628e-01 -7.08334744e-01
-1.28932118e-01 8.96555603e-01 -4.57748264e-01 7.61878639e-02
-6.77619517e-01 -7.73567021e-01 6.91721797e-01 3.51073086e-01
4.03223515e-01 -1.30361784e+00 -3.65625232e-01 2.94309497e-01
-9.40545857e-01 -1.11743796e+00 -8.51354897e-01 -2.32933104e-01
-8.65896344e-01 -9.40167665e-01 -7.25484550e-01 -6.28322124e-01
2.44983450e-01 4.37888443e-01 1.24480557e+00 1.92877725e-02
9.15332362e-02 5.36546230e-01 -6.59607410e-01 -5.21578252e-01
-8.45121205e-01 7.49515533e-01 -1.32217824e-01 -4.03286397e-01
1.73901677e-01 -5.12983799e-01 -5.08723021e-01 1.38604850e-01
-6.72526658e-01 1.10397361e-01 6.64985299e-01 6.71036184e-01
-2.86990494e-01 -3.87625426e-01 1.09005868e+00 -1.22318327e+00
1.67795086e+00 -3.43511164e-01 1.49485603e-01 1.97985470e-01
-4.21061963e-01 -1.48174435e-01 3.04118931e-01 -2.24466056e-01
-1.16924655e+00 -8.27254891e-01 -1.29860759e-01 5.48963428e-01
5.53819351e-02 5.45621812e-01 3.38705480e-01 5.58650374e-01
6.89296901e-01 3.63030791e-01 9.27523077e-02 -4.61299092e-01
2.50259638e-01 1.28663063e+00 3.38040173e-01 -5.23636580e-01
3.42295527e-01 2.08677828e-01 -6.69404268e-01 -1.15653336e+00
-1.19173443e+00 -8.39106143e-01 -3.65786046e-01 -2.83820271e-01
4.99339044e-01 -7.76478827e-01 -7.22733617e-01 5.18189788e-01
-1.62769985e+00 -1.16166055e-01 -8.95488560e-02 9.89860892e-02
-4.65680361e-01 1.01638246e+00 -1.00825977e+00 -7.31053114e-01
-1.10240602e+00 -7.82070100e-01 1.20019972e+00 3.75919163e-01
-1.23372436e+00 -9.16391492e-01 1.94327846e-01 8.58578146e-01
3.33092332e-01 -9.06032845e-02 7.94711351e-01 -1.11695242e+00
-2.29094565e-01 -3.47422510e-01 -1.76843062e-01 2.93655843e-01
3.34152341e-01 6.27638102e-02 -6.26723588e-01 -1.30227059e-01
-7.52954111e-02 -3.65438819e-01 1.06150925e+00 4.05431956e-01
7.40931869e-01 -7.29065061e-01 -3.56286556e-01 -2.67330229e-01
5.20416737e-01 -2.13544399e-01 4.76826608e-01 2.19500586e-01
5.33678055e-01 7.67787933e-01 4.62735742e-01 7.46806920e-01
6.48922861e-01 2.66750365e-01 -2.31625482e-01 2.75319964e-01
-9.97488648e-02 -1.76016584e-01 4.05458182e-01 1.33766508e+00
8.76525138e-03 -6.58903718e-01 -6.72613919e-01 4.46649760e-01
-2.05753922e+00 -1.15024567e+00 -3.50670129e-01 1.83852994e+00
1.16735530e+00 3.66098315e-01 3.18448514e-01 6.01370819e-02
8.00024986e-01 6.00506783e-01 -1.33020222e-01 -7.95630515e-01
1.21550083e-01 -7.94553235e-02 -1.34357875e-02 6.12915397e-01
-9.79167223e-01 8.10578465e-01 6.07044077e+00 8.65808308e-01
-7.16404259e-01 -5.09844422e-02 7.07867920e-01 2.13105217e-01
-2.46482417e-01 -3.46080097e-03 -9.99682069e-01 5.08868039e-01
1.10596621e+00 -7.37877190e-01 -5.84592745e-02 7.04817951e-01
3.64033878e-01 -3.89680862e-01 -1.20249593e+00 6.88482046e-01
3.77404839e-01 -1.53623652e+00 2.58424878e-01 -7.76732489e-02
8.89660835e-01 -9.79846492e-02 -2.93297470e-01 4.52416390e-01
2.85672635e-01 -6.52791739e-01 2.51649171e-01 3.79427671e-01
2.21242726e-01 -5.15245974e-01 9.84077632e-01 5.33733368e-01
-7.73427725e-01 2.21326515e-01 -1.96066603e-01 -3.20252776e-01
3.15071106e-01 7.10058093e-01 -1.11286128e+00 5.46451271e-01
2.24286422e-01 8.91763508e-01 -6.47494972e-01 8.32542539e-01
-1.66537806e-01 9.12010968e-01 5.32569773e-02 -5.31403720e-01
2.06090689e-01 -2.57987175e-02 8.92931879e-01 1.61833394e+00
-3.71823758e-02 9.95066464e-02 2.18841657e-01 2.84674436e-01
-4.63112324e-01 2.74638563e-01 -2.94537336e-01 -2.30481148e-01
5.18957913e-01 1.37583494e+00 -8.23341012e-01 -6.77676976e-01
-1.12563692e-01 8.77360702e-01 1.92841768e-01 3.41543138e-01
-4.29794103e-01 -6.37211859e-01 3.91405553e-01 -7.37526044e-02
1.93572659e-02 -2.07425609e-01 -6.41396224e-01 -1.06618822e+00
7.60580823e-02 -9.89782095e-01 3.54446173e-01 -6.31291568e-01
-1.37884307e+00 6.46604896e-01 6.79667667e-02 -8.76111448e-01
-5.08717477e-01 -3.39755230e-03 -1.25843942e+00 6.30253315e-01
-1.20485699e+00 -7.67847538e-01 -3.35016578e-01 -1.42211363e-01
1.06968343e+00 -9.29788128e-02 5.23251951e-01 -4.63602804e-02
-6.62010372e-01 4.15913582e-01 1.06514946e-01 2.87799798e-02
1.01773310e+00 -1.19232655e+00 7.35729933e-01 3.21794689e-01
-1.42968372e-01 7.35701323e-01 1.09761226e+00 -7.56565809e-01
-9.71835017e-01 -8.44813287e-01 1.47051692e+00 -8.34554791e-01
8.12099874e-01 -2.56426245e-01 -9.51615691e-01 6.67694092e-01
8.14558625e-01 -1.13929439e+00 9.99144077e-01 4.07663018e-01
2.62789465e-02 2.05647558e-01 -5.13158143e-01 7.53143549e-01
8.26199532e-01 -3.14376384e-01 -1.02706134e+00 5.98154247e-01
8.08921695e-01 -1.47493660e-01 -6.35239422e-01 4.63406593e-02
4.47547197e-01 -9.02729690e-01 5.35742879e-01 -5.58259785e-01
9.04226184e-01 3.66485059e-01 4.25082475e-01 -1.40564537e+00
-3.34137715e-02 -1.14253533e+00 -1.28686443e-01 1.50084269e+00
5.01308858e-01 -6.86860979e-01 8.05006206e-01 3.17637563e-01
-2.62106389e-01 -6.63230062e-01 -5.73548019e-01 -2.48122036e-01
-2.56274864e-02 -1.29078895e-01 2.09221840e-01 6.31427586e-01
5.72919965e-01 1.07206178e+00 -3.32565814e-01 -4.04842317e-01
5.93981087e-01 4.29077029e-01 1.12990701e+00 -1.36332202e+00
1.58733472e-01 -7.81923592e-01 1.77940175e-01 -1.29808581e+00
3.51325721e-01 -6.87682450e-01 -1.32779544e-02 -2.04400206e+00
6.60338938e-01 3.18418816e-02 3.55357826e-01 1.52044803e-01
-6.73472881e-01 -5.38993999e-02 -4.48459163e-02 5.64559221e-01
-1.02905405e+00 7.07205653e-01 1.17747951e+00 -4.83253784e-02
-2.59420514e-01 5.79440773e-01 -1.19777477e+00 6.56903505e-01
8.91692877e-01 -5.23737311e-01 -3.16856802e-02 -1.69104151e-02
2.82600909e-01 3.17477793e-01 4.52063382e-02 -4.54459548e-01
4.70751643e-01 8.67999196e-02 -2.83514168e-02 -8.79522800e-01
6.10658936e-02 1.54952750e-01 -5.72055519e-01 3.61746877e-01
-8.90567482e-01 3.24079454e-01 -2.08472088e-02 6.57002747e-01
-2.28506565e-01 -3.92259359e-01 2.29973108e-01 -2.03395784e-01
-1.78921774e-01 -1.57628104e-01 -7.73148119e-01 5.01890838e-01
5.63741982e-01 -1.01744033e-01 -6.44729614e-01 -1.03219628e+00
-3.95523906e-01 4.88365233e-01 1.77814811e-01 5.15334725e-01
3.94462794e-01 -7.63562560e-01 -1.27592301e+00 -3.79615515e-01
5.30047640e-02 -2.33761415e-01 2.16314495e-01 8.26230466e-01
-4.28370804e-01 5.95298767e-01 6.71546459e-02 -4.47853595e-01
-1.52715862e+00 -1.11813538e-01 -4.08300012e-01 -6.39942884e-01
-6.37401581e-01 7.27456093e-01 1.58840802e-03 -2.43637338e-01
3.74066293e-01 -2.47714698e-01 -4.02823687e-01 6.15235269e-01
7.42231071e-01 8.65494311e-01 3.27445753e-02 -2.99098998e-01
-1.06253937e-01 1.03921272e-01 -6.40757382e-01 -1.38085514e-01
1.07861221e+00 -4.25938994e-01 -4.21950251e-01 5.78440607e-01
1.23822200e+00 3.37671667e-01 -6.63184226e-01 -3.54747772e-01
3.56448889e-01 -1.20917439e-01 -4.19880003e-01 -6.03607416e-01
-1.61010668e-01 5.90228796e-01 -3.94240350e-01 7.83989549e-01
6.60704315e-01 3.05465609e-01 1.19130147e+00 6.52189910e-01
-2.36642554e-01 -1.03515935e+00 5.17202377e-01 8.66275907e-01
1.05082226e+00 -1.27092826e+00 4.08722758e-01 -3.42691630e-01
-9.99469221e-01 1.08113527e+00 2.97175974e-01 1.97181627e-01
2.11077780e-01 3.74821620e-03 5.02138324e-02 -2.08553895e-01
-1.16084695e+00 1.32749975e-01 2.23556980e-01 2.36387342e-01
8.27235818e-01 -1.62357856e-02 -7.02858388e-01 5.86354434e-01
-7.09255397e-01 -3.35184783e-01 1.13150632e+00 9.00417984e-01
-6.34831607e-01 -1.04080689e+00 -2.12280169e-01 8.36952925e-01
-6.63441598e-01 -9.08244699e-02 -9.21666443e-01 3.49034488e-01
-9.05727446e-01 1.37854612e+00 -6.87942803e-02 -9.35800523e-02
2.40064114e-01 1.87916905e-01 1.41351432e-01 -1.00014246e+00
-9.50431764e-01 -5.16508147e-02 7.29015350e-01 3.41995694e-02
-3.86775255e-01 -9.90240991e-01 -8.96592855e-01 -5.64886272e-01
-3.55530947e-01 8.37428808e-01 5.94224155e-01 8.91677499e-01
4.37907308e-01 6.82578564e-01 6.97724044e-01 -8.49829376e-01
-7.86983728e-01 -1.57859695e+00 -1.83033288e-01 2.52405345e-01
3.56504291e-01 -7.32953846e-03 -5.81899107e-01 7.40914047e-02] | [12.462868690490723, 9.344890594482422] |
af276a6f-ca04-4028-bcc2-265970df0545 | intelligent-sampling-for-surrogate-modeling | 2306.04066 | null | https://arxiv.org/abs/2306.04066v1 | https://arxiv.org/pdf/2306.04066v1.pdf | Intelligent sampling for surrogate modeling, hyperparameter optimization, and data analysis | Sampling techniques are used in many fields, including design of experiments, image processing, and graphics. The techniques in each field are designed to meet the constraints specific to that field such as uniform coverage of the range of each dimension or random samples that are at least a certain distance apart from each other. When an application imposes new constraints, for example, by requiring samples in a non-rectangular domain or the addition of new samples to an existing set, a common solution is to modify the algorithm currently in use, often with less than satisfactory results. As an alternative, we propose the concept of intelligent sampling, where we devise algorithms specifically tailored to meet our sampling needs, either by creating new algorithms or by modifying suitable algorithms from other fields. Surprisingly, both qualitative and quantitative comparisons indicate that some relatively simple algorithms can be easily modified to meet the many sampling requirements of surrogate modeling, hyperparameter optimization, and data analysis; these algorithms outperform their more sophisticated counterparts currently in use, resulting in better use of time and computer resources. | ['Chandrika Kamath'] | 2023-06-06 | null | null | null | null | ['hyperparameter-optimization'] | ['methodology'] | [ 4.37102526e-01 -6.65247366e-02 -3.38188499e-01 -2.10779920e-01
-4.30584490e-01 -4.74381834e-01 4.03012156e-01 2.89267898e-01
-5.59848070e-01 9.99146163e-01 -5.34757785e-02 -2.49507666e-01
-4.17028576e-01 -1.02625072e+00 -3.42928469e-01 -6.08521223e-01
2.31427744e-01 7.51548946e-01 3.55313152e-01 7.48054907e-02
6.68451667e-01 9.27373290e-01 -1.67290545e+00 -3.93895119e-01
9.28131998e-01 7.51375079e-01 6.40761554e-02 2.19618693e-01
-2.29314223e-01 -2.33907640e-01 -5.20590127e-01 3.07038743e-02
2.89205253e-01 -4.49267805e-01 -4.75478768e-01 4.27916527e-01
-1.74812779e-01 -4.10799459e-02 3.41105461e-01 8.00263047e-01
4.90680158e-01 4.16847587e-01 7.35142708e-01 -9.96815979e-01
-1.26487926e-01 2.00118646e-01 -6.32073879e-01 -2.55281091e-01
4.90122646e-01 -1.68501923e-03 4.07492399e-01 -4.52879071e-01
3.50357980e-01 8.72743309e-01 5.44096768e-01 3.64127785e-01
-1.57131541e+00 -4.27510411e-01 2.20685825e-01 -3.30470175e-01
-1.59937704e+00 -5.18513560e-01 7.54461229e-01 -3.44579607e-01
4.25438553e-01 6.69986844e-01 6.16035223e-01 6.38811231e-01
-7.09402934e-02 2.70525426e-01 8.68150413e-01 -7.22428977e-01
8.15157235e-01 5.62811971e-01 -1.77835777e-01 2.01610133e-01
6.51106238e-01 3.33846658e-02 -1.79752275e-01 -5.59992135e-01
8.98209751e-01 -9.16270763e-02 -2.95428008e-01 -9.49756742e-01
-1.05742884e+00 9.81724262e-01 -1.87850058e-01 5.63056409e-01
-2.84471452e-01 -3.95821482e-01 2.54968375e-01 3.58466096e-02
5.05921423e-01 1.04212439e+00 -4.80507791e-01 2.44981609e-02
-1.02965510e+00 5.53438962e-01 7.69077957e-01 8.66042435e-01
9.98211801e-01 -1.80656835e-01 6.97466657e-02 6.64024830e-01
1.38804197e-01 2.23838627e-01 5.76345980e-01 -7.84886599e-01
1.22070163e-01 7.52925873e-01 5.59319198e-01 -8.83379161e-01
-3.91068429e-01 -4.16701019e-01 -5.85531294e-01 2.65408874e-01
5.46761632e-01 -5.52336909e-02 -6.73315585e-01 1.59023154e+00
6.76317453e-01 -1.23245150e-01 -3.39559197e-01 6.59599483e-01
1.59669518e-01 4.71692592e-01 -6.99692667e-02 -7.18149900e-01
1.03581560e+00 -2.46392563e-01 -5.01043797e-01 -4.98475693e-02
5.31229675e-01 -8.48784089e-01 1.32453847e+00 5.94065726e-01
-1.12369847e+00 -3.38966846e-01 -9.28348303e-01 3.70126605e-01
-4.40532148e-01 -1.36396527e-01 5.90766907e-01 1.00520349e+00
-6.51291192e-01 5.18369138e-01 -5.46777546e-01 -1.81671470e-01
3.01057491e-02 4.26881820e-01 2.41547208e-02 2.44133826e-02
-7.35746980e-01 7.66765475e-01 1.81114629e-01 -1.42894015e-01
-1.47839129e-01 -7.61938632e-01 -6.50426865e-01 9.98827145e-02
3.97127002e-01 -6.36359096e-01 7.77177155e-01 -8.65816832e-01
-1.59445548e+00 5.41829824e-01 -3.55574824e-02 -1.63344264e-01
6.53027773e-01 7.67184943e-02 -2.41676778e-01 -1.86753437e-01
-8.93913731e-02 2.63088971e-01 7.68787146e-01 -1.11673582e+00
-2.36957029e-01 -4.89722729e-01 -1.16296515e-01 1.62393957e-01
-4.16693479e-01 3.53972726e-02 -3.84938002e-01 -5.37226558e-01
1.69840708e-01 -7.14941680e-01 -7.10826278e-01 -1.13147926e-02
-2.46308625e-01 1.77243035e-02 7.76387334e-01 -1.26697496e-01
1.54029334e+00 -2.01867771e+00 1.08533479e-01 7.11708903e-01
2.38146298e-02 3.44383568e-01 1.14770561e-01 3.82216334e-01
-2.85551953e-03 1.95126086e-01 -4.04969752e-01 -1.01799808e-01
-5.26692420e-02 1.38927802e-01 1.62902921e-02 5.77978432e-01
-7.44322687e-02 1.37433872e-01 -7.51602829e-01 -4.88228559e-01
3.29543710e-01 1.98957995e-01 -7.62093902e-01 -7.53045827e-02
-3.45166415e-01 4.19919193e-01 -7.67828524e-01 3.66112918e-01
6.30215228e-01 -1.09440409e-01 1.86053053e-01 1.11355998e-01
-2.52102166e-01 5.50843924e-02 -1.83885944e+00 1.08730745e+00
-4.24948245e-01 2.33667880e-01 -2.72888839e-02 -1.08218801e+00
1.19102407e+00 1.96785331e-01 7.23364174e-01 -4.62473601e-01
8.94887894e-02 2.71442384e-01 -2.14019999e-01 -3.70698482e-01
5.02098441e-01 -6.53617010e-02 1.82792991e-01 6.17358744e-01
-6.62777245e-01 -3.16735506e-01 4.19654369e-01 -3.06206316e-01
7.19198823e-01 -8.75966400e-02 6.57044351e-01 -3.62455040e-01
4.78677332e-01 -4.17506099e-02 4.98887211e-01 5.28442085e-01
-3.91818658e-02 5.24403811e-01 3.64828050e-01 -3.14792812e-01
-1.13324356e+00 -7.66759932e-01 -5.37793398e-01 5.61041296e-01
9.51561183e-02 -2.83155650e-01 -7.18530715e-01 -4.29928750e-01
8.51491541e-02 6.06363118e-01 -5.58932483e-01 1.61167085e-01
-4.79618341e-01 -7.99885809e-01 -3.49888466e-02 2.33870983e-01
3.99870962e-01 -7.40533054e-01 -1.02174687e+00 3.57893050e-01
3.02648157e-01 -5.86037695e-01 -3.89258087e-01 7.88819417e-02
-1.02628529e+00 -9.53700602e-01 -7.66289830e-01 -4.15938437e-01
8.93594205e-01 8.65489319e-02 7.67859757e-01 9.35881361e-02
-2.32958242e-01 3.11761051e-01 -3.32385302e-01 -2.50049710e-01
-2.16419205e-01 2.40770668e-01 -2.56880596e-02 -9.97645035e-02
1.93714663e-01 -5.48335314e-01 -4.28272605e-01 5.96302986e-01
-1.11219800e+00 -2.65877277e-01 3.56300235e-01 6.73543513e-01
6.17055595e-01 2.05329567e-01 7.20700920e-01 -8.60087931e-01
7.45443523e-01 -4.53747481e-01 -8.78228784e-01 2.08727658e-01
-8.93517375e-01 2.60341763e-01 6.23832166e-01 -7.64814854e-01
-8.16797197e-01 2.03157932e-01 1.35184392e-01 -1.31796092e-01
-2.20533267e-01 5.79766393e-01 -4.16010737e-01 -2.16313884e-01
9.14390028e-01 1.27101000e-02 2.66685784e-01 -4.73846585e-01
-9.44777578e-03 6.79831445e-01 5.62274791e-02 -7.28122294e-01
5.32467902e-01 2.72860229e-01 1.01939477e-01 -9.06482518e-01
-2.90804207e-01 -3.76421452e-01 -4.87874478e-01 -1.07701503e-01
4.44461107e-01 -1.48621500e-01 -4.27026361e-01 1.82328567e-01
-6.85591400e-01 -2.51335382e-01 -5.18813968e-01 5.00726640e-01
-5.01973689e-01 4.17061478e-01 4.21498805e-01 -8.21727097e-01
1.27037436e-01 -1.26608884e+00 7.88084149e-01 2.85961628e-01
-6.65626168e-01 -1.13157332e+00 8.08854252e-02 6.65640235e-02
4.96877611e-01 4.53431427e-01 9.30972815e-01 -6.74316704e-01
-2.27329016e-01 -4.53460842e-01 2.32142270e-01 1.04891375e-01
3.27622056e-01 2.66894311e-01 -5.55414915e-01 -4.64324027e-01
8.28361064e-02 1.30543992e-01 2.20798776e-01 5.43350637e-01
1.43691671e+00 -2.90043145e-01 -6.18800998e-01 5.13396561e-01
1.38020575e+00 5.34383059e-01 6.35730803e-01 4.58241731e-01
-1.40935732e-02 6.82868540e-01 5.97564161e-01 8.04911375e-01
-7.27536306e-02 8.65203440e-01 -2.33851969e-02 -1.26001701e-01
4.53869075e-01 -2.93781925e-02 -3.15553159e-01 9.97087210e-02
2.73826540e-01 -5.90089373e-02 -8.84505272e-01 4.73117173e-01
-1.74108827e+00 -7.22188890e-01 1.00759156e-01 2.84520674e+00
7.94432282e-01 2.47399867e-01 3.14688325e-01 3.57351840e-01
6.13046646e-01 5.27005419e-02 -6.16202593e-01 -5.25101125e-01
3.11607450e-01 2.23898545e-01 5.19902170e-01 4.91271526e-01
-7.03764558e-01 1.89635560e-01 7.31404352e+00 6.92652285e-01
-1.09155381e+00 -5.30553639e-01 6.83917701e-01 -1.30188063e-01
-7.12348461e-01 2.38414153e-01 -7.65779078e-01 6.43704772e-01
6.66130900e-01 -3.76768380e-01 4.98524576e-01 7.04299092e-01
4.93670791e-01 -6.66439176e-01 -1.05069375e+00 7.02859104e-01
-7.67701715e-02 -1.26631868e+00 -3.52821164e-02 2.19675794e-01
6.41474187e-01 -8.28237236e-01 -8.55745922e-04 -5.17115444e-02
-2.33314428e-02 -9.59880114e-01 2.37125307e-01 5.72439075e-01
6.80623591e-01 -7.70470619e-01 3.64778280e-01 5.38463235e-01
-7.77151406e-01 2.21657351e-01 -2.29316190e-01 -1.62213728e-01
-6.43910766e-02 8.23441744e-01 -7.71303654e-01 3.67310524e-01
4.01470721e-01 6.39403518e-03 -3.51438791e-01 1.34176767e+00
3.24692816e-01 3.05344284e-01 -5.99258661e-01 -3.79839033e-01
-1.50916129e-01 -4.86061454e-01 6.35935545e-01 7.25983202e-01
4.38439935e-01 -1.66954771e-02 1.33661181e-01 9.56963837e-01
4.72058475e-01 3.88006002e-01 -7.02599347e-01 1.64519519e-01
9.07999694e-01 9.30894673e-01 -7.12126493e-01 -3.49415541e-01
-3.83118093e-01 3.42461258e-01 -1.10777579e-01 4.27996308e-01
-7.42583513e-01 -6.85271800e-01 4.73132908e-01 5.68372190e-01
2.29956388e-01 -2.03666061e-01 -7.71609068e-01 -7.72181988e-01
7.25174099e-02 -9.18212771e-01 2.27247402e-01 -5.42063355e-01
-9.10254717e-01 2.48851717e-01 4.74247694e-01 -1.30613816e+00
-4.62084889e-01 -4.39110547e-01 -4.92832541e-01 1.10919225e+00
-9.39063370e-01 -4.24239904e-01 -1.94876954e-01 4.89414006e-01
1.60425678e-01 -1.81425005e-01 8.33790302e-01 2.44485423e-01
-4.45624530e-01 3.95858437e-01 3.06990296e-01 -5.13322473e-01
5.64596117e-01 -8.21351290e-01 -7.61418194e-02 7.05700994e-01
-3.11837226e-01 7.40774155e-01 9.99509692e-01 -5.37771404e-01
-1.15353680e+00 -5.98510683e-01 9.45186555e-01 -3.36360419e-03
3.09833169e-01 -2.54735827e-01 -9.39052522e-01 2.51574010e-01
-3.54644269e-01 -2.76192904e-01 6.57664359e-01 3.28047007e-01
1.11045688e-01 -2.17853308e-01 -1.48632932e+00 7.67952204e-01
6.47908866e-01 -4.83134091e-02 -2.56032705e-01 1.31115198e-01
1.58988848e-01 -3.36786419e-01 -8.28507960e-01 4.85177159e-01
4.80473816e-01 -9.40588832e-01 8.15463781e-01 -4.08976853e-01
-7.08844513e-02 -5.25026143e-01 -1.52375102e-01 -1.30994892e+00
-1.58346787e-01 -8.55054438e-01 2.50519902e-01 1.05524361e+00
4.38050181e-01 -8.97849977e-01 9.79214609e-01 1.09201789e+00
1.72904372e-01 -8.45575631e-01 -8.44618320e-01 -7.34906614e-01
5.68776280e-02 -1.86169893e-01 8.90829027e-01 8.85249317e-01
9.79791954e-02 1.72534660e-01 -9.18834209e-02 -2.20446009e-02
4.69561547e-01 2.04438999e-01 1.06134057e+00 -1.36624825e+00
-3.58966947e-01 -5.01542449e-01 -2.51174867e-01 -9.13834035e-01
-2.16202930e-01 -2.67516136e-01 -2.02310517e-01 -1.46301150e+00
-3.88554893e-02 -8.82338047e-01 2.75244296e-01 1.26839533e-01
1.34728057e-02 1.45096574e-02 -1.75637111e-01 -3.22010890e-02
1.98067408e-02 3.40963066e-01 1.32413721e+00 2.25761637e-01
-6.90099359e-01 4.29623574e-01 -8.30285668e-01 7.41081774e-01
7.38346875e-01 -2.15445176e-01 -6.08217955e-01 -1.02910204e-02
3.77708405e-01 2.88980395e-01 -6.29602224e-02 -1.07794333e+00
-4.14846763e-02 -5.27684152e-01 5.51130176e-01 -3.78193736e-01
5.89109547e-02 -1.10820878e+00 6.29267037e-01 2.02333316e-01
-2.57760555e-01 -5.91618419e-02 1.28821850e-01 2.10002020e-01
-2.34885048e-02 -7.13798404e-01 9.61062551e-01 -5.35741821e-02
-2.26281479e-01 4.28664051e-02 -4.65452105e-01 -1.51244372e-01
1.34607077e+00 -7.19679654e-01 -1.08717522e-02 -4.07864630e-01
-7.09802091e-01 1.87748857e-02 8.91268671e-01 -9.23326463e-02
2.97601074e-01 -1.07901943e+00 -2.76373446e-01 5.37946463e-01
1.11845154e-02 1.41936898e-01 -2.81971451e-02 7.22077250e-01
-3.16797882e-01 2.80099154e-01 -5.87307364e-02 -3.91880274e-01
-1.07637429e+00 7.08512187e-01 2.27108330e-01 -1.49178773e-01
-3.38060886e-01 2.40199864e-01 2.43374775e-03 -3.13616782e-01
2.95341074e-01 -5.03146410e-01 -4.21065725e-02 1.25010625e-01
3.55946839e-01 5.85864544e-01 1.73137531e-01 -7.03259781e-02
-1.90749079e-01 5.84613383e-01 2.07458124e-01 -1.56763241e-01
1.16903019e+00 -1.35160923e-01 9.21271518e-02 3.14462572e-01
9.32170391e-01 3.40432078e-01 -1.17650449e+00 -1.26436427e-01
-1.23883873e-01 -7.04537570e-01 -3.56674418e-02 -6.15638435e-01
-8.06963146e-01 2.99615264e-01 3.76964897e-01 5.82958341e-01
1.39658916e+00 -2.77087748e-01 2.09764317e-01 4.74440493e-02
4.60384756e-01 -1.17142928e+00 -1.57871425e-01 1.90705672e-01
8.67276549e-01 -8.30402076e-01 5.02394915e-01 -2.48930067e-01
-4.40641969e-01 1.07670605e+00 4.60987002e-01 -1.06117234e-01
7.19995558e-01 3.37522328e-01 -3.01371902e-01 1.11502476e-01
-3.98665786e-01 8.69018957e-02 3.16121995e-01 5.60563624e-01
4.09616202e-01 -5.04219979e-02 -8.72569442e-01 1.94951624e-01
6.59549795e-03 1.92787006e-01 4.23825681e-01 1.02027428e+00
-5.89372575e-01 -1.40836787e+00 -7.38378406e-01 6.74710035e-01
-7.51378909e-02 2.30645448e-01 -2.94243425e-01 1.02172840e+00
-1.10451188e-02 8.43122482e-01 2.98819542e-01 1.96267590e-02
3.20391983e-01 -5.85020296e-02 4.43897128e-01 -4.48231012e-01
-2.29876712e-01 2.48052567e-01 2.31302127e-01 -2.85799384e-01
-4.45377558e-01 -9.97074604e-01 -9.58177984e-01 -3.79245430e-01
-5.57986915e-01 2.74941415e-01 6.43556833e-01 8.81046057e-01
3.17649722e-01 2.58573979e-01 7.64576256e-01 -7.39382327e-01
-5.55034518e-01 -7.70257711e-01 -7.12022424e-01 1.02303371e-01
1.53435528e-01 -9.03949261e-01 -3.89160961e-01 -1.04604714e-01] | [7.128853797912598, 4.366183757781982] |
c29c690d-9c05-4268-a11f-9f1348fbea35 | large-discourse-treebanks-from-scalable | 2212.06038 | null | https://arxiv.org/abs/2212.06038v1 | https://arxiv.org/pdf/2212.06038v1.pdf | Large Discourse Treebanks from Scalable Distant Supervision | Discourse parsing is an essential upstream task in Natural Language Processing with strong implications for many real-world applications. Despite its widely recognized role, most recent discourse parsers (and consequently downstream tasks) still rely on small-scale human-annotated discourse treebanks, trying to infer general-purpose discourse structures from very limited data in a few narrow domains. To overcome this dire situation and allow discourse parsers to be trained on larger, more diverse and domain-independent datasets, we propose a framework to generate "silver-standard" discourse trees from distant supervision on the auxiliary task of sentiment analysis. | ['Giuseppe Carenini', 'Patrick Huber'] | 2022-10-18 | null | null | null | null | ['discourse-parsing'] | ['natural-language-processing'] | [ 4.34267789e-01 1.00370860e+00 -4.34013188e-01 -7.01705992e-01
-9.50062871e-01 -7.74266660e-01 9.91510749e-01 5.11973560e-01
-4.90117759e-01 1.27655447e+00 8.51998389e-01 -7.64742851e-01
3.64997745e-01 -7.67924607e-01 -3.37728679e-01 -2.99885690e-01
2.41884097e-01 5.63737810e-01 6.07283652e-01 -7.28339136e-01
3.28863412e-01 -1.75959811e-01 -1.14724004e+00 6.48287177e-01
8.76991391e-01 2.94561714e-01 2.98777997e-01 6.96459830e-01
-3.90823036e-01 1.26659453e+00 -1.10043800e+00 -7.37046063e-01
-3.18027884e-01 -7.65493751e-01 -1.47204220e+00 -5.53824231e-02
-1.40875399e-01 -1.08487725e-01 2.18929976e-01 9.89196539e-01
2.27348343e-01 -5.36734983e-02 4.93763655e-01 -5.42867005e-01
-6.68167889e-01 9.18076932e-01 -2.69258440e-01 3.60738575e-01
6.00782692e-01 1.94952972e-02 1.35692060e+00 -2.76218355e-01
1.13348770e+00 1.31318486e+00 2.86952347e-01 8.42400014e-01
-1.07995296e+00 -1.62195917e-02 4.10548866e-01 3.65002528e-02
-3.63859445e-01 -5.88569641e-01 9.27934587e-01 -4.76082802e-01
1.19960713e+00 3.64590108e-01 2.01067165e-01 1.59408569e+00
-1.75084427e-01 7.61322200e-01 1.09733009e+00 -7.02261209e-01
2.06186071e-01 2.19165072e-01 2.70368397e-01 1.56491444e-01
7.66358152e-02 -5.93933880e-01 -4.53373462e-01 3.74094099e-02
2.80791670e-01 -9.74277437e-01 -3.56476486e-01 2.22736970e-01
-1.15579331e+00 1.12517750e+00 1.04632884e-01 8.45757365e-01
-2.27354556e-01 -4.12485331e-01 1.02520871e+00 4.47241336e-01
7.72239029e-01 6.98546529e-01 -3.92190546e-01 -5.13943255e-01
-3.95883650e-01 3.40121835e-01 1.31243658e+00 5.52587569e-01
1.82404652e-01 -1.89069107e-01 -2.03816220e-01 7.07403958e-01
2.90590644e-01 1.16171375e-01 5.61925232e-01 -7.75121689e-01
1.02150929e+00 8.38938653e-01 8.17459971e-02 -8.01719546e-01
-4.37684238e-01 7.31004477e-02 -6.17375314e-01 -1.50127411e-01
9.30512786e-01 -4.72832501e-01 -3.30577433e-01 1.69181097e+00
5.84672034e-01 -5.22256494e-01 5.97596824e-01 9.85611498e-01
1.04951727e+00 6.73129737e-01 3.39778274e-01 -3.75972062e-01
1.67580497e+00 -9.35506463e-01 -8.62601042e-01 -6.02999866e-01
9.91675735e-01 -6.44535840e-01 1.22711968e+00 2.27628037e-01
-1.02513731e+00 -1.55800536e-01 -1.01483381e+00 -5.56575537e-01
-1.04505345e-01 -1.52812585e-01 5.09914339e-01 7.06032753e-01
-7.18641937e-01 2.93475688e-01 -6.99702561e-01 -6.15815334e-02
3.97552431e-01 -1.27090871e-01 -2.63242275e-01 1.80364162e-01
-1.49738681e+00 1.33591819e+00 6.16937101e-01 1.69320121e-01
-3.48486364e-01 -3.86753708e-01 -1.19083071e+00 -2.12223470e-01
7.64653027e-01 -5.05903244e-01 1.61566007e+00 -1.20980084e+00
-1.87393463e+00 1.43324637e+00 -3.60763609e-01 -7.56394625e-01
5.97192645e-01 -4.07561809e-01 -1.00141034e-01 1.34992048e-01
2.05452517e-01 -5.31558543e-02 5.71813226e-01 -8.91903579e-01
-5.79949796e-01 -3.70483547e-01 5.31639397e-01 3.52803528e-01
-2.17121854e-01 4.75746155e-01 1.43676892e-01 -3.93733174e-01
-3.40865970e-01 -5.39695084e-01 -3.54695052e-01 -5.57221234e-01
-6.14161432e-01 -7.51084089e-01 7.58884966e-01 -8.99318457e-01
1.30991054e+00 -1.84919465e+00 4.25188959e-01 -5.37666857e-01
2.04982594e-01 6.79558694e-01 8.39040428e-02 3.64032418e-01
1.92126960e-01 6.00964986e-02 -3.82555455e-01 -4.38578337e-01
-1.35446519e-01 3.64357740e-01 -4.58655298e-01 1.88366666e-01
6.67053759e-01 7.80946076e-01 -1.16573608e+00 -6.48848832e-01
2.87969202e-01 5.63101284e-02 -2.96693534e-01 4.19178814e-01
-9.13393974e-01 1.01414084e+00 -7.98657238e-01 8.47436190e-02
1.31308153e-01 -3.53044450e-01 8.60919893e-01 4.10207093e-01
-2.82430738e-01 1.06098533e+00 -6.58743799e-01 1.61850512e+00
-5.23854434e-01 1.02215087e+00 3.59011143e-01 -1.52454829e+00
8.74092519e-01 5.10516942e-01 -7.13075027e-02 -6.70166135e-01
3.90825719e-01 2.22655475e-01 3.11656892e-01 -8.31320405e-01
6.36953056e-01 -4.20092970e-01 -5.92636347e-01 4.56500381e-01
3.09919519e-03 -4.95658889e-02 5.26980102e-01 7.74574801e-02
1.08466601e+00 2.01171651e-01 7.01975405e-01 -2.15445131e-01
1.05807674e+00 5.71214974e-01 4.96808112e-01 1.20240368e-01
-2.69101173e-01 6.11537576e-01 1.05805314e+00 -3.76009762e-01
-1.06969297e+00 -4.30889010e-01 -1.78707629e-01 1.47987354e+00
-1.50541753e-01 -4.21716422e-01 -9.53852296e-01 -7.60496795e-01
-4.40354407e-01 9.83429730e-01 -4.64661688e-01 4.90528136e-01
-1.44140267e+00 -6.48258209e-01 6.50829315e-01 2.74509043e-01
4.38698828e-01 -1.37133813e+00 -7.81099200e-01 4.68503624e-01
-3.95405233e-01 -1.46440101e+00 4.47896093e-01 -1.11698387e-02
-4.74143952e-01 -1.43129539e+00 -6.11833334e-01 -8.84549081e-01
1.98125899e-01 -1.66690826e-01 1.33943987e+00 1.53792545e-01
4.00045663e-01 -3.00765067e-01 -5.59271216e-01 -6.06856465e-01
-1.17762113e+00 4.21456009e-01 -5.62471628e-01 -3.88320357e-01
5.93612254e-01 -1.54703200e-01 -9.69216973e-02 -2.46401533e-01
-6.18799448e-01 1.97408989e-01 8.76316354e-02 9.44640458e-01
-3.79078910e-02 -6.03112519e-01 1.03430641e+00 -1.56528068e+00
1.03814912e+00 -4.29809719e-01 -5.74251056e-01 2.21506998e-01
-4.13782820e-02 6.15935326e-02 9.51676548e-01 7.27472687e-03
-1.72017443e+00 -6.05966270e-01 -6.89591825e-01 6.34587884e-01
-3.80686641e-01 7.07664251e-01 -3.40441436e-01 9.00386751e-01
8.44232917e-01 -2.19900057e-01 -1.27035156e-01 -3.04003179e-01
4.63302225e-01 6.49360657e-01 5.70012033e-01 -7.48247445e-01
4.67103153e-01 1.69492468e-01 -2.58613378e-01 -1.18864858e+00
-1.42364454e+00 -1.89462990e-01 -8.18766892e-01 7.12616891e-02
1.41423452e+00 -9.21791077e-01 -4.63946223e-01 1.92580223e-01
-1.59435058e+00 -5.58193862e-01 -2.42360115e-01 1.27041683e-01
-3.17620158e-01 4.09698457e-01 -7.18946695e-01 -9.67246473e-01
-3.30409348e-01 -8.04738104e-01 9.57578897e-01 3.02593231e-01
-8.38922441e-01 -1.36619139e+00 3.42040896e-01 5.80566764e-01
1.72406230e-02 5.48613012e-01 1.02958059e+00 -1.01291072e+00
-1.74299002e-01 3.20780218e-01 -1.77486226e-01 4.88480389e-01
1.19071184e-02 6.51871646e-03 -1.02874446e+00 1.95167333e-01
3.18189561e-01 -6.67268991e-01 4.66500849e-01 1.13536440e-01
5.02782166e-01 -2.82090187e-01 -1.57007068e-01 -1.16146125e-01
1.04471874e+00 -2.40592267e-02 4.31908548e-01 7.20217288e-01
4.86022949e-01 1.20174491e+00 8.65901232e-01 9.76276547e-02
5.97733557e-01 2.60503948e-01 -2.03023124e-02 2.04948321e-01
-1.75044909e-01 -1.49629325e-01 3.09826165e-01 1.00912261e+00
-1.27545089e-01 -2.94165730e-01 -1.11827219e+00 9.87005115e-01
-2.03841758e+00 -9.20567811e-01 -5.73825181e-01 1.48691320e+00
1.39396787e+00 5.95831990e-01 3.34913237e-03 1.61112726e-01
8.00459027e-01 7.60480702e-01 -3.41246665e-01 -7.30012834e-01
-3.57389838e-01 2.40948200e-01 -1.36548787e-01 5.44429243e-01
-1.23301315e+00 1.07064629e+00 6.01830959e+00 3.59088063e-01
-9.94368136e-01 2.50509858e-01 6.30014658e-01 2.91211754e-01
-3.04962814e-01 5.04086241e-02 -7.07950711e-01 3.85510743e-01
1.28625512e+00 -3.90542030e-01 -3.46654773e-01 9.29237366e-01
1.88288242e-01 -3.64645809e-01 -1.31115448e+00 3.35321039e-01
-2.05034643e-01 -1.71518075e+00 -4.16585416e-01 -2.50619560e-01
6.28551722e-01 -2.77701262e-02 -3.74941140e-01 5.18354595e-01
5.30495942e-01 -1.04619896e+00 6.33365631e-01 -2.37168118e-01
6.13742888e-01 -2.68765062e-01 9.22361076e-01 8.91011596e-01
-6.56081617e-01 1.48581475e-01 -2.08764210e-01 -7.62187839e-01
7.51654208e-01 5.86016774e-01 -9.82976496e-01 5.64329028e-01
3.63735318e-01 5.43776095e-01 -2.06471309e-01 1.39654264e-01
-8.61133933e-01 8.24873507e-01 -7.39003792e-02 -4.61178273e-01
3.47691208e-01 -1.58648491e-01 6.23350501e-01 1.28745580e+00
-3.27466905e-01 3.53742480e-01 2.48944592e-02 5.31972766e-01
-1.90904006e-01 7.56554678e-02 -6.51884317e-01 -1.87773839e-01
4.04865175e-01 9.92869139e-01 -5.90685844e-01 -4.75948513e-01
-6.42280579e-01 5.63657880e-01 6.52603984e-01 -1.64923921e-01
-5.05809009e-01 -1.34871587e-01 6.08681500e-01 -1.99627671e-02
5.68815805e-02 -3.39654475e-01 -5.84999263e-01 -1.17644656e+00
2.78555095e-01 -1.08857977e+00 4.58173156e-01 -2.70031422e-01
-1.29242969e+00 7.72316813e-01 -3.44310515e-02 -8.27460468e-01
-8.44955027e-01 -7.73893774e-01 -9.42832291e-01 7.91475177e-01
-1.84866798e+00 -1.10988057e+00 1.86736479e-01 3.02940369e-01
9.31270957e-01 -2.54086237e-02 1.16073740e+00 -4.50498648e-02
-5.47777534e-01 2.59744003e-02 -1.50517374e-01 5.98960638e-01
7.77645528e-01 -1.44903588e+00 5.15102863e-01 8.06474805e-01
-5.08336090e-02 4.94365573e-01 1.06268537e+00 -4.53451782e-01
-1.06239831e+00 -6.74603343e-01 1.37597239e+00 -8.61045718e-01
1.14727592e+00 -5.12646496e-01 -1.23926950e+00 6.93918526e-01
6.96953058e-01 -5.01049757e-01 8.14308226e-01 5.79653025e-01
-5.60802184e-02 4.50023174e-01 -9.09604490e-01 4.63811874e-01
8.03226709e-01 -5.88514328e-01 -1.56627810e+00 3.77303302e-01
9.07066703e-01 -8.38108957e-01 -9.74750280e-01 1.55414328e-01
7.35177696e-02 -1.01025367e+00 4.77358878e-01 -8.92293692e-01
1.05337262e+00 5.42534366e-02 2.32289374e-01 -1.00678933e+00
5.43689013e-01 -6.67644799e-01 -6.52846843e-02 1.56361544e+00
6.79578722e-01 -5.41454434e-01 7.94218183e-01 7.57989943e-01
-3.06414187e-01 -4.03668910e-01 -9.64848936e-01 -3.50969702e-01
5.91749549e-01 -3.08218509e-01 1.57448009e-01 1.08300531e+00
7.23245621e-01 1.28983450e+00 -4.52889577e-02 -2.53872871e-01
1.05087027e-01 3.88839811e-01 9.31578755e-01 -1.40759325e+00
-1.76587760e-01 -4.67396200e-01 1.99546397e-01 -1.11534047e+00
5.83625674e-01 -5.70738733e-01 2.96902001e-01 -1.62021852e+00
-1.42677993e-01 -3.36529523e-01 4.93633330e-01 1.29061535e-01
-3.78496230e-01 -2.40993351e-01 6.13504499e-02 2.94664614e-02
-6.75651908e-01 5.27961969e-01 1.34539580e+00 4.56407629e-02
-3.05484205e-01 6.54925704e-02 -9.27746832e-01 1.10419726e+00
6.03334069e-01 -3.19129825e-01 -3.55767518e-01 -6.51091635e-01
3.81094605e-01 2.27015838e-01 -6.04081042e-02 -3.02600473e-01
1.41892254e-01 -3.32594663e-01 -1.94390133e-01 -3.39025915e-01
1.07158534e-01 -2.15173945e-01 -5.58516562e-01 1.59168914e-01
-5.96685410e-01 -1.56362399e-01 3.09013445e-02 3.04521829e-01
-6.78217053e-01 -6.87811196e-01 4.16550487e-01 -3.03736269e-01
-7.54180670e-01 -5.21317303e-01 -5.02644777e-01 6.47234797e-01
1.26964521e+00 1.92636680e-02 -8.57295454e-01 -2.40569249e-01
-6.81521237e-01 2.62281716e-01 3.53512883e-01 4.22126651e-01
9.95264724e-02 -6.48044705e-01 -1.04010987e+00 -2.49758050e-01
-1.22250065e-01 8.47049832e-01 3.05637233e-02 5.08257210e-01
-5.56634247e-01 6.75244570e-01 -5.97703345e-02 -4.93950456e-01
-1.28817284e+00 2.38303035e-01 -1.71714813e-01 -9.41463470e-01
-7.42927015e-01 8.55258703e-01 6.83081299e-02 -2.75213718e-01
-1.42814219e-01 -2.59556025e-01 -7.10180044e-01 2.72533923e-01
6.49336576e-01 -7.53284618e-02 1.58451393e-03 -7.94908881e-01
-2.96511441e-01 -1.31295234e-01 1.11965373e-01 -7.93495476e-02
1.57422054e+00 -2.79754460e-01 -2.78010249e-01 6.01218402e-01
7.89984107e-01 2.18255237e-01 -1.26047134e+00 -2.21601754e-01
8.19053888e-01 -1.50284678e-01 -4.46637779e-01 -3.26457083e-01
-2.65947491e-01 9.48613048e-01 -5.87289572e-01 9.34209287e-01
7.37863183e-01 2.91699111e-01 7.73925006e-01 4.98642594e-01
2.59589344e-01 -1.39297271e+00 -2.68021058e-02 9.00907934e-01
9.56719458e-01 -1.39774275e+00 4.83032241e-02 -6.19071960e-01
-1.01121461e+00 1.16049814e+00 6.48772955e-01 -2.26847008e-01
2.28964537e-01 3.31313968e-01 3.39779019e-01 -2.75693208e-01
-8.02660644e-01 -5.14292791e-02 -2.61994675e-02 8.77029657e-01
1.02667820e+00 -1.62909463e-01 -6.94230676e-01 7.77778983e-01
-4.75825429e-01 -3.12319428e-01 6.58568561e-01 9.83063638e-01
-4.15632516e-01 -1.53482640e+00 -3.00227493e-01 3.57681178e-02
-7.79447317e-01 7.02446923e-02 -5.49735665e-01 1.04898715e+00
-4.13097173e-01 1.10752869e+00 9.71216038e-02 4.88282710e-01
3.44721884e-01 4.58802469e-02 4.01911557e-01 -1.07156324e+00
-8.58470023e-01 -2.60533720e-01 1.25023210e+00 -2.34893546e-01
-1.13079631e+00 -8.15232038e-01 -1.37664306e+00 -3.76720637e-01
-1.16133533e-01 3.06404054e-01 4.32865828e-01 1.53735316e+00
-1.70992780e-02 6.86360419e-01 1.57287791e-01 -7.33271897e-01
-6.02074027e-01 -1.27025831e+00 -3.90795730e-02 5.89168191e-01
4.04528886e-01 -2.49619260e-01 1.35469297e-02 2.36883864e-01] | [10.823064804077148, 9.378169059753418] |
995d736f-f743-4a4c-9903-02494885f782 | eigensubspace-of-temporal-difference-dynamics | 2306.16750 | null | https://arxiv.org/abs/2306.16750v1 | https://arxiv.org/pdf/2306.16750v1.pdf | Eigensubspace of Temporal-Difference Dynamics and How It Improves Value Approximation in Reinforcement Learning | We propose a novel value approximation method, namely Eigensubspace Regularized Critic (ERC) for deep reinforcement learning (RL). ERC is motivated by an analysis of the dynamics of Q-value approximation error in the Temporal-Difference (TD) method, which follows a path defined by the 1-eigensubspace of the transition kernel associated with the Markov Decision Process (MDP). It reveals a fundamental property of TD learning that has remained unused in previous deep RL approaches. In ERC, we propose a regularizer that guides the approximation error tending towards the 1-eigensubspace, resulting in a more efficient and stable path of value approximation. Moreover, we theoretically prove the convergence of the ERC method. Besides, theoretical analysis and experiments demonstrate that ERC effectively reduces the variance of value functions. Among 26 tasks in the DMControl benchmark, ERC outperforms state-of-the-art methods for 20. Besides, it shows significant advantages in Q-value approximation and variance reduction. Our code is available at https://sites.google.com/view/erc-ecml23/. | ['Setareh Maghsudi', 'Meng Fang', 'Tianyi Zhou', 'Qiang He'] | 2023-06-29 | null | null | null | null | ['reinforcement-learning-1'] | ['methodology'] | [-5.53845108e-01 1.32838011e-01 -3.09126198e-01 2.19924375e-01
-9.33264017e-01 -3.44677150e-01 3.33389670e-01 -1.45992398e-01
-5.63277960e-01 1.03926158e+00 3.31863075e-01 -3.74644399e-01
-3.39775175e-01 -4.81521726e-01 -8.47708702e-01 -1.02935290e+00
-1.16287939e-01 1.86966404e-01 -1.07724272e-01 -3.66495848e-01
1.61319792e-01 3.43852311e-01 -9.49570835e-01 -2.09153995e-01
8.36782873e-01 1.12754548e+00 -9.58899409e-02 3.03125411e-01
6.15955293e-02 9.53002155e-01 -4.20449406e-01 -2.10197687e-01
3.66198987e-01 -6.43973589e-01 -8.55689168e-01 -3.50626230e-01
-2.50444770e-01 -4.33555871e-01 -4.18642372e-01 1.24850428e+00
6.64339781e-01 5.84848404e-01 7.62538612e-01 -1.24864578e+00
-5.12499809e-01 6.60829723e-01 -6.26866996e-01 9.68241617e-02
-2.43842173e-02 3.12391996e-01 9.55948651e-01 -7.09358513e-01
5.39136469e-01 1.32319438e+00 5.72721064e-01 8.37136447e-01
-1.52209401e+00 -5.38805425e-01 3.74663770e-01 1.56618029e-01
-1.20304072e+00 -1.87406495e-01 7.92705119e-01 -3.08713168e-01
7.65352190e-01 -2.26533860e-01 9.15159166e-01 1.22671139e+00
3.16621393e-01 1.16717315e+00 1.22089589e+00 -1.93334237e-01
6.03210807e-01 -1.50556892e-01 -1.71133965e-01 7.34868586e-01
-3.82498751e-04 4.65482771e-01 -3.27152729e-01 -1.42058581e-01
1.02359200e+00 -1.71815887e-01 -1.57004490e-01 -5.77943981e-01
-1.08648038e+00 9.18286860e-01 2.59504110e-01 3.84345762e-02
-7.80291080e-01 6.18249476e-01 5.38946211e-01 4.56950277e-01
5.17315030e-01 3.66980076e-01 -3.61288458e-01 -6.09921992e-01
-5.85300565e-01 6.03052974e-01 7.14857876e-01 6.67929649e-01
3.93661678e-01 4.24327642e-01 -6.12179995e-01 6.45247281e-01
2.56776720e-01 6.89956725e-01 5.58297098e-01 -1.52361035e+00
2.60839403e-01 2.73997486e-01 3.95074636e-01 -5.17044723e-01
-3.26682031e-01 -6.45473838e-01 -7.94019043e-01 3.53540957e-01
6.95180774e-01 -4.23443705e-01 -3.37118030e-01 1.91533208e+00
4.93598759e-01 7.03437775e-02 1.31115839e-01 9.84002471e-01
6.52679354e-02 5.65759838e-01 -7.99730942e-02 -5.12564361e-01
8.54930520e-01 -8.88358712e-01 -8.62588644e-01 4.11473364e-01
4.89017010e-01 -1.70546770e-01 1.28258097e+00 4.90274459e-01
-1.24444818e+00 -4.14820284e-01 -6.90122664e-01 2.69646555e-01
2.51040369e-01 -7.25801215e-02 2.55084425e-01 1.82704404e-01
-9.72390771e-01 1.07436740e+00 -1.02655876e+00 2.00003386e-01
4.00670797e-01 1.69866294e-01 1.31331980e-01 4.34911847e-01
-1.42866051e+00 7.90395141e-01 2.76623249e-01 4.69426475e-02
-1.35896993e+00 -6.42599285e-01 -4.70457137e-01 -4.90137823e-02
7.66644716e-01 -4.12086457e-01 1.84332836e+00 -9.67344761e-01
-2.19817400e+00 2.13824123e-01 -1.72964111e-02 -7.33660400e-01
8.53575826e-01 -4.45778161e-01 4.82836217e-02 1.83367550e-01
3.81842852e-02 4.25987601e-01 1.13210177e+00 -9.87535596e-01
-3.86889100e-01 -2.22304702e-01 4.40333150e-02 2.43816227e-01
-2.28921533e-01 -4.42889661e-01 -1.25168040e-01 -7.75324583e-01
-2.69242376e-01 -9.71955061e-01 -4.66823846e-01 -2.03621343e-01
-2.89684236e-01 -6.09504759e-01 2.01339707e-01 -5.63397706e-01
1.51889896e+00 -2.08943582e+00 4.74292427e-01 1.84637666e-01
2.36760914e-01 1.68334156e-01 -5.26812486e-02 5.84394157e-01
8.62345845e-02 -1.43763917e-02 -1.45707592e-01 -1.73780262e-01
4.50340360e-01 3.00802346e-02 -4.23393935e-01 6.87363863e-01
-9.67878625e-02 9.19013321e-01 -1.09871542e+00 -2.66076297e-01
-9.51940045e-02 3.84651363e-01 -5.54097116e-01 2.01056272e-01
-4.41937327e-01 6.04384184e-01 -7.73214579e-01 3.21002454e-01
4.56274599e-01 -1.56930760e-01 2.41557792e-01 1.13138177e-01
-4.22738671e-01 1.80316940e-01 -1.12732542e+00 1.54273951e+00
-3.60889196e-01 1.87779963e-01 -7.15473518e-02 -1.21684468e+00
8.95158350e-01 2.87046820e-01 6.44509852e-01 -9.26925063e-01
1.57622129e-01 2.72092760e-01 -2.74727970e-01 -2.74875849e-01
3.63837689e-01 -1.21056691e-01 1.10355891e-01 4.86370921e-01
-4.91020903e-02 9.55554470e-02 2.41648436e-01 6.24838322e-02
7.88476944e-01 5.41206121e-01 3.77356589e-01 -4.70512688e-01
5.95614254e-01 -3.34203064e-01 7.31686056e-01 7.34095216e-01
-6.14992797e-01 -8.81254673e-02 1.09988976e+00 -3.47376376e-01
-9.36087132e-01 -1.14516020e+00 -1.38539419e-01 9.40060794e-01
-1.46664912e-02 -3.31248730e-01 -9.59012091e-01 -8.19942474e-01
2.68401057e-01 8.24414015e-01 -7.58284867e-01 -3.31899673e-01
-5.30484676e-01 -1.57957107e-01 4.47799742e-01 5.56656957e-01
7.25784898e-01 -1.08638883e+00 -7.27652431e-01 4.16808575e-01
-2.11145282e-01 -6.88162446e-01 -5.84908843e-01 9.98669192e-02
-1.12634277e+00 -7.54483044e-01 -1.04728460e+00 -3.30381840e-01
3.20438594e-01 -2.41737485e-01 9.17364359e-01 -3.22230905e-01
3.54499072e-01 6.65466428e-01 -2.96647102e-01 -2.75653809e-01
-2.66336858e-01 -1.58179738e-02 2.44539440e-01 -1.98061136e-03
1.94215460e-03 -4.83553648e-01 -8.31351817e-01 2.69733816e-01
-4.42580551e-01 -7.36244172e-02 3.42536628e-01 1.05622709e+00
1.03431606e+00 -1.90314263e-01 6.68640792e-01 -5.27734339e-01
1.01832366e+00 -5.00104725e-01 -1.03065956e+00 -1.04863355e-02
-9.76931632e-01 6.72193944e-01 9.20299172e-01 -4.74285334e-01
-1.01526511e+00 -6.07952178e-02 -1.44879341e-01 -7.34539628e-01
2.56050795e-01 3.73081446e-01 3.95784348e-01 2.82995313e-01
5.45295537e-01 3.78221512e-01 3.39107066e-01 -5.39672136e-01
3.80852908e-01 1.68875694e-01 1.52233005e-01 -9.29387450e-01
5.22935987e-01 3.95400524e-01 1.96862131e-01 -5.98539829e-01
-8.40641022e-01 -8.51403773e-02 -1.27100736e-01 -4.82016176e-01
6.82953298e-01 -8.03528011e-01 -1.30004120e+00 4.58737254e-01
-9.63936567e-01 -9.36844647e-01 -8.41104507e-01 6.10248387e-01
-1.11531270e+00 1.50843605e-01 -7.79726326e-01 -1.17133963e+00
-4.40808088e-01 -9.82626915e-01 6.97780371e-01 1.32578745e-01
9.34173092e-02 -9.81281877e-01 4.55700189e-01 -1.50342733e-01
3.51910263e-01 4.66995835e-01 6.84545219e-01 -3.04849118e-01
-1.27834871e-01 3.46200466e-01 2.09830880e-01 5.86180210e-01
-3.44564527e-01 -2.46490642e-01 -6.02149487e-01 -5.31830549e-01
9.25140381e-02 -5.00450671e-01 9.12289441e-01 7.64300406e-01
1.17051339e+00 -5.49811959e-01 1.26369298e-01 6.09184563e-01
1.29774892e+00 4.29280102e-01 4.97481823e-01 5.68900347e-01
3.93237025e-01 3.85604203e-01 7.09572911e-01 1.01741350e+00
3.44201088e-01 6.01522088e-01 4.90560830e-01 1.53996304e-01
3.62984449e-01 -5.02164125e-01 7.59787142e-01 5.70295572e-01
-5.20102024e-01 2.34912068e-01 -6.76677823e-01 3.73794228e-01
-2.13599682e+00 -1.00213134e+00 1.28329828e-01 2.23787570e+00
9.95816529e-01 3.08742765e-02 6.84953094e-01 5.43756559e-02
4.47316825e-01 1.33910537e-01 -1.17049932e+00 -5.28986752e-01
2.35582829e-01 1.69259146e-01 5.06074607e-01 4.46769685e-01
-8.17041814e-01 9.08073962e-01 5.77401638e+00 1.02103043e+00
-8.73322606e-01 1.22994140e-01 4.72171456e-01 -1.81607887e-01
-1.42126039e-01 -3.07117075e-01 -7.08696961e-01 5.34302056e-01
9.79194880e-01 -4.16001350e-01 9.81267333e-01 1.03036201e+00
6.28700674e-01 8.42859000e-02 -7.23981857e-01 8.46710622e-01
-6.64908528e-01 -1.18074429e+00 -4.33682859e-01 2.41521329e-01
8.61180425e-01 -3.77789773e-02 3.36726278e-01 7.28728592e-01
4.62941945e-01 -6.90717816e-01 1.05295312e+00 6.51615202e-01
8.30085397e-01 -1.16345692e+00 4.85430032e-01 2.61010915e-01
-1.14043546e+00 -4.81428027e-01 -4.83254284e-01 -6.40522763e-02
-4.96404991e-02 2.55214900e-01 -3.19202900e-01 3.48522305e-01
6.60022438e-01 8.92605543e-01 3.42346318e-02 4.55959857e-01
-5.03993928e-01 7.60769606e-01 9.73302796e-02 -2.56376982e-01
5.84787726e-01 -5.50110877e-01 7.42182136e-01 7.49452949e-01
9.13743153e-02 1.01993315e-01 8.72934386e-02 1.05643570e+00
-6.49567768e-02 6.66630492e-02 -4.39972430e-01 -2.72509217e-01
4.26239580e-01 8.72923970e-01 -4.55947667e-01 -1.14895806e-01
-1.10973120e-01 7.75370598e-01 6.16476059e-01 6.83011830e-01
-1.06302524e+00 -2.95668513e-01 8.38113368e-01 -1.83061752e-02
5.16275167e-01 -2.58269638e-01 1.24069475e-01 -9.75162387e-01
5.20596206e-02 -1.10944080e+00 2.18833432e-01 -2.71898448e-01
-1.06323075e+00 5.60628235e-01 2.87843645e-02 -1.32861221e+00
-5.17013371e-01 -4.19831932e-01 -2.85095334e-01 5.98193467e-01
-1.50125599e+00 -3.96210551e-01 2.81245232e-01 9.74382401e-01
5.02136648e-01 -1.88768253e-01 6.45559669e-01 2.93407193e-03
-7.67237425e-01 7.01564372e-01 8.00504804e-01 2.04444351e-03
3.41577947e-01 -1.45199811e+00 2.67512769e-01 5.04757702e-01
-2.87374794e-01 5.28678536e-01 6.03311419e-01 -3.64828974e-01
-1.42708707e+00 -8.84871423e-01 1.89387992e-01 -2.75755152e-02
9.11730766e-01 2.80358014e-03 -7.12862074e-01 4.04844671e-01
-1.69252157e-02 -8.92102122e-02 2.25489095e-01 -1.39690414e-01
-1.63016707e-01 -2.67165810e-01 -9.71244276e-01 7.41530359e-01
8.58260274e-01 -4.88411635e-01 -2.50826299e-01 1.13130167e-01
8.75557125e-01 -5.15707314e-01 -9.52009201e-01 2.58604474e-02
7.47711599e-01 -9.47529733e-01 8.20049644e-01 -8.51126134e-01
5.61757803e-01 2.87987757e-02 -9.22385901e-02 -1.62558734e+00
-3.57572675e-01 -1.14454365e+00 -7.82767951e-01 7.14676082e-01
9.57103148e-02 -8.35802555e-01 5.18788338e-01 2.13280544e-01
-1.06998309e-01 -1.32557738e+00 -1.09078574e+00 -1.33715296e+00
5.43241203e-01 -3.37950230e-01 3.60020489e-01 4.81636137e-01
8.22684914e-02 2.54989807e-02 -4.42294180e-01 -3.36934537e-01
8.72755051e-01 4.98926491e-02 4.08389330e-01 -8.45157504e-01
-5.90966225e-01 -5.59201062e-01 2.34752670e-01 -1.12820506e+00
5.27593315e-01 -6.93311095e-01 -2.13598106e-02 -1.39814126e+00
-6.60982206e-02 -1.51374400e-01 -6.82318866e-01 3.21145058e-01
3.16920951e-02 -3.13056320e-01 5.20778656e-01 2.33340204e-01
-7.82899559e-01 1.19914162e+00 1.68176746e+00 1.06750920e-01
-4.76892084e-01 1.35565251e-01 -4.59406614e-01 6.61632061e-01
1.19543839e+00 -6.50758445e-01 -4.65712279e-01 -1.05639413e-01
2.89999902e-01 4.26445037e-01 2.69995719e-01 -8.46504509e-01
-2.16898710e-01 -3.08411896e-01 1.14658915e-01 -4.12469506e-01
1.03535950e-01 -4.85200495e-01 -1.87209770e-01 9.94535387e-01
-7.02689171e-01 2.57286578e-01 1.14427820e-01 7.60568082e-01
1.54841701e-02 -1.68286085e-01 9.57348704e-01 -3.89706045e-02
-4.43239331e-01 3.51649612e-01 -5.49649715e-01 6.91809356e-01
7.82105565e-01 2.18648806e-01 2.64221244e-02 -4.57540721e-01
-5.76557159e-01 3.99298936e-01 1.13772206e-01 -1.21507645e-01
6.74999774e-01 -1.52800202e+00 -5.50213039e-01 -7.90175200e-02
-3.51156712e-01 -1.94322422e-01 2.54801065e-01 1.18228483e+00
-1.70171350e-01 2.53005445e-01 -2.46426404e-01 -2.98456222e-01
-5.92592418e-01 4.81647968e-01 5.77392340e-01 -6.74353302e-01
-7.47125506e-01 5.66746116e-01 -1.85584426e-01 -2.25383505e-01
4.43776459e-01 -4.07528609e-01 -1.93130344e-01 9.66166854e-02
4.52384382e-01 8.60910833e-01 -3.61802399e-01 -1.95241362e-01
-1.48910150e-01 3.61175001e-01 8.87193754e-02 -5.70411205e-01
1.35482585e+00 -7.90015236e-03 1.75039172e-01 5.66016853e-01
1.14421213e+00 -3.61793101e-01 -1.95971906e+00 -2.23639578e-01
-1.54006225e-03 -1.25824496e-01 1.52122363e-01 -6.06783569e-01
-1.10299003e+00 7.15610623e-01 7.12572873e-01 4.75182123e-02
1.03402686e+00 -3.93151820e-01 8.80679190e-01 6.07966959e-01
4.39818114e-01 -1.46261215e+00 3.62798661e-01 7.82955170e-01
1.06371319e+00 -9.28898990e-01 -2.17508987e-01 5.24365842e-01
-1.13525200e+00 1.01248670e+00 4.86025393e-01 -3.49204540e-01
6.61243200e-01 5.60091510e-02 -2.28831604e-01 6.47665188e-02
-9.00953293e-01 -2.45345473e-01 -1.23404630e-01 3.94272983e-01
2.80462414e-01 2.38719210e-01 -6.15169406e-01 6.07713163e-01
1.43936142e-01 1.42427802e-01 2.53370970e-01 7.55916953e-01
-2.66260177e-01 -9.39815700e-01 -1.22981161e-01 3.94769572e-03
-5.47995508e-01 9.02363751e-03 8.09656605e-02 8.45801651e-01
-4.48252231e-01 7.03268588e-01 -1.65779427e-01 -2.76646674e-01
4.51990902e-01 -1.44214556e-02 5.37719011e-01 3.62382643e-02
-4.32894349e-01 2.59067506e-01 -2.48045385e-01 -1.08123231e+00
-1.68286845e-01 -6.82996094e-01 -1.50308037e+00 -3.65664989e-01
5.53073473e-02 4.36233521e-01 3.99706364e-01 6.61678433e-01
3.39742810e-01 5.25810599e-01 9.17050898e-01 -6.98035300e-01
-1.51581514e+00 -7.61675954e-01 -8.46880078e-01 2.22447380e-01
5.51105559e-01 -8.76050651e-01 -4.17376280e-01 -3.40979964e-01] | [4.091449737548828, 2.3552799224853516] |
3cf8be71-1ed8-4f7a-8a27-c4d801595ad6 | sigmorphon-2019-task-2-system-description | null | null | https://aclanthology.org/W19-4211 | https://aclanthology.org/W19-4211.pdf | Sigmorphon 2019 Task 2 system description paper: Morphological analysis in context for many languages, with supervision from only a few | This paper presents the UNT HiLT+Ling system for the Sigmorphon 2019 shared Task 2: Morphological Analysis and Lemmatization in Context. Our core approach focuses on the morphological tagging task; part-of-speech tagging and lemmatization are treated as secondary tasks. Given the highly multilingual nature of the task, we propose an approach which makes minimal use of the supplied training data, in order to be extensible to languages without labeled training data for the morphological inflection task. Specifically, we use a parallel Bible corpus to align contextual embeddings at the verse level. The aligned verses are used to build cross-language translation matrices, which in turn are used to map between embedding spaces for the various languages. Finally, we use sets of inflected forms, primarily from a high-resource language, to induce vector representations for individual UniMorph tags. Morphological analysis is performed by matching vector representations to embeddings for individual tokens. While our system results are dramatically below the average system submitted for the shared task evaluation campaign, our method is (we suspect) unique in its minimal reliance on labeled training data. | ['Suleyman Olcay Polat', 'Brad Aiken', 'Rodney Nielsen', 'Taraka Rama', 'Jared Kelly', 'Alexis Palmer'] | 2019-08-01 | null | null | null | ws-2019-8 | ['morphological-tagging'] | ['natural-language-processing'] | [ 6.94941282e-02 -1.04999794e-02 -8.76086131e-02 -4.59795862e-01
-1.08198714e+00 -1.17837489e+00 7.78634250e-01 4.87212032e-01
-9.65308309e-01 5.13792217e-01 5.99270642e-01 -6.66369617e-01
2.63122112e-01 -4.93821651e-01 -2.96135217e-01 -4.61154014e-01
1.33000940e-01 7.08164811e-01 -9.44606885e-02 -3.38944405e-01
7.11725280e-02 4.93610799e-01 -8.17381561e-01 2.84253180e-01
7.88210750e-01 2.99455315e-01 2.74018586e-01 5.32563746e-01
-3.88880938e-01 2.78984398e-01 -3.97756904e-01 -9.27183628e-01
3.64249140e-01 -2.44591445e-01 -9.76455986e-01 -4.13415069e-03
5.37240326e-01 3.28718811e-01 -7.97180086e-02 1.03817999e+00
4.52754498e-01 1.21111847e-01 9.28221226e-01 -6.96789563e-01
-7.85144329e-01 1.08156121e+00 -2.01806307e-01 4.35112059e-01
1.52292982e-01 -1.23018980e-01 1.67340088e+00 -1.24159265e+00
9.24140751e-01 1.13915133e+00 6.19363427e-01 4.38525259e-01
-1.57076454e+00 -2.94916153e-01 -2.83681359e-02 -2.02220362e-02
-1.35911500e+00 -5.55043817e-01 6.17088199e-01 -5.49842954e-01
1.13944542e+00 1.13023229e-01 4.87004489e-01 9.73236501e-01
1.68226436e-01 6.02577090e-01 1.24071038e+00 -8.92123580e-01
-9.44340006e-02 4.47629929e-01 1.50456265e-01 6.13909900e-01
8.05339441e-02 -2.60205925e-01 -2.47785360e-01 2.82091624e-03
4.69453603e-01 -6.71732605e-01 9.67872590e-02 -1.91504031e-01
-1.56339550e+00 8.83152604e-01 1.37226582e-01 7.26076126e-01
-3.27893883e-01 -1.20737478e-01 6.70669436e-01 4.21384335e-01
6.44120276e-01 7.71186769e-01 -6.41733825e-01 -3.34756114e-02
-9.89876211e-01 1.06874906e-01 8.83429408e-01 7.91286528e-01
8.48178983e-01 8.98894072e-02 8.95846263e-02 1.11828697e+00
1.48586959e-01 3.47785264e-01 6.18591309e-01 -6.00372195e-01
7.40483642e-01 4.11962241e-01 -1.95648566e-01 -6.67747557e-01
-3.22505057e-01 -1.33389518e-01 -1.79683194e-01 -1.44002780e-01
5.23462832e-01 -3.31152976e-01 -7.34228492e-01 1.74067283e+00
3.53777885e-01 -5.06000757e-01 3.11143339e-01 4.27275389e-01
4.62381512e-01 6.54782116e-01 4.19144005e-01 3.57044749e-02
1.66674709e+00 -7.32556641e-01 -7.03588009e-01 -3.91612679e-01
1.08936620e+00 -1.23797286e+00 1.22451377e+00 1.33301108e-03
-1.14391100e+00 -3.91203552e-01 -1.12637365e+00 -2.84434736e-01
-8.37310910e-01 2.74335086e-01 6.13231599e-01 4.90743011e-01
-1.06082642e+00 5.29792309e-01 -8.34332883e-01 -6.96386993e-01
-6.37021214e-02 3.13220769e-01 -5.72123885e-01 9.60157737e-02
-1.11783481e+00 1.42870271e+00 6.48880899e-01 -1.80809617e-01
-2.62533754e-01 -4.84908998e-01 -1.15857017e+00 -2.98304141e-01
-1.15245715e-01 -1.29500479e-01 1.10801041e+00 -7.58600235e-01
-1.16957068e+00 1.37315881e+00 5.77206574e-02 -2.82917619e-01
3.57140936e-02 -2.14719144e-03 -5.17383993e-01 6.28495887e-02
2.45085910e-01 7.16654122e-01 2.70804673e-01 -1.09870303e+00
-5.85616410e-01 -1.50023594e-01 -2.05915093e-01 2.99548179e-01
-5.29244840e-01 7.41436303e-01 -3.16056937e-01 -8.88144493e-01
2.23737016e-01 -1.10057986e+00 -2.48212084e-01 -6.75210655e-01
-3.02887022e-01 -4.28901255e-01 3.88062477e-01 -1.16277099e+00
1.20401251e+00 -2.11367869e+00 3.68890911e-01 1.44793585e-01
-2.97803909e-01 1.15848951e-01 -3.34634960e-01 8.08699965e-01
-2.31604099e-01 3.55875164e-01 -3.30239564e-01 -6.41806602e-01
2.59411365e-01 4.89874989e-01 -2.07748264e-01 5.34068882e-01
5.58333337e-01 9.15995538e-01 -8.84549439e-01 -9.15359080e-01
8.54598135e-02 1.69641331e-01 -5.68250060e-01 3.61195393e-02
3.12735364e-02 9.20904502e-02 -7.85631463e-02 6.48614943e-01
2.72676468e-01 6.19100511e-01 5.66589177e-01 -1.78424984e-01
-5.30275404e-01 8.32435489e-01 -1.01978564e+00 1.79915047e+00
-9.69409287e-01 4.69563067e-01 8.41500834e-02 -7.98750341e-01
8.36938202e-01 4.77045804e-01 4.18157101e-01 -2.60035932e-01
1.91879377e-01 6.42040431e-01 2.50526994e-01 -3.97328645e-01
7.70858884e-01 -3.66727412e-01 -7.20508039e-01 5.75623512e-01
5.88752091e-01 -4.36886102e-01 4.97411788e-01 1.45244464e-01
9.68472064e-01 3.97760838e-01 5.84852159e-01 -6.65327966e-01
3.67708355e-01 3.71543556e-01 6.94001555e-01 -7.73370266e-04
3.86255118e-03 3.82068902e-01 2.20644131e-01 -4.12536770e-01
-1.36217690e+00 -1.26408005e+00 -5.03573120e-01 1.32204819e+00
-4.64543164e-01 -5.80563843e-01 -7.79123247e-01 -9.17422652e-01
-1.96891457e-01 8.33220840e-01 -6.21358335e-01 3.61667573e-01
-1.06740975e+00 -8.38025689e-01 8.37422788e-01 5.29190958e-01
-1.77414089e-01 -1.42333126e+00 -1.33075997e-01 5.30920923e-01
-1.20386474e-01 -1.20574200e+00 -5.87814033e-01 5.89135706e-01
-5.96961200e-01 -7.31865764e-01 -3.10721487e-01 -1.45516753e+00
5.10847270e-01 -4.28269356e-01 1.12130094e+00 -1.94063902e-01
-1.00232199e-01 7.82336518e-02 -4.31778431e-01 -2.84647763e-01
-6.95705235e-01 4.20289963e-01 2.55011857e-01 -7.29230121e-02
5.69636285e-01 -4.29612428e-01 3.98879647e-02 -9.46310386e-02
-9.40681636e-01 -3.85501325e-01 6.72856748e-01 8.20345283e-01
5.48370361e-01 -4.25822765e-01 3.09169441e-01 -1.03793395e+00
4.60586607e-01 -3.86054069e-01 -3.82889748e-01 1.81988955e-01
-2.24466115e-01 1.40252367e-01 6.34983182e-01 -5.34252882e-01
-1.00850427e+00 1.16931714e-01 -3.75315428e-01 1.62688285e-01
-8.45469907e-02 5.54769874e-01 -4.24069196e-01 1.17138632e-01
6.08958244e-01 -6.18534870e-02 -2.45063573e-01 -8.95236552e-01
7.12429464e-01 9.14679050e-01 7.26873815e-01 -8.21625054e-01
9.74583626e-01 1.10059436e-02 -2.48870924e-01 -9.63757455e-01
-6.16751790e-01 -5.28513551e-01 -1.27573717e+00 3.06636006e-01
1.12642121e+00 -7.90530741e-01 1.16883002e-01 1.69546660e-02
-1.15391922e+00 -3.56923252e-01 -4.37616259e-01 6.08806610e-01
-5.21028638e-01 3.53930891e-01 -9.95388567e-01 -3.68679315e-01
-3.32608223e-01 -1.07440126e+00 8.88444901e-01 -2.89663583e-01
-6.10313296e-01 -1.51453710e+00 4.70941961e-01 1.15964241e-01
8.32093731e-02 1.12177603e-01 1.51609337e+00 -1.25431335e+00
-2.62262858e-02 -1.79146662e-01 7.75689259e-02 5.45504510e-01
1.63178444e-01 7.35013634e-02 -7.18636930e-01 -2.23887950e-01
-2.50718385e-01 -3.61018986e-01 4.90579814e-01 -1.88209742e-01
4.53705043e-01 -3.22775096e-01 -1.32963574e-02 6.67033613e-01
1.41617370e+00 -1.58172235e-01 3.33423346e-01 5.49353063e-01
7.99843788e-01 8.69302630e-01 5.22454977e-01 -2.00781357e-02
5.51510036e-01 6.64655566e-01 -1.33064970e-01 -4.96982150e-02
-3.04624528e-01 -3.43412846e-01 7.55161583e-01 1.70348001e+00
2.71906644e-01 5.04080430e-02 -1.11463737e+00 1.12844408e+00
-1.27924931e+00 -7.74531484e-01 -2.28333876e-01 2.00534964e+00
1.22252643e+00 -1.06893882e-01 -3.22263427e-02 -2.06391793e-02
5.61481655e-01 2.18735188e-01 2.96284020e-01 -9.62369800e-01
-1.58021972e-01 7.38687932e-01 5.39414108e-01 6.66310668e-01
-1.23647559e+00 1.58917129e+00 6.29870653e+00 8.08982909e-01
-7.83788323e-01 4.53272492e-01 -2.61757057e-02 2.44102791e-01
-4.44277108e-01 3.32890779e-01 -8.32299232e-01 2.47710779e-01
1.23741853e+00 -8.44857916e-02 3.91928613e-01 5.45089245e-01
-1.21224537e-01 1.37319207e-01 -1.18051898e+00 4.53763336e-01
9.06501040e-02 -1.03951573e+00 -1.01598859e-01 1.74333796e-01
6.44026458e-01 3.67971033e-01 -5.08124419e-02 3.81490618e-01
6.60194576e-01 -7.28331983e-01 7.80084312e-01 -7.27868751e-02
1.00466692e+00 -8.85776520e-01 7.28560388e-01 8.35680217e-02
-1.26014018e+00 3.45245570e-01 -4.70597297e-01 2.39574984e-01
2.82633871e-01 2.13514119e-01 -1.14202797e+00 4.53528136e-01
1.17854424e-01 4.94114816e-01 -7.34706759e-01 7.14686990e-01
-5.63286185e-01 8.34940732e-01 -2.61457682e-01 -1.82647556e-02
5.07537186e-01 -4.06367093e-01 5.43984294e-01 1.89448190e+00
1.95404440e-01 -4.15903986e-01 5.06802142e-01 4.23914760e-01
-1.25404850e-01 9.13130224e-01 -6.67857766e-01 -3.09632629e-01
5.94967008e-01 1.57637715e+00 -7.60947704e-01 -3.80106539e-01
-4.54130173e-01 1.01958811e+00 6.30058944e-01 4.61107530e-02
-4.75436628e-01 -7.30959594e-01 8.11017811e-01 -4.92455438e-02
3.28571320e-01 -6.37070656e-01 -3.57741565e-01 -1.19367194e+00
-1.43374011e-01 -8.25918198e-01 3.87328565e-01 -2.83659309e-01
-1.40543747e+00 7.94128478e-01 1.27957826e-02 -8.68625998e-01
-3.78761590e-01 -8.82168770e-01 -6.71935558e-01 1.15444088e+00
-1.28567159e+00 -1.49512649e+00 5.87396622e-01 3.05222929e-01
5.64992189e-01 -2.77708501e-01 1.17331827e+00 3.37287873e-01
-5.05435288e-01 6.82595611e-01 9.03398842e-02 6.45874143e-01
9.87908363e-01 -1.59240746e+00 7.99525201e-01 1.07828200e+00
7.54585207e-01 8.79669309e-01 6.36304557e-01 -6.54383302e-01
-1.17346406e+00 -1.19058633e+00 1.82308006e+00 -8.20942223e-01
1.42567134e+00 -7.95471191e-01 -8.52111518e-01 9.61955488e-01
4.96968538e-01 3.59232188e-03 1.06927514e+00 4.25359994e-01
-3.79145622e-01 3.02283257e-01 -8.96737039e-01 7.82973230e-01
9.00656879e-01 -7.47874558e-01 -1.15110695e+00 4.65717971e-01
6.25857234e-01 -3.26587036e-02 -1.16956544e+00 -1.17860615e-01
3.22940409e-01 -1.24164157e-01 5.95322251e-01 -8.00037861e-01
3.08781952e-01 -8.73288289e-02 -5.78492701e-01 -1.63702619e+00
-3.06695521e-01 -4.99712199e-01 5.62058687e-01 1.79620194e+00
8.23906898e-01 -3.78067493e-01 4.43892956e-01 9.51932929e-03
-3.79224628e-01 -4.71630186e-01 -1.03621066e+00 -8.97409856e-01
5.82102954e-01 -4.53498513e-01 4.24964428e-01 1.19163811e+00
3.36271554e-01 6.01226211e-01 9.84705165e-02 1.02027878e-01
4.78317916e-01 -3.23308706e-02 3.84340197e-01 -1.06857145e+00
-2.86018431e-01 -2.95026094e-01 -5.13585567e-01 -4.55616802e-01
5.99968791e-01 -1.62869513e+00 2.75595963e-01 -1.37473071e+00
-6.21603541e-02 -7.16574132e-01 -3.40357780e-01 7.51245499e-01
-1.16714425e-01 6.03448451e-01 2.93288648e-01 1.84469730e-01
-1.18762307e-01 2.31829166e-01 5.99322796e-01 4.46528979e-02
-1.35221675e-01 -4.61677700e-01 -5.66963494e-01 8.33407521e-01
8.14216077e-01 -7.80074000e-01 1.42605230e-01 -7.25576937e-01
3.89335081e-02 -3.26048762e-01 -1.09727897e-01 -8.68778527e-01
-1.59374326e-01 -1.58886641e-01 2.01261923e-01 -3.67506176e-01
2.67945707e-01 -6.10024035e-01 -2.28823557e-01 1.89419776e-01
-2.40923405e-01 7.61605322e-01 1.65354639e-01 -1.17974162e-01
-1.20615065e-01 -6.39641941e-01 8.14354062e-01 -1.97927609e-01
-6.24517322e-01 1.38769463e-01 -6.64070606e-01 3.92115384e-01
7.36530483e-01 -6.37045428e-02 1.28699929e-01 3.76611978e-01
-8.26678932e-01 -1.06091000e-01 9.04600799e-01 2.96849102e-01
8.57536867e-02 -1.45857751e+00 -8.79743934e-01 2.18730479e-01
2.81387866e-01 -5.02811253e-01 -3.84687692e-01 5.85201085e-01
-5.17479837e-01 2.30509311e-01 -2.75651127e-01 -1.81401111e-02
-1.21425331e+00 5.83594143e-01 -9.63537693e-02 -4.45994109e-01
-5.10898650e-01 7.69275367e-01 -1.51714444e-01 -7.94402897e-01
-2.35140324e-01 -1.23349085e-01 -1.26219407e-01 5.06922662e-01
2.30694413e-01 -4.72055562e-02 3.75562638e-01 -1.16847444e+00
-3.86178762e-01 3.46127927e-01 -1.77326575e-01 -7.37367749e-01
1.57213044e+00 9.36736260e-03 -2.67736048e-01 5.87890744e-01
1.40083206e+00 8.09154630e-01 -6.74311757e-01 -2.70140409e-01
6.29539788e-01 -6.04837127e-02 -1.34760022e-01 -6.47005439e-01
-5.85034907e-01 9.79208827e-01 -4.27402519e-02 -1.03136711e-01
6.00092232e-01 3.63074951e-02 8.32266986e-01 3.87447834e-01
2.33232260e-01 -1.44552112e+00 -3.74254286e-01 9.26536441e-01
5.91799378e-01 -8.55705619e-01 -9.40136462e-02 -2.66176164e-01
-7.92549968e-01 9.73038137e-01 1.41467884e-01 -3.25232714e-01
3.12118858e-01 4.09169942e-01 5.25458992e-01 -8.29520002e-02
-5.65447330e-01 -4.02893305e-01 2.02708289e-01 6.36802137e-01
9.01418984e-01 1.52349204e-01 -7.56213307e-01 4.11179364e-01
-7.03512192e-01 -8.25320780e-01 5.09004414e-01 8.63584340e-01
-2.73792326e-01 -1.84918773e+00 -2.42461830e-01 2.22466677e-01
-8.44712198e-01 -6.05535507e-01 -5.27641177e-01 1.09354007e+00
3.99318546e-01 6.47194743e-01 1.55509025e-01 -2.86638737e-01
2.51397610e-01 7.21480370e-01 6.22960627e-01 -1.23073936e+00
-9.59051073e-01 -1.55389523e-02 6.10618293e-01 -1.04301266e-01
-1.95050240e-02 -1.14869273e+00 -1.08337307e+00 2.14379819e-04
5.68578066e-03 4.02137011e-01 9.57967818e-01 9.24837589e-01
-2.24149838e-01 2.13711143e-01 4.18142587e-01 -9.32641804e-01
-7.40967751e-01 -1.16661870e+00 -6.21195257e-01 5.97847402e-01
-2.35940769e-01 -3.92241865e-01 -2.62097269e-01 1.88626021e-01] | [10.561917304992676, 10.027105331420898] |
407e0afd-3911-4e02-8d3d-9797298e50ee | code-recommendation-for-open-source-software | 2210.08332 | null | https://arxiv.org/abs/2210.08332v3 | https://arxiv.org/pdf/2210.08332v3.pdf | Code Recommendation for Open Source Software Developers | Open Source Software (OSS) is forming the spines of technology infrastructures, attracting millions of talents to contribute. Notably, it is challenging and critical to consider both the developers' interests and the semantic features of the project code to recommend appropriate development tasks to OSS developers. In this paper, we formulate the novel problem of code recommendation, whose purpose is to predict the future contribution behaviors of developers given their interaction history, the semantic features of source code, and the hierarchical file structures of projects. Considering the complex interactions among multiple parties within the system, we propose CODER, a novel graph-based code recommendation framework for open source software developers. CODER jointly models microscopic user-code interactions and macroscopic user-project interactions via a heterogeneous graph and further bridges the two levels of information through aggregation on file-structure graphs that reflect the project hierarchy. Moreover, due to the lack of reliable benchmarks, we construct three large-scale datasets to facilitate future research in this direction. Extensive experiments show that our CODER framework achieves superior performance under various experimental settings, including intra-project, cross-project, and cold-start recommendation. We will release all the datasets, code, and utilities for data retrieval upon the acceptance of this work. | ['Wei Wang', 'Yizhou Sun', 'Yanqiao Zhu', 'Yunsheng Bai', 'Yiqiao Jin'] | 2022-10-15 | null | null | null | null | ['graph-mining'] | ['graphs'] | [-6.86015785e-01 -3.22167993e-01 -1.80864438e-01 -1.53677151e-01
-1.55672491e-01 -5.57120264e-01 7.05338567e-02 9.16529745e-02
4.48291302e-01 -5.52506968e-02 2.97490478e-01 -3.56460065e-01
-4.62354630e-01 -6.97140276e-01 -3.62831265e-01 -1.15748204e-01
-6.16395921e-02 -2.18699515e-01 5.40089250e-01 -2.19490960e-01
8.10228884e-01 2.82544512e-02 -1.48320341e+00 2.39594400e-01
9.95087028e-01 4.81988370e-01 6.12175584e-01 4.26478654e-01
-2.93471873e-01 1.07407856e+00 -1.37488633e-01 -8.51888537e-01
1.65183425e-01 -1.80553779e-01 -6.87180459e-01 -1.15843825e-01
-6.36357740e-02 -9.45569947e-02 -5.40239930e-01 1.25003970e+00
2.31358007e-01 -1.01934969e-01 3.69855464e-01 -1.16478479e+00
-1.09377754e+00 8.44845772e-01 -8.49629462e-01 2.96034217e-01
3.16552460e-01 4.26643193e-02 1.34445500e+00 -9.99881506e-01
7.22698748e-01 7.22977817e-01 5.98099232e-01 2.31350914e-01
-7.01646984e-01 -5.13838887e-01 2.17241988e-01 2.18060896e-01
-1.33678138e+00 -3.93170938e-02 5.44842362e-01 -1.18121457e+00
9.56140697e-01 7.26220533e-02 5.30302823e-01 5.59315801e-01
5.16999006e-01 2.32617080e-01 4.37117964e-01 -1.16923869e-01
2.15899423e-01 8.16352069e-02 5.40284574e-01 6.77964091e-01
4.01527464e-01 -4.86105382e-01 -7.32820332e-01 -6.99772239e-01
4.82416272e-01 5.29315472e-01 -3.15053552e-01 -2.32183471e-01
-9.33492124e-01 4.48284298e-01 1.74371049e-01 5.00920057e-01
-1.78968295e-01 3.22935075e-01 1.11359462e-01 4.71794635e-01
6.30105317e-01 1.42723083e-01 -3.64158690e-01 -4.60421860e-01
-6.16158366e-01 -1.35842115e-01 8.95208478e-01 1.23864210e+00
1.03407252e+00 -5.31438947e-01 1.81877792e-01 8.84600461e-01
8.88646543e-01 1.49917081e-01 2.60713965e-01 -7.77859330e-01
6.13008082e-01 1.03269398e+00 -6.91315159e-02 -1.31708562e+00
2.13069633e-01 -6.41978323e-01 -2.42627203e-01 -2.71847367e-01
1.37223780e-01 4.39270176e-02 9.90650877e-02 1.20365965e+00
1.46184593e-01 2.36882776e-01 -4.12845463e-01 5.60316384e-01
5.53728640e-01 5.13998091e-01 -2.11630344e-01 -1.44961178e-01
9.76874471e-01 -9.59462881e-01 -2.80117929e-01 1.14013530e-01
8.61958921e-01 -1.00107896e+00 9.23121154e-01 1.73496664e-01
-8.06610823e-01 -2.58129358e-01 -5.39411962e-01 2.19569564e-01
-2.34573763e-02 2.38462836e-01 7.30278671e-01 5.08414328e-01
-1.14481378e+00 6.57438159e-01 -8.30149591e-01 -4.13648158e-01
5.17953455e-01 1.36808604e-01 -3.36262546e-02 -4.06432033e-01
-6.01444781e-01 3.08576018e-01 -4.78213817e-01 -1.96451828e-01
-7.02828705e-01 -8.74403775e-01 -2.00307950e-01 5.40584087e-01
2.21604913e-01 -3.73067230e-01 8.69238973e-01 -7.11426675e-01
-9.15642858e-01 3.52532059e-01 2.13515058e-01 4.44057435e-01
6.03693090e-02 -1.36846498e-01 -1.47428423e-01 -3.13415080e-01
1.50575355e-01 -4.99481261e-01 3.29664886e-01 -8.95432532e-01
-5.65515518e-01 -4.77225512e-01 2.76674598e-01 -2.64184594e-01
-1.06581700e+00 5.51876545e-01 -7.21527815e-01 -4.94193822e-01
-1.94239736e-01 -8.12641859e-01 -1.94019005e-01 -1.18786588e-01
-2.23527238e-01 -6.16249919e-01 2.84922540e-01 -6.99665010e-01
1.71114266e+00 -2.38613725e+00 2.52861679e-01 2.96066105e-01
5.71987748e-01 -2.83759564e-01 -3.25331539e-01 8.36160064e-01
2.32927337e-01 3.40662360e-01 1.63738355e-01 8.66714772e-03
-7.23966137e-02 -3.76395971e-01 -1.19511463e-01 4.32034492e-01
-3.34638834e-01 6.43784165e-01 -1.04956150e+00 -3.27782542e-01
-5.30080855e-01 2.02690363e-01 -7.36384332e-01 4.14817691e-01
6.64052144e-02 -6.94700852e-02 -1.03636289e+00 6.59756124e-01
5.22262275e-01 -8.05820644e-01 3.77599865e-01 4.15119231e-01
-1.87924027e-01 3.27373296e-01 -7.67117441e-01 1.64826882e+00
-6.22112632e-01 5.73660731e-01 -9.05239061e-02 -7.58794129e-01
1.02086616e+00 6.53589889e-02 5.09385407e-01 -1.60370886e-01
-2.41556510e-01 3.14245343e-01 4.40647975e-02 -7.04063356e-01
2.28586301e-01 4.60144848e-01 4.59285453e-02 1.09462762e+00
-1.39482260e-01 3.00246626e-01 1.97741449e-01 7.50070333e-01
1.81147194e+00 9.28793773e-02 4.18597348e-02 -4.63320732e-01
3.69474709e-01 -4.50421721e-01 5.90469599e-01 4.65854168e-01
1.58463195e-01 2.90612996e-01 9.14247811e-01 -3.93971413e-01
-5.48552573e-01 -7.68555701e-01 1.33964077e-01 1.37195170e+00
2.54289657e-01 -8.23940694e-01 -7.55554497e-01 -1.03718042e+00
5.80517128e-02 1.94575116e-01 -3.96663636e-01 5.28625213e-02
-3.28930110e-01 -5.25990307e-01 -7.61671970e-03 3.35469693e-01
-3.10152322e-01 -7.66096532e-01 -3.02086025e-02 1.85629666e-01
-2.02489689e-01 -6.72732830e-01 -9.75910604e-01 -4.59137231e-01
-8.15267861e-01 -1.53936851e+00 -5.93958020e-01 -6.05989277e-01
9.75799263e-01 7.48723209e-01 1.09204876e+00 1.00429189e+00
-3.40278447e-01 6.73416615e-01 -8.47779572e-01 1.85982063e-01
-3.03748071e-01 1.36800364e-01 -2.40137771e-01 2.46756058e-03
3.75319391e-01 -8.27094078e-01 -6.59469724e-01 5.44776261e-01
-4.50646669e-01 -1.17818259e-01 3.52117211e-01 4.20371234e-01
2.35300481e-01 1.99742764e-01 4.65863585e-01 -9.33849514e-01
5.79487145e-01 -1.24513674e+00 -5.51279604e-01 5.43095648e-01
-7.88342834e-01 -2.72781581e-01 6.26023352e-01 -1.18034676e-01
-1.23641849e+00 -3.28650743e-01 2.86025703e-01 7.39063397e-02
2.28790537e-01 9.72810209e-01 7.73010124e-03 -2.78395116e-01
4.20235664e-01 7.51313716e-02 -4.37658787e-01 -8.90838206e-01
6.37837201e-02 1.21227014e+00 3.67965139e-02 -8.11101496e-01
9.48712111e-01 2.73367733e-01 -4.17362481e-01 -1.92903131e-01
-6.59762025e-01 -7.02506483e-01 -4.54340249e-01 -4.59828913e-01
4.36118543e-01 -8.94002140e-01 -2.93397844e-01 3.31969231e-01
-1.31562793e+00 -1.28256395e-01 1.87451541e-01 4.59209293e-01
-1.05069809e-01 6.45903587e-01 -8.07975173e-01 -5.79609573e-01
-2.14695394e-01 -1.22894669e+00 5.50055385e-01 2.34001026e-01
1.51656240e-01 -9.73727107e-01 3.86201501e-01 6.92212641e-01
6.27586901e-01 -1.95037082e-01 1.07427907e+00 -3.57662350e-01
-1.22819591e+00 -1.60992101e-01 -4.33916241e-01 2.38863096e-01
2.24355370e-01 5.37445784e-01 -3.67796123e-01 -4.45786774e-01
-5.33424690e-02 2.10935697e-01 4.69748884e-01 -7.19922259e-02
1.32991779e+00 -1.73941888e-02 -4.03974593e-01 4.29439455e-01
1.48108888e+00 -1.31170675e-01 5.10052145e-01 1.43383771e-01
9.21924233e-01 8.67421031e-01 6.73458695e-01 1.09602964e+00
8.34017992e-01 6.06449008e-01 6.14368856e-01 6.07853591e-01
6.68566823e-02 -6.02945685e-02 6.27532840e-01 1.82034051e+00
-5.09542048e-01 -2.54913539e-01 -1.35740387e+00 7.49690831e-01
-1.95456624e+00 -7.87507653e-01 -7.08055258e-01 2.14256835e+00
5.90051532e-01 -2.00217083e-01 -5.36054559e-02 -4.86413270e-01
9.92811978e-01 -2.14084178e-01 -4.31512296e-01 8.67930651e-02
5.39293587e-01 -1.27944797e-01 1.06663324e-01 -1.30068675e-01
-2.27105319e-01 3.28623205e-01 5.07445383e+00 7.90232360e-01
-6.38254404e-01 5.05092442e-01 3.81026328e-01 4.48325127e-02
-6.97261930e-01 5.32144666e-01 -6.08985484e-01 9.59022880e-01
9.02835190e-01 -6.14287019e-01 7.93084025e-01 1.04976773e+00
9.51752439e-03 1.35613963e-01 -1.16063917e+00 6.26718938e-01
-6.29163906e-02 -1.39070153e+00 -3.82155746e-01 3.25432748e-01
9.52939987e-01 6.87032461e-01 -3.12082142e-01 1.53131336e-01
5.13622999e-01 -3.79109085e-01 6.26657307e-01 7.14209974e-01
4.63290960e-01 -5.70610404e-01 6.46650851e-01 5.43217599e-01
-1.47096932e+00 -3.74590129e-01 -6.84442639e-01 -2.06882477e-01
-3.35981190e-01 8.09881449e-01 -5.63056953e-03 6.43387735e-01
9.69527185e-01 1.28837276e+00 -8.11816216e-01 1.25380731e+00
-5.34626581e-02 6.73377991e-01 2.88638860e-01 -1.10129174e-02
-4.25319076e-01 -3.22406590e-01 1.87240213e-01 1.10014057e+00
9.81946170e-01 -1.47322444e-02 -1.51210297e-02 1.15728378e+00
-3.04724842e-01 5.11920154e-01 -5.05513191e-01 -3.34596902e-01
7.15729356e-01 1.50367677e+00 -6.69636369e-01 1.84680223e-01
-1.15285957e+00 3.04838359e-01 6.04514420e-01 2.31688038e-01
-5.02417862e-01 -6.66835129e-01 7.24937618e-01 4.75444704e-01
2.75501668e-01 -7.41192251e-02 -8.95620063e-02 -1.42410886e+00
3.62427592e-01 -6.24681771e-01 1.09598473e-01 -5.47262311e-01
-1.68160367e+00 4.11269128e-01 -4.80458587e-01 -1.39886427e+00
4.38211143e-01 -1.33232325e-01 -1.18817890e+00 7.09962904e-01
-1.38985217e+00 -8.34103346e-01 -5.45407712e-01 1.59699097e-01
2.44246468e-01 -5.95288634e-01 2.00855851e-01 4.26170617e-01
-7.30650485e-01 4.36249346e-01 5.25597930e-01 -8.84448290e-02
5.04024804e-01 -1.00729561e+00 5.82900286e-01 9.60935891e-01
1.93640083e-01 1.19495106e+00 1.23985820e-01 -8.25431883e-01
-1.80808902e+00 -1.11897409e+00 8.90969992e-01 -5.84202349e-01
1.20723403e+00 -4.37428802e-01 -9.68417048e-01 3.95642221e-01
1.17668852e-01 2.65598953e-01 9.36047912e-01 3.57194930e-01
-4.52181727e-01 -4.11007255e-02 -5.95943809e-01 8.84481147e-02
1.37965882e+00 -6.61119103e-01 -9.04302821e-02 5.80030620e-01
6.00055277e-01 1.07418075e-01 -1.48109102e+00 -1.16804481e-01
2.46426880e-01 -1.10171115e+00 5.00868917e-01 -3.30477446e-01
8.64132404e-01 -2.37864181e-01 -1.02583446e-01 -1.13814580e+00
-5.90151787e-01 -6.93311214e-01 -1.47186518e-01 1.69780767e+00
2.24418253e-01 -5.76277137e-01 5.86466730e-01 6.76728189e-01
-7.28498921e-02 -9.94292498e-01 -4.98427898e-01 -6.99533105e-01
-3.92173044e-02 -2.56345540e-01 8.86478007e-01 1.00498271e+00
5.71028948e-01 -7.10064471e-02 -1.58162996e-01 1.70851246e-01
6.62458479e-01 6.38531804e-01 7.65374064e-01 -1.44703245e+00
-8.56739104e-01 -4.11609054e-01 -4.38529313e-01 -9.91765380e-01
3.52056652e-01 -1.34248841e+00 -1.70117453e-01 -1.51615179e+00
1.14314198e+00 -8.36628020e-01 -5.60096025e-01 2.90307581e-01
-3.42134714e-01 -1.94960907e-01 9.48120095e-03 8.61151218e-01
-1.08354223e+00 2.96826005e-01 9.70605731e-01 1.50859877e-01
1.22448206e-02 2.06037998e-01 -9.77782309e-01 5.42887866e-01
3.85069996e-01 -8.71038973e-01 -3.91937137e-01 -8.05440664e-01
1.08108556e+00 3.71802092e-01 1.32882044e-01 -6.06008649e-01
2.96104640e-01 -3.75849515e-01 -4.93669897e-01 -1.68135449e-01
-6.12311006e-01 -5.51124394e-01 3.30790699e-01 2.66753525e-01
-2.48003677e-01 -7.27058053e-02 -5.46893299e-01 9.04134274e-01
-1.21003464e-01 -6.75300717e-01 2.50672638e-01 1.69710293e-02
-3.16225618e-01 7.26235628e-01 -3.29394877e-01 2.54665405e-01
1.00819886e+00 1.58511147e-01 -7.54046023e-01 -1.76855065e-02
-3.30632955e-01 2.55102962e-01 1.06352150e+00 8.31736028e-01
4.57915276e-01 -1.08464479e+00 -6.19953930e-01 -2.57271647e-01
4.20847893e-01 -3.84944975e-01 3.43867242e-01 8.99077833e-01
-4.89030987e-01 -5.57007417e-02 8.11236501e-02 -7.66350627e-02
-1.24157298e+00 2.86962569e-01 -4.88881916e-02 -1.70328096e-01
-4.32464302e-01 8.48016381e-01 4.69092458e-01 -3.79606366e-01
-2.15601232e-02 9.40177590e-02 -2.79429376e-01 -3.65309119e-01
5.61890125e-01 6.96578979e-01 -6.80746734e-02 -4.12657857e-01
-3.49484563e-01 6.02152050e-01 -2.55029500e-01 8.20862412e-01
1.85816348e+00 -2.80686796e-01 -7.35625088e-01 2.41738260e-01
9.94829655e-01 2.96952039e-01 -1.11568570e+00 -4.17224407e-01
2.68855155e-01 -9.14082468e-01 1.22301236e-01 -4.59195018e-01
-1.59158909e+00 9.38560545e-01 8.08710232e-02 4.99896407e-01
6.40611231e-01 3.96939427e-01 6.66178286e-01 2.03749254e-01
9.15124953e-01 -9.23519075e-01 3.72690827e-01 3.12583238e-01
5.97256124e-01 -8.16462457e-01 1.47070244e-01 -7.04783142e-01
-2.22830504e-01 1.24532366e+00 8.62877607e-01 7.88971633e-02
9.55887973e-01 3.08272243e-01 -2.57106274e-01 -5.64597070e-01
-1.19715142e+00 3.78660589e-01 7.80046079e-03 4.71720874e-01
8.49769056e-01 1.33915227e-02 -2.90596008e-01 1.03077388e+00
4.85827148e-01 4.11058813e-02 1.05409455e+00 8.51258576e-01
-4.46576685e-01 -1.50118494e+00 4.89661135e-02 8.20981801e-01
-6.85566306e-01 -2.81959534e-01 -2.59437531e-01 -1.32976510e-02
1.78697750e-01 9.08300519e-01 -3.56912494e-01 -6.12368405e-01
6.76016957e-02 -4.50213850e-01 1.16026290e-01 -1.13685095e+00
-7.21625984e-01 -3.67647231e-01 -2.30342060e-01 -5.14885902e-01
-3.88330072e-02 -8.63051295e-01 -1.17434669e+00 -4.48250324e-01
-9.31222975e-01 2.68392533e-01 7.99578786e-01 8.05503905e-01
1.04088318e+00 4.46507335e-01 1.09446311e+00 -3.63014340e-01
-6.72717571e-01 -6.37613237e-01 -8.74145746e-01 2.78385192e-01
-4.53145891e-01 -7.69112468e-01 -5.00519514e-01 2.12223917e-01] | [7.556730270385742, 7.983922958374023] |
59e2f5bc-000e-4486-a626-6d4171102d4c | mfe-ner-multi-feature-fusion-embedding-for | 2109.07877 | null | https://arxiv.org/abs/2109.07877v1 | https://arxiv.org/pdf/2109.07877v1.pdf | MFE-NER: Multi-feature Fusion Embedding for Chinese Named Entity Recognition | Pre-trained language models lead Named Entity Recognition (NER) into a new era, while some more knowledge is needed to improve their performance in specific problems. In Chinese NER, character substitution is a complicated linguistic phenomenon. Some Chinese characters are quite similar for sharing the same components or having similar pronunciations. People replace characters in a named entity with similar characters to generate a new collocation but referring to the same object. It becomes even more common in the Internet age and is often used to avoid Internet censorship or just for fun. Such character substitution is not friendly to those pre-trained language models because the new collocations are occasional. As a result, it always leads to unrecognizable or recognition errors in the NER task. In this paper, we propose a new method, Multi-Feature Fusion Embedding for Chinese Named Entity Recognition (MFE-NER), to strengthen the language pattern of Chinese and handle the character substitution problem in Chinese Named Entity Recognition. MFE fuses semantic, glyph, and phonetic features together. In the glyph domain, we disassemble Chinese characters into components to denote structure features so that characters with similar structures can have close embedding space representation. Meanwhile, an improved phonetic system is also proposed in our work, making it reasonable to calculate phonetic similarity among Chinese characters. Experiments demonstrate that our method improves the overall performance of Chinese NER and especially performs well in informal language environments. | ['Kui Meng', 'Jiatong Li'] | 2021-09-16 | mfe-ner-multi-feature-fusion-embedding-for-1 | https://openreview.net/forum?id=5N4bCRdqHAw | https://openreview.net/pdf?id=5N4bCRdqHAw | null | ['chinese-named-entity-recognition'] | ['natural-language-processing'] | [-3.26228827e-01 -6.12841427e-01 3.57417837e-02 -3.83565962e-01
-3.92394662e-01 -6.11612022e-01 3.90483648e-01 -4.32535028e-03
-7.67067432e-01 6.59667969e-01 6.97680056e-01 -2.58899420e-01
3.69793802e-01 -9.48770761e-01 -1.76539317e-01 -5.46479762e-01
5.63132405e-01 -4.35407087e-02 1.84493378e-01 -1.56219378e-01
2.33381107e-01 5.77008665e-01 -8.18951845e-01 1.76893592e-01
1.34144402e+00 4.50814456e-01 5.01092911e-01 1.00978054e-01
-1.00942969e+00 3.19067806e-01 -9.29443777e-01 -8.82360578e-01
-1.99385975e-02 -4.66468334e-01 -7.36885309e-01 -8.98519084e-02
-2.64310956e-01 4.47902456e-02 -3.62508982e-01 1.37304544e+00
7.17724860e-01 1.97254524e-01 4.79203552e-01 -7.56570399e-01
-1.17801046e+00 9.67978179e-01 -3.10057312e-01 -7.42416382e-02
2.40129098e-01 -4.40332472e-01 7.82280385e-01 -9.84432876e-01
3.53329718e-01 1.16555822e+00 8.43375206e-01 7.26602256e-01
-7.74727583e-01 -7.42396832e-01 1.58751890e-01 1.61887705e-01
-1.72917700e+00 -4.80769798e-02 4.78847086e-01 5.40149063e-02
6.52736127e-01 3.49628299e-01 4.15573806e-01 1.00212264e+00
1.38332825e-02 8.01777542e-01 7.26078391e-01 -3.19652349e-01
-2.18532551e-02 4.73319590e-01 1.03708774e-01 1.58488035e-01
5.17503023e-01 -3.80477488e-01 -1.39469886e-02 -1.99545752e-02
6.25455856e-01 4.64323163e-01 -3.94825995e-01 3.03867012e-01
-1.48540068e+00 7.56521463e-01 3.79089832e-01 9.96997595e-01
-2.40680277e-01 -7.63009265e-02 4.09002692e-01 -8.73901546e-02
4.11116816e-02 4.38232481e-01 -4.76879746e-01 -4.22630221e-01
-5.03365099e-01 -3.14638883e-01 8.15047562e-01 1.15021837e+00
7.02218473e-01 1.59931794e-01 8.96723121e-02 1.10033476e+00
1.66639715e-01 5.57056963e-01 1.06415427e+00 -3.86518687e-01
2.87203252e-01 7.17031717e-01 4.21091579e-02 -1.24104059e+00
-1.52602971e-01 -3.62622619e-01 -9.56036866e-01 -5.12071490e-01
5.83466701e-02 -4.12421972e-01 -6.85171068e-01 1.53005409e+00
1.80164680e-01 2.01415777e-01 2.73970097e-01 7.03626812e-01
7.22433925e-01 1.17584562e+00 3.24208289e-01 -2.24892469e-03
1.60303617e+00 -9.56248581e-01 -9.78171229e-01 -1.12240449e-01
8.23603868e-01 -1.07009971e+00 1.10709703e+00 -8.70435089e-02
-2.87783682e-01 -5.73524237e-01 -7.63514757e-01 -6.14666604e-02
-8.41316342e-01 4.14636463e-01 5.05985558e-01 1.05084252e+00
-4.81583178e-01 5.18613279e-01 -5.90947151e-01 -4.07784253e-01
1.15939207e-01 7.28864819e-02 -4.32976127e-01 -1.15269199e-02
-1.51795328e+00 8.33762765e-01 8.17531824e-01 3.04234385e-01
-3.70812416e-02 -3.46757084e-01 -9.09422576e-01 2.57740319e-01
1.27090901e-01 -3.74473520e-02 8.75018060e-01 -7.20907867e-01
-1.32361841e+00 3.71123642e-01 -2.90426075e-01 1.35660335e-01
1.01166554e-01 -3.31897289e-01 -1.45316541e+00 -3.09717387e-01
2.09853724e-01 4.05729145e-01 3.98464203e-01 -9.02535141e-01
-7.20494688e-01 -1.68960784e-02 -3.98899764e-01 1.90458849e-01
-8.82659972e-01 3.07866842e-01 -6.35757804e-01 -1.13361466e+00
2.10818008e-01 -6.55075908e-01 -2.24979088e-01 -3.87985796e-01
-5.56728065e-01 -4.18056369e-01 9.59296465e-01 -7.45787203e-01
1.72929740e+00 -2.45907879e+00 -1.66092575e-01 1.39006242e-01
-1.67613730e-01 6.11157358e-01 -9.43914875e-02 5.90820372e-01
5.10464236e-02 7.97073126e-01 -3.89145464e-01 -7.20482767e-02
-4.83278371e-02 2.78895795e-01 -2.41473511e-01 8.13909173e-02
3.00258368e-01 7.53027380e-01 -8.75288904e-01 -5.83832085e-01
-5.74831327e-04 5.45237958e-01 -1.52370423e-01 7.34445527e-02
3.73665959e-01 1.60191134e-01 -6.59091711e-01 5.02315879e-01
9.11258936e-01 1.17570003e-02 1.17960423e-01 -2.58613497e-01
-3.98828834e-01 2.94934601e-01 -1.48545766e+00 1.35079575e+00
-6.30670547e-01 2.46016532e-01 -5.23565173e-01 -7.37352610e-01
1.18100882e+00 4.36869770e-01 -2.81627737e-02 -3.57089847e-01
1.86360538e-01 6.31282389e-01 -1.38902813e-01 -4.42782342e-01
6.72486961e-01 -1.41845912e-01 -3.85240138e-01 1.76431805e-01
-1.71853334e-01 2.09324881e-02 3.59800085e-02 -7.02306181e-02
6.92403018e-01 -1.57836393e-01 4.85381871e-01 -1.17983289e-01
8.67610037e-01 -1.97981205e-02 1.09680367e+00 4.63459015e-01
-1.36882335e-01 5.39282501e-01 1.07382707e-01 -1.27693370e-01
-1.12542367e+00 -9.95146990e-01 -3.97933096e-01 8.10661972e-01
3.68020952e-01 -5.12403488e-01 -6.54119372e-01 -8.67577374e-01
-3.12409669e-01 8.68247032e-01 -1.78956136e-01 -2.91004390e-01
-8.37795377e-01 -5.24521470e-01 1.12042630e+00 6.60242796e-01
9.47857738e-01 -1.26757908e+00 2.57247657e-01 5.72811246e-01
-1.70478851e-01 -9.91066813e-01 -1.00716448e+00 -6.19204566e-02
-6.89493895e-01 -5.77067673e-01 -1.35499072e+00 -1.41726768e+00
6.34464979e-01 3.05924475e-01 3.89008820e-01 7.58036077e-02
1.97525024e-01 -1.22766532e-01 -8.35364759e-01 -2.91522413e-01
-4.36140895e-01 2.32217863e-01 2.69040108e-01 2.17339899e-02
7.33866930e-01 -2.54095256e-01 -1.98300943e-01 4.80542809e-01
-1.18773890e+00 -2.90182918e-01 6.31250620e-01 9.71887171e-01
3.77801090e-01 3.19921523e-01 4.04114455e-01 -6.35574639e-01
5.83167136e-01 -4.82524127e-01 -2.11918890e-01 5.07139862e-01
-3.66760850e-01 1.51316464e-01 1.24013686e+00 -8.33582640e-01
-1.30595577e+00 -1.32841170e-01 -4.11880881e-01 -1.02620445e-01
-3.23929220e-01 7.92953491e-01 -8.07039678e-01 1.29208639e-01
2.63253480e-01 6.14352047e-01 -4.49722230e-01 -8.74618530e-01
3.55831981e-01 1.12374365e+00 4.68615383e-01 -3.44095975e-01
8.19800615e-01 1.58803999e-01 -5.70892990e-01 -9.73598480e-01
-4.69094753e-01 -5.12204707e-01 -4.71164584e-01 3.15712154e-01
1.08269417e+00 -8.54934573e-01 -4.46180344e-01 7.54549742e-01
-1.53916633e+00 4.75953996e-01 6.47201166e-02 8.96059573e-01
3.07228684e-01 9.12324250e-01 -7.59441137e-01 -5.54711819e-01
-1.43851221e-01 -8.98355246e-01 6.45655096e-01 9.13053095e-01
6.90481365e-02 -1.05892408e+00 5.09937257e-02 -1.93071306e-01
4.49291021e-01 -3.56321305e-01 8.54241550e-01 -1.06279671e+00
-1.18185006e-01 -2.01604545e-01 -2.55406260e-01 6.48251057e-01
5.88040173e-01 -6.50592847e-03 -5.80082536e-01 1.52226174e-02
-1.08566850e-01 3.26651692e-01 6.54080570e-01 -3.44018936e-01
1.05837786e+00 -3.03498328e-01 -3.48800868e-01 5.61537385e-01
1.34468472e+00 6.12890244e-01 8.50170374e-01 4.81169373e-01
9.36620831e-01 4.35442328e-01 3.51544529e-01 3.63253385e-01
2.20935047e-01 4.22815442e-01 -2.14500800e-01 -8.38975012e-02
6.85380995e-02 -5.78575552e-01 5.20890653e-01 1.35079384e+00
1.56804547e-01 -3.12866658e-01 -8.71035516e-01 6.35655701e-01
-1.40440011e+00 -1.00062239e+00 -1.76590249e-01 2.06189013e+00
1.14700758e+00 -1.20534539e-01 -4.73487973e-01 -9.84207094e-02
1.45520210e+00 7.65171228e-03 -3.18665981e-01 -3.53609979e-01
-5.11438429e-01 2.83641666e-01 4.35627371e-01 3.19209881e-02
-9.89078522e-01 1.15683794e+00 4.61753416e+00 1.35094774e+00
-1.33850241e+00 1.35940850e-01 3.44711155e-01 9.67715979e-01
-5.67889094e-01 2.52337396e-01 -1.13571501e+00 8.52569461e-01
6.05645299e-01 -4.83017921e-01 2.23408639e-01 9.73348200e-01
-5.62900119e-02 3.38232696e-01 -4.98767227e-01 1.09988999e+00
8.30710828e-02 -1.22095406e+00 3.14022183e-01 -1.38504848e-01
6.74944639e-01 -2.04172730e-01 -2.22440898e-01 3.42075288e-01
8.90384391e-02 -8.01478326e-01 6.10826254e-01 4.17516440e-01
6.60963058e-01 -7.97816217e-01 1.03776276e+00 2.34370157e-01
-1.53512919e+00 1.45636261e-01 -7.27648020e-01 3.36316317e-01
5.35693467e-01 4.93805200e-01 -3.42687935e-01 7.19296455e-01
5.48098803e-01 6.30447090e-01 -5.53047001e-01 1.41344762e+00
-3.54722768e-01 5.69012880e-01 -3.94247025e-01 -6.20226145e-01
3.21173549e-01 -3.67674530e-01 3.82140577e-01 1.35956705e+00
7.64998853e-01 1.61336362e-01 -1.10218063e-01 7.44925976e-01
-1.21872105e-01 7.31515467e-01 -3.55920941e-01 -6.17680311e-01
8.10692132e-01 1.16674972e+00 -7.49661207e-01 -4.94716376e-01
-6.41223073e-01 1.48429811e+00 1.07988395e-01 3.04994434e-01
-8.97111654e-01 -1.29321837e+00 6.62326634e-01 -4.79976296e-01
5.49700737e-01 -3.52945328e-01 -2.05178738e-01 -1.59002185e+00
-2.16775835e-02 -7.18412161e-01 2.00648353e-01 -4.64521289e-01
-1.62380302e+00 7.91058958e-01 -5.88172197e-01 -1.51332819e+00
2.99750686e-01 -5.30204058e-01 -9.65914190e-01 9.43014562e-01
-1.30242515e+00 -8.12165022e-01 -7.96314850e-02 5.01986384e-01
3.40852410e-01 -1.25348166e-01 1.02746010e+00 6.78740799e-01
-8.72485816e-01 9.38547254e-01 5.57151139e-01 8.36279333e-01
7.12148428e-01 -8.96881759e-01 5.34683406e-01 1.05955374e+00
3.64287436e-01 1.21483099e+00 2.37565786e-01 -8.91163170e-01
-8.91375005e-01 -1.06236625e+00 1.45085955e+00 -1.11993477e-02
5.77127695e-01 -2.22892791e-01 -1.17046595e+00 3.51333678e-01
6.15071841e-02 -2.24378005e-01 9.06218231e-01 -1.88598141e-01
-2.28705376e-01 -6.88611157e-03 -8.18748891e-01 9.18858767e-01
8.30795288e-01 -6.66763783e-01 -9.66967344e-01 7.95677602e-02
9.64894950e-01 7.39322901e-02 -7.35066652e-01 1.76147535e-01
2.79412270e-01 -5.82771719e-01 5.27657568e-01 -6.82077825e-01
-6.81728274e-02 -7.84083605e-01 -1.89783439e-01 -1.24738455e+00
-3.95865083e-01 -4.73230630e-01 4.83056188e-01 1.83516967e+00
6.52829409e-01 -7.63900220e-01 5.21469414e-01 4.00695652e-01
-3.40867311e-01 -2.31283665e-01 -9.50424612e-01 -1.08359492e+00
1.53304920e-01 -2.78561801e-01 1.02622843e+00 1.28205347e+00
7.51987025e-02 2.38591537e-01 -3.53530735e-01 2.66634077e-01
3.06961741e-02 -1.62137270e-01 3.28376889e-01 -9.39474344e-01
1.79534018e-01 -4.23878044e-01 -4.59413648e-01 -1.26397991e+00
2.45436326e-01 -7.86745548e-01 -8.23199470e-03 -1.36411810e+00
-4.46975566e-02 -6.73021734e-01 -3.39743495e-01 5.52002072e-01
-4.40092087e-01 1.78555753e-02 2.50024587e-01 2.87853926e-01
-2.84710348e-01 8.46291006e-01 1.00084352e+00 -3.39660719e-02
-3.12597811e-01 -1.29254758e-02 -7.35613525e-01 5.82316160e-01
8.46111059e-01 -6.28183603e-01 2.97706574e-01 -6.26724958e-01
1.70339495e-01 -3.28162670e-01 -1.49030611e-01 -1.05074751e+00
3.60732377e-01 -1.43566594e-01 5.27991593e-01 -3.81682664e-01
-1.88890304e-02 -1.06949317e+00 1.35245875e-01 4.95650828e-01
4.94039059e-02 2.46377528e-01 -1.05163222e-02 3.73210698e-01
-4.38378215e-01 -6.33103251e-01 5.17968178e-01 -7.62794018e-02
-1.21796441e+00 1.85223222e-01 -5.54496229e-01 -7.16848671e-02
8.80723119e-01 -4.37249184e-01 -1.44574195e-01 -2.67058551e-01
-5.64951718e-01 1.65415585e-01 4.60047454e-01 6.52392030e-01
6.01579547e-01 -1.65546083e+00 -5.67903757e-01 1.53520375e-01
1.41949654e-01 -3.39969099e-01 3.86038363e-01 4.66637671e-01
-8.20953906e-01 1.78019539e-01 -5.30801564e-02 1.04212873e-02
-8.45430672e-01 4.66704905e-01 1.49137631e-01 5.96452355e-02
-3.85732144e-01 7.04920828e-01 -8.40610918e-03 -8.21066141e-01
3.67586017e-02 -1.77511156e-01 -5.62219858e-01 -3.46569233e-02
6.88187301e-01 3.71168733e-01 -2.50906825e-01 -9.29204285e-01
-5.22721052e-01 6.11630142e-01 -2.41994113e-01 -5.24650887e-02
9.82083738e-01 -2.43685991e-01 -3.80409688e-01 3.55538160e-01
1.42264938e+00 7.96600044e-01 -3.86632025e-01 -2.04746723e-01
1.79902941e-01 -3.87239307e-01 -2.90277898e-01 -4.27527010e-01
-8.68268967e-01 7.86075830e-01 4.16645199e-01 1.57796621e-01
8.50257158e-01 -7.57043213e-02 1.35684454e+00 8.00311923e-01
3.02986741e-01 -1.19858313e+00 -3.36592913e-01 8.63028526e-01
3.20464224e-01 -9.68345940e-01 -2.93489456e-01 -5.10646999e-01
-9.64843750e-01 1.38286102e+00 6.26317978e-01 3.96985374e-02
7.41981626e-01 -3.91666107e-02 3.60668480e-01 4.05883163e-01
1.56775206e-01 -2.00775698e-01 9.95121896e-02 5.82140565e-01
4.53988194e-01 -3.33456062e-02 -9.45114136e-01 1.00578511e+00
-2.13110983e-01 -4.96990353e-01 6.46242857e-01 8.45494568e-01
-4.38673615e-01 -1.56224573e+00 -3.44887137e-01 2.14001939e-01
-5.93534887e-01 -5.01457334e-01 -1.71627015e-01 6.06830180e-01
3.40581417e-01 1.03612351e+00 9.44428965e-02 -5.77164829e-01
2.56315649e-01 8.54993016e-02 -3.24981809e-01 -6.54290915e-01
-4.38292563e-01 7.34579563e-02 -8.77547786e-02 -1.44844830e-01
-2.56977499e-01 -4.40981239e-01 -1.79108489e+00 -3.78313482e-01
-8.45576525e-01 7.32550979e-01 6.65855229e-01 7.76519597e-01
4.28169191e-01 1.03532344e-01 9.82398570e-01 -9.66942608e-02
-4.26866055e-01 -7.67584920e-01 -8.51337552e-01 3.61848831e-01
-3.48518103e-01 -2.32067212e-01 -3.94258469e-01 -1.43700838e-01] | [9.841055870056152, 9.805038452148438] |
21f97291-83a1-4596-bb08-adf456d62efa | learning-semantic-relationship-among | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Fu_Learning_Semantic_Relationship_Among_Instances_for_Image-Text_Matching_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Fu_Learning_Semantic_Relationship_Among_Instances_for_Image-Text_Matching_CVPR_2023_paper.pdf | Learning Semantic Relationship Among Instances for Image-Text Matching | Image-text matching, a bridge connecting image and language, is an important task, which generally learns a holistic cross-modal embedding to achieve a high-quality semantic alignment between the two modalities. However, previous studies only focus on capturing fragment-level relation within a sample from a particular modality, e.g., salient regions in an image or text words in a sentence, where they usually pay less attention to capturing instance-level interactions among samples and modalities, e.g., multiple images and texts. In this paper, we argue that sample relations could help learn subtle differences for hard negative instances, and thus transfer shared knowledge for infrequent samples should be promising in obtaining better holistic embeddings. Therefore, we propose a novel hierarchical relation modeling framework (HREM), which explicitly capture both fragment- and instance-level relations to learn discriminative and robust cross-modal embeddings. Extensive experiments on Flickr30K and MS-COCO show our proposed method outperforms the state-of-the-art ones by 4%-10% in terms of rSum. | ['Yongdong Zhang', 'Yan Song', 'Zhendong Mao', 'Zheren Fu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['network-embedding', 'cross-modal-retrieval', 'text-matching', 'multimodal-deep-learning'] | ['methodology', 'miscellaneous', 'natural-language-processing', 'natural-language-processing'] | [ 1.07612617e-01 -3.30457389e-01 -3.95306408e-01 -5.34940839e-01
-9.24420714e-01 -2.80400127e-01 7.78007746e-01 3.54361176e-01
-3.18666041e-01 2.49961495e-01 4.33574021e-01 2.36645877e-01
-1.86835006e-01 -6.54460490e-01 -7.11974382e-01 -6.99351609e-01
2.88786590e-01 -9.84041858e-03 1.23077080e-01 5.22867106e-02
-3.15772579e-03 8.52460265e-02 -1.38286877e+00 6.88111842e-01
6.83547497e-01 1.21639419e+00 2.34985247e-01 1.12542575e-02
-3.23614419e-01 7.75384724e-01 -5.90807721e-02 -6.25054061e-01
-1.24065243e-01 -4.57030475e-01 -5.86806118e-01 3.65280032e-01
9.23728049e-01 -1.07624993e-01 -6.22215033e-01 1.36185980e+00
5.41296005e-01 1.13980539e-01 6.99342370e-01 -1.23136294e+00
-9.93885398e-01 5.11853993e-01 -1.06087124e+00 1.74571976e-01
2.84591824e-01 1.46860898e-01 1.26605344e+00 -1.14002120e+00
5.17319381e-01 1.39358771e+00 3.52683842e-01 3.71885329e-01
-1.26956725e+00 -7.70919859e-01 3.50535870e-01 5.11719406e-01
-1.60609078e+00 -4.21125829e-01 1.05311227e+00 -4.34469849e-01
6.25415206e-01 3.58146876e-01 5.02047777e-01 1.07160902e+00
6.44144341e-02 1.16574049e+00 1.16473556e+00 -2.54748702e-01
-2.22279400e-01 4.89577532e-01 2.46661305e-02 5.31790733e-01
2.73363348e-02 -3.77852768e-01 -7.92246044e-01 1.42294496e-01
5.74512780e-01 3.30085516e-01 -5.28134942e-01 -3.57635885e-01
-1.53404796e+00 5.78582644e-01 6.70976698e-01 6.09251022e-01
-2.00709283e-01 -6.58273548e-02 6.03730023e-01 1.90869421e-01
5.74637532e-01 2.30870806e-02 -3.21868122e-01 9.34500247e-02
-7.07810581e-01 -8.15662369e-02 1.76013961e-01 1.06378984e+00
1.05437171e+00 -4.79230732e-01 -3.17848027e-01 1.08076823e+00
4.89530116e-01 5.83718300e-01 3.93106967e-01 -4.22062159e-01
8.85848880e-01 1.05791378e+00 -3.00681084e-01 -1.54865181e+00
8.58186036e-02 -3.72888595e-01 -1.04880261e+00 -4.80801970e-01
2.64291409e-02 4.44111466e-01 -4.92174298e-01 1.63951993e+00
5.10422170e-01 3.08324814e-01 4.74802665e-02 1.00768912e+00
1.10523999e+00 6.27110660e-01 1.61728188e-01 -2.75183916e-01
1.74860883e+00 -1.12716389e+00 -7.45243073e-01 -4.16475862e-01
4.99112725e-01 -8.66366863e-01 1.33689177e+00 1.02386670e-02
-8.31951916e-01 -7.22895145e-01 -6.60316348e-01 -3.00470382e-01
-4.28355157e-01 2.43287787e-01 4.20189083e-01 7.06122890e-02
-4.29704309e-01 2.54759997e-01 -4.56906438e-01 -4.21386421e-01
5.07205665e-01 -5.44719882e-02 -6.86170280e-01 -7.36032903e-01
-1.27900481e+00 6.55726135e-01 3.27526927e-01 1.24197327e-01
-7.01066256e-01 -6.81901395e-01 -1.04290712e+00 -1.01394467e-01
4.71643418e-01 -4.78955775e-01 6.41254783e-01 -1.10240746e+00
-8.76291394e-01 1.16357851e+00 -3.94720048e-01 9.62907895e-02
2.93173879e-01 -2.33526379e-01 -6.02074265e-01 3.62914741e-01
9.52598229e-02 6.70928121e-01 9.05525327e-01 -1.39327276e+00
-5.25582433e-01 -6.39041185e-01 1.86053917e-01 4.62641329e-01
-8.15379083e-01 -6.74336479e-05 -7.76151776e-01 -6.53451383e-01
6.43196627e-02 -5.45126081e-01 2.82768250e-01 1.77318320e-01
-2.51122236e-01 -5.32785237e-01 9.98832822e-01 -7.23415792e-01
1.20348465e+00 -2.29942250e+00 3.07912737e-01 -1.89410686e-01
8.06529522e-02 -7.93467648e-03 -3.62874269e-01 5.05283892e-01
-1.99103132e-01 4.96245064e-02 -2.66710389e-03 -4.73940432e-01
-8.38074237e-02 1.94718868e-01 -2.09501103e-01 5.71647167e-01
3.43121290e-01 1.08920538e+00 -1.05295408e+00 -8.94104421e-01
1.80278808e-01 5.37219584e-01 -1.31039470e-01 3.24556500e-01
-7.11051971e-02 4.18969661e-01 -4.01839793e-01 7.61704385e-01
7.22788870e-01 -3.85443598e-01 -2.43535470e-02 -9.91533518e-01
1.89098954e-01 -2.51575131e-02 -9.46339309e-01 2.03098774e+00
-6.49856687e-01 5.89465857e-01 -1.83503866e-01 -1.27970862e+00
7.16642559e-01 2.22353294e-01 4.02729928e-01 -1.07772183e+00
6.71074167e-02 9.27638356e-03 -3.04750383e-01 -8.66146326e-01
3.40994984e-01 -1.80929348e-01 -5.61076589e-02 2.08367512e-01
8.69080126e-02 1.80500314e-01 7.30241602e-03 2.40854904e-01
6.44260347e-01 -4.25553247e-02 2.55146027e-01 -1.43449977e-01
6.98012412e-01 -2.83896863e-01 4.78965998e-01 2.97468245e-01
-1.93911269e-01 6.04851723e-01 3.36340696e-01 -5.28700985e-02
-8.08372319e-01 -8.62964392e-01 -2.63270319e-01 1.04745281e+00
7.80080616e-01 -4.93155807e-01 -3.46154213e-01 -8.05726826e-01
-5.65037765e-02 4.67519790e-01 -7.89456129e-01 -2.68215626e-01
-3.42573822e-01 -6.13219917e-01 1.71114266e-01 3.78024638e-01
7.67717779e-01 -8.34001541e-01 4.32780609e-02 -1.50006711e-01
-5.67812204e-01 -1.50719786e+00 -8.69606793e-01 -3.02906156e-01
-7.29803443e-01 -1.06314075e+00 -6.94345474e-01 -9.88285482e-01
7.81847417e-01 6.94450438e-01 1.25154269e+00 -1.01322290e-02
-2.36698449e-01 5.44846594e-01 -5.90208948e-01 1.44554004e-01
5.76047935e-02 -1.44226149e-01 -2.13670522e-01 6.08104289e-01
6.68493211e-01 -3.44723225e-01 -8.12364280e-01 3.25747281e-01
-1.15091479e+00 2.58834988e-01 7.95887232e-01 9.93622065e-01
8.43778372e-01 1.90335065e-01 2.13441044e-01 -5.97863317e-01
4.19121832e-01 -6.80047989e-01 4.40448225e-02 7.25712538e-01
-2.63367563e-01 -1.63293540e-01 6.30392671e-01 -7.11656034e-01
-1.00315273e+00 -9.53351557e-02 2.35754266e-01 -7.04877973e-01
-2.49138728e-01 6.49981976e-01 -6.25012517e-01 7.36488402e-02
2.86021024e-01 6.36154890e-01 -1.21459231e-01 -3.44371736e-01
5.21121621e-01 7.88668036e-01 3.03307116e-01 -7.36237705e-01
7.59216607e-01 7.02244937e-01 -1.26984760e-01 -7.82774448e-01
-1.02870715e+00 -8.28538775e-01 -6.17845953e-01 -2.94285715e-01
8.52627277e-01 -1.29301775e+00 -3.18270564e-01 3.96962792e-01
-9.73862886e-01 -4.25645411e-02 -1.16175160e-01 4.56508219e-01
-2.10662410e-01 4.95778680e-01 -5.01264513e-01 -4.93578881e-01
-1.18104927e-01 -1.02224910e+00 1.56721222e+00 4.32659984e-01
2.55261391e-01 -1.03863549e+00 -7.02992231e-02 6.77025795e-01
-1.45008797e-02 8.98447558e-02 7.13556409e-01 -4.95397925e-01
-6.95497155e-01 -1.39812946e-01 -7.40061402e-01 4.66198862e-01
4.08252090e-01 -1.42059654e-01 -8.40475321e-01 -3.35982680e-01
-3.92520763e-02 -6.31128371e-01 9.21089947e-01 -2.12681726e-01
1.27080464e+00 -3.43269169e-01 -4.44774538e-01 2.85943955e-01
1.50373685e+00 -2.91245610e-01 7.56630063e-01 6.27490729e-02
1.04958117e+00 8.05954635e-01 9.78752971e-01 2.15815693e-01
6.97576761e-01 8.69974017e-01 3.44811291e-01 -1.31602004e-01
-1.62065700e-01 -3.56096268e-01 4.41002011e-01 1.10835278e+00
2.35733375e-01 4.23864499e-02 -6.71220362e-01 7.82029092e-01
-2.07718015e+00 -1.04321849e+00 -1.77472278e-01 1.92221951e+00
1.17142713e+00 -2.48878196e-01 -9.47948396e-02 -1.44825250e-01
7.91729152e-01 3.82058144e-01 -5.28478742e-01 3.77376348e-01
-2.31183603e-01 -2.79355347e-01 7.17723221e-02 2.07609892e-01
-1.11669171e+00 7.98621058e-01 3.91087031e+00 1.32953429e+00
-1.12548113e+00 3.33704770e-01 7.71072745e-01 -1.33472040e-01
-5.30377030e-01 -1.40935659e-01 -6.77322626e-01 5.81928551e-01
4.45797920e-01 8.56067762e-02 3.06998104e-01 6.27125919e-01
4.76320609e-02 1.41942175e-02 -1.23446381e+00 1.45705402e+00
5.23307562e-01 -1.03311586e+00 3.63529235e-01 1.34734772e-02
8.03825319e-01 -2.84005970e-01 1.65609166e-01 2.98105508e-01
-2.98936129e-01 -1.04817855e+00 5.91188967e-01 7.07769930e-01
8.06543052e-01 -7.43186295e-01 8.20531964e-01 2.77674764e-01
-1.61763155e+00 3.13106954e-01 -4.14879590e-01 3.72729361e-01
1.91652570e-02 8.39964211e-01 -4.47841287e-02 9.24392760e-01
9.04297769e-01 1.24353790e+00 -6.90617919e-01 5.69902599e-01
-9.64494720e-02 2.68742859e-01 -5.72256418e-03 1.61893740e-01
2.12414369e-01 -1.84263840e-01 2.06763104e-01 1.27296960e+00
1.48553774e-01 1.63172737e-01 2.46270403e-01 7.11502016e-01
-1.96636856e-01 3.49698573e-01 -5.27778983e-01 -2.43281960e-01
3.95249277e-01 1.47156703e+00 -3.72426867e-01 -2.73891121e-01
-8.92464161e-01 1.19467115e+00 5.43321967e-01 3.70507240e-01
-9.04223084e-01 -1.96509749e-01 7.38414943e-01 -3.94994728e-02
3.78581852e-01 -8.13401490e-02 -9.11449362e-03 -1.69559360e+00
3.94345760e-01 -9.85236049e-01 4.73083854e-01 -7.10353613e-01
-1.83788562e+00 3.89397979e-01 -2.00887844e-01 -1.53217006e+00
3.11935484e-01 -5.01420856e-01 -2.88970321e-01 9.17197704e-01
-1.68828368e+00 -1.67151022e+00 -7.56796598e-01 8.12762022e-01
5.29534876e-01 7.17519373e-02 4.79929417e-01 6.37902737e-01
-5.80974340e-01 8.02816987e-01 4.79408773e-03 3.37544918e-01
9.04730380e-01 -9.03568029e-01 -2.55496204e-01 6.96254969e-01
4.85903084e-01 7.37015367e-01 3.93621325e-01 -3.93127233e-01
-1.59629929e+00 -1.34867764e+00 7.42296517e-01 -3.53204072e-01
7.23576128e-01 -2.83139616e-01 -1.16917098e+00 4.62200314e-01
3.86502922e-01 5.09452343e-01 8.43236864e-01 1.99263394e-01
-7.54866898e-01 -6.01608932e-01 -9.35976386e-01 7.50652790e-01
1.14483404e+00 -1.09730792e+00 -5.77962637e-01 2.06748009e-01
6.98527753e-01 -3.30147684e-01 -1.13162482e+00 5.55884063e-01
4.60410655e-01 -9.27677453e-01 9.08169031e-01 -5.00516295e-01
7.56309807e-01 -5.63778758e-01 -5.00501573e-01 -1.19521677e+00
-2.04917192e-01 5.52356578e-02 -3.55828643e-01 1.73232055e+00
-3.77277918e-02 -3.47551137e-01 3.65295470e-01 1.48346841e-01
1.46698460e-01 -9.51502442e-01 -8.86933804e-01 -6.20450556e-01
-2.72671729e-02 -5.22370934e-01 5.17550707e-01 1.40971243e+00
-7.46458545e-02 5.38763404e-01 -4.92275506e-01 3.79624873e-01
6.34202540e-01 5.33505380e-01 7.66999424e-01 -6.95231080e-01
-2.30840117e-01 -4.53439534e-01 -5.88088274e-01 -1.27416384e+00
3.62738937e-01 -8.14830005e-01 -1.01092674e-01 -1.45136571e+00
8.27798903e-01 -3.24328840e-01 -5.46762288e-01 2.73774683e-01
-6.22022212e-01 3.94318193e-01 1.08478991e-02 3.26326132e-01
-9.46585238e-01 9.28925276e-01 1.44838440e+00 -6.31832361e-01
2.88888812e-01 -6.37699842e-01 -7.37508535e-01 5.12729108e-01
4.11639899e-01 -3.42669278e-01 -5.27696550e-01 -5.01061678e-01
3.07867020e-01 -1.76509649e-01 5.60604990e-01 -6.81149364e-01
3.16995174e-01 -3.45615298e-01 3.44738334e-01 -5.64523339e-01
2.39869833e-01 -1.25878060e+00 5.40085025e-02 9.38161165e-02
-4.94393915e-01 -2.20733583e-01 -1.26754520e-02 9.19087589e-01
-8.00677359e-01 3.12012974e-02 6.58611774e-01 -6.10691160e-02
-8.86096179e-01 6.37803853e-01 5.35186768e-01 3.33455294e-01
1.08392668e+00 -1.09837450e-01 -3.59867156e-01 -2.24285245e-01
-4.06290412e-01 3.70746076e-01 4.89865780e-01 7.41571069e-01
7.84978092e-01 -1.85444903e+00 -6.35418415e-01 -4.57797684e-02
7.83904970e-01 7.98586011e-02 7.98569918e-01 1.08766854e+00
2.19939761e-02 2.30838224e-01 1.19098946e-01 -8.06950271e-01
-1.46160579e+00 5.78957200e-01 3.31006229e-01 -2.06105590e-01
-3.83657008e-01 9.32841301e-01 5.71136892e-01 -3.57158929e-01
1.09047771e-01 -1.04094729e-01 -1.70664549e-01 4.69507247e-01
7.12566853e-01 -1.27079204e-01 -1.98874012e-01 -1.09171963e+00
-6.31527424e-01 8.92860293e-01 -2.17787504e-01 2.53742486e-01
1.05564010e+00 -4.36449975e-01 -3.15940738e-01 7.67860711e-01
1.62961960e+00 -1.22757405e-01 -1.14710271e+00 -9.23950851e-01
-2.95549154e-01 -8.47908676e-01 -4.93516214e-02 -4.80690956e-01
-1.29672360e+00 1.17199326e+00 5.52283406e-01 -4.72134864e-03
1.32214558e+00 4.97943342e-01 8.91533375e-01 9.24705714e-02
2.19662622e-01 -8.85577619e-01 5.30794382e-01 1.16187692e-01
9.62187231e-01 -1.55392027e+00 3.41159068e-02 -3.87102842e-01
-7.42255330e-01 1.06739342e+00 9.03120577e-01 5.74610867e-02
6.02529049e-01 -4.29552585e-01 -1.24466948e-01 -2.04577923e-01
-7.28891909e-01 -3.21082622e-01 7.92955875e-01 4.23671335e-01
5.44997156e-01 5.48058413e-02 -8.50331187e-02 5.62415659e-01
4.62661743e-01 -2.37688571e-01 -1.54793277e-01 6.53991401e-01
-1.38865620e-01 -9.96868193e-01 -1.36126965e-01 4.63760912e-01
-3.28673124e-01 -2.80554116e-01 -2.76846558e-01 6.82750702e-01
3.82688999e-01 7.69326568e-01 1.61668912e-01 -5.72455943e-01
3.84598643e-01 -2.27084950e-01 5.91895044e-01 -5.30423641e-01
-2.29163006e-01 1.49341626e-03 -6.27798960e-02 -5.73014796e-01
-9.55333412e-01 -7.12692618e-01 -8.99459481e-01 -1.99982613e-01
-3.23072851e-01 -7.27753490e-02 1.17930859e-01 1.06731308e+00
3.99987429e-01 4.54932183e-01 7.44261920e-01 -5.71442664e-01
-1.84333205e-01 -9.33639169e-01 -7.07629204e-01 8.44918728e-01
2.82446116e-01 -5.52186131e-01 -4.60024655e-01 1.32243484e-01] | [10.820562362670898, 1.3179210424423218] |
7d5749ec-ca69-4100-883e-3e062efc3b87 | uncovering-the-potential-of-chatgpt-for | 2305.08391 | null | https://arxiv.org/abs/2305.08391v1 | https://arxiv.org/pdf/2305.08391v1.pdf | Uncovering the Potential of ChatGPT for Discourse Analysis in Dialogue: An Empirical Study | Large Language Models (LLMs) like ChatGPT have proven a great shallow understanding of many traditional NLP tasks, such as translation, summarization, etc. However, its performance on high-level understanding, such as dialogue discourse analysis task that requires a higher level of understanding and reasoning, remains less explored. This study investigates ChatGPT's capabilities in three dialogue discourse tasks: topic segmentation, discourse relation recognition, and discourse parsing, of varying difficulty levels. To adapt ChatGPT to these tasks, we propose discriminative and generative paradigms and introduce the Chain of Thought (COT) approach to improve ChatGPT's performance in more difficult tasks. The results show that our generative paradigm allows ChatGPT to achieve comparative performance in the topic segmentation task comparable to state-of-the-art methods but reveals room for improvement in the more complex tasks of discourse relation recognition and discourse parsing. Notably, the COT can significantly enhance ChatGPT's performance with the help of understanding complex structures in more challenging tasks. Through a series of case studies, our in-depth analysis suggests that ChatGPT can be a good annotator in topic segmentation but has difficulties understanding complex rhetorical structures. We hope these findings provide a foundation for future research to refine dialogue discourse analysis approaches in the era of LLMs. | ['Feng Jiang', 'Yaxin Fan'] | 2023-05-15 | null | null | null | null | ['discourse-parsing'] | ['natural-language-processing'] | [ 3.66529197e-01 9.80110943e-01 -1.34953156e-01 -2.71823049e-01
-1.05873108e+00 -7.28584170e-01 1.10177672e+00 4.36760366e-01
-3.66595536e-02 6.33873940e-01 7.46738851e-01 -8.47916603e-01
8.41794834e-02 -5.70376277e-01 -5.47644980e-02 -2.53047377e-01
1.21273026e-01 8.34463358e-01 3.84211302e-01 -4.90540713e-01
4.96647149e-01 -1.92107782e-01 -9.24025416e-01 8.59452367e-01
1.00326788e+00 2.71972209e-01 2.02354208e-01 7.84798920e-01
-6.49261951e-01 1.36498439e+00 -1.02955234e+00 -3.86310220e-01
-4.07518655e-01 -7.22758234e-01 -1.75326490e+00 7.80266374e-02
9.91325825e-02 -2.10737139e-01 1.00374557e-01 4.24560517e-01
2.89744318e-01 9.06996429e-02 5.77192128e-01 -9.24376309e-01
-2.97033012e-01 9.83122945e-01 -1.26390964e-01 3.03656876e-01
8.65290165e-01 3.14474888e-02 1.12997270e+00 -3.51137370e-01
8.37459922e-01 1.78328001e+00 6.89460218e-01 5.68211079e-01
-1.18618989e+00 -1.67583734e-01 3.16715330e-01 1.31479666e-01
-6.24660134e-01 -4.31849122e-01 4.98940974e-01 -4.98077571e-01
1.46152318e+00 4.48960125e-01 4.78428900e-01 1.10802925e+00
5.02087362e-02 1.13785803e+00 1.18617868e+00 -5.50529897e-01
-1.82140060e-02 1.37137190e-01 5.05255997e-01 3.96091998e-01
-3.38326603e-01 -7.35400259e-01 -6.29196942e-01 -1.89084157e-01
4.19203401e-01 -6.88454688e-01 -2.53471017e-01 4.28037465e-01
-1.34991622e+00 1.19882035e+00 -1.32443070e-01 7.69633174e-01
-5.14750034e-02 -2.79743701e-01 7.72830248e-01 5.43743074e-01
8.41239333e-01 9.59935606e-01 -3.01263183e-01 -9.47376847e-01
-7.38117814e-01 4.35655802e-01 1.74686837e+00 9.18058813e-01
2.25136429e-01 -2.51829714e-01 -3.91959190e-01 9.31671977e-01
1.46120816e-01 1.25145847e-02 2.41704285e-01 -1.10691690e+00
9.08218086e-01 1.02930200e+00 -2.91322917e-01 -8.06370020e-01
-5.04031360e-01 2.78012246e-01 -3.77361149e-01 -2.91816682e-01
7.44259596e-01 -3.29157978e-01 -3.64488542e-01 1.45167005e+00
3.18711281e-01 -4.68058765e-01 2.54756242e-01 6.00587368e-01
1.17707419e+00 1.00530064e+00 2.04795375e-01 -4.12842304e-01
1.76553655e+00 -1.03530502e+00 -9.07251179e-01 -4.85306650e-01
1.29058349e+00 -9.67623293e-01 9.87307668e-01 2.19970241e-01
-1.19922304e+00 -2.27444455e-01 -5.65246701e-01 -5.28106809e-01
-4.66832630e-02 -9.14603937e-03 9.42027569e-01 6.94700718e-01
-9.12699699e-01 2.53427595e-01 -8.14256191e-01 -6.00289524e-01
3.54706854e-01 7.79764354e-02 -7.83211142e-02 9.09411628e-03
-1.18372667e+00 1.29160094e+00 3.40714812e-01 -1.78275272e-01
-4.63388979e-01 -5.04469395e-01 -1.00862420e+00 8.30411762e-02
7.92975426e-01 -5.09949028e-01 1.72090900e+00 -5.60441911e-01
-1.72821593e+00 9.19246435e-01 -3.15972835e-01 -5.39853990e-01
4.54304755e-01 -4.37607795e-01 2.49113634e-01 2.36043260e-01
2.46786013e-01 5.02064466e-01 2.68110335e-01 -8.45760167e-01
-4.45731550e-01 -1.04019798e-01 4.28534538e-01 5.97566485e-01
5.57595724e-03 4.68870223e-01 -1.13628305e-01 -2.77793795e-01
2.72646062e-02 -7.98546195e-01 5.32537124e-05 -5.44996917e-01
-6.18103266e-01 -8.23558211e-01 1.03956878e+00 -9.40795064e-01
1.42475355e+00 -1.76431537e+00 3.76504004e-01 -3.54115367e-01
4.07868773e-01 4.27630335e-01 2.46828154e-01 1.05835390e+00
3.69711757e-01 5.12458324e-01 -1.45189881e-01 -5.44112206e-01
1.66183963e-01 3.43821317e-01 -1.53240293e-01 -1.48326516e-01
3.72076392e-01 1.18558609e+00 -8.22486639e-01 -6.35156453e-01
2.31260166e-01 7.31775463e-02 -3.82223159e-01 3.08254510e-01
-7.34035254e-01 5.89956403e-01 -5.77074409e-01 2.52583772e-01
-1.25191391e-01 -5.13244271e-01 6.00333154e-01 3.26540858e-01
-2.35142142e-01 1.14875579e+00 -6.29441023e-01 1.55389929e+00
-3.85539144e-01 1.26725769e+00 2.26120353e-01 -9.85243797e-01
7.15004742e-01 6.37559891e-01 -2.83842832e-02 -3.03992987e-01
2.32866570e-01 -5.54731674e-02 5.74551225e-01 -7.40380287e-01
7.75877535e-01 -1.66942433e-01 -2.58995533e-01 1.00025177e+00
-1.08462527e-01 -4.31702167e-01 3.93466175e-01 6.08658254e-01
1.09144151e+00 -1.40016034e-01 3.70285600e-01 -3.06886315e-01
4.49234068e-01 5.61768055e-01 9.43769887e-02 6.95842206e-01
-6.44253660e-03 3.21187079e-01 1.14250886e+00 -7.43417963e-02
-1.03489256e+00 -4.29804802e-01 2.01851763e-02 1.30799055e+00
-1.84398398e-01 -7.62463629e-01 -1.10919189e+00 -6.29229069e-01
-3.79407436e-01 1.01170528e+00 -3.13219190e-01 4.13513929e-01
-1.05374777e+00 -7.90509522e-01 8.53790462e-01 3.27435017e-01
7.44225919e-01 -1.26276958e+00 -5.53627968e-01 3.02022785e-01
-6.45187795e-01 -1.35133517e+00 -6.96433932e-02 -1.24807023e-01
-8.12568665e-01 -1.20832884e+00 -4.13126767e-01 -8.53709042e-01
2.39217415e-01 1.39170721e-01 1.06293619e+00 1.80260107e-01
1.19339935e-01 4.44096178e-01 -6.81966543e-01 -4.44860548e-01
-1.23427963e+00 6.86371386e-01 -6.64193749e-01 -7.53850639e-01
3.50684434e-01 -3.41043353e-01 -1.15956314e-01 2.43611753e-01
-7.53765881e-01 5.85243404e-01 4.36971188e-02 7.99144089e-01
-3.35631341e-01 -2.50563055e-01 6.73981547e-01 -1.34930718e+00
1.36817598e+00 -4.15013105e-01 1.18006580e-01 1.99509993e-01
-2.46580049e-01 -1.31656915e-01 3.16185057e-01 -4.49885488e-01
-1.47281408e+00 -9.21855569e-01 -3.57297719e-01 4.62893486e-01
-1.14070021e-01 7.47591197e-01 5.16300313e-02 3.05465132e-01
6.05785012e-01 -1.23743311e-01 2.47516707e-01 -4.75239247e-01
3.65488172e-01 7.30632424e-01 6.61790743e-02 -8.18392098e-01
3.06045830e-01 -6.58767521e-02 -4.21720058e-01 -1.13456571e+00
-8.72182190e-01 -5.04807591e-01 -5.30583084e-01 -8.78947824e-02
1.06209099e+00 -7.66979396e-01 -8.39436650e-01 3.38391006e-01
-1.46656239e+00 -8.58938873e-01 -1.73232406e-01 1.96946025e-01
-4.65473175e-01 6.74146533e-01 -1.07554448e+00 -9.61670041e-01
-3.54337960e-01 -1.07763147e+00 1.10891807e+00 2.96279281e-01
-1.09756458e+00 -1.54485190e+00 -1.93113610e-02 9.93040860e-01
2.77393728e-01 1.82255954e-01 1.29409432e+00 -1.21533132e+00
-5.54326475e-01 1.70703843e-01 -1.22757450e-01 1.96382090e-01
-1.22523038e-02 -1.70873165e-01 -9.04741287e-01 3.79318604e-03
3.61205190e-01 -4.93537575e-01 4.68709737e-01 7.12738335e-02
3.69540513e-01 -5.52390397e-01 -3.23471248e-01 -1.26460701e-01
5.87236881e-01 3.94004643e-01 6.43761814e-01 4.46182668e-01
6.08613133e-01 1.03925943e+00 5.87086678e-01 3.59902717e-02
8.28946292e-01 6.00922108e-01 -1.36598989e-01 1.60736293e-02
-1.81127205e-01 -5.63068911e-02 4.58021343e-01 1.22770131e+00
-2.06730999e-02 -5.72841287e-01 -1.30909586e+00 4.80438530e-01
-2.14279413e+00 -8.75602722e-01 -3.76691967e-01 1.29408586e+00
1.03876245e+00 1.75192803e-01 1.39834926e-01 1.38535753e-01
4.11504775e-01 4.93623555e-01 -5.63713200e-02 -1.03459334e+00
-4.84419391e-02 1.61796585e-02 -4.04598057e-01 6.38669312e-01
-8.90266836e-01 1.09379518e+00 6.31411791e+00 8.84997070e-01
-7.36118972e-01 2.16811761e-01 7.39629805e-01 3.80401820e-01
-2.36156628e-01 2.84137607e-01 -7.22289920e-01 2.74334788e-01
9.82578278e-01 -1.98629409e-01 1.43074691e-01 6.89189315e-01
2.18103454e-01 -5.92402160e-01 -1.30310905e+00 3.56927454e-01
7.56578222e-02 -1.51046300e+00 -8.86906758e-02 6.64091855e-02
4.21520144e-01 -3.39777738e-01 -5.06747723e-01 6.86580420e-01
2.89530754e-01 -1.17343068e+00 1.91222504e-01 5.65456785e-03
3.31937522e-01 -2.07644358e-01 8.78970861e-01 8.07459891e-01
-8.27153504e-01 1.99140683e-01 1.72465846e-01 -6.58855379e-01
5.90799570e-01 6.15592599e-02 -1.44102252e+00 5.76218963e-01
9.98318940e-02 4.57892090e-01 -3.42920721e-01 4.31494385e-01
-4.32819366e-01 1.06047797e+00 -2.54538059e-01 -3.35896373e-01
5.05568624e-01 -1.37461036e-01 7.93694258e-01 1.54902411e+00
-2.79067129e-01 6.04206920e-01 5.87265790e-01 7.22955585e-01
1.01587936e-01 1.68358773e-01 -3.82125974e-01 -3.94460261e-01
4.36101824e-01 1.03429306e+00 -9.36254561e-01 -5.56240737e-01
-1.98719785e-01 5.19342542e-01 2.06901446e-01 9.26860198e-02
-3.72378767e-01 -4.36691046e-02 2.55003154e-01 7.39116892e-02
-1.28765404e-01 -3.81029129e-01 -5.86549401e-01 -1.05372155e+00
-5.33008240e-02 -1.29295468e+00 3.10415030e-01 -6.18789196e-01
-1.02704048e+00 5.59220195e-01 4.56209421e-01 -4.82357442e-01
-8.05338383e-01 -4.57208872e-01 -1.03268695e+00 8.20306540e-01
-9.94961977e-01 -1.31667137e+00 2.77819224e-02 6.61173761e-02
1.28169262e+00 6.07417412e-02 9.60131705e-01 -1.95697963e-01
-5.02823114e-01 1.28566548e-01 -2.89393991e-01 4.31151658e-01
4.74422514e-01 -1.32267213e+00 6.06892943e-01 4.30159807e-01
-1.09264348e-02 7.31449366e-01 6.78443909e-01 -7.31690824e-01
-1.21472073e+00 -4.66372728e-01 1.14231718e+00 -7.40900874e-01
8.36393535e-01 -3.68522018e-01 -1.28322434e+00 8.10544431e-01
7.03997016e-01 -1.20749092e+00 8.41235697e-01 6.22964442e-01
6.71151374e-03 7.84078598e-01 -8.54931533e-01 5.44666886e-01
7.89670527e-01 -7.62225926e-01 -1.43653321e+00 5.54290891e-01
8.51641119e-01 -8.05956006e-01 -1.23265231e+00 8.23622048e-02
3.77829492e-01 -7.62029767e-01 6.40273213e-01 -5.88352621e-01
7.51639783e-01 1.77094147e-01 2.43378177e-01 -1.11825407e+00
2.21396610e-01 -1.07468867e+00 -2.32349690e-02 1.63600039e+00
4.88125563e-01 -5.69942057e-01 6.51221573e-01 8.58323991e-01
-3.47047865e-01 -8.11953664e-01 -8.94187391e-01 -4.08425510e-01
3.26599598e-01 -4.27518636e-01 4.49668430e-02 1.13261175e+00
8.42177808e-01 1.17372036e+00 -1.08641032e-02 -3.83058876e-01
1.22549292e-02 1.67865366e-01 9.65471387e-01 -1.25107515e+00
-2.11870939e-01 -4.86549646e-01 4.29682098e-02 -1.18317282e+00
2.82747954e-01 -7.39768207e-01 -9.14157461e-03 -2.02453661e+00
1.12455934e-01 -2.75417805e-01 9.29545522e-01 3.71514410e-01
-3.57247055e-01 -3.76167417e-01 5.10167658e-01 3.09785008e-01
-7.93038249e-01 3.61157805e-01 1.42008305e+00 -9.36166495e-02
-5.08960962e-01 8.32405984e-02 -7.93760121e-01 9.21042681e-01
7.81459928e-01 -3.06597948e-01 -5.30140996e-01 -3.07038665e-01
-8.28275606e-02 4.27795380e-01 -5.91226332e-02 -4.78268802e-01
2.57942080e-01 -9.67214927e-02 -1.97618842e-01 -4.56622392e-01
4.37323451e-01 1.37777124e-02 -1.67196199e-01 2.54516661e-01
-5.86986542e-01 -3.19166891e-02 4.24501300e-01 3.20282057e-02
-3.86063844e-01 -3.65969181e-01 1.73626274e-01 -3.92126739e-01
-4.19243127e-01 -5.46151459e-01 -1.01201844e+00 6.19501173e-01
7.85484374e-01 -4.13567603e-01 -8.09935153e-01 -7.29683816e-01
-7.37881362e-01 5.90427399e-01 3.18304777e-01 3.25388789e-01
2.60513365e-01 -6.27553582e-01 -7.78997004e-01 -3.74226719e-01
-4.15042698e-01 5.37563443e-01 3.33049186e-02 9.87069070e-01
-4.58580047e-01 7.92549133e-01 1.99092343e-01 -7.81147480e-01
-1.68263650e+00 -2.60881558e-02 -1.91756878e-02 -9.34780538e-01
-7.93728709e-01 6.86202347e-01 3.86139274e-01 -3.28352988e-01
1.00117259e-01 -5.13514757e-01 -5.91129601e-01 4.20352817e-01
4.86887991e-01 4.33089375e-01 -2.24128515e-01 -4.61992383e-01
-1.05939560e-01 1.64540887e-01 -4.34488952e-01 -2.99893469e-01
1.15599442e+00 -4.29842055e-01 -2.69775748e-01 6.82567120e-01
7.32803226e-01 -2.20268145e-01 -8.92166972e-01 -1.46376491e-01
5.70775330e-01 -8.53312239e-02 -4.10451114e-01 -9.24043953e-01
-4.90522496e-02 9.76123869e-01 -3.58391404e-01 9.68587279e-01
6.53666675e-01 3.74278069e-01 8.23658586e-01 6.00280166e-01
1.30388945e-01 -9.81512427e-01 2.04253227e-01 1.14384449e+00
8.75636041e-01 -1.19806123e+00 1.16916008e-01 -8.29251289e-01
-1.00029695e+00 1.26393855e+00 5.46834230e-01 2.69928098e-01
8.23641382e-03 2.97892511e-01 9.09770802e-02 -5.60494363e-01
-1.05047405e+00 -6.91609308e-02 2.06815273e-01 4.18447405e-01
8.21274459e-01 -6.14361316e-02 -4.38429981e-01 3.68889034e-01
-4.98936385e-01 -3.64241987e-01 6.47125602e-01 1.03464007e+00
-5.24378002e-01 -1.27195740e+00 -3.83849561e-01 3.82895142e-01
-5.59331119e-01 -2.88079530e-01 -9.04150188e-01 1.21723723e+00
-3.79868150e-01 1.26473296e+00 4.55791540e-02 -5.19151315e-02
1.83738485e-01 5.59555829e-01 3.53489310e-01 -1.22378409e+00
-1.22285986e+00 1.72019780e-01 1.24689126e+00 -2.10391238e-01
-7.43869305e-01 -8.17645729e-01 -1.20981264e+00 -5.25496662e-01
-4.10347193e-01 6.19847715e-01 3.94925147e-01 1.51180422e+00
1.11052014e-01 6.51503205e-01 -5.86864501e-02 -6.18987978e-01
-1.80893719e-01 -1.44197142e+00 1.52311593e-01 2.40583867e-01
8.53137895e-02 -3.67217332e-01 -2.77854875e-02 1.25464067e-01] | [10.96152400970459, 9.36782455444336] |
d6a4a6bc-511a-4969-961b-1dc106765bd0 | botied-multi-objective-bayesian-optimization | 2306.00344 | null | https://arxiv.org/abs/2306.00344v1 | https://arxiv.org/pdf/2306.00344v1.pdf | BOtied: Multi-objective Bayesian optimization with tied multivariate ranks | Many scientific and industrial applications require joint optimization of multiple, potentially competing objectives. Multi-objective Bayesian optimization (MOBO) is a sample-efficient framework for identifying Pareto-optimal solutions. We show a natural connection between non-dominated solutions and the highest multivariate rank, which coincides with the outermost level line of the joint cumulative distribution function (CDF). We propose the CDF indicator, a Pareto-compliant metric for evaluating the quality of approximate Pareto sets that complements the popular hypervolume indicator. At the heart of MOBO is the acquisition function, which determines the next candidate to evaluate by navigating the best compromises among the objectives. Multi-objective acquisition functions that rely on box decomposition of the objective space, such as the expected hypervolume improvement (EHVI) and entropy search, scale poorly to a large number of objectives. We propose an acquisition function, called BOtied, based on the CDF indicator. BOtied can be implemented efficiently with copulas, a statistical tool for modeling complex, high-dimensional distributions. We benchmark BOtied against common acquisition functions, including EHVI and random scalarization (ParEGO), in a series of synthetic and real-data experiments. BOtied performs on par with the baselines across datasets and metrics while being computationally efficient. | ['Kyunghyun Cho', 'Stephen Ra', 'Michael Maser', 'Nataša Tagasovska', 'Ji Won Park'] | 2023-06-01 | null | null | null | null | ['bayesian-optimization'] | ['methodology'] | [-1.71312258e-01 -3.40389252e-01 -3.35544907e-02 -1.12381339e-01
-1.06829691e+00 -9.12436485e-01 2.83015162e-01 4.46867496e-01
-4.55240309e-01 9.83292818e-01 3.71099636e-02 -4.48149405e-02
-9.89604235e-01 -7.49062538e-01 -5.83392620e-01 -9.64600146e-01
-3.30009639e-01 7.82478392e-01 1.59964934e-01 8.26070681e-02
6.10749006e-01 3.40957105e-01 -1.64608037e+00 -1.00210615e-01
1.13550746e+00 1.27897263e+00 1.71603858e-01 6.29979730e-01
1.49311963e-02 -2.13424206e-01 -7.74774134e-01 -4.95481849e-01
8.44645500e-02 -1.12549558e-01 -4.51542735e-01 -5.05133688e-01
-2.06022605e-01 1.17114708e-01 7.50538468e-01 9.62830126e-01
5.37396193e-01 5.89982346e-02 1.01927471e+00 -1.75886750e+00
-1.82055086e-01 4.96830642e-01 -5.82167029e-01 -9.49242339e-03
2.46768519e-01 1.20462969e-01 1.20348048e+00 -8.42847347e-01
4.48432088e-01 1.54021549e+00 6.75799906e-01 -1.09607361e-01
-1.43463254e+00 -1.75732926e-01 -2.69645512e-01 -1.05996253e-02
-1.41620779e+00 1.39771365e-02 4.83485848e-01 -5.41401863e-01
4.21040952e-01 6.56795144e-01 3.76496822e-01 5.90600669e-01
7.04315186e-01 6.60851061e-01 1.15542531e+00 -1.24037564e-01
6.51247263e-01 1.24038011e-01 8.78723860e-02 3.48383307e-01
7.64578462e-01 3.10782731e-01 -5.21238327e-01 -7.58045852e-01
1.43932223e-01 -2.97256589e-01 -2.68315673e-01 -6.24034405e-01
-1.02921104e+00 8.49224389e-01 3.92768346e-02 3.06071714e-03
-5.06150365e-01 1.53532982e-01 7.60961026e-02 -6.04528189e-02
4.04170990e-01 8.39634955e-01 -4.99051332e-01 -4.38547343e-01
-8.67409229e-01 5.60031950e-01 9.28036332e-01 6.30734444e-01
5.90438545e-01 -2.02736095e-01 -7.96225309e-01 5.51423669e-01
5.20434618e-01 5.63165665e-01 5.59866056e-03 -1.12707019e+00
3.30049932e-01 6.15267217e-01 5.36318183e-01 -7.97098517e-01
-4.35302675e-01 -9.22222614e-01 -4.30900931e-01 3.48477364e-01
7.03032494e-01 -2.10721642e-01 -4.12952632e-01 1.65726471e+00
3.94592136e-01 -3.04429770e-01 -7.45965913e-02 7.62171865e-01
2.87674189e-01 6.33958638e-01 -1.93690702e-01 -5.00564873e-01
1.15112829e+00 -4.41138029e-01 -5.40379643e-01 1.77352235e-01
2.00709522e-01 -6.44188344e-01 1.04339373e+00 7.29579806e-01
-1.21465099e+00 1.92856304e-02 -1.05354059e+00 7.02935398e-01
-4.02359486e-01 -1.73326135e-01 2.71501869e-01 9.23821330e-01
-8.89242113e-01 8.25974941e-01 -5.22664428e-01 2.18330801e-01
3.73943269e-01 3.66559088e-01 1.85575485e-01 7.40432888e-02
-7.41686940e-01 8.75531435e-01 4.47019428e-01 -1.77071780e-01
-9.34191048e-01 -9.80974674e-01 -5.27263463e-01 4.53475714e-01
5.73579013e-01 -5.80470145e-01 7.31211424e-01 -3.52148890e-01
-1.53842473e+00 2.13612616e-01 1.30111694e-01 -3.04534912e-01
6.12065315e-01 -9.80264321e-02 4.87501081e-03 -1.03366807e-01
-8.96360576e-02 3.65617931e-01 6.95860922e-01 -1.53740025e+00
-5.58529794e-01 -3.79882544e-01 -2.36675709e-01 -7.60540068e-02
-2.96886981e-01 -5.66743650e-02 -1.45069584e-01 -3.94856095e-01
-2.40654200e-01 -5.08458197e-01 -2.20371425e-01 -4.87894237e-01
-5.43278217e-01 -3.99432361e-01 2.96634227e-01 -5.31024575e-01
1.73478556e+00 -1.83512855e+00 4.79397118e-01 6.45846248e-01
1.24404589e-02 -2.26001531e-01 -9.46706682e-02 3.90237212e-01
3.42696786e-01 3.53111565e-01 -5.64965844e-01 -1.81412220e-01
6.05760574e-01 5.19422106e-02 -8.84298701e-03 6.45723581e-01
2.10119665e-01 7.24938154e-01 -1.00929308e+00 -4.67115253e-01
-9.64375660e-02 1.76069319e-01 -5.77624500e-01 -1.74362306e-03
-4.32802618e-01 2.01028347e-01 -2.89505839e-01 8.36883664e-01
8.18609118e-01 -2.87826121e-01 -1.74959563e-02 7.23945815e-03
-2.92054296e-01 -2.04013079e-01 -1.56000829e+00 1.04928875e+00
-2.00039715e-01 3.16735387e-01 1.73521996e-01 -8.21654499e-01
1.05968857e+00 -1.73606753e-01 7.76101589e-01 -3.47587824e-01
3.95447463e-01 4.24207091e-01 -1.21216483e-01 -1.88718271e-02
5.52309513e-01 7.99855143e-02 -3.34136009e-01 2.61561662e-01
5.67079037e-02 -5.85382655e-02 6.20329738e-01 -2.34878570e-01
9.82117176e-01 2.05795854e-01 4.40995336e-01 -9.05890822e-01
5.34918129e-01 -2.29774863e-01 6.08736038e-01 6.16882563e-01
-6.54576868e-02 5.51513731e-01 1.03086114e+00 2.76759565e-02
-7.48418868e-01 -1.43476319e+00 -2.93900907e-01 9.64008212e-01
3.35093468e-01 -3.24495792e-01 -5.20717025e-01 -4.98060405e-01
5.40139675e-01 9.40507710e-01 -4.13282663e-01 4.59582694e-02
-2.52689719e-01 -1.08612716e+00 2.21919775e-01 1.29259929e-01
-9.37423557e-02 -4.93833780e-01 -1.01867235e+00 3.56187671e-01
-9.34123620e-02 -6.83683038e-01 -2.64323622e-01 3.00363362e-01
-7.94715881e-01 -1.12687433e+00 -9.71522152e-01 4.92037125e-02
3.88706833e-01 -4.70912158e-01 1.45502412e+00 -3.81447136e-01
-1.88568950e-01 4.97006923e-01 -1.28113538e-01 -6.92364037e-01
-2.94343531e-01 -1.99915141e-01 -3.52380574e-02 1.56223923e-01
-1.29762515e-01 -5.53669333e-01 -5.34971237e-01 6.28644586e-01
-8.22641373e-01 -6.70985460e-01 3.88522685e-01 7.90164053e-01
9.61908937e-01 2.17792630e-01 5.41799307e-01 -9.91462767e-02
9.48856473e-01 -6.33918643e-01 -1.33257794e+00 5.43661356e-01
-6.67088687e-01 5.86381197e-01 5.22447169e-01 -3.37338835e-01
-5.93338072e-01 -3.39920163e-01 2.69771457e-01 -3.27634752e-01
3.47692311e-01 6.92663789e-01 -2.03782171e-01 1.25476688e-01
4.59571928e-01 -1.40997216e-01 -1.15493551e-01 -6.34432077e-01
6.27486929e-02 3.95145476e-01 4.34117079e-01 -9.10714030e-01
5.22884309e-01 2.94860601e-01 6.05240583e-01 -5.74247122e-01
-6.14629686e-01 -3.88715625e-01 -7.41945803e-02 -4.37065244e-01
5.92896938e-01 -9.81319547e-02 -1.30603194e+00 1.94632225e-02
-1.12795293e+00 2.77369171e-02 -4.37901109e-01 5.64180493e-01
-6.51684463e-01 2.23407298e-01 3.76958400e-01 -1.30634928e+00
-2.60258853e-01 -1.20216227e+00 1.13098109e+00 3.72519135e-01
-1.55141219e-01 -9.49217260e-01 3.08108032e-01 -4.98180278e-02
3.96489620e-01 1.04151690e+00 1.08815408e+00 -5.16819715e-01
-4.10442531e-01 -1.27622098e-01 4.01043892e-02 2.69859612e-01
-3.79886329e-01 4.23704207e-01 -4.95925546e-01 -4.11333680e-01
-4.58316579e-02 1.77213792e-02 7.43948579e-01 9.33453441e-01
1.13851917e+00 -3.13154817e-01 -3.68558288e-01 5.30782104e-01
1.72976601e+00 3.39212775e-01 4.17561531e-01 4.49209243e-01
-5.88392206e-02 7.93590188e-01 8.17535877e-01 9.04283524e-01
3.00121963e-01 7.42879450e-01 8.44797969e-01 4.55459505e-01
4.54157770e-01 1.96139246e-01 6.13388181e-01 3.74967813e-01
-2.31447108e-02 -6.25124633e-01 -1.13030648e+00 4.51286793e-01
-1.85683346e+00 -6.46068811e-01 -1.21147186e-01 2.60805511e+00
5.30555487e-01 2.63798296e-01 5.53618371e-01 2.03994364e-01
7.62850821e-01 -3.00956845e-01 -5.90825260e-01 -6.19936466e-01
-3.00515294e-01 1.72943324e-01 7.15149820e-01 6.85477078e-01
-8.69026721e-01 -4.02919725e-02 6.20769405e+00 1.09195137e+00
-7.01941848e-01 5.67782857e-02 6.30377293e-01 -2.78229177e-01
-4.78578687e-01 -4.68153208e-02 -8.55891943e-01 9.02497470e-01
9.46507215e-01 -7.08103061e-01 4.87102151e-01 8.02251637e-01
1.21872842e-01 -3.52900833e-01 -1.10484767e+00 7.56120563e-01
-2.80053318e-01 -1.25370812e+00 -4.24883306e-01 4.21116531e-01
8.68043542e-01 -2.13172168e-01 3.14653635e-01 -3.91527265e-02
3.72105509e-01 -1.12256670e+00 7.95152128e-01 8.32361341e-01
4.51273084e-01 -1.20318627e+00 9.02014792e-01 1.85562804e-01
-1.06164241e+00 -4.05674458e-01 -1.20468915e-01 5.23170352e-01
2.87795633e-01 9.25906360e-01 -5.39088607e-01 8.99023116e-01
8.20664048e-01 9.32561308e-02 -3.53313506e-01 1.74100745e+00
2.26105258e-01 3.35101962e-01 -7.07461715e-01 -4.81025815e-01
2.62372047e-01 -4.48546618e-01 1.22573042e+00 1.03092551e+00
7.49377549e-01 -4.51698363e-01 -3.03002968e-02 1.24095523e+00
2.91818768e-01 6.91632628e-02 -5.78003526e-02 -1.76241934e-01
5.72495639e-01 1.00808406e+00 -7.61468410e-01 1.44838661e-01
2.76176423e-01 4.44473773e-02 -2.10345015e-01 2.57562876e-01
-9.98393238e-01 -5.57729959e-01 6.55231059e-01 -2.98356950e-01
4.41118360e-01 -1.62959844e-01 -6.88462377e-01 -4.98642743e-01
1.75167233e-01 -6.88379288e-01 6.02886200e-01 -3.75584036e-01
-1.34097111e+00 3.58287811e-01 4.50789332e-01 -1.21765375e+00
-2.00214565e-01 -7.63395548e-01 -6.28691316e-01 8.11637521e-01
-1.40676188e+00 -2.74674892e-01 -2.67800003e-01 1.80585697e-01
3.39320898e-02 -1.31898373e-02 5.07259250e-01 2.34227069e-02
-5.20607352e-01 4.89668429e-01 6.88731670e-01 -7.73935080e-01
4.04066980e-01 -1.43964648e+00 -3.41950536e-01 4.95086044e-01
-3.24725688e-01 1.88829809e-01 1.21338701e+00 -5.08498490e-01
-1.48540843e+00 -7.02178776e-01 4.11012381e-01 -4.75113422e-01
5.97736537e-01 -1.39220178e-01 -4.77300704e-01 -2.34657466e-01
-1.25877783e-02 -4.13112760e-01 6.12100899e-01 -3.59173715e-02
1.95963204e-01 -2.87914932e-01 -1.42242098e+00 4.79027361e-01
6.22499049e-01 4.19989794e-01 -3.68037105e-01 2.29714990e-01
5.59898138e-01 -3.38940546e-02 -1.22180676e+00 7.46423244e-01
6.90040708e-01 -1.12560332e+00 1.09805048e+00 -3.28177154e-01
4.99487370e-01 -3.17696154e-01 -5.35978138e-01 -1.39033186e+00
-1.86585352e-01 -9.65088308e-01 -4.27619725e-01 1.19805002e+00
4.35091466e-01 -7.08671391e-01 6.18670523e-01 2.80377388e-01
9.21296328e-02 -1.42808890e+00 -1.32045937e+00 -1.48187172e+00
1.73690557e-01 -2.90041625e-01 1.11853683e+00 2.90950388e-01
-3.02913249e-01 -9.47801247e-02 1.30677698e-02 3.74918804e-02
1.14619398e+00 2.17074692e-01 5.77951849e-01 -1.57013059e+00
-4.62244689e-01 -1.02492976e+00 -1.63980216e-01 -4.27815109e-01
-2.75487095e-01 -7.66963303e-01 1.02875933e-01 -1.48395443e+00
3.91369238e-02 -2.31622487e-01 -4.37184721e-01 1.76682509e-02
-1.39695078e-01 -1.91772670e-01 2.42166594e-01 -1.16087988e-01
-7.04689503e-01 8.77837539e-01 1.12294257e+00 1.44828465e-02
-5.22752464e-01 9.70036760e-02 -6.65533066e-01 8.50214183e-01
6.98816180e-01 -7.55133510e-01 -6.34220466e-02 -1.83184315e-02
5.52966356e-01 1.11557379e-01 2.70767957e-01 -8.67505074e-01
6.12959787e-02 -5.28498054e-01 2.06447184e-01 -9.85681951e-01
3.06917578e-01 -8.10965419e-01 3.08778584e-01 5.00709295e-01
-3.12543698e-02 2.65462995e-01 1.70356572e-01 5.03077865e-01
-7.60473609e-02 -4.27447826e-01 7.07300901e-01 3.08103859e-01
-1.27769291e-01 -3.36590549e-03 -7.92209432e-02 2.56656766e-01
1.20421672e+00 -2.90139288e-01 -5.28191626e-01 -2.15330616e-01
-3.76669526e-01 7.19574690e-01 2.48992234e-01 1.57827929e-01
5.48005462e-01 -1.33760321e+00 -1.00933290e+00 -2.87093401e-01
-1.00286700e-01 -2.60766029e-01 -1.46280318e-01 1.04514194e+00
-3.55225682e-01 3.71852040e-01 -1.28837883e-01 -8.30057204e-01
-1.07501006e+00 2.63037443e-01 2.24912986e-01 -8.63777876e-01
1.55353218e-01 7.39253342e-01 -8.02490264e-02 -2.88429379e-01
3.80318493e-01 -3.58971238e-01 -1.66664928e-01 4.69888896e-01
2.17978448e-01 1.29625535e+00 -6.00374117e-02 -1.34221852e-01
-7.07955718e-01 5.10789394e-01 8.31913471e-01 -4.32295561e-01
1.60699463e+00 3.30956951e-02 -3.09147507e-01 2.94788092e-01
1.14809346e+00 -3.66495084e-03 -1.23366475e+00 3.26017439e-01
4.31391746e-01 -5.61880469e-01 6.90588653e-02 -1.14123464e+00
-6.60776556e-01 5.04449904e-01 4.69964802e-01 5.40848136e-01
1.18549037e+00 -2.35851556e-01 2.93376625e-01 8.75971243e-02
3.18877965e-01 -1.10941434e+00 1.82072327e-01 3.83678377e-01
1.22025728e+00 -8.24275196e-01 2.29896963e-01 -1.43667638e-01
-5.02724648e-01 1.08702683e+00 2.61041343e-01 -1.96741670e-02
7.52050698e-01 5.08015871e-01 -5.48733950e-01 -2.59671770e-02
-7.42126465e-01 -8.12531859e-02 7.64754295e-01 3.84802938e-01
-1.46431401e-01 1.76533520e-01 -6.94248140e-01 1.02187645e+00
-3.62732589e-01 -3.50706339e-01 2.60525405e-01 8.13240647e-01
-5.02516866e-01 -1.04470992e+00 -9.63815987e-01 4.87809956e-01
-5.46724916e-01 1.97142035e-01 -2.71196365e-01 6.78340197e-01
9.23389792e-02 1.03991771e+00 8.08943994e-03 -1.78315639e-01
2.44528487e-01 -1.15257353e-01 3.88567746e-01 -2.25160532e-02
-6.11548543e-01 1.51746750e-01 1.04521737e-01 -7.30165124e-01
-1.17955729e-01 -7.75588274e-01 -7.36248732e-01 -2.80242622e-01
-4.81350720e-01 3.41662109e-01 9.52888012e-01 8.37965250e-01
3.58841419e-01 5.55245161e-01 7.21942663e-01 -8.35519016e-01
-9.45798695e-01 -5.72989821e-01 -5.24625063e-01 -1.39712453e-01
2.40626246e-01 -1.01990354e+00 -5.41761041e-01 -7.14099705e-01] | [6.05806303024292, 3.6391899585723877] |
5cda383a-d10a-4a2b-a92a-8f7c465f1b33 | can-we-use-diffusion-probabilistic-models-for | 2302.14503 | null | https://arxiv.org/abs/2302.14503v1 | https://arxiv.org/pdf/2302.14503v1.pdf | Can We Use Diffusion Probabilistic Models for 3D Motion Prediction? | After many researchers observed fruitfulness from the recent diffusion probabilistic model, its effectiveness in image generation is actively studied these days. In this paper, our objective is to evaluate the potential of diffusion probabilistic models for 3D human motion-related tasks. To this end, this paper presents a study of employing diffusion probabilistic models to predict future 3D human motion(s) from the previously observed motion. Based on the Human 3.6M and HumanEva-I datasets, our results show that diffusion probabilistic models are competitive for both single (deterministic) and multiple (stochastic) 3D motion prediction tasks, after finishing a single training process. In addition, we find out that diffusion probabilistic models can offer an attractive compromise, since they can strike the right balance between the likelihood and diversity of the predicted future motions. Our code is publicly available on the project website: https://sites.google.com/view/diffusion-motion-prediction. | ['Dongheui Lee', 'Esteve Valls Mascaro', 'Hyemin Ahn'] | 2023-02-28 | null | null | null | null | ['motion-prediction'] | ['computer-vision'] | [-1.38036877e-01 1.11895382e-01 -3.97371858e-01 -5.21943755e-02
-4.88413483e-01 -4.20813918e-01 1.12040043e+00 -2.98666269e-01
-3.46325547e-01 3.96467179e-01 6.31139457e-01 -2.77964592e-01
-7.09158108e-02 -6.65569723e-01 -3.62672418e-01 -6.02212608e-01
-1.41171172e-01 4.91565138e-01 5.22163808e-01 -5.66599816e-02
3.94638509e-01 4.92100507e-01 -1.43168557e+00 1.87762186e-01
5.08369446e-01 4.00984257e-01 3.81570309e-01 1.03033960e+00
3.99945378e-02 1.01961231e+00 -3.58869165e-01 -3.93178344e-01
1.07990555e-01 -3.55822831e-01 -8.58732224e-01 1.07561098e-02
-3.05651482e-02 -2.15819076e-01 -6.62643135e-01 6.62816346e-01
5.99889696e-01 4.15250987e-01 8.95826161e-01 -1.52305806e+00
-6.67359173e-01 5.49190938e-01 -5.68703711e-01 3.45507711e-01
7.08962440e-01 4.55177575e-01 8.79840195e-01 -8.61064434e-01
1.02616405e+00 1.33005381e+00 6.26863122e-01 8.53575408e-01
-9.02938604e-01 -2.13324860e-01 -7.07523748e-02 2.91258901e-01
-1.44100034e+00 -4.04893637e-01 7.53655910e-01 -8.88133228e-01
1.04271984e+00 1.23276822e-01 6.67692900e-01 1.57077599e+00
5.79009295e-01 1.22019517e+00 8.57666016e-01 -2.53410459e-01
1.92498133e-01 -3.36445458e-02 -1.13757201e-01 4.94087547e-01
6.46435143e-03 4.14992958e-01 -7.23517716e-01 -3.12128276e-01
6.70805275e-01 -2.10472271e-01 -2.13752508e-01 -4.82459337e-01
-1.35559845e+00 8.20266247e-01 1.69643238e-01 3.40537935e-01
-6.18868053e-01 4.71093863e-01 -1.44026160e-01 -3.53004426e-01
5.96072853e-01 -3.90348621e-02 -1.63970158e-01 -5.88717759e-01
-1.09387016e+00 8.13206375e-01 6.47775710e-01 8.17416728e-01
3.78362030e-01 -1.48230076e-01 -4.18209910e-01 4.48424429e-01
7.13272572e-01 6.58620477e-01 5.78623891e-01 -1.35720098e+00
1.09557457e-01 8.80973563e-02 4.78781372e-01 -1.12321889e+00
-2.06912220e-01 3.78793664e-02 -7.63074458e-01 2.53070951e-01
4.84066814e-01 -3.42440456e-01 -7.80969143e-01 1.40607929e+00
4.47085261e-01 1.87787965e-01 -4.61594425e-02 8.36817443e-01
4.00625944e-01 9.91082370e-01 2.48411477e-01 3.59234512e-02
9.08285916e-01 -1.12910378e+00 -5.10886431e-01 -2.52695680e-01
6.36131704e-01 -9.34766471e-01 7.14558125e-01 1.85725763e-01
-1.09335315e+00 -6.92041576e-01 -5.92780530e-01 1.05995081e-01
-1.05708959e-02 -7.16651455e-02 5.92880666e-01 7.26593971e-01
-1.37827146e+00 7.24775970e-01 -1.01537001e+00 -5.23381293e-01
3.61224174e-01 -3.31140868e-02 -1.31371230e-01 -1.43857762e-01
-9.58048642e-01 8.88891935e-01 1.24189116e-01 -6.06564945e-03
-8.83122444e-01 -2.68999189e-01 -6.24303997e-01 -4.65511709e-01
1.79329202e-01 -1.27088058e+00 1.42237413e+00 -5.75513184e-01
-1.40356123e+00 7.81016350e-01 -4.16975826e-01 -6.95906639e-01
1.00783026e+00 -4.29026216e-01 4.63847071e-02 -6.08156063e-03
2.67927676e-01 1.29693758e+00 8.20695102e-01 -1.17544019e+00
-7.60755002e-01 -1.34073809e-01 -2.71072239e-01 2.80279547e-01
4.70340289e-02 -9.12529826e-02 -5.53829312e-01 -8.67973864e-01
-2.40401030e-01 -1.31889522e+00 -6.49848044e-01 -1.48293778e-01
-3.16921562e-01 -4.76196885e-01 6.08998120e-01 -4.72951889e-01
1.05776703e+00 -1.83443534e+00 3.30705613e-01 6.44298866e-02
1.27445430e-01 -3.98480073e-02 8.68862867e-03 4.11808878e-01
2.50300884e-01 3.62623066e-01 -2.49836609e-01 -4.90455478e-01
-2.47928202e-02 -3.21954377e-02 -1.69978395e-01 3.67281109e-01
8.91948417e-02 1.16719484e+00 -8.99477959e-01 -5.66908419e-01
1.66462705e-01 6.33506298e-01 -4.06156451e-01 8.02581385e-02
-2.51689970e-01 6.89943254e-01 -6.23822629e-01 3.51250887e-01
3.57176870e-01 -3.20238143e-01 5.38982041e-02 3.05310845e-01
-1.44670466e-02 -2.48492733e-01 -9.77720201e-01 1.70871294e+00
7.54707307e-03 8.22043478e-01 -6.09615266e-01 -3.21604103e-01
7.80958831e-01 1.82622150e-01 8.45824718e-01 -2.64927357e-01
-1.24008590e-02 2.83063278e-02 -1.81893244e-01 -4.71447021e-01
9.35083687e-01 -1.70240879e-01 1.46172447e-02 6.50146425e-01
-1.93338320e-01 -2.09056854e-01 8.19382966e-02 3.05710465e-01
1.10114622e+00 6.23504579e-01 2.10323870e-01 -1.04187749e-01
2.89309651e-01 4.76328939e-01 2.68336296e-01 9.30206716e-01
-5.33091187e-01 8.36391628e-01 1.59838736e-01 -4.23399448e-01
-1.23826540e+00 -1.14097703e+00 4.00531441e-01 5.92793822e-01
2.40979150e-01 -3.96291256e-01 -6.30450666e-01 -7.06262827e-01
-2.13274375e-01 9.58302975e-01 -6.77246332e-01 -1.86063331e-02
-5.30310631e-01 -7.08309233e-01 4.03542787e-01 5.37108660e-01
2.07480654e-01 -1.12235212e+00 -8.40825140e-01 1.26097724e-01
-4.27093625e-01 -9.66527462e-01 -4.83117223e-01 -6.08380973e-01
-9.45666969e-01 -7.18393028e-01 -1.51705170e+00 -2.84025669e-01
1.46637961e-01 5.88471234e-01 9.84401166e-01 -5.76421320e-02
-1.43793628e-01 8.65334809e-01 -5.08238375e-01 -2.89210975e-01
-5.87347865e-01 2.08652038e-02 1.34102657e-01 -1.55373305e-01
2.50829190e-01 -3.45071673e-01 -8.78919661e-01 4.19038773e-01
-7.89654016e-01 1.48786336e-01 3.43029499e-01 3.67839724e-01
4.31008607e-01 1.29710451e-01 1.08331047e-01 -5.38985789e-01
6.57886267e-01 -7.87572861e-01 -2.38145426e-01 -8.53455886e-02
-5.92406273e-01 -4.55166996e-02 -3.11993867e-01 -5.13262630e-01
-1.26098156e+00 2.57596225e-01 -2.04253599e-01 -4.93561357e-01
-3.44571322e-01 3.44082505e-01 3.58533710e-01 3.08680892e-01
8.15496624e-01 3.18249792e-01 -1.08333848e-01 -1.99054658e-01
7.46333599e-01 2.62394071e-01 1.96640804e-01 -3.25801700e-01
6.25052571e-01 8.73731554e-01 -1.50870517e-01 -8.78469944e-01
-3.94055963e-01 -5.48904419e-01 -7.32054532e-01 -6.36753380e-01
1.21278739e+00 -8.62247884e-01 -2.53250867e-01 7.13887990e-01
-1.38823223e+00 -7.34751403e-01 -2.26596713e-01 5.29344559e-01
-8.86616409e-01 6.32598341e-01 -5.54866970e-01 -1.19284022e+00
-3.81282270e-02 -1.07649815e+00 1.05732417e+00 2.01785043e-01
-7.99619198e-01 -1.22088325e+00 3.05770755e-01 4.39272434e-01
2.80378610e-01 9.59215388e-02 1.95934907e-01 -2.13797152e-01
-8.25147390e-01 -2.41435215e-01 1.14071704e-01 3.20834666e-02
-1.81583270e-01 3.57184500e-01 -7.37380028e-01 6.10033870e-02
-7.97110870e-02 1.19979598e-01 1.02271843e+00 8.39557409e-01
4.83307451e-01 -9.28042755e-02 -7.71866977e-01 1.72278866e-01
9.43455517e-01 1.22477792e-01 7.88388431e-01 4.79318947e-01
8.18659306e-01 8.82318556e-01 8.44452441e-01 5.99222600e-01
7.31856644e-01 7.23112106e-01 8.51723924e-02 3.49624306e-01
-4.06447321e-01 -5.34352958e-01 5.17456114e-01 5.12436211e-01
-3.70796889e-01 -7.85480499e-01 -1.44732523e+00 7.43705034e-01
-2.03447008e+00 -1.15608299e+00 -5.38093507e-01 1.85280812e+00
3.29454362e-01 2.33189508e-01 4.63406652e-01 -1.87420249e-01
7.33748376e-01 3.52326542e-01 -2.26062492e-01 -9.03696045e-02
-2.43454382e-01 -3.12522858e-01 4.48715210e-01 6.27825260e-01
-1.04388559e+00 1.14706850e+00 6.64365721e+00 9.91090298e-01
-6.45654023e-01 1.69606373e-01 7.91184187e-01 -5.15754968e-02
-5.08627295e-01 1.68969437e-01 -9.66342449e-01 4.03291017e-01
9.18275595e-01 -3.29889208e-01 -1.34412915e-01 7.59798527e-01
5.31243980e-01 -3.62603188e-01 -8.18399727e-01 9.54272330e-01
-1.03864737e-01 -1.55088937e+00 2.86160439e-01 4.31985617e-01
9.69748497e-01 2.43218943e-01 2.77977020e-01 -1.10853286e-02
4.60629344e-01 -9.59921181e-01 1.09909046e+00 9.80212450e-01
1.10602975e-01 -4.71185923e-01 6.71593487e-01 7.45614529e-01
-8.40007365e-01 8.87452587e-02 -9.75872502e-02 4.57979552e-02
8.96619976e-01 6.52139664e-01 -7.45284259e-01 4.79544878e-01
7.36265361e-01 8.36739063e-01 -5.00161946e-01 1.15588510e+00
-2.08190799e-01 5.60808837e-01 -2.33260334e-01 -1.48346052e-01
2.98848957e-01 7.98194185e-02 9.53431129e-01 1.19644773e+00
8.08640540e-01 5.17172106e-02 2.87602004e-02 7.98356712e-01
4.40808564e-01 9.85953435e-02 -8.88380349e-01 -1.57945976e-02
6.64511994e-02 7.29945600e-01 -8.65569890e-01 -2.50034779e-01
3.25721689e-02 1.25735652e+00 -9.50097665e-02 3.78239632e-01
-8.68360579e-01 3.09869468e-01 3.82105738e-01 1.74859986e-01
2.77750373e-01 -7.38114178e-01 -2.97914058e-01 -9.36365783e-01
-2.31082618e-01 -3.81074578e-01 2.05842376e-01 -1.10862446e+00
-1.19370091e+00 6.27806783e-01 2.69887239e-01 -1.08956301e+00
-6.03538692e-01 -3.73762012e-01 -5.40838718e-01 6.99898541e-01
-1.05256832e+00 -1.12973189e+00 1.00420713e-02 2.77815789e-01
7.94872522e-01 -6.16675243e-03 3.67547482e-01 -5.05559891e-02
-1.21743113e-01 -1.05672358e-02 -2.55512387e-01 -3.67168814e-01
5.04918993e-01 -1.01291227e+00 9.19503689e-01 7.70726204e-01
5.16538322e-01 1.55283809e-01 9.98100579e-01 -1.07993019e+00
-1.01854074e+00 -8.82252455e-01 1.27065468e+00 -1.05700302e+00
6.90709651e-01 1.75128028e-01 -6.82294667e-01 4.26874667e-01
1.73971489e-01 -4.50924873e-01 5.25653839e-01 -2.84909934e-01
1.64492920e-01 6.51879013e-01 -7.34044850e-01 9.83267248e-01
1.25806046e+00 -1.60638854e-01 -3.45134974e-01 3.61970097e-01
5.33086956e-01 -1.93352520e-01 -6.95632517e-01 3.52877289e-01
6.41867220e-01 -1.22367811e+00 1.14086246e+00 -2.27789402e-01
6.28419399e-01 -1.82764754e-01 -1.47549823e-01 -1.05030954e+00
-4.36026573e-01 -7.89396286e-01 -3.71821314e-01 1.00705326e+00
5.59452832e-01 -1.52926102e-01 1.20009243e+00 6.40304744e-01
3.27316314e-01 -6.17291451e-01 -7.95304537e-01 -7.68377781e-01
2.31685221e-01 -9.13863957e-01 1.21457838e-01 6.32040501e-01
-3.56410176e-01 2.22094297e-01 -7.43173063e-01 3.16286385e-02
6.04581892e-01 -9.82433185e-02 1.10214090e+00 -9.22389507e-01
-3.47947687e-01 -6.96955860e-01 -4.32914346e-01 -1.54661131e+00
4.15505022e-02 -6.06462538e-01 1.98660359e-01 -1.73648369e+00
1.95905969e-01 -2.15754643e-01 1.24844462e-01 3.40361036e-02
-3.41107219e-01 4.97334935e-02 4.81818527e-01 6.81374073e-01
-5.61673105e-01 6.97483480e-01 1.29931045e+00 -3.02962959e-02
-3.64305884e-01 3.74885410e-01 -3.34966421e-01 8.40303421e-01
1.41867113e+00 -6.04587018e-01 -4.32025433e-01 -4.95934725e-01
9.67089757e-02 2.40614966e-01 4.48518753e-01 -1.03494346e+00
4.42549080e-01 -3.60333771e-01 2.29891613e-01 -8.38126123e-01
6.38670027e-01 -2.41358876e-01 2.73815989e-01 4.93137687e-01
-2.72607625e-01 3.03617209e-01 2.03480590e-02 1.05916870e+00
7.47049376e-02 -1.68563709e-01 3.35765839e-01 -3.01213533e-01
-1.12978685e+00 3.84825349e-01 -9.19691265e-01 -1.30939052e-01
1.15007627e+00 -3.81519824e-01 3.11355237e-02 -8.33847046e-01
-9.37179387e-01 4.54951227e-02 6.34129703e-01 7.98955381e-01
8.85445058e-01 -1.22541571e+00 -8.99548352e-01 -4.44145441e-01
1.75078714e-03 -2.93520182e-01 6.67464674e-01 8.00161421e-01
-5.60362995e-01 3.99915308e-01 -9.56242383e-02 -8.14208508e-01
-1.13623345e+00 5.86576700e-01 2.48747170e-02 -1.72090426e-01
-5.49883366e-01 1.07173121e+00 5.63869923e-02 -1.70286283e-01
-1.14503540e-01 9.11053792e-02 -1.84952497e-01 -2.83087105e-01
3.71581554e-01 7.25959480e-01 -8.07761908e-01 -1.02170336e+00
-3.71285588e-01 5.34220695e-01 2.09504664e-01 -8.39156926e-01
9.95985508e-01 -5.25958598e-01 2.70304888e-01 6.69994116e-01
7.42822349e-01 9.64423716e-02 -1.46647358e+00 1.64674908e-01
4.20000404e-01 -4.48428124e-01 -1.35643184e-01 -4.80694622e-01
-6.55789316e-01 8.84202421e-01 5.22622824e-01 1.83514357e-01
6.36642635e-01 2.93806136e-01 8.69599760e-01 -1.84626263e-02
4.08480644e-01 -7.45646238e-01 1.10044621e-01 3.94416720e-01
7.14118242e-01 -1.13382292e+00 -4.99648601e-03 -4.54416603e-01
-1.19667232e+00 7.85408437e-01 3.13597649e-01 -6.35430366e-02
7.31400549e-01 -5.69642372e-02 6.68962821e-02 -1.44309238e-01
-9.58234310e-01 -2.83121347e-01 4.64108795e-01 9.00146961e-01
5.11884987e-01 1.97997689e-01 -1.28352180e-01 2.45726109e-01
-1.49149835e-01 2.26718903e-01 6.41136706e-01 8.87109220e-01
-4.54027891e-01 -1.22552347e+00 -4.01630193e-01 -1.24058329e-01
-3.54286671e-01 -3.16495565e-03 -4.99237061e-01 5.72153270e-01
2.65973601e-02 1.02727878e+00 -1.95901692e-01 -4.30393457e-01
4.57982011e-02 -5.06046861e-02 5.00505924e-01 -4.02057171e-01
-5.13839871e-02 1.05074048e-01 6.46660030e-02 -6.11080408e-01
-7.33569920e-01 -1.03306663e+00 -9.64469492e-01 -4.52872515e-01
1.02824688e-01 -6.02293760e-04 5.99361479e-01 5.75965941e-01
5.04555762e-01 6.73399195e-02 2.26003170e-01 -1.04430866e+00
-3.38396996e-01 -8.09114754e-01 -4.03613299e-01 2.82847852e-01
-1.21666528e-01 -5.92613041e-01 -1.35900199e-01 4.76334721e-01] | [7.3195600509643555, -0.10635749995708466] |
4945bb43-7f8e-45fe-a53a-c8a1bf63bfef | label-efficient-learning-in-agriculture-a | 2305.14691 | null | https://arxiv.org/abs/2305.14691v1 | https://arxiv.org/pdf/2305.14691v1.pdf | Label-Efficient Learning in Agriculture: A Comprehensive Review | The past decade has witnessed many great successes of machine learning (ML) and deep learning (DL) applications in agricultural systems, including weed control, plant disease diagnosis, agricultural robotics, and precision livestock management. Despite tremendous progresses, one downside of such ML/DL models is that they generally rely on large-scale labeled datasets for training, and the performance of such models is strongly influenced by the size and quality of available labeled data samples. In addition, collecting, processing, and labeling such large-scale datasets is extremely costly and time-consuming, partially due to the rising cost in human labor. Therefore, developing label-efficient ML/DL methods for agricultural applications has received significant interests among researchers and practitioners. In fact, there are more than 50 papers on developing and applying deep-learning-based label-efficient techniques to address various agricultural problems since 2016, which motivates the authors to provide a timely and comprehensive review of recent label-efficient ML/DL methods in agricultural applications. To this end, we first develop a principled taxonomy to organize these methods according to the degree of supervision, including weak supervision (i.e., active learning and semi-/weakly- supervised learning), and no supervision (i.e., un-/self- supervised learning), supplemented by representative state-of-the-art label-efficient ML/DL methods. In addition, a systematic review of various agricultural applications exploiting these label-efficient algorithms, such as precision agriculture, plant phenotyping, and postharvest quality assessment, is presented. Finally, we discuss the current problems and challenges, as well as future research directions. A well-classified paper list can be accessed at https://github.com/DongChen06/Label-efficient-in-Agriculture. | ['Xiaobo Tan', 'Daniel Morris', 'Yanbo Huang', 'Zhaojian Li', 'Xinda Qi', 'Dong Chen', 'Jiajia Li'] | 2023-05-24 | null | null | null | null | ['plant-phenotyping', 'active-learning', 'active-learning'] | ['computer-vision', 'methodology', 'natural-language-processing'] | [ 3.89117450e-01 -7.52391368e-02 -9.99765515e-01 -3.17866325e-01
-1.52961269e-01 -8.03896725e-01 -3.39930877e-02 7.60394573e-01
-1.47951189e-02 6.64223611e-01 -5.96747696e-01 -5.51276743e-01
-1.19868226e-01 -1.12631345e+00 -6.64324224e-01 -9.27109838e-01
-1.84171692e-01 3.32986355e-01 -1.69568822e-01 -8.88419449e-02
-2.98803061e-01 6.90674067e-01 -1.79042828e+00 1.18680470e-01
1.11552823e+00 1.01825643e+00 5.63447237e-01 3.47805917e-01
-2.66118735e-01 4.88555908e-01 -4.18968201e-01 -1.07620157e-01
2.97096539e-02 -4.77295846e-01 -5.59318542e-01 3.05918723e-01
-1.92436799e-01 -3.83027971e-01 3.69690150e-01 1.16852832e+00
6.11885488e-01 -2.14674219e-01 5.49741209e-01 -1.38661122e+00
-6.86839819e-01 6.07148230e-01 -8.37963462e-01 -4.13198024e-01
-1.41363814e-01 -4.92710844e-02 8.19746554e-01 -8.14064503e-01
4.78718936e-01 8.64761412e-01 5.89458525e-01 2.38114655e-01
-1.23304462e+00 -9.24976587e-01 3.67579281e-01 6.81161433e-02
-1.03020430e+00 -3.96055758e-01 5.50928652e-01 -4.94965851e-01
5.15120804e-01 -4.81237881e-02 9.96625483e-01 7.81924546e-01
2.45861351e-01 1.33001697e+00 9.39879119e-01 -5.99015296e-01
3.89792770e-01 -1.78551510e-01 1.62778720e-01 7.83282340e-01
5.47376156e-01 4.29340541e-01 -3.32377374e-01 -1.03828952e-01
5.63992739e-01 2.39161521e-01 -3.16554271e-02 -7.25149274e-01
-9.38612998e-01 9.87838745e-01 5.85361779e-01 8.42599571e-02
-4.93237257e-01 -3.94704700e-01 5.85146964e-01 2.96390146e-01
8.17784011e-01 6.92581534e-01 -1.09663320e+00 5.49439847e-01
-8.59981894e-01 3.06831133e-02 6.80275023e-01 1.25460732e+00
1.04905450e+00 1.38164967e-01 6.79231286e-02 9.15600181e-01
3.60878825e-01 8.43086004e-01 -1.18954293e-01 -9.97502923e-01
-2.59465158e-01 7.39724398e-01 1.81015909e-01 -9.69338536e-01
-5.66660464e-01 -3.44972998e-01 -1.19078958e+00 2.59635478e-01
1.50146499e-01 -2.59019047e-01 -9.75843310e-01 1.56335723e+00
4.22014713e-01 -1.81449279e-01 2.93389591e-03 7.11958647e-01
1.03488564e+00 7.07233131e-01 4.72945005e-01 -8.01747262e-01
1.18782115e+00 -1.09732234e+00 -1.13193953e+00 -2.85257459e-01
9.63483274e-01 -7.48531997e-01 7.10851490e-01 4.34389383e-01
-6.61766827e-01 -5.59189677e-01 -1.00862539e+00 1.03694489e-02
-8.29181433e-01 6.71078861e-01 1.15006483e+00 4.37497854e-01
-5.79612792e-01 5.59733868e-01 -1.04719102e+00 -5.67645490e-01
6.65316880e-01 2.16676697e-01 -2.73197174e-01 -3.58246267e-01
-1.10040987e+00 9.24106359e-01 5.69482267e-01 4.81490403e-01
-1.17065787e+00 -6.68687999e-01 -9.29653108e-01 -2.46870056e-01
6.25566423e-01 -1.15403064e-01 1.23239696e+00 -8.84231508e-01
-1.68049634e+00 1.32091618e+00 -1.68553084e-01 -1.34282619e-01
-4.66229692e-02 -5.06852388e-01 -3.19508851e-01 -1.11102484e-01
2.55788773e-01 7.74637222e-01 4.23978478e-01 -1.41004896e+00
-9.40588474e-01 -5.80651999e-01 8.59292448e-02 6.98836939e-03
-3.37581247e-01 1.94348432e-02 1.46613777e-01 -4.93806034e-01
2.71113366e-01 -8.82694900e-01 -3.32485557e-01 5.60513139e-01
-1.26854017e-01 -1.53962478e-01 9.53960061e-01 -4.74991083e-01
9.69054759e-01 -2.16452003e+00 -3.66611034e-02 -2.41389975e-01
2.10245013e-01 7.73223221e-01 -4.38115805e-01 3.53908986e-01
-4.77253199e-02 3.99035066e-02 -4.93847609e-01 2.45860606e-01
-2.84496546e-01 3.31361175e-01 -1.41867802e-01 5.48580170e-01
6.44881606e-01 1.05837917e+00 -1.58543348e+00 -3.25068295e-01
5.58099210e-01 2.26893555e-02 1.70757800e-01 4.82174814e-01
-5.31183124e-01 5.13396978e-01 -4.33929026e-01 1.35653496e+00
9.31799769e-01 -2.32853234e-01 2.67823160e-01 -3.02873760e-01
-4.60357338e-01 -2.56680071e-01 -6.86188996e-01 1.52916276e+00
-3.14247012e-01 3.30345273e-01 4.57799792e-01 -1.50095546e+00
1.15142643e+00 4.55189049e-01 6.50369585e-01 -5.28232634e-01
1.11179098e-01 6.38751328e-01 -2.70509034e-01 -5.15999317e-01
5.15527688e-02 2.63050497e-01 2.21396819e-01 1.26573727e-01
2.35050619e-01 -1.42419055e-01 4.83429193e-01 -2.71599561e-01
6.76597655e-01 6.23789370e-01 6.85040653e-01 -4.51170772e-01
2.79105932e-01 4.69484597e-01 9.38653469e-01 5.14279604e-01
-5.74460566e-01 1.30160436e-01 3.92821521e-01 -4.29633081e-01
-7.19183624e-01 -5.55426478e-01 -2.38706574e-01 1.58522213e+00
3.47999245e-01 -5.92414215e-02 -3.92508626e-01 -6.62729263e-01
2.26772144e-01 4.14634943e-01 -4.82120961e-01 -2.27536380e-01
-3.50284725e-01 -1.31405711e+00 4.61574435e-01 5.04375339e-01
6.36304557e-01 -1.27083206e+00 -5.07379115e-01 4.63102072e-01
-7.64852241e-02 -9.00188744e-01 4.74088222e-01 1.04140127e+00
-1.04628265e+00 -1.22369659e+00 -8.37703466e-01 -1.15439963e+00
5.96869767e-01 5.50612688e-01 1.32075202e+00 -1.16238356e-01
-3.00814748e-01 -3.47263962e-01 -8.88215959e-01 -1.06099403e+00
-3.35570782e-01 2.46156335e-01 -1.40994892e-01 -5.96683621e-01
6.38589919e-01 -2.29823478e-02 -4.26028520e-01 3.27614427e-01
-7.69888997e-01 -1.96084362e-02 9.34814036e-01 1.03943074e+00
9.51382577e-01 1.91128030e-01 7.83858776e-01 -1.09321749e+00
1.32052973e-03 -5.31312466e-01 -8.64492536e-01 7.67929912e-01
-7.06429780e-01 -4.84441400e-01 5.03953278e-01 -4.27691042e-01
-8.20263863e-01 5.21455050e-01 7.81914045e-04 8.91284496e-02
-4.37786490e-01 9.57847953e-01 -5.10424912e-01 -1.13777600e-01
6.92695796e-01 -2.90045500e-01 1.13637224e-01 -2.71769732e-01
4.18883443e-01 6.49332881e-01 1.41781375e-01 -4.11851913e-01
4.52997088e-01 7.32758045e-02 2.32365578e-01 -9.86590385e-01
-1.45638502e+00 -4.69263017e-01 -7.68875539e-01 -2.84716487e-01
6.61607862e-01 -9.55621660e-01 -3.68314058e-01 1.03671920e+00
-9.23305869e-01 -7.38569140e-01 -3.33351165e-01 6.30639136e-01
-2.86442757e-01 1.74720630e-01 -6.25772059e-01 -6.90023243e-01
-4.98774856e-01 -1.17021036e+00 1.10845423e+00 3.38990271e-01
-2.18853094e-02 -9.09237206e-01 -2.28450939e-01 1.04527213e-01
3.14938039e-01 6.17850065e-01 1.03883457e+00 -2.71517962e-01
-8.97252411e-02 -2.96160191e-01 -3.91509265e-01 4.65553284e-01
6.97256744e-01 2.95131654e-01 -1.17084372e+00 -3.75880569e-01
-3.16619068e-01 -8.78677845e-01 6.91868842e-01 8.72212768e-01
1.05699062e+00 2.78536290e-01 -5.56750774e-01 6.59054339e-01
1.14820695e+00 4.34260905e-01 1.33243307e-01 1.02286264e-01
5.75806022e-01 8.27048063e-01 1.37278271e+00 2.78148472e-01
-1.50479106e-02 8.62386897e-02 5.89835882e-01 -7.92536736e-01
1.34643048e-01 -5.67078479e-02 1.47253156e-01 7.53541827e-01
2.95778154e-03 -5.92225730e-01 -7.69944251e-01 4.41446662e-01
-1.96942461e+00 -5.25805354e-01 -2.47223020e-01 1.92984104e+00
1.05532157e+00 -2.35641435e-01 -1.59539163e-01 4.59235549e-01
8.45652878e-01 8.54507089e-02 -1.01411653e+00 -1.48834080e-01
-5.32376409e-01 2.21198514e-01 7.58945525e-01 1.53992370e-01
-1.68555522e+00 1.29475999e+00 6.17766237e+00 7.96482861e-01
-1.28088808e+00 -1.71703354e-01 5.98859847e-01 5.39600253e-01
3.01346302e-01 -2.55115479e-02 -6.63634002e-01 2.26703212e-02
4.53824461e-01 8.11629221e-02 1.40726700e-01 1.08958066e+00
1.60079286e-01 -3.22237462e-01 -1.05378985e+00 4.81068999e-01
-3.16423595e-01 -9.73821759e-01 -1.78071409e-01 -2.22856075e-01
8.86172116e-01 -7.15033412e-02 -2.88031906e-01 2.52889574e-01
6.25690758e-01 -6.62164688e-01 3.85434359e-01 -5.29816374e-02
9.50584888e-01 -5.86717308e-01 1.02001703e+00 4.23807502e-01
-1.36792779e+00 -1.62995741e-01 -7.16734469e-01 -1.07708544e-01
-4.54921164e-02 1.29741800e+00 -1.78595245e-01 8.63180578e-01
8.58777583e-01 1.25671768e+00 -3.03343356e-01 8.32448542e-01
-5.45421898e-01 7.56365240e-01 -1.53422803e-01 1.36597723e-01
1.19812265e-01 -3.36690366e-01 -5.39302342e-02 1.05580175e+00
2.36296594e-01 1.93784401e-01 6.92846715e-01 6.32346749e-01
1.99788269e-02 3.47865038e-02 -7.16849685e-01 -5.72417438e-01
6.65450931e-01 1.39944744e+00 -1.02178776e+00 -1.32873520e-01
-3.72099191e-01 5.65510511e-01 -1.08932806e-02 3.02054584e-01
-3.77948463e-01 -5.34292459e-01 3.51311326e-01 -2.36750364e-01
-2.01853782e-01 -2.18213558e-01 -2.63831019e-01 -8.63737643e-01
-3.65782350e-01 -9.46167350e-01 2.25592390e-01 -5.23894310e-01
-1.22020519e+00 2.87432987e-02 3.52049917e-02 -1.04620421e+00
2.64423072e-01 -8.42061460e-01 1.24702737e-01 7.19882727e-01
-1.88441098e+00 -1.29092646e+00 -7.23102629e-01 -1.58314109e-01
5.04601777e-01 -3.41822535e-01 1.37513280e+00 3.11023921e-01
-7.75306225e-01 2.17806935e-01 4.07098383e-01 1.40312761e-01
7.93596208e-01 -8.61930251e-01 1.69820070e-01 5.80078602e-01
-3.40293497e-01 3.12480092e-01 3.40109646e-01 -7.53526390e-01
-1.38589287e+00 -1.54576111e+00 8.51599872e-01 7.72764534e-02
4.79053050e-01 -3.13375652e-01 -7.88038552e-01 5.89299023e-01
-6.10263050e-02 2.43350267e-01 8.16408038e-01 2.05326732e-02
1.69802368e-01 -2.82111168e-01 -1.20593250e+00 1.26933277e-01
1.02260947e+00 -2.26378992e-01 1.77006543e-01 6.62237048e-01
6.61543667e-01 -3.64335209e-01 -9.60681796e-01 1.03202164e+00
5.35302281e-01 -4.60115194e-01 7.07389772e-01 -4.57210511e-01
5.17517269e-01 -2.46992365e-01 3.91422631e-03 -1.26976264e+00
-7.99297631e-01 -2.27556914e-01 -2.40926683e-01 1.21980488e+00
2.08369195e-01 -4.54286337e-01 6.77060127e-01 -1.60613120e-01
-3.78936172e-01 -8.97587359e-01 -3.58957760e-02 -7.63349354e-01
7.92512596e-02 1.12449318e-01 6.57087982e-01 1.28277290e+00
-2.91037351e-01 1.70052186e-01 -2.82041997e-01 2.57031232e-01
7.68880427e-01 3.85264844e-01 4.46689397e-01 -1.54620492e+00
3.69670361e-01 -1.97453633e-01 -1.45103157e-01 -9.66717005e-01
4.06198263e-01 -7.92776644e-01 3.79126281e-01 -1.74425435e+00
6.87030181e-02 -6.35288239e-01 -1.97763428e-01 9.70728040e-01
-3.72883350e-01 -1.74819734e-02 -3.35359126e-01 2.28560850e-01
-2.94627249e-01 4.53789562e-01 1.30957496e+00 -3.99979383e-01
-4.16187674e-01 2.15069801e-01 -4.76516485e-01 6.56738162e-01
1.12156785e+00 -5.09191751e-01 -4.74673331e-01 -7.05450535e-01
3.29952449e-01 -1.87135130e-01 7.53495693e-02 -4.62026983e-01
-2.64595687e-01 -6.04244351e-01 5.40175319e-01 -8.00345898e-01
-3.02046061e-01 -8.56147587e-01 -5.95876984e-02 7.02239513e-01
-4.94810134e-01 -1.88566521e-01 1.84730932e-01 3.39887977e-01
-3.26866597e-01 -3.07097495e-01 1.01100934e+00 -3.70914996e-01
-9.73288238e-01 5.04845142e-01 -5.16268671e-01 -1.59465939e-01
1.20084739e+00 2.64214892e-02 -3.57362509e-01 2.62537211e-01
-6.23920083e-01 6.51180744e-01 3.29704881e-01 4.42308456e-01
2.53886610e-01 -1.12412989e+00 -7.98006177e-01 2.30289832e-01
4.12165880e-01 6.06958687e-01 -6.69499561e-02 6.35158241e-01
-7.99851835e-01 4.33182776e-01 -4.63598937e-01 -7.25846708e-01
-1.01621032e+00 7.04688430e-01 4.22805957e-02 -9.78288651e-02
-2.08887994e-01 6.70322001e-01 2.18530640e-01 -8.70745301e-01
4.31338549e-01 -3.02638769e-01 -4.18507576e-01 3.69640052e-01
1.40233099e-01 5.46744049e-01 2.58337289e-01 -1.60481438e-01
-3.92498314e-01 4.55843538e-01 1.90499276e-01 6.36528254e-01
1.40634108e+00 1.04687661e-01 -3.40105593e-01 6.59393370e-01
8.97845030e-01 -7.04552472e-01 -1.09199679e+00 -3.90759200e-01
2.29805917e-01 -1.41004287e-02 2.38869473e-01 -8.54223132e-01
-1.36609149e+00 9.59808528e-01 8.73754323e-01 1.77913666e-01
1.33010626e+00 -3.10727030e-01 4.87817496e-01 7.41995513e-01
5.66430092e-01 -1.12679899e+00 -9.38044935e-02 4.82118458e-01
5.75629532e-01 -1.73738062e+00 3.99352282e-01 -7.08605766e-01
-3.72904181e-01 9.84418094e-01 6.30652547e-01 2.13193536e-01
9.87434864e-01 5.40392280e-01 4.44384187e-01 -7.30596185e-02
-3.99343371e-01 -3.51188332e-01 -3.03004175e-01 9.25250649e-01
9.99980807e-01 4.14570242e-01 -2.01910377e-01 2.57640928e-01
3.96939218e-01 5.20362794e-01 -9.49769281e-03 1.55826414e+00
-6.18024409e-01 -1.22249663e+00 -9.43765342e-02 5.55267751e-01
-1.93914577e-01 1.30643263e-01 -4.42064345e-01 4.55991566e-01
5.15234709e-01 1.26988292e+00 -2.58990347e-01 -7.04773515e-02
3.03767562e-01 -1.81162894e-01 3.64217103e-01 -9.52189445e-01
-3.54130536e-01 9.71882045e-02 -1.48165271e-01 -3.34348351e-01
-1.01699698e+00 -4.57511932e-01 -1.02146375e+00 -2.46619463e-01
-9.92818892e-01 1.69291291e-02 6.22583807e-01 6.69922471e-01
3.07210177e-01 3.80605251e-01 8.33129704e-01 -8.43643785e-01
-3.50703746e-01 -9.12526369e-01 -8.48089516e-01 -8.32549855e-02
2.45360196e-01 -9.53737557e-01 -1.62621200e-01 2.83436865e-01] | [9.157144546508789, -1.535788655281067] |
647d0671-bc8f-448c-b752-8414151eab5e | cross-domain-few-shot-classification-via-2 | 2208.08015 | null | https://arxiv.org/abs/2208.08015v1 | https://arxiv.org/pdf/2208.08015v1.pdf | Cross-Domain Few-Shot Classification via Inter-Source Stylization | Cross-Domain Few Shot Classification (CDFSC) leverages prior knowledge learned from a supervised auxiliary dataset to solve a target task with limited supervised information available, where the auxiliary and target datasets come from the different domains. It is challenging due to the domain shift between these datasets. Inspired by Multisource Domain Adaptation (MDA), the recent works introduce the multiple domains to improve the performance. However, they, on the one hand, evaluate only on the benchmark with natural images, and on the other hand, they need many annotations even in the source domains can be costly. To address the above mentioned issues, this paper explore a new Multisource CDFSC setting (MCDFSC) where only one source domain is fully labeled while the rest source domains remain unlabeled. These sources are from different fileds, means they are not only natural images. Considering the inductive bias of CNNs, this paper proposed Inter-Source stylization network (ISSNet) for this new MCDFSC setting. It transfers the styles of unlabeled sources to labeled source, which expands the distribution of labeled source and further improves the model generalization ability. Experiments on 8 target datasets demonstrate ISSNet effectively suppresses the performance degradation caused by different domains. | ['Li Liu', 'Huali Xu'] | 2022-08-17 | null | null | null | null | ['cross-domain-few-shot'] | ['computer-vision'] | [ 2.97945261e-01 -1.23026073e-01 -4.74758595e-01 -2.12561652e-01
-5.59662104e-01 -6.09038293e-01 6.22955620e-01 -2.10822120e-01
-4.09664392e-01 1.06732953e+00 2.95245945e-01 2.33858556e-01
1.49616286e-01 -6.94896698e-01 -7.26555586e-01 -6.49829805e-01
6.54280186e-01 3.36142510e-01 7.42718756e-01 -3.53346497e-01
-3.77323776e-02 -2.36055344e-01 -1.30104041e+00 3.16204637e-01
1.29437268e+00 1.03415573e+00 4.59056646e-01 -1.60549685e-01
-4.63391602e-01 7.81895280e-01 -7.00820625e-01 -3.15669507e-01
3.87727141e-01 -6.54654682e-01 -5.89360118e-01 3.79161835e-02
3.26031819e-02 -3.10688257e-01 -2.15548515e-01 1.27736688e+00
4.96705920e-01 1.55305400e-01 8.15893710e-01 -1.56303966e+00
-1.22186458e+00 5.56626320e-01 -8.22364688e-01 1.50812253e-01
-9.93440449e-02 1.24634996e-01 5.95632017e-01 -9.54197705e-01
7.33573854e-01 1.34943080e+00 3.87479097e-01 7.70642638e-01
-1.04968059e+00 -1.02331316e+00 4.12259281e-01 2.98765063e-01
-1.18211079e+00 -2.59316832e-01 1.03389239e+00 -4.88741159e-01
1.85406193e-01 -3.40731770e-01 1.48073703e-01 1.70167589e+00
-3.72412533e-01 1.00113571e+00 1.39759207e+00 -3.77770543e-01
3.84377837e-01 5.83874941e-01 3.21326166e-01 6.21155463e-02
1.77325651e-01 1.20311044e-02 -3.95902038e-01 1.21782310e-01
6.13348007e-01 1.31573662e-01 -4.74250436e-01 -5.65840840e-01
-1.15555370e+00 9.02677536e-01 3.69966537e-01 2.67550468e-01
1.26072571e-01 -6.20244741e-01 6.72936976e-01 4.36206311e-01
5.53116918e-01 1.79607078e-01 -6.37316406e-01 1.22468799e-01
-4.48534042e-01 -1.04813598e-01 6.33179009e-01 1.35587692e+00
8.22820485e-01 1.42255083e-01 -6.58823401e-02 1.17107666e+00
-2.90856492e-02 6.55216277e-01 8.43398392e-01 -5.05734682e-01
8.22014987e-01 7.33627856e-01 8.12402442e-02 -7.96265006e-01
3.76285724e-02 -4.46723104e-01 -8.82578850e-01 -1.47252413e-03
4.99971241e-01 -2.68252760e-01 -8.87403071e-01 2.01808310e+00
3.01501989e-01 3.90426308e-01 4.03436750e-01 1.03796303e+00
1.06411636e+00 6.19720995e-01 2.20608920e-01 -1.62322745e-01
1.12833202e+00 -1.25175047e+00 -5.95447481e-01 -4.11323220e-01
5.32778919e-01 -5.50108433e-01 1.47019815e+00 1.88007608e-01
-3.46489042e-01 -7.55074441e-01 -1.20388627e+00 2.86903251e-02
-6.97731674e-01 1.61051691e-01 1.47669077e-01 2.92624712e-01
-4.06928420e-01 3.68830919e-01 -2.38708571e-01 -4.24840719e-01
5.86320698e-01 -8.87703821e-02 -3.20437849e-01 -4.53778267e-01
-1.70489204e+00 7.27037072e-01 7.34388232e-01 -4.68028843e-01
-8.40663433e-01 -8.52217436e-01 -7.99521863e-01 -3.74385342e-02
8.17145407e-01 -1.06722407e-01 9.87093806e-01 -1.51095784e+00
-1.35782671e+00 7.49365926e-01 1.21026576e-01 -2.91777551e-01
6.10942602e-01 -1.95450544e-01 -6.04835391e-01 -4.62319590e-02
5.71807504e-01 6.95410669e-01 9.55468178e-01 -1.31009877e+00
-6.07448757e-01 -3.08985412e-01 6.41684607e-02 2.21477583e-01
-6.03759527e-01 -3.14968407e-01 -4.39841300e-01 -8.47752810e-01
-4.54620034e-01 -8.11075091e-01 2.45714977e-01 -1.76809490e-01
-1.93045810e-01 -2.16943562e-01 1.14894664e+00 -3.54734242e-01
9.88021016e-01 -2.29017067e+00 1.15791589e-01 -2.78803557e-01
-2.01647840e-02 5.08328259e-01 -4.10021871e-01 1.31514326e-01
-1.91201895e-01 -7.04644620e-02 -3.59924078e-01 -6.43836483e-02
-1.41039953e-01 3.60666841e-01 -4.44587559e-01 9.13227051e-02
3.61132473e-01 7.73937345e-01 -9.79252815e-01 -7.24348068e-01
4.47173901e-02 4.52168621e-02 -2.49536648e-01 1.64393395e-01
-2.87547231e-01 7.48787940e-01 -5.14781654e-01 5.13368070e-01
1.11720312e+00 -3.34592849e-01 -1.57696754e-01 -1.88659742e-01
2.76212841e-01 -3.03670555e-01 -1.23334253e+00 2.00910902e+00
-3.39240849e-01 1.60963371e-01 -1.23647727e-01 -1.21124744e+00
9.87639248e-01 1.48599789e-01 3.11654389e-01 -8.98724616e-01
2.03764230e-01 2.71441877e-01 -4.08380814e-02 -3.45410943e-01
2.54895799e-02 -4.71174717e-01 -2.65763700e-01 3.06743056e-01
3.90437871e-01 3.36656868e-01 2.22040251e-01 2.58740902e-01
7.48139560e-01 4.11022544e-01 3.61659557e-01 -3.70047241e-01
6.26658142e-01 1.59399107e-01 1.16875434e+00 3.96071315e-01
-5.79574168e-01 8.04632723e-01 5.56063533e-01 -7.86931068e-02
-8.85017276e-01 -1.11034548e+00 -1.58889130e-01 1.19553709e+00
7.43326306e-01 -1.34658497e-02 -7.32769370e-01 -1.20341682e+00
-4.39301617e-02 5.68904102e-01 -8.40113997e-01 -4.38638747e-01
-2.54509509e-01 -6.01619244e-01 3.99840355e-01 6.47881389e-01
9.09007788e-01 -9.26805139e-01 -4.14012194e-01 1.15653530e-01
-2.90780991e-01 -1.34269166e+00 -7.66854763e-01 1.57263324e-01
-5.89651883e-01 -1.23213446e+00 -1.22554696e+00 -8.67714584e-01
4.85767692e-01 4.74068344e-01 9.54629004e-01 -5.18087327e-01
1.01449706e-01 6.63350448e-02 -8.47758353e-01 -5.43708026e-01
-2.47291088e-01 1.39169812e-01 2.03754440e-01 2.93136120e-01
7.17877030e-01 -6.33504093e-01 -3.78232718e-01 5.49997151e-01
-1.08552074e+00 9.81424749e-02 4.82824981e-01 1.13726711e+00
4.03590947e-01 1.29628241e-01 1.13432467e+00 -1.30904996e+00
3.91579181e-01 -1.00485957e+00 -3.27084869e-01 2.95644104e-01
-4.94984388e-01 4.25665602e-02 1.10851240e+00 -8.51597011e-01
-1.54556370e+00 -1.06091537e-01 3.63388270e-01 -8.53564441e-01
-3.86044919e-01 1.92778334e-01 -6.78225040e-01 3.06221753e-01
8.51985574e-01 1.85985461e-01 -1.91601068e-02 -7.11961091e-01
2.47267991e-01 7.87602425e-01 5.04322708e-01 -6.16287172e-01
8.92907560e-01 4.48427498e-01 -4.48614031e-01 -6.43127680e-01
-1.09671164e+00 -4.85012472e-01 -7.76774466e-01 8.02990645e-02
7.87344098e-01 -1.11854482e+00 2.21299529e-01 6.70647562e-01
-1.17002964e+00 -2.57990777e-01 -3.64518970e-01 3.96648943e-01
-1.88878864e-01 3.06263953e-01 -2.56808728e-01 -3.95988554e-01
-5.36109973e-03 -1.02756572e+00 9.00318265e-01 4.95354861e-01
2.53257584e-02 -1.06523323e+00 -1.49587497e-01 3.85619774e-02
2.25481704e-01 4.52707052e-01 1.02631998e+00 -9.82432604e-01
-7.38399625e-02 2.30456606e-01 -5.10876238e-01 6.67162120e-01
4.84422535e-01 -6.19731247e-01 -1.00466323e+00 -1.88661143e-01
1.65002689e-01 -6.95084572e-01 8.43000472e-01 5.31129912e-02
8.55692267e-01 -7.23690316e-02 -2.94153512e-01 3.88925374e-01
1.51604843e+00 3.55612546e-01 3.99750739e-01 4.37884420e-01
7.10047781e-01 7.79446125e-01 9.79108155e-01 3.71267736e-01
2.53227562e-01 3.95810723e-01 1.98533922e-01 -8.50077048e-02
-3.11257154e-01 -4.96715099e-01 4.45808738e-01 7.24676371e-01
3.09595764e-01 -1.38347194e-01 -7.68499255e-01 7.40144551e-01
-1.78307271e+00 -6.38844073e-01 1.85156703e-01 1.96022487e+00
9.78816986e-01 2.20993012e-01 2.05208451e-01 -6.67217225e-02
1.12204862e+00 1.26231030e-01 -1.17736340e+00 2.34965220e-01
-4.44980264e-01 9.61621031e-02 4.64072138e-01 -4.47577350e-02
-1.16184473e+00 9.18015599e-01 4.26831341e+00 1.25588667e+00
-1.05386865e+00 5.14466643e-01 4.93661225e-01 -1.61072120e-01
-1.56380206e-01 -7.75629133e-02 -7.62720704e-01 1.05355513e+00
5.85317075e-01 -2.52704203e-01 1.52769655e-01 1.15751302e+00
-2.43999973e-01 -1.12947673e-01 -9.67476606e-01 1.08590424e+00
2.05861956e-01 -8.78737569e-01 1.81293599e-02 -1.51122257e-01
9.20910180e-01 5.03902212e-02 3.91793214e-02 8.82134736e-01
4.54062968e-01 -5.47901273e-01 6.65813863e-01 1.85207233e-01
1.15950048e+00 -7.35591829e-01 7.65719950e-01 6.29148841e-01
-1.10858524e+00 -2.48333797e-01 -7.63498366e-01 1.36507764e-01
-4.55998629e-02 4.13924068e-01 -3.28958511e-01 6.41587615e-01
8.44284356e-01 1.06816900e+00 -6.94582820e-01 6.46537364e-01
-2.69483053e-03 5.02924919e-01 -6.84848428e-02 2.81430095e-01
1.13274485e-01 -2.25417838e-01 4.23153758e-01 9.04503405e-01
3.04487973e-01 1.61189020e-01 3.68045419e-01 8.67993116e-01
-2.19252601e-01 1.03985853e-01 -8.89742374e-01 1.64973468e-01
5.63420534e-01 9.46417511e-01 -6.23602450e-01 -3.45800221e-01
-6.67950928e-01 1.04512453e+00 3.12230080e-01 5.08783996e-01
-9.82005179e-01 -6.08497143e-01 4.28325385e-01 1.26916632e-01
1.78889364e-01 2.67357349e-01 -2.68751860e-01 -1.64056122e+00
1.28474593e-01 -8.48559678e-01 5.12524009e-01 -4.97309119e-01
-1.83003676e+00 4.57761735e-01 1.93702042e-01 -1.77282226e+00
1.49056777e-01 -5.07288516e-01 -4.65015560e-01 8.41803074e-01
-1.91236460e+00 -1.17104626e+00 -4.09486681e-01 9.30604696e-01
9.03372288e-01 -5.17075717e-01 5.58197021e-01 4.97323424e-01
-5.55939555e-01 6.70566440e-01 4.75044608e-01 3.07192087e-01
1.29119670e+00 -1.04538512e+00 6.32047132e-02 8.44305992e-01
-3.21437001e-01 4.57048833e-01 3.10779989e-01 -6.86628103e-01
-9.29749787e-01 -1.34810555e+00 1.45736456e-01 -2.81412423e-01
6.63342535e-01 -3.50472093e-01 -1.31223726e+00 5.34402311e-01
3.25366825e-01 1.61507085e-01 7.11209297e-01 -2.34285742e-01
-6.62290215e-01 -2.56967604e-01 -1.09494960e+00 3.64867151e-01
1.17040718e+00 -3.15004200e-01 -8.33272815e-01 1.30244270e-01
9.14818525e-01 -1.91520751e-01 -6.12598836e-01 4.07740533e-01
7.61648733e-03 -8.21590245e-01 6.61393881e-01 -6.61813259e-01
7.71314085e-01 -3.73765379e-01 -9.19200927e-02 -1.69276154e+00
-1.20123416e-01 6.64839298e-02 -1.24189399e-01 1.80427027e+00
1.32027045e-01 -6.98486209e-01 6.15338564e-01 4.36624825e-01
-2.04571579e-02 -3.13389093e-01 -8.53933692e-01 -1.28962219e+00
4.67371225e-01 -5.33262594e-03 7.32991159e-01 1.25319397e+00
-1.04372911e-01 8.41793060e-01 -6.65398598e-01 -1.82509303e-01
5.65040052e-01 3.63830984e-01 6.71118379e-01 -1.46432054e+00
-2.06504822e-01 -9.33002755e-02 -1.11794323e-02 -1.03528869e+00
4.61017638e-01 -9.28102314e-01 5.93965240e-02 -1.09088004e+00
2.75430560e-01 -6.29011512e-01 -4.14239287e-01 4.68451709e-01
-3.42990398e-01 1.56874128e-03 3.56712192e-01 3.72678846e-01
-6.91178620e-01 9.57509696e-01 1.35847044e+00 -2.68467456e-01
-1.24412350e-01 -3.14498276e-01 -1.03796601e+00 7.69344985e-01
7.44541883e-01 -5.52250326e-01 -9.56715047e-01 -6.17185593e-01
-4.33774620e-01 -1.14841618e-01 1.74300492e-01 -1.18793428e+00
6.81709573e-02 -2.93057680e-01 3.78840387e-01 -3.96410763e-01
1.87699631e-01 -9.28315699e-01 -2.48064503e-01 1.60686180e-01
-2.29484573e-01 -4.98386353e-01 1.47154033e-01 1.05851614e+00
-4.47428107e-01 -2.74965465e-01 1.33861148e+00 -1.88821688e-01
-1.26392150e+00 2.95093805e-01 3.45133692e-01 7.47976303e-01
1.41678071e+00 -1.09772488e-01 -4.65563923e-01 -1.26709998e-01
-5.58836818e-01 3.92145365e-01 5.12443483e-01 6.84720218e-01
3.58980596e-01 -1.74965656e+00 -6.34945750e-01 1.00446805e-01
5.46658695e-01 2.60750115e-01 4.57139879e-01 6.17653966e-01
8.86600465e-02 2.06960797e-01 -5.64965606e-01 -5.14942765e-01
-9.00593817e-01 9.79432166e-01 5.43278046e-02 -2.04572424e-01
-4.29645985e-01 7.39418328e-01 7.64993608e-01 -5.09567559e-01
3.41843674e-03 8.59701559e-02 -4.55252498e-01 4.52091157e-01
5.41459560e-01 3.86947393e-01 -1.95269927e-01 -7.07725465e-01
-1.78010106e-01 5.78145385e-01 -2.92991996e-01 1.10132478e-01
1.17800224e+00 -2.66959488e-01 3.25408697e-01 5.73030114e-01
1.34540570e+00 -3.25658470e-01 -1.57865810e+00 -8.93384516e-01
8.05913657e-02 -5.32760978e-01 -4.76418823e-01 -8.18560719e-01
-1.14802706e+00 1.15381205e+00 5.71156442e-01 -2.10858330e-01
1.24542511e+00 -1.10645726e-01 1.05450368e+00 1.39872134e-02
5.15360475e-01 -1.57072139e+00 3.56651574e-01 7.38475919e-01
6.17491782e-01 -1.67440689e+00 -3.61918956e-01 -3.54867429e-01
-1.31481767e+00 8.62605870e-01 1.20871770e+00 -1.83799833e-01
6.17712259e-01 -3.84856872e-02 -5.04431687e-02 3.22824776e-01
-4.97256994e-01 -3.27204645e-01 -1.08656265e-01 1.05415547e+00
1.16267465e-02 -1.27632082e-01 -2.86825091e-01 1.34655285e+00
2.74890780e-01 1.94178671e-01 5.52248836e-01 7.12207437e-01
-2.36922219e-01 -1.18628967e+00 -3.43238056e-01 4.06725049e-01
-1.93820775e-01 -2.09531728e-02 -3.38086307e-01 8.65522861e-01
6.85478449e-01 9.51463103e-01 -2.63785720e-01 -3.44796389e-01
5.14409244e-01 1.41390175e-01 -5.38524278e-02 -7.74797261e-01
-5.62507249e-02 -2.16590129e-02 -2.98550367e-01 -1.14225395e-01
-5.36045909e-01 -4.94916677e-01 -1.16762340e+00 8.61112773e-02
-1.96850836e-01 2.77362466e-01 -2.95441244e-02 8.19284201e-01
3.98821205e-01 4.67223674e-01 7.22221851e-01 -4.17912602e-01
-7.43819952e-01 -1.12041664e+00 -8.84645045e-01 7.57748246e-01
2.67588526e-01 -1.27410436e+00 -2.99325138e-01 2.15760782e-01] | [10.260570526123047, 2.9254136085510254] |
431d14c8-74a9-4200-8aa5-8cc84b69d07d | impact-of-face-image-quality-estimation-on | 2209.15489 | null | https://arxiv.org/abs/2209.15489v1 | https://arxiv.org/pdf/2209.15489v1.pdf | Impact of Face Image Quality Estimation on Presentation Attack Detection | Non-referential face image quality assessment methods have gained popularity as a pre-filtering step on face recognition systems. In most of them, the quality score is usually designed with face matching in mind. However, a small amount of work has been done on measuring their impact and usefulness on Presentation Attack Detection (PAD). In this paper, we study the effect of quality assessment methods on filtering bona fide and attack samples, their impact on PAD systems, and how the performance of such systems is improved when training on a filtered (by quality) dataset. On a Vision Transformer PAD algorithm, a reduction of 20% of the training dataset by removing lower quality samples allowed us to improve the BPCER by 3% in a cross-dataset test. | ['Christoph Busch', 'Juan E. Tapia', 'Diego Pasmino', 'Carlos Aravena'] | 2022-09-30 | null | null | null | null | ['image-quality-estimation', 'face-image-quality', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.16135398e-01 -1.07262462e-01 1.19655170e-01 -4.39095646e-01
-5.10830998e-01 -5.46534359e-01 6.05610073e-01 2.06846356e-01
-3.88627112e-01 2.65037328e-01 -2.62584500e-02 -3.62648726e-01
-1.69277042e-01 -9.15484607e-01 -4.73339587e-01 -5.12446702e-01
-6.65495470e-02 1.23560317e-02 3.72301489e-01 -9.47652757e-02
1.90425992e-01 8.17557037e-01 -1.64790559e+00 4.80631977e-01
4.18650925e-01 1.05393159e+00 -3.94639730e-01 3.89142543e-01
-6.08281093e-03 2.79196292e-01 -1.10576284e+00 -9.14772928e-01
5.34491181e-01 -4.80221897e-01 -4.92747873e-01 -2.17453111e-03
9.02275264e-01 -4.40987349e-01 -2.17103302e-01 9.38621998e-01
8.01053882e-01 -3.89709562e-01 5.49196482e-01 -1.24209452e+00
-8.35531279e-02 4.85291183e-01 -3.59638900e-01 4.50601190e-01
5.57822168e-01 2.25474715e-01 5.48185229e-01 -9.64254856e-01
3.97144616e-01 1.33870232e+00 5.71070135e-01 2.93797106e-01
-1.42968857e+00 -1.01030254e+00 -4.36942548e-01 4.22228158e-01
-1.47370660e+00 -6.37784481e-01 6.05646193e-01 -3.93340498e-01
5.45397103e-01 2.84153402e-01 5.51229358e-01 6.96262240e-01
3.65229957e-02 1.13724492e-01 1.21225953e+00 -7.26674438e-01
1.30627275e-01 4.63613659e-01 3.45237643e-01 4.31457222e-01
4.45868850e-01 2.20672503e-01 -5.85273862e-01 -2.44291335e-01
5.50570965e-01 -5.58881938e-01 -1.80773765e-01 -1.35251477e-01
-2.51259536e-01 7.80889034e-01 -1.86036695e-02 5.13234198e-01
-3.06352470e-02 -2.28594080e-01 2.99761713e-01 5.26874363e-01
3.37589651e-01 4.32419568e-01 -8.73466209e-02 5.86813129e-02
-1.14775288e+00 2.61519961e-02 8.04430783e-01 2.97315270e-01
5.72494686e-01 5.37535511e-02 -5.65595746e-01 8.13505769e-01
2.04320084e-02 5.51718771e-01 1.59415975e-01 -5.18323123e-01
8.05612057e-02 6.61715209e-01 -2.10331813e-01 -1.05125058e+00
-2.09862784e-01 -1.85433060e-01 -3.59887302e-01 5.79424322e-01
7.91213274e-01 -1.27079725e-01 -8.17005634e-01 1.45479870e+00
1.08667627e-01 8.25660750e-02 -3.31666380e-01 6.22262955e-01
4.90464568e-01 3.00860345e-01 4.69674356e-02 -3.01701397e-01
1.47641289e+00 -1.50688171e-01 -5.64311624e-01 2.17613265e-01
2.92063475e-01 -1.43569899e+00 1.07590711e+00 7.84275651e-01
-9.85532045e-01 -6.89667165e-01 -1.22513056e+00 4.21431690e-01
-3.30494970e-01 2.24776193e-01 1.62927806e-01 1.75775826e+00
-9.38300073e-01 5.80874383e-01 -2.55487680e-01 -2.71512419e-01
3.85560274e-01 5.85679829e-01 -5.54817438e-01 -7.47288316e-02
-1.11774445e+00 1.10348046e+00 -2.24111408e-01 -4.76656675e-01
-6.89871132e-01 -7.48562694e-01 -3.17993015e-01 1.32270858e-01
2.58807093e-01 -3.28833330e-03 8.97367954e-01 -8.47746789e-01
-1.44237673e+00 9.15156126e-01 7.20272809e-02 -6.07071400e-01
6.34787381e-01 -8.99221674e-02 -8.81672323e-01 3.08053911e-01
-3.65198880e-01 3.51518840e-01 1.29067636e+00 -8.23376179e-01
-3.33123922e-01 -4.47995037e-01 1.76481590e-01 -4.98302698e-01
-4.28407609e-01 4.17745322e-01 -3.15690935e-01 -6.16097629e-01
-3.26107115e-01 -7.52767861e-01 4.03708130e-01 7.91138411e-02
2.02263713e-01 -7.16902241e-02 9.62409496e-01 -7.50991523e-01
1.21823752e+00 -2.33353996e+00 -5.30783832e-01 5.72526872e-01
-1.53601512e-01 8.83449137e-01 -2.02715784e-01 2.63524920e-01
-2.47322559e-01 1.83588937e-01 3.04431349e-01 3.51398885e-02
-1.43083751e-01 -1.69927225e-01 -2.90036798e-01 4.71282452e-01
4.57729042e-01 2.76529640e-01 -2.71458030e-01 -2.89697438e-01
4.47992086e-01 9.25110638e-01 -6.74389362e-01 1.61947638e-01
2.55138010e-01 8.49454254e-02 9.44531336e-02 4.36232269e-01
9.29542422e-01 2.26058081e-01 2.95511425e-01 -4.67615277e-01
1.34133965e-01 2.18673810e-01 -1.38399124e+00 9.04313743e-01
-3.83949012e-01 7.70127118e-01 -1.51747182e-01 -6.04540050e-01
1.02676272e+00 4.30929482e-01 3.95160884e-01 -1.00542843e+00
2.73701161e-01 2.27119084e-02 4.11309540e-01 -2.14340061e-01
2.84264535e-01 -2.82632355e-02 3.46756697e-01 3.82142127e-01
2.11746663e-01 3.21121305e-01 4.13216889e-01 -3.53551582e-02
9.49528456e-01 -3.92543644e-01 2.44623180e-02 -3.24789345e-01
8.52168858e-01 -4.66091603e-01 1.70610413e-01 4.73549038e-01
-2.28825092e-01 6.69586003e-01 6.87230289e-01 -2.16773838e-01
-8.60529840e-01 -1.00092208e+00 -4.95649397e-01 7.05476820e-01
-8.56680647e-02 -8.07236612e-01 -1.27326930e+00 -6.56574845e-01
-6.89414749e-03 5.57098806e-01 -3.95353287e-01 -4.16983396e-01
-5.20096958e-01 -8.98515582e-01 8.54351521e-01 1.79906964e-01
4.36102867e-01 -8.85242641e-01 -6.86063409e-01 -1.48406625e-02
1.89383030e-01 -1.04430163e+00 -3.59727263e-01 -2.12158561e-01
-6.33527040e-01 -1.42264295e+00 -4.35977697e-01 -2.56246477e-01
6.58420324e-01 2.45403066e-01 9.69724536e-01 3.09918225e-01
-5.32665312e-01 4.30787891e-01 -3.14910531e-01 -3.85971665e-01
-4.81287509e-01 -3.81055593e-01 -2.34242086e-03 4.21481043e-01
4.92120355e-01 -1.88152343e-01 -3.61830384e-01 7.35222340e-01
-8.84739697e-01 -6.03407681e-01 4.78216887e-01 5.57531476e-01
1.53908238e-01 2.06830755e-01 2.99034625e-01 -6.55253708e-01
8.28978479e-01 9.72385854e-02 -7.63275921e-01 3.11762422e-01
-8.10527623e-01 1.76405162e-02 2.77135819e-01 -6.53654993e-01
-8.27682137e-01 -1.72679052e-01 -2.27622628e-01 -3.09087843e-01
1.35283787e-02 9.36627388e-03 -3.44816893e-01 -6.19478524e-01
7.42542088e-01 -2.21957967e-01 3.38295043e-01 -5.58906436e-01
-2.44524032e-02 5.08223772e-01 1.58589736e-01 -3.97044152e-01
7.93533146e-01 1.15106948e-01 1.78645492e-01 -8.25594187e-01
-1.54005334e-01 -2.18675330e-01 -4.19024408e-01 -4.62993830e-01
3.39454353e-01 -5.85987687e-01 -7.94081151e-01 5.98282754e-01
-7.58834839e-01 -6.80546537e-02 -3.19907032e-02 4.39735115e-01
1.49070648e-02 4.98280674e-01 -2.78757811e-01 -8.26982200e-01
-2.74007231e-01 -1.21003115e+00 6.32019758e-01 1.45766631e-01
-3.11299622e-01 -4.68300730e-01 -2.47395471e-01 3.36425781e-01
6.21824443e-01 -4.39504758e-02 7.85904706e-01 -3.95449072e-01
-1.60602331e-01 -4.97123927e-01 -2.19994843e-01 7.09874570e-01
7.16323853e-02 1.00175083e-01 -1.30092967e+00 -3.40720773e-01
1.01695031e-01 2.24084452e-01 6.09267056e-01 1.50770798e-01
7.88470030e-01 -1.66613176e-01 4.03575376e-02 1.87441230e-01
1.37113774e+00 2.11192489e-01 9.46542025e-01 5.50624495e-03
-2.81969868e-02 6.85020924e-01 6.27283216e-01 3.71238947e-01
-4.85640466e-01 1.08407116e+00 2.09929958e-01 7.80643299e-02
-4.73019838e-01 6.45468310e-02 5.44210970e-01 -1.26985282e-01
7.05694258e-02 -6.50235116e-02 -7.57462263e-01 1.07058324e-01
-1.02619135e+00 -1.10566866e+00 -1.55377118e-02 2.37812877e+00
5.46102166e-01 3.78076702e-01 5.25367260e-01 8.09634566e-01
8.12264204e-01 -1.94198832e-01 1.09355830e-01 -4.27307367e-01
-4.06408608e-02 6.67904377e-01 4.48330522e-01 3.57838899e-01
-7.06238449e-01 4.09221530e-01 6.79191351e+00 8.51541221e-01
-1.54500580e+00 8.33607838e-02 6.91874504e-01 -1.27053335e-01
1.04572833e-01 -1.09522440e-01 -9.77554321e-01 7.17239499e-01
1.00161648e+00 -4.10372429e-02 2.42628425e-01 7.33828008e-01
1.24235630e-01 -3.59413475e-01 -8.69962811e-01 1.21197033e+00
3.90021890e-01 -9.86330450e-01 3.69364023e-02 3.22551876e-01
-3.00330874e-02 -6.56532228e-01 2.11288750e-01 1.80450946e-01
-4.02373105e-01 -1.10871065e+00 5.82568407e-01 2.44060621e-01
7.04061687e-01 -8.95227313e-01 8.95576060e-01 -2.14421421e-01
-8.27885449e-01 -6.08180575e-02 -1.78997561e-01 1.25289872e-01
-1.09781086e-01 7.05878913e-01 -1.04504859e+00 2.54436105e-01
6.38570607e-01 -3.17924172e-01 -1.00756025e+00 1.14634752e+00
-3.74912396e-02 8.20095539e-01 -5.35298944e-01 2.06056029e-01
-3.55115622e-01 -2.76869405e-02 3.56157154e-01 1.14176381e+00
3.03610355e-01 -3.48764546e-02 -3.72631788e-01 5.66126347e-01
-1.74565464e-02 3.12271446e-01 -2.42196888e-01 1.05807364e-01
3.10589194e-01 1.34153652e+00 -7.25834370e-01 -1.41395882e-01
-4.17757571e-01 4.81178939e-01 -2.22363114e-01 -2.01414570e-01
-8.18183780e-01 -2.44689286e-01 6.07982457e-01 7.13283002e-01
4.07943815e-01 1.27900884e-01 -1.30191877e-01 -6.01383984e-01
7.35658184e-02 -1.09527731e+00 5.06134331e-01 -5.06477356e-01
-9.93130505e-01 6.20660901e-01 -1.99310877e-03 -1.06641185e+00
-2.02046320e-01 -7.30861425e-01 -5.65140247e-01 1.00297070e+00
-1.08079028e+00 -8.37695122e-01 -3.32305819e-01 5.85834324e-01
1.17827870e-01 -4.32913929e-01 6.97039306e-01 7.59688437e-01
-3.65951806e-01 1.15270329e+00 -4.18139100e-01 2.30596542e-01
8.43284905e-01 -6.88550115e-01 1.45567596e-01 1.05046380e+00
3.32340091e-01 5.68939924e-01 6.90972090e-01 -3.05510789e-01
-1.22502041e+00 -6.90128505e-01 6.19205356e-01 -3.50421399e-01
1.51817828e-01 -1.91830441e-01 -9.29576278e-01 -5.21648489e-02
7.32271671e-02 5.83919026e-02 6.88957989e-01 7.55036473e-02
-7.42268264e-01 -5.37162781e-01 -1.68774235e+00 3.53758276e-01
4.79177475e-01 -6.02174938e-01 -2.85175830e-01 -2.36121804e-01
-1.30253866e-01 2.43422352e-02 -1.00616038e+00 5.80241024e-01
7.55162597e-01 -1.28936028e+00 1.05456114e+00 -3.19675691e-02
-1.43646240e-01 -2.80332744e-01 7.36847743e-02 -9.30006921e-01
-1.99114218e-01 -3.95440102e-01 9.31704342e-02 1.44546318e+00
1.51418090e-01 -3.78769159e-01 8.12254429e-01 4.82335389e-01
4.11870122e-01 -2.68238872e-01 -8.13739538e-01 -7.92337000e-01
-2.99551547e-01 -2.94039935e-01 6.39151752e-01 5.45634568e-01
-1.33209825e-01 2.46344611e-01 -2.80042082e-01 5.19389659e-02
4.92461354e-01 -3.26538593e-01 8.31956983e-01 -1.28642881e+00
-1.87420025e-01 -4.67914432e-01 -8.40726256e-01 5.83595596e-02
-3.35771859e-01 -5.14192164e-01 -6.60508096e-01 -6.87428474e-01
-7.53974319e-02 -7.17414841e-02 -2.79627472e-01 4.06952649e-01
1.03001380e-02 7.27108121e-01 6.00386143e-01 -1.08590379e-01
-1.71337312e-03 -2.22091019e-01 6.64042115e-01 -1.42740726e-01
1.39126750e-02 5.86739779e-02 -4.37552303e-01 3.57103109e-01
8.64972353e-01 -5.02676785e-01 -3.65415007e-01 1.28349485e-02
-8.49537253e-02 -9.76024792e-02 3.88786882e-01 -1.43634188e+00
-7.13922009e-02 3.03851485e-01 6.83738530e-01 -3.64990145e-01
4.73512888e-01 -8.77448738e-01 4.55614477e-01 6.14387512e-01
-3.66334505e-02 5.18266261e-02 4.40457851e-01 -1.36253806e-02
-2.15540931e-01 -4.73161310e-01 1.05481756e+00 2.89320499e-01
-4.84704077e-01 -5.39205633e-02 -2.60776728e-01 -2.22063914e-01
1.01435184e+00 -4.52768117e-01 -1.52142540e-01 -2.52812684e-01
-6.04787648e-01 -5.29672921e-01 4.93201762e-01 4.39824492e-01
3.98016036e-01 -1.05834293e+00 -6.67602420e-01 7.34144032e-01
-3.78742144e-02 -9.82909262e-01 8.54661241e-02 7.27454424e-01
-2.66942620e-01 2.78733820e-01 -7.72503793e-01 -4.58293170e-01
-2.06950307e+00 4.73733187e-01 2.89495468e-01 -7.78287649e-05
-1.80761367e-01 5.43960750e-01 -1.76003978e-01 2.38995597e-01
3.42965990e-01 2.43787766e-01 -3.42313290e-01 4.57433403e-01
8.28808010e-01 5.04868269e-01 6.81583226e-01 -5.91037571e-01
-4.46026057e-01 5.01994371e-01 -5.32659739e-02 -2.96171457e-01
1.04745758e+00 3.26805383e-01 -1.62216704e-02 -1.94825560e-01
1.16026616e+00 3.62753659e-01 -7.64867544e-01 3.87671962e-02
-2.75240503e-02 -8.31409454e-01 1.46979004e-01 -7.20030725e-01
-1.24142623e+00 8.86589527e-01 1.35435688e+00 4.24763620e-01
1.29561079e+00 -4.12957489e-01 2.82676637e-01 -6.22353293e-02
3.43068272e-01 -8.79974961e-01 2.19032645e-01 1.22702949e-01
7.26270080e-01 -9.15126383e-01 1.03123531e-01 -6.87647343e-01
-2.50352472e-01 1.26715887e+00 4.54962790e-01 -1.21937774e-01
8.49348485e-01 3.75566423e-01 2.32208949e-02 -5.72272837e-02
-3.93640429e-01 -2.24124938e-02 3.73278260e-01 7.06547379e-01
4.70142066e-01 -1.31980956e-01 -6.84203327e-01 4.81226057e-01
-1.62849829e-01 1.11825541e-01 4.94685918e-01 6.80848360e-01
-2.32597277e-01 -1.49712980e+00 -7.61378765e-01 6.13195539e-01
-8.29671681e-01 2.53709275e-02 -5.07849872e-01 7.21075058e-01
2.84811586e-01 1.14175308e+00 1.23183928e-01 -5.00761390e-01
5.47945797e-01 2.89541095e-01 7.97212303e-01 -3.02070320e-01
-9.82184470e-01 -7.51436129e-02 1.89035058e-01 -3.98879528e-01
-2.79308259e-01 -7.29834557e-01 -6.66588545e-01 -6.33544445e-01
-5.68806946e-01 1.61404103e-01 7.96841621e-01 6.73465133e-01
3.55102450e-01 3.90985340e-01 6.14857435e-01 -3.19198817e-01
-5.60964286e-01 -9.00566101e-01 -4.40934986e-01 5.39970100e-01
-8.86482969e-02 -7.91758358e-01 -3.74451697e-01 5.29506197e-03] | [13.042490005493164, 0.9620637893676758] |
867fc792-687f-4b3c-8b40-6332d7c03a55 | trap-attention-monocular-depth-estimation | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ning_Trap_Attention_Monocular_Depth_Estimation_With_Manual_Traps_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ning_Trap_Attention_Monocular_Depth_Estimation_With_Manual_Traps_CVPR_2023_paper.pdf | Trap Attention: Monocular Depth Estimation With Manual Traps | Predicting a high quality depth map from a single image is a challenging task, because it exists infinite possibility to project a 2D scene to the corresponding 3D scene. Recently, some studies introduced multi-head attention (MHA) modules to perform long-range interaction, which have shown significant progress in regressing the depth maps.The main functions of MHA can be loosely summarized to capture long-distance information and report the attention map by the relationship between pixels. However, due to the quadratic complexity of MHA, these methods can not leverage MHA to compute depth features in high resolution with an appropriate computational complexity. In this paper, we exploit a depth-wise convolution to obtain long-range information, and propose a novel trap attention, which sets some traps on the extended space for each pixel, and forms the attention mechanism by the feature retention ratio of convolution window, resulting in that the quadratic computational complexity can be converted to linear form. Then we build an encoder-decoder trap depth estimation network, which introduces a vision transformer as the encoder, and uses the trap attention to estimate the depth from single image in the decoder. Extensive experimental results demonstrate that our proposed network can outperform the state-of-the-art methods in monocular depth estimation on datasets NYU Depth-v2 and KITTI, with significantly reduced number of parameters. Code is available at: https://github.com/ICSResearch/TrapAttention. | ['Hongping Gan', 'Chao Ning'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['monocular-depth-estimation'] | ['computer-vision'] | [ 7.69247711e-02 -1.44513026e-01 1.79515071e-02 -4.71428305e-01
-5.49160898e-01 -6.21308722e-02 4.99697179e-01 -6.12204611e-01
-4.93885875e-01 4.00279999e-01 2.80787379e-01 -1.56989619e-02
1.74134746e-01 -9.90165532e-01 -9.64095473e-01 -8.49666893e-01
3.50753367e-01 7.71724880e-02 4.26413000e-01 1.24021508e-01
4.87437427e-01 3.28914434e-01 -1.55755138e+00 3.08416337e-01
8.95080388e-01 1.28807676e+00 8.23692620e-01 6.84816301e-01
-1.28527984e-01 9.89482343e-01 -2.05269754e-01 -4.65498380e-02
2.49644756e-01 -2.38031596e-01 -5.32205343e-01 -1.37744591e-01
3.88700426e-01 -1.18465173e+00 -8.25708508e-01 9.68928754e-01
7.28493571e-01 -9.97592360e-02 5.01906216e-01 -9.84340131e-01
-8.11213970e-01 1.96623400e-01 -8.65670204e-01 2.53004044e-01
4.11515325e-01 2.56881922e-01 6.79655194e-01 -1.22441423e+00
3.39583963e-01 1.28825772e+00 2.37898454e-01 5.97323775e-01
-6.62803590e-01 -7.80067325e-01 3.06385249e-01 3.32581848e-01
-1.40363419e+00 -4.08768088e-01 6.41496897e-01 -3.32735866e-01
1.34312022e+00 -7.69949779e-02 7.09955215e-01 8.90779555e-01
3.02118242e-01 9.18236613e-01 9.22820508e-01 -2.51606017e-01
-9.85332280e-02 -1.23883776e-01 1.82247981e-02 7.96939194e-01
2.68230122e-02 1.84590057e-01 -5.72209895e-01 5.97939372e-01
1.33604586e+00 2.48621032e-01 -6.22172773e-01 -1.49119765e-01
-1.06654823e+00 7.39778757e-01 9.45336640e-01 4.43309136e-02
-1.64695591e-01 2.83757508e-01 4.65515591e-02 1.41732851e-02
5.71786046e-01 5.58118448e-02 -3.56846333e-01 -2.05718186e-02
-5.59518516e-01 2.79723350e-02 2.40100920e-01 9.62769568e-01
1.06906092e+00 -2.68858969e-01 -1.15774579e-01 7.08361149e-01
5.03546476e-01 5.29207170e-01 6.41910613e-01 -8.31984043e-01
6.29893661e-01 5.08817255e-01 -1.15708061e-01 -7.28096008e-01
-3.32093209e-01 -2.34662682e-01 -8.85575771e-01 4.41569716e-01
2.11089045e-01 -1.42432889e-03 -1.02268636e+00 1.55896974e+00
1.61234021e-01 2.42427260e-01 4.98908646e-02 1.30033267e+00
1.02254140e+00 9.53934729e-01 -4.52255189e-01 -1.45049915e-01
1.11953151e+00 -1.30280375e+00 -6.13696516e-01 -6.03769243e-01
3.53843510e-01 -4.83488530e-01 1.11013556e+00 3.36317569e-01
-1.30396378e+00 -7.09864497e-01 -1.13460350e+00 -9.18998837e-01
-2.16508731e-01 5.22806421e-02 7.39162087e-01 9.73824859e-02
-1.13614261e+00 4.89501446e-01 -8.91628206e-01 -2.02022478e-01
6.24948978e-01 4.12728071e-01 -2.45185405e-01 -4.40271854e-01
-9.72090065e-01 6.77276790e-01 2.38137335e-01 4.54984695e-01
-1.05885851e+00 -6.12492383e-01 -1.04157233e+00 -4.94147092e-02
1.57569945e-01 -9.68330860e-01 1.13633120e+00 -5.85520804e-01
-1.55225480e+00 8.44704092e-01 -3.99222702e-01 -1.26169875e-01
3.98941994e-01 -5.11096835e-01 1.46549448e-01 1.60062432e-01
2.72415042e-01 1.02691555e+00 6.97948873e-01 -9.13660586e-01
-7.94776797e-01 -6.54298961e-01 3.12515944e-01 6.36544526e-01
-2.54230022e-01 -1.92029402e-01 -8.47707093e-01 -1.06053196e-01
5.32932937e-01 -4.60181326e-01 -1.61315218e-01 2.83137083e-01
-2.77385086e-01 -4.22009788e-02 6.35186076e-01 -3.65549862e-01
1.08567381e+00 -2.12423873e+00 3.31850857e-01 -3.73570293e-01
4.76576358e-01 -2.67855860e-02 1.04649134e-01 -2.85508186e-02
7.84465149e-02 -1.18028961e-01 -2.35404521e-01 -6.27095580e-01
-4.31672275e-01 1.26119882e-01 -3.08114320e-01 4.60788757e-01
9.69090164e-02 1.06206000e+00 -8.34550023e-01 -4.09870982e-01
5.37075102e-01 5.70083201e-01 -6.68313324e-01 5.36410153e-01
-1.22951053e-01 3.83611470e-01 -4.44376558e-01 6.67985678e-01
1.03426564e+00 -2.95971334e-01 -4.25382614e-01 -3.61100793e-01
-5.22728503e-01 3.38065356e-01 -7.48242319e-01 2.09763122e+00
-5.44391394e-01 7.49797761e-01 -1.11465961e-01 -7.49118865e-01
8.50029290e-01 -1.66132346e-01 2.56682873e-01 -9.19631362e-01
3.07342440e-01 2.89275765e-01 -1.68584749e-01 -7.32018888e-01
2.09288597e-01 3.44993547e-02 2.06819788e-01 1.55114338e-01
4.33699461e-03 -3.25962067e-01 -2.67974138e-01 3.18706818e-02
9.69583333e-01 2.55050182e-01 7.49023408e-02 1.43814057e-01
5.48070073e-01 -4.60555553e-01 4.62236732e-01 4.96550739e-01
-1.01736866e-01 1.01934278e+00 2.67545581e-01 -4.58482027e-01
-9.06160235e-01 -1.03894496e+00 -1.99674666e-01 5.98455667e-01
5.59687495e-01 -1.19214214e-01 -6.74468040e-01 -4.39534426e-01
-1.31533012e-01 1.59152269e-01 -7.82022834e-01 -1.46507472e-01
-4.84980464e-01 -5.48933089e-01 2.11273715e-01 6.24816716e-01
1.09583235e+00 -1.15832424e+00 -7.58160353e-01 -1.78404339e-02
-3.62778217e-01 -1.13671911e+00 -5.85592270e-01 2.66343772e-01
-8.01597297e-01 -8.70828569e-01 -9.23355043e-01 -8.90862465e-01
6.19606733e-01 6.30632043e-01 7.95799255e-01 -5.68561479e-02
-2.11926818e-01 -2.97249784e-03 -1.13547213e-01 -3.66166830e-01
5.65421999e-01 7.92646781e-02 -1.92007378e-01 7.50337448e-03
4.84944105e-01 -7.64567018e-01 -1.07052088e+00 2.06798807e-01
-7.56792188e-01 5.45673370e-01 8.86555433e-01 7.63612628e-01
6.52611971e-01 -2.41707966e-01 2.01930285e-01 -5.27331114e-01
1.04566015e-01 -4.46250409e-01 -6.71903789e-01 -2.18881547e-01
-3.37629884e-01 1.22391596e-01 4.21546549e-01 -2.42716148e-01
-1.10795140e+00 1.08155340e-01 -3.33210647e-01 -7.08818257e-01
6.56535849e-02 2.58073032e-01 -4.38015252e-01 -2.50497878e-01
2.73866445e-01 4.76498961e-01 -7.78040215e-02 -4.22615021e-01
2.53332883e-01 7.41185606e-01 4.95899141e-01 -1.85290515e-01
5.39284647e-01 6.75359488e-01 -6.70479015e-02 -5.54949820e-01
-1.15417027e+00 -3.45372081e-01 -6.96114600e-01 -1.52671799e-01
1.15884089e+00 -1.27426183e+00 -8.20921063e-01 9.26391184e-01
-1.35008371e+00 -5.59678555e-01 7.59799927e-02 6.32673860e-01
-5.70941210e-01 2.40522698e-01 -7.55735099e-01 -5.71534753e-01
-3.29876393e-01 -1.24264455e+00 1.24324965e+00 4.94907320e-01
3.46698731e-01 -7.24976540e-01 -9.53430459e-02 3.98244709e-01
2.41484627e-01 1.39060589e-02 6.13245606e-01 3.79826367e-01
-1.26771772e+00 2.87969172e-01 -7.41376102e-01 2.90418714e-01
-7.62444064e-02 -2.45811179e-01 -1.44930410e+00 -2.08849370e-01
2.42658496e-01 -3.73990059e-01 1.21035504e+00 8.22760165e-01
1.46533895e+00 -2.39458354e-03 -3.18177700e-01 1.38650823e+00
1.56210542e+00 2.72240847e-01 1.02339876e+00 3.47196817e-01
1.17221820e+00 4.70208585e-01 6.58991039e-01 4.11387533e-01
7.17893422e-01 5.65196693e-01 7.44060099e-01 -1.80212677e-01
-2.41608396e-01 -4.33824390e-01 3.31114829e-01 8.18388522e-01
-8.40286016e-02 -8.21827278e-02 -6.75878346e-01 4.82087731e-01
-1.78948665e+00 -8.21508288e-01 -1.54482216e-01 2.08209157e+00
6.37251318e-01 2.74420530e-01 -4.08266336e-01 -8.65625665e-02
3.93996775e-01 2.81146586e-01 -9.06177104e-01 -2.34364375e-01
-2.01785222e-01 5.87921776e-02 5.63137114e-01 7.95125186e-01
-8.09286356e-01 9.12871301e-01 5.75085354e+00 6.83192551e-01
-1.24357736e+00 7.99220279e-02 6.68534935e-01 -4.53478605e-01
-2.37699926e-01 -1.29081085e-01 -9.38338459e-01 5.59435487e-01
5.25208235e-01 1.86261237e-02 4.27975029e-01 7.47441530e-01
2.45430440e-01 -3.59309107e-01 -1.32678723e+00 1.52349210e+00
2.59166270e-01 -1.10551047e+00 3.60887945e-02 2.18061417e-01
6.67542517e-01 3.28076422e-01 2.14985669e-01 1.75726950e-01
-1.64689153e-01 -1.14236307e+00 6.56513631e-01 6.30589426e-01
1.03197610e+00 -6.81513548e-01 6.82136476e-01 4.71754789e-01
-1.28686750e+00 -4.18143839e-01 -8.18675697e-01 -5.32705903e-01
9.94184986e-02 6.49427056e-01 -3.53522986e-01 3.77539366e-01
1.01125169e+00 1.18260682e+00 -4.51410443e-01 1.01142287e+00
-2.60910809e-01 -9.95267183e-02 -1.92331597e-01 2.93475036e-02
1.99171349e-01 -2.84394354e-01 5.66257834e-02 7.03889489e-01
5.76613903e-01 4.44720894e-01 -1.77631989e-01 1.02940774e+00
-9.81171131e-02 -1.66247964e-01 -7.46378899e-01 5.40889680e-01
3.07083100e-01 1.08160627e+00 -2.70132422e-01 -2.33469084e-01
-6.43912375e-01 1.32782066e+00 6.93265140e-01 3.36353511e-01
-9.71866131e-01 -4.92894292e-01 7.85148859e-01 2.92376429e-01
2.46455893e-01 -9.78186578e-02 -4.75608319e-01 -1.30026400e+00
3.23172361e-01 -3.36834222e-01 7.49084651e-02 -1.45643616e+00
-8.71876717e-01 6.53697968e-01 -2.44730562e-01 -1.27131259e+00
4.57939245e-02 -6.89527690e-01 -5.86131454e-01 1.10072410e+00
-1.92170250e+00 -9.99567628e-01 -9.85661328e-01 7.90714383e-01
7.20733404e-01 2.52894700e-01 4.31363314e-01 4.55700457e-01
-7.33150661e-01 4.60849077e-01 -1.82793990e-01 1.00881997e-02
6.70421004e-01 -9.93090928e-01 4.02289897e-01 7.58239806e-01
-1.98768020e-01 3.35359722e-01 1.62295997e-01 -4.28967416e-01
-1.37193739e+00 -1.01392424e+00 7.87459671e-01 -5.62095523e-01
2.39432380e-01 -5.64216614e-01 -8.47586334e-01 7.54363060e-01
9.17827263e-02 2.78082728e-01 3.18916515e-03 -3.04978132e-01
-1.70572251e-01 -3.50195616e-01 -8.43210459e-01 4.44613516e-01
1.43391955e+00 -6.80971324e-01 -3.20963174e-01 1.41495049e-01
1.05919433e+00 -6.84145987e-01 -5.35340667e-01 2.75168777e-01
5.72101295e-01 -1.48063922e+00 1.01988101e+00 3.53763580e-01
9.72617924e-01 -4.14577782e-01 -1.16731353e-01 -1.00584745e+00
-3.94864291e-01 -2.16247603e-01 -1.59383833e-01 9.38282430e-01
1.39131621e-01 -7.18843699e-01 7.87464738e-01 4.82664466e-01
-4.61936712e-01 -1.10841334e+00 -8.55173290e-01 -3.00512433e-01
6.02160916e-02 -3.90847296e-01 5.03489852e-01 4.43306684e-01
-2.24051207e-01 5.34672737e-01 -4.34936076e-01 3.79336298e-01
6.12384081e-01 2.06261091e-02 7.13515639e-01 -9.90279555e-01
-2.90773273e-01 -3.73020679e-01 -4.20201838e-01 -1.96255231e+00
-6.11100458e-02 -5.67839563e-01 1.93320215e-01 -1.73495424e+00
3.83589774e-01 -2.46876329e-01 -1.14618272e-01 2.08593056e-01
-2.25872293e-01 2.57359862e-01 -1.49827544e-02 2.78413534e-01
-4.86310214e-01 8.32948983e-01 1.60919440e+00 3.56229916e-02
-1.18929997e-01 -2.31746212e-01 -6.37740493e-01 7.84942031e-01
4.74147141e-01 -1.51571155e-01 -6.39899254e-01 -1.03770697e+00
1.03758924e-01 1.89273506e-01 4.32599843e-01 -1.19516528e+00
5.05330563e-01 -4.88786474e-02 7.18373358e-01 -8.91652703e-01
7.06646383e-01 -6.87347591e-01 -2.60355771e-01 2.64339238e-01
-4.09643538e-02 -1.74670175e-01 1.36113716e-02 3.12011033e-01
-3.37111801e-01 -6.88307360e-03 8.00213337e-01 -2.50391990e-01
-1.02254593e+00 8.59945297e-01 2.07778424e-01 -1.27723098e-01
8.91976953e-01 -4.64123368e-01 -4.17735636e-01 -3.30865443e-01
-3.22398156e-01 3.33281368e-01 4.77590829e-01 3.45645458e-01
1.08966041e+00 -1.37897944e+00 -5.62216997e-01 6.64509058e-01
2.49368306e-02 7.07357287e-01 5.87890625e-01 7.28404880e-01
-6.44120932e-01 4.55247372e-01 -3.89869988e-01 -8.35610151e-01
-9.06134486e-01 5.00996649e-01 5.77879488e-01 1.58739351e-02
-8.31681490e-01 1.25377524e+00 1.03051078e+00 -1.15440905e-01
1.89895600e-01 -5.80864787e-01 -1.64905608e-01 -1.66562930e-01
8.23561847e-01 -1.55833671e-02 -1.44824386e-01 -3.98848504e-01
-3.29359412e-01 1.31704223e+00 -1.45387337e-01 -1.00064136e-01
1.28357935e+00 -5.72032273e-01 4.38336225e-04 4.52156186e-01
1.52748764e+00 -3.60673815e-01 -1.86586094e+00 -2.14083388e-01
-7.47272193e-01 -9.53068256e-01 3.71839255e-01 -4.61957663e-01
-1.24287677e+00 1.32487535e+00 5.61781406e-01 -2.24346116e-01
1.40608418e+00 -3.53069380e-02 8.16189289e-01 1.17785662e-01
3.51884812e-01 -6.98733866e-01 3.17089587e-01 6.91757679e-01
9.62200046e-01 -1.60719812e+00 -1.64359242e-01 -3.98091465e-01
-3.79966915e-01 9.85672772e-01 1.21773386e+00 -1.74410582e-01
5.87374926e-01 3.78051549e-01 -9.58799093e-04 -1.59348235e-01
-7.66708016e-01 -2.76509374e-01 -6.54939413e-02 6.47467911e-01
3.39625657e-01 -2.99714357e-01 1.25011072e-01 5.99484146e-01
-9.88042578e-02 7.86908641e-02 5.40494144e-01 4.88490820e-01
-5.50057352e-01 -6.47315979e-01 -9.03058499e-02 3.00625235e-01
-1.57705620e-01 -4.43718821e-01 7.28709251e-02 5.40969253e-01
3.23663592e-01 7.42586672e-01 3.94676149e-01 -4.72396582e-01
3.16070229e-01 -3.44394565e-01 5.44834733e-01 -6.24471843e-01
-5.06653525e-02 4.87102866e-02 -3.65530223e-01 -9.02779937e-01
-1.88348144e-01 -3.86587799e-01 -1.26359189e+00 -3.74663949e-01
-1.73741162e-01 -3.78247082e-01 5.07520258e-01 8.26377809e-01
2.96299964e-01 6.89659536e-01 8.40653539e-01 -1.15909445e+00
-1.51180267e-01 -1.06775570e+00 -4.92298841e-01 8.63370448e-02
6.50573254e-01 -7.10453391e-01 -4.81644988e-01 -2.56787479e-01] | [9.074688911437988, -2.4780542850494385] |
0f678ce7-61c8-4315-8d59-e20b301e7d60 | mets-cov-a-dataset-of-medical-entity-and | 2209.13773 | null | https://arxiv.org/abs/2209.13773v1 | https://arxiv.org/pdf/2209.13773v1.pdf | METS-CoV: A Dataset of Medical Entity and Targeted Sentiment on COVID-19 Related Tweets | The COVID-19 pandemic continues to bring up various topics discussed or debated on social media. In order to explore the impact of pandemics on people's lives, it is crucial to understand the public's concerns and attitudes towards pandemic-related entities (e.g., drugs, vaccines) on social media. However, models trained on existing named entity recognition (NER) or targeted sentiment analysis (TSA) datasets have limited ability to understand COVID-19-related social media texts because these datasets are not designed or annotated from a medical perspective. This paper releases METS-CoV, a dataset containing medical entities and targeted sentiments from COVID-19-related tweets. METS-CoV contains 10,000 tweets with 7 types of entities, including 4 medical entity types (Disease, Drug, Symptom, and Vaccine) and 3 general entity types (Person, Location, and Organization). To further investigate tweet users' attitudes toward specific entities, 4 types of entities (Person, Organization, Drug, and Vaccine) are selected and annotated with user sentiments, resulting in a targeted sentiment dataset with 9,101 entities (in 5,278 tweets). To the best of our knowledge, METS-CoV is the first dataset to collect medical entities and corresponding sentiments of COVID-19-related tweets. We benchmark the performance of classical machine learning models and state-of-the-art deep learning models on NER and TSA tasks with extensive experiments. Results show that the dataset has vast room for improvement for both NER and TSA tasks. METS-CoV is an important resource for developing better medical social media tools and facilitating computational social science research, especially in epidemiology. Our data, annotation guidelines, benchmark models, and source code are publicly available (https://github.com/YLab-Open/METS-CoV) to ensure reproducibility. | ['Jie Yang', 'Jiageng Wu', 'Zhiyang Teng', 'Zichang Su', 'Yining Hua', 'Zhijiang Guo', 'Dading Chong', 'Zeqiang Wang', 'Peilin Zhou'] | 2022-09-28 | null | null | null | null | ['epidemiology'] | ['medical'] | [-2.48689920e-01 6.21410040e-03 -4.44238365e-01 -2.31777608e-01
-4.31271344e-01 -5.22627711e-01 5.79251111e-01 1.10615575e+00
-8.35954309e-01 8.63731205e-01 6.29549682e-01 -3.34498465e-01
2.91926503e-01 -1.07939816e+00 -4.67708498e-01 -4.26602244e-01
-2.69549061e-02 7.98620462e-01 -1.26224682e-01 -5.42917728e-01
-6.02659099e-02 2.64540851e-01 -9.09708321e-01 2.61086881e-01
9.00816560e-01 5.33059835e-01 -7.28055760e-02 5.92790127e-01
-1.89640597e-01 8.47139120e-01 -6.80764318e-01 -4.78113115e-01
-3.36875528e-01 -1.46087110e-01 -8.31688285e-01 -6.58126414e-01
-2.83922464e-01 -8.98892432e-02 4.03944775e-02 8.78612041e-01
6.48088276e-01 -3.31129372e-01 8.22342634e-01 -1.06244063e+00
-7.37388015e-01 5.80489814e-01 -1.40370056e-01 4.45930749e-01
4.43780392e-01 -2.02407315e-01 9.12018478e-01 -7.23200858e-01
1.22773933e+00 7.41878450e-01 1.07028782e+00 5.53843200e-01
-4.56973732e-01 -7.56027699e-01 -2.92318851e-01 -2.34059483e-01
-1.47415769e+00 -1.00771628e-01 1.10257737e-01 -9.51145709e-01
1.06545115e+00 3.82044971e-01 7.22012579e-01 1.26432514e+00
3.98450047e-01 4.82960135e-01 9.74153578e-01 1.77754834e-01
6.74403831e-02 5.70733786e-01 2.99475700e-01 5.86282074e-01
6.06464028e-01 -2.98059285e-01 -4.28732708e-02 -6.72866404e-01
3.28009836e-02 5.59353471e-01 -2.32444867e-01 5.24625897e-01
-1.30314457e+00 1.02253211e+00 5.08371413e-01 5.56924105e-01
-8.12273324e-01 -5.65175116e-01 6.41713142e-01 1.64865945e-02
1.01589727e+00 8.18641484e-01 -9.82425928e-01 6.26914650e-02
-6.02508903e-01 2.12744981e-01 9.86315370e-01 6.40123963e-01
5.18491149e-01 -3.40609312e-01 -2.43482769e-01 7.33207405e-01
1.19348697e-01 1.06072199e+00 3.95348430e-01 1.72990467e-03
3.89296353e-01 1.00616837e+00 1.66085079e-01 -1.45243752e+00
-1.19250143e+00 -5.12192726e-01 -1.01246035e+00 -1.07346439e+00
-1.29897371e-01 -9.19327915e-01 -8.73086452e-01 1.70516789e+00
6.31288469e-01 5.04784226e-01 2.93541670e-01 4.93696988e-01
1.78317678e+00 6.64317906e-01 4.97720718e-01 -1.18691519e-01
1.96630943e+00 -4.08091873e-01 -8.35916162e-01 8.16073641e-02
1.07172322e+00 -7.74397194e-01 3.64913940e-01 -3.83356094e-01
-6.03040636e-01 -3.88197526e-02 -3.49227756e-01 1.56762347e-01
-1.43048179e+00 -1.64094150e-01 3.82455140e-01 5.11682272e-01
-8.00390482e-01 7.14230025e-03 -6.30743861e-01 -7.68576801e-01
4.33622986e-01 1.96651921e-01 -3.00036222e-01 3.64364952e-01
-1.91863298e+00 9.55718160e-01 1.65496066e-01 -9.99622270e-02
-5.70971966e-01 -9.67708230e-01 -9.90126848e-01 -2.55341172e-01
1.01802282e-01 -8.92194450e-01 9.71392572e-01 -2.14530736e-01
-6.05088711e-01 1.08396173e+00 -2.08706692e-01 -3.34572852e-01
-6.20003566e-02 2.81923730e-02 -1.07133996e+00 -6.73325956e-02
5.31533480e-01 4.76001561e-01 6.11662231e-02 -6.94133759e-01
-6.92312479e-01 -2.24192187e-01 -9.30983052e-02 -1.54436991e-01
-6.37245238e-01 5.06981075e-01 -1.13809422e-01 -6.40431643e-01
-8.30070555e-01 -1.16776967e+00 -4.12434250e-01 -1.03203154e+00
-7.79239953e-01 -4.19018000e-01 3.19182634e-01 -6.48844898e-01
1.45987606e+00 -1.81124139e+00 -2.82658815e-01 5.03927730e-02
6.34447515e-01 5.23889780e-01 6.90213665e-02 8.21923614e-01
-4.43806387e-02 6.91568017e-01 1.07864499e-01 -1.48964942e-01
-3.61091048e-01 3.79183777e-02 -2.29124159e-01 4.49487358e-01
2.69745350e-01 1.15667474e+00 -1.21736395e+00 -4.07958269e-01
-1.96617916e-01 8.14001322e-01 -5.54521620e-01 1.53097764e-01
-1.44360349e-01 5.21405101e-01 -1.05431712e+00 7.58354247e-01
5.24603069e-01 -8.39775801e-01 2.47236639e-02 -3.05404365e-01
-2.36707330e-01 3.85140270e-01 -3.13106149e-01 6.70073926e-01
-3.25256586e-01 5.28266549e-01 -2.69014150e-01 -4.35437471e-01
5.60142517e-01 6.62342072e-01 9.75089848e-01 -2.67615259e-01
6.20824516e-01 1.95540249e-01 -2.90474981e-01 -6.96924388e-01
7.33073056e-01 3.57387774e-02 -5.55280030e-01 3.54876637e-01
-3.32729757e-01 4.02136773e-01 1.05798326e-01 3.05951566e-01
1.08296621e+00 -8.13986182e-01 5.70033669e-01 -1.52979463e-01
3.29898506e-01 5.72752595e-01 6.21339679e-01 3.85891795e-01
-1.52277321e-01 1.48130342e-01 3.25150371e-01 -4.85609800e-01
-7.61579096e-01 -5.52252948e-01 -5.38991392e-01 1.05961180e+00
-1.06684759e-01 -5.62364876e-01 -6.80164337e-01 -5.73880017e-01
-1.47099093e-01 4.14290458e-01 -9.31987524e-01 3.75212520e-01
-4.12200511e-01 -1.34214163e+00 8.45114410e-01 1.43685535e-01
2.98330545e-01 -1.20777202e+00 -3.27941984e-01 2.84888566e-01
-6.73239231e-01 -1.29493749e+00 -5.93987703e-01 -2.15029150e-01
-1.83953419e-01 -1.41251111e+00 -7.54103065e-01 -8.65004122e-01
8.14628780e-01 1.81298237e-02 1.21376836e+00 4.27881517e-02
-1.54986337e-01 3.31361830e-01 -5.41766763e-01 -1.02411199e+00
-3.38332295e-01 4.69952852e-01 1.42627135e-01 -2.83549994e-01
7.13118792e-01 1.88804224e-01 -7.11525559e-01 2.73294270e-01
-8.92750740e-01 -4.75765392e-02 2.07015097e-01 5.95065832e-01
4.31362182e-01 -2.16588620e-02 8.76328409e-01 -1.52516711e+00
7.05692053e-01 -1.41787446e+00 -5.84905259e-02 4.53005582e-02
-5.13367236e-01 -6.34351552e-01 5.85359931e-01 -1.35924593e-01
-8.39566171e-01 -5.08079827e-01 -5.06812751e-01 3.83828104e-01
-2.26054803e-01 1.15330708e+00 5.74109852e-01 4.63453174e-01
7.48304427e-01 -1.13952570e-01 -4.64380562e-01 -2.70453155e-01
-2.93803331e-03 1.24739563e+00 -1.21590234e-02 1.43997058e-01
5.06340861e-01 4.77446586e-01 -4.21504617e-01 -9.45333958e-01
-1.28737128e+00 -9.89349127e-01 -1.47414850e-02 7.47789741e-02
1.49464989e+00 -1.38974857e+00 -7.71615624e-01 6.02357745e-01
-1.15533042e+00 4.28001657e-02 1.28335357e-01 6.12567425e-01
3.10966671e-01 -1.84474394e-01 -9.74744380e-01 -3.76601756e-01
-8.75330627e-01 -9.20074940e-01 9.81623173e-01 2.92730838e-01
-6.32827520e-01 -1.44577336e+00 6.58640563e-01 5.67635417e-01
5.67830503e-01 6.68770909e-01 6.69630289e-01 -1.36750758e+00
1.49495021e-01 -3.17168623e-01 -1.81659639e-01 -2.85821140e-01
5.36528707e-01 -4.99330945e-02 -5.93045652e-01 -1.68610141e-01
-2.05705434e-01 -2.86827981e-01 7.39837408e-01 3.43004674e-01
6.07612014e-01 -6.05690420e-01 -8.18749011e-01 3.00832629e-01
1.07455707e+00 2.42310420e-01 1.98597133e-01 4.23215419e-01
7.82022476e-01 5.19394219e-01 5.49520135e-01 6.69794440e-01
1.01735044e+00 3.17007601e-01 1.66384608e-01 -5.94139934e-01
5.04594266e-01 -7.73519129e-02 3.00156862e-01 9.81057286e-01
-2.28371307e-01 -6.46523893e-01 -1.34472179e+00 8.24026763e-01
-1.40602398e+00 -9.45998192e-01 -3.35285604e-01 1.67162645e+00
1.04667377e+00 -2.87000835e-01 6.92505315e-02 -6.49059892e-01
8.26627195e-01 1.60299003e-01 -4.70757872e-01 -3.83382082e-01
-2.31482938e-01 1.70037568e-01 5.86241782e-01 2.13739201e-01
-1.37615275e+00 8.47954810e-01 5.54262924e+00 5.47822475e-01
-1.21771657e+00 3.00418645e-01 7.08719730e-01 2.87724704e-01
-4.07151639e-01 -5.81255376e-01 -1.29303253e+00 5.54556549e-01
1.19212306e+00 -9.44668204e-02 -6.63371012e-02 5.78091443e-01
3.03429246e-01 5.15223026e-01 -4.96607006e-01 4.17424440e-01
9.72553343e-02 -1.78656864e+00 -2.54241973e-01 1.93452299e-01
1.02594149e+00 9.25468028e-01 1.20620616e-01 4.12700683e-01
3.38256299e-01 -9.43448961e-01 1.55302119e-02 5.10089338e-01
8.85976017e-01 -4.24953490e-01 1.39108491e+00 1.90432429e-01
-1.16794479e+00 3.31735075e-01 -1.98790327e-01 4.38236862e-01
3.45106423e-01 7.37323582e-01 -1.35276473e+00 5.22302508e-01
7.90234208e-01 1.04256213e+00 -3.28443944e-01 5.86889744e-01
1.05533324e-01 7.65852988e-01 -2.09977746e-01 -4.40255582e-01
4.45744276e-01 2.14495435e-01 3.85667145e-01 1.69470167e+00
3.93861234e-01 2.96151400e-01 2.09817126e-01 4.26089585e-01
-4.72257674e-01 6.11627936e-01 -6.63233936e-01 -7.14887679e-01
4.25496250e-01 1.45080662e+00 -7.61484623e-01 -5.65872490e-01
-5.05212307e-01 3.36280406e-01 5.72956987e-02 2.72869796e-01
-1.02280569e+00 -3.70168447e-01 7.05595791e-01 3.81805331e-01
1.12964004e-01 2.41045579e-01 2.64134496e-01 -1.24801505e+00
-7.55573988e-01 -7.49438584e-01 7.19336450e-01 -4.18088526e-01
-1.64632487e+00 9.82713819e-01 -2.64599323e-01 -1.16418648e+00
-1.75090283e-01 -4.59217936e-01 -3.90349448e-01 5.59926093e-01
-1.57810807e+00 -1.22656024e+00 -1.76042184e-01 5.52859604e-01
9.05604511e-02 -1.91477433e-01 1.11680067e+00 5.16935468e-01
-7.57604718e-01 4.04985785e-01 1.98603123e-01 6.16522253e-01
8.85959685e-01 -9.17784333e-01 3.96206796e-01 -5.85147403e-02
-4.85259980e-01 9.70585585e-01 5.76787591e-01 -1.09852064e+00
-8.55352342e-01 -1.67563820e+00 1.71715438e+00 -8.60664010e-01
1.01033962e+00 -1.55461088e-01 -6.47671640e-01 8.33155334e-01
5.00885010e-01 -4.55834538e-01 1.62035322e+00 2.99103767e-01
-1.22600205e-01 3.93650323e-01 -1.29520094e+00 6.06071115e-01
6.09274149e-01 -5.27297139e-01 -5.86789489e-01 9.69403565e-01
9.54604447e-01 -4.31783170e-01 -1.37679017e+00 4.90505636e-01
3.07926774e-01 -4.42089587e-01 9.82177913e-01 -1.01273715e+00
4.79172975e-01 -6.75185397e-02 1.41779810e-01 -1.39687610e+00
-1.92893401e-01 -3.28901738e-01 2.61219978e-01 1.01406503e+00
9.65835154e-01 -1.11449385e+00 4.74472165e-01 3.23013097e-01
-7.41454884e-02 -9.94457245e-01 -4.32095349e-01 -1.56117722e-01
-8.61128941e-02 -1.87355563e-01 8.44191134e-01 1.73170936e+00
1.80603594e-01 4.37222362e-01 -4.66442138e-01 1.91068053e-01
-1.00872926e-01 1.40857354e-01 5.70118845e-01 -1.21119261e+00
2.02091083e-01 -2.72630572e-01 -3.28402184e-02 -4.14310098e-01
4.04316671e-02 -8.93324912e-01 -3.90056074e-01 -1.78797841e+00
4.53995198e-01 -6.60507262e-01 -5.13854623e-01 7.39182472e-01
-2.57310063e-01 3.61572832e-01 -2.46847928e-01 2.64218211e-01
-7.19475150e-01 1.99741259e-01 1.06491661e+00 -3.34266126e-01
-3.53976727e-01 -2.81670131e-02 -9.25790131e-01 8.02927673e-01
9.19544697e-01 -7.63600826e-01 3.71787022e-03 -1.81520194e-01
8.76247406e-01 -1.78262889e-01 1.44716939e-02 -2.56930530e-01
3.63680482e-01 -2.23781407e-01 2.29456946e-01 -8.14670205e-01
1.12764955e-01 -5.02867937e-01 3.38008374e-01 7.86403120e-01
-2.08087683e-01 2.13304818e-01 1.65304765e-01 3.25867742e-01
-2.74276286e-01 1.79515421e-01 2.51260310e-01 -1.90504137e-02
-3.04947674e-01 6.66458786e-01 -7.84503460e-01 4.94982690e-01
9.48900759e-01 2.45093971e-01 -9.47025359e-01 -2.26678804e-01
-7.19535291e-01 5.35087645e-01 2.85609663e-01 3.99724960e-01
4.06024575e-01 -1.00442779e+00 -1.04946065e+00 -1.25976101e-01
4.07643199e-01 -1.24653243e-01 6.00522816e-01 1.08165216e+00
-5.54882050e-01 7.41014242e-01 7.88629502e-02 -2.98569411e-01
-1.17077196e+00 5.44741452e-01 9.90404487e-02 -6.61837518e-01
-4.65947911e-02 6.55658841e-01 2.48268336e-01 -9.19220269e-01
-3.35027575e-01 -2.45074838e-01 -9.05282319e-01 5.65910459e-01
5.87746203e-01 3.65192711e-01 -1.57555178e-01 -1.09032309e+00
-7.83224702e-01 4.95391428e-01 -1.77706212e-01 6.51966691e-01
1.58103240e+00 7.45531246e-02 -4.74606454e-01 1.69993788e-01
1.29590654e+00 4.48152453e-01 1.02202080e-01 -1.40728340e-01
-1.15583189e-01 1.65847957e-01 -2.20389426e-01 -8.01258981e-01
-9.77179766e-01 3.27788323e-01 2.86172271e-01 5.61513484e-01
9.24371779e-01 3.01388681e-01 1.29684961e+00 2.69443214e-01
6.57402053e-02 -8.26611876e-01 -3.06548864e-01 7.68874586e-01
5.85422158e-01 -1.34716988e+00 -1.61494881e-01 -3.73890132e-01
-7.41380990e-01 5.11921823e-01 2.82812893e-01 4.88317490e-01
1.27360177e+00 2.32561633e-01 2.87703902e-01 -7.91164517e-01
-6.19185567e-01 -2.27636382e-01 3.93363357e-01 2.28515834e-01
7.62058079e-01 5.03378868e-01 -3.69193763e-01 8.39424908e-01
-3.17015588e-01 -9.81930941e-02 5.61105132e-01 6.51445031e-01
-2.21426159e-01 -9.52248931e-01 -1.09955423e-01 8.49090397e-01
-1.10658574e+00 -5.27693987e-01 -4.47713703e-01 5.90337694e-01
5.51442564e-01 1.17477524e+00 -9.93974134e-02 -4.27942574e-01
2.87537038e-01 -3.19287211e-01 -4.88527000e-01 -8.99059057e-01
-1.02381599e+00 -1.34320438e-01 4.57879126e-01 -6.61743954e-02
-8.67013335e-01 -5.43276370e-01 -1.52249610e+00 -5.36113262e-01
-2.85186648e-01 4.97329116e-01 4.95302230e-01 8.09608102e-01
8.06463480e-01 4.61490929e-01 5.51827312e-01 -4.13303897e-02
3.68815601e-01 -8.78755629e-01 -2.49872521e-01 5.17699480e-01
3.30636352e-01 -3.82481396e-01 -2.39432365e-01 -2.10534751e-01] | [8.455174446105957, 9.264328956604004] |
9be45d5d-80ce-44c5-83e7-d8ed87b08015 | a-motion-assessment-method-for-reference | 2306.17434 | null | https://arxiv.org/abs/2306.17434v1 | https://arxiv.org/pdf/2306.17434v1.pdf | A Motion Assessment Method for Reference Stack Selection in Fetal Brain MRI Reconstruction Based on Tensor Rank Approximation | Purpose: Slice-to-volume registration and super-resolution reconstruction (SVR-SRR) is commonly used to generate 3D volumes of the fetal brain from 2D stacks of slices acquired in multiple orientations. A critical initial step in this pipeline is to select one stack with the minimum motion as a reference for registration. An accurate and unbiased motion assessment (MA) is thus crucial for successful selection. Methods: We presented a MA method that determines the minimum motion stack based on 3D low-rank approximation using CANDECOMP/PARAFAC (CP) decomposition. Compared to the current 2D singular value decomposition (SVD) based method that requires flattening stacks into matrices to obtain ranks, in which the spatial information is lost, the CP-based method can factorize 3D stack into low-rank and sparse components in a computationally efficient manner. The difference between the original stack and its low-rank approximation was proposed as the motion indicator. Results: Compared to SVD-based methods, our proposed CP-based MA demonstrated higher sensitivity in detecting small motion with a lower baseline bias. Experiments on randomly simulated motion illustrated that the proposed CP method achieved a higher success rate of 95.45% in identifying the minimum motion stack, compared to SVD-based method with a success rate of 58.18%. We further demonstrated that combining CP-based MA with existing SRR-SVR pipeline significantly improved 3D volume reconstruction. Conclusion: The proposed CP-based MA method showed superior performance compared to SVD-based methods with higher sensitivity to motion, success rate, and lower baseline bias, and can be used as a prior step to improve fetal brain reconstruction. | ['Dan Wu', 'Guangbin Wang', 'Sun Yi', 'Cong Sun', 'Tianshu Zheng', 'Jiwei Sun', 'Wen Shi', 'Haoan Xu'] | 2023-06-30 | null | null | null | null | ['super-resolution', 'mri-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 6.75399303e-02 -3.42611521e-01 1.83679745e-01 -2.26763964e-01
-8.77394319e-01 -4.66106325e-01 1.73207045e-01 1.07261799e-01
-3.52499902e-01 2.75792927e-01 4.24462885e-01 -1.35202389e-02
-3.08123827e-01 -5.32900691e-01 -3.77134562e-01 -8.51255298e-01
-5.01423478e-01 4.13706899e-01 4.18126583e-01 -5.05564660e-02
2.44032577e-01 7.25269675e-01 -1.08685923e+00 3.21615785e-01
7.13738799e-01 5.04781723e-01 5.59765995e-01 5.14669240e-01
-9.44969654e-02 3.88280332e-01 -2.32689589e-01 1.73441932e-01
2.74004728e-01 -4.27967757e-01 -7.68612325e-01 -2.80910820e-01
5.22868395e-01 -4.28415060e-01 5.09575307e-02 7.08800256e-01
8.16604972e-01 2.72680730e-01 6.51552439e-01 -5.54455638e-01
-7.80823082e-02 3.56264949e-01 -1.04603863e+00 6.80520475e-01
5.24373591e-01 -3.44606549e-01 3.28124046e-01 -1.58337498e+00
8.62725139e-01 7.87058532e-01 8.23569775e-01 5.65938532e-01
-1.38164353e+00 -7.90966511e-01 -4.11867082e-01 -7.68902972e-02
-1.52324009e+00 -3.54700178e-01 7.35533178e-01 -8.62999439e-01
7.22295463e-01 5.06462991e-01 9.67874706e-01 4.31252897e-01
2.78051257e-01 1.97565109e-01 1.04531217e+00 -1.06562428e-01
-5.64350896e-02 -4.53965336e-01 -2.18614098e-02 7.67434299e-01
4.90691215e-01 -2.33418062e-01 -6.12743258e-01 -4.05865312e-01
1.22213745e+00 -3.35746080e-01 -4.38633829e-01 -2.76783228e-01
-1.68419671e+00 5.66046357e-01 1.61661148e-01 8.36703956e-01
-6.24925375e-01 -1.46986946e-01 1.47403643e-01 -2.78761029e-01
3.07390213e-01 3.64186168e-01 7.14626089e-02 -2.95821335e-02
-1.58677328e+00 4.99458313e-02 3.74400169e-01 5.74054122e-01
3.18277150e-01 2.21783370e-01 -3.21368575e-01 9.26177084e-01
5.48514068e-01 4.48284745e-01 6.73201621e-01 -1.09130979e+00
3.72647613e-01 3.15949053e-01 -1.73658937e-01 -1.25772548e+00
-8.75581145e-01 -4.57347721e-01 -1.13051713e+00 -3.00864913e-02
2.95915306e-01 1.77209690e-01 -9.43712652e-01 1.41355860e+00
5.82664728e-01 6.51568234e-01 -1.00131474e-01 1.20793533e+00
1.24109924e+00 5.09022713e-01 -3.99345577e-01 -8.08240414e-01
1.30460310e+00 -3.31113189e-01 -8.05405498e-01 2.57624716e-01
4.61668134e-01 -9.60110962e-01 5.96762538e-01 2.46894747e-01
-1.26151013e+00 -1.22331232e-01 -1.05373490e+00 2.83432335e-01
5.90257525e-01 1.19408116e-01 4.38934684e-01 6.50973201e-01
-1.32456255e+00 7.28328824e-01 -1.30908978e+00 -2.25571748e-02
1.71397790e-01 3.71210665e-01 -8.20660889e-01 -4.72020023e-02
-6.70284152e-01 9.75612044e-01 -1.21272936e-01 3.86397213e-01
-7.61694968e-01 -1.08316207e+00 -8.44859123e-01 -3.63531798e-01
-2.99394727e-01 -5.90186179e-01 7.25401759e-01 1.82477571e-02
-1.33550203e+00 8.30445647e-01 -6.36461973e-01 1.74604952e-02
4.80539620e-01 1.07995220e-01 -1.96507558e-01 7.09532738e-01
3.26933295e-01 1.85155943e-01 7.19954193e-01 -1.22600973e+00
4.37644981e-02 -3.73540521e-01 -7.20320463e-01 1.41125917e-01
-2.51153857e-02 2.44997188e-01 -4.30295974e-01 -7.35909402e-01
1.35232890e+00 -9.85683918e-01 -3.60904098e-01 -1.98028520e-01
2.47559354e-01 3.11050266e-01 2.90446043e-01 -1.20769775e+00
1.33661890e+00 -1.81640553e+00 1.01307146e-01 4.95236725e-01
5.87634325e-01 2.32419938e-01 -2.75561549e-02 7.16067031e-02
-3.89016300e-01 8.51469561e-02 -3.17331254e-01 -1.40822038e-01
-9.40724730e-01 -2.66846836e-01 2.63763368e-01 9.88399088e-01
3.75251323e-02 4.51287061e-01 -1.21096861e+00 -7.64871597e-01
1.50827125e-01 9.13160503e-01 -6.97760820e-01 1.40703574e-01
8.33920240e-01 1.04108393e+00 -2.26761147e-01 4.01470691e-01
9.44918692e-01 -2.67462265e-02 3.33051443e-01 -6.19390666e-01
-3.24409842e-01 -2.37280596e-03 -1.41589582e+00 1.80600965e+00
-2.27263257e-01 5.95812857e-01 1.62960485e-01 -4.84531552e-01
1.15401614e+00 7.23304391e-01 1.11084259e+00 -3.37519318e-01
-8.55526328e-02 5.19759178e-01 7.88285807e-02 -6.57216132e-01
2.11450785e-01 -3.07459354e-01 6.44273043e-01 1.75635964e-01
-5.03470078e-02 -2.30662465e-01 1.37818813e-01 1.27896577e-01
1.00054395e+00 2.59010345e-01 1.66449383e-01 -5.60424268e-01
7.85745800e-01 -2.62195319e-01 9.44984734e-01 3.35701942e-01
-4.41192426e-02 1.35261905e+00 2.57298976e-01 -4.91064459e-01
-6.49536371e-01 -9.50759530e-01 -5.77500880e-01 2.46160075e-01
1.29361097e-02 -2.45565027e-01 -5.62340677e-01 -1.95261985e-01
-3.42553705e-01 3.23912024e-01 -5.37386835e-01 3.63692164e-01
-1.09277189e+00 -1.05856061e+00 3.03130895e-01 2.42157921e-01
3.85125615e-02 -4.33973521e-01 -6.24376655e-01 2.25713849e-01
-5.79198658e-01 -9.64559555e-01 -6.61474228e-01 -4.94380265e-01
-1.30453348e+00 -8.35843384e-01 -1.18322623e+00 -7.76468873e-01
1.12921965e+00 4.89196181e-01 6.74976349e-01 1.25641793e-01
-2.52370954e-01 1.95238441e-01 -2.77650476e-01 1.62904873e-01
-3.66656661e-01 -6.01704657e-01 3.87236536e-01 9.92423743e-02
-3.99330080e-01 -7.87514925e-01 -1.10429680e+00 3.42712015e-01
-7.93038070e-01 1.89847067e-01 3.06070447e-01 7.64671326e-01
7.34393597e-01 -4.28120941e-01 2.86260188e-01 -5.23808241e-01
3.76534134e-01 -3.38864833e-01 -5.75493753e-01 4.60640714e-02
-5.37121832e-01 9.01492387e-02 7.61773139e-02 -3.84779006e-01
-9.38194752e-01 9.94490758e-02 -1.70879409e-01 -6.30505204e-01
3.37868065e-01 5.50173283e-01 4.03471857e-01 -4.96282220e-01
6.55834496e-01 2.31922999e-01 4.19604599e-01 -5.04558682e-01
-2.46228337e-01 1.57084405e-01 4.66513306e-01 -3.64276826e-01
7.00768948e-01 6.36900365e-01 6.67620361e-01 -1.02688897e+00
-2.08869517e-01 -8.06235611e-01 -1.03136122e+00 -4.65126872e-01
9.59879518e-01 -7.98916638e-01 -5.20193994e-01 2.53422111e-01
-1.22664356e+00 1.85921893e-01 3.08956325e-01 1.21704102e+00
-2.12787122e-01 7.37293005e-01 -5.91895461e-01 -7.17115521e-01
-7.84294248e-01 -1.44237614e+00 7.07946062e-01 -5.67687452e-02
-3.19217831e-01 -7.33639121e-01 3.02567840e-01 6.08249962e-01
5.05995512e-01 7.32377589e-01 5.95696032e-01 -3.09481084e-01
-3.51471663e-01 -2.18296230e-01 5.85652478e-02 -3.35680358e-02
1.83697879e-01 9.47567225e-02 -6.81918204e-01 -3.63961637e-01
2.14412987e-01 7.16609061e-01 4.03687090e-01 7.46982753e-01
5.05996883e-01 -2.49563307e-02 -8.11779276e-02 8.96243453e-01
1.36562955e+00 2.60044932e-01 4.81275082e-01 2.60549247e-01
7.69226789e-01 5.45445204e-01 7.57589519e-01 5.45084655e-01
3.28222513e-01 7.79768229e-01 1.72960237e-01 -1.50019929e-01
-3.83482486e-01 -1.42263277e-02 6.68658093e-02 1.42566633e+00
-8.07984293e-01 7.82405436e-01 -1.34294188e+00 6.35119677e-01
-1.56926751e+00 -8.29933286e-01 -6.14063203e-01 2.70091009e+00
7.15506196e-01 -3.75578731e-01 1.43363118e-01 5.24885207e-02
6.56226695e-01 3.18600871e-02 1.09935656e-01 -6.07752353e-02
1.16093956e-01 2.71309074e-02 3.87199968e-01 5.24399340e-01
-8.40169966e-01 3.96672845e-01 6.32683897e+00 3.53850752e-01
-1.27530265e+00 2.99450010e-01 1.03981696e-01 -1.35614946e-01
-3.99069726e-01 -9.50244218e-02 -5.46347678e-01 2.60543376e-01
7.30071008e-01 -1.91195548e-01 3.29221487e-01 2.74663627e-01
6.54546857e-01 -1.27060160e-01 -7.98650146e-01 1.10681450e+00
1.42724350e-01 -1.59179318e+00 -1.18060768e-01 1.01090092e-02
6.28835857e-01 -2.31552310e-02 -3.68095249e-01 -3.67234379e-01
-4.82946038e-01 -9.79609549e-01 5.32659829e-01 6.30811214e-01
9.06989038e-01 -6.12284422e-01 8.83284390e-01 2.14982003e-01
-1.43608415e+00 4.65605319e-01 -3.05417210e-01 2.93010354e-01
2.55309552e-01 7.83388555e-01 -1.28868949e+00 6.95926547e-01
7.51376510e-01 5.19643605e-01 -1.38381988e-01 1.23337936e+00
1.00977987e-01 6.77009225e-01 -3.04435402e-01 4.70267236e-01
-2.20921338e-01 -6.47830546e-01 1.00238013e+00 1.37594938e+00
6.90447330e-01 4.74694818e-01 -1.82295322e-01 6.27024472e-01
4.91993725e-01 4.92467403e-01 -4.12850469e-01 3.42606783e-01
4.25845653e-01 1.47593689e+00 -1.14192069e+00 6.97263777e-02
-4.58431184e-01 5.50762713e-01 -9.07972362e-03 2.50619918e-01
-5.87213159e-01 -1.52497552e-03 3.78248572e-01 5.43592989e-01
1.31371260e-01 -4.98044670e-01 -3.62408847e-01 -1.11213756e+00
1.08898021e-01 -5.95426738e-01 3.36192578e-01 -4.42739516e-01
-6.79747283e-01 7.54301131e-01 2.15348393e-01 -1.55016065e+00
-1.84599504e-01 -9.31241736e-02 -5.04976273e-01 1.14682686e+00
-1.07392180e+00 -9.71587121e-01 -3.69239658e-01 2.90888965e-01
1.98380202e-01 6.01941645e-02 8.13569963e-01 3.53512079e-01
-4.87218291e-01 3.72182459e-01 -8.58136863e-02 1.55921906e-01
5.09000957e-01 -9.64971364e-01 -1.73930123e-01 1.09863424e+00
-1.22138835e-01 1.07746089e+00 6.52902186e-01 -9.28097665e-01
-1.54825437e+00 -7.57676005e-01 7.46631920e-01 -3.07580203e-01
3.61380309e-01 1.82654053e-01 -1.07181644e+00 3.35701138e-01
-3.15519571e-01 3.33524078e-01 8.89355123e-01 -3.97899777e-01
1.11367591e-01 -1.47400588e-01 -1.48662758e+00 3.50557685e-01
8.00912678e-01 2.41854023e-02 -6.27970755e-01 5.23635633e-02
3.48483443e-01 -9.70543742e-01 -1.50262225e+00 7.29418874e-01
7.90291309e-01 -7.31532991e-01 1.25287402e+00 -2.20549549e-03
3.19346815e-01 -6.94259167e-01 1.48574844e-01 -8.89202237e-01
-6.41472518e-01 -7.23300159e-01 -1.67085696e-02 1.07455325e+00
3.13696295e-01 -5.42531848e-01 5.99787295e-01 7.00422823e-01
-1.71548977e-01 -7.81944931e-01 -1.31401467e+00 -4.49142098e-01
-1.57539546e-01 -1.89731643e-01 3.46356034e-01 1.13650012e+00
-6.36072457e-02 -1.11994855e-01 -2.21954301e-01 6.03230655e-01
8.90834928e-01 1.57869369e-01 2.74210125e-01 -9.99913216e-01
3.54994051e-02 -2.36336350e-01 -6.76259518e-01 -3.75709891e-01
-2.76104897e-01 -1.21019208e+00 -2.22882559e-03 -1.83344901e+00
2.85378277e-01 -4.59937692e-01 -3.69533718e-01 2.08117306e-01
-1.90638512e-01 4.74720240e-01 1.46892071e-01 4.38355237e-01
1.05132081e-01 1.30277099e-02 1.63521385e+00 3.15978587e-01
-5.41212201e-01 4.67827395e-02 -2.87148386e-01 6.32802010e-01
3.80666494e-01 -5.41760981e-01 -2.22910181e-01 -3.50498468e-01
1.44361779e-01 6.54598475e-01 -1.05779313e-01 -8.95384431e-01
1.80931672e-01 -1.95073962e-01 5.40703952e-01 -8.82971346e-01
2.97901899e-01 -5.36063492e-01 6.68756962e-01 5.45936942e-01
1.30265042e-01 7.30711818e-02 1.67339399e-01 6.38566613e-02
-9.96326506e-02 -3.25872183e-01 7.98824072e-01 -6.44522980e-02
-3.69817108e-01 3.48228335e-01 -3.21293712e-01 -2.11830214e-01
7.61458337e-01 -4.58037406e-01 1.60967648e-01 -2.37203807e-01
-9.24584627e-01 -3.84057090e-02 2.78206438e-01 2.95199025e-02
1.27357650e+00 -1.26836133e+00 -1.10113716e+00 2.51875281e-01
-3.58507097e-01 2.89997816e-01 4.75641489e-01 1.61781180e+00
-1.18535006e+00 2.32704088e-01 -2.15282559e-01 -1.03213203e+00
-1.59292269e+00 5.69732040e-02 2.39613593e-01 -1.79981142e-01
-9.51835155e-01 8.59790862e-01 -2.17782725e-02 -2.08209068e-01
-3.12837750e-01 -2.91046798e-01 -7.43031621e-01 1.38877705e-01
6.81564808e-01 7.38979816e-01 3.60208631e-01 -1.17966342e+00
-8.40906084e-01 1.10129905e+00 3.70852649e-01 -2.54028916e-01
1.73181891e+00 -1.68777913e-01 -5.32326400e-01 2.71170348e-01
1.12643468e+00 4.46858734e-01 -7.98507333e-01 1.96277156e-01
-6.44028336e-02 -7.50424385e-01 3.56938541e-01 -1.92582533e-01
-1.30831563e+00 6.39000952e-01 7.51165986e-01 -3.42880398e-01
1.06833184e+00 -2.45433167e-01 4.53232169e-01 -4.29149270e-01
6.77207887e-01 -4.76002783e-01 -2.10185915e-01 2.51142532e-01
1.34914017e+00 -9.77045894e-01 4.71080273e-01 -7.00569391e-01
-6.45629585e-01 1.36439693e+00 3.41758490e-01 -3.26327235e-02
6.18010581e-01 2.73568600e-01 1.25251815e-01 -2.78892785e-01
-2.00326487e-01 2.61594385e-01 8.52923274e-01 4.68417943e-01
9.85481918e-01 1.27594098e-02 -7.66704082e-01 5.04297853e-01
-2.06385970e-01 -5.61061688e-03 6.07058346e-01 8.34745646e-01
-1.14296041e-01 -7.57134974e-01 -7.08160698e-01 4.48598623e-01
-7.52814114e-01 -1.29798263e-01 4.76994514e-01 2.69266367e-01
-1.51677117e-01 7.13020265e-01 -1.35644093e-01 -2.34264329e-01
2.54884571e-01 -1.65847778e-01 6.63899958e-01 -6.38855934e-01
-5.86959183e-01 5.89622557e-01 6.83742166e-02 -8.76177669e-01
-4.52944666e-01 -9.21600163e-01 -1.64482427e+00 1.40491098e-01
-2.70114779e-01 7.18857497e-02 9.07095671e-01 6.42674208e-01
2.61126459e-01 3.53981614e-01 7.51662850e-01 -1.11932182e+00
5.28373830e-02 -7.27935314e-01 -5.17090738e-01 2.90566355e-01
5.57233274e-01 -7.08484828e-01 -4.10118550e-01 4.62969877e-02] | [13.779532432556152, -2.463430643081665] |
d61d33bc-b9b3-4cfa-93e7-8cb4605fe171 | deep-multi-patch-aggregation-network-for | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Lu_Deep_Multi-Patch_Aggregation_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Lu_Deep_Multi-Patch_Aggregation_ICCV_2015_paper.pdf | Deep Multi-Patch Aggregation Network for Image Style, Aesthetics, and Quality Estimation | This paper investigates problems of image style, aesthetics, and quality estimation, which require fine-grained details from high-resolution images, utilizing deep neural network training approach. Existing deep convolutional neural networks mostly extracted one patch such as a down-sized crop from each image as a training example. However, one patch may not always well represent the entire image, which may cause ambiguity during training. We propose a deep multi-patch aggregation network training approach, which allows us to train models using multiple patches generated from one image. We achieve this by constructing multiple, shared columns in the neural network and feeding multiple patches to each of the columns. More importantly, we propose two novel network layers (statistics and sorting) to support aggregation of those patches. The proposed deep multi-patch aggregation network integrates shared feature learning and aggregation function learning into a unified framework. We demonstrate the effectiveness of the deep multi-patch aggregation network on the three problems, i.e., image style recognition, aesthetic quality categorization, and image quality estimation. Our models trained using the proposed networks significantly outperformed the state of the art in all three applications. | ['Radomir Mech', 'Xin Lu', 'Xiaohui Shen', 'Zhe Lin', 'James Z. Wang'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['image-quality-estimation', 'aesthetics-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 3.92189980e-01 -4.30414677e-01 6.53561279e-02 -5.53246915e-01
-6.26172662e-01 -3.54131758e-01 2.46008299e-02 6.68699369e-02
-1.09540410e-01 4.11924452e-01 -4.20297356e-03 1.09255180e-01
-1.67589635e-01 -1.07644570e+00 -9.62602079e-01 -5.74660540e-01
3.30154479e-01 -8.10128748e-02 -1.66825294e-01 -5.79618849e-02
4.06369060e-01 5.30900180e-01 -1.68289018e+00 5.95573902e-01
1.18338323e+00 1.71812487e+00 3.37317050e-01 5.00688612e-01
-3.66637141e-01 5.62187135e-01 -7.73805857e-01 -5.65880954e-01
2.34397665e-01 -2.73102641e-01 -6.29474998e-01 5.49049020e-01
6.37425184e-01 -6.39171779e-01 1.01496046e-02 1.17715871e+00
4.05384272e-01 -7.28405938e-02 4.92345363e-01 -1.25336361e+00
-1.14420128e+00 3.75772148e-01 -8.45012844e-01 -1.99944735e-01
1.00305025e-02 1.96548358e-01 1.30704892e+00 -8.01571250e-01
2.85905957e-01 1.29074001e+00 5.81431329e-01 2.16097489e-01
-1.04457593e+00 -7.57377148e-01 1.31268084e-01 6.19371645e-02
-1.22615218e+00 -2.67384291e-01 1.01477718e+00 -2.87075698e-01
5.09285748e-01 1.95594579e-01 7.66943276e-01 9.10741746e-01
1.87684953e-01 1.09810567e+00 1.25967109e+00 -1.87144071e-01
1.54477790e-01 3.60346735e-02 -6.04387894e-02 7.54720330e-01
3.28260958e-01 -1.03249840e-01 -4.34723854e-01 -3.34351696e-02
1.15667033e+00 5.78502864e-02 -2.26238202e-02 -1.26169458e-01
-1.01829767e+00 8.13949645e-01 7.05111206e-01 2.43154585e-01
-3.26084167e-01 7.16955885e-02 3.13542277e-01 2.16048703e-01
4.72604752e-01 4.32180583e-01 -5.19796073e-01 2.21216783e-01
-9.07582819e-01 9.54041854e-02 4.47071791e-01 9.10466909e-01
1.11258590e+00 -2.56800167e-02 -3.15411806e-01 1.19070756e+00
5.79755418e-02 3.99063200e-01 3.38747650e-01 -9.67545629e-01
3.16040903e-01 9.78568733e-01 -1.06982119e-01 -1.46719158e+00
-2.78421849e-01 -6.72283411e-01 -1.30483437e+00 4.42668468e-01
1.49012327e-01 2.97642965e-02 -6.87553823e-01 1.61617100e+00
1.09428316e-01 -2.39526778e-01 -2.16396470e-02 8.40701520e-01
1.03795791e+00 6.74930811e-01 -1.53132519e-02 8.06568936e-02
1.20730627e+00 -1.30983031e+00 -5.08336186e-01 -2.04640567e-01
3.43038023e-01 -9.36709404e-01 1.41924667e+00 7.10541725e-01
-1.15005589e+00 -9.24152076e-01 -1.15453398e+00 -2.17974395e-01
-3.63447964e-01 7.92095542e-01 7.29466736e-01 3.78250182e-01
-1.01992822e+00 8.29443276e-01 -1.61958441e-01 -2.94242054e-02
9.15016294e-01 1.27645314e-01 -4.40461725e-01 -1.03411108e-01
-7.25002646e-01 2.61184543e-01 4.37370390e-02 1.42005846e-01
-5.79554677e-01 -3.48579854e-01 -8.60496283e-01 2.98642367e-01
1.61099374e-01 -7.28370011e-01 8.03554296e-01 -1.73577106e+00
-1.50692677e+00 8.51184726e-01 5.00846654e-03 -6.77346019e-03
4.28377315e-02 -3.80042940e-01 -4.58734691e-01 2.85564344e-02
1.56289190e-01 8.11109722e-01 1.01945162e+00 -1.44653404e+00
-5.46887577e-01 -3.39906037e-01 2.40509674e-01 9.44571197e-02
-6.03986561e-01 -1.59220412e-01 -4.77434188e-01 -1.01926434e+00
-1.74379703e-02 -4.51518923e-01 -1.83132827e-01 2.85389364e-01
-5.30765474e-01 -1.09333098e-01 5.61554611e-01 -6.82408154e-01
1.32429600e+00 -2.26840043e+00 1.26226917e-01 1.09794132e-01
2.87085772e-01 1.44938119e-02 -5.13377190e-01 -1.40857771e-01
1.20700635e-01 2.76679188e-01 -2.40757912e-01 -4.28755164e-01
7.86039606e-02 8.25110227e-02 2.62949001e-02 9.45171267e-02
6.31016374e-01 8.64101350e-01 -6.46385133e-01 -5.12695670e-01
1.36036426e-01 2.91495383e-01 -6.38595223e-01 4.68105108e-01
-2.35030457e-01 2.32669920e-01 -3.30421716e-01 9.45609331e-01
9.66773033e-01 -5.69049478e-01 -2.12405235e-01 -6.88009799e-01
5.27874380e-02 -1.54299885e-01 -1.29475868e+00 2.01028943e+00
-6.49745524e-01 4.99138713e-01 -2.82075554e-02 -8.84522200e-01
1.26684487e+00 -1.42446041e-01 4.01102245e-01 -9.57323372e-01
2.73153484e-01 2.56340891e-01 -1.27870634e-01 -4.13683116e-01
6.31778121e-01 1.25671849e-01 -1.74240649e-01 6.77810669e-01
1.25737309e-01 2.54758209e-01 3.57264280e-02 -1.43478870e-01
7.44019747e-01 8.55121538e-02 1.70819089e-01 -6.67601898e-02
5.45729876e-01 -3.05241764e-01 6.29586756e-01 6.07871413e-01
-6.39000162e-02 7.91658580e-01 5.45195580e-01 -7.15779960e-01
-1.26941514e+00 -7.38136530e-01 1.03997104e-02 1.18195105e+00
3.02219480e-01 -3.04247618e-01 -8.15747857e-01 -6.98282897e-01
-4.18341942e-02 4.55727093e-02 -8.72295141e-01 -7.59105012e-02
-3.11725557e-01 -6.80191576e-01 2.62134194e-01 5.65198600e-01
9.28474844e-01 -1.33291209e+00 -5.17176211e-01 8.74043778e-02
-1.67527854e-01 -1.03980684e+00 -5.51120996e-01 -6.85955025e-03
-6.77673280e-01 -1.04617035e+00 -6.75093472e-01 -7.83960283e-01
6.85672402e-01 3.33130300e-01 1.43465149e+00 3.93100649e-01
-3.37623417e-01 2.82989703e-02 -3.96558940e-01 -1.14571333e-01
-8.09234083e-02 3.38001288e-02 -3.63103867e-01 4.56333309e-01
7.09794983e-02 -5.21181881e-01 -8.90509248e-01 1.97602093e-01
-1.16335642e+00 2.19034165e-01 1.25401437e+00 1.09356248e+00
8.24763358e-01 2.56727368e-01 4.87547994e-01 -7.03007877e-01
7.08683014e-01 -1.95651799e-01 -2.61978298e-01 3.97288978e-01
-3.15568924e-01 1.10959306e-01 6.24615312e-01 -2.82673001e-01
-9.19914126e-01 -2.83909552e-02 -2.71171600e-01 -4.21196342e-01
-3.86645526e-01 3.67881417e-01 -6.51620865e-01 -2.75832534e-01
4.43587571e-01 8.97561684e-02 -8.43174942e-03 -6.11209273e-01
3.25549632e-01 7.05186188e-01 3.55675101e-01 -6.28702760e-01
4.57507968e-01 1.94242880e-01 -1.38953343e-01 -4.95821118e-01
-1.00093055e+00 -6.03393279e-02 -5.77153504e-01 -5.30031584e-02
8.13255668e-01 -8.73209476e-01 -5.56156754e-01 8.14095795e-01
-1.23642111e+00 -2.26171687e-01 -2.60743797e-01 -1.08606972e-01
-4.39643949e-01 3.60496700e-01 -5.75678885e-01 -5.53509235e-01
-5.77499568e-01 -1.32566035e+00 1.50941718e+00 3.81659567e-01
1.08281944e-04 -7.81931937e-01 -3.47019225e-01 1.42508954e-01
4.88825589e-01 4.86208528e-01 1.01650310e+00 -9.49773714e-02
-4.48881447e-01 -1.56616587e-02 -7.13325262e-01 7.27892697e-01
3.37073386e-01 1.30499616e-01 -1.05664039e+00 -6.34960532e-02
-2.27071792e-01 -6.38774633e-01 9.10269201e-01 4.92180198e-01
1.88257813e+00 -3.37676853e-01 1.37723178e-01 9.35093105e-01
1.62361526e+00 -5.68042174e-02 6.55733883e-01 5.75674474e-01
8.08344305e-01 5.53603172e-01 4.92735714e-01 6.76639497e-01
2.58743137e-01 4.11443889e-01 5.65348625e-01 -5.80625296e-01
-1.49308845e-01 9.85580385e-02 9.22504440e-02 5.91885328e-01
1.81131706e-01 -8.58339518e-02 -3.42276573e-01 2.90390968e-01
-1.58216894e+00 -6.61650717e-01 2.03872144e-01 1.83297479e+00
6.55768037e-01 1.89729869e-01 1.09670326e-01 2.58570373e-01
7.58026004e-01 1.90033704e-01 -7.23136842e-01 -4.16132361e-01
-5.32508135e-01 3.92122388e-01 8.46119672e-02 1.86780468e-01
-1.25971127e+00 7.46633232e-01 5.89235640e+00 1.00066173e+00
-1.14728343e+00 -1.53209165e-01 1.29300165e+00 1.48709372e-01
-5.53877413e-01 -3.76752585e-01 -4.87840086e-01 4.38499361e-01
2.75456876e-01 2.25401357e-01 5.36357999e-01 9.07260239e-01
1.85893048e-02 -1.18254930e-01 -9.07433331e-01 1.30155861e+00
1.16519682e-01 -1.23055780e+00 5.08948088e-01 -5.30039482e-02
8.72392893e-01 -5.63181520e-01 4.28842694e-01 5.94589263e-02
6.69235811e-02 -1.27309227e+00 7.36486852e-01 5.29169917e-01
9.17495787e-01 -8.89063179e-01 7.25405812e-01 6.09744936e-02
-1.20996428e+00 -3.09166998e-01 -5.95055759e-01 -1.56932678e-02
-3.46890837e-01 9.06855762e-01 2.14793906e-01 6.30787313e-01
9.89838302e-01 9.56049025e-01 -9.00068939e-01 9.90517259e-01
1.70576930e-01 3.93325649e-02 -1.06396908e-02 8.21386725e-02
2.16441989e-01 -2.10092455e-01 3.49493399e-02 9.30842340e-01
5.74191451e-01 -1.79565087e-01 4.95032854e-02 1.17251122e+00
-3.04478914e-01 2.62905985e-01 -3.17489892e-01 -6.77360818e-02
1.81585088e-01 1.66803646e+00 -6.12883985e-01 -3.62682521e-01
-5.11451662e-01 1.23676908e+00 5.91173887e-01 1.29073098e-01
-6.54243588e-01 -6.22796893e-01 7.03747034e-01 -1.78906024e-01
3.40558708e-01 5.27741201e-02 -7.58954287e-01 -1.15497172e+00
2.25449085e-01 -1.02451301e+00 1.69174016e-01 -9.66567636e-01
-1.67952514e+00 9.40701485e-01 -6.45082712e-01 -1.35965943e+00
2.47007832e-01 -9.01084840e-01 -7.68318772e-01 9.19981539e-01
-1.54218590e+00 -1.28222382e+00 -8.13553452e-01 7.08125710e-01
4.79285628e-01 -1.50028020e-01 8.89941394e-01 2.76961952e-01
-7.30908096e-01 7.42148757e-01 -5.57462568e-04 1.80582762e-01
7.48290062e-01 -1.39519989e+00 2.45370939e-01 6.39881849e-01
-7.69945700e-03 3.92316252e-01 1.73014238e-01 -2.83196062e-01
-1.21831608e+00 -1.23665857e+00 3.90797615e-01 5.52721433e-02
3.23078632e-01 -2.13868111e-01 -8.31484616e-01 2.07653850e-01
3.72035354e-01 9.52879414e-02 7.84943640e-01 1.67371064e-01
-5.08774102e-01 -4.43267852e-01 -1.16172671e+00 3.29673618e-01
9.58198905e-01 -2.69903302e-01 -6.92506433e-02 4.56400178e-02
6.65360510e-01 -1.78812101e-01 -9.97841060e-01 4.57128078e-01
7.39001572e-01 -1.28693449e+00 8.19141567e-01 -2.24565849e-01
1.06606245e+00 -1.86326176e-01 -2.94252604e-01 -1.39872682e+00
-5.92880547e-01 -7.87251666e-02 6.66067982e-03 1.30590701e+00
2.81194836e-01 -2.28997454e-01 7.30651021e-01 2.82954067e-01
-1.19440317e-01 -1.11825240e+00 -3.05103987e-01 -2.84213185e-01
1.00503728e-01 -2.41049647e-01 1.15626752e+00 6.62335634e-01
-5.79579473e-01 4.46208596e-01 -4.47937191e-01 4.68088053e-02
6.46558166e-01 5.13073981e-01 7.81323433e-01 -1.25657260e+00
-4.31623220e-01 -9.29110289e-01 -2.67412990e-01 -1.05691528e+00
-5.03525101e-02 -4.16583568e-01 -6.31632358e-02 -1.72896719e+00
4.58689094e-01 -6.37113929e-01 -4.86173064e-01 3.47319335e-01
-4.05001074e-01 4.47893798e-01 7.66875371e-02 8.38972926e-02
-7.56104231e-01 6.34759009e-01 1.73054326e+00 -5.54468572e-01
-3.59954946e-02 -1.95043892e-01 -1.18905675e+00 4.92805183e-01
8.76851916e-01 4.44212928e-02 -2.36877143e-01 -7.84227192e-01
4.19284642e-01 -2.47848138e-01 4.73114491e-01 -1.22971129e+00
-4.09576222e-02 -1.56828418e-01 1.05333221e+00 -4.43603516e-01
1.08639032e-01 -8.91790688e-01 6.69450462e-02 9.53235552e-02
-3.68025512e-01 -6.08050916e-03 3.12971145e-01 3.42122793e-01
-3.50819737e-01 -1.39326826e-02 8.76514494e-01 -3.30644786e-01
-7.32550800e-01 5.82005978e-01 2.61029422e-01 -3.27665150e-01
7.29942918e-01 -1.22781515e-01 -3.63908261e-01 -1.19540229e-01
-8.11986625e-01 -4.28766347e-02 5.59422314e-01 4.42194879e-01
8.78634334e-01 -1.80588019e+00 -5.50280392e-01 4.20686901e-01
2.52143294e-01 1.95466697e-01 5.62027037e-01 4.48751986e-01
-4.73748028e-01 -7.76951835e-02 -7.58968055e-01 -7.19118118e-01
-9.98156309e-01 5.72619557e-01 2.85695940e-01 -3.15504849e-01
-3.32299799e-01 8.56410086e-01 4.87278491e-01 -3.48010093e-01
1.89402059e-01 -3.22310120e-01 -3.48171085e-01 -6.83211759e-02
5.01023531e-01 7.33873621e-02 7.16963410e-02 -6.24716043e-01
1.25827556e-02 1.01158845e+00 -6.05164655e-02 2.71694839e-01
1.33381557e+00 -2.34036714e-01 -3.64864975e-01 3.50502908e-01
1.40516162e+00 -2.68133968e-01 -1.30844235e+00 -3.21383178e-01
-2.98876882e-01 -6.32073045e-01 1.25864059e-01 -8.76789033e-01
-1.48003316e+00 1.02414823e+00 6.40043676e-01 1.80523038e-01
1.77121007e+00 -3.24590594e-01 1.02370560e+00 1.76879078e-01
2.32463717e-01 -1.21646988e+00 6.31496727e-01 2.30118677e-01
1.00048494e+00 -1.48431683e+00 -1.45714179e-01 -2.27336600e-01
-5.59897780e-01 1.34150159e+00 9.61835027e-01 -3.91557723e-01
5.82559586e-01 1.13470890e-01 4.25563455e-02 -1.39024436e-01
-5.19264221e-01 -1.16516419e-01 5.95812142e-01 2.91639686e-01
4.41665590e-01 2.26359233e-01 -4.18606550e-02 1.03474343e+00
-1.97219312e-01 -1.07454680e-01 2.48225719e-01 8.21115911e-01
-6.93247139e-01 -1.01522732e+00 -2.57696629e-01 6.82317317e-01
-3.17735553e-01 -2.10844278e-01 -4.17649180e-01 3.91580492e-01
5.73399186e-01 7.24068820e-01 2.27851823e-01 -8.10067177e-01
4.21550423e-01 -3.18864524e-01 4.81494874e-01 -3.11199576e-01
-5.47560275e-01 8.30519944e-02 -3.02066356e-01 -7.45640218e-01
-4.97740626e-01 -3.13733757e-01 -4.89469230e-01 -5.19405186e-01
-1.52724862e-01 -2.72407979e-01 5.33669055e-01 7.79893875e-01
4.35017496e-01 8.66339147e-01 9.53577101e-01 -9.06530857e-01
-1.90316424e-01 -1.02863133e+00 -7.90499449e-01 4.93493170e-01
4.60891783e-01 -5.84668875e-01 -2.18903691e-01 5.35745583e-02] | [11.518881797790527, -1.0880171060562134] |
51baa05c-6f31-4fca-a155-9dc24c648774 | unsupervised-learning-of-object-landmarks | 1806.07823 | null | http://arxiv.org/abs/1806.07823v2 | http://arxiv.org/pdf/1806.07823v2.pdf | Unsupervised Learning of Object Landmarks through Conditional Image Generation | We propose a method for learning landmark detectors for visual objects (such
as the eyes and the nose in a face) without any manual supervision. We cast
this as the problem of generating images that combine the appearance of the
object as seen in a first example image with the geometry of the object as seen
in a second example image, where the two examples differ by a viewpoint change
and/or an object deformation. In order to factorize appearance and geometry, we
introduce a tight bottleneck in the geometry-extraction process that selects
and distils geometry-related features. Compared to standard image generation
problems, which often use generative adversarial networks, our generation task
is conditioned on both appearance and geometry and thus is significantly less
ambiguous, to the point that adopting a simple perceptual loss formulation is
sufficient. We demonstrate that our approach can learn object landmarks from
synthetic image deformations or videos, all without manual supervision, while
outperforming state-of-the-art unsupervised landmark detectors. We further show
that our method is applicable to a large variety of datasets - faces, people,
3D objects, and digits - without any modifications. | ['Tomas Jakab', 'Hakan Bilen', 'Andrea Vedaldi', 'Ankush Gupta'] | 2018-06-20 | unsupervised-learning-of-object-landmarks-1 | http://papers.nips.cc/paper/7657-unsupervised-learning-of-object-landmarks-through-conditional-image-generation | http://papers.nips.cc/paper/7657-unsupervised-learning-of-object-landmarks-through-conditional-image-generation.pdf | neurips-2018-12 | ['unsupervised-facial-landmark-detection'] | ['computer-vision'] | [ 4.09402937e-01 5.61543286e-01 4.10893112e-01 -2.96193868e-01
-7.02304125e-01 -8.04653883e-01 8.25297952e-01 -9.20488983e-02
-3.32297176e-01 3.72441113e-01 -2.06502527e-01 1.97529513e-02
3.39341730e-01 -6.09990060e-01 -1.04268026e+00 -6.64584458e-01
1.60112098e-01 6.18538916e-01 1.90447822e-01 -5.70814200e-02
2.23951310e-01 9.64434266e-01 -1.48628056e+00 -1.34654567e-01
3.53248328e-01 8.78446341e-01 -2.20003605e-01 7.88572431e-01
1.07883245e-01 2.63634264e-01 -5.29079974e-01 -6.36308193e-01
7.14704275e-01 -4.77174431e-01 -7.54147291e-01 5.55067778e-01
1.05928934e+00 -5.01312554e-01 -2.63430566e-01 1.00958967e+00
4.16510910e-01 1.14439115e-01 8.64043653e-01 -1.27616298e+00
-7.22986996e-01 1.14010712e-02 -6.69441938e-01 -3.53410989e-01
6.04417861e-01 1.43246710e-01 8.10017288e-01 -1.13669097e+00
9.39625084e-01 1.42446089e+00 6.50809944e-01 7.19463646e-01
-1.72199965e+00 -2.90928155e-01 1.21720649e-01 -3.59423578e-01
-1.42780495e+00 -6.38239264e-01 1.02559960e+00 -5.75628877e-01
5.69847882e-01 1.59388836e-02 5.52601576e-01 1.01243675e+00
-4.29574400e-02 5.41724801e-01 9.91918743e-01 -6.54361904e-01
3.14912021e-01 3.78678590e-02 -4.78597552e-01 8.79754245e-01
1.48249432e-01 1.46006137e-01 -1.55375496e-01 -4.55390252e-02
1.13093352e+00 1.20191328e-01 -1.18395597e-01 -1.02018106e+00
-1.18360925e+00 6.75792634e-01 5.29154837e-01 -1.92410901e-01
-1.45664990e-01 3.30848068e-01 -1.76205814e-01 2.14339599e-01
3.40751201e-01 4.70559299e-01 -2.79501170e-01 3.28904301e-01
-8.80811870e-01 4.89779770e-01 6.87660158e-01 1.17639136e+00
1.04723775e+00 4.86092605e-02 1.79270685e-01 4.42171574e-01
2.35488504e-01 5.74254870e-01 3.67585868e-02 -1.09469819e+00
1.99339643e-01 4.43034291e-01 2.56213695e-01 -8.74939620e-01
-1.44914493e-01 -9.74763371e-03 -3.90740603e-01 8.04141819e-01
7.65245080e-01 -1.01032391e-01 -1.27049971e+00 1.93020916e+00
6.20280445e-01 2.13882804e-01 -1.45692110e-01 7.59022951e-01
5.70070982e-01 2.52840012e-01 -1.90706462e-01 9.87141207e-02
1.32304096e+00 -7.06418276e-01 -2.40274921e-01 -2.84241319e-01
1.66211560e-01 -8.21748495e-01 8.77253354e-01 1.43023416e-01
-1.47489047e+00 -5.24775326e-01 -1.08334541e+00 -3.44317645e-01
-4.07644063e-01 -2.41714530e-03 3.77230823e-01 6.09179676e-01
-1.20259023e+00 7.10836232e-01 -8.44165325e-01 -4.96059209e-01
3.78580272e-01 5.10539293e-01 -6.13075137e-01 3.28675524e-04
-5.72330236e-01 8.11857343e-01 -1.81319099e-02 -5.59239239e-02
-1.01229060e+00 -6.18183494e-01 -1.18791163e+00 -1.31175205e-01
3.36413920e-01 -7.98876166e-01 1.34289908e+00 -1.24711061e+00
-1.57069349e+00 1.26746595e+00 -1.20562792e-01 -5.53098843e-02
7.85984099e-01 -5.31155728e-02 4.68550585e-02 3.25296581e-01
3.31793563e-03 1.00176156e+00 1.36518407e+00 -1.61650383e+00
-2.10722685e-01 -6.36422038e-01 3.40149850e-01 2.05235422e-01
2.31104568e-01 -5.26125021e-02 -5.54745615e-01 -7.43601680e-01
1.44018546e-01 -1.06569660e+00 -3.08354706e-01 6.28300667e-01
-5.80388010e-01 1.55963093e-01 9.58662391e-01 -4.91662025e-01
2.89568037e-01 -2.10630512e+00 3.15254480e-01 3.51095766e-01
1.95170462e-01 1.37821242e-01 -3.76480699e-01 2.75571674e-01
-1.97769612e-01 1.20442465e-01 -4.03555304e-01 -7.83222377e-01
-8.27129483e-02 1.83067530e-01 -2.99733669e-01 5.99301398e-01
6.80525422e-01 1.06752062e+00 -9.65203643e-01 -3.80582184e-01
2.65528321e-01 6.93437994e-01 -7.07083941e-01 2.34906286e-01
-3.12901884e-01 7.01512754e-01 -3.89650285e-01 5.32472253e-01
7.38400161e-01 -1.82286561e-01 1.06473468e-01 -1.55960619e-01
1.45350263e-01 1.62157923e-01 -1.30870020e+00 1.88357663e+00
-3.62007648e-01 3.20516825e-01 2.69562364e-01 -9.02472556e-01
7.48233020e-01 1.78158760e-01 2.73712128e-01 -2.78129160e-01
2.09798906e-02 8.88718590e-02 -1.48845911e-01 -2.66506046e-01
8.94969255e-02 -2.02227741e-01 3.42103988e-02 4.86943781e-01
2.58598059e-01 -6.94972754e-01 -7.76364207e-02 1.86417088e-01
8.05487275e-01 5.35504460e-01 1.18048154e-01 8.36249888e-02
2.19909638e-01 -2.96595037e-01 3.02773386e-01 5.20523906e-01
1.48443967e-01 9.73386407e-01 5.54318547e-01 -3.64117682e-01
-1.29956210e+00 -1.41594505e+00 -1.42528966e-01 6.78098261e-01
1.58757508e-01 -8.49741399e-02 -9.67803776e-01 -8.27205896e-01
3.09771895e-01 4.57861960e-01 -8.11571717e-01 -1.01407312e-01
-6.71054840e-01 -1.90082878e-01 3.26335281e-01 5.21254182e-01
1.26786858e-01 -1.00271857e+00 -5.58467507e-01 -1.46132305e-01
3.72220665e-01 -1.14281666e+00 -7.87198067e-01 -9.48574543e-02
-7.51984417e-01 -1.09371734e+00 -7.74149716e-01 -9.25943077e-01
1.30756807e+00 7.84695968e-02 1.13619423e+00 6.67420402e-02
-7.09921241e-01 6.60188973e-01 -9.41776484e-03 -2.30330423e-01
-3.86866540e-01 -4.23284769e-01 4.90802750e-02 2.51141667e-01
-2.36737542e-02 -8.03406119e-01 -6.98745251e-01 3.40535104e-01
-1.03185725e+00 1.39253447e-02 4.22627330e-01 7.52817750e-01
6.17252409e-01 -2.98730075e-01 2.18359947e-01 -8.62083018e-01
1.41653314e-01 -1.13676168e-01 -7.69377351e-01 9.70128626e-02
-2.64360875e-01 1.95549354e-01 6.44181192e-01 -4.92667705e-01
-8.43743265e-01 6.18279994e-01 6.90907836e-02 -5.38828433e-01
-3.33655089e-01 -1.02963172e-01 -4.06984121e-01 -4.12813395e-01
6.99242413e-01 5.97593449e-02 2.67085046e-01 -4.50966418e-01
7.15702772e-01 1.66841701e-01 8.04304302e-01 -6.60686135e-01
1.34985709e+00 7.54898012e-01 3.80015552e-01 -7.80242741e-01
-5.55324018e-01 -1.33630976e-01 -1.08690929e+00 1.32278651e-01
8.00804794e-01 -6.83312774e-01 -6.15347326e-01 4.02368695e-01
-1.28343916e+00 -2.26845816e-01 -5.34090579e-01 1.75734445e-01
-8.82581770e-01 3.80349040e-01 -4.35947984e-01 -5.85249901e-01
1.83219865e-01 -1.21218586e+00 1.52712011e+00 1.17691927e-01
-2.62823999e-02 -9.90223408e-01 -1.90886378e-01 8.73300135e-02
1.19514793e-01 5.95414937e-01 9.04670417e-01 -4.71478939e-01
-9.28812146e-01 -2.97552198e-01 -4.48360927e-02 4.89371270e-01
2.97978014e-01 1.67756259e-01 -1.01699960e+00 -3.17451924e-01
-1.34216622e-01 -5.05293608e-01 5.62954664e-01 1.46132275e-01
9.91877973e-01 -4.07345623e-01 -3.39606464e-01 8.15659344e-01
1.24761522e+00 1.35677025e-01 6.73930883e-01 -2.71167129e-01
8.48540306e-01 7.87839293e-01 1.84462935e-01 1.93500429e-01
5.92909269e-02 8.78379047e-01 4.76350933e-01 -3.97210628e-01
-2.12014094e-01 -5.02835035e-01 2.95115918e-01 8.89962614e-02
-1.15471534e-01 -3.15886922e-02 -7.34582484e-01 4.55140471e-01
-1.51048350e+00 -8.00879538e-01 3.61343801e-01 2.41252208e+00
7.80096233e-01 -1.14208259e-01 2.07832649e-01 2.47702841e-02
5.56490779e-01 2.21437365e-02 -7.20304847e-01 -2.45529681e-01
1.71872348e-01 5.07030904e-01 3.40505213e-01 7.22728074e-01
-1.08868647e+00 9.22693551e-01 6.96303511e+00 2.91881830e-01
-1.05867100e+00 -2.22946003e-01 5.91819227e-01 -1.10595003e-02
-4.31796521e-01 1.40034035e-01 -5.42979360e-01 1.24508537e-01
3.30425560e-01 8.38595554e-02 5.25747061e-01 9.01660502e-01
-9.25291926e-02 1.65231749e-01 -1.68505847e+00 7.86595285e-01
3.14742774e-01 -1.02020907e+00 2.44201049e-01 8.83393213e-02
7.31269300e-01 -4.71126407e-01 3.82839173e-01 -1.88534558e-01
3.18075061e-01 -1.24947286e+00 8.57464552e-01 4.04995620e-01
1.08137488e+00 -4.95497942e-01 1.76620916e-01 4.46943901e-02
-9.63106692e-01 2.93623537e-01 -2.01302111e-01 1.78292096e-01
9.54689160e-02 1.55107126e-01 -8.81901979e-01 3.06543559e-01
3.41080427e-01 4.15867746e-01 -5.44577956e-01 8.44808280e-01
-4.09806639e-01 9.77178812e-02 -4.51759100e-01 3.31813008e-01
5.36527000e-02 -2.75651455e-01 6.36576474e-01 1.00001884e+00
2.77581900e-01 -8.97890981e-03 1.21360585e-01 1.16133285e+00
-3.04392040e-01 -4.85440716e-02 -8.65902483e-01 1.62274688e-01
3.42250943e-01 1.28196573e+00 -6.98172629e-01 -1.69517383e-01
-4.24626827e-01 1.26338148e+00 2.50497967e-01 4.93681341e-01
-6.45175278e-01 -4.54727858e-01 6.14617050e-01 4.76915270e-01
5.78836143e-01 -2.22964317e-01 -3.52215916e-02 -1.09351563e+00
2.31738314e-01 -7.54223466e-01 -1.33183360e-01 -9.28807080e-01
-1.12570989e+00 4.77791518e-01 6.12796061e-02 -1.19684875e+00
-5.20432293e-01 -8.37757409e-01 -6.54414237e-01 8.98752034e-01
-1.33329380e+00 -1.37016010e+00 -3.29937249e-01 6.52572453e-01
2.46613741e-01 1.26617461e-01 8.08794498e-01 -2.84269750e-02
-1.67628333e-01 6.40910745e-01 -1.97651297e-01 2.78192788e-01
6.43885195e-01 -1.42616713e+00 7.78499126e-01 7.25076318e-01
3.69305521e-01 6.47224307e-01 5.83178341e-01 -3.44662786e-01
-1.46210825e+00 -1.07085454e+00 6.77724063e-01 -7.76541770e-01
3.18604469e-01 -7.85932541e-01 -4.94865626e-01 8.73069227e-01
1.51379947e-02 4.72671479e-01 1.66359156e-01 -2.56719410e-01
-6.39672875e-01 -3.63662746e-03 -1.39035535e+00 7.78140068e-01
1.26501930e+00 -5.76208115e-01 -6.05139077e-01 3.51438850e-01
5.14694571e-01 -4.69369173e-01 -6.27584398e-01 2.76518792e-01
5.30734539e-01 -8.74164939e-01 1.25724280e+00 -8.16649914e-01
3.03416669e-01 -5.27027071e-01 -6.13832250e-02 -1.30631685e+00
-1.06206119e-01 -9.55772340e-01 4.59388345e-02 1.02702391e+00
3.51402342e-01 -4.68819737e-01 7.58169472e-01 6.48494005e-01
6.06875978e-02 -5.74026823e-01 -9.21092629e-01 -6.46981537e-01
4.75109853e-02 -5.41903675e-02 5.52414596e-01 7.31045425e-01
-3.04204851e-01 3.24897945e-01 -2.20670953e-01 2.36225888e-01
6.90822423e-01 1.46490470e-01 1.10965264e+00 -1.14873719e+00
-3.46493572e-01 -3.00806224e-01 -7.76243627e-01 -1.21272063e+00
1.70671374e-01 -7.66065300e-01 4.59063649e-02 -1.23677385e+00
7.84615893e-03 -4.18596655e-01 1.30234897e-01 5.09372234e-01
-9.37701613e-02 4.76542354e-01 1.83916077e-01 1.48923090e-02
-1.90864637e-01 4.65701163e-01 1.45323992e+00 -1.11825235e-01
-8.14802200e-02 -2.38768533e-02 -6.41002357e-01 9.19116080e-01
3.05177778e-01 -3.06974739e-01 -4.54487354e-01 -4.12197918e-01
-8.56414363e-02 6.22710511e-02 7.62319088e-01 -7.29958594e-01
-5.00685498e-02 -6.38831183e-02 7.35990047e-01 -7.01979399e-02
6.69584811e-01 -8.84588957e-01 1.78455055e-01 2.51340210e-01
-2.46368453e-01 1.69913441e-01 1.00089967e-01 5.60794711e-01
7.93115571e-02 -1.74003690e-01 9.91075456e-01 -2.46618703e-01
-2.12733880e-01 5.28990149e-01 1.19780220e-01 7.73275420e-02
1.17045486e+00 -3.51654530e-01 -9.90153775e-02 -5.09413600e-01
-8.82021666e-01 -1.38813734e-01 1.04412949e+00 3.29128116e-01
7.12997556e-01 -1.42427289e+00 -6.15605474e-01 6.08702242e-01
5.25166094e-02 4.30141985e-01 -1.56683102e-01 5.51848114e-01
-6.19092464e-01 1.17639853e-02 -1.59849107e-01 -6.02086008e-01
-1.15218413e+00 8.09667587e-01 5.86737394e-01 6.00834452e-02
-5.71064234e-01 8.54561746e-01 6.91485882e-01 -4.31859195e-01
2.06900209e-01 -2.67121702e-01 3.29519302e-01 -2.10540101e-01
3.24880481e-01 -5.50750941e-02 -2.62886733e-02 -8.15698743e-01
-3.17477137e-01 9.52660143e-01 -3.84045169e-02 -2.78502107e-01
1.07196057e+00 8.10264125e-02 1.29930988e-01 2.51647741e-01
1.24324346e+00 3.71356428e-01 -1.65865850e+00 -1.54745579e-01
-3.36780280e-01 -6.40201151e-01 -3.83389145e-01 -5.26901245e-01
-1.07560480e+00 9.14053738e-01 3.56207252e-01 -1.71978597e-03
9.41520154e-01 1.35273382e-01 4.74362522e-01 3.18941683e-01
3.56926441e-01 -7.99235523e-01 4.13824350e-01 1.99811041e-01
1.14250541e+00 -1.09313190e+00 -7.71377981e-02 -5.86070657e-01
-3.54906559e-01 1.00932562e+00 5.21356761e-01 -4.67525214e-01
7.05541551e-01 2.03846574e-01 7.91484788e-02 -1.01054922e-01
-3.56571585e-01 -8.99616107e-02 5.46889484e-01 7.88946152e-01
2.86717772e-01 -1.55267686e-01 2.55464673e-01 -1.44067392e-01
-1.34181082e-01 -3.18896383e-01 3.49381924e-01 9.93734300e-01
-4.87994924e-02 -1.35504448e+00 -3.39425296e-01 1.15386374e-01
-4.47537422e-01 -5.29829040e-02 -6.74631715e-01 1.01192987e+00
1.61046192e-01 5.12960196e-01 2.57263839e-01 -1.55983627e-01
4.06969905e-01 1.81452170e-01 9.50629175e-01 -8.75727892e-01
-4.49894071e-02 1.54719763e-02 -1.76629573e-01 -6.60308719e-01
-4.85148638e-01 -6.91249907e-01 -1.11709392e+00 1.91370898e-03
-2.41428673e-01 -2.18151391e-01 6.54836893e-01 7.44953930e-01
3.35691303e-01 1.05978429e-01 6.90065742e-01 -1.39510489e+00
-5.69016159e-01 -6.60284281e-01 -5.53788662e-01 8.12924802e-01
6.24260068e-01 -8.90058696e-01 -4.66024369e-01 4.65139687e-01] | [11.69188117980957, -0.6408027410507202] |
ad1917c3-d4d2-41d1-95e2-402695784132 | learning-a-3d-morphable-face-reflectance | 2303.11686 | null | https://arxiv.org/abs/2303.11686v1 | https://arxiv.org/pdf/2303.11686v1.pdf | Learning a 3D Morphable Face Reflectance Model from Low-cost Data | Modeling non-Lambertian effects such as facial specularity leads to a more realistic 3D Morphable Face Model. Existing works build parametric models for diffuse and specular albedo using Light Stage data. However, only diffuse and specular albedo cannot determine the full BRDF. In addition, the requirement of Light Stage data is hard to fulfill for the research communities. This paper proposes the first 3D morphable face reflectance model with spatially varying BRDF using only low-cost publicly-available data. We apply linear shiness weighting into parametric modeling to represent spatially varying specular intensity and shiness. Then an inverse rendering algorithm is developed to reconstruct the reflectance parameters from non-Light Stage data, which are used to train an initial morphable reflectance model. To enhance the model's generalization capability and expressive power, we further propose an update-by-reconstruction strategy to finetune it on an in-the-wild dataset. Experimental results show that our method obtains decent rendering results with plausible facial specularities. Our code is released \href{https://yxuhan.github.io/ReflectanceMM/index.html}{\textcolor{magenta}{here}}. | ['Feng Xu', 'Zhibo Wang', 'Yuxuan Han'] | 2023-03-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Han_Learning_a_3D_Morphable_Face_Reflectance_Model_From_Low-Cost_Data_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Han_Learning_a_3D_Morphable_Face_Reflectance_Model_From_Low-Cost_Data_CVPR_2023_paper.pdf | cvpr-2023-1 | ['face-model', 'inverse-rendering'] | ['computer-vision', 'computer-vision'] | [ 2.14746550e-01 -3.52806449e-02 1.78202152e-01 -6.47632957e-01
-3.27055871e-01 -3.59991878e-01 2.81132251e-01 -1.10905051e+00
3.45676094e-01 6.27833664e-01 4.94023226e-03 -8.22546110e-02
1.42782331e-01 -8.98034513e-01 -5.54681957e-01 -8.74092817e-01
5.26414692e-01 1.76249281e-01 -1.48177221e-01 -2.17478067e-01
-1.29713595e-01 8.24718237e-01 -1.86535740e+00 3.57764453e-01
1.07627416e+00 9.89237428e-01 6.52461722e-02 5.38983703e-01
-2.50030547e-01 5.13269424e-01 -3.61363173e-01 -4.60747153e-01
6.25392973e-01 -6.08201563e-01 -2.39381880e-01 2.03121141e-01
9.44120884e-01 -7.48059630e-01 -4.36703488e-02 1.02203608e+00
5.66152334e-01 1.46441519e-01 6.58246219e-01 -1.08697867e+00
-1.15776682e+00 -1.98341832e-01 -9.96747613e-01 -4.81847882e-01
3.55120569e-01 1.84503093e-01 6.43315852e-01 -1.00492740e+00
5.37876785e-01 1.51559496e+00 4.13320124e-01 1.00874436e+00
-1.16719627e+00 -9.72282469e-01 8.22333246e-02 -1.63470492e-01
-1.33063781e+00 -6.77483380e-01 1.21242440e+00 -1.83307573e-01
2.48587504e-01 7.40195870e-01 8.42251003e-01 9.94694471e-01
1.00459447e-02 3.85650396e-01 1.53681612e+00 -2.25032791e-01
-7.70094171e-02 9.47002321e-02 -2.61837304e-01 1.02758765e+00
2.39405289e-01 2.94953465e-01 -6.06264472e-01 -3.67252082e-01
1.08303809e+00 9.74596888e-02 -4.59479421e-01 -5.44156507e-03
-4.63723391e-01 4.44528431e-01 3.31865937e-01 -3.02653611e-01
-2.04219565e-01 1.03521004e-01 -4.53831822e-01 4.81625408e-01
9.85327899e-01 8.09307992e-02 -4.42714900e-01 2.53633946e-01
-6.83446109e-01 2.54129916e-01 5.05220115e-01 9.37023878e-01
1.28315806e+00 3.22247624e-01 3.29050183e-01 1.36629236e+00
6.78004265e-01 1.11537182e+00 -9.39010382e-02 -1.35193348e+00
-3.43216844e-02 4.86730844e-01 2.35422567e-01 -8.39649677e-01
-1.63367704e-01 -6.07393682e-02 -5.61491966e-01 6.26890182e-01
1.64227113e-01 -5.54199666e-02 -1.11701834e+00 1.48590660e+00
8.30289721e-01 2.69209862e-01 -2.60482550e-01 1.21552646e+00
1.20185316e+00 6.37628496e-01 -3.42268825e-01 -2.70541608e-01
1.10475576e+00 -6.84230149e-01 -7.51435459e-01 1.40636861e-01
1.41653776e-01 -1.14508355e+00 1.41368353e+00 4.49446023e-01
-1.14691067e+00 -3.82219851e-01 -6.59793973e-01 -3.96500826e-01
1.86294585e-01 3.68935429e-02 7.54666150e-01 9.57555830e-01
-1.15503240e+00 2.07252622e-01 -7.74367988e-01 -4.68053157e-03
5.77867568e-01 1.84719220e-01 -1.01113886e-01 -4.05575633e-01
-8.73849273e-01 4.39075798e-01 -5.48762798e-01 3.44242424e-01
-6.96269870e-01 -1.04517257e+00 -6.85285568e-01 -5.24826229e-01
-2.84701940e-02 -8.53575289e-01 9.10771072e-01 -1.42456746e+00
-2.23084641e+00 1.07807600e+00 -5.24437129e-01 7.77773082e-01
5.19439340e-01 -6.69398084e-02 -4.62931335e-01 9.97881964e-02
-4.21777070e-01 3.76837999e-01 9.93024886e-01 -2.05551362e+00
-1.85862265e-03 -6.35674179e-01 1.50047749e-01 2.34139308e-01
-2.07785472e-01 1.91433936e-01 -4.37820733e-01 -4.76102531e-01
2.68811285e-01 -9.19324994e-01 2.80997872e-01 6.04270041e-01
-2.91968435e-01 3.03110301e-01 8.55820656e-01 -7.78086424e-01
8.29795539e-01 -2.04063964e+00 -1.96705818e-01 2.39615098e-01
1.40066294e-03 7.00803325e-02 -4.31695223e-01 1.03921428e-01
-1.76162869e-02 1.73125751e-02 -2.14610592e-01 -4.25788045e-01
-6.94458410e-02 3.27376090e-02 -2.04450637e-01 6.97198927e-01
-3.57448384e-02 7.93994427e-01 -5.16083479e-01 -4.12561238e-01
9.96624976e-02 1.20000184e+00 -5.77867448e-01 3.74457926e-01
-2.38652200e-01 7.62492776e-01 -4.99065876e-01 1.14344251e+00
1.44310558e+00 1.50367349e-01 -9.92390066e-02 -2.38751173e-01
-8.07996616e-02 -1.83688775e-01 -1.02390265e+00 1.54322243e+00
-7.14883804e-01 3.93020183e-01 6.10378206e-01 -3.20735537e-02
1.29443169e+00 -6.03725612e-02 4.58049834e-01 -7.33445466e-01
1.87020615e-01 2.39800856e-01 -3.83353025e-01 -5.07544935e-01
2.02028245e-01 -4.96432573e-01 6.41241312e-01 3.60612035e-01
-5.04010558e-01 -6.18853807e-01 -5.98699212e-01 -4.03299928e-01
5.07010639e-01 6.82840943e-01 -2.92430580e-01 -3.00534517e-01
4.65332329e-01 -4.31115001e-01 5.42879820e-01 -4.46240827e-02
2.73918331e-01 9.23582613e-01 1.79786444e-01 -4.87411797e-01
-5.77172399e-01 -1.26162374e+00 -4.51636255e-01 1.16331995e+00
2.02080727e-01 5.04213907e-02 -8.90933096e-01 -1.54495358e-01
7.93229789e-02 6.94911838e-01 -7.85216391e-01 4.35168520e-02
-5.12506008e-01 -7.59387493e-01 2.30733305e-01 -1.36122867e-01
6.00134671e-01 -8.75245810e-01 -3.49581748e-01 -5.21192491e-01
-3.62496823e-02 -7.19751418e-01 -4.59998161e-01 -7.70897329e-01
-7.93808401e-01 -9.10591006e-01 -7.30100870e-01 -4.21900809e-01
1.05679941e+00 5.05044162e-01 9.61667120e-01 3.22832942e-01
-5.38937390e-01 5.89264750e-01 -1.94768339e-01 -5.94091475e-01
-2.60107994e-01 -5.61706424e-01 -2.22435787e-01 4.86465454e-01
5.84729575e-02 -7.12405503e-01 -1.00778127e+00 6.05616748e-01
-1.01604819e+00 3.06881636e-01 2.13820464e-03 3.51256102e-01
7.55195975e-01 -4.39273953e-01 1.93067595e-01 -1.08940971e+00
3.24369639e-01 -2.08620518e-01 -8.71543288e-01 7.44866952e-02
-4.06802118e-01 -2.65284121e-01 4.30173129e-01 -5.17332792e-01
-1.75404429e+00 -1.87583685e-01 -9.57286805e-02 -6.75484419e-01
-1.81335926e-01 -6.03141636e-02 -5.54479182e-01 -3.06335926e-01
6.41739428e-01 9.41576287e-02 1.27686352e-01 -6.70602202e-01
3.12234819e-01 6.01201713e-01 2.16667801e-02 -7.85345197e-01
1.08790410e+00 1.05630863e+00 2.42801711e-01 -1.21776319e+00
-8.28090072e-01 4.82055061e-02 -3.97716999e-01 -5.03517687e-01
4.96968925e-01 -8.62167537e-01 -8.39111447e-01 7.47892141e-01
-1.03712332e+00 -7.93772936e-01 -2.37155601e-01 1.54401422e-01
-5.17012239e-01 2.28724986e-01 -4.37793702e-01 -9.89986479e-01
-2.76330262e-01 -8.87065887e-01 1.24582136e+00 2.51422524e-01
2.96125561e-01 -7.68907070e-01 2.37557054e-01 7.43109524e-01
4.68077391e-01 5.56063831e-01 6.56040013e-01 6.40071511e-01
-7.06586719e-01 1.53532967e-01 -3.40963781e-01 4.38988864e-01
6.11806691e-01 4.98343855e-01 -1.44318104e+00 -1.47098884e-01
-6.35020956e-02 -1.66527852e-01 8.06008101e-01 3.55913192e-01
1.50076842e+00 -3.63562316e-01 1.18776225e-01 1.27860332e+00
1.37905216e+00 -4.45994996e-02 8.03647220e-01 -1.07399300e-01
1.05608964e+00 9.03555453e-01 6.14888966e-01 6.88487411e-01
4.53434855e-01 5.91071427e-01 6.91153765e-01 -3.54100913e-01
-6.80686831e-01 -1.08050682e-01 4.25558656e-01 5.85972369e-01
-7.16838181e-01 -1.80751294e-01 -5.09297490e-01 -2.38774940e-02
-1.23282921e+00 -9.30331349e-01 -3.38462681e-01 2.25279069e+00
1.10710394e+00 -7.46565282e-01 -2.89937139e-01 -4.63104278e-01
4.99496341e-01 1.82489783e-01 -5.90633810e-01 -3.69709432e-01
-3.51651549e-01 3.72090012e-01 2.06921488e-01 9.58607435e-01
-4.25153941e-01 1.05089152e+00 5.03109455e+00 5.05617559e-01
-1.44904959e+00 1.18717648e-01 5.84706128e-01 -4.43788081e-01
-1.26525104e+00 -1.77760571e-01 -7.27427721e-01 3.85522604e-01
6.12616599e-01 2.91443408e-01 8.51163864e-01 3.66524309e-01
7.42658913e-01 -1.43134436e-02 -5.87646723e-01 1.09058964e+00
2.33552709e-01 -8.20016026e-01 3.76119554e-01 2.14681536e-01
1.00703156e+00 -2.11347193e-01 4.13022071e-01 -3.11549425e-01
2.26862222e-01 -1.10452187e+00 6.16758406e-01 1.05016041e+00
1.40327668e+00 -6.06568098e-01 -9.58223119e-02 4.22308929e-02
-9.46881115e-01 3.19373578e-01 -5.97448409e-01 2.74541438e-01
-9.33467075e-02 7.33550549e-01 -3.49738657e-01 3.70555997e-01
6.05598629e-01 5.32976568e-01 -3.35636497e-01 6.53690636e-01
-4.37852949e-01 5.05868196e-01 -3.63692015e-01 1.79312780e-01
-5.55443943e-01 -7.11550772e-01 4.87632066e-01 8.81338358e-01
5.46464801e-01 5.76386929e-01 -2.47181550e-01 1.13627017e+00
-2.44167432e-01 2.15571642e-01 -4.79879528e-01 4.77585644e-01
2.31060237e-01 1.33243680e+00 -1.72385231e-01 2.41181388e-01
-3.38989615e-01 1.00774431e+00 1.27612129e-01 8.66729140e-01
-8.77774298e-01 -1.70222819e-02 1.02108836e+00 6.36324286e-01
-1.66287273e-01 -2.24578798e-01 -2.78833032e-01 -1.27409995e+00
-5.67896198e-03 -7.54870176e-01 -2.44338494e-02 -1.18971729e+00
-1.18189394e+00 4.17868525e-01 3.60617600e-02 -8.95063102e-01
3.48062277e-01 -6.78715885e-01 -3.52394313e-01 1.20436978e+00
-2.23297811e+00 -1.74664474e+00 -8.80171537e-01 9.41298783e-01
2.46619880e-01 8.73210579e-02 8.81037652e-01 2.29423225e-01
-4.58223075e-01 4.90067035e-01 -2.12117992e-02 -2.73362130e-01
9.89842296e-01 -7.89266884e-01 -1.42884767e-02 3.99451405e-01
-2.63831079e-01 7.44046032e-01 5.38751125e-01 -5.92469215e-01
-1.90403080e+00 -1.04396248e+00 1.59122169e-01 -3.99158716e-01
2.40476042e-01 -4.30970252e-01 -9.13689673e-01 5.04416049e-01
1.43034924e-02 4.09175307e-01 8.05832803e-01 -1.74938068e-01
-7.90686846e-01 -5.63455999e-01 -1.45466781e+00 7.86774337e-01
1.32684326e+00 -2.93157339e-01 6.05270714e-02 2.59472936e-01
5.07115543e-01 -4.67466325e-01 -8.58252347e-01 4.13061827e-01
9.82375562e-01 -1.20533061e+00 8.56551826e-01 -1.79345176e-01
4.12457049e-01 -3.53028536e-01 -2.67618537e-01 -1.18758678e+00
-9.59298313e-02 -1.03136063e+00 -1.46950167e-02 1.18086290e+00
3.15316409e-01 -1.06022596e+00 6.78304434e-01 8.08404326e-01
-1.95710286e-01 -7.00279355e-01 -6.17661059e-01 -4.44156289e-01
6.05403110e-02 -2.76900411e-01 1.05425107e+00 1.02216446e+00
-8.68080199e-01 -3.63998771e-01 -3.59806716e-01 3.27274591e-01
8.63279164e-01 5.09863973e-01 1.09503746e+00 -1.45304239e+00
-8.12000856e-02 -1.71764508e-01 2.35061288e-01 -9.57476854e-01
4.69535559e-01 -7.64634192e-01 -9.57277343e-02 -1.43959367e+00
2.12913156e-01 -6.98308110e-01 1.52969519e-02 5.75539827e-01
-9.77924746e-03 6.16781116e-01 -1.15553170e-01 2.37935424e-01
1.09632239e-01 7.23306477e-01 1.95305777e+00 9.95002538e-02
-3.29433709e-01 3.63526237e-03 -7.25970507e-01 7.39469945e-01
9.38008726e-01 -1.62837982e-01 -6.23360753e-01 -8.15402806e-01
3.69761407e-01 -1.46175504e-01 5.62223375e-01 -4.31330860e-01
-4.38989222e-01 -8.77536297e-01 4.07383859e-01 -1.15563236e-01
8.48082781e-01 -8.72049630e-01 5.57861447e-01 -9.51338783e-02
1.87251028e-02 -3.98980111e-01 2.23834679e-01 2.27976128e-01
3.23703498e-01 6.54078573e-02 1.14489520e+00 -2.18252093e-01
-1.17412046e-01 7.78814077e-01 2.20820710e-01 -1.15928933e-01
6.95448697e-01 -4.71779317e-01 -5.83029866e-01 -3.95010233e-01
-3.05121154e-01 -3.75760227e-01 1.07974446e+00 2.91432917e-01
7.81488478e-01 -1.19816184e+00 -9.63284552e-01 5.05622923e-01
6.78296387e-02 -3.87523249e-02 4.94176209e-01 4.13703471e-01
-7.69875765e-01 -2.81654179e-01 -1.62991062e-01 -3.46449345e-01
-1.51669490e+00 -1.49903558e-02 6.67048752e-01 6.08202457e-01
-3.01361680e-01 9.21948969e-01 5.41845977e-01 -6.58788621e-01
-3.19022626e-01 -2.16303110e-01 1.59307450e-01 -8.79171640e-02
4.93822962e-01 5.44500172e-01 -1.38301831e-02 -7.67774880e-01
-1.03090324e-01 1.09637284e+00 5.01081049e-01 2.98807044e-02
1.41801775e+00 -2.61444092e-01 -6.29869580e-01 3.53745162e-01
1.16615510e+00 5.93789876e-01 -1.48711824e+00 1.10426448e-01
-1.03606558e+00 -9.88684118e-01 1.33805946e-01 -8.38199258e-01
-1.45148933e+00 8.18968773e-01 4.14116889e-01 -1.34067789e-01
1.45593834e+00 -9.46498886e-02 6.03787899e-01 4.01887931e-02
3.76431376e-01 -7.02650785e-01 -2.55266614e-02 1.89943120e-01
1.33128917e+00 -9.86521780e-01 2.79794157e-01 -8.49101305e-01
-3.52281690e-01 1.10628724e+00 6.66111767e-01 3.46460454e-02
9.33910429e-01 3.42679530e-01 4.76402014e-01 -2.11183563e-01
-4.35274512e-01 4.26888615e-02 4.58520234e-01 7.26186693e-01
4.29168403e-01 1.36067584e-01 -1.01164103e-01 9.17894840e-02
-3.46142501e-01 -1.01530403e-01 5.13298810e-01 4.25652057e-01
-2.41215974e-01 -1.09616899e+00 -4.97587889e-01 2.69853830e-01
-3.21481764e-01 -9.33644325e-02 -4.12772745e-01 4.88969356e-01
1.35007247e-01 8.88979852e-01 -1.12582415e-01 -4.33245786e-02
3.16075802e-01 -1.75566480e-01 7.82585323e-01 -5.80033123e-01
-1.55894920e-01 3.58726531e-01 -1.06853740e-02 -6.70069695e-01
-5.78951716e-01 -6.17457509e-01 -1.09171176e+00 -5.20706356e-01
-2.25136131e-01 -3.79173905e-01 7.52587259e-01 1.24402702e-01
3.84904563e-01 1.17057197e-01 9.64785874e-01 -8.76005530e-01
-2.33901218e-02 -6.82949960e-01 -8.14501345e-01 2.92834252e-01
4.89255220e-01 -7.68758953e-01 -5.71309447e-01 2.77545601e-01] | [12.693463325500488, -0.3595455586910248] |
1d2fddeb-686e-42d0-9823-35c5363da6b1 | exploring-font-independent-features-for-scene | 2009.07447 | null | https://arxiv.org/abs/2009.07447v1 | https://arxiv.org/pdf/2009.07447v1.pdf | Exploring Font-independent Features for Scene Text Recognition | Scene text recognition (STR) has been extensively studied in last few years. Many recently-proposed methods are specially designed to accommodate the arbitrary shape, layout and orientation of scene texts, but ignoring that various font (or writing) styles also pose severe challenges to STR. These methods, where font features and content features of characters are tangled, perform poorly in text recognition on scene images with texts in novel font styles. To address this problem, we explore font-independent features of scene texts via attentional generation of glyphs in a large number of font styles. Specifically, we introduce trainable font embeddings to shape the font styles of generated glyphs, with the image feature of scene text only representing its essential patterns. The generation process is directed by the spatial attention mechanism, which effectively copes with irregular texts and generates higher-quality glyphs than existing image-to-image translation methods. Experiments conducted on several STR benchmarks demonstrate the superiority of our method compared to the state of the art. | ['Zhouhui Lian', 'Yizhi Wang'] | 2020-09-16 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 5.87999940e-01 -6.61051512e-01 1.25112861e-01 -3.93420517e-01
-1.15292631e-01 -6.93623006e-01 8.31858814e-01 -2.30292261e-01
-1.73017412e-01 3.09528232e-01 4.58025485e-01 -2.37731323e-01
3.19918036e-01 -7.00806558e-01 -7.77913332e-01 -6.29794955e-01
6.65824890e-01 3.01356643e-01 2.11923808e-01 -4.66518581e-01
6.15100980e-01 5.91204703e-01 -1.39293993e+00 6.87510729e-01
8.51849675e-01 7.12268949e-01 4.36439812e-01 6.97254539e-01
-8.14813435e-01 5.67440271e-01 -7.59451032e-01 -3.25416118e-01
3.67635816e-01 -3.51245046e-01 -2.98172474e-01 7.12760150e-01
9.27143574e-01 -2.45285943e-01 -6.80818141e-01 1.08187962e+00
5.17598331e-01 -1.15306176e-01 7.83448577e-01 -7.14412868e-01
-1.64582264e+00 7.70057619e-01 -8.34826767e-01 6.39422461e-02
4.25090462e-01 2.89300829e-01 6.69299603e-01 -1.43553877e+00
6.00504935e-01 1.43737400e+00 1.96665719e-01 5.35674810e-01
-1.16532016e+00 -4.86962318e-01 2.79058099e-01 3.75692621e-02
-1.41499770e+00 -2.49953210e-01 1.07408893e+00 -4.05425787e-01
7.40328014e-01 5.62081337e-01 4.92396355e-01 1.44461405e+00
1.92079663e-01 1.08733130e+00 1.09423840e+00 -7.25356758e-01
-1.46613032e-01 3.27189088e-01 -1.05908610e-01 5.01576066e-01
1.41068816e-01 -3.74900132e-01 -6.47079170e-01 3.22505981e-01
1.00727165e+00 3.68992537e-02 -4.04840380e-01 -5.67854464e-01
-1.64356112e+00 6.21681631e-01 1.45871863e-01 1.62477702e-01
2.08599612e-01 -1.51208654e-01 4.92221564e-01 3.22243035e-01
3.84897768e-01 3.99203539e-01 -1.72892272e-01 -4.58286935e-03
-7.79648781e-01 7.41991922e-02 3.84875596e-01 1.44839871e+00
3.14960629e-01 5.00132143e-01 -5.15495777e-01 1.10608482e+00
-3.37140448e-02 9.63508546e-01 7.85620511e-01 -2.07898617e-02
1.17115438e+00 8.74433637e-01 -1.13449320e-01 -1.24483943e+00
-2.14388549e-01 -6.75563887e-02 -1.03107178e+00 -2.18291596e-01
2.68443227e-01 6.93648532e-02 -8.95278633e-01 1.19364440e+00
6.82381094e-02 -5.33723533e-01 -1.12242065e-01 8.93676281e-01
7.33856738e-01 8.16649020e-01 -4.47510242e-01 4.17956300e-02
1.23000133e+00 -1.27140164e+00 -9.86825764e-01 -3.06612760e-01
4.51398104e-01 -1.26418972e+00 1.84779251e+00 2.82302737e-01
-1.01847625e+00 -9.61412132e-01 -1.01465487e+00 -2.39252537e-01
-6.13711238e-01 5.65932512e-01 1.94696978e-01 8.59671772e-01
-8.72294366e-01 1.66400149e-01 -1.41926453e-01 -4.95053560e-01
3.65141392e-01 5.26445322e-02 -5.90162575e-02 -8.77546147e-02
-8.78143787e-01 5.01964569e-01 3.15697879e-01 1.44858718e-01
-2.03117982e-01 -3.26466829e-01 -7.82467365e-01 3.29691172e-02
2.25920796e-01 -3.78280252e-01 7.87054300e-01 -1.30207181e+00
-1.59109533e+00 7.09728718e-01 5.43805920e-02 9.37498286e-02
8.64026487e-01 -2.43199334e-01 -6.52144611e-01 1.26706451e-01
-2.67600808e-02 7.75737643e-01 1.56958067e+00 -1.06966591e+00
-2.05184981e-01 -2.50408083e-01 -4.05424237e-01 5.96677244e-01
-1.10514021e+00 2.63608128e-01 -6.31680012e-01 -1.17990828e+00
2.36420818e-02 -6.01114571e-01 8.68147984e-02 2.74111956e-01
-5.63321471e-01 1.80344637e-02 1.27774274e+00 -3.31416249e-01
1.21283305e+00 -2.14372826e+00 2.16804579e-01 -1.52702421e-01
-6.86666295e-02 2.58086175e-01 -3.41063768e-01 6.52697742e-01
-3.56487297e-02 1.00849830e-01 1.07546404e-01 -2.16998085e-01
1.31939456e-01 -1.76642668e-02 -8.17892253e-01 1.82884365e-01
4.15731609e-01 1.11936593e+00 -4.90336567e-01 -7.21734822e-01
5.02790332e-01 2.42766276e-01 -2.98015594e-01 1.10925697e-01
-3.41228873e-01 3.41694206e-01 -6.66679204e-01 6.83944225e-01
7.88651049e-01 -3.14364612e-01 3.13761383e-02 -1.95546955e-01
-2.15530872e-01 -1.20173842e-01 -9.35057461e-01 1.61191010e+00
-2.23124668e-01 9.42796767e-01 -4.84719306e-01 -5.52406073e-01
1.35248220e+00 -1.30430683e-01 -2.66369041e-02 -9.65508282e-01
2.25752875e-01 4.39883443e-03 -2.20669851e-01 -6.92154825e-01
1.06952226e+00 4.83892530e-01 -1.17227949e-01 4.97169882e-01
-5.57455838e-01 -3.32950741e-01 2.60695040e-01 1.72550648e-01
5.37120581e-01 1.77568391e-01 -8.73432904e-02 -3.10071260e-01
5.26371479e-01 -1.31665558e-01 -1.52406573e-01 8.54551017e-01
8.42685923e-02 1.03752720e+00 3.69537264e-01 -7.07936347e-01
-1.63423312e+00 -8.50217283e-01 -2.24870399e-01 1.05689645e+00
2.71990925e-01 -1.84831187e-01 -7.17436075e-01 -5.86290896e-01
-9.40697566e-02 5.55009604e-01 -7.63114333e-01 9.80962366e-02
-9.38683748e-01 -6.26061559e-01 4.81824219e-01 6.84267104e-01
6.90423667e-01 -1.23218453e+00 -6.19403064e-01 3.09166536e-02
2.80381236e-02 -1.51228786e+00 -1.23432195e+00 -1.64002433e-01
-7.58357108e-01 -6.26274586e-01 -1.10402560e+00 -1.07519126e+00
1.05284607e+00 6.29053175e-01 8.15496266e-01 -1.68266624e-01
-5.16483784e-01 3.91296118e-01 -6.58526123e-01 -1.17868520e-01
-3.46855313e-01 -4.40518484e-02 -3.80665772e-02 4.94277656e-01
2.19586134e-01 -9.10533518e-02 -4.88115072e-01 4.13320899e-01
-1.30252075e+00 6.11087441e-01 5.57234943e-01 9.78295982e-01
2.51427084e-01 -2.06280559e-01 7.32917264e-02 -1.05384278e+00
9.21345949e-01 3.99227196e-04 -5.24836898e-01 4.89636779e-01
-3.01751435e-01 1.58883035e-02 1.25191510e+00 -9.64639068e-01
-9.78215456e-01 2.55182758e-02 3.75865757e-01 -5.85600734e-01
-1.66995481e-01 7.49170259e-02 -2.43153065e-01 -7.46908486e-02
6.76909626e-01 1.09526217e+00 -3.80466759e-01 -5.15813231e-01
4.03188884e-01 8.85756314e-01 2.33744770e-01 -7.72303045e-01
1.01880217e+00 5.00646472e-01 -2.19999224e-01 -1.14317393e+00
-4.92125362e-01 -1.68096256e-02 -9.31036294e-01 -1.77258581e-01
6.73305988e-01 -5.94437122e-01 -1.38293818e-01 6.40250206e-01
-1.18944848e+00 -3.03965300e-01 -2.69699752e-01 5.68953417e-02
-4.58139181e-01 8.85839939e-01 -4.64429498e-01 -5.26995659e-01
-2.62535691e-01 -1.19204509e+00 1.50350070e+00 1.73762515e-01
7.69381076e-02 -8.59055042e-01 -2.07857564e-01 1.20527588e-01
4.57322478e-01 3.31538580e-02 1.28985763e+00 -3.27855885e-01
-7.86966503e-01 4.87432890e-02 -7.66090930e-01 3.56228985e-02
3.19557458e-01 1.95329428e-01 -8.02918077e-01 -2.63448030e-01
-2.21438810e-01 -2.32483700e-01 7.00277805e-01 6.66707233e-02
1.57097220e+00 -4.04573143e-01 -2.91521624e-02 7.90239453e-01
1.28530991e+00 1.11625925e-01 6.32749557e-01 4.54265118e-01
1.09848678e+00 4.65812296e-01 4.89065289e-01 5.87253988e-01
1.13013998e-01 8.63487184e-01 1.57400638e-01 -3.36199313e-01
-2.91369349e-01 -4.71169561e-01 5.09976327e-01 1.01500475e+00
4.55026805e-01 -5.73585987e-01 -6.38280034e-01 2.52864540e-01
-1.65134799e+00 -6.87918484e-01 -3.67866963e-01 1.98247504e+00
9.11527336e-01 1.68648064e-01 -1.45395979e-01 1.85842831e-02
8.51903319e-01 5.88979125e-01 -5.40897429e-01 -7.45866239e-01
-8.44318330e-01 -3.52626517e-02 5.50897181e-01 -1.56481758e-01
-1.05351579e+00 1.25829399e+00 6.32897663e+00 1.11008954e+00
-1.21537316e+00 -3.86116296e-01 5.87551117e-01 7.12797865e-02
-3.44197422e-01 -2.68728137e-01 -1.01366532e+00 6.06235445e-01
3.18192989e-01 7.22842067e-02 6.02668881e-01 6.52827621e-01
2.73877442e-01 3.44958127e-01 -1.08442819e+00 1.15384269e+00
7.12659478e-01 -1.26283586e+00 7.25288928e-01 -2.30250731e-01
1.13399196e+00 -3.98233801e-01 6.48627937e-01 6.37031198e-02
-3.88761945e-02 -1.00169551e+00 1.01005936e+00 3.92001390e-01
1.34228098e+00 -3.71597677e-01 8.11275244e-02 1.45331725e-01
-1.17494953e+00 -7.07426891e-02 -8.44678998e-01 1.70064151e-01
-1.04560450e-01 2.93825239e-01 -7.05972254e-01 3.39143842e-01
6.08236253e-01 8.24566662e-01 -1.15711033e+00 6.40115857e-01
-1.07280491e-02 2.97644049e-01 -2.44415347e-02 -6.07103407e-01
4.88664925e-01 -2.51600713e-01 3.96752983e-01 1.44327891e+00
3.98123085e-01 -3.66809994e-01 1.40493944e-01 1.11955094e+00
-2.04334974e-01 6.74068928e-01 -9.28145885e-01 -2.85652131e-01
2.35601366e-01 1.12481511e+00 -8.05354476e-01 -3.58709037e-01
-6.36333406e-01 1.36875427e+00 1.94513455e-01 5.59375286e-01
-7.80802786e-01 -5.69103539e-01 1.11669190e-01 6.01562187e-02
4.64510471e-01 -4.87772644e-01 -6.22661829e-01 -1.36660302e+00
4.78076339e-01 -1.18228030e+00 -2.56547593e-02 -1.17087233e+00
-1.25049067e+00 7.64154971e-01 -2.86228061e-01 -1.57572770e+00
3.59476119e-01 -1.02820373e+00 -6.20971441e-01 7.47919261e-01
-1.43121159e+00 -1.49296677e+00 -3.43259841e-01 6.82872057e-01
1.45778298e+00 -4.63787794e-01 4.83591676e-01 6.77402988e-02
-6.41944826e-01 8.84821236e-01 7.16002643e-01 2.82941997e-01
9.80361581e-01 -1.09732127e+00 8.43474507e-01 8.27306032e-01
3.80124003e-01 5.72488904e-01 4.45906371e-01 -7.16365099e-01
-1.78647172e+00 -1.33294630e+00 7.40177572e-01 -5.26776016e-01
7.28884399e-01 -9.12687004e-01 -8.91815007e-01 3.88093621e-01
4.68718678e-01 -2.32268453e-01 2.72711873e-01 -3.68726283e-01
-7.54707992e-01 -1.65251624e-02 -5.08493185e-01 1.33332610e+00
1.03626466e+00 -3.31288040e-01 -4.05717909e-01 6.23383939e-01
5.95255196e-01 -6.36469603e-01 -3.34792674e-01 1.16129950e-01
4.66016948e-01 -9.10706341e-01 8.82841229e-01 -5.56640685e-01
6.13870740e-01 -2.89565802e-01 -1.56056255e-01 -9.56717491e-01
-2.65910983e-01 -6.72264159e-01 2.38253489e-01 1.08275199e+00
2.91992724e-01 -4.65643317e-01 6.02704167e-01 2.57852614e-01
-1.13464035e-01 -4.56004620e-01 -5.46383500e-01 -7.19649792e-01
1.97938904e-01 1.03066131e-01 7.87144423e-01 1.07869649e+00
-1.13573380e-01 2.32135966e-01 -7.08481133e-01 -1.93012655e-01
2.94279963e-01 4.55522031e-01 1.00429034e+00 -7.32792020e-01
-1.11800306e-01 -7.13523090e-01 -2.05891550e-01 -1.49947202e+00
-1.04843467e-01 -7.31699288e-01 -2.31082290e-01 -1.05044866e+00
3.17309678e-01 -1.46712333e-01 2.42631927e-01 2.05464616e-01
-2.55893141e-01 1.99704915e-01 4.19725180e-01 3.42196435e-01
-5.88032782e-01 8.93153250e-01 1.89937711e+00 -5.45639634e-01
-7.25997835e-02 -4.25389469e-01 -5.91911972e-01 4.60442185e-01
8.21354628e-01 -4.80651632e-02 -4.54570293e-01 -1.05480886e+00
3.12769264e-01 -2.78798163e-01 1.09404683e-01 -8.20448697e-01
2.13276431e-01 -3.90455306e-01 8.50356460e-01 -9.60279763e-01
2.56137818e-01 -7.42649317e-01 -2.77984023e-01 3.02842092e-02
-7.44237006e-01 5.45652747e-01 2.15609357e-01 5.56707978e-01
3.68599519e-02 -2.25164995e-01 7.36177564e-01 -2.49520335e-02
-3.71941477e-01 1.80963397e-01 -4.46023703e-01 1.32351309e-01
7.54641116e-01 -7.20782757e-01 -4.67717618e-01 -1.47686005e-01
-1.53513849e-01 4.52271057e-03 7.57603645e-01 9.30023789e-01
9.31287587e-01 -1.46918964e+00 -7.70006716e-01 6.65995657e-01
3.82133782e-01 -1.34891734e-01 2.35874251e-01 4.93540913e-01
-7.54248083e-01 5.47685623e-01 -3.06773901e-01 -6.72578633e-01
-1.19063306e+00 8.80265474e-01 -1.11853480e-01 -6.15119115e-02
-9.90251899e-01 5.11512280e-01 6.24019206e-01 -3.20498288e-01
1.85638517e-01 -4.17319447e-01 -2.41568044e-01 -3.24232727e-01
4.74439651e-01 -3.45391184e-02 -1.49834111e-01 -6.01904333e-01
3.29620391e-01 1.15129685e+00 -5.61699331e-01 1.26489073e-01
9.26797271e-01 -3.86415303e-01 1.83334962e-01 4.21687543e-01
9.08685684e-01 1.92093998e-01 -1.29771698e+00 -4.55504894e-01
-1.72842115e-01 -9.59228873e-01 -4.24314737e-01 -6.79417670e-01
-9.18906510e-01 1.36489534e+00 4.76268798e-01 1.61212444e-01
1.02463651e+00 -6.54103875e-01 7.26472318e-01 5.29347420e-01
1.10503756e-01 -1.39415109e+00 5.54953873e-01 5.98839283e-01
1.15473723e+00 -1.15668440e+00 -1.40902728e-01 -3.76262188e-01
-9.87763345e-01 1.63349199e+00 8.62289190e-01 -3.81433265e-03
4.40340377e-02 2.26217642e-01 1.65256307e-01 8.84686261e-02
-5.01576781e-01 1.62153810e-01 3.48870635e-01 5.75717509e-01
3.72180074e-01 -5.10806292e-02 -3.84067893e-02 1.07556373e-01
-1.17968641e-01 -5.12783527e-01 6.70452476e-01 7.33177543e-01
-4.16349113e-01 -1.16994119e+00 -6.24278724e-01 4.84969109e-01
-8.30529481e-02 -4.85004246e-01 -7.76561677e-01 6.80885494e-01
-2.68199891e-01 5.16205728e-01 4.75412123e-02 -2.23688707e-01
3.91324401e-01 3.20951594e-03 4.77913946e-01 -5.12522936e-01
-5.86971998e-01 4.30121094e-01 -3.75288814e-01 -1.72149435e-01
5.72818071e-02 -6.41751587e-01 -7.66430914e-01 -1.01302028e-01
-3.29295963e-01 -3.23680401e-01 5.13959348e-01 4.47497606e-01
3.49738657e-01 6.01444721e-01 1.04820716e+00 -7.58870363e-01
-7.08344579e-01 -8.70828509e-01 -5.21499157e-01 5.72131932e-01
-6.66575357e-02 -3.16169441e-01 -8.49321932e-02 3.25163782e-01] | [11.908699035644531, 1.8202598094940186] |
515db804-a856-4fe5-a35c-840d68cd405d | resolving-head-on-conflicts-for-multi-agent | 2007.03575 | null | https://arxiv.org/abs/2007.03575v1 | https://arxiv.org/pdf/2007.03575v1.pdf | Resolving Head-On Conflicts for Multi-Agent Path Finding with Conflict-Based Search | Conflict-Based Search (CBS) is a popular framework for solving the Multi-Agent Path Finding problem. Some of the conflicts incur a foreseeable conflict in one or both of the children nodes when splitting on them. This paper introduces a new technique, namely the head-on technique that finds out such conflicts, so they can be processed more efficiently by resolving the conflict with the potential conflict all together in one split. The proposed technique applies to all CBS-based solvers. Experimental results show that the head-on technique improves the state-of-the-art MAPF solver CBSH. | ['Lun Yang'] | 2020-07-07 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 4.93873656e-02 4.11045939e-01 -4.14858490e-01 -1.08359613e-01
-2.33108416e-01 -4.73937422e-01 3.33842069e-01 3.99081767e-01
-3.40517282e-01 1.60001421e+00 -1.98825106e-01 -3.36426616e-01
-6.79822803e-01 -1.12018573e+00 -1.86083987e-01 -6.75068200e-01
-5.31490684e-01 1.16905451e+00 1.12464654e+00 -5.86461604e-01
3.79507571e-01 4.03065950e-01 -1.24312389e+00 4.09065753e-01
8.59501481e-01 5.25109887e-01 1.85119912e-01 1.35657683e-01
-4.63735729e-01 6.89248979e-01 -8.06511879e-01 -2.33463004e-01
4.10736978e-01 -3.68397236e-01 -1.36704099e+00 -3.16005051e-01
-1.11625582e-01 -7.88641199e-02 2.58445799e-01 1.01817572e+00
-1.76838017e-03 2.34808903e-02 2.01694205e-01 -2.15897012e+00
4.74471420e-01 1.09941602e+00 -1.01917934e+00 6.13006890e-01
8.11042726e-01 -3.58715177e-01 6.83182836e-01 -2.98806816e-01
1.20071054e+00 1.16040564e+00 1.54690713e-01 2.35871851e-01
-8.02261114e-01 -7.56998479e-01 5.83364427e-01 1.01351380e+00
-1.28392541e+00 -5.09635434e-02 5.39207458e-01 1.68120816e-01
1.47184229e+00 7.39776015e-01 8.94214272e-01 3.07318896e-01
6.91635787e-01 3.23356539e-01 1.44694591e+00 -4.31580991e-01
5.05530059e-01 -8.80239159e-02 1.70747235e-01 4.94681537e-01
7.21615553e-01 2.90406674e-01 -6.28859580e-01 -5.22509098e-01
4.18809295e-01 -5.97297311e-01 -2.06926707e-02 -1.63457379e-01
-9.74269390e-01 1.04621732e+00 3.04000705e-01 3.96301776e-01
-4.48493630e-01 -4.96235639e-02 2.76050061e-01 3.66365552e-01
6.14384227e-02 3.03523481e-01 -4.71146643e-01 1.70851171e-01
-8.78311038e-01 6.39609873e-01 1.05628502e+00 9.29149508e-01
7.24945247e-01 -4.37020183e-01 2.66069502e-01 3.95892896e-02
2.17566103e-01 -3.18606533e-02 -3.27043682e-01 -8.15756202e-01
5.40544093e-01 7.17749894e-01 2.26705596e-01 -1.04564762e+00
-7.88658500e-01 -3.48520190e-01 -4.02132273e-01 7.37406850e-01
4.26179245e-02 -7.11599812e-02 -6.68967485e-01 1.49306917e+00
1.07133901e+00 4.78041738e-01 1.01573490e-01 8.90410364e-01
6.17900550e-01 1.11160219e+00 -1.70130864e-01 -9.47669148e-01
1.19072306e+00 -1.39727759e+00 -6.35309994e-01 -2.26008013e-01
3.43649745e-01 -7.64257610e-01 -2.77219474e-01 6.91229343e-01
-1.39438772e+00 3.44547868e-01 -1.11436856e+00 6.04269743e-01
-4.16823089e-01 -8.83990109e-01 8.84387136e-01 3.42361242e-01
-1.15590191e+00 9.39951316e-02 -6.54890835e-01 -3.85414749e-01
-2.42748708e-01 8.29840183e-01 -5.41080654e-01 -3.34118932e-01
-1.10593510e+00 1.25689638e+00 6.13321960e-01 -1.59144793e-02
-6.66705787e-01 -3.65096480e-01 -4.23838735e-01 3.97169814e-02
1.23575640e+00 -7.34207273e-01 1.09651458e+00 -5.50985396e-01
-1.14993632e+00 6.38366699e-01 -2.40486905e-01 -2.78936267e-01
4.05591160e-01 5.22887886e-01 -5.03173947e-01 1.39747739e-01
4.96591806e-01 3.97441149e-01 8.98512155e-02 -1.37552786e+00
-1.19590533e+00 -2.38676935e-01 3.89754176e-01 1.46876663e-01
4.13302511e-01 5.33040702e-01 -2.44095623e-01 -1.10952772e-01
2.94350922e-01 -8.71220171e-01 -8.18561733e-01 -4.77860004e-01
-3.73704344e-01 -4.53193277e-01 6.48915291e-01 -1.88765004e-01
1.54435742e+00 -1.38636863e+00 5.15072584e-01 6.52723849e-01
-4.17941855e-03 5.99137805e-02 -3.29052746e-01 1.16199577e+00
2.82196030e-02 -5.92187755e-02 -3.66609753e-03 4.20763314e-01
-2.40884960e-01 4.82155800e-01 1.20142691e-01 4.02362347e-01
-1.65142030e-01 1.34783313e-01 -9.27910089e-01 -7.93429077e-01
-8.72859135e-02 -3.10354680e-01 -5.81997573e-01 -5.09694777e-02
-4.10623044e-01 2.94116437e-01 -5.08576691e-01 4.89571333e-01
1.13095462e+00 2.65602559e-01 7.85031676e-01 3.07952702e-01
-8.96363318e-01 4.16265309e-01 -1.92673063e+00 1.31784296e+00
9.78710204e-02 -1.87204499e-03 5.25198221e-01 -9.12733912e-01
6.30962729e-01 2.22796783e-01 6.76961303e-01 -7.89562881e-01
1.94357838e-02 2.73753494e-01 2.10045144e-01 -2.57664263e-01
1.54727131e-01 -1.69957001e-02 -1.30643442e-01 6.66296184e-01
-4.45433170e-01 1.85775474e-01 9.29194272e-01 3.32146019e-01
1.32347977e+00 -2.58911490e-01 6.26990438e-01 -6.60891473e-01
1.00162077e+00 7.12096512e-01 1.29788554e+00 6.21917486e-01
-1.00894250e-01 -4.12548184e-01 5.42423010e-01 -7.54189909e-01
-3.97426754e-01 -6.64472103e-01 3.57522935e-01 8.52931261e-01
7.61640191e-01 -7.08497882e-01 -5.07313907e-01 -5.79283535e-01
-1.34656861e-01 7.66101062e-01 -4.01924461e-01 2.39798933e-01
-1.09503055e+00 -3.63072842e-01 -1.27950814e-02 1.24643251e-01
3.44027132e-01 -1.01804221e+00 -1.17512858e+00 7.64150977e-01
-2.68828332e-01 -7.53229141e-01 1.00612782e-01 1.92083940e-01
-5.67958713e-01 -1.48830283e+00 1.03291906e-01 -9.47401524e-01
5.45383394e-01 5.43749988e-01 8.53033423e-01 6.80307388e-01
-2.12017909e-01 -1.63620055e-01 -5.60174584e-01 -3.57965797e-01
-2.82231450e-01 3.23884636e-02 -1.16272502e-01 -6.17369592e-01
-4.54027578e-03 -5.60923815e-01 -1.92626953e-01 6.74092650e-01
-4.25414920e-01 1.16693862e-01 2.09594220e-01 4.96828973e-01
5.16752422e-01 9.71900165e-01 5.07525623e-01 -6.76024497e-01
6.24300718e-01 -5.71752369e-01 -1.03518796e+00 5.51276386e-01
-6.57907128e-01 -1.22722529e-01 3.46607804e-01 -5.89034371e-02
-1.01142490e+00 -1.64903909e-01 4.36473757e-01 9.96781960e-02
1.16288634e-02 8.79716694e-01 -2.87196646e-03 -3.92422587e-01
-1.66196767e-02 -3.09598565e-01 -4.16882008e-01 -2.07291067e-01
-9.23629701e-02 -5.61721958e-02 2.99233705e-01 -7.12508023e-01
6.43817246e-01 2.48870254e-01 6.79791987e-01 -9.99227241e-02
-8.40850100e-02 -4.08442765e-01 -2.45136425e-01 -4.69789058e-01
3.79508644e-01 -3.47549230e-01 -8.94075572e-01 1.66518927e-01
-1.37198353e+00 -5.86131327e-02 3.22409391e-01 2.88916439e-01
-2.28386909e-01 1.68348923e-01 -5.59612989e-01 -9.95382309e-01
-1.98234528e-01 -1.18482816e+00 1.92834049e-01 5.08977294e-01
-3.43280435e-01 -5.26188374e-01 6.36050284e-01 1.65916607e-01
3.05917144e-01 4.70842719e-01 1.12632406e+00 -7.69880950e-01
-6.50758684e-01 2.53058732e-01 3.72698829e-02 -9.30668592e-01
-2.08467975e-01 2.03686357e-01 1.77782521e-01 -4.45803732e-01
-1.63650259e-01 2.58762151e-01 2.49902055e-01 3.30198973e-01
5.59736080e-02 -5.44263363e-01 -1.00433826e+00 -4.42500189e-02
1.75930297e+00 8.03354859e-01 5.78737736e-01 9.90709543e-01
-2.19260797e-01 9.08963025e-01 1.24400520e+00 5.16132176e-01
5.64707696e-01 9.61920381e-01 8.74248266e-01 6.15907982e-02
2.62351811e-01 3.41167092e-01 1.71554729e-01 4.14193660e-01
-2.79875100e-01 -6.19544983e-01 -1.11685455e+00 4.71945107e-01
-2.36855388e+00 -9.69348311e-01 -7.13972628e-01 1.77067375e+00
7.29825854e-01 3.56753647e-01 3.56563687e-01 3.60253483e-01
1.03670263e+00 -1.37656569e-01 -1.54356167e-01 -1.13864410e+00
8.85877311e-02 2.82080621e-01 4.16359961e-01 9.09530282e-01
-6.12314880e-01 1.06027997e+00 6.79987097e+00 4.33151186e-01
-9.36927021e-01 1.22254878e-01 -1.13743886e-01 -4.59689684e-02
-2.18205690e-01 5.26791751e-01 -6.73395455e-01 7.66423568e-02
5.44901609e-01 -6.82407439e-01 5.83025873e-01 6.01355076e-01
2.00920358e-01 -8.31839263e-01 -8.75975668e-01 2.74187326e-01
-6.15682006e-02 -1.36301124e+00 -1.45650208e-01 2.39483967e-01
6.78605139e-01 -5.05458891e-01 -5.44806361e-01 -1.62702352e-01
3.01747710e-01 -7.60890007e-01 9.09453332e-01 -2.31733933e-01
4.61730026e-02 -1.22473300e+00 9.03896868e-01 5.58518350e-01
-1.34022343e+00 -2.10589826e-01 -2.36451045e-01 -6.37657404e-01
7.09174693e-01 4.04516608e-01 -7.98018515e-01 1.13249946e+00
7.02109873e-01 -6.51603490e-02 2.02811375e-01 1.58092248e+00
-3.88484985e-01 -1.31222472e-01 -6.39232874e-01 -9.46605876e-02
6.53759658e-01 -5.14526308e-01 9.54169154e-01 1.00184107e+00
2.37749398e-01 7.01170623e-01 5.31464934e-01 6.33292317e-01
7.87359595e-01 1.45275071e-02 -1.85020164e-01 4.10776645e-01
8.79694104e-01 9.48140025e-01 -1.11914086e+00 -2.46088654e-01
-3.13118488e-01 4.27851915e-01 9.12682489e-02 -9.61552262e-02
-9.66002584e-01 -1.99687764e-01 4.82402802e-01 6.41390458e-02
2.43553892e-01 -1.33178264e-01 -1.16386332e-01 -4.35254157e-01
-7.16428012e-02 -1.01332450e+00 9.20233965e-01 -5.24362624e-01
-8.29041719e-01 7.38900065e-01 7.20449805e-01 -7.00518906e-01
-2.36064941e-01 -2.92209029e-01 -1.09198272e+00 5.58187425e-01
-1.56446981e+00 -9.16051924e-01 -2.68412586e-02 8.76570284e-01
5.61013401e-01 8.84422809e-02 8.20398450e-01 1.96714401e-01
-7.75310040e-01 4.81241457e-02 -6.54920399e-01 -7.23761797e-01
2.47015908e-01 -6.62592053e-01 -1.48173243e-01 1.23994493e+00
-3.99947226e-01 4.65543807e-01 1.00456882e+00 -9.53126669e-01
-1.37994075e+00 -2.67555386e-01 1.16475403e+00 6.26641154e-01
5.18084288e-01 2.81633437e-01 -4.39936996e-01 6.91756725e-01
7.67463386e-01 -4.75979328e-01 3.88319731e-01 -3.88063304e-02
3.26868296e-02 -7.04920590e-02 -1.61676490e+00 3.93944234e-01
9.86295700e-01 5.53219140e-01 -7.24536955e-01 1.98352903e-01
3.92983317e-01 -3.45839202e-01 -3.68093938e-01 6.28998578e-01
2.64552474e-01 -1.27978015e+00 7.14320898e-01 -6.15976870e-01
1.63280740e-01 -6.66191399e-01 1.38282269e-01 -1.29236758e+00
-6.76347792e-01 -7.41333604e-01 2.55065918e-01 9.86557722e-01
5.72956800e-01 -9.18899179e-01 8.09073150e-01 5.74195087e-01
-1.83035672e-01 -7.81900883e-01 -1.67345977e+00 -7.55598009e-01
-2.05575213e-01 1.70144781e-01 1.14073849e+00 1.23315251e+00
8.03202569e-01 -3.34821530e-02 -1.88277721e-01 4.85124648e-01
7.46759474e-01 5.75296640e-01 4.49961483e-01 -1.38472927e+00
-2.47531980e-01 -4.33911562e-01 -2.22927034e-01 -1.45915955e-01
1.01887710e-01 -6.07459545e-01 -1.33715808e-01 -1.95015645e+00
1.78632587e-01 -6.84943438e-01 -1.77502185e-01 7.75526702e-01
2.76696712e-01 -4.21007574e-01 4.32576180e-01 -6.60959929e-02
-7.30897844e-01 -1.89109355e-01 1.13203692e+00 -1.08492896e-01
-3.50127548e-01 -1.34047702e-01 -2.45328099e-01 6.64643407e-01
9.30468619e-01 -9.99101877e-01 -3.78345698e-01 -2.37631649e-01
5.25365353e-01 8.22765946e-01 -1.08308740e-01 -8.05754662e-01
8.73040438e-01 -1.28153372e+00 -5.78286171e-01 -7.79794216e-01
1.70497134e-01 -1.14790273e+00 9.46072459e-01 1.11477971e+00
6.73514009e-02 5.00496686e-01 2.86213279e-01 4.53788675e-02
-1.81536749e-01 -5.49668431e-01 3.95101815e-01 -3.35739881e-01
-8.64623606e-01 -2.75054932e-01 -7.54651129e-01 -4.08790261e-01
1.75857639e+00 -1.87185138e-01 -8.62031102e-01 -1.53255850e-01
-5.68393946e-01 8.31103086e-01 6.99226409e-02 2.30761394e-02
5.25067270e-01 -9.72566128e-01 -6.97188318e-01 -1.77643865e-01
-3.67642432e-01 -1.04588792e-01 -7.28280097e-02 9.33945358e-01
-7.00509667e-01 6.01812243e-01 -8.86375487e-01 1.81493908e-02
-1.68160439e+00 8.55105162e-01 1.37357533e-01 -7.78982878e-01
-4.14977163e-01 9.68667448e-01 -2.30090588e-01 1.14395149e-01
8.06426406e-02 7.98197910e-02 -3.82982433e-01 2.67498754e-03
5.56628287e-01 1.01157284e+00 -1.06906876e-01 -4.39976662e-01
-1.17630601e+00 6.05799198e-01 -3.41779068e-02 -3.76943529e-01
1.58976305e+00 -1.12964906e-01 -9.23287988e-01 -3.30105454e-01
2.38300651e-01 1.69368908e-01 -1.99735895e-01 1.71262741e-01
1.18391879e-01 -6.91896200e-01 8.62409770e-02 -1.14601672e+00
-1.09472477e+00 -3.07953637e-02 -1.94824278e-01 3.51548225e-01
1.18691778e+00 -1.69456750e-02 7.16251731e-01 1.59397468e-01
1.33543551e+00 -1.10504210e+00 -5.06064832e-01 4.04972285e-01
9.19838905e-01 -4.33035076e-01 4.05116826e-01 -1.58189559e+00
-3.17179143e-01 1.10800266e+00 1.01151013e+00 -5.41431569e-02
2.89056987e-01 7.81632900e-01 -3.65010768e-01 -3.08562905e-01
-1.23345804e+00 -5.56251928e-02 -6.44489825e-01 6.89862192e-01
-3.56523722e-01 1.33552477e-01 -1.32088923e+00 4.67746019e-01
8.46510101e-03 -1.90723762e-02 9.84833539e-01 1.61678624e+00
-7.72990048e-01 -1.71691191e+00 -8.38479638e-01 -2.54749537e-01
-5.53786531e-02 2.52632231e-01 -7.79933989e-01 9.76704359e-01
4.31934059e-01 1.55983377e+00 -5.99740259e-02 -2.13853747e-01
3.70341331e-01 -2.43130431e-01 8.36238921e-01 -3.44931304e-01
-9.31095243e-01 9.58583131e-02 9.30905819e-01 -8.77701700e-01
-6.85711443e-01 -6.81828141e-01 -1.75059664e+00 -7.89425552e-01
-5.07011473e-01 5.68382740e-01 4.39672947e-01 8.79846990e-01
1.86830238e-02 2.14475796e-01 5.88148355e-01 -6.07518196e-01
-1.27867520e-01 -1.17383294e-01 -3.88462216e-01 -3.52249146e-01
-1.31917611e-01 -1.00723469e+00 -1.48764729e-01 -6.95076406e-01] | [4.984013080596924, 1.8396251201629639] |
1b6ba259-43ca-472e-bf7f-bd7bdbe817df | real-world-video-for-zoom-enhancement-based | 2306.13875 | null | https://arxiv.org/abs/2306.13875v1 | https://arxiv.org/pdf/2306.13875v1.pdf | Real-World Video for Zoom Enhancement based on Spatio-Temporal Coupling | In recent years, single-frame image super-resolution (SR) has become more realistic by considering the zooming effect and using real-world short- and long-focus image pairs. In this paper, we further investigate the feasibility of applying realistic multi-frame clips to enhance zoom quality via spatio-temporal information coupling. Specifically, we first built a real-world video benchmark, VideoRAW, by a synchronized co-axis optical system. The dataset contains paired short-focus raw and long-focus sRGB videos of different dynamic scenes. Based on VideoRAW, we then presented a Spatio-Temporal Coupling Loss, termed as STCL. The proposed STCL is intended for better utilization of information from paired and adjacent frames to align and fuse features both temporally and spatially at the feature level. The outperformed experimental results obtained in different zoom scenarios demonstrate the superiority of integrating real-world video dataset and STCL into existing SR models for zoom quality enhancement, and reveal that the proposed method can serve as an advanced and viable tool for video zoom. | ['Jinyue Yan', 'Ryosuke Shibasaki', 'Zekun Cai', 'Xiaodan Shi', 'Haoran Zhang', 'Yinqiang Zheng', 'Zhiling Guo'] | 2023-06-24 | null | null | null | null | ['image-super-resolution', 'super-resolution'] | ['computer-vision', 'computer-vision'] | [ 2.99265951e-01 -7.33946443e-01 1.15180865e-01 -2.26136237e-01
-3.35356802e-01 -6.95876554e-02 3.47556561e-01 -3.52711827e-01
-2.92245150e-01 7.07454443e-01 2.90770441e-01 2.19088420e-01
-3.50102007e-01 -6.04731560e-01 -6.23633087e-01 -5.85488737e-01
-2.33621433e-01 -3.81663859e-01 8.13622534e-01 -3.88422370e-01
5.02933681e-01 5.66791475e-01 -1.90427101e+00 3.40193480e-01
8.45354021e-01 8.98414791e-01 8.66037965e-01 5.28527200e-01
1.06437482e-01 1.05893695e+00 -6.09930575e-01 4.78303619e-02
4.96092707e-01 -4.06443030e-01 -6.22275293e-01 2.90907323e-01
7.18861401e-01 -6.76041901e-01 -4.85194832e-01 9.68230546e-01
5.40055633e-01 4.91062343e-01 -3.32882792e-01 -1.28459620e+00
-4.77782488e-01 4.01095927e-01 -9.93997335e-01 9.74091649e-01
7.50190854e-01 4.28527594e-01 6.37051821e-01 -8.32738280e-01
1.07537591e+00 1.41985548e+00 3.03096473e-01 2.95467943e-01
-1.04754508e+00 -9.79918122e-01 -6.79945275e-02 6.06009185e-01
-1.34205067e+00 -4.97725338e-01 6.45383835e-01 -2.35294893e-01
7.00356603e-01 2.69643307e-01 6.43343329e-01 5.47481358e-01
4.89989042e-01 4.23845708e-01 1.22300398e+00 -2.91316092e-01
-4.07155864e-02 -4.74075638e-02 -2.21677616e-01 2.49545008e-01
9.06156227e-02 3.79522353e-01 -1.06567705e+00 3.19996923e-01
1.27451360e+00 1.86881542e-01 -8.37621748e-01 -2.24892259e-01
-1.54068041e+00 2.62090862e-01 5.01541436e-01 4.83869076e-01
-1.57379761e-01 -1.49343058e-01 2.77156919e-01 3.63050938e-01
4.50276375e-01 4.08492684e-01 3.77152339e-02 -2.02599972e-01
-1.41910315e+00 1.46240130e-01 -6.88574016e-02 1.19071031e+00
6.67848706e-01 1.43628374e-01 -2.09769920e-01 6.60007894e-01
9.84300449e-02 4.81832981e-01 5.30914903e-01 -1.43402553e+00
4.51881528e-01 3.24302018e-01 3.20118099e-01 -1.23978293e+00
-2.20750734e-01 -2.22309932e-01 -8.42368364e-01 3.97550464e-01
1.31124184e-01 4.00170982e-01 -3.48339319e-01 1.47171652e+00
3.30206871e-01 6.37300432e-01 -1.03365548e-01 1.37078476e+00
8.45871329e-01 8.61985326e-01 -2.18196258e-01 -6.89118862e-01
1.14108074e+00 -1.12269604e+00 -1.21043968e+00 2.88630635e-01
7.82913342e-02 -1.00671375e+00 1.21080136e+00 4.89035159e-01
-1.52201986e+00 -1.04089117e+00 -1.12990928e+00 -2.10742787e-01
8.72380063e-02 -1.86962709e-01 1.82237059e-01 6.65347874e-02
-1.14426851e+00 6.08276725e-01 -7.07911432e-01 -3.03294897e-01
2.13324711e-01 1.54990517e-02 -5.64821959e-01 -4.80438799e-01
-1.27760410e+00 6.03873134e-01 4.84501839e-01 -8.72173905e-02
-7.31918097e-01 -1.01084137e+00 -5.98284066e-01 -2.65569203e-02
4.69512373e-01 -6.32467091e-01 8.15373421e-01 -7.35175788e-01
-1.46589732e+00 4.76436436e-01 -2.74746776e-01 -3.02555710e-01
4.24409986e-01 -4.85718817e-01 -7.32903838e-01 6.32726312e-01
2.19200160e-02 5.30716002e-01 8.37646604e-01 -1.19366157e+00
-9.11056459e-01 -1.91841200e-01 2.29316920e-01 4.25594777e-01
-2.38671750e-01 4.61745292e-01 -5.44077158e-01 -5.56609571e-01
-8.46020207e-02 -4.66473192e-01 1.00888386e-01 1.98121130e-01
-1.63509767e-03 3.97342324e-01 1.28591275e+00 -6.30909741e-01
1.68214524e+00 -2.24995208e+00 2.86893040e-01 -4.91541862e-01
4.36908662e-01 4.74882543e-01 -4.14238423e-02 4.03089136e-01
-4.66817021e-02 -2.05847561e-01 1.61211863e-01 -2.81591833e-01
-8.41336846e-01 -9.99964178e-02 -1.53934151e-01 3.82516146e-01
-7.78945461e-02 5.46578765e-01 -1.06120682e+00 -6.85625911e-01
7.44659662e-01 6.70936704e-01 -5.31099617e-01 3.78532618e-01
1.61516219e-01 8.01255465e-01 -5.58545999e-02 3.92075151e-01
1.18000591e+00 -3.48485500e-01 -1.66562125e-01 -4.64429498e-01
-5.64212084e-01 -1.42443866e-01 -1.28251660e+00 1.84992218e+00
-4.30464298e-01 9.66442883e-01 3.12878788e-02 -2.55783021e-01
7.60805249e-01 -1.59263745e-01 5.99535704e-01 -1.05197847e+00
-1.41481429e-01 1.11118004e-01 -2.31559008e-01 -6.43436909e-01
9.44489777e-01 6.22487143e-02 6.66462302e-01 -5.54306880e-02
-7.27062672e-02 -2.66432147e-02 3.99474055e-01 4.55202669e-01
8.76228392e-01 1.97124064e-01 4.10423160e-01 -3.11975479e-01
9.15023267e-01 -3.21628273e-01 4.90324736e-01 4.16943163e-01
-1.95528299e-01 9.81988072e-01 -2.85443515e-02 -2.55014211e-01
-1.18072534e+00 -1.11452699e+00 -1.86706468e-01 8.13374579e-01
8.58883142e-01 -5.62541366e-01 -4.30421799e-01 1.23274438e-01
-2.76105225e-01 4.25240815e-01 -2.66915232e-01 6.32802024e-02
-6.98463321e-01 -3.17628652e-01 4.29822877e-02 1.14258453e-01
1.08074939e+00 -8.80085766e-01 -1.12491071e+00 2.03768238e-01
-3.75649124e-01 -1.50831747e+00 -6.57683313e-01 -5.61565340e-01
-8.15370560e-01 -1.07538438e+00 -9.12743747e-01 -4.09196466e-01
1.97884426e-01 1.13887501e+00 1.15012753e+00 1.39309704e-01
-3.57422531e-01 1.14224210e-01 -6.55169666e-01 4.61436719e-01
-2.29508355e-01 -3.91430914e-01 2.44658031e-02 2.16455579e-01
-1.29193127e-01 -7.13681698e-01 -1.04128468e+00 7.89986312e-01
-1.29629815e+00 5.61664164e-01 2.89439052e-01 6.57850504e-01
4.26762342e-01 2.46582255e-01 2.82752395e-01 -4.48889464e-01
3.16878110e-01 -2.58759618e-01 -6.57361865e-01 1.76742941e-01
-4.22155380e-01 -3.07487100e-01 5.93653202e-01 -2.76464432e-01
-1.40126801e+00 -6.47568107e-01 2.33987927e-01 -8.43851209e-01
1.54098541e-01 8.41166601e-02 -3.70062947e-01 -1.99787259e-01
3.77260059e-01 3.64513069e-01 7.37800226e-02 -4.81811851e-01
1.97491243e-01 6.96652770e-01 6.79484725e-01 -1.54680669e-01
6.47301733e-01 8.17094505e-01 1.29400447e-01 -7.50487745e-01
-4.56181645e-01 -7.25774646e-01 -4.98434186e-01 -4.44826901e-01
8.38223100e-01 -1.04385400e+00 -7.38764763e-01 6.99978411e-01
-9.68092263e-01 -1.43512681e-01 -2.86309123e-01 5.51913917e-01
-5.19895554e-01 7.95141101e-01 -7.64308095e-01 -4.31812912e-01
-2.49115005e-01 -1.41538703e+00 1.20169866e+00 3.40039313e-01
5.14816582e-01 -7.19593644e-01 -1.29441082e-01 4.11136717e-01
7.42455184e-01 8.86477530e-02 2.67522305e-01 2.86096364e-01
-1.08309007e+00 3.77363294e-01 -7.07031608e-01 3.43262047e-01
3.83347720e-01 8.54904801e-02 -5.75147271e-01 -6.24338627e-01
1.03420340e-01 6.93213493e-02 5.56141138e-01 3.36929172e-01
8.91526878e-01 3.67778651e-02 -1.26260743e-01 9.86932874e-01
1.76724219e+00 2.77572572e-01 1.04844475e+00 6.45136535e-01
7.07306325e-01 3.07537735e-01 1.40179932e+00 6.25953615e-01
1.07090086e-01 1.20014989e+00 4.41998661e-01 -6.61855936e-02
-5.50007761e-01 -1.48480237e-01 3.27979237e-01 8.73331666e-01
-3.94679010e-01 -3.24248821e-01 -4.74418312e-01 2.24520013e-01
-1.70649350e+00 -1.18845010e+00 -6.31690398e-02 2.31464696e+00
6.67885721e-01 -6.91054612e-02 1.11138280e-02 1.03445694e-01
8.45407724e-01 3.72084081e-01 -4.30295676e-01 2.03382224e-01
-5.52200913e-01 -1.46683991e-01 3.91676813e-01 4.69249785e-01
-7.28818119e-01 7.80401826e-01 5.87609863e+00 1.12852812e+00
-1.41809535e+00 2.42038846e-01 4.42789942e-01 -5.08862078e-01
-1.94502592e-01 -1.19531592e-02 -7.94950187e-01 8.85609567e-01
8.23222756e-01 -3.76330674e-01 5.85959077e-01 3.43964577e-01
8.26410115e-01 -4.82208699e-01 -8.08287144e-01 1.36108541e+00
1.10052153e-01 -1.48432600e+00 1.38200149e-01 9.98921320e-02
8.99596035e-01 -2.10507259e-01 1.57024160e-01 -2.82606810e-01
-1.95345268e-01 -7.73398876e-01 6.95338368e-01 4.99838412e-01
1.17367077e+00 -5.37456274e-01 5.58474660e-01 7.52642080e-02
-1.49722195e+00 -1.94279134e-01 -4.00636792e-01 1.14285707e-01
6.01131737e-01 4.98268694e-01 -1.99140817e-01 1.01951981e+00
1.12292135e+00 1.23878419e+00 -7.75368333e-01 1.21313298e+00
3.57914925e-01 3.71039174e-02 -2.42835775e-01 6.48880064e-01
-8.74923170e-02 -2.50593871e-01 7.03624606e-01 9.39261079e-01
6.74564898e-01 1.28124014e-01 -1.49748385e-01 6.56839669e-01
3.94870609e-01 -5.69847487e-02 -3.71681958e-01 4.20967877e-01
6.68935120e-01 1.14460528e+00 -4.15879488e-01 -3.82502496e-01
-6.10735297e-01 9.89868164e-01 -1.94634914e-01 4.62721974e-01
-1.05643296e+00 -2.05773532e-01 5.86614847e-01 5.88341296e-01
2.08983481e-01 -2.70428896e-01 2.06612900e-01 -1.52753484e+00
-9.31831375e-02 -8.06653976e-01 1.69003159e-01 -1.47678781e+00
-7.64839768e-01 7.81397641e-01 4.04163301e-01 -1.89396369e+00
-5.26885651e-02 -2.14028150e-01 -3.38174403e-01 7.20688522e-01
-1.87635231e+00 -1.02565110e+00 -9.69162226e-01 8.84160995e-01
9.50743496e-01 -2.54834499e-02 -1.13186082e-02 7.39260912e-01
-3.91477376e-01 3.51445556e-01 3.13094884e-01 -5.46416163e-01
1.03447711e+00 -6.57529294e-01 -2.77604666e-02 1.09582067e+00
-2.47639209e-01 5.54169476e-01 9.92483795e-01 -4.52314019e-01
-1.26095998e+00 -8.98999870e-01 2.94771880e-01 -7.00585544e-03
4.03804958e-01 -2.05332544e-02 -9.54096198e-01 4.91671115e-01
4.75702018e-01 3.34894896e-01 4.92261462e-02 -6.66282892e-01
3.33097316e-02 -3.80782455e-01 -1.32660544e+00 4.30772692e-01
1.30003405e+00 -2.47528642e-01 -2.44469315e-01 -1.58151075e-01
1.18105531e+00 -4.91152912e-01 -1.08518136e+00 5.09823859e-01
3.82880390e-01 -1.68633091e+00 1.13915288e+00 -1.79547165e-02
7.53830791e-01 -6.87427342e-01 -2.39364758e-01 -1.12976325e+00
-3.13697398e-01 -7.93808877e-01 1.09062879e-03 1.29673719e+00
-4.63012934e-01 -4.31880444e-01 3.86227548e-01 1.61825106e-01
-9.48731527e-02 -5.46005487e-01 -9.17500913e-01 -7.87520051e-01
-5.66283941e-01 -5.65658920e-02 6.24277294e-01 8.32937419e-01
-9.19584706e-02 9.28425044e-03 -7.71420479e-01 6.46978393e-02
7.26514697e-01 4.30681929e-02 8.69834661e-01 -8.03363621e-01
-2.91802704e-01 -1.12658530e-01 -7.62486696e-01 -1.39106858e+00
-4.63097841e-01 -1.66169405e-01 -3.46939206e-01 -1.03725278e+00
3.97488356e-01 -1.89634234e-01 -4.09721255e-01 -4.00388330e-01
-3.04935426e-01 4.21306074e-01 7.02141047e-01 5.53109467e-01
-8.76734674e-01 5.14930844e-01 1.76500392e+00 2.30729148e-01
-1.58889174e-01 -3.73108536e-01 -2.41696894e-01 3.70444268e-01
3.97198170e-01 8.63083899e-02 -6.47266626e-01 -4.32336122e-01
1.20477825e-01 5.63730061e-01 3.70963126e-01 -1.28574216e+00
3.21124017e-01 -2.48799309e-01 3.81319039e-02 -6.99572682e-01
5.75843573e-01 -9.76646364e-01 5.62628984e-01 1.05637528e-01
-2.34803751e-01 2.46096060e-01 1.63955063e-01 5.99324405e-01
-5.95095038e-01 2.96455845e-02 1.09465981e+00 1.53301831e-03
-1.04632115e+00 1.66981027e-01 1.49516553e-01 -2.03527678e-02
1.20950544e+00 -5.38783371e-01 -6.31934941e-01 -4.08084929e-01
-4.97726440e-01 1.13895088e-01 8.83212268e-01 4.38718230e-01
1.16763091e+00 -1.25219274e+00 -5.54460227e-01 4.20455873e-01
1.25093441e-02 -5.12398407e-02 9.28966880e-01 9.78395402e-01
-9.47678506e-01 3.04920524e-01 -8.36707115e-01 -8.44437778e-01
-1.64160383e+00 9.23889995e-01 1.84031129e-01 -2.41037250e-01
-9.04819965e-01 6.98669732e-01 4.57935244e-01 3.52356344e-01
5.24646901e-02 -3.70153040e-01 -2.57566661e-01 -3.33666682e-01
1.03652072e+00 6.76424444e-01 -2.97775209e-01 -8.26303065e-01
-1.52297303e-01 8.61666262e-01 1.45898089e-02 -7.21390396e-02
1.28964078e+00 -1.03900993e+00 -1.96159668e-02 4.55156773e-01
1.08906817e+00 1.03131287e-01 -1.59104753e+00 -3.75688672e-01
-4.12862003e-01 -1.44470918e+00 5.49644381e-02 -3.69634390e-01
-1.40207946e+00 8.45730841e-01 8.77708375e-01 9.96594038e-03
1.66088641e+00 -2.40736648e-01 9.90438700e-01 -3.35272700e-01
8.21203828e-01 -7.74189174e-01 3.63800615e-01 -1.57463364e-02
8.50156367e-01 -1.21178591e+00 3.67876828e-01 -7.78778911e-01
-6.27917290e-01 1.10304356e+00 7.63722301e-01 -6.90037981e-02
3.23422968e-01 2.06733972e-01 -1.68641970e-01 5.20567484e-02
-8.65670741e-01 8.46863762e-02 -1.77354682e-02 4.45851713e-01
2.86475927e-01 -4.77019429e-01 -3.16966116e-01 -1.74604595e-01
9.95855555e-02 5.22073209e-01 9.17479932e-01 6.95294619e-01
-4.64501083e-01 -6.66713774e-01 -4.68337953e-01 6.82605281e-02
-2.76475906e-01 -1.23465642e-01 5.37007511e-01 8.98020029e-01
1.88442603e-01 9.52530265e-01 1.19952440e-01 -5.77365875e-01
1.41183034e-01 -6.90578699e-01 5.25767565e-01 -1.31248966e-01
-2.68818319e-01 2.12466925e-01 -3.04763049e-01 -1.06020498e+00
-9.74281251e-01 -4.71381605e-01 -1.05674088e+00 -5.78078747e-01
-2.96124607e-01 -6.82970360e-02 2.64420569e-01 4.20736820e-01
5.37811518e-01 7.22625613e-01 7.52991259e-01 -1.12446964e+00
-1.18016645e-01 -9.69979167e-01 -8.99446666e-01 7.08029866e-01
4.80549186e-01 -8.53436053e-01 -5.92152953e-01 2.65841693e-01] | [11.000901222229004, -2.0100250244140625] |
13bbec89-6d19-4245-ac63-e9bab9f635ca | supervised-word-sense-disambiguation-on | null | null | https://aclanthology.org/2021.paclic-1.12 | https://aclanthology.org/2021.paclic-1.12.pdf | Supervised Word Sense Disambiguation on Taiwan Hakka Polysemy with Neural Network Models: A Case Study of BUN, TUNG and LAU | null | ['Yanhong Chen', 'Chia-Hung Lin', 'Jyi-Shane Liu', 'Hsiao-Ling Hsu', 'Huei-ling Lai'] | null | null | https://aclanthology.org/2021.paclic-1.78 | https://aclanthology.org/2021.paclic-1.78.pdf | paclic-2021-11 | ['word-sense-disambiguation'] | ['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] | [-7.346590518951416, 3.6740894317626953] |
632c85f6-221e-44bd-be12-b1ba59d5f7e6 | monotone-comparative-statics-for-submodular | 2304.12171 | null | https://arxiv.org/abs/2304.12171v1 | https://arxiv.org/pdf/2304.12171v1.pdf | Monotone comparative statics for submodular functions, with an application to aggregated deferred acceptance | We propose monotone comparative statics results for maximizers of submodular functions, as opposed to maximizers of supermodular functions as in the classical theory put forth by Veinott, Topkis, Milgrom, and Shannon among others. We introduce matrons, a natural structure that is dual to sublattices that generalizes existing structures such as matroids and polymatroids in combinatorial optimization and M-sets in discrete convex analysis. Our monotone comparative statics result is based on a natural order on matrons, which is dual in some sense to Veinott's strong set order on sublattices. As an application, we propose a deferred acceptance algorithm that operates in the case of divisible goods, and we study its convergence properties. | ['Maxime Sylvestre', 'Yu-Wei Hsieh', 'Alfred Galichon'] | 2023-04-24 | null | null | null | null | ['combinatorial-optimization'] | ['methodology'] | [ 2.74127215e-01 5.66180527e-01 -5.19722044e-01 -4.28622633e-01
-9.31797922e-02 -1.04409683e+00 2.60405332e-01 8.69497210e-02
-2.96334982e-01 8.59175920e-01 3.06900054e-01 -2.22236186e-01
-9.08318877e-01 -1.18688095e+00 -9.43592310e-01 -5.16885817e-01
-5.22385418e-01 8.11368406e-01 -2.62598693e-01 -5.70296228e-01
1.36709183e-01 3.11636090e-01 -1.18112433e+00 2.39271551e-01
9.54290450e-01 7.82563150e-01 6.45825416e-02 2.91536003e-01
-2.24141747e-01 3.97315979e-01 -3.54689434e-02 -5.99704683e-01
1.06498635e+00 -2.64321506e-01 -7.24550486e-01 5.48169196e-01
4.85344678e-01 -1.20754294e-01 -3.28672826e-01 1.39052355e+00
-3.00528139e-01 -1.96880788e-01 6.88607037e-01 -1.74166358e+00
-7.78898299e-01 1.28563511e+00 -8.44375312e-01 -6.12525567e-02
1.48169473e-01 -3.74166876e-01 1.73100507e+00 -5.96341252e-01
4.60698098e-01 1.48599553e+00 2.59044141e-01 4.06514704e-01
-1.46275067e+00 -2.12456495e-01 2.81558484e-01 -2.03060895e-01
-1.01314223e+00 -2.09595636e-01 2.20517159e-01 -2.33166113e-01
3.51209074e-01 8.80918622e-01 8.09694231e-01 1.99545294e-01
3.91029179e-01 1.11592245e+00 1.23748755e+00 -3.89336139e-01
3.55196357e-01 -1.77709833e-01 4.01609004e-01 4.22242880e-01
1.14234245e+00 -9.51526985e-02 -6.01640522e-01 -3.09649318e-01
8.43365073e-01 1.83990970e-01 -2.61880308e-01 -7.45894432e-01
-1.18866515e+00 8.64401340e-01 1.54440105e-01 1.59169585e-01
-5.95972955e-01 3.18032861e-01 -1.00395732e-01 9.21535969e-01
3.44145417e-01 4.95494217e-01 -1.67320207e-01 3.40715617e-01
-7.46863902e-01 2.17213243e-01 1.22169435e+00 1.55862367e+00
3.78620356e-01 -2.16332808e-01 -6.38310313e-02 2.75590658e-01
4.27126676e-01 4.91101086e-01 -5.08397579e-01 -1.04908216e+00
7.75841415e-01 7.64918149e-01 2.27166101e-01 -6.54400408e-01
-5.95451772e-01 -5.44544339e-01 -8.14750791e-01 -1.55976370e-01
4.77627873e-01 8.25183243e-02 -5.31347990e-01 1.87194824e+00
-5.46174571e-02 -3.29741359e-01 -4.76059318e-02 7.97969162e-01
-2.36994223e-04 6.66381836e-01 -6.30797386e-01 -8.71773303e-01
9.38253701e-01 -4.98947501e-01 -6.80916846e-01 1.70127124e-01
4.07564282e-01 -4.56167102e-01 5.97407281e-01 6.64314032e-01
-1.50085592e+00 3.17823946e-01 -9.45651174e-01 2.01576337e-01
9.00637805e-02 -5.95804989e-01 9.94951010e-01 1.25841284e+00
-1.37494552e+00 3.05257142e-01 -3.48342776e-01 -1.85287073e-01
4.42711502e-01 8.68245065e-01 -3.18682157e-02 -6.91771805e-02
-6.74814165e-01 6.29393578e-01 4.51729745e-01 4.38884348e-02
-1.05253446e+00 -7.32453406e-01 -6.95299089e-01 1.49061501e-01
8.27852786e-01 -6.34699702e-01 1.07462573e+00 -7.56156564e-01
-1.26809824e+00 9.59408462e-01 3.48654509e-01 -6.79933786e-01
6.21606946e-01 4.25607897e-02 3.55322026e-02 -2.33954996e-01
-2.38723103e-02 2.55025029e-01 3.12011927e-01 -1.23889732e+00
-7.13799894e-01 -7.50773668e-01 8.95098090e-01 5.12015045e-01
-4.56410021e-01 -6.21359237e-02 3.52564514e-01 -2.49956533e-01
5.63434124e-01 -7.94387519e-01 -7.12714493e-01 -2.08799392e-01
-5.18427014e-01 -1.08530305e-01 2.45562219e-03 2.08249297e-02
1.18229139e+00 -1.83397806e+00 6.45698845e-01 7.36055553e-01
5.54982424e-01 -8.83044899e-01 -3.16717416e-01 6.92533851e-01
2.63835043e-02 3.09687018e-01 -3.88809115e-01 1.29418358e-01
5.41928768e-01 4.15594369e-01 -1.76320970e-01 1.08176959e+00
-3.62104446e-01 8.48402441e-01 -7.61825144e-01 5.65304304e-04
-2.62700081e-01 -8.16883624e-01 -6.78682685e-01 -3.86563450e-01
-6.57503068e-01 -3.62860471e-01 -3.84737581e-01 8.55201066e-01
1.11344862e+00 -2.15665072e-01 7.76530087e-01 2.97207624e-01
-3.16352874e-01 8.90526474e-02 -1.52534556e+00 1.14569414e+00
-1.15065828e-01 7.89242312e-02 7.86358654e-01 -1.03174484e+00
8.03623259e-01 1.16943978e-01 9.26713288e-01 -4.50922221e-01
1.52682856e-01 4.30061489e-01 1.02088399e-01 2.89401356e-02
8.38058949e-01 -3.32950741e-01 -3.67571115e-01 6.03392482e-01
-3.53096545e-01 -2.71802396e-01 5.27964950e-01 5.66070259e-01
8.76963556e-01 -5.54743588e-01 3.25666577e-01 -1.25430226e+00
4.53544825e-01 2.02303901e-02 6.46474123e-01 9.19900119e-01
4.61497545e-01 3.47502083e-01 8.15165997e-01 -2.82848835e-01
-1.23873305e+00 -1.47855449e+00 -3.13914835e-01 1.11773908e+00
6.32257402e-01 -5.21409094e-01 -3.62200618e-01 -2.71327943e-01
4.69056308e-01 6.23938501e-01 -5.51622331e-01 3.30106288e-01
-2.61229098e-01 -1.04285944e+00 7.08852038e-02 1.93849206e-01
2.34981984e-01 -3.95876765e-01 -6.89613000e-02 1.45349026e-01
2.35627696e-01 -6.75298214e-01 -8.08029473e-01 2.09056929e-01
-1.24703383e+00 -9.50871229e-01 -6.49707854e-01 -6.21833622e-01
8.43427300e-01 3.89783859e-01 1.29477882e+00 -9.65177193e-02
2.48930156e-01 6.50718033e-01 2.31975839e-02 -6.35003626e-01
-2.67844349e-01 -2.12586880e-01 5.84216118e-01 2.34844387e-01
3.33646573e-02 -5.75951159e-01 -3.82451236e-01 4.94555861e-01
-1.27411985e+00 -4.95265536e-02 4.00907099e-01 5.73055923e-01
5.87692380e-01 1.76608622e-01 4.31338102e-01 -1.10740912e+00
6.60763085e-01 -8.04300785e-01 -1.44986439e+00 4.45568144e-01
-6.23817861e-01 8.21435601e-02 3.82899761e-01 9.36419964e-02
-6.50674462e-01 -6.54406771e-02 6.88274384e-01 1.64712474e-01
8.29454958e-01 7.32377887e-01 -4.58500266e-01 -1.82528421e-01
4.01386976e-01 2.29758427e-01 2.02595428e-01 -3.54381114e-01
4.22345698e-01 6.52437747e-01 1.37424037e-01 -8.96993518e-01
8.25461686e-01 8.38527977e-01 6.40097737e-01 -8.96572411e-01
-4.67487067e-01 -3.51910740e-01 -3.15612882e-01 -1.79313168e-01
4.11992460e-01 -5.63283324e-01 -1.21026218e+00 2.55573869e-01
-9.83968496e-01 1.76593229e-01 -6.50195003e-01 3.98482770e-01
-1.07399023e+00 2.63589591e-01 -5.10972440e-01 -1.18154657e+00
8.98741782e-02 -5.58386028e-01 5.59131920e-01 6.94390386e-02
4.01398450e-01 -8.06866884e-01 1.31024942e-01 4.38921541e-01
2.79706091e-01 1.86599687e-01 9.31349039e-01 -6.19312406e-01
-1.32626510e+00 2.60410070e-01 4.09925692e-02 1.65835753e-01
2.72564795e-02 -3.65062773e-01 -7.65528530e-02 -4.74179864e-01
2.27423444e-01 -5.16864657e-02 8.41001272e-01 7.27170944e-01
6.83372736e-01 -1.08412421e+00 -2.06726253e-01 3.50893736e-01
1.60331082e+00 1.07432730e-01 6.45487070e-01 3.21007669e-01
2.19887033e-01 6.65451407e-01 4.38698202e-01 7.96166599e-01
4.66696858e-01 3.02071303e-01 8.24527919e-01 2.51523018e-01
8.81600142e-01 1.05472840e-02 3.77875119e-01 9.24861193e-01
-1.81574315e-01 -3.35660368e-01 -2.39950657e-01 6.68303609e-01
-2.09783149e+00 -1.07059610e+00 -3.73491049e-01 2.40122771e+00
9.03633773e-01 -1.34634436e-03 7.17312157e-01 -1.39991373e-01
8.82058620e-01 2.62187980e-02 -2.57472962e-01 -5.82976699e-01
-6.42931044e-01 -9.67845395e-02 1.26677990e+00 6.92899108e-01
-7.20099449e-01 3.47808212e-01 7.06945419e+00 6.20054603e-01
-1.84227973e-01 -4.75974241e-03 6.95748448e-01 -5.48238635e-01
-1.20112169e+00 2.05613390e-01 -6.27243936e-01 2.20071808e-01
2.76702434e-01 -8.82911742e-01 8.38575184e-01 6.99552774e-01
5.40311337e-02 -1.29365757e-01 -1.59508681e+00 7.92263329e-01
-1.12583503e-01 -1.47194564e+00 -4.04484011e-02 7.00271070e-01
1.34637678e+00 -2.98381597e-01 2.27092922e-01 -2.99402297e-01
7.25167334e-01 -1.29868996e+00 6.86956763e-01 2.51135260e-01
2.64080167e-01 -8.87820899e-01 3.78765672e-01 4.03006732e-01
-8.65569472e-01 -3.72862190e-01 -8.47783566e-01 -3.15751851e-01
3.70803773e-01 7.25119829e-01 -5.33232450e-01 7.59245634e-01
8.39444175e-02 2.93356508e-01 3.31922948e-01 1.32355535e+00
4.31881964e-01 3.48762184e-01 -8.64989460e-01 -1.77010119e-01
2.30727956e-01 -8.41551363e-01 9.19064045e-01 7.77549505e-01
5.15613109e-02 4.17203337e-01 2.77223319e-01 9.03095245e-01
-6.03113890e-01 2.19689324e-01 -7.23679781e-01 -1.27218157e-01
3.34019810e-01 1.17555678e+00 -7.36438572e-01 3.96149866e-02
-4.93783563e-01 1.28975779e-01 -3.94474626e-01 5.84204271e-02
-5.26045918e-01 3.49302262e-01 6.52840018e-01 5.25754154e-01
-1.01334870e-01 -3.00306886e-01 -7.27042615e-01 -1.08141518e+00
2.05879807e-01 -8.42714071e-01 8.10363054e-01 1.88429236e-01
-1.48259342e+00 -5.66140339e-02 3.98477137e-01 -1.00913632e+00
2.02198699e-01 -7.01687098e-01 -4.59783047e-01 4.87179518e-01
-1.02428150e+00 -7.47021258e-01 3.19014996e-01 5.20724475e-01
1.58397406e-01 -3.37451580e-03 9.05004591e-02 -6.56634495e-02
6.67041540e-02 2.14666948e-01 4.88932371e-01 -4.98350143e-01
-3.07230234e-01 -1.61302912e+00 8.26615542e-02 7.10787892e-01
2.97976822e-01 5.29337168e-01 7.34298766e-01 -5.91456831e-01
-2.12638593e+00 -5.79184532e-01 5.80232203e-01 -2.87078053e-01
8.49610567e-01 -4.59171504e-01 -2.51385093e-01 8.37703586e-01
3.86941820e-01 -6.13129675e-01 4.43965405e-01 3.47996116e-01
2.67521027e-05 -5.35221219e-01 -1.33185136e+00 6.47674978e-01
1.60155952e+00 6.82403669e-02 -3.79662454e-01 8.55754256e-01
6.60214245e-01 -2.08725914e-01 -9.43557858e-01 4.91871446e-01
5.15299082e-01 -7.56317377e-01 6.61695004e-01 -7.79888272e-01
2.00023472e-01 -1.74590304e-01 -6.95330143e-01 -9.38997805e-01
-5.06027699e-01 -1.08714163e+00 2.51844943e-01 6.35831833e-01
5.78147054e-01 -7.88031399e-01 1.02286756e+00 5.17941177e-01
-1.53807297e-01 -7.30570316e-01 -1.24090040e+00 -1.34917593e+00
2.70909458e-01 -1.39975362e-02 6.63482845e-01 8.58934343e-01
4.88316298e-01 2.61834532e-01 -6.01001024e-01 5.35591319e-03
1.12994325e+00 3.99374545e-01 5.05954385e-01 -1.33829165e+00
-6.12786770e-01 -5.78330338e-01 -2.80571669e-01 -1.21510601e+00
-2.63903558e-01 -1.49657011e+00 -1.26686648e-01 -1.48867011e+00
5.88822365e-01 -6.22117460e-01 -3.14901292e-01 1.89409763e-01
7.57028878e-01 6.22533960e-03 7.12641060e-01 1.31674096e-01
-8.93865585e-01 5.12326539e-01 1.47396004e+00 -3.42355877e-01
-5.46945632e-01 3.74751747e-01 -1.17418468e+00 4.49864089e-01
5.49663723e-01 -3.60670477e-01 -4.05134678e-01 -3.80532652e-01
9.43045557e-01 4.09303188e-01 -4.96721193e-02 -3.53879631e-01
2.94198722e-01 -9.49970841e-01 -1.59804910e-01 -5.93485475e-01
1.59867361e-01 -9.66442406e-01 5.04469752e-01 7.25888133e-01
-3.67634416e-01 2.87645400e-01 -2.11140051e-01 3.75572175e-01
2.27695495e-01 -5.95374048e-01 5.36321163e-01 -2.60777652e-01
-1.41523570e-01 4.94586229e-01 -2.51013786e-01 1.95391387e-01
1.26696801e+00 -3.70875537e-01 -5.21147132e-01 -6.95257843e-01
-4.58117813e-01 7.48169065e-01 4.97962654e-01 1.74491610e-02
6.00695193e-01 -1.20214093e+00 -1.15350652e+00 -2.94706523e-01
-1.35137796e-01 -1.00488640e-01 -1.35708064e-01 1.07786036e+00
-3.63853514e-01 4.64104056e-01 -4.38445896e-01 -4.35603589e-01
-9.50958669e-01 4.59495395e-01 2.36163065e-02 -1.61877394e-01
-4.29397449e-02 6.92433298e-01 6.57568932e-01 -1.83364764e-01
-1.41692162e-03 -3.43522757e-01 2.05660835e-01 5.87260164e-03
3.14105451e-02 7.32214987e-01 -2.99676180e-01 -1.68578088e-01
-5.01807809e-01 4.06028284e-03 -1.83064833e-01 -4.84612256e-01
1.67938662e+00 -2.10631311e-01 -1.02834749e+00 2.77243078e-01
4.15210932e-01 3.29252779e-01 -6.14422739e-01 -1.05565496e-01
2.94607967e-01 -6.70218885e-01 -5.72724402e-01 -3.81497175e-01
-8.25760782e-01 -5.96789345e-02 -1.41591104e-02 8.59740734e-01
1.09820104e+00 2.86534965e-01 3.19908381e-01 6.27210915e-01
9.63072538e-01 -1.29445541e+00 -1.90593019e-01 3.94699723e-01
1.07396209e+00 -4.64546978e-01 2.40872040e-01 -6.72529936e-01
-3.04243565e-01 1.03781700e+00 3.03594589e-01 -2.46594951e-01
6.62396312e-01 5.74777007e-01 -6.98885918e-01 -1.31335944e-01
-9.15857911e-01 -2.57132292e-01 1.32045329e-01 2.41503194e-01
8.90566260e-02 6.82033300e-01 -1.10430908e+00 6.89287245e-01
-2.17313170e-01 -1.27122059e-01 1.15872121e+00 9.86659288e-01
-7.81386852e-01 -1.18008029e+00 -6.05357289e-01 9.85911131e-01
-3.09792101e-01 -3.08895141e-01 -6.35052025e-01 5.76177895e-01
-1.28889099e-01 9.26152587e-01 1.11655965e-01 -2.78079867e-01
1.19975686e-01 -6.09275997e-01 9.77828085e-01 -8.06750953e-01
-3.75670642e-01 4.81336564e-02 1.82920560e-01 -1.83165073e-01
-3.85531038e-01 -7.10553348e-01 -7.59975016e-01 -8.01202238e-01
-5.84635377e-01 2.49490112e-01 3.38561028e-01 7.45815575e-01
-2.19485924e-01 -2.32684135e-01 9.47387516e-01 6.43726066e-02
-1.20843017e+00 -5.27638197e-01 -1.27521491e+00 -9.72493440e-02
5.66736050e-02 -2.18921319e-01 -4.47029322e-02 -4.87423599e-01] | [6.600698947906494, 4.836030960083008] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.