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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
c992763e-d241-4467-b973-4c7ac96781b8 | on-target-representation-in-continuous-output | null | null | https://aclanthology.org/2022.repl4nlp-1.24 | https://aclanthology.org/2022.repl4nlp-1.24.pdf | On Target Representation in Continuous-output Neural Machine Translation | Continuous generative models proved their usefulness in high-dimensional data, such as image and audio generation. However, continuous models for text generation have received limited attention from the community. In this work, we study continuous text generation using Transformers for neural machine translation (NMT). We argue that the choice of embeddings is crucial for such models, so we aim to focus on one particular aspect”:" target representation via embeddings. We explore pretrained embeddings and also introduce knowledge transfer from the discrete Transformer model using embeddings in Euclidean and non-Euclidean spaces. Our results on the WMT Romanian-English and English-Turkish benchmarks show such transfer leads to the best-performing continuous model. | ['Vlad Niculae', 'Evgeniia Tokarchuk'] | null | null | null | null | repl4nlp-acl-2022-5 | ['audio-generation'] | ['audio'] | [ 3.50298852e-01 3.56909335e-01 -5.69815636e-02 5.94669282e-02
-9.11984682e-01 -5.74626863e-01 1.39051640e+00 -3.27535480e-01
-2.29877666e-01 8.29556584e-01 5.46082377e-01 -4.02048856e-01
1.92314722e-02 -8.16258311e-01 -8.44362438e-01 -5.62572002e-01
3.82970095e-01 8.10383737e-01 -2.95123488e-01 -5.74984848e-01
-4.88505028e-02 4.10093293e-02 -1.19060445e+00 3.50021899e-01
7.62929499e-01 7.15599716e-01 -6.52733594e-02 6.33135021e-01
-1.77190974e-01 4.89459038e-01 -6.97710693e-01 -8.41468573e-01
3.51485699e-01 -8.24585974e-01 -9.61246669e-01 -1.42423958e-01
1.96924463e-01 -2.29394644e-01 -4.94390309e-01 7.54196465e-01
8.54784310e-01 -1.55598791e-02 1.20080566e+00 -1.26010990e+00
-1.43733144e+00 1.03090143e+00 -4.62201191e-03 -8.50716457e-02
2.74293661e-01 7.68875564e-03 1.18421948e+00 -1.38090968e+00
1.00058818e+00 1.33504343e+00 5.76578617e-01 7.87786186e-01
-1.54486442e+00 -3.38527799e-01 -1.81951925e-01 2.09474564e-01
-1.28098953e+00 -4.57966000e-01 6.96092010e-01 -2.58683711e-01
1.03224540e+00 1.69003591e-01 5.34193575e-01 1.86854136e+00
2.67547637e-01 7.86724925e-01 1.16995728e+00 -6.65723920e-01
4.67123277e-02 3.63891989e-01 -6.24743819e-01 2.29027659e-01
4.33888398e-02 1.17284067e-01 -7.60275483e-01 1.00572616e-01
7.09530592e-01 -4.32078302e-01 -2.64694333e-01 -3.03439230e-01
-1.64689100e+00 1.19745350e+00 2.82696187e-01 6.31360173e-01
-2.98485786e-01 4.52002347e-01 2.86808193e-01 7.07797945e-01
7.77016521e-01 7.71330357e-01 -2.40903556e-01 -5.61541736e-01
-8.37050796e-01 3.62773716e-01 7.45872736e-01 1.19805431e+00
3.74868214e-01 3.33301663e-01 -4.36590195e-01 1.01435494e+00
-4.72592423e-03 6.75149083e-01 8.18276227e-01 -5.40880263e-01
7.11214304e-01 1.18495554e-01 9.78403240e-02 -6.81934774e-01
1.11243185e-02 -4.63937670e-01 -8.27695668e-01 -9.25554037e-02
1.92390338e-01 -3.43673587e-01 -9.56748188e-01 1.71928477e+00
5.00301756e-02 3.94758694e-02 4.19133097e-01 8.02565396e-01
6.52833402e-01 8.54549468e-01 -3.80677879e-01 -4.14434001e-02
1.00882077e+00 -7.72447824e-01 -8.41136575e-01 5.74259413e-03
6.24952912e-01 -1.04148650e+00 1.13992846e+00 2.26640642e-01
-1.26835430e+00 -4.23440576e-01 -8.39773178e-01 -3.01414460e-01
-6.51123047e-01 3.99582356e-01 2.26376250e-01 5.44597685e-01
-1.26632655e+00 5.60674787e-01 -5.98280430e-01 -4.06046659e-01
6.77659223e-03 2.47846574e-01 -3.38052124e-01 -2.39060409e-02
-1.62996984e+00 1.11988103e+00 2.33075649e-01 2.85602100e-02
-7.43164420e-01 -5.84194422e-01 -6.24146223e-01 -1.54305637e-01
-1.38947666e-01 -1.14439058e+00 1.26422894e+00 -7.61246622e-01
-1.94045269e+00 5.90573430e-01 9.11924914e-02 -7.67634392e-01
9.16526258e-01 -3.24647933e-01 -1.73274070e-01 9.88728032e-02
-2.00793073e-02 1.06235731e+00 1.04145432e+00 -1.09940040e+00
-7.01231733e-02 -1.10292405e-01 1.42100276e-02 2.04591647e-01
-7.61818290e-01 -9.91692990e-02 5.85104637e-02 -1.12347031e+00
-2.03041613e-01 -1.24724281e+00 9.75320786e-02 -3.99607807e-01
-4.23735023e-01 -3.29088688e-01 8.72927368e-01 -6.26425683e-01
1.03250539e+00 -1.85040057e+00 5.47484338e-01 -2.45224193e-01
-1.39572889e-01 3.84537019e-02 -2.96614558e-01 9.44173932e-01
-1.12693839e-01 3.30010653e-01 3.49100940e-02 -6.52454436e-01
4.54331875e-01 2.05535814e-01 -7.40300417e-01 -2.28390582e-02
6.67633593e-01 1.35460806e+00 -6.57169700e-01 -4.03907955e-01
6.98477924e-02 8.92876685e-01 -7.00446546e-01 1.91564724e-01
-3.61493587e-01 3.45146745e-01 -3.27374011e-01 1.65776044e-01
3.09480190e-01 -8.66719242e-03 -7.31099471e-02 -1.53044388e-01
-7.75935575e-02 5.58081567e-01 -8.05161178e-01 1.95606041e+00
-7.35407114e-01 8.18461001e-01 -4.24862087e-01 -7.68497765e-01
9.29898441e-01 7.04750299e-01 2.75290340e-01 -6.43851817e-01
1.56992137e-01 3.42671484e-01 7.71761537e-02 -2.07438454e-01
9.64126170e-01 -4.53591257e-01 -9.63187292e-02 6.31726325e-01
3.61694694e-01 -5.44257760e-01 2.16130450e-01 2.01094285e-01
9.10742283e-01 2.82021701e-01 -2.47644559e-01 -1.24048732e-01
-5.26292361e-02 -4.44010878e-03 -9.74054188e-02 3.91412556e-01
4.28707302e-01 1.00833392e+00 5.38688242e-01 -2.13310104e-02
-1.23097336e+00 -1.22233355e+00 -7.00203851e-02 1.01744902e+00
-3.42063546e-01 -5.64972997e-01 -8.49988341e-01 -5.27333677e-01
-2.43584424e-01 1.03488421e+00 -9.14297879e-01 -4.73681748e-01
-6.23703539e-01 -6.78992808e-01 7.64826119e-01 5.99815607e-01
1.55022502e-01 -9.09322083e-01 -3.26546431e-01 3.07152480e-01
-5.00894547e-01 -1.22215497e+00 -5.89666665e-01 2.42176190e-01
-9.03488517e-01 -3.49949658e-01 -1.25323200e+00 -6.14720702e-01
3.72311622e-01 -1.37908041e-01 1.04690599e+00 -6.98930323e-01
-3.14952470e-02 5.23437858e-01 -6.05914056e-01 -4.36915457e-01
-7.96216547e-01 4.70302522e-01 1.72771648e-01 1.20998383e-01
3.71761546e-02 -6.79979742e-01 -3.49734455e-01 1.33056492e-01
-1.04367447e+00 3.90696138e-01 6.92721367e-01 1.06296003e+00
2.73353308e-01 -5.18509686e-01 7.96046019e-01 -6.06255770e-01
1.24161208e+00 -3.54154736e-01 -2.01501638e-01 1.12498157e-01
-6.50834978e-01 4.22441304e-01 7.73875833e-01 -7.65168369e-01
-9.40644860e-01 -4.16867137e-01 -5.71075901e-02 -7.40139246e-01
2.66397715e-01 5.22268295e-01 3.46239060e-02 3.61504346e-01
8.65555942e-01 5.41724741e-01 -6.00778870e-02 -3.75089496e-01
8.36217523e-01 6.60360336e-01 2.65016705e-01 -8.00257146e-01
9.37337816e-01 1.05376057e-01 -1.48369037e-02 -7.60825813e-01
-2.28866577e-01 2.72641957e-01 -6.26734078e-01 1.25004649e-01
9.04032946e-01 -6.84741259e-01 5.27619682e-02 1.01887882e-01
-1.47954607e+00 -4.05190557e-01 -6.66734695e-01 5.44440567e-01
-9.71537590e-01 -3.79973650e-02 -6.92442417e-01 -6.02016807e-01
-5.28662086e-01 -1.13591075e+00 1.39028788e+00 -2.87872076e-01
-3.56877208e-01 -1.27510071e+00 4.10994500e-01 9.13543627e-02
7.75479198e-01 1.61564738e-01 9.37906325e-01 -6.72111988e-01
-5.27795792e-01 -4.81333882e-02 4.67213523e-03 5.37774682e-01
9.41484943e-02 -1.66838467e-01 -9.76397872e-01 -4.18359302e-02
-2.54995644e-01 -4.71473873e-01 8.96169424e-01 6.96429089e-02
7.43521750e-01 -2.87159085e-01 -3.60061191e-02 4.35460627e-01
1.13911629e+00 -7.94758126e-02 7.54703403e-01 -4.63953279e-02
6.56849384e-01 4.60002780e-01 2.96543419e-01 2.74804413e-01
3.12207669e-01 8.43127847e-01 1.76942274e-01 -2.35804953e-02
-5.30002415e-01 -6.41319156e-01 7.07187116e-01 1.32355940e+00
-9.05311853e-02 -3.05059075e-01 -8.31286550e-01 7.14080989e-01
-1.67562783e+00 -8.65160525e-01 -1.30211096e-03 1.98472548e+00
1.19300640e+00 4.05269265e-02 -9.48728919e-02 2.00305462e-01
4.55650449e-01 -3.49648148e-02 -1.64785504e-01 -6.30640984e-01
-3.24529886e-01 6.80285275e-01 1.66039869e-01 3.15465868e-01
-7.78651774e-01 1.07114530e+00 6.50461674e+00 9.13297117e-01
-1.16610074e+00 3.89983624e-01 3.81422460e-01 -6.69996366e-02
-5.92483938e-01 -1.77332491e-01 -6.45319045e-01 3.92659098e-01
1.27597117e+00 -2.95613438e-01 3.54457647e-01 4.60305601e-01
1.31285787e-01 5.77728271e-01 -1.44095016e+00 9.55143332e-01
1.23028919e-01 -1.48978150e+00 5.52923679e-01 3.22991699e-01
1.01629210e+00 -3.87092978e-02 5.08465827e-01 5.29914081e-01
3.13040406e-01 -1.29131567e+00 8.98306131e-01 3.87719423e-01
1.10263610e+00 -5.66884875e-01 3.77186060e-01 5.23477141e-03
-8.58921528e-01 2.39973322e-01 -3.05240452e-01 2.73578644e-01
1.38185337e-01 4.75381494e-01 -1.21579814e+00 7.70478427e-01
1.21086702e-01 5.67006767e-01 -4.60342526e-01 4.22679335e-01
-1.40440211e-01 6.93152726e-01 -1.35639668e-01 -9.12935808e-02
3.23814154e-01 -2.25259885e-01 5.23420393e-01 1.28059947e+00
8.19844782e-01 -3.77195477e-01 -2.70759910e-01 1.18776822e+00
-3.15347791e-01 1.70911655e-01 -1.00339079e+00 -4.71977830e-01
9.90453586e-02 9.13845420e-01 -3.31822574e-01 -3.10012728e-01
-9.89461690e-02 1.30349970e+00 1.74612790e-01 3.66736591e-01
-9.93810236e-01 -3.97772133e-01 6.07721329e-01 1.52026162e-01
3.22906166e-01 -4.93270427e-01 -2.49277368e-01 -1.35075283e+00
-7.67868478e-03 -8.14194262e-01 -7.27975518e-02 -9.67776299e-01
-1.17617559e+00 8.13222706e-01 2.29074508e-01 -1.28040564e+00
-9.08948898e-01 -7.42513657e-01 -4.43024009e-01 8.92043412e-01
-1.29871690e+00 -1.47524953e+00 1.40825137e-01 6.18200243e-01
6.40930355e-01 -1.88663572e-01 1.00663829e+00 3.01087767e-01
-1.78472877e-01 7.48855591e-01 2.92079121e-01 1.31436422e-01
8.52357686e-01 -1.42795062e+00 7.58946061e-01 6.12778068e-01
5.90430975e-01 6.35443807e-01 8.48203003e-01 -3.59001219e-01
-1.64885092e+00 -1.20989525e+00 1.11440694e+00 -7.70792007e-01
6.50893092e-01 -6.53314292e-01 -4.40513194e-01 7.13910043e-01
6.94327116e-01 -2.40981951e-01 5.19749582e-01 -9.77461562e-02
-4.56778198e-01 1.55416965e-01 -8.38520944e-01 9.17545617e-01
1.18815231e+00 -6.48781478e-01 -5.39969862e-01 2.83095479e-01
1.00245321e+00 -2.60842651e-01 -9.93727207e-01 1.93881959e-01
4.39438075e-01 -6.04301453e-01 8.97169232e-01 -5.55657327e-01
8.48252118e-01 4.80414815e-02 -2.16569409e-01 -1.94589901e+00
-1.33489696e-02 -8.10757220e-01 -1.96508188e-02 1.29051435e+00
7.19794393e-01 -7.13615119e-01 4.58448321e-01 1.70457438e-01
-2.22392827e-01 -7.41424203e-01 -1.17442501e+00 -8.82287025e-01
7.99579561e-01 -5.34675896e-01 6.55633926e-01 9.66123402e-01
-2.80653015e-02 8.05620074e-01 -6.02178693e-01 -5.33074200e-01
1.99188575e-01 4.97373827e-02 9.96418595e-01 -9.48341250e-01
-2.81957418e-01 -5.44567406e-01 -2.82830000e-01 -1.05109656e+00
9.73296762e-02 -1.21886277e+00 -2.30067015e-01 -1.74258530e+00
-1.05405129e-01 -3.12767804e-01 1.26379162e-01 2.66448975e-01
2.01806083e-01 3.89882177e-01 2.31848910e-01 -9.14976820e-02
-2.38835990e-01 1.24723899e+00 1.32153249e+00 -2.36184165e-01
1.01714335e-01 -3.21508408e-01 -6.28952384e-01 2.97818054e-02
8.82477701e-01 -3.35054666e-01 -5.89031518e-01 -8.69202852e-01
3.69519174e-01 -5.23442738e-02 3.20265949e-01 -8.75848413e-01
-2.63146944e-02 -3.04516908e-02 2.09894761e-01 -1.41859606e-01
7.59412229e-01 -5.81275582e-01 2.00883791e-01 2.03945056e-01
-6.15663290e-01 3.00085753e-01 -1.24643715e-02 3.54253203e-01
-3.20515782e-01 -7.88766369e-02 4.30162996e-01 1.16212279e-01
7.22860619e-02 1.86265171e-01 -1.80200651e-01 2.80804455e-01
6.99354947e-01 -2.52432108e-01 -2.32323378e-01 -6.21607423e-01
-7.91827321e-01 -2.69452244e-01 2.87305325e-01 9.10663247e-01
7.18958199e-01 -1.88739896e+00 -1.16645515e+00 7.76696801e-02
1.88371807e-01 -3.78210753e-01 -2.85187870e-01 9.42883492e-01
-1.35401711e-01 6.46856785e-01 -6.45393655e-02 -5.21730363e-01
-8.82251799e-01 3.76171440e-01 3.38466823e-01 -3.62174809e-01
-4.03578877e-01 5.15079319e-01 6.91507235e-02 -4.59049731e-01
-1.39013350e-01 -6.51997685e-01 1.29538491e-01 2.07567543e-01
1.19593248e-01 2.29901239e-01 2.24598467e-01 -6.73005044e-01
7.33624026e-02 3.83780718e-01 1.72224324e-02 -7.87111104e-01
1.17071116e+00 4.45141457e-02 1.22970872e-01 5.72713077e-01
1.44943249e+00 -1.77184865e-01 -7.76944280e-01 -2.31792539e-01
-1.59960896e-01 -2.60073662e-01 -1.07732058e-01 -6.29522741e-01
-6.92013741e-01 1.34125113e+00 4.29842949e-01 3.15809041e-01
7.70854771e-01 -1.13745935e-01 9.18223202e-01 3.18430573e-01
4.25489128e-01 -1.10543513e+00 4.65361327e-01 6.92648590e-01
1.29538476e+00 -8.22663426e-01 -5.95540822e-01 1.26350954e-01
-6.73957229e-01 1.13666153e+00 2.05043480e-01 -3.83252203e-02
4.70916569e-01 1.12182431e-01 -3.76477465e-02 1.85544401e-01
-1.13615394e+00 -1.47471666e-01 3.86365414e-01 7.67217934e-01
7.29411781e-01 1.70146108e-01 -3.40026408e-01 4.50024277e-01
-7.92026520e-01 4.53381613e-02 5.71671188e-01 6.53389096e-01
-1.06009748e-02 -1.56838524e+00 -4.09698933e-01 2.19292194e-01
-4.80001926e-01 -3.72697979e-01 -7.34926224e-01 6.56473875e-01
-4.89898846e-02 8.83872211e-01 -1.21954367e-01 -5.27433991e-01
2.89833337e-01 4.53013003e-01 7.93611526e-01 -5.86195529e-01
-7.30415344e-01 4.37628068e-02 2.40801334e-01 3.51654328e-02
-1.90171480e-01 -5.63461721e-01 -8.56360853e-01 -2.20305562e-01
-2.35940650e-01 2.96312898e-01 8.00948501e-01 5.68220139e-01
5.78943372e-01 5.44847131e-01 5.92511356e-01 -8.41779828e-01
-9.08559144e-01 -1.37470520e+00 -1.52134106e-01 4.77761298e-01
6.39991835e-02 -5.49791634e-01 -2.02418894e-01 2.28781700e-01] | [11.845649719238281, 9.591038703918457] |
1d94d8d0-efe4-435d-9129-d3f6fab8aaff | no-reference-video-quality-assessment-based | null | null | https://www.mdpi.com/2079-9292/10/22/2768 | https://www.mdpi.com/2079-9292/10/22/2768 | No-Reference Video Quality Assessment Based on Benford’s Law and Perceptual Features | No-reference video quality assessment (NR-VQA) has piqued the scientific community’s interest throughout the last few decades, owing to its importance in human-centered interfaces. The goal of NR-VQA is to predict the perceptual quality of digital videos without any information about their distortion-free counterparts. Over the past few decades, NR-VQA has become a very popular research topic due to the spread of multimedia content and video databases. For successful video quality evaluation, creating an effective video representation from the original video is a crucial step. In this paper, we propose a powerful feature vector for NR-VQA inspired by Benford’s law. Specifically, it is demonstrated that first-digit distributions extracted from different transform domains of the video volume data are quality-aware features and can be effectively mapped onto perceptual quality scores. Extensive experiments were carried out on two large, authentically distorted VQA benchmark databases. | ['Domonkos Varga'] | 2021-11-12 | null | null | null | electronics-2021-11 | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 1.74682662e-01 -6.82485044e-01 -9.99869108e-02 -4.13885683e-01
-1.02080834e+00 -3.95721346e-01 3.07184339e-01 1.58850968e-01
-1.60742223e-01 5.61805785e-01 4.37086344e-01 8.55102465e-02
-2.81277508e-01 -5.64621568e-01 -3.05433810e-01 -6.29536450e-01
-3.37800235e-01 -3.80746335e-01 2.30837539e-01 -3.07814687e-01
5.10057867e-01 5.08214712e-01 -1.59893513e+00 2.56693244e-01
9.72448289e-01 1.53380835e+00 7.85848498e-02 7.78017163e-01
2.65435427e-01 7.72438347e-01 -6.92610025e-01 -6.97517514e-01
3.08143288e-01 -5.49683034e-01 -4.03324187e-01 1.60958767e-01
2.88964480e-01 -4.15627778e-01 -8.41817677e-01 1.14762771e+00
4.45341557e-01 1.67845726e-01 5.90095222e-01 -1.33593500e+00
-4.37745839e-01 -8.61100107e-02 -2.81633973e-01 4.99695957e-01
5.50916612e-01 1.85939044e-01 1.28042471e+00 -8.79404128e-01
4.12991434e-01 1.13825274e+00 2.77932495e-01 2.37588748e-01
-9.25534666e-01 -5.05618811e-01 -2.11072996e-01 1.02899027e+00
-1.39841485e+00 -4.51400846e-01 9.36963141e-01 -4.67162311e-01
5.90265930e-01 3.56127977e-01 6.47430956e-01 7.36136556e-01
5.64796448e-01 7.71122754e-01 8.34735632e-01 -1.89763010e-01
3.36196482e-01 -1.18583202e-01 -4.53372657e-01 4.10459161e-01
1.04390822e-01 5.71264625e-02 -5.95150769e-01 5.63478237e-03
8.70453537e-01 -2.01415449e-01 -6.67701364e-01 -7.31974423e-01
-1.07063818e+00 6.05360270e-01 2.13914618e-01 3.84883791e-01
-3.75813633e-01 -3.08064763e-02 5.47738671e-01 2.74117738e-01
1.66456938e-01 3.18007916e-01 -8.56169015e-02 -7.28793681e-01
-1.02428758e+00 1.87882096e-01 2.76827037e-01 7.92728066e-01
3.14335018e-01 2.83642828e-01 -1.33142382e-01 7.20593274e-01
2.50549287e-01 6.22513115e-01 5.09756505e-01 -1.33697712e+00
4.46823120e-01 3.53805959e-01 3.02989483e-01 -1.45181954e+00
2.32443530e-02 -2.84943879e-01 -1.01098788e+00 2.45320916e-01
3.87175977e-01 2.05593273e-01 -2.65030950e-01 1.40460896e+00
-3.26887779e-02 1.43723693e-02 9.02242735e-02 1.15629101e+00
5.44393182e-01 9.05367434e-01 -1.60296634e-01 -5.04647255e-01
9.97143626e-01 -4.71776158e-01 -8.92599046e-01 3.16747278e-01
2.08816826e-02 -7.33987987e-01 8.64118934e-01 8.89004886e-01
-1.22693932e+00 -9.28643525e-01 -1.34772456e+00 1.52832508e-01
1.95086196e-01 -1.93468988e-01 2.50973940e-01 7.82512188e-01
-1.04980528e+00 5.96012175e-01 -5.34289837e-01 -7.81256054e-03
4.99553144e-01 1.82385758e-01 -4.59093124e-01 -2.56982148e-01
-1.21412063e+00 5.57857454e-01 5.88695295e-02 -6.40508682e-02
-1.08935404e+00 -4.35308844e-01 -6.73380733e-01 1.99878916e-01
4.55729276e-01 -2.51876831e-01 9.21730459e-01 -1.10422325e+00
-1.42310631e+00 4.51553822e-01 -1.42537609e-01 -2.14795113e-01
4.58147228e-01 -1.78580269e-01 -1.01295888e+00 6.15829051e-01
-1.87016845e-01 1.20982811e-01 1.20172977e+00 -1.13936639e+00
-8.24671090e-01 -3.58364284e-01 1.33882180e-01 1.93078339e-01
-3.81487370e-01 -7.17444019e-03 -6.55807078e-01 -9.42221820e-01
9.07981992e-02 -4.21064913e-01 1.35487676e-01 2.42059946e-01
3.40075530e-02 3.50076221e-02 5.84795475e-01 -8.86568427e-01
1.44045913e+00 -2.29344225e+00 3.49766076e-01 2.22645596e-01
3.59541386e-01 4.22604918e-01 -2.64347523e-01 2.88043082e-01
1.54935420e-01 -1.37517345e-03 -9.70476121e-02 3.39995921e-01
-1.34520650e-01 -7.64699280e-02 -1.43601656e-01 5.56252539e-01
3.22041065e-01 5.91351330e-01 -9.08839941e-01 -5.25719643e-01
2.42690772e-01 6.52958930e-01 -6.44960940e-01 4.23102677e-01
1.26222879e-01 4.07264173e-01 -2.78170347e-01 5.72434127e-01
6.42503858e-01 -9.10249650e-02 1.04031213e-01 -6.19862497e-01
1.71925537e-02 -1.53216928e-01 -1.05179548e+00 1.64725900e+00
-2.83578634e-01 9.72286820e-01 -4.36808914e-02 -9.54264700e-01
8.62806439e-01 5.60596943e-01 8.09965730e-01 -1.33914733e+00
2.68262655e-01 2.30091363e-01 3.55670154e-02 -5.04557848e-01
5.62985718e-01 -1.21472022e-02 1.00694105e-01 -9.10512358e-02
1.16969377e-01 -1.77188605e-01 2.74871588e-01 8.51807296e-02
9.66290534e-01 -5.68517596e-02 5.79972923e-01 8.96922499e-02
9.79478300e-01 -4.42961186e-01 7.29354799e-01 2.34352037e-01
-6.87028825e-01 7.43370593e-01 3.07951301e-01 -1.44003063e-01
-1.21090639e+00 -1.40700209e+00 -5.31154051e-02 5.59510410e-01
4.12072241e-01 -4.64553684e-01 -7.00261414e-01 -3.83817643e-01
-2.37711549e-01 3.33823383e-01 -2.62521893e-01 -2.95521200e-01
-6.04336858e-01 -1.30676907e-02 3.39002669e-01 3.24500561e-01
6.93170309e-01 -8.19117963e-01 -6.61508083e-01 3.31101269e-01
-4.65588361e-01 -1.07701862e+00 -5.62351584e-01 -4.76877570e-01
-8.37942302e-01 -1.21145320e+00 -1.01976597e+00 -4.02147323e-01
4.60981429e-02 4.11780864e-01 1.11515880e+00 -2.51562204e-02
-2.77293563e-01 4.87992346e-01 -6.64143860e-01 9.75338742e-02
-4.40718144e-01 -4.84397054e-01 1.99698761e-01 3.40777606e-01
2.17291415e-01 -5.28830171e-01 -8.05332065e-01 5.91979146e-01
-1.02661467e+00 -3.81435335e-01 5.67357123e-01 6.27023876e-01
6.96917176e-01 3.54808033e-01 6.12094581e-01 -1.46663055e-01
5.15208006e-01 -2.52676666e-01 -7.02694774e-01 2.62291759e-01
-4.33982134e-01 -1.81699082e-01 7.33524382e-01 -1.31787091e-01
-9.32218194e-01 -5.96136212e-01 -2.99724936e-01 -4.88319963e-01
3.00047994e-02 3.35508615e-01 -8.94837499e-01 -1.32579252e-01
3.69211376e-01 3.37023079e-01 -9.08682868e-02 -2.17200086e-01
2.49441117e-01 7.47145891e-01 7.89873183e-01 -2.92409092e-01
9.03838933e-01 3.15301448e-01 9.24511477e-02 -1.04370880e+00
-1.88529387e-01 -6.35436475e-01 -3.59495550e-01 -5.59171379e-01
7.67904937e-01 -8.59566152e-01 -7.30870783e-01 4.51127350e-01
-9.03224051e-01 3.17261487e-01 -1.52077213e-01 7.31027901e-01
-6.61349416e-01 7.51433671e-01 -4.23755229e-01 -6.08840287e-01
-1.56089678e-01 -1.45504403e+00 5.49031913e-01 2.62814850e-01
3.73424515e-02 -7.92584836e-01 -6.02224730e-02 3.58539194e-01
4.60150450e-01 -1.92353912e-02 8.76973391e-01 -1.21469609e-01
-7.60070980e-01 -1.66313231e-01 -3.93040895e-01 7.76511848e-01
4.81898397e-01 4.29329053e-02 -7.90697217e-01 -3.75040352e-01
2.75485426e-01 -5.57340793e-02 3.88797402e-01 4.94617373e-01
1.04509079e+00 -2.01494738e-01 2.59718120e-01 6.11508250e-01
1.52329350e+00 6.01372898e-01 9.27981913e-01 1.10689014e-01
4.52208012e-01 1.82318792e-01 9.38392580e-01 6.72785640e-01
4.70196307e-02 9.61763263e-01 5.28133154e-01 2.71709293e-01
-3.97780947e-02 -5.69214225e-02 4.00845885e-01 1.12520814e+00
-3.37082326e-01 -3.67552996e-01 -5.17627656e-01 2.83757776e-01
-1.26521611e+00 -1.18130624e+00 2.15227202e-01 2.28166103e+00
5.89504302e-01 2.28634268e-01 9.95476320e-02 6.92278087e-01
5.88354826e-01 2.33222783e-01 -5.56363881e-01 -2.08520755e-01
-2.76749015e-01 7.90265873e-02 5.41074574e-02 1.45077884e-01
-9.96952772e-01 2.34804198e-01 5.73580217e+00 9.51704979e-01
-1.01699340e+00 -4.76338677e-02 6.12632811e-01 3.94129641e-02
-2.39076763e-01 -4.45678383e-01 -2.07587793e-01 5.03296018e-01
9.83342230e-01 -6.10022068e-01 3.74401987e-01 7.00219214e-01
5.12397826e-01 -1.17270663e-01 -1.01747572e+00 1.57378638e+00
2.22945958e-01 -1.14024198e+00 3.08609694e-01 8.66801962e-02
6.08245432e-01 -5.44447064e-01 3.95287603e-01 5.74997887e-02
-4.18276429e-01 -9.50277984e-01 7.81435192e-01 5.08038819e-01
1.05204916e+00 -1.01377809e+00 8.42689395e-01 -1.78494453e-01
-1.41901052e+00 -1.36421278e-01 -4.52962369e-01 1.90283313e-01
3.40549946e-01 5.58927059e-01 -1.59229353e-01 7.62800932e-01
8.08783174e-01 8.80164146e-01 -5.05263507e-01 1.49652910e+00
1.20967351e-01 6.37363970e-01 4.12568718e-01 2.15937391e-01
-5.55755384e-02 -3.42523515e-01 6.50410950e-01 7.84849465e-01
6.42503500e-01 3.95260453e-01 -2.71742225e-01 3.50470901e-01
-1.65424883e-01 3.35166276e-01 -2.81127930e-01 -2.47749731e-01
1.53806046e-01 9.98182893e-01 -3.19015861e-01 -9.97834578e-02
-6.81743562e-01 9.96515274e-01 -3.60077620e-01 2.59970009e-01
-9.20015454e-01 -4.80880290e-01 1.02250910e+00 4.41038795e-02
4.10932511e-01 -2.62282252e-01 1.93452165e-01 -1.15321338e+00
3.52483168e-02 -1.30090070e+00 7.46733621e-02 -7.91185081e-01
-1.20928133e+00 7.25281715e-01 -2.88010508e-01 -1.85265338e+00
-2.19845101e-01 -4.18672025e-01 -3.97724360e-02 7.57555902e-01
-1.47125304e+00 -4.20571178e-01 -5.29599786e-01 7.28872299e-01
7.43982196e-01 -2.93942660e-01 6.25940204e-01 6.11047268e-01
-8.78365636e-02 6.49084032e-01 2.95259982e-01 2.46616647e-01
6.75097167e-01 -8.56821477e-01 1.98568448e-01 9.24794376e-01
1.41673625e-01 2.48253718e-01 8.41407418e-01 -2.71657348e-01
-1.68751466e+00 -8.41049433e-01 3.06228340e-01 4.04571034e-02
4.76774663e-01 1.35556921e-01 -1.06442225e+00 -5.75846396e-02
1.65305734e-01 6.93400428e-02 6.02893353e-01 -6.36463284e-01
-3.53266150e-01 -6.09299421e-01 -1.14934111e+00 3.84899318e-01
8.06240678e-01 -6.96647823e-01 -3.57206523e-01 -3.45822364e-01
5.57447791e-01 -1.01310447e-01 -1.06309807e+00 4.46371764e-01
5.98748088e-01 -1.36134267e+00 9.65503037e-01 -1.32777154e-01
4.97989953e-01 -3.21370006e-01 -6.36496425e-01 -1.28584278e+00
-2.75116622e-01 -4.39660937e-01 -2.06181899e-01 1.04558969e+00
-1.64191395e-01 -1.17864296e-01 5.51319659e-01 1.95714921e-01
4.11857385e-03 -5.14828503e-01 -1.02231824e+00 -7.84746110e-01
-2.78591335e-01 -5.14394045e-01 5.54958284e-01 4.85325009e-01
-4.60338257e-02 -8.78548995e-02 -4.77479875e-01 6.77480251e-02
7.38956988e-01 -1.95156902e-01 5.85977912e-01 -1.15497577e+00
-3.62072468e-01 -5.62158823e-01 -1.12955976e+00 -1.14918172e+00
-2.48299494e-01 -3.87732446e-01 9.67652444e-03 -1.35056257e+00
1.73691884e-01 -1.57314882e-01 -6.11335039e-01 -3.66879821e-01
-1.45667985e-01 4.39314425e-01 2.65628308e-01 6.77167997e-02
-7.30642319e-01 8.32904875e-01 1.40702534e+00 -3.96905690e-01
7.52555504e-02 1.35934353e-02 -3.88191819e-01 4.09357905e-01
6.28766596e-01 1.33530982e-02 -6.47431791e-01 -2.22058237e-01
2.99019456e-01 5.41185498e-01 1.64200217e-01 -1.51288807e+00
-1.62494659e-01 -1.78126872e-01 3.26940387e-01 -3.69537383e-01
5.21080434e-01 -9.91168380e-01 1.65798575e-01 1.85895607e-01
-1.36031643e-01 2.43329421e-01 1.33082923e-03 6.16754591e-01
-8.12503040e-01 1.74820915e-01 8.53051424e-01 1.80624053e-01
-1.09327126e+00 3.97217453e-01 -2.48579547e-01 1.66837156e-01
9.95727479e-01 -2.67791897e-01 -6.08379357e-02 -7.48890400e-01
-3.55393738e-01 -2.13620260e-01 6.95616245e-01 5.26737809e-01
1.22325146e+00 -1.41879535e+00 -7.26815820e-01 1.48047492e-01
2.04891324e-01 -6.30255163e-01 6.02890909e-01 5.83894491e-01
-5.70963204e-01 5.48743367e-01 -5.77886045e-01 -6.22949779e-01
-1.41404247e+00 6.50265992e-01 1.10019788e-01 1.50396014e-02
-4.18919772e-01 4.68414754e-01 -6.30267721e-04 5.47345102e-01
2.23652974e-01 -2.02023521e-01 -4.83334988e-01 -1.39394104e-01
9.98495102e-01 6.49500132e-01 2.39219949e-01 -1.17686439e+00
-2.45404556e-01 5.24041414e-01 2.64520705e-01 -1.71043515e-01
1.15126145e+00 -4.45616394e-01 2.74256945e-01 4.07524556e-01
1.49881995e+00 1.21952705e-01 -1.32065427e+00 -1.45337045e-01
-8.47049877e-02 -9.91023242e-01 1.23184621e-01 -5.28396547e-01
-1.28649020e+00 1.00155890e+00 9.69223559e-01 1.77020922e-01
1.53306067e+00 -3.76946777e-01 7.62018263e-01 -3.11191212e-02
7.83422410e-01 -9.29949820e-01 4.78307575e-01 1.06556118e-01
9.66396093e-01 -1.17395580e+00 5.43645583e-02 -3.33218038e-01
-6.67361140e-01 9.93706346e-01 2.11327210e-01 2.31805623e-01
5.14395237e-01 -2.48732835e-01 1.94928676e-01 2.23360896e-01
-4.68507320e-01 1.51600361e-01 6.15398884e-01 7.76233494e-01
4.33807045e-01 -4.33956720e-02 -2.07812995e-01 4.85850513e-01
1.25380205e-02 1.46243155e-01 5.36191583e-01 6.21090174e-01
-4.96178776e-01 -1.05985510e+00 -4.26367998e-01 4.92463052e-01
-5.47938049e-01 9.99433398e-02 3.26347560e-01 4.71096069e-01
-1.07239418e-01 1.36460888e+00 -6.36460558e-02 -6.08775795e-01
4.36964780e-01 -4.17701691e-01 6.20118141e-01 -2.88732741e-02
-1.38077810e-01 5.40573820e-02 -3.59999329e-01 -1.01885927e+00
-5.40260375e-01 -6.00591600e-01 -1.10423028e+00 -3.83428276e-01
3.20004895e-02 3.38007152e-01 6.34634256e-01 7.26156592e-01
1.65477186e-01 4.52855259e-01 1.07986164e+00 -8.44860435e-01
-4.04296994e-01 -5.73523164e-01 -7.73516119e-01 7.09224820e-01
4.46699768e-01 -7.14693725e-01 -2.92887926e-01 1.99469492e-01] | [11.74120044708252, -1.8338137865066528] |
794d0b05-d2f0-47d3-a188-5da581b4bcc1 | towards-predicting-fine-finger-motions-from | 2202.05204 | null | https://arxiv.org/abs/2202.05204v2 | https://arxiv.org/pdf/2202.05204v2.pdf | Towards Predicting Fine Finger Motions from Ultrasound Images via Kinematic Representation | A central challenge in building robotic prostheses is the creation of a sensor-based system able to read physiological signals from the lower limb and instruct a robotic hand to perform various tasks. Existing systems typically perform discrete gestures such as pointing or grasping, by employing electromyography (EMG) or ultrasound (US) technologies to analyze muscle states. While estimating finger gestures has been done in the past by detecting prominent gestures, we are interested in detection, or inference, done in the context of fine motions that evolve over time. Examples include motions occurring when performing fine and dexterous tasks such as keyboard typing or piano playing. We consider this task as an important step towards higher adoption rates of robotic prostheses among arm amputees, as it has the potential to dramatically increase functionality in performing daily tasks. To this end, we present an end-to-end robotic system, which can successfully infer fine finger motions. This is achieved by modeling the hand as a robotic manipulator and using it as an intermediate representation to encode muscles' dynamics from a sequence of US images. We evaluated our method by collecting data from a group of subjects and demonstrating how it can be used to replay music played or text typed. To the best of our knowledge, this is the first study demonstrating these downstream tasks within an end-to-end system. | ['Alex M. Bronstein', 'Alon Wolf', 'Oren Salzman', 'Dean Zadok'] | 2022-02-10 | null | null | null | null | ['electromyography-emg'] | ['medical'] | [ 3.18274945e-01 1.66513160e-01 -3.37892532e-01 6.33191764e-02
-3.40253562e-01 -8.17476332e-01 3.69805664e-01 -8.96805465e-01
-3.70358050e-01 6.00560248e-01 2.92660773e-01 -8.76367465e-02
-2.38306984e-01 -8.77823457e-02 -5.87425590e-01 -3.24217379e-01
-2.45602816e-01 4.58324105e-01 2.03212604e-01 -3.13356549e-01
2.77840793e-01 7.21552253e-01 -1.33916724e+00 4.07925099e-01
3.43346447e-01 7.00715601e-01 6.28688514e-01 7.94057369e-01
3.11749011e-01 6.83044076e-01 -5.06030083e-01 1.29993930e-01
3.51819038e-01 -2.37415150e-01 -6.82588577e-01 -1.32830903e-01
3.46254297e-02 -7.00726271e-01 -5.85429072e-01 5.81241190e-01
8.33224297e-01 8.10230672e-02 5.37050843e-01 -8.44625890e-01
-2.45501965e-01 6.36912525e-01 -1.14582919e-01 -1.54504403e-01
6.15286112e-01 6.93651497e-01 6.24976516e-01 -3.51636797e-01
8.52259159e-01 1.15951610e+00 4.67372090e-01 1.01959455e+00
-1.20544159e+00 -5.95077753e-01 -2.74301916e-01 8.20852891e-02
-7.72481561e-01 -6.51117504e-01 9.09815133e-01 -7.28524625e-01
1.01356709e+00 4.25918460e-01 8.00170362e-01 1.57003319e+00
3.63045275e-01 1.06201148e+00 7.74266005e-01 -4.16163146e-01
6.09710813e-02 -6.40109003e-01 -1.58243015e-01 1.43392965e-01
-2.27868557e-03 1.76029950e-01 -5.42836428e-01 -7.14103281e-02
1.33449173e+00 -1.26789380e-02 -6.71927989e-01 -4.83743958e-02
-1.55603504e+00 -9.15240496e-02 2.89007604e-01 3.98140073e-01
-1.13945127e+00 8.09167922e-01 1.76181540e-01 4.29173380e-01
-2.87820309e-01 7.25028455e-01 -4.79233354e-01 -1.02391648e+00
-5.87233484e-01 4.06534165e-01 1.06849229e+00 7.54062653e-01
-4.16404933e-01 -2.21836552e-01 -3.31426710e-01 7.13821769e-01
2.39938483e-01 1.60647035e-01 6.25460088e-01 -1.57176363e+00
4.31986988e-01 4.92649138e-01 3.48968923e-01 -2.42922619e-01
-5.45033157e-01 3.09571829e-02 -1.75493777e-01 8.46878767e-01
6.94024026e-01 -2.43956015e-01 -8.70917022e-01 1.66314626e+00
2.04401389e-02 -2.76057810e-01 -1.73427790e-01 1.31391060e+00
5.95228411e-02 -1.70656681e-01 1.61411509e-01 -4.61357981e-02
1.33480740e+00 -1.60085931e-01 -7.25259483e-01 -1.98381573e-01
2.04816177e-01 -5.38754821e-01 1.09488487e+00 1.00421071e+00
-1.05130661e+00 -4.88906354e-01 -7.75545359e-01 1.56107888e-01
1.55759469e-01 2.51403481e-01 6.73752725e-01 3.23404074e-01
-5.28104782e-01 1.19656622e+00 -1.52556968e+00 -3.56355578e-01
1.23253137e-01 8.66922319e-01 -3.12806815e-01 3.48040968e-01
-8.69462490e-01 1.06297791e+00 1.69629604e-01 2.67872453e-01
-2.93416679e-01 -5.55535257e-01 -2.26051703e-01 -3.27988535e-01
1.52567387e-01 -8.51034641e-01 1.33718121e+00 -4.77129817e-01
-2.03961897e+00 7.53571093e-01 9.14167166e-02 -7.55034271e-04
7.83821881e-01 -6.82973742e-01 -1.03231944e-01 1.59653559e-01
-3.96005630e-01 5.65673351e-01 7.78316498e-01 -6.99015796e-01
5.04960231e-02 -7.63652444e-01 -6.85177669e-02 1.73633501e-01
-2.59850770e-01 1.66466814e-02 -4.64956135e-01 -8.92489851e-01
2.66360670e-01 -1.38461387e+00 1.84092328e-01 5.04601955e-01
-4.15383488e-01 -2.42717460e-01 6.62371755e-01 -9.78210926e-01
7.98728228e-01 -1.94918549e+00 7.90729821e-01 2.09954783e-01
4.09256890e-02 2.17457592e-01 1.33213580e-01 5.47177374e-01
1.02391131e-01 -3.00065458e-01 1.84093952e-01 2.61607289e-01
2.18605474e-02 2.66587436e-01 -1.27854243e-01 2.44128719e-01
-6.52845064e-03 1.05972886e+00 -7.94453144e-01 -4.13578898e-02
3.29761177e-01 4.06925112e-01 -3.11809659e-01 2.63150454e-01
-3.52665395e-01 1.04244995e+00 -4.67342854e-01 7.36422300e-01
-1.44532695e-01 1.85001940e-01 5.55086911e-01 -5.14408290e-01
1.06429227e-01 3.01612347e-01 -9.56938326e-01 2.09126782e+00
-3.64355743e-01 5.74843705e-01 4.53671455e-01 -7.55082667e-01
7.70249069e-01 7.73345709e-01 9.09417450e-01 -3.72100711e-01
5.04189849e-01 4.85152870e-01 5.06066918e-01 -1.17404294e+00
-7.24756792e-02 5.90091338e-03 1.79046467e-01 5.03269851e-01
-2.02617764e-01 -3.65271449e-01 -6.94280192e-02 -4.52982098e-01
1.49699712e+00 1.06309593e+00 -3.94172706e-02 2.06749052e-01
-1.02390954e-02 2.68148571e-01 -5.22138663e-02 3.81815761e-01
-9.19739902e-02 5.59911609e-01 8.64056125e-02 1.09913468e-01
-8.57262313e-01 -1.20079124e+00 1.88403457e-01 7.98682928e-01
-2.74851114e-01 5.29624075e-02 -5.57672501e-01 2.61064082e-01
5.10023057e-01 4.31682080e-01 -1.33834749e-01 -2.36508608e-01
-8.68281484e-01 5.14844209e-02 5.92914462e-01 8.84709895e-01
8.59090611e-02 -1.64097977e+00 -1.17864859e+00 5.43279827e-01
-1.94229245e-01 -1.07559621e+00 -1.00636169e-01 1.66910604e-01
-1.19178224e+00 -1.07128143e+00 -1.08697653e+00 -7.62749195e-01
2.57120430e-01 -3.04871798e-01 4.08950359e-01 -4.89718586e-01
-5.55219889e-01 6.52591348e-01 -4.39040452e-01 -3.08287889e-01
-4.36862558e-01 -2.35255510e-02 5.10299087e-01 -5.27511179e-01
2.03884840e-01 -1.16392088e+00 -7.53450334e-01 3.38922858e-01
-5.29964387e-01 9.31812376e-02 1.09023821e+00 5.90440094e-01
3.96022409e-01 -6.87289476e-01 3.34531486e-01 -1.39335841e-01
1.30545676e+00 -1.40581727e-01 2.57013664e-02 8.14955682e-02
-3.96310501e-02 2.41244867e-01 4.30332810e-01 -1.03904247e+00
-8.85052383e-01 5.99262416e-01 -2.45226964e-01 -6.90284133e-01
-2.06940919e-01 4.70498204e-01 1.68317065e-01 -4.30867858e-02
8.68686259e-01 1.95694089e-01 6.28355742e-01 -6.30121469e-01
4.07307416e-01 1.29876387e+00 1.27020657e+00 -8.68121743e-01
2.70387292e-01 2.64522731e-01 2.52107114e-01 -8.21255505e-01
1.64077133e-01 -5.51631749e-01 -1.10723674e+00 -5.67570984e-01
4.19385761e-01 -3.15623820e-01 -1.45408034e+00 5.65852284e-01
-1.06619132e+00 -8.58433485e-01 -1.12448826e-01 9.76579368e-01
-1.13079727e+00 2.68123806e-01 -8.42026651e-01 -9.99604404e-01
-4.96423125e-01 -9.94996965e-01 1.24499929e+00 -1.31825387e-01
-1.20743823e+00 -1.67308956e-01 -3.95262130e-02 2.45667070e-01
3.02657753e-01 5.65813422e-01 4.58656937e-01 -6.78208843e-02
-1.40133500e-01 -6.05448008e-01 4.70176369e-01 1.31798923e-01
3.39253068e-01 -4.57870543e-01 -7.47316122e-01 -3.54174554e-01
-4.83115809e-03 -5.61336458e-01 1.58955529e-01 3.61796707e-01
1.17320657e+00 4.08052132e-02 -3.71351391e-01 1.22963049e-01
7.07004488e-01 3.60480964e-01 8.86200726e-01 5.55465966e-02
5.52506864e-01 8.75054836e-01 6.08816147e-01 1.20878674e-01
-2.90747941e-01 1.12331986e+00 1.72945932e-01 7.12507904e-01
-4.11829680e-01 4.87837866e-02 4.87816662e-01 4.55877185e-01
-9.52524424e-01 3.24888587e-01 -7.34886050e-01 1.18459433e-01
-1.88245726e+00 -9.12350178e-01 -1.09405175e-01 1.97420466e+00
9.49745834e-01 1.35640949e-01 1.40826553e-01 3.58334601e-01
2.41652653e-01 -6.57209694e-01 -1.25553441e+00 -6.23540804e-02
5.55719316e-01 5.62440455e-01 2.55006790e-01 5.98868355e-04
-3.87726814e-01 7.73445845e-01 5.79941320e+00 2.07034290e-01
-1.34437430e+00 -2.56437033e-01 -4.73117650e-01 -4.05653954e-01
4.85825717e-01 -3.14302653e-01 -1.37001276e-01 6.12172961e-01
6.28381193e-01 4.12695795e-01 8.24207067e-01 6.16172433e-01
5.97620726e-01 -1.91792883e-02 -1.61198306e+00 1.06587100e+00
-3.33658189e-01 -8.31387520e-01 -5.62299728e-01 4.73813340e-02
6.07737489e-02 9.74603742e-02 -3.41365129e-01 3.02164517e-02
-6.54314682e-02 -9.04128671e-01 6.67087615e-01 9.57532287e-01
1.02975523e+00 8.02236199e-02 1.24745257e-02 6.66652262e-01
-9.62334931e-01 -2.29776606e-01 1.89144790e-01 -4.19967413e-01
3.91217411e-01 -4.69685718e-02 -8.35990787e-01 2.02759691e-02
3.76040399e-01 5.18833995e-01 2.21796602e-01 8.30353439e-01
-3.80733848e-01 4.72843915e-01 -4.98111278e-01 -4.02531743e-01
-3.98007393e-01 2.80265622e-02 9.10578370e-01 7.08251297e-01
3.44883174e-01 3.15406740e-01 -1.30633697e-01 1.13284397e+00
2.12800831e-01 -6.12562299e-01 -4.22456831e-01 -4.43313479e-01
3.01760763e-01 9.71917212e-01 -3.47865671e-01 7.33554885e-02
8.38289112e-02 1.50075865e+00 -3.75106670e-02 4.22954589e-01
-2.59307325e-01 -6.61525190e-01 8.00094068e-01 -3.87794226e-02
-1.77939400e-01 -9.03650522e-01 -3.51534992e-01 -1.00740600e+00
7.53803909e-01 -8.58427584e-01 -1.67962044e-01 -1.21407509e+00
-9.41877663e-01 -1.29449129e-01 -1.27181470e-01 -1.47129059e+00
-9.87152100e-01 -1.22707725e+00 -4.39749539e-01 1.01073778e+00
-5.73702693e-01 -6.80518508e-01 -4.74933088e-01 6.30893767e-01
6.27691269e-01 2.25917473e-01 9.36631024e-01 -4.56190668e-02
-7.86791295e-02 1.67884961e-01 -1.95663035e-01 6.58665821e-02
6.79913163e-01 -9.81127858e-01 2.48738363e-01 3.89838576e-01
-2.20650420e-01 1.09908962e+00 7.31547654e-01 -8.49654198e-01
-2.13338494e+00 -1.69380352e-01 1.44644037e-01 -5.65181792e-01
5.66260278e-01 -1.85715884e-01 -5.10018289e-01 6.65160120e-01
-5.37061870e-01 -1.59883097e-01 1.55267656e-01 -2.35352859e-01
2.00718582e-01 2.12922975e-01 -9.78791177e-01 9.25636649e-01
1.66046762e+00 -5.60851157e-01 -1.04613650e+00 3.10399860e-01
7.90749863e-02 -6.04471028e-01 -1.28802001e+00 4.28936958e-01
1.77344680e+00 -2.50526041e-01 1.08642673e+00 -6.96707785e-01
4.33759272e-01 -1.58295155e-01 4.94335182e-02 -1.44240081e+00
-1.63647234e-01 -8.34980667e-01 -7.09138274e-01 8.15903544e-01
-3.68754268e-02 -4.36295867e-01 9.88550305e-01 8.25230479e-01
-6.60348609e-02 -5.89533091e-01 -8.20501328e-01 -1.01162064e+00
-2.05430120e-01 -6.49487376e-01 1.83118537e-01 2.70589083e-01
8.73630822e-01 -2.03703240e-01 -4.31601554e-01 -5.73663302e-02
3.90293241e-01 3.36797200e-02 9.56643522e-01 -1.35262382e+00
-7.33286619e-01 -6.28087103e-01 -6.00142300e-01 -1.11106849e+00
-1.08013786e-01 -8.22097778e-01 2.61210769e-01 -1.55719459e+00
-2.00509652e-01 -2.84091562e-01 -7.43714347e-02 4.71561849e-01
2.65079170e-01 7.04901963e-02 1.41869545e-01 8.57578814e-01
3.33797187e-01 -6.64514452e-02 1.52021158e+00 2.09460855e-02
-5.68966329e-01 3.54106426e-01 -2.36595944e-01 3.63248289e-01
9.93656576e-01 -1.40007392e-01 -1.33929163e-01 -1.00960530e-01
-1.38339475e-01 4.96068120e-01 5.18170238e-01 -1.09302449e+00
3.50557864e-01 -6.56251563e-04 6.25101209e-01 -2.21110523e-01
5.58529258e-01 -7.96943307e-01 6.97630107e-01 1.09960687e+00
-4.53224003e-01 -4.68082190e-01 -1.22401953e-01 2.91431606e-01
8.50383937e-02 9.16245431e-02 1.50048835e-02 -2.42209703e-01
-8.94568384e-01 -2.04580083e-01 -4.01689529e-01 -5.31027615e-01
9.29486275e-01 -6.65492892e-01 -2.75329966e-02 -1.63639873e-01
-1.22574532e+00 9.98333003e-03 1.81065455e-01 7.79707074e-01
4.56116199e-01 -1.11715174e+00 -3.99611175e-01 3.03007543e-01
-3.03376708e-02 -4.02243078e-01 -1.92273051e-01 1.04944873e+00
-4.64990199e-01 4.47223395e-01 -7.88962305e-01 -6.71252012e-01
-1.21074259e+00 -6.92900196e-02 2.65891463e-01 1.67146236e-01
-1.01305747e+00 3.53655219e-01 -5.92183709e-01 -3.78979117e-01
4.54521745e-01 -4.70372885e-01 -7.66515881e-02 -5.03959298e-01
1.32010922e-01 3.80332768e-01 -2.42206156e-02 -1.38549041e-02
-1.70159370e-01 6.16868675e-01 5.10843277e-01 -5.05198956e-01
1.48517895e+00 2.55060196e-01 1.14086621e-01 4.89610136e-01
6.89258158e-01 -3.11951607e-01 -1.51928198e+00 2.71016747e-01
-3.45284827e-02 -2.79779345e-01 -2.50430107e-01 -1.28415167e+00
-6.40634477e-01 7.66564548e-01 7.00218856e-01 -1.89676896e-01
1.03003597e+00 2.02206507e-01 8.80692661e-01 8.33517373e-01
8.84126067e-01 -1.21905243e+00 1.48862123e-01 -2.02910211e-02
1.47128141e+00 -7.76090264e-01 -1.81783333e-01 -3.63383919e-01
-2.49080881e-01 1.54979849e+00 3.51354510e-01 -2.62746453e-01
4.14083332e-01 6.14599228e-01 -1.72160476e-01 -2.57016778e-01
-1.93724453e-01 -6.67295083e-02 6.32946312e-01 5.36052048e-01
6.80765331e-01 3.51200759e-01 -7.16608703e-01 6.10141039e-01
-2.74480343e-01 1.04669869e+00 8.18871856e-02 1.56057930e+00
-1.84407651e-01 -1.19331026e+00 -4.08291191e-01 8.38045239e-01
-4.49320525e-01 5.38046181e-01 -5.62485099e-01 7.52344131e-01
-3.27888340e-01 7.32738078e-01 -2.54442543e-01 -6.94462121e-01
7.55019665e-01 4.21415597e-01 1.24418938e+00 -3.94953460e-01
-5.10844529e-01 -8.79249275e-02 1.48791641e-01 -1.08235991e+00
-3.12494248e-01 -8.02744269e-01 -1.38305843e+00 2.17034295e-01
4.41397987e-02 -6.97019637e-01 1.10915709e+00 7.18373179e-01
2.66490340e-01 8.57802510e-01 -1.20870702e-01 -1.84278166e+00
-1.10755241e+00 -1.43865216e+00 -5.93574703e-01 7.70004511e-01
3.72469164e-02 -9.08644795e-01 2.23548096e-02 1.37709230e-01] | [6.794686317443848, 0.1579710990190506] |
c7a7c7eb-1cca-4f52-958a-02f5ec6ab14b | a-simple-baseline-for-direct-2d-multi-person | 2302.01110 | null | https://arxiv.org/abs/2302.01110v2 | https://arxiv.org/pdf/2302.01110v2.pdf | DirectMHP: Direct 2D Multi-Person Head Pose Estimation with Full-range Angles | Existing head pose estimation (HPE) mainly focuses on single person with pre-detected frontal heads, which limits their applications in real complex scenarios with multi-persons. We argue that these single HPE methods are fragile and inefficient for Multi-Person Head Pose Estimation (MPHPE) since they rely on the separately trained face detector that cannot generalize well to full viewpoints, especially for heads with invisible face areas. In this paper, we focus on the full-range MPHPE problem, and propose a direct end-to-end simple baseline named DirectMHP. Due to the lack of datasets applicable to the full-range MPHPE, we firstly construct two benchmarks by extracting ground-truth labels for head detection and head orientation from public datasets AGORA and CMU Panoptic. They are rather challenging for having many truncated, occluded, tiny and unevenly illuminated human heads. Then, we design a novel end-to-end trainable one-stage network architecture by joint regressing locations and orientations of multi-head to address the MPHPE problem. Specifically, we regard pose as an auxiliary attribute of the head, and append it after the traditional object prediction. Arbitrary pose representation such as Euler angles is acceptable by this flexible design. Then, we jointly optimize these two tasks by sharing features and utilizing appropriate multiple losses. In this way, our method can implicitly benefit from more surroundings to improve HPE accuracy while maintaining head detection performance. We present comprehensive comparisons with state-of-the-art single HPE methods on public benchmarks, as well as superior baseline results on our constructed MPHPE datasets. Datasets and code are released in https://github.com/hnuzhy/DirectMHP. | ['Hongtao Lu', 'Fei Jiang', 'Huayi Zhou'] | 2023-02-02 | null | null | null | null | ['head-detection', 'head-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-4.02510405e-01 2.71653980e-01 2.30428815e-01 -5.64364135e-01
-1.00443649e+00 -2.07319021e-01 2.36832976e-01 -6.14822626e-01
-4.62041676e-01 5.41612566e-01 3.43643248e-01 2.05931023e-01
3.20411444e-01 -4.03800189e-01 -8.04686725e-01 -7.42932498e-01
-3.98487262e-02 6.05966389e-01 2.21383154e-01 -9.86781940e-02
-3.09782267e-01 5.42698860e-01 -1.79680669e+00 -1.45435005e-01
5.40159643e-01 7.72450030e-01 5.99257164e-02 5.20132244e-01
5.98116815e-01 2.43306503e-01 -5.32644868e-01 -6.42419100e-01
2.74353892e-01 9.89390165e-02 -3.21590424e-01 1.03827812e-01
9.78462219e-01 -7.08599865e-01 -5.15296876e-01 7.28388131e-01
1.30011654e+00 4.18804847e-02 5.09189785e-01 -1.71074009e+00
-4.17624742e-01 2.57446945e-01 -9.33010817e-01 -2.25304693e-01
7.35467553e-01 4.19239789e-01 6.97678447e-01 -1.33108163e+00
1.91122547e-01 1.58586156e+00 1.02950847e+00 8.23299885e-01
-6.54689431e-01 -1.00821102e+00 4.10848081e-01 1.33731678e-01
-1.83424652e+00 -9.18036222e-01 6.94220304e-01 -1.71965122e-01
7.20900476e-01 2.37319589e-01 4.81688440e-01 1.32280993e+00
-2.80792236e-01 1.14071858e+00 9.05719340e-01 -9.30678993e-02
-1.79510593e-01 -7.22380877e-02 2.85728071e-02 8.89652848e-01
2.52828807e-01 1.24419257e-02 -5.01394868e-01 -3.36888693e-02
3.70705515e-01 -1.37274675e-02 -6.27150178e-01 -2.95118213e-01
-8.29721451e-01 7.29431093e-01 5.24757564e-01 -3.09713781e-01
-2.32774258e-01 -3.05919886e-01 1.88639984e-01 -3.72949362e-01
3.86554360e-01 -3.34611207e-01 -4.96976942e-01 2.99484015e-01
-9.63295341e-01 6.12724304e-01 7.42070556e-01 1.30286264e+00
4.03762996e-01 -1.67224392e-01 -3.15008223e-01 8.86026144e-01
6.40005529e-01 9.31101382e-01 1.13406859e-01 -7.77732432e-01
4.99446958e-01 2.83409148e-01 5.45661375e-02 -7.54875720e-01
-9.66158748e-01 -3.82465661e-01 -7.34667540e-01 -8.71800482e-02
4.40581650e-01 -2.98662245e-01 -9.85775709e-01 1.99966931e+00
8.35699379e-01 -2.93224715e-02 -3.94888967e-01 1.10146487e+00
1.26366556e+00 4.91150737e-01 7.21951202e-03 -2.14760423e-01
1.65095484e+00 -1.13953030e+00 -6.05370998e-01 -6.21686101e-01
3.29327196e-01 -4.66314286e-01 1.08505511e+00 3.04165155e-01
-1.09462786e+00 -3.37295294e-01 -6.94248676e-01 -3.81490767e-01
-2.31907859e-01 4.85354036e-01 3.56724918e-01 6.89594686e-01
-1.31282628e+00 -6.38766736e-02 -7.51665890e-01 -4.33858752e-01
4.49823797e-01 7.19840944e-01 -3.98476809e-01 -1.48472548e-01
-9.81840014e-01 7.64185429e-01 -1.43721411e-02 4.88263369e-01
-7.40114212e-01 -5.17238915e-01 -1.05898666e+00 6.19615093e-02
5.45562863e-01 -8.23616385e-01 1.46943378e+00 -2.23154843e-01
-1.53365254e+00 9.76216197e-01 -3.98298144e-01 -6.87054172e-02
8.61662924e-01 -5.25494099e-01 -2.42117956e-01 -1.28567085e-01
2.30884314e-01 9.44229603e-01 7.64982879e-01 -1.25431299e+00
-5.58470905e-01 -8.29591274e-01 -2.25848392e-01 1.90776348e-01
-3.28823060e-01 3.54065955e-01 -9.81096804e-01 -3.40030253e-01
-6.49857596e-02 -1.08672297e+00 1.63664728e-01 2.61826038e-01
-7.70862341e-01 -4.88674521e-01 9.14327085e-01 -9.19854641e-01
1.00730932e+00 -1.95567679e+00 -1.77339390e-01 -1.68250293e-01
3.15322995e-01 2.68005729e-01 -1.58149302e-02 -9.72660482e-02
1.70425009e-02 -3.55128556e-01 -1.62007868e-01 -1.14675844e+00
2.75926083e-01 -6.99681565e-02 -4.42521907e-02 8.78512263e-01
-1.15804449e-01 7.75415301e-01 -5.63899100e-01 -8.31447959e-01
-2.32207980e-02 9.25233364e-01 -7.58047998e-01 4.51045722e-01
1.55284718e-01 4.78168339e-01 -3.16769570e-01 1.05383098e+00
1.17870438e+00 -5.91052733e-02 -1.88629195e-01 -3.87974799e-01
-1.23705134e-01 1.63127072e-02 -1.16068280e+00 1.23306763e+00
-3.23378980e-01 3.13638240e-01 4.83274341e-01 -3.59271497e-01
6.32079899e-01 3.54828268e-01 2.36578763e-01 -4.19107139e-01
2.93295920e-01 3.43782902e-02 -2.71652877e-01 -4.24822301e-01
3.12472045e-01 1.11519024e-02 -6.12108856e-02 1.87871773e-02
4.70383018e-02 2.54787266e-01 -5.90491034e-02 -7.45970532e-02
6.99060082e-01 1.39575630e-01 2.41241828e-01 -3.52980532e-02
6.05415225e-01 -8.60540509e-01 8.93096566e-01 4.34702635e-01
-5.58750927e-01 1.20032024e+00 1.88485980e-01 -1.20715387e-01
-6.01897776e-01 -1.04397094e+00 -2.78158695e-01 1.59262240e+00
-6.08740933e-02 -3.60230416e-01 -1.05607021e+00 -6.63800657e-01
-3.56018450e-03 4.76461291e-01 -5.14808476e-01 2.13984042e-01
-9.53359425e-01 -1.06120324e+00 6.15214586e-01 7.52747297e-01
5.88359833e-01 -9.60540593e-01 -3.10949564e-01 2.23659202e-02
-4.60238606e-01 -1.38032639e+00 -8.46964002e-01 -3.15612227e-01
-1.69000402e-01 -9.36767817e-01 -1.20082688e+00 -7.23422050e-01
5.52955508e-01 3.01105261e-01 1.08286941e+00 2.78607998e-02
-1.90782890e-01 4.16470915e-01 -3.70565988e-03 -6.57105207e-01
3.05783987e-01 1.71079129e-01 4.59223479e-01 4.29084599e-02
3.49988669e-01 -4.22559530e-01 -1.07880723e+00 5.30439198e-01
-2.04444647e-01 -6.87694922e-02 5.93465626e-01 4.98540431e-01
3.52497548e-01 -5.24526119e-01 5.01314938e-01 -4.90048259e-01
1.15828961e-01 -3.82149041e-01 -6.00338995e-01 3.26762974e-01
-1.56679600e-01 -4.60456640e-01 3.56570333e-01 -3.63974869e-01
-1.19175696e+00 3.22910219e-01 -5.60269892e-01 -4.01445746e-01
-2.72765666e-01 -3.03851038e-01 -9.04803932e-01 -4.72196937e-02
3.21776003e-01 1.67419150e-01 -1.01114817e-01 -6.05556309e-01
1.24009863e-01 7.87104130e-01 8.33529949e-01 -4.64292079e-01
7.02954650e-01 4.96853381e-01 -1.24124810e-01 -8.90240550e-01
-9.43064392e-01 -4.47880358e-01 -5.55594087e-01 -1.65012807e-01
8.24868560e-01 -1.25826168e+00 -1.10412347e+00 8.89512062e-01
-1.29171729e+00 -2.05827817e-01 4.23596412e-01 2.73849517e-01
-3.10364515e-01 4.04716820e-01 -6.81130469e-01 -1.04727256e+00
-7.04003990e-01 -1.11994028e+00 1.86387265e+00 3.75478566e-01
5.35592996e-02 -4.36239570e-01 -1.59305677e-01 5.39048195e-01
1.75321236e-01 1.26646474e-01 8.51527005e-02 -4.72542405e-01
-4.70212221e-01 -4.02988911e-01 -2.14130208e-01 -9.19856206e-02
-2.83637851e-01 -7.38241225e-02 -1.44910371e+00 -8.39995742e-01
-1.40545174e-01 -3.89467448e-01 7.84817457e-01 5.43337107e-01
1.35345471e+00 -4.28361654e-01 -5.91988683e-01 1.02379429e+00
8.17475677e-01 -2.84459442e-01 4.21396941e-01 1.77428752e-01
1.03083134e+00 8.15058172e-01 4.26811904e-01 6.32204890e-01
1.09211218e+00 9.33809757e-01 3.65020573e-01 -2.17718557e-01
-2.23611444e-01 -4.03574318e-01 7.11868525e-01 5.09097219e-01
-6.46478385e-02 -5.68434536e-01 -7.75510550e-01 2.96079546e-01
-1.60688210e+00 -6.51435614e-01 1.09994030e-02 2.15786529e+00
5.28950810e-01 -7.91274756e-02 6.96297228e-01 -1.00881867e-01
9.95343983e-01 1.56994700e-01 -6.30682588e-01 4.35151875e-01
-1.36856765e-01 -1.68883234e-01 3.87082040e-01 3.19594443e-01
-1.28968382e+00 8.64029586e-01 5.41524267e+00 6.06845737e-01
-1.06374848e+00 3.82192463e-01 5.83278954e-01 -3.96762758e-01
3.19616824e-01 -5.03417253e-01 -1.64850903e+00 5.48800290e-01
6.76247418e-01 3.47276032e-01 8.82373303e-02 9.56366956e-01
1.01069584e-01 1.82980895e-01 -1.18718612e+00 1.21224964e+00
3.96256775e-01 -3.56043816e-01 -3.18037391e-01 2.03614548e-01
1.99142948e-01 1.36038139e-01 2.47536108e-01 6.46346927e-01
-3.18069309e-02 -9.58508134e-01 8.85615945e-01 2.84760326e-01
8.36298168e-01 -5.34768522e-01 5.86953223e-01 4.00288016e-01
-1.50009537e+00 -1.79653645e-01 -2.80000120e-01 2.63401419e-01
4.74515080e-01 2.45756373e-01 -6.01630151e-01 2.16700375e-01
8.98917615e-01 1.80766746e-01 -8.48786473e-01 1.20965195e+00
-4.09815460e-01 3.01834643e-01 -8.87563527e-01 2.73470193e-01
-2.78270036e-01 3.28892797e-01 5.29768109e-01 1.25359631e+00
4.65588212e-01 2.01171830e-01 2.48034671e-01 5.86974382e-01
-2.14669839e-01 -4.86676767e-02 -4.59652752e-01 7.75831521e-01
6.77305400e-01 1.37900972e+00 -4.20141727e-01 -9.08092484e-02
-6.99700296e-01 7.95671999e-01 4.11134452e-01 2.92025626e-01
-1.17482054e+00 -1.68427885e-01 5.62147677e-01 4.23809022e-01
3.28066230e-01 2.08033428e-01 1.54982641e-01 -1.31227732e+00
2.67238021e-01 -8.14294100e-01 4.09298837e-01 -6.97169542e-01
-1.31906927e+00 7.30744660e-01 1.79914996e-01 -1.02135766e+00
-2.67073125e-01 -6.84215963e-01 -8.38551581e-01 6.53367519e-01
-1.44791961e+00 -1.85191619e+00 -6.10624254e-01 8.29768181e-01
5.36459923e-01 1.32272929e-01 5.09830415e-01 5.95803499e-01
-1.06879330e+00 1.48585331e+00 -3.97051871e-01 3.32977653e-01
1.05198526e+00 -1.11742246e+00 4.60608542e-01 7.48817503e-01
-4.14240658e-01 5.90733826e-01 7.98371851e-01 -4.82373238e-01
-1.31553221e+00 -1.18523490e+00 7.72256076e-01 -5.70761323e-01
2.12910548e-02 -8.10654283e-01 -9.79993045e-01 9.85925138e-01
-1.25161141e-01 4.51289088e-01 3.89320046e-01 2.24999651e-01
-3.05894881e-01 -1.97760627e-01 -1.25010717e+00 6.20156288e-01
1.37998235e+00 -2.66270012e-01 -4.13241535e-01 5.31462371e-01
6.06182277e-01 -6.20897174e-01 -5.45470834e-01 6.18208528e-01
6.86707079e-01 -1.02803934e+00 1.22461784e+00 -1.13964103e-01
-1.16847143e-01 -2.41725266e-01 -7.54073044e-05 -9.70206022e-01
-1.83747023e-01 -6.06583595e-01 -4.05967295e-01 1.46703303e+00
1.84621707e-01 -6.73841774e-01 9.43235874e-01 7.87731588e-01
-2.53091663e-01 -8.40717614e-01 -9.14680839e-01 -6.43012166e-01
9.65983644e-02 -1.18125580e-01 8.65325868e-01 5.02273738e-01
-4.23017144e-01 4.26787972e-01 -6.38966501e-01 5.84375262e-01
1.01171291e+00 -7.57327676e-02 1.11817789e+00 -1.13434124e+00
-2.53285229e-01 -2.70007133e-01 -2.28457913e-01 -1.33127367e+00
3.83186847e-01 -4.19999570e-01 3.83648008e-01 -1.14002192e+00
4.57076997e-01 -9.18542594e-02 6.69767633e-02 6.50942504e-01
-4.05050009e-01 2.43383080e-01 3.21013153e-01 1.40047550e-01
-6.66708708e-01 8.92323434e-01 1.15624022e+00 1.78433493e-01
-2.82257423e-03 2.91616410e-01 -6.40870094e-01 1.13413358e+00
4.47299510e-01 -3.21600109e-01 -1.48943201e-01 -4.72189009e-01
-2.76703715e-01 1.20317966e-01 6.37037337e-01 -1.07405138e+00
5.26832759e-01 2.53545612e-01 6.86258972e-01 -8.66061270e-01
7.29900599e-01 -3.95444661e-01 -1.93430543e-01 1.08518429e-01
2.25433260e-01 1.94070697e-01 1.11440718e-01 3.36309761e-01
9.04023796e-02 1.81549594e-01 9.28846002e-01 2.13568099e-02
-5.34045935e-01 7.82028139e-01 2.25187063e-01 1.48494527e-01
8.28195155e-01 -7.52212331e-02 -3.51293385e-01 -5.53657293e-01
-6.90101862e-01 6.41160786e-01 4.47370887e-01 4.25278276e-01
5.92276692e-01 -1.16778553e+00 -8.27880919e-01 3.49978209e-01
3.93559672e-02 3.87463242e-01 3.95983100e-01 1.00645804e+00
-1.71523690e-01 3.14043671e-01 4.00583968e-02 -6.19902968e-01
-1.54928100e+00 4.89048213e-01 5.14583826e-01 2.90195625e-02
-7.15353012e-01 9.92910922e-01 8.25381935e-01 -8.07073712e-01
6.41116977e-01 -4.37640119e-03 -2.79081821e-01 3.92016470e-02
7.71938860e-01 2.86412150e-01 4.79597673e-02 -1.20716691e+00
-6.07521415e-01 7.61400819e-01 -1.52247176e-01 1.26784906e-01
1.27003300e+00 -4.37173069e-01 2.02617437e-01 -1.86882421e-01
1.37185931e+00 3.36296260e-02 -1.54083681e+00 -1.21664450e-01
-3.77940357e-01 -3.33432853e-01 -4.13958803e-02 -5.30889869e-01
-1.40071857e+00 9.38698828e-01 9.08366323e-01 -5.12613654e-01
1.11869061e+00 2.10272953e-01 1.08346212e+00 3.73186976e-01
4.55662996e-01 -8.53465736e-01 -8.37011784e-02 4.85070229e-01
1.04570115e+00 -1.52347112e+00 7.08227009e-02 -6.68994248e-01
-5.44360220e-01 6.80908918e-01 1.18770945e+00 3.52464288e-01
5.94400346e-01 2.75190890e-01 -9.98440832e-02 -1.69864297e-01
-3.62641454e-01 -3.94258887e-01 2.69645989e-01 6.18665695e-01
4.99796361e-01 8.11783299e-02 7.87735283e-02 9.25892711e-01
-6.65042400e-01 -9.67832506e-02 1.03329822e-01 8.27896237e-01
-3.39039326e-01 -6.82661176e-01 -7.98865259e-01 1.98648155e-01
-5.57705402e-01 2.07238600e-01 -1.34723186e-01 1.04084945e+00
2.53269792e-01 7.36302137e-01 -2.35892870e-02 -2.27518454e-01
6.20580673e-01 -5.24793379e-02 5.23788691e-01 -4.22097743e-01
-1.85455039e-01 2.78548837e-01 -9.31765139e-02 -5.04837871e-01
-5.65805584e-02 -8.64230394e-01 -1.10181153e+00 -5.31775713e-01
-3.31034958e-01 -2.32687995e-01 2.57007629e-01 8.41827810e-01
2.58960515e-01 1.43738374e-01 5.40902257e-01 -1.51197565e+00
-7.64705658e-01 -1.18660283e+00 -5.43484449e-01 3.01587850e-01
5.53012013e-01 -1.09100330e+00 -4.33616668e-01 -2.82307893e-01] | [13.621349334716797, 0.3228411376476288] |
94888115-46ee-461b-a16f-ac17a5cf31aa | inverse-reinforcement-learning-from-diverse | 2207.14299 | null | https://arxiv.org/abs/2207.14299v2 | https://arxiv.org/pdf/2207.14299v2.pdf | Graph Inverse Reinforcement Learning from Diverse Videos | Research on Inverse Reinforcement Learning (IRL) from third-person videos has shown encouraging results on removing the need for manual reward design for robotic tasks. However, most prior works are still limited by training from a relatively restricted domain of videos. In this paper, we argue that the true potential of third-person IRL lies in increasing the diversity of videos for better scaling. To learn a reward function from diverse videos, we propose to perform graph abstraction on the videos followed by temporal matching in the graph space to measure the task progress. Our insight is that a task can be described by entity interactions that form a graph, and this graph abstraction can help remove irrelevant information such as textures, resulting in more robust reward functions. We evaluate our approach, GraphIRL, on cross-embodiment learning in X-MAGICAL and learning from human demonstrations for real-robot manipulation. We show significant improvements in robustness to diverse video demonstrations over previous approaches, and even achieve better results than manual reward design on a real robot pushing task. Videos are available at https://sateeshkumar21.github.io/GraphIRL . | ['Xiaolong Wang', 'Rishabh Jangir', 'Nicklas Hansen', 'Jonathan Zamora', 'Sateesh Kumar'] | 2022-07-28 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 5.08421659e-02 2.29932368e-01 -3.25881064e-01 -1.64930686e-01
-6.31617606e-01 -5.81820667e-01 4.01023656e-01 -5.21583915e-01
-5.08793712e-01 8.13333690e-01 4.51125234e-01 1.23583145e-01
-3.17566127e-01 -2.91548278e-02 -1.21693468e+00 -4.84970212e-01
-6.35012031e-01 2.66929686e-01 2.25629285e-01 -4.13540244e-01
2.70646572e-01 3.40633273e-01 -1.47707939e+00 2.78960288e-01
6.01880252e-01 4.94778067e-01 5.04541159e-01 9.04462218e-01
6.08967721e-01 1.25942302e+00 -4.56054926e-01 1.08315416e-01
5.31261146e-01 -5.44213474e-01 -6.90812290e-01 1.18719272e-01
4.49950248e-01 -7.91873872e-01 -8.34588587e-01 9.17094767e-01
4.46156144e-01 5.59551835e-01 5.01834035e-01 -1.82423317e+00
-7.35643387e-01 8.81619155e-01 -3.90887588e-01 -6.31155223e-02
7.04825282e-01 5.66832960e-01 9.96932328e-01 -4.88268107e-01
1.15288126e+00 1.59304917e+00 5.08773625e-01 1.14062548e+00
-1.31870806e+00 -6.67178094e-01 3.75115991e-01 5.31098247e-01
-9.40909088e-01 -4.05053079e-01 6.15991831e-01 -4.95321393e-01
9.93575573e-01 -6.47736946e-03 8.73478651e-01 1.52586401e+00
1.50320455e-01 1.25078511e+00 9.51011717e-01 -2.23006055e-01
1.34939075e-01 -2.18139768e-01 -2.09974661e-01 9.51232910e-01
1.83850769e-02 5.33438563e-01 -7.04403043e-01 1.40060723e-01
1.01662445e+00 9.26083885e-03 -4.76771533e-01 -1.20640159e+00
-1.44192255e+00 6.72044098e-01 6.33142054e-01 -8.49766061e-02
-3.06392610e-01 6.87489092e-01 4.32321191e-01 7.21702218e-01
-1.59297332e-01 7.86266744e-01 -3.58300149e-01 -2.77015269e-01
-2.59899050e-01 6.26194119e-01 8.44607234e-01 1.33028615e+00
4.73293424e-01 6.59579365e-03 -3.50411199e-02 5.38407803e-01
8.35326016e-02 4.23567206e-01 3.88260454e-01 -1.76354611e+00
4.50638503e-01 2.13834509e-01 2.12141633e-01 -6.64373457e-01
-4.50067073e-01 1.38648123e-01 -1.14870101e-01 8.38440478e-01
6.39795482e-01 -2.08397523e-01 -8.93381059e-01 1.98201978e+00
2.09099948e-01 -5.18398397e-02 1.59245238e-01 1.20069242e+00
6.18013322e-01 2.56972253e-01 2.30558999e-02 2.88860083e-01
9.15373921e-01 -1.30582440e+00 -5.61609566e-01 -1.52615532e-01
9.04256165e-01 -3.03753763e-01 1.14742815e+00 6.05315208e-01
-8.78753364e-01 -4.88315433e-01 -1.04913485e+00 -1.02160625e-01
-8.91849026e-02 2.69787870e-02 6.69116259e-01 -4.69990540e-03
-8.40039968e-01 1.07527566e+00 -9.52652156e-01 -5.36600411e-01
2.63840765e-01 5.89239180e-01 -7.49306321e-01 -3.04071486e-01
-9.47215080e-01 9.72319484e-01 2.93861628e-01 -1.32614180e-01
-1.36350036e+00 -4.08844709e-01 -1.06248021e+00 -1.59786507e-01
9.37619030e-01 -6.45514250e-01 1.33120084e+00 -8.49666059e-01
-1.57696486e+00 4.09572750e-01 4.82202679e-01 -4.24025446e-01
6.41698241e-01 -6.20620728e-01 1.38065532e-01 3.56149107e-01
8.67691711e-02 1.13976240e+00 1.10679400e+00 -1.30391407e+00
-5.52895427e-01 -2.40196601e-01 5.30817986e-01 5.64633548e-01
-6.81968033e-02 -3.02281350e-01 -3.25953603e-01 -5.36742985e-01
-2.88664579e-01 -1.50960028e+00 -1.94933087e-01 2.13580444e-01
-1.18301570e-01 -3.52805704e-01 1.01499033e+00 -6.89651191e-01
5.51520228e-01 -2.22769403e+00 8.54619861e-01 -2.04951540e-01
1.66989088e-01 -1.73420161e-01 -5.72407186e-01 4.93268788e-01
-4.26798649e-02 -9.40515772e-02 2.42073551e-01 -7.11696446e-02
1.96362883e-01 4.04350191e-01 -9.83995721e-02 7.02015221e-01
2.20322579e-01 9.23758686e-01 -1.34470820e+00 -4.07204092e-01
1.79093122e-01 2.93686777e-01 -9.71239984e-01 2.88376987e-01
-3.75893980e-01 7.74165511e-01 -5.71427882e-01 5.98993897e-01
5.54369763e-02 4.08764556e-03 1.39203891e-01 4.11473885e-02
1.22691512e-01 -8.89331475e-02 -1.06719172e+00 2.03643775e+00
-1.00777149e-01 6.62674665e-01 3.93732399e-01 -1.13990700e+00
5.93067527e-01 1.21843234e-01 7.39797950e-01 -5.72200418e-01
-1.19481573e-03 1.77607656e-01 3.33936036e-01 -8.43722999e-01
5.14237523e-01 2.70802826e-01 -2.08618850e-01 3.01230788e-01
3.27238232e-01 -6.31433845e-01 5.43817401e-01 3.50158483e-01
1.36293805e+00 1.04987025e+00 -3.29252444e-02 -5.70520051e-02
1.25447214e-01 3.30642581e-01 3.76459420e-01 9.95756209e-01
-5.97677886e-01 3.63468140e-01 5.70237994e-01 -1.87237456e-01
-1.17297149e+00 -9.95555282e-01 4.71461862e-01 1.17765105e+00
3.16898286e-01 -3.37637335e-01 -7.24476755e-01 -8.37871909e-01
2.39567503e-01 4.44280595e-01 -6.42253816e-01 -4.06858027e-01
-9.75237906e-01 4.40403968e-02 2.71933377e-01 5.25880694e-01
2.07744911e-01 -1.52079165e+00 -8.10935378e-01 -1.03606828e-01
-1.04418270e-01 -1.17312372e+00 -6.47521794e-01 1.46912053e-01
-8.36853623e-01 -1.22613657e+00 -7.66390562e-01 -7.56969273e-01
5.98181605e-01 3.74892414e-01 8.07548404e-01 2.29778383e-02
-5.26190400e-01 1.09117091e+00 -7.17247784e-01 -1.47751734e-01
-4.76042002e-01 -1.98817700e-01 3.24518234e-01 -5.75366437e-01
-1.36827067e-01 -4.13929701e-01 -6.07792020e-01 3.30246508e-01
-5.69180787e-01 -8.22132677e-02 6.30098522e-01 1.25208902e+00
1.63003296e-01 -3.25151145e-01 3.52998376e-01 -3.22092652e-01
6.94897056e-01 -2.13083029e-01 -5.48453450e-01 -2.53063757e-02
-2.26164579e-01 2.82616138e-01 5.68104625e-01 -8.84921789e-01
-8.16569686e-01 4.29406822e-01 3.47762406e-01 -8.32561374e-01
5.06301224e-02 1.89013168e-01 3.21106404e-01 -2.48424694e-01
7.89962590e-01 -1.12500124e-01 3.42776775e-01 -3.58667113e-02
6.39726698e-01 1.19071856e-01 4.21279848e-01 -8.58012378e-01
6.96704865e-01 3.50847691e-01 -7.34412018e-03 -6.75498903e-01
-3.40111554e-01 -3.65386903e-01 -6.30783200e-01 -5.31470358e-01
8.55607212e-01 -9.96070623e-01 -1.11339366e+00 9.65876281e-02
-7.61515856e-01 -1.10468709e+00 -4.23462182e-01 9.44144607e-01
-1.39879823e+00 3.90531033e-01 -7.29730070e-01 -7.01749623e-01
1.84865266e-01 -1.29850364e+00 1.12715769e+00 1.38271958e-01
-4.23200816e-01 -5.44993818e-01 -1.05193026e-01 2.05757156e-01
-4.07485701e-02 2.05352679e-01 5.61699986e-01 -3.44497144e-01
-9.70812678e-01 -4.61166874e-02 3.95010002e-02 3.47526193e-01
-1.44049942e-01 -2.93491602e-01 -4.83529419e-01 -6.44158244e-01
-2.54196316e-01 -9.28070307e-01 7.41823792e-01 3.74984801e-01
8.93419504e-01 -1.36304706e-01 -3.28755289e-01 4.52439755e-01
1.03638709e+00 1.48103639e-01 5.13529539e-01 5.17239928e-01
7.44500458e-01 7.08497286e-01 1.20121241e+00 4.11097705e-01
2.00168073e-01 7.77354240e-01 7.02287674e-01 1.53551087e-01
-4.63201106e-01 -4.95041281e-01 9.11057353e-01 2.57630318e-01
-2.73555696e-01 -8.53625014e-02 -4.11699295e-01 5.06522834e-01
-2.22844863e+00 -1.18304312e+00 2.97728717e-01 1.89546776e+00
4.73064899e-01 -9.49292034e-02 4.45414722e-01 -2.35598162e-01
5.75932503e-01 -1.32447466e-01 -9.17212903e-01 -1.38592377e-01
1.88186511e-01 -2.10661843e-01 6.85271978e-01 4.49341208e-01
-9.06460106e-01 1.02845454e+00 6.19064140e+00 5.03990650e-01
-8.71872783e-01 -1.76149607e-01 -1.35985821e-01 -4.12389100e-01
3.01135987e-01 3.39194573e-02 -3.85952175e-01 1.41062021e-01
2.62100428e-01 -1.20926134e-01 1.05135763e+00 1.12783384e+00
3.20816636e-01 -1.23863712e-01 -1.47695804e+00 8.40205312e-01
5.47209792e-02 -8.62724066e-01 -3.28121901e-01 2.92779040e-03
6.09141469e-01 2.42691368e-01 -4.77802800e-03 8.73700082e-01
7.78717816e-01 -7.47517943e-01 9.06438708e-01 4.88772422e-01
5.24486840e-01 -3.79337698e-01 1.55988246e-01 4.94035274e-01
-8.78357351e-01 -5.92066050e-01 -2.31105581e-01 -1.57036275e-01
1.24615664e-03 -4.11071151e-01 -9.08551574e-01 2.74805367e-01
9.59311545e-01 9.46258545e-01 -2.51135260e-01 7.19166577e-01
-3.25821251e-01 1.39859403e-02 -1.84086874e-01 -3.18478942e-01
1.64581522e-01 -1.46597475e-01 8.80609870e-01 8.52061272e-01
3.03265840e-01 8.39222595e-02 5.47567189e-01 7.82631397e-01
-1.02566918e-02 -1.43289983e-01 -1.02918601e+00 -2.80150592e-01
6.90503642e-02 1.04577756e+00 -4.58911091e-01 -2.91529149e-01
-2.39239290e-01 1.10932720e+00 6.66084230e-01 5.20309508e-01
-1.00125086e+00 -2.93586940e-01 6.54361069e-01 -1.31057337e-01
5.06283104e-01 -6.12164676e-01 3.97732913e-01 -1.23728430e+00
9.01373774e-02 -1.15858352e+00 3.04525077e-01 -1.08443630e+00
-9.64417040e-01 2.15475515e-01 2.82640517e-01 -1.48724484e+00
-5.22808969e-01 -9.57260311e-01 -1.34606555e-01 2.03885436e-01
-1.14807630e+00 -9.81153131e-01 -3.54383618e-01 7.30925441e-01
7.86837459e-01 -1.36985883e-01 5.68936586e-01 1.40559273e-02
-1.45706609e-01 2.32354358e-01 -3.98856681e-03 -1.87844336e-02
9.65010881e-01 -1.36814761e+00 1.48686722e-01 4.87261325e-01
-4.45920043e-02 5.26222825e-01 9.06559765e-01 -7.21485794e-01
-1.90657055e+00 -5.23064673e-01 -3.79391275e-02 -6.21190012e-01
9.08022821e-01 -4.56378073e-01 -5.11691988e-01 1.01517296e+00
1.79717809e-01 1.65003076e-01 -9.12313908e-02 -4.93428558e-02
-2.96616197e-01 2.78751433e-01 -9.99005020e-01 1.06162786e+00
1.60093963e+00 -3.62941504e-01 -6.10669076e-01 5.12216926e-01
7.78051496e-01 -6.34931326e-01 -8.65536034e-01 3.43193918e-01
8.65090847e-01 -7.28258908e-01 1.05616426e+00 -7.88022220e-01
6.48952484e-01 -3.86762410e-01 -4.08827839e-03 -1.71178126e+00
-4.45133954e-01 -7.75806189e-01 -2.74061263e-01 4.53561872e-01
1.54632434e-01 -3.67971987e-01 8.24825168e-01 4.84017074e-01
-1.92011297e-01 -3.48381281e-01 -6.10988259e-01 -1.25142026e+00
-1.42546445e-01 -1.02259003e-01 1.01782633e-02 7.23255873e-01
5.47122240e-01 3.04095864e-01 -8.37662935e-01 7.09346356e-03
6.09624386e-01 -1.35248974e-01 1.37381041e+00 -8.93306315e-01
-5.25755823e-01 -3.64560574e-01 -3.52822274e-01 -1.17809761e+00
3.89067173e-01 -8.50267708e-01 4.68029171e-01 -1.46008825e+00
2.04428092e-01 -2.33045936e-01 -4.22178954e-02 5.34491479e-01
5.74565008e-02 5.85512444e-02 6.38170242e-01 1.26761749e-01
-8.88981521e-01 7.06000328e-01 1.96410918e+00 -2.26700708e-01
-2.90201277e-01 -3.94311041e-01 -1.90530807e-01 5.34742534e-01
8.54860842e-01 -5.32005250e-01 -5.44819236e-01 -4.62205023e-01
-1.11342028e-01 2.58135885e-01 5.06654978e-01 -1.00553298e+00
1.00559264e-01 -1.68906122e-01 2.33173683e-01 -8.08256194e-02
6.58384264e-01 -8.73678565e-01 3.50629464e-02 8.52328360e-01
-5.72131097e-01 -5.17117083e-02 2.32497960e-01 8.86981905e-01
1.54305980e-01 -2.91420847e-01 3.82308424e-01 -4.01366770e-01
-1.06388485e+00 1.35514319e-01 -2.50870615e-01 -5.27340919e-02
1.06677139e+00 -1.37410909e-01 -4.11068439e-01 -7.71726608e-01
-1.13434100e+00 4.81564820e-01 6.68628871e-01 8.82953942e-01
6.12795889e-01 -1.23728263e+00 -5.16674697e-01 -3.21052712e-03
1.12993918e-01 -4.53137517e-01 2.81252742e-01 9.34639812e-01
-2.78198361e-01 7.88092688e-02 -7.08228707e-01 -6.51817918e-01
-1.45005167e+00 7.66826332e-01 1.85377985e-01 -4.95646149e-02
-1.03820181e+00 6.87420607e-01 3.90373588e-01 -5.33825338e-01
4.69940722e-01 -3.90677452e-01 -1.18193356e-02 -4.44904715e-01
1.29322007e-01 3.16321641e-01 -6.29613221e-01 -3.80547911e-01
-1.54861018e-01 4.97724682e-01 -5.48644140e-02 -2.51396656e-01
1.48025513e+00 -5.70992082e-02 3.67008597e-01 1.33266106e-01
1.01800323e+00 -1.73117146e-01 -1.90552390e+00 8.64541233e-02
-4.03514430e-02 -5.20218849e-01 -3.44547480e-01 -6.33398116e-01
-7.38547802e-01 4.48573291e-01 4.33271050e-01 8.68957564e-02
7.46142447e-01 1.12549372e-01 4.32171047e-01 9.62404430e-01
6.13910854e-01 -1.31305981e+00 7.56224394e-01 5.81258357e-01
1.41603923e+00 -1.37020051e+00 2.83639669e-01 -2.07338527e-01
-9.39813197e-01 1.17751014e+00 7.78551459e-01 -5.84264994e-01
1.61159858e-01 1.37216300e-01 -1.97751746e-01 -2.44808927e-01
-7.21306205e-01 -4.55086887e-01 4.61967960e-02 1.01673198e+00
9.85943004e-02 -1.08997850e-02 -3.36723924e-02 1.42701045e-01
-1.20721519e-01 1.74169481e-01 7.73738503e-01 1.16626096e+00
-3.78079474e-01 -9.83302057e-01 -2.89768785e-01 1.82734296e-01
-5.48508428e-02 3.70945483e-01 -4.13171828e-01 1.33821726e+00
-2.46486068e-01 7.36757100e-01 -3.61216187e-01 -4.22091365e-01
5.65648794e-01 -2.21155405e-01 1.18309891e+00 -5.70047557e-01
-1.06096506e-01 -5.89319132e-03 3.13006878e-01 -1.08477342e+00
-4.63760525e-01 -9.60466683e-01 -1.33291316e+00 -6.06917031e-02
-2.23969653e-01 -1.10398911e-01 6.40840590e-01 4.93299991e-01
2.21694306e-01 6.19164288e-01 2.95149297e-01 -1.47300863e+00
-7.81519055e-01 -7.93079793e-01 -4.58715558e-01 6.79256082e-01
6.24757707e-01 -1.16574550e+00 -5.16414940e-01 1.51842237e-01] | [4.575743198394775, 0.8225477933883667] |
63e4efe1-a8be-4c2e-9230-1a9b99e101f4 | xformal-a-benchmark-for-multilingual | 2104.04108 | null | https://arxiv.org/abs/2104.04108v1 | https://arxiv.org/pdf/2104.04108v1.pdf | XFORMAL: A Benchmark for Multilingual Formality Style Transfer | We take the first step towards multilingual style transfer by creating and releasing XFORMAL, a benchmark of multiple formal reformulations of informal text in Brazilian Portuguese, French, and Italian. Results on XFORMAL suggest that state-of-the-art style transfer approaches perform close to simple baselines, indicating that style transfer is even more challenging when moving multilingual. | ['Joel Tetreault', 'Ke Zhang', 'Di Lu', 'Eleftheria Briakou'] | 2021-04-08 | null | null | null | null | ['formality-style-transfer'] | ['natural-language-processing'] | [-1.50335729e-01 5.19240424e-02 -2.11466849e-01 -5.13697326e-01
-1.23000085e+00 -1.24418974e+00 9.60355282e-01 -3.83363515e-01
-8.20731163e-01 1.72740078e+00 5.13379395e-01 -6.21233940e-01
4.42091137e-01 -3.90193999e-01 -6.87795818e-01 -3.24971341e-02
2.97920287e-01 8.81958306e-01 -4.13628221e-02 -9.63507175e-01
6.83664950e-03 2.83021718e-01 -5.93389988e-01 7.26346612e-01
9.57454681e-01 2.99881041e-01 4.89603765e-02 7.14179695e-01
-3.47230524e-01 5.64439297e-01 -1.03498018e+00 -1.13513577e+00
6.80123642e-02 -3.96487772e-01 -1.42626107e+00 -5.30198872e-01
7.70033002e-01 -2.59732157e-01 -3.39210719e-01 9.63025093e-01
5.82823575e-01 -4.00774851e-02 6.81163192e-01 -7.08298147e-01
-1.25053525e+00 1.09732187e+00 -1.06395297e-01 4.52483714e-01
7.42023945e-01 1.62369996e-01 9.58253920e-01 -1.04582715e+00
1.14524281e+00 1.72584689e+00 7.35837281e-01 8.10173631e-01
-1.35002053e+00 -6.41734362e-01 2.65695423e-01 -8.47338438e-02
-1.13073957e+00 -4.66586024e-01 3.35105598e-01 -2.27720901e-01
1.25103998e+00 2.00160548e-01 3.58244658e-01 1.72669482e+00
3.63656074e-01 9.42544043e-01 1.42674077e+00 -3.76719296e-01
-4.75132912e-01 3.09175551e-01 -1.32910311e-01 2.60937870e-01
9.37127788e-03 4.41126935e-02 -3.48505795e-01 5.53305969e-02
5.98082542e-01 -7.97192872e-01 -3.30800325e-01 4.26058114e-01
-1.65710926e+00 5.60621500e-01 2.98298299e-01 7.24788308e-01
1.77878946e-01 -2.01925322e-01 7.93558180e-01 9.32108521e-01
5.66783249e-01 8.14950347e-01 -9.36459124e-01 -6.20447874e-01
-9.62959290e-01 3.25593680e-01 7.79922903e-01 1.43651068e+00
6.46943092e-01 9.83124692e-03 -2.17896223e-01 9.00612295e-01
-3.71734910e-02 7.42101312e-01 4.68686223e-01 -7.89818227e-01
1.23469806e+00 -1.05027944e-01 2.92090714e-01 -1.19496465e-01
-1.54953137e-01 -3.27138424e-01 -5.31524062e-01 2.14039031e-02
6.67178154e-01 -4.94483650e-01 -6.62949800e-01 1.61005127e+00
-3.23864043e-01 -6.55008495e-01 2.58367717e-01 3.04551035e-01
6.66278720e-01 8.04211557e-01 1.56836182e-01 7.88341090e-02
1.06428933e+00 -1.09811115e+00 -7.59790540e-01 -3.08052570e-01
7.48037934e-01 -1.26025414e+00 1.65048099e+00 3.59385937e-01
-1.45237994e+00 -8.93834293e-01 -1.07807076e+00 -4.94758099e-01
-6.10254705e-01 2.30234992e-02 2.66965806e-01 6.33688569e-01
-1.07843304e+00 7.08793879e-01 -4.71005201e-01 -3.81724656e-01
2.67062485e-01 -5.26710376e-02 -6.87365472e-01 -1.06329143e-01
-1.41426826e+00 1.13736689e+00 6.78452849e-01 -1.12419061e-01
-4.76171851e-01 -1.02259266e+00 -7.96464443e-01 -3.41262668e-01
-2.63974786e-01 -5.77877760e-01 1.63166392e+00 -1.07632411e+00
-1.60531771e+00 1.06477690e+00 -2.73556188e-02 -1.31799549e-01
9.39593017e-01 -7.96537876e-01 -9.62023735e-01 -3.29569429e-01
3.19105804e-01 8.57063770e-01 4.21379209e-01 -1.06434464e+00
-9.51556146e-01 -6.86637759e-02 3.95947158e-01 3.62766683e-01
-2.35043660e-01 4.46049929e-01 -3.40884253e-02 -1.08907235e+00
-9.35008883e-01 -8.24887395e-01 1.74575448e-01 -4.35129732e-01
-2.91333526e-01 -3.31386089e-01 7.43143678e-01 -1.17119956e+00
1.44177115e+00 -2.03303981e+00 4.61004555e-01 -2.36380965e-01
-2.81179070e-01 5.36328375e-01 -3.91234726e-01 6.46614790e-01
-1.30885905e-02 4.48521495e-01 -1.46432638e-01 -4.61823940e-01
1.21365532e-01 4.54650730e-01 -3.12432468e-01 2.30016410e-02
3.17733109e-01 1.09489381e+00 -1.29057121e+00 -3.58664602e-01
-1.47001296e-01 3.73054296e-01 -5.75342655e-01 -7.96402805e-03
-1.65125191e-01 1.16220200e+00 -4.76268232e-02 3.30452979e-01
3.87191772e-01 3.38040680e-01 1.71092600e-01 -7.72534981e-02
-3.07934523e-01 8.63562524e-01 -7.93666005e-01 2.33827710e+00
-6.71350241e-01 8.44385445e-01 -1.93499580e-01 -2.91346997e-01
6.63415790e-01 6.02237046e-01 -6.30269423e-02 -6.91368163e-01
-1.53810680e-01 6.81340396e-01 -5.23463599e-02 -1.27669871e-01
7.94559538e-01 -3.02390635e-01 -5.78685880e-01 3.74485433e-01
4.36851561e-01 -5.62358618e-01 6.63928390e-01 1.35348395e-01
5.60841858e-01 6.23212457e-01 2.71185696e-01 -8.12809706e-01
8.34329367e-01 -2.35801131e-01 3.90788138e-01 6.47672713e-01
-3.30673158e-02 5.96381783e-01 2.53066361e-01 -4.92368966e-01
-1.07303655e+00 -1.43207264e+00 -1.42277271e-01 1.26446056e+00
-5.66317379e-01 -5.82019746e-01 -9.82370853e-01 -1.51670456e+00
-2.05202147e-01 7.94759572e-01 -5.66638649e-01 2.83568501e-01
-1.11353254e+00 -3.78189594e-01 1.00545025e+00 6.17587864e-01
5.67168593e-01 -1.23596215e+00 1.67271614e-01 4.15946662e-01
-3.27156931e-01 -1.09425640e+00 -9.47789907e-01 -1.67591333e-01
-6.05078578e-01 -8.23959410e-01 -1.08737040e+00 -9.75510240e-01
2.14355752e-01 -4.50452030e-01 1.75257790e+00 -2.16910884e-01
4.63480562e-01 1.62738040e-01 -3.14684004e-01 -3.82072210e-01
-8.73736203e-01 8.25789988e-01 -8.34759995e-02 -6.26699388e-01
4.24695835e-02 -3.78382474e-01 -1.85734555e-01 -3.26023065e-03
-7.05004096e-01 -2.39219308e-01 4.24479097e-01 8.99126709e-01
1.66843846e-01 -6.60317898e-01 7.57877231e-01 -1.34051430e+00
8.69320929e-01 -2.08423004e-01 -2.42504120e-01 4.77484941e-01
-3.28271568e-01 2.32878372e-01 9.38602746e-01 -2.08899319e-01
-1.52509654e+00 -4.86048251e-01 -5.11908293e-01 3.19204420e-01
-2.67352581e-01 3.75485867e-01 -3.91167492e-01 1.46762103e-01
8.75867069e-01 -1.47578001e-01 -4.83382612e-01 -7.50707030e-01
9.16164815e-01 7.04541087e-01 7.69960821e-01 -1.07053828e+00
7.54583538e-01 -2.76966877e-02 -5.51915824e-01 -7.57603645e-01
-9.27441180e-01 2.13513318e-02 -1.09137464e+00 1.04545675e-01
6.40601158e-01 -9.39819336e-01 1.02244519e-01 4.00191098e-01
-1.39874709e+00 -6.18617415e-01 -5.43007612e-01 2.41590813e-01
-4.95594323e-01 3.07663023e-01 -1.04969764e+00 2.45702639e-01
-4.54921722e-01 -1.23877585e+00 1.11731386e+00 -9.41676870e-02
-6.73632622e-01 -1.50890076e+00 5.72176099e-01 1.06470756e-01
4.94663030e-01 -3.78266945e-02 8.98351729e-01 -3.24560732e-01
1.62063912e-01 2.76573002e-01 -2.00694412e-01 6.92724049e-01
4.60894108e-01 1.27068814e-02 -7.26173401e-01 -6.39378905e-01
-4.39218581e-01 -4.96311605e-01 4.33090210e-01 -2.18081012e-01
4.11195010e-01 -3.35925758e-01 -2.07086448e-02 8.04950833e-01
1.06996274e+00 1.25835329e-01 6.78215325e-01 5.09749055e-01
7.72694528e-01 2.72856712e-01 5.96534252e-01 -6.96562305e-02
6.59504771e-01 6.90851569e-01 -4.82092381e-01 -2.29846776e-01
-6.48013353e-01 -3.16763014e-01 7.06472218e-01 1.32341087e+00
-1.46027923e-01 -7.35546127e-02 -8.82518768e-01 5.14306605e-01
-1.42736113e+00 -7.67636240e-01 -7.27774873e-02 1.87649739e+00
1.52722406e+00 3.73532236e-01 2.83746570e-01 -1.65972695e-01
4.31682825e-01 -1.90979056e-02 -1.30498437e-02 -9.44546282e-01
-5.06186187e-01 7.53634393e-01 2.30889916e-01 9.52556431e-01
-1.05081356e+00 1.53720474e+00 7.85941315e+00 6.17257893e-01
-1.26225698e+00 2.49375880e-01 4.47817832e-01 -3.86073790e-03
-5.77683210e-01 -1.45129114e-01 -8.83949101e-01 4.95619893e-01
1.02972543e+00 -3.06138158e-01 7.28656292e-01 3.68629396e-01
-1.02285989e-01 5.34522414e-01 -1.41670430e+00 6.08574748e-01
-1.41589776e-01 -1.11892831e+00 4.58507895e-01 -2.23489657e-01
1.25081670e+00 4.65569377e-01 -4.65946198e-02 8.40252459e-01
6.52994573e-01 -8.89905572e-01 1.08098745e+00 3.72303277e-01
1.16494572e+00 -8.64061058e-01 5.81480086e-01 -1.65997714e-01
-1.11722922e+00 4.13846254e-01 -1.99123815e-01 1.49095282e-02
4.30323571e-01 2.12900698e-01 -5.10360122e-01 9.88136113e-01
6.95454657e-01 1.03797150e+00 -1.05895269e+00 4.92010176e-01
-6.05645001e-01 4.97673035e-01 3.15191820e-02 3.96255106e-01
4.69351679e-01 1.59667060e-02 5.43154538e-01 1.99780166e+00
3.06737423e-01 -7.36465156e-01 1.71817183e-01 3.92265975e-01
-3.32818806e-01 3.27447027e-01 -8.45507443e-01 -9.53962677e-04
3.34850371e-01 9.58658159e-01 -2.88873374e-01 -6.16503358e-01
-6.93537652e-01 1.30053747e+00 6.86094224e-01 5.90139568e-01
-7.82603264e-01 -5.70966363e-01 7.35277116e-01 -2.47529954e-01
1.18724771e-01 -3.80652070e-01 -2.73841351e-01 -1.46454728e+00
1.42862931e-01 -1.39311814e+00 4.31167245e-01 -5.86791396e-01
-1.49632466e+00 1.02568781e+00 -6.42833710e-02 -9.61301744e-01
-4.40471143e-01 -8.72055471e-01 -4.42844808e-01 1.15157664e+00
-1.49805474e+00 -1.36606503e+00 1.87173173e-01 7.51112759e-01
7.32803166e-01 -3.40701014e-01 1.10981417e+00 6.34513736e-01
-7.08847195e-02 1.05336714e+00 2.42159456e-01 2.56588876e-01
1.39750862e+00 -1.64194345e+00 1.13804591e+00 7.98268318e-01
2.91572660e-01 5.21147370e-01 4.76142794e-01 -6.43515587e-01
-5.99056244e-01 -9.68624115e-01 1.17779577e+00 -7.89859772e-01
9.97885823e-01 -3.78110528e-01 -6.93358064e-01 1.16477013e+00
1.04397273e+00 -2.67146140e-01 4.51811105e-01 1.98058411e-01
-3.82360190e-01 8.56026039e-02 -1.03783607e+00 7.64800191e-01
1.22077262e+00 -9.59533095e-01 -9.23695445e-01 1.25008658e-01
6.05449617e-01 -4.03088480e-01 -1.46269417e+00 3.85763139e-01
6.43164277e-01 -5.24537146e-01 8.08588922e-01 -9.37439144e-01
3.51390451e-01 -4.55209240e-02 -1.36784509e-01 -2.06612754e+00
-3.57848287e-01 -1.10952604e+00 2.95443714e-01 1.56026936e+00
8.93556118e-01 -6.54200912e-01 2.03410774e-01 -2.52321720e-01
-2.99392343e-01 -3.50123048e-02 -8.68454814e-01 -9.38607574e-01
1.06531405e+00 -1.05622418e-01 5.61352611e-01 1.28246498e+00
-7.78933316e-02 7.16776133e-01 -1.96896240e-01 -5.85600972e-01
1.05868168e-01 -5.37813604e-01 7.01804042e-01 -1.12176752e+00
-3.46303165e-01 -7.43288577e-01 9.93590429e-02 -8.72813761e-01
5.27210832e-01 -1.21694064e+00 -2.80884147e-01 -1.45785594e+00
-3.71385485e-01 -3.17555755e-01 -2.50637919e-01 3.66529733e-01
-4.49451238e-01 5.59676886e-01 4.03778732e-01 -7.10617052e-03
-4.44817007e-01 3.92006427e-01 1.60955763e+00 -3.92702729e-01
-6.07626066e-02 -2.28686765e-01 -7.14475870e-01 5.07061303e-01
9.45792317e-01 -2.66907990e-01 -2.13027745e-01 -1.18276501e+00
2.68221051e-01 -2.33013362e-01 -4.88132000e-01 -8.67917776e-01
-4.75602627e-01 1.56421840e-01 4.04013276e-01 -3.12787145e-01
-7.44299740e-02 -3.62317860e-01 -7.75046200e-02 2.63328493e-01
-3.59182656e-01 8.38349938e-01 6.05066359e-01 -3.44996989e-01
-3.42234433e-01 -9.48105901e-02 6.16937160e-01 -3.68444026e-01
-3.96247506e-01 9.25607607e-03 -4.91967738e-01 7.69972682e-01
4.30399805e-01 2.47823760e-01 -5.56946695e-01 -1.57068551e-01
-8.68256629e-01 -3.52045447e-02 5.46321273e-01 9.57681179e-01
-1.05901614e-01 -1.61373568e+00 -1.25877595e+00 2.03014299e-01
1.78649619e-01 -3.52531731e-01 -4.66086924e-01 4.09929633e-01
-7.72057533e-01 7.29586959e-01 -5.94597042e-01 -2.24278018e-01
-8.80284667e-01 2.32990310e-01 3.01012099e-01 -7.35298455e-01
-4.03669149e-01 6.16586447e-01 9.72239897e-02 -1.05573046e+00
-2.37347670e-02 -2.78589040e-01 -3.03702569e-03 -4.47030142e-02
6.01814449e-01 5.65623105e-01 2.45570511e-01 -6.20714009e-01
-2.39749283e-01 5.42271256e-01 -3.74897361e-01 -7.07734764e-01
9.18674767e-01 -1.46015137e-01 -1.75111696e-01 6.45223975e-01
1.07880187e+00 5.51602185e-01 -1.01896393e+00 -5.29378392e-02
3.51685166e-01 -4.98837195e-02 -4.38645542e-01 -1.31738698e+00
-6.89689934e-01 7.88139522e-01 2.98755229e-01 -1.80564702e-01
7.30284929e-01 -1.97183937e-01 8.18361640e-01 5.77086568e-01
4.68674839e-01 -1.23656094e+00 -3.47804517e-01 1.37232184e+00
1.22141862e+00 -9.64405358e-01 -3.67477238e-01 5.57012744e-02
-4.87212211e-01 1.16884768e+00 5.60389519e-01 -4.12954204e-02
6.16147339e-01 3.56871486e-01 5.07515609e-01 4.17269170e-01
-4.35238808e-01 1.44474208e-01 4.87457365e-01 7.00263262e-01
1.15228724e+00 2.59954542e-01 -6.16102159e-01 4.14885342e-01
-8.74946415e-01 -4.99978326e-02 2.77622730e-01 9.87812459e-01
2.11773962e-01 -1.90938818e+00 -2.31501684e-01 -4.94872220e-02
-9.22622085e-01 -4.95611846e-01 -5.71619093e-01 1.13365376e+00
-1.21899107e-02 7.90615797e-01 -1.91689983e-01 -7.17895254e-02
6.29543185e-01 2.77261645e-01 1.08261478e+00 -7.35274315e-01
-1.17060506e+00 1.56385023e-02 6.15315020e-01 -4.56253976e-01
-1.65390119e-01 -6.56746864e-01 -8.39579940e-01 -3.10237229e-01
2.74398386e-01 3.18173647e-01 4.21368062e-01 1.15921640e+00
6.60863668e-02 6.31295264e-01 3.42453659e-01 -1.04448283e+00
-2.60731369e-01 -1.38836443e+00 -1.64358467e-02 7.41234541e-01
2.51476139e-01 -1.70239106e-01 1.14896009e-03 1.78921938e-01] | [11.431241035461426, 10.004508972167969] |
dfcea4b9-16f2-4061-953d-65cf76c344e9 | document-level-relation-extraction-with-dual | null | null | https://aclanthology.org/2020.coling-main.143 | https://aclanthology.org/2020.coling-main.143.pdf | Document-level Relation Extraction with Dual-tier Heterogeneous Graph | Document-level relation extraction (RE) poses new challenges over its sentence-level counterpart since it requires an adequate comprehension of the whole document and the multi-hop reasoning ability across multiple sentences to reach the final result. In this paper, we propose a novel graph-based model with Dual-tier Heterogeneous Graph (DHG) for document-level RE. In particular, DHG is composed of a structure modeling layer followed by a relation reasoning layer. The major advantage is that it is capable of not only capturing both the sequential and structural information of documents but also mixing them together to benefit for multi-hop reasoning and final decision-making. Furthermore, we employ Graph Neural Networks (GNNs) based message propagation strategy to accumulate information on DHG. Experimental results demonstrate that the proposed method achieves state-of-the-art performance on two widely used datasets, and further analyses suggest that all the modules in our model are indispensable for document-level RE. | ['Li Guo', 'Wang Yubin', 'Hengzhu Tang', 'Tingwen Liu', 'Xiaobo Shu', 'Bowen Yu', 'Zhenyu Zhang'] | 2020-12-01 | null | null | null | coling-2020-8 | ['document-level-relation-extraction'] | ['natural-language-processing'] | [ 3.19984347e-01 4.08273160e-01 -3.03872108e-01 -8.02215487e-02
-7.22725809e-01 -2.94864625e-01 6.56949759e-01 8.51849318e-01
-1.66102529e-01 5.02710819e-01 1.84610471e-01 -6.25896037e-01
-3.18341911e-01 -1.43774426e+00 -4.67449874e-01 -1.55628413e-01
-2.27940947e-01 4.18465137e-01 4.36007440e-01 -4.07230765e-01
1.08195864e-01 4.97277617e-01 -9.54095721e-01 4.72047716e-01
1.14242399e+00 1.01754069e+00 1.89764857e-01 8.74026418e-01
-5.39768934e-01 1.64734674e+00 -5.57845891e-01 -7.58458853e-01
-3.85833740e-01 -3.14300448e-01 -1.21440935e+00 1.06635191e-01
-1.82254482e-02 -2.86826462e-01 -5.56577742e-01 1.02607203e+00
1.66344255e-01 6.83838427e-02 5.74815512e-01 -8.97394061e-01
-5.77225149e-01 1.32570386e+00 -7.25683033e-01 2.01202229e-01
5.66858649e-01 -3.69770706e-01 1.44425344e+00 -6.60787642e-01
6.92850471e-01 1.07883668e+00 3.89085829e-01 6.13028854e-02
-6.41225636e-01 -5.14056981e-01 5.96049845e-01 4.64516073e-01
-1.24204516e+00 -2.28538454e-01 9.57233787e-01 -1.61301643e-01
1.49508142e+00 3.04006934e-01 5.52649617e-01 6.42742157e-01
4.10361260e-01 8.63797367e-01 5.47453821e-01 -4.36492294e-01
-9.47605297e-02 -1.86686158e-01 5.92372894e-01 9.70320225e-01
5.14466465e-01 -8.33976865e-01 -4.30259258e-01 8.36752057e-02
3.45717877e-01 -5.42724803e-02 -1.91877767e-01 3.40442926e-01
-8.90177429e-01 6.19427919e-01 6.02018297e-01 5.42654514e-01
-5.69151700e-01 -9.91029516e-02 2.41593361e-01 1.15621984e-01
5.77995479e-01 1.58219948e-01 -2.50130922e-01 3.22660178e-01
-8.22536647e-01 -7.72283673e-02 1.07288575e+00 1.17025650e+00
5.37496030e-01 -2.88139731e-01 -6.42509282e-01 6.35602474e-01
3.73030037e-01 1.55467138e-01 1.76643029e-01 -4.81203586e-01
1.22058976e+00 1.20582831e+00 -3.91653895e-01 -1.48301268e+00
-6.05319500e-01 -8.72113347e-01 -1.44696963e+00 -5.70515990e-01
-1.03767030e-01 -1.21916577e-01 -6.74228966e-01 1.41945827e+00
2.60186851e-01 -6.51455149e-02 2.86181599e-01 3.76155466e-01
1.49738967e+00 8.11833978e-01 4.60441373e-02 -3.03036451e-01
1.49769843e+00 -1.24987698e+00 -9.19839203e-01 -4.69436258e-01
6.27425194e-01 -4.91681516e-01 4.30829048e-01 2.89960623e-01
-1.28901911e+00 -4.55928743e-01 -1.39696825e+00 -4.15215671e-01
-3.87654752e-01 7.59765953e-02 7.06787109e-01 1.18663840e-01
-1.03473079e+00 4.05726612e-01 -4.85386819e-01 -1.12708926e-01
3.45489740e-01 2.11519137e-01 -1.97563171e-01 -2.09077850e-01
-1.47788978e+00 7.52698302e-01 7.19551921e-01 5.73640108e-01
-4.02207524e-01 -2.74537653e-01 -9.74785686e-01 5.61636865e-01
7.78390706e-01 -8.28754902e-01 1.18294609e+00 1.34445503e-01
-1.23929846e+00 3.47954005e-01 -3.24210733e-01 -2.90962100e-01
3.21662098e-01 3.55853029e-02 -6.75068259e-01 4.14909393e-01
-1.07345283e-01 2.75112991e-03 4.53859568e-01 -1.13253188e+00
-8.32797885e-01 -4.46328074e-01 4.06834185e-01 2.63676733e-01
-4.50882792e-01 -5.79830706e-02 -9.49625254e-01 -4.61451799e-01
3.80737305e-01 -3.07497263e-01 -7.30006024e-02 -7.76368558e-01
-1.05557394e+00 -5.18697262e-01 4.58171517e-01 -1.09587467e+00
2.02692103e+00 -1.51243675e+00 2.28899270e-01 3.92995715e-01
9.85455930e-01 2.11964339e-01 -2.10459217e-01 1.05468893e+00
1.66323930e-01 2.17529163e-01 -6.28086478e-02 -2.42797494e-01
-5.73014580e-02 -1.01302221e-01 -7.32122958e-02 -6.46797046e-02
3.36631328e-01 1.40070307e+00 -9.68242586e-01 -8.22574675e-01
-1.80059761e-01 2.21876845e-01 -9.54345614e-02 1.43832684e-01
-3.75213861e-01 -1.48599287e-02 -5.84072948e-01 5.60343146e-01
8.24935496e-01 -8.03046644e-01 8.00753891e-01 -2.96576172e-01
4.06156540e-01 4.44678754e-01 -1.08176434e+00 1.24411380e+00
-3.53524506e-01 2.81499207e-01 -4.56412584e-02 -8.85154426e-01
9.61138964e-01 9.66815725e-02 2.36679390e-01 -8.37968946e-01
1.58868477e-01 -1.48541868e-01 -3.08250170e-02 -5.04233062e-01
7.07674325e-01 1.69702932e-01 -1.28540590e-01 5.91339290e-01
-4.15157564e-02 5.31974658e-02 6.93922460e-01 9.30380940e-01
1.44974339e+00 -4.66163516e-01 3.98690939e-01 2.51666516e-01
9.82111335e-01 -8.26843828e-02 4.48951691e-01 8.29367518e-01
4.41565782e-01 1.33427784e-01 8.80031765e-01 7.28959888e-02
-5.77311933e-01 -6.50368989e-01 5.59147477e-01 5.15735745e-01
2.35522926e-01 -8.65375757e-01 -6.73914850e-01 -9.45730746e-01
-1.79166451e-01 8.83043051e-01 -2.47148439e-01 -1.55045584e-01
-5.96248746e-01 -7.05447793e-01 4.76281554e-01 6.25179172e-01
7.86995530e-01 -9.28162873e-01 3.22850019e-01 4.60203111e-01
-7.03562856e-01 -1.65093136e+00 -2.99539030e-01 -4.52753678e-02
-8.70436132e-01 -1.31380224e+00 -2.77220458e-01 -8.60710979e-01
7.99836636e-01 5.90195656e-01 1.22638226e+00 4.84344959e-01
2.02809781e-01 8.74595717e-02 -6.76952720e-01 -4.81993631e-02
-6.00797117e-01 5.41518569e-01 -6.10476315e-01 -8.53314176e-02
2.69452006e-01 -5.11589885e-01 -3.31370980e-01 -8.36700052e-02
-8.99519503e-01 4.44210470e-01 8.55008304e-01 4.98959482e-01
3.82876545e-01 8.42742920e-01 5.24742424e-01 -1.28405261e+00
1.27993822e+00 -4.83928651e-01 -3.13154727e-01 8.52808475e-01
-9.57880855e-01 -6.75799623e-02 7.54075766e-01 -1.23646138e-02
-1.24710488e+00 -6.30631924e-01 -1.35792181e-01 4.94092315e-01
2.92196780e-01 1.21825624e+00 -1.85572967e-01 3.03649753e-01
5.04088514e-02 3.33584249e-01 -2.80811280e-01 -3.65439385e-01
3.71025294e-01 8.69896472e-01 3.89571190e-01 -4.41163361e-01
8.20343018e-01 8.42517167e-02 1.67811394e-01 -5.49010396e-01
-1.08890438e+00 -2.22335249e-01 -6.42356873e-01 -2.07985401e-01
7.61702955e-01 -8.40721726e-01 -1.01812613e+00 6.12914860e-01
-1.44046175e+00 -9.84551609e-02 1.14415295e-01 1.50416180e-01
4.84827869e-02 5.47761261e-01 -1.20169401e+00 -8.22674990e-01
-8.71697009e-01 -8.62435937e-01 7.76871383e-01 2.54646450e-01
1.87568702e-02 -1.02084196e+00 -3.62922788e-01 7.60647178e-01
-4.56774235e-03 1.90215614e-02 1.37336063e+00 -6.37112617e-01
-8.45602930e-01 -4.21953499e-01 -7.94673443e-01 1.93053424e-01
1.95405856e-01 -5.07570654e-02 -6.65720642e-01 -1.00609481e-01
-1.92293882e-01 2.47507375e-02 8.59530091e-01 -2.66720541e-03
1.03779674e+00 -4.33163106e-01 -3.22621971e-01 1.87408686e-01
1.32335150e+00 1.74482539e-01 5.96552789e-01 1.91534087e-01
1.05230260e+00 6.47981763e-01 4.35656995e-01 1.74147338e-01
1.20124221e+00 3.04936856e-01 2.50356406e-01 -2.08755925e-01
-3.43034148e-01 -4.63029176e-01 8.40828046e-02 1.60065770e+00
2.79537160e-02 -9.38794374e-01 -8.72457683e-01 2.23764375e-01
-2.05497265e+00 -8.36336911e-01 -5.94058394e-01 1.52582836e+00
8.45530868e-01 4.93675023e-01 -2.38613695e-01 4.77261156e-01
6.33412540e-01 3.63282323e-01 -2.41291150e-01 -4.27846283e-01
-3.93107012e-02 -3.60531397e-02 2.00476274e-01 6.24666035e-01
-8.31726551e-01 8.97224247e-01 5.80443335e+00 1.09531069e+00
-5.40951729e-01 -1.77574337e-01 5.31451821e-01 4.78727192e-01
-5.47962546e-01 -2.40335194e-03 -9.30499673e-01 -5.24717430e-03
7.30506718e-01 -3.47878188e-01 3.79970163e-01 2.15728655e-01
-6.35912642e-02 -1.32457510e-01 -8.65851283e-01 5.52276552e-01
1.15009636e-01 -1.53840661e+00 3.48044008e-01 -1.61802456e-01
3.88616681e-01 -2.99254119e-01 -6.21596396e-01 2.91533053e-01
4.40103233e-01 -8.81734610e-01 4.44090933e-01 5.70123494e-01
6.09202981e-01 -8.79977524e-01 9.83867288e-01 6.24620378e-01
-1.68448079e+00 -1.95279140e-02 -5.47562242e-02 -6.08506575e-02
8.69067311e-02 9.35988426e-01 -9.16604578e-01 1.55029404e+00
1.83354110e-01 8.24241936e-01 -7.76111841e-01 4.10358012e-01
-6.90690279e-01 3.71734113e-01 -2.85566654e-02 -3.82849216e-01
2.20510259e-01 -1.04382329e-01 2.43399352e-01 1.46454489e+00
1.13119431e-01 2.43792623e-01 2.50345379e-01 5.55018604e-01
-3.82698745e-01 9.92318913e-02 -2.68602312e-01 -3.34274620e-01
4.73240197e-01 1.48692703e+00 -8.59399498e-01 -5.31273305e-01
-5.46882153e-01 7.88025677e-01 9.50047851e-01 5.21414816e-01
-4.81344104e-01 -8.38485956e-01 -4.18623723e-02 -1.53066799e-01
3.77492011e-01 -4.40286279e-01 -2.17598498e-01 -1.24012768e+00
3.12156558e-01 -7.09788084e-01 7.16902196e-01 -6.98690057e-01
-1.18386686e+00 8.95376503e-01 -1.12371199e-01 -9.25691426e-01
-1.89998135e-01 -2.92282462e-01 -6.17136717e-01 9.03519571e-01
-1.89681077e+00 -1.41535842e+00 -4.46676135e-01 4.14178729e-01
3.04413050e-01 7.03373849e-02 5.63858330e-01 3.99456412e-01
-9.12069082e-01 5.46675980e-01 -3.63497943e-01 5.48545301e-01
2.07279231e-02 -1.23738682e+00 5.71839631e-01 1.16411150e+00
2.64589190e-01 6.65333807e-01 1.83890074e-01 -1.04358160e+00
-1.78445292e+00 -1.07237780e+00 1.54022348e+00 -2.65322365e-02
7.85890698e-01 -3.08399975e-01 -8.49330902e-01 6.74613833e-01
1.26767680e-01 -5.03596127e-01 3.86200100e-01 3.26617777e-01
-3.36202353e-01 -3.46277028e-01 -8.09918344e-01 5.53108275e-01
1.10700130e+00 -5.48645198e-01 -5.00401855e-01 3.15866977e-01
1.06622183e+00 -4.87296462e-01 -1.10859716e+00 5.39694726e-01
2.58843273e-01 -7.55971491e-01 8.33349228e-01 -4.51281369e-01
8.00095320e-01 -2.49970824e-01 3.15155759e-02 -9.72809255e-01
-5.92604816e-01 -4.94694650e-01 -8.70854259e-01 1.60417533e+00
7.34735966e-01 -6.18593276e-01 5.09300113e-01 4.05831754e-01
9.46168825e-02 -1.04483700e+00 -4.47952181e-01 -5.48491061e-01
-3.60714257e-01 -6.88812673e-01 8.21822703e-01 6.65541232e-01
3.01401317e-01 9.29681361e-01 -2.07469523e-01 3.90697271e-01
6.50824249e-01 4.81802970e-01 5.27023792e-01 -1.18453383e+00
-1.80039957e-01 -4.59929347e-01 5.96079007e-02 -1.17074180e+00
1.61523968e-01 -1.28414571e+00 -2.11005211e-01 -2.68447804e+00
4.83330905e-01 -1.73986971e-01 -4.06105518e-01 3.80003959e-01
-3.67677003e-01 -2.74582863e-01 1.09220833e-01 1.05759455e-02
-8.83588433e-01 5.04522920e-01 1.68716228e+00 -5.02932370e-01
-1.25481188e-01 -2.72230129e-03 -1.17832541e+00 5.18251896e-01
5.25456011e-01 -4.06485379e-01 -6.88058496e-01 -5.52042186e-01
5.62246501e-01 4.17045504e-01 3.72582488e-02 -6.27626896e-01
7.75780380e-01 -1.78172022e-01 2.93173313e-01 -9.64872301e-01
-4.10116948e-02 -6.26638055e-01 -1.74034625e-01 3.57671231e-01
-3.41177672e-01 -6.03348576e-02 -7.08820745e-02 6.26156032e-01
-4.51129049e-01 -1.92307144e-01 1.08862191e-01 -2.29563080e-02
-4.45758492e-01 4.26197082e-01 -2.47460738e-01 1.50094107e-01
6.60322607e-01 1.12842165e-01 -7.88934112e-01 -5.73344946e-01
-3.66821855e-01 6.01851285e-01 -2.83543915e-01 4.09627497e-01
7.44283557e-01 -1.04177094e+00 -8.22698236e-01 -2.35369671e-02
1.21520869e-01 2.87563890e-01 3.27686489e-01 6.52948618e-01
-4.30394948e-01 5.50096035e-01 5.74519396e-01 -5.19741923e-02
-1.23974478e+00 5.11003494e-01 -9.98206586e-02 -1.17805851e+00
-7.77527153e-01 7.43892372e-01 -2.80315667e-01 -3.39154243e-01
2.69725651e-01 -4.67273533e-01 -6.63333416e-01 1.85271174e-01
5.72980940e-01 4.29052234e-01 2.57697701e-01 -3.42440188e-01
-4.14056301e-01 4.15375471e-01 -3.11324656e-01 1.02437466e-01
1.32731700e+00 -2.91185647e-01 -7.62214959e-01 1.43420607e-01
9.05506670e-01 9.68165249e-02 -4.94639963e-01 -5.75492978e-01
1.84401006e-01 1.33152753e-01 2.26098180e-01 -9.36469138e-01
-9.73215103e-01 6.97461724e-01 -5.62044561e-01 5.96132815e-01
1.26909399e+00 3.93311158e-02 1.23855448e+00 6.38396502e-01
2.63628602e-01 -1.00878310e+00 -1.85728241e-02 7.14078307e-01
8.34300935e-01 -9.53555405e-01 4.59780037e-01 -9.94184911e-01
-4.16052759e-01 1.26812100e+00 5.53640187e-01 3.27250332e-01
5.85674524e-01 3.49283397e-01 -1.78139195e-01 -4.67452198e-01
-8.55548382e-01 -2.46448338e-01 6.58678770e-01 2.39718303e-01
3.03376734e-01 7.38004828e-03 -3.81307095e-01 9.62981224e-01
-5.88138513e-02 4.52480949e-02 3.71561795e-01 1.01142097e+00
-3.21275234e-01 -1.26546061e+00 1.66891202e-01 6.27144277e-01
-3.19324642e-01 -3.00178438e-01 -5.63075960e-01 5.65843642e-01
-3.08555961e-01 1.54088950e+00 -3.08749288e-01 -8.98702443e-01
3.49337488e-01 -1.46455750e-01 4.94583368e-01 -6.34290695e-01
-7.26039827e-01 -1.20728821e-01 7.44618297e-01 -3.24551642e-01
-2.70222217e-01 -1.02589633e-02 -1.55661356e+00 -5.55190742e-01
-3.66871387e-01 2.76672274e-01 4.81379807e-01 1.19912899e+00
4.85653341e-01 1.20602965e+00 5.57835400e-01 -2.97772307e-02
-2.15310916e-01 -1.16127992e+00 -6.08646750e-01 8.05661678e-02
2.09751621e-01 -6.70959875e-02 -8.26530680e-02 -2.82941788e-01] | [9.262062072753906, 8.564370155334473] |
b320f6e7-8e07-4f04-8cab-1041699e55fe | learning-word-representations-from-scarce-and | null | null | https://aclanthology.org/P15-1104 | https://aclanthology.org/P15-1104.pdf | Learning Word Representations from Scarce and Noisy Data with Embedding Subspaces | null | ["M{\\'a}rio Silva", 'Ramon Astudillo', 'Wang Ling', 'Silvio Amir', 'Isabel Trancoso'] | 2015-07-01 | learning-word-representations-from-scarce-and-1 | https://aclanthology.org/P15-1104 | https://aclanthology.org/P15-1104.pdf | ijcnlp-2015-7 | ['twitter-sentiment-analysis'] | ['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] | [-7.468672275543213, 3.5347490310668945] |
d007ad6e-d656-4b66-8849-701bd7a884e1 | cellular-network-speech-enhancement-removing | 2301.09027 | null | https://arxiv.org/abs/2301.09027v1 | https://arxiv.org/pdf/2301.09027v1.pdf | Cellular Network Speech Enhancement: Removing Background and Transmission Noise | The primary objective of speech enhancement is to reduce background noise while preserving the target's speech. A common dilemma occurs when a speaker is confined to a noisy environment and receives a call with high background and transmission noise. To address this problem, the Deep Noise Suppression (DNS) Challenge focuses on removing the background noise with the next-generation deep learning models to enhance the target's speech; however, researchers fail to consider Voice Over IP (VoIP) applications their transmission noise. Focusing on Google Meet and its cellular application, our work achieves state-of-the-art performance on the Google Meet To Phone Track of the VoIP DNS Challenge. This paper demonstrates how to beat industrial performance and achieve 1.92 PESQ and 0.88 STOI, as well as superior acoustic fidelity, perceptual quality, and intelligibility in various metrics. | ['Ojas Bhargave', 'Joseph Konan', 'Shikhar Agnihotri', 'Haohui Liu', 'Hamza Khalid', 'Amanda Shu'] | 2023-01-22 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [ 5.65041900e-02 -3.20795864e-01 3.25404443e-02 1.25020832e-01
-9.57132339e-01 -3.97033393e-01 5.48192747e-02 -3.32448781e-01
-2.58693039e-01 5.73228538e-01 5.58193743e-01 -7.34501541e-01
-4.72398624e-02 -4.62786406e-01 -1.14476442e-01 -8.60138655e-01
6.60345554e-02 -1.79508060e-01 -4.30431627e-02 -4.83124673e-01
-2.23049849e-01 5.08180618e-01 -1.41410160e+00 3.20708930e-01
5.80691934e-01 1.00758982e+00 2.56737411e-01 1.18049359e+00
-5.27990535e-02 7.11004078e-01 -1.13182080e+00 -1.93443149e-01
2.26912156e-01 -3.37413162e-01 -3.79798800e-01 -1.08956099e-01
3.57569546e-01 -5.31185567e-01 -8.89126778e-01 1.13596547e+00
1.51681983e+00 1.29449978e-01 6.40082136e-02 -1.18522131e+00
-3.33408237e-01 6.13005042e-01 -3.25904369e-01 7.09106803e-01
-1.21772863e-01 4.22131985e-01 8.97879958e-01 -7.44309425e-01
-1.91341683e-01 1.19356775e+00 8.85311365e-01 7.21866429e-01
-1.21586084e+00 -9.28641498e-01 -3.47004086e-01 1.83673844e-01
-1.26485050e+00 -1.33131504e+00 6.14927530e-01 7.19600171e-02
1.05925262e+00 3.38923395e-01 2.69678354e-01 1.05960882e+00
-1.60681158e-01 6.17416441e-01 3.29633147e-01 -3.03048700e-01
2.70073503e-01 2.68079638e-02 -4.74989824e-02 -6.33779243e-02
-2.29334295e-01 5.00280440e-01 -6.67193174e-01 -2.91727409e-02
2.69985706e-01 -6.54080093e-01 -5.48981309e-01 5.78624249e-01
-7.00627923e-01 4.18088108e-01 -4.57677692e-02 3.75742674e-01
-5.15835166e-01 2.84239173e-01 3.90028119e-01 5.07677674e-01
3.51759493e-01 2.06723481e-01 -4.35010761e-01 -7.51861751e-01
-1.13064206e+00 5.14201447e-02 9.26712275e-01 9.78903592e-01
2.18881950e-01 9.20181692e-01 -2.37303302e-01 1.27866375e+00
3.35121185e-01 7.16592371e-01 1.95378929e-01 -1.32599604e+00
6.19604588e-01 -5.97761989e-01 1.74934380e-02 -7.15004802e-01
-1.66532248e-01 -1.19820225e+00 -1.06021130e+00 2.56094694e-01
3.47952321e-02 -6.69949114e-01 -7.71332204e-01 1.73339379e+00
-3.73783149e-02 3.54169875e-01 3.40011895e-01 5.83262205e-01
9.12877679e-01 9.83947694e-01 -1.17772475e-01 -4.41915870e-01
8.82242560e-01 -1.06397200e+00 -1.22770786e+00 -2.93378770e-01
1.54977739e-01 -9.60469127e-01 9.28158164e-01 6.95591688e-01
-1.32034612e+00 -7.73877144e-01 -1.06047201e+00 3.15923959e-01
2.48748288e-01 -3.25660586e-01 1.07502937e-02 1.59350693e+00
-1.50979769e+00 3.45981032e-01 -3.37161720e-01 -3.11276149e-02
3.40743303e-01 4.83541310e-01 2.56178290e-01 1.65518939e-01
-1.21827376e+00 3.64182562e-01 -3.71923715e-01 -1.70934215e-01
-9.77279723e-01 -1.13082266e+00 -4.50032949e-01 6.35155916e-01
1.68372512e-01 -4.53274459e-01 1.70573628e+00 -5.81571102e-01
-1.73920214e+00 4.01046664e-01 -2.09831387e-01 -5.50837278e-01
3.49224329e-01 -2.25771055e-01 -1.23573208e+00 7.70174786e-02
-1.37891397e-01 3.69200408e-01 9.21123743e-01 -1.33579183e+00
-9.29964483e-01 -1.06788658e-01 -3.22912723e-01 9.54069942e-02
-5.32121420e-01 2.84652352e-01 -5.41213453e-01 -6.60730004e-01
5.44001535e-02 -5.16501963e-01 6.71202540e-02 -5.30312538e-01
-3.26555073e-01 1.55261710e-01 1.41573942e+00 -1.01541853e+00
1.43436575e+00 -2.56713343e+00 -5.63816607e-01 3.38567048e-02
1.49809316e-01 8.82373810e-01 -3.00902128e-01 9.20040384e-02
4.86059561e-02 6.24111414e-01 4.90055680e-02 -5.08256376e-01
2.87949275e-02 9.49115958e-03 -4.11519676e-01 1.12108566e-01
-2.07849085e-01 3.31027597e-01 -5.75991988e-01 -1.13564301e-02
2.36767352e-01 7.64264226e-01 -7.53968179e-01 2.00396225e-01
3.97470385e-01 2.99736500e-01 1.59684703e-01 5.82288206e-01
9.86392558e-01 3.40609670e-01 -1.11373566e-01 -2.44066492e-01
-1.60911664e-01 7.12285280e-01 -1.19862199e+00 1.10400069e+00
-6.61618888e-01 1.06019640e+00 1.07810068e+00 -5.49120784e-01
9.73183632e-01 7.58413196e-01 5.46326458e-01 -8.88121665e-01
-6.25028647e-03 5.42384267e-01 2.26797849e-01 -4.89749044e-01
4.16519076e-01 -3.01747620e-01 4.85194504e-01 2.56892247e-03
-5.61738051e-02 -2.63843030e-01 -2.82420009e-01 -2.88519170e-02
1.34180260e+00 -7.46469676e-01 -2.06943974e-01 -2.81928450e-01
3.94440770e-01 -7.51359522e-01 7.51425922e-01 9.27341580e-01
-8.66828799e-01 4.77427930e-01 2.17213586e-01 3.10703665e-01
-1.19332898e+00 -1.34245682e+00 1.32619599e-02 1.01437962e+00
-1.19795009e-01 -3.63617651e-02 -8.92580152e-01 3.34783420e-02
-2.99031377e-01 9.29666638e-01 3.02899778e-01 -1.12668298e-01
-6.20652556e-01 -4.33307171e-01 1.22204971e+00 2.70088673e-01
8.77404034e-01 -7.98478305e-01 1.13341659e-01 4.98019040e-01
-5.04051447e-01 -1.39152265e+00 -7.80597866e-01 2.72226483e-01
-4.18305278e-01 -1.60611719e-01 -7.67116785e-01 -8.99580777e-01
-3.71417314e-01 4.75098670e-01 1.00090921e+00 -1.60623580e-01
7.02893138e-02 3.50944102e-01 -9.97833312e-02 -4.55568194e-01
-9.02506351e-01 1.12529598e-01 3.75692725e-01 -7.98822269e-02
2.98062414e-01 -9.65092421e-01 -4.74891633e-01 3.79494637e-01
-6.96529925e-01 -7.47466981e-01 2.16264471e-01 6.84670985e-01
5.76500371e-02 8.41913939e-01 1.18088317e+00 1.98676437e-01
9.42240059e-01 -3.46610725e-01 -3.32430661e-01 -3.89576793e-01
-3.51925224e-01 -7.03900874e-01 4.85887319e-01 -3.62582564e-01
-1.06362617e+00 -3.38646412e-01 -8.81533444e-01 -2.37770498e-01
-1.82272509e-01 1.30413368e-01 -7.16730475e-01 -1.44430494e-03
6.08867645e-01 2.26737872e-01 -2.59939991e-02 -5.76272547e-01
-1.78527489e-01 1.58287036e+00 7.39975214e-01 -1.16158761e-01
6.23962402e-01 2.71491379e-01 -3.37602705e-01 -1.54432738e+00
-3.05390060e-01 -5.98484039e-01 4.65737730e-02 -4.83338945e-02
4.93767411e-01 -1.04482496e+00 -8.23288321e-01 6.05876088e-01
-1.21148312e+00 -4.60339248e-01 -3.31131011e-01 4.38501060e-01
-4.47793871e-01 5.45642614e-01 -8.87840331e-01 -1.23881125e+00
-6.81032896e-01 -1.16610539e+00 6.55699432e-01 2.25143701e-01
-8.12515244e-03 -4.91215229e-01 -1.28570765e-01 4.62619722e-01
1.12258792e+00 -5.82096994e-01 6.73612475e-01 -4.17146742e-01
-2.33683586e-01 1.99608758e-01 -1.31843194e-01 1.00451398e+00
4.46433991e-01 -2.59254187e-01 -1.56348264e+00 -3.48163277e-01
4.79505926e-01 2.39359677e-01 7.16204345e-01 7.63426661e-01
9.60481703e-01 -1.79239035e-01 2.52610054e-02 6.79945230e-01
9.89751875e-01 8.19153547e-01 1.10747814e+00 6.65232092e-02
1.83084980e-01 3.99081171e-01 2.17911042e-02 5.42017400e-01
-2.94397622e-01 6.53516293e-01 5.19777179e-01 -1.59164757e-01
-7.18509138e-01 5.12379482e-02 4.57023501e-01 9.61736679e-01
4.89946336e-01 -8.43074858e-01 -9.36199725e-01 8.77448618e-01
-1.03910637e+00 -1.10224521e+00 -1.54990390e-01 1.96074116e+00
7.69144535e-01 1.79984212e-01 6.83635995e-02 6.43867552e-01
1.00969076e+00 2.20400006e-01 -3.63316447e-01 -4.43431646e-01
-4.65119660e-01 2.84562171e-01 5.11491537e-01 9.07529712e-01
-9.73532319e-01 7.50562429e-01 6.87053823e+00 1.16475105e+00
-1.13842440e+00 2.80848771e-01 8.20949733e-01 -1.84528962e-01
-1.33438170e-01 -7.47921407e-01 -7.20558941e-01 4.33908969e-01
1.53276908e+00 -7.43660480e-02 6.91902578e-01 6.40302300e-01
8.84064257e-01 3.92036051e-01 -6.70971990e-01 1.23616624e+00
-2.62233585e-01 -1.23868668e+00 -2.55889833e-01 3.56082708e-01
5.07587433e-01 3.25762868e-01 4.86265391e-01 5.97910464e-01
-3.86678204e-02 -1.18897355e+00 4.82263297e-01 1.04961127e-01
9.05472696e-01 -1.05494511e+00 7.13586390e-01 2.09170386e-01
-1.02415454e+00 -4.29913074e-01 -1.05270967e-01 1.56591043e-01
4.29952294e-01 7.69894361e-01 -8.82054746e-01 -3.11864857e-02
9.19989765e-01 -4.20190580e-02 3.47738892e-01 1.32042849e+00
7.08037615e-02 1.19522321e+00 -3.59699309e-01 2.31613204e-01
2.23467603e-01 3.46592188e-01 1.09607875e+00 1.57051492e+00
5.70529938e-01 9.34651792e-02 -1.59967363e-01 6.15167201e-01
-5.04777312e-01 -1.93471298e-01 -4.37332332e-01 9.62858126e-02
8.50846827e-01 7.22563744e-01 1.32607535e-01 -1.93600908e-01
-2.20568925e-01 7.82844961e-01 -7.58082688e-01 6.96844697e-01
-8.68101001e-01 -7.43370652e-01 1.50015104e+00 1.35752425e-01
3.58460903e-01 -3.65037531e-01 -6.28568351e-01 -4.24264461e-01
-7.11935535e-02 -1.36521924e+00 -3.00345361e-01 -6.18305922e-01
-1.01814044e+00 5.51170766e-01 -7.91338980e-01 -9.06321764e-01
2.24943459e-01 -3.49846810e-01 -6.64453328e-01 1.14900899e+00
-1.59803677e+00 -4.01292801e-01 -2.18888283e-01 5.40401161e-01
8.63779724e-01 -5.62585890e-01 5.73882818e-01 1.22149825e+00
-3.66997153e-01 9.47158456e-01 4.99010056e-01 7.43831992e-02
4.61802691e-01 -8.96499634e-01 7.28152573e-01 8.83731425e-01
-2.79084712e-01 1.93452626e-01 9.52900290e-01 -4.45020437e-01
-1.01496935e+00 -1.00280094e+00 9.68280077e-01 2.29728460e-01
4.94633317e-01 -1.88137040e-01 -8.74906242e-01 9.79804993e-02
4.40533161e-01 -1.84268624e-01 5.96534967e-01 -1.87048495e-01
-1.71957314e-02 -4.83149737e-01 -1.36128855e+00 8.68683219e-01
9.10375297e-01 -7.28532135e-01 -1.22712245e-02 -3.32657918e-02
1.26201379e+00 -2.15541020e-01 -5.07785797e-01 2.13633314e-01
2.07889274e-01 -9.86096501e-01 9.96779263e-01 -4.59472030e-01
-1.61435127e-01 -2.83699483e-01 -7.66841054e-01 -1.42705393e+00
-2.85595506e-01 -1.48608983e+00 -2.21607927e-02 1.63616157e+00
5.15588880e-01 -2.89116621e-01 9.09093499e-01 2.74714857e-01
-5.94865620e-01 -1.09399363e-01 -1.28053653e+00 -1.05250180e+00
-3.96317169e-02 -9.03090358e-01 6.55271590e-01 6.10403121e-01
-2.77621537e-01 3.61767679e-01 -6.39645219e-01 5.12298405e-01
6.19236231e-01 -1.05404425e+00 6.24092758e-01 -7.60781288e-01
-3.54084969e-01 -5.73547959e-01 -2.06195027e-01 -1.40808153e+00
-1.35034174e-01 -3.11936945e-01 2.20426232e-01 -1.34083855e+00
-4.92806196e-01 -1.91776291e-01 -4.81980473e-01 -4.02937718e-02
1.83483317e-01 6.05209656e-02 2.90322214e-01 -2.38489613e-01
-1.31643534e-01 6.81956232e-01 7.44410813e-01 -5.39238811e-01
-5.89628875e-01 4.75455880e-01 -1.01024616e+00 6.28826499e-01
1.11524546e+00 -3.70966762e-01 -3.07751000e-01 -5.65817475e-01
-3.43647212e-01 2.33673275e-01 1.32039785e-01 -1.35860229e+00
3.47247511e-01 5.60304187e-02 -3.51214409e-01 -5.89289546e-01
5.87269843e-01 -7.92345464e-01 -1.08922124e-01 4.46215272e-01
-2.96183705e-01 -3.59172732e-01 5.79454660e-01 4.26944256e-01
-3.59715015e-01 2.32788511e-02 1.22397351e+00 2.85166293e-01
-4.42302495e-01 1.33824527e-01 -9.52652514e-01 1.68470636e-01
3.07879031e-01 -1.22059651e-01 -3.78251761e-01 -1.17353451e+00
-8.58827889e-01 -8.11003894e-02 -2.84170866e-01 3.19612473e-01
7.00842738e-01 -1.14120662e+00 -1.06852210e+00 1.34848237e-01
-6.08934104e-01 -5.58386981e-01 5.77455223e-01 6.02314949e-01
-4.66065794e-01 4.48815048e-01 1.59014300e-01 -5.99966466e-01
-1.45266986e+00 1.47827774e-01 9.84236538e-01 7.41752461e-02
-4.55713868e-01 1.01677728e+00 -4.20241095e-02 -2.12330222e-01
9.30423737e-01 -2.82218754e-02 9.36867148e-02 -4.12649781e-01
8.77306283e-01 9.76751566e-01 5.52727520e-01 -4.16385263e-01
-2.00316042e-01 -4.34137732e-02 3.40642035e-01 -3.15168113e-01
1.00192523e+00 -7.13459015e-01 1.20657377e-01 -9.48329046e-02
1.48219359e+00 5.94798326e-01 -1.03097415e+00 -3.67946446e-01
-2.32237563e-01 -4.82556790e-01 6.75824881e-01 -8.76684189e-01
-1.18110323e+00 1.07217121e+00 1.19212174e+00 5.12195766e-01
1.37479246e+00 -3.95367980e-01 1.43480873e+00 3.53726625e-01
-1.18484229e-01 -1.24312532e+00 3.18320245e-01 8.90444100e-01
7.43075192e-01 -9.85460758e-01 -6.97954059e-01 -2.74504542e-01
-3.74412566e-01 9.01169240e-01 3.41040581e-01 6.37261569e-01
8.82871032e-01 8.36565137e-01 4.58004951e-01 3.45645159e-01
-5.62903702e-01 -2.96042293e-01 -3.69617075e-01 1.04600072e+00
4.09334689e-01 -1.47992130e-02 4.00292456e-01 5.20149350e-01
-4.67709064e-01 -3.60241443e-01 5.10911584e-01 3.88217688e-01
-9.86852825e-01 -8.58306468e-01 -6.15813017e-01 3.34149629e-01
-9.91216183e-01 -6.38342142e-01 -5.46210594e-02 2.67256945e-01
1.27766863e-01 1.99614120e+00 2.03372296e-02 -6.34605527e-01
6.32059872e-01 -7.94710368e-02 -4.05350924e-01 -3.12333882e-01
-7.41256714e-01 5.42929173e-01 4.86100048e-01 -1.66567296e-01
-4.12440151e-02 -3.84930938e-01 -1.07528651e+00 -9.52524126e-01
-1.72652185e-01 1.70044228e-01 8.45220208e-01 6.68982029e-01
4.47206587e-01 1.15057731e+00 8.44187319e-01 -4.58800167e-01
-6.67809844e-01 -8.49232554e-01 -8.77921283e-01 1.30465711e-02
1.08038914e+00 8.07000026e-02 -7.38316119e-01 -1.69957146e-01] | [14.947736740112305, 5.989822864532471] |
ebc73d29-9517-4ff3-9002-52c0e562a924 | type-enriched-hierarchical-contrastive | 2208.10081 | null | https://arxiv.org/abs/2208.10081v1 | https://arxiv.org/pdf/2208.10081v1.pdf | Type-enriched Hierarchical Contrastive Strategy for Fine-Grained Entity Typing | Fine-grained entity typing (FET) aims to deduce specific semantic types of the entity mentions in text. Modern methods for FET mainly focus on learning what a certain type looks like. And few works directly model the type differences, that is, let models know the extent that one type is different from others. To alleviate this problem, we propose a type-enriched hierarchical contrastive strategy for FET. Our method can directly model the differences between hierarchical types and improve the ability to distinguish multi-grained similar types. On the one hand, we embed type into entity contexts to make type information directly perceptible. On the other hand, we design a constrained contrastive strategy on the hierarchical structure to directly model the type differences, which can simultaneously perceive the distinguishability between types at different granularity. Experimental results on three benchmarks, BBN, OntoNotes, and FIGER show that our method achieves significant performance on FET by effectively modeling type differences. | ['Yu Luo', 'Zhou Fang', 'Shuang Zeng', 'Ning Jing', 'Haijin Liang', 'Xinyu Zuo'] | 2022-08-22 | null | https://aclanthology.org/2022.coling-1.212 | https://aclanthology.org/2022.coling-1.212.pdf | coling-2022-10 | ['entity-typing'] | ['natural-language-processing'] | [-3.05706710e-01 1.65466130e-01 -4.64658111e-01 -5.71801126e-01
-2.99895614e-01 -8.28956544e-01 4.97954160e-01 4.88955706e-01
-5.04379034e-01 5.82913160e-01 3.44530106e-01 -3.38092685e-01
3.42972249e-01 -1.18666470e+00 -7.11485684e-01 -4.21549469e-01
3.34794849e-01 4.16903645e-01 5.55865288e-01 -2.73138791e-01
3.05607826e-01 -2.10145831e-01 -1.59726381e+00 8.90801132e-01
1.11962736e+00 1.08653784e+00 3.64286155e-01 3.02435696e-01
-7.28524983e-01 8.21108341e-01 -7.51714051e-01 -5.14146864e-01
-1.30952597e-01 4.19964381e-02 -1.06947386e+00 -3.96446317e-01
4.69860077e-01 -1.36722341e-01 9.24673453e-02 1.34912300e+00
2.01103643e-01 -2.06019282e-01 8.03600967e-01 -1.34478271e+00
-7.85876215e-01 1.18867290e+00 -3.81824017e-01 1.03608258e-01
4.99037117e-01 -7.34724775e-02 1.63804579e+00 -9.19744551e-01
6.35795057e-01 1.59814775e+00 7.90088117e-01 7.27181256e-01
-1.14374411e+00 -6.48237884e-01 7.03684568e-01 4.67655331e-01
-1.20656490e+00 -1.54885277e-01 5.68402827e-01 -6.24787748e-01
6.23184562e-01 6.10943854e-01 2.98887014e-01 8.70864630e-01
-2.66457438e-01 1.00910187e+00 1.70008028e+00 -3.40383291e-01
2.84890503e-01 2.08055183e-01 8.03137779e-01 4.94024664e-01
4.62323517e-01 -1.74648404e-01 -2.89246500e-01 -1.30609915e-01
5.15507281e-01 -1.79928601e-01 -2.00521991e-01 7.50884488e-02
-1.31576383e+00 4.19476151e-01 5.59905112e-01 4.35158640e-01
1.40767666e-02 -1.56730235e-01 4.54504341e-01 1.35109484e-01
2.64784008e-01 6.16242826e-01 -9.38416779e-01 6.80797771e-02
-3.93325955e-01 3.87456536e-01 8.85640681e-01 1.32688868e+00
1.06391811e+00 -5.37805140e-01 -5.26505470e-01 8.72044802e-01
4.22006726e-01 3.20486188e-01 4.93975550e-01 -6.32237136e-01
5.61298788e-01 1.08979595e+00 2.52327293e-01 -8.40292275e-01
-4.78783995e-01 -6.64292574e-02 -6.94764674e-01 -5.97942285e-02
7.56526411e-01 -9.77589861e-02 -8.29892695e-01 1.78536320e+00
5.49148798e-01 2.09639192e-01 -7.23081157e-02 7.95823812e-01
1.14474666e+00 3.81311178e-01 2.23926425e-01 1.63470581e-01
1.90277731e+00 -6.55139744e-01 -6.76172733e-01 -2.95572132e-01
1.02062249e+00 -3.16766948e-01 1.25173318e+00 -1.99490070e-01
-7.66208649e-01 -3.81252527e-01 -8.97715986e-01 -2.69071996e-01
-8.67248833e-01 -2.44712636e-01 5.69041193e-01 6.31482482e-01
-3.76068115e-01 3.73684883e-01 -5.75411975e-01 1.61667272e-01
5.71742393e-02 4.93838303e-02 7.14721484e-03 1.34123757e-01
-1.66873777e+00 6.01238489e-01 7.20556438e-01 3.67765576e-02
-9.78618562e-02 -9.99762714e-01 -8.94874156e-01 2.66140312e-01
6.27071083e-01 -9.10136163e-01 1.36017883e+00 -9.36735153e-01
-9.14968669e-01 8.06383789e-01 -4.77430910e-01 -2.23475173e-01
2.45101154e-01 2.04050411e-02 -5.42186916e-01 -3.58267903e-01
1.85306057e-01 3.68633986e-01 2.45848700e-01 -1.41109502e+00
-1.23255205e+00 -3.11413318e-01 7.94876575e-01 1.84264794e-01
-2.44982883e-01 3.21956463e-02 -2.03720137e-01 -7.40469694e-01
1.50737301e-01 -6.73132420e-01 1.68978069e-02 -1.82546839e-01
-4.24863487e-01 -7.35809624e-01 4.68272477e-01 -3.63051176e-01
1.56773889e+00 -1.94161701e+00 1.24836890e-02 -1.23316282e-02
6.97055757e-01 -9.21860337e-04 3.67937565e-01 6.53901428e-04
2.78662294e-01 5.89946568e-01 3.37245725e-02 7.66381919e-02
4.75748569e-01 5.45893371e-01 -2.20292971e-01 -1.98406383e-01
-1.27378106e-01 9.07924414e-01 -1.07750356e+00 -5.91157734e-01
-2.75659889e-01 -1.27473310e-01 -8.08478355e-01 3.09034228e-01
-5.58182955e-01 -7.92961568e-03 -6.34359837e-01 3.96953613e-01
8.02543402e-01 -3.45490485e-01 5.25121272e-01 -6.46989644e-01
-1.62120342e-01 9.28085327e-01 -1.39472103e+00 1.03811657e+00
-6.38538301e-01 1.51399538e-01 1.87524203e-02 -6.27250075e-01
5.15201271e-01 1.78894833e-01 -2.18895242e-01 -6.84145868e-01
-2.11056426e-01 3.29820275e-01 1.10226035e-01 -3.29673350e-01
7.72453010e-01 -3.07234317e-01 -5.77027798e-01 9.23344567e-02
-3.69346708e-01 4.86414343e-01 1.13792561e-01 1.61383167e-01
7.43613660e-01 -2.23271772e-02 4.09240723e-01 -5.40908337e-01
3.61336470e-01 -1.32533252e-01 1.38629138e+00 8.69279921e-01
-1.18704215e-02 2.05513939e-01 7.30024815e-01 -3.62928480e-01
-6.50985420e-01 -1.05692303e+00 -6.32293075e-02 1.43777657e+00
8.07874024e-01 -7.24411786e-01 -6.98024929e-01 -1.22350883e+00
2.65413243e-02 5.98413944e-01 -8.16378295e-01 2.25536823e-01
-7.72838354e-01 -5.83632767e-01 6.75553799e-01 8.40966046e-01
7.14216888e-01 -9.74651515e-01 -2.17935219e-01 6.53878227e-02
-7.51725495e-01 -1.12605429e+00 -4.98370171e-01 7.60890171e-02
-4.41545486e-01 -1.02612996e+00 -4.05470759e-01 -9.55663264e-01
5.35234094e-01 -6.57849014e-02 1.34373724e+00 3.46112579e-01
4.48974013e-01 -3.33788604e-01 -4.52607602e-01 -3.39623362e-01
-3.00953418e-01 2.81340212e-01 -8.34152997e-02 -1.43869504e-01
4.48676437e-01 -3.56898934e-01 -4.55706507e-01 6.05639637e-01
-5.25046825e-01 2.98102230e-01 5.41404784e-01 8.69506121e-01
4.14849967e-01 1.90231323e-01 4.16491598e-01 -1.71939230e+00
1.87568665e-01 -5.17799377e-01 -3.26986581e-01 6.18920028e-01
-7.91982651e-01 4.95292693e-01 8.11293066e-01 -4.30355936e-01
-1.50576723e+00 -7.03158319e-01 -2.50065744e-01 2.15985164e-01
-3.51123244e-01 7.33839452e-01 -9.06099796e-01 4.50386167e-01
3.81134599e-01 -1.27138263e-02 -8.88542354e-01 -6.20094359e-01
3.49319607e-01 8.46849918e-01 6.37361348e-01 -1.35916293e+00
5.96308887e-01 1.62018076e-01 -3.16773951e-01 -4.32252645e-01
-1.09366488e+00 -5.31143963e-01 -7.30436027e-01 5.75139336e-02
7.41606712e-01 -8.25390577e-01 -5.99366605e-01 8.10821593e-01
-9.92256641e-01 -6.36144638e-01 2.06907824e-01 -7.80755505e-02
-4.95321564e-02 6.26275957e-01 -8.22051585e-01 -3.71485502e-01
4.93043959e-02 -7.30282784e-01 1.14282990e+00 2.34955832e-01
-1.81417882e-01 -1.26744354e+00 -2.78127789e-01 4.73179035e-02
7.19521716e-02 2.62045320e-02 1.53877735e+00 -6.64143145e-01
-5.73339522e-01 3.84667635e-01 -5.34725964e-01 -1.99048087e-01
2.19217360e-01 -1.55330226e-01 -8.33108425e-01 1.85473368e-01
-3.67202908e-01 8.66704658e-02 7.02020943e-01 -1.37283549e-01
1.19050694e+00 -9.40508604e-01 -5.54188669e-01 5.11212707e-01
1.13043499e+00 4.40079086e-02 7.14217663e-01 4.79738623e-01
1.20105970e+00 5.38211584e-01 8.32646608e-01 1.68527231e-01
1.03032422e+00 8.46008718e-01 -2.47519221e-02 9.03364494e-02
-1.29881278e-01 -4.65913445e-01 -2.88886088e-03 8.00387084e-01
-7.45752230e-02 -1.24437176e-01 -7.87880421e-01 4.90128994e-01
-1.90180242e+00 -1.04560173e+00 -3.55921507e-01 1.99231589e+00
1.39847529e+00 4.75616127e-01 1.41942859e-01 8.36772919e-02
1.01192725e+00 7.49593601e-02 -2.80490845e-01 -1.83900729e-01
-1.21735325e-02 -2.29024306e-01 4.39636230e-01 4.81573373e-01
-8.84175122e-01 9.90553081e-01 5.40206671e+00 1.01331151e+00
-8.89286041e-01 -1.69067696e-01 3.08000028e-01 5.14827251e-01
-7.74834633e-01 2.66798615e-01 -1.22144318e+00 9.46450233e-01
3.67992550e-01 -3.27222258e-01 4.73875441e-02 6.58751965e-01
-3.79051208e-01 4.32889760e-02 -1.31607592e+00 5.84859848e-01
-6.03518963e-01 -1.21999896e+00 3.52929793e-02 7.49045983e-02
3.20538014e-01 -5.17091751e-01 -2.15501025e-01 7.31416285e-01
5.54749072e-01 -6.23966038e-01 1.11384916e+00 3.02445441e-01
7.03000486e-01 -5.44226110e-01 6.58263862e-01 5.35503507e-01
-1.54746282e+00 1.72262609e-01 -3.85995954e-01 -3.61806870e-01
-2.81692773e-01 6.57874405e-01 -5.76269090e-01 7.10571468e-01
5.61416209e-01 5.75230598e-01 -6.59314692e-01 7.76124239e-01
-3.73222440e-01 5.74368298e-01 -1.91278160e-02 -2.75767922e-01
-1.43679515e-01 2.01984569e-01 4.48251486e-01 1.54142797e+00
-1.95864346e-02 3.16449493e-01 4.40046668e-01 8.22864950e-01
-1.70109406e-01 -2.86186427e-01 1.21827396e-02 3.11564088e-01
1.10806739e+00 1.24695385e+00 -5.09890556e-01 -8.19924593e-01
-6.22144997e-01 6.63445473e-01 5.99594593e-01 1.82920575e-01
-6.93454266e-01 -4.90794271e-01 8.81611168e-01 7.78379664e-02
3.57384235e-01 -8.65996704e-02 -6.38456583e-01 -1.62452018e+00
2.16375709e-01 -8.67900789e-01 8.78746808e-01 -5.63956797e-01
-1.56941891e+00 2.98709512e-01 1.20207466e-01 -9.56594467e-01
-4.97680306e-02 -8.17271292e-01 -2.96744972e-01 8.58947337e-01
-1.46490717e+00 -1.10267341e+00 -3.13791096e-01 3.11217815e-01
3.96661401e-01 5.54292977e-01 7.38810122e-01 2.84835935e-01
-5.78091085e-01 1.12363803e+00 -3.53937149e-01 6.31731689e-01
6.66183293e-01 -2.00529265e+00 5.93949437e-01 7.86132336e-01
-1.16555192e-01 1.14818919e+00 7.34565437e-01 -6.09383285e-01
-1.06224728e+00 -1.14041615e+00 1.25267804e+00 -4.93483454e-01
6.86636865e-01 -4.95833278e-01 -1.35374868e+00 8.33539784e-01
-1.06574275e-01 -6.36112839e-02 5.37283361e-01 9.37196732e-01
-1.34522772e+00 5.38554527e-02 -1.14525020e+00 7.59357452e-01
1.32671118e+00 -7.90562451e-01 -1.17137086e+00 -2.09662154e-01
7.30804920e-01 -5.36719024e-01 -1.06329048e+00 3.67279619e-01
7.91455448e-01 -9.19221699e-01 7.05126762e-01 -5.27719319e-01
4.07355636e-01 -4.87887770e-01 -2.56227553e-01 -1.56673717e+00
-5.20705700e-01 -2.56411493e-01 -2.11330414e-01 1.78581953e+00
5.50986111e-01 -8.88474643e-01 6.07934535e-01 9.92807865e-01
-2.83621103e-01 -5.52075565e-01 -4.34984237e-01 -1.00432360e+00
3.23777020e-01 -8.88314843e-02 1.14793980e+00 1.27437127e+00
4.02719289e-01 5.09607255e-01 1.35962620e-01 5.32183826e-01
4.18498665e-01 6.70624793e-01 5.33628464e-01 -1.32502472e+00
-5.48323929e-01 -6.70479357e-01 -3.17404777e-01 -1.44829273e+00
2.64396876e-01 -1.01023471e+00 2.03602716e-01 -1.44994676e+00
2.78364688e-01 -1.22504234e+00 -5.31591594e-01 6.59692526e-01
-8.54965985e-01 -1.63255453e-01 1.04228728e-01 2.36967519e-01
-5.67067623e-01 -9.40283462e-02 1.06860530e+00 -2.43224502e-01
2.06400439e-01 -1.75425425e-01 -1.12450790e+00 9.84833539e-01
4.70405966e-01 -4.10033613e-01 -1.77972019e-02 -5.95199287e-01
6.66515768e-01 -1.31111488e-01 3.32180321e-01 -6.63006842e-01
2.45197594e-01 -5.40556371e-01 2.20066365e-02 -4.66273218e-01
-1.34249285e-01 -5.38456500e-01 -1.65669560e-01 1.66836858e-01
-6.12401307e-01 -2.93799881e-02 -1.00155234e-01 4.18358445e-01
-1.72821864e-01 -2.01959744e-01 5.28663993e-01 -1.20097421e-01
-1.45963967e+00 2.48069167e-01 -6.54512048e-02 7.30415642e-01
3.67607325e-01 1.71422422e-01 -7.90793121e-01 2.80811936e-01
-5.28916657e-01 2.98888564e-01 7.88245499e-01 4.75431353e-01
9.20007303e-02 -1.43169165e+00 -4.97210860e-01 3.56282480e-02
5.08470774e-01 1.09636150e-01 1.40376657e-01 4.73478049e-01
1.74495224e-02 -3.26791592e-02 -1.08943276e-01 -4.29082960e-01
-1.51561201e+00 7.50403941e-01 4.59812492e-01 -3.90961409e-01
-5.05170763e-01 8.18180084e-01 9.56757724e-01 -5.72348893e-01
2.12526321e-02 -7.55628049e-01 -3.10640603e-01 1.54692203e-01
6.25987232e-01 1.52121291e-01 4.66866083e-02 -4.36064959e-01
-5.07943690e-01 3.99118274e-01 -4.33711916e-01 1.12143062e-01
7.96519101e-01 -1.65247753e-01 -2.70760775e-01 7.98962772e-01
1.06055844e+00 5.09660065e-01 -1.05329478e+00 -4.82092768e-01
4.86080259e-01 -4.03792948e-01 -3.14851224e-01 -1.01940572e+00
-7.43635356e-01 5.24108112e-01 -4.16573398e-02 7.16395557e-01
9.36272800e-01 1.09628618e-01 9.32416260e-01 4.46785957e-01
6.65860832e-01 -9.56531286e-01 -4.48130488e-01 8.25445712e-01
3.08820993e-01 -8.87174189e-01 -2.65326947e-01 -1.02143204e+00
-4.17184830e-01 7.39392221e-01 7.58164227e-01 1.18417233e-01
4.56903934e-01 6.36765778e-01 1.54368505e-01 -2.95573864e-02
-8.02570760e-01 -3.66622597e-01 4.86989141e-01 6.65402949e-01
7.15906143e-01 5.70427358e-01 -4.40049410e-01 9.56804514e-01
-3.27234387e-01 -2.97496438e-01 5.41564226e-01 9.04387176e-01
-4.38562274e-01 -1.25913203e+00 -1.19466186e-01 4.77168918e-01
-5.92162669e-01 -4.83566195e-01 -4.62994814e-01 7.05710232e-01
3.06665540e-01 8.85595500e-01 2.35278402e-02 -4.45746332e-01
3.50371152e-01 -1.95091181e-02 6.62776947e-01 -7.63190448e-01
-7.95260668e-01 -3.17588508e-01 6.36641026e-01 -2.74671614e-01
-2.52970338e-01 -6.45538032e-01 -1.39552128e+00 -3.60856891e-01
-3.18217397e-01 2.85995722e-01 1.33493543e-01 1.24611282e+00
2.82727540e-01 4.83481795e-01 5.63825905e-01 -1.33379638e-01
-4.04248983e-01 -8.20833385e-01 -6.32950962e-01 8.11955154e-01
3.69879514e-01 -9.90634441e-01 -3.24933618e-01 -6.07945509e-02] | [9.648143768310547, 8.762304306030273] |
15349305-5b00-4c18-bf3e-9aba8d26cf8f | analysis-of-recent-trends-in-face-recognition | 2304.11725 | null | https://arxiv.org/abs/2304.11725v1 | https://arxiv.org/pdf/2304.11725v1.pdf | Analysis of Recent Trends in Face Recognition Systems | With the tremendous advancements in face recognition technology, face modality has been widely recognized as a significant biometric identifier in establishing a person's identity rather than any other biometric trait like fingerprints that require contact sensors. However, due to inter-class similarities and intra-class variations, face recognition systems generate false match and false non-match errors respectively. Recent research focuses on improving the robustness of extracted features and the pre-processing algorithms to enhance recognition accuracy. Since face recognition has been extensively used for several applications ranging from law enforcement to surveillance systems, the accuracy and performance of face recognition must be the finest. In this paper various face recognition systems are discussed and analysed like RPRV, LWKPCA, SVM Model, LTrP based SPM and a deep learning framework for recognising images from CCTV. All these face recognition methods, their implementations and performance evaluations are compared to derive the best outcome for future developmental works. | ['Krishnendu K. S'] | 2023-04-23 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [ 4.97232258e-01 -2.33612686e-01 1.94634683e-02 -6.98443770e-01
-2.03959003e-01 -4.46848691e-01 6.82616293e-01 -3.10512543e-01
-1.94218069e-01 6.23623669e-01 -2.07136616e-01 -2.60673523e-01
-3.41957390e-01 -7.19819903e-01 -1.41915083e-01 -8.65938485e-01
1.94578573e-01 2.10394531e-01 -1.13850310e-01 2.12637782e-01
5.74148655e-01 1.12354064e+00 -1.87656713e+00 2.89334476e-01
4.75792974e-01 1.24956000e+00 -4.38527137e-01 5.06961405e-01
-1.88506067e-01 1.95113957e-01 -3.92191947e-01 -4.44540560e-01
3.64233822e-01 -1.87137544e-01 -6.51708186e-01 -1.73430108e-02
6.25478745e-01 -3.83964509e-01 -1.26702860e-01 7.83290088e-01
7.69786179e-01 -2.00629398e-01 8.19021106e-01 -1.15348959e+00
-5.84798515e-01 -6.36185482e-02 -7.79828310e-01 1.46018714e-01
7.33611703e-01 -6.84111705e-03 6.67360669e-04 -8.85617495e-01
1.74912959e-01 1.24081898e+00 8.73847067e-01 6.15790546e-01
-8.91264379e-01 -8.55473518e-01 -6.26474977e-01 2.68472046e-01
-1.36077464e+00 -8.45711708e-01 5.45645058e-01 -3.54512304e-01
9.51053679e-01 3.29361200e-01 2.49947503e-01 8.47295463e-01
3.07080477e-01 2.68331528e-01 1.57665455e+00 -5.86146712e-01
-1.92879438e-01 4.92937326e-01 2.55728632e-01 7.17408359e-01
1.19579233e-01 3.79495859e-01 -4.83510703e-01 -2.46695936e-01
8.19365323e-01 4.13973778e-02 1.48935961e-02 6.92219585e-02
-4.75953400e-01 3.70833367e-01 5.23779131e-02 5.36247432e-01
-5.41588187e-01 -3.42310816e-01 3.57208073e-01 3.81598711e-01
-1.13131240e-01 -2.59296596e-01 -2.95945138e-01 -1.77508861e-01
-9.76176977e-01 -1.49068028e-01 6.89572036e-01 6.09566271e-01
4.39145654e-01 1.09782510e-01 -2.15251029e-01 1.03732097e+00
5.11781633e-01 8.04495752e-01 4.20178235e-01 -4.82176304e-01
2.07022973e-03 7.81137645e-01 1.21729262e-02 -1.15707755e+00
-2.60846406e-01 -4.46161665e-02 -7.88217247e-01 2.31934831e-01
3.58097970e-01 1.77847443e-03 -1.07506347e+00 1.08665693e+00
3.33139598e-01 3.53083253e-01 -7.29694143e-02 4.34739232e-01
8.70111585e-01 2.95705408e-01 1.14961430e-01 -1.05450377e-01
1.52516019e+00 -2.82755703e-01 -4.53699648e-01 2.13856369e-01
-1.67906150e-01 -1.15676117e+00 4.26211149e-01 3.76567602e-01
-6.95127904e-01 -7.72783101e-01 -7.98271239e-01 2.80467004e-01
-5.42892218e-01 3.60244125e-01 3.43903661e-01 1.50591218e+00
-1.02700031e+00 4.12701428e-01 -4.37306076e-01 -8.27251971e-01
7.83369660e-01 1.02549195e+00 -7.76072085e-01 8.54563341e-03
-7.49777734e-01 8.90170395e-01 4.01435159e-02 4.71542001e-01
-3.99492592e-01 -2.98244238e-01 -4.37001526e-01 -1.04663312e-01
-1.51985258e-01 -6.02152236e-02 8.34543765e-01 -9.46764290e-01
-1.61439824e+00 1.28334451e+00 -5.81246257e-01 -2.33541638e-01
3.47502410e-01 2.13984624e-01 -7.37834513e-01 -4.62023914e-02
-4.24513698e-01 1.32375941e-01 1.11571407e+00 -8.26753795e-01
-4.81336862e-01 -8.55904102e-01 -4.61257279e-01 -2.68920243e-01
-4.73814830e-02 6.04050040e-01 3.58357638e-01 -2.83265635e-02
2.31315449e-01 -7.73638427e-01 6.23091757e-01 -1.36776924e-01
-2.22058237e-01 -4.77100343e-01 1.47596991e+00 -7.52382159e-01
5.64049244e-01 -2.10656929e+00 -5.96233249e-01 5.50954044e-01
-2.84732997e-01 9.28461432e-01 7.55167753e-02 5.78218520e-01
2.16650311e-03 -2.88319401e-02 7.01387152e-02 -8.60326341e-04
-2.71986485e-01 5.89328967e-02 -1.16115034e-01 6.40135825e-01
4.18053299e-01 6.52562797e-01 -2.19053134e-01 -4.55940872e-01
5.07554889e-01 9.19802606e-01 1.18585780e-01 1.80688202e-01
6.57098293e-01 3.99770051e-01 -5.03478348e-01 1.18574929e+00
1.14591968e+00 2.28212208e-01 -8.35462436e-02 -4.25118893e-01
-6.63063452e-02 -2.78331518e-01 -1.14611006e+00 7.89863229e-01
-2.67132789e-01 6.03231847e-01 1.73852429e-01 -1.19545829e+00
1.26273739e+00 6.99908555e-01 3.54486495e-01 -6.11922860e-01
4.10210699e-01 4.11697388e-01 -7.84076676e-02 -8.62464547e-01
1.71294451e-01 -1.81230474e-02 5.97314477e-01 3.67614001e-01
8.60295817e-03 4.77196276e-01 -3.70388478e-01 -5.75379670e-01
6.11833215e-01 9.64478776e-02 3.46343279e-01 -2.45417535e-01
1.20744741e+00 -3.89315784e-01 2.93513894e-01 4.94897634e-01
-5.88141978e-01 5.54308295e-01 4.68345322e-02 -5.53950787e-01
-6.23600841e-01 -9.17519391e-01 -8.13287258e-01 5.93136430e-01
-2.63146698e-01 5.47971845e-01 -7.67839491e-01 -7.23058105e-01
2.65788525e-01 -1.53132351e-02 -4.62352157e-01 2.00000945e-02
-4.26932633e-01 -6.89504504e-01 9.72895026e-01 3.48258078e-01
1.02245760e+00 -1.19423652e+00 -4.72020239e-01 3.36941443e-02
4.38367844e-01 -8.46647918e-01 1.08271884e-02 -2.61557162e-01
-7.58477628e-01 -1.44552648e+00 -6.66715443e-01 -8.84186506e-01
8.19306076e-01 1.12418815e-01 5.23022354e-01 3.96171361e-01
-6.08452916e-01 4.79218960e-01 -1.83262840e-01 -5.89747965e-01
-3.07241559e-01 -1.56477153e-01 3.09146374e-01 3.66759688e-01
1.15984368e+00 -4.37877506e-01 -6.78947628e-01 4.03544068e-01
-5.63906848e-01 -6.46089613e-01 6.02259636e-01 7.51419663e-01
2.43107546e-02 -5.20716533e-02 7.91238546e-01 -8.30102205e-01
6.51121438e-01 -2.30687052e-01 -4.43789333e-01 5.36788702e-01
-5.97303569e-01 -3.39772791e-01 3.70126724e-01 2.71931440e-02
-1.17110574e+00 1.57954648e-01 -2.56665140e-01 1.69739515e-01
-6.81329846e-01 2.16034323e-01 -1.28216550e-01 -7.69523501e-01
4.54484224e-01 6.02384865e-01 4.23610777e-01 -4.10205483e-01
-3.75232995e-01 1.37370765e+00 4.27100092e-01 -4.32315052e-01
7.08171606e-01 3.01783264e-01 1.43942729e-01 -1.31182933e+00
-9.13282260e-02 -5.72553396e-01 -8.82855892e-01 -3.85444820e-01
6.83344603e-01 -4.94115233e-01 -1.24951816e+00 1.16538298e+00
-1.07691431e+00 6.66998327e-01 3.58641148e-01 2.16963768e-01
9.72243678e-03 4.68865216e-01 -4.43508059e-01 -1.30980957e+00
-6.45270884e-01 -1.17723393e+00 1.11411905e+00 7.33219206e-01
1.19030671e-02 -7.24941254e-01 -1.83063895e-01 5.58593154e-01
8.16488087e-01 1.44312263e-01 6.99176788e-01 -6.29275024e-01
-2.51603752e-01 -6.54333770e-01 -4.98402119e-01 4.02997106e-01
6.95989668e-01 3.62738043e-01 -1.54073703e+00 -2.83162206e-01
1.16746992e-01 -1.87607229e-01 4.57109839e-01 2.77103275e-01
9.82710361e-01 -1.15096077e-01 -4.83666182e-01 4.48558182e-01
1.40851986e+00 7.91272163e-01 9.83835399e-01 5.18757268e-04
3.06313753e-01 6.52794838e-01 9.65390205e-02 2.53739089e-01
-6.94644526e-02 4.35387760e-01 1.68380644e-02 2.06643745e-01
1.14488013e-01 8.98833498e-02 1.50630310e-01 1.90538794e-01
-5.18907666e-01 1.44283965e-01 -8.54993820e-01 -2.35667434e-02
-1.29953051e+00 -1.28891063e+00 -5.13739623e-02 2.41899610e+00
4.33006644e-01 -5.29943943e-01 1.19915381e-01 5.84165394e-01
8.16611052e-01 -1.93767399e-01 -3.55864614e-01 -6.57608628e-01
-3.24467830e-02 6.80553436e-01 3.24675471e-01 2.19558835e-01
-1.09265041e+00 5.89012504e-01 6.57644367e+00 3.32198709e-01
-1.62846649e+00 -1.50357574e-01 8.08057606e-01 4.05491173e-01
4.97561425e-01 -4.52710301e-01 -9.13467348e-01 4.85405594e-01
7.99189866e-01 2.87246376e-01 3.42724442e-01 5.93439519e-01
1.66616082e-01 -2.37758309e-01 -1.03669596e+00 1.27571332e+00
1.05953701e-01 -9.91146266e-01 -5.20962887e-02 1.97792768e-01
5.16759455e-01 -3.82174522e-01 2.79829651e-01 2.70875264e-02
-4.85692531e-01 -1.42186201e+00 5.58176562e-02 5.82651436e-01
7.14269698e-01 -7.06728458e-01 1.10743284e+00 -3.09124552e-02
-1.04462004e+00 -2.62610205e-02 -4.28911150e-01 -6.19218275e-02
-2.34723896e-01 1.91750556e-01 -8.93060446e-01 4.08150285e-01
5.03025770e-01 3.12373251e-01 -4.71799314e-01 9.37415481e-01
2.50330418e-01 4.75910753e-01 -4.27603632e-01 -5.31186275e-02
-1.16823241e-01 -1.89756781e-01 1.31134391e-01 1.12899101e+00
5.04209042e-01 2.55875051e-01 -2.25835308e-01 7.11414099e-01
9.30093005e-02 1.34887338e-01 -6.54418886e-01 -1.07292339e-01
6.72271252e-01 1.29239166e+00 -5.58578670e-01 -2.84503736e-02
-4.31207687e-01 8.01314056e-01 -1.65839523e-01 2.35633001e-01
-4.90964860e-01 -2.62027174e-01 6.73863053e-01 3.17907482e-01
1.80288121e-01 -2.91829128e-02 -2.29398906e-01 -6.33185446e-01
1.86147504e-02 -1.05133736e+00 1.34600982e-01 -2.22517267e-01
-1.13706732e+00 6.10922039e-01 -2.82163233e-01 -8.72709215e-01
-1.68579221e-01 -9.26250994e-01 -8.22512925e-01 1.42338979e+00
-1.22150385e+00 -1.45146251e+00 -3.98466170e-01 4.76767987e-01
1.22588433e-01 -7.97386944e-01 9.51991916e-01 3.47786367e-01
-6.28751159e-01 9.15530503e-01 1.57080844e-01 2.80240119e-01
4.77491856e-01 -5.10001063e-01 7.16631711e-02 6.52337015e-01
-2.11982355e-01 9.05047596e-01 3.03291082e-01 -5.26469111e-01
-1.78735638e+00 -5.91125011e-01 8.22996080e-01 -4.01541799e-01
-6.05012551e-02 6.23895898e-02 -7.60239363e-01 3.55778188e-01
1.97306037e-01 5.18161468e-02 6.89574063e-01 5.33595271e-02
-2.86015242e-01 -5.51456094e-01 -1.88819337e+00 -7.10823163e-02
5.12812614e-01 -9.01268482e-01 -3.17620814e-01 -6.35093078e-02
-6.02338076e-01 -8.02355185e-02 -8.68893862e-01 4.78445113e-01
1.21122062e+00 -1.32480276e+00 9.16809440e-01 -3.90620619e-01
-2.74638403e-02 -2.57447124e-01 -2.31363475e-02 -5.80002844e-01
-2.45112985e-01 -1.58391163e-01 2.96844333e-01 1.59191608e+00
9.31315348e-02 -1.01020741e+00 1.04926312e+00 9.29113090e-01
5.38400531e-01 -6.51394904e-01 -1.12782860e+00 -6.32140815e-01
-3.87642890e-01 2.69433726e-02 7.50922978e-01 8.58482003e-01
-2.98845947e-01 -1.32135779e-01 -4.20070022e-01 2.86229074e-01
8.36605728e-01 -1.32446736e-01 5.26002467e-01 -1.47827077e+00
1.59346417e-01 -2.80575842e-01 -9.67766702e-01 -2.07642302e-01
-7.78790116e-02 -4.62176114e-01 -1.83331132e-01 -1.05051136e+00
2.38457218e-01 -5.02370656e-01 -3.77632678e-01 4.75873560e-01
1.72538221e-01 5.81538677e-01 -1.16723441e-01 4.68781143e-02
3.28866661e-01 6.34621177e-03 8.47802103e-01 -1.08475089e-01
7.75846140e-03 3.32996041e-01 -3.64524215e-01 3.80872726e-01
8.04579616e-01 -1.36842027e-01 -2.91774035e-01 -2.25561559e-01
-2.84188718e-01 -4.59321626e-02 3.95991892e-01 -1.12146616e+00
2.92446196e-01 -5.19324886e-03 9.60736156e-01 -4.13646191e-01
3.21909338e-01 -1.01508760e+00 3.29514444e-01 4.33159858e-01
1.72729880e-01 -4.14673574e-02 2.80283868e-01 8.43653083e-02
-2.60456711e-01 -2.98759401e-01 1.02278721e+00 -1.59882568e-02
-6.17362678e-01 3.49029452e-01 -3.63767475e-01 -7.06519544e-01
1.06748676e+00 -1.01277363e+00 -3.15506160e-01 -6.90655857e-02
-4.36448216e-01 -3.15571070e-01 2.53513962e-01 4.41171259e-01
8.42187166e-01 -1.07958794e+00 -6.60888731e-01 8.09976637e-01
-1.79364309e-01 -5.68001628e-01 2.84173131e-01 7.26475537e-01
-5.27067125e-01 6.59455240e-01 -7.16178596e-01 -6.00823700e-01
-2.00172520e+00 1.06137350e-01 4.83805150e-01 3.91571283e-01
-2.07701400e-02 8.59168291e-01 -3.79685432e-01 -3.28391820e-01
2.22945780e-01 1.04280181e-01 -4.82506245e-01 -1.54580504e-01
7.57569373e-01 5.68145573e-01 2.88085431e-01 -1.01373172e+00
-7.11783886e-01 9.39007819e-01 -2.85294086e-01 3.94690156e-01
1.08421230e+00 2.58441120e-02 -4.88594413e-01 3.54895927e-02
1.06624341e+00 -2.17593715e-01 -6.60794795e-01 -3.30386832e-02
1.12252466e-01 -7.60605454e-01 -4.68402244e-02 -9.80608642e-01
-1.05522239e+00 8.81480098e-01 1.34728098e+00 2.11755201e-01
1.18293977e+00 -5.43792009e-01 4.35258210e-01 3.78559411e-01
6.30120218e-01 -9.53746021e-01 -4.48029667e-01 3.01313162e-01
7.46672034e-01 -1.55586529e+00 -5.50601631e-02 -3.58707935e-01
-8.33328739e-02 1.38092315e+00 4.63974893e-01 -3.43960263e-02
9.65236783e-01 2.70634025e-01 1.52126789e-01 -1.59587324e-01
-1.13834381e-01 3.32912385e-01 5.22819698e-01 1.04095531e+00
1.04787135e+00 -1.27891898e-01 -3.56944472e-01 1.74128309e-01
4.97739278e-02 4.39958155e-01 -5.97723350e-02 8.89733613e-01
-4.01919067e-01 -1.39927495e+00 -6.60261393e-01 8.39461803e-01
-7.33437777e-01 3.17052901e-01 -5.71726143e-01 5.88525832e-01
1.52730465e-01 1.02929783e+00 -1.46237716e-01 -5.17768025e-01
1.85501620e-01 4.49694157e-01 7.95895159e-01 -2.34924838e-01
-7.77202368e-01 -4.97293591e-01 1.56745617e-03 -3.07737768e-01
-6.95617318e-01 -6.96001530e-01 -9.39435124e-01 -7.58506894e-01
-3.06414694e-01 -5.92201240e-02 8.89814973e-01 1.17942679e+00
3.79705399e-01 -1.23694219e-01 7.07179964e-01 -6.36452019e-01
-6.16073787e-01 -1.01990700e+00 -6.22055352e-01 1.96346730e-01
1.76241875e-01 -6.11627460e-01 7.44309276e-02 3.27620432e-02] | [13.258362770080566, 0.9296302199363708] |
212e464c-9f02-4463-a7d8-35b6232295c0 | assessing-mortality-prediction-through | 2207.10872 | null | https://arxiv.org/abs/2207.10872v1 | https://arxiv.org/pdf/2207.10872v1.pdf | Assessing mortality prediction through different representation models based on concepts extracted from clinical notes | Recent years have seen particular interest in using electronic medical records (EMRs) for secondary purposes to enhance the quality and safety of healthcare delivery. EMRs tend to contain large amounts of valuable clinical notes. Learning of embedding is a method for converting notes into a format that makes them comparable. Transformer-based representation models have recently made a great leap forward. These models are pre-trained on large online datasets to understand natural language texts effectively. The quality of a learning embedding is influenced by how clinical notes are used as input to representation models. A clinical note has several sections with different levels of information value. It is also common for healthcare providers to use different expressions for the same concept. Existing methods use clinical notes directly or with an initial preprocessing as input to representation models. However, to learn a good embedding, we identified the most essential clinical notes section. We then mapped the extracted concepts from selected sections to the standard names in the Unified Medical Language System (UMLS). We used the standard phrases corresponding to the unique concepts as input for clinical models. We performed experiments to measure the usefulness of the learned embedding vectors in the task of hospital mortality prediction on a subset of the publicly available Medical Information Mart for Intensive Care (MIMIC-III) dataset. According to the experiments, clinical transformer-based representation models produced better results with getting input generated by standard names of extracted unique concepts compared to other input formats. The best-performing models were BioBERT, PubMedBERT, and UmlsBERT, respectively. | ['Maryam Lotfi Shahreza', 'Nasser Ghadiri', 'Hoda Memarzadeh'] | 2022-07-22 | null | null | null | null | ['mortality-prediction'] | ['medical'] | [ 8.19819281e-04 2.09183455e-01 -2.62543291e-01 -1.65145621e-01
-8.96469235e-01 -5.17652214e-01 3.57446730e-01 1.00607562e+00
-6.11840963e-01 5.05257487e-01 8.93961728e-01 -4.45771068e-01
-4.39014554e-01 -9.97002423e-01 -3.05604190e-01 -4.16255027e-01
-2.30361167e-02 5.61226070e-01 -2.69469708e-01 -2.32463792e-01
6.67715305e-03 2.59431183e-01 -9.27920938e-01 9.01747525e-01
6.81886971e-01 7.53009558e-01 1.01142362e-01 5.13186574e-01
-5.05683959e-01 9.57836330e-01 -6.05736494e-01 -5.09979725e-01
2.40674227e-01 -5.45631647e-01 -8.07630301e-01 -5.46336710e-01
-3.02662641e-01 -6.87079411e-03 -4.18288112e-01 7.21620142e-01
6.41743124e-01 -1.35730445e-01 1.02169371e+00 -9.23306286e-01
-7.89820015e-01 8.21850002e-01 7.58697912e-02 2.56958175e-02
5.59191942e-01 -1.11767359e-01 1.00028431e+00 -8.35836470e-01
9.39683497e-01 8.97077858e-01 7.52120912e-01 6.78018570e-01
-9.51775551e-01 -5.49285114e-01 -4.43777621e-01 1.51631281e-01
-1.23929799e+00 2.81673465e-02 3.42150480e-01 -7.95684636e-01
1.12683678e+00 4.12855327e-01 6.29097939e-01 1.04140568e+00
6.21431947e-01 3.71920466e-01 6.78562284e-01 -4.11844790e-01
1.33554032e-02 6.59190476e-01 4.08735484e-01 6.11733735e-01
4.31377500e-01 -1.04326233e-01 -1.95797756e-01 -6.85757637e-01
3.33907783e-01 7.48378634e-01 -5.99043190e-01 -2.32830197e-02
-1.48348308e+00 1.06976151e+00 5.74411988e-01 6.11792266e-01
-6.05759561e-01 -3.11087191e-01 7.21843362e-01 2.82351613e-01
7.99849778e-02 8.24722409e-01 -6.15697503e-01 -1.23316787e-01
-6.19177938e-01 2.39380058e-02 7.93768764e-01 9.00611699e-01
2.06182420e-01 -6.29298210e-01 -4.94475096e-01 7.91094124e-01
1.16845392e-01 1.90873414e-01 9.93643165e-01 -3.93133998e-01
6.99876904e-01 1.04216194e+00 5.46116605e-02 -9.22594070e-01
-3.51593316e-01 -3.27807069e-01 -1.07687962e+00 -4.29040462e-01
1.45972416e-01 -7.73602799e-02 -8.20505321e-01 1.30814791e+00
-9.56307575e-02 -8.98836553e-02 7.35673130e-01 3.87196243e-01
1.13232052e+00 7.76616275e-01 3.96368414e-01 -7.56753087e-02
1.64115489e+00 -5.77550650e-01 -9.87105906e-01 2.87186265e-01
1.30842757e+00 -6.75834596e-01 6.94038570e-01 1.17396787e-01
-7.01339841e-01 -2.36389011e-01 -7.63214052e-01 -8.60086351e-04
-7.58016825e-01 1.82059720e-01 1.46799788e-01 2.53892928e-01
-8.31547916e-01 7.62058914e-01 -7.23168254e-01 -3.86909932e-01
4.28390086e-01 1.37312725e-01 -7.90784478e-01 -1.35688886e-01
-1.38941646e+00 1.15941727e+00 3.88281703e-01 -3.68007243e-01
-3.03092808e-01 -1.07148790e+00 -1.11354923e+00 3.35096687e-01
-1.09421775e-01 -9.81522202e-01 8.58685315e-01 -5.23569047e-01
-8.03110301e-01 8.15531850e-01 1.33632943e-01 -4.60326701e-01
2.80064553e-01 -1.80851340e-01 -5.21602809e-01 2.84339696e-01
4.60190922e-02 2.96718925e-01 2.67386734e-01 -7.93414891e-01
-4.09620374e-01 -2.88578153e-01 -9.64095965e-02 -2.24401936e-01
-7.23916292e-01 1.25225022e-01 -1.48872748e-01 -8.71141434e-01
-1.68260843e-01 -7.01988041e-01 -4.15283918e-01 -3.02267224e-02
-2.40177527e-01 -2.00264588e-01 2.50897765e-01 -8.97059143e-01
1.66701770e+00 -2.18861532e+00 1.40141711e-01 1.05010808e-01
4.67632413e-01 3.37824583e-01 -4.58343178e-02 9.65707958e-01
-4.53558117e-01 5.18485844e-01 -2.53301710e-01 6.56473078e-03
-2.91841805e-01 3.40508521e-02 -2.29333490e-01 4.72240113e-02
2.82653928e-01 9.05278504e-01 -1.06756830e+00 -5.58860242e-01
6.91190138e-02 7.85345018e-01 -6.96894944e-01 5.76268315e-01
2.45376691e-01 8.06752518e-02 -5.65105498e-01 4.32373643e-01
1.19214743e-01 -4.78292197e-01 2.70959765e-01 -3.97889346e-01
2.58851677e-01 3.81406516e-01 -6.61132336e-01 1.39818633e+00
-6.02851629e-01 2.82254100e-01 -6.49373412e-01 -7.37058520e-01
7.99826920e-01 9.30413067e-01 9.51876819e-01 -3.12667936e-01
1.25061408e-01 1.60559803e-01 5.67887537e-02 -8.48089993e-01
1.70370579e-01 -4.06858772e-01 -1.29703194e-01 3.41735333e-01
-3.22097838e-02 1.79217353e-01 -1.48287445e-01 4.30516213e-01
1.49728620e+00 -3.63997310e-01 9.52848315e-01 -1.14933671e-02
4.31950003e-01 1.52773947e-01 5.40596604e-01 2.60625631e-01
3.39883119e-02 6.66600883e-01 4.82818037e-01 -5.72960079e-01
-7.26261020e-01 -9.72845435e-01 -4.11801428e-01 3.53954792e-01
-4.10042763e-01 -9.79243100e-01 -4.22815710e-01 -9.05798614e-01
1.21847406e-01 6.93805575e-01 -9.68174994e-01 -4.28165853e-01
-3.40086490e-01 -6.39744699e-01 5.81936419e-01 8.22795510e-01
-2.62608051e-01 -1.33991361e+00 -8.75330567e-01 5.31680048e-01
-2.31008291e-01 -9.08744574e-01 -5.67044675e-01 2.92154670e-01
-9.69691575e-01 -1.36317217e+00 -9.77104545e-01 -6.98104024e-01
8.34679782e-01 -2.74919420e-01 1.01712549e+00 1.58522621e-01
-5.35221994e-01 3.65509778e-01 -7.37290561e-01 -6.34804010e-01
-7.01446414e-01 3.11423605e-03 -9.01695564e-02 -1.59118116e-01
5.63912809e-01 3.05492952e-02 -5.80539048e-01 -1.79853320e-01
-1.35754561e+00 -7.70343048e-03 4.98509377e-01 1.09996390e+00
4.19637710e-01 -3.28113109e-01 6.24817193e-01 -1.17216361e+00
9.66380239e-01 -7.91523993e-01 1.59143936e-02 3.62522542e-01
-7.48035431e-01 4.78726327e-01 9.27341282e-01 -2.80677229e-01
-3.69391054e-01 -1.45218372e-01 -3.38468373e-01 -4.38145936e-01
-1.35492366e-02 9.39434111e-01 2.37076923e-01 6.10475123e-01
7.39196062e-01 1.23497337e-01 -6.53182436e-03 -5.69816887e-01
5.46818674e-02 1.09151471e+00 1.06387893e-02 -2.45447546e-01
4.51732218e-01 3.35959457e-02 -2.47005165e-01 -5.27178407e-01
-5.37763655e-01 -7.81100929e-01 -5.32406330e-01 4.31763142e-01
1.09009850e+00 -7.14493692e-01 -2.87297577e-01 -1.48078784e-01
-1.09988439e+00 1.47899747e-01 -4.28687036e-01 7.54307270e-01
-2.69052535e-01 3.28655928e-01 -5.95247149e-01 -4.26820517e-01
-8.10837686e-01 -1.06725776e+00 7.93407083e-01 -2.31154189e-01
-5.34752905e-01 -1.04848456e+00 3.68487984e-01 -7.75453672e-02
2.82902509e-01 3.74713331e-01 1.64721882e+00 -1.05227447e+00
1.28543451e-01 -4.22183305e-01 -6.22188114e-02 4.46069270e-01
7.17838943e-01 -6.95431456e-02 -7.44594574e-01 -1.79562449e-01
-1.76158801e-01 9.37319845e-02 7.97276795e-01 1.29027188e-01
1.26129818e+00 -5.68096340e-01 -6.52592361e-01 6.05614841e-01
1.59623504e+00 4.78983492e-01 7.06042171e-01 4.02791500e-01
4.42915708e-01 6.60758495e-01 5.32497466e-01 4.96073335e-01
3.24447483e-01 3.97420615e-01 -3.38166803e-02 -9.16901231e-02
3.61358672e-01 -2.18540013e-01 2.51395583e-01 1.18902183e+00
1.11825399e-01 -2.26399511e-01 -1.18836725e+00 7.23677516e-01
-1.56538689e+00 -9.24704552e-01 1.96222588e-01 2.21105886e+00
1.18715823e+00 -1.54243425e-01 -2.19332516e-01 2.44422629e-01
1.72847450e-01 -3.83465081e-01 -3.67025621e-02 -5.40268481e-01
3.56318802e-01 5.24998307e-01 3.47980767e-01 8.07958096e-02
-7.46119797e-01 3.73169392e-01 5.71438503e+00 2.59302467e-01
-9.74870384e-01 3.58912088e-02 3.83520722e-01 3.69332656e-02
-4.03430313e-01 -3.82927865e-01 -6.21422887e-01 6.08456135e-01
1.33646345e+00 -4.58006203e-01 -2.11861908e-01 6.89619482e-01
3.56114432e-02 4.80381906e-01 -1.44536912e+00 1.16198039e+00
-1.10024899e-01 -1.68203378e+00 5.41968524e-01 -2.48196013e-02
3.98270547e-01 -2.15860575e-01 -1.26338273e-01 3.07144612e-01
5.26745617e-02 -1.24431241e+00 1.64934114e-01 7.61065066e-01
1.03087449e+00 -5.44639587e-01 1.29247773e+00 1.51416302e-01
-1.18584311e+00 -1.55422956e-01 -4.44721133e-01 3.74926776e-01
-3.79654653e-02 2.86668271e-01 -1.30625439e+00 8.65521491e-01
6.86873376e-01 8.99551868e-01 -4.61069375e-01 9.84782815e-01
1.51591823e-01 5.16265690e-01 1.03227362e-01 8.53019357e-02
1.62419245e-01 -3.02393711e-03 4.23132926e-02 1.55588591e+00
5.90570569e-01 3.85243654e-01 -5.26345819e-02 6.33143842e-01
-3.00425410e-01 6.67813957e-01 -9.99087512e-01 -6.15992785e-01
2.77636379e-01 1.10224950e+00 -3.25589865e-01 -6.84205115e-01
-5.41295528e-01 6.90363348e-01 -3.87159809e-02 1.17369354e-01
-6.06151223e-01 -6.53415442e-01 8.25064659e-01 4.55129355e-01
-2.86260508e-02 3.09504390e-01 -6.28615394e-02 -1.14000881e+00
-2.71947354e-01 -1.02350950e+00 8.59310150e-01 -5.71999848e-01
-1.41502130e+00 1.05550206e+00 -7.06524476e-02 -1.72203577e+00
-4.16184485e-01 -6.86911583e-01 -2.15497687e-01 1.00092208e+00
-1.41975713e+00 -7.21478999e-01 -1.82308584e-01 5.81737101e-01
3.57703328e-01 -3.66255790e-01 1.57842410e+00 4.71380889e-01
-2.42129043e-01 7.32335150e-01 2.06297159e-01 7.15565979e-01
9.13141608e-01 -1.17511833e+00 8.42843112e-03 1.17019817e-01
-2.90645864e-02 1.16733766e+00 3.34768534e-01 -6.13850057e-01
-1.08597505e+00 -1.37260544e+00 1.36474121e+00 -8.22837770e-01
3.31328601e-01 1.51980603e-02 -1.09715319e+00 7.44766951e-01
2.00690344e-01 -1.63220286e-01 1.37684846e+00 -3.15615982e-01
-3.72212768e-01 1.05099380e-01 -1.15509689e+00 3.48192155e-01
5.90579510e-01 -6.36703312e-01 -1.23718691e+00 3.44022036e-01
9.17333126e-01 -2.35121734e-02 -1.40114594e+00 3.89046907e-01
4.39736933e-01 -1.46877766e-01 1.00807607e+00 -1.22057223e+00
7.65656888e-01 -1.59357898e-02 -6.93056881e-02 -1.56690300e+00
-2.70190954e-01 -2.01196477e-01 9.12249163e-02 7.74803340e-01
6.39835775e-01 -7.72875190e-01 2.31723085e-01 6.03849471e-01
8.92588347e-02 -1.06854129e+00 -4.56098795e-01 -3.17278683e-01
1.70061797e-01 -1.99517887e-02 7.13839889e-01 1.19070590e+00
4.65632856e-01 1.96549848e-01 -8.65966082e-03 1.77884810e-02
1.30712628e-01 9.99798402e-02 3.98060411e-01 -1.38633168e+00
-1.17646754e-01 -1.88070863e-01 -6.92809343e-01 -1.91141188e-01
-2.11721376e-01 -1.43680799e+00 -3.11302394e-01 -2.03071189e+00
4.80497956e-01 -4.08648312e-01 -9.04195189e-01 7.37459421e-01
-2.96448827e-01 -1.36709988e-01 1.68587938e-01 9.43610594e-02
-4.84797433e-02 2.76558340e-01 9.89123225e-01 -3.84782374e-01
-3.21661890e-01 -9.20375958e-02 -8.57605815e-01 4.67590630e-01
6.86695516e-01 -9.82339025e-01 -2.61509061e-01 -2.15189219e-01
3.11419100e-01 3.63452733e-01 -6.74598366e-02 -7.61208057e-01
3.71325426e-02 -3.14666405e-02 2.81527340e-01 -3.27805370e-01
5.12892306e-02 -1.14079988e+00 2.55962759e-01 8.22119892e-01
-6.71304405e-01 4.51011509e-01 2.69839704e-01 2.86132962e-01
-4.63572204e-01 -3.49831194e-01 5.18637836e-01 -2.02320173e-01
-2.87249595e-01 2.85053998e-01 -3.57164055e-01 7.58648664e-02
9.64256644e-01 -4.79279645e-02 8.63729864e-02 -1.54477656e-01
-8.65913212e-01 -1.26554631e-03 2.60456353e-01 6.13057852e-01
9.87689912e-01 -1.15533817e+00 -8.12719405e-01 2.84975529e-01
6.14983380e-01 -3.33328277e-01 -5.73776253e-02 7.76162446e-01
-6.73682690e-01 6.96221769e-01 -2.70146042e-01 -2.48831391e-01
-1.40696788e+00 7.71707714e-01 2.91277975e-01 -5.59436977e-01
-9.56793368e-01 4.66155201e-01 2.59456992e-01 -5.06166399e-01
1.60640717e-01 -9.13079977e-01 -6.94142878e-01 2.72237480e-01
7.88421929e-01 6.35733688e-03 2.48229280e-01 -4.58815515e-01
-5.04891872e-01 5.41763484e-01 -1.47446558e-01 2.09819421e-01
1.74577439e+00 4.79840249e-01 -1.19092725e-01 3.53824079e-01
1.57009196e+00 4.46391404e-02 -1.88679025e-01 -2.41595313e-01
2.38569319e-01 -2.83844292e-01 -1.48190647e-01 -7.92850137e-01
-9.29182053e-01 9.80391860e-01 6.96038783e-01 -1.68621317e-01
1.02084732e+00 -9.49085951e-02 7.96410978e-01 4.52551842e-01
3.35894793e-01 -6.24063194e-01 -9.02871266e-02 3.09956133e-01
8.70321512e-01 -1.11124694e+00 -2.54447937e-01 -1.79196805e-01
-8.52665603e-01 1.34456563e+00 7.11160377e-02 2.51515388e-01
1.02121222e+00 3.36730748e-01 4.06547427e-01 -3.04331452e-01
-8.18133175e-01 1.87228203e-01 3.86671066e-01 4.44087148e-01
7.98770249e-01 2.02365205e-01 -5.00892818e-01 1.13048315e+00
-2.89485697e-02 2.21332669e-01 5.33404648e-01 9.65716958e-01
-4.16962206e-02 -1.32648551e+00 -1.88482717e-01 8.99012685e-01
-8.15455079e-01 -4.41325605e-01 -2.84804076e-01 4.95108455e-01
1.50511721e-02 7.04849899e-01 -2.92739093e-01 -5.53444564e-01
5.57068169e-01 5.10660768e-01 3.09524033e-02 -1.14779913e+00
-1.18104959e+00 -4.49542463e-01 -2.09479496e-01 -5.23207784e-01
-7.44313598e-02 -2.76521772e-01 -1.62540877e+00 1.21444628e-01
-8.45038444e-02 5.31856418e-01 3.55072647e-01 5.70047677e-01
6.09315276e-01 8.00071239e-01 3.79092872e-01 3.67561840e-02
-7.29541659e-01 -9.55714345e-01 -3.39513451e-01 7.66668260e-01
3.35179150e-01 -3.42382878e-01 -2.13057667e-01 2.32490584e-01] | [8.012188911437988, 7.045865058898926] |
924156f6-34b0-4594-a632-e09763dcc1fc | multi-target-normal-behaviour-models-for-wind | 2012.03074 | null | https://arxiv.org/abs/2012.03074v3 | https://arxiv.org/pdf/2012.03074v3.pdf | Multi-target normal behaviour models for wind farm condition monitoring | The trend towards larger wind turbines and remote locations of wind farms fuels the demand for automated condition monitoring strategies that can reduce the operating cost and avoid unplanned downtime. Normal behaviour modelling has been introduced to detect anomalous deviations from normal operation based on the turbine's SCADA data. A growing number of machine learning models of the normal behaviour of turbine subsystems are being developed by wind farm managers to this end. However, these models need to be kept track of, be maintained and require frequent updates. This research explores multi-target models as a new approach to capturing a wind turbine's normal behaviour. We present an overview of multi-target regression methods, motivate their application and benefits in wind turbine condition monitoring, and assess their performance in a wind farm case study. We find that multi-target models are advantageous in comparison to single-target modelling in that they can reduce the cost and effort of practical condition monitoring without compromising on the accuracy. We also outline some areas of future research. | ['Angela Meyer'] | 2020-12-05 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [-1.30393326e-01 -5.65640628e-01 7.15623200e-02 -1.94475830e-01
6.88406974e-02 -6.68584526e-01 2.68710732e-01 3.87866855e-01
1.94927782e-01 4.66534257e-01 -3.72434378e-01 -5.26282251e-01
-6.19379044e-01 -8.65168154e-01 2.47516096e-01 -9.79647040e-01
-5.07951140e-01 1.48272336e-01 1.97318017e-01 -2.35723719e-01
3.74792446e-03 7.40611255e-01 -1.76012683e+00 4.08724584e-02
6.35616362e-01 7.53266931e-01 8.56180191e-02 9.90193188e-01
5.69588304e-01 5.26841044e-01 -1.07740664e+00 4.64282542e-01
3.18801016e-01 -2.38465697e-01 -4.67767119e-01 1.11874580e-01
-2.82136679e-01 -4.94959414e-01 4.29284453e-01 8.28594744e-01
6.42467797e-01 2.39516720e-01 5.77398360e-01 -1.41131282e+00
2.69301713e-01 1.14105284e-01 -5.55028737e-01 8.29147637e-01
-1.88727316e-03 1.56171337e-01 7.59024501e-01 -2.67615288e-01
-2.24180564e-01 6.99376285e-01 8.84493709e-01 -4.83831465e-02
-1.35230207e+00 -3.27987075e-01 2.61361361e-01 -8.71334150e-02
-1.10814548e+00 -2.18956858e-01 5.67536592e-01 -5.37569284e-01
1.57770300e+00 4.72046673e-01 7.71018922e-01 4.51010257e-01
7.23883450e-01 2.01982036e-01 1.07827282e+00 -3.41813326e-01
4.88864243e-01 -2.12215975e-01 -2.11735919e-01 8.77746120e-02
7.02959239e-01 5.87351859e-01 -2.35871449e-01 -1.73750833e-01
5.99620461e-01 -2.19441846e-01 -3.17866892e-01 -4.91456300e-01
-6.17886126e-01 7.61944711e-01 -1.35042280e-01 5.36998272e-01
-6.48186862e-01 -6.45467022e-04 8.33644807e-01 8.72544408e-01
8.30870688e-01 7.29250669e-01 -1.27771330e+00 -4.16102767e-01
-1.08972573e+00 1.62618518e-01 7.74457693e-01 4.93546605e-01
9.03026983e-02 9.75386739e-01 4.50695723e-01 4.51553255e-01
3.01909029e-01 5.65404117e-01 6.90223694e-01 -4.98195916e-01
-7.83861503e-02 2.22764850e-01 5.16229033e-01 -8.44018698e-01
-5.96951544e-01 -6.61396623e-01 -7.69738674e-01 6.85789645e-01
5.26313446e-02 -5.19564807e-01 -4.70575035e-01 9.38394189e-01
2.92821795e-01 1.97137907e-01 -6.69106022e-02 3.13674480e-01
-2.57209748e-01 6.29694045e-01 -6.66516572e-02 -8.10627580e-01
7.29847789e-01 -4.80933428e-01 -9.44902658e-01 -6.15507700e-02
1.13468099e+00 -8.31470847e-01 4.56017315e-01 8.97612929e-01
-7.49418199e-01 -6.80750906e-01 -1.10577536e+00 1.03199756e+00
-6.35308743e-01 2.44835779e-01 3.49054098e-01 1.02898443e+00
-9.76176143e-01 1.08370495e+00 -1.51934695e+00 -2.97509462e-01
-7.86150917e-02 2.18977585e-01 8.52251332e-03 6.09885275e-01
-8.27709019e-01 1.56420302e+00 4.54988390e-01 5.74995875e-01
-8.43657911e-01 -7.81501174e-01 -7.82666028e-01 8.99975002e-02
1.76228862e-02 -2.43229672e-01 1.67629695e+00 -3.90221447e-01
-1.35846770e+00 1.08306959e-01 1.66197360e-01 -6.03858769e-01
1.50650054e-01 -5.87592840e-01 -1.11477101e+00 -6.07779473e-02
3.94240469e-02 -4.39071685e-01 1.07924807e+00 -8.58896494e-01
-1.24132395e+00 -3.23422849e-02 -4.35988933e-01 -5.29177561e-02
-6.96121529e-02 1.72054514e-01 8.49425912e-01 -3.71622860e-01
-1.16246149e-01 -5.52238822e-01 -4.76527780e-01 -5.74190676e-01
1.28520787e-01 -4.81389873e-02 1.44863164e+00 -6.40034735e-01
1.73398161e+00 -1.80974102e+00 -3.45755488e-01 2.02439919e-01
-3.02985311e-01 7.16530025e-01 2.23517165e-01 8.94060731e-01
-5.18320322e-01 1.94630697e-01 1.93957195e-01 2.18086630e-01
-2.70266086e-01 5.56698143e-01 -4.40265507e-01 6.84934676e-01
4.64491129e-01 1.34700924e-01 -8.56471837e-01 4.11931962e-01
9.35049772e-01 -6.58573061e-02 3.13020833e-02 2.50765830e-01
-5.86629137e-02 2.47906744e-01 -3.94391775e-01 4.96061891e-01
5.70563376e-01 3.70207101e-01 1.05990410e-01 -4.47695376e-03
-4.63017493e-01 1.66123778e-01 -1.21860397e+00 6.58385515e-01
-8.52885187e-01 3.37760627e-01 2.86063045e-01 -1.46188152e+00
9.99937296e-01 1.02487767e+00 8.04708540e-01 -1.86424524e-01
4.80093956e-02 8.06104839e-02 3.40923965e-01 -7.07968771e-01
2.14281604e-01 -3.13390940e-01 1.62210047e-01 4.37230229e-01
-1.43201426e-01 -4.97072369e-01 3.48059773e-01 -4.93892252e-01
1.08341479e+00 3.66306156e-01 8.76312256e-01 -6.55741572e-01
4.20183957e-01 2.01717809e-01 7.72189319e-01 2.08540216e-01
-1.39236495e-01 -2.05331609e-01 3.48408282e-01 -6.80775464e-01
-7.65311182e-01 -7.97728419e-01 -2.66862482e-01 6.72639012e-01
-2.86975503e-01 -5.61915517e-01 -1.47866458e-01 -3.98626685e-01
1.91011086e-01 1.03723836e+00 -3.59973371e-01 -2.61519998e-01
-3.97905916e-01 -1.14169621e+00 1.81937754e-01 7.81860054e-01
1.74699560e-01 -6.94266200e-01 -1.35793853e+00 5.69500804e-01
3.75093013e-01 -7.57447064e-01 2.96998531e-01 7.72293091e-01
-1.38173342e+00 -1.15384221e+00 8.30969512e-02 -2.50549465e-01
5.12568533e-01 4.41712111e-01 1.13657343e+00 2.37465590e-01
-3.66179109e-01 2.59106338e-01 -7.06499100e-01 -8.00987184e-01
-5.44822931e-01 -1.50227934e-01 5.90648174e-01 -5.65014720e-01
3.00670654e-01 -7.63090372e-01 -1.76062167e-01 6.15062416e-01
-6.38772905e-01 -6.73047960e-01 4.73766923e-01 1.05901194e+00
2.71367878e-01 1.29509687e+00 1.05275726e+00 -6.10501111e-01
6.78523540e-01 -6.79245591e-01 -1.08272398e+00 1.64958194e-01
-1.43110096e+00 -3.03917050e-01 9.41609621e-01 -3.03979993e-01
-1.16960800e+00 -1.96363479e-02 2.29449138e-01 -2.62616664e-01
-6.40715718e-01 8.87499392e-01 1.30895138e-01 1.83146730e-01
4.40530747e-01 -2.27217942e-01 1.37470782e-01 -7.82208681e-01
2.17946284e-02 6.94025934e-01 2.66560525e-01 -2.63118893e-02
1.18335116e+00 5.45670614e-02 2.22591206e-01 -1.23025525e+00
-5.45688689e-01 -9.50357914e-01 -9.94526505e-01 -3.20181727e-01
2.31670648e-01 -9.75366592e-01 -3.55201364e-01 5.47930896e-01
-4.49541628e-01 -4.99835134e-01 -3.48500103e-01 5.95117331e-01
-3.43227357e-01 2.53249615e-01 -3.39562386e-01 -1.19724560e+00
-3.91571134e-01 -5.85900247e-01 7.71882117e-01 1.15924701e-01
-4.83210444e-01 -1.56666398e+00 2.40481570e-01 -2.40072802e-01
6.69440627e-01 5.36004841e-01 7.22939432e-01 -5.79626083e-01
1.45454764e-01 -8.21879864e-01 4.55857992e-01 7.14266598e-01
8.93249035e-01 5.29265583e-01 -9.37882900e-01 -8.18842411e-01
4.18680549e-01 1.39031529e-01 -5.90760224e-02 4.62559193e-01
3.70951146e-01 -1.19874880e-01 -3.47195178e-01 1.66022316e-01
1.44801700e+00 3.62051517e-01 2.89312415e-02 4.69680905e-01
-4.52536866e-02 4.96603459e-01 1.49137151e+00 6.57989323e-01
-3.74362379e-01 4.52883452e-01 6.48881674e-01 -2.73205101e-01
5.70504129e-01 2.15445220e-01 6.79818869e-01 8.27325046e-01
-2.06068158e-01 -1.72286302e-01 -6.66514218e-01 8.14130247e-01
-1.67026019e+00 -1.15155554e+00 -5.72349191e-01 2.18681979e+00
1.72343865e-01 2.10082829e-01 1.24836691e-01 6.76276147e-01
5.62376738e-01 9.33458582e-02 -2.44354546e-01 -9.30526376e-01
2.82448709e-01 4.07540888e-01 4.91657645e-01 2.48349324e-01
-1.32935071e+00 5.23820281e-01 6.96507311e+00 4.02688324e-01
-1.14718211e+00 -8.89782459e-02 7.80397952e-02 -1.12060696e-01
5.29032350e-01 -2.50072628e-02 -5.45624137e-01 3.24459463e-01
1.38638031e+00 -5.05209982e-01 -1.78836491e-02 1.22666049e+00
1.10578275e+00 -5.61148405e-01 -8.63463163e-01 1.87940866e-01
-4.77613688e-01 -5.65034747e-01 -6.10664546e-01 1.79692402e-01
7.21621275e-01 9.82765853e-02 -6.28520727e-01 2.13531375e-01
3.22081834e-01 -7.02425718e-01 3.07947427e-01 1.46789402e-01
3.61182481e-01 -7.38377690e-01 1.37601900e+00 3.98460835e-01
-1.64003611e+00 -3.58686537e-01 -3.91033858e-01 -8.26347351e-01
6.41241372e-01 8.44713151e-01 -1.05418873e+00 1.14120960e+00
1.10821867e+00 8.86156082e-01 -2.70472020e-01 1.24311972e+00
-1.42289087e-01 1.10135484e+00 -5.66350937e-01 1.67235672e-01
5.14985137e-02 -2.60069937e-01 4.28169698e-01 7.05254316e-01
4.75321740e-01 -4.01559055e-01 2.89199829e-01 2.60310352e-01
1.12086773e+00 -2.94506580e-01 -9.66963351e-01 1.07863940e-01
6.11491382e-01 1.47916317e+00 -8.57903302e-01 6.54592961e-02
-3.99621189e-01 5.14381468e-01 -5.26096940e-01 4.45792079e-03
-5.33255160e-01 -3.30560446e-01 9.58155870e-01 -6.97300658e-02
5.43260396e-01 -4.88287419e-01 -3.99226755e-01 -5.01673937e-01
1.14077087e-02 -6.96181178e-01 4.38764304e-01 -6.65919125e-01
-1.23641777e+00 2.47956023e-01 3.25061440e-01 -1.68934906e+00
-8.45065176e-01 -6.77839041e-01 -1.28564763e+00 9.33221102e-01
-1.43049443e+00 -7.53943920e-01 -3.89613807e-02 -9.08995420e-02
7.75409460e-01 -1.08773343e-01 1.38024783e+00 -8.58475715e-02
-7.79036343e-01 -1.49860710e-01 3.67521882e-01 -2.62532949e-01
2.81714886e-01 -1.44143522e+00 5.54010451e-01 1.23871386e+00
-1.15378246e-01 2.05829158e-01 1.07353544e+00 -8.49600673e-01
-1.08544207e+00 -1.20709682e+00 5.20796537e-01 -5.47273993e-01
1.04569447e+00 2.43910745e-01 -7.85379946e-01 6.51953340e-01
2.18152925e-01 -1.40696287e-01 8.36439967e-01 2.23063856e-01
3.56625259e-01 -3.14004034e-01 -1.15491009e+00 3.43235023e-02
2.10038751e-01 -9.15361047e-02 -6.39247954e-01 1.58424675e-01
-1.10452406e-01 -1.40688419e-01 -1.46556747e+00 5.52982628e-01
2.53277898e-01 -8.34691823e-01 6.12726033e-01 -6.67281330e-01
-2.79434741e-01 -7.03387320e-01 2.82238275e-01 -2.17561793e+00
-5.12787402e-01 -9.20315325e-01 -3.03012818e-01 1.22175074e+00
3.42098475e-01 -7.90171206e-01 2.80382454e-01 2.37540916e-01
-3.91225785e-01 -7.03677058e-01 -1.02915931e+00 -1.39662695e+00
1.06227992e-04 -2.92624772e-01 6.22233272e-01 1.03195882e+00
2.34938547e-01 2.98850030e-01 -3.87478262e-01 9.30230796e-01
1.16628438e-01 5.86359128e-02 4.63997811e-01 -1.43524170e+00
2.15626191e-02 -1.57690540e-01 -5.43791771e-01 -2.39225507e-01
-3.30934376e-01 5.93642239e-03 1.01805143e-01 -1.28442383e+00
-9.33757484e-01 -1.87088549e-01 -1.13435134e-01 5.35106122e-01
-1.26369158e-02 -2.67133474e-01 -3.92646760e-01 6.89319596e-02
2.16769293e-01 4.18547451e-01 5.46788931e-01 2.59334534e-01
-2.96410948e-01 8.26850295e-01 -2.31164992e-01 7.43726790e-01
1.45959294e+00 -3.31252098e-01 -7.86228418e-01 -2.69397676e-01
1.56000718e-01 3.36301811e-02 4.71716281e-03 -1.10168111e+00
1.30563583e-02 -3.05999041e-01 4.56627041e-01 -7.98453808e-01
-1.18032835e-01 -1.23855269e+00 3.64368528e-01 9.57868516e-01
3.06288004e-01 8.31087351e-01 4.34380352e-01 6.46907270e-01
-2.58603811e-01 -4.87691045e-01 6.39915407e-01 1.18231200e-01
-6.46413088e-01 -1.41408846e-01 -1.15502691e+00 -5.42762160e-01
1.51200378e+00 -2.64132649e-01 -2.41731599e-01 -1.02892049e-01
-7.44037807e-01 5.41969776e-01 1.36847362e-01 5.11888325e-01
1.95129395e-01 -6.78287029e-01 -5.03909469e-01 4.73129839e-01
-2.36620560e-01 -6.85577327e-03 2.13172093e-01 9.48027849e-01
-7.12216735e-01 6.24651015e-01 -2.77526289e-01 -6.01960242e-01
-1.36760950e+00 5.94979048e-01 7.56727934e-01 -4.13092494e-01
-5.85147142e-01 3.66095155e-01 -6.57299101e-01 -1.20653763e-01
-3.48982632e-01 -6.28256619e-01 -2.72161394e-01 1.87870026e-01
5.10966837e-01 8.66663694e-01 6.98498309e-01 -2.89039224e-01
-1.80491358e-01 3.52855682e-01 6.46150038e-02 4.14246589e-01
1.35261440e+00 -2.88552999e-01 -1.21291965e-01 7.16156304e-01
3.71099561e-01 -1.94953799e-01 -1.07432675e+00 4.73961532e-01
4.17920619e-01 -4.99376744e-01 4.71543163e-01 -9.27484512e-01
-1.07092047e+00 6.00025296e-01 8.38939846e-01 1.03166032e+00
1.37223971e+00 -7.58120775e-01 2.26284295e-01 4.09104228e-01
4.63371545e-01 -1.31471086e+00 -4.15440291e-01 2.81862497e-01
5.68468571e-01 -9.13962424e-01 2.55997777e-01 -4.05940473e-01
-4.08046067e-01 1.25227690e+00 6.83589637e-01 -2.66712531e-02
1.15834188e+00 6.89341307e-01 3.95738214e-01 -2.67791152e-01
-1.25062013e+00 -1.70684382e-01 -4.74639297e-01 1.39415812e+00
4.38136101e-01 3.54381263e-01 -4.81349938e-02 2.59272337e-01
-2.08895072e-01 -1.09386984e-02 6.74629748e-01 1.24620795e+00
-5.11928558e-01 -1.22376299e+00 -2.92051047e-01 9.64288116e-01
-5.03603518e-01 1.94205359e-01 8.89875442e-02 9.71386850e-01
6.09661005e-02 1.63136125e+00 1.15850694e-01 -4.87045288e-01
5.22140622e-01 2.08646595e-01 -1.69094309e-01 -8.95165622e-01
-6.66436017e-01 1.97666898e-01 4.95865285e-01 -3.34650576e-01
-4.58852917e-01 -1.05334163e+00 -7.93835044e-01 -1.77770883e-01
-1.02960432e+00 4.36072916e-01 6.01855814e-01 1.01465297e+00
1.12223253e-01 8.17668557e-01 1.21768332e+00 -7.14000881e-01
-1.06554139e+00 -1.31725156e+00 -1.01269758e+00 -2.81921893e-01
3.14309508e-01 -9.12394643e-01 -7.44799256e-01 1.86220586e-01] | [6.477437973022461, 2.507061719894409] |
c3a00751-7dda-4137-9a69-9c2d3210a655 | studyformer-attention-based-and-dynamic-multi | 2302.11840 | null | https://arxiv.org/abs/2302.11840v1 | https://arxiv.org/pdf/2302.11840v1.pdf | StudyFormer : Attention-Based and Dynamic Multi View Classifier for X-ray images | Chest X-ray images are commonly used in medical diagnosis, and AI models have been developed to assist with the interpretation of these images. However, many of these models rely on information from a single view of the X-ray, while multiple views may be available. In this work, we propose a novel approach for combining information from multiple views to improve the performance of X-ray image classification. Our approach is based on the use of a convolutional neural network to extract feature maps from each view, followed by an attention mechanism implemented using a Vision Transformer. The resulting model is able to perform multi-label classification on 41 labels and outperforms both single-view models and traditional multi-view classification architectures. We demonstrate the effectiveness of our approach through experiments on a dataset of 363,000 X-ray images. | ['Andre Dourson', 'Diane Wilson', 'Michael Fitzke', 'Lucas Wannenmacher'] | 2023-02-23 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 3.08906794e-01 1.03477947e-02 -2.41199657e-01 -6.78214252e-01
-1.14205861e+00 -4.15873051e-01 6.32772923e-01 2.99249977e-01
-2.31537104e-01 1.07962973e-01 5.49079441e-02 -2.62495428e-01
-1.19639434e-01 -7.86710799e-01 -6.62429631e-01 -5.05045116e-01
4.30039287e-01 6.12127244e-01 1.78459167e-01 1.74897805e-01
9.80335381e-03 4.77090240e-01 -1.22481167e+00 8.00088227e-01
3.04098040e-01 1.24372494e+00 3.82464439e-01 8.92365515e-01
1.13237649e-01 9.92038548e-01 -3.91399980e-01 -2.53123462e-01
-7.00616231e-03 -4.86537993e-01 -8.95773947e-01 4.72922772e-01
5.65015435e-01 -6.44809425e-01 -3.59972388e-01 7.13654041e-01
5.55050969e-01 -7.87574798e-02 6.51772499e-01 -7.76231170e-01
-6.33267283e-01 3.10132504e-01 -6.10260189e-01 5.26263118e-01
3.79417211e-01 -1.60435632e-01 1.03145325e+00 -9.43276644e-01
4.57897335e-01 9.56934273e-01 5.38151979e-01 5.33063054e-01
-9.90492105e-01 -2.06380546e-01 6.54935464e-02 3.03128421e-01
-8.43686342e-01 -2.66196281e-01 7.21271515e-01 -3.77268881e-01
9.37096894e-01 1.89436480e-01 6.75511003e-01 1.03379226e+00
5.40280223e-01 7.68243968e-01 1.46372354e+00 -4.94959772e-01
2.80007496e-02 1.48322657e-01 2.47926906e-01 1.09520376e+00
-8.01561400e-02 6.99990662e-03 -3.43680501e-01 -4.86077160e-01
6.49547637e-01 6.25578642e-01 -1.93204805e-01 -3.77117515e-01
-1.03871965e+00 8.25504243e-01 7.03318596e-01 4.33601320e-01
-6.64816082e-01 1.32156864e-01 3.51134688e-01 8.35099518e-02
5.72403729e-01 3.84504259e-01 -1.96468815e-01 2.49951109e-01
-6.11313343e-01 -8.08910429e-02 3.58948886e-01 4.22730654e-01
2.33827651e-01 -5.09734809e-01 6.86797407e-03 1.01997530e+00
2.78783977e-01 4.02961224e-01 6.37908399e-01 -7.65812874e-01
4.17452484e-01 8.20403099e-01 -4.35038596e-01 -6.65199280e-01
-8.34460497e-01 -2.73592800e-01 -5.43206513e-01 2.63177216e-01
1.86153054e-01 1.27922416e-01 -1.25383544e+00 1.22049892e+00
1.70415789e-01 -2.17885539e-01 1.20471623e-02 8.62739801e-01
9.01428103e-01 2.40037486e-01 4.80161645e-02 1.44567966e-01
1.60314643e+00 -1.14596236e+00 -4.78389889e-01 -9.37021673e-02
4.80084062e-01 -5.76762438e-01 8.59082937e-01 5.73388159e-01
-1.33487701e+00 -5.80791891e-01 -1.08348393e+00 1.22121610e-02
-2.29040414e-01 4.20552731e-01 3.35723728e-01 4.99518752e-01
-7.85968244e-01 4.29057598e-01 -1.17567289e+00 -3.83503228e-01
5.63716054e-01 4.24762934e-01 -5.08162856e-01 -4.46963251e-01
-5.91790438e-01 1.00266969e+00 1.26367047e-01 -1.47005603e-01
-6.79742157e-01 -5.55820584e-01 -5.39245725e-01 2.78296947e-01
3.55431020e-01 -9.14978862e-01 1.26216233e+00 -8.08923662e-01
-1.29932725e+00 8.82353067e-01 2.65329685e-02 -8.21011066e-02
3.13598394e-01 -3.34631085e-01 -3.63821208e-01 8.96380067e-01
2.34439988e-02 4.90275741e-01 6.30425155e-01 -1.49843860e+00
-7.83604085e-01 -7.09754407e-01 4.15355653e-01 1.72938108e-01
-2.45023653e-01 7.76918828e-02 -5.14447033e-01 -4.55328643e-01
3.90839428e-01 -1.13916600e+00 -1.63519472e-01 -2.37003770e-02
-4.01284397e-01 -1.36612818e-01 7.32758880e-01 -6.51308477e-01
7.15383232e-01 -2.14036012e+00 1.49880439e-01 2.00799480e-01
5.54729879e-01 2.24019233e-02 3.20050329e-01 2.92030007e-01
-2.57835567e-01 -1.79443121e-01 -1.55417249e-01 -2.75885373e-01
-4.35189664e-01 2.32526436e-01 -8.72710273e-02 4.02066290e-01
7.88397864e-02 8.02233577e-01 -5.87255120e-01 -5.05336821e-01
3.91152561e-01 4.51525539e-01 -3.85668367e-01 2.64214426e-01
-5.06175123e-02 4.71018821e-01 -5.14378905e-01 6.43617034e-01
1.39013782e-01 -9.67787385e-01 3.04051191e-01 -2.00984880e-01
3.62675339e-01 2.60242820e-01 -5.41484177e-01 1.73730528e+00
-7.99360693e-01 3.20558459e-01 -9.44747478e-02 -8.69216383e-01
3.80073518e-01 7.15871811e-01 7.77128041e-01 -4.84687388e-01
2.88669616e-01 -2.71377619e-02 -3.45856994e-02 -7.08867252e-01
-1.40109509e-01 -2.96953589e-01 1.67928323e-01 1.03539312e+00
1.03456810e-01 1.61942386e-03 1.94040546e-03 5.28053455e-02
1.26661921e+00 -1.65875778e-01 4.10620838e-01 2.90217549e-01
6.12798870e-01 -9.35836136e-02 -6.59540892e-02 7.63883412e-01
2.18120106e-02 9.59899545e-01 2.17677817e-01 -8.98591042e-01
-8.40829194e-01 -9.69402969e-01 -2.83608615e-01 1.04550540e+00
2.64779832e-02 -2.84311086e-01 -6.13184988e-01 -1.27608836e+00
-1.40731364e-01 4.95728195e-01 -8.32321286e-01 -6.21290728e-02
-7.15998948e-01 -7.13424385e-01 3.84954333e-01 9.23302472e-01
3.77587736e-01 -9.37851727e-01 -1.06684041e+00 8.88893232e-02
-1.93486750e-01 -1.17218566e+00 -2.76675552e-01 3.47536713e-01
-9.46546495e-01 -1.39108431e+00 -6.64996028e-01 -5.51733732e-01
8.91330302e-01 3.96102339e-01 1.12944877e+00 3.91026199e-01
-4.46583718e-01 8.48184049e-01 -2.81438231e-01 -2.88264871e-01
-5.93908012e-01 1.80765018e-01 -2.40142778e-01 1.01459749e-01
2.48823151e-01 -3.75464559e-01 -6.65352404e-01 1.65274635e-01
-9.40166056e-01 1.02111839e-01 7.73107588e-01 1.06127584e+00
6.51824117e-01 -3.97091955e-01 4.43394482e-01 -1.29157758e+00
3.37130100e-01 -4.46902543e-01 -2.68031299e-01 5.61380029e-01
-6.57680690e-01 -1.53008133e-01 6.93734586e-01 -1.22748695e-01
-9.67228413e-01 1.51139200e-01 -2.88641244e-01 -6.83244288e-01
-5.28436720e-01 4.42882806e-01 2.61665612e-01 -2.19968036e-01
4.82781023e-01 1.21171765e-01 -7.88008347e-02 -5.17680705e-01
2.02621877e-01 8.16181004e-01 2.85272598e-01 -1.38232261e-01
1.02774881e-01 6.59917772e-01 3.25439535e-02 -5.50590932e-01
-1.16777587e+00 -4.91244793e-01 -7.23784447e-01 -3.58557403e-01
1.15608633e+00 -7.67391324e-01 -4.54361707e-01 2.47216672e-01
-8.64130139e-01 2.25515127e-01 5.23896813e-02 8.42103541e-01
-6.89062297e-01 2.35532939e-01 -8.70479107e-01 -2.80245870e-01
-3.73807281e-01 -1.49478185e+00 1.20513070e+00 -6.03642091e-02
-1.39323762e-02 -9.37499523e-01 4.43902053e-02 8.28792453e-01
1.82875082e-01 -6.99846223e-02 1.35412991e+00 -8.13618004e-01
-4.20692593e-01 -2.86667943e-01 -2.03780517e-01 2.48830229e-01
2.71445394e-01 -4.14957672e-01 -1.25776684e+00 -3.50503385e-01
4.75308567e-01 -5.27290642e-01 1.02059019e+00 4.77258146e-01
1.38431704e+00 1.36504263e-01 -5.72089016e-01 4.94289666e-01
1.49087846e+00 4.01827574e-01 2.33690500e-01 1.57288879e-01
8.23866606e-01 4.17310655e-01 2.41403729e-01 2.13860914e-01
4.91956562e-01 6.27143323e-01 5.05032480e-01 -3.09026629e-01
-9.49986503e-02 1.55407682e-01 -2.25551575e-01 7.52604663e-01
-2.01822102e-01 -2.73821592e-01 -1.08018541e+00 2.31538549e-01
-1.44117355e+00 -7.05097139e-01 1.32574305e-01 1.67697048e+00
2.72267580e-01 8.51790309e-02 -4.53031957e-02 1.35502264e-01
4.74065304e-01 1.83588862e-02 -5.44756114e-01 -7.52615109e-02
4.39435989e-01 2.59568751e-01 3.13558996e-01 1.69412538e-01
-1.31448090e+00 1.66338772e-01 7.43738747e+00 6.90656677e-02
-1.40513527e+00 3.24813396e-01 7.23784685e-01 -1.84930667e-01
-6.46940321e-02 -3.96125227e-01 -3.52905869e-01 1.50938258e-01
5.91875017e-01 3.94009501e-01 5.64424172e-02 9.01728809e-01
-3.58257771e-01 -2.30818674e-01 -1.28763247e+00 9.51087058e-01
6.11436903e-01 -1.24461973e+00 1.18805123e-02 7.01369569e-02
5.47040343e-01 2.75000215e-01 1.97885573e-01 -1.97358847e-01
4.94958945e-02 -8.14405501e-01 3.12760293e-01 4.57815200e-01
5.94013691e-01 -6.28415525e-01 7.03944921e-01 3.21649790e-01
-7.81022310e-01 -2.84020871e-01 7.03842267e-02 3.69126499e-01
6.61017150e-02 1.49772391e-01 -9.53482330e-01 7.51824796e-01
7.08270788e-01 5.36548257e-01 -7.17535436e-01 7.15494931e-01
3.89366746e-02 5.03010988e-01 -1.21060364e-01 3.15975517e-01
3.15564513e-01 4.43384089e-02 1.55657053e-01 8.06682169e-01
1.99355006e-01 4.54202861e-01 3.96459371e-01 3.85689408e-01
2.17309203e-02 4.61895764e-02 -7.00881422e-01 2.79548377e-01
-1.48371726e-01 1.24591005e+00 -9.88549888e-01 -6.88181221e-01
-1.04289448e+00 7.95987487e-01 3.41911614e-01 1.21716082e-01
-7.83275783e-01 6.36227131e-02 -6.56805784e-02 3.30882221e-02
4.58571076e-01 1.25262022e-01 -2.26076424e-01 -1.11010003e+00
-2.18693423e-03 -8.92163992e-01 6.19536996e-01 -9.85977292e-01
-1.23873150e+00 9.86919582e-01 6.51526451e-03 -1.20021725e+00
-3.73036712e-01 -7.04817116e-01 -3.00910175e-01 5.90421736e-01
-1.35088384e+00 -1.39127564e+00 -3.82849842e-01 4.57794368e-01
6.11650825e-01 -3.57912928e-01 1.08476496e+00 4.57858324e-01
-2.54975438e-01 2.70035863e-01 3.01982882e-03 4.16281186e-02
6.17428720e-01 -1.47044098e+00 1.36798754e-01 3.47325623e-01
4.77781713e-01 3.85777563e-01 1.72585517e-01 -2.88752258e-01
-1.28053534e+00 -9.25324500e-01 5.37344992e-01 -6.59920037e-01
1.01670437e-01 1.54592261e-01 -9.40836906e-01 9.02849555e-01
3.40587914e-01 3.17942798e-01 1.13725996e+00 5.13424724e-02
-3.18620920e-01 8.45345855e-02 -9.64083135e-01 3.50508513e-03
5.40354609e-01 -6.50178611e-01 -7.38443673e-01 4.45339829e-01
8.73891413e-02 -6.27081156e-01 -1.09007502e+00 2.96383351e-01
5.90404570e-01 -1.14531648e+00 1.09410453e+00 -5.92082977e-01
7.00397432e-01 1.52648240e-01 -1.97802395e-01 -1.26967919e+00
-3.02833974e-01 4.31359887e-01 -3.86484130e-03 5.02588093e-01
3.25536400e-01 -5.46525955e-01 5.92352331e-01 2.95047790e-01
-2.39568740e-01 -1.10545695e+00 -7.48493493e-01 -1.40443787e-01
-2.05797300e-01 -1.85874268e-01 3.36047560e-01 8.37910771e-01
-1.45309567e-01 5.19557714e-01 -2.55701095e-02 2.39534974e-01
3.73916805e-01 5.57163537e-01 3.40858310e-01 -1.12655079e+00
-7.66479969e-01 -2.24791408e-01 -3.72169524e-01 -6.63751304e-01
-1.12057906e-02 -1.01145482e+00 -3.13891560e-01 -1.72456086e+00
5.93151748e-01 -4.73930418e-01 -4.97343391e-01 5.41549742e-01
-3.03450942e-01 5.85764885e-01 2.40531519e-01 3.07704002e-01
-5.68829060e-01 1.51062146e-01 1.22332883e+00 -1.85702085e-01
1.07141681e-01 2.21325651e-01 -4.99542981e-01 1.12102425e+00
6.09401107e-01 -5.82138598e-01 -3.89290482e-01 -6.30699933e-01
-1.33013390e-02 4.92102206e-01 3.88217628e-01 -1.12751484e+00
1.30353734e-01 2.47094765e-01 7.89911389e-01 -8.58755648e-01
6.96370661e-01 -9.92801964e-01 -6.48664683e-02 6.09816372e-01
-5.26087224e-01 3.92698079e-01 -9.45496261e-02 7.15642691e-01
-3.08739185e-01 -3.03239554e-01 7.37080693e-01 -6.50477290e-01
-2.06924766e-01 8.56324732e-02 -2.40105048e-01 -2.39854246e-01
1.20298827e+00 1.85293734e-01 -2.27639481e-01 -5.84719479e-02
-1.07795000e+00 1.05780615e-02 3.04495275e-01 3.88285816e-01
7.38965154e-01 -1.21900225e+00 -4.38793391e-01 3.38001490e-01
3.59913349e-01 -2.20066428e-01 1.22596346e-01 7.91287124e-01
-6.04357302e-01 4.89484131e-01 -2.73118228e-01 -8.88510287e-01
-1.55440044e+00 6.91964746e-01 4.36969876e-01 -4.95974451e-01
-1.03314829e+00 4.46826071e-01 5.21082759e-01 -2.53056377e-01
3.60146388e-02 -3.83493066e-01 -4.53750223e-01 -4.94340770e-02
5.71664989e-01 1.44668534e-01 4.70879018e-01 -6.41704261e-01
-1.92906931e-01 7.85706520e-01 -4.65535223e-01 -2.47077286e-01
1.49432790e+00 -1.94819465e-01 3.65045480e-02 7.76927233e-01
1.17573845e+00 -3.01007748e-01 -7.83743739e-01 -3.21100682e-01
-2.10415095e-01 -3.36676508e-01 1.97624698e-01 -8.45178068e-01
-1.16579151e+00 1.12081969e+00 7.69812524e-01 3.13003093e-01
1.17796087e+00 3.28881621e-01 7.79178858e-01 4.56688106e-01
1.93301618e-01 -7.33427763e-01 4.12729055e-01 6.71969354e-02
6.46419287e-01 -1.40565038e+00 1.44623592e-01 -2.94716299e-01
-6.71523154e-01 1.34729159e+00 5.89160264e-01 -6.62271380e-02
6.48824573e-01 -1.33631080e-02 4.59142387e-01 -7.81471848e-01
-9.95939851e-01 -1.58040356e-02 2.55178422e-01 1.50935352e-01
5.14882565e-01 -1.34359545e-03 1.22966416e-01 3.04140061e-01
3.71891469e-01 -2.18179030e-03 3.32278401e-01 1.21532357e+00
-2.53847927e-01 -1.10485458e+00 -4.62353468e-01 8.86326909e-01
-1.00633025e+00 1.52166933e-01 -4.02450472e-01 7.08736956e-01
5.56334220e-02 7.54144609e-01 1.64403617e-02 -1.98408306e-01
3.15970957e-01 3.54333669e-01 7.18174636e-01 -7.21582234e-01
-8.31159890e-01 2.50878274e-01 -1.26996562e-01 -5.14000654e-01
-6.66780353e-01 -5.97791195e-01 -1.02632117e+00 2.31755361e-01
-4.05733347e-01 -2.74840683e-01 6.39745414e-01 1.07575870e+00
1.15839072e-01 9.68421638e-01 8.53013098e-01 -4.05391991e-01
-5.77662468e-01 -7.51269460e-01 -4.64920104e-01 4.20818806e-01
5.37537277e-01 -7.18645811e-01 1.10771880e-02 1.84491396e-01] | [15.13623046875, -1.8959739208221436] |
45720039-8df8-46bd-b55d-aa2953a04d0b | semantic-segmentation-assisted-scene | 2109.11453 | null | https://arxiv.org/abs/2109.11453v1 | https://arxiv.org/pdf/2109.11453v1.pdf | Semantic Segmentation-assisted Scene Completion for LiDAR Point Clouds | Outdoor scene completion is a challenging issue in 3D scene understanding, which plays an important role in intelligent robotics and autonomous driving. Due to the sparsity of LiDAR acquisition, it is far more complex for 3D scene completion and semantic segmentation. Since semantic features can provide constraints and semantic priors for completion tasks, the relationship between them is worth exploring. Therefore, we propose an end-to-end semantic segmentation-assisted scene completion network, including a 2D completion branch and a 3D semantic segmentation branch. Specifically, the network takes a raw point cloud as input, and merges the features from the segmentation branch into the completion branch hierarchically to provide semantic information. By adopting BEV representation and 3D sparse convolution, we can benefit from the lower operand while maintaining effective expression. Besides, the decoder of the segmentation branch is used as an auxiliary, which can be discarded in the inference stage to save computational consumption. Extensive experiments demonstrate that our method achieves competitive performance on SemanticKITTI dataset with low latency. Code and models will be released at https://github.com/jokester-zzz/SSA-SC. | ['Hongbo Zhang', 'Feng Wen', 'Wanlong Li', 'Yong liu', 'Tianxin Huang', 'Xin Kong', 'Hao Zou', 'Xuemeng Yang'] | 2021-09-23 | null | null | null | null | ['3d-semantic-scene-completion'] | ['computer-vision'] | [ 2.36176923e-01 2.11960170e-02 -9.75683853e-02 -7.40274906e-01
-2.92253435e-01 -3.54327053e-01 1.67135105e-01 -4.15808633e-02
-3.68518591e-01 2.40248576e-01 -9.26507041e-02 -4.03068244e-01
-4.94437618e-03 -9.30164278e-01 -7.66960919e-01 -5.44811964e-01
1.81986988e-01 4.33608145e-01 5.52720070e-01 -6.54715253e-03
2.74100095e-01 5.18191099e-01 -1.42468846e+00 -1.36056855e-01
1.04864371e+00 1.31204557e+00 8.70126307e-01 2.58072525e-01
-6.16173685e-01 4.15592283e-01 -4.21022102e-02 -6.40971819e-03
3.41571361e-01 2.84365658e-02 -5.76896906e-01 2.89537311e-01
1.04899704e-01 -6.03996456e-01 -4.78984147e-01 1.22112918e+00
2.19643399e-01 1.81516856e-01 2.11933941e-01 -1.25142932e+00
-2.08466295e-02 2.41203547e-01 -6.22298837e-01 -1.51130378e-01
1.28854960e-01 1.71881706e-01 7.22155809e-01 -9.76624846e-01
3.13390493e-01 1.36804461e+00 4.02637899e-01 2.22756162e-01
-7.12702632e-01 -8.22028399e-01 5.22870123e-01 2.41341799e-01
-1.42926264e+00 -4.06532168e-01 9.28382397e-01 -1.62217095e-01
6.66760802e-01 9.86048486e-03 7.96051860e-01 4.75769430e-01
-1.36596963e-01 1.14271557e+00 7.11669803e-01 1.32395759e-01
2.22068697e-01 -1.89574540e-01 2.25184962e-01 7.57231414e-01
1.13981143e-01 -9.95045900e-02 -3.30393732e-01 2.81812400e-01
9.08013046e-01 3.96601170e-01 -7.54858255e-02 -4.76519853e-01
-1.00353408e+00 8.33909333e-01 8.18072557e-01 -3.12447846e-01
-3.42168778e-01 3.62520158e-01 1.99055851e-01 -8.41235518e-02
4.53279227e-01 -2.31464729e-01 -6.02205992e-01 1.27802780e-02
-8.15668285e-01 2.87682801e-01 5.57772696e-01 1.43252587e+00
1.27292705e+00 -1.63319647e-01 1.36904687e-01 8.46886933e-01
5.54092109e-01 6.34288192e-01 -1.16886735e-01 -1.06714118e+00
5.82289517e-01 7.22976744e-01 -9.39021483e-02 -7.62364745e-01
-4.90518570e-01 -4.21786815e-01 -7.79872060e-01 1.35793863e-02
3.20924036e-02 5.64489998e-02 -1.17136323e+00 1.29951668e+00
7.43629038e-01 5.43993115e-01 -1.76686034e-01 1.28518963e+00
9.68582571e-01 6.66602194e-01 -6.59128604e-03 1.03609227e-01
1.37889886e+00 -1.09814680e+00 -3.95080358e-01 -5.59041619e-01
4.15298223e-01 -7.20732391e-01 8.81320715e-01 1.60824463e-01
-8.52268219e-01 -5.59659898e-01 -1.02483571e+00 -8.04659903e-01
-7.39237154e-03 9.11848769e-02 9.44672763e-01 -3.64571926e-04
-7.18408108e-01 2.21458301e-01 -1.17049682e+00 -2.39541128e-01
8.85786176e-01 2.33067885e-01 -8.05493444e-02 -5.83537221e-01
-8.54856551e-01 4.84830827e-01 5.59429407e-01 4.04418230e-01
-9.52504456e-01 -4.74393070e-01 -1.19823563e+00 -8.60852562e-03
5.88244677e-01 -7.11807966e-01 1.30202234e+00 -2.54361928e-01
-1.23998630e+00 6.35485172e-01 -4.95409518e-01 -3.14834803e-01
4.55537379e-01 -4.29256171e-01 1.14431337e-01 2.16525063e-01
4.15881842e-01 9.38119888e-01 6.57205045e-01 -1.18600583e+00
-8.78869116e-01 -6.04027450e-01 2.30534405e-01 4.98689771e-01
2.14324087e-01 -5.30177355e-01 -1.18850386e+00 -1.79406032e-01
9.01489258e-01 -8.27609777e-01 -5.72994351e-01 2.81254023e-01
-5.93389630e-01 -2.15194300e-01 9.53071356e-01 -6.22778535e-01
7.06764877e-01 -2.39812183e+00 5.69505729e-02 1.81814134e-01
2.64242262e-01 -7.30134696e-02 8.60387757e-02 1.14918482e-02
3.53040725e-01 -1.01397760e-01 -6.88471437e-01 -5.88405490e-01
-8.34607184e-02 4.31695282e-01 -3.24677616e-01 4.29807633e-01
2.71404356e-01 8.90582860e-01 -7.65375912e-01 -5.85192323e-01
4.83087063e-01 4.12700146e-01 -6.78940296e-01 2.29750127e-01
-4.03364301e-01 5.30158937e-01 -1.09935308e+00 8.22031796e-01
1.17969227e+00 -1.37317657e-01 -3.73873174e-01 -1.41079932e-01
-3.85158598e-01 5.96929252e-01 -1.17501903e+00 2.44975877e+00
-4.05830920e-01 3.32893193e-01 3.83274496e-01 -1.12527800e+00
9.76437807e-01 -2.13180915e-01 2.57873029e-01 -6.33254945e-01
3.08604807e-01 1.54917270e-01 -3.70914489e-01 -4.08612400e-01
4.63005483e-01 9.90822613e-02 -1.94753468e-01 3.03627569e-02
-1.75293088e-01 -6.92549825e-01 -1.74262635e-02 3.21463913e-01
7.22054541e-01 4.95483071e-01 -1.35273397e-01 -1.45716608e-01
5.83382905e-01 3.71117443e-01 9.11802709e-01 3.64248425e-01
-7.90509135e-02 5.68806887e-01 3.51244718e-01 -3.06928366e-01
-7.09061205e-01 -1.16680419e+00 -3.82776707e-02 5.07401407e-01
9.37122643e-01 -1.59041837e-01 -6.49560988e-01 -3.42649281e-01
2.60551181e-02 6.53689206e-01 -1.16375476e-01 -1.25575602e-01
-5.84027767e-01 -2.35612810e-01 1.42043307e-01 4.67493027e-01
9.57821786e-01 -7.93269753e-01 -6.50859594e-01 2.13770092e-01
-2.90451348e-01 -1.42853355e+00 -3.72265220e-01 2.57906288e-01
-1.16632307e+00 -1.09022987e+00 -2.62432933e-01 -8.20969462e-01
7.70875633e-01 8.60149980e-01 6.87396586e-01 2.76724190e-01
-2.25533694e-01 4.10736874e-02 -2.96240658e-01 -4.29826617e-01
2.52061218e-01 2.24138200e-01 -2.71407485e-01 -1.58019453e-01
4.05189484e-01 -7.47254789e-01 -7.46145487e-01 2.16241613e-01
-8.56181502e-01 5.60205579e-01 6.35484278e-01 4.66510624e-01
1.05928206e+00 8.03467557e-02 1.42556150e-02 -7.30876923e-01
-4.76202369e-02 -5.01923859e-01 -8.83574665e-01 -3.96711409e-01
-1.01884902e-01 -1.25841245e-01 4.36678469e-01 2.04709411e-01
-1.17600596e+00 5.34378827e-01 -4.86704379e-01 -7.05099642e-01
-3.14961731e-01 4.54240203e-01 -5.84672451e-01 2.14199439e-01
-9.28861201e-02 3.27849984e-01 1.02425560e-01 -7.00676322e-01
3.37869585e-01 5.86338699e-01 4.96636659e-01 -3.94230306e-01
9.43726659e-01 7.58688331e-01 -4.61077280e-02 -9.28546309e-01
-1.09949911e+00 -5.95077872e-01 -6.64176404e-01 -1.07379898e-01
1.00798488e+00 -1.32664430e+00 -5.72910070e-01 5.56038082e-01
-1.31304145e+00 -3.95842701e-01 -1.27192169e-01 5.51684201e-01
-3.84229153e-01 3.42328876e-01 -4.05493766e-01 -6.01203859e-01
-1.47966549e-01 -1.29239881e+00 1.34818554e+00 6.05484366e-01
3.89365911e-01 -5.63979566e-01 -4.94581223e-01 4.98167157e-01
-1.46199865e-02 3.45395356e-02 7.36147940e-01 -1.86527833e-01
-1.24922562e+00 -9.01251957e-02 -6.50999367e-01 2.15601102e-01
-6.40932620e-02 -3.03077340e-01 -9.92275476e-01 6.57628197e-03
1.71632349e-01 -9.75840986e-02 1.07279384e+00 3.38054895e-01
1.42844748e+00 1.55635282e-01 -4.32196617e-01 1.05080974e+00
1.38998878e+00 6.74775168e-02 5.20941675e-01 9.73026454e-02
9.45335209e-01 5.92316747e-01 9.44600165e-01 3.69599044e-01
8.83641720e-01 3.16233307e-01 8.72794390e-01 -3.69432196e-02
-1.23249285e-01 -5.26911795e-01 1.03307620e-01 9.10074890e-01
3.42954427e-01 9.53072831e-02 -8.86761844e-01 3.44563097e-01
-1.90321970e+00 -5.48694611e-01 -3.61474961e-01 1.96415007e+00
4.25587833e-01 3.14010292e-01 -3.97476017e-01 7.84991607e-02
7.00416028e-01 2.72102088e-01 -9.98151481e-01 2.59511918e-02
1.93428680e-01 1.66242167e-01 5.20555496e-01 6.55162871e-01
-1.03927505e+00 1.35586262e+00 4.20639706e+00 9.77702260e-01
-1.01380146e+00 1.19836092e-01 3.81636441e-01 8.93067382e-03
-3.94397408e-01 3.99586260e-01 -8.92823577e-01 3.56633842e-01
1.63794070e-01 1.91104218e-01 3.96987349e-01 9.78422582e-01
4.47014242e-01 -5.51189423e-01 -7.93799877e-01 1.06633437e+00
-2.42526397e-01 -1.11534572e+00 -3.28196660e-02 3.94832566e-02
3.43151629e-01 4.58800465e-01 -4.27863061e-01 1.88833594e-01
2.21803278e-01 -7.50136852e-01 8.55896533e-01 4.02110875e-01
7.11137772e-01 -8.64230454e-01 4.53720152e-01 7.15430975e-01
-1.57144785e+00 1.28411159e-01 -6.69071317e-01 -3.72390449e-01
4.05951321e-01 1.04936659e+00 -5.43607354e-01 7.55363643e-01
7.36811519e-01 1.05531013e+00 -2.03992829e-01 1.11164498e+00
-6.37859464e-01 3.78588349e-01 -6.40550137e-01 1.55154914e-01
3.61112714e-01 -5.84141076e-01 4.88971680e-01 8.09940994e-01
3.09213936e-01 3.64918172e-01 6.37318969e-01 9.05628324e-01
-3.74115296e-02 -2.21118256e-01 -5.15020132e-01 1.81676134e-01
7.37955272e-01 1.38376415e+00 -9.55047727e-01 -2.79940277e-01
-3.91926289e-01 1.09650218e+00 2.73393869e-01 3.17084640e-01
-8.60202432e-01 -5.59754729e-01 8.76818001e-01 -8.40700790e-02
3.50653678e-01 -7.60843337e-01 -6.45952880e-01 -1.13266242e+00
1.34017959e-01 -1.13720432e-01 -7.32972194e-03 -7.96136498e-01
-7.75412023e-01 2.23118529e-01 -1.10123985e-01 -1.21215534e+00
4.88726676e-01 -2.95243889e-01 -4.97894943e-01 9.45375800e-01
-1.95900559e+00 -1.12373292e+00 -8.75939846e-01 6.28049135e-01
7.46126354e-01 4.73169893e-01 3.09509307e-01 3.52943987e-01
-5.76544940e-01 -4.95886765e-02 -3.67952645e-01 1.43237397e-01
1.02186672e-01 -8.56584668e-01 5.10605216e-01 9.26208377e-01
-2.93820381e-01 2.69543380e-01 2.01377735e-01 -7.70112574e-01
-1.51840460e+00 -1.41599107e+00 5.97839236e-01 2.59721521e-02
1.91073745e-01 -4.36619908e-01 -8.18908453e-01 5.58885455e-01
-3.36563170e-01 -1.31252706e-01 2.91958004e-01 -2.10072517e-01
-1.46089062e-01 -1.74370632e-01 -8.77602458e-01 5.76739490e-01
1.54072845e+00 -3.70599985e-01 -4.90444362e-01 3.48371446e-01
1.27746630e+00 -7.52366900e-01 -5.01117408e-01 5.25891483e-01
2.46041417e-01 -8.62025738e-01 9.65026081e-01 1.12062514e-01
5.52766919e-01 -7.42509186e-01 -2.21986070e-01 -8.50076377e-01
-5.06795272e-02 -2.38846317e-01 1.21824130e-01 9.89933729e-01
1.16348349e-01 -5.38920879e-01 8.84407997e-01 4.41389412e-01
-7.55487502e-01 -7.64533877e-01 -9.06573355e-01 -5.17940700e-01
-2.99953312e-01 -9.62212503e-01 7.80584157e-01 5.58246315e-01
-6.21477902e-01 4.81033415e-01 -2.07880735e-02 5.22599399e-01
7.87772000e-01 4.98537004e-01 1.10204315e+00 -1.24192333e+00
1.89164922e-01 -3.15655112e-01 -2.00962916e-01 -1.96166301e+00
8.68003145e-02 -1.11384332e+00 4.49565768e-01 -1.99165118e+00
-3.98410372e-02 -9.02948856e-01 1.15637086e-01 4.89062876e-01
-1.35431841e-01 6.34912029e-02 1.26740098e-01 3.59983623e-01
-5.94212830e-01 9.65342224e-01 1.60978365e+00 -2.26041991e-02
-2.96949208e-01 3.62264961e-02 -6.26690507e-01 8.92372012e-01
9.22708035e-01 -4.68211055e-01 -6.00005329e-01 -8.70015264e-01
-1.25352710e-01 1.67393908e-01 4.53106761e-01 -9.55371261e-01
3.86001110e-01 -2.19463661e-01 3.16367924e-01 -1.16810775e+00
6.52717113e-01 -8.75575066e-01 -7.03035668e-02 4.98192072e-01
1.80131838e-01 -4.10677314e-01 7.67076313e-02 5.90200603e-01
-1.82578772e-01 -2.18303710e-01 5.76398432e-01 -3.11919898e-01
-9.60520685e-01 9.76651669e-01 -2.02242080e-02 -7.27818832e-02
8.89818549e-01 -3.14074755e-01 -2.61262134e-02 -6.04098216e-02
-4.53702182e-01 9.52429950e-01 5.89689553e-01 3.91263336e-01
8.54385734e-01 -1.07383132e+00 -4.91870761e-01 4.05076772e-01
3.60586867e-02 1.35661113e+00 5.22419095e-01 9.36290264e-01
-8.42923522e-01 2.82090992e-01 3.37415151e-02 -1.03550839e+00
-7.48375118e-01 1.62051350e-01 -2.28422787e-02 1.63618177e-01
-8.16544294e-01 9.71813858e-01 4.92195368e-01 -5.92570901e-01
1.85019583e-01 -4.31792557e-01 6.00796603e-02 -1.24528766e-01
2.12559164e-01 1.98100463e-01 -1.32482603e-01 -4.77904320e-01
-3.50298852e-01 6.60697401e-01 -7.32704848e-02 4.46409397e-02
1.17694104e+00 -4.26852047e-01 -3.07049960e-01 2.74970114e-01
1.14760303e+00 -3.42932373e-01 -1.54002035e+00 -2.77729303e-01
-1.46256134e-01 -5.25510907e-01 2.79951692e-01 -2.56407797e-01
-1.27042830e+00 1.25112402e+00 2.16697365e-01 -1.77673072e-01
1.16968513e+00 7.35689476e-02 1.21233559e+00 4.00152475e-01
6.09462619e-01 -9.48172569e-01 -2.76198894e-01 8.50100756e-01
6.60318613e-01 -1.16859162e+00 1.66512027e-01 -9.85283852e-01
-4.09721285e-01 8.47188056e-01 7.44276822e-01 -2.75999397e-01
8.07313979e-01 5.25167063e-02 -7.77789280e-02 -2.27794796e-01
-2.95564979e-01 -4.87959683e-01 -2.08196655e-01 4.58647877e-01
-2.59421706e-01 1.00135751e-01 -1.62675202e-01 7.17832148e-01
-2.29920357e-01 -4.29275408e-02 2.32581660e-01 1.05410266e+00
-7.39659250e-01 -8.67572725e-01 -7.67011121e-02 4.36627746e-01
7.14369416e-02 -1.96406409e-01 6.37118742e-02 5.84115326e-01
3.33215624e-01 9.48508918e-01 1.52282596e-01 -4.10829097e-01
3.98638040e-01 -1.32269830e-01 1.38804212e-01 -8.96577239e-01
1.30712822e-01 3.11121166e-01 -1.97405033e-02 -9.25498843e-01
-3.14417660e-01 -6.85811520e-01 -1.89732170e+00 -1.42500579e-01
-3.00929844e-01 7.07157031e-02 1.01158130e+00 1.03120577e+00
4.70218688e-01 5.25209308e-01 7.19716787e-01 -1.01506269e+00
-6.34508207e-02 -7.63625681e-01 -4.17659402e-01 -2.47832518e-02
2.90012240e-01 -7.92951822e-01 -2.08251446e-01 -3.71761955e-02] | [8.361595153808594, -2.881075143814087] |
582af57f-5869-40bf-9f81-848e34b6fc0c | infrared-and-visible-image-fusion-based-on | 2201.10739 | null | https://arxiv.org/abs/2201.10739v1 | https://arxiv.org/pdf/2201.10739v1.pdf | Infrared and visible image fusion based on Multi-State Contextual Hidden Markov Model | The traditional two-state hidden Markov model divides the high frequency coefficients only into two states (large and small states). Such scheme is prone to produce an inaccurate statistical model for the high frequency subband and reduces the quality of fusion result. In this paper, a fine-grained multi-state contextual hidden Markov model (MCHMM) is proposed for infrared and visible image fusion in the non-subsampled Shearlet domain, which takes full consideration of the strong correlations and level of details of NSST coefficients. To this end, an accurate soft context variable is designed correspondingly from the perspective of context correlation. Then, the statistical features provided by MCHMM are utilized for the fusion of high frequency subbands. To ensure the visual quality, a fusion strategy based on the difference in regional energy is proposed as well for lowfrequency subbands. Experimental results demonstrate that the proposed method can achieve a superior performance compared with other fusion methods in both subjective and objective aspects. | ['Xiao-Jun Wu', 'Zhancheng Zhang', 'Anqi Wang', 'Yuting Jiang', 'Xiaoqing Luo'] | 2022-01-26 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 1.79128185e-01 -6.21636033e-01 -2.52018571e-01 -1.77911773e-01
-7.51828849e-01 4.06732820e-02 3.86656612e-01 3.56599428e-02
-7.67953545e-02 5.81184387e-01 2.94105023e-01 1.92254871e-01
-1.84070915e-01 -8.51046741e-01 1.44300178e-01 -1.47153497e+00
1.66518971e-01 -5.39120257e-01 4.17552680e-01 3.57525982e-02
-3.04526407e-02 9.28137749e-02 -1.63767529e+00 3.02262217e-01
8.33228767e-01 1.19202995e+00 3.46986264e-01 5.62118828e-01
1.53518701e-02 7.02485383e-01 -4.43707734e-01 1.48815706e-01
7.70779699e-02 -7.25099921e-01 -2.69201398e-01 1.19211420e-01
-6.69581145e-02 -1.28608122e-01 -1.02889426e-01 1.47612774e+00
5.70855558e-01 2.89831817e-01 3.88232023e-01 -8.82345140e-01
-4.06106025e-01 3.20551872e-01 -7.61805773e-01 4.78799671e-01
3.78990173e-01 9.14616436e-02 7.02605426e-01 -7.22031653e-01
2.26361245e-01 1.18041635e+00 6.04355216e-01 -7.55779445e-02
-1.07551789e+00 -7.24930227e-01 -3.94354621e-03 5.58359027e-01
-1.49658835e+00 -2.50960886e-01 1.12235510e+00 -3.99976850e-01
4.47013259e-01 3.91792417e-01 7.77934849e-01 6.63353384e-01
6.76058233e-01 2.46332392e-01 1.86836910e+00 -5.10210991e-01
2.55830199e-01 -3.98878366e-01 3.21301788e-01 6.38910174e-01
3.72969359e-01 4.62486386e-01 -6.27565682e-01 -2.84741223e-01
5.10497570e-01 2.85682410e-01 -5.38655102e-01 1.03468500e-01
-1.23501122e+00 4.95922923e-01 1.54596537e-01 6.88237131e-01
-5.79456210e-01 -1.91560596e-01 2.81788409e-02 -1.32628977e-02
3.42439771e-01 -3.28821987e-01 -1.72675610e-01 2.60670841e-01
-1.10728657e+00 -1.34921268e-01 2.99040228e-01 2.82602370e-01
9.73361731e-01 1.79657608e-01 -3.16353112e-01 3.10251176e-01
6.46023452e-01 8.30233216e-01 2.08352014e-01 -8.83312047e-01
6.25674352e-02 3.18185657e-01 1.37558773e-01 -1.09850752e+00
-2.33476460e-01 -5.33205807e-01 -1.28552842e+00 3.91176790e-01
-9.59342048e-02 -1.06604628e-01 -9.50473428e-01 1.49298155e+00
5.52883208e-01 4.62332726e-01 4.37796891e-01 1.06723702e+00
1.03533804e+00 9.79078531e-01 3.71958971e-01 -9.84434724e-01
1.45784426e+00 -2.48994410e-01 -1.44131851e+00 2.88656235e-01
-5.82301393e-02 -1.21605587e+00 4.32447672e-01 4.24240977e-01
-9.43707526e-01 -1.13346744e+00 -1.17311764e+00 3.23913574e-01
1.03589945e-01 1.58786178e-01 3.71568680e-01 5.35662353e-01
-6.49730682e-01 4.03597802e-01 -1.02809274e+00 -2.53726952e-02
-1.17787071e-01 1.54669564e-02 -1.34018632e-02 -1.92371022e-03
-1.46470380e+00 7.74387479e-01 5.24967790e-01 4.43259805e-01
-2.21789166e-01 -2.34903783e-01 -7.93717921e-01 9.37518179e-02
1.61258891e-01 -6.15997493e-01 7.78000116e-01 -7.71941066e-01
-1.39345729e+00 1.25324264e-01 -2.79820114e-01 2.95255408e-02
-4.44236062e-02 1.61748171e-01 -8.34849060e-01 5.00580370e-01
4.55436326e-04 -9.72236469e-02 1.13944292e+00 -1.32410073e+00
-6.75740898e-01 -2.64372289e-01 -3.60002577e-01 3.05037290e-01
-5.76016121e-02 -1.09982453e-01 -1.92930419e-02 -9.21719670e-01
6.01826549e-01 -6.46339834e-01 -1.32743537e-01 -3.50048751e-01
-2.51791567e-01 -2.54979543e-02 1.15670788e+00 -9.00631189e-01
1.42358160e+00 -2.34600520e+00 6.09979667e-02 2.78698891e-01
6.16283529e-02 1.89012185e-01 4.41943765e-01 3.12870741e-01
-1.32317305e-01 -2.77548015e-01 -3.41464162e-01 7.86531866e-02
-4.92622644e-01 1.22150242e-01 7.56535539e-03 5.09476721e-01
-8.65789726e-02 3.70420516e-01 -8.66001368e-01 -8.80607903e-01
5.29412448e-01 7.85033107e-01 -9.41023976e-02 8.86853337e-02
3.48133057e-01 7.51832962e-01 -6.10716403e-01 6.35385931e-01
1.03478086e+00 -3.30202654e-02 8.44090655e-02 -9.16055739e-01
-4.48553950e-01 -4.36147064e-01 -1.46276450e+00 1.43977249e+00
-1.97120726e-01 5.34325764e-02 2.17719838e-01 -6.86779559e-01
9.72791612e-01 6.77427113e-01 8.17680538e-01 -4.79150116e-01
2.62196273e-01 -9.54164714e-02 -1.62876293e-01 -5.35121262e-01
3.44539195e-01 -6.96600735e-01 -5.34569994e-02 4.98544835e-02
-1.38166156e-02 -6.19533062e-02 -1.19530842e-01 -1.84433669e-01
5.43328226e-01 7.82383084e-02 5.74849129e-01 -2.59409040e-01
7.96568036e-01 -3.67724538e-01 1.06385076e+00 3.99511129e-01
-4.07764375e-01 5.63818395e-01 -1.49934486e-01 -1.78874239e-01
-6.09773219e-01 -9.22382236e-01 -7.33125433e-02 5.07946908e-01
6.67010665e-01 -3.58820647e-01 -3.48964989e-01 -2.05319718e-01
-3.71418089e-01 3.75911236e-01 -3.30903083e-01 -3.60013783e-01
-4.53287512e-01 -1.05029988e+00 -9.71772522e-02 1.35520577e-01
1.08179128e+00 -7.75271535e-01 -7.33858526e-01 2.08671406e-01
-4.67281550e-01 -1.15963852e+00 -1.88489065e-01 -1.13233536e-01
-8.32927227e-01 -9.86193538e-01 -6.31600201e-01 -6.86868191e-01
2.23913148e-01 6.02323532e-01 5.22774220e-01 2.90776461e-01
-5.44773825e-02 3.02531272e-02 -6.91387296e-01 -1.28819957e-01
-3.88545066e-01 -4.36890602e-01 -2.13256359e-01 5.31762600e-01
1.11533202e-01 -4.59200144e-01 -7.71258831e-01 1.65379956e-01
-9.34842646e-01 2.28307739e-01 8.72587562e-01 8.21007907e-01
7.68125117e-01 9.48810935e-01 3.08795631e-01 -1.51140690e-01
2.13591829e-01 -1.16840832e-01 -5.39371789e-01 1.52556643e-01
-6.10135138e-01 -1.17223956e-01 5.75557590e-01 -1.50513500e-01
-1.70689607e+00 -3.57765541e-03 2.53870264e-02 -3.13934714e-01
-5.74088842e-02 6.32716835e-01 -1.18818648e-01 -1.85067013e-01
1.38491407e-01 5.20269632e-01 3.03286817e-02 -5.75203419e-01
-4.30075042e-02 4.98060971e-01 6.29841745e-01 -3.33951443e-01
7.92520702e-01 7.56687343e-01 4.17525619e-01 -7.93150902e-01
-8.58243287e-01 -6.09348476e-01 -4.39729512e-01 -4.08030033e-01
1.22137225e+00 -1.10445869e+00 -6.51158273e-01 6.93292260e-01
-8.60218048e-01 2.49375939e-01 -2.12646231e-01 1.12984586e+00
-3.34167451e-01 8.81656826e-01 -8.00074816e-01 -1.05643189e+00
-2.86524475e-01 -1.19836426e+00 1.00509238e+00 5.20353675e-01
2.28295639e-01 -7.67614305e-01 2.26410981e-02 2.87012607e-01
2.58944273e-01 4.70878959e-01 7.55739570e-01 2.61800677e-01
-5.55103898e-01 7.64061287e-02 -4.09846976e-02 4.22641695e-01
2.70644665e-01 1.73734456e-01 -8.87921631e-01 -2.50956804e-01
6.01291418e-01 2.02582806e-01 1.00123668e+00 7.43583322e-01
7.36615539e-01 -7.74137527e-02 -1.52875260e-01 3.87596339e-01
1.70692074e+00 1.63023263e-01 6.64241910e-01 1.39312986e-02
5.10077357e-01 3.03742677e-01 1.04024851e+00 5.71457207e-01
1.40984342e-01 5.19575119e-01 2.49798939e-01 -4.55519795e-01
-4.02001381e-01 1.75324112e-01 4.78955865e-01 1.28629768e+00
-4.80212241e-01 1.00670725e-01 -5.32656848e-01 3.78170460e-01
-1.79523087e+00 -1.39648747e+00 -3.87690514e-01 2.07283902e+00
8.16820443e-01 5.29932603e-02 -4.41989183e-01 6.13887191e-01
1.13327742e+00 5.15689671e-01 -7.86471218e-02 2.70588677e-02
-4.44897771e-01 1.97848901e-01 4.10234958e-01 5.99033475e-01
-1.08010852e+00 3.00116450e-01 5.98631573e+00 1.02275658e+00
-1.02249336e+00 3.04127634e-01 3.22211534e-01 3.24817598e-01
-1.64164707e-01 4.11737919e-01 -5.64118385e-01 8.37453425e-01
5.97017288e-01 1.13759317e-01 -1.12790307e-02 1.35831535e-01
5.67203283e-01 -7.38558173e-01 -1.10691309e-01 9.76718664e-01
-2.66967684e-01 -8.51565301e-01 -1.09925546e-01 -7.31398836e-02
6.86784089e-01 -4.99523848e-01 7.37797245e-02 -1.22468606e-01
6.02651946e-02 -5.15523314e-01 7.11563885e-01 9.42641377e-01
6.74418747e-01 -8.97172034e-01 1.04419518e+00 4.55278784e-01
-1.72378290e+00 -7.12459683e-02 -9.58033055e-02 3.83307561e-02
4.69337672e-01 1.01260293e+00 -4.49991189e-02 1.17452502e+00
8.67961764e-01 6.58267438e-01 -1.82719260e-01 7.72229254e-01
-1.98343322e-01 7.05605686e-01 -1.85325235e-01 5.06341338e-01
-4.52877842e-02 -4.20444906e-01 6.44197166e-01 1.11070108e+00
4.87304062e-01 7.71075428e-01 5.53612471e-01 3.36001903e-01
7.23842919e-01 -6.93868473e-02 -1.14982784e-01 4.27183032e-01
2.65260875e-01 1.30989158e+00 -7.11125135e-01 -6.12829804e-01
-5.00586331e-01 6.25823736e-01 -6.57544494e-01 4.62193578e-01
-7.44081020e-01 -1.50660604e-01 2.44017735e-01 -1.68765754e-01
4.00297552e-01 -4.79688287e-01 -3.07311714e-01 -1.29148078e+00
-1.45782545e-01 -5.39050758e-01 7.15458035e-01 -8.80485952e-01
-1.02968645e+00 4.51305717e-01 2.20658332e-01 -1.56741130e+00
1.28794298e-01 1.40787140e-01 -5.24112642e-01 1.09550643e+00
-1.91897678e+00 -1.39044940e+00 -6.12696171e-01 7.93034554e-01
3.68872225e-01 2.35825390e-01 6.23998046e-01 2.02836290e-01
-5.23362279e-01 1.00175273e-02 1.46070607e-02 -1.87464997e-01
5.32329381e-01 -6.66139483e-01 -3.46494347e-01 1.29961216e+00
-2.97032803e-01 3.55182290e-01 1.02125371e+00 -1.02213633e+00
-1.04445720e+00 -8.70264530e-01 7.30768144e-01 2.00338602e-01
1.90869376e-01 3.02528322e-01 -9.24511731e-01 8.00635442e-02
3.11317861e-01 2.19075292e-01 6.51687562e-01 -4.55977678e-01
6.09488934e-02 -1.12066321e-01 -1.17583942e+00 1.88121468e-01
4.79981244e-01 -5.36298633e-01 -5.53553522e-01 -6.40234947e-02
5.30916810e-01 -8.76241326e-02 -1.36921859e+00 9.43532467e-01
4.60398734e-01 -1.23682928e+00 9.17481959e-01 2.87728637e-01
-8.50205421e-02 -9.13901329e-01 -3.72696161e-01 -9.59465086e-01
-8.10713530e-01 -6.62245214e-01 -3.09097856e-01 1.16428483e+00
-2.63095528e-01 -5.70242763e-01 4.14958388e-01 -1.24917991e-01
1.78981610e-02 -5.24836719e-01 -9.90288496e-01 -6.10354841e-01
-7.26567805e-01 -2.74467915e-02 3.53423238e-01 8.82184565e-01
-2.40471467e-01 1.54054686e-01 -3.06969047e-01 6.52346075e-01
1.03728962e+00 6.16339982e-01 9.42047313e-02 -1.18597817e+00
-1.27683938e-01 5.73731139e-02 -4.67281312e-01 -5.17771244e-01
-1.20678976e-01 -4.61690575e-01 -1.50105748e-02 -1.48946285e+00
3.76725733e-01 -3.39168280e-01 -7.35418677e-01 1.70998067e-01
-5.70475280e-01 4.66056228e-01 1.13905989e-01 6.24483407e-01
-3.56504321e-01 6.57391906e-01 1.36893153e+00 -6.39639888e-03
6.38634786e-02 -6.43796474e-02 -3.26053917e-01 6.64783001e-01
6.46224916e-01 -2.84226120e-01 -1.34237021e-01 1.41614005e-01
-2.25766420e-01 3.92724574e-01 5.45335770e-01 -1.33623171e+00
1.80331126e-01 -4.74265456e-01 5.53923011e-01 -9.40100670e-01
5.28638601e-01 -9.70899045e-01 4.87551451e-01 6.17037535e-01
1.27767593e-01 -1.99848592e-01 -1.75857499e-01 6.76033080e-01
-5.86635828e-01 2.07390502e-01 1.15058815e+00 -6.43426627e-02
-8.87245357e-01 1.58338845e-01 -3.81459624e-01 -5.48206210e-01
1.06894636e+00 -3.99573535e-01 -1.03721410e-01 -2.55001009e-01
-8.82497132e-01 -1.60329923e-01 4.23528582e-01 -9.32716355e-02
6.85204446e-01 -1.42178857e+00 -6.19592786e-01 3.90337795e-01
5.63318096e-02 -2.97744244e-01 9.76871431e-01 1.13874829e+00
-1.25191823e-01 1.17528908e-01 -1.63813129e-01 -7.38949895e-01
-1.47461903e+00 5.87694943e-01 -1.48622049e-02 -3.76402855e-01
-6.71534121e-01 3.61531973e-01 8.71724635e-02 3.55966955e-01
-3.72071266e-01 -4.46809232e-01 -5.50980389e-01 1.33951694e-01
5.04117668e-01 4.47087616e-01 2.13076966e-03 -1.29015064e+00
-1.26640320e-01 1.08106863e+00 4.75713164e-01 -1.12225696e-01
9.27367091e-01 -5.78897417e-01 -1.60600051e-01 3.29297900e-01
1.08855867e+00 1.73441455e-01 -1.38043523e+00 -2.38198280e-01
-5.14338791e-01 -5.16282678e-01 6.63130939e-01 -5.03788888e-01
-1.06045663e+00 7.35908687e-01 9.92373943e-01 3.51657569e-01
1.59294343e+00 -2.55710900e-01 1.10779619e+00 -3.42338145e-01
3.54961693e-01 -1.00432014e+00 -2.28481606e-01 -5.78525551e-02
3.28280687e-01 -9.19246435e-01 5.19233286e-01 -7.76140392e-01
-3.75319719e-01 1.12795055e+00 1.54418737e-01 2.53497660e-02
8.32042217e-01 2.79005736e-01 1.30564883e-01 -2.04195112e-01
-5.23354948e-01 -3.95683527e-01 2.83998191e-01 4.79372084e-01
1.36432514e-01 7.62211606e-02 -4.87353474e-01 2.45025203e-01
1.75971851e-01 3.71866375e-02 3.80625725e-01 1.23059118e+00
-8.03920865e-01 -8.42608809e-01 -9.99768138e-01 1.86688825e-01
-7.06383407e-01 7.99781233e-02 5.03971100e-01 2.47322902e-01
4.47611481e-01 1.29884994e+00 -3.20346415e-01 -4.34739649e-01
1.00492463e-02 -1.27761483e-01 4.11866397e-01 -2.46749818e-01
-2.66288131e-01 6.26359224e-01 -1.65200725e-01 -5.43867171e-01
-1.04450965e+00 -6.54942155e-01 -1.21847522e+00 -1.26398996e-01
-7.54449487e-01 2.74074584e-01 5.44361889e-01 8.01975667e-01
9.24603269e-02 4.80806947e-01 8.51390362e-01 -8.98632288e-01
-3.91377777e-01 -7.28002369e-01 -8.45693648e-01 3.31963688e-01
5.91133118e-01 -8.19579422e-01 -5.61165929e-01 3.83230269e-01] | [10.544173240661621, -1.9588836431503296] |
9b965b48-9156-4915-9382-da6b90ff834d | toward-grammatical-error-detection-from | 1906.01154 | null | https://arxiv.org/abs/1906.01154v6 | https://arxiv.org/pdf/1906.01154v6.pdf | Detecting Local Insights from Global Labels: Supervised & Zero-Shot Sequence Labeling via a Convolutional Decomposition | We propose a new, more actionable view of neural network interpretability and data analysis by leveraging the remarkable matching effectiveness of representations derived from deep networks, guided by an approach for class-conditional feature detection. The decomposition of the filter-ngram interactions of a convolutional neural network and a linear layer over a pre-trained deep network yields a strong binary sequence labeler, with flexibility in producing predictions at -- and defining loss functions for -- varying label granularities, from the fully-supervised sequence labeling setting to the challenging zero-shot sequence labeling setting, in which we seek token-level predictions but only have document-level labels for training. From this sequence-labeling layer we derive dense representations of the input that can then be matched to instances from training, or a support set with known labels. Such introspection with inference-time decision rules provides a means, in some settings, of making local updates to the model by altering the labels or instances in the support set without re-training the full model. Finally, we construct a particular K-nearest neighbors (K-NN) model from matched exemplar representations that approximates the original model's predictions and is at least as effective a predictor with respect to the ground-truth labels. This additionally yields interpretable heuristics at the token level for determining when predictions are less likely to be reliable, and for screening input dissimilar to the support set. In effect, we show that we can transform the deep network into a simple weighting over exemplars and associated labels, yielding an introspectable -- and modestly updatable -- version of the original model. | ['Allen Schmaltz'] | 2019-06-04 | null | null | null | null | ['grammatical-error-detection'] | ['natural-language-processing'] | [ 8.20429146e-01 6.74046993e-01 -4.97756451e-01 -8.57897460e-01
-9.54109430e-01 -7.63391078e-01 5.53365469e-01 4.05027419e-01
-4.74189550e-01 7.96886861e-01 2.10346460e-01 -3.66005629e-01
-1.50631340e-02 -8.09676170e-01 -1.02633965e+00 -6.90297842e-01
-2.81445161e-02 7.69737601e-01 -4.63862270e-02 7.76965022e-02
2.17424124e-01 5.44199944e-01 -1.66668046e+00 9.32157815e-01
3.87630403e-01 1.07110977e+00 6.38599098e-02 7.79658616e-01
2.87402645e-02 9.17623818e-01 -3.94624770e-01 -3.88426095e-01
3.11521679e-01 -3.20733190e-01 -1.03026617e+00 2.63537496e-01
5.16546369e-01 -3.41181844e-01 -2.39894018e-01 9.90076184e-01
-2.95660235e-02 2.83869296e-01 7.86078572e-01 -9.07951713e-01
-9.56051171e-01 9.55351174e-01 2.67929398e-02 1.00613624e-01
1.31307170e-01 4.78467286e-01 1.41732383e+00 -7.38346398e-01
7.55931497e-01 1.14921486e+00 8.87632012e-01 7.48827159e-01
-1.63278413e+00 -3.49172056e-01 4.72093761e-01 1.08479269e-01
-9.81543303e-01 -6.91993952e-01 2.47395158e-01 -4.23754245e-01
1.18892574e+00 4.23879206e-01 2.22616747e-01 1.23817360e+00
-3.16805243e-01 6.47308469e-01 8.49372625e-01 -6.32739365e-01
4.64450747e-01 2.64409304e-01 5.54007828e-01 9.56170201e-01
8.19099471e-02 8.54251459e-02 -4.92964238e-01 -3.85875493e-01
3.19240481e-01 1.97431281e-01 -2.62915224e-01 -7.09226280e-02
-1.02978611e+00 6.10137403e-01 4.47906673e-01 1.75312117e-01
-3.71833414e-01 3.06912929e-01 6.29967391e-01 4.29499447e-01
5.21182537e-01 8.08150351e-01 -9.32441354e-01 1.51766956e-01
-9.84586000e-01 1.64277405e-01 8.29852343e-01 8.88653696e-01
1.06413090e+00 3.35714500e-03 -5.07477403e-01 6.32589579e-01
4.69510537e-03 6.27746657e-02 5.78599036e-01 -9.33508635e-01
2.26189986e-01 5.13010800e-01 1.63898125e-01 -5.53588986e-01
-2.54016489e-01 -4.75393116e-01 -5.78234673e-01 2.65923161e-02
3.24394971e-01 -2.07896885e-02 -1.19841921e+00 2.13842177e+00
5.91838397e-02 2.72366315e-01 1.17184035e-01 7.26211667e-01
3.42159480e-01 7.39292502e-01 6.93056211e-02 -1.36250764e-01
1.19382977e+00 -7.58664548e-01 -9.60324705e-02 -3.80563080e-01
1.15193343e+00 -7.06089512e-02 1.12959146e+00 3.33013654e-01
-1.00182772e+00 -5.33509672e-01 -1.01564085e+00 -1.76029444e-01
-4.19231743e-01 -4.33344990e-02 5.03547251e-01 3.12351942e-01
-1.26238263e+00 1.12997413e+00 -6.62569940e-01 -2.86377251e-01
5.14991760e-01 3.64266336e-01 -3.20504487e-01 -1.85631827e-01
-1.30353808e+00 1.06777787e+00 7.84636617e-01 -4.12845872e-02
-1.15856957e+00 -8.05011034e-01 -8.36107254e-01 3.35828543e-01
4.03315514e-01 -3.74861896e-01 1.52847564e+00 -1.41307926e+00
-1.15345955e+00 1.07867408e+00 -4.48710591e-01 -7.33291268e-01
2.07956389e-01 1.60880253e-01 -3.45454603e-01 -7.03401268e-02
1.55741170e-01 6.70871854e-01 8.06823909e-01 -1.26941538e+00
-7.66831875e-01 -2.92455584e-01 1.72879621e-01 6.19860068e-02
-1.18386857e-01 -2.50168175e-01 3.00736502e-02 -3.41489047e-01
7.41460398e-02 -9.24354136e-01 -3.39087516e-01 7.61421248e-02
-7.30965614e-01 -2.37364605e-01 4.73475099e-01 -5.16944826e-01
9.89901185e-01 -2.03984690e+00 -9.51905847e-02 4.17460740e-01
3.05101484e-01 2.61001825e-01 -3.28944772e-01 3.24678659e-01
-2.33925372e-01 3.52861851e-01 -3.88305336e-01 -4.64116365e-01
1.14272282e-01 5.24083793e-01 -7.61775076e-01 4.08356190e-01
6.22246027e-01 8.08933556e-01 -1.13185835e+00 -9.68233570e-02
-2.37741172e-01 -9.29606408e-02 -5.36726356e-01 2.87202060e-01
-7.13733435e-01 2.31527016e-02 -3.63586277e-01 5.00607133e-01
3.07755440e-01 -6.12085700e-01 4.16704327e-01 8.27256590e-02
1.59876466e-01 5.47937274e-01 -9.20576453e-01 1.44051063e+00
-4.93944585e-01 5.62664449e-01 -8.24567676e-02 -1.31513071e+00
8.42017055e-01 2.63231367e-01 -5.88334985e-02 -2.09046900e-01
-1.63910508e-01 2.36233920e-01 -6.45181956e-03 -4.51737791e-01
4.82013673e-01 -3.83616477e-01 -5.17838672e-02 7.75334239e-01
3.20820510e-01 4.00035977e-01 1.61383413e-02 2.37874612e-01
1.38147461e+00 1.80865034e-01 3.50230753e-01 -6.64343387e-02
2.10631371e-01 7.95625225e-02 6.79213822e-01 1.34903347e+00
-1.45693906e-02 5.03034055e-01 5.80450237e-01 -7.44370461e-01
-1.29918420e+00 -9.33506787e-01 -1.84830129e-01 1.53513026e+00
-3.48123014e-01 -2.26963550e-01 -5.42502463e-01 -9.04312372e-01
2.36870155e-01 1.00476980e+00 -8.90986145e-01 -4.64297801e-01
-4.95456010e-01 -2.59196848e-01 6.96856320e-01 5.43738306e-01
-1.79828089e-02 -1.27009809e+00 -3.36985886e-01 2.53892094e-01
1.03459656e-01 -7.17969596e-01 -2.55292803e-01 1.13554680e+00
-6.82308733e-01 -7.68564403e-01 -5.86014502e-02 -7.67563879e-01
8.50418329e-01 -3.46082121e-01 1.18133688e+00 3.11441272e-01
-9.27133113e-02 -1.08397610e-01 -2.69146979e-01 -6.04980029e-02
-9.14743304e-01 1.19040720e-01 1.14537559e-01 1.09914236e-01
4.28127944e-01 -4.91429180e-01 -3.22056741e-01 2.73159035e-02
-9.52213585e-01 -1.09110773e-02 5.59918523e-01 1.02864552e+00
5.90773821e-01 -2.62503117e-01 6.90647244e-01 -1.52344275e+00
3.80708396e-01 -6.51352108e-01 -3.38766664e-01 4.80690479e-01
-6.96647346e-01 6.78819358e-01 1.18997467e+00 -5.60519397e-01
-9.24789011e-01 6.15758225e-02 -1.22041360e-01 -5.87042928e-01
-4.27660733e-01 5.61875880e-01 -4.78361510e-02 2.35845953e-01
1.15547073e+00 3.76709014e-01 -1.07320160e-01 -4.72629547e-01
7.09780157e-01 7.49320209e-01 7.93688834e-01 -8.85910094e-01
5.02148092e-01 3.12131941e-01 -2.18261912e-01 -2.06990793e-01
-1.28524625e+00 -2.38800883e-01 -7.41561592e-01 2.60073334e-01
5.21873116e-01 -5.85125208e-01 -5.94816983e-01 1.62460208e-01
-1.31068981e+00 -5.29123545e-01 -6.60648823e-01 -9.84528586e-02
-6.70029342e-01 1.05406813e-01 -8.63719106e-01 -6.30761027e-01
-9.41674411e-02 -1.06481528e+00 9.14790988e-01 -1.59032300e-01
-5.56593359e-01 -9.79229867e-01 -4.77908477e-02 -5.47308326e-02
1.64055914e-01 -1.27773574e-02 1.40973830e+00 -1.53890419e+00
-5.41068196e-01 -4.04752225e-01 -9.05703157e-02 6.28343105e-01
-1.69927254e-01 -4.39851992e-02 -1.38715959e+00 -2.22465962e-01
-7.91839436e-02 -7.68598914e-01 1.13125110e+00 3.60787846e-02
1.48553574e+00 -7.65579164e-01 -5.73793471e-01 5.75463176e-01
1.12093186e+00 -2.87245531e-02 3.19343597e-01 2.92586386e-01
4.70577985e-01 4.78346139e-01 3.50372344e-01 4.69672948e-01
-1.69449583e-01 4.02784288e-01 4.43781435e-01 1.40715763e-01
1.86133668e-01 -6.32876217e-01 1.79812700e-01 3.10354829e-01
3.47961515e-01 -4.21325773e-01 -7.41696954e-01 4.92757082e-01
-1.95810378e+00 -1.23042405e+00 4.10427958e-01 2.31397223e+00
1.23697889e+00 2.85671115e-01 -1.98093235e-01 1.75922345e-02
8.40807617e-01 1.77776799e-01 -9.23755229e-01 -6.88313007e-01
1.54184401e-01 1.94265872e-01 5.87634563e-01 5.37095964e-01
-1.10389543e+00 9.76603925e-01 6.56997824e+00 9.07558084e-01
-1.05378103e+00 -1.58541240e-02 1.03778303e+00 -1.42408371e-01
-6.87531769e-01 1.63663253e-01 -9.61788058e-01 5.02089858e-01
1.35037720e+00 1.17877632e-01 6.08412743e-01 1.05441904e+00
-1.66374464e-02 2.98579484e-01 -1.89448464e+00 4.87062424e-01
-7.01577514e-02 -1.87952685e+00 3.40995401e-01 -1.73535373e-03
5.64323962e-01 1.34127647e-01 8.80090520e-02 4.23636913e-01
8.60961437e-01 -1.08352971e+00 9.29185808e-01 5.87505758e-01
1.07659352e+00 -5.85426331e-01 3.99254650e-01 7.09495008e-01
-5.56378961e-01 -3.37387651e-01 -5.76468766e-01 -2.76025951e-01
-1.67206809e-01 5.35130560e-01 -1.36497176e+00 -1.89869199e-02
2.03700677e-01 6.65365458e-01 -7.69883320e-02 3.65627497e-01
-2.46885896e-01 8.42453778e-01 -1.87175691e-01 -6.97601512e-02
4.87534970e-01 1.40409902e-01 2.17883140e-01 1.57549918e+00
-2.63856314e-02 2.00785741e-01 3.46345961e-01 1.06922197e+00
-4.05574471e-01 -5.22492766e-01 -6.43251896e-01 -1.72071099e-01
7.99154758e-01 1.21911991e+00 -4.23768997e-01 -7.86751211e-01
-1.54647201e-01 9.04144764e-01 1.05747271e+00 5.30199587e-01
-4.54516172e-01 -2.45641500e-01 8.81043255e-01 4.47941991e-03
3.52172196e-01 3.00124943e-01 -3.27580243e-01 -1.04192662e+00
-7.87527934e-02 -9.12555277e-01 4.31508482e-01 -7.01128781e-01
-1.40175676e+00 5.61378300e-01 -3.75460327e-01 -8.17231834e-01
-5.95123410e-01 -7.92539358e-01 -5.20285308e-01 1.05538881e+00
-1.21921611e+00 -8.89748275e-01 3.16825390e-01 2.20891416e-01
4.10531759e-01 -4.79316823e-02 1.12629282e+00 -3.07354718e-01
-4.28101152e-01 9.29000616e-01 3.33950996e-01 2.49845386e-01
3.93355936e-01 -1.20531166e+00 5.44098198e-01 7.96587944e-01
3.27940732e-01 7.99556375e-01 6.16027474e-01 -4.36295629e-01
-1.08770132e+00 -1.53639877e+00 1.02356899e+00 -5.78003287e-01
6.90710008e-01 -4.29141074e-01 -1.33346915e+00 1.30775881e+00
-3.08047593e-01 2.30129883e-01 7.66293943e-01 4.28122133e-01
-8.34248304e-01 1.57078534e-01 -1.34409702e+00 5.73002636e-01
1.25135112e+00 -8.98299575e-01 -6.88162565e-01 5.88638186e-01
1.01719260e+00 -3.33411872e-01 -4.92524952e-01 1.11801520e-01
3.68155956e-01 -7.37359583e-01 6.79900885e-01 -1.49272811e+00
6.42422020e-01 -4.31710966e-02 -3.34446281e-01 -1.38530612e+00
-8.23664665e-01 -3.75139385e-01 2.02233773e-02 9.66383040e-01
8.18778694e-01 -5.66106856e-01 9.63741779e-01 9.08837438e-01
-6.01047933e-01 -1.10175705e+00 -7.55041957e-01 -4.97963428e-01
-1.64444357e-01 -5.85931361e-01 6.03851616e-01 1.08860326e+00
1.89014032e-01 2.33820856e-01 -1.83214456e-01 3.13268840e-01
3.18370521e-01 1.89010054e-01 2.85880655e-01 -9.97991621e-01
-7.75369167e-01 -2.92018682e-01 -4.56691623e-01 -1.13795507e+00
5.98008394e-01 -1.52309501e+00 4.40140575e-01 -1.21460176e+00
2.86316097e-01 -6.34976625e-01 -6.73841119e-01 1.02329886e+00
-1.03398293e-01 -6.72578514e-02 -5.29205836e-02 3.92223686e-01
-6.59672141e-01 -4.86111492e-02 7.21584141e-01 -1.19989298e-01
1.17308803e-01 2.29700049e-03 -8.33644271e-01 7.37262309e-01
6.06642425e-01 -6.76850677e-01 -2.66615808e-01 -3.27812314e-01
2.88364053e-01 7.56401122e-02 3.92981708e-01 -5.29117584e-01
1.45926639e-01 -2.19590798e-01 5.31574905e-01 -1.47240803e-01
2.04833850e-01 -4.70045924e-01 -5.44061102e-02 3.82006943e-01
-1.49169648e+00 -3.28686446e-01 -1.11761786e-01 7.56750703e-01
2.35176340e-01 -6.90599203e-01 7.62576401e-01 -4.76760060e-01
-8.98907959e-01 3.31390411e-01 -3.70349914e-01 1.12768635e-01
5.72774887e-01 -3.71617109e-01 -5.18712521e-01 -2.20433459e-01
-1.01026011e+00 9.77815762e-02 5.83571672e-01 -2.64028255e-02
4.75378573e-01 -1.02553940e+00 -5.19787312e-01 3.42645228e-01
3.40223134e-01 9.69720259e-02 -8.36821869e-02 2.17678800e-01
-9.65591371e-02 3.56180608e-01 2.18903646e-02 -3.97359222e-01
-6.83481574e-01 7.05980420e-01 4.87765253e-01 -2.81495601e-01
-4.02767599e-01 1.14954579e+00 3.15974146e-01 -7.99439490e-01
2.11341649e-01 -5.47887504e-01 3.56727064e-01 -3.18537384e-01
5.31252146e-01 2.26720665e-02 1.82199582e-01 -2.55185276e-01
-2.77712494e-01 -2.61792183e-01 -4.73391652e-01 1.04922161e-01
1.20728743e+00 1.91927731e-01 -1.24367028e-01 7.28151023e-01
1.29238963e+00 -4.12652820e-01 -1.52216113e+00 -5.94101071e-01
1.63920462e-01 -2.44237304e-01 -3.58788341e-01 -1.07884645e+00
-6.19366884e-01 8.11225712e-01 7.41263926e-02 1.88592896e-01
6.46013379e-01 1.45959914e-01 4.00849313e-01 9.80246067e-01
1.13216415e-01 -8.89057577e-01 2.37639491e-02 6.89676881e-01
5.34017384e-01 -1.07037032e+00 -2.81670004e-01 9.41348374e-02
-5.29950738e-01 1.07562017e+00 5.99891126e-01 -1.48101389e-01
2.94032484e-01 3.15670073e-01 -7.39759132e-02 -4.95956428e-02
-1.32694888e+00 1.34747669e-01 4.47352156e-02 6.84475243e-01
2.93236911e-01 2.24072233e-01 4.17139679e-01 6.21263266e-01
-3.00611146e-02 -1.02053694e-01 3.41221601e-01 6.85568273e-01
-9.25669909e-01 -7.64371514e-01 -1.26619544e-02 1.01835978e+00
-2.11086050e-01 -4.31613654e-01 -2.95281798e-01 4.44066972e-01
2.63019204e-01 6.09394908e-01 2.18278512e-01 -5.52873671e-01
6.01240955e-02 6.81170821e-01 8.95199552e-02 -1.15186024e+00
-6.19444132e-01 -5.57873845e-01 3.82848233e-01 -4.48500901e-01
8.46448392e-02 -5.70544958e-01 -1.52017295e+00 -2.15118214e-01
-1.79880753e-01 1.26659825e-01 3.35640818e-01 1.31394637e+00
5.19975841e-01 1.69002861e-01 6.35073125e-01 -9.17393327e-01
-1.19999790e+00 -7.86992729e-01 -6.26344800e-01 6.89506650e-01
5.63539267e-01 -2.77469754e-01 -4.94241744e-01 1.74527749e-01] | [9.751568794250488, 7.05945348739624] |
965ce0bf-c3ed-4dce-a5b9-ce64ebf305e1 | lade-the-first-comprehensive-last-mile | 2306.10675 | null | https://arxiv.org/abs/2306.10675v1 | https://arxiv.org/pdf/2306.10675v1.pdf | LaDe: The First Comprehensive Last-mile Delivery Dataset from Industry | Real-world last-mile delivery datasets are crucial for research in logistics, supply chain management, and spatio-temporal data mining. Despite a plethora of algorithms developed to date, no widely accepted, publicly available last-mile delivery dataset exists to support research in this field. In this paper, we introduce \texttt{LaDe}, the first publicly available last-mile delivery dataset with millions of packages from the industry. LaDe has three unique characteristics: (1) Large-scale. It involves 10,677k packages of 21k couriers over 6 months of real-world operation. (2) Comprehensive information. It offers original package information, such as its location and time requirements, as well as task-event information, which records when and where the courier is while events such as task-accept and task-finish events happen. (3) Diversity. The dataset includes data from various scenarios, including package pick-up and delivery, and from multiple cities, each with its unique spatio-temporal patterns due to their distinct characteristics such as populations. We verify LaDe on three tasks by running several classical baseline models per task. We believe that the large-scale, comprehensive, diverse feature of LaDe can offer unparalleled opportunities to researchers in the supply chain community, data mining community, and beyond. The dataset homepage is publicly available at https://huggingface.co/datasets/Cainiao-AI/LaDe. | ['Huaiyu Wan', 'Youfang Lin', 'Roger Zimmermann', 'Liuqing Yang', 'Yuxuan Liang', 'Junhong Lou', 'Jianbin Zhen', 'Ergang Shan', 'Yutong Xia', 'Xiaowei Mao', 'Haoyuan Hu', 'Haomin Wen', 'Lixia Wu'] | 2023-06-19 | null | null | null | null | ['management'] | ['miscellaneous'] | [-5.95183253e-01 -9.62657213e-01 -3.45522761e-01 -5.68323374e-01
-7.11693943e-01 -8.30178261e-01 4.55886573e-01 3.94238949e-01
-2.40530819e-02 5.70369184e-01 2.48224914e-01 -3.84391963e-01
-6.74797595e-01 -8.79499078e-01 -6.99332476e-01 -7.63068795e-01
-6.96142733e-01 8.14981341e-01 -2.64753938e-01 -2.86301553e-01
1.19558886e-01 4.16641176e-01 -1.10669374e+00 2.64637917e-01
7.61986613e-01 1.25446939e+00 5.45975566e-01 2.02230349e-01
6.66418374e-02 4.58361089e-01 -4.44060981e-01 -3.71036083e-01
4.26967084e-01 6.79264739e-02 -2.63248086e-01 3.75931375e-02
-1.52911499e-01 -6.65561035e-02 -3.18537176e-01 4.71490800e-01
3.27465326e-01 -1.32778972e-01 5.19639313e-01 -1.93717277e+00
-8.22287023e-01 6.03784382e-01 -6.14513457e-01 3.49512011e-01
3.26454490e-01 4.03549463e-01 9.20978129e-01 -9.17230666e-01
7.28520215e-01 7.70177007e-01 6.83114469e-01 -4.46785450e-01
-1.05235076e+00 -7.67281950e-01 1.41833842e-01 3.13032269e-02
-1.32915616e+00 -9.88267735e-02 6.58072770e-01 -5.81770301e-01
8.56168926e-01 2.45793283e-01 5.21397710e-01 1.05135798e+00
6.75752044e-01 6.91102982e-01 1.13891804e+00 1.54230013e-01
1.22525707e-01 4.40087318e-02 1.32691234e-01 3.24796319e-01
3.44761871e-02 -1.72504872e-01 -5.60519695e-01 -3.34024638e-01
5.56690276e-01 6.11934423e-01 1.52739897e-01 -1.96853623e-01
-1.44810700e+00 7.88981378e-01 1.62811577e-01 2.13865519e-01
-6.93406105e-01 1.78937837e-01 5.93816817e-01 5.70001841e-01
6.31430387e-01 1.09635741e-01 -9.50829387e-01 -3.71873766e-01
-4.31086034e-01 7.87048817e-01 9.92643952e-01 1.53760004e+00
7.92757690e-01 -1.94901630e-01 1.12307072e-01 5.12451768e-01
4.63157967e-02 7.22270370e-01 -1.64593786e-01 -6.80165350e-01
1.08647323e+00 5.53915620e-01 2.08662570e-01 -1.14519668e+00
-7.90849626e-01 -3.70976329e-01 -7.71557271e-01 -3.93825829e-01
6.64201498e-01 -2.70257324e-01 -5.51687121e-01 1.21632075e+00
7.70307630e-02 -2.98119962e-01 -3.28002006e-01 8.16480517e-01
7.20771849e-01 8.02787840e-01 5.18915989e-02 -2.98180044e-01
1.51493335e+00 -6.59001231e-01 -8.16802084e-01 -7.03920797e-02
5.67398548e-01 -9.43929791e-01 8.58830571e-01 4.72399145e-01
-7.18169689e-01 -1.31747037e-01 -2.75756538e-01 4.61702257e-01
-6.77586854e-01 -7.74453282e-02 1.19532442e+00 2.85906941e-01
-4.86609310e-01 1.80627048e-01 -3.94494742e-01 -4.73347902e-01
4.89847124e-01 7.89039433e-02 -4.30615604e-01 -4.91662979e-01
-1.03743529e+00 5.69981873e-01 -1.54002577e-01 2.36566767e-01
-8.30138266e-01 -1.03996384e+00 -7.26077378e-01 -9.40670744e-02
7.89988816e-01 -2.34928519e-01 9.48776305e-01 -1.63343772e-01
-5.33794940e-01 3.70333642e-01 2.88488921e-02 1.50011748e-01
3.45701903e-01 4.44571413e-02 -7.46856630e-01 -4.39205945e-01
3.82153630e-01 -1.54853687e-01 2.73814768e-01 -1.18108153e+00
-7.74351835e-01 -6.03367984e-01 -3.44788015e-01 -1.81822643e-01
2.26362702e-02 5.58581352e-01 -5.05746543e-01 -8.38182926e-01
-5.69775477e-02 -1.03513086e+00 -2.39334792e-01 -7.57537544e-01
-5.21067500e-01 -6.35815501e-01 4.95860606e-01 -7.29494095e-01
1.45822823e+00 -2.13583302e+00 -3.47501338e-01 1.83797136e-01
5.20262942e-02 -4.46409196e-01 -2.01331735e-01 1.29234588e+00
1.55911669e-01 -2.35178154e-02 9.28586572e-02 -2.43447155e-01
3.65360737e-01 1.01923309e-01 -2.23792091e-01 7.05296993e-01
2.60042589e-06 1.14952421e+00 -9.95397627e-01 -9.76635218e-02
1.92691639e-01 -7.11148530e-02 -1.06535800e-01 5.40996678e-02
-1.83168069e-01 4.18169349e-01 -7.44161248e-01 1.27915764e+00
8.09854865e-01 -8.46921504e-02 1.20298035e-01 -1.70756012e-01
-3.43424708e-01 2.18025714e-01 -9.90282655e-01 1.55657721e+00
-4.72234875e-01 4.98213112e-01 2.60372341e-01 -7.54825592e-01
8.99950624e-01 7.26934224e-02 9.96569037e-01 -9.09655988e-01
1.28316820e-01 2.66254842e-01 -1.47804007e-01 -6.51006699e-01
5.29108405e-01 2.19484329e-01 -9.58091378e-01 6.78659141e-01
-5.53893328e-01 2.10253984e-01 5.90905011e-01 1.31445751e-01
1.54887187e+00 -3.35415266e-02 -2.27967978e-01 -3.59558791e-01
-3.87712002e-01 3.76835167e-01 7.96078265e-01 4.37669128e-01
-2.28617728e-01 1.87128216e-01 5.66628695e-01 -8.22476208e-01
-8.95636976e-01 -1.24208629e+00 -2.54884064e-01 1.38225329e+00
5.97422272e-02 -3.37300479e-01 -1.43208876e-01 -4.70962882e-01
6.49424732e-01 4.56072032e-01 -3.03469896e-01 6.00018501e-01
-8.29127848e-01 -9.11452174e-01 3.73727940e-02 5.47766268e-01
3.15305382e-01 -1.09466636e+00 -1.95643485e-01 6.97676361e-01
-3.25571924e-01 -1.18455756e+00 -6.44649446e-01 3.29375356e-01
-5.90398967e-01 -1.21266139e+00 -4.69003171e-01 -7.67226338e-01
5.47996521e-01 3.15692604e-01 1.58703244e+00 -3.55591089e-01
-4.62460428e-01 -2.87884530e-02 -4.73526657e-01 -4.80821669e-01
2.82353275e-02 1.52401015e-01 1.77257642e-01 -2.07365617e-01
6.71657205e-01 -5.69562256e-01 -5.25239766e-01 7.45890141e-01
-5.36389887e-01 -3.79757494e-01 7.31928110e-01 1.70492962e-01
4.24566507e-01 6.16535902e-01 1.19296730e+00 -8.57658088e-01
6.23984993e-01 -1.27230322e+00 -6.36705756e-01 1.45609275e-01
-5.20709157e-01 -7.24762499e-01 6.07741475e-01 -1.34707257e-01
-8.16772819e-01 -9.96209010e-02 1.66415155e-01 -5.61860055e-02
-3.25658560e-01 7.66430616e-01 -2.06340656e-01 6.38518095e-01
-1.24043792e-01 3.61720026e-01 -1.03813581e-01 -9.20300245e-01
1.71136037e-01 7.90853918e-01 4.43765521e-01 -5.60399652e-01
8.37401330e-01 2.39584431e-01 -2.28961378e-01 -3.50144267e-01
-1.26412481e-01 -7.19834268e-01 -4.85414177e-01 -1.71273947e-01
5.93265414e-01 -9.32113290e-01 -1.23668683e+00 8.69750261e-01
-9.86913621e-01 -4.99493569e-01 3.10316622e-01 3.47024590e-01
-3.72966975e-01 -2.49465764e-01 -8.42997670e-01 -7.23598719e-01
2.80272327e-02 -1.00848305e+00 8.86827111e-01 -3.33895206e-01
-9.93288606e-02 -1.14561594e+00 -3.25143725e-01 5.61916292e-01
6.01137996e-01 6.68640912e-01 9.95035172e-01 -6.17423236e-01
-8.12678277e-01 -2.81236708e-01 -1.72866613e-01 -1.03650846e-01
4.30737883e-01 -1.85559750e-01 -1.56663448e-01 -4.54532623e-01
-3.46898377e-01 -1.04871765e-02 4.22707707e-01 4.44860011e-01
1.04994953e+00 -3.99668962e-01 -5.40768266e-01 2.72477627e-01
1.41041374e+00 4.64932770e-01 3.10334861e-01 4.09141928e-01
4.41931248e-01 7.93019414e-01 1.24606359e+00 8.44771802e-01
9.35626149e-01 7.97851741e-01 5.31281531e-01 -1.93937823e-01
4.44370270e-01 1.00817084e-01 3.69103611e-01 9.25451159e-01
3.83652821e-02 -4.88931090e-01 -1.08254588e+00 1.00002253e+00
-1.90464199e+00 -8.84782135e-01 -4.61017191e-01 1.83583391e+00
5.96028745e-01 2.32468583e-02 5.57527840e-01 -4.13872153e-02
4.17296767e-01 1.45044178e-01 -4.13727999e-01 -3.40369582e-01
-4.47266130e-03 -5.38104475e-01 9.13878679e-01 -1.47715569e-01
-1.05391955e+00 4.63609576e-01 5.67876196e+00 8.33145678e-01
-6.23247504e-01 1.82875812e-01 7.18190253e-01 -4.20528173e-01
-2.80618310e-01 -1.82828233e-01 -9.08184886e-01 1.18557572e+00
7.20201313e-01 -9.27482024e-02 4.80228424e-01 8.83658469e-01
4.97805685e-01 -2.23137051e-01 -1.33982289e+00 7.74967611e-01
-2.86131762e-02 -1.47873020e+00 -6.09263301e-01 5.45993149e-01
7.33848691e-01 2.37867445e-01 1.28116921e-01 2.67499954e-01
4.97544497e-01 -8.52812409e-01 8.52654219e-01 3.69700223e-01
5.79766691e-01 -1.00237846e+00 8.47426236e-01 5.06652296e-01
-1.49991763e+00 -3.54394615e-01 1.17922433e-01 -1.77501798e-01
6.15082324e-01 1.13540828e+00 -6.47979438e-01 9.10227716e-01
1.07811916e+00 7.15687215e-01 -2.14378834e-01 8.70606899e-01
4.74973977e-01 4.44385558e-01 -2.85551876e-01 -7.72860423e-02
3.75755131e-01 -5.50262988e-01 3.19880933e-01 1.24764812e+00
4.34648722e-01 -2.38060787e-01 6.66495860e-01 6.16443694e-01
-4.61114682e-02 -1.10678725e-01 -8.07529509e-01 -4.50726748e-02
9.40475106e-01 1.28032768e+00 -7.00351775e-01 2.53022611e-01
-7.12076664e-01 4.87947732e-01 -1.99024528e-01 5.43892860e-01
-1.01388335e+00 -4.55322266e-01 1.01718354e+00 5.70540369e-01
2.01445088e-01 -6.99466705e-01 -7.71473572e-02 -5.29249728e-01
4.12529558e-01 -6.44027948e-01 3.27472508e-01 -5.16281903e-01
-1.86896360e+00 4.86652613e-01 1.95595860e-01 -1.08439851e+00
2.07885355e-02 -5.51299751e-01 -5.20677328e-01 8.79288495e-01
-1.58363283e+00 -1.34713423e+00 -2.35494539e-01 4.58516777e-01
6.57519579e-01 -4.40841943e-01 3.39664251e-01 8.68250668e-01
-7.78571427e-01 3.33306760e-01 3.94523889e-01 4.43136334e-01
6.12320483e-01 -1.02270567e+00 6.59333467e-01 2.61230201e-01
-4.34171766e-01 7.19324768e-01 5.03483534e-01 -9.60451126e-01
-2.33120608e+00 -1.39030707e+00 1.14211798e+00 -1.02640831e+00
1.05639505e+00 -7.64898241e-01 -3.68654966e-01 1.01815057e+00
-2.06726491e-02 -2.10342300e-03 6.68730617e-01 3.52891833e-01
-5.09637520e-02 -6.85414791e-01 -9.31550622e-01 5.11423014e-02
9.90328670e-01 -7.20509142e-02 -2.07019970e-01 7.47535348e-01
5.84840417e-01 -1.65300712e-01 -1.34237063e+00 2.73857385e-01
4.54103380e-01 -6.46843076e-01 7.77059615e-01 -4.11812395e-01
2.89177716e-01 7.58557320e-02 -3.17694157e-01 -1.34540629e+00
-4.48607981e-01 -6.76207900e-01 1.27407312e-01 1.69787025e+00
5.80171525e-01 -8.96335542e-01 5.97011030e-01 6.31030917e-01
-4.38222170e-01 -1.27517366e+00 -1.03236115e+00 -1.20961070e+00
-1.33545563e-01 -6.98689878e-01 1.26468134e+00 9.77337003e-01
-2.32210562e-01 -1.11890137e-01 -5.42341769e-01 -3.33276205e-02
5.61714232e-01 7.20186055e-01 1.01822114e+00 -1.13331056e+00
1.84955075e-03 -1.98284864e-01 1.06555484e-01 -8.84040117e-01
-1.73948377e-01 -9.46082711e-01 -1.70809068e-02 -1.85443723e+00
2.50888735e-01 -1.48576224e+00 -1.65946096e-01 5.81435442e-01
3.69274974e-01 1.79356664e-01 1.13224514e-01 5.17423391e-01
-8.57260883e-01 1.03721172e-01 1.19268346e+00 9.28442851e-02
-1.35600999e-01 1.51975766e-01 -6.26524746e-01 5.16312301e-01
7.28736162e-01 -6.04508579e-01 3.96280847e-02 -5.73521078e-01
6.05745494e-01 1.99934140e-01 2.57298827e-01 -3.05142462e-01
1.75491899e-01 -6.30222440e-01 2.26585820e-01 -1.00948596e+00
1.42097339e-01 -1.02195072e+00 6.25460923e-01 2.84877926e-01
3.28802556e-01 8.55014443e-01 -1.46012202e-01 7.65615344e-01
-4.93667908e-02 3.15448076e-01 -2.76525497e-01 -2.80778408e-01
-8.61925900e-01 6.91000998e-01 -4.67191488e-01 -7.43603799e-04
1.39424956e+00 5.47365192e-03 -7.17517316e-01 -1.04973979e-01
-4.26369548e-01 7.43487179e-01 4.06006247e-01 7.57805169e-01
2.85801768e-01 -1.57169461e+00 -1.04717004e+00 6.86303675e-02
3.58134568e-01 -1.17257379e-01 2.85458416e-01 1.05837870e+00
-4.14632201e-01 5.34705937e-01 -9.04813334e-02 -5.96639514e-01
-8.77941668e-01 6.58783793e-01 -5.38852751e-01 -3.08181912e-01
-5.30517161e-01 3.48340034e-01 -1.57702863e-01 -5.17958462e-01
1.94405660e-01 -4.65447664e-01 2.55618185e-01 4.92796421e-01
3.02503049e-01 7.03377962e-01 2.06902236e-01 -4.78935003e-01
-6.94198966e-01 3.96250159e-01 9.74668413e-02 5.28906703e-01
1.91596782e+00 -2.49668241e-01 -3.95998389e-01 4.57689553e-01
1.21220767e+00 1.56047922e-02 -1.11282361e+00 -3.63916695e-01
3.52072328e-01 -8.37102234e-01 -3.25790346e-01 -1.01434767e+00
-1.20181108e+00 2.67547905e-01 5.97452298e-02 5.99325180e-01
1.22406924e+00 4.82798159e-01 1.34908116e+00 1.49708316e-01
8.86025250e-01 -1.07140112e+00 1.22227386e-01 4.48271930e-01
9.68877673e-01 -1.33976138e+00 -1.16588101e-01 -5.37194967e-01
-7.42916703e-01 3.99046242e-01 1.85965776e-01 1.04410194e-01
8.17038536e-01 6.33448005e-01 4.75445800e-02 -5.88650405e-01
-9.11666393e-01 5.25729135e-02 -3.12209696e-01 4.65073287e-01
1.44184917e-01 4.40768272e-01 -1.41937286e-01 1.17755973e+00
-6.86592683e-02 -1.02125898e-01 3.22766632e-01 1.14838588e+00
-1.41756058e-01 -1.03823960e+00 -2.11095646e-01 7.57606328e-01
-4.10695672e-01 -6.96744546e-02 1.77059323e-01 1.04720962e+00
2.99070150e-01 1.15670013e+00 3.89342487e-01 -2.40174308e-01
5.26611030e-01 -2.06415474e-01 1.20777689e-01 -4.89649981e-01
-9.11557019e-01 -1.65137611e-02 3.79359841e-01 -6.10407412e-01
1.26129106e-01 -9.08215106e-01 -1.18622911e+00 -1.01524186e+00
-9.78551880e-02 8.81093889e-02 9.44143295e-01 5.50948501e-01
5.89248121e-01 4.09933448e-01 1.32090628e+00 -7.66023457e-01
-4.07087147e-01 -8.81067097e-01 -1.04161882e+00 6.09448075e-01
1.48538098e-01 -8.65256786e-01 -2.29316607e-01 -1.01028167e-01] | [7.087385654449463, 2.8577821254730225] |
0dd0af01-6f78-410f-a6ce-779c495b458a | learning-to-steer-by-mimicking-features-from | 1811.02759 | null | http://arxiv.org/abs/1811.02759v1 | http://arxiv.org/pdf/1811.02759v1.pdf | Learning to Steer by Mimicking Features from Heterogeneous Auxiliary Networks | The training of many existing end-to-end steering angle prediction models
heavily relies on steering angles as the supervisory signal. Without learning
from much richer contexts, these methods are susceptible to the presence of
sharp road curves, challenging traffic conditions, strong shadows, and severe
lighting changes. In this paper, we considerably improve the accuracy and
robustness of predictions through heterogeneous auxiliary networks feature
mimicking, a new and effective training method that provides us with much
richer contextual signals apart from steering direction. Specifically, we train
our steering angle predictive model by distilling multi-layer knowledge from
multiple heterogeneous auxiliary networks that perform related but different
tasks, e.g., image segmentation or optical flow estimation. As opposed to
multi-task learning, our method does not require expensive annotations of
related tasks on the target set. This is made possible by applying contemporary
off-the-shelf networks on the target set and mimicking their features in
different layers after transformation. The auxiliary networks are discarded
after training without affecting the runtime efficiency of our model. Our
approach achieves a new state-of-the-art on Udacity and Comma.ai, outperforming
the previous best by a large margin of 12.8% and 52.1%, respectively.
Encouraging results are also shown on Berkeley Deep Drive (BDD) dataset. | ['Chen Change Loy', 'Yuenan Hou', 'Zheng Ma', 'Chunxiao Liu'] | 2018-11-07 | null | null | null | null | ['steering-control'] | ['computer-vision'] | [ 2.84884572e-02 1.31855039e-02 -3.32960814e-01 -5.92061162e-01
-6.01103008e-01 -5.48388720e-01 4.75419223e-01 -4.21527356e-01
-4.37092841e-01 9.03564334e-01 1.90313965e-01 -4.59338129e-01
2.16812432e-01 -7.08983004e-01 -9.91012335e-01 -8.18867505e-01
-7.62540624e-02 3.01953018e-01 3.94928724e-01 -3.89310718e-01
-7.64444917e-02 6.59416616e-01 -1.50505352e+00 1.13656983e-01
8.01939070e-01 1.42610157e+00 7.74685200e-03 9.12951410e-01
1.79002106e-01 8.72719526e-01 -2.81551987e-01 -2.82172322e-01
5.30728042e-01 8.78517851e-02 -3.40276480e-01 -2.47392029e-01
7.87113905e-01 -6.07028902e-01 -8.06182742e-01 3.73375565e-01
7.74362385e-01 6.35507852e-02 3.37779254e-01 -1.56594861e+00
-2.89868712e-01 2.76114732e-01 -3.94713938e-01 7.03119561e-02
-3.40823233e-01 5.25530636e-01 9.24821198e-01 -8.42151642e-01
4.84995723e-01 1.08051026e+00 7.37762928e-01 4.26106155e-01
-1.13689554e+00 -8.85382116e-01 2.39512593e-01 4.86777097e-01
-1.05064702e+00 -9.81482446e-01 7.38480031e-01 -2.75814712e-01
8.06867301e-01 1.71997756e-01 3.39954466e-01 1.42099535e+00
-3.15062962e-02 9.69073772e-01 1.07891738e+00 6.96190819e-02
-5.97481243e-02 3.26176703e-01 -1.38143785e-02 7.14472055e-01
-1.26819208e-01 2.91690320e-01 -6.95887566e-01 3.48380476e-01
5.45102298e-01 -3.23659152e-01 -3.63367349e-01 -6.51217163e-01
-1.20063138e+00 5.00502348e-01 5.21150887e-01 -5.44465721e-01
-2.21659645e-01 5.76095283e-01 5.51643789e-01 2.79677242e-01
3.71321768e-01 1.35607481e-01 -7.72542953e-01 -3.99688363e-01
-7.25127101e-01 -5.61975650e-02 8.51525068e-01 1.07172263e+00
8.94204199e-01 2.75851756e-01 -3.20471227e-01 7.27146387e-01
3.75330329e-01 9.18204486e-01 -5.79248555e-02 -1.24899638e+00
8.68536472e-01 -6.09357543e-02 1.54610172e-01 -1.00951970e+00
-6.60213470e-01 -8.53846669e-01 -9.55635309e-01 2.47674584e-01
6.40590370e-01 -5.33063650e-01 -9.32636380e-01 1.81458235e+00
2.71966845e-01 5.19031525e-01 4.49697785e-02 1.01400983e+00
7.62643516e-01 6.47779167e-01 -2.30534852e-01 1.85129762e-01
7.21112788e-01 -1.36640608e+00 -3.41997951e-01 -4.60778266e-01
6.88352287e-01 -6.16576791e-01 1.10658216e+00 5.64444482e-01
-7.73898184e-01 -6.17441654e-01 -1.11047268e+00 -1.42261326e-01
-2.89265513e-01 3.01641703e-01 7.75083125e-01 7.78577268e-01
-1.07958174e+00 4.95253116e-01 -5.57017922e-01 -3.50496694e-02
7.48722494e-01 3.84424627e-01 -1.84426978e-01 -9.14602652e-02
-1.24713933e+00 9.04141605e-01 -2.44406715e-01 1.29912019e-01
-9.50414062e-01 -1.02963006e+00 -7.11201668e-01 -2.64222529e-02
6.55390203e-01 -7.01662719e-01 1.05751395e+00 -8.89256239e-01
-2.05924511e+00 3.98522824e-01 -1.31455645e-01 -6.32691681e-01
8.95649612e-01 -6.06622338e-01 -4.17914540e-01 1.20196836e-02
-1.80419713e-01 7.12431967e-01 9.34366524e-01 -1.35597456e+00
-7.30817854e-01 -2.11845547e-01 -4.95583937e-03 -1.45058669e-02
-1.20523222e-01 -5.87366343e-01 -8.77431929e-01 -4.20505732e-01
-2.91678399e-01 -1.33162010e+00 -2.22335711e-01 5.26746929e-01
-4.80819017e-01 1.90111935e-01 1.17191231e+00 -4.24141973e-01
9.15380120e-01 -2.10084462e+00 -2.29401767e-01 4.01857495e-01
3.80910903e-01 3.10062319e-01 -4.89513725e-01 -9.29632410e-02
2.35120222e-01 -9.89618897e-02 2.27807403e-01 -4.50968862e-01
1.23474881e-01 3.82693559e-01 -4.65077311e-01 6.26996875e-01
1.37783155e-01 8.65716577e-01 -8.52107167e-01 -1.87431425e-01
3.84534180e-01 6.02474213e-01 -7.64077544e-01 1.04675613e-01
2.42138244e-02 6.66218519e-01 -2.82155037e-01 4.59240735e-01
5.16575694e-01 -2.59862363e-01 -1.50526255e-01 -3.49249512e-01
-6.72170008e-03 3.74118179e-01 -1.29876089e+00 1.67268312e+00
-1.15475714e+00 1.13706112e+00 7.83567280e-02 -9.93329823e-01
9.19268906e-01 7.55313411e-02 5.75061500e-01 -1.00494158e+00
4.91036065e-02 2.15749353e-01 9.73592699e-02 -3.88741076e-01
4.74220037e-01 3.40303779e-01 -1.67668745e-01 2.13456973e-01
-9.58400071e-02 -8.37186649e-02 -4.34298106e-02 2.16105029e-01
9.58870769e-01 4.67629999e-01 -2.17238888e-01 9.24169552e-03
4.51447874e-01 -3.12191576e-01 7.84810424e-01 7.49383152e-01
-3.91790897e-01 5.33672512e-01 5.50808311e-01 -4.98387218e-01
-1.13369048e+00 -1.12119591e+00 -4.38815616e-02 1.21245456e+00
1.70017451e-01 -2.02105284e-01 -1.91427886e-01 -5.65968752e-01
2.87309140e-01 7.24971771e-01 -5.81240118e-01 -1.78993925e-01
-5.78170180e-01 -6.90672755e-01 6.82251215e-01 7.23346293e-01
6.68164194e-01 -4.55693871e-01 -4.34381485e-01 2.80063897e-01
-1.86616063e-01 -1.64834702e+00 -4.44680840e-01 2.22145632e-01
-5.81369281e-01 -8.65443707e-01 -4.95777458e-01 -1.83777541e-01
1.46846682e-01 4.21963781e-01 1.23164988e+00 -2.20372587e-01
-5.29934019e-02 1.26345739e-01 1.69971585e-02 -3.78354728e-01
-2.33219326e-01 3.00243407e-01 8.95158723e-02 1.09540068e-01
-1.38011966e-02 -7.55969107e-01 -7.87784278e-01 5.68648219e-01
-2.18622208e-01 3.38503063e-01 7.60236084e-01 7.51022220e-01
5.54289401e-01 -6.25845313e-01 6.25433862e-01 -9.24342275e-01
6.15327209e-02 -5.23955524e-01 -6.32447660e-01 -1.50966063e-01
-7.28776515e-01 3.03363413e-01 9.57212508e-01 -1.99123368e-01
-1.01991248e+00 2.29419768e-01 2.89434250e-02 -5.30214250e-01
-2.37902835e-01 4.07018960e-02 -2.23570496e-01 -1.45280361e-01
6.97232425e-01 1.62163842e-02 1.15394793e-01 -1.54659137e-01
6.51585042e-01 6.81484759e-01 6.94923997e-01 -2.62761146e-01
1.11277151e+00 5.90150356e-01 4.52884763e-01 -6.45053864e-01
-1.04319894e+00 -4.18437123e-01 -6.64612770e-01 -3.37777346e-01
4.92777348e-01 -1.24274778e+00 -7.81744421e-01 6.38643682e-01
-8.32231224e-01 -7.95218885e-01 -2.56019458e-02 3.40569139e-01
-6.78044438e-01 2.03030750e-01 -2.23310530e-01 -3.83714318e-01
-3.65642875e-01 -1.14698184e+00 8.98270130e-01 1.26763955e-01
2.07467169e-01 -7.97948837e-01 -2.10438937e-01 6.37524486e-01
8.50115299e-01 3.03943247e-01 5.79501748e-01 -4.25008059e-01
-8.22951078e-01 -1.97927549e-01 -5.42415023e-01 5.06857932e-01
1.40100522e-02 1.05230249e-01 -1.20175004e+00 -5.39985336e-02
-7.64190674e-01 -4.86494064e-01 1.05456793e+00 3.26332688e-01
1.49061835e+00 -1.98457077e-01 -3.26735169e-01 1.05506980e+00
1.22159183e+00 -2.54466683e-01 6.14745140e-01 4.03508484e-01
1.06185377e+00 4.38258559e-01 6.33550227e-01 5.77857792e-01
7.07559764e-01 7.02090561e-01 6.88312113e-01 -3.25535059e-01
-4.29786295e-01 5.83035983e-02 4.30661857e-01 3.03233385e-01
-1.19825028e-01 -5.01059592e-01 -1.00686228e+00 4.14873123e-01
-1.92090797e+00 -8.95001471e-01 -3.26396108e-01 2.08525681e+00
7.94584811e-01 5.17158628e-01 -1.09389741e-02 1.74422413e-02
3.00656497e-01 4.06602859e-01 -7.54456937e-01 -3.09874564e-01
-1.95548415e-01 5.10553122e-02 1.11415088e+00 5.59892237e-01
-1.26197183e+00 1.00642049e+00 5.28986740e+00 6.39790058e-01
-1.52966177e+00 -2.41775438e-02 9.36455071e-01 -5.43667734e-01
-2.50435770e-01 -1.51785925e-01 -8.48365128e-01 4.30603325e-01
1.11545992e+00 4.14457321e-01 4.95423764e-01 8.63548636e-01
6.59386992e-01 -8.54008943e-02 -9.65702176e-01 9.00789261e-01
-1.45963291e-02 -1.40474808e+00 -3.84599924e-01 -5.18317036e-02
7.11040258e-01 7.31339753e-01 2.34533444e-01 5.74710548e-01
4.36411619e-01 -9.94942725e-01 5.65535426e-01 6.79552019e-01
9.62276995e-01 -7.12818563e-01 7.48448730e-01 1.40444010e-01
-9.25246418e-01 -1.01096034e-01 3.97077296e-03 8.44306499e-02
2.77697951e-01 7.33098567e-01 -8.33348036e-01 2.59754330e-01
6.82555854e-01 8.77058089e-01 -4.63599801e-01 1.02748823e+00
-2.86099374e-01 7.31651723e-01 -6.58768833e-01 7.72439614e-02
4.37916160e-01 4.25621159e-02 5.17921269e-01 1.09932637e+00
1.29807830e-01 -2.35617369e-01 3.87149751e-02 3.87044370e-01
-3.64082396e-01 -1.48409247e-01 -4.62694228e-01 5.91185331e-01
3.83086324e-01 1.43622684e+00 -4.97078598e-02 -2.84129620e-01
-7.61174083e-01 7.48037100e-01 1.71590492e-01 4.70653415e-01
-1.37231541e+00 -2.26956829e-01 9.24743474e-01 1.22722715e-01
4.68756437e-01 -3.57154608e-01 -4.49078172e-01 -1.05009580e+00
-1.93713102e-02 -5.66979170e-01 -2.03125402e-01 -6.16070271e-01
-9.85759020e-01 3.98932070e-01 -3.87973130e-01 -1.32703984e+00
-2.42367551e-01 -6.05702639e-01 -6.10302150e-01 7.61614025e-01
-2.29109335e+00 -1.13711166e+00 -7.95521617e-01 6.88021660e-01
3.88910890e-01 -3.02364498e-01 4.62219417e-01 5.60036838e-01
-7.32382655e-01 7.80830681e-01 3.40355098e-01 1.17310263e-01
1.35923362e+00 -1.24736178e+00 3.63689125e-01 8.53840768e-01
-1.13812797e-01 -1.59844205e-01 7.80737698e-01 1.94803756e-02
-1.44991767e+00 -1.36876464e+00 4.98767883e-01 -3.77546132e-01
8.54562461e-01 -4.82309878e-01 -4.36187506e-01 4.47836787e-01
-1.77956343e-01 7.98500240e-01 5.82594216e-01 2.31291339e-01
-4.43338811e-01 -6.66340947e-01 -6.44809723e-01 6.69376493e-01
1.24114716e+00 -3.44794482e-01 1.91839576e-01 2.86615491e-01
5.16726434e-01 -6.26994848e-01 -6.87890410e-01 4.18808222e-01
7.27268696e-01 -9.60057497e-01 1.11767685e+00 -6.63507223e-01
5.11253655e-01 -1.49058610e-01 -2.45243102e-01 -1.33868170e+00
-1.67661652e-01 -8.42771590e-01 -5.48383951e-01 8.49194109e-01
7.96878099e-01 -6.38592243e-01 1.01460087e+00 6.10518634e-01
-5.33336103e-01 -9.66091156e-01 -8.58457386e-01 -6.28522277e-01
-4.69576381e-02 -8.01437020e-01 3.68510067e-01 6.06243014e-01
-5.44216871e-01 4.38538462e-01 -7.62273610e-01 2.92229414e-01
5.51351428e-01 1.97987214e-01 1.41013575e+00 -1.14134598e+00
-3.05747986e-01 -3.34766358e-01 -2.60530114e-01 -1.47739863e+00
1.76889539e-01 -6.19344294e-01 2.24946171e-01 -1.17779732e+00
-3.96488219e-01 -9.08735216e-01 -3.13840210e-01 5.94384015e-01
3.03735165e-03 4.59986091e-01 3.34515870e-01 5.53437136e-02
-7.44483054e-01 8.06674004e-01 1.50834787e+00 -1.32413968e-01
-3.14877003e-01 3.71623755e-01 -5.60802460e-01 6.48381174e-01
1.06640697e+00 -2.05632910e-01 -4.59601432e-01 -5.91504693e-01
2.23650679e-01 4.62572649e-02 5.75984001e-01 -1.08214152e+00
2.03917935e-01 -1.25424817e-01 3.79175723e-01 -5.88102639e-01
4.89535838e-01 -8.89016330e-01 -3.66930872e-01 2.10099578e-01
-4.32562865e-02 -3.57304960e-01 4.07243639e-01 6.77103579e-01
1.41417667e-01 4.68416452e-01 8.27283442e-01 3.58420968e-01
-1.11722326e+00 6.35533452e-01 -1.36732355e-01 1.41078830e-01
9.09076810e-01 -2.55346566e-01 -5.41866839e-01 -7.09207892e-01
-7.45463848e-01 4.16168004e-01 1.00337222e-01 7.45194435e-01
2.57907957e-01 -1.12381446e+00 -6.87707603e-01 5.71741089e-02
1.65310539e-02 1.56438738e-01 1.61644384e-01 1.16175747e+00
-2.96359628e-01 5.10576606e-01 -2.40251482e-01 -7.40480602e-01
-9.54945505e-01 1.25076711e-01 4.22929913e-01 -1.48120925e-01
-3.75522137e-01 6.59855723e-01 -4.84569632e-02 -3.65386367e-01
2.66651183e-01 -3.07018399e-01 -5.12746070e-03 1.16044674e-02
1.12902205e-02 4.29554462e-01 8.52052867e-02 -6.60660446e-01
-3.45312446e-01 3.90510380e-01 4.08501439e-02 2.70019978e-01
1.53113019e+00 -4.58517402e-01 5.26620865e-01 2.02319682e-01
1.31831026e+00 1.43795803e-01 -1.97137916e+00 -4.47102875e-01
-5.09012938e-01 -4.97719407e-01 5.48999786e-01 -1.01960158e+00
-1.55762887e+00 9.58093166e-01 7.28726327e-01 -3.30974847e-01
1.05177367e+00 -2.68106520e-01 1.03775167e+00 6.80664539e-01
2.40874499e-01 -1.33634353e+00 1.03569031e-01 7.40212023e-01
7.68885612e-01 -1.82064879e+00 -3.87057029e-02 -4.51015711e-01
-6.56332254e-01 1.13883960e+00 8.37067902e-01 -2.48341393e-02
7.98540056e-01 4.38439280e-01 2.71049082e-01 2.98284233e-01
-9.53152299e-01 -1.64403468e-01 4.10264760e-01 6.02331400e-01
1.33742392e-01 -1.05135694e-01 3.24292034e-01 4.77226675e-01
-2.53560007e-01 -3.53287570e-02 7.04967141e-01 4.53688413e-01
-2.89981663e-01 -6.96784139e-01 3.14938277e-02 5.89973509e-01
-1.17900334e-01 -2.39581406e-01 9.44260284e-02 8.92719626e-01
2.41017453e-02 7.14670897e-01 1.33225277e-01 -5.69435060e-01
3.70755076e-01 -3.07716101e-01 -9.97212529e-03 -2.00357690e-01
-1.75223246e-01 -2.87469804e-01 5.34831643e-01 -1.10169256e+00
-3.07149887e-01 -6.50397241e-01 -1.13034666e+00 -6.46963775e-01
-1.58197358e-01 -5.74567378e-01 6.39629364e-01 1.00904298e+00
6.67108059e-01 6.22683585e-01 9.83511448e-01 -1.00779700e+00
-4.52721834e-01 -7.31670499e-01 -1.01075217e-01 4.37612683e-02
6.70716345e-01 -7.18929589e-01 -3.38802367e-01 -1.75983116e-01] | [8.198147773742676, -1.4124958515167236] |
a420d34d-94e7-423e-ac10-e17b6a82a532 | deep-single-image-portrait-relighting | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Zhou_Deep_Single-Image_Portrait_Relighting_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Zhou_Deep_Single-Image_Portrait_Relighting_ICCV_2019_paper.pdf | Deep Single-Image Portrait Relighting | Conventional physically-based methods for relighting portrait images need to solve an inverse rendering problem, estimating face geometry, reflectance and lighting. However, the inaccurate estimation of face components can cause strong artifacts in relighting, leading to unsatisfactory results. In this work, we apply a physically-based portrait relighting method to generate a large scale, high quality, "in the wild" portrait relighting dataset (DPR). A deep Convolutional Neural Network (CNN) is then trained using this dataset to generate a relit portrait image by using a source image and a target lighting as input. The training procedure regularizes the generated results, removing the artifacts caused by physically-based relighting methods. A GAN loss is further applied to improve the quality of the relit portrait image. Our trained network can relight portrait images with resolutions as high as 1024 x 1024. We evaluate the proposed method on the proposed DPR datset, Flickr portrait dataset and Multi-PIE dataset both qualitatively and quantitatively. Our experiments demonstrate that the proposed method achieves state-of-the-art results. Please refer to https://zhhoper.github.io/dpr.html for dataset and code.
| [' David W. Jacobs', ' Kalyan Sunkavalli', ' Sunil Hadap', 'Hao Zhou'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['single-image-portrait-relighting'] | ['computer-code'] | [ 3.67748857e-01 -1.93308443e-01 3.75968516e-01 -3.57684851e-01
-7.96107292e-01 -4.00306940e-01 3.67405444e-01 -6.76118195e-01
1.99270591e-01 8.91694784e-01 -2.00839937e-02 -5.51322103e-02
2.64939815e-01 -1.19482911e+00 -9.35534298e-01 -7.60539353e-01
4.63665068e-01 9.92803425e-02 -3.72726470e-01 -3.12542528e-01
5.61037660e-02 6.90035641e-01 -1.45600080e+00 2.75241226e-01
1.06689358e+00 1.02843499e+00 -5.88893294e-02 7.10391223e-01
1.14222474e-01 6.03620589e-01 -7.98825085e-01 -7.43769825e-01
5.45041680e-01 -7.16714084e-01 -4.12844867e-01 3.13909292e-01
8.89245689e-01 -7.76488125e-01 -5.86325705e-01 8.87836218e-01
4.39467698e-01 4.67828326e-02 5.70107102e-01 -1.42981148e+00
-1.36765289e+00 1.73912019e-01 -1.04924309e+00 -4.27853197e-01
3.94114554e-01 3.09837610e-01 5.71308970e-01 -7.81453073e-01
6.02749586e-01 1.49399781e+00 8.46065819e-01 6.74108922e-01
-1.29036438e+00 -1.00679708e+00 -3.49072129e-01 -1.64999276e-01
-1.77435875e+00 -4.96763170e-01 1.27977061e+00 3.45588587e-02
2.67941803e-01 6.11337125e-01 7.89255500e-01 1.15860653e+00
2.52396315e-01 3.89950216e-01 1.39979947e+00 -1.21795163e-01
-2.63743959e-02 1.35435820e-01 -4.89646554e-01 7.45364010e-01
-6.53423592e-02 4.82963510e-02 -4.82581496e-01 3.79576907e-02
1.13305271e+00 4.58494090e-02 -3.97946894e-01 2.71209210e-01
-9.18965220e-01 6.32546246e-01 9.00172412e-01 3.21721397e-02
-1.32943094e-01 4.53210026e-01 3.34570967e-02 1.35863930e-01
7.60276973e-01 -1.27781192e-02 2.04540744e-01 2.89277047e-01
-1.06154442e+00 4.64552157e-02 4.40167904e-01 8.80365133e-01
8.09584856e-01 3.53489041e-01 -1.87938791e-02 1.06384528e+00
4.11250710e-01 9.18531060e-01 -1.69152975e-01 -1.00057793e+00
2.70481646e-01 3.58251125e-01 9.99896452e-02 -1.21917677e+00
1.11737713e-01 -1.10642508e-01 -1.32896197e+00 4.65670109e-01
5.26668876e-02 -1.78211287e-01 -1.10252202e+00 1.29286373e+00
6.34227157e-01 1.68005422e-01 -1.98696721e-02 1.13852072e+00
1.17950046e+00 1.12562108e+00 -2.63399363e-01 7.85956830e-02
1.06585920e+00 -7.66545594e-01 -8.91604722e-01 1.37549683e-01
5.47804534e-02 -9.75113511e-01 1.26849687e+00 4.74177241e-01
-9.08237576e-01 -5.16885221e-01 -9.26210225e-01 -6.15684867e-01
-8.47629160e-02 4.37372208e-01 3.74519855e-01 6.29226327e-01
-1.06887555e+00 6.48977041e-01 -3.60692948e-01 -2.14108825e-01
8.03657651e-01 7.45439678e-02 -2.18276575e-01 -2.82250375e-01
-9.70785081e-01 4.14905250e-01 -9.47825052e-03 4.17767674e-01
-1.12741554e+00 -8.99604559e-01 -7.89380908e-01 -2.35224515e-01
-8.85699391e-02 -4.68015194e-01 6.06808960e-01 -1.42335415e+00
-1.64915586e+00 8.65517795e-01 1.42704204e-01 -6.31394610e-02
6.65066004e-01 2.34067626e-02 -5.47457278e-01 1.22442603e-01
-1.70495421e-01 8.46629918e-01 1.05000818e+00 -1.74832761e+00
3.50245796e-02 -1.63003802e-01 -1.64916664e-01 1.22025087e-01
-1.09269500e-01 -2.89683223e-01 -3.76780093e-01 -7.79028594e-01
-2.24226490e-01 -6.86373889e-01 3.84312153e-01 5.81180871e-01
-7.50910223e-01 3.47173959e-01 1.16292572e+00 -9.96279836e-01
6.95544243e-01 -2.29782319e+00 -1.37409061e-01 5.07926904e-02
1.13067307e-01 8.47565085e-02 -5.03934264e-01 4.24764246e-01
-5.68282530e-02 3.35758328e-01 -3.20767462e-01 -6.25040710e-01
-1.74481854e-01 2.19155028e-01 -3.08730394e-01 6.36127472e-01
2.25199386e-01 8.71713102e-01 -5.40494621e-01 -5.93567610e-01
4.25439030e-01 1.33566511e+00 -2.12813661e-01 1.83863670e-01
-2.54182994e-01 6.46322191e-01 -1.54411927e-01 8.67743492e-01
1.29810703e+00 1.20995104e-01 -1.95514679e-01 -5.02836764e-01
-1.66817363e-02 -2.58851767e-01 -8.05410147e-01 1.70359337e+00
-5.87978125e-01 9.01739597e-01 2.47872863e-02 -2.69135475e-01
1.32953894e+00 5.98425977e-02 2.39926115e-01 -7.97350764e-01
3.42671961e-01 4.84652855e-02 -5.61860621e-01 -3.87407571e-01
4.61613208e-01 -2.76055932e-01 1.77848503e-01 2.05977052e-01
-3.13867271e-01 -4.27175164e-01 -2.26287290e-01 5.39298775e-03
6.56020522e-01 4.54809427e-01 -2.67463446e-01 -5.58638200e-02
2.87775785e-01 -1.36427358e-02 4.38320875e-01 -5.91188259e-02
1.95761874e-01 1.15230560e+00 3.77997279e-01 -6.61124468e-01
-1.30079317e+00 -1.06070852e+00 -2.69007295e-01 4.85357910e-01
2.32005373e-01 3.75396982e-02 -1.04933786e+00 -4.67674404e-01
-9.13603678e-02 8.40247393e-01 -1.02185941e+00 -3.58340405e-02
-3.62434506e-01 -6.33667171e-01 5.15654683e-01 -4.62299399e-03
1.31164801e+00 -1.00193334e+00 -4.26471420e-02 -2.71153927e-01
-2.95203179e-01 -8.71435881e-01 -7.88934886e-01 -7.15430737e-01
-5.38971961e-01 -8.08075964e-01 -8.27344298e-01 -5.55865407e-01
8.21756542e-01 4.32714254e-01 1.16190660e+00 3.90767157e-01
-5.05848408e-01 4.54949178e-02 -2.00541750e-01 -3.61199304e-02
-5.25340080e-01 -3.05823177e-01 -3.48992854e-01 5.71983278e-01
-3.41155112e-01 -7.04388738e-01 -9.64677513e-01 4.33128893e-01
-1.20149624e+00 6.99602485e-01 4.74811047e-01 5.96189618e-01
6.88221812e-01 2.21554413e-01 3.07166815e-01 -6.22546077e-01
2.87329704e-01 -2.50159174e-01 -5.97879648e-01 1.85964733e-01
-1.01007581e-01 -3.83066416e-01 7.76883304e-01 -5.77835262e-01
-1.31563270e+00 -1.80823967e-01 -3.18502635e-01 -7.62978613e-01
-1.27443671e-01 -1.43348932e-01 -4.61628646e-01 -4.31519806e-01
6.29247367e-01 2.89813519e-01 -5.04681952e-02 -3.19204330e-01
2.90975630e-01 6.36878014e-01 3.66420895e-01 -2.65787601e-01
1.16443264e+00 7.34190106e-01 1.07876852e-01 -9.44287956e-01
-7.25459039e-01 5.45619726e-01 -2.86855936e-01 -6.02572381e-01
9.03935730e-01 -8.74883831e-01 -8.43962014e-01 6.22823298e-01
-9.74257708e-01 -7.63902307e-01 -4.01899636e-01 -1.03429355e-01
-3.73604506e-01 7.36760497e-02 -5.93629658e-01 -8.27161193e-01
-5.78200042e-01 -9.26607966e-01 1.38423812e+00 5.32484591e-01
2.65419126e-01 -7.66862333e-01 3.66135351e-02 7.59711206e-01
3.95329505e-01 9.95675385e-01 5.40946066e-01 5.00643671e-01
-8.29369187e-01 3.07681262e-02 -5.57960689e-01 3.59451652e-01
3.13908249e-01 2.86215156e-01 -1.08874738e+00 -1.70370802e-01
-2.54623741e-01 -3.70409995e-01 5.08544922e-01 2.00193495e-01
1.31246829e+00 -5.45778155e-01 1.16797127e-01 1.08282614e+00
1.89009702e+00 -2.85242256e-02 1.12775481e+00 -1.57913178e-01
1.15703034e+00 4.04082745e-01 5.42046607e-01 4.49464798e-01
3.20219308e-01 4.94540006e-01 5.71214497e-01 -5.29382288e-01
-7.64749348e-01 -6.60414279e-01 2.97490776e-01 4.08336252e-01
-1.05157621e-01 -6.67853534e-01 -5.46477795e-01 3.11697692e-01
-1.31421340e+00 -6.95790589e-01 -3.61402601e-01 2.08398724e+00
8.48049521e-01 -4.88920033e-01 -1.91535637e-01 1.84800643e-02
7.31594801e-01 1.56818151e-01 -7.76435733e-01 -3.44002753e-01
-2.69146472e-01 -2.14193929e-02 3.58390898e-01 5.32949269e-01
-5.39285958e-01 9.30915356e-01 4.94997644e+00 1.00706112e+00
-1.35015678e+00 1.49986878e-01 1.27741611e+00 -1.43347308e-01
-7.49959111e-01 -2.59507209e-01 -3.53340805e-01 4.39181715e-01
5.79670906e-01 2.27877378e-01 9.43980694e-01 4.46877956e-01
5.28133690e-01 -2.59619445e-01 -6.34786308e-01 1.36950326e+00
4.92823631e-01 -1.27535784e+00 3.97666246e-01 2.30531633e-01
9.64859247e-01 -4.75214183e-01 4.16535735e-01 -1.31276369e-01
1.43883809e-01 -1.36521411e+00 6.18853033e-01 8.27927232e-01
1.41045582e+00 -1.03247833e+00 3.13021302e-01 -1.24032833e-01
-1.03824842e+00 4.15593624e-01 -3.92907351e-01 3.85072351e-01
-1.99279431e-02 7.84715414e-01 -4.82557118e-01 6.13504469e-01
7.02246010e-01 8.53959441e-01 -7.08133221e-01 5.72904885e-01
-9.22705829e-02 5.03833711e-01 -2.19956368e-01 1.92171067e-01
-1.37963384e-01 -6.36523962e-01 2.81314880e-01 6.50164187e-01
5.57662725e-01 1.94181412e-01 -3.88837337e-01 1.25304019e+00
-6.50982678e-01 9.99165475e-02 -7.13018775e-01 6.50787055e-02
1.27056360e-01 1.55313289e+00 -6.39009535e-01 -2.44538426e-01
5.96612468e-02 1.50932288e+00 7.84422904e-02 5.59996843e-01
-1.33945191e+00 -2.87197590e-01 4.47694480e-01 3.71465832e-01
-4.65851836e-02 -4.06032391e-02 -2.46514112e-01 -1.21487987e+00
-9.01815295e-02 -9.21719193e-01 -2.14254975e-01 -1.53569233e+00
-1.28624082e+00 7.15424478e-01 -1.93123981e-01 -1.15738022e+00
4.77824152e-01 -1.90385625e-01 -7.15157509e-01 8.91777754e-01
-1.46868503e+00 -1.71030915e+00 -8.94708216e-01 6.67980313e-01
5.01284122e-01 2.99668729e-01 6.31056607e-01 3.57425272e-01
-7.36949563e-01 5.48574388e-01 2.31304318e-01 1.58369988e-01
7.62217402e-01 -9.92114663e-01 2.90102303e-01 6.67305052e-01
-5.27197532e-02 4.52745035e-02 6.31371200e-01 -7.28005946e-01
-1.86017752e+00 -1.47290039e+00 2.08154947e-01 -1.52740434e-01
5.48608378e-02 -5.43350339e-01 -7.43195057e-01 5.02662539e-01
6.37702763e-01 2.17299387e-01 5.35863757e-01 -5.96202016e-01
-3.48962307e-01 -7.22725868e-01 -1.83489966e+00 7.27567434e-01
9.92404580e-01 -2.91925609e-01 1.74059302e-01 4.45658207e-01
7.00770199e-01 -3.97395551e-01 -1.09654033e+00 -8.01494718e-02
7.11817324e-01 -1.09594345e+00 9.16134715e-01 4.09780592e-01
8.49784851e-01 -4.55656230e-01 -4.72245142e-02 -1.30683100e+00
-1.91610418e-02 -8.35402846e-01 2.75938213e-01 1.63114142e+00
5.36808036e-02 -7.20799327e-01 5.30569494e-01 5.13494670e-01
1.64074123e-01 -6.00264370e-01 -8.17202508e-01 -3.64815235e-01
2.65608095e-02 -1.88887697e-02 1.03197503e+00 8.46356452e-01
-9.79033291e-01 2.45655239e-01 -8.12611222e-01 1.09530091e-01
1.11542630e+00 1.95516303e-01 9.97093201e-01 -8.39402199e-01
1.99084915e-02 1.08961836e-01 5.74532934e-02 -5.68315208e-01
-1.03956714e-01 -4.94251013e-01 -5.19083701e-02 -1.49050188e+00
1.88476574e-02 -5.97575188e-01 3.20041597e-01 4.17249173e-01
1.19083606e-01 1.27419090e+00 2.25426152e-01 1.19688183e-01
-2.96245385e-02 9.05789375e-01 1.93075120e+00 -1.76146269e-01
-6.95728064e-02 -5.28663874e-01 -7.00527430e-01 3.23505402e-01
1.07489800e+00 -3.47348124e-01 -4.05824453e-01 -6.42620027e-01
2.43644595e-01 2.05902889e-01 7.98799813e-01 -1.07679939e+00
-3.82625282e-01 -1.87450498e-01 9.85459983e-01 -5.81116438e-01
7.66548038e-01 -8.56133342e-01 9.26145017e-01 1.92527622e-01
-1.21119007e-01 -2.50439346e-01 3.30421567e-01 3.80489111e-01
1.80105969e-01 2.06359744e-01 1.20112634e+00 1.09951973e-01
1.77837815e-02 5.46893716e-01 1.48431078e-01 -1.75767034e-01
1.04217172e+00 -3.05130363e-01 -3.82139772e-01 -4.80946273e-01
-1.87227860e-01 -1.89767882e-01 8.66032481e-01 1.44248873e-01
9.72230613e-01 -1.73493981e+00 -9.11369562e-01 3.39233428e-01
-1.76409349e-01 3.76234315e-02 7.13795364e-01 4.76480484e-01
-9.92841780e-01 -3.98927003e-01 -3.50921839e-01 -4.09809470e-01
-1.35026324e+00 4.66398001e-01 4.21739966e-01 2.24131033e-01
-8.47224474e-01 4.28157508e-01 2.38758802e-01 -3.92301559e-01
-1.59592316e-01 -1.93320274e-01 2.07930222e-01 -2.32034832e-01
4.80225950e-01 1.87324136e-01 -3.07923198e-01 -7.96598256e-01
-4.20131013e-02 9.39938784e-01 2.74422735e-01 -1.39654115e-01
1.31851053e+00 -3.17597836e-01 -1.95592910e-01 1.91352904e-01
1.36644053e+00 2.72931196e-02 -1.47566569e+00 2.06796363e-01
-1.20274496e+00 -1.01871741e+00 2.96732485e-01 -8.85200322e-01
-1.59470046e+00 9.08086598e-01 7.40888000e-01 -2.08289418e-02
1.55602622e+00 -3.88299197e-01 1.16459620e+00 -1.10604100e-01
3.09517413e-01 -6.30977631e-01 1.43976092e-01 -2.14506298e-01
1.44362903e+00 -1.20077181e+00 1.47992402e-01 -5.26759863e-01
-6.33808851e-01 1.05179906e+00 5.14782548e-01 -4.86458510e-01
4.29046512e-01 1.64826080e-01 1.38641670e-01 -7.13963285e-02
-2.99284279e-01 2.34707370e-01 1.66608587e-01 5.55603027e-01
1.98368162e-01 5.85629530e-02 1.19787388e-01 -1.39359936e-01
-3.61602873e-01 1.74765348e-01 5.45432568e-01 4.76941824e-01
1.16196632e-01 -7.51122236e-01 -8.68003368e-01 3.27861220e-01
-1.22802936e-01 -1.10598601e-01 -4.66862261e-01 8.26882899e-01
2.90929765e-01 9.03337419e-01 1.74941137e-01 -2.68570125e-01
4.45761532e-02 -4.43897545e-01 6.91343069e-01 -8.74936879e-02
-5.70100129e-01 1.35347456e-01 9.80357826e-03 -4.52068448e-01
-3.15028667e-01 -1.04531638e-01 -8.88323009e-01 -1.03713644e+00
-6.78257644e-02 -2.07966596e-01 6.72408581e-01 3.88861895e-01
3.83888990e-01 5.68653524e-01 9.87185180e-01 -1.00953877e+00
3.51262152e-01 -9.60214913e-01 -6.74675882e-01 4.11817133e-01
3.56594831e-01 -3.44640374e-01 -3.16157550e-01 2.97775358e-01] | [12.396029472351074, -0.41851750016212463] |
346aa442-f0ba-4c82-8db7-cb494c9a548d | sinet-a-scale-insensitive-convolutional | 1804.00433 | null | http://arxiv.org/abs/1804.00433v2 | http://arxiv.org/pdf/1804.00433v2.pdf | SINet: A Scale-insensitive Convolutional Neural Network for Fast Vehicle Detection | Vision-based vehicle detection approaches achieve incredible success in
recent years with the development of deep convolutional neural network (CNN).
However, existing CNN based algorithms suffer from the problem that the
convolutional features are scale-sensitive in object detection task but it is
common that traffic images and videos contain vehicles with a large variance of
scales. In this paper, we delve into the source of scale sensitivity, and
reveal two key issues: 1) existing RoI pooling destroys the structure of small
scale objects, 2) the large intra-class distance for a large variance of scales
exceeds the representation capability of a single network. Based on these
findings, we present a scale-insensitive convolutional neural network (SINet)
for fast detecting vehicles with a large variance of scales. First, we present
a context-aware RoI pooling to maintain the contextual information and original
structure of small scale objects. Second, we present a multi-branch decision
network to minimize the intra-class distance of features. These lightweight
techniques bring zero extra time complexity but prominent detection accuracy
improvement. The proposed techniques can be equipped with any deep network
architectures and keep them trained end-to-end. Our SINet achieves
state-of-the-art performance in terms of accuracy and speed (up to 37 FPS) on
the KITTI benchmark and a new highway dataset, which contains a large variance
of scales and extremely small objects. | ['Pheng-Ann Heng', 'Hao Chen', 'Yongjie Xiao', 'Xuemiao Xu', 'Xiaowei Hu', 'Shengfeng He', 'Jing Qin'] | 2018-04-02 | null | null | null | null | ['fast-vehicle-detection'] | ['computer-vision'] | [ 8.73000994e-02 -4.89508420e-01 -6.18616678e-02 -3.40013415e-01
-4.80335742e-01 -3.99817616e-01 2.83357233e-01 -2.55595863e-01
-6.23311162e-01 1.79167151e-01 -3.22489738e-01 -2.84239441e-01
1.42659202e-01 -8.49179804e-01 -9.27057743e-01 -6.53545976e-01
-3.10809821e-01 -1.34255692e-01 1.28043485e+00 -5.01396000e-01
2.06804380e-01 8.11629951e-01 -1.65680528e+00 3.06115478e-01
5.95668077e-01 1.43248463e+00 2.17926785e-01 6.57430232e-01
-5.03671095e-02 8.09748709e-01 -5.67755938e-01 -2.33339369e-01
6.85849607e-01 6.08598292e-02 -2.01569036e-01 -1.37691915e-01
9.30674672e-01 -6.67888820e-01 -8.26163650e-01 1.19443858e+00
5.93643010e-01 -1.53252065e-01 3.73581588e-01 -1.52846742e+00
-5.05297303e-01 2.38367662e-01 -9.00220513e-01 9.70963001e-01
-4.94894922e-01 2.80829489e-01 5.67482054e-01 -9.76124644e-01
2.83776641e-01 1.46317780e+00 8.74409437e-01 2.39755824e-01
-8.08776975e-01 -9.77294266e-01 7.16356859e-02 4.66340750e-01
-1.56515837e+00 -4.39088821e-01 7.14463174e-01 -3.51844162e-01
1.00763619e+00 1.01947412e-01 3.94358933e-01 6.85642779e-01
3.92071515e-01 6.77737653e-01 9.13618088e-01 -2.36871671e-02
7.20063895e-02 9.62724164e-02 2.93794759e-02 7.22284377e-01
5.82025766e-01 1.53382257e-01 -2.33100146e-01 3.26977432e-01
8.82641435e-01 3.02435368e-01 2.64223158e-01 -4.61018711e-01
-1.12557411e+00 8.67792189e-01 8.47233713e-01 8.09261054e-02
-4.60090861e-03 4.52848196e-01 6.88849866e-01 3.06772977e-01
1.58992454e-01 -3.06515340e-02 -5.09949982e-01 1.47533357e-01
-6.70576990e-01 1.90046296e-01 1.51661858e-01 1.02837694e+00
7.63103664e-01 2.39140615e-01 -3.28415573e-01 6.19014561e-01
6.41749427e-02 7.87918568e-01 3.96813989e-01 -6.75106645e-01
7.61200964e-01 7.32884288e-01 -8.30605179e-02 -1.34269452e+00
-5.93071997e-01 -4.39678401e-01 -8.07546377e-01 6.51260734e-01
3.93744737e-01 -1.05078228e-01 -9.52739954e-01 1.62935543e+00
4.03948188e-01 8.79445449e-02 8.93719569e-02 1.00032353e+00
1.12930393e+00 5.83134353e-01 4.59310114e-02 2.77397949e-02
1.65416181e+00 -9.45041001e-01 -4.85917181e-01 -5.65379918e-01
4.31234121e-01 -6.87710226e-01 7.43737578e-01 -9.80786383e-02
-8.01784754e-01 -8.93388748e-01 -1.32670856e+00 -1.36052325e-01
-6.22671485e-01 2.99061239e-01 6.28268242e-01 8.43398750e-01
-8.22754562e-01 3.84264529e-01 -6.61247671e-01 -5.82924008e-01
7.53014326e-01 3.28370333e-01 -2.71496683e-01 9.03502181e-02
-9.81740296e-01 8.00240636e-01 2.99820393e-01 2.14009389e-01
-7.55892754e-01 -5.11149049e-01 -6.55528188e-01 1.48799300e-01
5.22116303e-01 -9.56029147e-02 9.18806493e-01 -1.09518158e+00
-9.85248327e-01 5.41370690e-01 -4.15386539e-03 -5.35797775e-01
7.57652700e-01 -3.48621234e-02 -4.56818461e-01 3.27210128e-01
3.64146620e-01 7.81461239e-01 8.43645334e-01 -8.13806534e-01
-1.16580093e+00 -3.36073518e-01 4.24599051e-02 -1.23898663e-01
-4.12394851e-01 3.66915643e-01 -4.19517308e-01 -5.63846052e-01
1.61322072e-01 -8.32260311e-01 -3.02494895e-02 4.29836601e-01
-6.87048584e-02 -3.67395461e-01 1.27947736e+00 -3.46242815e-01
1.08843517e+00 -2.34414768e+00 -6.04631722e-01 -2.47985289e-01
3.03967625e-01 5.79937458e-01 -2.14849636e-01 6.02846332e-02
2.47812718e-02 6.32648468e-02 9.81386900e-02 4.50312160e-02
-1.66453853e-01 1.21869496e-04 -3.42885971e-01 6.27075195e-01
5.84698319e-01 9.29436803e-01 -6.66969061e-01 -7.02919900e-01
3.01742613e-01 3.49269539e-01 -2.86678731e-01 2.62091774e-02
3.53555501e-01 -1.23603195e-01 -4.80714411e-01 8.89042974e-01
1.14245105e+00 -3.57892700e-02 -2.97327369e-01 -4.55890387e-01
-5.03176689e-01 -1.95811808e-01 -1.37946022e+00 9.67799246e-01
1.62493557e-01 9.58539963e-01 -1.00565448e-01 -9.42182004e-01
8.68129969e-01 -1.50535971e-01 3.06442797e-01 -8.71769011e-01
3.50120157e-01 3.79849851e-01 1.44033954e-01 -6.78050160e-01
4.34519112e-01 1.58527195e-01 -1.68261752e-02 -1.77303240e-01
-1.95163116e-01 3.17915291e-01 3.91574711e-01 1.53539956e-01
1.17658854e+00 -4.25519824e-01 2.07245916e-01 -2.65947849e-01
3.11351776e-01 -9.32221934e-02 7.96525419e-01 7.85135329e-01
-7.74453998e-01 6.60163581e-01 6.30666077e-01 -7.97037005e-01
-1.10988104e+00 -8.65832925e-01 -2.38690361e-01 1.26875556e+00
5.89048564e-01 3.61155234e-02 -4.81944531e-01 -7.94047892e-01
1.83342516e-01 -6.29893504e-03 -7.99465060e-01 -1.78325176e-01
-8.82752478e-01 -7.19196200e-01 6.94814384e-01 1.06459117e+00
1.00560153e+00 -8.41389775e-01 -9.65445817e-01 1.27320245e-01
1.99070156e-01 -1.53375554e+00 -5.91067731e-01 1.59482405e-01
-6.95332646e-01 -1.10337567e+00 -3.99689019e-01 -1.10661113e+00
5.10757446e-01 1.10257447e+00 7.05728471e-01 7.08250154e-04
-5.58866322e-01 -1.74309954e-01 -1.73865035e-01 -5.68366349e-01
8.46124161e-03 -8.32698122e-02 3.42330858e-02 9.78816897e-02
4.05484915e-01 -3.26566756e-01 -1.00451374e+00 6.54078543e-01
-7.63185143e-01 -1.24779582e-01 8.61967266e-01 5.14730453e-01
3.43764752e-01 1.81572378e-01 6.68118715e-01 -1.09910607e-01
5.81550598e-02 -1.91704929e-01 -8.23125124e-01 4.54292223e-02
-3.79786670e-01 -2.56770253e-01 4.56973970e-01 -6.70886278e-01
-7.50780761e-01 -3.26385759e-02 1.13773443e-01 -4.25108999e-01
-3.31137702e-02 -2.44997248e-01 -1.37935206e-01 -4.64853376e-01
7.04684377e-01 1.71535164e-01 -4.18867245e-02 -2.55600631e-01
1.88920960e-01 6.44344330e-01 3.90258789e-01 -5.44988364e-03
8.83931935e-01 8.50297928e-01 1.36252210e-01 -8.25165510e-01
-5.86410701e-01 -5.32449543e-01 -6.20426953e-01 -3.08499187e-01
8.91607881e-01 -1.14776194e+00 -5.99962771e-01 6.16750717e-01
-1.10500336e+00 -7.93244988e-02 -1.19269699e-01 3.53566110e-01
-1.40616551e-01 4.00834829e-01 -6.13512456e-01 -6.48625970e-01
-5.42696178e-01 -1.23595190e+00 1.03709280e+00 5.25004804e-01
5.18491268e-01 -2.89298058e-01 -4.36690688e-01 8.87477845e-02
8.61839950e-01 3.00360471e-01 5.37746489e-01 -5.04327059e-01
-8.13325465e-01 -5.02733946e-01 -1.08559716e+00 3.64165425e-01
-1.31238103e-01 7.40619898e-02 -7.50318825e-01 -5.16994774e-01
-3.92787546e-01 -1.86318442e-01 1.22817528e+00 4.07968372e-01
9.84350145e-01 -9.29014981e-02 -4.31986898e-01 6.44905329e-01
1.59214783e+00 1.82692960e-01 5.67667305e-01 5.67911923e-01
7.82377422e-01 3.97076666e-01 4.15430129e-01 1.39139339e-01
3.01396012e-01 6.11434877e-01 6.40810311e-01 -2.31117606e-01
-4.73569363e-01 3.19240317e-02 3.68269682e-01 2.65016288e-01
-3.06448876e-03 6.21465333e-02 -5.81833720e-01 6.56445742e-01
-1.82685125e+00 -1.02065611e+00 -2.04081491e-01 2.09109163e+00
3.60492706e-01 5.30189157e-01 3.78002048e-01 -7.43194520e-02
9.14296746e-01 1.16352133e-01 -7.03606963e-01 -1.97300047e-01
-3.68018746e-01 -3.45340908e-01 1.06147027e+00 -1.65564194e-01
-1.50136089e+00 9.84611392e-01 5.83038521e+00 9.46688294e-01
-1.44619811e+00 1.97400957e-01 5.07981002e-01 -1.30091548e-01
5.63967705e-01 -3.45722049e-01 -1.14647865e+00 4.63174909e-01
6.22285604e-01 9.86665040e-02 -5.09970523e-02 1.35492337e+00
1.54220164e-01 1.37659023e-02 -6.68997943e-01 8.91856074e-01
4.77576815e-02 -1.22271597e+00 -1.87305048e-01 -1.18620582e-01
6.33313119e-01 5.39391935e-01 1.00666489e-02 5.65522015e-01
-1.52142152e-01 -7.38179445e-01 1.07919562e+00 -9.91973579e-02
7.71663785e-01 -8.74844551e-01 9.19613957e-01 7.13155046e-02
-1.68417788e+00 -4.92605656e-01 -8.55206132e-01 6.82523921e-02
-1.63215041e-01 3.42288315e-01 -4.38703090e-01 1.09723993e-02
1.12157762e+00 4.36745197e-01 -1.00639045e+00 1.32628715e+00
1.37910903e-01 4.40899104e-01 -3.64950627e-01 -1.53492421e-01
3.11091572e-01 2.47035503e-01 5.23963928e-01 1.64278197e+00
3.38719189e-01 -1.03793003e-01 2.34302148e-01 6.53871119e-01
-8.55029821e-02 -1.00143321e-01 -4.60248709e-01 3.34735781e-01
5.43665826e-01 1.46605289e+00 -1.21594191e+00 -4.69147712e-01
-6.55769289e-01 5.92424393e-01 2.49263585e-01 2.57006466e-01
-1.15347111e+00 -6.79721951e-01 7.49231040e-01 2.79537082e-01
9.85804617e-01 -2.13171899e-01 -2.71662891e-01 -8.51982236e-01
2.37407759e-01 -5.42937219e-01 1.22084439e-01 -3.90189320e-01
-1.03382337e+00 5.39015770e-01 -1.00891657e-01 -1.34956300e+00
4.01686579e-01 -8.38209927e-01 -7.56320834e-01 3.10071200e-01
-1.87519693e+00 -1.20884502e+00 -6.01538301e-01 4.35333759e-01
7.62842894e-01 -1.83843672e-01 3.82530242e-02 4.85530019e-01
-8.69256854e-01 8.32031965e-01 1.10775076e-01 6.23119175e-01
6.80720270e-01 -9.59474862e-01 6.00639760e-01 1.06473911e+00
-4.06141013e-01 4.22843784e-01 5.39949119e-01 -5.54774523e-01
-1.61278248e+00 -1.50879252e+00 6.52650297e-01 -1.80744216e-01
6.43623233e-01 -4.12866861e-01 -7.51381278e-01 4.62787151e-01
2.80497074e-02 7.52638936e-01 -5.63394018e-02 -3.71399909e-01
-6.92986906e-01 -7.15426803e-01 -1.16553462e+00 3.83670717e-01
1.05999911e+00 -1.96990475e-01 -2.98789680e-01 2.78366923e-01
7.75604248e-01 -2.47379974e-01 -2.84154594e-01 7.51924038e-01
6.77346051e-01 -1.08890200e+00 1.04275525e+00 -4.47121352e-01
1.68790668e-01 -6.70353591e-01 -2.10803166e-01 -5.64562082e-01
-5.29189169e-01 -3.22816312e-01 -2.54358016e-02 1.04681134e+00
1.31285787e-01 -6.46083832e-01 6.31205738e-01 2.45280281e-01
-1.76968351e-01 -7.34686136e-01 -1.29692817e+00 -1.05114532e+00
-8.48732069e-02 -2.67855257e-01 4.54545856e-01 4.30675596e-01
-5.19613445e-01 2.41320103e-01 -2.89301306e-01 3.97262126e-01
7.40281403e-01 1.77048966e-01 8.51749241e-01 -1.10716462e+00
2.73539752e-01 -5.62911808e-01 -1.00712657e+00 -1.01792395e+00
-5.02476633e-01 -3.26714069e-01 2.69036144e-01 -1.34308708e+00
4.05556560e-01 -2.81864792e-01 -4.62192386e-01 4.42718208e-01
-2.52541721e-01 7.21886873e-01 2.49703661e-01 1.30793303e-01
-1.04084837e+00 2.48903885e-01 9.85963643e-01 -1.45269886e-01
4.29022387e-02 -1.18099421e-01 -5.52757084e-01 7.83050299e-01
8.30095947e-01 -5.58203936e-01 -6.76937103e-02 -3.24544430e-01
5.15662953e-02 -4.03826356e-01 6.20688736e-01 -1.26345193e+00
3.14781368e-01 -8.14452171e-02 7.21407235e-01 -8.47165585e-01
-8.45243596e-03 -6.87726021e-01 -4.19402897e-01 6.28772557e-01
7.91908894e-03 1.21609077e-01 4.73789543e-01 6.63733304e-01
-5.42600779e-03 2.31091961e-01 1.12056816e+00 1.11852430e-01
-1.24333584e+00 3.58322799e-01 -4.44713950e-01 -5.36032319e-02
1.19882524e+00 -5.74497283e-01 -4.98189390e-01 2.10111380e-01
4.50971387e-02 1.53625160e-01 1.78445309e-01 8.07039857e-01
5.49138308e-01 -1.43692136e+00 -7.57950187e-01 3.46643358e-01
1.95177928e-01 -1.18623860e-01 4.79704708e-01 9.22585130e-01
-6.09975159e-01 4.61786807e-01 -4.57595974e-01 -7.47924387e-01
-1.37961257e+00 6.76866829e-01 5.05925596e-01 1.62517458e-01
-6.75851882e-01 9.03565407e-01 4.46526617e-01 -1.38270119e-02
2.43599355e-01 -5.79626083e-01 -2.85028577e-01 8.43262151e-02
8.79963994e-01 4.50467974e-01 9.53179002e-02 -7.72488296e-01
-5.30660689e-01 7.66542673e-01 -1.34270087e-01 6.62672877e-01
1.26786304e+00 -2.01217934e-01 1.74440220e-01 5.50676025e-02
1.38285995e+00 -2.22772792e-01 -1.72640967e+00 -2.69430846e-01
-2.94444025e-01 -5.64933002e-01 1.70670465e-01 -3.73981386e-01
-1.40026045e+00 7.75025606e-01 1.24091649e+00 2.87804157e-01
9.34631109e-01 3.80350202e-02 9.35601890e-01 3.48795176e-01
1.66646793e-01 -1.22272837e+00 3.33424151e-01 4.28798884e-01
7.61855781e-01 -1.66201091e+00 1.03465937e-01 -4.44938153e-01
-4.39924568e-01 1.31094992e+00 8.99509490e-01 -3.75406832e-01
3.80861372e-01 3.06596845e-01 -1.96502004e-02 -1.22845814e-01
-4.11453635e-01 -5.80473840e-01 3.09295863e-01 5.15861392e-01
-1.19686238e-01 -2.47858576e-02 -2.02104241e-01 3.67776155e-01
2.93960899e-01 -2.15487331e-01 2.83417434e-01 9.58336532e-01
-1.11820960e+00 -2.81610131e-01 -3.95320237e-01 3.07835609e-01
-3.99271578e-01 6.74273968e-02 -7.71714225e-02 9.71086860e-01
5.35187960e-01 1.00867462e+00 2.25296706e-01 -4.33749110e-01
5.71563840e-01 -4.55323607e-01 1.41674057e-01 -7.36115575e-02
-2.35314399e-01 3.26956213e-02 -2.07380116e-01 -7.31117189e-01
-2.09601983e-01 -4.46128070e-01 -1.25495434e+00 -3.53629589e-01
-4.67588574e-01 -3.79278690e-01 7.10300744e-01 8.27476978e-01
4.37049806e-01 5.49254715e-01 6.85230851e-01 -9.90114510e-01
-6.56599700e-01 -9.94914174e-01 -5.18205285e-01 2.93211460e-01
5.83400905e-01 -7.15700865e-01 -3.95641774e-01 -2.08774865e-01] | [8.610787391662598, -0.5739319324493408] |
8cd6feed-068e-474d-88be-ab6b3a2063f6 | playing-chess-with-limited-look-ahead | 2007.02130 | null | https://arxiv.org/abs/2007.02130v1 | https://arxiv.org/pdf/2007.02130v1.pdf | Playing Chess with Limited Look Ahead | We have seen numerous machine learning methods tackle the game of chess over the years. However, one common element in these works is the necessity of a finely optimized look ahead algorithm. The particular interest of this research lies with creating a chess engine that is highly capable, but restricted in its look ahead depth. We train a deep neural network to serve as a static evaluation function, which is accompanied by a relatively simple look ahead algorithm. We show that our static evaluation function has encoded some semblance of look ahead knowledge, and is comparable to classical evaluation functions. The strength of our chess engine is assessed by comparing its proposed moves against those proposed by Stockfish. We show that, despite strict restrictions on look ahead depth, our engine recommends moves of equal strength in roughly $83\%$ of our sample positions. | ['Arman Maesumi'] | 2020-07-04 | null | null | null | null | ['game-of-chess'] | ['playing-games'] | [-3.25701326e-01 -1.35272413e-01 -1.11793287e-01 -2.53683180e-01
-4.58339214e-01 -9.55665290e-01 5.54570556e-01 7.25219101e-02
-9.84803617e-01 5.85494399e-01 -8.14033449e-02 -6.71281934e-01
-5.67015767e-01 -1.08640289e+00 -7.56664693e-01 -4.05141592e-01
-2.04493567e-01 6.54333770e-01 8.22343290e-01 -1.08211851e+00
8.42755914e-01 3.96129668e-01 -1.46754622e+00 5.41344881e-01
4.91353840e-01 1.06084907e+00 -2.18615774e-03 1.11967278e+00
1.76742882e-01 1.07094240e+00 -9.10412371e-01 -9.01088119e-01
7.71300018e-01 -4.61687058e-01 -1.20354533e+00 -7.54618406e-01
2.25333899e-01 -3.90909523e-01 -2.11842567e-01 9.40370858e-01
5.67348719e-01 2.73551255e-01 4.51712787e-01 -9.47185218e-01
-3.02153617e-01 1.04054928e+00 -2.96715379e-01 6.03658915e-01
3.01027954e-01 3.53199542e-01 1.23711252e+00 -3.36814940e-01
4.60701317e-01 4.83149946e-01 9.09228504e-01 4.25107747e-01
-6.21662557e-01 -3.56507391e-01 -1.21956348e-01 4.43712205e-01
-1.22354913e+00 -1.65319756e-01 6.22668505e-01 -2.84590065e-01
1.11506248e+00 3.43837619e-01 9.12378073e-01 4.19349521e-01
3.62487376e-01 7.66477048e-01 8.15058172e-01 -6.17707670e-01
5.85542500e-01 -1.55100688e-01 2.89058536e-02 6.74372554e-01
7.72109255e-02 4.17765647e-01 -4.25916582e-01 3.78400564e-01
5.81579864e-01 -3.53280336e-01 1.31136343e-01 -4.76461262e-01
-7.49666572e-01 8.07713151e-01 5.74631631e-01 5.33234000e-01
-4.93084677e-02 2.99234390e-01 4.11262453e-01 5.74303329e-01
-8.20430219e-02 1.21248960e+00 -5.32259524e-01 -6.08919084e-01
-1.36287868e+00 5.38177848e-01 9.24063981e-01 4.65088516e-01
6.31272852e-01 -2.26505417e-02 3.97930443e-02 1.24112800e-01
-2.63836056e-01 1.58556346e-02 7.47872353e-01 -1.10583222e+00
5.20240068e-01 6.27062976e-01 2.70250231e-01 -1.11894751e+00
-5.34893215e-01 -7.66264856e-01 -4.38229233e-01 7.99682438e-01
6.46879017e-01 -2.75018781e-01 -2.58261979e-01 1.44816995e+00
6.02145270e-02 -1.59048840e-01 -7.21258372e-02 8.44358504e-01
3.72783989e-01 4.47588652e-01 -5.44955015e-01 2.50125676e-01
8.44737589e-01 -1.19411278e+00 -4.06563692e-02 -2.32100099e-01
9.86360312e-01 -3.94773930e-01 1.24579930e+00 6.96811676e-01
-1.51413417e+00 -6.28946245e-01 -1.39509773e+00 1.97972506e-01
-5.94758153e-01 -1.13026217e-01 8.31748009e-01 8.95642519e-01
-1.20856011e+00 1.05169129e+00 -5.89019597e-01 1.48035660e-01
-9.87097472e-02 5.28280675e-01 -8.67434144e-02 4.86101061e-01
-1.31187654e+00 9.96333480e-01 8.04730237e-01 2.04580780e-02
-6.02144837e-01 -3.69969308e-01 -5.27832508e-01 4.26604182e-01
7.21388400e-01 -4.97025818e-01 1.59926820e+00 -9.45760965e-01
-1.73521090e+00 8.51410031e-01 4.86260593e-01 -1.04284012e+00
7.49356508e-01 -6.22530058e-02 -7.15348423e-02 -1.48953855e-01
-4.49155532e-02 3.63595963e-01 3.68445158e-01 -7.09696293e-01
-1.20976067e+00 -8.89540464e-02 9.25240159e-01 2.35496253e-01
-1.90121666e-01 -3.11074257e-02 -4.60784972e-01 -3.36485684e-01
-4.14788246e-01 -7.81257689e-01 -3.06452543e-01 -3.58418941e-01
-1.54390484e-01 1.62879061e-02 -6.88584596e-02 -1.39237866e-01
1.71992540e+00 -1.73683071e+00 7.80138746e-02 2.73706943e-01
1.65211439e-01 4.73123372e-01 -1.23808533e-01 4.87223566e-01
-2.61824578e-02 1.41536027e-01 7.03620762e-02 2.94657856e-01
4.49520439e-01 1.08790873e-02 -2.92934567e-01 2.29727015e-01
-2.64982074e-01 1.10946858e+00 -9.93295789e-01 -1.16184689e-01
2.88003068e-02 -3.16727132e-01 -9.11855578e-01 5.40438481e-02
-1.73380092e-01 -2.14722067e-01 -2.05534041e-01 4.41314250e-01
4.15245742e-01 -2.83799738e-01 1.73374429e-01 1.05123289e-01
-5.40055990e-01 3.45132917e-01 -1.29625869e+00 1.59634638e+00
-1.90303385e-01 5.98235965e-01 -3.37121010e-01 -1.03607428e+00
6.10862672e-01 -3.42606336e-01 1.99693739e-01 -1.11556935e+00
3.11973453e-01 3.71578455e-01 3.77409905e-01 -2.34797269e-01
1.07938707e+00 8.71013030e-02 -1.77563220e-01 4.34685171e-01
-2.09019825e-01 -1.18119083e-01 6.83170497e-01 -6.39174646e-03
1.21092200e+00 1.07406139e-01 4.38291341e-01 -2.01410264e-01
5.52167237e-01 3.31359088e-01 3.24998856e-01 1.17254555e+00
-3.68861377e-01 2.96677381e-01 6.03012323e-01 -1.15456271e+00
-9.95041013e-01 -8.96644890e-01 3.47149819e-01 1.56120205e+00
3.56078833e-01 -5.97255766e-01 -8.12532246e-01 -7.86804855e-01
-2.35858843e-01 5.45319259e-01 -9.57601964e-01 -2.96825945e-01
-8.38723361e-01 -6.79564416e-01 8.93554986e-01 6.67329609e-01
6.23751998e-01 -1.05692315e+00 -1.38000393e+00 1.67671487e-01
3.42830122e-01 -4.24151570e-01 -3.30314875e-01 4.24998015e-01
-5.45620918e-01 -1.27079201e+00 -8.08055222e-01 -6.12563372e-01
1.27741262e-01 -7.90145695e-02 1.47843397e+00 2.32030645e-01
-7.53792673e-02 -6.93940045e-03 -5.80891728e-01 -3.32669139e-01
-3.99728477e-01 6.15709305e-01 -9.42848176e-02 -6.29793108e-01
6.83965608e-02 -3.49664003e-01 -8.34748209e-01 3.24727744e-01
-6.91510141e-01 -4.36520651e-02 7.30471969e-01 9.26607072e-01
2.83706307e-01 3.63258421e-01 1.26247630e-01 -9.39381778e-01
8.15949202e-01 -1.73571944e-01 -1.00425339e+00 3.00655991e-01
-8.06178212e-01 3.65543693e-01 9.29608583e-01 -1.53310165e-01
-5.58235824e-01 -8.04464817e-02 -4.60933983e-01 2.42616639e-01
1.86056852e-01 5.76681793e-01 2.63026029e-01 -3.83010179e-01
1.19573939e+00 3.26560378e-01 -4.08437073e-01 -3.27591330e-01
4.42413419e-01 1.85939282e-01 6.69840872e-01 -5.04172385e-01
6.41549587e-01 1.90650627e-01 -3.90412286e-02 -1.54014409e-01
-6.29594743e-01 -2.03062683e-01 -6.07143342e-01 -2.23666623e-01
4.02687579e-01 -2.45124653e-01 -1.35139763e+00 4.68409002e-01
-7.07220674e-01 -6.99715972e-01 -5.29291749e-01 1.29574627e-01
-7.71937907e-01 3.63383234e-01 -5.34095109e-01 -6.19416952e-01
-2.48413652e-01 -1.08594704e+00 6.22353673e-01 3.19089413e-01
-1.47478729e-01 -8.69847178e-01 4.32640016e-01 7.06339255e-02
5.51839769e-01 -5.32635786e-02 6.39192820e-01 -8.10452461e-01
-4.92398381e-01 -2.99380571e-01 2.44420134e-02 1.63287610e-01
-3.05402875e-01 -1.23984821e-01 -4.67862606e-01 -2.26798102e-01
-2.22996697e-01 -2.25490361e-01 8.26537192e-01 4.14451778e-01
1.02144659e+00 -3.30431998e-01 7.43470853e-04 9.95803177e-01
1.42418551e+00 5.89053392e-01 6.04802847e-01 1.23477507e+00
1.67171191e-02 3.36449534e-01 5.27937531e-01 4.83031452e-01
2.95933157e-01 7.13117957e-01 6.86279833e-01 2.43006721e-01
1.53120786e-01 -2.81462431e-01 1.58664733e-01 4.61927861e-01
-3.74603897e-01 -2.77693093e-01 -1.06865358e+00 4.63117599e-01
-1.72656178e+00 -1.13999057e+00 2.46958345e-01 2.15438581e+00
7.39692211e-01 7.13062942e-01 4.86464918e-01 4.73183453e-01
2.44798630e-01 1.97784305e-01 -4.24893141e-01 -7.85260797e-01
-1.05833121e-01 4.42740977e-01 7.06662297e-01 4.05503184e-01
-1.11142719e+00 9.06583965e-01 7.80379438e+00 1.03252947e+00
-1.08491194e+00 -4.88413632e-01 4.86973464e-01 -3.92662287e-01
-3.73497754e-02 -6.12426847e-02 -6.78670883e-01 6.15769088e-01
8.72099340e-01 -3.28726083e-01 5.22036076e-01 1.06377518e+00
-2.74464101e-01 -2.32964709e-01 -1.00194621e+00 7.64830589e-01
1.51518919e-02 -1.82733738e+00 -1.27801627e-01 -9.94113758e-02
6.26799405e-01 -1.80881649e-01 3.59168261e-01 6.65567815e-01
6.07925892e-01 -1.03313875e+00 1.00301909e+00 4.46190029e-01
5.01956224e-01 -1.02364588e+00 9.62319374e-01 5.97617447e-01
-1.12947023e+00 -4.45604056e-01 -3.12516034e-01 -6.52144670e-01
-1.51213527e-01 -9.52262282e-02 -7.91228235e-01 4.80132222e-01
8.96974921e-01 8.88049453e-02 -6.65429056e-01 1.35502589e+00
-1.51605532e-01 3.71710241e-01 -2.59656638e-01 -5.66401243e-01
6.35400176e-01 1.21665724e-01 2.37673894e-01 1.12490749e+00
2.56719440e-01 -2.55829729e-02 -2.04723477e-01 5.98598897e-01
3.00868321e-02 4.93700355e-02 -2.74159908e-01 3.71765159e-02
3.89579594e-01 8.46704721e-01 -7.45478511e-01 -2.47575060e-01
4.48779501e-02 9.42287266e-01 4.43229169e-01 -2.68457532e-01
-8.85174334e-01 -6.58043206e-01 1.96530297e-01 2.14413106e-01
5.53370118e-01 2.17345338e-02 -3.52407157e-01 -8.77360106e-01
-2.61040498e-02 -9.09317672e-01 5.24331629e-01 -7.54761457e-01
-9.14616048e-01 7.64588177e-01 -2.33274817e-01 -1.12883174e+00
-7.13015378e-01 -9.11540329e-01 -8.25549662e-01 6.17885590e-01
-1.31074119e+00 -8.98826182e-01 -2.26023644e-02 2.65583664e-01
2.91871637e-01 -5.66579103e-01 5.97804546e-01 3.10522169e-01
-2.93976814e-01 9.99619901e-01 3.02867860e-01 1.47566840e-01
3.80900115e-01 -1.62011623e+00 7.43669152e-01 7.76874661e-01
3.00305039e-01 5.48811376e-01 9.08138871e-01 -1.86863557e-01
-1.08355117e+00 -4.63191062e-01 7.22968280e-01 -4.37931001e-01
7.47208297e-01 1.56715021e-01 -5.33399582e-01 5.78187287e-01
7.71790668e-02 -4.11477268e-01 4.13272798e-01 2.65678555e-01
-2.15916187e-01 -2.51398563e-01 -8.32944930e-01 6.59618855e-01
9.02151883e-01 -1.89637840e-01 -8.36890280e-01 -1.66436926e-01
2.47590303e-01 -7.76588738e-01 -3.27281475e-01 2.35656008e-01
9.43126857e-01 -1.69464731e+00 1.03618097e+00 -9.55369055e-01
3.84028077e-01 -1.96130827e-01 -1.47510471e-03 -1.23834229e+00
-5.29256284e-01 -5.88834226e-01 3.77736683e-03 6.10863209e-01
5.51369309e-01 -2.55390227e-01 1.48741496e+00 5.52956760e-01
-2.11206540e-01 -1.07923400e+00 -8.55794191e-01 -1.07569456e+00
4.12898362e-01 -3.46478611e-01 6.54100537e-01 6.84082329e-01
1.39678612e-01 2.76266206e-02 -4.76645380e-01 -1.37102738e-01
9.82302353e-02 4.22253460e-01 8.64684343e-01 -1.09321535e+00
-9.38134313e-01 -1.12551939e+00 -4.57139075e-01 -1.26566398e+00
-2.78809428e-01 -7.11025476e-01 -1.67356268e-01 -1.37675452e+00
-3.20990719e-02 -2.32609183e-01 -5.03225803e-01 3.85891080e-01
1.25321209e-01 2.70798355e-01 3.11521143e-01 -7.38044456e-02
-1.05932665e+00 3.71364988e-02 9.31249678e-01 -2.11760968e-01
-1.97867706e-01 3.36115748e-01 -8.52819562e-01 1.05483627e+00
8.76518667e-01 -3.05390686e-01 -3.36531878e-01 -6.10246658e-01
1.24768329e+00 -9.93540660e-02 4.74414695e-03 -1.40305316e+00
7.43189573e-01 -1.62975267e-01 2.30038717e-01 -5.11771500e-01
2.32531860e-01 -4.59815294e-01 -1.52075812e-01 8.24856341e-01
-4.81026202e-01 5.14705241e-01 1.57214597e-01 1.04971584e-02
-1.42173976e-01 -6.94936037e-01 5.48506618e-01 -3.47018450e-01
-1.21915948e+00 -1.72168650e-02 -2.07922906e-01 3.38320434e-01
9.80980158e-01 -6.68077230e-01 -1.63997844e-01 -3.41466844e-01
-4.70602661e-01 9.40365344e-02 5.50828159e-01 -4.48125824e-02
3.26645941e-01 -9.35130775e-01 -4.67656583e-01 4.78690900e-02
-1.01955071e-01 -3.35291952e-01 3.06797445e-01 4.11119193e-01
-1.17640388e+00 5.13728321e-01 -5.85232317e-01 5.12744896e-02
-1.17239714e+00 7.49026656e-01 6.96064174e-01 -6.68594718e-01
-2.69556522e-01 1.20324230e+00 -2.22500399e-01 -3.83151859e-01
2.08388373e-01 -3.85687172e-01 -1.64654970e-01 -1.28766105e-01
7.24951148e-01 5.05215943e-01 1.55705959e-01 -1.15171619e-01
-3.51026475e-01 5.30014277e-01 1.28199279e-01 -1.59223422e-01
1.40481699e+00 2.70250708e-01 3.26582313e-01 1.86669886e-01
5.43373585e-01 3.45653325e-01 -1.10462892e+00 1.58955768e-01
1.58729419e-01 -4.62009907e-01 -2.46013895e-01 -1.26904356e+00
-8.44177365e-01 9.01599288e-01 3.60489219e-01 4.97317940e-01
1.20680082e+00 -3.53910923e-01 7.41731584e-01 9.46071446e-01
6.84090793e-01 -1.15835392e+00 -4.99717630e-02 9.01278853e-01
4.82521564e-01 -1.13841736e+00 -1.50083140e-01 2.82131225e-01
-4.61048424e-01 1.29357910e+00 5.35899818e-01 -4.28558499e-01
2.65279293e-01 5.00060618e-01 -1.04460016e-01 -2.11942658e-01
-6.55216396e-01 -2.90264696e-01 2.47822836e-01 4.46582407e-01
1.97839662e-01 -1.78916633e-01 -4.63098854e-01 9.74770188e-01
-8.04431677e-01 1.43016100e-01 4.72189069e-01 9.26380098e-01
-7.67880559e-01 -1.09600413e+00 -1.74381286e-01 1.41427204e-01
-5.66057920e-01 -2.41239950e-01 -6.71028137e-01 1.05987000e+00
3.44118625e-01 5.89116812e-01 2.29260214e-02 -8.24100316e-01
4.47579741e-01 -2.05974832e-01 5.35547853e-01 -2.35862136e-01
-1.09554863e+00 -5.19483984e-01 -2.36855038e-02 -6.12649024e-01
4.03675772e-02 -1.01174541e-01 -1.13756645e+00 -7.36101091e-01
-1.40477717e-01 5.37936747e-01 4.41074908e-01 8.59016359e-01
1.28316477e-01 4.92883295e-01 4.57366854e-01 -6.24817669e-01
-8.01582456e-01 -3.83282632e-01 -6.82554066e-01 1.72654852e-01
5.49405366e-02 -4.76280361e-01 -1.89931676e-01 -3.74364555e-01] | [3.4365477561950684, 1.4183486700057983] |
ab4d713a-044d-40d0-a702-2adf8a4c5332 | a-large-scale-homography-benchmark-1 | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Barath_A_Large-Scale_Homography_Benchmark_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Barath_A_Large-Scale_Homography_Benchmark_CVPR_2023_paper.pdf | A Large-Scale Homography Benchmark | We present a large-scale dataset of Planes in 3D, Pi3D, of roughly 1000 planes observed in 10 000 images from the 1DSfM dataset, and HEB, a large-scale homography estimation benchmark leveraging Pi3D. The applications of the Pi3D dataset are diverse, e.g. training or evaluating monocular depth, surface normal estimation and image matching algorithms. The HEB dataset consists of 226 260 homographies and includes roughly 4M correspondences. The homographies link images that often undergo significant viewpoint and illumination changes. As applications of HEB, we perform a rigorous evaluation of a wide range of robust estimators and deep learning-based correspondence filtering methods, establishing the current state-of-the-art in robust homography estimation. We also evaluate the uncertainty of the SIFT orientations and scales w.r.t. the ground truth coming from the underlying homographies and provide codes for comparing uncertainty of custom detectors. | ['Jiri Matas', 'Wolfgang Förstner', 'Michal Polic', 'Dmytro Mishkin', 'Daniel Barath'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['homography-estimation'] | ['computer-vision'] | [-5.83946854e-02 -2.20290676e-01 -1.01390423e-03 -2.28711188e-01
-9.87651765e-01 -6.89185023e-01 7.95503736e-01 -3.00038844e-01
-6.42896444e-03 1.00184500e-01 3.04451257e-01 3.01395804e-01
-1.97227105e-01 -7.13463306e-01 -1.31976473e+00 -3.68193954e-01
-2.01308072e-01 9.70372021e-01 4.07988012e-01 -1.22501001e-01
6.88847661e-01 8.19656670e-01 -1.54906940e+00 -9.84856207e-03
2.36576125e-01 1.20449460e+00 -1.74091890e-01 5.67841291e-01
4.87034619e-01 -8.36378336e-02 -3.67970288e-01 -5.51674545e-01
9.26880360e-01 1.08928449e-01 -3.65882784e-01 3.81596386e-01
1.67821705e+00 -7.96668947e-01 -8.53534281e-01 1.13329017e+00
4.49997813e-01 -1.81384161e-01 7.60292113e-01 -1.39836800e+00
-4.26880807e-01 -1.61731213e-01 -6.57994151e-01 -8.33106712e-02
7.22386122e-01 3.29522550e-01 9.04747188e-01 -1.04991651e+00
1.28104699e+00 1.60568750e+00 9.17101026e-01 3.90297621e-02
-1.17459536e+00 -5.71533203e-01 -3.32849294e-01 1.04032509e-01
-1.47674680e+00 -4.34719950e-01 5.01300931e-01 -7.00841963e-01
1.23470724e+00 -1.94901913e-01 6.92853808e-01 1.13499594e+00
5.56121886e-01 5.11956871e-01 7.54811645e-01 -2.88438767e-01
2.11209245e-02 -2.85301477e-01 -1.40300691e-01 7.65180290e-01
6.37482405e-01 6.13832176e-01 -1.15257585e+00 -1.70012578e-01
1.51271677e+00 -2.39748076e-01 -3.00579518e-01 -1.15395546e+00
-1.53450155e+00 4.58834499e-01 3.87849331e-01 -5.98490596e-01
-2.46302895e-02 2.00070962e-01 1.00795396e-01 5.56488216e-01
1.24400213e-01 5.80241680e-01 -2.40829065e-01 1.54615641e-02
-6.61831319e-01 4.69039053e-01 6.81948304e-01 1.93784094e+00
1.03428316e+00 -3.58235508e-01 3.60809594e-01 3.98565173e-01
3.10144782e-01 9.03273523e-01 -7.86597654e-02 -1.46920872e+00
6.56032145e-01 3.34216326e-01 1.21940650e-01 -1.37314594e+00
-3.09092402e-01 -2.60749254e-02 -5.63562572e-01 3.70345205e-01
4.70615268e-01 4.51424003e-01 -7.53071427e-01 1.13694727e+00
3.02637458e-01 -7.60706887e-02 -2.96043009e-01 8.68216217e-01
6.52496517e-01 9.48111489e-02 -1.04281628e+00 4.62738037e-01
9.99653041e-01 -7.28030622e-01 -1.92338094e-01 -3.03131372e-01
2.44094104e-01 -1.22998571e+00 6.09133720e-01 7.12101042e-01
-1.19547617e+00 -4.26947623e-01 -1.31631386e+00 -5.64564943e-01
-2.49204144e-01 -2.55074985e-02 5.26780009e-01 1.66683793e-01
-9.96256948e-01 8.56533587e-01 -6.91261590e-01 -4.86742377e-01
2.06625447e-01 3.71388406e-01 -7.38847077e-01 -3.54188323e-01
-7.19169617e-01 8.19339216e-01 3.73033844e-02 -2.49584123e-01
-7.44618893e-01 -9.36850905e-01 -1.04275072e+00 -2.27752998e-01
1.39433488e-01 -8.87763023e-01 1.09921825e+00 -1.85994163e-01
-1.19822931e+00 1.50887513e+00 -5.83338775e-02 -3.20794076e-01
9.28948402e-01 -5.54151058e-01 -1.80220932e-01 3.47969979e-01
1.26715928e-01 7.38681734e-01 8.74372363e-01 -9.59320366e-01
-6.25972927e-01 -5.83196759e-01 -3.14678818e-01 1.91625327e-01
3.89438629e-01 -3.69840115e-01 -7.47380614e-01 -3.55644435e-01
7.72549510e-01 -8.69947791e-01 1.29434630e-01 4.25140679e-01
-6.06087625e-01 2.72981375e-01 4.23414439e-01 -5.15353858e-01
4.26222205e-01 -2.16502213e+00 1.63837120e-01 5.19738495e-01
3.21110189e-01 -5.37209868e-01 -7.76485726e-02 5.49824238e-01
1.48069009e-01 -3.14790934e-01 3.42214048e-01 -4.48671848e-01
3.83058935e-01 -1.42405063e-01 -3.59631628e-01 1.10550797e+00
-4.52317894e-02 7.01330543e-01 -6.00809634e-01 -1.03877917e-01
6.23251259e-01 4.48799878e-01 -6.28459632e-01 1.57618478e-01
-5.10069393e-02 2.24239141e-01 3.60933989e-02 9.17301595e-01
1.13074601e+00 -2.20828190e-01 -2.66004771e-01 -8.74187171e-01
-2.06049517e-01 3.75149161e-01 -1.39247119e+00 1.95795000e+00
-4.44488563e-02 1.05788279e+00 -2.41193064e-02 -1.76422268e-01
9.54667926e-01 -2.79447764e-01 6.36699200e-01 -6.81865275e-01
2.37996116e-01 4.39887077e-01 -3.12478632e-01 1.83731273e-01
6.51493311e-01 5.09803534e-01 8.72883275e-02 -4.12088335e-02
3.46388757e-01 -8.96482646e-01 -5.64780831e-02 1.50014058e-01
1.12276256e+00 2.16987997e-01 4.68112409e-01 -4.72614259e-01
7.17532560e-02 5.69806211e-02 3.23387742e-01 6.96596444e-01
2.22184118e-02 1.12516928e+00 3.09036106e-01 -5.14306664e-01
-1.55190885e+00 -1.44395638e+00 -5.05149424e-01 -1.21619023e-01
6.40709519e-01 -4.29562628e-01 -3.53645325e-01 7.48101920e-02
7.61914492e-01 -1.37992501e-01 -4.51496035e-01 -1.79186776e-01
-2.72381216e-01 -1.37999609e-01 1.36889562e-01 3.89084488e-01
8.09396684e-01 -2.64266491e-01 -4.66329545e-01 -1.18572690e-01
3.06735963e-01 -1.55175769e+00 -6.78787649e-01 -1.77913427e-01
-8.42140377e-01 -1.51658475e+00 -4.02564228e-01 -5.44897437e-01
6.72654927e-01 5.11706352e-01 1.64831424e+00 -2.69275665e-01
-5.22207379e-01 7.64114201e-01 8.61727446e-02 8.39553177e-02
-1.52602702e-01 -2.37128526e-01 2.28719130e-01 -4.32020307e-01
5.43290615e-01 -5.99095881e-01 -7.14127660e-01 9.85745788e-01
-3.67386341e-01 -3.95009145e-02 5.24402440e-01 6.36158586e-01
9.15183306e-01 -1.65386975e-01 -6.05694294e-01 -4.76678371e-01
-1.62764624e-01 -3.04591842e-02 -1.59317207e+00 1.01925209e-01
-4.64639693e-01 1.12862978e-03 -2.96210796e-02 -1.94681481e-01
-5.96151054e-01 1.44716680e-01 1.75654158e-01 -8.12767744e-01
-4.43735644e-02 -1.89605579e-01 -3.60065877e-01 -8.44424665e-01
6.90491319e-01 -1.56566992e-01 -1.24141477e-01 -4.58545953e-01
3.50992203e-01 1.94710165e-01 1.09926760e+00 -5.63097894e-01
1.23793602e+00 9.22914863e-01 5.82165360e-01 -9.97723043e-01
-4.85448569e-01 -8.07054400e-01 -1.01582348e+00 6.66172756e-03
5.95244944e-01 -1.24775398e+00 -8.39565992e-01 8.35407853e-01
-1.26926160e+00 -1.86108798e-01 -2.83023622e-02 6.52090907e-01
-9.93516505e-01 5.05202472e-01 -6.15588307e-01 -3.53668593e-02
-3.54550555e-02 -1.33723485e+00 1.68970013e+00 -7.61719123e-02
-2.11910293e-01 -7.59872258e-01 6.16162345e-02 1.48384362e-01
-1.47304028e-01 3.68333578e-01 6.18919790e-01 1.60478968e-02
-1.53243244e+00 -2.85681814e-01 -3.47630650e-01 5.12499548e-02
1.86369158e-02 2.91199803e-01 -8.35099339e-01 -4.87627029e-01
-1.45333484e-01 -2.38744929e-01 6.63026154e-01 5.50431907e-01
5.27607560e-01 9.85920802e-02 -2.37514347e-01 1.49731636e+00
1.54975986e+00 -1.61601171e-01 8.57056141e-01 8.75271797e-01
8.39758098e-01 5.39218664e-01 7.42479622e-01 4.78003949e-01
2.07373247e-01 9.80989277e-01 4.92725223e-01 7.38688782e-02
-3.19398165e-01 -5.21830797e-01 1.16713017e-01 5.47688365e-01
1.38957202e-01 1.01620927e-01 -8.21362555e-01 3.15514266e-01
-1.48492074e+00 -5.58311164e-01 -2.27052376e-01 2.65782881e+00
5.19187033e-01 2.62146145e-01 -2.15000257e-01 -2.58995771e-01
5.26722014e-01 4.50186059e-02 -8.07088971e-01 2.12283388e-01
-4.43702221e-01 1.60458665e-02 1.18347728e+00 6.20378435e-01
-1.27817655e+00 1.00376070e+00 7.00112247e+00 4.86650050e-01
-8.27648878e-01 -5.68247259e-01 -1.03790602e-02 1.63766891e-02
-1.50000870e-01 7.01295733e-02 -1.35362494e+00 6.72529787e-02
3.26545954e-01 -1.99569345e-01 5.50004363e-01 8.56408238e-01
-3.89797062e-01 -2.26521164e-01 -1.59238482e+00 1.51415002e+00
3.30606550e-01 -1.59210968e+00 6.24907464e-02 4.57866788e-01
1.19011605e+00 5.28473616e-01 1.82879135e-01 -4.03726578e-01
3.70008707e-01 -6.59475982e-01 7.77776897e-01 4.68513399e-01
9.63925719e-01 -5.36521733e-01 4.63234425e-01 -1.54909328e-01
-1.03643644e+00 4.65868205e-01 -9.09456670e-01 3.27291399e-01
1.64809912e-01 6.41867876e-01 -6.12415016e-01 4.87108201e-01
9.43197310e-01 1.23208535e+00 -7.04482198e-01 1.31075180e+00
-2.44237348e-01 -2.27963969e-01 -6.31762624e-01 5.63793838e-01
-5.60779944e-02 -3.44271332e-01 6.97947741e-01 6.91909611e-01
4.90029544e-01 -1.98761120e-01 -6.30809441e-02 8.41323793e-01
-2.92365253e-01 -3.61044824e-01 -9.12949562e-01 2.61958867e-01
6.08116448e-01 8.95932198e-01 -4.70037699e-01 -1.17212921e-01
-4.77953374e-01 9.37331915e-01 -7.89898559e-02 2.80465871e-01
-4.65754271e-01 -2.45457664e-01 1.19988501e+00 2.75656402e-01
3.23192626e-01 -5.86074233e-01 -1.73684686e-01 -1.40827262e+00
1.07842147e-01 -9.40441728e-01 -1.23358453e-02 -1.04751003e+00
-1.52318871e+00 2.72672415e-01 1.77449420e-01 -1.53894234e+00
-2.44341582e-01 -1.22420180e+00 -2.61685461e-01 7.14481771e-01
-1.31615591e+00 -8.80485713e-01 -8.17865670e-01 5.63520193e-01
3.27501595e-01 -3.00096899e-01 3.10768008e-01 2.25500762e-01
1.02900624e-01 5.35655260e-01 3.53369117e-01 -1.52800232e-02
1.10274744e+00 -1.12298739e+00 1.30307245e+00 8.28228712e-01
1.62066028e-01 7.41194248e-01 7.95808375e-01 -7.75825679e-01
-1.96632826e+00 -6.27478063e-01 3.74227881e-01 -1.03517842e+00
7.76614904e-01 -6.27229869e-01 -5.71971536e-01 1.01633966e+00
-1.93710521e-01 4.42458838e-02 -7.89112821e-02 -2.44933248e-01
-6.37534380e-01 -1.25560567e-01 -1.16646695e+00 5.06644905e-01
1.61220157e+00 -8.35945606e-01 -4.03453916e-01 4.82134402e-01
5.88431895e-01 -1.13365543e+00 -1.14636803e+00 5.28702199e-01
8.41485679e-01 -1.43722498e+00 1.54318953e+00 -1.70484304e-01
4.23609674e-01 -1.70310482e-01 -5.99605024e-01 -1.07689250e+00
-2.61024237e-01 -7.65257955e-01 8.23895559e-02 7.75896788e-01
-1.59569785e-01 -6.71746016e-01 1.04408324e+00 5.22586524e-01
-2.09572107e-01 -1.93242595e-01 -1.02208257e+00 -1.11933589e+00
-2.36707881e-01 -1.91343129e-02 6.31988466e-01 7.78491437e-01
-5.58905840e-01 -1.26144394e-01 -3.02334577e-01 4.92980599e-01
1.43554139e+00 2.58042991e-01 1.48554683e+00 -1.14575338e+00
-4.29627776e-01 -4.57927436e-01 -1.17054021e+00 -1.61326432e+00
-4.26310860e-02 -5.01002133e-01 -7.11623505e-02 -9.25624907e-01
-1.19413815e-01 4.29968387e-02 5.24996459e-01 -1.28956154e-01
5.36681414e-01 1.44977883e-01 -2.14926928e-01 4.62375790e-01
-2.12623641e-01 4.12869304e-01 1.17962182e+00 -1.55845568e-01
1.42206296e-01 -1.44521311e-01 1.26460478e-01 8.50983739e-01
3.10351253e-01 -4.55073044e-02 -2.06580341e-01 -7.41849542e-01
3.18146944e-01 -1.28966168e-01 5.62786400e-01 -1.15513158e+00
4.75426644e-01 1.01016276e-01 6.64045215e-01 -1.17411244e+00
5.27442157e-01 -8.44766974e-01 4.72803086e-01 1.71366125e-01
-1.94675028e-02 3.12445492e-01 2.03262374e-01 4.00883645e-01
-2.07334787e-01 2.29003102e-01 7.62727678e-01 -1.93429753e-01
-1.00155497e+00 6.23520434e-01 4.12305623e-01 3.66345227e-01
5.11763573e-01 -5.96003294e-01 -7.89107084e-01 -3.46383929e-01
-1.23698100e-01 3.98433320e-02 1.40432239e+00 4.45858389e-01
8.99922848e-01 -1.44950986e+00 -3.33960384e-01 7.21034706e-01
4.38866675e-01 1.82855010e-01 -2.11580284e-02 5.32246470e-01
-1.04365361e+00 3.76903385e-01 -5.92189848e-01 -1.00705159e+00
-1.09836078e+00 2.71238238e-01 4.11252856e-01 3.07869196e-01
-9.27276909e-01 8.05952907e-01 4.39514667e-01 -3.27243805e-01
4.65911716e-01 -5.01621664e-01 6.16799891e-01 -4.40874845e-01
2.87539333e-01 5.63093781e-01 2.65665829e-01 -6.66971564e-01
-4.20651197e-01 1.31889904e+00 1.89714372e-01 -1.43407449e-01
1.15210199e+00 -1.62172675e-01 -1.07110783e-01 7.62541518e-02
1.47039497e+00 2.84497421e-02 -1.64988518e+00 -1.93888009e-01
-7.68751055e-02 -1.17286205e+00 1.68506373e-02 -3.11048210e-01
-8.65731537e-01 8.41395140e-01 5.59950531e-01 -6.38574302e-01
6.86109722e-01 9.80263352e-02 6.41602159e-01 7.08130956e-01
1.00295341e+00 -8.25246632e-01 1.06541030e-01 5.45328796e-01
1.03287244e+00 -1.40836072e+00 5.18733442e-01 -7.28248358e-01
-6.66405261e-02 1.32630193e+00 5.90297759e-01 -5.39336085e-01
5.20808220e-01 2.99069464e-01 -3.08234361e-03 -5.07730424e-01
-3.87831360e-01 1.72468454e-01 6.70880675e-01 7.81506836e-01
-1.54728174e-01 -3.61797631e-01 4.32813853e-01 -2.07848027e-01
-6.38193667e-01 -4.17882144e-01 4.16523367e-01 6.85034096e-01
-2.23001674e-01 -9.04723048e-01 -5.96886694e-01 2.18431782e-02
3.31069350e-01 -6.26608264e-03 -5.65289259e-01 1.14609921e+00
-3.59231144e-01 4.32770312e-01 2.08790064e-01 -6.02754772e-01
6.77433968e-01 -6.61912799e-01 1.06119764e+00 -2.78545648e-01
6.00433350e-03 1.16123453e-01 5.87496422e-02 -1.27508879e+00
-2.12859482e-01 -8.06344271e-01 -7.01196074e-01 -6.27150476e-01
1.76364519e-02 -4.39782113e-01 6.77423120e-01 3.95651102e-01
2.74786174e-01 -2.00722709e-01 4.94929552e-01 -1.30455184e+00
-6.41188979e-01 -4.96047080e-01 -8.26771200e-01 6.34821773e-01
4.52873230e-01 -9.82760847e-01 -7.96495318e-01 -1.44302964e-01] | [8.103046417236328, -2.3380258083343506] |
b9419abc-8d2d-41a9-bd62-117b75d17d27 | syntactically-guided-generative-embeddings-1 | 2101.11530 | null | https://arxiv.org/abs/2101.11530v2 | https://arxiv.org/pdf/2101.11530v2.pdf | Syntactically Guided Generative Embeddings for Zero-Shot Skeleton Action Recognition | We introduce SynSE, a novel syntactically guided generative approach for Zero-Shot Learning (ZSL). Our end-to-end approach learns progressively refined generative embedding spaces constrained within and across the involved modalities (visual, language). The inter-modal constraints are defined between action sequence embedding and embeddings of Parts of Speech (PoS) tagged words in the corresponding action description. We deploy SynSE for the task of skeleton-based action sequence recognition. Our design choices enable SynSE to generalize compositionally, i.e., recognize sequences whose action descriptions contain words not encountered during training. We also extend our approach to the more challenging Generalized Zero-Shot Learning (GZSL) problem via a confidence-based gating mechanism. We are the first to present zero-shot skeleton action recognition results on the large-scale NTU-60 and NTU-120 skeleton action datasets with multiple splits. Our results demonstrate SynSE's state of the art performance in both ZSL and GZSL settings compared to strong baselines on the NTU-60 and NTU-120 datasets. The code and pretrained models are available at https://github.com/skelemoa/synse-zsl | ['Ravi Kiran Sarvadevabhatla', 'Divyanshu Sharma', 'Pranay Gupta'] | 2021-01-27 | syntactically-guided-generative-embeddings | https://arxiv.org/pdf/2101.11530.pdf | https://arxiv.org/pdf/2101.11530.pdf | null | ['zero-shot-skeletal-action-recognition'] | ['computer-vision'] | [ 5.36520898e-01 1.27458900e-01 -5.48021019e-01 -2.79342651e-01
-1.26001143e+00 -2.31440097e-01 7.44456172e-01 -4.08296019e-01
-2.99950808e-01 4.87911105e-01 8.63854766e-01 3.26926082e-01
1.31153837e-01 -5.66298842e-01 -6.29596949e-01 -6.85933352e-01
-3.13806050e-02 3.96101505e-01 3.15562874e-01 -3.05511933e-02
-3.54543142e-02 3.06955241e-02 -1.55478418e+00 6.78504884e-01
1.78657100e-01 6.54244602e-01 1.09725051e-01 9.15525258e-01
5.32020554e-02 7.91199625e-01 -2.05697209e-01 -2.67125845e-01
1.57369375e-01 -8.23195815e-01 -6.93502188e-01 3.08251113e-01
6.82318330e-01 -2.91144162e-01 -5.76030433e-01 7.48373210e-01
8.21420729e-01 3.59320611e-01 7.75509775e-01 -1.41435051e+00
-9.90044475e-01 5.47854900e-01 -4.28675652e-01 3.17496240e-01
5.88036895e-01 6.55547500e-01 1.36198688e+00 -1.13266432e+00
1.08575761e+00 1.42028356e+00 4.67581242e-01 1.18165135e+00
-1.18571532e+00 -3.82215232e-01 2.02283040e-01 3.51417392e-01
-1.08179915e+00 -6.50553942e-01 5.05218983e-01 -5.98505914e-01
1.29687047e+00 -1.28528640e-01 6.32700562e-01 2.01236701e+00
2.00097591e-01 1.29760325e+00 6.25568092e-01 -3.69975269e-01
5.16453564e-01 -7.64147103e-01 1.59804165e-01 9.90285516e-01
-1.32879138e-01 1.24663413e-01 -1.23038208e+00 -5.18397838e-02
7.57553816e-01 8.73100683e-02 -9.18291062e-02 -7.23806322e-01
-1.44034362e+00 9.15043652e-01 7.51588047e-02 1.89341336e-01
-3.63214344e-01 5.74612558e-01 7.29844451e-01 1.29764020e-01
3.21189284e-01 5.20878918e-02 -3.27654094e-01 -6.77033424e-01
-6.58464491e-01 8.59143659e-02 4.92520154e-01 1.06827748e+00
3.67966175e-01 2.15014920e-01 -5.44700801e-01 8.06736469e-01
2.19005942e-01 4.02243465e-01 7.83157706e-01 -1.07061362e+00
4.67898190e-01 1.28042266e-01 -3.71544957e-01 -1.38332561e-01
3.38867716e-02 -4.02105320e-03 -4.81949031e-01 2.71020740e-01
1.03990324e-01 6.39047520e-03 -1.44299638e+00 1.99800265e+00
2.75956511e-01 3.80178601e-01 1.58444121e-01 7.67443597e-01
9.47034955e-01 4.80380923e-01 3.76637578e-01 5.63192442e-02
1.22432888e+00 -1.11825657e+00 -5.12821376e-01 -4.22453970e-01
6.01543307e-01 -2.82230943e-01 1.42423332e+00 1.16340578e-01
-8.69213879e-01 -5.35617054e-01 -9.11388159e-01 -1.94925815e-01
-1.77285478e-01 -3.86473052e-02 4.77287829e-01 2.16893375e-01
-1.00516784e+00 5.28133273e-01 -1.43235624e+00 -6.47355616e-01
7.68907785e-01 -5.78447580e-02 -5.53025603e-01 -1.53683051e-01
-1.16993034e+00 5.99642694e-01 4.19015139e-01 -5.21313548e-01
-1.39890814e+00 -5.31945884e-01 -1.34501910e+00 -6.89273700e-02
7.04033792e-01 -8.67626727e-01 1.43766940e+00 -6.70036972e-01
-1.37274659e+00 9.56817627e-01 -7.87675530e-02 -3.87305051e-01
2.18022078e-01 -3.48250240e-01 -2.74670064e-01 3.40232790e-01
4.41140831e-01 9.00645018e-01 8.47426057e-01 -7.35266209e-01
-3.59073371e-01 -3.10989290e-01 -1.26240611e-01 1.66218191e-01
3.48334312e-02 -3.33368927e-02 -4.00604665e-01 -9.26342785e-01
-1.27141848e-01 -8.51727188e-01 -8.38129371e-02 2.97828794e-01
-3.70223194e-01 -4.47627991e-01 5.60768962e-01 -4.71885234e-01
1.00192046e+00 -2.35145974e+00 4.93820548e-01 -3.49520743e-01
-2.08642390e-02 1.69984922e-01 -5.44138312e-01 7.56015301e-01
-2.87772626e-01 -1.94519937e-01 -4.64724422e-01 -4.93515611e-01
3.24333549e-01 5.95440030e-01 -1.97860822e-01 3.79397213e-01
2.34886125e-01 1.35808396e+00 -1.14225972e+00 -6.07002676e-01
3.69225442e-01 4.11420077e-01 -5.58122694e-01 8.48633051e-02
-4.06374574e-01 1.53436184e-01 -4.46641296e-01 1.02730739e+00
-3.10334235e-01 -3.69902521e-01 -3.45685296e-02 -7.76340812e-02
3.31956446e-01 1.44150421e-01 -7.65105724e-01 2.38619637e+00
-2.62289494e-01 3.60687405e-01 -3.57667893e-01 -7.05868661e-01
3.96798730e-01 4.90021169e-01 5.67320168e-01 -3.81949902e-01
6.30891025e-02 -1.04836255e-01 -2.06805736e-01 -7.31525779e-01
-1.44035267e-02 -5.16362965e-01 -4.62021112e-01 6.70219481e-01
8.22470009e-01 1.59020469e-01 4.27296579e-01 4.57363039e-01
1.46303391e+00 7.77888536e-01 7.86757708e-01 1.61340788e-01
1.19672135e-01 -2.75465846e-01 6.76223814e-01 6.14136100e-01
-4.08318788e-01 7.24410117e-01 4.99721318e-01 -1.61536083e-01
-9.13324535e-01 -1.63486648e+00 2.62010366e-01 1.48545194e+00
-2.10487947e-01 -6.83685184e-01 -5.04725575e-01 -9.40221667e-01
1.74264647e-02 8.17805588e-01 -8.61463606e-01 -3.09396982e-01
-4.81453925e-01 -1.27065361e-01 6.30524397e-01 1.03654671e+00
1.13006040e-01 -1.50755584e+00 -8.86630893e-01 6.35319427e-02
-3.77214588e-02 -9.70847905e-01 -6.78569734e-01 1.21505052e-01
-7.01717019e-01 -1.17850327e+00 -7.59326220e-01 -5.61799407e-01
3.57877702e-01 -2.21362814e-01 1.06101787e+00 -4.20783132e-01
-7.73056865e-01 8.98697913e-01 -6.81043208e-01 5.00738956e-02
-2.96968430e-01 -1.38788655e-01 -5.32312132e-02 -1.62463918e-01
4.78130847e-01 -6.98065579e-01 -4.47851062e-01 -7.27137085e-03
-9.01656210e-01 1.11715250e-01 6.94805026e-01 9.60680068e-01
7.73473799e-01 -8.86003077e-01 4.40179616e-01 -6.52891576e-01
4.32574540e-01 -5.10273099e-01 2.02577636e-01 3.57365012e-01
-2.14405894e-01 2.73879826e-01 1.09447218e-01 -6.16083264e-01
-7.53339052e-01 2.60410517e-01 -3.05762261e-01 -8.64004552e-01
-4.13032234e-01 2.56113708e-01 -3.14105779e-01 5.07482946e-01
7.00405061e-01 4.59499836e-01 2.71595418e-02 -4.26045328e-01
7.81847537e-01 5.08489430e-01 7.75292873e-01 -5.81991315e-01
4.25772846e-01 4.73507196e-01 -1.92866877e-01 -8.13777447e-01
-9.55514014e-01 -5.29199123e-01 -7.52770483e-01 -2.66990155e-01
1.31406343e+00 -9.41272020e-01 -2.49447078e-01 4.06168401e-01
-7.76140094e-01 -6.41402721e-01 -7.93000937e-01 4.92318004e-01
-1.26767778e+00 5.95163941e-01 -7.53588676e-01 -6.56331003e-01
-3.93801868e-01 -9.10524309e-01 1.58678913e+00 -1.04249634e-01
-7.72285700e-01 -8.18297505e-01 5.56846082e-01 3.40941966e-01
-8.86048004e-02 5.04040241e-01 7.21703708e-01 -6.20842755e-01
-3.82131517e-01 4.19857949e-02 2.90091187e-01 4.22834992e-01
3.54926646e-01 -2.00344101e-01 -7.17188716e-01 -2.91145533e-01
-3.31654608e-01 -1.00859559e+00 1.24507225e+00 2.54665107e-01
9.45448458e-01 -9.93342698e-03 -1.31709293e-01 4.55208749e-01
1.15862548e+00 5.80944605e-02 6.63647354e-01 -5.43164797e-02
6.32028520e-01 2.01383814e-01 7.52025366e-01 5.84269226e-01
-5.46715036e-02 6.12514675e-01 2.69066274e-01 2.97361284e-01
-5.63946486e-01 -6.21350527e-01 7.85202861e-01 4.34407651e-01
-7.77133368e-03 -2.74283826e-01 -7.47144043e-01 7.97601283e-01
-2.00031805e+00 -1.38234007e+00 4.15162712e-01 1.81220353e+00
9.36958253e-01 -7.97445420e-03 1.97247565e-01 -1.99847266e-01
5.73395371e-01 7.10277319e-01 -7.80313730e-01 -1.78908542e-01
1.57084897e-01 7.41057217e-01 1.30443946e-01 4.39728260e-01
-1.29557669e+00 1.07471097e+00 5.81050014e+00 9.48504567e-01
-7.58613765e-01 5.38986325e-01 8.53113756e-02 -6.50425315e-01
-1.39338121e-01 -2.42487445e-01 -6.37271106e-01 4.86510426e-01
8.87699008e-01 -9.00815874e-02 1.49494678e-01 8.90127718e-01
-1.32996947e-01 -1.40868956e-02 -1.41763091e+00 1.13975739e+00
5.02761245e-01 -1.28208971e+00 -8.10138974e-03 -4.97312397e-02
5.65376461e-01 2.98196346e-01 -7.54838511e-02 5.87609410e-01
5.51470935e-01 -8.50570083e-01 7.11030900e-01 5.06563127e-01
1.05630970e+00 -3.86686534e-01 1.19767450e-01 1.65777102e-01
-1.17950797e+00 -2.48334426e-02 -6.48512393e-02 1.60395399e-01
7.16339707e-01 8.49088356e-02 -4.85242605e-01 3.63123596e-01
3.94480884e-01 1.18446696e+00 -3.34013730e-01 4.76567447e-01
-8.13699782e-01 6.56725585e-01 2.85413954e-02 8.56961235e-02
3.14824820e-01 1.90898269e-01 6.18262768e-01 1.04294574e+00
3.21430683e-01 2.05613583e-01 2.52893627e-01 7.58032441e-01
8.34137872e-02 -2.08426908e-01 -7.10521519e-01 -5.40595710e-01
9.22977179e-02 8.62622201e-01 -4.29103464e-01 -5.32095313e-01
-4.34278876e-01 1.34114528e+00 2.20665216e-01 3.16419065e-01
-8.78307581e-01 -1.01732843e-01 1.04493141e+00 5.41187488e-02
7.72275448e-01 -1.43996522e-01 2.04190657e-01 -1.33214962e+00
-1.15257852e-01 -8.19055796e-01 7.75342286e-01 -8.99695575e-01
-1.43550241e+00 2.52678275e-01 1.63137212e-01 -1.29806685e+00
-7.54249990e-01 -6.07143462e-01 -5.30492544e-01 3.42033148e-01
-7.30793417e-01 -1.41464329e+00 3.54807451e-02 6.09006822e-01
1.22898638e+00 -2.88275629e-01 1.02936053e+00 4.00449447e-02
-3.99526417e-01 4.32384819e-01 -2.77896553e-01 2.71905988e-01
7.10515976e-01 -1.05946326e+00 7.24308491e-01 8.83325458e-01
5.85664630e-01 5.39863348e-01 5.08268058e-01 -9.84104335e-01
-1.26563251e+00 -1.08528113e+00 7.96804309e-01 -6.11406326e-01
7.86982358e-01 -5.80161273e-01 -5.29293954e-01 1.07830501e+00
2.11264744e-01 1.93366647e-01 1.06559420e+00 7.77927414e-02
-6.85841203e-01 4.24269170e-01 -7.69937932e-01 5.95685899e-01
1.79440320e+00 -8.41395855e-01 -1.08281517e+00 3.14598113e-01
6.80981219e-01 -2.26465493e-01 -7.25497723e-01 2.72039503e-01
7.40897238e-01 -8.00714016e-01 1.00112891e+00 -1.07963181e+00
8.67075682e-01 -3.94559614e-02 -4.55423951e-01 -1.26184595e+00
-3.81938428e-01 -4.19988573e-01 -4.83149886e-01 8.39355290e-01
2.52744168e-01 -2.59279788e-01 7.78396130e-01 2.07024306e-01
-4.37474787e-01 -1.00234282e+00 -1.14886391e+00 -1.00959992e+00
-1.78543389e-01 -5.11274993e-01 6.75823167e-02 6.71963394e-01
2.34418064e-01 4.21803832e-01 -5.96145034e-01 -3.77599865e-01
6.96751833e-01 1.57340139e-01 7.52845585e-01 -6.97114110e-01
-8.42119992e-01 -2.20955670e-01 -7.82442153e-01 -9.22613919e-01
4.73453462e-01 -1.01402307e+00 3.27176213e-01 -1.77695239e+00
4.26714540e-01 5.50404787e-01 -3.19171876e-01 1.02306330e+00
-4.92271483e-02 2.30879322e-01 2.36877888e-01 -3.63708511e-02
-8.94307375e-01 1.01769829e+00 1.05788231e+00 -1.78600580e-01
2.28089258e-01 -3.29025388e-01 -4.69437242e-01 5.26218534e-01
6.01734281e-01 -3.52684945e-01 -8.22067499e-01 -2.74818778e-01
-1.71769500e-01 4.52978350e-02 6.06265664e-01 -9.66576099e-01
-6.52629435e-02 -2.26942167e-01 2.25148872e-01 -5.31930864e-01
5.64553916e-01 -2.43953645e-01 3.50333638e-02 5.36380947e-01
-5.86103499e-01 -3.27959120e-01 -2.07236797e-01 8.48001659e-01
-8.69500712e-02 -7.97546506e-02 7.12819517e-01 -3.46878141e-01
-1.24520338e+00 5.53529561e-01 -4.02807742e-01 4.20405447e-01
1.24321067e+00 -2.70845830e-01 -2.26013899e-01 -3.54142755e-01
-1.27784622e+00 1.37521356e-01 5.59358656e-01 6.99675739e-01
9.09122705e-01 -1.55150568e+00 -5.21305561e-01 2.49411374e-01
6.28386021e-01 -2.86196321e-01 3.85813236e-01 7.44238377e-01
-3.65279354e-02 2.12275311e-01 -3.76745343e-01 -4.19568181e-01
-1.48625827e+00 4.41139698e-01 1.54413609e-02 -1.27130806e-01
-1.05469096e+00 1.15795946e+00 3.02216381e-01 -2.77698487e-01
3.09677541e-01 -3.19856584e-01 4.06976879e-01 -6.66232333e-02
5.37333131e-01 3.67399096e-01 -4.85000670e-01 -6.10635221e-01
-5.35905719e-01 3.18842471e-01 1.31864876e-01 -5.18017709e-01
1.48499393e+00 1.36410788e-01 3.63735974e-01 7.83767819e-01
1.13201761e+00 -4.86956626e-01 -1.44097900e+00 -2.17774689e-01
-6.81129172e-02 -4.95254338e-01 -3.47713709e-01 -7.19389319e-01
-8.68231833e-01 8.73807847e-01 5.13120890e-01 -6.14477396e-01
8.50244164e-01 7.60237992e-01 9.64397073e-01 2.12304220e-01
5.07988870e-01 -1.10730326e+00 5.81784308e-01 4.55627739e-01
9.03258681e-01 -1.05484605e+00 -1.91617802e-01 -1.76293299e-01
-9.83460903e-01 7.81618536e-01 5.68065584e-01 -1.97163537e-01
3.05746049e-01 2.94539124e-01 -1.17853090e-01 -5.33835232e-01
-1.02579951e+00 -6.01410866e-01 1.09974958e-01 6.33637071e-01
4.50652361e-01 8.50569680e-02 -2.44865417e-01 4.67454284e-01
1.12767503e-01 3.03247273e-01 1.33790880e-01 1.35786879e+00
-5.02657831e-01 -1.30340552e+00 1.71437338e-01 4.19344246e-01
-1.11286394e-01 -6.60286173e-02 -5.64987659e-01 7.23645151e-01
1.08290270e-01 4.59496021e-01 1.33339213e-02 -3.28602165e-01
2.43927270e-01 7.25208819e-01 8.87274802e-01 -1.17233300e+00
-8.78005382e-03 1.32479042e-01 2.80489624e-01 -1.13283622e+00
-5.56376576e-01 -1.02762842e+00 -1.52402067e+00 3.60228449e-01
3.95853817e-01 -3.54806155e-01 9.29931328e-02 9.21137273e-01
3.65308344e-01 4.86119568e-01 1.01264678e-01 -7.18685687e-01
-7.83776760e-01 -1.07020950e+00 -6.42786980e-01 6.48420632e-01
9.49403569e-02 -9.31666970e-01 -2.01939061e-01 3.53397161e-01] | [8.733853340148926, 0.9293658137321472] |
ee35cf5d-4142-4ccc-8f23-7570ad3a2b30 | adaptive-speech-quality-aware-complex-neural | 2210.16791 | null | https://arxiv.org/abs/2210.16791v3 | https://arxiv.org/pdf/2210.16791v3.pdf | Adaptive Speech Quality Aware Complex Neural Network for Acoustic Echo Cancellation with Supervised Contrastive Learning | Acoustic echo cancellation (AEC) is designed to remove echoes, reverberation, and unwanted added sounds from the microphone signal while maintaining the quality of the near-end speaker's speech. This paper proposes adaptive speech quality complex neural networks to focus on specific tasks for real-time acoustic echo cancellation. In specific, we propose a complex modularize neural network with different stages to focus on feature extraction, acoustic separation, and mask optimization receptively. Furthermore, we adopt the contrastive learning framework and novel speech quality aware loss functions to further improve the performance. The model is trained with 72 hours for pre-training and then 72 hours for fine-tuning. The proposed model outperforms the state-of-the-art performance. | ['Hantao Huang', 'Xiaoxi Yu', 'Bozhong Liu'] | 2022-10-30 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [ 1.10849598e-02 -3.80286336e-01 7.72272587e-01 -4.64894354e-01
-1.05965042e+00 -1.53548762e-01 1.20687686e-01 -3.32297474e-01
-7.31020629e-01 2.51022905e-01 5.25341809e-01 -3.52096081e-01
2.52573919e-02 -1.82617947e-01 -4.82573122e-01 -5.73080420e-01
-2.21386135e-01 -4.02491003e-01 1.07588530e-01 -2.23033518e-01
-8.46968889e-02 4.70350623e-01 -1.48205817e+00 2.41944388e-01
1.00057590e+00 1.14493740e+00 3.71061504e-01 1.07104528e+00
4.34617400e-02 4.89139944e-01 -1.00434959e+00 -3.02069962e-01
2.04559892e-01 -3.31095785e-01 1.51686341e-01 -3.51329178e-01
5.47608912e-01 -2.93032438e-01 -4.91467923e-01 1.05237699e+00
1.31106532e+00 5.45252681e-01 3.85543883e-01 -8.65210295e-01
-4.34944093e-01 5.35976052e-01 -8.48924834e-03 3.13479513e-01
-7.73785412e-02 2.63066858e-01 5.82429707e-01 -1.47884619e+00
-2.07509637e-01 1.27371395e+00 9.34976459e-01 7.27657020e-01
-5.89467287e-01 -1.15029049e+00 1.71402872e-01 3.27121347e-01
-1.22382498e+00 -1.42584741e+00 9.18685496e-01 5.15984697e-03
1.09763658e+00 4.88037109e-01 3.88688624e-01 1.04021394e+00
1.47203133e-01 5.17270803e-01 7.49614835e-01 -5.12444973e-01
3.02827090e-01 -1.22336753e-01 8.08816850e-02 3.95038158e-01
-4.20870006e-01 5.96216142e-01 -5.89537680e-01 5.35984151e-02
1.94840193e-01 -2.76903182e-01 -5.38384557e-01 4.43576187e-01
-5.60948491e-01 3.32639307e-01 2.43052930e-01 3.10878128e-01
-2.79598415e-01 2.56285220e-01 3.96103352e-01 3.97427857e-01
5.89566112e-01 3.01908076e-01 -4.55835938e-01 -6.13844432e-02
-1.11369431e+00 6.28054291e-02 7.54237950e-01 8.33507776e-01
1.68574765e-01 9.58175838e-01 -3.47555697e-01 1.45685124e+00
5.15284181e-01 9.75797296e-01 7.27089107e-01 -9.27669883e-01
6.04565620e-01 -4.30100828e-01 -1.59334943e-01 -9.87075269e-01
-3.25430602e-01 -1.03559327e+00 -1.09951842e+00 2.35464364e-01
-2.81628668e-01 -6.21502638e-01 -1.05327034e+00 1.68006790e+00
6.62640929e-02 5.82343817e-01 1.69455688e-02 1.00589752e+00
1.00476325e+00 1.03287160e+00 -2.71918058e-01 -4.10247952e-01
8.52836132e-01 -1.43244064e+00 -1.21160853e+00 -3.84716511e-01
-2.18971178e-01 -9.88019109e-01 9.52085018e-01 4.81471121e-01
-1.33792520e+00 -9.65607941e-01 -1.41156602e+00 1.08074844e-01
-1.49772868e-01 -1.22438498e-01 1.15055457e-01 1.05969882e+00
-1.10096407e+00 5.20630479e-01 -7.06173360e-01 6.11919880e-01
-1.58603072e-01 1.92357942e-01 -2.95671038e-02 8.95335972e-02
-1.27940977e+00 5.48499644e-01 -1.77843213e-01 5.38070083e-01
-9.94216323e-01 -7.72815764e-01 -7.98198402e-01 3.98116320e-01
1.43237010e-01 -3.57867360e-01 1.39724851e+00 -8.89369965e-01
-2.14188528e+00 1.65405609e-02 -1.81060940e-01 -3.76458049e-01
3.39867860e-01 -6.32432163e-01 -1.26693237e+00 -2.37271756e-01
-4.13353056e-01 2.79398859e-01 1.32292390e+00 -1.19264150e+00
-5.70304513e-01 -1.10734385e-02 -5.28650641e-01 3.20359319e-01
-4.12473232e-01 3.25846076e-01 -7.31846929e-01 -9.50322270e-01
2.33384043e-01 -4.44375068e-01 -2.32663661e-01 -3.71718913e-01
-7.25585744e-02 2.24167377e-01 5.25816619e-01 -1.09086847e+00
1.60657084e+00 -2.56491899e+00 -2.51747996e-01 1.60069004e-01
5.13628311e-02 7.56060958e-01 -5.57006896e-01 7.22724153e-03
-1.41410902e-01 -1.64064527e-01 -2.06704319e-01 -9.08179164e-01
2.50222325e-01 -2.63266921e-01 -2.42890686e-01 3.42693388e-01
1.92217112e-01 3.99420500e-01 -4.02385831e-01 -2.22784489e-01
1.84011221e-01 7.62476861e-01 -6.87615693e-01 5.69317102e-01
4.54940706e-01 1.91192091e-01 3.29942077e-01 3.62205356e-01
1.10692334e+00 6.71920359e-01 -3.26377451e-01 -4.53443006e-02
-1.97314575e-01 4.51781601e-01 -1.52872646e+00 1.39469373e+00
-9.38173950e-01 6.45944893e-01 1.03165853e+00 -5.64229488e-01
9.88447368e-01 5.66290319e-01 3.24033084e-03 -1.16484392e+00
6.58648536e-02 5.74415982e-01 2.08393127e-01 -2.82692671e-01
3.65280241e-01 -1.69707030e-01 3.49841803e-01 -1.68942958e-02
1.52634278e-01 -5.93219474e-02 -3.47874850e-01 -3.13725829e-01
1.04792881e+00 -5.11089444e-01 -2.03237399e-01 -1.02323830e-01
7.77225971e-01 -9.56044674e-01 9.32173133e-01 6.19391561e-01
-4.78989452e-01 7.19614327e-01 -4.57422644e-01 1.77058607e-01
-6.87147558e-01 -1.17875850e+00 7.95053989e-02 1.41784072e+00
-9.51169506e-02 -2.02266723e-01 -9.60547507e-01 -1.54352948e-01
-2.66575664e-01 7.21440434e-01 -6.35238644e-03 -2.11518630e-01
-9.49844480e-01 -2.16696307e-01 8.07921886e-01 3.19078475e-01
5.61527610e-01 -1.06259048e+00 -3.16694453e-02 4.26995665e-01
-1.81958914e-01 -8.05948555e-01 -1.16966379e+00 2.85202652e-01
-4.18099076e-01 -3.20630938e-01 -9.83032405e-01 -9.52021658e-01
1.46216094e-01 2.89762974e-01 9.81600165e-01 -5.12113571e-02
1.18437279e-02 1.88562408e-01 -1.82016745e-01 -6.37333274e-01
-4.03542697e-01 -1.57633141e-01 3.70420218e-01 1.85652480e-01
-1.96064085e-01 -7.52632797e-01 -7.29033589e-01 3.79367381e-01
-7.10909545e-01 -3.45880508e-01 4.78549004e-01 8.00798953e-01
3.52017760e-01 3.89046192e-01 7.52719164e-01 -2.46070534e-01
1.13310862e+00 -8.53821561e-02 -4.64519233e-01 2.99194772e-02
-5.80786228e-01 -2.64217824e-01 8.70891273e-01 -6.29398704e-01
-1.39468145e+00 -3.07230979e-01 -8.43684793e-01 -4.91828710e-01
3.28530930e-02 1.70291319e-01 -5.00592887e-01 7.24781305e-02
5.06708264e-01 1.35991096e-01 -2.12907925e-01 -7.90129960e-01
4.29152809e-02 1.08752763e+00 7.39737272e-01 -8.46239254e-02
7.42217004e-01 -2.07566991e-01 -3.93121779e-01 -9.03199255e-01
-3.12281191e-01 -5.72354078e-01 2.50167307e-02 -2.31053546e-01
3.34977269e-01 -1.06498516e+00 -5.42006910e-01 8.98936152e-01
-1.23665643e+00 -3.36931527e-01 3.69446725e-02 9.08839285e-01
-8.23917612e-02 1.08883850e-01 -8.58981490e-01 -1.17958450e+00
-8.68964195e-01 -1.07643449e+00 6.68657601e-01 1.87572151e-01
3.68937701e-01 -5.08012474e-01 1.66626751e-01 4.28965315e-02
1.14009535e+00 -6.14259958e-01 3.61339688e-01 -3.79461974e-01
-2.57487565e-01 -4.17684130e-02 1.19715467e-01 1.06235933e+00
2.65157707e-02 -2.24722788e-01 -1.48688149e+00 -3.76794249e-01
3.77853692e-01 2.15997577e-01 9.50661242e-01 6.14129305e-01
8.98470700e-01 -4.78782207e-01 1.35655880e-01 8.97645295e-01
1.01171482e+00 7.13879824e-01 6.87998414e-01 -2.13305488e-01
3.66187543e-01 4.94775146e-01 2.40257084e-01 2.48902038e-01
-1.53545424e-01 5.55790246e-01 4.04153652e-02 -3.38320166e-01
-5.95109522e-01 -6.40172139e-02 6.62477493e-01 1.73633933e+00
3.56510580e-01 -4.69784945e-01 -4.88665789e-01 3.70092481e-01
-1.30168557e+00 -8.51478934e-01 1.46593466e-01 2.05392265e+00
7.83104897e-01 1.97166160e-01 -3.05162162e-01 3.50644648e-01
7.87316442e-01 3.54150325e-01 -5.24150789e-01 -6.14315093e-01
-1.41918659e-01 6.15692675e-01 3.51697773e-01 1.08226120e+00
-1.00427663e+00 7.19584465e-01 6.46550655e+00 1.16522563e+00
-1.34048212e+00 3.18937778e-01 5.16088963e-01 -3.35194737e-01
-3.50985676e-01 -6.40831351e-01 -6.26102209e-01 3.41639102e-01
1.21596634e+00 2.22875714e-01 8.49303067e-01 5.73950112e-01
3.84345531e-01 4.58138883e-01 -7.32210636e-01 1.25497913e+00
2.13227764e-01 -7.35983908e-01 -5.05200148e-01 -3.12878489e-01
4.02213961e-01 6.76788837e-02 3.91497433e-01 8.03865135e-01
-1.70316145e-01 -1.09280705e+00 9.55740273e-01 6.80700600e-01
7.15978324e-01 -8.67710888e-01 5.63647568e-01 2.59517998e-01
-1.28458989e+00 -2.58669108e-01 -2.20900312e-01 2.42812075e-02
1.77488625e-01 8.68342340e-01 -5.10102153e-01 2.42237046e-01
7.66374052e-01 -4.35312715e-04 -2.76418090e-01 1.28194988e+00
-1.52226258e-03 9.37204659e-01 -4.14731175e-01 -9.39864367e-02
7.97382370e-02 1.12484053e-01 7.16577411e-01 1.68846917e+00
5.00615180e-01 7.42990747e-02 -1.66698635e-01 4.02752906e-01
-1.68548897e-01 2.11704805e-01 -8.97422209e-02 3.77947450e-01
6.76669359e-01 9.12809432e-01 1.40677495e-02 -8.01256448e-02
-2.84955412e-01 1.06006765e+00 2.93999910e-02 6.97754025e-01
-8.81985426e-01 -1.06697547e+00 7.04885185e-01 -2.59674489e-01
2.25575596e-01 -3.16031128e-01 -1.57136440e-01 -7.43724227e-01
2.89913446e-01 -1.03472936e+00 -3.94094467e-01 -6.81615651e-01
-1.03800094e+00 9.87163186e-01 -6.14181936e-01 -9.00074363e-01
-1.58072755e-01 -6.96096003e-01 -9.03374612e-01 1.21635222e+00
-1.58940256e+00 -4.26206321e-01 -6.98500797e-02 5.22069633e-01
7.34552920e-01 -4.28496689e-01 7.49097526e-01 9.82707500e-01
-6.38929248e-01 1.08361793e+00 4.46105868e-01 8.75466615e-02
9.21471119e-01 -1.00443852e+00 7.18601227e-01 1.11492431e+00
-1.67868793e-01 6.27483904e-01 6.93288982e-01 -3.04570675e-01
-1.03316605e+00 -1.07522082e+00 9.37997937e-01 2.41047189e-01
2.18528435e-01 -6.79707825e-01 -8.88058066e-01 -9.06621069e-02
3.82218122e-01 4.26151864e-02 6.13781929e-01 8.09872150e-02
-3.99953395e-01 -7.64514267e-01 -9.28222656e-01 6.20803297e-01
1.00523877e+00 -6.16435409e-01 -3.94355357e-01 -2.20894784e-01
1.19896269e+00 -3.76866430e-01 -3.84879410e-01 5.29294550e-01
6.05485976e-01 -9.89754498e-01 9.79588330e-01 -1.46333605e-01
-1.02668710e-01 -1.02942072e-01 -4.68279690e-01 -1.62548935e+00
-4.56946105e-01 -1.04030049e+00 -8.94001350e-02 1.44376969e+00
6.89782679e-01 -6.50210261e-01 3.47424954e-01 4.16097105e-01
-8.97083580e-01 -4.73978430e-01 -9.95071828e-01 -7.69737422e-01
6.10881634e-02 -8.60049069e-01 6.46800995e-01 4.10305232e-01
-3.89565885e-01 4.97529298e-01 -6.07137203e-01 4.02620256e-01
4.56654787e-01 -5.55306733e-01 3.54186654e-01 -8.02819192e-01
-5.67274928e-01 -6.32866561e-01 2.59694517e-01 -1.28925359e+00
6.76593510e-03 -3.29906702e-01 7.30202973e-01 -1.14257216e+00
-5.46075284e-01 -2.60154456e-01 -6.86380506e-01 -9.56270248e-02
-4.42722082e-01 -8.10289904e-02 4.27578449e-01 -4.32456702e-01
-5.04029989e-01 1.02002645e+00 1.07423759e+00 -3.75659287e-01
-4.98916715e-01 4.24123526e-01 -4.50925171e-01 5.20796180e-01
7.55270004e-01 -4.74168777e-01 -4.33949828e-01 -6.86026573e-01
-1.81734964e-01 2.65137911e-01 7.09184706e-02 -1.39084184e+00
5.42003572e-01 2.30627403e-01 1.54654428e-01 -5.74552834e-01
6.12398207e-01 -9.33604002e-01 -3.54768410e-02 3.31723064e-01
-5.76606274e-01 -1.20707735e-01 4.07661140e-01 4.28542912e-01
-4.57255125e-01 -2.88837641e-01 9.28446293e-01 1.20475464e-01
-1.43409267e-01 2.05664724e-01 -6.17908478e-01 8.17197412e-02
1.60648659e-01 1.81158289e-01 -8.97189882e-03 -6.77358150e-01
-6.60472572e-01 9.31632519e-02 -4.63285863e-01 3.73954743e-01
9.58634317e-01 -1.31729090e+00 -9.69851911e-01 5.21907210e-01
-3.37245554e-01 -2.92088181e-01 6.78794742e-01 4.17633146e-01
-2.49464050e-01 4.09288853e-01 1.37872219e-01 -9.36746374e-02
-1.29526615e+00 3.44616890e-01 8.76270473e-01 2.14541052e-02
-3.47914010e-01 1.16121495e+00 2.19403595e-01 -6.99138284e-01
1.04761028e+00 -3.37314785e-01 -1.88401386e-01 -4.70488995e-01
8.24271917e-01 7.59936094e-01 4.24313694e-01 -4.60253507e-01
-2.92346269e-01 3.71940017e-01 1.14096433e-01 -6.37699485e-01
1.01164520e+00 -3.79060775e-01 1.98522270e-01 2.98799843e-01
1.45104003e+00 6.66707397e-01 -1.11918843e+00 -3.50966454e-01
-3.31349164e-01 -2.04337835e-01 5.01050472e-01 -1.16608512e+00
-1.15479851e+00 1.16815579e+00 1.26457858e+00 -9.14301127e-02
1.62514114e+00 -7.23962963e-01 1.03390467e+00 4.01663095e-01
-1.21962808e-01 -1.35172498e+00 1.84282109e-01 8.54396462e-01
1.19386768e+00 -9.18501794e-01 -5.67621708e-01 -1.14423357e-01
-3.58306676e-01 8.02744925e-01 5.74667931e-01 -8.27769190e-02
1.06614304e+00 6.49850547e-01 4.88510430e-01 3.96537602e-01
-6.34653568e-01 1.58879580e-03 5.50151289e-01 6.37716174e-01
4.70846176e-01 7.26338923e-02 -5.59870037e-04 1.04630828e+00
-2.97577530e-01 -5.07049739e-01 -4.36716117e-02 4.14327025e-01
-7.08059907e-01 -6.82586014e-01 -5.77160597e-01 6.62855059e-02
-7.06237853e-01 -6.91689312e-01 -1.78829968e-01 -1.89717021e-02
2.12728798e-01 1.60352993e+00 3.33607122e-02 -5.57606816e-01
8.90253901e-01 1.48649991e-01 2.93080751e-02 -1.98735163e-01
-8.33932638e-01 7.90033221e-01 1.10579535e-01 -3.85212779e-01
4.24097963e-02 -1.83948874e-01 -1.06629574e+00 -1.75293922e-01
-5.35092652e-01 1.64697766e-01 9.36074495e-01 5.38583875e-01
4.03756738e-01 1.22536314e+00 1.04216194e+00 -9.18736517e-01
-6.49337947e-01 -1.36814368e+00 -6.69751644e-01 9.49658304e-02
9.79753017e-01 -1.69113114e-01 -6.26335323e-01 -2.14282498e-01] | [14.999635696411133, 5.944546699523926] |
16a55983-f9d5-4372-998d-de9880e5ef26 | who-wins-the-game-of-thrones-how-sentiments | 2003.07683 | null | https://arxiv.org/abs/2003.07683v1 | https://arxiv.org/pdf/2003.07683v1.pdf | Who Wins the Game of Thrones? How Sentiments Improve the Prediction of Candidate Choice | This paper analyzes how candidate choice prediction improves by different psychological predictors. To investigate this question, it collected an original survey dataset featuring the popular TV series "Game of Thrones". The respondents answered which character they anticipated to win in the final episode of the series, and explained their choice of the final candidate in free text from which sentiments were extracted. These sentiments were compared to feature sets derived from candidate likeability and candidate personality ratings. In our benchmarking of 10-fold cross-validation in 100 repetitions, all feature sets except the likeability ratings yielded a 10-11% improvement in accuracy on the holdout set over the base model. Treating the class imbalance with synthetic minority oversampling (SMOTE) increased holdout set performance by 20-34% but surprisingly not testing set performance. Taken together, our study provides a quantified estimation of the additional predictive value of psychological predictors. Likeability ratings were clearly outperformed by the feature sets based on personality, emotional valence, and basic emotions. | ['Chaehan So'] | 2020-02-29 | null | null | null | null | ['holdout-set'] | ['computer-vision'] | [-1.19153477e-01 3.64738017e-01 -6.05957210e-01 -6.95303202e-01
-7.62339175e-01 -3.58414829e-01 3.99256259e-01 5.37829161e-01
-4.92201895e-01 8.00611019e-01 3.28442067e-01 2.14974016e-01
-2.24415809e-01 -8.31865489e-01 -1.35813370e-01 -3.08064282e-01
1.98348284e-01 4.12188649e-01 -3.37669790e-01 -7.63217270e-01
5.39918542e-01 -7.42029548e-02 -1.80728376e+00 4.18395251e-01
6.68820918e-01 1.45607758e+00 -5.38768351e-01 2.74743646e-01
3.31662118e-01 9.98626649e-01 -5.45851767e-01 -1.14947474e+00
2.69955218e-01 -4.19376284e-01 -6.08174741e-01 1.07835688e-01
2.58793235e-01 -4.62884493e-02 3.82979549e-02 6.91299796e-01
4.39684927e-01 3.27342480e-01 4.55360979e-01 -1.22300494e+00
-3.80093724e-01 8.90488803e-01 -3.21071565e-01 -1.87985420e-01
8.03844512e-01 -1.24219377e-02 1.57289553e+00 -8.86408865e-01
6.31693244e-01 9.66215968e-01 6.80413365e-01 3.81467611e-01
-1.44364679e+00 -8.76530170e-01 -7.80730471e-02 1.35936230e-01
-1.09501886e+00 -3.74367625e-01 8.69919002e-01 -3.63581628e-01
5.82735837e-01 6.83195233e-01 8.81600201e-01 1.09316969e+00
1.74956903e-01 5.34237504e-01 1.60802090e+00 -1.68167397e-01
4.18509662e-01 9.22364473e-01 5.06667078e-01 1.66275948e-01
2.45363608e-01 1.21680032e-02 -6.95201993e-01 -4.69874203e-01
1.27974227e-01 -3.60915542e-01 7.85396248e-02 2.28660733e-01
-8.52452099e-01 1.12996960e+00 2.10366428e-01 -9.48347747e-02
-5.14701724e-01 -3.93465847e-01 3.61419141e-01 6.73632085e-01
6.52576268e-01 1.14095879e+00 -5.48955858e-01 -4.67288703e-01
-9.33542252e-01 5.66299736e-01 1.03008997e+00 3.73112798e-01
8.56245399e-01 -3.88488173e-02 -9.54371765e-02 1.01155853e+00
-2.92901218e-01 3.02745938e-01 1.00028110e+00 -1.01145029e+00
2.02564672e-02 7.61303484e-01 2.36756504e-01 -1.48039329e+00
-7.16125011e-01 -8.21834147e-01 -4.53783989e-01 -2.70597376e-02
3.31921995e-01 -4.22747761e-01 -1.45710990e-01 1.68712091e+00
5.31856604e-02 -6.33302867e-01 -2.93782622e-01 8.13225031e-01
8.96385431e-01 2.45635793e-01 9.23608840e-02 -4.07217056e-01
1.46174026e+00 -5.64349949e-01 -3.75206381e-01 -3.59688908e-01
5.21297216e-01 -5.28596938e-01 1.16591179e+00 8.70894432e-01
-1.08458853e+00 -9.18830335e-01 -1.08867073e+00 3.60635191e-01
-4.79627438e-02 2.90607717e-02 8.96389961e-01 1.08183670e+00
-7.15425849e-01 8.57148051e-01 -4.00127731e-02 -1.73288867e-01
1.01494849e-01 6.34425104e-01 -3.82333457e-01 3.11895758e-01
-1.35427177e+00 8.69120538e-01 1.09462170e-02 -5.91304123e-01
-1.94052398e-01 -5.42402983e-01 -6.57437503e-01 1.02254190e-01
2.07565904e-01 -3.78566623e-01 9.77810502e-01 -1.69687116e+00
-1.48056829e+00 1.31657970e+00 -1.31078605e-02 -4.21263695e-01
3.17736119e-01 -5.90544529e-02 -5.84345520e-01 -3.51824760e-01
2.40059048e-01 4.55593318e-01 6.20669603e-01 -8.14931810e-01
-5.13141274e-01 -3.71283323e-01 1.70564398e-01 1.45966411e-01
-4.76504922e-01 1.45888329e-01 1.65066510e-01 -4.86842632e-01
3.82998656e-03 -1.17979884e+00 -3.35911602e-01 -9.31957364e-01
-2.36603916e-01 -1.74375564e-01 -2.62919724e-01 -4.04028744e-01
1.47555375e+00 -2.06632972e+00 -1.28966719e-01 3.63327503e-01
2.93605715e-01 -3.25655043e-01 1.26860470e-01 4.42364693e-01
-3.04909736e-01 3.54568720e-01 5.54659903e-01 -3.48728858e-02
2.90921211e-01 -3.10485512e-01 -1.59801722e-01 2.95177817e-01
9.59782973e-02 5.21855593e-01 -5.61087370e-01 -2.20109746e-01
-1.44288763e-01 -1.02608196e-01 -8.90817106e-01 -1.41165406e-01
1.61919847e-01 -5.04158549e-02 -2.25145236e-01 5.48358083e-01
4.43045020e-01 -2.72409260e-01 2.50971407e-01 1.69206671e-02
1.06115311e-01 6.28445566e-01 -8.02020490e-01 9.27510619e-01
-1.60721347e-01 6.20935082e-01 -5.70978642e-01 -6.02987766e-01
1.69081700e+00 4.73022349e-02 4.48868185e-01 -9.50365484e-01
3.31777871e-01 3.12903941e-01 3.13863873e-01 -2.98870683e-01
9.53727007e-01 -6.13849938e-01 -8.56542230e-01 1.82204530e-01
-2.39116877e-01 -1.86078265e-01 7.13756979e-02 -6.96718507e-03
9.31204379e-01 -1.35733649e-01 4.84376192e-01 -2.20814824e-01
4.89206761e-01 2.99892664e-01 8.49178731e-01 5.43177366e-01
-3.17902356e-01 4.74443644e-01 1.02923751e+00 -5.80739200e-01
-7.75635362e-01 -5.45726061e-01 -1.14903674e-01 1.42678559e+00
-1.05419993e-01 -6.34741962e-01 -4.27611023e-01 -3.86565179e-01
6.67039678e-02 1.06618369e+00 -9.28247452e-01 -3.98487985e-01
-1.17068840e-02 -1.06887758e+00 3.64091605e-01 1.51498988e-01
2.19413042e-01 -8.67622852e-01 -4.27083403e-01 -7.50262523e-03
-4.21202272e-01 -7.14222431e-01 1.54096842e-01 3.30273181e-01
-6.49116695e-01 -7.40568638e-01 -2.34889105e-01 -4.41012979e-01
8.56279731e-02 -1.90914527e-01 1.49483836e+00 -1.13775037e-01
2.58442551e-01 1.21792713e-02 -4.64634925e-01 -5.04017055e-01
-6.40350878e-02 2.93377757e-01 2.30596989e-01 2.46711507e-01
9.13627267e-01 -6.13966703e-01 -3.88345450e-01 3.20741504e-01
-2.40401596e-01 -9.85945240e-02 3.17870796e-01 7.43026733e-01
3.40872228e-01 1.36712650e-02 9.47768688e-01 -1.06063747e+00
8.68575811e-01 -8.16264033e-01 -1.83341503e-02 -4.07700539e-01
-9.24453497e-01 -2.35762984e-01 5.64974010e-01 -4.61438626e-01
-9.12390709e-01 -1.00313880e-01 -7.28837922e-02 1.65604696e-01
1.20726572e-02 7.18172848e-01 2.21867815e-01 2.16939554e-01
9.97829318e-01 -2.81238556e-01 -7.39909783e-02 -1.33906931e-01
-1.71691120e-01 4.38502491e-01 1.94362521e-01 -4.74993050e-01
5.36539674e-01 9.38981920e-02 -1.37251198e-01 -4.80263144e-01
-1.02399421e+00 -2.64867783e-01 -3.34814250e-01 -3.62356275e-01
4.32081133e-01 -1.04568470e+00 -1.08510506e+00 -3.22362827e-03
-4.20953274e-01 6.71173781e-02 -4.22817498e-01 7.18398213e-01
-4.36972797e-01 -2.53398359e-01 -5.01106143e-01 -9.18711483e-01
-6.64211586e-02 -8.03520918e-01 4.28600997e-01 2.95371205e-01
-1.13289905e+00 -7.47248888e-01 2.67609984e-01 7.65835464e-01
2.41266146e-01 4.89320874e-01 8.22049737e-01 -1.24378586e+00
3.08047801e-01 -7.07796156e-01 2.39013940e-01 1.12675086e-01
-2.57483900e-01 -6.20484948e-02 -1.07857251e+00 7.30976015e-02
2.68716484e-01 -7.92741597e-01 6.57479584e-01 3.48043889e-01
7.70071089e-01 -1.95361778e-01 1.66089669e-01 2.24357650e-01
1.41280675e+00 9.08166692e-02 6.49106026e-01 7.39332974e-01
8.31266940e-02 9.21045244e-01 8.56823683e-01 9.99339461e-01
4.70475376e-01 5.76749086e-01 1.69414759e-01 2.00885892e-01
5.67602336e-01 -2.18732283e-01 6.26829922e-01 6.83970213e-01
-2.21490920e-01 5.16568534e-02 -6.27488971e-01 2.59132832e-01
-1.50031745e+00 -1.08393645e+00 -3.19017231e-01 2.20764947e+00
4.97351587e-01 5.42675972e-01 6.65163159e-01 5.99523604e-01
5.22087276e-01 1.48526341e-01 -4.56574082e-01 -8.92324388e-01
-3.93388331e-01 2.70880193e-01 5.17564833e-01 3.37694824e-01
-7.90396571e-01 6.21207952e-01 6.58123589e+00 8.37905288e-01
-1.15001428e+00 -2.59857059e-01 1.42629099e+00 -4.73663151e-01
-4.53758568e-01 2.32158713e-02 -6.60961449e-01 3.62486184e-01
1.38615203e+00 -3.30683053e-01 2.50097781e-01 1.07983863e+00
2.74934620e-01 -4.03696358e-01 -9.15941358e-01 9.55783308e-01
2.57722646e-01 -9.22157645e-01 -3.88137728e-01 2.51173377e-01
7.28484511e-01 -3.18242490e-01 4.23274487e-01 9.49364066e-01
1.41860232e-01 -1.22139859e+00 7.85735607e-01 5.35083592e-01
5.82446396e-01 -1.02811754e+00 9.32871699e-01 3.77479583e-01
-3.13033968e-01 -2.54322290e-01 -3.71979445e-01 -1.02984369e+00
-3.94955635e-01 6.47030830e-01 -8.56609523e-01 1.99544817e-01
5.40516615e-01 5.00085592e-01 -9.72646952e-01 4.73683447e-01
2.64227092e-01 9.11707819e-01 -1.05607204e-01 -3.65169495e-01
1.03836812e-01 -1.19349033e-01 2.19263598e-01 7.95002460e-01
1.93250537e-01 2.20965683e-01 -5.78474775e-02 5.44669330e-01
-8.51780772e-02 5.31311214e-01 -2.14520305e-01 -4.82278280e-02
3.18237394e-01 1.47690761e+00 -6.65597677e-01 -2.64813542e-01
-3.86764050e-01 5.53158998e-01 2.37103358e-01 -2.25433543e-01
-9.87138093e-01 -3.82914543e-01 5.33573389e-01 5.38741708e-01
-7.13751540e-02 5.63388109e-01 -9.22708213e-01 -1.08774269e+00
-1.72054529e-01 -1.22046840e+00 2.36825988e-01 -6.31727874e-01
-1.43173099e+00 7.60161459e-01 -3.17437172e-01 -1.11619234e+00
-2.02467352e-01 -4.24893111e-01 -5.96453071e-01 7.73670077e-01
-7.33303607e-01 -5.26486576e-01 -3.55468601e-01 1.54502243e-01
2.25341231e-01 -1.77605763e-01 9.32011485e-01 1.75928235e-01
-4.95699883e-01 7.50985563e-01 -1.55370548e-01 3.73251699e-02
9.99803722e-01 -1.11614633e+00 -3.41621488e-02 2.51098932e-03
-2.30048791e-01 4.63661462e-01 9.42990005e-01 -5.06177366e-01
-8.30712378e-01 -5.52013934e-01 1.16135681e+00 -5.86300611e-01
5.38522661e-01 -8.64047185e-03 -4.44609165e-01 4.69339222e-01
-3.57307419e-02 -5.52469313e-01 1.31371260e+00 6.43914759e-01
-1.74025133e-01 -2.12297261e-01 -1.17957151e+00 5.04556060e-01
4.17470932e-01 -3.70970249e-01 -4.32255417e-01 -3.01350541e-02
2.67511725e-01 1.16285365e-02 -1.18924129e+00 2.22035527e-01
1.03558683e+00 -1.60636985e+00 6.94526851e-01 -6.87047958e-01
1.04659307e+00 3.99548382e-01 -5.20783365e-01 -1.25353014e+00
-8.12975824e-01 -5.51681161e-01 7.80231357e-01 1.25874853e+00
7.72537768e-01 -4.48106438e-01 1.21526647e+00 1.12019420e+00
2.21929312e-01 -1.17388296e+00 -4.84862238e-01 -2.83470064e-01
2.23683938e-01 -6.60718501e-01 6.98866546e-01 1.03445232e+00
6.40093565e-01 6.94644094e-01 -5.90629339e-01 -6.80355251e-01
2.62337565e-01 6.86511546e-02 8.27914059e-01 -1.56894445e+00
-5.03932953e-01 -5.90594888e-01 -4.89251286e-01 -4.96164024e-01
1.09663695e-01 -8.89550865e-01 -4.75488663e-01 -9.44866300e-01
3.30760002e-01 -3.75640780e-01 -4.18784559e-01 7.70134777e-02
-2.99913079e-01 5.02980947e-01 2.25769982e-01 9.56799090e-02
-5.51757336e-01 3.82919163e-01 9.40373182e-01 1.29277408e-01
-4.15470302e-01 2.60025442e-01 -1.39330244e+00 8.93339097e-01
9.31691229e-01 -3.90810132e-01 -2.50172973e-01 4.39769417e-01
8.51597846e-01 5.14965057e-01 7.53620714e-02 -1.09772122e+00
-2.76090711e-01 -4.12664622e-01 7.51260698e-01 -1.85565561e-01
8.67753983e-01 -3.42557341e-01 3.43047112e-01 3.33065212e-01
-9.37443495e-01 2.62824982e-01 1.73131060e-02 2.32933089e-01
-1.87819436e-01 -1.23205774e-01 5.06702423e-01 -1.98541239e-01
-3.92623037e-01 -1.63706020e-01 -5.55097401e-01 8.13351795e-02
9.34746504e-01 -4.49252576e-01 -2.04474106e-02 -9.51540589e-01
-8.79963219e-01 -3.07145298e-01 6.42363131e-01 3.54041755e-01
3.23005795e-01 -1.36495805e+00 -9.99523997e-01 9.63847414e-02
5.12752645e-02 -9.72778559e-01 1.49706230e-01 1.09189808e+00
-9.92528573e-02 2.97818065e-01 -5.77483475e-01 -5.13370987e-03
-1.19064343e+00 3.41314912e-01 1.03127390e-01 -3.97315502e-01
2.76472449e-01 9.72955406e-01 -5.29484451e-02 -3.77299875e-01
-1.88507453e-01 -5.93215413e-02 -4.70006138e-01 6.46614134e-01
2.48505548e-01 6.57711625e-01 -3.23929228e-02 -8.61865401e-01
-3.77465874e-01 1.10029340e-01 6.12705685e-02 -3.93320024e-01
1.24407303e+00 -1.61479384e-01 -2.58024931e-02 9.93336618e-01
1.10441232e+00 5.69401205e-01 -5.30659020e-01 1.33651003e-01
1.39064297e-01 -5.08817852e-01 -1.42539546e-01 -8.41430426e-01
-8.12533677e-01 2.94713169e-01 7.45450109e-02 4.97155696e-01
9.19072509e-01 -4.67532039e-01 5.27927995e-01 1.89513221e-01
3.68429095e-01 -1.24769282e+00 -2.20590155e-03 3.86699736e-01
7.04191625e-01 -1.45632541e+00 2.66506374e-01 -1.85188577e-01
-1.35897171e+00 1.01258028e+00 6.93606079e-01 -5.50918460e-01
6.80816054e-01 -3.50739539e-01 -7.60481954e-02 -1.39896259e-01
-1.26670742e+00 1.36203349e-01 3.09985340e-01 2.48489112e-01
8.65279853e-01 5.30616522e-01 -7.07589090e-01 1.57730699e+00
-1.06094706e+00 -1.90887615e-01 8.48479927e-01 2.29541346e-01
-4.92473572e-01 -8.86216462e-01 -3.63507271e-01 9.77549970e-01
-8.08493018e-01 1.00497352e-02 -6.88021660e-01 7.64459908e-01
3.32976609e-01 1.35227573e+00 1.28423214e-01 -1.30037546e+00
2.18107760e-01 9.29225609e-02 1.35516711e-02 -3.93599719e-01
-1.36919260e+00 -5.09383678e-02 8.08433950e-01 -4.47752506e-01
-2.23050803e-01 -8.92319083e-01 -8.09518576e-01 -7.94014871e-01
-3.04506034e-01 5.19764125e-01 3.31292152e-01 6.35612726e-01
3.66720185e-02 3.02007735e-01 1.00600529e+00 -6.21361971e-01
-7.18670905e-01 -9.69878316e-01 -8.43546808e-01 6.91245854e-01
-9.93357524e-02 -5.18628597e-01 -6.39361739e-01 -3.45028281e-01] | [9.39883804321289, 10.198674201965332] |
37841ffa-a5a3-451a-ae8f-df7eff8273f8 | csvc-net-code-switched-voice-command | null | null | https://ieeexplore.ieee.org/document/9564183 | https://ieeexplore.ieee.org/document/9564183 | CSVC-Net: Code-Switched Voice Command Classification using Deep CNN-LSTM Network | Colloquial Bengali has adopted many English words due to colonial influence. In conversational Bengali, it is quite common to speak in a mixture of English and Bengali, a phenomenon termed Code-switching (CS). To build a Voice Command Classifier in this era, when the usage of CS is ever-increasing, it is often necessary to map a single base command to its many different variants - spoken in multiple mixtures of languages. The works done with Bengali Speech have been primarily focused on single word classification and mostly incompetent in understanding the complex semantic relationships displayed in sentences. This paper proposes ‘CSVC-Net’, a CNN-LSTM based architecture for classifying spoken commands that exhibit code-switching between Bengali and English. To effectively reflect the scenario, it also presents a newly curated dataset named ‘Banglish’ containing 3,840 audio files of spoken computer commands belonging to 11 classes, considering 64 variations in total. The proposed pipeline passes the input audio signal through a series of appropriate transformation and augmentation steps enabling the model to achieve an accuracy of 92.08% on the curated dataset. Furthermore, the robustness of the proposed model has been justified by comparing with different architectures and tested under different noise levels with promising accuracy, which shows the applicability of the model in real-life scenarios. | ['Md. Hasanul Kabir', 'Sabbir Ahmed', 'Fariha Ishrat Rahman', 'Arowa Yasmeen'] | 2021-08-17 | null | null | null | international-conference-on-informatics-1 | ['voice-query-recognition'] | ['speech'] | [ 1.22919366e-01 -2.36937612e-01 6.37811542e-01 -4.73526746e-01
-6.69433236e-01 -5.04493713e-01 5.98484874e-01 -3.07228994e-02
-5.09909332e-01 4.48448747e-01 1.30810663e-01 -4.50592101e-01
1.03617184e-01 -4.75135207e-01 -3.58697355e-01 -5.96906543e-01
-1.10297903e-01 4.82175857e-01 3.28881405e-02 -4.68653262e-01
2.35489875e-01 4.76333171e-01 -1.77082121e+00 5.20979285e-01
8.07457745e-01 6.60331011e-01 4.69144076e-01 1.00256515e+00
-3.06584418e-01 7.75655031e-01 -1.03066647e+00 -3.04106534e-01
-1.85694620e-01 -7.22131789e-01 -8.59140396e-01 -1.19247288e-01
3.31991673e-01 6.01843074e-02 -1.36227056e-01 8.87852788e-01
7.44232297e-01 2.11716831e-01 4.86566216e-01 -7.70071328e-01
-4.15704876e-01 1.11023510e+00 2.55496353e-01 2.95075208e-01
4.74198312e-01 -2.15750396e-01 8.33731592e-01 -1.05172431e+00
3.60587180e-01 1.05814624e+00 5.94604731e-01 5.18573582e-01
-1.04444623e+00 -6.64422512e-01 -2.06703156e-01 6.14563704e-01
-1.69424140e+00 -6.04616940e-01 6.08403563e-01 -3.59074324e-01
1.27215803e+00 6.76646590e-01 7.14225173e-01 1.08884013e+00
8.73212814e-02 6.21360004e-01 1.32669663e+00 -6.36223912e-01
1.80117548e-01 5.18134475e-01 -6.27374798e-02 1.18894964e-01
-5.44986308e-01 -4.13041800e-01 -6.64345801e-01 1.43033087e-01
2.71901153e-02 -3.66163164e-01 -5.00489950e-01 2.63198912e-01
-1.05328369e+00 7.70008862e-01 1.21227123e-01 9.41381752e-01
-3.73701394e-01 -3.29749256e-01 5.98074853e-01 5.64122915e-01
2.96658605e-01 2.67728329e-01 -4.71261144e-01 -5.45596659e-01
-1.07607651e+00 1.88933417e-01 1.07914853e+00 1.00204051e+00
2.21279830e-01 6.87319458e-01 4.10501540e-01 1.33163702e+00
8.80280659e-02 3.04102957e-01 8.03099096e-01 -1.95461109e-01
2.95310229e-01 3.43264371e-01 -4.97680217e-01 -9.01921213e-01
-4.69002962e-01 -5.71555078e-01 -7.89474070e-01 -3.66338715e-02
1.75698698e-01 -2.07774118e-01 -8.36277544e-01 1.44266665e+00
3.53221921e-03 -2.41559893e-01 3.09024274e-01 7.45532155e-01
1.03562105e+00 9.07240927e-01 -2.98104465e-01 -1.90932885e-01
1.14539731e+00 -6.53300047e-01 -9.96619701e-01 2.40987808e-01
3.10929269e-01 -1.08432949e+00 1.38243520e+00 7.70574152e-01
-8.71881306e-01 -6.14602864e-01 -1.13849771e+00 2.92481512e-01
-7.00698376e-01 1.56909257e-01 -1.02071665e-01 8.74704421e-01
-1.12119424e+00 2.73134083e-01 -3.88325006e-01 -7.89969742e-01
-8.38440508e-02 2.99533516e-01 -3.82749945e-01 2.26381421e-01
-1.31134784e+00 1.12942207e+00 5.38100541e-01 2.65812546e-01
-8.86555493e-01 -2.76984781e-01 -5.00540733e-01 3.10415551e-02
1.84853673e-02 3.19968671e-01 1.09126735e+00 -9.95680749e-01
-1.84432387e+00 6.55064464e-01 -9.61119533e-02 -4.19003546e-01
5.82561374e-01 -1.73399717e-01 -9.16189611e-01 -1.93293408e-01
-4.86335576e-01 1.46817908e-01 5.76475859e-01 -1.18558884e+00
-8.60786676e-01 -2.57599920e-01 -1.51125014e-01 3.03469747e-01
-3.47442061e-01 3.87655914e-01 -1.83028698e-01 -6.99948072e-01
2.83963323e-01 -9.00969028e-01 4.13069189e-01 -9.93860602e-01
-5.26584089e-01 -1.15481459e-01 8.62248778e-01 -9.64102685e-01
1.30860519e+00 -2.21085072e+00 2.07569316e-01 1.56331405e-01
-2.29235500e-01 4.00372416e-01 9.64610428e-02 8.24618757e-01
-3.69775966e-02 -5.46564832e-02 -4.34492022e-01 -2.80246109e-01
-1.59914702e-01 2.72258341e-01 -2.96047390e-01 4.31010932e-01
2.68397599e-01 3.15818071e-01 -4.37710434e-01 -2.66741455e-01
1.38649926e-01 7.37126052e-01 -1.54965550e-01 3.99394989e-01
1.94688395e-01 6.13993466e-01 1.18158564e-01 6.91839218e-01
6.99278176e-01 8.09487998e-01 3.32098037e-01 3.11280182e-03
-5.59624791e-01 4.33112383e-01 -1.03708160e+00 1.42828751e+00
-7.08056629e-01 1.00113273e+00 2.96924770e-01 -1.13450754e+00
1.19033086e+00 6.07065678e-01 -1.82256978e-02 -6.02459490e-01
3.70083123e-01 6.68612242e-01 4.82129902e-01 -6.72816932e-01
5.06250381e-01 -9.29966345e-02 9.11682565e-03 1.81493387e-01
4.47787672e-01 -6.10766053e-01 -1.42819270e-01 -3.28769863e-01
8.21009099e-01 -1.20934583e-01 1.26518101e-01 -3.97212923e-01
8.67928326e-01 -2.32134108e-02 1.17699102e-01 5.91789305e-01
-4.23914790e-02 6.66742325e-01 1.43599480e-01 -2.61978388e-01
-9.34184372e-01 -8.61510754e-01 -3.81131172e-01 1.29577696e+00
-5.89744687e-01 -1.17919691e-01 -9.34232354e-01 -3.13383713e-02
-5.73370039e-01 9.95013118e-01 -4.06486452e-01 4.74458113e-02
-7.92766869e-01 -5.65382481e-01 1.05581367e+00 -6.01694472e-02
6.86210990e-01 -1.23557210e+00 -6.31223381e-01 4.13827479e-01
-2.09491253e-01 -1.16670549e+00 -3.56196687e-02 6.11923218e-01
-3.30328912e-01 -8.67721081e-01 -5.69577694e-01 -9.49383318e-01
1.70283653e-02 -1.19348779e-01 9.64260757e-01 -9.93997827e-02
-1.14041559e-01 1.05087474e-01 -9.27562654e-01 -6.95275247e-01
-1.02177024e+00 4.34908569e-01 2.46093407e-01 2.38704205e-01
3.79357249e-01 -5.10828972e-01 1.44241244e-01 6.05872907e-02
-8.65330398e-01 -2.65046954e-01 4.02719676e-01 7.01983869e-01
-1.19055130e-01 1.84490252e-03 6.10403121e-01 -7.79921234e-01
7.86395609e-01 -4.77404118e-01 -1.34766370e-01 1.10822007e-01
-2.13888943e-01 -5.11673391e-01 9.02924001e-01 -5.38571477e-01
-9.05561745e-01 -9.24402475e-02 -6.28318548e-01 -2.32376475e-02
-4.33082253e-01 7.17255473e-01 -3.70849848e-01 1.03634357e-01
4.50905949e-01 8.10344517e-01 -1.07497901e-01 -5.11639059e-01
2.38296568e-01 1.53687298e+00 6.48576021e-01 -2.81709045e-01
5.67643821e-01 2.92712334e-03 -6.30900919e-01 -1.56213355e+00
-1.95892721e-01 -3.01641792e-01 -9.03456271e-01 -4.64969516e-01
9.91268873e-01 -7.17058897e-01 -5.57294786e-01 1.10012591e+00
-1.22376561e+00 -2.44634032e-01 1.63878024e-01 5.35944879e-01
-3.12106907e-01 7.87375271e-02 -5.24456203e-01 -1.00423443e+00
-1.37093306e-01 -1.31792092e+00 2.61990219e-01 5.22982068e-02
-3.36154312e-01 -7.68051505e-01 -1.70370296e-01 5.17694987e-02
8.83859575e-01 -1.57316715e-01 1.11748314e+00 -1.36629450e+00
8.15633871e-03 -1.87721446e-01 3.29147041e-01 6.91955268e-01
4.57661718e-01 2.82295972e-01 -1.34137392e+00 -2.68248141e-01
2.96291590e-01 -3.34919989e-01 4.12694395e-01 -1.32568792e-01
7.06264079e-01 -1.88136816e-01 4.67202306e-01 3.31888169e-01
1.02229178e+00 8.33030701e-01 4.32464093e-01 3.76557261e-01
5.33800781e-01 8.74499083e-01 3.26384783e-01 2.77802944e-01
3.15631241e-01 7.77302682e-01 2.56361097e-01 1.22927494e-01
4.56097489e-03 1.50986150e-01 5.78042924e-01 1.93264449e+00
-3.30007374e-02 -2.19964296e-01 -1.17074776e+00 5.37781298e-01
-1.18002176e+00 -7.44782329e-01 -1.59975052e-01 2.17620873e+00
9.67807472e-01 1.04063965e-01 -1.97834126e-03 5.06085098e-01
7.76213944e-01 6.55384511e-02 -1.01324521e-01 -1.06216812e+00
-3.67258310e-01 5.21777868e-01 -1.39408363e-02 5.71386278e-01
-1.06510997e+00 1.03548467e+00 5.72120619e+00 8.80660236e-01
-1.62931764e+00 2.58386940e-01 1.83512226e-01 1.84953690e-01
-5.73672354e-02 -2.81188846e-01 -4.92020577e-01 5.10380208e-01
1.52366269e+00 -2.84688622e-02 5.13776183e-01 5.09383023e-01
2.86559761e-01 2.99705099e-02 -8.70371938e-01 9.06326652e-01
5.91983080e-01 -9.98216748e-01 1.99821949e-01 -3.26069653e-01
3.23886484e-01 2.68483341e-01 -4.63402905e-02 3.94452900e-01
8.06625560e-03 -1.27860761e+00 1.15296865e+00 3.02398890e-01
7.98628807e-01 -9.78298008e-01 1.06438839e+00 3.92451137e-01
-1.08934796e+00 -1.59000218e-01 -2.38763273e-01 -6.32528663e-02
-9.97368395e-02 -5.53641170e-02 -1.27833605e+00 5.02233565e-01
9.45553303e-01 2.19638556e-01 -4.64165181e-01 6.96822643e-01
1.33894160e-01 9.93037641e-01 -2.59628087e-01 -2.70251036e-01
3.04751396e-01 -2.81211406e-01 7.36974001e-01 1.76554739e+00
5.25814831e-01 -6.06227331e-02 -3.42435211e-01 6.29563153e-01
3.10188860e-01 4.56290781e-01 -5.65510631e-01 -1.22390777e-01
7.83970296e-01 9.78765488e-01 -7.17844963e-01 -2.33308509e-01
-3.51289481e-01 1.06847477e+00 -3.34498249e-02 3.40541244e-01
-7.93175459e-01 -6.51384652e-01 5.33068478e-01 -2.00467974e-01
2.06797510e-01 -3.44945878e-01 4.60606478e-02 -7.76738167e-01
-9.79684144e-02 -1.23970592e+00 -4.98401299e-02 -5.51490724e-01
-8.29556108e-01 1.46523976e+00 -1.08365685e-01 -1.23099589e+00
-3.51434588e-01 -4.95200783e-01 -6.47184432e-01 1.03365684e+00
-9.47057724e-01 -1.39277768e+00 -3.31579179e-01 5.08121431e-01
1.21194601e+00 -7.47144520e-01 1.29972410e+00 6.09672606e-01
-4.89469498e-01 6.22645795e-01 2.13220850e-01 2.14690596e-01
6.09865665e-01 -1.32235599e+00 2.75675446e-01 8.63402486e-01
2.73900926e-01 4.66314048e-01 9.31441665e-01 -1.66802540e-01
-1.15795743e+00 -8.21030736e-01 9.20196652e-01 -8.15944225e-02
6.62207723e-01 -8.10035050e-01 -1.18031132e+00 4.05030429e-01
6.18859053e-01 -6.28998160e-01 7.70351708e-01 -2.81432122e-01
-1.15749769e-01 -1.44449696e-01 -8.11010242e-01 4.38760161e-01
4.84849960e-01 -8.60145032e-01 -7.55041361e-01 -3.74956168e-02
6.74113452e-01 -3.18949640e-01 -7.62530267e-01 1.65305167e-01
4.33439016e-01 -1.23905897e+00 3.18306446e-01 -5.16167939e-01
1.81256887e-02 -6.90722391e-02 -6.00250602e-01 -1.55040097e+00
2.99946517e-01 -4.09646660e-01 4.08069372e-01 1.63937354e+00
4.26471949e-01 -6.50556266e-01 1.73107207e-01 1.03151277e-01
-5.83623171e-01 -5.22433281e-01 -1.36504841e+00 -4.27525759e-01
1.47111610e-01 -7.39531338e-01 5.57113290e-01 1.27223384e+00
-1.30826443e-01 1.74465343e-01 -2.56275624e-01 -3.91617306e-02
1.37523527e-03 -3.91725957e-01 6.86538935e-01 -1.12447262e+00
2.76343711e-02 -5.31745672e-01 -5.91411054e-01 -4.16492492e-01
1.87311739e-01 -9.01024163e-01 2.28149489e-01 -1.04390824e+00
-3.21431667e-01 -2.89669484e-01 -1.03507847e-01 3.35841745e-01
2.26027250e-01 3.42293829e-01 3.59334707e-01 1.84956059e-01
1.72637165e-01 4.14508194e-01 5.33684671e-01 -1.60547033e-01
-2.83697039e-01 1.98283106e-01 -3.33097652e-02 6.13583982e-01
8.11517417e-01 -3.08775872e-01 -2.24159062e-01 -2.96073854e-01
-1.29029807e-02 -9.34767947e-02 -4.52992953e-02 -1.37666070e+00
2.83767819e-01 2.81760812e-01 -2.05197912e-02 -7.63418913e-01
4.04396504e-01 -8.31537247e-01 4.01458979e-01 5.17275274e-01
-2.96320498e-01 1.99475542e-01 2.70858258e-01 -6.48492277e-02
-6.90309346e-01 -3.07165742e-01 9.05911446e-01 -9.74064022e-02
-7.68387198e-01 -4.20660496e-01 -1.08576345e+00 -2.26652995e-01
7.80128419e-01 -5.12737036e-01 8.72141272e-02 -4.45141912e-01
-8.90526474e-01 -2.45488241e-01 2.28284728e-02 6.80745602e-01
6.79089665e-01 -9.60761666e-01 -1.00905240e+00 4.17408496e-01
5.62748546e-03 -3.13543260e-01 3.76634508e-01 8.92027915e-01
-9.68407452e-01 4.71951455e-01 -3.88402969e-01 -6.17901385e-01
-1.38870871e+00 -1.25654966e-01 4.97917980e-01 5.31106412e-01
-4.01752710e-01 9.37100112e-01 -4.88457739e-01 -8.81595910e-01
3.35028321e-01 -2.67039210e-01 -7.42026389e-01 6.16646171e-01
1.85050443e-01 3.45217943e-01 5.41964412e-01 -1.12195802e+00
-4.51422960e-01 2.96048552e-01 6.05072975e-02 -4.02111679e-01
1.31533265e+00 -2.40652233e-01 -4.94022071e-01 1.24806345e+00
1.14784288e+00 2.89592832e-01 -5.29227138e-01 1.27000865e-02
2.50638366e-01 -2.46013448e-01 -1.78825200e-01 -7.48682022e-01
-8.42388630e-01 9.88182425e-01 8.81765306e-01 4.60829794e-01
1.04668593e+00 -1.79712966e-01 4.77035522e-01 5.59500933e-01
2.33770311e-01 -1.25071800e+00 -4.32231098e-01 9.62907910e-01
1.07800937e+00 -8.48293960e-01 -5.55376291e-01 1.44948632e-01
-8.25892627e-01 1.31404805e+00 3.86428028e-01 1.84675366e-01
5.99350035e-01 3.06904465e-01 6.77398086e-01 1.09503806e-01
-5.21183372e-01 -1.89244628e-01 -1.11826451e-03 6.93703830e-01
8.48145783e-01 8.00363347e-02 -3.37757915e-01 6.19757950e-01
-9.19273138e-01 -4.49117988e-01 1.02157938e+00 7.22654939e-01
-3.22782427e-01 -1.02165985e+00 -4.95289922e-01 2.77429610e-01
-7.45500743e-01 -2.89389819e-01 -5.14994323e-01 9.06429648e-01
2.11040467e-01 1.30618918e+00 3.03891003e-01 -7.61555970e-01
5.52086890e-01 5.70327520e-01 -2.55596191e-02 -5.68756521e-01
-9.45891500e-01 1.80734232e-01 1.25042975e-01 -9.49057713e-02
-7.20056891e-01 -5.56978166e-01 -1.19312823e+00 -2.15546370e-01
-3.39508325e-01 4.88769710e-01 1.07096744e+00 1.00352871e+00
-8.19046497e-02 7.79252231e-01 7.00460196e-01 -8.12379718e-01
-5.55019498e-01 -1.59087384e+00 -7.58347392e-01 1.74446419e-01
4.28066075e-01 -3.27173889e-01 -3.82248044e-01 8.52978826e-02] | [14.153274536132812, 6.523498058319092] |
dd81ad46-9407-493d-a9e1-ca24a5a5aa4d | post-processing-independent-evaluation-of | 2306.15440 | null | https://arxiv.org/abs/2306.15440v1 | https://arxiv.org/pdf/2306.15440v1.pdf | Post-Processing Independent Evaluation of Sound Event Detection Systems | Due to the high variation in the application requirements of sound event detection (SED) systems, it is not sufficient to evaluate systems only in a single operating mode. Therefore, the community recently adopted the polyphonic sound detection score (PSDS) as an evaluation metric, which is the normalized area under the PSD receiver operating characteristic (PSD-ROC). It summarizes the system performance over a range of operating modes resulting from varying the decision threshold that is used to translate the system output scores into a binary detection output. Hence, it provides a more complete picture of the overall system behavior and is less biased by specific threshold tuning. However, besides the decision threshold there is also the post-processing that can be changed to enter another operating mode. In this paper we propose the post-processing independent PSDS (piPSDS) as a generalization of the PSDS. Here, the post-processing independent PSD-ROC includes operating points from varying post-processings with varying decision thresholds. Thus, it summarizes even more operating modes of an SED system and allows for system comparison without the need of implementing a post-processing and without a bias due to different post-processings. While piPSDS can in principle combine different types of post-processing, we hear, as a first step, present median filter independent PSDS (miPSDS) results for this year's DCASE Challenge Task4a systems. Source code is publicly available in our sed_scores_eval package (https://github.com/fgnt/sed_scores_eval). | ['Romain Serizel', 'Reinhold Haeb-Umbach', 'Janek Ebbers'] | 2023-06-27 | null | null | null | null | ['sound-event-detection'] | ['audio'] | [ 1.24705113e-01 -5.13924062e-01 4.71467227e-01 -3.62104535e-01
-1.06966031e+00 -8.78535628e-01 3.99488032e-01 4.62645918e-01
-5.49234092e-01 1.44436777e-01 -2.73871068e-02 -4.38515037e-01
-2.96584696e-01 -5.39426088e-01 -3.35590780e-01 -6.73003078e-01
-3.10260326e-01 -1.00919761e-01 9.01973486e-01 5.06423265e-02
1.72309965e-01 5.12385607e-01 -2.01628995e+00 4.71925259e-01
4.68597353e-01 1.04715288e+00 3.90817732e-01 1.11855459e+00
2.28717282e-01 -2.27797985e-01 -9.71645415e-01 4.19499502e-02
2.08976865e-01 -4.46669370e-01 -2.12512061e-01 -6.46533251e-01
1.40813127e-01 -1.50963649e-01 4.55652215e-02 9.68862474e-01
9.97220576e-01 1.23236530e-01 5.91828644e-01 -1.33009708e+00
7.82990754e-02 4.39094961e-01 5.84544763e-02 5.93034685e-01
5.96266091e-01 2.93780774e-01 8.83771002e-01 -7.39936650e-01
5.73280565e-02 1.07697463e+00 6.84852481e-01 1.29534677e-02
-1.38142347e+00 -7.56293535e-01 -4.41025555e-01 6.36051372e-02
-1.43722224e+00 -3.01822662e-01 6.00036383e-01 -5.83164454e-01
9.21098232e-01 5.62324941e-01 4.73513484e-01 9.80457783e-01
2.62339950e-01 3.46468985e-01 1.18137002e+00 -4.44542736e-01
5.29623151e-01 1.46584034e-01 3.68833452e-01 -8.21327493e-02
2.39034250e-01 2.91201174e-01 -4.27378386e-01 -5.02731681e-01
3.67340505e-01 -3.45283121e-01 -3.58400106e-01 1.68963805e-01
-8.66590083e-01 4.17718530e-01 -3.94738391e-02 3.89680743e-01
-1.40288875e-01 5.36626056e-02 4.74917561e-01 5.75642049e-01
2.24210054e-01 4.17231351e-01 -5.24043918e-01 -5.85978270e-01
-1.20300519e+00 3.15692246e-01 9.40876424e-01 5.35975814e-01
4.67097938e-01 1.05136447e-02 -4.00461733e-01 9.26082611e-01
2.18330666e-01 6.30337715e-01 3.70300978e-01 -8.05799305e-01
1.42329171e-01 2.51407605e-02 9.59338993e-02 -7.02409744e-01
-7.79378951e-01 -3.11136872e-01 -2.40309879e-01 3.32481802e-01
5.71910918e-01 -3.50212157e-01 -6.83319986e-01 1.63274074e+00
-4.26680110e-02 -1.59294710e-01 -1.02908991e-01 7.49537230e-01
6.94798827e-01 6.29672885e-01 6.51692897e-02 -3.40051562e-01
1.56326091e+00 -2.79239774e-01 -5.74801147e-01 -7.74212256e-02
3.53838831e-01 -1.11053121e+00 1.26171780e+00 6.35486722e-01
-8.31672907e-01 -6.23789668e-01 -1.34926724e+00 5.31774998e-01
-4.84506667e-01 2.70218253e-02 6.52335733e-02 1.03109705e+00
-1.07942557e+00 6.97932780e-01 -9.77653921e-01 -4.80091125e-01
-2.60941833e-01 2.49490991e-01 -2.30911989e-02 5.27247727e-01
-1.33290625e+00 6.56094909e-01 1.05543338e-01 -3.38524878e-01
-6.33318365e-01 -8.36699665e-01 -4.98386115e-01 1.62501991e-01
2.54158109e-01 -1.24847390e-01 1.62965906e+00 -3.47849756e-01
-1.46919203e+00 4.22101051e-01 1.00973248e-01 -3.17329556e-01
5.54833055e-01 -1.82160005e-01 -7.97061086e-01 1.78967729e-01
-7.32077807e-02 1.41598806e-01 9.36746061e-01 -9.24317479e-01
-6.04216397e-01 -1.21530797e-03 -9.65384617e-02 -1.30532205e-01
-7.89248720e-02 4.21255052e-01 -2.41779253e-01 -5.90172410e-01
6.98998049e-02 -8.60221982e-01 1.27405494e-01 -3.67435753e-01
-2.63408244e-01 4.06514630e-02 6.72050714e-01 -5.41399181e-01
1.67932570e+00 -2.56016254e+00 -5.26347041e-01 1.63436696e-01
-2.13046953e-01 3.20785701e-01 -4.30907197e-02 7.27406442e-01
-3.01663995e-01 1.58391967e-01 -3.31830621e-01 -2.53239244e-01
1.39662549e-01 -1.80349693e-01 -3.11173052e-01 2.43234381e-01
2.69679457e-01 1.09681852e-01 -7.67396629e-01 -1.27076924e-01
4.32787240e-01 4.33254659e-01 -5.16702116e-01 9.12284106e-02
1.78899243e-01 1.87559500e-01 -5.73777482e-02 2.55126625e-01
7.40739048e-01 3.97307664e-01 -1.22478664e-01 -2.54581511e-01
-5.27394533e-01 5.61916113e-01 -1.51684713e+00 1.25638747e+00
-4.88725156e-01 7.42932558e-01 8.30633566e-03 -7.19792902e-01
9.25941586e-01 5.08892059e-01 4.69420105e-01 -4.79810208e-01
2.05491960e-01 4.29272056e-01 3.47003847e-01 -4.47946221e-01
3.33505154e-01 -6.02133125e-02 -2.86351174e-01 4.77940530e-01
1.79393053e-01 -4.68295693e-01 3.75684261e-01 -5.12211174e-02
1.37342250e+00 -1.89534545e-01 2.48046204e-01 -4.39013600e-01
3.61813813e-01 -3.81381810e-01 2.10872188e-01 8.58899653e-01
-3.32012773e-01 9.93005157e-01 6.56026602e-01 2.74239689e-01
-6.20926142e-01 -1.64776886e+00 -8.19456637e-01 1.04264843e+00
-1.30075142e-01 -5.91248930e-01 -7.23913252e-01 -1.76532805e-01
5.12837879e-02 8.85207951e-01 -2.50999868e-01 -2.51267016e-01
-2.54104942e-01 -8.27927113e-01 1.01175046e+00 2.56275624e-01
2.35009447e-01 -1.11060679e+00 -9.67946172e-01 1.63995266e-01
-2.20078994e-02 -1.02244616e+00 -3.59999329e-01 6.63184285e-01
-6.37535214e-01 -7.31214821e-01 -5.53110540e-01 -2.46078491e-01
4.83151637e-02 1.61539093e-01 8.36860597e-01 -2.10406154e-01
-2.33251035e-01 6.57267034e-01 -5.80934644e-01 -6.94236577e-01
-6.38113260e-01 -1.52782723e-01 3.39428157e-01 -2.22207218e-01
1.52577043e-01 -8.88558626e-01 -7.33206570e-01 4.14595872e-01
-1.03827500e+00 -5.35282612e-01 4.42785293e-01 3.14199895e-01
4.14784133e-01 2.09365994e-01 7.77988672e-01 -4.22780454e-01
9.38777685e-01 -5.08266687e-01 -5.91653228e-01 -7.68674538e-02
-4.87884998e-01 -8.17859694e-02 6.08908415e-01 -4.39036429e-01
-6.92924559e-01 -2.03864142e-01 -4.93862212e-01 -1.69966549e-01
-4.40360099e-01 3.43983203e-01 -1.23672426e-01 3.94158095e-01
8.76121461e-01 1.03261517e-02 -1.14096224e-01 -5.92277467e-01
-7.11296946e-02 1.09665823e+00 4.53649461e-01 -3.26188475e-01
6.92770898e-01 1.19713828e-01 -3.01010787e-01 -1.00585699e+00
-4.35860366e-01 -7.07463324e-01 -3.94998282e-01 -3.10139269e-01
7.38697052e-01 -6.27471030e-01 -4.73955840e-01 5.56071281e-01
-9.58277106e-01 -2.04025745e-01 -2.38016501e-01 7.12232888e-01
-5.12704015e-01 3.12700510e-01 -4.07254666e-01 -1.19817770e+00
-3.30203891e-01 -9.53109443e-01 8.98970902e-01 3.77907604e-01
-5.23741245e-01 -6.11551046e-01 2.61927694e-01 -3.27603310e-01
3.77403408e-01 1.61645740e-01 5.63972890e-01 -8.20231497e-01
1.54721141e-01 -4.06052649e-01 7.76111856e-02 6.50615215e-01
2.27940902e-01 3.79293233e-01 -1.42926931e+00 -2.49550134e-01
2.19454542e-01 4.71991718e-01 6.37093306e-01 5.08003175e-01
9.94718134e-01 5.63169047e-02 -5.78122512e-02 2.79105097e-01
1.20157146e+00 6.05558455e-01 6.84769034e-01 2.87939817e-01
-5.80418408e-02 5.29040575e-01 7.08362520e-01 4.65294659e-01
-1.04992606e-01 8.19612622e-01 2.26550236e-01 3.01770922e-02
-1.31440833e-01 -8.51244479e-02 7.94102788e-01 5.72228849e-01
1.17469691e-01 -3.75621498e-01 -8.17865551e-01 3.60524625e-01
-1.27589023e+00 -9.78035092e-01 -3.61532092e-01 2.70516920e+00
6.29887223e-01 4.87156808e-01 4.67172682e-01 7.03858793e-01
8.17666888e-01 1.16252974e-01 -2.50002503e-01 -8.28370571e-01
2.21293896e-01 5.88061690e-01 3.70669901e-01 3.60838413e-01
-1.11053491e+00 1.89341769e-01 6.13234568e+00 9.35320556e-01
-1.19640744e+00 1.20266870e-01 -6.36495650e-02 -1.13129839e-01
1.61535010e-01 -1.67266820e-02 -6.00723028e-01 7.27883875e-01
1.51567638e+00 -3.38776439e-01 9.66842175e-02 6.55104458e-01
5.32705605e-01 -6.08765483e-01 -1.06532848e+00 9.59441364e-01
-3.72052163e-01 -5.90973973e-01 -3.56993228e-01 6.88382145e-03
8.92148539e-03 3.66195701e-02 1.26047194e-01 3.95317167e-01
-3.37133646e-01 -3.50911319e-01 9.81340349e-01 2.95691282e-01
9.64602888e-01 -5.07307768e-01 7.51963556e-01 2.16304377e-01
-1.30293727e+00 -2.25802153e-01 -4.43547927e-02 -1.61223203e-01
3.35833102e-01 8.26438546e-01 -8.79291773e-01 5.62145114e-01
1.01007748e+00 6.65023848e-02 -4.39099103e-01 1.41148615e+00
-4.71685529e-02 8.95083845e-01 -6.26391709e-01 -6.82710335e-02
-3.26169670e-01 1.73937544e-01 9.25490677e-01 1.67948771e+00
7.73412466e-01 -4.16414216e-02 -1.98881239e-01 6.94224894e-01
6.00662231e-01 6.22610413e-02 -2.21463084e-01 -2.33793519e-02
7.49405086e-01 1.28269207e+00 -8.17164361e-01 -1.07485307e-02
-2.37556800e-01 5.14643133e-01 -5.16900897e-01 2.85274565e-01
-9.44332123e-01 -8.87224317e-01 7.55634785e-01 4.28184271e-01
1.22894265e-01 -2.87848026e-01 -2.56345063e-01 -5.21756232e-01
7.02599213e-02 -4.80997056e-01 5.15651405e-01 -7.57185280e-01
-1.13948822e+00 5.39269388e-01 5.64215183e-01 -1.69865501e+00
-1.74744427e-01 -5.02460182e-01 -9.55159962e-01 8.20711613e-01
-1.02208662e+00 -1.46009624e-01 -2.52875865e-01 2.11724386e-01
3.73558581e-01 2.18158558e-01 8.83381188e-01 5.59755981e-01
-3.90524596e-01 6.53009236e-01 5.05682491e-02 -2.01966584e-01
9.92958307e-01 -1.26805663e+00 3.49235952e-01 8.92848194e-01
-5.40114939e-03 5.15828550e-01 1.21500981e+00 -3.42538208e-01
-8.18988562e-01 -7.51517892e-01 5.56444824e-01 -4.02066082e-01
7.82466173e-01 -4.43451464e-01 -9.84758317e-01 3.28461677e-02
-1.90036789e-01 -2.93908179e-01 7.72741914e-01 2.93786656e-02
-1.79080710e-01 -3.54236960e-01 -9.73082542e-01 4.03207421e-01
6.55120194e-01 -5.59081435e-01 -6.12422943e-01 -1.10203966e-01
6.74959421e-01 -2.90586829e-01 -1.01345170e+00 4.72979575e-01
6.38740122e-01 -1.45736909e+00 8.30338001e-01 2.64230192e-01
-7.81254768e-02 -5.56196630e-01 -2.50556141e-01 -1.15744722e+00
3.44913304e-02 -6.18700206e-01 2.53115624e-01 1.38972485e+00
5.87646067e-01 -9.24698353e-01 5.95611600e-05 3.89487922e-01
-4.77279276e-01 -5.62328577e-01 -1.03306258e+00 -1.21157348e+00
-1.34262860e-01 -1.26231170e+00 4.04700369e-01 3.78648520e-01
1.26905590e-01 -7.60811344e-02 1.50726274e-01 4.39695925e-01
3.25769335e-01 -2.54063159e-01 5.67180395e-01 -1.25877476e+00
-5.50844312e-01 -4.68481869e-01 -5.03627539e-01 -5.84938467e-01
-6.77098036e-01 -7.42850125e-01 2.56753117e-01 -1.16346300e+00
-1.76604107e-01 -4.20117974e-01 -7.47024179e-01 4.16127145e-01
-1.13806399e-02 2.15906829e-01 3.75136077e-01 1.42431125e-01
-1.73889562e-01 7.83322901e-02 6.49135828e-01 2.52187908e-01
-6.77069843e-01 6.05719328e-01 -4.32036579e-01 6.99076891e-01
9.93570566e-01 -8.34360182e-01 -3.68468434e-01 3.32742445e-02
9.13013071e-02 4.53863926e-02 5.88995755e-01 -1.55647540e+00
5.77077270e-02 3.01798791e-01 2.34248415e-02 -6.22840106e-01
3.54826212e-01 -4.99710560e-01 3.71246904e-01 6.76853836e-01
4.47896756e-02 -7.97222480e-02 5.91412783e-01 3.74751896e-01
-1.53125376e-01 -3.44432652e-01 8.54710400e-01 3.81239802e-01
-4.60853815e-01 -3.04188073e-01 -8.19930255e-01 -1.10243827e-01
7.97888815e-01 -4.16628391e-01 -3.24954420e-01 -3.01556051e-01
-8.98785889e-01 -1.07437856e-01 2.13641495e-01 3.64382565e-01
2.23420367e-01 -1.02523386e+00 -4.51993793e-01 2.26490363e-01
1.87810272e-01 -5.25027394e-01 5.36833942e-01 1.03182065e+00
-3.86452466e-01 2.27178082e-01 -1.49396285e-01 -8.25366437e-01
-1.36978889e+00 2.19995320e-01 3.75908703e-01 3.16916928e-02
-3.93207520e-01 6.10796452e-01 3.11485138e-02 -1.31708488e-01
1.61878899e-01 -6.74653649e-01 -2.14058831e-01 2.78864056e-01
5.94342709e-01 5.47535658e-01 4.63872164e-01 -2.74223864e-01
-6.35825396e-01 3.87461811e-01 4.25269216e-01 -5.86209953e-01
1.01829755e+00 -3.23068202e-02 2.04581395e-01 9.01790619e-01
1.06501603e+00 2.03779787e-01 -9.78618205e-01 2.41345584e-01
1.59104094e-02 -2.30259299e-01 -1.71213858e-02 -7.36291111e-01
-5.96854568e-01 8.07303488e-01 1.03660214e+00 8.42912674e-01
1.51457477e+00 1.60785690e-02 4.34618980e-01 -7.79996102e-04
3.30141217e-01 -1.09321928e+00 -1.77467987e-01 4.74433273e-01
1.01780534e+00 -6.53291047e-01 -1.11319959e-01 -2.92067468e-01
-3.58613819e-01 1.14839089e+00 1.92549244e-01 -3.71030346e-02
1.07423913e+00 6.25365853e-01 2.16885522e-01 1.00146435e-01
-5.40049791e-01 -2.42659807e-01 2.35335529e-01 6.03281736e-01
4.88937676e-01 1.41155332e-01 -6.26352191e-01 1.13435507e+00
-6.11681700e-01 -3.11232746e-01 5.02916098e-01 8.45273077e-01
-6.01586223e-01 -1.23182905e+00 -5.90275645e-01 6.67944133e-01
-5.59403956e-01 -4.80470732e-02 -1.23561092e-01 5.90899348e-01
1.26313671e-01 1.41217148e+00 9.51074660e-02 -8.05538833e-01
8.62519145e-01 1.76200882e-01 1.72355369e-01 -7.07577586e-01
-8.19346666e-01 1.41524032e-01 4.21661325e-02 -4.42226231e-01
-2.08075233e-02 -9.64787006e-01 -1.36330295e+00 -2.24127155e-02
-3.94930989e-01 2.15328813e-01 7.78808475e-01 6.34948313e-01
2.44848654e-01 8.99675548e-01 6.40269637e-01 -8.73241484e-01
-5.57223141e-01 -1.07602119e+00 -7.65474498e-01 7.17682764e-02
2.60357589e-01 -7.34755456e-01 -1.01589835e+00 -1.57690033e-01] | [15.278253555297852, 5.371995449066162] |
e3ff1804-931f-463f-9b63-78d74cf57ce1 | filtered-guided-diffusion-fast-filter | 2306.17141 | null | https://arxiv.org/abs/2306.17141v1 | https://arxiv.org/pdf/2306.17141v1.pdf | Filtered-Guided Diffusion: Fast Filter Guidance for Black-Box Diffusion Models | Recent advances in diffusion-based generative models have shown incredible promise for Image-to-Image translation and editing. Most recent work in this space relies on additional training or architecture-specific adjustments to the diffusion process. In this work, we show that much of this low-level control can be achieved without additional training or any access to features of the diffusion model. Our method simply applies a filter to the input of each diffusion step based on the output of the previous step in an adaptive manner. Notably, this approach does not depend on any specific architecture or sampler and can be done without access to internal features of the network, making it easy to combine with other techniques, samplers, and diffusion architectures. Furthermore, it has negligible cost to performance, and allows for more continuous adjustment of guidance strength than other approaches. We show FGD offers a fast and strong baseline that is competitive with recent architecture-dependent approaches. Furthermore, FGD can also be used as a simple add-on to enhance the structural guidance of other state-of-the-art I2I methods. Finally, our derivation of this method helps to understand the impact of self attention, a key component of other recent architecture-specific I2I approaches, in a more architecture-independent way. Project page: https://github.com/jaclyngu/FilteredGuidedDiffusion | ['Abe Davis', 'Zeqi Gu'] | 2023-06-29 | null | null | null | null | ['image-to-image-translation', 'image-to-image-translation'] | ['computer-vision', 'miscellaneous'] | [ 3.37879330e-01 1.64331332e-01 -6.06861338e-03 -1.88156605e-01
-5.83129466e-01 -5.81558585e-01 1.11375034e+00 -9.04397517e-02
-5.14061570e-01 3.97480339e-01 3.24499398e-01 -4.20524627e-01
1.90207869e-01 -7.12671638e-01 -6.95793509e-01 -7.58657157e-01
2.63304889e-01 6.35835528e-01 5.96424401e-01 -4.57445711e-01
1.90418422e-01 6.28357351e-01 -1.01615584e+00 2.32995436e-01
5.75924337e-01 7.73647368e-01 3.57432365e-01 9.10602868e-01
-1.34932652e-01 5.42013586e-01 -4.89113539e-01 -2.94037133e-01
2.54737347e-01 -6.79955423e-01 -6.91200078e-01 -6.98221251e-02
6.42231405e-01 -3.45998555e-01 -3.87615144e-01 8.19108784e-01
7.57831991e-01 4.53696921e-02 7.66172111e-01 -7.82835245e-01
-1.08189845e+00 4.80269998e-01 -6.15443170e-01 5.31235039e-01
-1.38005659e-01 4.30720627e-01 7.27620721e-01 -9.41146910e-01
8.15432310e-01 1.15955031e+00 3.13064605e-01 8.48373055e-01
-1.57074904e+00 -6.23742104e-01 4.04109955e-01 -7.17491731e-02
-1.04458427e+00 -8.05947423e-01 6.66386366e-01 -4.39570904e-01
9.81708288e-01 1.28456228e-03 7.97707617e-01 1.13826644e+00
7.68820122e-02 8.47706795e-01 9.80421901e-01 -5.09094596e-01
7.82570690e-02 1.89292774e-01 -1.39634609e-01 7.77217507e-01
-6.80465102e-02 1.49699345e-01 -3.18938911e-01 6.70735016e-02
1.06033123e+00 -1.95504561e-01 -2.96126008e-01 -4.84404922e-01
-1.16567171e+00 8.42038333e-01 5.46538651e-01 3.81157547e-01
-1.26644894e-01 5.14338434e-01 1.36901915e-01 4.30599660e-01
6.25410199e-01 4.06824589e-01 -2.95861363e-01 -1.15771711e-01
-1.16708982e+00 2.04146042e-01 6.55331731e-01 7.71537006e-01
6.02064252e-01 2.67573774e-01 -3.90821576e-01 6.46395266e-01
2.48925224e-01 4.66304749e-01 4.33902651e-01 -1.21825957e+00
2.10207999e-01 2.81329393e-01 -9.52088311e-02 -7.60805845e-01
-2.86205739e-01 -6.39328003e-01 -8.25990438e-01 5.71278989e-01
5.01487494e-01 -1.51019767e-01 -1.08120823e+00 1.86585772e+00
2.69298196e-01 3.62399146e-02 -2.96523064e-01 7.22835362e-01
4.58219707e-01 4.49247330e-01 -2.38572106e-01 1.67740613e-01
1.03244209e+00 -1.22482109e+00 -3.52549523e-01 -2.87774354e-01
3.57960612e-01 -9.53769505e-01 1.25359774e+00 2.79646188e-01
-1.45818317e+00 -3.78342599e-01 -1.01291680e+00 -1.91357493e-01
-1.99504286e-01 -4.21211049e-02 6.57019854e-01 6.43845022e-01
-1.63547361e+00 7.00072348e-01 -9.60088313e-01 -3.73950928e-01
5.81553698e-01 4.50046033e-01 -1.72125362e-02 -7.55324215e-02
-8.72800827e-01 8.40045571e-01 -8.71279910e-02 -1.16488226e-01
-9.03918564e-01 -7.44031847e-01 -5.96415043e-01 -8.28350633e-02
4.31397408e-01 -1.18425977e+00 1.34103775e+00 -1.16306734e+00
-1.85275483e+00 7.40296960e-01 -2.26750255e-01 -5.82692266e-01
6.56879425e-01 -1.62620634e-01 8.50357711e-02 1.04075521e-01
-2.28398755e-01 1.06582451e+00 1.12307084e+00 -9.33506727e-01
-2.32040852e-01 -2.09443375e-01 2.57123321e-01 2.38404095e-01
-3.78500938e-01 -7.72185326e-02 -9.96406615e-01 -7.69380152e-01
-1.66312143e-01 -1.23030043e+00 -2.82113373e-01 3.83457392e-01
-1.76780969e-01 5.22567406e-02 6.50837600e-01 -3.87360066e-01
1.28325617e+00 -2.08127069e+00 4.17546779e-01 1.36687905e-01
4.12241697e-01 5.01396298e-01 -4.55625296e-01 4.28433090e-01
2.11875230e-01 3.04446906e-01 -3.57866466e-01 -5.71608722e-01
-9.99880284e-02 -2.76274383e-02 -1.44840300e-01 2.93514609e-01
3.15523058e-01 1.06371546e+00 -7.97660172e-01 -2.08353549e-01
1.58113524e-01 7.64897168e-01 -9.82956886e-01 -9.73580405e-02
-2.93996871e-01 5.56905448e-01 -1.91746667e-01 -4.65054577e-03
2.36524835e-01 -3.83701205e-01 2.99493759e-03 -1.98594213e-01
-2.74797808e-02 4.52309251e-01 -9.14082468e-01 1.98410046e+00
-5.92435539e-01 8.33220780e-01 1.43042117e-01 -8.15379143e-01
6.29402697e-01 1.38500288e-01 2.09019244e-01 -6.62879646e-01
3.67603824e-02 1.75163761e-01 2.77785718e-01 7.26631507e-02
2.97422767e-01 1.05749354e-01 3.41783315e-01 7.54972994e-01
2.07557350e-01 -3.88901442e-01 3.39861035e-01 6.44783378e-01
1.03406298e+00 3.00466537e-01 -2.96582468e-03 -2.67448843e-01
3.05080593e-01 -3.37500453e-01 2.33748347e-01 8.85702670e-01
6.79047406e-02 9.03787553e-01 3.33054096e-01 -1.95060641e-01
-1.37133372e+00 -9.23574328e-01 1.12543650e-01 9.22657728e-01
-1.72634438e-01 -3.90686274e-01 -9.21441257e-01 -6.93910360e-01
-2.24291831e-01 5.81731796e-01 -7.88428128e-01 -1.62660927e-01
-6.08063221e-01 -6.98881090e-01 5.30814409e-01 6.44422412e-01
3.71866882e-01 -9.39732492e-01 -2.09714726e-01 3.00076634e-01
2.00106293e-01 -7.75281429e-01 -9.05824900e-01 8.41739699e-02
-9.48323250e-01 -7.32489109e-01 -9.75234687e-01 -5.19372642e-01
6.66099548e-01 3.26544076e-01 1.15378320e+00 1.72721535e-01
-2.67677139e-02 5.26527762e-01 2.47783326e-02 -3.82170647e-01
-6.29398942e-01 5.35267770e-01 -2.02110693e-01 -1.81780770e-01
2.08346937e-02 -7.41785586e-01 -1.03464985e+00 3.76700073e-01
-1.00141132e+00 2.22651631e-01 5.80236912e-01 8.54883611e-01
5.37538886e-01 -4.18555975e-01 4.54968840e-01 -1.02554643e+00
7.71154106e-01 -1.62130445e-01 -5.20403147e-01 -1.44818023e-01
-8.82957935e-01 3.96610081e-01 4.01028126e-01 -5.41336060e-01
-9.69506681e-01 -1.24622829e-01 -3.00566494e-01 -3.32846135e-01
1.64425105e-01 3.10299873e-01 1.07923970e-02 -2.61034936e-01
7.88574457e-01 1.27841771e-01 2.87275910e-01 -5.17998397e-01
8.52764964e-01 2.83412308e-01 3.20909083e-01 -4.55062628e-01
8.37848306e-01 6.22356415e-01 -1.96037233e-01 -6.56251967e-01
-6.49480343e-01 -5.07186018e-02 -6.19386077e-01 1.52174607e-01
6.41738832e-01 -7.42233515e-01 -3.45453918e-01 6.33405328e-01
-1.00100708e+00 -1.02567971e+00 -4.88288641e-01 1.73352137e-01
-5.46811879e-01 1.78803563e-01 -7.84002483e-01 -2.98962682e-01
-4.13358003e-01 -1.27017915e+00 8.09100032e-01 1.46448404e-01
-3.35585445e-01 -1.30199170e+00 2.45069936e-02 2.04183936e-01
9.98897970e-01 -2.55688876e-01 7.61602640e-01 -5.36746323e-01
-8.54503393e-01 2.40082853e-02 -1.89523160e-01 5.01726806e-01
-3.08938138e-02 1.71932369e-01 -7.94961274e-01 -4.16124940e-01
-5.01580425e-02 -1.89840987e-01 1.13308072e+00 4.86359417e-01
9.26386118e-01 -3.38746399e-01 -2.42535368e-01 7.64414549e-01
1.11132574e+00 -7.76486471e-02 6.91719115e-01 4.10207301e-01
7.80790269e-01 2.47626126e-01 2.39475612e-02 3.05460058e-02
4.93313223e-01 9.48858500e-01 1.27330452e-01 -4.58767295e-01
-8.26987028e-01 -8.71141404e-02 4.05534774e-01 7.02318847e-01
-8.77057016e-02 -2.59208798e-01 -8.37071359e-01 5.89149892e-01
-1.63275266e+00 -1.02747095e+00 1.05201699e-01 2.10938954e+00
1.00276029e+00 3.35941732e-01 9.49730426e-02 -1.50031775e-01
3.40509146e-01 3.27948451e-01 -7.54636705e-01 -4.93362606e-01
-7.94429928e-02 3.14137012e-01 5.10099649e-01 8.30436349e-01
-8.97091329e-01 1.12771833e+00 6.80145884e+00 9.68051851e-01
-1.24935770e+00 2.95659155e-01 6.80504560e-01 -5.23971558e-01
-6.96970463e-01 -2.16234699e-02 -7.96731830e-01 2.43022606e-01
9.72222090e-01 -5.51284775e-02 7.03726411e-01 4.75191206e-01
2.37772390e-01 -1.16055652e-01 -1.15256763e+00 6.37132287e-01
9.23477039e-02 -1.62702429e+00 4.96005937e-02 2.63742149e-01
7.73195028e-01 4.32173699e-01 4.42292422e-01 9.61453989e-02
5.88935375e-01 -9.02736783e-01 7.20469773e-01 5.05797923e-01
8.26315761e-01 -4.43021417e-01 2.66578734e-01 2.91634470e-01
-8.36143494e-01 7.58801922e-02 -3.01401943e-01 1.35327235e-01
1.88902259e-01 5.26958585e-01 -6.81118190e-01 1.14569418e-01
4.12738174e-01 4.76047337e-01 -5.48687339e-01 8.43510926e-01
-4.34946597e-01 8.96411896e-01 -5.23999512e-01 1.66776642e-01
3.77776384e-01 -1.39205083e-01 6.75796092e-01 1.37269521e+00
2.28276089e-01 -1.57923788e-01 -1.62557885e-02 8.75357687e-01
-2.24998891e-01 4.31829393e-02 -6.65757060e-01 2.15566196e-02
2.59371221e-01 1.04383385e+00 -5.57852626e-01 -4.45219934e-01
-5.20974874e-01 1.28789580e+00 5.13309836e-01 4.50917095e-01
-7.58590698e-01 -1.40307009e-01 6.68827236e-01 4.78606403e-01
6.03949070e-01 -4.49071318e-01 -2.00748205e-01 -1.14615500e+00
-1.91985205e-01 -9.90282655e-01 7.16104209e-02 -6.74362063e-01
-1.04542816e+00 6.15857422e-01 -1.52936578e-01 -6.78232968e-01
-3.74059856e-01 -2.77061105e-01 -5.71373463e-01 1.07414293e+00
-1.53817499e+00 -1.17420983e+00 1.73046160e-02 6.57654107e-01
5.96789181e-01 -8.17600787e-02 7.56712794e-01 2.78846443e-01
-5.16519785e-01 7.93406785e-01 1.66650161e-01 6.24053217e-02
9.85530913e-01 -1.14929926e+00 9.06583726e-01 9.37544703e-01
3.73630881e-01 8.06127131e-01 5.68885386e-01 -4.72848624e-01
-1.20387709e+00 -9.43978429e-01 5.70134461e-01 -7.83534169e-01
7.61093676e-01 -4.07457590e-01 -8.44040811e-01 6.65693641e-01
4.68269914e-01 -1.36071533e-01 4.69739079e-01 2.07655266e-01
-2.26946801e-01 -1.66164964e-01 -7.12988257e-01 7.90017784e-01
1.15180862e+00 -4.17853445e-01 -1.96837559e-01 1.36877239e-01
7.04920173e-01 -5.14850318e-01 -5.75842202e-01 1.14748161e-02
5.49677789e-01 -9.69405293e-01 1.01695395e+00 -4.07266885e-01
3.33171278e-01 -2.41054192e-01 1.23288520e-01 -1.41261041e+00
-6.68754935e-01 -9.28861558e-01 -2.63311446e-01 1.17641962e+00
8.11362386e-01 -7.08042741e-01 7.08945394e-01 7.26302028e-01
-1.99894428e-01 -8.08300555e-01 -5.70656121e-01 -6.44244611e-01
3.98816437e-01 -4.17600602e-01 3.93226326e-01 6.25467002e-01
-4.54552948e-01 6.03430271e-01 -3.18158031e-01 -1.68196082e-01
4.52343702e-01 -1.36580706e-01 9.56431448e-01 -8.95400643e-01
-7.27716148e-01 -8.63570690e-01 -2.15891182e-01 -1.55109191e+00
-2.49072388e-01 -1.06182134e+00 -3.60187083e-01 -1.74024701e+00
1.83237847e-02 -5.39799273e-01 -2.82555521e-01 5.27875304e-01
-1.92420661e-01 6.15357995e-01 5.25089622e-01 5.60048163e-01
-2.65372992e-01 5.79352856e-01 1.50687957e+00 -2.40366936e-01
-3.28196883e-01 1.81430094e-02 -9.23975408e-01 7.66440809e-01
8.95938635e-01 -4.61281687e-01 -7.61555493e-01 -8.51020098e-01
1.85205936e-01 -3.62120599e-01 2.28319049e-01 -8.56122851e-01
3.20822477e-01 1.05163574e-01 2.30434954e-01 -6.20668940e-02
2.41554677e-01 -4.01640534e-01 2.39925887e-02 3.22468996e-01
-3.23659152e-01 1.23911738e-01 3.49772274e-01 5.63198507e-01
7.86985159e-02 -1.55638576e-01 8.86367202e-01 -1.85479060e-01
-4.48765904e-01 4.83846784e-01 -4.63356048e-01 1.50325224e-01
6.78864241e-01 -1.98228449e-01 -2.38679439e-01 -4.95611817e-01
-7.20907092e-01 -8.17326605e-02 8.05763543e-01 2.82844901e-01
3.23411077e-01 -1.08245087e+00 -7.25623310e-01 1.47886455e-01
-2.59050608e-01 -2.82126684e-02 2.72385702e-02 1.05895531e+00
-2.91042626e-01 2.29093850e-01 -5.56388795e-02 -5.08282185e-01
-9.72073853e-01 5.59943378e-01 3.54169190e-01 -4.32346344e-01
-8.08497965e-01 9.87339735e-01 5.34389555e-01 -9.51809436e-02
3.07089910e-02 -1.72386080e-01 9.18168128e-02 -7.35678971e-02
6.05571568e-01 -8.02127570e-02 2.79790610e-02 -2.66932547e-01
-2.36135125e-01 4.89425987e-01 -4.16401774e-01 -6.79685473e-01
1.36538815e+00 -2.96890438e-01 1.88775286e-01 3.56053650e-01
8.91005576e-01 1.61336392e-01 -1.61358559e+00 -3.63352716e-01
-4.60235000e-01 -2.39411771e-01 3.14553350e-01 -1.01409769e+00
-1.35467803e+00 1.15506744e+00 5.01326859e-01 -3.30200121e-02
1.05188203e+00 -7.49055222e-02 6.56521261e-01 1.09419130e-01
4.65122387e-02 -9.68066275e-01 1.85666770e-01 4.28296894e-01
9.86416698e-01 -1.02891719e+00 -2.25905124e-02 -6.92612380e-02
-5.92924654e-01 8.50949705e-01 3.82342607e-01 -1.98357627e-01
6.23759806e-01 5.42302370e-01 2.04851046e-01 -1.35681376e-01
-9.18492198e-01 -1.77106813e-01 4.73608792e-01 7.60928571e-01
6.23019755e-01 -3.51402193e-01 -2.96155810e-01 -2.25894582e-02
6.28790334e-02 3.40865105e-02 4.14676666e-01 7.57638454e-01
-3.57076973e-01 -1.45802569e+00 4.24313582e-02 2.85057247e-01
-3.79771054e-01 -6.40217006e-01 -3.48809153e-01 6.95840776e-01
-2.97801912e-01 6.27950132e-01 -2.49745604e-02 -2.57429741e-02
1.96034044e-01 9.64512751e-02 7.83046186e-01 -6.63905680e-01
-6.60677671e-01 3.12376440e-01 -3.12834186e-03 -6.67143762e-01
-3.59147191e-01 -6.03691518e-01 -9.84279811e-01 -4.61669534e-01
-3.47617209e-01 -2.21605450e-01 7.05419898e-01 7.13575065e-01
8.82721961e-01 5.93374550e-01 1.98098272e-01 -1.07010341e+00
-3.56960028e-01 -8.36083651e-01 -9.12140533e-02 3.31110865e-01
2.51704335e-01 -4.37972814e-01 -1.02721371e-01 5.42283244e-02] | [11.350852966308594, -0.22509580850601196] |
a8302d7a-b8c9-4ee2-a4cf-e8d85717b6d0 | co-learning-with-pre-trained-networks | 2212.07585 | null | https://arxiv.org/abs/2212.07585v1 | https://arxiv.org/pdf/2212.07585v1.pdf | Co-Learning with Pre-Trained Networks Improves Source-Free Domain Adaptation | Source-free domain adaptation aims to adapt a source model trained on fully-labeled source domain data to a target domain with unlabeled target domain data. Source data is assumed inaccessible due to proprietary or privacy reasons. Existing works use the source model to pseudolabel target data, but the pseudolabels are unreliable due to data distribution shift between source and target domain. In this work, we propose to leverage an ImageNet pre-trained feature extractor in a new co-learning framework to improve target pseudolabel quality for finetuning the source model. Benefits of the ImageNet feature extractor include that it is not source-biased and it provides an alternate view of features and classification decisions different from the source model. Such pre-trained feature extractors are also publicly available, which allows us to readily leverage modern network architectures that have strong representation learning ability. After co-learning, we sharpen predictions of non-pseudolabeled samples by entropy minimization. Evaluation on 3 benchmark datasets show that our proposed method can outperform existing source-free domain adaptation methods, as well as unsupervised domain adaptation methods which assume joint access to source and target data. | ['Chuan-Sheng Foo', 'Li Shen', 'Wenyu Zhang'] | 2022-12-15 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 5.11152446e-01 2.49827534e-01 -8.25363338e-01 -7.71528661e-01
-1.02634597e+00 -9.49918568e-01 7.51746595e-01 2.00113486e-02
-4.82352257e-01 1.13645136e+00 1.93378270e-01 -4.90261912e-02
2.80183941e-01 -7.88940549e-01 -8.72347891e-01 -5.52779973e-01
4.97163922e-01 5.75398684e-01 3.78933288e-02 -6.74329624e-02
1.49901379e-02 2.16843471e-01 -1.02069783e+00 3.62771481e-01
9.80503619e-01 1.18724835e+00 -8.52965638e-02 1.75861046e-01
-2.16421515e-01 5.71341932e-01 -5.32354653e-01 -4.54061896e-01
8.15266848e-01 -5.17652512e-01 -7.45412946e-01 9.71659049e-02
4.68841821e-01 -4.04073268e-01 -1.94264323e-01 1.16822457e+00
2.42972568e-01 -5.52388392e-02 1.04896832e+00 -1.37877691e+00
-1.22911203e+00 5.33337235e-01 -4.34740424e-01 -1.23400480e-01
-9.33097862e-03 2.17803091e-01 8.37525785e-01 -8.30968142e-01
6.92634225e-01 1.02221215e+00 4.31859076e-01 8.58135939e-01
-1.38832283e+00 -1.25655138e+00 9.88508016e-02 1.96964085e-01
-1.42188585e+00 -5.77873170e-01 8.11474562e-01 -3.61513287e-01
4.61799562e-01 -3.10378850e-01 1.49075657e-01 1.76453173e+00
-7.65094012e-02 6.54820979e-01 1.35001099e+00 -5.61201632e-01
5.38937628e-01 8.27407956e-01 -2.57718973e-02 2.46681690e-01
2.94369161e-01 4.90807533e-01 -4.27350372e-01 -3.72905314e-01
7.07495749e-01 2.36622110e-01 -1.98179364e-01 -1.09909391e+00
-1.19865203e+00 1.10487628e+00 6.01847470e-01 -1.02607727e-01
-1.08956233e-01 -3.98082584e-01 3.70182782e-01 5.87852716e-01
4.54654902e-01 5.15665770e-01 -9.30520892e-01 3.82097065e-01
-6.33818209e-01 1.53822300e-04 8.06548297e-01 1.37868440e+00
1.07616401e+00 -9.88172963e-02 -5.78916930e-02 1.00912797e+00
2.20841929e-01 8.09400380e-01 9.74408805e-01 -9.10182655e-01
5.67310810e-01 6.70564950e-01 1.54104367e-01 -4.26298499e-01
1.14019141e-02 -3.45590800e-01 -8.51900280e-01 2.78123677e-01
6.27135456e-01 -2.87845492e-01 -1.21298921e+00 1.86531007e+00
3.13804507e-01 1.72811598e-01 5.82165182e-01 6.19810283e-01
7.00505376e-01 3.29432905e-01 2.09658071e-01 7.60681704e-02
1.04852402e+00 -9.63010848e-01 -3.94204140e-01 -4.75962698e-01
4.23903674e-01 -2.79290140e-01 1.06860268e+00 1.81778580e-01
-4.63051200e-01 -5.33389449e-01 -1.28014576e+00 1.69078112e-01
-6.82264745e-01 -4.47667316e-02 3.85483831e-01 7.64309704e-01
-6.76928759e-01 4.48177010e-01 -5.30236304e-01 -4.54202294e-01
1.01827013e+00 3.45164120e-01 -7.39705265e-01 -3.08887362e-01
-1.39316893e+00 8.86127591e-01 8.02303076e-01 -7.32043803e-01
-1.02205086e+00 -8.45012426e-01 -1.02118194e+00 -5.63709140e-02
4.21803057e-01 -5.01603901e-01 1.29800618e+00 -1.55799448e+00
-1.69362307e+00 1.03113532e+00 1.15840055e-01 -6.33006155e-01
3.88244361e-01 1.33037806e-01 -7.91653693e-01 6.29026815e-02
3.08350563e-01 7.75955915e-01 1.26427662e+00 -1.25193012e+00
-7.88899481e-01 -3.16204727e-01 -2.51431048e-01 2.32277900e-01
-7.33596087e-01 -6.06075943e-01 -6.87931031e-02 -4.94535536e-01
-3.47729117e-01 -8.58898818e-01 -2.15088755e-01 9.29530070e-04
-4.07432675e-01 1.74909219e-01 8.00402045e-01 -1.83373258e-01
7.72746921e-01 -2.16305518e+00 -3.83718163e-01 2.76402980e-01
1.81775808e-01 4.93773878e-01 -4.03943866e-01 1.20351963e-01
-2.88577139e-01 9.22657326e-02 -4.09413218e-01 -3.32247950e-02
-1.13219917e-01 2.40770891e-01 -4.71608162e-01 3.91811460e-01
3.10005009e-01 7.65692115e-01 -9.74385381e-01 -4.31671828e-01
1.38782710e-01 1.81437403e-01 -4.70711708e-01 3.22142154e-01
-1.82794824e-01 6.39675915e-01 -4.66707170e-01 6.49115741e-01
9.43352401e-01 -5.12259305e-01 2.11178169e-01 -5.99038191e-02
5.14544010e-01 1.96754977e-01 -9.16315317e-01 1.88952529e+00
-5.15764713e-01 1.67141214e-01 -1.37493536e-01 -1.16193843e+00
1.24686241e+00 2.76547641e-01 3.29624921e-01 -6.67869925e-01
5.16538043e-03 3.45847905e-01 -2.18871981e-01 -6.83326554e-03
1.19536310e-01 -2.52868086e-01 -3.14740419e-01 4.64950889e-01
5.80138564e-01 7.66738057e-02 -2.13766038e-01 2.62414850e-02
1.06426620e+00 1.58217981e-01 6.76469803e-01 -6.39718547e-02
4.32732254e-01 2.17232317e-01 8.13708961e-01 7.50879347e-01
-4.61732566e-01 7.58962393e-01 1.17414571e-01 -3.11999083e-01
-1.09258330e+00 -1.30943012e+00 -2.80248731e-01 1.18129170e+00
9.96630359e-03 2.64921077e-02 -5.83490133e-01 -1.49463034e+00
1.74346328e-01 7.64439106e-01 -6.95796847e-01 -6.41475558e-01
-2.68146336e-01 -3.49830627e-01 5.35111904e-01 6.43454254e-01
6.60051465e-01 -8.60852540e-01 -1.61929235e-01 2.17333570e-01
-1.21290088e-01 -1.00496936e+00 -6.45473897e-01 5.98033547e-01
-8.84368718e-01 -1.02300048e+00 -8.66528511e-01 -7.44304955e-01
7.73398399e-01 5.93639351e-03 1.02043056e+00 -7.25083888e-01
2.25872591e-01 3.06041449e-01 -4.57226038e-01 -5.92282057e-01
-6.54124618e-01 3.77168924e-01 1.26474530e-01 2.13582978e-01
1.08087015e+00 -6.45103395e-01 -4.81776178e-01 3.67972910e-01
-9.75962162e-01 -4.02001917e-01 7.07746089e-01 1.14460397e+00
6.79907858e-01 -2.66294450e-01 9.75385904e-01 -1.38947248e+00
4.43595201e-01 -9.43426609e-01 -5.74990094e-01 2.61017412e-01
-1.09441066e+00 2.53812462e-01 9.18487430e-01 -6.78805292e-01
-1.29290247e+00 4.69792396e-01 2.87680626e-01 -7.17906058e-01
-6.22076809e-01 1.83589026e-01 -5.56712866e-01 1.47141933e-01
1.24813271e+00 6.73208758e-02 1.28643304e-01 -5.13176739e-01
5.91519058e-01 9.05098915e-01 5.80350935e-01 -4.39124525e-01
1.01473105e+00 4.68931019e-01 -3.27012509e-01 -1.33984879e-01
-1.11724973e+00 -4.27568614e-01 -9.66458023e-01 5.69973052e-01
5.40936649e-01 -1.28585625e+00 2.59289276e-02 2.17098251e-01
-7.80552983e-01 -3.47161025e-01 -6.12630486e-01 4.54459727e-01
-6.33466423e-01 8.62488002e-02 -4.52598967e-02 -1.64190605e-01
-1.75427496e-01 -7.32719600e-01 7.30703235e-01 2.54028857e-01
-1.14998609e-01 -1.32589734e+00 7.08374754e-02 9.05851796e-02
4.60765928e-01 1.98973492e-01 7.21943736e-01 -1.66085851e+00
-2.45376393e-01 -2.65175432e-01 -2.77557969e-01 7.54020333e-01
6.39356673e-01 -5.86548924e-01 -1.09249139e+00 -4.01011407e-01
-3.34235206e-02 -6.87519431e-01 8.48217845e-01 1.76548645e-01
1.05010378e+00 -6.28803492e-01 -4.97931719e-01 8.33337665e-01
1.47434235e+00 1.08162835e-01 3.74837637e-01 4.46641207e-01
6.13329351e-01 4.61413175e-01 7.15574503e-01 4.50001866e-01
4.90985125e-01 4.11668062e-01 3.29919249e-01 -1.59879640e-01
-2.15356007e-01 -6.68280005e-01 4.30826485e-01 1.11684613e-01
6.69045866e-01 -9.50510427e-02 -8.80372047e-01 7.90901721e-01
-1.65762234e+00 -7.34083891e-01 5.08758068e-01 2.37232018e+00
1.24446094e+00 -7.84968119e-03 2.00053796e-01 -4.04632926e-01
7.84619987e-01 -2.01467633e-01 -1.27295172e+00 -1.04672171e-01
-4.79909033e-02 3.92887533e-01 1.08766079e+00 1.37885451e-01
-1.31205690e+00 1.01244271e+00 6.08421040e+00 8.20319772e-01
-1.23678541e+00 3.50267828e-01 5.70573270e-01 -1.12443548e-02
-2.43433639e-01 -5.78704588e-02 -9.49357092e-01 4.11370218e-01
1.17648888e+00 -4.80895132e-01 2.75137573e-01 1.42100573e+00
-5.38873851e-01 3.78417850e-01 -1.50542259e+00 1.02918839e+00
7.42913261e-02 -1.20284891e+00 5.48988059e-02 2.62559235e-01
8.30555677e-01 3.16875905e-01 2.53446996e-01 6.49795115e-01
1.01025248e+00 -1.02267408e+00 4.11982328e-01 1.78242773e-01
1.34036756e+00 -7.01312184e-01 6.22103333e-01 3.75600636e-01
-7.72060096e-01 -2.88395286e-01 -6.30856454e-01 1.87904656e-01
-2.22968340e-01 4.35193747e-01 -1.36093438e+00 4.66401130e-01
5.51378369e-01 9.63658094e-01 -6.41665399e-01 7.10189939e-01
-2.31461316e-01 6.68173254e-01 -2.19327882e-01 4.33745861e-01
-1.22220010e-01 6.62928671e-02 2.53035665e-01 1.04909015e+00
3.49431217e-01 -1.17042840e-01 3.51796538e-01 9.93152857e-01
-4.33209151e-01 5.14901318e-02 -1.09053934e+00 -6.03386424e-02
8.06383312e-01 1.17790055e+00 -2.68924683e-01 -4.77336526e-01
-5.48969209e-01 1.10840547e+00 3.02944958e-01 5.12506485e-01
-5.74075520e-01 -4.31806535e-01 8.27035904e-01 4.47653569e-02
3.77929032e-01 2.81968892e-01 -5.45320034e-01 -1.43559432e+00
-3.29185516e-01 -1.03297853e+00 7.07689881e-01 -3.27580452e-01
-1.94447613e+00 6.65496945e-01 9.68985632e-02 -1.67474234e+00
-4.95986551e-01 -6.84784174e-01 -2.86569297e-01 1.05558336e+00
-1.88500905e+00 -1.29899085e+00 -7.12904483e-02 1.12939453e+00
3.49037170e-01 -7.79968977e-01 1.15329957e+00 9.92413834e-02
-1.98779389e-01 1.21296716e+00 4.55111742e-01 4.57231760e-01
1.43106592e+00 -1.26710725e+00 2.29899719e-01 5.86251915e-01
-1.27949249e-02 6.35992646e-01 2.34792292e-01 -5.75914383e-01
-8.32461476e-01 -1.62665331e+00 5.21259189e-01 -8.30399215e-01
4.24360543e-01 -4.21458930e-01 -9.19633746e-01 9.90428925e-01
2.14261919e-01 4.89488274e-01 1.14036834e+00 3.67331714e-03
-9.31243837e-01 -3.74355257e-01 -1.69669902e+00 2.31628820e-01
8.00097466e-01 -4.17685598e-01 -7.58721054e-01 2.97568172e-01
7.07079470e-01 -2.65646636e-01 -8.37705433e-01 3.33111584e-01
2.87115723e-01 -6.01575315e-01 1.06907880e+00 -7.19179869e-01
4.01702762e-01 -1.96232513e-01 -2.70226330e-01 -1.67210567e+00
-3.51919860e-01 -1.60502508e-01 -1.43938497e-01 1.41787434e+00
6.16229177e-01 -9.22455251e-01 8.41141224e-01 8.18622470e-01
2.04504967e-01 -1.36892542e-01 -8.15752387e-01 -1.08090711e+00
5.06788671e-01 -4.75542992e-02 9.25076067e-01 1.31103182e+00
-5.60301431e-02 5.20081282e-01 -3.97745103e-01 1.71390221e-01
8.18528712e-01 1.69310063e-01 8.54316354e-01 -1.39998496e+00
-1.65316954e-01 -5.37220798e-02 -1.73208434e-02 -1.01017475e+00
4.83544886e-01 -1.23108590e+00 -1.03274405e-01 -1.03584576e+00
1.21689171e-01 -6.61251307e-01 -7.91212022e-01 8.67546737e-01
1.25480732e-02 2.18972340e-01 -1.04545310e-01 3.81271333e-01
-5.41619658e-01 6.33637130e-01 1.06309795e+00 -3.04261744e-01
-2.77846813e-01 3.77145782e-02 -1.24652398e+00 6.59716606e-01
8.34024549e-01 -8.10099602e-01 -6.38881087e-01 -2.16317356e-01
-3.84095848e-01 -3.40421021e-01 2.58736074e-01 -9.10163760e-01
-5.47101572e-02 -4.63654220e-01 7.03649759e-01 -8.82001594e-03
1.04677968e-01 -1.21749377e+00 -1.51305199e-01 8.86468068e-02
-6.49167418e-01 -6.76740587e-01 -3.67096229e-03 8.74848902e-01
-2.44976401e-01 -2.52053708e-01 1.14848292e+00 -2.10602269e-01
-9.03303623e-01 5.10380983e-01 1.15247652e-01 1.81232080e-01
1.11324322e+00 -2.48033196e-01 -3.55302602e-01 -4.84611601e-01
-5.50298631e-01 8.95658880e-02 7.78907955e-01 5.71603119e-01
4.49313581e-01 -1.70553827e+00 -6.70951664e-01 4.66992676e-01
7.35608935e-01 -7.09573925e-02 -7.63457194e-02 2.18225524e-01
2.34754652e-01 3.35510314e-01 -5.17590106e-01 -5.47630191e-01
-6.95845306e-01 1.00546849e+00 1.28083453e-01 -2.74395019e-01
-3.00733566e-01 8.25134695e-01 5.10935307e-01 -1.05238652e+00
5.78797534e-02 -4.47458699e-02 -1.01225108e-01 -7.50004053e-02
5.28403401e-01 -4.29763645e-02 -6.09492399e-02 -4.73933846e-01
-3.14722389e-01 1.70072690e-01 -4.02943015e-01 -4.44375984e-02
1.14404297e+00 -3.77498090e-01 4.73586768e-01 2.68973798e-01
1.42936301e+00 -1.67124644e-01 -1.59809971e+00 -8.80493402e-01
1.19019831e-02 -5.17192006e-01 -2.56836236e-01 -1.31372154e+00
-1.01912248e+00 8.19250345e-01 8.90352845e-01 -2.71938443e-01
1.30832791e+00 -6.55297786e-02 6.63451731e-01 4.89891917e-01
4.50411975e-01 -1.24643660e+00 1.43311292e-01 5.47447145e-01
6.69573307e-01 -1.76233375e+00 -2.84592718e-01 -1.31770119e-01
-1.07280743e+00 8.42715323e-01 9.35526133e-01 -1.21312514e-01
8.19224954e-01 -1.52176684e-02 4.24374610e-01 3.28059316e-01
-5.96316755e-01 -9.01470259e-02 2.48107180e-01 1.36245131e+00
8.01590234e-02 -2.88734697e-02 1.13781095e-01 9.86519575e-01
-5.63192880e-03 2.52754629e-01 4.93884593e-01 7.58918524e-01
-1.47608638e-01 -1.72544241e+00 -4.46370721e-01 5.58676362e-01
-3.19399476e-01 -1.08754456e-01 -4.85291332e-01 6.39345109e-01
1.54737562e-01 8.75451505e-01 -1.15893349e-01 -3.30268174e-01
2.86977619e-01 3.74428421e-01 1.04464710e-01 -9.64809060e-01
-2.79697835e-01 -2.56738871e-01 -2.71176755e-01 -4.06317830e-01
-2.56713152e-01 -4.48805004e-01 -1.03543591e+00 1.07438512e-01
-2.15179354e-01 5.84380962e-02 4.78263319e-01 7.28969455e-01
9.02201176e-01 5.47912903e-02 7.70475984e-01 -4.57485825e-01
-1.01105487e+00 -9.23609197e-01 -6.30411506e-01 7.76125491e-01
4.91004944e-01 -7.55573392e-01 -2.33883068e-01 4.21061099e-01] | [10.374053955078125, 3.10821533203125] |
320ea123-ba48-4e33-84d2-98c4d5200aff | exploiting-sentence-level-representations-for | 2106.07316 | null | https://arxiv.org/abs/2106.07316v2 | https://arxiv.org/pdf/2106.07316v2.pdf | Exploiting Sentence-Level Representations for Passage Ranking | Recently, pre-trained contextual models, such as BERT, have shown to perform well in language related tasks. We revisit the design decisions that govern the applicability of these models for the passage re-ranking task in open-domain question answering. We find that common approaches in the literature rely on fine-tuning a pre-trained BERT model and using a single, global representation of the input, discarding useful fine-grained relevance signals in token- or sentence-level representations. We argue that these discarded tokens hold useful information that can be leveraged. In this paper, we explicitly model the sentence-level representations by using Dynamic Memory Networks (DMNs) and conduct empirical evaluation to show improvements in passage re-ranking over fine-tuned vanilla BERT models by memory-enhanced explicit sentence modelling on a diverse set of open-domain QA datasets. We further show that freezing the BERT model and only training the DMN layer still comes close to the original performance, while improving training efficiency drastically. This indicates that the usual fine-tuning step mostly helps to aggregate the inherent information in a single output token, as opposed to adapting the whole model to the new task, and only achieves rather small gains. | ['Avishek Anand', 'Fabian Beringer', 'Jurek Leonhardt'] | 2021-06-14 | null | null | null | null | ['passage-ranking', 'passage-re-ranking'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.95556924e-01 1.98178515e-01 1.08874738e-01 -3.86380792e-01
-1.26069105e+00 -7.29412913e-01 7.25906134e-01 6.27384245e-01
-8.26296210e-01 7.52735376e-01 7.05090582e-01 -4.93249714e-01
-1.70675293e-01 -6.65748119e-01 -7.18976676e-01 -3.74495506e-01
2.03307951e-03 6.45353854e-01 6.82056189e-01 -7.53701091e-01
3.91545236e-01 9.03044548e-03 -1.46087790e+00 9.31818068e-01
1.07516241e+00 6.43219471e-01 4.50795531e-01 9.25636590e-01
-4.04363513e-01 1.06785357e+00 -8.78687799e-01 -4.72324014e-01
-1.35302737e-01 -2.89083660e-01 -1.55347252e+00 -6.65811062e-01
5.85449159e-01 -2.43330121e-01 -2.96050966e-01 6.12586677e-01
5.65067053e-01 6.86726153e-01 4.63214606e-01 -1.88632682e-01
-8.48756969e-01 7.94900358e-01 2.03082059e-02 6.81981027e-01
4.10864949e-01 -3.98242138e-02 1.35375023e+00 -7.13415027e-01
5.69155514e-01 1.24250543e+00 6.15856886e-01 6.70440674e-01
-1.33304501e+00 1.38100192e-01 3.80881429e-01 4.74220544e-01
-9.49703872e-01 -5.81040323e-01 4.49100673e-01 -1.19107135e-01
1.67728043e+00 4.02990669e-01 1.87708408e-01 9.91702020e-01
3.25832158e-01 7.60284543e-01 9.12240863e-01 -8.52900207e-01
1.00226171e-01 -1.23749770e-01 6.73628926e-01 3.28767687e-01
7.70751014e-02 -2.36380965e-01 -4.34845150e-01 -2.79658616e-01
2.08295047e-01 -2.11781010e-01 -1.80824399e-01 -2.34142289e-01
-1.06137347e+00 9.24269021e-01 5.99144399e-01 6.98923707e-01
-2.11871997e-01 1.41457424e-01 7.10438550e-01 7.65948236e-01
5.40023983e-01 8.83580863e-01 -9.24615085e-01 -2.62437403e-01
-1.03532541e+00 3.34284782e-01 7.38451421e-01 5.04708230e-01
5.98198116e-01 -2.77463287e-01 -8.23477685e-01 1.18933797e+00
8.51086527e-02 9.74930599e-02 7.44683623e-01 -1.01284790e+00
6.53260171e-01 3.91643435e-01 2.27178514e-01 -4.48509604e-01
-3.44248801e-01 -5.82596779e-01 -3.91903222e-01 -3.80650789e-01
4.43445265e-01 2.99882088e-02 -9.84642386e-01 1.98620200e+00
-3.23699303e-02 -3.21722955e-01 8.97073224e-02 5.94739497e-01
7.40945220e-01 6.83313608e-01 4.09163415e-01 -1.61832746e-03
1.25940943e+00 -1.27624118e+00 -4.25600767e-01 -4.99288738e-01
1.08183956e+00 -6.63075387e-01 1.52774513e+00 2.04705507e-01
-1.31394219e+00 -5.28054953e-01 -1.02721584e+00 -6.76180065e-01
-4.59478647e-01 -2.26853073e-01 3.17972094e-01 3.72707933e-01
-1.40905285e+00 9.96274352e-01 -6.07034206e-01 -4.70995247e-01
6.49876520e-02 4.62399811e-01 1.08971046e-02 -2.78607041e-01
-1.59761858e+00 1.28690386e+00 3.26680809e-01 1.00885043e-02
-6.62281036e-01 -4.96851563e-01 -7.05573857e-01 3.80682290e-01
2.91366756e-01 -1.17766678e+00 1.65183389e+00 -7.83432841e-01
-1.51760232e+00 7.27524221e-01 -5.77050865e-01 -6.24997377e-01
2.86790244e-02 -4.00012434e-01 -2.60994077e-01 2.29389697e-01
3.74161378e-02 4.44460630e-01 6.02232158e-01 -9.89115119e-01
-3.78824472e-01 -1.03716657e-01 6.66089356e-01 2.05333814e-01
-3.67229223e-01 7.34703541e-02 -7.37164021e-02 -4.63403463e-01
-2.22400770e-01 -5.54348052e-01 -2.14534327e-01 -7.09061801e-01
5.20792454e-02 -5.26439071e-01 1.38256297e-01 -8.10036182e-01
1.49257052e+00 -1.74821556e+00 1.86169177e-01 -1.73203707e-01
-1.76481623e-02 5.69739699e-01 -7.08531618e-01 7.42017269e-01
1.15919873e-01 1.85235068e-01 -2.23478034e-01 -5.29854655e-01
-1.14882179e-02 3.67459774e-01 -5.18533230e-01 -1.55072942e-01
2.56658226e-01 1.26340020e+00 -1.13315141e+00 -3.68284494e-01
-8.50580912e-03 2.11126000e-01 -6.98616683e-01 5.45897633e-02
-5.76795042e-01 2.71294177e-01 -3.41792613e-01 5.46007603e-02
4.09980804e-01 -3.45923394e-01 2.77308911e-01 6.83842748e-02
2.17413023e-01 1.27642894e+00 -6.39092505e-01 2.00944853e+00
-8.69430006e-01 5.62357485e-01 -8.71823505e-02 -9.85268414e-01
5.11675000e-01 2.01134071e-01 -7.90845752e-02 -1.17156601e+00
-1.22117124e-01 2.57912099e-01 8.19858983e-02 -4.48473632e-01
9.53568697e-01 -2.69752532e-01 -1.51811421e-01 4.77039665e-01
4.10550654e-01 -3.62561494e-02 2.82088280e-01 4.14581627e-01
1.24026585e+00 8.76032263e-02 1.57397985e-01 -3.95872712e-01
6.51903629e-01 1.14933781e-01 9.60160270e-02 1.18769109e+00
-7.20937327e-02 5.29202998e-01 2.63564616e-01 -1.27710208e-01
-1.00606811e+00 -1.05750942e+00 -8.01522806e-02 1.78060830e+00
-5.62643297e-02 -5.34717381e-01 -7.06336021e-01 -7.51814723e-01
-2.14906424e-01 1.04591966e+00 -6.57846570e-01 -4.53268081e-01
-1.01619112e+00 -7.34093547e-01 4.86586571e-01 4.43820298e-01
1.11211330e-01 -1.17839134e+00 -3.88994426e-01 4.99550015e-01
-6.11275136e-01 -6.95425034e-01 -1.89377546e-01 4.22974706e-01
-1.16672719e+00 -6.12186849e-01 -6.98628485e-01 -8.64814520e-01
3.28464866e-01 3.47271383e-01 1.78754818e+00 3.95512283e-01
1.14717178e-01 3.46255779e-01 -6.38455570e-01 -8.77842121e-03
-4.47588205e-01 6.14335656e-01 -3.04322481e-01 -5.46600640e-01
3.94072354e-01 -4.57300454e-01 -5.78917205e-01 -7.85474628e-02
-8.56085181e-01 -3.02456826e-01 4.67779905e-01 1.23424780e+00
1.90578327e-01 -6.26868725e-01 1.00691366e+00 -8.68358612e-01
1.06914437e+00 -4.15393800e-01 2.02917904e-02 5.65534651e-01
-4.37245697e-01 4.84898120e-01 7.57757843e-01 -1.84210688e-01
-1.19668913e+00 -7.28459299e-01 -5.60361922e-01 3.24475825e-01
5.10512590e-02 7.05861747e-01 1.09672733e-01 2.01167107e-01
8.35811138e-01 1.13675550e-01 -3.72179061e-01 -7.68538594e-01
7.63302028e-01 6.10455394e-01 2.89543688e-01 -8.04677308e-01
4.00856555e-01 1.43172204e-01 -4.13574219e-01 -4.95244920e-01
-1.42981791e+00 -6.28287375e-01 -7.03498244e-01 3.22295576e-01
5.65279722e-01 -7.59438574e-01 -4.00423735e-01 1.30669272e-03
-1.30754757e+00 -5.54097235e-01 -6.74112201e-01 1.99221298e-01
-4.59639996e-01 5.54858446e-01 -9.41660821e-01 -5.66822171e-01
-5.06631672e-01 -7.27189779e-01 1.03981996e+00 6.68730540e-03
-3.26285601e-01 -1.13428509e+00 5.24833441e-01 5.02641737e-01
8.24773729e-01 -4.89498496e-01 1.35094595e+00 -8.86379719e-01
-4.45565104e-01 -2.42500138e-02 -1.06440902e-01 4.58047450e-01
-2.93717116e-01 -5.10754168e-01 -1.26954484e+00 -2.54488349e-01
2.02037632e-01 -4.92274523e-01 1.55876291e+00 1.88003823e-01
9.77052867e-01 -2.45918706e-01 -1.13536291e-01 1.35769695e-01
1.25630748e+00 -2.58288473e-01 6.54384255e-01 6.28941953e-01
2.90760130e-01 4.83394802e-01 4.35303122e-01 3.00797280e-02
6.30137086e-01 5.38403094e-01 1.99257642e-01 2.61971503e-01
-2.49017492e-01 -3.27901810e-01 4.63544428e-01 1.14957082e+00
-1.13608809e-02 -3.40585500e-01 -6.93079472e-01 9.05379415e-01
-1.87737906e+00 -1.10949230e+00 8.28333125e-02 2.29983878e+00
1.14574289e+00 2.70044595e-01 -9.05353203e-02 -1.11858025e-01
3.52323294e-01 4.16594326e-01 -2.48222888e-01 -9.82687593e-01
-2.34841034e-01 6.87308371e-01 7.20099807e-02 9.08281386e-01
-8.61084640e-01 1.18613887e+00 7.04022503e+00 9.65898931e-01
-8.37431371e-01 5.99241555e-01 4.23249245e-01 -2.34895855e-01
-5.96482515e-01 1.62446231e-01 -5.94853163e-01 1.16460524e-01
1.58134937e+00 -1.08927377e-01 4.63755339e-01 4.82545644e-01
6.52984157e-02 -2.37978160e-01 -1.17382050e+00 3.88659865e-01
6.63249493e-02 -1.34245670e+00 3.62420440e-01 -2.98336089e-01
6.63513184e-01 4.22096342e-01 -2.06059203e-01 9.03679252e-01
4.07171309e-01 -1.01998711e+00 4.91494447e-01 6.75038278e-01
2.69879073e-01 -4.59187299e-01 8.91342461e-01 5.43659866e-01
-7.66849339e-01 -2.93010324e-01 -8.09309721e-01 -4.02848303e-01
1.59134820e-01 3.88689846e-01 -7.44836152e-01 7.15564013e-01
5.22223711e-01 3.36351871e-01 -8.97108674e-01 9.91611660e-01
-1.37058392e-01 8.81933630e-01 -1.58699140e-01 -1.91106990e-01
3.18163395e-01 2.39392102e-01 4.22086358e-01 1.42664051e+00
-9.52942483e-03 -2.45375797e-01 -2.35606506e-01 5.66986740e-01
-1.29277721e-01 1.79064214e-01 -1.82236224e-01 1.51566207e-01
2.40639016e-01 9.61459696e-01 -1.67686626e-01 -6.26638591e-01
-3.97790045e-01 1.04886460e+00 9.21366692e-01 3.96461964e-01
-5.22999704e-01 -3.92693400e-01 4.41128105e-01 3.47533487e-02
5.46995699e-01 -4.12877381e-01 -2.11960182e-01 -1.40035450e+00
1.22510508e-01 -7.06689537e-01 5.76226592e-01 -6.47649944e-01
-1.26940238e+00 8.36993098e-01 -1.36737540e-01 -8.45229089e-01
-5.64383566e-01 -5.37551105e-01 -6.38964474e-01 1.08363473e+00
-1.98124015e+00 -8.82523596e-01 1.88125029e-01 4.34692621e-01
6.29379690e-01 2.60123760e-01 1.10710013e+00 2.47714981e-01
-1.75112262e-01 6.76161766e-01 6.42856359e-01 -2.39715818e-02
8.44284236e-01 -1.33060563e+00 6.27043068e-01 7.11396277e-01
3.68682116e-01 1.15149128e+00 6.23961270e-01 -3.21733534e-01
-1.03316474e+00 -8.28420103e-01 1.63965762e+00 -9.50448215e-01
6.35612607e-01 -4.32204485e-01 -1.11632347e+00 5.96089661e-01
6.01692379e-01 -2.72143394e-01 5.91739655e-01 5.72060764e-01
-3.67638171e-01 2.18503624e-02 -8.71962130e-01 4.81261522e-01
1.03244948e+00 -8.89179111e-01 -1.49140739e+00 3.11623722e-01
1.10624838e+00 -1.18818834e-01 -6.48600757e-01 1.87898666e-01
2.48774245e-01 -8.88948560e-01 1.17668188e+00 -1.13794351e+00
3.47311318e-01 5.32528609e-02 -1.88702554e-01 -1.45346677e+00
-5.55674493e-01 -3.71037245e-01 -3.42261404e-01 1.08698416e+00
5.86795688e-01 -5.00031650e-01 4.31664795e-01 4.98078793e-01
-3.27759176e-01 -7.87380040e-01 -1.11218441e+00 -7.20948279e-01
7.13779926e-01 -3.38060319e-01 4.96652395e-01 5.27542293e-01
2.79843807e-01 8.33676934e-01 -5.03729545e-02 -2.57255971e-01
8.53733346e-02 1.07209161e-01 3.42776269e-01 -1.23317218e+00
-4.87424195e-01 -4.07355607e-01 1.01714320e-01 -1.59036827e+00
9.86304507e-02 -1.12609994e+00 3.35882939e-02 -2.01172829e+00
2.21169829e-01 -3.16766322e-01 -6.62442029e-01 1.87546879e-01
-5.65549016e-01 -3.45057412e-03 1.95460662e-01 1.80311292e-01
-1.08176899e+00 5.36210775e-01 1.23520875e+00 -2.18948424e-01
-1.53119490e-01 -6.02185354e-02 -9.46395457e-01 4.29162562e-01
7.88287759e-01 -4.42255527e-01 -3.91704053e-01 -1.09023964e+00
4.52577442e-01 -1.52841181e-01 3.92176002e-01 -6.92939520e-01
1.78299785e-01 1.16447903e-01 2.71823585e-01 -3.60022992e-01
4.76742119e-01 -3.37930620e-01 -7.61319160e-01 2.94739395e-01
-8.73942971e-01 4.83875833e-02 3.21763515e-01 5.82643569e-01
-3.51863831e-01 -6.29421234e-01 4.65461314e-01 -4.41600502e-01
-7.19007850e-01 -2.32585773e-01 -4.81779575e-01 5.41206241e-01
1.97822288e-01 -5.61680719e-02 -5.92874110e-01 -3.81809562e-01
-9.52423573e-01 2.33901128e-01 1.33648589e-01 5.59583306e-01
1.97701737e-01 -9.29105937e-01 -7.16528475e-01 -1.16893597e-01
9.61637720e-02 -2.11784452e-01 4.68708932e-01 6.87707305e-01
-1.20174453e-01 9.93981123e-01 1.41358063e-01 -3.96632373e-01
-8.48411024e-01 5.92092574e-01 3.98007214e-01 -9.70617831e-01
-4.49390739e-01 1.03953552e+00 2.87815351e-02 -6.72368407e-01
-1.95271466e-02 -6.39731705e-01 -4.65437859e-01 2.00398698e-01
4.39809501e-01 1.74166113e-01 5.93045533e-01 -3.24843913e-01
-3.97760361e-01 3.30884844e-01 -4.31519002e-01 -2.20075756e-01
1.28223205e+00 -4.81925368e-01 -1.16732895e-01 5.30522048e-01
1.16240311e+00 1.13428384e-01 -7.40658045e-01 -5.35708070e-01
3.50703835e-01 -2.19520971e-01 -1.45726234e-01 -1.18502212e+00
-2.87826538e-01 1.23944736e+00 1.59941733e-01 2.20133051e-01
8.50314677e-01 5.19954301e-02 9.74518716e-01 8.66102517e-01
4.38357621e-01 -1.21221972e+00 9.00023952e-02 1.11393464e+00
8.65072608e-01 -9.51791763e-01 -2.74356306e-01 1.61518723e-01
-5.44295728e-01 9.70711887e-01 4.41088706e-01 -2.86391139e-01
3.30241144e-01 -3.22778821e-01 5.78026325e-02 4.15696241e-02
-1.19375646e+00 -2.29674220e-01 3.05506676e-01 3.87915283e-01
6.79679811e-01 -9.90435630e-02 -4.59590882e-01 7.77406037e-01
-2.60875642e-01 -2.19962467e-02 4.07889992e-01 9.98081386e-01
-9.39690471e-01 -1.37201881e+00 -1.03088431e-01 6.43127680e-01
-4.98565495e-01 -7.91642785e-01 -3.10430765e-01 4.84090745e-01
-6.10329211e-02 1.14864218e+00 3.15546729e-02 -1.29894137e-01
3.98355722e-01 5.72378993e-01 6.92023337e-01 -8.62114549e-01
-1.04838514e+00 -3.44775915e-01 4.68614936e-01 -3.82314295e-01
-2.60063320e-01 -4.33140010e-01 -1.02565420e+00 3.57288420e-02
-3.48558754e-01 6.05347693e-01 2.35403955e-01 1.19257486e+00
6.85950160e-01 5.97753644e-01 2.57637262e-01 -5.15240669e-01
-1.19168282e+00 -1.26861119e+00 -1.21228620e-01 3.52576584e-01
6.91451788e-01 -4.43616152e-01 -3.40417325e-01 -2.48746842e-01] | [11.326737403869629, 8.003156661987305] |
15a0236d-586c-4fb1-9c67-e7b9438a83b4 | test-time-adaptation-with-clip-reward-for | 2305.18010 | null | https://arxiv.org/abs/2305.18010v1 | https://arxiv.org/pdf/2305.18010v1.pdf | Test-Time Adaptation with CLIP Reward for Zero-Shot Generalization in Vision-Language Models | Misalignment between the outputs of a vision-language (VL) model and task goal hinders its deployment. This issue can worsen when there are distribution shifts between the training and test data. To address this problem, prevailing fully test-time adaptation~(TTA) methods bootstrap themselves through entropy minimization. However, minimizing the entropy of the predictions makes the model overfit to incorrect output distributions of itself. In this work, we propose TTA with feedback to avoid such overfitting and align the model with task goals. Specifically, we adopt CLIP as reward model to provide feedback for VL models during test time in various tasks, including image classification, image-text retrieval, and image captioning. Given a single test sample, the model aims to maximize CLIP reward through reinforcement learning. We adopt a reward design with the average CLIP score of sampled candidates as the baseline. This design is simple and surprisingly effective when combined with various task-specific sampling strategies. The entire system is flexible, allowing the reward model to be extended with multiple CLIP models. Plus, a momentum buffer can be used to memorize and leverage the learned knowledge from multiple test samples. Extensive experiments demonstrate that our method significantly improves different VL models after TTA. | ['Yi Yang', 'Linchao Zhu', 'Xiaohan Wang', 'Shuai Zhao'] | 2023-05-29 | null | null | null | null | ['image-captioning'] | ['computer-vision'] | [ 2.41763040e-01 -1.55813187e-01 -3.88237357e-01 -5.71873069e-01
-9.52837706e-01 -4.36904043e-01 3.53182375e-01 -4.00905907e-01
-4.42905605e-01 7.47503340e-01 -1.01032436e-01 -2.94716299e-01
2.90587038e-01 -4.05389965e-01 -8.89562666e-01 -6.63075268e-01
3.66925210e-01 4.01333392e-01 1.51465714e-01 3.03226173e-01
3.65930021e-01 6.53462559e-02 -1.28088725e+00 3.66865784e-01
1.17209005e+00 1.15760875e+00 8.61789942e-01 5.64443350e-01
-2.50068486e-01 6.64416432e-01 -7.45181859e-01 -4.21416909e-01
2.85572588e-01 -5.64628541e-01 -4.89562988e-01 2.67828405e-01
5.84251642e-01 -4.18782741e-01 -2.87073225e-01 1.20836270e+00
5.89502752e-01 4.02709506e-02 5.46944082e-01 -1.41144860e+00
-9.36412513e-01 4.71887320e-01 -6.28140509e-01 1.50127066e-02
1.15365259e-01 6.27701998e-01 7.79834151e-01 -1.10393906e+00
3.61027628e-01 1.16681671e+00 4.62911904e-01 7.45191395e-01
-1.22641015e+00 -7.92109430e-01 6.34706438e-01 3.44136119e-01
-1.24015212e+00 -4.44702417e-01 6.58035815e-01 -3.93584967e-01
8.24533999e-01 1.11091010e-01 6.97504818e-01 1.19523132e+00
2.14799121e-01 1.24405301e+00 1.03820312e+00 -3.80536407e-01
3.22630852e-01 4.75375891e-01 -1.69533432e-01 5.37638009e-01
4.49869782e-02 -7.04204366e-02 -5.99205494e-01 -1.22124888e-01
6.37210846e-01 -4.15213294e-02 -3.99576485e-01 -4.79063928e-01
-1.00290465e+00 7.52851605e-01 4.00736719e-01 -1.84644267e-01
-3.01138014e-01 1.91755638e-01 1.82828650e-01 3.14336777e-01
2.49298185e-01 6.07263684e-01 -5.14058173e-01 -1.89814353e-04
-1.02249742e+00 2.76981086e-01 5.06811321e-01 1.04583454e+00
6.84477091e-01 2.25550562e-01 -8.17086756e-01 1.05856955e+00
3.81930768e-01 8.28841448e-01 7.70254016e-01 -9.46620226e-01
6.79269433e-01 3.15851986e-01 1.08548433e-01 -5.47922671e-01
1.57624453e-01 -5.06687880e-01 -5.10979712e-01 3.04405093e-01
3.32139164e-01 -1.83293626e-01 -1.24634135e+00 1.91266835e+00
5.80607317e-02 2.49606341e-01 -2.74818223e-02 9.34383154e-01
3.36477518e-01 5.22511542e-01 2.85597588e-03 -4.99336243e-01
8.25362086e-01 -1.17860734e+00 -6.71211004e-01 -7.37605333e-01
3.23670954e-01 -7.43638337e-01 1.43488550e+00 3.93810570e-01
-1.16379976e+00 -5.88140428e-01 -1.02579701e+00 2.36651793e-01
5.48138395e-02 2.62384385e-01 2.07461253e-01 2.27156997e-01
-1.00486505e+00 4.36903805e-01 -6.95707619e-01 7.29111815e-03
5.70516586e-01 3.77429008e-01 1.66872159e-01 -2.30011761e-01
-8.36032569e-01 7.65764773e-01 3.33660513e-01 5.47149181e-02
-8.63100588e-01 -4.19339329e-01 -6.79441631e-01 4.23935615e-02
4.31255579e-01 -6.63904488e-01 1.62221575e+00 -1.20997500e+00
-1.57769465e+00 4.42671627e-01 -3.50420743e-01 -5.53789556e-01
6.50794148e-01 -3.90461028e-01 -4.75702062e-02 -1.94809169e-01
7.43964314e-02 1.06821227e+00 1.24573493e+00 -1.17094338e+00
-5.13053477e-01 -4.60782200e-02 -2.72989362e-01 4.80872482e-01
-3.78089249e-01 -2.61084586e-01 -7.30063379e-01 -5.05104303e-01
-1.83750853e-01 -8.89589071e-01 -2.42208302e-01 6.90700039e-02
-4.15724963e-01 -1.28338456e-01 6.00618184e-01 -3.81980002e-01
1.49114788e+00 -2.17856526e+00 -4.77607213e-02 1.25777379e-01
-5.48273847e-02 2.51584738e-01 -4.17510748e-01 -4.76417044e-04
2.06803694e-01 1.11186504e-01 -2.09796935e-01 -3.35362315e-01
-1.40858188e-01 1.96765989e-01 -5.26586771e-01 3.62006463e-02
4.73212212e-01 1.04728866e+00 -8.12389970e-01 -5.47200680e-01
-6.40772050e-03 3.99675295e-02 -7.84207225e-01 5.65298855e-01
-6.66615963e-01 3.26508611e-01 -6.46193922e-01 7.05259144e-01
6.74491525e-01 -4.90280539e-01 6.99509531e-02 1.57074973e-01
1.26483247e-01 7.80026540e-02 -7.86946893e-01 1.45999622e+00
-2.58052766e-01 6.81222975e-01 -1.93572596e-01 -7.95829892e-01
1.02512503e+00 1.23789432e-02 1.75362036e-01 -7.32893169e-01
-1.38750330e-01 1.79884493e-01 1.56641021e-01 -7.53817379e-01
3.34426165e-01 1.53037101e-01 1.41921967e-01 1.52038693e-01
3.87641648e-03 -2.33779207e-01 -1.69909686e-01 -1.28292382e-01
7.94390142e-01 2.63352603e-01 2.30842128e-01 1.86013252e-01
2.74944544e-01 -5.91006055e-02 7.16162503e-01 1.09088647e+00
-3.72391731e-01 7.59121776e-01 3.30686420e-01 -2.02495456e-01
-1.08461285e+00 -1.06805491e+00 -1.15275450e-01 1.20224559e+00
8.05046782e-02 3.72211598e-02 -6.70613706e-01 -9.59998965e-01
1.58803940e-01 9.48048055e-01 -4.10562724e-01 -4.27581906e-01
-5.94389319e-01 -5.81780493e-01 1.73322305e-01 5.46215355e-01
5.69671988e-01 -1.24948919e+00 -4.60231692e-01 1.86007679e-01
-2.46491939e-01 -8.04064095e-01 -9.11047041e-01 2.56189167e-01
-8.41712892e-01 -8.37502897e-01 -8.24485302e-01 -8.86301339e-01
7.78817475e-01 3.17376137e-01 1.02717197e+00 -1.47834659e-01
2.42714528e-02 2.71509677e-01 -1.96655646e-01 -6.51170194e-01
-2.54736304e-01 1.25286840e-02 1.99434590e-02 -5.43569736e-02
3.78259540e-01 -1.81570694e-01 -7.59377182e-01 3.68623108e-01
-7.31622756e-01 9.79410261e-02 7.13984311e-01 1.17958188e+00
7.31360555e-01 -5.21945477e-01 7.76930273e-01 -6.00290239e-01
8.24270308e-01 -5.35671949e-01 -6.58389628e-01 6.30389333e-01
-9.88270879e-01 2.09633887e-01 6.70334399e-01 -9.18745577e-01
-8.95367026e-01 5.35736308e-02 1.34520248e-01 -9.77394819e-01
2.02048510e-01 4.54740018e-01 1.38637153e-02 -1.78591646e-02
5.44831991e-01 4.46723670e-01 1.84419408e-01 -2.17841089e-01
1.31681249e-01 7.32310414e-01 4.01695579e-01 -4.75329250e-01
6.06814325e-01 -1.45160019e-01 -5.36186218e-01 -3.60221088e-01
-1.01326513e+00 -2.18201250e-01 -1.37182206e-01 -3.51206213e-01
5.90782166e-01 -8.84236276e-01 -3.29918772e-01 3.46230567e-01
-9.56481814e-01 -6.63130701e-01 -2.74081141e-01 5.90917110e-01
-7.00469136e-01 1.18155770e-01 -1.53255895e-01 -9.27867174e-01
-2.98529685e-01 -1.46171808e+00 9.00752604e-01 6.37642145e-01
7.91123733e-02 -6.92451239e-01 -9.86652821e-02 9.92111266e-02
5.27334392e-01 -3.01278383e-01 7.02701926e-01 -6.56751394e-01
-7.93751359e-01 -5.49082235e-02 -1.17109373e-01 6.34055972e-01
-1.68728620e-01 -9.51905698e-02 -1.01509988e+00 -5.56382656e-01
6.03130199e-02 -8.56917739e-01 9.25341427e-01 7.27257490e-01
1.65676928e+00 -4.64838773e-01 -2.01313064e-01 6.55515850e-01
1.27640939e+00 3.69534284e-01 5.37647605e-01 4.27869499e-01
4.85224873e-01 2.80166835e-01 9.63653088e-01 4.68611389e-01
1.75109833e-01 6.68325424e-01 5.18131435e-01 2.28085354e-01
-5.94727546e-02 -4.15120572e-01 8.20538819e-01 7.71783173e-01
3.48913044e-01 -3.79846662e-01 -7.89655864e-01 3.69381875e-01
-1.97052610e+00 -8.92563760e-01 4.69595820e-01 2.50717354e+00
1.08051348e+00 3.06464970e-01 -1.41312525e-01 -6.42631292e-01
5.46330631e-01 1.12154156e-01 -1.25641143e+00 -2.41916329e-01
5.07432409e-02 -1.46247193e-01 4.44498301e-01 4.41522241e-01
-9.04001772e-01 9.89418089e-01 7.00371170e+00 9.04115319e-01
-1.30045533e+00 -1.23354211e-01 7.95620978e-01 -4.40887958e-01
-4.66454148e-01 -1.68590292e-01 -9.31786478e-01 5.78302205e-01
7.74999678e-01 -4.19985354e-01 6.82496488e-01 1.03026843e+00
1.69514403e-01 -6.03152774e-02 -1.20249701e+00 1.02980721e+00
1.23341419e-01 -1.00367069e+00 -1.18848551e-02 -3.60788256e-02
7.19766796e-01 4.17979300e-01 4.91904408e-01 7.82761216e-01
3.40221256e-01 -8.81252766e-01 8.51366222e-01 6.85368657e-01
9.43489134e-01 -4.52513754e-01 5.34734428e-01 4.98686939e-01
-7.60487199e-01 -3.56193483e-01 -6.23997569e-01 2.50717133e-01
-9.60851833e-02 2.68534005e-01 -1.24264526e+00 7.17215315e-02
5.65647483e-01 4.37322527e-01 -6.71303689e-01 1.43222320e+00
-1.21193670e-01 7.29148865e-01 -6.43571317e-02 -2.61488259e-01
2.44747251e-01 -1.76300257e-01 5.82690358e-01 1.12439775e+00
4.90829587e-01 -4.64307934e-01 6.39317930e-01 9.23948765e-01
-1.68929905e-01 5.90308495e-02 -7.17477143e-01 1.18133001e-01
7.51384974e-01 8.43958735e-01 -2.07640037e-01 -2.75726736e-01
-3.46589297e-01 9.49618161e-01 5.94181240e-01 7.43677318e-01
-9.64097559e-01 -5.64002320e-02 5.43888927e-01 -1.24890551e-01
4.26781893e-01 3.31724547e-02 -3.17458153e-01 -1.16754496e+00
2.27955908e-01 -7.93793499e-01 1.26688555e-01 -9.29405153e-01
-1.46396661e+00 5.63312173e-01 7.05349967e-02 -1.56689942e+00
-5.26837170e-01 -4.30376559e-01 -6.16500795e-01 9.75642800e-01
-1.71948314e+00 -7.03237414e-01 -1.04809299e-01 4.41334218e-01
1.06369925e+00 -4.11718786e-01 5.03268719e-01 -1.11373723e-01
-6.72677219e-01 1.01894832e+00 2.00438991e-01 -1.08594112e-01
1.08028543e+00 -1.29010892e+00 1.16638452e-01 6.67640030e-01
5.40980585e-02 3.98407072e-01 6.12999439e-01 -6.36896312e-01
-1.14055264e+00 -1.18970394e+00 5.08922696e-01 -2.76330143e-01
4.62189019e-01 -2.11178929e-01 -1.16155827e+00 5.58036327e-01
3.10658544e-01 1.51154667e-01 5.08337915e-01 -8.55325609e-02
-2.20636547e-01 -2.68705040e-01 -9.19346988e-01 6.48080826e-01
7.31494963e-01 -5.37818134e-01 -4.69973445e-01 3.94771159e-01
8.87994945e-01 -5.91549158e-01 -4.42666382e-01 3.83750796e-01
5.15720308e-01 -6.91916406e-01 7.33555615e-01 -5.33384264e-01
3.34943026e-01 -2.75876135e-01 -7.80327022e-02 -1.42533064e+00
-4.08300161e-01 -3.97022992e-01 -3.25604618e-01 1.02393830e+00
6.22651815e-01 -4.77163821e-01 6.81963086e-01 7.45906353e-01
-2.11392954e-01 -1.03302050e+00 -8.50390792e-01 -8.09937656e-01
-2.97864582e-02 -3.85773331e-01 3.75662267e-01 6.19950354e-01
-1.00009657e-01 2.20877826e-01 -4.44895089e-01 1.33973211e-01
5.29348016e-01 1.37646198e-01 6.13313735e-01 -8.65383625e-01
-5.46150982e-01 -5.51268518e-01 1.15264349e-01 -1.42236280e+00
6.04343899e-02 -8.56246889e-01 4.63298112e-01 -1.15392399e+00
4.47123557e-01 -4.63435262e-01 -5.75740457e-01 5.69011211e-01
-5.11844337e-01 -1.93905041e-01 4.96853083e-01 5.43698072e-01
-8.32264721e-01 7.79231429e-01 1.53753495e+00 -2.13006586e-01
-4.22971904e-01 3.15677762e-01 -6.43314540e-01 6.51755810e-01
9.23587680e-01 -4.87160146e-01 -6.86325252e-01 -6.80500388e-01
-3.97996753e-02 1.30658954e-01 2.12735236e-01 -9.20383573e-01
3.37489069e-01 -4.90441054e-01 6.08746588e-01 -5.82140207e-01
1.06343329e-01 -6.69477522e-01 -2.82561600e-01 3.04046124e-01
-7.09664226e-01 7.25729689e-02 1.85693949e-01 7.70250618e-01
-2.05555081e-01 -4.66027468e-01 7.15426624e-01 -8.84843320e-02
-5.59658408e-01 3.26879442e-01 -9.01114717e-02 7.93043301e-02
1.00323272e+00 -2.31594726e-01 -3.43461931e-01 -3.73450786e-01
-5.63298583e-01 8.49876583e-01 4.19615626e-01 5.27660251e-01
9.57493246e-01 -1.44326937e+00 -6.29996419e-01 2.94126332e-01
2.21535563e-01 2.55599350e-01 3.59942280e-02 6.25492215e-01
2.27589607e-02 3.03210467e-01 -1.06322393e-01 -7.79060721e-01
-1.01283920e+00 4.86068040e-01 3.40292782e-01 -3.98403525e-01
-2.36846313e-01 8.78762186e-01 3.80640775e-01 -2.62568206e-01
6.21812940e-01 -2.55028367e-01 -1.79048464e-01 -2.51028806e-01
5.79488218e-01 -1.12757057e-01 -3.16620499e-01 1.23278730e-01
-6.89090788e-02 3.02380949e-01 -5.88735640e-01 -1.78884566e-01
1.08919775e+00 -2.52293348e-01 3.68618816e-01 6.32967770e-01
8.75118256e-01 -1.89302310e-01 -1.94374406e+00 -3.72375071e-01
-2.04358771e-01 -4.96804923e-01 8.22862517e-03 -1.04914880e+00
-8.47286463e-01 7.99721777e-01 5.69099903e-01 -2.89416709e-03
1.21054518e+00 -1.35658592e-01 4.66919094e-01 5.55073023e-01
2.54390210e-01 -1.27280700e+00 5.32894850e-01 5.04328787e-01
1.04090118e+00 -1.47870779e+00 -2.42467910e-01 4.68296632e-02
-1.09664547e+00 1.02453232e+00 1.26160383e+00 -3.02904490e-02
2.41512716e-01 8.88181850e-02 -2.65786820e-03 2.90693849e-01
-1.14851236e+00 -1.24046862e-01 2.86712587e-01 6.31358683e-01
1.58754379e-01 -1.51434708e-02 -3.57674155e-03 4.19663191e-01
7.14983344e-02 2.72196025e-01 2.78140187e-01 8.43143344e-01
-6.99656069e-01 -1.03267646e+00 -3.24954718e-01 7.65080869e-01
-1.76261395e-01 -2.65104175e-01 -1.30433977e-01 4.86432284e-01
-1.16763380e-03 6.57378793e-01 9.19005796e-02 -4.93323565e-01
1.95152134e-01 2.70405322e-01 2.95462996e-01 -5.90154827e-01
-3.23845029e-01 3.42276722e-01 -3.47257197e-01 -4.81978714e-01
-1.16928257e-01 -8.16444874e-01 -9.74624395e-01 2.29120150e-01
-5.25317907e-01 -7.13023022e-02 4.60368931e-01 7.92686582e-01
4.64315832e-01 5.12182176e-01 9.31517422e-01 -5.94539344e-01
-1.21437633e+00 -1.07500088e+00 -2.57885545e-01 1.90753400e-01
4.10769016e-01 -5.34912825e-01 -1.98372543e-01 2.79783159e-02] | [9.837273597717285, 2.8782670497894287] |
4d568dfb-cccb-4cfe-8e8d-1504d54ac712 | multi-view-priors-for-learning-detectors-from | 1312.6095 | null | http://arxiv.org/abs/1312.6095v2 | http://arxiv.org/pdf/1312.6095v2.pdf | Multi-View Priors for Learning Detectors from Sparse Viewpoint Data | While the majority of today's object class models provide only 2D bounding
boxes, far richer output hypotheses are desirable including viewpoint,
fine-grained category, and 3D geometry estimate. However, models trained to
provide richer output require larger amounts of training data, preferably well
covering the relevant aspects such as viewpoint and fine-grained categories. In
this paper, we address this issue from the perspective of transfer learning,
and design an object class model that explicitly leverages correlations between
visual features. Specifically, our model represents prior distributions over
permissible multi-view detectors in a parametric way -- the priors are learned
once from training data of a source object class, and can later be used to
facilitate the learning of a detector for a target class. As we show in our
experiments, this transfer is not only beneficial for detectors based on
basic-level category representations, but also enables the robust learning of
detectors that represent classes at finer levels of granularity, where training
data is typically even scarcer and more unbalanced. As a result, we report
largely improved performance in simultaneous 2D object localization and
viewpoint estimation on a recent dataset of challenging street scenes. | ['Michael Stark', 'Peter Gehler', 'Bojan Pepik', 'Bernt Schiele'] | 2013-12-20 | null | null | null | null | ['viewpoint-estimation'] | ['computer-vision'] | [ 7.30781704e-02 1.24980807e-01 -3.14241141e-01 -4.78576660e-01
-7.57943630e-01 -7.89415777e-01 7.52672970e-01 3.32052588e-01
-1.44719079e-01 2.42844149e-01 2.75976032e-01 -1.24947220e-01
2.54157543e-01 -8.06663215e-01 -8.69827867e-01 -4.29100871e-01
1.01585209e-01 6.00204110e-01 6.83007717e-01 5.02992533e-02
3.20214838e-01 7.54664958e-01 -1.86314189e+00 4.02958870e-01
4.77389753e-01 1.09018159e+00 3.32787275e-01 3.77719104e-01
1.24791302e-01 4.55441236e-01 -2.52784729e-01 -1.66211903e-01
4.18739647e-01 5.77959418e-02 -2.93606550e-01 3.75235558e-01
9.52271461e-01 -5.62175035e-01 -1.76587954e-01 9.00021136e-01
2.84659415e-01 4.89162691e-02 1.14043164e+00 -1.16027176e+00
-3.79067034e-01 7.29826167e-02 -6.25266373e-01 1.41534284e-01
1.92088038e-01 2.94920623e-01 1.22293854e+00 -1.03337014e+00
4.86368001e-01 1.33986521e+00 5.61940491e-01 3.71235341e-01
-1.39767659e+00 -6.31772935e-01 7.46711612e-01 1.33203138e-02
-1.38536012e+00 -5.39062381e-01 7.05741107e-01 -7.35490799e-01
9.70562577e-01 -1.41260862e-01 6.52348399e-01 9.79994714e-01
-1.14310635e-02 9.65650260e-01 9.47430193e-01 -1.98450536e-01
3.19019854e-01 3.85413796e-01 1.57897398e-02 6.70819044e-01
5.34377158e-01 2.81117290e-01 -4.30651426e-01 -1.18564637e-02
1.05524385e+00 1.12644702e-01 -1.80937313e-02 -1.15645516e+00
-1.14602256e+00 9.72806752e-01 7.18257368e-01 -1.97431788e-01
-2.05409825e-01 6.18048981e-02 8.33391845e-02 -1.11867949e-01
5.35826981e-01 3.63817036e-01 -4.93759692e-01 2.85248160e-01
-8.97513211e-01 2.73544550e-01 5.55904269e-01 1.27203214e+00
9.11885142e-01 -1.22904390e-01 -2.53465593e-01 6.50908351e-01
5.16149998e-01 5.30093431e-01 -1.17708527e-01 -9.75898385e-01
5.69521129e-01 6.46638691e-01 3.25769067e-01 -8.68559837e-01
-4.16634470e-01 -5.39803326e-01 -3.70772302e-01 4.63390321e-01
5.57543457e-01 3.77384901e-01 -1.15785062e+00 1.83408642e+00
4.90572244e-01 -1.07004821e-01 -3.29953700e-01 1.02771974e+00
5.83671987e-01 4.01479870e-01 3.16597223e-01 3.01136822e-01
1.57796669e+00 -6.57567263e-01 3.82831581e-02 -7.05991566e-01
4.10070240e-01 -6.27914667e-01 8.68428528e-01 2.20585242e-01
-1.02779150e+00 -6.89567208e-01 -9.56499875e-01 -2.76005507e-01
-3.75003308e-01 2.93731660e-01 7.18768179e-01 5.42139411e-01
-9.43591654e-01 1.99928015e-01 -7.48660386e-01 -5.51895916e-01
6.85487866e-01 2.39704117e-01 -2.60341585e-01 -2.16061130e-01
-7.41568029e-01 1.14437413e+00 3.63083690e-01 -3.72506827e-01
-1.12204111e+00 -6.36587083e-01 -9.53943551e-01 2.58517802e-01
2.90422112e-01 -8.81573617e-01 1.19023681e+00 -5.41981101e-01
-1.08692527e+00 1.12208641e+00 -7.63308555e-02 -2.05125079e-01
4.32585627e-01 -9.46791321e-02 2.65799891e-02 1.46415591e-01
2.76770234e-01 1.06888020e+00 1.15773761e+00 -1.41777229e+00
-8.95300329e-01 -5.69047213e-01 3.43884289e-01 3.78640562e-01
-6.36867285e-02 -3.43272746e-01 -3.86992425e-01 -4.01127666e-01
3.40070903e-01 -8.03146839e-01 -7.52163529e-02 4.99985605e-01
-2.17826411e-01 -3.51818144e-01 6.74343228e-01 -4.12772447e-01
6.86591268e-01 -2.27825594e+00 -1.88840050e-02 3.74799632e-02
2.23839670e-01 -5.54291680e-02 1.89496689e-02 1.58025831e-01
8.18358138e-02 -3.56205702e-02 1.00541033e-01 -1.91582546e-01
2.54304707e-02 -7.56122395e-02 -4.34778333e-01 6.81149483e-01
5.07898629e-01 6.47430062e-01 -8.06211293e-01 -4.13327754e-01
5.35287917e-01 2.73372978e-01 -9.50934887e-01 1.24011613e-01
-3.91808808e-01 3.85558665e-01 -6.00428581e-01 5.79353809e-01
6.16689682e-01 -5.29991567e-01 -1.46148056e-01 -5.12029886e-01
-5.82195967e-02 4.16081101e-01 -1.27828074e+00 1.58979642e+00
-5.92255473e-01 4.69100833e-01 3.03709246e-02 -8.21058095e-01
8.24031174e-01 -4.70610857e-02 2.04848930e-01 -3.18429381e-01
1.03345633e-01 1.97225343e-02 -8.40598494e-02 -1.98576063e-01
4.98992801e-01 -8.35202858e-02 -1.46679819e-01 2.58554101e-01
8.82196650e-02 -4.50173050e-01 -2.43209414e-02 1.72046512e-01
8.38849127e-01 3.37764323e-01 6.09181404e-01 -1.40673712e-01
1.28293172e-01 4.69213724e-02 3.45609516e-01 9.19228256e-01
-3.28184694e-01 7.38636971e-01 2.63281047e-01 -2.64300466e-01
-1.28678191e+00 -1.38905466e+00 -4.01501268e-01 1.20367515e+00
3.99800926e-01 -1.38254836e-01 -2.12186173e-01 -7.15765834e-01
3.76299769e-01 7.03550160e-01 -5.67044914e-01 -9.41680595e-02
-3.04490864e-01 -4.23721254e-01 3.78952399e-02 8.64779770e-01
2.80883580e-01 -5.61163187e-01 -8.06171715e-01 3.27041969e-02
1.40573576e-01 -1.26450121e+00 -4.25272882e-01 4.04189765e-01
-9.70148146e-01 -1.18058133e+00 -4.44593906e-01 -7.08069503e-01
7.76464760e-01 6.91259027e-01 1.23995614e+00 -1.77290127e-01
-3.92465323e-01 5.29418051e-01 -1.62799656e-01 -4.38858926e-01
-1.11546963e-01 -1.78591609e-02 -1.38155110e-02 -5.73234893e-02
4.58104461e-01 -5.29729843e-01 -7.34821439e-01 3.06223959e-01
-4.67386276e-01 2.20806122e-01 8.52913082e-01 6.59126699e-01
5.26878834e-01 -2.94850528e-01 3.61760050e-01 -7.57928252e-01
-1.10058263e-01 -4.23769176e-01 -8.20049286e-01 9.30199176e-02
-2.12892950e-01 1.76911175e-01 3.56538057e-01 -4.31931883e-01
-1.00713110e+00 3.00101548e-01 1.73137292e-01 -4.70960140e-01
-5.95067382e-01 -8.56459141e-02 -2.87862003e-01 4.79486473e-02
8.58876228e-01 1.18479811e-01 -3.55327636e-01 -4.85392720e-01
6.35055602e-01 5.60640872e-01 1.52689844e-01 -6.48162663e-01
8.57842922e-01 7.87024438e-01 -3.78574841e-02 -5.86362481e-01
-1.17075181e+00 -7.52742767e-01 -7.94604659e-01 -3.06354836e-02
1.00214124e+00 -1.46260786e+00 -4.07791078e-01 8.77619162e-02
-1.19299316e+00 -1.86029613e-01 -2.79033303e-01 5.97153306e-01
-7.44112372e-01 -3.68356407e-02 -3.24047059e-01 -6.32344425e-01
2.73252189e-01 -1.07344663e+00 1.56830227e+00 7.07337707e-02
-1.27956718e-01 -7.96182752e-01 -2.84034520e-01 1.60159722e-01
1.37340739e-01 2.23772041e-02 1.07449818e+00 -5.80726385e-01
-1.12580109e+00 -2.46772125e-01 -5.36420286e-01 1.50440499e-01
5.26824147e-02 -1.97726250e-01 -1.21631622e+00 -3.76433522e-01
-1.54701531e-01 -3.56124282e-01 1.00336874e+00 4.43900824e-01
1.21492040e+00 1.12756059e-01 -6.58174396e-01 5.13583183e-01
1.33691037e+00 -7.28101656e-02 1.20444953e-01 9.67570692e-02
5.85957646e-01 6.29642367e-01 5.84588110e-01 3.87414962e-01
4.81631160e-01 8.99486303e-01 5.19293845e-01 9.55979973e-02
-4.10771400e-01 -4.93531138e-01 7.29551688e-02 2.71644503e-01
1.38604611e-01 -7.37524703e-02 -8.69844437e-01 6.77062750e-01
-1.64965034e+00 -1.02773857e+00 3.03786844e-01 2.24490476e+00
5.61348200e-01 4.28737819e-01 2.40526900e-01 -2.52438724e-01
8.13131213e-01 5.17487340e-02 -8.46038282e-01 7.39170760e-02
1.90956935e-01 -1.56358987e-01 6.45751357e-01 3.39197338e-01
-1.38096189e+00 9.17090893e-01 6.34877110e+00 6.31230354e-01
-8.28897476e-01 -8.63104984e-02 4.89448786e-01 -6.59669787e-02
-2.36654460e-01 1.14402942e-01 -1.29247296e+00 1.80980533e-01
3.55083436e-01 2.64368176e-01 1.98961496e-01 1.23961210e+00
1.75585598e-02 -2.17498377e-01 -1.57248390e+00 9.15851235e-01
2.88884751e-02 -1.20746219e+00 7.96960741e-02 3.06926221e-01
5.59422255e-01 1.71747610e-01 5.33494428e-02 4.59950775e-01
4.46270496e-01 -7.56986558e-01 1.06994343e+00 1.51838362e-01
7.36742198e-01 -5.36911428e-01 1.10791475e-01 6.87614381e-01
-1.24899626e+00 -2.68403143e-01 -5.67291379e-01 -1.56775057e-01
-4.72734571e-02 3.85868728e-01 -9.73538101e-01 1.33672282e-01
6.27622545e-01 8.84977996e-01 -6.75938964e-01 1.19004607e+00
-3.58884275e-01 2.48483524e-01 -3.68607163e-01 5.54546192e-02
1.86643470e-02 1.45721436e-01 3.67055923e-01 9.57590461e-01
2.42196381e-01 1.77956372e-01 5.06162286e-01 8.25243533e-01
1.08377427e-01 -2.64345676e-01 -7.75940597e-01 3.71956378e-01
7.61363387e-01 1.24872112e+00 -8.26918483e-01 -3.24031204e-01
-6.92499220e-01 4.35156256e-01 6.72059357e-01 2.96163678e-01
-7.76452243e-01 3.48737575e-02 7.11049139e-01 4.68119293e-01
8.97692680e-01 -4.13693577e-01 -4.30647999e-01 -1.37143385e+00
-2.53685027e-01 -5.61750114e-01 3.10049176e-01 -8.37385714e-01
-1.48805344e+00 1.80485442e-01 3.87537748e-01 -1.36767960e+00
-3.07447582e-01 -1.09910572e+00 -3.54835421e-01 8.33915830e-01
-1.47109210e+00 -1.35586059e+00 -2.61585206e-01 3.93439084e-01
5.68228304e-01 7.66250677e-03 6.28291726e-01 1.60692558e-01
-1.01092942e-01 3.42214674e-01 -1.31404400e-01 8.29267353e-02
6.57914340e-01 -1.13728297e+00 3.66873801e-01 7.12833047e-01
2.78927177e-01 5.66871703e-01 6.26420557e-01 -4.80460823e-01
-1.09342456e+00 -1.17309785e+00 4.87301379e-01 -9.28643584e-01
4.36227471e-01 -6.57446384e-01 -5.84705293e-01 7.71333873e-01
-3.98601383e-01 1.86382204e-01 4.63966161e-01 4.43755209e-01
-8.04815531e-01 -7.76707474e-03 -9.91150856e-01 4.83160287e-01
1.19435310e+00 -7.51677752e-01 -6.56338513e-01 3.27376395e-01
4.29675668e-01 -4.84154373e-01 -5.87487042e-01 4.15897548e-01
4.62641925e-01 -1.00691080e+00 1.28858650e+00 -5.73548675e-01
1.39942676e-01 -5.26937366e-01 -5.65193295e-01 -1.17267036e+00
-5.06567419e-01 2.62302309e-01 -2.87686765e-01 8.84034514e-01
1.58724263e-01 -3.34407389e-01 7.42700756e-01 4.51143175e-01
-2.64947236e-01 -6.42330945e-01 -9.37307000e-01 -7.48736918e-01
2.09485237e-02 -5.10777414e-01 4.49911088e-01 5.65253317e-01
-4.22665179e-01 5.14061630e-01 -2.88796369e-02 4.74559903e-01
8.45348060e-01 5.31190455e-01 8.61214936e-01 -1.34793806e+00
-4.38070744e-01 -4.80135888e-01 -6.10460639e-01 -1.52123606e+00
1.76065695e-02 -8.60141098e-01 1.80818141e-01 -1.49820054e+00
3.07534546e-01 -6.51657045e-01 -4.20095287e-02 4.28770572e-01
-6.71502873e-02 1.82031229e-01 2.82907635e-01 3.55876446e-01
-6.05513692e-01 4.45667207e-01 1.24894524e+00 5.94707578e-03
9.55622271e-02 2.36816823e-01 -7.79854536e-01 9.20109510e-01
5.15878797e-01 -2.87863791e-01 -5.59378207e-01 -6.20819569e-01
-1.02049783e-01 -5.91626354e-02 8.29516828e-01 -9.51928318e-01
9.80673954e-02 -2.58664459e-01 8.41561615e-01 -9.23054099e-01
6.86835468e-01 -8.65085602e-01 -3.55982542e-01 1.35149494e-01
-2.17295796e-01 -3.91568214e-01 1.48212552e-01 8.89357388e-01
1.38383761e-01 3.02933082e-02 9.49446321e-01 -3.36657792e-01
-9.93332505e-01 4.86042440e-01 -4.37084362e-02 1.18111826e-01
1.00843358e+00 -5.31208277e-01 -2.91159689e-01 -3.07926327e-01
-8.45918715e-01 1.07261062e-01 7.95034409e-01 4.96410608e-01
5.19411743e-01 -1.24728096e+00 -4.24985379e-01 4.09364671e-01
4.49947000e-01 2.28773490e-01 -4.20482643e-02 4.30047780e-01
-1.36859426e-02 4.55939889e-01 -1.92136124e-01 -1.06485927e+00
-7.58887708e-01 7.49750853e-01 2.50056893e-01 1.44315153e-01
-5.43789625e-01 8.99593294e-01 9.73892450e-01 -4.14706290e-01
3.04849178e-01 -4.19641346e-01 -6.13530241e-02 1.05017744e-01
2.78728396e-01 4.44772467e-02 -6.07272200e-02 -5.98624408e-01
-4.63342845e-01 7.07261443e-01 -8.85662809e-02 4.45650481e-02
1.25342691e+00 -2.66919911e-01 4.41110462e-01 4.15247321e-01
9.96080220e-01 -1.65872484e-01 -2.04943013e+00 -3.17655891e-01
-1.98222697e-02 -6.54175878e-01 6.88878447e-02 -6.79842055e-01
-5.93334734e-01 1.13210094e+00 5.53944826e-01 -2.98004020e-02
7.88294196e-01 5.39430201e-01 1.88694775e-01 2.96321720e-01
6.16658926e-01 -9.65749383e-01 3.57272089e-01 5.08090496e-01
6.38980746e-01 -1.45380282e+00 1.14223987e-01 -5.83402157e-01
-3.58597159e-01 9.42418098e-01 8.66328120e-01 -2.59240329e-01
5.97217500e-01 6.97038174e-02 -1.94201648e-01 -2.49527752e-01
-5.84957123e-01 -3.89180362e-01 4.62308824e-01 8.21518779e-01
2.88780570e-01 -3.54808271e-02 5.40378153e-01 3.73447627e-01
-2.08581723e-02 -4.21223313e-01 2.03957886e-01 7.89289892e-01
-9.15751576e-01 -8.06383431e-01 -3.77672523e-01 5.90866446e-01
1.41084537e-01 -5.00236414e-02 4.31022234e-03 9.82588649e-01
3.23601604e-01 5.62067211e-01 3.61704111e-01 -5.43716475e-02
3.52104813e-01 -1.25439987e-01 8.65675211e-01 -1.25282037e+00
6.79718032e-02 -5.93961566e-04 -8.95274803e-02 -5.21875501e-01
-3.26817691e-01 -9.14245367e-01 -8.82621288e-01 4.47499231e-02
-3.85794967e-01 -2.33757019e-01 5.30678630e-01 9.65698123e-01
3.07616264e-01 2.19808906e-01 5.33874869e-01 -1.31474721e+00
-8.15729439e-01 -7.04797924e-01 -5.47233522e-01 3.52666348e-01
3.98417860e-01 -1.19103706e+00 -2.96725780e-01 -3.80251370e-02] | [7.789819240570068, -2.8108577728271484] |
df138635-6249-483d-b1c9-a311d927f982 | decouple-learning-for-parameterized-image | 1807.08186 | null | http://arxiv.org/abs/1807.08186v2 | http://arxiv.org/pdf/1807.08186v2.pdf | Decouple Learning for Parameterized Image Operators | Many different deep networks have been used to approximate, accelerate or
improve traditional image operators, such as image smoothing, super-resolution
and denoising. Among these traditional operators, many contain parameters which
need to be tweaked to obtain the satisfactory results, which we refer to as
"parameterized image operators". However, most existing deep networks trained
for these operators are only designed for one specific parameter configuration,
which does not meet the needs of real scenarios that usually require flexible
parameters settings. To overcome this limitation, we propose a new decouple
learning algorithm to learn from the operator parameters to dynamically adjust
the weights of a deep network for image operators, denoted as the base network.
The learned algorithm is formed as another network, namely the weight learning
network, which can be end-to-end jointly trained with the base network.
Experiments demonstrate that the proposed framework can be successfully applied
to many traditional parameterized image operators. We provide more analysis to
better understand the proposed framework, which may inspire more promising
research in this direction. Our codes and models have been released in
https://github.com/fqnchina/DecoupleLearning | ['Dong-Dong Chen', 'Baoquan Chen', 'Qingnan Fan', 'Nenghai Yu', 'Lu Yuan', 'Gang Hua'] | 2018-07-21 | decouple-learning-for-parameterized-image-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Qingnan_Fan_Learning_to_Learn_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Qingnan_Fan_Learning_to_Learn_ECCV_2018_paper.pdf | eccv-2018-9 | ['image-smoothing'] | ['computer-vision'] | [ 2.26240039e-01 -3.26219112e-01 -1.02365069e-01 -4.47461963e-01
-1.22161463e-01 -2.06423119e-01 1.08364560e-01 -2.86531538e-01
-5.02311885e-01 2.66601413e-01 -3.60258482e-02 -1.71899080e-01
-5.55305555e-02 -6.26191676e-01 -6.21147335e-01 -9.41596150e-01
2.35149115e-01 2.38053612e-02 3.37303340e-01 -1.25657916e-01
1.77758083e-01 3.38055879e-01 -1.22662008e+00 6.84429482e-02
1.08094442e+00 8.98361981e-01 4.38030779e-01 3.45568448e-01
-1.46368951e-01 5.12422860e-01 -3.40578705e-01 -1.65963948e-01
3.11952472e-01 -4.00667489e-01 -3.28760296e-01 1.69398144e-01
9.81157944e-02 -5.03256977e-01 -3.86833429e-01 1.48622715e+00
6.88453615e-01 1.95455223e-01 2.70119011e-01 -1.08607161e+00
-8.77345562e-01 8.43608201e-01 -7.92428374e-01 2.66091287e-01
-6.69938549e-02 2.89283395e-01 6.41009629e-01 -7.80392647e-01
1.58741578e-01 1.37519860e+00 6.43753529e-01 6.07892454e-01
-9.58832443e-01 -9.69749928e-01 4.65802789e-01 3.35347146e-01
-1.50595820e+00 -4.75593090e-01 1.01559651e+00 -1.25854060e-01
3.70298296e-01 -5.90879694e-02 4.84457433e-01 8.78596485e-01
-1.57301441e-01 9.67182338e-01 8.80127370e-01 -2.13285223e-01
-1.05987757e-01 1.76124692e-01 -3.95991839e-02 7.91507363e-01
1.61036983e-01 -1.56373173e-01 -2.10895956e-01 1.64896443e-01
1.18451786e+00 2.22113058e-01 -6.69393718e-01 -4.67899084e-01
-1.14729178e+00 7.07846165e-01 7.29448080e-01 2.26877227e-01
-3.28952253e-01 1.77422404e-01 5.93073666e-01 2.54663080e-01
4.40562844e-01 -9.27916169e-03 -4.96867627e-01 -2.72081830e-02
-6.03767216e-01 1.54520929e-01 3.69509429e-01 7.88408041e-01
9.79939640e-01 2.71878928e-01 -2.99916454e-02 1.08526182e+00
2.78709084e-01 2.61867732e-01 6.53543770e-01 -8.62347305e-01
4.53052163e-01 6.20023489e-01 2.44533733e-01 -1.09823298e+00
-5.36556900e-01 -5.56540072e-01 -1.07693589e+00 1.00863114e-01
1.35529459e-01 -4.39402550e-01 -8.70584667e-01 1.66961110e+00
4.49433386e-01 7.15643108e-01 -1.02152653e-01 1.20893359e+00
7.96863675e-01 9.50321794e-01 3.50748971e-02 -1.12463132e-01
1.26345336e+00 -1.19675791e+00 -6.53069377e-01 -2.51498401e-01
4.13308859e-01 -7.76108503e-01 1.37936997e+00 4.58150953e-01
-1.19789267e+00 -7.24378586e-01 -1.07940793e+00 -1.27866924e-01
-9.21759456e-02 4.12239820e-01 6.77674294e-01 3.06421965e-01
-9.68583763e-01 3.62094551e-01 -1.02476883e+00 -1.96837381e-01
5.12379706e-01 3.79247218e-01 9.49127972e-02 1.33380875e-01
-1.16881716e+00 5.85383773e-01 7.17652082e-01 6.67017579e-01
-9.32052672e-01 -5.42047977e-01 -4.92893040e-01 1.74113810e-01
5.86530566e-01 -5.76715529e-01 1.24278247e+00 -1.09212112e+00
-1.58808064e+00 6.63800418e-01 -6.30414858e-02 -4.09086108e-01
4.98954415e-01 -2.96443820e-01 -3.67079079e-01 -2.95856073e-02
-2.09189177e-01 4.80527461e-01 9.84611928e-01 -1.22926688e+00
-7.13908434e-01 4.95359162e-03 1.97242677e-01 2.50815362e-01
-8.78082395e-01 2.09610492e-01 -9.92146730e-01 -9.15916741e-01
1.25961686e-02 -8.45347166e-01 -2.76436955e-01 2.11815313e-01
-2.94281125e-01 -1.84508681e-01 9.59566891e-01 -4.78751898e-01
1.38009226e+00 -2.27213049e+00 3.75601053e-01 7.08163008e-02
2.46210024e-01 6.53714120e-01 -2.81283706e-01 1.72577471e-01
-2.19967470e-01 -6.06802739e-02 -3.53322268e-01 -2.97803223e-01
-1.81915089e-01 2.43225902e-01 -1.64858207e-01 4.26426083e-01
2.88962238e-02 5.88977933e-01 -8.02163780e-01 -4.62580562e-01
3.10982704e-01 6.50577009e-01 -4.12902832e-01 2.25685626e-01
-1.68294042e-01 4.35140371e-01 -7.72714794e-01 3.96519959e-01
1.07687926e+00 -2.66522318e-01 -5.13292737e-02 -6.64362371e-01
-1.17351435e-01 -2.54656166e-01 -1.47844195e+00 1.57133520e+00
-4.99339372e-01 1.91030890e-01 4.62634414e-01 -1.32056439e+00
1.10551023e+00 2.05260202e-01 4.31390047e-01 -3.97921294e-01
2.63085812e-01 1.69575706e-01 1.08824819e-01 -7.30506241e-01
1.58104911e-01 1.88275091e-02 3.50869596e-01 2.44774163e-01
-1.77390948e-01 2.91744143e-01 2.51756936e-01 -2.40134690e-02
5.62371373e-01 9.06612948e-02 4.35127355e-02 -4.06393707e-02
1.06936836e+00 -3.74788046e-01 9.71557498e-01 4.31527793e-01
-2.93687940e-01 5.29958129e-01 4.83760923e-01 -6.54934406e-01
-9.24600542e-01 -9.31812763e-01 -1.07897535e-01 1.11667371e+00
4.80999768e-01 -1.08718067e-01 -7.72550583e-01 -5.60675144e-01
-1.53011337e-01 2.19247743e-01 -4.16130841e-01 -2.37826690e-01
-7.55456865e-01 -8.89928341e-01 3.26338619e-01 5.97725689e-01
9.97286320e-01 -1.19239318e+00 -2.67075092e-01 2.78199047e-01
4.33196360e-03 -1.06117165e+00 -8.25697064e-01 -1.28366813e-01
-8.47047031e-01 -8.49033415e-01 -8.64977419e-01 -1.06666970e+00
7.88102508e-01 4.64614600e-01 5.91185033e-01 5.59464037e-01
1.82995453e-01 -3.49265523e-02 -4.20262754e-01 -3.00839663e-01
-1.90052763e-01 3.55925471e-01 5.70734255e-02 4.57038850e-01
1.37372538e-01 -8.47963750e-01 -8.43680680e-01 3.72253746e-01
-1.26492608e+00 2.30292693e-01 8.60133529e-01 8.25506508e-01
5.55490613e-01 1.89255625e-01 5.59342504e-01 -1.00807583e+00
8.52335274e-01 -2.59229660e-01 -7.22081184e-01 2.88295299e-01
-4.91470873e-01 1.17418230e-01 8.74612153e-01 -7.41986156e-01
-1.28219450e+00 1.75111562e-01 -1.71418771e-01 -5.63628793e-01
3.78015377e-02 6.55060887e-01 -2.59201378e-01 -4.12695736e-01
4.29027379e-01 2.57508695e-01 -1.64643507e-02 -6.09293044e-01
3.93312275e-01 6.82007194e-01 7.31186569e-01 -5.37163556e-01
9.70577180e-01 3.64363760e-01 -3.20703775e-01 -3.14955413e-01
-8.64787877e-01 -2.64595419e-01 -2.64504224e-01 -1.19387932e-01
7.42636800e-01 -9.61827457e-01 -6.59463584e-01 7.70207167e-01
-1.11241579e+00 -3.88751775e-01 2.30708480e-01 4.59681571e-01
-2.57981926e-01 4.28031087e-01 -6.96192980e-01 -5.00635266e-01
-3.64209116e-01 -1.43603384e+00 8.04480016e-01 8.20115149e-01
3.84233952e-01 -1.16984797e+00 -2.82773554e-01 1.00722983e-01
6.55340135e-01 1.48187205e-01 6.95655644e-01 -1.82043478e-01
-5.39888680e-01 -6.66757748e-02 -5.04343510e-01 6.12818420e-01
2.39587322e-01 2.09134221e-02 -5.77658594e-01 -4.76642668e-01
1.40540048e-01 -3.58035415e-01 9.56624985e-01 2.54824489e-01
1.67707205e+00 -2.61397421e-01 -2.47220695e-01 1.21194160e+00
1.38386738e+00 2.41176024e-01 6.08141363e-01 4.61885870e-01
8.16140950e-01 1.42846420e-01 3.35102618e-01 5.89769661e-01
2.80884326e-01 5.80806017e-01 4.02544707e-01 -2.65618891e-01
3.10714804e-02 -1.36790834e-02 2.56303787e-01 8.96616459e-01
-2.62964368e-01 -1.09805278e-01 -7.78258085e-01 3.07220191e-01
-1.95438886e+00 -6.95736766e-01 1.29229948e-01 1.72247577e+00
9.84064519e-01 2.15767443e-01 -5.09586185e-02 -6.00123927e-02
9.43211377e-01 5.26708186e-01 -8.47473919e-01 -3.17976505e-01
2.11762488e-02 1.01692438e-01 4.18615609e-01 4.09085542e-01
-1.16586161e+00 9.78928149e-01 5.03777122e+00 9.20002759e-01
-1.44815528e+00 1.00889958e-01 4.36459750e-01 1.07651688e-01
-1.96943820e-01 7.86935091e-02 -6.82605803e-01 6.20770693e-01
4.70508724e-01 -3.28685343e-01 6.46313608e-01 8.95112574e-01
7.00499415e-01 4.39957529e-01 -6.59516454e-01 1.11983240e+00
-1.31010234e-01 -1.05460346e+00 1.44286945e-01 -4.08899039e-01
7.43834794e-01 -1.43221870e-01 1.71873271e-01 2.81390399e-01
-3.84089909e-02 -6.71916842e-01 3.42370957e-01 6.46616459e-01
4.70216900e-01 -8.21932077e-01 7.04356909e-01 3.07023227e-01
-1.34386194e+00 -1.74130574e-01 -6.21372163e-01 1.65454313e-01
1.57816678e-01 5.03448904e-01 -2.22621620e-01 4.76785719e-01
6.76967442e-01 9.34779823e-01 -4.90214527e-01 1.30782425e+00
-3.18239093e-01 5.17601371e-01 -1.61055759e-01 9.79700014e-02
3.05759728e-01 -3.14839989e-01 2.32567459e-01 9.81383145e-01
4.14300799e-01 3.71763706e-02 3.24278921e-01 7.50293136e-01
-2.08316684e-01 4.73114662e-02 -3.60182561e-02 7.43486658e-02
5.61607778e-01 1.48705018e+00 -6.80785894e-01 -2.92034775e-01
-6.01983070e-01 1.00238895e+00 2.98848480e-01 5.58518171e-01
-1.16840267e+00 -6.45407915e-01 4.98026162e-01 -1.45630673e-01
3.01569849e-01 -2.24364251e-01 -7.31713548e-02 -1.37230432e+00
8.48979354e-02 -1.02440798e+00 2.14845344e-01 -8.87077570e-01
-1.22284973e+00 4.72003996e-01 7.16483071e-02 -1.30008411e+00
3.12303036e-01 -6.73166215e-01 -9.82808888e-01 9.03300107e-01
-1.53739846e+00 -9.16858912e-01 -7.47269750e-01 6.52549267e-01
4.87533957e-01 -5.03669046e-02 3.63212198e-01 6.24140561e-01
-1.09054804e+00 7.74953604e-01 1.22481495e-01 2.11660355e-01
7.98136830e-01 -8.89333546e-01 1.57130480e-01 9.13227081e-01
-2.20301181e-01 6.41275644e-01 5.91515541e-01 -2.55294472e-01
-1.34265673e+00 -1.09308600e+00 2.62140989e-01 3.85190606e-01
6.92516863e-01 -1.27114534e-01 -1.19572079e+00 7.88249850e-01
2.64062911e-01 3.31608593e-01 6.64220303e-02 -1.93981022e-01
-2.77825985e-02 -6.97297454e-01 -8.61489058e-01 7.18328357e-01
9.25180256e-01 -1.65662244e-01 -2.16980502e-01 3.84070545e-01
8.56295407e-01 -7.80149460e-01 -7.32889771e-01 3.97192150e-01
2.82569587e-01 -8.38108242e-01 9.91558611e-01 -3.26169044e-01
2.86173970e-01 -6.12304330e-01 3.63957763e-01 -1.22094524e+00
-3.61048669e-01 -5.45444548e-01 -8.83336440e-02 1.30366099e+00
2.82487571e-01 -8.21933985e-01 8.55621517e-01 4.80384946e-01
-2.49150857e-01 -1.06221378e+00 -6.34480059e-01 -5.55838943e-01
-1.37096435e-01 -1.75300300e-01 8.59223068e-01 9.16137695e-01
-5.93155265e-01 2.28472576e-01 -6.24065042e-01 4.61543649e-01
4.70105559e-01 -9.14612412e-02 6.93954706e-01 -8.89407456e-01
-2.50481308e-01 -5.93138397e-01 -2.52816975e-01 -1.36578822e+00
-4.74073254e-02 -6.92124009e-01 5.05755730e-02 -1.34949851e+00
1.45398170e-01 -5.94596684e-01 -5.76160789e-01 4.55275536e-01
-4.61694270e-01 4.36897017e-02 3.49900834e-02 3.11166435e-01
-4.96357054e-01 8.04684281e-01 1.54157424e+00 -1.44894764e-01
-3.67335826e-01 1.02072962e-01 -9.22288120e-01 9.64994013e-01
1.01214838e+00 -2.27334738e-01 -7.08118796e-01 -9.77512121e-01
1.99265666e-02 -7.23195300e-02 2.75397480e-01 -1.16716218e+00
3.71756673e-01 -3.35237414e-01 2.33921871e-01 -2.36967251e-01
6.87674507e-02 -8.42963457e-01 -5.23419641e-02 3.07245016e-01
-3.75581563e-01 2.22987860e-01 1.77330121e-01 4.37661886e-01
-2.44581833e-01 -3.89646709e-01 9.05076802e-01 -9.05790180e-02
-8.99729073e-01 7.35665083e-01 7.84277543e-02 -3.59550193e-02
9.36124980e-01 4.60693166e-02 -1.23168238e-01 -4.37791228e-01
-6.50497377e-01 5.43369830e-01 2.71424115e-01 3.03794473e-01
7.25696862e-01 -1.39773142e+00 -6.99619889e-01 7.49081299e-02
-1.98172897e-01 2.43652374e-01 4.56501603e-01 8.98058116e-01
-8.58405054e-01 -7.26299360e-02 -1.11131310e-01 -4.58283961e-01
-1.05207300e+00 7.27133512e-01 5.74246585e-01 -1.55038059e-01
-6.29087210e-01 8.20853770e-01 1.87455177e-01 -2.65313268e-01
4.62434649e-01 -3.54975283e-01 -2.14873701e-01 -2.95589745e-01
5.48187912e-01 1.00748174e-01 -3.37764710e-01 -4.01224524e-01
-2.37538636e-01 7.48365521e-01 -2.82698542e-01 2.63786703e-01
1.41892076e+00 -3.34059000e-01 -3.29443723e-01 1.34447411e-01
1.23126447e+00 -1.50130153e-01 -1.48313260e+00 -5.59809923e-01
-4.07703876e-01 -3.94199491e-01 1.82114333e-01 -3.68112355e-01
-1.73316419e+00 9.30531323e-01 8.98654580e-01 1.83538318e-01
1.69424987e+00 -4.96804982e-01 1.11731553e+00 4.28845674e-01
5.65172955e-02 -1.05764163e+00 4.63958643e-03 2.84430146e-01
8.35805595e-01 -1.12889373e+00 2.94748656e-02 -2.59053349e-01
-6.12519503e-01 1.30088115e+00 9.53622758e-01 -3.27055961e-01
7.08866119e-01 2.22576588e-01 2.94351965e-01 1.63797185e-01
-4.68013376e-01 -5.84548302e-02 -3.49605270e-02 3.66792858e-01
4.05318558e-01 -1.68955639e-01 -5.16585290e-01 5.58436394e-01
1.22234508e-01 3.35544646e-01 3.24619353e-01 5.23773372e-01
-4.69367057e-01 -1.28888977e+00 -2.69298226e-01 3.01927269e-01
-2.99962670e-01 8.15731846e-03 2.21917063e-01 6.72274768e-01
5.54424345e-01 5.96512318e-01 -9.35682803e-02 -3.96392524e-01
3.92498761e-01 -3.04443240e-01 1.32095039e-01 -3.60863566e-01
-3.87028605e-01 2.16698591e-02 -2.29270607e-01 -3.95698845e-01
-7.06797719e-01 -3.94151807e-01 -1.27192044e+00 -4.35128212e-01
-2.25886077e-01 2.95742489e-02 1.45262077e-01 7.40486026e-01
1.52893439e-01 6.37455821e-01 6.80763066e-01 -8.99246931e-01
-5.14870703e-01 -9.00519371e-01 -3.95781934e-01 3.90572786e-01
2.24879935e-01 -6.73633933e-01 -2.76351333e-01 1.17039241e-01] | [11.057214736938477, -1.3483364582061768] |
6dfcc9e6-6bfa-4bfb-991d-2bdc10f4f0e7 | multiple-human-3d-pose-estimation-from | null | null | https://link.springer.com/article/10.1007/s11042-017-5133-8 | https://link.springer.com/article/10.1007/s11042-017-5133-8 | Multiple human 3d pose estimation from multiview images | Multiple human 3D pose estimation is a challenging task. It is mainly because of large variations in the scale and pose of humans, fast motions, multiple persons in the scene, and arbitrary number of visible body parts due to occlusion or truncation. Some of these ambiguities can be resolved by using multiview images. This is due to the fact that more evidences of body parts would be available in multiple views. In this work, a novel method for multiple human 3D pose estimation using evidences in multiview images is proposed. The proposed method utilizes a fully connected pairwise conditional random field that contains two types of pairwise terms. The first pairwise term encodes the spatial dependencies among human body joints based on an articulated human body configuration. The second pairwise term is based on the output of a 2D deep part detector. An approximate inference is then performed using the loopy belief propagation algorithm. The proposed method is evaluated on the Campus, Shelf, Utrecht Multi-Person Motion benchmark, Human3.6M, KTH Football II, and MPII Cooking datasets. Experimental results indicate that the proposed method achieves substantial improvements over the existing state-of-the-art methods in terms of the probability of correct pose and the mean per joint position error performance measures. | ['Shohreh Kasaei', 'Sara Ershadi-Nasab', 'Esmaeil Sanaei', 'Erfan Noury'] | 2017-09-04 | null | null | null | null | ['3d-multi-person-pose-estimation'] | ['computer-vision'] | [-1.59402505e-01 -1.19943105e-01 2.42180433e-02 -1.54896557e-01
-5.63942790e-01 -2.54451782e-01 4.14818406e-01 -9.72801968e-02
-4.54267889e-01 5.97895563e-01 2.03128412e-01 6.79563046e-01
7.01772347e-02 -2.76859730e-01 -7.53145754e-01 -6.13775969e-01
-5.88061772e-02 1.01352370e+00 5.92706561e-01 -1.35362640e-01
-1.07098684e-01 1.95102140e-01 -1.39017284e+00 -3.68745849e-02
3.30124617e-01 6.83281660e-01 1.67903662e-01 5.87702930e-01
5.03962755e-01 3.96606386e-01 -5.37531435e-01 -4.50402051e-01
3.04566503e-01 -1.81565210e-01 -5.10606349e-01 4.80033010e-01
5.84049165e-01 -4.89844978e-01 -1.11222945e-01 9.37835336e-01
6.98640704e-01 4.75452006e-01 6.62871540e-01 -1.29734337e+00
2.09943607e-01 1.14191517e-01 -9.29362297e-01 -1.34640023e-01
7.72279918e-01 2.37384979e-02 7.68136263e-01 -8.89253438e-01
8.22211862e-01 1.73136616e+00 5.81720173e-01 2.69499034e-01
-9.32043016e-01 -3.34537506e-01 4.99233231e-02 3.22140068e-01
-1.45817232e+00 -1.41407818e-01 6.91829383e-01 -6.69148862e-01
6.96378410e-01 -2.92993784e-02 9.08647180e-01 9.93810594e-01
5.71690619e-01 7.63922930e-01 8.57572198e-01 -4.58585590e-01
1.78973358e-02 -1.74979195e-01 -2.81014331e-02 1.08255076e+00
4.76729065e-01 -9.51836258e-02 -6.61725938e-01 -2.96905428e-01
8.15754592e-01 6.93291724e-02 -1.89719379e-01 -9.23476458e-01
-1.37151980e+00 6.35259092e-01 2.39348218e-01 -1.35238379e-01
-4.27405536e-01 1.75983340e-01 3.17801148e-01 -5.01227140e-01
1.51662454e-01 -3.27229798e-01 -3.43816191e-01 6.46972060e-02
-9.18630958e-01 6.95126832e-01 6.30835652e-01 8.73106956e-01
3.32408279e-01 -2.22614542e-01 1.23275053e-02 6.24291420e-01
8.58749747e-01 6.56282187e-01 2.76905119e-01 -1.10242355e+00
6.61956310e-01 4.71275538e-01 4.11713243e-01 -1.23144484e+00
-5.16010702e-01 -1.96942016e-01 -8.51120234e-01 2.00641930e-01
5.54455340e-01 -1.66958068e-02 -8.79753828e-01 1.55983651e+00
9.34700310e-01 -3.82712066e-01 -3.34824145e-01 1.27283561e+00
6.71406567e-01 3.78967226e-01 -1.53913712e-02 4.43924144e-02
1.58440721e+00 -1.03033137e+00 -8.22855294e-01 -4.44044977e-01
-9.04100686e-02 -8.20720255e-01 2.20506400e-01 4.17571604e-01
-1.17760360e+00 -8.71136487e-01 -1.04068899e+00 1.14183061e-01
4.50539961e-02 2.15720698e-01 1.02103509e-01 5.35416007e-01
-4.20535326e-01 4.23911452e-01 -1.13555682e+00 -4.48050559e-01
-1.34172231e-01 2.38762081e-01 -6.58852041e-01 -2.80376732e-01
-1.04235506e+00 1.08557165e+00 4.04380977e-01 3.95182967e-01
-7.22819924e-01 -7.60137439e-02 -1.17285764e+00 -1.86529577e-01
3.74571264e-01 -1.17037070e+00 9.09896255e-01 -4.11452711e-01
-1.30352855e+00 6.77386224e-01 -6.15166910e-02 -1.52028665e-01
9.29954350e-01 -6.71680272e-01 4.09792885e-02 3.99661154e-01
2.51320302e-01 6.41148984e-01 8.36662948e-01 -1.20269632e+00
-3.55465114e-01 -8.94375682e-01 -4.09507513e-01 6.54910386e-01
3.67258430e-01 -2.93166935e-01 -9.24151003e-01 -6.04750276e-01
4.49336082e-01 -1.36919880e+00 -2.98505902e-01 7.92848617e-02
-4.91133213e-01 -1.19115859e-01 5.87358057e-01 -1.04802418e+00
7.01664388e-01 -1.81921053e+00 7.62064099e-01 1.96202263e-01
-8.70721862e-02 -3.87939587e-02 3.19580227e-01 3.56821179e-01
3.96821678e-01 -4.18544114e-01 8.53911191e-02 -5.09477258e-01
5.37008122e-02 2.42366686e-01 4.30146247e-01 8.75051379e-01
-3.28743994e-01 4.47091430e-01 -6.02982402e-01 -9.07158792e-01
4.83253092e-01 6.70125067e-01 -2.98008621e-01 1.39753878e-01
5.97645715e-02 7.74873793e-01 -3.51103097e-01 6.01547718e-01
6.32611215e-01 -2.03891203e-01 3.98103923e-01 -5.23174286e-01
9.30359736e-02 -2.52170593e-01 -1.79155278e+00 2.09774947e+00
1.99522763e-01 -1.93999540e-02 1.43723920e-01 -7.90488720e-01
5.64988434e-01 6.28104329e-01 5.60902297e-01 -7.48399198e-02
2.03802779e-01 3.21299359e-02 -2.22191080e-01 -3.85431856e-01
4.26392943e-01 -1.02011189e-01 -1.68319538e-01 -7.40584731e-02
3.45049173e-01 1.05622426e-01 2.55150706e-01 2.66510341e-02
6.71167910e-01 6.80105448e-01 6.72174096e-01 -1.10464521e-01
4.01508570e-01 -2.01070413e-01 8.84099722e-01 4.86103117e-01
-4.78788853e-01 6.57215774e-01 -2.44633351e-02 -4.43235606e-01
-9.16521966e-01 -1.26809359e+00 1.07232206e-01 5.36683559e-01
4.06454414e-01 -4.47696567e-01 -5.52961409e-01 -3.97083908e-01
-4.48230207e-02 1.11576825e-01 -3.12863290e-01 9.90757719e-02
-6.65072262e-01 -5.16009271e-01 1.22189179e-01 6.31322205e-01
7.14159608e-01 -6.75888777e-01 -1.12296450e+00 1.88494742e-01
-8.66547763e-01 -1.43553555e+00 -3.60725403e-01 -2.97096729e-01
-9.89893615e-01 -1.11968946e+00 -1.10831261e+00 -5.30988276e-01
4.32848006e-01 2.35565051e-01 9.92218256e-01 -1.83166772e-01
-4.84001279e-01 6.76060855e-01 -2.21664578e-01 -1.96786955e-01
-7.66830593e-02 -4.49973196e-01 3.72126430e-01 -5.71568199e-02
9.42151155e-03 -2.98593909e-01 -8.10276091e-01 5.36754847e-01
-2.88394243e-01 1.11098282e-01 5.66905022e-01 8.66278112e-01
7.82651842e-01 1.03710964e-01 -1.96769699e-01 -2.02552140e-01
-1.70121849e-01 -9.45251212e-02 -4.73775715e-01 3.07074368e-01
3.60481590e-02 4.43167612e-03 -9.14736986e-02 -3.34598809e-01
-1.18604851e+00 5.11045516e-01 1.37071624e-01 -4.10697162e-01
-2.96371043e-01 2.32258216e-01 -4.13322151e-01 1.08003430e-01
2.92577118e-01 -7.25831166e-02 1.05738021e-01 -4.19455945e-01
2.23434269e-01 1.26471490e-01 5.36879420e-01 -6.08524323e-01
6.10712826e-01 6.30954683e-01 4.53317970e-01 -9.27376986e-01
-5.79003334e-01 -8.07168543e-01 -1.02013397e+00 -7.01708078e-01
1.26711798e+00 -1.25817895e+00 -7.36388087e-01 6.07188225e-01
-1.28953266e+00 2.06446663e-01 3.37041467e-01 9.93553996e-01
-6.10921144e-01 9.36014354e-01 -7.05720663e-01 -9.42728698e-01
-3.32681805e-01 -1.32285857e+00 1.39086902e+00 5.61614484e-02
-4.59691972e-01 -6.93141520e-01 1.56013072e-01 7.36764669e-01
-3.56672853e-01 6.76526964e-01 4.70884144e-01 -1.47247046e-01
-4.72865582e-01 -4.61205870e-01 3.63893479e-01 1.13174193e-01
-1.12887330e-01 -2.29549274e-01 -3.89509469e-01 -4.32219863e-01
-2.13058442e-01 -2.67529070e-01 4.87346411e-01 8.05913210e-01
2.34693184e-01 3.37066054e-02 -4.81177926e-01 9.40802619e-02
1.31158137e+00 -2.28760079e-01 3.91896605e-01 2.32097298e-01
9.10029590e-01 7.85364985e-01 8.85770917e-01 7.62767792e-01
5.37082732e-01 1.12890840e+00 5.76086581e-01 2.52037853e-01
-1.89194679e-02 -1.49381459e-01 3.73709172e-01 7.02952325e-01
-6.61942720e-01 -1.03198230e-01 -8.51014972e-01 3.63212854e-01
-2.13077497e+00 -1.08056402e+00 -3.41158003e-01 2.37855315e+00
2.42509618e-01 1.26339495e-01 2.71765798e-01 1.37508601e-01
9.55522120e-01 -7.28424415e-02 -2.44627848e-01 2.64151484e-01
1.83366403e-01 -1.92489997e-01 4.21906412e-01 3.49652529e-01
-1.14893305e+00 4.34056848e-01 5.27209187e+00 4.30607319e-01
-3.69119704e-01 8.76069590e-02 8.28919336e-02 2.65745185e-02
4.62710083e-01 -1.53651059e-01 -9.65234935e-01 2.56163895e-01
2.89027929e-01 4.96177197e-01 -1.39543355e-01 6.57533407e-01
4.40798141e-02 -5.17144680e-01 -8.80770206e-01 1.06231058e+00
3.40185016e-01 -5.50542176e-01 -8.46624970e-02 1.11023754e-01
5.78058839e-01 -2.72032470e-01 -3.13831210e-01 -1.90944634e-02
3.82588804e-02 -5.24719715e-01 1.06757700e+00 7.06331909e-01
3.85859869e-02 -7.88659036e-01 7.50332415e-01 7.45800555e-01
-1.51208282e+00 1.62530750e-01 -2.47116879e-01 5.34180515e-02
6.03781283e-01 5.89285076e-01 -5.46242416e-01 7.81405807e-01
9.79464650e-01 3.61790925e-01 -3.45298827e-01 1.08884180e+00
-3.32261235e-01 1.84686661e-01 -5.22386730e-01 2.51985341e-01
-2.26016670e-01 -1.14901997e-01 8.71231556e-01 8.57244730e-01
3.42136949e-01 8.56641605e-02 6.38402820e-01 4.41919118e-01
4.56483722e-01 -6.32631630e-02 -3.48844022e-01 5.92463970e-01
3.68118435e-02 1.34046257e+00 -7.78425813e-01 -3.17511320e-01
-3.64133090e-01 1.32012451e+00 1.22788697e-01 1.78509146e-01
-9.97603536e-01 2.26404458e-01 9.46632251e-02 1.10052444e-01
4.56200898e-01 -6.19189918e-01 3.74654323e-01 -1.28185821e+00
3.36583823e-01 -8.32419515e-01 5.38413107e-01 -7.44563043e-01
-1.05480087e+00 1.62728772e-01 4.75099236e-01 -1.26300538e+00
-4.85022038e-01 -4.80243564e-01 -9.99308527e-02 6.49810851e-01
-6.82293236e-01 -1.27436090e+00 -3.49561900e-01 5.93982458e-01
7.12488592e-01 1.79206431e-01 7.07386732e-01 2.75885284e-01
-4.03413922e-01 2.30481863e-01 -3.47505808e-01 8.31950083e-02
9.39632535e-01 -1.15722668e+00 1.56294361e-01 7.68535316e-01
9.20299441e-02 4.38214093e-01 9.76872325e-01 -9.22371030e-01
-1.39976728e+00 -6.44955516e-01 7.90135443e-01 -4.94932562e-01
-4.77737077e-02 -1.25332505e-01 -3.67763162e-01 7.95950294e-01
6.34530187e-03 1.40554503e-01 6.08524323e-01 -1.32250339e-01
-8.42089672e-03 2.60211498e-01 -1.10030532e+00 3.48600596e-01
8.17871809e-01 2.23463610e-01 -7.56181777e-01 1.99131683e-01
1.41538322e-01 -7.06358969e-01 -9.11478758e-01 6.85289323e-01
1.00161707e+00 -9.28798378e-01 1.33134723e+00 -1.79023594e-01
4.00692336e-02 -5.38910091e-01 -3.74046743e-01 -1.04064906e+00
-3.60235155e-01 -9.85110030e-02 -2.73954719e-01 9.23671722e-01
-1.49213627e-01 -1.48794288e-03 7.64645875e-01 5.90582311e-01
2.63597041e-01 -3.18356335e-01 -1.09712791e+00 -6.78904831e-01
-4.85599220e-01 3.74362394e-02 -9.65141430e-02 4.75225002e-01
-1.85778216e-01 4.44377333e-01 -8.52463961e-01 4.98582512e-01
1.35588324e+00 1.43873826e-01 1.05470514e+00 -1.36381400e+00
-7.10316598e-01 2.76079088e-01 -8.35436642e-01 -1.05945408e+00
-1.46800816e-01 -2.88581431e-01 2.71328509e-01 -1.61036932e+00
5.13230979e-01 2.99724758e-01 1.98003426e-01 1.95396468e-02
-3.11946988e-01 1.04647689e-01 4.22021121e-01 2.27051750e-01
-8.51102471e-01 5.46745181e-01 1.18702781e+00 -9.45653766e-02
2.82376595e-02 1.69915706e-01 3.52350742e-01 1.16648543e+00
5.00407040e-01 -4.49082524e-01 -4.94475588e-02 -2.47046098e-01
5.00071757e-02 6.03945136e-01 9.02264595e-01 -1.36080098e+00
2.50145167e-01 7.62972161e-02 8.94642651e-01 -1.11969244e+00
9.13459420e-01 -7.60299444e-01 5.57965696e-01 7.69380510e-01
1.73152015e-01 2.46009618e-01 -7.76782706e-02 9.92653370e-01
-1.93238348e-01 -3.22895348e-02 6.67531312e-01 -6.59737349e-01
-7.28714585e-01 1.07210733e-01 -3.25654417e-01 -1.59644932e-01
8.22420180e-01 -1.98672056e-01 2.89594054e-01 -3.96514177e-01
-1.08549023e+00 2.13025481e-01 2.08000034e-01 3.65145952e-01
6.74668670e-01 -1.42793453e+00 -6.17995620e-01 -1.96710184e-01
4.16884348e-02 1.45714775e-01 5.25737226e-01 1.10266554e+00
-4.18273538e-01 3.13640654e-01 -4.43868250e-01 -1.01845944e+00
-1.68581963e+00 2.90257096e-01 2.30045140e-01 -4.68209893e-01
-5.67250073e-01 5.52122235e-01 1.04137771e-01 -4.24591959e-01
2.87339449e-01 4.09945520e-03 -1.13693655e-01 -1.53412357e-01
2.57753402e-01 8.43586504e-01 -1.68069750e-01 -1.28183734e+00
-5.93933940e-01 1.00922167e+00 2.78407991e-01 -4.01467711e-01
8.80043745e-01 -3.67131323e-01 1.01924852e-01 4.34544086e-01
9.77417886e-01 -4.27499190e-02 -1.14667308e+00 -1.32988766e-01
-4.90862042e-01 -4.48741794e-01 -4.31721091e-01 -5.80078483e-01
-7.48650432e-01 7.54173934e-01 8.26233149e-01 -5.29223263e-01
7.13722110e-01 -7.58015737e-02 6.86589301e-01 2.32384518e-01
7.26936877e-01 -1.21526313e+00 1.73912466e-01 1.53057352e-01
8.95305932e-01 -1.35768425e+00 5.10517776e-01 -6.14809752e-01
-6.25934243e-01 1.09735513e+00 6.97087824e-01 -1.54693993e-02
5.29290795e-01 -1.44904733e-01 -1.47877678e-01 -1.80392832e-01
-2.57728994e-01 -1.02084734e-01 5.68989277e-01 3.89496416e-01
3.11592817e-01 1.29349634e-01 -5.14181077e-01 4.48465705e-01
1.76344097e-01 -1.44552380e-01 9.30024832e-02 1.11656129e+00
-5.19150436e-01 -9.74994838e-01 -9.62501049e-01 -1.71935648e-01
-6.31935537e-01 5.31241298e-01 -1.73466697e-01 1.03950942e+00
3.00252914e-01 9.60616827e-01 -3.40182573e-01 -1.65512085e-01
4.89632875e-01 9.68815684e-02 1.04206336e+00 -2.88423449e-01
-1.76618874e-01 5.39087832e-01 -2.31586602e-02 -7.01972067e-01
-8.28704655e-01 -9.08971488e-01 -1.29967213e+00 1.06745787e-01
-3.98559272e-01 -5.72926253e-02 6.18453503e-01 9.93650019e-01
1.72444824e-02 2.07416356e-01 -8.07174966e-02 -1.45612252e+00
-5.90663612e-01 -9.53363955e-01 -5.01059413e-01 6.70982301e-01
1.88120566e-02 -1.20900309e+00 -2.18256295e-01 2.97211677e-01] | [7.035489559173584, -1.0003266334533691] |
2992d0e9-08dd-4c93-aa0d-76eb000eecde | adaptive-estimators-show-information | 1902.09037 | null | https://arxiv.org/abs/1902.09037v2 | https://arxiv.org/pdf/1902.09037v2.pdf | Adaptive Estimators Show Information Compression in Deep Neural Networks | To improve how neural networks function it is crucial to understand their learning process. The information bottleneck theory of deep learning proposes that neural networks achieve good generalization by compressing their representations to disregard information that is not relevant to the task. However, empirical evidence for this theory is conflicting, as compression was only observed when networks used saturating activation functions. In contrast, networks with non-saturating activation functions achieved comparable levels of task performance but did not show compression. In this paper we developed more robust mutual information estimation techniques, that adapt to hidden activity of neural networks and produce more sensitive measurements of activations from all functions, especially unbounded functions. Using these adaptive estimation techniques, we explored compression in networks with a range of different activation functions. With two improved methods of estimation, firstly, we show that saturation of the activation function is not required for compression, and the amount of compression varies between different activation functions. We also find that there is a large amount of variation in compression between different network initializations. Secondary, we see that L2 regularization leads to significantly increased compression, while preventing overfitting. Finally, we show that only compression of the last layer is positively correlated with generalization. | ["Cian O'Donnell", 'Ivan Chelombiev', 'Conor Houghton'] | 2019-02-24 | adaptive-estimators-show-information-1 | https://openreview.net/forum?id=SkeZisA5t7 | https://openreview.net/pdf?id=SkeZisA5t7 | iclr-2019-5 | ['mutual-information-estimation', 'l2-regularization'] | ['methodology', 'methodology'] | [ 3.25799406e-01 1.30631268e-01 1.04459755e-01 -4.86433953e-01
-2.12976903e-01 -5.30282259e-01 4.84771192e-01 2.37396374e-01
-9.68453586e-01 7.87925124e-01 1.54992312e-01 -3.80460292e-01
-5.84668517e-01 -6.69353604e-01 -7.18597651e-01 -7.37666905e-01
-2.91085809e-01 3.04018766e-01 2.28818133e-01 -8.99378285e-02
3.73299956e-01 6.76331699e-01 -1.66666412e+00 5.15679717e-01
4.89927471e-01 8.53120685e-01 2.75086254e-01 8.15691650e-01
-5.80362827e-02 8.05714667e-01 -8.77699852e-01 -2.05156147e-01
4.02606547e-01 -6.88485205e-01 -1.04448032e+00 -3.15734088e-01
4.89579141e-01 -1.68020308e-01 -5.07813990e-01 9.33949232e-01
4.27301228e-01 3.19298565e-01 6.87508047e-01 -8.77097070e-01
-3.08239967e-01 9.23580706e-01 -1.77544251e-01 6.49052799e-01
-2.95749336e-01 9.53771845e-02 9.85202730e-01 -3.93178284e-01
2.68951088e-01 1.14283395e+00 7.25749075e-01 7.34978795e-01
-1.52317142e+00 -6.14912033e-01 -3.19162995e-04 -6.53817728e-02
-1.19686723e+00 -7.02057242e-01 4.17872012e-01 -1.72413886e-01
1.13096166e+00 4.32956189e-01 7.78234839e-01 7.07293034e-01
1.04600139e-01 6.15655303e-01 8.91232431e-01 -5.12425303e-01
2.03929245e-01 2.29339063e-01 1.88768536e-01 6.20621979e-01
4.87083286e-01 -3.39261349e-03 -4.36287612e-01 9.94605348e-02
8.33425522e-01 -1.36103183e-02 -5.52625597e-01 -1.60176054e-01
-1.06479871e+00 9.51725721e-01 4.24882084e-01 7.80480266e-01
-1.61468372e-01 3.96560490e-01 4.75625575e-01 8.44528854e-01
3.10498357e-01 9.56970572e-01 -6.83655143e-01 -2.67729938e-01
-1.03884184e+00 -3.22116688e-02 1.02714920e+00 6.31442308e-01
7.00077355e-01 2.41785929e-01 5.83720468e-02 9.42297399e-01
-6.81374595e-02 -6.97001144e-02 8.37813854e-01 -1.35518575e+00
3.99066776e-01 3.50835711e-01 -4.93134111e-01 -1.01857615e+00
-6.81125283e-01 -9.01062548e-01 -9.28416193e-01 1.68223351e-01
8.34533811e-01 -2.57269442e-01 -7.17217803e-01 2.21574092e+00
-4.49616134e-01 -3.72675508e-01 1.96231574e-01 5.94997644e-01
3.72760326e-01 3.54275584e-01 -1.58149093e-01 -3.57439131e-01
8.75539899e-01 -4.89372283e-01 -6.26802921e-01 -4.00242537e-01
1.06987250e+00 -4.43106085e-01 9.70260680e-01 4.77622867e-01
-1.52824461e+00 -3.85136932e-01 -1.32090461e+00 1.29512146e-01
-4.42027509e-01 -3.69365245e-01 6.10647142e-01 6.04184568e-01
-1.43659019e+00 1.28625274e+00 -7.64186859e-01 -2.59024709e-01
5.67257881e-01 7.20573783e-01 -4.40569133e-01 1.09789163e-01
-1.20214498e+00 1.13450253e+00 9.37264442e-01 -1.00223877e-01
-3.99220228e-01 -5.78439295e-01 -6.34328067e-01 5.32865524e-01
-2.43586879e-02 -5.78941822e-01 1.02858865e+00 -1.38301015e+00
-1.06948495e+00 5.75447738e-01 -2.54425462e-02 -7.93610156e-01
3.33522648e-01 -7.40443394e-02 5.87225147e-02 1.66561902e-01
-6.28837466e-01 1.02580345e+00 5.90102434e-01 -1.07421756e+00
-2.64466017e-01 -4.58669424e-01 -1.74869806e-01 3.16409975e-01
-7.37738490e-01 -3.50455850e-01 -1.19938634e-01 -2.83171147e-01
4.60548550e-01 -7.28526056e-01 -4.81787324e-02 -1.17879391e-01
1.05659105e-01 -1.04890995e-01 5.19926608e-01 -4.49968457e-01
1.22238934e+00 -2.08614349e+00 1.23950079e-01 4.73884046e-01
5.81445992e-01 2.14269906e-01 -1.86484411e-01 1.00011833e-01
-3.15528542e-01 4.52063322e-01 -3.37638170e-01 -1.24065705e-01
-1.09449275e-01 5.31987369e-01 6.65119886e-02 4.91551280e-01
4.16615568e-02 7.15460718e-01 -6.61709964e-01 -2.91000634e-01
5.49019128e-03 5.60708165e-01 -8.26654375e-01 -2.16769308e-01
1.61004305e-01 -1.58807710e-01 2.64201313e-01 8.06021038e-03
3.60592753e-01 -3.54070336e-01 3.27647030e-01 -6.05430193e-02
9.21479799e-03 4.11748677e-01 -8.95019412e-01 1.46919477e+00
-2.40734726e-01 1.36041546e+00 -3.52507532e-02 -1.41339087e+00
8.52951884e-01 2.32924134e-01 5.17068446e-01 -8.13104808e-01
2.67091066e-01 1.62316829e-01 8.23776424e-01 -1.76019117e-01
1.56429052e-01 -9.83795002e-02 5.18406630e-01 5.83942890e-01
3.52084368e-01 4.34875004e-02 4.01235580e-01 1.70417234e-01
1.44641709e+00 -4.83765543e-01 -2.17209989e-03 -4.62224126e-01
1.52737617e-01 -4.03676152e-01 2.24765316e-01 9.99297082e-01
-1.08557686e-01 5.87436438e-01 8.11213672e-01 -3.77374619e-01
-1.34400678e+00 -6.91728115e-01 -3.55127782e-01 1.19504976e+00
-3.87611449e-01 -4.67900842e-01 -9.41851497e-01 -3.62044901e-01
-1.22944146e-01 7.96212375e-01 -7.45426238e-01 -6.21309936e-01
-4.58259255e-01 -9.49467421e-01 7.28316307e-01 4.17150766e-01
5.78139186e-01 -1.02437639e+00 -8.20667148e-01 2.87961233e-02
1.51566803e-01 -7.01369345e-01 -1.13569528e-01 7.63475776e-01
-1.50172365e+00 -8.58395457e-01 -6.60894454e-01 -7.00447798e-01
7.89226294e-01 1.21678144e-01 1.18834209e+00 4.31715250e-01
-2.55048405e-02 2.21680328e-01 3.97481062e-02 -2.45767206e-01
-4.90127623e-01 4.08018589e-01 -2.04568207e-02 -6.91528678e-01
3.01127106e-01 -8.13422978e-01 -6.12775743e-01 2.70210743e-01
-1.14965463e+00 -9.66337845e-02 9.61005032e-01 8.09113324e-01
7.82336518e-02 1.76960200e-01 3.43973488e-01 -8.53073955e-01
1.01112449e+00 -3.85493666e-01 -2.36320436e-01 -7.13044256e-02
-9.54582751e-01 8.49008441e-01 5.75062037e-01 -5.46720266e-01
-6.21619642e-01 -1.44544736e-01 -1.34223938e-01 -3.19532245e-01
-8.52473429e-04 3.62511992e-01 2.31829956e-01 -7.30639398e-02
1.10073113e+00 2.08512723e-01 4.27282304e-01 -3.30420822e-01
9.12355855e-02 2.54600644e-01 2.10142240e-01 -3.02394181e-01
1.66553706e-01 7.78511465e-02 -2.57070735e-02 -8.53467286e-01
-5.92964113e-01 -2.58796722e-01 -5.08930862e-01 -9.18498784e-02
4.81308728e-01 -3.92662853e-01 -9.53419089e-01 1.20142579e-01
-1.15712011e+00 -5.78132033e-01 -5.07145226e-01 7.10295975e-01
-5.77732265e-01 8.07495937e-02 -7.39477873e-01 -7.51783311e-01
-2.40811020e-01 -1.08293724e+00 2.68878460e-01 -1.00724749e-01
-3.11838716e-01 -1.24512017e+00 -7.45068789e-02 -2.40547657e-01
6.67897046e-01 -2.56911039e-01 9.77935791e-01 -9.83680427e-01
-9.49783698e-02 -4.22103256e-02 -3.12087655e-01 7.43951261e-01
1.57063007e-02 -2.47272864e-01 -9.49866235e-01 -3.95488769e-01
2.13718042e-01 -3.14874649e-01 1.29737353e+00 5.34915686e-01
1.41651237e+00 -7.63592184e-01 -7.70036057e-02 7.70149589e-01
1.25663030e+00 1.46652251e-01 9.60104406e-01 3.57290506e-01
3.57368112e-01 7.42739081e-01 -3.40930909e-01 9.98533219e-02
-3.23408872e-01 2.28790477e-01 3.26205283e-01 -3.21660563e-02
4.33721095e-02 9.01912823e-02 2.11046055e-01 1.03675532e+00
-2.22670212e-01 -1.90372482e-01 -7.56974876e-01 2.83521682e-01
-1.57271600e+00 -1.19815814e+00 8.93681776e-03 2.45775199e+00
1.18489528e+00 6.64615929e-01 -2.36579287e-03 4.38399553e-01
4.11443025e-01 -3.98888662e-02 -5.81768036e-01 -5.30066431e-01
-1.72140092e-01 2.34762177e-01 9.05169666e-01 6.75974786e-01
-5.41209519e-01 5.60934067e-01 7.61125708e+00 6.83578432e-01
-8.97317231e-01 -4.37520966e-02 8.04119885e-01 -4.36704755e-01
-2.73284853e-01 -1.40441477e-01 -4.50335056e-01 4.69194978e-01
1.38360107e+00 9.21881273e-02 6.25826955e-01 5.95999300e-01
-1.41149372e-01 -1.02463171e-01 -1.14717841e+00 8.61754060e-01
-1.32304683e-01 -1.23640978e+00 -1.82496279e-01 1.54292986e-01
5.98311007e-01 1.70370221e-01 1.65386256e-02 2.59343147e-01
1.56082407e-01 -1.12086713e+00 2.33872026e-01 5.62119842e-01
3.66403103e-01 -1.00022578e+00 7.38163173e-01 5.19710183e-01
-6.68964565e-01 -1.35862023e-01 -7.47881532e-01 -2.24377826e-01
-5.09700000e-01 5.39961457e-01 -8.40658069e-01 -2.20373571e-01
4.72957373e-01 3.48849148e-01 -6.83410823e-01 9.84578788e-01
1.29092142e-01 6.76406205e-01 -5.70638835e-01 -1.58931613e-01
1.48563325e-01 -1.22914780e-02 2.05325007e-01 1.39922822e+00
2.55540222e-01 -5.16882055e-02 -5.65071762e-01 8.37927341e-01
-1.69836715e-01 -5.55163100e-02 -8.23341191e-01 -1.20855756e-01
4.27589893e-01 8.39744806e-01 -7.36303926e-01 -3.16997439e-01
-2.19008565e-01 7.98195601e-01 5.45977235e-01 4.10951853e-01
-4.55017835e-01 -5.18165648e-01 4.06899691e-01 1.49543688e-01
1.16407819e-01 -4.17107284e-01 -4.29894030e-01 -8.25523019e-01
-2.57673472e-01 -7.16692567e-01 2.65720725e-01 -4.26893532e-01
-7.88496137e-01 4.98615354e-01 7.95495212e-02 -5.50150275e-01
-4.44705993e-01 -6.41328156e-01 -2.90481746e-01 8.68290603e-01
-8.92454684e-01 -6.36930694e-04 -9.83886942e-02 3.71712685e-01
4.39642280e-01 -1.56292439e-01 7.18943417e-01 2.88642794e-01
-3.38046253e-01 7.82887995e-01 3.14718455e-01 1.95154876e-01
3.71820390e-01 -1.17730820e+00 1.29827082e-01 4.27913874e-01
3.56731236e-01 9.98866975e-01 8.45684826e-01 -3.03323984e-01
-1.08381844e+00 -6.56199396e-01 6.23375952e-01 -1.45017922e-01
2.80934602e-01 -2.24558800e-01 -1.32404137e+00 7.03989148e-01
2.52354950e-01 -2.84476310e-01 6.54058993e-01 3.00411493e-01
-2.30051696e-01 -5.64489551e-02 -9.52090740e-01 5.49519360e-01
1.05705941e+00 -3.06689352e-01 -4.52645332e-01 3.24357420e-01
6.25476778e-01 1.88073516e-02 -8.05989325e-01 2.87640125e-01
7.20283329e-01 -1.22227561e+00 8.28470170e-01 -7.59686768e-01
3.09774727e-01 3.34548444e-01 -1.53658554e-01 -1.27011204e+00
-4.62925255e-01 -2.67346501e-01 -3.32328826e-01 7.11933494e-01
8.03112149e-01 -8.24990630e-01 9.99089122e-01 8.52128088e-01
-9.84918773e-02 -8.31793010e-01 -7.04801679e-01 -8.14378679e-01
2.85852760e-01 -3.94163519e-01 1.94602106e-02 1.01126790e+00
3.34055901e-01 4.43606436e-01 7.12180743e-03 -4.95044857e-01
4.09035772e-01 -6.82888508e-01 2.26341054e-01 -1.29070246e+00
-4.37898993e-01 -1.03828335e+00 -5.79672873e-01 -1.05051661e+00
9.17046815e-02 -9.74816203e-01 -8.78166184e-02 -1.21118093e+00
1.93625852e-01 -2.84792274e-01 -4.23161060e-01 4.78286862e-01
1.36404693e-01 1.71389267e-01 1.82626560e-01 4.51800764e-01
-3.09809864e-01 1.58707276e-01 1.11775804e+00 1.04530141e-01
-3.38232368e-01 3.56314518e-02 -8.01491499e-01 6.94876492e-01
1.19698513e+00 -4.54590499e-01 -7.24637568e-01 -6.49282336e-01
4.72896934e-01 -2.38345027e-01 8.84527564e-02 -1.39030516e+00
3.72863621e-01 2.41085589e-01 8.27399254e-01 -2.60492057e-01
3.27061623e-01 -7.04624474e-01 -1.06330767e-01 7.77040005e-01
-1.05288565e+00 2.99268961e-01 4.88234401e-01 2.92889655e-01
1.22742362e-01 -8.11351538e-01 9.58655238e-01 -5.06244361e-01
-2.20768899e-01 -2.61906331e-04 -4.70750391e-01 1.89225256e-01
5.20821929e-01 -2.66207606e-01 -1.86337546e-01 -4.46578175e-01
-7.12570190e-01 -1.13072813e-01 3.79004508e-01 -1.45743191e-01
7.33651519e-01 -1.18905151e+00 -7.05915391e-01 2.86885321e-01
-5.98395884e-01 -7.63119012e-02 -1.81724533e-01 9.21293199e-01
-5.43523848e-01 5.37806213e-01 -2.75907934e-01 -5.85365236e-01
-1.09516490e+00 3.87764424e-01 5.70144951e-01 -4.00456600e-02
-2.87966698e-01 9.70882714e-01 5.77386133e-02 -6.60730302e-02
4.78369147e-01 -3.36060286e-01 -1.33433342e-01 2.15476498e-01
5.47012806e-01 3.79604548e-01 2.87434965e-01 -1.37568176e-01
-4.00510132e-02 1.53646514e-01 -4.22091842e-01 -4.63819236e-01
1.19377124e+00 1.84981212e-01 5.39845899e-02 6.32184625e-01
1.68259823e+00 -5.51323414e-01 -1.41130054e+00 -1.18279763e-01
6.66525215e-02 -3.05873156e-01 3.21805358e-01 -4.30060685e-01
-1.37064147e+00 9.29430842e-01 6.30422592e-01 6.85376525e-01
1.08560038e+00 -2.13416994e-01 3.25278044e-01 8.57192099e-01
4.89469767e-02 -1.20810997e+00 1.60332277e-01 6.50256753e-01
6.65198386e-01 -1.05271900e+00 1.76590964e-01 2.47728020e-01
-3.20314556e-01 1.33602715e+00 5.57536066e-01 -2.60608494e-02
3.54108840e-01 2.73851633e-01 -5.05629480e-01 -2.74994224e-01
-1.04015815e+00 -9.38292071e-02 1.81435257e-01 4.03600216e-01
6.77904785e-01 -3.94225568e-01 -2.41421953e-01 1.43082896e-02
-4.48388517e-01 -3.30250889e-01 3.07399958e-01 9.53384578e-01
-8.71398747e-01 -7.24637210e-01 -1.05281912e-01 8.64418507e-01
-4.23369408e-01 -4.83252406e-01 -3.95251334e-01 1.04036343e+00
-2.06313446e-01 5.51860571e-01 3.92335743e-01 -3.32044125e-01
1.79694628e-03 4.08736914e-01 8.22231650e-01 -3.79837781e-01
-4.60411221e-01 -2.19511792e-01 -1.61761850e-01 -4.64073479e-01
-2.00836033e-01 -5.55975795e-01 -1.23155928e+00 -6.66743696e-01
-5.61105847e-01 3.40281010e-01 8.84391725e-01 9.12496686e-01
1.02531828e-01 7.45472074e-01 3.02862287e-01 -7.05228508e-01
-7.22486854e-01 -1.17526150e+00 -5.15372932e-01 4.43042219e-01
4.39664036e-01 -4.51393038e-01 -8.09947252e-01 -3.89024876e-02] | [8.22033977508545, 3.4473588466644287] |
0d22f2ca-a496-4523-b8ca-b364f7e230d7 | topological-deep-learning-a-review-of-an | 2302.03836 | null | https://arxiv.org/abs/2302.03836v1 | https://arxiv.org/pdf/2302.03836v1.pdf | Topological Deep Learning: A Review of an Emerging Paradigm | Topological data analysis (TDA) provides insight into data shape. The summaries obtained by these methods are principled global descriptions of multi-dimensional data whilst exhibiting stable properties such as robustness to deformation and noise. Such properties are desirable in deep learning pipelines but they are typically obtained using non-TDA strategies. This is partly caused by the difficulty of combining TDA constructs (e.g. barcode and persistence diagrams) with current deep learning algorithms. Fortunately, we are now witnessing a growth of deep learning applications embracing topologically-guided components. In this survey, we review the nascent field of topological deep learning by first revisiting the core concepts of TDA. We then explore how the use of TDA techniques has evolved over time to support deep learning frameworks, and how they can be integrated into different aspects of deep learning. Furthermore, we touch on TDA usage for analyzing existing deep models; deep topological analytics. Finally, we discuss the challenges and future prospects of topological deep learning. | ['Lars Petersson', 'Vivien Rolland', 'Zeeshan Hayder', 'James Nichols', 'Abdelwahed Khamis', 'Ali Zia'] | 2023-02-08 | null | null | null | null | ['topological-data-analysis'] | ['graphs'] | [-6.41759932e-01 1.73297487e-02 1.70278717e-02 -2.17507124e-01
-3.17765743e-01 -8.03719819e-01 9.89815474e-01 4.17531669e-01
1.40662074e-01 3.63886327e-01 2.90785968e-01 -4.88151371e-01
-6.74634635e-01 -1.05168378e+00 -8.53193462e-01 -7.19982445e-01
-1.00748825e+00 6.66250288e-01 3.30187500e-01 -4.06466961e-01
4.71939206e-01 9.41735506e-01 -1.48865902e+00 4.04771775e-01
6.24743521e-01 1.11006904e+00 -2.97621340e-01 7.34618962e-01
-3.36329579e-01 6.21535301e-01 -1.69611871e-01 -4.36475687e-02
2.35955715e-01 1.77730799e-01 -1.15444911e+00 -5.34814298e-01
8.41889679e-01 -3.66403103e-01 -6.54190898e-01 4.65169817e-01
3.00772846e-01 7.06735700e-02 9.44457114e-01 -1.21362317e+00
-5.95021009e-01 2.94937998e-01 -3.55304599e-01 3.57712626e-01
7.24850874e-03 4.94169593e-02 1.19706428e+00 -9.06447828e-01
8.50891292e-01 1.43034625e+00 1.21851587e+00 1.55533835e-01
-1.41741002e+00 2.06113961e-02 -6.83789421e-03 2.20872328e-01
-8.45613778e-01 -4.38686967e-01 1.05334234e+00 -7.82149971e-01
9.99883771e-01 1.02948666e-01 8.29379976e-01 9.25920129e-01
5.87911427e-01 9.28193629e-01 1.06816947e+00 -1.48287937e-01
2.52873689e-01 -6.26048148e-01 1.51090577e-01 9.91396725e-01
1.51097536e-01 8.12326595e-02 -4.08545583e-01 6.91120028e-02
9.72968638e-01 -2.65818864e-01 4.10630256e-01 -1.23910558e+00
-1.25979149e+00 6.11346662e-01 7.39837050e-01 2.90419519e-01
2.58215591e-02 4.79419887e-01 8.65695834e-01 4.07388479e-01
4.58051473e-01 5.59583068e-01 -5.85860550e-01 -2.60328382e-01
-9.18534577e-01 6.40888095e-01 5.31984806e-01 7.20018506e-01
7.48100460e-01 -7.35191852e-02 3.08927000e-01 7.13526785e-01
2.85703152e-01 1.50334612e-01 3.79036590e-02 -9.71331656e-01
1.31578535e-01 7.10125983e-01 -2.42280260e-01 -1.45564604e+00
-9.27299440e-01 -3.22767854e-01 -8.70673478e-01 4.43502486e-01
6.30915046e-01 4.51823503e-01 -9.14961994e-01 1.50913477e+00
1.62823111e-01 -6.96185380e-02 -4.22255039e-01 5.48369288e-01
9.71876621e-01 3.18158001e-01 -3.09004694e-01 4.21192497e-01
8.99855077e-01 -6.14568651e-01 -2.63144910e-01 3.05019349e-01
1.10477400e+00 -4.34311092e-01 1.04716456e+00 3.43090028e-01
-1.10663831e+00 -3.75230521e-01 -1.02077508e+00 -6.68407500e-01
-1.03949451e+00 -1.43783092e-01 9.13927972e-01 3.20646793e-01
-1.51627922e+00 1.00841391e+00 -1.28797352e+00 -6.29813790e-01
9.54650819e-01 4.33481872e-01 -4.34781313e-01 3.13835740e-01
-9.46405411e-01 8.42158854e-01 1.23683423e-01 -2.27866564e-02
-8.24588418e-01 -9.15014923e-01 -6.86283648e-01 -9.76702198e-02
-1.94583401e-01 -7.20834970e-01 9.57018673e-01 -6.30318895e-02
-1.05300379e+00 1.04157341e+00 1.12743340e-01 -5.61236560e-01
8.10648143e-01 -1.36896431e-01 7.57531300e-02 1.24892980e-01
-8.30268860e-02 7.28543758e-01 3.01431388e-01 -1.10670435e+00
-9.40134376e-02 -3.61656964e-01 2.28640079e-01 -2.69237965e-01
-2.18894675e-01 -4.45786506e-01 1.42697379e-01 -4.16892141e-01
3.65868479e-01 -7.03856528e-01 -1.43491905e-02 5.73821604e-01
-3.10581684e-01 -4.12347347e-01 1.34514964e+00 -2.51399934e-01
9.61368024e-01 -1.84121478e+00 3.68083328e-01 1.40163153e-01
8.29907596e-01 2.75236696e-01 -4.09271121e-02 7.03130722e-01
-2.68031232e-04 3.97713244e-01 -1.21734731e-01 -3.04463655e-01
2.59783119e-01 2.10851450e-02 -4.70115095e-01 6.85075223e-01
2.07280159e-01 1.29397631e+00 -9.64795172e-01 -5.49113631e-01
5.75724244e-01 4.38441753e-01 -6.92374825e-01 -2.94849634e-01
-3.87231529e-01 3.86050731e-01 -2.66501248e-01 7.65158832e-01
8.43099713e-01 -3.06606323e-01 5.33051789e-02 -3.01602364e-01
-4.41492140e-01 6.75306857e-01 -6.33804977e-01 1.81784809e+00
-3.27118456e-01 1.24748445e+00 2.10589156e-01 -1.44775951e+00
1.01755607e+00 -2.44851366e-01 8.52281868e-01 -8.97455454e-01
1.49473608e-01 4.33654666e-01 1.93783920e-02 -3.34084183e-01
3.74434412e-01 8.51291418e-02 2.67496314e-02 5.59782445e-01
1.67714417e-01 1.78321116e-02 -4.57508117e-02 2.98699975e-01
1.17957187e+00 3.79347920e-01 -1.68687090e-01 -7.81377494e-01
1.73332557e-01 1.58907771e-01 9.33902711e-02 5.84898412e-01
-5.26946366e-01 3.85868728e-01 9.29109752e-01 -1.24570477e+00
-1.56417084e+00 -1.33225906e+00 -6.72385216e-01 1.01248741e+00
-5.68790734e-02 -6.35370433e-01 -4.73742783e-01 -4.39678311e-01
3.89271408e-01 -2.75429875e-01 -1.06008613e+00 4.41361144e-02
-9.09507215e-01 -4.97068137e-01 8.35790396e-01 6.90739810e-01
4.89623696e-01 -8.74175131e-01 -5.74560881e-01 5.38334064e-02
2.05420941e-01 -7.93847144e-01 3.43708634e-01 3.74764800e-02
-1.25020683e+00 -1.22076738e+00 -4.16728079e-01 -8.49462032e-01
2.63243139e-01 2.79201180e-01 1.18220270e+00 2.32038900e-01
-3.86585623e-01 3.59228641e-01 -1.99809093e-02 3.53171006e-02
-1.97068438e-01 4.50322866e-01 4.48904693e-01 -5.72218299e-01
9.38460082e-02 -1.02473354e+00 -7.65290082e-01 1.45298868e-01
-7.84796119e-01 1.31254166e-01 3.60876709e-01 4.84853834e-01
2.92147458e-01 -8.61789137e-02 6.09058261e-01 -4.40724462e-01
6.47676229e-01 -4.82863337e-01 -3.60959470e-01 1.16176240e-01
-3.50823462e-01 1.67589530e-01 7.12724805e-01 -2.22618822e-02
-4.57624018e-01 -4.07019615e-01 -8.83314610e-02 -3.09440136e-01
-2.09105074e-01 5.79605460e-01 1.30281910e-01 -4.20444667e-01
5.81849039e-01 5.64452435e-04 4.20653999e-01 -5.66576302e-01
6.29086375e-01 1.83144331e-01 4.23390269e-01 -1.02387750e+00
5.32773852e-01 8.57700467e-01 5.67782938e-01 -9.93741810e-01
-4.07925040e-01 -1.02504149e-01 -1.34612918e+00 -4.26079899e-01
4.98194993e-01 -3.20266038e-01 -8.64721417e-01 6.54554069e-01
-1.03209758e+00 -5.65050542e-01 -1.10229470e-01 -1.12570614e-01
-1.00653601e+00 3.78845453e-01 -6.45969808e-01 -4.05762583e-01
-1.06516987e-01 -1.21910179e+00 1.04166877e+00 -1.33708000e-01
-2.08486214e-01 -1.81907523e+00 4.21220630e-01 1.55587550e-02
6.42551363e-01 8.80314291e-01 1.27515447e+00 -4.77974683e-01
-7.81484663e-01 -9.59937721e-02 -3.74520689e-01 -1.51520297e-01
3.78019102e-02 4.34360683e-01 -8.98294449e-01 -3.57192457e-01
-6.19133770e-01 -4.26739246e-01 1.10501802e+00 6.64827824e-01
1.27042413e+00 -1.91319332e-01 -4.99507815e-01 9.73331392e-01
1.16108465e+00 -4.53938216e-01 3.50818425e-01 6.73327804e-01
8.55640113e-01 8.43389213e-01 2.57709980e-01 -1.51294665e-02
6.26186371e-01 6.71678364e-01 6.75753951e-01 -1.45953909e-01
-3.04204285e-01 -3.17046285e-01 -2.84464490e-02 1.03857803e+00
-4.90543758e-03 3.65291417e-01 -1.65093720e+00 7.15241134e-01
-1.98383594e+00 -9.51012850e-01 -5.11889696e-01 1.77987957e+00
4.26745474e-01 2.86458433e-01 5.32372177e-01 1.55797631e-01
2.86101252e-01 4.78620440e-01 -7.68636525e-01 -6.74999595e-01
-3.36864382e-01 5.20559121e-03 2.80183256e-01 4.39602405e-01
-1.34228969e+00 9.97300625e-01 7.30368567e+00 4.74313289e-01
-1.29887474e+00 -3.67659748e-01 5.22312164e-01 8.50449726e-02
-3.78198117e-01 1.52262021e-02 -1.78914607e-01 2.08343938e-01
8.54867935e-01 -2.91916542e-03 2.88582444e-01 6.40496790e-01
2.67932296e-01 2.57507533e-01 -1.24994922e+00 7.34896362e-01
-4.97168928e-01 -2.15340114e+00 6.63881004e-02 5.80965340e-01
5.16857862e-01 4.67193335e-01 4.32105988e-01 1.00159146e-01
2.35649973e-01 -1.16720855e+00 5.95694602e-01 6.67022407e-01
8.68948340e-01 -5.09884298e-01 3.94206494e-01 1.10200077e-01
-1.35385513e+00 -1.83562815e-01 -4.38929826e-01 -2.67522722e-01
-1.18065886e-01 5.99041581e-01 -5.47195196e-01 4.88231182e-01
8.58774245e-01 1.31790543e+00 -8.86096537e-01 1.19144201e+00
5.25204360e-01 2.20245644e-01 -4.25166190e-01 2.42410842e-02
5.44267237e-01 -2.34172955e-01 5.68594277e-01 1.42372406e+00
5.57184219e-02 -3.92465979e-01 -1.78896636e-01 1.11420214e+00
-8.79820958e-02 -1.13130763e-01 -1.18884182e+00 -1.40251324e-01
5.02136469e-01 1.11195564e+00 -9.88835454e-01 -5.77662252e-02
-2.09464476e-01 1.93917707e-01 7.22371817e-01 9.10613984e-02
-4.65782642e-01 -3.62033963e-01 9.00562465e-01 4.90511149e-01
7.85759613e-02 -8.38448584e-01 -8.34602237e-01 -1.02979445e+00
-4.16571014e-02 -5.71053565e-01 1.40196651e-01 -6.87133729e-01
-1.32152092e+00 3.25410962e-02 -4.67137806e-03 -1.00375783e+00
2.52535462e-01 -1.16320634e+00 -8.96691561e-01 3.71167898e-01
-1.23491108e+00 -1.37677801e+00 -1.03986099e-01 2.17865303e-01
2.30626225e-01 -1.81383416e-02 6.64228022e-01 1.05194584e-01
-3.46730739e-01 2.46470556e-01 5.24266720e-01 3.18941325e-01
5.12242973e-01 -1.59874058e+00 1.03860807e+00 3.64252776e-01
-1.14845127e-01 9.90246415e-01 5.57243705e-01 -4.72188354e-01
-1.57895577e+00 -7.91734695e-01 6.59834862e-01 -9.21341181e-01
1.09131110e+00 -8.64719510e-01 -1.13791764e+00 6.58806086e-01
-1.04325697e-01 2.45864570e-01 3.83336931e-01 5.60421109e-01
-6.28964961e-01 -6.01699352e-02 -1.08126569e+00 5.48892617e-01
1.22654617e+00 -7.69848406e-01 -4.81082112e-01 8.05231258e-02
4.47459996e-01 -3.67056787e-01 -1.00033510e+00 4.51447368e-01
1.01105535e+00 -1.33279037e+00 1.26797950e+00 -9.71951663e-01
4.53039348e-01 -2.39190459e-01 6.91123903e-02 -1.11858940e+00
-5.79147279e-01 -5.33204854e-01 -5.12612700e-01 8.08852375e-01
-1.02286953e-02 -4.59211230e-01 8.36181164e-01 3.10142487e-01
-4.30025667e-01 -1.10023606e+00 -1.09127462e+00 -8.28221977e-01
1.07379341e+00 -3.15505087e-01 6.40957415e-01 1.17813015e+00
3.43925476e-01 -1.10126741e-01 7.67921060e-02 -3.11495215e-01
7.13763297e-01 2.73876339e-02 8.30102921e-01 -1.82313120e+00
4.83578920e-01 -1.17813873e+00 -9.02236760e-01 -1.13852620e+00
6.72674477e-02 -1.12301707e+00 -4.20507640e-01 -1.78265285e+00
-2.12614592e-02 -6.35941982e-01 -2.03949496e-01 3.53640020e-01
5.46450973e-01 1.14865609e-01 -1.41161993e-01 4.01186436e-01
-7.23660350e-01 5.34847617e-01 1.31050503e+00 -2.90089790e-02
-1.08093940e-01 -3.05807561e-01 -3.69749993e-01 7.87002802e-01
9.93573725e-01 1.50800124e-01 -2.76972562e-01 -6.28741384e-01
7.25762427e-01 -3.44429195e-01 6.99635983e-01 -1.09021223e+00
3.66711646e-01 1.60870180e-02 4.22334224e-01 -9.60430682e-01
6.64432868e-02 -4.25491631e-01 -4.07563657e-01 2.46811822e-01
-4.18519646e-01 3.53689492e-01 4.35401589e-01 5.79973578e-01
5.58638535e-02 6.33293033e-01 9.03966904e-01 -1.31411627e-02
-4.54676956e-01 3.97766441e-01 -2.75651306e-01 8.24548863e-03
7.03081727e-01 -4.39239204e-01 -9.16679323e-01 -1.45808607e-01
-7.44188964e-01 3.36619586e-01 8.38041663e-01 4.71627027e-01
5.71710408e-01 -1.62176669e+00 -3.44437748e-01 1.12608097e-01
7.39044249e-02 2.42670596e-01 1.84422597e-01 1.10510862e+00
-1.02604949e+00 7.74093628e-01 -6.32768929e-01 -9.18215275e-01
-7.81914473e-01 4.56350833e-01 5.87961555e-01 -1.60535611e-02
-1.08046234e+00 5.59820414e-01 9.36126634e-02 -7.92727590e-01
2.76178360e-01 -5.95033109e-01 -9.39415023e-03 9.62586924e-02
9.74093825e-02 6.86087489e-01 3.18417668e-01 -6.03782713e-01
-7.13685513e-01 7.79071987e-01 -7.59288743e-02 1.69823036e-01
1.75471485e+00 -9.08995196e-02 -4.84134704e-01 6.23729825e-01
1.41382825e+00 -2.39034116e-01 -1.30564916e+00 -2.16432307e-02
2.92288423e-01 -5.31146154e-02 1.12455105e-02 -6.13213122e-01
-8.30623984e-01 1.52125013e+00 3.91609013e-01 7.58336663e-01
5.76841772e-01 2.59453088e-01 7.51906395e-01 4.52781409e-01
2.94630140e-01 -9.64221120e-01 3.92712742e-01 8.72233927e-01
9.47346210e-01 -9.32245135e-01 -9.96179879e-03 2.57363617e-01
1.68501303e-01 1.54515386e+00 4.41195041e-01 -5.58863819e-01
1.07626212e+00 1.12230219e-01 -6.43726215e-02 -7.54062116e-01
-8.58420670e-01 3.63454595e-02 3.75912666e-01 7.06340373e-01
6.32959127e-01 2.12573800e-02 4.20871042e-02 -1.36353791e-01
-5.90685248e-01 -4.43512946e-01 5.80014765e-01 8.64117563e-01
-7.65132368e-01 -8.94181550e-01 5.88015951e-02 4.56097573e-01
-5.25949448e-02 1.53272584e-01 -7.81559289e-01 8.94849598e-01
-6.57778084e-02 3.32485318e-01 3.52186829e-01 -4.80535358e-01
-2.02008076e-02 1.60547122e-02 5.00935793e-01 -2.20825315e-01
-7.98430387e-03 -2.55954027e-01 -1.63266003e-01 -7.44611144e-01
-2.35991448e-01 -7.35810220e-01 -1.13040555e+00 -1.05224431e+00
2.00524732e-01 -1.35778964e-01 7.66866624e-01 8.61271918e-01
6.40805125e-01 3.77148867e-01 2.91268915e-01 -1.03391397e+00
6.71944246e-02 -7.31987715e-01 -3.18824619e-01 6.13758303e-02
9.44659173e-01 -1.01327121e+00 -3.19741547e-01 -2.61408538e-01] | [6.773507118225098, 5.882699489593506] |
faa6fee9-5676-49c3-b19a-018e490860f2 | differential-viewpoints-for-ground-terrain | 2009.11072 | null | https://arxiv.org/abs/2009.11072v1 | https://arxiv.org/pdf/2009.11072v1.pdf | Differential Viewpoints for Ground Terrain Material Recognition | Computational surface modeling that underlies material recognition has transitioned from reflectance modeling using in-lab controlled radiometric measurements to image-based representations based on internet-mined single-view images captured in the scene. We take a middle-ground approach for material recognition that takes advantage of both rich radiometric cues and flexible image capture. A key concept is differential angular imaging, where small angular variations in image capture enables angular-gradient features for an enhanced appearance representation that improves recognition. We build a large-scale material database, Ground Terrain in Outdoor Scenes (GTOS) database, to support ground terrain recognition for applications such as autonomous driving and robot navigation. The database consists of over 30,000 images covering 40 classes of outdoor ground terrain under varying weather and lighting conditions. We develop a novel approach for material recognition called texture-encoded angular network (TEAN) that combines deep encoding pooling of RGB information and differential angular images for angular-gradient features to fully leverage this large dataset. With this novel network architecture, we extract characteristics of materials encoded in the angular and spatial gradients of their appearance. Our results show that TEAN achieves recognition performance that surpasses single view performance and standard (non-differential/large-angle sampling) multiview performance. | ['Ko Nishino', 'Jia Xue', 'Hang Zhang', 'Kristin J. Dana'] | 2020-09-22 | null | null | null | null | ['material-recognition'] | ['computer-vision'] | [ 7.21363783e-01 -3.32152873e-01 1.60500661e-01 -5.95260918e-01
-8.12860966e-01 -6.64533377e-01 3.02185655e-01 -3.23352456e-01
-7.76126087e-02 3.16072822e-01 -1.22767337e-01 -1.10752776e-01
-3.87145072e-01 -1.41436625e+00 -1.01568711e+00 -5.48830450e-01
1.17942113e-02 3.55033726e-01 1.80888936e-01 -5.17491102e-01
1.82067633e-01 9.41139638e-01 -1.93125081e+00 3.37537915e-01
4.73318338e-01 1.61706972e+00 2.24561974e-01 7.17422783e-01
1.43499494e-01 2.49429643e-01 -3.60145681e-02 -6.08084947e-02
8.13609660e-01 3.53186309e-01 -2.76591927e-01 1.53914154e-01
1.13842237e+00 -6.11352086e-01 -5.22380114e-01 5.30079007e-01
4.66222078e-01 -5.34528829e-02 7.08426237e-01 -8.08710098e-01
-6.11231089e-01 -1.22211888e-01 -5.42633295e-01 -1.33910388e-01
5.63642740e-01 1.86157048e-01 1.07767129e+00 -1.01672971e+00
5.26765764e-01 9.87562418e-01 8.94148052e-01 1.50554225e-01
-1.22706008e+00 -4.13769335e-01 3.08788437e-02 1.20351717e-01
-1.41349721e+00 -4.66362298e-01 9.19990838e-01 -3.91866326e-01
1.12017584e+00 4.56445932e-01 1.12950838e+00 1.03736830e+00
3.66691321e-01 5.78060806e-01 1.33705473e+00 -2.33028784e-01
7.03163221e-02 -2.37839535e-01 -3.40823889e-01 1.00387716e+00
5.58283865e-01 1.10523783e-01 -9.54855442e-01 8.46203938e-02
9.78379548e-01 2.70701230e-01 -3.22060585e-01 -7.06094801e-01
-1.09229982e+00 3.95362318e-01 6.76102817e-01 -6.67168856e-01
-4.72459257e-01 1.21883780e-01 -1.88642398e-01 3.98052871e-01
5.12952328e-01 4.26948398e-01 -5.04726291e-01 1.04644768e-01
-7.45454252e-01 1.15655534e-01 4.67129230e-01 8.88133764e-01
1.29171836e+00 3.26709658e-01 4.37006295e-01 1.01430321e+00
4.02424723e-01 1.47343647e+00 4.51809689e-02 -9.39909816e-01
4.69208807e-01 7.73565769e-01 1.41327810e-02 -1.19748175e+00
-5.16177475e-01 -9.00570676e-02 -4.93908346e-01 4.74650830e-01
2.15908587e-01 3.03067505e-01 -1.02460110e+00 1.32794547e+00
2.78950512e-01 -5.96604764e-01 -2.10363530e-02 8.73732388e-01
8.40285718e-01 1.54955968e-01 -7.75927901e-01 5.16540885e-01
1.34144819e+00 -5.59725702e-01 1.07781157e-01 -4.76584733e-01
1.47451818e-01 -6.21373892e-01 1.25780928e+00 5.93606591e-01
-7.82454312e-01 -2.01854423e-01 -1.48809814e+00 -1.23559460e-01
-6.28550768e-01 8.98646638e-02 8.01806688e-01 7.25373209e-01
-8.84110749e-01 3.74166310e-01 -7.50355601e-01 -3.03285182e-01
6.42201424e-01 2.33495489e-01 -4.93703485e-01 -6.45012975e-01
-7.33788729e-01 2.97395617e-01 -3.25028986e-01 2.35679910e-01
-7.45178461e-01 -6.03622973e-01 -9.62569773e-01 -5.17415166e-01
4.27416444e-01 -8.99766326e-01 6.56989455e-01 -8.97892952e-01
-1.56563735e+00 1.08523357e+00 9.79857370e-02 -1.64769605e-01
5.10416329e-01 -3.73987257e-01 -3.90992314e-01 5.09167075e-01
1.06561981e-01 5.31441808e-01 9.17972624e-01 -1.24572730e+00
-1.47101179e-01 -7.45682657e-01 4.07815017e-02 4.07434702e-01
-2.23338604e-01 -6.42701149e-01 -3.07693094e-01 -3.08694065e-01
7.73062766e-01 -9.54899073e-01 1.54460698e-01 5.79288185e-01
-4.69277620e-01 5.61478376e-01 7.97226548e-01 -4.54815209e-01
1.34377196e-01 -1.92304456e+00 -2.26561159e-01 4.02867913e-01
2.39596188e-01 -4.20639843e-01 -6.85563833e-02 3.89490634e-01
3.01404476e-01 -3.50622654e-01 7.62077346e-02 -7.94322565e-02
-2.81296354e-02 -2.09933370e-02 -3.53633195e-01 6.64121568e-01
1.51246876e-01 8.33395064e-01 -4.18387651e-01 -1.62202269e-01
4.24847245e-01 4.79613006e-01 -6.48365915e-01 1.06072843e-01
-2.75769413e-01 8.57983679e-02 -3.97851050e-01 1.46870482e+00
1.01295733e+00 -2.55124032e-01 -2.00055148e-02 -7.51834750e-01
-1.24130383e-01 1.44342914e-01 -1.09696805e+00 1.83683360e+00
-6.47087038e-01 8.26749086e-01 3.04415405e-01 -5.56599557e-01
1.23181486e+00 -3.71937841e-01 6.37452245e-01 -9.66867983e-01
-1.46514624e-01 3.39761198e-01 -6.24640822e-01 -3.98912698e-01
9.27587152e-01 8.70605335e-02 2.83593982e-02 1.70987606e-01
-2.87010700e-01 -6.67601526e-01 -4.16298509e-01 8.21774360e-03
1.10307014e+00 5.61503768e-01 -1.79157510e-01 -1.05701417e-01
-1.51260383e-02 1.32565424e-02 1.62272468e-01 7.45049000e-01
1.83185920e-01 9.59679484e-01 -2.48093188e-01 -7.82748520e-01
-1.06304538e+00 -1.46966434e+00 -2.31433496e-01 8.98211956e-01
3.79225105e-01 -2.92894512e-01 -2.53036082e-01 -1.71007589e-02
4.17547017e-01 -2.65690714e-01 -6.98957384e-01 -5.14955446e-02
-4.30305451e-01 -7.40984917e-01 3.83274466e-01 6.43009841e-01
1.08348787e+00 -6.25137925e-01 -9.49623764e-01 -2.26265229e-02
-8.06632042e-02 -1.26369143e+00 1.38136506e-01 3.73549193e-01
-7.74178922e-01 -1.05843616e+00 -4.07012343e-01 -1.96980551e-01
3.17977101e-01 8.29658806e-01 1.34253597e+00 -3.44834834e-01
-8.31314027e-01 1.06368911e+00 -2.12639138e-01 -3.78932983e-01
3.23087215e-01 -3.98380160e-02 1.46490902e-01 1.68865785e-01
6.73952997e-02 -9.31663096e-01 -9.73211884e-01 4.98976618e-01
-5.88985562e-01 3.92181337e-01 8.98734629e-01 4.77482110e-01
9.45570648e-01 -4.95162040e-01 -4.77994978e-02 -4.77005690e-01
1.99045762e-02 -2.47730091e-01 -5.89744329e-01 2.06113502e-01
-5.20671546e-01 -2.49211803e-01 2.65000790e-01 -1.00201778e-01
-7.42078900e-01 -7.70342201e-02 1.20294794e-01 -3.76270175e-01
-3.92116010e-02 4.30021316e-01 -2.31763124e-01 -5.81664801e-01
8.16794336e-01 3.77016217e-01 2.45187674e-02 -1.45691827e-01
1.79699510e-01 6.82294488e-01 5.80444217e-01 -8.91403913e-01
7.58222640e-01 1.02861691e+00 2.49738336e-01 -1.33942986e+00
-7.90918529e-01 -1.97638527e-01 -4.89244699e-01 -6.03112161e-01
6.18052065e-01 -1.29241014e+00 -9.71344352e-01 8.82461727e-01
-5.60143828e-01 -7.15714097e-01 -2.18247384e-01 3.20454508e-01
-7.34181941e-01 1.50880456e-01 -5.15027940e-01 -8.24982762e-01
-4.31078255e-01 -9.58123565e-01 1.77665293e+00 9.66161415e-02
2.26627603e-01 -6.06135666e-01 -2.61048645e-01 6.35054290e-01
6.64656341e-01 5.45560598e-01 6.76742494e-01 4.08691436e-01
-1.18553710e+00 -2.01120287e-01 -4.67574954e-01 1.26315877e-01
3.16444159e-01 3.13685089e-03 -1.30167222e+00 -2.68039852e-01
-3.76503497e-01 -7.31287837e-01 1.01442957e+00 3.83468062e-01
1.10428870e+00 2.05124661e-01 -5.85292988e-02 1.20816183e+00
1.65188265e+00 -3.63572866e-01 7.22690642e-01 7.91314602e-01
1.08591175e+00 5.32253087e-01 4.69872475e-01 4.77535278e-01
5.40193558e-01 7.99860358e-01 7.08799839e-01 -2.51732022e-01
-1.26632214e-01 -1.22346766e-01 4.85870779e-01 4.62699383e-01
-5.23345470e-01 -8.54386687e-02 -9.30948555e-01 1.68810785e-01
-1.15126622e+00 -7.84685254e-01 -2.35224944e-02 2.17696667e+00
4.30972785e-01 1.71999380e-01 -9.35994461e-02 -6.57489970e-02
2.35344663e-01 3.48206729e-01 -1.00250769e+00 -1.27285555e-01
-7.39937484e-01 1.17333911e-01 1.00221896e+00 2.34184891e-01
-9.33185577e-01 6.95779741e-01 6.47291613e+00 4.71273720e-01
-1.54197013e+00 -5.44741333e-01 3.58918875e-01 -2.74535790e-02
-5.21456301e-01 -4.99710917e-01 -7.33640254e-01 -2.00244650e-01
7.59494007e-01 4.35946584e-01 6.84524596e-01 1.00700390e+00
-6.32441491e-02 -2.51095146e-01 -7.45185435e-01 1.25668037e+00
2.86743432e-01 -1.34685421e+00 1.35625154e-01 3.59687746e-01
4.72791582e-01 7.37281084e-01 5.15803218e-01 -3.05159986e-01
3.38602632e-01 -8.34359467e-01 9.63899136e-01 7.52513468e-01
1.32724595e+00 -1.91541150e-01 1.27871826e-01 -1.44333467e-01
-1.56694949e+00 -1.42173590e-02 -4.77217108e-01 8.00080076e-02
-2.17254177e-01 6.86180234e-01 -7.82194972e-01 8.27853739e-01
1.09907925e+00 1.13162422e+00 -7.17534184e-01 5.28163970e-01
2.08923742e-01 2.78053463e-01 -7.82999635e-01 1.71253562e-01
1.31665659e-03 -4.21440512e-01 3.59583586e-01 8.59130919e-01
3.94208938e-01 -3.48647207e-01 1.58862621e-01 9.37435210e-01
-5.80973849e-02 -1.59157827e-01 -1.09080350e+00 1.93746388e-01
4.92921382e-01 1.44183910e+00 -8.52544546e-01 -2.12727830e-01
-3.79613996e-01 8.10752928e-01 2.18614101e-01 2.16021419e-01
-4.30660129e-01 -1.92362383e-01 7.36478329e-01 3.85196090e-01
4.51483428e-01 -7.22614288e-01 -2.53587991e-01 -1.14074564e+00
4.04001743e-01 -7.64333785e-01 -1.29146814e-01 -1.30847824e+00
-1.26626062e+00 4.30939198e-01 -1.19400531e-01 -1.34633553e+00
4.01397943e-01 -1.14739776e+00 -1.16890945e-01 7.78829753e-01
-1.61854672e+00 -1.52728426e+00 -1.20430756e+00 6.44196272e-01
4.37250584e-01 -8.85145813e-02 9.23871040e-01 6.33730888e-02
-2.51054406e-01 2.80813307e-01 2.95622975e-01 -1.28218353e-01
7.36910164e-01 -1.21754265e+00 3.10629696e-01 3.85028988e-01
6.64162822e-03 3.99800301e-01 3.93340915e-01 -5.17577589e-01
-2.20735335e+00 -1.09312093e+00 -2.56422430e-01 -6.12250149e-01
3.06578875e-01 -5.73563874e-01 -4.91127253e-01 5.74688375e-01
-3.83871257e-01 7.70584643e-02 7.80402303e-01 3.47915702e-02
-8.88184905e-01 -5.89044333e-01 -1.10141003e+00 4.20583606e-01
1.49035966e+00 -8.23081911e-01 1.14702247e-02 3.88857901e-01
3.56443405e-01 -6.40892863e-01 -9.08615947e-01 5.92514813e-01
1.34169042e+00 -1.05384398e+00 1.25951862e+00 4.43683453e-02
4.86276507e-01 -1.41459405e-01 -1.06601346e+00 -1.07043719e+00
-6.33364916e-02 -2.13691831e-01 3.89184475e-01 6.97249830e-01
3.20466667e-01 -9.71681833e-01 1.15137589e+00 4.37085688e-01
-4.93994713e-01 -6.57892823e-01 -7.60158598e-01 -8.30394804e-01
-2.79592752e-01 -6.54903769e-01 7.10012138e-01 6.18393004e-01
-6.81470513e-01 1.75316297e-02 -1.47328481e-01 3.28321725e-01
7.98795521e-01 7.41880059e-01 1.10125709e+00 -1.36724830e+00
-3.05158943e-01 -4.68566492e-02 -7.21411526e-01 -1.24548817e+00
-1.63651586e-01 -8.18520963e-01 2.73153912e-02 -1.47122717e+00
-7.23354444e-02 -6.79593384e-01 1.07850477e-01 3.32616389e-01
4.24179673e-01 8.47331405e-01 -3.15299511e-01 2.73218155e-01
-6.10709548e-01 6.76603913e-01 1.12064242e+00 -4.37271863e-01
1.88938156e-02 -4.17589158e-01 -4.77908194e-01 5.77514410e-01
6.82486773e-01 1.72721073e-01 -1.80545583e-01 -7.62657225e-01
7.66704321e-01 -1.79179490e-01 6.47898138e-01 -1.45296061e+00
-9.91143137e-02 -3.26565295e-01 7.71237314e-01 -6.20537937e-01
9.98008668e-01 -8.04861069e-01 2.27216348e-01 4.00281064e-02
1.24789216e-01 2.14633830e-02 1.71265379e-01 7.79645622e-01
1.95367873e-01 6.57401621e-01 4.72582519e-01 -2.25855082e-01
-8.01638782e-01 4.13730115e-01 -2.50921965e-01 -2.32421055e-01
5.03570020e-01 -8.54488909e-01 -7.73957133e-01 -3.81146133e-01
-3.92547667e-01 -2.13994175e-01 1.07397532e+00 7.77300373e-02
9.86041129e-01 -1.41233170e+00 -4.62483853e-01 6.72863603e-01
5.74295402e-01 1.88909069e-01 4.24851865e-01 7.69195616e-01
-8.44148576e-01 9.41472948e-02 -5.39340496e-01 -1.00171232e+00
-1.09320450e+00 -2.95005351e-01 6.01517856e-01 2.45606020e-01
-9.19140160e-01 7.89545894e-01 1.25510901e-01 -6.75154686e-01
-3.97122324e-01 -3.94588262e-01 3.69154453e-01 -1.90390408e-01
1.14427805e-01 2.31319800e-01 5.53666651e-01 -5.82032144e-01
-1.93483233e-01 1.16834986e+00 1.76653951e-01 -1.75615802e-01
1.81491518e+00 -2.39696532e-01 -9.06290933e-02 7.06122220e-01
1.05750918e+00 1.11435093e-01 -1.53976846e+00 -3.86484444e-01
-6.93665683e-01 -7.25086212e-01 1.67944849e-01 -7.10442305e-01
-9.96301532e-01 8.37877572e-01 8.71699631e-01 1.07125118e-01
1.02886724e+00 -2.17743650e-01 6.37102246e-01 9.88519669e-01
9.07679677e-01 -1.24867666e+00 3.66736263e-01 4.63926226e-01
9.35053706e-01 -1.30705333e+00 2.97914594e-01 -5.96750319e-01
-6.52581379e-02 1.45879960e+00 5.21761239e-01 -1.71326116e-01
6.33356094e-01 2.87903130e-01 1.98465675e-01 -7.46836662e-01
-4.73140568e-01 1.46134291e-02 2.80815870e-01 7.32348263e-01
-3.52002345e-02 3.83638531e-01 7.88455129e-01 7.43519664e-02
-5.35261691e-01 -3.81830782e-01 3.09962958e-01 1.22307503e+00
-5.83355844e-01 -5.99182487e-01 -5.22965372e-01 8.81168365e-01
1.15772597e-01 -7.32022664e-03 -4.58001196e-01 6.01796448e-01
-9.78539884e-02 6.57157838e-01 1.22934029e-01 -7.68334806e-01
5.44145048e-01 -4.68514897e-02 8.45995843e-01 -3.63378853e-01
1.43544272e-01 -1.66938797e-01 3.27451080e-01 -9.78854358e-01
-6.21294975e-01 -7.36092389e-01 -6.22060239e-01 -3.23006749e-01
-3.22773606e-01 -6.00074232e-01 1.05367374e+00 6.71566188e-01
3.77950370e-01 4.24303114e-01 7.39728928e-01 -1.46525466e+00
-3.68185155e-02 -6.36812270e-01 -9.79245663e-01 1.48195758e-01
3.63754392e-01 -1.00861561e+00 -3.37359011e-01 -7.96401352e-02] | [9.47161865234375, -2.7694756984710693] |
f6a20fbb-1e6b-4d74-87ed-da87de16d6b4 | counter-gap-counterfactual-bias-evaluation | 2302.05674 | null | https://arxiv.org/abs/2302.05674v1 | https://arxiv.org/pdf/2302.05674v1.pdf | Counter-GAP: Counterfactual Bias Evaluation through Gendered Ambiguous Pronouns | Bias-measuring datasets play a critical role in detecting biased behavior of language models and in evaluating progress of bias mitigation methods. In this work, we focus on evaluating gender bias through coreference resolution, where previous datasets are either hand-crafted or fail to reliably measure an explicitly defined bias. To overcome these shortcomings, we propose a novel method to collect diverse, natural, and minimally distant text pairs via counterfactual generation, and construct Counter-GAP, an annotated dataset consisting of 4008 instances grouped into 1002 quadruples. We further identify a bias cancellation problem in previous group-level metrics on Counter-GAP, and propose to use the difference between inconsistency across genders and within genders to measure bias at a quadruple level. Our results show that four pre-trained language models are significantly more inconsistent across different gender groups than within each group, and that a name-based counterfactual data augmentation method is more effective to mitigate such bias than an anonymization-based method. | ['Oana-Maria Camburu', 'Thomas Lukasiewicz', 'Vid Kocijan', 'Zhongbin Xie'] | 2023-02-11 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [ 9.30919200e-02 2.40161121e-01 -7.09667146e-01 -8.52588892e-01
-7.55084872e-01 -6.67789280e-01 1.18055260e+00 3.83189976e-01
-6.84806049e-01 1.26509082e+00 8.46181333e-01 -2.73819417e-01
8.72185733e-03 -8.29388380e-01 -7.06802070e-01 -2.51347691e-01
2.26359978e-01 5.58115423e-01 -3.55975509e-01 -3.72571349e-01
5.49128652e-01 1.66626766e-01 -1.07279134e+00 1.72304600e-01
1.25697815e+00 7.19367936e-02 -7.38646567e-01 2.34064132e-01
-2.16845036e-01 5.64400792e-01 -8.14142644e-01 -1.06062973e+00
1.78428695e-01 -5.43101430e-01 -1.03771138e+00 -7.30044663e-01
1.05148470e+00 -2.94520795e-01 -7.99731165e-02 1.18303347e+00
8.72334480e-01 -2.26656690e-01 7.84196019e-01 -1.54639304e+00
-4.82775480e-01 1.36964500e+00 -8.49342167e-01 3.51127446e-01
3.96371901e-01 -8.96118805e-02 9.56789315e-01 -5.72893798e-01
8.59070599e-01 1.76811647e+00 9.65616167e-01 1.01027215e+00
-1.72137272e+00 -1.46341169e+00 2.75614262e-01 -1.54506475e-01
-1.08882308e+00 -7.03453422e-01 9.43482220e-01 -5.16122162e-01
2.10159048e-01 5.58209121e-01 2.07369938e-01 1.86943269e+00
-1.11085430e-01 4.36559796e-01 1.40266609e+00 -3.02636713e-01
1.04335770e-02 2.30100572e-01 4.32584703e-01 4.19982463e-01
8.77295434e-01 1.83313251e-01 -9.34496224e-01 -8.17355335e-01
2.98540682e-01 -2.06451297e-01 -2.84456432e-01 -4.67187673e-01
-1.50263512e+00 1.04978216e+00 3.16433400e-01 3.72234643e-01
1.24417752e-01 -1.11430325e-02 6.42590344e-01 2.64418602e-01
7.55120099e-01 8.54898870e-01 -3.91330808e-01 -2.93253008e-02
-1.23474324e+00 8.64560246e-01 6.80957913e-01 7.37031341e-01
5.79546928e-01 -3.12184662e-01 -6.17880702e-01 1.05022991e+00
-7.25490227e-02 8.57540369e-01 5.60333312e-01 -9.53459918e-01
9.84240770e-01 6.26060665e-01 4.44828659e-01 -1.32495570e+00
-6.40575171e-01 -3.01303834e-01 -1.05600810e+00 -3.64261121e-01
9.89478946e-01 -2.47543901e-01 -4.53707993e-01 2.37963653e+00
3.96908134e-01 -2.80749321e-01 -1.14847422e-01 7.24081993e-01
7.76048183e-01 -1.11481205e-01 3.49182814e-01 -1.74910054e-01
1.25486970e+00 -4.48566139e-01 -7.14690864e-01 -3.26169372e-01
1.04971850e+00 -5.32125592e-01 1.11359179e+00 -1.10649906e-01
-8.69743824e-01 -2.41016716e-01 -9.88927007e-01 5.75658977e-02
-3.88340116e-01 -3.00712168e-01 7.23351955e-01 1.03706574e+00
-6.75540209e-01 5.56967199e-01 -9.50917304e-02 -3.44777286e-01
5.89897990e-01 1.84764519e-01 -5.25617063e-01 -1.53864333e-02
-1.59074771e+00 7.69169986e-01 1.42059594e-01 -3.39877903e-01
-5.32237232e-01 -1.26626992e+00 -9.25312459e-01 -2.70993024e-01
3.34712684e-01 -7.66718388e-01 8.80275309e-01 -8.21521223e-01
-5.35045624e-01 1.49427092e+00 -4.62763101e-01 -5.55794835e-01
1.08122683e+00 -1.23587385e-01 -5.31414509e-01 -5.86471081e-01
6.19461596e-01 6.19596720e-01 5.58125496e-01 -1.51451874e+00
-5.15089035e-01 -7.65500724e-01 -1.52754053e-01 -3.36952694e-02
-4.07999247e-01 2.92248577e-01 3.06283623e-01 -8.70725214e-01
-1.96786877e-02 -7.72933900e-01 -1.25282690e-01 -5.19636035e-01
-7.04853892e-01 -1.94205612e-01 3.55641276e-01 -7.17059255e-01
1.42133200e+00 -1.73032796e+00 -2.55254537e-01 2.44467527e-01
4.24175173e-01 2.13899627e-01 -9.91686732e-02 4.73398529e-02
-2.57142425e-01 7.31070578e-01 -2.92382389e-01 -3.51577997e-01
9.34882089e-02 -1.21321738e-01 -6.08012736e-01 5.12065351e-01
-2.06868090e-02 5.59081137e-01 -1.21505451e+00 -7.34695733e-01
-3.95811498e-01 -1.17984846e-01 -7.72344112e-01 -6.59001768e-02
2.26882085e-01 4.00168955e-01 -7.50170052e-02 6.39217377e-01
1.16597307e+00 5.18456995e-01 3.74212325e-01 -9.69572663e-02
-1.90949097e-01 6.77831769e-01 -9.21548247e-01 1.48350728e+00
-3.93563420e-01 5.84220946e-01 8.67163539e-02 -7.83719063e-01
1.17165065e+00 -1.81062922e-01 1.93807125e-01 -7.54611492e-01
1.07915886e-02 4.62977856e-01 8.14188495e-02 -1.14979513e-01
9.19678092e-01 -2.72933692e-01 -4.95058179e-01 6.16630137e-01
-1.69015452e-01 -6.10634834e-02 3.20976973e-01 1.74173370e-01
8.24454367e-01 -3.65954489e-01 2.64474720e-01 -5.16096175e-01
5.27012169e-01 -1.18556030e-01 9.92052317e-01 1.21880937e+00
-5.71323216e-01 7.08035827e-01 9.55232143e-01 -4.15350914e-01
-9.44598675e-01 -8.74471724e-01 -3.33892435e-01 1.12561786e+00
-2.22698469e-02 -4.05616939e-01 -7.14056075e-01 -1.40223897e+00
5.03199637e-01 1.06190753e+00 -9.45550919e-01 -4.09476906e-01
-6.91252530e-01 -1.19613433e+00 1.07757807e+00 3.27288300e-01
7.40847468e-01 -4.72155750e-01 -7.69408941e-02 -1.95791155e-01
-6.89770758e-01 -6.20861828e-01 -5.77975929e-01 -3.44138950e-01
-6.60722494e-01 -1.19752252e+00 -6.28580391e-01 -2.51297027e-01
5.65715373e-01 -4.57774103e-02 1.59253991e+00 2.97135375e-02
2.21976936e-01 -8.18496719e-02 1.84107557e-01 -9.35933053e-01
-5.95013976e-01 3.38074178e-01 2.72476107e-01 -1.31686911e-01
8.71391118e-01 -3.81689191e-01 -5.26466787e-01 4.54545557e-01
-4.10387754e-01 -2.72710860e-01 1.10876322e-01 1.00138927e+00
-1.55404434e-01 -6.52148068e-01 7.83134520e-01 -1.45950019e+00
9.17964399e-01 -5.10146201e-01 -2.16983512e-01 3.87431146e-03
-9.63720679e-01 1.29410736e-02 3.22490364e-01 -3.35754782e-01
-1.19662869e+00 -6.82514250e-01 1.76186115e-01 3.29059064e-01
5.08110691e-03 1.06037492e-02 -3.26892078e-01 2.85811990e-01
9.96267676e-01 -3.12963486e-01 -6.31171104e-04 -3.66506696e-01
2.58889467e-01 9.09540892e-01 5.64147174e-01 -8.97049248e-01
7.17029870e-01 6.73834980e-01 -1.63873762e-01 -2.46303931e-01
-8.46065462e-01 -1.84066817e-01 -6.79291546e-01 4.42834347e-02
3.93962264e-01 -9.59349155e-01 -5.32283008e-01 3.90991896e-01
-1.20425904e+00 -1.54494405e-01 -9.00673941e-02 4.34124202e-01
-1.36798009e-01 1.40946940e-01 -1.08280540e-01 -7.28696287e-01
-3.99795115e-01 -5.48759460e-01 9.86992836e-01 -3.14478856e-03
-9.80254233e-01 -9.07206416e-01 3.95760030e-01 6.12679183e-01
1.54387131e-01 4.10195649e-01 7.80559659e-01 -9.50395405e-01
1.89058289e-01 -1.42905548e-01 -4.90130752e-01 -3.65349911e-02
1.64596230e-01 -7.30589358e-03 -9.38671112e-01 -3.43869716e-01
-3.35267246e-01 -2.68527269e-01 1.04566920e+00 1.92872390e-01
1.12194324e+00 -3.61106843e-01 -6.42025948e-01 3.79775614e-01
1.03410077e+00 -1.69248551e-01 3.75471801e-01 3.78243208e-01
7.93562770e-01 9.40662622e-01 6.94943249e-01 1.79383412e-01
6.39169097e-01 5.98233640e-01 9.40902978e-02 -1.11453451e-01
8.62857550e-02 -6.82125866e-01 1.77524790e-01 1.72456831e-01
4.85987199e-04 1.04092821e-01 -1.20332372e+00 1.05222726e+00
-1.68129265e+00 -1.23185468e+00 -3.50912362e-01 2.38699245e+00
1.26029158e+00 2.65058041e-01 2.44982764e-01 1.55425236e-01
1.05495501e+00 4.52316016e-01 -2.49031126e-01 -6.56693518e-01
-3.56309891e-01 -2.63214678e-01 7.97447264e-01 4.76350367e-01
-9.13196981e-01 8.01982045e-01 6.91174603e+00 5.13182759e-01
-9.16641295e-01 1.66847073e-02 8.25135469e-01 -3.40587258e-01
-7.84244120e-01 -1.60483515e-03 -7.70476222e-01 7.65247762e-01
7.73983002e-01 -4.08580720e-01 1.92632079e-02 5.32429099e-01
2.54202545e-01 -6.08086474e-02 -1.52235043e+00 1.09463465e+00
1.51076466e-01 -1.19786656e+00 2.55400896e-01 1.79202482e-01
9.24263954e-01 -3.30153346e-01 2.21365839e-02 3.13721508e-01
6.43527806e-01 -1.07471848e+00 9.10824955e-01 3.99220794e-01
7.37238646e-01 -7.89013624e-01 9.20403481e-01 2.54750639e-01
-3.20734203e-01 -1.30447224e-01 -3.12230796e-01 -2.47690648e-01
-1.91354021e-01 1.00927579e+00 -8.38603675e-01 5.63628316e-01
8.12063396e-01 4.35819417e-01 -6.88427508e-01 4.51607734e-01
2.75774859e-02 7.85383999e-01 -1.03705358e-02 -2.29550935e-02
-1.90488338e-01 -3.73465158e-02 6.21450722e-01 1.33508420e+00
8.75813290e-02 -3.37664664e-01 -4.13199723e-01 1.06092000e+00
-6.47458792e-01 4.64789718e-02 -8.76933992e-01 1.10517055e-01
9.24300909e-01 9.61009860e-01 -2.05538735e-01 -3.98736179e-01
-1.50565416e-01 4.88746762e-01 4.55085933e-01 1.61519215e-01
-6.71838284e-01 -3.35750610e-01 1.16799998e+00 7.88081884e-02
-4.87449616e-01 1.85842320e-01 -8.22830260e-01 -1.23233128e+00
-2.01871283e-02 -1.19727886e+00 7.22951531e-01 -3.59446198e-01
-1.35850573e+00 -1.03883058e-01 1.58540294e-01 -7.74922371e-01
-3.77133638e-01 -1.28939688e-01 -5.24081051e-01 1.05730259e+00
-1.22384381e+00 -9.93909478e-01 -3.14495116e-01 3.34742069e-01
-5.61104119e-02 -1.71771303e-01 6.91487014e-01 4.23806727e-01
-6.58342421e-01 1.16808748e+00 -1.38016075e-01 4.86936808e-01
1.57340920e+00 -1.36049366e+00 3.59674573e-01 7.49574423e-01
-3.55445892e-02 1.10375655e+00 1.03014016e+00 -7.86701024e-01
-6.14899695e-01 -1.02399230e+00 1.62586546e+00 -1.00623739e+00
3.43453139e-01 -7.21550047e-01 -6.51498199e-01 6.81547701e-01
7.74922222e-02 -2.89836019e-01 8.27787876e-01 7.15674698e-01
-8.33944917e-01 -3.92322361e-01 -1.58173430e+00 8.67121935e-01
1.65960705e+00 -6.51147783e-01 -9.65161324e-01 2.03467272e-02
5.64732254e-01 -3.80086422e-01 -5.45875907e-01 5.74420393e-01
4.78133768e-01 -1.10491836e+00 8.52621794e-01 -1.11510384e+00
7.46719718e-01 8.59411061e-02 4.58441786e-02 -1.56817400e+00
-1.19315714e-01 -5.41218460e-01 3.91952664e-01 2.08093071e+00
5.15459001e-01 -8.26277077e-01 8.07449698e-01 8.52167666e-01
4.80273575e-01 -1.63364291e-01 -8.85573030e-01 -4.72533494e-01
7.05070913e-01 -1.33421525e-01 1.30595040e+00 1.77167213e+00
8.13017488e-02 1.70006514e-01 -4.21210647e-01 -1.37608618e-01
8.00205231e-01 2.57368922e-01 1.22637820e+00 -1.36611211e+00
4.50250238e-01 -4.65701550e-01 -5.24392538e-02 -3.53637248e-01
6.67035699e-01 -1.01610124e+00 -2.69425035e-01 -9.86967921e-01
6.15224063e-01 -6.39367402e-01 -1.21920347e-01 3.05200964e-01
-4.88873154e-01 2.18096346e-01 -4.79330011e-02 4.00597714e-02
-7.04360232e-02 4.17118609e-01 9.31085110e-01 -4.02603030e-01
4.24597040e-03 -1.45889446e-01 -1.43230200e+00 7.28923738e-01
6.40055954e-01 -7.02023923e-01 -6.93026781e-02 -2.82946259e-01
3.71706158e-01 -4.95443314e-01 5.63110828e-01 -4.58370149e-01
-1.26667812e-01 -3.21875006e-01 4.74455863e-01 -3.87569904e-01
-3.18784684e-01 -3.45388204e-01 -2.98042566e-01 4.22847629e-01
-8.28321457e-01 -2.33299192e-02 -2.75874441e-03 3.91634524e-01
-1.45752862e-01 -3.05157690e-03 5.35270035e-01 -1.24110758e-01
-2.05232874e-01 -2.28264127e-02 3.29221860e-02 6.87408507e-01
4.83734459e-01 2.53929377e-01 -6.83877766e-01 -2.04053566e-01
-1.37198240e-01 4.06726509e-01 4.70295697e-01 6.73319936e-01
-1.07496671e-01 -1.60681939e+00 -1.15090179e+00 1.58236817e-01
3.85900915e-01 -3.48549902e-01 -4.52249194e-04 7.07080007e-01
1.03992216e-01 3.59050393e-01 -2.23211646e-01 -3.16783875e-01
-1.52603650e+00 2.11892471e-01 3.88772339e-01 -4.64918226e-01
2.26411283e-01 8.79226148e-01 -7.67375110e-03 -1.08340347e+00
1.04689719e-02 -2.05611050e-01 -1.32900670e-01 4.84195262e-01
4.31639254e-01 6.25250697e-01 5.86379282e-02 -6.66916907e-01
-5.70748925e-01 1.62032813e-01 -1.84579343e-01 -1.05134174e-01
9.42916036e-01 -2.01549545e-01 -3.79931867e-01 3.69598389e-01
9.09618378e-01 6.66076899e-01 -6.25265479e-01 -1.56505376e-01
2.33177558e-01 -8.84803832e-01 -3.66922587e-01 -9.37229693e-01
-7.08377838e-01 4.49302346e-01 2.80596107e-01 1.29547939e-01
5.10882795e-01 -2.41950944e-01 2.83223689e-01 -1.20482281e-01
3.73046607e-01 -1.32678556e+00 -4.26483661e-01 2.70132422e-01
1.06614435e+00 -1.63906121e+00 1.35219976e-01 -5.28862953e-01
-2.88845211e-01 4.39353198e-01 8.05381179e-01 1.96706802e-01
2.91738749e-01 -7.98844695e-02 2.17693076e-01 -6.34832354e-03
-4.64031428e-01 2.41925627e-01 2.20064551e-01 6.96563423e-01
8.16833854e-01 1.41244531e-01 -1.04848623e+00 9.69587803e-01
-8.95717800e-01 -2.18964159e-01 4.95560765e-01 4.20658857e-01
3.11558723e-01 -1.16083443e+00 -7.12542355e-01 6.66565001e-01
-5.88819206e-01 -9.59571451e-02 -1.04521143e+00 9.68725204e-01
3.42154086e-01 9.30938482e-01 2.69823492e-01 -3.92966121e-01
5.15215456e-01 2.73310244e-01 3.61734271e-01 -1.82826370e-01
-7.83997059e-01 -9.38727319e-01 7.22087622e-01 -7.07517505e-01
-5.92810333e-01 -9.51494753e-01 -7.49188602e-01 -7.88671374e-01
-3.69420610e-02 2.66699255e-01 3.84981006e-01 8.28381419e-01
3.39508384e-01 -1.85654089e-02 5.62041819e-01 -4.10106212e-01
-8.41762722e-01 -1.11457288e+00 -4.70543057e-01 1.03843999e+00
4.21490788e-01 -7.83353865e-01 -6.27514303e-01 -4.68215346e-01] | [9.362992286682129, 10.242742538452148] |
c79284a9-32cd-4c24-9cc7-ca83852dceeb | order-disorder-imitation-adversarial-attacks | 2209.06506 | null | https://arxiv.org/abs/2209.06506v2 | https://arxiv.org/pdf/2209.06506v2.pdf | Order-Disorder: Imitation Adversarial Attacks for Black-box Neural Ranking Models | Neural text ranking models have witnessed significant advancement and are increasingly being deployed in practice. Unfortunately, they also inherit adversarial vulnerabilities of general neural models, which have been detected but remain underexplored by prior studies. Moreover, the inherit adversarial vulnerabilities might be leveraged by blackhat SEO to defeat better-protected search engines. In this study, we propose an imitation adversarial attack on black-box neural passage ranking models. We first show that the target passage ranking model can be transparentized and imitated by enumerating critical queries/candidates and then train a ranking imitation model. Leveraging the ranking imitation model, we can elaborately manipulate the ranking results and transfer the manipulation attack to the target ranking model. For this purpose, we propose an innovative gradient-based attack method, empowered by the pairwise objective function, to generate adversarial triggers, which causes premeditated disorderliness with very few tokens. To equip the trigger camouflages, we add the next sentence prediction loss and the language model fluency constraint to the objective function. Experimental results on passage ranking demonstrate the effectiveness of the ranking imitation attack model and adversarial triggers against various SOTA neural ranking models. Furthermore, various mitigation analyses and human evaluation show the effectiveness of camouflages when facing potential mitigation approaches. To motivate other scholars to further investigate this novel and important problem, we make the experiment data and code publicly available. | ['Xiaozhong Liu', 'Wei Lu', 'XiaoFeng Wang', 'Changlong Sun', 'Kaisong Song', 'Di Tang', 'Yangyang Kang', 'Jiawei Liu'] | 2022-09-14 | null | null | null | null | ['passage-ranking'] | ['natural-language-processing'] | [ 3.68342996e-01 -6.13960624e-02 -9.48894322e-02 -2.14869559e-01
-8.95882726e-01 -1.03845513e+00 7.36496389e-01 -3.72039944e-01
-4.91147637e-01 5.97450554e-01 2.63677090e-01 -4.62428302e-01
-2.56417006e-01 -7.89531410e-01 -9.11888361e-01 -4.81204510e-01
-3.89973223e-02 6.32864684e-02 3.04019302e-01 -6.59241319e-01
2.98447639e-01 1.41985297e-01 -1.02491355e+00 4.11066115e-01
1.17675900e+00 7.01318085e-01 7.58974114e-04 6.48833513e-01
1.96702927e-01 1.19551921e+00 -1.01575518e+00 -1.02573943e+00
4.41255808e-01 -2.11780384e-01 -6.92967892e-01 -1.03977489e+00
4.95952278e-01 -8.67374420e-01 -8.35690916e-01 1.30115211e+00
8.27176034e-01 8.46367925e-02 4.48604703e-01 -1.19754469e+00
-9.08480346e-01 1.13959360e+00 -1.79506540e-01 2.70330310e-01
3.42619359e-01 2.47148409e-01 1.18201816e+00 -4.85122263e-01
3.36046517e-01 1.22758913e+00 3.86850715e-01 9.03621435e-01
-6.00660205e-01 -1.14619887e+00 3.34472895e-01 2.99481809e-01
-1.01472139e+00 -1.62954807e-01 9.81593847e-01 -5.77377230e-02
7.05682576e-01 6.19529963e-01 1.51492700e-01 1.69448578e+00
2.22727507e-01 1.14594710e+00 7.85009086e-01 5.95283508e-02
-1.90180063e-01 -5.45461886e-02 2.18456537e-01 8.42265427e-01
1.55184604e-02 4.69006479e-01 -3.91109794e-01 -4.91612852e-01
2.89004803e-01 1.13469571e-01 -1.56862333e-01 1.11414418e-01
-9.05667186e-01 8.65426421e-01 6.28836393e-01 -3.05887125e-02
9.07466114e-02 3.77288043e-01 7.06783056e-01 4.64111209e-01
4.07875925e-01 7.07467973e-01 -5.70135117e-01 2.54085776e-03
-5.82422495e-01 2.88324296e-01 5.47310412e-01 6.67392492e-01
6.49756268e-02 4.06979509e-02 -7.18393087e-01 9.28400993e-01
2.59582639e-01 8.23553741e-01 4.82982844e-01 -6.78062677e-01
6.72152221e-01 4.99296188e-01 -2.22222492e-01 -1.15226889e+00
-1.43519938e-01 -6.83305025e-01 -7.36741602e-01 -2.34362572e-01
1.85263529e-01 -2.67504632e-01 -6.92405701e-01 2.05824757e+00
-8.73598456e-03 2.17306882e-01 1.26448229e-01 8.75996590e-01
5.87310553e-01 5.67977607e-01 -2.90951096e-02 1.08389936e-01
1.15396857e+00 -1.07891893e+00 -4.26218152e-01 -1.84754804e-02
6.13349140e-01 -4.94254380e-01 1.54165566e+00 3.70369703e-01
-9.67094421e-01 -1.80392832e-01 -1.28659868e+00 -2.23823097e-02
-4.34386104e-01 -1.03620715e-01 5.75166106e-01 7.53022492e-01
-5.73247969e-01 6.58760309e-01 -6.23605072e-01 1.28206357e-01
3.88369322e-01 4.08453673e-01 1.35887906e-01 1.14161320e-01
-2.11224604e+00 8.22256327e-01 2.49041766e-01 2.77940124e-01
-1.37918973e+00 -5.95886290e-01 -3.98842007e-01 6.86989203e-02
5.59157073e-01 -6.93664074e-01 1.16508472e+00 -6.98570013e-01
-1.76548445e+00 3.98719847e-01 4.76408541e-01 -4.60369527e-01
9.05682683e-01 -6.42331421e-01 -4.97052997e-01 5.45498133e-02
-1.89712062e-01 9.81209874e-02 9.59603310e-01 -9.21904683e-01
-2.23443568e-01 -1.06812239e-01 5.96444547e-01 1.80445135e-01
-9.34830368e-01 2.88100779e-01 -2.72550583e-01 -1.03512299e+00
-5.00167549e-01 -8.43588769e-01 -2.02042282e-01 -5.13136446e-01
-1.06997085e+00 -8.13966915e-02 6.07969999e-01 -6.95474625e-01
1.56139565e+00 -2.04133701e+00 1.88347638e-01 3.57033074e-01
2.63980180e-01 5.74745834e-01 -4.53104556e-01 3.82236063e-01
2.74254084e-01 6.72969639e-01 -1.69700965e-01 3.68679315e-02
2.73335934e-01 -1.98936045e-01 -9.45559800e-01 1.20157294e-01
-2.90747155e-02 1.21765018e+00 -1.05614614e+00 -3.26823413e-01
-2.87873894e-01 2.89558232e-01 -7.89551437e-01 2.91393220e-01
-3.05353642e-01 9.58546028e-02 -8.96458745e-01 4.74762291e-01
4.49298501e-01 5.63389063e-02 -6.73071146e-02 9.80917737e-03
2.71572948e-01 4.96446073e-01 -5.19556761e-01 1.29591310e+00
-2.77761549e-01 3.63788426e-01 -3.49312961e-01 -6.42590821e-01
6.51772439e-01 -3.19775566e-02 2.70727463e-02 -6.90888047e-01
1.52132556e-01 2.17084929e-01 -9.71256383e-03 -4.48942661e-01
5.56498945e-01 2.59804875e-01 -4.99521732e-01 4.89104062e-01
-3.73847157e-01 1.37844086e-01 -2.21551850e-01 6.35681629e-01
1.47874093e+00 -1.28228083e-01 -3.03784847e-01 1.53926924e-01
6.69535697e-01 -1.26158014e-01 3.84436935e-01 1.32451010e+00
-2.02502608e-01 3.36112499e-01 6.50233150e-01 -2.28339031e-01
-8.33860099e-01 -1.17441893e+00 7.88734257e-02 1.42250645e+00
2.92170882e-01 -4.60894376e-01 -9.55944359e-01 -1.34068096e+00
-1.89968750e-01 4.78634715e-01 -5.27786136e-01 -8.83076310e-01
-7.28300333e-01 -7.08548725e-01 1.43382740e+00 2.85942316e-01
8.19478333e-01 -1.21578455e+00 -1.73714712e-01 4.31522913e-03
-2.41677970e-01 -7.49093533e-01 -7.23366022e-01 -2.25804538e-01
-4.67760295e-01 -9.96228933e-01 -6.03297353e-01 -5.68978548e-01
4.64573860e-01 8.43295604e-02 8.81008506e-01 5.16072929e-01
4.63816673e-02 1.55664578e-01 -4.62409347e-01 -3.32514465e-01
-4.18272346e-01 4.25908923e-01 2.15272099e-01 -3.79912168e-01
5.83160073e-02 -4.52993274e-01 -6.76431239e-01 4.42551404e-01
-1.06463766e+00 -2.00901061e-01 6.07824087e-01 1.00525665e+00
1.28086761e-01 -4.55870070e-02 7.31701195e-01 -1.04621732e+00
1.32589793e+00 -5.35240591e-01 -5.61614096e-01 6.04376793e-01
-4.58363652e-01 1.47009462e-01 1.14800942e+00 -8.91642869e-01
-7.51085043e-01 -4.80711937e-01 -2.44478568e-01 -5.36857307e-01
4.17311043e-01 5.99623919e-01 -4.03372139e-01 1.12057440e-02
7.12211192e-01 2.60884821e-01 -2.73478091e-01 -2.99222529e-01
5.12971640e-01 5.95730364e-01 3.77357006e-01 -6.97096467e-01
1.30252016e+00 1.39493287e-01 -4.40231323e-01 -2.20331758e-01
-9.28092837e-01 2.96355724e-01 1.33637726e-01 -2.00026453e-01
6.57246888e-01 -4.83429581e-01 -9.13628817e-01 6.60405695e-01
-1.32855272e+00 -2.28973597e-01 2.60287941e-01 8.60466510e-02
-1.25519484e-01 5.23443758e-01 -8.12656820e-01 -6.52799070e-01
-7.57727087e-01 -1.13625944e+00 6.78580284e-01 6.09983467e-02
1.96119606e-01 -5.54372370e-01 1.17125005e-01 5.19946098e-01
6.45160794e-01 3.69347408e-02 1.16847098e+00 -1.08243906e+00
-6.43179178e-01 -4.64469165e-01 -6.61306903e-02 4.62647080e-01
-4.20520663e-01 1.22092729e-02 -9.56006765e-01 -2.52471983e-01
-7.53361434e-02 -4.75265384e-01 9.66867745e-01 -1.70186445e-01
1.45971155e+00 -8.26637030e-01 -1.14074573e-01 5.36252499e-01
8.07859123e-01 1.24268569e-01 6.26243830e-01 4.74352837e-01
8.74302864e-01 5.18253446e-01 4.84318376e-01 8.40973035e-02
1.40965253e-01 7.18966722e-01 6.45816028e-01 1.39789730e-01
2.90345669e-01 -7.97308266e-01 8.14482570e-01 9.58151519e-01
2.45141447e-01 -4.26640838e-01 -5.95067322e-01 -4.46814895e-02
-1.87428284e+00 -1.22929168e+00 2.26213083e-01 2.06618834e+00
9.40939784e-01 5.22399604e-01 -8.29151496e-02 -1.12599745e-01
6.83160067e-01 3.26843143e-01 -8.34526181e-01 -4.81307656e-01
-1.04263812e-01 -1.02904826e-01 4.74575311e-01 5.06556273e-01
-1.05086410e+00 1.29372084e+00 6.08079338e+00 1.17261934e+00
-1.01212645e+00 1.96396455e-01 6.37204230e-01 -3.36214602e-01
-6.83013380e-01 -1.51405320e-01 -6.78783655e-01 7.13109255e-01
7.72292852e-01 -3.66320133e-01 7.94143438e-01 8.37870598e-01
4.95116264e-02 6.72716320e-01 -8.70461345e-01 4.23808098e-01
2.92986110e-02 -1.05496621e+00 4.03191626e-01 -1.98464189e-02
5.43502152e-01 6.31245002e-02 5.97555876e-01 8.68826568e-01
6.21727824e-01 -1.07482016e+00 6.09201193e-01 4.96321887e-01
5.81658483e-01 -8.41773450e-01 6.15197539e-01 4.29661632e-01
-7.51816988e-01 -3.66433084e-01 -5.43919027e-01 9.11704898e-02
-4.99608321e-03 3.01935405e-01 -4.47991312e-01 4.68442082e-01
4.68523562e-01 3.26959103e-01 -6.58006370e-01 7.32809722e-01
-6.34600043e-01 8.40587199e-01 -2.45898724e-01 -5.54413497e-01
2.20706984e-01 1.05142623e-01 7.93706775e-01 8.93502951e-01
-6.93707392e-02 -1.33736536e-01 -9.52961575e-03 8.83427203e-01
-5.52019894e-01 1.49626657e-01 -8.11132848e-01 -1.20134301e-01
7.48401701e-01 1.01168954e+00 -3.49737555e-02 -2.23837689e-01
6.93773329e-02 1.09134805e+00 5.93670130e-01 5.28474629e-01
-1.33306837e+00 -7.22272217e-01 5.85195303e-01 -3.32140118e-01
-2.07363725e-01 1.28001468e-02 -1.26210481e-01 -1.33050752e+00
2.23284781e-01 -1.24507582e+00 3.18598777e-01 -4.44785237e-01
-1.46618736e+00 7.55915880e-01 3.63217779e-02 -9.92777109e-01
-1.13362297e-01 -4.25778806e-01 -9.54381406e-01 5.74564278e-01
-1.46398723e+00 -9.69295263e-01 2.36112356e-01 7.19339192e-01
4.16770190e-01 -5.19312024e-01 5.74748397e-01 4.16049153e-01
-1.01232612e+00 1.49424124e+00 1.76517427e-01 4.28462237e-01
6.00206196e-01 -9.03731644e-01 4.32881027e-01 9.23678100e-01
3.40734310e-02 1.18116248e+00 5.69111586e-01 -7.40512490e-01
-1.43428123e+00 -1.09939218e+00 5.77817202e-01 -7.16169477e-01
8.61468434e-01 -6.31056666e-01 -8.31209183e-01 3.69808793e-01
-1.92021169e-02 -3.81580770e-01 5.46520114e-01 1.27180992e-03
-6.92887723e-01 -6.31510317e-02 -1.06739283e+00 1.16318524e+00
1.30046892e+00 -8.90371978e-01 -5.25541365e-01 5.71248233e-01
1.40110111e+00 -2.10508674e-01 -4.49673921e-01 6.54813051e-01
7.37585902e-01 -6.15305066e-01 1.20183301e+00 -1.04757202e+00
7.79440224e-01 4.25465824e-03 -1.24587096e-01 -1.11879027e+00
-7.87907690e-02 -8.21452975e-01 -3.60253960e-01 1.29015291e+00
7.23736465e-01 -6.95494533e-01 7.50268400e-01 5.49473286e-01
-7.14141428e-02 -9.27821875e-01 -8.68785739e-01 -9.33265924e-01
5.74644744e-01 -3.46288413e-01 7.20529020e-01 8.11783552e-01
-4.10576686e-02 3.15507233e-01 -8.25275183e-01 2.62499958e-01
4.34940130e-01 -3.70908022e-01 6.85693502e-01 -7.65229881e-01
-4.83657986e-01 -6.86871231e-01 5.54174893e-02 -1.07466888e+00
5.09650409e-01 -1.07415855e+00 -7.53224790e-02 -1.00837624e+00
2.20889896e-01 -2.73419380e-01 -7.12444544e-01 5.19938350e-01
-6.65577054e-01 9.68931913e-02 3.14663202e-01 3.84151578e-01
-7.44616270e-01 7.40521908e-01 1.03149223e+00 -4.75042701e-01
-2.38708891e-02 5.89020401e-02 -9.11029577e-01 5.81342041e-01
9.80483353e-01 -7.20960796e-01 -6.54491544e-01 -6.46454871e-01
6.97980702e-01 -1.16696648e-01 3.29736590e-01 -6.07814789e-01
2.89883256e-01 1.42544210e-02 -1.12995625e-01 -2.66122073e-01
9.47269499e-02 -5.64349651e-01 -4.20727462e-01 6.08455956e-01
-9.70709920e-01 2.29625940e-01 -1.97532438e-02 6.87794924e-01
8.31341222e-02 -2.40845576e-01 3.75738889e-01 1.48836136e-01
-3.34916919e-01 3.64459038e-01 -2.30984524e-01 2.22196311e-01
8.10104847e-01 2.38316566e-01 -7.52735376e-01 -4.33309346e-01
-1.92721397e-01 4.72655445e-01 1.90405339e-01 7.11032748e-01
7.43632734e-01 -1.17073631e+00 -8.09775412e-01 4.18808199e-02
-9.99761745e-02 -5.02764463e-01 2.45069191e-01 4.76055831e-01
-2.90019006e-01 2.31763154e-01 2.03419238e-01 4.82136868e-02
-1.13830888e+00 7.48851597e-01 5.33341110e-01 -7.00137377e-01
-3.75603139e-01 9.50447321e-01 1.66825324e-01 -7.88290262e-01
5.52994490e-01 9.85776037e-02 -2.76381850e-01 -3.92013043e-01
4.90636289e-01 3.53195786e-01 -2.44162962e-01 -5.97137446e-03
-3.31174374e-01 2.81931102e-01 -6.26331866e-01 -1.10576183e-01
7.47000217e-01 2.37473547e-01 -4.64835614e-02 -1.70809761e-01
1.26363623e+00 3.63560528e-01 -9.41603899e-01 -1.30364597e-01
-1.69389285e-02 -2.33781576e-01 -2.70037800e-01 -1.34759283e+00
-9.80820239e-01 8.16841543e-01 3.93174320e-01 3.51360232e-01
1.02401435e+00 -2.94052005e-01 1.12142551e+00 7.57491946e-01
1.97124615e-01 -7.67380595e-01 3.46558511e-01 8.66743624e-01
9.36797678e-01 -8.43500257e-01 -3.20007145e-01 -1.41766131e-01
-6.12971127e-01 5.64937294e-01 7.99074411e-01 -2.49942854e-01
3.54360342e-01 4.56756167e-03 1.42071694e-01 -7.84764290e-02
-7.93354332e-01 3.60527247e-01 2.79015660e-01 2.50213265e-01
8.54450613e-02 -3.63366380e-02 -4.99792725e-01 1.09853029e+00
-4.68159199e-01 -4.27357256e-01 2.26989210e-01 6.29015267e-01
-3.11502427e-01 -1.20332301e+00 -1.05735138e-01 5.05469978e-01
-8.98859859e-01 -6.84855938e-01 -8.01242471e-01 3.74593049e-01
-2.71670759e-01 1.04008973e+00 -5.83402455e-01 -1.13811624e+00
4.34556514e-01 -1.08941793e-01 -3.83034162e-02 -2.86226541e-01
-1.15654445e+00 -4.93143827e-01 1.50671214e-01 -6.21886373e-01
3.54187608e-01 -1.88372836e-01 -1.00541258e+00 -4.04364377e-01
-4.96085495e-01 3.46322328e-01 5.51070631e-01 8.13784719e-01
5.29069543e-01 6.11070395e-01 1.01989174e+00 -4.00042802e-01
-1.35581517e+00 -8.64109695e-01 -2.21502949e-02 3.30678403e-01
2.07046092e-01 -4.14477706e-01 -6.34769678e-01 -6.96906507e-01] | [6.040886878967285, 8.112648010253906] |
57ce1c13-4b3c-4778-af63-ddad2efcb31a | 2d-human-pose-estimation-with-explicit | 2212.02163 | null | https://arxiv.org/abs/2212.02163v1 | https://arxiv.org/pdf/2212.02163v1.pdf | 2D Human Pose Estimation with Explicit Anatomical Keypoints Structure Constraints | Recently, human pose estimation mainly focuses on how to design a more effective and better deep network structure as human features extractor, and most designed feature extraction networks only introduce the position of each anatomical keypoint to guide their training process. However, we found that some human anatomical keypoints kept their topology invariance, which can help to localize them more accurately when detecting the keypoints on the feature map. But to the best of our knowledge, there is no literature that has specifically studied it. Thus, in this paper, we present a novel 2D human pose estimation method with explicit anatomical keypoints structure constraints, which introduces the topology constraint term that consisting of the differences between the distance and direction of the keypoint-to-keypoint and their groundtruth in the loss object. More importantly, our proposed model can be plugged in the most existing bottom-up or top-down human pose estimation methods and improve their performance. The extensive experiments on the benchmark dataset: COCO keypoint dataset, show that our methods perform favorably against the most existing bottom-up and top-down human pose estimation methods, especially for Lite-HRNet, when our model is plugged into it, its AP scores separately raise by 2.9\% and 3.3\% on COCO val2017 and test-dev2017 datasets. | ['Yuhua Qian', 'Yapeng Chen', 'Ming Zhang', 'Zilong Wang', 'Zhangjian Ji'] | 2022-12-05 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [-4.12713110e-01 1.48477331e-01 -2.74618685e-01 -8.68275538e-02
-4.50279295e-01 -2.03796700e-01 1.74852327e-01 6.96083680e-02
-7.53300071e-01 5.77079535e-01 2.30837047e-01 3.05993229e-01
-2.65973687e-01 -4.82630700e-01 -6.05225086e-01 -5.00518262e-01
-2.44059071e-01 4.26318496e-01 6.19013727e-01 -3.73214841e-01
6.63692364e-03 4.90112036e-01 -1.01021707e+00 -2.49716848e-01
4.25432265e-01 1.13547325e+00 1.21177070e-01 6.02847226e-02
4.94409919e-01 3.88851464e-01 -5.45144022e-01 -4.48194057e-01
3.17551553e-01 -1.12025045e-01 -6.40505612e-01 -4.24628735e-01
3.49407583e-01 -3.65573764e-01 -6.79609478e-01 7.58973718e-01
9.80207860e-01 -5.42580150e-02 4.27492082e-01 -1.18858325e+00
-2.69811898e-01 3.20799887e-01 -7.57775068e-01 1.71607509e-01
2.91234255e-01 -8.47403258e-02 7.46286929e-01 -9.72417593e-01
7.96825171e-01 9.35094237e-01 1.04507029e+00 5.15589237e-01
-5.68501115e-01 -8.18927109e-01 1.25270516e-01 2.90192038e-01
-1.58551490e+00 -7.45598078e-02 1.02197051e+00 -3.46513003e-01
5.91380119e-01 1.84843540e-01 9.29204702e-01 1.07714355e+00
5.24263918e-01 9.52233851e-01 8.87122273e-01 -2.53213465e-01
-1.60609305e-01 -7.94951618e-02 -7.17461854e-02 9.78818119e-01
1.89261869e-01 1.44171342e-01 -5.38627207e-01 -4.02991325e-02
1.10496438e+00 2.09122568e-01 -3.73694986e-01 -7.17420042e-01
-1.37126291e+00 6.13430023e-01 1.12864959e+00 2.54691154e-01
-2.90698498e-01 9.93613154e-02 4.26796377e-01 -3.22969556e-01
1.56309694e-01 4.92175758e-01 -4.76964325e-01 2.50830557e-02
-6.49948120e-01 4.23616111e-01 4.20309037e-01 9.38993454e-01
2.42755935e-01 -4.97079432e-01 -3.25196505e-01 6.38579071e-01
3.09495747e-01 3.15588266e-01 6.14688814e-01 -5.03166020e-01
5.24311781e-01 7.65232623e-01 -8.85278657e-02 -1.48854613e+00
-9.72916067e-01 -6.89637840e-01 -8.71307194e-01 -2.06149459e-01
2.81900257e-01 -2.24794164e-01 -9.29164350e-01 1.66173983e+00
5.38240314e-01 -9.15656984e-02 -5.82524061e-01 1.29955184e+00
1.02265573e+00 2.06357360e-01 -1.09153457e-01 6.67591840e-02
1.55048168e+00 -9.80419695e-01 -6.35046422e-01 -3.35203201e-01
5.24660647e-01 -6.04814827e-01 8.95870507e-01 4.06770915e-01
-8.49034548e-01 -6.34714484e-01 -1.15048230e+00 -5.73608465e-02
-2.97779262e-01 4.91560161e-01 7.88992643e-01 2.55480379e-01
-5.38800836e-01 5.20570993e-01 -9.90181386e-01 -2.23053813e-01
3.15323979e-01 4.42857444e-01 -7.15342104e-01 2.23434642e-01
-1.42523646e+00 1.02308297e+00 3.95684361e-01 6.09975040e-01
-5.50101995e-01 -4.93572742e-01 -7.82640517e-01 -3.97727549e-01
7.36199677e-01 -7.61302948e-01 1.01058996e+00 -2.99570769e-01
-1.32967639e+00 8.03977549e-01 1.97900876e-01 -2.80086815e-01
7.95876920e-01 -6.40739381e-01 -2.77016014e-01 2.45548487e-01
1.75780356e-01 5.51445842e-01 4.14373726e-01 -9.81068254e-01
-5.18456280e-01 -7.90574312e-01 4.42389622e-02 2.47860000e-01
-2.27178276e-01 -2.00363696e-01 -9.17775869e-01 -8.29578161e-01
5.00833213e-01 -1.13137603e+00 -3.34386975e-01 4.51770723e-01
-7.09621608e-01 -2.50284046e-01 6.39866590e-01 -7.53531039e-01
1.40897667e+00 -2.00957656e+00 3.74787420e-01 3.47313344e-01
3.78216147e-01 2.26486176e-01 2.89877385e-01 3.25690567e-01
4.15020771e-02 -1.60900384e-01 5.73723353e-02 -8.77535790e-02
-1.06545709e-01 -8.77571702e-02 2.75859803e-01 8.42972159e-01
-2.88461775e-01 1.03505611e+00 -6.45052612e-01 -7.65947223e-01
2.52156913e-01 5.38831890e-01 -5.45483828e-01 8.04631487e-02
5.25013864e-01 3.18717867e-01 -6.33715749e-01 7.05857038e-01
5.65022945e-01 -8.98137018e-02 -1.63854688e-01 -6.63186014e-01
1.16605937e-01 -1.66204229e-01 -1.15259504e+00 2.04513311e+00
-1.86188698e-01 2.11716473e-01 -1.40047461e-01 -5.51812708e-01
7.14500666e-01 1.36536375e-01 7.05053508e-01 -7.46322691e-01
2.91750282e-01 1.21998154e-01 6.41310886e-02 -3.49541366e-01
1.69188246e-01 5.76154254e-02 -2.86258608e-01 -2.06942096e-01
2.55417544e-02 2.41099879e-01 -1.58124641e-01 1.03661522e-01
8.10292900e-01 3.95864099e-02 4.43098158e-01 -1.47523627e-01
5.19829631e-01 -2.58050948e-01 6.57562077e-01 5.53883970e-01
-4.05818611e-01 7.88556218e-01 4.39590454e-01 -6.24063909e-01
-7.45460987e-01 -1.00014484e+00 -2.68826395e-01 8.44965458e-01
3.77674758e-01 -6.91882133e-01 -7.51303077e-01 -1.03787351e+00
1.16043694e-01 -5.02884276e-02 -8.95003021e-01 -2.99848288e-01
-8.17496479e-01 -4.71601725e-01 6.24786377e-01 7.99373507e-01
6.56215727e-01 -1.01537549e+00 -8.78031611e-01 1.02185803e-02
-2.85987377e-01 -1.05388558e+00 -5.60442269e-01 9.19959322e-02
-5.69568455e-01 -1.21304691e+00 -1.13358736e+00 -7.98641860e-01
7.28217065e-01 -2.58060008e-01 5.49936354e-01 1.10327363e-01
-5.05094528e-01 6.08068556e-02 -4.45616901e-01 -3.38223159e-01
5.61885059e-01 4.12096411e-01 1.02742054e-01 -4.24445063e-01
7.65004829e-02 -2.30827361e-01 -9.85119283e-01 6.87274098e-01
-3.74102503e-01 7.97197223e-02 8.50287497e-01 8.18834305e-01
5.92965066e-01 -9.19952542e-02 1.93930075e-01 -5.77795267e-01
3.70747536e-01 -8.64446685e-02 -1.53354540e-01 1.86058626e-01
-3.32552999e-01 -9.51164663e-02 4.65801805e-01 -3.52770686e-01
-5.21607757e-01 4.23088580e-01 -3.91817212e-01 -3.41240406e-01
-2.24844553e-02 4.07478482e-01 -2.93453902e-01 -3.20555508e-01
5.31858563e-01 1.88725039e-01 -2.49952711e-02 -3.76064062e-01
4.09840979e-02 3.68039250e-01 6.06084049e-01 -3.89624357e-01
8.30125093e-01 4.96270269e-01 1.24947488e-01 -4.49396431e-01
-8.37051988e-01 -4.90871429e-01 -9.55235541e-01 -2.41199866e-01
8.74419749e-01 -7.74225056e-01 -8.52387249e-01 4.72613722e-01
-9.25897300e-01 8.76392499e-02 1.51313588e-01 6.49413764e-01
-5.28003514e-01 4.00063932e-01 -5.06981790e-01 -3.95400107e-01
-6.78099632e-01 -1.27371645e+00 1.27956772e+00 1.74016982e-01
-2.36747101e-01 -6.52170420e-01 6.97615966e-02 1.13238871e-01
2.53508389e-01 5.99894762e-01 6.15710557e-01 -5.62690973e-01
-1.32711187e-01 -7.50157535e-01 -1.49494916e-01 6.48830608e-02
-5.73890321e-02 -2.50595391e-01 -5.28597176e-01 -4.86925393e-01
-1.31515160e-01 -1.77709594e-01 5.55298507e-01 4.76121902e-01
1.26079535e+00 -9.10405517e-02 -7.85847425e-01 8.40872109e-01
1.12224269e+00 -4.40931320e-02 5.81054568e-01 5.51251531e-01
9.20371711e-01 4.88703310e-01 8.00656080e-01 3.27891082e-01
3.63923728e-01 1.04675603e+00 4.86165583e-01 -3.03299963e-01
-6.90289959e-02 -5.08316338e-01 -7.15605095e-02 7.25324392e-01
-3.23254436e-01 1.47080347e-01 -9.11145747e-01 2.92681396e-01
-1.83279634e+00 -5.30212045e-01 -3.98195698e-04 2.04749322e+00
6.12778902e-01 3.91009390e-01 2.64371395e-01 6.21058978e-02
5.15300035e-01 1.79753155e-01 -5.20902276e-01 5.03766298e-01
3.05585116e-01 1.50772288e-01 6.75637543e-01 -1.21540483e-02
-1.37726820e+00 8.52147341e-01 5.79844618e+00 8.49553645e-01
-1.20107186e+00 -5.80210388e-02 1.57630026e-01 -1.91356897e-01
4.02621597e-01 -2.45360538e-01 -9.38728988e-01 4.74182189e-01
1.30442604e-01 1.73664987e-01 -2.09107935e-01 1.01961863e+00
6.31036162e-02 1.63416937e-02 -1.07670212e+00 1.29409933e+00
1.75172180e-01 -8.50713491e-01 -9.40944552e-02 1.42581731e-01
1.48393303e-01 -3.39179337e-01 -6.23311624e-02 4.18468475e-01
-4.52259004e-01 -1.00758874e+00 6.57170892e-01 6.26693904e-01
4.97142017e-01 -1.12049508e+00 1.02532709e+00 3.93650651e-01
-1.42549109e+00 -2.06813831e-02 -3.58510584e-01 1.31619796e-01
1.10193238e-01 2.52927512e-01 -7.29362309e-01 7.41693854e-01
9.24152732e-01 6.36248112e-01 -8.88663054e-01 1.32005465e+00
-4.26640928e-01 6.15808405e-02 -3.25207561e-01 -9.65112671e-02
1.34282887e-01 5.20991981e-01 3.44292372e-01 9.36378479e-01
1.72315314e-01 -2.15710301e-04 4.70885098e-01 4.42194015e-01
9.74090248e-02 4.04222250e-01 -2.71165729e-01 4.39660668e-01
3.26782435e-01 1.40176940e+00 -7.98914671e-01 1.19142666e-01
-4.24820296e-02 1.05212760e+00 3.53006840e-01 3.62276360e-02
-1.18162930e+00 -7.52803922e-01 2.09897771e-01 5.88813424e-01
5.82366101e-02 -7.72852227e-02 2.94199623e-02 -9.95935798e-01
2.80419260e-01 -7.82391250e-01 4.79185939e-01 -7.03430116e-01
-1.10799444e+00 4.86684382e-01 1.87941477e-01 -1.26307523e+00
-1.39608830e-01 -8.19325447e-01 -3.63083333e-01 5.51677883e-01
-8.46275568e-01 -1.28314209e+00 -5.79065323e-01 7.23334551e-01
1.65049180e-01 5.50703034e-02 5.62626898e-01 3.20142090e-01
-5.55041075e-01 1.01674378e+00 -3.66630763e-01 7.09602773e-01
9.37847316e-01 -9.66978610e-01 2.67704010e-01 5.97912669e-01
-5.55502623e-02 9.02910113e-01 5.29253721e-01 -6.62050903e-01
-1.34814382e+00 -6.49339259e-01 4.53520089e-01 -5.15886843e-01
3.63309711e-01 -5.69626331e-01 -5.82344115e-01 6.23096943e-01
-3.12308848e-01 3.45760912e-01 3.44098926e-01 2.10705012e-01
-5.11410572e-02 -1.35727823e-01 -1.03926349e+00 5.74090958e-01
1.22585487e+00 -1.60934538e-01 -6.79211855e-01 3.39568585e-01
6.24355793e-01 -8.28970015e-01 -9.93480325e-01 9.34704125e-01
1.04263473e+00 -7.93125987e-01 1.19785285e+00 -3.30925018e-01
2.95787334e-01 -3.40273589e-01 8.65344778e-02 -1.15986633e+00
-4.93415177e-01 -1.79927662e-01 -1.21172138e-01 8.25216770e-01
1.08537629e-01 -3.61167043e-01 1.02407408e+00 1.73361495e-01
-8.14768076e-02 -1.34312820e+00 -1.17428589e+00 -5.90271771e-01
-1.69409416e-03 -9.02838483e-02 4.05772269e-01 7.25143611e-01
9.25325206e-04 1.26536340e-01 -5.68140328e-01 2.01466709e-01
4.85846430e-01 -2.63651043e-01 9.39598560e-01 -1.26200688e+00
-2.11360216e-01 -3.55749965e-01 -8.71066391e-01 -1.08949721e+00
-2.01528758e-01 -6.96169198e-01 2.83626141e-03 -1.57883692e+00
2.23269224e-01 -2.81717151e-01 -3.39466155e-01 6.27650976e-01
-1.39124557e-01 1.81430638e-01 1.21912919e-01 1.61833763e-01
-5.08508801e-01 5.25845587e-01 1.56929362e+00 1.41167164e-01
-1.01604015e-01 9.19794757e-03 -5.47496259e-01 9.71135855e-01
4.47254717e-01 -5.29530168e-01 -3.03060561e-01 -1.33120745e-01
1.93843797e-01 -2.47901976e-02 6.18657112e-01 -1.35065389e+00
5.26217699e-01 1.17281370e-01 9.15883839e-01 -8.72631729e-01
4.20610607e-01 -8.65163386e-01 7.99493641e-02 7.64580011e-01
-1.36887357e-01 1.08658276e-01 4.88345809e-02 3.45847696e-01
-1.29727945e-01 1.16165847e-01 7.14830697e-01 -3.55407208e-01
-7.39477277e-01 6.51690900e-01 2.98016399e-01 1.48505479e-01
1.34652126e+00 -2.19748333e-01 -2.84997933e-02 -1.73775196e-01
-7.93378651e-01 2.77534842e-01 2.03916177e-01 5.25336623e-01
8.60526800e-01 -1.45387816e+00 -4.74468082e-01 1.15490369e-01
2.23298460e-01 1.96941659e-01 3.17392141e-01 1.13797843e+00
-6.47993326e-01 4.97575819e-01 -2.92119414e-01 -6.38565421e-01
-1.20794320e+00 4.67696041e-01 6.75197482e-01 -3.20734769e-01
-8.42801929e-01 8.98440301e-01 3.41745615e-01 -5.90402782e-01
4.63823289e-01 -2.17936993e-01 -2.11716905e-01 4.70869951e-02
3.66287798e-01 2.98971593e-01 6.40682951e-02 -8.30248892e-01
-8.73401225e-01 9.66028452e-01 -1.94816396e-01 1.71611935e-01
1.27648449e+00 2.26115882e-01 8.47209468e-02 2.49751255e-01
1.34576750e+00 9.77668464e-02 -9.93428409e-01 -9.71545354e-02
-3.63279909e-01 -3.89185607e-01 -1.37774184e-01 -9.05688941e-01
-1.26941121e+00 7.92964697e-01 9.57937598e-01 -3.86164904e-01
9.42632198e-01 1.10063389e-01 9.63641047e-01 3.14114302e-01
7.10057437e-01 -1.09112227e+00 1.80106491e-01 3.09853584e-01
1.04588437e+00 -1.05568933e+00 3.14232469e-01 -5.08650362e-01
-5.42162836e-01 9.59815383e-01 1.08879805e+00 -4.38082755e-01
7.98392296e-01 1.10993229e-01 -9.49666426e-02 -5.02674580e-01
-5.64017445e-02 1.34974346e-02 7.68365562e-01 4.13285524e-01
4.66384709e-01 3.52417901e-02 -5.46822369e-01 9.45719600e-01
-5.38051963e-01 -1.12794004e-01 -1.27856150e-01 1.04100561e+00
-4.55754668e-01 -7.30178535e-01 -3.05608898e-01 2.90033072e-01
-5.08783102e-01 1.29651055e-01 -3.87393057e-01 1.41535330e+00
3.64762306e-01 2.37863511e-01 -2.83198416e-01 -6.46174371e-01
8.96729887e-01 -2.70533621e-01 4.87689525e-01 -3.94936651e-01
-5.93351007e-01 1.24334216e-01 -2.52189428e-01 -8.78222466e-01
-6.21349365e-02 -2.84197956e-01 -1.44406044e+00 -4.34909724e-02
-5.63272715e-01 5.40751703e-02 4.88677204e-01 8.26239645e-01
1.12605631e-01 5.17350197e-01 2.65598744e-01 -9.56528306e-01
-4.15776908e-01 -9.10818040e-01 -4.87387121e-01 4.44076300e-01
1.89371426e-02 -1.18868923e+00 -1.08879022e-01 -6.48339450e-01] | [7.133866310119629, -0.7379562258720398] |
c9b94a00-7a9e-45e5-ae93-56f2f025a7c4 | visual-composite-set-detection-using-part-and | 2105.02170 | null | https://arxiv.org/abs/2105.02170v2 | https://arxiv.org/pdf/2105.02170v2.pdf | Visual Relationship Detection Using Part-and-Sum Transformers with Composite Queries | Computer vision applications such as visual relationship detection and human object interaction can be formulated as a composite (structured) set detection problem in which both the parts (subject, object, and predicate) and the sum (triplet as a whole) are to be detected in a hierarchical fashion. In this paper, we present a new approach, denoted Part-and-Sum detection Transformer (PST), to perform end-to-end visual composite set detection. Different from existing Transformers in which queries are at a single level, we simultaneously model the joint part and sum hypotheses/interactions with composite queries and attention modules. We explicitly incorporate sum queries to enable better modeling of the part-and-sum relations that are absent in the standard Transformers. Our approach also uses novel tensor-based part queries and vector-based sum queries, and models their joint interaction. We report experiments on two vision tasks, visual relationship detection and human object interaction and demonstrate that PST achieves state of the art results among single-stage models, while nearly matching the results of custom designed two-stage models. | ['Stefano Soatto', 'Vijay Mahadevan', 'Yuting Zhang', 'Haofu Liao', 'Zhuowen Tu', 'Qi Dong'] | 2021-05-05 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Dong_Visual_Relationship_Detection_Using_Part-and-Sum_Transformers_With_Composite_Queries_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Dong_Visual_Relationship_Detection_Using_Part-and-Sum_Transformers_With_Composite_Queries_ICCV_2021_paper.pdf | iccv-2021-1 | ['visual-relationship-detection'] | ['computer-vision'] | [ 2.97814757e-01 1.29502654e-01 1.11415938e-01 -3.62825274e-01
-4.98534232e-01 -6.21442139e-01 8.95906270e-01 3.51516396e-01
-2.29526371e-01 -1.76065508e-02 2.48720441e-02 -1.91818058e-01
6.22421838e-02 -2.87896901e-01 -6.57817423e-01 -3.04656476e-01
-1.29562232e-03 8.39748204e-01 9.37757969e-01 -1.66240662e-01
2.89439429e-02 4.91080374e-01 -1.94748628e+00 8.00528824e-01
5.44977486e-01 9.90342915e-01 2.43089840e-01 8.94341111e-01
-3.62013578e-02 1.07602179e+00 -4.90638852e-01 -7.02398002e-01
1.46858424e-01 -1.86518654e-01 -8.85987341e-01 4.99738634e-01
1.05452776e+00 -5.39268374e-01 -1.01426505e-01 9.49137747e-01
3.72606516e-01 -9.25707072e-02 6.55094266e-01 -1.63231707e+00
-6.72418058e-01 5.74808359e-01 -9.60535407e-01 1.98474407e-01
6.44644976e-01 2.55395591e-01 1.42562425e+00 -1.34243429e+00
5.71393192e-01 1.82373559e+00 2.49700293e-01 1.42085642e-01
-1.53641033e+00 -3.60247225e-01 5.04903555e-01 4.24267977e-01
-1.23410559e+00 -5.44992328e-01 6.27370834e-01 -7.39623308e-01
1.62000024e+00 4.73034441e-01 7.78665185e-01 4.95260179e-01
-1.06472701e-01 1.16736543e+00 6.50648057e-01 -4.74204063e-01
-1.05939850e-01 2.85301864e-01 4.65029776e-01 9.08353269e-01
2.59180337e-01 -4.00743112e-02 -5.89729786e-01 -2.56917119e-01
6.49687588e-01 -1.06520541e-02 -1.38241753e-01 -8.85420918e-01
-1.36993361e+00 4.83852208e-01 5.58686137e-01 4.95986938e-02
-3.80287826e-01 3.37258905e-01 3.02275717e-01 1.22673571e-01
1.69033408e-01 -2.78864894e-02 -4.07826424e-01 5.01694918e-01
-6.27189875e-01 4.33883786e-01 8.66861522e-01 1.28861737e+00
6.75051749e-01 -3.66006404e-01 -8.68129551e-01 8.79360020e-01
6.26160383e-01 2.09926322e-01 -7.07044452e-03 -8.61213923e-01
3.49651456e-01 9.52671647e-01 1.58486202e-01 -8.94988954e-01
-1.19243570e-01 -5.61709583e-01 -4.30441111e-01 3.36032450e-01
2.79719144e-01 5.39708197e-01 -1.10960996e+00 1.67056584e+00
7.57827401e-01 -2.53929466e-01 -3.25154662e-01 9.14224505e-01
1.24117088e+00 1.99932709e-01 2.87232220e-01 -9.00236145e-02
2.07504606e+00 -1.31565630e+00 -5.79851329e-01 -3.13997865e-01
4.80524868e-01 -7.65381575e-01 8.75172973e-01 2.57900894e-01
-1.46214104e+00 -7.91431487e-01 -7.67863452e-01 -7.12164760e-01
-3.26310039e-01 2.07315564e-01 5.92567563e-01 1.72228932e-01
-1.33731425e+00 1.31125198e-04 -7.13939369e-01 -5.57816863e-01
4.56440717e-01 3.94040912e-01 -3.70526195e-01 -4.14187647e-03
-5.33518374e-01 9.28126752e-01 2.99680501e-01 1.04881957e-01
-9.54450071e-01 -3.44605982e-01 -8.76663089e-01 2.88268059e-01
5.96899450e-01 -1.26457024e+00 1.55018461e+00 -6.14807963e-01
-8.25591207e-01 1.29116118e+00 -4.45809484e-01 -2.03132823e-01
3.87501448e-01 -2.14017630e-01 2.26562768e-01 2.58759499e-01
2.53385693e-01 1.02562749e+00 8.92470896e-01 -1.54305005e+00
-7.72718370e-01 -4.86253619e-01 2.82905072e-01 5.07084846e-01
-1.11582853e-01 6.15205228e-01 -1.07788575e+00 -3.44801426e-01
2.17215136e-01 -6.22658134e-01 -2.39418335e-02 7.66236007e-01
-5.86716294e-01 -6.31029367e-01 9.26539540e-01 -5.00872433e-01
9.93983984e-01 -2.05484772e+00 3.96913767e-01 -4.35027629e-02
8.38146985e-01 2.09253699e-01 -1.58322036e-01 3.93116564e-01
-1.65183350e-01 -8.66729096e-02 -8.25643092e-02 -9.48435903e-01
7.34299123e-02 2.23368064e-01 2.64882203e-02 1.63659960e-01
4.69745517e-01 1.02625608e+00 -8.79696548e-01 -9.11612153e-01
1.71462938e-01 4.14538175e-01 -4.94987190e-01 2.95187652e-01
-4.59109217e-01 -8.21930692e-02 -1.63481429e-01 1.03704298e+00
5.71024239e-01 -5.85027158e-01 1.33213643e-02 -7.84368217e-01
-3.06128431e-02 1.36872306e-01 -1.04830706e+00 1.14274895e+00
1.78054482e-01 5.51811278e-01 2.46155396e-01 -8.00530553e-01
4.95472163e-01 1.79654464e-01 2.90516853e-01 -5.82697034e-01
-1.23067036e-01 -2.91423112e-01 3.68491143e-01 -5.49810052e-01
4.16329026e-01 3.28109801e-01 1.97205350e-01 6.33774400e-02
1.86160371e-01 -8.84318724e-02 4.88780081e-01 6.33487165e-01
1.03763640e+00 3.93781513e-01 5.66652954e-01 1.16131842e-01
4.26564723e-01 1.51082203e-01 1.97573811e-01 8.82622242e-01
-3.32639307e-01 5.97831905e-01 7.22139001e-01 -1.54874325e-01
-9.03429627e-01 -1.03797197e+00 2.62084782e-01 1.42597616e+00
2.34208122e-01 -5.78376293e-01 -2.52264887e-01 -5.91024995e-01
1.71381265e-01 4.75999504e-01 -5.48048735e-01 8.05376545e-02
-4.77537900e-01 -4.30464387e-01 1.11518912e-01 8.06484640e-01
2.33163491e-01 -1.06864977e+00 -9.39691901e-01 1.53545633e-01
-1.85826108e-01 -1.50283682e+00 -7.54500747e-01 1.67634085e-01
-4.22360301e-01 -1.20206690e+00 -4.39376682e-01 -8.47481430e-01
5.56469440e-01 5.25993228e-01 1.34846151e+00 1.18331231e-01
-5.97021341e-01 6.44716680e-01 -1.54937193e-01 -5.63716412e-01
-1.07859671e-02 -7.53442526e-01 -3.15353036e-01 2.31552288e-01
1.38316616e-01 -2.62471318e-01 -5.76584280e-01 3.71460348e-01
-6.62456036e-01 3.16742688e-01 7.93076932e-01 7.79743969e-01
5.38251519e-01 -3.27747524e-01 -2.58017093e-01 -5.53904712e-01
2.47126162e-01 -4.97379079e-02 -5.60305059e-01 5.31833887e-01
-2.34168038e-01 -1.35659456e-01 2.84862053e-02 -7.29287565e-01
-9.08247530e-01 3.38746816e-01 2.78255403e-01 -8.90779138e-01
-6.51923418e-02 2.59073913e-01 -2.47122779e-01 -8.02822262e-02
4.86297816e-01 2.36751199e-01 -2.14145944e-01 -4.90673393e-01
5.66292048e-01 2.68106729e-01 5.01024485e-01 -2.97369301e-01
5.52104473e-01 5.93089461e-01 6.95978384e-03 -6.70230448e-01
-7.40855396e-01 -8.94969642e-01 -7.35846400e-01 -3.19504708e-01
9.21740830e-01 -1.05917084e+00 -1.16819417e+00 4.32835519e-01
-1.71170163e+00 -7.37795979e-02 -3.44325036e-01 1.48515880e-01
-3.07623237e-01 4.86390471e-01 -6.26105785e-01 -1.13130844e+00
-3.15727592e-01 -1.18771946e+00 1.67233789e+00 3.92889678e-02
-1.61712974e-01 -5.16275942e-01 -3.51298541e-01 4.30126786e-01
1.31232530e-01 1.14891231e-01 9.33450103e-01 -7.77808785e-01
-1.00390160e+00 2.49077920e-02 -7.29635954e-01 1.21485628e-01
-2.10359201e-01 8.14734474e-02 -8.94246340e-01 -2.59409428e-01
-1.01846434e-01 -3.73494148e-01 1.06580400e+00 1.50641501e-01
7.96741247e-01 -1.57643467e-01 -6.48823082e-01 2.29606688e-01
1.23081839e+00 1.64965972e-01 3.51597250e-01 -1.87411979e-01
1.10752189e+00 8.31320703e-01 5.35551667e-01 3.84334743e-01
7.96542883e-01 9.69980896e-01 7.00995445e-01 -4.03342694e-01
-5.68824649e-01 1.00739576e-01 3.20079207e-01 1.01914860e-01
5.31578436e-02 -3.79090697e-01 -9.22333598e-01 6.82184517e-01
-2.16221714e+00 -1.07822311e+00 -5.45899332e-01 1.96027005e+00
6.34019792e-01 2.76388973e-01 4.48803633e-01 1.03896363e-02
6.67269349e-01 -1.07958294e-01 -5.12524009e-01 -2.36591443e-01
-4.14417982e-02 -1.13392927e-01 2.10662223e-02 4.32770491e-01
-1.05115485e+00 9.99945521e-01 6.43474579e+00 4.66015399e-01
-5.71085393e-01 1.31697223e-01 3.25070202e-01 2.66141351e-02
-9.79467034e-02 2.77756959e-01 -1.05674219e+00 -1.80404842e-01
1.10562995e-01 2.44905457e-01 1.74235493e-01 7.09928453e-01
-4.63528596e-02 -3.62710744e-01 -1.73329973e+00 1.13354516e+00
2.71726638e-01 -8.70661676e-01 3.74657750e-01 5.74484393e-02
-1.78660527e-02 -1.69024602e-01 -6.98812604e-02 1.97363630e-01
2.92342275e-01 -5.87586045e-01 1.23392582e+00 6.13824427e-01
5.99857450e-01 -4.55363840e-02 2.88242579e-01 2.60825187e-01
-1.66986525e+00 -9.08867344e-02 4.68363352e-02 -8.81501138e-02
2.16914732e-02 3.10627669e-01 -9.51620221e-01 4.96261418e-01
8.60606611e-01 6.43071830e-01 -7.73273408e-01 1.24611485e+00
-2.81723868e-02 1.27658680e-01 -4.40066129e-01 -3.42753641e-02
-1.50784463e-01 1.29280061e-01 8.81116331e-01 1.19756711e+00
-3.27605903e-01 3.22347023e-02 6.09429955e-01 1.15489316e+00
2.08069265e-01 -1.71542659e-01 -3.45024645e-01 1.25791833e-01
3.83938402e-01 1.26286304e+00 -7.42015541e-01 -7.32890606e-01
-6.29569352e-01 1.08889961e+00 5.05684376e-01 4.44645822e-01
-7.09281266e-01 -9.74409655e-02 5.12534440e-01 1.37798637e-01
7.74370730e-01 -1.93344995e-01 -8.81659314e-02 -1.10452044e+00
3.82471323e-01 -7.44349122e-01 6.38868928e-01 -1.02541089e+00
-1.33418179e+00 4.29243773e-01 4.89985198e-01 -1.05052376e+00
-9.86273065e-02 -6.81053817e-01 -5.35209239e-01 8.33811402e-01
-1.27457750e+00 -1.76248658e+00 -3.26432616e-01 8.58712435e-01
5.46462595e-01 3.57326031e-01 5.53641319e-01 4.73535620e-02
-5.87086022e-01 4.44001555e-01 -7.58162618e-01 1.34295419e-01
4.04182315e-01 -1.37307787e+00 3.37804824e-01 9.51129138e-01
1.90375343e-01 7.28536546e-01 7.97450602e-01 -6.92861199e-01
-1.32073772e+00 -8.92517686e-01 1.06863117e+00 -6.56996131e-01
5.04580200e-01 -7.03044116e-01 -7.75922596e-01 8.72275233e-01
3.53310913e-01 2.63925344e-01 2.48174086e-01 6.34336397e-02
-6.97624207e-01 -7.25058243e-02 -1.01741266e+00 6.50001943e-01
1.34040475e+00 -6.15817785e-01 -6.87333107e-01 4.99385297e-01
9.58029926e-01 -2.98165232e-01 -4.83355582e-01 6.06339097e-01
6.94244146e-01 -1.16468954e+00 1.28020775e+00 -4.32203084e-01
1.68028861e-01 -6.18879020e-01 -1.45855278e-01 -5.45937061e-01
-6.96042657e-01 -4.31972116e-01 -6.56816721e-01 1.16527379e+00
1.80085689e-01 -3.73481244e-01 4.27397549e-01 5.47538161e-01
-1.74375981e-01 -6.65730059e-01 -6.00275040e-01 -5.61664343e-01
-7.95591056e-01 -1.85344085e-01 6.80181012e-02 5.77954888e-01
-1.54973924e-01 8.41633737e-01 -1.37946397e-01 2.20087066e-01
9.30022597e-01 2.60481566e-01 7.59380102e-01 -1.19515455e+00
-8.70948017e-01 -6.66454852e-01 -4.65136200e-01 -1.17467046e+00
-3.37211668e-01 -6.64676011e-01 7.74189681e-02 -1.90745103e+00
5.95262170e-01 8.64512026e-02 -1.18255615e-01 9.02268767e-01
-3.52999479e-01 2.00591475e-01 3.74618173e-01 3.56370628e-01
-8.83967936e-01 4.52681422e-01 1.20051014e+00 -3.78006399e-01
-1.74101889e-01 -1.85490623e-01 -6.80715561e-01 7.18049765e-01
3.93753842e-04 -2.75437742e-01 -4.21361178e-01 -4.45910990e-01
-8.15463215e-02 1.65583745e-01 9.57305551e-01 -7.02422976e-01
6.47994161e-01 9.87429246e-02 3.94384950e-01 -1.07975137e+00
7.81853497e-01 -8.12960684e-01 -3.00639607e-02 4.40835118e-01
-3.25045705e-01 1.66565016e-01 1.41879663e-01 5.31594515e-01
-1.33833051e-01 3.89579833e-01 6.09122455e-01 -2.27474183e-01
-7.56262898e-01 2.41204798e-01 -6.34544492e-02 -3.05944771e-01
1.08191597e+00 -4.18006897e-01 -3.68889153e-01 -1.54676378e-01
-1.09290683e+00 5.57426989e-01 1.83872581e-01 2.91021794e-01
1.00824940e+00 -1.04164612e+00 -8.53157818e-01 -1.16080213e-02
6.51456773e-01 1.09249331e-01 5.24711683e-02 1.04613876e+00
-6.87796697e-02 1.62495330e-01 4.22384590e-02 -1.09109795e+00
-1.89921296e+00 1.01091719e+00 3.18754822e-01 -2.33351558e-01
-4.94182050e-01 1.23895919e+00 7.00315058e-01 -5.82017712e-02
5.94762504e-01 -4.72592860e-01 -3.25491011e-01 1.75529331e-01
4.87248421e-01 -9.73560847e-03 -1.96394205e-01 -7.72238314e-01
-6.99389517e-01 6.45976245e-01 -2.78301895e-01 -1.50664106e-01
8.78544807e-01 -2.04821482e-01 -4.75024611e-01 4.86536622e-01
9.31183100e-01 -2.37429827e-01 -8.34275663e-01 -5.71368754e-01
-1.27399698e-01 -3.32242101e-01 -2.36138448e-01 -8.89264345e-01
-6.98566735e-01 8.35339129e-01 4.55914110e-01 3.19206297e-01
1.14223182e+00 6.62429094e-01 2.44548693e-01 2.65171766e-01
9.92561653e-02 -4.49766368e-01 3.71081829e-01 3.20655525e-01
1.26945877e+00 -1.16558588e+00 4.37837839e-02 -1.05821359e+00
-5.82301259e-01 7.36935198e-01 1.03047228e+00 6.32086396e-02
5.62971532e-01 3.88349652e-01 -4.11201984e-01 -5.43242991e-01
-1.29006910e+00 -9.60296631e-01 7.08786845e-01 5.16577840e-01
5.01073956e-01 -1.28807053e-01 4.08147201e-02 -7.01092854e-02
3.36825728e-01 -3.14286232e-01 6.71700388e-02 1.11626625e+00
-3.71366709e-01 -1.00414193e+00 -4.15775776e-01 4.28387105e-01
-1.78787902e-01 -3.25568736e-01 -7.55861998e-01 6.95514739e-01
3.28495353e-01 1.07665491e+00 1.38881788e-01 -3.64407897e-01
5.64536750e-01 -8.97316635e-02 7.28083372e-01 -9.04506207e-01
-8.58177125e-01 4.17144537e-01 3.40358734e-01 -7.88059831e-01
-5.95458746e-01 -6.77011490e-01 -1.06314945e+00 3.89832318e-01
-4.87062186e-01 -4.29213405e-01 3.86399448e-01 8.88673842e-01
3.44493628e-01 6.81236327e-01 9.31022391e-02 -1.01646662e+00
-5.43232441e-01 -9.96722579e-01 -2.09993243e-01 6.68214619e-01
6.24086976e-01 -7.56016254e-01 1.59575939e-02 2.26609439e-01] | [10.058732986450195, 1.557834267616272] |
03a3cbe8-0916-4b0d-add7-ddb199a1b68f | recurrent-models-of-visual-attention | 1406.6247 | null | http://arxiv.org/abs/1406.6247v1 | http://arxiv.org/pdf/1406.6247v1.pdf | Recurrent Models of Visual Attention | Applying convolutional neural networks to large images is computationally
expensive because the amount of computation scales linearly with the number of
image pixels. We present a novel recurrent neural network model that is capable
of extracting information from an image or video by adaptively selecting a
sequence of regions or locations and only processing the selected regions at
high resolution. Like convolutional neural networks, the proposed model has a
degree of translation invariance built-in, but the amount of computation it
performs can be controlled independently of the input image size. While the
model is non-differentiable, it can be trained using reinforcement learning
methods to learn task-specific policies. We evaluate our model on several image
classification tasks, where it significantly outperforms a convolutional neural
network baseline on cluttered images, and on a dynamic visual control problem,
where it learns to track a simple object without an explicit training signal
for doing so. | ['Koray Kavukcuoglu', 'Volodymyr Mnih', 'Nicolas Heess', 'Alex Graves'] | 2014-06-24 | recurrent-models-of-visual-attention-1 | http://papers.nips.cc/paper/5542-recurrent-models-of-visual-attention | http://papers.nips.cc/paper/5542-recurrent-models-of-visual-attention.pdf | neurips-2014-12 | ['hard-attention'] | ['methodology'] | [ 6.61603153e-01 8.87480974e-02 -4.00718212e-01 -2.24940702e-01
-5.31830072e-01 -6.40706956e-01 6.29491329e-01 -2.04684481e-01
-8.77071381e-01 4.33447242e-01 -8.62268955e-02 -4.01294917e-01
1.61638007e-01 -6.83223605e-01 -1.06301725e+00 -7.49844849e-01
-1.06207319e-01 3.20304871e-01 3.97943676e-01 -7.53692016e-02
3.48080546e-01 6.97700262e-01 -1.18217492e+00 3.77086520e-01
1.63735017e-01 1.22956491e+00 4.55237955e-01 9.24671650e-01
3.63334268e-01 1.12457883e+00 -4.07093078e-01 3.27745140e-01
4.39300448e-01 -3.70558172e-01 -7.36683428e-01 4.51021612e-01
7.34353840e-01 -4.01557952e-01 -4.87387896e-01 9.65401471e-01
3.57108861e-01 2.33983338e-01 5.49814701e-01 -8.05958688e-01
-7.75932193e-01 2.25477651e-01 -3.62340838e-01 7.14504719e-01
-2.16854289e-01 2.27433816e-01 7.95181990e-01 -8.34165990e-01
6.98827922e-01 1.28799355e+00 5.42220354e-01 4.53759223e-01
-1.59216630e+00 -4.10106182e-01 5.26884675e-01 -3.31618525e-02
-8.10638070e-01 -4.46844846e-01 5.39034426e-01 -3.61248285e-01
1.30058742e+00 -1.52686849e-01 7.43661582e-01 1.04626012e+00
1.57343999e-01 8.17104578e-01 1.06095195e+00 -4.02160347e-01
3.03514600e-01 -3.80144387e-01 -2.21341848e-01 8.94177735e-01
-9.51864123e-02 3.93400222e-01 -3.84470403e-01 6.40692040e-02
1.53167713e+00 1.89217776e-01 -1.44071490e-01 -4.90281880e-01
-1.43732548e+00 7.64687896e-01 7.59501874e-01 1.86418831e-01
-4.50040728e-01 7.99100399e-01 3.55883777e-01 5.16529441e-01
2.61591911e-01 5.86701214e-01 -7.24549770e-01 1.04412906e-01
-8.90687883e-01 2.28891701e-01 3.55170548e-01 7.14113235e-01
7.42104709e-01 4.67002571e-01 -3.57397765e-01 6.40160084e-01
-2.16768906e-01 4.25941944e-01 8.23857009e-01 -1.32916367e+00
2.90694177e-01 3.78968179e-01 2.55420089e-01 -7.68285751e-01
-5.02956092e-01 -3.03373158e-01 -9.21857595e-01 6.88897133e-01
5.34221172e-01 -3.76725078e-01 -1.29363859e+00 1.72005713e+00
-1.57370880e-01 8.81791040e-02 -5.99852838e-02 9.14409459e-01
1.91814572e-01 7.01950490e-01 -4.76519167e-02 -3.37884612e-02
1.26295578e+00 -1.18905127e+00 -4.45302635e-01 -6.20046079e-01
2.30855882e-01 -3.64995360e-01 9.26994801e-01 3.07039887e-01
-1.29846871e+00 -7.58541524e-01 -1.01148355e+00 -2.19434798e-01
-4.33571398e-01 4.50218856e-01 5.55401742e-01 9.21094343e-02
-1.36310327e+00 6.76559091e-01 -1.06001759e+00 -1.77087680e-01
4.60094690e-01 6.28071487e-01 -2.07550719e-01 2.48050377e-01
-7.68212736e-01 8.38184834e-01 4.58622545e-01 2.09359065e-01
-1.19816875e+00 -5.53123236e-01 -9.33612823e-01 2.96284378e-01
3.47402513e-01 -6.35069370e-01 1.46054447e+00 -1.75408685e+00
-1.66530359e+00 7.12103605e-01 -1.20058447e-01 -8.25946987e-01
4.68882948e-01 -2.37654388e-01 -7.81731531e-02 4.48946506e-01
-1.50072843e-01 1.15276098e+00 1.41562808e+00 -8.03126812e-01
-7.44654536e-01 -2.28535056e-01 2.47544914e-01 2.00355247e-01
-3.05526238e-02 1.72004327e-01 -4.11391407e-01 -1.00960124e+00
1.30362093e-01 -1.22454321e+00 -6.92120135e-01 3.59327406e-01
1.09962877e-02 1.94553321e-03 1.14535987e+00 -3.51465046e-01
5.95455706e-01 -2.02164793e+00 3.53146613e-01 5.35683930e-02
4.21969332e-02 2.53112048e-01 -3.98708314e-01 2.56156120e-02
-1.69240296e-01 3.66288200e-02 -1.84326828e-01 1.21175818e-01
-4.41869199e-01 1.50262401e-01 -5.02676129e-01 5.44505954e-01
6.28111005e-01 1.18171632e+00 -8.80580425e-01 -9.94291380e-02
2.55920380e-01 5.07036805e-01 -7.15067387e-01 1.27824962e-01
-5.04982769e-01 4.08851653e-01 -3.89140069e-01 3.78690302e-01
2.27622762e-01 -8.18166614e-01 3.63463640e-01 1.50406957e-02
-3.92378820e-03 9.23136249e-02 -9.21736896e-01 1.64105844e+00
-4.88057733e-01 1.15244436e+00 1.05272904e-01 -1.28372705e+00
6.40225232e-01 1.86148033e-01 4.76013720e-01 -8.48136067e-01
8.15586839e-03 -2.05544457e-01 1.04780756e-02 -4.04358834e-01
3.89067441e-01 1.70086101e-01 -2.20523831e-02 5.31741679e-01
1.14076078e-01 1.73830241e-01 1.57294631e-01 -1.12860091e-01
1.24130070e+00 2.71908045e-01 2.72020638e-01 -1.33646592e-01
2.01560423e-01 1.48137072e-02 4.80786741e-01 8.87605548e-01
-1.63556650e-01 5.21860540e-01 4.89999503e-01 -1.00601995e+00
-1.20922565e+00 -8.91446292e-01 1.16698034e-01 1.41750646e+00
4.32772674e-02 4.76543512e-03 -6.02733433e-01 -6.16057992e-01
-2.16813218e-02 8.67448524e-02 -8.17694247e-01 -4.57434542e-02
-1.00012839e+00 -4.03770149e-01 3.44310373e-01 8.33809137e-01
6.95527256e-01 -1.59258854e+00 -1.29113328e+00 3.36413950e-01
6.08843863e-02 -1.28942680e+00 -6.06541395e-01 4.96451855e-01
-1.06683838e+00 -1.07863581e+00 -6.48301184e-01 -1.08605456e+00
1.02169609e+00 1.96661785e-01 1.22777474e+00 6.13013804e-02
-2.99613953e-01 2.97879368e-01 -4.53659967e-02 -3.65362227e-01
-1.36191454e-02 2.41422728e-01 -8.24271739e-02 -8.08620378e-02
7.04554990e-02 -3.17609400e-01 -8.03595126e-01 2.80833334e-01
-9.12288189e-01 9.94274989e-02 8.14479351e-01 1.21575570e+00
7.42034912e-01 -4.48451415e-02 4.49481040e-01 -9.73707438e-01
4.77342218e-01 3.45032774e-02 -9.79487002e-01 4.00706902e-02
-3.87289971e-01 3.19594234e-01 7.89178491e-01 -8.05241108e-01
-7.79401362e-01 4.69262987e-01 3.56354922e-01 -7.11820126e-01
-1.36632130e-01 2.91498601e-01 4.73091006e-01 -2.11558521e-01
6.73865855e-01 3.09600294e-01 8.84405449e-02 -2.26987958e-01
4.83584315e-01 -3.93540375e-02 6.81491017e-01 -3.36057127e-01
6.42564297e-01 5.05734444e-01 -2.64690947e-02 -6.08653903e-01
-7.60104954e-01 -7.20278919e-02 -9.62152243e-01 2.02680333e-03
8.64403725e-01 -9.23439264e-01 -7.98440158e-01 2.91787773e-01
-1.03325367e+00 -8.71932209e-01 -3.69219095e-01 3.93287003e-01
-8.86542678e-01 -1.79568172e-01 -7.85237789e-01 -4.25490230e-01
-1.01251811e-01 -1.16000795e+00 9.48905110e-01 1.02580376e-01
-3.29531357e-02 -9.30925250e-01 -2.53225058e-01 -2.26795062e-01
6.62549138e-01 3.02991390e-01 7.51390040e-01 -1.74217775e-01
-1.01546800e+00 -1.86188832e-01 -2.99739897e-01 2.75594890e-01
1.08287297e-01 -1.13722950e-01 -6.96216583e-01 -6.06725037e-01
-2.36576989e-01 -6.28193140e-01 1.24234223e+00 7.96838760e-01
1.61115110e+00 -7.03579545e-01 -2.34314874e-01 6.25410736e-01
1.39356911e+00 1.07972071e-01 5.05885243e-01 4.47194606e-01
4.08977121e-01 2.79404849e-01 4.26678389e-01 7.95000419e-02
1.03982175e-02 6.10430717e-01 5.79998553e-01 -4.21123296e-01
-1.22989520e-01 -1.19815603e-01 3.66951734e-01 1.57048360e-01
-3.25925708e-01 1.05465405e-01 -6.58351600e-01 4.93017823e-01
-2.05920196e+00 -1.23067057e+00 7.31683731e-01 2.05873752e+00
7.37393022e-01 2.21486360e-01 2.51333773e-01 -2.37748906e-01
5.62222362e-01 4.15010780e-01 -9.44509149e-01 -4.28367645e-01
-8.18620622e-02 3.18390459e-01 8.65573287e-01 3.89512926e-01
-1.29074359e+00 1.33937061e+00 8.18402672e+00 3.99039209e-01
-1.62793732e+00 -2.59919673e-01 8.39283347e-01 -3.86711478e-01
2.98964739e-01 -2.15422556e-01 -5.31163096e-01 9.57878157e-02
9.21913922e-01 1.54582739e-01 7.13340342e-01 8.57215464e-01
3.17307353e-01 1.00291014e-01 -1.19789648e+00 7.83118546e-01
-6.29284978e-02 -1.77506852e+00 3.88384354e-03 -3.23674735e-03
6.94406807e-01 4.02599126e-01 4.51744080e-01 2.47301787e-01
6.29462957e-01 -1.35744846e+00 6.72456861e-01 3.96013796e-01
7.30061471e-01 -5.22108376e-01 1.39390782e-01 3.79893780e-01
-9.62213814e-01 -4.80248868e-01 -4.08075780e-01 -3.10360998e-01
-1.90477803e-01 -6.90336525e-02 -6.10596299e-01 -3.39075446e-01
9.58870232e-01 7.81628251e-01 -5.87823629e-01 6.85862124e-01
-1.78210437e-01 4.95604008e-01 -1.68723792e-01 1.14843577e-01
6.11296177e-01 -1.61446482e-01 1.01986960e-01 1.26593065e+00
9.59211662e-02 1.20868750e-01 3.64518046e-01 7.77734160e-01
-3.76426488e-01 -2.72701740e-01 -7.26368308e-01 4.32447530e-02
3.11526507e-01 1.24788392e+00 -7.98069179e-01 -5.39887965e-01
-4.37535524e-01 1.12822354e+00 7.42016852e-01 8.48904967e-01
-5.81966102e-01 -2.78546780e-01 4.76132989e-01 1.58960409e-02
9.99492705e-01 -5.66200018e-01 -1.64971486e-01 -1.06909359e+00
-1.05562136e-01 -9.13188398e-01 2.28866756e-01 -8.35888624e-01
-7.46466219e-01 5.75062871e-01 -1.92743272e-01 -1.09017384e+00
-7.11344957e-01 -9.00918901e-01 -4.68611598e-01 5.94189703e-01
-1.70908451e+00 -8.89571607e-01 -4.15473469e-02 6.30422413e-01
7.59192765e-01 -2.64096349e-01 7.37637162e-01 -1.74683869e-01
-5.70491374e-01 3.72963220e-01 2.56770067e-02 5.47407806e-01
4.06063765e-01 -1.23474979e+00 5.57230890e-01 8.33311021e-01
3.74124676e-01 5.40476680e-01 4.92683887e-01 -8.57941508e-02
-1.29371905e+00 -1.24050939e+00 5.07406592e-01 -3.00525188e-01
4.93088484e-01 -4.00974929e-01 -7.99695671e-01 1.19485056e+00
5.47180355e-01 7.20564544e-01 1.45919070e-01 -1.14126295e-01
-5.25717020e-01 -2.52576675e-02 -9.78235185e-01 6.02442622e-01
9.00540531e-01 -5.83017111e-01 -3.45781893e-01 3.93785715e-01
7.07068026e-01 -6.65508151e-01 -6.25403047e-01 2.32441261e-01
4.35926467e-01 -7.89152086e-01 1.18177605e+00 -9.58798051e-01
5.34246266e-01 -2.79246420e-01 1.16195835e-01 -1.28703594e+00
-7.36108541e-01 -4.83051598e-01 -3.27150226e-01 3.29785526e-01
4.75867927e-01 -4.08772260e-01 9.27473187e-01 4.83159482e-01
2.14732051e-01 -7.74296820e-01 -6.60475850e-01 -5.89941621e-01
6.41734675e-02 -2.71233879e-02 1.20978102e-01 6.20037079e-01
-3.64881873e-01 3.24811697e-01 -3.68466645e-01 2.81830579e-01
3.82522792e-01 3.89934778e-01 5.43538988e-01 -8.32759380e-01
-3.89218420e-01 -5.82562029e-01 -3.01357895e-01 -1.37277329e+00
2.69324422e-01 -5.47532380e-01 -4.25739326e-02 -1.12740374e+00
-1.15381561e-01 -2.17195034e-01 -5.42420566e-01 9.25444782e-01
1.59331411e-01 3.75607401e-01 2.22036064e-01 2.60495603e-01
-6.03521585e-01 2.79674172e-01 1.30382109e+00 -3.84384990e-01
-2.30713308e-01 9.83869508e-02 -4.81566757e-01 7.98259735e-01
7.95866132e-01 -3.35284382e-01 -4.56206113e-01 -5.54334223e-01
1.92220047e-01 1.95774987e-01 5.28376460e-01 -8.92061114e-01
1.88028127e-01 -3.52459759e-01 9.13697720e-01 -3.72539788e-01
2.97813147e-01 -8.95469606e-01 -3.94489050e-01 6.84667289e-01
-9.31476176e-01 4.46650535e-01 3.23230088e-01 7.10305393e-01
-1.23340204e-01 4.10152748e-02 1.07004726e+00 -4.55389827e-01
-9.64742184e-01 4.98727709e-01 -7.00731397e-01 -1.16386391e-01
1.02413261e+00 -1.79805666e-01 -1.58559740e-01 -3.48340213e-01
-7.36124873e-01 4.59223613e-02 4.91982609e-01 5.83711505e-01
6.47970200e-01 -1.33060420e+00 -4.23335552e-01 3.42101038e-01
-1.07901029e-01 7.89137185e-02 -2.57831812e-01 3.55688512e-01
-5.54468334e-01 4.56955969e-01 -4.71754014e-01 -7.05063820e-01
-1.05065835e+00 9.55291390e-01 6.15707576e-01 -3.23343784e-01
-9.41098869e-01 5.00850558e-01 2.13726282e-01 -1.13337055e-01
3.54829580e-01 -5.61561644e-01 -1.80139303e-01 -2.46086299e-01
6.94961727e-01 -8.76691863e-02 -1.38151318e-01 -4.30308074e-01
3.63995023e-02 6.06001854e-01 -2.49397218e-01 -1.81772843e-01
1.31351554e+00 -8.20003897e-02 -2.01578606e-02 2.58381337e-01
1.11612475e+00 -6.20888948e-01 -2.02279329e+00 -5.33934832e-01
8.80424678e-02 -3.15052956e-01 2.55617946e-01 -7.13666260e-01
-1.33588231e+00 7.87456393e-01 6.34198785e-01 -8.14938769e-02
1.20812154e+00 -2.38877699e-01 4.67220634e-01 7.98036814e-01
1.59008965e-01 -1.25478768e+00 6.80487275e-01 7.20330358e-01
9.91896808e-01 -1.48903596e+00 -1.18596539e-01 1.02108561e-01
-6.23138428e-01 1.28394163e+00 7.00969279e-01 -7.22958982e-01
6.98205173e-01 4.80677515e-01 1.70462877e-01 -1.59834430e-01
-1.10094929e+00 -1.81957096e-01 3.04496646e-01 6.63480282e-01
3.22860926e-01 -2.41406649e-01 2.67480403e-01 -2.32644424e-01
5.97034988e-04 7.98359290e-02 3.67430568e-01 9.35629129e-01
-6.08693004e-01 -7.52766848e-01 -1.55497298e-01 4.30355549e-01
-5.46943545e-01 -7.31930956e-02 -1.81828767e-01 6.99646175e-01
-1.69594586e-01 4.92224127e-01 5.02056837e-01 -8.46030340e-02
-2.67175213e-02 -4.20092903e-02 6.08622909e-01 -5.69500387e-01
-5.85488915e-01 1.54440179e-01 -3.37452173e-01 -8.62358510e-01
-6.58063412e-01 -5.60643971e-01 -1.19636273e+00 -1.93151124e-02
8.56072083e-02 -3.42406660e-01 4.71182883e-01 8.02474082e-01
3.78986567e-01 6.80163562e-01 5.81630468e-01 -1.18695712e+00
-5.41704655e-01 -7.50614822e-01 -2.38099352e-01 3.29805821e-01
8.91953468e-01 -3.14480126e-01 5.27454279e-02 3.74799818e-01] | [9.347434997558594, 1.082413911819458] |
70b8a279-7b33-4b3f-83f3-9e91705af174 | text2facegan-face-generation-from-fine | 1911.11378 | null | https://arxiv.org/abs/1911.11378v1 | https://arxiv.org/pdf/1911.11378v1.pdf | Text2FaceGAN: Face Generation from Fine Grained Textual Descriptions | Powerful generative adversarial networks (GAN) have been developed to automatically synthesize realistic images from text. However, most existing tasks are limited to generating simple images such as flowers from captions. In this work, we extend this problem to the less addressed domain of face generation from fine-grained textual descriptions of face, e.g., "A person has curly hair, oval face, and mustache". We are motivated by the potential of automated face generation to impact and assist critical tasks such as criminal face reconstruction. Since current datasets for the task are either very small or do not contain captions, we generate captions for images in the CelebA dataset by creating an algorithm to automatically convert a list of attributes to a set of captions. We then model the highly multi-modal problem of text to face generation as learning the conditional distribution of faces (conditioned on text) in same latent space. We utilize the current state-of-the-art GAN (DC-GAN with GAN-CLS loss) for learning conditional multi-modality. The presence of more fine-grained details and variable length of the captions makes the problem easier for a user but more difficult to handle compared to the other text-to-image tasks. We flipped the labels for real and fake images and added noise in discriminator. Generated images for diverse textual descriptions show promising results. In the end, we show how the widely used inceptions score is not a good metric to evaluate the performance of generative models used for synthesizing faces from text. | ['Yi Yu', 'Shailesh Kumar Jha', 'Rajiv Ratn Shah', 'Manraj Singh Grover', 'Ajit Kumar', 'Osaid Rehman Nasir'] | 2019-11-26 | null | null | null | null | ['text-to-face-generation'] | ['computer-vision'] | [ 5.45457661e-01 2.63083100e-01 3.45307082e-01 -6.73352182e-01
-1.01244116e+00 -8.24448764e-01 9.56093132e-01 -8.69019330e-01
6.55930340e-02 1.21985567e+00 1.31233245e-01 7.29456097e-02
3.72542351e-01 -9.10224617e-01 -1.06343114e+00 -7.67421901e-01
4.90664035e-01 9.35566366e-01 -5.64054847e-01 -1.13277085e-01
-1.75576374e-01 6.13200426e-01 -1.58232152e+00 6.50940180e-01
4.63975549e-01 8.41047287e-01 -1.15575247e-01 7.19554484e-01
-5.76816872e-02 6.29329562e-01 -8.52185428e-01 -1.12789476e+00
2.85370409e-01 -9.88381863e-01 -5.68370461e-01 2.58390665e-01
1.03173089e+00 -5.18965721e-01 -1.35940135e-01 1.01319444e+00
5.93575954e-01 -3.03655058e-01 1.15470409e+00 -1.68163002e+00
-1.06186950e+00 3.73197049e-01 -5.79828084e-01 -3.73159528e-01
5.41316569e-01 3.37630689e-01 5.64155757e-01 -7.93981552e-01
8.30793917e-01 1.80659831e+00 6.10794008e-01 1.18835986e+00
-1.35520518e+00 -9.21103776e-01 -4.31814939e-01 -3.44572723e-01
-1.24993551e+00 -7.60514975e-01 5.74461699e-01 -5.13504505e-01
3.65771532e-01 2.49566555e-01 1.23650491e-01 2.03444934e+00
-1.46485707e-02 3.91272992e-01 1.41998720e+00 -3.50993752e-01
3.30244265e-02 3.32000464e-01 -7.84836888e-01 7.68403947e-01
1.50379106e-01 1.08212136e-01 -3.77767205e-01 -1.96881324e-01
7.86295176e-01 -2.51276702e-01 4.69597243e-02 1.21964201e-01
-8.76783550e-01 1.21968365e+00 1.86894983e-01 1.64349571e-01
-3.43027174e-01 3.32944423e-01 1.83301255e-01 2.26419732e-01
4.05599266e-01 3.33955407e-01 1.56132340e-01 1.61782086e-01
-1.11780095e+00 5.58050931e-01 6.99238360e-01 1.05745125e+00
5.98952591e-01 4.72551286e-01 -5.70936263e-01 7.05482841e-01
1.67834312e-01 9.90913033e-01 2.30655581e-01 -8.05504978e-01
3.58796865e-01 7.72530138e-02 1.43010780e-01 -8.17684829e-01
2.23652661e-01 1.42408833e-02 -8.67768109e-01 3.94199222e-01
4.87309813e-01 -4.92674440e-01 -1.41693330e+00 1.92002237e+00
1.08674183e-01 1.02945969e-01 9.82298031e-02 7.32921958e-01
9.86935198e-01 7.28217483e-01 1.55933991e-01 1.36153623e-02
1.31476355e+00 -7.67098129e-01 -7.54088700e-01 -3.30827445e-01
-5.91144785e-02 -9.50824201e-01 1.09236038e+00 -2.29385402e-02
-1.08020556e+00 -5.55240989e-01 -6.37984335e-01 2.11443775e-03
-4.03259337e-01 2.73826450e-01 4.63634551e-01 1.10001969e+00
-1.20182765e+00 2.77791739e-01 -3.32237005e-01 -3.27503800e-01
8.73196006e-01 2.73228437e-01 -7.14782178e-01 -3.31928879e-01
-1.04863501e+00 8.50469768e-01 1.99534059e-01 -6.62685260e-02
-1.47056353e+00 -3.87373030e-01 -8.78993928e-01 -6.83742538e-02
3.62016857e-02 -8.36583197e-01 9.79277194e-01 -1.60736680e+00
-1.41902435e+00 1.23601425e+00 -2.10443929e-01 -3.31816226e-01
8.65124941e-01 9.69870463e-02 -3.78917634e-01 1.39961287e-01
3.30861032e-01 1.26282024e+00 1.55336988e+00 -1.74237216e+00
-1.07874423e-01 -3.16093445e-01 4.84610535e-02 -1.01360708e-01
-9.29728970e-02 1.53563142e-01 -9.10840854e-02 -7.64067471e-01
-6.57385409e-01 -1.06030023e+00 3.47726136e-01 1.00889474e-01
-6.13043845e-01 1.16516642e-01 1.06401241e+00 -8.53966832e-01
4.41079915e-01 -1.91401577e+00 7.73038864e-02 -1.64255247e-01
-9.72882286e-02 1.52016819e-01 -4.72307712e-01 3.33396584e-01
-1.46099508e-01 4.46234077e-01 -3.87792498e-01 -8.12972665e-01
7.11399391e-02 2.81613439e-01 -4.96631682e-01 3.20503324e-01
6.42870009e-01 1.01445258e+00 -5.83916664e-01 -5.83750963e-01
-4.76700850e-02 8.86224985e-01 -4.21616077e-01 4.71506059e-01
-4.18526649e-01 7.38853574e-01 -2.33555540e-01 8.09620380e-01
8.31690550e-01 1.23559907e-01 -2.80720085e-01 -1.16500854e-01
5.35800040e-01 -4.00615096e-01 -6.65058613e-01 1.38605630e+00
-5.32365918e-01 5.73122799e-01 8.40651840e-02 -5.30445337e-01
8.73984814e-01 4.69970405e-01 4.77923714e-02 -4.11954820e-01
1.39869183e-01 8.15680623e-02 -1.62840307e-01 -4.54532087e-01
3.05330783e-01 -7.10575879e-01 -1.63049653e-01 4.14948672e-01
3.34781021e-01 -5.79409361e-01 1.63010687e-01 -1.29982848e-02
7.55446374e-01 4.92239743e-01 -2.33242899e-01 1.49415180e-01
2.35415116e-01 -1.65989593e-01 1.29950851e-01 5.52743733e-01
1.76802650e-01 1.20503640e+00 5.97370028e-01 -1.44290581e-01
-1.44843364e+00 -1.22811604e+00 -7.00505264e-03 8.47494364e-01
-5.04657567e-01 6.56364067e-03 -1.17284930e+00 -8.06163430e-01
-5.54795451e-02 8.90950382e-01 -9.92668450e-01 -8.20155255e-03
-3.48807037e-01 -7.77994037e-01 1.07932556e+00 1.96868211e-01
5.81519783e-01 -1.34979582e+00 -8.37269574e-02 -1.22156240e-01
-2.76667178e-01 -1.28341448e+00 -4.44668114e-01 -4.26096767e-01
-3.07964683e-01 -7.99409568e-01 -9.75359857e-01 -7.98856676e-01
8.99657249e-01 -4.05989915e-01 1.30444276e+00 -3.21161717e-01
-4.56772953e-01 4.29780215e-01 -2.28323728e-01 -5.37764788e-01
-9.42491770e-01 -3.29426914e-01 -1.45299554e-01 4.16790962e-01
2.47712173e-02 -3.29741120e-01 -2.76756525e-01 1.41168162e-01
-1.19822598e+00 9.96105224e-02 5.02084136e-01 9.72701252e-01
3.30054879e-01 -1.14715323e-01 8.24080169e-01 -1.25289595e+00
5.95563531e-01 -5.33583224e-01 -4.43082839e-01 2.76270568e-01
-1.33511990e-01 5.43460771e-02 7.98969984e-01 -5.47640800e-01
-1.32085550e+00 1.86140448e-01 -2.60721445e-01 -6.93124115e-01
-4.15230095e-01 -6.74270093e-02 -3.60898703e-01 -8.39531608e-03
6.35383427e-01 3.23203444e-01 4.43498082e-02 -4.72323336e-02
5.11820436e-01 7.49582291e-01 6.89771235e-01 -7.57228971e-01
1.10247171e+00 5.52683115e-01 2.05530047e-01 -6.97979093e-01
-6.15966797e-01 4.32963222e-01 -3.59852731e-01 -1.62271529e-01
1.04418266e+00 -9.67054546e-01 -5.89024723e-01 4.88512069e-01
-1.35193050e+00 -2.23979369e-01 -1.53791353e-01 -1.13046803e-01
-8.22210014e-01 -3.32485773e-02 -4.39761162e-01 -8.60132039e-01
-3.96041662e-01 -1.10638011e+00 1.53342366e+00 2.28280559e-01
-8.52848589e-02 -9.29920435e-01 -1.42301753e-01 7.32706606e-01
5.11947334e-01 9.45569277e-01 7.89771259e-01 -3.25319856e-01
-4.01695549e-01 -1.65752441e-01 -2.13753298e-01 5.52297652e-01
1.78216994e-01 1.68535143e-01 -1.44183815e+00 -3.42034966e-01
-8.94242302e-02 -8.80129576e-01 6.73293412e-01 1.27866358e-01
9.98225749e-01 -6.09463096e-01 -1.18607171e-01 6.01889968e-01
1.51153946e+00 3.79028618e-02 1.02278697e+00 -2.50702441e-01
8.15999448e-01 8.11067581e-01 3.60145271e-01 3.34723800e-01
-5.37730716e-02 4.96963024e-01 4.91758525e-01 -1.89801618e-01
-4.44774002e-01 -5.17446399e-01 5.56145608e-01 -1.53124243e-01
5.27527370e-02 -7.54058242e-01 -6.18918777e-01 4.88773793e-01
-1.24857867e+00 -1.28196430e+00 8.77815485e-02 1.84426498e+00
9.27616477e-01 -1.67710036e-01 -6.31345622e-03 -3.16743404e-01
1.05141592e+00 -9.97001603e-02 -3.66831750e-01 -4.98797446e-01
-3.71430695e-01 4.91778642e-01 2.40282536e-01 3.97529781e-01
-1.05423963e+00 1.11655998e+00 5.89854956e+00 9.89021778e-01
-1.20940232e+00 2.52085298e-01 1.18452382e+00 -3.86692882e-02
-3.78778636e-01 -1.97255030e-01 -8.59853625e-01 7.51361609e-01
9.64905560e-01 1.49995774e-01 5.99743426e-01 6.19603455e-01
1.13465153e-01 1.78471029e-01 -1.28299522e+00 1.15467012e+00
5.69506943e-01 -1.16818953e+00 5.83492637e-01 -8.10538325e-03
1.17873466e+00 -4.50510353e-01 5.43704510e-01 1.57874256e-01
3.97280931e-01 -1.87756538e+00 7.94183433e-01 5.29609501e-01
1.61990786e+00 -6.45472288e-01 5.94835520e-01 2.46747658e-02
-6.51947498e-01 1.83729231e-01 -2.92474300e-01 3.96766782e-01
1.55435413e-01 3.16936880e-01 -1.10617507e+00 3.38140219e-01
3.81589234e-01 1.69217035e-01 -6.13599539e-01 4.69955534e-01
-1.64200753e-01 4.44731504e-01 -2.08890378e-01 2.44015366e-01
2.24530816e-01 -3.02112009e-02 2.98411071e-01 1.08472466e+00
6.29430532e-01 -2.48478279e-01 -2.67549694e-01 1.35552609e+00
-5.56165695e-01 -1.98178232e-01 -1.07150817e+00 -4.43970203e-01
2.95017332e-01 1.36897767e+00 -5.24319768e-01 -1.75142199e-01
-2.30055794e-01 1.28964114e+00 1.07489444e-01 3.47488940e-01
-1.05239475e+00 -1.56556800e-01 3.84888232e-01 3.91201019e-01
2.47312322e-01 3.00279230e-01 -2.25828752e-01 -9.47162271e-01
-7.21449554e-02 -1.17342794e+00 5.94462156e-02 -1.25328970e+00
-1.57609630e+00 1.07410812e+00 2.16872003e-02 -8.70691836e-01
-7.08730280e-01 -5.69075465e-01 -5.94223142e-01 1.13518977e+00
-1.17604756e+00 -2.03490996e+00 -4.36894923e-01 7.24421084e-01
6.38089478e-01 -5.25877416e-01 1.12746596e+00 2.11840957e-01
-1.53400764e-01 7.96127498e-01 -5.29971533e-02 3.23868424e-01
9.85732257e-01 -1.03940487e+00 4.48834687e-01 7.31388450e-01
9.61927176e-02 3.06538194e-01 7.99692750e-01 -7.82530189e-01
-1.20276022e+00 -1.36866415e+00 6.24924362e-01 -7.45034933e-01
2.23209798e-01 -6.44377828e-01 -3.67726386e-01 8.49745572e-01
3.86891991e-01 -1.93643682e-02 4.78314281e-01 -5.54005325e-01
-3.74146819e-01 5.23216762e-02 -1.74194920e+00 5.26000917e-01
8.09629261e-01 -5.60786068e-01 -2.13057607e-01 4.25026417e-01
3.02227974e-01 -2.14632541e-01 -6.06849194e-01 1.39579043e-01
5.11841655e-01 -8.51719022e-01 8.32747817e-01 -6.34874761e-01
9.03232217e-01 -1.22856282e-01 -1.57764420e-01 -1.29527843e+00
5.04017249e-02 -6.74463153e-01 4.37241465e-01 1.78001404e+00
4.03247654e-01 -3.37404341e-01 8.76953781e-01 5.66363931e-01
2.19669074e-01 -2.40183532e-01 -8.65263641e-01 -5.53891718e-01
2.02796072e-01 1.38395399e-01 6.45297945e-01 8.94666672e-01
-9.27467942e-01 5.39210439e-01 -8.26949716e-01 -1.00533046e-01
7.18298972e-01 6.78286031e-02 9.13391292e-01 -1.02987111e+00
-2.87401170e-01 -1.22686801e-02 -2.93883324e-01 -1.83618233e-01
6.44400895e-01 -8.68127286e-01 -6.29124716e-02 -1.19309461e+00
3.43485206e-01 -3.39601398e-01 5.82319736e-01 5.45678496e-01
-3.24671268e-02 9.21514571e-01 2.92799503e-01 -8.26604068e-02
2.94941347e-02 5.34582555e-01 1.51002121e+00 -2.86403626e-01
4.92141277e-01 -1.21914245e-01 -7.13091552e-01 4.90430623e-01
6.57044113e-01 -6.76100433e-01 -4.88984108e-01 -4.68432933e-01
5.48708029e-02 2.06048191e-01 7.21440017e-01 -8.42924356e-01
-3.48931372e-01 -1.39254048e-01 8.00560951e-01 1.10061122e-02
7.44591117e-01 -7.38379061e-01 6.26929522e-01 1.37980357e-01
-2.31479988e-01 2.36915257e-02 2.20120341e-01 3.21257323e-01
-1.76449075e-01 -3.40665698e-01 1.05526292e+00 -3.65879893e-01
-1.72356620e-01 4.13170815e-01 4.59788479e-02 2.26898149e-01
1.12987709e+00 -1.61905125e-01 -4.16657180e-01 -8.32512021e-01
-7.30470300e-01 -2.47949570e-01 6.70751691e-01 4.80895042e-01
6.38512433e-01 -1.49224877e+00 -1.27398348e+00 2.42392913e-01
-9.12931487e-02 -1.38678536e-01 1.43020794e-01 8.16167146e-02
-6.68801188e-01 1.87720791e-01 -7.54909277e-01 -3.84231925e-01
-1.41529655e+00 5.10537803e-01 3.25421214e-01 -5.44516034e-02
-1.68341454e-02 8.30350637e-01 4.56280082e-01 -2.42150798e-01
-1.46356329e-01 3.29006582e-01 4.24398202e-03 9.69126970e-02
3.65694255e-01 4.30800347e-03 -1.17147818e-01 -1.08288300e+00
-7.21635371e-02 4.19690400e-01 6.20681196e-02 -3.02773446e-01
1.33032167e+00 1.61039382e-01 -1.53580189e-01 7.77233467e-02
1.20040774e+00 3.89728360e-02 -1.34705424e+00 3.90211433e-01
-7.16264725e-01 -5.79695284e-01 -4.41493809e-01 -1.04233968e+00
-1.12194550e+00 9.18151379e-01 7.19730437e-01 -4.51604612e-02
1.03445697e+00 -1.40193719e-02 6.16010666e-01 2.55119838e-02
2.63035655e-01 -5.49043477e-01 3.44852924e-01 1.61964834e-01
1.16888809e+00 -1.35544515e+00 -2.83174574e-01 -4.11480784e-01
-9.29381728e-01 9.03348684e-01 6.88300371e-01 -4.70512472e-02
3.19912955e-02 4.91482228e-01 6.36224151e-02 -3.80821228e-02
-4.95312274e-01 -1.75094623e-02 1.36438981e-01 1.05469441e+00
3.85246426e-01 3.94359976e-02 1.62493423e-01 2.53419995e-01
-5.34158170e-01 -5.28465360e-02 6.84357762e-01 5.11229694e-01
1.88376009e-01 -1.25266731e+00 -7.25145459e-01 5.43072045e-01
-8.28385353e-01 -1.84726626e-01 -6.86350465e-01 6.23365998e-01
3.92078459e-01 9.14747000e-01 5.43364100e-02 -1.14306100e-01
-2.17539877e-01 2.92797148e-01 8.16971004e-01 -6.36135519e-01
-5.02802849e-01 -2.23282844e-01 2.17935592e-01 -6.09838497e-03
-3.74742001e-01 -5.50572813e-01 -7.28336215e-01 -5.32022476e-01
-2.41493564e-02 -9.80977267e-02 7.54275441e-01 7.39625633e-01
1.85340181e-01 2.82905966e-01 5.32319486e-01 -9.52862501e-01
-4.14126128e-01 -1.13226187e+00 -4.12179381e-01 9.04656351e-01
1.65260807e-01 -6.26918435e-01 -2.97754973e-01 6.38979197e-01] | [12.24974250793457, -0.09776004403829575] |
6bca4c5c-63d6-408c-a2d9-b8abd5f3a4e1 | inspecting-state-of-the-art-performance-and | 2011.09257 | null | https://arxiv.org/abs/2011.09257v3 | https://arxiv.org/pdf/2011.09257v3.pdf | Inspecting state of the art performance and NLP metrics in image-based medical report generation | Several deep learning architectures have been proposed over the last years to deal with the problem of generating a written report given an imaging exam as input. Most works evaluate the generated reports using standard Natural Language Processing (NLP) metrics (e.g. BLEU, ROUGE), reporting significant progress. In this article, we contrast this progress by comparing state of the art (SOTA) models against weak baselines. We show that simple and even naive approaches yield near SOTA performance on most traditional NLP metrics. We conclude that evaluation methods in this task should be further studied towards correctly measuring clinical accuracy, ideally involving physicians to contribute to this end. | ['Sergio Uribe', 'Cecilia Besa', 'Pablo Messina', 'Denis Parra', 'Pablo Pino'] | 2020-11-18 | null | null | null | null | ['medical-report-generation'] | ['medical'] | [ 3.10243309e-01 8.09164286e-01 -2.82737434e-01 -5.97746551e-01
-1.53158116e+00 -5.63395381e-01 7.01577961e-01 6.83509886e-01
-5.68803430e-01 1.00026584e+00 7.18839288e-01 -4.84588742e-01
-8.59631822e-02 -6.92396998e-01 -5.46554565e-01 -3.37952942e-01
6.73611555e-03 7.12091982e-01 -4.07898836e-02 9.60959494e-02
3.86977762e-01 2.81232834e-01 -8.24267626e-01 7.22573340e-01
5.75808525e-01 7.77476251e-01 -3.06808233e-01 1.09320915e+00
-1.58168554e-01 1.46383643e+00 -8.69264901e-01 -9.14418221e-01
-1.36841759e-01 -5.99976063e-01 -1.15988934e+00 -1.17497176e-01
7.12428749e-01 -4.86258030e-01 -1.63465247e-01 8.32224488e-01
7.66609967e-01 -4.92618605e-02 7.47710288e-01 -8.12493801e-01
-7.50017762e-01 7.97062933e-01 -3.06736857e-01 7.11887598e-01
7.19528913e-01 2.74265110e-01 7.29467690e-01 -6.22020185e-01
9.69201207e-01 9.86651897e-01 6.02753460e-01 5.76342762e-01
-9.15235460e-01 -3.88379425e-01 -3.08624327e-01 1.40280440e-01
-8.67338240e-01 -1.59627587e-01 4.90177572e-01 -5.42082012e-01
1.03426278e+00 3.12577039e-01 2.21602231e-01 1.46253288e+00
7.72575676e-01 1.03969049e+00 1.31537688e+00 -4.10091758e-01
2.51450092e-01 1.73012823e-01 3.93452823e-01 9.52324331e-01
5.27346909e-01 -2.84424592e-02 -4.16873276e-01 -9.40515026e-02
5.15613735e-01 -5.80635250e-01 -2.26448417e-01 -9.77733452e-03
-1.56404102e+00 1.07510102e+00 5.11449337e-01 5.15980780e-01
-6.23462915e-01 5.47731817e-02 7.88907409e-01 1.40787899e-01
5.44436216e-01 9.12196338e-01 -9.19084623e-02 -3.23392332e-01
-1.29193354e+00 6.46319807e-01 9.86322641e-01 8.52138877e-01
-2.47427404e-01 -2.03715548e-01 -9.33102369e-01 4.59552705e-01
3.12441848e-02 -2.98499204e-02 7.54157603e-01 -9.80185926e-01
6.49217069e-01 3.37314993e-01 1.05697006e-01 -8.95301819e-01
-7.57786274e-01 -5.95924437e-01 -8.88496578e-01 3.45275849e-02
2.17248783e-01 -2.33634114e-01 -1.13449514e+00 1.23135948e+00
-1.05095662e-01 -5.55028692e-02 1.93695977e-01 8.67696047e-01
1.50689685e+00 6.06835067e-01 4.23184454e-01 -3.39289755e-01
1.64190054e+00 -9.47173476e-01 -1.04664612e+00 -6.02941737e-02
7.68515706e-01 -9.20766175e-01 9.39411581e-01 6.99661851e-01
-1.78723943e+00 -5.22427976e-01 -1.05481946e+00 -3.13372910e-01
-3.15005094e-01 3.51460278e-01 6.55711889e-01 6.27948701e-01
-1.37760162e+00 5.34535229e-01 -9.25087035e-01 -2.24755168e-01
5.89735925e-01 2.28698745e-01 -2.88905770e-01 -1.67814016e-01
-1.09570491e+00 1.15448391e+00 4.28781599e-01 -1.21085532e-01
-1.02694309e+00 -1.01702809e+00 -9.47367251e-01 8.29363614e-02
4.43924010e-01 -1.10146034e+00 2.07461500e+00 -2.57352591e-01
-1.31389463e+00 1.46948719e+00 4.12605330e-02 -1.03050816e+00
8.89093935e-01 -7.81757310e-02 -2.82001585e-01 4.04968858e-01
1.26554370e-01 8.81815195e-01 2.70291474e-02 -1.04554093e+00
-5.02515435e-01 -6.88415170e-02 1.56944826e-01 -1.14173032e-01
1.34088069e-01 2.57253319e-01 -1.49201959e-01 -7.73114383e-01
-3.70303035e-01 -8.68905842e-01 -4.86853361e-01 -2.85457879e-01
-7.28849232e-01 -6.00691199e-01 -7.27804676e-02 -6.22601509e-01
1.23877406e+00 -1.61248827e+00 -1.78907186e-01 -2.73910344e-01
5.41327119e-01 2.59194136e-01 -3.21585499e-02 2.65413910e-01
-3.37306768e-01 3.84315163e-01 -2.45152891e-01 -4.70249832e-01
-8.64468217e-02 -1.95756536e-02 -3.14897984e-01 1.92874417e-01
5.50460100e-01 1.10510075e+00 -1.12925065e+00 -8.60831320e-01
9.77602750e-02 2.85763562e-01 -2.87540466e-01 3.39248121e-01
-1.92829624e-01 2.03882486e-01 -3.95698637e-01 2.55634457e-01
2.46780574e-01 -6.31324232e-01 1.13573447e-01 -1.00319367e-03
1.67265415e-01 7.59915471e-01 -5.76593161e-01 1.71492708e+00
-6.30337059e-01 6.96356535e-01 -2.99605131e-01 -7.80862451e-01
7.10749090e-01 7.60964036e-01 5.94866157e-01 -5.46944439e-01
3.66653681e-01 2.04368815e-01 1.89712152e-01 -6.43204987e-01
5.93976021e-01 -1.74108937e-01 -5.79306856e-03 6.56425238e-01
3.75815392e-01 -2.27754116e-01 4.52726036e-01 4.87246990e-01
1.37470925e+00 -1.47240311e-01 9.70473468e-01 -1.48953721e-01
5.49156010e-01 3.87081772e-01 5.98646626e-02 1.24994314e+00
-3.50492418e-01 1.03099346e+00 6.36642456e-01 -7.82638490e-01
-1.06960857e+00 -1.09267294e+00 -9.28513557e-02 6.48066401e-01
-6.23177111e-01 -4.22288835e-01 -8.84605110e-01 -1.12279975e+00
-4.38251019e-01 9.78195667e-01 -8.76051188e-01 -4.39831167e-02
-5.65320194e-01 -7.84163952e-01 5.76141000e-01 8.01781952e-01
2.17121407e-01 -1.53949130e+00 -9.05670166e-01 5.16008556e-01
-1.34141460e-01 -1.25720489e+00 -2.28953212e-01 6.99630678e-02
-9.60188687e-01 -8.39455843e-01 -1.12522209e+00 -6.28882766e-01
4.36023980e-01 -3.63924474e-01 1.76159823e+00 -8.70759562e-02
-5.20113707e-01 5.04965782e-01 -3.93838972e-01 -8.34674835e-01
-9.50684309e-01 1.81929424e-01 -3.94974202e-01 -5.39157927e-01
5.95646083e-01 -7.58091435e-02 -7.18858063e-01 -4.09527868e-01
-1.23535979e+00 8.53101313e-02 9.06170607e-01 7.89366484e-01
6.59548104e-01 -5.15294671e-01 4.13276345e-01 -1.24257743e+00
1.23341203e+00 -4.50576842e-01 -9.75498483e-02 1.51751295e-01
-7.17508316e-01 1.93985671e-01 5.08904636e-01 -1.03968225e-01
-1.04073024e+00 -1.30861938e-01 -3.32533509e-01 -1.03392310e-01
-6.40311599e-01 7.26899564e-01 6.68229043e-01 3.37026596e-01
1.00307739e+00 3.17839742e-01 -1.24513984e-01 -3.31653953e-02
1.50216356e-01 4.37526852e-01 6.98899746e-01 -2.89413154e-01
2.47365445e-01 2.80303210e-01 -1.02046560e-02 -1.29984811e-01
-1.57653165e+00 -2.84527928e-01 -4.54354137e-01 -2.17320979e-01
9.77549851e-01 -6.36739612e-01 -5.96397460e-01 -3.85367244e-01
-1.42355275e+00 -6.52539283e-02 -3.73248398e-01 6.65631890e-01
-8.46338689e-01 2.18055785e-01 -9.80837464e-01 -6.59031034e-01
-8.21198642e-01 -1.28508389e+00 1.36919177e+00 2.29000777e-01
-7.76197314e-01 -1.02688754e+00 5.22117436e-01 5.71319580e-01
3.51524234e-01 5.89693785e-01 6.98208094e-01 -7.96972454e-01
-2.83092678e-01 -4.53122824e-01 -3.63433331e-01 1.23374827e-01
-1.40840203e-01 -1.91080540e-01 -9.40306723e-01 -5.14233895e-02
1.62954047e-01 -6.00558877e-01 9.80620325e-01 7.01077223e-01
1.58220220e+00 -3.41990352e-01 -3.24666262e-01 1.94943234e-01
1.22520995e+00 2.59730309e-01 6.99675024e-01 2.20199332e-01
1.93680331e-01 4.74530101e-01 5.98213851e-01 1.41331062e-01
5.11521578e-01 9.74028558e-02 9.09531638e-02 -2.03145251e-01
-4.40319508e-01 -1.37964888e-02 1.17642753e-01 8.04359615e-01
-6.37865141e-02 -5.29009342e-01 -1.28746402e+00 6.62534535e-01
-1.59264994e+00 -6.82428777e-01 -8.99477117e-03 1.59136724e+00
1.08946383e+00 4.67147380e-01 -1.20170116e-01 -1.02868162e-01
1.60564169e-01 1.70282155e-01 -2.36299872e-01 -8.23134124e-01
2.72630900e-01 4.98424917e-01 2.65504003e-01 2.87717521e-01
-1.13199914e+00 3.97630870e-01 7.16156244e+00 4.96448487e-01
-1.10995018e+00 3.18863511e-01 1.11813009e+00 -9.93341729e-02
-9.61879939e-02 -6.90789104e-01 -5.20983040e-01 1.68588221e-01
1.37743318e+00 -2.64473140e-01 -3.31079006e-01 9.07471418e-01
2.59765029e-01 -3.81343886e-02 -1.45921755e+00 9.07095134e-01
4.20240462e-01 -1.71874237e+00 1.13831483e-01 4.46843803e-02
7.22001970e-01 1.10972539e-01 6.37173653e-02 5.33342063e-01
7.74373114e-01 -1.14087737e+00 3.02025288e-01 6.78510606e-01
7.46811509e-01 -4.64849174e-01 1.20020449e+00 -1.26426458e-01
-4.16138083e-01 3.80580127e-01 -1.29849717e-01 2.73080289e-01
1.73542157e-01 6.98989034e-01 -1.66042864e+00 8.17011297e-01
4.36504155e-01 3.71661305e-01 -6.52913809e-01 1.25835657e+00
-2.74978161e-01 1.01818132e+00 8.01371038e-03 -1.96746796e-01
7.44684637e-01 4.49046940e-01 4.12506610e-01 1.67646897e+00
2.50160873e-01 1.10608205e-01 2.52802074e-01 1.03060579e+00
-4.07502353e-01 4.22676146e-01 -7.45086908e-01 -2.04648048e-01
-2.95109183e-01 1.32968521e+00 -9.79240239e-01 -8.46126676e-01
-3.02956998e-01 6.44562304e-01 2.64274240e-01 -1.47151589e-01
-8.46532583e-01 -1.70913890e-01 3.77615243e-02 1.44357339e-01
-1.28938735e-01 -1.76814888e-02 -5.74887276e-01 -8.59236538e-01
1.24125287e-01 -8.92459393e-01 9.37340021e-01 -1.07036602e+00
-1.47966909e+00 9.35910463e-01 -1.29072949e-01 -1.26424611e+00
-6.28969491e-01 -8.06788146e-01 -5.39188147e-01 5.11798501e-01
-1.57858670e+00 -6.88180625e-01 -2.80435890e-01 1.10506974e-01
8.21123183e-01 -9.88752991e-02 1.07009590e+00 8.73562172e-02
-2.17065185e-01 6.28625035e-01 -4.51712191e-01 4.40826803e-01
8.79533350e-01 -1.68951178e+00 2.56210983e-01 7.00041890e-01
2.14547426e-01 5.98970532e-01 7.26695716e-01 -5.60750961e-01
-8.71804476e-01 -9.94017959e-01 1.06190932e+00 -8.51229072e-01
5.75878799e-01 7.08476678e-02 -6.50454402e-01 6.60520077e-01
7.21255839e-01 -1.03768587e-01 9.50920045e-01 -2.08215974e-02
-1.04936287e-01 1.63185209e-01 -1.17553294e+00 5.43929815e-01
5.18452525e-01 -2.33977407e-01 -7.79131413e-01 8.25106144e-01
8.98240387e-01 -7.44355261e-01 -1.10250092e+00 5.54971695e-01
2.50117004e-01 -7.91700900e-01 7.00273454e-01 -1.05667269e+00
1.43473279e+00 1.16645716e-01 5.38408875e-01 -1.37290502e+00
-2.68254608e-01 -4.59057808e-01 -1.28035039e-01 7.41253793e-01
6.89880729e-01 1.72959194e-02 8.99151087e-01 4.78271604e-01
-4.17108148e-01 -1.00536263e+00 -6.65790737e-01 -3.26916754e-01
3.95013630e-01 -5.89457393e-01 2.62809247e-01 9.46335435e-01
1.21315718e-01 4.74334359e-01 -7.37305656e-02 -9.13648084e-02
3.38541210e-01 -6.20504878e-02 3.86576772e-01 -9.34538066e-01
-1.74070410e-02 -6.13643110e-01 -4.29610223e-01 -3.85564625e-01
-1.24108851e-01 -1.06237948e+00 1.13241799e-01 -1.97302198e+00
4.54474479e-01 1.62071168e-01 -4.30658907e-01 4.20274943e-01
-3.70745778e-01 2.65260965e-01 8.38981941e-02 -2.52274483e-01
-9.38051641e-01 -8.16528723e-02 1.50680339e+00 -3.29797059e-01
6.27995133e-02 -1.06192097e-01 -9.84443724e-01 5.25027692e-01
7.55271494e-01 -6.73445642e-01 -1.23923637e-01 -4.39226031e-01
4.24759954e-01 3.81173402e-01 2.56013960e-01 -1.04411745e+00
2.12114871e-01 5.05978763e-02 4.54301804e-01 -7.08669007e-01
-1.29289091e-01 -3.62269580e-01 -4.73687500e-01 5.60573339e-01
-1.11428642e+00 2.88554490e-01 2.58192241e-01 2.31921718e-01
-5.80158651e-01 -3.66832972e-01 4.32944208e-01 -6.27653539e-01
-1.67778447e-01 1.59227267e-01 -4.65094388e-01 3.37695540e-03
9.16614592e-01 2.53145397e-01 -3.75154823e-01 -6.02412760e-01
-1.05087495e+00 2.10790873e-01 -3.47824156e-01 4.46479648e-01
6.84786320e-01 -1.04182744e+00 -1.37557197e+00 -4.81942326e-01
2.55030245e-01 8.23115185e-02 2.86082834e-01 1.02136672e+00
-9.76014078e-01 1.05551255e+00 -2.89531462e-02 -4.67781216e-01
-1.04513907e+00 7.01004028e-01 2.52380013e-01 -1.34767163e+00
-6.72468841e-01 7.52258182e-01 5.13095781e-03 -3.14449996e-01
2.81133741e-01 -8.17638457e-01 -3.55480909e-01 -1.65137388e-02
9.80023623e-01 4.27166931e-02 5.67301095e-01 1.28462613e-01
-1.66049212e-01 -2.27827996e-01 -5.34537971e-01 -1.28226429e-01
1.42607844e+00 3.65803063e-01 1.75911382e-01 3.65344405e-01
9.95499671e-01 -1.60142019e-01 -4.52513725e-01 -4.33415361e-02
2.20856994e-01 1.38236284e-01 2.12404698e-01 -1.34710598e+00
-8.76793206e-01 9.53147352e-01 6.90724492e-01 4.04296637e-01
9.22775149e-01 8.36246163e-02 7.32389390e-01 4.55438107e-01
2.15112776e-01 -7.37765253e-01 7.04708546e-02 1.69364497e-01
9.34987485e-01 -1.55085564e+00 7.26715624e-02 -8.51923227e-02
-8.61761332e-01 1.35826802e+00 3.73848557e-01 -3.60720783e-01
3.07478845e-01 5.85683942e-01 2.11104274e-01 -5.43461978e-01
-8.50563467e-01 -1.50174182e-02 5.49481273e-01 3.47124428e-01
1.21823645e+00 2.99539924e-01 -7.74442494e-01 5.43420553e-01
-4.87097383e-01 5.19459546e-01 1.01064765e+00 8.60336423e-01
1.51755288e-02 -8.54262352e-01 -5.03631353e-01 9.37453151e-01
-1.05013192e+00 -1.32361650e-01 -6.52752876e-01 7.91110158e-01
-2.01898217e-01 7.53779292e-01 -8.78459215e-03 7.48881996e-02
3.54270995e-01 2.37000942e-01 5.56003213e-01 -9.85393345e-01
-8.99982154e-01 -2.91711092e-01 3.47764611e-01 -7.38110900e-01
-4.42925423e-01 -4.93173182e-01 -1.08165085e+00 1.27674475e-01
3.13115329e-01 -9.69489012e-03 5.53705692e-01 7.85343289e-01
1.88312158e-01 1.14317405e+00 -7.67327473e-02 -2.74524093e-01
-6.64331377e-01 -1.33196652e+00 -5.39057814e-02 5.03768146e-01
3.31052572e-01 -3.37305844e-01 1.07807517e-01 2.38463461e-01] | [15.035369873046875, -1.3583585023880005] |
4f253181-ed76-459d-8f78-2c1c243840a3 | reinforced-self-attention-network-a-hybrid-of | 1801.10296 | null | http://arxiv.org/abs/1801.10296v2 | http://arxiv.org/pdf/1801.10296v2.pdf | Reinforced Self-Attention Network: a Hybrid of Hard and Soft Attention for Sequence Modeling | Many natural language processing tasks solely rely on sparse dependencies
between a few tokens in a sentence. Soft attention mechanisms show promising
performance in modeling local/global dependencies by soft probabilities between
every two tokens, but they are not effective and efficient when applied to long
sentences. By contrast, hard attention mechanisms directly select a subset of
tokens but are difficult and inefficient to train due to their combinatorial
nature. In this paper, we integrate both soft and hard attention into one
context fusion model, "reinforced self-attention (ReSA)", for the mutual
benefit of each other. In ReSA, a hard attention trims a sequence for a soft
self-attention to process, while the soft attention feeds reward signals back
to facilitate the training of the hard one. For this purpose, we develop a
novel hard attention called "reinforced sequence sampling (RSS)", selecting
tokens in parallel and trained via policy gradient. Using two RSS modules, ReSA
efficiently extracts the sparse dependencies between each pair of selected
tokens. We finally propose an RNN/CNN-free sentence-encoding model, "reinforced
self-attention network (ReSAN)", solely based on ReSA. It achieves
state-of-the-art performance on both Stanford Natural Language Inference (SNLI)
and Sentences Involving Compositional Knowledge (SICK) datasets. | ['Tao Shen', 'Sen Wang', 'Jing Jiang', 'Chengqi Zhang', 'Tianyi Zhou', 'Guodong Long'] | 2018-01-31 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 3.03709298e-01 1.12337060e-01 -1.05147921e-02 -6.64007664e-01
-8.76913548e-01 -2.97793984e-01 6.20834887e-01 2.28132412e-01
-7.34867036e-01 9.52881634e-01 3.05280030e-01 -2.98654288e-01
2.86679268e-01 -7.77746379e-01 -9.50752139e-01 -6.82321787e-01
2.41570577e-01 3.59981984e-01 1.50831595e-01 -4.82911199e-01
1.27547294e-01 -4.51961718e-02 -1.35667348e+00 5.58454275e-01
1.14908803e+00 6.78303540e-01 1.01619995e+00 5.87964058e-01
-5.57482362e-01 1.20671749e+00 -4.67390776e-01 -3.26710552e-01
-1.81692034e-01 -3.77500921e-01 -1.02187836e+00 -4.84005094e-01
-3.54987034e-03 -2.45287642e-01 -9.86323357e-02 1.03590024e+00
5.80604255e-01 2.76657581e-01 4.25585657e-01 -6.26903415e-01
-8.75074089e-01 1.26292098e+00 -3.78001153e-01 4.15486485e-01
4.62292403e-01 2.26732001e-01 1.38041377e+00 -9.64707375e-01
4.33420300e-01 1.28802085e+00 4.00799781e-01 6.65683210e-01
-9.88288462e-01 -4.12022322e-01 6.11352742e-01 3.83420944e-01
-1.07362795e+00 -5.32031059e-01 7.74562895e-01 -2.37819403e-01
1.62060928e+00 3.27801615e-01 3.02429557e-01 1.31336606e+00
3.22002918e-01 1.13328159e+00 8.78233373e-01 -6.00118041e-01
3.23580682e-01 4.80339266e-02 4.27971184e-01 6.91474378e-01
-2.63275087e-01 -2.44899526e-01 -5.54326296e-01 -1.24616316e-02
3.39286983e-01 6.21493459e-02 -2.84720771e-02 5.51458716e-01
-1.13103092e+00 8.65347862e-01 3.96706849e-01 6.14761412e-01
-6.77684009e-01 1.57297835e-01 4.48268890e-01 2.94655621e-01
4.64413881e-01 1.64888740e-01 -6.23289466e-01 -1.80486307e-01
-6.89195931e-01 -3.06500625e-02 5.17646670e-01 8.74271035e-01
7.50462532e-01 3.05685364e-02 -6.86200678e-01 1.07331479e+00
3.52902025e-01 2.90630847e-01 8.41632068e-01 -3.41823280e-01
7.27828264e-01 4.97634768e-01 -1.14787519e-01 -6.15586877e-01
-1.75557107e-01 -4.56727833e-01 -1.04459465e+00 -1.73684746e-01
-5.75029058e-03 -2.62938052e-01 -8.96617949e-01 1.92045569e+00
1.23537742e-01 1.52834758e-01 7.07919672e-02 8.72883260e-01
8.16199780e-01 9.58182931e-01 4.36388165e-01 -1.70545027e-01
1.45871234e+00 -1.14185572e+00 -6.72706723e-01 -7.15456069e-01
5.70152879e-01 -5.93117356e-01 1.21106875e+00 8.71064812e-02
-1.42358541e+00 -6.61714494e-01 -7.63600290e-01 -4.20421124e-01
-4.60787773e-01 1.52055860e-01 3.17915142e-01 -6.32448345e-02
-9.60752428e-01 6.59303248e-01 -7.26361215e-01 -6.27479702e-02
3.01046789e-01 3.89358670e-01 1.08830556e-02 1.87372789e-02
-1.63473201e+00 9.03240383e-01 2.60000706e-01 4.90436792e-01
-8.19366932e-01 -3.58332634e-01 -1.11575162e+00 4.84728456e-01
4.19659942e-01 -8.68038177e-01 1.25222027e+00 -1.05202651e+00
-1.87778473e+00 6.68839574e-01 -5.36067784e-01 -6.53752387e-01
2.36910194e-01 -5.34329772e-01 -7.68429711e-02 -2.14075670e-02
2.78955579e-01 6.45835459e-01 8.93021822e-01 -9.32445824e-01
-3.23581070e-01 2.33661346e-02 1.73705012e-01 1.51065663e-01
-1.26084194e-01 4.10655528e-01 -1.76190317e-01 -7.45552123e-01
-4.18697178e-01 -5.75803459e-01 -3.95410389e-01 -6.64905488e-01
-6.64794505e-01 -6.73650265e-01 4.59684998e-01 -6.72060013e-01
1.36476731e+00 -1.94638801e+00 3.76094043e-01 -3.07092015e-02
-9.17892233e-02 6.71428382e-01 -3.68019998e-01 3.84331703e-01
5.26120095e-03 1.87713459e-01 -4.61479396e-01 -7.37398803e-01
6.85392879e-03 4.60175037e-01 -3.98149669e-01 -1.67222351e-01
7.68747330e-01 1.08888388e+00 -1.35289907e+00 -5.76552212e-01
1.47646204e-01 3.86835307e-01 -6.68788314e-01 5.05160093e-01
-5.00486553e-01 2.37779200e-01 -5.15508354e-01 3.23814601e-01
4.50501233e-01 -4.22105521e-01 3.93228561e-01 1.50097251e-01
-1.89794242e-01 9.48626041e-01 -7.56898165e-01 1.71601057e+00
-8.56143475e-01 1.81187272e-01 1.72171533e-01 -9.54353273e-01
9.43371177e-01 3.61871094e-01 -4.68753614e-02 -5.85881591e-01
2.73723006e-01 4.31872718e-02 4.46453765e-02 -8.00100923e-01
5.91729701e-01 -2.06354529e-01 -9.34157446e-02 3.99063855e-01
1.79540128e-01 1.75068468e-01 3.59788239e-01 4.02893931e-01
1.21694088e+00 2.37200081e-01 3.75246167e-01 -1.24427579e-01
9.13593709e-01 -5.70324540e-01 8.45423937e-01 9.36925173e-01
-7.73653314e-02 6.25359356e-01 4.60793972e-01 -1.83535516e-01
-8.49833429e-01 -8.61056387e-01 2.44176105e-01 1.58981776e+00
-1.93314329e-01 -5.41107774e-01 -6.44632459e-01 -1.06756592e+00
-1.00538157e-01 8.12816679e-01 -6.76203072e-01 -1.19073994e-01
-7.97428608e-01 -5.19242108e-01 3.12920004e-01 7.50015020e-01
4.74491417e-01 -1.89634562e+00 -2.77700156e-01 5.82364500e-01
-3.56405318e-01 -1.06460607e+00 -6.51993275e-01 4.87162709e-01
-3.93840611e-01 -6.39538348e-01 -4.87311125e-01 -1.02215350e+00
6.80393815e-01 1.13352537e-01 1.23932183e+00 2.49059647e-01
9.77603905e-03 -1.30451456e-01 -6.23385489e-01 -2.92266369e-01
-3.24756235e-01 3.32941204e-01 -1.41323209e-02 1.96440488e-01
3.22976768e-01 -5.31796098e-01 -3.99148285e-01 -1.89629227e-01
-8.03973436e-01 9.47319716e-02 9.80961800e-01 1.27969909e+00
5.15213728e-01 -4.19587463e-01 9.48229790e-01 -1.03239286e+00
7.33182669e-01 -8.32416713e-01 -1.00829326e-01 3.47304970e-01
-6.39899671e-02 4.04751569e-01 1.32920146e+00 -2.81696737e-01
-1.30740976e+00 -5.60359135e-02 -6.42007828e-01 -3.07995826e-01
-2.14942142e-01 7.26186931e-01 -1.53010227e-02 5.17923355e-01
2.55155414e-01 4.84587818e-01 -1.46133021e-01 -5.42130828e-01
1.89487144e-01 7.01121986e-01 3.45498472e-01 -7.23000526e-01
3.67552936e-01 7.69467801e-02 -5.28900445e-01 -8.05259466e-01
-1.24728942e+00 -3.75363320e-01 -4.42745060e-01 1.67100430e-01
9.96614993e-01 -7.04107344e-01 -7.01313674e-01 3.36425215e-01
-1.35620785e+00 -5.05437374e-01 -4.08374906e-01 4.51114774e-01
-3.45278889e-01 5.16736150e-01 -8.82786572e-01 -8.68920863e-01
-7.44625211e-01 -1.09345007e+00 1.08772814e+00 2.80019253e-01
-2.09790677e-01 -8.85655344e-01 -5.13449050e-02 3.65734160e-01
5.12012303e-01 -3.62000436e-01 6.66714966e-01 -6.97170019e-01
-4.35998470e-01 -6.45856708e-02 -2.34852254e-01 7.31649637e-01
-6.19278960e-02 -8.20519589e-03 -9.90086555e-01 -1.49632528e-01
-6.98861703e-02 -5.68886399e-01 1.16832530e+00 4.31120098e-01
1.32973051e+00 -4.64511067e-01 -1.06415085e-01 1.85440123e-01
1.17006516e+00 -3.17810774e-02 5.82774878e-01 1.13166049e-01
6.12569451e-01 3.04754347e-01 4.25706029e-01 4.80662972e-01
5.31613886e-01 3.15034598e-01 3.11714023e-01 5.05319834e-02
8.34816173e-02 -2.69709557e-01 6.88786507e-01 1.35298371e+00
-6.77912906e-02 -2.50702620e-01 -5.86437523e-01 5.20191610e-01
-2.00773787e+00 -1.15616715e+00 7.72076026e-02 1.86229157e+00
1.28161025e+00 6.25529110e-01 -2.38483667e-01 -1.08487047e-01
7.41831243e-01 2.40239382e-01 -3.53693128e-01 -7.49622762e-01
-2.92586982e-01 5.24085999e-01 6.59578070e-02 8.77328515e-01
-9.26823080e-01 1.18277395e+00 5.28574944e+00 9.62634027e-01
-9.44355011e-01 1.66903973e-01 5.11941195e-01 -4.72754762e-02
-5.22022307e-01 1.11417331e-01 -9.86066699e-01 8.26633751e-01
1.00529408e+00 9.90020186e-02 3.70774180e-01 7.88536191e-01
3.15102339e-01 -1.38557434e-01 -9.03771281e-01 4.04651672e-01
-7.66879022e-02 -1.39522099e+00 1.57702416e-01 -4.41288173e-01
5.02460420e-01 1.90628543e-01 -2.58813590e-01 7.62916088e-01
4.40486968e-01 -8.73382211e-01 7.04967916e-01 6.43373072e-01
4.52567011e-01 -5.34074903e-01 9.63047743e-01 7.23290384e-01
-1.20803475e+00 -9.18312371e-02 -4.25698668e-01 -3.73374969e-01
2.60806531e-01 7.17820108e-01 -7.98271239e-01 5.96266031e-01
5.69865108e-01 7.10164130e-01 -3.00893635e-01 5.33989966e-01
-6.87022328e-01 9.52219009e-01 -2.72814780e-01 -4.77828920e-01
5.05028486e-01 -1.22035533e-01 5.14335454e-01 1.75062549e+00
-1.53251410e-01 1.65526763e-01 1.67693779e-01 9.71782863e-01
-1.95023358e-01 1.19971499e-01 -3.27180922e-01 1.92154109e-01
4.67814863e-01 1.26637042e+00 -3.35442185e-01 -7.62643814e-01
-4.13684100e-01 8.59704971e-01 9.78124559e-01 2.58811772e-01
-7.87588716e-01 -6.54959261e-01 4.78807688e-01 -2.87035584e-01
5.76347232e-01 -1.03316240e-01 -3.17395210e-01 -1.30338073e+00
2.10985348e-01 -5.65500379e-01 3.00236225e-01 -7.37913430e-01
-1.56214511e+00 7.28029668e-01 -3.69336575e-01 -8.63708138e-01
-3.08204561e-01 -2.84358948e-01 -9.55953717e-01 1.20415044e+00
-1.95659316e+00 -9.67702866e-01 1.05981022e-01 4.73883271e-01
1.02710545e+00 7.74048045e-02 7.95495152e-01 2.34633416e-01
-1.01922786e+00 6.47112429e-01 -7.54563650e-03 2.62476891e-01
5.22280633e-01 -1.45096409e+00 4.86598164e-01 7.88996935e-01
-1.49468288e-01 8.60715568e-01 3.51728976e-01 -7.03199267e-01
-1.21787155e+00 -1.08521509e+00 1.54545748e+00 4.15193737e-02
5.24542987e-01 -4.88555372e-01 -1.14098704e+00 8.11795831e-01
5.46850801e-01 -1.28253205e-02 5.19675791e-01 2.16874212e-01
-1.93104088e-01 6.56199753e-02 -8.79609346e-01 4.62260425e-01
8.11369896e-01 -5.53806424e-01 -1.00021315e+00 3.50284517e-01
8.94052207e-01 -1.90876722e-01 -6.59296751e-01 2.61047035e-01
1.60444975e-01 -8.41753125e-01 8.54080617e-01 -5.90404272e-01
8.28290343e-01 -1.66385040e-01 9.10863280e-02 -1.30732429e+00
-7.08403945e-01 -7.81189084e-01 -1.92178860e-01 1.34956133e+00
5.26669443e-01 -6.45190716e-01 4.58692253e-01 1.85502112e-01
-5.88619232e-01 -1.15338457e+00 -8.78685296e-01 -6.64025962e-01
8.24975297e-02 -3.68558049e-01 5.48848808e-01 9.37112570e-01
2.78108090e-01 8.12695265e-01 -3.14193279e-01 1.19066305e-01
2.98093915e-01 5.34946658e-02 2.10946575e-01 -6.88674986e-01
-5.69472849e-01 -6.15899026e-01 1.25529438e-01 -1.36172700e+00
5.06451011e-01 -1.13286924e+00 4.52987254e-01 -1.57751560e+00
2.87850499e-01 -4.87270564e-01 -5.03261447e-01 7.52466202e-01
-7.79182792e-01 -1.45488694e-01 2.16999203e-01 -2.47003213e-02
-8.18069279e-01 9.26004827e-01 1.09699464e+00 -1.85498819e-01
-1.88565284e-01 -1.84630007e-01 -6.24071538e-01 7.39635050e-01
8.40923369e-01 -3.40931773e-01 -2.05302015e-02 -6.22687519e-01
2.88793266e-01 3.13500911e-02 1.45301640e-01 -6.43239617e-01
2.26962537e-01 -2.31716394e-01 2.01234996e-01 -7.42933214e-01
2.31029645e-01 -2.69392788e-01 -6.45609677e-01 3.42437625e-01
-7.40565836e-01 -8.31198767e-02 9.08901766e-02 4.04091418e-01
-2.36009598e-01 -5.49950302e-01 5.33240438e-01 -3.56190562e-01
-4.78807062e-01 1.30176842e-01 -5.21675110e-01 1.68020323e-01
7.29508162e-01 3.78954083e-01 -1.74733907e-01 -3.20024416e-02
-8.35360467e-01 5.55579364e-01 -2.59869158e-01 4.17613715e-01
7.00475872e-01 -1.16637790e+00 -9.34364736e-01 3.63436788e-01
-2.55245835e-01 4.15488183e-01 4.90129948e-01 6.58843517e-01
-1.06286064e-01 4.18644369e-01 4.04496305e-02 -3.02415639e-01
-1.10006464e+00 5.89345276e-01 -6.40791357e-02 -7.10886896e-01
-4.39289391e-01 1.21645129e+00 7.69839734e-02 -4.78797883e-01
2.15484351e-01 -6.23787344e-01 -2.48034254e-01 -4.87718433e-02
6.22443616e-01 4.80218529e-04 4.82657216e-02 -3.45020026e-01
-4.51787174e-01 1.78007320e-01 -4.89548028e-01 1.06046207e-01
1.41516876e+00 -1.44665211e-01 -2.26886421e-01 5.01604021e-01
1.09926343e+00 1.46162957e-02 -1.21022034e+00 -5.09925783e-01
-1.62111178e-01 -1.50683850e-01 -1.68973044e-01 -7.80542195e-01
-7.08951294e-01 1.00680649e+00 -3.00874621e-01 5.08176088e-01
9.68289375e-01 1.86986383e-02 1.12025857e+00 3.57021421e-01
-1.27639055e-01 -1.19665420e+00 9.94323790e-02 1.09973931e+00
9.69998658e-01 -1.24266303e+00 -3.14618856e-01 -2.11755067e-01
-8.22350979e-01 9.94790435e-01 7.24553704e-01 -3.29230100e-01
5.56916535e-01 2.93839753e-01 -3.11274409e-01 1.85385272e-01
-1.23673391e+00 -3.42428684e-01 1.97730605e-02 2.22328410e-01
7.49910951e-01 5.08757904e-02 -3.77147079e-01 8.48479033e-01
-1.65038019e-01 8.95681009e-02 2.17376038e-01 1.21036851e+00
-6.33051932e-01 -1.21331120e+00 -1.13910876e-01 5.76381087e-01
-4.26274955e-01 -6.08021736e-01 -1.50577173e-01 1.05265550e-01
1.99865103e-01 9.00404632e-01 9.18559432e-02 -3.62072378e-01
9.61140022e-02 2.63386726e-01 2.29311988e-01 -8.99184287e-01
-1.16732574e+00 -1.51636273e-01 2.36370012e-01 -2.90296793e-01
-2.64121175e-01 -6.75112665e-01 -1.39568007e+00 -1.95354417e-01
-2.62016892e-01 1.97201744e-01 4.93637949e-01 1.26954269e+00
5.60384572e-01 8.69018316e-01 8.07771027e-01 -8.59542608e-01
-9.25974011e-01 -1.22591209e+00 -2.39212498e-01 4.22997832e-01
4.03388023e-01 -2.84599721e-01 -2.07897112e-01 -5.81425987e-02] | [10.9798583984375, 8.706693649291992] |
cac80aec-b136-4789-b650-0b0c07f1db55 | heterogeneous-tri-stream-clustering-network | 2301.04451 | null | https://arxiv.org/abs/2301.04451v1 | https://arxiv.org/pdf/2301.04451v1.pdf | Heterogeneous Tri-stream Clustering Network | Contrastive deep clustering has recently gained significant attention with its ability of joint contrastive learning and clustering via deep neural networks. Despite the rapid progress, previous works mostly require both positive and negative sample pairs for contrastive clustering, which rely on a relative large batch-size. Moreover, they typically adopt a two-stream architecture with two augmented views, which overlook the possibility and potential benefits of multi-stream architectures (especially with heterogeneous or hybrid networks). In light of this, this paper presents a new end-to-end deep clustering approach termed Heterogeneous Tri-stream Clustering Network (HTCN). The tri-stream architecture in HTCN consists of three main components, including two weight-sharing online networks and a target network, where the parameters of the target network are the exponential moving average of that of the online networks. Notably, the two online networks are trained by simultaneously (i) predicting the instance representations of the target network and (ii) enforcing the consistency between the cluster representations of the target network and that of the two online networks. Experimental results on four challenging image datasets demonstrate the superiority of HTCN over the state-of-the-art deep clustering approaches. The code is available at https://github.com/dengxiaozhi/HTCN. | ['Chang-Dong Wang', 'Dong Huang', 'Xiaozhi Deng'] | 2023-01-11 | null | null | null | null | ['deep-clustering', 'deep-clustering'] | ['miscellaneous', 'natural-language-processing'] | [-1.77895263e-01 -2.58820206e-01 7.32863322e-02 -3.47306341e-01
-6.37229323e-01 -3.06407273e-01 6.63029909e-01 -6.87201843e-02
-3.88825208e-01 7.43614789e-03 -5.37082180e-02 1.21227257e-01
-3.86929326e-03 -4.75214928e-01 -7.16341794e-01 -9.70337987e-01
-1.75712958e-01 5.92851341e-01 1.70271978e-01 8.74946937e-02
-1.33269385e-01 -1.12785444e-01 -1.51433837e+00 4.66499537e-01
6.07132971e-01 1.28456569e+00 1.79744288e-01 4.98553902e-01
7.01256245e-02 7.63335228e-01 -3.19778681e-01 -4.39397544e-01
4.33979988e-01 -5.43103039e-01 -3.16618800e-01 3.78263235e-01
1.23594470e-01 -1.97904065e-01 -6.71777129e-01 1.07172680e+00
6.87293231e-01 2.52552181e-01 3.90167594e-01 -1.63894057e+00
-7.11382627e-01 7.41085052e-01 -7.25137234e-01 1.48717258e-02
-2.40188912e-01 2.27112368e-01 1.01555288e+00 -1.00247777e+00
3.64608526e-01 1.00028729e+00 5.00366151e-01 3.54028314e-01
-1.05545366e+00 -9.89435196e-01 5.81227541e-01 4.80613291e-01
-1.63969159e+00 -5.31761765e-01 8.98469627e-01 -4.95529860e-01
6.95661664e-01 -2.75501102e-01 8.47330809e-01 1.05255151e+00
-3.19852382e-01 9.36094940e-01 6.72863245e-01 3.32988352e-02
3.60807985e-01 6.44673258e-02 -1.77471831e-01 4.14091855e-01
4.77977619e-02 -8.95924345e-02 -4.84578967e-01 2.37594638e-03
6.42566741e-01 5.49585998e-01 -2.50152767e-01 -7.07827747e-01
-1.05200315e+00 8.28072667e-01 6.72213972e-01 2.77474493e-01
-5.29790878e-01 2.06486240e-01 6.34383380e-01 6.92877322e-02
5.59236348e-01 -1.32720202e-01 -3.27462196e-01 6.80865049e-02
-9.99838114e-01 -1.20801993e-01 6.20618641e-01 1.05767632e+00
5.98521829e-01 -4.33669165e-02 1.20344438e-01 8.26144338e-01
3.41983318e-01 1.40105173e-01 6.57020986e-01 -7.90483415e-01
6.20718658e-01 6.79786861e-01 -2.04122543e-01 -8.95646751e-01
-1.44277036e-01 -6.23025715e-01 -1.22406960e+00 -1.16140842e-01
3.04225296e-01 -1.85071349e-01 -5.80302417e-01 1.73521280e+00
5.13715804e-01 4.97560114e-01 -9.90454853e-02 9.86389637e-01
5.38733184e-01 8.35148692e-01 -1.62505209e-01 -1.01834051e-01
9.20200646e-01 -1.17656791e+00 -5.19348145e-01 -9.86675397e-02
3.37620825e-01 -5.15134335e-01 8.03400397e-01 4.02344465e-01
-1.22089994e+00 -5.10067880e-01 -1.11176825e+00 6.12593852e-02
-3.44467789e-01 4.30674441e-02 3.16589266e-01 9.20156389e-02
-1.14044058e+00 6.15025103e-01 -1.12178707e+00 -1.59769356e-01
8.05916607e-01 4.29074436e-01 -1.42865330e-01 -2.79336244e-01
-1.00826812e+00 1.72567815e-01 4.03593630e-01 4.67793137e-01
-8.93598557e-01 -5.69173813e-01 -6.65507078e-01 3.18470746e-01
4.02481616e-01 -5.35968304e-01 1.05826318e+00 -1.50670993e+00
-1.42686915e+00 5.62615693e-01 -6.42107353e-02 -3.02118480e-01
7.95106769e-01 -1.00561917e-01 -1.93336368e-01 5.19476712e-01
7.89397731e-02 7.91776299e-01 9.40385759e-01 -1.37221968e+00
-7.50781775e-01 -4.93647963e-01 -3.77133071e-01 4.45286512e-01
-6.50368512e-01 -1.14369571e-01 -1.03145015e+00 -4.41821575e-01
2.14665569e-02 -8.51605058e-01 -1.59632295e-01 2.74144351e-01
-4.47265208e-01 -3.60171616e-01 8.90687585e-01 -4.11024600e-01
1.16487360e+00 -2.52698851e+00 2.54866362e-01 2.18114138e-01
4.70076889e-01 3.20936531e-01 -2.30221555e-01 3.91535640e-01
-4.05901849e-01 -4.38378230e-02 -2.81711489e-01 -6.04277611e-01
1.24111600e-01 -1.10712692e-01 3.14469449e-02 7.24710941e-01
1.94784954e-01 7.82168388e-01 -9.77895141e-01 -3.34336013e-01
9.96056795e-02 6.85370207e-01 -3.77804071e-01 2.73160100e-01
4.27288748e-02 2.66068310e-01 -1.15648948e-01 3.70484024e-01
6.50323629e-01 -6.86117947e-01 4.30398434e-01 -1.64741099e-01
-3.48542035e-02 7.11993724e-02 -1.20913684e+00 1.51084137e+00
4.32600500e-04 4.41710204e-01 3.93731177e-01 -1.15366900e+00
5.92447877e-01 5.93185782e-01 7.39893675e-01 -6.55595124e-01
1.73017427e-01 1.43530220e-01 1.99678481e-01 -3.14990908e-01
6.70439154e-02 -2.04564765e-01 2.37805173e-01 6.48879766e-01
1.31216154e-01 5.21598399e-01 2.84174681e-01 5.73272228e-01
7.64403582e-01 -1.76798835e-01 -1.99290574e-01 -5.91199379e-03
3.23318005e-01 -4.10882443e-01 7.12965429e-01 4.64754969e-01
-4.32259142e-01 8.12191725e-01 7.08963513e-01 -2.17443973e-01
-1.14836085e+00 -9.90813732e-01 3.07095945e-01 1.04240406e+00
1.88855737e-01 -3.30071121e-01 -5.72299540e-01 -7.09808230e-01
-1.18048526e-01 1.89851180e-01 -6.51651084e-01 -1.16366901e-01
-4.35136080e-01 -6.59202397e-01 3.36172462e-01 6.35107934e-01
5.48364580e-01 -7.37526834e-01 -3.45993280e-01 1.23399906e-01
-2.89744616e-01 -1.02630293e+00 -6.71742260e-01 1.74336672e-01
-6.94166839e-01 -1.19155586e+00 -6.85183525e-01 -9.08111572e-01
7.23492563e-01 6.76030159e-01 8.84854376e-01 3.35585088e-01
2.30082273e-02 3.25658232e-01 -4.11991954e-01 -2.41022706e-01
-2.14191936e-02 1.53899118e-01 -1.98152699e-02 4.83537823e-01
3.35973501e-01 -6.87528789e-01 -1.01195288e+00 2.31285602e-01
-1.12866580e+00 1.11525074e-01 4.87289906e-01 7.32276857e-01
5.67670166e-01 2.22967625e-01 7.01333404e-01 -9.63616788e-01
1.91141814e-01 -9.13427114e-01 -4.41176921e-01 4.15011235e-02
-5.73423326e-01 -4.99382198e-01 9.50772166e-01 -4.64947701e-01
-8.17046762e-01 1.43504828e-01 2.14844923e-02 -1.07436550e+00
9.34763104e-02 5.46708226e-01 -3.01097304e-01 5.31295657e-01
2.33261332e-01 1.84098259e-01 2.65325189e-01 -9.70657691e-02
3.98074716e-01 5.52879930e-01 4.50574934e-01 -3.05891722e-01
8.75521243e-01 7.43917167e-01 -4.87748414e-01 -5.08020222e-01
-6.79012597e-01 -7.80924916e-01 -7.17004657e-01 -4.92668837e-01
7.85433352e-01 -1.34279382e+00 -5.98538816e-01 8.40589166e-01
-7.28855371e-01 -6.16039932e-01 -2.70758182e-01 4.41237599e-01
-3.08485985e-01 2.89836913e-01 -8.21738243e-01 -6.06057048e-01
-2.18270779e-01 -9.44813311e-01 8.38398337e-01 1.81897417e-01
1.85680613e-01 -1.06871605e+00 -3.44427466e-01 2.87127912e-01
2.76133806e-01 8.42449218e-02 7.56699562e-01 -8.46895874e-01
-5.39370239e-01 -2.90338516e-01 -3.31157237e-01 4.97044444e-01
-3.96637358e-02 2.70567894e-01 -1.09529805e+00 -6.04096889e-01
5.89816198e-02 -4.45330590e-01 8.77785385e-01 4.14505571e-01
1.29996562e+00 -3.35277498e-01 -1.56558663e-01 7.14381516e-01
1.56906295e+00 1.26053169e-01 4.96206611e-01 1.83093920e-01
9.68464494e-01 6.51061296e-01 1.97090447e-01 5.52024186e-01
8.14704597e-01 3.12020600e-01 7.01658785e-01 -1.07395947e-01
2.59299129e-02 -1.97034761e-01 4.36193705e-01 1.21429121e+00
2.64608949e-01 -2.25843176e-01 -9.06339347e-01 7.50545859e-01
-2.20638633e+00 -1.11593366e+00 -1.84109192e-02 2.22091150e+00
6.07608378e-01 6.72965497e-02 3.08199674e-01 1.01278588e-01
1.01674294e+00 2.24656209e-01 -9.61840212e-01 1.77695960e-01
-7.03660697e-02 -2.16900826e-01 -1.27213949e-03 1.42487094e-01
-1.22545385e+00 5.03279567e-01 4.25685358e+00 7.02998936e-01
-1.27449000e+00 2.89333731e-01 9.02142107e-01 -5.64793348e-01
-4.22202013e-02 -1.37728572e-01 -4.30552363e-01 7.01052129e-01
9.41308498e-01 -1.00198559e-01 4.28757340e-01 6.55761957e-01
1.23457126e-01 7.74894133e-02 -1.21395242e+00 8.91356349e-01
6.75884560e-02 -1.16947019e+00 1.25001660e-02 2.72222469e-03
7.55615234e-01 4.79962587e-01 3.52878690e-01 3.61325294e-01
2.07651958e-01 -6.36475801e-01 9.43952978e-01 2.31677979e-01
6.28668129e-01 -9.13761854e-01 7.38795280e-01 4.06166196e-01
-1.36448574e+00 -2.33608499e-01 -1.62218899e-01 4.39858101e-02
-1.93804324e-01 6.27286315e-01 -3.53205144e-01 5.62000394e-01
1.03131223e+00 1.15713251e+00 -5.21019518e-01 9.61137772e-01
3.81559506e-02 6.49997056e-01 -4.09394830e-01 2.09317058e-01
4.98691350e-01 -5.09758532e-01 1.29677132e-01 1.21006811e+00
2.21209630e-01 6.41972497e-02 3.00803602e-01 7.43652761e-01
-4.16176707e-01 -9.31957588e-02 -3.53322417e-01 -6.91021159e-02
7.74171054e-01 1.46565580e+00 -8.75925422e-01 -5.37819922e-01
-5.74443817e-01 9.09155548e-01 5.30809343e-01 4.69641685e-01
-8.15557241e-01 -2.34684810e-01 3.71095240e-01 3.96834454e-03
7.66203523e-01 -7.58023188e-02 -2.00780407e-01 -1.20880175e+00
2.15261713e-01 -7.64682174e-01 5.87537944e-01 -4.65339363e-01
-1.72716236e+00 5.15314579e-01 -1.98590696e-01 -1.55666375e+00
1.48123115e-01 -1.78464264e-01 -8.59061062e-01 6.23718739e-01
-1.41886473e+00 -1.14928114e+00 -3.63443792e-01 8.35684538e-01
4.50510949e-01 -3.09926253e-02 3.76202524e-01 6.99514508e-01
-1.13307297e+00 6.53520167e-01 4.13706481e-01 3.83654386e-01
7.20104992e-01 -1.22098196e+00 6.69025555e-02 9.08655941e-01
-2.10154384e-01 5.34578264e-01 1.71477720e-01 -3.64391387e-01
-1.16022336e+00 -1.59576273e+00 5.13223469e-01 -1.11162916e-01
7.20129609e-01 -6.15470648e-01 -1.05976224e+00 5.47687173e-01
4.19122368e-01 3.02412629e-01 8.86784196e-01 -1.27561122e-01
-4.91260439e-01 -3.05444270e-01 -7.18625724e-01 4.68063414e-01
7.37339973e-01 -5.04834712e-01 1.30648449e-01 3.60000700e-01
5.45399904e-01 -1.60172239e-01 -8.47478509e-01 4.01059277e-02
4.57397997e-01 -1.20764303e+00 6.63170338e-01 -1.86863258e-01
7.91877925e-01 -3.39808375e-01 -8.98095593e-02 -1.30806661e+00
-3.99074942e-01 -3.51505548e-01 -2.29170546e-01 1.23134816e+00
2.82421499e-01 -5.22501230e-01 7.94980109e-01 4.37738657e-01
-1.93602711e-01 -1.08846378e+00 -7.54439712e-01 -6.56459808e-01
1.53727606e-01 -1.48029566e-01 3.77096087e-01 1.23233688e+00
-3.73260044e-02 6.54361546e-01 -1.72412232e-01 1.34150267e-01
5.94760120e-01 1.03184797e-01 5.77934921e-01 -1.06955230e+00
-3.73210549e-01 -5.86708248e-01 -1.08534679e-01 -9.02563632e-01
6.30750135e-02 -1.09732580e+00 1.58738956e-01 -1.45498776e+00
5.18398166e-01 -4.03559864e-01 -4.35046822e-01 3.61498713e-01
-4.19018835e-01 1.50898755e-01 4.17328835e-01 5.75910866e-01
-1.05192947e+00 9.40947354e-01 9.11940455e-01 -7.17788786e-02
-7.31246546e-02 -5.94330467e-02 -7.54220963e-01 7.31588125e-01
8.57915521e-01 -5.71451724e-01 -6.66299164e-01 -6.02254987e-01
2.38644645e-01 -1.14872023e-01 3.50926191e-01 -8.82807672e-01
7.52898335e-01 3.88153940e-01 5.44733703e-01 -6.27696812e-01
3.01655114e-01 -9.24849212e-01 -4.86991787e-03 4.08891737e-01
-3.54899853e-01 2.90067941e-01 -8.32244530e-02 1.06311500e+00
-4.22935754e-01 2.59633362e-01 9.48556900e-01 -1.98805332e-01
-3.32669020e-01 8.36687148e-01 -2.72230715e-01 1.35987699e-01
1.08581614e+00 -1.43033355e-01 -3.26566577e-01 -5.07253051e-01
-6.16442859e-01 7.20453203e-01 4.41328615e-01 3.17692965e-01
7.17714429e-01 -1.32887816e+00 -6.55676365e-01 1.89759597e-01
-1.60218384e-02 5.36342919e-01 5.90789437e-01 1.22004557e+00
-3.04908723e-01 -9.38048437e-02 2.79143713e-02 -7.71050572e-01
-8.20746601e-01 7.15730727e-01 3.98880690e-01 -2.75517344e-01
-6.30990744e-01 9.42550480e-01 4.99240220e-01 -5.13327599e-01
5.15637815e-01 2.81893611e-01 -1.00432094e-02 3.13697457e-01
2.90901005e-01 4.53822017e-01 8.99049714e-02 -4.57926512e-01
-2.92528391e-01 9.26872715e-02 -3.89185131e-01 -1.36277869e-01
1.57098866e+00 -1.77534595e-01 -2.43007481e-01 7.34745502e-01
1.45813584e+00 -6.93253219e-01 -1.64051998e+00 -4.29782778e-01
-2.59793222e-01 -1.67566717e-01 -6.52993843e-02 -3.92189384e-01
-1.68844533e+00 1.13723445e+00 5.50296605e-01 1.13372438e-01
1.28832233e+00 -1.07051820e-01 9.28508043e-01 1.30520329e-01
-8.41023847e-02 -1.30671799e+00 2.90683359e-01 2.68574417e-01
5.91781378e-01 -1.39029825e+00 -1.34451836e-01 -5.12580760e-03
-6.92110121e-01 8.70264769e-01 8.31982136e-01 -3.95240366e-01
1.01096821e+00 4.29513641e-02 1.33753151e-01 -3.49882752e-01
-1.26750195e+00 -6.28336519e-02 -3.69056575e-02 3.86240155e-01
3.22659194e-01 1.72471292e-02 1.77821875e-01 6.48947358e-01
2.94343084e-01 -2.08623096e-01 3.38269591e-01 8.01220596e-01
-6.49088249e-02 -7.35947490e-01 -8.95325243e-02 5.60714602e-01
-3.69480222e-01 -5.37978206e-03 -4.64268625e-01 7.97942042e-01
1.45504147e-01 1.08146846e+00 3.09719145e-01 -5.37316620e-01
1.75156608e-01 -1.07184432e-01 8.20469707e-02 -5.83575606e-01
-9.29811954e-01 4.78446871e-01 -5.12544394e-01 -6.24481678e-01
-5.29662669e-01 -9.09338295e-01 -1.16911137e+00 -2.47560918e-01
-3.34352940e-01 1.77773058e-01 4.70121741e-01 7.14483142e-01
5.48526347e-01 4.78626907e-01 1.00414693e+00 -1.08541965e+00
-5.65709710e-01 -9.36991394e-01 -7.04064071e-01 5.73451161e-01
2.04952270e-01 -2.83179611e-01 -4.29193050e-01 1.82376131e-01] | [9.117911338806152, 3.327848196029663] |
343cd165-93b4-486b-b976-0ced44932ed1 | author-profiling-for-abuse-detection | null | null | https://aclanthology.org/C18-1093 | https://aclanthology.org/C18-1093.pdf | Author Profiling for Abuse Detection | The rapid growth of social media in recent years has fed into some highly undesirable phenomena such as proliferation of hateful and offensive language on the Internet. Previous research suggests that such abusive content tends to come from users who share a set of common stereotypes and form communities around them. The current state-of-the-art approaches to abuse detection are oblivious to user and community information and rely entirely on textual (i.e., lexical and semantic) cues. In this paper, we propose a novel approach to this problem that incorporates community-based profiling features of Twitter users. Experimenting with a dataset of 16k tweets, we show that our methods significantly outperform the current state of the art in abuse detection. Further, we conduct a qualitative analysis of model characteristics. We release our code, pre-trained models and all the resources used in the public domain. | ['Ekaterina Shutova', 'Helen Yannakoudakis', 'Marco del Tredici', 'Pushkar Mishra'] | 2018-08-01 | author-profiling-for-abuse-detection-1 | https://aclanthology.org/C18-1093 | https://aclanthology.org/C18-1093.pdf | coling-2018-8 | ['abuse-detection'] | ['natural-language-processing'] | [-2.71973073e-01 -2.98759133e-01 -5.46506763e-01 -1.70739457e-01
-5.03183484e-01 -7.74796128e-01 7.44606078e-01 7.39152193e-01
-4.68768895e-01 4.96701598e-01 5.92502713e-01 -4.28136677e-01
2.70591229e-01 -7.52243698e-01 -8.46738145e-02 -5.10104410e-02
-1.70552462e-01 8.94652456e-02 2.15094239e-01 -4.57937002e-01
6.44334733e-01 2.58314073e-01 -8.89904022e-01 3.14304292e-01
9.41403508e-01 5.16712487e-01 -6.44893944e-01 2.50493705e-01
-1.99083313e-01 1.05914104e+00 -7.01065838e-01 -9.40069616e-01
-1.23661272e-01 -1.97075531e-01 -9.11862850e-01 8.74808151e-03
1.99430928e-01 -4.44404125e-01 -4.76621628e-01 1.37405002e+00
2.00313926e-01 -1.30558401e-01 5.78074753e-01 -1.26124001e+00
-5.66749811e-01 7.25443244e-01 -1.02613997e+00 6.01710021e-01
4.70171720e-01 7.28191361e-02 9.54035759e-01 -6.69668376e-01
9.10720587e-01 1.16696084e+00 1.04188800e+00 2.40204170e-01
-1.17163610e+00 -1.07051134e+00 -1.38718411e-01 2.58544385e-02
-1.35038960e+00 -4.91412371e-01 9.72158015e-01 -9.80849981e-01
6.76654994e-01 1.46132946e-01 6.76244676e-01 1.54964685e+00
-3.00110728e-01 5.91089308e-01 1.15901721e+00 -3.02044481e-01
1.12610951e-01 3.98489922e-01 4.67207193e-01 6.56867266e-01
6.28078282e-01 -6.05827987e-01 -4.36437428e-01 -1.08215070e+00
2.10038275e-01 1.66022301e-01 1.17862031e-01 -1.31324038e-01
-2.22705603e-01 1.51316464e+00 2.31093422e-01 8.09488893e-01
-1.57905594e-01 -2.11602837e-01 8.10057223e-01 2.71362271e-02
9.90717173e-01 5.73751926e-01 9.86664221e-02 -4.83704329e-01
-1.07448161e+00 3.60959113e-01 9.34854150e-01 4.00681615e-01
6.60347283e-01 -4.47571188e-01 4.02036875e-01 1.17289507e+00
4.29371297e-01 1.95790425e-01 5.18620968e-01 -6.91217959e-01
6.52449191e-01 8.29154730e-01 -7.28208795e-02 -1.69026363e+00
-2.99955070e-01 -3.79903495e-01 -3.78803521e-01 -3.42519492e-01
5.69328845e-01 -3.80144596e-01 -4.30604041e-01 1.61609435e+00
3.03468615e-01 -1.98309161e-02 -7.49694824e-01 4.62137073e-01
8.61614794e-02 3.22790444e-01 2.30887830e-01 -1.12042643e-01
9.67158914e-01 -4.36328769e-01 -6.95454836e-01 -2.32204676e-01
1.12735939e+00 -5.95692158e-01 7.29236126e-01 2.93482751e-01
-8.34111571e-01 4.06690001e-01 -8.06772828e-01 1.57865435e-01
-6.35806918e-01 -7.08157837e-01 6.35938525e-01 1.23499632e+00
-5.68450511e-01 7.05636978e-01 -7.02272117e-01 -8.47808182e-01
9.05164480e-01 -8.22013542e-02 -3.40649128e-01 -1.53666465e-02
-1.00541103e+00 9.15537715e-01 6.86857626e-02 -4.81918246e-01
-4.58710492e-01 -4.73232538e-01 -7.25685358e-01 -4.06465083e-01
5.06192684e-01 -6.29506335e-02 1.22950685e+00 -1.06475866e+00
-8.85301232e-01 1.04768014e+00 6.37151226e-02 -2.67106324e-01
4.39724028e-01 -4.60467279e-01 -4.60648775e-01 2.03573942e-01
3.68844658e-01 -2.14099646e-01 7.24235475e-01 -1.20808363e+00
-2.47381240e-01 -7.47635543e-01 3.27654108e-02 -4.23744828e-01
-1.17433655e+00 9.32134509e-01 7.18794614e-02 -7.07939684e-01
-3.09190542e-01 -9.28474784e-01 -1.53287515e-01 -1.60757989e-01
-7.08595455e-01 -1.09433256e-01 9.81066167e-01 -6.59953177e-01
2.07286143e+00 -2.11381602e+00 -3.56330420e-03 4.82092708e-01
6.62452042e-01 5.03681481e-01 4.51869339e-01 1.04006743e+00
1.97874442e-01 9.77355719e-01 -2.59243101e-01 -6.97441816e-01
-1.95936769e-01 1.63505167e-01 -2.95904309e-01 9.54922020e-01
-1.25536650e-01 4.51938868e-01 -1.26356876e+00 -6.06275439e-01
-1.94143534e-01 1.55870363e-01 -8.18681419e-01 -8.25823247e-02
1.56610280e-01 2.04188719e-01 -7.35978544e-01 5.88854373e-01
5.82499921e-01 -3.15291107e-01 2.90944993e-01 4.32020605e-01
-3.29891145e-01 5.65968275e-01 -4.40126330e-01 1.35393512e+00
-2.47493424e-02 7.43277609e-01 3.87049198e-01 -5.76250374e-01
4.66715634e-01 1.34745118e-04 4.75899637e-01 -2.03198180e-01
5.63054860e-01 2.84724563e-01 1.02786548e-01 -6.56036317e-01
4.14163083e-01 -1.11555919e-01 -2.15715379e-01 8.35154414e-01
-2.05787309e-02 3.27861011e-01 1.81111515e-01 6.53518140e-01
1.28531110e+00 -5.25910199e-01 4.89594102e-01 -2.22409442e-01
1.79965243e-01 1.20772839e-01 2.85842180e-01 6.96049988e-01
-5.57936490e-01 2.04091460e-01 7.72419333e-01 -1.03765190e-01
-1.17010629e+00 -5.81265569e-01 -1.12536885e-01 1.21474731e+00
-1.55851573e-01 -9.04932797e-01 -1.06682491e+00 -8.04160535e-01
2.38788322e-01 6.40728235e-01 -6.70433879e-01 -3.06384023e-02
-5.60669005e-01 -7.79421866e-01 1.02689219e+00 1.04853116e-01
3.77438605e-01 -5.81402242e-01 -2.69118696e-01 1.97929174e-01
-2.27888823e-01 -1.09576666e+00 -2.08058149e-01 -5.83170235e-01
-6.48564696e-01 -1.17079353e+00 -3.07522923e-01 6.01790324e-02
4.06500697e-01 4.71459210e-01 7.59472132e-01 5.73285043e-01
-2.68079609e-01 -8.02410915e-02 -7.13640332e-01 -4.54513937e-01
-4.61573184e-01 5.04252911e-01 1.21252067e-01 2.72620380e-01
6.51857495e-01 -9.88906085e-01 -2.01016203e-01 -2.28569303e-02
-1.00709128e+00 -3.50330085e-01 -5.54038584e-02 3.38401407e-01
-5.76560378e-01 -3.31595182e-01 4.03469563e-01 -1.51305449e+00
9.11615491e-01 -1.33708930e+00 5.45074455e-02 -4.65487212e-01
-5.01731753e-01 -5.14354229e-01 4.46573436e-01 -3.96035314e-01
-7.04193652e-01 -6.59540653e-01 -1.68611437e-01 -3.42429042e-01
-2.23135620e-01 7.70661592e-01 3.30731660e-01 -1.06116138e-01
9.84064341e-01 -3.68658572e-01 7.91880041e-02 -7.86058545e-01
9.98628289e-02 1.13343823e+00 -9.28985104e-02 -4.07174826e-01
1.07476556e+00 6.61718786e-01 -4.46041852e-01 -1.19334769e+00
-1.10181630e+00 -8.97990584e-01 -5.61452746e-01 -2.44835481e-01
5.76327682e-01 -6.47324443e-01 -3.88726294e-01 8.14483464e-01
-1.20121562e+00 -1.45976678e-01 4.98375952e-01 2.38346066e-02
-6.55654911e-03 7.73842573e-01 -1.08187318e+00 -1.19928300e+00
-1.04770608e-01 -6.11420691e-01 4.80576128e-01 -8.55705664e-02
-9.37548935e-01 -1.14250052e+00 6.11961424e-01 6.23949826e-01
6.08065665e-01 6.49456620e-01 8.54744494e-01 -1.27480924e+00
1.44918606e-01 -2.92134792e-01 -2.58665979e-01 9.87388790e-02
2.37048999e-03 2.52054036e-01 -8.97916317e-01 -1.88135058e-01
-8.00052956e-02 -4.39900905e-01 6.39953732e-01 -3.65463853e-01
9.30829763e-01 -8.40933323e-01 -5.62462389e-01 8.78451541e-02
1.40787506e+00 -3.07792753e-01 3.78817886e-01 6.04554594e-01
8.46693695e-01 8.31652224e-01 4.62237559e-02 9.92835879e-01
2.49531746e-01 4.60938007e-01 3.80883634e-01 1.84242934e-01
5.63525975e-01 -5.23519695e-01 1.97870940e-01 7.66432941e-01
1.66105870e-02 -2.55967289e-01 -1.32105851e+00 7.78315127e-01
-1.90844524e+00 -1.35168207e+00 -2.90142417e-01 1.78800583e+00
9.37292337e-01 3.01571220e-01 8.17348361e-01 3.03295553e-01
7.99924970e-01 4.60625410e-01 2.76295189e-02 -6.75886095e-01
2.21144006e-01 1.53474346e-01 6.59264445e-01 5.80150068e-01
-1.04953051e+00 8.24396372e-01 6.51146030e+00 8.74685466e-01
-9.43686724e-01 5.44599116e-01 5.68453252e-01 -2.12972477e-01
-2.57894367e-01 -1.09471731e-01 -5.66589952e-01 7.38811493e-01
9.93027687e-01 -1.89948440e-01 4.73372012e-01 7.89849043e-01
1.85475543e-01 -2.02934265e-01 -6.43736541e-01 7.93753028e-01
3.01620901e-01 -1.17608929e+00 -3.96120310e-01 5.44888794e-01
5.99365354e-01 3.37238371e-01 3.13012823e-02 1.97609156e-01
4.35427755e-01 -9.86355722e-01 6.42145932e-01 -2.87362002e-02
3.53634477e-01 -6.94339573e-01 5.01237750e-01 6.38072431e-01
-3.65560085e-01 -4.31821346e-01 8.25463012e-02 -4.82862413e-01
3.35889935e-01 6.22001767e-01 -6.51382387e-01 2.15762436e-01
5.73680639e-01 9.93442059e-01 -7.49647915e-01 8.37789357e-01
-3.03798262e-02 1.14607072e+00 -3.90881658e-01 -1.24416582e-01
4.89144534e-01 -1.58254787e-01 7.61450052e-01 1.64312017e+00
-1.22999743e-01 6.58390000e-02 6.17801771e-02 7.73087323e-01
-1.96885780e-01 3.22027504e-01 -1.10873294e+00 -5.89674950e-01
5.76859772e-01 1.35303020e+00 -4.34225589e-01 -2.54186779e-01
-7.06887960e-01 5.98590374e-01 8.11145961e-01 -9.20209847e-03
-5.54714203e-01 -1.71718583e-01 8.66777837e-01 8.49869370e-01
-1.10397890e-01 -4.77247685e-01 -3.07229280e-01 -1.30079436e+00
-1.55907691e-01 -9.56271648e-01 3.45488101e-01 -2.93231636e-01
-1.80536950e+00 1.95979089e-01 1.96407601e-01 -5.57886660e-01
-4.27396119e-01 -4.48334724e-01 -6.83264852e-01 5.33509731e-01
-1.10567570e+00 -9.74118292e-01 4.79932167e-02 4.20513213e-01
6.45924285e-02 9.47256237e-02 6.11462772e-01 6.27168477e-01
-7.87808239e-01 4.89719063e-01 8.07708427e-02 7.74300933e-01
6.12783432e-01 -7.70483196e-01 2.90228099e-01 7.03687727e-01
-1.34029895e-01 1.01856053e+00 9.49413776e-01 -9.55109239e-01
-9.53962266e-01 -7.41661191e-01 1.05627072e+00 -8.01254034e-01
1.69163334e+00 -6.77587271e-01 -9.91243482e-01 8.19375038e-01
3.00193906e-01 -3.08013737e-01 1.30177414e+00 5.66372991e-01
-8.86146247e-01 6.11362159e-01 -1.28744388e+00 5.91887236e-01
1.34317291e+00 -7.57809758e-01 -4.90847290e-01 4.09586996e-01
1.09486923e-01 1.03862360e-01 -7.25413561e-01 -7.90659189e-02
5.92154980e-01 -1.08035839e+00 4.03980315e-01 -9.98194456e-01
6.71713710e-01 4.24340695e-01 -5.85308485e-02 -1.15836871e+00
-3.53471339e-01 -7.82191932e-01 -9.96611044e-02 1.42333126e+00
3.30339968e-01 -7.34862685e-01 6.22668147e-01 7.90868640e-01
1.78561017e-01 -6.96091473e-01 -7.36161828e-01 -5.43866873e-01
4.05765474e-01 -2.97546595e-01 9.75700393e-02 1.66876113e+00
7.09547281e-01 3.47359568e-01 -3.53706777e-01 -1.86564997e-01
5.83810925e-01 -3.90743673e-01 8.38220954e-01 -1.36351740e+00
-1.23059280e-01 -6.64485037e-01 -4.31856066e-01 -4.02651399e-01
2.42858231e-01 -6.87388897e-01 -6.15090847e-01 -9.64324594e-01
7.70744801e-01 -4.62165505e-01 1.40721321e-01 3.66112620e-01
5.50481789e-02 4.72832888e-01 2.64033824e-01 5.72014213e-01
-7.14656413e-01 7.28601888e-02 5.98680317e-01 1.50213897e-01
-1.22755572e-01 -5.06613910e-01 -8.88078213e-01 9.65141535e-01
8.78549397e-01 -8.56029510e-01 7.57049546e-02 -1.21523693e-01
6.75442576e-01 -4.29935694e-01 2.99574733e-01 -7.32964635e-01
-3.03010084e-02 -2.29327500e-01 -1.19027356e-02 -1.08546764e-01
4.03076351e-01 -3.82704675e-01 -1.66551352e-01 3.58892471e-01
-1.58751860e-01 1.22309938e-01 -4.05371934e-03 4.74116653e-01
7.18800630e-03 -5.39958179e-01 8.30938220e-01 -2.25312963e-01
8.65878984e-02 2.16300428e-01 -9.11371827e-01 3.76688540e-01
9.86861408e-01 -7.97122940e-02 -6.18747115e-01 -6.43228114e-01
-4.32983309e-01 -3.11060641e-02 9.14524376e-01 2.90459424e-01
7.32130185e-03 -1.02043295e+00 -7.48902023e-01 -2.22085774e-01
1.59592316e-01 -7.99194992e-01 -2.02268481e-01 8.87947142e-01
-3.63489330e-01 9.63877589e-02 -1.40464932e-01 -1.03613697e-01
-1.15630293e+00 5.22238553e-01 1.09547853e-01 -2.67847806e-01
-1.32135123e-01 4.08229947e-01 -3.98716718e-01 -2.15187773e-01
-3.71284366e-01 4.14892107e-01 -3.59787077e-01 4.68532145e-01
6.50825918e-01 7.41029024e-01 -2.54095703e-01 -1.15143704e+00
-3.74894857e-01 -1.07563576e-02 -4.28208977e-01 -3.50032091e-01
1.36005318e+00 1.44930044e-02 -4.22146410e-01 6.00412667e-01
1.59775388e+00 6.37577653e-01 -2.50789762e-01 -2.09262103e-01
3.56114417e-01 -1.14810050e+00 -1.99786037e-01 -3.17739934e-01
-6.50137424e-01 6.32823408e-01 -2.63755530e-01 9.09640372e-01
3.58601063e-01 1.63286433e-01 9.66887474e-01 -2.00916622e-02
4.66579914e-01 -9.97963488e-01 2.67285436e-01 5.83469987e-01
4.07401323e-01 -1.04905808e+00 1.90597996e-01 -8.62797678e-01
-3.94423068e-01 7.45251179e-01 4.70415980e-01 -3.30051631e-01
9.48081732e-01 2.09438819e-02 -1.12851918e-01 -4.01056141e-01
-3.88535559e-01 8.50863680e-02 -2.92177409e-01 3.35985333e-01
6.50136828e-01 -3.77014950e-02 -7.03417242e-01 5.78652501e-01
-1.48640260e-01 -2.19344690e-01 8.10949922e-01 1.15385628e+00
-5.88670552e-01 -1.16137600e+00 -2.79488802e-01 4.92745459e-01
-1.30313623e+00 -2.99051143e-02 -1.02784240e+00 8.79492879e-01
-3.44664007e-02 1.19212079e+00 -6.33497387e-02 -5.15382707e-01
-1.01573743e-01 1.34860158e-01 2.36449882e-01 -9.21022713e-01
-1.05531490e+00 -2.63409704e-01 5.73063850e-01 -4.74613428e-01
-3.14182580e-01 -1.12021661e+00 -6.96245849e-01 -1.00841677e+00
-4.32465196e-01 -6.82236347e-03 5.02347529e-01 1.01455307e+00
4.66458023e-01 -5.18276989e-01 7.12826312e-01 -5.57354033e-01
-5.04205227e-01 -1.22133768e+00 -5.02602816e-01 9.00725305e-01
2.63465077e-01 -7.40455091e-01 -5.33605516e-01 -4.59978074e-01] | [8.625264167785645, 10.456000328063965] |
a666747a-c4a0-46bd-b33f-8ca4cabbd2c1 | copula-based-conformal-prediction-for-multi | 2101.12002 | null | https://arxiv.org/abs/2101.12002v1 | https://arxiv.org/pdf/2101.12002v1.pdf | Copula-based conformal prediction for Multi-Target Regression | There are relatively few works dealing with conformal prediction for multi-task learning issues, and this is particularly true for multi-target regression. This paper focuses on the problem of providing valid (i.e., frequency calibrated) multi-variate predictions. To do so, we propose to use copula functions applied to deep neural networks for inductive conformal prediction. We show that the proposed method ensures efficiency and validity for multi-target regression problems on various data sets. | ['Sylvain Rousseau', 'Sébastien Destercke', 'Soundouss Messoudi'] | 2021-01-28 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [ 6.28801212e-02 -5.21973856e-02 -2.82693267e-01 -5.31530678e-01
-1.20921183e+00 -2.85608619e-01 3.53467613e-01 9.88920927e-02
-2.12033242e-01 1.02256036e+00 -6.89369291e-02 -1.29718244e-01
-5.82278669e-01 -8.56640875e-01 -8.00946176e-01 -7.15093017e-01
1.07570719e-02 5.75032830e-01 -7.87838101e-02 9.36713591e-02
1.11399628e-01 2.29935169e-01 -1.16339815e+00 2.71175534e-01
7.96633363e-01 9.17141497e-01 -6.22994341e-02 5.76196611e-01
4.09290224e-01 4.73219067e-01 -5.15810430e-01 -8.69509876e-01
-2.53745526e-01 -1.04353234e-01 -4.85015333e-01 -9.56887841e-01
2.42530659e-01 2.39028633e-01 2.07374841e-01 7.00761199e-01
4.80070978e-01 2.21814841e-01 1.42114472e+00 -1.33409941e+00
-9.33847070e-01 5.56965768e-01 -7.21345305e-01 3.93046021e-01
-1.85928732e-01 -7.54943848e-01 1.13049769e+00 -1.01000869e+00
-1.90192834e-01 1.19105673e+00 1.18768346e+00 3.93430442e-01
-1.08922362e+00 -8.59435558e-01 -2.38347784e-01 5.32167368e-02
-1.39088428e+00 -1.65615767e-01 8.19199741e-01 -3.99184525e-01
7.08487749e-01 -1.46494173e-02 1.89767882e-01 1.69292533e+00
1.25674772e+00 5.63076973e-01 1.21975207e+00 -4.20991451e-01
8.60275701e-02 3.24022502e-01 2.06146121e-01 1.51640952e-01
1.17793113e-01 3.17592293e-01 -1.76950857e-01 -4.25483376e-01
5.04983962e-01 8.87254998e-02 9.46888998e-02 -1.23800077e-01
-7.76591063e-01 1.23535943e+00 1.20819226e-01 3.55891496e-01
-2.74891436e-01 3.65770489e-01 3.75931740e-01 1.99647561e-01
1.09447026e+00 9.85526666e-02 -6.27971947e-01 2.39792280e-02
-8.47735465e-01 3.87184143e-01 6.57084227e-01 8.26396644e-01
2.72393078e-01 3.61204408e-02 -3.98433805e-01 1.08418119e+00
1.54381096e-01 7.40222335e-01 3.86522889e-01 -2.25292996e-01
2.82096952e-01 -3.70196223e-01 -1.19080275e-01 -1.12607861e+00
-9.28929865e-01 -6.71470821e-01 -1.12649858e+00 -5.20765223e-02
2.32584298e-01 -5.29148042e-01 -5.17046988e-01 1.84162307e+00
4.48451005e-02 1.71910256e-01 7.15389177e-02 4.83071089e-01
6.34950817e-01 5.98915398e-01 7.65898049e-01 -2.13619709e-01
1.16393209e+00 -8.21075857e-01 -7.28590250e-01 -2.17555523e-01
5.13485789e-01 -7.66799271e-01 8.40529442e-01 2.20229909e-01
-6.64719820e-01 -7.37284303e-01 -8.55689466e-01 1.97007149e-01
-4.86544788e-01 1.77169785e-01 7.38235772e-01 8.15866232e-01
-7.35923946e-01 5.07696748e-01 -5.69350362e-01 -7.78502375e-02
4.82145905e-01 8.25820938e-02 -1.86524183e-01 4.58610840e-02
-1.47254956e+00 1.29918659e+00 3.52211565e-01 -2.49228179e-01
-5.76054931e-01 -1.08605778e+00 -7.78909385e-01 2.97343165e-01
-2.18019143e-01 -5.33246398e-01 1.19457436e+00 -1.07979989e+00
-1.12513745e+00 4.79950041e-01 1.93066988e-02 -3.20928961e-01
3.22617471e-01 -3.19112241e-01 -5.41982889e-01 -3.91496927e-01
-6.74351603e-02 2.36363500e-01 8.97831857e-01 -9.42000449e-01
-6.79945469e-01 -4.40846801e-01 -2.82650352e-01 2.24119164e-02
-6.61305130e-01 1.17651485e-01 1.41338870e-01 -1.03134370e+00
-4.55726534e-01 -9.95454252e-01 -5.42374663e-02 -5.26150823e-01
-8.86951834e-02 -5.77788115e-01 6.28600001e-01 -8.82777095e-01
1.13627005e+00 -2.19729733e+00 1.03989430e-01 1.94717884e-01
-3.03801913e-02 -1.85520589e-01 -1.61453053e-01 2.95079678e-01
-3.80023509e-01 -4.23523970e-02 -9.29303393e-02 -6.19668901e-01
-8.61974284e-02 -2.48575564e-02 -4.41195786e-01 4.83241767e-01
2.32278883e-01 1.17935026e+00 -3.64277035e-01 -5.33089638e-01
-1.14420980e-01 7.16008246e-01 -2.24346116e-01 -1.61874201e-02
-2.92537361e-02 3.02264005e-01 -4.60930586e-01 6.06111288e-01
7.77239025e-01 -2.88768500e-01 -8.68929848e-02 -9.53592584e-02
2.00503662e-01 -2.76955336e-01 -5.70217371e-01 1.21478283e+00
-8.96320224e-01 5.91928601e-01 -7.17689455e-01 -1.31114399e+00
1.09152150e+00 4.32313770e-01 6.70399189e-01 -5.85969925e-01
1.48056180e-03 1.93197697e-01 -2.16415510e-01 3.21929231e-02
3.29398811e-01 -6.31911337e-01 -4.26428229e-01 3.60516906e-01
2.08949074e-01 1.66466713e-01 -4.40918684e-01 -5.29617429e-01
6.10306263e-01 2.35685244e-01 6.08083367e-01 -6.75356746e-01
2.12547660e-01 -2.40320593e-01 3.98233920e-01 5.56382596e-01
3.99118140e-02 6.19892716e-01 2.88747936e-01 -4.99544889e-01
-1.23695135e+00 -1.08593464e+00 -8.43080163e-01 1.32318151e+00
-3.09378564e-01 -1.43380594e-02 -5.44223428e-01 -4.49380487e-01
2.34789401e-01 9.97617841e-01 -1.04852116e+00 -2.80617356e-01
-3.27219546e-01 -1.12544250e+00 4.11563635e-01 9.50734854e-01
-4.79886644e-02 -9.29334164e-01 -3.78259927e-01 3.87248337e-01
-1.28360301e-01 -9.03154731e-01 -3.19470197e-01 4.85841721e-01
-5.84992528e-01 -9.99882281e-01 -1.02572203e+00 -8.05079162e-01
1.32007346e-01 -1.74472407e-01 1.16080630e+00 -5.57298422e-01
9.43589881e-02 4.64454174e-01 -2.11142212e-01 -9.12633479e-01
-2.80157745e-01 3.22262853e-01 1.81408450e-02 -1.25544826e-02
6.08736932e-01 -6.51730657e-01 -1.37686998e-01 4.60545897e-01
-6.97455466e-01 -4.96650599e-02 5.08390069e-01 1.01737940e+00
8.10533226e-01 4.05750237e-02 1.22524261e+00 -8.69201124e-01
1.15390253e+00 -7.57596314e-01 -6.02688372e-01 7.03427911e-01
-7.22527444e-01 -1.56153455e-01 8.79695058e-01 -7.52837837e-01
-1.08422029e+00 -2.18570545e-01 -9.96674374e-02 -5.23583829e-01
-5.45462370e-02 7.65654743e-01 3.55524272e-01 -1.99141487e-01
7.07790792e-01 -3.64852510e-02 -3.12214047e-01 -3.29712868e-01
5.29839471e-02 6.61267161e-01 2.29395777e-01 -4.99056250e-01
5.07862985e-01 -3.67315374e-02 4.80242044e-01 -2.98751682e-01
-1.08314788e+00 1.47685185e-01 -6.66646719e-01 -2.94617504e-01
9.41974819e-01 -9.45636928e-01 -6.80392325e-01 3.68118614e-01
-1.04839551e+00 -3.32829773e-01 1.61922798e-01 7.12743104e-01
-8.99354458e-01 -1.25304297e-01 -3.82517189e-01 -8.80080223e-01
-7.22277999e-01 -6.61858499e-01 8.67130160e-01 1.60816923e-01
-8.24978575e-02 -1.53633618e+00 5.27044177e-01 -7.02491328e-02
7.25308776e-01 3.44254464e-01 1.14343119e+00 -9.28072453e-01
1.36447713e-01 -2.79527038e-01 -8.88823867e-02 2.04890400e-01
-2.37879828e-02 -4.51044030e-02 -1.17148983e+00 -2.81086147e-01
-7.03244098e-03 -3.58221442e-01 7.99542248e-01 1.02820003e+00
1.48235250e+00 -9.03173611e-02 -4.03767318e-01 6.08303607e-01
1.43580317e+00 1.51499704e-01 4.59291071e-01 8.51509124e-02
4.92496818e-01 5.35104394e-01 8.31671536e-01 4.99539196e-01
6.36183918e-01 7.99124479e-01 -1.45855416e-02 -9.70044807e-02
2.35739514e-01 -2.32000649e-01 1.77684706e-03 5.69142163e-01
2.25982890e-02 -3.58192205e-01 -1.11434853e+00 4.33833301e-01
-1.96409881e+00 -9.03834462e-01 -1.52890727e-01 2.02095103e+00
6.72757864e-01 -7.22348616e-02 3.91138485e-03 -3.22028279e-01
6.52397990e-01 -1.00195706e-01 -5.19108772e-01 -4.83332992e-01
-2.69217193e-01 5.53035617e-01 4.55877483e-01 1.47901282e-01
-1.57528400e+00 6.14847600e-01 6.98443365e+00 1.13317192e+00
-9.21165943e-01 5.20023525e-01 1.03771102e+00 7.45820478e-02
-2.99200892e-01 -3.32078338e-01 -8.81155074e-01 4.07099217e-01
1.30125511e+00 -4.52027261e-01 1.38339370e-01 9.74758089e-01
-3.02100450e-01 4.63218689e-02 -1.08583248e+00 9.45062995e-01
1.76793084e-01 -1.01205540e+00 -4.96320277e-01 -1.54528826e-01
7.65911043e-01 -5.95084876e-02 4.89351064e-01 6.30964816e-01
1.33057669e-01 -1.42035365e+00 5.02557516e-01 7.52765298e-01
7.30408430e-01 -1.24446726e+00 8.84400368e-01 3.36431056e-01
-1.08752036e+00 -2.01289147e-01 -7.23713338e-01 3.62490296e-01
8.54029432e-02 8.02415371e-01 -3.11312675e-01 7.57076323e-01
7.77478993e-01 7.58538187e-01 -5.00744581e-01 1.04717863e+00
2.55765051e-01 5.04261196e-01 -1.61550447e-01 -3.52522284e-01
-1.55254066e-01 -4.81527038e-02 -2.43076570e-02 1.07911026e+00
8.04796636e-01 -2.58774489e-01 -1.52067170e-01 7.26775169e-01
1.76760107e-01 4.75856572e-01 -6.74164712e-01 3.95278484e-01
1.31508693e-01 1.14571273e+00 -2.75104523e-01 1.73872650e-01
-7.03564346e-01 6.60191953e-01 6.29972994e-01 2.62591362e-01
-1.24526453e+00 -1.59165710e-01 3.07452023e-01 -3.75649452e-01
2.38248691e-01 1.76884569e-02 -6.79948032e-01 -8.41693699e-01
-1.65565223e-01 -4.84788179e-01 7.17969656e-01 -7.76094913e-01
-2.09490204e+00 4.98666704e-01 3.41467440e-01 -1.03520787e+00
-2.76843786e-01 -6.91376209e-01 -7.07995594e-01 1.12459576e+00
-1.62543738e+00 -1.58539271e+00 1.37041742e-02 6.90104246e-01
1.61286160e-01 -3.90384942e-01 1.15537298e+00 3.35011572e-01
-2.05194637e-01 1.00594115e+00 4.61430550e-01 -1.72982305e-01
9.96735752e-01 -1.06883562e+00 3.88843298e-01 1.86378285e-01
9.85708535e-02 1.71707720e-01 5.10174096e-01 -6.02514386e-01
-9.78242040e-01 -1.18494654e+00 9.05659258e-01 -7.08726287e-01
6.99572027e-01 -2.69440025e-01 -8.87534440e-01 7.74913907e-01
7.25670382e-02 -2.04867229e-01 1.18717647e+00 6.12320304e-01
-2.62156844e-01 -2.08456114e-01 -1.01508367e+00 1.77484214e-01
3.05620253e-01 -3.63673955e-01 -3.53688598e-01 6.10909104e-01
6.19240284e-01 -3.69233787e-01 -1.38670707e+00 6.57015979e-01
7.65515447e-01 -4.26552385e-01 1.02340472e+00 -6.82719529e-01
7.72036254e-01 5.16255498e-01 -3.46044362e-01 -1.71899188e+00
-5.92199743e-01 1.31441072e-01 -1.00204922e-01 9.38894033e-01
6.93129063e-01 -7.37790346e-01 2.56318092e-01 6.06932342e-01
-1.16242722e-01 -9.69916880e-01 -1.17595637e+00 -6.76723897e-01
9.30137694e-01 -4.52362537e-01 4.71077442e-01 9.85843539e-01
-3.73014659e-02 2.39284992e-01 -9.25051153e-01 9.17015374e-02
7.55478561e-01 2.89681941e-01 4.44909602e-01 -1.53349960e+00
-3.04956019e-01 -2.60133952e-01 -1.42973596e-02 -3.47507358e-01
5.99239409e-01 -7.07663536e-01 -9.61965472e-02 -1.11165404e+00
5.20767033e-01 -6.03901684e-01 -5.97307682e-01 3.69211465e-01
-2.99036771e-01 1.02011934e-01 -9.29815173e-02 -6.19079545e-03
-1.44911692e-01 8.08246017e-01 1.01277101e+00 7.85633922e-02
2.12418750e-01 7.40870893e-01 -7.02279687e-01 2.89804280e-01
1.32189870e+00 -6.65983260e-01 -4.01225835e-01 -1.65159151e-01
2.49165744e-01 2.45670319e-01 3.67323220e-01 -1.03327215e+00
1.11852504e-01 -3.03441793e-01 7.55973637e-01 -5.16282439e-01
3.95682901e-01 -8.84823501e-01 3.72680664e-01 1.51832163e-01
-3.92481863e-01 4.44061667e-01 5.38297892e-01 8.15714777e-01
-1.04003586e-01 -2.87871361e-01 1.09507906e+00 3.47702473e-01
-1.17812768e-01 4.66107070e-01 3.66007313e-02 2.39985418e-02
1.25324965e+00 3.44556153e-01 -3.68899912e-01 -3.87547076e-01
-6.50251150e-01 7.32926801e-02 -1.96473494e-01 7.59918392e-01
5.56513786e-01 -1.67034018e+00 -8.83188963e-01 6.34330213e-02
3.13943595e-01 -5.67476928e-01 4.28671986e-01 7.88118601e-01
1.63600013e-01 6.51370227e-01 -2.78717846e-01 -4.03070122e-01
-1.07275534e+00 6.09662533e-01 5.47369599e-01 -4.54489172e-01
-4.15610909e-01 6.55108750e-01 2.86164612e-01 -4.37206537e-01
4.90950793e-03 -3.00427433e-02 -5.61399043e-01 1.27399251e-01
3.86253685e-01 1.11518711e-01 -2.64790468e-02 -3.66934657e-01
-2.84069836e-01 7.74284124e-01 -1.32441372e-02 6.25615343e-02
1.56114566e+00 2.03251809e-01 1.67231839e-02 7.16212213e-01
1.32972550e+00 -2.51030326e-01 -9.86174226e-01 -2.45819882e-01
-5.21152373e-03 -1.36836708e-01 1.61427602e-01 -8.18993151e-01
-9.14519310e-01 9.95179832e-01 7.97078013e-01 7.41738603e-02
1.08303797e+00 1.09730624e-02 3.05763006e-01 2.09744781e-01
4.83888209e-01 -1.13966954e+00 -4.69380431e-02 4.64734286e-01
1.04702652e+00 -1.37817168e+00 -3.01948041e-02 5.10839857e-02
-7.84857512e-01 1.23048627e+00 6.65099204e-01 -2.79411860e-02
1.32205868e+00 1.65813491e-01 -2.58214831e-01 -7.48984143e-03
-8.07826459e-01 3.87928396e-01 7.55473912e-01 8.45255017e-01
6.03998363e-01 3.55354220e-01 -3.44201177e-01 1.06122899e+00
-1.11180440e-01 -1.97180733e-01 1.65837139e-01 3.37313920e-01
-2.02157751e-01 -9.98858571e-01 -3.46335471e-01 7.05959618e-01
-6.26850545e-01 -2.44361639e-01 1.20738991e-01 7.25584269e-01
-3.83793950e-01 8.07020307e-01 1.85836971e-01 -3.11340481e-01
1.46886244e-01 1.30721003e-01 4.21734691e-01 -2.20999226e-01
-5.03565192e-01 -7.13415667e-02 -6.52995184e-02 -4.97560017e-02
-4.47084814e-01 -7.05505073e-01 -5.70895731e-01 -3.63121599e-01
-2.20987245e-01 -2.16092080e-01 7.63831556e-01 7.67883599e-01
1.11292407e-01 6.97288573e-01 8.10927570e-01 -5.11813939e-01
-6.97618842e-01 -1.29316413e+00 -6.25826955e-01 9.10781473e-02
1.41400963e-01 -9.68967021e-01 -2.64213920e-01 -3.45450163e-01] | [7.8377604484558105, 4.040994644165039] |
17659bf9-ab7f-4b71-9fed-ca4fa3279707 | to-compute-or-not-to-compute-adaptive-smart | 2209.02166 | null | https://arxiv.org/abs/2209.02166v2 | https://arxiv.org/pdf/2209.02166v2.pdf | To Compute or not to Compute? Adaptive Smart Sensing in Resource-Constrained Edge Computing | We consider a network of smart sensors for edge computing application that sample a signal of interest and send updates to a base station for remote global monitoring. Sensors are equipped with sensing and compute, and can either send raw data or process them on-board before transmission. Limited hardware resources at the edge generate a fundamental latency-accuracy trade-off: raw measurements are inaccurate but timely, whereas accurate processed updates are available after computational delay. Also, if sensor on-board processing entails data compression, latency caused by wireless communication might be higher for raw measurements. Hence, one needs to decide when sensors should transmit raw measurements or rely on local processing to maximize overall network performance. To tackle this sensing design problem, we model an estimation-theoretic optimization framework that embeds computation and communication delays, and propose a Reinforcement Learning-based approach to dynamically allocate computational resources at each sensor. Effectiveness of our proposed approach is validated through numerical simulations with case studies motivated by the Internet of Drones and self-driving vehicles. | ['Paolo Dini', 'Francesco Zanini', 'Giovanni Peserico', 'Luca Ballotta'] | 2022-09-05 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 6.79667592e-01 2.74286360e-01 -3.68652642e-01 -3.05528283e-01
-7.01905489e-01 -2.29571730e-01 1.06410064e-01 3.66004348e-01
-5.90805531e-01 7.16910958e-01 -3.64575356e-01 -1.49128333e-01
-2.77489066e-01 -1.24755800e+00 -6.30741835e-01 -6.45680070e-01
-5.41103482e-01 1.84054911e-01 1.91715732e-01 7.36035034e-02
5.43304235e-02 1.60603404e-01 -1.30597413e+00 -4.07831699e-01
5.62787831e-01 1.96045864e+00 2.12876409e-01 1.02725899e+00
3.47238421e-01 3.73781651e-01 -6.80097938e-01 2.26669535e-01
4.30899590e-01 -1.39418259e-01 7.86822811e-02 2.01319993e-01
-2.76739806e-01 -6.04545772e-01 -1.75029393e-02 1.19689465e+00
5.27243555e-01 -1.79241061e-01 2.13696346e-01 -1.65163255e+00
2.69349605e-01 6.87135994e-01 -6.01215124e-01 3.23328793e-01
2.43639931e-01 1.29732907e-01 5.59777021e-01 -3.17078441e-01
2.85580903e-01 4.48529065e-01 6.89491868e-01 2.74373591e-01
-8.65038633e-01 -7.43478954e-01 1.40572533e-01 1.91915914e-01
-1.15562558e+00 -8.52966666e-01 9.59155321e-01 -3.03163156e-02
6.63397670e-01 2.55656302e-01 1.03224075e+00 3.58923733e-01
5.57220697e-01 3.50400329e-01 7.64189601e-01 -8.64202902e-03
9.07691896e-01 -1.71225592e-01 -2.57313162e-01 2.81730950e-01
7.92012155e-01 3.81568313e-01 -6.56723022e-01 -4.95795384e-02
5.43246031e-01 3.01816225e-01 -2.83291489e-01 -6.77918969e-03
-1.13348329e+00 3.40143025e-01 2.85547435e-01 8.06941278e-03
-1.17330408e+00 7.07203686e-01 3.00664335e-01 7.20109224e-01
3.56418163e-01 -1.86261714e-01 -4.57513899e-01 -1.58562243e-01
-1.14211369e+00 -1.76936850e-01 6.99594676e-01 1.22694135e+00
7.33683527e-01 1.38014257e-01 2.10346833e-01 -9.65026468e-02
4.23894972e-01 9.77952361e-01 2.92761981e-01 -9.19682443e-01
5.39419711e-01 1.85306281e-01 4.10191774e-01 -1.08201587e+00
-7.96409011e-01 -3.08833241e-01 -1.32426298e+00 2.84337133e-01
5.27174473e-02 -9.32228684e-01 -3.71400684e-01 1.38826048e+00
5.71711302e-01 6.96045816e-01 3.09867471e-01 1.16332185e+00
-3.13218720e-02 7.52153158e-01 -1.34962007e-01 -7.21049190e-01
1.18547332e+00 -3.67812663e-01 -9.33816433e-01 -3.10642362e-01
3.94556761e-01 -5.01155674e-01 2.43908390e-01 5.23345709e-01
-1.38989437e+00 -4.61756200e-01 -1.58110929e+00 5.87720811e-01
-2.75087774e-01 1.38430342e-01 2.25257963e-01 7.68902302e-01
-1.15813875e+00 4.89331543e-01 -1.10987699e+00 -1.07419968e-01
3.52118025e-03 5.78352928e-01 3.02915543e-01 7.02705607e-02
-8.87188792e-01 1.71528772e-01 3.98539096e-01 3.80767375e-01
-8.45723212e-01 -8.49299252e-01 -4.88309681e-01 -6.68834150e-03
5.88213146e-01 -8.28273177e-01 1.22018111e+00 -9.54280853e-01
-1.75950384e+00 -1.01719767e-01 -4.97022131e-03 -9.46966767e-01
4.91500080e-01 1.04334734e-01 -7.35588074e-01 1.97503835e-01
-4.06435519e-01 2.05325797e-01 1.14717138e+00 -8.01218867e-01
-1.01590312e+00 -3.91124427e-01 4.56782691e-02 -6.13412820e-02
-3.24594766e-01 -5.63537061e-01 1.01129115e-01 -2.07119763e-01
2.94633687e-01 -7.79653847e-01 -6.01265609e-01 1.13449000e-01
-3.51265296e-02 1.91089302e-01 1.03312826e+00 -2.20979989e-01
1.17415226e+00 -1.85736668e+00 -3.90460730e-01 6.81618512e-01
3.19279701e-01 -8.59733671e-02 8.31983984e-02 3.35595101e-01
4.94603455e-01 -1.81933746e-01 4.30810591e-03 -3.77087772e-01
-2.58510023e-01 1.73958376e-01 -1.24605864e-01 6.24707639e-01
6.60814159e-03 3.99264932e-01 -9.57115650e-01 -2.99892783e-01
2.73184121e-01 3.61260921e-01 -3.96540880e-01 1.27514616e-01
-2.29702413e-01 2.73131162e-01 -8.57930541e-01 5.94606817e-01
7.41541922e-01 -3.20336938e-01 3.16559941e-01 -2.78241187e-01
-2.01060593e-01 2.53359467e-01 -1.68304718e+00 1.40741825e+00
-8.52081358e-01 4.55594480e-01 1.02426565e+00 -1.13564456e+00
8.24284375e-01 4.29315269e-01 8.70389163e-01 -7.04882979e-01
4.55490232e-01 1.66554600e-01 -4.50233042e-01 -2.88731813e-01
6.42381072e-01 2.12791398e-01 -8.82413760e-02 5.85490286e-01
-8.13575745e-01 -8.26458335e-02 -4.24664915e-02 1.60390120e-02
1.67693210e+00 -4.00226593e-01 8.14101249e-02 -2.14154929e-01
4.51098084e-01 -1.14420056e-02 8.26023936e-01 5.22167981e-01
-2.69738108e-01 -3.56405169e-01 -2.73990314e-02 -2.65231609e-01
-7.70942688e-01 -7.43391216e-01 4.40956682e-01 6.35572612e-01
5.95429778e-01 -3.43119621e-01 -5.58319867e-01 3.66097055e-02
1.08798151e-03 2.37263009e-01 -1.77992284e-01 -2.15062976e-01
-1.47989169e-01 -3.74393344e-01 -5.19336499e-02 2.90189147e-01
6.45022571e-01 -4.43513364e-01 -1.74138165e+00 8.63308847e-01
1.45520568e-01 -1.24253345e+00 -3.92204672e-01 1.19317167e-01
-1.01915920e+00 -1.09432018e+00 -9.69780162e-02 -2.48137295e-01
9.06237245e-01 6.63685918e-01 9.40682352e-01 1.06553070e-01
-8.61001015e-02 8.08026552e-01 -3.33782881e-01 -6.85959041e-01
-6.55896589e-02 -1.37341708e-01 1.40809506e-01 3.34378809e-01
8.98018554e-02 -6.01151466e-01 -9.69970882e-01 9.04484391e-02
-1.08414650e+00 -1.14816152e-01 6.46797657e-01 4.54384089e-01
9.03691232e-01 7.49529123e-01 7.61806965e-01 -6.52150750e-01
7.52058923e-01 -5.61484158e-01 -1.47009897e+00 3.73343043e-02
-6.10532939e-01 -1.98021717e-02 7.35479414e-01 -1.74223468e-01
-6.01837456e-01 4.90335912e-01 3.46776426e-01 -2.54646689e-01
3.48556757e-01 4.35161442e-01 -1.79540142e-01 -3.48486394e-01
8.21551010e-02 -1.10997051e-01 -2.16195390e-01 2.14976043e-01
-2.24175639e-02 7.89864182e-01 4.61180121e-01 -2.28614315e-01
6.60139680e-01 6.47558749e-01 3.87766808e-01 -1.01976061e+00
-1.10365152e-02 -3.18749189e-01 3.36160101e-02 -6.22904658e-01
4.02011722e-01 -1.13555753e+00 -1.26836872e+00 1.41953722e-01
-9.54849064e-01 -3.29941720e-01 -1.50656238e-01 6.17088139e-01
-5.20267427e-01 9.75521840e-03 -1.68010369e-01 -1.09968281e+00
-5.51517427e-01 -9.41029072e-01 1.02356875e+00 3.21659476e-01
-2.35241745e-02 -8.55184495e-01 -2.92917192e-01 -1.96411252e-01
8.42285931e-01 5.31633019e-01 4.61581796e-02 1.95908565e-02
-1.16700864e+00 -5.47243714e-01 -1.63458407e-01 -1.08051240e-01
4.69136350e-02 -1.01269744e-01 -5.43300629e-01 -5.78218997e-01
-5.11253849e-02 1.13010824e-01 4.24165964e-01 6.38356268e-01
1.17587161e+00 -4.96498227e-01 -5.57967126e-01 5.13462782e-01
1.84572399e+00 1.02297314e-01 1.50859922e-01 -2.51103401e-01
2.87324730e-02 -4.02936265e-02 8.80133808e-01 1.23317683e+00
5.33506155e-01 2.64017612e-01 9.51308846e-01 3.21062893e-01
4.50251460e-01 -8.76108930e-02 5.29720783e-01 7.16573179e-01
1.66600436e-01 -5.90926886e-01 -2.80949056e-01 4.62714434e-01
-1.74063385e+00 -6.19412780e-01 1.43724144e-01 2.55415773e+00
3.02914679e-01 5.01099646e-01 1.45102918e-01 5.68407953e-01
5.93004405e-01 1.44365475e-01 -8.72928798e-01 -3.82464528e-02
3.71721715e-01 -5.07272705e-02 1.38284802e+00 4.55953598e-01
-6.03212953e-01 1.45252898e-01 5.66018200e+00 3.88539344e-01
-1.35278940e+00 1.84012428e-01 4.55789715e-01 -2.58585960e-01
-2.36987606e-01 -2.05179021e-01 -6.22581422e-01 7.48625040e-01
1.51000822e+00 -3.73366386e-01 5.70253074e-01 7.53494740e-01
8.44463825e-01 -7.08419621e-01 -1.14762819e+00 1.22932100e+00
-4.24433202e-01 -1.43138409e+00 -8.27969372e-01 1.43220782e-01
7.46204734e-01 1.76526293e-01 -2.10907981e-01 -3.25492322e-01
3.55950385e-01 -3.53121102e-01 4.79036450e-01 6.14214420e-01
6.28894687e-01 -7.45542586e-01 6.21921480e-01 5.02520919e-01
-1.61642182e+00 -1.71949923e-01 -6.12327039e-01 -4.57554370e-01
4.02346045e-01 1.44332874e+00 -7.01568127e-01 3.92859042e-01
4.57182854e-01 6.16975427e-01 2.39280894e-01 9.71254170e-01
2.22000983e-02 5.38619697e-01 -1.01500773e+00 -5.80818653e-01
-1.51293889e-01 -2.77426839e-01 6.19389355e-01 6.36136770e-01
7.80961931e-01 5.23820817e-01 5.11462212e-01 4.32540298e-01
-1.64891422e-01 -3.95181656e-01 -6.49659872e-01 9.98525694e-02
1.11704528e+00 1.43912685e+00 -8.14673781e-01 -5.69022536e-01
-4.90810692e-01 6.06698751e-01 -2.93134511e-01 3.44575346e-01
-6.97934628e-01 -5.79248369e-01 8.42224181e-01 2.73134649e-01
1.87908784e-01 -5.23937225e-01 -4.31271255e-01 -6.05928779e-01
1.75711542e-01 -2.16101557e-01 1.54384196e-01 -5.21636426e-01
-7.08208382e-01 3.05228889e-01 -4.36096877e-01 -1.44447768e+00
-1.62911013e-01 -9.45985392e-02 -5.75718522e-01 1.21018335e-01
-1.83270800e+00 -3.17823499e-01 -6.57731175e-01 7.40940392e-01
3.03808630e-01 2.95940906e-01 2.59632289e-01 5.91390848e-01
-3.80781174e-01 3.36349219e-01 1.45568356e-01 -2.57914424e-01
1.38000205e-01 -6.08595073e-01 -1.08563518e-02 9.80502009e-01
-3.73276353e-01 -1.83087781e-01 8.60378444e-01 -7.02576280e-01
-2.39347959e+00 -1.18914700e+00 3.99544269e-01 3.87498438e-01
6.79795980e-01 -2.20350120e-02 -1.26557767e-01 4.09369111e-01
2.22344100e-01 4.59497303e-01 3.57351065e-01 -7.17279077e-01
5.36861479e-01 -7.91523993e-01 -1.53423762e+00 2.17588991e-01
8.67197752e-01 -1.05117425e-01 2.18923569e-01 1.99485108e-01
3.96151721e-01 -2.66624957e-01 -1.04062486e+00 1.81544796e-01
3.05161089e-01 -6.39371395e-01 4.54567105e-01 1.87819600e-01
-2.78699636e-01 -4.94433761e-01 -1.51565313e-01 -1.49791193e+00
2.41025358e-01 -1.12385952e+00 -1.31591901e-01 9.12167132e-01
2.22131655e-01 -8.64965796e-01 1.14112425e+00 6.87068045e-01
-2.23818347e-02 -2.20025942e-01 -1.24348843e+00 -5.47014952e-01
-9.32218015e-01 -8.82263958e-01 9.75583971e-01 3.29090983e-01
2.61952549e-01 9.45932642e-02 -1.85216621e-01 8.95198643e-01
1.05807638e+00 -1.92714427e-02 8.31776261e-01 -1.28026676e+00
-1.35661557e-01 4.79280390e-02 -5.09344816e-01 -1.47595394e+00
-4.02187645e-01 -1.80580631e-01 3.47394556e-01 -1.16763425e+00
-8.03865373e-01 -8.21769178e-01 -1.91782832e-01 5.42146675e-02
4.20372605e-01 8.90283957e-02 5.97743429e-02 -1.65831372e-01
-8.86195362e-01 3.16221088e-01 8.30292821e-01 -7.16983378e-02
-2.51614451e-01 3.94125760e-01 -6.23927638e-02 5.36107719e-01
9.77761388e-01 -3.36686790e-01 -7.47071385e-01 -4.35617834e-01
7.50242829e-01 8.71961594e-01 2.27261186e-01 -1.67321157e+00
8.47006977e-01 -2.31472239e-01 2.27875710e-01 -7.43395686e-01
5.00369430e-01 -1.95381594e+00 3.57907951e-01 8.80883455e-01
-3.34006436e-02 3.90492201e-01 -1.92798153e-01 1.04685891e+00
-2.96604019e-02 2.94798940e-01 5.25791109e-01 3.23142320e-01
-6.18256390e-01 5.43132365e-01 -7.41218567e-01 -2.31976241e-01
1.51667452e+00 -1.82404593e-01 -4.15544547e-02 -7.34137595e-01
-5.46689212e-01 6.67156696e-01 7.84632191e-02 2.32790951e-02
7.24393487e-01 -1.07417572e+00 -3.92889470e-01 5.86327612e-01
-1.93777964e-01 -2.30776556e-02 1.79584861e-01 9.63011503e-01
-3.25083554e-01 1.96115032e-01 2.34704167e-02 -5.81504107e-01
-1.02517653e+00 4.10697699e-01 1.77478790e-01 -7.36693591e-02
-1.19502567e-01 5.05738318e-01 -8.38138223e-01 3.31907719e-01
3.37184519e-01 -1.02341425e+00 3.54486376e-01 -1.30780786e-01
6.86581552e-01 7.87203014e-01 1.93088412e-01 1.47799049e-02
-1.00100659e-01 5.94459236e-01 5.83087623e-01 -1.18754126e-01
1.10084271e+00 -8.34084034e-01 2.13554740e-01 1.13048792e-01
8.78391325e-01 -2.18094051e-01 -1.39452851e+00 -4.51565057e-01
-5.75540066e-02 -3.56141269e-01 7.51111448e-01 -1.86006635e-01
-1.53111553e+00 3.71982813e-01 7.28692174e-01 8.04524899e-01
1.79749262e+00 -4.95911330e-01 1.13395417e+00 4.09105211e-01
8.45548153e-01 -1.57252324e+00 -2.54324794e-01 -5.51695265e-02
2.51539461e-02 -1.09501183e+00 -6.90145604e-03 -5.36643922e-01
5.59923649e-02 1.00470197e+00 2.49816746e-01 -2.47764573e-01
1.17581367e+00 8.50008011e-01 -1.72210842e-01 5.11361547e-02
-1.09307086e+00 -1.46423757e-01 -6.06169462e-01 6.04979575e-01
-2.13729098e-01 3.82532775e-01 -3.60710740e-01 3.83356601e-01
-4.12336737e-02 5.24644911e-01 7.34451294e-01 1.09007525e+00
-8.31142962e-01 -8.65326703e-01 -5.62667012e-01 5.18206954e-01
-3.07815611e-01 3.69224101e-01 3.06593210e-01 2.73251414e-01
-3.06096189e-02 1.39985204e+00 5.09101987e-01 -5.32164335e-01
6.32411316e-02 -3.93538356e-01 2.96346098e-02 4.26553115e-02
-6.11712299e-02 9.02370065e-02 1.03451259e-01 -8.15133750e-01
-5.98044991e-01 -4.55935121e-01 -1.48240840e+00 -6.17166460e-01
-2.00228289e-01 3.24887991e-01 1.30500615e+00 8.44235003e-01
7.59413481e-01 8.32542419e-01 1.02653182e+00 -8.82474840e-01
-5.90532124e-01 -4.45561230e-01 -6.45867825e-01 -4.61934268e-01
8.02564025e-01 -2.30229840e-01 -2.95513272e-01 -5.98106347e-02] | [5.9357686042785645, 1.5816491842269897] |
3eef10a4-dd42-4c70-8426-3988b45619d1 | stochastic-shield-a-probabilistic-approach | 2105.06512 | null | https://arxiv.org/abs/2105.06512v1 | https://arxiv.org/pdf/2105.06512v1.pdf | Stochastic-Shield: A Probabilistic Approach Towards Training-Free Adversarial Defense in Quantized CNNs | Quantized neural networks (NN) are the common standard to efficiently deploy deep learning models on tiny hardware platforms. However, we notice that quantized NNs are as vulnerable to adversarial attacks as the full-precision models. With the proliferation of neural networks on small devices that we carry or surround us, there is a need for efficient models without sacrificing trust in the prediction in presence of malign perturbations. Current mitigation approaches often need adversarial training or are bypassed when the strength of adversarial examples is increased. In this work, we investigate how a probabilistic framework would assist in overcoming the aforementioned limitations for quantized deep learning models. We explore Stochastic-Shield: a flexible defense mechanism that leverages input filtering and a probabilistic deep learning approach materialized via Monte Carlo Dropout. We show that it is possible to jointly achieve efficiency and robustness by accurately enabling each module without the burden of re-retraining or ad hoc fine-tuning. | ['Partha Maji', 'René de Jong', 'Sangwon Ha', 'Lorena Qendro'] | 2021-05-13 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 1.61715552e-01 1.97099805e-01 -9.65751559e-02 -2.64292717e-01
-8.06867540e-01 -8.92410338e-01 5.60535967e-01 -1.39050409e-01
-5.85880756e-01 6.67599976e-01 -6.20916337e-02 -7.42169797e-01
-7.91739381e-04 -8.29282701e-01 -1.21058953e+00 -5.96612275e-01
-8.75594616e-02 -3.65830660e-02 2.84751326e-01 -2.11484924e-01
-1.00924812e-01 6.22153223e-01 -1.13451195e+00 8.51206556e-02
5.02173364e-01 1.09451246e+00 -5.01227617e-01 9.09159005e-01
2.77034134e-01 7.86357105e-01 -9.58535135e-01 -8.19523573e-01
6.25856042e-01 1.64457142e-01 -2.40516812e-01 -5.68164110e-01
5.65926552e-01 -7.43643224e-01 -7.97096610e-01 1.35110176e+00
6.14463091e-01 -4.32038009e-01 2.31010482e-01 -1.28531373e+00
-5.73018193e-01 1.10974264e+00 -1.20313011e-01 5.84717467e-02
-3.57120216e-01 7.40414202e-01 5.77966332e-01 -1.96171358e-01
2.23375097e-01 1.20055819e+00 9.53734219e-01 9.47569132e-01
-1.21545494e+00 -9.17547405e-01 4.23318744e-02 -4.12464350e-01
-1.17310524e+00 -7.14050949e-01 6.56819642e-01 -9.07202661e-02
1.00213540e+00 1.52990803e-01 2.59546518e-01 1.75239944e+00
6.02779150e-01 6.45095348e-01 7.13509321e-01 -2.11632058e-01
6.72232628e-01 1.45411193e-01 1.99918106e-01 5.61335206e-01
5.89406848e-01 4.88741487e-01 -3.38715613e-01 -6.10345721e-01
5.86075306e-01 2.10344821e-01 -5.54181375e-02 -1.06197141e-01
-6.63372636e-01 7.65065908e-01 5.50927579e-01 -9.23469570e-03
-1.47264600e-01 1.15284038e+00 7.52670228e-01 3.04550588e-01
1.91236183e-01 4.96520966e-01 -8.05145085e-01 -1.55639708e-01
-9.43098962e-01 2.19409823e-01 8.71994376e-01 7.37266481e-01
4.48061794e-01 4.68238205e-01 4.86326069e-02 -2.52975188e-02
2.22575933e-01 4.31482941e-01 2.65552670e-01 -9.38235044e-01
2.81705827e-01 1.31228656e-01 -8.92808139e-02 -5.54203987e-01
-1.40238643e-01 -7.59583592e-01 -8.30810726e-01 6.06669903e-01
3.28903556e-01 -5.86942434e-01 -1.05351400e+00 1.99362671e+00
1.39508158e-01 3.82246941e-01 2.11634383e-01 4.61741567e-01
5.85571937e-02 1.88303143e-01 2.01032132e-01 4.44385171e-01
1.04696381e+00 -5.24125457e-01 -4.37715769e-01 -3.13575119e-01
4.83676881e-01 -2.77254820e-01 1.02766824e+00 5.18667877e-01
-8.07622254e-01 -9.90706235e-02 -1.68786716e+00 -4.19600531e-02
-4.23354089e-01 -3.02178442e-01 8.58222544e-01 1.60646069e+00
-1.11991894e+00 8.89434636e-01 -1.48859894e+00 2.52415717e-01
8.47546458e-01 8.35050344e-01 -1.42743602e-01 2.05850869e-01
-1.27267122e+00 6.66912436e-01 1.74961746e-01 -2.09260993e-02
-1.30905056e+00 -8.79060328e-01 -6.47025347e-01 2.26864025e-01
-1.02731161e-01 -8.27691197e-01 1.19323003e+00 -8.65640461e-01
-1.66515994e+00 2.84644723e-01 4.19381499e-01 -1.23858154e+00
7.29155481e-01 -4.91217166e-01 -1.35479733e-01 -4.67518717e-02
-6.08924925e-01 5.24505317e-01 1.14585793e+00 -8.98178458e-01
-5.58863133e-02 -3.10302079e-01 4.17696297e-01 -5.19075930e-01
-8.70133162e-01 -4.01218645e-02 3.34200226e-02 -6.32763207e-01
-3.37402880e-01 -1.01124060e+00 -4.89596695e-01 2.91209340e-01
-6.29773021e-01 2.56199986e-01 1.12830055e+00 -2.55660355e-01
7.78098345e-01 -2.22193170e+00 -3.82615864e-01 2.63095349e-01
4.21002746e-01 6.23705804e-01 4.80191689e-03 2.61663701e-02
2.70011961e-01 5.71040034e-01 -9.70383454e-03 -5.11151195e-01
3.60876322e-01 2.76911020e-01 -9.30693030e-01 4.41628516e-01
3.45763713e-01 9.83009577e-01 -7.56054282e-01 -2.26592198e-02
4.80342880e-02 9.35574055e-01 -8.22350144e-01 -2.04765439e-01
-5.49614191e-01 5.81433736e-02 -4.52505469e-01 6.04238391e-01
6.46663845e-01 -1.43339604e-01 1.63998187e-01 8.77262186e-03
4.22356218e-01 4.82056767e-01 -9.54242647e-01 1.34141195e+00
-4.96944129e-01 4.67207700e-01 3.80228996e-01 -3.64775419e-01
6.33654296e-01 2.21947700e-01 -1.82145655e-01 -2.53636897e-01
5.15632987e-01 1.67652011e-01 -1.63802087e-01 2.52655357e-01
3.09196562e-01 -9.74787548e-02 -4.25284177e-01 3.77485365e-01
1.35490328e-01 -1.03869937e-01 -6.21739268e-01 2.80237705e-01
1.50288558e+00 -2.03787491e-01 -1.10869743e-01 -5.76757193e-02
-1.53054476e-01 -2.84260035e-01 6.33588791e-01 1.22444260e+00
-4.16406482e-01 2.49001652e-01 7.43789494e-01 -4.11852688e-01
-1.08627856e+00 -1.25947118e+00 -6.27215132e-02 8.06348264e-01
-2.29317248e-01 -3.11968982e-01 -9.91568029e-01 -6.69243217e-01
1.19372316e-01 6.10062897e-01 -6.28818452e-01 -6.60557806e-01
-2.45596215e-01 -7.83962309e-01 1.49791861e+00 6.88820541e-01
4.09599543e-01 -5.17888188e-01 -5.13356090e-01 8.21612179e-02
7.88873494e-01 -1.01329291e+00 -1.55605271e-01 7.83858120e-01
-8.31597626e-01 -4.76855367e-01 -1.54376909e-01 -1.68867022e-01
3.82875174e-01 -1.74475759e-01 8.68125021e-01 1.35353416e-01
-2.26922974e-01 1.86430782e-01 3.95858586e-02 -6.30726457e-01
-6.35885894e-01 3.10502410e-01 5.28026104e-01 -3.91792417e-01
9.48277339e-02 -1.03377998e+00 -5.72674751e-01 -2.84857869e-01
-1.07132864e+00 -4.89754826e-01 6.38059616e-01 9.24583852e-01
3.28523278e-01 2.42797047e-01 4.58255768e-01 -9.21804786e-01
5.67148328e-01 -4.36649501e-01 -9.65259731e-01 1.05318837e-01
-4.79575753e-01 3.56648266e-01 1.11650360e+00 -7.25700855e-01
-5.35208762e-01 1.35348618e-01 -4.38783139e-01 -8.76727998e-01
2.76630260e-02 2.58678216e-02 -4.37117428e-01 -6.44739628e-01
1.01901996e+00 -1.32161081e-01 -2.72979021e-01 -2.02647388e-01
6.55445695e-01 4.59794849e-01 5.41204154e-01 -7.32916296e-01
1.04736316e+00 5.09930670e-01 1.32969677e-01 -4.40807760e-01
-5.42547107e-01 4.47042614e-01 -1.62483424e-01 2.63834894e-01
6.99058473e-01 -1.02111554e+00 -9.59884644e-01 5.53912401e-01
-9.98110712e-01 -5.22275507e-01 -3.17077667e-01 1.46240100e-01
-1.42652571e-01 1.73529714e-01 -9.67922032e-01 -8.54902506e-01
-6.43631399e-01 -1.37575805e+00 8.49886477e-01 2.88285673e-01
-1.88223384e-02 -8.63871932e-01 -1.89974472e-01 -9.46331117e-03
8.16357911e-01 2.87674218e-01 7.85005867e-01 -7.70817757e-01
-9.11105096e-01 -7.23072529e-01 1.14444569e-01 6.44593298e-01
-2.53928959e-01 1.51011929e-01 -1.46783233e+00 -4.13758546e-01
1.06491640e-01 -6.55081689e-01 9.06583726e-01 1.96719468e-02
1.37811494e+00 -6.01582885e-01 -7.27858767e-02 1.06383157e+00
1.43995357e+00 -3.23204249e-01 5.59975922e-01 1.30334809e-01
7.04620838e-01 -2.78911106e-02 -2.56553352e-01 3.47507030e-01
-6.44030645e-02 2.51677573e-01 8.05025339e-01 2.40307331e-01
1.35011464e-01 -5.41945875e-01 5.44831932e-01 2.05613256e-01
4.72641021e-01 -2.17334300e-01 -8.93075287e-01 2.78678209e-01
-1.31972134e+00 -8.79039884e-01 3.47292572e-01 2.23392582e+00
9.93758082e-01 9.36688840e-01 -2.48879626e-01 2.65329003e-01
3.01232696e-01 1.48704112e-01 -8.40135455e-01 -5.92695415e-01
3.62242796e-02 5.92749298e-01 1.17556417e+00 4.77705628e-01
-1.28000236e+00 9.37324941e-01 6.71451712e+00 9.93472695e-01
-1.42446494e+00 2.34760702e-01 9.61100936e-01 -4.11126167e-01
-4.21823323e-01 -4.12845835e-02 -1.01755691e+00 4.51883286e-01
1.58243680e+00 4.30371799e-02 4.15536374e-01 1.30158532e+00
-2.80933022e-01 3.91990453e-01 -1.10157001e+00 4.98774379e-01
-5.17664611e-01 -1.46981633e+00 -6.80060834e-02 6.10705651e-02
6.92238569e-01 4.52448696e-01 7.43335009e-01 4.08959210e-01
8.87919009e-01 -1.19589508e+00 7.70459056e-01 1.82414010e-01
7.78418124e-01 -9.59969401e-01 6.08517468e-01 3.73437941e-01
-5.91280401e-01 -1.88983634e-01 -4.86663789e-01 -6.71513751e-02
-1.71011910e-01 7.13699698e-01 -8.28185439e-01 -1.68787632e-02
5.85500360e-01 -4.33314852e-02 -5.52264988e-01 6.73190892e-01
-3.87562305e-01 9.90044057e-01 -6.91416442e-01 -1.45277336e-01
2.55753011e-01 4.92561042e-01 4.19692427e-01 9.72139418e-01
5.68609871e-02 -4.25199866e-01 -1.21936381e-01 8.81983936e-01
-3.59443426e-01 -7.48064995e-01 -6.26894116e-01 -2.31204972e-01
7.40522742e-01 1.04903579e+00 -5.02450883e-01 -1.71897590e-01
-1.41990408e-01 8.92184258e-01 3.20187509e-01 1.86678693e-01
-9.25102174e-01 -3.86850327e-01 1.15043414e+00 -1.41799703e-01
1.91810772e-01 -4.10605282e-01 -7.59978116e-01 -1.08729255e+00
-5.43805137e-02 -9.83030558e-01 -9.18695182e-02 -2.56894588e-01
-1.34351277e+00 5.90609610e-01 -5.73045611e-01 -8.21538270e-01
-1.59872726e-01 -7.62174606e-01 -6.72448516e-01 6.97737992e-01
-1.27491772e+00 -1.20652366e+00 3.02696437e-01 4.67038393e-01
-1.70701250e-01 -6.15076236e-02 9.23044026e-01 2.07362160e-01
-5.62818468e-01 1.47779024e+00 3.95588487e-01 2.53682196e-01
3.92260551e-01 -1.13551414e+00 9.07253981e-01 1.08914340e+00
-6.07517362e-02 1.08239484e+00 1.08659983e+00 -5.52082777e-01
-1.70965052e+00 -1.33507657e+00 3.23548168e-01 -6.42007232e-01
1.05272639e+00 -9.15701985e-01 -7.95094013e-01 8.82804453e-01
-1.69722319e-01 4.23550665e-01 6.73690856e-01 -4.86150980e-02
-8.68356287e-01 -1.82756439e-01 -1.49342692e+00 9.13315415e-01
8.07778001e-01 -1.01273358e+00 -1.31826580e-01 3.17802966e-01
1.41267741e+00 -3.91935289e-01 -7.72704244e-01 2.50596642e-01
5.20832598e-01 -8.18481088e-01 1.13173950e+00 -5.38705945e-01
6.93991855e-02 -9.18720216e-02 -3.57255638e-01 -9.06679213e-01
1.84693798e-01 -1.24392819e+00 -5.81809521e-01 1.15119255e+00
4.09650624e-01 -7.35908747e-01 1.19859397e+00 1.02696145e+00
-5.42220753e-03 -5.57155728e-01 -1.16639853e+00 -8.43261540e-01
5.37607253e-01 -7.13173628e-01 8.22200835e-01 6.01807475e-01
-2.21769229e-01 -1.56037316e-01 -4.87789780e-01 7.07650959e-01
8.71751845e-01 -7.67221451e-01 6.55289233e-01 -7.09542871e-01
-7.70769358e-01 -3.31487030e-01 -7.96342075e-01 -7.37997055e-01
1.80587813e-01 -6.01712644e-01 -8.61094892e-02 -6.14926755e-01
-3.01416159e-01 -4.38821703e-01 -5.62864423e-01 5.79055548e-01
-6.08718023e-02 2.90484190e-01 8.27176496e-02 -7.26860091e-02
-5.06085515e-01 4.97512817e-01 4.72074926e-01 -1.57546908e-01
-9.29093512e-04 1.06702231e-01 -8.44436288e-01 6.38721764e-01
9.64958131e-01 -7.57278621e-01 -4.83431160e-01 -7.21736073e-01
3.46910864e-01 -1.43597126e-01 6.62537038e-01 -1.37581289e+00
5.08576691e-01 3.68371874e-01 4.24334615e-01 -8.24316293e-02
3.59689802e-01 -7.96609521e-01 7.49921575e-02 7.37971246e-01
-3.26184273e-01 -1.33305758e-01 3.47449958e-01 8.19412470e-01
3.02985251e-01 3.11137666e-03 8.23450983e-01 1.04262076e-01
-1.07291639e-01 4.49073136e-01 -4.30925518e-01 -1.75009772e-01
8.58278751e-01 2.21265942e-01 -5.87237060e-01 -1.59741506e-01
-6.22020245e-01 -4.65039304e-03 7.47009933e-01 2.06568912e-01
5.16514897e-01 -9.32885528e-01 -2.12780938e-01 2.29424015e-01
-2.64562994e-01 -1.49709120e-01 2.39172474e-01 2.69928724e-01
-5.22079527e-01 2.90197253e-01 -6.20436408e-02 -1.37901336e-01
-9.06164050e-01 7.10219681e-01 6.87480628e-01 -3.53078663e-01
-4.71542925e-01 1.20933437e+00 -1.36904374e-01 -4.39763308e-01
5.96235693e-01 -4.19543028e-01 5.57507813e-01 -7.19471931e-01
7.15513468e-01 2.25441484e-03 2.33471513e-01 7.81989992e-02
-3.97864074e-01 -1.24985397e-01 -2.74152786e-01 -2.17645735e-01
1.19159973e+00 3.20912153e-01 9.12729800e-02 1.38034105e-01
1.08993971e+00 1.02914646e-01 -1.69204795e+00 -8.98563340e-02
-1.63035735e-01 -5.70579134e-02 3.11697990e-01 -7.43001819e-01
-9.77116227e-01 1.08818936e+00 7.55081654e-01 3.85310240e-02
8.35194170e-01 -4.11601901e-01 1.05770516e+00 6.63149893e-01
6.32840395e-01 -5.67827344e-01 -1.80365816e-01 5.61536372e-01
2.08033323e-01 -9.06856418e-01 -1.71906978e-01 -3.58735062e-02
-1.29719362e-01 8.65628660e-01 4.21783000e-01 -5.48521280e-01
9.24307048e-01 1.06728506e+00 4.50127833e-02 9.07574743e-02
-1.06148863e+00 4.06113982e-01 -2.13872790e-01 6.83964133e-01
6.01682663e-02 1.23077229e-01 2.92714447e-01 1.00752246e+00
-3.81475627e-01 -1.24713443e-01 5.25928140e-01 1.08963001e+00
-3.73298436e-01 -1.12496972e+00 -3.49215508e-01 3.88707846e-01
-1.04351211e+00 -2.92123258e-01 -2.95843840e-01 4.63598937e-01
1.12391576e-01 8.07733297e-01 -1.81437507e-01 -9.23159063e-01
-1.65146410e-01 5.85394399e-03 3.53306174e-01 -3.60893697e-01
-9.58211780e-01 -3.53508174e-01 -1.41144633e-01 -5.99318981e-01
5.17622173e-01 -2.67445445e-01 -1.07088387e+00 -6.27518594e-01
-2.80523926e-01 -1.48372009e-01 8.79341841e-01 8.02060962e-01
6.02380335e-01 6.13946080e-01 4.56042618e-01 -9.02406931e-01
-1.39454567e+00 -4.27322298e-01 -3.35121244e-01 -2.29787990e-01
6.38868988e-01 -2.64897376e-01 -6.51307464e-01 -3.41054678e-01] | [5.706996917724609, 7.7302374839782715] |
dbb92058-f5da-4615-80c4-0f5c664a0bae | item-tagging-for-information-retrieval-a | 2008.11567 | null | https://arxiv.org/abs/2008.11567v1 | https://arxiv.org/pdf/2008.11567v1.pdf | Item Tagging for Information Retrieval: A Tripartite Graph Neural Network based Approach | Tagging has been recognized as a successful practice to boost relevance matching for information retrieval (IR), especially when items lack rich textual descriptions. A lot of research has been done for either multi-label text categorization or image annotation. However, there is a lack of published work that targets at item tagging specifically for IR. Directly applying a traditional multi-label classification model for item tagging is sub-optimal, due to the ignorance of unique characteristics in IR. In this work, we propose to formulate item tagging as a link prediction problem between item nodes and tag nodes. To enrich the representation of items, we leverage the query logs available in IR tasks, and construct a query-item-tag tripartite graph. This formulation results in a TagGNN model that utilizes heterogeneous graph neural networks with multiple types of nodes and edges. Different from previous research, we also optimize both full tag prediction and partial tag completion cases in a unified framework via a primary-dual loss mechanism. Experimental results on both open and industrial datasets show that our TagGNN approach outperforms the state-of-the-art multi-label classification approaches. | ['Ruiming Tang', 'Jieming Zhu', 'Xi Xiao', 'Biao Lu', 'Xiuqiang He', 'Kelong Mao'] | 2020-08-26 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 3.86350334e-01 4.90930341e-02 -8.83775771e-01 -4.01363015e-01
-1.07512295e+00 -3.91735733e-01 2.99638301e-01 3.64121974e-01
-4.15162921e-01 5.83148658e-01 1.80672914e-01 -4.52433713e-02
-3.22050571e-01 -5.27365923e-01 -4.62888092e-01 -4.66022670e-01
1.65389642e-01 5.65279663e-01 2.27426603e-01 -4.47671860e-02
8.31135809e-02 -2.74405275e-02 -1.29414082e+00 4.56366897e-01
6.22991979e-01 1.19084680e+00 7.98458457e-02 1.31161481e-01
-2.80139178e-01 8.60900342e-01 -1.50130957e-01 -8.20297182e-01
1.80555835e-01 -7.77428374e-02 -1.20475733e+00 -9.47169065e-02
6.73024476e-01 2.31573917e-03 -1.71647146e-01 1.08903432e+00
6.56496823e-01 1.71389416e-01 6.72416329e-01 -1.28349233e+00
-1.13066328e+00 9.01656151e-01 -1.00734186e+00 -1.61077261e-01
2.28308111e-01 -8.41389894e-01 1.74424744e+00 -8.25251460e-01
3.53660733e-01 1.30865455e+00 8.30255151e-01 5.23138225e-01
-1.00988030e+00 -6.74879134e-01 2.71842957e-01 3.06245051e-02
-1.56789756e+00 1.99731439e-02 6.97758973e-01 -2.22463191e-01
7.22096384e-01 7.10072648e-03 1.67775422e-01 8.78784657e-01
1.23259053e-01 1.03204691e+00 9.25525606e-01 -7.26403177e-01
-3.19903374e-01 2.93738008e-01 5.45211256e-01 6.98606372e-01
4.52695817e-01 -2.23660201e-01 -3.45484495e-01 -4.31774050e-01
4.30039793e-01 2.95739233e-01 -3.81822251e-02 -6.60501122e-01
-7.84789562e-01 1.06196928e+00 5.04781067e-01 4.36884046e-01
-2.74522424e-01 3.80636424e-01 4.44294810e-01 2.04050139e-01
9.73987281e-01 2.88716674e-01 -3.81331086e-01 5.73770583e-01
-6.31812692e-01 -1.09896854e-01 7.10525393e-01 1.13107777e+00
9.87792432e-01 -4.64459300e-01 -4.77043897e-01 1.25405133e+00
6.64703667e-01 2.16923431e-01 3.37845802e-01 -5.42639494e-01
4.91021216e-01 7.62297392e-01 -6.73593730e-02 -1.02691126e+00
-6.68339133e-01 -8.95785987e-01 -7.99367964e-01 -4.89080876e-01
-3.89646254e-02 -2.18467768e-02 -7.06412554e-01 1.69074214e+00
3.68437231e-01 2.98125982e-01 -4.54277620e-02 8.98054540e-01
8.81537259e-01 3.09606075e-01 7.01744258e-01 -1.03555039e-01
1.71274793e+00 -1.24185121e+00 -7.00514138e-01 -4.41788822e-01
1.21641946e+00 -7.77152002e-01 8.75345945e-01 4.95265126e-02
-7.60615587e-01 -4.20265794e-01 -7.13875890e-01 -1.34519592e-01
-4.80236858e-01 1.89514056e-01 1.10621071e+00 6.83190107e-01
-1.09135067e+00 1.64967433e-01 -8.75382796e-02 -3.59074682e-01
3.54923755e-01 5.88776171e-01 -3.89066249e-01 -4.85465497e-01
-1.57876456e+00 6.75475657e-01 5.89622557e-01 -1.08304657e-01
-3.93899679e-01 -5.88142455e-01 -7.95737743e-01 1.16514400e-01
6.33352757e-01 -8.96948218e-01 1.25163472e+00 -1.00552201e+00
-7.30907083e-01 9.93815720e-01 1.34612948e-01 -1.76588058e-01
-1.76271915e-01 -1.65100008e-01 -5.68033397e-01 5.94464727e-02
1.84159905e-01 7.38558471e-01 5.67168474e-01 -1.57894063e+00
-8.95212710e-01 -4.48916018e-01 3.54755163e-01 5.19921184e-01
-8.03677201e-01 2.83693671e-01 -7.63301194e-01 -6.72311783e-01
2.20463537e-02 -1.01449001e+00 -1.13308661e-01 -2.22426191e-01
-3.73683065e-01 -6.30059838e-01 4.21235710e-01 -3.50047529e-01
1.45085335e+00 -1.90004134e+00 -1.61402345e-01 2.18627319e-01
3.04542631e-01 1.15046814e-01 -3.01820099e-01 4.51857209e-01
1.08702257e-01 2.53470212e-01 4.15753841e-01 -8.13101709e-01
1.62016436e-01 5.58724105e-02 -8.35232511e-02 2.90071458e-01
-2.89089918e-01 1.22061622e+00 -9.34086621e-01 -1.01893902e+00
-3.17929655e-01 5.59601545e-01 -4.75132972e-01 -5.83910495e-02
-8.30681920e-02 6.49362653e-02 -8.74810815e-01 9.71514940e-01
5.67274928e-01 -9.50221837e-01 3.46609056e-01 -4.67389196e-01
4.02528644e-01 4.94837053e-02 -1.06102765e+00 1.98961937e+00
-5.93204081e-01 -1.21250525e-01 -5.26822396e-02 -1.13989162e+00
6.09241605e-01 5.31470835e-01 9.28987026e-01 -7.30666220e-01
6.21003099e-02 1.37120977e-01 -6.36551857e-01 -1.99799210e-01
8.44665051e-01 -9.63619798e-02 -3.05946827e-01 5.64532518e-01
2.63034135e-01 6.72772825e-01 1.51816413e-01 4.52249378e-01
9.50220227e-01 2.83067506e-02 2.77580500e-01 -1.43387041e-03
3.25369745e-01 -6.82259500e-02 2.94147491e-01 1.05322063e+00
1.54138386e-01 3.74015868e-01 1.82569362e-02 -2.58295745e-01
-6.10194266e-01 -6.08066916e-01 -9.69165862e-02 1.85852134e+00
4.19194072e-01 -2.84340352e-01 -2.39697352e-01 -1.11463416e+00
1.39056584e-02 2.89097607e-01 -5.81388235e-01 -2.66304970e-01
-3.59536737e-01 -1.00824177e+00 4.24623132e-01 4.18195754e-01
2.09681109e-01 -1.05675697e+00 2.17060968e-01 2.87326783e-01
-4.20180351e-01 -1.18718505e+00 -8.28158736e-01 3.32733631e-01
-6.59911752e-01 -8.48042667e-01 -7.24905789e-01 -1.19078350e+00
5.43988347e-01 7.81373560e-01 1.40407443e+00 3.79848570e-01
-2.10039183e-01 7.44543433e-01 -8.18720341e-01 -2.49748528e-01
-1.01368867e-01 4.81400192e-01 -1.59203380e-01 2.24799737e-01
5.48695028e-01 -1.62110090e-01 -7.12758303e-01 5.88498831e-01
-1.10800385e+00 -4.15538698e-02 8.76774967e-01 9.66274023e-01
6.99487627e-01 -2.83111811e-01 9.74906266e-01 -1.12968528e+00
6.11590266e-01 -7.89620399e-01 -3.48791957e-01 8.67106080e-01
-1.24432182e+00 3.86692137e-02 1.29734501e-01 -5.18966079e-01
-1.07048380e+00 3.58544402e-02 4.11085896e-02 -8.88330266e-02
-4.66968631e-03 8.79859269e-01 2.52936017e-02 -3.76841456e-01
2.39363268e-01 -2.51571238e-01 -4.88260180e-01 -6.64605856e-01
4.50147986e-01 8.39275479e-01 -7.61401132e-02 -5.08474410e-01
3.42539579e-01 2.95114487e-01 2.17827618e-01 -6.22840002e-02
-1.74860024e+00 -1.38761353e+00 -4.89125311e-01 -1.82484925e-01
8.37654769e-01 -1.12557459e+00 -7.81359375e-01 1.01902917e-01
-9.98063028e-01 1.17325589e-01 2.96839364e-02 6.46197140e-01
-2.66438723e-01 5.75831175e-01 -7.87971556e-01 -6.64190829e-01
-7.40618944e-01 -8.45350683e-01 1.41211283e+00 2.94885151e-02
2.76531786e-01 -1.34599710e+00 2.34810114e-01 7.44467258e-01
2.92150319e-01 -5.51900089e-01 9.97042775e-01 -9.42945004e-01
-4.97975737e-01 -4.60449308e-01 -6.75838351e-01 9.50487852e-02
7.99005926e-02 -1.00364184e+00 -8.82694900e-01 -5.25549412e-01
-5.64418495e-01 -8.35901499e-01 1.07807970e+00 2.68703341e-01
9.27955031e-01 -3.67845863e-01 -8.19485664e-01 3.47792417e-01
1.79909742e+00 -7.13904724e-02 2.39480197e-01 3.03816527e-01
9.51125383e-01 5.36764622e-01 9.25138831e-01 4.42380190e-01
7.11969376e-01 1.18437719e+00 5.74840665e-01 -3.40766221e-01
-1.82353035e-01 -3.43765080e-01 -6.60886467e-02 6.17105305e-01
1.62733361e-01 -1.01767778e+00 -6.62144840e-01 5.05665779e-01
-2.03293586e+00 -7.41906106e-01 -2.26395920e-01 2.03549790e+00
7.88988173e-01 -3.10245275e-01 -1.50002807e-01 -3.09960842e-01
8.46294045e-01 -8.80638417e-03 -2.00788707e-01 1.54072538e-01
-1.34861335e-01 1.96275730e-02 7.78810918e-01 1.08958930e-01
-1.54570735e+00 1.02536500e+00 5.67480707e+00 1.17889059e+00
-4.99567419e-01 4.66433764e-01 3.96697402e-01 3.37235570e-01
-2.81398684e-01 -1.53458500e-02 -1.32349384e+00 8.30320194e-02
7.18490660e-01 1.50948003e-01 6.67823032e-02 7.65163481e-01
-3.18850160e-01 1.84399143e-01 -8.31512034e-01 9.42322314e-01
4.76241261e-01 -8.91524434e-01 3.33140016e-01 1.89383522e-01
6.80987000e-01 -9.56234783e-02 1.11629538e-01 5.08875608e-01
5.30781627e-01 -6.05922043e-01 3.55883181e-01 3.86305809e-01
6.73268080e-01 -4.62749988e-01 1.04822505e+00 1.74070582e-01
-1.56660855e+00 -2.14142904e-01 -5.15649617e-01 4.08062428e-01
2.60502905e-01 5.09460866e-01 -7.10505545e-01 1.03535199e+00
5.63276768e-01 6.90170288e-01 -7.16092825e-01 1.21667755e+00
1.48330972e-01 5.92814088e-01 1.45110395e-02 3.05712633e-02
3.10463488e-01 -6.83876574e-02 3.66957746e-02 1.20809853e+00
3.87325525e-01 -1.50884718e-01 6.89747214e-01 1.34774476e-01
-5.98623276e-01 7.49694407e-01 -4.13511664e-01 1.81398749e-01
3.56523365e-01 1.74741709e+00 -6.15287423e-01 -1.52808979e-01
-1.05938256e+00 1.01459575e+00 5.92310369e-01 2.42457166e-01
-7.01990843e-01 -5.40307760e-02 1.63390204e-01 -2.05083728e-01
1.20462500e-01 2.19076946e-01 6.51639253e-02 -9.81735110e-01
-1.60330310e-01 -3.19891006e-01 1.07064140e+00 -7.72556067e-01
-1.69970834e+00 4.24764782e-01 8.55928957e-02 -1.31453121e+00
-1.82304010e-02 -4.42360699e-01 1.39055058e-01 7.12301850e-01
-1.96490657e+00 -2.01019573e+00 -1.72498673e-01 5.12020469e-01
2.27091134e-01 -4.08532321e-02 9.00375485e-01 9.65772808e-01
-2.95821458e-01 1.04807138e+00 8.95496905e-02 4.59629446e-02
1.11041892e+00 -1.08760488e+00 -1.81132972e-01 2.89303660e-01
4.84068096e-01 4.25806582e-01 1.68740228e-01 -7.79733598e-01
-1.19449902e+00 -1.38114810e+00 1.16934264e+00 -2.38913819e-01
5.73159218e-01 -2.43738592e-01 -8.27731609e-01 6.88568234e-01
2.90047765e-01 2.16067404e-01 1.05853546e+00 4.62144166e-01
-5.53467751e-01 -9.63188335e-02 -8.92922640e-01 1.70897812e-01
1.19774747e+00 -3.50323826e-01 -1.66573465e-01 6.79641604e-01
9.63568091e-01 7.69422725e-02 -1.11717224e+00 3.68927807e-01
5.44870138e-01 -1.76730707e-01 1.29025161e+00 -6.39658988e-01
5.32741547e-02 -6.37835935e-02 -1.59744158e-01 -1.09889913e+00
-7.05848336e-01 -7.41656646e-02 -5.51905297e-02 1.52531707e+00
6.20223343e-01 -3.82603735e-01 7.95041800e-01 3.79590452e-01
-3.55059117e-01 -6.15402520e-01 -6.49196446e-01 -7.66248584e-01
-1.88650623e-01 -1.55443043e-01 3.20211023e-01 9.92858469e-01
-9.52854101e-03 7.70546317e-01 -9.41517115e-01 1.38773695e-01
8.00095022e-01 2.62437612e-01 2.11679950e-01 -1.73831975e+00
-2.75221050e-01 -1.50843129e-01 -1.07736267e-01 -1.19112766e+00
5.75487673e-01 -1.42133701e+00 -8.32256228e-02 -1.87205195e+00
8.65449369e-01 -9.85983372e-01 -8.72136056e-01 9.34918940e-01
-3.28192174e-01 5.94192624e-01 -1.46668972e-02 5.16355455e-01
-1.48452997e+00 3.36264014e-01 1.03861558e+00 -3.92774671e-01
9.26769376e-02 -8.69205315e-03 -8.60012949e-01 3.39971453e-01
5.86842060e-01 -8.60192060e-01 -3.76053780e-01 -5.87858617e-01
8.47399771e-01 1.15088090e-01 2.75717199e-01 -6.05710030e-01
4.33949053e-01 -1.26949344e-02 -6.64857542e-03 -5.03904283e-01
4.51401830e-01 -1.16746056e+00 4.82989214e-02 -7.15777650e-02
-6.93011940e-01 -5.46307378e-02 -2.43743256e-01 8.64768386e-01
-2.62708217e-01 -5.53352416e-01 3.70390803e-01 -1.86174423e-01
-8.04753780e-01 7.28096724e-01 1.31829187e-01 1.78000629e-01
6.55844927e-01 5.88325076e-02 -4.40908194e-01 -2.75028855e-01
-4.34887975e-01 4.00248766e-01 4.62848693e-02 6.53504431e-01
3.61877501e-01 -1.69154942e+00 -5.48407674e-01 -3.73389482e-01
7.08423793e-01 -4.92340595e-01 4.45553005e-01 7.76554883e-01
4.32788044e-01 7.37996876e-01 2.14298725e-01 -2.92028457e-01
-1.36332369e+00 9.14628565e-01 -6.21055067e-02 -9.79714990e-01
-1.83448151e-01 6.54433489e-01 3.95238757e-01 -3.34328294e-01
3.05697858e-01 4.23949182e-01 -7.16983914e-01 3.25861782e-01
2.28300348e-01 1.39836922e-01 2.27966517e-01 -7.30969727e-01
-1.88665062e-01 7.75249958e-01 -2.91258842e-01 2.49492556e-01
1.07974029e+00 -6.20813429e-01 -1.51507959e-01 2.10192576e-01
1.33084130e+00 -3.03836852e-01 -3.98581475e-01 -8.06591988e-01
2.75091588e-01 -1.32086769e-01 2.66225338e-01 -9.24436390e-01
-1.17388678e+00 6.13668501e-01 7.45546460e-01 2.59675324e-01
1.04642248e+00 4.80404824e-01 8.03237855e-01 3.75125796e-01
8.49779129e-01 -1.07555509e+00 2.70972878e-01 3.60755593e-01
4.89966393e-01 -1.50242960e+00 -1.29367515e-01 -5.57193696e-01
-5.35038471e-01 6.36383176e-01 6.79165363e-01 5.48822165e-01
7.04834938e-01 -2.47012198e-01 -1.65159665e-02 -4.22094166e-01
-5.10370851e-01 -5.26894689e-01 7.77452588e-01 3.79455209e-01
7.35007226e-01 -2.20309924e-02 -4.61815447e-01 7.27145731e-01
7.14049160e-01 -4.33066441e-03 -5.09105064e-02 9.45867002e-01
-4.74766821e-01 -1.53114390e+00 -9.02345777e-02 7.90824831e-01
-9.76876915e-01 -4.78497773e-01 -8.51530731e-02 5.02918720e-01
7.49865994e-02 1.11391175e+00 -2.37651050e-01 -3.63189161e-01
1.13820441e-01 6.53264374e-02 2.18161911e-01 -9.12152171e-01
-7.91798949e-01 6.43406287e-02 1.74418092e-01 -3.05165529e-01
-8.13407421e-01 -2.93360293e-01 -1.00368571e+00 2.41335511e-01
-9.28552926e-01 4.43513900e-01 7.45253444e-01 7.82624245e-01
4.70509589e-01 3.85362267e-01 5.35260797e-01 -3.95095319e-01
-5.89154661e-01 -1.03306997e+00 -8.64865124e-01 6.54239357e-01
-5.08806221e-02 -9.05065775e-01 -1.44911066e-01 -2.40414172e-01] | [9.6884765625, 4.430361270904541] |
baacf6d5-4d22-4839-a04a-7a8592bc614b | leveraging-video-coding-knowledge-for-deep | 2302.13594 | null | https://arxiv.org/abs/2302.13594v1 | https://arxiv.org/pdf/2302.13594v1.pdf | Leveraging Video Coding Knowledge for Deep Video Enhancement | Recent advancements in deep learning techniques have significantly improved the quality of compressed videos. However, previous approaches have not fully exploited the motion characteristics of compressed videos, such as the drastic change in motion between video contents and the hierarchical coding structure of the compressed video. This study proposes a novel framework that leverages the low-delay configuration of video compression to enhance the existing state-of-the-art method, BasicVSR++. We incorporate a context-adaptive video fusion method to enhance the final quality of compressed videos. The proposed approach has been evaluated in the NTIRE22 challenge, a benchmark for video restoration and enhancement, and achieved improvements in both quantitative metrics and visual quality compared to the previous method. | ['Van-Quang Nguyen', 'Thuong Nguyen Canh', 'Thong Bach'] | 2023-02-27 | null | null | null | null | ['video-enhancement', 'video-restoration'] | ['computer-vision', 'computer-vision'] | [ 2.16198027e-01 -7.44802892e-01 -2.07091942e-01 -1.94425315e-01
-6.47920310e-01 -1.95909932e-01 3.68337393e-01 1.95319988e-02
-3.45761836e-01 4.85753059e-01 7.61507630e-01 -1.91051275e-01
-8.34448040e-02 -4.67839122e-01 -7.45502770e-01 -4.79614794e-01
-5.32629728e-01 -4.47491437e-01 2.38561705e-01 -2.52140939e-01
3.84907156e-01 2.21852526e-01 -1.69491374e+00 9.28695381e-01
4.77182746e-01 1.25275373e+00 4.91369873e-01 9.29030299e-01
1.91757992e-01 1.22662818e+00 -2.54162937e-01 7.05212653e-02
4.24475163e-01 -2.20692188e-01 -6.97546244e-01 3.32566381e-01
6.58647418e-01 -1.09043229e+00 -1.24530256e+00 9.11263049e-01
5.13504386e-01 1.77218497e-01 5.65131158e-02 -9.07246709e-01
-6.79719687e-01 5.39695203e-01 -7.00160265e-01 7.59174049e-01
4.29731518e-01 3.85247134e-02 8.53353202e-01 -6.95964098e-01
6.85970604e-01 1.28223729e+00 5.66252708e-01 3.77042741e-01
-8.68883431e-01 -5.26227891e-01 -6.39311820e-02 1.10573590e+00
-1.21390772e+00 -7.16041863e-01 5.33560514e-01 -2.75809437e-01
1.37735689e+00 -5.06262034e-02 6.96538806e-01 7.92060435e-01
4.22944427e-01 8.45856428e-01 6.56796336e-01 -3.93459469e-01
5.42587452e-02 -6.47093654e-01 -2.77162850e-01 4.55876291e-01
-1.30381480e-01 4.08161014e-01 -7.18424559e-01 3.68341178e-01
6.55602574e-01 1.83009088e-01 -6.38812363e-01 -2.48545811e-01
-1.04374719e+00 5.89270294e-01 4.02076602e-01 4.18623239e-01
-5.41842759e-01 4.32045937e-01 7.58192480e-01 3.99248511e-01
3.91394466e-01 -1.74507678e-01 -4.13521290e-01 -4.44472998e-01
-1.72692084e+00 2.56319672e-01 3.55445027e-01 5.99066317e-01
4.35925066e-01 4.49815601e-01 -4.31592733e-01 6.32749319e-01
5.79177260e-01 2.29176596e-01 4.88035828e-01 -1.65901697e+00
5.41906595e-01 2.71549433e-01 -2.49809958e-02 -1.24931550e+00
8.94933492e-02 -3.35882336e-01 -7.75319099e-01 2.93423355e-01
-3.20924699e-01 2.66380757e-01 -9.43056941e-01 1.48573983e+00
-8.81933514e-03 5.22336245e-01 1.99104995e-01 9.45015669e-01
9.83999550e-01 8.96149099e-01 -1.06163725e-01 -2.28802785e-01
7.79396236e-01 -1.30474150e+00 -9.86336052e-01 1.06305957e-01
2.93769926e-01 -8.75372410e-01 5.04118741e-01 6.70221210e-01
-1.55560708e+00 -9.24153030e-01 -1.35809624e+00 -1.50120914e-01
2.80301064e-01 -1.86815813e-01 2.95226455e-01 4.53683853e-01
-1.65342867e+00 9.62107956e-01 -9.43174899e-01 2.10194625e-02
7.17911005e-01 3.77362669e-01 -4.69449371e-01 -6.53304696e-01
-1.08637893e+00 4.24808621e-01 4.95745242e-01 -1.11100934e-01
-1.30873621e+00 -9.44330513e-01 -8.55014086e-01 3.80603820e-01
3.72433782e-01 -5.52551448e-01 1.37782228e+00 -1.04555380e+00
-1.15434206e+00 2.88617551e-01 -1.10744253e-01 -8.25056791e-01
3.42203677e-01 -5.87805986e-01 -4.20439512e-01 7.99566269e-01
-3.50937963e-01 9.83189344e-01 9.31559563e-01 -1.17870653e+00
-7.90760994e-01 5.59395626e-02 2.63241082e-01 1.75468519e-01
-4.06691611e-01 1.18634000e-01 -7.62226701e-01 -8.64952385e-01
-2.01368511e-01 -4.03411776e-01 -1.28795013e-01 2.13539258e-01
2.18638241e-01 3.11504632e-01 1.44028389e+00 -1.17304933e+00
1.37694180e+00 -2.33505344e+00 1.98610350e-01 -2.51789927e-01
3.78003687e-01 6.71563625e-01 -4.44017798e-01 4.42402959e-01
-5.62629290e-02 -3.38552482e-02 -5.88976592e-02 -5.29728413e-01
-3.47943276e-01 2.00598165e-01 -3.03978086e-01 4.05546129e-01
-2.19733045e-01 6.40655756e-01 -8.18868220e-01 -5.47868490e-01
5.91342032e-01 1.09501088e+00 -1.02341688e+00 3.85732502e-01
-6.05992861e-02 2.44958162e-01 1.02991782e-01 5.42418897e-01
8.36181641e-01 -2.62772560e-01 2.83970684e-01 -5.93297601e-01
-5.67203015e-02 2.03675568e-01 -1.03829253e+00 2.04890275e+00
-2.88765401e-01 1.01485896e+00 3.11912239e-01 -1.01798022e+00
3.66512239e-01 6.41504288e-01 8.89923513e-01 -9.13516343e-01
1.56786129e-01 1.88825894e-02 -8.36305693e-02 -6.68020248e-01
8.49566638e-01 3.00368786e-01 7.49655843e-01 -7.37309754e-02
1.17554121e-01 3.08170527e-01 4.58273828e-01 4.85513091e-01
1.26565528e+00 5.06873786e-01 2.82814711e-01 1.02362754e-02
5.89152038e-01 -3.60440373e-01 4.97555882e-01 6.20074809e-01
-4.41148788e-01 7.93899298e-01 -1.39986770e-02 -2.79085219e-01
-1.29212546e+00 -9.92738843e-01 2.20876649e-01 7.89075911e-01
1.51863679e-01 -7.50843942e-01 -7.01029956e-01 -3.90976101e-01
-2.69695342e-01 2.86128372e-01 -3.04200768e-01 -2.22813293e-01
-7.30880082e-01 -2.81642437e-01 3.27139795e-01 6.89480066e-01
6.90361321e-01 -9.58314478e-01 -7.98361063e-01 4.12996709e-01
-5.60171485e-01 -1.63456428e+00 -3.88982177e-01 -4.64535445e-01
-1.04989827e+00 -9.42758679e-01 -5.44553041e-01 -6.85133100e-01
-6.31380826e-02 8.42560649e-01 1.20674729e+00 5.37992597e-01
-2.61653632e-01 5.24152458e-01 -8.73332024e-01 4.41056520e-01
-5.06449103e-01 -4.97469902e-01 -2.20069945e-01 -3.39145213e-02
-1.52625561e-01 -6.87633753e-01 -1.04260445e+00 -7.48498067e-02
-1.29810691e+00 1.57008395e-02 4.75146443e-01 6.67291164e-01
4.73220468e-01 4.43599641e-01 3.68791312e-01 -1.15531459e-01
1.47575244e-01 -4.92840767e-01 -2.14511648e-01 -1.78291947e-02
-5.74328125e-01 -1.71826839e-01 5.45655966e-01 -1.82811260e-01
-9.81736064e-01 -3.03780884e-01 -3.98261696e-01 -8.13234448e-01
3.89374569e-02 6.35731041e-01 -4.76234127e-03 -2.45280668e-01
-2.68324371e-03 3.04644734e-01 -4.42810059e-02 -5.59140623e-01
1.41464740e-01 6.81735456e-01 8.97549093e-01 -1.07914038e-01
3.42304349e-01 6.35235667e-01 1.00274973e-01 -7.63887942e-01
-3.56149256e-01 -4.87094343e-01 -3.20274025e-01 -3.41353804e-01
6.45116389e-01 -1.30866492e+00 -4.35681701e-01 6.17503047e-01
-1.21909201e+00 -3.16757619e-01 -1.08238541e-01 6.42494917e-01
-6.33105516e-01 1.20376158e+00 -1.08904326e+00 -2.35333130e-01
-6.02385521e-01 -1.45517647e+00 8.00412357e-01 1.53860934e-02
2.65621454e-01 -9.25043643e-01 7.46205598e-02 4.60860670e-01
9.21585083e-01 1.36999428e-01 7.34698653e-01 1.80283152e-02
-7.34848440e-01 3.79561745e-02 -4.10843700e-01 8.11857998e-01
5.32130264e-02 1.42458326e-03 -6.43677711e-01 -8.43512297e-01
-3.25176232e-02 -2.50771582e-01 1.22726893e+00 6.62617683e-01
1.26560104e+00 -2.50529408e-01 -2.28803921e-02 9.29073572e-01
1.71758842e+00 2.25485384e-01 1.08420205e+00 4.73480910e-01
5.67122340e-01 1.49470150e-01 5.64798295e-01 7.44003713e-01
3.85628551e-01 8.16471934e-01 8.08766365e-01 5.37381545e-02
-6.48831666e-01 3.34266275e-02 6.41538620e-01 9.26551163e-01
-3.12610596e-01 -5.39019287e-01 -5.40950000e-01 5.42963326e-01
-1.81080925e+00 -1.17141008e+00 1.86988518e-01 1.81335640e+00
6.05339944e-01 -6.66932017e-02 -3.65563244e-01 4.44293588e-01
5.94850123e-01 5.54532766e-01 -3.03744555e-01 -1.93814844e-01
-1.54555976e-01 2.44865045e-01 2.88448215e-01 3.92740130e-01
-1.14901638e+00 7.26055384e-01 7.15821648e+00 9.10694003e-01
-1.14549494e+00 1.94274962e-01 2.48234242e-01 -4.02053177e-01
7.63604343e-02 -8.59595463e-02 -3.76059860e-01 3.93070728e-01
1.17320395e+00 4.29632841e-03 5.63334227e-01 4.85072970e-01
6.67546749e-01 3.34673040e-02 -1.06863606e+00 1.07221127e+00
3.41323555e-01 -1.82493293e+00 2.85416007e-01 4.12493013e-02
8.23749304e-01 1.91957101e-01 2.69272536e-01 1.70342386e-01
-1.56289995e-01 -8.72667134e-01 8.00687730e-01 4.25968617e-01
7.70589352e-01 -6.93468094e-01 8.90724301e-01 -1.22124910e-01
-1.50830269e+00 -2.59768754e-01 -1.68029472e-01 -1.00506872e-01
5.85118413e-01 4.40733552e-01 -2.58907199e-01 7.71101654e-01
1.21035576e+00 1.24704874e+00 -5.55430293e-01 1.19063139e+00
1.58515811e-01 7.44551539e-01 1.69792116e-01 9.37114418e-01
3.84796619e-01 2.65187413e-01 6.86688960e-01 1.41102064e+00
5.71986079e-01 -9.12620500e-02 1.66195348e-01 2.44376659e-01
-1.94965988e-01 -7.08582252e-02 -1.91936582e-01 -8.63619521e-02
1.53618187e-01 9.91067052e-01 -2.48251647e-01 -5.05355835e-01
-7.85385787e-01 8.85448217e-01 -1.44757241e-01 3.31795096e-01
-9.22112584e-01 9.42502022e-02 6.59872890e-01 2.21044332e-01
1.04071808e+00 -4.64930207e-01 4.69226629e-01 -1.14282763e+00
-1.26831029e-02 -1.41044915e+00 2.89861768e-01 -8.82819474e-01
-6.95179105e-01 6.28080964e-01 -1.13217182e-01 -1.24835515e+00
-1.98268384e-01 -2.99167931e-01 -2.95013279e-01 3.48846704e-01
-1.86236358e+00 -1.00466239e+00 -4.76807922e-01 7.98569918e-01
9.39336121e-01 -3.94779354e-01 3.69871944e-01 8.62306893e-01
-2.80962735e-01 4.70931530e-01 3.39111745e-01 -1.74025849e-01
6.99274600e-01 -6.88068807e-01 1.48806989e-01 1.41679513e+00
-1.09543040e-01 2.27773115e-01 7.40678072e-01 -5.57025611e-01
-1.56616628e+00 -1.10114813e+00 4.49571490e-01 4.58186954e-01
3.12374234e-01 1.85371861e-01 -8.88819933e-01 4.69301969e-01
7.54585981e-01 2.77497858e-01 5.29176712e-01 -6.44126654e-01
-5.24634361e-01 -1.84630558e-01 -1.18636036e+00 2.99604774e-01
8.31868470e-01 -4.98451352e-01 -3.45145464e-01 -5.67793325e-02
1.08288908e+00 -2.85261750e-01 -8.45351100e-01 7.52445996e-01
5.60083210e-01 -1.32191908e+00 1.22667062e+00 -3.77168238e-01
1.09075403e+00 -3.41737539e-01 -9.53100860e-01 -1.00978136e+00
-4.81507093e-01 -5.01428604e-01 -8.75117719e-01 9.78704870e-01
-4.66187596e-01 2.95632005e-01 6.74600601e-01 1.58168469e-02
-3.95356178e-01 -7.99839318e-01 -9.17881429e-01 -5.62400043e-01
-3.81265610e-01 -4.71455991e-01 3.23543131e-01 5.78352213e-01
-3.68782729e-01 -1.68867350e-01 -9.27865565e-01 1.18448511e-01
6.55360043e-01 -1.98679686e-01 2.35971928e-01 -5.50957918e-01
-6.64507329e-01 -2.09856808e-01 -8.13456178e-01 -1.23215699e+00
-6.34287149e-02 -6.66813135e-01 -1.10845417e-01 -1.75618172e+00
4.40371633e-01 1.31507084e-01 -7.05935359e-01 2.14689791e-01
4.26238263e-03 6.21587873e-01 5.99152684e-01 1.11677021e-01
-1.04996872e+00 7.13981330e-01 9.73261297e-01 -3.03590268e-01
2.78240442e-02 -7.17416346e-01 -4.83348250e-01 4.63810384e-01
5.81995845e-01 -2.93529660e-01 -3.47152859e-01 -9.20563400e-01
1.09230198e-01 2.80789584e-01 4.78638023e-01 -1.41693676e+00
3.12679261e-01 1.90956742e-01 3.95917565e-01 -8.31858099e-01
3.77239645e-01 -8.52855146e-01 1.63439929e-01 6.53755426e-01
-2.94001222e-01 3.31988454e-01 3.79602313e-01 7.05316424e-01
-5.85782290e-01 -2.21979618e-02 9.56189156e-01 4.33078147e-02
-1.22627389e+00 3.81566882e-01 -5.29944956e-01 -1.91448495e-01
8.02245796e-01 -1.77388325e-01 -2.39812985e-01 -6.36478543e-01
-4.05799508e-01 -4.20097038e-02 4.40620869e-01 5.63430130e-01
1.23656797e+00 -1.18279123e+00 -9.03159440e-01 5.65476492e-02
-4.32678372e-01 -6.17019117e-01 8.27787578e-01 8.39820445e-01
-7.59794831e-01 3.59505177e-01 -6.83013916e-01 -6.49505556e-01
-1.54488325e+00 8.26068342e-01 2.34786883e-01 -3.88545066e-01
-8.65778029e-01 4.53939974e-01 -1.13035642e-01 4.67773020e-01
3.81124079e-01 -7.69090429e-02 -4.55071092e-01 -3.58661771e-01
1.01587737e+00 6.63329959e-01 6.47279695e-02 -8.16085398e-01
-1.30374908e-01 4.77850497e-01 -3.59716296e-01 3.19029726e-02
1.56653678e+00 -6.57903433e-01 -7.03541040e-02 -1.97568029e-01
1.45628631e+00 -2.34854758e-01 -1.52868986e+00 -2.58726388e-01
-4.30655837e-01 -9.54618871e-01 7.22308934e-01 -6.10190272e-01
-1.71319282e+00 9.09022450e-01 1.09602833e+00 -2.00793281e-01
1.57999992e+00 -4.43188250e-01 1.35407913e+00 3.16477865e-02
2.34706804e-01 -8.12789977e-01 3.05510104e-01 4.08322543e-01
7.07249701e-01 -1.19455802e+00 2.79020816e-01 -1.73294142e-01
-3.01062852e-01 1.18836546e+00 2.52355814e-01 -9.11391005e-02
6.19406939e-01 4.46680754e-01 -1.19603023e-01 1.57432929e-01
-1.05326676e+00 1.75822362e-01 2.59446770e-01 6.07281804e-01
5.52457273e-01 -4.23928171e-01 -2.77375847e-01 8.22420269e-02
3.84399056e-01 5.39348841e-01 6.40520811e-01 1.13484597e+00
-4.64946061e-01 -1.06865180e+00 -1.86571866e-01 2.48277187e-01
-9.11604047e-01 -3.82098764e-01 3.14191163e-01 3.57080907e-01
2.10133895e-01 1.34064972e+00 6.93410486e-02 -6.96066618e-01
3.23647782e-02 -5.90744972e-01 6.31573677e-01 -3.63462567e-02
-4.87119198e-01 2.35622749e-01 -1.66334078e-01 -1.16608560e+00
-9.87424850e-01 -4.11190957e-01 -1.16032577e+00 -5.47704756e-01
2.09876269e-01 -1.66065052e-01 6.13616526e-01 9.72675085e-01
6.10308468e-01 8.17799985e-01 5.91291428e-01 -1.26960516e+00
-3.70701760e-01 -6.35651827e-01 -3.52310836e-01 4.20111179e-01
8.35338414e-01 -5.27134836e-01 -3.16947371e-01 3.95467550e-01] | [11.362659454345703, -1.711758017539978] |
25bce792-2d15-476b-b3ea-89c40cc3a5d6 | optimizing-fiducial-marker-placement-for | 2211.01513 | null | https://arxiv.org/abs/2211.01513v2 | https://arxiv.org/pdf/2211.01513v2.pdf | Optimizing Fiducial Marker Placement for Improved Visual Localization | Adding fiducial markers to a scene is a well-known strategy for making visual localization algorithms more robust. Traditionally, these marker locations are selected by humans who are familiar with visual localization techniques. This paper explores the problem of automatic marker placement within a scene. Specifically, given a predetermined set of markers and a scene model, we compute optimized marker positions within the scene that can improve accuracy in visual localization. Our main contribution is a novel framework for modeling camera localizability that incorporates both natural scene features and artificial fiducial markers added to the scene. We present optimized marker placement (OMP), a greedy algorithm that is based on the camera localizability framework. We have also designed a simulation framework for testing marker placement algorithms on 3D models and images generated from synthetic scenes. We have evaluated OMP within this testbed and demonstrate an improvement in the localization rate by up to 20 percent on four different scenes. | ['John J. Leonard', 'Sudipta N. Sinha', 'Victor Fragoso', 'Joseph DeGol', 'Qiangqiang Huang'] | 2022-11-02 | null | null | null | null | ['visual-localization'] | ['computer-vision'] | [ 1.75529588e-02 -3.47050726e-01 -1.99892551e-01 -2.47061431e-01
-7.47483730e-01 -9.54621613e-01 4.54914898e-01 2.80856073e-01
-4.20862466e-01 4.04014677e-01 -2.16543674e-01 -3.67644131e-01
2.85116434e-01 -2.65089184e-01 -8.49191785e-01 -1.09746419e-01
-3.75086606e-01 1.28369927e-01 3.75652105e-01 4.32501733e-01
4.70873743e-01 1.12442148e+00 -9.94205773e-01 -6.13700688e-01
7.52622664e-01 7.17219532e-01 4.79796648e-01 1.03856516e+00
3.94211560e-01 6.72280550e-01 -7.03602195e-01 1.57412082e-01
4.36035067e-01 -2.34134585e-01 -3.27603608e-01 5.27868330e-01
6.44095480e-01 -1.40285805e-01 1.64969582e-02 1.01365662e+00
3.38624001e-01 2.73782253e-01 3.77671808e-01 -1.42074513e+00
-3.83986123e-02 1.58137500e-01 -9.14664805e-01 8.37887675e-02
7.41490960e-01 1.16680353e-03 3.12008083e-01 -6.78932607e-01
6.33881986e-01 8.26738954e-01 9.52713847e-01 1.84843287e-01
-1.27039802e+00 -1.52935967e-01 -6.33395761e-02 -4.54744518e-01
-1.88218510e+00 -4.21746880e-01 7.01358080e-01 -5.81623197e-01
5.41272700e-01 4.26110268e-01 3.84753644e-01 2.80178756e-01
2.62674510e-01 8.32164958e-02 1.02132368e+00 -1.01078224e+00
3.52739155e-01 4.89241391e-01 -1.24006635e-02 1.02564001e+00
4.80363697e-01 1.43372625e-01 -2.99438238e-01 -4.28563923e-01
1.23998928e+00 -2.80749172e-01 -4.80374426e-01 -1.20748186e+00
-1.53080750e+00 6.17733657e-01 6.23454988e-01 3.55980992e-02
-1.78531945e-01 6.92306757e-01 -5.36278300e-02 -5.34924090e-01
-9.98364855e-03 7.55660176e-01 1.85632184e-01 -1.39160275e-01
-9.62267518e-01 -7.23447353e-02 6.86163127e-01 1.31257832e+00
8.28731418e-01 3.82444300e-02 -1.85759626e-02 5.63679695e-01
4.40436363e-01 8.10774267e-01 8.60468447e-02 -1.31178212e+00
1.48013696e-01 2.37148240e-01 7.06535935e-01 -1.36406755e+00
-3.51554751e-01 -2.09250063e-01 -8.85463580e-02 5.03724575e-01
3.62527370e-01 -5.29216230e-01 -8.75637591e-01 1.56441450e+00
5.17789841e-01 4.60519761e-01 -1.40540943e-01 1.07715261e+00
1.47674292e-01 4.66587812e-01 -1.55728638e-01 5.19704558e-02
9.80141580e-01 -9.96459126e-01 -4.05053347e-01 -3.03283006e-01
5.86902976e-01 -1.15766549e+00 8.36703479e-01 7.14087412e-02
-9.06856894e-01 -5.07772446e-01 -1.31010306e+00 1.69695824e-01
-6.14275150e-02 5.12207508e-01 6.11959994e-01 1.00250256e+00
-1.38295174e+00 1.14693925e-01 -8.67488205e-01 -7.86035419e-01
-2.35435680e-01 4.75826859e-01 -3.07045430e-01 1.79877266e-01
-3.85482281e-01 1.09833312e+00 6.61617666e-02 -8.36460441e-02
-9.98846114e-01 -3.86720806e-01 -1.09395409e+00 -2.41884917e-01
2.02691928e-01 -5.33732057e-01 1.32326210e+00 -7.41800189e-01
-1.38353240e+00 8.20524931e-01 -4.38221097e-01 -4.80511159e-01
5.59542358e-01 -1.49189502e-01 -1.86448023e-01 3.07676405e-01
4.15566593e-01 6.47702038e-01 7.08635032e-01 -1.78549194e+00
-3.74168694e-01 6.75752759e-02 1.03910573e-01 4.73090619e-01
3.35001975e-01 1.79434538e-01 -8.19062769e-01 -4.18195993e-01
1.07638665e-01 -1.10819626e+00 -6.13945305e-01 2.08443716e-01
-5.69295704e-01 7.03966260e-01 5.63860953e-01 -3.71553093e-01
8.89604032e-01 -2.18742990e+00 -3.52377594e-01 6.40248299e-01
5.79288229e-02 -1.11351639e-01 1.34509981e-01 1.80384234e-01
-1.93631239e-02 7.50464797e-02 3.66832137e-01 -4.19183582e-01
-1.89859942e-01 -1.97528079e-01 -1.68019205e-01 8.71770978e-01
-5.52506030e-01 3.92743289e-01 -9.49709773e-01 -4.25276965e-01
7.87214339e-01 2.64548331e-01 -2.99339175e-01 1.69182971e-01
1.62413836e-01 2.18518332e-01 -3.19115728e-01 6.82218432e-01
8.47407103e-01 -3.44544768e-01 4.36843373e-02 -2.09693313e-02
-3.34588647e-01 -1.37107998e-01 -1.42043161e+00 1.58867705e+00
-4.63754684e-01 7.78413236e-01 1.82956941e-02 -2.53432896e-02
9.68878329e-01 -5.06652594e-02 5.99166930e-01 -1.29476711e-01
2.47998670e-01 6.82341009e-02 -3.97216886e-01 1.95013043e-02
9.37447190e-01 3.49414438e-01 -1.64372265e-01 3.22142929e-01
-3.63020599e-01 -3.99767697e-01 3.34202056e-03 2.35771939e-01
9.75386083e-01 9.76415630e-03 5.59270024e-01 -3.69938403e-01
3.25249165e-01 5.57426870e-01 2.41730452e-01 1.01805806e+00
-3.38391811e-01 8.07254136e-01 3.40405256e-02 -2.45683566e-01
-1.07469964e+00 -1.07409847e+00 -1.31922260e-01 4.77733910e-01
9.87008393e-01 -4.97241795e-01 -8.88647079e-01 -2.98520029e-01
3.86273339e-02 5.10543764e-01 -5.41863620e-01 1.87552437e-01
-2.56915092e-01 -2.04731032e-01 4.92714673e-01 4.56527114e-01
5.08203149e-01 -2.57026047e-01 -1.25899076e+00 -6.94338232e-02
1.99257657e-01 -1.30855274e+00 -7.20970571e-01 1.02992028e-01
-5.31943023e-01 -1.23938704e+00 -6.35710597e-01 -5.40916562e-01
1.20684755e+00 9.32274342e-01 7.36112654e-01 -2.83922758e-02
-2.68195659e-01 9.73444164e-01 -2.60782063e-01 -3.29080701e-01
-2.53835231e-01 -3.03531259e-01 1.05145276e-01 -1.83953926e-01
-1.04987703e-01 2.21487552e-01 -3.61995101e-01 6.16350770e-01
-2.95967281e-01 9.69055295e-02 1.52918190e-01 4.19763178e-01
6.85485423e-01 -3.41719121e-01 -4.87658918e-01 -5.54273129e-01
4.63349640e-01 -2.38838404e-01 -1.34685051e+00 3.44806194e-01
-1.37476593e-01 -1.48451403e-01 2.36987397e-01 -4.08460289e-01
-6.63000822e-01 4.52557504e-01 5.91836214e-01 -6.43392861e-01
-3.89770344e-02 3.03223073e-01 9.93962586e-02 -1.14899635e+00
9.16530728e-01 -2.67470181e-01 -1.63685530e-01 -6.22986956e-03
4.90338117e-01 4.12185431e-01 6.59958303e-01 -6.49304688e-01
8.46572638e-01 5.53011715e-01 1.54832363e-01 -8.28253090e-01
-2.60813475e-01 -6.29433692e-01 -6.58239007e-01 -3.89667213e-01
6.18697762e-01 -9.56636608e-01 -6.57823145e-01 1.54662222e-01
-1.25102484e+00 -5.50535440e-01 -5.13732582e-02 7.34537840e-01
-6.63074136e-01 5.08870244e-01 -1.59299478e-01 -7.02818811e-01
2.12322742e-01 -1.37181389e+00 1.11915433e+00 6.31966889e-01
-2.77865916e-01 -1.20525444e+00 1.98034540e-01 -1.69524655e-01
2.83855766e-01 7.03349292e-01 1.50743023e-01 -2.05367450e-02
-1.05643368e+00 -3.34948123e-01 -1.46010652e-01 -1.87573731e-01
1.89530000e-01 1.63171932e-01 -9.55478430e-01 -4.02956873e-01
-2.75951594e-01 2.79246658e-01 7.36077502e-02 5.42480648e-01
8.42807114e-01 5.30756228e-02 -8.02326083e-01 8.67536247e-01
1.76585674e+00 5.55427313e-01 3.37714911e-01 7.44331121e-01
6.30399168e-01 -2.66803294e-01 1.01610303e+00 5.05575478e-01
3.12571615e-01 1.03413892e+00 4.13094580e-01 -4.40990031e-01
-2.67260280e-02 -5.16537905e-01 1.32501185e-01 1.23736709e-01
1.21804230e-01 -3.39827687e-01 -1.07214200e+00 6.45241678e-01
-1.58726799e+00 -4.46489930e-01 -1.75160959e-01 2.52472520e+00
3.55890542e-01 -2.91915059e-01 -3.01867165e-02 -3.31219316e-01
1.05613506e+00 2.78277397e-02 -2.94599891e-01 -4.57184464e-01
2.95647711e-01 -3.16200048e-01 1.33868265e+00 1.11654854e+00
-1.24512434e+00 9.74087119e-01 7.15621758e+00 1.56225875e-01
-1.41873050e+00 -4.89429682e-02 2.59339362e-01 4.49436992e-01
5.51318638e-02 3.92868429e-01 -9.17506635e-01 4.44611728e-01
6.13406777e-01 -5.04649758e-01 4.70182747e-01 1.35349822e+00
3.13080966e-01 -8.33320379e-01 -9.69465971e-01 1.20524299e+00
4.76470321e-01 -1.39019024e+00 -4.46580917e-01 7.37519935e-02
7.84279048e-01 8.16773549e-02 7.92842582e-02 -4.79340851e-01
5.50786614e-01 -8.09340060e-01 8.83187950e-01 3.21570545e-01
8.80214989e-01 -5.39066494e-01 4.72709268e-01 1.65560529e-01
-1.12965071e+00 3.85684162e-01 -5.86439252e-01 1.52777269e-01
2.61593491e-01 2.46944726e-01 -1.58932400e+00 3.47579241e-01
3.88238579e-01 3.19745034e-01 -1.06067312e+00 2.05495095e+00
-2.62264460e-01 1.35146320e-01 -6.15228832e-01 -1.01122953e-01
1.69571653e-01 -1.79888960e-02 5.62576294e-01 1.08442485e+00
5.62704504e-01 -4.64068413e-01 4.21122372e-01 6.62938714e-01
2.26116806e-01 -7.25678727e-02 -1.18684816e+00 5.49338341e-01
8.14418137e-01 1.18075931e+00 -9.92162168e-01 -6.57862574e-02
-1.66803628e-01 9.69489992e-01 3.98030132e-02 5.57647228e-01
-1.19214106e+00 -5.93224823e-01 4.34974879e-01 3.06946546e-01
-5.31327389e-02 -7.87552059e-01 -3.65188271e-01 -1.04045713e+00
-1.79441392e-01 -4.49798077e-01 -4.55865972e-02 -1.24802721e+00
-3.32501769e-01 5.18501699e-01 2.16832668e-01 -1.52877533e+00
-3.01767111e-01 -4.57297564e-01 -3.55348438e-01 9.63090658e-01
-1.16857898e+00 -1.00613034e+00 -9.33681250e-01 5.73412478e-01
5.18585965e-02 1.72453210e-01 6.15680873e-01 1.17482513e-01
-1.60578534e-01 5.19482315e-01 1.30401462e-01 1.94380701e-01
7.69127488e-01 -1.24382579e+00 2.85069108e-01 1.48505783e+00
3.68142277e-01 1.01138794e+00 7.85867691e-01 -5.44808984e-01
-1.30132759e+00 -1.07950914e+00 4.86462325e-01 -6.62657857e-01
4.89577651e-01 -4.10682887e-01 -2.97324479e-01 1.11936235e+00
2.84676343e-01 3.20593983e-01 3.33479404e-01 -2.35135823e-01
1.46257356e-01 -8.97890627e-02 -1.24652791e+00 7.83093631e-01
5.60784757e-01 -4.51783597e-01 -4.78667356e-02 4.58093256e-01
2.89999902e-01 -1.22663426e+00 -5.44379711e-01 1.45178989e-01
2.00271040e-01 -8.71133983e-01 9.97202933e-01 2.82329738e-01
-5.66306055e-01 -1.11385369e+00 -2.03128189e-01 -1.47242558e+00
-1.33274555e-01 -9.83501554e-01 4.34620500e-01 1.04378641e+00
1.37300789e-01 -4.74859387e-01 7.75903463e-01 7.88656712e-01
8.42428878e-02 -7.59606659e-02 -6.21955633e-01 -7.59652436e-01
-6.81747198e-01 -3.58950615e-01 3.32127392e-01 7.40348816e-01
3.89019065e-02 1.48333181e-02 -4.37292665e-01 7.45282531e-01
6.83335483e-01 -5.01965620e-02 1.34742916e+00 -5.60539901e-01
-9.47435275e-02 -3.47237289e-02 -8.49202573e-01 -1.26177609e+00
-1.66924849e-01 -4.51366723e-01 3.71181697e-01 -1.43268561e+00
-1.60798198e-03 -1.03131378e+00 2.52469897e-01 2.43252695e-01
-1.03584938e-02 2.46671021e-01 4.38692957e-01 7.87942484e-03
-7.11382687e-01 -1.14220150e-01 7.24835813e-01 2.18950897e-01
-1.88381702e-01 -9.49281752e-02 -3.74061346e-01 8.85700941e-01
8.15623164e-01 -3.53783488e-01 -4.83678043e-01 -7.22640932e-01
-2.00963706e-01 1.71605974e-01 6.36463404e-01 -1.43235731e+00
5.64021528e-01 -2.52538413e-01 5.43636203e-01 -4.41848427e-01
4.43390310e-01 -1.13812757e+00 2.90821880e-01 3.46256226e-01
-4.80442159e-02 5.88933408e-01 4.80196893e-01 3.54729950e-01
-4.44752090e-02 -4.43680644e-01 9.99360621e-01 -4.18781824e-02
-1.13655543e+00 -1.00878708e-01 -3.06261510e-01 -2.31214747e-01
1.73418283e+00 -3.91099900e-01 -4.12649810e-01 -5.80979884e-01
-5.66094697e-01 2.15686187e-02 1.35803056e+00 4.87867296e-02
6.90529406e-01 -1.36145484e+00 8.35157409e-02 2.72334605e-01
2.03936145e-01 -2.85664022e-01 -3.88433605e-01 7.80008256e-01
-1.51878989e+00 4.70396012e-01 -3.04635540e-02 -1.02498972e+00
-1.45122552e+00 7.33564675e-01 7.00800121e-01 3.18361074e-01
-1.50759637e-01 6.58048809e-01 -2.78984644e-02 -1.71352789e-01
2.94089109e-01 -5.90996563e-01 4.15193439e-01 -7.59255111e-01
3.27175647e-01 2.93896943e-01 -2.45024949e-01 -8.08367729e-01
-6.28278971e-01 1.02872670e+00 3.71909052e-01 -4.15788502e-01
5.13488114e-01 -3.88448119e-01 8.39847922e-02 3.08260709e-01
8.70306492e-01 8.12125921e-01 -1.28258073e+00 2.19293788e-01
1.77302659e-01 -1.14513075e+00 -1.35760710e-01 -6.28038824e-01
-5.98428607e-01 5.27009547e-01 7.62636125e-01 -1.37993336e-01
6.22053981e-01 -2.02972293e-01 -5.97966947e-02 2.48821706e-01
1.08959556e+00 -7.02377379e-01 -7.13668950e-03 2.30019122e-01
5.10619879e-01 -1.04588866e+00 2.08207354e-01 -7.53139257e-01
-6.64838374e-01 9.27180350e-01 6.31278992e-01 -3.21996987e-01
3.10965598e-01 6.22534990e-01 4.70392764e-01 1.16760843e-01
6.27373829e-02 -8.08003694e-02 1.79658402e-02 8.06666434e-01
1.75641000e-01 1.44331098e-01 1.39747560e-01 -3.35122555e-01
1.33654729e-01 -4.57459223e-03 9.11150753e-01 1.03721404e+00
-5.20659924e-01 -8.25696945e-01 -9.93132949e-01 -2.44249195e-01
-1.90911725e-01 1.74968809e-01 -2.39559859e-01 1.10822237e+00
1.85274554e-03 1.00395429e+00 -1.36458173e-01 -1.66097745e-01
1.00215882e-01 -6.34828985e-01 6.33493841e-01 -7.28810430e-01
-2.08894163e-01 2.58708000e-01 1.17935181e-01 -6.71582341e-01
-3.27454776e-01 -5.99391997e-01 -1.27806461e+00 -1.89398378e-01
-4.33458060e-01 4.99540478e-01 9.07285929e-01 2.45249525e-01
5.55140316e-01 1.99958712e-01 7.25924134e-01 -1.02993786e+00
-7.68715218e-02 -3.77737910e-01 -6.15110099e-01 -7.89740980e-02
4.34338540e-01 -8.27806056e-01 -2.57354915e-01 2.07388401e-01] | [7.6637468338012695, -2.137322187423706] |
e8643cf2-5a8e-4ffb-ad52-741fff1b0519 | network-traffic-analysis-based-iot-device | 2009.04682 | null | https://arxiv.org/abs/2009.04682v1 | https://arxiv.org/pdf/2009.04682v1.pdf | Network Traffic Analysis based IoT Device Identification | Device identification is the process of identifying a device on Internet without using its assigned network or other credentials. The sharp rise of usage in Internet of Things (IoT) devices has imposed new challenges in device identification due to a wide variety of devices, protocols and control interfaces. In a network, conventional IoT devices identify each other by utilizing IP or MAC addresses, which are prone to spoofing. Moreover, IoT devices are low power devices with minimal embedded security solution. To mitigate the issue in IoT devices, fingerprint (DFP) for device identification can be used. DFP identifies a device by using implicit identifiers, such as network traffic (or packets), radio signal, which a device used for its communication over the network. These identifiers are closely related to the device hardware and software features. In this paper, we exploit TCP/IP packet header features to create a device fingerprint utilizing device originated network packets. We present a set of three metrics which separate some features from a packet which contribute actively for device identification. To evaluate our approach, we used publicly accessible two datasets. We observed the accuracy of device genre classification 99.37% and 83.35% of accuracy in the identification of an individual device from IoT Sentinel dataset. However, using UNSW dataset device type identification accuracy reached up to 97.78%. | ['Emeroylariffion Abas', 'Sandhya Aneja', 'Nagender Aneja', 'Rajarshi Roy Chowdhury'] | 2020-09-10 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 6.09066129e-01 -8.95905793e-02 -7.27535069e-01 -1.91337153e-01
-1.25121266e-01 -1.09803951e+00 4.78309840e-01 2.49901682e-01
-2.73417145e-01 6.96435630e-01 -7.35392496e-02 -5.05453229e-01
-2.90136456e-01 -9.01520908e-01 -1.21341564e-01 -5.14627099e-01
1.26161993e-01 2.46484220e-01 3.77098918e-01 5.17099023e-01
1.86351001e-01 7.28316784e-01 -1.45779574e+00 3.22486050e-02
1.06842212e-01 1.67021739e+00 -2.83552915e-01 4.86749858e-01
-3.94240648e-01 2.12692648e-01 -9.49416876e-01 -3.02014798e-01
3.14954549e-01 1.83717832e-01 -7.17922449e-01 -4.76182342e-01
-2.57430553e-01 -3.33343446e-01 2.14582980e-01 1.06020832e+00
5.17188072e-01 -5.71733057e-01 4.52253044e-01 -1.62196946e+00
-1.06763951e-01 7.32125163e-01 -4.35639590e-01 3.09251428e-01
8.47079992e-01 -8.01400095e-02 3.91938031e-01 -7.10045099e-02
3.60312223e-01 8.99705112e-01 9.88978207e-01 4.63067949e-01
-9.13801372e-01 -1.09199142e+00 -6.75418139e-01 -3.74941677e-01
-1.55759335e+00 -3.73193562e-01 9.49824274e-01 -5.88818789e-01
6.61882699e-01 3.62962306e-01 1.79808125e-01 1.33114076e+00
1.92789927e-01 -1.55954838e-01 8.80934536e-01 -2.29628384e-01
2.31686488e-01 5.81895709e-01 4.02947396e-01 1.56650707e-01
7.95967340e-01 2.04556376e-01 -4.46496457e-02 -6.58760905e-01
3.91231179e-01 4.56668317e-01 1.45013675e-01 4.35800403e-02
-1.21460295e+00 3.94240469e-01 -7.51125291e-02 6.01084590e-01
-3.84669930e-01 -6.83750361e-02 7.00222731e-01 1.47597209e-01
-2.28575155e-01 1.25910118e-01 -8.20749640e-01 -5.35423398e-01
-2.63282984e-01 -6.25861883e-01 8.83222997e-01 9.82680917e-01
5.82894027e-01 -3.36459547e-01 2.96461970e-01 4.57044482e-01
5.26118159e-01 8.01893115e-01 4.90155041e-01 -7.18065321e-01
3.63369256e-01 7.58373201e-01 1.94736943e-01 -1.22996473e+00
-5.48775136e-01 -4.66335006e-02 -1.00818181e+00 -1.69651777e-01
2.35838234e-01 -8.69935378e-02 -1.94399893e-01 1.60179806e+00
3.53597403e-01 3.60815972e-01 -1.65280342e-01 -4.40940708e-02
6.48741543e-01 2.30098858e-01 4.81491655e-01 -1.66667745e-01
1.71040404e+00 -2.34426139e-03 -7.02889264e-01 4.91710633e-01
3.77256870e-01 -8.36744249e-01 4.62471396e-01 1.72753334e-01
-5.80696762e-01 -7.46612549e-01 -1.08163095e+00 6.72983229e-01
-9.02725697e-01 -1.59873471e-01 4.52864200e-01 1.54959381e+00
-6.28489077e-01 4.19289380e-01 -4.52557951e-01 -8.25168848e-01
4.24316913e-01 1.09123445e+00 -1.79387346e-01 5.12550414e-01
-8.89643133e-01 1.08343050e-01 6.07759356e-01 -4.52863872e-01
-1.45709917e-01 -4.03354406e-01 -5.08075774e-01 -1.82749361e-01
-2.40141362e-01 -5.62176406e-01 8.44907761e-01 -6.87913537e-01
-1.59663630e+00 8.22129846e-01 8.42363834e-02 -2.64239371e-01
1.52238742e-01 3.47855985e-01 -1.47172093e+00 2.17445150e-01
2.55163431e-01 2.37807885e-01 1.00236678e+00 -8.27327728e-01
-7.00726748e-01 -3.64829570e-01 -7.33429343e-02 -8.90708447e-01
-5.83765149e-01 3.18240225e-01 1.32364094e-01 -3.48871648e-01
-3.79498303e-03 -1.09289682e+00 5.49630880e-01 -3.93802494e-01
-7.11679637e-01 -3.25177640e-01 1.42513406e+00 -2.53595054e-01
1.37590981e+00 -2.33382559e+00 -7.08269417e-01 5.11192262e-01
3.13372850e-01 6.52384222e-01 3.13027531e-01 3.60430896e-01
1.70904264e-01 8.15893352e-01 2.36376449e-01 -2.19586089e-01
5.92604391e-02 1.42806351e-01 -3.46948445e-01 3.15268487e-01
-1.87657446e-01 5.92740655e-01 -9.27896023e-01 -5.22541285e-01
5.45601547e-01 6.46690786e-01 -6.80393213e-03 -1.63619235e-01
4.34687465e-01 8.25049937e-01 -1.07810664e+00 8.75313461e-01
9.29696202e-01 -2.42108688e-01 2.64647245e-01 -5.54503143e-01
-2.22037300e-01 4.48299825e-01 -1.08885932e+00 1.21328044e+00
-5.19265354e-01 8.69028345e-02 -8.44545215e-02 -5.87519705e-01
8.59286904e-01 7.16884255e-01 8.15662742e-01 -4.43971008e-01
6.56541705e-01 4.81191278e-01 -4.15598929e-01 -5.22943497e-01
-1.50904641e-01 5.68136573e-01 -3.58061075e-01 5.81569254e-01
-2.74417818e-01 6.74474955e-01 -2.26695836e-01 -2.63208896e-01
1.67933762e+00 -4.13623601e-01 6.12109244e-01 -2.92043418e-01
1.04628062e+00 -4.96256143e-01 2.37798080e-01 9.28877592e-01
-5.48997998e-01 -2.69565303e-02 -1.50330484e-01 -5.88710904e-01
-5.59975743e-01 -1.21193397e+00 -4.84813154e-01 4.50797558e-01
2.06509098e-01 -3.01595420e-01 -3.41387868e-01 -8.85213852e-01
1.43416524e-01 -1.46066159e-01 -3.40541750e-01 4.29740846e-02
-5.77172995e-01 -3.24778289e-01 1.13847339e+00 3.55330884e-01
8.74600589e-01 -9.41872239e-01 -6.11738145e-01 3.04246664e-01
1.30438298e-01 -1.60935998e+00 -3.54519725e-01 -6.33247867e-02
-8.58587742e-01 -1.21267283e+00 -5.77143840e-02 -7.37017751e-01
4.30879891e-01 2.39096880e-01 1.02511120e+00 2.30477437e-01
-7.87706152e-02 4.75571066e-01 -2.62054056e-01 -4.05018389e-01
-5.62609017e-01 3.66590798e-01 5.08346796e-01 4.29568827e-01
8.58564556e-01 -7.66764700e-01 -6.35108590e-01 8.15075278e-01
-6.65923297e-01 -6.39197528e-01 2.26837412e-01 1.56477824e-01
2.80202776e-01 5.51365793e-01 6.38372481e-01 -5.85465908e-01
4.57787514e-01 -8.03227484e-01 -5.39187253e-01 -4.99059558e-02
-4.53916460e-01 -9.86945704e-02 6.97218597e-01 -7.09667802e-01
-3.20478916e-01 1.95641518e-01 -1.91605046e-01 1.83278024e-01
-6.45580769e-01 -1.25374720e-01 -5.77493846e-01 -5.09442031e-01
1.60855353e-01 -2.80379146e-01 -4.42243606e-01 -7.77734041e-01
-4.24368203e-01 1.41224682e+00 3.82741630e-01 -4.43447202e-01
9.03963566e-01 6.47587478e-01 1.43253580e-01 -9.19539690e-01
-3.42119366e-01 -7.80293167e-01 -4.19360042e-01 1.67327076e-02
9.18181956e-01 -5.02172112e-01 -1.74882007e+00 5.51314771e-01
-1.27080703e+00 4.54340935e-01 -1.07686684e-01 4.86408383e-01
2.13592395e-01 4.08267409e-01 -4.27091956e-01 -8.59695673e-01
-5.17575264e-01 -1.06305206e+00 1.20222485e+00 2.52687901e-01
-6.91948771e-01 -9.83184934e-01 -1.22312661e-02 1.24132730e-01
8.65562081e-01 5.37342072e-01 7.25413859e-01 -6.90118611e-01
-2.45104536e-01 -6.82515740e-01 -5.04269719e-01 3.66273783e-02
8.18939567e-01 -1.09336436e-01 -1.06331944e+00 -1.78471476e-01
2.32171416e-01 3.73809427e-01 -1.02330402e-01 9.42200124e-02
1.26922452e+00 -3.04707050e-01 -8.09702575e-01 7.76711106e-01
1.51868618e+00 6.46444082e-01 5.18869281e-01 3.81939113e-01
7.63518393e-01 3.19072783e-01 4.80528511e-02 5.66448152e-01
7.18812495e-02 8.75247240e-01 4.02514130e-01 3.03530186e-01
1.09628491e-01 -3.28451335e-01 1.30606636e-01 3.53164852e-01
5.25837801e-02 -4.42105472e-01 -8.18637133e-01 7.22999275e-02
-1.17532623e+00 -9.09198165e-01 -2.87853539e-01 2.45955777e+00
2.11622432e-01 3.68665218e-01 4.13051128e-01 7.67623007e-01
1.15910351e+00 -1.72641233e-01 -4.90211934e-01 -4.37996954e-01
9.74452496e-02 4.51744527e-01 1.13649964e+00 1.77715525e-01
-1.10622478e+00 1.59016237e-01 5.80065918e+00 5.48735440e-01
-1.12739944e+00 4.13657010e-01 1.94486693e-01 6.27623260e-01
7.15803728e-02 -3.57992083e-01 -9.21583593e-01 1.17925084e+00
1.18006837e+00 3.07703227e-01 3.15835327e-01 6.40766919e-01
2.68673927e-01 1.31384745e-01 -8.64089072e-01 1.48924363e+00
-4.76480871e-01 -1.08566439e+00 -1.76412687e-01 5.35842299e-01
3.54598343e-01 -1.14744283e-01 1.91964749e-02 -1.95987180e-01
-6.84087127e-02 -5.68643451e-01 2.94572413e-01 -2.71888189e-02
1.01216984e+00 -5.76716244e-01 9.80892539e-01 -1.32849559e-01
-1.81188989e+00 -2.25244045e-01 9.70775411e-02 -1.23838872e-01
2.44035676e-01 5.95038593e-01 -7.54657269e-01 3.13019931e-01
7.24772394e-01 5.37285507e-01 -4.02486920e-01 7.77792215e-01
9.75842476e-02 6.16877913e-01 -8.13788474e-01 -1.35071769e-01
-3.57074648e-01 1.97498485e-01 3.35342973e-01 8.62472177e-01
4.52877939e-01 -2.92241186e-01 -6.58469796e-02 5.25216162e-01
-2.05835119e-01 -1.85032010e-01 -9.38748479e-01 -1.16805978e-01
1.04471231e+00 1.12730312e+00 -9.02099729e-01 -1.11459643e-01
-4.23862040e-01 8.37121606e-01 -6.59391046e-01 4.54161651e-02
-8.19847107e-01 -6.00505412e-01 9.31613147e-01 1.56746000e-01
6.67349696e-02 -2.07742125e-01 -1.93623558e-01 -6.88542366e-01
-7.36370161e-02 -4.60063219e-01 1.66082576e-01 -1.23561256e-01
-1.38285363e+00 5.64630151e-01 -1.26062497e-01 -1.53459907e+00
-1.50243774e-01 -5.86988270e-01 -4.75867510e-01 6.10594988e-01
-1.14418960e+00 -1.05119765e+00 -5.44896305e-01 8.47034633e-01
-1.87303841e-01 -1.23910718e-01 1.11446011e+00 8.40118885e-01
-4.99275833e-01 7.07512736e-01 2.75043435e-02 4.07935858e-01
2.96210140e-01 -7.36120343e-01 5.57198226e-01 3.76878709e-01
-2.84817129e-01 9.58971322e-01 3.10826540e-01 -7.06892610e-01
-1.66353059e+00 -7.91020095e-01 9.37520862e-01 -6.51862502e-01
6.50943577e-01 -3.51403892e-01 -1.30006790e-01 5.97103059e-01
-1.84055075e-01 2.15756521e-01 1.01280844e+00 -4.27236736e-01
-4.01092112e-01 -4.35691327e-01 -1.91816723e+00 -8.18049610e-02
1.01104414e+00 -8.64472210e-01 4.19911603e-03 -6.50897026e-02
5.96920192e-01 2.89167315e-01 -1.20498264e+00 4.99606937e-01
1.10476661e+00 -8.90497565e-01 1.16314769e+00 2.18980554e-02
-6.81350470e-01 -5.30269563e-01 -3.70808929e-01 -1.44099131e-01
3.39920213e-03 -9.91545737e-01 -2.53668606e-01 2.05330849e+00
1.51405737e-01 -1.24705613e+00 9.00736570e-01 5.28657496e-01
6.65692985e-01 3.18451151e-02 -8.99854541e-01 -1.08740306e+00
-7.30078340e-01 -5.27149081e-01 1.38719285e+00 9.78363395e-01
-1.55435190e-01 2.68142283e-01 -2.78547972e-01 4.63862926e-01
7.67505527e-01 -6.22793078e-01 6.26489162e-01 -1.98297536e+00
-2.16607794e-01 -2.14432791e-01 -8.99055898e-01 -7.67118514e-01
-1.43916979e-01 -4.67381060e-01 -7.99032569e-01 -8.04127634e-01
-3.04764718e-01 -1.00846684e+00 -6.24695718e-01 1.66437194e-01
8.38882685e-01 6.95991158e-01 -1.99061751e-01 2.89896697e-01
-3.85832012e-01 -5.46032906e-01 2.85547704e-01 -4.13895965e-01
-3.33355814e-01 5.70975602e-01 -2.61485219e-01 7.21170843e-01
1.39452302e+00 -8.10132623e-01 -4.01139349e-01 -2.75872648e-01
2.62428731e-01 -1.48033038e-01 3.35652560e-01 -1.33141446e+00
-1.95225812e-02 3.68707746e-01 1.65091634e-01 -3.70286196e-01
2.10514873e-01 -1.65944457e+00 7.72887290e-01 9.04871166e-01
2.89756715e-01 2.96281457e-01 -1.24985371e-02 4.99949008e-01
1.13896102e-01 -2.56052643e-01 4.95426387e-01 2.21862227e-01
-3.58253300e-01 3.13526481e-01 -3.68979663e-01 -4.41353083e-01
1.03174496e+00 -5.89755118e-01 -5.16900957e-01 1.91775501e-01
-3.22747022e-01 -4.89897549e-01 8.35528553e-01 4.72230673e-01
1.37417927e-01 -1.25333083e+00 3.54270190e-01 5.16526639e-01
1.97755620e-01 -7.55827188e-01 -2.43494049e-01 6.20377123e-01
-4.51570153e-01 6.11244380e-01 -2.85032809e-01 -6.63918495e-01
-1.41141164e+00 7.98879683e-01 5.83272837e-02 -1.11007921e-01
-8.62978101e-02 2.39458993e-01 -5.17647088e-01 -1.53456718e-01
4.36003089e-01 -4.55359995e-01 -2.60507464e-01 -3.33874710e-02
8.58820975e-01 7.63493359e-01 1.42976955e-01 -8.97157609e-01
-9.28367794e-01 1.26514447e+00 4.14891422e-01 2.28320196e-01
6.47907913e-01 -4.01906580e-01 -1.34389549e-01 8.97093192e-02
1.76383090e+00 4.01894897e-01 -3.68427813e-01 1.99544936e-01
3.23827356e-01 -3.20924431e-01 -3.06416631e-01 -5.56484282e-01
-1.01765013e+00 4.21436429e-01 1.21282470e+00 1.27551413e+00
9.96217191e-01 -8.29540938e-02 1.24541712e+00 -2.22055651e-02
1.21844089e+00 -4.33430731e-01 -4.23009694e-01 5.23593053e-02
-1.92162320e-01 -1.29220176e+00 -3.56273979e-01 -6.74603343e-01
3.23927850e-01 1.11765921e+00 -1.25007689e-01 3.07087481e-01
1.25755572e+00 4.92665917e-01 -9.51255411e-02 -1.48685172e-01
3.08868021e-01 5.24485298e-02 -2.37463698e-01 1.21717942e+00
7.29584843e-02 2.22919375e-01 -2.76829153e-01 4.67779160e-01
-1.31602630e-01 2.54744798e-01 3.80315781e-02 1.08809471e+00
-1.82933420e-01 -1.75960112e+00 -2.95212835e-01 4.09654945e-01
-1.09902680e+00 3.18439901e-01 -2.71178454e-01 4.04994637e-01
5.57292938e-01 1.62567735e+00 -1.36384806e-02 -9.55359876e-01
9.76271778e-02 -5.62256016e-02 9.20261294e-02 -2.40911931e-01
-8.37757528e-01 -8.01491320e-01 3.96935008e-02 -5.74717224e-01
-8.02743077e-01 -3.20032418e-01 -1.04266953e+00 -7.56601453e-01
-1.50061607e-01 3.86770852e-02 1.13014710e+00 9.33034241e-01
6.43201888e-01 1.66391924e-01 9.02365029e-01 -6.36117399e-01
-1.94970444e-02 -5.07858753e-01 -5.97757339e-01 4.58988339e-01
4.87649560e-01 -7.66711235e-01 -6.92277372e-01 -1.03681847e-01] | [5.18638277053833, 7.133514881134033] |
22c61720-c31d-4d49-be11-ba71cef036d6 | data-augmentation-for-leaf-segmentation-and | 1903.08583 | null | http://arxiv.org/abs/1903.08583v1 | http://arxiv.org/pdf/1903.08583v1.pdf | Data Augmentation for Leaf Segmentation and Counting Tasks in Rosette Plants | Deep learning techniques involving image processing and data analysis are
constantly evolving. Many domains adapt these techniques for object
segmentation, instantiation and classification. Recently, agricultural
industries adopted those techniques in order to bring automation to farmers
around the globe. One analysis procedure required for automatic visual
inspection in this domain is leaf count and segmentation. Collecting labeled
data from field crops and greenhouses is a complicated task due to the large
variety of crops, growth seasons, climate changes, phenotype diversity, and
more, especially when specific learning tasks require a large amount of labeled
data for training. Data augmentation for training deep neural networks is well
established, examples include data synthesis, using generative semi-synthetic
models, and applying various kinds of transformations. In this paper we propose
a method that preserves the geometric structure of the data objects, thus
keeping the physical appearance of the data-set as close as possible to imaged
plants in real agricultural scenes. The proposed method provides state of the
art results when applied to the standard benchmark in the field, namely, the
ongoing Leaf Segmentation Challenge hosted by Computer Vision Problems in Plant
Phenotyping. | ['Alon Zvirin', 'Yaron Honen', 'Ron Kimmel', 'Dmitry Kuznichov'] | 2019-03-20 | null | null | null | null | ['plant-phenotyping'] | ['computer-vision'] | [ 5.69762170e-01 -5.15739806e-03 -4.36469018e-02 -3.13733160e-01
6.53492212e-02 -9.20040011e-01 7.84679428e-02 5.53526461e-01
-4.98317704e-02 4.19415027e-01 -6.32389784e-01 -3.06129187e-01
-1.23229600e-01 -1.17413008e+00 -6.61009133e-01 -7.45990098e-01
1.04746707e-01 6.78464115e-01 1.99249133e-01 -1.45185769e-01
-1.64024740e-01 9.74875271e-01 -1.64332449e+00 2.51342151e-02
9.72729146e-01 1.03870296e+00 6.61603510e-01 7.84726441e-01
-4.40504283e-01 2.24268988e-01 -4.96468544e-01 -9.88786761e-03
3.33948255e-01 -4.63171989e-01 -6.88591123e-01 8.83658946e-01
1.61268160e-01 -1.60878956e-01 3.64822447e-01 1.20404541e+00
1.89431682e-01 -1.47384137e-01 6.33653224e-01 -1.21101534e+00
-7.51145661e-01 5.81348479e-01 -8.83186996e-01 -3.14087331e-01
-4.19424266e-01 5.10678850e-02 5.96678495e-01 -3.26503187e-01
4.68729436e-01 1.05937517e+00 4.78194684e-01 1.70011654e-01
-1.47810924e+00 -1.77216634e-01 5.39330766e-02 1.65764675e-01
-9.98241901e-01 6.44956343e-03 7.28310168e-01 -5.07039607e-01
1.20662533e-01 2.77778208e-01 7.29704380e-01 5.39227664e-01
-2.01796070e-02 7.79274523e-01 8.82842302e-01 -5.31560421e-01
4.53458965e-01 8.89179260e-02 2.40428507e-01 5.61802626e-01
4.77633744e-01 -1.04624733e-01 2.77761102e-01 2.03867659e-01
5.57493985e-01 4.40140702e-02 -2.28499100e-01 -7.73567080e-01
-7.48509586e-01 7.07684338e-01 5.90496778e-01 3.55377197e-01
-6.00545883e-01 -2.59176344e-01 3.28914434e-01 -1.77689397e-03
3.85541707e-01 3.21388215e-01 -8.25469255e-01 5.65213561e-01
-9.12616611e-01 1.83615312e-01 8.49205017e-01 1.06989491e+00
7.49017656e-01 1.90368861e-01 -2.44747847e-02 7.04016387e-01
2.67757207e-01 5.57812572e-01 2.26197600e-01 -1.00947094e+00
-1.84810981e-01 1.03087020e+00 -2.66095046e-02 -9.12300229e-01
-5.49846768e-01 -2.85470814e-01 -1.21252692e+00 5.40065885e-01
6.48428738e-01 -7.72532895e-02 -1.20704710e+00 1.51269579e+00
4.54910457e-01 -4.04007673e-01 4.35732827e-02 5.64375937e-01
9.11343157e-01 6.96701050e-01 2.76904464e-01 -1.58237040e-01
1.35323012e+00 -4.64655340e-01 -6.06741369e-01 -4.26356494e-01
5.48602521e-01 -7.97511578e-01 9.27069902e-01 5.27671337e-01
-7.98494816e-01 -7.70290136e-01 -9.70374346e-01 1.18648633e-01
-7.27806032e-01 7.71738112e-01 7.53482163e-01 6.56506956e-01
-6.88930988e-01 6.38914883e-01 -8.02234769e-01 -6.99381948e-01
8.64096582e-01 3.42992008e-01 -3.85256171e-01 -4.69965041e-02
-4.64273959e-01 5.71927249e-01 8.75224173e-01 6.31394684e-01
-8.34067166e-01 -5.95211029e-01 -6.63527310e-01 2.31696337e-01
5.21345198e-01 -2.07825795e-01 9.87429798e-01 -1.09978342e+00
-1.51789069e+00 1.28955543e+00 3.98888975e-01 -4.81467009e-01
2.89997965e-01 -1.17985994e-01 1.50392670e-03 -2.98453122e-01
-1.31819814e-01 9.25892532e-01 8.71390998e-01 -1.39856267e+00
-6.30421817e-01 -8.21428657e-01 -2.06519678e-01 -1.64458096e-01
-2.38877788e-01 -1.63600907e-01 -5.79829663e-02 -3.25122774e-01
3.25587809e-01 -1.03867185e+00 -4.32918489e-01 3.83252114e-01
-5.03220260e-01 2.02126727e-01 1.47239387e+00 -8.61455381e-01
3.43227416e-01 -1.99156666e+00 1.87644184e-01 5.01669832e-02
5.95953539e-02 8.00407767e-01 -2.18191206e-01 -6.64536580e-02
-1.20869301e-01 -1.63282230e-02 -7.62070060e-01 7.07032159e-02
-2.33553112e-01 3.95490795e-01 -3.67232822e-02 3.45090836e-01
5.16560614e-01 8.38298380e-01 -4.85993832e-01 -3.39756578e-01
4.86201525e-01 2.34422743e-01 -1.20748222e-01 4.06542003e-01
-8.32590580e-01 5.30696690e-01 -3.66042644e-01 8.76754642e-01
1.13021827e+00 -5.83708547e-02 2.49657035e-01 -1.62712842e-01
-2.95954317e-01 -6.32966757e-01 -1.01126969e+00 1.38418067e+00
-1.53208852e-01 7.41699874e-01 4.24303740e-01 -1.42206466e+00
1.25839913e+00 1.26750693e-02 4.30910647e-01 -2.26906702e-01
4.30993110e-01 6.78967759e-02 1.86849341e-01 -4.56355840e-01
1.76897034e-01 6.19878292e-01 4.52125192e-01 2.75147539e-02
1.08289242e-01 -6.00047052e-01 7.55137682e-01 -2.46029764e-01
5.80200076e-01 5.58727503e-01 2.59306729e-01 -3.55907887e-01
3.64658147e-01 4.74713147e-01 6.72764480e-01 3.07374835e-01
-1.48128852e-01 5.29668093e-01 5.79509556e-01 -4.23779249e-01
-1.16026044e+00 -5.54488957e-01 -7.82866329e-02 9.17862952e-01
-6.70041740e-02 3.29462796e-01 -1.07596505e+00 -5.48229992e-01
-8.90032426e-02 5.95788300e-01 -5.84797978e-01 -2.18842313e-01
-3.31820726e-01 -1.24743712e+00 2.29955748e-01 5.19927204e-01
8.38425696e-01 -1.66469097e+00 -1.14325809e+00 2.58508146e-01
1.41323864e-01 -1.37189662e+00 2.57494569e-01 6.62890851e-01
-9.72762346e-01 -1.08559823e+00 -7.35497296e-01 -1.00028050e+00
7.33108044e-01 1.88256398e-01 1.11267066e+00 -1.66623611e-02
-9.44669366e-01 -2.53224909e-01 -5.44368207e-01 -8.90970588e-01
-7.19825029e-01 3.66739511e-01 -6.44397855e-01 -2.02906281e-02
2.09533513e-01 -4.16035831e-01 -2.78053164e-01 7.00631961e-02
-1.21626973e+00 1.11547455e-01 7.63215661e-01 7.17441857e-01
8.33066523e-01 2.19330877e-01 2.16299593e-01 -9.90456343e-01
1.79699019e-01 -1.28949896e-01 -1.33465183e+00 6.72250509e-01
-1.74396247e-01 -2.16731191e-01 6.18270218e-01 -3.42881948e-01
-9.67645705e-01 6.09984696e-01 1.53568059e-01 1.85497940e-01
-9.47590888e-01 4.72296208e-01 -8.03726912e-01 -1.89389646e-01
7.49105871e-01 7.73047879e-02 1.51656881e-01 -4.21492606e-01
5.04687428e-01 3.24609399e-01 6.42184913e-01 -1.34251714e-01
8.80893588e-01 4.40537691e-01 3.59098673e-01 -1.27380121e+00
-7.84598589e-01 -1.76130652e-01 -1.15694237e+00 -6.65941089e-02
9.17902112e-01 -2.47905895e-01 -6.55556142e-01 1.02911472e+00
-1.20082772e+00 -5.58677554e-01 -5.10151625e-01 2.58241147e-01
-3.98255497e-01 3.18791777e-01 -1.86320305e-01 -5.12097657e-01
-5.37333369e-01 -1.12212610e+00 9.04205322e-01 7.80402720e-01
2.22615555e-01 -8.51314962e-01 -2.25567684e-01 -3.02887298e-02
3.25193942e-01 6.35634959e-01 1.04236221e+00 -3.00601929e-01
-4.17920709e-01 -3.68405402e-01 -7.48644531e-01 6.74797118e-01
4.81722623e-01 6.73108578e-01 -1.00313771e+00 4.12292080e-03
-5.74162193e-02 -4.25482512e-01 6.30022705e-01 9.60855842e-01
1.49723911e+00 2.05233455e-01 -2.68040657e-01 7.26513743e-01
1.29215467e+00 4.62755054e-01 4.81458247e-01 1.04632415e-01
7.97286332e-01 1.05549920e+00 7.62914717e-01 4.12740380e-01
-8.61008465e-02 3.35871935e-01 8.15627992e-01 -6.92824602e-01
-3.05595878e-03 3.93539488e-01 -2.72584319e-01 5.70015647e-02
6.14218898e-02 -5.87820172e-01 -9.05921042e-01 4.69571292e-01
-1.74857867e+00 -6.21342719e-01 -4.81943488e-01 1.98826313e+00
5.89521170e-01 -2.90153525e-03 -1.00440554e-01 4.95446324e-01
6.96546793e-01 -1.82242975e-01 -8.30577910e-01 -2.70605206e-01
-4.58949149e-01 2.28213876e-01 9.09398258e-01 4.79856879e-02
-1.41990554e+00 1.16310573e+00 5.38843298e+00 4.01220828e-01
-1.17044008e+00 -2.70503163e-01 1.01858807e+00 8.27671528e-01
2.51514673e-01 6.91523030e-02 -6.77911103e-01 -7.50893503e-02
4.12343144e-01 2.44452730e-01 3.09337944e-01 8.58079493e-01
1.74041286e-01 -3.51280540e-01 -7.87583828e-01 5.25703013e-01
-1.47039890e-01 -9.58071649e-01 -1.10003471e-01 -3.99177410e-02
6.08296871e-01 -3.41879845e-01 -6.79284185e-02 -1.54979736e-01
4.47455585e-01 -7.64523387e-01 3.96028072e-01 1.65552959e-01
4.04844135e-01 -4.85891491e-01 6.59379482e-01 3.85407478e-01
-1.17427754e+00 -1.32037461e-01 -6.49245560e-01 3.74721944e-01
-1.75057799e-01 9.08120632e-01 -7.29524612e-01 4.47203875e-01
7.56865203e-01 4.59618032e-01 -9.76193249e-01 1.21867311e+00
-9.70368385e-02 6.68172479e-01 -4.86576766e-01 1.83128059e-01
1.86978206e-01 -5.99713981e-01 2.04451010e-01 7.79900610e-01
2.45122164e-01 -2.10399225e-01 3.68403971e-01 1.12295008e+00
4.82327864e-03 1.04019269e-01 -5.66228330e-01 -4.64754134e-01
-1.15751252e-02 1.65407252e+00 -1.40909517e+00 2.83881817e-02
-1.11396154e-02 1.01363659e+00 -7.45959878e-02 1.53290078e-01
-5.36761820e-01 -3.45174581e-01 1.32958755e-01 -5.84661029e-02
2.24251688e-01 -3.99296016e-01 -3.03571135e-01 -4.08051103e-01
-1.18470728e-01 -8.09967101e-01 -2.31095441e-02 -6.12165868e-01
-9.39133167e-01 3.52173418e-01 -1.68592371e-02 -8.07452738e-01
-7.76837170e-02 -1.04515100e+00 -4.88933563e-01 6.05734527e-01
-1.38828766e+00 -1.48544455e+00 -9.94384110e-01 7.94726238e-02
6.81439459e-01 -2.09894851e-01 1.10934782e+00 1.62025794e-01
-6.17240489e-01 -1.83830764e-02 2.38731548e-01 1.30086884e-01
2.70556301e-01 -1.19823742e+00 5.54898679e-01 9.87052202e-01
2.75767773e-01 -3.64465207e-01 4.75495368e-01 -4.14535105e-01
-1.20546031e+00 -1.29995358e+00 4.69888121e-01 1.39554188e-01
2.34575972e-01 -3.62720877e-01 -1.06535399e+00 5.11873066e-01
1.43289834e-01 2.09707528e-01 2.97612131e-01 -3.71161073e-01
2.98122406e-01 -3.38898450e-01 -1.22075963e+00 3.51161748e-01
5.36744356e-01 2.06821442e-01 1.63317263e-01 6.18463159e-01
5.06554067e-01 -3.54759425e-01 -4.81111318e-01 7.19601810e-01
3.34174901e-01 -7.63534129e-01 7.39555836e-01 -4.99080598e-01
1.88628554e-01 -3.78228724e-01 -1.10987291e-01 -1.36268306e+00
-3.60326767e-01 -4.44795996e-01 2.42065534e-01 1.45034277e+00
2.41941154e-01 -1.87319845e-01 1.00409138e+00 3.46315384e-01
-2.11102478e-02 -2.02874795e-01 -6.68109730e-02 -4.86177951e-01
-7.57396221e-02 9.38071162e-02 6.44038498e-01 7.45972514e-01
-1.12993908e+00 -8.27503055e-02 6.47071097e-03 3.98773968e-01
6.68314457e-01 4.35151845e-01 7.56506443e-01 -1.81748962e+00
-6.08049855e-02 -4.27766621e-01 -4.85774577e-01 -3.30011994e-01
1.21302307e-01 -6.26365423e-01 2.43966475e-01 -1.55124032e+00
-4.10853401e-02 -4.13591236e-01 3.64759505e-01 5.08826375e-01
-1.58033863e-01 5.15365787e-02 1.52566984e-01 -3.41930449e-01
1.93362370e-01 2.80248642e-01 1.27702391e+00 -4.81005847e-01
-5.05883753e-01 4.64817882e-01 -4.60834473e-01 8.67544413e-01
1.15503061e+00 -3.30493659e-01 -5.10177612e-01 -6.46755278e-01
-1.49323598e-01 -4.04704332e-01 4.67441946e-01 -1.15677154e+00
-2.33355403e-01 -3.42218786e-01 6.91042840e-01 -7.82283366e-01
1.62095040e-01 -1.11604702e+00 1.93295062e-01 6.63280487e-01
-2.19831288e-01 -1.13178693e-01 4.37822074e-01 1.55084312e-01
-3.42851356e-02 -6.88114107e-01 1.15606177e+00 -2.70182848e-01
-8.93800795e-01 4.21755463e-01 -3.53241771e-01 -7.93205574e-02
1.33090901e+00 -2.22535223e-01 -1.96236536e-01 -1.65303592e-02
-8.39942992e-01 1.28083721e-01 1.97460964e-01 4.12006378e-01
2.70312250e-01 -7.59483337e-01 -8.65636468e-01 3.76527727e-01
1.26045898e-01 3.93980145e-01 1.15570456e-01 4.57154006e-01
-1.16505694e+00 1.38060957e-01 -7.88127005e-01 -7.64305532e-01
-1.35832191e+00 4.85663295e-01 1.94732383e-01 -2.47427016e-01
-4.14135247e-01 7.26896286e-01 3.03692609e-01 -5.60757935e-01
2.88220406e-01 -8.24361086e-01 -4.41123784e-01 1.47884563e-01
2.11165965e-01 2.05891043e-01 2.72314340e-01 -2.88974702e-01
7.81069174e-02 4.35803086e-01 1.66787609e-01 4.52143520e-01
1.45491278e+00 2.07945734e-01 -2.27645427e-01 3.81809443e-01
5.55799901e-01 -6.47419810e-01 -1.17763186e+00 -9.72475335e-02
2.71814853e-01 -2.68012881e-01 1.33622557e-01 -6.53470635e-01
-1.46156585e+00 1.04115927e+00 1.12896657e+00 7.79300570e-01
1.42428315e+00 -3.04202825e-01 3.93445969e-01 6.53741539e-01
8.17265734e-03 -1.10952735e+00 -4.90623057e-01 2.76303798e-01
8.49538565e-01 -1.56476986e+00 -1.77719429e-01 -6.67024493e-01
-4.80103225e-01 1.14847469e+00 6.40100718e-01 -4.80774157e-02
7.36316144e-01 4.77988511e-01 1.94972739e-01 -2.31095054e-03
-1.85777530e-01 -5.68186700e-01 4.62949090e-02 1.09506559e+00
5.61366320e-01 2.43930474e-01 -9.58541930e-02 1.42414659e-01
1.47353068e-01 -3.26952040e-02 3.95621032e-01 1.06845081e+00
-4.79452729e-01 -1.27855623e+00 -5.02721488e-01 4.22740370e-01
-1.75505504e-01 2.70593524e-01 -7.31811404e-01 7.28306711e-01
3.27643365e-01 7.29965568e-01 2.19105035e-02 2.93752253e-01
3.57078910e-01 -1.32357970e-01 5.66740036e-01 -7.22490728e-01
-3.43832046e-01 -2.97031384e-02 -4.52820897e-01 -7.31327236e-02
-4.58520085e-01 -5.82018256e-01 -1.07939291e+00 -3.60469520e-02
-4.63009655e-01 -3.90081137e-01 1.07601690e+00 8.33480060e-01
1.69446021e-01 8.19615543e-01 6.24690711e-01 -9.03113246e-01
-2.79696494e-01 -9.25908804e-01 -7.63883889e-01 3.25617552e-01
-3.84210162e-02 -4.71541375e-01 2.59908319e-01 6.98021472e-01] | [9.139488220214844, -1.5344096422195435] |
ae9631d5-c954-4d71-92c5-03435fe977cc | image-blind-denoising-using-dual | 2304.01620 | null | https://arxiv.org/abs/2304.01620v1 | https://arxiv.org/pdf/2304.01620v1.pdf | Image Blind Denoising Using Dual Convolutional Neural Network with Skip Connection | In recent years, deep convolutional neural networks have shown fascinating performance in the field of image denoising. However, deeper network architectures are often accompanied with large numbers of model parameters, leading to high training cost and long inference time, which limits their application in practical denoising tasks. In this paper, we propose a novel dual convolutional blind denoising network with skip connection (DCBDNet), which is able to achieve a desirable balance between the denoising effect and network complexity. The proposed DCBDNet consists of a noise estimation network and a dual convolutional neural network (CNN). The noise estimation network is used to estimate the noise level map, which improves the flexibility of the proposed model. The dual CNN contains two branches: a u-shaped sub-network is designed for the upper branch, and the lower branch is composed of the dilated convolution layers. Skip connections between layers are utilized in both the upper and lower branches. The proposed DCBDNet was evaluated on several synthetic and real-world image denoising benchmark datasets. Experimental results have demonstrated that the proposed DCBDNet can effectively remove gaussian noise in a wide range of levels, spatially variant noise and real noise. With a simple model structure, our proposed DCBDNet still can obtain competitive denoising performance compared to the state-of-the-art image denoising models containing complex architectures. Namely, a favorable trade-off between denoising performance and model complexity is achieved. Codes are available at https://github.com/WenCongWu/DCBDNet. | ['Yungang Zhang', 'Peng Liang', 'Guannan Lv', 'Shicheng Liao', 'Wencong Wu'] | 2023-04-04 | null | null | null | null | ['noise-estimation'] | ['medical'] | [-2.73146462e-02 -4.60880965e-01 5.57020128e-01 -2.36336976e-01
-3.37605715e-01 -8.67557451e-02 3.62944752e-01 -1.56118363e-01
-6.95801139e-01 4.23684537e-01 -5.88951930e-02 -1.76416442e-01
9.81867611e-02 -9.05166388e-01 -4.93135959e-01 -1.24969077e+00
2.46208459e-01 -4.44355339e-01 5.73226929e-01 -3.06560785e-01
-9.08663049e-02 3.39618832e-01 -1.28432941e+00 -1.23765925e-02
1.07949317e+00 1.49716389e+00 5.94893217e-01 3.47681135e-01
-1.33535862e-01 5.19118309e-01 -5.49296677e-01 -4.32673126e-01
3.05226237e-01 -3.73344004e-01 9.04418156e-03 -1.69725135e-01
3.08184680e-02 -5.96343815e-01 -8.06860924e-01 1.39384055e+00
9.67723429e-01 4.08193544e-02 1.58702061e-01 -8.55175376e-01
-5.73046148e-01 5.06235421e-01 -5.99606693e-01 3.47314715e-01
-4.19603199e-01 3.95977288e-01 5.15160799e-01 -5.00174820e-01
-6.46393225e-02 1.36348641e+00 1.01621616e+00 5.21311283e-01
-1.21499979e+00 -9.36167419e-01 2.19618604e-01 3.06233644e-01
-1.35034835e+00 -2.45725185e-01 8.41651618e-01 -1.62855729e-01
5.46439469e-01 1.50397408e-03 5.74544609e-01 1.12748539e+00
3.23150396e-01 4.40491825e-01 1.01500237e+00 -1.02002593e-02
8.45686793e-02 -2.02061549e-01 5.07749282e-02 5.06604016e-01
5.31602502e-01 9.40338969e-02 -3.90248536e-03 1.39139697e-01
1.08094728e+00 1.73625931e-01 -5.39833724e-01 -3.40592898e-02
-7.58373320e-01 6.90857053e-01 1.02084601e+00 3.86069983e-01
-4.17346179e-01 3.76667768e-01 6.29403412e-01 2.57390112e-01
4.57444698e-01 -1.59924015e-01 -3.14985871e-01 1.12528965e-01
-6.89558387e-01 1.78003088e-01 5.69067895e-01 6.00513697e-01
5.54107249e-01 2.83943415e-01 -2.33350158e-01 1.11210775e+00
3.47870618e-01 3.36455107e-01 5.19950271e-01 -7.51095295e-01
4.49766427e-01 3.39117676e-01 -1.30314857e-01 -1.19996977e+00
-3.53065878e-01 -7.64671504e-01 -1.70817852e+00 3.52317870e-01
2.08747342e-01 -3.38327140e-01 -1.00143027e+00 1.62957978e+00
1.06792837e-01 2.24070609e-01 4.29776311e-02 1.17992055e+00
1.04142952e+00 8.05334091e-01 1.30722240e-01 -7.32641220e-02
1.31907630e+00 -9.45073545e-01 -8.91943634e-01 -2.46255040e-01
1.20515816e-01 -9.05735910e-01 7.83291399e-01 5.34192502e-01
-1.13639331e+00 -9.00330663e-01 -1.21458089e+00 -3.12663883e-01
-1.19948715e-01 4.42055136e-01 4.19642240e-01 4.47615981e-01
-9.97145772e-01 7.68616021e-01 -1.00828922e+00 -8.51651132e-02
6.38751924e-01 2.84777522e-01 -2.31154338e-01 -3.00543845e-01
-1.22390866e+00 5.72138608e-01 3.42980027e-01 1.09636307e+00
-1.07708645e+00 -3.15588027e-01 -7.46382117e-01 3.68435115e-01
2.02885166e-01 -6.44794881e-01 1.08218801e+00 -8.68489861e-01
-1.45578551e+00 3.86899590e-01 6.18509501e-02 -2.52492994e-01
6.93983257e-01 -3.75892401e-01 -5.24035394e-01 9.45717283e-03
-8.74839276e-02 2.93068916e-01 8.75153899e-01 -1.28265119e+00
-4.91757393e-01 -2.77299762e-01 -6.48050308e-02 -3.83912772e-02
-5.35464704e-01 -3.36339593e-01 -6.69306993e-01 -9.68270540e-01
4.85069007e-01 -4.67359990e-01 -3.88162732e-01 2.29484230e-01
-2.62776464e-01 -4.78781275e-02 1.00408053e+00 -7.88079858e-01
1.32249951e+00 -2.35489440e+00 5.84967434e-02 4.75721993e-03
1.45409763e-01 7.64015019e-01 -2.32562259e-01 1.36330277e-01
-1.12316348e-01 1.48662776e-01 -2.45451480e-01 -3.23661298e-01
-2.86209404e-01 2.87249744e-01 3.57662104e-02 4.80664849e-01
2.30811268e-01 5.46973884e-01 -6.61336303e-01 -1.62194327e-01
2.26660103e-01 8.15553665e-01 -3.62307012e-01 3.35822344e-01
4.38817125e-03 3.88442099e-01 -4.60205942e-01 3.62292051e-01
1.28906560e+00 1.39739931e-01 -1.18846726e-02 -6.28594756e-01
-1.52423888e-01 -8.55272356e-03 -1.47407281e+00 1.52474487e+00
-5.46339273e-01 4.66926843e-01 7.60214329e-01 -9.83934939e-01
1.17314935e+00 2.81232148e-01 -9.83840823e-02 -9.11673546e-01
4.67365801e-01 4.16802883e-01 1.19867317e-01 -6.72997117e-01
-4.61965017e-02 -3.47269833e-01 4.02331233e-01 -1.14145637e-01
4.70276624e-02 5.71730249e-02 1.96682021e-01 -1.65889695e-01
1.02363157e+00 -3.01757399e-02 -6.00638203e-02 -4.12277550e-01
8.93800437e-01 -5.13186514e-01 1.01875448e+00 7.06373751e-01
-2.11391717e-01 6.74049258e-01 6.24754965e-01 -4.99773979e-01
-8.40847135e-01 -9.39397275e-01 -1.37849554e-01 5.15551150e-01
4.95773673e-01 -4.78581302e-02 -8.02737772e-01 -2.10700467e-01
-2.21996963e-01 2.13912413e-01 -4.34708297e-01 -3.54253054e-01
-6.91604376e-01 -8.39589119e-01 6.48011446e-01 5.13493836e-01
1.39781392e+00 -9.41673994e-01 -2.19350293e-01 2.24482670e-01
-2.12127641e-01 -9.75511491e-01 -4.84498769e-01 3.19988161e-01
-1.00563490e+00 -8.49900126e-01 -7.77862728e-01 -1.19599855e+00
6.26598001e-01 4.89541918e-01 8.05051506e-01 4.36700851e-01
-1.01453260e-01 -3.73252898e-01 -2.00536028e-01 -3.53021324e-01
-1.35972202e-01 -2.60533951e-02 -2.72425205e-01 1.98978737e-01
2.45991141e-01 -9.21323478e-01 -1.04737413e+00 3.98067534e-01
-1.33916295e+00 -9.74433497e-02 8.41436982e-01 1.21352184e+00
5.00545561e-01 4.19076502e-01 4.76456225e-01 -5.23185670e-01
7.04741120e-01 -2.82610595e-01 -8.91395152e-01 -1.27012759e-01
-3.31671804e-01 -7.01593384e-02 8.69954109e-01 -3.02994519e-01
-1.30795252e+00 -1.54922858e-01 -6.59794569e-01 -3.10431749e-01
-5.46996556e-02 5.32524645e-01 -5.09086192e-01 2.92178988e-02
4.15363610e-01 2.70274490e-01 3.16109091e-01 -9.41486895e-01
-7.93844759e-02 7.00716615e-01 6.64015889e-01 -3.04145545e-01
7.73718536e-01 4.02878612e-01 8.39258428e-04 -7.95489013e-01
-6.95357442e-01 -2.87926137e-01 -4.06116426e-01 2.44324617e-02
8.01605403e-01 -1.26881039e+00 -7.64414132e-01 1.18113673e+00
-1.24887204e+00 -3.33250791e-01 1.83339000e-01 4.25720304e-01
2.66566034e-02 3.71063918e-01 -9.92734790e-01 -5.60812891e-01
-5.19393444e-01 -1.37755549e+00 7.11564958e-01 5.25612652e-01
4.83609945e-01 -9.19526041e-01 -5.08818507e-01 8.56373236e-02
6.79067254e-01 1.67944282e-01 9.09325242e-01 -1.88324377e-01
-4.22991037e-01 -1.66703671e-01 -5.11742711e-01 1.12087607e+00
9.30801630e-02 -2.80434668e-01 -9.01891291e-01 -4.43977535e-01
3.50831509e-01 -1.94209307e-01 1.38352537e+00 5.88954449e-01
1.51960468e+00 -9.45984870e-02 -2.17049792e-02 9.07689154e-01
1.76974678e+00 2.07107246e-01 1.11156344e+00 4.30092275e-01
6.10225499e-01 1.62866741e-01 5.27801141e-02 2.33510584e-01
-4.22106460e-02 3.17315787e-01 7.65180230e-01 -4.15780991e-01
-3.66259873e-01 2.33903509e-02 1.92268729e-01 8.82563949e-01
-8.42698514e-02 -2.27745354e-01 -6.48401201e-01 4.73047376e-01
-1.71062315e+00 -7.25386918e-01 -3.53617281e-01 1.79846346e+00
8.34771216e-01 2.89311439e-01 -3.59210998e-01 2.38298655e-01
7.42983162e-01 1.47125468e-01 -5.01651704e-01 -2.85321802e-01
-2.55985945e-01 3.12336147e-01 4.37635988e-01 2.97372848e-01
-1.20639825e+00 5.34504592e-01 4.62983370e+00 1.12185061e+00
-1.20850635e+00 2.00987175e-01 6.85978234e-01 1.25895739e-01
-3.86713678e-03 -3.29730392e-01 -6.08478963e-01 5.62826097e-01
4.08137172e-01 3.40203434e-01 3.13497186e-01 5.65543175e-01
5.28357029e-01 -1.01401433e-01 -5.14189780e-01 1.02346146e+00
-3.12845558e-01 -1.10081792e+00 1.20653950e-01 -2.44423985e-01
6.24326289e-01 1.48882638e-04 -8.57582241e-02 1.41764954e-01
-1.37641653e-02 -9.45780396e-01 5.75988650e-01 4.62278962e-01
4.51666802e-01 -8.28036845e-01 1.33904040e+00 4.02003407e-01
-1.10026300e+00 -4.11769420e-01 -6.44750118e-01 -1.77986681e-01
2.92575452e-02 1.04796469e+00 2.16018021e-01 5.87198317e-01
1.20268357e+00 7.64266789e-01 -3.65796447e-01 1.23685288e+00
-4.17927802e-01 6.74527943e-01 -1.12596303e-01 2.23845288e-01
3.83795857e-01 -6.55447841e-01 3.68640304e-01 1.26209021e+00
3.94640476e-01 1.77156717e-01 -3.10030635e-02 7.70879507e-01
-3.12609464e-01 -2.64598697e-01 -1.24311030e-01 3.37200910e-01
2.55634815e-01 1.49994302e+00 -6.17431462e-01 -1.62543550e-01
-4.90884244e-01 8.81812513e-01 1.15003794e-01 5.95288217e-01
-7.59971082e-01 -7.62912631e-01 8.49277377e-01 -3.15355952e-03
6.54733300e-01 -2.26086155e-01 -5.77038169e-01 -9.51290071e-01
1.86467275e-01 -8.09837580e-01 1.45220757e-01 -6.41783476e-01
-1.40435970e+00 8.82341325e-01 -5.17025113e-01 -1.06224763e+00
6.68901920e-01 -8.02858949e-01 -8.13437939e-01 1.33051765e+00
-1.68330336e+00 -9.53465223e-01 -9.31963205e-01 4.58804965e-01
3.87953609e-01 1.96871515e-02 5.09785831e-01 6.96416557e-01
-9.91641283e-01 3.76900256e-01 3.93060505e-01 3.54429394e-01
5.55607200e-01 -9.79323804e-01 1.74686074e-01 1.10977173e+00
-6.58507764e-01 6.15098596e-01 5.47922492e-01 -4.91619974e-01
-1.01725125e+00 -1.19048393e+00 2.56372035e-01 5.29957235e-01
4.19862837e-01 -4.05584365e-01 -1.18328500e+00 2.89065808e-01
2.01459274e-01 3.54048640e-01 1.11106904e-02 -2.91970283e-01
-2.77609348e-01 -7.65983999e-01 -1.14374518e+00 6.01149082e-01
8.40660334e-01 -5.98630048e-02 -1.52264953e-01 -5.85869886e-02
6.79309964e-01 -3.88717562e-01 -7.64299512e-01 4.77426887e-01
4.14335847e-01 -1.06083083e+00 9.53239560e-01 1.80449337e-02
6.44505858e-01 -5.00601172e-01 4.21137810e-02 -1.39608347e+00
-5.15935957e-01 -3.48067433e-01 4.22171913e-02 1.32888067e+00
9.38888788e-02 -7.02593207e-01 4.24935102e-01 4.61844206e-02
-4.11678135e-01 -1.05335808e+00 -9.92098272e-01 -7.13529170e-01
8.47896785e-02 -1.35214597e-01 3.61633539e-01 4.27956641e-01
-8.42082024e-01 3.02416533e-01 -2.83308327e-01 2.79343456e-01
6.31392300e-01 -3.39322388e-01 3.73485088e-01 -1.15597963e+00
-7.68816993e-02 -5.23114085e-01 -3.66331905e-01 -1.43639600e+00
-2.64804631e-01 -4.87930149e-01 1.95765063e-01 -1.74317527e+00
4.70775776e-02 -4.42634761e-01 -4.27332193e-01 3.66088420e-01
-3.39967310e-01 4.44331348e-01 -2.23518675e-03 -4.00051055e-03
-6.64007217e-02 8.88034284e-01 1.49626291e+00 -2.08221406e-01
-7.11758509e-02 1.16617475e-02 -7.50806510e-01 8.56564999e-01
9.90099490e-01 -2.30338082e-01 -2.78353542e-01 -1.01213241e+00
-1.83011696e-01 -2.42698342e-01 7.41405666e-01 -1.14338601e+00
3.49884093e-01 2.98022449e-01 6.00738704e-01 -3.59161764e-01
2.81333148e-01 -1.03594816e+00 1.49165019e-01 8.07923079e-01
-3.80364656e-02 -1.17900349e-01 3.10995936e-01 5.91215670e-01
-5.88265181e-01 -2.34714851e-01 1.11957145e+00 -1.80078030e-01
-7.42941916e-01 3.08167458e-01 -1.86231762e-01 -3.07017684e-01
7.71173477e-01 -1.08993158e-01 -3.59683514e-01 -1.92741916e-01
-5.52795172e-01 3.19937557e-01 3.08484491e-02 3.14465821e-01
7.53538966e-01 -1.30845630e+00 -6.49024665e-01 2.11471051e-01
-4.18027341e-01 3.77338707e-01 6.14805222e-01 9.73495722e-01
-8.63606215e-01 -9.23491716e-02 -2.33739927e-01 -4.72441971e-01
-1.06928957e+00 2.53652513e-01 7.60392487e-01 -1.52868077e-01
-7.27387786e-01 9.37952638e-01 1.99709237e-01 -3.29020053e-01
4.74156410e-01 -2.66035348e-01 -1.15664072e-01 -1.00120693e-01
4.89184529e-01 5.21797419e-01 2.78113037e-01 -4.09598291e-01
-1.00519903e-01 4.13241416e-01 3.65205780e-02 4.11621213e-01
1.63684464e+00 -2.01197147e-01 -5.28868139e-01 -2.51599923e-02
1.29936159e+00 -4.48962599e-01 -1.49019873e+00 -2.93337077e-01
-4.14079428e-01 -3.79109919e-01 5.52387774e-01 -7.95030415e-01
-1.67478967e+00 1.05223691e+00 9.97659028e-01 7.39000365e-02
1.80810547e+00 -5.34084320e-01 1.16139960e+00 1.25673458e-01
-1.76749956e-02 -8.89957845e-01 1.19572163e-01 4.18055266e-01
9.20790195e-01 -1.15741587e+00 -3.60315852e-02 -5.37614226e-01
-1.64545581e-01 1.28657877e+00 8.96394253e-01 -3.08953822e-01
8.28437150e-01 3.02578539e-01 2.87871778e-01 -7.34087601e-02
-2.91612357e-01 -3.08923684e-02 -6.75734356e-02 3.81528944e-01
3.53089720e-01 -2.55385756e-01 -5.54662764e-01 1.01405215e+00
1.37891918e-01 -3.51669937e-02 3.71654361e-01 6.87536418e-01
-5.03362596e-01 -8.70787680e-01 -3.69909555e-01 4.47200775e-01
-6.38698876e-01 -2.23334834e-01 2.83591956e-01 6.54585660e-01
6.43658161e-01 1.16758025e+00 1.58444136e-01 -3.21572065e-01
6.28906250e-01 -4.80101138e-01 5.62161915e-02 -1.50478899e-01
-7.50111878e-01 2.99073756e-01 -2.32653633e-01 -4.64738190e-01
-3.28340560e-01 -2.24041000e-01 -9.12606180e-01 -4.42518294e-01
-4.10660326e-01 1.67784140e-01 5.01338959e-01 6.20757341e-01
2.09032148e-01 9.68822241e-01 5.10608792e-01 -8.75873089e-01
-7.08939135e-01 -1.19631541e+00 -7.51075506e-01 2.38158181e-01
5.58344483e-01 -4.72207755e-01 -3.14228565e-01 -2.76107229e-02] | [11.40324878692627, -2.376570224761963] |
ad022e2c-9cb5-4ac6-993c-4bdc90ddd65b | urbangiraffe-representing-urban-scenes-as | 2303.14167 | null | https://arxiv.org/abs/2303.14167v2 | https://arxiv.org/pdf/2303.14167v2.pdf | UrbanGIRAFFE: Representing Urban Scenes as Compositional Generative Neural Feature Fields | Generating photorealistic images with controllable camera pose and scene contents is essential for many applications including AR/VR and simulation. Despite the fact that rapid progress has been made in 3D-aware generative models, most existing methods focus on object-centric images and are not applicable to generating urban scenes for free camera viewpoint control and scene editing. To address this challenging task, we propose UrbanGIRAFFE, which uses a coarse 3D panoptic prior, including the layout distribution of uncountable stuff and countable objects, to guide a 3D-aware generative model. Our model is compositional and controllable as it breaks down the scene into stuff, objects, and sky. Using stuff prior in the form of semantic voxel grids, we build a conditioned stuff generator that effectively incorporates the coarse semantic and geometry information. The object layout prior further allows us to learn an object generator from cluttered scenes. With proper loss functions, our approach facilitates photorealistic 3D-aware image synthesis with diverse controllability, including large camera movement, stuff editing, and object manipulation. We validate the effectiveness of our model on both synthetic and real-world datasets, including the challenging KITTI-360 dataset. | ['Yiyi Liao', 'Yue Wang', 'Rong Xiong', 'Hanlei Guo', 'Yifei Yang', 'Yuanbo Yang'] | 2023-03-24 | null | null | null | null | ['3d-aware-image-synthesis'] | ['computer-vision'] | [ 3.13278824e-01 -3.78305279e-02 2.55007923e-01 -1.18801571e-01
-3.72220844e-01 -8.57661009e-01 9.08636570e-01 -4.81341183e-01
1.27038300e-01 4.96353656e-01 2.32638061e-01 1.16234468e-02
2.28963066e-02 -1.01904535e+00 -1.03349781e+00 -6.18137240e-01
5.96546710e-01 7.61339903e-01 2.18014926e-01 -2.66586155e-01
1.05359279e-01 8.24693024e-01 -1.66394496e+00 -1.61173686e-01
9.62533355e-01 7.38124907e-01 5.40520728e-01 8.32925618e-01
-1.19865343e-01 8.97821248e-01 -5.71611643e-01 -2.14030415e-01
5.44593751e-01 -4.32056934e-01 -3.57533872e-01 8.69389653e-01
6.13276362e-01 -3.20908308e-01 -3.38988632e-01 8.24255407e-01
4.07367021e-01 2.76290923e-01 7.15455174e-01 -1.25080538e+00
-8.92128229e-01 1.30351439e-01 -7.03030407e-01 -3.50955248e-01
2.47707620e-01 5.28174102e-01 7.33532906e-01 -8.46937656e-01
7.55821526e-01 1.47009861e+00 2.55513698e-01 3.61526459e-01
-1.55615962e+00 -6.01027191e-01 4.03466910e-01 -2.15001434e-01
-1.51245177e+00 -2.16279864e-01 9.45773542e-01 -5.43101430e-01
5.22530437e-01 4.25399274e-01 9.52942908e-01 9.85024512e-01
1.37427077e-01 6.53808355e-01 9.75619435e-01 -2.23381549e-01
4.70955521e-01 6.82990402e-02 -5.08226573e-01 5.08321404e-01
1.12356000e-01 8.90472606e-02 -2.57578343e-01 1.17990494e-01
1.45970750e+00 1.81582421e-01 -2.50493705e-01 -9.53688204e-01
-1.35734785e+00 7.20658839e-01 3.76287550e-01 -3.20650756e-01
-2.02102751e-01 4.20303941e-01 -1.91774219e-01 -1.62858754e-01
5.14019430e-01 7.26852000e-01 -1.24060653e-01 2.21205816e-01
-7.62426436e-01 5.59839427e-01 5.58188915e-01 1.74709833e+00
7.01492667e-01 3.19997013e-01 -3.41912121e-01 7.09787607e-01
1.22164041e-01 1.00026536e+00 -2.38145545e-01 -1.27838218e+00
2.94948757e-01 5.62113523e-01 3.66225988e-01 -9.41920280e-01
3.93485017e-02 -3.67480040e-01 -9.91899014e-01 2.63968676e-01
4.43240926e-02 5.82933612e-02 -1.30593753e+00 1.57709944e+00
6.23760045e-01 1.64205506e-01 -3.47011805e-01 9.38093722e-01
7.44305193e-01 7.42903352e-01 -1.08502589e-01 1.00021079e-01
1.05717981e+00 -9.67945278e-01 -5.40792108e-01 -2.92623013e-01
1.47909224e-01 -8.51157606e-01 1.41465175e+00 2.93273926e-01
-1.22450268e+00 -3.56383473e-01 -7.02266634e-01 -4.02147323e-01
-8.62603784e-02 -8.95643234e-03 8.00682485e-01 4.56622958e-01
-9.57922876e-01 -8.60623419e-02 -5.35049379e-01 -1.22640625e-01
4.69172418e-01 2.32215505e-02 -1.96681961e-01 -4.64860618e-01
-6.92629457e-01 6.62896335e-01 1.49121165e-01 5.90582490e-02
-1.26307130e+00 -9.67568517e-01 -1.09089696e+00 8.55408683e-02
5.94619513e-01 -1.27842069e+00 9.57445085e-01 -4.35678452e-01
-1.61045492e+00 7.52812803e-01 4.96835709e-02 7.48587623e-02
7.16920435e-01 -2.15781089e-02 1.72960162e-01 -3.40861976e-02
1.91778764e-01 9.00384128e-01 1.00556302e+00 -1.65725005e+00
-3.46311063e-01 -1.39543578e-01 3.75109494e-01 6.11833155e-01
2.58389831e-01 -4.24946398e-01 -6.98728979e-01 -7.62300432e-01
1.07167684e-01 -1.09324801e+00 -5.06555736e-01 1.47963256e-01
-6.81016386e-01 2.80982465e-01 7.41987586e-01 -1.12582713e-01
7.20854461e-01 -2.22867942e+00 4.52676356e-01 4.06556241e-02
1.73797697e-01 -2.80027151e-01 -2.10908145e-01 3.87725323e-01
6.56437278e-02 1.13182560e-01 -2.33889848e-01 -6.25962496e-01
1.06251583e-01 2.49126092e-01 -5.33010006e-01 3.04266900e-01
2.08056554e-01 9.77530956e-01 -9.48834538e-01 -3.45774412e-01
8.63334596e-01 6.64653718e-01 -9.88739371e-01 2.82947540e-01
-7.01451421e-01 7.94953525e-01 -5.05273640e-01 5.72261810e-01
8.63504231e-01 -3.22243214e-01 -1.73244238e-01 -7.84950927e-02
-5.94366975e-02 -2.82956455e-02 -1.24149334e+00 1.81352937e+00
-7.80592561e-01 4.54594731e-01 5.21409810e-02 -2.78438389e-01
8.22731137e-01 -1.49754465e-01 2.10169137e-01 -3.94112527e-01
1.90449685e-01 -2.03030884e-01 -4.90742087e-01 -2.24229038e-01
7.09574342e-01 -8.91606212e-02 -1.97759166e-01 2.32094303e-01
-2.79518068e-01 -1.39319587e+00 1.78402290e-02 4.32302743e-01
7.64536977e-01 2.89935201e-01 4.70257699e-02 -2.36900032e-01
1.21311232e-01 7.94282630e-02 2.52914667e-01 5.62389910e-01
5.42070746e-01 1.33708870e+00 2.67402261e-01 -3.21002215e-01
-1.33360338e+00 -1.42266285e+00 3.03244591e-02 6.53239369e-01
5.88419497e-01 -1.59853518e-01 -7.02326775e-01 -1.09911479e-01
-9.84426886e-02 1.06386924e+00 -6.74799144e-01 -1.66667446e-01
-5.30303657e-01 -4.39800352e-01 -1.22398891e-01 3.70433569e-01
4.54227060e-01 -9.72729146e-01 -5.32774270e-01 2.16223183e-03
-2.45267019e-01 -1.20121622e+00 -9.99979436e-01 -1.85957834e-01
-6.40121818e-01 -8.78420353e-01 -5.92013478e-01 -5.26172400e-01
9.27353919e-01 6.74595892e-01 1.30523062e+00 -2.17624545e-01
-5.59468508e-01 5.29843271e-01 -1.47515103e-01 -3.00833583e-01
-3.24722558e-01 -1.96012452e-01 -2.78712034e-01 5.38031310e-02
-4.37708676e-01 -7.39658892e-01 -7.45382905e-01 3.95003408e-01
-1.10260558e+00 7.13093162e-01 3.46431792e-01 5.86810529e-01
8.94102156e-01 9.91314128e-02 -6.29377961e-02 -8.51479053e-01
9.91952568e-02 -2.75707692e-01 -8.55778694e-01 4.76787910e-02
-9.19247344e-02 -5.02481274e-02 5.99993467e-01 -6.67757988e-01
-1.22988200e+00 1.11306034e-01 2.28931457e-01 -9.43729460e-01
-1.49955109e-01 -2.43682340e-01 -6.04390502e-01 7.18787163e-02
4.99887615e-01 4.09919173e-01 -3.33171785e-01 -2.17508972e-01
8.39658380e-01 1.24257736e-01 5.80731153e-01 -7.90219724e-01
1.07783139e+00 7.88414359e-01 1.62251458e-01 -9.47009325e-01
-7.35972941e-01 -6.84228167e-02 -4.55838054e-01 -1.51362389e-01
1.00480819e+00 -1.10630453e+00 -5.39154828e-01 5.01577973e-01
-1.14105976e+00 -6.29945278e-01 -7.01057136e-01 1.31319612e-01
-8.72545958e-01 7.16759358e-03 -2.23347783e-01 -6.37289822e-01
9.83375162e-02 -1.29046857e+00 1.64049721e+00 2.20339727e-02
-1.39215186e-01 -7.75678992e-01 -1.85941666e-01 4.24774259e-01
2.59561628e-01 5.80268085e-01 8.80546093e-01 5.47703385e-01
-1.51120198e+00 1.20130312e-02 -1.95500478e-01 5.03796265e-02
1.99485436e-01 1.24398030e-01 -7.47899532e-01 8.38612318e-02
-1.39059216e-01 -1.40195593e-01 4.64638382e-01 5.54491162e-01
1.34078610e+00 -2.63196677e-01 -2.25075766e-01 1.04559648e+00
1.38357627e+00 1.15523055e-01 7.83633471e-01 -9.61322431e-03
1.16689074e+00 4.57999885e-01 6.13777399e-01 5.93530774e-01
5.30676961e-01 7.52659976e-01 7.12508380e-01 -2.03449264e-01
-4.57015544e-01 -7.34765351e-01 -1.12983905e-01 3.48681480e-01
6.62987009e-02 -6.50353372e-01 -5.77270865e-01 4.95292723e-01
-1.54628825e+00 -9.04751122e-01 -1.51072323e-01 2.05865550e+00
6.74770713e-01 -2.08176762e-01 -2.14084044e-01 -3.23033154e-01
6.12828732e-01 2.74275601e-01 -6.60447538e-01 5.31921312e-02
-2.39358440e-01 -4.06346098e-03 4.78691727e-01 6.35403931e-01
-8.30858588e-01 1.13180971e+00 6.00234556e+00 9.49573457e-01
-9.52656865e-01 -1.38278991e-01 7.32429862e-01 -3.83948475e-01
-1.00339901e+00 1.41734900e-02 -8.50328624e-01 4.24877107e-01
9.44469497e-02 -7.83051550e-02 6.61920011e-01 8.98679435e-01
5.56646109e-01 -2.18044147e-01 -1.07992852e+00 1.17600226e+00
1.01831891e-01 -1.55964136e+00 4.23098981e-01 1.68487653e-01
1.14465141e+00 -2.70041913e-01 2.59191781e-01 -2.37590317e-02
7.70314515e-01 -9.76259410e-01 1.31782377e+00 4.86932486e-01
1.10151398e+00 -7.30331779e-01 -5.30780740e-02 4.29657221e-01
-1.05023563e+00 1.80485502e-01 -3.51982474e-01 1.36271238e-01
4.37231690e-01 6.16256356e-01 -7.25949883e-01 3.31605881e-01
4.90298539e-01 6.06337905e-01 -2.77640671e-01 7.98744678e-01
-3.54686469e-01 1.14655837e-01 -4.44566965e-01 1.84766993e-01
1.72650889e-02 -6.01153612e-01 5.42584240e-01 7.39426792e-01
3.72585922e-01 3.06656003e-01 2.57984340e-01 1.35095668e+00
-6.94300830e-02 -1.98449954e-01 -8.37480903e-01 1.44931406e-01
5.97936034e-01 1.11102092e+00 -8.70164216e-01 -1.68458790e-01
8.97291452e-02 1.03597724e+00 1.35526225e-01 4.67882842e-01
-1.10308111e+00 2.14642063e-02 7.24180758e-01 5.41836381e-01
3.95242542e-01 -5.40273786e-01 -5.98631859e-01 -1.32160997e+00
-7.92036057e-02 -6.71643615e-01 -3.26705754e-01 -1.30634844e+00
-9.12030160e-01 5.40495455e-01 3.39497060e-01 -1.23486376e+00
-1.43755227e-01 -2.15660349e-01 -4.34267402e-01 7.50938594e-01
-1.07033050e+00 -1.41889453e+00 -6.30692065e-01 5.22955418e-01
8.36531818e-01 2.87134618e-01 4.44045573e-01 1.07109353e-01
-3.61978978e-01 8.47605914e-02 -1.53723285e-01 -2.30392307e-01
4.58363771e-01 -1.12022769e+00 5.78661084e-01 7.82821953e-01
9.44333076e-02 5.05840182e-01 7.65314817e-01 -6.42116845e-01
-1.61093962e+00 -1.48760533e+00 2.72051692e-01 -9.19778943e-01
2.31399104e-01 -8.71952295e-01 -3.50427985e-01 7.10101902e-01
1.06859878e-01 6.25312924e-02 3.11017074e-02 -3.45007002e-01
-3.03057820e-01 -4.19083796e-02 -1.02784562e+00 1.02288079e+00
1.58855224e+00 -1.72101289e-01 -8.47565830e-02 4.50411230e-01
1.08482611e+00 -7.14545965e-01 -5.95205843e-01 2.88859874e-01
2.46675193e-01 -7.75096059e-01 1.27641070e+00 -1.38383016e-01
6.03275597e-01 -6.07072890e-01 -2.12515950e-01 -1.47276378e+00
-4.00260389e-01 -6.62531614e-01 1.24636970e-01 1.08399057e+00
8.96211043e-02 -3.22913885e-01 7.10262597e-01 6.96963727e-01
-4.07212973e-01 -4.81347948e-01 -3.94013643e-01 -6.90116286e-01
-1.15603715e-01 -4.44626719e-01 8.53843212e-01 7.36784816e-01
-8.70534837e-01 4.07984555e-01 -5.52430689e-01 2.07800150e-01
7.38791168e-01 5.01645327e-01 1.36613119e+00 -8.84560764e-01
-3.43565613e-01 -3.74448478e-01 -2.03229889e-01 -1.41537130e+00
-4.32893708e-02 -4.85609651e-01 2.12115556e-01 -1.57845950e+00
2.86058813e-01 -7.66990900e-01 7.60578454e-01 6.54685944e-02
-1.40267760e-01 3.58691305e-01 3.02558213e-01 7.98526406e-02
-4.48828846e-01 1.01177204e+00 1.95109987e+00 -1.39076218e-01
-3.05654258e-01 -1.79307565e-01 -7.69257367e-01 6.63137794e-01
3.54444146e-01 -1.01824954e-01 -9.55210924e-01 -7.66629159e-01
2.21540824e-01 5.78354150e-02 6.57259941e-01 -7.97005117e-01
-1.01285793e-01 -6.50806725e-01 4.21554446e-01 -8.14468086e-01
7.32188582e-01 -8.78208697e-01 7.82554388e-01 -8.30016434e-02
-1.40917540e-01 -1.86836034e-01 8.04670677e-02 6.66852713e-01
1.65401429e-01 3.53348553e-01 9.29596007e-01 -2.34861821e-01
-4.82805341e-01 8.01865995e-01 3.40089649e-02 2.43580014e-01
1.18727207e+00 -3.67812872e-01 -1.32833302e-01 -5.53361714e-01
-5.52216470e-01 2.83930928e-01 1.05518413e+00 5.45124829e-01
6.71604455e-01 -1.43424463e+00 -5.38005292e-01 5.01098454e-01
2.09518209e-01 9.08482015e-01 5.86047053e-01 3.12595606e-01
-8.56597304e-01 1.98472753e-01 1.06917061e-01 -9.69253302e-01
-9.37083900e-01 7.35484242e-01 2.30723202e-01 1.91810235e-01
-8.05809617e-01 8.47783208e-01 1.23112488e+00 -6.06701493e-01
-2.38670676e-04 -5.37049115e-01 3.07462215e-01 -4.73011643e-01
2.64674217e-01 1.81470186e-01 -2.77846187e-01 -4.25234109e-01
9.92535576e-02 7.73719668e-01 2.09019244e-01 -1.64816305e-01
1.22127545e+00 -3.77166748e-01 8.73362795e-02 3.24998021e-01
8.27475548e-01 1.95393458e-01 -1.79238868e+00 8.30168799e-02
-8.74064684e-01 -8.43768597e-01 -1.09960428e-02 -4.87644285e-01
-1.02365530e+00 7.41952658e-01 3.62965697e-03 -1.54325858e-01
9.62266207e-01 1.77734464e-01 5.58908045e-01 -1.22546135e-02
8.08688879e-01 -6.43146932e-01 2.19418138e-01 4.19885188e-01
1.18793547e+00 -8.96601796e-01 -8.87688324e-02 -9.53263760e-01
-6.50689065e-01 4.94436175e-01 6.57621026e-01 -3.54818404e-01
4.32967514e-01 3.75835896e-01 -2.22687647e-01 -2.29540363e-01
-6.29467547e-01 -4.29474637e-02 1.89991549e-01 6.00047231e-01
-5.44498004e-02 1.64966494e-01 3.79776984e-01 7.36214519e-02
-4.95593965e-01 -1.18352167e-01 6.64576709e-01 5.68621397e-01
-2.16283888e-01 -7.74992526e-01 -4.75641727e-01 1.57755956e-01
1.75217077e-01 -1.78675264e-01 -1.91207796e-01 9.33678448e-01
1.78040266e-01 6.69045746e-01 2.64085233e-01 -1.30437493e-01
4.47985470e-01 -5.63781261e-01 7.89186299e-01 -1.02865398e+00
4.24946025e-02 2.66820550e-01 -1.64045632e-01 -5.38187861e-01
-1.91548198e-01 -5.82121789e-01 -9.21608746e-01 -4.79260355e-01
-2.36606747e-01 -2.37459555e-01 5.20122468e-01 6.10911548e-01
4.62439686e-01 6.56347573e-01 7.44232297e-01 -1.23841131e+00
-1.51735112e-01 -6.18605793e-01 -7.25733042e-01 4.32419628e-01
2.08587870e-01 -9.16232347e-01 -4.21919763e-01 2.85176069e-01] | [9.223445892333984, -3.1164424419403076] |
8ab2a445-bd34-4baf-b644-802d63fc3416 | dadmatools-natural-language-processing | null | null | https://aclanthology.org/2022.naacl-demo.13 | https://aclanthology.org/2022.naacl-demo.13.pdf | DadmaTools: Natural Language Processing Toolkit for Persian Language | We introduce DadmaTools, an open-source Python Natural Language Processing toolkit for the Persian language. The toolkit is a neural pipeline based on spaCy for several text processing tasks, including normalization, tokenization, lemmatization, part-of-speech, dependency parsing, constituency parsing, chunking, and ezafe detecting. DadmaTools relies on fine-tuning of ParsBERT using the PerDT dataset for most of the tasks. Dataset module and embedding module are included in DadmaTools that support different Persian datasets, embeddings, and commonly used functions for them. Our evaluations show that DadmaTools can attain state-of-the-art performance on multiple NLP tasks. The source code is freely available at https://github.com/Dadmatech/DadmaTools. | ['Mohammad Taher Pilehvar', 'Mohamad Bagher Sajadi', 'Najmeh Zare', 'Mohammad Karrabi', 'Romina Etezadi'] | null | null | null | null | naacl-acl-2022-7 | ['constituency-parsing'] | ['natural-language-processing'] | [-6.57246828e-01 1.94393814e-01 -1.65438384e-01 -7.30574548e-01
-7.61093676e-01 -9.08444881e-01 5.14640749e-01 5.99940181e-01
-9.30337965e-01 5.69130838e-01 6.86648607e-01 -2.62169927e-01
4.30756897e-01 -7.65027344e-01 -2.08474308e-01 -2.91022450e-01
1.20641664e-01 7.79227912e-01 1.13372758e-01 -3.21081072e-01
2.80003518e-01 4.47868854e-01 -7.51334488e-01 5.69622934e-01
7.17130840e-01 4.09506708e-01 1.43564001e-01 8.62257898e-01
-6.81679249e-01 7.83745408e-01 -4.18210506e-01 -8.70329916e-01
1.76013529e-01 3.86368632e-01 -1.14001417e+00 -1.06842756e+00
1.99671298e-01 -1.88786581e-01 -4.65230554e-01 8.65102172e-01
6.70184374e-01 -1.22017064e-03 3.29277307e-01 -1.04798985e+00
-7.52133787e-01 1.55326009e+00 -1.21785887e-01 6.42272890e-01
2.87495047e-01 1.18581370e-01 1.22805882e+00 -1.09720147e+00
7.33355463e-01 1.63348210e+00 8.15682411e-01 4.47245210e-01
-9.89047289e-01 -7.11966038e-01 -2.63933331e-01 1.71297222e-01
-1.07491910e+00 -6.59779608e-01 2.50845373e-01 -2.60965645e-01
1.68807817e+00 3.00044157e-02 6.75063506e-02 1.05611539e+00
5.59359372e-01 8.80841851e-01 7.86980748e-01 -2.97655374e-01
5.39821126e-02 -4.11972493e-01 1.10910881e+00 6.11282170e-01
1.32610634e-01 -2.90647507e-01 -4.83838201e-01 -9.55632851e-02
3.50089699e-01 -4.54929203e-01 3.98095399e-01 6.60752833e-01
-1.14613688e+00 8.23143899e-01 7.50582889e-02 4.83331650e-01
-4.77749795e-01 -9.77078155e-02 8.77067506e-01 -6.00188365e-03
4.52387065e-01 4.87973094e-01 -9.51342940e-01 -6.15967453e-01
-7.21706152e-01 1.31582052e-01 1.14708912e+00 9.54993904e-01
6.88168705e-01 -7.49133527e-02 -3.22668850e-01 1.16911852e+00
1.72833636e-01 2.56203532e-01 6.23579025e-01 -9.59644496e-01
8.95723283e-01 6.83775187e-01 -2.38324881e-01 -3.89038503e-01
-6.96125209e-01 3.61231953e-01 -6.76372945e-01 -2.98150569e-01
4.08193886e-01 -5.11625111e-01 -9.43831742e-01 1.52773893e+00
4.05960143e-01 -3.11504364e-01 3.37540239e-01 5.30334353e-01
1.27339816e+00 1.04717052e+00 6.12884402e-01 1.96274281e-01
1.79530787e+00 -1.03496075e+00 -7.38928378e-01 -5.53578556e-01
6.82189405e-01 -1.06852496e+00 1.13738573e+00 2.94373184e-01
-1.09329271e+00 -2.53875732e-01 -6.64004028e-01 -9.22431827e-01
-7.21363485e-01 4.57149565e-01 6.41872227e-01 3.31212312e-01
-6.99745595e-01 8.27770233e-01 -1.31123412e+00 -7.07241774e-01
3.28284264e-01 9.69347358e-02 -8.60947311e-01 2.33288482e-03
-1.25108159e+00 1.01845586e+00 1.05446362e+00 -6.77207336e-02
-3.43938351e-01 -7.58594990e-01 -1.13971448e+00 5.44984788e-02
-1.60647318e-01 -2.17380866e-01 1.58759391e+00 -3.40024829e-01
-1.43816555e+00 1.19264638e+00 -2.11189196e-01 -5.65105438e-01
1.50516406e-02 -7.85515726e-01 -4.40547645e-01 -1.18184738e-01
4.33998436e-01 6.41789258e-01 2.81522065e-01 -2.58394241e-01
-4.21298414e-01 -4.43651289e-01 -2.23512232e-01 -2.40094624e-02
7.02785701e-02 8.45683753e-01 -5.27876258e-01 -7.05315113e-01
-1.62466779e-01 -6.52943254e-01 -1.00192085e-01 -6.17019653e-01
-7.47254491e-01 -7.49945700e-01 5.61813176e-01 -1.33469272e+00
1.29272997e+00 -2.30674601e+00 -2.37151101e-01 -4.62363064e-01
-1.35772407e-01 6.60372734e-01 -4.33273137e-01 1.01418138e+00
-2.69412905e-01 3.19176137e-01 -4.70567167e-01 -5.97784042e-01
3.91375273e-01 6.07565045e-01 -1.49952203e-01 3.71531427e-01
7.04599202e-01 9.34987128e-01 -7.10028529e-01 -6.33730471e-01
2.01560542e-01 4.85531926e-01 -3.72361958e-01 3.95339429e-01
-1.93023458e-02 -5.82727836e-03 -8.12985748e-02 7.61721492e-01
1.08110309e+00 5.73096931e-01 3.79016727e-01 -1.27195075e-01
-6.42587066e-01 1.16705346e+00 -1.11950612e+00 1.79571235e+00
-3.63023281e-01 3.81212711e-01 3.52740884e-01 -6.78199470e-01
6.36248291e-01 1.46809384e-01 1.49392998e-02 -2.95523643e-01
4.02953148e-01 2.53683757e-02 -1.72809530e-02 -6.68617010e-01
8.04443538e-01 2.56834418e-01 -5.55455923e-01 4.87323791e-01
7.55691171e-01 1.29852995e-01 6.98830783e-01 4.94986743e-01
1.41086340e+00 1.35330642e-02 9.03371751e-01 -3.27148795e-01
2.48039931e-01 1.22495502e-01 1.14823663e+00 2.28285968e-01
-2.46204600e-01 3.98758560e-01 7.09401727e-01 -4.01847899e-01
-1.05942285e+00 -1.41600299e+00 -2.40369782e-01 1.63437569e+00
-7.28490949e-01 -8.14958036e-01 -8.90119374e-01 -6.78692937e-01
8.46547559e-02 1.18012238e+00 -4.86651838e-01 4.19537991e-01
-9.76713002e-01 -8.83431911e-01 1.30341148e+00 7.65423894e-01
5.90009332e-01 -1.70788646e+00 -2.45198056e-01 4.99798357e-01
-1.55991673e-01 -1.34591925e+00 -4.70067143e-01 4.69046682e-01
-6.36850238e-01 -1.16547179e+00 -3.54877338e-02 -1.10389912e+00
1.35190845e-01 -4.21877921e-01 1.33899748e+00 -4.47051466e-01
-2.68679947e-01 7.48679489e-02 -4.28358346e-01 -4.95511919e-01
-4.55446720e-01 4.06192154e-01 1.63719907e-01 -7.32891262e-01
7.79068649e-01 -8.19952130e-01 -1.26692250e-01 -3.71487796e-01
-8.70852709e-01 -3.04756939e-01 6.48026407e-01 4.63231891e-01
4.52974409e-01 -3.43294472e-01 2.92653352e-01 -1.15210688e+00
8.38126898e-01 -6.10513210e-01 -6.03611588e-01 4.83064465e-02
-2.06240207e-01 2.15791687e-01 8.68033290e-01 -8.23406205e-02
-1.22427773e+00 2.30967194e-01 -7.54613578e-01 1.67717353e-01
-6.11802578e-01 4.83051777e-01 -5.88316381e-01 7.54343748e-01
4.33765650e-01 -3.15212198e-02 -4.11388159e-01 -9.20784056e-01
9.64535654e-01 9.07058060e-01 9.42715585e-01 -6.57289445e-01
3.95112008e-01 -5.90837598e-02 -6.55423820e-01 -9.48279858e-01
-6.98518872e-01 -6.19347692e-01 -1.19408560e+00 5.94487965e-01
1.10869861e+00 -8.76448929e-01 -5.11673868e-01 6.37465835e-01
-1.63990188e+00 -4.35849994e-01 -1.97237417e-01 3.09240460e-01
3.82540338e-02 3.88113230e-01 -1.38176632e+00 -4.12644982e-01
-1.17763627e+00 -6.31917238e-01 7.09706664e-01 3.53049904e-01
-5.54172635e-01 -8.81766677e-01 5.78965783e-01 2.65771389e-01
1.05988428e-01 -2.21728981e-01 8.52878869e-01 -1.33520412e+00
9.95283276e-02 -8.05588998e-03 -2.95204490e-01 6.49153769e-01
-9.25336704e-02 3.81144643e-01 -8.98076773e-01 8.55793804e-02
-4.10186678e-01 -4.16700214e-01 9.66146529e-01 4.20628607e-01
7.60434330e-01 -6.53934896e-01 -9.29721594e-02 7.03274190e-01
1.17031682e+00 7.29540959e-02 7.54523933e-01 4.29750413e-01
7.36067891e-01 6.03237152e-01 6.21416390e-01 4.94060904e-01
8.14973176e-01 -9.87616368e-03 1.45855129e-01 4.77411151e-01
1.02914341e-01 -2.00202078e-01 7.13301063e-01 1.19447267e+00
3.43143940e-01 -8.69071856e-02 -1.35630941e+00 8.24701905e-01
-1.73646665e+00 -8.74372184e-01 -3.48783493e-01 1.53278756e+00
1.05883062e+00 -2.01689526e-01 8.44471902e-02 -1.21735111e-01
6.32807136e-01 2.96988338e-01 -3.20813537e-01 -1.28792846e+00
-1.74065545e-01 8.90212655e-01 5.76476932e-01 5.16973436e-01
-1.50491071e+00 1.76198077e+00 5.35718489e+00 5.97913384e-01
-9.50716197e-01 3.96390080e-01 2.55339652e-01 2.49606177e-01
1.77382961e-01 1.79762691e-02 -1.27009702e+00 2.94926167e-01
1.25163960e+00 -1.30460322e-01 3.80574375e-01 9.59963202e-01
1.71198577e-01 -1.00825191e-01 -9.64273691e-01 7.17329443e-01
-2.27243885e-01 -1.20122731e+00 -2.61301368e-01 -4.86755461e-01
-7.95312524e-02 9.32669103e-01 -6.51177883e-01 7.08288431e-01
7.77507126e-01 -9.73542094e-01 4.68745351e-01 -3.74275707e-02
5.45969605e-01 -7.63536632e-01 7.96348333e-01 1.85765222e-01
-1.21392488e+00 6.47469983e-02 -5.55294454e-01 3.27319168e-02
4.32191223e-01 9.47408438e-01 -1.07953501e+00 3.99880886e-01
8.69472563e-01 7.21481979e-01 -6.21581376e-01 6.68926418e-01
-9.14978921e-01 1.13706028e+00 -4.89229202e-01 1.59278572e-01
5.92610091e-02 -1.76217720e-01 6.27278328e-01 2.17490554e+00
-5.93080334e-02 3.87527525e-01 -7.82172605e-02 5.85310161e-01
-2.50322640e-01 4.47302997e-01 -2.03849703e-01 -5.09247899e-01
1.05737054e+00 1.78238702e+00 -5.86629927e-01 -3.21518302e-01
-6.24834523e-02 1.02089298e+00 7.76706100e-01 -2.06452280e-01
-7.04008639e-01 -8.26449573e-01 9.45878625e-01 -1.94257617e-01
3.01668078e-01 -5.92707276e-01 -4.44923311e-01 -1.34390283e+00
-2.48949811e-01 -7.15101182e-01 9.20325339e-01 -5.04928052e-01
-1.50965893e+00 7.06271887e-01 -5.80808744e-02 -4.39368695e-01
-1.28765821e-01 -8.66152942e-01 -1.19777596e+00 7.90999353e-01
-1.07660115e+00 -1.18097758e+00 2.31288522e-01 6.22813523e-01
6.53031945e-01 -1.39477432e-01 1.23187089e+00 3.74302149e-01
-1.14472783e+00 4.25899357e-01 -6.88273758e-02 9.88203704e-01
1.08808100e+00 -1.57789040e+00 1.15745556e+00 1.05315518e+00
1.55388806e-02 7.62668729e-01 2.62846321e-01 -7.08786190e-01
-1.31753087e+00 -1.42629158e+00 1.54848933e+00 -3.45888793e-01
9.15372014e-01 -6.16295457e-01 -9.44170833e-01 1.18705571e+00
7.57498264e-01 -9.35702771e-02 8.79807055e-01 3.56122941e-01
-5.16410887e-01 2.86140926e-02 -1.32017744e+00 2.39218667e-01
7.01299787e-01 -4.25120652e-01 -1.29059136e+00 3.45356725e-02
6.84161544e-01 -3.66670340e-01 -1.01828372e+00 -1.42096519e-01
4.27869499e-01 -5.83469868e-01 6.56059563e-01 -6.87419891e-01
6.14996374e-01 -1.78778302e-02 -2.64680713e-01 -1.46580255e+00
-5.85119188e-01 -4.35716271e-01 1.24653421e-01 2.16060424e+00
5.88423967e-01 -7.50029147e-01 3.06119323e-01 4.63040829e-01
-4.39445347e-01 -2.09761202e-01 -1.09481359e+00 -5.83560586e-01
5.00591755e-01 -7.11349189e-01 4.07208830e-01 1.26055777e+00
2.68844366e-01 7.55205333e-01 2.12048143e-01 3.98587972e-01
4.00750518e-01 -2.51051635e-01 5.48992455e-01 -1.11080003e+00
-2.78583113e-02 -1.20185740e-01 -1.33681089e-01 -3.78347754e-01
5.05782425e-01 -1.23070848e+00 7.50494525e-02 -1.91496778e+00
-1.68395981e-01 -3.13020945e-01 -8.24420825e-02 1.31060421e+00
4.92320023e-02 -1.11548945e-01 4.74853426e-01 -1.63365551e-03
-4.17230517e-01 3.18178117e-01 2.75947332e-01 -1.42041489e-01
-3.60548615e-01 -4.47847754e-01 -7.29190409e-01 8.97795558e-01
1.76207566e+00 -9.12955463e-01 3.92386287e-01 -8.03650737e-01
3.02854534e-02 -3.68401438e-01 -1.56914636e-01 -8.80530715e-01
4.96883877e-03 -4.85532843e-02 2.14772329e-01 -7.92111814e-01
3.02172273e-01 -1.48683041e-02 -4.01726186e-01 2.53828168e-01
5.48281521e-02 6.46091461e-01 7.27134407e-01 -3.61726731e-01
-1.09728143e-01 -3.06237370e-01 8.78894210e-01 -1.57011688e-01
-8.62519145e-01 2.06680242e-02 -7.17758477e-01 5.79832792e-01
5.43354452e-01 5.51607430e-01 -6.69214785e-01 3.32511961e-01
-6.48651302e-01 3.34374487e-01 -1.64233204e-02 4.75347549e-01
2.99805999e-01 -1.04373777e+00 -9.40455437e-01 8.71747956e-02
-2.00097710e-02 1.61040142e-01 -2.03677919e-02 6.42762542e-01
-1.00795841e+00 3.88329774e-01 -4.00555879e-01 5.52395284e-02
-1.42734408e+00 2.49936908e-01 -9.49758515e-02 -6.72004759e-01
-3.93244296e-01 8.42037380e-01 -3.29179853e-01 -1.18520463e+00
-1.35165006e-01 -7.59408772e-01 -3.33334059e-01 2.78556734e-01
6.44949377e-01 6.65004730e-01 8.91820416e-02 -6.98155642e-01
-8.74420345e-01 -4.58028838e-02 -3.23641211e-01 -1.54563561e-01
1.73403478e+00 2.46947646e-01 -7.45104015e-01 3.92043233e-01
8.27931106e-01 4.79669094e-01 -6.71985626e-01 1.69656351e-01
3.72681409e-01 3.65228802e-01 -2.35036880e-01 -1.10901082e+00
-7.86459625e-01 9.70198929e-01 1.41115412e-01 -9.99150425e-02
8.84165168e-01 1.22218333e-01 1.06955612e+00 5.62077701e-01
-6.55965880e-02 -1.46966255e+00 -7.27742970e-01 1.48373735e+00
8.79976392e-01 -9.24004078e-01 -1.96721315e-01 -3.63056272e-01
-7.94566929e-01 1.24189889e+00 6.24927819e-01 -4.35669005e-01
8.83395612e-01 6.52641773e-01 4.64973092e-01 8.63437057e-02
-7.30146289e-01 6.99261727e-04 -2.07977891e-01 4.86265093e-01
9.44812000e-01 4.11045402e-01 -7.83477783e-01 1.54696953e+00
-7.97245622e-01 -3.20361316e-01 4.65621024e-01 9.47103262e-01
-2.63143033e-01 -1.63111997e+00 -4.44388390e-01 1.92747191e-01
-9.42475259e-01 -7.47178376e-01 -6.83079422e-01 8.71916831e-01
2.35454127e-01 8.63185465e-01 2.40096092e-01 -1.63360074e-01
4.40409988e-01 4.25950736e-01 2.71614760e-01 -9.55362678e-01
-9.50215042e-01 -3.29498798e-01 6.52054012e-01 -6.89156830e-01
7.34163960e-03 -7.46326864e-01 -1.88047218e+00 -5.20879626e-01
3.84149879e-01 1.74951598e-01 4.12953854e-01 7.01140344e-01
6.73022628e-01 4.97444235e-02 -1.39056826e-02 -8.65982890e-01
-1.79482758e-01 -1.54743457e+00 -5.01086056e-01 4.70927879e-02
-2.47248575e-01 -1.50690163e-02 -7.22291693e-02 -7.40534589e-02] | [10.466817855834961, 9.992916107177734] |
7d6623cd-9714-4c73-b46e-2102f12d8327 | egocentric-audio-visual-noise-suppression | 2211.03643 | null | https://arxiv.org/abs/2211.03643v2 | https://arxiv.org/pdf/2211.03643v2.pdf | Egocentric Audio-Visual Noise Suppression | This paper studies audio-visual noise suppression for egocentric videos -- where the speaker is not captured in the video. Instead, potential noise sources are visible on screen with the camera emulating the off-screen speaker's view of the outside world. This setting is different from prior work in audio-visual speech enhancement that relies on lip and facial visuals. In this paper, we first demonstrate that egocentric visual information is helpful for noise suppression. We compare object recognition and action classification-based visual feature extractors and investigate methods to align audio and visual representations. Then, we examine different fusion strategies for the aligned features, and locations within the noise suppression model to incorporate visual information. Experiments demonstrate that visual features are most helpful when used to generate additive correction masks. Finally, in order to ensure that the visual features are discriminative with respect to different noise types, we introduce a multi-task learning framework that jointly optimizes audio-visual noise suppression and video-based acoustic event detection. This proposed multi-task framework outperforms the audio-only baseline on all metrics, including a 0.16 PESQ improvement. Extensive ablations reveal the improved performance of the proposed model with multiple active distractors, overall noise types, and across different SNRs. | ['Kaustubh Kalgaonkar', 'Yang Liu', 'Egor Lakomkin', 'Ju Lin', 'Weipeng He', 'Roshan Sharma'] | 2022-11-07 | null | null | null | null | ['action-classification'] | ['computer-vision'] | [ 2.89787829e-01 -4.71364170e-01 1.29415244e-01 -9.10207629e-02
-1.37182498e+00 -5.00229299e-01 3.85806739e-01 -2.09247187e-01
-4.96358842e-01 2.33967438e-01 8.33213329e-01 1.81374714e-01
4.59709689e-02 -1.72493346e-02 -6.60525382e-01 -1.00756788e+00
1.45456314e-01 -6.61496341e-01 1.46022558e-01 -5.08574024e-03
1.76677182e-01 3.40833217e-01 -1.92293227e+00 5.50833821e-01
2.47708216e-01 9.97410059e-01 4.14991617e-01 1.23043346e+00
3.65256578e-01 6.14973783e-01 -8.65896165e-01 -9.27827135e-03
3.13114643e-01 -2.60961920e-01 -9.17364433e-02 6.36472404e-02
8.34061086e-01 -3.73190075e-01 -6.07795835e-01 1.14520824e+00
1.14484715e+00 3.41168106e-01 4.81398076e-01 -1.48611951e+00
-1.96526676e-01 2.77227670e-01 -7.69561589e-01 4.53149289e-01
6.92690194e-01 3.41321796e-01 8.32352400e-01 -1.08356476e+00
2.15924412e-01 1.38670290e+00 4.40963745e-01 6.37394786e-01
-1.27955651e+00 -8.58643472e-01 2.22422406e-01 5.83025038e-01
-1.55013001e+00 -1.28675771e+00 9.25901115e-01 -2.37787455e-01
9.17361081e-01 4.36855048e-01 3.20999473e-01 1.41982079e+00
2.45508971e-03 6.48430407e-01 6.40231550e-01 -5.36992550e-01
1.15969047e-01 1.74631208e-01 -2.19354883e-01 1.22002028e-01
-5.75601347e-02 1.21059589e-01 -1.18516314e+00 -3.78711522e-01
3.40035021e-01 -3.73449296e-01 -5.73426962e-01 -2.55223423e-01
-1.05987608e+00 4.79817897e-01 -3.18208992e-01 2.43636146e-01
-3.28946054e-01 4.83008921e-01 5.18111944e-01 1.66208386e-01
4.83652562e-01 -6.06897213e-02 -2.21545354e-01 -2.86072165e-01
-8.80515873e-01 -1.58701893e-02 3.62683743e-01 8.95221591e-01
3.48456115e-01 5.98609269e-01 -6.11669362e-01 1.07902431e+00
3.61758143e-01 7.72239864e-01 3.65915447e-01 -1.23421562e+00
5.02960443e-01 -2.49213219e-01 5.01357466e-02 -9.09854650e-01
-2.12183818e-01 -4.10968065e-01 -3.26742947e-01 3.62426043e-01
1.01829953e-02 -3.13618779e-01 -7.06571937e-01 2.12547660e+00
3.64120185e-01 6.27972662e-01 7.46587738e-02 9.89907801e-01
7.90908694e-01 4.77290004e-01 2.16529503e-01 -5.73654711e-01
1.45297027e+00 -7.41463900e-01 -1.24319458e+00 -2.46907607e-01
1.69422358e-01 -8.75800192e-01 8.70560050e-01 5.15124500e-01
-1.12755358e+00 -6.41525149e-01 -1.00297332e+00 2.00461552e-01
1.25606403e-01 2.63621867e-01 -3.91739234e-02 1.02299738e+00
-1.28887045e+00 4.96196467e-03 -6.09065771e-01 -2.74887532e-01
1.34108007e-01 2.34181151e-01 -3.12890738e-01 1.13281071e-01
-1.04773259e+00 6.86629295e-01 -1.08734563e-01 -4.29937020e-02
-1.37588978e+00 -4.91335064e-01 -1.05075145e+00 2.42561504e-01
4.84058946e-01 -5.53945363e-01 1.39177024e+00 -1.11275983e+00
-1.36376345e+00 3.53874803e-01 -8.15191329e-01 -4.06662017e-01
2.82830149e-01 -3.96502912e-01 -6.89038813e-01 5.22040844e-01
5.16223349e-02 6.55131280e-01 1.44381773e+00 -1.42436409e+00
-6.14831090e-01 -1.25901416e-01 -2.55091399e-01 6.78135455e-01
-5.51364005e-01 4.44720328e-01 -7.86167204e-01 -9.47508991e-01
-8.06928873e-02 -6.10540628e-01 1.47443399e-01 1.12420386e-02
-2.59050280e-01 1.68969020e-01 1.05421853e+00 -6.82472408e-01
1.17539442e+00 -2.59530258e+00 -6.50903285e-02 1.95714563e-01
1.15344226e-01 1.61891878e-01 -5.00832438e-01 1.13340586e-01
-2.75096208e-01 -5.55690229e-02 2.64430165e-01 -7.34472752e-01
-2.48548821e-01 -2.36691043e-01 -1.25450298e-01 6.75074875e-01
4.60461387e-03 3.72204453e-01 -5.25982797e-01 -4.74315405e-01
3.97996575e-01 8.82495344e-01 -7.03434289e-01 1.08617470e-01
3.81349683e-01 2.37077191e-01 -5.18819019e-02 7.08061099e-01
8.40607643e-01 4.21471745e-01 -6.36408255e-02 -5.28685272e-01
1.23519458e-01 1.42446160e-01 -1.47914410e+00 1.67330325e+00
-4.48567301e-01 7.94562578e-01 7.12731361e-01 -7.41129935e-01
4.89929825e-01 6.83521926e-01 5.21196008e-01 -8.03430200e-01
-2.29750238e-02 -1.31991431e-01 4.38132286e-02 -5.86717248e-01
3.21609527e-01 8.74420628e-02 1.83270290e-01 2.18920365e-01
2.56452680e-01 -2.76187677e-02 -4.57242168e-02 3.28275800e-01
1.08316815e+00 -1.85119569e-01 2.16039032e-01 1.61146238e-01
6.44389093e-01 -6.49907947e-01 5.82723737e-01 9.50476050e-01
-6.84103847e-01 7.85019696e-01 2.29288101e-01 4.24953282e-01
-6.92780674e-01 -1.14146984e+00 1.47527814e-01 1.48770225e+00
1.99255958e-01 -7.10232794e-01 -8.95102382e-01 -3.95538360e-01
-2.51521379e-01 7.72729337e-01 -3.39628190e-01 -3.39560151e-01
-4.05915469e-01 -2.97302544e-01 7.40799963e-01 4.62870687e-01
9.03306678e-02 -8.22808146e-01 -4.48712707e-01 -7.92493224e-02
-6.32462442e-01 -1.16892815e+00 -8.45784724e-01 1.75230294e-01
-2.44806632e-01 -8.72319877e-01 -5.94396412e-01 -5.69791734e-01
2.67852426e-01 8.80457819e-01 5.25401652e-01 -4.72199678e-01
-2.29723215e-01 1.38182068e+00 -9.95005965e-02 -5.75995922e-01
-2.77400315e-01 -6.56554341e-01 4.29426491e-01 4.10558194e-01
3.49132180e-01 -6.06158853e-01 -6.47257984e-01 1.88212380e-01
-6.35908961e-01 -3.55599105e-01 3.31146121e-01 7.15307474e-01
3.86041671e-01 -4.66990136e-02 5.64686477e-01 9.90421921e-02
6.63212359e-01 -3.17388564e-01 -1.32257849e-01 -1.69527084e-02
2.15466996e-03 -4.45437700e-01 2.56201297e-01 -7.71112144e-01
-1.22918117e+00 1.75495967e-01 -5.14592156e-02 -9.04136062e-01
-3.35481465e-01 -4.54371460e-02 -6.25549674e-01 -1.22405224e-01
6.56488061e-01 2.13425204e-01 -2.37253513e-02 -3.06446970e-01
3.34017068e-01 1.01936853e+00 4.53330696e-01 -1.37637764e-01
5.38932323e-01 5.89014232e-01 -2.37709835e-01 -1.24251449e+00
-3.27752918e-01 -7.77625561e-01 -1.61163643e-01 -4.91389811e-01
7.07990527e-01 -1.47002780e+00 -7.47863352e-01 3.77207965e-01
-1.21345496e+00 4.09747846e-02 -5.45179658e-02 7.71129310e-01
-6.47010446e-01 5.86181223e-01 -4.17043328e-01 -1.29956055e+00
-1.72585532e-01 -1.41082144e+00 1.31324840e+00 -3.51745542e-03
-1.59641966e-01 -4.66341197e-01 -7.65115172e-02 4.53604192e-01
4.06449407e-01 -4.02262181e-01 2.81235129e-01 -5.31478405e-01
-2.98515052e-01 8.95227492e-02 8.82221013e-02 5.45155406e-01
3.24788123e-01 -1.95114896e-01 -1.74245107e+00 -3.29691827e-01
3.42941523e-01 -1.13126867e-01 1.02363372e+00 8.81734014e-01
1.17852688e+00 -2.78727353e-01 -3.06832939e-01 5.90045154e-01
9.76453602e-01 5.58831155e-01 5.57353556e-01 6.85820878e-02
5.33632159e-01 6.08717084e-01 5.98467886e-01 5.88716924e-01
2.29924470e-02 1.00603068e+00 5.55697381e-01 -1.68543562e-01
-5.08914649e-01 -1.04589343e-01 7.74199009e-01 4.80643600e-01
4.44362938e-01 -4.96849924e-01 -4.47824091e-01 6.57376111e-01
-1.61887825e+00 -1.19560552e+00 1.57543570e-01 2.21616697e+00
5.22199273e-01 -1.10562652e-01 1.58641607e-01 2.68974781e-01
1.03943658e+00 2.86510140e-01 -4.46836114e-01 -2.09959388e-01
-3.53643835e-01 9.65015069e-02 2.95770228e-01 7.01762021e-01
-1.45443416e+00 5.90470076e-01 6.54953337e+00 1.07469225e+00
-1.05975151e+00 3.63015622e-01 3.06829542e-01 -8.90146375e-01
-1.61628857e-01 -4.00708556e-01 -6.38723612e-01 3.14110518e-01
1.02087450e+00 -7.32664764e-02 4.54524338e-01 6.32616699e-01
9.31892037e-01 -3.23007375e-01 -1.00194728e+00 1.50818300e+00
6.73107624e-01 -7.57070780e-01 -7.60417357e-02 -2.23641954e-02
2.28923500e-01 -1.00192346e-01 3.47008348e-01 5.51952645e-02
-2.17521399e-01 -8.18819106e-01 1.03804374e+00 3.25067848e-01
7.55884588e-01 -8.08767736e-01 2.33597293e-01 -7.24035874e-02
-1.40853345e+00 -4.56938326e-01 -2.10829619e-02 2.82767028e-01
2.34983474e-01 2.62674600e-01 -6.08651757e-01 2.53278673e-01
1.06444049e+00 3.80177885e-01 -4.04955298e-01 1.15061963e+00
-1.44702882e-01 7.34992981e-01 -2.82230347e-01 3.36848468e-01
-2.90641725e-01 2.64061034e-01 1.18997455e+00 1.47474110e+00
3.43046874e-01 -6.72519356e-02 5.76340966e-03 4.67101157e-01
2.07454011e-01 2.03489929e-01 -8.57202411e-01 3.66076261e-01
5.38983166e-01 1.12674165e+00 -2.69021571e-01 -1.14500687e-01
-6.83901966e-01 9.24768984e-01 -2.06254721e-01 7.81072199e-01
-1.03429985e+00 -4.92918283e-01 1.16438532e+00 -1.37554720e-01
4.64288980e-01 -1.23250507e-01 -2.93264370e-02 -9.37388062e-01
1.03208683e-01 -1.02994716e+00 2.66227543e-01 -1.07902265e+00
-6.65967584e-01 4.26914781e-01 -1.74414031e-02 -1.41640604e+00
-1.92745522e-01 -2.55563289e-01 -6.18927360e-01 9.08270240e-01
-1.19192719e+00 -1.00180602e+00 -1.95457757e-01 8.67102742e-01
8.00480425e-01 -8.23641941e-02 5.61166525e-01 5.49135864e-01
-4.78080481e-01 9.26332355e-01 3.34547050e-02 -1.90429646e-03
1.15290177e+00 -7.91276455e-01 2.14630477e-02 1.21041632e+00
2.57655978e-01 3.83485228e-01 8.49276960e-01 -6.91000879e-01
-1.39026403e+00 -9.58461761e-01 4.52357888e-01 -1.37150779e-01
3.97053808e-01 -7.31539667e-01 -7.32076108e-01 4.72212553e-01
4.00604129e-01 1.92923769e-02 7.29314446e-01 -1.72385648e-01
-2.87157357e-01 -3.97211581e-01 -9.92511094e-01 7.93376505e-01
9.26560163e-01 -8.34386528e-01 -3.48679125e-01 5.05290516e-02
7.76417315e-01 -1.30620405e-01 -1.44728199e-01 3.52743447e-01
6.26391470e-01 -8.27711821e-01 1.11751986e+00 -3.41664374e-01
-3.70881051e-01 -5.92506647e-01 -5.76444209e-01 -1.28539050e+00
-1.93343416e-01 -8.40329587e-01 -1.42946362e-01 1.51672804e+00
2.29535729e-01 -3.13586086e-01 3.43433678e-01 9.93056372e-02
-1.89258561e-01 5.98835610e-02 -1.31272113e+00 -5.91911197e-01
-7.34957755e-01 -9.36341047e-01 -4.79205810e-02 5.52652180e-01
-5.11610694e-02 2.29324073e-01 -6.98595822e-01 4.52874005e-01
5.84163547e-01 -6.01494491e-01 6.86626613e-01 -6.28707707e-01
-2.18171194e-01 -1.78536341e-01 -4.41190034e-01 -8.24914217e-01
1.50928617e-01 -4.41413105e-01 1.69212446e-01 -1.18243897e+00
1.77087292e-01 4.30556267e-01 -5.43389678e-01 3.22518528e-01
-1.46741435e-01 2.74036467e-01 2.63724566e-01 -1.93914652e-01
-5.71083903e-01 7.53763855e-01 7.31075466e-01 -3.24304432e-01
-3.52799803e-01 -9.05131549e-02 -8.91708255e-01 6.85786903e-01
5.24680793e-01 -4.45161134e-01 -5.67168236e-01 -3.21768492e-01
-3.08941305e-01 1.55123204e-01 5.41529238e-01 -1.07610869e+00
4.98493046e-01 1.11122197e-02 4.01305854e-01 -6.02644026e-01
9.93883848e-01 -8.25043797e-01 -7.36820549e-02 6.11113869e-02
-3.98579866e-01 -1.82335839e-01 6.43711507e-01 8.96906912e-01
-2.89055437e-01 -1.85293518e-02 7.93320656e-01 1.30272016e-01
-4.74211246e-01 -1.48532957e-01 -9.14897382e-01 -6.26391247e-02
9.27104831e-01 -3.16624165e-01 -1.80082560e-01 -9.75578904e-01
-9.20935571e-01 1.13083329e-02 -1.92940608e-02 4.92055863e-01
7.84761488e-01 -1.32286382e+00 -7.67043471e-01 1.94755435e-01
1.54736172e-02 -7.68419862e-01 7.52713323e-01 9.71653819e-01
2.35820547e-01 2.85620660e-01 7.66791701e-02 -9.57322180e-01
-1.94146717e+00 5.53688824e-01 5.42075813e-01 5.53754210e-01
-4.48555499e-01 8.53341639e-01 7.39458442e-01 3.37450892e-01
8.57702792e-01 -8.56176540e-02 -3.56300473e-01 1.53329164e-01
8.69458735e-01 5.74415326e-01 1.73901841e-01 -8.84045422e-01
-5.52881777e-01 5.53292990e-01 1.02520257e-01 -6.61676109e-01
1.00711405e+00 -6.56017900e-01 5.49355865e-01 4.52403188e-01
1.04365849e+00 6.96289957e-01 -1.26765037e+00 -1.96398422e-01
-4.07591343e-01 -5.67495167e-01 3.32789034e-01 -6.77535653e-01
-9.81435955e-01 9.59051967e-01 1.14767432e+00 -1.77829683e-01
1.53285861e+00 -1.80341318e-01 1.21221259e-01 7.10892454e-02
2.58425865e-02 -1.13409066e+00 4.78193969e-01 1.60049871e-01
1.05293834e+00 -9.88231659e-01 -1.38820663e-01 -4.84392881e-01
-9.04200375e-01 8.49884391e-01 6.84658587e-01 3.90975237e-01
5.62683821e-01 5.77566087e-01 2.86031276e-01 9.54567716e-02
-8.72408330e-01 -4.32556540e-01 3.35611939e-01 1.00662792e+00
1.45707071e-01 -3.35263401e-01 3.01496625e-01 6.92142546e-01
1.11510791e-01 -5.83852589e-01 4.64211702e-01 7.29683161e-01
-4.73773092e-01 -6.31069481e-01 -7.97312677e-01 1.10190600e-01
-7.10954010e-01 -2.89030999e-01 -3.75483364e-01 3.43190491e-01
4.30343263e-02 1.67017233e+00 5.19079417e-02 -3.92599225e-01
4.58750069e-01 2.59516120e-01 2.60506392e-01 -3.76342982e-01
-5.30547619e-01 1.03048229e+00 1.27193570e-01 -8.27397466e-01
-4.86391187e-01 -8.69895816e-01 -7.24866509e-01 2.29871482e-01
-5.46072721e-01 2.61650160e-02 7.69678652e-01 5.93589246e-01
5.57979047e-01 7.44364917e-01 6.85854673e-01 -1.04498136e+00
-3.07026535e-01 -1.05088615e+00 -6.35012865e-01 2.09309831e-01
7.70370960e-01 -7.77044415e-01 -8.88533890e-01 -2.70329509e-02] | [14.439260482788086, 5.1395134925842285] |
336c69e2-9e3d-4e42-a9f8-88a18541990b | metric-type-identification-for-multi-level | 2102.00819 | null | https://arxiv.org/abs/2102.00819v1 | https://arxiv.org/pdf/2102.00819v1.pdf | Metric-Type Identification for Multi-Level Header Numerical Tables in Scientific Papers | Numerical tables are widely used to present experimental results in scientific papers. For table understanding, a metric-type is essential to discriminate numbers in the tables. We introduce a new information extraction task, metric-type identification from multi-level header numerical tables, and provide a dataset extracted from scientific papers consisting of header tables, captions, and metric-types. We then propose two joint-learning neural classification and generation schemes featuring pointer-generator-based and BERT-based models. Our results show that the joint models can handle both in-header and out-of-header metric-type identification problems. | ['Hiroya Takamura', 'Manabu Okumura', 'Hidetaka Kamigaito', 'Lya Hulliyyatus Suadaa'] | 2021-02-01 | null | https://aclanthology.org/2021.eacl-main.267 | https://aclanthology.org/2021.eacl-main.267.pdf | eacl-2021-2 | ['metric-type-identification'] | ['natural-language-processing'] | [-5.47726192e-02 -1.89096898e-01 -6.18441880e-01 -3.93876940e-01
-1.04306126e+00 -9.21951175e-01 4.27380592e-01 8.02701354e-01
8.18806961e-02 1.09826517e+00 3.36924009e-02 -9.53103602e-01
-3.01836580e-01 -1.23811936e+00 -9.41225767e-01 6.96088374e-02
-1.60949603e-01 4.99797195e-01 -2.77809322e-01 2.38551006e-01
9.55204308e-01 8.12079430e-01 -1.37545168e+00 7.41469085e-01
8.46268892e-01 1.37782025e+00 -1.36957482e-01 8.55992615e-01
-9.38412428e-01 9.17176962e-01 -1.27337968e+00 -6.55226767e-01
1.10890381e-01 -1.46890789e-01 -6.77152574e-01 -4.87508267e-01
6.79557502e-01 -1.65804118e-01 -5.81543386e-01 9.36252415e-01
3.34356695e-01 -2.08878800e-01 1.08174157e+00 -1.74896681e+00
-1.02927327e+00 1.40114725e+00 -5.07247329e-01 2.70764261e-01
6.30470932e-01 -3.43131453e-01 1.41181457e+00 -7.20592976e-01
6.03789568e-01 1.31877506e+00 7.63951182e-01 1.74824849e-01
-8.82502615e-01 -7.89266050e-01 -1.51371537e-02 4.42445159e-01
-1.29312932e+00 -6.91765696e-02 1.04427254e+00 -6.19446039e-01
7.14004099e-01 6.01724148e-01 1.49568945e-01 1.00116646e+00
3.35596025e-01 7.31550157e-01 8.36861610e-01 -3.94955844e-01
1.82689145e-01 5.00399023e-02 3.14269543e-01 5.28394639e-01
1.07780921e+00 -6.26490355e-01 -6.77166343e-01 -9.06390920e-02
1.02396715e+00 -3.94754350e-01 7.39597827e-02 -2.11496368e-01
-1.63689959e+00 5.79913855e-01 1.88983291e-01 2.48272985e-01
8.82743448e-02 -1.35359183e-01 6.10572994e-01 5.05758151e-02
-1.09143451e-01 1.00166166e+00 -6.69505060e-01 -3.14708114e-01
-8.53368521e-01 4.49280649e-01 1.17845440e+00 1.72424901e+00
2.71128565e-01 -8.54870453e-02 -7.95149803e-01 7.21354306e-01
-2.08472982e-02 4.26578432e-01 3.65301609e-01 -6.66322708e-01
1.05962014e+00 1.00357342e+00 1.14670925e-01 -1.21339703e+00
-4.75784540e-01 -4.64076877e-01 -1.15306818e+00 -4.15408641e-01
7.31993496e-01 -5.80380624e-03 -4.94260579e-01 1.40836072e+00
-1.28163695e-01 -5.50999820e-01 -3.57525125e-02 1.02549091e-01
1.55024958e+00 7.97155857e-01 -1.77488953e-01 -5.23642218e-03
1.65019798e+00 -5.28218329e-01 -1.27784383e+00 3.44394624e-01
8.17457974e-01 -6.28321230e-01 8.85810256e-01 4.26185846e-01
-1.23846042e+00 -8.36674750e-01 -1.38751447e+00 -6.84738100e-01
-1.40188217e+00 6.99657261e-01 5.24614632e-01 6.62449121e-01
-6.11633062e-01 8.18472028e-01 -2.21332282e-01 3.75966489e-01
5.13907909e-01 7.36204386e-02 -2.28864297e-01 4.12592888e-01
-1.28550589e+00 6.42100513e-01 4.10897791e-01 4.38566282e-02
9.77792293e-02 -1.03954792e+00 -1.02471352e+00 1.83434725e-01
1.42722338e-01 -4.92994279e-01 1.08074522e+00 5.84829152e-01
-8.42328131e-01 8.31090212e-01 -3.02114785e-02 -1.70966357e-01
3.73580754e-01 1.86457604e-01 -5.35219252e-01 -2.79933095e-01
1.77119032e-01 3.64896983e-01 2.37720698e-01 -1.23400927e+00
-6.43453896e-01 -2.19237268e-01 -8.93432945e-02 -1.02948107e-01
-3.33079278e-01 -2.39282876e-01 -4.39516157e-01 -9.70168173e-01
1.83934465e-01 -3.28967094e-01 2.98781484e-01 2.07766220e-01
-1.26792204e+00 -5.94380319e-01 4.64225799e-01 -8.82230937e-01
1.97378027e+00 -1.58228111e+00 -1.38430923e-01 1.39862210e-01
5.39319158e-01 -2.40448847e-01 2.33831227e-01 2.02110782e-01
-7.48645216e-02 6.09370589e-01 -6.81880489e-02 5.92761077e-02
5.49760878e-01 -2.83944398e-01 -3.21317524e-01 -1.28141567e-01
8.79185274e-02 8.16890657e-01 -7.43754983e-01 -9.37006891e-01
-9.99807194e-02 -5.40302955e-02 -2.53374904e-01 4.60303813e-01
-2.46113911e-01 -2.84259945e-01 -3.72907728e-01 1.12316227e+00
6.98323548e-01 -3.28459769e-01 1.99147165e-01 -7.07122207e-01
-1.43975914e-01 6.68258190e-01 -1.52381861e+00 1.25916457e+00
-3.89477283e-01 5.39527774e-01 -4.10196871e-01 -8.27253282e-01
1.26706505e+00 -2.25973040e-01 2.00942934e-01 -6.41424298e-01
1.54027238e-01 4.03203577e-01 8.68768767e-02 -2.70500064e-01
7.88645029e-01 8.09271395e-01 -7.02242732e-01 2.82950908e-01
-6.85649589e-02 -3.08512688e-01 1.25232041e+00 4.50987309e-01
8.45799506e-01 1.25395492e-01 3.82323384e-01 -1.82216302e-01
9.58472550e-01 -3.28093886e-01 5.25052845e-01 1.13575637e+00
1.33874431e-01 6.17076814e-01 1.37418556e+00 -5.36607563e-01
-1.16677737e+00 -1.18812490e+00 -5.73928237e-01 8.25820029e-01
-3.57959479e-01 -8.29026997e-01 -5.15915573e-01 -7.01898754e-01
5.75125277e-01 8.66803169e-01 -7.77339220e-01 -1.08793400e-01
-8.00071180e-01 -5.56085706e-01 6.32648766e-01 1.11247742e+00
2.85567939e-01 -9.59437668e-01 -3.09147239e-01 5.55409305e-02
-2.74272621e-01 -1.21292555e+00 -4.25147176e-01 6.56729043e-01
-7.07789958e-01 -1.10328424e+00 -5.85256875e-01 -8.76643479e-01
5.63067973e-01 -4.02094871e-01 1.47455060e+00 -2.97551960e-01
-4.02373850e-01 -1.87447488e-01 -8.69338773e-03 -6.75178468e-01
-2.83212721e-01 3.41602534e-01 -1.38773650e-01 -5.11373818e-01
3.81610602e-01 -2.11764738e-01 1.14685930e-01 -3.22628208e-02
-5.40581107e-01 -5.57414182e-02 6.76459730e-01 5.36195815e-01
5.47667325e-01 -1.44962773e-01 3.69134068e-01 -9.12670434e-01
7.42345273e-01 -3.35071117e-01 -8.60781491e-01 6.45542562e-01
-5.24200797e-01 5.27686715e-01 1.12292504e+00 -1.71820670e-01
-4.55558807e-01 -4.29707855e-01 7.59339407e-02 -6.47493685e-03
6.60719946e-02 5.39920151e-01 -8.91660035e-01 2.65433967e-01
4.19125289e-01 2.85607815e-01 -2.73035169e-01 -7.09124684e-01
4.66416240e-01 1.00767982e+00 9.53212440e-01 -1.06198621e+00
8.77929389e-01 -2.62744009e-01 4.24837142e-01 -3.35281670e-01
-7.65408516e-01 -8.91707651e-03 -1.06472802e+00 1.54610485e-01
5.32819986e-01 -6.65734887e-01 -9.52934682e-01 2.85949439e-01
-1.32113302e+00 4.67191860e-02 -1.15641885e-01 1.50439695e-01
-5.12993574e-01 2.04771250e-01 -9.32853699e-01 -4.00463909e-01
-2.67156661e-01 -1.16643059e+00 8.83371115e-01 2.56707549e-01
-4.84809190e-01 -1.07705402e+00 -3.23074460e-01 3.32023241e-02
1.75745726e-01 2.40918979e-01 1.88302803e+00 -9.63951170e-01
-5.94512582e-01 -4.13065135e-01 -8.30926418e-01 -2.24936649e-01
3.57364565e-01 4.00338650e-01 -4.94151354e-01 2.91541368e-01
-6.12360954e-01 -2.28708848e-01 5.25909483e-01 1.88966170e-01
2.26561022e+00 -6.42221868e-01 -5.39826155e-01 9.22984421e-01
1.17261183e+00 5.08314967e-01 4.67021972e-01 6.38017118e-01
1.07498884e+00 3.78842354e-01 1.66467085e-01 7.36164510e-01
6.37397110e-01 3.45538378e-01 -5.71943559e-02 2.12404698e-01
-1.80695653e-01 -3.92327905e-01 -3.82754683e-01 9.41746235e-01
5.18046319e-01 -4.77977157e-01 -9.38225329e-01 4.46876973e-01
-1.24304628e+00 -6.36301279e-01 -3.98273885e-01 2.13672805e+00
1.46512043e+00 6.36283815e-01 1.83196619e-01 8.04211020e-01
8.34176660e-01 4.18613702e-02 -4.34178382e-01 -7.41321683e-01
-3.00014019e-01 -1.04059046e-03 5.70500016e-01 -1.67625435e-02
-1.39369631e+00 3.36829662e-01 7.19506502e+00 8.64926040e-01
-6.10134363e-01 -7.58853436e-01 8.90756667e-01 3.71643305e-01
-3.73844624e-01 -4.41959560e-01 -1.36255085e+00 7.52120852e-01
1.01156890e+00 -6.99767768e-01 -4.43162769e-02 8.81533742e-01
-4.70996588e-01 2.73306876e-01 -1.78368819e+00 1.44081974e+00
9.86746326e-02 -1.78944206e+00 7.97177970e-01 -2.64863491e-01
2.58470744e-01 -1.19664299e+00 2.42660612e-01 5.72411716e-01
9.12321433e-02 -1.07472885e+00 8.00471008e-01 5.26356757e-01
1.31817913e+00 -7.94166982e-01 7.93470800e-01 -2.54336149e-01
-1.33698750e+00 -1.56383943e-02 -4.57600415e-01 2.46200878e-02
-3.81783158e-01 8.56424749e-01 -3.66963029e-01 7.17016518e-01
5.63066065e-01 8.72853279e-01 -1.17663682e+00 1.22153974e+00
-6.01370744e-02 1.79977909e-01 -8.42128769e-02 -4.93518829e-01
-1.91316932e-01 -1.15696356e-01 9.70472991e-02 1.48795390e+00
4.38912481e-01 -3.58125627e-01 -1.63445279e-01 1.34735096e+00
-7.36167610e-01 5.68585582e-02 -4.80282396e-01 -4.61015165e-01
1.02811718e+00 1.34867847e+00 -1.04975951e+00 -5.92168927e-01
-2.62271166e-01 1.99757710e-01 2.94825017e-01 3.21531966e-02
-8.17838907e-01 -1.41209352e+00 2.35118508e-01 -1.66041166e-01
2.26344511e-01 -2.09097296e-01 -1.42689085e+00 -9.56533194e-01
3.73824030e-01 -7.98225045e-01 6.43360436e-01 -7.93609917e-01
-1.38663757e+00 3.37364882e-01 2.83556938e-01 -1.34204304e+00
-3.10387760e-01 -1.19968343e+00 -3.43206197e-01 8.26466084e-01
-1.05456424e+00 -7.19939172e-01 -2.30284616e-01 -3.47464457e-02
1.48224056e-01 -3.05007935e-01 8.01890910e-01 3.39334637e-01
-9.69257236e-01 1.32263887e+00 3.16072345e-01 1.03369677e+00
5.99599659e-01 -2.01541901e+00 7.11098552e-01 6.18418872e-01
7.63757303e-02 1.00882638e+00 4.66054827e-01 -7.88574696e-01
-1.86591971e+00 -1.00129831e+00 1.22334456e+00 -7.20609963e-01
8.18069220e-01 -8.86706948e-01 -9.57747936e-01 5.17120063e-01
-1.42456889e-01 -3.48870456e-02 6.55671060e-01 2.06643790e-01
-8.44683647e-01 -4.03358579e-01 -1.04294086e+00 5.57562709e-01
9.73865330e-01 -5.15881836e-01 -7.43590832e-01 5.24828911e-01
5.52618682e-01 -7.56502867e-01 -1.03836656e+00 2.22281694e-01
5.26232541e-01 -4.22428459e-01 1.18490648e+00 -4.05225247e-01
1.00094008e+00 -2.06500977e-01 -6.07871115e-02 -9.85261500e-01
-5.05408823e-01 -5.16617060e-01 -8.85213554e-01 1.43885374e+00
7.47985125e-01 -4.91393693e-02 8.63312602e-01 4.59076852e-01
-2.27025244e-02 -7.37853765e-01 -6.37294114e-01 -1.04486930e+00
6.07966006e-01 -1.40297189e-01 1.11336446e+00 8.67410183e-01
2.70893395e-01 2.81270683e-01 1.39398158e-01 -4.19538558e-01
7.55215406e-01 4.36922610e-01 6.20613933e-01 -1.42036426e+00
2.13005081e-01 -1.19217801e+00 -3.16824079e-01 -8.76903892e-01
-3.37626375e-02 -1.05406857e+00 -2.40998909e-01 -1.92421985e+00
-3.81251089e-02 -4.58987862e-01 -4.18881208e-01 4.33395177e-01
-2.78143674e-01 1.40136397e-02 1.78093612e-02 -1.78899076e-02
-5.12757361e-01 1.46989867e-01 1.11283708e+00 -4.31707799e-01
8.91052932e-02 -2.00123638e-01 -8.35605264e-01 2.37780437e-01
5.65550685e-01 -3.23654473e-01 -2.38728127e-03 -5.01735322e-02
4.81230706e-01 1.88829780e-01 -9.04897526e-02 -1.30707181e+00
2.82619774e-01 -1.47166312e-01 1.05368459e+00 -1.41721594e+00
-2.25759268e-01 -1.84781983e-01 -6.51539862e-01 3.56645912e-01
-9.93268132e-01 6.26409888e-01 2.81733811e-01 1.29587755e-01
2.59176251e-02 -1.81748271e-01 2.87729174e-01 -1.51594400e-01
-3.62354249e-01 5.14287278e-02 -3.33755255e-01 2.77665734e-01
5.17297626e-01 -1.58617899e-01 -5.63457012e-01 5.49621955e-02
-3.18557501e-01 2.19546482e-01 -5.28180078e-02 5.95328987e-01
4.54157978e-01 -1.54141653e+00 -6.26892328e-01 2.95357913e-01
3.58041525e-01 -1.42488154e-02 -2.91289598e-01 1.77647412e-01
-7.08020687e-01 9.07340407e-01 -5.13455629e-01 -2.52001524e-01
-6.82111263e-01 9.24606800e-01 -1.91213772e-01 -4.29819524e-01
-2.78717160e-01 9.84028935e-01 -4.92688000e-01 -7.59702623e-01
8.70898724e-01 -9.52539861e-01 -6.00138843e-01 4.22945857e-01
6.18578970e-01 5.64300656e-01 4.51782525e-01 -2.12840773e-02
-3.54162753e-01 3.35604787e-01 -2.85792589e-01 1.03608742e-01
6.78471267e-01 3.37877750e-01 -3.29850316e-01 1.00995529e+00
1.40200329e+00 1.51791751e-01 -4.70191360e-01 -8.52848440e-02
3.04792523e-01 -2.19403148e-01 -2.24148259e-01 -1.13005376e+00
-5.58460116e-01 8.47042024e-01 1.80185586e-02 1.80523291e-01
5.98281741e-01 -5.25343478e-01 7.90708721e-01 7.79129207e-01
2.20102638e-01 -1.00248516e+00 -1.15055971e-01 5.20090163e-01
9.67747152e-01 -9.16463852e-01 2.90617287e-01 -5.60856164e-01
-4.56366092e-02 1.70010889e+00 8.36885035e-01 2.55767852e-01
5.98224163e-01 8.62846911e-01 -1.42765120e-01 3.21772337e-01
-7.66653776e-01 2.63810605e-01 6.97912693e-01 6.94261491e-01
9.72200096e-01 -2.76810136e-02 -3.21297765e-01 1.10235894e+00
-8.78730357e-01 -2.31459036e-01 9.97021854e-01 1.31478381e+00
-5.50742932e-02 -1.09402597e+00 -7.26138830e-01 1.11759198e+00
-3.78295273e-01 -1.76507711e-01 -6.49404824e-01 8.44981015e-01
-1.31230325e-01 7.36347079e-01 5.56734800e-01 -4.31218117e-01
2.12036982e-01 2.81964928e-01 6.21490717e-01 -2.69610077e-01
-3.79048407e-01 -4.25487667e-01 -4.10041632e-03 -2.78906614e-01
3.32444340e-01 -5.07647514e-01 -1.19526052e+00 -3.47104549e-01
1.93538696e-01 4.86114949e-01 7.55632639e-01 5.15246332e-01
3.30520213e-01 1.05104494e+00 4.58109528e-01 -2.53393978e-01
-6.44261777e-01 -1.06074739e+00 -5.90261698e-01 3.01974833e-01
3.80812079e-01 -7.88785160e-01 -4.36791688e-01 -1.16219960e-01] | [11.647747039794922, 3.107576847076416] |
44f58288-f3e4-4c43-b011-f609a9746a49 | stereo-matching-with-cost-volume-based-sparse | 2201.11937 | null | https://arxiv.org/abs/2201.11937v1 | https://arxiv.org/pdf/2201.11937v1.pdf | Stereo Matching with Cost Volume based Sparse Disparity Propagation | Stereo matching is crucial for binocular stereo vision. Existing methods mainly focus on simple disparity map fusion to improve stereo matching, which require multiple dense or sparse disparity maps. In this paper, we propose a simple yet novel scheme, termed feature disparity propagation, to improve general stereo matching based on matching cost volume and sparse matching feature points. Specifically, our scheme first calculates a reliable sparse disparity map by local feature matching, and then refines the disparity map by propagating reliable disparities to neighboring pixels in the matching cost domain. In addition, considering the gradient and multi-scale information of local disparity regions, we present a $\rho$-Census cost measure based on the well-known AD-Census, which guarantees the robustness of cost volume even without the cost aggregation step. Extensive experiments on Middlebury stereo benchmark V3 demonstrate that our scheme achieves promising performance comparable to state-of-the-art methods. | ['Xiaojiang Peng', 'Wei Xue'] | 2022-01-28 | null | null | null | null | ['stereo-matching-1'] | ['computer-vision'] | [ 1.96434543e-01 -5.41361034e-01 -6.42390251e-02 -5.17956197e-01
-6.56974792e-01 7.50509789e-03 4.31663692e-01 2.21338533e-02
-3.23724270e-01 6.76007688e-01 3.29991192e-01 9.98270363e-02
6.21009544e-02 -1.01200187e+00 -6.40321970e-01 -5.41392982e-01
3.58947337e-01 -2.66671814e-02 6.39927089e-01 -1.68817028e-01
8.26180816e-01 2.91954607e-01 -2.08654881e+00 1.81450546e-01
1.38607371e+00 1.23968732e+00 3.99594963e-01 2.50270903e-01
-1.46922618e-01 4.84858274e-01 -1.55401781e-01 -2.58687496e-01
7.35276103e-01 -2.20515683e-01 -4.65499699e-01 3.95589918e-02
9.56405878e-01 -6.92542493e-01 -4.50764269e-01 1.42352486e+00
6.56040788e-01 -2.05345750e-02 2.05921069e-01 -8.48533154e-01
-1.87297672e-01 -4.97657098e-02 -9.09886539e-01 1.55829534e-01
7.79091120e-01 3.15708607e-01 7.55922019e-01 -1.11279440e+00
6.33049548e-01 1.25874579e+00 7.52101660e-01 1.34599552e-01
-1.12651169e+00 -8.58280241e-01 1.05028413e-01 2.55001724e-01
-1.59893727e+00 -3.99100453e-01 8.41989517e-01 -5.18797219e-01
9.10374403e-01 8.86670426e-02 7.87572801e-01 3.10104378e-02
3.16248119e-01 4.69662994e-01 1.38167572e+00 -2.83140808e-01
8.60616844e-03 -3.96406889e-01 -2.77040657e-02 7.50662804e-01
2.28808254e-01 7.69594073e-01 -7.95097232e-01 -2.20273390e-01
1.08399141e+00 1.93596348e-01 -6.56303704e-01 -5.52266777e-01
-1.40056491e+00 7.62580931e-01 7.57303834e-01 3.32229622e-02
-4.25075471e-01 4.30327319e-02 1.28195852e-01 1.28898665e-01
5.16002953e-01 -1.28544837e-01 -1.53566688e-01 8.49161074e-02
-9.15547311e-01 4.45565820e-01 2.53251255e-01 9.90884364e-01
1.63233328e+00 -2.99327105e-01 -1.63174033e-01 7.46968031e-01
2.33018994e-01 5.72608292e-01 2.89307058e-01 -1.22502244e+00
8.33593786e-01 8.31479073e-01 1.91329662e-02 -1.16415024e+00
-1.57594144e-01 1.80843491e-02 -9.61829841e-01 6.42899215e-01
1.81792200e-01 1.84457943e-01 -8.67656469e-01 1.29439938e+00
5.22413433e-01 5.31630039e-01 -2.34717011e-01 1.07149327e+00
9.34509337e-01 4.26525891e-01 -3.85912389e-01 -6.91913664e-02
7.42641985e-01 -7.73917615e-01 -3.33742499e-01 3.50939929e-02
1.91132903e-01 -1.10670638e+00 6.72232270e-01 6.59222156e-02
-1.36873996e+00 -8.82921278e-01 -1.07588971e+00 -4.62208331e-01
-5.33678569e-02 -5.17549276e-01 7.61962712e-01 4.20825064e-01
-1.24964285e+00 5.16890228e-01 -4.59586471e-01 -4.21296544e-02
3.23327988e-01 5.17647147e-01 -1.05962344e-01 -5.39338112e-01
-9.21221316e-01 6.29708886e-01 2.28107437e-01 -8.63728970e-02
-3.38569552e-01 -7.61125743e-01 -1.33489776e+00 -1.75285786e-01
-1.67007804e-01 -1.28015292e+00 8.51002157e-01 -3.55905056e-01
-1.36306930e+00 1.00302255e+00 -8.72143149e-01 -2.22768664e-01
5.49186587e-01 1.80966947e-02 7.30745718e-02 1.23092480e-01
3.48479331e-01 1.09190166e+00 7.32009828e-01 -1.11052084e+00
-1.24063694e+00 -5.35171092e-01 1.01526588e-01 5.10946035e-01
1.69329718e-01 -2.72547305e-01 -6.84650540e-01 -4.95502293e-01
7.56630898e-01 -3.18359345e-01 -4.98985171e-01 1.89934582e-01
-2.68916756e-01 -3.52332890e-02 4.38526720e-01 -3.68656546e-01
1.25910878e+00 -2.10745835e+00 6.24841712e-02 4.92456079e-01
4.16734278e-01 -2.18001187e-01 7.64002055e-02 -1.97435066e-01
2.84892291e-01 -3.99694026e-01 -3.43642890e-01 -2.95338660e-01
-2.06811279e-01 -1.14893593e-01 -1.21120840e-01 6.32388711e-01
3.00883222e-02 7.50606894e-01 -9.30107653e-01 -7.13782609e-01
8.08719039e-01 5.79040051e-01 -8.69258642e-01 4.00998332e-02
3.09595972e-01 5.79553306e-01 -4.10898060e-01 9.90133941e-01
1.34478319e+00 3.04778293e-02 -5.17312586e-01 -3.50278705e-01
-6.16880357e-01 2.50680864e-01 -1.37774730e+00 2.15917134e+00
-3.38514656e-01 5.21948099e-01 -1.44165680e-01 -5.68039000e-01
1.06086183e+00 -3.67620558e-01 5.60698211e-01 -1.02594936e+00
1.21773079e-01 5.85649192e-01 -3.92185688e-01 1.39799058e-01
7.92163312e-01 1.21302389e-01 1.04643762e-01 -7.57530108e-02
-4.13770825e-01 -6.46521330e-01 4.74652462e-02 -2.29021356e-01
7.68396378e-01 1.28192276e-01 3.66511822e-01 -4.02909756e-01
1.00519454e+00 -1.94235202e-02 7.25327194e-01 4.12680179e-01
-2.60755986e-01 1.13001072e+00 -1.31060243e-01 -4.37260628e-01
-9.73164201e-01 -1.19306684e+00 -5.67001998e-01 2.54519135e-01
1.12820268e+00 -1.51541561e-01 -4.97284442e-01 -4.39381786e-02
3.40874672e-01 -2.84626693e-01 -2.79353738e-01 3.07490200e-01
-7.27289200e-01 -4.01599318e-01 -1.54461503e-01 3.55426937e-01
1.10326123e+00 -5.22835255e-01 -5.65297008e-01 2.94300199e-01
-2.93419808e-01 -9.05597746e-01 -8.00110042e-01 -3.78726006e-01
-1.14734197e+00 -1.05293572e+00 -1.11750638e+00 -1.12844265e+00
7.47021019e-01 6.96175754e-01 1.11366141e+00 2.30797797e-01
-3.69262189e-01 -9.91266072e-02 1.91892967e-01 2.59968713e-02
3.07927877e-01 -1.70009062e-01 -2.48841569e-01 -1.06171496e-01
5.78702331e-01 -6.51377082e-01 -1.23568940e+00 6.14009023e-01
-5.66317737e-01 3.17077339e-01 3.17650944e-01 8.75046730e-01
1.04552138e+00 -2.08348438e-01 -1.25863850e-01 -3.78686845e-01
2.06820488e-01 -5.30292355e-02 -1.14846969e+00 -1.26696408e-01
-5.82990468e-01 3.02227978e-02 5.62446974e-02 7.89764225e-02
-9.60885644e-01 2.08575577e-01 -1.89823568e-01 -3.76133859e-01
9.13103968e-02 5.33125140e-02 -6.81695566e-02 -7.07564890e-01
3.59486073e-01 3.34253430e-01 -1.02619864e-01 -2.77564645e-01
1.71528175e-01 4.56576437e-01 8.69877934e-01 -4.44917023e-01
9.33779359e-01 1.05613065e+00 2.54446208e-01 -3.95996690e-01
-3.17530483e-01 -9.41185534e-01 -5.98032892e-01 -7.92147517e-02
7.33068705e-01 -1.37240660e+00 -1.03878653e+00 8.09019804e-01
-1.19028330e+00 1.35151431e-01 -1.57686755e-01 6.24354959e-01
-6.96235597e-01 5.78788817e-01 -5.89564860e-01 -3.02087933e-01
-2.07602605e-01 -1.32666802e+00 1.32629764e+00 5.33958733e-01
1.46005750e-01 -9.06680942e-01 2.17917800e-01 2.24217668e-01
5.56704044e-01 3.03261280e-01 4.57370579e-01 6.02056921e-01
-1.41270566e+00 5.17957583e-02 -8.13479543e-01 8.81628692e-02
2.07995296e-01 -3.01290572e-01 -1.10191905e+00 -6.59648106e-02
8.56786501e-03 1.64111946e-02 1.14276338e+00 8.58802557e-01
9.84797418e-01 4.04287726e-01 -3.01271945e-01 1.30659437e+00
1.77163208e+00 7.56069124e-02 9.54345524e-01 5.49816310e-01
7.90688813e-01 6.17453635e-01 8.96842420e-01 5.70843279e-01
7.43345499e-01 7.59416282e-01 2.90181607e-01 -4.41491306e-01
-3.05978030e-01 -2.98275232e-01 3.45324725e-02 7.61321366e-01
-2.17772305e-01 5.00607908e-01 -6.36347592e-01 3.71692836e-01
-1.80732429e+00 -9.12979484e-01 -2.31816247e-01 2.53113198e+00
7.95524955e-01 8.99716690e-02 -2.22639620e-01 1.83383122e-01
9.85953033e-01 3.55346709e-01 -5.49070299e-01 -1.03242621e-01
-4.98208225e-01 4.11997139e-01 7.14466095e-01 1.05752063e+00
-9.55711126e-01 7.74793267e-01 6.32376432e+00 9.15467680e-01
-1.00501919e+00 -7.97737166e-02 5.85559130e-01 8.49329084e-02
-6.58605933e-01 1.64159760e-01 -1.12020755e+00 6.25630558e-01
-1.48689738e-02 -5.56241095e-01 2.05916733e-01 6.24013364e-01
1.59687564e-01 -6.78987265e-01 -9.80723739e-01 1.55616713e+00
9.25064534e-02 -1.60445559e+00 6.48135170e-02 2.00544134e-01
1.19469368e+00 2.11635947e-01 9.41893533e-02 -1.28300592e-01
1.26091585e-01 -7.90729642e-01 3.67190540e-01 3.32773745e-01
9.66217637e-01 -7.76032329e-01 7.25840509e-01 8.69880915e-02
-1.77791214e+00 2.82207727e-01 -6.41287684e-01 -2.34113708e-01
2.69704908e-01 8.16900492e-01 -4.32743430e-02 7.14419544e-01
7.60201812e-01 1.22601807e+00 -4.57568467e-01 1.55141449e+00
9.13696289e-02 -5.21606386e-01 -3.71803492e-01 4.05261964e-01
9.16881040e-02 -4.06964421e-01 2.71863818e-01 8.91431093e-01
4.89405394e-01 9.23178419e-02 1.93358555e-01 7.95324564e-01
3.70330662e-02 7.96677917e-02 -7.96873510e-01 1.05919659e+00
4.59420860e-01 7.04705238e-01 -2.84402549e-01 -4.51074183e-01
-7.00005829e-01 1.07087946e+00 1.23209991e-01 2.47348905e-01
-4.69008386e-01 -5.25794506e-01 1.01384473e+00 6.66938201e-02
1.87182814e-01 -5.99901304e-02 -8.60678852e-01 -1.36960173e+00
3.35980147e-01 -3.64167154e-01 2.41776839e-01 -6.73641980e-01
-1.15888023e+00 4.72228080e-01 -1.69378191e-01 -1.96480310e+00
-3.03794265e-01 -2.30879769e-01 -5.87508917e-01 1.33807850e+00
-2.43339348e+00 -6.82347417e-01 -8.52809787e-01 1.18629467e+00
6.02941513e-01 9.84081030e-02 2.13733077e-01 5.38142920e-01
-5.80131682e-03 3.86602759e-01 -3.22317593e-02 -2.91078508e-01
5.96965253e-01 -7.80261219e-01 5.56671441e-01 8.30918968e-01
-4.74426597e-01 4.12916034e-01 3.34431231e-01 -5.40997148e-01
-1.02046919e+00 -8.61973882e-01 1.20137393e+00 -1.20295733e-01
2.89561957e-01 3.94303724e-02 -7.04562783e-01 9.91944596e-02
1.92777719e-02 1.29835546e-01 9.82726067e-02 -3.22213113e-01
-2.35407054e-01 -3.62377137e-01 -1.37265420e+00 4.52704400e-01
1.53067315e+00 -6.51288688e-01 -5.90277433e-01 -8.16060379e-02
5.89121997e-01 -8.36985946e-01 -6.46478832e-01 7.51952648e-01
6.43927872e-01 -1.72436571e+00 1.34025443e+00 3.69136661e-01
4.13219631e-01 -7.12404728e-01 -4.04798776e-01 -9.54809844e-01
-3.51335377e-01 -5.49106717e-01 3.08632225e-01 9.29554105e-01
8.04240908e-03 -9.19418812e-01 9.54993129e-01 5.77001095e-01
-3.14416736e-01 -5.16587794e-01 -1.24500358e+00 -6.65294349e-01
-2.57924318e-01 -2.77268082e-01 7.23466218e-01 7.75641620e-01
-1.05024509e-01 -2.89770573e-01 -7.43792951e-02 1.87186852e-01
1.18895268e+00 6.12810791e-01 9.50326324e-01 -1.10805500e+00
-4.62312661e-02 -6.99682832e-01 -9.68267441e-01 -1.76108241e+00
-2.34894119e-02 -5.88717699e-01 1.47479802e-01 -1.49426305e+00
3.68343115e-01 -6.28706813e-01 -1.91802815e-01 -2.62527704e-01
-4.30948704e-01 5.97745538e-01 1.71436518e-02 3.69827658e-01
-2.79869050e-01 5.38014948e-01 1.49191976e+00 -1.30305737e-01
-3.54648262e-01 -2.22656161e-01 -2.77794778e-01 8.59204471e-01
4.48886186e-01 -9.53461751e-02 -2.48272553e-01 -7.00974882e-01
-9.87446159e-02 -2.75584087e-02 3.17622244e-01 -1.21905148e+00
5.49122930e-01 -2.50175864e-01 4.09912944e-01 -1.12336981e+00
5.67525804e-01 -7.09045887e-01 5.59603013e-02 6.14552140e-01
2.29868621e-01 5.20549268e-02 -1.01525337e-02 2.99354315e-01
-8.78751159e-01 2.61741400e-01 9.51394498e-01 -2.84433868e-02
-1.23738682e+00 6.77489161e-01 3.64713728e-01 3.72111164e-02
9.95029330e-01 -7.53904879e-01 -4.24806714e-01 -1.25385299e-01
-1.11248009e-01 4.61848557e-01 1.08730197e+00 1.16878554e-01
1.11149263e+00 -1.57440996e+00 -8.09268832e-01 6.73967719e-01
3.37055832e-01 3.11973870e-01 3.91123384e-01 7.09591210e-01
-9.36212361e-01 4.77253914e-01 -4.07090753e-01 -1.18545616e+00
-1.23249769e+00 2.76254147e-01 3.51214051e-01 6.62379712e-02
-6.71198487e-01 8.92895460e-01 6.42023921e-01 -1.46046979e-02
2.52500534e-01 -6.36359036e-01 -2.06894223e-02 -2.96721876e-01
6.03256404e-01 4.10093516e-01 -1.63901635e-02 -6.54814005e-01
-3.78308296e-01 1.58588862e+00 1.59790635e-01 -9.68728401e-03
7.81959832e-01 -5.75320661e-01 -7.40570128e-02 2.81870123e-02
1.39610696e+00 6.39741719e-02 -1.42873299e+00 -4.58079726e-01
-3.50286782e-01 -1.50434399e+00 1.75872743e-01 1.26204342e-01
-1.18483031e+00 7.73032844e-01 7.79728174e-01 -3.81845474e-01
1.41546917e+00 -1.50563151e-01 1.11106157e+00 -7.69961104e-02
8.94832611e-01 -8.79478514e-01 -3.61704111e-01 4.68625218e-01
6.85761690e-01 -1.68293417e+00 2.03259498e-01 -8.35456371e-01
-1.20310895e-01 9.22751904e-01 7.31735408e-01 -2.79212058e-01
8.03667605e-01 1.76891431e-01 1.30225688e-01 -8.74484610e-03
-2.96696752e-01 -6.12818837e-01 3.05572599e-01 6.87939346e-01
2.60568887e-01 -2.38309368e-01 -5.77329338e-01 -3.39052856e-01
-1.51533678e-01 2.21771806e-01 9.17067379e-02 8.94991994e-01
-7.23345160e-01 -1.02565181e+00 -5.09540856e-01 2.78097361e-01
-8.17557275e-02 -4.94476646e-01 1.38322309e-01 4.85904723e-01
3.10406566e-01 9.02846873e-01 4.34606314e-01 -4.62468237e-01
4.98082042e-01 -6.25775516e-01 5.64408779e-01 -4.33440626e-01
-4.04762864e-01 9.60077494e-02 -2.36899123e-01 -1.07207203e+00
-8.37391555e-01 -5.47341287e-01 -1.13664317e+00 -7.39004254e-01
-6.07305542e-02 -3.00878465e-01 5.41460991e-01 4.94654030e-01
3.09417337e-01 -1.89669862e-01 1.03078139e+00 -1.36139727e+00
7.25973621e-02 -4.98002112e-01 -5.42487144e-01 5.27508736e-01
6.06717467e-01 -6.68939292e-01 -5.91606200e-01 -6.99113682e-03] | [9.067461013793945, -2.3990566730499268] |
d1a59858-dccb-4dfd-acf0-d960134f9d75 | unsupervised-domain-adaptation-with-2 | 2106.08752 | null | https://arxiv.org/abs/2106.08752v1 | https://arxiv.org/pdf/2106.08752v1.pdf | Unsupervised Domain Adaptation with Variational Approximation for Cardiac Segmentation | Unsupervised domain adaptation is useful in medical image segmentation. Particularly, when ground truths of the target images are not available, domain adaptation can train a target-specific model by utilizing the existing labeled images from other modalities. Most of the reported works mapped images of both the source and target domains into a common latent feature space, and then reduced their discrepancy either implicitly with adversarial training or explicitly by directly minimizing a discrepancy metric. In this work, we propose a new framework, where the latent features of both domains are driven towards a common and parameterized variational form, whose conditional distribution given the image is Gaussian. This is achieved by two networks based on variational auto-encoders (VAEs) and a regularization for this variational approximation. Both of the VAEs, each for one domain, contain a segmentation module, where the source segmentation is trained in a supervised manner, while the target one is trained unsupervisedly. We validated the proposed domain adaptation method using two cardiac segmentation tasks, i.e., the cross-modality (CT and MR) whole heart segmentation and the cross-sequence cardiac MR segmentation. Results show that the proposed method achieved better accuracies compared to two state-of-the-art approaches and demonstrated good potential for cardiac segmentation. Furthermore, the proposed explicit regularization was shown to be effective and efficient in narrowing down the distribution gap between domains, which is useful for unsupervised domain adaptation. Our code and data has been released via https://zmiclab.github.io/projects.html. | ['Xiahai Zhuang', 'Fuping Wu'] | 2021-06-16 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 4.22318399e-01 3.38551253e-01 -2.61825919e-01 -4.60608304e-01
-1.01230991e+00 -5.22287786e-01 3.93743098e-01 -1.27418667e-01
-4.74967122e-01 8.87751937e-01 -1.35899216e-01 -6.56162053e-02
1.77078143e-01 -7.01013863e-01 -7.56025732e-01 -1.03686452e+00
2.36052766e-01 7.58841574e-01 3.30243409e-01 2.21673757e-01
-1.33736700e-01 2.63167828e-01 -9.19186652e-01 1.09619752e-01
1.17463756e+00 7.00831175e-01 2.79961199e-01 4.93145823e-01
-2.28073541e-02 5.25876284e-01 -4.34292376e-01 -1.74436465e-01
2.05482468e-01 -7.00083137e-01 -8.72537851e-01 3.68012071e-01
-2.55956687e-02 -1.18802860e-01 -2.40319699e-01 1.09550309e+00
4.49865103e-01 1.38221890e-01 9.62072134e-01 -1.12321317e+00
-6.01935267e-01 3.66479516e-01 -4.15399730e-01 3.93927470e-02
-1.60981134e-01 1.04525872e-01 5.55367589e-01 -5.38118422e-01
8.02840114e-01 8.27125490e-01 4.04799104e-01 8.97744596e-01
-1.54433036e+00 -7.11720288e-01 -1.31274372e-01 -1.34734765e-01
-1.26992142e+00 -1.29221186e-01 1.02000248e+00 -8.83173943e-01
4.07589853e-01 -2.35817105e-01 4.89278853e-01 1.14949858e+00
1.60362914e-01 7.16794908e-01 1.26330829e+00 -3.46235693e-01
2.97534764e-01 5.37968218e-01 3.93385179e-02 5.01173377e-01
-5.21516092e-02 2.07487047e-01 2.58202627e-02 -3.07492226e-01
9.87403393e-01 -1.61622331e-01 -2.99199015e-01 -9.18987393e-01
-1.30920351e+00 1.07013524e+00 3.60077143e-01 5.16809225e-01
-4.31882113e-01 -2.17859402e-01 5.86104453e-01 -6.57142326e-02
5.23511767e-01 1.44230008e-01 -3.84418547e-01 1.90823153e-01
-1.13476896e+00 2.15038825e-02 6.47872567e-01 8.13319504e-01
7.48694599e-01 4.05745469e-02 -2.60096908e-01 9.02608454e-01
5.14353037e-01 4.81479377e-01 7.12930858e-01 -9.83683884e-01
3.09921592e-01 4.31845665e-01 -4.92922813e-02 -6.43495440e-01
-2.55797952e-01 -3.59360009e-01 -9.93014872e-01 1.40916780e-01
6.04711711e-01 -2.90076405e-01 -1.22364855e+00 2.07839298e+00
5.27453482e-01 4.00893390e-01 3.31575781e-01 9.96601999e-01
7.65515566e-01 5.67824244e-01 4.24978912e-01 -3.00497830e-01
1.14668858e+00 -8.77952218e-01 -7.22818613e-01 -1.10997871e-01
5.16572058e-01 -7.43059337e-01 9.80603278e-01 1.81831166e-01
-1.01579916e+00 -6.47240460e-01 -1.12506199e+00 9.84631702e-02
-6.58811852e-02 1.41815916e-01 1.56741574e-01 6.65733695e-01
-8.50189984e-01 4.79796857e-01 -1.23370063e+00 -2.12418705e-01
5.78412831e-01 3.08797836e-01 -2.97510922e-01 -9.99553688e-03
-1.35593867e+00 7.02861488e-01 5.79926848e-01 -2.53195584e-01
-1.17143488e+00 -6.64933145e-01 -9.35831487e-01 -2.02087671e-01
4.70350049e-02 -8.22400868e-01 1.01092982e+00 -1.21876371e+00
-1.58015347e+00 1.20041597e+00 4.73236963e-02 -5.45271277e-01
6.65554464e-01 3.85698006e-02 -3.09637934e-01 4.42884505e-01
3.78767937e-01 6.77543461e-01 1.19443572e+00 -1.35972583e+00
-1.71298280e-01 -3.49159807e-01 -2.52076000e-01 1.26193523e-01
-1.75657958e-01 -3.12864244e-01 -5.37939191e-01 -6.62622213e-01
7.19008446e-02 -9.72317159e-01 -3.10009092e-01 -1.48811907e-01
-2.95617789e-01 2.20854357e-02 6.91140532e-01 -8.74583900e-01
9.84564483e-01 -2.18513131e+00 4.28690404e-01 2.39437580e-01
1.20981254e-01 3.04541856e-01 8.64915028e-02 5.78784682e-02
-2.07728311e-01 9.85113010e-02 -8.77622128e-01 -3.18102002e-01
-1.77759707e-01 5.16705751e-01 -2.95324195e-02 6.97010100e-01
1.33171633e-01 6.94229364e-01 -9.11912799e-01 -9.20298636e-01
3.83791864e-01 6.10339224e-01 -5.35905719e-01 4.84058648e-01
-1.66982651e-01 1.37383330e+00 -5.40537119e-01 2.74952710e-01
7.70378470e-01 -3.33899766e-01 3.34621906e-01 -1.63188607e-01
3.12356889e-01 -1.34483844e-01 -1.03413701e+00 2.07714415e+00
-2.70606846e-01 1.40001461e-01 4.83944006e-02 -1.41366136e+00
9.44099367e-01 6.12422049e-01 8.80121052e-01 -3.75150084e-01
1.81939259e-01 3.34610581e-01 -1.02898017e-01 -5.01360536e-01
-1.01608887e-01 -5.68250358e-01 -2.41377763e-03 2.30347782e-01
4.54510123e-01 -2.82268763e-01 1.51928455e-01 1.06330857e-01
5.43047369e-01 4.15510088e-01 3.63451600e-01 -2.72865027e-01
9.18215096e-01 1.40729189e-01 7.79571354e-01 5.42690337e-01
-3.19923043e-01 7.69341290e-01 4.97832865e-01 -1.04210131e-01
-1.15119410e+00 -1.41832912e+00 -5.14878392e-01 6.24833941e-01
1.42578021e-01 9.48724374e-02 -1.04022801e+00 -8.91345084e-01
-2.21074775e-01 7.16227651e-01 -5.71052372e-01 -1.58197448e-01
-5.38658798e-01 -6.07047260e-01 6.72906458e-01 5.43717504e-01
5.61410427e-01 -1.03933144e+00 -4.69188005e-01 2.87987500e-01
-3.42451662e-01 -1.17761993e+00 -6.37512207e-01 4.94136363e-02
-1.22500682e+00 -1.09568369e+00 -1.13066983e+00 -8.56105745e-01
7.82504022e-01 -5.31249523e-01 1.06450117e+00 -3.17201555e-01
-7.78382272e-02 4.81900096e-01 -2.31752232e-01 -1.68823719e-01
-7.35701799e-01 1.32502899e-01 -4.75022057e-03 2.67508119e-01
2.46745244e-01 -6.34893477e-01 -7.04769731e-01 3.19571048e-01
-1.05799437e+00 -4.12691832e-02 4.91542906e-01 1.25781310e+00
9.76744831e-01 -3.48986387e-01 5.33457518e-01 -1.26401699e+00
3.53292137e-01 -5.70350468e-01 -5.14740109e-01 1.01821162e-02
-6.31865561e-01 -5.91514865e-03 5.27159572e-01 -4.23694640e-01
-1.24702978e+00 3.70040983e-01 -2.65663326e-01 -7.21996725e-01
-4.48776841e-01 4.35611397e-01 -1.40688598e-01 1.87753871e-01
7.20629156e-01 3.66809607e-01 3.22513312e-01 -3.11003536e-01
2.98408329e-01 5.86998224e-01 6.07764125e-01 -7.75377572e-01
7.09170282e-01 4.93130744e-01 -1.92084834e-01 -6.72905982e-01
-5.57381094e-01 -4.76920635e-01 -1.02825475e+00 -6.77001849e-03
1.31158650e+00 -9.29082751e-01 1.44349076e-02 5.51903844e-01
-9.97830033e-01 -5.38428962e-01 -5.28348565e-01 7.49999881e-01
-8.14306676e-01 5.55117428e-01 -5.42637765e-01 -4.18380231e-01
-1.90685153e-01 -1.38094831e+00 7.83298194e-01 2.09706426e-01
-5.01255766e-02 -1.45285988e+00 3.26229841e-01 4.61741418e-01
1.62195623e-01 5.44499040e-01 8.51583660e-01 -9.06246424e-01
-3.31431359e-01 4.77131754e-02 -5.27836159e-02 8.60035837e-01
2.45936424e-01 -3.36351126e-01 -9.09192681e-01 -4.48471755e-01
3.08591247e-01 -2.86479622e-01 6.15199268e-01 8.13325703e-01
9.79924202e-01 7.90755674e-02 -3.57448071e-01 7.76942551e-01
1.38069785e+00 3.01901460e-01 5.84004700e-01 1.40429765e-01
5.92207551e-01 3.71333808e-01 5.57077169e-01 2.17947319e-01
1.77057236e-01 6.08878613e-01 1.52543500e-01 -4.61686522e-01
-6.77905306e-02 -1.09739289e-01 9.48127955e-02 7.87231147e-01
8.43397155e-03 3.36484239e-02 -1.12183750e+00 7.56634474e-01
-1.81738997e+00 -5.69693863e-01 9.05365199e-02 2.14600229e+00
1.02050936e+00 6.53928593e-02 -1.05122244e-02 -2.08826751e-01
8.40421796e-01 1.47359921e-02 -7.15021253e-01 -2.37618938e-01
1.93601966e-01 3.27376455e-01 3.93372089e-01 4.67086583e-01
-1.33836162e+00 7.91108549e-01 5.60401964e+00 6.73462212e-01
-1.15021050e+00 5.72608709e-01 6.27579451e-01 3.33115011e-01
-3.70116085e-02 -1.25667691e-01 -3.19803774e-01 4.52923596e-01
9.09402788e-01 -2.53098458e-02 7.47295991e-02 8.39422703e-01
9.40007195e-02 9.37725231e-02 -1.07310820e+00 7.26837158e-01
-8.81987661e-02 -1.15269935e+00 -1.76761478e-01 -3.98487002e-02
9.02876496e-01 -5.06288856e-02 3.65508571e-02 3.80847305e-01
1.09229147e-01 -9.12239850e-01 2.93289691e-01 5.33932030e-01
1.16087091e+00 -5.90609610e-01 8.69073033e-01 5.10501564e-01
-9.85139310e-01 3.73605639e-01 -2.22848058e-01 4.05677408e-01
2.43498266e-01 3.88075709e-01 -7.67730534e-01 6.08699203e-01
5.32835126e-01 7.49874413e-01 -2.12744668e-01 6.26007378e-01
-3.36908072e-01 8.09743166e-01 -1.32001162e-01 6.06906712e-01
1.58043250e-01 -4.72653568e-01 7.32393622e-01 1.11476207e+00
5.62695973e-02 7.15158284e-02 4.54380125e-01 1.16139364e+00
3.21715735e-02 2.48283610e-01 -5.22518039e-01 -1.67071435e-03
1.84888303e-01 1.18268573e+00 -7.43485570e-01 -5.01941383e-01
-4.33316410e-01 1.02164710e+00 4.01697494e-02 5.35975218e-01
-1.08899701e+00 -1.59523278e-01 3.46978933e-01 5.46420738e-02
2.89617628e-01 -1.21926274e-02 -4.07984704e-01 -1.29311574e+00
-2.25948855e-01 -6.44386530e-01 5.51917732e-01 -5.51580548e-01
-1.47814786e+00 6.98216081e-01 2.84182996e-01 -1.32432163e+00
-3.84320706e-01 -3.77089709e-01 -3.83230031e-01 1.19152868e+00
-1.44830573e+00 -1.22996247e+00 -2.07741201e-01 9.45039809e-01
3.98020715e-01 -2.89750904e-01 9.19831336e-01 4.76088494e-01
-3.47269803e-01 5.58642507e-01 3.52241457e-01 4.53324467e-01
9.27454114e-01 -1.29033887e+00 -1.93490729e-01 7.73769140e-01
-3.09511665e-02 5.29449344e-01 5.18665016e-01 -6.56998098e-01
-5.95715880e-01 -1.13585782e+00 3.81396413e-01 -2.31346935e-01
4.17457998e-01 -2.28163674e-01 -1.11427951e+00 8.44778717e-01
2.00790927e-01 2.31380537e-01 7.50489354e-01 -3.65275711e-01
1.41096506e-02 1.33258373e-01 -1.43494153e+00 2.04110146e-01
5.25944054e-01 -4.47500736e-01 -7.88894117e-01 3.01200688e-01
6.02309108e-01 -8.11658800e-01 -1.32286203e+00 3.68917316e-01
2.63103098e-01 -7.69830108e-01 1.01399517e+00 -4.95114297e-01
5.20818949e-01 -2.86458164e-01 -2.70047151e-02 -1.36606050e+00
-1.48873553e-01 -1.68349028e-01 -1.13783628e-02 1.21401596e+00
2.40443587e-01 -7.16512024e-01 7.88754404e-01 4.51886237e-01
-2.52528578e-01 -6.10354483e-01 -1.02256680e+00 -5.43983698e-01
5.37692428e-01 -2.74017215e-01 2.46886760e-01 1.01893485e+00
-3.97011399e-01 2.67455727e-01 -2.14377955e-01 3.38302135e-01
7.79818892e-01 1.48093812e-02 5.17696559e-01 -1.01231420e+00
-3.03236455e-01 -9.16550532e-02 -3.49016070e-01 -8.23039412e-01
4.15742606e-01 -1.21691418e+00 7.53370151e-02 -1.37481034e+00
1.98623508e-01 -3.96428913e-01 -5.00559270e-01 3.21685553e-01
-8.93981829e-02 2.38593400e-01 1.71222221e-02 4.34243143e-01
-1.28928453e-01 5.10782838e-01 1.57634246e+00 -2.80208737e-01
-3.65831077e-01 2.04002216e-01 -3.88253152e-01 7.40942776e-01
9.36637819e-01 -7.13263392e-01 -5.93335032e-01 -1.67409956e-01
-5.73842764e-01 3.48611057e-01 4.31726307e-01 -9.66701329e-01
1.53770417e-01 8.73637665e-03 3.26298594e-01 -3.70863885e-01
1.79985806e-01 -8.94557595e-01 1.45786107e-01 4.29420114e-01
-3.29327166e-01 -6.36314154e-01 1.97128400e-01 5.97634435e-01
-4.93870646e-01 -3.40223730e-01 1.33677888e+00 -9.81929600e-02
-5.75079501e-01 3.84408236e-01 -2.42016092e-01 4.03813899e-01
1.23040879e+00 -2.26509139e-01 2.26525843e-01 -1.48027733e-01
-1.38042259e+00 1.40503019e-01 3.91629875e-01 4.25380021e-02
5.88397324e-01 -1.29503596e+00 -7.93747246e-01 4.02525604e-01
6.24896735e-02 2.64975458e-01 4.97148305e-01 1.06457245e+00
-4.17773008e-01 2.38960564e-01 -4.71478432e-01 -1.17781830e+00
-9.49810326e-01 5.66125929e-01 6.08116090e-01 -6.02181554e-01
-5.77175975e-01 6.04715288e-01 4.69418198e-01 -8.34141970e-01
-4.03901562e-02 -2.01649964e-01 -2.37277254e-01 -1.37152448e-01
2.53367200e-02 6.72792718e-02 -1.82153299e-01 -8.52361023e-01
-3.62909943e-01 7.19212353e-01 8.38417634e-02 -1.48174018e-01
1.14327610e+00 -1.29892662e-01 -5.32955080e-02 4.25345927e-01
1.24280953e+00 -3.04276764e-01 -1.44962847e+00 -4.44299042e-01
-1.80168346e-01 -2.13242531e-01 -4.77146916e-02 -7.34483540e-01
-1.20552349e+00 1.03653193e+00 9.62886214e-01 -1.00805946e-01
1.32706535e+00 1.22173972e-01 7.25486934e-01 -2.46304184e-01
1.98271945e-01 -1.02675629e+00 -4.33811173e-02 3.28175277e-01
6.34330571e-01 -1.41052997e+00 -1.09151617e-01 -4.23680574e-01
-9.81966615e-01 1.05166185e+00 5.04800260e-01 -3.03272247e-01
8.30983937e-01 9.51318536e-03 2.59739041e-01 -7.48164505e-02
-1.76012099e-01 3.56911123e-02 4.93346423e-01 9.02467191e-01
5.38001835e-01 1.61698937e-01 -4.47287679e-01 8.20211291e-01
2.14869484e-01 2.28802070e-01 1.77441001e-01 6.87191367e-01
4.79059033e-02 -1.38061190e+00 -4.09198910e-01 2.51250833e-01
-5.72399557e-01 1.12245448e-01 2.28200331e-01 9.59384501e-01
3.80873412e-01 5.41060984e-01 -8.36347044e-02 -4.44086492e-02
2.86443084e-01 2.88393021e-01 4.02485698e-01 -7.43475795e-01
-3.95090401e-01 3.44757527e-01 -3.89346212e-01 -4.02102530e-01
-7.22567499e-01 -8.01280379e-01 -1.39842570e+00 3.82692873e-01
-3.51165757e-02 2.30809897e-01 3.89189392e-01 8.56598020e-01
9.22945589e-02 7.55791306e-01 5.03574967e-01 -7.93197215e-01
-6.02862179e-01 -1.08521736e+00 -7.42200971e-01 6.99143529e-01
1.69062525e-01 -7.91342735e-01 -2.21985191e-01 6.31765425e-01] | [14.536225318908691, -2.008660078048706] |
5df4bbfe-2b34-4834-b62e-770517390871 | syntactic-structure-processing-in-the-brain | 2302.08589 | null | https://arxiv.org/abs/2302.08589v1 | https://arxiv.org/pdf/2302.08589v1.pdf | Syntactic Structure Processing in the Brain while Listening | Syntactic parsing is the task of assigning a syntactic structure to a sentence. There are two popular syntactic parsing methods: constituency and dependency parsing. Recent works have used syntactic embeddings based on constituency trees, incremental top-down parsing, and other word syntactic features for brain activity prediction given the text stimuli to study how the syntax structure is represented in the brain's language network. However, the effectiveness of dependency parse trees or the relative predictive power of the various syntax parsers across brain areas, especially for the listening task, is yet unexplored. In this study, we investigate the predictive power of the brain encoding models in three settings: (i) individual performance of the constituency and dependency syntactic parsing based embedding methods, (ii) efficacy of these syntactic parsing based embedding methods when controlling for basic syntactic signals, (iii) relative effectiveness of each of the syntactic embedding methods when controlling for the other. Further, we explore the relative importance of syntactic information (from these syntactic embedding methods) versus semantic information using BERT embeddings. We find that constituency parsers help explain activations in the temporal lobe and middle-frontal gyrus, while dependency parsers better encode syntactic structure in the angular gyrus and posterior cingulate cortex. Although semantic signals from BERT are more effective compared to any of the syntactic features or embedding methods, syntactic embedding methods explain additional variance for a few brain regions. | ['Bapi Raju Surampud', 'Manish Gupta', 'Mounika Marreddy', 'Subba Reddy Oota'] | 2023-02-16 | null | null | null | null | ['activity-prediction', 'dependency-parsing', 'activity-prediction'] | ['computer-vision', 'natural-language-processing', 'time-series'] | [-9.36116129e-02 2.69931048e-01 -1.11006707e-01 -5.96547604e-01
-2.89219797e-01 -5.80486298e-01 6.37838900e-01 4.75757003e-01
-5.79121828e-01 2.26063311e-01 1.06308675e+00 -2.88173199e-01
-2.93237656e-01 -8.05739462e-01 -4.42938089e-01 -6.70388579e-01
-4.62294698e-01 1.29742637e-01 8.12421516e-02 3.19137834e-02
2.09368289e-01 3.41664940e-01 -1.25853801e+00 3.00024927e-01
4.44079936e-01 8.43183279e-01 4.63143975e-01 4.58220452e-01
-4.36073929e-01 6.41563714e-01 -3.07648063e-01 -3.50559384e-01
-1.41623631e-01 -5.74238956e-01 -6.99855804e-01 -6.61227345e-01
-8.29208195e-02 -2.33662844e-01 -9.64604132e-03 1.19064450e+00
6.70830786e-01 -1.61564946e-01 4.99144614e-01 -5.88867545e-01
-8.07281375e-01 8.67551923e-01 -1.12105787e-01 9.37064946e-01
3.17058772e-01 -3.24407220e-02 1.61289573e+00 -8.40821207e-01
6.06580913e-01 1.45727885e+00 5.84131837e-01 3.81585240e-01
-1.40728939e+00 -6.74982965e-01 4.09904540e-01 7.88411275e-02
-6.02222919e-01 -4.15009201e-01 7.41369367e-01 -6.64197087e-01
1.23157644e+00 -1.11270458e-01 7.00956702e-01 1.33090794e+00
9.53141928e-01 2.62415737e-01 1.55218351e+00 -2.23517448e-01
3.61559808e-01 5.57067059e-02 1.13060832e+00 4.94031906e-01
3.79154384e-01 3.17767501e-01 -5.77502072e-01 -4.32376653e-01
3.39324385e-01 -1.35861754e-01 -1.59730926e-01 3.48943830e-01
-1.07952523e+00 1.27940381e+00 5.49544036e-01 6.19502008e-01
-5.64433753e-01 3.51381242e-01 6.82032585e-01 -6.87536523e-02
5.22813082e-01 6.92155242e-01 -8.45508516e-01 1.01210922e-01
-6.93126798e-01 8.62855464e-03 5.70240557e-01 1.90148130e-01
5.09733558e-01 1.58550978e-01 -2.86835074e-01 9.45659637e-01
7.59585500e-01 2.05210820e-01 7.88353622e-01 -6.22224510e-01
5.74969769e-01 2.13552490e-01 -5.36556184e-01 -1.14456606e+00
-1.18419337e+00 6.13067709e-02 -1.66207448e-01 -7.51301199e-02
3.37112874e-01 -3.31681848e-01 -5.97320497e-01 2.29439640e+00
2.08156779e-02 -4.88784999e-01 -2.30745479e-01 9.16737258e-01
9.01760340e-01 5.33196092e-01 8.51345956e-01 -2.69601464e-01
2.06093073e+00 -4.33788449e-01 -8.13170254e-01 -8.73871744e-01
8.03664625e-01 -2.88909972e-01 5.91346085e-01 -4.69870031e-01
-9.90828216e-01 -5.84273517e-01 -9.28452849e-01 -2.74577349e-01
-3.34169418e-01 -3.35920267e-02 9.38774824e-01 5.70605218e-01
-1.11918592e+00 4.27203268e-01 -9.15291071e-01 -4.01599526e-01
5.08385718e-01 3.64547163e-01 -6.03704751e-01 -4.76953387e-02
-1.32141304e+00 1.16971076e+00 3.24905634e-01 -6.97662309e-02
-4.12835240e-01 -7.31463671e-01 -1.12892151e+00 5.70644617e-01
-3.52934003e-01 -8.12657356e-01 6.23120725e-01 -1.29272759e+00
-1.13178468e+00 8.11825097e-01 -3.60645175e-01 -3.54473144e-02
-7.22458541e-01 1.64942130e-01 1.26156928e-02 3.60969245e-01
4.53395814e-01 8.17102671e-01 5.14716208e-01 -5.01266539e-01
-8.96929875e-02 -1.05633116e+00 4.00873199e-02 1.42226011e-01
-2.18877662e-02 3.82011980e-01 4.06540483e-01 -6.89420462e-01
5.40114641e-01 -8.00151527e-01 -8.82272944e-02 -1.05600357e-01
-7.65587464e-02 -7.30668724e-01 1.94852918e-01 -1.05305159e+00
8.63379717e-01 -2.38005280e+00 2.13661402e-01 -1.84732720e-01
2.98173726e-01 -4.52852845e-01 -2.47555032e-01 1.05221130e-01
-4.98407513e-01 8.44069719e-01 -8.96179751e-02 3.24728675e-02
2.04097882e-01 1.34442538e-01 -6.59962445e-02 5.80193520e-01
5.48201740e-01 1.03106570e+00 -6.36118412e-01 -1.48160338e-01
-2.44438335e-01 3.91676664e-01 -9.33592021e-01 -1.28362164e-01
1.64182648e-01 1.69513449e-01 -5.60572267e-01 3.75617713e-01
4.90255266e-01 1.53771117e-01 4.75542605e-01 -4.25313473e-01
-9.05375257e-02 1.05558503e+00 -5.11369646e-01 1.35791838e+00
-3.53073478e-01 9.15421426e-01 2.77878135e-01 -1.12428498e+00
6.73673987e-01 4.23638493e-01 7.67473578e-02 -1.05510926e+00
4.83773500e-01 5.87146096e-02 1.09204483e+00 -4.37948167e-01
-2.32821569e-01 -6.86127245e-01 -3.18881124e-01 6.86590791e-01
6.76251769e-01 4.41652298e-01 -4.25029211e-02 2.64078900e-02
1.40619528e+00 -9.18777734e-02 4.44405138e-01 -6.80406213e-01
-3.84406298e-02 -3.92856300e-01 6.92749381e-01 5.46991229e-01
-6.17203772e-01 2.64923602e-01 1.05282128e+00 -2.30374932e-01
-6.48216724e-01 -1.06915486e+00 -6.26782537e-01 1.69574046e+00
-2.82062203e-01 -4.59490210e-01 -6.27025425e-01 -4.66563553e-01
-1.52604014e-01 9.21652377e-01 -8.47125828e-01 -4.50701743e-01
-4.90632385e-01 -1.17489958e+00 2.74391383e-01 9.93591070e-01
2.20889717e-01 -1.28994668e+00 -9.25332665e-01 2.86126673e-01
1.04642823e-01 -1.24903262e+00 -2.47780904e-01 8.21462691e-01
-1.20313644e+00 -9.52376544e-01 3.39156538e-02 -7.65202880e-01
5.57631493e-01 -5.08587956e-02 9.03144419e-01 2.05697455e-02
-2.82242388e-01 6.03583813e-01 -4.38803941e-01 -2.78082460e-01
-1.60241127e-01 -7.64890015e-02 -8.89150351e-02 -4.65908110e-01
4.09930557e-01 -6.80017650e-01 -5.08960605e-01 -1.97688296e-01
-5.46375036e-01 -2.59333074e-01 6.65166497e-01 8.57547760e-01
3.53709430e-01 -1.93745017e-01 5.30980170e-01 -9.08493042e-01
8.77810597e-01 -8.56235087e-01 -6.68569654e-02 -3.52156878e-01
-3.73393357e-01 4.08925027e-01 4.82672155e-01 -1.35692656e-01
-1.20066464e+00 -2.06314832e-01 -4.70372558e-01 3.92666906e-01
-1.29176870e-01 7.43434846e-01 -8.06223452e-02 2.72069097e-01
4.66788143e-01 -1.08295165e-01 -1.20920837e-01 -1.50812075e-01
7.81245381e-02 1.20216370e-01 -9.17294919e-02 -7.82305658e-01
2.47019082e-01 1.74226224e-01 -1.32053152e-01 -9.28560257e-01
-8.47670853e-01 1.34250060e-01 -8.27002883e-01 1.79633781e-01
1.87791586e+00 -7.15918899e-01 -4.67927247e-01 -3.00698578e-01
-1.58179402e+00 -4.62878682e-02 1.25858244e-02 8.00599933e-01
-2.72964746e-01 1.54965341e-01 -8.43050241e-01 -4.62196410e-01
-2.93129236e-01 -1.50066018e+00 1.03488314e+00 -1.06774837e-01
-5.75209081e-01 -1.23771381e+00 1.31405771e-01 9.12430976e-03
3.68737191e-01 1.81214184e-01 1.60430825e+00 -1.03026700e+00
2.41629146e-02 -6.80851862e-02 -3.72273415e-01 -8.22760388e-02
-1.63502693e-01 -2.64379263e-01 -1.21046996e+00 1.72691405e-01
4.25757796e-01 1.12153627e-02 1.07977831e+00 8.45206738e-01
7.22228348e-01 -2.34863192e-01 -3.09182703e-01 4.14333642e-01
1.15753663e+00 2.04272404e-01 3.22967887e-01 1.31157011e-01
2.72741824e-01 1.10665226e+00 -1.35742888e-01 -5.89770935e-02
3.56111228e-01 3.75050187e-01 2.62594461e-01 4.22237307e-01
-4.76048440e-02 -3.27619165e-02 8.73962343e-01 7.27251828e-01
2.45055452e-01 2.72182256e-01 -7.97686994e-01 4.18601155e-01
-1.12744915e+00 -1.04061413e+00 -2.05194026e-01 1.78786206e+00
5.74894130e-01 2.70477623e-01 -3.64015013e-01 -7.47040808e-02
4.99639571e-01 4.75106686e-01 -3.10988277e-01 -1.01663399e+00
8.62941891e-03 5.82841873e-01 2.90470213e-01 2.00419217e-01
-9.45484459e-01 7.61513174e-01 6.13795042e+00 1.01498887e-01
-8.02015066e-01 6.45548880e-01 5.87804556e-01 1.24312840e-01
-2.98971325e-01 1.73592731e-01 -7.89442837e-01 3.20545733e-01
1.48559427e+00 1.28075078e-01 3.39583397e-01 8.30570102e-01
1.52122363e-01 -4.07117188e-01 -1.41890574e+00 5.09125113e-01
4.73722592e-02 -1.18288672e+00 -4.49127138e-01 1.34029850e-01
-9.05881301e-02 2.36555114e-01 -2.82052487e-01 6.16577566e-01
-4.82013300e-02 -9.03754115e-01 8.31778646e-01 1.14614703e-01
3.46466422e-01 -1.90585088e-02 8.53675783e-01 2.73095131e-01
-1.19740033e+00 -3.26670855e-01 -6.01493418e-01 -3.06176692e-01
2.72809982e-01 5.00342369e-01 -2.07616702e-01 -2.68787175e-01
8.82603884e-01 3.84361714e-01 -6.64759398e-01 1.86394483e-01
-6.16071403e-01 9.69277263e-01 -1.32180005e-01 -1.69052511e-01
1.98393330e-01 2.51830101e-01 5.14449537e-01 1.24348998e+00
-5.60820438e-02 5.10817587e-01 -2.27083251e-01 1.19030166e+00
1.13423645e-01 2.59389251e-01 -6.76989317e-01 -2.11344182e-01
5.92111886e-01 1.37020063e+00 -1.02328956e+00 2.17920206e-02
-5.71337640e-01 6.65357769e-01 6.46012247e-01 1.88246906e-01
-6.22237980e-01 -1.38267791e-02 7.36910820e-01 -3.90872620e-02
1.76820800e-01 -3.76164436e-01 -4.92221028e-01 -8.00339282e-01
5.11274934e-02 -2.08870441e-01 3.76312017e-01 -9.44544017e-01
-1.15434647e+00 4.94091064e-01 2.61759833e-02 -8.69445205e-02
5.36289252e-02 -8.93016815e-01 -8.23280096e-01 1.14606249e+00
-1.42230332e+00 -6.17171109e-01 1.73249483e-01 7.50776082e-02
7.50694335e-01 2.00595155e-01 1.27731729e+00 -1.03545584e-01
-7.14891374e-01 1.88071370e-01 -2.96354622e-01 3.07496846e-01
4.65722114e-01 -1.15462852e+00 2.98878215e-02 5.05898356e-01
1.06071271e-01 9.97300625e-01 2.20346346e-01 -4.86977637e-01
-1.17049873e+00 -6.95889771e-01 1.03214347e+00 -4.86956805e-01
5.89919806e-01 -6.82316661e-01 -9.66767251e-01 8.38206828e-01
1.91435963e-01 1.74038950e-03 1.05950236e+00 3.75442654e-01
-5.78738093e-01 2.84641474e-01 -1.04085386e+00 3.51321876e-01
8.69585454e-01 -3.89479131e-01 -1.04415798e+00 2.29884878e-01
9.71945107e-01 3.12309474e-01 -1.04437006e+00 5.18040918e-02
5.64348221e-01 -9.37342286e-01 9.49971318e-01 -8.11553895e-01
8.22221637e-01 3.12010884e-01 -3.93106520e-01 -1.40528131e+00
-1.17673731e+00 5.26380062e-01 3.85826319e-01 1.30906403e+00
6.10088468e-01 -8.52012575e-01 1.50500804e-01 8.62645149e-01
-1.69151440e-01 -5.16229331e-01 -1.20594811e+00 -2.57694453e-01
3.33892763e-01 -6.03110313e-01 9.81090739e-02 9.31622565e-01
1.90468445e-01 9.83541310e-01 5.17572105e-01 2.47000486e-01
6.55831173e-02 -2.01543912e-01 -2.32663468e-01 -1.28943563e+00
-2.78570622e-01 -5.81588268e-01 -4.20042217e-01 -2.73246139e-01
8.24761271e-01 -1.57241654e+00 -1.29902825e-01 -1.84250855e+00
4.55554754e-01 1.55834053e-02 -1.94074556e-01 7.94136882e-01
-4.90223020e-01 -3.89288843e-01 3.80859584e-01 8.15935954e-02
1.03677124e-01 4.32847977e-01 6.89838707e-01 3.02166734e-02
-1.40700802e-01 -4.97670203e-01 -1.09063792e+00 9.52705979e-01
8.51292729e-01 -8.86916697e-01 -2.39818260e-01 -6.53440952e-01
2.67107427e-01 2.74725467e-01 2.06588835e-01 -6.69915199e-01
-1.02456570e-01 3.33954930e-01 8.04112673e-01 2.67760843e-01
1.89649150e-01 -7.75885642e-01 -3.83665055e-01 5.54559827e-01
-3.82907838e-01 4.30292666e-01 5.17269433e-01 4.96465206e-01
8.26922506e-02 -4.30990279e-01 9.11528647e-01 -2.36794829e-01
-4.76810694e-01 -1.96325988e-01 -9.17424381e-01 3.31828266e-01
6.62870288e-01 -2.28876099e-01 -3.19091916e-01 6.68416694e-02
-1.15250826e+00 -2.79474407e-01 -4.79097128e-01 4.46318507e-01
3.31548303e-01 -1.22502553e+00 -5.29204249e-01 8.43246356e-02
-3.51444215e-01 -7.57708549e-01 1.38635114e-01 1.20633531e+00
-1.90142661e-01 4.24449623e-01 -2.63682693e-01 -2.77622461e-01
-8.86947572e-01 4.73687291e-01 3.44585180e-01 -2.03746960e-01
-4.54224914e-01 8.71304274e-01 7.62865305e-01 -3.68683577e-01
-3.95431846e-01 -4.93071854e-01 -5.23639202e-01 6.38084114e-01
4.96304989e-01 8.85630026e-02 -1.81602776e-01 -7.79692829e-01
-7.37876177e-01 4.61826771e-01 1.01721622e-01 -3.99330184e-02
1.60148537e+00 2.55098552e-01 -6.05168223e-01 5.91946483e-01
1.37867057e+00 -4.27673310e-02 -6.21871531e-01 1.91866025e-01
2.76356936e-01 7.92228356e-02 3.89864117e-01 -6.14789307e-01
-9.87935185e-01 1.16242158e+00 7.58891225e-01 3.70539464e-02
8.56098711e-01 4.15614218e-01 2.47720838e-01 -1.06159918e-01
2.09695935e-01 -9.84000444e-01 -2.00132474e-01 4.73926187e-01
8.60874951e-01 -1.12884772e+00 4.35427651e-02 -2.06873596e-01
-4.28425640e-01 1.41936588e+00 5.11304617e-01 -3.93589824e-01
1.21126175e+00 2.57653683e-01 -5.34008563e-01 -6.01481199e-01
-8.62449110e-01 -2.59911835e-01 9.20330212e-02 3.54771852e-01
1.03456366e+00 7.96026811e-02 -1.05563307e+00 1.56246912e+00
-3.52992684e-01 -8.59321535e-01 3.55183423e-01 7.48177528e-01
-6.69414639e-01 -7.68920898e-01 -3.71773839e-01 7.94560134e-01
-7.01639533e-01 -4.75906372e-01 -4.17958170e-01 8.11878204e-01
8.05777386e-02 7.59878814e-01 3.82462889e-01 -1.22656398e-01
3.57978493e-01 6.70744240e-01 2.43267730e-01 -1.27752137e+00
-8.19032669e-01 5.54863885e-02 2.97840208e-01 -8.01012278e-01
-3.27148974e-01 -1.08140242e+00 -1.57807529e+00 2.58591473e-01
-2.49884188e-01 -1.07610434e-01 7.96994746e-01 1.13715386e+00
3.65632385e-01 5.36924779e-01 3.46596152e-01 -1.02136588e+00
-2.54656136e-01 -1.06441140e+00 -5.22347450e-01 1.44218415e-01
-1.86386004e-01 -9.77150023e-01 -5.92277110e-01 -2.43088350e-01] | [10.35354995727539, 8.50393009185791] |
15fb337c-ddf6-4d8c-a6e4-252daf36893d | yolox-exceeding-yolo-series-in-2021 | 2107.08430 | null | https://arxiv.org/abs/2107.08430v2 | https://arxiv.org/pdf/2107.08430v2.pdf | YOLOX: Exceeding YOLO Series in 2021 | In this report, we present some experienced improvements to YOLO series, forming a new high-performance detector -- YOLOX. We switch the YOLO detector to an anchor-free manner and conduct other advanced detection techniques, i.e., a decoupled head and the leading label assignment strategy SimOTA to achieve state-of-the-art results across a large scale range of models: For YOLO-Nano with only 0.91M parameters and 1.08G FLOPs, we get 25.3% AP on COCO, surpassing NanoDet by 1.8% AP; for YOLOv3, one of the most widely used detectors in industry, we boost it to 47.3% AP on COCO, outperforming the current best practice by 3.0% AP; for YOLOX-L with roughly the same amount of parameters as YOLOv4-CSP, YOLOv5-L, we achieve 50.0% AP on COCO at a speed of 68.9 FPS on Tesla V100, exceeding YOLOv5-L by 1.8% AP. Further, we won the 1st Place on Streaming Perception Challenge (Workshop on Autonomous Driving at CVPR 2021) using a single YOLOX-L model. We hope this report can provide useful experience for developers and researchers in practical scenes, and we also provide deploy versions with ONNX, TensorRT, NCNN, and Openvino supported. Source code is at https://github.com/Megvii-BaseDetection/YOLOX. | ['Jian Sun', 'Zeming Li', 'Feng Wang', 'Songtao Liu', 'Zheng Ge'] | 2021-07-18 | null | null | null | null | ['real-time-object-detection'] | ['computer-vision'] | [-4.47813034e-01 -2.44215727e-01 -2.62371421e-01 -5.87958694e-02
-8.06369007e-01 -5.29133320e-01 -1.59650408e-02 -7.40127414e-02
-6.16299093e-01 2.41565511e-01 -3.73703629e-01 -3.67224663e-01
6.06945634e-01 -3.25302690e-01 -8.42114627e-01 -4.93583381e-01
-3.70124578e-01 2.91676279e-02 1.01267314e+00 -2.92139202e-01
2.92149205e-02 3.29261124e-01 -1.82438445e+00 2.41285428e-01
2.43187323e-01 1.35658944e+00 -1.77062186e-03 1.45493519e+00
5.30227780e-01 9.55439508e-01 -6.56311512e-01 3.61825973e-02
3.73718113e-01 4.16105650e-02 -2.34232694e-01 -4.19773519e-01
9.14751112e-01 -3.77205759e-01 -5.25668383e-01 7.29591906e-01
8.72279286e-01 -1.50415525e-01 3.89032848e-02 -1.70880127e+00
-2.11581975e-01 4.76708591e-01 -9.94754016e-01 8.72187614e-01
1.34575605e-01 3.71075451e-01 9.78287518e-01 -1.03017330e+00
3.19086462e-01 1.04649794e+00 8.92240047e-01 3.16788524e-01
-7.57858276e-01 -9.61158633e-01 -5.52303828e-02 5.77359378e-01
-1.47339404e+00 -7.56112635e-01 9.48962346e-02 -2.35497192e-01
1.14145935e+00 5.79147413e-02 3.32909286e-01 8.66673529e-01
3.98687333e-01 7.72582889e-01 8.44306350e-01 -1.84292525e-01
1.56270385e-01 -2.54671741e-02 3.12692732e-01 7.41615653e-01
1.39439702e-01 1.56139523e-01 -9.74170923e-01 9.63641622e-04
3.32943022e-01 -5.96304595e-01 -4.36273217e-03 2.19149180e-02
-1.13096821e+00 7.02396333e-01 3.74514222e-01 -1.98011056e-01
1.08404443e-01 7.05935955e-01 7.68407524e-01 2.16932505e-01
8.22820589e-02 -6.38813712e-04 -3.20368528e-01 -6.67138398e-01
-9.41066563e-01 2.71243751e-01 6.47037327e-01 1.39902103e+00
5.26168168e-01 3.13309461e-01 1.95178017e-01 6.60786569e-01
3.49317461e-01 8.72364700e-01 2.52244264e-01 -1.56967008e+00
4.67875034e-01 -2.91507870e-01 1.88331231e-01 -5.73758781e-01
-6.16900444e-01 -6.60349607e-01 -3.15007120e-01 4.25942659e-01
3.58782828e-01 -3.95338446e-01 -8.66405308e-01 1.35228336e+00
1.41787633e-01 3.12285841e-01 9.96510014e-02 1.16716123e+00
9.11839843e-01 9.11745071e-01 -1.05930023e-01 1.77434772e-01
1.62498045e+00 -1.48180306e+00 -2.69480318e-01 -5.64966261e-01
9.49179530e-01 -7.94997811e-01 1.02115190e+00 9.80088711e-01
-1.06758010e+00 -8.62442970e-01 -1.49111545e+00 -1.18937902e-01
6.68021962e-02 2.42495015e-01 3.49184155e-01 7.67794609e-01
-1.26706219e+00 6.03978038e-02 -9.60590422e-01 -5.96726120e-01
2.41680548e-01 3.60899121e-02 3.10726706e-02 -1.06833152e-01
-1.00059211e+00 5.54732919e-01 8.64982158e-02 -7.39466250e-02
-1.15718579e+00 -6.95003629e-01 -4.46928710e-01 -1.78285055e-02
5.33560812e-01 -2.55821079e-01 1.65417778e+00 -1.89404681e-01
-1.34942210e+00 7.68062055e-01 -9.23618078e-02 -1.15373373e+00
4.23624367e-01 -7.01185942e-01 -8.22488129e-01 5.25175512e-01
3.33078176e-01 1.13228118e+00 5.53871095e-01 -1.05296111e+00
-1.05531824e+00 8.41604620e-02 6.01381920e-02 8.97420272e-02
5.94309391e-03 1.02569364e-01 -6.93715155e-01 -3.42831039e-03
-1.05504632e-01 -1.02473021e+00 -8.08362439e-02 8.88679996e-02
-1.33963719e-01 -1.12006821e-01 9.90284801e-01 -1.72549203e-01
1.10578656e+00 -2.54467630e+00 -7.13480711e-01 -1.79972813e-01
5.67144573e-01 5.90834439e-01 -5.85677624e-02 3.75072062e-01
3.52865726e-01 -8.60405564e-02 2.94041783e-01 -4.69195724e-01
4.76008840e-02 1.90524697e-01 -2.05887407e-01 6.81549251e-01
-1.44367531e-01 5.74139774e-01 -8.63847077e-01 -4.11000878e-01
2.60348856e-01 2.26575196e-01 -7.38803029e-01 1.43856974e-02
7.93873612e-03 -1.97259244e-03 -1.30990118e-01 8.28947663e-01
8.15123022e-01 -2.26885900e-01 -5.03245831e-01 -9.46742892e-02
-5.03065884e-01 2.02458113e-01 -1.08643603e+00 1.55721354e+00
-4.06405181e-01 1.32759714e+00 2.04770744e-01 -5.94325542e-01
9.53849673e-01 1.63302675e-01 3.51939142e-01 -7.62948275e-01
2.54432976e-01 3.07668746e-01 -2.98818597e-03 -4.19476569e-01
8.05575848e-01 3.72012585e-01 -7.55283684e-02 -1.55104995e-01
-5.07147014e-02 9.73398983e-03 5.26948750e-01 6.40539169e-01
1.48109269e+00 3.92824374e-02 -6.10096380e-02 -1.28194481e-01
1.51668027e-01 6.60780221e-02 6.08234882e-01 1.22886193e+00
-8.65282714e-01 7.40312040e-01 5.46916485e-01 -1.01485930e-01
-9.62289393e-01 -1.24976528e+00 -3.80200177e-01 1.16668081e+00
3.59550625e-01 -8.86793435e-01 -6.70796573e-01 -1.53004825e-01
2.24975497e-02 6.41260386e-01 -1.61008477e-01 6.02308773e-02
-4.73553926e-01 -6.51438653e-01 1.26707458e+00 7.22418725e-01
8.00699353e-01 -8.07703316e-01 -1.17850399e+00 2.10107937e-01
1.93494856e-02 -1.78531182e+00 -4.58637357e-01 3.25339556e-01
-3.77695590e-01 -7.07409799e-01 -3.76577020e-01 -4.58997160e-01
-3.09615225e-01 6.65307522e-01 1.11504352e+00 -7.82582387e-02
-3.83916348e-01 2.04252616e-01 -6.82221353e-01 -4.23667908e-01
-1.30950868e-01 8.81747752e-02 1.92760944e-01 -2.32989714e-01
3.86558831e-01 -3.71554822e-01 -7.84169078e-01 5.82439661e-01
-4.72996026e-01 -3.15048277e-01 3.91950428e-01 4.00474489e-01
4.06296998e-01 -4.41340297e-01 5.46456218e-01 -5.70305765e-01
-1.13014698e-01 -6.15435421e-01 -8.25310946e-01 -4.98053312e-01
-6.77076936e-01 -4.54231113e-01 6.78855836e-01 -2.33519465e-01
-5.83022714e-01 -9.73361731e-02 -4.13875610e-01 -7.85796762e-01
-1.55288145e-01 2.10543070e-03 3.30510527e-01 -2.86859851e-02
9.55957174e-01 -1.58681914e-01 -3.13382626e-01 -2.31129825e-01
4.96269375e-01 8.09171319e-01 9.44844484e-01 -3.41075391e-01
4.84448731e-01 7.81001031e-01 -6.54217079e-02 -1.03369594e+00
-8.58243108e-01 -8.98016870e-01 -1.12486929e-01 -1.52596846e-01
9.45502996e-01 -1.41623914e+00 -1.05166006e+00 6.10071838e-01
-8.88761699e-01 -6.13460124e-01 -2.18945310e-01 6.13172531e-01
-5.30990124e-01 3.36820841e-01 -8.10827613e-01 -5.63814819e-01
-1.39607802e-01 -1.11034548e+00 1.16669416e+00 3.57210517e-01
1.14861622e-01 -4.04766142e-01 4.30946536e-02 4.66776878e-01
3.43569160e-01 -1.22077234e-01 -2.60545373e-01 -1.66405320e-01
-5.29385388e-01 -1.55373022e-01 -6.97589338e-01 5.36508024e-01
-7.31548965e-01 1.96086332e-01 -1.27545798e+00 -6.19994223e-01
-3.22217256e-01 -5.99639416e-01 1.06286120e+00 4.43990380e-01
8.94373715e-01 4.21948403e-01 -3.82046759e-01 9.67312992e-01
1.28115022e+00 2.22402841e-01 5.03152788e-01 5.77114940e-01
7.49530911e-01 -6.30471259e-02 1.01073503e+00 5.84271491e-01
6.36432528e-01 7.49660730e-01 7.00548708e-01 4.88101617e-02
-4.10612673e-01 -6.06753081e-02 7.97930419e-01 8.80642712e-01
8.81792605e-02 -8.50789070e-01 -9.65786457e-01 4.09422785e-01
-1.71460915e+00 -8.02899003e-01 -7.14780092e-01 1.89344490e+00
1.60294861e-01 8.30383480e-01 3.92813891e-01 -4.14737016e-02
4.28092062e-01 4.18637067e-01 -5.90127587e-01 -7.70635068e-01
-1.69753656e-01 -2.23999983e-03 1.26839840e+00 5.21570802e-01
-1.03925037e+00 1.10111046e+00 6.19684267e+00 9.31690395e-01
-1.13217056e+00 4.79311466e-01 4.02297884e-01 -4.15749401e-01
4.65592951e-01 1.28692016e-01 -1.45142221e+00 5.04668355e-01
1.57186687e+00 -7.95908421e-02 3.41221504e-02 1.01482999e+00
4.67401028e-01 -7.05670595e-01 -1.01034677e+00 1.22031999e+00
2.21542001e-01 -1.16682208e+00 -7.29173660e-01 1.88846439e-01
3.17103446e-01 9.86039162e-01 -1.27295002e-01 4.81732488e-01
2.90255994e-01 -8.23699355e-01 1.03614771e+00 9.70420241e-02
9.84084964e-01 -5.85882306e-01 7.99051583e-01 2.00748891e-01
-1.40300786e+00 -5.24369217e-02 -6.86705530e-01 -2.94234425e-01
4.40026373e-01 4.57554728e-01 -6.50045335e-01 2.91639984e-01
1.32751918e+00 8.41360331e-01 -6.62214100e-01 1.29136670e+00
-9.50384662e-02 1.14970744e+00 -8.03150415e-01 7.10257515e-02
4.61670399e-01 5.06844580e-01 6.78404927e-01 1.44311905e+00
3.06326210e-01 -1.09773174e-01 3.63389611e-01 1.65297970e-01
-6.14150874e-02 -4.08338070e-01 -3.09138238e-01 6.54180944e-01
8.11421633e-01 1.37765539e+00 -6.22529626e-01 -3.75352472e-01
-5.79332650e-01 7.63146460e-01 -1.62663013e-02 1.52643919e-01
-1.48865354e+00 -4.57521915e-01 6.21223330e-01 8.85912478e-02
6.74276769e-01 -5.92448115e-01 -5.20165004e-02 -9.37753439e-01
8.63270718e-04 -6.11828089e-01 2.93578029e-01 -1.12196255e+00
-7.85985887e-01 7.18254685e-01 3.84599417e-02 -1.37704110e+00
2.55853590e-02 -6.28484488e-01 -4.30103749e-01 4.80716884e-01
-1.67675519e+00 -5.30385673e-01 -5.77946365e-01 1.41110197e-01
6.61694348e-01 -7.60323927e-02 2.58833259e-01 7.38660932e-01
-5.73314726e-01 8.57168078e-01 8.40659440e-02 -1.39497310e-01
1.01415062e+00 -1.09879208e+00 7.37826347e-01 1.01424730e+00
1.00925326e-01 -2.19374940e-01 9.15472567e-01 -9.46415439e-02
-1.34568393e+00 -1.02317798e+00 4.33681518e-01 -6.08976007e-01
9.90487874e-01 -4.64869350e-01 -5.95624387e-01 7.69891202e-01
3.15759540e-01 6.82230711e-01 2.23208427e-01 -3.37882429e-01
-3.47277910e-01 -3.97593826e-01 -7.00190425e-01 3.68636876e-01
1.10055864e+00 -6.40382320e-02 -1.17655201e-02 3.03459376e-01
8.94839823e-01 -7.43270516e-01 -8.13838959e-01 9.85210389e-02
6.46465719e-01 -1.43204963e+00 7.64983296e-01 6.37125969e-02
-5.95697984e-02 -6.16245747e-01 -4.70846564e-01 -8.49888563e-01
-2.10647956e-01 -7.50391304e-01 -1.25665486e-01 1.08011627e+00
2.51325667e-01 -7.93693662e-01 9.84992385e-01 -4.09775168e-01
-7.65015364e-01 -6.21405005e-01 -1.12666154e+00 -1.18732882e+00
-7.43064955e-02 -8.81449997e-01 -9.71123129e-02 1.95599213e-01
-2.55242914e-01 4.71121758e-01 -4.35241401e-01 5.01237392e-01
6.74680650e-01 -3.88569951e-01 1.10618222e+00 -5.72653234e-01
-5.69324791e-01 -8.09149519e-02 -7.64447451e-01 -1.70767188e+00
-4.53950733e-01 -5.04589558e-01 1.66682407e-01 -1.01500666e+00
-2.78230995e-01 -5.79814911e-01 -1.69174016e-01 3.56364995e-01
2.72976398e-01 8.81962001e-01 6.27471805e-01 3.48140299e-01
-1.09577012e+00 2.73873329e-01 7.11027920e-01 2.23218203e-01
2.33360454e-02 -1.97428986e-01 -4.16173279e-01 8.04905117e-01
1.03220630e+00 -5.95978379e-01 -2.49056235e-01 -4.63010520e-01
7.70940110e-02 2.07488984e-02 4.60016489e-01 -1.85042238e+00
3.37542742e-01 3.68029624e-01 2.03865990e-02 -7.99594104e-01
6.79749846e-01 -3.89785439e-01 -3.12523216e-01 5.57286263e-01
1.21975705e-01 3.66090477e-01 4.86106455e-01 3.43295664e-01
-1.55812785e-01 -5.07374629e-02 8.71056795e-01 2.65872359e-01
-1.29027987e+00 1.96340173e-01 -7.12174296e-01 4.70628530e-01
1.09342074e+00 -3.41635555e-01 -1.07351387e+00 -4.47054267e-01
-6.43552840e-01 7.71847010e-01 3.25777650e-01 4.10431892e-01
4.64912981e-01 -1.00939476e+00 -7.86191940e-01 -4.37056497e-02
3.23675424e-01 -1.08905174e-01 2.57528126e-01 1.14206183e+00
-1.07627273e+00 4.71037358e-01 -1.10878490e-01 -1.18842947e+00
-1.27173555e+00 3.22466254e-01 3.29009503e-01 -4.25302424e-02
-7.82805562e-01 1.00842643e+00 1.17802434e-01 2.05292359e-01
1.83433503e-01 -3.72621864e-01 8.86149704e-02 -2.67165273e-01
4.71043259e-01 7.43664026e-01 2.47413605e-01 -6.26126468e-01
-6.76747441e-01 5.16461134e-01 3.56898010e-02 -1.13887765e-01
7.08788753e-01 -3.68926346e-01 5.11905909e-01 6.30348384e-01
1.02600002e+00 3.65859926e-01 -1.63192427e+00 5.79528436e-02
-2.08264083e-01 -3.31512213e-01 2.06991822e-01 -4.90583867e-01
-8.13678026e-01 1.04519713e+00 9.80840087e-01 2.97312587e-01
9.64284778e-01 3.11185837e-01 1.09662819e+00 2.53915370e-01
3.99534285e-01 -9.33626831e-01 8.59001204e-02 5.52240133e-01
4.21880633e-01 -1.15393341e+00 -2.53462885e-02 -5.15140414e-01
-4.77391690e-01 9.72526014e-01 9.06754315e-01 -4.28835481e-01
4.61075693e-01 7.42586851e-01 3.71304482e-01 -1.15956999e-01
-1.16291964e+00 -2.73545265e-01 -2.88638890e-01 4.99841452e-01
2.89786488e-01 1.10958584e-01 -8.08257833e-02 1.44745126e-01
-3.63999456e-01 -5.08243330e-02 9.82683957e-01 8.37951660e-01
-8.31522584e-01 -7.38899171e-01 -5.19486070e-01 2.66868114e-01
-3.85936469e-01 -1.68511555e-01 3.33258480e-01 9.33694184e-01
3.09247494e-01 1.34030247e+00 3.27426732e-01 -5.22692144e-01
5.05910099e-01 -4.44553733e-01 1.20657928e-01 -5.85003853e-01
-2.96467841e-01 2.83181906e-01 4.83459175e-01 -1.00990999e+00
-1.36395499e-01 -6.68229759e-01 -1.45520544e+00 -6.72373533e-01
-2.21056879e-01 4.60763909e-02 6.00870848e-01 5.23531616e-01
6.35041535e-01 6.22543514e-01 5.23399055e-01 -9.19224918e-01
-3.80771041e-01 -7.78698266e-01 -6.10028386e-01 -3.76551718e-01
5.23439944e-01 -6.86670303e-01 -5.43924689e-01 -1.85926128e-02] | [8.5990629196167, -0.3257101774215698] |
0047f076-eddf-4164-bb25-952146632337 | ill-posed-image-reconstruction-without-an | 2304.05589 | null | https://arxiv.org/abs/2304.05589v1 | https://arxiv.org/pdf/2304.05589v1.pdf | Ill-Posed Image Reconstruction Without an Image Prior | We consider solving ill-posed imaging inverse problems without access to an image prior or ground-truth examples. An overarching challenge in these inverse problems is that an infinite number of images, including many that are implausible, are consistent with the observed measurements. Thus, image priors are required to reduce the space of possible solutions to more desireable reconstructions. However, in many applications it is difficult or potentially impossible to obtain example images to construct an image prior. Hence inaccurate priors are often used, which inevitably result in biased solutions. Rather than solving an inverse problem using priors that encode the spatial structure of any one image, we propose to solve a set of inverse problems jointly by incorporating prior constraints on the collective structure of the underlying images. The key assumption of our work is that the underlying images we aim to reconstruct share common, low-dimensional structure. We show that such a set of inverse problems can be solved simultaneously without the use of a spatial image prior by instead inferring a shared image generator with a low-dimensional latent space. The parameters of the generator and latent embeddings are found by maximizing a proxy for the Evidence Lower Bound (ELBO). Once identified, the generator and latent embeddings can be combined to provide reconstructed images for each inverse problem. The framework we propose can handle general forward model corruptions, and we show that measurements derived from only a small number of ground-truth images ($\leqslant 150$) are sufficient for "prior-free" image reconstruction. We demonstrate our approach on a variety of convex and non-convex inverse problems, ranging from denoising, phase retrieval, and black hole video reconstruction. | ['Katherine L. Bouman', 'He Sun', 'Angela F. Gao', 'Oscar Leong'] | 2023-04-12 | null | null | null | null | ['image-reconstruction', 'video-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 5.16414046e-01 3.44702512e-01 2.33206257e-01 -2.16395333e-01
-9.32379127e-01 -6.47175968e-01 4.75264847e-01 -6.32883310e-01
-3.81907284e-01 6.89926565e-01 2.06439450e-01 -1.32420287e-02
-5.42720139e-01 -4.79358763e-01 -8.05049300e-01 -9.90380704e-01
2.79650867e-01 5.45758188e-01 -2.72664189e-01 -2.95213535e-02
2.70376354e-01 3.00852686e-01 -1.19141626e+00 -8.70549772e-03
5.95689178e-01 8.24055254e-01 6.95708275e-01 6.61887944e-01
4.02457237e-01 5.88619351e-01 -2.39320964e-01 -1.24563232e-01
4.54406738e-01 -6.24246359e-01 -6.98074400e-01 5.36144733e-01
8.17702264e-02 -5.23185849e-01 -4.32870924e-01 1.40443707e+00
2.53128320e-01 3.58865559e-02 7.48617768e-01 -1.07302940e+00
-7.41070509e-01 3.32981758e-02 -5.83465099e-01 -3.48279253e-02
3.03352237e-01 9.89924222e-02 9.77804303e-01 -8.84681344e-01
7.39534378e-01 9.24799502e-01 6.24599934e-01 3.71217906e-01
-1.56399918e+00 -1.45128444e-01 -1.18812881e-01 1.21425688e-01
-1.46362984e+00 -8.83780420e-01 8.91841650e-01 -3.05857718e-01
4.41461474e-01 4.26272899e-01 4.78425354e-01 1.01702619e+00
3.76712419e-02 3.10311228e-01 1.17558432e+00 -5.66857278e-01
2.96831071e-01 2.60273963e-01 -2.61090904e-01 6.48647904e-01
2.85270810e-01 8.62022862e-02 -5.43112159e-01 -2.50924110e-01
9.70325410e-01 -1.37165606e-01 -7.40305185e-01 -5.51001012e-01
-1.34372234e+00 7.19893873e-01 3.74031246e-01 3.68700288e-02
-4.77411449e-01 2.89054573e-01 -3.83671731e-01 2.30926767e-01
2.84290642e-01 5.85391104e-01 -2.33072385e-01 2.54186541e-01
-7.26506710e-01 1.06493764e-01 8.13723624e-01 9.09417391e-01
1.08333147e+00 -7.99156576e-02 4.90866661e-01 4.77083594e-01
6.19195938e-01 9.12740767e-01 -7.92569071e-02 -1.76540613e+00
2.56411135e-01 -1.83254302e-01 6.06263518e-01 -1.26007879e+00
8.62763375e-02 -3.38791162e-01 -9.65783000e-01 2.40408286e-01
5.51273584e-01 2.34530680e-02 -9.12949324e-01 2.06220341e+00
1.38361946e-01 2.94813216e-01 8.03036168e-02 1.07624197e+00
2.17141345e-01 7.49621630e-01 -6.14155114e-01 -5.42340577e-01
1.14797139e+00 -3.08565259e-01 -7.69679606e-01 -5.32433808e-01
1.96823463e-01 -8.12645137e-01 8.17869782e-01 5.59174776e-01
-1.20684147e+00 -7.34298304e-02 -1.13777363e+00 -2.22061619e-01
3.86183232e-01 -3.40572745e-02 1.44171551e-01 4.11979616e-01
-9.91710842e-01 3.80782038e-01 -7.70158112e-01 -7.14821517e-02
1.02946691e-01 2.82591671e-01 -4.56515729e-01 -4.73222494e-01
-5.96148074e-01 9.76902902e-01 -4.40382324e-02 3.86787981e-01
-9.59683776e-01 -5.44861972e-01 -7.11083651e-01 -1.63077399e-01
4.72391099e-01 -9.02202487e-01 8.85660291e-01 -9.32935059e-01
-1.18793356e+00 7.80837476e-01 -2.88861424e-01 3.31826881e-02
4.75339025e-01 6.27640635e-02 2.78506633e-02 4.98179048e-01
3.15479755e-01 4.43390965e-01 1.21977556e+00 -1.82122171e+00
-5.56833707e-02 -3.36494714e-01 2.96475962e-02 3.44600767e-01
7.19483942e-02 -4.84613478e-01 -5.97979665e-01 -2.09688947e-01
8.15450728e-01 -1.15858781e+00 -4.21471208e-01 2.52476722e-01
-4.77975309e-01 7.28479922e-01 7.49558628e-01 -7.55237699e-01
5.64151168e-01 -2.22053504e+00 6.16622031e-01 4.34825987e-01
3.40142667e-01 -4.02620345e-01 -3.50116611e-01 2.03726962e-01
4.84596454e-02 1.18235171e-01 -5.92525542e-01 -4.84274030e-01
-3.39336954e-02 7.35913515e-01 -5.29086113e-01 1.11036432e+00
-1.35251373e-01 7.43757427e-01 -8.00949037e-01 -3.11548322e-01
3.87322456e-01 5.83161712e-01 -7.23777771e-01 2.32631117e-01
-1.42101035e-01 1.01750600e+00 -4.81324166e-01 1.42642960e-01
8.10017824e-01 -5.33852816e-01 1.88885882e-01 -4.33606893e-01
4.92034219e-02 2.93045361e-02 -1.51544833e+00 1.81126523e+00
-5.02566934e-01 4.97089952e-01 6.18214130e-01 -1.24018002e+00
4.12811697e-01 3.26764673e-01 6.78658247e-01 -4.46428090e-01
-4.06491570e-02 1.68124288e-01 -5.02847508e-02 -6.67398214e-01
2.93775648e-01 -5.81766427e-01 8.18309337e-02 6.87685728e-01
2.78277732e-02 -5.68506062e-01 -2.51992047e-01 2.50185698e-01
1.14897525e+00 -1.16629398e-03 5.71122877e-02 -1.56816036e-01
2.19542667e-01 -1.19320385e-01 6.97273672e-01 8.26741815e-01
1.73362762e-01 1.17034566e+00 3.25815767e-01 -2.47755438e-01
-1.37970984e+00 -1.24806881e+00 -3.24101835e-01 2.35192344e-01
3.36558491e-01 -1.51781768e-01 -6.42352283e-01 -1.09416515e-01
-4.27353144e-01 5.57359517e-01 -4.21172738e-01 1.47133917e-01
-4.76064414e-01 -9.27052081e-01 1.26054451e-01 -2.31485218e-01
5.42496145e-01 -6.41829848e-01 -5.23076355e-01 2.61975192e-02
-8.05073142e-01 -1.40061152e+00 -3.71113956e-01 7.19431862e-02
-7.51897454e-01 -1.07510936e+00 -5.95463336e-01 -2.90177673e-01
1.14961565e+00 5.50186396e-01 1.07129872e+00 1.13851078e-01
-2.26333335e-01 8.39797080e-01 -7.57053681e-03 8.13881457e-02
-3.25202495e-01 -6.84882224e-01 1.60981610e-01 3.31848443e-01
-3.64691317e-01 -9.23315406e-01 -6.21810317e-01 3.09997410e-01
-1.05458558e+00 1.55507028e-01 5.93272626e-01 1.00693882e+00
7.44159043e-01 1.20334923e-01 7.60508925e-02 -5.66585958e-01
2.25778624e-01 -5.08742511e-01 -6.04943752e-01 2.03218713e-01
-4.05114502e-01 3.91154498e-01 3.12655360e-01 -4.00571436e-01
-1.19056189e+00 1.88088804e-01 -1.39925284e-02 -6.17550850e-01
5.34131564e-02 5.16552687e-01 -4.16803449e-01 -2.75442481e-01
6.74277961e-01 3.20961803e-01 2.94704944e-01 -4.42670345e-01
6.09753311e-01 3.04515749e-01 6.87918425e-01 -5.69305122e-01
1.01198244e+00 1.04661071e+00 3.34237367e-01 -1.25836313e+00
-9.38526273e-01 -3.11743557e-01 -5.65544546e-01 -5.43465465e-02
9.52724278e-01 -8.97783101e-01 -5.99649787e-01 1.17881052e-01
-1.26662123e+00 -1.29500479e-01 -4.66473818e-01 5.61117649e-01
-9.62637365e-01 7.16232479e-01 -2.60934204e-01 -8.24615300e-01
3.90175998e-01 -1.26998818e+00 1.05017662e+00 -2.75269300e-01
-3.61928865e-02 -9.65952814e-01 -5.16034774e-02 4.86360669e-01
1.37206495e-01 8.95171463e-02 7.26017892e-01 2.40147173e-01
-1.26885402e+00 1.75255630e-02 -1.94202527e-01 4.29149717e-01
3.69538218e-02 -5.01668394e-01 -9.58735228e-01 -3.12099755e-01
9.89398122e-01 -2.39270627e-01 7.47616768e-01 7.29571223e-01
1.02542675e+00 -4.99884933e-01 -1.68602020e-01 1.14075339e+00
1.53306758e+00 -1.94500804e-01 8.36949646e-01 -5.45232333e-02
6.03152215e-01 6.68922246e-01 1.56375617e-01 3.59709978e-01
2.52147973e-01 5.81082523e-01 6.73121035e-01 7.86109790e-02
8.36499315e-03 -8.24548453e-02 2.37273708e-01 8.17489326e-01
-2.35518575e-01 -3.85937840e-01 -6.64323688e-01 5.74861526e-01
-1.68691504e+00 -1.03313899e+00 -2.19966650e-01 2.28035784e+00
4.53299701e-01 -4.10782725e-01 -6.25284612e-01 3.21707465e-02
4.04817194e-01 1.79802313e-01 -6.56062007e-01 3.26373696e-01
-3.38216782e-01 4.75657508e-02 6.15601659e-01 8.13026011e-01
-8.02830637e-01 2.87640244e-01 6.80265188e+00 4.27039891e-01
-7.42101073e-01 3.16219836e-01 4.82681006e-01 1.04952827e-02
-7.78109550e-01 5.33696771e-01 -2.12866142e-01 3.02031398e-01
6.80923522e-01 -6.03620932e-02 9.12835240e-01 2.44489923e-01
2.40535513e-01 -4.28243726e-01 -1.18133914e+00 1.29156959e+00
1.42222688e-01 -1.40116501e+00 -1.47159562e-01 4.83831018e-01
7.98610330e-01 1.51546866e-01 2.61885434e-01 -5.14586866e-01
3.45364541e-01 -9.76402342e-01 7.80959368e-01 6.39529109e-01
8.84054601e-01 -2.44559065e-01 2.53458112e-01 6.78894758e-01
-6.48996472e-01 -5.42522632e-02 -4.41481531e-01 -1.30292892e-01
6.25419021e-01 8.19298863e-01 -2.26968870e-01 4.98615503e-01
6.33723319e-01 6.02037251e-01 7.99058005e-02 6.32408321e-01
-3.39616954e-01 2.49135047e-01 -6.90390229e-01 7.18353927e-01
-1.56339437e-01 -7.13672578e-01 7.86673188e-01 4.68825608e-01
7.92432904e-01 7.96917617e-01 2.88069490e-02 1.36919332e+00
1.71523169e-02 -5.32392263e-01 -7.78091967e-01 1.41577482e-01
2.19922349e-01 1.20396030e+00 -6.37762785e-01 -2.23876536e-02
-4.19526815e-01 1.04428244e+00 -1.29186764e-01 8.55274856e-01
-5.45444012e-01 3.55506033e-01 6.88736260e-01 9.48297456e-02
9.90878940e-02 -4.97246861e-01 -3.17792356e-01 -1.56918693e+00
1.95204079e-01 -8.98006856e-01 -4.62288745e-02 -1.05872428e+00
-1.30296087e+00 1.37607262e-01 1.62423775e-01 -1.14788234e+00
-4.20588285e-01 -4.46957856e-01 -3.54063600e-01 1.06238103e+00
-1.33724749e+00 -9.35235083e-01 -2.89914876e-01 5.80740511e-01
1.14848316e-01 2.80083150e-01 7.45643735e-01 -1.48345456e-02
-9.60278437e-02 -2.90723711e-01 2.70415574e-01 -1.93107218e-01
3.80505800e-01 -1.04676366e+00 -1.18344977e-01 9.86956894e-01
3.31132919e-01 5.60878873e-01 1.02861726e+00 -4.98359621e-01
-1.95820332e+00 -6.29636824e-01 4.55706447e-01 -5.57516396e-01
5.70730090e-01 -2.92109847e-01 -7.54155338e-01 1.01709425e+00
1.15781546e-01 4.48645890e-01 3.67402464e-01 -2.71617800e-01
-2.51942277e-01 1.31107509e-01 -1.13949907e+00 3.74280900e-01
1.02802396e+00 -7.92733669e-01 -3.99805069e-01 4.33488816e-01
1.93143472e-01 -3.66566092e-01 -6.50645196e-01 1.68621957e-01
3.19380522e-01 -1.04585063e+00 1.39737904e+00 -1.29710466e-01
4.99946654e-01 -5.25128067e-01 -6.60480976e-01 -1.19844019e+00
-1.44563392e-01 -6.53677881e-01 3.55405360e-02 6.65626705e-01
1.74794078e-01 -5.98876119e-01 7.12025642e-01 8.70075881e-01
1.94085017e-02 -1.78296223e-01 -1.11596382e+00 -7.09172964e-01
-1.42778262e-01 -6.05889857e-01 1.21059179e-01 9.55095649e-01
-4.41207558e-01 2.45529726e-01 -7.54824936e-01 7.25329101e-01
1.36714399e+00 5.89576922e-02 6.93329334e-01 -9.98166561e-01
-7.26404130e-01 1.63459554e-01 -1.94535270e-01 -1.48605943e+00
1.30392864e-01 -7.86564946e-01 4.73007709e-01 -1.37425554e+00
3.91632825e-01 -6.90830767e-01 1.67393312e-01 -5.34512987e-03
3.05918723e-01 3.77725512e-01 2.70590514e-01 5.84922194e-01
-2.78816789e-01 6.13170087e-01 1.27797902e+00 -1.30424723e-01
1.48356959e-01 -2.01419473e-01 -7.06227362e-01 9.65324819e-01
2.41114289e-01 -6.00263715e-01 -6.40146792e-01 -8.65449011e-01
5.87799668e-01 4.45083290e-01 7.60073483e-01 -6.61515236e-01
3.68319511e-01 -3.25872809e-01 3.82023871e-01 -2.00158849e-01
8.67538571e-01 -8.14621270e-01 7.91953266e-01 1.11687355e-01
-4.37471047e-02 -2.86290199e-01 -3.54543686e-01 7.39735246e-01
-6.35997355e-02 -6.36719823e-01 7.17586458e-01 -3.89852703e-01
-4.74437296e-01 2.52228260e-01 -1.30974457e-01 -7.66021153e-03
6.69824123e-01 -1.26527786e-01 -1.27287999e-01 -8.71274710e-01
-1.01810694e+00 -1.23247080e-01 7.87381172e-01 -1.87683284e-01
8.93917561e-01 -1.16050160e+00 -6.05925083e-01 3.77483666e-01
-1.45284384e-01 1.40882015e-01 3.36240828e-01 9.58981991e-01
-4.03216392e-01 -2.51316093e-03 -1.17419148e-02 -8.73871326e-01
-8.90277028e-01 4.46508944e-01 5.00173509e-01 9.31072906e-02
-8.77269745e-01 6.47989392e-01 5.91320336e-01 -4.24631298e-01
-2.19324261e-01 -6.54551238e-02 4.94847238e-01 -3.04140061e-01
3.91691595e-01 1.96910962e-01 -3.10056716e-01 -9.11586165e-01
-1.09296076e-01 8.10900629e-01 4.14422452e-01 -8.00928414e-01
1.63059878e+00 -6.69863343e-01 -4.07887429e-01 1.80044234e-01
1.32519794e+00 1.05908699e-01 -1.51380551e+00 -2.01645687e-01
-2.78246611e-01 -6.99169040e-01 3.67757797e-01 -3.16988677e-01
-1.09325898e+00 7.71066546e-01 3.28565627e-01 9.21314061e-02
1.29372990e+00 2.30466440e-01 4.38693404e-01 5.26929319e-01
5.19569218e-01 -7.90502369e-01 1.63928300e-01 2.93336421e-01
1.07452667e+00 -1.29239237e+00 3.09682667e-01 -4.19625133e-01
-1.72918990e-01 8.13625157e-01 -9.41543952e-02 -2.38210723e-01
8.00036728e-01 2.62401521e-01 -1.23130441e-01 -3.72710943e-01
-4.54686016e-01 -2.23643426e-02 1.60236910e-01 6.26197875e-01
-1.86889723e-01 -1.71708211e-01 1.20782904e-01 1.13337249e-01
-9.05381143e-02 -1.47798136e-01 8.97851467e-01 6.98432565e-01
-3.48837435e-01 -1.17695999e+00 -7.24289656e-01 2.18660772e-01
-2.92409807e-01 5.48034459e-02 -2.31130142e-02 6.37836635e-01
9.11390781e-02 9.51243877e-01 -5.43498695e-02 -8.21434930e-02
-1.00856848e-01 -3.35287482e-01 9.16270077e-01 -4.55288231e-01
4.78296369e-01 3.24905932e-01 -8.64576846e-02 -6.14128292e-01
-6.48205936e-01 -7.62499094e-01 -9.95055795e-01 -3.12850684e-01
-2.72093624e-01 5.78645542e-02 8.10390711e-01 1.02188563e+00
1.37142003e-01 8.57417285e-02 5.43998539e-01 -1.12976241e+00
-5.71174860e-01 -5.44648111e-01 -8.07705700e-01 4.58934575e-01
7.16074586e-01 -5.43743968e-01 -7.99459398e-01 3.34474683e-01] | [11.623950958251953, -2.3373100757598877] |
849d2925-efc5-47ea-bf45-eaa1d7a5f116 | image-based-3d-object-reconstruction-state-of | 1906.06543 | null | https://arxiv.org/abs/1906.06543v3 | https://arxiv.org/pdf/1906.06543v3.pdf | Image-based 3D Object Reconstruction: State-of-the-Art and Trends in the Deep Learning Era | 3D reconstruction is a longstanding ill-posed problem, which has been explored for decades by the computer vision, computer graphics, and machine learning communities. Since 2015, image-based 3D reconstruction using convolutional neural networks (CNN) has attracted increasing interest and demonstrated an impressive performance. Given this new era of rapid evolution, this article provides a comprehensive survey of the recent developments in this field. We focus on the works which use deep learning techniques to estimate the 3D shape of generic objects either from a single or multiple RGB images. We organize the literature based on the shape representations, the network architectures, and the training mechanisms they use. While this survey is intended for methods which reconstruct generic objects, we also review some of the recent works which focus on specific object classes such as human body shapes and faces. We provide an analysis and comparison of the performance of some key papers, summarize some of the open problems in this field, and discuss promising directions for future research. | ['Xian-Feng Han', 'Mohammed Bennamoun', 'Hamid Laga'] | 2019-06-15 | null | null | null | null | ['3d-object-reconstruction'] | ['computer-vision'] | [ 2.49254629e-01 1.16092004e-01 -3.50200161e-02 -3.79909217e-01
-1.24891922e-01 -1.91173673e-01 3.15391302e-01 -3.80061001e-01
-1.27917811e-01 2.55142808e-01 5.97323589e-02 2.89982539e-02
-4.63200584e-02 -8.16838801e-01 -6.14138126e-01 -6.10429347e-01
-1.37414530e-01 4.96660203e-01 -9.95683074e-02 1.93247944e-02
7.40737692e-02 1.19707930e+00 -1.72058547e+00 1.41525760e-01
2.42437929e-01 1.46949720e+00 -2.28108346e-01 4.49434012e-01
-3.78349535e-02 3.06681603e-01 -2.39142552e-01 -5.00577033e-01
3.13708484e-01 -2.15902060e-01 -6.43032968e-01 4.87534076e-01
6.53077364e-01 -5.67284048e-01 -5.18253028e-01 6.69746459e-01
8.97616506e-01 -9.01058763e-02 7.18887329e-01 -9.85441387e-01
-8.20455968e-01 -6.41223490e-02 -5.08557200e-01 -1.29449591e-01
2.75601476e-01 -1.15084670e-01 6.10922456e-01 -1.10750139e+00
5.69025934e-01 1.38754332e+00 1.06810951e+00 8.37672770e-01
-9.22895908e-01 -4.83859718e-01 -1.14619806e-02 9.15799066e-02
-1.17592013e+00 -3.87285203e-01 1.13741267e+00 -3.93571347e-01
1.02323294e+00 7.96130225e-02 9.14886177e-01 8.59619081e-01
1.25359893e-01 9.66021955e-01 8.67086470e-01 -4.48863894e-01
1.00397639e-01 -3.76878262e-01 -9.90354419e-02 1.05117559e+00
3.18455815e-01 2.02577084e-01 -4.47111934e-01 -1.73980340e-01
1.52306378e+00 -1.46012366e-01 -1.49056017e-01 -9.59835768e-01
-9.54920411e-01 9.47350502e-01 6.42095447e-01 2.01494724e-01
-3.72762322e-01 3.13217402e-01 2.59331226e-01 -6.17543422e-02
7.19131529e-01 9.80894491e-02 -3.83237362e-01 1.66706204e-01
-7.97656476e-01 4.25145864e-01 7.88065076e-01 7.10049450e-01
4.85196412e-01 5.31624138e-01 4.26572412e-01 9.81272817e-01
3.85444105e-01 4.93794918e-01 -2.72578932e-02 -1.11584151e+00
-7.80750513e-02 5.12064576e-01 -7.52727613e-02 -1.16084170e+00
-5.67476571e-01 -3.36380631e-01 -1.05479026e+00 5.86051762e-01
2.07630873e-01 1.10621573e-02 -8.94719481e-01 1.21919334e+00
4.91511106e-01 1.04467189e-02 -3.26496780e-01 1.07502484e+00
1.60816896e+00 3.33241135e-01 -2.80576199e-01 6.75791353e-02
1.07128620e+00 -7.44333327e-01 -5.51282465e-01 -4.13119607e-02
-1.68976486e-01 -7.60859430e-01 5.45454502e-01 2.73765087e-01
-1.47366273e+00 -6.10302687e-01 -1.00426424e+00 -3.42359751e-01
-2.17317820e-01 2.52061635e-01 7.91106224e-01 6.92079961e-01
-1.10031080e+00 9.01152134e-01 -1.00885367e+00 -5.28201044e-01
7.83125341e-01 5.15542924e-01 -3.40152264e-01 -1.84753209e-01
-7.01178491e-01 8.52224946e-01 -1.79146171e-01 3.03331465e-01
-6.56298518e-01 -5.63742638e-01 -9.69654620e-01 -2.63482779e-01
4.64067981e-03 -9.15693641e-01 1.30460882e+00 -7.11627901e-01
-1.61174560e+00 1.45254171e+00 1.31618036e-02 -2.15410218e-01
4.04712200e-01 -2.01480016e-01 -1.01376608e-01 2.51570940e-01
-3.27739060e-01 7.63542175e-01 8.42170000e-01 -1.26344144e+00
-1.54773891e-01 -6.51111960e-01 1.06307007e-01 1.73483908e-01
-9.82073043e-03 7.97655359e-02 -4.62780744e-01 -7.57059455e-01
5.91120064e-01 -8.51162374e-01 -3.20142657e-01 7.86984801e-01
-1.93574548e-01 -3.30980062e-01 6.78608358e-01 -3.22217673e-01
6.25226140e-01 -2.01912713e+00 2.94188082e-01 -6.72422200e-02
2.62419671e-01 1.50570288e-01 2.30832826e-02 2.74005771e-01
-2.07660168e-01 -1.23179145e-01 -6.37631655e-01 -8.45566213e-01
-1.02696598e-01 3.27828318e-01 -1.22069605e-01 8.40794921e-01
2.09550500e-01 1.07079148e+00 -4.60519433e-01 -2.26304606e-01
4.95115548e-01 1.06295788e+00 -3.86626601e-01 2.28048638e-01
-2.18714699e-02 5.63614368e-01 -2.58547932e-01 9.09333050e-01
8.71470690e-01 -3.03199738e-01 1.96037907e-02 -4.07890320e-01
-2.81691402e-02 5.57581857e-02 -1.12567759e+00 1.67793393e+00
-3.66518289e-01 4.82912034e-01 4.61484939e-01 -1.28331339e+00
1.13263178e+00 4.58143651e-01 8.40488374e-01 -4.87005502e-01
1.19964451e-01 3.70460004e-01 -2.51238793e-01 -5.89254558e-01
2.46929988e-01 -5.72673678e-01 2.20991343e-01 5.32682359e-01
4.57763672e-02 -4.30508286e-01 -4.79911417e-01 -3.50426972e-01
7.02319682e-01 5.75498164e-01 3.90793085e-01 -3.92205752e-02
4.79420573e-01 -1.43278360e-01 2.86939532e-01 3.54907691e-01
-1.70952678e-01 1.07838988e+00 -8.33909959e-02 -1.21677351e+00
-9.64631677e-01 -1.05645454e+00 -3.71708781e-01 5.78803301e-01
1.87530130e-01 -5.18307388e-02 -4.86834079e-01 -3.46440554e-01
2.31271043e-01 -4.02853591e-03 -8.51968288e-01 5.31610921e-02
-9.66584802e-01 -7.52238750e-01 3.36054504e-01 8.26178968e-01
7.05525398e-01 -1.43442249e+00 -9.18630183e-01 8.15395415e-02
-9.03705955e-02 -9.96720791e-01 2.08645493e-01 -1.31365180e-01
-1.55796528e+00 -1.17455637e+00 -1.09344065e+00 -9.21530068e-01
7.13331878e-01 2.62051523e-01 1.42391622e+00 2.93207109e-01
-4.59144711e-01 8.80938649e-01 -1.48990050e-01 -4.28670466e-01
-5.66914529e-02 -1.84633881e-01 9.78579894e-02 -8.48179758e-02
2.06110835e-01 -6.85029626e-01 -5.97748339e-01 2.26292908e-01
-9.14101243e-01 -7.19587039e-03 3.85719538e-01 5.34261584e-01
8.15688610e-01 -2.93931216e-01 3.17013979e-01 -4.47704285e-01
2.77437717e-01 -3.19084823e-01 -3.66031885e-01 1.23047465e-02
-1.70928180e-01 -2.06340551e-01 1.38239235e-01 -1.93678841e-01
-9.84439313e-01 3.50242138e-01 -6.75698042e-01 -4.56151158e-01
-4.01284009e-01 3.71896476e-01 -3.59288044e-03 -4.50089633e-01
5.55046499e-01 1.45420715e-01 2.61156797e-01 -9.76359785e-01
1.91947281e-01 3.40286016e-01 4.92394030e-01 -5.78343868e-01
5.32425642e-01 9.46321666e-01 2.35939920e-01 -1.06526077e+00
-8.40626538e-01 -2.53726691e-01 -9.27329719e-01 -4.86719102e-01
6.82736993e-01 -6.85513020e-01 -6.75025642e-01 7.58954287e-01
-1.41413486e+00 -2.31814981e-01 -3.77384901e-01 3.12009156e-01
-9.84902203e-01 3.78445655e-01 -7.93546379e-01 -7.93632746e-01
-5.72986603e-01 -1.23906255e+00 1.31368661e+00 2.32987970e-01
-7.68962279e-02 -1.17374337e+00 6.22935593e-02 3.29588592e-01
4.31434095e-01 6.52963758e-01 9.03146029e-01 4.28426033e-03
-6.58466578e-01 -2.85895944e-01 -2.46779889e-01 2.53557891e-01
1.93018973e-01 -1.53158188e-01 -1.05604994e+00 -2.45503828e-01
1.36350721e-01 -3.95828962e-01 8.13823760e-01 8.59533727e-01
1.42168069e+00 1.07472599e-01 -2.89825410e-01 9.07355011e-01
1.46134543e+00 -8.54598284e-02 5.56305349e-01 8.78954306e-02
6.26013279e-01 6.13651335e-01 9.12036970e-02 3.73051673e-01
3.40480328e-01 6.16371930e-01 9.68138874e-01 -2.18968987e-01
-4.73312855e-01 1.72502145e-01 -1.49956912e-01 8.04281533e-01
-7.73132503e-01 1.35877788e-01 -6.77530229e-01 3.60545188e-01
-1.57045245e+00 -5.69286525e-01 -2.09247097e-02 2.01568818e+00
3.94184411e-01 -2.72139281e-01 2.50173151e-01 3.28302234e-01
6.71556592e-01 1.44727170e-01 -7.07091689e-01 -2.63281405e-01
-1.88200951e-01 6.84583724e-01 8.73923823e-02 1.94522932e-01
-1.19865048e+00 7.27340877e-01 7.81906176e+00 4.18715566e-01
-1.32308745e+00 -1.65322751e-01 5.75979412e-01 1.79606646e-01
-1.71880890e-02 -5.94722927e-01 -5.20346820e-01 -1.43420368e-01
3.63231063e-01 3.62163246e-01 2.24843740e-01 9.17466998e-01
-9.52742994e-02 -4.41425592e-02 -1.03873956e+00 1.34460247e+00
4.89331394e-01 -1.50404477e+00 -1.18670668e-02 1.64065003e-01
7.68265605e-01 1.95548505e-01 2.54192501e-01 -1.69635579e-01
-1.11776426e-01 -1.25706756e+00 6.87254310e-01 6.25809014e-01
9.50596333e-01 -6.76739454e-01 4.61931348e-01 2.28006408e-01
-1.21226346e+00 1.79204285e-01 -7.09100246e-01 -3.54830563e-01
2.54523665e-01 7.04362810e-01 -1.79419786e-01 5.50243855e-01
1.02216434e+00 9.98899162e-01 -1.37140900e-01 1.36561286e+00
-1.67094752e-01 4.10379827e-01 -4.20159936e-01 8.32646415e-02
-3.06031778e-02 -2.33503133e-01 4.64695483e-01 8.89895618e-01
2.72451252e-01 3.42867613e-01 2.47255191e-02 9.90035176e-01
-2.27415994e-01 -5.48280105e-02 -7.16142058e-01 1.37046158e-01
-1.34954825e-01 1.07633400e+00 -8.67657781e-01 -1.08845204e-01
-5.23379028e-01 8.42918634e-01 2.03243330e-01 1.58786029e-01
-6.32936418e-01 -1.68288976e-01 6.97149873e-01 3.34715247e-01
3.63393992e-01 -6.95022523e-01 -6.16928518e-01 -9.65881169e-01
1.21167731e-02 -4.35031652e-01 2.63114840e-01 -8.32744181e-01
-1.42601264e+00 6.23806715e-01 1.75731033e-01 -1.17882776e+00
4.98441383e-02 -9.56497133e-01 -2.87122875e-01 5.46219707e-01
-1.56466973e+00 -1.03987026e+00 -3.32516879e-01 3.86544764e-01
5.40074587e-01 -1.13141425e-02 1.02779019e+00 2.11689502e-01
-2.96381801e-01 1.93688259e-01 -5.44974469e-02 2.52402514e-01
2.05883160e-01 -7.71510363e-01 6.20911002e-01 1.16772674e-01
8.11925307e-02 3.98510635e-01 3.62657338e-01 -5.29845834e-01
-1.56050169e+00 -7.24004447e-01 6.87686086e-01 -2.92193532e-01
-5.45329154e-02 -2.77045667e-01 -7.86790252e-01 6.94872558e-01
8.72047693e-02 4.21507418e-01 6.15139961e-01 -1.48098648e-01
-3.11598390e-01 8.99831802e-02 -1.38597357e+00 3.15128326e-01
1.27517831e+00 -1.58191577e-01 -5.40815353e-01 6.61448538e-02
2.61241555e-01 -8.20008397e-01 -8.36375475e-01 7.01171935e-01
9.75861669e-01 -1.41428924e+00 1.39966905e+00 -5.10534883e-01
5.12995660e-01 4.10997719e-02 -1.75130382e-01 -1.03881562e+00
-4.22723770e-01 -9.26674157e-02 -4.05274630e-01 4.03640449e-01
-3.96234483e-01 -3.66347104e-01 1.15170801e+00 5.07278383e-01
-2.66968071e-01 -1.29149425e+00 -1.00003088e+00 -5.67896008e-01
3.25143009e-01 -4.51039672e-01 5.16157508e-01 6.50148809e-01
-4.93662268e-01 9.17817503e-02 -3.86330009e-01 -1.91291660e-01
8.03290725e-01 5.46326518e-01 5.83249569e-01 -1.54238391e+00
1.92533627e-01 -3.75663221e-01 -5.39692521e-01 -1.38881946e+00
1.14298336e-01 -9.01360869e-01 -9.44026932e-03 -1.92386949e+00
-2.85054818e-02 -3.82498145e-01 1.51622310e-01 4.40693885e-01
3.75007540e-01 9.68463480e-01 2.23118868e-02 1.24782875e-01
-2.59428382e-01 6.72308803e-01 1.57679915e+00 -1.51316583e-01
1.06446609e-01 3.43233317e-01 -3.96722764e-01 1.00262737e+00
7.30384529e-01 -9.91931036e-02 -3.11830342e-02 -7.85192192e-01
2.11813882e-01 -1.11479409e-01 4.84382480e-01 -1.02099299e+00
-1.49007812e-01 1.07696660e-01 9.08336699e-01 -8.77392352e-01
8.72478664e-01 -1.00057673e+00 1.01876684e-01 4.74472940e-01
1.07059730e-02 2.32811533e-02 2.69622952e-01 3.07360172e-01
-3.04990299e-02 -1.55480266e-01 1.13423884e+00 -6.81933045e-01
-5.42380452e-01 7.10307062e-01 -3.55152279e-01 -3.95611674e-01
7.74457216e-01 -6.24226689e-01 3.90123278e-01 -4.58479524e-01
-1.04559875e+00 -2.74361819e-01 3.83956134e-01 2.97110707e-01
1.13540459e+00 -1.62373829e+00 -6.09671474e-01 3.59592944e-01
-7.44636655e-02 2.41935551e-01 3.50882292e-01 4.90443885e-01
-7.51254976e-01 3.15394580e-01 -4.86173540e-01 -7.33620703e-01
-1.24948764e+00 3.46032321e-01 6.78526044e-01 2.17704728e-01
-8.17235053e-01 8.95369232e-01 8.45397040e-02 -6.44635856e-01
4.30944085e-01 -2.39672959e-01 -2.43630543e-01 -4.19487715e-01
3.49257261e-01 5.02803445e-01 2.48561710e-01 -1.01685309e+00
-4.72258598e-01 1.13810432e+00 5.18998265e-01 2.49012038e-01
1.76673448e+00 -1.62465240e-05 -3.16590846e-01 2.96801537e-01
1.24525762e+00 -4.66914266e-01 -1.27060544e+00 -3.53977948e-01
-4.56372291e-01 -4.02176142e-01 -1.79404381e-03 -5.75679541e-01
-1.63656461e+00 1.31662762e+00 5.15494347e-01 -3.42693590e-02
1.18666017e+00 3.22441876e-01 7.82566547e-01 1.31386682e-01
5.01477599e-01 -6.28120363e-01 4.55390900e-01 6.87415123e-01
1.39928699e+00 -1.06820118e+00 3.59557629e-01 -6.35258675e-01
-5.15689515e-02 1.55748665e+00 4.98929381e-01 -6.22083485e-01
1.22521853e+00 2.06233293e-01 -9.45937112e-02 -4.75374132e-01
-2.66909022e-02 -1.46450192e-01 4.20505106e-01 9.13861811e-01
6.48925781e-01 -7.98938423e-02 5.96854575e-02 3.75509232e-01
-2.13377237e-01 9.97521952e-02 1.69050395e-01 9.31215227e-01
-3.23290229e-01 -1.01982558e+00 -5.54604053e-01 3.75614554e-01
-2.60164082e-01 3.12386870e-01 -4.39744592e-01 6.66219592e-01
2.60300785e-01 4.40606147e-01 3.36201340e-01 -6.91064000e-02
4.75562245e-01 -4.33029234e-02 1.20622742e+00 -5.42585492e-01
-2.95116514e-01 2.34637801e-02 -1.80930197e-01 -5.24370670e-01
-1.03557694e+00 -5.54253638e-01 -1.10084856e+00 -3.26286405e-01
-4.83898148e-02 -5.00648975e-01 8.55356812e-01 8.51245284e-01
2.56000221e-01 2.29049921e-01 2.50012487e-01 -1.54752862e+00
-2.57245749e-01 -7.18693316e-01 -6.27948940e-01 1.22014396e-01
3.71352881e-01 -1.12370396e+00 3.72999173e-04 2.83983257e-02] | [8.442591667175293, -3.488193988800049] |
b18817a8-1f7e-4bdc-b8e5-c9b870109c1f | gaittake-gait-recognition-by-temporal | 2207.03608 | null | https://arxiv.org/abs/2207.03608v2 | https://arxiv.org/pdf/2207.03608v2.pdf | GaitTAKE: Gait Recognition by Temporal Attention and Keypoint-guided Embedding | Gait recognition, which refers to the recognition or identification of a person based on their body shape and walking styles, derived from video data captured from a distance, is widely used in crime prevention, forensic identification, and social security. However, to the best of our knowledge, most of the existing methods use appearance, posture and temporal feautures without considering a learned temporal attention mechanism for global and local information fusion. In this paper, we propose a novel gait recognition framework, called Temporal Attention and Keypoint-guided Embedding (GaitTAKE), which effectively fuses temporal-attention-based global and local appearance feature and temporal aggregated human pose feature. Experimental results show that our proposed method achieves a new SOTA in gait recognition with rank-1 accuracy of 98.0% (normal), 97.5% (bag) and 92.2% (coat) on the CASIA-B gait dataset; 90.4% accuracy on the OU-MVLP gait dataset. | ['Kwang-Ju Kim', 'Hoang Le Uyen Thuc', 'Jenq-Neng Hwang', 'Cheng-Yen Yang', 'Yizhou Wang', 'Hung-Min Hsu'] | 2022-07-07 | null | null | null | null | ['gait-recognition'] | ['computer-vision'] | [-1.29578546e-01 -6.15119755e-01 -2.13983171e-02 -1.72774538e-01
-6.26789451e-01 -4.37821187e-02 2.75933981e-01 -3.09363417e-02
-4.10489142e-01 4.41325665e-01 2.48412341e-01 3.29230338e-01
-1.74367666e-01 -5.81828654e-01 -1.07280195e-01 -8.55300903e-01
-3.75145465e-01 3.05640101e-01 8.03376958e-02 -3.54231633e-02
2.50731081e-01 4.84518558e-01 -1.41858852e+00 -6.55622557e-02
4.74742234e-01 1.06887889e+00 -3.59954983e-01 6.92544580e-01
4.25191402e-01 4.26280558e-01 -3.56239021e-01 -7.44323552e-01
3.28176841e-02 -1.32122710e-01 -5.05181849e-01 -5.19256443e-02
4.13536996e-01 -5.66831410e-01 -6.57659054e-01 8.27650905e-01
8.43027830e-01 2.17907652e-01 7.58958459e-01 -1.22752595e+00
-8.03421617e-01 -2.40848154e-01 -1.00603318e+00 5.49408257e-01
6.69545949e-01 2.70419985e-01 6.28226280e-01 -9.81303215e-01
3.46039921e-01 1.18951690e+00 8.55768085e-01 4.81870294e-01
-8.25654149e-01 -5.46295106e-01 -1.06464460e-01 7.91976631e-01
-1.42811704e+00 -2.27743030e-01 9.66883242e-01 -5.82715869e-01
8.10676754e-01 1.32868057e-02 7.67363429e-01 1.18790007e+00
4.09784496e-01 7.74191797e-01 8.22268546e-01 -3.20947737e-01
-1.55240685e-01 -6.03938818e-01 4.71512198e-01 7.81040132e-01
5.98381400e-01 7.32286945e-02 -6.14319623e-01 -1.07774109e-01
6.37262762e-01 2.29529098e-01 -1.65686503e-01 -2.23653745e-02
-9.23410118e-01 4.88049626e-01 2.13834286e-01 3.08244616e-01
-5.27520001e-01 7.73605704e-02 6.11285329e-01 1.77164027e-03
3.35673481e-01 -2.95117199e-01 9.87628773e-02 -5.97239375e-01
-8.38463068e-01 2.32598588e-01 3.06171179e-01 4.12217408e-01
3.45455289e-01 3.22286755e-01 -2.69794941e-01 7.91359186e-01
4.06867146e-01 1.15017533e+00 7.29452252e-01 -5.60398519e-01
4.28008556e-01 5.68955362e-01 -5.76477498e-02 -1.84021640e+00
-3.69625986e-01 -1.57429231e-03 -8.65160704e-01 -1.11930594e-01
2.36023679e-01 7.08421618e-02 -8.56258273e-01 1.62525475e+00
2.25584209e-01 4.06712592e-01 -2.31849521e-01 9.64742959e-01
7.55035758e-01 3.33213419e-01 9.40053537e-02 -1.43941298e-01
1.35540509e+00 -5.18068433e-01 -7.62995839e-01 -2.38650143e-01
3.35866302e-01 -5.04126251e-01 5.69437325e-01 3.29442114e-01
-8.23692620e-01 -7.82858193e-01 -1.15416312e+00 2.16043681e-01
-3.52667838e-01 3.11561346e-01 2.30487511e-01 9.49333727e-01
-7.52764285e-01 4.84308392e-01 -1.03813303e+00 -6.36840343e-01
1.78199500e-01 4.16494519e-01 -7.31235206e-01 -2.32369065e-01
-1.01014900e+00 6.64486051e-01 1.53396487e-01 3.67073596e-01
-4.09258455e-01 -1.02469875e-02 -1.03356063e+00 -3.03597987e-01
-4.21761870e-02 -4.62938160e-01 5.17965615e-01 -2.03726843e-01
-1.23522508e+00 8.56599391e-01 -2.62377083e-01 -4.15885180e-01
3.87095451e-01 -5.56228280e-01 -7.21754432e-01 5.60963809e-01
3.66276711e-01 1.57233864e-01 1.03375971e+00 -5.94433784e-01
-2.29255661e-01 -9.94980395e-01 -5.02730787e-01 -3.95509154e-02
-5.37351549e-01 1.29982904e-01 -5.94832540e-01 -8.26084495e-01
7.93469772e-02 -8.61429691e-01 2.60444760e-01 -1.20905124e-01
-8.48580375e-02 -1.46296218e-01 1.29779959e+00 -1.47750390e+00
1.39485741e+00 -2.16579676e+00 2.04472005e-01 2.05494344e-01
8.64063725e-02 5.55824041e-01 1.49554675e-02 2.02581361e-01
-9.15867388e-02 -7.16168806e-02 -1.66547790e-01 -3.81129980e-01
-8.08345992e-03 -1.50438992e-03 8.79679173e-02 7.88255274e-01
5.99156246e-02 9.55489814e-01 -8.64853084e-01 -5.83440483e-01
5.07009268e-01 5.93643188e-01 -4.81961876e-01 1.01820707e-01
9.24631417e-01 2.76738435e-01 -4.47982550e-01 1.02463305e+00
5.48876882e-01 2.30095565e-01 -8.11855420e-02 -4.52071667e-01
3.41619313e-01 -5.06951272e-01 -1.09688723e+00 1.43842387e+00
1.03966214e-01 3.89887661e-01 -2.63066918e-01 -1.13426197e+00
1.04183376e+00 4.13533598e-01 9.21830177e-01 -7.16404319e-01
3.60356867e-01 1.62297025e-01 -1.74954817e-01 -8.07155788e-01
3.92169386e-01 1.25976548e-01 -2.30857074e-01 3.74580562e-01
1.69466406e-01 7.47127235e-01 3.97544831e-01 -8.71479809e-02
1.07266748e+00 9.66729745e-02 2.97136068e-01 1.18479639e-01
6.36295080e-01 -5.38526237e-01 7.29162633e-01 4.55830902e-01
-8.38065386e-01 7.27374434e-01 2.18635250e-04 -5.74802279e-01
-8.66274178e-01 -1.07059908e+00 1.65044576e-01 6.31353974e-01
-3.81918624e-03 -5.96740782e-01 -5.98176599e-01 -5.50293028e-01
1.36845559e-01 -9.69130993e-02 -8.59122515e-01 -5.73403239e-01
-9.38678086e-01 -8.09980452e-01 7.57622778e-01 9.21886027e-01
9.88100231e-01 -9.88190174e-01 -6.24636590e-01 2.53878325e-01
-4.29189533e-01 -1.08681655e+00 -7.97551572e-01 -7.42999732e-01
-7.01710582e-01 -1.24588454e+00 -9.86857176e-01 -5.43571770e-01
4.00842220e-01 1.68043658e-01 3.63080889e-01 -3.63507718e-02
-6.18998468e-01 9.22850966e-01 -5.37552178e-01 1.61307231e-01
3.72352213e-01 -3.75958443e-01 5.70386410e-01 5.97568333e-01
5.23464024e-01 -5.60351610e-01 -6.89357638e-01 3.29128772e-01
-3.84227484e-01 -4.46330726e-01 4.59105819e-01 1.01466060e+00
2.80366600e-01 -1.09134145e-01 3.65426928e-01 1.69368625e-01
6.17308319e-01 -3.27347189e-01 1.50554776e-01 3.08110207e-01
-2.74689138e-01 -3.43207091e-01 2.41811365e-01 -3.00726622e-01
-7.09534287e-01 -2.23362163e-01 -3.58182415e-02 -6.73708856e-01
-1.97446808e-01 4.79176998e-01 -1.68153509e-01 -2.23474652e-01
2.59300381e-01 5.91752648e-01 9.26409513e-02 -5.56066811e-01
-1.32499605e-01 7.13523388e-01 9.65740502e-01 -8.64941537e-01
7.04319835e-01 4.91040379e-01 1.19298220e-01 -9.01759088e-01
-1.85225904e-01 -7.21701980e-01 -8.53708446e-01 -5.93800306e-01
1.24160445e+00 -5.84326625e-01 -8.81551862e-01 9.67193902e-01
-9.04176652e-01 4.10200328e-01 2.36510351e-01 5.98922253e-01
-5.76911569e-01 1.05660903e+00 -6.90195262e-01 -1.10696793e+00
-6.69894755e-01 -8.87097120e-01 1.22050285e+00 1.92842960e-01
-3.31881970e-01 -7.33566463e-01 7.09209815e-02 5.82057774e-01
1.83868691e-01 7.02388048e-01 4.25728381e-01 -2.39477098e-01
2.78593171e-02 -7.22197235e-01 -2.90687591e-01 2.25912645e-01
4.38252836e-01 -2.71801930e-02 -6.39313757e-01 -6.19502306e-01
-1.90769747e-01 -2.14005366e-01 7.68094540e-01 2.98230261e-01
7.64718711e-01 -1.37642711e-01 -2.64889419e-01 4.86462682e-01
1.14984179e+00 5.70121586e-01 7.67833650e-01 3.38319957e-01
8.49698365e-01 3.60189587e-01 6.48467422e-01 5.63230515e-01
2.72792727e-01 9.77014780e-01 7.20874593e-02 3.95793259e-01
-7.98085928e-02 -1.29998684e-01 7.16521263e-01 9.76880074e-01
-7.20499814e-01 7.66170025e-02 -1.13719702e+00 6.70045853e-01
-2.11877418e+00 -1.38970423e+00 8.92830491e-02 2.17288041e+00
1.31706789e-01 -5.20194769e-02 5.19367158e-01 5.39860129e-01
1.01232231e+00 8.00135136e-02 -3.14397007e-01 -1.66673377e-01
-4.65210862e-02 1.91428959e-01 3.04356694e-01 7.03480095e-02
-1.38417721e+00 6.20085001e-01 5.76735640e+00 7.80512035e-01
-1.04092097e+00 3.06141134e-02 3.88979375e-01 1.01302288e-01
6.86274529e-01 -5.65568626e-01 -6.06180608e-01 8.05776596e-01
8.54068160e-01 -2.45641962e-01 8.18373784e-02 6.56843543e-01
2.70703793e-01 2.13616461e-01 -6.79132044e-01 1.51825905e+00
5.51408172e-01 -6.87551320e-01 7.95535594e-02 2.49817774e-01
6.88553080e-02 -6.26508057e-01 3.62394512e-01 2.25698516e-01
-2.17572570e-01 -8.59820306e-01 4.35996622e-01 7.78780699e-01
7.75860846e-01 -8.40612769e-01 9.57747102e-01 -9.53395516e-02
-1.79791629e+00 -1.28731579e-01 8.33442733e-02 2.74520498e-02
3.80241871e-01 6.23442940e-02 -2.86895245e-01 8.64718199e-01
9.81016576e-01 1.28294945e+00 -6.76179290e-01 1.19123280e+00
1.35967853e-02 6.58044815e-01 -2.84209788e-01 1.78412482e-01
4.12593819e-02 -6.60358891e-02 6.27138555e-01 1.20507169e+00
4.81942236e-01 2.35727564e-01 3.87490779e-01 2.45149821e-01
4.42885876e-01 8.52039009e-02 -5.89349151e-01 -7.35425055e-02
1.33985355e-01 8.58013630e-01 -5.52820504e-01 -1.46861851e-01
-2.56710678e-01 1.04386854e+00 -8.60208739e-03 1.57942072e-01
-1.04459667e+00 -5.19656718e-01 6.57383025e-01 -1.19242996e-01
2.36763775e-01 -4.89033312e-01 5.40404115e-03 -1.41700053e+00
2.80988872e-01 -6.45835698e-01 7.56902695e-01 -5.87200880e-01
-1.44367385e+00 3.08097601e-01 1.36622787e-01 -1.45156085e+00
-2.42869630e-01 -7.53085792e-01 -7.26088703e-01 5.12589633e-01
-7.07667053e-01 -1.45914793e+00 -3.62693548e-01 7.63974845e-01
3.12322199e-01 -4.26554829e-01 6.17882133e-01 6.18645251e-01
-9.08753574e-01 8.99367929e-01 -9.14319381e-02 7.39700019e-01
6.93439424e-01 -9.20034409e-01 3.54247123e-01 1.07685065e+00
-1.61930203e-01 8.20774555e-01 4.24828678e-01 -8.58149827e-01
-1.47017348e+00 -9.44765270e-01 8.07963490e-01 -3.20390612e-01
4.48047280e-01 1.40447810e-01 -7.55244255e-01 4.45941955e-01
-2.76822358e-01 2.24931352e-02 7.53695309e-01 -1.59715340e-01
-2.94371396e-01 -2.21715137e-01 -1.30487919e+00 4.74195242e-01
1.10267448e+00 -4.06673253e-01 -7.59680927e-01 -1.49127766e-01
4.79087830e-02 -1.09571129e-01 -1.12583351e+00 6.21247351e-01
9.94414032e-01 -6.78946614e-01 1.27120173e+00 -4.68437314e-01
1.85455475e-02 -3.91387373e-01 -4.91667241e-01 -7.84880042e-01
-6.52998269e-01 -3.73553447e-02 -3.75498533e-01 1.10028458e+00
-5.44489622e-01 -6.36822224e-01 7.79435039e-01 4.73001003e-01
3.80538590e-02 -4.23554689e-01 -1.29697096e+00 -1.05204999e+00
-2.62850016e-01 -4.07524407e-01 4.35200661e-01 8.61422479e-01
8.68841410e-02 -1.24377772e-01 -9.17289793e-01 1.92215353e-01
1.03586495e+00 -6.97899535e-02 7.41019666e-01 -1.19903553e+00
-1.50208950e-01 -2.67755598e-01 -1.41362572e+00 -4.01394784e-01
-1.07031815e-01 -4.81189460e-01 -4.52028483e-01 -1.29290390e+00
3.40901107e-01 2.91569889e-01 -5.83883822e-01 5.61136901e-01
-1.52008310e-01 5.88362873e-01 3.25742811e-01 1.68089554e-01
-6.25044942e-01 7.49148846e-01 8.14206362e-01 -4.69342202e-01
1.97171524e-01 -3.73385429e-01 -2.46897772e-01 5.98860323e-01
7.36179054e-01 -1.01100363e-01 -1.21967614e-01 -2.74567127e-01
-5.56499362e-01 6.13387562e-02 6.56478643e-01 -1.49349380e+00
1.97612807e-01 -3.74037288e-02 6.13780856e-01 -7.20319510e-01
8.04317772e-01 -4.37470913e-01 6.58400878e-02 7.64072239e-01
3.64725530e-01 3.50020975e-01 1.60085082e-01 8.04369926e-01
-3.24451625e-01 3.05402249e-01 5.84628999e-01 1.37205422e-01
-1.07778001e+00 5.62445581e-01 -3.89684200e-01 -1.73191845e-01
9.61933315e-01 -7.35525906e-01 -2.61204869e-01 -3.13311934e-01
-1.07564616e+00 4.28255498e-02 2.14766383e-01 6.07058942e-01
1.07574308e+00 -1.85878038e+00 -6.80865645e-01 2.98688412e-01
3.04417163e-01 -8.20159376e-01 6.20324433e-01 9.38894928e-01
-3.52799624e-01 4.68985289e-01 -7.05528975e-01 -6.16911590e-01
-1.58898294e+00 2.39566863e-01 1.93157122e-01 -3.27978343e-01
-6.19559169e-01 4.35164571e-01 -2.94815242e-01 -3.72173339e-02
-1.01832256e-01 2.46608481e-01 -4.13349628e-01 8.38482007e-02
6.64199769e-01 7.65832901e-01 1.70200691e-02 -1.40570962e+00
-8.50744069e-01 1.24940777e+00 -5.86037971e-02 -8.78702626e-02
1.12290943e+00 1.26975730e-01 9.60752144e-02 2.23146155e-01
1.28576469e+00 -2.77272701e-01 -8.11175108e-01 -1.82119720e-02
-9.96585935e-02 -7.86240578e-01 -2.27127194e-01 -3.87818605e-01
-1.08102739e+00 1.03843045e+00 1.06183863e+00 -2.54328102e-01
9.76923704e-01 -5.37576258e-01 1.10752165e+00 2.32824877e-01
7.97473192e-01 -1.04521847e+00 4.48106587e-01 2.92468399e-01
8.93207192e-01 -1.06107724e+00 -7.31545463e-02 -5.16450265e-03
-3.35269958e-01 1.11680806e+00 6.99689448e-01 -2.38530591e-01
5.55226505e-01 -3.24322671e-01 -3.31027865e-01 -7.12033287e-02
-2.89359629e-01 -1.97432891e-01 4.54882413e-01 8.86457562e-01
1.86400369e-01 1.14175297e-01 -4.49797183e-01 8.35025191e-01
1.93292852e-02 9.67577752e-03 4.97974642e-02 1.20241773e+00
-4.92123306e-01 -8.71551812e-01 -6.70212269e-01 3.71586233e-01
-4.34558302e-01 3.26261938e-01 -2.54571170e-01 6.89915895e-01
2.28390858e-01 9.44368005e-01 -1.31886512e-01 -8.43418956e-01
2.13151187e-01 1.60055533e-01 6.48838520e-01 -2.00266853e-01
5.33240754e-03 -6.30259067e-02 4.74981871e-03 -5.18311620e-01
-4.64526087e-01 -1.00732899e+00 -8.52056503e-01 -5.19761682e-01
1.09670274e-01 1.24053374e-01 1.84479371e-01 8.91385138e-01
3.31326038e-01 1.57871991e-01 3.39261979e-01 -1.03296399e+00
-1.70340091e-01 -8.64446461e-01 -6.89660251e-01 6.79547369e-01
1.99860230e-01 -1.25178862e+00 -1.60601422e-01 2.13364810e-01] | [14.291805267333984, 1.4306927919387817] |
32ebd2ed-4571-416f-ab66-9a0a53a5c7d3 | leveraging-inpainting-for-single-image-shadow | 2302.05361 | null | https://arxiv.org/abs/2302.05361v2 | https://arxiv.org/pdf/2302.05361v2.pdf | Leveraging Inpainting for Single-Image Shadow Removal | Fully-supervised shadow removal methods achieve the best restoration qualities on public datasets but still generate some shadow remnants. One of the reasons is the lack of large-scale shadow & shadow-free image pairs. Unsupervised methods can alleviate the issue but their restoration qualities are much lower than those of fully-supervised methods. In this work, we find that pretraining shadow removal networks on the image inpainting dataset can reduce the shadow remnants significantly: a naive encoder-decoder network gets competitive restoration quality w.r.t. the state-of-the-art methods via only 10% shadow & shadow-free image pairs. After analyzing networks with/without inpainting pre-training via the information stored in the weight (IIW), we find that inpainting pretraining improves restoration quality in non-shadow regions and enhances the generalization ability of networks significantly. Additionally, shadow removal fine-tuning enables networks to fill in the details of shadow regions. Inspired by these observations we formulate shadow removal as an adaptive fusion task that takes advantage of both shadow removal and image inpainting. Specifically, we develop an adaptive fusion network consisting of two encoders, an adaptive fusion block, and a decoder. The two encoders are responsible for extracting the feature from the shadow image and the shadow-masked image respectively. The adaptive fusion block is responsible for combining these features in an adaptive manner. Finally, the decoder converts the adaptive fused features to the desired shadow-free result. The extensive experiments show that our method empowered with inpainting outperforms all state-of-the-art methods. | ['Song Wang', 'Ivor Tsang', 'Wei Feng', 'Di Lin', 'Rabab Abdelfattah', 'Qing Guo', 'Xiaoguang Li'] | 2023-02-10 | null | null | null | null | ['shadow-removal', 'image-inpainting', 'image-shadow-removal'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 7.13236749e-01 6.25912324e-02 3.01666260e-02 -5.15172780e-01
-7.21176624e-01 5.56991063e-02 2.26193041e-01 -6.70379639e-01
-2.15487510e-01 8.94259155e-01 4.10223544e-01 -1.71391830e-01
1.62646249e-01 -7.38390625e-01 -1.02261388e+00 -1.12493038e+00
2.84249246e-01 4.63743024e-02 2.89565802e-01 -3.55440378e-01
-4.01476733e-02 4.45825696e-01 -1.60497177e+00 3.94908220e-01
1.37943506e+00 1.21520936e+00 7.74151802e-01 5.40392220e-01
-4.23972420e-02 8.68181705e-01 -7.79220223e-01 2.61574052e-02
5.13923705e-01 -4.11622703e-01 -3.80053441e-03 2.73405850e-01
5.89284360e-01 -7.48547018e-01 -7.42149949e-01 7.37395406e-01
5.32395959e-01 1.07473418e-01 4.97789860e-01 -1.14984024e+00
-1.09391022e+00 2.22080380e-01 -5.78209460e-01 -3.34853530e-02
-6.06338680e-02 1.04503572e-01 6.53180778e-01 -1.10476387e+00
5.09796441e-01 1.17662954e+00 8.18681717e-01 3.23008060e-01
-1.07629848e+00 -6.48745537e-01 -1.15988314e-01 2.04349026e-01
-1.23305106e+00 -5.26859760e-01 9.96077061e-01 1.54557541e-01
6.24737740e-01 1.79876238e-01 4.27274048e-01 9.86730993e-01
4.44591820e-01 8.94954026e-01 1.40653205e+00 -4.41101223e-01
4.01816405e-02 1.01176113e-01 -2.86090881e-01 7.37798750e-01
3.93941589e-02 4.62240338e-01 -8.34943473e-01 1.52878582e-01
7.23508537e-01 3.85313153e-01 -7.12407827e-01 -1.45931959e-01
-7.51694560e-01 6.05165362e-01 8.15322578e-01 -6.19369326e-03
-4.42547947e-01 2.41917253e-01 -6.71059489e-02 3.82569790e-01
6.55704200e-01 5.97868264e-02 -5.07946849e-01 6.56551361e-01
-1.24058568e+00 -1.02313511e-01 4.63610828e-01 6.62812412e-01
1.20897639e+00 3.22952062e-01 -2.05299228e-01 9.83733356e-01
1.72217876e-01 9.40489471e-01 3.18043351e-01 -9.92827356e-01
3.51366043e-01 3.84853959e-01 8.18072259e-02 -8.64883721e-01
-8.47858936e-02 -6.23173535e-01 -1.02357292e+00 5.24556398e-01
-4.55743782e-02 -1.06382914e-01 -1.25571001e+00 1.78369713e+00
1.09096222e-01 2.03152999e-01 2.98591763e-01 8.70513737e-01
8.92672539e-01 7.90833175e-01 -4.77599144e-01 -3.32143933e-01
1.19092000e+00 -1.43647695e+00 -9.48291838e-01 -6.45732820e-01
-4.34357347e-03 -9.87186432e-01 1.15274072e+00 1.96451157e-01
-8.52374434e-01 -7.38376796e-01 -1.36924636e+00 -1.37250990e-01
-3.37652832e-01 4.39515114e-01 5.25665462e-01 4.82703298e-01
-1.05721700e+00 6.53701484e-01 -3.75082493e-01 -1.77488267e-01
6.22649670e-01 2.40979850e-01 -3.36822689e-01 -3.09506029e-01
-1.08037734e+00 8.65226805e-01 2.18282547e-02 2.20378801e-01
-1.11108160e+00 -7.24285603e-01 -8.13765466e-01 -9.45647433e-03
6.09744787e-01 -5.36682904e-01 9.48880970e-01 -1.41805696e+00
-1.40359247e+00 3.72237772e-01 -2.65932918e-01 -5.36265850e-01
3.95286411e-01 -4.20955241e-01 -5.10401309e-01 2.27701604e-01
3.14980775e-01 6.57784998e-01 1.39449477e+00 -1.95037806e+00
-5.46673536e-01 -1.54946223e-02 -1.88033000e-01 4.97132391e-01
-3.56012106e-01 -4.71576273e-01 -5.93094826e-01 -9.57265437e-01
5.23468442e-02 -7.45773435e-01 8.30626413e-02 2.78367460e-01
-4.68267888e-01 4.67839956e-01 1.40610480e+00 -1.01833689e+00
7.67759264e-01 -2.27261591e+00 -1.71347156e-01 2.73971856e-02
1.56079575e-01 2.56445974e-01 -3.42929214e-01 4.37390447e-01
-2.71308441e-02 -3.65371257e-01 -7.28002369e-01 -7.14740276e-01
-2.52002448e-01 6.28806949e-01 -7.21108615e-01 4.32937562e-01
3.12667410e-03 8.15785944e-01 -5.25761306e-01 -4.78979260e-01
3.81482244e-01 7.72535264e-01 -1.44220591e-01 4.87457454e-01
-1.93483531e-01 3.31565738e-01 -7.72664174e-02 8.86241198e-01
1.09121740e+00 -1.76603690e-01 7.99832568e-02 -3.95120054e-01
6.86638281e-02 -4.53549810e-03 -8.91085982e-01 1.63501906e+00
-6.54259801e-01 9.79137540e-01 2.59659797e-01 -6.28225207e-01
8.86389315e-01 7.49751702e-02 2.78192371e-01 -1.03801656e+00
7.95563310e-03 2.73629755e-01 -2.93042690e-01 -3.24147910e-01
6.10234380e-01 -2.60094583e-01 3.36752951e-01 5.28697133e-01
1.87924635e-02 -8.00182000e-02 -1.43144786e-01 2.63774097e-01
1.12472892e+00 1.71379879e-01 -1.15336142e-01 -1.34630324e-02
4.01053965e-01 -4.43950206e-01 6.29712343e-01 8.39158833e-01
1.53771684e-01 9.86474574e-01 8.06254297e-02 -2.84588605e-01
-8.31624687e-01 -1.24290991e+00 1.19235590e-01 1.18272972e+00
4.00538296e-01 -9.67471302e-02 -7.47196734e-01 -6.92052901e-01
1.42000064e-01 5.89788437e-01 -8.38891566e-01 -1.74058348e-01
-4.98098046e-01 -7.31220901e-01 4.73965675e-01 4.98663664e-01
1.08122873e+00 -1.25432312e+00 -5.30437946e-01 -9.78893489e-02
-4.15925294e-01 -1.14095998e+00 -4.73766148e-01 4.95909989e-01
-6.80460691e-01 -9.61243987e-01 -7.30299890e-01 -7.29728103e-01
8.14225376e-01 8.65939200e-01 9.25673604e-01 5.03741682e-01
-2.70586520e-01 -3.67429852e-02 -3.68094265e-01 -2.98597574e-01
-3.71466488e-01 -2.61874884e-01 -1.01059258e-01 2.36774281e-01
-4.50837314e-01 -9.69847858e-01 -8.86744797e-01 2.55858809e-01
-1.24941397e+00 4.30672884e-01 1.20516241e+00 9.90519464e-01
4.51827615e-01 3.13520730e-01 3.13894808e-01 -9.76844549e-01
4.08903569e-01 -1.29206046e-01 -2.33653709e-01 2.46272683e-01
-9.98658061e-01 1.02798216e-01 7.55482674e-01 -2.05946624e-01
-1.60572088e+00 1.05798401e-01 6.94197938e-02 -3.68740469e-01
-2.43064482e-03 -9.60806757e-03 -3.55643153e-01 -4.00358081e-01
7.33050287e-01 5.57787299e-01 -2.68781111e-02 -5.27594030e-01
4.14387584e-01 5.39330482e-01 8.14871848e-01 -2.45707974e-01
1.27171671e+00 7.85552621e-01 -1.86358571e-01 -6.90098226e-01
-1.02909720e+00 -2.38482073e-01 -3.52439255e-01 -2.51206815e-01
4.90926325e-01 -1.00525963e+00 -1.84073687e-01 7.11534619e-01
-9.52231169e-01 -7.64131129e-01 -4.22995150e-01 -1.01723284e-01
-3.54579419e-01 5.25737286e-01 -4.24300432e-01 -6.42759025e-01
-4.07650113e-01 -1.02490246e+00 1.16867161e+00 4.15134311e-01
5.09377599e-01 -6.38893664e-01 -2.35285029e-01 5.05597949e-01
6.27659976e-01 1.26433164e-01 7.83854783e-01 1.05512820e-01
-8.02577138e-01 1.12136304e-01 -6.77854478e-01 8.84968460e-01
4.23829347e-01 -4.45264578e-01 -1.31290579e+00 -1.40575081e-01
1.18502587e-01 -2.31729358e-01 1.53667188e+00 4.91756141e-01
1.08730090e+00 -3.53213578e-01 -1.94111153e-01 8.19805264e-01
1.59464073e+00 2.10813852e-03 1.09101272e+00 2.89510727e-01
7.47172296e-01 3.43843818e-01 6.10904753e-01 2.79925227e-01
2.48188466e-01 4.62801784e-01 5.65418959e-01 -5.99419892e-01
-8.23223412e-01 -1.54977024e-01 6.50093257e-01 3.70357960e-01
-5.27769513e-02 -5.07058024e-01 -3.68936002e-01 4.42536384e-01
-1.82861710e+00 -8.21554244e-01 9.63216722e-02 1.98301029e+00
9.82917905e-01 -3.54743674e-02 -5.65138817e-01 1.34414211e-01
4.74671841e-01 7.15827525e-01 -5.19412816e-01 7.71382526e-02
-6.51391983e-01 3.31284463e-01 8.48913133e-01 8.07296753e-01
-8.33077729e-01 1.16952038e+00 5.62464952e+00 9.53656852e-01
-9.40937757e-01 3.09006155e-01 5.27844489e-01 1.40923038e-01
-3.67415547e-01 2.81823158e-01 -3.76279831e-01 4.17376786e-01
4.99479681e-01 6.78615689e-01 7.73112118e-01 5.86552978e-01
1.42457992e-01 -6.85038388e-01 -6.34657502e-01 7.85791516e-01
5.23465276e-01 -1.49803615e+00 -2.84805913e-02 -1.45524181e-03
1.00692737e+00 6.35326505e-02 6.36042506e-02 9.71656665e-02
6.47048950e-02 -9.72519398e-01 7.56295145e-01 8.14967155e-01
9.16354418e-01 -5.46502888e-01 8.64321291e-01 2.51498252e-01
-1.09744191e+00 -2.43270621e-01 -3.03541869e-01 -2.16102563e-02
2.31256977e-01 1.03283501e+00 -6.76885486e-01 7.19235659e-01
8.36758733e-01 5.49498022e-01 -6.13380373e-01 6.60303950e-01
-7.51377344e-01 4.26528066e-01 -1.47920132e-01 6.35542333e-01
-1.34077743e-01 -2.00852513e-01 3.11443835e-01 1.12179542e+00
1.78678185e-01 6.15981501e-03 -2.13778410e-02 6.44957364e-01
-2.77572006e-01 -3.03108454e-01 -4.30390745e-01 2.56742239e-01
3.65047514e-01 1.36118042e+00 -7.17654705e-01 -5.03160894e-01
-3.12186897e-01 1.46198082e+00 2.58568451e-02 6.61256731e-01
-9.11925912e-01 -3.18667918e-01 4.34871614e-01 8.84563401e-02
3.54169905e-01 4.51471796e-03 -4.09173340e-01 -8.62385452e-01
1.14421345e-01 -8.76527011e-01 -2.32054666e-01 -1.28596449e+00
-9.46413040e-01 6.79733872e-01 -3.92080992e-01 -1.03858805e+00
2.24268347e-01 -4.17526633e-01 -6.50632381e-01 7.65689909e-01
-2.04495716e+00 -1.37758446e+00 -9.57386255e-01 7.15870440e-01
6.42308652e-01 -2.20396414e-01 6.13383532e-01 3.77304018e-01
-4.79787886e-01 4.51228827e-01 3.00279915e-01 -7.72090927e-02
1.13326824e+00 -1.07604206e+00 1.45954743e-01 1.12573719e+00
-1.22771479e-01 1.83195621e-01 8.21286023e-01 -8.68188202e-01
-1.35758853e+00 -1.13358784e+00 6.60157979e-01 7.95476586e-02
2.24627972e-01 -3.63596648e-01 -8.33279431e-01 4.09051120e-01
3.72491449e-01 1.66685477e-01 2.62787730e-01 -4.54521894e-01
-5.49083948e-01 -6.13128304e-01 -1.12148547e+00 5.12355924e-01
1.02395523e+00 -7.06407130e-01 -3.67342889e-01 5.65639019e-01
8.07332754e-01 -4.33303177e-01 -3.39540511e-01 5.41148663e-01
5.00405610e-01 -1.51318562e+00 1.16929293e+00 3.74107659e-01
6.64413750e-01 -4.60445285e-01 -5.33245325e-01 -1.19195437e+00
-3.24754082e-02 -5.57449520e-01 -2.06529602e-01 1.18349922e+00
1.40826106e-01 -7.73719192e-01 7.91864395e-01 -6.42544869e-03
-5.59075892e-01 -8.15484107e-01 -6.65286601e-01 -6.64747417e-01
-5.50571263e-01 -2.92326659e-01 5.40010512e-01 4.12432760e-01
-9.23791349e-01 1.83148354e-01 -8.20415437e-01 4.17994678e-01
7.78790116e-01 4.69483763e-01 8.46364439e-01 -9.47036266e-01
-5.24796903e-01 7.11275414e-02 3.19672942e-01 -1.05773306e+00
2.09108815e-01 -4.85761642e-01 5.94600379e-01 -1.68229425e+00
4.07023877e-01 -4.35684592e-01 -1.73096895e-01 8.92384231e-01
-1.33608758e-01 7.21328616e-01 1.81955919e-01 4.66292858e-01
-2.46834725e-01 9.24662352e-01 1.39616275e+00 -1.03936762e-01
-1.68522760e-01 -1.45248890e-01 -7.85986900e-01 6.45914495e-01
7.99456298e-01 -5.89478552e-01 -4.67027277e-01 -4.61309969e-01
-1.32584095e-01 8.83076489e-02 7.28028238e-01 -1.16524363e+00
9.58070531e-02 -1.70033336e-01 7.18880296e-01 -6.31668925e-01
8.44402432e-01 -7.48697996e-01 3.31044532e-02 3.31869513e-01
1.07182831e-01 -3.76716554e-01 -4.25013900e-02 7.42201626e-01
-1.70745581e-01 1.05636150e-01 8.28213155e-01 -3.00922748e-02
-5.84616065e-01 6.65484965e-02 -3.06774020e-01 -3.64526063e-01
4.92423683e-01 -3.41978669e-01 -5.95260739e-01 -7.10183024e-01
-2.89398134e-01 -1.63923398e-01 6.27444804e-01 2.19948217e-01
8.99403572e-01 -1.10648763e+00 -5.29866755e-01 5.40526092e-01
-2.37362504e-01 1.60689913e-02 4.71514106e-01 8.45100522e-01
-4.63415593e-01 -8.29009786e-02 -3.24289054e-01 -2.25417942e-01
-1.23561072e+00 1.91434801e-01 2.64096439e-01 -1.74251705e-01
-7.60743976e-01 8.53802800e-01 6.68267369e-01 -3.44624430e-01
3.37041587e-01 -1.64859861e-01 2.83877909e-01 -1.89035207e-01
3.92202258e-01 2.07910955e-01 1.36955023e-01 -6.43143058e-01
-1.99814051e-01 4.19939846e-01 7.45223686e-02 -1.11023627e-01
1.44182193e+00 -1.28805906e-01 -2.55539685e-01 3.52354124e-02
1.03989780e+00 2.12942943e-01 -1.84360254e+00 -3.29998761e-01
-6.22274458e-01 -5.25365233e-01 2.38611490e-01 -1.11434901e+00
-1.58749497e+00 6.72546268e-01 9.01608646e-01 -1.12815410e-01
1.75698662e+00 -2.54100740e-01 1.16508663e+00 2.99999416e-01
2.54613459e-02 -1.21056366e+00 4.06530112e-01 4.07149971e-01
1.05235171e+00 -1.27531743e+00 3.78545821e-01 -5.98864794e-01
-5.40641069e-01 9.11565244e-01 4.48428929e-01 -4.30982560e-01
5.45587063e-01 5.19660354e-01 2.30623305e-01 -4.05267887e-02
-4.72478062e-01 -2.36132801e-01 2.72470504e-01 6.96802676e-01
-1.36728138e-01 5.63176386e-02 -3.34424190e-02 4.35787231e-01
-1.23391502e-01 -1.65350810e-01 4.32113469e-01 8.87024522e-01
-6.99557900e-01 -1.09145403e+00 -6.77171767e-01 4.25186872e-01
1.52174942e-02 -4.31616634e-01 -5.55975199e-01 6.91249847e-01
5.00526786e-01 1.05615199e+00 -3.40276480e-01 -5.97463012e-01
1.36596458e-02 -1.18635543e-01 4.24945563e-01 -3.54059488e-01
-4.40428942e-01 1.17130682e-01 -2.38083862e-02 -6.50094450e-01
-2.54947633e-01 -2.67456055e-01 -1.27929652e+00 -2.67749578e-01
-4.54873174e-01 -7.89305791e-02 5.73579848e-01 1.13391173e+00
3.63713652e-01 8.08314741e-01 7.36298740e-01 -1.21657813e+00
-1.36526078e-01 -9.71293271e-01 -6.97665870e-01 1.34609297e-01
6.93175256e-01 -7.27598608e-01 -4.89391536e-01 9.15616304e-02] | [10.846075057983398, -4.108641147613525] |
9cc21522-4fa2-44a9-9b2e-a9cf08776002 | neuromorphic-computing-with-deeply-scaled | 2103.13302 | null | https://arxiv.org/abs/2103.13302v2 | https://arxiv.org/pdf/2103.13302v2.pdf | Neuromorphic Computing with Ferroelectric FinFETs in the Presence of Temperature, Process Variation, Device Aging and Flicker Noise | This paper reports a comprehensive study on the impacts of temperature-change, process variation, flicker noise and device aging on the inference accuracy of pre-trained all-ferroelectric (FE) FinFET deep neural networks. Multiple-level-cell (MLC) operation with a novel adaptive-program-and-read algorithm with 100ns write pulse has been experimentally demonstrated in 5 nm thick hafnium zirconium oxide (HZO)-based FE-FinFET. With pre-trained neural network (NN) with 97.5% inference accuracy on MNIST dataset as baseline, device to device variation is shown to have negligible impact. Flicker noise characterization at various bias conditions depicts that drain current fluctuation is less than 0.7% with virtually no inference accuracy degradation. The conductance drift of a programmed cell, as an aftermath of temperature change, was captured by a compact model over a wide range of gate biases. Despite significant inference accuracy degradation at 233K for a NN trained at 300K, gate bias optimization for recovering the accuracy is demonstrated. Endurance above 10$^8$ cycles and extrapolated retention above 10 years are shown, which paves the way for edge device artificial intelligence with FE-FinFETs. | ['Darsen Lu', 'Yao-Jen Lee', 'Chung-Jun Su', 'Md. Aftab Baig', 'Wei-Xuan Bu', 'Bo-Han Qiu', 'Sourav De'] | 2021-03-05 | null | null | null | null | ['neural-network-simulation'] | ['computer-code'] | [ 3.27967912e-01 -3.56991142e-01 -2.55080909e-01 -1.94581375e-01
-4.39168932e-03 -2.46594712e-01 9.34972912e-02 2.68481106e-01
-7.59408057e-01 1.20183241e+00 -4.70397443e-01 -3.90707046e-01
-5.97165786e-02 -7.98816144e-01 -7.26701558e-01 -1.04314125e+00
1.65211067e-01 2.06684172e-01 2.38881186e-01 -2.44132951e-01
5.01178980e-01 4.03494686e-01 -1.34389496e+00 3.58964413e-01
7.04721212e-01 1.33889067e+00 -4.29743260e-01 6.81167066e-01
4.05481756e-01 5.91233850e-01 -1.03267598e+00 -3.33163023e-01
-2.69129336e-01 -8.69344920e-03 -3.54489714e-01 -1.14532089e+00
3.37595612e-01 -5.26418269e-01 -7.18295038e-01 6.72972500e-01
7.52821147e-01 -3.29923749e-01 9.28760588e-01 -1.03179324e+00
-6.37567043e-01 8.50955784e-01 -1.32708877e-01 5.47356546e-01
-3.79642367e-01 4.64603364e-01 1.35715842e-01 -6.93501830e-01
2.94136912e-01 3.91818792e-01 1.03129101e+00 7.56318033e-01
-1.20740902e+00 -9.36217785e-01 -8.55241239e-01 -1.18497245e-01
-1.53451848e+00 -6.74939752e-01 2.22374111e-01 -2.50001065e-02
2.15263915e+00 -1.74882058e-02 7.07114339e-01 1.24357879e+00
1.67937517e+00 -9.19875205e-02 1.15854740e+00 -3.43115777e-01
3.28444183e-01 2.94428393e-02 8.08419287e-01 3.39207858e-01
1.11857009e+00 3.00913215e-01 -9.54569697e-01 -1.80140212e-02
3.38035226e-01 1.82664860e-02 -9.17870849e-02 7.12050676e-01
-3.84169549e-01 -1.67105421e-01 4.88813341e-01 3.68671983e-01
-2.22328812e-01 1.14870405e+00 8.92328560e-01 3.62094641e-01
-3.49437892e-01 3.51697505e-01 -2.25481942e-01 -5.47247350e-01
-1.05818653e+00 1.50144890e-01 8.76776934e-01 7.70381212e-01
1.30543932e-01 8.57099235e-01 -2.74756700e-01 1.09525867e-01
2.18611941e-01 1.28359318e+00 2.59134442e-01 -4.41961586e-01
-5.18434448e-03 6.81416810e-01 1.99450150e-01 -5.18641353e-01
-3.43757749e-01 -2.84492642e-01 -1.21361291e+00 2.29488671e-01
3.46272141e-01 -2.07432851e-01 -1.41617501e+00 1.17711115e+00
-5.06065071e-01 -3.57436419e-01 6.41911104e-02 1.46608502e-01
6.99404836e-01 5.34343004e-01 1.84862390e-02 -1.90445244e-01
8.57980788e-01 -1.81893975e-01 -1.09926128e+00 5.51586365e-03
2.11806297e-01 2.98096716e-01 6.38302624e-01 2.87588805e-01
-1.07399631e+00 -7.26093292e-01 -1.81769955e+00 -2.67106324e-01
-4.39484090e-01 -1.03079351e-02 2.12635651e-01 1.64271152e+00
-1.11197269e+00 1.00401640e+00 -1.14069438e+00 -8.81663635e-02
7.47870803e-01 1.18560350e+00 2.11116910e-01 3.54073010e-02
-1.18372142e+00 5.00582874e-01 3.17951292e-01 3.86622965e-01
-9.95166481e-01 -9.02843356e-01 -2.01022148e-01 1.14340313e-01
-1.57698721e-01 -4.09398615e-01 6.88852429e-01 -6.01867855e-01
-1.54054487e+00 4.49811071e-01 -4.89495337e-01 -9.30336773e-01
2.32566129e-02 -4.13512766e-01 -9.95948315e-01 -3.07846427e-01
-3.68361503e-01 2.43625104e-01 6.18478775e-01 -7.43994951e-01
2.06732750e-01 -4.66160476e-01 -8.77296269e-01 -6.37357473e-01
-6.86003625e-01 -3.20995569e-01 9.46786925e-02 1.03490137e-01
-1.19940512e-01 -8.61261964e-01 2.01645792e-01 -5.32664537e-01
-2.17864260e-01 2.14080308e-02 8.72006834e-01 -2.62375981e-01
1.78955877e+00 -1.94779372e+00 -7.27539778e-01 6.28286600e-01
1.06694207e-01 4.29173291e-01 5.94653606e-01 5.34703314e-01
5.33403099e-01 3.18341583e-01 -3.13236117e-01 1.99728221e-01
-3.36748004e-01 2.65111518e-03 -2.04486102e-01 4.68064606e-01
1.69283524e-01 1.33235335e+00 -4.50638562e-01 3.16331089e-01
-1.55380473e-01 4.87710357e-01 -1.27473429e-01 -5.32245994e-01
2.97386438e-01 -1.49060665e-02 -1.76796019e-01 1.10082936e+00
6.51235878e-01 8.76865909e-03 3.14376801e-01 -5.28772950e-01
-2.41129294e-01 -1.70706045e-02 -4.97093290e-01 1.30441570e+00
4.56149615e-02 9.06867862e-01 -2.39004597e-01 -4.22345400e-01
1.49234092e+00 -3.91915776e-02 -2.15053707e-02 -1.51805234e+00
5.12075663e-01 6.97092056e-01 5.19169211e-01 -2.43401036e-01
8.43657851e-01 -3.38374227e-01 -2.62105763e-01 2.16812175e-03
-8.19499716e-02 2.73474097e-01 -2.03999892e-01 -1.47096086e-02
1.35731053e+00 -9.81560796e-02 -3.92121106e-01 -1.23604822e+00
6.54323101e-02 -9.04701836e-03 2.35823840e-01 1.23538554e+00
-4.20895427e-01 2.13140436e-03 4.02882785e-01 -2.95175791e-01
-1.14461088e+00 -1.17908823e+00 -6.77717268e-01 6.16138339e-01
2.13056594e-01 -4.00436312e-01 -7.85312772e-01 1.24587215e-01
5.98484837e-03 8.73072565e-01 -5.91615021e-01 -6.30961716e-01
-5.84266901e-01 -1.20435917e+00 1.64754021e+00 1.31617248e+00
9.37850356e-01 -6.71362758e-01 -1.31105316e+00 4.32928830e-01
7.18442619e-01 -9.39769149e-01 2.99708307e-01 1.16254902e+00
-1.30692756e+00 -3.11087370e-01 9.63722393e-02 -4.69675928e-01
5.16943693e-01 -9.59774375e-01 1.33111501e+00 2.37550303e-01
-4.91571337e-01 4.88533527e-02 2.46739432e-01 -6.36037171e-01
-3.63161206e-01 1.58144981e-01 4.18962836e-01 -5.78829229e-01
6.41035795e-01 -5.67723453e-01 -8.37530553e-01 3.08241118e-02
-4.13062811e-01 -6.12055719e-01 3.91397715e-01 1.05637038e+00
7.05169678e-01 3.69595796e-01 8.74287367e-01 -9.73617673e-01
3.34734827e-01 -9.84355062e-02 -1.52701691e-01 2.29334477e-02
-1.21651804e+00 2.34849274e-01 7.56744802e-01 -2.88837552e-01
-1.15396023e+00 -2.40344599e-01 -1.77099317e-01 -1.35385394e-01
7.94889852e-02 3.33467603e-01 -3.71187210e-01 7.37041533e-02
1.33058512e+00 9.54854414e-02 -1.09212331e-01 2.29470506e-01
-5.03134787e-01 6.95885241e-01 7.60070682e-01 -4.93721128e-01
2.72871315e-01 4.96727794e-01 4.95941728e-01 -9.14905071e-01
3.32538962e-01 4.00476545e-01 -5.86048484e-01 -4.28888142e-01
5.09817123e-01 -9.14490461e-01 -1.16757119e+00 1.08584118e+00
-4.58954841e-01 -7.91050255e-01 1.43826470e-01 -2.03067392e-01
4.85920519e-01 -6.64327085e-01 -9.68066156e-01 -1.24259746e+00
-1.16490519e+00 -8.01878035e-01 5.49554229e-01 6.25057101e-01
-2.47586429e-01 -1.20221436e+00 -5.96348524e-01 -1.46744000e-02
1.08639157e+00 3.27004403e-01 1.10860574e+00 -4.80950058e-01
-3.42876315e-01 -3.74036968e-01 -6.07781997e-03 3.04585546e-01
-1.47371218e-02 2.37289928e-02 -1.38613224e+00 -5.75283051e-01
1.67263776e-01 -2.57856995e-01 1.19046187e+00 6.16673589e-01
1.07150364e+00 5.34923136e-01 -8.90283883e-01 4.21112061e-01
1.93040121e+00 7.37384975e-01 1.16869104e+00 -3.49073440e-01
9.06257510e-01 -5.19082606e-01 -1.19993098e-01 1.81265920e-01
-7.02068880e-02 -9.64726806e-02 3.85538459e-01 4.11032468e-01
3.04130971e-01 -1.95465177e-01 5.78627884e-01 3.73541087e-01
-1.08092077e-01 -8.61224532e-01 -1.54618013e+00 2.83990324e-01
-8.58596742e-01 -6.18250787e-01 -9.37384963e-01 2.28197479e+00
6.31738365e-01 8.26557457e-01 -5.12321174e-01 3.25302631e-01
5.34889340e-01 -2.17261136e-01 -1.39273179e+00 -8.62834871e-01
-2.12236151e-01 8.55983078e-01 1.38651478e+00 4.71250191e-02
-4.41915870e-01 5.32940030e-01 5.70487547e+00 5.43514132e-01
-1.82979703e+00 -1.64159626e-01 1.06337321e+00 -2.34139994e-01
-3.86427104e-01 -4.26946878e-01 -1.42811501e+00 6.72456741e-01
2.04612565e+00 2.70593971e-01 -1.39307201e-01 1.53087437e-01
-2.37734020e-01 -6.47547483e-01 -1.28940368e+00 1.11987102e+00
-1.34280547e-01 -1.47658932e+00 -2.66114831e-01 3.08513880e-01
8.56336415e-01 4.55124453e-02 3.52084070e-01 5.31269908e-01
-3.75994414e-01 -1.81387186e+00 4.89812672e-01 8.69862616e-01
1.79014099e+00 -1.04437459e+00 1.01403570e+00 -7.32297301e-02
-7.24845052e-01 -4.79577571e-01 -2.86426216e-01 -3.43499631e-01
-2.00225577e-01 1.00773370e+00 -4.33274478e-01 2.25819170e-01
1.04062545e+00 3.83967847e-01 -4.59409535e-01 2.58867115e-01
7.10125327e-01 1.19027388e+00 -7.51634479e-01 -6.24562800e-01
-1.44428343e-01 2.43731067e-01 -5.79682970e-03 1.15841258e+00
3.11972171e-01 -1.80532075e-02 -7.56524146e-01 1.03149867e+00
-4.79325026e-01 -8.21196198e-01 -4.88665462e-01 -2.50903755e-01
9.55709994e-01 8.96758914e-01 -1.08483791e+00 -1.31874889e-01
3.05131912e-01 9.70521152e-01 -3.95587713e-01 3.36608857e-01
-8.14575613e-01 -6.93109334e-01 1.50720164e-01 7.49632776e-01
2.48125255e-01 -1.73867390e-01 -1.21976769e+00 -1.21170536e-01
8.29938948e-02 -7.81744197e-02 -3.75543833e-01 -1.54983178e-01
-4.93324697e-01 1.43795788e-01 -2.44007990e-01 -1.62340254e-01
4.07137424e-01 -8.78796339e-01 -1.03679764e+00 8.70738268e-01
-1.17038751e+00 -4.85858768e-01 -5.02383053e-01 -2.73350161e-02
-3.38786155e-01 7.35541508e-02 7.22820044e-01 1.57203302e-01
-6.67323887e-01 1.33541894e+00 7.74154544e-01 1.40915006e-01
4.93631482e-01 -9.23281372e-01 7.60503829e-01 5.32228112e-01
-1.05056989e+00 8.40741694e-01 5.10487795e-01 -1.14098120e+00
-2.27313685e+00 -1.34265935e+00 5.19842565e-01 -5.66459179e-01
1.26438543e-01 -7.17375994e-01 -1.30069089e+00 2.73943067e-01
3.19675803e-01 1.63302600e-01 5.32082975e-01 -3.84333581e-01
1.15392856e-01 -7.73953497e-01 -1.36643445e+00 3.47597331e-01
1.13834584e+00 -4.87203240e-01 3.55700433e-01 -6.09233558e-01
1.78542167e-01 -5.07957339e-01 -1.24694991e+00 1.02357328e+00
1.12894201e+00 -9.17163134e-01 3.38938504e-01 1.49922162e-01
1.27560765e-01 -1.59531869e-02 -2.69730389e-01 -4.60084498e-01
1.80354759e-01 -4.40382212e-01 -3.59522790e-01 1.32509232e+00
5.56568325e-01 -7.00248599e-01 1.03770077e+00 7.30999470e-01
-2.48136222e-01 -9.20807898e-01 -9.68686998e-01 -8.56779993e-01
5.77664077e-01 -2.89959818e-01 1.98898599e-01 -1.01873599e-01
-4.49678421e-01 3.35511208e-01 -1.29819751e-01 -1.20378390e-01
3.17805618e-01 -9.22427356e-01 7.83284828e-02 -1.20633733e+00
7.13876784e-02 1.05623379e-01 -3.83400381e-01 -5.89098454e-01
2.86765024e-02 -6.22419238e-01 9.41359922e-02 -1.00856817e+00
-9.26822200e-02 -3.71399671e-01 -6.45905077e-01 2.53160536e-01
-9.35781840e-03 1.78081766e-01 -4.99334216e-01 -2.33748779e-01
-1.75670773e-01 2.44587839e-01 5.62457502e-01 3.93707566e-02
-3.25528711e-01 -5.30687988e-01 -2.31546372e-01 1.77107807e-02
8.29494119e-01 -5.41664958e-01 -3.92935157e-01 -4.29392546e-01
6.61180079e-01 1.58160418e-01 4.39483613e-01 -1.75204182e+00
4.86729622e-01 6.02874160e-01 1.18665671e+00 -6.47667766e-01
2.09606051e-01 -6.20744765e-01 5.38509965e-01 8.10298145e-01
1.65332898e-01 2.63677537e-01 9.45620179e-01 5.68349719e-01
2.85331100e-01 -2.02263907e-01 7.00329959e-01 2.28707999e-01
-5.76409876e-01 6.89197797e-03 -5.79454958e-01 1.31764904e-01
4.86564696e-01 -1.01868522e+00 -5.10755002e-01 4.93142098e-01
-2.27224901e-01 -1.59695260e-02 6.07347608e-01 -2.02574909e-01
8.33549619e-01 -9.63695526e-01 -2.19659880e-01 7.28420019e-01
-9.85398144e-02 -1.23615913e-01 8.08432341e-01 9.07019734e-01
-5.83226681e-01 3.89519393e-01 -4.13610786e-01 -8.57345879e-01
-9.69141185e-01 -2.29701996e-01 8.45583498e-01 -1.61168903e-01
-3.24522138e-01 9.30672944e-01 -8.54662836e-01 3.98356050e-01
-1.70942739e-01 -4.74511772e-01 3.40073138e-01 -2.17285499e-01
2.73209244e-01 7.36564279e-01 5.02211511e-01 1.02918766e-01
-5.35960436e-01 3.91217262e-01 -2.51753718e-01 -9.23178270e-02
1.06898844e+00 1.32486954e-01 -7.41551220e-02 1.16599345e+00
1.07232106e+00 -5.37866890e-01 -1.22359836e+00 2.36470759e-01
1.70601293e-01 4.19651598e-01 3.20550472e-01 -1.22036147e+00
-7.49666333e-01 9.22591448e-01 1.39466512e+00 -4.04453635e-01
1.32019877e+00 -7.84075677e-01 9.64914024e-01 8.23152840e-01
1.97083414e-01 -1.42519355e+00 -1.17068291e-01 9.28379059e-01
1.05692908e-01 -7.27263808e-01 -5.39014116e-03 5.68081021e-01
-1.10362239e-01 1.41531646e+00 9.94171381e-01 -1.27875090e-01
9.02093410e-01 1.24811411e+00 -2.53904492e-01 -3.61890376e-01
-9.17573988e-01 7.03221142e-01 -2.43187413e-01 5.86489379e-01
4.37169105e-01 -1.01135887e-01 1.53121278e-01 6.46210849e-01
1.50324568e-01 4.18449342e-01 5.29310882e-01 1.19770038e+00
-4.26589042e-01 -4.83104199e-01 7.90120587e-02 1.18303633e+00
-5.16030490e-01 -3.53699833e-01 1.95810031e-02 6.56252205e-01
7.20869452e-02 6.70157671e-01 4.89352882e-01 -4.99872446e-01
1.48570478e-01 6.16244972e-01 5.68637252e-01 4.23293486e-02
-1.16962111e+00 -4.21942919e-01 -9.54456329e-02 -2.68824250e-01
1.12783216e-01 -2.17525423e-01 -1.92161608e+00 -5.49211979e-01
-4.14766312e-01 -2.61807114e-01 7.91175187e-01 7.60622740e-01
4.68330145e-01 1.00044692e+00 -9.89899263e-02 -3.20909142e-01
-1.60851449e-01 -8.92790735e-01 -7.75182247e-01 -2.57543534e-01
4.36213434e-01 -3.76428097e-01 -1.67819083e-01 -3.91560972e-01] | [8.264960289001465, 2.5404210090637207] |
d6785268-d63e-4229-b873-6f5cdf747b71 | associative-embedding-end-to-end-learning-for | 1611.05424 | null | http://arxiv.org/abs/1611.05424v2 | http://arxiv.org/pdf/1611.05424v2.pdf | Associative Embedding: End-to-End Learning for Joint Detection and Grouping | We introduce associative embedding, a novel method for supervising
convolutional neural networks for the task of detection and grouping. A number
of computer vision problems can be framed in this manner including multi-person
pose estimation, instance segmentation, and multi-object tracking. Usually the
grouping of detections is achieved with multi-stage pipelines, instead we
propose an approach that teaches a network to simultaneously output detections
and group assignments. This technique can be easily integrated into any
state-of-the-art network architecture that produces pixel-wise predictions. We
show how to apply this method to both multi-person pose estimation and instance
segmentation and report state-of-the-art performance for multi-person pose on
the MPII and MS-COCO datasets. | ['Alejandro Newell', 'Zhiao Huang', 'Jia Deng'] | 2016-11-16 | associative-embedding-end-to-end-learning-for-1 | http://papers.nips.cc/paper/6822-associative-embedding-end-to-end-learning-for-joint-detection-and-grouping | http://papers.nips.cc/paper/6822-associative-embedding-end-to-end-learning-for-joint-detection-and-grouping.pdf | neurips-2017-12 | ['2d-human-pose-estimation'] | ['computer-vision'] | [ 1.12777054e-01 2.91979492e-01 2.43114427e-01 -6.41731739e-01
-5.01612186e-01 -4.25300360e-01 5.32215893e-01 2.46391013e-01
-8.74855459e-01 2.25978345e-01 -8.07712078e-02 2.84589261e-01
3.72640081e-02 -6.12807691e-01 -8.67657840e-01 -2.66195387e-01
-3.80703390e-01 1.03443825e+00 6.36944830e-01 -7.55736977e-02
-1.91217676e-01 5.00312328e-01 -1.51573408e+00 3.64105701e-01
2.39430785e-01 5.50179303e-01 -1.33652017e-01 1.19486439e+00
2.92122394e-01 2.54831046e-01 -7.23341405e-01 -6.88029349e-01
2.69439697e-01 -3.50812543e-03 -8.33750069e-01 1.58004791e-01
1.19346976e+00 -3.81309867e-01 -3.18997443e-01 7.28292525e-01
7.82927334e-01 2.72218615e-01 6.69985712e-01 -1.27877116e+00
-1.37873247e-01 3.08332890e-01 -7.24400938e-01 3.48547131e-01
4.65768814e-01 1.73480496e-01 7.29450405e-01 -7.34226048e-01
5.40801108e-01 1.64315319e+00 1.00528336e+00 6.73019409e-01
-1.37949836e+00 -3.50700766e-01 3.00215364e-01 6.07847720e-02
-1.17668498e+00 -1.27090842e-01 3.31372261e-01 -7.01597095e-01
1.04458523e+00 1.66003525e-01 7.89654553e-01 8.80726635e-01
-5.03581315e-02 1.20170879e+00 8.65647793e-01 -4.21745002e-01
-1.18167140e-01 1.03562079e-01 3.76943350e-01 1.04148257e+00
3.82793427e-01 1.73381433e-01 -6.46077871e-01 1.41218156e-01
9.45596695e-01 -1.01833157e-01 3.56605589e-01 -3.23708057e-01
-1.26763368e+00 7.87858844e-01 7.19779730e-01 7.42297098e-02
-1.69876903e-01 6.92351460e-01 5.06345093e-01 -1.64672494e-01
4.17893946e-01 3.75987291e-01 -4.01520580e-01 4.44310129e-01
-1.36382556e+00 7.67633915e-01 6.84277058e-01 6.20460033e-01
4.56455976e-01 -3.05446982e-01 -4.48716104e-01 4.89163548e-01
4.25749063e-01 -6.55618683e-02 1.52311549e-01 -1.03289270e+00
1.95595846e-01 4.67587829e-01 2.31650800e-01 -5.80728352e-01
-8.71025681e-01 -3.81321996e-01 -3.94105196e-01 5.81224799e-01
7.84406304e-01 -4.06960368e-01 -1.17788434e+00 1.52954805e+00
5.69160938e-01 1.46412149e-01 -4.67299521e-01 1.09165680e+00
1.00098956e+00 3.42140466e-01 3.51697773e-01 5.75220227e-01
1.65374088e+00 -1.31647289e+00 -2.80853301e-01 -6.55170918e-01
2.02879786e-01 -3.80411446e-01 3.36401612e-01 3.02148014e-01
-1.21978533e+00 -1.07431185e+00 -1.07213867e+00 -1.97533578e-01
-5.60153663e-01 4.43567187e-01 8.21762979e-01 8.45095992e-01
-1.20290935e+00 7.56170154e-01 -1.10521924e+00 -6.20640934e-01
6.07679069e-01 8.63495290e-01 -4.48871315e-01 3.00162613e-01
-7.30798066e-01 1.02338493e+00 6.31125271e-01 1.55126438e-01
-8.20522964e-01 -4.32891130e-01 -7.86589622e-01 7.26103559e-02
2.75711596e-01 -1.05494285e+00 1.30725574e+00 -7.58712590e-01
-1.06330490e+00 1.15783441e+00 -1.89486910e-02 -6.74665272e-01
7.45502949e-01 -6.00509048e-01 -1.20025985e-01 2.07320318e-01
1.14810303e-01 1.58704114e+00 7.59831786e-01 -9.57118213e-01
-8.19527388e-01 -6.22825146e-01 1.42864183e-01 1.91493168e-01
-9.37391967e-02 4.03805077e-01 -8.09728324e-01 -3.20601314e-01
9.04925987e-02 -1.08949399e+00 -6.74529672e-01 4.44774330e-01
-6.11882985e-01 -6.22412860e-01 8.32433999e-01 -6.32488728e-01
5.29919446e-01 -1.81995308e+00 4.33423758e-01 -1.57505702e-02
2.93360710e-01 1.33048013e-01 -8.86139274e-02 -6.67220131e-02
-1.78016856e-01 -2.75519729e-01 -4.84975949e-02 -1.08010578e+00
2.60124117e-01 8.89354665e-03 3.08068216e-01 6.59198046e-01
3.78370047e-01 1.12089276e+00 -6.87985420e-01 -7.09931910e-01
5.55878341e-01 6.13487661e-01 -6.20857120e-01 1.10991389e-01
-2.57309377e-01 4.44995880e-01 -2.82947011e-02 4.53962713e-01
2.71694422e-01 -8.78547505e-02 9.07014385e-02 -1.90572038e-01
-8.40263888e-02 -1.10921219e-01 -1.58321869e+00 1.94842064e+00
9.49608162e-02 7.92539299e-01 4.79268953e-02 -8.56461585e-01
4.70142573e-01 1.93720490e-01 3.76203448e-01 -2.53618858e-03
2.12321684e-01 -2.02712238e-01 3.08768991e-02 -2.42407516e-01
7.43379712e-01 2.28007138e-01 -2.20102608e-01 3.24454010e-01
5.99562585e-01 2.84090400e-01 5.03025830e-01 1.62816167e-01
8.70797396e-01 5.60687900e-01 -4.89025228e-02 -1.15623161e-01
3.01996469e-01 -2.01203674e-02 2.89124995e-01 8.64751220e-01
-4.91679043e-01 6.63686931e-01 1.85383677e-01 -7.93883801e-01
-1.03630626e+00 -1.15523958e+00 1.09176589e-02 1.65867615e+00
-6.97401240e-02 -3.65346283e-01 -9.21708167e-01 -6.58088744e-01
1.99779972e-01 1.51204810e-01 -8.37336600e-01 3.41916621e-01
-7.65647352e-01 -8.98663282e-01 6.73607886e-01 9.99125242e-01
4.11265552e-01 -1.19062364e+00 -9.18742359e-01 3.76213312e-01
2.21272469e-01 -1.19195318e+00 -2.82125562e-01 3.01439017e-01
-5.39562881e-01 -1.10837400e+00 -8.42336178e-01 -9.77614343e-01
5.71462154e-01 -2.99077451e-01 1.30825019e+00 -4.60388549e-02
-7.94744492e-01 6.62478745e-01 1.73215821e-01 -5.53100169e-01
5.28087690e-02 2.28067622e-01 2.36244321e-01 -2.04452336e-01
4.46184665e-01 -3.48587245e-01 -8.44331920e-01 1.32488027e-01
-5.67734301e-01 -6.80606589e-02 6.22556806e-01 4.76567924e-01
5.84788501e-01 -2.10905448e-01 1.86230123e-01 -8.35552752e-01
3.24642748e-01 1.90758958e-01 -5.18720388e-01 1.91920951e-01
5.04854061e-02 -6.60129189e-02 -6.95984140e-02 -3.85682166e-01
-6.60748541e-01 8.95009100e-01 -2.73741812e-01 -3.73276949e-01
-6.73447967e-01 -5.57794645e-02 -7.26042315e-03 -2.41517112e-01
7.17782199e-01 -1.36910990e-01 -2.09545657e-01 -4.19903159e-01
7.93181896e-01 4.85108018e-01 1.12153423e+00 -4.44016844e-01
6.05612338e-01 4.80743617e-01 6.32875934e-02 -6.86409056e-01
-8.83941054e-01 -7.17581809e-01 -1.27818394e+00 -4.90738064e-01
1.48261476e+00 -1.05531824e+00 -1.01064348e+00 4.29546565e-01
-1.47045910e+00 -3.04276317e-01 -9.54052657e-02 2.82608271e-01
-5.09035587e-01 1.10070683e-01 -7.40115881e-01 -7.90478349e-01
-3.69413942e-01 -1.05571663e+00 1.37365496e+00 4.93486255e-01
-5.03871918e-01 -8.46195638e-01 8.55057836e-02 4.66110170e-01
2.20699552e-02 3.36226404e-01 1.97794527e-01 -7.38389254e-01
-5.72445393e-01 -5.73143601e-01 -1.85464337e-01 -1.45744845e-01
-6.16413534e-01 -1.99042901e-01 -1.17200243e+00 -4.31510419e-01
-7.12087154e-01 -4.61457610e-01 1.09003389e+00 6.84347093e-01
1.39714003e+00 1.04150593e-01 -7.23659277e-01 4.46508020e-01
1.11033630e+00 -4.71431345e-01 4.11712736e-01 3.65329802e-01
9.66861129e-01 7.81874955e-01 1.54269338e-01 1.20340727e-01
5.07165492e-01 1.04677677e+00 3.68972719e-01 -3.10628682e-01
-3.58283848e-01 -4.14191373e-02 3.58371474e-02 -9.96994376e-02
-2.73599058e-01 -3.55208083e-03 -8.26490879e-01 6.31164849e-01
-2.10851359e+00 -1.18302631e+00 -2.65759677e-01 1.88421428e+00
4.54771340e-01 3.07811439e-01 7.90965617e-01 -7.79504515e-03
8.76243174e-01 7.30993599e-02 -3.58075887e-01 -2.23567680e-01
2.26743340e-01 2.91784376e-01 5.55790186e-01 3.56538028e-01
-1.85828161e+00 1.02648199e+00 7.55772638e+00 3.65587622e-01
-6.79468930e-01 2.79421568e-01 5.92047751e-01 -2.61114359e-01
5.64278781e-01 -4.54530805e-01 -1.28951931e+00 2.10880727e-01
9.24504936e-01 7.13120818e-01 4.60744351e-02 9.10934627e-01
-5.26084453e-02 -1.62559047e-01 -1.53182745e+00 1.01062226e+00
1.21004142e-01 -1.17504728e+00 -3.52622718e-01 -2.01862261e-01
5.84704757e-01 -2.28168946e-02 6.05672039e-02 3.83035272e-01
3.31102073e-01 -1.06775689e+00 7.15883553e-01 4.27349836e-01
5.33069551e-01 -8.06057155e-01 3.99379045e-01 3.09959292e-01
-1.26353490e+00 -6.93091378e-02 -1.68661743e-01 -7.61183053e-02
2.37162367e-01 2.11511567e-01 -8.31378698e-01 2.79563248e-01
1.03672183e+00 4.79433954e-01 -1.03301418e+00 1.48873103e+00
-3.98478717e-01 1.07272550e-01 -4.88058835e-01 -1.08062224e-02
1.82355911e-01 1.59591377e-01 4.80005980e-01 1.66835368e+00
-1.80613220e-01 -2.77155638e-01 6.17242098e-01 7.85345554e-01
1.64115608e-01 -3.25371653e-01 -1.91441551e-01 3.71525466e-01
1.53883934e-01 1.48726177e+00 -1.10075521e+00 -6.05381489e-01
-3.10723543e-01 1.28289914e+00 5.01910150e-01 1.02234855e-01
-8.47198367e-01 -2.36672476e-01 6.22800767e-01 2.23280326e-01
6.48724735e-01 -4.79031116e-01 -3.41757864e-01 -7.19177842e-01
-2.36140907e-01 -4.14114118e-01 5.32904267e-01 -7.69996762e-01
-1.11638463e+00 3.66734236e-01 1.99324086e-01 -6.20649636e-01
-4.82473075e-01 -1.03818011e+00 -6.89612031e-01 6.95248663e-01
-9.50057805e-01 -1.52700520e+00 -3.01594883e-01 4.44748014e-01
5.81874251e-01 4.54710610e-02 9.12767708e-01 4.33948219e-01
-7.79207766e-01 6.28624737e-01 -5.19125164e-01 6.70047522e-01
6.28276467e-01 -1.72236109e+00 5.67776382e-01 1.06387866e+00
5.32027304e-01 3.64500195e-01 7.93791175e-01 -4.59670722e-01
-8.57217669e-01 -1.23022294e+00 7.73798287e-01 -8.60030830e-01
2.50299603e-01 -5.80798984e-01 -3.87219995e-01 1.14141500e+00
3.92142683e-01 1.84418738e-01 5.66542149e-01 5.36670387e-01
-1.49373710e-01 1.57200485e-01 -9.55869377e-01 5.48129559e-01
1.07803106e+00 -2.28383765e-01 -6.49446666e-01 7.13714123e-01
4.16320801e-01 -9.04403806e-01 -5.03988564e-01 1.67429224e-02
6.62379622e-01 -9.03326333e-01 1.44969785e+00 -7.32077837e-01
2.66073495e-01 -5.28137982e-01 2.08203167e-01 -1.07110858e+00
-4.84255731e-01 -2.35537887e-01 -2.73486197e-01 8.99218321e-01
2.98593760e-01 -3.90399694e-02 1.09244692e+00 6.55506730e-01
-1.27162337e-01 -5.41342258e-01 -9.02666986e-01 -3.92266184e-01
-1.38398066e-01 -3.24105769e-01 1.67064622e-01 2.98483670e-01
-4.94660318e-01 3.31325740e-01 -3.78648192e-01 4.71452624e-01
8.85496020e-01 -1.07226044e-01 9.39515293e-01 -1.38022864e+00
-6.14869297e-01 -3.63951802e-01 -9.17377949e-01 -1.05058324e+00
4.90510426e-02 -7.23226368e-01 1.82954162e-01 -1.72092140e+00
2.71061331e-01 -5.02697527e-02 -1.33005872e-01 6.38433933e-01
-2.07884386e-01 7.06376553e-01 3.59066755e-01 -5.50903082e-02
-1.09803152e+00 -8.32104832e-02 8.37371409e-01 -2.58996129e-01
-1.88306078e-01 6.02929704e-02 -3.04104298e-01 8.30121398e-01
4.64735389e-01 -4.54743326e-01 1.94864869e-01 -4.88844305e-01
9.41483025e-03 -2.96465069e-01 1.00825608e+00 -1.64666450e+00
5.83797991e-01 5.17575979e-01 1.24858117e+00 -9.40337360e-01
7.75220871e-01 -4.49135393e-01 -5.24979718e-02 6.54811978e-01
-2.74672478e-01 5.99651635e-02 3.75786185e-01 5.17564595e-01
1.13509156e-01 -1.18534125e-01 7.17669785e-01 -4.56138313e-01
-1.12264776e+00 3.18558574e-01 -1.84213921e-01 -3.21359605e-01
1.09123492e+00 -3.42623979e-01 -6.15301393e-02 -7.35457912e-02
-1.38904691e+00 4.38682646e-01 -6.27016136e-03 5.04450023e-01
4.53414112e-01 -1.24220049e+00 -8.08374047e-01 5.37399352e-02
-1.56814262e-01 9.36145261e-02 8.89277011e-02 8.23977053e-01
-5.20643950e-01 3.29711765e-01 -3.15122664e-01 -9.35272992e-01
-1.60072672e+00 2.37420604e-01 7.38393486e-01 -3.51668924e-01
-5.66074848e-01 1.21763265e+00 -9.25388187e-02 -5.27755737e-01
4.27583337e-01 1.51161388e-01 -2.31971696e-01 2.10759863e-01
7.34426022e-01 2.91419476e-01 -8.64187628e-02 -6.44436002e-01
-6.01148188e-01 5.64005733e-01 -8.46480727e-02 -2.61383891e-01
1.31416059e+00 9.71191451e-02 1.22106731e-01 2.24258319e-01
8.41031551e-01 -5.75418293e-01 -1.41163933e+00 1.01697028e-01
-1.67931199e-01 -1.81935072e-01 3.45950462e-02 -9.61288631e-01
-9.28342342e-01 9.49532330e-01 1.02201426e+00 -4.25172783e-02
5.63259304e-01 3.45432967e-01 3.81533563e-01 3.50017101e-01
2.40874439e-01 -1.31838953e+00 2.51354337e-01 3.49046141e-01
5.64928353e-01 -1.36969471e+00 1.55425370e-01 -4.80048180e-01
-3.57338071e-01 1.12373710e+00 9.42195475e-01 -4.33825046e-01
4.11700070e-01 3.19144815e-01 -1.10224150e-01 -3.53369147e-01
-1.77858502e-01 -6.40836716e-01 7.16061234e-01 6.49152040e-01
6.11638606e-01 2.34037295e-01 -6.83201104e-02 4.13035005e-01
-1.74153849e-01 -8.70435014e-02 1.63076937e-01 7.91167319e-01
-6.01565182e-01 -1.04964161e+00 -7.39186466e-01 4.34974521e-01
-4.85635757e-01 2.65808254e-01 -5.03921092e-01 8.73169720e-01
6.83562458e-01 6.24912441e-01 3.86971384e-01 -2.50205845e-01
2.16127068e-01 3.09706032e-01 8.42665493e-01 -8.97495270e-01
-9.20271635e-01 3.41379531e-02 1.07961655e-01 -6.51037693e-01
-7.31004953e-01 -8.86017859e-01 -1.12903333e+00 -1.56148329e-01
-1.63097754e-01 -4.39935833e-01 6.73445940e-01 1.07243276e+00
-1.18781924e-02 9.05894816e-01 -2.82897025e-01 -1.45781803e+00
-5.91861755e-02 -1.08361030e+00 -3.58694971e-01 3.01393777e-01
1.60445914e-01 -6.98583424e-01 2.55781889e-01 -4.07198854e-02] | [7.214658737182617, -0.7910486459732056] |
69a5be27-d33c-4547-8c54-ea8d0a48bc0c | ct-icp-real-time-elastic-lidar-odometry-with | 2109.12979 | null | https://arxiv.org/abs/2109.12979v2 | https://arxiv.org/pdf/2109.12979v2.pdf | CT-ICP: Real-time Elastic LiDAR Odometry with Loop Closure | Multi-beam LiDAR sensors are increasingly used in robotics, particularly with autonomous cars for localization and perception tasks, both relying on the ability to build a precise map of the environment. For this, we propose a new real-time LiDAR-only odometry method called CT-ICP (for Continuous-Time ICP), completed into a full SLAM with a novel loop detection procedure. The core of this method, is the introduction of the combined continuity in the scan matching, and discontinuity between scans. It allows both the elastic distortion of the scan during the registration for increased precision, and the increased robustness to high frequency motions from the discontinuity. We build a complete SLAM on top of this odometry, using a fast pure LiDAR loop detection based on elevation image 2D matching, providing a pose graph with loop constraints. To show the robustness of the method, we tested it on seven datasets: KITTI, KITTI-raw, KITTI-360, KITTI-CARLA, ParisLuco, Newer College, and NCLT in driving and high-frequency motion scenarios. Both the CT-ICP odometry and the loop detection are made available online. CT-ICP is currently first, among those giving access to a public code, on the KITTI odometry leaderboard, with an average Relative Translation Error (RTE) of 0.59% and an average time per scan of 60ms on a CPU with a single thread. | ['François Goulette', 'Bastien Jacquet', 'Jean-Emmanuel Deschaud', 'Pierre Dellenbach'] | 2021-09-27 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [-5.93560226e-02 -3.74131687e-02 1.47569761e-01 -4.25246447e-01
-5.82449913e-01 -3.74783576e-01 6.76325023e-01 2.22560927e-01
-7.05012500e-01 4.76084650e-01 -4.70595688e-01 -3.42062384e-01
-2.10149124e-01 -9.92305875e-01 -9.57579434e-01 -2.60932982e-01
-1.10641897e-01 1.25568676e+00 7.51763225e-01 -4.49101031e-01
3.53102773e-01 9.55124557e-01 -1.72894430e+00 -6.17697537e-01
9.28736508e-01 8.63431871e-01 4.30462569e-01 5.09095430e-01
2.71509197e-02 5.30350097e-02 1.11888163e-01 -2.03109905e-01
4.37925190e-01 3.12614888e-01 -1.28170401e-01 -1.87147141e-01
7.46099830e-01 -2.81924874e-01 -6.64396361e-02 8.43050778e-01
4.64063972e-01 -3.32165919e-02 4.08697188e-01 -1.53218579e+00
4.00555968e-01 -7.62180379e-03 -5.15397131e-01 -3.38504761e-01
4.94350940e-01 7.75001049e-02 5.58455110e-01 -1.23711503e+00
9.09757614e-01 1.18006635e+00 9.57355917e-01 -1.41871586e-01
-1.27195334e+00 -8.12860668e-01 -5.68227291e-01 3.36141855e-01
-1.59399104e+00 -3.51497680e-01 4.81492609e-01 -5.43635607e-01
8.16386878e-01 -5.49869202e-02 7.01044440e-01 5.75938582e-01
6.66936576e-01 -1.93075627e-01 1.08857501e+00 -1.88843086e-01
2.79986829e-01 3.42322588e-02 1.99461672e-02 7.57068098e-01
7.14785576e-01 4.48437989e-01 -5.19073188e-01 -4.50991988e-02
4.05373007e-01 1.72118425e-01 1.68951079e-01 -1.06512249e+00
-1.09406900e+00 8.60539496e-01 5.15192628e-01 -1.62390277e-01
-1.56112328e-01 1.98745623e-01 2.23383561e-01 3.33594203e-01
2.23669678e-01 -2.62688510e-02 -1.73651606e-01 -3.18889648e-01
-6.95627928e-01 4.86818194e-01 8.02161992e-01 1.34894288e+00
1.52959847e+00 -3.15535754e-01 6.25822306e-01 2.16925666e-01
3.93046379e-01 1.23766840e+00 2.15572745e-01 -8.39410067e-01
5.74440897e-01 6.27692521e-01 2.83384562e-01 -1.08232498e+00
-9.27093625e-01 -2.12852642e-01 -5.30321419e-01 6.32888556e-01
1.47687688e-01 -4.85002361e-02 -6.82189047e-01 1.28360808e+00
6.43491268e-01 1.38264999e-01 -4.53620255e-02 6.79995835e-01
3.70750874e-01 2.42859468e-01 -6.71993256e-01 -1.07745826e-01
1.34094810e+00 -5.29171526e-01 -5.50490797e-01 -5.28175175e-01
8.39822590e-01 -9.43018258e-01 7.10234225e-01 4.15283144e-01
-4.98818725e-01 -7.16816366e-01 -1.48270631e+00 -2.71383971e-01
-3.40223432e-01 -5.99425808e-02 1.38623238e-01 3.98821473e-01
-1.10035837e+00 4.81022686e-01 -9.06684399e-01 -6.38963878e-01
-2.26612687e-01 3.70488107e-01 -5.96700191e-01 -3.39141995e-01
-8.99362147e-01 1.46558249e+00 2.51473457e-01 -1.85701512e-02
-4.74834889e-01 -5.71846783e-01 -9.94220614e-01 -4.44002330e-01
4.25101161e-01 -4.99328673e-01 9.34458435e-01 9.96149778e-02
-1.60011649e+00 8.11990440e-01 -4.65610266e-01 -7.11033881e-01
8.70174050e-01 -4.94943827e-01 -1.65740162e-01 -1.16249308e-01
3.86833996e-01 6.87006354e-01 5.01569748e-01 -1.26937151e+00
-5.72036684e-01 -6.71420217e-01 -4.27984327e-01 4.74728309e-02
5.53939998e-01 -5.18294394e-01 -2.36146271e-01 2.08848625e-01
8.56581032e-01 -1.49667454e+00 -3.32216412e-01 1.91456322e-02
-6.25986606e-02 1.20762408e-01 9.87010539e-01 -2.18805060e-01
4.13037598e-01 -2.17868614e+00 -1.23035967e-01 3.81128520e-01
-1.85444728e-01 -1.07947461e-01 1.61313698e-01 5.31965137e-01
3.47386688e-01 -4.64087456e-01 -4.02420700e-01 -7.69545555e-01
-1.07942261e-01 5.64431608e-01 -3.62948954e-01 9.67876315e-01
-4.72871363e-02 6.28849089e-01 -7.76432037e-01 -2.89722145e-01
4.90954101e-01 4.45530683e-01 -4.04225320e-01 -1.47752061e-01
1.42773837e-01 5.35319865e-01 -1.40109867e-01 3.74121636e-01
1.21604133e+00 6.46009207e-01 -6.70764074e-02 -5.66865727e-02
-9.20114636e-01 5.29542089e-01 -1.72453463e+00 1.90217054e+00
-4.89397496e-01 5.67026317e-01 3.10147732e-01 -2.38428757e-01
1.40399301e+00 -1.14006802e-01 4.19962943e-01 -8.44319642e-01
1.02341197e-01 8.30898404e-01 -2.39095971e-01 -5.52793555e-02
9.64056611e-01 5.67973107e-02 -1.14634931e-01 2.95637667e-01
-2.77110040e-01 -9.84867036e-01 2.03463867e-01 2.30618771e-02
8.90793324e-01 3.31814259e-01 2.07903847e-01 -3.89762312e-01
5.08092403e-01 3.62835675e-01 4.47427124e-01 3.20431143e-01
-8.05272069e-03 4.36491936e-01 -5.37027679e-02 -2.37832978e-01
-1.16670227e+00 -1.09712195e+00 -4.86772835e-01 3.31131995e-01
5.93942761e-01 -3.28383923e-01 -2.09881485e-01 5.93191199e-02
5.93947470e-01 4.80218112e-01 -2.06589028e-01 -2.45947596e-02
-9.13385868e-01 -1.81167871e-01 3.89509529e-01 1.99675679e-01
5.82357705e-01 -5.06846130e-01 -1.05777085e+00 3.50958318e-01
-1.34843858e-02 -1.38226271e+00 -6.41943514e-02 3.59515309e-01
-1.08560061e+00 -1.00580192e+00 1.10405691e-01 -4.96407539e-01
3.03957134e-01 4.85728681e-01 7.52755523e-01 -1.91618592e-01
-1.81177527e-01 2.77238730e-02 -4.70203273e-02 -4.80864316e-01
-3.37537885e-01 -8.60927776e-02 4.61461008e-01 -3.27206105e-01
1.14447013e-01 -8.30454171e-01 -1.31613225e-01 7.54997790e-01
-3.71833205e-01 -7.68259317e-02 6.13329113e-01 3.78405541e-01
7.56399274e-01 -3.81900012e-01 -5.82091548e-02 -4.04401988e-01
-3.72672826e-02 -3.27870280e-01 -1.20654428e+00 -7.18879879e-01
-8.17665756e-01 -3.36319096e-02 1.32123336e-01 -2.42102128e-02
-5.77173591e-01 5.44085085e-01 -3.29170674e-01 -3.40601265e-01
-1.05284192e-01 3.87004167e-01 -6.58953935e-02 -4.60939169e-01
7.09165692e-01 6.95881918e-02 4.01084393e-01 -4.77946758e-01
5.88351965e-01 4.90857095e-01 7.32379794e-01 -2.57310957e-01
1.26411450e+00 9.02214766e-01 6.29498482e-01 -8.54281306e-01
-1.06383488e-01 -8.17954123e-01 -1.12169802e+00 -1.55581817e-01
6.46283031e-01 -1.06314075e+00 -7.93218851e-01 4.31963354e-01
-1.26280332e+00 -1.84508562e-01 -2.41624683e-01 7.99792469e-01
-7.00407088e-01 3.62524956e-01 -1.71482861e-01 -6.96685135e-01
-8.32495913e-02 -1.35290062e+00 1.11089289e+00 -9.56140608e-02
3.37218270e-02 -4.33213145e-01 4.50268209e-01 9.69165489e-02
2.31652796e-01 3.86258692e-01 1.98082283e-01 -7.63318092e-02
-9.60739315e-01 -3.23158860e-01 -1.89124331e-01 -2.34685495e-01
-1.27834409e-01 -6.77331677e-03 -8.38957608e-01 -4.70572352e-01
2.23692395e-02 2.00645067e-02 6.82043910e-01 -2.30258666e-02
-3.99398096e-02 2.39089966e-01 -5.86273849e-01 6.79535031e-01
1.74184692e+00 1.15718795e-02 5.34286857e-01 6.49361789e-01
5.81964910e-01 5.44304788e-01 1.20991433e+00 2.44877249e-01
7.22759783e-01 1.21662819e+00 9.88363028e-01 1.26859233e-01
1.38125151e-01 -2.36789718e-01 4.71647292e-01 9.46649194e-01
-6.49099573e-02 4.12763596e-01 -1.09492660e+00 4.21021700e-01
-1.81515825e+00 -4.55814153e-01 -8.83158565e-01 2.67738080e+00
2.84975648e-01 4.76095676e-01 -1.27158463e-01 2.74130017e-01
3.72289240e-01 -5.98783866e-02 -3.18612427e-01 -5.54448724e-01
2.32172489e-01 1.24061599e-01 1.27101839e+00 1.01661038e+00
-6.64942265e-01 8.64701629e-01 5.09943676e+00 2.39832401e-01
-1.30892873e+00 3.46482843e-01 -7.21630096e-01 3.14260274e-01
9.10455212e-02 3.45545560e-01 -1.24159682e+00 2.01951072e-01
1.16679907e+00 -1.55905858e-02 3.86503376e-02 1.11213768e+00
3.39193285e-01 -5.91501653e-01 -8.52830946e-01 8.36915731e-01
-4.94943224e-02 -1.03750956e+00 -4.62967187e-01 4.66514081e-01
2.60085791e-01 8.11576068e-01 -3.97366792e-01 1.52057558e-01
2.33705789e-01 -4.86199439e-01 1.10058153e+00 1.33502334e-01
8.14042866e-01 -8.13336790e-01 6.78460836e-01 7.32553840e-01
-1.63636315e+00 1.31539404e-01 -5.40273368e-01 -3.73855799e-01
5.06469369e-01 9.32982981e-01 -1.22071743e+00 1.06235838e+00
4.66613621e-01 6.69780254e-01 -5.20476878e-01 9.47885513e-01
-3.47782135e-01 -4.34652083e-02 -7.36201346e-01 1.76009238e-01
3.44404764e-02 -5.78095913e-01 8.77582490e-01 9.00332510e-01
7.56838620e-01 -4.14316386e-01 3.71996254e-01 6.70730829e-01
4.10977334e-01 -1.63732772e-03 -8.38329613e-01 8.69075000e-01
5.47114968e-01 1.23218465e+00 -4.22943741e-01 -9.69655290e-02
-8.43546391e-02 5.17015398e-01 3.28925467e-04 -2.30726793e-01
-8.70698094e-01 -3.59951884e-01 9.23140824e-01 4.18767244e-01
1.58876821e-01 -1.10343134e+00 -4.27424073e-01 -7.10501552e-01
2.26919010e-01 -1.69548973e-01 -2.77966559e-01 -8.04695964e-01
-4.33230758e-01 3.86221647e-01 1.11506708e-01 -1.56690586e+00
-4.97316033e-01 -5.71691573e-01 -2.42284819e-01 8.86240423e-01
-1.86484408e+00 -9.56871450e-01 -7.16953158e-01 5.43831289e-01
2.21216097e-01 4.63248134e-01 5.82474053e-01 5.88730931e-01
1.91693660e-02 -5.57179675e-02 -6.53934404e-02 -3.63998532e-01
8.53831112e-01 -1.00106502e+00 7.32748151e-01 8.36385906e-01
1.35892332e-02 4.73181605e-01 1.09602320e+00 -8.87709260e-01
-1.78126931e+00 -1.10006261e+00 1.15290654e+00 -4.67019826e-01
5.57968020e-01 -6.06193960e-01 -7.19916224e-01 8.60911369e-01
-3.35168779e-01 -1.66135971e-02 -3.34340513e-01 -1.96771875e-01
-1.67911127e-01 -4.35552955e-01 -1.27161372e+00 2.25094557e-01
1.20385432e+00 -3.52022052e-01 -5.86399376e-01 2.60541081e-01
7.72258699e-01 -1.06417823e+00 -7.46630847e-01 6.27781510e-01
6.44690335e-01 -1.11853075e+00 8.50618422e-01 6.52199149e-01
-3.22203279e-01 -8.55373502e-01 -1.52534544e-01 -1.05063140e+00
-5.73941059e-02 -4.93686944e-01 3.84329259e-01 8.97263229e-01
8.68507922e-02 -1.26576138e+00 6.77565932e-01 -1.14301100e-01
-5.65975666e-01 -2.17608333e-01 -1.60827732e+00 -9.29665387e-01
-3.05615783e-01 -9.35010135e-01 5.42398334e-01 4.56130087e-01
-3.00472736e-01 3.10240328e-01 -1.46520928e-01 6.47591591e-01
8.12476695e-01 3.89847308e-02 1.55556333e+00 -1.79240286e+00
1.30379409e-01 2.98344102e-02 -9.68551815e-01 -1.07436085e+00
-1.36599377e-01 -6.21980250e-01 4.42226857e-01 -1.30354726e+00
-5.09101868e-01 -6.32929444e-01 5.63960612e-01 1.40632808e-01
5.46369076e-01 2.37020403e-01 8.98563638e-02 4.81542856e-01
-1.75850990e-03 3.99281949e-01 8.12918425e-01 2.30305746e-01
-2.33098164e-01 1.44276157e-01 3.82442653e-01 7.95129955e-01
3.95791590e-01 -7.80640304e-01 -1.88841134e-01 -2.99704790e-01
4.88425195e-01 -6.49435297e-02 4.24339950e-01 -1.63456297e+00
5.01110792e-01 5.54302111e-02 -2.51933277e-01 -1.31846368e+00
7.50643611e-01 -1.07136273e+00 5.86361706e-01 1.00675225e+00
5.99258780e-01 4.94011045e-01 2.88974613e-01 4.73693103e-01
-2.12491378e-01 -5.29253893e-02 8.66550207e-01 2.00615287e-01
-8.86863053e-01 2.87598968e-01 -1.32055044e-01 -4.01054263e-01
1.05620074e+00 -5.12126267e-01 -2.21711069e-01 -1.63459197e-01
-3.19653094e-01 2.65616268e-01 1.01632833e+00 3.04656416e-01
5.39611816e-01 -1.21355498e+00 -5.77340662e-01 6.02403879e-01
3.95325035e-01 4.60127801e-01 -6.23676255e-02 1.27424371e+00
-8.38536918e-01 5.48489690e-01 -1.97116122e-01 -1.31375945e+00
-1.22056842e+00 3.32734793e-01 2.86236435e-01 -6.10477328e-02
-7.52836943e-01 3.44822526e-01 -3.50566804e-01 -9.05183673e-01
-1.61657199e-01 -8.41787219e-01 9.75171924e-02 4.33599018e-03
1.45495966e-01 6.13090098e-01 6.17018223e-01 -1.01875842e+00
-6.81018293e-01 1.29842031e+00 6.12153649e-01 -5.13813257e-01
1.00199366e+00 -4.78477865e-01 -1.38732180e-01 5.76476395e-01
1.01519346e+00 4.40689236e-01 -1.05846143e+00 9.02632717e-03
2.57088691e-01 -4.10217881e-01 -1.53738350e-01 -9.04371515e-02
-4.91127670e-01 8.07638764e-01 9.01326478e-01 -7.78111517e-02
4.52010065e-01 -3.41531426e-01 7.31877089e-01 4.22477603e-01
1.20113134e+00 -7.00292587e-01 -3.98002118e-01 1.05369854e+00
8.98539007e-01 -1.12516272e+00 3.06073964e-01 -6.48053825e-01
-1.35271862e-01 1.08269870e+00 1.58963606e-01 -3.84207547e-01
5.76952934e-01 5.23919463e-01 1.54416114e-01 -2.82069370e-02
-3.48131657e-01 -3.93460125e-01 -1.65666729e-01 5.79165518e-01
-1.53477222e-01 3.21932361e-02 -3.63947392e-01 -3.11141551e-01
-5.93785465e-01 1.02785863e-01 5.95174015e-01 1.24506867e+00
-8.82278502e-01 -1.11732101e+00 -5.48697650e-01 -1.25771061e-01
3.89603257e-01 1.99441314e-01 1.56044094e-02 1.12179828e+00
3.80266786e-01 8.50755513e-01 3.37286711e-01 -5.38476288e-01
8.95916402e-01 -3.97073142e-02 2.79608786e-01 -4.73238587e-01
-2.32216537e-01 -2.25716576e-01 1.62638903e-01 -1.06451750e+00
-2.25718439e-01 -7.88687289e-01 -1.60869610e+00 -4.02024359e-01
-3.57569575e-01 1.30467758e-01 1.46027350e+00 8.64665091e-01
5.36233306e-01 2.66307909e-02 4.93202120e-01 -1.25602722e+00
-5.11709094e-01 -9.51615572e-01 -6.03413165e-01 3.38447727e-02
3.91444296e-01 -1.06555843e+00 -4.73913252e-01 -4.89321262e-01] | [7.37793493270874, -2.183598518371582] |
2d7fced2-180c-47e2-9e25-f499a982a33e | addressing-the-rank-degeneration-in | 2306.11986 | null | https://arxiv.org/abs/2306.11986v1 | https://arxiv.org/pdf/2306.11986v1.pdf | Addressing the Rank Degeneration in Sequential Recommendation via Singular Spectrum Smoothing | Sequential recommendation (SR) investigates the dynamic user preferences modeling and generates the next-item prediction. The next item preference is typically generated by the affinity between the sequence and item representations. However, both sequence and item representations suffer from the rank degeneration issue due to the data sparsity problem. The rank degeneration issue significantly impairs the representations for SR. This motivates us to measure how severe is the rank degeneration issue and alleviate the sequence and item representation rank degeneration issues simultaneously for SR. In this work, we theoretically connect the sequence representation degeneration issue with the item rank degeneration, particularly for short sequences and cold items. We also identify the connection between the fast singular value decay phenomenon and the rank collapse issue in transformer sequence output and item embeddings. We propose the area under the singular value curve metric to evaluate the severity of the singular value decay phenomenon and use it as an indicator of rank degeneration. We further introduce a novel singular spectrum smoothing regularization to alleviate the rank degeneration on both sequence and item sides, which is the Singular sPectrum sMoothing for sequential Recommendation (SPMRec). We also establish a correlation between the ranks of sequence and item embeddings and the rank of the user-item preference prediction matrix, which can affect recommendation diversity. We conduct experiments on four benchmark datasets to demonstrate the superiority of SPMRec over the state-of-the-art recommendation methods, especially in short sequences. The experiments also demonstrate a strong connection between our proposed singular spectrum smoothing and recommendation diversity. | ['Philip S. Yu', 'Hao Peng', 'Zhiwei Liu', 'Ziwei Fan'] | 2023-06-21 | null | null | null | null | ['sequential-recommendation'] | ['miscellaneous'] | [-4.99766767e-02 -8.16023648e-01 -3.02089542e-01 -2.20043510e-01
-2.65025675e-01 -6.24955058e-01 3.31749737e-01 2.70300936e-02
-1.75185338e-01 2.65006244e-01 8.14546585e-01 -1.17400758e-01
-4.30183589e-01 -5.23492873e-01 -4.15143669e-01 -7.91270912e-01
-6.11115023e-02 -1.02081433e-01 2.66430259e-01 -5.00966311e-01
4.83821392e-01 -4.05008495e-02 -1.63173282e+00 4.04380322e-01
1.31667936e+00 1.11124480e+00 3.46041858e-01 1.89751044e-01
-2.94262171e-01 1.74944878e-01 -2.38279149e-01 -2.49288127e-01
3.97081614e-01 -4.40294832e-01 -2.94252783e-01 -1.17135093e-01
2.57172614e-01 -2.32231930e-01 -1.95719093e-01 1.17610240e+00
6.72640979e-01 4.95350063e-01 6.43941820e-01 -1.14676821e+00
-1.09020865e+00 7.79461563e-01 -7.74980366e-01 4.76932436e-01
3.36735755e-01 -8.20130110e-02 1.30256951e+00 -1.29738688e+00
2.44628191e-01 1.17844999e+00 7.10004568e-01 1.66806638e-01
-9.14933801e-01 -5.48554957e-01 3.88778955e-01 2.90069073e-01
-1.22923815e+00 7.57982582e-02 6.28659785e-01 -6.34521127e-01
4.90097791e-01 2.61056960e-01 4.96557593e-01 1.13637662e+00
-3.43020260e-03 8.44023585e-01 7.38485038e-01 3.69875529e-03
8.10212716e-02 5.36978943e-03 7.01348901e-01 1.24678046e-01
3.41898590e-01 2.00133562e-01 -3.86778682e-01 -3.28289121e-01
8.90795290e-01 5.03966928e-01 -3.76574010e-01 -2.71915376e-01
-1.05308402e+00 9.13978815e-01 3.95742983e-01 2.23640621e-01
-3.14075619e-01 -1.99154630e-01 5.65486610e-01 5.43929994e-01
1.85003296e-01 3.29779536e-01 -5.26945055e-01 -1.23370677e-01
-5.46710372e-01 1.42831832e-01 3.92089516e-01 7.00746477e-01
2.88182795e-01 1.29637256e-01 -4.69689757e-01 1.22745073e+00
4.36171710e-01 3.14952403e-01 9.89502192e-01 -4.63461339e-01
4.86651182e-01 5.47221601e-01 1.65133551e-01 -1.18419027e+00
-2.86678016e-01 -9.06385720e-01 -7.71821022e-01 -3.51387143e-01
3.32534015e-01 3.50583298e-03 -4.44488645e-01 1.79759336e+00
2.20349714e-01 4.22827452e-01 -9.00717974e-02 1.29227579e+00
8.33705604e-01 7.58764684e-01 1.27033934e-01 -5.49688280e-01
1.18607032e+00 -1.04047740e+00 -7.21363723e-01 1.06792092e-01
6.77674234e-01 -8.49240780e-01 1.60290861e+00 2.88268477e-01
-7.36080527e-01 -8.48750591e-01 -1.04949939e+00 2.40986466e-01
6.64280355e-02 1.53629795e-01 3.59889448e-01 4.97830212e-01
-4.14779127e-01 8.99616480e-01 -3.70191753e-01 -1.53307602e-01
-7.50258118e-02 2.07025856e-01 6.18025512e-02 -4.30989936e-02
-1.58886218e+00 6.26085281e-01 1.53561041e-01 -3.18519175e-02
-1.29942924e-01 -9.79227066e-01 -3.77103090e-01 1.63811177e-01
1.28229231e-01 -5.48557580e-01 6.86321139e-01 -1.07960403e+00
-1.16004550e+00 2.77636480e-02 2.00214058e-01 7.16150776e-02
2.18302503e-01 -5.40631771e-01 -1.03013134e+00 -6.72697604e-01
-1.37657568e-01 -1.23202637e-01 7.59652793e-01 -1.03730857e+00
-6.85993075e-01 -5.42435110e-01 -1.37795120e-01 4.49840844e-01
-8.72727990e-01 -1.35632172e-01 -3.77298504e-01 -1.20437729e+00
3.60889025e-02 -8.53350163e-01 -2.60296464e-01 -3.69984865e-01
1.04562797e-01 -4.24723923e-01 4.32513684e-01 -5.88913977e-01
1.95451975e+00 -2.57196903e+00 2.88211137e-01 3.58920515e-01
-2.01260567e-01 2.15065077e-01 -6.38280094e-01 4.51036632e-01
-1.63723398e-02 3.14709842e-02 1.46152005e-01 2.90654540e-01
2.16215346e-02 1.19019330e-01 -6.62401676e-01 3.01476568e-01
-2.58446604e-01 6.09877765e-01 -1.11565447e+00 -1.83960516e-02
-2.71860600e-01 4.91803259e-01 -7.62248337e-01 1.87080383e-01
1.34486826e-02 2.01453373e-01 -5.17971873e-01 2.82308877e-01
9.50623214e-01 -3.03494126e-01 2.72540659e-01 -6.34248197e-01
-2.16675967e-01 3.67819399e-01 -1.56000793e+00 1.46563756e+00
-1.92004517e-01 -2.30741784e-01 -3.57406259e-01 -5.78959048e-01
1.01534307e+00 -6.50460869e-02 3.59248936e-01 -8.51356149e-01
-3.09877042e-02 5.15220881e-01 9.47127044e-02 -5.01865506e-01
8.18024695e-01 -1.72240183e-01 3.20908278e-02 6.44373596e-01
-2.56548971e-01 7.88570523e-01 -7.60171339e-02 1.25852793e-01
7.98020422e-01 1.61867831e-02 1.36696935e-01 -5.60397863e-01
5.80792725e-01 -5.31404376e-01 6.79398775e-01 4.28825080e-01
2.54863081e-03 4.47730720e-01 2.71991849e-01 -2.73875773e-01
-8.52349222e-01 -1.04687989e+00 -1.53539956e-01 1.74332869e+00
6.50732398e-01 -5.25819182e-01 -7.36036524e-02 -8.35898340e-01
3.89542580e-01 5.50169945e-01 -4.25915956e-01 -6.03282034e-01
-5.26113391e-01 -8.17628503e-01 1.72071949e-01 7.81068563e-01
-1.00236200e-01 -7.89261878e-01 1.87843338e-01 2.03926161e-01
-9.30852368e-02 -5.34440696e-01 -1.33649206e+00 -1.20067216e-01
-1.05094743e+00 -6.20025992e-01 -7.16034174e-01 -7.51689494e-01
5.17761767e-01 6.43694937e-01 7.95986891e-01 1.60914347e-01
2.77677864e-01 9.22362208e-02 -8.67239416e-01 2.66689062e-01
-2.05102768e-02 -1.34273153e-02 6.12775862e-01 2.09528357e-01
1.88815877e-01 -6.11072063e-01 -1.05082357e+00 8.46908450e-01
-1.17690217e+00 -3.96037936e-01 4.67719257e-01 9.60683107e-01
5.90929568e-01 5.18190153e-02 7.64115453e-01 -6.06516421e-01
1.17027199e+00 -9.04535115e-01 -2.15311408e-01 2.31308922e-01
-9.94863391e-01 2.45166242e-01 9.55118895e-01 -8.01516116e-01
-7.74940491e-01 -4.30127382e-01 -2.55305976e-01 -4.04695034e-01
3.99836630e-01 7.66973972e-01 -1.39296040e-01 3.30655873e-01
5.66797018e-01 2.21373856e-01 -1.19329289e-01 -9.36470151e-01
4.71256852e-01 6.84457481e-01 1.59820840e-01 -3.50399643e-01
5.93377173e-01 5.41667417e-02 -4.45122987e-01 -5.39778590e-01
-8.81042361e-01 -7.63368189e-01 -1.41742751e-01 1.44034177e-01
2.10505664e-01 -7.51290977e-01 -4.38356698e-01 7.49367028e-02
-8.76474261e-01 1.88096955e-01 -2.24379942e-01 6.84339702e-01
-1.43725201e-01 7.96573102e-01 -6.23467386e-01 -6.77603781e-01
-5.36224067e-01 -1.01596785e+00 5.30915856e-01 1.60580650e-01
-1.96770534e-01 -8.41283917e-01 3.30039799e-01 2.02298071e-02
4.13059682e-01 -2.75048733e-01 1.07883835e+00 -7.17881382e-01
3.27168107e-02 1.38747543e-01 -3.11789423e-01 3.99709553e-01
5.55894524e-02 -2.04948649e-01 -4.11830246e-01 -6.16041243e-01
-3.62843797e-02 1.33448780e-01 8.25963259e-01 2.30642468e-01
9.88963962e-01 -3.75330508e-01 6.05301596e-02 4.87140149e-01
1.46857965e+00 1.93269566e-01 6.05791688e-01 2.27503821e-01
8.92512798e-01 4.27741408e-01 9.67302561e-01 7.99941421e-01
-3.85034643e-02 9.60517168e-01 1.70985281e-01 3.44343185e-01
-2.66518700e-03 -4.95691210e-01 7.29598343e-01 1.50411558e+00
-7.98308030e-02 -2.68101633e-01 -4.84961659e-01 2.57580370e-01
-2.09307623e+00 -8.07801485e-01 -4.40391570e-01 2.53300595e+00
7.55226016e-01 -1.45169675e-01 3.54023933e-01 1.77786708e-01
7.53527045e-01 -1.63874924e-02 -5.33760667e-01 -4.95483816e-01
-1.10322133e-01 -2.11757287e-01 4.01929349e-01 2.46631920e-01
-7.42229760e-01 6.20650232e-01 5.57586813e+00 1.18336642e+00
-1.01665521e+00 8.76297802e-02 2.37895325e-01 -2.10948348e-01
-8.44428897e-01 -3.26800644e-01 -9.04600799e-01 1.07123280e+00
7.74237454e-01 -2.12036848e-01 5.33916831e-01 7.82175362e-01
2.39649400e-01 5.96871614e-01 -8.78134787e-01 1.09512055e+00
8.00737590e-02 -9.65245008e-01 4.32865590e-01 1.04872070e-01
7.52441704e-01 -3.25930603e-02 4.90135700e-01 3.82511050e-01
1.71212062e-01 -7.57734060e-01 5.72481334e-01 4.51883525e-01
7.09097028e-01 -8.68105590e-01 7.18329549e-01 1.10998839e-01
-1.41893291e+00 -6.23866320e-01 -6.99476063e-01 7.97037706e-02
2.57767558e-01 8.62540543e-01 -1.97386622e-01 4.88557786e-01
4.70013529e-01 9.44288313e-01 -3.50912750e-01 1.26487303e+00
1.27759770e-01 5.12801349e-01 -9.52428579e-02 -2.78419126e-02
1.24607638e-01 -7.03007102e-01 6.00527227e-01 1.11144519e+00
7.48422980e-01 1.18702121e-01 1.63021907e-01 4.85945225e-01
6.05617613e-02 6.49130642e-01 -3.24889161e-02 -1.63714513e-01
6.23074353e-01 1.00218987e+00 -2.91855335e-01 2.34666951e-02
-4.77099866e-01 9.69871819e-01 2.30887815e-01 3.26729834e-01
-8.61015677e-01 -1.87909856e-01 1.00328457e+00 2.17299923e-01
7.46050477e-01 -1.40734643e-01 -4.51628685e-01 -1.23981321e+00
5.74365407e-02 -1.05706954e+00 5.15002370e-01 -1.14505596e-01
-1.85959864e+00 3.68205130e-01 -4.70266759e-01 -1.76753914e+00
2.28820488e-01 -3.06348234e-01 -6.17896318e-01 6.80190265e-01
-1.30153573e+00 -6.30496085e-01 -3.14390734e-02 5.44201851e-01
5.24315894e-01 -1.70300379e-01 5.06168962e-01 8.31066072e-01
-6.99306905e-01 1.09951186e+00 7.22450793e-01 -4.29700434e-01
7.26360977e-01 -7.87172377e-01 1.94295242e-01 4.97675538e-01
-5.08232377e-02 1.16962552e+00 7.16371775e-01 -6.88955188e-01
-1.64026988e+00 -1.07765424e+00 5.66689610e-01 -3.48718911e-01
9.15098011e-01 2.20167916e-03 -1.08556545e+00 9.37750190e-03
-3.77526462e-01 -1.96653038e-01 1.13478255e+00 4.16266620e-01
-7.20660448e-01 -2.09839314e-01 -7.72163033e-01 6.06934726e-01
1.31881237e+00 -3.39852214e-01 -5.44617534e-01 -1.15310751e-01
9.95578408e-01 2.48861626e-01 -1.06825781e+00 6.10787272e-01
9.80227113e-01 -8.47526014e-01 1.07221961e+00 -7.01907218e-01
5.15081584e-01 -4.90951389e-01 -4.38572586e-01 -1.39751303e+00
-1.11624551e+00 -3.25807214e-01 -3.91959012e-01 1.25609493e+00
4.35654312e-01 -3.88159215e-01 3.47640783e-01 1.96254462e-01
-6.80152327e-02 -8.04472387e-01 -5.36041558e-01 -1.06257367e+00
1.07486844e-01 -6.59686476e-02 7.76243448e-01 1.02592456e+00
2.12662190e-01 6.00598574e-01 -6.82502151e-01 -3.29691842e-02
3.38291645e-01 4.38037843e-01 3.28358144e-01 -1.27975750e+00
-4.72606957e-01 -5.88316381e-01 -9.70118791e-02 -1.42600977e+00
-3.52010220e-01 -9.60783780e-01 -1.89133823e-01 -1.17372954e+00
2.80785352e-01 -5.67455471e-01 -1.17989457e+00 -1.36405766e-01
-5.78992724e-01 2.37375847e-03 1.31714925e-01 6.68717623e-01
-6.97248757e-01 7.88000822e-01 1.28378296e+00 2.90682495e-01
-5.16195893e-01 -6.70734420e-02 -1.02740288e+00 4.09521252e-01
6.70325339e-01 -1.88165769e-01 -6.94294631e-01 -4.56991076e-01
6.75588489e-01 -2.24147648e-01 -2.44351998e-01 -6.98056698e-01
-4.03061695e-02 -2.27272585e-01 1.50793031e-01 -4.62019354e-01
-1.43578127e-01 -8.10849488e-01 1.74602896e-01 4.33901876e-01
-5.18910706e-01 4.79828492e-02 -3.76884006e-02 9.68373537e-01
3.10307145e-02 -1.16401361e-02 7.22586393e-01 2.86189646e-01
-7.42671251e-01 5.07056296e-01 -6.93561286e-02 3.87127697e-03
6.55047238e-01 -1.86671019e-01 -1.52591720e-01 -1.10605165e-01
-6.35131896e-01 2.98376709e-01 3.94410402e-01 9.53751087e-01
7.51166165e-01 -1.96842194e+00 -6.89408600e-01 4.38252151e-01
3.01676065e-01 -8.77106071e-01 5.20708144e-01 8.35573137e-01
1.21140078e-01 5.92912883e-02 -7.26450235e-02 -1.46028101e-01
-1.18490171e+00 7.44680822e-01 -1.51332885e-01 -3.03313464e-01
-3.87346298e-01 9.20278728e-01 3.35923761e-01 -3.47261816e-01
3.43300790e-01 -1.63908765e-01 -7.74075329e-01 3.79123598e-01
7.18329906e-01 6.85041904e-01 -1.64514109e-01 -5.36145926e-01
-3.27924967e-01 6.86795056e-01 -1.96868658e-01 2.69525170e-01
1.17839348e+00 -4.04421985e-01 -1.76675785e-02 4.32289183e-01
1.24589658e+00 1.14824012e-01 -9.99568343e-01 -4.21315759e-01
1.07664997e-02 -6.49101257e-01 -7.73997465e-03 -6.00049734e-01
-1.04158175e+00 5.48591077e-01 8.83866429e-01 2.65778869e-01
9.12001789e-01 -3.02313119e-01 1.39405394e+00 -9.26664919e-02
1.78893909e-01 -1.23792803e+00 2.09461048e-01 7.93423891e-01
9.54180121e-01 -9.67490494e-01 4.54529375e-02 -3.82063270e-01
-1.00807130e+00 8.92875314e-01 5.03505528e-01 -2.49814108e-01
8.58214438e-01 -9.95073989e-02 -9.55977663e-02 1.49397910e-01
-5.67101359e-01 1.15681272e-02 6.70400381e-01 2.12807566e-01
7.03983366e-01 2.83390492e-01 -1.01976097e+00 1.43376422e+00
-3.93207557e-02 -2.91880742e-02 8.13870803e-02 4.53525037e-01
-6.56279385e-01 -1.37379718e+00 -1.47647530e-01 6.30937099e-01
-2.63884634e-01 -3.79396826e-01 -2.27259055e-01 -8.27475563e-02
2.43064508e-01 9.06013012e-01 2.08088439e-02 -9.93948519e-01
5.04762590e-01 -1.42678708e-01 8.12651664e-02 -3.80069077e-01
-5.95662773e-01 1.96362033e-01 -1.23535126e-01 -5.57743013e-01
7.33793005e-02 -4.75496799e-01 -1.23494387e+00 -3.07992667e-01
-5.65323114e-01 3.55441242e-01 3.40973884e-01 6.21253371e-01
6.28015339e-01 2.90886581e-01 1.10025668e+00 -3.84708256e-01
-1.14708388e+00 -8.41045916e-01 -9.37986612e-01 9.47297513e-01
1.13118120e-01 -7.62757897e-01 -4.86813694e-01 -4.86168504e-01] | [10.104257583618164, 5.567850589752197] |
aab895ef-c26d-4fb3-8479-2743537ac9e4 | markerless-3d-human-pose-tracking-through | 2303.18119 | null | https://arxiv.org/abs/2303.18119v1 | https://arxiv.org/pdf/2303.18119v1.pdf | Markerless 3D human pose tracking through multiple cameras and AI: Enabling high accuracy, robustness, and real-time performance | Tracking 3D human motion in real-time is crucial for numerous applications across many fields. Traditional approaches involve attaching artificial fiducial objects or sensors to the body, limiting their usability and comfort-of-use and consequently narrowing their application fields. Recent advances in Artificial Intelligence (AI) have allowed for markerless solutions. However, most of these methods operate in 2D, while those providing 3D solutions compromise accuracy and real-time performance. To address this challenge and unlock the potential of visual pose estimation methods in real-world scenarios, we propose a markerless framework that combines multi-camera views and 2D AI-based pose estimation methods to track 3D human motion. Our approach integrates a Weighted Least Square (WLS) algorithm that computes 3D human motion from multiple 2D pose estimations provided by an AI-driven method. The method is integrated within the Open-VICO framework allowing simulation and real-world execution. Several experiments have been conducted, which have shown high accuracy and real-time performance, demonstrating the high level of readiness for real-world applications and the potential to revolutionize human motion capture. | ['Arash Ajoudani', 'Elena De Momi', 'Juan M. Gandarias', 'Mattia Leonori', 'Luca Fortini'] | 2023-03-31 | null | null | null | null | ['pose-tracking', '3d-human-pose-tracking'] | ['computer-vision', 'computer-vision'] | [-0.05514497 -0.25025064 -0.15244909 0.20987605 -0.32146654 -0.5213676
0.6187867 -0.16022551 -0.6487709 0.468544 -0.02291054 -0.15729976
0.19232412 -0.33902574 -0.19392845 -0.30752617 -0.2722912 0.47858772
0.66712666 -0.22509179 0.032602 0.90185523 -1.6144843 -0.3087157
0.4470013 0.7932389 0.03262857 0.97878003 0.4335892 0.35072145
-0.6071926 0.10010119 0.5097198 -0.23175672 -0.18305026 0.03043766
0.62393993 -0.3631134 -0.13519311 0.57469904 0.8295292 0.05174737
0.2589532 -1.5236114 0.28553078 -0.57228714 -0.59587985 -0.14155997
1.2000911 0.32539725 0.0778842 -0.5563992 0.8476884 1.0908642
1.0982223 0.70688826 -0.8894694 -0.20373188 -0.36392462 -0.23840204
-1.3972338 -0.2835416 0.7262115 -0.65260667 0.9163815 0.63883454
1.1893007 0.80580974 0.53012747 0.556072 0.8065327 -0.6503658
0.16857384 0.12855703 -0.35286278 0.7443634 0.68054914 0.38026947
-0.30175057 -0.19826774 1.1398973 0.23988178 -0.39741188 -1.1199555
-1.6182312 0.40198186 0.20143719 0.3188578 -0.5061427 0.4139568
0.33530408 -0.23540525 0.17137083 0.22264792 -0.11275624 -0.43533686
-1.0191188 0.53154707 0.675127 0.86710006 0.2541558 -0.06690885
0.08914654 0.09840555 0.51602924 0.8417048 0.5058959 -1.0632241
0.11713145 0.7582314 0.67267996 -1.3282677 -0.7974081 0.11041095
-0.49438 0.80413884 0.5998548 -0.29885703 -0.6829378 1.2434046
0.92205083 -0.09136247 -0.30020875 1.3582428 0.36795497 0.35753575
0.02541739 0.03573564 1.262384 -0.7497864 -0.8274457 -0.49114484
0.5895599 -0.76992464 0.6925089 0.54715 -1.0916091 -0.6946446
-1.2919432 0.13351275 -0.2039061 0.12105237 0.5548991 1.0181019
-0.88248825 0.4490008 -1.0740829 -0.725474 -0.28579128 0.588173
-0.669841 0.25045577 -0.98698914 1.3379813 0.35715833 0.29028487
-0.40624267 -0.39064392 -0.9811175 -0.562942 0.35063007 -1.0402365
1.1262792 -0.5151242 -1.7165865 0.99215925 0.12530796 -0.34446406
1.032 -0.69414276 -0.36914456 0.36202556 -0.11251224 0.641821
0.6792537 -1.237636 -0.4655908 -0.4214115 -0.15558521 0.35931662
-0.09015752 -0.24486981 -0.57091445 -0.53338534 0.15110426 -1.3428967
-0.42306373 0.56593865 0.08676228 0.35569203 0.9253099 -0.64774334
1.274996 -1.7857381 0.09708092 0.12357482 0.01455309 0.6899997
0.5044295 0.3833009 0.2509357 -0.45079413 0.16572645 -0.24508318
-0.13341038 0.07517037 0.24865785 0.73916906 -0.2957095 0.74187165
-0.9878525 -0.6295903 0.96618754 0.7074235 -0.08652017 0.32134128
0.06814554 0.53600425 -0.40152326 0.6493473 0.5059101 0.09103573
0.21640986 -0.08616552 -0.21386519 -0.36264756 -1.743114 1.949689
-0.3643746 0.44637606 0.22058004 -0.31631425 0.8466184 0.52344084
0.7840451 -0.535951 0.24393328 0.38755232 -0.41640693 -0.6538726
0.8273496 -0.03822415 -0.20028926 0.11946476 -0.26620418 -0.36432415
-0.23735474 -0.10284919 0.91993904 0.9041695 0.85096383 -0.06088797
0.6219099 0.4002353 0.18967055 0.24761008 -0.6174021 0.7658081
-0.2963552 -0.7082045 -1.1104598 -0.9602518 0.39586076 0.3398003
0.45936492 -0.24475831 -0.7802917 -0.27953202 0.11815716 0.21841708
-0.5119913 -0.04840842 -0.8116242 -0.1196164 0.41455427 0.5343923
0.18922651 -0.6693278 -2.0172994 0.34799427 0.11002116 -1.2289512
-0.33090684 -0.41174108 -0.9407519 -1.1098573 -1.1044376 -0.20449191
0.5657563 0.35969797 0.9478046 -0.01046483 -0.68478906 0.99052435
-0.25377944 -0.4765193 -0.28419554 -0.43410328 0.31298202 -0.26835212
0.05507946 -0.09558836 -0.7826647 0.57884836 -0.55586046 0.07336513
0.27188393 0.2892943 0.31918254 -0.6197396 -0.01477922 -0.20460887
0.43652225 -0.05750225 -0.859976 0.13081789 -0.3304883 -0.35719424
0.16098794 -0.65415204 -0.68665785 0.60595995 0.13257256 -0.59961087
-0.2723621 0.11049477 0.07217754 -0.16313991 0.81557924 -0.24664038
0.5696079 -0.07033177 0.29465383 0.5951511 0.92256397 -0.1637381
0.7316844 0.80177045 0.26507586 -0.82723826 -0.24536347 -0.8301234
-0.97572505 -0.79320925 0.9207757 -0.93457186 -0.97084314 0.39898634
-1.1164417 -0.14732812 -0.17700635 0.8091612 -0.68070495 0.5852943
0.05256097 -1.2636 -0.45572293 -1.1449499 1.2969843 0.3698877
-0.83111304 -0.98631436 0.40871194 0.23274937 0.35495842 0.9952188
-0.14082277 -0.02357231 -0.31129828 -0.8285968 0.3729252 -0.31026715
0.00837118 -0.17192328 -0.8360159 -0.4151814 -0.17671378 0.02676391
-0.16692358 0.44184417 0.0689465 0.16089344 -0.7749447 0.27655208
1.4819269 0.24311408 0.61280525 0.64524895 0.72678506 0.4185556
1.1132821 0.53139186 0.258428 1.3727154 0.45303702 -0.06919181
-0.1244849 -0.1575666 0.48796663 0.35649797 -0.34402236 0.1389011
-1.1970155 0.5191414 -1.9821748 -0.81791496 -0.45953482 2.7544527
0.29990616 0.128969 0.54214865 0.26886928 0.6100018 -0.20677282
-0.14502223 -0.44191605 0.5855024 -0.11357631 0.5753764 0.47681183
-1.0434843 0.556218 6.4320297 0.28719774 -1.2721297 0.02040612
-0.20422387 -0.09283496 0.22783394 -0.1220189 -0.71283895 0.31283012
0.8581718 -0.08742149 0.00733491 0.98095655 0.38099045 -0.63222003
-0.8606104 1.0449079 0.27239755 -1.0357054 -0.48976424 0.17469521
0.36866105 -0.32472035 -0.4306542 -0.15111898 -0.18795836 -0.6826298
0.85524076 0.66582334 0.7857526 -0.5972905 0.6462425 0.78120595
-1.283463 0.29849654 0.08666044 -0.26507163 0.5729786 0.27917618
-0.9798313 0.71874785 0.6068748 0.1531648 -0.5096596 1.322301
0.04840508 -0.14211349 -0.63967365 -0.22910559 0.03469264 -0.15709457
0.7454389 1.0813075 0.42500564 -0.15453486 0.33616018 0.3685764
0.86017287 -0.19755252 -0.97444594 0.3289032 0.17089716 1.3435293
-0.8664751 -0.21400118 -0.14299314 1.1373378 -0.2925155 -0.19434087
-1.049593 -0.29063335 0.45104325 0.5507161 0.01976408 -0.73183805
-0.22291112 -0.97213614 0.20752884 -0.6930326 0.28622314 -0.8908617
-0.48606786 0.49412844 0.30914813 -1.9160103 -0.7811126 -0.6380825
-0.23247528 0.70384884 -0.79741406 -1.3903273 -0.57686126 0.42810425
0.26105753 0.1555348 0.9324403 0.36531654 -0.06576258 0.2399574
-0.3930913 -0.14707717 0.7763908 -0.97653073 0.4539243 0.8560038
-0.08522677 0.5589291 0.8651588 -0.8405875 -1.9019626 -0.47299626
0.69750404 -0.931094 0.17630148 -0.3017627 -0.4594585 0.5067615
-0.02561954 0.45110124 0.24554609 -0.5140493 0.2387514 0.09164713
-1.426374 0.7135973 0.78304225 -0.0696002 -0.68503714 -0.15117712
0.19075271 -0.9581156 -0.8535756 0.51955223 1.071473 -1.1041014
1.2162174 -0.04115773 -0.32550362 -0.78930426 0.20821008 -1.0550771
-0.0665689 -0.96225214 -0.43851325 0.5897722 -0.11161305 -0.3682708
0.89407927 0.8141127 0.2040246 -0.39692003 -1.1548263 -0.81455404
-0.59773487 -0.5281791 0.13202819 0.6738798 0.19057448 0.02266205
-0.5635704 0.16489513 0.63770026 -0.14068124 1.4141306 -1.1936238
-0.02388194 -0.18183899 -1.054113 -0.59753287 -0.48727688 -0.23163365
0.04670716 -1.5444683 -0.2839259 -0.1637477 0.421701 0.25932387
-0.14447404 0.4058566 0.44104323 0.10114333 -0.4914562 -0.06799871
0.9754255 0.41281667 -0.18314254 0.02420489 0.20692752 0.9442256
0.4348118 -0.39065677 -0.25720677 -0.1662103 0.06524558 0.256949
0.72816294 -1.7329923 0.5086754 -0.0120197 0.5595907 -0.7145944
0.56383675 -1.4551448 0.73025167 0.98144454 0.3865525 0.40080005
0.37221044 0.3545605 0.02411915 -0.02910612 0.646582 -0.24728948
-0.89084786 0.07071276 -0.41115287 -0.21097997 1.7028193 -1.0221555
0.17965837 -0.4199532 -0.63274145 -0.06532896 0.9054411 0.5268569
0.68774176 -1.4257526 -0.2316942 0.3056423 0.12423453 -0.08281247
0.00904469 0.87111163 -1.1768457 0.5409063 -0.5294009 -1.0416384
-1.4430724 0.6719832 0.42771307 -0.07942791 -0.71181816 0.11376458
-0.4485326 -0.37554106 0.13481538 -0.0767833 -0.00847326 -0.1800783
0.46830696 0.78465503 0.0694696 -0.8207867 -0.60458493 1.043354
0.6052524 -0.5804354 0.94895476 -0.20271604 0.34233105 0.4808847
0.84610367 0.23665935 -1.4027585 0.33615026 0.21609728 -0.72212183
-0.20817585 -0.6844978 -0.59441054 0.68868995 0.9960382 -0.02118802
0.89609176 -0.5253572 0.72577786 0.02468731 0.85930043 -1.0994936
-0.14979883 0.00686852 0.7205974 -1.1065733 0.4045651 -0.4263443
-0.6033297 1.2211992 0.5382039 -0.14096196 0.34917668 0.67190605
0.4591195 -0.13662294 -0.2132575 -0.1196847 0.46939096 0.7809612
0.45855227 0.05377656 -0.47534716 -0.01919697 0.1976927 0.57550085
0.3428562 1.5085018 -0.30094504 -1.0436589 -0.99666095 -0.39182255
-0.33032206 0.69397277 -0.25217956 1.2591454 0.09327915 0.7388079
-0.16995013 -0.22222112 0.6297706 -0.07579848 0.78974867 -0.29681963
-0.8645162 0.2881593 0.12003534 -0.7945354 -0.47235185 -0.67583835
-1.1882281 -0.07515475 -0.2995398 -0.09991844 0.99017286 0.5430374
0.5027425 0.13954528 0.16449751 -1.5645792 -0.4224415 -0.65188706
-0.16311291 0.49512744 0.3449054 -0.8375641 0.12582578 0.10930613] | [7.272627353668213, -0.9229164719581604] |
5e9479b2-d4d9-415c-a285-9ae915ce9185 | deep-multi-view-learning-for-tire | 2203.12451 | null | https://arxiv.org/abs/2203.12451v1 | https://arxiv.org/pdf/2203.12451v1.pdf | Deep Multi-View Learning for Tire Recommendation | We are constantly using recommender systems, often without even noticing. They build a profile of our person in order to recommend the content we will most likely be interested in. The data representing the users, their interactions with the system or the products may come from different sources and be of a various nature. Our goal is to use a multi-view learning approach to improve our recommender system and improve its capacity to manage multi-view data. We propose a comparative study between several state-of-the-art multi-view models applied to our industrial data. Our study demonstrates the relevance of using multi-view learning within recommender systems. | ['Bruno Canitia', 'Khalid Benabdeslem', 'Kilian Bourhis', 'Thomas Ranvier'] | 2022-03-23 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [-4.00534123e-01 -2.08855629e-01 -2.36645162e-01 -5.96096814e-01
-2.19887376e-01 -6.92476869e-01 6.82747722e-01 2.79150102e-02
2.74491370e-01 2.35270619e-01 6.33190334e-01 2.52105165e-02
-4.54598904e-01 -8.21131229e-01 -1.92706391e-01 -1.57875076e-01
1.40211850e-01 9.64977086e-01 3.31051111e-01 -8.05644155e-01
4.99097437e-01 4.51067865e-01 -2.07261848e+00 1.00659192e+00
2.59009182e-01 6.25389576e-01 3.86639357e-01 7.86025465e-01
-1.03710026e-01 8.67206752e-01 -4.49675083e-01 -5.53626120e-01
2.52792716e-01 1.47937080e-02 -8.41248572e-01 2.82800198e-01
4.87962216e-01 -3.72161031e-01 -2.62409542e-02 5.59059680e-01
3.86425257e-01 3.03429842e-01 6.51909113e-01 -1.12009788e+00
-4.78173375e-01 5.92845380e-01 -1.32950217e-01 2.00813264e-01
9.58344221e-01 -5.68038523e-01 1.12720895e+00 -9.08823371e-01
9.38249111e-01 1.24391723e+00 4.11892116e-01 4.97006178e-01
-7.50008225e-01 -7.78818727e-02 5.17616451e-01 6.45203829e-01
-8.65761638e-01 -4.14833546e-01 6.69804573e-01 -6.61579967e-01
7.58699417e-01 4.96180952e-01 6.00148559e-01 1.11456919e+00
2.73623168e-01 8.03961515e-01 9.27556932e-01 -3.67695630e-01
1.31035289e-02 7.39791393e-01 1.89692825e-01 4.50613052e-01
-2.14964658e-01 -1.34598017e-01 -4.65323746e-01 -5.14885902e-01
5.96953988e-01 6.60357356e-01 4.63272706e-02 -7.05468059e-01
-7.62687564e-01 7.73335218e-01 -1.86184525e-01 4.45493937e-01
-3.20653379e-01 -4.98864055e-01 3.69281709e-01 6.57325029e-01
2.89485782e-01 4.55648452e-01 -6.58754110e-01 -3.29045981e-01
-7.05233157e-01 3.37244183e-01 1.31375647e+00 1.06628370e+00
5.98326325e-01 -4.79480684e-01 2.54451036e-01 1.07579303e+00
5.74772120e-01 2.06307620e-01 3.89394611e-01 -9.59152579e-01
2.76340216e-01 8.20903838e-01 3.18575650e-01 -7.29173779e-01
-1.47096023e-01 -2.29504764e-01 -2.76640981e-01 2.88962185e-01
2.21319214e-01 -2.36270949e-02 -3.87114674e-01 9.18577075e-01
1.05222195e-01 -1.45082891e-01 1.36068324e-02 7.92685151e-01
9.02958512e-01 4.13073957e-01 -3.60386699e-01 -3.92084956e-01
1.14521623e+00 -1.04309821e+00 -5.28445780e-01 4.13005710e-01
5.57956696e-01 -1.19127679e+00 7.69341230e-01 1.08359408e+00
-1.22169876e+00 -9.67854977e-01 -1.00089276e+00 3.28307629e-01
-5.93052447e-01 3.89028877e-01 4.40523833e-01 6.91922784e-01
-6.84509218e-01 8.63506675e-01 -4.76456881e-01 -6.78407192e-01
-2.77206093e-01 4.24841970e-01 -2.21394181e-01 -1.35544196e-01
-9.45139587e-01 1.03670609e+00 -1.35783494e-01 -3.61815035e-01
-9.05976534e-01 -6.11354768e-01 -2.41384685e-01 6.27820119e-02
6.16040826e-01 -7.97036171e-01 1.19922161e+00 -7.95812905e-01
-1.32988882e+00 3.67476404e-01 4.40424532e-02 -7.79095739e-02
2.63016164e-01 -4.65729326e-01 -1.03420329e+00 -2.71976531e-01
-2.85107493e-01 -4.03378367e-01 7.62884915e-01 -1.45827699e+00
-1.11661696e+00 -4.81162697e-01 6.82712376e-01 3.87519002e-01
-3.02691072e-01 5.14432304e-02 -5.35873473e-01 -2.45402113e-01
-3.23570013e-01 -1.01572180e+00 -3.56464505e-01 -6.84323072e-01
4.75948825e-02 -2.58608460e-01 1.08043742e+00 -3.36087346e-01
1.38541341e+00 -1.75740921e+00 4.47458684e-01 3.64350319e-01
1.33380443e-01 3.12648296e-01 1.82285249e-01 1.09038007e+00
3.40330303e-02 5.82196414e-02 8.53887796e-01 -2.91051686e-01
-1.66207939e-01 2.38831878e-01 -1.95470214e-01 1.84725046e-01
-7.39836812e-01 5.89952692e-02 -7.90922940e-01 -3.30841303e-01
4.00100440e-01 4.06129718e-01 -5.94051719e-01 5.83814740e-01
-9.45628062e-02 1.82506159e-01 -7.19279289e-01 3.89225096e-01
2.16902852e-01 -3.80683631e-01 5.62104464e-01 -2.61334062e-01
-1.53422445e-01 1.31661445e-01 -1.60857332e+00 1.54514098e+00
-9.43917036e-01 5.76274469e-02 -2.09102910e-02 -6.68951392e-01
7.80365467e-01 3.59393895e-01 8.89298439e-01 -2.20235303e-01
-2.44109910e-02 -1.70462251e-01 1.54851794e-01 -7.34083116e-01
7.23574281e-01 1.56257391e-01 2.06789091e-01 8.83746028e-01
2.01378837e-01 6.02052450e-01 3.27504128e-01 3.67709786e-01
7.06437469e-01 2.59315550e-01 5.13918519e-01 -1.47885114e-01
1.03728676e+00 -2.84087956e-01 2.33780444e-01 5.30581713e-01
4.82574403e-01 3.38316411e-01 3.91872406e-01 -8.64318252e-01
-9.49555457e-01 -6.94453180e-01 7.73377717e-02 1.64030576e+00
-1.63783893e-01 -1.01233339e+00 -2.66053259e-01 -1.00095487e+00
-9.18740779e-02 7.60312438e-01 -5.60594440e-01 2.50847876e-01
-2.08030030e-01 -6.60053492e-02 -3.93334150e-01 3.62365067e-01
-2.40486830e-01 -7.96734273e-01 -5.28134406e-01 3.89327347e-01
1.26811922e-01 -7.60861337e-01 -4.90216076e-01 -1.64146334e-01
-8.77125502e-01 -1.14722896e+00 -2.84354031e-01 -3.32394481e-01
5.29152691e-01 4.89536285e-01 1.50899243e+00 6.21878132e-02
9.94204283e-02 9.97092545e-01 -9.67625260e-01 -4.58827317e-01
-6.24878049e-01 1.06462743e-02 3.76932442e-01 1.21735342e-01
4.37357903e-01 -7.37424672e-01 -5.48071504e-01 8.54820907e-01
-4.89850253e-01 -3.24711919e-01 2.96453893e-01 5.56075096e-01
5.14591217e-01 2.05506191e-01 2.40203708e-01 -1.90361369e+00
7.50102878e-01 -5.98140121e-01 -3.71982813e-01 6.30445361e-01
-9.30226505e-01 -1.11999497e-01 7.11430013e-01 -4.00703043e-01
-1.08107448e+00 7.78358150e-03 -5.64424247e-02 -5.55484891e-01
-5.19533753e-01 5.51681042e-01 -2.55643100e-01 5.98839186e-02
5.53503036e-01 -1.86824817e-02 -1.63303971e-01 -8.17819238e-01
6.35012925e-01 8.58831048e-01 -6.51248246e-02 -4.29477394e-01
3.19010913e-01 3.69069576e-01 -1.24933183e-01 -6.82271123e-01
-8.40542436e-01 -1.07364631e+00 -1.06550932e+00 -5.99772274e-01
3.93930793e-01 -7.10375428e-01 -7.97090054e-01 -9.39368457e-02
-7.11140990e-01 2.88196921e-01 -1.25244498e-01 3.53955686e-01
-6.94009781e-01 2.89701909e-01 -3.40771735e-01 -7.86028743e-01
-2.06790939e-01 -1.03753662e+00 5.73903441e-01 -1.78721547e-02
-2.74145007e-01 -1.24712467e+00 3.15430939e-01 6.75622821e-01
4.09024417e-01 -3.06425810e-01 8.11401844e-01 -1.23726416e+00
-3.11127722e-01 -2.20218569e-01 3.80858779e-01 1.65528089e-01
4.20685649e-01 1.85828522e-01 -7.42790163e-01 -4.27586257e-01
-2.20216773e-02 8.38364437e-02 3.11655968e-01 1.06206648e-01
1.06773019e+00 -1.39541179e-01 -5.20759642e-01 1.22925177e-01
1.51853800e+00 3.02658558e-01 4.63057935e-01 1.60939068e-01
7.48449385e-01 8.31252098e-01 9.55765367e-01 8.85715663e-01
4.04750079e-01 1.13410640e+00 4.33234066e-01 2.28329584e-01
1.83987439e-01 -1.96218088e-01 3.43198329e-01 1.04556990e+00
-5.92173219e-01 -3.91254246e-01 -5.22303164e-01 2.98111498e-01
-2.19229269e+00 -1.38048863e+00 -2.66005844e-01 2.27771330e+00
-1.00713670e-02 1.83394909e-01 7.56564438e-01 -1.37650415e-01
3.63977492e-01 -6.89519420e-02 -3.98490340e-01 -8.57613921e-01
6.86726689e-01 -2.94534504e-01 2.37925023e-01 2.26807356e-01
-1.05132556e+00 4.76573974e-01 7.00250721e+00 2.60141015e-01
-7.25635290e-01 3.36809047e-02 1.69978946e-01 -3.31193209e-01
-3.44138265e-01 3.48933712e-02 -1.07329845e+00 1.84866771e-01
1.32023120e+00 -3.54671061e-01 3.99961054e-01 1.13853300e+00
9.36690420e-02 2.86199227e-02 -1.43283236e+00 1.04477513e+00
6.36673212e-01 -1.30159724e+00 1.06935382e-01 5.57910085e-01
7.49029756e-01 7.90327117e-02 7.65107274e-02 9.96699259e-02
4.99980688e-01 -4.90987271e-01 2.00175792e-01 1.10580254e+00
9.49074477e-02 -9.66461957e-01 7.76185989e-01 7.39485562e-01
-1.21373153e+00 -3.12345684e-01 -5.09192467e-01 -3.21579352e-02
2.18362749e-01 2.70875067e-01 -8.92392755e-01 8.83113503e-01
5.81332564e-01 1.06294489e+00 -4.68798876e-01 8.88155580e-01
3.48072410e-01 1.47211254e-01 3.01494151e-01 -7.66918957e-02
-2.47235417e-01 -4.57147479e-01 3.71930420e-01 9.15622830e-01
5.18001795e-01 2.17316657e-01 6.06044888e-01 1.30616054e-01
3.40920806e-01 5.21230102e-01 -9.11273301e-01 2.76104033e-01
1.62272915e-01 1.53235269e+00 -2.95376629e-01 -4.33803499e-01
-1.14406145e+00 6.96924806e-01 1.05962202e-01 -5.78133091e-02
-3.10218066e-01 1.17788360e-01 6.62456214e-01 5.99953532e-01
5.82951486e-01 2.54615713e-02 4.04794931e-01 -1.24844289e+00
-3.42640847e-01 -1.20186913e+00 5.24717748e-01 -5.57393849e-01
-1.71777427e+00 8.05895209e-01 1.41517490e-01 -1.80025184e+00
-6.20171726e-01 -5.51339388e-01 -4.60581809e-01 8.27515304e-01
-9.87081409e-01 -1.26596832e+00 7.76380971e-02 7.66079962e-01
8.45013440e-01 -1.05696809e+00 1.03244483e+00 4.15612876e-01
-3.07920668e-02 2.23047435e-01 2.52730697e-01 -4.58642423e-01
1.06626129e+00 -1.43568254e+00 1.27208650e-01 2.38190383e-01
5.75391889e-01 1.00237036e+00 7.46942997e-01 -4.72378910e-01
-1.59641433e+00 -5.55622816e-01 6.19452477e-01 -9.00032043e-01
6.25192106e-01 -1.46692857e-01 -5.19347548e-01 6.87937677e-01
3.33927870e-01 -2.68856972e-01 1.49043059e+00 9.44396734e-01
-2.30836421e-01 -1.76098362e-01 -1.01373422e+00 2.39328250e-01
8.03354383e-01 -3.72329831e-01 -8.27021539e-01 3.60777795e-01
-3.77679728e-02 -1.61224395e-01 -1.45006657e+00 6.65279627e-02
9.17737246e-01 -1.38059282e+00 1.08470178e+00 -9.86533880e-01
4.03939486e-01 -1.16894528e-01 -4.24700409e-01 -1.59199131e+00
-8.99304748e-01 -2.56581992e-01 -5.16969204e-01 1.34592819e+00
4.07795250e-01 -1.88549668e-01 9.56900716e-01 5.18406451e-01
8.49080384e-02 -7.18639493e-01 -3.28692913e-01 -4.64449883e-01
-4.44111228e-01 -1.81130484e-01 6.41684055e-01 6.41071141e-01
3.85045707e-01 5.75332761e-01 -9.86149669e-01 1.51594996e-01
5.26136160e-01 7.44011760e-01 9.43303525e-01 -1.80239606e+00
-7.48153031e-01 -6.07160106e-02 -2.31347889e-01 -8.11412811e-01
-2.40159750e-01 -6.81714356e-01 -5.81167936e-01 -1.70538807e+00
1.77480683e-01 -2.91069686e-01 -6.25461578e-01 -6.62113503e-02
2.62948036e-01 -3.14958751e-01 2.95230120e-01 5.29908538e-01
-8.78889441e-01 1.84763186e-02 1.25177026e+00 1.75212860e-01
-3.31742197e-01 8.43685150e-01 -8.84235680e-01 1.08021295e+00
4.36858535e-01 -2.68401027e-01 -8.70447457e-01 5.34046292e-02
4.75057840e-01 3.94061804e-01 -4.72132713e-01 -1.03837574e+00
4.00111526e-01 -1.48902446e-01 3.83125246e-01 -9.66859460e-01
4.92483526e-01 -1.23482800e+00 4.57988530e-01 1.87141031e-01
-3.12668711e-01 3.10263753e-01 -4.19277966e-01 7.18596816e-01
-4.43033278e-02 -5.67820847e-01 1.51365697e-01 -7.10633814e-01
-8.73606682e-01 3.55232179e-01 -4.38753128e-01 -5.82086146e-01
9.58322406e-01 7.64556751e-02 -2.35011131e-01 -6.86012626e-01
-1.42339158e+00 2.18779117e-01 5.20767272e-01 8.98062468e-01
5.79475999e-01 -1.23345196e+00 -4.54753757e-01 1.22730657e-01
4.73307163e-01 -1.17742753e+00 3.02931160e-01 2.43177742e-01
-3.27627808e-02 3.16231608e-01 -5.50204992e-01 -1.20568544e-01
-1.88277829e+00 1.04229128e+00 -1.06689185e-01 -4.78229463e-01
-4.08969760e-01 4.70775425e-01 -2.43758589e-01 -4.39901710e-01
1.86623901e-01 3.85527581e-01 -1.35998178e+00 5.79558671e-01
6.98475361e-01 7.01942563e-01 7.66497850e-02 -7.81571031e-01
-1.29895657e-01 5.49594104e-01 -6.31847203e-01 6.97667897e-02
1.71950221e+00 -4.61851388e-01 1.35625079e-02 1.00678611e+00
7.91322470e-01 2.64478773e-01 -6.53006732e-01 -1.88586831e-01
-2.53145128e-01 -6.54849052e-01 -1.02020204e-01 -8.26375663e-01
-9.94137764e-01 6.08585060e-01 8.59216213e-01 7.70535707e-01
8.59523773e-01 1.46944463e-01 5.40439844e-01 6.21256888e-01
9.50393915e-01 -1.39018309e+00 6.57126904e-02 2.98093796e-01
8.63395035e-01 -1.08825827e+00 6.43996716e-01 -4.11086798e-01
-1.02943087e+00 1.43951130e+00 6.30521536e-01 -1.36604100e-01
1.35562897e+00 1.10877082e-01 2.21371949e-01 -4.85830069e-01
-1.44921184e+00 -3.08168054e-01 3.69207859e-01 6.56915903e-01
7.45919883e-01 -4.14063688e-03 -4.09852743e-01 6.48624778e-01
-6.29841238e-02 5.02943248e-02 8.47641349e-01 8.72268319e-01
-4.52626079e-01 -2.16319799e+00 -1.25091925e-01 8.04842472e-01
-5.84229827e-01 4.05120969e-01 -4.37941760e-01 3.28005672e-01
3.30176532e-01 1.11863816e+00 -4.23827291e-01 -7.72135079e-01
9.10008550e-01 -1.37205152e-02 4.55273658e-01 -1.09908676e+00
-9.50507343e-01 4.34776366e-01 4.11094636e-01 -7.01696575e-01
-7.57719040e-01 -1.22075844e+00 -6.22131467e-01 -2.27908060e-01
-2.52802849e-01 2.31674641e-01 5.88130116e-01 8.60263705e-01
3.48275810e-01 5.70152998e-01 9.79859173e-01 -5.71866274e-01
-4.08656090e-01 -8.36792171e-01 -8.33612502e-01 5.58216572e-01
-8.62344056e-02 -7.13283896e-01 1.77807719e-01 3.14795792e-01] | [10.042545318603516, 5.776186943054199] |
c69f19be-7617-400d-98ca-92337d60722e | tinto-multisensor-benchmark-for-3d | 2305.09928 | null | https://arxiv.org/abs/2305.09928v1 | https://arxiv.org/pdf/2305.09928v1.pdf | Tinto: Multisensor Benchmark for 3D Hyperspectral Point Cloud Segmentation in the Geosciences | The increasing use of deep learning techniques has reduced interpretation time and, ideally, reduced interpreter bias by automatically deriving geological maps from digital outcrop models. However, accurate validation of these automated mapping approaches is a significant challenge due to the subjective nature of geological mapping and the difficulty in collecting quantitative validation data. Additionally, many state-of-the-art deep learning methods are limited to 2D image data, which is insufficient for 3D digital outcrops, such as hyperclouds. To address these challenges, we present Tinto, a multi-sensor benchmark digital outcrop dataset designed to facilitate the development and validation of deep learning approaches for geological mapping, especially for non-structured 3D data like point clouds. Tinto comprises two complementary sets: 1) a real digital outcrop model from Corta Atalaya (Spain), with spectral attributes and ground-truth data, and 2) a synthetic twin that uses latent features in the original datasets to reconstruct realistic spectral data (including sensor noise and processing artifacts) from the ground-truth. The point cloud is dense and contains 3,242,964 labeled points. We used these datasets to explore the abilities of different deep learning approaches for automated geological mapping. By making Tinto publicly available, we hope to foster the development and adaptation of new deep learning tools for 3D applications in Earth sciences. The dataset can be accessed through this link: https://doi.org/10.14278/rodare.2256. | ['Michael Heizmann', 'Richard Gloaguen', 'Moritz Kirsch', 'Raimon Tolosana-Delgado', 'Pedram Ghamisi', 'Sandra Lorenz', 'Samuel T. Thiele', 'Ahmed J. Afifi'] | 2023-05-17 | null | null | null | null | ['point-cloud-segmentation'] | ['computer-vision'] | [-1.67265415e-01 1.11760855e-01 2.75097042e-01 -3.33479226e-01
-1.03686821e+00 -5.23200750e-01 6.15417182e-01 4.45785820e-01
-3.09873372e-01 7.67333210e-01 1.27856791e-01 -2.57238805e-01
-1.36564165e-01 -1.46484768e+00 -9.02052164e-01 -7.09020376e-01
-4.66574967e-01 9.02105689e-01 1.32884860e-01 -3.73962194e-01
-9.18697938e-02 7.65399814e-01 -1.63755155e+00 1.18987091e-01
9.88788664e-01 8.44752669e-01 6.16196394e-01 2.64778525e-01
-1.72023818e-01 2.99888849e-01 -1.52151525e-01 7.05798762e-03
4.44668740e-01 3.02647322e-01 -5.35786271e-01 -1.25572458e-01
7.48262107e-01 -3.98496628e-01 -2.27608219e-01 9.51201975e-01
6.83645844e-01 -8.89275596e-02 6.49140477e-01 -8.02089036e-01
-4.64585751e-01 5.45313716e-01 -4.91262406e-01 -1.91843182e-01
-1.62136897e-01 1.25275910e-01 8.97461414e-01 -1.06973445e+00
3.60457689e-01 1.02288759e+00 1.23168123e+00 -3.21415663e-02
-1.41804767e+00 -7.00054169e-01 -2.99671352e-01 5.79481497e-02
-1.46239591e+00 -3.96331251e-01 6.02048576e-01 -9.83680665e-01
8.63718092e-01 -1.70326203e-01 8.67539942e-01 8.05008411e-01
-6.06610924e-02 4.37223047e-01 1.36082828e+00 -1.99856356e-01
3.08352441e-01 -3.40538174e-01 -3.10146809e-01 4.60782826e-01
3.07252496e-01 3.95749927e-01 -4.73228067e-01 -2.45047331e-01
7.68070161e-01 -3.98147777e-02 -9.84319597e-02 -3.12195182e-01
-8.37503076e-01 8.71234775e-01 6.62874877e-01 7.17212856e-02
-6.30871236e-01 2.43455648e-01 1.77259162e-01 2.19783664e-01
8.06434453e-01 1.37435943e-01 -4.17780519e-01 7.46040642e-02
-1.31568599e+00 5.49637616e-01 5.52040577e-01 6.05275035e-01
1.17065525e+00 3.73359323e-01 7.68685520e-01 9.61403131e-01
3.19326609e-01 1.09738684e+00 1.72373913e-02 -8.83516669e-01
4.00800586e-01 5.55110991e-01 1.41541019e-01 -1.06299782e+00
-5.07317305e-01 -2.96643108e-01 -8.85607839e-01 6.55985773e-01
2.08928749e-01 6.26718393e-03 -9.65258121e-01 1.33159041e+00
-2.56444793e-02 -8.45512562e-03 1.50264367e-01 8.68043363e-01
8.95363033e-01 6.03734195e-01 6.34771064e-02 4.28218752e-01
8.50385785e-01 1.20239416e-02 -2.77862996e-01 -4.08103198e-01
6.36139870e-01 -3.78194153e-01 1.07071805e+00 1.47668526e-01
-6.84385002e-01 -5.22605062e-01 -1.16072619e+00 1.15606748e-01
-7.52157271e-01 2.32350022e-01 7.21930325e-01 2.34280184e-01
-1.11162972e+00 6.61872804e-01 -1.03053820e+00 -3.11919212e-01
7.12378025e-01 9.20897648e-02 -5.65709710e-01 1.11320332e-01
-1.39021385e+00 9.96248543e-01 5.65720975e-01 4.24262464e-01
-1.10211444e+00 -8.70526910e-01 -8.56300116e-01 4.09316011e-02
3.72473300e-02 -1.62792757e-01 8.38226199e-01 -6.65330887e-01
-7.49733031e-01 1.27816379e+00 4.36329871e-01 -4.96323943e-01
7.63799906e-01 -2.45096639e-01 -1.82366788e-01 -1.73507690e-01
4.83281642e-01 5.12208879e-01 4.52605307e-01 -1.39909434e+00
-4.60018545e-01 -6.48596704e-01 -3.96137327e-01 -2.09078282e-01
-3.59367952e-02 -4.70173866e-01 7.40410537e-02 -5.41201949e-01
4.11064237e-01 -8.28123450e-01 -1.30295679e-01 1.27011448e-01
-9.60308462e-02 1.82350636e-01 5.50751686e-01 -9.36148047e-01
4.42047417e-01 -1.99871063e+00 -1.03872925e-01 2.89860815e-01
1.09695561e-01 2.24099055e-01 -1.09438427e-01 6.20530486e-01
-6.32421970e-02 2.75507599e-01 -8.41944456e-01 -1.69823587e-01
-1.66836083e-01 3.71573269e-01 -4.29022282e-01 5.95531881e-01
1.77937806e-01 6.53703451e-01 -8.15528274e-01 -2.54218966e-01
6.17525816e-01 4.18668211e-01 -6.04781928e-03 3.21820341e-02
-4.09317255e-01 2.49825671e-01 -3.62074882e-01 7.60743380e-01
1.27767825e+00 2.22189635e-01 1.08569585e-01 -1.09747358e-01
-5.83084941e-01 1.45371780e-01 -1.12509489e+00 1.66425264e+00
-6.43887341e-01 8.32888722e-01 3.52991857e-02 -8.29083622e-01
1.43332493e+00 2.53957570e-01 5.32621384e-01 -9.13782656e-01
-1.88791186e-01 6.61870241e-01 -2.19651014e-01 -6.43438876e-01
5.45625389e-01 -4.70539331e-01 8.17452297e-02 2.61848509e-01
8.41345191e-02 -8.06005836e-01 -2.70809293e-01 9.98688629e-04
6.65511966e-01 3.50550264e-01 -1.74539521e-01 -5.10375381e-01
2.01529674e-02 3.17082971e-01 3.61417085e-01 6.85708523e-01
1.68277770e-01 8.85392547e-01 1.86352879e-01 -8.89679372e-01
-1.48748302e+00 -1.27141941e+00 -5.90305328e-01 4.22089726e-01
-2.82880336e-01 4.03565392e-02 -1.07370965e-01 -5.77032119e-02
6.28826559e-01 5.78299701e-01 -7.85154045e-01 1.93328783e-01
-2.00524002e-01 -1.08881307e+00 9.11803007e-01 5.20991683e-01
5.18540382e-01 -8.81558597e-01 -6.26059353e-01 3.98030281e-01
-2.03585774e-01 -8.48247051e-01 8.26861441e-01 3.05039406e-01
-1.00786006e+00 -1.01267672e+00 -2.38011286e-01 -1.44908249e-01
1.89563304e-01 -9.59514000e-04 1.11237395e+00 2.61630975e-02
-8.97240173e-03 -8.85102972e-02 -3.99445653e-01 -6.28614366e-01
-3.86462629e-01 2.19517320e-01 1.46392226e-01 -3.37306648e-01
4.07959998e-01 -9.15615201e-01 -6.46707565e-02 3.13030966e-02
-8.86414587e-01 1.09617658e-01 5.50267220e-01 6.10758841e-01
6.23873711e-01 -3.87018286e-02 5.42225838e-01 -6.38802767e-01
1.00723766e-01 -6.34558976e-01 -9.39181387e-01 2.31719613e-02
-2.44689494e-01 -8.14262182e-02 3.07581276e-01 2.37853050e-01
-9.87620234e-01 2.61384577e-01 -4.42220688e-01 -2.10589707e-01
-3.26444954e-01 1.19417894e+00 7.71021917e-02 -2.24382341e-01
1.03874385e+00 2.54248828e-01 9.14947912e-02 -7.08917379e-01
4.73165028e-02 8.32674325e-01 7.95486271e-01 -6.48227274e-01
9.76209879e-01 7.39763796e-01 1.10077493e-01 -1.36053145e+00
-7.23883808e-01 -2.41762504e-01 -9.52205300e-01 -4.06791598e-01
5.78149557e-01 -1.30498028e+00 -1.74534649e-01 7.26463795e-01
-7.41896749e-01 -8.53862703e-01 -7.29790851e-02 3.96869957e-01
-4.99525875e-01 2.00688973e-01 -1.64741620e-01 -8.34845722e-01
-2.58793414e-01 -7.72723377e-01 1.21598995e+00 -1.30070329e-01
1.17017806e-01 -1.06851733e+00 5.16452074e-01 1.83833584e-01
4.87837285e-01 9.22492921e-01 7.88850367e-01 -2.03347132e-01
-5.34521282e-01 -4.16657537e-01 -2.49181867e-01 2.93249279e-01
3.52394849e-01 2.05113649e-01 -1.41399610e+00 -1.67895377e-01
-1.50027856e-01 -5.73646843e-01 1.27665496e+00 3.82857502e-01
9.90005612e-01 1.74839750e-01 -2.70513706e-02 7.77052939e-01
1.71689796e+00 -2.09886536e-01 6.80004537e-01 7.05202818e-01
6.31112754e-01 6.98424697e-01 5.54344416e-01 5.07340789e-01
2.75879174e-01 5.05685866e-01 1.02900422e+00 -2.89060891e-01
1.80334985e-01 -3.14119905e-01 -1.86057910e-01 3.32709759e-01
-5.30255497e-01 8.66370872e-02 -1.77590847e+00 8.79954398e-01
-1.80444980e+00 -9.36412156e-01 -5.70315003e-01 2.14418674e+00
6.27299666e-01 -4.83540371e-02 -3.29252243e-01 2.58989871e-01
3.58698875e-01 4.10349280e-01 -3.43518645e-01 2.90782630e-01
-7.74488449e-01 1.98595420e-01 9.06742454e-01 6.11920297e-01
-1.27343965e+00 1.15652227e+00 5.07976532e+00 4.33833003e-01
-1.40130615e+00 1.09477811e-01 1.91362962e-01 1.34680390e-01
-4.17138785e-01 -9.32298005e-02 -4.21803385e-01 3.26255292e-01
9.04821038e-01 1.58949345e-01 1.01749644e-01 8.31928730e-01
6.72837079e-01 -1.84214950e-01 -7.61399329e-01 8.34014714e-01
-2.34471545e-01 -1.89063036e+00 -2.45257378e-01 2.75678039e-01
7.01892197e-01 9.30000246e-01 -3.55984867e-01 -7.47934431e-02
5.11459827e-01 -9.96890008e-01 9.31735992e-01 7.02466190e-01
9.06844199e-01 -4.67805654e-01 9.94146407e-01 1.87152892e-01
-1.01045477e+00 1.24784753e-01 -5.33137083e-01 -3.52644414e-01
-2.71500554e-03 9.22464192e-01 -8.07069659e-01 8.20080400e-01
1.12080073e+00 9.99973476e-01 -6.36726677e-01 1.19233799e+00
-3.04005235e-01 6.50432944e-01 -6.36367798e-01 4.64087099e-01
3.47425103e-01 -3.14640343e-01 3.15428078e-01 1.01092398e+00
5.33561945e-01 6.19840808e-02 1.18073653e-02 9.56605315e-01
-7.80715346e-02 -2.21289337e-01 -9.15663898e-01 1.79271445e-01
8.03127885e-01 9.70671296e-01 -4.31626350e-01 -5.02919331e-02
-2.49234468e-01 2.35591114e-01 2.44273946e-01 4.47076149e-02
-3.63041222e-01 -1.69602469e-01 5.80652535e-01 3.66177440e-01
-4.80958410e-02 -5.28604329e-01 -7.06726074e-01 -9.58962142e-01
1.76445432e-02 -6.31731808e-01 1.60646573e-01 -1.08525777e+00
-1.19586182e+00 3.74561787e-01 9.94612724e-02 -1.33331740e+00
8.53408352e-02 -6.08687103e-01 -5.14510214e-01 1.09102237e+00
-1.63710058e+00 -1.44686842e+00 -9.29550290e-01 1.01887666e-01
1.19301587e-01 -1.59638494e-01 8.02545309e-01 2.82857656e-01
7.24698454e-02 -2.96697140e-01 6.16305113e-01 1.35153055e-01
5.01945019e-01 -1.19252896e+00 7.33678877e-01 6.12298906e-01
-8.44437107e-02 -1.12406261e-01 6.58156753e-01 -7.01512039e-01
-1.11188221e+00 -1.34767842e+00 6.99741662e-01 -3.72680813e-01
8.73914242e-01 -3.90372753e-01 -1.43977094e+00 8.38503778e-01
-5.60024798e-01 5.41834906e-03 5.05912960e-01 7.16977101e-03
-1.35130048e-01 -3.75865400e-01 -1.15074921e+00 1.52705491e-01
6.58982635e-01 -6.63104951e-01 -5.97683549e-01 4.24722016e-01
4.71944474e-02 -3.75749558e-01 -9.69515026e-01 5.08822024e-01
7.04205632e-01 -1.03650427e+00 9.32307422e-01 -5.76558001e-02
8.37019205e-01 -3.53896856e-01 -5.51019728e-01 -1.41796637e+00
-1.75300121e-01 3.12473178e-01 5.23269475e-01 1.00512755e+00
2.50051558e-01 -5.32338142e-01 9.62358356e-01 2.00622514e-01
-5.58765471e-01 -1.32583797e-01 -1.07139301e+00 -8.61789882e-01
5.89568794e-01 -7.63007998e-01 9.11225975e-01 1.17887115e+00
-6.01381183e-01 -3.48692000e-01 -2.07724035e-01 6.65448785e-01
8.32694530e-01 2.07073510e-01 7.70940721e-01 -1.85850251e+00
2.34413773e-01 -2.46384263e-01 -5.17633080e-01 -2.03919023e-01
1.85579106e-01 -8.37441564e-01 6.93601593e-02 -1.64238167e+00
-3.52225125e-01 -1.07299614e+00 1.86815903e-01 9.99351561e-01
1.81141898e-01 2.91485399e-01 -1.05437994e-01 6.04454815e-01
4.34048772e-01 9.93055165e-01 6.31499767e-01 -4.20881987e-01
-1.07370913e-02 -1.79484650e-01 1.59979641e-01 7.88743854e-01
9.79017675e-01 -6.82059169e-01 1.08226702e-01 -1.12476110e+00
4.33009982e-01 1.13235474e-01 7.16749430e-01 -1.31618071e+00
-1.38733640e-01 -2.89339006e-01 4.45798099e-01 -1.03819180e+00
3.78797084e-01 -8.59561920e-01 5.47889173e-01 4.78212297e-01
2.75642425e-02 -3.53882283e-01 4.05124545e-01 1.04443043e-01
-3.07448566e-01 -2.78073341e-01 8.67361546e-01 -3.84688139e-01
-9.60406721e-01 4.49686974e-01 -2.91762471e-01 -2.99827397e-01
5.21412373e-01 -3.21613878e-01 -2.12909624e-01 -1.22124463e-01
-6.87185347e-01 3.14226359e-01 7.92288899e-01 2.14909971e-01
6.38021708e-01 -1.15178740e+00 -1.26312554e+00 3.22691262e-01
3.70562881e-01 5.55113673e-01 3.91842365e-01 3.22095931e-01
-1.16363561e+00 1.05903000e-01 -6.43947303e-01 -8.07545125e-01
-7.06348181e-01 -1.97954878e-01 7.34440148e-01 3.39841306e-01
-7.32475519e-01 6.45992219e-01 -2.78330088e-01 -9.69793856e-01
-3.36886823e-01 -1.76819175e-01 -4.49920446e-02 2.92765200e-01
1.60454094e-01 1.60240874e-01 5.41718304e-01 -7.70157695e-01
-2.76382118e-01 4.69472051e-01 5.77195644e-01 -5.92232794e-02
2.06956768e+00 1.85248777e-01 -2.38475278e-01 6.97197020e-01
6.56817496e-01 -1.99037790e-01 -1.27838147e+00 -2.92499155e-01
1.70690730e-01 -8.44321966e-01 3.96983564e-01 -7.07713425e-01
-1.22971284e+00 1.11859179e+00 8.84259880e-01 6.87142387e-02
6.95986390e-01 -1.77251697e-02 2.57730335e-01 5.70013881e-01
4.73052979e-01 -9.99973893e-01 -3.32581520e-01 5.49947321e-01
1.20353639e+00 -1.50760520e+00 1.94917396e-01 7.68446550e-02
-1.81479990e-01 1.29020822e+00 3.36465180e-01 -1.50366664e-01
7.66436219e-01 3.15160900e-01 1.86883047e-01 -5.20767093e-01
-4.35327172e-01 -2.78533638e-01 -3.52437884e-01 9.11266029e-01
3.61181378e-01 3.89480084e-01 2.35441253e-01 2.36995876e-01
-5.67333341e-01 2.84983754e-01 7.47503340e-01 8.35287750e-01
-5.54963768e-01 -8.10639620e-01 -6.68347120e-01 5.49817979e-01
6.25986084e-02 -1.26140565e-01 -1.95539236e-01 8.71157169e-01
2.51319945e-01 4.75084394e-01 4.18562293e-01 -2.89649159e-01
2.39371255e-01 -2.12485969e-01 1.40543550e-01 -6.15700841e-01
-1.17005117e-01 -1.99401289e-01 1.86135009e-01 -5.33722453e-02
-5.09166539e-01 -9.76437628e-01 -9.83642995e-01 -4.09217298e-01
8.24160725e-02 1.77232131e-01 9.82006073e-01 9.00706470e-01
8.90235305e-02 1.62124455e-01 3.19829762e-01 -1.45378113e+00
-3.01342010e-01 -1.21138358e+00 -8.55215013e-01 1.34653822e-01
4.31882977e-01 -9.36047435e-01 -2.99226105e-01 -6.57275990e-02] | [9.393776893615723, -1.4091452360153198] |
30870e5b-8b3a-4f70-98bb-f80a9151ff1d | materials-representation-and-transfer | 2106.02225 | null | https://arxiv.org/abs/2106.02225v3 | https://arxiv.org/pdf/2106.02225v3.pdf | Materials Representation and Transfer Learning for Multi-Property Prediction | The adoption of machine learning in materials science has rapidly transformed materials property prediction. Hurdles limiting full capitalization of recent advancements in machine learning include the limited development of methods to learn the underlying interactions of multiple elements, as well as the relationships among multiple properties, to facilitate property prediction in new composition spaces. To address these issues, we introduce the Hierarchical Correlation Learning for Multi-property Prediction (H-CLMP) framework that seamlessly integrates (i) prediction using only a material's composition, (ii) learning and exploitation of correlations among target properties in multi-target regression, and (iii) leveraging training data from tangential domains via generative transfer learning. The model is demonstrated for prediction of spectral optical absorption of complex metal oxides spanning 69 3-cation metal oxide composition spaces. H-CLMP accurately predicts non-linear composition-property relationships in composition spaces for which no training data is available, which broadens the purview of machine learning to the discovery of materials with exceptional properties. This achievement results from the principled integration of latent embedding learning, property correlation learning, generative transfer learning, and attention models. The best performance is obtained using H-CLMP with Transfer learning (H-CLMP(T)) wherein a generative adversarial network is trained on computational density of states data and deployed in the target domain to augment prediction of optical absorption from composition. H-CLMP(T) aggregates multiple knowledge sources with a framework that is well-suited for multi-target regression across the physical sciences. | ['John M. Gregoire', 'Carla P. Gomes', 'Dan Guevarra', 'Shufeng Kong'] | 2021-06-04 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [ 6.54572606e-01 -1.70995116e-01 -2.52226412e-01 -5.20969555e-02
-1.04207754e+00 -2.41811797e-01 5.36160767e-01 2.46469557e-01
-2.57302020e-02 8.93809557e-01 1.06616765e-01 -3.06984186e-01
-5.95719457e-01 -1.12516797e+00 -8.06612313e-01 -1.15904868e+00
8.31164122e-02 6.00068152e-01 1.16785668e-01 -2.93834299e-01
3.05793315e-01 5.84874272e-01 -1.47027457e+00 4.73628849e-01
1.06755793e+00 1.35161614e+00 2.43823469e-01 4.06645030e-01
5.18685319e-02 6.77877724e-01 2.31480747e-02 -2.56925486e-02
4.25700936e-03 -2.09678769e-01 -6.81520462e-01 -2.63923377e-01
1.65482938e-01 1.41016409e-01 -8.44416097e-02 6.71524882e-01
3.43660295e-01 9.57101732e-02 1.29344416e+00 -1.16059625e+00
-1.13563323e+00 4.01341975e-01 -2.09155127e-01 -1.67210266e-01
4.27003235e-01 3.19016188e-01 1.28834796e+00 -7.53595352e-01
3.93198788e-01 7.39899278e-01 7.73546934e-01 3.39347064e-01
-1.69278347e+00 -6.51538014e-01 -2.37977639e-01 4.11434948e-01
-1.10039210e+00 -7.72102401e-02 1.14478910e+00 -8.15252483e-01
1.49173772e+00 1.61235154e-01 8.44440699e-01 1.20818877e+00
4.24496472e-01 4.14611042e-01 1.26313758e+00 -5.43876708e-01
2.22193807e-01 3.26797247e-01 -7.04389289e-02 4.05260503e-01
4.27262299e-02 5.24096191e-01 -7.06085324e-01 -2.63374269e-01
4.37903374e-01 4.42417823e-02 -4.26208340e-02 -5.02662182e-01
-9.29091632e-01 7.37083793e-01 5.41879475e-01 2.80078143e-01
-5.35246432e-01 -2.47347690e-02 2.55570292e-01 4.24258262e-01
7.63078213e-01 1.06310451e+00 -8.28058600e-01 7.19043016e-02
-8.43003690e-01 2.68003017e-01 7.07816958e-01 7.05919206e-01
1.20569301e+00 3.30108494e-01 7.65481070e-02 7.56724000e-01
2.85951257e-01 6.50253832e-01 3.75211537e-01 -4.82596368e-01
3.18431437e-01 6.65552616e-01 -3.25436927e-02 -3.44067305e-01
-3.49654317e-01 -2.31986433e-01 -4.86204177e-01 4.63540435e-01
1.00180253e-01 -5.27593493e-02 -6.68121934e-01 1.49744391e+00
2.69425541e-01 -1.51471868e-01 1.54044092e-01 3.44036996e-01
5.06261170e-01 7.54251957e-01 2.58569419e-01 -1.10421285e-01
8.71317983e-01 -8.39973152e-01 -1.44443572e-01 1.04981385e-01
3.84620845e-01 -5.40038466e-01 1.04693341e+00 3.62835109e-01
-9.65418935e-01 -6.56119466e-01 -1.27346504e+00 -2.09218971e-02
-8.91769648e-01 -2.29807645e-01 1.06591761e+00 5.94109833e-01
-6.41274333e-01 1.13310122e+00 -8.25747371e-01 1.72314614e-01
6.15972579e-01 9.16046560e-01 -4.54223573e-01 -3.63749228e-02
-1.11443889e+00 8.72643471e-01 4.66390997e-01 -1.13103040e-01
-7.76790261e-01 -1.36386418e+00 -7.18789339e-01 -8.78079142e-03
3.69211793e-01 -8.00618529e-01 5.66973567e-01 -1.00246131e+00
-1.93067861e+00 2.57319212e-01 2.07873583e-01 9.76986811e-03
1.94703236e-01 -2.56027609e-01 -6.34394884e-01 -4.60355222e-04
9.60046947e-02 3.33061516e-01 1.04597390e+00 -1.14913869e+00
-1.75640866e-01 -1.66229606e-01 -5.10564625e-01 -1.27505630e-01
-4.16857272e-01 -3.23588878e-01 3.89562011e-01 -3.44633222e-01
5.89815825e-02 -8.46564889e-01 -6.37314916e-02 -3.09870034e-01
-3.79642844e-01 -2.26849094e-01 9.57980752e-01 -3.22678953e-01
5.46992421e-01 -1.91407657e+00 4.11505818e-01 4.11657751e-01
1.98853999e-01 4.65154760e-02 -2.38520622e-01 9.31423724e-01
-5.28072655e-01 -6.64323121e-02 -1.92856938e-01 1.27843946e-01
-2.89591011e-02 -1.65581867e-01 -2.65522003e-01 2.94238597e-01
7.62808442e-01 1.28821015e+00 -7.85559297e-01 9.53329653e-02
3.70899707e-01 6.45855427e-01 -8.25504780e-01 3.04001093e-01
-7.23000467e-01 6.54872954e-01 -6.09696984e-01 8.94326866e-01
4.42776114e-01 -5.18630266e-01 -5.43685816e-02 -5.12733817e-01
-3.52739304e-01 4.23336864e-01 -7.07318604e-01 1.42426574e+00
-4.79714364e-01 1.70070618e-01 -5.27327120e-01 -1.06129396e+00
9.58126903e-01 3.55900764e-01 1.11730683e+00 -6.89191282e-01
-1.36031881e-01 3.86466801e-01 2.09908918e-01 -6.62741601e-01
1.69611156e-01 -6.71866238e-01 1.37292340e-01 4.23796803e-01
4.74324584e-01 -5.42141020e-01 -2.21017644e-01 -3.25898081e-01
1.04040384e+00 5.14452636e-01 2.53839523e-01 -2.51980633e-01
5.31804681e-01 -1.81080952e-01 8.03224519e-02 4.89905000e-01
2.22805962e-01 7.91971758e-02 3.15534234e-01 -1.97788402e-01
-1.58620465e+00 -1.33044398e+00 -4.67178792e-01 1.11741769e+00
-1.62438199e-01 -2.15038344e-01 -1.24167182e-01 -3.47794265e-01
3.69792491e-01 5.11314571e-01 -8.84230793e-01 -5.36530912e-01
-2.91967869e-01 -9.92615342e-01 9.41548403e-03 6.57254815e-01
2.36380339e-01 -1.29694378e+00 2.78307088e-02 4.23587561e-01
5.65811813e-01 -9.75595295e-01 5.66884317e-02 6.67813301e-01
-7.28583515e-01 -1.11797285e+00 -2.79696077e-01 -4.58415389e-01
2.87164807e-01 -2.89570749e-01 7.49696434e-01 -3.15568447e-01
-5.60622990e-01 5.63131094e-01 -1.79110050e-01 -5.34818292e-01
-7.71468759e-01 3.21098596e-01 1.64022967e-01 -3.78087373e-03
2.40792245e-01 -1.13373482e+00 -4.64791119e-01 6.01151623e-02
-6.71227217e-01 1.31421313e-01 1.03326154e+00 1.08102632e+00
6.14877641e-01 1.65485218e-01 9.14706230e-01 -9.92434800e-01
3.72766793e-01 -6.51874483e-01 -5.62350929e-01 4.25528944e-01
-1.18224823e+00 4.67602521e-01 7.04175293e-01 -7.41825461e-01
-1.02860129e+00 -2.11215705e-01 -2.98295707e-01 -1.52601555e-01
-1.25765622e-01 4.86055642e-01 -3.74529004e-01 -2.44083911e-01
6.23128414e-01 2.23367646e-01 -1.11805499e-01 -3.93279046e-01
2.76992112e-01 4.76680845e-01 3.36743146e-01 -1.07669771e+00
8.53481650e-01 9.04657692e-02 5.00425994e-01 -9.33155775e-01
-8.31508458e-01 -3.09785187e-01 -9.33784604e-01 -6.77669942e-02
9.71447170e-01 -6.75277293e-01 -9.15102303e-01 4.02800918e-01
-5.38432658e-01 -4.06768590e-01 -5.83977342e-01 6.37389660e-01
-8.54155004e-01 2.65688188e-02 -5.69169641e-01 -6.51332021e-01
-6.02279663e-01 -1.21939850e+00 1.03135157e+00 -5.63895181e-02
-1.71907380e-01 -1.29348826e+00 1.01743624e-01 4.83981311e-01
5.73542595e-01 3.41084778e-01 1.68680632e+00 -7.12361991e-01
-7.86235094e-01 -3.13735127e-01 -1.06700230e-02 5.23896635e-01
5.83738208e-01 2.10024863e-02 -1.11083674e+00 -1.82226077e-01
-5.95282502e-02 -6.43273890e-01 7.90557981e-01 3.96512121e-01
1.15111697e+00 4.18061167e-02 -2.81674504e-01 6.34276867e-01
1.56128514e+00 2.22015083e-01 7.43929982e-01 3.73658121e-01
9.41493332e-01 3.60621482e-01 1.55061945e-01 2.53775090e-01
-3.31415012e-02 6.31499052e-01 2.44957641e-01 -8.35040468e-04
-3.15875448e-02 -4.31529105e-01 2.72821695e-01 8.56459618e-01
-4.86228228e-01 1.49983838e-01 -7.72736013e-01 -1.23506570e-02
-1.39794183e+00 -1.07830000e+00 4.39541116e-02 2.02666426e+00
9.05253530e-01 2.64484137e-01 -1.77914929e-03 2.38349840e-01
3.50667775e-01 4.10399437e-02 -1.05523312e+00 -1.40910342e-01
-2.33995710e-02 7.07933843e-01 3.78600091e-01 3.81942451e-01
-9.03104603e-01 8.41617346e-01 6.72720432e+00 7.05147207e-01
-1.25634408e+00 8.77596736e-02 2.74959743e-01 2.86250502e-01
-6.95468783e-01 6.58169240e-02 -7.75995910e-01 3.33469361e-01
9.62625980e-01 1.41124487e-01 5.61419725e-01 7.20711768e-01
-1.14156067e-01 1.01744696e-01 -1.35639238e+00 5.80872476e-01
-1.18342444e-01 -1.46954465e+00 -1.19094933e-04 1.69886053e-01
9.54369485e-01 1.22256950e-01 5.38735867e-01 3.46661299e-01
-5.45930378e-02 -9.43706214e-01 4.71149564e-01 7.18145192e-01
9.98257339e-01 -3.99202287e-01 3.46429706e-01 -2.00411063e-02
-8.85797262e-01 -1.68251842e-01 -1.82824999e-01 6.57401979e-02
-1.04812868e-01 4.22358125e-01 -9.44231987e-01 7.96911657e-01
3.92560869e-01 9.22874808e-01 -2.75419444e-01 6.68661237e-01
-1.22827962e-01 6.13128066e-01 -1.26061201e-01 -5.96053414e-02
-6.15334101e-02 -5.86660087e-01 3.57586861e-01 6.86381757e-01
3.02980810e-01 -2.79134095e-01 3.57484035e-02 1.31228447e+00
6.48006722e-02 2.99251787e-02 -5.08751512e-01 -4.04213786e-01
9.51697454e-02 1.02189076e+00 -3.54251117e-01 -1.06059730e-01
-5.78464925e-01 6.56440437e-01 3.63170207e-01 3.50942761e-01
-7.03375876e-01 8.74256641e-02 6.25212431e-01 4.09341812e-01
4.53948677e-01 -1.62456155e-01 -2.48849154e-01 -7.99641669e-01
-2.37942770e-01 -5.93437672e-01 -4.82808948e-02 -8.46754014e-01
-1.78355300e+00 3.48040551e-01 5.09107597e-02 -1.02029669e+00
-7.85964206e-02 -1.05703831e+00 -4.43484128e-01 1.08530486e+00
-1.72665393e+00 -1.60442603e+00 8.72681811e-02 4.03643131e-01
1.35321632e-01 -4.97933656e-01 1.10092938e+00 1.74468279e-01
-3.86447579e-01 3.80537510e-01 3.88427526e-01 -4.52074617e-01
5.72847545e-01 -1.46271372e+00 5.31898029e-02 7.40373209e-02
-7.14368895e-02 1.64505705e-01 6.26183212e-01 -7.32007265e-01
-1.52172935e+00 -9.36121643e-01 2.67285019e-01 -5.77594280e-01
1.13618731e+00 -4.56045628e-01 -1.05367267e+00 3.60651493e-01
-2.21205875e-01 9.16773304e-02 1.21110070e+00 3.68381500e-01
-5.85004330e-01 -1.40311226e-01 -9.17311490e-01 3.56627166e-01
6.09515071e-01 -1.06507337e+00 -3.40918511e-01 5.43949962e-01
6.90868556e-01 -1.06476203e-01 -1.53756285e+00 5.98956764e-01
7.27289438e-01 -6.87881291e-01 1.22028708e+00 -8.08872163e-01
7.84323037e-01 8.31983387e-02 -3.75140101e-01 -1.16428280e+00
-4.08680409e-01 -4.68816906e-01 -1.77194148e-01 9.30487514e-01
7.96601892e-01 -7.54241943e-01 7.91375875e-01 9.23444331e-01
-4.18430746e-01 -1.04403162e+00 -8.75079215e-01 -7.53798366e-01
5.07785439e-01 -4.81538892e-01 6.50368392e-01 7.41414011e-01
-2.07925275e-01 4.94063139e-01 -3.53666127e-01 2.30955094e-01
3.14502865e-01 3.42279553e-01 4.52948153e-01 -1.39472163e+00
-7.52436638e-01 -3.10423911e-01 -4.21540111e-01 -5.19623816e-01
3.63954484e-01 -1.16882658e+00 -3.65850151e-01 -1.19466043e+00
3.33827525e-01 -6.13842845e-01 -5.87980986e-01 5.17138183e-01
-3.88718322e-02 -5.53957671e-02 -1.28895864e-01 2.43191928e-01
-2.39550233e-01 1.03282499e+00 1.47025061e+00 -3.69990259e-01
-3.50298196e-01 1.83747724e-01 -6.23633087e-01 1.85964137e-01
6.15059257e-01 -3.87276322e-01 -3.10273319e-01 3.47482055e-01
4.42581475e-01 -7.49194250e-02 2.70040900e-01 -1.29186213e+00
-7.09514245e-02 -2.06765845e-01 6.24667406e-01 -2.85440475e-01
5.84668398e-01 -9.27639246e-01 2.30438337e-01 3.05406362e-01
-1.91113517e-01 -3.22922170e-01 2.05874667e-01 5.24600267e-01
2.31126067e-03 -1.61116391e-01 5.47823429e-01 1.03487857e-01
-5.25842726e-01 5.76751113e-01 -4.28853976e-03 -3.17646325e-01
9.11092162e-01 -3.63850027e-01 -3.34022164e-01 1.62836596e-01
-8.09997499e-01 -3.00290525e-01 3.51434201e-01 3.00179660e-01
4.64434505e-01 -1.26025116e+00 -3.80641133e-01 5.25908232e-01
2.65143365e-01 -1.72461674e-01 3.91074717e-01 7.00010717e-01
-1.41717613e-01 3.85574192e-01 -3.25663835e-01 -4.86985236e-01
-8.85878265e-01 8.11380565e-01 3.54348511e-01 -3.44218016e-01
-5.19679785e-01 6.29635096e-01 -1.73254276e-03 -4.17249262e-01
-6.90275013e-01 -1.54299676e-01 1.59249324e-02 -8.41059089e-02
4.75109592e-02 3.86001408e-01 1.93756819e-01 -4.25304204e-01
8.33102018e-02 6.58807337e-01 -2.60277718e-01 2.02399418e-01
1.94723260e+00 4.18326110e-01 -1.53125167e-01 8.05616021e-01
1.41461051e+00 9.86753702e-02 -1.59033632e+00 -1.95564732e-01
2.10944470e-02 -2.03832816e-02 7.49529451e-02 -1.00427604e+00
-5.98760128e-01 6.57021523e-01 6.45661175e-01 1.20430119e-01
9.51475859e-01 2.28361249e-01 6.42921329e-01 3.53645593e-01
2.39591092e-01 -1.06471610e+00 5.50094545e-01 2.84634203e-01
8.04074645e-01 -1.40032279e+00 4.21546161e-01 -4.47477520e-01
-3.24404210e-01 1.31953573e+00 6.37176692e-01 1.78777967e-02
9.74357486e-01 2.49783218e-01 -3.94626230e-01 -5.16145766e-01
-7.71522284e-01 -3.71353850e-02 6.73260629e-01 6.58383369e-01
4.61302072e-01 1.35665700e-01 4.28953797e-01 4.14734334e-01
-1.73306525e-01 -2.41933420e-01 -6.40437752e-02 7.54127920e-01
-2.15107471e-01 -1.69220781e+00 -8.96040723e-02 7.74044633e-01
-1.35182500e-01 -2.98464835e-01 2.79848650e-03 7.24084735e-01
4.02453303e-01 4.92625177e-01 -7.92452544e-02 -4.10670072e-01
1.32091567e-01 3.34332198e-01 8.30478251e-01 -7.39750803e-01
-3.84954631e-01 -8.58165249e-02 -2.74239272e-01 -2.53882289e-01
-5.19014001e-01 -7.82250822e-01 -8.89007866e-01 -1.03516933e-02
-6.02653086e-01 4.33840230e-02 7.50795245e-01 1.13732493e+00
1.09191544e-01 8.51144373e-01 7.20582545e-01 -1.07667267e+00
-5.00163794e-01 -1.01453197e+00 -8.07486176e-01 3.51656586e-01
3.50356638e-01 -8.89878869e-01 -2.35559523e-01 -2.37125888e-01] | [5.209752082824707, 5.43380069732666] |
82b548ca-0e9b-4dfc-8f71-e1ef63fdfc7a | where-is-my-uri | null | null | https://www.researchgate.net/publication/325529570_Where_is_My_URI | https://svn.aksw.org/papers/2018/ESWC_WIMU/public.pdf | Where is my URI? | One of the Semantic Web foundations is the possibility to dereference URIs to let applications negotiate their semantic content. However, this exploitation is often infeasible as the availability of such information depends on the reliability of networks, services, and human factors. Moreover, it has been shown that around 90% of the information published as Linked Open Data is available as data dumps and more than 60% of endpoints are offline. To this end, we propose a Web service called Where is my URI?. Our service aims at indexing URIs and their use in order to let Linked Data consumers find the respective RDF data source, in case such information cannot be retrieved from the URI alone. We rank the corresponding datasets by following the rationale upon which a dataset contributes to the definition of a URI proportionally to the number of literals. We finally describe potential use-cases of applications that can immediately benefit from our simple yet useful service. | ['Axel-Cyrille Ngonga Ngomo', 'Andre Valdestilhas', 'Markus Nentwig', 'Edgard Marx', 'Tommaso Soru', 'Muhammad Saleem'] | 2018-06-15 | null | null | null | european-semantic-web-conference-2018-6 | ['rdf-dataset-discovery'] | ['knowledge-base'] | [-1.26011714e-01 5.92679977e-01 -4.58237648e-01 -3.81394863e-01
-4.76606756e-01 -9.68571842e-01 5.72053671e-01 6.80924237e-01
-3.09443116e-01 8.25649738e-01 1.52018905e-01 -1.59457505e-01
-5.53870499e-01 -1.18093789e+00 -5.73686719e-01 -3.64999734e-02
1.89950407e-01 5.78011572e-01 8.41668069e-01 -3.80501151e-01
1.32174045e-01 4.57399338e-01 -1.86824727e+00 2.95138180e-01
7.49661207e-01 1.43648696e+00 -4.27186042e-02 -4.96608168e-02
-8.92378867e-01 3.49293739e-01 -4.51198727e-01 -5.43484330e-01
1.40587136e-01 1.89674366e-02 -1.06165302e+00 -4.80706275e-01
-5.74364848e-02 -1.75391838e-01 1.77757517e-01 8.91794801e-01
1.51809067e-01 -2.28529125e-01 1.64531961e-01 -1.60085213e+00
-3.67334694e-01 7.87460327e-01 -5.58291636e-02 -1.83984071e-01
5.88974655e-01 -5.62651694e-01 1.03251147e+00 -2.17355862e-01
1.46725988e+00 3.54523420e-01 2.15440691e-02 2.82808602e-01
-5.97965896e-01 -2.94756442e-01 -2.71607518e-01 7.68635869e-02
-1.08875954e+00 -5.40929854e-01 3.55184346e-01 6.92689791e-02
5.73175132e-01 8.15393567e-01 3.70631844e-01 4.95772451e-01
-2.93753386e-01 2.20737159e-01 7.47765005e-01 -4.77237642e-01
5.24670482e-01 3.11510801e-01 1.54219285e-01 5.16252965e-02
9.72843051e-01 -5.77627361e-01 -5.53243697e-01 -4.11516845e-01
1.08025558e-01 -2.23335698e-01 -3.39601278e-01 -6.70457423e-01
-8.90213788e-01 2.22400248e-01 3.16329688e-01 6.31339908e-01
-5.65889359e-01 4.49452735e-03 3.68498743e-01 2.55814701e-01
5.71532212e-02 2.21129939e-01 -5.67790568e-01 -2.10000843e-01
-5.89881122e-01 7.22125322e-02 1.29941726e+00 1.26686454e+00
7.51319528e-01 -7.30739236e-01 6.03654087e-01 5.44016719e-01
6.42378151e-01 1.07913770e-01 -4.32072766e-02 -1.09735477e+00
2.47498795e-01 9.48281705e-01 5.85147262e-01 -4.77894634e-01
-3.53499167e-02 3.76405381e-02 -4.26067878e-03 1.38500035e-01
4.12183046e-01 3.67899537e-01 -5.22544906e-02 1.33741212e+00
6.76336825e-01 -3.69899571e-01 9.57422107e-02 7.07971215e-01
7.31270373e-01 1.60001829e-01 2.84083754e-01 -1.69373795e-01
1.33935130e+00 1.23488940e-01 -9.26450372e-01 3.43832880e-01
3.54757249e-01 -5.73541224e-01 4.99744743e-01 7.71396011e-02
-1.02549970e+00 3.37586075e-01 -1.00273192e+00 -1.84700396e-02
-9.40314651e-01 -8.62303495e-01 4.23198760e-01 5.87083280e-01
-8.01052570e-01 5.25637627e-01 -4.33629572e-01 -8.44760418e-01
6.94937408e-02 2.89638311e-01 -5.21877289e-01 1.39978692e-01
-1.32829428e+00 5.68635046e-01 5.18995523e-01 -2.70163625e-01
-2.40881611e-02 -4.09152180e-01 -4.81077492e-01 2.18069881e-01
4.80596066e-01 -5.25477648e-01 8.70086014e-01 -7.99857557e-01
-8.45987856e-01 9.63244915e-01 1.97099954e-01 -4.31641757e-01
5.98073184e-01 2.46936664e-01 -9.09730434e-01 2.67073333e-01
5.22039056e-01 2.56104261e-01 6.25489354e-02 -1.44177401e+00
-1.38591075e+00 -7.40354717e-01 5.07356644e-01 -1.60919935e-01
-4.89292473e-01 2.46059850e-01 -8.22979808e-01 1.95429891e-01
2.51893133e-01 -4.75577295e-01 3.31896931e-01 1.78241223e-01
2.08951160e-02 -3.10833871e-01 8.90487731e-01 -4.98816937e-01
1.39294183e+00 -1.92624199e+00 -2.99562454e-01 5.54167092e-01
3.28061804e-02 1.03329048e-01 2.79030800e-01 8.94107282e-01
3.39783579e-01 7.99677372e-01 8.92535523e-02 6.42865837e-01
3.22221577e-01 4.43331182e-01 -1.73220173e-01 3.45488749e-02
-4.94748026e-01 3.58073056e-01 -9.32799876e-01 -3.74139369e-01
-2.56738067e-01 4.53898638e-01 -1.35969177e-01 -1.06311880e-01
-5.29548049e-01 1.55690581e-01 -5.86775839e-01 8.20831954e-01
5.62622845e-01 -2.39211634e-01 1.00953579e+00 -5.05825162e-01
-4.96766806e-01 5.85168123e-01 -1.24351931e+00 1.67682230e+00
-4.38935608e-01 -7.35604689e-02 3.83336246e-01 -3.37869942e-01
6.24080777e-01 6.40356898e-01 9.25592899e-01 -9.50582922e-01
-2.70000935e-01 8.22621107e-01 -6.52300179e-01 -4.40283418e-01
5.61772227e-01 6.22834921e-01 -2.29380921e-01 6.25858486e-01
-4.13188994e-01 4.14244175e-01 6.28144622e-01 2.60956675e-01
1.21801257e+00 7.92114496e-01 2.35373527e-01 -1.44830868e-01
3.79247308e-01 -2.08719626e-01 5.46010733e-01 3.86980385e-01
2.11222976e-01 4.12005074e-02 7.27838337e-01 -3.70583594e-01
-1.06618655e+00 -9.62876558e-01 -5.00533462e-01 6.91836774e-01
4.93615568e-01 -5.69994926e-01 -6.74673498e-01 -8.81493926e-01
9.22280177e-02 8.29942703e-01 -2.15640217e-01 4.23666984e-01
-2.59974986e-01 -7.91248530e-02 1.12479702e-01 1.80617481e-01
2.50890166e-01 -9.33544219e-01 -1.07624841e+00 1.95089996e-01
-4.80428576e-01 -1.18695915e+00 2.17050642e-01 -1.88748166e-01
-4.58447278e-01 -1.36742222e+00 -2.68066432e-02 -5.39245792e-02
3.97820264e-01 1.44175231e-01 1.14578617e+00 4.04766113e-01
3.81876267e-02 5.89274406e-01 -7.46799171e-01 -2.13421434e-01
-2.70516127e-01 2.06638470e-01 -4.13331687e-01 -5.70467897e-02
4.41515446e-01 -7.20363200e-01 -6.14763141e-01 4.49087799e-01
-1.35834956e+00 -2.81719357e-01 -1.83288291e-01 -1.71788007e-01
5.44984818e-01 8.21269006e-02 7.84639537e-01 -1.00556672e+00
3.56255829e-01 -9.54413474e-01 -9.35889304e-01 4.90482062e-01
-8.60942364e-01 1.47858977e-01 1.48201566e-02 3.92405689e-01
-9.29850161e-01 -4.16331261e-01 3.98103669e-02 1.20898031e-01
-2.19419859e-02 7.52145469e-01 -6.81551933e-01 2.26453707e-01
1.25880823e-01 -6.03163064e-01 -1.57286003e-01 -7.36319005e-01
4.84192461e-01 8.08190048e-01 4.05095220e-02 -6.12198055e-01
4.90435213e-01 7.77135074e-01 7.28027448e-02 -4.32062596e-01
-3.28481674e-01 -5.13160467e-01 -2.48259440e-01 -3.53464335e-01
7.51950324e-01 -1.82847813e-01 -1.17053187e+00 -2.29230076e-01
-1.01598048e+00 1.71300679e-01 -4.92943555e-01 -5.75778112e-02
-4.37432349e-01 2.50256211e-01 -6.40226007e-02 -8.15555513e-01
-1.59615755e-01 -7.78913140e-01 5.99928319e-01 9.50995237e-02
-3.76919210e-01 -4.58478093e-01 -2.79675663e-01 4.83476281e-01
5.33126771e-01 3.07207078e-01 1.06079304e+00 -8.02276790e-01
-8.57282400e-01 -4.49238002e-01 -4.00795788e-01 -2.56596416e-01
1.92969650e-01 1.85584545e-01 -5.88767350e-01 1.88787486e-02
-5.07761657e-01 1.98219657e-01 -2.40074590e-01 -3.84052098e-01
7.81596303e-01 -3.07659417e-01 -4.17152196e-01 1.87158540e-01
1.89800501e+00 3.03007305e-01 9.29395914e-01 8.28094423e-01
2.48622119e-01 1.01696777e+00 5.10042489e-01 5.16331851e-01
5.42764962e-01 9.81459558e-01 8.94728124e-01 4.85492736e-01
-1.54025763e-01 -3.93016368e-01 -1.84671968e-01 4.37668681e-01
-9.18418914e-02 -4.31244582e-01 -7.49477148e-01 5.51324069e-01
-1.89325595e+00 -7.82624245e-01 -3.78107995e-01 2.74260521e+00
4.53543603e-01 9.34529155e-02 -4.06603254e-02 1.08925126e-01
6.40640497e-01 -1.65028855e-01 -4.02798563e-01 -1.38038933e-01
9.34691206e-02 1.62169430e-02 8.58613312e-01 1.91740885e-01
-1.78002343e-01 4.16268140e-01 5.26349783e+00 1.85585946e-01
-6.48437619e-01 4.24839079e-01 -1.54824378e-02 -4.98551689e-02
-9.45122361e-01 6.53687537e-01 -8.96299839e-01 7.73074269e-01
1.23692083e+00 -4.53952104e-01 4.41718251e-01 5.54184258e-01
2.42322206e-01 -3.64607245e-01 -8.85748744e-01 4.04273659e-01
-4.02193725e-01 -1.43878043e+00 -8.19075946e-03 5.30519485e-01
1.99088991e-01 -1.64028779e-02 -5.60503602e-01 -3.74066532e-01
2.13642418e-01 -2.62521207e-01 9.59855795e-01 6.80339158e-01
9.35068846e-01 -6.64608300e-01 4.65775728e-01 3.22210416e-02
-1.12220430e+00 6.04703277e-02 7.01138601e-02 5.91159105e-01
3.61877203e-01 6.42352462e-01 -3.75543565e-01 1.06944811e+00
9.06178594e-01 4.03892389e-03 -7.55569264e-02 1.06013453e+00
-8.50067958e-02 1.03738137e-01 -6.55146480e-01 1.14629455e-01
-1.69398129e-01 -4.54375595e-01 2.49160573e-01 5.26094794e-01
6.45428598e-01 -4.95380424e-02 -3.24881464e-01 5.01948357e-01
-4.24537778e-01 3.85396898e-01 -5.20142078e-01 -1.57405779e-01
9.71817911e-01 1.07764244e+00 -7.88989723e-01 -9.54960734e-02
-7.48050690e-01 6.26172960e-01 -1.70424283e-02 4.46653485e-01
-6.21207356e-01 -5.88987887e-01 7.25943625e-01 9.53991294e-01
1.99586481e-01 9.04124975e-02 1.22106679e-01 -1.06566596e+00
4.93202209e-01 -6.76168621e-01 6.49908066e-01 -9.47664082e-01
-9.37651694e-01 4.52375293e-01 -9.67700407e-02 -1.08949232e+00
-4.71349359e-01 -1.83299854e-01 2.27298051e-01 5.74121356e-01
-1.75979090e+00 -1.10902441e+00 -2.02372223e-01 3.57455522e-01
-2.29894981e-01 3.32526028e-01 9.61273909e-01 7.96832800e-01
1.71788484e-02 -6.58472627e-02 1.47044852e-01 -1.69735178e-01
6.32586896e-01 -8.23543787e-01 2.01645922e-02 6.20758116e-01
8.51178691e-02 6.08425736e-01 5.46780705e-01 -8.74731123e-01
-1.52335489e+00 -4.89538610e-01 1.31929994e+00 -2.08553880e-01
8.27793837e-01 -2.56386455e-02 -7.87850499e-01 7.14858949e-01
2.11903214e-01 2.18641832e-01 8.22692215e-01 4.72231545e-02
-4.71315235e-01 -6.88401759e-01 -1.47542965e+00 3.02180678e-01
1.29149401e+00 -3.95663470e-01 -1.74326196e-01 1.81599095e-01
5.11495471e-01 8.88024345e-02 -1.42179251e+00 4.64102179e-01
7.15579212e-01 -1.10242772e+00 6.74281716e-01 -5.83922386e-01
5.68530299e-02 -5.37202537e-01 -5.55134416e-01 -5.77302754e-01
3.89609545e-01 -6.83302641e-01 -2.33711481e-01 1.73687136e+00
4.18475091e-01 -8.38256419e-01 5.73484123e-01 1.38595927e+00
2.33975083e-01 -1.70728832e-01 -1.09775591e+00 -7.12707341e-01
-5.41392982e-01 -4.76639748e-01 1.25989139e+00 9.67333078e-01
2.42773920e-01 -3.66519243e-01 2.68072128e-01 2.04469249e-01
7.47021079e-01 1.76609784e-01 5.17600536e-01 -1.79976094e+00
1.67427167e-01 -1.42054679e-02 -3.49344850e-01 -3.06740880e-01
-1.27566427e-01 -9.74747419e-01 -7.06142426e-01 -2.18660402e+00
-2.75444061e-01 -1.08102024e+00 -3.15353900e-01 5.11717737e-01
7.02929676e-01 5.86003996e-02 2.35752240e-01 6.55088425e-01
-5.99931419e-01 -2.46869877e-01 5.92381477e-01 3.34436864e-01
4.05096933e-02 -5.77440523e-02 -8.25502872e-01 4.35600549e-01
5.85585535e-01 -6.69438958e-01 -4.05785888e-01 -3.41094285e-01
1.13079214e+00 3.71239871e-01 5.59410639e-02 -8.22764277e-01
4.16093707e-01 -3.11660767e-01 -2.48047709e-01 -3.20056885e-01
1.01996675e-01 -1.51507378e+00 8.72784376e-01 -9.47044045e-02
-2.71711171e-01 -2.89262623e-01 -4.91716951e-01 3.39965135e-01
-1.57395393e-01 -7.50753820e-01 2.01391652e-01 -2.55654931e-01
-6.80563331e-01 3.17546159e-01 -4.24506105e-02 6.69202581e-02
1.11484742e+00 -1.20674185e-01 -3.57607931e-01 -8.59464705e-02
-6.64438128e-01 1.30547017e-01 1.01462364e+00 5.39958298e-01
1.34500461e-02 -1.02083886e+00 -1.44798055e-01 -9.55919176e-02
3.43059689e-01 -4.49637115e-01 -1.71624631e-01 3.49335670e-01
-4.14133370e-01 3.07912767e-01 -2.97666281e-01 2.25787871e-02
-9.57632959e-01 7.13469207e-01 1.36344850e-01 1.79138482e-01
-4.67906058e-01 -1.48086891e-01 -6.90594733e-01 -1.58361524e-01
1.29181266e-01 1.65147051e-01 1.68276913e-02 5.11170268e-01
5.63344479e-01 4.60936129e-01 4.83488709e-01 -5.24965584e-01
-5.06895244e-01 2.33326778e-01 2.96816617e-01 -4.84916180e-01
1.40005016e+00 -4.70938087e-01 -6.17484510e-01 3.16395730e-01
9.85177934e-01 4.61415470e-01 -6.44219279e-01 -1.00387949e-02
4.01877433e-01 -5.57904601e-01 -1.61449730e-01 -8.92982304e-01
-9.93617177e-01 7.33844414e-02 2.99505174e-01 9.92674530e-01
9.70407844e-01 2.22514734e-01 8.69719565e-01 1.52655035e-01
1.12848949e+00 -1.23553693e+00 -8.49714696e-01 -1.87558919e-01
6.20925546e-01 -6.94533646e-01 -8.35141242e-02 -7.39112079e-01
-1.70367196e-01 1.24989522e+00 -8.72692317e-02 6.30237222e-01
6.86036587e-01 3.35677147e-01 6.35738969e-02 -5.59782445e-01
-5.87450862e-01 -3.04719627e-01 -4.35065359e-01 6.85939372e-01
3.91285747e-01 -1.00973450e-01 -1.11710632e+00 3.20130944e-01
2.96401471e-01 5.40627658e-01 4.98078525e-01 1.30768359e+00
-3.94067645e-01 -1.77112758e+00 -1.17244199e-01 2.10694715e-01
-8.10486555e-01 1.28093064e-01 -6.12999126e-03 6.88571393e-01
1.56586125e-01 9.42623019e-01 2.39279076e-01 3.75089169e-01
4.93787766e-01 2.04586953e-01 -1.08294562e-02 -3.33421379e-01
-5.10136127e-01 -3.22747171e-01 6.44476831e-01 -7.26958454e-01
-4.79854256e-01 -6.86186850e-01 -1.78840327e+00 -2.11268082e-01
7.72863161e-04 6.47390664e-01 1.42207730e+00 5.70769012e-01
6.56525910e-01 5.59707843e-02 2.80902177e-01 1.05475023e-01
-1.69497922e-01 -9.42976773e-03 -4.36482847e-01 6.15502417e-01
-4.50692236e-01 -4.06146228e-01 -2.15071946e-01 1.33830696e-01] | [9.106501579284668, 7.748978614807129] |
39e8201b-e9ba-4873-b868-551cad668dcd | optical-flow-for-video-super-resolution-a | 2203.10462 | null | https://arxiv.org/abs/2203.10462v1 | https://arxiv.org/pdf/2203.10462v1.pdf | Optical Flow for Video Super-Resolution: A Survey | Video super-resolution is currently one of the most active research topics in computer vision as it plays an important role in many visual applications. Generally, video super-resolution contains a significant component, i.e., motion compensation, which is used to estimate the displacement between successive video frames for temporal alignment. Optical flow, which can supply dense and sub-pixel motion between consecutive frames, is among the most common ways for this task. To obtain a good understanding of the effect that optical flow acts in video super-resolution, in this work, we conduct a comprehensive review on this subject for the first time. This investigation covers the following major topics: the function of super-resolution (i.e., why we require super-resolution); the concept of video super-resolution (i.e., what is video super-resolution); the description of evaluation metrics (i.e., how (video) superresolution performs); the introduction of optical flow based video super-resolution; the investigation of using optical flow to capture temporal dependency for video super-resolution. Prominently, we give an in-depth study of the deep learning based video super-resolution method, where some representative algorithms are analyzed and compared. Additionally, we highlight some promising research directions and open issues that should be further addressed. | ['Junsong Yuan', 'Baoxin Li', 'Shifu Zhang', 'Yuanzhong Liu', 'Wei Xie', 'Hongyan Li', 'Zhigang Tu'] | 2022-03-20 | null | null | null | null | ['video-super-resolution', 'motion-compensation'] | ['computer-vision', 'computer-vision'] | [ 3.96380097e-01 -5.70791781e-01 -3.86329859e-01 -1.29917106e-02
-4.37708437e-01 -1.64565057e-01 2.55736232e-01 -5.96414745e-01
-2.38462433e-01 1.02243721e+00 3.76780242e-01 3.40217531e-01
4.45367815e-03 -6.34235561e-01 -5.09839594e-01 -8.43025744e-01
-9.14405808e-02 -3.59304428e-01 5.15642047e-01 -3.34903568e-01
4.00934964e-01 4.81347233e-01 -1.81692088e+00 4.37214792e-01
7.43972361e-01 6.91925764e-01 4.42401886e-01 7.54250407e-01
2.77712885e-02 1.02095222e+00 -4.21997130e-01 -6.06697164e-02
1.09482326e-01 -8.03159118e-01 -1.05921650e+00 2.28947639e-01
9.30397928e-01 -9.24151301e-01 -4.39016700e-01 1.27294600e+00
2.49376073e-01 4.77122426e-01 2.01633871e-01 -7.87109137e-01
-6.74941182e-01 4.74121600e-01 -7.00085282e-01 1.48910189e+00
5.25668144e-01 -1.50712699e-01 6.68467462e-01 -9.04873252e-01
8.74063075e-01 1.20431662e+00 1.73812300e-01 7.30755270e-01
-9.70088065e-01 -6.09072566e-01 4.52825986e-02 7.10097671e-01
-1.22124159e+00 -5.34840584e-01 6.87843204e-01 -3.90348405e-01
5.18385828e-01 1.87624454e-01 5.40934801e-01 8.25877845e-01
3.99735093e-01 5.47606587e-01 9.58309531e-01 -1.80070326e-01
2.74716336e-02 -1.53129175e-01 1.23179145e-02 3.88660878e-01
1.40962660e-01 3.46827775e-01 -8.23952675e-01 2.57483333e-01
1.73437870e+00 -6.64219037e-02 -6.07457280e-01 5.64859137e-02
-1.12865531e+00 5.51519036e-01 2.03521267e-01 7.18978405e-01
-3.61933649e-01 -8.69340003e-02 3.50783497e-01 1.36938319e-01
4.69436914e-01 6.34092912e-02 -5.56636825e-02 -1.18417308e-01
-1.27537572e+00 2.56289124e-01 9.72908735e-02 8.00691664e-01
6.83944285e-01 6.75494194e-01 -1.54230639e-01 7.39965320e-01
-9.66574475e-02 8.57525244e-02 6.22070014e-01 -1.67787457e+00
2.74322391e-01 -1.59177721e-01 3.24889839e-01 -9.32862043e-01
-1.44396439e-01 -3.05457532e-01 -1.04124486e+00 3.41624856e-01
3.13810229e-01 6.92376569e-02 -4.27951932e-01 1.62201071e+00
2.61116743e-01 9.64600325e-01 6.15198947e-02 1.39035094e+00
1.00820851e+00 7.82056272e-01 -5.95393442e-02 -9.94361699e-01
1.34641612e+00 -9.56397831e-01 -1.08211517e+00 3.80668230e-02
-1.65442705e-01 -8.28242540e-01 4.55530673e-01 1.37925208e-01
-1.55494094e+00 -1.00833642e+00 -9.21190262e-01 -3.06475699e-01
1.81282789e-01 -3.90149832e-01 4.43472773e-01 1.98528662e-01
-1.26989150e+00 8.24834704e-01 -8.68117213e-01 -2.72975743e-01
1.46642357e-01 1.38192087e-01 -3.55746031e-01 -1.55091792e-01
-1.52493548e+00 7.76822448e-01 4.39358443e-01 -4.19125222e-02
-7.37274170e-01 -7.70409405e-01 -7.50909805e-01 -2.34245360e-02
3.20240289e-01 -8.02412748e-01 1.09947252e+00 -1.21516562e+00
-1.32917070e+00 6.60952687e-01 -7.01389015e-01 -4.20390904e-01
4.46366489e-01 -2.18239009e-01 -6.01199806e-01 6.92376375e-01
1.09623492e-01 3.90101641e-01 1.23866260e+00 -1.17818284e+00
-1.19724000e+00 -2.78438956e-01 2.90811270e-01 4.24835116e-01
-1.67916000e-01 3.73885512e-01 -5.11861980e-01 -8.57730687e-01
-5.38541786e-02 -6.41903579e-01 -8.83787200e-02 -1.92834064e-01
2.42782190e-01 -5.88531792e-02 9.00118113e-01 -9.84856546e-01
1.65431261e+00 -2.02903032e+00 4.36982453e-01 -6.10497773e-01
5.93022048e-01 4.95639116e-01 7.08093643e-02 3.02172322e-02
-2.11497471e-01 -1.82747003e-02 -1.81186289e-01 -1.33031934e-01
-9.56175506e-01 -2.65517794e-02 -2.19767317e-01 3.66569847e-01
-5.47116399e-02 6.69263661e-01 -1.10635638e+00 -7.20909894e-01
6.63984001e-01 1.09496999e+00 -2.90681541e-01 7.32439458e-02
3.54504079e-01 1.01977170e+00 -4.49566990e-01 2.94349194e-01
8.16785991e-01 -2.59687364e-01 -7.32785761e-02 -5.32748520e-01
-7.01025248e-01 -7.87916705e-02 -1.30817246e+00 1.45406246e+00
-1.93982705e-01 9.26343739e-01 2.25298822e-01 -5.81854761e-01
5.09621263e-01 3.79176408e-01 8.32813025e-01 -6.84020758e-01
-9.98250172e-02 6.17566332e-02 -3.48918051e-01 -6.52360559e-01
9.08342659e-01 -1.53589800e-01 6.15445018e-01 2.09622443e-01
-1.77798182e-01 4.17588353e-01 5.26962399e-01 2.34416407e-02
6.28079355e-01 1.48747280e-01 5.83826363e-01 -1.30795809e-02
9.88808036e-01 -2.51178294e-01 8.07393134e-01 2.70234764e-01
-6.94886565e-01 7.31864810e-01 -1.04389213e-01 -6.93445325e-01
-1.21583152e+00 -8.74986529e-01 -4.53406647e-02 9.67369378e-01
6.87024772e-01 -1.36369258e-01 -8.55125308e-01 1.64974947e-04
-4.04876441e-01 1.59477979e-01 -6.26754224e-01 1.61296129e-01
-1.12984943e+00 -5.45047998e-01 7.94666857e-02 6.56013310e-01
9.27207232e-01 -1.08508873e+00 -7.51794875e-01 1.64816558e-01
-7.60432184e-01 -1.67292511e+00 -6.78987682e-01 -6.83816314e-01
-1.33195233e+00 -9.96424079e-01 -9.90192056e-01 -8.81590605e-01
2.82468587e-01 1.11355448e+00 1.01839626e+00 9.59499702e-02
-1.06567815e-01 2.12683856e-01 -3.89680207e-01 4.01826441e-01
-3.31433922e-01 -3.38334024e-01 1.63503647e-01 1.64508343e-01
2.53886282e-01 -6.33637190e-01 -8.00291240e-01 3.24613720e-01
-1.07885551e+00 1.50517464e-01 3.93169731e-01 6.17395937e-01
8.85947824e-01 3.49569649e-01 1.80029482e-01 -6.31040514e-01
5.18568993e-01 -1.89270645e-01 -5.23796380e-01 -2.13685948e-02
-1.00514412e-01 -1.26401350e-01 6.99561238e-01 -2.68888801e-01
-1.42488849e+00 -2.15907618e-01 1.34920090e-01 -7.66059220e-01
-4.03200462e-03 1.65385053e-01 9.13967937e-02 -2.96106517e-01
4.06410694e-01 5.45851409e-01 -1.35673404e-01 -3.79022449e-01
2.85954237e-01 3.24450314e-01 7.02705145e-01 -2.76703648e-02
7.27398694e-01 9.93565500e-01 1.00959852e-01 -1.21339130e+00
-9.00727808e-01 -6.34281218e-01 -1.06308973e+00 -3.73703241e-01
1.09517753e+00 -9.25670981e-01 -4.95235831e-01 4.03903514e-01
-1.14722788e+00 2.65028235e-02 -4.56340574e-02 7.16902554e-01
-6.78219557e-01 6.89942956e-01 -1.03919697e+00 -5.68376064e-01
-2.78210342e-01 -1.29487383e+00 8.11184883e-01 6.61452353e-01
4.52866443e-02 -1.11606193e+00 1.33769810e-01 6.61115825e-01
5.70559144e-01 3.77995759e-01 2.63625026e-01 1.80283770e-01
-9.26754594e-01 5.07994354e-01 -5.18729448e-01 4.39154387e-01
3.75350118e-01 1.09261218e-02 -8.10156524e-01 -4.54963416e-01
3.09582472e-01 3.69652092e-01 7.94928849e-01 9.18310881e-01
8.28470051e-01 -1.48113072e-01 -7.87101984e-02 9.59744692e-01
1.59591997e+00 3.60268414e-01 9.00127947e-01 4.16957885e-01
7.82347918e-01 5.02929330e-01 7.75233448e-01 3.10664892e-01
2.78553247e-01 9.21428084e-01 3.38355720e-01 1.56425625e-01
-7.22433805e-01 7.87852556e-02 2.80535132e-01 8.97107959e-01
-1.04194129e+00 1.94468692e-01 -2.58757919e-01 2.03672200e-01
-1.73111153e+00 -1.62143242e+00 -2.39927366e-01 2.11036444e+00
6.41141951e-01 -2.12667942e-01 1.00270860e-01 1.75270766e-01
1.27537262e+00 5.98210990e-01 -6.24019980e-01 -5.07826395e-02
-3.26079279e-01 -4.08057794e-02 1.98216259e-01 8.65999997e-01
-1.14811814e+00 1.05704117e+00 6.26501942e+00 6.05784357e-01
-1.25183237e+00 2.13764578e-01 4.17317182e-01 -7.17100571e-04
9.66124758e-02 -2.89959639e-01 -8.31475019e-01 6.42077863e-01
6.49214029e-01 -3.80173266e-01 7.34918058e-01 4.12081480e-01
8.56254101e-01 -1.18564174e-01 -8.39869440e-01 1.37430525e+00
2.67199397e-01 -1.53454423e+00 3.45216721e-01 -6.88771158e-02
1.00709212e+00 -3.04355115e-01 8.87562707e-02 -2.38442838e-01
-2.92499095e-01 -8.62801433e-01 3.87223363e-01 4.42606449e-01
1.13136458e+00 -7.63414502e-01 5.54625869e-01 -2.41654739e-02
-1.62815499e+00 -9.84764695e-02 -4.99917030e-01 -8.44485611e-02
5.54356635e-01 5.15363514e-01 8.81128833e-02 7.39257514e-01
1.07151234e+00 1.19245458e+00 -1.99417397e-02 1.03613842e+00
-1.24874584e-01 1.80149183e-01 3.35919350e-01 6.68336570e-01
-6.34098873e-02 -2.52650708e-01 9.30908561e-01 1.13696718e+00
2.06574023e-01 3.93921971e-01 -4.54067960e-02 7.05497980e-01
5.05476892e-02 -1.60442725e-01 -1.12531051e-01 2.42761850e-01
5.16812563e-01 1.16096699e+00 -5.31430304e-01 -4.05231476e-01
-7.32801080e-01 1.17195296e+00 -2.58947220e-02 6.17675304e-01
-7.25285411e-01 -1.00323707e-01 9.50111151e-01 2.34363139e-01
2.76038110e-01 -2.75164783e-01 -3.84218642e-03 -1.40105677e+00
-6.84139803e-02 -6.63627863e-01 5.09624124e-01 -9.43281174e-01
-7.86325872e-01 7.60880172e-01 4.84638773e-02 -1.61080384e+00
-4.94391352e-01 -1.25352994e-01 -2.49839097e-01 1.01677847e+00
-2.00996304e+00 -7.34939873e-01 -8.04648638e-01 7.93557942e-01
1.19488871e+00 2.14738011e-01 2.87883073e-01 4.97346580e-01
-5.75011075e-01 2.11908922e-01 4.86831181e-02 1.73368648e-01
6.80679798e-01 -7.00426102e-01 1.99908167e-01 1.33787179e+00
-1.61278456e-01 4.80022073e-01 9.59029317e-01 -6.20215058e-01
-1.10458517e+00 -9.32403684e-01 6.97596431e-01 -2.76150525e-01
4.50304925e-01 4.43555534e-01 -1.20848644e+00 6.45002961e-01
8.89412165e-02 1.55005008e-01 2.67324567e-01 -7.89021909e-01
-1.02978507e-02 -1.24679409e-01 -1.12459052e+00 3.86226565e-01
1.05901194e+00 -5.02047002e-01 -4.80812222e-01 -2.28647932e-01
7.96562493e-01 -5.71163476e-01 -9.82981682e-01 2.32535914e-01
5.92874110e-01 -1.54722059e+00 1.21324039e+00 -3.03670079e-01
7.28951812e-01 -5.53364158e-01 -2.09158752e-02 -1.01874566e+00
-8.73343229e-01 -5.97257376e-01 -6.68953598e-01 9.83999074e-01
-4.62161303e-01 -3.11100006e-01 6.37622476e-01 4.52728957e-01
6.34667650e-02 -4.99960840e-01 -7.73696244e-01 -5.25330782e-01
-2.24141046e-01 2.73915008e-02 2.16816530e-01 1.17318225e+00
-3.32663208e-01 1.00883231e-01 -5.91486990e-01 2.80810475e-01
1.06837642e+00 3.00856214e-02 2.59340227e-01 -1.01856768e+00
1.77824404e-02 -4.35686499e-01 -4.66750920e-01 -1.15829480e+00
2.95925271e-02 -2.95204461e-01 -4.45723504e-01 -1.52343476e+00
4.50769097e-01 3.78470421e-01 -3.65170658e-01 -3.64512712e-01
-4.68948811e-01 4.12363499e-01 2.33162075e-01 7.25467443e-01
-4.13996607e-01 1.28884241e-01 1.79120517e+00 2.86743730e-01
-3.97709489e-01 8.99765119e-02 -6.75326049e-01 6.54905379e-01
7.47819066e-01 1.35006517e-01 -4.28868800e-01 -6.05347216e-01
-1.42231435e-01 5.61112285e-01 2.24237293e-01 -6.97475016e-01
2.97455192e-01 -4.37063098e-01 3.60330015e-01 -3.71148914e-01
4.23640043e-01 -5.98238468e-01 4.08267342e-02 2.89919466e-01
-3.41179043e-01 2.21636757e-01 -1.96945351e-02 4.80410129e-01
-6.74636126e-01 -3.79453748e-02 1.28645265e+00 -5.05994141e-01
-1.39229286e+00 5.49365520e-01 -2.91553825e-01 2.30807327e-02
1.01630616e+00 -6.13443434e-01 -2.95065373e-01 -4.69965458e-01
-6.99226379e-01 9.27784815e-02 5.06285608e-01 5.16047120e-01
9.87982452e-01 -1.21332347e+00 -9.69425499e-01 -1.81584563e-02
-4.21145201e-01 -1.10151045e-01 8.57548892e-01 7.75234044e-01
-5.68132877e-01 3.83611172e-01 -5.87559640e-01 -4.68456268e-01
-1.65700555e+00 7.65955925e-01 4.74244386e-01 -1.01152800e-01
-8.02753210e-01 6.31380916e-01 3.99893254e-01 8.02707314e-01
8.16378519e-02 -2.41305493e-03 -8.91431153e-01 -6.07499480e-02
1.43412888e+00 1.00632811e+00 -3.34043235e-01 -1.15998042e+00
-2.74394691e-01 1.03625858e+00 -7.71797374e-02 5.40081821e-02
1.10370851e+00 -8.80351305e-01 -3.70610386e-01 3.09219658e-01
1.04779720e+00 -2.21722275e-01 -1.64858472e+00 -2.93291301e-01
-4.17565793e-01 -8.43813300e-01 1.67534053e-01 -2.22058877e-01
-1.37707472e+00 8.98668170e-01 5.73197484e-01 2.31603876e-01
1.47126603e+00 -1.27585977e-01 9.43588436e-01 -4.28235173e-01
5.66260993e-01 -9.70275402e-01 8.72756466e-02 4.29814011e-01
7.91065216e-01 -1.18500948e+00 1.84726924e-01 -8.05650353e-01
-4.77576911e-01 1.34051967e+00 6.33573651e-01 -1.59311756e-01
4.32117701e-01 1.28803283e-01 -5.24170734e-02 2.08145142e-01
-5.19259214e-01 -1.69842094e-01 2.11899459e-01 6.72991037e-01
6.27228618e-01 -2.40311354e-01 -3.21819723e-01 8.76935869e-02
7.24408999e-02 4.28806752e-01 8.33700001e-01 5.97982705e-01
-6.27876103e-01 -6.87091410e-01 -6.46433830e-01 9.61918384e-03
-6.96286619e-01 2.84003243e-02 1.54566333e-01 4.33091581e-01
1.18756397e-02 9.80874062e-01 7.79231340e-02 -3.97150338e-01
1.61954686e-01 -5.53168237e-01 6.28987074e-01 -3.34196031e-01
-1.09683268e-01 9.85679030e-02 -3.95044178e-01 -1.00645173e+00
-1.21534228e+00 -5.33737123e-01 -1.18500650e+00 -5.73742807e-01
1.25789687e-01 2.75357962e-02 1.96596235e-01 8.14386010e-01
1.69618815e-01 6.89238369e-01 6.34585381e-01 -1.24761605e+00
2.22276431e-02 -5.82908452e-01 -6.93224728e-01 5.42197227e-01
7.93811023e-01 -5.77862859e-01 -5.81485748e-01 5.04495084e-01] | [11.003952026367188, -1.9156601428985596] |
83c8a1b4-7786-4fc6-98cd-313ae6e8d6ec | high-resolution-depth-maps-imaging-via | 2104.01530 | null | https://arxiv.org/abs/2104.01530v3 | https://arxiv.org/pdf/2104.01530v3.pdf | High-resolution Depth Maps Imaging via Attention-based Hierarchical Multi-modal Fusion | Depth map records distance between the viewpoint and objects in the scene, which plays a critical role in many real-world applications. However, depth map captured by consumer-grade RGB-D cameras suffers from low spatial resolution. Guided depth map super-resolution (DSR) is a popular approach to address this problem, which attempts to restore a high-resolution (HR) depth map from the input low-resolution (LR) depth and its coupled HR RGB image that serves as the guidance. The most challenging problems for guided DSR are how to correctly select consistent structures and propagate them, and properly handle inconsistent ones. In this paper, we propose a novel attention-based hierarchical multi-modal fusion (AHMF) network for guided DSR. Specifically, to effectively extract and combine relevant information from LR depth and HR guidance, we propose a multi-modal attention based fusion (MMAF) strategy for hierarchical convolutional layers, including a feature enhance block to select valuable features and a feature recalibration block to unify the similarity metrics of modalities with different appearance characteristics. Furthermore, we propose a bi-directional hierarchical feature collaboration (BHFC) module to fully leverage low-level spatial information and high-level structure information among multi-scale features. Experimental results show that our approach outperforms state-of-the-art methods in terms of reconstruction accuracy, running speed and memory efficiency. | ['Xiangyang Ji', 'Zhiwen Chen', 'Debin Zhao', 'Junjun Jiang', 'Xianming Liu', 'Zhiwei Zhong'] | 2021-04-04 | null | null | null | null | ['depth-map-super-resolution'] | ['computer-vision'] | [ 2.56709397e-01 -5.01160085e-01 8.40568170e-02 -5.43459535e-01
-1.17106366e+00 -8.68482292e-02 3.06374401e-01 -6.07298315e-02
-2.16820672e-01 3.86113316e-01 4.50701863e-01 4.08218414e-01
-4.79195058e-01 -1.00108552e+00 -4.48767573e-01 -7.49218345e-01
4.58318800e-01 -2.38231872e-03 4.76575911e-01 -3.61378044e-01
4.05639976e-01 7.04521835e-01 -2.00228095e+00 6.08237445e-01
9.59730446e-01 1.33567834e+00 6.88548386e-01 4.07754064e-01
-3.99153531e-02 9.77433383e-01 -5.60091138e-02 1.96721731e-03
4.80059423e-02 -1.93772569e-01 -7.30543852e-01 2.67306268e-01
6.16109431e-01 -6.49861097e-01 -6.27801180e-01 1.24272835e+00
4.87589628e-01 3.51462632e-01 1.48260280e-01 -9.55755651e-01
-6.75529659e-01 7.22140521e-02 -7.72417486e-01 3.64327461e-01
6.74991846e-01 5.00359423e-02 7.71522522e-01 -1.15154660e+00
5.70135295e-01 1.39817750e+00 1.66583285e-01 3.85158092e-01
-8.06840241e-01 -6.16651952e-01 2.97591478e-01 4.46760148e-01
-1.48775446e+00 -4.47438121e-01 9.23819244e-01 -1.99958637e-01
8.46166670e-01 2.74662197e-01 5.11847436e-01 6.61066651e-01
1.26217470e-01 5.44215262e-01 1.15235472e+00 -5.22220246e-02
3.19066942e-02 -2.45976061e-01 -6.86500296e-02 6.05647504e-01
4.28031497e-02 6.40934780e-02 -8.82777572e-01 1.73160538e-01
1.19231951e+00 4.92457896e-01 -5.51537514e-01 -2.56094545e-01
-1.32894838e+00 5.34522176e-01 8.19273174e-01 4.19933230e-01
-6.28419042e-01 -4.89006937e-02 -1.19042769e-01 -2.49000057e-03
3.70517224e-01 1.09821528e-01 -3.25497597e-01 1.30767137e-01
-4.96402413e-01 4.00836430e-02 -1.10783391e-01 7.81353593e-01
1.14985549e+00 -2.68593788e-01 -1.51861653e-01 9.74822223e-01
5.33321381e-01 3.81547391e-01 5.38813531e-01 -1.24024570e+00
5.81495285e-01 9.45613325e-01 1.38999231e-03 -1.27569294e+00
-3.98501754e-01 -3.37969884e-02 -1.16550434e+00 9.95874330e-02
-1.58928826e-01 6.14397407e-01 -9.52970028e-01 1.40465128e+00
5.97126842e-01 9.32840034e-02 7.34120384e-02 1.40568149e+00
1.37417841e+00 6.08439386e-01 -4.08331484e-01 -2.40237325e-01
1.18698001e+00 -7.88779616e-01 -6.30248606e-01 -2.18813673e-01
8.42924491e-02 -6.04147494e-01 8.42476547e-01 3.77078205e-01
-1.18684375e+00 -7.55895495e-01 -9.48844314e-01 -7.08159328e-01
-2.53226936e-01 -3.10646221e-02 5.52173615e-01 -2.42881626e-02
-9.82314050e-01 4.58528429e-01 -6.80473566e-01 -1.43452168e-01
4.19622213e-01 3.52393329e-01 -7.25813568e-01 -8.42941523e-01
-1.03490353e+00 5.62280476e-01 2.81894594e-01 4.12425935e-01
-7.02773452e-01 -5.52463889e-01 -1.05139291e+00 -2.00604066e-01
3.18814993e-01 -6.52425945e-01 8.08823645e-01 -5.13041914e-01
-1.34055829e+00 7.75413930e-01 -3.64670336e-01 2.73361325e-01
1.27898082e-01 -1.38923645e-01 -4.42304283e-01 3.70025158e-01
1.89051181e-01 4.49730009e-01 6.65412903e-01 -1.46411312e+00
-1.18243206e+00 -8.94404054e-01 2.22979426e-01 6.51944339e-01
-1.65195063e-01 -1.47115797e-01 -7.78174281e-01 -2.95180738e-01
1.05771959e+00 -2.82578766e-01 -2.63648510e-01 -3.43924612e-02
-1.54465467e-01 -5.04220948e-02 6.61171436e-01 -7.07186818e-01
1.05032539e+00 -2.15025282e+00 5.30147910e-01 1.15924165e-01
4.35713559e-01 -1.73619345e-01 -6.44878969e-02 -4.91670519e-02
1.18919551e-01 -2.34200478e-01 -1.44764811e-01 -3.64187509e-01
-4.14362818e-01 2.24908173e-01 -1.13088228e-01 5.29976368e-01
7.96523914e-02 7.82293558e-01 -1.03147221e+00 -4.62103337e-01
6.22170269e-01 8.59711230e-01 -4.02397841e-01 4.61572677e-01
2.02017128e-01 8.24751973e-01 -5.54222226e-01 9.83796179e-01
9.44007993e-01 -3.07050914e-01 -2.99619764e-01 -8.02968919e-01
-2.59690583e-01 1.65156081e-01 -1.29584789e+00 2.25329375e+00
-5.73187590e-01 2.99283773e-01 1.62838139e-02 -4.56725240e-01
1.05052710e+00 -2.53769130e-01 6.30038202e-01 -1.25621963e+00
9.77710187e-02 2.34114617e-01 -4.23265725e-01 -4.36711967e-01
8.96265924e-01 1.33184016e-01 3.47688086e-02 1.15615450e-01
-3.49005908e-02 -1.90161869e-01 -2.90597558e-01 1.52161360e-01
9.16211784e-01 6.03033416e-02 1.86513066e-01 3.21888924e-01
8.04897904e-01 -3.56499910e-01 7.42446065e-01 2.38972515e-01
1.21329064e-02 1.06164074e+00 -1.83233067e-01 -4.08953071e-01
-7.90365219e-01 -9.31772649e-01 -1.69366337e-02 7.82935619e-01
7.77176917e-01 -2.88106441e-01 -2.32103139e-01 -3.31942916e-01
-5.41298464e-02 2.17367470e-01 -7.20331550e-01 -2.43853033e-01
-5.56056499e-01 -6.03092432e-01 -1.85152888e-01 5.28122127e-01
8.78970623e-01 -8.91289234e-01 -6.14787519e-01 5.79206944e-02
-5.51494300e-01 -1.23565865e+00 -3.33904594e-01 -6.24042004e-02
-8.08706284e-01 -1.11105418e+00 -5.27409732e-01 -4.22832817e-01
6.42343581e-01 9.39817369e-01 9.98131454e-01 -1.03858290e-02
-2.58236796e-01 4.00020927e-01 -5.49695134e-01 3.25405777e-01
9.58601758e-02 -2.10839376e-01 -1.90146133e-01 4.11268979e-01
1.64532274e-01 -5.39540827e-01 -8.21370542e-01 4.76085007e-01
-1.18965769e+00 2.60657936e-01 7.97491670e-01 6.39857888e-01
1.08753300e+00 1.90500841e-01 2.60748863e-01 -3.68477106e-01
6.09613582e-02 -3.59985292e-01 -5.42853296e-01 2.61748403e-01
-2.89633363e-01 -1.44724518e-01 3.21819663e-01 5.12600988e-02
-1.09729552e+00 1.22206807e-01 -7.32592344e-02 -9.05959010e-01
-7.67288432e-02 3.07363808e-01 -6.39501870e-01 -2.87519932e-01
9.60282087e-02 5.24195492e-01 -2.92887419e-01 -6.06711030e-01
4.13039684e-01 7.53800392e-01 8.38177383e-01 -2.95086563e-01
6.75184667e-01 7.82740414e-01 5.25640920e-02 -6.18900776e-01
-1.05000091e+00 -6.34391189e-01 -7.23308623e-01 -1.74546421e-01
9.59548771e-01 -1.26507568e+00 -6.02485657e-01 6.48504794e-01
-1.00062585e+00 3.01669333e-02 -2.05353931e-01 5.08544207e-01
-5.67600667e-01 3.01724732e-01 -4.75372434e-01 -6.90009773e-01
-2.06438184e-01 -1.12487996e+00 1.51268005e+00 5.69717109e-01
5.75334013e-01 -5.69238126e-01 -1.84089854e-01 5.83196282e-01
3.64829451e-01 3.44459027e-01 4.89361286e-01 8.67720619e-02
-1.22625029e+00 9.24704075e-02 -7.31614709e-01 2.19533443e-01
3.79306108e-01 -2.49653757e-01 -9.04330254e-01 -2.38554299e-01
5.44854775e-02 -2.46673316e-01 9.58586991e-01 4.05250788e-01
1.30461395e+00 1.02037422e-01 -1.22939192e-01 1.04090166e+00
1.63587558e+00 -8.91713984e-03 7.71134794e-01 4.10529912e-01
1.35345352e+00 5.99524856e-01 1.05818057e+00 5.54220915e-01
9.31310177e-01 8.16664994e-01 7.51001596e-01 -1.28655180e-01
-2.16073826e-01 -1.88318834e-01 1.94058836e-01 8.43981385e-01
-9.83805507e-02 1.99361682e-01 -6.06997132e-01 4.22883272e-01
-2.00056911e+00 -8.32332015e-01 -1.03610739e-01 2.30058742e+00
5.82145691e-01 -2.43261352e-01 -2.24424407e-01 1.43377706e-01
7.80887961e-01 2.55611330e-01 -6.79984629e-01 2.32829109e-01
-5.19864619e-01 -9.65611264e-02 3.53702903e-01 4.66433167e-01
-9.75386620e-01 6.22588754e-01 4.70084524e+00 8.29168022e-01
-1.01934886e+00 2.04325885e-01 4.91549760e-01 -1.87252596e-01
-5.11503398e-01 -2.88512379e-01 -7.90897429e-01 3.68277431e-01
3.21291536e-01 3.44229117e-02 5.63916862e-01 5.57756543e-01
1.74433097e-01 -3.44067127e-01 -1.02927899e+00 1.63477170e+00
3.09265971e-01 -1.23457289e+00 1.11720435e-01 4.38009910e-02
8.33723605e-01 8.53279009e-02 6.07834011e-02 -7.10819885e-02
1.03913896e-01 -9.44808483e-01 5.64163506e-01 9.78797853e-01
8.81577194e-01 -1.04473221e+00 7.35261858e-01 4.46238928e-02
-1.73780644e+00 -3.49346310e-01 -5.69425642e-01 3.16449702e-01
9.75400433e-02 6.66170061e-01 1.07702374e-01 1.15053523e+00
1.14408934e+00 1.25727820e+00 -7.20912457e-01 9.50832784e-01
-2.58460790e-02 -5.29557765e-01 -1.24165662e-01 6.47424102e-01
4.42456082e-02 -1.57651842e-01 2.92207450e-01 6.16790295e-01
4.85797852e-01 6.28955901e-01 5.54263927e-02 7.56930351e-01
9.25696045e-02 -1.45565897e-01 -3.45337182e-01 3.94289434e-01
4.86507237e-01 1.42194438e+00 -5.15802383e-01 -1.89626560e-01
-5.65518618e-01 1.26167238e+00 3.83030325e-01 1.68385491e-01
-6.60864174e-01 -1.40816361e-01 8.48975539e-01 6.45279661e-02
2.86855668e-01 -1.13015302e-01 -2.15037376e-01 -1.40696907e+00
1.63999334e-01 -6.82601750e-01 5.61831176e-01 -1.29941702e+00
-1.28577077e+00 8.36813211e-01 -2.92577773e-01 -1.63111377e+00
8.55639726e-02 -1.14232764e-01 -4.53143045e-02 1.05288219e+00
-1.97183108e+00 -1.33014810e+00 -1.21163559e+00 1.09055173e+00
5.11452317e-01 1.67778820e-01 3.65867972e-01 5.38375139e-01
-4.75611299e-01 2.41101339e-01 -2.68868450e-02 -1.04790434e-01
4.90864426e-01 -9.32608247e-01 7.79436454e-02 9.32103097e-01
-1.53603032e-01 3.41324359e-01 1.36561915e-01 -4.68281090e-01
-1.74263799e+00 -1.37697530e+00 4.19301152e-01 -3.06490898e-01
6.03316315e-02 7.52187893e-02 -1.01026332e+00 2.90563285e-01
-4.91648346e-01 5.28243661e-01 1.96105197e-01 -3.46891433e-01
-4.26030397e-01 -3.91776800e-01 -1.21308029e+00 2.09588245e-01
1.27646136e+00 -7.34592021e-01 -3.39994758e-01 5.67433164e-02
9.60805357e-01 -7.77814567e-01 -1.15680873e+00 7.12021053e-01
4.90557611e-01 -1.48235583e+00 1.13805604e+00 7.52135441e-02
6.35490954e-01 -7.35375285e-01 -8.74475002e-01 -1.00862980e+00
-4.22720164e-01 -1.30779937e-01 -2.63725251e-01 1.31920159e+00
-2.96952367e-01 -5.13038576e-01 4.91253048e-01 6.59518838e-01
-3.43270183e-01 -9.01613235e-01 -9.54449177e-01 -3.22538465e-01
-5.04733086e-01 -3.84899229e-01 7.57133007e-01 8.92204940e-01
-4.68466043e-01 1.69264778e-01 -3.06232423e-01 6.04691267e-01
8.28526437e-01 5.61793268e-01 6.63166344e-01 -1.18916988e+00
2.93348008e-03 -3.42089027e-01 -6.61468327e-01 -1.07302666e+00
-1.56879470e-01 -6.34376049e-01 9.92727503e-02 -1.97025907e+00
3.90152693e-01 -4.91247267e-01 -4.70840394e-01 2.62753427e-01
-2.19617680e-01 5.03391623e-01 4.41687889e-02 3.05707783e-01
-9.44164991e-01 8.61776471e-01 1.31131995e+00 8.66239145e-02
-3.35456848e-01 -3.79465729e-01 -7.60394335e-01 5.78241646e-01
2.45734543e-01 -2.31309682e-01 -3.04566324e-01 -5.49264014e-01
1.73190251e-01 3.51816684e-01 5.09250343e-01 -1.04997563e+00
3.94344479e-01 -2.97563136e-01 6.49204612e-01 -1.04483652e+00
6.27474964e-01 -9.21310604e-01 1.45938754e-01 -4.35731001e-02
-7.25294799e-02 3.74520756e-02 -1.64871037e-01 5.86984813e-01
-4.95973200e-01 4.65984374e-01 9.39836204e-01 -1.95124656e-01
-1.09421158e+00 8.72546017e-01 2.66642570e-01 -1.71946749e-01
8.86696458e-01 -4.12415922e-01 -5.12347460e-01 -2.82588631e-01
-4.67456669e-01 3.01159233e-01 8.26496243e-01 6.67324901e-01
1.43958652e+00 -1.56601465e+00 -6.79573715e-01 4.82163310e-01
4.60666448e-01 7.27152824e-01 9.36112583e-01 9.00658727e-01
-3.00817668e-01 1.86679676e-01 -4.39490795e-01 -8.57699394e-01
-1.11385000e+00 3.59267920e-01 4.17023927e-01 -3.11722327e-02
-7.53387809e-01 9.12393510e-01 4.02667850e-01 -1.75900370e-01
3.45588699e-02 -3.84377003e-01 -3.44937533e-01 6.61253855e-02
9.76251423e-01 3.56349558e-01 3.25577855e-01 -1.14547217e+00
-5.19339561e-01 1.14253128e+00 -4.81981412e-02 3.03539075e-03
1.41931880e+00 -8.44615221e-01 -3.55414838e-01 4.69765395e-01
1.26903117e+00 -3.24415177e-01 -1.44276881e+00 -6.66352630e-01
-5.34199238e-01 -1.13644004e+00 5.81457675e-01 -5.20136476e-01
-1.46747673e+00 7.58506894e-01 8.16317558e-01 -1.70881748e-01
1.68645012e+00 1.58977240e-01 8.57759655e-01 -6.92786602e-03
7.12606013e-01 -7.65982926e-01 2.80305952e-01 2.45984197e-01
9.83806670e-01 -1.54784679e+00 2.21692920e-01 -3.68293673e-01
-5.40199518e-01 1.03147388e+00 9.25797641e-01 -7.19545409e-02
5.13981402e-01 -1.70620367e-01 -1.03154726e-01 -3.39265406e-01
-6.31436765e-01 -4.52129215e-01 4.64958370e-01 6.28090560e-01
-1.56120928e-02 -3.02927941e-01 1.93017349e-01 6.47199810e-01
1.71507180e-01 -1.34253383e-01 4.26862806e-01 8.10846269e-01
-5.96765101e-01 -5.89055240e-01 -4.34842139e-01 3.60199302e-01
-1.10409647e-01 -3.94868515e-02 -1.21763252e-01 3.59753847e-01
2.97136545e-01 1.07455301e+00 2.25144640e-01 -8.30712438e-01
4.04047221e-01 -5.87392926e-01 6.04357958e-01 -4.71646279e-01
-2.62051702e-01 3.19026969e-03 -3.94235909e-01 -1.21431637e+00
-8.36505890e-01 -5.26269495e-01 -1.24261081e+00 -2.51385361e-01
-1.28864616e-01 -3.23519230e-01 6.43347561e-01 9.03137565e-01
4.92210567e-01 6.42988622e-01 1.02122128e+00 -1.27458715e+00
2.07717746e-01 -7.43489504e-01 -7.82336175e-01 4.17598933e-01
6.75238788e-01 -8.95856678e-01 -3.34189326e-01 -2.28271261e-01] | [9.802844047546387, -2.298464298248291] |
3cd69128-86d5-421b-a11f-eb18f7406419 | learning-geometry-image-representation-for-3d | 2011.14289 | null | https://arxiv.org/abs/2011.14289v1 | https://arxiv.org/pdf/2011.14289v1.pdf | Learning geometry-image representation for 3D point cloud generation | We study the problem of generating point clouds of 3D objects. Instead of discretizing the object into 3D voxels with huge computational cost and resolution limitations, we propose a novel geometry image based generator (GIG) to convert the 3D point cloud generation problem to a 2D geometry image generation problem. Since the geometry image is a completely regular 2D array that contains the surface points of the 3D object, it leverages both the regularity of the 2D array and the geodesic neighborhood of the 3D surface. Thus, one significant benefit of our GIG is that it allows us to directly generate the 3D point clouds using efficient 2D image generation networks. Experiments on both rigid and non-rigid 3D object datasets have demonstrated the promising performance of our method to not only create plausible and novel 3D objects, but also learn a probabilistic latent space that well supports the shape editing like interpolation and arithmetic. | ['Yuxuan Liu', 'Yaolin Hou', 'Pengjie Tao', 'Yuchun Huang', 'Lei Wang'] | 2020-11-29 | null | null | null | null | ['point-cloud-generation'] | ['computer-vision'] | [ 5.10825552e-02 4.72973108e-01 3.53447616e-01 -6.81241751e-02
-5.90869069e-01 -6.92282140e-01 9.45117235e-01 -2.59495318e-01
1.95494026e-01 4.19075131e-01 -4.24643680e-02 -1.72406092e-01
3.50508392e-02 -1.31960607e+00 -1.06525064e+00 -6.05955422e-01
3.93048748e-02 1.00641131e+00 7.56736025e-02 3.55503820e-02
2.75146633e-01 1.12285614e+00 -1.55359590e+00 -2.15287283e-01
9.92995322e-01 9.94353354e-01 5.58326300e-03 4.32565331e-01
-3.78899485e-01 -2.17643902e-01 -2.86727279e-01 -5.70371971e-02
6.51485145e-01 -4.84656766e-02 -4.07730341e-01 4.09890473e-01
3.75112653e-01 -2.56881267e-01 -6.43636733e-02 8.55917454e-01
3.04993123e-01 4.69535440e-02 1.11509633e+00 -1.11974978e+00
-8.70174408e-01 7.67198056e-02 -8.06412518e-01 -6.84565783e-01
4.09209907e-01 1.04836471e-01 5.78074455e-01 -1.18545985e+00
7.23323882e-01 1.40789700e+00 6.26486421e-01 3.17390263e-01
-1.18344855e+00 -6.27228200e-01 -9.32794660e-02 -6.27176404e-01
-1.64382279e+00 -5.95549345e-02 1.14467907e+00 -7.23471403e-01
5.19555867e-01 3.13426048e-01 9.34810281e-01 8.23765337e-01
7.80857876e-02 3.73833239e-01 9.43823636e-01 -3.61545473e-01
4.05074239e-01 -1.55191451e-01 -3.12178701e-01 6.48392320e-01
2.32783124e-01 6.85163811e-02 -1.19086593e-01 -4.77400094e-01
1.71213508e+00 1.42325059e-01 8.11187644e-03 -7.19706059e-01
-1.32114673e+00 7.22947896e-01 3.43184680e-01 -1.71086028e-01
-4.91336226e-01 2.36744344e-01 -2.04362005e-01 -3.43420476e-01
8.65395486e-01 3.84210736e-01 -1.24021940e-01 2.22174302e-01
-7.12331474e-01 6.81836903e-01 6.18367136e-01 1.42187095e+00
9.57779527e-01 6.24428280e-02 -1.96077496e-01 5.24579942e-01
5.68229616e-01 7.88805187e-01 -1.35893017e-01 -1.18218148e+00
2.95357138e-01 9.38828528e-01 2.89377421e-01 -1.12569165e+00
-5.91243058e-02 -2.17106432e-01 -1.02223361e+00 5.43482900e-01
5.13726845e-02 1.94209233e-01 -1.06742132e+00 1.42872560e+00
7.71244884e-01 4.04887229e-01 -1.31613314e-01 7.85933852e-01
6.51036501e-01 7.45415509e-01 -4.35347915e-01 6.29485622e-02
1.09503508e+00 -4.24422920e-01 -1.18556790e-01 3.42817366e-01
9.04334784e-02 -6.21052921e-01 9.63120103e-01 -8.68531018e-02
-1.45459747e+00 -3.97163153e-01 -8.70478749e-01 -2.99168497e-01
1.65339351e-01 -2.40462467e-01 6.46528840e-01 2.74047315e-01
-9.16477084e-01 4.34396476e-01 -8.07555974e-01 1.06702566e-01
7.55101919e-01 2.15675339e-01 -2.54448980e-01 1.20794680e-02
-5.71482599e-01 3.43832374e-01 1.52073681e-01 1.16268173e-02
-8.11935484e-01 -1.04044461e+00 -7.88238347e-01 4.32675481e-02
1.39323458e-01 -1.36705613e+00 9.44252908e-01 -3.31024975e-01
-1.63404322e+00 1.05586362e+00 -2.61153162e-01 -6.56626001e-02
6.63242221e-01 3.32553424e-02 3.91254067e-01 1.03665426e-01
7.37028569e-02 7.55276799e-01 9.81032968e-01 -1.54254389e+00
-1.53139472e-01 -5.14246225e-01 -7.03402534e-02 3.88324022e-01
3.07495415e-01 -5.29258013e-01 -3.85360718e-01 -6.95649147e-01
5.92794061e-01 -7.94327259e-01 -3.79424393e-01 2.90974110e-01
-7.53418863e-01 -4.04533833e-01 8.01419437e-01 -2.35025078e-01
3.75446409e-01 -2.12938499e+00 1.40706599e-01 5.84716141e-01
2.92813420e-01 -1.88387990e-01 -5.47243021e-02 1.59630179e-01
1.42216810e-03 5.47142982e-01 -5.03974080e-01 -4.49490309e-01
3.64409313e-02 1.60175011e-01 -5.84813595e-01 3.72241318e-01
4.62946653e-01 1.11175239e+00 -7.94845164e-01 -2.74376720e-01
3.59388351e-01 8.09655845e-01 -8.00255597e-01 2.37887174e-01
-7.30919361e-01 9.08757508e-01 -9.93968129e-01 5.23347616e-01
1.03597462e+00 -2.78761029e-01 -5.13891935e-01 -5.29137179e-02
-2.77163029e-01 2.14056671e-01 -1.19464016e+00 1.99149072e+00
-3.75809997e-01 -2.48683572e-01 -1.29244272e-02 -4.57825363e-01
1.23952341e+00 3.14451873e-01 6.58797383e-01 -2.56422311e-01
-9.55234766e-02 1.96243867e-01 -6.31804883e-01 4.44959337e-03
3.66342276e-01 -2.79762357e-01 -2.00777296e-02 5.36329567e-01
-4.78255153e-01 -1.16250420e+00 -5.98845661e-01 1.02231845e-01
6.09059751e-01 4.49948907e-01 -1.48027807e-01 -3.68344486e-01
2.94730872e-01 -3.12885009e-02 4.26471144e-01 4.82198566e-01
7.09044755e-01 1.13137937e+00 4.17759597e-01 -3.25181544e-01
-1.47145224e+00 -1.55519176e+00 -1.78183556e-01 -9.96657833e-02
3.32256734e-01 -1.09999798e-01 -7.86354601e-01 -2.53268272e-01
2.77307123e-01 9.19766426e-01 -5.23190498e-01 -1.34604014e-02
-7.02830851e-01 -3.55050802e-01 1.25378966e-01 2.82267720e-01
2.89361298e-01 -7.23256588e-01 -3.86264563e-01 -3.88211943e-02
1.54337794e-01 -8.95966470e-01 -6.95807338e-01 -4.75936770e-01
-1.26328635e+00 -8.20764482e-01 -6.22724831e-01 -6.88490927e-01
1.19444871e+00 4.69419122e-01 1.07514882e+00 6.91541284e-03
-2.35437199e-01 4.88948554e-01 -6.64057350e-03 -4.47370529e-01
-4.66777623e-01 -1.68396533e-01 2.10866537e-02 6.83859289e-02
-2.10872442e-01 -1.18061149e+00 -5.48624396e-01 2.97641993e-01
-9.51301098e-01 5.83440244e-01 1.86261535e-01 4.66596246e-01
1.33688331e+00 2.35142216e-01 2.20606133e-01 -6.29949331e-01
2.75152475e-01 -5.26029408e-01 -9.56771612e-01 -1.56423301e-01
-1.58817455e-01 7.73455948e-02 2.85866380e-01 -2.42160395e-01
-9.53988969e-01 2.75483996e-01 -8.85185599e-02 -7.52953708e-01
-2.33376026e-01 3.21103744e-02 -1.97053760e-01 -1.92383230e-02
4.55943584e-01 2.23991394e-01 9.26173702e-02 -7.15995967e-01
6.83706462e-01 3.56505126e-01 3.75660837e-01 -1.03802729e+00
1.36065090e+00 9.58187222e-01 5.03760278e-01 -8.27924550e-01
-5.21769047e-01 5.13213314e-03 -8.17670643e-01 9.18774232e-02
1.08871412e+00 -8.78578901e-01 -6.92001700e-01 4.78636056e-01
-1.65877223e+00 -1.95883825e-01 -6.97642863e-01 1.45952180e-01
-9.45579469e-01 3.60286027e-01 -2.57565588e-01 -7.68992603e-01
-4.25275207e-01 -1.23209870e+00 1.59451962e+00 -4.27219160e-02
-8.30753222e-02 -7.56250620e-01 -7.17996657e-02 1.26037017e-01
5.57617247e-02 8.89968634e-01 1.30598664e+00 8.58585238e-02
-1.36381269e+00 -1.43990651e-01 -2.40383044e-01 8.31223577e-02
3.38109136e-01 1.33100078e-01 -8.09571207e-01 5.26657738e-02
2.26026818e-01 3.58068906e-02 2.13246122e-01 4.96150374e-01
1.48069394e+00 -3.35443199e-01 -3.32847893e-01 9.80327904e-01
1.33591902e+00 -4.97283787e-02 5.52598298e-01 -1.92616776e-01
1.02889645e+00 5.84303141e-01 1.56322747e-01 4.90823418e-01
5.03817797e-01 7.85284162e-01 5.97247601e-01 -1.07919596e-01
-2.15154305e-01 -7.36151516e-01 -8.17578286e-02 8.45158637e-01
-4.22895461e-01 5.75436465e-02 -9.22796309e-01 2.72385210e-01
-1.47828496e+00 -5.38893759e-01 -4.09268111e-01 2.44944167e+00
6.87875628e-01 -2.16002539e-01 -7.12944642e-02 -1.82059661e-01
7.19725668e-01 -1.92856863e-01 -6.68996871e-01 1.18846437e-02
-1.36289626e-01 3.35757405e-01 1.85031369e-01 5.47805965e-01
-4.99090225e-01 8.12170565e-01 6.09371662e+00 7.57353902e-01
-8.04903209e-01 -2.60784715e-01 6.38089597e-01 -6.41305745e-02
-1.05830228e+00 -2.33289488e-02 -8.63330364e-01 3.64065379e-01
2.43787333e-01 -3.36120933e-01 4.25144136e-01 8.35760593e-01
3.89734119e-01 2.94164896e-01 -1.17469525e+00 1.21132016e+00
-2.20219225e-01 -1.72469008e+00 7.17270494e-01 4.57937658e-01
1.05263209e+00 -1.27858266e-01 1.74656942e-01 -3.37671965e-01
4.12983179e-01 -1.25476694e+00 1.15800905e+00 6.36055648e-01
1.20064938e+00 -7.58138597e-01 -5.07966541e-02 6.71186626e-01
-1.02153969e+00 6.03212178e-01 -4.78366196e-01 3.61838564e-02
2.88472980e-01 1.02169061e+00 -8.55120778e-01 6.80201113e-01
5.66973925e-01 6.13763869e-01 -8.41297209e-02 9.25489724e-01
-3.52446020e-01 1.46577165e-01 -7.34493911e-01 4.50487703e-01
1.92838497e-02 -7.61608541e-01 9.14448380e-01 5.56789875e-01
7.98285127e-01 4.25091237e-01 3.70088816e-02 1.75086570e+00
-1.92398652e-01 -5.02538532e-02 -1.06192231e+00 2.53646731e-01
7.71570921e-01 1.03346074e+00 -6.52677774e-01 -2.40588952e-02
9.16966889e-03 7.64304221e-01 1.45614415e-01 3.75947952e-01
-6.38403535e-01 -1.27603754e-01 7.88027883e-01 4.99781698e-01
3.67509812e-01 -8.23774517e-01 -8.22838187e-01 -9.18717921e-01
3.20512295e-01 -3.36186796e-01 -2.65005589e-01 -1.10412061e+00
-1.46191895e+00 4.73708034e-01 1.57529369e-01 -1.16523755e+00
-9.12608877e-02 -4.36910421e-01 -6.52355909e-01 1.30829203e+00
-1.48423064e+00 -1.32385468e+00 -2.59290606e-01 5.89674890e-01
1.32951766e-01 1.95719674e-01 6.92321301e-01 -3.01179022e-01
2.44056080e-02 6.09746724e-02 -1.42646343e-01 -1.75745681e-01
-3.08868978e-02 -1.01762187e+00 9.56552625e-01 4.62151945e-01
-5.43421879e-02 6.93704307e-01 1.14159621e-01 -8.71416688e-01
-1.69551814e+00 -1.21045351e+00 5.06867826e-01 -9.08188462e-01
1.21331200e-01 -7.44849563e-01 -9.07629490e-01 5.94002306e-01
-3.70291442e-01 -1.13500342e-01 2.86610425e-01 -3.65886837e-01
-3.70565474e-01 2.51544386e-01 -1.30965984e+00 7.26807296e-01
1.48901796e+00 -2.91032821e-01 -5.13515592e-01 5.41413069e-01
1.07603467e+00 -8.03250015e-01 -9.58230913e-01 5.22528827e-01
1.47278860e-01 -7.89831638e-01 1.41216385e+00 -3.36632490e-01
5.26844382e-01 -5.45130491e-01 -6.58064634e-02 -1.22720349e+00
-3.65942240e-01 -8.63375068e-01 -3.75342295e-02 1.08603144e+00
1.43442661e-01 -5.47920287e-01 9.06275749e-01 8.28319609e-01
-4.94598299e-01 -7.48049855e-01 -1.11882913e+00 -7.03378499e-01
4.30828124e-01 -6.20910048e-01 1.17607772e+00 8.50954831e-01
-7.51635611e-01 1.07945839e-03 1.07706726e-01 3.69528353e-01
9.09903944e-01 4.56361860e-01 1.00027585e+00 -1.51215518e+00
-1.02127805e-01 -3.82593215e-01 -2.51126438e-01 -1.41582584e+00
1.39860898e-01 -1.26343012e+00 -6.15887865e-02 -1.56752825e+00
-1.17634088e-01 -1.15837204e+00 4.55583334e-01 8.30496475e-02
9.77087915e-02 6.19925670e-02 -7.13614821e-02 4.81189877e-01
9.84813944e-02 9.70827699e-01 1.90192449e+00 2.00951621e-01
-3.92743230e-01 1.14834398e-01 -7.70394742e-01 7.86453068e-01
3.34580064e-01 -2.98274308e-01 -4.02485251e-01 -9.41724002e-01
3.26834738e-01 9.79892090e-02 6.25811994e-01 -5.92043340e-01
-1.84399173e-01 -3.83956701e-01 3.81687194e-01 -9.76396978e-01
5.31350374e-01 -8.45069706e-01 5.66980004e-01 -1.53263658e-01
1.19181670e-01 -1.04966693e-01 5.66362590e-02 5.61785340e-01
2.33024999e-01 4.10593115e-02 8.47897887e-01 -4.12790507e-01
1.06697351e-01 8.87176514e-01 2.08307445e-01 -5.86216636e-02
1.12233484e+00 -5.13113618e-01 4.25818153e-02 -1.10226616e-01
-4.15142000e-01 6.52876869e-02 1.09484768e+00 2.79104382e-01
9.17549670e-01 -1.71055579e+00 -8.15174878e-01 6.34270430e-01
-4.45016436e-02 1.30366957e+00 9.77115184e-02 1.56355962e-01
-6.48395836e-01 1.57467768e-01 6.74687838e-03 -1.01948965e+00
-5.50445318e-01 3.61353964e-01 3.00430626e-01 2.61498988e-01
-1.15278494e+00 8.16861868e-01 6.84306204e-01 -7.41512120e-01
-1.25700638e-01 -5.58110297e-01 2.98134714e-01 -5.24191439e-01
2.69635558e-01 3.54138643e-01 -1.39639333e-01 -5.38263440e-01
-3.96709368e-02 1.12681508e+00 3.33601892e-01 -2.45077580e-01
1.49548137e+00 1.43228322e-01 -4.09619451e-01 3.20055515e-01
9.17842746e-01 1.35049105e-01 -1.42195094e+00 -9.60484147e-02
-3.54537785e-01 -7.55162001e-01 -3.92096266e-02 -1.70521975e-01
-8.38847518e-01 8.96455109e-01 -3.77325788e-02 -4.62476024e-03
5.92277586e-01 2.62896001e-01 8.62608552e-01 2.48924531e-02
7.30246186e-01 -4.80630219e-01 1.28865140e-02 6.72967315e-01
1.37886524e+00 -5.23768783e-01 -5.10026366e-02 -9.22812045e-01
-1.32659063e-01 1.10015821e+00 2.44562164e-01 -5.01937687e-01
7.33234107e-01 1.01676956e-02 -3.32716256e-01 -4.33784664e-01
-2.38915861e-01 4.25124288e-01 4.10728216e-01 7.88945258e-01
4.03737463e-02 6.68343678e-02 2.13202789e-01 3.63421202e-01
-5.29296875e-01 8.16415343e-03 4.29787964e-01 5.67857027e-01
-1.31758243e-01 -1.17907536e+00 -5.34755707e-01 2.49200374e-01
2.38384932e-01 5.47766872e-02 -1.10196181e-01 4.49811429e-01
1.42428562e-01 2.95844913e-01 4.39511895e-01 -1.11852638e-01
4.96121854e-01 -5.11659607e-02 6.37996972e-01 -1.03681505e+00
6.27865642e-02 -1.65569130e-02 -4.57709312e-01 -5.08460224e-01
-1.06613845e-01 -6.50699914e-01 -1.53500187e+00 -2.04624295e-01
-2.75895558e-02 6.82242066e-02 1.01500547e+00 5.65912664e-01
8.43464911e-01 -1.92996487e-02 9.36026514e-01 -1.34754169e+00
-5.51240087e-01 -5.11695802e-01 -6.14699543e-01 5.21044672e-01
1.48224518e-01 -9.03421283e-01 -2.98886448e-01 6.06741309e-02] | [8.720512390136719, -3.6427557468414307] |
04aa2270-dcd2-45ef-8074-56d52bb48424 | mv-fcos3d-multi-view-camera-only-4d-object | 2207.12716 | null | https://arxiv.org/abs/2207.12716v1 | https://arxiv.org/pdf/2207.12716v1.pdf | MV-FCOS3D++: Multi-View Camera-Only 4D Object Detection with Pretrained Monocular Backbones | In this technical report, we present our solution, dubbed MV-FCOS3D++, for the Camera-Only 3D Detection track in Waymo Open Dataset Challenge 2022. For multi-view camera-only 3D detection, methods based on bird-eye-view or 3D geometric representations can leverage the stereo cues from overlapped regions between adjacent views and directly perform 3D detection without hand-crafted post-processing. However, it lacks direct semantic supervision for 2D backbones, which can be complemented by pretraining simple monocular-based detectors. Our solution is a multi-view framework for 4D detection following this paradigm. It is built upon a simple monocular detector FCOS3D++, pretrained only with object annotations of Waymo, and converts multi-view features to a 3D grid space to detect 3D objects thereon. A dual-path neck for single-frame understanding and temporal stereo matching is devised to incorporate multi-frame information. Our method finally achieves 49.75% mAPL with a single model and wins 2nd place in the WOD challenge, without any LiDAR-based depth supervision during training. The code will be released at https://github.com/Tai-Wang/Depth-from-Motion. | ['Wenwei Zhang', 'Xinge Zhu', 'Chenming Zhu', 'Qing Lian', 'Tai Wang'] | 2022-07-26 | null | null | null | null | ['stereo-matching-1'] | ['computer-vision'] | [-2.20387325e-01 -6.67370409e-02 -2.89462924e-01 -3.31463218e-01
-8.33753049e-01 -7.39898443e-01 5.41234553e-01 -4.67597485e-01
-3.99593651e-01 -4.01476286e-02 -4.01657522e-02 -1.99337855e-01
3.79449815e-01 -7.77431965e-01 -9.07938421e-01 -2.88511068e-01
3.69118869e-01 5.21525681e-01 9.15304244e-01 -3.21366400e-01
1.58046052e-01 5.71116745e-01 -1.84517503e+00 3.82319123e-01
3.47796142e-01 1.16809583e+00 5.68777025e-01 8.66530418e-01
1.17172539e-01 4.38051671e-01 1.83390632e-01 -8.65664184e-02
7.84802914e-01 1.66179672e-01 -3.95521849e-01 3.32420439e-01
1.10078168e+00 -8.67535055e-01 -6.88383460e-01 7.23800361e-01
6.52616501e-01 -5.89571409e-02 4.39037383e-01 -1.26946449e+00
-9.49829593e-02 -2.85629630e-01 -8.86081338e-01 2.80250520e-01
8.32102120e-01 4.55004632e-01 8.68756473e-01 -1.51754272e+00
8.52546096e-01 1.45583165e+00 7.46983469e-01 4.65313584e-01
-8.15499485e-01 -5.78458965e-01 2.89023012e-01 1.46998493e-02
-1.30008829e+00 -4.45320547e-01 7.38972545e-01 -4.88845468e-01
1.35335827e+00 -2.31038053e-02 7.49788046e-01 9.88002360e-01
-5.92345521e-02 1.06445289e+00 9.43745852e-01 -2.00432464e-01
-2.35722542e-01 -1.80447847e-02 -7.11393282e-02 1.03287470e+00
1.14859045e-01 6.10368371e-01 -7.22766340e-01 2.73547947e-01
9.52689707e-01 3.30282658e-01 -2.40154132e-01 -1.01764429e+00
-1.03703594e+00 5.74554980e-01 7.16808856e-01 -2.41340727e-01
5.15973940e-02 1.45794570e-01 1.49146944e-01 8.57224092e-02
6.71299994e-01 -8.58320147e-02 -6.62958622e-01 7.12979957e-02
-7.50532508e-01 2.55843908e-01 2.77940780e-01 1.36464190e+00
1.01903629e+00 -2.44744644e-01 2.81002879e-01 4.63359684e-01
4.99992162e-01 8.29866827e-01 -8.80682543e-02 -1.30531132e+00
8.40619028e-01 8.56939435e-01 2.10100248e-01 -5.41819632e-01
-5.48359632e-01 -2.66798019e-01 -2.55154222e-01 5.06261706e-01
2.75693089e-01 6.77655116e-02 -1.17715216e+00 1.17093289e+00
7.94940770e-01 -3.87713388e-02 -2.26283222e-01 1.31088793e+00
9.82030571e-01 3.48773122e-01 -6.92107916e-01 4.54960257e-01
1.21821868e+00 -1.24432683e+00 -3.82987261e-02 -5.89216590e-01
7.97023594e-01 -7.26539135e-01 8.79472613e-01 2.89277613e-01
-1.21839392e+00 -5.96591592e-01 -1.03097689e+00 -6.89394176e-01
-3.71706456e-01 8.66729319e-02 4.42040205e-01 4.65693831e-01
-1.18456364e+00 3.90300574e-03 -8.45780134e-01 -4.84689951e-01
4.84451652e-01 1.32759467e-01 -5.82054377e-01 -5.42016327e-01
-8.43870103e-01 9.29886162e-01 3.08312178e-01 -6.81372955e-02
-1.06569266e+00 -8.28991830e-01 -1.19142282e+00 -2.85183311e-01
6.60222590e-01 -1.28631675e+00 1.27330685e+00 -3.11930388e-01
-1.10938740e+00 1.55917239e+00 -3.01741004e-01 -3.42061043e-01
8.21431458e-01 -5.70728779e-01 -3.57214473e-02 2.82970309e-01
3.99439961e-01 1.00397491e+00 6.14258826e-01 -1.33851182e+00
-1.08336854e+00 -8.11057210e-01 3.73414963e-01 5.45497358e-01
1.79585963e-01 -1.58109099e-01 -9.76577699e-01 -2.10398167e-01
6.06373966e-01 -9.64897215e-01 -2.49991313e-01 5.49048603e-01
-4.11623210e-01 -5.70147261e-02 9.87360775e-01 -2.58421987e-01
6.22753561e-01 -2.06942081e+00 4.78520468e-02 -1.79395095e-01
4.12008971e-01 3.34674120e-02 -7.96953738e-02 3.15113753e-01
2.35636532e-01 -3.11620235e-01 3.40733351e-03 -6.87083721e-01
-1.05187275e-01 5.61947562e-03 -1.99174330e-01 5.93060732e-01
1.47551194e-01 8.31604481e-01 -1.01086676e+00 -3.06839108e-01
7.57075906e-01 4.61152941e-01 -9.23844218e-01 3.18177879e-01
-2.05658972e-01 4.07109737e-01 -3.09130996e-01 9.57318902e-01
9.84741688e-01 -2.10651323e-01 -3.08640540e-01 -3.57082754e-01
-4.28304315e-01 3.90038490e-01 -1.20845008e+00 2.31905746e+00
-6.23219013e-01 6.16203129e-01 2.77064055e-01 -5.03942847e-01
7.73962855e-01 -1.56945571e-01 4.67178434e-01 -6.71876609e-01
1.43896401e-01 4.28779185e-01 -6.05524719e-01 -2.86853313e-01
4.72664654e-01 2.58058548e-01 -6.64661229e-02 1.10226229e-01
9.73521098e-02 -5.82059622e-01 -3.84802856e-02 2.41092220e-01
1.11642849e+00 5.21123350e-01 2.19405219e-01 1.40959233e-01
4.87940550e-01 1.67824537e-01 4.31291789e-01 6.13114655e-01
-1.03768103e-01 1.11505413e+00 -5.04277870e-02 -5.22682130e-01
-1.03225946e+00 -1.39250302e+00 -1.45293161e-01 6.70972168e-01
5.46421587e-01 -4.59993750e-01 -2.27206632e-01 -8.26517344e-01
3.19102734e-01 3.12747717e-01 -3.38720024e-01 2.10168194e-02
-5.42398453e-01 -1.48091525e-01 2.95467854e-01 6.48538589e-01
5.97529352e-01 -2.42093936e-01 -1.04192185e+00 -2.00395584e-02
-2.91109607e-02 -1.58538640e+00 -6.26122892e-01 4.15662944e-01
-9.59015369e-01 -1.30905831e+00 -7.18358457e-01 -6.00941181e-01
3.48024964e-01 1.21015704e+00 1.06971204e+00 -1.84425429e-01
-3.97398412e-01 6.67489469e-01 -2.88573980e-01 -2.94301987e-01
1.19434401e-01 5.12056192e-03 1.06865518e-01 -4.20314908e-01
4.05131459e-01 -5.77667117e-01 -1.00111043e+00 5.96353889e-01
-3.60502571e-01 2.61076659e-01 4.42931145e-01 6.19086385e-01
6.71522617e-01 -5.81304014e-01 -3.10641885e-01 -4.05888706e-01
-5.61115384e-01 -3.39202702e-01 -8.33320379e-01 -1.95520073e-01
-3.87416840e-01 -2.94076204e-01 9.87357050e-02 1.00047395e-01
-8.44840527e-01 5.77402771e-01 -1.33371294e-01 -1.19593549e+00
-2.16418311e-01 -3.16324323e-01 -3.46746475e-01 -1.91715688e-01
6.20000780e-01 2.64798731e-01 -1.02804027e-01 -6.46341562e-01
6.11895621e-01 6.45876586e-01 6.26092672e-01 -1.77735269e-01
9.59733844e-01 1.12639880e+00 -1.04278460e-01 -5.94170034e-01
-1.33034110e+00 -1.02478266e+00 -9.92864311e-01 -2.85701215e-01
9.92364407e-01 -1.72008038e+00 -3.89599502e-01 5.28656602e-01
-1.32482076e+00 -2.47471496e-01 -9.83377844e-02 5.91145039e-01
-7.70127475e-01 3.27989817e-01 -4.59377348e-01 -7.23000228e-01
-1.88077226e-01 -1.10311210e+00 1.76597786e+00 -3.22341397e-02
2.74072319e-01 -6.09832168e-01 -3.78464460e-02 8.06268454e-01
-1.10394202e-01 1.19843669e-01 1.12488762e-01 -6.52834494e-03
-1.16129816e+00 -2.14452580e-01 -3.99170488e-01 2.31920987e-01
-1.78538486e-01 -3.17781717e-01 -1.40057385e+00 -3.02688420e-01
-1.75713692e-02 -3.75764370e-01 1.28157771e+00 3.64784449e-01
8.30055296e-01 3.62340271e-01 -5.38314760e-01 1.18332410e+00
1.31274784e+00 -1.19267806e-01 1.79506063e-01 4.94367659e-01
1.10555780e+00 4.00118679e-01 8.76945972e-01 4.34622228e-01
9.82409656e-01 9.68308806e-01 1.00401664e+00 -2.03886598e-01
-3.60894114e-01 -5.81476390e-01 4.82323647e-01 3.36475194e-01
-9.64877661e-03 -1.87111020e-01 -1.05328906e+00 4.93688911e-01
-1.69996095e+00 -7.97892094e-01 -4.88944799e-01 1.98940563e+00
2.05289900e-01 3.52598816e-01 1.95190713e-01 -8.76231268e-02
5.30756891e-01 2.84966975e-01 -7.36232340e-01 1.18145712e-01
-1.95139304e-01 -2.06385851e-01 8.46544385e-01 6.44259870e-01
-1.21301508e+00 1.04416144e+00 5.07661581e+00 5.90981305e-01
-8.52564991e-01 2.93252885e-01 1.92343861e-01 -4.87890691e-01
-2.57856250e-01 1.25764117e-01 -1.33702123e+00 1.44678280e-01
2.44000673e-01 3.35147113e-01 -3.37646762e-03 1.01877391e+00
3.29787970e-01 -3.00653070e-01 -1.24661219e+00 1.39556539e+00
8.32072049e-02 -1.44756591e+00 -2.27705210e-01 1.83308125e-01
7.46507406e-01 7.83361614e-01 -7.75275975e-02 1.06040269e-01
1.62250817e-01 -4.72663164e-01 9.73004699e-01 9.38302875e-02
8.77772868e-01 -4.50121105e-01 3.76664311e-01 7.75450170e-01
-1.58667576e+00 -1.50611445e-01 -3.72800767e-01 -1.74965337e-01
3.70151788e-01 6.64900661e-01 -7.30898798e-01 7.44388759e-01
1.00039411e+00 1.22238708e+00 -4.56644714e-01 1.04012263e+00
-2.18295097e-01 -1.67770565e-01 -7.10445762e-01 5.92962563e-01
2.69368768e-01 7.65383691e-02 7.97140241e-01 9.00665760e-01
5.73304176e-01 -9.68057215e-02 4.10444587e-01 6.44761801e-01
8.32286328e-02 -5.70766568e-01 -9.80095506e-01 7.67501891e-01
4.83972847e-01 1.04380214e+00 -4.83422101e-01 -2.71271527e-01
-7.50481963e-01 1.19672024e+00 2.88877845e-01 4.43163626e-02
-8.48819971e-01 6.42919689e-02 8.83854091e-01 6.43208981e-01
6.44865572e-01 -4.73100036e-01 -1.83449194e-01 -1.53256822e+00
2.32205376e-01 -3.00127566e-01 4.01456267e-01 -1.08225143e+00
-1.19062197e+00 3.55894327e-01 -1.24093276e-02 -1.61249876e+00
-8.56649429e-02 -8.21919858e-01 -3.45635355e-01 6.79120600e-01
-1.90137076e+00 -1.34198070e+00 -5.52780330e-01 8.02109301e-01
7.42765903e-01 1.38226375e-01 3.22790772e-01 5.14528394e-01
-2.44985417e-01 3.37073952e-01 -2.85313040e-01 -5.96676916e-02
8.01563680e-01 -1.19773364e+00 6.85366154e-01 8.56032491e-01
1.12138815e-01 4.13138382e-02 3.19785118e-01 -4.33944434e-01
-1.72911191e+00 -1.16551256e+00 7.60895491e-01 -1.05111039e+00
3.84688020e-01 -6.77986979e-01 -4.24254596e-01 6.22083843e-01
-2.37681732e-01 4.97022122e-01 7.55082890e-02 -2.18313009e-01
-5.63835740e-01 -1.46889821e-01 -9.76849258e-01 2.67871112e-01
1.94336903e+00 -6.46029592e-01 -5.11621654e-01 3.47733170e-01
9.57238972e-01 -1.10069489e+00 -6.69066548e-01 4.89578754e-01
4.81720835e-01 -1.40523112e+00 1.50011396e+00 -2.30755523e-01
2.20404610e-01 -6.48613214e-01 -4.79837388e-01 -8.58187437e-01
5.40840328e-02 -3.60552043e-01 -3.24715227e-01 6.71325564e-01
1.22488521e-01 -3.49733472e-01 1.15315843e+00 1.24895461e-01
-6.09855473e-01 -6.04013562e-01 -1.06997645e+00 -9.52901363e-01
-3.68008137e-01 -8.69195521e-01 2.29626402e-01 6.11065090e-01
-5.24107337e-01 4.70010340e-01 -2.75092304e-01 4.80696678e-01
8.00383687e-01 4.90098536e-01 1.31162798e+00 -1.03385675e+00
-2.35875025e-01 -1.26814276e-01 -6.33813798e-01 -1.85932755e+00
-2.24865571e-01 -9.95263696e-01 -1.29672334e-01 -1.53469574e+00
-2.02900395e-02 -2.19842210e-01 9.34021398e-02 1.96502075e-01
2.77934745e-02 3.94781739e-01 3.39867711e-01 2.91574210e-01
-7.45232284e-01 6.36185348e-01 1.32541108e+00 1.73610881e-01
-7.67371878e-02 -9.81327984e-03 -2.65776426e-01 1.01807380e+00
4.78356659e-01 -4.56645757e-01 -4.26525593e-01 -7.99209118e-01
1.63581088e-01 2.71649837e-01 9.41289842e-01 -1.11431265e+00
3.77141088e-01 7.38897473e-02 5.24847269e-01 -1.39743555e+00
8.55669022e-01 -9.29847956e-01 -1.51816085e-01 4.45614845e-01
3.14776540e-01 1.08535849e-01 1.06852926e-01 7.21830785e-01
-1.00157216e-01 2.30724528e-01 6.84561074e-01 -4.53066856e-01
-1.24430621e+00 7.18975961e-01 5.55175357e-02 2.35453978e-01
1.10563111e+00 -5.81540704e-01 -2.48782292e-01 -1.06878646e-01
-7.41887808e-01 6.81741476e-01 9.52697396e-01 6.28714025e-01
8.67568493e-01 -1.24181533e+00 -5.02196014e-01 2.63739616e-01
4.89350408e-01 7.10051358e-01 5.12109697e-01 8.26795697e-01
-5.21107137e-01 5.27054250e-01 5.36767729e-02 -1.18582487e+00
-1.04610014e+00 4.87341315e-01 5.16121328e-01 -2.13289354e-02
-9.60872710e-01 9.42416668e-01 5.72549582e-01 -7.10838318e-01
1.16477475e-01 -4.42457825e-01 2.97097176e-01 -3.47415991e-02
3.71570557e-01 3.16320449e-01 2.49522492e-01 -5.37951827e-01
-6.52236223e-01 1.11526060e+00 6.65399209e-02 -5.34636853e-03
1.29351103e+00 -5.61102033e-01 3.67756814e-01 4.10465598e-01
1.33199227e+00 -2.64492869e-01 -1.76403534e+00 -3.63275349e-01
-3.08095306e-01 -8.00646484e-01 3.72689068e-01 -4.48027343e-01
-9.88978446e-01 1.16962886e+00 7.46525943e-01 -4.12999898e-01
8.93914521e-01 3.13917696e-01 8.78884494e-01 3.89432102e-01
6.76976621e-01 -8.53141487e-01 1.64642364e-01 7.67534256e-01
6.80333078e-01 -1.74506712e+00 -6.28863871e-02 -6.86165810e-01
-2.89570004e-01 1.15260553e+00 1.06121361e+00 -2.28779525e-01
6.65873170e-01 3.04493546e-01 -2.43308544e-02 -4.26159739e-01
-8.52861643e-01 -6.17419302e-01 2.92667478e-01 7.76035726e-01
-2.84189787e-02 -2.26004139e-01 4.72854525e-01 2.33803540e-01
-1.37622757e-02 -1.32255167e-01 3.62432063e-01 9.44676697e-01
-5.75700521e-01 -8.68517876e-01 -3.27718318e-01 1.15399808e-01
1.62233368e-01 -5.73163144e-02 -1.97877154e-01 9.76170003e-01
4.06464756e-01 7.36990929e-01 1.38252020e-01 -5.64062536e-01
7.23337591e-01 -1.99814245e-01 7.07801998e-01 -8.75387609e-01
-1.86900243e-01 5.95009811e-02 7.88691416e-02 -1.18830776e+00
-4.33862299e-01 -7.19969869e-01 -1.01248085e+00 -3.09638143e-01
-2.19178990e-01 -6.22167826e-01 5.48800588e-01 6.85565174e-01
5.68829179e-01 7.71827064e-03 7.47585177e-01 -1.52870154e+00
-3.16791534e-01 -4.02808368e-01 -2.96238154e-01 -3.84784043e-02
6.72164142e-01 -9.06510174e-01 -4.13328111e-01 -2.91231722e-01] | [7.836630821228027, -2.6108896732330322] |
1ba901bb-f115-4bc3-a6eb-7f67d96b9b18 | application-of-knowledge-distillation-to | 2210.16611 | null | https://arxiv.org/abs/2210.16611v2 | https://arxiv.org/pdf/2210.16611v2.pdf | Application of Knowledge Distillation to Multi-task Speech Representation Learning | Model architectures such as wav2vec 2.0 and HuBERT have been proposed to learn speech representations from audio waveforms in a self-supervised manner. When they are combined with downstream tasks such as keyword spotting and speaker verification, they provide state-of-the-art performance. However, these models use a large number of parameters, the smallest version of which has 95 million parameters. This constitutes a challenge for edge AI device deployments. In this paper, we investigate the application of knowledge distillation to speech representation learning (SRL) models followed by joint fine-tuning with multiple downstream voice-activated tasks. In our experiments on two such tasks, our approach results in nearly 75% reduction in model size while suffering only 0.1% accuracy and 0.9% equal error rate degradation compared to the full-size model. In addition, we show that fine-tuning the SRL models results in a significant performance boost compared to using frozen SRL models. | ['Erik Visser', 'Shuhua Zhang', 'Van Nguyen', 'Mine Kerpicci'] | 2022-10-29 | null | null | null | null | ['keyword-spotting', 'speaker-verification'] | ['speech', 'speech'] | [ 2.90424198e-01 2.76457250e-01 -1.50285572e-01 -2.94780016e-01
-1.30056274e+00 -6.28874123e-01 6.08712614e-01 -1.48214456e-02
-3.79426479e-01 5.55662334e-01 4.82760787e-01 -5.04716337e-01
7.95070678e-02 -2.48804212e-01 -6.09988093e-01 -3.89736623e-01
-7.28642866e-02 4.52315122e-01 8.74102786e-02 -1.30476370e-01
-1.98922589e-01 3.25330526e-01 -1.38902831e+00 3.21234196e-01
2.17799246e-01 1.22786081e+00 -6.95360303e-02 8.55945289e-01
-8.27429891e-02 7.76177347e-01 -8.35649192e-01 -1.53752029e-01
-4.29253615e-02 -1.05011657e-01 -8.76120567e-01 -3.07888001e-01
3.47180635e-01 -2.25359648e-01 -6.55633211e-01 5.51642656e-01
1.00548017e+00 2.58617759e-01 4.35213089e-01 -1.04740524e+00
-5.76218545e-01 1.19208908e+00 -2.51841128e-01 3.61414462e-01
9.88347009e-02 6.77966624e-02 1.14275217e+00 -9.08573329e-01
1.16259180e-01 1.04381192e+00 7.50111938e-01 8.26942325e-01
-1.28446937e+00 -8.96522582e-01 4.52480465e-03 2.88869858e-01
-1.69962883e+00 -1.36459947e+00 7.56192982e-01 -1.31808415e-01
1.55892098e+00 2.45256901e-01 1.73002273e-01 1.19856453e+00
-4.52912152e-01 7.70392656e-01 5.02088010e-01 -3.89261872e-01
3.66472751e-01 2.29196846e-01 1.11127213e-01 3.76747012e-01
-2.02971008e-02 -7.62474118e-03 -9.75144446e-01 -2.28847682e-01
5.04644394e-01 -7.07787812e-01 -2.04230651e-01 2.31403977e-01
-8.68424773e-01 6.85598075e-01 2.08615541e-01 4.08756554e-01
-3.04044574e-01 4.94796127e-01 6.48364186e-01 2.69164145e-01
6.06905282e-01 4.84012246e-01 -5.98975480e-01 -6.82612300e-01
-1.17649591e+00 -4.82924990e-02 8.57861698e-01 9.71891344e-01
2.26845235e-01 8.61696839e-01 -2.67329603e-01 1.16978145e+00
2.06091553e-01 3.16306859e-01 8.12224865e-01 -7.28329301e-01
4.66877341e-01 -3.66046987e-02 -2.76114255e-01 -3.94136518e-01
-3.81901890e-01 -7.71720231e-01 -5.10346174e-01 -1.91345751e-01
2.55143959e-02 -4.12127852e-01 -1.26704049e+00 1.87111497e+00
-6.86555430e-02 7.83333302e-01 3.17289054e-01 6.17552578e-01
1.10589325e+00 9.27870035e-01 -9.39897553e-04 -1.47346020e-01
1.16630352e+00 -1.10310030e+00 -7.80285001e-01 -5.32399356e-01
5.56064725e-01 -7.06104159e-01 8.56885433e-01 4.79362518e-01
-1.32170987e+00 -4.84480321e-01 -1.18426907e+00 -5.47675975e-02
-2.33043462e-01 1.27145365e-01 5.29026806e-01 9.87369120e-01
-1.33758748e+00 3.29096377e-01 -7.29090869e-01 -2.17470065e-01
4.95188147e-01 5.56167901e-01 -1.91394523e-01 2.57357627e-01
-1.21330976e+00 8.49186420e-01 2.37991065e-01 -2.00714990e-01
-1.24173284e+00 -9.95780528e-01 -8.26582849e-01 3.74975085e-01
2.32285291e-01 -3.61289680e-01 1.54844630e+00 -4.94425327e-01
-1.84083140e+00 5.79628587e-01 -3.17694962e-01 -9.04024839e-01
-8.32682941e-03 -3.39565039e-01 -8.84416461e-01 -8.94882753e-02
-5.28862894e-01 6.32154942e-01 1.03609467e+00 -9.06474829e-01
-4.02367949e-01 1.01733685e-01 -1.53120637e-01 -2.74046175e-02
-6.87958062e-01 2.44110450e-01 -6.65217102e-01 -7.75512099e-01
-3.55197996e-01 -7.43445754e-01 -6.07183464e-02 -2.95154631e-01
-1.72202840e-01 -2.73559213e-01 7.98246622e-01 -8.70155513e-01
1.51641989e+00 -2.28636432e+00 1.43237282e-02 4.14170176e-02
5.63387088e-02 6.77861452e-01 -4.83895838e-01 2.64097124e-01
-3.09920534e-02 2.26611733e-01 -1.25066280e-01 -7.78418005e-01
-3.74641456e-02 9.20406580e-02 -5.11900723e-01 1.83496088e-01
2.89882183e-01 9.20261681e-01 -5.76383948e-01 1.94967575e-02
-4.48360443e-02 9.20707583e-01 -6.13375664e-01 1.55969769e-01
-3.97133380e-02 1.52075559e-01 -3.14672990e-03 4.46482658e-01
2.48364881e-01 -2.45743524e-02 9.01025757e-02 -2.11349264e-01
1.92058012e-01 9.40645635e-01 -9.02067542e-01 2.01880860e+00
-8.19756269e-01 9.62985575e-01 1.66497961e-01 -9.25043523e-01
9.44152355e-01 6.64478242e-01 3.08724493e-01 -4.42138523e-01
2.45077699e-01 4.93899360e-02 1.25322834e-01 -9.06499550e-02
4.34712470e-01 -1.40773147e-01 -9.25502852e-02 4.43881124e-01
5.49640656e-01 -3.40159237e-01 -2.63148427e-01 1.72821924e-01
1.27486765e+00 -3.38189125e-01 1.74221560e-01 -7.07504150e-05
2.62417465e-01 -4.38925117e-01 4.61438119e-01 6.74532413e-01
4.35365997e-02 6.75007701e-01 1.31683126e-01 -9.23878178e-02
-7.52144694e-01 -1.04512453e+00 -1.02677651e-01 9.98758316e-01
-3.43995064e-01 -8.58970761e-01 -8.04860473e-01 -3.48945886e-01
-5.20977788e-02 1.09354341e+00 -2.75833488e-01 -5.58334827e-01
-4.74091828e-01 -7.17255950e-01 1.36571586e+00 7.53470659e-01
2.00733319e-01 -9.94726539e-01 -3.24434698e-01 2.13350266e-01
1.43961355e-01 -1.47124720e+00 -3.93042952e-01 2.99775690e-01
-6.07984066e-01 -3.05041105e-01 -6.78150713e-01 -6.45315170e-01
3.23158093e-02 -1.40347779e-01 1.06049240e+00 -3.15360725e-01
-2.13873565e-01 4.10397708e-01 -3.60973507e-01 -4.73143071e-01
-4.42199081e-01 3.15089732e-01 5.04911423e-01 1.78977922e-01
1.57723367e-01 -5.97474277e-01 -1.80876285e-01 -7.60324746e-02
-6.38827741e-01 -3.36818129e-01 4.43693846e-01 8.20473194e-01
3.78076702e-01 -9.53054652e-02 1.09559667e+00 -6.38655782e-01
8.92495751e-01 -4.58561569e-01 -3.26872200e-01 1.78542733e-01
-7.94978917e-01 1.08705990e-01 2.37106010e-01 -7.00347126e-01
-9.21641767e-01 4.66960529e-03 -4.78070050e-01 -7.70715475e-01
-8.77484083e-02 4.67633188e-01 -7.81918876e-03 2.11031642e-02
6.41097665e-01 2.18370602e-01 -9.13459733e-02 -6.69544935e-01
6.40834272e-01 1.12412763e+00 5.30843556e-01 -3.06389481e-01
6.55567288e-01 5.07086404e-02 -5.12980402e-01 -1.15782046e+00
-5.88763595e-01 -3.24187547e-01 -1.58866614e-01 8.98501128e-02
7.08198607e-01 -1.19715440e+00 -6.18867815e-01 3.16480398e-01
-1.09289813e+00 -5.81612289e-01 -5.39160073e-01 5.34475148e-01
-5.73820591e-01 6.59387782e-02 -6.80926621e-01 -1.11565852e+00
-7.32611954e-01 -9.79970753e-01 1.02073896e+00 2.22374033e-02
-4.41854209e-01 -7.98493147e-01 -4.70078103e-02 3.48379403e-01
9.65054750e-01 -5.64423800e-01 7.52170324e-01 -1.02672017e+00
-7.57763460e-02 -1.80889398e-01 4.54225093e-02 5.81035614e-01
1.20054945e-01 -5.14528036e-01 -1.75374687e+00 -3.56433004e-01
-1.50806189e-01 -4.48839307e-01 9.97815132e-01 3.82599503e-01
1.28709400e+00 -1.46909118e-01 -1.51067168e-01 6.95878625e-01
7.01615989e-01 2.67034650e-01 3.72395873e-01 -2.55681515e-01
5.86437106e-01 2.29878798e-01 1.29813671e-01 3.52550179e-01
1.13157168e-01 9.47236657e-01 1.67620499e-02 8.19906965e-02
-7.78744042e-01 -3.00856292e-01 4.85445499e-01 1.21897066e+00
1.53099909e-01 -3.26352388e-01 -9.69173729e-01 7.09379256e-01
-1.60389292e+00 -7.05034256e-01 3.33770633e-01 2.24222016e+00
9.80517209e-01 2.66082853e-01 7.73512125e-02 3.05048406e-01
4.44140494e-01 2.62812793e-01 -7.06051171e-01 -6.13872230e-01
-1.38988495e-01 6.92187130e-01 4.35612410e-01 5.83652496e-01
-9.49589849e-01 1.26200795e+00 7.12052441e+00 1.09671867e+00
-1.30467141e+00 4.91468459e-01 4.82415199e-01 -5.59409738e-01
-2.90881723e-01 -2.40674525e-01 -8.31497967e-01 2.55060852e-01
1.69301152e+00 -3.13166022e-01 8.25589120e-01 6.31057143e-01
-3.65094244e-02 4.47528929e-01 -1.01512396e+00 1.60554624e+00
4.24197167e-01 -1.43284357e+00 -7.27856383e-02 -1.07117459e-01
5.04640102e-01 4.52910691e-01 1.71188116e-02 6.49065733e-01
1.45402431e-01 -1.27031291e+00 7.38013089e-01 1.31361187e-01
1.21602261e+00 -7.89952397e-01 3.86906415e-01 1.55856535e-01
-1.19827437e+00 -1.57366157e-01 -1.91449150e-01 1.85018703e-02
6.36719763e-01 5.21783590e-01 -1.19894135e+00 2.43255705e-01
7.08199024e-01 3.50035846e-01 -2.59920537e-01 9.14449632e-01
-1.48891106e-01 1.36544478e+00 -4.59862649e-01 1.62371233e-01
-6.49546310e-02 5.24195135e-01 5.44668853e-01 1.40708983e+00
3.37576628e-01 -1.17181249e-01 -2.30122805e-01 5.80552280e-01
-3.95642608e-01 -6.05030842e-02 -6.26938224e-01 -3.85291636e-01
9.06202435e-01 9.73981619e-01 -6.08403496e-02 -4.51401174e-01
-3.01044613e-01 9.54608142e-01 3.99276376e-01 4.25822973e-01
-7.90915966e-01 -4.28318977e-01 1.07877243e+00 -3.55260856e-02
4.00145531e-01 -2.46138901e-01 -2.85203546e-01 -8.36248815e-01
-1.92450672e-01 -7.14582145e-01 2.03394413e-01 -6.38678908e-01
-1.13619149e+00 1.00582206e+00 -1.65248320e-01 -8.38467240e-01
-7.19706655e-01 -4.90976393e-01 -5.88425517e-01 9.16238546e-01
-1.43151951e+00 -1.11802077e+00 1.57481417e-01 4.12266403e-01
9.17342544e-01 -6.23713911e-01 1.27900279e+00 6.19177759e-01
-5.40090144e-01 1.20081246e+00 -1.87510729e-01 5.30431755e-02
5.92529655e-01 -8.97053540e-01 8.81229520e-01 6.12839460e-01
6.17329597e-01 4.47958946e-01 4.72032428e-01 -2.94475526e-01
-1.38177669e+00 -1.08679581e+00 9.58044410e-01 -4.56204742e-01
7.38585114e-01 -6.11973584e-01 -8.97712767e-01 7.10359991e-01
3.51162195e-01 1.45487264e-01 8.36976707e-01 4.92179245e-01
-5.34387231e-01 -2.00581759e-01 -8.55699420e-01 4.86926943e-01
1.16530085e+00 -9.56486523e-01 -4.70948726e-01 8.07856768e-02
1.19186449e+00 -2.87268192e-01 -8.98767591e-01 4.64817107e-01
5.08834660e-01 -2.53899455e-01 1.10456514e+00 -8.20274472e-01
-1.75392240e-01 -1.25673816e-01 -2.92379558e-01 -1.41575122e+00
-2.61657000e-01 -1.00449705e+00 -6.47902548e-01 1.62920916e+00
8.92677665e-01 -4.63664681e-01 4.59785491e-01 4.11887974e-01
-4.23623860e-01 -6.01153731e-01 -1.22380769e+00 -1.03021729e+00
1.18865490e-01 -1.03774667e+00 5.01584411e-01 7.61754572e-01
1.76982090e-01 7.43408024e-01 -5.21663129e-01 2.09246263e-01
3.44864786e-01 -3.84287536e-01 5.44316232e-01 -9.69842970e-01
-5.31742871e-01 -5.31957150e-01 -4.62064743e-01 -1.07108951e+00
3.11162025e-01 -1.02651811e+00 5.37947267e-02 -1.33996642e+00
-2.33023718e-01 -4.79410648e-01 -5.92425704e-01 7.56427884e-01
1.26271963e-01 2.69460678e-01 2.46310204e-01 -1.42204940e-01
-3.36142063e-01 5.23710251e-01 2.43349656e-01 -3.66007090e-01
-4.07241404e-01 -4.16739769e-02 -8.02297175e-01 4.40580636e-01
1.16080105e+00 -3.31589520e-01 -6.73476338e-01 -7.68624604e-01
-1.77748099e-01 -1.13371789e-01 4.55797240e-02 -1.13300967e+00
3.74398887e-01 4.30198640e-01 9.02960673e-02 -1.24450207e-01
1.16874254e+00 -3.90456378e-01 -1.18463181e-01 3.43496054e-02
-5.56736827e-01 -2.20289648e-01 7.72856534e-01 4.25029248e-01
-2.71494508e-01 -1.96178451e-01 4.97646421e-01 3.54551464e-01
-7.12521195e-01 -2.47408990e-02 -5.14239728e-01 1.09115161e-01
6.91522479e-01 2.30469137e-01 -2.30296150e-01 -7.36957490e-01
-8.27671885e-01 -2.52236985e-02 -3.38122874e-01 8.70782912e-01
6.76250815e-01 -1.24755204e+00 -8.62195015e-01 3.83402109e-01
9.92991850e-02 -1.75774962e-01 1.65229216e-01 6.48117900e-01
1.81801602e-01 6.16438150e-01 3.66202086e-01 -4.08225358e-01
-1.23761749e+00 3.38773400e-01 1.43379897e-01 2.38653235e-02
-5.29467642e-01 1.31938779e+00 1.23153776e-02 -4.10743207e-01
6.51459217e-01 -2.84002215e-01 -1.22083426e-01 -6.71780715e-03
7.32274115e-01 2.91198462e-01 4.19204175e-01 -5.53038120e-01
-6.31277204e-01 3.52872163e-01 -6.90709203e-02 -5.07961988e-01
1.22060025e+00 9.77259502e-02 4.82338846e-01 5.76974809e-01
1.02988005e+00 1.88581258e-01 -9.29992855e-01 -3.64046872e-01
9.23308451e-03 4.72290479e-02 5.68491697e-01 -9.71192658e-01
-1.17639303e+00 1.05769169e+00 6.79287195e-01 5.63203879e-02
1.09914160e+00 4.47968185e-01 9.12577033e-01 4.13032919e-01
2.83970386e-01 -9.90524471e-01 -3.09654206e-01 5.91565251e-01
8.63399088e-01 -8.84171069e-01 -3.49761575e-01 -1.60062611e-01
-7.70830393e-01 6.31017625e-01 3.71685356e-01 7.46030062e-02
8.79262626e-01 7.21355975e-01 8.45803469e-02 4.08041710e-03
-1.21684158e+00 -4.24033761e-01 4.39693093e-01 6.62508249e-01
4.64940965e-01 1.65915102e-01 1.63338959e-01 1.01964641e+00
-3.75024617e-01 -1.09005071e-01 1.19293071e-01 7.26514280e-01
-2.92901278e-01 -1.01852345e+00 -1.18346304e-01 4.90780324e-01
-4.05704439e-01 -4.66825545e-01 -3.03308159e-01 2.20051095e-01
-3.29793692e-01 1.33867013e+00 2.28866190e-01 -8.21968555e-01
3.34989876e-01 5.08724988e-01 3.37146938e-01 -7.54177392e-01
-7.01756716e-01 1.77537695e-01 5.89435995e-01 -5.07087886e-01
1.18724287e-01 -5.44732332e-01 -1.38695097e+00 1.35863394e-01
-4.10046369e-01 1.32739469e-01 9.23075974e-01 8.05363536e-01
7.87145078e-01 1.03108573e+00 3.76919419e-01 -7.35186875e-01
-7.80468345e-01 -1.20827866e+00 -4.33543116e-01 -6.29152954e-02
4.59923297e-01 -5.34316123e-01 -3.41746956e-01 2.95813172e-03] | [14.354606628417969, 6.377509117126465] |
de81caa9-0a95-4afd-97aa-06ab44f93b43 | velocity-continuation-with-fourier-neural | 2203.14386 | null | https://arxiv.org/abs/2203.14386v1 | https://arxiv.org/pdf/2203.14386v1.pdf | Velocity continuation with Fourier neural operators for accelerated uncertainty quantification | Seismic imaging is an ill-posed inverse problem that is challenged by noisy data and modeling inaccuracies -- due to errors in the background squared-slowness model. Uncertainty quantification is essential for determining how variability in the background models affects seismic imaging. Due to the costs associated with the forward Born modeling operator as well as the high dimensionality of seismic images, quantification of uncertainty is computationally expensive. As such, the main contribution of this work is a survey-specific Fourier neural operator surrogate to velocity continuation that maps seismic images associated with one background model to another virtually for free. While being trained with only 200 background and seismic image pairs, this surrogate is able to accurately predict seismic images associated with new background models, thus accelerating seismic imaging uncertainty quantification. We support our method with a realistic data example in which we quantify seismic imaging uncertainties using a Fourier neural operator surrogate, illustrating how variations in background models affect the position of reflectors in a seismic image. | ['Felix J. Herrmann', 'Mathias Louboutin', 'Ali Siahkoohi'] | 2022-03-27 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 7.32420087e-01 6.86861202e-03 8.08675110e-01 -2.26450071e-01
-1.38491273e+00 -3.95800561e-01 5.72549105e-01 -2.25951701e-01
-6.54355884e-01 6.24164164e-01 4.26139235e-01 -1.30256370e-01
-3.93818289e-01 -6.89892292e-01 -8.20064008e-01 -8.66059780e-01
-2.08039418e-01 4.33700114e-01 4.40187901e-01 1.06043413e-01
4.75365222e-01 5.84582269e-01 -1.20702672e+00 6.74664453e-02
3.95893067e-01 1.00437164e+00 1.23732559e-01 7.71057069e-01
-6.33746684e-02 8.58723879e-01 -6.42205417e-01 8.64211395e-02
1.92697391e-01 -3.26384127e-01 -5.08753121e-01 -1.53177544e-01
2.13223249e-01 -6.34926796e-01 -4.90866512e-01 9.80520427e-01
6.05735779e-01 4.64112222e-01 9.35937166e-01 -5.68818331e-01
-1.04015380e-01 6.55919790e-01 -6.99004829e-01 5.05616307e-01
6.79252818e-02 7.45986924e-02 4.16120142e-01 -9.54099000e-01
4.61654782e-01 9.32978392e-01 1.19015682e+00 1.34097010e-01
-1.47231877e+00 -3.26472580e-01 -2.10733518e-01 -4.90973294e-02
-1.37867296e+00 -6.24886215e-01 8.83229911e-01 -8.37053597e-01
7.22561955e-01 3.54234189e-01 1.52531713e-01 8.49684358e-01
-4.33190307e-03 1.60723805e-01 1.00597978e+00 -2.61484891e-01
4.22535002e-01 -2.69021511e-01 5.12172170e-02 1.98499665e-01
1.48006767e-01 4.07270879e-01 -6.53182030e-01 -2.46078327e-01
9.92275417e-01 -5.48011899e-01 -5.28719008e-01 2.40679146e-04
-1.11294377e+00 5.81117332e-01 2.76050359e-01 1.46261007e-01
-2.90837467e-01 7.48696327e-01 2.58524157e-02 -1.11452989e-01
4.98506099e-01 4.74671513e-01 -6.98494166e-02 -3.36771488e-01
-1.42686987e+00 2.73757339e-01 6.15751922e-01 2.86076933e-01
7.16172814e-01 6.17367148e-01 2.25413725e-01 4.21736062e-01
4.72413778e-01 8.80185544e-01 6.78897351e-02 -1.43543935e+00
1.38555527e-01 -2.40398273e-01 2.57026672e-01 -1.09133947e+00
-5.10249555e-01 -3.85145009e-01 -5.59178293e-01 6.74761117e-01
7.85563827e-01 -2.10706905e-01 -7.73837924e-01 1.38207972e+00
3.84984054e-02 5.44824958e-01 1.12375118e-01 7.42771685e-01
5.57281971e-01 6.14596009e-01 -1.64089128e-01 -8.35806280e-02
9.74808633e-01 -1.02029681e-01 -5.82311153e-01 -4.13164735e-01
3.65738362e-01 -7.36565173e-01 7.33329177e-01 2.33963698e-01
-1.15291345e+00 -1.30639285e-01 -1.00524044e+00 1.97118670e-01
1.95762679e-01 -4.15201336e-01 1.91952437e-01 5.86346686e-01
-8.53123724e-01 1.00445342e+00 -1.09617090e+00 2.59692311e-01
1.97101101e-01 -1.44269941e-02 -1.43002734e-01 9.41055119e-02
-1.17619729e+00 9.04138625e-01 1.28439128e-01 3.21475446e-01
-8.13511610e-01 -1.10206211e+00 -8.51766586e-01 -2.52164509e-02
1.73839673e-01 -2.71915078e-01 1.18114400e+00 -6.34470880e-01
-1.13247180e+00 1.74740463e-01 4.28480841e-02 -2.98200577e-01
9.75394607e-01 1.55370504e-01 -2.68930912e-01 3.78216624e-01
1.59548353e-02 5.34513555e-02 1.08311605e+00 -1.55873418e+00
-2.83979982e-01 -1.20325178e-01 -4.27634358e-01 -1.10886984e-01
4.13637757e-01 -1.26949642e-02 -8.74050781e-02 -6.40765369e-01
6.45890236e-01 -5.90792358e-01 -4.27425534e-01 -4.44847904e-02
-2.76411902e-02 9.32504714e-01 5.10125935e-01 -1.14863551e+00
9.52676117e-01 -2.26660633e+00 -1.73057631e-01 5.62440157e-01
3.91451456e-02 -4.02177125e-01 7.71913677e-02 2.57894993e-01
-9.05724168e-02 2.26365060e-01 -8.85906398e-01 -8.28713477e-02
-1.11070096e-01 7.98628405e-02 -3.58315766e-01 5.44720531e-01
4.04536813e-01 3.41819406e-01 -8.21533203e-01 -1.42017782e-01
2.86547065e-01 8.32532942e-01 -2.87756175e-01 -1.77952886e-01
1.80955399e-02 7.65584230e-01 -1.84713408e-01 4.18974608e-01
1.12012160e+00 7.20525235e-02 -3.89409006e-01 -3.99173677e-01
-4.38585043e-01 -2.02892408e-01 -1.60084331e+00 1.34177184e+00
-4.02228624e-01 7.01882899e-01 2.26847395e-01 -7.81495273e-01
8.91205013e-01 2.05971718e-01 4.13302630e-01 -6.20781779e-01
-1.27500340e-01 6.85879171e-01 -7.55458996e-02 -6.47696674e-01
5.23368001e-01 -5.92303157e-01 2.34161705e-01 3.00952286e-01
-2.71907032e-01 -7.91211665e-01 -1.41698807e-01 1.33877292e-01
1.20572567e+00 3.13378036e-01 -3.58752608e-01 -4.00599808e-01
1.27915233e-01 -2.32216772e-02 3.06110978e-01 9.62191284e-01
1.55139536e-01 1.36030519e+00 5.01246810e-01 -2.93721348e-01
-1.10934222e+00 -1.16816020e+00 -5.09105921e-01 2.30184644e-01
-1.35355160e-01 3.47944915e-01 -5.55055916e-01 1.38875753e-01
-3.65576744e-02 1.04587436e+00 -6.93589091e-01 -1.43954337e-01
-6.32008493e-01 -1.10821688e+00 7.77947485e-01 5.15097916e-01
4.04380232e-01 -5.50241947e-01 -9.77809668e-01 4.12998497e-01
-3.14748764e-01 -1.04548335e+00 -1.90497458e-01 3.43368053e-01
-1.01158381e+00 -9.41563666e-01 -1.06540060e+00 4.61785160e-02
6.31200075e-01 -1.45631462e-01 9.90201652e-01 -2.23729655e-01
-3.23235154e-01 7.42682397e-01 -3.11237574e-02 -3.15981030e-01
-7.46298730e-01 -7.85311639e-01 -2.21640408e-01 1.05767950e-01
-2.85605013e-01 -5.67977071e-01 -5.34917474e-01 2.11277962e-01
-1.23947155e+00 -2.99563497e-01 9.48286727e-02 8.57389808e-01
4.63144064e-01 3.87197226e-01 -1.04211718e-01 -5.37739694e-01
4.00447071e-01 -4.16954815e-01 -9.21249688e-01 -3.72290574e-02
-1.96514994e-01 3.58583003e-01 -1.58833727e-01 -3.09031874e-01
-1.60422707e+00 1.79715827e-03 -2.55071640e-01 -7.16487169e-02
1.15704350e-01 8.51736307e-01 3.70199829e-01 -4.98088121e-01
1.10741508e+00 2.60981880e-02 5.22841103e-02 -3.46049964e-01
-9.58607793e-02 2.46753559e-01 9.92911875e-01 -6.92295969e-01
7.46483207e-01 8.04689944e-01 5.28876483e-01 -1.41427791e+00
-4.86422330e-01 -3.13391715e-01 -4.62681472e-01 -6.19457722e-01
7.76271105e-01 -8.70419681e-01 -4.18248326e-01 6.70911849e-01
-1.07405174e+00 -3.62480402e-01 -5.43281734e-01 8.79183650e-01
-4.56095040e-01 6.49281561e-01 -4.57973748e-01 -1.26718640e+00
3.81453574e-01 -1.41453695e+00 9.78022516e-01 -3.71888541e-02
-3.40931453e-02 -1.03872287e+00 -4.54839729e-02 6.47047609e-02
6.31556094e-01 8.04503918e-01 6.50610387e-01 8.46096203e-02
-6.60808563e-01 -3.41815472e-01 -2.35242665e-01 4.49460268e-01
-2.33612686e-01 3.01416159e-01 -1.32443392e+00 1.71128690e-01
7.09726334e-01 -4.52634059e-02 1.07151604e+00 9.75896597e-01
6.56588197e-01 6.39392734e-02 1.58889487e-01 8.20665300e-01
1.71266484e+00 1.76323175e-01 8.35590422e-01 5.92510879e-01
4.23670948e-01 8.39860380e-01 8.93665329e-02 5.54215491e-01
-1.83984727e-01 2.65532851e-01 4.94617194e-01 6.89044446e-02
-8.03452581e-02 1.75447449e-01 3.97271067e-02 5.03006577e-01
-4.14123803e-01 -4.40750569e-02 -1.28795385e+00 4.60136771e-01
-1.43874621e+00 -1.08325708e+00 -6.88049734e-01 2.48596549e+00
3.34949136e-01 1.73745930e-01 -4.89336520e-01 2.10702449e-01
4.66962248e-01 1.42670393e-01 -5.14571667e-01 1.85217634e-01
-4.19165850e-01 1.92606777e-01 8.79741371e-01 1.01291084e+00
-8.39158475e-01 9.20437574e-02 7.03552437e+00 4.64864671e-01
-9.43281651e-01 -1.46074653e-01 6.10291421e-01 1.01878941e-01
-7.92656958e-01 4.73576449e-02 -2.50978738e-01 4.89220589e-01
9.13523853e-01 5.07884212e-02 4.91598517e-01 1.43060356e-01
3.46246272e-01 -7.85730600e-01 -7.13392377e-01 8.98387134e-01
-7.66383559e-02 -1.53315437e+00 -2.47754231e-01 -2.45938450e-02
5.67480326e-01 3.97344977e-02 6.58646822e-02 -2.97638148e-01
1.20725058e-01 -9.22302783e-01 1.03393006e+00 1.00886559e+00
7.86472440e-01 -7.26236999e-01 7.29493499e-01 2.21424162e-01
-7.41997480e-01 2.43386269e-01 -2.58658707e-01 -6.29616976e-02
6.29333079e-01 7.88583994e-01 -4.79841858e-01 4.29053247e-01
6.87933624e-01 7.50242472e-02 -2.13191643e-01 8.80889952e-01
1.06517904e-01 7.40097225e-01 -7.32647598e-01 5.24264872e-01
2.46394411e-01 -3.88839394e-01 6.14565194e-01 1.07040381e+00
7.98352301e-01 3.99054140e-01 -3.89894307e-01 1.16242874e+00
4.23953950e-01 -5.25623202e-01 -4.53773230e-01 -6.64913729e-02
1.41139030e-01 7.77739584e-01 -9.61881280e-01 1.78954154e-01
-3.25720072e-01 5.99805593e-01 -3.73061419e-01 6.86639786e-01
-6.28700078e-01 -2.14306444e-01 2.25939557e-01 3.37009400e-01
6.41936436e-02 -3.76329601e-01 -3.87320787e-01 -6.98386371e-01
7.26717860e-02 -5.01726925e-01 9.61258486e-02 -9.98957574e-01
-1.00248003e+00 3.90183300e-01 4.02778059e-01 -1.03080332e+00
-3.74474853e-01 -5.60626805e-01 -5.61566293e-01 1.32033193e+00
-1.34854639e+00 -6.43334091e-01 -3.97782981e-01 1.20785415e-01
2.37336084e-02 1.88162342e-01 6.70473754e-01 2.18406633e-01
-2.43662782e-02 -1.03389964e-01 6.85939908e-01 -1.21439375e-01
3.93294603e-01 -1.11980045e+00 6.71857715e-01 9.44098234e-01
-1.51007414e-01 1.93196088e-01 1.30379939e+00 -7.48728454e-01
-1.12792325e+00 -4.94514734e-01 3.86332929e-01 -4.39496547e-01
1.06611705e+00 1.22133337e-01 -1.30099475e+00 4.08720404e-01
-3.53365302e-01 2.90023267e-01 4.49601591e-01 -6.20822430e-01
3.82511765e-02 1.43962532e-01 -1.27330959e+00 4.71369416e-01
5.59085011e-01 -7.04581201e-01 -4.86484349e-01 1.46819698e-02
3.58626992e-01 -6.01182103e-01 -8.48042130e-01 5.20152807e-01
5.37870109e-01 -1.24486995e+00 1.20186555e+00 3.94600853e-02
5.38520575e-01 -4.40667689e-01 -4.34974641e-01 -1.19961214e+00
-1.32543385e-01 -5.13949633e-01 1.63543090e-01 9.73976433e-01
3.59527260e-01 -6.70036495e-01 7.35321045e-01 1.09272599e+00
4.70025977e-03 -4.52110767e-02 -1.30831647e+00 -9.44051445e-01
1.37146786e-01 -9.97674465e-01 1.89466506e-01 6.17418110e-01
-5.76279283e-01 -3.23582977e-01 -1.34575188e-01 6.80122674e-01
1.15428627e+00 -6.36012197e-01 2.53105283e-01 -1.21549368e+00
-5.75602889e-01 -3.63911182e-01 -2.52272427e-01 -6.30018353e-01
-4.33632165e-01 -2.95101196e-01 4.85535830e-01 -1.23067069e+00
-1.79051071e-01 -4.91780668e-01 4.90452647e-02 -3.37233275e-01
3.34334373e-03 2.95316190e-01 -9.45780799e-03 2.83226371e-01
5.59533715e-01 2.30088979e-01 8.07632506e-01 -1.59323037e-01
4.17604893e-02 -1.79594867e-02 4.26305085e-02 1.18432581e+00
4.77639943e-01 -8.04380476e-01 -2.62637049e-01 -7.98271477e-01
6.60488725e-01 5.93583047e-01 6.92613959e-01 -1.23760068e+00
2.07090050e-01 9.31043252e-02 4.88074720e-01 -5.66774905e-01
3.65022451e-01 -8.72019112e-01 8.35445762e-01 5.05437434e-01
-1.82982296e-01 -1.41982481e-01 4.66118634e-01 6.41018331e-01
-2.70117432e-01 -1.08140671e+00 8.30715775e-01 -4.21799362e-01
-8.94836247e-01 -1.47492155e-01 -3.68681073e-01 2.58246124e-01
4.72841084e-01 -3.18589151e-01 -1.28410012e-01 -5.18663406e-01
-8.58178854e-01 -4.67081934e-01 4.19794351e-01 -4.63186264e-01
5.96724153e-01 -9.41894114e-01 -8.72579098e-01 3.04066092e-01
-9.70086604e-02 1.17942668e-01 7.32087314e-01 7.67748177e-01
-1.21892488e+00 -2.47207150e-01 1.05555683e-01 -7.81086206e-01
-7.91024983e-01 1.14923529e-01 7.50818789e-01 5.03086410e-02
-6.33939862e-01 1.28375494e+00 5.25712818e-02 -1.06059179e-01
1.32262334e-01 -4.18225557e-01 1.98525444e-01 -2.08628308e-02
7.59081006e-01 9.16342616e-01 7.81835914e-02 -6.63304389e-01
-1.80522934e-01 7.94667959e-01 5.74968398e-01 -8.42115283e-01
1.17479122e+00 -8.53231847e-02 5.19913658e-02 8.57838571e-01
9.47162747e-01 1.11497633e-01 -1.77018201e+00 -5.28463647e-02
2.08216414e-01 -5.08491874e-01 5.65204740e-01 -6.18507981e-01
-7.00533926e-01 9.44776416e-01 6.96340263e-01 8.52775201e-02
8.72946143e-01 -3.50683063e-01 2.15617970e-01 3.75074953e-01
2.48341441e-01 -1.10017979e+00 -7.80493245e-02 3.85312915e-01
1.06136405e+00 -1.18770421e+00 2.82755286e-01 -1.25249475e-01
-5.29652119e-01 1.32742441e+00 -1.47083789e-01 -1.80728771e-02
7.68333375e-01 9.01060462e-01 3.39423299e-01 -1.56725124e-01
1.54543817e-01 1.78420380e-01 -2.74267211e-03 6.74552381e-01
1.09692767e-01 -4.04846281e-01 3.06319535e-01 4.95834798e-02
1.95429370e-01 5.18561304e-02 8.90215695e-01 1.01578021e+00
-3.62752646e-01 -5.60042381e-01 -1.05619729e+00 3.14768910e-01
-3.21954399e-01 -2.19542846e-01 2.95213640e-01 6.85724497e-01
-1.70327067e-01 6.18813097e-01 3.38524759e-01 8.90851989e-02
2.98705578e-01 -6.65448830e-02 4.79204118e-01 -1.08525738e-01
-1.68584734e-01 3.57233763e-01 7.88641796e-02 -3.18208992e-01
-4.53162700e-01 -9.78992224e-01 -1.49268520e+00 -1.14146054e-01
-3.12888771e-01 -5.02293836e-03 9.37447131e-01 7.05520868e-01
-1.99562401e-01 8.80943120e-01 1.95622176e-01 -1.37016547e+00
-1.01716077e+00 -7.26597369e-01 -9.46598411e-01 8.89775902e-02
5.70612848e-01 -5.80732822e-01 -1.10529757e+00 1.79880008e-01] | [6.843433380126953, 3.366564989089966] |
a6edad6a-43f2-45ac-b066-6a16370f4e91 | studenteval-a-benchmark-of-student-written | 2306.04556 | null | https://arxiv.org/abs/2306.04556v1 | https://arxiv.org/pdf/2306.04556v1.pdf | StudentEval: A Benchmark of Student-Written Prompts for Large Language Models of Code | Code LLMs are being rapidly deployed and there is evidence that they can make professional programmers more productive. Current benchmarks for code generation measure whether models generate correct programs given an expert prompt. In this paper, we present a new benchmark containing multiple prompts per problem, written by a specific population of non-expert prompters: beginning programmers. StudentEval contains 1,749 prompts for 48 problems, written by 80 students who have only completed one semester of Python programming. Our students wrote these prompts while working interactively with a Code LLM, and we observed very mixed success rates. We use StudentEval to evaluate 5 Code LLMs and find that StudentEval is a better discriminator of model performance than existing benchmarks. We analyze the prompts and find significant variation in students' prompting techniques. We also find that nondeterministic LLM sampling could mislead students into thinking that their prompts are more (or less) effective than they actually are, which has implications for how to teach with Code LLMs. | ['Carolyn Jane Anderson', 'Molly Q Feldman', 'Arjun Guha', 'Yangtian Zi', 'Sydney Nguyen', 'Hannah McLean Babe'] | 2023-06-07 | null | null | null | null | ['code-generation'] | ['computer-code'] | [-5.12286695e-03 2.62527347e-01 -8.83890241e-02 -5.30538857e-01
-9.09673393e-01 -1.14609528e+00 3.66548806e-01 5.30027688e-01
-1.72636226e-01 4.40019161e-01 -2.67853178e-02 -1.11313879e+00
9.24756154e-02 -6.96678519e-01 -8.75670493e-01 -1.02634378e-01
2.43860438e-01 2.44406879e-01 3.09680045e-01 -1.14051871e-01
9.68494713e-01 1.13560870e-01 -1.60120881e+00 5.51037729e-01
1.59081101e+00 -3.05797905e-01 4.11591917e-01 1.27380347e+00
-4.89654481e-01 1.53537595e+00 -1.08622551e+00 -1.81318775e-01
-1.60411611e-01 -3.76495302e-01 -1.19850135e+00 -5.09848833e-01
9.85500038e-01 -1.21748932e-01 1.04653329e-01 1.00278485e+00
3.02359700e-01 1.00528272e-02 3.76325488e-01 -1.44969285e+00
-4.81382310e-01 1.01537728e+00 -5.87814391e-01 3.18445981e-01
8.80132854e-01 4.66613322e-01 7.40658522e-01 -5.34454882e-01
5.08925855e-01 1.10754752e+00 9.20512199e-01 6.57085180e-01
-1.69086814e+00 -7.79457867e-01 -2.09398404e-01 -1.46591604e-01
-1.06149960e+00 -6.86267018e-02 2.20461443e-01 -1.21360683e+00
9.32334781e-01 3.98157656e-01 8.19953978e-01 8.82945776e-01
3.98566335e-01 7.30876088e-01 1.37285721e+00 -6.38662875e-01
7.51397535e-02 6.78307652e-01 6.87447131e-01 6.48575544e-01
3.91737670e-01 -6.16633110e-02 -3.99245620e-01 -7.74342179e-01
5.82195401e-01 -3.03321779e-01 -2.27333650e-01 1.03926331e-01
-1.13367867e+00 8.00430298e-01 -2.03342289e-01 2.43775219e-01
1.86776221e-01 4.34464961e-01 1.30693182e-01 5.71460187e-01
-3.76593381e-01 1.15114069e+00 -5.94248176e-01 -9.47339296e-01
-8.19957495e-01 7.63260663e-01 1.32697904e+00 1.34186137e+00
6.51888907e-01 7.68407434e-02 -1.42795995e-01 6.74864948e-01
2.00440720e-01 3.04798335e-01 5.37204385e-01 -1.29319143e+00
3.83586496e-01 9.00736511e-01 1.68376088e-01 -7.80525863e-01
-8.86215568e-02 -1.58153534e-01 2.06937626e-01 5.53231955e-01
7.53672361e-01 -4.37088668e-01 -4.00803238e-01 1.47298133e+00
-2.57250547e-01 -2.31401682e-01 -1.33669406e-01 3.02254051e-01
8.89618516e-01 7.57175326e-01 4.00644600e-01 2.13040531e-01
1.15593016e+00 -1.06927943e+00 -7.26822838e-02 -4.16333556e-01
1.06393409e+00 -1.18811429e+00 1.59999931e+00 9.42615211e-01
-1.30821884e+00 -6.42498493e-01 -6.96592569e-01 6.28359541e-02
1.43041655e-01 -5.62498905e-02 6.85406148e-01 9.90421116e-01
-1.15910339e+00 6.03697002e-01 -5.15511036e-01 -1.38599247e-01
-1.75116826e-02 6.77338988e-02 1.40543878e-01 -5.37343621e-02
-5.17794371e-01 7.41332293e-01 8.38179737e-02 -9.43093777e-01
-1.05722463e+00 -1.28479064e+00 -5.48957288e-01 2.61082619e-01
-1.10070072e-01 -1.13883086e-01 2.09922028e+00 -9.10200000e-01
-1.15134585e+00 9.56571043e-01 -2.80750543e-03 2.09341437e-01
6.05850339e-01 -5.18967323e-02 1.56088263e-01 -3.94803792e-01
2.61100739e-01 4.80317533e-01 2.53590852e-01 -1.25529993e+00
-5.50666034e-01 4.21689719e-01 4.10800576e-01 -1.25078395e-01
-1.32944852e-01 3.72099876e-01 4.42700982e-01 -4.04998392e-01
-9.06767920e-02 -8.59820545e-01 -2.55752534e-01 -5.12055337e-01
-1.69515371e-01 -7.14184582e-01 2.40486443e-01 -2.88142532e-01
1.41963840e+00 -1.76285911e+00 -3.64267915e-01 1.95467606e-01
5.43644607e-01 5.64358942e-02 -3.40968490e-01 6.63320065e-01
-4.98224735e-01 5.87728202e-01 -3.80646507e-03 3.71449679e-01
2.33310923e-01 -2.76788116e-01 -4.68623608e-01 -2.90964879e-02
-1.55578777e-01 5.78166187e-01 -1.14611161e+00 -4.49803680e-01
-4.88922223e-02 -3.15042622e-02 -9.48453486e-01 6.13698184e-01
-2.91639864e-01 2.53063738e-01 -2.15770289e-01 4.73321229e-01
5.39244831e-01 -3.19187135e-01 9.48709026e-02 9.30868387e-01
-4.75702554e-01 7.28160799e-01 -9.46511567e-01 1.35696888e+00
-6.73946261e-01 1.04377401e+00 -2.58353680e-01 -4.89828706e-01
1.11419666e+00 3.18518579e-01 -1.88081324e-01 -4.04315919e-01
-2.87696600e-01 5.37458003e-01 5.97278953e-01 -7.48855591e-01
5.65640926e-01 2.91427225e-01 -5.55216596e-02 9.71961737e-01
-1.58512801e-01 -7.76416898e-01 4.80871350e-01 5.63957751e-01
1.53504610e+00 2.48168432e-03 1.93436638e-01 -8.32786918e-01
2.16911748e-01 3.11942905e-01 2.78021872e-01 1.55518389e+00
-1.05234876e-01 3.62209201e-01 9.98325884e-01 -3.67849320e-01
-1.07012522e+00 -8.89628828e-01 2.44030524e-02 1.72622764e+00
-3.55873317e-01 -8.53311718e-01 -8.22157085e-01 -6.11101806e-01
-6.66980818e-02 1.28592849e+00 -1.33171856e-01 -1.57667994e-01
-4.80799973e-01 -4.18917000e-01 6.55577421e-01 4.80598897e-01
-1.49382472e-01 -1.06708133e+00 -9.81202900e-01 4.00455803e-01
-4.43846658e-02 -5.99632204e-01 -4.42737848e-01 4.01653767e-01
-8.44805777e-01 -1.10819650e+00 -3.84440690e-01 -9.40972924e-01
1.14702225e+00 2.37891495e-01 1.79492891e+00 1.08880270e+00
-4.56674427e-01 7.13121533e-01 -2.81256765e-01 -5.22379935e-01
-1.10002351e+00 2.05509767e-01 -3.72908682e-01 -1.33267033e+00
6.41409874e-01 -5.49562871e-01 -1.43000662e-01 1.51437148e-01
-6.33112490e-01 2.87630975e-01 4.40501213e-01 5.95162630e-01
-4.99133229e-01 -2.66505778e-01 3.55222076e-01 -1.42678082e+00
1.06677580e+00 -6.07213259e-01 -7.61865675e-01 3.27376604e-01
-6.91281736e-01 1.06213577e-01 7.63375282e-01 -5.78835070e-01
-1.00119853e+00 -2.38877535e-01 -1.90561518e-01 6.12613559e-01
-5.07438302e-01 4.35667068e-01 5.90025842e-01 -5.75924218e-01
1.54035401e+00 1.98836595e-01 -3.98345351e-01 -1.81384161e-01
-2.99402297e-01 4.47908372e-01 7.14970604e-02 -1.66390014e+00
7.18502641e-01 -6.79752588e-01 -6.99550509e-01 -6.82316780e-01
-3.48935604e-01 -1.65187433e-01 1.37955740e-01 -3.05528402e-01
2.45015383e-01 -7.60372043e-01 -1.20829391e+00 4.86518919e-01
-1.43278575e+00 -1.11949301e+00 2.54263915e-02 2.44029671e-01
-3.60063821e-01 1.28530160e-01 -9.10528004e-01 -7.32669294e-01
2.07593352e-01 -1.72589374e+00 2.76069671e-01 7.99409032e-01
-1.00110424e+00 -9.07332122e-01 3.95725638e-01 3.05289000e-01
7.07170248e-01 -1.73700601e-01 1.31681550e+00 -8.62726688e-01
-6.21600032e-01 4.84763831e-02 1.65901557e-02 -7.84746651e-03
-3.98724735e-01 7.10442543e-01 -9.29309666e-01 -1.19249746e-01
-1.84342399e-01 -6.66987777e-01 1.70080677e-01 -7.39486590e-02
1.26404214e+00 -5.25514662e-01 -1.44854337e-01 1.55585095e-01
1.44841743e+00 3.31767052e-01 4.72490281e-01 4.48843241e-01
4.92838353e-01 6.30634427e-01 2.63249338e-01 3.91690791e-01
4.62677002e-01 1.34925067e-01 -7.58734420e-02 4.37505662e-01
4.34344202e-01 -3.27915668e-01 7.49021471e-01 9.51501846e-01
6.03439882e-02 9.46034044e-02 -1.71312201e+00 6.91547275e-01
-1.40609252e+00 -7.34089792e-01 -9.06356692e-01 2.05140996e+00
1.52388418e+00 3.39874268e-01 8.52378458e-02 -2.44122893e-01
3.45445365e-01 -2.98065633e-01 6.98435456e-02 -9.99257863e-01
6.30693436e-01 5.55744886e-01 3.10780287e-01 8.35514605e-01
-2.19214663e-01 6.79154813e-01 7.34175634e+00 4.90732223e-01
-8.79751742e-01 -8.82716030e-02 4.63560879e-01 3.60433936e-01
-1.01719224e+00 3.66908640e-01 -9.76428688e-01 5.57696760e-01
1.33504641e+00 -6.52265549e-01 4.84557152e-01 1.40089023e+00
-1.11316971e-01 -6.09068155e-01 -1.57409179e+00 4.91236597e-01
-2.13156313e-01 -1.15998292e+00 -5.40359378e-01 -2.29638129e-01
1.24342537e+00 -4.65887278e-01 -1.58763781e-01 8.60785782e-01
1.14981484e+00 -1.28812170e+00 8.16848099e-01 3.24192911e-01
2.06289113e-01 -5.51040053e-01 2.96173126e-01 8.57851863e-01
-6.77530646e-01 -2.94004399e-02 -3.05825502e-01 -6.64292395e-01
-7.26992190e-01 3.71900827e-01 -1.49959314e+00 -4.71616119e-01
5.85178792e-01 8.05002917e-03 -1.15606248e+00 1.40476573e+00
-4.38516468e-01 1.06983078e+00 8.82998947e-03 -5.39051652e-01
-1.52040988e-01 1.76783457e-01 2.50636756e-01 1.49352968e+00
2.97860473e-01 2.09006578e-01 3.15649420e-01 1.39903879e+00
3.16301703e-01 2.73458846e-02 -4.43932354e-01 -2.94144481e-01
8.85829031e-01 1.23276746e+00 -7.20267653e-01 -5.64773381e-01
-4.05422896e-01 4.62544322e-01 2.50191033e-01 4.01675552e-01
-6.18735075e-01 -8.37632954e-01 5.16552091e-01 1.87719211e-01
-6.67102695e-01 -3.17696899e-01 -8.11333776e-01 -8.19525957e-01
-2.10456908e-01 -1.47105682e+00 -2.51905382e-01 -8.42050195e-01
-9.61236298e-01 1.11113645e-01 5.54658063e-02 -6.69710994e-01
-1.79851234e-01 -6.20294392e-01 -1.24470568e+00 1.23363912e+00
-8.49035501e-01 -4.55242805e-02 -9.23327804e-01 -1.17540799e-01
5.19979119e-01 -6.87262416e-02 7.49687433e-01 1.00694694e-01
-2.17160210e-01 6.97043359e-01 -3.85279149e-01 -5.23981079e-02
9.66665804e-01 -1.77993798e+00 8.02984476e-01 7.14590728e-01
-3.23908776e-01 1.37992811e+00 1.04084826e+00 -6.17037892e-01
-1.27414382e+00 -5.26969373e-01 9.89723146e-01 -8.10529172e-01
3.91420782e-01 -1.79081723e-01 -1.04242289e+00 1.01916051e+00
4.10351962e-01 -4.46666867e-01 9.15673852e-01 9.96408612e-02
-3.53303879e-01 4.50282574e-01 -8.62809658e-01 6.72629356e-01
6.33600414e-01 -5.57243586e-01 -8.40138376e-01 5.65452158e-01
5.27419627e-01 -8.47981930e-01 -7.48207808e-01 -1.95283547e-01
4.31146055e-01 -1.32704937e+00 5.55919528e-01 -4.67381746e-01
1.00940287e+00 -5.70171960e-02 3.02804083e-01 -1.32592893e+00
-1.10195562e-01 -6.83408082e-01 3.14369857e-01 1.27909803e+00
3.72479081e-01 -6.41341150e-01 7.50482559e-01 1.24793589e+00
-3.48917216e-01 -5.07073045e-01 1.86553404e-01 -5.29997051e-01
6.25562668e-01 -3.78251165e-01 4.58560973e-01 1.28294659e+00
5.07367611e-01 -6.89352155e-02 2.05125988e-01 -3.00239086e-01
4.98544484e-01 -1.05302133e-01 1.12211072e+00 -1.42863810e+00
-6.19540274e-01 -7.87042677e-01 1.15166679e-01 -1.21531117e+00
3.23845036e-02 -1.08903146e+00 3.34029794e-01 -1.13466334e+00
5.32131612e-01 -7.59602129e-01 5.86226404e-01 6.66739762e-01
-4.05261964e-01 -4.29924816e-01 1.28742322e-01 -8.44809413e-02
-3.96636307e-01 -3.00602019e-01 9.14930344e-01 2.00923979e-01
-2.15996012e-01 -9.26060081e-02 -1.00994730e+00 9.71229434e-01
5.99182189e-01 -6.91622853e-01 -3.10083002e-01 -4.40893531e-01
7.90024221e-01 2.97906607e-01 3.53458136e-01 -1.29325581e+00
4.41097468e-01 -7.16714740e-01 2.91503400e-01 5.05939946e-02
-7.69716978e-01 -4.69415277e-01 5.77893667e-02 7.47553408e-01
-8.59332860e-01 4.41473305e-01 4.40568447e-01 -2.48929039e-01
1.60436466e-01 -1.25428522e+00 7.86234856e-01 -4.62237358e-01
-5.09761691e-01 -4.26510364e-01 -1.09002113e+00 6.83657527e-01
1.05950606e+00 -1.43240705e-01 -7.28433251e-01 -1.30626112e-01
-1.97630107e-01 2.53771931e-01 7.37828255e-01 2.74867654e-01
3.03485185e-01 -9.39862788e-01 -6.28945470e-01 8.97123590e-02
1.17945097e-01 -1.63378537e-01 -2.74471310e-03 7.11469352e-01
-1.23211694e+00 3.18136275e-01 -2.45248824e-01 -6.94120705e-01
-1.26833868e+00 -6.21459410e-02 1.27530426e-01 1.02583215e-01
-2.85974205e-01 1.31242251e+00 5.38850203e-02 -1.04284286e+00
3.28686386e-01 -5.48332512e-01 1.77096024e-01 -1.62827536e-01
7.58945286e-01 5.58512926e-01 -6.25629351e-02 4.36441630e-01
-1.63085550e-01 1.27593085e-01 -3.42139602e-01 5.03527857e-02
1.08817971e+00 6.43854678e-01 -1.66104972e-01 6.77288294e-01
7.65491426e-01 5.06597996e-01 -9.03520286e-01 2.48339057e-01
3.77623618e-01 -7.01978564e-01 -7.44028747e-01 -9.31805849e-01
-3.94595653e-01 8.67493272e-01 1.42757684e-01 4.17297721e-01
3.50753784e-01 -2.69714683e-01 1.49408013e-01 5.26732028e-01
5.54122567e-01 -8.65292668e-01 4.43144530e-01 8.60408843e-01
6.75390720e-01 -1.25513649e+00 -2.48041153e-01 -2.55181253e-01
-3.82019997e-01 1.54850662e+00 1.51039767e+00 -1.32420257e-01
4.50341776e-02 4.39811110e-01 4.79553547e-03 -3.46375294e-02
-1.11330140e+00 8.70524824e-01 -1.75244689e-01 4.51548696e-01
1.51487756e+00 1.31413430e-01 -5.46602964e-01 7.04791248e-01
-6.92038536e-01 3.83302495e-02 1.56938529e+00 1.29139996e+00
-8.19007576e-01 -1.20666981e+00 -7.86479115e-01 4.90646750e-01
-3.04087639e-01 -4.41013008e-01 -4.13040757e-01 5.70713520e-01
-6.04938231e-02 8.43214691e-01 -1.63759038e-01 -3.77241939e-01
4.84388359e-02 4.76161629e-01 6.49338067e-01 -1.30430615e+00
-1.22319245e+00 -8.24215651e-01 -6.76639974e-02 -2.18598410e-01
3.60912442e-01 -6.82393193e-01 -1.31645465e+00 -8.84078979e-01
-1.52457044e-01 5.08286595e-01 3.88609171e-01 5.94315886e-01
4.12785150e-02 5.65070212e-01 2.22956449e-01 -4.24212933e-01
-8.26893985e-01 -8.28622580e-01 -3.06050479e-02 -4.36781570e-02
2.48279378e-01 -1.60105020e-01 -4.19285595e-01 1.91302016e-01] | [9.26189136505127, 7.439583778381348] |
56dbe940-9934-49c2-8d8c-8aaf655299d7 | partial-auc-optimization-based-deep-speaker | 1911.08077 | null | https://arxiv.org/abs/1911.08077v1 | https://arxiv.org/pdf/1911.08077v1.pdf | Partial AUC optimization based deep speaker embeddings with class-center learning for text-independent speaker verification | Deep embedding based text-independent speaker verification has demonstrated superior performance to traditional methods in many challenging scenarios. Its loss functions can be generally categorized into two classes, i.e., verification and identification. The verification loss functions match the pipeline of speaker verification, but their implementations are difficult. Thus, most state-of-the-art deep embedding methods use the identification loss functions with softmax output units or their variants. In this paper, we propose a verification loss function, named the maximization of partial area under the Receiver-operating-characteristic (ROC) curve (pAUC), for deep embedding based text-independent speaker verification. We also propose a class-center based training trial construction method to improve the training efficiency, which is critical for the proposed loss function to be comparable to the identification loss in performance. Experiments on the Speaker in the Wild (SITW) and NIST SRE 2016 datasets show that the proposed pAUC loss function is highly competitive with the state-of-the-art identification loss functions. | ['Xiao-Lei Zhang', 'Zhongxin Bai', 'Jingdong Chen'] | 2019-11-19 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 1.32661453e-02 -3.26645344e-01 2.24642269e-02 -9.30444479e-01
-1.13145244e+00 -5.71075261e-01 3.99900138e-01 -1.17978361e-02
-5.57336509e-01 2.76692927e-01 1.07192539e-01 -5.92365801e-01
2.80807950e-02 -2.43612975e-01 -5.09347916e-01 -8.27663481e-01
6.74291998e-02 1.65682286e-01 -8.76644030e-02 5.10535873e-02
6.61019683e-02 3.02273929e-01 -1.55415833e+00 -3.81426401e-02
7.80415654e-01 1.21653926e+00 -3.27928901e-01 5.44374824e-01
5.38021252e-02 3.64273459e-01 -6.88596249e-01 -6.93957090e-01
1.17250055e-01 -2.36377597e-01 -4.53018934e-01 -4.37756240e-01
4.10839766e-01 -4.75411713e-01 -2.28189930e-01 9.38733160e-01
1.21422756e+00 9.43291932e-02 6.05089664e-01 -1.42084253e+00
-7.30476022e-01 7.16273546e-01 -2.21950918e-01 -6.80777105e-03
1.37907729e-01 -9.08758193e-02 1.02485764e+00 -1.06330144e+00
2.38423273e-02 1.19634616e+00 9.99394298e-01 9.13300455e-01
-1.14789295e+00 -9.41856802e-01 5.34216687e-02 5.47904789e-01
-1.52888620e+00 -8.17274809e-01 8.52225006e-01 -4.19761926e-01
7.19103336e-01 2.27905765e-01 -1.22352786e-01 1.14038646e+00
-1.79398403e-01 9.13870156e-01 8.10596108e-01 -5.38363814e-01
1.92402735e-01 6.13306582e-01 2.80513287e-01 4.80848402e-01
-2.16339417e-02 3.00138265e-01 -5.93347013e-01 -1.06878132e-01
1.62308261e-01 -2.82899976e-01 -3.56240422e-01 -7.23478273e-02
-9.60669577e-01 9.61334050e-01 2.74886757e-01 1.80947587e-01
-5.22522293e-02 4.20759283e-02 6.56095982e-01 2.88494200e-01
4.34341341e-01 -1.17317423e-01 -2.93168604e-01 -3.20844948e-02
-1.09715521e+00 -2.60373531e-03 6.57268226e-01 5.37798822e-01
4.30659264e-01 3.99714381e-01 -3.99520874e-01 1.10387564e+00
7.94462979e-01 3.63699853e-01 6.80061221e-01 -2.84950525e-01
5.62515557e-01 4.08827394e-01 -1.58796962e-02 -4.42612171e-01
-1.48912102e-01 -6.21353090e-01 -7.14746356e-01 3.04800689e-01
3.58153999e-01 -1.37649804e-01 -6.93666279e-01 1.93995190e+00
3.81922275e-01 1.80306032e-01 1.90156758e-01 9.11616147e-01
1.02877843e+00 4.45254803e-01 1.30446330e-01 1.80616364e-01
1.44452262e+00 -8.23181450e-01 -9.93624806e-01 -9.25027430e-02
5.00696957e-01 -8.83213460e-01 9.26653385e-01 2.58082729e-02
-8.33746195e-01 -9.69242334e-01 -1.51004493e+00 -5.97315319e-02
-4.12469506e-01 4.51931328e-01 1.33852020e-01 1.26551032e+00
-1.05782485e+00 3.72411191e-01 -7.21145272e-01 3.09937391e-02
2.32362196e-01 4.87286448e-01 -3.85910600e-01 1.45097390e-01
-1.44545019e+00 7.52909541e-01 1.53879970e-01 3.22686374e-01
-1.10625076e+00 -8.03470612e-01 -1.07861865e+00 3.65737230e-01
-1.67500213e-01 -3.21773201e-01 1.25273526e+00 -5.83536029e-01
-1.88432002e+00 9.29575264e-01 -2.82123387e-01 -4.74470139e-01
5.07800639e-01 -1.12695657e-01 -5.36551297e-01 -3.75105411e-01
-2.32489169e-01 4.31548387e-01 1.04280913e+00 -1.15775526e+00
-2.91507214e-01 -6.62477076e-01 -1.95551723e-01 -2.60187864e-01
-8.03992808e-01 3.91110420e-01 -1.63766950e-01 -3.05977017e-01
-6.07004389e-02 -5.90524971e-01 2.46037096e-01 9.50576887e-02
-4.76598769e-01 -5.55043101e-01 1.00617588e+00 -1.07192695e+00
1.30646861e+00 -2.55982947e+00 -3.29571038e-01 1.84779558e-02
-2.80912947e-02 4.59769785e-01 -1.56708121e-01 2.66051203e-01
-1.63823307e-01 7.08888918e-02 -2.43245780e-01 -1.15877748e+00
5.89269400e-01 -1.96742803e-01 -1.23071030e-01 7.23012090e-01
2.56357700e-01 6.83798969e-01 -4.73025471e-01 -3.03421140e-01
1.44659787e-01 9.08798575e-01 -2.02412665e-01 3.53958219e-01
4.50859040e-01 2.39745721e-01 -2.59463321e-02 6.54894769e-01
9.70108092e-01 2.33630508e-01 -8.54023546e-02 -2.21827239e-01
-8.12440142e-02 6.40769005e-01 -1.20142078e+00 1.50705075e+00
-6.51423931e-01 7.04143822e-01 4.60146308e-01 -9.46038485e-01
1.17130804e+00 6.87755406e-01 8.16773400e-02 -2.66131371e-01
3.16572487e-01 4.03807163e-01 9.81716663e-02 -1.27633825e-01
1.99695021e-01 -2.83371657e-01 2.15941638e-01 5.31447768e-01
2.08312780e-01 4.10538763e-01 -3.37249219e-01 -2.10506916e-01
5.84611952e-01 -2.45145902e-01 -6.23787008e-02 -8.69417638e-02
1.12974155e+00 -7.95552731e-01 7.42415488e-01 5.20147145e-01
-6.72746301e-01 6.00030243e-01 2.26699039e-01 -6.77520037e-02
-8.38791430e-01 -1.04679012e+00 -4.69310254e-01 9.26376581e-01
-2.68480759e-02 -3.91755141e-02 -9.99191523e-01 -7.66619742e-01
1.98994860e-01 6.39566302e-01 -5.46444833e-01 -2.34486476e-01
-4.55560774e-01 -4.70226765e-01 1.12304187e+00 7.16434896e-01
5.86521029e-01 -7.40629733e-01 -1.99321061e-02 2.30037257e-01
-3.09501830e-02 -1.11439216e+00 -8.63586605e-01 1.51883274e-01
-4.50569868e-01 -7.19821393e-01 -9.12520468e-01 -9.61945117e-01
4.40190762e-01 -3.10572702e-02 5.56687832e-01 -2.14934364e-01
-5.92365768e-03 1.57809258e-01 -2.73638606e-01 -7.04943120e-01
-6.67061508e-01 7.68611431e-02 5.10305822e-01 4.82704043e-01
6.74605250e-01 -2.48945564e-01 -6.47657096e-01 5.66275001e-01
-7.03223705e-01 -6.85791969e-01 2.82210052e-01 1.20697880e+00
3.36258084e-01 -1.06906533e-01 8.15570295e-01 -1.16426840e-01
7.11513221e-01 2.75959857e-02 -7.00943291e-01 4.85982656e-01
-7.27879763e-01 1.78824648e-01 6.23542309e-01 -5.75302243e-01
-9.17415619e-01 5.48293069e-03 -5.62319756e-01 -4.67638582e-01
-3.92233878e-02 3.95381987e-01 -5.27332306e-01 -2.42484853e-01
4.44817841e-01 4.99184906e-01 1.57239750e-01 -6.57006800e-01
-4.38315570e-02 1.28409851e+00 4.11484510e-01 -2.04487294e-01
8.16209376e-01 1.62370875e-01 -3.31792384e-01 -7.14717746e-01
-4.39058423e-01 -8.30629528e-01 -3.58065158e-01 -7.02629536e-02
8.16300213e-01 -8.97802949e-01 -1.16013801e+00 6.72409713e-01
-1.31776643e+00 5.90143539e-02 -5.92939444e-02 6.56217992e-01
-2.31345266e-01 6.11384571e-01 -4.53075498e-01 -1.33155859e+00
-8.92373025e-01 -1.47810674e+00 1.24834239e+00 2.07824469e-01
1.51274949e-01 -9.70257819e-01 2.13452987e-02 4.74573582e-01
7.88958073e-01 -2.42185220e-02 6.35295630e-01 -1.01335394e+00
-1.38843983e-01 -6.73749924e-01 -1.65478766e-01 9.83363211e-01
2.11129785e-02 -1.96624286e-02 -1.67405033e+00 -5.79119444e-01
2.76356012e-01 -2.30409741e-01 9.19478238e-01 3.78578901e-01
1.16684461e+00 -1.93051383e-01 -7.52277821e-02 7.73414552e-01
1.26807058e+00 1.49924710e-01 6.76684618e-01 1.70762464e-01
3.22210342e-01 5.86799085e-01 1.03962891e-01 3.52166772e-01
3.85180384e-01 1.00570452e+00 1.67081654e-01 -1.72214195e-01
-1.01965144e-01 -2.54342854e-01 7.78384209e-01 8.90740633e-01
3.64483386e-01 -2.06874996e-01 -7.73987353e-01 4.83783662e-01
-1.57367814e+00 -1.02925789e+00 1.73817784e-01 2.58875537e+00
6.81820631e-01 5.61146699e-02 1.57230869e-01 7.38816142e-01
9.17950988e-01 9.20763984e-02 -5.49920380e-01 -6.52902544e-01
-6.19478598e-02 1.61864728e-01 4.13088202e-01 6.61590874e-01
-1.20337129e+00 5.08405626e-01 5.36291885e+00 7.57566988e-01
-1.35116303e+00 5.00033379e-01 4.35320497e-01 3.64647865e-01
-9.35242623e-02 -2.83868551e-01 -1.32651281e+00 5.15557349e-01
1.21988487e+00 -9.70089808e-02 1.33811131e-01 9.49804902e-01
1.96921751e-01 6.39140189e-01 -1.30452609e+00 1.17174065e+00
2.80211776e-01 -8.16367328e-01 -2.94730335e-01 -2.41482221e-02
3.86990100e-01 -6.98356889e-03 3.64336818e-01 5.83810091e-01
-6.37831166e-02 -9.72392023e-01 8.92431080e-01 4.21693325e-02
1.05379307e+00 -5.73013365e-01 1.12642181e+00 5.81010059e-02
-1.24321699e+00 -2.18050912e-01 -1.93947658e-01 4.75296915e-01
2.85326481e-01 4.39158320e-01 -9.30717826e-01 5.26229441e-01
4.50576246e-01 3.97084594e-01 -2.73858309e-01 1.11740637e+00
-2.80260235e-01 8.25813353e-01 -2.27767155e-01 -9.26073045e-02
7.88819194e-02 1.83963686e-01 4.34470952e-01 1.30535209e+00
2.49608234e-01 -5.51876724e-01 -2.34179452e-01 1.05873382e+00
-2.96791196e-01 4.24380638e-02 -1.95076793e-01 -4.05303612e-02
6.34559214e-01 1.08364058e+00 9.60028172e-02 -1.11563265e-01
-2.00373545e-01 9.27325189e-01 1.02114484e-01 2.93407977e-01
-9.99227345e-01 -8.18005919e-01 9.37737465e-01 -8.43620673e-02
5.16042531e-01 -4.48238477e-02 -4.21947300e-01 -1.10118759e+00
4.05261546e-01 -7.87791312e-01 1.97966620e-01 3.43643539e-02
-1.49312508e+00 7.26056159e-01 -4.18357968e-01 -1.39729214e+00
-8.16467926e-02 -7.59752989e-01 -9.46944952e-01 1.26644886e+00
-1.98657978e+00 -1.29760325e+00 -3.00882757e-01 4.59749132e-01
3.44966799e-01 -3.79608065e-01 9.54445839e-01 7.87433326e-01
-8.57835770e-01 1.83111572e+00 3.82430941e-01 4.94737208e-01
7.65734613e-01 -1.02400649e+00 5.93197048e-01 8.12123656e-01
7.19531551e-02 6.94417596e-01 3.48476440e-01 5.83431832e-02
-1.35500014e+00 -8.87166262e-01 1.07533395e+00 -1.54051602e-01
4.03407127e-01 -8.11412692e-01 -9.37761128e-01 2.74816751e-01
-1.25846878e-01 -6.20950535e-02 9.66748476e-01 1.57159641e-01
-7.73684800e-01 -5.40765762e-01 -1.36383235e+00 4.29555364e-02
5.25840104e-01 -9.61799204e-01 -1.92960218e-01 1.58465043e-01
5.87212145e-01 -1.02575794e-01 -8.01873863e-01 4.88814116e-01
8.81719530e-01 -7.73407161e-01 1.12231040e+00 -4.67764199e-01
-4.34394367e-02 -3.50603521e-01 -2.52007812e-01 -1.10797453e+00
-2.20976993e-01 -5.35301387e-01 -1.42304242e-01 1.68676305e+00
6.90014601e-01 -9.44494307e-01 6.53803170e-01 5.82335114e-01
-2.07007140e-01 -6.53352201e-01 -1.43097794e+00 -1.20240796e+00
3.67293111e-03 -3.89859945e-01 7.90304124e-01 8.03063750e-01
-1.65768743e-01 1.87080771e-01 -2.49073461e-01 3.75789553e-01
8.57321203e-01 -3.84235024e-01 6.58705413e-01 -1.12402213e+00
-2.30417907e-01 -5.44994414e-01 -8.22960436e-01 -9.47027624e-01
4.42352474e-01 -8.92935455e-01 1.66588992e-01 -1.40580654e+00
3.35907713e-02 -4.47322786e-01 -6.70290411e-01 5.23999691e-01
-3.85435283e-01 -5.50555103e-02 -2.14736443e-04 3.13057527e-02
-9.86908302e-02 1.05395961e+00 6.08918309e-01 -5.66901982e-01
-1.88853219e-01 2.16552615e-01 -4.25372809e-01 8.83928314e-02
8.03404689e-01 -4.39020008e-01 -9.13233161e-02 -4.15335000e-01
-5.55429935e-01 -3.48177671e-01 1.51153773e-01 -1.02033174e+00
2.80505031e-01 5.66393375e-01 8.09674487e-02 -4.97858554e-01
5.33455789e-01 -7.30445325e-01 -1.96176067e-01 5.59401691e-01
-4.73112822e-01 -1.46531522e-01 3.02035928e-01 3.56240094e-01
-5.74032545e-01 -2.89340377e-01 1.00412464e+00 5.82926989e-01
-2.64365733e-01 2.18330011e-01 8.44383240e-02 -2.17193946e-01
6.23251617e-01 -2.91955143e-01 -1.76002979e-01 -4.01196718e-01
-2.71659225e-01 1.55479312e-01 -5.45220785e-02 5.41279256e-01
7.36823916e-01 -1.40384591e+00 -1.24404967e+00 4.73715544e-01
3.26220334e-01 -5.05933464e-01 2.69526809e-01 8.23261559e-01
-2.73964882e-01 5.16454399e-01 2.31402680e-01 -6.37121499e-01
-1.50671077e+00 4.19094980e-01 5.31620264e-01 -2.54769325e-01
-4.01739404e-02 1.06530309e+00 1.25826001e-01 -8.67310762e-01
8.39767814e-01 -1.76470697e-01 2.45368090e-02 -1.25557601e-01
1.00192475e+00 4.21538204e-01 4.29114729e-01 -7.17990220e-01
-7.19258428e-01 3.58681709e-01 -1.66029304e-01 -2.13457674e-01
1.08097804e+00 1.52258232e-01 2.03904048e-01 1.93575889e-01
1.65904510e+00 -1.67289928e-01 -9.76633310e-01 -3.56671721e-01
-1.30985394e-01 -3.05718899e-01 3.55728894e-01 -8.08045506e-01
-8.83701444e-01 1.28838062e+00 1.24738455e+00 -1.14812385e-02
8.86835575e-01 -3.17989796e-01 1.01004028e+00 1.20086089e-01
-1.84817705e-02 -9.27639842e-01 -3.57577741e-01 3.85046959e-01
8.18871081e-01 -1.60232615e+00 -3.74453276e-01 -1.17663227e-01
-3.35354805e-01 1.09556675e+00 4.09901053e-01 4.37702447e-01
8.92466307e-01 2.10592791e-01 3.46362412e-01 1.59468979e-01
-3.11732978e-01 6.96856007e-02 6.24894381e-01 5.33368051e-01
6.88426554e-01 1.74208358e-01 -3.56639415e-01 8.81658375e-01
-1.07489973e-01 -2.52920091e-01 -1.30240262e-01 4.65126455e-01
-9.04465690e-02 -1.28957748e+00 -4.76263702e-01 -2.03194451e-02
-5.25193989e-01 -4.95431460e-02 -2.96396762e-01 3.92896235e-01
-2.67151855e-02 1.30443132e+00 -2.32064530e-01 -7.29656518e-01
4.23826426e-01 3.75673562e-01 1.71361446e-01 -1.88686252e-01
-8.04484725e-01 -3.13121289e-01 -1.04125172e-01 -1.36716098e-01
-1.57902464e-01 -4.87061977e-01 -9.38285887e-01 -3.29740614e-01
-9.86483872e-01 1.04118533e-01 1.50930333e+00 6.29537106e-01
3.43059510e-01 2.39834636e-01 1.10537326e+00 -6.04645431e-01
-1.18961716e+00 -1.24789464e+00 -5.72621942e-01 3.83802295e-01
6.17652714e-01 -6.61106646e-01 -7.19048619e-01 -1.80010825e-01] | [14.275091171264648, 6.060482978820801] |
a1746700-0c95-4d0c-8d87-486105e58267 | survey-of-aspect-based-sentiment-analysis | 2204.05232 | null | https://arxiv.org/abs/2204.05232v4 | https://arxiv.org/pdf/2204.05232v4.pdf | Survey of Aspect-based Sentiment Analysis Datasets | Aspect-based sentiment analysis (ABSA) is a natural language processing problem that requires analyzing user-generated reviews to determine: a) The target entity being reviewed, b) The high-level aspect to which it belongs, and c) The sentiment expressed toward the targets and the aspects. Numerous yet scattered corpora for ABSA make it difficult for researchers to identify corpora best suited for a specific ABSA subtask quickly. This study aims to present a database of corpora that can be used to train and assess autonomous ABSA systems. Additionally, we provide an overview of the major corpora for ABSA and its subtasks and highlight several features that researchers should consider when selecting a corpus. Finally, we discuss the advantages and disadvantages of current collection approaches and make recommendations for future corpora creation. This survey examines 65 publicly available ABSA datasets covering over 25 domains, including 45 English and 20 other languages datasets. | ['Thamar Solorio', 'Nedim Lipka', 'Franck Dernoncourt', 'Siva Uday Sampreeth Chebolu'] | 2022-04-11 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 8.96861702e-02 9.23705474e-02 -4.16061997e-01 -8.46834123e-01
-7.98173070e-01 -9.66051817e-01 9.21140373e-01 6.92297637e-01
-4.13662910e-01 5.23075163e-01 2.79127836e-01 -4.46412534e-01
2.03773201e-01 -6.90887392e-01 -4.17458385e-01 -3.82178515e-01
2.46450230e-01 6.70045674e-01 7.94163346e-02 -7.04943776e-01
5.41177571e-01 2.02971593e-01 -1.76292038e+00 5.75542450e-01
7.22597241e-01 8.58302653e-01 -4.81810793e-02 5.24119616e-01
-5.61216295e-01 7.34633625e-01 -1.13181973e+00 -8.88541877e-01
6.61084632e-05 -3.70271623e-01 -9.82362688e-01 2.14756504e-01
5.00795424e-01 4.64878768e-01 6.02838397e-01 8.69252741e-01
1.40638620e-01 6.58091903e-02 7.23248243e-01 -1.27453399e+00
-4.52947170e-01 7.12888777e-01 -4.13451314e-01 3.14698040e-01
6.47979975e-01 -2.86560744e-01 1.59313977e+00 -9.16364193e-01
8.88369381e-01 1.05369270e+00 4.53942150e-01 5.05405903e-01
-5.68610609e-01 -4.14240032e-01 6.03805482e-01 -2.76553422e-01
-7.32754767e-01 -5.23261368e-01 6.36589527e-01 -4.97397989e-01
1.50004387e+00 3.56709272e-01 9.58871067e-01 9.07465875e-01
4.83155221e-01 1.05824912e+00 1.36896884e+00 -6.98674321e-01
4.44330961e-01 6.02196693e-01 7.07931519e-01 2.41479442e-01
6.59216762e-01 -7.81490922e-01 -9.38032448e-01 -4.31450933e-01
-3.29613984e-01 -7.64580250e-01 3.39622378e-01 -2.79842436e-01
-1.04574299e+00 9.80658591e-01 -3.18841964e-01 4.54744458e-01
-5.19862890e-01 -5.14707088e-01 6.10281229e-01 4.21308190e-01
6.49966657e-01 1.07619655e+00 -1.00913322e+00 -3.54073793e-01
-6.66243315e-01 7.18448222e-01 1.14205861e+00 1.18288159e+00
6.99460328e-01 3.15512829e-02 1.82887524e-01 9.84833598e-01
5.83174169e-01 6.60958886e-01 5.90000212e-01 -5.77519715e-01
5.47278881e-01 1.06605911e+00 2.37716362e-01 -1.03963375e+00
-3.73825580e-01 5.45836650e-02 1.19918801e-01 1.01655900e-01
3.39285433e-01 -2.14965656e-01 -6.89303935e-01 1.29761744e+00
3.96812916e-01 -9.43342149e-01 5.76240242e-01 6.13088787e-01
1.04541707e+00 7.01751888e-01 1.50402278e-01 -1.70141652e-01
2.02170348e+00 -9.60585535e-01 -8.29180658e-01 -1.04181206e+00
7.26709843e-01 -1.25854671e+00 1.15377665e+00 4.05567139e-01
-1.12114918e+00 -1.62315920e-01 -1.23702502e+00 -1.17322914e-01
-1.05034518e+00 2.57604450e-01 8.83156478e-01 8.08687985e-01
-9.04400408e-01 5.30097680e-03 -5.95791698e-01 -7.10275233e-01
8.99837688e-02 1.81901872e-01 -1.60910279e-01 2.72371739e-01
-1.15644085e+00 9.66788292e-01 6.56015947e-02 -2.76588142e-01
-3.44621450e-01 -3.31976622e-01 -1.20525098e+00 -3.82174551e-01
1.76765621e-01 -3.44185919e-01 1.57322335e+00 -1.39490879e+00
-1.14763725e+00 1.39149380e+00 -6.30410075e-01 -4.64462876e-01
-1.96485445e-01 -4.06009227e-01 -6.91783607e-01 -1.29597783e-01
6.97915614e-01 3.98602307e-01 6.27816617e-01 -9.34823871e-01
-1.23665714e+00 -5.67137301e-01 5.83293796e-01 6.88766003e-01
-4.35929447e-01 6.88086331e-01 -4.03593212e-01 -5.77628911e-01
-3.64121735e-01 -8.95821393e-01 -1.84989363e-01 -7.03875363e-01
-2.93370515e-01 -5.88898361e-01 5.30221701e-01 -3.17336380e-01
1.20482349e+00 -1.99080634e+00 -1.90036207e-01 5.99809475e-02
-3.48892100e-02 1.30027279e-01 -1.97166964e-01 6.90323412e-01
7.89289922e-02 1.67707622e-01 4.39582467e-02 -1.20183721e-01
-1.25171930e-01 -1.11156695e-01 -4.27046627e-01 5.72505184e-02
2.17609406e-01 5.28973043e-01 -9.36265469e-01 -4.55537140e-01
-1.75354630e-01 1.78313971e-01 -1.01688258e-01 1.48280263e-01
-2.23493338e-01 -1.18482992e-01 -8.26911032e-01 9.33113039e-01
2.82875836e-01 -1.38730302e-01 1.73436642e-01 -6.77096769e-02
-3.31000417e-01 9.14285064e-01 -8.47307742e-01 1.08909678e+00
-6.43697500e-01 9.61647689e-01 3.45497690e-02 -5.92774093e-01
1.33462369e+00 1.77124649e-01 3.76371980e-01 -5.64714789e-01
3.59985858e-01 3.65751475e-01 1.79411303e-02 -3.34829837e-01
1.11944151e+00 -1.62906379e-01 -5.69056630e-01 7.10737050e-01
-8.30735713e-02 -4.74544406e-01 9.44850862e-01 4.09949154e-01
8.15984309e-01 -1.22877002e-01 7.86231220e-01 -7.50698864e-01
8.55902374e-01 7.14988470e-01 5.65382540e-01 4.47528720e-01
-4.03426617e-01 2.90954351e-01 6.03325427e-01 -7.38647461e-01
-8.75129163e-01 -4.07346725e-01 -6.52806014e-02 1.29174900e+00
8.36625844e-02 -8.99688363e-01 -6.34768605e-01 -8.47319484e-01
-1.41136870e-01 1.02284753e+00 -8.88397455e-01 2.39041194e-01
-2.33713880e-01 -8.15988302e-01 9.43612382e-02 3.64922047e-01
6.77208304e-02 -1.19089365e+00 -5.84410787e-01 -6.02480099e-02
-3.27100039e-01 -1.21044040e+00 -2.40459338e-01 2.28722796e-01
-6.46003246e-01 -1.22381485e+00 -2.07192987e-01 -1.03305912e+00
6.05545282e-01 4.41557318e-01 1.54726493e+00 -2.05083072e-01
2.12593630e-01 4.53827143e-01 -7.51432478e-01 -1.27270293e+00
-5.62504947e-01 3.92283738e-01 -3.81473475e-03 -1.63222894e-01
1.37184966e+00 1.72998101e-01 -2.23611966e-01 1.95470437e-01
-5.07643819e-01 -2.26334929e-01 3.57806534e-01 2.67157167e-01
5.82314491e-01 3.68410237e-02 5.83690524e-01 -1.52072263e+00
1.18459010e+00 -4.22766954e-01 -3.38483006e-01 3.54929805e-01
-8.01120400e-01 -5.35606921e-01 4.80870068e-01 2.33094487e-02
-1.15216267e+00 -3.63911279e-02 -8.21697041e-02 5.85176170e-01
-2.24865526e-01 7.54743397e-01 -7.36988261e-02 3.59179378e-01
8.44356120e-01 -5.06595075e-02 -2.05099195e-01 2.11195531e-03
1.10288672e-01 9.11361098e-01 -2.39293933e-01 -4.52077836e-01
3.44943941e-01 4.62121457e-01 -6.23150706e-01 -1.07688355e+00
-1.40793669e+00 -8.64262342e-01 -4.71027613e-01 -4.43887532e-01
7.38887906e-01 -1.11505532e+00 4.49401066e-02 4.25752431e-01
-9.74183977e-01 -1.96205452e-02 -3.34114879e-01 3.22388411e-01
-3.68797511e-01 -9.06668827e-02 -4.44987148e-01 -7.02887535e-01
-8.60180438e-01 -1.32713008e+00 9.45297658e-01 4.85975832e-01
-1.08933270e+00 -1.02054644e+00 4.17539597e-01 1.00184631e+00
1.59138352e-01 -1.29389539e-01 7.36271501e-01 -1.14503706e+00
2.58077800e-01 -5.70838153e-01 3.15548956e-01 2.43585959e-01
4.17671323e-01 5.49598932e-01 -9.89786029e-01 -1.87192470e-01
1.06602408e-01 -5.07515669e-01 2.29344875e-01 4.43054318e-01
1.87760636e-01 -1.49211422e-01 -2.04530135e-01 -1.46390751e-01
1.00999784e+00 5.56756020e-01 2.45888382e-01 1.05319691e+00
3.71742576e-01 9.37992811e-01 1.57034695e+00 1.38276607e-01
7.56805837e-01 2.20217198e-01 1.64306292e-03 2.49942228e-01
2.10753426e-01 3.94174233e-02 7.65907049e-01 1.42770469e+00
1.62443910e-02 -3.71371567e-01 -9.72769141e-01 1.07194519e+00
-1.43870389e+00 -5.34520686e-01 -2.90978879e-01 1.68157387e+00
8.40040445e-01 4.00252551e-01 1.63610905e-01 -1.92263946e-01
4.24231976e-01 5.75520992e-01 -4.54173058e-01 -1.17052448e+00
-3.67312342e-01 -1.37359589e-01 -1.91042274e-02 3.20451975e-01
-1.39558160e+00 1.12742972e+00 6.99167156e+00 4.17215496e-01
-7.72875369e-01 -1.08785458e-01 7.55577326e-01 1.17377922e-01
-5.54298341e-01 1.58547983e-01 -1.10897183e+00 2.78990828e-02
1.18973398e+00 -6.28711700e-01 -1.05280943e-01 1.29785478e+00
8.31050575e-02 -2.89674938e-01 -9.65312302e-01 4.37722743e-01
5.26271760e-01 -1.06669891e+00 3.39444607e-01 -2.42251948e-01
1.01064801e+00 2.85883963e-01 -5.13471849e-02 3.46346974e-01
4.76212233e-01 -6.11245155e-01 8.47955704e-01 -1.41555130e-01
3.26430947e-01 -7.66192675e-01 1.06619000e+00 -1.48000777e-01
-1.10952711e+00 3.29126805e-01 -4.82162058e-01 -2.26856142e-01
3.17432061e-02 4.15156037e-01 -6.58308446e-01 3.18812370e-01
1.02282715e+00 1.02848256e+00 -8.50485981e-01 4.66864347e-01
-3.53946209e-01 5.89711845e-01 9.43346620e-02 -7.66566992e-01
5.31390429e-01 -4.93825495e-01 7.52197623e-01 1.33443522e+00
6.26334250e-02 -1.19711667e-01 -1.03822656e-01 1.40717506e-01
-1.27097629e-02 8.84586573e-01 -9.17988837e-01 -4.32717264e-01
4.04258728e-01 1.54944956e+00 -1.00881672e+00 -4.87782419e-01
-8.99146914e-01 3.19081515e-01 1.21694051e-01 2.41662025e-01
-1.99664801e-01 -7.62975037e-01 8.95019114e-01 -2.97846030e-02
1.45860776e-01 8.76866356e-02 -4.95491266e-01 -1.27566600e+00
1.58321127e-01 -1.49656773e+00 6.27559006e-01 -8.85353863e-01
-1.22976530e+00 1.08040178e+00 -2.87043244e-01 -1.35834551e+00
-2.51420587e-01 -7.64985323e-01 -6.89151227e-01 6.10522926e-01
-1.42037916e+00 -1.15802193e+00 -6.50414452e-03 8.84250104e-02
9.74427760e-01 -5.73768497e-01 8.12905312e-01 -1.05550207e-01
-4.27443624e-01 1.59451917e-01 -2.47258589e-01 7.77592659e-02
1.06001210e+00 -1.34221673e+00 7.85891175e-01 7.66180396e-01
1.86894923e-01 8.22007179e-01 1.02791989e+00 -7.68845558e-01
-1.31819963e+00 -7.98654318e-01 1.69247890e+00 -9.10922706e-01
1.06926644e+00 -3.06379765e-01 -5.92164338e-01 8.72512817e-01
7.32756078e-01 -8.80380094e-01 1.12888801e+00 7.17922330e-01
-3.02649766e-01 -2.69494593e-01 -9.24046457e-01 6.98804319e-01
1.77776292e-01 -5.57242692e-01 -8.88837755e-01 2.33761489e-01
4.51182872e-01 -2.48254135e-01 -7.25688279e-01 1.78113386e-01
5.08017719e-01 -7.67350316e-01 4.52160448e-01 -7.36161053e-01
6.56893134e-01 -1.88218579e-01 -5.41503951e-02 -1.46143460e+00
1.16594434e-02 -3.46089900e-01 3.01804066e-01 1.46139002e+00
1.17981684e+00 -5.63581169e-01 7.22480536e-01 1.03573310e+00
-2.87842061e-02 -9.13561583e-01 -4.11122203e-01 -3.06890637e-01
2.27011472e-01 -6.22361720e-01 6.32058263e-01 1.16320586e+00
4.91943777e-01 8.36718082e-01 1.75522551e-01 -2.24789940e-02
1.43974721e-01 5.83171010e-01 8.71479690e-01 -8.81541073e-01
3.23767871e-01 -5.44834137e-01 -2.40971223e-01 -5.24830163e-01
-2.79280916e-02 -6.58360422e-01 9.75270793e-02 -1.68666911e+00
1.77663684e-01 -2.25795224e-01 -1.18362375e-01 3.99937063e-01
-2.53786832e-01 1.07925475e-01 -7.08779320e-02 8.31537768e-02
-7.51796007e-01 2.65703261e-01 8.49982738e-01 -2.59797752e-01
-2.59192199e-01 1.88837126e-01 -1.35058033e+00 1.13112175e+00
1.03212869e+00 -3.65694195e-01 -5.40048242e-01 -4.82822448e-01
9.62244809e-01 -5.77907979e-01 -6.50432348e-01 -4.59013283e-01
2.09123284e-01 -4.88649577e-01 -4.25408874e-03 -7.74141014e-01
2.27605030e-01 -5.06563246e-01 -5.74820995e-01 -1.58953760e-02
-4.62973028e-01 6.16878271e-01 3.48034918e-01 2.18730360e-01
-6.78695977e-01 -6.03768349e-01 5.07093310e-01 -1.40747443e-01
-7.48843014e-01 -1.43424407e-01 -1.14475894e+00 4.65422392e-01
1.06184709e+00 5.36283143e-02 -1.75891668e-01 -4.04846370e-01
-2.36236587e-01 3.11053842e-01 5.73587537e-01 7.61946678e-01
4.18666452e-01 -1.02071738e+00 -5.82822442e-01 -5.67868948e-02
7.11308181e-01 -1.39587104e-01 -3.49982665e-03 3.66352469e-01
-5.68368554e-01 6.45388782e-01 -1.39473423e-01 -8.23784843e-02
-1.59582710e+00 1.89023644e-01 -6.98253438e-02 -3.84600192e-01
-2.04235420e-01 7.95451999e-01 -1.36061218e-02 -6.96414769e-01
-1.53390080e-01 -1.02778122e-01 -1.06688309e+00 6.27799630e-01
7.31931150e-01 6.15588352e-02 3.39884520e-01 -1.16162813e+00
-5.66047609e-01 4.16058362e-01 -5.07429659e-01 -1.45774677e-01
1.27452552e+00 -3.85548383e-01 -3.39738995e-01 9.90117848e-01
8.08005393e-01 3.06239903e-01 -3.07161987e-01 -1.03987396e-01
1.62616059e-01 -2.65567750e-01 -1.06490143e-01 -8.26680064e-01
-9.39142287e-01 5.67382872e-01 -6.43920973e-02 6.62662387e-01
9.70880687e-01 1.39317796e-01 5.16636312e-01 6.26835346e-01
2.77901053e-01 -1.64322758e+00 -1.91457763e-01 7.46298611e-01
9.57808495e-01 -1.25748527e+00 3.67079556e-01 -3.60392660e-01
-1.36350060e+00 9.72436726e-01 8.28961432e-01 -3.75944190e-02
7.57394791e-01 1.52038068e-01 8.86433959e-01 -7.26916730e-01
-1.08379388e+00 -5.43489233e-02 3.33235472e-01 4.77577329e-01
8.60473990e-01 1.67693481e-01 -7.73156106e-01 7.95854926e-01
-7.75159955e-01 -6.35885358e-01 9.13106859e-01 1.21600091e+00
-2.58529663e-01 -1.16810310e+00 -1.26105756e-01 7.83115804e-01
-7.65208066e-01 -2.21715216e-02 -1.00798512e+00 7.94018149e-01
-3.73704940e-01 1.27478421e+00 -7.99102262e-02 -1.37669235e-01
6.07268214e-01 5.22119030e-02 -1.81980222e-01 -8.88637722e-01
-9.72977698e-01 1.82154290e-02 8.74167919e-01 -3.64319772e-01
-9.32822049e-01 -1.21999955e+00 -1.01791942e+00 3.86416656e-03
-4.01520610e-01 6.52498841e-01 9.10928905e-01 1.04196167e+00
3.43136787e-01 7.11663812e-02 4.64353114e-01 -9.96029228e-02
3.11122811e-03 -9.80511904e-01 -5.70128798e-01 4.02897626e-01
5.08312322e-02 -2.86626458e-01 -2.65474409e-01 2.00493678e-01] | [11.280167579650879, 6.8393354415893555] |
e30fccdb-1a2c-4ee3-a857-e8b266c162a8 | latent-predictor-networks-for-code-generation | 1603.06744 | null | http://arxiv.org/abs/1603.06744v2 | http://arxiv.org/pdf/1603.06744v2.pdf | Latent Predictor Networks for Code Generation | Many language generation tasks require the production of text conditioned on
both structured and unstructured inputs. We present a novel neural network
architecture which generates an output sequence conditioned on an arbitrary
number of input functions. Crucially, our approach allows both the choice of
conditioning context and the granularity of generation, for example characters
or tokens, to be marginalised, thus permitting scalable and effective training.
Using this framework, we address the problem of generating programming code
from a mixed natural language and structured specification. We create two new
data sets for this paradigm derived from the collectible trading card games
Magic the Gathering and Hearthstone. On these, and a third preexisting corpus,
we demonstrate that marginalising multiple predictors allows our model to
outperform strong benchmarks. | ['Fumin Wang', 'Tomáš Kočiský', 'Edward Grefenstette', 'Wang Ling', 'Karl Moritz Hermann', 'Phil Blunsom', 'Andrew Senior'] | 2016-03-22 | latent-predictor-networks-for-code-generation-1 | https://aclanthology.org/P16-1057 | https://aclanthology.org/P16-1057.pdf | acl-2016-8 | ['card-games'] | ['playing-games'] | [ 5.09039104e-01 2.67505765e-01 -3.51139344e-02 -3.79718035e-01
-8.02445889e-01 -8.66674781e-01 9.40632105e-01 -7.71253780e-02
-3.40035498e-01 9.54423368e-01 1.86329201e-01 -7.49790609e-01
2.67272651e-01 -1.07883465e+00 -8.85345340e-01 -1.38743848e-01
-7.62952715e-02 6.52687252e-01 -5.97157851e-02 -3.95341843e-01
1.92988306e-01 -6.10289164e-02 -1.71238494e+00 6.12076104e-01
8.75116050e-01 6.63744390e-01 3.71290416e-01 1.03133929e+00
-2.37755477e-01 1.12650180e+00 -7.07540870e-01 -3.58768493e-01
3.32863063e-01 -4.80845034e-01 -7.91055560e-01 -2.65271425e-01
2.97408342e-01 -3.92218560e-01 1.74114674e-01 6.74757421e-01
2.89648741e-01 1.27168685e-01 6.41562343e-01 -1.11712277e+00
-6.14029527e-01 1.33498585e+00 6.88206032e-02 -2.04384267e-01
4.45020705e-01 4.93864924e-01 1.30242741e+00 -6.16190493e-01
8.00289929e-01 1.04248703e+00 5.31897187e-01 7.47481763e-01
-1.69571280e+00 -5.39524376e-01 -1.92546938e-02 -7.65494227e-01
-1.12922239e+00 -5.34493446e-01 3.24602395e-01 -7.23751545e-01
1.23879528e+00 5.72188906e-02 6.22221649e-01 1.30592060e+00
1.14437573e-01 7.37367511e-01 9.36294436e-01 -7.00948119e-01
3.16832602e-01 1.82852238e-01 5.28119467e-02 7.81171560e-01
9.70100090e-02 4.09620613e-01 -2.06516892e-01 -4.57758427e-01
7.68154263e-01 -3.73119652e-01 -6.82355762e-02 -2.46356964e-01
-1.18812239e+00 9.96552527e-01 -1.47336110e-01 -1.37504069e-02
-8.63505062e-04 5.86620927e-01 4.47952777e-01 4.80323851e-01
1.13544554e-01 1.03857613e+00 -5.12969077e-01 -4.39299017e-01
-9.35807824e-01 6.62325621e-01 1.24702525e+00 1.29560804e+00
6.93224669e-01 1.81968555e-01 -3.81047845e-01 5.91544509e-01
1.24893278e-01 2.75575101e-01 6.31247401e-01 -8.42331231e-01
7.87912607e-01 3.93696934e-01 3.09371352e-01 -3.87554199e-01
-1.21985003e-01 -1.26534626e-01 -4.92029160e-01 4.36991364e-01
6.57041311e-01 -7.78543472e-01 -1.08350849e+00 1.99575913e+00
-2.38044947e-01 -3.00724983e-01 1.14541501e-01 4.63541985e-01
4.70206529e-01 6.86290622e-01 5.29886484e-02 1.36595175e-01
9.84164417e-01 -8.34518850e-01 -1.15074627e-01 -3.96995246e-01
7.52720535e-01 -4.80065942e-01 1.25652659e+00 7.35607445e-01
-1.47803974e+00 -4.72271621e-01 -1.05683219e+00 -1.67229548e-01
-3.85515600e-01 4.42843139e-02 1.07459521e+00 5.48646867e-01
-1.22022724e+00 6.96396410e-01 -7.33326674e-01 2.04510265e-03
4.66543920e-02 5.66487551e-01 -1.80904746e-01 1.18189171e-01
-9.91863430e-01 6.44229591e-01 7.48788238e-01 -1.86654598e-01
-9.51362252e-01 -6.19606614e-01 -8.71874571e-01 2.21194237e-01
2.87656784e-01 -8.65922868e-01 1.78511000e+00 -9.78776157e-01
-1.74594283e+00 6.99617684e-01 2.89130211e-01 -5.80807209e-01
6.79502487e-01 -8.44029337e-03 7.58900046e-02 -3.42788458e-01
1.43795433e-02 7.80163407e-01 7.01342821e-01 -9.10213113e-01
-6.12735391e-01 2.13383168e-01 3.24561357e-01 -1.05658285e-01
1.44972533e-01 1.57319620e-01 -1.64558977e-01 -6.41870856e-01
-5.01122534e-01 -8.20597231e-01 -4.55682606e-01 -4.71507430e-01
-4.02230263e-01 -1.78806305e-01 -6.84052706e-02 -5.62135458e-01
1.11356592e+00 -1.96149969e+00 3.06161284e-01 2.94931829e-01
1.12763800e-01 -2.61909813e-01 -3.34771335e-01 5.09891272e-01
-2.71034185e-02 4.51899111e-01 -3.68007004e-01 -1.85864851e-01
6.07704520e-01 1.59532487e-01 -6.53263688e-01 -1.92508236e-01
6.79358184e-01 1.08359063e+00 -9.03481305e-01 -3.27041566e-01
-2.91230768e-01 4.67170775e-02 -1.13233662e+00 3.49826485e-01
-9.71977174e-01 1.33568019e-01 -3.86075974e-01 3.54487985e-01
1.19860448e-01 -2.64904708e-01 3.31058592e-01 6.43205762e-01
-1.82832256e-01 7.49743223e-01 -1.19122481e+00 1.86050522e+00
-7.77935684e-01 3.92464578e-01 7.64580816e-03 -4.87156034e-01
9.90008891e-01 3.67422044e-01 1.37734283e-02 -4.76655096e-01
9.60628614e-02 3.21061283e-01 1.76173985e-01 -1.98976845e-01
8.59487355e-01 -2.82045364e-01 -6.52250886e-01 1.00635684e+00
1.73152298e-01 -7.03173697e-01 7.28926659e-01 2.97486633e-01
1.22144413e+00 4.64957982e-01 2.24869162e-01 -2.77921200e-01
9.48444158e-02 5.92107438e-02 4.82968181e-01 1.13404083e+00
5.16288698e-01 6.34929538e-01 9.34563041e-01 -5.46204090e-01
-1.33047199e+00 -9.23620343e-01 1.01836406e-01 1.41904330e+00
-5.28360903e-01 -6.18821442e-01 -7.29687452e-01 -6.08002007e-01
7.23029627e-03 9.96868730e-01 -5.51460564e-01 -2.23907195e-02
-9.41690445e-01 -6.45637572e-01 7.35281885e-01 5.77689528e-01
-1.22074649e-01 -1.53110015e+00 -7.63128698e-01 3.52430284e-01
1.92564741e-01 -7.76776016e-01 -4.84300256e-01 6.81378603e-01
-6.68122768e-01 -5.69942653e-01 -3.04342568e-01 -8.72285485e-01
5.44970036e-01 -6.04128659e-01 1.78632486e+00 2.43543461e-01
-2.35026047e-01 1.31686041e-02 -1.29875630e-01 -2.18501419e-01
-1.05138481e+00 5.22173941e-01 -2.97995299e-01 -5.12091994e-01
3.89665887e-02 -7.99548686e-01 3.71557847e-03 -2.18334898e-01
-1.12037313e+00 6.41957939e-01 6.82437003e-01 1.07658720e+00
6.35843873e-02 -2.00662971e-01 4.31101412e-01 -1.27119768e+00
8.84199202e-01 -6.08769715e-01 -9.76269782e-01 2.29178056e-01
-4.53887314e-01 6.02140129e-01 1.06292188e+00 -4.28514600e-01
-9.87025857e-01 1.25148878e-01 -1.99863520e-02 1.37620032e-01
-2.46703014e-01 6.60845816e-01 -9.25202295e-02 3.84700090e-01
9.24370170e-01 2.73444921e-01 -2.13389024e-01 -2.63530493e-01
7.03398883e-01 4.74291116e-01 7.14365780e-01 -1.33070087e+00
7.27188706e-01 -1.37225017e-01 -2.84346431e-01 -3.05328459e-01
-1.38757929e-01 3.06449354e-01 -4.91421640e-01 4.46397334e-01
5.44171393e-01 -7.18806148e-01 -4.79820609e-01 2.17289373e-01
-1.44005692e+00 -9.95852411e-01 -4.44574058e-01 1.70371026e-01
-8.46955299e-01 -1.69645235e-01 -8.15546751e-01 -7.82816887e-01
-3.01519603e-01 -1.13921404e+00 8.49480093e-01 -1.03236757e-01
-6.86199546e-01 -8.83886218e-01 1.71748027e-01 -1.72119871e-01
4.74105924e-01 2.66306937e-01 1.31517160e+00 -7.74623036e-01
-8.91891599e-01 -1.17397450e-01 -5.07259136e-03 2.77809292e-01
-1.05964124e-01 2.43787542e-01 -6.67096257e-01 -8.31544995e-02
-3.28346163e-01 -7.73390532e-01 6.56590819e-01 -2.06427962e-01
9.59471524e-01 -5.01742065e-01 -5.66484220e-02 8.14492583e-01
1.29937136e+00 8.57593417e-02 5.59383869e-01 1.86637804e-01
4.27640438e-01 3.96374047e-01 -1.27603533e-02 5.56603491e-01
2.28425935e-01 3.84396970e-01 1.43436074e-01 2.57010043e-01
2.06728101e-01 -6.09020531e-01 4.45986390e-01 6.00876451e-01
2.25698501e-02 -2.43993953e-01 -1.13303936e+00 5.54023564e-01
-1.54200447e+00 -9.69405711e-01 3.00444871e-01 2.13385296e+00
1.42939794e+00 3.18237573e-01 6.10878468e-02 -1.19373344e-01
4.04796690e-01 -1.07526518e-01 -3.32355410e-01 -6.73498452e-01
5.45609407e-02 9.17503536e-01 2.73087829e-01 6.53438628e-01
-9.05879855e-01 8.72246265e-01 7.33284855e+00 6.34145558e-01
-1.11656630e+00 -2.77893364e-01 5.11908293e-01 -3.41846436e-01
-8.72982025e-01 5.75578660e-02 -8.15841138e-01 6.00967646e-01
1.19711185e+00 -2.76800156e-01 9.65035796e-01 7.99581647e-01
-1.24844559e-01 -5.11124693e-02 -1.67491663e+00 4.11599874e-01
-1.68826595e-01 -1.35541499e+00 1.16198376e-01 7.69096017e-02
7.33620703e-01 3.86535074e-03 2.02249810e-02 7.66564369e-01
1.38812590e+00 -1.46526039e+00 1.12766063e+00 3.82571489e-01
9.32220340e-01 -5.82896233e-01 2.49313712e-01 5.62613130e-01
-6.61579669e-01 -2.41707265e-01 7.61051625e-02 -4.53887016e-01
1.01135001e-02 2.53872633e-01 -1.13263750e+00 1.57688707e-01
5.38646020e-02 2.64792204e-01 -4.80817467e-01 6.08493209e-01
-4.60411429e-01 6.08263791e-01 -2.60598123e-01 -1.90633938e-01
2.44160250e-01 -1.96755573e-01 1.80333763e-01 1.40014994e+00
3.89599502e-01 -8.14642534e-02 2.35652253e-01 1.50099480e+00
-1.12935178e-01 -1.56031445e-01 -7.41758704e-01 -3.42058361e-01
2.70083100e-01 1.17863750e+00 -4.77612764e-01 -3.96752834e-01
-3.43504250e-01 5.97183406e-01 5.88279128e-01 4.30201173e-01
-7.26112962e-01 -6.45299733e-01 3.46181870e-01 7.50223873e-03
4.84487772e-01 -4.56845284e-01 -4.88515854e-01 -1.26134050e+00
6.59563467e-02 -1.26234639e+00 2.41639480e-01 -6.09843612e-01
-1.07675028e+00 6.14036083e-01 7.87806660e-02 -7.19974816e-01
-1.14164352e+00 -6.66358948e-01 -7.60426760e-01 1.12629664e+00
-1.06414843e+00 -9.30534840e-01 1.16430961e-01 9.39916745e-02
3.02963614e-01 -3.12465996e-01 9.32103693e-01 -4.53842245e-03
-4.24896121e-01 6.04111254e-01 -1.89785957e-01 3.49982828e-01
3.68394136e-01 -1.65158248e+00 9.39307988e-01 6.96210802e-01
2.02154934e-01 1.04441285e+00 6.38580441e-01 -4.62374926e-01
-1.40747321e+00 -1.02562881e+00 8.58529568e-01 -6.46148086e-01
7.74748325e-01 -8.62630785e-01 -5.02178609e-01 9.64169681e-01
4.05359685e-01 -3.96069884e-01 6.08369112e-01 8.91157165e-02
-4.46639955e-01 3.00266355e-01 -8.63342881e-01 7.92630613e-01
9.96637642e-01 -4.70983535e-01 -6.96748614e-01 5.65154143e-02
8.87756884e-01 -6.18827641e-01 -6.85254753e-01 9.05157328e-02
6.46270931e-01 -8.01382065e-01 5.34654975e-01 -9.58887517e-01
1.16134441e+00 -9.22982097e-02 -2.04438418e-01 -1.40793121e+00
-2.18938351e-01 -1.14770591e+00 -4.13352624e-02 1.29682899e+00
9.42867160e-01 -4.34826195e-01 7.63429046e-01 1.05599129e+00
-2.02087924e-01 -5.14283478e-01 -3.62747818e-01 -4.80451524e-01
6.44990027e-01 -5.65059841e-01 9.87405181e-01 5.68037808e-01
2.78326958e-01 5.97231567e-01 -1.77700400e-01 -3.64244133e-01
1.01477407e-01 3.74907374e-01 9.67680216e-01 -1.06503654e+00
-1.02552176e+00 -6.43849194e-01 1.24776512e-01 -1.08155465e+00
2.77024418e-01 -1.24269271e+00 3.99108976e-01 -1.12921751e+00
9.07405913e-02 -7.46038377e-01 4.57125530e-02 6.04706824e-01
-7.56299645e-02 -4.20135930e-02 2.20564649e-01 -2.59655952e-01
-3.74745309e-01 2.83655971e-01 9.22091961e-01 -1.61381856e-01
-2.91999131e-01 -6.39865026e-02 -8.88285339e-01 5.05592287e-01
7.61577547e-01 -2.96888620e-01 -4.81738657e-01 -7.48358786e-01
8.20947587e-01 3.04628432e-01 2.80747682e-01 -7.65499294e-01
1.27055779e-01 -4.79363680e-01 1.46412060e-01 -1.05246916e-01
4.10526879e-02 -2.96437800e-01 1.70028925e-01 1.82920814e-01
-9.00228620e-01 2.65665829e-01 2.62384355e-01 2.23216966e-01
1.11293949e-01 -4.45270658e-01 3.88700098e-01 -6.00023568e-01
-4.84480262e-01 5.37263267e-02 -5.02506256e-01 4.98207569e-01
6.66339874e-01 1.95824653e-02 -3.73758227e-01 -2.02921271e-01
-4.66096520e-01 1.61704570e-01 6.75383389e-01 3.14552039e-01
2.77588814e-01 -1.19451594e+00 -8.02805901e-01 6.10353112e-01
-8.62716883e-02 3.03430200e-01 -4.78409499e-01 1.73293039e-01
-6.44205213e-01 3.69118720e-01 -3.20996076e-01 -2.07409129e-01
-6.34142697e-01 3.91698897e-01 1.87685132e-01 -5.82540095e-01
-3.64026785e-01 6.64411664e-01 1.07059062e-01 -7.89145350e-01
-1.07725253e-02 -8.76121461e-01 2.20633551e-01 -3.22090805e-01
4.03432846e-01 -3.28232460e-02 -7.81367049e-02 7.07962289e-02
2.49059871e-01 -2.18387708e-01 -6.88721761e-02 -5.83596528e-01
1.42177391e+00 4.78230834e-01 -3.75174344e-01 5.47024131e-01
7.83055186e-01 1.65356666e-01 -1.27899444e+00 -2.64008850e-04
1.86352313e-01 -1.91222206e-01 -4.86394674e-01 -9.74073529e-01
-6.28768682e-01 6.96685910e-01 -2.55813926e-01 3.82884085e-01
7.94996738e-01 -3.24461073e-01 5.87827086e-01 5.59757888e-01
5.82928181e-01 -8.39117825e-01 7.20443502e-02 1.01389623e+00
7.41878271e-01 -9.49909866e-01 -4.74577874e-01 2.63454542e-02
-4.64385480e-01 1.17229354e+00 6.60829306e-01 -3.00040305e-01
4.78910320e-02 9.16742980e-01 -1.56765327e-01 9.68492925e-02
-1.49131811e+00 1.57826871e-01 7.83167034e-03 5.01465619e-01
7.59712815e-01 1.43856525e-01 -9.74088609e-02 1.03285468e+00
-8.70228827e-01 2.77796865e-01 7.88529456e-01 9.99517918e-01
-1.71913385e-01 -1.51823556e+00 -2.12011442e-01 7.69469798e-01
-5.16209304e-01 -6.03115261e-01 -3.33488047e-01 7.95584381e-01
2.62868404e-01 5.96074760e-01 1.02436163e-01 -1.76788285e-01
3.58140729e-02 4.51277345e-01 5.70655644e-01 -1.27670097e+00
-8.97196651e-01 -2.99112022e-01 4.03590441e-01 -1.16026886e-01
2.89277881e-01 -6.19497836e-01 -1.00463128e+00 -2.19264641e-01
-7.02698380e-02 3.27192605e-01 3.22205603e-01 6.63569212e-01
1.78414434e-01 4.15714920e-01 3.52756262e-01 -7.99022198e-01
-9.85651553e-01 -9.04837072e-01 -6.00952208e-01 4.01551843e-01
3.49853933e-01 -5.06947041e-02 -1.47973195e-01 3.70647281e-01] | [8.194466590881348, 7.5525407791137695] |
5af2891d-be61-42e3-b2f8-b7d9e03bea66 | investigating-the-contribution-of | null | null | https://aclanthology.org/W14-1505 | https://aclanthology.org/W14-1505.pdf | Investigating the Contribution of Distributional Semantic Information for Dialogue Act Classification | null | ['Matthew Purver', 'Dmitrijs Milajevs'] | 2014-04-01 | null | null | null | ws-2014-4 | ['dialogue-act-classification'] | ['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] | [-7.379645824432373, 3.7125189304351807] |
e394f459-36b9-4548-9c57-fb789cea9833 | mcscset-a-specialist-annotated-dataset-for | 2210.11720 | null | https://arxiv.org/abs/2210.11720v1 | https://arxiv.org/pdf/2210.11720v1.pdf | MCSCSet: A Specialist-annotated Dataset for Medical-domain Chinese Spelling Correction | Chinese Spelling Correction (CSC) is gaining increasing attention due to its promise of automatically detecting and correcting spelling errors in Chinese texts. Despite its extensive use in many applications, like search engines and optical character recognition systems, little has been explored in medical scenarios in which complex and uncommon medical entities are easily misspelled. Correcting the misspellings of medical entities is arguably more difficult than those in the open domain due to its requirements of specificdomain knowledge. In this work, we define the task of Medical-domain Chinese Spelling Correction and propose MCSCSet, a large scale specialist-annotated dataset that contains about 200k samples. In contrast to the existing open-domain CSC datasets, MCSCSet involves: i) extensive real-world medical queries collected from Tencent Yidian, ii) corresponding misspelled sentences manually annotated by medical specialists. To ensure automated dataset curation, MCSCSet further offers a medical confusion set consisting of the commonly misspelled characters of given Chinese medical terms. This enables one to create the medical misspelling dataset automatically. Extensive empirical studies have shown significant performance gaps between the open-domain and medical-domain spelling correction, highlighting the need to develop high-quality datasets that allow for Chinese spelling correction in specific domains. Moreover, our work benchmarks several representative Chinese spelling correction models, establishing baselines for future work. | ['Yefeng Zheng', 'Yujiu Yang', 'Bang Liu', 'Siheng Li', 'Yi Liu', 'Jianguang Zheng', 'Ruihui Zhao', 'Zijing Ou', 'Zhihao Ye', 'Wangjie Jiang'] | 2022-10-21 | null | null | null | null | ['spelling-correction'] | ['natural-language-processing'] | [ 7.32681751e-01 -1.63637176e-01 5.76442247e-03 -1.86010107e-01
-1.21366704e+00 -5.19630194e-01 3.31663668e-01 6.39104068e-01
-8.34203780e-01 8.29442441e-01 3.12717497e-01 -4.72111344e-01
1.13769583e-01 -3.37459385e-01 -2.89786696e-01 -4.99870956e-01
3.89057308e-01 7.32889533e-01 3.23603511e-01 -6.89118579e-02
4.32353646e-01 1.03455864e-01 -9.90252435e-01 7.32869387e-01
1.70824015e+00 3.61992270e-01 5.02659500e-01 9.77622867e-01
-9.12879854e-02 5.84241331e-01 -7.75878370e-01 -7.09328115e-01
-6.30018041e-02 -5.15221000e-01 -1.00758123e+00 1.26382066e-02
3.19859415e-01 2.06421360e-01 -4.88384478e-02 1.27445924e+00
6.89812303e-01 -2.82758623e-01 5.29509783e-01 -5.74489057e-01
-9.84499454e-01 3.11808944e-01 -2.46434554e-01 3.93364370e-01
4.29058611e-01 -1.18815795e-01 7.79464483e-01 -7.56737292e-01
8.09577942e-01 6.49800897e-01 7.79536307e-01 9.83344615e-01
-6.32946253e-01 -4.60391939e-01 -3.03568304e-01 1.54528961e-01
-1.45519614e+00 -1.31590888e-01 -4.39585187e-02 -2.78340071e-01
8.65971744e-01 7.70151913e-01 1.51775599e-01 8.49461973e-01
1.55446902e-01 1.10293555e+00 7.60699391e-01 -8.74736845e-01
2.74258759e-02 3.40109974e-01 6.74443841e-02 3.17823946e-01
6.22906625e-01 -3.41752172e-01 -1.56714842e-01 -3.29634935e-01
2.87130713e-01 1.64221376e-01 -5.61964869e-01 2.18695581e-01
-1.49252152e+00 5.21738708e-01 4.07031216e-02 5.70126295e-01
-9.38944444e-02 -4.26803648e-01 6.43893361e-01 1.91555321e-01
5.36794424e-01 1.01527786e+00 -8.27546954e-01 -1.76685408e-01
-1.19995821e+00 2.15805948e-01 6.24469757e-01 1.34627116e+00
1.40671253e-01 -6.25417888e-01 -4.83211905e-01 1.02161634e+00
-1.85212225e-01 6.63736820e-01 1.02167797e+00 -2.22308502e-01
7.72158861e-01 7.37604141e-01 1.77723795e-01 -1.05514228e+00
-3.67017388e-01 -4.70922559e-01 -7.32811630e-01 -9.08333123e-01
4.37755883e-01 -1.53833181e-01 -8.97490680e-01 1.08904445e+00
8.50766376e-02 1.25629485e-01 2.66376913e-01 8.03345501e-01
7.79506505e-01 1.58301428e-01 7.00537860e-02 -2.07801133e-01
1.61042452e+00 -9.51101720e-01 -8.83329272e-01 -2.78958023e-01
1.21123731e+00 -1.17375970e+00 9.78985310e-01 3.44651341e-01
-6.86544180e-01 5.99672925e-03 -4.73199666e-01 -1.37973294e-01
-4.62169051e-01 3.76107812e-01 4.06167507e-01 8.71403039e-01
-6.90231085e-01 3.05984437e-01 -7.66512692e-01 -5.57515323e-01
4.51038510e-01 -2.04550531e-02 -2.14748457e-01 -4.40338194e-01
-1.27253640e+00 9.55862999e-01 4.57210600e-01 -1.78063735e-01
-3.24052125e-01 -1.13590455e+00 -8.15901160e-01 -1.29222333e-01
4.39058274e-01 -5.18294871e-01 1.51891708e+00 -8.89479220e-01
-5.95085680e-01 1.30389273e+00 -2.39947975e-01 -5.16293406e-01
5.87915778e-01 -3.45136017e-01 -1.08315480e+00 2.72838265e-01
3.92158121e-01 1.49440259e-01 5.17298222e-01 -6.31097555e-01
-9.34442341e-01 -1.27786666e-01 -5.72230399e-01 1.22506343e-01
-4.23035681e-01 3.92113686e-01 -8.69966984e-01 -1.34091866e+00
-1.03783444e-01 -1.16628969e+00 -4.69464272e-01 -3.73994112e-01
-6.64955974e-01 1.55012980e-02 2.74458677e-01 -9.31772768e-01
1.99653864e+00 -2.11806178e+00 -3.93207401e-01 1.26042232e-01
2.67741889e-01 9.60171878e-01 -7.87586719e-02 4.82245445e-01
-5.44342361e-02 3.09848726e-01 -6.88460469e-01 -2.10376799e-01
-3.99137735e-01 1.38662368e-01 -1.51321009e-01 3.43568236e-01
2.51041144e-01 9.66395736e-01 -1.32202673e+00 -8.10742795e-01
-1.88696489e-01 -5.18462695e-02 -7.96049416e-01 -1.49859816e-01
-8.23172107e-02 1.92996606e-01 -4.48999435e-01 8.56416643e-01
7.62023926e-01 -4.36137140e-01 1.52875930e-01 2.97867060e-01
7.68058300e-02 3.83467853e-01 -8.69512081e-01 1.59422946e+00
-1.50168240e-01 5.16936779e-01 -3.91867518e-01 -5.12104869e-01
5.19297957e-01 4.43474203e-01 3.46159190e-01 -8.15437675e-01
-5.15316948e-02 5.98109663e-01 -1.72617525e-01 -8.54337633e-01
8.50996912e-01 -7.64728561e-02 -3.68557841e-01 2.17001602e-01
-3.16531211e-01 -9.80558544e-02 5.04111290e-01 5.35962284e-01
1.17133605e+00 -4.10906494e-01 5.32538056e-01 -2.54977494e-01
8.22946548e-01 3.99865299e-01 7.95787394e-01 9.22300935e-01
-3.86490494e-01 1.05890501e+00 -1.97636671e-02 -2.17457622e-01
-9.09352243e-01 -5.06727934e-01 -4.02120590e-01 6.34934008e-01
-1.47347555e-01 -7.80636668e-01 -1.10236216e+00 -9.60509479e-01
-1.47417262e-01 6.48782551e-01 -5.40182233e-01 -5.73578030e-02
-6.74772084e-01 -8.84254992e-01 1.15324020e+00 4.01224941e-01
2.16578841e-01 -8.97124946e-01 -5.28677523e-01 1.83394790e-01
-5.72615504e-01 -1.19766736e+00 -1.28630888e+00 -7.17139989e-02
-6.52217686e-01 -1.52561736e+00 -1.18959093e+00 -1.04573011e+00
1.04709804e+00 2.63846159e-01 1.09912324e+00 5.44260144e-01
-7.68140078e-01 1.21123008e-01 -8.35406899e-01 -7.87513971e-01
-8.21167827e-01 1.65430576e-01 -5.26265241e-02 -2.96995252e-01
9.77197766e-01 3.95249039e-01 -6.06887817e-01 1.96173221e-01
-1.15772879e+00 2.96385251e-02 6.72276378e-01 8.46109092e-01
4.86530900e-01 -2.34616995e-01 2.69728065e-01 -1.47397554e+00
7.15229273e-01 -4.03438777e-01 -3.16757083e-01 6.15195036e-01
-6.75461233e-01 -6.55358359e-02 5.79187334e-01 -2.36581609e-01
-1.00274897e+00 -1.97363384e-02 -2.58623749e-01 2.42373943e-01
-2.82660306e-01 7.11452842e-01 3.43948632e-01 1.67912006e-01
8.66063595e-01 4.99145299e-01 -2.65819371e-01 -7.42045760e-01
-6.35173544e-02 1.03621519e+00 8.13443065e-01 -8.37792084e-02
5.49487174e-01 1.87178135e-01 -6.28151178e-01 -9.34676707e-01
-9.55141544e-01 -9.79565263e-01 -6.92757845e-01 3.45265657e-01
8.45411181e-01 -8.02360117e-01 -2.28289247e-01 5.17296076e-01
-9.75190282e-01 1.18464939e-01 6.65014163e-02 4.66916263e-01
6.60163583e-03 9.40882385e-01 -8.54872465e-01 -4.18734729e-01
-5.49893498e-01 -9.83629048e-01 1.13510728e+00 1.36620730e-01
-6.87016785e-01 -9.43566740e-01 2.04958931e-01 5.46451271e-01
8.58167559e-02 -5.70342764e-02 8.81203711e-01 -9.67850387e-01
-3.12587358e-02 -5.74066579e-01 -1.29787460e-01 2.04176113e-01
2.30259806e-01 -1.49138570e-01 -5.44238448e-01 -2.72992641e-01
-2.28735313e-01 -2.43262481e-02 7.38671541e-01 1.18245356e-01
1.31713510e+00 -2.96705812e-01 -5.49762785e-01 5.38133085e-01
1.14243114e+00 1.70350686e-01 7.21747875e-01 4.45897996e-01
4.69139576e-01 2.15303764e-01 9.15757954e-01 5.62075317e-01
3.49531323e-01 3.10548156e-01 -1.91131219e-01 -1.65109336e-01
-6.23587407e-02 -2.74317741e-01 -1.85751989e-01 1.09850001e+00
2.18488604e-01 -2.14483947e-01 -1.33141458e+00 8.94101024e-01
-1.46220016e+00 -7.49437988e-01 -5.71853161e-01 2.24831390e+00
1.62769866e+00 -1.44909650e-01 -3.35318118e-01 2.04416707e-01
7.74461687e-01 -5.20724416e-01 -2.86966741e-01 -3.86551231e-01
-1.92793354e-01 2.62547582e-01 7.91954577e-01 2.42167398e-01
-1.16345382e+00 9.83785510e-01 5.77907324e+00 1.08218443e+00
-9.60312307e-01 -2.72935312e-02 5.15803456e-01 5.34169041e-02
-1.62860498e-01 -3.86500388e-01 -1.03726602e+00 8.14021230e-01
6.89612925e-01 3.20145413e-02 -4.99765202e-02 6.82865977e-01
1.32094964e-01 -2.51489460e-01 -9.09395874e-01 1.07823825e+00
4.76129293e-01 -1.52729630e+00 6.35309517e-02 -2.74561554e-01
1.11010683e+00 6.62320629e-02 -5.88146113e-02 1.84539720e-01
2.20047101e-01 -7.77200222e-01 4.94172215e-01 4.13944513e-01
1.18734467e+00 -5.59743106e-01 9.54169691e-01 5.74802458e-01
-5.86466908e-01 9.84029174e-02 -3.30902159e-01 2.57545322e-01
-9.67127830e-02 7.45477498e-01 -1.16712582e+00 4.43558007e-01
6.33363068e-01 6.68816686e-01 -1.02247405e+00 1.48518348e+00
-2.85869867e-01 7.23159730e-01 3.97768877e-02 -1.57506093e-01
2.12508678e-01 2.68361092e-01 2.82865912e-01 1.81346583e+00
2.98685908e-01 2.30024815e-01 -1.46490082e-01 4.11626637e-01
-1.44547289e-02 3.48408163e-01 -4.22242582e-02 -3.35075498e-01
5.75821579e-01 9.19396281e-01 -8.28438461e-01 -4.25443232e-01
-4.38332349e-01 1.41183531e+00 5.15323356e-02 1.49001375e-01
-6.74729347e-01 -7.81834364e-01 7.33759820e-01 -1.15788072e-01
1.32921010e-01 1.48632750e-01 -5.76972425e-01 -1.51486695e+00
-6.47136047e-02 -1.29629815e+00 7.48226881e-01 -3.32265705e-01
-1.19406259e+00 5.51687419e-01 -5.20261288e-01 -1.50151992e+00
4.57098857e-02 -5.43562889e-01 -1.78952143e-01 7.95331895e-01
-1.72723424e+00 -7.82806098e-01 -1.74001530e-01 6.46899462e-01
8.20657313e-01 -5.43707283e-03 9.57036078e-01 6.76876843e-01
-6.36917233e-01 1.13553417e+00 4.63909090e-01 4.29684371e-01
1.34469521e+00 -1.23680198e+00 5.63322663e-01 1.14891183e+00
3.57354097e-02 8.89879584e-01 8.17938983e-01 -1.12864029e+00
-8.26189697e-01 -1.28003752e+00 1.94792521e+00 -8.88863623e-01
3.47282410e-01 -1.84114620e-01 -1.05032945e+00 3.42002809e-01
-1.57162428e-01 -1.34937361e-01 9.97925639e-01 -3.44945759e-01
-3.21338996e-02 3.03341776e-01 -1.10356843e+00 6.42533123e-01
7.94480860e-01 -6.17228746e-01 -7.66118765e-01 6.36047304e-01
6.55871809e-01 -6.50657654e-01 -4.44266826e-01 1.71288639e-01
1.42336935e-01 -3.24994594e-01 7.54274845e-01 -8.22498262e-01
6.17129862e-01 -3.10404241e-01 2.34187171e-01 -1.44810426e+00
-3.09438020e-01 -5.26746154e-01 4.13287640e-01 9.98059988e-01
8.04642797e-01 -2.11731300e-01 6.30856335e-01 9.06508029e-01
-4.20028001e-01 -4.48869675e-01 -6.59679472e-01 -5.14187992e-01
2.11430907e-01 -4.82262254e-01 5.25204897e-01 1.18631399e+00
4.19665188e-01 -2.88815767e-01 -3.73712480e-01 2.28193477e-01
1.13049544e-01 -1.94006413e-01 3.46352279e-01 -9.08185601e-01
-2.58866162e-03 -2.09927738e-01 -9.13632382e-03 -7.66034424e-01
-2.71436512e-01 -1.04867625e+00 2.59055912e-01 -1.33829808e+00
4.72292513e-01 -7.11509585e-01 2.70973891e-02 2.83209920e-01
-8.43167186e-01 4.36320871e-01 1.57601330e-02 2.61401325e-01
-8.99326921e-01 -3.36620770e-02 1.16451943e+00 2.96827648e-02
-1.30546745e-02 1.56367913e-01 -8.47492516e-01 8.31815898e-01
5.33745885e-01 -6.98793590e-01 6.86120540e-02 -6.48373723e-01
3.75141710e-01 -2.31747046e-01 -3.90182026e-02 -7.78938591e-01
3.65504742e-01 -3.65969390e-01 1.35355651e-01 -6.51439965e-01
-4.56052929e-01 -6.61442101e-01 -3.03527415e-01 7.28868067e-01
-6.49845719e-01 1.53637856e-01 7.10047856e-02 6.06166840e-01
-1.33902773e-01 -5.05556047e-01 6.86450243e-01 -4.42015171e-01
-7.92733967e-01 1.00835748e-01 -8.22639763e-01 6.04177475e-01
1.00273192e+00 -3.60471636e-01 -2.76290745e-01 1.96909621e-01
-4.57947731e-01 3.04607987e-01 4.92790222e-01 5.27577817e-01
5.03346801e-01 -9.10099089e-01 -8.19523096e-01 1.81885704e-01
5.90109646e-01 -2.29516439e-02 3.87646377e-01 8.54693890e-01
-1.07854080e+00 9.33448255e-01 3.27092201e-01 -2.88105398e-01
-1.55153775e+00 7.40900815e-01 6.22532889e-02 -4.22003746e-01
-6.29834652e-01 1.12462378e+00 4.20978256e-02 -5.33812821e-01
1.99980944e-01 -7.26132035e-01 -1.84355736e-01 -7.80157000e-02
1.17266500e+00 2.77298510e-01 6.51161969e-01 -4.22620356e-01
-5.83447397e-01 1.10078298e-01 -4.89026338e-01 3.99728000e-01
9.99336123e-01 -3.55858535e-01 -1.60562098e-01 -8.31023604e-02
9.26444888e-01 3.49012226e-01 -4.46831226e-01 -5.36242008e-01
6.62204146e-01 -5.66015542e-01 -2.97300428e-01 -1.11897922e+00
-5.68251610e-01 5.67992210e-01 2.67918199e-01 -4.13868278e-01
1.15817022e+00 -1.49951041e-01 1.14170671e+00 4.02112603e-01
2.45102629e-01 -1.51674414e+00 -8.87817740e-02 7.15139389e-01
5.97650588e-01 -1.35362136e+00 -1.60478801e-01 -4.61436629e-01
-9.84610200e-01 8.77787471e-01 3.38486254e-01 2.87194967e-01
3.19413692e-01 1.93982914e-01 2.81477720e-01 1.60907935e-02
-4.62072194e-01 -6.44539744e-02 4.93531227e-01 6.68667138e-01
6.91698670e-01 2.56797194e-01 -7.83205211e-01 9.28358495e-01
-2.40411565e-01 2.11709559e-01 7.61980414e-01 9.59940851e-01
-2.10097864e-01 -1.18494940e+00 -4.71003085e-01 7.63102829e-01
-9.51953053e-01 -7.55465329e-01 -3.82485718e-01 3.37834209e-01
2.07776278e-01 9.28139508e-01 -1.61413968e-01 1.20083705e-01
2.05693170e-01 -1.11305103e-01 -1.24551123e-03 -1.04353619e+00
-9.08333659e-01 -7.77973160e-02 1.17533550e-01 -3.21619511e-01
-6.29972368e-02 -9.93849933e-01 -1.29098678e+00 -2.36626640e-02
-5.07983387e-01 4.02480990e-01 2.47017145e-01 9.90185380e-01
4.51941639e-01 3.68755996e-01 7.74944499e-02 1.00325890e-01
-4.93646383e-01 -7.73672283e-01 -5.24555624e-01 8.50675881e-01
4.41378325e-01 -3.95187549e-02 1.00289576e-01 2.79534400e-01] | [10.839962005615234, 10.572408676147461] |
7fff9aa4-d5ba-4832-ad9e-9239af031653 | a-nuclear-norm-model-for-multi-frame-super | 1704.06196 | null | http://arxiv.org/abs/1704.06196v1 | http://arxiv.org/pdf/1704.06196v1.pdf | A Nuclear-norm Model for Multi-Frame Super-Resolution Reconstruction from Video Clips | We propose a variational approach to obtain super-resolution images from
multiple low-resolution frames extracted from video clips. First the
displacement between the low-resolution frames and the reference frame are
computed by an optical flow algorithm. Then a low-rank model is used to
construct the reference frame in high-resolution by incorporating the
information of the low-resolution frames. The model has two terms: a 2-norm
data fidelity term and a nuclear-norm regularization term. Alternating
direction method of multipliers is used to solve the model. Comparison of our
methods with other models on synthetic and real video clips show that our
resulting images are more accurate with less artifacts. It also provides much
finer and discernable details. | ['Rui Zhao', 'Raymond H. Chan'] | 2017-04-17 | null | null | null | null | ['multi-frame-super-resolution'] | ['computer-vision'] | [ 1.67694926e-01 -1.95197657e-01 -1.27743557e-01 -1.39360055e-01
-8.90125453e-01 -1.73898265e-01 4.16665465e-01 -6.88651741e-01
-3.98220390e-01 1.07173789e+00 5.57202280e-01 7.11578250e-01
-7.83404261e-02 -4.15313423e-01 -4.91335809e-01 -6.85496092e-01
-7.06734732e-02 -1.68775663e-01 4.17978585e-01 -4.35196385e-02
3.66872519e-01 3.85148078e-01 -1.38302743e+00 3.96589696e-01
5.32977104e-01 7.58276403e-01 5.45996130e-01 8.29316616e-01
3.41928452e-01 1.09333098e+00 -6.07866943e-02 1.53002098e-01
5.31156957e-01 -6.46736085e-01 -9.89107192e-01 4.56758767e-01
7.15878308e-01 -6.27280056e-01 -6.89448714e-01 1.27158475e+00
2.91137576e-01 5.66222072e-01 2.65384227e-01 -5.83887875e-01
-6.25927210e-01 2.11907644e-02 -1.04153693e+00 6.85415745e-01
8.63607287e-01 -3.74597996e-01 9.00391698e-01 -1.27075255e+00
1.18700171e+00 1.46718550e+00 5.21678507e-01 4.56884503e-01
-1.43093991e+00 -1.87097833e-01 -1.22039698e-01 3.93591374e-01
-1.53714037e+00 -5.80241323e-01 8.95822167e-01 -4.64173675e-01
3.59086812e-01 2.54136443e-01 4.83245373e-01 8.75345826e-01
2.36232087e-01 3.40485871e-01 1.01876807e+00 -3.19149435e-01
-1.13518275e-01 -1.89780951e-01 -2.58876890e-01 7.99981356e-01
9.83117446e-02 3.66665334e-01 -7.99657941e-01 -2.32336000e-01
1.58896053e+00 -3.74708921e-02 -7.55230606e-01 -1.16718322e-01
-1.37101805e+00 6.17479503e-01 1.22791588e-01 5.02118707e-01
-7.14638472e-01 -2.20819563e-02 -5.27013279e-02 -2.82929569e-01
4.43511635e-01 2.29134306e-01 1.32226601e-01 2.58578938e-02
-1.28953564e+00 1.86137915e-01 4.22660828e-01 7.85651267e-01
9.06168997e-01 4.62590694e-01 -3.17874327e-02 8.14284742e-01
3.88322830e-01 2.85104632e-01 4.03422862e-01 -1.89019084e+00
2.81430036e-01 -1.49308383e-01 5.46444237e-01 -1.43099523e+00
-6.30969927e-02 -2.16502137e-02 -1.07719934e+00 3.72636229e-01
2.39141971e-01 1.26790211e-01 -4.81248736e-01 1.51142263e+00
3.17300081e-01 8.33991468e-01 -1.56958103e-02 1.49484968e+00
6.44523978e-01 9.12693977e-01 -4.06533360e-01 -8.80310416e-01
1.14623022e+00 -8.19905519e-01 -1.32709193e+00 1.90280259e-01
-2.82786906e-01 -8.95373821e-01 4.88290489e-01 3.57624888e-01
-1.78019202e+00 -9.11315024e-01 -1.11107874e+00 -3.60344797e-01
3.46485883e-01 -4.02162783e-02 -2.37138569e-03 -4.31815274e-02
-1.02241039e+00 8.29065681e-01 -8.80591333e-01 1.38339430e-01
1.88454345e-01 -4.88082729e-02 -6.05645359e-01 -3.18460912e-02
-1.23560691e+00 7.65119016e-01 1.07042655e-01 1.66881293e-01
-8.84149730e-01 -6.96018815e-01 -9.69720960e-01 -2.93683708e-01
9.30826813e-02 -7.37089872e-01 6.85798109e-01 -9.96082485e-01
-1.65158904e+00 7.16422319e-01 -5.53219974e-01 -1.22514576e-01
5.05668044e-01 -2.23026127e-01 -3.46220315e-01 8.00549388e-01
1.20008186e-01 3.54621887e-01 1.38927746e+00 -1.48734820e+00
-7.45904028e-01 7.28023574e-02 1.19547367e-01 3.64148647e-01
2.51662195e-01 -9.64329951e-03 -4.95590478e-01 -9.02072549e-01
3.85642856e-01 -7.19001234e-01 -2.43654072e-01 8.64649564e-02
2.43650116e-02 4.77861553e-01 7.97319591e-01 -1.06463277e+00
1.36602569e+00 -2.08954120e+00 7.28398561e-01 -1.48049504e-01
4.29045707e-01 1.00776121e-01 -4.18372266e-02 -1.36873499e-01
-6.69159070e-02 -1.99305162e-01 -2.64370143e-01 -2.06032127e-01
-7.45188951e-01 2.96327174e-01 -1.59976125e-01 6.95497334e-01
8.55945796e-02 5.58843613e-01 -1.16588068e+00 -7.89165974e-01
4.23077792e-01 1.16754675e+00 -6.39389694e-01 3.64252955e-01
4.78899211e-01 9.68918622e-01 -4.27211314e-01 2.09131554e-01
7.20578074e-01 -2.71676481e-01 -6.69127107e-02 -7.50803530e-01
-4.28443879e-01 -1.34036049e-01 -1.59101832e+00 1.95510328e+00
-5.74348867e-02 7.21634328e-01 7.05311239e-01 -7.61145949e-01
5.79668581e-01 4.18902338e-01 8.17205548e-01 -4.51473653e-01
6.29789801e-03 -1.35541588e-01 -4.73336518e-01 -7.99681544e-01
6.72592580e-01 -2.61424333e-01 5.04205167e-01 2.67197937e-02
-4.98517752e-02 -4.95583452e-02 5.43869913e-01 2.51657844e-01
6.12488747e-01 5.68981588e-01 2.41556481e-01 -3.54612201e-01
1.05460155e+00 -3.53038251e-01 1.05617678e+00 3.49893272e-01
-2.15551406e-01 1.20379317e+00 4.77342531e-02 -6.39144897e-01
-1.32586479e+00 -9.95407224e-01 -3.05964410e-01 5.63391149e-01
3.99366260e-01 -4.89458442e-01 -6.91523433e-01 5.25724888e-02
-4.63442981e-01 5.35283377e-03 -6.00908637e-01 2.65381545e-01
-9.72777247e-01 -4.82659727e-01 -7.73767978e-02 2.81780392e-01
7.25392282e-01 -6.17636800e-01 -6.31982088e-01 1.57837883e-01
-8.04245770e-01 -1.66961932e+00 -8.07485044e-01 -7.52085388e-01
-1.00946271e+00 -1.10765934e+00 -8.65291476e-01 -8.05547059e-01
7.91447878e-01 4.62065727e-01 1.03026426e+00 -1.29203442e-02
-3.01235706e-01 3.12792778e-01 -1.55227289e-01 5.50038576e-01
-2.25940198e-01 -8.20597231e-01 2.54928350e-01 4.91258234e-01
-3.29027295e-01 -5.83883643e-01 -7.04621434e-01 3.31450164e-01
-1.03503978e+00 2.59606421e-01 2.01528773e-01 1.03374624e+00
1.10514307e+00 1.54196069e-01 1.84805654e-02 -5.38396120e-01
1.98111624e-01 -1.99424818e-01 -8.21073830e-01 -1.57145113e-01
-2.57425278e-01 -1.28611233e-02 6.22025669e-01 -4.41222548e-01
-1.39125586e+00 1.78253397e-01 1.75152108e-01 -9.16479886e-01
2.80627578e-01 8.64360332e-02 1.41762272e-01 -2.17451841e-01
5.26861429e-01 1.85446218e-01 -4.50668186e-02 -4.54683602e-01
3.57687056e-01 1.07477240e-01 9.82419252e-01 -5.20556927e-01
9.70532775e-01 1.02372706e+00 2.71618903e-01 -1.21770048e+00
-1.06095254e+00 -4.17921722e-01 -1.04137015e+00 -2.94419229e-01
1.17059076e+00 -1.07248771e+00 -5.10639369e-01 2.61430472e-01
-1.10992932e+00 1.50959685e-01 -4.05617774e-01 9.79481339e-01
-8.26736212e-01 7.99070895e-01 -1.01521933e+00 -5.29204369e-01
5.32445684e-03 -1.25071359e+00 7.97492445e-01 3.05686295e-01
3.18491459e-02 -1.10103905e+00 2.44327977e-01 4.35133308e-01
2.73934931e-01 5.26671588e-01 1.97227910e-01 5.92641413e-01
-9.35533464e-01 2.28423640e-01 -2.61235327e-01 6.05751097e-01
1.99098811e-01 1.10745832e-01 -7.87900090e-01 -4.33594942e-01
4.22900796e-01 5.69537282e-02 5.81457376e-01 7.83900738e-01
8.07476580e-01 -4.50676292e-01 -1.36309722e-02 1.03594446e+00
1.72690082e+00 -2.36341711e-02 8.51917982e-01 1.02907844e-01
9.44489062e-01 4.50304627e-01 7.09500492e-01 3.72607470e-01
2.65888453e-01 8.08470070e-01 1.13042034e-01 -2.15430707e-02
-3.74038845e-01 -4.09005061e-02 3.22987229e-01 9.95605707e-01
-9.06184077e-01 4.91569787e-01 -3.53286833e-01 4.89849538e-01
-2.00087905e+00 -1.56314516e+00 -2.39544213e-01 2.04456782e+00
7.45140076e-01 -2.32423767e-01 2.66818386e-02 1.75257593e-01
8.82410467e-01 4.94815677e-01 -2.84214169e-01 5.37039153e-02
-3.61622512e-01 -9.54760239e-02 2.34920472e-01 1.16698372e+00
-9.35497761e-01 6.82643831e-01 7.61271143e+00 6.01230264e-01
-8.26189995e-01 1.54456601e-01 3.40826958e-01 -1.83672652e-01
-1.00885928e-01 2.32453961e-02 -5.08300483e-01 4.76212531e-01
6.38438046e-01 -3.63312602e-01 6.32044613e-01 3.67471844e-01
6.79008305e-01 -1.94090188e-01 -8.20654571e-01 1.33526397e+00
3.20285290e-01 -1.70012927e+00 1.68086991e-01 -5.85940145e-02
1.12097776e+00 -3.43692899e-01 -2.37670280e-02 -5.07048070e-01
-6.02870509e-02 -7.68314302e-01 7.46985614e-01 1.13864088e+00
9.42454755e-01 -7.11863399e-01 3.21295023e-01 -3.41372676e-02
-1.50369895e+00 7.25512654e-02 -5.99277496e-01 -9.66425017e-02
5.90700924e-01 4.89601851e-01 1.74519688e-01 5.89792788e-01
9.42519188e-01 1.18636119e+00 5.84096350e-02 6.55876219e-01
-1.64285630e-01 1.25784398e-04 2.36602221e-02 8.54197383e-01
3.45035754e-02 -8.08317959e-01 1.00405610e+00 1.03216839e+00
2.64908731e-01 8.00844967e-01 2.00815335e-01 8.11560810e-01
1.10259876e-01 -1.81016102e-01 -4.12442148e-01 4.60435808e-01
2.16707036e-01 1.50614941e+00 -2.44819805e-01 -4.20647115e-01
-6.40082717e-01 1.11209536e+00 8.06237310e-02 7.74551928e-01
-6.33392632e-01 -3.40315513e-02 7.36374438e-01 2.47305900e-01
2.14142680e-01 -4.22138363e-01 1.14599064e-01 -1.64836001e+00
-5.68705015e-02 -6.90064371e-01 3.61729980e-01 -1.17313528e+00
-9.86432672e-01 7.11865067e-01 1.63404997e-02 -1.54973388e+00
-3.70305121e-01 -2.33056575e-01 -2.22865298e-01 1.00811076e+00
-1.64037871e+00 -7.11094916e-01 -4.81898487e-01 9.52646673e-01
6.03155553e-01 6.50535375e-02 5.11290431e-01 3.87719065e-01
-3.44798297e-01 -3.84000055e-02 1.05682455e-01 3.00837040e-01
5.94437301e-01 -9.90698874e-01 -2.62162268e-01 1.20375180e+00
-7.99989924e-02 4.39524233e-01 7.99708962e-01 -4.20144737e-01
-1.23627710e+00 -7.60702610e-01 6.91441000e-01 -3.89747351e-01
4.62597907e-01 1.37984067e-01 -1.26903558e+00 7.45067835e-01
2.60486335e-01 5.31216621e-01 2.61196584e-01 -7.30419159e-01
-1.62208408e-01 -1.67749003e-02 -1.29045498e+00 2.76228130e-01
8.58050704e-01 -6.68018341e-01 -7.80592382e-01 -1.24729075e-01
4.99883890e-01 -6.34384155e-01 -1.17685735e+00 2.60430634e-01
6.14482164e-01 -1.08456993e+00 1.34188104e+00 -1.92738369e-01
6.35802269e-01 -6.80092216e-01 -3.14528108e-01 -1.05822968e+00
-9.49567080e-01 -1.04334962e+00 -3.99057537e-01 8.33325088e-01
-6.32618964e-02 -1.66242942e-01 4.85154450e-01 5.44535816e-01
2.53062606e-01 -4.47385579e-01 -9.52914596e-01 -4.31000292e-01
-3.46634775e-01 6.18969388e-02 -8.28365460e-02 1.14948761e+00
4.02187780e-02 3.62330675e-01 -9.28662002e-01 2.98781991e-01
1.37718689e+00 4.84715635e-03 2.94473320e-01 -1.04744554e+00
6.96333405e-03 -9.42889825e-02 -4.59149718e-01 -1.07321429e+00
1.79653302e-01 -3.70440125e-01 -9.32514668e-02 -1.38535547e+00
3.35503310e-01 1.53979033e-01 -2.73003280e-01 -1.96817622e-01
-1.55064791e-01 5.16174614e-01 5.56545742e-02 4.71687466e-01
-4.51918811e-01 4.25488114e-01 1.68958318e+00 2.62739658e-01
-1.85716122e-01 -4.02760237e-01 -2.48033673e-01 1.10332608e+00
2.52466023e-01 -3.16612571e-01 -2.17292219e-01 -3.94552469e-01
-3.14186653e-03 6.18053138e-01 2.01426387e-01 -9.15558815e-01
2.24687949e-01 -3.96619231e-01 6.57751858e-01 -4.72290516e-01
6.54273927e-01 -7.13733852e-01 3.55578035e-01 1.61463857e-01
-4.11234766e-01 4.94800247e-02 -2.53651708e-01 6.03456140e-01
-5.65217018e-01 -1.35920063e-01 1.39266205e+00 -3.26803923e-01
-7.22674966e-01 5.41137695e-01 -2.67517686e-01 1.16501421e-01
7.06869364e-01 -3.85780215e-01 1.97839644e-02 -4.74631131e-01
-1.05672336e+00 -1.45293176e-01 5.28462350e-01 5.02148867e-01
1.05688930e+00 -1.68796861e+00 -1.00986910e+00 3.13487053e-01
-4.40134764e-01 9.33919027e-02 4.30457085e-01 9.10871148e-01
-6.96918309e-01 1.15221687e-01 -5.70964158e-01 -6.88345611e-01
-1.23597229e+00 5.68573594e-01 4.60850060e-01 -1.07581757e-01
-9.61251676e-01 5.52721918e-01 3.38381559e-01 4.57584620e-01
-6.06816486e-02 2.33518686e-02 -5.44894755e-01 6.65473118e-02
1.23268807e+00 9.22869205e-01 -6.19282246e-01 -1.58796191e+00
-2.52814978e-01 1.22321248e+00 1.75233841e-01 -3.64188015e-01
1.38260829e+00 -6.98829830e-01 -2.81265587e-01 4.28587258e-01
1.45853615e+00 1.11114055e-01 -1.82797778e+00 -3.85192275e-01
-2.30669662e-01 -1.16525674e+00 3.71814311e-01 1.00221016e-01
-1.31682873e+00 4.65352982e-01 4.59768027e-01 5.10731302e-02
1.33455825e+00 -3.89375836e-01 6.99338615e-01 -2.17312291e-01
4.13806289e-01 -1.22872257e+00 2.40165383e-01 3.85797918e-01
1.10786748e+00 -1.20164108e+00 4.53757286e-01 -4.83794630e-01
-3.48830163e-01 1.14385581e+00 2.30192766e-01 -5.73897600e-01
5.72488964e-01 1.72704950e-01 -1.58926807e-02 1.70311749e-01
-6.59423947e-01 -7.05047399e-02 5.22399724e-01 6.03047132e-01
4.36301917e-01 -5.07370114e-01 -4.58261430e-01 6.06937967e-02
2.91497767e-01 1.94076180e-01 9.02757823e-01 5.42021811e-01
-3.73464614e-01 -7.36347198e-01 -6.93043947e-01 -1.39342606e-01
-8.36039305e-01 9.93628502e-02 3.46861184e-01 3.58815163e-01
-1.13288976e-01 9.74448144e-01 9.50702652e-02 -1.10424586e-01
3.50400716e-01 -3.23802203e-01 8.21054578e-01 -1.40868410e-01
1.71083789e-02 6.25312269e-01 -1.69167399e-01 -1.26856470e+00
-1.17533350e+00 -5.64349115e-01 -1.28876996e+00 -3.27305645e-01
-5.53040691e-02 3.02865326e-01 2.54260510e-01 4.49263066e-01
2.11201563e-01 4.18550193e-01 8.43804598e-01 -1.47475505e+00
1.16228916e-01 -6.23908520e-01 -7.89048493e-01 7.72139609e-01
8.79558265e-01 -6.04935348e-01 -7.59519756e-01 7.55524397e-01] | [10.999505996704102, -2.008357048034668] |
d12a270a-4a8b-44ba-9b94-0e2da38729ce | taxonomy-of-aisecops-threat-modeling-for | 2305.11189 | null | https://arxiv.org/abs/2305.11189v1 | https://arxiv.org/pdf/2305.11189v1.pdf | Taxonomy of AISecOps Threat Modeling for Cloud Based Medical Chatbots | Artificial Intelligence (AI) is playing a vital role in all aspects of technology including cyber security. Application of Conversational AI like the chatbots are also becoming very popular in the medical field to provide timely and immediate medical assistance to patients in need. As medical chatbots deal with a lot of sensitive information, the security of these chatbots is crucial. To secure the confidentiality, integrity, and availability of cloud-hosted assets like these, medical chatbots can be monitored using AISecOps (Artificial Intelligence for Secure IT Operations). AISecOPs is an emerging field that integrates three different but interrelated domains like the IT operation, AI, and security as one domain, where the expertise from all these three domains are used cohesively to secure the cyber assets. It considers cloud operations and security in a holistic framework to collect the metrics required to assess the security threats and train the AI models to take immediate actions. This work is focused on applying the STRIDE threat modeling framework to model the possible threats involved in each component of the chatbot to enable the automatic threat detection using the AISecOps techniques. This threat modeling framework is tailored to the medical chatbots that involves sensitive data sharing but could also be applied for chatbots used in other sectors like the financial services, public sector, and government sectors that are concerned with security and compliance. | ['Subash Chandran', 'Sharon Priya S', 'Aisha Banu', 'Ruby Annette J'] | 2023-05-18 | null | null | null | null | ['chatbot', 'chatbot'] | ['methodology', 'natural-language-processing'] | [-3.49827707e-02 3.14237148e-01 2.12405041e-01 2.98210651e-01
-2.39803717e-01 -6.76647604e-01 5.63473463e-01 7.04725325e-01
-3.58050823e-01 3.65800828e-01 5.12938723e-02 -6.44680798e-01
-7.15260386e-01 -7.81356752e-01 2.65935093e-01 -6.60177946e-01
1.12425879e-01 6.55546188e-01 3.04249227e-01 -6.51294827e-01
3.63920093e-01 6.76267684e-01 -9.68091369e-01 6.93679869e-01
5.00329137e-01 1.09451640e+00 -3.69693726e-01 7.84933627e-01
5.60315733e-04 1.32115018e+00 -8.95217955e-01 -6.65857613e-01
2.86505014e-01 1.47780672e-01 -1.20963407e+00 -5.59341729e-01
-6.28742218e-01 -4.58841585e-02 4.70595717e-01 1.19455791e+00
1.70829386e-01 -3.74618143e-01 3.39103192e-01 -1.68651056e+00
3.87401611e-01 4.23057765e-01 -1.55369882e-02 2.66827822e-01
4.06375706e-01 3.21752608e-01 4.73637760e-01 -5.53538166e-02
5.91073871e-01 8.75180960e-01 5.16282737e-01 4.82568681e-01
-3.45773995e-01 -7.66721308e-01 -3.32357168e-01 4.27065223e-01
-8.02400231e-01 4.17894185e-01 6.40904844e-01 -6.63885057e-01
8.41653407e-01 6.04353845e-01 3.76333892e-01 9.87714529e-01
5.03806651e-01 1.41528785e-01 9.99912977e-01 -4.67642367e-01
4.06999916e-01 8.35481942e-01 4.46036071e-01 1.93210855e-01
3.67255598e-01 2.38070693e-02 -2.51573771e-01 -6.07203901e-01
2.34976247e-01 4.17825878e-01 3.27844471e-01 3.53979796e-01
-1.09476519e+00 1.04777038e+00 3.61799538e-01 1.20356131e+00
-8.57063115e-01 -1.55784175e-01 9.18395042e-01 2.88592488e-01
2.27716371e-01 8.92888188e-01 -2.83483475e-01 -5.11718392e-01
-8.41083899e-02 -1.12773143e-01 1.13816762e+00 4.54312533e-01
4.87033688e-02 -3.51415306e-01 1.87129572e-01 -5.13346232e-02
1.40369922e-01 -1.03921942e-01 -1.22403756e-01 -8.45732629e-01
3.66393775e-01 1.23464406e+00 1.82728440e-01 -1.09575093e+00
-3.27756464e-01 1.92816891e-02 -7.74061561e-01 2.76580602e-01
6.88312203e-02 -4.13497269e-01 -3.12124312e-01 1.00401926e+00
7.39051640e-01 1.28447235e-01 3.33217233e-01 4.46890265e-01
6.00460291e-01 5.23528099e-01 5.00431955e-01 -3.53963301e-02
2.04581547e+00 -3.33035767e-01 -8.64742875e-01 1.07712701e-01
6.19919002e-01 -8.45098495e-01 3.99438530e-01 6.89321101e-01
-7.68767357e-01 1.91259921e-01 -9.04371262e-01 3.98719698e-01
-9.96263981e-01 -7.20379531e-01 5.90662777e-01 1.01916254e+00
-7.79384315e-01 -5.86946122e-02 -6.40305042e-01 -4.07706201e-01
8.34962204e-02 7.58918583e-01 -2.65320212e-01 1.01968132e-01
-1.65144408e+00 1.45464492e+00 4.27252084e-01 -8.61500949e-02
-6.21272326e-01 -5.68550289e-01 -5.40342927e-01 1.83977798e-01
2.29536489e-01 -4.44009334e-01 9.46812391e-01 -6.83987856e-01
-1.00750601e+00 5.01580656e-01 9.50367451e-01 -6.57574713e-01
3.22254866e-01 7.09029734e-02 -2.13783875e-01 4.74468887e-01
-2.73628831e-01 -1.78668663e-01 2.00845227e-01 -9.06418741e-01
-7.07388878e-01 -4.80352461e-01 2.56569386e-01 -4.78562042e-02
-6.20062590e-01 1.00470686e+00 4.18571174e-01 -9.70487893e-02
-5.17329156e-01 -7.56697416e-01 -4.00348902e-01 -4.13863927e-01
-1.54690623e-01 2.05019433e-02 1.34981096e+00 -7.18790948e-01
9.79166031e-01 -1.78265810e+00 -2.16377169e-01 5.02455831e-01
2.83255666e-01 1.01088977e+00 4.48734790e-01 1.05329144e+00
2.02289388e-01 2.21182987e-01 8.18583928e-03 2.05435887e-01
-3.16865176e-01 -3.11465729e-02 -2.87632309e-02 -4.72558923e-02
1.72328696e-01 2.97539443e-01 -9.47413564e-01 -5.12366593e-01
4.86692131e-01 5.25877714e-01 -2.11008161e-01 3.63515019e-01
-4.14022505e-02 5.71118057e-01 -8.97120833e-01 6.77606106e-01
4.45484012e-01 3.65606025e-02 4.70501155e-01 2.60844886e-01
-9.75312144e-02 8.10886472e-02 -5.99080503e-01 7.36763239e-01
-5.11716187e-01 1.72072858e-01 4.35431063e-01 -1.03251648e+00
8.53956163e-01 1.19599235e+00 8.89278769e-01 -1.23935297e-01
7.45144904e-01 1.46603540e-01 3.63125205e-01 -7.86904275e-01
1.77665815e-01 -2.35532358e-01 -3.10498655e-01 7.89633095e-01
-1.83028169e-02 -7.34255537e-02 -5.12108147e-01 4.66157436e-01
1.34126759e+00 -5.43855190e-01 6.55884206e-01 2.59492211e-02
9.72170830e-01 3.75173211e-01 3.12577993e-01 1.16273046e-01
-6.18441761e-01 -2.74019539e-01 7.09656835e-01 -7.64697492e-01
-5.63283145e-01 -4.68770355e-01 9.22255218e-02 4.21444714e-01
-1.43528581e-01 -2.92257756e-01 -9.74033415e-01 -1.13948619e+00
-4.66577917e-01 6.21730387e-01 -5.32996833e-01 -3.58802706e-01
-1.98195472e-01 -3.60435486e-01 7.43146181e-01 1.45627946e-01
5.05451739e-01 -1.50799012e+00 -1.33833253e+00 4.12891150e-01
-3.15402210e-01 -1.35283673e+00 1.77336618e-01 1.07185014e-01
-3.42095733e-01 -1.59414506e+00 -1.74054280e-02 -3.11774343e-01
2.83965230e-01 -7.46118277e-03 6.91155016e-01 5.52362978e-01
-7.43747950e-01 6.17118120e-01 -7.17581749e-01 -1.52997696e+00
-1.08985794e+00 -9.20391083e-02 -1.13536762e-02 5.12389652e-02
8.22228253e-01 -3.66734177e-01 -4.74136025e-01 3.91626596e-01
-1.16012585e+00 -4.21163708e-01 1.83828294e-01 6.40372336e-01
-2.77055383e-01 1.72948003e-01 7.08303750e-01 -1.12301874e+00
7.81335652e-01 -9.26934481e-01 -4.37804163e-01 4.25249398e-01
-1.09989896e-01 -5.86787462e-01 5.39069116e-01 -4.06319618e-01
-8.67041886e-01 -2.76034415e-01 -4.40482087e-02 -2.48368338e-01
-4.00946438e-01 5.65203428e-01 -3.92950997e-02 -2.61095345e-01
6.76153481e-01 -3.43571544e-01 3.63862485e-01 -7.68956393e-02
-4.56162989e-01 1.21935630e+00 -1.87495172e-01 -2.71675080e-01
7.12409675e-01 4.95974004e-01 1.58100367e-01 -5.77318847e-01
-4.20331359e-01 -9.13001716e-01 -1.64518282e-01 -6.02831542e-01
1.21325123e+00 1.00454710e-01 -1.75155759e+00 2.59611994e-01
-1.47326159e+00 2.00547606e-01 1.31916344e-01 4.93908703e-01
-8.03761631e-02 1.44175917e-01 -6.69590354e-01 -1.51227129e+00
-1.09315026e+00 -1.10913861e+00 5.11024833e-01 1.53474182e-01
-4.00811940e-01 -9.65528548e-01 1.18016526e-01 9.07746792e-01
6.30922437e-01 5.49163282e-01 1.06261349e+00 -1.41850114e+00
-4.31971580e-01 -5.69288373e-01 2.18282379e-02 5.61255515e-01
4.11906481e-01 -1.29128387e-02 -7.30738282e-01 1.49288997e-01
8.19785833e-01 -2.61334836e-01 -3.64296705e-01 -2.57156044e-01
4.99255270e-01 -7.83663809e-01 -2.31965855e-01 -2.96194583e-01
1.23659515e+00 9.92016375e-01 8.40418041e-01 5.46279967e-01
1.88720033e-01 1.48423958e+00 1.19767249e+00 4.78621155e-01
8.13584477e-02 4.83771414e-01 9.77477908e-01 2.86081657e-02
9.56723094e-01 4.48241562e-01 2.35386401e-01 4.93002146e-01
-3.04095894e-01 3.07456583e-01 -1.53933215e+00 4.32433397e-01
-2.08339167e+00 -1.10287464e+00 6.56932127e-03 2.00046873e+00
4.77312803e-01 1.10476933e-01 1.41300902e-01 5.72402775e-01
5.45015574e-01 -4.70223069e-01 -2.73536146e-02 -1.10063112e+00
8.52797747e-01 4.93032545e-01 2.11389393e-01 2.54690737e-01
-8.96415710e-01 8.07840049e-01 4.75452518e+00 4.12736893e-01
-1.04204452e+00 2.93175817e-01 6.91088438e-01 3.13310325e-01
2.04431489e-01 2.39286982e-02 -3.29229683e-01 3.25578809e-01
1.20235217e+00 -3.21881622e-02 3.33412856e-01 9.67201054e-01
2.87564129e-01 -2.19835013e-01 -4.83970672e-01 4.37630504e-01
-2.29948014e-01 -1.57453358e+00 -3.54310900e-01 1.49980679e-01
1.99799910e-01 -2.96301931e-01 -1.19698018e-01 -2.47429669e-01
3.83585930e-01 -9.21268046e-01 4.08475511e-02 3.70695472e-01
1.81642100e-01 -1.14715779e+00 1.59026420e+00 7.06339300e-01
-7.15577304e-01 -3.74780446e-01 9.64690596e-02 -8.44430253e-02
9.55543518e-02 -4.57018837e-02 -1.52590775e+00 8.68423283e-01
1.07951927e+00 -2.11647198e-01 7.45040700e-02 6.64500296e-01
1.24269962e-01 4.00172740e-01 -9.93173644e-02 -3.60101908e-01
4.75984782e-01 -2.80917794e-01 5.11131883e-01 8.30345154e-01
7.74745345e-02 5.20217597e-01 -1.78167447e-01 4.51661319e-01
5.96886873e-01 1.23343952e-01 -9.38539803e-01 -2.12953568e-01
4.24469441e-01 1.43084228e+00 -8.31738293e-01 -7.62557536e-02
-1.51634634e-01 3.54415238e-01 -3.78802061e-01 -1.64241165e-01
-6.56221390e-01 -6.45040512e-01 9.53257382e-01 1.44784957e-01
-2.67527521e-01 1.21234342e-01 -4.38644290e-01 -3.96882385e-01
-2.52584726e-01 -1.46681237e+00 7.42645919e-01 -3.96915853e-01
-1.18364751e+00 1.11963224e+00 1.10157415e-01 -1.42621303e+00
-2.15993121e-01 -6.90501034e-01 -9.59502816e-01 6.41444087e-01
-9.64997351e-01 -1.66587317e+00 -2.36725271e-01 1.02310145e+00
-8.79145116e-02 -3.28282416e-01 1.22530711e+00 5.66393323e-03
-1.84106454e-01 6.56354725e-02 -6.49027884e-01 2.00157642e-01
3.17406565e-01 -7.08915651e-01 -2.48803094e-01 7.22907126e-01
-1.91545784e-01 9.25826669e-01 5.17721057e-01 -6.95373356e-01
-1.26007438e+00 -6.58524334e-01 9.53507960e-01 -5.99059820e-01
7.58083045e-01 -4.55363542e-01 -6.82957232e-01 5.62959135e-01
5.55198729e-01 -4.74425703e-01 9.90588069e-01 -8.20123851e-02
-1.48711637e-01 -7.90176839e-02 -1.90470660e+00 4.41839546e-01
8.67600217e-02 -5.11509001e-01 -5.53986371e-01 6.26961768e-01
6.73703313e-01 3.61534692e-02 -9.29947317e-01 8.84934291e-02
-1.03397435e-02 -1.01685488e+00 6.86964214e-01 -1.04681909e+00
6.07421175e-02 -8.12337697e-02 2.93926328e-01 -7.22085655e-01
3.55187714e-01 -1.04764891e+00 2.48988152e-01 1.00637090e+00
9.52182561e-02 -1.14994812e+00 9.11277235e-01 1.13168097e+00
2.42417380e-01 -5.74069619e-01 -1.09647107e+00 -3.44854623e-01
-2.15007961e-01 -4.42543894e-01 6.47401929e-01 1.28197479e+00
8.20641458e-01 9.21317339e-02 -2.00784490e-01 3.85159969e-01
3.59404355e-01 -6.36799097e-01 5.79811871e-01 -1.44829953e+00
1.50599582e-02 1.01965539e-01 -9.58027124e-01 7.36182749e-01
-5.74055374e-01 -3.13108921e-01 -2.81722307e-01 -1.43337131e+00
-4.58778813e-02 -5.77773511e-01 -3.78207743e-01 6.15972757e-01
4.10088971e-02 -2.16035768e-01 4.63281035e-01 -5.48696145e-02
-1.02272645e-01 -2.43777305e-01 7.13856339e-01 2.70886160e-03
-9.63754952e-02 4.09675449e-01 -8.21079969e-01 6.92697406e-01
9.61446643e-01 -7.93958068e-01 -3.95151526e-01 4.07304972e-01
2.92375952e-01 4.28612649e-01 2.91535288e-01 -9.78962064e-01
7.20534682e-01 -5.76442063e-01 -6.03757799e-01 -1.95005447e-01
3.82650763e-01 -1.74754405e+00 3.52886200e-01 1.17887390e+00
5.30269779e-02 1.68128997e-01 6.60714805e-02 6.78968132e-02
-5.38678229e-01 -3.01020384e-01 7.68839121e-01 -2.71522611e-01
5.04196435e-02 8.48161876e-02 -7.83363283e-01 -4.85454381e-01
1.85734844e+00 -1.70220390e-01 -5.85022271e-01 -7.14451551e-01
-5.15128672e-01 2.75401354e-01 -2.48757042e-02 1.38359651e-01
5.58895290e-01 -5.86223483e-01 -4.28988874e-01 -1.93904638e-01
1.91676263e-02 -1.93210378e-01 4.47138995e-01 9.31847036e-01
-8.63288999e-01 5.13839066e-01 -6.27100945e-01 -1.00012846e-01
-1.88227677e+00 9.65882778e-01 2.84075052e-01 -9.58803296e-01
-1.47931144e-01 4.03559029e-01 -1.66022152e-01 -1.56259656e-01
4.33419466e-01 2.16981143e-01 -7.49885023e-01 6.45555034e-02
1.00602841e+00 4.67271388e-01 1.74046099e-01 -3.84322613e-01
-6.35548413e-01 -2.98651326e-02 -1.88477278e-01 -3.12566042e-01
1.43718612e+00 5.09036422e-01 -5.05622566e-01 8.09377246e-03
6.94323599e-01 -1.53562099e-01 -1.75547794e-01 1.82095259e-01
6.57283723e-01 -2.01833129e-01 -2.56517202e-01 -1.10976160e+00
-8.68939102e-01 1.03691483e+00 3.00664425e-01 7.31570840e-01
1.14646912e+00 -4.05254573e-01 6.46573246e-01 2.38621518e-01
7.52778411e-01 -1.01272345e+00 2.37800121e-01 4.07499880e-01
8.35710466e-01 -8.64465773e-01 -3.11729193e-01 -6.73718572e-01
-1.36316001e+00 1.03912795e+00 6.09556675e-01 1.83863223e-01
8.89045238e-01 6.52933180e-01 5.23180604e-01 -6.30879343e-01
-7.03542650e-01 1.06092364e-01 -1.26326948e-01 9.09433126e-01
9.38060284e-02 -7.81708509e-02 -3.56918305e-01 7.83068597e-01
2.96421170e-01 1.09991193e-01 6.29886210e-01 1.15522349e+00
-8.13222528e-02 -1.62400186e+00 -7.73986578e-01 1.57215044e-01
-9.84793484e-01 2.44942103e-02 -1.06105578e+00 7.78277397e-01
2.73648232e-01 1.24435234e+00 -3.26867849e-01 -8.10321033e-01
2.17356846e-01 9.99157280e-02 -4.74835485e-01 -6.14553511e-01
-2.05972505e+00 -5.00633419e-01 4.98682976e-01 -3.70967746e-01
-3.22388589e-01 -3.40802103e-01 -1.09809673e+00 -5.39101899e-01
-1.57252386e-01 7.16288865e-01 1.20874357e+00 7.09489286e-01
2.82872766e-01 3.40581715e-01 8.25990021e-01 -1.62604749e-01
-3.21268111e-01 -9.03415143e-01 -2.05182314e-01 2.11932495e-01
-2.28791535e-02 -3.11309934e-01 -2.07278833e-01 -6.29650176e-01] | [5.402652740478516, 7.151310443878174] |
15a2c47b-2198-44f8-9072-2fb2456bb5bc | optimization-of-rule-based-energy-management | 2207.06450 | null | https://arxiv.org/abs/2207.06450v1 | https://arxiv.org/pdf/2207.06450v1.pdf | Optimization of rule-based energy management strategies for hybrid vehicles using dynamic programming | Reducing energy consumption is a key focus for hybrid electric vehicle (HEV) development. The popular vehicle dynamic model used in many energy management optimization studies does not capture the vehicle dynamics that the in-vehicle measurement system does. However, feedback from the measurement system is what the vehicle controller actually uses to manage energy consumption. Therefore, the optimization solely using the model does not represent what the vehicle controller sees in the vehicle. This paper reports the utility factor-weighted energy consumption using a rule-based strategy under a real-world representative drive cycle. In addition, the vehicle test data was used to perform the optimization approach. By comparing results from both rule-based and optimization-based strategies, the areas for further improving rule-based strategy are discussed. Furthermore, recent development of OBD raises a concern about the increase of energy consumption. This paper investigates the energy consumption increase with extensive OBD usage. | ['Yang Xu', 'Vivek Kumar', 'Sumanth Reddy Dadam', 'Ewan Pritchard', 'Di Zhu'] | 2022-07-08 | null | null | null | null | ['energy-management'] | ['time-series'] | [-4.80445653e-01 1.84646294e-01 -5.82635641e-01 -1.07839577e-01
-3.51377428e-02 -3.01135063e-01 5.08059025e-01 1.81643158e-01
-1.77253723e-01 6.64111912e-01 -3.16039532e-01 -7.40142465e-01
-2.60073841e-01 -1.04686642e+00 -4.59347546e-01 -8.23880196e-01
3.69454354e-01 1.34776086e-01 1.31561771e-01 -1.82294428e-01
6.82135820e-02 7.42483556e-01 -1.65089512e+00 -5.91362774e-01
6.29151642e-01 7.45359719e-01 4.56753731e-01 4.21438277e-01
8.69408473e-02 4.15100485e-01 -4.57526028e-01 3.40314448e-01
1.28171295e-01 -6.78218901e-01 -3.13454509e-01 5.98008260e-02
-6.31395459e-01 -5.17407060e-01 -2.10756674e-01 9.41234291e-01
1.97240844e-01 3.16769749e-01 6.80504203e-01 -1.91984689e+00
3.48478526e-01 9.24339518e-02 -1.96043923e-02 2.42373392e-01
-3.28826696e-01 3.87618393e-01 3.23371410e-01 -3.45266223e-01
4.44218248e-01 8.98055971e-01 -2.46802215e-02 5.15236676e-01
-9.81939852e-01 -6.41417682e-01 -1.44257873e-01 5.98730326e-01
-1.46003175e+00 -3.23979437e-01 9.69443202e-01 -3.11692357e-01
1.58135688e+00 4.83178020e-01 1.16203809e+00 1.09845914e-01
7.08919287e-01 2.90885210e-01 1.07167268e+00 -3.75518709e-01
5.28182924e-01 6.26826227e-01 3.50314945e-01 1.66472867e-01
8.08846176e-01 4.93926764e-01 3.60383630e-01 1.68494776e-01
-2.20118955e-01 -3.64187658e-01 1.44242586e-04 -4.24416631e-01
-1.41416728e-01 7.25423038e-01 -3.11316341e-01 4.41199362e-01
-5.05681872e-01 3.48305613e-01 3.80455256e-01 2.19432518e-01
5.51479943e-02 3.94896954e-01 -4.38875705e-01 -5.10287106e-01
-9.52671528e-01 2.65255719e-01 7.89887488e-01 8.89743388e-01
8.91638637e-01 6.90666914e-01 1.37981415e-01 2.81578302e-01
7.98112035e-01 7.54156947e-01 1.17425313e-02 -1.21237624e+00
-3.20552647e-01 8.80110741e-01 2.31762096e-01 -7.04524159e-01
-5.40370405e-01 -7.90469646e-02 -2.05159813e-01 6.36237383e-01
-5.20410910e-02 -3.52402210e-01 -7.77197480e-01 1.40313447e+00
3.78254294e-01 -3.17620039e-01 2.54407674e-01 6.18696511e-01
3.18382472e-01 9.75822687e-01 1.98434740e-01 -9.05935466e-01
1.20906079e+00 -5.51100850e-01 -1.51151073e+00 -2.38916334e-02
8.40233386e-01 -3.64239275e-01 3.13666135e-01 1.49949566e-01
-1.12080228e+00 -1.71085104e-01 -1.64117360e+00 3.45237851e-01
-7.39311337e-01 -2.26057902e-01 5.21781780e-02 1.01363313e+00
-9.58247662e-01 3.19516629e-01 -7.70079732e-01 -4.57859695e-01
-1.33616239e-01 3.33621949e-01 1.65610060e-01 2.91985929e-01
-1.20888543e+00 1.73543656e+00 3.55230212e-01 -1.25228390e-01
-9.22172785e-01 -8.12240958e-01 -7.63831198e-01 -5.50692575e-03
3.65541875e-01 -1.64596103e-02 1.46373200e+00 -3.50252002e-01
-1.60648370e+00 2.26321653e-01 -2.10618168e-01 -5.07317245e-01
3.68580222e-01 2.93419063e-01 -8.22428048e-01 -7.04822764e-02
-3.20436329e-01 7.91118592e-02 6.69755042e-02 -1.25962555e+00
-9.09266114e-01 1.15193121e-01 -1.31367147e-01 -6.62032217e-02
1.31340474e-01 -1.95587650e-01 -1.38990670e-01 2.63502121e-01
-6.58696234e-01 -1.11730039e+00 -1.13098301e-01 -5.45379400e-01
-1.65163632e-02 -7.52869308e-01 1.51881158e+00 -4.34515268e-01
1.56670260e+00 -1.92388022e+00 -3.01886320e-01 5.86346447e-01
-3.32581162e-01 2.27627322e-01 2.65116036e-01 5.93637526e-01
1.03798874e-01 3.11182171e-01 1.93584606e-01 1.90018803e-01
3.36374074e-01 4.13474947e-01 1.49454951e-01 4.62815255e-01
1.19420595e-01 5.04812479e-01 -5.80510974e-01 -2.69554764e-01
9.08818185e-01 3.82677823e-01 -2.59296030e-01 2.74177760e-01
-1.93292424e-01 -2.98190325e-01 -5.02595961e-01 2.97968775e-01
6.13046467e-01 5.71224391e-01 4.47481096e-01 -6.01796925e-01
-6.57020926e-01 6.27523381e-03 -1.11928511e+00 7.55688071e-01
-5.71284294e-01 9.77156341e-01 2.80618101e-01 -1.11578190e+00
7.67589748e-01 2.98320919e-01 9.23629403e-01 -1.23954141e+00
4.91559446e-01 3.96189660e-01 2.50378937e-01 -5.52367687e-01
6.98416173e-01 3.25707681e-02 5.58991469e-02 2.61393815e-01
-1.06429435e-01 -3.47364753e-01 5.23207843e-01 -1.07118890e-01
7.97240019e-01 5.03294775e-03 2.88699865e-01 -8.56468141e-01
5.08883476e-01 5.46227038e-01 7.79539347e-01 -1.30766451e-01
-3.65869433e-01 -6.73101366e-01 4.63259101e-01 2.04079479e-01
-1.38062823e+00 -3.94052029e-01 -3.94266337e-01 2.84532696e-01
7.47092426e-01 -4.26252335e-01 -8.93752277e-01 -1.70820460e-01
3.01419757e-03 1.73261929e+00 -1.75970554e-01 -7.53408611e-01
-5.75437248e-01 -6.11134589e-01 -1.00822754e-01 2.35854253e-01
3.27732325e-01 -2.61803150e-01 -1.37412488e+00 5.20236552e-01
2.46813402e-01 -8.41242492e-01 -2.63576448e-01 1.79243371e-01
-5.55011749e-01 -1.11611032e+00 1.39214352e-01 -1.77502707e-01
5.14512539e-01 3.21551591e-01 8.65931571e-01 3.19700360e-01
1.22252507e-02 6.17884934e-01 4.26721387e-02 -1.11584878e+00
-7.89139390e-01 -2.11213976e-01 1.31056532e-01 -3.90919566e-01
9.79797721e-01 1.80606797e-01 -5.03326833e-01 6.13691688e-01
-3.89305919e-01 -1.69971376e-04 1.81989297e-01 1.06956743e-01
7.22107887e-01 9.15736079e-01 8.66419852e-01 -1.93335131e-01
6.40148461e-01 -6.91093981e-01 -9.36834872e-01 6.24179728e-02
-1.55192864e+00 2.22992033e-01 3.02200884e-01 -1.20041661e-01
-1.08634794e+00 -7.83622712e-02 4.98229489e-02 -4.06013876e-01
-5.13583347e-02 1.09870911e-01 -7.41961718e-01 2.05236092e-01
-2.84597814e-01 9.93623137e-02 4.70968664e-01 -3.32283348e-01
-3.08116764e-01 6.54840291e-01 9.46901143e-02 -7.42316842e-02
8.96606266e-01 -6.03730939e-02 4.88231987e-01 -8.97365987e-01
4.39912319e-01 -8.14534202e-02 9.28511620e-02 -9.62850392e-01
1.03772902e+00 -7.19014704e-01 -1.00099313e+00 2.30433643e-01
-8.18514585e-01 -5.28665304e-01 -3.09125185e-01 7.24454582e-01
-4.90539312e-01 -3.43348563e-01 2.99145132e-01 -1.28913522e+00
-7.93980882e-02 -1.52948976e+00 3.60516757e-01 5.97964525e-01
-2.28579268e-01 -8.65253448e-01 1.20105386e-01 -7.38292113e-02
4.52186406e-01 4.00639474e-01 9.77035522e-01 -2.39161327e-01
-7.58933544e-01 -2.21202284e-01 3.28319997e-01 4.34170038e-01
2.32455265e-02 3.30368787e-01 -5.43403327e-01 -5.07747293e-01
7.81341344e-02 5.32662749e-01 3.63681763e-02 3.37388217e-01
6.84103429e-01 -6.06464520e-02 -6.49352670e-01 5.44782914e-02
2.17125559e+00 1.14582789e+00 6.40423238e-01 5.06287515e-01
6.28974065e-02 7.22978771e-01 1.18810141e+00 2.31076032e-01
4.87265319e-01 8.91934752e-01 6.79698050e-01 1.71661586e-01
-5.81627525e-03 -1.26249388e-01 5.09956002e-01 5.25660574e-01
1.54669378e-02 -4.16116327e-01 -6.96851552e-01 5.81574261e-01
-1.59873617e+00 -8.21435094e-01 -4.64032322e-01 1.94975269e+00
3.28323036e-01 1.94658145e-01 1.44459516e-01 4.24780250e-01
5.04708588e-01 -3.43123078e-01 -5.98077118e-01 -1.24467051e+00
3.13323200e-01 -3.65548939e-01 1.22695386e+00 4.48058188e-01
-2.22407028e-01 -1.58536714e-02 6.30609846e+00 8.79683375e-01
-1.04189014e+00 1.32646039e-01 1.14435799e-01 -2.36948535e-01
-4.49433208e-01 3.70826107e-03 -8.54709625e-01 7.86842465e-01
1.90663171e+00 -1.25323451e+00 3.20438951e-01 7.87470698e-01
1.14305937e+00 -7.51867294e-01 -1.06894982e+00 5.91942489e-01
-3.51164728e-01 -1.14135122e+00 -4.40065831e-01 6.27711236e-01
4.18595403e-01 -2.07070917e-01 -6.20152533e-01 3.59729230e-01
1.80870260e-03 -7.67502904e-01 9.61658299e-01 6.60578191e-01
4.78099555e-01 -1.33076787e+00 9.21151698e-01 4.24589366e-01
-1.32474196e+00 -9.55448896e-02 -9.35996696e-02 8.30052271e-02
6.00240231e-01 2.61279404e-01 -6.60306454e-01 4.61924136e-01
6.08479142e-01 4.82129399e-03 -3.27762544e-01 6.99328244e-01
2.53927439e-01 7.64808953e-01 -5.14970362e-01 -5.58797538e-01
1.18680082e-01 -6.60673320e-01 6.78196609e-01 1.00732553e+00
4.67914462e-01 1.91811457e-01 -3.23258549e-01 7.35323906e-01
4.64678854e-01 -1.02357939e-01 -7.98790693e-01 -2.68812120e-01
6.84700072e-01 1.23284757e+00 -4.60329622e-01 -2.61489987e-01
-5.08197069e-01 4.23077727e-04 -7.40525246e-01 3.53000671e-01
-1.18658566e+00 -5.13261735e-01 1.19972610e+00 4.82185364e-01
-1.37568228e-02 -2.01924190e-01 -2.22656876e-01 -1.39498308e-01
-2.49961808e-01 -8.07987228e-02 -3.10712665e-01 -6.98607147e-01
-2.99931794e-01 -5.38821556e-02 8.34429085e-01 -1.02519596e+00
-4.62013632e-01 -6.44538641e-01 -8.09414089e-01 7.77283490e-01
-1.51523030e+00 -3.90610665e-01 -2.39963889e-01 -1.26413908e-02
6.49804354e-01 -1.55894428e-01 3.25354695e-01 6.58918798e-01
-9.82092321e-01 2.89150864e-01 4.81169313e-01 -5.99578321e-01
5.29746078e-02 -9.88202333e-01 -5.23341835e-01 8.88502002e-01
-1.14530969e+00 -1.10316321e-01 1.39663851e+00 -5.64302623e-01
-2.20080853e+00 -1.09964168e+00 6.14594579e-01 -8.89533013e-02
3.49136591e-01 2.58103888e-02 -7.86399901e-01 1.81304112e-01
9.72425699e-01 -6.35235667e-01 2.34574020e-01 -8.56357217e-01
9.79381919e-01 -4.55167025e-01 -1.52934003e+00 5.03732502e-01
4.40600485e-01 -2.10766762e-01 -2.67590761e-01 -3.64891768e-01
4.51358169e-01 4.39209007e-02 -9.71402645e-01 4.82350737e-01
5.24724901e-01 -3.23138833e-01 2.70929277e-01 -1.88684121e-01
-3.17194343e-01 -5.68490267e-01 -1.13805413e-01 -1.50609577e+00
-4.15911749e-02 -4.25813109e-01 -4.31143373e-01 1.46800101e+00
3.90860200e-01 -7.11056292e-01 4.76066381e-01 1.01210785e+00
-3.05797219e-01 -7.45330989e-01 -1.12416136e+00 -1.03045404e+00
1.45184517e-01 -6.52004957e-01 7.66869307e-01 3.91175240e-01
8.15560594e-02 2.63267532e-02 1.23252913e-01 1.04279980e-01
8.33728194e-01 -5.19632220e-01 6.08312547e-01 -7.91856647e-01
5.10044217e-01 -5.04439235e-01 -4.50572938e-01 -2.01636013e-02
1.72442645e-01 -6.24926388e-01 1.57834157e-01 -1.89743161e+00
-2.85744220e-02 1.22473679e-01 -1.91955939e-01 3.96094099e-02
3.77338439e-01 -2.22414553e-01 3.20194364e-01 -2.98913658e-01
1.11215472e-01 6.21933341e-01 8.28460038e-01 -3.61057907e-01
-1.39550671e-01 -2.53246605e-01 -2.23390609e-01 5.42617798e-01
1.20783710e+00 -6.12941802e-01 -8.83269191e-01 1.93129927e-01
6.65652007e-02 -1.98791653e-01 2.48445779e-01 -9.26869690e-01
1.53654233e-01 -7.04791725e-01 -3.65387499e-01 -9.89461660e-01
4.48640734e-02 -1.77674615e+00 1.16710663e+00 8.95411015e-01
3.84991556e-01 1.26746759e-01 4.18517381e-01 4.03163522e-01
2.18153317e-02 -2.52201527e-01 1.04761732e+00 5.63341200e-01
-1.06326056e+00 -1.50565729e-01 -1.18869174e+00 -5.82830667e-01
1.95849359e+00 -7.31706381e-01 -3.14001143e-01 8.89049247e-02
-2.35859945e-01 8.13987434e-01 5.45505941e-01 5.96274495e-01
1.40277669e-01 -1.50639760e+00 -3.16908330e-01 -2.81097405e-02
-9.80058759e-02 -5.81831932e-01 2.90913343e-01 9.23119485e-01
-4.12222683e-01 7.34190464e-01 -2.47298777e-01 -4.12595332e-01
-1.30933213e+00 5.69063723e-01 9.03838277e-01 4.78706583e-02
-2.90896714e-01 -3.95115793e-01 -2.28300691e-01 1.77178025e-01
-2.88617432e-01 -1.22803472e-01 -1.99813426e-01 2.55916357e-01
2.09484622e-01 1.18996572e+00 1.30534753e-01 -6.52165055e-01
-5.62498391e-01 3.63345146e-01 6.47526741e-01 -1.11859508e-01
1.00240469e+00 -5.08716583e-01 1.55359969e-01 6.74117088e-01
1.41151357e+00 -4.88188952e-01 -1.16024303e+00 5.16670942e-01
3.95580158e-02 -4.99591865e-02 7.97208190e-01 -6.82828963e-01
-1.11430621e+00 3.18772763e-01 1.07248127e+00 4.96585816e-01
1.30701602e+00 -3.25624824e-01 3.63196820e-01 2.89519370e-01
4.68406200e-01 -2.10288572e+00 -6.97251976e-01 1.26192942e-01
7.76806712e-01 -7.18378723e-01 8.35024839e-05 -2.89486516e-02
-2.26686314e-01 1.03052533e+00 9.04110551e-01 7.11795092e-02
1.21173739e+00 7.40650415e-01 -3.67421865e-01 -1.13710247e-01
-9.85006273e-01 -2.82600094e-02 -1.82680935e-01 2.76746571e-01
-1.09706827e-01 2.06312194e-01 -9.97989118e-01 5.27268648e-01
2.56762385e-01 2.25209951e-01 9.12453473e-01 1.07364237e+00
-6.82567716e-01 -9.90751505e-01 -3.38583797e-01 5.23927152e-01
-1.85905695e-01 8.13853741e-01 1.45965591e-01 1.32918203e+00
2.93303341e-01 1.35568571e+00 3.70783567e-01 -5.27605772e-01
9.32281852e-01 2.35168830e-01 1.71246510e-02 -4.82426323e-02
-1.01375684e-01 -2.14986280e-01 5.52900732e-01 -6.16468370e-01
-5.75277984e-01 -6.85193181e-01 -1.79647493e+00 -7.40862012e-01
-6.04234457e-01 6.64729655e-01 1.40402412e+00 9.67464745e-01
3.44083339e-01 9.97760892e-01 1.03002679e+00 -4.83453482e-01
-2.01010332e-01 -7.94264913e-01 -7.99989998e-01 -2.14169323e-01
2.40679234e-01 -9.10427690e-01 -6.51376486e-01 -2.89412677e-01] | [5.5767903327941895, 2.160926580429077] |
39b2bebd-c375-4366-bbb5-8c9aaff4f06d | a-trigger-sense-memory-flow-framework-for | 2101.10213 | null | https://arxiv.org/abs/2101.10213v3 | https://arxiv.org/pdf/2101.10213v3.pdf | A Trigger-Sense Memory Flow Framework for Joint Entity and Relation Extraction | Joint entity and relation extraction framework constructs a unified model to perform entity recognition and relation extraction simultaneously, which can exploit the dependency between the two tasks to mitigate the error propagation problem suffered by the pipeline model. Current efforts on joint entity and relation extraction focus on enhancing the interaction between entity recognition and relation extraction through parameter sharing, joint decoding, or other ad-hoc tricks (e.g., modeled as a semi-Markov decision process, cast as a multi-round reading comprehension task). However, there are still two issues on the table. First, the interaction utilized by most methods is still weak and uni-directional, which is unable to model the mutual dependency between the two tasks. Second, relation triggers are ignored by most methods, which can help explain why humans would extract a relation in the sentence. They're essential for relation extraction but overlooked. To this end, we present a Trigger-Sense Memory Flow Framework (TriMF) for joint entity and relation extraction. We build a memory module to remember category representations learned in entity recognition and relation extraction tasks. And based on it, we design a multi-level memory flow attention mechanism to enhance the bi-directional interaction between entity recognition and relation extraction. Moreover, without any human annotations, our model can enhance relation trigger information in a sentence through a trigger sensor module, which improves the model performance and makes model predictions with better interpretation. Experiment results show that our proposed framework achieves state-of-the-art results by improves the relation F1 to 52.44% (+3.2%) on SciERC, 66.49% (+4.9%) on ACE05, 72.35% (+0.6%) on CoNLL04 and 80.66% (+2.3%) on ADE. | ['Weiming Lu', 'Yechun Tang', 'Xinyin Ma', 'Yongliang Shen'] | 2021-01-25 | null | null | null | null | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [ 2.12849766e-01 6.36811078e-01 -2.02409044e-01 -4.59956676e-01
-5.46585858e-01 -3.62456232e-01 5.23097694e-01 3.16060454e-01
-6.62909150e-01 6.78187728e-01 2.34749556e-01 -5.12901127e-01
5.76259494e-02 -1.05038130e+00 -8.72939467e-01 -1.60451338e-01
-8.09645001e-03 2.93881118e-01 3.42889220e-01 -2.65666276e-01
-4.75078402e-03 5.56711406e-02 -1.30984581e+00 5.86004555e-01
1.14691663e+00 7.92230725e-01 2.77107865e-01 5.86439192e-01
-3.00273389e-01 8.80182922e-01 -7.18174398e-01 -6.80035114e-01
-3.32322031e-01 -1.88121393e-01 -1.24767303e+00 -3.97878587e-01
-2.58052796e-01 -3.87162924e-01 -4.72353905e-01 8.81644666e-01
3.43809515e-01 3.37266363e-02 5.68802238e-01 -1.01081336e+00
-7.37274170e-01 1.24598873e+00 -6.75232768e-01 1.46438748e-01
4.38038766e-01 -1.20531842e-01 9.38579023e-01 -9.52194870e-01
4.58105236e-01 1.15011585e+00 4.17506903e-01 4.05917615e-01
-8.90684009e-01 -7.68806338e-01 4.71123964e-01 4.96837914e-01
-1.44788134e+00 -5.39878905e-01 3.74526232e-01 -2.68553883e-01
1.70572317e+00 4.97662306e-01 3.36229742e-01 1.07908666e+00
2.03747317e-01 1.02503717e+00 8.33243430e-01 -4.03817385e-01
-2.10002944e-01 1.09937154e-01 7.74143159e-01 5.17263055e-01
2.28701770e-01 -7.14676008e-02 -7.23554432e-01 2.63303310e-01
5.03813922e-01 -2.20024586e-01 -4.03564841e-01 4.90320563e-01
-1.03910768e+00 2.90286452e-01 4.62859094e-01 3.71976256e-01
-3.98102134e-01 -2.51699746e-01 2.15428039e-01 7.86041394e-02
2.76251525e-01 3.75426739e-01 -9.03424144e-01 -1.93263650e-01
-6.65991783e-01 3.64917330e-02 9.49365854e-01 1.25046563e+00
6.52719855e-01 -3.73414993e-01 -4.75359321e-01 6.20634735e-01
5.22584260e-01 4.65230107e-01 3.58602852e-01 -2.05081120e-01
7.14489937e-01 1.04751575e+00 3.83962505e-02 -1.12912738e+00
-7.06631362e-01 -7.60737777e-01 -1.03987086e+00 -5.01567543e-01
1.38072655e-01 -1.57571256e-01 -8.30648959e-01 1.91454029e+00
2.02751100e-01 1.57995582e-01 2.32124940e-01 6.70058012e-01
1.33753645e+00 5.71226060e-01 4.89123940e-01 -4.84730691e-01
1.85990226e+00 -1.22007072e+00 -1.16053963e+00 -5.79203963e-01
8.76644790e-01 -6.53358281e-01 7.04631686e-01 7.24067241e-02
-1.17197943e+00 -7.23527670e-01 -1.17943883e+00 -3.92216533e-01
-3.60942543e-01 2.75234461e-01 7.62225389e-01 2.49354169e-01
-4.21647131e-01 3.94082159e-01 -1.01136220e+00 -1.19484738e-01
1.89069897e-01 5.11807203e-01 -3.52469862e-01 3.57941017e-02
-1.84975529e+00 1.26144791e+00 8.13536286e-01 4.33442503e-01
-2.88448483e-01 -5.96581221e-01 -9.18374836e-01 2.68033922e-01
5.98107815e-01 -6.84630036e-01 1.10337341e+00 -3.72477144e-01
-1.28368258e+00 7.00326622e-01 -5.67734003e-01 -5.11199176e-01
1.79449961e-01 -8.24981749e-01 -6.58929050e-01 -2.52538532e-01
-4.79230453e-04 4.52345371e-01 7.55696818e-02 -8.65993500e-01
-6.57457709e-01 -3.26192081e-01 8.62077996e-02 3.32053334e-01
-1.65920094e-01 2.13114977e-01 -7.62492716e-01 -4.41147864e-01
2.00434268e-01 -5.68143010e-01 -6.62060827e-03 -7.30098426e-01
-6.94099367e-01 -4.98533040e-01 5.18390894e-01 -1.12745273e+00
1.90468371e+00 -1.93650448e+00 2.36251161e-01 -1.20878942e-01
3.73615295e-01 5.61358213e-01 1.54024670e-02 3.68862510e-01
-1.12491667e-01 4.51302618e-01 -1.38775989e-01 -3.98968726e-01
-1.12911984e-01 9.28311050e-02 -1.49995089e-01 -1.62670642e-01
7.00764120e-01 1.24781954e+00 -8.51594210e-01 -4.79074985e-01
-1.77800313e-01 4.78675812e-01 -2.43509591e-01 4.90893453e-01
-7.25945225e-03 2.79450387e-01 -4.25037116e-01 4.91865307e-01
7.58926153e-01 -3.49060506e-01 5.15126050e-01 -4.60441679e-01
3.46381366e-02 8.23124290e-01 -1.37425756e+00 1.57134140e+00
-3.74390155e-01 2.58784264e-01 -2.37155229e-01 -7.26669967e-01
9.18301582e-01 3.96203130e-01 -5.55729493e-02 -7.20417142e-01
1.27342433e-01 1.39207169e-01 2.82784522e-01 -6.71437025e-01
5.64320862e-01 1.65689513e-01 -8.01863223e-02 2.50518769e-01
2.22537667e-01 6.17477477e-01 1.86213091e-01 3.39454532e-01
1.31287372e+00 3.18651348e-01 5.82560837e-01 4.26583625e-02
6.17243469e-01 -2.66617119e-01 8.50831747e-01 6.50634229e-01
1.19282350e-01 1.84183300e-01 5.05914390e-01 -1.10837601e-01
-3.69641572e-01 -7.85623491e-01 -1.12351425e-01 1.02776265e+00
2.90491313e-01 -8.03746283e-01 -6.04540646e-01 -7.34944999e-01
-3.66105705e-01 8.47212970e-01 -5.40776134e-01 -5.97411454e-01
-7.33998954e-01 -9.91002321e-01 7.11484313e-01 9.27339673e-01
8.56743991e-01 -1.00693536e+00 -3.71963829e-01 4.17822659e-01
-6.49787843e-01 -1.35285413e+00 -5.84560037e-02 4.19977814e-01
-6.26521051e-01 -9.13648069e-01 -7.21236542e-02 -6.52775586e-01
4.23439831e-01 7.85629265e-04 1.34538364e+00 2.53536582e-01
2.88465083e-01 -3.07597965e-01 -4.93153542e-01 -3.40648770e-01
-1.08529337e-01 4.85126317e-01 -4.75526750e-02 -1.58707932e-01
8.49506557e-01 -4.29642946e-01 -3.43560964e-01 3.64451021e-01
-5.70980251e-01 5.68685651e-01 9.90217626e-01 8.44877303e-01
4.06646639e-01 -9.76592749e-02 7.07688808e-01 -9.94406104e-01
4.43355113e-01 -6.26577437e-01 -1.48031414e-01 6.16961479e-01
-7.52115369e-01 1.25119731e-01 2.58221239e-01 -3.83860677e-01
-1.31849813e+00 -1.29745781e-01 -1.60038546e-01 1.50695860e-01
-2.04603761e-01 7.79897213e-01 -8.39660645e-01 6.20051682e-01
3.66639107e-01 9.50551480e-02 -5.52803993e-01 -5.33121884e-01
3.99923265e-01 8.76624584e-01 6.38602197e-01 -4.97410506e-01
5.59099019e-01 -2.23786131e-01 -3.27189207e-01 -2.96911329e-01
-1.12140608e+00 -3.03589284e-01 -7.43978560e-01 2.66009152e-01
1.10958838e+00 -1.13744998e+00 -9.41631436e-01 5.14453828e-01
-1.51997387e+00 -9.32700038e-02 1.62757754e-01 4.85903203e-01
1.20963484e-01 1.85438663e-01 -8.75516891e-01 -9.31350052e-01
-5.73560357e-01 -1.03031933e+00 9.43017423e-01 6.04219615e-01
-4.91215199e-01 -6.95466280e-01 -4.60021049e-01 5.55655837e-01
2.47510731e-01 -1.03806905e-01 9.60019827e-01 -9.84721363e-01
-6.98706567e-01 -1.98553633e-02 -5.82166016e-01 6.18640929e-02
-1.32527472e-02 -1.85758144e-01 -1.12451530e+00 1.46597832e-01
-1.11095436e-01 -1.33364853e-02 8.23576987e-01 -2.04232395e-01
8.90350163e-01 -2.86283553e-01 -7.43153453e-01 4.00694370e-01
8.62597048e-01 3.10485452e-01 8.78360927e-01 2.13406876e-01
8.58065903e-01 6.42823637e-01 6.65087938e-01 1.13522433e-01
9.28493500e-01 7.05346704e-01 1.28597572e-01 4.88114320e-02
-1.70750797e-01 -4.26444978e-01 3.87220949e-01 1.09163833e+00
-1.69945538e-01 -4.30947989e-01 -1.07734334e+00 3.97800058e-01
-2.08996654e+00 -5.90554178e-01 -6.35578156e-01 1.90831423e+00
1.34774506e+00 3.68588895e-01 -4.03962404e-01 2.24399090e-01
6.32049382e-01 -1.61650896e-01 -2.87805915e-01 -2.61642516e-01
-1.80623502e-01 3.66645008e-01 1.76917419e-01 5.39334059e-01
-1.15925670e+00 1.17688811e+00 4.91546202e+00 8.13799083e-01
-8.10664117e-01 -1.94682535e-02 4.87547576e-01 2.80818909e-01
-1.23565674e-01 1.47868678e-01 -1.23939848e+00 2.70941854e-01
1.10780966e+00 -2.79794317e-02 -3.16769294e-02 4.36852634e-01
-2.21526593e-01 -1.82013586e-01 -1.37558091e+00 9.10364032e-01
-2.12365240e-01 -1.04835522e+00 -1.12691857e-01 -1.73795894e-01
1.79264739e-01 -3.76644790e-01 -3.75549287e-01 8.67604494e-01
7.41463155e-02 -1.27185595e+00 3.69606733e-01 6.43028617e-01
5.29381931e-01 -6.38994455e-01 1.12256789e+00 6.04396880e-01
-1.31977284e+00 1.38749078e-01 -1.74482688e-01 -3.44291151e-01
4.01183069e-01 8.73159468e-01 -6.81949317e-01 1.05677724e+00
5.84279358e-01 4.80071783e-01 -5.90058982e-01 5.73123097e-01
-7.44627237e-01 6.86917484e-01 -3.85674894e-01 -3.17669548e-02
-2.65889764e-01 2.33857438e-01 4.27204639e-01 1.48526204e+00
-3.43875661e-02 3.78333837e-01 1.32130831e-01 8.28029871e-01
-2.54087448e-01 1.12965263e-01 -1.44980490e-01 -3.19909230e-02
6.83850884e-01 1.30782413e+00 -4.36579317e-01 -4.65492964e-01
-3.69919211e-01 9.45567489e-01 7.25346804e-01 3.34921360e-01
-1.08143747e+00 -6.05963230e-01 2.92333633e-01 -1.07959986e-01
1.95999354e-01 -1.39506832e-01 -5.57421148e-01 -1.33643425e+00
3.47009867e-01 -7.48885751e-01 2.91343093e-01 -6.42127037e-01
-9.38205183e-01 8.15673470e-01 -1.06116729e-02 -6.75699770e-01
-1.42437562e-01 -4.02006358e-01 -4.64652419e-01 1.07569742e+00
-1.35131562e+00 -1.30355263e+00 -2.88719743e-01 1.84141546e-01
2.05994740e-01 2.09537357e-01 1.08103037e+00 6.81699753e-01
-1.00883305e+00 9.35172319e-01 -7.73636639e-01 5.60223341e-01
6.51666820e-01 -1.19347441e+00 4.99136299e-01 1.14366031e+00
1.13328926e-01 1.05305350e+00 3.67680609e-01 -8.49477530e-01
-1.31319869e+00 -1.05048585e+00 1.62522447e+00 -5.25183737e-01
4.65197384e-01 -3.52898866e-01 -1.27341235e+00 8.11843634e-01
2.56348222e-01 -1.73306778e-01 7.94355392e-01 6.57403767e-01
-4.82549399e-01 1.17920414e-01 -7.97636688e-01 5.69153070e-01
1.29870629e+00 -4.87222403e-01 -9.20420349e-01 -2.69357190e-02
1.06361961e+00 -5.58361053e-01 -1.11529553e+00 8.26607525e-01
4.12622511e-01 -6.16104066e-01 7.49135911e-01 -8.96893203e-01
5.45515537e-01 -4.48662668e-01 -2.24270388e-01 -1.01034343e+00
-4.31560338e-01 -5.52788734e-01 -9.38659191e-01 1.78337455e+00
8.75685036e-01 -4.60305125e-01 2.43771583e-01 9.52557981e-01
-1.31559193e-01 -1.12219346e+00 -5.75801551e-01 -4.81924832e-01
-1.00441530e-01 -4.60441411e-01 7.50771999e-01 9.27630126e-01
1.64257377e-01 1.15204859e+00 -2.18346462e-01 4.95456100e-01
9.33158323e-02 1.23859502e-01 5.35243988e-01 -1.01964653e+00
-5.09470761e-01 -2.31590793e-01 5.25817694e-03 -1.43720078e+00
1.81358084e-01 -7.91633785e-01 -6.54337257e-02 -1.67621803e+00
3.26572925e-01 -3.40706229e-01 -4.14230168e-01 8.20414245e-01
-8.64208281e-01 -3.31022143e-01 1.02937847e-01 3.15357782e-02
-6.59319699e-01 5.34167647e-01 1.07679939e+00 3.22348326e-02
-2.59566218e-01 -1.56083897e-01 -9.81930435e-01 6.67891681e-01
7.15408087e-01 -4.57332581e-01 -2.56284863e-01 -6.85279012e-01
4.24932003e-01 1.41166225e-01 2.70042811e-02 -6.88805580e-01
4.53629136e-01 4.31314409e-02 4.37454939e-01 -6.25571609e-01
1.41653016e-01 -4.56316352e-01 -2.72360221e-02 2.39745244e-01
-1.76467434e-01 -3.83943096e-02 2.39355966e-01 4.01193708e-01
-3.05425018e-01 -7.69258738e-02 3.00092965e-01 1.19183347e-01
-6.41388953e-01 -1.09316707e-02 -6.83742166e-02 1.23052485e-01
8.54998648e-01 2.13179275e-01 -5.23939908e-01 2.25681160e-03
-8.99890363e-01 5.30236423e-01 -1.88307881e-01 5.61327219e-01
4.50137526e-01 -1.09273505e+00 -8.09618831e-01 1.33132741e-01
5.02296388e-02 3.88631046e-01 3.03213626e-01 9.31294203e-01
1.03574798e-01 3.01614434e-01 1.10301495e-01 -2.20655143e-01
-1.21853507e+00 5.51564157e-01 1.63858533e-01 -9.77520823e-01
-3.01599979e-01 1.10029638e+00 1.63387343e-01 -3.62172484e-01
1.83631107e-01 -3.83733869e-01 -5.44093370e-01 2.17135191e-01
7.45786667e-01 2.28059158e-01 2.51593679e-01 -3.86903524e-01
-7.29117393e-01 1.76125094e-01 -5.02077758e-01 1.79968908e-01
1.19846952e+00 -1.70352578e-01 -3.82322758e-01 3.06152552e-01
8.55205417e-01 -3.13446042e-03 -6.54330850e-01 -3.14006835e-01
2.75039762e-01 1.32452354e-01 -1.17373876e-01 -1.25636601e+00
-7.82644987e-01 9.72081423e-01 2.08825156e-01 6.64584339e-02
1.10098851e+00 1.44926580e-02 1.04410303e+00 4.14811432e-01
1.91314280e-01 -8.76566410e-01 -4.27285373e-01 8.85207534e-01
8.68134022e-01 -1.23133337e+00 -1.56373009e-02 -9.30971801e-01
-5.39260685e-01 8.84168446e-01 1.04202151e+00 2.79325008e-01
6.44010007e-01 5.98947167e-01 -2.29309455e-01 3.24554667e-02
-1.04561782e+00 -2.76504487e-01 5.98782420e-01 3.72076899e-01
8.88050497e-01 2.70350873e-01 -6.53094292e-01 1.63966560e+00
-7.78722093e-02 -9.30406302e-02 1.73357025e-01 8.55934858e-01
-2.63059527e-01 -1.27993071e+00 7.75589570e-02 3.28837842e-01
-5.81155300e-01 -4.99917716e-01 -3.44302535e-01 5.90836048e-01
2.06457496e-01 1.27601802e+00 -1.79170743e-01 -7.27976382e-01
5.12432396e-01 3.82203072e-01 2.55684525e-01 -6.94595873e-01
-8.45748305e-01 -9.51261818e-03 6.33239567e-01 -4.99266207e-01
-1.64580926e-01 -2.47304425e-01 -1.61973202e+00 -1.39492646e-01
-6.48844957e-01 8.72360468e-02 2.68910021e-01 1.26122630e+00
7.25000501e-01 1.05973804e+00 1.21237122e-01 -1.33665845e-01
-3.49666893e-01 -1.28252935e+00 -4.13415171e-02 2.10944071e-01
-1.70354337e-01 -7.31901765e-01 -1.87960789e-02 -4.15644087e-02] | [9.288496017456055, 8.734658241271973] |
aa0afb0d-58ee-452b-95b8-7edb3bb26aa0 | implicit-anatomical-rendering-for-medical | 2304.03209 | null | https://arxiv.org/abs/2304.03209v1 | https://arxiv.org/pdf/2304.03209v1.pdf | Implicit Anatomical Rendering for Medical Image Segmentation with Stochastic Experts | Integrating high-level semantically correlated contents and low-level anatomical features is of central importance in medical image segmentation. Towards this end, recent deep learning-based medical segmentation methods have shown great promise in better modeling such information. However, convolution operators for medical segmentation typically operate on regular grids, which inherently blur the high-frequency regions, i.e., boundary regions. In this work, we propose MORSE, a generic implicit neural rendering framework designed at an anatomical level to assist learning in medical image segmentation. Our method is motivated by the fact that implicit neural representation has been shown to be more effective in fitting complex signals and solving computer graphics problems than discrete grid-based representation. The core of our approach is to formulate medical image segmentation as a rendering problem in an end-to-end manner. Specifically, we continuously align the coarse segmentation prediction with the ambiguous coordinate-based point representations and aggregate these features to adaptively refine the boundary region. To parallelly optimize multi-scale pixel-level features, we leverage the idea from Mixture-of-Expert (MoE) to design and train our MORSE with a stochastic gating mechanism. Our experiments demonstrate that MORSE can work well with different medical segmentation backbones, consistently achieving competitive performance improvements in both 2D and 3D supervised medical segmentation methods. We also theoretically analyze the superiority of MORSE. | ['James S. Duncan', 'Lawrence Staib', 'Yifei Min', 'Weicheng Dai', 'Chenyu You'] | 2023-04-06 | null | null | null | null | ['neural-rendering'] | ['computer-vision'] | [ 4.01478946e-01 3.13423276e-01 -6.02460876e-02 -4.52292174e-01
-1.20953310e+00 -1.68861255e-01 1.96155459e-01 2.49410853e-01
-4.06244606e-01 3.29804927e-01 6.20193146e-02 -1.44507438e-01
-1.36972338e-01 -7.93982804e-01 -4.90997821e-01 -7.88448513e-01
-9.44381356e-02 5.16446292e-01 2.68665344e-01 -8.29508826e-02
8.19647759e-02 6.34791851e-01 -1.11598587e+00 3.86820972e-01
1.17901897e+00 9.92855310e-01 2.72049665e-01 5.89709818e-01
-1.91684231e-01 5.21851718e-01 -3.08279634e-01 -7.80926049e-02
9.95737910e-02 -4.62596923e-01 -8.85351241e-01 3.21050823e-01
3.75267863e-01 -2.15023905e-02 3.27101648e-02 1.02000880e+00
5.98460615e-01 7.10974112e-02 7.51125634e-01 -4.85010207e-01
-3.54602367e-01 2.69336283e-01 -8.02914321e-01 1.74049020e-01
-1.98358014e-01 1.36111543e-01 8.10500383e-01 -7.47480750e-01
4.96858776e-01 9.95148361e-01 1.02346134e+00 4.17850971e-01
-1.51769483e+00 -1.82906538e-01 1.70320064e-01 -3.56170654e-01
-1.27930319e+00 -3.29663604e-02 9.88527417e-01 -5.58372557e-01
5.06733418e-01 3.83583635e-01 7.20101297e-01 4.52449799e-01
4.41525221e-01 8.95108104e-01 1.21290231e+00 -2.21991986e-01
2.66106874e-01 -2.67154694e-01 2.82991290e-01 9.62097406e-01
-8.58026668e-02 -4.10906598e-02 -1.41414851e-01 -2.29794502e-01
1.29786921e+00 3.39144394e-02 -3.34202439e-01 -3.82075012e-01
-1.09210682e+00 1.02344811e+00 9.96893466e-01 4.67947364e-01
-7.43712842e-01 3.49225730e-01 2.13433161e-01 -3.22762102e-01
6.38301492e-01 4.24748689e-01 -2.10303813e-01 1.92124024e-01
-1.27011645e+00 1.02677837e-01 4.27233011e-01 3.79843622e-01
6.09848559e-01 -2.42636964e-01 -2.52313286e-01 9.47179675e-01
3.54317784e-01 -4.15484346e-02 4.17690575e-01 -1.06971717e+00
-5.70579134e-02 4.78056222e-01 -2.07042828e-01 -8.08037281e-01
-7.27030218e-01 -8.21878433e-01 -1.16848505e+00 3.72503877e-01
3.54287982e-01 9.22066867e-02 -1.33942568e+00 1.55045569e+00
5.59100270e-01 3.09873134e-01 -2.43368208e-01 1.10908651e+00
6.88570976e-01 4.59345192e-01 3.70410681e-01 -6.76368997e-02
1.48792696e+00 -8.04758549e-01 -5.70458114e-01 -8.89145657e-02
7.32949078e-01 -4.93783951e-01 1.12408054e+00 2.22232908e-01
-1.28082621e+00 -5.13736308e-01 -8.22144926e-01 -1.28917798e-01
2.86793690e-02 -2.66150851e-02 7.30928898e-01 5.89614987e-01
-1.10766625e+00 7.16526806e-01 -1.33730340e+00 2.37348497e-01
9.29852068e-01 5.54076493e-01 4.07413440e-03 1.32106215e-01
-9.37978864e-01 5.74795902e-01 2.18380928e-01 2.43564069e-01
-3.88182431e-01 -9.57475901e-01 -8.95228505e-01 -1.08272053e-01
1.26104161e-01 -1.11605346e+00 1.04546869e+00 -8.94264340e-01
-1.38545251e+00 1.20767403e+00 -1.51371390e-01 -5.46019077e-01
4.93183196e-01 1.29574323e-02 1.25807449e-01 4.18610334e-01
1.72817394e-01 8.87029588e-01 6.77269876e-01 -1.37419963e+00
-4.21100169e-01 -5.52782118e-01 -1.43052608e-01 2.46017009e-01
8.12997222e-02 -3.21037620e-01 -5.59739947e-01 -9.76759076e-01
5.07986128e-01 -6.77780271e-01 -9.46916938e-01 1.24979690e-01
-3.65506798e-01 7.88764581e-02 3.96343142e-01 -7.85484850e-01
1.06520164e+00 -2.06072807e+00 2.30673745e-01 4.69554931e-01
5.17759800e-01 1.92043534e-03 2.01202706e-01 -3.00649107e-01
5.53084612e-02 9.65001136e-02 -7.92847335e-01 -4.21518266e-01
-1.82894677e-01 2.24110857e-01 4.94893156e-02 4.99713898e-01
2.95973092e-01 1.25100005e+00 -8.30685496e-01 -7.01520264e-01
4.08849210e-01 8.27113688e-01 -9.62316513e-01 4.47320379e-02
-1.13366403e-01 1.07857108e+00 -7.38047957e-01 6.53168857e-01
5.97977340e-01 -5.89902401e-01 -7.41896629e-02 -4.36121315e-01
-2.84302677e-03 1.23022288e-01 -7.93816507e-01 2.10891032e+00
-6.82251751e-01 2.92933702e-01 4.09966230e-01 -1.20348907e+00
5.05210340e-01 3.47647488e-01 9.61825311e-01 -5.80590546e-01
2.65245259e-01 3.79895091e-01 -5.11759482e-02 -2.38215387e-01
9.92077142e-02 -4.84656394e-01 -5.39148189e-02 8.88226852e-02
-6.39522895e-02 -4.70074117e-01 -2.51032412e-01 -1.05574593e-01
8.98767233e-01 1.44056410e-01 1.87561154e-01 -3.87642533e-01
4.91667300e-01 1.36538446e-01 5.08839428e-01 7.27434099e-01
-6.12756237e-02 9.46520865e-01 3.74674678e-01 -4.32570398e-01
-6.33181870e-01 -1.09255016e+00 -4.75596160e-01 7.15419233e-01
2.31998339e-01 -1.69567689e-01 -1.17991662e+00 -6.88609838e-01
-2.41291046e-01 4.71750259e-01 -8.19439828e-01 3.02824825e-02
-9.31257188e-01 -9.44809198e-01 2.11066917e-01 5.02816916e-01
2.66347885e-01 -1.02517700e+00 -9.49043810e-01 6.30245984e-01
-1.26194596e-01 -8.93271625e-01 -5.13332844e-01 3.18737626e-01
-1.19361222e+00 -9.04321134e-01 -1.06449354e+00 -8.06622028e-01
8.37693036e-01 -9.02623609e-02 1.28110707e+00 1.89423949e-01
-6.63046002e-01 2.30932102e-01 -9.20987800e-02 -1.59645617e-01
-3.63367379e-01 1.26734406e-01 -6.76166356e-01 -1.87986512e-02
-2.57309765e-01 -7.65268207e-01 -8.71308625e-01 2.12582707e-01
-1.04853725e+00 4.88640517e-01 7.01334774e-01 1.06077349e+00
1.24596858e+00 -9.30249542e-02 2.64546186e-01 -1.17274594e+00
4.16158617e-01 -2.05949187e-01 -4.56990391e-01 1.60720155e-01
-1.96083158e-01 -1.31268026e-02 3.77989084e-01 -2.25497603e-01
-7.13091910e-01 2.65241176e-01 -6.60103321e-01 -3.82334203e-01
-2.94377655e-01 8.65135431e-01 1.09543979e-01 -2.91451216e-01
5.04782975e-01 2.02478990e-01 6.80181906e-02 -3.89650047e-01
3.28326911e-01 2.15833038e-01 7.01955616e-01 -7.15000689e-01
4.13164407e-01 7.43706107e-01 2.08896637e-01 -7.33592510e-01
-1.14079237e+00 -4.90203023e-01 -7.17455387e-01 -2.17647910e-01
1.32602394e+00 -6.74577594e-01 -5.44745862e-01 4.25543785e-01
-1.06735802e+00 -6.56384647e-01 -5.90783477e-01 3.42660666e-01
-8.76636744e-01 2.22302273e-01 -7.56373823e-01 -5.13687670e-01
-4.72282678e-01 -1.74251556e+00 1.56871080e+00 2.42482021e-01
-2.27826282e-01 -1.20835137e+00 -3.66554335e-02 3.21250558e-01
3.67435098e-01 7.70541012e-01 9.63610947e-01 -1.99369535e-01
-5.90784073e-01 -1.11230046e-01 -2.14551896e-01 1.55705988e-01
8.49332958e-02 -4.18449163e-01 -9.06713188e-01 -6.38303310e-02
2.64104635e-01 -1.53240681e-01 9.11408067e-01 1.08893335e+00
1.70175171e+00 2.02423751e-01 -2.93902278e-01 1.14294052e+00
1.45713723e+00 -8.99055004e-02 4.99953330e-01 -5.22128753e-02
9.48774815e-01 5.93222201e-01 4.25180823e-01 1.92884654e-01
3.30051184e-01 7.40587652e-01 3.84919792e-01 -7.83156693e-01
-3.23564649e-01 7.22673722e-03 -2.63427615e-01 7.91966200e-01
1.38304234e-02 1.71046257e-01 -1.04043126e+00 3.51217508e-01
-1.74467397e+00 -4.55249429e-01 -2.20333040e-01 1.86588836e+00
1.00647831e+00 9.81983766e-02 1.16906781e-02 -2.50835847e-02
5.51372290e-01 4.77807224e-02 -4.41161394e-01 5.49859516e-02
1.30373746e-01 6.80926800e-01 4.34336901e-01 5.88275015e-01
-1.34362125e+00 8.50898206e-01 5.41331720e+00 1.05688596e+00
-1.36134410e+00 4.09588933e-01 1.14511538e+00 2.09494039e-01
-2.17099667e-01 -2.19355747e-01 -4.37246382e-01 4.00759846e-01
5.28263092e-01 3.47606182e-01 1.26449019e-01 4.88042593e-01
2.34173715e-01 -1.38906136e-01 -9.66104269e-01 1.06182086e+00
-2.90207118e-01 -1.70084286e+00 -5.56958616e-02 2.13060617e-01
7.56092846e-01 6.12192079e-02 2.44048059e-01 3.59829962e-02
6.48241192e-02 -1.28438246e+00 8.25386524e-01 4.99075830e-01
7.85223663e-01 -5.70421755e-01 4.84104395e-01 3.38358819e-01
-1.14124882e+00 3.46794754e-01 -8.84592235e-02 4.56663519e-01
4.50077266e-01 6.74957752e-01 -5.53332806e-01 6.16728067e-01
5.01909554e-01 6.03577912e-01 -2.23673061e-01 1.19638157e+00
1.73202511e-02 5.29088795e-01 -3.57706577e-01 5.07001221e-01
6.20527089e-01 -3.22273642e-01 4.16235089e-01 1.09555721e+00
1.20654106e-01 3.29368711e-01 3.60629141e-01 1.14337540e+00
1.24630354e-01 1.26682028e-01 -2.00459547e-02 2.84560531e-01
-1.42285258e-01 1.30439484e+00 -1.35032094e+00 -3.62301350e-01
-1.67163759e-01 8.59666288e-01 2.75103599e-01 2.78241396e-01
-8.96032453e-01 5.98250702e-02 3.74135077e-01 3.48790377e-01
3.03701490e-01 -2.12603673e-01 -7.88902462e-01 -9.02498364e-01
-1.27892032e-01 -5.75159550e-01 3.75355273e-01 -3.79871041e-01
-1.20499933e+00 6.78016663e-01 -2.70969719e-01 -1.04915714e+00
-5.37431752e-03 -4.37543899e-01 -5.10900557e-01 1.16076171e+00
-1.55407107e+00 -1.24488699e+00 -3.16647172e-01 4.18004602e-01
4.97131795e-01 4.88798708e-01 8.59831631e-01 3.51129651e-01
-2.79429674e-01 4.10102278e-01 -1.78159148e-01 2.84455746e-01
2.49946371e-01 -1.40292454e+00 2.89721221e-01 5.06609440e-01
1.24742173e-01 4.83831406e-01 3.99213195e-01 -6.49284840e-01
-9.55488443e-01 -1.16226399e+00 2.80742645e-01 -1.52999893e-01
3.26391995e-01 -1.95706308e-01 -1.02898204e+00 2.76586503e-01
-2.06880644e-01 5.88973880e-01 6.17688835e-01 -1.02010205e-01
1.51547596e-01 2.75242299e-01 -1.23138380e+00 5.41590214e-01
8.61491382e-01 -4.02974635e-01 -4.58437979e-01 3.68690789e-01
5.25025189e-01 -9.37451839e-01 -1.16364264e+00 6.67588055e-01
2.74006248e-01 -9.55749750e-01 1.24361622e+00 -3.16136986e-01
4.76002008e-01 -2.16578737e-01 1.07697658e-01 -1.16134489e+00
-3.27583700e-01 -5.66892684e-01 5.76998331e-02 6.24766946e-01
1.98312983e-01 -5.44022501e-01 1.00227845e+00 5.23633063e-01
-6.34752452e-01 -1.35000277e+00 -1.11301720e+00 -3.13359171e-01
3.14243078e-01 -5.83857059e-01 2.71252483e-01 8.22033465e-01
-4.43783343e-01 -6.77095074e-03 1.17120646e-01 2.53220499e-01
7.40607858e-01 2.95028001e-01 3.12628895e-01 -1.26101375e+00
-4.70514476e-01 -8.79541576e-01 -3.57199937e-01 -1.20108151e+00
8.17936063e-02 -1.11984253e+00 1.29975408e-01 -1.74580789e+00
2.70521641e-02 -9.98693585e-01 -4.87960577e-01 3.19102854e-01
-5.33611774e-01 5.46962202e-01 -8.21129978e-02 3.03898692e-01
-3.66192251e-01 4.18810278e-01 1.68119109e+00 -8.24610814e-02
-2.85360396e-01 1.07397877e-01 -6.00315213e-01 8.80079806e-01
5.52400351e-01 -4.38153535e-01 -1.84560239e-01 -3.84991378e-01
-2.32237175e-01 3.59930128e-01 7.20371485e-01 -9.94915783e-01
5.37501238e-02 2.39176393e-01 4.51288253e-01 -6.85516834e-01
3.28214258e-01 -7.85907626e-01 4.30558771e-02 5.46261609e-01
-5.07753193e-01 -3.33940834e-01 1.22550443e-01 4.13727432e-01
-3.81305367e-01 -1.31595477e-01 1.00282180e+00 -2.68517911e-01
-3.64967346e-01 6.52953863e-01 -1.40970454e-01 2.35443398e-01
8.35730493e-01 -3.37173253e-01 5.12392759e-01 7.10826041e-03
-1.20720220e+00 8.01824331e-02 4.69582319e-01 -9.12322029e-02
5.67884564e-01 -1.07031178e+00 -6.49282277e-01 1.53619289e-01
-2.45569453e-01 6.19237006e-01 5.52054882e-01 1.38446379e+00
-7.56911278e-01 2.18021080e-01 9.72775072e-02 -1.07308960e+00
-9.03755307e-01 1.93763420e-01 7.99972594e-01 -7.04014719e-01
-1.01628900e+00 1.10139251e+00 8.00319791e-01 -2.95862705e-01
1.18170105e-01 -6.89639330e-01 -9.27023739e-02 -1.35157943e-01
1.77191004e-01 -1.76933035e-01 2.52929777e-01 -6.93806350e-01
-2.84996212e-01 8.10946107e-01 -3.44302803e-02 -1.62785172e-01
1.41397381e+00 1.66115612e-01 -4.36286032e-02 2.57207394e-01
1.29166603e+00 -5.70206791e-02 -1.36150229e+00 -2.70329416e-01
-2.55917269e-03 -2.21541792e-01 5.10012686e-01 -7.29191601e-01
-1.32710552e+00 9.66146827e-01 7.24541068e-01 7.64599666e-02
1.17392921e+00 1.12953804e-01 1.14320207e+00 -4.22270209e-01
2.87173390e-01 -8.16528320e-01 -1.79889515e-01 1.60697386e-01
6.73211336e-01 -1.22098267e+00 -5.79229519e-02 -9.01118636e-01
-4.25576776e-01 1.03569758e+00 2.26910502e-01 -3.27177584e-01
7.76659667e-01 4.19689178e-01 6.35772422e-02 -5.81539392e-01
-1.43362716e-01 -2.90161073e-01 6.58297658e-01 3.49719822e-01
6.42136991e-01 2.05815762e-01 -1.86886102e-01 6.23807847e-01
2.12524086e-02 7.02895597e-03 -1.18246682e-01 9.37944353e-01
-1.69466928e-01 -1.03308165e+00 -3.61370087e-01 4.89114881e-01
-6.51688337e-01 -2.02821702e-01 1.12945817e-01 6.40855312e-01
2.39747941e-01 3.07150126e-01 1.87923521e-01 1.16831549e-01
2.49429792e-01 -3.04468781e-01 6.10647798e-01 -7.47824192e-01
-8.94321918e-01 5.75859547e-01 -4.04147029e-01 -6.99970007e-01
-3.50589603e-01 -6.23141885e-01 -1.70756149e+00 4.21727896e-01
-3.96527983e-02 1.98768266e-02 5.29393435e-01 8.33567917e-01
1.96088955e-01 9.74072874e-01 4.44868594e-01 -1.17507291e+00
-3.25052023e-01 -4.28615063e-01 -3.78636986e-01 4.53923345e-01
3.21528554e-01 -5.65096676e-01 1.97101310e-02 -2.00745575e-02] | [14.41737174987793, -2.3663878440856934] |
902402e6-d081-4e29-ae1e-86653fe6e04c | cleanclip-mitigating-data-poisoning-attacks | 2303.03323 | null | https://arxiv.org/abs/2303.03323v2 | https://arxiv.org/pdf/2303.03323v2.pdf | CleanCLIP: Mitigating Data Poisoning Attacks in Multimodal Contrastive Learning | Multimodal contrastive pretraining has been used to train multimodal representation models, such as CLIP, on large amounts of paired image-text data. However, previous studies have revealed that such models are vulnerable to backdoor attacks. Specifically, when trained on backdoored examples, CLIP learns spurious correlations between the embedded backdoor trigger and the target label, aligning their representations in the joint embedding space. Injecting even a small number of poisoned examples, such as 75 examples in 3 million pretraining data, can significantly manipulate the model's behavior, making it difficult to detect or unlearn such correlations. To address this issue, we propose CleanCLIP, a finetuning framework that weakens the learned spurious associations introduced by backdoor attacks by independently re-aligning the representations for individual modalities. We demonstrate that unsupervised finetuning using a combination of multimodal contrastive and unimodal self-supervised objectives for individual modalities can significantly reduce the impact of the backdoor attack. Additionally, we show that supervised finetuning on task-specific labeled image data removes the backdoor trigger from the CLIP vision encoder. We show empirically that CleanCLIP maintains model performance on benign examples while erasing a range of backdoor attacks on multimodal contrastive learning. | ['Kai-Wei Chang', 'Aditya Grover', 'Fan Yin', 'Yu Yang', 'Nishad Singhi', 'Hritik Bansal'] | 2023-03-06 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 5.09595931e-01 7.23074079e-02 -3.39176953e-01 -2.68149406e-01
-1.02831531e+00 -1.23714983e+00 8.16432655e-01 -1.32017151e-01
-4.55837935e-01 4.81805116e-01 1.65806621e-01 -1.70023710e-01
1.90816715e-01 -3.08967829e-01 -1.42945373e+00 -6.78281009e-01
-9.13432389e-02 7.15484023e-02 -1.46944925e-01 -3.20088118e-02
4.74531427e-02 2.46700242e-01 -1.37383687e+00 6.66525543e-01
5.63025177e-01 5.52363694e-01 -1.87287465e-01 9.07645345e-01
3.14189941e-01 7.42503464e-01 -8.46948683e-01 -8.13844562e-01
3.82906616e-01 -2.69420892e-01 -3.83526683e-01 -1.07721984e-01
1.22898686e+00 -5.95752895e-01 -7.59753346e-01 1.28743100e+00
2.84607232e-01 -1.24117360e-01 6.79377973e-01 -1.58826327e+00
-9.04357195e-01 8.04627776e-01 -6.42360926e-01 1.23075023e-01
9.08464119e-02 5.88728487e-01 1.02887666e+00 -7.41324663e-01
6.10451102e-01 1.26137066e+00 5.41234970e-01 9.02778983e-01
-1.74217749e+00 -1.10640645e+00 -8.94768313e-02 -1.28856733e-01
-1.29040766e+00 -6.51735902e-01 5.46006203e-01 -3.96187335e-01
7.53933251e-01 4.12981063e-01 1.16819568e-01 1.91105103e+00
2.20195904e-01 5.86699009e-01 1.11769581e+00 -3.58482361e-01
1.84495505e-02 3.98924083e-01 1.79434538e-01 9.07830179e-01
1.30472660e-01 6.21315122e-01 -8.46823871e-01 -4.37762618e-01
3.69298995e-01 -5.18009439e-02 -2.65024185e-01 -4.54338223e-01
-1.13343120e+00 9.40266192e-01 4.93015170e-01 -6.29432127e-02
2.29929626e-01 4.21920180e-01 3.96957397e-01 3.84540737e-01
-8.48397538e-02 8.31560910e-01 -3.00578266e-01 5.94663434e-02
-5.75740516e-01 -4.17889096e-02 6.02420032e-01 8.61235499e-01
8.43959689e-01 -9.92542952e-02 -2.61323541e-01 5.79079747e-01
5.72815537e-02 8.79734576e-01 4.53491390e-01 -8.67320478e-01
7.84962058e-01 2.74987310e-01 -3.06460887e-01 -9.36985314e-01
-2.23672073e-02 -2.79015571e-01 -5.20178556e-01 1.61089152e-01
4.47524846e-01 -4.17666972e-01 -1.07927024e+00 2.31353402e+00
-3.63390446e-02 1.53054044e-01 1.96108043e-01 7.38508761e-01
5.58526933e-01 5.72687566e-01 1.83390886e-01 2.31732666e-01
1.26525152e+00 -8.71142507e-01 -6.92495227e-01 -3.17285538e-01
5.55101573e-01 -5.66247344e-01 1.45233452e+00 3.14107120e-01
-8.46728921e-01 -1.96576625e-01 -1.36378086e+00 2.68016197e-02
-5.69469988e-01 -3.17345798e-01 5.79092801e-01 9.42930698e-01
-5.92250645e-01 3.88196945e-01 -7.75306284e-01 -1.23756438e-01
6.49844825e-01 5.31449854e-01 -9.52602327e-01 -3.46742839e-01
-1.15932763e+00 7.82325864e-01 2.03887120e-01 -6.75580502e-02
-1.45089352e+00 -7.39546001e-01 -1.03328192e+00 1.51406839e-01
5.11081576e-01 -5.08622944e-01 9.54619825e-01 -1.02929413e+00
-1.16029322e+00 7.18253195e-01 3.45617086e-02 -4.77601647e-01
3.63668084e-01 -3.38641703e-01 -3.68886948e-01 2.64406890e-01
-7.61110783e-02 9.64021444e-01 1.68474364e+00 -1.61835778e+00
-4.45975289e-02 -3.12014729e-01 -8.41226354e-02 1.66486964e-01
-8.26888978e-01 -1.58659726e-01 -4.09683645e-01 -5.82006991e-01
-4.44673359e-01 -1.09061027e+00 4.64448482e-02 -3.03769618e-01
-9.78512466e-01 4.57201242e-01 8.95568728e-01 -3.15561771e-01
8.23849320e-01 -2.48221922e+00 2.20120877e-01 4.61924195e-01
2.97077119e-01 1.01787217e-01 -6.92859471e-01 3.68449003e-01
-2.74339348e-01 3.82258862e-01 -2.69985378e-01 -5.54478765e-01
2.06235319e-01 4.51881796e-01 -8.02648187e-01 6.40275419e-01
2.27851897e-01 9.86113310e-01 -7.21412122e-01 -2.73518831e-01
7.94131160e-02 5.61083555e-01 -7.90935338e-01 3.55382204e-01
-1.30284250e-01 1.92637771e-01 2.39655688e-01 6.98795617e-01
6.73604488e-01 1.26625095e-02 7.25055560e-02 -5.00056088e-01
4.98641729e-01 -1.41175479e-01 -5.56250870e-01 1.44519460e+00
-2.33281747e-01 7.88293183e-01 -1.03925504e-01 -5.47845423e-01
2.24403471e-01 7.42201954e-02 -5.80204614e-02 -5.96010566e-01
5.87001480e-02 -1.05360173e-01 -1.74742080e-02 -6.24510050e-01
4.92771536e-01 -3.30325961e-01 -3.39513302e-01 4.77934241e-01
4.84775007e-01 -1.10971622e-01 -1.61223739e-01 5.82881570e-01
1.32086015e+00 -4.29472804e-01 -3.35184157e-01 3.41953903e-01
-2.59069856e-02 -1.11303039e-01 1.19437732e-01 1.20310509e+00
-1.86288610e-01 7.27580667e-01 6.87827170e-01 1.62856117e-01
-9.81116474e-01 -1.56953228e+00 -4.21780087e-02 1.28500342e+00
1.55673131e-01 -7.07827389e-01 -9.85893130e-01 -1.21369112e+00
2.55297869e-01 7.25251913e-01 -1.01476228e+00 -7.96180904e-01
-1.46444783e-01 -6.95451200e-01 1.24418330e+00 3.62533152e-01
2.72962779e-01 -6.24098122e-01 -1.64398566e-01 -4.16643351e-01
4.38039303e-02 -1.18188047e+00 -7.26482034e-01 5.58852077e-01
-5.26397228e-01 -1.02844501e+00 -3.08739036e-01 -3.97888541e-01
1.04771900e+00 1.05435945e-01 8.46452415e-01 -9.67114568e-02
-3.02082896e-01 9.42675710e-01 -5.13251238e-02 -7.86513239e-02
-6.72748446e-01 -6.07161457e-03 1.32292613e-01 1.39829680e-01
1.57048538e-01 -4.59672719e-01 -1.01541117e-01 2.88974255e-01
-1.43790948e+00 -3.04796427e-01 5.66101909e-01 1.26209950e+00
2.16671422e-01 6.46507293e-02 2.72124648e-01 -9.96020675e-01
7.84804165e-01 -4.39577669e-01 -4.99005228e-01 3.02257717e-01
-1.95093885e-01 3.42508674e-01 4.96409446e-01 -1.07557571e+00
-6.41819656e-01 4.17564884e-02 2.45253950e-01 -1.07523894e+00
-1.23881228e-01 3.70445997e-01 -2.72160292e-01 -2.18571916e-01
1.01698470e+00 1.27101049e-01 5.92968166e-02 -1.52964994e-01
8.35601807e-01 3.65823954e-01 9.51151013e-01 -6.84639215e-01
1.34009099e+00 3.97924602e-01 -2.07344517e-01 -7.59880722e-01
-7.49385357e-01 1.85755819e-01 -1.89521909e-01 -9.01580825e-02
7.80018806e-01 -9.13317025e-01 -7.95323133e-01 3.18536162e-01
-9.06971276e-01 -3.71285409e-01 -1.13109060e-01 3.42330873e-01
-3.22390735e-01 4.18610036e-01 -6.82374418e-01 -5.13988018e-01
9.52143371e-02 -1.29813838e+00 8.07528019e-01 1.46720886e-01
-3.25790107e-01 -7.93108761e-01 1.45218283e-01 5.83130419e-01
2.34418616e-01 7.08189756e-02 1.23426628e+00 -8.70794594e-01
-5.15082896e-01 -3.53659272e-01 -6.93191439e-02 6.53626859e-01
8.24485719e-02 2.46207304e-02 -1.42183173e+00 -5.24278820e-01
-1.06328376e-01 -9.48106825e-01 1.12156248e+00 -2.27440253e-01
1.02481735e+00 -6.41254783e-01 -1.60937548e-01 7.46777177e-01
1.06149018e+00 -2.60935456e-01 5.59480190e-01 2.58450300e-01
9.81023610e-01 4.50664669e-01 1.27524823e-01 4.82230000e-02
-1.41479701e-01 4.09617305e-01 7.76167750e-01 6.12387806e-02
3.42336856e-02 -7.36781716e-01 8.10954809e-01 2.48344049e-01
4.99218225e-01 -1.70113012e-01 -6.60249233e-01 2.66935915e-01
-1.56543088e+00 -9.80814517e-01 6.35417700e-01 2.37516928e+00
1.06421709e+00 2.52112091e-01 -1.17549986e-01 -2.10546538e-01
7.01742172e-01 1.42063200e-01 -6.94331288e-01 -3.60119313e-01
-3.56843442e-01 2.02282757e-01 8.69800031e-01 6.33900046e-01
-1.25876355e+00 1.10345697e+00 6.67140865e+00 6.81131542e-01
-1.22869754e+00 1.51488647e-01 3.84412199e-01 -5.39840519e-01
-5.31508207e-01 -1.07100233e-01 -5.96908808e-01 4.70302522e-01
8.93253446e-01 2.04203725e-01 8.43242288e-01 7.18626082e-01
-4.41369981e-01 3.75951678e-02 -1.59313321e+00 1.07208836e+00
5.49501717e-01 -1.25308454e+00 3.27350765e-01 3.25406820e-01
7.40270019e-01 1.06954083e-01 8.73621404e-01 4.71945465e-01
2.97911704e-01 -1.43930674e+00 6.72326863e-01 2.52907425e-01
8.05747211e-01 -8.22022557e-01 3.94739538e-01 1.19082853e-01
-2.57637531e-01 -3.08835477e-01 -2.52981097e-01 3.80336523e-01
-2.84526587e-01 1.76149905e-01 -9.94085371e-01 -8.21897835e-02
4.30793434e-01 2.49842674e-01 -1.07442248e+00 5.37496567e-01
-2.71676779e-01 6.89110458e-01 -3.02482724e-01 3.84491563e-01
1.26230344e-01 2.18435973e-01 7.44926095e-01 1.25820506e+00
-8.44898745e-02 -4.97743338e-01 -2.46922001e-01 1.01315975e+00
-5.88863552e-01 -4.01964188e-01 -1.14095008e+00 -6.11706913e-01
4.75611717e-01 1.09355116e+00 -1.01452433e-01 -2.28941113e-01
-1.79616243e-01 1.19340551e+00 5.58745146e-01 6.24195516e-01
-9.75767791e-01 -2.71363497e-01 7.25448072e-01 -1.86380014e-01
2.73832142e-01 -1.43954545e-01 -1.68021798e-01 -1.37809813e+00
-1.56382933e-01 -1.43303216e+00 4.34418172e-01 -8.13282371e-01
-1.54092014e+00 3.36244583e-01 -2.60797311e-02 -9.07521367e-01
-1.78845614e-01 -6.09086335e-01 -5.16121268e-01 5.86110115e-01
-1.06240427e+00 -1.24270868e+00 8.50559622e-02 9.65391994e-01
-1.11684324e-02 -3.81369591e-01 9.53376651e-01 8.08400512e-02
-7.93220222e-01 1.35551810e+00 3.06342058e-02 2.59329081e-01
1.17820001e+00 -1.21627939e+00 -7.01242536e-02 8.23655665e-01
4.65797812e-01 1.10897517e+00 8.41188908e-01 -5.84455609e-01
-1.80162227e+00 -9.14035976e-01 2.01123834e-01 -8.17553461e-01
8.18128169e-01 -8.01171601e-01 -9.17094290e-01 9.34783399e-01
4.03730392e-01 1.56661138e-01 9.63738441e-01 1.55627221e-01
-1.33999336e+00 -1.95077255e-01 -1.12913418e+00 9.81929243e-01
7.29898930e-01 -1.24334097e+00 -6.31824970e-01 3.46072674e-01
7.50132620e-01 -2.90178508e-01 -7.05437720e-01 3.16000551e-01
6.50568545e-01 -7.72957683e-01 9.74689126e-01 -1.03251624e+00
6.29859626e-01 -1.01621486e-01 -5.58764637e-01 -1.29129815e+00
2.81122804e-01 -7.93043613e-01 -3.82301897e-01 1.23404217e+00
6.15804195e-01 -5.54323375e-01 7.52789080e-01 7.52924383e-01
1.18905410e-01 -9.61408094e-02 -9.81475711e-01 -7.80367970e-01
2.01255381e-01 -4.73532200e-01 2.05736116e-01 1.17532623e+00
2.60166138e-01 5.25805235e-01 -7.36280978e-01 6.65726364e-01
8.29892993e-01 -4.00465339e-01 1.12152481e+00 -3.73287082e-01
-5.55978775e-01 -1.76327080e-01 -3.48358572e-01 -6.35610104e-01
4.14989740e-01 -8.93377542e-01 3.64173502e-02 -4.39255357e-01
4.98882622e-01 -1.20817900e-01 -4.74621683e-01 7.70072341e-01
-1.81503311e-01 4.69708085e-01 4.72008973e-01 1.36152104e-01
-4.22344714e-01 4.45838690e-01 1.02635026e+00 -6.27081573e-01
8.23727325e-02 -5.00574052e-01 -8.27805340e-01 4.72288579e-01
5.27169466e-01 -7.84469068e-01 -5.41954100e-01 -4.18846041e-01
3.71690899e-01 -3.71066004e-01 6.67060971e-01 -7.87456572e-01
2.75027186e-01 -1.76878460e-02 4.75265920e-01 -1.77258432e-01
6.84429228e-01 -9.02528107e-01 -1.14233255e-01 2.97062457e-01
-8.87627184e-01 -7.56537020e-02 5.32047033e-01 7.33472049e-01
-3.32102999e-02 -2.41524294e-01 6.99199677e-01 1.40935004e-01
-2.62060940e-01 3.33797187e-02 -3.64141375e-01 1.25272855e-01
8.48532856e-01 1.53155476e-01 -8.89170051e-01 -3.76549214e-01
-6.28300488e-01 1.09271765e-01 5.20121932e-01 4.60460097e-01
7.04883814e-01 -1.30963707e+00 -2.39961892e-01 6.06123388e-01
4.47331786e-01 -4.84065622e-01 3.72013807e-01 6.53667629e-01
-4.41825092e-02 -7.77928531e-03 -2.15586320e-01 -7.04675257e-01
-1.42313969e+00 9.04105783e-01 4.00789171e-01 7.26151541e-02
-3.00966769e-01 8.54179144e-01 6.14441872e-01 -5.88486612e-01
4.67358083e-01 -1.05555713e-01 3.56327474e-01 -1.57064293e-02
5.02895176e-01 -1.86909705e-01 -1.96585163e-01 -5.38830996e-01
-2.48652831e-01 3.08240294e-01 -6.02353334e-01 -5.72690010e-01
9.01971638e-01 1.52621880e-01 1.05220862e-01 2.94251621e-01
1.62611079e+00 4.36411470e-01 -1.45429313e+00 -4.59345542e-02
-3.92828971e-01 -5.54850638e-01 -5.64234555e-02 -1.02320659e+00
-8.10697258e-01 9.55668092e-01 7.06754863e-01 4.85815890e-02
7.85871565e-01 5.39015159e-02 6.25370324e-01 8.19026768e-01
-1.10479593e-01 -8.70260596e-01 5.44902682e-01 3.21037978e-01
6.35922134e-01 -1.29447556e+00 -2.60529995e-01 -1.55046850e-01
-8.42670619e-01 7.49529183e-01 7.39899039e-01 -2.54931711e-02
2.95355350e-01 4.02918994e-01 1.42646343e-01 -2.09913671e-01
-8.01547289e-01 2.08545208e-01 1.76158503e-01 8.58288169e-01
-2.34356940e-01 2.85375547e-02 7.35999405e-01 5.21501839e-01
-1.70106471e-01 -5.47051966e-01 5.50144672e-01 9.62302864e-01
-4.04467620e-02 -8.77761662e-01 -6.93667591e-01 2.15113282e-01
-4.67906654e-01 -3.53523076e-01 -7.17629433e-01 8.37916017e-01
3.85073060e-03 7.50287473e-01 4.52061892e-02 -8.70057106e-01
2.13443920e-01 2.06376150e-01 7.71710098e-01 -5.56335390e-01
-6.14025891e-01 -1.17902026e-01 4.37842049e-02 -5.34179688e-01
2.98257116e-02 -3.39739233e-01 -8.25886548e-01 -2.61098325e-01
-2.43331611e-01 -5.54257259e-03 6.33845210e-01 9.99628305e-01
2.78925836e-01 1.99046567e-01 7.19018638e-01 -7.68009365e-01
-9.27163899e-01 -7.24509895e-01 -5.32918751e-01 6.96512997e-01
7.65995324e-01 -6.87160432e-01 -7.84549415e-01 8.41981247e-02] | [5.855416774749756, 7.896878719329834] |
d9287634-f55e-453b-be13-e068e25e7a03 | neural-multigrid-memory-for-computational | 2306.12545 | null | https://arxiv.org/abs/2306.12545v2 | https://arxiv.org/pdf/2306.12545v2.pdf | Neural Multigrid Memory For Computational Fluid Dynamics | Turbulent flow simulation plays a crucial role in various applications, including aircraft and ship design, industrial process optimization, and weather prediction. In this paper, we propose an advanced data-driven method for simulating turbulent flow, representing a significant improvement over existing approaches. Our methodology combines the strengths of Video Prediction Transformer (VPTR) (Ye & Bilodeau, 2022) and Multigrid Architecture (MgConv, MgResnet) (Ke et al., 2017). VPTR excels in capturing complex spatiotemporal dependencies and handling large input data, making it a promising choice for turbulent flow prediction. Meanwhile, Multigrid Architecture utilizes multiple grids with different resolutions to capture the multiscale nature of turbulent flows, resulting in more accurate and efficient simulations. Through our experiments, we demonstrate the effectiveness of our proposed approach, named MGxTransformer, in accurately predicting velocity, temperature, and turbulence intensity for incompressible turbulent flows across various geometries and flow conditions. Our results exhibit superior accuracy compared to other baselines, while maintaining computational efficiency. Our implementation in PyTorch is available publicly at https://github.com/Combi2k2/MG-Turbulent-Flow | ['Truong Son Hy', 'Nguyen Tri Nguyen', 'Tri Huynh', 'Tuan Anh Nguyen', 'Minh Chau Vu', 'Duc Minh Nguyen'] | 2023-06-21 | null | null | null | null | ['video-prediction'] | ['computer-vision'] | [-5.68332195e-01 -1.15204239e+00 1.93012685e-01 2.35192537e-01
-1.15275159e-01 -5.37918210e-01 6.06107116e-01 3.91788557e-02
1.54613823e-01 7.93504000e-01 4.16044623e-01 -6.13108933e-01
-2.06385449e-01 -8.08877110e-01 -1.36676744e-01 -8.24677825e-01
-1.98799074e-01 -9.46509019e-02 1.97750032e-01 -2.15305299e-01
1.95273519e-01 6.53279543e-01 -1.32248914e+00 1.40428945e-01
9.07867789e-01 1.14043856e+00 7.57892057e-02 9.88871515e-01
-6.63676765e-03 1.23024833e+00 2.22392809e-02 -5.27974889e-02
3.52932453e-01 -3.21266294e-01 -6.18645728e-01 5.04555814e-02
1.31649911e-01 -2.39657745e-01 -3.76512319e-01 6.14480972e-01
4.92751837e-01 6.71600103e-01 5.62354326e-01 -7.25045323e-01
-9.48281214e-02 1.61161106e-02 -4.98417646e-01 5.79694211e-01
-2.54138820e-02 8.66904080e-01 7.45501757e-01 -9.58706141e-01
4.11108762e-01 1.20473564e+00 6.66078866e-01 3.01301003e-01
-9.04368103e-01 -5.57725072e-01 3.26939188e-02 4.09436636e-02
-1.27906525e+00 -4.22535986e-01 5.95828652e-01 -6.95412755e-01
8.24445724e-01 6.88740194e-01 9.50099409e-01 8.61786783e-01
5.70416451e-01 4.73214149e-01 1.13682628e+00 5.99270239e-02
3.21541756e-01 -2.06429973e-01 -2.40473911e-01 4.74302143e-01
1.05119191e-01 4.66062665e-01 -4.73104835e-01 -2.61925429e-01
1.24041402e+00 -8.11555907e-02 -4.20882076e-01 1.47847861e-01
-1.09870946e+00 4.83419597e-01 2.23550200e-01 1.09531514e-01
-4.77139890e-01 1.95491701e-01 5.69686770e-01 7.20929056e-02
9.18866813e-01 5.08471668e-01 -6.24274492e-01 -7.50422180e-01
-7.79395461e-01 5.42651474e-01 6.87561154e-01 7.94665515e-01
3.53843510e-01 4.25404072e-01 -4.25620265e-02 7.39574909e-01
2.00041488e-01 4.70626742e-01 5.72623909e-01 -1.30034077e+00
3.42856973e-01 8.66848975e-02 2.40911216e-01 -1.13834393e+00
-2.11014524e-01 -5.17505705e-01 -1.43539226e+00 1.56956598e-01
1.08149506e-01 -4.83396828e-01 -4.29953068e-01 1.04262590e+00
5.02622247e-01 1.06854272e+00 3.29766989e-01 1.17057133e+00
6.74015105e-01 1.17326546e+00 2.98412234e-01 -4.06193286e-01
1.07785738e+00 -1.34389639e+00 -4.76897091e-01 5.25586829e-02
5.94994307e-01 -1.12335563e+00 7.88658440e-01 3.70798260e-01
-1.05282354e+00 -7.06747234e-01 -3.93231153e-01 1.24732725e-01
-8.76147486e-03 -2.63909400e-01 8.50369871e-01 2.24942803e-01
-9.15139437e-01 9.51455116e-01 -1.03880262e+00 -1.79831032e-02
2.77861953e-02 -2.02508882e-01 5.37606664e-02 2.16521859e-01
-9.94316697e-01 2.95642257e-01 -8.64857957e-02 3.52453202e-01
-6.46231890e-01 -1.09363198e+00 -7.53343105e-01 1.09362938e-01
1.56549364e-01 -1.05468762e+00 1.29004323e+00 -4.66749340e-01
-1.81018853e+00 5.96354194e-02 -3.48073483e-01 -2.45503440e-01
8.94873261e-01 -3.24155092e-01 -3.50631177e-01 1.36694714e-01
-2.87914515e-01 -3.07113081e-02 4.81774330e-01 -1.07656372e+00
-4.35662419e-01 1.86000869e-01 -2.27477103e-01 5.22509038e-01
-7.12090507e-02 -1.71792388e-01 -3.98449033e-01 -8.79592657e-01
-1.93636805e-01 -8.47357213e-01 -7.23867714e-01 -2.57203668e-01
-1.52026042e-01 8.18419978e-02 9.02351081e-01 -4.50047672e-01
1.44712555e+00 -2.04859042e+00 1.08892769e-01 -4.21163589e-02
2.00487092e-01 5.20565927e-01 2.06897452e-01 7.32953906e-01
1.26074672e-01 1.98317230e-01 -2.23357722e-01 -4.05505806e-01
-2.51853973e-01 -1.22414514e-01 -5.34334242e-01 4.15527225e-01
-3.88475619e-02 8.67074013e-01 -9.68343258e-01 -2.35711962e-01
7.81963468e-01 4.72437859e-01 -6.20916843e-01 2.81223893e-01
-8.64038765e-02 1.05543697e+00 -6.04167283e-01 4.98894840e-01
9.02076244e-01 -3.79545510e-01 2.37206057e-01 -1.25467747e-01
-6.29719436e-01 3.53883430e-02 -1.30615711e+00 1.17178655e+00
-8.63392889e-01 4.39216554e-01 4.07798111e-01 -5.70274651e-01
7.97358215e-01 5.32709241e-01 7.72102356e-01 -5.05131662e-01
1.62662446e-01 8.84370580e-02 -2.62608171e-01 -6.15373373e-01
7.15470672e-01 -2.10138619e-01 3.29933673e-01 9.17149708e-02
-5.23332417e-01 -5.58882654e-01 3.01004976e-01 2.29442045e-01
9.09793675e-01 2.03608528e-01 1.31399363e-01 -6.61159098e-01
5.71099520e-01 -1.35077283e-01 6.83555305e-01 3.71852130e-01
-3.98033649e-01 3.10602903e-01 3.04249406e-01 -6.88281775e-01
-8.48881721e-01 -7.52997875e-01 -2.79057115e-01 4.39205736e-01
4.50492710e-01 -8.80442321e-01 -3.83053064e-01 -2.22018268e-02
2.14470237e-01 2.63913453e-01 -2.62048215e-01 2.40370363e-01
-6.62168622e-01 -7.19894350e-01 2.49701485e-01 4.34511453e-01
7.58996785e-01 -9.95457590e-01 -3.70143056e-01 3.45429122e-01
-1.36249334e-01 -1.42845631e+00 -3.04792374e-01 -4.03789759e-01
-1.10640371e+00 -1.11612785e+00 -3.32410067e-01 -3.54968488e-01
1.70689449e-01 4.34007198e-01 1.12746966e+00 2.54537970e-01
-2.46982485e-01 2.57127762e-01 -2.54058033e-01 2.68100481e-03
-2.07338631e-01 -2.43193880e-01 2.04472169e-01 -4.82665151e-02
-2.50177860e-01 -5.65050066e-01 -9.08599496e-01 2.69757301e-01
-8.89087796e-01 4.60879087e-01 -3.47675160e-02 4.66526628e-01
5.12243688e-01 3.64617914e-01 1.62466615e-01 -7.98063159e-01
6.92380905e-01 -7.23229468e-01 -8.66023600e-01 -2.42065549e-01
-3.26244593e-01 -3.84394765e-01 1.19930899e+00 -1.80021465e-01
-1.19694877e+00 -4.27687943e-01 -3.37378114e-01 -8.21408868e-01
-2.65844643e-01 5.61925411e-01 2.92985976e-01 -9.95202735e-02
1.08534455e-01 1.79923654e-01 -2.08987683e-01 -6.85917854e-01
-1.48848921e-01 3.86242747e-01 5.13989814e-02 -8.22332561e-01
6.53992534e-01 3.03550720e-01 4.46428627e-01 -1.12557232e+00
-3.05193335e-01 -6.22306466e-01 -4.75594789e-01 -4.60077226e-01
5.55755258e-01 -1.04036093e+00 -1.04857206e+00 7.99307048e-01
-8.03238273e-01 -7.30459750e-01 -5.41152507e-02 7.28554845e-01
-3.69161904e-01 5.18341124e-01 -1.26239502e+00 -8.73595357e-01
-2.23424792e-01 -1.29719579e+00 9.56565320e-01 1.74570605e-01
-4.02565151e-02 -1.53315604e+00 1.59848675e-01 1.72117755e-01
8.34272087e-01 5.59740961e-01 4.94144708e-01 3.36035401e-01
-7.16338992e-01 4.78193581e-01 -9.78052318e-02 2.00978994e-01
3.01605791e-01 4.42726433e-01 -6.66365147e-01 -3.18463653e-01
-9.68036801e-02 -4.98532578e-02 8.60517204e-01 5.74519455e-01
1.42887735e+00 -2.96285927e-01 -1.56117037e-01 9.82303381e-01
1.49575067e+00 -1.33953005e-01 3.21687639e-01 5.20361103e-02
7.63526618e-01 2.67394662e-01 6.14862382e-01 7.94709980e-01
1.69757456e-01 5.66190839e-01 2.28056610e-01 -2.82921225e-01
-3.12294476e-02 1.02510735e-01 2.01857522e-01 1.39255607e+00
-5.17485797e-01 -3.55152160e-01 -9.90279913e-01 2.67107517e-01
-1.84434545e+00 -9.65564668e-01 -5.93082905e-01 1.83865213e+00
5.00917554e-01 -1.20035686e-01 -8.89451131e-02 -1.61329180e-01
2.91240185e-01 2.04411313e-01 -3.58710974e-01 -5.89675367e-01
1.86786234e-01 3.03146124e-01 5.18754125e-01 8.29390705e-01
-1.29414308e+00 1.07665110e+00 5.94599867e+00 7.18214214e-01
-1.42754781e+00 -1.67143464e-01 8.87381315e-01 -1.53549358e-01
5.41260876e-02 -1.56118914e-01 -6.67040169e-01 7.35653639e-01
1.12702417e+00 -3.10113519e-01 4.13422316e-01 4.54974949e-01
1.15174520e+00 -1.07944548e-01 -2.55501062e-01 7.05761135e-01
-5.58891892e-01 -1.59645700e+00 -4.91857193e-02 2.55261026e-02
9.16747987e-01 1.96319427e-02 4.85800989e-02 7.27634132e-02
2.45305449e-01 -7.87041545e-01 3.59121799e-01 5.46439290e-01
5.73309541e-01 -7.23907113e-01 6.96346223e-01 2.17469513e-01
-1.86153817e+00 1.57688424e-01 -2.31177419e-01 -4.87372041e-01
3.98025572e-01 8.20714295e-01 -2.69322306e-01 9.85888422e-01
7.31375277e-01 1.25787878e+00 -1.03298865e-01 1.08012962e+00
1.75553575e-01 1.00833535e+00 -1.92630172e-01 1.78875476e-01
4.10162896e-01 -6.21046305e-01 6.67082548e-01 1.26905358e+00
5.52489460e-01 5.79146981e-01 5.91005266e-01 3.67584556e-01
6.99755177e-02 3.53424966e-01 -4.40706998e-01 1.82200283e-01
1.90007687e-01 1.38393545e+00 -7.01062977e-01 -5.51710486e-01
-3.31042707e-01 5.29350996e-01 -1.73637290e-02 6.19910300e-01
-8.63177240e-01 2.63856049e-03 1.55042398e+00 6.27750950e-03
2.57907152e-01 -7.25250244e-01 -2.79277056e-01 -1.44697845e+00
-2.84034222e-01 -6.71103716e-01 2.92121768e-01 -6.15971148e-01
-1.32784641e+00 6.25882804e-01 -5.07332804e-03 -1.63188028e+00
4.09357175e-02 -8.20506036e-01 -7.65777647e-01 9.90226567e-01
-1.86616755e+00 -8.04882169e-01 -5.06981075e-01 3.80490988e-01
8.31881106e-01 8.82529765e-02 5.82164526e-01 2.76854068e-01
-8.67544591e-01 -3.21822643e-01 4.93314028e-01 -7.30459243e-02
4.62251782e-01 -1.13879752e+00 7.08367586e-01 9.46618438e-01
-4.63662565e-01 5.29519022e-01 6.82079792e-01 -7.87629187e-01
-1.58345532e+00 -1.47893906e+00 4.07771975e-01 -1.03869267e-01
9.02103066e-01 -1.14326164e-01 -1.03127086e+00 2.85562724e-01
2.22234040e-01 3.80472243e-01 5.24288476e-01 -1.90838650e-01
3.14171284e-01 -4.24422584e-02 -7.88360000e-01 6.10731483e-01
8.05845857e-01 -4.96364161e-02 2.34847993e-01 3.59543622e-01
4.52547401e-01 -7.92770743e-01 -1.15557301e+00 5.56088388e-01
3.11612815e-01 -1.12404025e+00 8.35965693e-01 -5.24496615e-01
8.37534547e-01 -3.65496069e-01 1.01496279e-01 -1.49812639e+00
-4.69362915e-01 -7.21180260e-01 -3.49118829e-01 1.00606751e+00
-8.65991414e-02 -6.23107255e-01 4.86179918e-01 4.69727784e-01
-1.20148890e-01 -9.91748691e-01 -7.26471543e-01 -6.20079696e-01
2.78661132e-01 -5.74698925e-01 6.33029699e-01 1.10148835e+00
-1.13379173e-01 -8.22886974e-02 -6.42525673e-01 4.94375199e-01
5.05708277e-01 3.92811298e-01 7.62989581e-01 -9.28926826e-01
-3.10283482e-01 -4.33545679e-01 -2.66355366e-01 -1.21017098e+00
1.17908463e-01 -6.65855646e-01 -1.05854839e-01 -1.33163095e+00
-2.33024597e-01 -4.38994974e-01 -1.95823893e-01 -8.88191536e-02
-3.25017333e-01 4.68482226e-01 2.99964488e-01 3.33626807e-01
-3.81549031e-01 6.46402836e-01 1.71695113e+00 2.84455568e-01
-2.30462864e-01 2.93910294e-03 -1.53391898e-01 5.59319377e-01
1.09845877e+00 2.70891160e-01 -1.52997404e-01 -4.94657695e-01
-4.27532434e-01 2.73774296e-01 3.16083759e-01 -1.27010083e+00
6.97234496e-02 -6.38212681e-01 4.04174894e-01 -4.58424315e-02
3.22657138e-01 -7.54981279e-01 1.60743713e-01 6.33652866e-01
2.91275568e-02 5.14734149e-01 8.18511784e-01 3.56372446e-01
-3.36076647e-01 3.37031364e-01 7.67351627e-01 -4.16521668e-01
-8.50060344e-01 7.17901409e-01 -7.63854086e-01 -2.26407461e-02
9.76781726e-01 1.20994933e-01 -3.26963246e-01 -3.27846766e-01
-5.56546748e-01 3.72735173e-01 5.07302463e-01 3.43447238e-01
4.75424260e-01 -1.09658682e+00 -6.26848519e-01 3.77545059e-01
-4.39332008e-01 1.57482009e-02 6.05051994e-01 9.21337187e-01
-1.29335356e+00 3.68972003e-01 1.65405855e-01 -5.34394681e-01
-1.22015715e+00 1.31653264e-01 4.80490983e-01 -5.07073879e-01
-7.78105795e-01 7.14513659e-01 3.80771369e-01 -3.56493711e-01
-3.99481058e-01 -5.14967263e-01 7.28617609e-02 -5.40571809e-01
4.29874539e-01 6.92671061e-01 7.07870126e-02 -6.13335609e-01
-4.60788570e-02 7.20758915e-01 4.82352644e-01 2.32165903e-01
1.26212776e+00 -2.78849095e-01 -2.51302999e-02 4.58365291e-01
9.29269195e-01 2.79784590e-01 -1.55130947e+00 1.00704737e-01
-4.86334324e-01 -7.84507990e-01 4.28742915e-01 -5.14141977e-01
-1.27132213e+00 7.88107455e-01 9.20850635e-02 3.38842899e-01
1.20833838e+00 -6.23030961e-01 1.20708871e+00 -2.13361010e-01
2.37928629e-01 -5.30649960e-01 -2.12510511e-01 9.47031736e-01
4.53168511e-01 -1.03747618e+00 1.55031607e-01 -8.22108030e-01
-7.85472333e-01 1.17707551e+00 9.26074862e-01 1.03665888e-01
8.66370976e-01 6.29941404e-01 4.56233054e-01 1.09019153e-01
-1.20623195e+00 -1.80640407e-02 3.55218127e-02 -1.52468085e-01
6.44187093e-01 1.08090408e-01 -1.25222608e-01 -9.62462723e-02
-9.27015543e-02 3.72674763e-02 3.15458506e-01 1.08110499e+00
-1.00343488e-01 -1.02044475e+00 -3.75334680e-01 4.28315133e-01
-3.85392934e-01 -2.22585723e-01 2.20888197e-01 4.60873604e-01
-1.53090432e-01 9.70899940e-01 2.29022995e-01 -1.93246856e-01
1.80694714e-01 -2.50280172e-01 5.83010130e-02 -2.74065197e-01
-6.96975589e-01 2.15129033e-01 -1.60598695e-01 -6.09148443e-01
-4.31215018e-01 -5.12469947e-01 -9.92855191e-01 -8.56436312e-01
3.39570604e-02 7.32513487e-01 2.90737301e-01 8.36571038e-01
6.30657673e-01 8.17782223e-01 8.99999499e-01 -1.16268051e+00
1.89224929e-01 -8.57338488e-01 -6.77819788e-01 4.97800797e-01
3.57104450e-01 -7.36809969e-01 -4.83651936e-01 1.40441015e-01] | [6.588656902313232, 3.1986076831817627] |
7de93ba1-ac9e-4666-adf4-9630980a04b2 | how-important-are-activation-functions-in | 2209.02681 | null | https://arxiv.org/abs/2209.02681v6 | https://arxiv.org/pdf/2209.02681v6.pdf | How important are activation functions in regression and classification? A survey, performance comparison, and future directions | Inspired by biological neurons, the activation functions play an essential part in the learning process of any artificial neural network commonly used in many real-world problems. Various activation functions have been proposed in the literature for classification as well as regression tasks. In this work, we survey the activation functions that have been employed in the past as well as the current state-of-the-art. In particular, we present various developments in activation functions over the years and the advantages as well as disadvantages or limitations of these activation functions. We also discuss classical (fixed) activation functions, including rectifier units, and adaptive activation functions. In addition to discussing the taxonomy of activation functions based on characterization, a taxonomy of activation functions based on applications is presented. To this end, the systematic comparison of various fixed and adaptive activation functions is performed for classification data sets such as the MNIST, CIFAR-10, and CIFAR- 100. In recent years, a physics-informed machine learning framework has emerged for solving problems related to scientific computations. For this purpose, we also discuss various requirements for activation functions that have been used in the physics-informed machine learning framework. Furthermore, various comparisons are made among different fixed and adaptive activation functions using various machine learning libraries such as TensorFlow, Pytorch, and JAX. | ['George Em Karniadakis', 'Ameya D. Jagtap'] | 2022-09-06 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-2.56758898e-01 -3.55590075e-01 8.23708624e-02 -4.81593490e-01
9.20061022e-02 -2.97925025e-01 4.79658157e-01 1.90489307e-01
-8.75847995e-01 9.47677076e-01 -2.89449662e-01 -1.48106650e-01
-4.37748611e-01 -1.02518582e+00 -4.49932545e-01 -1.10340726e+00
-2.05182955e-01 2.10650459e-01 2.05635838e-02 -3.60689014e-01
6.15215003e-01 1.06146073e+00 -1.64649713e+00 1.45396307e-01
8.26845288e-01 1.00234997e+00 -5.09954430e-02 6.30282044e-01
-3.71988654e-01 6.57403827e-01 -1.06748211e+00 -2.24744938e-02
-9.65688843e-03 -1.43656865e-01 -8.07396233e-01 -6.49980307e-01
7.55925626e-02 5.00551879e-01 -4.83465701e-01 7.23298788e-01
6.98784173e-01 6.94950342e-01 9.77511585e-01 -1.21796417e+00
-7.26828694e-01 6.23294294e-01 -2.12383285e-01 8.22479963e-01
-2.15818241e-01 7.43067637e-02 5.98088741e-01 -1.22566295e+00
3.27381678e-02 9.77453411e-01 4.02175903e-01 8.23712468e-01
-1.01266718e+00 -8.10798883e-01 3.10724139e-01 5.25592983e-01
-1.54528618e+00 -5.31476811e-02 8.27447355e-01 -3.75099272e-01
1.33240950e+00 1.57489121e-01 7.09096134e-01 9.37595725e-01
3.02145123e-01 8.08984935e-01 9.51166630e-01 -6.54000759e-01
4.85377610e-01 -4.64582331e-02 7.88532436e-01 5.30426204e-01
2.95195431e-01 1.23383261e-01 -5.46100140e-01 -2.10926309e-01
9.95760977e-01 -3.96732062e-01 -4.42747116e-01 8.04556385e-02
-1.06288922e+00 9.92170334e-01 7.09435642e-01 7.81622946e-01
-3.53993356e-01 4.52768534e-01 5.10383606e-01 8.62201303e-02
5.31961441e-01 6.57586873e-01 -4.61119145e-01 3.05966467e-01
-5.66900253e-01 2.76836872e-01 7.49166846e-01 3.73102039e-01
6.06998324e-01 8.04973423e-01 -1.95125386e-01 1.03241837e+00
1.32917956e-01 1.41235933e-01 6.73068643e-01 -7.44970143e-01
3.97289358e-02 3.37534517e-01 -2.10888147e-01 -7.04017997e-01
-6.35295212e-01 -6.51232958e-01 -1.31557238e+00 5.53775787e-01
4.09908742e-01 -1.77361026e-01 -1.04220784e+00 1.43222260e+00
1.02013111e-01 4.40601259e-01 2.36297727e-01 9.54622567e-01
1.40243101e+00 9.46907043e-01 2.96847433e-01 -1.49772391e-02
1.43161714e+00 -8.15138161e-01 -5.39619684e-01 3.23348604e-02
4.00906950e-01 -5.38391411e-01 7.81645596e-01 6.65603817e-01
-1.06962216e+00 -9.16153610e-01 -9.55449641e-01 5.70190288e-02
-9.00611699e-01 3.57833616e-02 9.86714303e-01 8.14969957e-01
-9.45172250e-01 7.12247908e-01 -8.73792171e-01 -9.44278315e-02
1.98548570e-01 7.44385719e-01 -2.43125074e-02 4.80788648e-01
-1.51459897e+00 1.03647971e+00 7.26402462e-01 1.88794926e-01
-4.24713641e-01 -8.24329674e-01 -5.09843111e-01 2.55454153e-01
-9.12674665e-02 -7.67823875e-01 1.14051676e+00 -1.07868004e+00
-1.72930908e+00 6.24785304e-01 8.53714198e-02 -6.49575174e-01
-8.26075077e-02 -1.47173047e-01 -5.11858046e-01 8.15667287e-02
-6.84222579e-01 7.62604237e-01 4.92678910e-01 -8.92380834e-01
-4.41281021e-01 -1.74507692e-01 2.74242610e-01 2.72422493e-01
-3.60399544e-01 -2.26841792e-02 -1.97678462e-01 -1.08768988e+00
-1.86227784e-01 -6.81798220e-01 -2.73380518e-01 -1.29117161e-01
5.50723216e-03 -7.45026529e-01 5.37720323e-01 -2.56624352e-02
1.13452446e+00 -1.94919324e+00 2.55500168e-01 5.36759675e-01
8.20160955e-02 3.09452653e-01 1.07169030e-02 1.13911228e-03
-4.03423607e-01 -7.01040700e-02 -4.14994895e-01 2.30414584e-01
-4.41437066e-01 3.29156905e-01 -2.08697855e-01 4.14455801e-01
-2.12595370e-02 8.65889966e-01 -6.53100073e-01 -2.05544725e-01
4.97497797e-01 8.84793460e-01 -1.97399512e-01 -1.74635705e-02
1.16604142e-01 5.55325747e-01 -5.38582444e-01 5.15090227e-01
3.66194159e-01 4.79642418e-04 -3.43549430e-01 -2.09335357e-01
-2.60222524e-01 9.74293426e-02 -1.01697028e+00 1.46406841e+00
-5.99646747e-01 6.73279643e-01 -2.06784122e-02 -1.86225867e+00
1.21015012e+00 5.47228932e-01 6.78510845e-01 -5.62690020e-01
2.34768063e-01 3.55722994e-01 5.38854063e-01 -2.29426682e-01
2.74302989e-01 -8.66115242e-02 3.73237342e-01 3.27198505e-01
3.95362318e-01 5.70262447e-02 4.42505419e-01 -1.18436627e-01
7.00679421e-01 -1.31341830e-01 3.52102876e-01 -6.21218264e-01
1.24652469e+00 -1.86164513e-01 2.96839654e-01 6.59077644e-01
7.71704167e-02 3.46356809e-01 1.17205381e-02 -8.51096034e-01
-6.68947279e-01 -8.39258015e-01 -4.97942567e-01 1.02925944e+00
-7.19574466e-02 -7.42915943e-02 -7.83265531e-01 -2.57013381e-01
-3.93237807e-02 7.17314839e-01 -7.74586856e-01 -2.82484502e-01
-8.04453313e-01 -1.56647861e+00 8.10645461e-01 7.98105538e-01
7.75903881e-01 -1.60609329e+00 -7.30524004e-01 2.03232244e-01
1.81434572e-01 -8.77077639e-01 1.25035331e-01 5.12203157e-01
-1.31672394e+00 -1.01027966e+00 -9.15103376e-01 -7.09386647e-01
6.23394668e-01 -1.21955410e-01 1.22905934e+00 1.68872699e-01
-4.84774649e-01 5.04547715e-01 -1.78261667e-01 -5.52004278e-01
-8.07501674e-02 2.91178823e-01 1.90423593e-01 -9.25401300e-02
3.87569845e-01 -4.23852503e-01 -6.54003680e-01 1.59733519e-01
-9.29633558e-01 -3.55180740e-01 5.82471788e-01 8.17675591e-01
4.96090263e-01 4.51972485e-02 7.26762354e-01 -7.83149660e-01
8.13173234e-01 -5.04813015e-01 -6.52065754e-01 1.58180192e-01
-6.79273903e-01 3.50948602e-01 8.05861235e-01 -6.13993049e-01
-1.15282857e+00 -1.59519956e-01 -5.19195735e-01 -3.12662542e-01
-1.49569139e-01 5.36810756e-01 2.81255543e-01 -5.96432269e-01
1.08558047e+00 1.44285426e-01 -3.36505800e-01 -5.32369196e-01
1.40287757e-01 4.59475905e-01 5.42957067e-01 -9.79258001e-01
2.65500546e-01 1.96488678e-01 7.63525888e-02 -1.22502029e+00
-5.90221703e-01 -4.64215428e-01 -4.71752226e-01 -1.77425534e-01
7.97867417e-01 -4.03542995e-01 -8.87803257e-01 8.10306430e-01
-1.29385579e+00 -1.27077088e-01 -4.55937415e-01 8.46052766e-01
-3.21937978e-01 -1.47935776e-02 -7.69968688e-01 -7.95751989e-01
-9.01180804e-01 -1.32535851e+00 2.50383198e-01 6.33173406e-01
-1.73662752e-02 -1.42410898e+00 4.36140131e-03 -1.05355814e-01
7.18836784e-01 7.49112815e-02 1.02914870e+00 -6.15888298e-01
-9.96441171e-02 7.20041618e-02 8.26249421e-02 3.97608936e-01
-7.92186856e-02 1.12557635e-01 -1.17780912e+00 -9.67441499e-02
-1.04694348e-02 -1.56572729e-01 1.09028983e+00 8.56190264e-01
1.84138560e+00 1.07499376e-01 -3.45965475e-01 7.57660389e-01
1.26642227e+00 5.33824444e-01 6.84203744e-01 1.49691075e-01
2.66691387e-01 5.40013075e-01 8.21865350e-02 3.32598947e-02
-3.67233545e-01 4.11457956e-01 2.78173059e-01 -1.81882009e-01
1.34345487e-01 5.63261926e-01 1.06125258e-01 9.52244401e-01
-7.32499182e-01 -2.48442098e-01 -8.21417749e-01 7.24667013e-02
-1.64810765e+00 -9.22111273e-01 -2.93803573e-01 2.20895195e+00
6.76514804e-01 2.61268109e-01 -4.60610688e-02 5.04373729e-01
7.65882611e-01 -3.84264588e-02 -6.96648061e-01 -6.62348628e-01
-1.84158325e-01 1.03167188e+00 4.18683887e-01 3.97919804e-01
-1.11684644e+00 7.27488935e-01 6.92886543e+00 1.00328302e+00
-1.53267431e+00 8.06687251e-02 7.77427018e-01 4.04783897e-02
1.69877857e-01 -3.66434246e-01 -7.32436955e-01 3.58470768e-01
1.09461784e+00 8.76476988e-03 2.98140317e-01 8.85459900e-01
-9.97566953e-02 -1.63641110e-01 -8.18684161e-01 1.21057379e+00
-8.99228081e-02 -1.57466459e+00 -7.51791745e-02 -2.41871312e-01
4.95170444e-01 3.00423801e-01 -1.56255990e-01 4.62840736e-01
-6.61247224e-02 -1.27931857e+00 3.24734479e-01 4.75572050e-01
7.72952378e-01 -8.93016398e-01 8.14722717e-01 1.69371769e-01
-1.19060302e+00 -3.37091163e-02 -5.45465112e-01 -1.84094414e-01
-2.33506978e-01 9.56092298e-01 -2.46614113e-01 6.09423816e-01
7.92684138e-01 4.63665396e-01 -1.62145063e-01 1.34237373e+00
-3.96504670e-01 7.35943615e-01 -4.71335411e-01 -4.41849351e-01
2.06085831e-01 -2.95239687e-01 2.50227034e-01 1.38595057e+00
1.91023365e-01 4.33401614e-01 -3.04594070e-01 1.04152501e+00
-5.87820560e-02 1.96324080e-01 -2.39408448e-01 1.39801487e-01
4.33372766e-01 1.50437951e+00 -1.01430321e+00 -4.58190531e-01
-2.72348255e-01 5.73244393e-01 1.53150722e-01 6.03488684e-01
-9.22427595e-01 -7.44187951e-01 7.00969696e-01 -2.17629567e-01
-4.76433903e-01 -3.89862120e-01 -4.32095766e-01 -1.10431576e+00
-5.07818460e-01 -4.04464304e-01 4.37839061e-01 -6.57714605e-01
-1.22441351e+00 8.03604841e-01 4.98720050e-01 -6.93903565e-01
1.07036419e-01 -1.22511876e+00 -9.32596445e-01 1.06027341e+00
-1.28103328e+00 -3.23398918e-01 -4.07062382e-01 6.03342116e-01
6.49648368e-01 -4.16813344e-01 9.83215332e-01 2.78118521e-01
-7.07538903e-01 2.93039739e-01 -2.96911900e-03 2.64941484e-01
3.55975926e-01 -1.10866082e+00 4.11192387e-01 3.59611243e-01
2.93077648e-01 8.08466792e-01 4.54810709e-01 -8.89093149e-03
-1.06382287e+00 -5.87627828e-01 4.73149210e-01 1.72340482e-01
3.33688498e-01 -1.66040272e-01 -1.28154945e+00 2.84327209e-01
3.12900454e-01 4.11981106e-01 5.52648723e-01 6.83912933e-02
-1.15363248e-01 -2.85317808e-01 -1.01268351e+00 3.22643667e-01
7.17966318e-01 -7.58527741e-02 -5.26964188e-01 2.00179324e-01
9.89442617e-02 -3.75686884e-01 -7.33296394e-01 6.13859951e-01
5.42098641e-01 -7.21957147e-01 1.37296617e+00 -5.49631536e-01
-7.15054348e-02 -1.70103386e-01 2.65953153e-01 -1.54072797e+00
-4.42538112e-01 -2.25523204e-01 -3.88149321e-01 7.55476236e-01
1.86866522e-01 -9.79625881e-01 8.78236234e-01 4.32018280e-01
-2.15113595e-01 -1.13532305e+00 -9.57629919e-01 -7.25581884e-01
6.13157749e-01 -4.74969715e-01 1.94568932e-01 7.68257916e-01
-1.94508970e-01 5.57221711e-01 3.91051620e-01 -2.18335554e-01
4.11811382e-01 -1.26598403e-01 2.06614226e-01 -1.60130727e+00
-3.11920792e-02 -8.92829359e-01 -4.66282994e-01 -8.13865662e-01
2.61819094e-01 -1.03490114e+00 -4.74652708e-01 -1.43756723e+00
-1.99079216e-01 -8.26489449e-01 -8.13738465e-01 6.16153419e-01
-2.34064274e-03 2.72949100e-01 -1.07801720e-01 3.39342475e-01
2.04461887e-01 4.98848557e-01 9.49866414e-01 -1.69373870e-01
-1.82667315e-01 1.30414233e-01 -3.43078911e-01 7.72548079e-01
1.41753078e+00 -3.46347988e-01 -4.83587772e-01 -4.12095100e-01
1.31843075e-01 -6.02229297e-01 1.93643525e-01 -1.31206059e+00
3.21460754e-01 -7.83927888e-02 7.79442549e-01 -4.16691720e-01
5.21139205e-01 -5.74826837e-01 -2.41267979e-01 6.84283078e-01
-3.38374525e-01 2.97045708e-01 2.80810386e-01 1.82987042e-02
-4.04713452e-01 -7.16240227e-01 1.20082450e+00 -3.18037868e-01
-8.79310608e-01 1.69498950e-01 -5.33119559e-01 8.53715315e-02
1.08919919e+00 -1.77979499e-01 -1.78154588e-01 8.81410483e-03
-6.85878277e-01 1.24760412e-01 -3.36841583e-01 2.91244030e-01
8.00410867e-01 -1.03462803e+00 -6.67339861e-01 1.97093815e-01
-1.58517793e-01 6.84514493e-02 1.31366774e-01 9.22609270e-01
-7.25901306e-01 6.57308340e-01 -5.59508502e-01 -4.09153163e-01
-1.06279910e+00 4.71665591e-01 7.25257993e-01 -1.36696398e-01
-3.92694026e-01 9.61627603e-01 3.68258536e-01 -3.55089664e-01
3.61960262e-01 -4.91905004e-01 -7.04548836e-01 -1.78245604e-01
3.54885191e-01 5.34339488e-01 5.14609337e-01 -2.94384867e-01
-4.73352045e-01 7.17870533e-01 2.19483957e-01 -9.30276662e-02
1.17892241e+00 5.70045650e-01 -4.43855263e-02 5.37084520e-01
8.45840752e-01 -4.12675977e-01 -5.32831490e-01 -1.22722670e-01
-9.66631249e-02 2.63471901e-01 3.31786126e-01 -8.83466780e-01
-1.55081129e+00 1.28452921e+00 6.07780099e-01 1.04615822e-01
1.20422363e+00 -4.57823783e-01 3.77688199e-01 6.38448298e-01
3.74388546e-01 -1.26177907e+00 -9.01132822e-02 7.33275294e-01
9.43067610e-01 -8.51433456e-01 9.15130153e-02 -3.39508057e-01
-3.27080309e-01 1.60888195e+00 8.78038347e-01 -4.20045286e-01
9.65045035e-01 4.39119011e-01 -2.07757592e-01 -1.59282610e-01
-5.43285072e-01 6.27060188e-03 4.56690609e-01 4.54849839e-01
1.11072564e+00 -2.35027298e-01 -6.40786231e-01 4.37437385e-01
-1.79815292e-02 9.39364731e-02 1.18381210e-01 9.47417259e-01
-4.94722307e-01 -1.13896883e+00 -7.63528824e-01 3.78616571e-01
-5.76417863e-01 -2.27727950e-01 -2.80819744e-01 8.56272697e-01
1.84876218e-01 6.01688862e-01 1.95889205e-01 -3.06112841e-02
2.12875888e-01 1.79823443e-01 8.07807803e-01 -4.23419923e-01
-1.08920074e+00 -3.50630581e-01 -2.14751363e-01 1.75246899e-03
-6.13826215e-01 9.94692221e-02 -1.85004604e+00 -2.32543543e-01
-3.41900915e-01 6.15592360e-01 9.95606840e-01 8.07198286e-01
9.93629321e-02 9.06634986e-01 1.29817668e-02 -1.11436391e+00
-4.05972093e-01 -1.17298591e+00 -3.93564165e-01 -3.77170220e-02
-1.52745411e-01 -1.06807065e+00 -3.16729099e-01 -2.15257391e-01] | [8.30443000793457, 3.202078342437744] |
3e21da68-fb69-45b9-bad3-ed2711ed92ca | item-graph-convolution-collaborative | 2303.15946 | null | https://arxiv.org/abs/2303.15946v1 | https://arxiv.org/pdf/2303.15946v1.pdf | Item Graph Convolution Collaborative Filtering for Inductive Recommendations | Graph Convolutional Networks (GCN) have been recently employed as core component in the construction of recommender system algorithms, interpreting user-item interactions as the edges of a bipartite graph. However, in the absence of side information, the majority of existing models adopt an approach of randomly initialising the user embeddings and optimising them throughout the training process. This strategy makes these algorithms inherently transductive, curtailing their ability to generate predictions for users that were unseen at training time. To address this issue, we propose a convolution-based algorithm, which is inductive from the user perspective, while at the same time, depending only on implicit user-item interaction data. We propose the construction of an item-item graph through a weighted projection of the bipartite interaction network and to employ convolution to inject higher order associations into item embeddings, while constructing user representations as weighted sums of the items with which they have interacted. Despite not training individual embeddings for each user our approach achieves state of-the-art recommendation performance with respect to transductive baselines on four real-world datasets, showing at the same time robust inductive performance. | ['Aonghus Lawlor', 'Neil Hurley', 'Barry Smyth', 'Elias Tragos', 'Khalil Muhammad', "Edoardo D'Amico"] | 2023-03-28 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [ 1.77021310e-01 4.63619381e-01 -1.54467180e-01 -3.12388450e-01
2.97605097e-01 -7.97969341e-01 8.78673434e-01 2.65307397e-01
-4.18195873e-01 4.37593788e-01 5.65251470e-01 -5.80542982e-01
-3.06223541e-01 -1.25673139e+00 -7.62238622e-01 -5.07554114e-01
-1.11589447e-01 7.22173572e-01 -7.63727427e-02 -6.55804038e-01
-6.01914562e-02 3.16708535e-01 -1.42275918e+00 3.88973922e-01
6.68271184e-01 7.30242789e-01 -2.50199586e-01 7.36772537e-01
-1.19252168e-01 4.61791009e-01 -1.15604058e-01 -8.49552751e-01
3.18856984e-01 -3.99180084e-01 -7.14905083e-01 -3.70821618e-02
2.83380955e-01 -1.48552015e-01 -4.67721134e-01 7.66255081e-01
3.07593971e-01 4.10737276e-01 6.98998153e-01 -1.07973897e+00
-1.53099954e+00 9.42433059e-01 5.78505695e-02 -7.31648356e-02
3.94080579e-01 -1.19315080e-01 1.72825909e+00 -9.40742671e-01
4.78941262e-01 8.85172963e-01 6.97934568e-01 6.38969541e-01
-1.77018440e+00 -3.07420701e-01 2.74140984e-01 -5.92593141e-02
-9.83204961e-01 -1.87128946e-01 7.11379051e-01 -4.39359039e-01
9.82704639e-01 3.33216876e-01 7.91846931e-01 1.25913048e+00
-1.68109775e-01 5.58482945e-01 5.94723225e-01 -4.25655544e-01
1.49890691e-01 3.19043696e-01 2.84462005e-01 6.90961242e-01
3.12874854e-01 1.58140182e-01 -1.53245583e-01 -2.77868330e-01
8.11031342e-01 2.12872058e-01 -3.32611471e-01 -8.34689438e-01
-1.11864245e+00 9.80288208e-01 8.41991067e-01 4.22160238e-01
-3.74998152e-01 1.07166276e-03 1.77496880e-01 4.89769280e-01
5.10298014e-01 6.09762609e-01 -4.32635516e-01 3.40930492e-01
-3.57742250e-01 2.81949997e-01 1.04736817e+00 6.40536726e-01
7.21080184e-01 -1.42188504e-01 -2.10545287e-01 6.59565389e-01
4.51836973e-01 -6.12490438e-02 5.56535900e-01 -1.19569503e-01
1.69239983e-01 9.60176528e-01 8.51657018e-02 -1.04106236e+00
-3.94569069e-01 -7.65962303e-01 -6.55260265e-01 -1.97838545e-02
4.67789143e-01 -2.14093313e-01 -1.00076067e+00 1.91008437e+00
2.87386209e-01 5.43090045e-01 -7.33478218e-02 9.45319057e-01
6.36702418e-01 4.40196306e-01 2.80499607e-01 2.54094541e-01
1.07406461e+00 -7.75961459e-01 -2.76979148e-01 -6.03443459e-02
9.25051272e-01 -3.12872380e-01 1.06514621e+00 1.55651405e-01
-7.96748757e-01 -5.72386444e-01 -9.89157140e-01 -6.06282353e-02
-7.46968269e-01 -1.25416502e-01 9.24402654e-01 6.67342722e-01
-1.21074784e+00 8.29260945e-01 -4.45633888e-01 -3.75017166e-01
2.26923093e-01 7.65749812e-01 -4.87214088e-01 4.05578874e-02
-1.49724424e+00 8.38005126e-01 4.02668893e-01 1.94250852e-01
-4.05323774e-01 -7.59948313e-01 -7.90788710e-01 4.40302104e-01
1.75039917e-01 -9.71106589e-01 9.57038939e-01 -1.36072135e+00
-1.49237537e+00 5.33396125e-01 3.54332507e-01 -5.95645070e-01
2.02763349e-01 -2.22903173e-02 -6.44284785e-01 -4.74134654e-01
-5.27491987e-01 4.78548259e-01 7.00344563e-01 -1.18508041e+00
-3.82442027e-01 -3.13648969e-01 7.14388549e-01 2.78347552e-01
-5.68588614e-01 -4.84705776e-01 -2.69329339e-01 -4.42380428e-01
-1.78863600e-01 -1.11815512e+00 -4.19836193e-01 -2.80115306e-01
-8.90557766e-02 -4.73501682e-01 4.45028484e-01 -2.19895020e-01
1.28911185e+00 -2.05879164e+00 4.75374967e-01 5.12399673e-01
4.97446626e-01 5.56330323e-01 -4.76901978e-01 6.53121591e-01
-3.12952250e-01 7.87118152e-02 4.35594916e-02 -4.44956213e-01
2.85479009e-01 3.21611345e-01 -3.54313672e-01 3.90616328e-01
1.61454543e-01 1.34813142e+00 -1.21279216e+00 9.67073366e-02
2.95742929e-01 6.50451243e-01 -1.03059757e+00 4.45746481e-01
-3.91611099e-01 2.18556717e-01 -2.62975514e-01 -1.35672256e-01
3.69581699e-01 -3.44301283e-01 5.75689971e-01 -2.80450523e-01
2.68207282e-01 5.41409910e-01 -1.04090643e+00 1.56425691e+00
-6.20420754e-01 2.34887242e-01 -3.32260132e-01 -1.18627524e+00
6.58948779e-01 4.39861089e-01 4.68370646e-01 -6.01068497e-01
1.48837641e-01 3.47163854e-03 4.39669460e-01 -1.41925141e-01
5.45512795e-01 -2.02849254e-01 -3.07188444e-02 7.01582372e-01
4.75340575e-01 5.28997362e-01 1.82718728e-02 5.09312093e-01
1.15867615e+00 2.30361328e-01 7.27405697e-02 -2.18399316e-01
4.32351798e-01 -4.86421376e-01 1.83279812e-02 6.55499518e-01
3.62451106e-01 4.89545524e-01 4.16141510e-01 -6.63702548e-01
-9.82726693e-01 -1.16727579e+00 9.60580856e-02 1.51151168e+00
-8.52051601e-02 -5.97884119e-01 -5.70549607e-01 -1.05459809e+00
8.65750462e-02 6.29396319e-01 -1.17229462e+00 -4.93712217e-01
-4.56240177e-01 -6.66051984e-01 2.15347573e-01 3.72890383e-01
-1.72641754e-01 -1.14270747e+00 4.32785191e-02 4.39052582e-01
1.68955758e-01 -7.03391373e-01 -5.35843551e-01 3.43121499e-01
-5.92026412e-01 -9.75082994e-01 -3.83371770e-01 -6.90126598e-01
9.50551867e-01 2.28459373e-01 1.42653680e+00 3.51681978e-01
3.36760767e-02 5.15055478e-01 -4.99174148e-01 -1.21742837e-01
-1.43022701e-01 2.22114474e-01 1.19943380e-01 5.21317840e-01
5.73737144e-01 -8.43844831e-01 -8.48474205e-01 1.41317442e-01
-1.12914467e+00 -9.73463582e-04 4.36379224e-01 8.56806457e-01
7.15830773e-02 -3.91373485e-01 6.99654222e-01 -1.56977010e+00
9.39780354e-01 -9.50835347e-01 -2.75689274e-01 1.45018846e-01
-8.49579394e-01 1.94046900e-01 1.12403858e+00 -6.64438903e-01
-7.45566487e-01 -4.01712721e-03 -1.39693186e-01 -1.00207746e-01
-1.10286370e-01 5.77563345e-01 -7.38941431e-02 5.14792204e-02
8.93647730e-01 1.34195602e-02 -2.19731331e-01 -5.16788185e-01
8.99762154e-01 6.37895584e-01 1.46671027e-01 -3.91116142e-01
7.89771616e-01 2.45736092e-01 -1.96193457e-01 -4.97148931e-01
-9.73212361e-01 -5.72129905e-01 -8.06485355e-01 -8.42903852e-02
6.40856326e-01 -5.64982891e-01 -6.84631407e-01 -3.78737226e-02
-7.16396332e-01 -2.79676169e-01 -4.02801216e-01 4.19411123e-01
-1.91202372e-01 1.15975149e-01 -4.19159651e-01 -4.59258974e-01
-2.41927490e-01 -7.68880010e-01 6.26940191e-01 -6.86476976e-02
-3.52575153e-01 -1.46106255e+00 3.11316550e-01 1.52883604e-01
6.40149415e-01 -8.30783099e-02 1.06676376e+00 -1.07527542e+00
-3.02459180e-01 -3.26736122e-01 -3.04694116e-01 3.30370665e-01
1.94477186e-01 -3.70749742e-01 -8.63624573e-01 -3.17456931e-01
-5.57316780e-01 -6.86008483e-03 6.49290085e-01 -6.60061510e-03
9.37044978e-01 -3.96832556e-01 -2.13437989e-01 4.47203904e-01
1.39649034e+00 -2.96949118e-01 6.84290349e-01 -8.53703022e-02
1.04772317e+00 4.77608621e-01 -1.77796438e-01 2.46752337e-01
4.10659462e-01 6.60348833e-01 5.22160530e-01 -1.26306847e-01
-2.83202957e-02 -5.61609805e-01 2.89228857e-01 5.81581473e-01
-1.75320759e-01 -6.01840198e-01 -4.61811006e-01 6.16870582e-01
-2.01249313e+00 -8.26248944e-01 -2.44454786e-01 2.41346025e+00
7.03352809e-01 2.38641128e-01 3.10419232e-01 -6.89558610e-02
3.20150346e-01 1.08224675e-01 -2.53649592e-01 -6.78274810e-01
3.15954894e-01 6.20794952e-01 4.81009066e-01 6.15632713e-01
-8.85839164e-01 9.11268711e-01 5.83874083e+00 6.69551492e-02
-1.03828025e+00 -1.16395712e-01 2.10178643e-01 -8.22479874e-02
-7.15973735e-01 5.84900156e-02 -4.99553680e-01 3.53807598e-01
1.17433572e+00 4.65170629e-02 8.33336413e-01 4.95917946e-01
-2.79714037e-02 5.34181476e-01 -1.37801945e+00 4.55149382e-01
1.05967805e-01 -1.09935641e+00 3.04539651e-01 2.48683572e-01
6.23282015e-01 7.19006285e-02 9.45512205e-02 5.90730488e-01
6.99236989e-01 -1.17979264e+00 2.68561542e-01 5.39065361e-01
3.79671097e-01 -5.50952077e-01 5.95585704e-01 2.15646431e-01
-9.65863883e-01 -7.13003054e-02 -2.40442067e-01 -3.52314413e-01
-2.30744947e-02 4.99809712e-01 -1.06521964e+00 5.22030950e-01
1.79402038e-01 6.30118549e-01 -5.57068646e-01 8.00942719e-01
-3.09427291e-01 8.39302599e-01 -2.01845199e-01 -6.32612184e-02
4.88600880e-01 -6.04870498e-01 2.49194846e-01 1.21660781e+00
2.31048793e-01 3.30084525e-02 2.10189238e-01 8.25950563e-01
-4.70245630e-01 2.90266931e-01 -8.20374608e-01 -1.81732625e-01
1.30501658e-01 1.47055089e+00 -5.16631544e-01 -1.46919146e-01
-6.19057953e-01 1.02475595e+00 8.23279262e-01 4.06933635e-01
-7.52981484e-01 -2.26003051e-01 7.36610472e-01 4.09657389e-01
5.00774026e-01 -2.51444548e-01 -3.58091518e-02 -1.17628884e+00
-1.77487597e-01 -4.95741069e-01 3.26926202e-01 -4.02972221e-01
-1.48207271e+00 4.46899027e-01 -4.02597666e-01 -9.98185277e-01
-1.85052589e-01 -7.35231578e-01 -5.55285275e-01 9.65755880e-01
-1.22946227e+00 -1.37127364e+00 3.63699086e-02 5.52080512e-01
1.16279041e-02 5.69572113e-02 1.16666710e+00 4.73568112e-01
-3.80184412e-01 7.12718308e-01 1.47064384e-02 2.56025314e-01
5.71692467e-01 -1.57415891e+00 6.02324665e-01 5.57820976e-01
7.65447080e-01 1.14853346e+00 6.40529931e-01 -3.50516766e-01
-1.46973217e+00 -1.18222260e+00 1.00998533e+00 -8.61854851e-01
8.76900971e-01 -8.23801935e-01 -9.87893105e-01 9.87818718e-01
1.68678999e-01 2.52520055e-01 1.02325153e+00 7.85952985e-01
-6.08427584e-01 1.55216485e-01 -7.49343097e-01 8.55697095e-01
1.34801924e+00 -5.26039004e-01 -6.25178218e-01 2.75952071e-01
7.34441936e-01 -8.01615044e-02 -8.95799518e-01 2.22745985e-02
7.44123399e-01 -7.94817805e-01 1.09563828e+00 -1.21465743e+00
3.28526437e-01 -1.89771339e-01 1.52167276e-01 -1.74340808e+00
-1.03533792e+00 -4.57267910e-01 -3.76774251e-01 1.06081903e+00
7.57167161e-01 -6.97319329e-01 8.83681297e-01 6.82062149e-01
-1.47386789e-01 -7.95696318e-01 -4.32634681e-01 -3.54581743e-01
-1.46321759e-01 -2.93557793e-01 6.97249234e-01 1.01720345e+00
3.02748859e-01 8.58527720e-01 -5.32884419e-01 7.00405166e-02
3.06210369e-01 6.68110624e-02 6.79576278e-01 -1.50526893e+00
-6.16170704e-01 -3.25578600e-01 -4.13979322e-01 -9.95144546e-01
2.24720329e-01 -1.36213946e+00 -1.66085318e-01 -1.57752788e+00
-1.36886239e-01 -6.02675736e-01 -8.58056366e-01 4.33115631e-01
-1.70464545e-01 5.40732384e-01 -1.24904238e-01 -7.67185539e-02
-5.31354845e-01 2.99117237e-01 1.11143327e+00 5.67846633e-02
-2.52747178e-01 2.02725139e-02 -1.18425155e+00 4.04682785e-01
6.92142785e-01 -2.81517446e-01 -7.83905387e-01 -6.04520023e-01
7.76091874e-01 -5.49717367e-01 2.45763540e-01 -7.18736529e-01
6.18498325e-02 4.55873832e-02 4.02200073e-01 2.04680413e-01
2.15855762e-01 -1.12657976e+00 2.86884606e-01 1.70616522e-01
-5.78172207e-01 -1.13959849e-01 -9.79702622e-02 8.52360070e-01
1.93137914e-01 -1.41753554e-01 4.92019027e-01 -2.47635227e-02
-4.75877464e-01 5.42842150e-01 -1.40692279e-01 -2.73788512e-01
8.57366741e-01 -8.96459818e-02 -3.01549174e-02 -2.65054941e-01
-8.73621643e-01 -7.62935728e-02 4.44049329e-01 7.37460673e-01
3.10199529e-01 -1.40391624e+00 -5.18743932e-01 4.26897258e-01
3.33543956e-01 -4.28088814e-01 8.29191580e-02 5.27674496e-01
-1.02178715e-01 3.94661725e-01 -1.03298530e-01 -4.18580696e-02
-9.76968110e-01 8.52773845e-01 2.34746128e-01 -4.40303862e-01
-4.74461854e-01 8.67821038e-01 2.04857558e-01 -8.37277830e-01
4.56523187e-02 -2.01071426e-01 -5.09770811e-01 1.37129024e-01
2.55714893e-01 -9.39423591e-02 2.68235534e-01 -6.19520187e-01
4.27110121e-02 7.41104409e-02 -3.12282741e-01 1.52243778e-01
1.47907698e+00 8.62139314e-02 -2.34526508e-02 2.82391459e-01
1.35596263e+00 1.45878419e-01 -7.68022120e-01 -4.22180653e-01
-2.88028456e-02 -4.49666709e-01 4.81200553e-02 -8.32157016e-01
-8.97502482e-01 6.91995203e-01 3.70006293e-01 7.68396378e-01
6.68190062e-01 -1.22633033e-01 8.48818898e-01 2.77486086e-01
1.55181751e-01 -6.79379046e-01 -5.11783399e-02 4.99548614e-01
4.98621196e-01 -1.14555824e+00 -1.90944776e-01 -1.78840056e-01
-3.79004419e-01 9.07524526e-01 3.98015976e-01 -5.65408170e-01
8.03862154e-01 -1.20213814e-01 -3.00248832e-01 -2.15202332e-01
-8.19780231e-01 -3.26357096e-01 6.76496506e-01 5.44334114e-01
8.93366754e-01 3.44048291e-01 -4.10664886e-01 5.86541176e-01
-3.35385837e-02 2.70165168e-02 2.58760542e-01 5.38476527e-01
-1.38971880e-01 -1.57022464e+00 3.22797924e-01 8.05552781e-01
-4.27154779e-01 -4.14909095e-01 -5.65981746e-01 4.33951944e-01
8.26278999e-02 7.41692841e-01 1.13866164e-03 -7.14366794e-01
5.50335288e-01 2.04359636e-01 4.64627087e-01 -9.24900830e-01
-9.09518123e-01 -5.59746146e-01 1.71957165e-01 -3.70772809e-01
-1.07559934e-01 -2.43235022e-01 -9.40336645e-01 -3.08305949e-01
-4.47954357e-01 4.50895846e-01 4.50741708e-01 1.00325060e+00
6.08469427e-01 5.95875859e-01 5.95454276e-01 -9.49280620e-01
-5.79118729e-01 -9.39825356e-01 -5.17399907e-01 1.01218343e+00
1.83493540e-01 -6.45160198e-01 -2.39819676e-01 -2.10170113e-02] | [10.014036178588867, 5.672024726867676] |
ae1044a3-1314-4872-a544-57e1c0dea211 | on-matrix-factorizations-in-subspace | 2106.12016 | null | https://arxiv.org/abs/2106.12016v1 | https://arxiv.org/pdf/2106.12016v1.pdf | On Matrix Factorizations in Subspace Clustering | This article explores subspace clustering algorithms using CUR decompositions, and examines the effect of various hyperparameters in these algorithms on clustering performance on two real-world benchmark datasets, the Hopkins155 motion segmentation dataset and the Yale face dataset. Extensive experiments are done for a variety of sampling methods and oversampling parameters for these datasets, and some guidelines for parameter choices are given for practical applications. | ['Keaton Hamm', 'Reeshad Arian'] | 2021-06-22 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [-2.31587991e-01 -7.53486812e-01 -5.48268318e-01 -4.52196777e-01
-8.90782058e-01 -5.28552890e-01 4.75638747e-01 -3.93384427e-01
-7.33363748e-01 4.72733885e-01 1.73173442e-01 -1.27375588e-01
-1.60371661e-01 -1.04050092e-01 1.21931210e-01 -1.15181577e+00
-7.35762060e-01 4.95674938e-01 2.52853751e-01 3.47310692e-01
2.12367490e-01 6.33635938e-01 -1.45234382e+00 -6.70866445e-02
8.08662832e-01 5.70133269e-01 -2.31230989e-01 6.21779501e-01
3.28906000e-01 1.93094641e-01 -6.46034539e-01 -1.22460738e-01
5.92232168e-01 -4.31752384e-01 -7.13556886e-01 7.47850180e-01
4.35553849e-01 -4.36883867e-02 -4.86161858e-01 1.08387041e+00
6.02722585e-01 7.98369646e-01 1.15992785e+00 -1.52453995e+00
8.13469738e-02 4.88463253e-01 -9.63277161e-01 5.41117489e-01
2.38452271e-01 2.74436653e-01 7.15141892e-01 -7.95179665e-01
8.81848633e-01 1.50941145e+00 6.10477090e-01 4.65756059e-01
-1.37593281e+00 -5.80416322e-01 -3.09851587e-01 4.86353874e-01
-1.71675014e+00 -5.24186313e-01 7.70184755e-01 -4.14532155e-01
6.81825638e-01 3.11447084e-01 4.60083604e-01 8.26319575e-01
1.88578814e-01 7.58702874e-01 1.13324678e+00 -8.71083736e-02
7.87629843e-01 -1.19148821e-01 5.70838451e-01 3.26852143e-01
4.34433281e-01 -6.03565620e-03 -3.10597926e-01 -7.36685693e-01
7.95550287e-01 -4.55642611e-01 -2.83712268e-01 -1.04221869e+00
-1.09515893e+00 1.14153671e+00 6.57520071e-02 3.17221321e-02
1.27363279e-01 1.14505373e-01 4.40195143e-01 -1.37752622e-01
2.46426448e-01 1.16965234e-01 -1.64742798e-01 -7.14027137e-02
-1.17614615e+00 3.83638024e-01 7.81089544e-01 1.18125260e+00
4.37129617e-01 1.39269948e-01 1.70356795e-01 1.09318233e+00
2.72036821e-01 4.58869040e-01 5.96422017e-01 -1.65193641e+00
2.40062416e-01 2.12121338e-01 7.17453584e-02 -9.43987250e-01
-6.85512662e-01 8.20170417e-02 -8.47714305e-01 -2.50355080e-02
5.40125072e-01 -3.40173155e-01 -1.13258767e+00 1.30119240e+00
7.63338685e-01 4.73183483e-01 -9.96420011e-02 1.06473708e+00
7.14934886e-01 4.17580247e-01 -3.51849422e-02 -5.88460326e-01
1.06223750e+00 -7.80155718e-01 -6.64869249e-01 -3.63902785e-02
4.80910867e-01 -8.05658102e-01 8.41823637e-01 4.53767806e-01
-7.56031573e-01 -3.85257125e-01 -9.48412418e-01 7.58688003e-02
-1.91620201e-01 1.87011331e-01 7.61292577e-01 9.67963398e-01
-9.78452623e-01 5.61622024e-01 -1.49813449e+00 -7.71675646e-01
3.80567729e-01 5.29252589e-01 -5.75820133e-02 -7.85819069e-02
-5.63178599e-01 4.12110567e-01 5.33138812e-01 -1.24631830e-01
-5.96475244e-01 -2.91281730e-01 -7.06205487e-01 -1.87409610e-01
-5.88797033e-02 -3.84971827e-01 8.19771230e-01 -2.22239166e-01
-1.06034565e+00 6.98782861e-01 -3.54692549e-01 -3.11728299e-01
5.06763041e-01 6.91219568e-02 -5.26695848e-01 4.30215895e-01
2.78867446e-02 1.05776882e+00 9.04403508e-01 -1.15553451e+00
-2.70252228e-01 -5.99731863e-01 -5.64428806e-01 4.18170333e-01
-3.57248664e-01 1.31786257e-01 -8.99568617e-01 -7.86157131e-01
5.42097330e-01 -1.38915133e+00 -7.39554286e-01 -5.20289838e-01
-4.88706172e-01 -2.05422953e-01 1.14349747e+00 -2.00131252e-01
1.18140090e+00 -2.33038497e+00 4.12673622e-01 5.70011973e-01
-1.11720145e-01 -3.85559827e-01 -1.41954958e-01 2.89042220e-02
-7.83012956e-02 -8.12339932e-02 -4.45313483e-01 -2.68042177e-01
-1.85079008e-01 1.99422076e-01 -4.54606637e-02 1.04238105e+00
-5.44080317e-01 3.69306147e-01 -5.99472761e-01 -8.61585736e-01
3.34823012e-01 3.28085393e-01 -6.42726183e-01 -1.91139042e-01
1.97052598e-01 3.57802689e-01 -3.20294112e-01 7.22379625e-01
5.58569551e-01 -2.99224351e-02 3.22931737e-01 4.05853987e-02
2.77002126e-01 -3.28579992e-01 -1.55240011e+00 1.64022386e+00
4.24955308e-01 7.31702805e-01 3.21858197e-01 -6.21486008e-01
5.45913160e-01 -2.91083418e-02 8.40763211e-01 1.35291398e-01
3.71906042e-01 -7.79060572e-02 8.84364620e-02 -2.64732212e-01
5.32234788e-01 1.09729446e-01 4.37204167e-02 5.09300649e-01
-1.88971847e-01 -3.65134835e-01 8.02287400e-01 4.06933129e-01
9.10647690e-01 -2.32419103e-01 -5.49142435e-02 -8.28376174e-01
4.73595351e-01 3.76229823e-01 8.01530957e-01 4.03100163e-01
-8.08273435e-01 7.54627168e-01 4.01559770e-01 -4.48983043e-01
-9.14300680e-01 -8.94822776e-01 -7.24065900e-01 8.01944613e-01
5.16844213e-01 -5.11200547e-01 -1.09416354e+00 -3.93598258e-01
1.11382276e-01 6.15689158e-01 -5.61101854e-01 6.03097267e-02
-5.77927828e-01 -1.60803211e+00 5.63367784e-01 7.96068609e-01
5.45736313e-01 -7.62303174e-01 -7.98895180e-01 -3.29723954e-01
-2.20385462e-01 -1.51491046e+00 -7.09991872e-01 -1.23374667e-02
-1.40604818e+00 -1.38435531e+00 -3.97685111e-01 -7.15351045e-01
7.28838027e-01 5.42739987e-01 8.92372429e-01 -1.88354272e-02
-5.87722123e-01 8.14227998e-01 -1.97462425e-01 1.37385428e-01
1.92235075e-02 1.96991906e-01 4.77957785e-01 -2.69132346e-01
6.45219028e-01 -2.53565282e-01 -4.97449756e-01 6.56899750e-01
-7.47674346e-01 -5.42485714e-01 3.62307355e-02 8.37539494e-01
7.09171891e-01 5.04654408e-01 -1.56027287e-01 -8.25191438e-01
6.26838923e-01 -4.30013984e-01 -7.39202142e-01 -3.35593522e-02
-3.53835195e-01 -1.53988317e-01 9.41163227e-02 -4.94985402e-01
-1.01660907e+00 4.44749624e-01 5.15006602e-01 -9.03800488e-01
-2.02664047e-01 3.59989963e-02 -2.65332073e-01 -4.47154492e-02
7.92235017e-01 -8.91236663e-02 -8.77818540e-02 -2.96156734e-01
5.46958566e-01 4.53839391e-01 7.86696136e-01 -8.08874071e-01
7.94667959e-01 8.17851365e-01 8.41923524e-03 -1.34142685e+00
4.31862511e-02 -7.83380806e-01 -1.15142083e+00 -9.46925059e-02
9.33779836e-01 -8.69430959e-01 -2.78486818e-01 4.29309785e-01
-5.94196677e-01 -2.90182054e-01 -1.61870122e-02 6.69677198e-01
-8.10159266e-01 6.27830803e-01 -6.77855372e-01 -6.13597155e-01
1.24448605e-01 -1.57490325e+00 8.72157216e-01 2.80349433e-01
-3.06345105e-01 -1.07593751e+00 8.11645091e-02 4.96753633e-01
-8.24467093e-02 2.80485511e-01 8.88544858e-01 -4.21685249e-01
-4.36119884e-01 1.27589852e-01 1.91933453e-01 -1.89991966e-02
-7.50966966e-02 3.74753147e-01 -7.54169166e-01 -8.47016692e-01
1.81792304e-02 -1.94511205e-01 9.81051445e-01 8.79418731e-01
1.27339089e+00 5.99747077e-02 -7.94028461e-01 9.24956858e-01
1.27564549e+00 4.25294697e-01 5.02697051e-01 1.77893311e-01
6.71915472e-01 3.89074892e-01 7.34681189e-01 3.70540231e-01
-4.80331853e-03 4.80509460e-01 -7.20570460e-02 2.32990608e-01
1.45330250e-01 4.24307197e-01 8.60381464e-04 7.43425846e-01
9.26571041e-02 1.01494648e-01 -1.13517308e+00 5.41457474e-01
-1.76261008e+00 -9.75463271e-01 -5.99763542e-02 1.95723236e+00
4.44353282e-01 -1.97026268e-01 6.52090788e-01 1.81869552e-01
9.69372630e-01 3.21738183e-01 -9.24853623e-01 2.86285742e-03
-1.76667944e-01 -2.67939270e-01 6.83568478e-01 4.87194806e-01
-1.67980576e+00 1.18985474e+00 8.79140091e+00 8.16694558e-01
-7.85111725e-01 -1.44301772e-01 9.83390689e-01 -6.10324740e-01
9.81885344e-02 -1.53015926e-01 -7.87648559e-01 2.31943026e-01
8.30009043e-01 -2.65366554e-01 5.03374100e-01 9.60967243e-01
2.99322993e-01 -4.24081296e-01 -1.02314675e+00 1.25472999e+00
1.07061408e-01 -1.01963890e+00 -1.88648954e-01 1.32310688e-01
9.10768092e-01 1.52245745e-01 2.15234309e-01 -1.73886821e-01
4.37800407e-01 -8.46780777e-01 6.68521747e-02 -1.61393657e-01
6.77189708e-01 -1.17285728e+00 3.33387733e-01 -2.73783542e-02
-1.02091348e+00 -1.35079443e-01 -6.52091563e-01 4.08184111e-01
-1.77995250e-01 3.72797042e-01 -6.43901348e-01 9.02355388e-02
1.06616688e+00 5.93694210e-01 -8.03919256e-01 1.25556564e+00
2.72177104e-02 7.28636086e-01 -6.27118170e-01 1.06017783e-01
2.02874616e-02 -6.51859045e-01 7.06579149e-01 1.20414078e+00
-1.46029741e-01 6.67153776e-01 1.74020678e-01 4.28036690e-01
1.06164835e-01 -5.27474470e-02 -5.07456720e-01 1.49051070e-01
1.04401839e+00 1.37831926e+00 -1.51089323e+00 -2.55080819e-01
-1.34195849e-01 5.36384165e-01 -1.26190633e-01 8.95267069e-01
-7.73149371e-01 -3.26533556e-01 1.04769814e+00 -1.19431108e-01
3.78225148e-01 -7.26516306e-01 -5.25129676e-01 -1.25010443e+00
-6.39599562e-01 -9.96107221e-01 8.67011607e-01 -3.39558303e-01
-1.18762648e+00 3.79354745e-01 6.52361512e-01 -1.12080240e+00
-4.18132335e-01 -4.76732045e-01 -6.34166956e-01 1.93053871e-01
-4.51801956e-01 -3.62318426e-01 -1.61833584e-01 7.95958281e-01
5.51690161e-01 -4.44279552e-01 2.96668619e-01 -2.72813868e-02
-1.10948026e+00 5.59297860e-01 6.29779220e-01 2.98840225e-01
6.43069804e-01 -1.06264114e+00 3.10374111e-01 1.00589001e+00
3.30865592e-01 8.71258736e-01 8.95745814e-01 -4.49260414e-01
-1.59394276e+00 -9.13104594e-01 -2.02548638e-01 -5.87936878e-01
3.34555924e-01 -2.41908476e-01 -8.32870007e-01 8.01706672e-01
3.70692551e-01 -1.59457833e-01 8.33320498e-01 2.08170235e-01
-7.58279860e-02 1.89494699e-01 -1.48418665e+00 7.06803262e-01
1.03861463e+00 -2.51055998e-03 -2.75609910e-01 4.37530130e-01
2.42992729e-01 -5.60726643e-01 -1.07884836e+00 5.80979526e-01
3.34624082e-01 -9.79425609e-01 1.17955291e+00 -4.60594267e-01
-1.70780823e-01 -5.00206292e-01 -2.10641563e-01 -1.29562938e+00
-4.52185541e-01 -5.33359706e-01 -1.16033681e-01 9.84081984e-01
-6.10827468e-02 -2.53652632e-01 1.29896116e+00 9.95962143e-01
3.45396072e-01 -6.86539412e-01 -1.20609653e+00 -7.26930737e-01
2.97525465e-01 -4.41630930e-01 4.15103257e-01 1.12763894e+00
9.96850710e-03 4.88713562e-01 1.69194877e-01 -1.98924392e-02
1.26208949e+00 2.29996473e-01 9.88417685e-01 -7.92341828e-01
1.38881743e-01 -5.74989915e-01 -5.13620973e-01 -5.91763318e-01
3.25689107e-01 -5.21820784e-01 -2.90163696e-01 -9.03335750e-01
3.69091332e-01 -2.54414529e-01 1.76118433e-01 1.07502356e-01
-2.72757739e-01 3.32450032e-01 2.35643163e-01 5.03576100e-01
-6.74716711e-01 6.02494001e-01 7.71529377e-01 -2.80828804e-01
-4.14959490e-01 -2.18727008e-01 -3.18066597e-01 9.99892294e-01
5.30482113e-01 -1.32658899e-01 -8.46319318e-01 -1.83122456e-01
-9.19682860e-01 2.81382967e-02 -4.49974462e-02 -1.30387199e+00
4.42681640e-01 -3.10030371e-01 8.37074578e-01 -1.14550042e+00
4.51290369e-01 -7.17080653e-01 1.36472285e-01 3.12295496e-01
-1.98438823e-01 2.96344817e-01 2.84918249e-01 7.35107005e-01
1.70382007e-03 1.03487447e-01 1.43564832e+00 2.12625355e-01
-9.33234155e-01 2.28958324e-01 -4.27576035e-01 4.06962931e-01
1.46552289e+00 -4.03888285e-01 8.45405161e-02 -4.01077509e-01
-8.17954600e-01 6.80812418e-01 8.05898249e-01 3.89636338e-01
5.33835351e-01 -1.40931439e+00 -5.35622478e-01 3.52969617e-01
-2.75948852e-01 -9.79291424e-02 2.79779490e-02 7.68021405e-01
-6.51541471e-01 4.14673805e-01 -1.17952213e-01 -1.13503993e+00
-1.65602672e+00 5.65167725e-01 3.51959825e-01 3.83657783e-01
-3.19703579e-01 8.31621528e-01 -1.24118827e-01 -2.20318556e-01
4.93879110e-01 -1.14471197e-01 -1.28940448e-01 8.71134847e-02
3.60114485e-01 1.17262840e+00 -4.06587929e-01 -9.99541759e-01
-6.50056124e-01 5.82834423e-01 -8.89405683e-02 -3.99186313e-01
9.00346041e-01 -1.88574746e-01 -1.15321368e-01 3.22490990e-01
1.12592185e+00 -5.00006318e-01 -1.14856100e+00 1.08883465e-02
2.16671109e-01 -7.11058199e-01 -1.44880220e-01 -3.26297879e-02
-1.33655202e+00 6.03863776e-01 7.87822366e-01 -8.12950656e-02
1.22968483e+00 -8.03527609e-02 5.81924856e-01 3.71116400e-01
7.08639562e-01 -1.51362932e+00 -8.62391144e-02 2.52251595e-01
4.43684578e-01 -1.14330614e+00 6.15914106e-01 -6.52650476e-01
-8.63583446e-01 9.19961035e-01 7.63544679e-01 -2.98581034e-01
8.52195263e-01 2.83830434e-01 3.27544063e-01 -1.59814462e-01
-6.13381445e-01 1.02162376e-01 1.96705133e-01 5.49135387e-01
3.42503220e-01 1.01799652e-01 -5.89166105e-01 1.78782754e-02
-4.54022259e-01 -4.72689629e-01 6.91483259e-01 8.22771549e-01
-2.11365297e-01 -8.97783458e-01 -9.92812514e-01 4.37978178e-01
-2.97165573e-01 4.22591418e-01 -6.57359600e-01 9.79799747e-01
-2.27769136e-01 1.12133944e+00 2.88336664e-01 -2.78730184e-01
-1.69798568e-01 4.22691107e-02 4.82087970e-01 -3.10252577e-01
-2.12219760e-01 6.40315831e-01 -9.42784846e-02 -7.48129189e-01
-7.35993624e-01 -1.47534895e+00 -1.57486808e+00 -5.93049943e-01
-3.39141428e-01 3.87516022e-01 3.74201477e-01 5.65862775e-01
2.52672642e-01 -2.92163510e-02 6.08853698e-01 -9.38227475e-01
-4.79297996e-01 -8.47734034e-01 -8.79322410e-01 5.28035343e-01
-2.62864798e-01 -8.36656630e-01 -4.83294785e-01 1.11987211e-01] | [7.675237655639648, 4.489449501037598] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.