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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
19772a8f-f17a-4555-a74c-30fb364a6723 | commu-dataset-for-combinatorial-music | 2211.09385 | null | https://arxiv.org/abs/2211.09385v1 | https://arxiv.org/pdf/2211.09385v1.pdf | ComMU: Dataset for Combinatorial Music Generation | Commercial adoption of automatic music composition requires the capability of generating diverse and high-quality music suitable for the desired context (e.g., music for romantic movies, action games, restaurants, etc.). In this paper, we introduce combinatorial music generation, a new task to create varying background music based on given conditions. Combinatorial music generation creates short samples of music with rich musical metadata, and combines them to produce a complete music. In addition, we introduce ComMU, the first symbolic music dataset consisting of short music samples and their corresponding 12 musical metadata for combinatorial music generation. Notable properties of ComMU are that (1) dataset is manually constructed by professional composers with an objective guideline that induces regularity, and (2) it has 12 musical metadata that embraces composers' intentions. Our results show that we can generate diverse high-quality music only with metadata, and that our unique metadata such as track-role and extended chord quality improves the capacity of the automatic composition. We highly recommend watching our video before reading the paper (https://pozalabs.github.io/ComMU). | ['Seon Joo Kim', 'Sharang Han', 'Kwanho Park', 'Hyeonchan Hwang', 'Minjoo Ki', 'Hyolim Kang', 'Taehyun Kim', 'Lee Hyun'] | 2022-11-17 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 1.74219683e-01 -2.98854679e-01 1.60961188e-02 4.59161662e-02
-8.00145626e-01 -9.92565930e-01 3.48503739e-01 -2.36997351e-01
9.96184424e-02 6.45577013e-01 4.40242350e-01 1.07434250e-01
-5.02304137e-01 -7.12564290e-01 -5.82140982e-01 -6.31527901e-01
-1.07134618e-01 4.78622258e-01 1.03147969e-01 -4.53803778e-01
5.01361549e-01 8.36335942e-02 -1.86158061e+00 4.52411681e-01
7.79768646e-01 8.44568312e-01 4.75818068e-01 7.38253951e-01
-2.11391211e-01 6.13049805e-01 -8.16341281e-01 -4.22301888e-01
6.27728462e-01 -9.06021833e-01 -5.15642047e-01 -1.50399193e-01
4.93428737e-01 -9.41450745e-02 4.17995006e-02 1.08774304e+00
6.57155097e-01 1.70000181e-01 4.17545378e-01 -1.17038178e+00
-2.93200761e-01 1.49802411e+00 -1.21910542e-01 -4.07838792e-01
6.23800576e-01 1.09722339e-01 1.40434289e+00 -7.53986955e-01
8.56582761e-01 6.61692441e-01 7.29868412e-01 5.22699833e-01
-1.13753009e+00 -1.10558748e+00 -2.25205719e-01 4.93543111e-02
-1.65672767e+00 -5.65796018e-01 9.71414924e-01 -4.42332506e-01
1.43883973e-01 8.56757164e-01 1.04198551e+00 1.14200699e+00
-3.31330985e-01 6.70726717e-01 7.03243434e-01 -3.86615366e-01
2.31925994e-01 -4.99528944e-01 -3.34879726e-01 1.45638987e-01
1.36395678e-01 -1.97042041e-02 -8.37037385e-01 -1.76048860e-01
9.56605971e-01 -2.80345231e-01 -2.86612570e-01 -9.64635238e-02
-1.88163352e+00 3.88144523e-01 -1.55731142e-01 4.77383018e-01
-2.13743329e-01 2.43513495e-01 3.80759895e-01 2.39997432e-01
-3.60201538e-01 1.01265037e+00 -3.38811457e-01 -5.91205895e-01
-1.33767498e+00 8.61503184e-01 8.59099925e-01 1.37320518e+00
6.29942715e-01 1.93487540e-01 -2.45953947e-01 1.03483665e+00
-6.11147769e-02 5.28294265e-01 5.82617581e-01 -1.39703906e+00
1.62968054e-01 3.10224801e-01 1.53156698e-01 -8.49492371e-01
-6.89044818e-02 -4.85664994e-01 -9.57412899e-01 1.14280619e-01
4.47250575e-01 1.08478859e-01 -2.79365510e-01 1.75683558e+00
8.89218375e-02 3.31674069e-01 -1.96606457e-01 8.59464049e-01
9.56206858e-01 4.60697979e-01 -4.85377729e-01 -3.65085155e-01
1.27787089e+00 -7.90669501e-01 -7.33967721e-01 4.32774246e-01
1.68354794e-01 -1.27803993e+00 1.56735373e+00 8.33918691e-01
-1.26384103e+00 -7.31621146e-01 -1.09951317e+00 1.85590181e-02
2.37269774e-01 2.00347960e-01 6.81306660e-01 4.52305615e-01
-6.03778899e-01 8.83189857e-01 -3.52273166e-01 1.47002473e-01
2.39595817e-03 1.95417225e-01 1.72499437e-02 4.48184997e-01
-1.06865561e+00 -6.57915650e-03 5.48777342e-01 -1.64145485e-01
-8.54123950e-01 -8.06659937e-01 -4.22401726e-01 -1.41448915e-01
4.61361110e-01 -6.69111192e-01 1.46676099e+00 -1.04347265e+00
-1.85770464e+00 6.54366255e-01 1.74639165e-01 -6.79391772e-02
5.33337772e-01 -1.83479831e-01 -4.94015098e-01 -7.70898610e-02
2.93815583e-01 4.11995620e-01 7.69943237e-01 -1.18726850e+00
-7.23070323e-01 1.41812727e-01 4.71031219e-02 1.24364235e-01
-2.20206752e-01 -1.34982693e-03 -8.30229998e-01 -1.26419055e+00
2.26816311e-01 -1.06360745e+00 -2.08692983e-01 -5.34112692e-01
-7.57158875e-01 4.20818366e-02 2.25940362e-01 -3.79145503e-01
1.94009066e+00 -2.37956953e+00 2.72015929e-01 4.96098816e-01
1.09402006e-02 -2.68048614e-01 -1.88022897e-01 5.05089819e-01
6.89328685e-02 1.38373673e-01 -1.81126550e-01 5.93935661e-02
3.43775064e-01 -1.14327110e-02 -4.16574895e-01 -1.18671231e-01
-4.49736685e-01 6.31049097e-01 -1.04396212e+00 -5.73453009e-01
-2.77914137e-01 1.56977400e-02 -9.10811126e-01 8.79510194e-02
-5.32884359e-01 7.24072337e-01 -3.94238502e-01 7.41018176e-01
3.79310668e-01 3.39302671e-04 2.30319157e-01 -3.20445985e-01
-4.28736538e-01 5.46904564e-01 -1.81534743e+00 2.26022649e+00
-1.22239597e-01 2.84700423e-01 -1.01912558e-01 1.54989697e-02
1.12849331e+00 2.97248870e-01 7.30885744e-01 -2.72627741e-01
-2.98875757e-02 5.63988984e-01 1.79698899e-01 -2.75426239e-01
1.02742851e+00 -6.92430064e-02 -4.39383626e-01 6.43508971e-01
-1.59046024e-01 -5.51396549e-01 8.40748191e-01 1.82241336e-01
9.97777879e-01 4.09946531e-01 3.06305885e-01 -2.84052044e-01
2.62309372e-01 -2.87686400e-02 8.21958780e-01 7.55630851e-01
3.14348042e-01 1.02008796e+00 3.41578752e-01 -2.24629760e-01
-1.51431966e+00 -1.22771800e+00 2.91687949e-03 1.20080519e+00
5.07352911e-02 -1.15856147e+00 -5.65309286e-01 3.50502320e-02
-2.37110093e-01 5.76814115e-01 -2.07690313e-01 2.37777427e-01
-8.53692770e-01 -3.33616108e-01 8.78551781e-01 1.88422844e-01
3.01737309e-01 -1.33183682e+00 -2.46924162e-01 3.73568177e-01
-5.74866593e-01 -6.02717102e-01 -1.13647735e+00 -1.37656122e-01
-6.90803409e-01 -1.03369248e+00 -5.92549503e-01 -9.28041279e-01
1.82271153e-01 1.04989652e-02 1.31247604e+00 7.75021166e-02
-7.40511268e-02 -5.69520518e-02 -6.74683392e-01 -2.76899517e-01
-5.72143376e-01 3.79883796e-01 3.18840921e-01 3.68624814e-02
-1.72998860e-01 -1.05854642e+00 -4.72375154e-01 4.22847211e-01
-9.83287334e-01 5.09710968e-01 3.57783735e-01 5.65938473e-01
9.90447223e-01 1.75113261e-01 5.76142013e-01 -7.02486753e-01
6.73218727e-01 -1.79023221e-02 -3.66798908e-01 1.32199824e-02
-2.10377648e-01 -1.22205289e-02 6.46241963e-01 -8.06703568e-01
-6.67616427e-01 6.04324229e-02 -2.31552646e-01 -3.24397624e-01
2.30682176e-02 4.44321543e-01 -3.56792510e-01 4.41083014e-01
9.45554674e-01 3.31373811e-01 -3.85908782e-01 -7.38805413e-01
3.99079412e-01 6.18340313e-01 9.81259823e-01 -1.07091951e+00
9.18330610e-01 5.20774312e-02 -1.18094437e-01 -5.35028636e-01
-7.08738923e-01 -3.28698546e-01 -6.77725255e-01 -3.23458046e-01
5.11791527e-01 -8.14952612e-01 -8.99739861e-01 3.17294419e-01
-7.76104808e-01 -5.06853163e-01 -7.46042252e-01 5.78401268e-01
-9.22124624e-01 3.00176591e-01 -5.96655905e-01 -6.30227268e-01
-4.01482522e-01 -8.72796416e-01 9.13396955e-01 1.15671910e-01
-7.99302161e-01 -2.65486658e-01 3.36087108e-01 2.59949893e-01
6.25818446e-02 3.30039620e-01 5.83112776e-01 -2.23867640e-01
-7.38627970e-01 1.90556929e-01 3.78208518e-01 1.10912628e-01
3.33175570e-01 4.21751589e-01 -7.20620155e-01 -5.49236219e-03
-3.54256570e-01 -1.56795979e-01 4.01077479e-01 1.30431250e-01
1.24291146e+00 -4.26874608e-01 2.61506706e-01 8.67482483e-01
1.05268192e+00 2.12678090e-01 7.86598086e-01 4.99646664e-01
7.39562571e-01 1.24763370e-01 6.54818058e-01 8.55456591e-01
1.06590979e-01 9.66415405e-01 8.96915123e-02 4.60765064e-01
-3.49227041e-01 -6.29320800e-01 5.66068828e-01 1.53631592e+00
-6.90748990e-01 2.28151724e-01 -5.94006300e-01 3.94502968e-01
-1.90987647e+00 -1.49036884e+00 -3.35659236e-01 2.34990335e+00
1.46045876e+00 1.41719989e-02 4.52743739e-01 6.13582850e-01
6.71205997e-01 -9.41495672e-02 -2.31321961e-01 -1.03100762e-03
-4.07198489e-01 5.09348929e-01 2.08800077e-01 1.92785993e-01
-9.01635647e-01 8.25927079e-01 6.14887810e+00 1.26343226e+00
-8.32276762e-01 -1.44064918e-01 -6.71428069e-02 -5.56700706e-01
-7.50104129e-01 2.87215542e-02 -7.04484403e-01 6.14593089e-01
5.20106554e-01 -3.69044721e-01 9.02608633e-01 6.87545002e-01
3.20978284e-01 3.97202760e-01 -9.66479421e-01 1.25316322e+00
-3.02250516e-02 -1.45816505e+00 3.35779160e-01 -9.27916244e-02
1.03665054e+00 -4.38514799e-01 5.40232100e-02 2.73562253e-01
2.69553304e-01 -9.19525266e-01 1.45635366e+00 7.07980454e-01
1.13919473e+00 -7.63072729e-01 1.92920893e-01 1.44766808e-01
-1.42890501e+00 1.64321929e-01 -9.38146189e-02 -1.33093566e-01
2.40889937e-01 5.96343756e-01 -5.62767565e-01 7.55902052e-01
6.13193154e-01 7.68992126e-01 -4.52632844e-01 1.27771473e+00
-3.11425745e-01 7.48918831e-01 -4.16785032e-01 1.64002907e-02
-9.58237797e-02 -4.51527268e-01 7.53287137e-01 1.01858425e+00
9.86680150e-01 3.96611281e-02 2.81599700e-01 9.66296315e-01
-1.55115694e-01 4.93137509e-01 -1.93372771e-01 -1.99449882e-01
6.71183467e-01 1.08782756e+00 -6.21066988e-01 -3.34399611e-01
1.68312475e-01 8.06450784e-01 -2.26934627e-01 7.20264614e-02
-8.14427078e-01 -4.77546781e-01 6.28224432e-01 3.11087906e-01
2.23899677e-01 -3.38360876e-01 -5.41505516e-01 -1.34003472e+00
1.08890701e-02 -1.47292924e+00 2.73342222e-01 -9.41312730e-01
-1.12804985e+00 6.02044582e-01 -2.78035641e-01 -1.85888016e+00
-3.33546191e-01 -1.62225902e-01 -5.33121884e-01 5.82739949e-01
-6.49736166e-01 -1.04132819e+00 -3.89750570e-01 6.76279902e-01
5.27151763e-01 -4.43575978e-01 9.69873667e-01 6.03351653e-01
-2.16992691e-01 6.30286753e-01 -3.49076441e-03 1.78960070e-01
1.02540636e+00 -1.20498800e+00 2.86473513e-01 5.44961393e-01
8.02651584e-01 6.65960014e-01 8.42551112e-01 -7.05616355e-01
-1.32052994e+00 -9.05244052e-01 9.42790985e-01 -5.13061404e-01
6.55979156e-01 -2.48918042e-01 -5.07183373e-01 4.09718126e-01
-1.00183882e-01 -6.79512441e-01 1.07247400e+00 2.70676792e-01
-3.75017434e-01 -1.98381633e-01 -5.30419767e-01 1.06246948e+00
1.48800170e+00 -3.53964776e-01 -5.15872478e-01 1.18357055e-01
6.94521129e-01 -6.32102549e-01 -9.97526228e-01 4.15908009e-01
1.01335728e+00 -1.02595186e+00 8.64779115e-01 -2.24943608e-01
5.46184897e-01 -1.00358570e+00 -4.31001693e-01 -1.01526093e+00
-5.74192047e-01 -1.26033175e+00 8.82217214e-02 1.39647353e+00
5.06626844e-01 2.06436098e-01 6.48147464e-01 -7.76886567e-02
-4.92977500e-01 -2.34668076e-01 -4.76791859e-01 -9.77552354e-01
-3.27775896e-01 -8.40441823e-01 1.13623679e+00 1.04587245e+00
1.23763859e-01 1.65951222e-01 -7.19364703e-01 -1.56270936e-01
4.66833055e-01 6.41227067e-01 1.29604387e+00 -1.14065433e+00
-8.65134299e-01 -8.50975275e-01 -1.64489985e-01 -9.71818149e-01
-3.15197945e-01 -1.12828577e+00 -3.21069360e-02 -1.08785665e+00
2.17285916e-01 -9.31721866e-01 -2.57649332e-01 3.63816231e-01
-3.34609710e-02 6.67994142e-01 5.31711936e-01 5.97615242e-01
-6.27029300e-01 3.69357973e-01 1.63673067e+00 -1.13710828e-01
-6.41975522e-01 2.79339254e-01 -8.05811942e-01 5.58270454e-01
8.77553046e-01 -4.36311960e-01 -2.94589221e-01 -9.17255878e-02
8.25496018e-01 -9.94972214e-02 2.10652985e-02 -1.16117013e+00
5.48428968e-02 -3.80872577e-01 2.02212203e-02 -5.33392727e-01
1.34466082e-01 -3.25433642e-01 9.94429946e-01 2.37067521e-01
-4.19611871e-01 -8.49165395e-02 -8.59280527e-02 -2.96744350e-02
-3.36215645e-01 -4.79513049e-01 4.43666041e-01 -2.02467278e-01
-5.38927674e-01 2.58521289e-01 -5.17047942e-03 2.89376199e-01
5.95863879e-01 -1.93964928e-01 1.58041328e-01 -4.00433838e-01
-9.52690542e-01 -2.93829739e-01 7.49829113e-01 3.52416873e-01
2.22885579e-01 -1.90937841e+00 -9.23180938e-01 2.16604128e-01
3.03625703e-01 -1.40783656e-02 1.37235686e-01 5.41628242e-01
-7.18606770e-01 -2.33942047e-01 -2.55874097e-01 -3.35317940e-01
-1.23354197e+00 3.89324427e-01 -2.38068819e-01 -7.77092576e-02
-7.56848872e-01 5.79784036e-01 -7.88699314e-02 -4.54924613e-01
3.07679236e-01 -4.59554225e-01 -4.30567265e-02 1.26023456e-01
7.09251583e-01 2.80072123e-01 -1.69509828e-01 -4.16199714e-01
-1.53865563e-02 5.86358488e-01 6.60068810e-01 -5.23311079e-01
1.08247459e+00 1.73767284e-01 -2.64036030e-01 8.35093737e-01
4.90144223e-01 9.27454054e-01 -9.35958982e-01 1.66300997e-01
-6.94316030e-02 -4.96938109e-01 -5.67049503e-01 -7.31980503e-01
-8.22116733e-01 2.09706396e-01 -6.77175894e-02 2.45315030e-01
1.21939099e+00 -2.02206969e-01 9.09611344e-01 3.28818470e-01
6.84343338e-01 -1.09847116e+00 2.91564334e-02 7.21432090e-01
1.15139532e+00 -4.76707399e-01 -8.90334398e-02 -2.40100294e-01
-6.96849287e-01 1.09488130e+00 1.66952521e-01 -1.17190778e-01
4.25400645e-01 4.61993605e-01 1.54921383e-01 1.79530293e-01
-6.46884024e-01 -3.05099428e-01 4.47569996e-01 4.15101916e-01
5.93987823e-01 4.17917460e-01 -5.40115476e-01 1.17771113e+00
-1.32569432e+00 4.26928811e-02 5.39562047e-01 4.33931887e-01
-4.73346591e-01 -1.58573461e+00 -5.79503834e-01 3.60898405e-01
-3.59590262e-01 -3.57333958e-01 -3.80877286e-01 3.76334906e-01
6.97342455e-01 7.00419247e-01 -9.49407443e-02 -6.76402450e-01
3.22288871e-01 6.40364960e-02 6.26171529e-01 -5.95833540e-01
-8.48000526e-01 5.50652921e-01 2.95414716e-01 -4.23480809e-01
-3.20333123e-01 -8.21939349e-01 -1.27008140e+00 -4.37183917e-01
-2.57571209e-02 4.40949500e-01 4.87201571e-01 3.42264086e-01
1.90891981e-01 6.43015563e-01 5.62665999e-01 -9.34566081e-01
-2.26005673e-01 -1.04957902e+00 -9.35836434e-01 7.42457926e-01
-2.24314794e-01 -2.84383565e-01 -2.06801400e-01 6.10260427e-01] | [15.93899917602539, 5.4590020179748535] |
743b5622-c815-4e32-a944-d3e3c5738187 | qualitative-and-quantitative-analysis-of-1 | null | null | https://openreview.net/forum?id=BxSvC2DvNH9 | https://openreview.net/pdf?id=BxSvC2DvNH9 | Qualitative and Quantitative Analysis of Diversity in Cross-document Coreference Resolution Datasets | Established cross-document coreference resolution (CDCR) datasets contain manually annotated event-centric mentions of events and entities that form coreference chains with identity relations. In this paper, we qualitatively and quantitatively compare the annotation schemes of ECB+, a CDCR dataset with identity coreference relations, and NewsWCL50, a CDCR dataset with identity, bridging, and near-identity coreference relations. The analysis shows that coreference chains of NewsWCL50 are more lexically diverse ECB+ but annotating of NewsWCL50 leads to the lower inter-coder reliability. We propose a phrasing diversity metric (PD) that encounters for the diversity of full phrases unlike the previously proposed metrics. We discuss the different tasks that both CDCR datasets create, i.e., lexical disambiguation and lexical diversity challenges for CDCR models, and propose a direction for further CDCR evaluation. | ['Anonymous'] | 2021-10-16 | null | null | null | acl-arr-october-2021-10 | ['cross-document-coreference-resolution'] | ['natural-language-processing'] | [-2.77583450e-01 4.59843516e-01 -6.16118073e-01 -1.30341113e-01
-1.04098904e+00 -1.04595625e+00 7.03141749e-01 3.74300539e-01
-4.59521532e-01 8.50231469e-01 1.05510569e+00 -1.09102562e-01
-5.05168676e-01 -4.46558863e-01 -8.45314041e-02 -2.06039265e-01
3.93245369e-02 1.17744124e+00 1.41793385e-01 -5.68829238e-01
8.79758149e-02 4.89766628e-01 -1.33297253e+00 6.48363590e-01
6.30221426e-01 4.59874511e-01 4.86282222e-02 2.69546747e-01
-1.33814976e-01 9.75341797e-01 -6.05277836e-01 -9.53295887e-01
-1.54190525e-01 -1.30507713e-02 -1.51784718e+00 -8.13993156e-01
4.66612548e-01 6.50328934e-01 -5.72636127e-01 1.04040301e+00
8.22148263e-01 3.32402110e-01 4.53515410e-01 -1.20728254e+00
-5.85889459e-01 1.63912129e+00 -6.30943000e-01 6.03488088e-01
1.12291670e+00 -5.09916246e-01 1.63069439e+00 -7.41161466e-01
1.60539722e+00 1.69746995e+00 8.71455967e-01 7.30438054e-01
-1.31274605e+00 -1.07064152e+00 2.50175018e-02 4.62585121e-01
-1.76264501e+00 -5.08578479e-01 4.53588307e-01 -3.69608194e-01
1.41251373e+00 4.70108002e-01 -7.18406439e-02 1.56725311e+00
-4.23329324e-01 4.62158501e-01 7.93867946e-01 -3.56060088e-01
-2.16235518e-01 -1.87902868e-01 7.30078042e-01 5.12883589e-02
5.11599362e-01 3.45439941e-01 -9.25118327e-01 -4.04598087e-01
2.99657583e-01 -7.44907737e-01 -6.30881011e-01 -6.46587387e-02
-9.88898754e-01 7.78752863e-01 8.42727870e-02 6.50935531e-01
-4.48860787e-02 -5.54778397e-01 6.80580020e-01 3.13104391e-01
-2.92854216e-02 8.39085817e-01 -6.23005688e-01 -8.86616036e-02
-5.00985026e-01 2.69511163e-01 1.09800446e+00 1.49603724e+00
3.79207581e-01 -7.00200558e-01 -2.72252649e-01 1.27545750e+00
7.77920932e-02 4.44388241e-01 4.61104453e-01 -1.33891940e+00
6.60318553e-01 5.44313550e-01 2.46111348e-01 -1.15461349e+00
-8.40736449e-01 -1.52691096e-01 -5.74152470e-01 -6.64592266e-01
2.80580312e-01 4.48074490e-02 9.82901733e-03 2.00209022e+00
1.51206791e-01 -9.98448431e-02 7.18911409e-01 8.03200960e-01
1.57780886e+00 2.92034864e-01 4.04666513e-01 -6.01210833e-01
1.85995519e+00 -5.61104894e-01 -1.10593045e+00 2.40624323e-02
6.57262087e-01 -9.26738858e-01 5.97745299e-01 6.62962645e-02
-1.07780075e+00 -2.04877734e-01 -7.42670298e-01 -2.58798331e-01
-1.47315785e-01 -1.75490513e-01 5.02867341e-01 1.90206259e-01
-7.24683642e-01 3.27283144e-01 -2.75870711e-01 -6.43352389e-01
-4.13066834e-01 8.98515731e-02 -7.36902177e-01 2.61247009e-01
-1.95917499e+00 1.34142256e+00 1.01001608e+00 -5.17319560e-01
-3.88335049e-01 -9.89430964e-01 -9.73347366e-01 5.91363721e-02
5.07518411e-01 -2.02247903e-01 1.23459268e+00 -7.96700269e-02
-7.50578344e-01 1.51754880e+00 -2.39539836e-02 -4.35219884e-01
1.56695589e-01 -1.30271241e-01 -1.30989063e+00 2.28316411e-02
4.26496297e-01 5.68573892e-01 -2.71708667e-01 -1.33262813e+00
-1.02998698e+00 -6.45413026e-02 9.01123732e-02 3.71617943e-01
2.36847699e-01 4.93017852e-01 -3.60032022e-01 -7.15873659e-01
8.59353393e-02 -8.08428586e-01 3.17235231e-01 -1.14660490e+00
-4.43538725e-01 -8.22709262e-01 4.34166640e-01 -4.37054276e-01
1.81880462e+00 -2.15923452e+00 1.73395947e-01 1.12092882e-01
1.58502534e-01 1.58036485e-01 -2.84923613e-01 7.19896913e-01
-7.14790404e-01 2.78178632e-01 2.99941331e-01 1.42329475e-02
3.76787260e-02 2.43032634e-01 -7.91262805e-01 1.63956374e-01
-1.22411102e-01 7.02343762e-01 -1.23944592e+00 -9.23526227e-01
-3.01107675e-01 2.10133970e-01 -4.50144082e-01 -6.51500374e-02
-5.85357882e-02 2.61978567e-01 -1.53530002e-01 4.69057500e-01
3.81475806e-01 -6.79336786e-02 9.89252090e-01 -8.76543641e-01
-4.03386891e-01 8.57473671e-01 -1.21335375e+00 1.69750679e+00
-2.77307898e-01 5.19719899e-01 -1.91639662e-02 -4.47813690e-01
1.09280264e+00 7.37009048e-01 4.44705904e-01 -7.33970582e-01
1.25895366e-01 2.83671141e-01 7.71403983e-02 -3.21660876e-01
1.08839130e+00 2.64966600e-02 -7.31711328e-01 4.08556372e-01
3.87246162e-01 4.63781863e-01 4.68916446e-01 6.67381346e-01
1.23856831e+00 -2.46854022e-01 8.83651078e-01 -5.20653725e-01
5.89398980e-01 4.08178538e-01 1.25230634e+00 5.80896795e-01
-2.38555193e-01 3.30582768e-01 4.31034267e-01 -1.72332153e-01
-7.19921231e-01 -1.12075400e+00 -5.79877794e-01 1.16002858e+00
4.74825174e-01 -8.71968746e-01 -1.81563184e-01 -6.88035429e-01
-2.55621690e-02 1.07737446e+00 -5.70575595e-01 3.79874930e-02
-9.52563584e-01 -5.85835040e-01 1.25012732e+00 5.06853402e-01
1.28658682e-01 -1.11330986e+00 -2.29814216e-01 2.49866381e-01
-1.12578166e+00 -1.37540555e+00 -7.09713221e-01 2.27548569e-01
-1.50533035e-01 -1.63024342e+00 1.00840628e-01 -9.54534531e-01
-3.42768207e-02 -6.42042235e-02 1.76034105e+00 -8.11970606e-02
-2.05486760e-01 4.23245549e-01 -8.57383668e-01 1.78700551e-01
-5.29343247e-01 3.06539267e-01 2.27640808e-01 -8.00845981e-01
1.13797832e+00 -3.98108691e-01 -7.39764422e-02 6.41385555e-01
-3.25718284e-01 -4.12131637e-01 9.03455988e-02 8.32762361e-01
5.53840578e-01 -2.04790965e-01 7.01152682e-01 -1.18491983e+00
6.98068738e-01 -6.00755572e-01 -2.44554117e-01 6.45823419e-01
-5.47493815e-01 9.27598849e-02 9.32662375e-03 -5.97207367e-01
-1.43456352e+00 -3.48840654e-01 -1.51286393e-01 -1.39759913e-01
-1.00910082e-01 3.33594859e-01 -2.52622187e-01 4.71922249e-01
1.02122247e+00 -4.83337551e-01 -8.63286257e-01 -5.77571273e-01
6.71663642e-01 8.36540282e-01 1.26160097e+00 -1.35164285e+00
3.17430556e-01 1.54840261e-01 -5.55540383e-01 -3.39834780e-01
-1.12717652e+00 -9.73351777e-01 -6.46327555e-01 1.78623468e-01
8.24364007e-01 -1.16726506e+00 -1.06299055e+00 -3.53079230e-01
-1.54294968e+00 8.86528119e-02 -2.95253485e-01 5.01909554e-01
-4.62614000e-01 3.06172818e-01 -7.89140344e-01 -4.61343616e-01
-4.29004967e-01 -6.32822812e-01 8.02971661e-01 4.21674699e-02
-1.24344432e+00 -9.55164313e-01 6.42772555e-01 3.07255030e-01
-3.22550654e-01 1.47847757e-01 9.40938354e-01 -1.30761445e+00
2.20834985e-01 2.83835292e-01 -2.51325309e-01 -6.63475990e-01
-6.58391416e-02 -9.39529836e-02 -5.70670784e-01 -1.69494256e-01
-4.32147801e-01 -3.01427960e-01 5.42902827e-01 -4.27060910e-02
1.90723255e-01 -2.79350489e-01 -9.14302051e-01 3.38093817e-01
1.21209431e+00 5.82486153e-01 5.30687928e-01 5.98916352e-01
3.95099491e-01 7.62472689e-01 9.90198553e-01 3.00526142e-01
6.80198848e-01 9.68128860e-01 -3.02836061e-01 5.54189205e-01
-3.84604067e-01 -3.25980514e-01 1.50358662e-01 1.07160604e+00
-2.04312697e-01 -1.74019396e-01 -1.17793238e+00 7.89484203e-01
-1.88074601e+00 -1.44963849e+00 -3.69159222e-01 1.68103290e+00
1.52531445e+00 -2.42941543e-01 -9.50972438e-02 -9.56110656e-02
1.27296340e+00 -2.51940284e-02 -3.88747938e-02 -4.02465910e-01
-7.34089971e-01 1.78208113e-01 2.40670651e-01 8.15496743e-01
-1.02704906e+00 1.23310113e+00 6.47884083e+00 7.23131537e-01
-2.16916010e-01 3.04036200e-01 -2.98855573e-01 -2.57808864e-02
-4.45357084e-01 6.65606633e-02 -1.41755247e+00 2.09497571e-01
7.87667334e-01 -3.10930312e-01 2.97976404e-01 6.55691385e-01
-4.02285933e-01 2.37590477e-01 -1.39777184e+00 1.05553782e+00
-5.33757405e-03 -1.24453461e+00 -7.44291488e-03 -3.46959859e-01
6.91486895e-01 4.62990068e-02 -4.90355790e-01 4.25813109e-01
1.35286272e+00 -7.30044782e-01 6.42843485e-01 1.89804748e-01
9.96645987e-01 -7.89022684e-01 9.16158020e-01 -1.02459997e-01
-1.54949558e+00 -6.78010806e-02 -3.53982896e-01 4.75190729e-01
4.36486721e-01 3.23395580e-01 -2.19205126e-01 8.94296467e-01
1.02857280e+00 6.02345824e-01 -3.29433352e-01 5.03376544e-01
-4.05912906e-01 1.62478477e-01 3.17645818e-02 4.30918574e-01
-3.45034599e-01 2.09431425e-01 9.47787702e-01 1.72928357e+00
1.63022861e-01 6.70044959e-01 6.78253844e-02 6.48909926e-01
-2.58095711e-01 3.23735364e-02 -2.96995193e-01 1.36106074e-01
1.77433264e+00 1.23256636e+00 -3.71695608e-01 -1.39436856e-01
-4.31525260e-01 4.67730522e-01 7.61260152e-01 4.31183092e-02
-7.24584103e-01 -3.99026841e-01 1.00114942e+00 -3.20611715e-01
1.37697794e-02 3.96300703e-01 -3.32199200e-03 -1.32471693e+00
-6.70780241e-01 -1.22246039e+00 1.38624203e+00 -6.31125689e-01
-1.73073936e+00 7.83661127e-01 3.05503249e-01 -9.41735923e-01
-3.04667622e-01 -2.22979218e-01 -2.23759949e-01 5.43749571e-01
-1.15403652e+00 -8.98832679e-01 -1.25472471e-01 7.71343946e-01
1.19071126e-01 -2.24015221e-01 1.10608745e+00 6.31827712e-01
-4.54930067e-01 9.09213781e-01 -5.06292224e-01 7.48543799e-01
1.27576005e+00 -1.26581836e+00 3.16820294e-01 5.24882317e-01
1.44194767e-01 1.14658225e+00 1.03585625e+00 -8.84004414e-01
-6.47313058e-01 -9.60286379e-01 1.49732590e+00 -7.15667427e-01
9.06640828e-01 -8.50461074e-04 -9.33465004e-01 1.10420275e+00
4.28878486e-01 -4.67483610e-01 1.03972721e+00 8.94064844e-01
-1.11965847e+00 2.50577539e-01 -1.18762088e+00 5.70538521e-01
1.73580635e+00 -8.45673442e-01 -1.69248331e+00 -1.15805604e-01
9.31595922e-01 -6.15321338e-01 -1.52997577e+00 6.75891876e-01
2.40866631e-01 -5.59557319e-01 1.13276815e+00 -7.49671817e-01
-5.48400357e-02 -3.40169460e-01 -5.12847722e-01 -1.27570009e+00
-5.96430123e-01 -5.74617207e-01 -1.61547139e-01 2.05868745e+00
5.50337732e-01 -3.44990492e-01 4.68392596e-02 5.02096117e-01
-2.72785217e-01 3.27117085e-01 -9.68306541e-01 -8.33082378e-01
1.24257907e-01 -3.20178837e-01 7.06270456e-01 1.81051445e+00
1.00008166e+00 6.93930864e-01 -5.96559886e-03 3.69362205e-01
5.51669419e-01 5.25214434e-01 5.55782877e-02 -1.72509134e+00
6.75474703e-02 -4.73532706e-01 -2.52161492e-02 -4.58269626e-01
6.86566353e-01 -1.19656575e+00 -1.05001479e-01 -1.00800133e+00
5.12906432e-01 -5.72799504e-01 -8.71757865e-02 6.63045764e-01
-1.08189300e-01 -6.60337582e-02 6.84736371e-02 4.94330138e-01
-9.56787765e-01 2.80252341e-02 7.08527684e-01 -5.83276451e-02
-2.86040306e-01 -7.56837010e-01 -9.21993554e-01 7.40172684e-01
3.85756373e-01 -6.79215372e-01 -7.58069903e-02 -1.79495126e-01
5.59186697e-01 2.63894945e-01 -1.87484041e-01 -4.45926368e-01
5.40920138e-01 -3.73291299e-02 -9.29856822e-02 -6.62541866e-01
-1.14536591e-01 -5.42437613e-01 6.66714847e-01 1.04336977e-01
-7.23669708e-01 2.21545026e-01 5.40589504e-02 2.38382995e-01
-4.01931763e-01 -3.05549383e-01 6.97310090e-01 -8.05777758e-02
-9.00914133e-01 -1.79274261e-01 -1.55914709e-01 1.06242609e+00
6.98200107e-01 3.32422525e-01 -1.02856278e+00 5.14650578e-03
-1.06201959e+00 4.77438956e-01 3.78304780e-01 8.57954860e-01
1.46583378e-01 -1.56970739e+00 -7.95732319e-01 -5.39065659e-01
5.59295833e-01 -3.28581423e-01 1.23427324e-01 5.78564703e-01
5.75483143e-02 6.75728679e-01 -7.46081024e-02 -5.74941039e-02
-1.47549164e+00 8.41641366e-01 1.55506849e-01 -5.95785975e-01
-5.42386949e-01 6.62434161e-01 -4.34923545e-02 -7.36387134e-01
4.11844105e-01 1.22137427e-01 -8.65919411e-01 6.35723650e-01
6.19789064e-01 5.89101553e-01 -7.48643279e-02 -1.21888983e+00
-7.91704655e-01 4.61493522e-01 -2.47795567e-01 -8.59291777e-02
9.11190093e-01 -4.18535292e-01 -3.14263523e-01 1.42916173e-01
7.83325374e-01 4.16439682e-01 -3.98838103e-01 -5.23418546e-01
8.43106747e-01 -7.49339610e-02 -2.88923144e-01 -1.14571977e+00
-7.95662224e-01 2.57849991e-01 2.01470807e-01 1.50790453e-01
6.82376087e-01 6.56510472e-01 4.31827068e-01 3.48343462e-01
6.72644734e-01 -1.16514087e+00 -3.78108799e-01 9.58490789e-01
1.01886451e+00 -8.40031326e-01 -1.84192836e-01 -7.07062244e-01
-9.47068989e-01 7.45216489e-01 7.13352442e-01 2.70949125e-01
4.43599612e-01 6.59757495e-01 1.09274410e-01 -4.19260353e-01
-9.09313977e-01 -4.76448745e-01 1.90484360e-01 7.39555299e-01
7.89687276e-01 1.16296604e-01 -9.61252928e-01 1.39139366e+00
-5.96735597e-01 -4.42440122e-01 2.11460337e-01 2.14575708e-01
-7.13543221e-02 -1.28371608e+00 -2.66681135e-01 -1.62202179e-01
-6.57316267e-01 -3.01070690e-01 -6.45249248e-01 1.09816086e+00
1.84157878e-01 1.17418873e+00 3.10110807e-01 -4.54114646e-01
6.41317129e-01 1.31388128e-01 3.40833604e-01 -6.76462770e-01
-7.83373356e-01 -2.60666460e-01 9.17772889e-01 -6.66523516e-01
-7.44153440e-01 -8.50728333e-01 -1.56091249e+00 -5.12208641e-01
-4.85770255e-01 6.95390761e-01 -1.95815086e-01 7.95168400e-01
2.84023523e-01 2.06327006e-01 1.81980401e-01 -1.09488703e-01
-1.50934532e-01 -1.01905477e+00 -5.34177542e-01 9.89274442e-01
-2.51048625e-01 -1.01644528e+00 -3.85955483e-01 -1.86575189e-01] | [9.318414688110352, 9.532918930053711] |
3167450d-3b3a-4f00-9139-9019c0bf2cfa | exploring-continual-learning-for-code | 2307.02435 | null | https://arxiv.org/abs/2307.02435v1 | https://arxiv.org/pdf/2307.02435v1.pdf | Exploring Continual Learning for Code Generation Models | Large-scale code generation models such as Codex and CodeT5 have achieved impressive performance. However, libraries are upgraded or deprecated very frequently and re-training large-scale language models is computationally expensive. Therefore, Continual Learning (CL) is an important aspect that remains underexplored in the code domain. In this paper, we introduce a benchmark called CodeTask-CL that covers a wide range of tasks, including code generation, translation, summarization, and refinement, with different input and output programming languages. Next, on our CodeTask-CL benchmark, we compare popular CL techniques from NLP and Vision domains. We find that effective methods like Prompt Pooling (PP) suffer from catastrophic forgetting due to the unstable training of the prompt selection mechanism caused by stark distribution shifts in coding tasks. We address this issue with our proposed method, Prompt Pooling with Teacher Forcing (PP-TF), that stabilizes training by enforcing constraints on the prompt selection mechanism and leads to a 21.54% improvement over Prompt Pooling. Along with the benchmark, we establish a training pipeline that can be used for CL on code models, which we believe can motivate further development of CL methods for code models. Our code is available at https://github.com/amazon-science/codetaskcl-pptf | ['Bing Xiang', 'Mohit Bansal', 'Murali Krishna Ramanathan', 'Ramesh Nallapati', 'Parminder Bhatia', 'Xiaofei Ma', 'Ming Tan', 'Dejiao Zhang', 'Xiaopeng Li', 'Hantian Ding', 'Qing Sun', 'Prateek Yadav'] | 2023-07-05 | null | null | null | null | ['code-generation', 'continual-learning'] | ['computer-code', 'methodology'] | [-7.66206160e-02 -1.83472320e-01 -2.83114046e-01 -1.99087203e-01
-9.65653956e-01 -6.15505219e-01 5.66564977e-01 4.24608141e-02
-1.66516796e-01 5.92513263e-01 2.43273467e-01 -5.66736698e-01
3.83499712e-01 -3.68603796e-01 -9.42370117e-01 -3.84834468e-01
2.26812765e-01 4.63869143e-03 2.54230440e-01 2.88202595e-02
4.77513522e-01 2.31326111e-02 -1.36985135e+00 6.30518079e-01
1.13966846e+00 4.14762884e-01 7.53354192e-01 6.33340538e-01
-5.17398655e-01 1.17152596e+00 -5.95424652e-01 -4.41374391e-01
-3.30049805e-02 -3.71653974e-01 -1.00204432e+00 -3.68209779e-01
3.98639172e-01 -6.75557926e-02 -1.26796514e-01 1.09869564e+00
6.58156693e-01 -4.27466556e-02 3.18753660e-01 -1.28811514e+00
-1.14334941e+00 9.95307624e-01 -6.29458010e-01 6.23775050e-02
2.21445411e-01 3.41006756e-01 9.64004278e-01 -1.34345055e+00
6.77705467e-01 1.04730773e+00 8.30738842e-01 8.95550072e-01
-1.46660459e+00 -7.94146955e-01 4.22052667e-02 -6.41395822e-02
-1.49686229e+00 -5.72377205e-01 5.94172120e-01 -7.94082642e-01
1.42939675e+00 7.95241594e-02 3.03410172e-01 1.29895127e+00
4.27917421e-01 1.03918242e+00 8.21741760e-01 -3.15086871e-01
5.00360057e-02 3.08551848e-01 1.25272602e-01 8.99285555e-01
2.43360903e-02 -1.05864532e-01 -6.95555151e-01 -3.03791463e-01
4.58703399e-01 -4.51656170e-02 -4.14715886e-01 -8.07305053e-02
-1.27075291e+00 7.74609745e-01 3.29283804e-01 2.33615533e-01
-8.25819001e-02 4.25655097e-01 5.07025898e-01 3.16339135e-01
5.50320566e-01 8.91499400e-01 -8.71402025e-01 -4.90132928e-01
-1.14967918e+00 4.95857418e-01 6.69531465e-01 1.40360558e+00
8.36430073e-01 3.51956822e-02 -5.94015419e-01 1.00410426e+00
1.52664050e-01 3.01985890e-01 8.07906330e-01 -7.76364386e-01
5.93522906e-01 6.50353074e-01 -1.87629864e-01 -5.52143633e-01
-2.98149109e-01 -3.78103316e-01 -6.85057938e-01 -4.74707559e-02
1.32654890e-01 -3.01399350e-01 -6.45568848e-01 1.81909347e+00
-2.27195114e-01 1.81104392e-01 -8.75748321e-02 3.85533214e-01
9.92038727e-01 7.27139533e-01 1.62017092e-01 -5.38437255e-02
1.10694885e+00 -1.47294879e+00 -4.80952740e-01 -2.80354857e-01
7.98655510e-01 -1.09876800e+00 1.58588517e+00 2.99556702e-01
-1.00096345e+00 -6.46608293e-01 -7.31415212e-01 -2.95133352e-01
-2.16925129e-01 3.51041853e-01 7.24654675e-01 1.96471453e-01
-1.34346974e+00 4.59876418e-01 -9.61936772e-01 -4.46779966e-01
4.31111485e-01 9.84928682e-02 -7.89611340e-02 -1.99061800e-02
-7.68550336e-01 7.09078014e-01 3.48468751e-01 -3.27828377e-01
-1.01601291e+00 -1.10888016e+00 -7.65621364e-01 2.02417701e-01
2.24007040e-01 -6.07383370e-01 1.60632777e+00 -7.54452705e-01
-1.46834493e+00 7.35545814e-01 -3.52220744e-01 -3.76396447e-01
4.34236497e-01 -4.00100678e-01 -2.99530532e-02 -4.74412143e-01
1.15414687e-01 8.61902535e-01 9.04042363e-01 -1.09543860e+00
-3.86810750e-01 2.09064513e-01 1.15388446e-02 -1.56233504e-01
-3.90903503e-01 3.21899831e-01 -5.26042521e-01 -7.96815097e-01
-3.48426640e-01 -9.28987384e-01 -3.75939608e-02 -2.43839294e-01
-4.01604891e-01 -4.92107034e-01 4.60911274e-01 -5.59792876e-01
1.70552206e+00 -2.40059495e+00 2.33008623e-01 -5.19540310e-01
1.72286302e-01 1.68217540e-01 -4.52330232e-01 5.28997600e-01
-8.87185112e-02 4.05723959e-01 -4.12174016e-01 -6.42788947e-01
-2.12515276e-02 -2.57689785e-02 -6.58886373e-01 1.08460225e-01
3.91647398e-01 1.12389529e+00 -9.22971964e-01 -4.72494215e-01
-2.98614770e-01 2.47940570e-01 -1.02881968e+00 4.10314143e-01
-5.36301315e-01 2.72269398e-01 -2.66210824e-01 7.57721961e-01
5.26786566e-01 -5.95239580e-01 -2.63178021e-01 3.24126065e-01
-5.14143646e-01 3.62480491e-01 -4.25947756e-01 2.26596093e+00
-7.33081758e-01 7.15877891e-01 -2.22718030e-01 -4.49279964e-01
8.50639462e-01 2.78931051e-01 1.16084861e-02 -4.91201431e-01
-1.93719655e-01 2.84592360e-01 -8.65571350e-02 -6.76891744e-01
7.06982732e-01 2.85226941e-01 -1.68579295e-01 6.18930638e-01
3.87076825e-01 -4.77668434e-01 2.64231533e-01 4.15558428e-01
1.40410864e+00 3.35602760e-01 1.23435944e-01 -4.91289616e-01
3.40802938e-01 7.32450746e-03 3.55657130e-01 8.88939679e-01
-9.99679118e-02 7.81613410e-01 5.44814885e-01 -3.18007410e-01
-9.93357539e-01 -7.46611416e-01 -1.58791542e-01 1.49121499e+00
-2.48347133e-01 -9.04323578e-01 -6.91518843e-01 -7.58731842e-01
-8.73498544e-02 8.09959412e-01 -4.64796960e-01 -2.75560260e-01
-5.92316926e-01 -9.07345533e-01 7.20872223e-01 4.61421877e-01
2.78713763e-01 -1.37375724e+00 -4.05746996e-01 2.52150595e-01
-2.40528390e-01 -6.91946745e-01 -7.53936350e-01 3.96643996e-01
-7.14706957e-01 -8.05495858e-01 -7.36942947e-01 -8.96511137e-01
8.03348660e-01 2.34312877e-01 1.46104372e+00 4.70383167e-01
-1.35416046e-01 1.18761964e-01 -3.67794871e-01 -3.38868886e-01
-6.82831585e-01 7.18810141e-01 -2.61724651e-01 -5.91105044e-01
2.61648029e-01 -5.18630743e-01 -3.08947563e-01 -6.83933720e-02
-7.91078806e-01 4.18650568e-01 7.08985746e-01 1.03981769e+00
3.19936484e-01 -3.59208912e-01 5.32801747e-01 -1.03373981e+00
9.89394248e-01 -7.69981980e-01 -8.05359423e-01 4.01512593e-01
-6.67355239e-01 3.72114062e-01 7.73862362e-01 -5.72948456e-01
-1.22422683e+00 3.97390127e-02 -2.91830331e-01 -3.10505956e-01
1.06558047e-01 6.85955822e-01 2.36927167e-01 3.51929143e-02
9.83792901e-01 4.47650313e-01 -3.05937350e-01 -6.14288747e-01
3.17971498e-01 6.23531640e-01 3.85166436e-01 -1.05226898e+00
6.36499643e-01 -1.40280798e-02 -6.16548240e-01 -4.78476852e-01
-7.86985040e-01 -1.89855337e-01 -3.77561063e-01 1.79815814e-01
6.52576268e-01 -1.13829613e+00 -2.46508002e-01 5.58296621e-01
-1.64435303e+00 -8.26602936e-01 -8.10201690e-02 1.47809967e-01
-4.25607800e-01 9.35580209e-02 -8.86498332e-01 -3.96070987e-01
-4.99150097e-01 -1.42422593e+00 1.07362175e+00 2.55993128e-01
-4.09736753e-01 -8.47404838e-01 2.77878791e-01 4.83369455e-02
8.38479280e-01 -1.66053712e-01 1.11557472e+00 -3.96975309e-01
-7.69720793e-01 2.27573097e-01 -3.34897459e-01 3.93012673e-01
-3.96517999e-02 3.80639166e-01 -9.04912293e-01 -4.22954381e-01
-1.31051496e-01 -6.30312800e-01 9.87604320e-01 1.04170986e-01
1.60705817e+00 -2.43620977e-01 -2.39799052e-01 8.57529581e-01
1.35271597e+00 -4.76659052e-02 3.90513897e-01 1.91322401e-01
7.42656231e-01 2.01227456e-01 2.49466643e-01 5.60633242e-01
4.98287171e-01 5.09167850e-01 1.60109803e-01 3.56534161e-02
-4.08358037e-01 -4.15005922e-01 7.16298640e-01 1.25659025e+00
2.24183828e-01 -6.84455708e-02 -1.21338642e+00 6.07869804e-01
-1.90747094e+00 -6.06876373e-01 -1.62496477e-01 2.03995919e+00
1.48535991e+00 7.49173537e-02 -3.50556463e-01 -5.39453328e-01
4.20097142e-01 2.39955187e-02 -5.81706643e-01 -3.97069514e-01
1.14969321e-01 2.38191649e-01 2.38136426e-01 5.08991957e-01
-8.13424647e-01 1.25100434e+00 6.19509649e+00 1.00196183e+00
-1.31102109e+00 4.89038944e-01 5.15464902e-01 -3.12646553e-02
-4.28617030e-01 1.30405769e-01 -9.61728036e-01 7.12610066e-01
9.42066550e-01 -5.30662417e-01 7.39607453e-01 1.17099047e+00
-3.30958851e-02 -1.38562135e-02 -1.30053937e+00 9.39574122e-01
9.27351937e-02 -1.30157816e+00 -2.67970730e-02 -3.47652495e-01
1.15081155e+00 5.97804487e-01 2.66897846e-02 9.17224586e-01
6.40064597e-01 -8.87309194e-01 9.11845207e-01 3.41242522e-01
9.47996140e-01 -4.08402115e-01 5.24828434e-01 5.02779245e-01
-1.10152757e+00 -1.16352998e-01 -5.57459652e-01 -1.05793685e-01
-1.07604675e-01 6.31508052e-01 -7.87944257e-01 1.83385044e-01
6.73398077e-01 8.47664475e-01 -1.02930129e+00 1.10159147e+00
-5.31985581e-01 7.45971024e-01 1.15746364e-01 4.74336483e-02
1.82567891e-02 3.15294892e-01 1.94743797e-01 1.39403093e+00
6.60149217e-01 -4.07809436e-01 1.57281071e-01 1.47351301e+00
-4.14436638e-01 1.04740210e-01 -5.47856152e-01 -1.72055557e-01
5.29518545e-01 1.24011934e+00 -5.12344718e-01 -3.52820396e-01
-6.23054564e-01 9.11512494e-01 7.22339451e-01 4.45750028e-01
-1.01735115e+00 -4.02899951e-01 6.07136190e-01 -9.54746827e-02
3.99863720e-02 -2.24851757e-01 -3.09848011e-01 -1.50012267e+00
2.14916021e-02 -8.84949982e-01 -4.07527722e-02 -7.35597610e-01
-1.16416335e+00 8.14024389e-01 -1.13296853e-02 -9.98522162e-01
-1.03896946e-01 -3.48544449e-01 -7.70748079e-01 8.53910208e-01
-1.59507155e+00 -1.05731320e+00 -2.33768642e-01 3.83661091e-01
1.01635242e+00 -2.56418407e-01 7.45193422e-01 4.44452763e-01
-5.92570662e-01 6.79457247e-01 3.41421887e-02 -8.21197107e-02
9.05126512e-01 -1.32864654e+00 8.33726346e-01 9.12434101e-01
-5.19620953e-03 1.06169641e+00 5.53519726e-01 -6.39594078e-01
-1.35370409e+00 -1.38187933e+00 1.19038153e+00 -8.15485060e-01
7.98167169e-01 -8.02001178e-01 -1.09984446e+00 7.13082969e-01
4.81985301e-01 5.51719330e-02 5.22343755e-01 9.69898254e-02
-4.23375160e-01 2.01807037e-01 -7.50900924e-01 5.44525683e-01
1.09550464e+00 -6.26030624e-01 -4.27364439e-01 5.07689476e-01
1.18965268e+00 -5.10236740e-01 -4.98022348e-01 1.17545605e-01
3.20858121e-01 -7.33651280e-01 5.81817806e-01 -4.92822617e-01
8.86967003e-01 -2.52941638e-01 -1.10178636e-02 -1.40902615e+00
-3.92817587e-01 -8.05050790e-01 -1.99442610e-01 1.52500272e+00
6.47199810e-01 -5.14836073e-01 2.97651976e-01 4.43117559e-01
-4.59130943e-01 -8.35804224e-01 -6.04520440e-01 -7.23195612e-01
4.36669946e-01 -4.06362444e-01 6.28199041e-01 1.00082469e+00
1.41270518e-01 3.40043306e-01 -4.11298454e-01 -2.39489749e-01
1.59008235e-01 1.42182618e-01 8.98425400e-01 -9.14215446e-01
-6.14123702e-01 -5.95860004e-01 4.83786345e-01 -1.21137488e+00
4.86544341e-01 -1.26038384e+00 3.89090329e-01 -1.27183759e+00
5.34430027e-01 -4.55599755e-01 -1.43852755e-01 8.35425615e-01
-3.92333925e-01 -4.95658927e-02 2.38825902e-01 4.56220180e-01
-7.79799879e-01 6.41590238e-01 1.03717422e+00 -1.92561820e-01
-1.81805477e-01 -1.16234541e-01 -7.62629986e-01 5.30123651e-01
8.68514240e-01 -7.01386511e-01 -3.82686347e-01 -8.66559386e-01
6.04931533e-01 -1.76462263e-01 -5.71113005e-02 -8.92927468e-01
2.66284674e-01 -6.42960817e-02 -1.06110340e-02 -2.10412726e-01
-2.71597087e-01 -3.94746542e-01 -1.57018304e-02 4.80711430e-01
-5.64630032e-01 5.06110847e-01 4.70715940e-01 3.92248519e-02
-1.79840058e-01 -4.56198066e-01 7.07760155e-01 -3.76003355e-01
-6.05419695e-01 2.33576328e-01 -4.13206637e-01 2.86540121e-01
6.76183105e-01 4.53094542e-01 -7.62594044e-01 7.29697123e-02
-2.57306755e-01 3.88075501e-01 6.55321240e-01 8.17251563e-01
4.68875498e-01 -1.26376045e+00 -8.33571494e-01 3.14439446e-01
4.39230174e-01 7.86585510e-02 1.46882400e-01 8.71519208e-01
-4.93703723e-01 3.99134159e-01 6.97261542e-02 -4.78151739e-01
-1.06951690e+00 5.89079559e-01 2.77466662e-02 -4.44225788e-01
-4.24118578e-01 1.25542796e+00 3.07048082e-01 -7.47967422e-01
1.97517350e-01 -6.63350999e-01 1.50897160e-01 -1.95088029e-01
5.36325395e-01 -4.76971836e-05 1.23319104e-01 -2.19767690e-01
-1.66829631e-01 2.15319157e-01 -3.59491020e-01 3.18458140e-01
1.33570445e+00 8.26216266e-02 -5.48731387e-01 5.02937615e-01
1.07024550e+00 1.35814890e-01 -1.37759352e+00 -2.75659144e-01
1.35207012e-01 -4.69431996e-01 -1.71870381e-01 -9.56709802e-01
-9.16365802e-01 9.06027257e-01 3.10669899e-01 8.82071629e-02
9.27233756e-01 1.68533966e-01 6.69800341e-01 3.77800792e-01
5.47835469e-01 -7.45962441e-01 2.78503746e-01 9.78371620e-01
1.12020910e+00 -1.34363067e+00 -2.01546177e-01 8.29469785e-02
-5.19724667e-01 9.85565662e-01 1.04811513e+00 6.80178478e-02
4.64098126e-01 5.95787406e-01 -6.85348660e-02 -3.73862460e-02
-1.31865215e+00 2.03852341e-01 4.79304343e-02 2.39777401e-01
9.73852098e-01 -1.38806075e-01 -3.42962444e-01 7.47047842e-01
-2.86444038e-01 2.16659993e-01 6.28534555e-01 1.03867817e+00
-2.13736758e-01 -1.21489644e+00 -1.32194489e-01 4.32216436e-01
-5.03236771e-01 -7.32738853e-01 -1.66554794e-01 4.29066479e-01
2.79801309e-01 5.73388994e-01 -1.24015808e-01 -2.77387708e-01
8.61010477e-02 3.52759480e-01 3.11160743e-01 -1.17346680e+00
-8.57445896e-01 -1.97984368e-01 -4.02529895e-01 -4.87262070e-01
2.53444575e-02 -5.54769278e-01 -1.20080447e+00 -2.72770911e-01
-3.59956980e-01 8.24086964e-02 5.37453592e-01 5.60643613e-01
6.95075691e-01 7.34317780e-01 1.80315316e-01 -7.68102229e-01
-5.71109235e-01 -1.19250381e+00 -1.60846964e-01 2.46170610e-01
3.92692894e-01 -4.59844232e-01 -3.83471578e-01 3.87804329e-01] | [7.72246789932251, 7.894862651824951] |
238adb35-0099-4455-ae7a-24f0caada33a | epasad-ellipsoid-decision-boundary-based | 2204.04154 | null | https://arxiv.org/abs/2204.04154v1 | https://arxiv.org/pdf/2204.04154v1.pdf | EPASAD: Ellipsoid decision boundary based Process-Aware Stealthy Attack Detector | Due to the importance of Critical Infrastructure (CI) in a nation's economy, they have been lucrative targets for cyber attackers. These critical infrastructures are usually Cyber-Physical Systems (CPS) such as power grids, water, and sewage treatment facilities, oil and gas pipelines, etc. In recent times, these systems have suffered from cyber attacks numerous times. Researchers have been developing cyber security solutions for CIs to avoid lasting damages. According to standard frameworks, cyber security based on identification, protection, detection, response, and recovery are at the core of these research. Detection of an ongoing attack that escapes standard protection such as firewall, anti-virus, and host/network intrusion detection has gained importance as such attacks eventually affect the physical dynamics of the system. Therefore, anomaly detection in physical dynamics proves an effective means to implement defense-in-depth. PASAD is one example of anomaly detection in the sensor/actuator data, representing such systems' physical dynamics. We present EPASAD, which improves the detection technique used in PASAD to detect these micro-stealthy attacks, as our experiments show that PASAD's spherical boundary-based detection fails to detect. Our method EPASAD overcomes this by using Ellipsoid boundaries, thereby tightening the boundaries in various dimensions, whereas a spherical boundary treats all dimensions equally. We validate EPASAD using the dataset produced by the TE-process simulator and the C-town datasets. The results show that EPASAD improves PASAD's average recall by 5.8% and 9.5% for the two datasets, respectively. | ['Sandeep Kumar Shukla', 'Saurabh Kumar', 'Rachit Agarwal', 'Vikas Maurya'] | 2022-04-08 | null | null | null | null | ['network-intrusion-detection'] | ['miscellaneous'] | [-2.16035128e-01 -2.62414306e-01 1.84633955e-01 4.94768232e-01
2.76940409e-02 -9.20669436e-01 8.74077201e-01 4.18500513e-01
-6.15215823e-02 4.64700401e-01 -3.24161828e-01 -6.96052849e-01
-3.74974638e-01 -9.94963765e-01 -1.43160447e-01 -8.61589432e-01
-6.55345738e-01 1.01524957e-01 7.81089365e-01 -1.19534738e-01
5.08617282e-01 9.71094549e-01 -1.08518076e+00 -2.39206910e-01
3.14137131e-01 1.16619980e+00 -4.83391613e-01 5.18760741e-01
1.93521619e-01 1.45429850e-01 -1.09195960e+00 1.31821275e-01
3.85766864e-01 1.32943630e-01 -5.10497391e-01 -4.24333245e-01
-5.38029432e-01 -1.69950098e-01 -2.95449942e-01 1.28760827e+00
1.23854913e-01 -5.88487871e-02 5.81932724e-01 -2.02158546e+00
-2.70014312e-02 3.05519581e-01 -7.00380921e-01 3.95723313e-01
4.41202283e-01 3.43584299e-01 3.95193577e-01 -5.77567577e-01
1.27493337e-01 1.33178222e+00 4.49483246e-01 4.57840025e-01
-9.42388892e-01 -9.63921249e-01 -5.26934266e-02 6.39456734e-02
-1.34067512e+00 1.83033019e-01 5.95407188e-01 -3.47587138e-01
1.08798909e+00 4.55361366e-01 3.69473249e-01 1.23770034e+00
9.53224182e-01 3.56427431e-01 8.92007172e-01 -6.39342442e-02
6.76534176e-01 -2.06722364e-01 4.27698761e-01 6.98322654e-02
8.80888522e-01 5.04677355e-01 2.49742910e-01 -7.31720626e-01
6.40739858e-01 1.35859519e-01 -1.93773270e-01 2.36536339e-02
-8.94876301e-01 5.14033198e-01 -7.08061680e-02 4.83167559e-01
-5.22133172e-01 -1.01446740e-01 7.70433784e-01 3.27870369e-01
3.15960422e-02 6.37610197e-01 -6.76487505e-01 -3.01791012e-01
-3.27752717e-02 -5.19818813e-02 9.89876568e-01 6.01040184e-01
-1.19947590e-01 2.69732028e-01 3.23166043e-01 -1.23333745e-01
2.51543432e-01 8.00518274e-01 1.64852545e-01 -4.65377986e-01
1.02274671e-01 6.26903117e-01 2.80561119e-01 -1.26401293e+00
-5.84429562e-01 -1.21533535e-01 -1.17695749e+00 5.36100030e-01
1.81809038e-01 -2.19622731e-01 -8.48493218e-01 1.38950253e+00
4.88846809e-01 7.44451284e-01 3.30701351e-01 5.42910039e-01
-1.28781363e-01 7.66876340e-01 1.44190580e-01 -2.81257033e-01
1.41468704e+00 -2.58927852e-01 -9.08801615e-01 2.97284484e-01
3.96395385e-01 -5.95268190e-01 5.38923025e-01 7.76015937e-01
-4.51107949e-01 -1.61304682e-01 -1.24886739e+00 1.19606280e+00
-7.58866072e-01 -7.13506877e-01 2.97683120e-01 1.09943712e+00
-5.81550360e-01 4.68701661e-01 -1.12541056e+00 -3.64437342e-01
6.53369129e-02 2.75648385e-01 -2.67954469e-01 4.96721268e-01
-1.48729062e+00 1.05340505e+00 4.32902962e-01 -3.15412521e-01
-1.23320782e+00 -7.06174552e-01 -5.71347654e-01 1.18970424e-01
5.27896166e-01 -2.76958719e-02 7.83030868e-01 4.95176800e-02
-1.13100874e+00 4.60976958e-02 5.94816208e-01 -5.41372597e-01
1.61141753e-01 -9.12852064e-02 -1.31733978e+00 2.33829364e-01
-2.91575253e-01 -4.47595149e-01 8.16717863e-01 -1.00248289e+00
-5.46368122e-01 -3.14270258e-01 -2.58453842e-02 -6.25898421e-01
-3.29264969e-01 3.53781134e-01 5.62673509e-01 -3.68335217e-01
-1.03065558e-01 -9.27682579e-01 -3.27093124e-01 -4.58188385e-01
-8.02741587e-01 -2.49218628e-01 1.82999921e+00 -4.14529890e-01
1.30297184e+00 -2.27195811e+00 -4.11824554e-01 8.92507195e-01
8.08942169e-02 9.03823078e-01 2.02401489e-01 7.97173381e-01
-3.20060045e-01 3.28046381e-01 -1.31336078e-01 2.14801982e-01
8.24311525e-02 2.16139168e-01 -7.26886690e-01 6.90111697e-01
5.61776944e-02 9.55469012e-02 -8.63823652e-01 3.27383399e-01
4.56683189e-01 3.25876504e-01 -2.04911996e-02 1.20663866e-01
1.22832298e-01 3.81855011e-01 -8.45236123e-01 8.77749205e-01
8.82925034e-01 1.60250328e-02 -1.39265969e-01 -2.11639293e-02
-2.92807579e-01 -7.02712983e-02 -1.39110661e+00 6.62726820e-01
-7.07276985e-02 2.27699399e-01 2.64152735e-01 -1.05166256e+00
9.85631645e-01 7.17267036e-01 6.73990071e-01 -5.48573077e-01
3.96377057e-01 2.09526256e-01 7.86398128e-02 -3.67365330e-01
1.31091252e-02 3.35512668e-01 -4.02346462e-01 7.04388678e-01
-5.86898029e-01 -1.10125408e-01 -7.61844888e-02 2.68801779e-01
1.78692365e+00 -6.32761717e-01 4.76331413e-01 -4.93808955e-01
9.38970983e-01 -6.16903044e-03 6.63418829e-01 4.37117934e-01
-4.86744553e-01 -2.50187963e-01 5.08793771e-01 -5.18790722e-01
-8.95784259e-01 -1.45102346e+00 -2.51314461e-01 -1.35852009e-01
4.02387142e-01 -3.38153392e-01 -5.99852741e-01 -9.12924886e-01
1.51127771e-01 9.71419096e-01 -2.28046030e-01 -6.43164933e-01
-4.23036337e-01 -7.59487152e-01 9.85856652e-01 4.33234900e-01
5.54735780e-01 -7.89421439e-01 -7.68699765e-01 3.63659412e-01
4.34241146e-01 -1.38481498e+00 -6.71020970e-02 1.08430032e-02
-4.72240627e-01 -1.70935988e+00 1.44943133e-01 1.26917614e-02
6.07290864e-01 2.68964231e-01 5.35196841e-01 2.08931461e-01
-5.29684484e-01 5.43901205e-01 -2.71961421e-01 -6.80086911e-01
-7.96782255e-01 -5.64786017e-01 1.00603676e+00 -1.40895694e-01
5.32906890e-01 -6.99205339e-01 -5.61797142e-01 7.35517263e-01
-1.07739532e+00 -8.12514365e-01 2.08558336e-01 3.62873465e-01
-5.68811409e-03 7.77876556e-01 8.27442884e-01 -3.78212899e-01
8.89844596e-01 -7.80671716e-01 -1.09321809e+00 -2.22421791e-02
-6.89453363e-01 -2.69306570e-01 9.47895288e-01 -5.82849383e-01
-6.30003273e-01 -4.17934954e-01 2.27623433e-03 -5.03777981e-01
-5.43072224e-01 2.47245461e-01 -2.94266194e-01 -1.86394200e-01
4.60581183e-01 -1.70377903e-02 8.81273299e-02 -3.40850383e-01
-3.68245125e-01 6.59708023e-01 3.94388497e-01 -5.11742294e-01
1.21379590e+00 5.93654156e-01 4.23279226e-01 -9.71439481e-01
-2.31283948e-01 -4.50049251e-01 -2.66912160e-03 -1.15994886e-01
5.20208240e-01 -4.29191649e-01 -1.36848545e+00 8.22172463e-01
-1.09792817e+00 7.12186396e-02 3.72502990e-02 6.63366079e-01
2.81528924e-02 6.53615355e-01 -8.79928887e-01 -1.06407976e+00
-4.59234625e-01 -1.07440674e+00 5.25080383e-01 2.16848269e-01
-2.89847702e-01 -9.67562139e-01 3.47315967e-01 -4.01959240e-01
6.35786772e-01 6.55976534e-01 8.64197433e-01 -1.28353536e+00
-3.64745378e-01 -4.38353360e-01 -4.80150394e-02 5.58077753e-01
5.15161693e-01 4.51098144e-01 -7.48614550e-01 -6.86921895e-01
3.83880019e-01 3.99482578e-01 -1.04538403e-01 -5.18529527e-02
9.74810481e-01 -2.68375546e-01 -7.13853419e-01 1.92083567e-01
1.34422874e+00 9.64409053e-01 7.09958911e-01 4.07658935e-01
2.73165047e-01 3.85584831e-01 7.92635381e-01 6.61615729e-01
-3.38795006e-01 3.78168523e-01 1.07125318e+00 1.90857604e-01
7.20062256e-01 2.78361380e-01 6.36523843e-01 5.15804708e-01
1.04736820e-01 -2.86036074e-01 -1.06579030e+00 2.80449450e-01
-1.39626586e+00 -1.01271224e+00 -3.81613791e-01 2.29592872e+00
1.31884649e-01 5.67608654e-01 9.88204032e-03 6.94816828e-01
7.69213676e-01 -1.85865775e-01 -5.81409872e-01 -6.32698298e-01
2.30019554e-01 2.22257122e-01 5.88840187e-01 2.11319342e-01
-1.18927574e+00 4.46765065e-01 5.93023872e+00 6.92053676e-01
-1.03662479e+00 -1.28440991e-01 2.12007672e-01 5.90389073e-01
3.98243845e-01 -1.12848161e-02 -6.79278433e-01 7.21877992e-01
1.32697642e+00 -3.76734346e-01 -3.86112137e-03 8.49518001e-01
5.20637989e-01 -4.34569754e-02 -7.76059628e-01 4.30278540e-01
-4.46665168e-01 -8.25758994e-01 -2.49070525e-01 3.10208648e-01
3.01613241e-01 -2.73771971e-01 -8.63049850e-02 1.99818373e-01
6.34513199e-01 -8.02383006e-01 1.22890389e-02 1.37807190e-01
2.45834857e-01 -1.08523595e+00 9.23158407e-01 4.35677171e-01
-1.31555164e+00 -3.30873579e-01 -4.08195779e-02 -1.21363498e-01
5.13052583e-01 7.87336648e-01 -6.03259563e-01 7.54334509e-01
6.37103260e-01 1.98047280e-01 -6.25815913e-02 9.09129858e-01
-2.28769660e-01 6.71883821e-01 -7.17252433e-01 5.60995862e-02
2.28727102e-01 -2.15142325e-01 1.10676992e+00 6.96262181e-01
5.26395023e-01 2.41415769e-01 3.54743838e-01 5.94698489e-01
7.03924000e-01 -5.67822754e-01 -8.20736110e-01 -1.29109278e-01
8.98755372e-01 1.11554253e+00 -7.78796077e-01 -1.18946247e-01
-2.35904142e-01 4.46365774e-01 -6.86390102e-01 1.91118032e-01
-1.04581177e+00 -6.84808552e-01 1.23976517e+00 5.20987026e-02
-2.31507882e-01 -5.63918471e-01 -2.34111063e-02 -8.04031789e-01
-4.12822008e-01 -9.78382289e-01 6.55015647e-01 -3.06666285e-01
-1.39572573e+00 5.74451029e-01 1.65554777e-01 -1.45472157e+00
-1.88882232e-01 -7.67970443e-01 -9.71514583e-01 5.93872607e-01
-9.59615231e-01 -6.81895435e-01 -6.90289028e-03 8.79034638e-01
2.55023949e-02 -2.65100658e-01 1.12393832e+00 2.79057503e-01
-8.39415193e-01 2.04882652e-01 -8.07759464e-02 2.77684659e-01
2.61930615e-01 -8.54605377e-01 7.40182579e-01 1.33117950e+00
-4.18388575e-01 6.19433761e-01 1.03217137e+00 -8.61694872e-01
-1.47955084e+00 -7.52908826e-01 4.12451178e-02 -3.04880559e-01
1.25094461e+00 -3.25461626e-01 -1.00678372e+00 5.29017985e-01
1.61619723e-01 -2.24492531e-02 4.08069849e-01 -4.85899657e-01
-2.13005736e-01 5.63602448e-02 -1.66710675e+00 6.23384178e-01
5.88515401e-01 -2.65699536e-01 -5.91114700e-01 2.26670891e-01
7.85494924e-01 -1.10182073e-02 -9.31023836e-01 5.60811400e-01
-4.24819021e-03 -7.28711784e-01 1.08290517e+00 -5.64357281e-01
-4.73172098e-01 -5.42094290e-01 -8.03940669e-02 -1.41129053e+00
-8.26371387e-02 -8.33974302e-01 -5.15474319e-01 1.10997164e+00
-1.04323678e-01 -1.47432101e+00 3.44451338e-01 4.22855675e-01
-1.35316864e-01 -4.10048276e-01 -1.13168144e+00 -1.31834805e+00
-2.80992150e-01 -4.34023231e-01 9.95076656e-01 1.27613306e+00
4.65058535e-01 -1.12148136e-01 3.72064300e-02 1.05337024e+00
8.42148364e-01 -3.81385654e-01 7.68531859e-01 -1.34314549e+00
2.20460668e-02 -3.32081616e-01 -7.70178378e-01 -2.57389724e-01
-2.16943473e-01 -1.06399842e-02 -4.41047221e-01 -8.54899645e-01
-4.27532881e-01 -5.48852801e-01 -6.33668661e-01 4.61820871e-01
3.79203767e-01 -1.24395445e-01 -3.58732082e-02 -1.40781617e-02
1.08535672e-02 3.35163683e-01 7.63767183e-01 -1.66331157e-01
-5.75528033e-02 2.10580602e-01 4.15686555e-02 8.70723069e-01
1.17246974e+00 -5.87937355e-01 -5.10867596e-01 4.87848550e-01
-4.42121141e-02 1.92411557e-01 3.78620565e-01 -1.25839174e+00
2.81738758e-01 -5.42364120e-01 -9.94774923e-02 -5.64084589e-01
2.74532199e-01 -1.46830809e+00 3.69569451e-01 1.28400123e+00
5.85473657e-01 5.96058190e-01 4.31852281e-01 6.77546561e-01
-1.81947887e-01 7.32419342e-02 7.90329576e-01 3.23419243e-01
-6.17038906e-01 4.48099017e-01 -8.88077319e-01 -2.95521855e-01
1.76324069e+00 -8.77769813e-02 -8.40981781e-01 -1.54288441e-01
-5.71500897e-01 4.14044529e-01 2.22889289e-01 5.30670047e-01
6.88550591e-01 -1.02391660e+00 -2.34363750e-01 5.57865083e-01
-2.17597842e-01 -2.59125233e-01 1.43960714e-02 8.87526214e-01
-3.40495646e-01 3.79228890e-01 -3.41481805e-01 -4.90185201e-01
-1.36371469e+00 9.36948657e-01 3.05821866e-01 -5.06337225e-01
-6.21951580e-01 1.68829769e-01 -5.04147112e-02 -9.88814831e-02
1.33210748e-01 -2.89527625e-01 -1.23155855e-01 -3.03207636e-01
8.11630011e-01 7.95242846e-01 -4.63822447e-02 -2.43899092e-01
-6.91365540e-01 1.60308063e-01 -1.15900829e-01 1.32203445e-01
1.09360433e+00 1.45146906e-01 -1.76217526e-01 -2.80096829e-02
6.19242251e-01 4.94224988e-02 -8.95004392e-01 2.23216251e-01
3.84528518e-01 -4.83008802e-01 -2.02913195e-01 -7.35436141e-01
-1.02508092e+00 5.88189602e-01 4.68510062e-01 1.07364726e+00
1.01828599e+00 -4.38511282e-01 9.30534899e-01 2.11558416e-01
8.95641565e-01 -7.45802641e-01 2.80322760e-01 7.14897394e-01
4.21490967e-01 -5.97013533e-01 -7.05027804e-02 -6.09240532e-01
-1.41114607e-01 1.08586490e+00 6.62284732e-01 -4.07772154e-01
1.07458460e+00 9.86607730e-01 -2.18336701e-01 -4.02411729e-01
-6.44327700e-01 5.72076261e-01 -4.24359441e-01 7.54655838e-01
-5.82306564e-01 3.01719546e-01 -1.70295656e-01 5.38690627e-01
3.25334281e-01 -5.03768802e-01 9.09350276e-01 1.14320815e+00
-3.61372352e-01 -1.17035615e+00 -8.58119786e-01 1.15212791e-01
-5.73701680e-01 3.48751724e-01 -1.84143260e-01 1.00753438e+00
-3.83145422e-01 1.26368201e+00 1.80458456e-01 -6.74640656e-01
4.13508981e-01 -1.48215607e-01 -2.75175989e-01 -1.94068298e-01
-4.09541011e-01 -3.57354879e-01 -6.34316215e-03 -7.89265931e-01
6.84901178e-02 -4.96588737e-01 -1.63443935e+00 -7.56733835e-01
-3.63752604e-01 4.82811928e-01 7.76996672e-01 8.74202430e-01
4.44358557e-01 6.87234461e-01 9.96941805e-01 -4.18022633e-01
-1.00870740e+00 -8.20231199e-01 -8.33313286e-01 2.68088698e-01
2.22325370e-01 -1.03722394e+00 -9.83983994e-01 -7.52952456e-01] | [5.3653082847595215, 7.189121723175049] |
75e190ec-60db-48dc-8d1a-42e9f0bd747f | actor-context-actor-relation-network-for | 2006.07976 | null | https://arxiv.org/abs/2006.07976v3 | https://arxiv.org/pdf/2006.07976v3.pdf | Actor-Context-Actor Relation Network for Spatio-Temporal Action Localization | Localizing persons and recognizing their actions from videos is a challenging task towards high-level video understanding. Recent advances have been achieved by modeling direct pairwise relations between entities. In this paper, we take one step further, not only model direct relations between pairs but also take into account indirect higher-order relations established upon multiple elements. We propose to explicitly model the Actor-Context-Actor Relation, which is the relation between two actors based on their interactions with the context. To this end, we design an Actor-Context-Actor Relation Network (ACAR-Net) which builds upon a novel High-order Relation Reasoning Operator and an Actor-Context Feature Bank to enable indirect relation reasoning for spatio-temporal action localization. Experiments on AVA and UCF101-24 datasets show the advantages of modeling actor-context-actor relations, and visualization of attention maps further verifies that our model is capable of finding relevant higher-order relations to support action detection. Notably, our method ranks first in the AVA-Kineticsaction localization task of ActivityNet Challenge 2020, out-performing other entries by a significant margin (+6.71mAP). Training code and models will be available at https://github.com/Siyu-C/ACAR-Net. | ['Yu Liu', 'Mike Zheng Shou', 'Siyu Chen', 'Junting Pan', 'Hongsheng Li', 'Jing Shao'] | 2020-06-14 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Pan_Actor-Context-Actor_Relation_Network_for_Spatio-Temporal_Action_Localization_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Pan_Actor-Context-Actor_Relation_Network_for_Spatio-Temporal_Action_Localization_CVPR_2021_paper.pdf | cvpr-2021-1 | ['spatio-temporal-action-localization'] | ['computer-vision'] | [ 6.30521923e-02 1.32756799e-01 -3.44375461e-01 -3.79349351e-01
-2.29969800e-01 -4.61411119e-01 1.04600453e+00 1.37298658e-01
-4.90155309e-01 4.83919233e-01 8.41310382e-01 3.24527845e-02
-3.67148370e-01 -5.73477983e-01 -6.31958723e-01 -3.09889346e-01
-5.82440376e-01 2.77061969e-01 3.13857079e-01 -1.70486480e-01
-2.59951711e-01 3.89982373e-01 -1.30489719e+00 8.30702305e-01
3.82386088e-01 9.03125226e-01 -2.44146049e-01 5.74534416e-01
4.44184572e-01 1.61352706e+00 -3.82243335e-01 -5.16667426e-01
1.28938153e-01 -5.49134970e-01 -1.21396828e+00 -1.11157373e-01
5.64519823e-01 -3.41701955e-01 -7.66979933e-01 6.58805728e-01
2.15027496e-01 3.16375375e-01 5.14558256e-01 -1.50249207e+00
-5.01594424e-01 6.80332184e-01 -2.87826270e-01 6.46185100e-01
8.91367555e-01 2.34542370e-01 1.43678856e+00 -8.49751294e-01
9.20423687e-01 1.20913351e+00 6.35132313e-01 4.54725236e-01
-1.00981438e+00 -4.15261060e-01 5.80640137e-01 7.96797216e-01
-1.67064941e+00 -3.60521644e-01 4.40113097e-01 -4.90884960e-01
1.48456192e+00 4.46063697e-01 1.05485463e+00 1.25419724e+00
-1.46836653e-01 1.03611898e+00 5.85940480e-01 -5.35759218e-02
-2.67935038e-01 -3.30852896e-01 1.00327246e-01 9.04628873e-01
-5.12461476e-02 -3.34587663e-01 -9.49717581e-01 -1.75441839e-02
7.57075846e-01 2.50607338e-02 -3.83451670e-01 -2.00588211e-01
-1.64100838e+00 3.55020165e-01 6.70423329e-01 4.54213619e-01
-4.46027011e-01 4.36436296e-01 4.49204922e-01 3.33119668e-02
3.31528991e-01 5.10926604e-01 -3.11724335e-01 -3.06680083e-01
-4.97732222e-01 2.53199637e-01 6.32183611e-01 8.94468904e-01
3.15690666e-01 -5.60750008e-01 -5.71803629e-01 5.10433853e-01
2.64764249e-01 3.39176203e-03 4.02494892e-02 -1.08277798e+00
6.03519976e-01 9.42210138e-01 1.74931679e-02 -1.20514274e+00
-4.85755384e-01 -1.53316900e-01 -7.98523426e-01 -2.38397032e-01
5.23719728e-01 9.78310034e-02 -3.98379654e-01 1.71088779e+00
4.98769224e-01 7.37637937e-01 -1.04136027e-01 9.25330341e-01
1.02079880e+00 4.32368398e-01 3.11352402e-01 5.34841307e-02
1.67292893e+00 -1.09488237e+00 -6.33182228e-01 -1.06149361e-01
8.86758685e-01 -1.73414052e-01 6.60676956e-01 1.63154215e-01
-1.06148243e+00 -5.09451807e-01 -6.35596693e-01 -3.01588565e-01
-4.63519335e-01 3.37340206e-01 8.37147474e-01 -7.59601686e-03
-9.67944682e-01 4.00504142e-01 -8.35452259e-01 -7.45282471e-01
7.65060127e-01 2.76649803e-01 -8.91636789e-01 7.73616806e-02
-1.44083166e+00 9.70308661e-01 3.68467480e-01 3.11579168e-01
-8.25605273e-01 -6.40224278e-01 -8.89477968e-01 1.83746219e-01
7.64970243e-01 -7.64570057e-01 8.24833810e-01 -8.46230388e-01
-9.89537537e-01 8.66687775e-01 -2.02646464e-01 -6.94812059e-01
5.51043630e-01 -5.49650073e-01 -5.91830492e-01 4.93314505e-01
-9.39984396e-02 7.11371243e-01 2.37032056e-01 -7.23587990e-01
-7.92116225e-01 -1.41469032e-01 6.85333550e-01 4.06954139e-01
-3.20367634e-01 4.32929218e-01 -7.01245606e-01 -5.80510080e-01
-1.12770960e-01 -8.83569419e-01 -8.58983919e-02 1.14103876e-01
-2.90457428e-01 -6.64454043e-01 5.74096441e-01 -7.25579500e-01
1.47768712e+00 -1.92993534e+00 2.99122572e-01 1.82481900e-01
5.39294064e-01 4.00590599e-01 -1.25448748e-01 5.97510993e-01
-3.94029349e-01 3.39040160e-03 1.07324548e-01 -1.78329021e-01
-1.16952015e-02 1.59400940e-01 -7.86120370e-02 5.76631248e-01
4.78686571e-01 1.21645117e+00 -1.13687766e+00 -6.28420651e-01
2.25051135e-01 7.55233347e-01 -5.44704974e-01 -3.98592986e-02
-1.57523140e-01 4.80016649e-01 -3.49778622e-01 6.84803069e-01
-1.82651319e-02 -5.33218980e-01 5.20584464e-01 -5.47506928e-01
9.72565729e-03 4.15541798e-01 -1.12192941e+00 1.65125632e+00
-1.56809434e-01 9.03554082e-01 -1.72122091e-01 -9.80290294e-01
4.35953766e-01 4.30650383e-01 8.97697628e-01 -4.78036612e-01
1.28237102e-02 -3.40225816e-01 1.43466204e-01 -8.56247246e-01
2.86004275e-01 5.45583546e-01 9.38033499e-03 1.34437710e-01
2.66961575e-01 8.06786120e-01 5.43279111e-01 7.44254112e-01
1.51531577e+00 4.31612700e-01 5.67879677e-01 -1.09468170e-01
8.97090256e-01 -2.76965678e-01 5.88221848e-01 5.17031729e-01
-4.25761938e-01 4.02162045e-01 9.72296655e-01 -6.71059728e-01
-4.54385936e-01 -7.31879532e-01 2.32909709e-01 1.14841938e+00
1.94473252e-01 -1.21930873e+00 -4.23968554e-01 -1.10649312e+00
-1.27266198e-01 4.37006146e-01 -1.04396629e+00 3.65432985e-02
-8.67115259e-01 -2.76246637e-01 7.48673022e-01 8.27068031e-01
6.78514361e-01 -9.43139911e-01 -6.28293514e-01 -1.18507057e-01
-6.87346458e-01 -1.47099161e+00 -4.54462707e-01 -3.16956699e-01
-3.72720778e-01 -1.52555621e+00 -4.11020815e-01 -5.48257768e-01
5.65075994e-01 -1.16854860e-02 1.27197599e+00 3.00919414e-01
-3.15587163e-01 8.08853030e-01 -4.62162644e-01 1.05587952e-01
1.94265097e-01 -1.20436475e-01 5.25141023e-02 4.30722684e-01
5.10539293e-01 -5.01377344e-01 -7.07293868e-01 5.73909700e-01
-4.82168585e-01 2.28134602e-01 3.77785951e-01 6.37310684e-01
4.49117482e-01 1.02465495e-03 2.41862372e-01 -6.61744535e-01
1.88020632e-01 -6.35853410e-01 -2.18911260e-01 6.12653553e-01
-1.70126230e-01 -2.36439094e-01 3.40771198e-01 -3.76001030e-01
-1.16278136e+00 2.78864145e-01 3.52334268e-02 -4.30094749e-01
-2.93040931e-01 5.13624251e-01 -3.80980700e-01 3.34104151e-01
5.06469250e-01 -4.20294553e-02 -4.63215709e-01 -1.15475580e-01
4.85157788e-01 1.60529017e-01 5.48206091e-01 -5.66451073e-01
5.03508449e-01 7.34911501e-01 2.93604344e-01 -6.50101542e-01
-9.51439083e-01 -7.51147509e-01 -9.84559774e-01 -6.47459388e-01
1.25927925e+00 -1.11137402e+00 -1.19376528e+00 2.44265676e-01
-1.19679296e+00 -4.45444763e-01 -2.57996887e-01 5.78480244e-01
-3.97100925e-01 2.81566232e-01 -5.38934469e-01 -6.59219623e-01
7.21650720e-02 -7.89631963e-01 8.85092258e-01 -2.57298257e-02
-4.81315643e-01 -1.04131496e+00 -8.95349681e-02 6.45365298e-01
5.70202135e-02 4.04001743e-01 2.87111878e-01 -7.35544503e-01
-7.60050535e-01 -1.63821727e-01 -2.94432759e-01 3.35180061e-03
-3.53797190e-02 -8.15979615e-02 -5.56376159e-01 2.97704078e-02
-5.71551442e-01 -1.85473680e-01 1.03427672e+00 1.12948418e-01
1.02699542e+00 -3.49981755e-01 -5.70451856e-01 4.77222949e-01
7.58450806e-01 -1.40534669e-01 7.13666975e-01 1.14428267e-01
1.08168304e+00 6.79994643e-01 8.48977566e-01 4.49711680e-01
7.40940750e-01 9.91926730e-01 3.41817975e-01 5.63823394e-02
-3.89547348e-01 -4.91844654e-01 4.07334834e-01 1.34133399e-01
-6.87013507e-01 -3.71502161e-01 -1.08232319e+00 5.96181393e-01
-2.29044008e+00 -1.46295941e+00 -5.94880819e-01 1.77027214e+00
5.67451954e-01 -6.43350035e-02 4.30926532e-01 -2.10224733e-01
5.49185038e-01 3.99549335e-01 -2.20579907e-01 3.07130873e-01
-1.03973404e-01 -1.99634269e-01 2.50346363e-01 5.30814528e-01
-1.43837619e+00 1.02227354e+00 5.18497753e+00 6.75483286e-01
-4.90151376e-01 7.73760527e-02 5.05166650e-01 -3.95510197e-01
1.61312789e-01 9.58914459e-02 -8.31883252e-01 2.35252827e-01
6.00797474e-01 2.29110315e-01 3.78180087e-01 5.65950155e-01
2.13075846e-01 -2.05682710e-01 -1.56051958e+00 1.13385510e+00
3.55522275e-01 -1.42060828e+00 1.48376366e-02 1.56382453e-02
4.61898029e-01 -1.37144819e-01 -3.90677720e-01 2.59557188e-01
2.57856458e-01 -1.02213085e+00 6.61571026e-01 8.96739006e-01
5.12276292e-01 -4.67298508e-01 6.16666436e-01 7.09638968e-02
-1.78136241e+00 -4.83268127e-02 1.23295031e-01 -3.11355054e-01
3.60808223e-01 3.15232515e-01 -6.73900962e-01 6.89500153e-01
9.27468836e-01 1.43699193e+00 -7.46393561e-01 7.83110738e-01
-6.65164411e-01 6.05842471e-01 -2.45776966e-01 1.34422302e-01
8.37366879e-02 -9.34101269e-02 5.41705012e-01 1.39428520e+00
2.87220511e-03 5.89470506e-01 1.60223275e-01 7.35449135e-01
-1.37736604e-01 -1.39897972e-01 -5.11128783e-01 -7.99227729e-02
1.31379351e-01 1.23329163e+00 -7.51585841e-01 -5.60671329e-01
-5.90046525e-01 1.03037333e+00 5.31596899e-01 3.20468873e-01
-1.29734373e+00 -2.22630594e-02 8.02709758e-01 1.51591823e-01
2.53524184e-01 -1.80144891e-01 2.84834474e-01 -1.41912556e+00
2.36375645e-01 -8.24516416e-01 8.08212101e-01 -8.18314373e-01
-1.08790040e+00 4.99920219e-01 2.84464538e-01 -1.14249456e+00
-2.81531326e-02 -5.75805068e-01 -4.48482603e-01 4.85508978e-01
-9.92957652e-01 -1.58571827e+00 -4.19201910e-01 8.55667233e-01
4.19883937e-01 -1.90340336e-02 6.12854362e-01 5.38806200e-01
-7.08809137e-01 4.01828289e-01 -7.33953714e-01 6.25175714e-01
4.93973017e-01 -1.03366899e+00 3.34612101e-01 1.00650299e+00
6.68625891e-01 6.96855843e-01 4.43538070e-01 -7.05496669e-01
-1.11064231e+00 -1.01130605e+00 1.24298143e+00 -9.46801960e-01
8.98951352e-01 -5.85702300e-01 -7.04185009e-01 1.20802331e+00
3.04510236e-01 4.80559498e-01 7.91935921e-01 5.49051285e-01
-6.95585012e-01 -7.34319836e-02 -6.93678141e-01 7.31301606e-01
1.88399565e+00 -7.22562432e-01 -5.55011690e-01 5.18340647e-01
3.50380033e-01 -2.24465862e-01 -9.82226253e-01 4.74029690e-01
7.07347631e-01 -1.06345129e+00 1.31070006e+00 -8.65298748e-01
4.86146897e-01 -4.69135672e-01 4.02424596e-02 -7.81908929e-01
-4.69131827e-01 -5.81046820e-01 -6.70804739e-01 1.37572253e+00
3.74377459e-01 -4.12024349e-01 6.92385435e-01 7.05053270e-01
5.84032238e-02 -9.00856078e-01 -8.60606909e-01 -7.63760924e-01
-6.41296923e-01 -4.64784294e-01 3.15867960e-01 1.23694265e+00
5.13004959e-01 4.40293968e-01 -5.69276571e-01 3.10676187e-01
1.75684139e-01 -1.27839014e-01 7.64836848e-01 -8.30338538e-01
-6.02375507e-01 -4.66852248e-01 -5.89596391e-01 -1.29417229e+00
2.42656067e-01 -8.31604600e-01 -2.43206248e-01 -1.76914239e+00
3.03656429e-01 -1.67430222e-01 -3.39613169e-01 7.93870389e-01
-1.99592367e-01 2.53986478e-01 2.93487936e-01 2.30286524e-01
-1.44983244e+00 4.29778248e-01 8.48836482e-01 -1.31666288e-01
-7.66045973e-03 -5.18460795e-02 -2.25096047e-01 9.74481642e-01
6.30197108e-01 -2.84788579e-01 -4.58729684e-01 -4.26755905e-01
4.30989951e-01 -2.43812650e-02 8.46740842e-01 -1.01196575e+00
4.39344913e-01 -1.08539835e-01 3.74171793e-01 -5.17270505e-01
6.28662229e-01 -9.38485563e-01 2.87193477e-01 2.36806884e-01
-6.75595343e-01 3.47994715e-02 -2.12539822e-01 6.93867862e-01
-2.73847014e-01 1.97015449e-01 1.11360893e-01 -1.15154944e-01
-9.36422110e-01 3.23919237e-01 -3.42560321e-01 6.89994842e-02
1.29606736e+00 8.15769471e-03 -4.46189851e-01 -4.44252998e-01
-1.09418786e+00 3.35615754e-01 7.42813274e-02 4.53976542e-01
4.76063520e-01 -1.49808812e+00 -6.93281591e-01 -2.50543714e-01
2.95466155e-01 -2.91370809e-01 4.70921010e-01 1.27965355e+00
-3.29657555e-01 5.25606155e-01 2.12142561e-02 -4.07343268e-01
-1.77527249e+00 4.91849661e-01 4.01395142e-01 -5.39275229e-01
-6.14171445e-01 1.19445753e+00 2.61322081e-01 -4.74810898e-02
3.29532236e-01 -3.95155132e-01 -4.67795461e-01 1.90842137e-01
6.75692976e-01 5.59601426e-01 -4.07546997e-01 -1.17285109e+00
-8.84571552e-01 2.98695326e-01 2.43492395e-01 5.54504730e-02
1.32496536e+00 1.13255076e-01 -2.52394557e-01 1.86772212e-01
9.95752275e-01 -1.48596361e-01 -1.41566503e+00 -3.10491204e-01
7.75383785e-02 -6.85722113e-01 -3.23306054e-01 -8.98584187e-01
-1.09632170e+00 6.95311189e-01 1.75476000e-01 2.52127022e-01
1.02258956e+00 4.94770825e-01 2.66400367e-01 4.46685553e-01
1.96950436e-01 -9.03456807e-01 3.28953445e-01 3.87328684e-01
1.14738214e+00 -1.19799161e+00 2.45242834e-01 -5.74551642e-01
-7.84987867e-01 8.55670810e-01 9.01637852e-01 3.68180759e-02
6.39300644e-01 -1.03383727e-01 -3.90740365e-01 -5.35479724e-01
-9.47234154e-01 -5.61554611e-01 6.68791294e-01 4.84102845e-01
5.52246094e-01 -3.48465145e-02 -1.91374078e-01 4.19043690e-01
3.65537852e-01 3.78410965e-02 -3.45197283e-02 7.35016048e-01
1.18076190e-01 -1.01657259e+00 -1.15561567e-01 2.70427972e-01
-3.58079076e-01 -1.28536835e-01 -5.66421330e-01 7.63905466e-01
5.00517428e-01 9.61084068e-01 1.35351911e-01 -4.47332829e-01
4.45710868e-01 -5.27530201e-02 4.47287887e-01 -4.01596785e-01
-7.28042543e-01 -1.78418741e-01 6.44062877e-01 -1.07616830e+00
-9.30847049e-01 -1.08003902e+00 -1.25198889e+00 -1.68205902e-01
-4.99965996e-02 -1.48161456e-01 1.59614086e-02 9.18295085e-01
5.81542373e-01 6.17959857e-01 -1.30098267e-02 -5.95589936e-01
2.47768506e-01 -8.35907161e-01 -1.84548855e-01 6.67790473e-01
2.18027942e-02 -6.87633038e-01 -8.82399008e-02 3.03152084e-01] | [8.36176872253418, 0.6340634822845459] |
6be627fd-48d9-473f-822e-d1b384ad9b0b | hierarchical-reinforcement-learning-for-open | 1909.07547 | null | https://arxiv.org/abs/1909.07547v3 | https://arxiv.org/pdf/1909.07547v3.pdf | Hierarchical Reinforcement Learning for Open-Domain Dialog | Open-domain dialog generation is a challenging problem; maximum likelihood training can lead to repetitive outputs, models have difficulty tracking long-term conversational goals, and training on standard movie or online datasets may lead to the generation of inappropriate, biased, or offensive text. Reinforcement Learning (RL) is a powerful framework that could potentially address these issues, for example by allowing a dialog model to optimize for reducing toxicity and repetitiveness. However, previous approaches which apply RL to open-domain dialog generation do so at the word level, making it difficult for the model to learn proper credit assignment for long-term conversational rewards. In this paper, we propose a novel approach to hierarchical reinforcement learning, VHRL, which uses policy gradients to tune the utterance-level embedding of a variational sequence model. This hierarchical approach provides greater flexibility for learning long-term, conversational rewards. We use self-play and RL to optimize for a set of human-centered conversation metrics, and show that our approach provides significant improvements -- in terms of both human evaluation and automatic metrics -- over state-of-the-art dialog models, including Transformers. | ['Natasha Jaques', 'Abdelrhman Saleh', 'Rosalind Picard', 'Judy Hanwen Shen', 'Asma Ghandeharioun'] | 2019-09-17 | null | null | null | null | ['open-domain-dialog'] | ['natural-language-processing'] | [-6.10762369e-03 6.64241731e-01 -1.84472650e-01 -4.87356961e-01
-1.08571494e+00 -8.02915096e-01 7.97326922e-01 -1.41738564e-01
-2.70525545e-01 1.04644251e+00 7.47455716e-01 -3.51558059e-01
2.20300332e-01 -6.36234045e-01 -2.19747335e-01 -4.18007970e-01
1.06328391e-01 8.56774628e-01 4.63671377e-03 -7.08291888e-01
3.14754784e-01 1.62805319e-02 -1.01562285e+00 3.55679780e-01
8.71755838e-01 6.36728466e-01 1.78020880e-01 8.22877705e-01
-3.29023629e-01 1.16951478e+00 -9.19504166e-01 -6.24799788e-01
-6.08575977e-02 -6.82893813e-01 -1.03978312e+00 9.28909332e-02
-4.20771688e-02 -6.46329284e-01 -4.42265384e-02 7.83407867e-01
5.84379137e-01 5.35576880e-01 8.80541325e-01 -1.17824125e+00
-6.95778370e-01 9.58140194e-01 -4.07544188e-02 -2.31277257e-01
5.10214567e-01 5.42714298e-01 1.39635682e+00 -5.79671323e-01
6.12639427e-01 1.79420447e+00 4.39349353e-01 1.09450746e+00
-1.39115310e+00 -4.88184452e-01 1.90109581e-01 -3.31129462e-01
-6.81783915e-01 -4.64866310e-01 7.19142377e-01 -5.44953465e-01
9.64220643e-01 8.99894983e-02 2.72132784e-01 1.52466667e+00
9.50512066e-02 7.73455083e-01 1.04892826e+00 -2.85630405e-01
3.49486709e-01 4.06001985e-01 -7.73046389e-02 5.16615987e-01
-5.50707042e-01 3.61838400e-01 -4.60155457e-01 -5.33915102e-01
5.90696096e-01 -3.03589821e-01 -5.63269900e-03 -2.21407190e-01
-8.81458700e-01 1.51754367e+00 3.37980181e-01 1.93837300e-01
-2.33209208e-01 -8.73709619e-02 3.07618231e-01 4.16325957e-01
5.46617985e-01 1.14624727e+00 -4.14069802e-01 -4.61683214e-01
-5.92217207e-01 7.58230567e-01 1.39898968e+00 6.36453032e-01
6.04582310e-01 4.85789850e-02 -7.64956892e-01 1.27438140e+00
3.75909775e-01 2.09960014e-01 6.61730587e-01 -1.46253657e+00
4.10784483e-01 3.52628022e-01 4.03045535e-01 -4.27518904e-01
-3.50361556e-01 1.26715258e-01 -3.33703905e-01 3.30143154e-01
6.80321872e-01 -7.90962756e-01 -5.23100793e-01 2.05108595e+00
2.71258682e-01 -4.31343317e-01 4.07213628e-01 8.09569538e-01
7.09745407e-01 8.41337681e-01 2.15799689e-01 -3.07452559e-01
1.08942771e+00 -1.22844291e+00 -8.69855404e-01 -4.02853161e-01
4.68967170e-01 -6.16412461e-01 1.52647567e+00 2.45754510e-01
-1.25296628e+00 -3.02375406e-01 -7.86630929e-01 1.81889646e-02
-1.94400549e-01 -4.57876951e-01 3.08209896e-01 5.37864268e-01
-8.31743479e-01 7.41073608e-01 -3.13230753e-01 -1.76847711e-01
-7.82250911e-02 2.58829683e-01 2.84308404e-01 3.28798771e-01
-1.52564561e+00 1.19988775e+00 9.23660770e-02 -3.34822834e-01
-9.06464279e-01 -4.92214382e-01 -8.03222120e-01 1.94632590e-01
5.28908372e-01 -4.83089745e-01 2.08041573e+00 -8.22504938e-01
-2.23064804e+00 4.87281471e-01 1.05636150e-01 -4.35767353e-01
5.61269760e-01 -2.83799589e-01 9.04305652e-02 -1.15424000e-01
-3.25533077e-02 1.08532238e+00 8.60779226e-01 -1.17167759e+00
-2.88866967e-01 -1.39569901e-02 3.30208361e-01 6.82403505e-01
-4.82348293e-01 -4.69168164e-02 1.83746099e-01 -5.69240808e-01
-6.63810909e-01 -1.08032703e+00 -5.31857908e-01 -4.41700816e-01
-3.36238354e-01 -7.11033940e-01 4.80422974e-01 -6.12062275e-01
1.17072403e+00 -1.92174435e+00 3.55968475e-01 -1.82370052e-01
3.67248021e-02 1.09969877e-01 -8.68038014e-02 5.48427165e-01
4.91837442e-01 2.20454752e-01 -1.10218078e-01 -4.64229822e-01
3.35701942e-01 2.21511409e-01 -3.58790278e-01 -2.91780442e-01
2.19520569e-01 8.71620893e-01 -9.83002126e-01 -5.85665226e-01
2.84910239e-02 7.04402626e-02 -8.67983162e-01 8.93884063e-01
-9.97521460e-01 6.11769855e-01 -3.07441145e-01 7.62632266e-02
-4.21077898e-03 -2.33641446e-01 3.77236217e-01 4.27671880e-01
9.95000228e-02 6.19136751e-01 -7.05860376e-01 1.45439959e+00
-6.90523505e-01 3.60667557e-01 9.99366790e-02 -3.24163944e-01
1.03626251e+00 4.07928854e-01 1.72983557e-01 -3.61748338e-01
1.88391462e-01 -2.85064373e-02 6.87791593e-03 -4.66402829e-01
7.92516947e-01 -5.01964927e-01 -4.24459785e-01 7.67062902e-01
7.71001875e-02 -5.93300283e-01 -3.06744017e-02 3.94025981e-01
8.85000467e-01 1.77722834e-02 5.61427251e-02 3.68095748e-02
3.20537359e-01 1.15926994e-03 3.01562428e-01 8.61279011e-01
-2.06248909e-01 3.48609865e-01 8.18214417e-01 7.05750659e-02
-9.19558823e-01 -1.09716916e+00 2.75238574e-01 1.69953752e+00
-1.77424297e-01 -2.71925420e-01 -1.01395857e+00 -7.96063721e-01
-2.72032488e-02 1.15913379e+00 -2.71189153e-01 -2.06052914e-01
-5.16787529e-01 -3.98140073e-01 5.18482924e-01 3.12180638e-01
1.90744445e-01 -1.56144786e+00 -2.76842445e-01 5.19076765e-01
-5.58309376e-01 -1.07554412e+00 -7.79470444e-01 1.55875459e-01
-7.44317472e-01 -6.05525792e-01 -7.42504895e-01 -6.79035485e-01
2.03141123e-01 -2.16323555e-01 1.29701614e+00 -9.04182568e-02
1.34958595e-01 2.14203015e-01 -3.05443674e-01 -3.03420544e-01
-1.16992664e+00 2.44969413e-01 -1.28767923e-01 -2.61726499e-01
1.94515064e-01 -3.12935531e-01 -4.18622106e-01 3.99036467e-01
-5.34255624e-01 -1.49501801e-01 2.50699610e-01 1.32020545e+00
-1.38920665e-01 -6.99633002e-01 1.17339003e+00 -1.00231218e+00
1.68324482e+00 -5.39987504e-01 -5.28245747e-01 1.94396973e-01
-7.73363173e-01 3.48248363e-01 6.93762481e-01 -6.15539968e-01
-1.34329724e+00 -1.47316143e-01 -2.56539643e-01 -2.56074756e-01
-8.96777064e-02 2.97423780e-01 5.44923618e-02 3.82050753e-01
1.06866729e+00 -1.33951440e-01 3.66471738e-01 -3.22283655e-01
7.62759924e-01 9.57727075e-01 2.84512609e-01 -6.84961021e-01
4.04442161e-01 -2.43280724e-01 -6.25358939e-01 -7.38536954e-01
-1.05587912e+00 -2.13778555e-01 -1.92205876e-01 -2.42950022e-01
9.88601565e-01 -7.87430882e-01 -8.11026573e-01 1.25194922e-01
-1.06989777e+00 -9.65888500e-01 -2.53087908e-01 7.94243515e-02
-7.60197937e-01 2.46459052e-01 -9.50750053e-01 -1.06092715e+00
-4.22817290e-01 -1.17207479e+00 8.46881986e-01 3.71026427e-01
-8.84236455e-01 -1.12772036e+00 2.95764685e-01 6.42936647e-01
5.01003921e-01 -1.22919165e-01 8.98405850e-01 -9.88863826e-01
-3.34315091e-01 1.62168860e-01 2.37090975e-01 4.31546181e-01
-2.01094821e-02 -2.29052573e-01 -9.76899445e-01 -1.30412549e-01
8.88883974e-03 -1.23126030e+00 3.89787614e-01 2.13986948e-01
6.72527194e-01 -7.24349201e-01 5.11344112e-02 2.62308866e-02
6.27539635e-01 2.04287991e-01 3.47223818e-01 -2.17614062e-02
3.48532557e-01 9.36881840e-01 8.62150252e-01 6.69519305e-01
4.76796091e-01 8.04543138e-01 1.16336033e-01 1.62393019e-01
1.83750823e-01 -5.48845172e-01 6.13965392e-01 5.39842010e-01
4.20918554e-01 -2.89987057e-01 -4.92571503e-01 4.48941946e-01
-1.92698979e+00 -9.79765117e-01 3.18693817e-01 1.91654801e+00
1.37853134e+00 3.39893460e-01 5.30564725e-01 -4.96941209e-01
5.18321812e-01 3.68695855e-01 -6.50592566e-01 -8.99734080e-01
1.65137544e-01 2.66260467e-03 2.73895953e-02 1.05990982e+00
-7.99895167e-01 1.35007989e+00 6.80896616e+00 5.10729671e-01
-8.60892653e-01 1.94097281e-01 7.42321789e-01 -1.75764635e-01
-4.96362627e-01 6.65197968e-02 -8.59429955e-01 2.98347205e-01
1.15170240e+00 -1.86688587e-01 7.02566147e-01 9.78694081e-01
4.00480777e-01 1.16621166e-01 -1.23171163e+00 5.48989415e-01
-5.35294339e-02 -1.24480569e+00 -1.88852668e-01 1.62622362e-01
7.22559512e-01 -2.23639965e-01 -5.06379567e-02 9.17500794e-01
1.08978713e+00 -1.09241164e+00 4.01333719e-01 1.52887344e-01
3.84113848e-01 -6.93936050e-01 4.14511234e-01 6.74055815e-01
-3.90528411e-01 -1.88846588e-01 -2.33654335e-01 -1.58144340e-01
3.58469695e-01 -3.75550240e-02 -1.48409986e+00 -2.79724777e-01
1.67664006e-01 6.12448975e-02 -1.44019902e-01 2.62221307e-01
-4.13059235e-01 5.68417311e-01 2.76325326e-02 -7.53276289e-01
4.12454635e-01 -1.59000784e-01 4.61183220e-01 1.13175416e+00
-7.84698650e-02 1.26285538e-01 5.80473781e-01 1.02471292e+00
-2.70583421e-01 1.17752820e-01 -7.11644351e-01 -1.95033342e-01
5.92373073e-01 1.13136387e+00 -1.18266396e-01 -2.79847056e-01
-8.01359564e-02 9.60708857e-01 5.40658295e-01 2.03751832e-01
-6.86080813e-01 -1.94292456e-01 7.36319900e-01 -1.23439334e-01
1.71482265e-01 -1.69973388e-01 -2.19433770e-01 -9.67728972e-01
-4.67081159e-01 -1.28898478e+00 3.70978743e-01 -4.97812480e-01
-1.53415716e+00 5.94834924e-01 -1.26536831e-01 -7.11083472e-01
-1.20590758e+00 -4.20115381e-01 -7.39902020e-01 8.44702840e-01
-1.05610204e+00 -8.41571987e-01 1.26825497e-01 5.21369100e-01
1.06083024e+00 -3.37886244e-01 1.00664711e+00 -2.56823361e-01
-2.06674755e-01 7.47403800e-01 6.06289227e-03 5.79958111e-02
1.02354753e+00 -1.47582138e+00 3.50661576e-01 1.71067044e-01
5.86114638e-02 3.72803897e-01 1.03675222e+00 -5.96998513e-01
-9.22067821e-01 -6.88650370e-01 8.92472506e-01 -6.18687928e-01
8.03036392e-01 -5.32115519e-01 -7.95428455e-01 5.28568923e-01
5.41757345e-01 -8.05726707e-01 8.05796862e-01 5.17999589e-01
-1.90139309e-01 3.09862226e-01 -1.12952220e+00 9.88372982e-01
6.03330612e-01 -5.29258788e-01 -8.11243117e-01 6.48740411e-01
1.08755016e+00 -5.18246114e-01 -9.33697522e-01 -1.56516418e-01
3.81794214e-01 -8.26869011e-01 7.53508210e-01 -7.90505052e-01
7.19534457e-01 4.10334826e-01 8.53389725e-02 -1.81313980e+00
-2.32941404e-01 -1.19122875e+00 -1.96381614e-01 1.34370995e+00
8.10530841e-01 -3.05424094e-01 7.32743502e-01 8.71116877e-01
-1.40930668e-01 -7.23280251e-01 -5.24432719e-01 -5.89146316e-01
5.48800170e-01 -1.13287922e-02 4.18312132e-01 7.33916461e-01
6.98250413e-01 1.19286561e+00 -8.17799985e-01 -2.70126432e-01
3.18151623e-01 -4.27668635e-03 9.40401137e-01 -9.59272742e-01
-6.22341275e-01 -5.64156115e-01 4.48816478e-01 -1.37690508e+00
4.55645412e-01 -5.96580565e-01 5.87774277e-01 -1.31360292e+00
-2.26990476e-01 -4.15158510e-01 1.92063317e-01 1.97701633e-01
-3.83187056e-01 -2.82172620e-01 4.41102862e-01 5.96449636e-02
-5.63386202e-01 8.84218156e-01 1.36325300e+00 -8.07869285e-02
-6.21340573e-01 2.46017143e-01 -9.34267521e-01 5.67308962e-01
8.09014440e-01 -3.21800560e-01 -6.11553848e-01 1.56580878e-03
-1.04485795e-01 6.53773367e-01 6.44976050e-02 -5.57186663e-01
1.07894339e-01 -5.05943835e-01 -3.41140144e-02 -1.62176773e-01
7.26194501e-01 -1.61286458e-01 -4.14299339e-01 2.86705047e-01
-1.07542694e+00 -8.76888633e-02 -1.28884956e-01 5.47664285e-01
-9.94520932e-02 -5.14965117e-01 8.44756663e-01 -4.07500327e-01
-1.78490415e-01 -3.95616852e-02 -7.06482649e-01 5.84211826e-01
7.18016803e-01 2.07263052e-01 -3.68265957e-01 -1.20600688e+00
-7.85465121e-01 6.71125174e-01 2.98513293e-01 7.11627007e-01
4.46423531e-01 -1.05465674e+00 -7.48485923e-01 -1.74904883e-01
3.60389128e-02 -2.29769617e-01 -2.68693678e-02 1.35311067e-01
-8.13826025e-02 2.49471694e-01 3.60631719e-02 -3.69147837e-01
-1.07249236e+00 3.74848813e-01 4.14094001e-01 -7.90263057e-01
-3.51551026e-01 8.73353660e-01 1.48527935e-01 -8.37326467e-01
5.23944497e-01 1.73746049e-02 -2.78886557e-01 1.51290223e-01
3.41168880e-01 1.32610491e-02 -3.40439886e-01 -2.43571207e-01
2.12335110e-01 -1.81508034e-01 -2.31630936e-01 -8.59387159e-01
9.59609568e-01 -1.13889024e-01 4.00686324e-01 6.46406293e-01
9.16010857e-01 -8.78203735e-02 -1.69029725e+00 -2.04553083e-01
7.13960305e-02 -2.07283720e-01 -4.08753902e-01 -1.16544116e+00
-2.78105259e-01 8.36232603e-01 2.38432974e-01 5.05462587e-01
3.86878431e-01 6.11268841e-02 9.53260660e-01 6.25482738e-01
1.24358982e-01 -1.43654895e+00 8.48242342e-01 9.71829414e-01
9.91953075e-01 -1.43883991e+00 -4.35348243e-01 1.61603820e-02
-1.45966029e+00 9.13108230e-01 1.03341687e+00 5.58199510e-02
3.66419524e-01 5.74524999e-02 4.48296398e-01 -2.48629004e-02
-1.18853021e+00 1.34058058e-01 -5.54636009e-02 4.70184088e-01
7.09801555e-01 2.10360259e-01 -1.63764462e-01 3.97118539e-01
-5.93146145e-01 -3.10244620e-01 5.76074898e-01 7.97036648e-01
-6.18090451e-01 -1.33767462e+00 -3.15876842e-01 4.01381016e-01
-4.42143112e-01 -5.36341369e-02 -7.87075937e-01 3.99649382e-01
-5.35792828e-01 1.40791047e+00 -1.57793745e-01 -3.48497152e-01
2.18127415e-01 5.23450077e-01 2.66570657e-01 -8.24011207e-01
-1.09078169e+00 1.86472550e-01 6.67314053e-01 -2.42268652e-01
5.07766753e-02 -6.19434893e-01 -1.21568787e+00 -2.60473937e-01
-4.63786304e-01 4.09828126e-01 4.60691512e-01 9.88940060e-01
1.35794073e-01 3.93950552e-01 9.79483306e-01 -7.91680932e-01
-1.53756440e+00 -1.28770614e+00 -3.66828978e-01 6.40901089e-01
2.42654219e-01 -5.21221101e-01 -3.92040074e-01 -1.48804232e-01] | [12.82711124420166, 8.106203079223633] |
dcdcedff-ab0e-46d3-8f8b-fc4a8927bf5f | the-portiloop-a-deep-learning-based-open | 2107.13473 | null | https://arxiv.org/abs/2107.13473v3 | https://arxiv.org/pdf/2107.13473v3.pdf | The Portiloop: a deep learning-based open science tool for closed-loop brain stimulation | Closed-loop brain stimulation refers to capturing neurophysiological measures such as electroencephalography (EEG), quickly identifying neural events of interest, and producing auditory, magnetic or electrical stimulation so as to interact with brain processes precisely. It is a promising new method for fundamental neuroscience and perhaps for clinical applications such as restoring degraded memory function; however, existing tools are expensive, cumbersome, and offer limited experimental flexibility. In this article, we propose the Portiloop, a deep learning-based, portable and low-cost closed-loop stimulation system able to target specific brain oscillations. We first document open-hardware implementations that can be constructed from commercially available components. We also provide a fast, lightweight neural network model and an exploration algorithm that automatically optimizes the model hyperparameters to the desired brain oscillation. Finally, we validate the technology on a challenging test case of real-time sleep spindle detection, with results comparable to off-line expert performance on the Massive Online Data Annotation spindle dataset (MODA; group consensus). Software and plans are available to the community as an open science initiative to encourage further development and advance closed-loop neuroscience research. | ['Emily B. J. Coffey', 'Milo Sobral', "Xavier L'Heureux", 'Giovanni Beltrame', 'Hugo R. Jourde', 'Yann Bouteiller', 'Nicolas Valenchon'] | 2021-07-28 | null | null | null | null | ['spindle-detection'] | ['medical'] | [ 7.82211646e-02 -8.43652897e-03 1.41968504e-01 -2.31651604e-01
-4.43374366e-01 -5.90473831e-01 1.40198022e-01 2.07279567e-02
-3.59559268e-01 8.31108749e-01 2.19340697e-01 -3.47138166e-01
-4.33471471e-01 -1.31252468e-01 -5.99325359e-01 -6.06095314e-01
-3.73277903e-01 6.19087756e-01 2.09300563e-01 -4.41894680e-03
5.52064598e-01 3.72549534e-01 -1.75227487e+00 7.70288333e-02
7.28664994e-01 9.62331474e-01 6.14808083e-01 2.82333553e-01
3.66820306e-01 1.90949276e-01 -7.90651500e-01 3.75627100e-01
-1.57054350e-01 -4.55521762e-01 -6.36803448e-01 -3.00840765e-01
-1.25781745e-01 1.63916126e-01 8.04352574e-03 9.98924196e-01
1.26725304e+00 -1.66310117e-01 4.12438884e-02 -1.17154980e+00
-2.20033467e-01 5.92106402e-01 -2.16470122e-01 8.38650584e-01
4.24019814e-01 3.80047411e-01 3.43573660e-01 -7.09767938e-01
5.35909474e-01 5.35690665e-01 5.91979861e-01 5.85968435e-01
-1.49608719e+00 -1.02548468e+00 -4.09188926e-01 5.45630395e-01
-1.30544102e+00 -1.02140903e+00 5.49368083e-01 -3.30560952e-01
1.54338622e+00 1.27934992e-01 1.08599126e+00 1.45001996e+00
7.97262073e-01 2.48867944e-01 1.11492407e+00 -2.23965824e-01
7.81888306e-01 -9.78429690e-02 1.44124582e-01 3.19884509e-01
1.95689410e-01 -7.64314905e-02 -1.24431157e+00 -3.61672014e-01
5.66560745e-01 -4.22466964e-01 -6.47550404e-01 -8.87528285e-02
-1.38767421e+00 3.34113419e-01 1.77674349e-02 5.77752173e-01
-6.88656867e-01 1.72081128e-01 5.10166645e-01 3.64998817e-01
3.45941037e-01 9.57207859e-01 -7.68658459e-01 -6.89154565e-01
-1.11288357e+00 3.09002865e-02 7.44448245e-01 5.82974434e-01
5.29249668e-01 2.14147985e-01 -1.02010556e-01 6.94813251e-01
-7.90740177e-02 9.21705663e-02 1.14438367e+00 -1.02028573e+00
-2.07265750e-01 5.03454864e-01 7.96026364e-03 -6.59242988e-01
-1.36241436e+00 -5.83391070e-01 -7.42799819e-01 8.92936662e-02
-3.15600447e-02 -2.76273072e-01 -3.20766270e-01 1.58539844e+00
-9.15169641e-02 4.60888416e-01 -2.68505752e-01 1.01645768e+00
7.51396656e-01 2.95649529e-01 -1.84919223e-01 -4.49599296e-01
1.65732777e+00 -4.20767695e-01 -1.07463431e+00 -5.55026889e-01
5.52458584e-01 -3.02817553e-01 1.30513465e+00 7.07920551e-01
-1.22360671e+00 -7.98705518e-02 -1.09851050e+00 2.75357723e-01
-3.96847576e-01 1.25576124e-01 7.13254452e-01 5.60599625e-01
-1.52174997e+00 5.22513926e-01 -1.31819177e+00 -5.48576474e-01
5.05352855e-01 8.07827890e-01 -2.46192798e-01 6.89741373e-01
-1.01538074e+00 1.32018387e+00 3.24864596e-01 -1.14909798e-01
-8.35267067e-01 -9.42179501e-01 -2.60515124e-01 2.82537341e-01
8.89516994e-02 -6.25286102e-01 1.07804596e+00 -6.40644312e-01
-1.69994843e+00 1.00154006e+00 -5.06085232e-02 -7.53667474e-01
-4.78351980e-01 9.31860879e-02 -4.18556899e-01 1.03254952e-01
-4.01121266e-02 8.59113276e-01 6.64349735e-01 -5.06934583e-01
-3.27126402e-03 -6.02652371e-01 -3.11236173e-01 -1.70698557e-02
-5.62777162e-01 3.67260635e-01 -9.61528718e-02 -4.25291657e-01
-4.71386723e-02 -8.18218946e-01 2.65345037e-01 -2.57989883e-01
-2.17729792e-01 -4.63769101e-02 5.15087485e-01 -6.16789460e-01
1.10668671e+00 -2.05493331e+00 3.34131032e-01 -3.67711037e-02
2.01958209e-01 1.53706837e-02 -7.30639473e-02 2.97740817e-01
-4.51628119e-01 -1.39626175e-01 -1.89206868e-01 1.78508516e-02
6.14587776e-02 -2.48231202e-01 -9.88671258e-02 6.22713029e-01
-6.20830171e-02 8.39367628e-01 -6.93961263e-01 1.55722111e-01
2.90237591e-02 2.83115000e-01 -4.15750682e-01 2.39152372e-01
2.81704664e-02 5.01369119e-01 3.56818676e-01 6.42271340e-01
2.62955487e-01 -3.16065609e-01 9.04971585e-02 -3.24594140e-01
-2.51138389e-01 5.25865078e-01 -8.50563109e-01 2.07529593e+00
-4.80394423e-01 1.00369811e+00 3.01319331e-01 -1.10810435e+00
8.26069593e-01 5.51818013e-01 5.20968854e-01 -7.79439926e-01
5.00128984e-01 2.66063422e-01 2.52195686e-01 -7.48515189e-01
-2.26524919e-01 1.97045699e-01 2.57885933e-01 7.89395392e-01
4.71699834e-01 -2.51459539e-01 2.10536644e-01 -9.43428203e-02
1.61739910e+00 -9.92076248e-02 4.70276147e-01 -7.93937266e-01
5.46456203e-02 -2.57787555e-01 2.77980059e-01 3.42890739e-01
-2.01489210e-01 4.27766949e-01 4.97932792e-01 -2.91386604e-01
-6.30291700e-01 -7.25153685e-01 -3.72525632e-01 1.08001482e+00
-1.02633394e-01 -4.34865236e-01 -1.01205766e+00 2.02069089e-01
-5.10919392e-01 7.11137652e-01 -5.37301540e-01 -4.47512209e-01
1.40833296e-02 -9.42670286e-01 5.41584074e-01 3.48207295e-01
2.22785994e-01 -1.59834242e+00 -1.26149762e+00 3.56237531e-01
-3.87305737e-01 -9.97451961e-01 -3.19464684e-01 9.39596593e-01
-8.21981072e-01 -8.96070898e-01 -1.76376745e-01 -8.36343288e-01
2.94934183e-01 -1.84793651e-01 1.00664806e+00 -5.93447015e-02
-7.46140957e-01 1.34388492e-01 -6.47507608e-02 -5.29580414e-01
1.17818676e-01 2.31126711e-01 5.31864643e-01 -4.02029723e-01
6.43972576e-01 -1.21481776e+00 -6.74562752e-01 9.87027287e-02
-7.43093371e-01 1.17225952e-01 4.97517079e-01 6.91638350e-01
4.79700059e-01 -4.00115661e-02 1.14909887e+00 -3.24496388e-01
1.09687757e+00 -4.68831658e-01 -6.27759814e-01 -8.54582489e-02
-6.21519744e-01 -1.22249573e-02 4.86184329e-01 -4.20497537e-01
-5.79588115e-01 1.34896114e-01 -1.81360200e-01 -2.45879039e-01
-2.96043932e-01 6.22611701e-01 -1.85919844e-03 -1.64642990e-01
1.09740210e+00 4.81163353e-01 -1.11113816e-01 -1.36564061e-01
-1.73851743e-01 6.97845638e-01 8.21027756e-01 -5.22825569e-02
-3.16574350e-02 2.27858678e-01 -2.56681502e-01 -6.58610225e-01
-4.46008921e-01 -3.32263380e-01 -3.49910706e-01 -2.29170293e-01
6.80045128e-01 -8.39359045e-01 -1.05507004e+00 4.53059822e-01
-1.16772509e+00 -6.48115337e-01 1.84280097e-01 4.92802203e-01
-8.26778829e-01 -2.53696024e-01 -3.25938612e-01 -5.01777947e-01
-8.76127183e-01 -9.17257130e-01 1.06081414e+00 3.39016020e-01
-6.33897543e-01 -7.60883987e-01 3.45855594e-01 5.08613177e-02
5.88655412e-01 -2.18481779e-01 7.48610973e-01 -6.66556418e-01
-1.05161048e-01 -1.81686543e-02 1.21198826e-01 -8.09117854e-02
-2.48826984e-02 -3.82761598e-01 -1.07348275e+00 -3.24564278e-01
3.64338011e-01 -5.63837767e-01 1.65581509e-01 6.10416710e-01
1.62750804e+00 -5.99860400e-02 -4.41603333e-01 6.89206719e-01
1.11659932e+00 1.56845838e-01 6.88344300e-01 4.75663066e-01
-2.80712638e-02 4.07047212e-01 -1.13029279e-01 5.68489552e-01
4.86484244e-02 5.78476369e-01 3.14710736e-01 2.39661366e-01
1.01979278e-01 3.79164785e-01 3.74009907e-01 8.38993609e-01
1.87212393e-01 9.99899209e-02 -1.08858860e+00 5.68940580e-01
-1.59642208e+00 -8.45644593e-01 3.18455361e-02 2.10522604e+00
9.99187469e-01 6.85266405e-02 -9.20448918e-03 1.73999965e-01
4.79797155e-01 -4.73751575e-01 -8.28218460e-01 -3.50769669e-01
-7.89545700e-02 6.53182864e-01 2.38784730e-01 -3.58465612e-02
-4.20080662e-01 6.82828963e-01 6.72181606e+00 6.47358000e-01
-1.37393641e+00 5.53294659e-01 1.37409717e-01 -7.05592275e-01
8.18898082e-02 -4.10515815e-01 -3.80409956e-01 7.76861548e-01
1.63471508e+00 -4.09780532e-01 1.23928714e+00 4.22531515e-01
8.56146812e-01 -3.38075906e-01 -1.12251461e+00 1.37748051e+00
-1.36816353e-01 -1.69936252e+00 -8.38026583e-01 -1.42615899e-01
2.66816348e-01 5.48278153e-01 -2.68642694e-01 7.71034807e-02
-2.62441337e-01 -9.34186578e-01 5.96344113e-01 5.91473401e-01
7.61691689e-01 -7.37331390e-01 5.13623178e-01 5.41013956e-01
-6.27263725e-01 -3.25594544e-01 -1.98495016e-01 -1.35047063e-01
-8.94455835e-02 7.86415935e-01 -5.85027874e-01 -9.64505449e-02
1.05328250e+00 4.78742301e-01 -6.37153924e-01 1.41631949e+00
-2.11037938e-02 7.89409697e-01 -3.25880736e-01 -1.89428344e-01
-4.14933503e-01 1.19114973e-01 5.73813558e-01 1.13340592e+00
5.61970353e-01 1.02704667e-01 -4.10478652e-01 1.17925727e+00
-2.84686610e-02 -2.35670716e-01 -5.07974982e-01 -1.67088479e-01
7.30529308e-01 1.59788227e+00 -1.10929835e+00 6.40194267e-02
3.72926444e-02 9.09899116e-01 3.30058396e-01 1.53788388e-01
-7.82962024e-01 -6.19102061e-01 4.46328491e-01 -5.25987297e-02
-1.77477673e-01 -1.42866760e-01 -6.78624272e-01 -9.69114244e-01
-2.27036364e-02 -9.48637724e-01 -2.22884610e-01 -1.34087157e+00
-8.96395147e-01 8.05802643e-01 -1.20193794e-01 -8.58368158e-01
-3.24696869e-01 -4.67599452e-01 -7.61570394e-01 7.75879562e-01
-8.88868988e-01 -4.11578476e-01 -4.28873360e-01 6.05833530e-01
4.39644992e-01 -1.53862864e-01 1.16430342e+00 4.19096321e-01
-6.89121902e-01 1.26771197e-01 -4.40538377e-02 -6.63096130e-01
7.72912741e-01 -1.01625776e+00 1.14939407e-01 5.51726341e-01
-9.38031543e-03 6.15202248e-01 9.07887936e-01 -4.02178705e-01
-1.60458708e+00 -7.73818493e-01 7.37367928e-01 -1.37944445e-01
8.47763181e-01 -9.20521677e-01 -9.44702566e-01 4.44028676e-01
5.24737060e-01 -2.56015122e-01 7.92691827e-01 1.35986999e-01
4.04533088e-01 -2.99793929e-01 -1.09484720e+00 4.71384645e-01
1.02523339e+00 -4.54806238e-01 -5.93109548e-01 7.03496873e-01
4.56803083e-01 -3.79433483e-01 -6.30580604e-01 1.19457088e-01
3.77992183e-01 -1.00370669e+00 4.64332759e-01 1.39884315e-02
-9.69105810e-02 -1.39013097e-01 4.44013178e-01 -1.79160774e+00
-3.27434957e-01 -1.04675853e+00 -2.97027528e-01 7.20764816e-01
3.16198945e-01 -8.06266487e-01 5.09185433e-01 3.03057134e-01
-6.09615982e-01 -7.92473614e-01 -1.16304076e+00 -7.74477541e-01
-3.57024133e-01 -6.04682028e-01 4.41979557e-01 5.88318408e-01
9.35883105e-01 4.75269407e-01 1.09755318e-03 7.77589530e-02
2.06665337e-01 -2.58711576e-01 2.10927531e-01 -1.23589873e+00
-2.25409761e-01 -5.97204447e-01 -4.13773775e-01 -4.60620195e-01
3.70142132e-01 -9.58692253e-01 2.80006051e-01 -1.35799789e+00
2.42623761e-01 9.00137275e-02 -2.33658001e-01 7.95930803e-01
2.55155087e-01 4.45531338e-01 -4.40255642e-01 4.98013571e-02
-3.63398314e-01 4.31468099e-01 6.74212039e-01 1.03439964e-01
-4.25358891e-01 -2.55048394e-01 -9.09206510e-01 6.45471096e-01
1.06857312e+00 -6.54033899e-01 -5.84226549e-01 -3.88275415e-01
3.26607168e-01 1.22183137e-01 4.60659742e-01 -1.38027430e+00
7.75858402e-01 3.22941959e-01 3.82215679e-01 -2.66362101e-01
3.14820051e-01 -5.86624086e-01 2.16276258e-01 4.66763228e-01
-2.56924987e-01 1.27797410e-01 6.25597775e-01 2.38304570e-01
1.94952667e-01 -1.13712490e-01 7.88898289e-01 9.99775901e-02
-4.11789775e-01 3.56677510e-02 -1.02505982e+00 -1.27069309e-01
9.08251762e-01 -1.37788191e-01 -5.22016168e-01 -3.26644093e-01
-8.63752067e-01 5.45386113e-02 1.51672259e-01 4.79434997e-01
4.58010375e-01 -1.07316065e+00 -4.80543315e-01 5.26599169e-01
-9.45467576e-02 -5.49153686e-01 7.64223263e-02 1.29026389e+00
-1.76672623e-01 6.48138046e-01 -7.47398257e-01 -7.19142795e-01
-1.02174163e+00 4.51228529e-01 4.82928157e-01 3.01320255e-01
-7.02358305e-01 6.89666331e-01 -2.38113970e-01 -2.37179607e-01
4.25307184e-01 -3.55477124e-01 -1.90082416e-01 1.57981608e-02
7.45766163e-01 2.22839445e-01 9.97930348e-01 3.42668593e-02
-7.16037750e-01 -6.89240247e-02 5.44663489e-01 -2.50277042e-01
1.57916021e+00 -2.51366086e-02 -4.40149158e-01 6.19741976e-01
8.08438361e-01 -5.55296600e-01 -9.29252565e-01 4.29643154e-01
-1.34106549e-02 1.69771835e-01 4.98912305e-01 -1.07531393e+00
-1.00488210e+00 7.59172499e-01 1.04386032e+00 3.05324316e-01
1.63951790e+00 4.41008098e-02 5.08300662e-01 6.66353583e-01
6.28032982e-01 -1.35067499e+00 1.42087266e-01 2.58332342e-01
1.14242971e+00 -5.85935831e-01 -1.56669855e-01 2.15647101e-01
-2.53414780e-01 1.17378294e+00 7.43070900e-01 -5.37002794e-02
7.22435474e-01 8.31509292e-01 -2.11365700e-01 -6.76096797e-01
-1.28866136e+00 2.00372577e-01 -5.29875234e-02 8.38941813e-01
4.90361005e-01 -1.87959179e-01 -4.91918981e-01 1.06861293e+00
-3.82060140e-01 5.42321920e-01 5.96394658e-01 7.41986573e-01
-5.64147413e-01 -6.15971148e-01 -4.39186841e-01 9.81355190e-01
-2.61201411e-01 -2.65205145e-01 -2.96899647e-01 1.68148592e-01
1.09819226e-01 9.38701808e-01 2.54518181e-01 -3.83884609e-01
2.00726688e-01 3.58444959e-01 6.90476835e-01 -7.62662232e-01
-7.34241545e-01 1.07725851e-01 -1.11932836e-01 -8.33665848e-01
-2.14166075e-01 -7.18236744e-01 -1.48071003e+00 -7.80843124e-02
-3.21390390e-01 -1.25490710e-01 1.01976764e+00 1.17093873e+00
9.90853727e-01 8.40168059e-01 1.67722926e-01 -1.27018273e+00
-1.79955333e-01 -1.22232771e+00 -6.58499837e-01 -3.16385418e-01
4.83371019e-02 -8.41662765e-01 -3.83592635e-01 2.04428300e-01] | [13.31008529663086, 3.44884991645813] |
d6a52b4a-9c39-403c-b228-dbc0031be850 | clotho-an-audio-captioning-dataset | 1910.09387 | null | https://arxiv.org/abs/1910.09387v1 | https://arxiv.org/pdf/1910.09387v1.pdf | Clotho: An Audio Captioning Dataset | Audio captioning is the novel task of general audio content description using free text. It is an intermodal translation task (not speech-to-text), where a system accepts as an input an audio signal and outputs the textual description (i.e. the caption) of that signal. In this paper we present Clotho, a dataset for audio captioning consisting of 4981 audio samples of 15 to 30 seconds duration and 24 905 captions of eight to 20 words length, and a baseline method to provide initial results. Clotho is built with focus on audio content and caption diversity, and the splits of the data are not hampering the training or evaluation of methods. All sounds are from the Freesound platform, and captions are crowdsourced using Amazon Mechanical Turk and annotators from English speaking countries. Unique words, named entities, and speech transcription are removed with post-processing. Clotho is freely available online (https://zenodo.org/record/3490684). | ['Konstantinos Drossos', 'Tuomas Virtanen', 'Samuel Lipping'] | 2019-10-21 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 2.16287404e-01 2.50215232e-01 4.88281772e-02 -3.12909573e-01
-1.64597714e+00 -1.11879456e+00 4.02047127e-01 3.48157971e-03
-2.25659981e-01 7.53132641e-01 8.89369369e-01 -6.01823367e-02
3.65771770e-01 7.14516826e-03 -7.61528373e-01 -3.38376164e-01
1.56984068e-02 7.06459343e-01 -1.14879005e-01 -1.96622148e-01
-2.14767382e-01 -1.13148093e-01 -1.43417096e+00 6.04443729e-01
3.46015960e-01 1.10843480e+00 1.93489507e-01 1.02401102e+00
-9.74495858e-02 5.68129897e-01 -8.35824847e-01 -6.18432164e-01
6.48866175e-03 -5.36724567e-01 -9.70191479e-01 -2.45146245e-01
4.56658065e-01 -4.46895324e-02 -2.29438499e-01 6.90143347e-01
1.00721979e+00 -2.64724847e-02 4.12421793e-01 -1.55499661e+00
-5.92580497e-01 1.15134704e+00 1.62217632e-01 9.74557027e-02
9.29914534e-01 -2.34534852e-02 1.30125988e+00 -1.35913825e+00
7.08974242e-01 1.07610416e+00 6.61148190e-01 6.37107790e-01
-1.18502462e+00 -8.65388393e-01 -3.74579638e-01 9.21081528e-02
-1.75009620e+00 -1.19409263e+00 4.56608802e-01 -6.48262322e-01
7.60981739e-01 4.83313441e-01 4.24280226e-01 1.72011161e+00
-4.05139178e-01 6.62470996e-01 6.69972777e-01 -4.73090261e-01
1.11480944e-01 2.48162895e-01 -1.49763674e-01 6.52171746e-02
-3.72319221e-01 -2.58422136e-01 -1.20859730e+00 -4.23568964e-01
1.26635656e-01 -9.92722988e-01 -5.05632222e-01 2.50488073e-01
-1.78375781e+00 4.93154794e-01 -2.99848448e-02 3.31747234e-02
-3.68260324e-01 1.86102673e-01 6.79097533e-01 4.02924329e-01
3.99921417e-01 4.97161806e-01 -3.95607293e-01 -6.27225220e-01
-8.72160912e-01 3.11489940e-01 8.16055894e-01 1.32168186e+00
4.18629229e-01 1.61092028e-01 -2.92009324e-01 1.05275285e+00
8.30972195e-02 1.08752501e+00 6.63741350e-01 -1.03732574e+00
8.32440078e-01 -2.55094707e-01 5.49902081e-01 -6.92420900e-01
-2.01174766e-01 5.37306517e-02 -2.81547695e-01 -5.23553848e-01
1.70042962e-01 -5.16021073e-01 -7.36623585e-01 1.82356000e+00
1.68218594e-02 -4.02655154e-02 1.01901785e-01 9.47644532e-01
1.29372859e+00 9.94459212e-01 2.97830123e-02 -1.49245799e-01
1.44306016e+00 -8.89404714e-01 -1.08642447e+00 -2.71684527e-01
4.06965911e-01 -1.27094245e+00 1.20875967e+00 2.48386815e-01
-1.16392601e+00 -4.65639770e-01 -7.07992315e-01 -1.11408673e-01
-3.30133229e-01 1.50460169e-01 4.63946871e-02 4.64199215e-01
-9.98647928e-01 1.41332978e-02 -4.13516998e-01 -2.00737342e-01
-4.52949405e-02 5.73023148e-02 -6.24755144e-01 2.73639828e-01
-1.65680671e+00 5.43186307e-01 3.16168070e-01 -1.45651745e-02
-8.66121590e-01 -6.28120005e-01 -1.01870108e+00 -1.79200709e-01
2.20330104e-01 -3.42618495e-01 1.87429011e+00 -1.08120406e+00
-1.55992079e+00 7.92598844e-01 -3.55467975e-01 -4.08548057e-01
2.66044289e-01 -2.60772437e-01 -7.71413863e-01 4.25785810e-01
4.68635321e-01 1.03227162e+00 9.40967858e-01 -9.07659769e-01
-3.73542786e-01 2.19070718e-01 -3.81209970e-01 3.36379439e-01
-1.40897915e-01 5.99605680e-01 -3.66785109e-01 -8.76848876e-01
-1.70864463e-01 -1.12881124e+00 4.55847919e-01 -2.82024860e-01
-6.47921503e-01 4.85927574e-02 3.00804436e-01 -9.87804949e-01
1.15724599e+00 -2.54555821e+00 -7.51921013e-02 -1.75931960e-01
-3.08060676e-01 -1.32144436e-01 -4.58426923e-01 8.11109483e-01
-2.18860477e-01 2.56203771e-01 -1.92821056e-01 -6.75709367e-01
3.53701651e-01 -1.36153802e-01 -8.22866440e-01 2.66312182e-01
2.17789412e-01 8.69278252e-01 -9.98203337e-01 -6.11213923e-01
-8.06046426e-02 5.69879889e-01 -2.05716968e-01 3.30251634e-01
-7.09322020e-02 5.03111660e-01 8.17939192e-02 6.03785694e-01
2.96120256e-01 1.71435550e-01 -2.72719294e-01 -2.57672146e-02
-1.28334537e-01 8.17178607e-01 -1.08343744e+00 1.78033876e+00
-5.68418205e-01 1.10714340e+00 3.78415197e-01 -1.72774196e-01
8.46895576e-01 1.09364378e+00 3.95873964e-01 -3.60970616e-01
8.32859799e-02 5.25148034e-01 -3.31566423e-01 -6.75589740e-01
7.35409975e-01 -6.97696954e-02 -6.07477903e-01 5.10920763e-01
3.70675921e-01 -6.42923295e-01 1.50369719e-01 2.01473266e-01
8.89630914e-01 -9.31853727e-02 1.30486563e-01 1.55648559e-01
1.03766650e-01 2.33556479e-01 2.56132364e-01 5.62760115e-01
-1.48552790e-01 1.18057549e+00 2.21305996e-01 2.12760985e-01
-1.26716948e+00 -1.09413230e+00 -2.95378596e-01 1.30572963e+00
-3.52917820e-01 -7.81423688e-01 -8.18854332e-01 2.54562683e-02
-1.84752092e-01 7.05289483e-01 -4.79265064e-01 1.94363669e-01
-3.49048078e-01 9.01098698e-02 1.11881518e+00 2.85875440e-01
2.51480304e-02 -1.40116620e+00 -5.40325157e-02 3.27982098e-01
-1.03109539e+00 -1.48761988e+00 -1.04894054e+00 6.67404756e-02
-5.99009171e-02 -6.00172222e-01 -1.09519327e+00 -9.24499094e-01
9.05824229e-02 -5.69507033e-02 1.13304341e+00 -6.61515415e-01
7.81170800e-02 5.26864409e-01 -7.55438685e-01 -6.51575446e-01
-7.79808044e-01 3.39175195e-01 3.55420470e-01 1.16574824e-01
2.62177438e-01 -4.83298779e-01 -3.31107080e-02 3.28617930e-01
-4.74029034e-01 1.42022669e-01 2.87776113e-01 6.51503682e-01
5.77246428e-01 -7.47783422e-01 9.05433714e-01 -3.05592746e-01
8.87120962e-01 -6.11503482e-01 -1.07583985e-01 -8.56096148e-02
-5.35716787e-02 -4.54716831e-01 6.85046315e-01 -6.15184247e-01
-5.95214963e-01 2.17825100e-01 -1.89289927e-01 -4.54526514e-01
-3.48224819e-01 5.74099839e-01 -1.75805137e-01 5.44467628e-01
8.15351069e-01 2.33812183e-01 -1.15697443e-01 -6.07086420e-01
5.46284676e-01 1.52282774e+00 1.09407640e+00 -5.02401829e-01
6.06392562e-01 -7.49311270e-03 -6.87216759e-01 -8.96921575e-01
-7.68940270e-01 -4.69934136e-01 -4.76356715e-01 -3.40216249e-01
7.37194180e-01 -1.35446632e+00 -3.92989546e-01 3.23642373e-01
-1.34103560e+00 -2.33175918e-01 -3.38583946e-01 6.14284456e-01
-7.33707905e-01 -5.98875992e-02 -4.88697171e-01 -7.75011301e-01
-5.52897751e-01 -8.55738342e-01 1.26745868e+00 -1.90376922e-01
-7.74381280e-01 -3.84254456e-01 1.30176291e-01 4.16191280e-01
2.59819061e-01 1.68111652e-01 2.12903038e-01 -9.32737947e-01
1.27192080e-01 -2.28681833e-01 6.14735000e-02 2.96748877e-01
8.53745639e-02 2.96113342e-02 -1.39508951e+00 -4.99534234e-02
-4.47845101e-01 -7.51565337e-01 3.62751842e-01 1.00346595e-01
6.99161589e-01 -6.36697590e-01 1.37986869e-01 1.96516797e-01
7.14034677e-01 1.21550955e-01 2.17542171e-01 9.71026346e-02
4.14003760e-01 7.29322493e-01 6.17821574e-01 5.74424982e-01
2.97988892e-01 9.30479050e-01 1.80809125e-01 1.67727128e-01
-3.53836238e-01 -6.22522235e-01 8.19482565e-01 1.27239478e+00
5.49787343e-01 -3.99750739e-01 -1.21827602e+00 1.01423526e+00
-1.41293430e+00 -9.43843663e-01 -1.55405506e-01 2.10703254e+00
1.27892947e+00 -1.44629195e-01 3.29839677e-01 1.90534756e-01
1.10131466e+00 4.92469035e-02 -1.85951948e-01 -5.20583093e-01
-2.01021835e-01 6.23125993e-02 2.40319744e-01 6.93681359e-01
-9.42643762e-01 7.79324234e-01 6.31849480e+00 7.25335121e-01
-1.10571420e+00 3.67072135e-01 6.83920400e-04 -3.54148865e-01
-3.03730041e-01 -3.01170737e-01 -4.85833108e-01 6.34404004e-01
1.59452534e+00 -4.37970370e-01 7.28928328e-01 4.64525253e-01
3.53320390e-01 2.57303655e-01 -1.15887094e+00 1.12606835e+00
1.94908574e-01 -1.02332699e+00 -1.18922099e-01 -3.66186768e-01
4.11683708e-01 3.88167858e-01 1.13537669e-01 2.41716921e-01
-1.36037588e-01 -1.02044761e+00 1.71454942e+00 3.22265863e-01
1.34638393e+00 -4.76619244e-01 7.04505980e-01 3.23358066e-02
-1.17995524e+00 1.22152969e-01 5.38268834e-02 8.23413283e-02
5.85563004e-01 2.88935483e-01 -1.23668325e+00 4.20854717e-01
8.07266653e-01 4.70513880e-01 -4.99514937e-01 1.10674989e+00
-2.99044728e-01 1.09393156e+00 -4.25269723e-01 -1.12095162e-01
1.21087790e-01 4.27843243e-01 9.93819714e-01 1.66056311e+00
4.44534957e-01 -1.74250588e-01 2.18867958e-02 5.29339850e-01
-3.31099063e-01 4.35576469e-01 -6.38982475e-01 -5.24561703e-01
1.11387408e+00 9.50191736e-01 -3.29933465e-01 -2.64926046e-01
-1.32798061e-01 9.00330365e-01 -2.68536359e-01 4.01141703e-01
-9.28495467e-01 -9.03191030e-01 4.90187109e-01 1.06607430e-01
1.81714460e-01 -6.42495304e-02 4.11086008e-02 -1.07859433e+00
2.37310827e-01 -9.45677698e-01 1.73875719e-01 -1.49343359e+00
-1.16672468e+00 9.66510594e-01 4.65225764e-02 -1.55905294e+00
-6.45690918e-01 -1.81502402e-01 -1.67170316e-01 9.56044793e-01
-1.06969178e+00 -1.09904623e+00 -1.96198747e-01 4.62317705e-01
5.94386995e-01 -1.33049682e-01 1.09613335e+00 5.25526702e-01
-1.53750136e-01 6.09835267e-01 4.59930487e-02 3.00596446e-01
1.31917703e+00 -9.84048605e-01 6.98519528e-01 4.86072570e-01
2.93793350e-01 4.13600862e-01 1.09533989e+00 -4.42579716e-01
-9.20765340e-01 -1.21657240e+00 1.57079887e+00 -6.83004916e-01
1.03015268e+00 -9.21043336e-01 -7.10874319e-01 6.83191955e-01
4.10061747e-01 5.30406786e-03 8.69638264e-01 -1.08869888e-01
-4.16430295e-01 7.79358000e-02 -7.89815903e-01 4.24291521e-01
8.12679946e-01 -1.10368788e+00 -7.46195614e-01 4.40016806e-01
1.23565733e+00 -7.08840251e-01 -7.71772623e-01 -1.33829534e-01
5.89446902e-01 -2.07115278e-01 6.00118697e-01 -4.01218385e-01
2.34430328e-01 -3.57508481e-01 -4.88041133e-01 -1.26542163e+00
7.57139698e-02 -1.27874386e+00 1.86775953e-01 1.72901762e+00
8.29218566e-01 -3.06614965e-01 1.12945646e-01 2.45292962e-01
-5.23588657e-01 -2.18260083e-02 -1.32446945e+00 -9.55402255e-01
-4.68342341e-02 -8.40799510e-01 6.67859375e-01 9.37879562e-01
3.80981088e-01 6.02639616e-01 -4.98054653e-01 1.26126871e-01
1.25007719e-01 -7.80188739e-02 6.59914017e-01 -9.23739970e-01
3.50978076e-02 -2.21002456e-02 -2.03117087e-01 -6.57494783e-01
3.78443062e-01 -9.88957345e-01 5.18976748e-01 -1.28380263e+00
-2.28228197e-01 -1.02718867e-01 1.94978267e-01 9.83093858e-01
2.60227442e-01 6.03779733e-01 2.91178107e-01 3.78522575e-01
-5.49835861e-01 6.13626063e-01 7.64083564e-01 -3.21322531e-01
-2.44918570e-01 -8.35877955e-02 -5.31913042e-01 1.93974718e-01
8.75967264e-01 -7.97097206e-01 -1.84620261e-01 -4.84124184e-01
2.93065071e-01 2.82226026e-01 4.16314602e-01 -9.96237516e-01
1.18261442e-01 2.97679240e-03 -2.52992600e-01 -4.34159935e-01
6.99162543e-01 -5.04322708e-01 4.14403588e-01 -8.75754282e-02
-7.39026964e-01 2.98875779e-01 4.19501185e-01 1.53395727e-01
-5.20224333e-01 -4.58312988e-01 3.90123874e-01 1.66780308e-01
-3.30652177e-01 -7.16959611e-02 -7.33141482e-01 4.80551809e-01
5.08197904e-01 1.15448393e-01 -2.34249726e-01 -9.01421964e-01
-9.40955460e-01 1.07133314e-01 3.00234288e-01 7.52753139e-01
2.83222437e-01 -1.72319961e+00 -1.18643641e+00 -1.35839671e-01
5.27857840e-01 -2.60731250e-01 -3.96919660e-02 5.76421261e-01
-1.61984161e-01 7.34292269e-01 3.83468568e-02 -3.67950231e-01
-1.20056498e+00 2.41265059e-01 1.47256047e-01 6.35489643e-01
-4.60493207e-01 7.48652101e-01 -3.59430045e-01 -3.98359865e-01
4.65790302e-01 -1.60971150e-01 -9.12601426e-02 4.56320226e-01
7.05286622e-01 1.06635921e-01 2.21536830e-01 -1.17565262e+00
-4.98437971e-01 1.49758205e-01 3.98715138e-01 -1.05419695e+00
1.04442918e+00 -4.77703094e-01 8.62349570e-02 1.05454504e+00
1.22850251e+00 4.20996547e-01 -6.41128123e-01 -4.05886993e-02
-1.55261874e-01 -9.42724869e-02 -1.32557869e-01 -8.92286897e-01
-3.43122542e-01 8.56253803e-01 2.94209450e-01 3.93099278e-01
7.27780879e-01 2.14212716e-01 1.00089526e+00 3.33787113e-01
2.63048112e-01 -1.10543811e+00 -1.36874199e-01 7.48684466e-01
1.32039261e+00 -9.72875953e-01 -5.55972040e-01 -2.19429925e-01
-9.27291632e-01 9.12363589e-01 2.31732696e-01 5.04475415e-01
2.59298593e-01 2.08889678e-01 4.58556622e-01 2.22929806e-01
-7.73254156e-01 -1.63290203e-01 2.93885171e-01 7.92196751e-01
5.04192889e-01 1.44880280e-01 8.94299522e-02 1.08310997e+00
-9.95599270e-01 -1.95692137e-01 6.65793896e-01 5.44408083e-01
-1.88889518e-01 -8.50934982e-01 -6.17893755e-01 -2.84744743e-02
-6.85658157e-01 -3.76596123e-01 -7.50912607e-01 3.75270367e-01
-7.15460032e-02 1.33868897e+00 6.85886340e-03 -4.02943790e-01
3.21679085e-01 4.39221084e-01 -7.26360008e-02 -7.37203002e-01
-5.58442652e-01 5.67378178e-02 5.64483404e-01 -3.45756114e-01
-1.25059322e-01 -8.08157921e-01 -1.20547366e+00 4.21224795e-02
-2.54944742e-01 7.61848688e-01 9.80303943e-01 5.51281273e-01
5.92545629e-01 2.03088179e-01 7.05502748e-01 -9.81731117e-01
-3.16667080e-01 -1.33751798e+00 -4.20876026e-01 2.41507232e-01
6.65909410e-01 -2.84639448e-01 -7.03166902e-01 5.23518085e-01] | [15.288493156433105, 4.948275566101074] |
933c4c6c-5e31-4729-8fe7-fdc634079fe0 | informer-beyond-efficient-transformer-for | 2012.07436 | null | https://arxiv.org/abs/2012.07436v3 | https://arxiv.org/pdf/2012.07436v3.pdf | Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting | Many real-world applications require the prediction of long sequence time-series, such as electricity consumption planning. Long sequence time-series forecasting (LSTF) demands a high prediction capacity of the model, which is the ability to capture precise long-range dependency coupling between output and input efficiently. Recent studies have shown the potential of Transformer to increase the prediction capacity. However, there are several severe issues with Transformer that prevent it from being directly applicable to LSTF, including quadratic time complexity, high memory usage, and inherent limitation of the encoder-decoder architecture. To address these issues, we design an efficient transformer-based model for LSTF, named Informer, with three distinctive characteristics: (i) a $ProbSparse$ self-attention mechanism, which achieves $O(L \log L)$ in time complexity and memory usage, and has comparable performance on sequences' dependency alignment. (ii) the self-attention distilling highlights dominating attention by halving cascading layer input, and efficiently handles extreme long input sequences. (iii) the generative style decoder, while conceptually simple, predicts the long time-series sequences at one forward operation rather than a step-by-step way, which drastically improves the inference speed of long-sequence predictions. Extensive experiments on four large-scale datasets demonstrate that Informer significantly outperforms existing methods and provides a new solution to the LSTF problem. | ['Wancai Zhang', 'Hui Xiong', 'JianXin Li', 'Shuai Zhang', 'Jieqi Peng', 'Shanghang Zhang', 'Haoyi Zhou'] | 2020-12-14 | null | null | null | null | ['univariate-time-series-forecasting'] | ['time-series'] | [ 2.54283458e-01 -1.90895706e-01 -1.15490258e-01 -5.50363481e-01
-6.50114536e-01 -4.69510883e-01 2.32625455e-01 -2.63755083e-01
-4.16552685e-02 5.88768065e-01 1.86306536e-01 -6.04639888e-01
1.40378088e-01 -7.61700869e-01 -8.48846316e-01 -6.50367141e-01
-1.62865028e-01 4.39125061e-01 3.61023098e-02 -3.18370819e-01
1.61883041e-01 1.30736783e-01 -1.48330963e+00 4.35154021e-01
1.07962072e+00 1.36667430e+00 7.79451489e-01 8.00384700e-01
-3.23930204e-01 1.17763174e+00 -4.71480995e-01 -5.61964691e-01
4.89624143e-02 -4.41374362e-01 -7.75131702e-01 -2.60972023e-01
-1.87144354e-01 -2.55698532e-01 -2.60767072e-01 7.25090921e-01
5.13056517e-01 -1.07150324e-01 3.25940490e-01 -1.28603208e+00
-6.30655885e-01 9.69883502e-01 -4.70081478e-01 5.68189323e-01
2.35751599e-01 2.92956799e-01 1.17077959e+00 -7.23265827e-01
1.96614459e-01 1.11586368e+00 8.92457187e-01 2.16522217e-01
-1.03373313e+00 -8.72146130e-01 4.10696805e-01 4.65497583e-01
-1.35323656e+00 -4.18226361e-01 6.62742674e-01 -2.98274100e-01
1.94804275e+00 5.01276791e-01 5.87687254e-01 8.38728666e-01
3.20504248e-01 9.00822997e-01 6.04223311e-01 -1.56236365e-01
-4.86363396e-02 -4.14383709e-01 1.49002140e-02 4.50844049e-01
-4.69889969e-01 9.09676328e-02 -4.13976580e-01 8.96662846e-03
5.18806338e-01 8.98827016e-02 -1.91668391e-01 4.29758459e-01
-1.10218287e+00 6.12727106e-01 2.80112118e-01 3.73525321e-01
-2.64169276e-01 3.24468017e-01 6.67096674e-01 4.19875830e-01
6.80766463e-01 1.01489797e-01 -8.21540236e-01 -6.58735394e-01
-7.62551546e-01 -2.79705413e-03 7.67427146e-01 1.19210601e+00
4.40661877e-01 4.70386773e-01 -8.67930998e-04 8.53646696e-01
1.02291666e-01 6.92431331e-01 7.75326550e-01 -5.32954574e-01
8.48941803e-01 4.06170666e-01 -1.91689223e-01 -8.46582353e-01
-3.68724674e-01 -6.31517172e-01 -1.18641901e+00 -4.65989828e-01
-2.66419780e-02 -7.39068836e-02 -6.10635877e-01 1.90002310e+00
-1.68281168e-01 4.37803149e-01 -9.30843130e-02 6.25549018e-01
4.62474138e-01 1.35522342e+00 3.88489179e-02 -6.72132194e-01
9.86515820e-01 -1.08559048e+00 -5.87381244e-01 -5.83167434e-01
7.31120467e-01 -7.09580719e-01 1.15421665e+00 1.56780228e-01
-1.19692123e+00 -6.94082618e-01 -8.23728383e-01 -2.47780636e-01
-1.77638426e-01 -1.50057226e-01 4.88464892e-01 1.81921437e-01
-1.02204370e+00 5.79058230e-01 -7.56611645e-01 -6.73972666e-02
2.04912499e-02 4.68037993e-01 1.80333704e-01 9.67445746e-02
-1.40941262e+00 9.39247966e-01 4.79346007e-01 3.03429008e-01
-5.03837705e-01 -8.85996997e-01 -8.12695503e-01 5.65314770e-01
7.54272863e-02 -4.91897434e-01 1.44292450e+00 -9.31400597e-01
-1.61425507e+00 2.27907017e-01 -6.69203997e-01 -6.13284826e-01
2.53810585e-01 -2.39937559e-01 -7.07317531e-01 -3.99468839e-01
-1.47910073e-01 3.49290013e-01 6.49143219e-01 -5.60293257e-01
-7.64731646e-01 -1.14569508e-01 -4.89563912e-01 4.64184731e-02
-3.91538113e-01 7.04157799e-02 -4.50215250e-01 -8.47470999e-01
-1.93258330e-01 -8.94899189e-01 -4.11750346e-01 -5.87522805e-01
-3.61163646e-01 -4.74738389e-01 8.50885451e-01 -8.03309083e-01
1.88668704e+00 -2.08477163e+00 -2.96070799e-02 1.14794426e-01
-1.70852795e-01 4.13493484e-01 -1.47726119e-01 7.45050311e-01
-3.64476144e-01 1.38552055e-01 -3.57752383e-01 -1.95843160e-01
5.70879206e-02 4.02369171e-01 -8.24673176e-01 8.07359666e-02
2.28406996e-01 1.29208362e+00 -8.58075678e-01 -3.41398627e-01
9.75469425e-02 3.81311148e-01 -4.14366096e-01 3.52502733e-01
-4.51749504e-01 7.72204846e-02 -2.05948383e-01 2.95596719e-01
3.62875730e-01 -7.10678101e-01 3.55209738e-01 -1.19584762e-01
-2.41402835e-01 6.89593434e-01 -6.10516965e-01 1.47154868e+00
-6.24005437e-01 6.34784460e-01 -4.57965076e-01 -1.08285558e+00
9.58647251e-01 4.12471771e-01 5.24176240e-01 -9.70017850e-01
-7.25221038e-02 3.15255582e-01 -9.12350882e-03 -6.06698930e-01
3.04876477e-01 -2.20114619e-01 -1.72759742e-01 5.94701231e-01
-2.32455030e-01 6.43573999e-02 2.41958588e-01 -1.13482438e-01
9.75855708e-01 1.06498815e-01 2.74207234e-01 -1.58317640e-01
4.64159876e-01 -2.44446233e-01 8.67434800e-01 3.74885589e-01
1.89931497e-01 3.37905705e-01 4.04634684e-01 -7.51276195e-01
-1.33272278e+00 -7.07561314e-01 2.14496642e-01 1.33191979e+00
-2.52930224e-01 -5.87860644e-01 -4.26048845e-01 -4.54095900e-01
8.85495264e-03 1.02129912e+00 -2.71043479e-01 -5.94696626e-02
-1.13874686e+00 -8.75029862e-01 4.76757497e-01 8.71803880e-01
2.47241423e-01 -1.18407547e+00 -6.06145322e-01 6.85788393e-01
-6.31965399e-01 -1.02726901e+00 -9.28062558e-01 4.49961603e-01
-8.25484395e-01 -6.41278207e-01 -3.53893191e-01 -8.47313762e-01
2.90409029e-01 4.16896455e-02 1.44256413e+00 -1.83319464e-01
2.74809767e-02 -4.24012899e-01 -2.16706142e-01 -3.30971122e-01
-4.74948555e-01 3.32162261e-01 -8.50117356e-02 -2.81288207e-01
3.66830856e-01 -8.73976886e-01 -4.42793190e-01 3.65523428e-01
-7.07648635e-01 3.78693402e-01 7.15028822e-01 8.40103626e-01
5.04583538e-01 1.29870236e-01 9.23859060e-01 -7.44152248e-01
5.53441823e-01 -5.92498779e-01 -6.68156743e-01 3.70037675e-01
-1.01883733e+00 2.04868093e-01 1.24692822e+00 -4.04039621e-01
-9.60839510e-01 -3.97646204e-02 -7.89776027e-01 -4.95804429e-01
3.20889115e-01 8.12207937e-01 -5.16363792e-02 4.45768267e-01
1.39477462e-01 8.48306656e-01 1.33141503e-03 -5.83766580e-01
-2.86528841e-02 6.65773094e-01 4.51319933e-01 -2.87285239e-01
4.80950236e-01 -1.15481898e-01 -3.36436242e-01 -6.04026735e-01
-7.27760851e-01 -1.61504969e-01 -4.36058819e-01 2.48840787e-02
5.21267414e-01 -8.78345013e-01 -9.61370170e-01 4.59129781e-01
-1.23672080e+00 -5.07716954e-01 -1.33168772e-01 1.26278892e-01
-6.59475029e-01 3.10589433e-01 -9.97826815e-01 -8.23928058e-01
-8.45635653e-01 -9.83508766e-01 8.11846852e-01 -2.04110384e-01
-3.50392610e-01 -9.35018182e-01 -9.92124528e-02 5.53565286e-03
5.89441836e-01 -9.51281786e-02 1.30270207e+00 -5.53369343e-01
-5.62516809e-01 4.94800787e-03 -1.22210220e-01 4.43971038e-01
-1.75110728e-01 -1.91532582e-01 -9.22572970e-01 -3.20580781e-01
1.58744548e-02 -8.51510689e-02 6.37970507e-01 1.84569359e-01
1.68685007e+00 -6.85753167e-01 -3.70480746e-01 7.67856777e-01
1.26638496e+00 6.59808815e-01 7.30968952e-01 -1.48052230e-01
7.06703246e-01 1.23816535e-01 3.64901066e-01 6.04316115e-01
7.69482076e-01 4.55160290e-01 3.40624332e-01 -5.97214280e-03
1.64756671e-01 -5.07378936e-01 5.85577250e-01 1.81804097e+00
7.94906095e-02 -6.21821225e-01 -9.65623975e-01 5.06095588e-01
-1.95816422e+00 -1.27810740e+00 3.25167216e-02 1.96453917e+00
8.34853947e-01 2.74870157e-01 1.98835898e-02 3.88373226e-01
3.23428452e-01 3.91388267e-01 -8.01285207e-01 -6.96802735e-01
2.94530373e-02 1.46019414e-01 3.84067833e-01 3.04833770e-01
-6.98553860e-01 7.36031532e-01 7.33258390e+00 9.97733235e-01
-1.30894232e+00 6.66949302e-02 9.60610390e-01 -1.81742787e-01
-5.52948952e-01 -4.56829906e-01 -9.10528898e-01 9.72311854e-01
1.52151990e+00 -3.22505593e-01 5.51111221e-01 8.10882628e-01
2.06726313e-01 5.26256979e-01 -1.23377740e+00 1.05985487e+00
1.36093143e-03 -1.45485294e+00 4.81787883e-02 -1.30679086e-01
3.46386552e-01 3.61210525e-01 -4.50323820e-02 4.84054625e-01
3.53450626e-01 -1.05970359e+00 7.49057233e-01 3.31702560e-01
9.71586704e-01 -1.01552773e+00 6.00614607e-01 7.85140753e-01
-1.55963850e+00 -4.42067713e-01 -2.65115261e-01 -2.32077762e-01
4.21259999e-01 7.01314926e-01 -9.08823550e-01 4.24389392e-01
7.57257521e-01 9.83197749e-01 -5.61798289e-02 4.82724667e-01
1.48027003e-01 8.44441831e-01 -3.83779824e-01 -1.87731564e-01
1.62036255e-01 -1.93791255e-01 1.36433586e-01 1.50127351e+00
7.52315223e-01 3.15829426e-01 1.65944964e-01 5.82756042e-01
2.05032110e-01 -1.00813910e-01 -4.79094982e-01 -2.27787435e-01
5.56214929e-01 5.55559158e-01 -2.16977134e-01 -3.53455961e-01
-6.36490166e-01 8.85013998e-01 4.18370366e-01 3.38465899e-01
-1.10507059e+00 -1.19106703e-01 7.82750368e-01 3.83210741e-02
7.63172984e-01 -2.36334428e-01 -4.23494607e-01 -1.11618996e+00
1.82060376e-01 -8.93082559e-01 4.69142765e-01 -8.35649967e-01
-1.31906867e+00 9.25520837e-01 -3.81019026e-01 -1.24023008e+00
-7.60454416e-01 -1.85356662e-01 -4.22475725e-01 1.08695269e+00
-1.46426332e+00 -1.06710744e+00 1.85846105e-01 5.33362448e-01
9.67779219e-01 1.50427327e-01 9.13047791e-01 4.69622076e-01
-7.34265625e-01 7.29834735e-01 2.00787112e-01 -2.88078915e-02
7.97520876e-02 -9.79557812e-01 1.01465929e+00 9.50374722e-01
-1.15938261e-01 3.62823784e-01 5.55598795e-01 -5.79645872e-01
-1.65356553e+00 -1.28864229e+00 1.49089229e+00 -1.60219923e-01
6.99405074e-01 -4.64084238e-01 -1.00120354e+00 9.05363798e-01
1.88227892e-01 -1.17214315e-01 8.27915967e-01 1.22057647e-01
-3.80566835e-01 -3.96851629e-01 -5.56488454e-01 3.87856215e-01
1.13614070e+00 -6.46683753e-01 -3.32441896e-01 1.42394707e-01
9.52461004e-01 -4.30822015e-01 -1.06980860e+00 3.94391745e-01
7.14741468e-01 -8.29597652e-01 1.03443706e+00 -2.98752695e-01
6.30990028e-01 -1.32221714e-01 -2.36945510e-01 -1.36271393e+00
-7.19651282e-01 -8.69853735e-01 -4.98954356e-01 1.04253757e+00
6.15685403e-01 -7.80529797e-01 4.31099772e-01 4.57766503e-01
-4.45897967e-01 -1.12524498e+00 -8.23991895e-01 -8.01601231e-01
5.22953905e-02 -8.74120176e-01 1.10417509e+00 7.96906054e-01
2.63811439e-01 5.26944637e-01 -8.11821878e-01 7.40843788e-02
1.41144186e-01 6.78808451e-01 3.00967515e-01 -8.61528337e-01
-4.98313278e-01 -4.66925561e-01 9.15294737e-02 -1.61469591e+00
5.73896840e-02 -8.23073447e-01 2.26932809e-01 -1.26577139e+00
-1.89924687e-02 -5.60918152e-01 -4.72531080e-01 5.14106750e-01
-2.25081667e-01 -6.52946681e-02 2.48173445e-01 2.90722668e-01
-4.13459182e-01 5.65154195e-01 1.02938819e+00 7.64903799e-02
5.43115772e-02 2.38570794e-02 -5.40489554e-01 5.21446109e-01
7.09253252e-01 -2.76611000e-01 -6.10003054e-01 -7.75102317e-01
3.81427199e-01 5.99274218e-01 -1.52804762e-01 -7.34512568e-01
2.14165941e-01 -2.88699806e-01 3.24899167e-01 -8.84410799e-01
2.10431248e-01 -7.79091179e-01 4.71017599e-01 6.34154737e-01
-3.49533796e-01 7.67833054e-01 6.36120215e-02 4.05562282e-01
-2.08735809e-01 2.24380091e-01 5.10297894e-01 -4.22369912e-02
-7.77076066e-01 5.54147482e-01 -4.72121805e-01 6.38158619e-02
7.81698763e-01 3.07114627e-02 -2.85589129e-01 -4.49558228e-01
-2.70472467e-01 3.96572322e-01 7.80470371e-02 5.41385829e-01
4.56669390e-01 -1.33789527e+00 -6.67396307e-01 6.49320960e-01
-1.62739217e-01 1.03932664e-01 1.37069628e-01 6.63683891e-01
-2.32381091e-01 6.57642543e-01 1.06200390e-01 -5.84654868e-01
-9.97434437e-01 8.17405641e-01 1.62224039e-01 -6.28669620e-01
-8.30336392e-01 8.12822759e-01 1.85763925e-01 -2.05565244e-01
1.48161128e-01 -6.67067051e-01 9.16114226e-02 -1.75660670e-01
6.64837897e-01 2.93305278e-01 6.29509315e-02 -5.59924901e-01
-2.67402679e-01 4.63022351e-01 -9.36147943e-02 3.29849899e-01
1.58103275e+00 -3.35435361e-01 -8.99392739e-02 6.05822265e-01
1.36306465e+00 -4.44586039e-01 -1.32977486e+00 -3.30522805e-01
2.17503637e-01 -3.40662539e-01 -2.34547302e-01 -8.55374634e-01
-1.19809818e+00 9.47263122e-01 1.90109789e-01 5.56298494e-01
1.50466478e+00 -2.62864888e-01 1.53398490e+00 1.66200340e-01
3.32937062e-01 -8.85049045e-01 -2.07843006e-01 9.76533592e-01
8.59285414e-01 -8.72053504e-01 -4.36386317e-01 -3.66205603e-01
-6.24771416e-01 1.02014101e+00 5.59393644e-01 2.34030485e-01
6.69249952e-01 8.93793404e-01 -5.26201986e-02 1.70357913e-01
-1.53721952e+00 3.89245152e-02 1.74366593e-01 2.01261967e-01
6.49411559e-01 1.14642553e-01 -4.27948199e-02 7.43609905e-01
-4.50384527e-01 2.36166686e-01 1.23608829e-02 6.43140256e-01
-2.61906117e-01 -9.82694030e-01 7.47803152e-02 6.45593464e-01
-4.69569147e-01 -3.44207406e-01 2.08615348e-01 3.37122530e-01
4.19812948e-02 8.75635684e-01 4.22008932e-01 -6.26743317e-01
2.45381236e-01 3.76338810e-01 -2.44863592e-02 -1.58637807e-01
-7.68114626e-01 2.74370730e-01 1.91054583e-01 -6.31448686e-01
7.72694871e-02 -6.48076832e-01 -1.17512751e+00 -5.93510211e-01
-1.24814220e-01 1.30666137e-01 3.51359814e-01 1.03592837e+00
7.55159020e-01 5.92471063e-01 8.23336542e-01 -5.66192627e-01
-6.84914052e-01 -1.01423407e+00 -2.53843099e-01 3.10396433e-01
3.68742824e-01 -4.85606566e-02 -5.92438355e-02 2.22149953e-01] | [7.122158527374268, 2.910532236099243] |
5a25517d-7ee5-4af5-a00f-b7a18c9eedb8 | video-instance-segmentation-in-an-open-world | 2304.01200 | null | https://arxiv.org/abs/2304.01200v1 | https://arxiv.org/pdf/2304.01200v1.pdf | Video Instance Segmentation in an Open-World | Existing video instance segmentation (VIS) approaches generally follow a closed-world assumption, where only seen category instances are identified and spatio-temporally segmented at inference. Open-world formulation relaxes the close-world static-learning assumption as follows: (a) first, it distinguishes a set of known categories as well as labels an unknown object as `unknown' and then (b) it incrementally learns the class of an unknown as and when the corresponding semantic labels become available. We propose the first open-world VIS approach, named OW-VISFormer, that introduces a novel feature enrichment mechanism and a spatio-temporal objectness (STO) module. The feature enrichment mechanism based on a light-weight auxiliary network aims at accurate pixel-level (unknown) object delineation from the background as well as distinguishing category-specific known semantic classes. The STO module strives to generate instance-level pseudo-labels by enhancing the foreground activations through a contrastive loss. Moreover, we also introduce an extensive experimental protocol to measure the characteristics of OW-VIS. Our OW-VISFormer performs favorably against a solid baseline in OW-VIS setting. Further, we evaluate our contributions in the standard fully-supervised VIS setting by integrating them into the recent SeqFormer, achieving an absolute gain of 1.6\% AP on Youtube-VIS 2019 val. set. Lastly, we show the generalizability of our contributions for the open-world detection (OWOD) setting, outperforming the best existing OWOD method in the literature. Code, models along with OW-VIS splits are available at \url{https://github.com/OmkarThawakar/OWVISFormer}. | ['Fahad Shahbaz Khan', 'Mubarak Shah', 'Jorma Laaksonen', 'Salman Khan', 'Rao Muhammad Anwer', 'Hisham Cholakkal', 'Sanath Narayan', 'Omkar Thawakar'] | 2023-04-03 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 3.56199235e-01 1.79903790e-01 -4.35047537e-01 -2.24806458e-01
-1.02112317e+00 -7.98537731e-01 6.97958887e-01 -9.20692831e-02
-4.61929232e-01 7.25840271e-01 -7.09975883e-02 -8.48780721e-02
5.71790291e-03 -6.56484604e-01 -1.09993923e+00 -8.06286693e-01
-1.37398988e-01 4.79392141e-01 7.85256803e-01 2.87165552e-01
-1.86166018e-02 1.53610632e-01 -1.59831226e+00 4.74153817e-01
8.03882897e-01 1.17903852e+00 2.90973663e-01 6.10524356e-01
-4.45387401e-02 8.19607139e-01 -3.33762825e-01 -4.96894717e-01
5.71744800e-01 -3.42763931e-01 -9.49321270e-01 2.18782336e-01
8.33510578e-01 -3.86480093e-01 -5.04594326e-01 1.00345314e+00
4.37801450e-01 3.85490924e-01 6.37365878e-01 -1.36332273e+00
-5.04754305e-01 5.65898597e-01 -8.70980680e-01 6.66994274e-01
7.56139755e-02 4.95426536e-01 1.15151978e+00 -8.67411852e-01
8.58098388e-01 9.43794250e-01 5.70380688e-01 5.88803411e-01
-1.21363246e+00 -6.62745893e-01 6.95418060e-01 4.35074776e-01
-1.39281011e+00 -3.46574813e-01 5.92963994e-01 -4.63784218e-01
6.43663824e-01 3.37563425e-01 5.91726005e-01 1.32003915e+00
-2.08772764e-01 1.28318810e+00 1.16331470e+00 -1.16377898e-01
3.46885324e-01 3.71431410e-02 4.86552328e-01 7.48435974e-01
1.52945936e-01 4.99872081e-02 -5.35103023e-01 1.23805501e-01
6.56160951e-01 3.29188333e-04 -4.76791114e-01 -6.36796415e-01
-1.39450133e+00 3.31044793e-01 6.88056588e-01 8.71216580e-02
-2.22389400e-01 3.72328252e-01 3.90490025e-01 -1.07991453e-02
5.86663008e-01 1.05726361e-01 -5.45898855e-01 -7.03008100e-02
-1.13390315e+00 1.36673257e-01 5.44250727e-01 1.01768076e+00
6.90790057e-01 -1.58460438e-01 -5.35547733e-01 6.63897157e-01
2.10101679e-01 4.09156650e-01 3.02670777e-01 -1.13120866e+00
4.99112070e-01 3.78843367e-01 1.03364266e-01 -5.20664752e-01
-9.04412195e-02 -6.02155983e-01 -2.82567114e-01 1.38870865e-01
7.05939353e-01 -7.81949162e-02 -1.26657093e+00 1.83622026e+00
6.31421447e-01 9.82220352e-01 -7.66129792e-02 1.01557338e+00
1.01145017e+00 6.81176782e-01 2.84021437e-01 -1.37085542e-01
1.30772460e+00 -1.26961195e+00 -4.60017741e-01 -2.88425297e-01
3.10917079e-01 -2.59934455e-01 1.17612779e+00 3.70150268e-01
-1.03111339e+00 -5.51435590e-01 -7.88949013e-01 2.96598095e-02
-3.68945301e-01 -7.38630593e-02 5.51025391e-01 6.04951382e-01
-1.11486602e+00 4.20455575e-01 -9.08492088e-01 -3.47674638e-01
1.02405000e+00 1.43264189e-01 -1.94964513e-01 -1.56574443e-01
-9.36144948e-01 3.62879813e-01 4.72706586e-01 -4.50103125e-03
-1.45544434e+00 -8.91839802e-01 -7.78878987e-01 -8.45118687e-02
9.33539331e-01 -8.52619708e-01 1.15520811e+00 -1.43868792e+00
-1.04745591e+00 1.10479069e+00 -3.92712057e-01 -3.92727762e-01
8.17747235e-01 -1.64845183e-01 -2.79396832e-01 4.13331658e-01
2.83407450e-01 7.94750810e-01 7.51686633e-01 -1.64531696e+00
-9.71331418e-01 -3.62583339e-01 4.80260968e-01 2.39899278e-01
-1.10658251e-01 -2.26977855e-01 -9.09522831e-01 -7.50809371e-01
6.54419884e-02 -7.25022197e-01 -2.16555805e-03 3.53547871e-01
-5.02635896e-01 -1.95968404e-01 6.83000982e-01 -5.05417407e-01
1.18165398e+00 -2.32616234e+00 1.17582463e-01 -2.90029142e-02
5.27961969e-01 2.98102558e-01 -2.15285107e-01 -4.66032466e-03
-7.19985515e-02 -9.22389925e-02 -4.46445197e-01 -2.77462363e-01
8.70682672e-02 1.99771181e-01 -2.67256856e-01 5.98176301e-01
1.40908808e-01 1.06423521e+00 -1.17768502e+00 -5.63784122e-01
3.19700867e-01 3.29795331e-01 -5.30591965e-01 5.85372224e-02
-4.47222471e-01 2.74852365e-01 -3.66479039e-01 8.29396486e-01
6.41655684e-01 -3.53281558e-01 -7.78178200e-02 -2.14762658e-01
2.92720739e-02 -9.07056630e-02 -1.40546322e+00 1.65188134e+00
-1.69779375e-01 5.37417710e-01 8.98556411e-02 -9.62624609e-01
3.50058198e-01 1.33706644e-01 4.72130418e-01 -5.03040850e-01
1.51010379e-01 2.31976047e-01 -1.61953583e-01 -4.92187619e-01
2.13170364e-01 5.32666408e-02 2.65297651e-01 2.49697611e-01
3.07624161e-01 3.03266674e-01 3.47651303e-01 5.07064342e-01
1.12195659e+00 6.83700979e-01 7.41568953e-02 -3.86156112e-01
4.32572395e-01 -2.40513794e-02 6.94325387e-01 8.70193064e-01
-6.88961506e-01 8.24502707e-01 4.96721268e-01 -1.57205641e-01
-5.27515531e-01 -1.46496630e+00 -3.55289936e-01 1.07503319e+00
7.19382524e-01 -2.45270431e-02 -8.68636131e-01 -1.13488543e+00
2.18771026e-03 6.73101902e-01 -7.25328267e-01 1.09691568e-01
-4.64641631e-01 -5.45759916e-01 4.18542296e-01 5.96290231e-01
5.89109719e-01 -1.05969286e+00 -5.86311340e-01 7.71153206e-03
-3.80667597e-01 -1.20485306e+00 -6.24144793e-01 2.06333995e-01
-6.11823857e-01 -1.26765347e+00 -9.55059111e-01 -8.35325837e-01
5.58819234e-01 3.54343891e-01 1.02866268e+00 -9.93773714e-02
-2.18466580e-01 5.95912516e-01 -3.89869004e-01 -1.73016086e-01
1.07105248e-01 -8.07637647e-02 -2.35827282e-01 4.63488102e-01
2.35479161e-01 -2.52616674e-01 -9.52151537e-01 3.61902297e-01
-7.95869052e-01 2.01519638e-01 3.54523003e-01 6.85980678e-01
8.45393002e-01 -1.14739724e-01 4.90056664e-01 -1.06885910e+00
-1.61256239e-01 -8.02035928e-01 -3.46074611e-01 2.77477533e-01
-3.03326339e-01 -3.04415166e-01 3.39347810e-01 -4.51793939e-01
-1.24321425e+00 -4.56393994e-02 -1.60874709e-01 -5.99925160e-01
-4.47950274e-01 2.77423504e-04 -3.13591808e-01 2.02405468e-01
4.97775495e-01 2.41048679e-01 -4.07338381e-01 -1.74827561e-01
5.78258693e-01 5.02750993e-01 7.33678818e-01 -7.55665481e-01
7.90872693e-01 9.39955354e-01 -4.46103215e-01 -6.06393456e-01
-1.06595445e+00 -8.37874234e-01 -7.16652870e-01 -4.34702724e-01
9.63248014e-01 -1.17467010e+00 -2.98615038e-01 5.46059310e-01
-7.69308865e-01 -9.71169293e-01 -7.29841709e-01 2.38119721e-01
-6.52319491e-01 3.13810021e-01 -5.63246846e-01 -7.07395673e-01
-2.85647507e-03 -1.02334964e+00 1.35543549e+00 1.24150917e-01
-7.50438571e-02 -1.04866195e+00 -2.16328233e-01 5.46719253e-01
-1.12839974e-01 3.99117023e-01 4.35844362e-01 -5.79773307e-01
-8.51691306e-01 2.30470836e-01 -4.97610867e-01 1.77433088e-01
-9.40095335e-02 -1.61095202e-01 -1.33948934e+00 -2.13720843e-01
-4.56966549e-01 -2.54561305e-01 1.25970757e+00 6.55447602e-01
1.31523848e+00 -3.05647980e-02 -6.20205462e-01 8.74483168e-01
1.42082798e+00 9.51055884e-02 5.05334139e-01 3.22697312e-01
7.84518838e-01 4.83909726e-01 7.37817049e-01 2.81300187e-01
4.30321544e-01 7.07951009e-01 5.18454194e-01 -3.28154176e-01
-5.95992088e-01 -1.96314976e-01 4.23877567e-01 1.52604789e-01
-1.33012369e-01 -5.58167696e-01 -7.50523508e-01 9.01077151e-01
-1.90342820e+00 -1.04000115e+00 -2.72519618e-01 2.10078168e+00
7.69294083e-01 2.06017271e-01 3.64476711e-01 3.67391407e-02
6.42473280e-01 3.62481117e-01 -7.69910574e-01 1.03644572e-01
-1.34854421e-01 1.92079976e-01 4.86705929e-01 3.98526311e-01
-1.42514265e+00 9.91791725e-01 5.08201838e+00 9.91910279e-01
-8.75406802e-01 5.82274437e-01 9.08634782e-01 -3.79443437e-01
-2.77659863e-01 -1.22735173e-01 -7.98829079e-01 5.65792680e-01
5.43808758e-01 4.48489822e-02 1.99143961e-01 6.45519197e-01
2.45885879e-01 -3.10167760e-01 -1.15124285e+00 9.86574590e-01
2.26217344e-01 -1.24196529e+00 1.80004432e-03 -1.44171923e-01
8.72960865e-01 2.08031878e-01 1.81331076e-02 4.24099833e-01
2.41313830e-01 -5.55072069e-01 1.20767128e+00 3.13132852e-01
9.48107719e-01 -3.88500899e-01 5.00581324e-01 2.54746169e-01
-1.46311259e+00 -2.84216851e-01 5.12599349e-02 1.02735721e-01
3.07349205e-01 3.94513071e-01 -2.91241556e-01 6.16548836e-01
8.88559163e-01 1.00569093e+00 -5.03477931e-01 1.19189501e+00
-2.96474099e-01 9.48836267e-01 -2.25650907e-01 4.79176670e-01
4.64615434e-01 -8.71204957e-02 6.91406786e-01 1.34345114e+00
-1.33965358e-01 3.23903918e-01 3.33269477e-01 1.04905117e+00
-1.17640130e-01 -1.93671316e-01 -2.06142634e-01 2.55586386e-01
2.32421324e-01 1.17482066e+00 -1.19800937e+00 -5.36104620e-01
-4.51039463e-01 1.20421672e+00 1.79047689e-01 7.02068686e-01
-1.22473395e+00 -1.57817826e-01 6.51954889e-01 3.53128314e-01
5.70519745e-01 1.23321138e-01 -1.31884515e-01 -1.24317348e+00
4.17485684e-02 -5.79443991e-01 7.51994014e-01 -7.91904807e-01
-1.21386302e+00 4.58035856e-01 1.87437311e-01 -1.12383306e+00
3.07464898e-01 -5.93107462e-01 -4.51954633e-01 5.03719449e-01
-1.85388219e+00 -1.27851045e+00 -4.68372345e-01 7.16383398e-01
1.08883142e+00 2.01727957e-01 2.18546808e-01 5.38228512e-01
-8.49178970e-01 5.97790062e-01 2.16075182e-02 2.20010415e-01
5.70581257e-01 -1.39892757e+00 2.31595561e-01 1.14296913e+00
4.13908333e-01 3.03971648e-01 4.90851790e-01 -6.12753808e-01
-1.03470922e+00 -1.46341836e+00 4.15432394e-01 -7.54765272e-01
7.73724794e-01 -5.28632402e-01 -8.87351036e-01 8.55218768e-01
-5.91563806e-02 6.29536033e-01 2.23139361e-01 -1.76105261e-01
-3.54087591e-01 -7.58282691e-02 -1.09831119e+00 4.80690151e-01
1.61902833e+00 -2.93867767e-01 -5.28321564e-01 3.47147524e-01
7.47690022e-01 -6.13136113e-01 -5.28335154e-01 2.74840862e-01
3.89512479e-01 -9.08972323e-01 1.12524199e+00 -4.45241213e-01
4.21332181e-01 -5.36525011e-01 -1.20772406e-01 -9.25130129e-01
-1.82513565e-01 -5.22269666e-01 -4.79713500e-01 1.32482612e+00
2.68069446e-01 -6.10676169e-01 7.85610676e-01 3.40863526e-01
-2.97400713e-01 -9.36302304e-01 -9.58513677e-01 -9.96811867e-01
1.25508877e-02 -6.70464396e-01 2.46358141e-01 9.58570302e-01
-3.75150472e-01 -4.61920239e-02 -1.59379132e-02 3.71317029e-01
8.09325993e-01 1.09603465e-01 5.98990500e-01 -8.83640110e-01
-5.32484710e-01 -3.58551264e-01 -3.65373045e-01 -1.30484951e+00
1.69388056e-01 -1.13000810e+00 7.83841759e-02 -1.64400268e+00
5.37892640e-01 -6.53196573e-01 -6.60266578e-01 5.28670251e-01
-4.29208338e-01 6.51928365e-01 2.58020014e-01 1.66939333e-01
-1.11216116e+00 3.85042310e-01 1.25962949e+00 -2.05817163e-01
-1.31286696e-01 -3.78444940e-02 -5.81188619e-01 7.42783546e-01
4.05705631e-01 -3.15319419e-01 -5.71421862e-01 -3.78024340e-01
-2.50978231e-01 -2.10386723e-01 7.99673080e-01 -9.51970756e-01
-8.50578323e-02 -1.99830249e-01 2.09510252e-01 -3.72304559e-01
2.27349803e-01 -7.18991637e-01 1.03502326e-01 3.64572525e-01
-2.04354808e-01 -5.50304294e-01 2.02915356e-01 9.64913726e-01
-2.26066653e-02 -1.08980209e-01 8.62915039e-01 -1.65375635e-01
-1.48143148e+00 6.31484330e-01 1.48359034e-02 5.04002571e-01
1.56157088e+00 -7.02460170e-01 -4.56940770e-01 2.95746233e-02
-8.65947664e-01 4.41369712e-01 3.88265491e-01 4.26643074e-01
4.30499315e-01 -8.73864233e-01 -5.94261348e-01 3.85030881e-02
3.59375954e-01 3.35890576e-02 3.84696156e-01 1.19242382e+00
-3.33525807e-01 8.71153399e-02 1.63605332e-01 -8.40390384e-01
-1.17920208e+00 6.04442179e-01 5.07739723e-01 4.23307568e-02
-8.98764729e-01 1.14242017e+00 8.33038747e-01 -2.40162387e-01
4.66774225e-01 -2.51407564e-01 -7.82442763e-02 4.65342961e-02
4.89945859e-01 4.20048982e-01 -9.78014991e-02 -7.97118604e-01
-3.90562803e-01 4.57866192e-01 1.96567610e-01 4.07952517e-02
1.20027983e+00 -2.84841567e-01 3.52908820e-01 5.90929627e-01
9.46820796e-01 -1.86704129e-01 -1.78603530e+00 -3.35187763e-01
-1.99286729e-01 -7.13882506e-01 9.04480219e-02 -9.36768651e-01
-1.28739953e+00 7.15986848e-01 7.87043214e-01 -1.92286551e-01
1.06829727e+00 4.03877139e-01 7.39022136e-01 -1.53710499e-01
5.14743149e-01 -8.44776690e-01 7.84652457e-02 1.28465727e-01
5.16892612e-01 -1.38146138e+00 -1.13466114e-01 -6.78514719e-01
-7.40563512e-01 6.23308599e-01 8.03424954e-01 -1.11772381e-02
5.27590096e-01 2.01062411e-01 6.53152689e-02 -2.73436755e-01
-5.58051884e-01 -5.66102922e-01 3.91183078e-01 7.20045805e-01
1.11483812e-01 -6.29107980e-03 -7.15068132e-02 5.98209620e-01
3.38654816e-01 -1.24265840e-02 4.09493983e-01 6.90815449e-01
-2.58030027e-01 -6.24023855e-01 -2.63027459e-01 5.40212512e-01
-3.08673710e-01 -1.04447469e-01 -1.40830979e-01 8.28613222e-01
5.76364994e-01 7.54840910e-01 2.31180295e-01 2.19639428e-02
3.20597559e-01 -9.35528651e-02 4.05519247e-01 -7.57276833e-01
-4.04788256e-01 4.74517047e-02 1.33700194e-02 -8.43244672e-01
-6.32545292e-01 -9.20106411e-01 -1.39343143e+00 -1.14911562e-02
-3.91616255e-01 -2.08384156e-01 6.32741302e-02 7.99720287e-01
1.79999277e-01 7.28287339e-01 2.79373884e-01 -8.80810022e-01
-1.32884577e-01 -4.13853914e-01 -6.95509970e-01 7.47468650e-01
4.56770122e-01 -1.00175548e+00 -4.38138038e-01 3.98960888e-01] | [9.264726638793945, 0.06481468677520752] |
bb212395-8543-40fe-951c-637ddf9fecda | multi-step-greedy-policies-in-model-free-deep | null | null | https://openreview.net/forum?id=r1l7E1HFPH | https://openreview.net/pdf?id=r1l7E1HFPH | Multi-step Greedy Policies in Model-Free Deep Reinforcement Learning | Multi-step greedy policies have been extensively used in model-based Reinforcement Learning (RL) and in the case when a model of the environment is available (e.g., in the game of Go). In this work, we explore the benefits of multi-step greedy policies in model-free RL when employed in the framework of multi-step Dynamic Programming (DP): multi-step Policy and Value Iteration. These algorithms iteratively solve short-horizon decision problems and converge to the optimal solution of the original one. By using model-free algorithms as solvers of the short-horizon problems we derive fully model-free algorithms which are instances of the multi-step DP framework. As model-free algorithms are prone to instabilities w.r.t. the decision problem horizon, this simple approach can help in mitigating these instabilities and results in an improved model-free algorithms. We test this approach and show results on both discrete and continuous control problems. | ['Mohammad Ghavamzadeh', 'Manan Tomar', 'Yonathan Efroni'] | 2019-09-25 | null | null | null | null | ['game-of-go'] | ['playing-games'] | [ 1.16327129e-01 3.62153918e-01 -3.60204160e-01 2.87571490e-01
-6.47050083e-01 -5.05298197e-01 4.75806981e-01 1.92219645e-01
-8.32740247e-01 1.33930433e+00 -2.64917284e-01 -6.51820898e-01
-4.93375480e-01 -9.00757015e-01 -5.77047050e-01 -7.88446665e-01
-3.12503815e-01 7.35208035e-01 1.84857577e-01 -5.24643958e-01
2.86989987e-01 4.01376635e-01 -1.51573479e+00 -1.12768784e-01
8.38948846e-01 6.93233550e-01 5.06223857e-01 9.07042384e-01
-4.05905806e-02 8.34644258e-01 -3.06570888e-01 3.83565128e-01
5.33596635e-01 -5.77241898e-01 -8.73484969e-01 1.96874559e-01
-4.97924805e-01 -2.82331973e-01 2.39626810e-01 9.95779634e-01
5.85835695e-01 5.75130582e-01 1.08617671e-01 -1.15831029e+00
5.66482544e-01 4.25197273e-01 -5.15313983e-01 3.41413659e-03
4.59691554e-01 3.46364826e-01 7.41391182e-01 -2.95904100e-01
6.43310428e-01 1.52582169e+00 3.12617540e-01 5.55936277e-01
-1.45246339e+00 -1.88782275e-01 4.05630708e-01 1.76503643e-01
-9.34812903e-01 -1.65038079e-01 4.94287699e-01 -2.73619145e-01
1.04646611e+00 1.80884764e-01 9.48555052e-01 6.74123466e-01
3.54346454e-01 7.15494156e-01 1.76115012e+00 -7.77617097e-01
7.49754608e-01 1.34999603e-02 -2.94227153e-01 5.53341508e-01
6.33221492e-02 7.06878662e-01 -8.99922699e-02 -1.93079680e-01
9.42198277e-01 -3.01979184e-01 6.58407733e-02 -5.40139496e-01
-8.29882145e-01 1.07401669e+00 2.74851150e-03 1.39552534e-01
-7.38147914e-01 2.44776756e-01 2.17367217e-01 7.92181253e-01
3.43575120e-01 7.60895967e-01 -4.44995850e-01 -4.65381742e-01
-9.32820737e-01 9.39410806e-01 1.10295439e+00 5.37299335e-01
7.46906579e-01 2.97123283e-01 -2.06153870e-01 4.44831312e-01
3.10761929e-01 2.91537732e-01 3.34170938e-01 -1.25831306e+00
4.47788268e-01 1.91609681e-01 8.59205127e-01 -4.37380672e-01
-5.56216240e-01 -4.59784806e-01 -3.20800573e-01 7.91077077e-01
6.64101362e-01 -5.93330204e-01 -6.66634560e-01 1.63586116e+00
6.75755918e-01 1.66371182e-01 3.12341362e-01 6.88353777e-01
-4.15565401e-01 7.76477218e-01 -8.66692960e-02 -9.71833348e-01
7.60788918e-01 -8.61018360e-01 -6.81106567e-01 -2.02773809e-01
7.58138299e-01 -3.79814297e-01 8.52968037e-01 8.29883218e-01
-1.31348658e+00 -2.26422817e-01 -8.84204745e-01 7.43622780e-01
-2.13680774e-01 -2.63282061e-01 2.67556012e-01 5.13182402e-01
-1.11875749e+00 1.07409716e+00 -9.76654530e-01 -1.10737927e-01
-1.66974783e-01 5.51514447e-01 2.77623743e-01 2.01401785e-01
-1.13716614e+00 1.17649460e+00 7.93668151e-01 2.20152866e-02
-1.06378877e+00 -3.84063989e-01 -4.84868318e-01 3.21755710e-04
1.30413318e+00 -3.07114661e-01 1.79861546e+00 -1.08759630e+00
-2.02331138e+00 3.67016822e-01 8.82505625e-02 -7.83868909e-01
1.13542378e+00 -2.27468610e-01 2.44459644e-01 4.90198471e-02
-2.30644464e-01 2.40521193e-01 9.25750554e-01 -1.04007542e+00
-7.94657767e-01 4.46426123e-02 7.28909850e-01 4.05987233e-01
2.96722353e-01 -9.42841396e-02 3.43198091e-01 -2.02572733e-01
-3.53748649e-01 -1.17022967e+00 -9.59147632e-01 -4.04435843e-01
3.49397101e-02 -1.05980514e-02 4.80321407e-01 -3.58725548e-01
1.32089758e+00 -1.61596894e+00 4.76612210e-01 3.08703601e-01
-3.65275353e-01 3.82825464e-01 -5.99850640e-02 8.22498381e-01
-5.16586900e-02 7.70040601e-03 -7.37594068e-02 -2.29507446e-01
1.40655085e-01 7.87554562e-01 -2.28548065e-01 3.46121639e-01
-1.30987063e-01 6.25555336e-01 -1.01445484e+00 -4.90122437e-01
3.79203647e-01 -3.34502518e-01 -6.06146038e-01 2.37803310e-01
-8.66156816e-01 4.78988439e-01 -6.83740854e-01 6.27434626e-02
3.45363736e-01 3.37747633e-01 5.20643890e-01 5.65107584e-01
-6.90170050e-01 2.57938672e-02 -1.65747714e+00 1.32440424e+00
-7.25635946e-01 -1.12781818e-05 4.66900826e-01 -1.27773345e+00
6.61372721e-01 3.54485035e-01 5.76078951e-01 -8.14709246e-01
1.00821093e-01 1.05804592e-01 -6.56389678e-03 -2.83102125e-01
4.57712710e-01 -5.59704483e-01 7.93492347e-02 5.46959579e-01
-1.77083343e-01 -3.03679436e-01 6.07572436e-01 -2.39941508e-01
9.39097166e-01 4.67103779e-01 7.80927837e-01 -5.68513989e-01
7.38300145e-01 1.23980068e-01 6.89445436e-01 1.14004397e+00
-7.03580379e-02 1.52006606e-02 1.07112217e+00 -2.79691547e-01
-1.05343616e+00 -3.62680733e-01 3.05178493e-01 9.63864744e-01
-1.71477363e-01 -3.36002707e-01 -5.38963675e-01 -4.49601710e-01
-4.18979935e-02 9.00310099e-01 -5.94666600e-01 2.68506389e-02
-7.37012386e-01 -6.89953685e-01 -8.89154971e-02 8.99370760e-02
3.88883114e-01 -1.09036839e+00 -1.20399666e+00 9.19080973e-01
2.81047225e-01 -6.78006351e-01 2.69879818e-01 4.47678238e-01
-9.94231880e-01 -1.09264147e+00 -6.70804977e-01 -1.64270818e-01
4.44419473e-01 -1.44891381e-01 8.61714005e-01 -3.79773080e-02
1.27442896e-01 7.74976850e-01 -2.29513928e-01 -4.99801397e-01
-6.93632960e-01 2.69147102e-02 5.50022908e-02 -1.77993208e-01
-6.08285308e-01 -4.28175688e-01 -2.19059214e-01 2.81757355e-01
-7.85506129e-01 2.33283252e-01 2.53062755e-01 9.45451379e-01
6.76808357e-01 2.48282835e-01 6.43898606e-01 -8.13162804e-01
1.07501817e+00 -3.81291270e-01 -1.59721029e+00 3.49330783e-01
-8.65356922e-01 5.31756580e-01 8.96889150e-01 -6.32911623e-01
-1.00108778e+00 2.07697138e-01 6.89077228e-02 -3.61743808e-01
6.61592185e-02 6.80079043e-01 1.85518876e-01 -3.39512318e-01
3.28278065e-01 2.06990778e-01 2.34949723e-01 -4.84361589e-01
9.80303884e-02 9.96436104e-02 -1.44501209e-01 -9.58986878e-01
5.77106535e-01 -1.47015303e-01 5.40452480e-01 -6.78786278e-01
-6.75197423e-01 -2.14964241e-01 -3.43433350e-01 -4.22877729e-01
4.58620846e-01 -5.29959321e-01 -8.15916777e-01 4.77912754e-01
-7.46356010e-01 -9.68176007e-01 -6.68640137e-01 2.90146947e-01
-1.34551871e+00 1.69376493e-01 -1.53949112e-01 -1.34369183e+00
1.94550276e-01 -1.08970010e+00 3.55157554e-01 5.23597360e-01
1.79730862e-01 -1.27167702e+00 4.79141623e-01 -2.45861098e-01
3.60495895e-01 5.51967800e-01 7.03961253e-01 -4.67009962e-01
-3.42360139e-01 2.24135935e-01 5.33098221e-01 2.19664067e-01
-2.83188254e-01 -1.98277980e-01 -5.76415718e-01 -6.59828901e-01
1.03773490e-01 -3.73483241e-01 4.69326913e-01 4.48275805e-01
7.18460739e-01 -7.57020056e-01 -4.93779518e-02 1.23243645e-01
1.83051026e+00 4.77504075e-01 1.96209028e-01 8.28028738e-01
5.13983741e-02 5.40360391e-01 1.38239813e+00 1.05612612e+00
1.22262798e-01 7.59880841e-01 5.87499022e-01 1.50102451e-01
6.52749002e-01 -8.70526284e-02 4.87549782e-01 1.10326381e-02
-1.96876526e-01 8.23190883e-02 -9.58288610e-01 5.19200027e-01
-2.40523791e+00 -1.06711507e+00 2.13566497e-01 2.35673475e+00
9.97035682e-01 3.72365922e-01 6.26826346e-01 1.52622730e-01
4.65686321e-01 8.86742622e-02 -6.84862912e-01 -1.04313767e+00
2.67737120e-01 2.97293067e-01 7.00183749e-01 8.90512526e-01
-9.50560570e-01 9.40950692e-01 6.43608570e+00 7.68071771e-01
-1.11818266e+00 -1.88217293e-02 3.16762984e-01 -2.90678680e-01
1.16135158e-01 2.67853647e-01 -7.41017938e-01 3.93170297e-01
1.27340317e+00 -7.32959628e-01 9.58656549e-01 9.43758786e-01
8.94705713e-01 -7.76173949e-01 -9.44081724e-01 5.89414835e-01
-7.59340107e-01 -1.03229368e+00 -6.25113368e-01 3.66007894e-01
9.09881294e-01 -2.27386475e-01 -2.37851217e-01 6.28782153e-01
5.71466804e-01 -7.95205832e-01 7.94942319e-01 3.56803894e-01
2.65809089e-01 -1.22273970e+00 5.43318152e-01 9.15957272e-01
-8.78461957e-01 -5.67713141e-01 -2.31842756e-01 -6.23744130e-01
2.24036932e-01 2.53625542e-01 -8.11488807e-01 7.18585312e-01
9.15778652e-02 3.46990675e-01 -7.89195523e-02 1.20041788e+00
-1.50825500e-01 5.44162393e-01 -5.76674104e-01 -1.84127539e-01
8.84955406e-01 -5.30540466e-01 7.61795342e-01 8.69628370e-01
9.32706818e-02 3.66510265e-02 6.12356603e-01 6.66068733e-01
7.88522124e-01 -3.53102684e-02 -5.85865855e-01 -5.59588261e-02
-9.80972778e-03 9.74651158e-01 -6.69542909e-01 -2.53020942e-01
-1.69977412e-01 3.14439088e-01 3.15123498e-01 3.85875434e-01
-7.16464818e-01 -1.24147087e-01 5.69391727e-01 -4.89390790e-02
4.19193059e-01 -5.27469575e-01 1.18621491e-01 -7.70234883e-01
-1.51537716e-01 -1.09780312e+00 3.10778975e-01 -2.94244468e-01
-4.77772206e-01 9.04239938e-02 5.06731331e-01 -1.00180006e+00
-1.08503783e+00 -3.59391779e-01 -5.78054130e-01 7.19455361e-01
-1.64175725e+00 -3.96260619e-01 4.21751350e-01 5.81382692e-01
5.21719158e-01 1.65131271e-01 7.44554996e-01 -2.36156270e-01
-4.51112419e-01 1.69214532e-02 6.25450850e-01 -7.18320966e-01
9.29066390e-02 -1.66392303e+00 -2.32484996e-01 7.93040335e-01
-4.05765027e-01 1.24374487e-01 1.16054749e+00 -4.18632507e-01
-1.45491827e+00 -7.72151113e-01 2.66290367e-01 4.17136610e-01
8.49176764e-01 -7.42775202e-02 -5.74930251e-01 4.99295741e-01
-1.47394016e-02 -4.16399300e-01 -1.05240624e-02 -7.51098916e-02
6.70495093e-01 -1.29621342e-01 -1.24270022e+00 5.56142926e-01
4.74693835e-01 -6.28049374e-02 -4.30391014e-01 3.31887662e-01
3.83342594e-01 -6.23503089e-01 -6.50335610e-01 -1.04071042e-02
3.75752568e-01 -9.15171027e-01 7.78848231e-01 -7.02453434e-01
-5.92633933e-02 -3.87597121e-02 2.37304464e-01 -1.66540718e+00
-4.78739478e-02 -1.43852091e+00 -2.93428481e-01 8.27256739e-01
1.47066981e-01 -8.47832978e-01 5.31416774e-01 6.11590981e-01
2.00204089e-01 -9.42282856e-01 -1.29606247e+00 -1.03948748e+00
3.17475498e-01 -2.37375557e-01 3.05174738e-01 4.31578726e-01
2.46770997e-02 -2.38069206e-01 -5.54912746e-01 -1.16218172e-01
6.29637003e-01 2.83681482e-01 6.20627165e-01 -8.07099044e-01
-8.87151599e-01 -3.05375159e-01 1.53550848e-01 -7.80479908e-01
2.20748857e-01 -1.31798059e-01 1.06499173e-01 -1.47654760e+00
-3.14302295e-01 -7.67824709e-01 -3.34566504e-01 5.56264281e-01
2.91339103e-02 -7.44668841e-01 6.45382226e-01 -2.67906815e-01
-7.45764494e-01 5.03087997e-01 1.36702549e+00 1.91448167e-01
-6.93903804e-01 4.52631414e-01 -2.45190576e-01 5.87897360e-01
1.21482575e+00 -6.08918548e-01 -6.00316286e-01 9.87249538e-02
4.37596202e-01 8.11537445e-01 8.03145915e-02 -9.90050018e-01
1.35888560e-02 -9.41583395e-01 -4.09073085e-01 -4.46539581e-01
3.32673252e-01 -7.23234415e-01 9.91605520e-02 1.00709617e+00
-4.61922944e-01 1.39341027e-01 3.36267769e-01 5.05275667e-01
-3.97360623e-02 -8.05269599e-01 9.17159855e-01 -6.58231854e-01
-6.77195609e-01 -1.24196261e-01 -8.34405959e-01 1.27919883e-01
1.30863476e+00 -1.33977637e-01 1.91735297e-01 -5.47090530e-01
-1.14464200e+00 7.28644907e-01 3.68544400e-01 -3.29773240e-02
1.69957504e-01 -7.28906870e-01 -3.80848438e-01 -1.76785544e-01
-5.97985089e-01 -9.43633616e-02 -8.63451138e-02 9.75846827e-01
-3.66474420e-01 6.08278334e-01 -3.90014440e-01 -1.52404338e-01
-1.29021883e+00 6.27608716e-01 7.40006924e-01 -9.68510211e-01
-3.23107064e-01 2.20104754e-01 -1.63780376e-01 -1.14778370e-01
2.54988432e-01 -5.09700119e-01 -1.73171327e-01 1.15534484e-01
3.60150844e-01 6.13706052e-01 -3.31213593e-01 6.93536475e-02
-1.57691520e-02 3.36247921e-01 6.83604851e-02 -8.01381230e-01
1.45230198e+00 -2.46837839e-01 7.19796717e-02 5.61376810e-01
4.70443875e-01 -4.41198766e-01 -1.62426317e+00 -5.52758761e-02
3.14272076e-01 -2.47890741e-01 2.89598674e-01 -7.37612486e-01
-6.37981057e-01 6.42210722e-01 5.56762636e-01 4.08164352e-01
1.14643204e+00 -7.28830159e-01 7.35477209e-02 6.34582520e-01
7.65657067e-01 -1.53519714e+00 -2.16260806e-01 7.32408345e-01
8.06977749e-01 -7.45150447e-01 2.78328836e-01 8.29209760e-02
-5.59823513e-01 1.22593534e+00 4.67790812e-01 -3.73294204e-01
2.72247225e-01 1.55287638e-01 -3.19411159e-01 4.25981611e-01
-1.29437351e+00 -5.89623868e-01 -4.51675326e-01 5.20966530e-01
-2.74646252e-01 8.05903003e-02 -8.33787680e-01 1.64803341e-01
9.77974236e-02 3.70050997e-01 8.18557978e-01 1.51607108e+00
-6.87359393e-01 -1.59081054e+00 -5.91220677e-01 1.40312344e-01
-3.54679376e-01 2.62175620e-01 -1.96058169e-01 1.00762999e+00
-2.44006723e-01 9.41151559e-01 -3.60234052e-01 1.63882405e-01
1.76577106e-01 1.68515176e-01 8.38708937e-01 -3.95715505e-01
-7.95438945e-01 3.33669603e-01 4.88875568e-01 -9.60026443e-01
-3.79502535e-01 -7.18425870e-01 -1.23305190e+00 -1.03505075e-01
-2.16282427e-01 3.79319191e-01 5.83538353e-01 1.08238256e+00
1.37247995e-01 3.63193959e-01 8.63506734e-01 -9.49187994e-01
-1.24815130e+00 -6.14841163e-01 -6.73895359e-01 -3.19082320e-01
4.20199811e-01 -8.40775192e-01 -1.39568850e-01 -4.98632520e-01] | [4.191315650939941, 2.1818580627441406] |
1f8b9a67-8589-4164-afea-2c14e73b8e62 | svcnet-scribble-based-video-colorization | 2303.11591 | null | https://arxiv.org/abs/2303.11591v1 | https://arxiv.org/pdf/2303.11591v1.pdf | SVCNet: Scribble-based Video Colorization Network with Temporal Aggregation | In this paper, we propose a scribble-based video colorization network with temporal aggregation called SVCNet. It can colorize monochrome videos based on different user-given color scribbles. It addresses three common issues in the scribble-based video colorization area: colorization vividness, temporal consistency, and color bleeding. To improve the colorization quality and strengthen the temporal consistency, we adopt two sequential sub-networks in SVCNet for precise colorization and temporal smoothing, respectively. The first stage includes a pyramid feature encoder to incorporate color scribbles with a grayscale frame, and a semantic feature encoder to extract semantics. The second stage finetunes the output from the first stage by aggregating the information of neighboring colorized frames (as short-range connections) and the first colorized frame (as a long-range connection). To alleviate the color bleeding artifacts, we learn video colorization and segmentation simultaneously. Furthermore, we set the majority of operations on a fixed small image resolution and use a Super-resolution Module at the tail of SVCNet to recover original sizes. It allows the SVCNet to fit different image resolutions at the inference. Finally, we evaluate the proposed SVCNet on DAVIS and Videvo benchmarks. The experimental results demonstrate that SVCNet produces both higher-quality and more temporally consistent videos than other well-known video colorization approaches. The codes and models can be found at https://github.com/zhaoyuzhi/SVCNet. | ['Mengyang Liu', 'Yujia Zhang', 'Pengfei Xian', 'Wing-Yin Yu', 'Xuehui Wang', 'Kangcheng Liu', 'Lai-Man Po', 'Yuzhi Zhao'] | 2023-03-21 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 6.12022402e-03 -6.36072934e-01 -1.24884374e-01 -1.93356931e-01
-1.58335999e-01 -5.47008634e-01 2.20846925e-02 -5.84361970e-01
-4.14608240e-01 5.63106954e-01 -2.12403201e-02 -2.15862423e-01
2.07553953e-01 -7.31094956e-01 -8.69144082e-01 -7.29499936e-01
2.54605174e-01 -1.95752427e-01 5.12836218e-01 -6.29352406e-02
1.60898343e-01 4.30677295e-01 -1.19050038e+00 3.42329115e-01
1.14403200e+00 1.10582006e+00 2.61815608e-01 7.66216815e-01
-2.36154169e-01 6.75325871e-01 -4.93009061e-01 -2.39485458e-01
4.06871110e-01 -4.75817382e-01 -5.21764219e-01 3.20291460e-01
5.94161630e-01 -7.09360838e-01 -3.82417262e-01 1.36305869e+00
1.75029829e-01 2.77445167e-01 2.34458387e-01 -1.36114275e+00
-1.26106596e+00 5.38599432e-01 -1.10098112e+00 2.95849770e-01
2.83407569e-01 3.42239887e-01 6.06543243e-01 -9.88120735e-01
5.17050326e-01 1.54443538e+00 2.83980221e-01 6.75485909e-01
-1.12004745e+00 -1.10466421e+00 5.32052696e-01 3.13888729e-01
-1.43658102e+00 -1.61980346e-01 7.47446299e-01 -1.83218375e-01
1.63089201e-01 2.78835446e-01 9.91409123e-01 7.74481356e-01
-9.90246609e-02 9.16584849e-01 1.03753388e+00 -5.81768379e-02
2.34966710e-01 -3.28722984e-01 5.38738770e-03 8.86770427e-01
2.56437182e-01 1.69816196e-01 -4.44212645e-01 2.11403176e-01
1.46823001e+00 4.53670174e-01 -5.35713494e-01 -1.63869351e-01
-1.08611548e+00 4.97732103e-01 7.18979239e-01 2.33735099e-01
-1.65968642e-01 7.10317314e-01 3.13950360e-01 7.35048205e-02
3.52283329e-01 -1.27491072e-01 -4.05253828e-01 -1.57776661e-02
-1.06920850e+00 -6.79572672e-02 3.29373837e-01 1.29522753e+00
1.03240740e+00 3.02019477e-01 -4.58153158e-01 8.47768366e-01
2.02371478e-01 5.15071392e-01 1.33985534e-01 -1.34584069e+00
4.65377510e-01 3.80108953e-01 6.46941885e-02 -8.88400316e-01
2.46354491e-02 1.39787719e-01 -1.02081823e+00 4.22172368e-01
3.64221931e-01 -3.62522513e-01 -1.36403489e+00 1.48701620e+00
2.17205003e-01 6.83135927e-01 -1.62994757e-01 1.34485388e+00
7.29389846e-01 1.06416690e+00 2.53365964e-01 -9.96882543e-02
1.35644567e+00 -1.24717379e+00 -6.86425924e-01 -1.36913965e-02
-1.74778447e-01 -7.48418689e-01 1.30741179e+00 4.85583186e-01
-1.14938533e+00 -7.06074655e-01 -1.12685597e+00 -3.78279835e-01
-1.65628031e-01 2.54753828e-01 6.60993099e-01 4.31959808e-01
-1.20874774e+00 4.72267210e-01 -9.16961670e-01 -1.14570126e-01
5.59948385e-01 2.21561361e-02 -1.51559524e-02 -3.36453736e-01
-1.19580507e+00 1.71946079e-01 4.67675090e-01 2.92366326e-01
-8.93748701e-01 -7.03817427e-01 -6.90967441e-01 2.14605674e-01
4.33029801e-01 -6.43264651e-01 8.54502976e-01 -1.54673898e+00
-1.66135836e+00 4.54140395e-01 -1.60728544e-01 -1.83783229e-02
6.46486819e-01 -3.23829263e-01 -4.17899996e-01 3.82077187e-01
-1.16459891e-01 8.67347598e-01 1.14982080e+00 -1.44278932e+00
-9.70322609e-01 7.02346638e-02 1.87105209e-01 2.63341665e-01
-3.04437518e-01 6.17076308e-02 -1.43734586e+00 -1.03073919e+00
-2.02035066e-02 -8.22271883e-01 -1.45282269e-01 4.34395164e-01
-4.78032887e-01 5.97145129e-03 1.02017546e+00 -7.76050925e-01
1.37752235e+00 -2.36721325e+00 3.09530526e-01 2.56769389e-01
4.09973443e-01 2.23238781e-01 -3.61774564e-01 -8.69091824e-02
-2.30603397e-01 2.38814056e-01 -1.38205156e-01 -1.92173794e-02
-3.06496620e-01 1.23800956e-01 -5.87425604e-02 2.57912636e-01
1.01152629e-01 8.53965878e-01 -9.08893824e-01 -6.16537273e-01
4.11751240e-01 6.72213376e-01 -6.85368598e-01 1.05274692e-01
-2.27138042e-01 2.89211273e-01 -4.70442027e-01 7.07217336e-01
1.14793408e+00 -1.45149395e-01 8.89781490e-03 -6.77513599e-01
-1.56038448e-01 -5.61364651e-01 -1.34245861e+00 1.80787098e+00
-3.89097720e-01 6.04316592e-01 1.78763568e-01 -4.15720165e-01
6.99667156e-01 -3.89967263e-02 5.15112162e-01 -6.02307200e-01
2.56979674e-01 -8.27294439e-02 -4.81130481e-01 -3.78554463e-01
6.48057461e-01 2.71878242e-01 1.12398691e-01 2.45700255e-01
-2.07078278e-01 5.29525131e-02 4.60594654e-01 4.43058640e-01
4.89833444e-01 3.43197078e-01 -2.66253382e-01 -1.74543217e-01
6.41228795e-01 -3.23488176e-01 1.03527343e+00 3.23055655e-01
-1.72622442e-01 7.24700451e-01 4.54013824e-01 -3.88796389e-01
-1.03755724e+00 -1.25159681e+00 2.74459839e-01 1.24273336e+00
8.33899856e-01 -2.14103147e-01 -8.98651421e-01 -4.28958178e-01
-2.42196433e-02 4.73293126e-01 -6.78445041e-01 -1.37598574e-01
-6.10644639e-01 -4.28852707e-01 2.51980990e-01 6.87496960e-01
8.81410837e-01 -8.85867953e-01 -4.44141448e-01 -8.08321163e-02
-1.97522879e-01 -9.12892103e-01 -1.26333058e+00 -2.54167885e-01
-7.06119180e-01 -1.01292050e+00 -1.05849278e+00 -9.15985763e-01
7.61893451e-01 6.13915145e-01 8.11113596e-01 4.15245295e-01
-1.95956603e-01 1.96810171e-01 -5.66390336e-01 1.07889988e-01
-8.86776224e-02 -3.00758213e-01 -2.19407722e-01 2.69147933e-01
1.85454771e-01 -4.90495652e-01 -1.01256835e+00 2.84013659e-01
-1.28812861e+00 5.47373712e-01 5.18200159e-01 7.55493045e-01
6.61340237e-01 6.15019612e-02 -3.57913366e-03 -7.47533262e-01
4.59754080e-01 -1.13457471e-01 -8.05052817e-01 3.54811788e-01
-3.79770875e-01 3.46702337e-02 7.32713699e-01 -6.21361673e-01
-1.27396262e+00 3.40801328e-02 1.94016010e-01 -8.36564541e-01
8.70571285e-02 1.76922828e-01 -1.70782670e-01 -1.18358634e-01
1.44088134e-01 1.77831411e-01 -1.69577569e-01 -4.13424194e-01
8.96834552e-01 3.99920076e-01 8.14316690e-01 -6.41132712e-01
1.02674496e+00 5.59175313e-01 -4.88997459e-01 -4.03721422e-01
-5.26512802e-01 -1.98994819e-02 -4.63509768e-01 -4.08959061e-01
1.29104924e+00 -1.06913686e+00 -8.32728148e-01 6.05775476e-01
-1.02930844e+00 -5.95558822e-01 -4.46275771e-02 2.31983885e-01
-2.14837089e-01 5.76069593e-01 -1.04946995e+00 -3.15195262e-01
-4.82992142e-01 -1.09454906e+00 9.34564888e-01 9.13069904e-01
4.37926710e-01 -6.19558930e-01 -3.43396813e-01 -5.98899275e-02
2.70423830e-01 1.83735937e-01 7.52086043e-01 4.14172679e-01
-9.02011335e-01 3.55123699e-01 -8.12482059e-01 4.29738104e-01
1.89801842e-01 6.37036324e-01 -5.60444295e-01 -4.51581120e-01
-5.16897798e-01 2.33484507e-02 1.14086473e+00 5.08796096e-01
1.60402226e+00 -2.60193318e-01 -1.56902913e-02 1.20305359e+00
1.76205933e+00 5.18378794e-01 7.91086316e-01 3.09877425e-01
1.11826432e+00 6.45871162e-02 5.98566353e-01 4.72332239e-01
2.64983505e-01 4.19693559e-01 3.05401355e-01 -5.92704654e-01
-4.50997651e-01 -2.89944589e-01 4.16224331e-01 6.86523080e-01
-2.73960084e-01 -1.12965427e-01 -3.65485311e-01 1.72850266e-01
-1.77811050e+00 -1.08578718e+00 7.00576454e-02 1.85884690e+00
9.41953838e-01 -9.21261162e-02 3.42352851e-03 -2.36202285e-01
1.01314151e+00 2.89687514e-01 -8.07265639e-01 -2.50135452e-01
-2.00533524e-01 4.36368212e-02 7.18797505e-01 5.18516719e-01
-9.47754562e-01 1.22468781e+00 5.12588215e+00 9.32009816e-01
-1.24200296e+00 6.45587444e-02 9.47975516e-01 -1.40656367e-01
-4.48428780e-01 -8.31380710e-02 -3.80891293e-01 8.13782871e-01
3.45687605e-02 -1.49330810e-01 1.12967980e+00 6.35760844e-01
4.71991956e-01 -2.79226512e-01 -8.56876493e-01 1.28360009e+00
3.32831144e-02 -1.36052370e+00 2.40770072e-01 -5.31418741e-01
1.05627310e+00 -2.83488780e-01 2.89477080e-01 -2.47740652e-02
5.11863828e-01 -7.04442263e-01 1.12237918e+00 5.46249986e-01
1.23379028e+00 -8.79095137e-01 1.97070524e-01 -4.92008299e-01
-1.79950702e+00 -1.78123757e-01 -4.33785439e-01 4.41722780e-01
2.31564492e-01 3.67272496e-01 9.74352881e-02 4.61047977e-01
1.00881708e+00 1.06118214e+00 -6.19504392e-01 1.03081596e+00
-4.26888525e-01 3.88970524e-01 -1.23544455e-01 1.92782268e-01
2.44337782e-01 -6.66113675e-01 1.38102502e-01 1.22437239e+00
2.48776957e-01 3.21262896e-01 2.44565457e-01 1.10387611e+00
-1.64810434e-01 -1.73586056e-01 2.52453327e-01 1.63773775e-01
8.34774375e-01 1.38748395e+00 -7.76584983e-01 -6.60619497e-01
-5.71442544e-01 1.54881346e+00 6.79981187e-02 1.05943012e+00
-1.23331881e+00 -5.54403663e-01 7.53457844e-01 -3.09993118e-01
4.53009456e-01 -3.56025785e-01 -3.33195299e-01 -1.26694977e+00
-2.40060449e-01 -7.14183211e-01 4.51497346e-01 -1.19611943e+00
-1.07425702e+00 6.17073596e-01 -1.11454576e-01 -1.29134357e+00
4.21401978e-01 -5.51155090e-01 -8.16068232e-01 7.28255630e-01
-1.58021057e+00 -9.95375037e-01 -9.12675321e-01 9.46032941e-01
5.66126883e-01 1.92322746e-01 2.42977198e-02 5.58856606e-01
-9.50073242e-01 5.58766782e-01 1.72984630e-01 3.24305385e-01
8.22661877e-01 -1.15394747e+00 3.72429609e-01 1.25813520e+00
-2.50608504e-01 3.95417541e-01 4.18631196e-01 -6.54795349e-01
-1.43581438e+00 -1.38609695e+00 -8.72947499e-02 1.02787100e-01
3.87924373e-01 -1.58055335e-01 -8.50681186e-01 5.46628296e-01
2.93449461e-01 1.21307939e-01 1.67437270e-02 -4.41576511e-01
-4.11342174e-01 -4.43490028e-01 -7.66943216e-01 8.34773481e-01
9.89274204e-01 -3.71331751e-01 -1.70655072e-01 -5.37985712e-02
9.63443995e-01 -5.18418610e-01 -6.26000583e-01 -3.25980112e-02
4.53102261e-01 -9.94556487e-01 1.06528330e+00 -2.01219365e-01
6.06609702e-01 -8.04550827e-01 2.21423671e-01 -1.13431978e+00
-4.61330861e-01 -5.81998706e-01 2.51888663e-01 1.22748494e+00
9.58743542e-02 -3.04239929e-01 7.62961864e-01 7.70475686e-01
-1.38196692e-01 -5.81354260e-01 -5.22994816e-01 -6.15412951e-01
-7.62214065e-02 -1.61264420e-01 7.44648397e-01 8.02301943e-01
-3.02713394e-01 -7.10159689e-02 -5.92830002e-01 1.48799002e-01
6.80703461e-01 3.63153636e-01 6.36351168e-01 -7.12517440e-01
-1.23964384e-01 -6.08676016e-01 1.09892324e-01 -1.34487784e+00
-2.01983050e-01 -6.19920373e-01 3.50703932e-02 -1.52971303e+00
4.30023283e-01 -4.26544815e-01 -4.44722474e-01 4.92681950e-01
-5.60178876e-01 3.87172729e-01 5.94715238e-01 1.54117979e-02
-7.21427917e-01 5.07212222e-01 1.65657163e+00 -2.37755254e-01
-4.43275899e-01 -5.93546331e-01 -6.64679348e-01 6.84204102e-01
6.72474444e-01 -9.72532332e-02 -4.40017402e-01 -7.94195950e-01
-9.18678790e-02 9.72508639e-02 4.07811224e-01 -8.88999283e-01
2.16145068e-01 -6.11019850e-01 8.19758713e-01 -4.47828084e-01
1.40016139e-01 -8.43110681e-01 3.76796186e-01 4.56928968e-01
-2.66676068e-01 2.34714866e-01 8.56305435e-02 6.71550274e-01
-1.28548220e-01 1.48341313e-01 1.08859539e+00 -2.63842512e-02
-1.27833569e+00 8.57723773e-01 -2.59233952e-01 4.84285280e-02
1.10730112e+00 -2.95545012e-01 -4.09333050e-01 -3.25504988e-01
-5.37180424e-01 4.22769189e-01 7.83330619e-01 5.51323533e-01
9.52625394e-01 -1.57885921e+00 -5.32961309e-01 2.43057832e-01
-1.58925638e-01 3.03232465e-02 6.36812449e-01 5.80425918e-01
-1.05732739e+00 -1.78360775e-01 -5.57030797e-01 -3.23321491e-01
-1.17880225e+00 8.63808930e-01 3.42727989e-01 3.48211080e-01
-7.04578638e-01 1.02407217e+00 5.75050056e-01 4.26529616e-01
3.31086308e-01 -5.83759308e-01 -6.44368455e-02 -2.08710447e-01
5.91430068e-01 4.96302545e-01 -6.57679915e-01 -4.00516897e-01
-2.53767669e-01 8.81392062e-01 -1.18145920e-01 -1.14257187e-01
9.79632199e-01 -5.43731749e-01 -2.95868069e-01 -6.48475736e-02
1.21980464e+00 -1.56935956e-02 -1.94703245e+00 -2.54918098e-01
-5.25692403e-01 -8.65253508e-01 6.66078255e-02 -6.29705191e-01
-1.99284852e+00 7.24747777e-01 6.65054440e-01 -1.28137112e-01
1.64055359e+00 -2.98933089e-01 1.12954712e+00 -2.27747574e-01
1.76924929e-01 -1.23913419e+00 3.69144827e-01 2.34082714e-01
7.32587039e-01 -8.28140736e-01 2.74511296e-02 -6.12436056e-01
-8.78767014e-01 1.27172625e+00 9.47623909e-01 -3.19441915e-01
4.13010746e-01 2.62436301e-01 2.20691830e-01 1.11292839e-01
-4.25117880e-01 -9.05264392e-02 2.78851628e-01 4.52670217e-01
2.09227383e-01 1.13069206e-01 -3.95793557e-01 5.94814718e-01
3.11454087e-01 8.57536793e-02 6.44444287e-01 4.09736931e-01
-3.78029376e-01 -8.17900658e-01 -3.98001164e-01 2.14283064e-01
-1.76635236e-01 -3.24573517e-01 -3.15652825e-02 4.87858504e-01
3.16185087e-01 9.36622739e-01 2.15788275e-01 -5.84713936e-01
1.59109473e-01 -5.39581597e-01 2.34088510e-01 -9.70146134e-02
-2.44682938e-01 4.12355751e-01 -2.63769984e-01 -1.06904805e+00
-4.23701316e-01 -3.51123780e-01 -1.49839163e+00 -6.50357783e-01
-9.12869424e-02 1.01293854e-01 1.85010836e-01 4.20186967e-01
2.58461386e-01 8.41526270e-01 7.25921810e-01 -1.05135334e+00
8.40299577e-02 -5.11119485e-01 -7.32244730e-01 6.55951560e-01
3.44592988e-01 -3.90886635e-01 -1.39986262e-01 5.24138629e-01] | [11.113011360168457, -1.2280457019805908] |
84b4b3e9-07b1-4a30-be23-3c367e725259 | a-neural-template-matching-method-to-detect | 2209.11791 | null | https://arxiv.org/abs/2209.11791v1 | https://arxiv.org/pdf/2209.11791v1.pdf | A Neural Template Matching Method to Detect Knee Joint Areas | In this paper, new methods are considered to detect knee joint areas in bilateral PA fixed flexion knee X-ray images. The methods are of template matching type where the distance criterion is based on the negative normalized cross-correlation. The manual annotations are made on only one side of a single bilateral image when the templates are selected. The best matching patch search is formulated as an unconstrained continuous domain minimization problem. For the minimization problem different optimization methods are considered. The main method of the paper is a trainable optimizer where the method is taught to take zoomed and possibly rotated patches from its input images which look like the template. In the experiments, we compare the minimum values found by different optimization methods. We also look at some test images to examine the correspondence between the minimum value and how well the knee area is localized. It seems that making annotations only to a single image enables to detect knee joint areas quite precisely. | ['Juha Tiirola'] | 2022-09-23 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 1.43616036e-01 3.75525773e-01 -4.12616253e-01 -2.73147196e-01
-8.07154953e-01 -2.11032573e-02 1.95406973e-01 8.51116404e-02
-7.04329729e-01 7.55251169e-01 -1.25781223e-01 3.03818852e-01
-3.46747249e-01 -3.91533017e-01 -5.90656042e-01 -6.07507527e-01
-2.84073591e-01 7.77670383e-01 7.81951547e-01 -2.90849715e-01
3.26410562e-01 4.50209916e-01 -1.33233440e+00 3.34257007e-01
6.05239391e-01 5.41423678e-01 5.04166663e-01 4.31262732e-01
3.23898643e-01 7.90094957e-02 -8.01675141e-01 -7.59482086e-02
5.08232176e-01 -4.16469932e-01 -7.24082649e-01 4.72752094e-01
6.51941180e-01 5.43753710e-03 1.56588852e-01 1.08037889e+00
5.37996948e-01 8.06897953e-02 6.70725167e-01 -9.75028634e-01
2.29768082e-01 3.15981135e-02 -6.88889980e-01 3.00831348e-01
5.66272140e-01 -2.73987409e-02 7.84833610e-01 -9.05839026e-01
1.18856573e+00 9.10030544e-01 4.00547177e-01 3.93569589e-01
-1.36153662e+00 -1.43787459e-01 -2.58784652e-01 4.22351897e-01
-1.37001479e+00 -6.38412982e-02 7.76646137e-01 -5.05985260e-01
9.17154253e-01 5.25672913e-01 8.31165135e-01 4.15241808e-01
7.97597051e-01 6.07736826e-01 1.47289681e+00 -7.50450730e-01
1.76088467e-01 1.73458159e-01 -2.21317634e-02 6.59292877e-01
6.19185157e-02 3.77807356e-02 -3.32247525e-01 -5.96162584e-03
1.01214302e+00 -6.66226074e-02 -4.88166451e-01 -9.64524388e-01
-1.34547627e+00 7.08877802e-01 5.94508648e-01 7.71138310e-01
-4.17748868e-01 -9.32643190e-02 7.35192001e-02 1.85372889e-01
2.06758812e-01 6.98707461e-01 -2.68331826e-01 1.25293136e-01
-9.88258660e-01 3.69020849e-01 5.97138882e-01 4.41866904e-01
6.25324070e-01 -3.26814234e-01 2.23054588e-01 8.32056105e-01
2.44947195e-01 6.40080348e-02 5.53063452e-01 -7.98277318e-01
2.91130096e-01 6.55837297e-01 -1.49779126e-01 -8.82313907e-01
-4.11952585e-01 -3.71425241e-01 -2.40428478e-01 8.41834664e-01
7.47780979e-01 4.58315900e-03 -1.03931379e+00 1.14473748e+00
5.32657206e-01 -4.48464066e-01 -3.95408511e-01 1.27623820e+00
2.61811584e-01 2.39323631e-01 -2.07897007e-01 -4.27923858e-01
1.40306938e+00 -1.01460242e+00 -6.88071251e-01 -3.25322479e-01
5.59800088e-01 -1.08362067e+00 1.07823336e+00 4.25596982e-01
-1.13150597e+00 -5.40192664e-01 -1.36925185e+00 2.67278045e-01
-2.81576633e-01 3.79932880e-01 -8.53888318e-02 2.63899416e-01
-7.64485300e-01 9.39497888e-01 -8.28874350e-01 -4.28824872e-01
-2.92788863e-01 6.10364437e-01 -7.79885113e-01 3.21463823e-01
-8.97119462e-01 1.29192209e+00 2.70964265e-01 2.18475670e-01
-3.32018375e-01 -1.60383329e-01 -6.05088353e-01 -2.52288193e-01
4.93890047e-01 -5.60024679e-01 9.42080796e-01 -1.26556730e+00
-1.55473137e+00 1.56078684e+00 1.61071897e-01 -2.49916092e-01
8.97432804e-01 -2.57213205e-01 -1.02035150e-01 4.49059397e-01
2.20005631e-01 5.93117118e-01 8.42794061e-01 -1.29822850e+00
-3.45899016e-01 -5.07073045e-01 -1.79137975e-01 1.84682325e-01
3.35190475e-01 6.63135052e-02 -4.23924536e-01 -9.10110831e-01
4.48820442e-01 -1.05982959e+00 -4.41926032e-01 1.25619426e-01
-3.67648631e-01 -1.71141416e-01 6.79293633e-01 -8.37332010e-01
1.05351567e+00 -1.99518979e+00 3.76757979e-01 6.02612853e-01
4.61232476e-02 -2.25311264e-01 3.01597357e-01 3.05870801e-01
-4.09412980e-01 -4.02171522e-01 -2.02458397e-01 9.72254574e-02
-1.84877679e-01 2.85388172e-01 2.99622983e-01 8.51871610e-01
-1.79593354e-01 2.61229694e-01 -4.12906617e-01 -9.11644280e-01
1.40420765e-01 1.88654944e-01 -4.17250842e-01 2.93318868e-01
2.59705465e-02 3.38265717e-01 -2.68047810e-01 4.65855896e-01
4.71859276e-01 4.84916493e-02 4.01177078e-01 -5.30249953e-01
-3.09325214e-02 -4.27554213e-02 -1.59400666e+00 1.77798891e+00
-7.40169659e-02 5.45932710e-01 1.89798176e-01 -1.21670032e+00
9.51418996e-01 4.56233174e-01 5.69132507e-01 -7.30763972e-01
8.14009532e-02 6.06822550e-01 3.86685580e-01 -5.95624685e-01
1.37143925e-01 -1.85778588e-01 2.40729958e-01 1.58953175e-01
1.20968781e-01 -1.70730412e-01 4.27998960e-01 -2.98745841e-01
7.07465649e-01 1.84443310e-01 4.87663507e-01 -5.84122181e-01
6.76744878e-01 2.52991050e-01 3.49312276e-01 3.33030611e-01
-7.14108571e-02 7.97734439e-01 5.22533298e-01 -5.56277037e-01
-1.05676663e+00 -1.13358438e+00 -4.34314460e-01 5.24859428e-01
1.66181087e-01 -2.47587055e-01 -8.03059399e-01 -6.94971442e-01
-9.30674598e-02 5.99569902e-02 -6.29120231e-01 1.03243925e-01
-1.11531138e+00 -4.15703267e-01 -3.74145478e-01 3.91543746e-01
1.88260630e-01 -9.57225740e-01 -1.12518394e+00 7.38720149e-02
3.07927821e-02 -6.20772004e-01 -3.52305055e-01 3.41557115e-01
-1.34437084e+00 -1.10202837e+00 -1.26086438e+00 -1.10863364e+00
9.07271326e-01 -3.21538776e-01 1.05225003e+00 1.93884626e-01
-6.26826823e-01 2.53686696e-01 -5.57809658e-02 2.77035329e-02
-2.68674493e-01 -4.84060943e-02 -2.04045195e-02 -2.18138933e-01
-4.40308750e-02 -3.14333081e-01 -6.80823505e-01 6.94484591e-01
-5.61857879e-01 -2.07262769e-01 8.33035707e-01 7.90451229e-01
1.01321805e+00 -2.81969011e-01 -9.62238386e-02 -4.89443898e-01
6.02239072e-01 8.00482705e-02 -4.64654088e-01 3.43977094e-01
-5.78828454e-01 1.95743471e-01 1.34210408e-01 -5.09574711e-01
-5.70738673e-01 3.20257485e-01 1.50716295e-02 -3.72546643e-01
-2.26895228e-01 2.40401641e-01 -1.83094461e-02 -3.43705505e-01
6.70910835e-01 -1.68912902e-01 3.34783524e-01 -6.55063748e-01
-7.16091767e-02 3.04615378e-01 4.73125309e-01 -4.45638388e-01
4.25856143e-01 3.07211190e-01 2.59093884e-02 -8.82335842e-01
-2.22467408e-01 -5.88944614e-01 -8.88507485e-01 -6.87208116e-01
9.82245445e-01 -2.96204567e-01 -3.72637361e-01 -1.59461543e-01
-9.12079334e-01 -2.86596030e-01 -4.13231105e-01 1.03786802e+00
-9.26242948e-01 4.46977854e-01 -5.37755668e-01 -3.03125739e-01
-2.65450269e-01 -1.47666049e+00 1.20463920e+00 1.21169634e-01
-4.43817228e-01 -8.44553053e-01 4.84250039e-01 2.54498690e-01
-2.27717832e-01 1.86885074e-01 8.84406865e-01 -6.70270503e-01
-4.88492399e-01 -5.57406604e-01 3.58468622e-01 2.25795001e-01
-7.33075812e-02 -6.35726079e-02 -5.38474500e-01 -2.84693867e-01
2.55480886e-01 7.77112991e-02 5.52227736e-01 5.32647371e-01
7.59707153e-01 1.03491426e-01 -6.27730966e-01 2.91747361e-01
1.52415049e+00 3.75945300e-01 6.62341416e-01 8.08314502e-01
4.44396548e-02 6.66235328e-01 1.00657296e+00 -8.12843665e-02
-4.45780277e-01 1.33227623e+00 4.03008223e-01 -3.46226782e-01
-7.67083019e-02 1.75281689e-01 1.39739543e-01 7.42966831e-01
-5.45811772e-01 2.39632308e-01 -7.79452085e-01 4.02568907e-01
-1.63568580e+00 -6.74157083e-01 -5.82462214e-02 2.38374496e+00
7.93760061e-01 4.58333939e-01 2.82037795e-01 3.03955913e-01
6.45097256e-01 -5.07015996e-02 1.03518561e-01 -3.92105967e-01
1.50017321e-01 3.51477414e-01 3.44212860e-01 7.60768890e-01
-1.09490252e+00 3.10828328e-01 6.70422411e+00 8.28108728e-01
-1.21582425e+00 -5.82287312e-02 1.83916211e-01 -2.83027012e-02
2.42588043e-01 6.11828640e-02 -4.29883391e-01 4.81962144e-01
1.11428149e-01 1.25837326e-01 -8.79505575e-02 8.71503949e-01
2.14282032e-02 -8.47463906e-01 -1.04160857e+00 7.72092998e-01
1.21730290e-01 -9.68336463e-01 -3.63697857e-01 2.46915296e-01
4.92188901e-01 -4.03260350e-01 -2.06428900e-01 -2.31672138e-01
-5.06757081e-01 -9.54658806e-01 3.90426189e-01 8.37336361e-01
3.81854296e-01 -5.02604723e-01 8.08783591e-01 2.09140554e-01
-9.42054570e-01 3.37118208e-01 -2.77555406e-01 1.93032756e-01
2.27411613e-01 2.93950826e-01 -8.85225892e-01 2.87160844e-01
6.95267200e-01 1.97013691e-01 -6.92179978e-01 1.51868737e+00
-2.27202788e-01 9.65335295e-02 -6.15565419e-01 3.79317137e-03
1.18942402e-01 -5.50911844e-01 9.03892398e-01 9.59742367e-01
2.16152996e-01 -2.40373760e-01 3.28405440e-01 5.83537042e-01
6.03616834e-01 5.81040323e-01 -4.50401753e-01 5.80526948e-01
-3.07434231e-01 1.12077606e+00 -1.00053287e+00 -1.98461309e-01
-2.11244673e-01 9.50800538e-01 -4.13759705e-03 2.58070797e-01
-5.11811972e-01 -4.41781491e-01 2.18609035e-01 6.45814240e-01
1.72972575e-01 -1.76339805e-01 4.57530422e-03 -8.73663425e-01
3.52011353e-01 -8.92124474e-01 5.60299456e-01 -9.67175841e-01
-7.41592765e-01 4.55733925e-01 4.36287284e-01 -1.25111568e+00
-5.42319357e-01 -8.79521370e-01 -5.00523925e-01 7.56118774e-01
-6.76021934e-01 -5.49402237e-01 4.70200516e-02 5.40580094e-01
7.47399271e-01 -2.12438088e-02 5.29657245e-01 2.91897058e-01
-1.74582094e-01 1.64910927e-01 -7.43499026e-02 8.93420428e-02
8.13696384e-01 -1.57792592e+00 -4.84145582e-01 6.46680832e-01
2.90527850e-01 5.55550635e-01 1.17601061e+00 -7.84619451e-01
-8.63403976e-01 1.87976789e-02 9.82850552e-01 -1.47331730e-01
3.36931586e-01 -4.57194187e-02 -9.89328265e-01 4.74104345e-01
5.56115091e-01 -2.33571101e-02 1.90044791e-01 -2.28839159e-01
4.40572530e-01 -1.76926464e-01 -9.51508999e-01 4.22212929e-01
4.36244398e-01 -1.12932369e-01 -1.01100945e+00 5.88348448e-01
-2.49011397e-01 -6.57965899e-01 -1.08790350e+00 6.38067305e-01
5.79050362e-01 -9.81608450e-01 1.08480930e+00 -4.47732627e-01
2.49086767e-01 -4.48292792e-01 2.66298801e-01 -1.24710190e+00
1.05369277e-01 -1.98028952e-01 4.42388654e-01 4.05013293e-01
3.98351431e-01 -4.58765507e-01 1.05078590e+00 3.55433971e-02
1.73986048e-01 -1.09468114e+00 -1.23123813e+00 -7.57434309e-01
-2.22921386e-01 1.98474556e-01 -1.98145792e-01 6.87139094e-01
2.37882420e-01 1.50417581e-01 -8.00661370e-02 -4.36554588e-02
7.03086555e-01 4.14482564e-01 7.39077210e-01 -1.06163979e+00
-5.73411107e-01 -2.90357709e-01 -9.46894288e-01 -8.59251857e-01
-3.71559858e-01 -6.21886134e-01 -6.32194942e-03 -1.31812465e+00
-1.09183103e-01 -5.27273007e-02 2.15621386e-03 1.09737292e-01
9.40073375e-03 1.51874050e-01 2.29738519e-01 2.88512230e-01
1.15472404e-02 -3.15524451e-02 1.57239676e+00 3.85504849e-02
-2.46522978e-01 4.19407904e-01 5.19406259e-01 7.97964990e-01
7.13613153e-01 -6.06444716e-01 -1.43558592e-01 -3.68847251e-02
1.40423879e-01 3.21333051e-01 4.20013756e-01 -1.14912534e+00
1.63371325e-01 -2.03230362e-02 4.55633551e-01 -7.10753739e-01
4.47358608e-01 -1.03772211e+00 3.73717368e-01 7.96161771e-01
-1.08268231e-01 5.28253257e-01 -1.81594104e-01 5.61374165e-02
-3.54060501e-01 -6.59975529e-01 9.64332521e-01 -5.37843943e-01
-6.63796186e-01 -3.34994256e-01 -3.82632405e-01 -9.17761922e-02
1.17576575e+00 -6.78541303e-01 3.65567476e-01 -1.29527822e-01
-1.58588815e+00 -5.85002787e-02 6.28141165e-01 7.04718232e-02
6.20397389e-01 -1.17977595e+00 -3.05211276e-01 3.02954376e-01
-4.36026491e-02 -3.22542667e-01 -3.66471964e-03 1.28229511e+00
-8.60059977e-01 3.35317075e-01 -5.50922215e-01 -1.09826303e+00
-1.79917288e+00 6.08537376e-01 9.13592339e-01 -4.39817071e-01
-6.40339315e-01 3.76845479e-01 -1.62423383e-02 -5.38188927e-02
2.58801132e-01 -3.89237851e-01 -2.67377645e-01 -5.97087443e-02
6.03806935e-02 4.63337213e-01 2.68630207e-01 -6.70403659e-01
-3.70074242e-01 1.03543127e+00 1.36943728e-01 -3.36233318e-01
1.11691582e+00 2.71919489e-01 -1.83859542e-01 4.19767380e-01
1.19959581e+00 2.96400607e-01 -9.61800396e-01 1.05232589e-01
2.46177450e-01 -6.56878114e-01 -2.39629671e-01 -6.20377183e-01
-1.05591714e+00 8.56824219e-01 1.17890882e+00 9.48499292e-02
9.63084102e-01 1.00668661e-01 3.13021719e-01 8.89868736e-02
2.84990102e-01 -1.42928958e+00 2.02944815e-01 -2.23261714e-01
1.21203303e+00 -9.69210565e-01 5.04960597e-01 -5.42872012e-01
-5.21139860e-01 1.61206949e+00 8.07983756e-01 -6.79246247e-01
5.70545971e-01 3.79710168e-01 2.58926839e-01 -4.80747670e-01
-2.83384055e-01 -2.25738436e-01 6.83437526e-01 4.23191100e-01
3.97953182e-01 -7.53140524e-02 -1.15821302e+00 -5.13886213e-02
-1.08424634e-01 -3.86044718e-02 1.05704188e-01 1.16720366e+00
-6.39668167e-01 -1.24844432e+00 -6.79604292e-01 2.28185162e-01
-6.39884233e-01 4.49958146e-01 -4.23942208e-01 1.44196904e+00
1.66965649e-01 3.35626662e-01 1.09909616e-01 9.34919342e-02
7.16441154e-01 4.44379449e-02 8.97399008e-01 -4.70368177e-01
-6.24227583e-01 6.34919941e-01 1.12611577e-02 -7.42539585e-01
-5.70749223e-01 -6.10287130e-01 -1.13468432e+00 4.20686215e-01
-4.37497616e-01 3.54170471e-01 6.65948272e-01 8.75958323e-01
-2.45463073e-01 1.57678515e-01 1.79654017e-01 -7.70538568e-01
-5.74595451e-01 -7.17112601e-01 -6.60862088e-01 2.93411762e-01
6.77430034e-02 -9.82274830e-01 -2.90878862e-01 5.93390279e-02] | [13.952411651611328, -2.6635961532592773] |
3a646451-6322-4d1b-a478-c50eec61879a | egocol-egocentric-camera-pose-estimation-for | 2306.16606 | null | https://arxiv.org/abs/2306.16606v1 | https://arxiv.org/pdf/2306.16606v1.pdf | EgoCOL: Egocentric Camera pose estimation for Open-world 3D object Localization @Ego4D challenge 2023 | We present EgoCOL, an egocentric camera pose estimation method for open-world 3D object localization. Our method leverages sparse camera pose reconstructions in a two-fold manner, video and scan independently, to estimate the camera pose of egocentric frames in 3D renders with high recall and precision. We extensively evaluate our method on the Visual Query (VQ) 3D object localization Ego4D benchmark. EgoCOL can estimate 62% and 59% more camera poses than the Ego4D baseline in the Ego4D Visual Queries 3D Localization challenge at CVPR 2023 in the val and test sets, respectively. Our code is publicly available at https://github.com/BCV-Uniandes/EgoCOL | ['Pablo Arbeláez', 'Kevis-Kokitsi Maninis', 'Jordi Pont-Tuset', 'Maria Escobar', 'Cristhian Forigua'] | 2023-06-29 | null | null | null | null | ['pose-estimation', 'object-localization'] | ['computer-vision', 'computer-vision'] | [-8.30167234e-01 -2.82019854e-01 -3.79439533e-01 -1.26918137e-01
-1.24096143e+00 -1.37329006e+00 6.31201506e-01 -4.28853244e-01
-2.89077163e-01 -3.21980268e-02 5.32673180e-01 1.30395219e-01
4.33010757e-01 -1.65563539e-01 -9.62094665e-01 -3.41933072e-01
1.14528142e-01 5.58916509e-01 1.81654662e-01 5.25306463e-01
2.60846049e-01 8.07705522e-01 -1.13510287e+00 -1.47427201e-01
-1.27946749e-01 8.48305523e-01 2.00916260e-01 1.10399926e+00
3.70000780e-01 7.09637463e-01 -2.97862012e-02 -2.45241672e-01
5.39903581e-01 3.79698247e-01 -7.34883130e-01 2.26993382e-01
1.09469736e+00 -8.64333928e-01 -9.98711407e-01 1.10347593e+00
7.14363575e-01 5.35643920e-02 7.54349232e-01 -1.12045383e+00
-5.18160045e-01 -8.48183632e-02 -9.40033257e-01 4.90785211e-01
1.01470709e+00 3.00209105e-01 9.41727877e-01 -1.52511108e+00
1.02180004e+00 1.29247904e+00 7.67924309e-01 2.16537118e-01
-1.08644986e+00 -6.22621298e-01 -1.96342133e-02 1.73148051e-01
-2.02273393e+00 -7.81370401e-01 5.54254889e-01 -4.70086038e-01
1.21131289e+00 -1.77269742e-01 5.44944286e-01 1.45346797e+00
-6.45894036e-02 1.03245044e+00 5.59640706e-01 -3.36338207e-02
-1.47487642e-02 -1.26615241e-01 -6.14789911e-02 6.64463818e-01
2.62883186e-01 5.77039830e-02 -7.14412510e-01 -1.98640555e-01
8.85851622e-01 4.00317684e-02 -2.29545340e-01 -1.14604902e+00
-1.37593985e+00 6.07329130e-01 6.33051097e-01 -1.85656741e-01
-3.08245867e-01 6.81953847e-01 2.08828568e-01 -9.15050730e-02
5.82114100e-01 2.58361191e-01 -4.38820243e-01 -3.68374556e-01
-3.19622010e-01 2.42588818e-01 7.55025387e-01 1.98153627e+00
7.84646928e-01 -2.03977361e-01 6.10979497e-02 4.44282532e-01
5.51802039e-01 1.29008389e+00 -1.02720588e-01 -1.48352540e+00
4.79840130e-01 2.60599136e-01 3.97916079e-01 -1.15329683e+00
-1.30504280e-01 -2.36374199e-01 -7.49412552e-02 -4.45665568e-01
1.98234737e-01 1.42162725e-01 -8.48197937e-01 1.48380172e+00
6.99591815e-01 3.50438356e-01 -1.65979967e-01 1.16964006e+00
1.05595076e+00 3.22580904e-01 -1.97636738e-01 4.57828909e-01
1.42275155e+00 -8.68920803e-01 -3.09410155e-01 -3.92802000e-01
6.30877078e-01 -1.01967204e+00 5.18537343e-01 2.93338927e-03
-9.33130980e-01 -2.25318357e-01 -6.47622168e-01 -3.82686466e-01
-2.10389733e-01 2.85259932e-01 4.85353112e-01 5.60095966e-01
-1.25600517e+00 -3.20716709e-01 -9.30455446e-01 -7.60887861e-01
5.42493403e-01 1.39864951e-01 -8.53403091e-01 -5.11122227e-01
-6.46380365e-01 8.30875218e-01 1.06224850e-01 -2.40626186e-01
-1.66243994e+00 -6.34362757e-01 -1.12660611e+00 -1.65242702e-01
4.10501719e-01 -8.47844481e-01 1.46219277e+00 -2.15911176e-02
-8.79292965e-01 1.34322131e+00 -3.54779989e-01 -3.29279512e-01
5.71521342e-01 -6.07222080e-01 4.68043052e-02 3.95425111e-01
4.52438563e-01 8.24560523e-01 5.37986875e-01 -1.09228206e+00
-5.01684904e-01 -7.07742274e-01 -1.60088912e-02 5.50391376e-01
3.99836451e-01 -9.94274467e-02 -1.52678537e+00 -2.36955926e-01
5.88643014e-01 -1.28655529e+00 2.43821684e-02 1.09661430e-01
-4.62248206e-01 -1.92720428e-01 8.48179460e-01 -3.10328305e-01
2.69833267e-01 -2.24877119e+00 1.83461398e-01 2.53428165e-02
6.23014867e-01 -2.77956754e-01 -2.17190847e-01 2.30703682e-01
-1.96594163e-03 -2.27313995e-01 6.04734123e-01 -8.86890233e-01
5.67277856e-02 -1.63676172e-01 -2.05569372e-01 1.08616638e+00
-3.17728043e-01 1.08783364e+00 -1.04004443e+00 -2.96421438e-01
6.55756772e-01 7.24069118e-01 -7.47440875e-01 1.38265431e-01
9.05082151e-02 2.91253299e-01 -5.87887406e-01 1.24368572e+00
8.67205262e-01 -4.04973179e-01 -1.80761069e-01 -5.66004217e-01
1.19920433e-01 1.09437136e-02 -1.04123306e+00 2.28398585e+00
-1.71790347e-01 9.52142417e-01 -1.29025504e-01 -1.60153270e-01
6.40860617e-01 1.26165912e-01 5.44240475e-01 -2.55150408e-01
3.69580835e-01 1.98590010e-02 -1.02899730e+00 -3.63293260e-01
5.92244267e-01 7.39063263e-01 -2.61884630e-01 1.63098529e-01
4.87942487e-01 -4.42344517e-01 -2.07238078e-01 5.44806719e-01
1.48973107e+00 2.86745220e-01 3.60534430e-01 1.45788572e-03
2.47771338e-01 9.74243283e-02 1.74731627e-01 8.42556596e-01
-6.00907564e-01 9.36497867e-01 2.62963414e-01 -3.75321239e-01
-1.11544836e+00 -1.58003700e+00 2.63329092e-02 5.10546446e-01
6.44726932e-01 -6.49109781e-01 -5.52493751e-01 -8.27932060e-01
2.74793863e-01 4.35059309e-01 -4.03795689e-01 1.68549538e-01
-1.57606721e-01 6.82698041e-02 5.48431873e-01 3.73860061e-01
4.32781219e-01 -2.82658190e-01 -3.13924193e-01 -3.74626607e-01
-4.12463754e-01 -1.81040549e+00 -9.70563412e-01 -2.50684004e-02
-5.78828514e-01 -1.48254728e+00 -7.10022867e-01 -5.34513772e-01
6.01820111e-01 1.25169504e+00 1.12545824e+00 -6.04141235e-01
-3.19992006e-01 1.37460423e+00 -4.29238826e-01 -2.23029017e-01
1.36323810e-01 2.27414280e-01 4.75098163e-01 -2.54303873e-01
1.07950234e+00 -3.88624489e-01 -8.91316652e-01 5.73029578e-01
3.19053642e-02 -2.83964366e-01 2.88151085e-01 1.32592127e-01
9.71521676e-01 -8.15109372e-01 -2.51203477e-01 -4.63095427e-01
-2.10344866e-01 -7.95006573e-01 -1.03135157e+00 -2.64297873e-01
-1.53948441e-01 -3.77469212e-01 -3.11439127e-01 -1.43785492e-01
-4.38542932e-01 7.18046367e-01 1.09014198e-01 -1.59065163e+00
-3.11955035e-01 -2.62614995e-01 -2.09612966e-01 -4.61211145e-01
6.47179782e-01 2.43727937e-01 -2.73195207e-01 -4.79007721e-01
5.72409511e-01 5.01157463e-01 6.80089831e-01 -3.24410737e-01
1.05309618e+00 7.34993398e-01 -1.91412330e-01 -6.92737401e-01
-7.21260369e-01 -1.11400163e+00 -6.26282215e-01 -3.16527933e-01
1.13513505e+00 -1.89599252e+00 -8.45031917e-01 2.57442832e-01
-1.44533598e+00 -2.51918975e-02 -2.80116200e-02 8.47695827e-01
-9.18146908e-01 3.20899099e-01 -2.41721049e-01 -4.28591162e-01
-1.17041260e-01 -1.41493487e+00 1.84487998e+00 -8.80376250e-03
-1.01371728e-01 -6.26240790e-01 2.44087592e-01 4.02977198e-01
-8.91045406e-02 1.33021250e-01 -7.42534250e-02 -4.34500396e-01
-1.35687113e+00 -7.28827536e-01 -5.36859632e-01 -1.23373181e-01
-3.41867842e-02 -4.79553282e-01 -1.22840762e+00 -4.58367407e-01
-1.45992488e-01 -2.16611102e-01 6.25207722e-01 4.74144220e-01
8.25621665e-01 9.43603963e-02 -5.81801772e-01 1.18427336e+00
1.49937916e+00 -2.35343486e-01 3.62787724e-01 1.86866745e-01
8.96689773e-01 -6.80491934e-03 7.46138811e-01 4.84527290e-01
6.72663569e-01 9.06428516e-01 8.38917851e-01 4.44219083e-01
-2.37287194e-01 -6.54988229e-01 2.75644213e-01 3.07049006e-01
3.83483581e-02 -3.84481281e-01 -1.04001176e+00 8.57133865e-01
-1.59316349e+00 -8.34140360e-01 -3.31417508e-02 1.86850822e+00
7.00699389e-02 -3.02696168e-01 7.29776220e-03 -7.42163718e-01
7.73876369e-01 4.41272348e-01 -6.84378743e-01 4.97895092e-01
-2.11059168e-01 -5.23915470e-01 1.22936094e+00 5.60479999e-01
-1.33416760e+00 1.27126276e+00 6.36429071e+00 5.35851657e-01
-4.63039160e-01 5.40189981e-01 2.32250646e-01 -5.20252705e-01
7.14380592e-02 1.42013729e-01 -1.24066579e+00 1.96162820e-01
5.74140728e-01 6.52841032e-02 5.26424050e-01 1.41062951e+00
-5.86157925e-02 -3.17028314e-02 -1.15053618e+00 1.81227171e+00
3.19676459e-01 -1.58075082e+00 -1.33336306e-01 4.68843222e-01
6.63003623e-01 1.15096581e+00 -6.57915100e-02 4.98197339e-02
3.61371279e-01 -5.72341800e-01 8.93501103e-01 7.60136843e-02
8.75685871e-01 -5.74499309e-01 5.20118415e-01 1.20127112e-01
-1.40623379e+00 1.32417127e-01 -4.84978080e-01 4.30943847e-01
2.26822048e-01 2.44794432e-02 -1.04045308e+00 2.45791733e-01
1.25428307e+00 1.16514289e+00 -7.20493793e-01 1.04734862e+00
-3.18591475e-01 1.71515211e-01 -5.90146840e-01 1.89569980e-01
2.12453246e-01 1.08012423e-01 1.23040807e+00 1.01559091e+00
2.98843384e-01 1.17038399e-01 -4.45604511e-02 7.47080743e-01
-3.89822632e-01 -3.14293146e-01 -1.32182550e+00 2.89414793e-01
9.21816587e-01 1.10669541e+00 -4.72484976e-01 1.19882435e-01
-4.13555533e-01 1.09202087e+00 3.05951118e-01 5.80811858e-01
-1.06486571e+00 -6.13158680e-02 1.12677753e+00 6.28793612e-02
5.25742233e-01 -6.30514324e-01 3.40119928e-01 -1.48081434e+00
-1.48847133e-01 -4.52075750e-01 1.50097579e-01 -1.35418141e+00
-1.10990298e+00 2.79752582e-01 2.16093495e-01 -1.44045866e+00
-2.62000084e-01 -6.07399762e-01 1.12959174e-02 6.37507677e-01
-1.03983188e+00 -1.33796716e+00 -7.66226053e-01 7.80281544e-01
5.95097661e-01 -2.63437420e-01 4.34323758e-01 3.71210486e-01
-2.23718099e-02 5.31038582e-01 1.35930419e-01 4.09263223e-01
8.47441852e-01 -1.11587942e+00 9.78984833e-01 7.84077466e-01
4.43987280e-01 5.51913440e-01 5.20912886e-01 -4.89761919e-01
-2.25013781e+00 -1.13641775e+00 5.66952169e-01 -1.58956969e+00
5.59292138e-01 -8.18203211e-01 -8.44602808e-02 1.23171580e+00
4.62343059e-02 3.96430492e-01 2.00681746e-01 2.44903248e-02
-7.11092472e-01 1.32533669e-01 -9.58111823e-01 5.92325091e-01
1.47464335e+00 -9.02905524e-01 -3.20995539e-01 6.74410164e-01
9.85782444e-01 -1.02311504e+00 -8.51563811e-01 1.17399201e-01
5.13871729e-01 -7.39669025e-01 1.50503397e+00 -1.34045973e-01
-1.22717805e-01 -6.69791102e-01 -8.96951973e-01 -7.80654490e-01
-1.43713489e-01 -5.41043341e-01 -3.97675723e-01 8.09214234e-01
-2.09494177e-02 -4.85332578e-01 1.08144534e+00 2.09837914e-01
1.27142414e-01 -2.23019883e-01 -1.14601088e+00 -7.95943081e-01
-6.51757717e-01 -6.93402529e-01 4.56004083e-01 5.60265601e-01
-7.57021964e-01 2.69853324e-01 -3.48139971e-01 6.28613830e-01
1.05415559e+00 -1.28152370e-01 1.57336783e+00 -6.12966537e-01
-8.17306899e-03 -3.62117402e-02 -1.05586779e+00 -1.61382627e+00
3.76495808e-01 -8.64730060e-01 -1.77516177e-01 -1.08064044e+00
4.72138643e-01 -5.64803518e-02 1.28938317e-01 1.25851601e-01
3.81523222e-01 6.95981026e-01 2.37018943e-01 4.91588205e-01
-1.07837582e+00 4.44736123e-01 7.89021313e-01 -2.03238398e-01
1.17877431e-01 -1.15364172e-01 -3.86265159e-01 6.40555739e-01
2.41841033e-01 -6.15137875e-01 -3.21978509e-01 -9.33061779e-01
4.14018556e-02 1.08788103e-01 7.82878458e-01 -9.16362762e-01
4.54106420e-01 2.12548211e-01 7.59782910e-01 -1.25619471e+00
8.80970478e-01 -9.76642430e-01 3.64600360e-01 2.30142638e-01
6.20307550e-02 2.66744971e-01 1.68145582e-01 1.09397221e+00
2.92959716e-02 1.52301729e-01 5.57870030e-01 -3.08755964e-01
-1.12652302e+00 8.16936195e-01 -4.82717790e-02 2.12289810e-01
1.16462815e+00 -2.68730432e-01 -5.13711810e-01 -6.51458740e-01
-4.85482424e-01 3.84622455e-01 9.83359993e-01 7.81103551e-01
7.81658649e-01 -1.42148793e+00 -5.07438481e-01 8.13399479e-02
7.03264654e-01 8.31274986e-02 2.62837857e-01 8.77923846e-01
-9.46806073e-01 7.00804293e-01 2.49445111e-01 -1.43489230e+00
-1.23347998e+00 6.18986964e-01 4.96644825e-01 4.36540842e-01
-5.64328015e-01 9.28906977e-01 5.22259533e-01 -4.99476671e-01
3.54658276e-01 3.90298432e-03 3.90668064e-01 -2.43785739e-01
4.62313712e-01 2.98032075e-01 -2.63270050e-01 -1.06967986e+00
-8.87162864e-01 1.09197998e+00 -1.70945916e-02 1.21755488e-02
1.15711176e+00 -6.65357709e-01 2.79649109e-01 1.00732982e-01
1.59710968e+00 7.20475465e-02 -1.39606357e+00 -2.59320617e-01
-3.42372000e-01 -1.02876186e+00 2.33723208e-01 -1.96143672e-01
-8.66514683e-01 5.33266664e-01 8.86341274e-01 -4.57052529e-01
5.65894067e-01 7.51772225e-01 3.27553809e-01 7.07198024e-01
7.98271716e-01 -5.88605940e-01 5.13675846e-02 6.84972584e-01
7.94124067e-01 -1.52876413e+00 1.18781388e-01 -2.88801938e-01
-5.08069932e-01 6.57520652e-01 5.88596642e-01 -4.43181992e-01
5.92903435e-01 -3.59040871e-02 -3.02903410e-02 -4.67518151e-01
-5.01360714e-01 -1.73306569e-01 1.70465857e-01 7.43321896e-01
-1.16030619e-01 -1.53000727e-01 8.49465609e-01 -2.53206585e-02
1.92442223e-01 -3.87922764e-01 3.16724777e-01 7.20473468e-01
-1.16912439e-01 -3.69525999e-01 -4.22977209e-01 -2.34466106e-01
-1.88304946e-01 6.47364780e-02 -6.17708862e-01 1.10240674e+00
-3.55043560e-01 7.85036922e-01 1.06413931e-01 -4.75442648e-01
3.47620964e-01 -3.21896642e-01 6.51790738e-01 -5.07910550e-01
1.30853221e-01 2.52405107e-01 -8.18095133e-02 -1.30988371e+00
-3.73065889e-01 -8.63367379e-01 -7.44789004e-01 -5.14042914e-01
-2.94364840e-01 -1.58516079e-01 9.41847980e-01 4.04204279e-01
8.74461710e-01 -2.09869206e-01 6.74605727e-01 -1.43878841e+00
-3.69364023e-01 -6.57722473e-01 -4.26086813e-01 1.75325915e-01
5.38242817e-01 -8.88796628e-01 -6.41861737e-01 -1.19913228e-01] | [7.534502983093262, -2.390181303024292] |
159aeb3a-8312-4a09-ba8b-e110bf6c524e | towards-in-context-scene-understanding | 2306.01667 | null | https://arxiv.org/abs/2306.01667v1 | https://arxiv.org/pdf/2306.01667v1.pdf | Towards In-context Scene Understanding | In-context learning$\unicode{x2013}$the ability to configure a model's behavior with different prompts$\unicode{x2013}$has revolutionized the field of natural language processing, alleviating the need for task-specific models and paving the way for generalist models capable of assisting with any query. Computer vision, in contrast, has largely stayed in the former regime: specialized decoders and finetuning protocols are generally required to perform dense tasks such as semantic segmentation and depth estimation. In this work we explore a simple mechanism for in-context learning of such scene understanding tasks: nearest neighbor retrieval from a prompt of annotated features. We propose a new pretraining protocol$\unicode{x2013}$leveraging attention within and across images$\unicode{x2013}$which yields representations particularly useful in this regime. The resulting Hummingbird model, suitably prompted, performs various scene understanding tasks without modification while approaching the performance of specialists that have been finetuned for each task. Moreover, Hummingbird can be configured to perform new tasks much more efficiently than finetuned models, raising the possibility of scene understanding in the interactive assistant regime. | ['Olivier J. Hénaff', 'Relja Arandjelović', 'Nikhil Parthasarathy', 'David Steiner', 'Ivana Balažević'] | 2023-06-02 | null | null | null | null | ['scene-understanding'] | ['computer-vision'] | [ 4.18644667e-01 3.08324099e-01 2.63982415e-01 -5.98958373e-01
-5.96308053e-01 -6.49894238e-01 5.61343491e-01 2.01184094e-01
-9.47888494e-01 4.03648406e-01 -3.33270550e-01 -4.80268598e-01
-4.04641390e-01 -8.29630852e-01 -7.41071284e-01 -6.58975244e-01
6.96026310e-02 5.98422527e-01 4.20579642e-01 -4.01569664e-01
3.79697829e-01 3.82997245e-01 -1.95313513e+00 2.28877008e-01
8.88069510e-01 7.76972890e-01 1.10171878e+00 8.89915347e-01
-1.60448939e-01 3.43439758e-01 -6.74652100e-01 -2.29706392e-01
1.00912936e-01 -1.54872060e-01 -9.54589188e-01 -8.20159614e-02
6.19184434e-01 -2.68296689e-01 -2.92695969e-01 8.19253981e-01
4.03753281e-01 3.57768655e-01 5.72992206e-01 -7.81494081e-01
-4.58648086e-01 4.25058246e-01 -2.88267910e-01 5.03435314e-01
2.95659155e-01 4.05534029e-01 9.30335581e-01 -4.94568855e-01
4.49581236e-01 1.03061843e+00 3.10853779e-01 6.80644572e-01
-1.01103687e+00 -3.90193462e-01 4.59577173e-01 4.56830591e-01
-1.13227940e+00 -1.59945890e-01 6.35071754e-01 -3.02667141e-01
9.79380608e-01 2.31723294e-01 6.91766560e-01 1.00752938e+00
-2.19489321e-01 1.11799669e+00 9.54992592e-01 -5.39439201e-01
1.88466534e-01 1.21468462e-01 5.45949899e-02 8.62060487e-01
8.99422355e-03 1.87717393e-01 -6.29161775e-01 4.70556200e-01
9.23297405e-01 -8.90624449e-02 -2.93495387e-01 -3.56288940e-01
-1.23638833e+00 6.89728379e-01 6.83990002e-01 3.39159340e-01
-1.88795224e-01 3.99538167e-02 3.34787071e-01 3.59862119e-01
9.19802636e-02 9.06558812e-01 -5.06757319e-01 -2.48167310e-02
-8.20451498e-01 4.14166376e-02 5.15442967e-01 1.08965874e+00
1.03005004e+00 -4.15907316e-02 -4.92275842e-02 8.80891919e-01
-2.07526069e-02 4.50818568e-01 4.41275179e-01 -9.44483399e-01
2.55217791e-01 7.44131207e-01 -8.43880791e-03 -7.04340696e-01
-6.64041877e-01 -4.27441955e-01 -4.18270558e-01 1.13300107e-01
5.83826363e-01 -8.47834125e-02 -1.12526679e+00 1.76972890e+00
4.37612295e-01 -3.73992659e-02 -4.09745164e-02 8.61270428e-01
5.70240498e-01 7.04752505e-01 2.44467273e-01 1.88263535e-01
1.27898300e+00 -8.77561569e-01 1.97832901e-02 -5.49537718e-01
6.90607727e-01 -4.79450941e-01 1.66786838e+00 4.89666611e-01
-6.47035360e-01 -7.98570335e-01 -9.47852552e-01 -1.81719542e-01
-6.58181489e-01 -1.75109461e-01 8.66509497e-01 4.11640584e-01
-1.25544167e+00 5.80825865e-01 -7.08275080e-01 -6.22429669e-01
4.52423722e-01 6.15036845e-01 -3.69962722e-01 -4.00953293e-01
-9.47048545e-01 1.02206337e+00 5.53708971e-01 2.71615803e-01
-1.14868188e+00 -5.66817164e-01 -9.04869616e-01 -4.16612513e-02
4.95506376e-01 -6.45365953e-01 1.20055676e+00 -8.56673837e-01
-1.12513328e+00 1.25794768e+00 1.48642346e-01 -3.55593801e-01
2.12081775e-01 -1.81752205e-01 -1.98019788e-01 2.99198389e-01
2.15782732e-01 1.10304332e+00 8.31262112e-01 -1.22730243e+00
-6.84154749e-01 -6.92971826e-01 5.47743022e-01 3.90289545e-01
-2.18991145e-01 -3.87996733e-01 -5.85534811e-01 -3.70703995e-01
1.41904026e-01 -8.52037728e-01 -1.88725322e-01 2.04794960e-05
-1.82240605e-01 -2.29137734e-01 7.18647718e-01 -2.39706129e-01
8.88855994e-01 -2.02066493e+00 1.21433482e-01 -2.60723960e-02
-1.83721203e-02 4.22891200e-01 -1.77139312e-01 2.70926446e-01
2.76343167e-01 -8.97916630e-02 -4.04726356e-01 -2.01936617e-01
-5.39318062e-02 6.19589210e-01 -7.29148102e-04 1.70378551e-01
-5.40094040e-02 7.18405366e-01 -9.29590046e-01 -5.65082908e-01
3.78641367e-01 3.05520773e-01 -6.81472361e-01 2.97457367e-01
-5.89353800e-01 5.95814228e-01 -5.16340077e-01 5.31474888e-01
3.34875673e-01 -1.55001193e-01 -1.38295593e-03 9.67968255e-02
-1.51559860e-01 2.84139905e-02 -8.53871822e-01 2.19454002e+00
-7.85797834e-01 6.47829235e-01 3.34090143e-01 -1.19123912e+00
9.07959580e-01 -1.13213845e-02 1.37970448e-01 -8.85537028e-01
3.32028329e-01 -2.67679598e-02 -5.14612608e-02 -7.09701240e-01
5.99274993e-01 -1.50586069e-01 -3.05472165e-01 3.61222237e-01
2.39798084e-01 -5.85908175e-01 1.51970237e-01 1.20548360e-01
1.10380125e+00 3.59048694e-01 -2.38114428e-02 -3.65301698e-01
3.46827477e-01 3.26357156e-01 7.41896182e-02 1.11992610e+00
-3.68373543e-01 4.62415844e-01 1.09204806e-01 -5.38261712e-01
-8.25564682e-01 -1.04782522e+00 -2.15215161e-01 1.75242054e+00
3.45307589e-01 -1.14062596e-02 -9.87063229e-01 -6.14873767e-01
-1.07756771e-01 8.33928227e-01 -7.08301306e-01 -3.43359470e-01
-5.46518207e-01 -4.69644427e-01 4.25667256e-01 4.49387848e-01
7.93761730e-01 -1.44500637e+00 -1.16843867e+00 1.32930383e-01
-1.42702326e-01 -9.41326261e-01 -1.84560314e-01 8.25612247e-01
-8.65627110e-01 -9.77484167e-01 -4.93051052e-01 -8.10155809e-01
6.42543495e-01 4.31290269e-01 1.12789154e+00 2.34655783e-01
-5.49707353e-01 8.75506163e-01 -5.59535921e-01 -4.53089654e-01
-1.07745558e-01 4.09906924e-01 -1.54343799e-01 -2.27844387e-01
5.03020287e-01 -6.26814783e-01 -7.18642056e-01 2.24443734e-01
-1.23452210e+00 1.11157909e-01 6.51074290e-01 6.88614786e-01
5.42527437e-01 -1.87389240e-01 4.59236771e-01 -1.10172927e+00
2.01241076e-01 -1.79345131e-01 -5.18059433e-01 1.67083070e-01
-4.63769257e-01 3.00510198e-01 6.52904749e-01 -2.64234871e-01
-1.05578709e+00 1.50107026e-01 -2.36573413e-01 -5.40698431e-02
-5.46169579e-01 2.65113324e-01 -4.35002178e-01 -3.38319726e-02
8.90821457e-01 2.83656359e-01 -3.47189069e-01 -6.13426089e-01
4.78681773e-01 5.66465080e-01 9.36956286e-01 -6.46709442e-01
7.26910293e-01 3.67983639e-01 -3.92525673e-01 -1.06643128e+00
-9.26276147e-01 -5.79247296e-01 -1.00321734e+00 -2.45001003e-01
9.84604061e-01 -7.75720835e-01 -7.81390548e-01 4.55764651e-01
-9.24916744e-01 -8.09389234e-01 -2.61754930e-01 2.27021649e-01
-7.00871170e-01 1.41270980e-02 -4.83818948e-02 -5.54711759e-01
1.11727364e-01 -9.45869505e-01 1.06004119e+00 5.11424541e-01
-1.88549310e-01 -8.91916037e-01 -1.11160271e-01 6.62230730e-01
3.09052378e-01 -1.28200695e-01 1.12753427e+00 -5.92513740e-01
-6.82182610e-01 -2.89049655e-01 -2.57930130e-01 1.31585672e-01
-9.77582037e-02 -4.55966473e-01 -1.34352398e+00 -1.60286397e-01
-1.59891158e-01 -3.77582848e-01 1.06043506e+00 1.86816081e-01
1.38413370e+00 -1.06021471e-01 -3.62190038e-01 7.04349101e-01
1.26418960e+00 3.05247098e-01 4.82332855e-01 5.92510700e-01
6.10825241e-01 6.67542934e-01 5.19322813e-01 2.13395119e-01
6.49030864e-01 5.32912910e-01 7.88485467e-01 -2.85578836e-02
1.13919586e-01 -3.52954000e-01 -3.08481883e-03 8.68018195e-02
-1.95534024e-02 -1.78275555e-01 -1.02104056e+00 6.35388792e-01
-1.49169099e+00 -8.20293128e-01 3.08759332e-01 2.18593478e+00
7.73798227e-01 1.49102837e-01 -1.67585805e-01 5.32460772e-02
5.08432686e-01 3.63657683e-01 -7.38867700e-01 -5.25492489e-01
8.36929753e-02 3.60409290e-01 2.84076542e-01 6.79519236e-01
-1.17552423e+00 1.40903902e+00 5.12581968e+00 5.79607129e-01
-1.19855368e+00 -1.27545968e-01 3.88552874e-01 -3.29747833e-02
-3.61126035e-01 -4.31637354e-02 -8.23320806e-01 2.53199637e-01
7.87262321e-01 2.81625062e-01 4.13426310e-01 9.51615334e-01
9.89387259e-02 -7.59089947e-01 -1.14513671e+00 9.85784233e-01
1.48583874e-01 -1.06254339e+00 6.34970441e-02 -2.12499171e-01
3.62391412e-01 -3.79834906e-03 2.06100821e-01 5.97818315e-01
2.39248112e-01 -9.71729994e-01 4.72228646e-01 3.77577782e-01
6.50073767e-01 -5.46499908e-01 2.61871606e-01 5.85558355e-01
-8.35899889e-01 -3.87401909e-01 -3.62130523e-01 -5.38064465e-02
-5.25452830e-02 2.61929542e-01 -1.07220781e+00 1.13603763e-01
9.09789920e-01 9.22462121e-02 -7.84150720e-01 9.40332294e-01
-2.07557291e-01 3.60838503e-01 -4.08955753e-01 -1.46993250e-01
3.45100909e-01 -2.11284980e-02 3.13904911e-01 1.25665462e+00
1.32458493e-01 3.03687304e-01 4.71141413e-02 4.98272747e-01
8.05646777e-02 -5.90036809e-02 -6.45239592e-01 2.07601398e-01
3.24996501e-01 1.04252923e+00 -8.97792637e-01 -2.46939272e-01
-5.75636700e-02 1.21085835e+00 4.11314815e-01 3.04007262e-01
-5.49558401e-01 -5.09719789e-01 4.78493124e-01 3.10812861e-01
4.73231792e-01 -5.21796286e-01 -2.94450730e-01 -9.80023384e-01
-1.75557777e-01 -7.03851521e-01 4.22336668e-01 -9.41542268e-01
-7.05594242e-01 8.49737585e-01 1.93738684e-01 -7.55133450e-01
-1.77304476e-01 -8.40035319e-01 -3.35522890e-01 5.18863857e-01
-1.41913927e+00 -1.14159584e+00 -5.73648751e-01 8.24738443e-01
6.32292926e-01 4.46508601e-02 7.55302489e-01 8.59234929e-02
-3.87719512e-01 3.93177509e-01 -9.19051170e-02 -3.10082305e-02
5.88861883e-01 -1.40412438e+00 2.00790480e-01 6.78480983e-01
4.20543194e-01 3.65023643e-01 7.17302442e-01 -2.57425547e-01
-1.26605284e+00 -8.58234942e-01 7.07690716e-01 -5.19009829e-01
2.50610799e-01 -5.12563586e-01 -8.59891117e-01 5.20122707e-01
-1.08522378e-01 -2.53100008e-01 5.66571116e-01 1.85095593e-01
-3.49005163e-01 -3.42131317e-01 -1.05351949e+00 6.60505950e-01
1.19170034e+00 -7.98895121e-01 -5.57922006e-01 2.78918713e-01
7.27021456e-01 -3.25300872e-01 -3.34258318e-01 2.84529060e-01
3.53485078e-01 -1.21027052e+00 9.03673708e-01 -4.69896138e-01
1.14719160e-01 -2.06991553e-01 -2.32873544e-01 -1.04984403e+00
-1.06455490e-01 -3.77486527e-01 3.82954448e-01 7.96577692e-01
5.36176622e-01 -2.80319273e-01 8.99079204e-01 7.73030519e-01
-5.24779856e-01 -5.12069881e-01 -8.01521659e-01 -5.37231326e-01
-2.90051866e-02 -7.70512938e-01 3.54720592e-01 5.64814448e-01
-2.04234555e-01 2.95059055e-01 -2.50779395e-03 3.04158688e-01
4.24849421e-01 1.96403742e-01 7.34407306e-01 -1.18232000e+00
-5.12580335e-01 -3.03878427e-01 -3.20855588e-01 -1.56137466e+00
8.68214965e-02 -8.16850483e-01 3.85202527e-01 -1.52670324e+00
-1.19530104e-01 -6.95776165e-01 -8.03195611e-02 6.46270871e-01
-1.37179062e-01 2.40621358e-01 2.30355069e-01 1.72374304e-02
-6.73951864e-01 2.45035008e-01 1.40135825e+00 -1.59414276e-01
-2.42188424e-01 1.64389864e-01 -8.70166183e-01 6.90291464e-01
5.76668620e-01 -2.11814091e-01 -5.88190198e-01 -8.32387924e-01
1.42667651e-01 -3.29645686e-02 7.20905900e-01 -1.28561699e+00
5.07319033e-01 -1.86157934e-02 3.66812736e-01 -4.09403682e-01
4.96592671e-01 -8.12253952e-01 -3.53535742e-01 1.11900471e-01
-3.23661059e-01 -1.10572666e-01 3.41578245e-01 6.01788938e-01
-9.52390432e-02 -5.51548898e-01 7.97920167e-01 -5.26437640e-01
-1.32565928e+00 2.57087886e-01 -5.65121472e-01 -7.74421031e-03
9.17612672e-01 -6.58765376e-01 -1.01406112e-01 -2.59410262e-01
-1.00920415e+00 3.03899348e-01 4.47507590e-01 2.64778107e-01
4.77032155e-01 -4.95750546e-01 -2.65998334e-01 2.15717539e-01
2.80126929e-01 5.58928251e-01 5.67479670e-01 5.42963445e-01
-4.55226004e-01 5.03669083e-01 -2.40225777e-01 -6.55699015e-01
-8.55196357e-01 5.75789988e-01 3.75037938e-01 8.37891623e-02
-3.85795861e-01 1.28358507e+00 5.42522550e-01 -5.40830016e-01
3.84258896e-01 -2.52760738e-01 -1.97546795e-01 1.80415720e-01
3.67831796e-01 1.96698513e-02 1.35846600e-01 -3.38897943e-01
-1.54674485e-01 4.85703111e-01 -1.77692682e-01 -4.62121107e-02
1.32350349e+00 -3.79495203e-01 1.77452356e-01 3.63497108e-01
9.08867717e-01 -5.78278303e-01 -1.65561306e+00 -7.95341134e-02
9.14663672e-02 -3.90263677e-01 -7.18563842e-03 -1.19472456e+00
-7.40589619e-01 1.15074492e+00 6.16714478e-01 -1.93927642e-02
1.25540018e+00 2.21443906e-01 5.59071720e-01 8.54476094e-01
6.82118237e-01 -1.07860804e+00 2.18281448e-01 7.06933796e-01
6.18983746e-01 -1.33774734e+00 -3.80768925e-01 1.14548109e-01
-6.34244740e-01 9.36632156e-01 8.20623934e-01 -2.38224361e-02
4.14362758e-01 9.52295661e-02 1.93517968e-01 -3.04243237e-01
-4.33152556e-01 -5.21805823e-01 -3.48286470e-03 9.59014654e-01
1.65081009e-01 -7.78859574e-03 2.69310355e-01 4.51255947e-01
-3.54588032e-01 -2.86583334e-01 3.51325542e-01 8.50340724e-01
-9.55827057e-01 -9.42379653e-01 -2.20519960e-01 3.31939697e-01
1.58191949e-01 -2.86997080e-01 -3.71269733e-01 1.12118673e+00
4.14172858e-01 8.54208410e-01 1.61507681e-01 -2.64540613e-01
3.99905056e-01 2.68140763e-01 6.02544785e-01 -8.82858455e-01
-6.25770032e-01 -1.87033057e-01 3.49633209e-02 -3.85040015e-01
-2.15517491e-01 -5.53362250e-01 -1.31411886e+00 -4.29933406e-02
-1.49511397e-01 2.04437703e-01 6.43753052e-01 1.06588101e+00
1.29064217e-01 3.33901137e-01 3.52846771e-01 -1.12461662e+00
-2.34305277e-01 -8.17264020e-01 -5.56321263e-01 6.30046502e-02
2.23039106e-01 -6.18674755e-01 -1.14650995e-01 2.17404276e-01] | [9.91576099395752, 1.4951705932617188] |
77da8f60-bd56-43ac-a5db-adc3040e220d | supervised-learning-of-the-next-best-view-for | 1905.05833 | null | https://arxiv.org/abs/1905.05833v1 | https://arxiv.org/pdf/1905.05833v1.pdf | Supervised Learning of the Next-Best-View for 3D Object Reconstruction | Motivated by the advances in 3D sensing technology and the spreading of low-cost robotic platforms, 3D object reconstruction has become a common task in many areas. Nevertheless, the selection of the optimal sensor pose that maximizes the reconstructed surface is a problem that remains open. It is known in the literature as the next-best-view planning problem. In this paper, we propose a novel next-best-view planning scheme based on supervised deep learning. The scheme contains an algorithm for automatic generation of datasets and an original three-dimensional convolutional neural network (3D-CNN) used to learn the next-best-view. Unlike previous work where the problem is addressed as a search, the trained 3D-CNN directly predicts the sensor pose. We present a comparison of the proposed network against a similar net, and we present several experiments of the reconstruction of unknown objects validating the effectiveness of the proposed scheme. | ['Hind Taud', 'J. Irving Vasquez-Gomez', 'Carolina Reta', 'Miguel Mendoza', 'Luis Enrique Sucar'] | 2019-05-14 | null | null | null | null | ['3d-object-reconstruction'] | ['computer-vision'] | [ 2.83552080e-01 2.83165306e-01 2.22510491e-02 -4.42507803e-01
-5.96895099e-01 -4.67587084e-01 4.21111852e-01 -3.27802777e-01
-2.26041913e-01 2.74871439e-01 1.27795249e-01 -2.33941749e-02
-3.35706472e-01 -8.83961976e-01 -1.03726709e+00 -4.18428510e-01
1.11427121e-01 7.50731111e-01 3.34694207e-01 -1.50490448e-01
6.42172098e-01 1.00827730e+00 -1.59766877e+00 4.37027887e-02
2.66807795e-01 1.51417804e+00 6.24289036e-01 3.73873532e-01
1.91608533e-01 2.11394966e-01 -7.64661003e-03 3.28589566e-02
7.85886586e-01 -7.94089958e-02 -7.78805435e-01 3.13964665e-01
2.64215261e-01 -5.02593935e-01 -5.43898493e-02 8.01802337e-01
6.46684766e-01 -1.85599953e-01 5.44140995e-01 -9.88423347e-01
-8.95700082e-02 1.07117034e-01 -2.94391841e-01 -3.55794102e-01
4.18103069e-01 -1.98937833e-01 8.22351098e-01 -1.14027309e+00
8.48095834e-01 1.24202478e+00 6.40799284e-01 5.48943222e-01
-8.69535029e-01 -1.76267073e-01 -1.81883767e-01 3.82722802e-02
-1.24236131e+00 -2.19541252e-01 1.25874865e+00 -4.82536137e-01
8.53831053e-01 -2.98338205e-01 8.11432064e-01 9.08342063e-01
4.07132685e-01 6.89185202e-01 9.81808543e-01 -4.92955744e-01
4.08763617e-01 -2.29090184e-01 -5.07205427e-01 6.73004925e-01
2.11581513e-01 3.43163610e-01 -5.00053883e-01 1.00051627e-01
1.12024581e+00 1.42304510e-01 -2.84889098e-02 -1.23243701e+00
-1.19693220e+00 7.75545239e-01 5.49532473e-01 6.82579055e-02
-6.58797026e-01 2.20995665e-01 -2.65036616e-02 -1.37889981e-01
4.23088700e-01 5.44629753e-01 -6.33727968e-01 5.14529124e-02
-6.41193926e-01 4.53562796e-01 8.43201280e-01 8.69609058e-01
7.32610464e-01 -1.03649544e-02 3.68743151e-01 5.81374466e-01
6.16933048e-01 3.23769152e-01 -6.10943772e-02 -1.38914585e+00
3.59398723e-01 9.05955434e-01 5.48852742e-01 -9.41958845e-01
-6.01157546e-01 -4.21219170e-01 -4.98333424e-01 5.50102174e-01
2.99663991e-01 -2.48611659e-01 -7.78175056e-01 1.17565095e+00
5.05602419e-01 -2.81242967e-01 1.20557688e-01 1.17184222e+00
5.12488008e-01 2.92234480e-01 -7.42363930e-01 8.69076103e-02
8.71069133e-01 -7.56283045e-01 -3.55441988e-01 -5.00864804e-01
-4.80746143e-02 -6.54723346e-01 4.61324304e-01 5.96270800e-01
-1.03333974e+00 -5.57545125e-01 -1.28290868e+00 5.90138249e-02
-1.70324773e-01 3.30858566e-02 5.33617556e-01 1.00541674e-01
-9.47157383e-01 7.29148507e-01 -7.93608546e-01 -6.78745091e-01
4.83527005e-01 4.57439810e-01 -4.95430827e-01 -2.01663911e-01
-7.65915930e-01 1.16772878e+00 6.87701762e-01 1.98981076e-01
-1.10402095e+00 -2.59333462e-01 -8.30332696e-01 -1.87608317e-01
5.63226938e-01 -8.25954258e-01 1.40560400e+00 -5.39463341e-01
-1.88445914e+00 1.00525451e+00 1.14581250e-01 -3.53617430e-01
5.34301817e-01 -3.95403415e-01 2.00558558e-01 1.86805353e-01
6.66415552e-03 7.79064238e-01 6.92881167e-01 -1.67829037e+00
-5.02005219e-01 -7.94323444e-01 4.38159049e-01 4.22262192e-01
1.58771008e-01 -4.93057251e-01 -5.37975729e-01 -1.98504999e-01
9.69059110e-01 -1.05181396e+00 -5.47375023e-01 4.83124405e-01
-4.19573337e-01 -6.03023730e-02 7.32532620e-01 -3.51475835e-01
4.12360311e-01 -1.76182985e+00 2.76896030e-01 1.71968190e-04
-1.70404077e-01 -1.30129248e-01 1.75817192e-01 6.04244113e-01
2.27666184e-01 -2.27115676e-01 -2.00785384e-01 -3.34533811e-01
-1.52189329e-01 2.90220916e-01 5.82071207e-02 5.42488158e-01
2.03300379e-02 7.73336470e-01 -7.42189825e-01 -2.87173629e-01
4.99809802e-01 3.36993515e-01 -5.64626217e-01 4.99796897e-01
-6.09387994e-01 6.04611695e-01 -7.45101869e-01 7.51093626e-01
7.07873583e-01 -3.44320506e-01 1.18248761e-01 -3.40197802e-01
-2.78569967e-01 -5.28485281e-03 -1.25500655e+00 2.32198024e+00
-4.94033396e-01 1.36461481e-01 2.28738770e-01 -9.86975729e-01
1.40861225e+00 1.33170158e-01 8.32846999e-01 -3.96591783e-01
4.00203198e-01 3.59024495e-01 -3.04199219e-01 -7.27749169e-01
4.54796672e-01 -3.98251563e-01 -1.58423603e-01 2.81437367e-01
-1.75514277e-02 -5.41948497e-01 -5.08738995e-01 -3.87044013e-01
8.66297960e-01 8.68139803e-01 6.24379814e-01 -1.33338436e-01
4.31678861e-01 4.23524618e-01 3.64831984e-01 3.22245061e-01
6.96088672e-02 9.15757239e-01 4.50256206e-02 -7.31288075e-01
-1.10442042e+00 -7.65172303e-01 -1.88404545e-02 2.58066267e-01
6.58410490e-01 2.64574975e-01 -5.30163348e-01 -5.27296484e-01
2.25154459e-01 4.60003674e-01 -4.88670528e-01 2.81094700e-01
-6.33017302e-01 -7.36632496e-02 -8.92111138e-02 4.68630195e-01
5.92402637e-01 -1.07632267e+00 -1.39572537e+00 2.05403045e-01
3.76025401e-02 -1.28540301e+00 -1.10013504e-03 1.93501770e-01
-1.21143782e+00 -1.23923218e+00 -5.48706353e-01 -6.81964040e-01
6.37242317e-01 3.32058191e-01 1.07973671e+00 -3.29052746e-01
-2.25134362e-02 4.94720876e-01 -4.17236626e-01 -7.99215019e-01
-3.42098325e-01 1.35384381e-01 -7.05715716e-02 -1.39142394e-01
9.99802798e-02 -7.24650383e-01 -7.57949829e-01 2.86599547e-01
-6.35315001e-01 3.06383222e-01 7.31054366e-01 5.51593363e-01
9.12487745e-01 -1.61464900e-01 4.73545760e-01 -5.90500355e-01
1.65866911e-01 -3.26128215e-01 -9.76595342e-01 3.38882133e-02
-5.41390061e-01 1.57365546e-01 3.77745658e-01 4.49236557e-02
-9.68343019e-01 9.57525611e-01 -2.32797876e-01 -7.24145591e-01
-4.71278369e-01 4.37293053e-01 -3.64322692e-01 -1.90031007e-01
4.45623517e-01 1.73011154e-01 2.33221203e-01 -7.09424138e-01
3.25785488e-01 5.57485104e-01 1.62169844e-01 -1.36897475e-01
7.01510251e-01 8.32528591e-01 2.87147105e-01 -5.98321855e-01
-1.21572959e+00 -3.92558366e-01 -1.02990377e+00 -6.17366433e-01
1.06280088e+00 -7.74828494e-01 -8.12433481e-01 3.91338319e-01
-1.54103911e+00 1.06966551e-02 -2.64794558e-01 4.82009381e-01
-1.11709344e+00 1.44943595e-02 1.35955602e-01 -8.87820661e-01
-3.76997143e-01 -1.27327156e+00 1.41860676e+00 6.92299288e-03
2.65744496e-02 -7.25230217e-01 -3.92502965e-03 5.14301479e-01
2.13707194e-01 6.58836067e-01 6.34653986e-01 -4.36493337e-01
-1.11160374e+00 -1.53156400e-01 7.80888274e-02 8.87710080e-02
-9.35333129e-03 -6.30636096e-01 -1.01249552e+00 -1.18138634e-01
1.96258217e-01 -2.90271282e-01 4.56900895e-01 5.29179096e-01
9.95900035e-01 4.10998203e-02 -5.52820742e-01 6.00772023e-01
1.87683761e+00 3.47180426e-01 4.38709289e-01 4.93504673e-01
4.29662913e-01 5.81021249e-01 7.45342791e-01 6.78262472e-01
4.20322329e-01 8.79965007e-01 1.43173385e+00 2.07022071e-01
1.32517427e-01 -5.11026919e-01 -5.87947406e-02 5.93183756e-01
-4.50814776e-02 -1.95639193e-01 -9.70220208e-01 5.91248155e-01
-1.79166150e+00 -7.05050111e-01 2.99635381e-01 2.04312515e+00
1.85462028e-01 2.13615537e-01 -1.29873604e-01 1.31174117e-01
4.52658564e-01 1.16495363e-01 -9.33294237e-01 -1.26289666e-01
2.66051024e-01 -6.40203878e-02 6.06964171e-01 4.70439970e-01
-1.01865709e+00 6.82324648e-01 6.23172426e+00 1.04303047e-01
-1.22078919e+00 -1.80325657e-01 1.26980677e-01 1.93141505e-01
-1.87355548e-01 2.02042609e-03 -8.91269028e-01 -7.02870637e-02
1.60437495e-01 3.61080408e-01 3.45385313e-01 1.02196002e+00
2.61436790e-01 -3.17683220e-01 -1.34221625e+00 1.05361509e+00
3.77903223e-01 -1.46773612e+00 -6.64486885e-02 1.16015896e-01
7.44426608e-01 1.36014506e-01 -4.56359774e-01 -2.23528966e-01
2.21420661e-01 -6.84617639e-01 1.00145352e+00 7.10089684e-01
4.65662152e-01 -6.13240898e-01 6.94371045e-01 9.98813391e-01
-9.53863680e-01 -1.82862058e-01 -2.72065848e-01 -1.56610161e-01
3.92058074e-01 6.21845245e-01 -1.19959116e+00 8.45804930e-01
7.95604289e-01 7.61476099e-01 -1.98509172e-01 1.08670008e+00
-1.00392058e-01 -6.32589981e-02 -3.06760132e-01 -1.60063222e-01
6.51411861e-02 -1.78885221e-01 7.39162982e-01 4.59091276e-01
6.48424149e-01 8.44711885e-02 3.20704401e-01 9.43632364e-01
-2.03953460e-02 -2.21103981e-01 -1.11436939e+00 2.58477271e-01
2.76074946e-01 1.07439351e+00 -7.59299040e-01 1.81708723e-01
-2.97052950e-01 7.90019691e-01 3.79708201e-01 -1.18158028e-01
-3.08791667e-01 2.98971366e-02 1.89965844e-01 3.88021916e-01
6.40330970e-01 -4.08532739e-01 -5.18583179e-01 -7.69684494e-01
1.43143713e-01 -3.70173723e-01 -1.97184756e-02 -1.10341609e+00
-1.22674131e+00 4.76023436e-01 1.68485537e-01 -1.76737785e+00
-3.73541564e-01 -7.17573166e-01 -6.27066791e-02 6.34081185e-01
-1.71860576e+00 -1.12503970e+00 -4.31562454e-01 2.22132429e-01
9.03430521e-01 -7.18695074e-02 8.16576660e-01 -1.04814768e-01
1.75753206e-01 -2.19325975e-01 -1.29355669e-01 -1.37860849e-01
2.00614601e-01 -9.63546157e-01 1.26628548e-01 4.97393489e-01
-2.40040258e-01 5.36535829e-02 5.92590451e-01 -6.58553302e-01
-1.88203502e+00 -9.63458121e-01 6.76662743e-01 -2.66447037e-01
1.67740896e-01 -3.13937962e-01 -3.72540116e-01 6.01827383e-01
1.77522510e-01 1.95598453e-01 1.78174451e-02 -3.40032786e-01
2.40981847e-01 -1.31466940e-01 -1.32457793e+00 2.23219663e-01
1.30168939e+00 -1.40143082e-01 -5.90471089e-01 1.52907401e-01
5.98864436e-01 -8.32956433e-01 -8.85505855e-01 6.87491059e-01
8.23391795e-01 -1.15947104e+00 9.89235938e-01 -1.50581270e-01
6.49371147e-01 -2.43324578e-01 -4.96218652e-01 -1.35182631e+00
-7.19528943e-02 -3.31866801e-01 -7.28480369e-02 6.11495912e-01
2.04307154e-01 -3.57459724e-01 1.22751176e+00 2.36312658e-01
-4.47904170e-01 -1.11406362e+00 -9.97820497e-01 -5.15711546e-01
-2.58646071e-01 -3.64260972e-01 4.93584394e-01 4.04504865e-01
-4.53601032e-01 3.25965017e-01 -2.67734766e-01 5.35769165e-01
7.40481496e-01 7.26833165e-01 8.48121047e-01 -1.65483081e+00
-1.12709522e-01 1.21379038e-02 -4.36301261e-01 -1.47286248e+00
5.20694405e-02 -7.98015416e-01 5.67169845e-01 -2.05330205e+00
-1.53943330e-01 -4.10787225e-01 1.80437237e-01 9.73081738e-02
5.27307928e-01 -2.02840298e-01 1.11752741e-01 6.53553531e-02
-4.40412998e-01 5.87098002e-01 1.70024467e+00 1.24211214e-01
-1.15471572e-01 2.41461813e-01 -4.31593508e-01 9.24549818e-01
8.06596518e-01 -3.87390554e-01 -3.57560039e-01 -5.90794563e-01
3.84897202e-01 6.46789312e-01 3.32100540e-01 -1.19928575e+00
3.24685574e-01 -1.69468567e-01 4.15269434e-01 -1.27284098e+00
8.06545377e-01 -1.46455038e+00 2.10391268e-01 4.99862671e-01
-9.26475003e-02 5.59911057e-02 -9.85524133e-02 6.47819996e-01
-9.52683315e-02 -4.14275199e-01 7.77724028e-01 -7.13762641e-01
-8.44815612e-01 4.04988259e-01 7.05434158e-02 -3.56578648e-01
1.21372259e+00 -6.49540484e-01 2.86786288e-01 -3.09999228e-01
-6.64367855e-01 1.34186685e-01 5.94144166e-01 4.91843015e-01
1.07022905e+00 -1.26354182e+00 -4.30430502e-01 2.43611440e-01
1.68157429e-01 6.94754124e-01 -5.58255687e-02 2.06044167e-01
-5.60870051e-01 3.61173362e-01 -3.13194573e-01 -9.17868555e-01
-7.62283981e-01 4.76579666e-01 5.89082420e-01 -6.29121363e-02
-4.94798839e-01 5.01102030e-01 -1.75446719e-01 -8.49959075e-01
2.14714155e-01 -1.43766150e-01 -3.48888487e-01 -3.35070908e-01
-2.60444544e-02 3.79026860e-01 1.96743533e-01 -6.29789591e-01
-1.62671000e-01 8.78932834e-01 4.97490227e-01 -6.18715025e-02
1.91071689e+00 -8.59075636e-02 4.75794785e-02 4.42480117e-01
1.01028764e+00 -4.61327732e-01 -1.53521192e+00 -1.28375188e-01
1.01690531e-01 -3.37092757e-01 6.62938431e-02 -6.60575390e-01
-9.93180394e-01 8.85166526e-01 7.32721388e-01 5.64733520e-02
8.87502968e-01 1.72215983e-01 6.93299413e-01 4.79723126e-01
1.01469111e+00 -9.01004970e-01 2.31392160e-01 6.94819450e-01
1.37594187e+00 -1.46877766e+00 2.20976830e-01 -5.58287978e-01
-1.92439884e-01 1.23075628e+00 8.27649534e-01 -3.78147066e-01
1.00509632e+00 2.47413218e-01 8.53529051e-02 -6.57531738e-01
-3.01567018e-01 -8.13498646e-02 1.93640828e-01 5.41186333e-01
-8.30486715e-02 -1.34811401e-01 7.14714453e-02 3.29234689e-01
-1.68552279e-01 2.96475172e-01 2.92162210e-01 1.27509248e+00
-5.72831392e-01 -1.01462424e+00 -4.59203988e-01 3.76754254e-01
-1.33161888e-01 6.84037745e-01 -3.35988134e-01 7.94570327e-01
3.46332669e-01 7.72192597e-01 -2.00784430e-02 -4.95736182e-01
8.52182209e-01 8.63877032e-03 7.11890399e-01 -6.67502880e-01
-3.10840547e-01 -2.11009625e-02 -1.41105175e-01 -7.43796289e-01
-9.23226774e-01 -6.56290352e-01 -1.17845356e+00 3.49530607e-01
-3.55332345e-01 -4.47620660e-01 1.01823628e+00 1.06647766e+00
5.40820360e-01 1.07347988e-01 8.50622475e-01 -1.56312764e+00
-7.09640563e-01 -8.20821822e-01 -4.50923473e-01 2.79511124e-01
3.39441359e-01 -1.00018644e+00 9.49077532e-02 -8.77900273e-02] | [8.153594970703125, -2.8240268230438232] |
abc85d61-35a2-450a-aa2b-9e324ebc98d4 | temporal-dynamic-quantization-for-diffusion | 2306.02316 | null | https://arxiv.org/abs/2306.02316v1 | https://arxiv.org/pdf/2306.02316v1.pdf | Temporal Dynamic Quantization for Diffusion Models | The diffusion model has gained popularity in vision applications due to its remarkable generative performance and versatility. However, high storage and computation demands, resulting from the model size and iterative generation, hinder its use on mobile devices. Existing quantization techniques struggle to maintain performance even in 8-bit precision due to the diffusion model's unique property of temporal variation in activation. We introduce a novel quantization method that dynamically adjusts the quantization interval based on time step information, significantly improving output quality. Unlike conventional dynamic quantization techniques, our approach has no computational overhead during inference and is compatible with both post-training quantization (PTQ) and quantization-aware training (QAT). Our extensive experiments demonstrate substantial improvements in output quality with the quantized diffusion model across various datasets. | ['Eunhyeok Park', 'HyungJun Kim', 'Daehyun Ahn', 'Jungwon Lee', 'Junhyuk So'] | 2023-06-04 | null | null | null | null | ['quantization'] | ['methodology'] | [ 3.08526456e-01 -3.42515439e-01 -3.62532705e-01 -3.07378203e-01
-8.70038390e-01 -5.38832903e-01 7.02785313e-01 8.79333466e-02
-4.71419573e-01 6.51561141e-01 -2.72823013e-02 -3.19792628e-01
-4.45216745e-02 -7.98502564e-01 -3.45967382e-01 -8.18815589e-01
-1.20128281e-01 3.36858809e-01 4.38348681e-01 8.47952142e-02
2.70470381e-01 2.34982938e-01 -1.38347602e+00 -1.61693133e-02
1.03998530e+00 1.11826456e+00 3.09679508e-01 9.52574492e-01
3.92521322e-02 4.92419600e-01 -7.86657989e-01 -4.21848297e-01
2.22004816e-01 -5.37569582e-01 -3.06763589e-01 2.06357446e-02
3.86632383e-01 -5.47482967e-01 -5.44699848e-01 1.11569023e+00
7.49196887e-01 1.64172426e-01 5.40135920e-01 -1.21844137e+00
-9.52345014e-01 5.18018425e-01 -5.38969159e-01 6.53332233e-01
-3.07624582e-02 2.37314582e-01 8.85608673e-01 -6.81942165e-01
5.19485772e-01 1.24738407e+00 6.90630436e-01 5.36899388e-01
-1.33334196e+00 -5.59539735e-01 -3.06049385e-03 3.08727860e-01
-1.66676486e+00 -4.34159726e-01 4.84705865e-01 -1.94654852e-01
1.20718217e+00 5.74952550e-02 8.92939210e-01 8.32157731e-01
3.61375362e-01 7.72122562e-01 8.34443569e-01 -2.73050815e-01
6.38262093e-01 -3.38704288e-01 -3.72946531e-01 7.12586462e-01
-1.89654101e-02 8.02986845e-02 -6.98142469e-01 -1.00888819e-01
1.27448153e+00 -2.49264762e-01 -3.08407605e-01 -8.38884562e-02
-1.16357422e+00 7.89771914e-01 4.34169322e-01 7.27377506e-03
-4.70236182e-01 7.20126092e-01 3.14360768e-01 1.46261305e-01
2.48898432e-01 1.87550738e-01 -1.08089849e-01 -7.72659123e-01
-1.31993783e+00 2.36536980e-01 4.47210610e-01 7.41856217e-01
5.03620505e-01 4.26422983e-01 -3.76932651e-01 9.41519260e-01
3.27625573e-01 5.71330726e-01 8.53273511e-01 -1.29425204e+00
2.81092227e-01 2.04701006e-01 -8.94423276e-02 -9.51396763e-01
-1.42159343e-01 -3.85369688e-01 -1.05675924e+00 1.23715386e-01
2.02109456e-01 -8.62387717e-02 -1.20945442e+00 1.73997915e+00
3.12252730e-01 2.67788231e-01 -7.60293901e-02 7.44842112e-01
3.93049151e-01 7.81552672e-01 1.49603367e-01 -3.37018251e-01
1.00328696e+00 -7.65833318e-01 -9.28443611e-01 4.38160636e-02
3.59255433e-01 -5.03162742e-01 1.00168741e+00 4.31495965e-01
-1.12976158e+00 -4.74416435e-01 -1.19447684e+00 -1.27878964e-01
-1.06010474e-01 -5.42017147e-02 8.96288514e-01 8.90960097e-01
-1.31990552e+00 6.23344898e-01 -1.25167239e+00 -2.64029149e-02
6.33153200e-01 6.37446165e-01 3.23501408e-01 1.69017270e-01
-1.11559761e+00 5.55931807e-01 2.83994317e-01 1.80826150e-02
-5.61383605e-01 -4.97719526e-01 -6.58544719e-01 1.27888322e-01
3.74343209e-02 -7.05639362e-01 1.48082888e+00 -5.87634683e-01
-1.91081882e+00 2.47124314e-01 -3.77238721e-01 -8.76137376e-01
6.41936362e-01 -2.23455532e-03 -2.08011150e-01 3.96547198e-01
-7.78857246e-02 1.07227886e+00 1.13338184e+00 -6.85081005e-01
-5.01403034e-01 -3.09266657e-01 -1.66186884e-01 3.07086289e-01
-5.24565220e-01 -5.30755997e-01 -9.44486439e-01 -9.66226876e-01
2.03277126e-01 -8.37575376e-01 -2.72904158e-01 3.46837550e-01
-1.25685990e-01 -2.13632956e-01 1.06630254e+00 -3.43575746e-01
1.62055767e+00 -1.99027729e+00 -7.53679499e-02 1.12398885e-01
1.99440330e-01 4.53896463e-01 1.59063980e-01 1.35097787e-01
6.14592075e-01 1.33103803e-01 -2.66360134e-01 -4.67458338e-01
-9.64214355e-02 5.42356074e-01 -3.43541116e-01 2.26804480e-01
8.05100873e-02 1.23276842e+00 -1.13733053e+00 -7.01268256e-01
3.13798428e-01 7.67824292e-01 -6.60695016e-01 -2.63495803e-01
-1.88668400e-01 1.86006218e-01 -2.72426456e-01 7.16989040e-01
4.44913477e-01 -6.19480550e-01 2.74497569e-01 -1.36304498e-01
1.88115835e-01 1.76928714e-01 -9.86766636e-01 1.71058750e+00
-3.41169268e-01 9.17970777e-01 -1.35419890e-01 -6.14051938e-01
6.65834308e-01 3.31357956e-01 3.73504519e-01 -9.20353532e-01
-2.68796273e-03 3.48571911e-02 -6.19521774e-02 -3.31233372e-03
7.39850819e-01 5.74843809e-02 2.75289625e-01 4.35282320e-01
-5.19008785e-02 -4.11371469e-01 3.24998260e-01 2.37828135e-01
8.75709832e-01 -1.15025446e-01 1.81743845e-01 5.56876659e-02
-7.21087009e-02 -3.00129980e-01 5.32164216e-01 8.04954469e-01
-3.49592358e-01 6.67588413e-01 2.14266852e-01 -1.33178875e-01
-1.09019125e+00 -1.18666196e+00 -2.67247230e-01 8.37544203e-01
1.60505816e-01 -7.97310770e-01 -7.57258356e-01 -3.35920244e-01
-5.30601628e-02 4.25284833e-01 -5.04068077e-01 -1.68109000e-01
-4.39675689e-01 -8.60077620e-01 7.25557029e-01 7.31834888e-01
8.13688695e-01 -6.71591818e-01 -9.23284233e-01 4.53103632e-01
-5.65194823e-02 -9.25186455e-01 -6.07797325e-01 3.96739133e-02
-1.29402077e+00 -5.26909411e-01 -7.93672562e-01 -4.00347173e-01
4.69609410e-01 1.24733597e-01 8.16234708e-01 5.21590374e-02
-2.62718856e-01 2.92001128e-01 1.40365241e-02 -2.85349846e-01
-2.83258677e-01 -1.90146789e-02 2.87207798e-03 -3.50396961e-01
7.20660463e-02 -6.09898388e-01 -8.95914197e-01 9.37234163e-02
-8.87026310e-01 -3.09477048e-03 4.58482683e-01 1.02318096e+00
9.24867928e-01 3.99035215e-01 5.13797879e-01 -5.29070079e-01
9.25939143e-01 -1.40068993e-01 -6.37035549e-01 1.29574165e-01
-1.02317262e+00 1.79035947e-01 3.52010518e-01 -6.98816895e-01
-9.32583451e-01 -4.77974713e-02 -1.65877342e-01 -5.92270195e-01
5.22755802e-01 3.42854291e-01 2.58534878e-01 -1.18597671e-01
6.44787610e-01 3.25422198e-01 -8.24093297e-02 -9.91717502e-02
4.81080204e-01 5.44172168e-01 6.75107956e-01 -2.68291801e-01
4.96545196e-01 4.01266813e-01 -1.25245571e-01 -8.75407875e-01
-2.99333423e-01 6.01801742e-03 -4.98664260e-01 -1.08610176e-01
6.19364798e-01 -8.32468450e-01 -6.55532956e-01 5.95022500e-01
-9.58923936e-01 -5.17793834e-01 -3.19359958e-01 3.65515143e-01
-5.32278478e-01 4.25222009e-01 -8.11063290e-01 -9.23736393e-01
-5.70645034e-01 -1.14424312e+00 1.13151872e+00 3.85774225e-01
-2.96198785e-01 -1.16763282e+00 -1.40805095e-01 -1.04005992e-01
5.46248019e-01 8.10207054e-02 7.47589111e-01 3.08488369e-01
-7.01259613e-01 -8.87978300e-02 -1.04948297e-01 1.50062308e-01
2.31535152e-01 1.18436262e-01 -7.80426323e-01 -2.65669554e-01
-2.64912993e-01 -3.26988965e-01 7.94970751e-01 7.71770895e-01
1.36437511e+00 -3.10124159e-01 -3.65297794e-01 7.17348397e-01
1.35668099e+00 3.29025447e-01 6.33446455e-01 1.16219833e-01
5.40331185e-01 -3.61499220e-01 3.94655943e-01 5.60599923e-01
3.40814471e-01 7.92231798e-01 2.51277626e-01 1.04392856e-01
-4.34029788e-01 -5.28085947e-01 1.38512924e-01 8.25434148e-01
-7.03190863e-02 -4.53165412e-01 -8.60944569e-01 4.93937284e-01
-1.74815130e+00 -9.31862056e-01 2.88923979e-01 2.21570230e+00
1.11467326e+00 3.48463923e-01 -1.12852175e-02 3.14980954e-01
4.16401833e-01 1.03049129e-01 -8.96818876e-01 -5.32133639e-01
-6.88645989e-02 2.00640917e-01 6.38100803e-01 4.99108762e-01
-8.69917035e-01 8.51463795e-01 8.34782124e+00 9.86618817e-01
-1.31960130e+00 1.27401873e-01 7.30675817e-01 -3.74564558e-01
-3.07301491e-01 -4.93150771e-01 -8.66503060e-01 6.96450770e-01
1.05441499e+00 -2.89023310e-01 4.81421918e-01 6.80923641e-01
2.12391019e-01 -1.94331616e-01 -7.88181007e-01 1.42921138e+00
-1.90546289e-01 -1.60270524e+00 3.40941250e-01 1.93287447e-01
9.44055080e-01 1.16604663e-01 5.72618365e-01 5.93358576e-02
3.56987298e-01 -9.82125521e-01 7.38926291e-01 1.70960307e-01
1.04325676e+00 -7.28779495e-01 3.53202313e-01 3.37829232e-01
-1.18074715e+00 -1.21552855e-01 -4.35012102e-01 -7.40428194e-02
2.79725403e-01 6.15912974e-01 -1.04608810e+00 -3.32213342e-02
6.51461780e-01 6.04955494e-01 -5.50944567e-01 9.11957681e-01
-2.64726698e-01 8.27879429e-01 -5.99489212e-01 -1.38376430e-01
2.70019233e-01 -1.51749432e-01 3.29108477e-01 1.14499378e+00
4.46896642e-01 1.65454835e-01 -5.48267066e-02 6.70739055e-01
-9.59033147e-02 -2.83328474e-01 -3.48365188e-01 -2.19379976e-01
1.00241888e+00 7.93208122e-01 -1.02668667e+00 -6.60458922e-01
-1.07429244e-01 1.22295475e+00 2.49228835e-01 3.93189788e-01
-9.07023191e-01 -2.96116829e-01 6.52351201e-01 -1.15936503e-01
7.81831861e-01 -7.32159436e-01 -4.81084257e-01 -9.48385537e-01
-1.13758044e-02 -5.86944282e-01 1.12299792e-01 -5.77834725e-01
-8.76601696e-01 4.73547786e-01 -9.97099504e-02 -1.07085335e+00
-7.26857424e-01 -1.58468619e-01 -2.00882807e-01 7.48130798e-01
-1.22465670e+00 -7.88642108e-01 -1.44553751e-01 4.59918618e-01
6.70166492e-01 1.94403350e-01 8.51880550e-01 2.51783311e-01
-3.27579230e-01 9.05092061e-01 3.63607407e-01 -2.41899639e-01
2.87639856e-01 -1.38128924e+00 7.21853256e-01 7.06350148e-01
3.34726244e-01 6.85986996e-01 5.55761755e-01 -5.31270504e-01
-1.55068088e+00 -9.84175622e-01 7.01904535e-01 -3.01144838e-01
4.57890868e-01 -1.06252506e-01 -9.23605263e-01 3.96306753e-01
5.35837598e-02 -3.90836690e-03 5.83055496e-01 -1.86790511e-01
-2.72991180e-01 -5.32733649e-02 -1.14091849e+00 7.18056917e-01
9.55336750e-01 -7.33966410e-01 -1.99696898e-01 1.13195710e-01
5.12887239e-01 -6.43159866e-01 -1.03617132e+00 1.80191711e-01
6.96132660e-01 -7.09994495e-01 9.94332135e-01 1.76839605e-01
1.74972534e-01 -3.21621329e-01 -8.43708441e-02 -1.17074108e+00
-3.95827144e-01 -8.65256667e-01 -7.35760510e-01 9.25487220e-01
2.05272183e-01 -4.22884703e-01 8.63689601e-01 6.58118784e-01
3.08578104e-01 -9.62870538e-01 -1.11371076e+00 -6.61283970e-01
-2.15285555e-01 -5.42800188e-01 5.04832208e-01 6.06576264e-01
-2.33768612e-01 3.33418250e-01 -2.96974599e-01 5.86156361e-02
6.38896227e-01 -2.68024337e-02 4.57058668e-01 -8.86593640e-01
-5.45442998e-01 -5.43805957e-01 -6.45776749e-01 -1.73029363e+00
-4.72986013e-01 -5.68518400e-01 1.21306498e-02 -1.57604539e+00
-1.98458403e-01 -6.21027648e-01 -1.96643591e-01 4.77584094e-01
-2.62753338e-01 8.57222617e-01 1.62646875e-01 4.02477562e-01
-6.09828055e-01 7.19658554e-01 1.41133237e+00 -1.67950749e-01
-4.39706415e-01 -2.54480720e-01 -3.11545372e-01 4.67136592e-01
6.47251725e-01 -2.16235265e-01 -9.14370716e-01 -7.54514217e-01
9.46076885e-02 9.18204617e-03 1.49130777e-01 -1.17042410e+00
3.57599169e-01 6.70572594e-02 6.29518032e-01 -6.48615658e-01
6.61742985e-01 -3.18466008e-01 1.23550914e-01 6.03140891e-01
-2.34015509e-01 2.24081337e-01 2.79274851e-01 8.47237051e-01
-2.14166045e-01 6.39807880e-02 7.99110711e-01 1.44435927e-01
-7.97696710e-01 4.43806201e-01 -5.96108973e-01 -1.42670184e-01
7.70645618e-01 -5.68445385e-01 1.94940101e-02 -5.91571510e-01
-5.93458056e-01 8.73158500e-02 5.36519408e-01 2.49149680e-01
7.59504020e-01 -1.45280397e+00 -1.83565035e-01 2.94794917e-01
-2.87584692e-01 1.09332673e-01 1.33097053e-01 5.92319548e-01
-4.24775064e-01 3.07630032e-01 -2.92663574e-02 -9.27492917e-01
-1.18670499e+00 2.65832722e-01 1.80785462e-01 -1.45778373e-01
-7.78291285e-01 9.56125557e-01 -2.10155517e-01 3.12419027e-01
4.09205437e-01 -5.60182869e-01 2.05684185e-01 -1.47845462e-01
7.47236669e-01 4.16722149e-01 2.50433423e-02 -2.38878503e-01
-8.34227800e-02 5.27587533e-01 -1.23760387e-01 -5.08274138e-01
8.74473572e-01 -2.74968594e-01 3.08809936e-01 5.25064766e-01
1.03841460e+00 -3.80225450e-01 -1.66941154e+00 -1.64749250e-01
-3.00483108e-01 -5.26980460e-01 4.99573499e-01 -6.22074366e-01
-9.74240482e-01 9.01916683e-01 8.58657837e-01 2.58365035e-01
1.21880686e+00 -3.52920443e-01 1.16042721e+00 4.37534690e-01
4.66993421e-01 -1.17252028e+00 8.09612423e-02 3.82845640e-01
4.76913244e-01 -9.74501669e-01 6.99231029e-02 -2.53060937e-01
-5.39361358e-01 9.73655879e-01 2.90926158e-01 1.72208115e-01
6.40609562e-01 4.57773209e-01 2.15371579e-01 7.44349062e-02
-9.47972536e-01 1.79684550e-01 5.09638600e-02 8.24956894e-01
2.53851920e-01 7.72267804e-02 -2.56387353e-01 -1.06608376e-01
-3.71438980e-01 2.99265474e-01 9.42103192e-02 1.02531910e+00
-4.97627169e-01 -1.05547988e+00 -1.43090367e-01 5.69239974e-01
-3.01148862e-01 -1.31382197e-01 -2.24087667e-03 3.80814195e-01
4.40559350e-03 8.95588100e-01 3.62588525e-01 -3.28601003e-01
-1.79991156e-01 -1.04836069e-01 7.59158075e-01 -2.62277603e-01
-2.83412546e-01 1.95314884e-01 -3.90834510e-01 -6.17391109e-01
-3.66321295e-01 -5.59657633e-01 -1.47426760e+00 -3.73936296e-01
-3.51025373e-01 1.23689689e-01 8.08995187e-01 7.58008182e-01
6.41150057e-01 5.27188659e-01 2.78239548e-01 -7.86823452e-01
-7.18630612e-01 -8.74553382e-01 -5.00749826e-01 1.30471036e-01
3.16226959e-01 -7.66397297e-01 9.49301720e-02 3.10981333e-01] | [11.125394821166992, -0.42279669642448425] |
86233e2f-6b6c-43a2-a2ba-8bc1ebe80926 | the-usfd-spoken-language-translation-system | 1509.03870 | null | http://arxiv.org/abs/1509.03870v1 | http://arxiv.org/pdf/1509.03870v1.pdf | The USFD Spoken Language Translation System for IWSLT 2014 | The University of Sheffield (USFD) participated in the International Workshop
for Spoken Language Translation (IWSLT) in 2014. In this paper, we will
introduce the USFD SLT system for IWSLT. Automatic speech recognition (ASR) is
achieved by two multi-pass deep neural network systems with adaptation and
rescoring techniques. Machine translation (MT) is achieved by a phrase-based
system. The USFD primary system incorporates state-of-the-art ASR and MT
techniques and gives a BLEU score of 23.45 and 14.75 on the English-to-French
and English-to-German speech-to-text translation task with the IWSLT 2014 data.
The USFD contrastive systems explore the integration of ASR and MT by using a
quality estimation system to rescore the ASR outputs, optimising towards better
translation. This gives a further 0.54 and 0.26 BLEU improvement respectively
on the IWSLT 2012 and 2014 evaluation data. | ['Ghada Alharbi', 'Mortaza Doulaty', 'Lucia Specia', 'Raymond W. M. Ng', 'Oscar Saz', 'Kashif Shah', 'Rama Doddipatla', 'Wilker Aziz', 'Thomas Hain', 'Madina Hasan'] | 2015-09-13 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 3.98607939e-01 2.21377864e-01 -1.10509964e-02 -5.07865787e-01
-1.62079298e+00 -5.71822166e-01 8.68823826e-01 -4.50823605e-02
-7.35880077e-01 9.21052814e-01 5.18424273e-01 -8.03744316e-01
1.06157601e-01 -2.61656046e-01 -6.24869943e-01 -2.97965884e-01
4.07039523e-01 1.02407217e+00 -1.23354718e-01 -4.19832557e-01
-9.04433951e-02 2.72126645e-01 -9.97629583e-01 7.70683408e-01
8.70820284e-01 6.20264828e-01 4.13999885e-01 1.13485849e+00
-1.21220194e-01 3.16567510e-01 -6.77996755e-01 -3.32365841e-01
2.90753305e-01 -4.34883088e-01 -1.21895850e+00 -2.08677381e-01
5.25678813e-01 -1.30656868e-01 -3.44652802e-01 5.89379013e-01
9.32534397e-01 4.47969697e-02 3.77998650e-01 -6.10218287e-01
-7.46759951e-01 9.18525279e-01 -1.85452953e-01 3.18196684e-01
5.23078620e-01 -9.26207379e-02 9.14869308e-01 -1.42776632e+00
5.69969893e-01 1.37407303e+00 4.09933746e-01 8.04657936e-01
-1.21261215e+00 -3.67313266e-01 -3.91330689e-01 3.71233881e-01
-1.20135415e+00 -1.03551817e+00 -3.93921658e-02 5.65862395e-02
1.93458188e+00 4.07382846e-01 2.48874292e-01 1.14633906e+00
2.80155003e-01 8.71145070e-01 1.08169019e+00 -9.44088101e-01
1.87871918e-01 2.20157295e-01 -2.24084929e-01 2.54760325e-01
-3.38938296e-01 1.42509133e-01 -8.27756345e-01 9.84272659e-02
4.65834230e-01 -6.47235215e-01 -1.09508827e-01 5.00997305e-01
-1.52998066e+00 6.44253194e-01 -8.68491770e-05 5.82523227e-01
-6.12315178e-01 -1.93371117e-01 6.63488865e-01 9.03535545e-01
7.55652666e-01 3.52604330e-01 -9.00394917e-01 -4.35103297e-01
-1.28184569e+00 -8.71041119e-02 8.41821909e-01 9.11965847e-01
3.11641693e-01 3.62569332e-01 -5.69034815e-01 1.36074233e+00
4.44131345e-01 9.88215148e-01 8.52235973e-01 -8.47332716e-01
8.44847143e-01 6.90587051e-03 8.10399130e-02 -6.70230836e-02
-1.59542233e-01 -5.32141328e-01 -6.37418270e-01 -2.22831905e-01
7.74429590e-02 -3.38152677e-01 -1.33373106e+00 1.64083946e+00
-7.67531544e-02 -2.33026907e-01 6.18202567e-01 8.78125727e-01
9.11855400e-01 1.20832932e+00 -1.46625951e-01 -4.52229977e-01
1.07135391e+00 -1.29588366e+00 -9.87317502e-01 -4.41858321e-01
8.53761196e-01 -1.25331724e+00 1.05691111e+00 1.40229672e-01
-1.52090478e+00 -5.52615106e-01 -8.21224570e-01 3.79783697e-02
-1.98230445e-01 5.28107643e-01 -2.74850219e-01 5.72328985e-01
-1.78807318e+00 4.38462377e-01 -8.79790545e-01 -8.36512744e-01
-2.56941229e-01 5.53875685e-01 -5.15859485e-01 3.39529873e-03
-1.63658869e+00 1.39093828e+00 1.74383596e-01 1.40093490e-01
-6.24889016e-01 -2.39508376e-01 -8.02108049e-01 -9.86379683e-02
-1.28024250e-01 -4.63128060e-01 1.84342158e+00 -1.09165657e+00
-2.16092229e+00 9.89311337e-01 -6.80053353e-01 -8.27013016e-01
4.33733612e-01 -2.41451010e-01 -6.55652583e-01 4.86284196e-02
1.98001266e-01 6.91108644e-01 4.10839111e-01 -6.35812879e-01
-7.08801270e-01 -2.56469280e-01 -6.86135173e-01 5.51788867e-01
1.52239427e-02 6.41009212e-01 -4.65906598e-02 -4.41924632e-01
7.26671666e-02 -9.38681304e-01 1.55302897e-01 -7.65780866e-01
-2.68772542e-01 -3.75643581e-01 5.08111835e-01 -1.28808904e+00
1.31784296e+00 -1.72827411e+00 3.40189636e-01 -3.01335633e-01
-5.67622423e-01 7.98193634e-01 -6.32635772e-01 8.99643481e-01
-8.77392590e-02 -1.60999969e-02 -2.89885312e-01 -7.51038194e-01
1.59395300e-02 4.38781381e-01 -2.09306002e-01 2.04438865e-01
4.12143588e-01 1.01143146e+00 -6.12905562e-01 -7.04779103e-02
1.68577358e-01 3.68778408e-01 1.39360666e-01 4.11357462e-01
1.36470094e-01 2.64213204e-01 5.16248196e-02 3.09142172e-01
4.47700500e-01 2.98256725e-01 -8.01141262e-02 3.91898364e-01
-4.42758650e-01 1.18703258e+00 -5.29206157e-01 1.87717509e+00
-6.97250962e-01 7.61235535e-01 8.90054256e-02 -7.33666718e-01
1.09797633e+00 9.15526092e-01 -1.20095260e-01 -9.19138610e-01
1.45039886e-01 8.95438313e-01 1.56994566e-01 -2.69764364e-01
6.17802978e-01 -3.08947027e-01 8.10560137e-02 6.27528667e-01
4.06300753e-01 -1.47255212e-01 1.18236229e-01 5.05233333e-02
9.98449922e-01 1.59031466e-01 1.23927832e-01 -3.81369114e-01
6.12225890e-01 3.61687727e-02 2.77210116e-01 7.23862708e-01
-2.43060708e-01 7.58221090e-01 -3.06679428e-01 -2.57853657e-01
-1.37309277e+00 -1.00027668e+00 1.83527134e-02 1.10160625e+00
-7.99994409e-01 -1.10338032e-01 -1.09156406e+00 -4.78576630e-01
-5.76170862e-01 1.27356744e+00 -3.28668833e-01 -1.52009383e-01
-8.11793327e-01 -4.66510266e-01 8.34865093e-01 1.66231439e-01
4.35529917e-01 -1.33400512e+00 -2.12414674e-02 5.78682244e-01
-6.19614899e-01 -1.28159487e+00 -7.73887336e-01 3.38577300e-01
-7.10000515e-01 -5.17140701e-02 -1.21741617e+00 -9.76855457e-01
-8.97303149e-02 8.27386752e-02 1.04405582e+00 -4.11872685e-01
3.57101887e-01 3.22136357e-02 -5.18660843e-01 -4.09277320e-01
-1.33987343e+00 6.57147884e-01 4.72273827e-01 -2.27301121e-01
6.80234730e-01 -2.81472474e-01 -9.73728895e-02 2.29777083e-01
-4.71993864e-01 2.33800337e-01 8.31177354e-01 8.89092803e-01
3.34705174e-01 -9.17857349e-01 8.45410109e-01 -4.22192395e-01
7.09320784e-01 -7.53906891e-02 -2.59585083e-01 3.54788035e-01
-8.27567399e-01 3.63553315e-02 5.70817590e-01 -2.68055767e-01
-1.05441499e+00 -1.49343938e-01 -7.14391172e-01 -1.92319795e-01
-3.44801068e-01 5.16450226e-01 1.34265656e-02 3.04010868e-01
7.34159470e-01 5.39870799e-01 -1.11921623e-01 -4.49358314e-01
1.66814268e-01 1.43477964e+00 3.94730210e-01 -1.26108959e-01
4.32262689e-01 -4.54131186e-01 -5.43784678e-01 -7.75362313e-01
-6.68673635e-01 -4.13602740e-01 -7.07966328e-01 -2.36634612e-02
7.36071765e-01 -8.79763544e-01 3.78386751e-02 6.01824880e-01
-1.53678346e+00 -5.83340228e-01 -2.04308152e-01 6.69636548e-01
-6.55434966e-01 1.09378904e-01 -9.01340425e-01 -8.80644917e-01
-1.04167926e+00 -1.36663508e+00 1.34726727e+00 -1.04001202e-01
-5.72378755e-01 -8.02170098e-01 2.90709525e-01 4.70565647e-01
6.67049885e-01 -6.30678654e-01 5.70437253e-01 -1.06029487e+00
9.00721326e-02 -7.03392029e-02 -4.18184809e-02 7.03635752e-01
4.13849652e-02 -3.93503815e-01 -9.74053919e-01 -4.74891782e-01
-2.62271278e-02 -2.06041098e-01 5.38296700e-01 4.82550800e-01
3.74979489e-02 -4.11295682e-01 2.15735346e-01 1.73960313e-01
1.04522526e+00 3.75296205e-01 5.48981786e-01 5.95271647e-01
3.24073046e-01 6.32690728e-01 5.44394732e-01 -5.35519235e-02
3.35611552e-01 8.48040223e-01 -2.26998195e-01 1.36556908e-01
-3.92373621e-01 -5.31304665e-02 9.01718855e-01 1.41971600e+00
2.84812987e-01 -5.04430532e-01 -1.19857299e+00 7.82970846e-01
-1.72635865e+00 -4.93543714e-01 -4.08433706e-01 2.14791489e+00
1.07369375e+00 8.34055766e-02 -3.86720486e-02 6.40297607e-02
8.04830730e-01 -8.41624588e-02 -1.33635232e-03 -1.22799385e+00
-1.90194070e-01 5.13156295e-01 5.27595758e-01 8.19744527e-01
-5.59180260e-01 1.38587463e+00 6.03042459e+00 9.00669873e-01
-1.22340715e+00 4.95413393e-01 6.29056036e-01 -1.08801723e-01
-5.18703274e-02 -3.20426643e-01 -8.34662795e-01 1.87857464e-01
2.13902044e+00 -2.60320276e-01 7.88175702e-01 1.55428246e-01
6.57210886e-01 1.31732047e-01 -8.51950645e-01 7.08101511e-01
2.01427303e-02 -1.14197505e+00 9.48975757e-02 -1.86492845e-01
8.29345584e-01 8.22065234e-01 -2.20833775e-02 4.85222310e-01
7.48703033e-02 -1.03642845e+00 8.49060714e-01 2.22199425e-01
1.06590378e+00 -8.83119404e-01 1.10860658e+00 4.81556058e-01
-6.86749220e-01 2.68114984e-01 -3.35301459e-01 1.11073032e-01
2.80057341e-01 3.01822066e-01 -1.37111712e+00 7.02041626e-01
5.88961303e-01 4.59833771e-01 -1.37040153e-01 5.42663634e-01
-3.84457678e-01 1.11112988e+00 -3.64547729e-01 -1.50025323e-01
6.70084953e-01 -5.26466742e-02 9.41463351e-01 1.67899799e+00
5.91627479e-01 -3.09150845e-01 -1.94261864e-01 5.11608958e-01
-1.81958318e-01 3.81804228e-01 -2.85279870e-01 -1.59196943e-01
5.00077188e-01 8.89071465e-01 -1.47155136e-01 -4.31855053e-01
-5.51444590e-02 1.33659804e+00 2.08015651e-01 4.82278258e-01
-2.75132090e-01 -3.72374862e-01 5.51273406e-01 3.80585268e-02
1.07012890e-01 -3.60687137e-01 -4.22119319e-01 -9.18610930e-01
5.55363111e-02 -1.12782049e+00 -1.12429298e-01 -8.87760937e-01
-9.23644900e-01 1.27757287e+00 -2.80727595e-01 -8.60651493e-01
-8.08259010e-01 -3.77785832e-01 -4.62013900e-01 1.53423297e+00
-1.49594975e+00 -1.27488697e+00 5.26236773e-01 6.40371442e-02
1.24748361e+00 -6.54042244e-01 1.07324541e+00 2.88432747e-01
-3.16359103e-01 8.24482322e-01 4.32361424e-01 -9.39297751e-02
9.96809304e-01 -1.19439924e+00 1.25690842e+00 8.39138865e-01
1.18319407e-01 4.83303696e-01 8.41198742e-01 -5.46661019e-01
-1.25091290e+00 -1.32992935e+00 1.94678795e+00 -4.43354607e-01
6.48207188e-01 -3.48332971e-01 -8.62882376e-01 5.67540407e-01
8.46374154e-01 -6.69109404e-01 4.57510114e-01 -2.60533571e-01
5.04295677e-02 1.61836460e-01 -1.06035507e+00 5.02471268e-01
7.37322450e-01 -7.31802046e-01 -7.39816666e-01 3.22069198e-01
1.09558952e+00 -4.13679540e-01 -7.40731180e-01 3.25395077e-01
4.17505503e-01 -4.40453529e-01 5.22666752e-01 -4.79164600e-01
3.58312964e-01 -1.44043574e-02 -3.74459594e-01 -1.78804672e+00
-2.48158649e-01 -1.00708425e+00 1.73582762e-01 1.06088638e+00
1.03866196e+00 -6.58598840e-01 2.38762945e-01 6.75073490e-02
-5.41042387e-01 -5.78724444e-01 -1.59564340e+00 -8.04690599e-01
3.23850244e-01 -4.77985859e-01 3.79128128e-01 4.95218068e-01
-1.52099550e-01 8.79162371e-01 -3.62640113e-01 -1.17726885e-02
1.32004559e-01 -5.70236325e-01 3.51005197e-01 -7.78511584e-01
-1.67735294e-01 -4.12915170e-01 3.31317596e-02 -8.96253705e-01
2.24774778e-01 -1.05541766e+00 4.03758407e-01 -1.85442317e+00
-8.51840526e-02 3.07634950e-01 -2.13007331e-01 5.89520931e-01
2.05243118e-02 2.98717082e-01 3.52919511e-02 2.01118752e-01
-2.41354346e-01 6.29582465e-01 9.71486628e-01 1.30361337e-02
-1.91632241e-01 7.26984069e-02 -2.25728810e-01 -7.55451024e-02
1.14746642e+00 -6.33159757e-01 -2.45787343e-03 -8.97132337e-01
-2.16815874e-01 3.98586780e-01 -2.04411700e-01 -6.71298265e-01
1.21974573e-01 1.82457268e-01 1.00893356e-01 -6.87883914e-01
2.90644199e-01 -3.32887858e-01 -6.31744042e-02 3.54678631e-01
-6.20278120e-01 3.38253587e-01 5.89510977e-01 1.22512151e-02
-2.71410018e-01 -3.28744531e-01 9.95648861e-01 -7.24862665e-02
-1.04648046e-01 -4.80708070e-02 -1.11247694e+00 -2.31217802e-01
2.66686112e-01 -1.21315017e-01 -2.18091737e-02 -6.31009221e-01
-7.22356379e-01 1.70497596e-01 1.13091290e-01 7.13788986e-01
6.06160641e-01 -1.26948130e+00 -1.52037132e+00 3.91345799e-01
-1.56960376e-02 -4.04773682e-01 -3.56225520e-02 1.01781881e+00
-2.87485033e-01 9.75806952e-01 -1.06944935e-02 -5.12648284e-01
-1.44654417e+00 -2.02946559e-01 4.33809102e-01 -4.43420470e-01
-1.77571610e-01 7.52246439e-01 -4.33190286e-01 -1.05460715e+00
4.35212888e-02 9.68304798e-02 -5.16778119e-02 -2.47826815e-01
5.13318002e-01 3.59043598e-01 6.53779268e-01 -9.53897119e-01
-3.25029284e-01 9.54848230e-02 -3.04594547e-01 -1.02466047e+00
1.35931885e+00 -5.76629043e-01 -1.57799050e-01 5.72778404e-01
1.17366767e+00 -2.29066640e-01 -5.48921704e-01 -4.51161444e-01
2.91873515e-01 8.45936462e-02 3.55812669e-01 -1.48667181e+00
-3.70803416e-01 1.10434496e+00 6.69899702e-01 4.77107875e-02
1.04476488e+00 -2.36531124e-01 1.27870929e+00 6.33336484e-01
2.52676129e-01 -1.29939103e+00 -6.12307787e-01 1.27473402e+00
1.12605977e+00 -1.25328243e+00 -7.70934939e-01 1.67711928e-01
-5.93608379e-01 1.29980493e+00 1.32215142e-01 2.31950179e-01
9.31854942e-04 1.57483995e-01 5.02050281e-01 4.27531630e-01
-1.03025806e+00 -8.41413736e-02 4.93474007e-01 3.39421839e-01
8.25603604e-01 2.40812689e-01 -5.23628831e-01 3.37395787e-01
-4.24759537e-01 -5.77334724e-02 4.02297825e-01 6.83886647e-01
-5.05482852e-01 -1.42013431e+00 -3.34993750e-01 2.28967980e-01
-6.76146269e-01 -5.92062473e-01 -7.50726998e-01 3.71041417e-01
-5.64648211e-01 1.42492676e+00 -1.07852489e-01 -4.79988217e-01
4.33354586e-01 4.73615468e-01 2.64005184e-01 -9.11526799e-01
-1.00158453e+00 5.24587035e-01 6.28181875e-01 -3.59661251e-01
-1.69084042e-01 -8.98704171e-01 -1.19326425e+00 -2.03260928e-01
-2.80079156e-01 4.88579512e-01 1.10489011e+00 1.12182319e+00
4.65230525e-01 4.15409148e-01 7.20828593e-01 -6.29602253e-01
-6.93054438e-01 -1.70122814e+00 -1.28593966e-01 -3.04543614e-01
4.41710413e-01 2.02749193e-01 -2.35750273e-01 -1.55508444e-01] | [14.451739311218262, 7.136730194091797] |
b1a6b2c8-a97c-495d-8b67-828d53e5f769 | learning-cross-modal-context-graph-for-visual | null | null | https://arxiv.org/pdf/1911.09042.pdf | https://arxiv.org/pdf/1911.09042.pdf | Learning Cross-modal Context Graph for Visual Grounding | Visual grounding is a ubiquitous building block in many vision-language tasks and yet remains challenging due to large variations in visual and linguistic features of grounding entities, strong context effect and the resulting semantic ambiguities. Prior works typically focus on learning representations of individual phrases with limited context information. To address their limitations, this paper proposes a language-guided graph representation to capture the global context of grounding entities and their relations, and develop a cross-modal graph matching strategy for the multiple-phrase visual grounding task. In particular, we introduce a modular graph neural network to compute context-aware representations of phrases and object proposals respectively via message propagation, followed by a graph-based matching module to generate globally consistent localization of grounding phrases. We train the entire graph neural network jointly in a two-stage strategy and evaluate it on the Flickr30K Entities benchmark. Extensive experiments show that our method outperforms the prior state of the arts by a sizable margin, evidencing the efficacy of our grounding framework. Code is available at https://github.com/youngfly11/LCMCG-PyTorch. | ['Yongfei Liu; Bo Wan; Xiaodan Zhu; Xuming He'] | 2020-02-13 | null | null | null | aaai-2020-2020-2 | ['phrase-grounding', 'natural-language-visual-grounding'] | ['natural-language-processing', 'reasoning'] | [ 8.50324929e-02 1.97820619e-01 -2.14727089e-01 -2.37946287e-01
-7.80250847e-01 -6.88867748e-01 7.25393414e-01 4.24857795e-01
-2.44744509e-01 3.55388105e-01 3.55933905e-01 -3.30599695e-01
1.41085938e-01 -7.84769416e-01 -8.45014691e-01 -3.59186679e-01
-4.17145304e-02 2.38150969e-01 3.94798934e-01 -3.47909182e-01
1.69687793e-01 8.87682512e-02 -1.35730338e+00 3.61127973e-01
7.66315103e-01 8.22548211e-01 4.48509037e-01 6.31933510e-01
-3.56810689e-01 6.91156864e-01 -1.83647245e-01 -6.21915817e-01
1.59935638e-01 -2.49049708e-01 -8.77858996e-01 1.85005948e-01
1.06412196e+00 -3.24616544e-02 -5.41760385e-01 1.27152169e+00
4.00988728e-01 1.79100141e-01 3.89074296e-01 -1.48921514e+00
-1.22298944e+00 6.77899361e-01 -6.18862510e-01 1.24510892e-01
4.78326112e-01 5.85315488e-02 1.69982672e+00 -1.07500601e+00
8.30011189e-01 1.19712877e+00 5.74018121e-01 3.74710262e-01
-1.09838152e+00 -5.71917117e-01 6.66337311e-01 2.35141203e-01
-1.73939586e+00 -7.99009427e-02 8.23385835e-01 -5.35982549e-01
1.07999778e+00 -1.13348197e-02 6.98454678e-01 9.23687577e-01
2.86291000e-02 7.00429261e-01 7.95153379e-01 -3.96664053e-01
-4.84621301e-02 -6.45696819e-02 3.40854287e-01 1.19260585e+00
4.48444515e-01 -1.84188977e-01 -6.61830783e-01 -1.53630689e-01
6.60125613e-01 2.04143487e-02 -3.51185590e-01 -5.49074531e-01
-1.42962992e+00 7.29359925e-01 1.13235235e+00 3.59532297e-01
-3.33273083e-01 6.32571459e-01 1.05197005e-01 9.41662118e-02
4.12967116e-01 2.24220693e-01 1.41715948e-02 5.78271389e-01
-8.33920896e-01 4.96976346e-01 4.62732732e-01 1.20772088e+00
1.04416549e+00 -3.78732741e-01 -5.15393674e-01 5.46907008e-01
8.07773352e-01 5.58219373e-01 7.72977322e-02 -3.80537659e-01
8.66797090e-01 9.03137445e-01 9.50485468e-02 -1.69923067e+00
-3.46471727e-01 -4.50118035e-01 -6.72403455e-01 -2.56197780e-01
5.27083986e-02 1.52969554e-01 -1.06609416e+00 1.72846711e+00
4.97891814e-01 6.71115160e-01 -4.85071428e-02 1.02305591e+00
1.46393514e+00 6.24111056e-01 4.75937575e-01 3.80854577e-01
1.57866585e+00 -1.18888843e+00 -5.19485712e-01 -5.41818261e-01
4.45015132e-01 -7.39244580e-01 1.02160537e+00 -2.23811999e-01
-8.26838791e-01 -5.47064781e-01 -9.63933825e-01 -3.42252076e-01
-5.97442746e-01 1.11474320e-01 6.90349638e-01 1.79125890e-01
-1.36991131e+00 3.25547814e-01 -4.93989140e-01 -6.69452429e-01
4.41792518e-01 1.58866122e-01 -4.24057037e-01 -9.12019238e-02
-1.17937732e+00 6.93200052e-01 7.16313243e-01 4.18293089e-01
-6.89088345e-01 -5.39220691e-01 -1.12272847e+00 -6.04748391e-02
4.05057311e-01 -1.09004128e+00 9.19910669e-01 -6.20456338e-01
-7.64942646e-01 1.23736477e+00 -7.68367872e-02 -4.58495259e-01
2.53373444e-01 -1.55431181e-01 -4.02950674e-01 1.32138982e-01
3.08154792e-01 9.07081783e-01 7.35773802e-01 -1.35070086e+00
-7.88566947e-01 -2.02572346e-01 4.66613889e-01 3.94421667e-01
-2.62116734e-02 -6.72421604e-02 -8.83826435e-01 -6.34959459e-01
3.13862234e-01 -9.58836436e-01 -3.41931462e-01 -7.47908354e-02
-6.36764526e-01 -2.87178516e-01 3.89098734e-01 -6.10173881e-01
1.27476358e+00 -2.00605440e+00 2.88691401e-01 1.85417920e-01
4.83607560e-01 -5.01964008e-03 -3.39657038e-01 6.90815210e-01
1.55891329e-01 1.34117514e-01 -1.67561099e-01 -3.99657696e-01
2.34699458e-01 1.04186237e-01 -4.93364990e-01 3.93751949e-01
4.78982568e-01 1.41523111e+00 -1.11457312e+00 -5.61330974e-01
8.25634524e-02 5.63736916e-01 -4.74191040e-01 2.47365236e-01
-4.91233021e-01 1.72284573e-01 -5.11274517e-01 7.88009524e-01
5.44182003e-01 -7.78859615e-01 1.89339057e-01 -5.05556762e-01
1.12290874e-01 6.61083832e-02 -1.42573118e+00 2.18073797e+00
-2.01168701e-01 4.67623681e-01 -1.11063167e-01 -7.93619633e-01
8.45862329e-01 1.75183013e-01 -2.81341989e-02 -6.06602907e-01
2.59264540e-02 9.59780440e-02 -4.83987927e-01 -3.70954245e-01
8.87605786e-01 1.27342045e-01 -1.24363028e-01 1.07476093e-01
2.71963596e-01 3.43169086e-02 9.96110067e-02 6.43300235e-01
9.12100911e-01 1.65441126e-01 2.80283004e-01 -1.09275609e-01
4.09785748e-01 5.91627657e-02 2.58240908e-01 8.14816177e-01
-1.28157869e-01 7.72727430e-01 2.40148991e-01 -2.87069947e-01
-6.58914208e-01 -8.36049557e-01 3.67728204e-01 1.16188431e+00
7.77086079e-01 -7.23681569e-01 -4.00739521e-01 -8.33298922e-01
1.66122451e-01 4.35763270e-01 -6.43553734e-01 6.88250512e-02
-4.64134544e-01 -4.20161724e-01 3.15806538e-01 4.30723459e-01
5.12364388e-01 -1.02289736e+00 -2.35527560e-01 5.70703335e-02
-3.08285594e-01 -1.50054002e+00 -5.58471620e-01 -2.00197339e-01
-5.16680777e-01 -1.19289696e+00 -5.45476198e-01 -1.01603281e+00
7.08757102e-01 6.12517595e-01 1.46597064e+00 5.57118714e-01
-2.37273976e-01 6.97813153e-01 -3.17194700e-01 -1.58964127e-01
-5.00554517e-02 1.53054699e-01 -5.19008934e-01 1.82724506e-01
2.96806842e-01 -2.94617295e-01 -8.60281825e-01 4.80084941e-02
-7.35532463e-01 3.27857494e-01 4.93008494e-01 5.97098053e-01
7.68650174e-01 -3.62520695e-01 1.72576621e-01 -8.32231104e-01
6.16407931e-01 -6.38286829e-01 -8.01267922e-01 5.08376658e-01
-5.54369152e-01 4.91826274e-02 -4.36034761e-02 -5.04675210e-02
-6.73108578e-01 1.12916259e-02 1.24262258e-01 -3.67526591e-01
-2.31818743e-02 5.69202304e-01 -2.28744503e-02 -2.06812933e-01
5.01265347e-01 9.34614465e-02 -5.36078870e-01 -1.68891743e-01
8.82863224e-01 2.38806129e-01 5.78467667e-01 -5.83378911e-01
1.13658035e+00 4.41554189e-01 -3.04981582e-02 -5.15449226e-01
-8.94162178e-01 -8.34340930e-01 -4.94852185e-01 -2.00922489e-01
1.15005898e+00 -1.32606673e+00 -3.07727158e-01 1.07916348e-01
-1.26807129e+00 -2.84265935e-01 1.29007891e-01 1.79077044e-01
-2.51324713e-01 4.51518983e-01 -3.08243930e-01 -4.22179133e-01
-5.84647596e-01 -9.94699061e-01 1.52971721e+00 3.01502585e-01
1.59415707e-01 -1.02643228e+00 2.02766910e-01 3.97851795e-01
1.50226042e-01 4.01143581e-01 7.20457792e-01 -5.59664488e-01
-1.12238610e+00 -2.19843220e-02 -7.03023911e-01 -1.21857055e-01
-8.00184608e-02 -3.02818604e-02 -8.63697290e-01 -2.65215248e-01
-9.07626450e-01 -2.72863448e-01 1.06600952e+00 1.60923779e-01
6.56677067e-01 -1.62800506e-01 -5.10373592e-01 6.38266444e-01
1.82903397e+00 -4.00270730e-01 2.94235051e-01 2.67005265e-01
1.26923609e+00 3.76868665e-01 5.83810210e-01 1.86743662e-01
9.12409544e-01 7.59259164e-01 8.72344375e-01 -2.74003804e-01
-3.97447258e-01 -5.97145379e-01 -7.22754970e-02 6.25607193e-01
8.22580513e-03 -4.97325331e-01 -1.12369227e+00 8.49784911e-01
-2.20363522e+00 -8.58630180e-01 -1.80837318e-01 1.85284448e+00
4.56347466e-01 -7.60274380e-02 -3.50697339e-02 -4.50711340e-01
1.04139125e+00 5.41351259e-01 -5.76557629e-02 -3.31117474e-02
-8.99830535e-02 -2.72624716e-02 4.24859643e-01 6.35970533e-01
-1.31209886e+00 1.46255648e+00 5.11919451e+00 5.60023248e-01
-9.28364098e-01 1.03029780e-01 2.32189834e-01 3.14422995e-01
-6.97927773e-01 1.78684577e-01 -9.07607853e-01 1.20450832e-01
3.17351758e-01 -1.07481018e-01 2.96939522e-01 6.63751245e-01
-8.61237496e-02 2.68856645e-01 -9.08732891e-01 1.14653957e+00
1.13885805e-01 -1.67062151e+00 3.07112247e-01 -4.25549150e-02
7.56896436e-01 3.62967879e-01 -1.25691190e-01 2.71679401e-01
3.71604413e-01 -9.80274022e-01 9.11317408e-01 3.68568420e-01
4.18329030e-01 -4.13109571e-01 5.29121876e-01 -9.23653841e-02
-1.75008249e+00 1.45677939e-01 -4.19927299e-01 8.63698497e-02
2.18531162e-01 2.84287900e-01 -9.30717945e-01 9.47515845e-01
6.03412032e-01 8.22431624e-01 -8.65555584e-01 9.46464419e-01
-5.79822361e-01 4.77658153e-01 -1.39411733e-01 -5.91229349e-02
6.61825836e-01 -1.12414144e-01 5.79062998e-01 1.43998551e+00
1.80559561e-01 7.16316840e-03 4.58516181e-01 9.85565066e-01
-2.71955878e-01 3.78404349e-01 -6.88522518e-01 -2.25458384e-01
5.67442536e-01 1.61695325e+00 -9.77243006e-01 -2.67485917e-01
-6.60112262e-01 9.48600590e-01 7.51438022e-01 5.92651904e-01
-9.42762256e-01 -1.81393236e-01 4.31970209e-01 3.78020480e-02
5.36410749e-01 -4.11552668e-01 1.82247803e-01 -1.36732149e+00
1.31370187e-01 -6.13700688e-01 6.49982154e-01 -9.99596536e-01
-1.22529447e+00 6.49128973e-01 -1.69873089e-02 -9.39840138e-01
8.57996121e-02 -5.69284022e-01 -5.75709879e-01 7.40878820e-01
-1.88777292e+00 -1.83835804e+00 -8.33568811e-01 7.32632995e-01
2.37853736e-01 4.06771302e-02 6.00564837e-01 2.91093022e-01
-3.73598933e-01 4.15692866e-01 -5.64068735e-01 3.43973726e-01
4.65616465e-01 -1.36262000e+00 7.66224623e-01 1.14629257e+00
8.58870625e-01 6.37913823e-01 6.65610373e-01 -6.91042006e-01
-1.49933124e+00 -1.38051248e+00 1.08607197e+00 -4.37832624e-01
9.51323092e-01 -5.77773452e-01 -9.29994404e-01 8.27530086e-01
4.72613394e-01 3.93099666e-01 4.86572653e-01 1.49456233e-01
-6.69335067e-01 1.44086033e-01 -6.94778621e-01 7.90164948e-01
1.36079884e+00 -7.65481949e-01 -6.62455261e-01 4.14832234e-01
1.08042216e+00 -6.17879748e-01 -5.60842097e-01 3.32171828e-01
2.98287183e-01 -7.08055377e-01 1.12641764e+00 -5.82143307e-01
3.56425166e-01 -7.95010924e-01 -4.22687918e-01 -8.90171289e-01
-5.05500674e-01 -4.83103037e-01 -3.44178528e-02 1.41855872e+00
4.56311911e-01 -3.85126561e-01 6.73350394e-01 3.01717550e-01
-4.94908839e-02 -5.47035277e-01 -6.07415557e-01 -3.97923529e-01
-3.18747312e-01 -3.98245454e-01 6.60627782e-01 9.75272834e-01
-2.02824950e-01 4.94739473e-01 -1.68428898e-01 7.81530738e-01
6.46123827e-01 5.35298467e-01 7.89105833e-01 -9.18989301e-01
-2.95129001e-01 -2.99003720e-01 -7.09344923e-01 -9.46960807e-01
3.12876701e-01 -1.31938803e+00 -3.81248631e-02 -2.06896091e+00
2.57271051e-01 -4.02847767e-01 -5.37726700e-01 5.64270258e-01
-4.87199366e-01 5.07897139e-01 4.73988593e-01 5.09433858e-02
-1.08588445e+00 4.14721549e-01 1.06998730e+00 -4.28447694e-01
3.96241322e-02 -4.06211972e-01 -7.28480637e-01 5.18348157e-01
6.71164453e-01 -3.23523968e-01 -6.10170722e-01 -7.78448522e-01
7.24789977e-01 -3.19115311e-01 9.42666829e-01 -7.61623919e-01
3.59643728e-01 -9.35714021e-02 -1.11853875e-01 -5.93400776e-01
7.92223364e-02 -8.16374123e-01 2.28999883e-01 1.90753952e-01
-3.50191265e-01 2.40138248e-01 2.44331554e-01 1.01730251e+00
-3.07541549e-01 -5.86705543e-02 2.06176847e-01 -2.14703932e-01
-1.33354723e+00 6.18218362e-01 2.86072433e-01 4.23394501e-01
8.43110621e-01 -3.16661708e-02 -4.71441209e-01 -1.64134502e-01
-5.78455865e-01 3.55589211e-01 4.49285477e-01 6.16246998e-01
6.05491102e-01 -1.57727873e+00 -7.39238501e-01 -7.99064785e-02
6.09926701e-01 1.00473635e-01 1.83327213e-01 5.81991851e-01
-5.94095707e-01 2.79107839e-01 1.40781000e-01 -7.96058893e-01
-1.32494915e+00 5.74781835e-01 2.74962783e-01 -2.69622236e-01
-6.95569515e-01 1.23169339e+00 3.79436523e-01 -2.01900810e-01
1.55780450e-01 -5.28781533e-01 -3.20324600e-01 1.24569632e-01
3.16614360e-01 -1.05892003e-01 3.77097577e-02 -9.38586652e-01
-5.91890812e-01 9.10806894e-01 3.65596302e-02 8.36164355e-02
9.92152214e-01 -2.99326301e-01 -8.77747014e-02 1.17806502e-01
9.92958009e-01 -1.13434300e-01 -9.57209229e-01 -6.36728764e-01
2.05027223e-01 -4.24385071e-01 -8.97273142e-03 -4.86722112e-01
-9.86771107e-01 6.88403070e-01 4.43511099e-01 1.45523593e-01
9.19333696e-01 4.34856057e-01 5.65462172e-01 3.46097410e-01
3.82503688e-01 -6.70756817e-01 1.58532560e-01 2.46406257e-01
1.05629659e+00 -1.42955935e+00 6.83382601e-02 -6.78145289e-01
-5.80871820e-01 7.60929763e-01 5.78891397e-01 -2.86156803e-01
4.75150257e-01 -3.23528111e-01 9.64827165e-02 -6.18709743e-01
-5.94213843e-01 -6.45030379e-01 8.63278747e-01 4.96283203e-01
3.68068486e-01 1.09354191e-01 -9.04593021e-02 2.99468368e-01
-5.63126709e-03 -2.33866170e-01 5.26878126e-02 7.63558447e-01
-4.45177555e-01 -9.08746362e-01 -1.46184089e-02 -1.55352339e-01
-3.02897871e-01 -5.22024989e-01 -4.06738907e-01 8.47222686e-01
2.14971900e-01 9.11570430e-01 -1.35731444e-01 -2.70586282e-01
3.59110326e-01 -1.39220491e-01 3.52734238e-01 -8.42844069e-01
-6.73368275e-01 -1.02629885e-01 1.12038456e-01 -8.89941335e-01
-6.41745985e-01 -2.58441299e-01 -1.28087926e+00 -1.88090783e-02
-4.24602240e-01 -9.20165777e-02 4.55172509e-01 7.78164923e-01
6.85394049e-01 4.10497963e-01 4.54116277e-02 -7.57220089e-01
-4.44624089e-02 -5.33609629e-01 -3.20848197e-01 7.48655260e-01
2.60956764e-01 -5.89006603e-01 -1.73409238e-01 2.33707996e-03] | [10.450016021728516, 1.508952260017395] |
5e400278-6bbc-4012-8e62-e5584af20863 | local-prediction-aggregation-a-frustratingly | 2205.04183 | null | https://arxiv.org/abs/2205.04183v3 | https://arxiv.org/pdf/2205.04183v3.pdf | Attracting and Dispersing: A Simple Approach for Source-free Domain Adaptation | We propose a simple but effective source-free domain adaptation (SFDA) method. Treating SFDA as an unsupervised clustering problem and following the intuition that local neighbors in feature space should have more similar predictions than other features, we propose to optimize an objective of prediction consistency. This objective encourages local neighborhood features in feature space to have similar predictions while features farther away in feature space have dissimilar predictions, leading to efficient feature clustering and cluster assignment simultaneously. For efficient training, we seek to optimize an upper-bound of the objective resulting in two simple terms. Furthermore, we relate popular existing methods in domain adaptation, source-free domain adaptation and contrastive learning via the perspective of discriminability and diversity. The experimental results prove the superiority of our method, and our method can be adopted as a simple but strong baseline for future research in SFDA. Our method can be also adapted to source-free open-set and partial-set DA which further shows the generalization ability of our method. Code is available in https://github.com/Albert0147/AaD_SFDA. | ['Shangling Jui', 'Joost Van de Weijer', 'Kai Wang', 'Yaxing Wang', 'Shiqi Yang'] | 2022-05-09 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [-4.08602476e-01 -2.00014621e-01 -4.77320254e-01 -7.70666599e-01
-7.53614902e-01 -4.28474277e-01 5.12697399e-01 -4.86774221e-02
-2.10840389e-01 6.27254903e-01 2.88977802e-01 1.64048761e-01
-3.92754316e-01 -6.83133662e-01 -4.19543207e-01 -8.54051173e-01
-1.22510917e-01 5.36755085e-01 2.33470663e-01 -1.01511240e-01
1.50492162e-01 5.88338256e-01 -1.45947647e+00 1.05578706e-01
9.16861117e-01 8.44972908e-01 2.51390368e-01 1.93485245e-01
2.65917510e-01 4.00457293e-01 -4.43329990e-01 2.18463806e-03
4.54859942e-01 -4.91663963e-01 -1.07929480e+00 2.25602046e-01
2.35754594e-01 -5.57592586e-02 -7.68166557e-02 7.77644575e-01
4.99041826e-01 5.95033646e-01 1.04918492e+00 -1.50865567e+00
-1.03965998e+00 2.17361376e-01 -3.72672379e-01 6.24594744e-03
2.01413050e-01 -1.31468758e-01 1.24517882e+00 -1.06486130e+00
6.15847051e-01 1.00875604e+00 7.30386436e-01 5.86004674e-01
-1.33681214e+00 -6.59286678e-01 1.68922380e-01 2.31485710e-01
-1.65411448e+00 -4.65126008e-01 8.43482494e-01 -4.87291873e-01
8.21401000e-01 1.80151194e-01 4.30605590e-01 1.00645995e+00
-2.04569787e-01 7.16713905e-01 8.69072437e-01 -5.13480842e-01
4.60496902e-01 3.26893926e-01 2.55504608e-01 3.74821186e-01
8.83685648e-02 1.13611832e-01 -3.62321615e-01 -4.09947604e-01
7.59367645e-01 2.69033432e-01 -1.21104971e-01 -8.97303283e-01
-1.19528568e+00 1.15852106e+00 5.59057117e-01 2.75878608e-01
-1.80721328e-01 -3.57495070e-01 5.71283326e-02 4.95840639e-01
5.53255498e-01 5.58700383e-01 -7.37172842e-01 1.31179821e-02
-6.97505772e-01 4.44268346e-01 7.12124288e-01 9.88789976e-01
1.07891297e+00 -4.33801502e-01 1.76785085e-02 1.23214602e+00
3.69371891e-01 2.87882626e-01 8.63811135e-01 -1.24265206e+00
4.12833951e-02 5.86694300e-01 1.95817798e-01 -8.15822542e-01
-3.93103927e-01 -2.69568145e-01 -5.96009433e-01 7.25810826e-02
3.39681834e-01 -1.40947476e-01 -6.23396218e-01 1.78093600e+00
5.84502637e-01 3.71143222e-01 6.08225316e-02 1.00849712e+00
4.22768980e-01 4.62263852e-01 -9.71292704e-02 -8.55394080e-02
9.05106485e-01 -1.11032224e+00 -2.24391699e-01 1.13584206e-01
8.32850277e-01 -7.48712063e-01 1.25423098e+00 1.90133691e-01
-6.86773539e-01 -5.48253596e-01 -9.45987463e-01 -2.01845039e-02
-3.68831217e-01 3.49654034e-02 7.04350829e-01 3.12004954e-01
-9.54949915e-01 6.48512840e-01 -9.73347545e-01 -6.41246438e-01
4.64678675e-01 5.03617585e-01 -4.10656333e-01 2.92075351e-02
-1.01618624e+00 7.60178864e-01 4.18600917e-01 -4.87972379e-01
-4.43085402e-01 -8.38131785e-01 -7.10193634e-01 -1.13565400e-01
1.23079754e-01 -7.19650745e-01 1.36526000e+00 -1.08848202e+00
-1.52773845e+00 7.28434563e-01 -3.65161479e-01 -2.88621694e-01
5.99269606e-02 -2.32379064e-01 -5.77684343e-01 -5.10842949e-02
3.21907759e-01 7.20739841e-01 5.73880553e-01 -1.11051512e+00
-7.11057246e-01 -2.26312011e-01 -2.31723607e-01 4.10932302e-01
-7.46549845e-01 -7.68313706e-02 -2.43585095e-01 -6.89153016e-01
6.89031929e-02 -9.13151205e-01 -2.72811949e-01 -3.66443954e-03
-2.09429130e-01 -4.46650237e-01 6.88043833e-01 -2.13603407e-01
1.42590356e+00 -2.35859537e+00 -4.43616067e-04 5.86286366e-01
1.37302265e-01 1.17837697e-01 -2.39013746e-01 4.35505897e-01
-6.61068410e-02 1.54775940e-02 -3.85572076e-01 -6.52646571e-02
9.08065066e-02 2.04964310e-01 -9.88932103e-02 4.58740711e-01
3.18355232e-01 7.06558049e-01 -9.21103299e-01 -4.27426159e-01
1.10977083e-01 3.58649909e-01 -7.80896246e-01 1.42877921e-01
2.45151129e-02 4.34173346e-01 -7.16975629e-01 4.46623892e-01
6.26537800e-01 -5.66521168e-01 4.22848724e-02 8.28210711e-02
4.88807447e-02 3.50164622e-01 -1.20728862e+00 1.71842802e+00
-1.87249973e-01 3.20936769e-01 -3.19751441e-01 -1.17988741e+00
9.76121426e-01 -3.02073564e-02 6.57945871e-01 -4.47811544e-01
-1.42586127e-01 2.10279316e-01 -4.34881216e-03 -2.19589323e-01
4.08704072e-01 5.12292162e-02 -6.06409833e-02 4.20711547e-01
2.30770409e-01 9.98478848e-03 2.55115181e-02 2.30403438e-01
9.32224751e-01 1.87923551e-01 6.02392733e-01 -4.78266776e-01
4.26503062e-01 5.78741021e-02 6.56258225e-01 5.90335667e-01
-3.51342469e-01 6.44870818e-01 5.33258207e-02 -3.16982836e-01
-1.10602331e+00 -1.20686579e+00 -5.80268264e-01 1.31204605e+00
5.49287684e-02 -4.97741133e-01 -5.52278042e-01 -9.29278910e-01
2.06939772e-01 5.20827770e-01 -7.18986332e-01 -1.08978376e-01
-2.97739863e-01 -5.21255612e-01 2.77084589e-01 5.60333669e-01
3.22781593e-01 -7.87791669e-01 -1.71386041e-02 -4.51492742e-02
-4.97404486e-02 -5.72792530e-01 -6.30030632e-01 2.26363197e-01
-9.40242350e-01 -8.81004810e-01 -8.06333780e-01 -1.08042848e+00
6.84905887e-01 5.10016620e-01 1.05046034e+00 -2.87571251e-02
4.20457078e-03 3.97933543e-01 -5.88159621e-01 -2.11934879e-01
-2.45016083e-01 3.58563095e-01 3.80641997e-01 -7.27538392e-02
9.10734415e-01 -7.86283910e-01 -7.00789094e-01 7.43098080e-01
-6.85415685e-01 -3.48712742e-01 4.15823013e-01 9.86670971e-01
8.49740565e-01 -2.14140251e-01 8.35655570e-01 -1.01101875e+00
6.76240802e-01 -9.03563678e-01 -3.38123500e-01 1.50183037e-01
-8.20916295e-01 5.33975624e-02 7.04718053e-01 -3.49046916e-01
-9.36482072e-01 3.50291401e-01 -7.76895601e-03 -4.16369557e-01
-5.97162247e-01 3.70644003e-01 -1.65083632e-01 3.47366855e-02
1.06552815e+00 2.01198339e-01 1.30418897e-01 -5.76288223e-01
5.21176279e-01 8.72571051e-01 3.14038932e-01 -6.62924111e-01
8.30689251e-01 5.12181222e-01 -2.77796537e-01 -6.60144985e-01
-8.15087736e-01 -9.75416005e-01 -1.00993061e+00 3.01121920e-01
4.20468092e-01 -1.10213363e+00 -2.23230630e-01 1.50599033e-01
-4.68057454e-01 -4.29487437e-01 -3.78514707e-01 5.11290193e-01
-6.96014464e-01 4.78556216e-01 -2.67731935e-01 -3.40607762e-01
-8.22421536e-02 -6.96773291e-01 8.32850456e-01 2.16300383e-01
-5.09113491e-01 -1.31316638e+00 4.46537286e-01 -3.92602272e-02
2.44409680e-01 8.74533132e-02 6.18283808e-01 -1.22341990e+00
-2.84486953e-02 -4.42412719e-02 1.56037003e-01 4.17725891e-01
5.04571140e-01 6.18518740e-02 -9.06079590e-01 -3.01597834e-01
-2.51575023e-01 -3.98759931e-01 7.92512655e-01 4.66181278e-01
1.14693177e+00 -3.18807125e-01 -5.01262128e-01 6.22453928e-01
1.16653597e+00 -5.25697246e-02 2.93589950e-01 5.34556985e-01
5.42242825e-01 3.91360730e-01 1.04251277e+00 8.01146448e-01
5.45581639e-01 9.23799813e-01 -1.38132825e-01 -1.60221010e-01
-7.33918250e-02 -1.21582791e-01 3.74075472e-01 8.46957505e-01
9.40370336e-02 6.52299374e-02 -1.02257586e+00 7.61789083e-01
-2.11422133e+00 -1.05895400e+00 -5.25434688e-02 2.33036828e+00
8.93342733e-01 -3.90271187e-01 6.35158062e-01 -9.18814689e-02
6.78052783e-01 -3.48645806e-01 -5.46944678e-01 -3.01692069e-01
-8.52975622e-02 1.61830008e-01 1.46373346e-01 4.64405447e-01
-1.40405381e+00 9.43468869e-01 6.26113510e+00 1.07732034e+00
-9.39486444e-01 1.33829907e-01 5.88014483e-01 -2.29981408e-01
-2.00536489e-01 -8.65196288e-02 -7.95505881e-01 5.85440814e-01
9.28810060e-01 -2.82967001e-01 4.99742866e-01 1.24777842e+00
1.76885352e-01 1.07979730e-01 -1.22138035e+00 7.58434176e-01
-1.37942359e-01 -1.09397590e+00 -8.08053613e-02 7.58322254e-02
1.01656997e+00 2.12509513e-01 1.48373291e-01 3.87236387e-01
5.27808607e-01 -8.39851022e-01 3.09180081e-01 2.25225151e-01
5.40388644e-01 -8.78939569e-01 5.13360560e-01 3.07137787e-01
-1.16871500e+00 -1.22158900e-01 -6.99646294e-01 -2.00402215e-01
-3.32851321e-01 5.10900080e-01 -8.74173284e-01 5.64174175e-01
8.32163393e-01 1.15421760e+00 -6.43258154e-01 1.25380933e+00
-1.15676105e-01 6.48591101e-01 -4.14598703e-01 6.97053298e-02
1.59960583e-01 -2.24463612e-01 2.39872530e-01 1.23564267e+00
3.56849045e-01 5.05142957e-02 4.07122642e-01 6.77706361e-01
3.10621560e-02 3.38355392e-01 -6.33092403e-01 2.64852703e-01
9.77060020e-01 1.17252517e+00 -4.66258109e-01 -1.99695468e-01
-6.46170557e-01 1.10621369e+00 6.47097349e-01 4.38826889e-01
-6.99204504e-01 -4.18815821e-01 1.00342786e+00 1.08668193e-01
5.56769788e-01 -1.00321293e-01 -3.65448147e-01 -1.24925101e+00
3.14826774e-03 -7.19175518e-01 6.53450668e-01 -3.84349406e-01
-1.89029419e+00 4.65095341e-01 8.30742270e-02 -1.78093374e+00
-2.89454252e-01 -5.46093762e-01 -5.37753880e-01 6.74445868e-01
-1.52590072e+00 -1.00628448e+00 -1.37081727e-01 1.15825057e+00
3.91973227e-01 -3.28420103e-01 1.07632482e+00 2.40083650e-01
-4.56046969e-01 9.38189626e-01 7.78544962e-01 1.50713205e-01
1.23199642e+00 -1.18775487e+00 2.11557657e-01 6.17306948e-01
2.05479026e-01 7.83245683e-01 3.43972415e-01 -4.40507352e-01
-9.32407439e-01 -1.29702497e+00 9.30107892e-01 -6.11133695e-01
7.47799993e-01 -2.64773577e-01 -1.02141607e+00 7.56306827e-01
-8.78086761e-02 1.49779171e-01 1.16725814e+00 5.06099164e-01
-4.25463051e-01 -1.02478124e-01 -1.27094483e+00 4.18437123e-01
1.05743945e+00 -2.58139044e-01 -6.74725175e-01 4.35093254e-01
4.65711236e-01 -1.07122950e-01 -1.28783154e+00 2.09994599e-01
4.71181542e-01 -8.83923173e-01 9.53554451e-01 -5.99189222e-01
3.22991937e-01 -3.47426564e-01 -2.76747227e-01 -1.50948536e+00
-8.98549914e-01 -3.87911618e-01 -1.16198473e-01 1.36153495e+00
6.19897068e-01 -8.24552655e-01 7.13213325e-01 6.68005764e-01
-1.65237963e-01 -7.20588148e-01 -8.05637360e-01 -1.23052394e+00
5.52972257e-01 -1.42204791e-01 7.39086866e-01 1.36437106e+00
3.45636904e-01 2.04290748e-01 -2.60918796e-01 2.71021515e-01
4.10036087e-01 2.08561897e-01 7.66417503e-01 -1.51291871e+00
-3.61580610e-01 -4.49690729e-01 -3.53082925e-01 -1.18034518e+00
2.14363948e-01 -1.06071413e+00 -2.93938220e-02 -1.06397808e+00
4.03846592e-01 -7.89485216e-01 -6.22921646e-01 7.03907490e-01
-2.07257226e-01 1.70839787e-01 6.79727197e-02 7.64102697e-01
-8.64191234e-01 5.89644372e-01 9.44072127e-01 2.59269029e-01
-4.36792821e-01 5.18169962e-02 -7.99111366e-01 5.79023361e-01
9.93630230e-01 -5.66027761e-01 -3.88318390e-01 -3.46283168e-02
-3.05852562e-01 -4.97939438e-01 9.98622999e-02 -9.25886452e-01
8.83810520e-02 -3.93976212e-01 6.05862677e-01 4.05605324e-02
2.65095234e-01 -8.72686863e-01 -3.14407870e-02 1.71443269e-01
-3.60258937e-01 -1.33404583e-01 -9.80514959e-02 5.03603578e-01
-4.00388956e-01 -1.52242452e-01 7.80129194e-01 7.50172287e-02
-1.08996427e+00 3.70421976e-01 -3.24066728e-01 7.30460733e-02
1.15065563e+00 -3.64549100e-01 -1.89426094e-01 -3.33042771e-01
-7.57001638e-01 3.71197879e-01 8.38214576e-01 3.77674252e-01
6.00128889e-01 -1.67442155e+00 -7.36373067e-01 2.89663047e-01
4.78409916e-01 -2.15357304e-01 -2.12945342e-02 8.01053286e-01
-1.27630904e-01 3.42041850e-01 -2.63441294e-01 -7.26165771e-01
-1.02414453e+00 5.92327297e-01 1.63271084e-01 -3.92540134e-02
-4.20368463e-01 9.70585465e-01 2.35435471e-01 -9.08217847e-01
4.42071110e-02 -5.21112978e-02 -9.45238173e-02 -2.67582417e-01
3.92859966e-01 3.83013904e-01 -1.33531272e-01 -6.54709041e-01
-5.42753637e-01 5.79177737e-01 -2.59400129e-01 9.31710452e-02
1.46125984e+00 -3.29349369e-01 2.07528040e-01 3.30712497e-01
1.26757097e+00 -4.23709536e-03 -1.34079707e+00 -4.44635928e-01
1.52916059e-01 -5.84561646e-01 -1.17452495e-01 -8.10190558e-01
-7.52684534e-01 5.94374716e-01 6.64972067e-01 7.81409442e-02
1.33958554e+00 2.49478146e-01 4.26279277e-01 4.89771426e-01
1.93116218e-01 -1.25056624e+00 9.33265686e-02 4.96198207e-01
7.12403238e-01 -1.44281232e+00 -5.75655177e-02 -3.52899253e-01
-1.04213762e+00 9.53303456e-01 5.96210837e-01 -4.50645030e-01
7.95759737e-01 -5.72446249e-02 7.49793276e-02 1.28403589e-01
-6.80182397e-01 -2.21302316e-01 4.40032899e-01 8.68692100e-01
7.13791907e-01 8.54043886e-02 -2.57065117e-01 7.75419235e-01
-1.78327486e-01 -6.52267635e-02 1.17474377e-01 7.81031847e-01
-4.38737273e-01 -1.45490527e+00 -1.94432259e-01 4.13858175e-01
-2.79683352e-01 -1.04212463e-01 -5.21361113e-01 8.92340779e-01
7.03634694e-02 9.50669348e-01 1.27030462e-01 -3.39204758e-01
1.37864590e-01 3.37674171e-01 2.00439319e-01 -7.26889729e-01
-3.66187841e-01 7.38022253e-02 -1.36539087e-01 -5.52637875e-01
-4.88419205e-01 -9.24673200e-01 -1.32122993e+00 -3.18711370e-01
-4.18598443e-01 4.04355317e-01 2.81793773e-01 7.05092788e-01
8.65092993e-01 -7.85107911e-02 9.80196834e-01 -6.46740437e-01
-5.79005182e-01 -9.15164590e-01 -7.08712220e-01 7.78315008e-01
1.29714936e-01 -7.66403139e-01 -2.06363425e-01 2.13203251e-01] | [10.2052001953125, 3.029291868209839] |
99b6045e-5f2f-40ba-9902-a08d5231739e | hybrid-neural-networks-for-on-device | 2112.05893 | null | https://arxiv.org/abs/2112.05893v1 | https://arxiv.org/pdf/2112.05893v1.pdf | Hybrid Neural Networks for On-device Directional Hearing | On-device directional hearing requires audio source separation from a given direction while achieving stringent human-imperceptible latency requirements. While neural nets can achieve significantly better performance than traditional beamformers, all existing models fall short of supporting low-latency causal inference on computationally-constrained wearables. We present DeepBeam, a hybrid model that combines traditional beamformers with a custom lightweight neural net. The former reduces the computational burden of the latter and also improves its generalizability, while the latter is designed to further reduce the memory and computational overhead to enable real-time and low-latency operations. Our evaluation shows comparable performance to state-of-the-art causal inference models on synthetic data while achieving a 5x reduction of model size, 4x reduction of computation per second, 5x reduction in processing time and generalizing better to real hardware data. Further, our real-time hybrid model runs in 8 ms on mobile CPUs designed for low-power wearable devices and achieves an end-to-end latency of 17.5 ms. | ['Shyamnath Gollakota', 'Hao Zhang', 'Maruchi Kim', 'Anran Wang'] | 2021-12-11 | hybrid-neural-networks-for-on-device-1 | https://directionalhearing.cs.washington.edu/ | https://arxiv.org/pdf/2112.05893 | aaai-2022-2 | ['directional-hearing', 'audio-source-separation', 'real-time-directional-hearing'] | ['audio', 'audio', 'audio'] | [ 3.72695446e-01 1.95502684e-01 -1.06663078e-01 -4.86112833e-01
-9.33818638e-01 -1.78336203e-01 1.57808408e-01 -2.87519712e-02
-5.44277370e-01 6.15106165e-01 9.28436697e-01 -7.34326065e-01
-2.40820870e-01 -7.46117234e-01 -6.18844807e-01 -3.37391794e-01
-4.88405645e-01 1.01756595e-01 3.27290595e-01 4.36475843e-01
-2.73106903e-01 -5.98222576e-02 -1.47836113e+00 3.00064534e-01
1.24183170e-01 1.27163184e+00 -3.08878049e-02 1.12186277e+00
7.08405554e-01 7.13585377e-01 -5.83471179e-01 8.55448246e-02
-1.69592753e-01 8.77235606e-02 -2.33811736e-01 -9.59453046e-01
7.41071999e-01 -7.58851767e-01 -6.08284414e-01 5.11826038e-01
1.39300025e+00 -1.94936901e-01 -7.73345754e-02 -1.07513762e+00
-2.04371512e-01 9.39260483e-01 -3.88119459e-01 3.78756046e-01
7.45707393e-01 -5.49622923e-02 9.64319825e-01 -6.09878302e-01
-6.78402781e-02 1.40198970e+00 1.11403406e+00 4.38858241e-01
-1.29825890e+00 -1.26120782e+00 1.17725685e-01 1.88476332e-02
-1.36350358e+00 -1.03708196e+00 4.32285011e-01 1.58001050e-01
1.20432341e+00 6.29965842e-01 5.93678355e-01 1.38524079e+00
1.78283602e-01 6.14799917e-01 8.72939527e-01 -1.55645028e-01
5.76401830e-01 -7.16388464e-01 7.71186501e-02 4.64278966e-01
1.78351149e-01 3.70070070e-01 -1.48168409e+00 -5.98840237e-01
6.36110961e-01 -3.67719382e-01 -4.16204929e-01 4.80349749e-01
-1.33869433e+00 2.59199709e-01 3.92953545e-01 -1.76001146e-01
-4.18317735e-01 1.04461884e+00 2.76820838e-01 -8.37821066e-02
3.79222304e-01 -1.21513605e-01 -4.76291925e-01 -6.27959907e-01
-1.25706756e+00 3.91525537e-01 9.17428136e-01 9.47528601e-01
-2.16254562e-01 4.23421681e-01 -1.30218595e-01 4.59533244e-01
6.16997242e-01 9.39452231e-01 1.04395159e-01 -9.31011021e-01
3.74486446e-01 -6.86830878e-02 2.90133851e-03 -9.40374494e-01
-9.12462652e-01 -7.32068360e-01 -9.62915659e-01 -9.28395987e-02
2.34859958e-01 -5.60618639e-01 -8.18912625e-01 2.02435040e+00
2.85574853e-01 5.14649451e-01 -4.37157303e-01 8.50890100e-01
7.54262388e-01 6.48787200e-01 1.86980650e-01 -3.45952332e-01
1.40859616e+00 -3.85425180e-01 -8.22330177e-01 -5.04432261e-01
-2.15504356e-02 -6.82692826e-01 1.06711757e+00 6.69615507e-01
-1.28918064e+00 -3.40775460e-01 -1.40840662e+00 -1.53893098e-01
4.07979786e-01 -4.62184921e-02 1.05436039e+00 1.12636173e+00
-1.08539224e+00 1.51917771e-01 -1.25746882e+00 -2.40014642e-02
4.36542332e-01 7.84075737e-01 8.17139521e-02 2.40169913e-01
-1.00589836e+00 8.90785232e-02 -1.11277819e-01 2.15023011e-01
-1.01966941e+00 -1.44385350e+00 -4.01384622e-01 1.07502550e-01
1.62408635e-01 -1.04759276e+00 1.58675754e+00 -2.11314157e-01
-1.43463922e+00 -4.19549905e-02 -3.46694082e-01 -5.86182535e-01
2.18914822e-01 -8.31143260e-01 -7.69967377e-01 -2.59769231e-01
-1.22680016e-01 5.47192812e-01 5.95934987e-01 -8.47534776e-01
-6.86445534e-01 -3.16318303e-01 -1.51043981e-01 -2.10557744e-01
-6.94996834e-01 -2.28644814e-02 -3.58518332e-01 -1.04095650e+00
1.98630810e-01 -1.03535950e+00 -6.46750107e-02 1.60905480e-01
-4.98660356e-01 1.18607432e-01 6.29148364e-01 -6.59395456e-01
1.48216748e+00 -2.08614349e+00 -4.05013055e-01 2.06065387e-01
3.74550492e-01 -8.31500068e-02 1.07456096e-01 1.88001215e-01
-5.43724485e-02 -3.93452086e-02 1.35834783e-01 -3.80304217e-01
-8.63793641e-02 7.61839077e-02 -4.93111014e-01 5.07310569e-01
-4.53562915e-01 6.00672960e-01 -8.06298316e-01 -1.43355250e-01
-8.61329958e-02 7.51045346e-01 -1.05747736e+00 1.78617045e-01
2.29556728e-02 1.05396993e-01 -1.38118342e-01 4.98836279e-01
4.98878270e-01 -4.59563807e-02 4.27549034e-01 -4.94862616e-01
-6.95039928e-02 9.86055374e-01 -1.58772850e+00 1.83399737e+00
-8.78116310e-01 7.95079350e-01 2.40465596e-01 -1.12407848e-01
3.83889526e-01 5.60579538e-01 3.70907068e-01 -7.96825647e-01
9.91761312e-02 -2.36251522e-02 1.07689708e-01 -2.35141814e-01
3.49388808e-01 1.00824468e-01 -4.49231602e-02 7.86359727e-01
-1.20622419e-01 4.07699436e-01 -4.30653363e-01 4.74072248e-02
1.59633636e+00 -1.05324231e-01 4.23566662e-02 -4.21008319e-01
-4.42300647e-01 -7.56249249e-01 5.58296323e-01 9.16111112e-01
1.39809042e-01 3.74911577e-01 -9.42772031e-02 -4.21742082e-01
-4.65299904e-01 -1.44328380e+00 3.38717736e-02 1.44226193e+00
-1.02185570e-01 -8.09519887e-01 -6.57737672e-01 1.63442999e-01
-1.01904452e-01 8.71934712e-01 -4.27286506e-01 5.84382452e-02
-5.45912147e-01 -9.18956995e-01 1.29129100e+00 1.03806269e+00
4.70755845e-01 -5.79925060e-01 -1.24725747e+00 4.40057278e-01
-3.49115461e-01 -9.27886724e-01 -2.52252012e-01 3.09015810e-01
-8.23772669e-01 -4.00450408e-01 -4.99705821e-02 -1.82392389e-01
1.28328592e-01 1.35466486e-01 1.05994093e+00 -2.29596838e-01
-1.69420555e-01 1.53607011e-01 1.69912785e-01 -8.62037539e-01
4.23275948e-01 -4.45737764e-02 6.46850705e-01 -2.49989644e-01
9.25441459e-02 -1.27264011e+00 -9.37047780e-01 1.58274144e-01
-5.18127561e-01 2.96091344e-02 4.38382208e-01 7.03527689e-01
3.23077559e-01 1.08198576e-01 4.20559764e-01 -3.01038742e-01
6.52906775e-01 -2.76581675e-01 -2.21928403e-01 -1.98452890e-01
-7.34616339e-01 1.16961166e-01 1.72591403e-01 -6.38536513e-01
-1.24233508e+00 -7.93265365e-03 -5.37345946e-01 1.75802350e-01
1.21361554e-01 4.86005068e-01 -1.14497371e-01 2.41370559e-01
8.01208496e-01 -2.20581457e-01 -4.06857282e-01 -4.79701161e-01
4.09045637e-01 6.96748734e-01 9.01244700e-01 -4.20447797e-01
4.27089870e-01 6.22617662e-01 -1.21935509e-01 -7.04284668e-01
-6.09950840e-01 -2.65959278e-02 -8.15557595e-03 -1.21978633e-01
6.06531382e-01 -1.34660840e+00 -1.40605950e+00 3.12977374e-01
-1.19291472e+00 -5.48033297e-01 1.24421798e-01 5.75987160e-01
-2.91122466e-01 -3.11943382e-01 -6.06444597e-01 -1.10476613e+00
-8.32377136e-01 -6.54347122e-01 1.26749229e+00 -6.48733303e-02
-9.53876913e-01 -5.83577275e-01 9.95754614e-04 1.52964056e-01
4.70047295e-01 -9.77245122e-02 6.36556685e-01 9.25161391e-02
-3.85085553e-01 -2.26384655e-01 -1.34737357e-01 -1.87950388e-01
-7.57130831e-02 -4.46960509e-01 -1.47350466e+00 -3.06779355e-01
-6.43675625e-02 -1.15001887e-01 5.62750995e-01 7.73348451e-01
1.15283537e+00 -5.10473251e-01 -6.08408332e-01 5.92539907e-01
9.72050488e-01 1.59705356e-01 6.21815085e-01 -2.17328072e-01
6.56515360e-01 -2.64756642e-02 2.46163577e-01 5.79412043e-01
5.11817157e-01 7.95131981e-01 4.00431395e-01 -3.17636490e-01
-5.28106689e-01 -5.43336451e-01 3.22371155e-01 9.11041677e-01
-1.92744583e-01 -4.96860534e-01 -9.47135806e-01 5.18516958e-01
-1.83776295e+00 -8.84579301e-01 -2.61741012e-01 2.28922129e+00
1.04672420e+00 4.75643933e-01 1.73493728e-01 4.64844555e-01
3.67122851e-02 1.00102037e-01 -4.38922524e-01 -2.22542688e-01
3.33553553e-01 5.55879772e-01 7.80123234e-01 8.40100884e-01
-8.85783136e-01 2.73361981e-01 7.19925499e+00 7.16344833e-01
-1.20265770e+00 5.10368884e-01 3.52498174e-01 -1.04351699e+00
-2.43737727e-01 -3.98633540e-01 -6.88232780e-01 3.37447047e-01
1.46414232e+00 1.75394014e-01 3.68444949e-01 8.45595360e-01
4.37172711e-01 -1.23012193e-01 -1.25320065e+00 1.20202506e+00
-1.58249155e-01 -1.26774585e+00 -3.64807576e-01 1.47768959e-01
2.06139028e-01 1.38238132e-01 2.53980309e-02 -1.97964281e-01
3.28244954e-01 -9.93879676e-01 1.12604380e+00 2.46322528e-01
1.04419041e+00 -8.33116472e-01 4.44398582e-01 2.17584986e-02
-1.12843943e+00 -2.20572114e-01 3.29808854e-02 -8.56908917e-01
5.57295442e-01 1.06046677e+00 -8.41587365e-01 -4.03143233e-03
1.29771984e+00 -1.52263530e-02 -2.92958677e-01 6.99275851e-01
-2.75773048e-01 1.25977099e+00 -8.36456180e-01 -4.07842577e-01
-3.26474488e-01 9.85723078e-01 5.72694242e-01 1.37739611e+00
2.72415698e-01 1.86665729e-01 -1.54925406e-01 3.77895862e-01
1.10229403e-01 -6.20915174e-01 -2.25130215e-01 5.48331320e-01
1.03908229e+00 7.47320294e-01 -3.31298620e-01 -6.70284554e-02
-1.31928064e-02 6.04393601e-01 -2.16370389e-01 1.74875721e-01
-1.12550020e+00 -4.06033903e-01 8.68003488e-01 2.45118544e-01
-7.03888759e-02 -4.52533811e-01 -9.77747917e-01 -4.62850392e-01
1.11801386e-01 -8.12027633e-01 2.23284755e-02 -8.78093958e-01
-8.47257495e-01 5.19352019e-01 -2.60307156e-02 -7.65356302e-01
-2.13031277e-01 -2.95486808e-01 -2.61352986e-01 7.95704246e-01
-6.17380261e-01 -1.22927999e+00 -4.43105221e-01 6.77348316e-01
2.89886624e-01 3.65330011e-01 1.30549669e+00 8.35042417e-01
-2.52320349e-01 1.12209427e+00 -4.40968424e-01 -2.40961581e-01
7.35463679e-01 -1.01504576e+00 8.70412707e-01 1.00829184e+00
5.56105934e-02 1.20017183e+00 1.09951544e+00 -6.42269135e-01
-1.65910721e+00 -9.48357046e-01 1.02978110e+00 -4.86308545e-01
6.87499583e-01 -9.23369169e-01 -3.32686424e-01 5.54148912e-01
3.12671155e-01 -1.29551098e-01 9.61337864e-01 1.05248618e+00
-6.44547403e-01 -5.73911905e-01 -6.66668296e-01 1.01270783e+00
1.47965372e+00 -5.96285999e-01 -2.75878072e-01 1.38130471e-01
8.37769806e-01 -5.11049449e-01 -8.82453322e-01 5.38437605e-01
1.29019034e+00 -7.44201660e-01 1.17308199e+00 -1.14796333e-01
1.80407420e-01 -3.58695298e-01 -2.60502517e-01 -8.35754275e-01
-4.62219417e-01 -1.10663772e+00 -5.03606796e-01 1.12913060e+00
4.95948732e-01 -3.67496669e-01 7.92541921e-01 6.71177626e-01
-1.58812016e-01 -7.54169822e-01 -1.33224678e+00 -5.58856726e-01
-6.14032388e-01 -1.57553101e+00 4.47029054e-01 3.73223126e-01
6.15204424e-02 6.66466534e-01 -7.25158095e-01 5.57518721e-01
8.38816226e-01 -4.10968155e-01 6.35052085e-01 -9.24363136e-01
-6.16806388e-01 -1.04851380e-01 -3.35527033e-01 -1.49281752e+00
-4.32022601e-01 -2.19425544e-01 2.81642556e-01 -1.27955604e+00
-1.22567862e-01 -6.67086244e-01 -3.12638581e-01 9.33713138e-01
4.71170945e-03 6.62250876e-01 -8.08317363e-02 -3.97463977e-01
-2.32913792e-01 2.67619461e-01 3.63968998e-01 -1.80160493e-01
-2.80658931e-01 5.38287051e-02 -8.42674196e-01 9.43333805e-01
4.87098247e-01 -6.26746774e-01 -8.61282885e-01 -9.18133020e-01
7.57246971e-01 1.22410588e-01 7.20688641e-01 -1.50248146e+00
8.39621782e-01 1.64533362e-01 4.03033257e-01 -5.46022654e-01
5.76765418e-01 -5.87765574e-01 4.09621030e-01 5.33344626e-01
-3.69442672e-01 2.84709521e-02 6.03978515e-01 6.70541406e-01
2.78009802e-01 5.83959222e-01 2.17972636e-01 5.25208116e-01
-1.49857566e-01 -3.16359289e-02 -5.05303860e-01 -2.97487170e-01
2.94893652e-01 7.27515493e-04 -3.30122441e-01 -7.09213495e-01
-5.47478676e-01 -3.06018621e-01 -4.39982831e-01 6.30933702e-01
5.50271451e-01 -1.31233525e+00 -4.23864782e-01 -2.03435402e-02
-2.27671102e-01 -8.35321844e-02 3.00382644e-01 5.93781590e-01
-2.44310588e-01 4.81756300e-01 1.53215513e-01 -7.03997314e-01
-1.56743073e+00 2.18497172e-01 1.43934600e-02 1.60063133e-01
-5.15021205e-01 1.50195312e+00 7.58059174e-02 1.06213465e-01
7.31531799e-01 -5.67726076e-01 6.46809578e-01 -2.32514739e-01
9.45186079e-01 7.17311144e-01 3.35227966e-01 1.62283361e-01
-7.47694850e-01 1.44109294e-01 3.54601234e-01 -6.17874384e-01
1.22460890e+00 5.40509000e-02 -8.07732809e-03 4.71996695e-01
8.14022839e-01 4.44166124e-01 -1.21292913e+00 7.88666904e-02
-3.28054070e-01 -1.33543253e-01 7.53583610e-01 -1.28030777e+00
-8.85087132e-01 6.76344514e-01 1.27788734e+00 -6.55819848e-02
1.48064983e+00 -2.43580744e-01 9.81537104e-01 2.63345897e-01
6.91537797e-01 -7.34429896e-01 -9.95213985e-02 9.42348391e-02
8.54690135e-01 -4.95979458e-01 3.44608009e-01 -3.62160653e-01
1.13298878e-01 7.01428056e-01 2.90795743e-01 1.21837400e-01
1.04780114e+00 1.31044579e+00 -5.64244203e-02 -3.58472206e-02
-9.39619601e-01 3.53999376e-01 2.86037624e-01 6.82035983e-01
6.00689113e-01 4.23312217e-01 -1.63180336e-01 8.40317726e-01
-6.35621011e-01 2.25689918e-01 -4.13172320e-03 8.26897144e-01
-6.32173717e-02 -7.90475607e-01 -3.19704086e-01 3.86687219e-01
-6.31002605e-01 -5.43230951e-01 1.69870630e-02 3.25403243e-01
1.87766597e-01 1.61096478e+00 3.24788362e-01 -8.69278491e-01
4.10239190e-01 -4.25079554e-01 4.39791918e-01 -2.56832778e-01
-6.58954501e-01 3.79110843e-01 5.35235107e-01 -9.34363425e-01
-2.33204931e-01 -5.43107986e-01 -1.06596577e+00 -6.63129568e-01
-1.64430007e-01 -4.72050488e-01 8.74216974e-01 6.69797003e-01
7.76090562e-01 1.17648506e+00 3.31618749e-02 -9.56320643e-01
-2.10002512e-01 -1.11097395e+00 -6.92590401e-02 -3.17951113e-01
5.36338508e-01 -5.97427070e-01 5.83155453e-02 1.19468130e-01] | [14.98604679107666, 5.670225143432617] |
1f6a17aa-c4c7-432d-b52f-7eefd101518c | pix3d-dataset-and-methods-for-single-image-3d | 1804.04610 | null | http://arxiv.org/abs/1804.04610v1 | http://arxiv.org/pdf/1804.04610v1.pdf | Pix3D: Dataset and Methods for Single-Image 3D Shape Modeling | We study 3D shape modeling from a single image and make contributions to it
in three aspects. First, we present Pix3D, a large-scale benchmark of diverse
image-shape pairs with pixel-level 2D-3D alignment. Pix3D has wide applications
in shape-related tasks including reconstruction, retrieval, viewpoint
estimation, etc. Building such a large-scale dataset, however, is highly
challenging; existing datasets either contain only synthetic data, or lack
precise alignment between 2D images and 3D shapes, or only have a small number
of images. Second, we calibrate the evaluation criteria for 3D shape
reconstruction through behavioral studies, and use them to objectively and
systematically benchmark cutting-edge reconstruction algorithms on Pix3D.
Third, we design a novel model that simultaneously performs 3D reconstruction
and pose estimation; our multi-task learning approach achieves state-of-the-art
performance on both tasks. | ['Chengkai Zhang', 'Zhoutong Zhang', 'Xiuming Zhang', 'Jiajun Wu', 'Xingyuan Sun', 'William T. Freeman', 'Tianfan Xue', 'Joshua B. Tenenbaum'] | 2018-04-12 | pix3d-dataset-and-methods-for-single-image-3d-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Sun_Pix3D_Dataset_and_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Sun_Pix3D_Dataset_and_CVPR_2018_paper.pdf | cvpr-2018-6 | ['3d-shape-modeling', 'viewpoint-estimation'] | ['computer-vision', 'computer-vision'] | [-2.92134136e-02 -3.47652704e-01 -3.96118641e-01 -3.41387182e-01
-1.15529013e+00 -6.74126506e-01 4.76443708e-01 -3.49740416e-01
1.47986999e-02 2.10079432e-01 4.08390582e-01 -9.74866897e-02
1.03101104e-01 -4.66074049e-01 -1.02803838e+00 -2.60858297e-01
5.01926541e-01 1.01309919e+00 2.28921905e-01 2.86069023e-03
2.45260313e-01 6.25815570e-01 -1.32379138e+00 2.38855347e-01
4.67935622e-01 1.19205999e+00 2.59647965e-01 2.63748318e-01
6.77968264e-02 -2.59806901e-01 7.95131549e-02 -4.44678247e-01
6.72870994e-01 1.35770082e-01 -4.96077359e-01 3.26818913e-01
9.54005361e-01 -5.76468825e-01 -1.05329052e-01 8.29753518e-01
9.20614779e-01 -4.65164840e-01 8.90067995e-01 -1.22508979e+00
-8.36194038e-01 -7.42689967e-02 -1.01110637e+00 -2.43585795e-01
2.72614777e-01 3.64628762e-01 8.17918599e-01 -1.20791662e+00
9.95762885e-01 1.44013333e+00 8.83820236e-01 4.54909861e-01
-1.74076867e+00 -8.53168190e-01 7.31714219e-02 -3.11104953e-01
-1.24534953e+00 -5.53808272e-01 1.07508826e+00 -5.94734907e-01
7.87993252e-01 -3.78926168e-03 9.45422351e-01 1.38359284e+00
1.03979763e-02 1.04017830e+00 1.40375006e+00 3.02773784e-03
-8.98753107e-02 -3.26228917e-01 -1.72396675e-01 6.21272147e-01
3.25549453e-01 4.40843552e-01 -4.62085187e-01 6.79380372e-02
1.46549654e+00 -3.75323385e-01 -9.04478482e-04 -1.06755161e+00
-1.33938944e+00 2.75529981e-01 2.28829905e-01 -5.75637631e-02
-3.11715871e-01 3.80554289e-01 3.02159280e-01 2.43192434e-01
8.15122366e-01 2.20668465e-01 -6.50081217e-01 1.32790534e-02
-6.74358666e-01 6.28585041e-01 2.17601955e-01 1.32518768e+00
5.42559862e-01 7.39429221e-02 -2.11325094e-01 1.01496100e+00
3.50310832e-01 9.57406461e-01 1.44165292e-01 -1.14435601e+00
5.60667992e-01 3.77493143e-01 1.83401912e-01 -1.18911242e+00
-6.17307544e-01 -4.91021961e-01 -7.26094484e-01 2.82328665e-01
5.08656323e-01 2.48179987e-01 -6.97586536e-01 1.82877958e+00
5.00881970e-01 2.22505867e-01 -5.88788033e-01 1.03403473e+00
1.07126749e+00 1.47286355e-01 -2.41062224e-01 1.48258880e-01
1.20047259e+00 -7.56440997e-01 -1.44278720e-01 -3.96799505e-01
2.78916091e-01 -8.68834615e-01 1.26386750e+00 2.33376101e-01
-1.44016504e+00 -5.63264906e-01 -7.13715553e-01 -2.87246019e-01
-5.48910722e-02 3.81283849e-01 6.67602062e-01 5.45400262e-01
-1.00631440e+00 2.79986382e-01 -5.23513973e-01 -1.26453713e-01
8.88776481e-01 2.07443699e-01 -5.41340232e-01 -3.40670720e-02
-3.90429944e-01 7.80147016e-01 -3.00736219e-01 -2.75329471e-01
-9.62009668e-01 -1.08739316e+00 -9.15892899e-01 -4.58293825e-01
1.74381480e-01 -1.18366766e+00 1.21449673e+00 -4.31430668e-01
-1.15126884e+00 1.76557457e+00 -1.30165666e-01 -4.71221022e-02
5.18438876e-01 -1.23505965e-01 1.23695314e-01 -1.68379143e-01
1.71460986e-01 1.10980844e+00 9.01636362e-01 -1.82721043e+00
3.59213911e-02 -9.79107857e-01 -3.30565095e-01 2.01934204e-01
4.76008952e-02 -2.61380881e-01 -8.89612675e-01 -9.20127690e-01
4.82656747e-01 -9.27952886e-01 -6.21104538e-02 6.14036679e-01
-4.61514592e-01 -9.78983939e-03 5.82231224e-01 -5.30149758e-01
6.30834460e-01 -1.93956459e+00 2.40617961e-01 1.06214862e-02
3.87152076e-01 1.84892341e-01 -5.86327016e-01 -9.36767235e-02
-5.76320179e-02 1.73500791e-01 -7.13996822e-03 -8.97844195e-01
1.10147484e-01 -4.24753688e-02 -3.19382727e-01 3.86769384e-01
-3.89390625e-02 1.53952599e+00 -5.73522210e-01 -5.25220096e-01
2.49802291e-01 2.45728537e-01 -7.18362510e-01 -4.82305624e-02
-2.97705621e-01 6.62543774e-01 -5.15860915e-01 9.52410758e-01
9.44468439e-01 -3.08014810e-01 -2.95301765e-01 -8.78622949e-01
7.21632242e-02 -3.63295898e-02 -9.49670136e-01 2.16506124e+00
-4.95440066e-01 3.86510998e-01 1.27494395e-01 -8.66030455e-01
7.52083421e-01 5.42885484e-03 1.00969160e+00 -9.81422365e-01
2.41589040e-01 3.16362679e-01 -3.51696491e-01 -3.64563882e-01
1.23973481e-01 3.05122565e-02 -2.02118121e-02 7.05257058e-01
-1.49377614e-01 -9.37425911e-01 -5.12858510e-01 -3.58791739e-01
5.58508217e-01 6.68069184e-01 5.34140319e-02 -1.72638386e-01
-9.85847227e-03 -1.79704428e-01 5.92064500e-01 3.01401258e-01
-3.49326283e-01 9.98796761e-01 2.57984102e-01 -4.74176258e-01
-1.49062991e+00 -1.34098780e+00 -4.07993942e-01 4.91408437e-01
4.48747903e-01 -2.09703326e-01 -5.93362272e-01 -5.23084998e-01
3.92897874e-01 3.86653662e-01 -4.43779349e-01 -5.56950644e-02
-3.78448159e-01 -4.39916760e-01 3.59585971e-01 5.81277192e-01
5.31703293e-01 -8.53900909e-01 -2.38377452e-01 -1.04309268e-01
-1.66910127e-01 -1.24471903e+00 -1.02529943e+00 -4.17211592e-01
-1.17859697e+00 -1.10849261e+00 -7.23376572e-01 -9.19531763e-01
7.04632580e-01 7.11356282e-01 1.49740899e+00 -1.07055932e-01
-3.24930310e-01 7.14622319e-01 5.76375574e-02 -4.91720259e-01
-9.61753726e-02 -3.09818953e-01 1.24860600e-01 -2.79911906e-01
5.44909835e-02 -9.99625325e-01 -5.78852773e-01 8.92253697e-01
-4.49425995e-01 5.51995695e-01 7.43131399e-01 7.06915855e-01
1.34573090e+00 -3.35405797e-01 6.66166604e-01 -7.45653093e-01
4.73566264e-01 1.21687293e-01 -7.87533343e-01 1.76842019e-01
-5.15190721e-01 -6.43135235e-02 -2.63980217e-02 -4.31178778e-01
-8.66477251e-01 3.25117469e-01 -2.70311862e-01 -8.13421965e-01
-1.44767731e-01 -1.54145239e-02 -3.87068063e-01 -3.68972123e-01
5.55806279e-01 3.63381207e-01 1.96710721e-01 -6.78382218e-01
4.33528870e-01 2.93640167e-01 6.14773810e-01 -9.30693269e-01
9.67988372e-01 5.66761613e-01 2.26981938e-01 -6.28175795e-01
-8.41093719e-01 -3.41941237e-01 -8.79176497e-01 -3.87824535e-01
6.72386110e-01 -1.08451915e+00 -8.20255280e-01 9.00434375e-01
-1.19379508e+00 -5.81673622e-01 -9.38239098e-02 3.02972466e-01
-1.12680769e+00 1.76411673e-01 -4.00419474e-01 -4.28843617e-01
-4.63329047e-01 -1.39024806e+00 1.88588047e+00 -3.50861371e-01
-9.31267440e-02 -7.40481198e-01 1.39446976e-02 8.47917378e-01
2.61733800e-01 2.41020083e-01 1.14741480e+00 -5.66124171e-02
-8.24341059e-01 -9.20526758e-02 -5.67067981e-01 1.00136315e-03
1.61755998e-02 -3.31384450e-01 -8.48733664e-01 -2.80417204e-01
-1.51816860e-01 -8.08409750e-01 6.28317535e-01 8.24850142e-01
1.75212157e+00 9.72859487e-02 -4.58618015e-01 9.11302686e-01
1.13944805e+00 -2.14875907e-01 5.76723635e-01 -1.13426231e-01
8.58332515e-01 4.30173874e-01 5.11310995e-01 2.44446471e-01
6.96013808e-01 1.15349090e+00 5.30770719e-01 -6.28716648e-02
-7.33552575e-01 -7.24446416e-01 -1.48135787e-02 8.89287770e-01
-8.55710059e-02 1.73664585e-01 -6.84576511e-01 4.56845522e-01
-1.56483579e+00 -6.23477221e-01 -1.09937951e-01 2.35117197e+00
9.11008596e-01 1.31907612e-01 3.35295826e-01 -3.38260472e-01
3.76551658e-01 5.84895276e-02 -9.76119280e-01 3.78929436e-01
-2.93299854e-01 4.84994575e-02 4.24981415e-01 3.49829108e-01
-1.15058959e+00 9.45601046e-01 6.97704840e+00 1.03218043e+00
-1.03286850e+00 1.54109241e-03 8.08841646e-01 -8.38890448e-02
-7.67029643e-01 -2.23883405e-01 -7.14722514e-01 2.39161327e-01
-3.73862162e-02 5.38147502e-02 4.60063636e-01 8.17402422e-01
2.23075211e-01 5.48017621e-02 -1.33238339e+00 1.64582145e+00
2.41723582e-01 -1.49062967e+00 2.38311663e-01 3.46303612e-01
9.55412030e-01 1.78052783e-01 3.20283741e-01 -3.94121483e-02
4.25740853e-02 -1.05713761e+00 1.00270367e+00 4.68876511e-01
1.14077592e+00 -3.91581088e-01 5.43204620e-02 6.07495308e-01
-1.21577704e+00 3.74877959e-01 -3.17228794e-01 4.02874410e-01
2.70022154e-01 5.47836125e-01 -2.97668487e-01 3.02678168e-01
5.64925909e-01 9.92171168e-01 -4.92722631e-01 1.26715493e+00
1.31863624e-01 7.55023211e-02 -2.05458865e-01 3.74953479e-01
-3.40358973e-01 -3.78610730e-01 7.30561376e-01 5.72090864e-01
3.02525699e-01 2.70671900e-02 3.64968538e-01 1.28627300e+00
-3.65654349e-01 5.92995398e-02 -9.27773297e-01 2.14647561e-01
5.55037498e-01 1.12954962e+00 -5.24567366e-01 1.10333465e-01
-3.86361361e-01 7.68557966e-01 2.87868470e-01 6.11415505e-02
-6.56150222e-01 5.95613003e-01 8.49074244e-01 4.64267045e-01
1.09584644e-01 -5.17169833e-01 -8.14352572e-01 -9.87387598e-01
9.47000235e-02 -7.24234521e-01 -2.04061985e-01 -1.39405572e+00
-1.60055757e+00 1.21385485e-01 2.14336991e-01 -1.51791871e+00
1.07025757e-01 -7.66845524e-01 -3.15874428e-01 5.63866258e-01
-1.41444075e+00 -1.57260275e+00 -2.13542148e-01 3.72934639e-01
6.52686477e-01 -1.79263890e-01 7.52926767e-01 4.70434546e-01
-3.69358659e-01 7.59420931e-01 -2.84473449e-01 1.21730745e-01
7.67499208e-01 -8.45692694e-01 8.85065734e-01 2.68504739e-01
4.62051958e-01 2.72745460e-01 2.19899565e-01 -7.72594571e-01
-1.97696888e+00 -1.10721540e+00 3.60831916e-01 -9.80499566e-01
1.38975576e-01 -5.31628788e-01 -5.63567936e-01 7.17394888e-01
-3.08003008e-01 2.90782809e-01 3.14660013e-01 1.73211414e-02
-5.42902231e-01 -1.45269141e-01 -1.16695166e+00 8.44263434e-01
1.79226184e+00 -5.32077253e-01 -3.16875249e-01 3.87448937e-01
6.98288202e-01 -8.03979039e-01 -9.50166762e-01 5.73001862e-01
9.19111907e-01 -8.82105350e-01 1.68637621e+00 -6.08403921e-01
5.56944311e-01 -1.39379188e-01 -3.06902915e-01 -1.41399097e+00
-4.08718169e-01 -2.29381382e-01 -2.78891120e-02 1.01880050e+00
3.92698556e-01 -2.97918320e-01 1.12909615e+00 3.59415382e-01
-3.13698053e-01 -1.06605327e+00 -8.70844960e-01 -9.70987082e-01
2.31830329e-01 -8.38363051e-01 5.87373614e-01 6.40747547e-01
-6.42749667e-01 3.31791878e-01 -4.85201955e-01 -1.94650918e-01
9.93525922e-01 7.90624380e-01 1.25107431e+00 -1.37201679e+00
-1.02078311e-01 -8.06208789e-01 -2.66104043e-01 -1.70974982e+00
3.96234632e-01 -1.12541783e+00 -4.61050421e-02 -1.33019841e+00
4.08989280e-01 -7.50748813e-01 4.06831831e-01 4.56823409e-01
2.28735983e-01 4.59926665e-01 1.90511458e-02 2.89949000e-01
-3.07816893e-01 8.51686358e-01 2.10508323e+00 -2.56422073e-01
-4.05363068e-02 1.56310603e-01 -7.21027911e-01 8.07761312e-01
3.08338493e-01 -1.56907842e-01 -6.65953219e-01 -7.81880736e-01
2.13456988e-01 1.26804724e-01 7.40443170e-01 -4.90025163e-01
-2.62816977e-02 -3.53101611e-01 6.93563342e-01 -1.05628192e+00
7.48684824e-01 -1.05152702e+00 2.03399003e-01 8.95429626e-02
-1.58188462e-01 -8.30955047e-04 2.73433715e-01 6.28419995e-01
2.74969101e-01 2.70413190e-01 8.16423118e-01 -2.57881969e-01
-6.73254251e-01 9.96485412e-01 4.28871512e-01 4.27317053e-01
8.66718650e-01 -4.85114723e-01 -7.32453987e-02 -3.44794542e-01
-3.10022563e-01 2.69648045e-01 8.15210164e-01 6.37556016e-01
8.60243201e-01 -1.99852002e+00 -7.79131234e-01 3.68211269e-01
4.12343323e-01 2.09083155e-01 2.00611383e-01 9.59298134e-01
-2.14636281e-01 2.75732160e-01 -4.66726691e-01 -9.57082808e-01
-1.19649255e+00 5.36738336e-01 3.75657141e-01 7.89292231e-02
-7.91150928e-01 6.93268001e-01 5.03417969e-01 -9.74144757e-01
1.54638603e-01 -4.90421563e-01 1.11557484e-01 -1.97617799e-01
9.24232677e-02 3.51317115e-02 8.76643695e-03 -8.93499613e-01
-2.50020325e-01 1.55679548e+00 4.04960603e-01 -4.90715615e-02
1.60851204e+00 -1.12995185e-01 2.69423813e-01 2.95184255e-01
1.20524168e+00 -5.93919195e-02 -1.50636005e+00 -4.80582833e-01
-4.43847418e-01 -8.62258852e-01 1.24088958e-01 -7.46771932e-01
-1.38449490e+00 7.23426878e-01 4.36060578e-01 -4.83135700e-01
1.03051281e+00 3.05186629e-01 1.06088006e+00 2.65778899e-01
8.04127276e-01 -1.03246319e+00 4.70591903e-01 3.07383895e-01
1.46213377e+00 -1.43641150e+00 8.30892920e-02 -7.71671832e-01
-5.03125846e-01 6.29455805e-01 8.00456226e-01 -1.26339644e-01
9.07073081e-01 2.41447940e-01 -2.45350897e-01 -3.77971917e-01
-5.99370718e-01 -1.14049032e-01 7.70644665e-01 9.42170143e-01
9.54132825e-02 6.06775247e-02 5.20202331e-03 7.63205647e-01
-1.89017355e-01 -9.20587927e-02 -3.90298888e-02 3.47799629e-01
-1.56502843e-01 -1.19091988e+00 -3.42448324e-01 4.68492955e-01
2.14156434e-01 1.91765204e-01 -5.28721273e-01 6.46679699e-01
-1.09801978e-01 3.26447755e-01 -4.74208593e-02 -4.45344657e-01
6.31367683e-01 -1.62684247e-01 1.06270301e+00 -4.22937900e-01
-1.19114593e-01 2.64386475e-01 4.20572832e-02 -7.46186435e-01
-3.77138078e-01 -8.85303915e-01 -8.95982862e-01 -2.21907809e-01
-5.84289730e-02 -8.96723509e-01 7.81082749e-01 7.99724221e-01
6.69586599e-01 4.70312387e-02 6.27370059e-01 -1.33417594e+00
-5.07219553e-01 -5.56988597e-01 -4.38386202e-01 6.72802687e-01
5.72788231e-02 -1.13423753e+00 1.93532854e-01 -4.66735773e-02] | [8.506453514099121, -3.1909584999084473] |
162712d2-3593-4544-8181-bef7ba092c81 | compressing-sentence-representation-via | null | null | https://openreview.net/forum?id=n3cvM4Phez9 | https://openreview.net/pdf?id=n3cvM4Phez9 | Compressing Sentence Representation via Homomorphic Projective Distillation | How to learn highly compact yet effective sentence representation? Pre-trained language models have been effective in many NLP tasks. However, these models are often huge and produce large sentence embeddings. Moreover, there is a big performance gap between large and small models. In this paper, we propose Homomorphic Projective Distillation (HPD) to learn compressed sentence embeddings. Our method augments a small Transformer encoder model with learnable projection layers to produce compact representations while mimicking a large pre-trained language model to retain the sentence representation quality. We evaluate our method with different model sizes on both semantic textual similarity (STS) and semantic retrieval (SR) tasks. Experiments show that our method achieves 2.7-4.5 points performance gain on STS tasks compared with previous best representations of the same size. In SR tasks, our method improves retrieval speed (8.2×) and memory usage (8.0×) compared with state-of-the-art large models. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['semantic-retrieval'] | ['natural-language-processing'] | [ 1.62698194e-01 1.00625359e-01 -2.46473074e-01 -1.41782224e-01
-1.49082780e+00 -2.42292419e-01 7.61150181e-01 3.62844050e-01
-6.52118623e-01 5.42730033e-01 7.71447837e-01 -2.84831643e-01
2.21899033e-01 -8.37854266e-01 -6.80159450e-01 -2.48784825e-01
7.84529746e-02 4.45660591e-01 6.22920506e-02 -3.48919481e-01
3.66582304e-01 1.93138823e-01 -1.13016903e+00 5.53446352e-01
7.72609830e-01 8.30462992e-01 5.59180617e-01 6.60331368e-01
-2.89617330e-01 6.04798615e-01 -5.51984668e-01 -4.27008688e-01
3.75895888e-01 -1.36380931e-02 -9.28203702e-01 -4.19403911e-01
6.34767234e-01 -6.00503862e-01 -9.90935028e-01 9.25875008e-01
7.55472839e-01 9.59722325e-02 6.15835011e-01 -6.59104884e-01
-1.36805916e+00 6.83683753e-01 -6.38166189e-01 2.35175848e-01
2.49875411e-01 -2.92887956e-01 1.32965147e+00 -1.19030952e+00
5.40507317e-01 1.37721944e+00 5.24372458e-01 5.07705867e-01
-1.12652731e+00 -5.56195557e-01 -2.66345888e-01 1.13762602e-01
-1.48929918e+00 -5.06634891e-01 3.79799128e-01 2.03562558e-01
1.63774407e+00 1.94880471e-01 2.72513002e-01 1.09459901e+00
4.08458740e-01 8.54672432e-01 6.95460021e-01 -5.62207460e-01
5.71145229e-02 4.27780777e-01 1.28260925e-01 6.74115896e-01
4.52869833e-01 -3.61269087e-01 -3.86337280e-01 -2.91387260e-01
5.92053831e-01 4.95774090e-01 -1.32631212e-01 -1.52214691e-01
-9.55184460e-01 1.15108502e+00 5.55996776e-01 3.62689614e-01
-2.58013278e-01 3.72833163e-01 7.52228916e-01 5.43050051e-01
6.03079915e-01 7.47489750e-01 -3.19348544e-01 -2.25821286e-01
-9.47542608e-01 1.85373560e-01 7.42292762e-01 9.90232527e-01
5.15527725e-01 -2.04611674e-01 -4.18100059e-01 1.20533979e+00
4.22400273e-02 6.99036241e-01 8.88785779e-01 -7.70520210e-01
9.94975328e-01 3.38343173e-01 -2.32514083e-01 -1.01991570e+00
1.30792201e-01 -4.01700348e-01 -8.95221353e-01 -6.15160108e-01
-2.09891200e-01 2.98948139e-01 -6.63416386e-01 1.47195697e+00
-3.48860294e-01 1.01489954e-01 4.72070485e-01 6.47628307e-01
6.60875440e-01 1.20125926e+00 -4.70766388e-02 1.39802158e-01
1.41636240e+00 -1.43493712e+00 -4.76924360e-01 -4.89999086e-01
1.21951067e+00 -7.99176872e-01 1.38386548e+00 -8.61403272e-02
-1.30485165e+00 -4.51303184e-01 -1.13897657e+00 -6.51022077e-01
-3.08379829e-01 3.23993295e-01 6.26697719e-01 4.37229276e-01
-1.17673683e+00 6.46647274e-01 -8.16223204e-01 -4.65383381e-01
5.30928850e-01 2.65972048e-01 -4.49635953e-01 -6.19147182e-01
-1.28028488e+00 1.04586482e+00 3.15238178e-01 -4.10224080e-01
-7.18176961e-01 -9.11022902e-01 -9.53893244e-01 7.28082836e-01
-9.87961143e-02 -7.49773145e-01 1.18594122e+00 -1.82437435e-01
-1.40655315e+00 7.27136791e-01 -1.26118749e-01 -8.60226870e-01
-6.35977313e-02 -6.30512178e-01 -1.36058554e-01 6.28795028e-01
3.24293450e-02 7.79109776e-01 6.82660103e-01 -5.70128679e-01
-1.45033095e-02 -2.23691568e-01 2.26702645e-01 3.83135200e-01
-1.16577637e+00 2.51398742e-01 -6.04741991e-01 -6.93985701e-01
-7.78194964e-02 -8.30867887e-01 -2.97317684e-01 2.69185573e-01
2.22209492e-03 -3.32628965e-01 6.81712627e-01 -7.29909599e-01
1.33310950e+00 -2.18403125e+00 9.43585485e-02 -3.77997875e-01
1.76111162e-01 5.55496335e-01 -6.53714240e-01 9.09893513e-01
2.26941362e-01 1.90210223e-01 -2.21360683e-01 -8.31611037e-01
2.29655176e-01 2.16569319e-01 -8.29084158e-01 1.83021039e-01
1.81271762e-01 1.16618526e+00 -8.05617332e-01 -4.96924520e-01
7.43193403e-02 6.33184016e-01 -8.12352121e-01 2.63334364e-01
-4.47305385e-03 -5.61216354e-01 -5.09300470e-01 3.43185186e-01
7.81423569e-01 -4.30467665e-01 2.22357333e-01 7.42355511e-02
3.95972550e-01 8.39433253e-01 -5.36025882e-01 2.30252099e+00
-1.01781905e+00 7.99919069e-01 -2.86868811e-01 -9.94497895e-01
9.52939451e-01 2.16940254e-01 1.60941660e-01 -8.85345161e-01
-4.38100845e-02 3.22673202e-01 -5.32504857e-01 -3.12567115e-01
1.03275764e+00 -2.26513058e-01 -2.44930536e-01 8.19294870e-01
2.29615062e-01 -2.02248812e-01 -6.99227378e-02 7.11220384e-01
1.42154002e+00 -4.62851644e-01 8.34792182e-02 -1.65471748e-01
4.33384895e-01 -4.61361080e-01 1.07062832e-01 7.71806180e-01
1.20880321e-01 8.22625279e-01 5.02660751e-01 -2.58898407e-01
-1.37938654e+00 -1.03077376e+00 -1.35526672e-01 9.34239805e-01
9.96601880e-02 -9.13148582e-01 -4.63383675e-01 -4.72609937e-01
6.61786944e-02 6.80033624e-01 -3.10698956e-01 -6.49091542e-01
-6.67181432e-01 -5.76108217e-01 8.45666885e-01 8.80918086e-01
4.87224400e-01 -9.27791715e-01 -2.86612272e-01 8.38856958e-03
3.03243008e-02 -1.28238535e+00 -7.94437826e-01 -2.25940943e-01
-1.07193959e+00 -4.25990880e-01 -9.04877961e-01 -1.03463972e+00
6.56970382e-01 7.57608414e-01 1.06051707e+00 -3.99210006e-02
-3.46442938e-01 2.25523904e-01 -6.18920624e-01 -1.95366815e-01
-2.37310797e-01 3.82291853e-01 8.57947543e-02 -5.92824042e-01
6.09567940e-01 -4.41187918e-01 -6.76503658e-01 -2.83907890e-01
-1.20823991e+00 -1.62592441e-01 9.23860967e-01 1.14046013e+00
4.75926727e-01 -3.08150351e-01 5.38520873e-01 -7.67098784e-01
1.09949982e+00 -4.83625501e-01 -2.56396592e-01 4.54661131e-01
-8.94431531e-01 4.40454721e-01 8.11881185e-01 -3.59609842e-01
-8.18804502e-01 -4.63802338e-01 -1.60157681e-01 -6.39182687e-01
3.41312170e-01 6.13836586e-01 3.26100677e-01 9.46603939e-02
6.31560385e-01 4.47270334e-01 1.80283293e-01 -7.66010642e-01
4.29431826e-01 1.01554620e+00 3.72582972e-02 -5.48649728e-01
7.02778995e-01 3.50433111e-01 -2.95673937e-01 -7.88916767e-01
-8.80982816e-01 -6.21721387e-01 -3.67563456e-01 8.34417939e-01
4.66006160e-01 -1.32303798e+00 6.54219389e-02 -1.00238301e-01
-1.34507334e+00 -1.58282090e-02 -5.35127640e-01 6.24728858e-01
-2.53975391e-01 6.97510540e-01 -1.08632958e+00 -4.03497010e-01
-1.13847280e+00 -8.34080696e-01 1.52141285e+00 -1.40354916e-01
4.18522395e-02 -8.23876739e-01 3.46704900e-01 3.33919376e-01
7.71383286e-01 -5.29735744e-01 7.95269370e-01 -6.50805593e-01
-5.18116713e-01 -4.97284889e-01 -5.27997017e-01 6.45847976e-01
-2.99220413e-01 -6.02196813e-01 -8.90759885e-01 -7.94825792e-01
5.77695817e-02 -7.93114722e-01 1.29010713e+00 -2.68918537e-02
1.47317719e+00 -5.16431093e-01 -2.55294621e-01 5.57774961e-01
1.54569435e+00 -3.75151515e-01 9.71986890e-01 2.38133267e-01
2.99382418e-01 2.19034895e-01 4.08669531e-01 3.43203574e-01
1.51437908e-01 6.64498508e-01 -1.36522233e-01 2.19757035e-01
-2.08843932e-01 -7.15418160e-01 7.05153465e-01 1.39494586e+00
3.95729065e-01 -2.27508157e-01 -5.18378973e-01 5.11687458e-01
-1.60066950e+00 -8.96534443e-01 5.60176671e-01 2.16146302e+00
8.40220749e-01 1.51373848e-01 -5.07546544e-01 -6.48815604e-03
2.41814896e-01 5.62902153e-01 -2.19465345e-01 -7.90192902e-01
5.96945826e-03 7.98916459e-01 4.54109192e-01 4.75637734e-01
-7.75388002e-01 1.03212678e+00 5.90155315e+00 1.15863287e+00
-1.01836669e+00 3.11905175e-01 2.53447026e-01 -4.82447475e-01
-5.07802367e-01 -2.57980563e-02 -8.21975231e-01 1.95385575e-01
1.27872181e+00 -5.59200406e-01 1.64401963e-01 9.57765520e-01
-3.47357303e-01 4.45828915e-01 -1.09968865e+00 1.17108977e+00
6.25390410e-01 -1.61073196e+00 5.11019051e-01 8.91990960e-02
7.59917438e-01 2.25267351e-01 1.77206412e-01 8.71646225e-01
-1.76418781e-01 -1.17191231e+00 7.25153014e-02 1.63388759e-01
1.01919603e+00 -5.77341080e-01 8.82541716e-01 2.70504326e-01
-1.05179310e+00 -7.03940466e-02 -1.21227908e+00 -7.24733248e-02
2.18934357e-01 4.10469681e-01 -7.63211787e-01 3.53858143e-01
4.18928206e-01 8.76115441e-01 -6.45407021e-01 7.88547933e-01
-2.00939029e-01 4.51792985e-01 -3.11374694e-01 -3.59769970e-01
4.20931011e-01 -1.75270543e-01 2.60761410e-01 1.24565065e+00
5.40323019e-01 6.27425015e-02 -1.87757939e-01 7.13774681e-01
-6.96810782e-01 2.18501776e-01 -9.29569721e-01 -3.84726554e-01
6.62457943e-01 9.49782848e-01 1.49429059e-02 -5.64918697e-01
-5.96855283e-01 1.37443352e+00 6.93229318e-01 1.70735285e-01
-6.25790000e-01 -7.55904794e-01 6.57463610e-01 1.68215483e-02
5.44793665e-01 -2.88471699e-01 -1.90656930e-02 -1.61312389e+00
3.64793867e-01 -6.11889362e-01 3.14095229e-01 -6.36318803e-01
-1.38276899e+00 5.61294734e-01 -2.78899580e-01 -1.21103597e+00
-1.06841289e-01 -5.24583220e-01 -5.42485416e-01 7.98988521e-01
-1.74608648e+00 -1.12551475e+00 2.05422416e-01 1.56905293e-01
8.15104067e-01 -2.30825707e-01 1.35432208e+00 5.31528771e-01
-3.40371579e-01 9.54040825e-01 7.59567618e-01 7.36363828e-02
7.44571865e-01 -9.46282685e-01 5.49252987e-01 4.89540339e-01
4.21034932e-01 1.00857377e+00 1.07837483e-01 -2.36309752e-01
-1.78286123e+00 -1.12460399e+00 1.42628109e+00 -3.40886265e-01
6.97114289e-01 -5.76000810e-01 -9.54892874e-01 6.47622526e-01
4.50236440e-01 -1.88736022e-02 8.16997468e-01 2.49565512e-01
-7.63167262e-01 -1.49339885e-01 -1.00805891e+00 6.31753862e-01
8.68331313e-01 -1.07815564e+00 -1.07356322e+00 6.70548856e-01
1.35829329e+00 -1.51405752e-01 -9.54493463e-01 8.21116865e-02
4.78240281e-01 -4.05317336e-01 1.44528043e+00 -7.75292277e-01
8.51832747e-01 3.80293190e-01 -3.51120949e-01 -1.09977210e+00
-3.16450030e-01 -3.28198791e-01 -4.59995449e-01 7.80664444e-01
3.90137076e-01 -6.78277075e-01 7.39084244e-01 6.57776475e-01
-1.24279685e-01 -1.14231598e+00 -8.55989873e-01 -1.03289962e+00
6.37139022e-01 -1.21174417e-01 5.86976171e-01 5.93265891e-01
4.17416155e-01 7.44138479e-01 -3.05867374e-01 -1.12258554e-01
3.86830658e-01 2.68436551e-01 5.40159941e-01 -7.62929261e-01
-3.83052319e-01 -3.32510710e-01 -4.58912373e-01 -1.62907517e+00
3.76755416e-01 -1.29074335e+00 -4.30030942e-01 -1.62562144e+00
6.92220211e-01 -3.39062840e-01 -4.45508868e-01 1.39862940e-01
-1.84028700e-01 8.74576941e-02 2.87938416e-01 2.24636957e-01
-7.64995396e-01 1.21795094e+00 1.13497221e+00 -3.90690953e-01
1.08304977e-01 -5.31396806e-01 -8.89646888e-01 1.21411793e-01
9.43047106e-01 -4.98895139e-01 -5.82485795e-01 -1.07423592e+00
1.09073564e-01 -8.93163010e-02 9.74028930e-02 -9.04412627e-01
2.47785434e-01 4.21143472e-01 3.14736962e-02 -4.60731477e-01
7.66828835e-01 -4.82038856e-01 -6.43630743e-01 6.42669141e-01
-8.47882569e-01 1.35371104e-01 1.79434851e-01 6.76860750e-01
-4.36016738e-01 -4.93608057e-01 3.90108466e-01 -1.10073768e-01
-2.84807533e-01 4.09386992e-01 -1.16020469e-02 2.48837829e-01
6.39301658e-01 3.35634947e-01 -5.19954622e-01 -4.27673429e-01
-3.45126055e-02 2.40108535e-01 2.60472029e-01 4.93531346e-01
1.19261491e+00 -1.41551650e+00 -7.94467807e-01 2.00918570e-01
1.85427696e-01 -1.56327873e-01 1.52143285e-01 3.72288674e-01
-6.64628923e-01 1.11038578e+00 7.70285279e-02 -9.28514302e-02
-1.17845786e+00 6.28074288e-01 -2.53552943e-01 -6.56551003e-01
-9.47356105e-01 9.58591104e-01 4.08681631e-02 -5.33728719e-01
2.27021202e-01 -2.59459317e-01 -8.57262872e-03 -3.77042174e-01
8.86763155e-01 3.17905009e-01 -3.37814800e-02 -7.15147108e-02
-2.06379145e-02 4.74827051e-01 -6.94869816e-01 7.76641145e-02
1.62966645e+00 -5.98079413e-02 -2.52455115e-01 -1.56041663e-02
2.17968416e+00 -9.58249643e-02 -5.94620764e-01 -6.14942789e-01
-2.80939937e-01 -7.97317326e-01 1.29084945e-01 -1.89917728e-01
-6.82181060e-01 1.41773665e+00 3.13903302e-01 -1.18471250e-01
8.28995883e-01 1.33402616e-01 1.68136370e+00 1.02975535e+00
4.23971742e-01 -1.14393020e+00 4.48714584e-01 6.45863891e-01
1.06068671e+00 -1.15405917e+00 2.34462202e-01 -2.16986164e-01
-6.10511065e-01 1.02246630e+00 4.28288400e-01 -5.03995717e-01
5.08624375e-01 -2.58240588e-02 -4.25460935e-01 -2.41660431e-01
-1.11478126e+00 2.56006271e-01 1.88563183e-01 1.30099788e-01
4.02424753e-01 -1.32917827e-02 -4.38629150e-01 4.61864620e-01
-1.78713024e-01 -3.22575122e-02 3.26826990e-01 1.01567507e+00
-4.79214758e-01 -1.34836030e+00 4.63527516e-02 7.94171393e-01
-4.09088403e-01 -7.72283673e-01 -9.21655893e-02 3.63970786e-01
-7.97938943e-01 5.61926425e-01 2.16766804e-01 -4.47667658e-01
1.84858769e-01 1.26118332e-01 4.82363254e-01 -9.44778502e-01
-3.56703162e-01 -4.38127816e-01 1.05578415e-01 -6.19625688e-01
1.56723976e-01 -3.78261954e-01 -1.07543266e+00 -3.91440243e-01
-3.58905226e-01 3.39750051e-01 5.33374131e-01 4.74623471e-01
8.55638921e-01 1.03693344e-01 7.32255936e-01 -5.36721945e-01
-1.41358125e+00 -1.15628958e+00 -6.12065792e-01 3.87776762e-01
2.18802579e-02 -1.93791226e-01 -4.63958442e-01 -3.68683547e-01] | [11.033600807189941, 8.396317481994629] |
dbc92207-658a-4f94-bd60-8a3a732287df | bridging-the-training-inference-gap-for-dense | 2210.13678 | null | https://arxiv.org/abs/2210.13678v1 | https://arxiv.org/pdf/2210.13678v1.pdf | Bridging the Training-Inference Gap for Dense Phrase Retrieval | Building dense retrievers requires a series of standard procedures, including training and validating neural models and creating indexes for efficient search. However, these procedures are often misaligned in that training objectives do not exactly reflect the retrieval scenario at inference time. In this paper, we explore how the gap between training and inference in dense retrieval can be reduced, focusing on dense phrase retrieval (Lee et al., 2021) where billions of representations are indexed at inference. Since validating every dense retriever with a large-scale index is practically infeasible, we propose an efficient way of validating dense retrievers using a small subset of the entire corpus. This allows us to validate various training strategies including unifying contrastive loss terms and using hard negatives for phrase retrieval, which largely reduces the training-inference discrepancy. As a result, we improve top-1 phrase retrieval accuracy by 2~3 points and top-20 passage retrieval accuracy by 2~4 points for open-domain question answering. Our work urges modeling dense retrievers with careful consideration of training and inference via efficient validation while advancing phrase retrieval as a general solution for dense retrieval. | ['William Yang Wang', 'Yashar Mehdad', 'Yizhe Zhang', 'Wenhan Xiong', 'Barlas Oguz', 'Jinhyuk Lee', 'Gyuwan Kim'] | 2022-10-25 | null | null | null | null | ['passage-retrieval', 'open-domain-question-answering'] | ['natural-language-processing', 'natural-language-processing'] | [-8.44425932e-02 -1.35227084e-01 -3.26556176e-01 -1.15563683e-01
-1.70395863e+00 -6.41831577e-01 2.86415279e-01 2.58965254e-01
-6.28600121e-01 9.19609010e-01 1.25633895e-01 -3.59030634e-01
-5.45145273e-01 -1.11982822e+00 -9.78009284e-01 -3.30920994e-01
2.25592598e-01 1.07597375e+00 2.33940661e-01 -4.67889667e-01
3.09213817e-01 2.59727716e-01 -1.46059549e+00 3.20831001e-01
9.72554743e-01 9.91531670e-01 4.11700070e-01 7.61805713e-01
-2.63955116e-01 4.77143407e-01 -7.93024421e-01 -6.42146468e-01
8.96433070e-02 -6.00349195e-02 -9.40824270e-01 -7.12759972e-01
6.03889525e-01 -5.46311796e-01 -4.52010304e-01 8.94670904e-01
7.43842840e-01 3.48993003e-01 6.30635917e-01 -6.16483092e-01
-8.51689160e-01 8.44469488e-01 -3.05377215e-01 3.42101902e-01
5.52168310e-01 -3.26711774e-01 1.64100993e+00 -8.88618350e-01
3.95132065e-01 1.12982428e+00 6.88291252e-01 4.20413524e-01
-1.02792382e+00 -7.66799510e-01 -8.24063197e-02 2.28529930e-01
-1.80543554e+00 -3.93376738e-01 3.44301373e-01 1.20701447e-01
1.29474890e+00 6.13222241e-01 4.12419885e-01 9.38268840e-01
-2.56410152e-01 9.69793439e-01 4.83265281e-01 -6.05019212e-01
-9.16124955e-02 1.75603941e-01 5.05786777e-01 5.03103971e-01
3.85751456e-01 1.97609991e-01 -4.21812117e-01 -4.99936551e-01
4.47623432e-01 -1.86823592e-01 -4.13341433e-01 1.72619224e-01
-9.45145547e-01 8.71464074e-01 5.72344899e-01 1.16426274e-01
-2.79811621e-01 2.22809389e-01 3.02620620e-01 4.25740957e-01
2.92431682e-01 9.49476600e-01 -5.37928641e-01 -5.42307124e-02
-1.21953404e+00 8.22392404e-01 9.60185111e-01 1.10205066e+00
8.04111004e-01 -3.88224930e-01 -5.97182214e-01 1.20025218e+00
3.97634357e-01 9.59459186e-01 3.91060650e-01 -9.87594426e-01
5.96649289e-01 2.13014066e-01 1.47324681e-01 -1.06951284e+00
-1.71141494e-02 -7.84414828e-01 -6.70422375e-01 -7.70039320e-01
1.89514369e-01 2.45432720e-01 -7.76336491e-01 1.66915429e+00
-3.89239192e-02 -9.19092260e-03 -5.34834042e-02 7.65137553e-01
8.81447971e-01 1.02209759e+00 1.26414493e-01 -1.16646640e-01
1.40747893e+00 -9.59757149e-01 -6.98009729e-01 -1.09997004e-01
9.02761579e-01 -7.64930367e-01 1.34867072e+00 1.68117538e-01
-1.44802415e+00 -4.78854209e-01 -9.05193925e-01 -4.32266235e-01
-3.44801962e-01 -1.29521430e-01 5.97555280e-01 3.48758787e-01
-1.13722849e+00 4.21707094e-01 -4.48948979e-01 -1.74843408e-02
7.70567507e-02 4.65466410e-01 -8.11521616e-03 -3.33917499e-01
-1.83626199e+00 1.00120294e+00 5.77083766e-01 1.09531254e-01
-6.26856863e-01 -9.76732552e-01 -6.34840906e-01 5.64621329e-01
2.09733769e-01 -1.15181756e+00 1.51741815e+00 9.44714546e-02
-9.81779516e-01 6.50889099e-01 -2.45723814e-01 -5.15355527e-01
-3.03332116e-02 -5.84764600e-01 -2.62792408e-01 2.14715749e-01
6.38612062e-02 6.55501962e-01 5.30545413e-01 -1.13343036e+00
-4.88955945e-01 -1.49737760e-01 2.99927831e-01 4.85914141e-01
-5.66175222e-01 -5.65656833e-02 -1.05249000e+00 -5.75774491e-01
6.60479888e-02 -7.41716564e-01 8.71961936e-02 -4.68584180e-01
-3.58308017e-01 -4.78006840e-01 3.12158436e-01 -6.73040986e-01
1.80982852e+00 -1.80112648e+00 1.00059696e-01 5.20842671e-01
3.43008190e-01 2.29039297e-01 -3.81785929e-01 3.88633221e-01
5.70875645e-01 2.25707740e-01 4.22181226e-02 -4.17729527e-01
2.63593107e-01 3.47037822e-01 -8.90351534e-01 -7.51175955e-02
-1.06002958e-02 1.33320081e+00 -8.07280421e-01 -8.67574692e-01
-3.24307114e-01 3.45052242e-01 -8.95127594e-01 3.76785487e-01
-4.51678038e-01 -1.91319183e-01 -7.58462906e-01 8.67362916e-01
3.55327368e-01 -7.61574924e-01 -1.80308327e-01 -1.78680822e-01
4.86870706e-01 6.49123549e-01 -7.83962786e-01 1.90919137e+00
-5.91180861e-01 4.84391034e-01 -3.63062590e-01 -8.34278524e-01
8.56229186e-01 2.88955390e-01 1.49864867e-01 -1.14560318e+00
-1.63296193e-01 4.76356089e-01 -4.10285413e-01 -3.54328066e-01
1.11781669e+00 -2.11712807e-01 -2.42795065e-01 6.12288833e-01
3.25513259e-02 -3.02107006e-01 4.41176951e-01 4.23928410e-01
1.16856778e+00 -2.52353787e-01 -3.37915897e-01 -1.22188106e-02
3.14401984e-01 3.05916239e-02 3.25724095e-01 1.35259604e+00
2.40923300e-01 7.20505893e-01 2.00607572e-02 -2.33372618e-02
-1.23152220e+00 -1.14577436e+00 -4.12858009e-01 1.25764239e+00
6.53101653e-02 -4.70207989e-01 -5.01855850e-01 -3.31402063e-01
2.64959693e-01 7.83529162e-01 -2.87301838e-01 -3.67545754e-01
-8.68567705e-01 -7.30880380e-01 9.69104707e-01 5.40894628e-01
3.90798032e-01 -1.02771616e+00 5.93315624e-03 8.46040770e-02
-6.41895354e-01 -9.42866921e-01 -2.73425162e-01 1.20985769e-01
-1.07314980e+00 -7.19520330e-01 -1.04892409e+00 -9.19517994e-01
5.61273634e-01 2.84907877e-01 1.83846200e+00 5.88769317e-01
7.95767549e-03 2.28147298e-01 -4.10319328e-01 -9.52573121e-02
-1.99961185e-01 6.94654644e-01 -1.50203391e-03 -9.47667718e-01
5.99627256e-01 -4.34326410e-01 -7.26094842e-01 2.34827816e-01
-8.44263554e-01 -1.76316053e-01 8.02858651e-01 9.61630583e-01
7.31545091e-01 -1.36019766e-01 6.56615019e-01 -8.45251143e-01
1.09934223e+00 -5.63712001e-01 -5.98406374e-01 8.17286313e-01
-9.58224237e-01 2.51614660e-01 3.11467171e-01 -3.09737712e-01
-7.75710881e-01 -8.48804474e-01 -4.29940552e-01 -5.46795011e-01
4.09976512e-01 8.51400554e-01 3.68123859e-01 3.21796276e-02
9.03541982e-01 2.86406249e-01 -2.71710813e-01 -6.91804886e-01
4.23426360e-01 6.54955029e-01 1.97660133e-01 -1.13492596e+00
8.99686396e-01 -1.20001905e-01 -4.46590811e-01 -3.08685005e-01
-1.27797449e+00 -7.01407194e-01 -4.18211557e-02 1.42423689e-01
3.26435447e-01 -9.70773399e-01 -7.00515628e-01 -7.62528181e-02
-1.08713913e+00 -9.10023674e-02 -4.20463741e-01 4.89294112e-01
-3.41952413e-01 3.62740815e-01 -9.71857011e-01 -4.75667000e-01
-9.39254820e-01 -1.07418740e+00 1.41341496e+00 4.74270061e-02
-2.95650393e-01 -8.64101052e-01 4.27721471e-01 6.17491484e-01
7.52721846e-01 -6.62608504e-01 1.16024387e+00 -6.75139248e-01
-8.59602928e-01 -5.61309755e-01 -3.99480253e-01 3.38558376e-01
-4.02195811e-01 -4.56344694e-01 -9.52403188e-01 -3.47381413e-01
-1.48320302e-01 -8.27780306e-01 9.30059969e-01 2.29662061e-01
1.37463474e+00 -1.68826342e-01 -2.54599959e-01 5.55576384e-01
1.35542977e+00 -1.11976154e-01 7.65255630e-01 4.94390339e-01
4.67416286e-01 1.78298071e-01 6.82053506e-01 9.07138065e-02
1.82662651e-01 5.82427025e-01 -1.54262140e-01 2.66499907e-01
-5.86510636e-02 -5.39769650e-01 -6.89853877e-02 1.32434857e+00
7.65754357e-02 -4.29462522e-01 -9.35316205e-01 4.23106462e-01
-1.64600658e+00 -9.00553703e-01 3.87483627e-01 2.15252495e+00
1.38676596e+00 2.32289106e-01 -3.31050724e-01 -2.10801195e-02
5.29469490e-01 1.02160564e-02 -4.17475104e-01 -1.23517096e-01
-7.14074224e-02 7.14497566e-01 2.71902889e-01 5.99858224e-01
-6.40193343e-01 9.22585905e-01 7.15570211e+00 1.31247115e+00
-5.99302590e-01 -1.08443767e-01 5.47384202e-01 -3.62470090e-01
-8.21016967e-01 -5.37312962e-02 -1.39108908e+00 2.75756359e-01
1.22935665e+00 -2.61294961e-01 4.58916873e-01 7.08470285e-01
-4.59759504e-01 1.58931077e-01 -1.16388369e+00 9.62364495e-01
4.63300757e-02 -1.42412472e+00 5.16462803e-01 -1.48879126e-01
6.20077550e-01 3.22980992e-02 1.54559806e-01 1.11762071e+00
-9.33155417e-03 -1.15820312e+00 3.68573070e-01 6.35348022e-01
6.58558965e-01 -5.21873891e-01 6.83288991e-01 4.27756786e-01
-8.15249324e-01 1.73394695e-01 -6.43972933e-01 2.80503452e-01
2.35577404e-01 6.08838379e-01 -5.34940481e-01 3.08027416e-01
6.80461884e-01 3.28610539e-01 -4.22393799e-01 1.07138300e+00
-6.07531471e-03 5.74147761e-01 -6.96251333e-01 -3.75202149e-01
2.58101821e-01 8.32321644e-02 3.96937966e-01 1.14835227e+00
3.95367384e-01 1.37401432e-01 -2.29303420e-01 9.13009048e-01
-5.11352479e-01 2.25607324e-02 -4.22254354e-01 -1.16750762e-01
8.51202011e-01 8.63171875e-01 1.50732007e-02 -4.92020965e-01
-2.83198237e-01 5.48610747e-01 5.38126349e-01 5.78573346e-01
-8.10887933e-01 -5.28452575e-01 2.75501668e-01 -1.55662447e-02
2.94179887e-01 7.23262429e-02 -1.08192824e-01 -1.24413419e+00
2.02419952e-01 -1.05030656e+00 5.57673693e-01 -8.33476603e-01
-1.35899997e+00 4.36816752e-01 2.76039988e-01 -9.92694318e-01
-4.40610439e-01 -3.36154073e-01 1.32428613e-02 1.17716956e+00
-2.01023102e+00 -6.86295569e-01 -2.69090924e-02 5.44693708e-01
4.50160742e-01 -3.45390812e-02 1.13351846e+00 8.82345438e-01
-2.53178149e-01 1.14820099e+00 2.32918784e-01 8.69881138e-02
5.89024484e-01 -1.04781604e+00 2.01661021e-01 2.15458363e-01
3.70992184e-01 1.27434361e+00 5.38851142e-01 -4.25457507e-01
-1.62714183e+00 -6.72661483e-01 1.12501323e+00 -5.53259492e-01
7.45530665e-01 -1.08040497e-01 -1.27183020e+00 5.11466682e-01
-1.30614694e-02 -2.32496142e-01 8.59392405e-01 8.51857305e-01
-4.69263077e-01 -2.00706184e-01 -7.96589613e-01 6.34647012e-01
8.86896491e-01 -1.01835263e+00 -1.04451919e+00 6.40150726e-01
1.09125996e+00 -5.75297892e-01 -1.18969274e+00 8.07834685e-01
5.92122018e-01 -2.66965091e-01 1.38820267e+00 -7.48674095e-01
4.23761815e-01 9.33899432e-02 -4.40829635e-01 -8.36047888e-01
-2.66298801e-01 -1.04581028e-01 -8.45292330e-01 8.25571716e-01
8.12479615e-01 -4.28961605e-01 8.44309747e-01 6.64431036e-01
-4.87845764e-02 -1.07250059e+00 -6.20602787e-01 -6.74965680e-01
4.29749161e-01 -4.41850156e-01 6.18454456e-01 5.89194119e-01
-1.70492262e-01 4.77442682e-01 -1.89734176e-02 1.03555217e-01
5.02599895e-01 3.57492656e-01 4.95693356e-01 -1.00152194e+00
-5.86688519e-01 -4.30996746e-01 1.16481632e-01 -1.77836359e+00
2.30559811e-01 -9.07279730e-01 1.40612945e-01 -1.44525135e+00
4.61329788e-01 -1.04542184e+00 -6.51258826e-01 1.64484933e-01
-3.71103644e-01 2.72547066e-01 -1.63132235e-01 6.82970226e-01
-9.32097197e-01 6.72439218e-01 1.36971271e+00 -3.44166994e-01
2.06111610e-01 -1.04228556e-01 -8.40935767e-01 9.03250054e-02
6.57910168e-01 -5.40542305e-01 -5.81731796e-01 -8.91450465e-01
9.30131972e-01 1.97718561e-01 1.69914722e-01 -8.06078374e-01
5.09534121e-01 3.53705466e-01 1.70087516e-01 -8.54182184e-01
5.37787139e-01 -6.77196205e-01 -1.75638378e-01 9.72093120e-02
-8.36923361e-01 1.08815685e-01 4.00619775e-01 6.38884604e-01
-6.25653625e-01 -5.94571710e-01 1.20348185e-01 -3.89425367e-01
-5.08577704e-01 3.37214798e-01 2.00574294e-01 5.87376952e-01
3.26601267e-01 1.40331805e-01 -5.34938157e-01 -2.55664885e-01
-5.40625036e-01 5.80175817e-01 2.65059005e-02 2.61573225e-01
7.02740490e-01 -1.28997612e+00 -7.22134292e-01 7.47029558e-02
7.22151101e-02 2.09140688e-01 2.07031280e-01 4.92948920e-01
-6.10529721e-01 1.01401114e+00 4.19498801e-01 -5.64155519e-01
-9.21280265e-01 3.25115800e-01 6.30902201e-02 -8.84724975e-01
-5.99047959e-01 1.22213459e+00 -8.46318975e-02 -4.84523952e-01
6.11648917e-01 -8.82547721e-02 -6.32292032e-02 -5.84299900e-02
6.25947297e-01 5.56826629e-02 2.80530542e-01 9.52334516e-03
3.07511613e-02 5.74286938e-01 -6.84512377e-01 -1.29671022e-01
1.06556606e+00 -3.75036187e-02 -1.87348217e-01 2.38461286e-01
1.66143894e+00 -9.87917855e-02 -3.91204894e-01 -5.25435030e-01
-1.45133182e-01 -4.71803159e-01 2.96812743e-01 -7.82326698e-01
-8.62211525e-01 7.53232718e-01 2.44159564e-01 2.43780598e-01
1.20615351e+00 5.29988334e-02 1.48590291e+00 1.24419320e+00
5.16509712e-01 -1.03631556e+00 -2.37976555e-02 8.07547390e-01
8.00001502e-01 -1.19356287e+00 9.74533185e-02 6.79637045e-02
-9.04585198e-02 7.46104360e-01 4.81492460e-01 -8.51243287e-02
5.59813440e-01 -2.82421876e-02 -2.67883390e-01 -5.16399264e-01
-8.62898529e-01 4.88671996e-02 5.51135421e-01 -3.16835493e-02
3.58018458e-01 -1.82283103e-01 -2.99283952e-01 2.88081229e-01
-4.49858397e-01 1.05536208e-01 -4.50150937e-01 8.71649086e-01
-6.37944698e-01 -1.09583127e+00 -2.25487441e-01 8.26355875e-01
-4.64641809e-01 -7.11208820e-01 1.68678924e-01 7.26106226e-01
-5.78800917e-01 5.93169093e-01 8.12427998e-02 -3.20351124e-01
2.94672966e-01 2.17566296e-01 4.66398329e-01 -6.29021823e-01
-5.31728745e-01 -2.41396829e-01 3.60666960e-01 -4.14492369e-01
-7.67903626e-02 -1.59579575e-01 -1.04523027e+00 -5.85347056e-01
-9.31816220e-01 8.25973332e-01 4.62340534e-01 8.20424616e-01
4.08639163e-01 3.53422284e-01 4.03902084e-01 -2.59330213e-01
-1.20954669e+00 -1.05995154e+00 -4.96584982e-01 3.11484545e-01
1.36579931e-01 -5.34794152e-01 -6.02200031e-01 -3.37153554e-01] | [11.458056449890137, 7.696648120880127] |
3f1b8080-c0a4-45b6-a10c-6a111f2e983e | multi-domain-dialogue-state-tracking-with-top | null | null | https://aclanthology.org/2022.sigdial-1.24 | https://aclanthology.org/2022.sigdial-1.24.pdf | Multi-Domain Dialogue State Tracking with Top-K Slot Self Attention | As an important component of task-oriented dialogue systems, dialogue state tracking is designed to track the dialogue state through the conversations between users and systems. Multi-domain dialogue state tracking is a challenging task, in which the correlation among different domains and slots needs to consider. Recently, slot self-attention is proposed to provide a data-driven manner to handle it. However, a full-support slot self-attention may involve redundant information interchange. In this paper, we propose a top-k attention-based slot self-attention for multi-domain dialogue state tracking. In the slot self-attention layers, we force each slot to involve information from the other k prominent slots and mask the rest out. The experimental results on two mainstream multi-domain task-oriented dialogue datasets, MultiWOZ 2.0 and MultiWOZ 2.4, present that our proposed approach is effective to improve the performance of multi-domain dialogue state tracking. We also find that the best result is obtained when each slot interchanges information with only a few slots. | ['Takahiro Shinozaki', 'Sheng Li', 'Jiyi Li', 'Longfei Yang'] | null | null | null | null | sigdial-acl-2022-9 | ['dialogue-state-tracking', 'task-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing'] | [-3.25005166e-02 1.69339150e-01 -3.76188546e-01 -3.23831022e-01
-5.27383745e-01 -5.60621977e-01 8.55464697e-01 7.40218386e-02
-3.40948045e-01 7.48937249e-01 6.78329349e-01 -2.27676719e-01
5.87380268e-02 -2.27034077e-01 3.38215679e-01 -4.08346385e-01
2.19197512e-01 8.27042878e-01 6.78510308e-01 -1.07615817e+00
3.88414860e-01 -2.13707507e-01 -9.64078069e-01 4.19395685e-01
9.15737212e-01 6.19273961e-01 3.19297403e-01 7.08494008e-01
-8.08259070e-01 7.82757223e-01 -7.55271494e-01 -1.23880327e-01
-9.14828032e-02 -6.68101788e-01 -1.59135199e+00 3.13368529e-01
-2.33567610e-01 -4.26988602e-01 -2.89830148e-01 7.91953146e-01
6.90096736e-01 4.10440236e-01 3.28282326e-01 -1.22561145e+00
-2.35672519e-02 6.58862054e-01 -4.01821911e-01 4.08427328e-01
5.27285337e-01 7.52520710e-02 1.02034760e+00 -4.65751678e-01
5.33151984e-01 1.55286884e+00 2.71645486e-01 1.03501570e+00
-7.63274550e-01 -3.90504301e-01 4.51556504e-01 -2.44760755e-02
-8.75172973e-01 -7.56708503e-01 9.28570688e-01 -3.52070332e-01
1.06000018e+00 3.64181995e-01 3.29846889e-01 7.27308691e-01
-1.75897896e-01 1.15157759e+00 9.33727801e-01 -6.11952603e-01
-1.18407808e-01 3.13844591e-01 5.70008636e-01 3.59339893e-01
-6.25686586e-01 -5.67193151e-01 -6.52551353e-01 -5.08275986e-01
5.83389342e-01 -2.11574063e-01 1.82141718e-02 -1.69441495e-02
-1.13977873e+00 1.00116456e+00 -1.56038180e-01 5.26780307e-01
-1.87445432e-01 -5.14052987e-01 9.83378470e-01 3.08728546e-01
7.52618313e-01 4.39157814e-01 -7.83728302e-01 -7.29753971e-01
-3.92495334e-01 2.88995296e-01 9.58573341e-01 8.99726927e-01
6.45040214e-01 -3.47447693e-01 -6.89044654e-01 1.54798400e+00
3.54302794e-01 -6.72541419e-03 8.58789146e-01 -7.98455715e-01
6.60664201e-01 9.19399619e-01 3.65174562e-01 -5.70818007e-01
-6.68296039e-01 3.59225750e-01 -8.25955749e-01 -3.92059624e-01
5.45824766e-01 -5.56708157e-01 -3.90901834e-01 1.67456985e+00
7.01291919e-01 -2.04635635e-01 3.18851858e-01 8.25084865e-01
1.02175319e+00 7.14513838e-01 2.88647443e-01 -3.64662856e-01
1.92788506e+00 -1.19540453e+00 -1.34900463e+00 -4.36459899e-01
7.28143632e-01 -9.28253174e-01 1.13207448e+00 -1.76239625e-01
-9.27075565e-01 -5.05617201e-01 -6.39866710e-01 -1.70129210e-01
-2.50635982e-01 4.03368622e-02 5.22984922e-01 4.95306879e-01
-6.39413059e-01 1.08250760e-01 -4.72885787e-01 -5.19309700e-01
-1.65023491e-01 9.91722122e-02 -1.66086748e-01 3.92489314e-01
-1.90974438e+00 9.57364917e-01 6.30828738e-01 2.84609646e-02
-2.87868589e-01 -1.85450405e-01 -7.93571174e-01 -4.81822304e-02
6.07589066e-01 -1.11967728e-01 1.83601689e+00 -4.62468266e-01
-1.98608446e+00 7.51378238e-01 -4.75396901e-01 -1.77896068e-01
3.87939900e-01 -2.62792170e-01 -4.62354690e-01 -1.27397269e-01
1.89863473e-01 2.86402017e-01 4.33080435e-01 -8.32904756e-01
-9.17822838e-01 -2.20835179e-01 3.43640029e-01 8.21341157e-01
-3.38436842e-01 2.42230386e-01 -7.76409984e-01 -6.47086725e-02
1.24916494e-01 -9.83638108e-01 -2.22335637e-01 -5.08848310e-01
-6.32659197e-01 -9.61827576e-01 9.49944615e-01 -5.50298274e-01
1.58196735e+00 -2.09813404e+00 -7.24880323e-02 -3.68657112e-01
3.25440794e-01 5.63273609e-01 1.34457618e-01 7.67994404e-01
2.01940179e-01 -2.22663373e-01 4.98290323e-02 -4.47915614e-01
1.30059302e-01 1.51989177e-01 -5.30729108e-02 2.05745757e-01
1.49353176e-01 6.49797022e-01 -8.62075925e-01 -5.95910788e-01
3.23527366e-01 -2.29370192e-01 -4.88675296e-01 7.04973400e-01
-4.40668881e-01 4.95190740e-01 -5.15657067e-01 1.26441389e-01
5.24103820e-01 -4.70631629e-01 4.82826561e-01 -1.56824380e-01
-2.93870240e-01 8.04445267e-01 -1.01090026e+00 1.99367344e+00
-4.25541788e-01 5.82692742e-01 2.09348783e-01 -7.47209430e-01
1.02365804e+00 6.60236657e-01 5.29549956e-01 -7.40823328e-01
2.18430653e-01 -8.57326910e-02 2.01405898e-01 -5.08606970e-01
1.09438634e+00 -3.49367708e-02 -6.48202837e-01 7.18734503e-01
8.96687433e-02 1.47214711e-01 2.62805343e-01 3.84460896e-01
7.64734447e-01 -4.49434757e-01 6.20106459e-01 -1.98914617e-01
9.21625197e-01 2.24903107e-01 6.87871575e-01 6.36974156e-01
-7.46228576e-01 9.57009122e-02 6.02581799e-01 -1.33651286e-01
-9.22802985e-01 -3.21250081e-01 7.23640397e-02 1.86250949e+00
3.65911156e-01 -4.94788826e-01 -7.22026467e-01 -7.92013764e-01
-1.53792799e-01 3.07096153e-01 -4.18692857e-01 -2.86277622e-01
-4.00549144e-01 -6.70981467e-01 5.57151139e-01 8.32091272e-02
1.04198754e+00 -1.18207288e+00 -2.10945874e-01 6.16646111e-01
-6.72805905e-01 -1.10575247e+00 -7.46009588e-01 1.97078690e-01
-4.02582347e-01 -1.08189893e+00 -8.94090712e-01 -8.29681635e-01
1.98070928e-01 4.02141988e-01 9.73210573e-01 -2.79134531e-02
1.84295923e-01 1.77489802e-01 -4.69968051e-01 -2.24816683e-03
-6.74001515e-01 6.36940777e-01 8.87624696e-02 -1.52964637e-01
6.61907017e-01 2.61773150e-02 -4.26258117e-01 7.15570807e-01
-4.41048026e-01 2.46551484e-01 3.48884985e-02 1.28431511e+00
-3.30798119e-01 -1.51714608e-01 8.61013114e-01 -1.28456497e+00
1.37743759e+00 -4.96251792e-01 -1.69516101e-01 3.09398264e-01
-2.90154159e-01 5.21449670e-02 5.47122300e-01 -3.87026161e-01
-1.43779552e+00 -1.24289691e-01 -3.62606108e-01 1.89976186e-01
-3.44041109e-01 3.64834547e-01 -2.85738319e-01 3.40130061e-01
4.86161917e-01 4.79303479e-01 2.44171456e-01 -7.22845018e-01
3.45411003e-01 1.25398135e+00 3.45320940e-01 -5.91569185e-01
2.88012713e-01 -7.26481155e-02 -8.32505584e-01 -9.90081668e-01
-7.91584134e-01 -1.14163435e+00 -6.30955875e-01 -1.83756277e-01
8.17283154e-01 -8.26144636e-01 -1.12877905e+00 6.45879507e-01
-1.31055439e+00 -3.89941394e-01 1.79812208e-01 7.15688542e-02
-3.67797494e-01 7.12242365e-01 -8.44075739e-01 -1.16388679e+00
-5.22571504e-01 -1.09031570e+00 8.42424750e-01 5.20497739e-01
-5.48118055e-01 -1.18656790e+00 4.14758533e-01 4.56643105e-01
4.16010708e-01 -5.48234880e-01 7.38812208e-01 -1.36645210e+00
1.91963553e-01 1.23373076e-01 -2.34347016e-01 3.87526234e-03
6.89333677e-01 -4.40650076e-01 -9.45253491e-01 -2.39700288e-01
-1.31795347e-01 -5.01429915e-01 3.22572321e-01 -5.74028231e-02
4.95799899e-01 -3.13947618e-01 -1.65421396e-01 -2.62443572e-01
6.29971623e-01 5.61726153e-01 1.93917513e-01 3.64032120e-01
5.34704387e-01 9.00006056e-01 1.11630535e+00 7.39939630e-01
7.53722727e-01 9.15724576e-01 -4.85014804e-02 -3.03277016e-01
2.01781571e-01 -8.84121507e-02 2.31829897e-01 1.10402572e+00
5.91684520e-01 -2.23789573e-01 -1.02451944e+00 5.91575503e-01
-2.24658060e+00 -8.24802101e-01 -4.90827933e-02 1.78115940e+00
1.32019973e+00 3.32148105e-01 4.90777045e-01 -4.18715999e-02
1.01657093e+00 5.16930401e-01 -5.80693305e-01 -4.80955362e-01
1.62101775e-01 -4.45444852e-01 1.09237041e-02 7.24587381e-01
-1.32929027e+00 1.43730414e+00 6.13884878e+00 9.34293032e-01
-8.72347891e-01 1.25193715e-01 3.31534386e-01 3.49727899e-01
7.86347017e-02 -1.73252970e-01 -1.20086491e+00 7.25480199e-01
8.59134614e-01 -5.33015728e-01 6.24711588e-02 8.04439366e-01
2.58857846e-01 -3.84156674e-01 -6.40747368e-01 8.02143395e-01
-3.36214900e-01 -1.11095679e+00 -2.97845423e-01 -1.52452543e-01
3.00485075e-01 -2.77772099e-01 -3.15278649e-01 6.84681654e-01
6.62433326e-01 -4.73188221e-01 3.60971391e-02 -8.73855203e-02
5.53229332e-01 -5.92728436e-01 7.05925524e-01 6.85779750e-01
-1.31201851e+00 1.24795072e-01 -2.46022016e-01 -1.61808014e-01
3.14229101e-01 6.50815740e-02 -1.45898664e+00 3.59308511e-01
2.95569092e-01 5.72173715e-01 -5.66130094e-02 7.91515410e-01
2.88657278e-01 4.30760890e-01 -1.38860881e-01 -2.92066246e-01
4.73677218e-01 -1.03288935e-02 4.43381011e-01 1.34263575e+00
-3.76027256e-01 1.94424599e-01 6.84390604e-01 2.93291539e-01
1.96210459e-01 2.76352674e-01 -4.28225726e-01 -4.77721952e-02
6.90008521e-01 1.20563364e+00 -4.54541326e-01 -5.26114523e-01
-4.99566436e-01 1.18425333e+00 2.32251287e-01 -3.85370813e-02
-4.29673821e-01 -4.29540783e-01 1.23191166e+00 -2.93667465e-01
-1.55192330e-01 -2.41624802e-01 2.99259182e-02 -1.05819917e+00
-4.46995020e-01 -8.86461616e-01 6.17729723e-01 -3.74104947e-01
-1.19790375e+00 6.76516771e-01 -7.52588138e-02 -1.23038137e+00
-5.57655513e-01 -2.25478530e-01 -6.23394549e-01 1.23484540e+00
-1.54716122e+00 -8.16163421e-01 -1.05061807e-01 7.86399007e-01
1.28149128e+00 -3.60372961e-01 1.10867095e+00 3.11518818e-01
-7.40567267e-01 6.16003692e-01 1.36852115e-01 5.26822388e-01
1.01099908e+00 -1.34300411e+00 6.82866633e-01 3.88027549e-01
-5.22836030e-01 7.10560560e-01 7.25629210e-01 -6.83649600e-01
-9.36465740e-01 -5.71300507e-01 9.58246946e-01 -9.47955549e-02
6.11139536e-01 -3.90130728e-01 -1.16726601e+00 4.31401551e-01
5.87064683e-01 -4.53871965e-01 9.14891958e-01 5.95051467e-01
1.29417092e-01 3.62555653e-01 -8.90190840e-01 6.27597988e-01
6.66046262e-01 -8.59536290e-01 -1.00195098e+00 3.41601968e-01
8.70041549e-01 -8.39619398e-01 -8.76976848e-01 -1.04757175e-01
2.45415568e-01 -7.73596466e-01 7.02122629e-01 -9.00518894e-01
2.97749005e-02 -1.66692972e-01 3.34757507e-01 -1.47946215e+00
-2.96492726e-01 -1.01655948e+00 -6.25214428e-02 1.54229212e+00
2.97456682e-01 -4.86840785e-01 8.85632575e-01 9.52436388e-01
-1.45246893e-01 -1.26668632e-01 -9.39274371e-01 -3.62861395e-01
-3.55953909e-02 1.22864403e-01 5.45078158e-01 1.15082824e+00
7.73457944e-01 1.03261268e+00 -9.41342592e-01 -2.08548680e-01
3.19783157e-03 1.05143011e-01 9.66968477e-01 -1.37217474e+00
8.39680880e-02 -2.72386611e-01 2.63390660e-01 -1.86198926e+00
3.00628185e-01 -3.41519326e-01 4.14820999e-01 -1.25091636e+00
4.03723903e-02 -5.61060667e-01 -6.82431236e-02 4.68630165e-01
-6.38012290e-01 -5.30196011e-01 1.81251511e-01 3.34960788e-01
-1.10601532e+00 8.14720392e-01 1.30746818e+00 -1.25417992e-01
-6.05750144e-01 2.91552573e-01 -6.66775048e-01 4.65165704e-01
9.85659182e-01 -2.08269924e-01 -4.56294119e-01 9.57831815e-02
-4.12649751e-01 6.15821362e-01 -4.44570780e-01 -5.08959711e-01
6.63480341e-01 -3.17555100e-01 -2.62967139e-01 -7.09193230e-01
6.05854690e-01 -7.24774182e-01 -5.64262927e-01 3.72536957e-01
-6.85034752e-01 -1.34472087e-01 3.73445481e-01 3.35642308e-01
-1.83197096e-01 -2.95515448e-01 6.73874259e-01 -3.08776498e-01
-1.22291374e+00 1.08953647e-01 -7.86152065e-01 4.49309617e-01
6.66695356e-01 1.07196160e-01 -5.82517207e-01 -5.19713759e-01
-7.03631222e-01 7.57538319e-01 -8.02880302e-02 7.65173435e-01
1.57587767e-01 -1.34789646e+00 -4.50482965e-01 9.73694101e-02
3.94401789e-01 -6.11070264e-03 5.95438659e-01 3.93646896e-01
1.26839146e-01 6.96926236e-01 -3.13426167e-01 -6.59564555e-01
-1.61422348e+00 -2.46898551e-02 3.46097857e-01 -6.67022824e-01
-3.74687642e-01 7.06295669e-01 7.89187253e-02 -8.34826767e-01
2.85597861e-01 1.20295212e-02 -7.63700306e-01 4.42484558e-01
6.86972082e-01 1.40115365e-01 -1.34869069e-01 -7.54706085e-01
-2.98957616e-01 9.46079940e-02 -6.32907510e-01 -2.40681544e-01
7.49914289e-01 -7.97292531e-01 1.88900549e-02 7.69067168e-01
9.69533563e-01 -2.66733557e-01 -1.25953352e+00 -8.63118947e-01
3.20501387e-01 -4.09496695e-01 -1.26131475e-01 -8.12148690e-01
-5.05184293e-01 7.19969273e-01 2.99250513e-01 8.98729861e-01
6.94651723e-01 -9.38779265e-02 9.52914298e-01 4.06375796e-01
1.58282489e-01 -1.44607604e+00 2.71539956e-01 1.30038989e+00
5.41517258e-01 -1.65188646e+00 -4.56468999e-01 -3.27074200e-01
-1.13242710e+00 9.26596403e-01 1.34103692e+00 4.96167451e-01
4.97591108e-01 -1.19550332e-01 5.17444313e-01 -2.27747083e-01
-1.04551601e+00 -4.80554879e-01 -1.86922941e-02 3.63329858e-01
6.75575078e-01 -1.85474247e-01 -4.69457835e-01 6.50172532e-01
1.45130083e-01 -2.68045604e-01 3.27121317e-01 1.07367003e+00
-8.17893088e-01 -1.42152500e+00 -3.20755571e-01 2.96826899e-01
-3.56342703e-01 1.73673630e-02 -5.45838773e-01 5.76405764e-01
-6.55136049e-01 1.26813626e+00 1.17889784e-01 -3.82223338e-01
5.57652056e-01 5.19545436e-01 -2.54986793e-01 -8.19126070e-01
-9.88432229e-01 2.32923687e-01 5.10232449e-01 -2.86765307e-01
-3.75883013e-01 -4.76762384e-01 -1.21394646e+00 -5.23089468e-01
-4.23395783e-01 7.18590021e-01 1.40823916e-01 1.02155888e+00
4.16409016e-01 5.80990255e-01 8.69948387e-01 -4.69857484e-01
-6.03453159e-01 -1.37999880e+00 -6.57459795e-01 3.98153216e-01
4.41789180e-01 -8.46715987e-01 -3.60814556e-02 -3.46597940e-01] | [12.751171112060547, 7.848485946655273] |
0eb57beb-1870-4ee4-add0-21f537310a47 | towards-unsupervised-deep-graph-structure | 2201.06367 | null | https://arxiv.org/abs/2201.06367v1 | https://arxiv.org/pdf/2201.06367v1.pdf | Towards Unsupervised Deep Graph Structure Learning | In recent years, graph neural networks (GNNs) have emerged as a successful tool in a variety of graph-related applications. However, the performance of GNNs can be deteriorated when noisy connections occur in the original graph structures; besides, the dependence on explicit structures prevents GNNs from being applied to general unstructured scenarios. To address these issues, recently emerged deep graph structure learning (GSL) methods propose to jointly optimize the graph structure along with GNN under the supervision of a node classification task. Nonetheless, these methods focus on a supervised learning scenario, which leads to several problems, i.e., the reliance on labels, the bias of edge distribution, and the limitation on application tasks. In this paper, we propose a more practical GSL paradigm, unsupervised graph structure learning, where the learned graph topology is optimized by data itself without any external guidance (i.e., labels). To solve the unsupervised GSL problem, we propose a novel StrUcture Bootstrapping contrastive LearnIng fraMEwork (SUBLIME for abbreviation) with the aid of self-supervised contrastive learning. Specifically, we generate a learning target from the original data as an "anchor graph", and use a contrastive loss to maximize the agreement between the anchor graph and the learned graph. To provide persistent guidance, we design a novel bootstrapping mechanism that upgrades the anchor graph with learned structures during model learning. We also design a series of graph learners and post-processing schemes to model the structures to learn. Extensive experiments on eight benchmark datasets demonstrate the significant effectiveness of our proposed SUBLIME and high quality of the optimized graphs. | ['Shirui Pan', 'Hao Peng', 'Hongxu Chen', 'Daokun Zhang', 'Yu Zheng', 'Yixin Liu'] | 2022-01-17 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [ 1.57074928e-01 2.91916698e-01 -2.46608973e-01 -3.90972883e-01
-1.65145442e-01 -4.16668922e-01 3.34909499e-01 2.67728060e-01
-1.97251573e-01 5.98215997e-01 -1.15922101e-01 -2.35038474e-01
-3.40793341e-01 -9.93623316e-01 -7.51466990e-01 -8.08929145e-01
-9.54833627e-02 2.89217651e-01 3.03572595e-01 -1.53964922e-01
5.88450506e-02 3.16130370e-01 -9.35613930e-01 -2.16134623e-01
1.25839126e+00 8.65407825e-01 3.48997653e-01 1.68691337e-01
-3.38247776e-01 6.52979374e-01 -4.52704042e-01 -1.96435228e-01
2.73315787e-01 -4.96937066e-01 -4.71349120e-01 2.91714787e-01
8.04332271e-02 4.85173613e-03 -4.57767397e-01 1.26936710e+00
4.60889250e-01 4.82828952e-02 4.58489954e-01 -1.43344355e+00
-5.79303205e-01 7.76533663e-01 -6.60667479e-01 1.13204375e-01
-3.38340104e-02 3.06729436e-01 1.13913131e+00 -6.85209394e-01
4.80304807e-01 1.12925661e+00 6.95531547e-01 3.97713453e-01
-1.24504328e+00 -7.17970610e-01 6.58452094e-01 8.66440013e-02
-1.30284894e+00 -1.85835540e-01 1.34052598e+00 -2.49199659e-01
3.05675566e-01 -1.03413887e-01 6.27950728e-01 9.46269274e-01
-5.58603741e-02 6.64160311e-01 9.51021135e-01 -1.95760116e-01
2.57634044e-01 5.66976406e-02 1.07028469e-01 9.89892542e-01
4.16384995e-01 1.02888919e-01 -1.65136501e-01 2.20623445e-02
7.29939818e-01 -3.04226521e-02 -3.63342673e-01 -8.18134606e-01
-7.98270643e-01 6.96972787e-01 1.01305187e+00 1.07668601e-01
-1.24709837e-01 -3.76687162e-02 4.21561390e-01 3.75094265e-01
4.06923354e-01 2.61197627e-01 -3.31937462e-01 3.82017821e-01
-4.67143089e-01 -1.20298527e-01 6.65219247e-01 8.75348270e-01
9.82805550e-01 1.69770434e-01 -3.18116583e-02 7.89505303e-01
4.51825410e-01 1.59396723e-01 3.26602936e-01 -1.51125580e-01
7.40406692e-01 1.07326424e+00 -3.84076476e-01 -1.47836864e+00
-5.29989600e-01 -8.87302279e-01 -1.24043369e+00 -4.58667688e-02
2.11184993e-01 -2.04748183e-01 -9.92325664e-01 1.90969062e+00
4.82666016e-01 4.29099560e-01 -4.34525609e-02 8.05463254e-01
8.89715791e-01 6.48957312e-01 4.24675345e-02 -2.48868540e-01
5.68879545e-01 -1.07025051e+00 -4.96470064e-01 -3.75092268e-01
8.37429166e-01 -1.34247869e-01 1.31280625e+00 2.17195645e-01
-6.88846052e-01 -5.35018802e-01 -1.11773598e+00 3.54136497e-01
-2.45895505e-01 -4.49748486e-02 5.87143302e-01 3.80573928e-01
-9.40825284e-01 7.21970975e-01 -7.12209642e-01 -2.72415429e-01
5.36932826e-01 4.94268686e-01 -3.42439801e-01 -6.36577010e-02
-1.05980754e+00 3.52420539e-01 8.59546363e-01 4.80875164e-01
-8.90065551e-01 -3.16503346e-01 -9.24441576e-01 2.50839740e-01
8.59074473e-01 -4.26086992e-01 7.27333844e-01 -1.24332929e+00
-1.43129754e+00 5.42811036e-01 4.20883656e-01 -3.28971863e-01
3.99854422e-01 1.77134916e-01 -3.80991071e-01 -4.32895608e-02
-9.31935385e-03 2.16408804e-01 8.36446524e-01 -1.42094409e+00
-3.39888871e-01 -2.83019930e-01 1.70659021e-01 2.58169055e-01
-5.80825090e-01 -5.80432475e-01 -5.23428142e-01 -6.59806609e-01
3.21320236e-01 -7.94801831e-01 -4.27323490e-01 -1.75608322e-01
-6.55063510e-01 -2.43656531e-01 8.18057358e-01 -3.83160025e-01
1.43587005e+00 -2.21921587e+00 1.62141487e-01 6.12920463e-01
5.70161223e-01 3.60674709e-01 -3.98689121e-01 2.87040859e-01
-1.57498866e-01 2.60805815e-01 -5.00548363e-01 -1.34693414e-01
-2.69232780e-01 3.00672680e-01 3.81202884e-02 2.53368467e-01
2.98088312e-01 1.00032675e+00 -1.12827444e+00 -6.43921137e-01
6.71727806e-02 1.02450684e-01 -5.30185163e-01 4.97699171e-01
-2.74285644e-01 5.37190855e-01 -7.07714200e-01 4.35520738e-01
7.33747602e-01 -6.48669481e-01 4.73101556e-01 -3.35493892e-01
2.70159692e-01 6.23571686e-02 -1.21235251e+00 1.50163198e+00
-3.46987367e-01 1.65425222e-02 2.03188434e-01 -1.45653808e+00
1.22563148e+00 -9.14212689e-02 3.11572820e-01 -5.03233492e-01
8.27481076e-02 9.73616689e-02 2.89259285e-01 -4.32404488e-01
-1.49004698e-01 -1.87117327e-02 1.42207935e-01 3.19823861e-01
5.14265373e-02 8.73728991e-02 1.30495474e-01 3.61611098e-01
1.11604106e+00 3.22961621e-02 3.10552686e-01 -3.08122218e-01
7.19517469e-01 -2.63388038e-01 8.33317637e-01 5.50287187e-01
-1.51757315e-01 3.83510083e-01 7.65578091e-01 -4.65180784e-01
-6.76742494e-01 -9.11304116e-01 3.18358153e-01 8.49433184e-01
4.01868403e-01 -5.77813506e-01 -6.66345894e-01 -1.24203992e+00
-1.53904915e-01 1.91470325e-01 -4.50891823e-01 -5.47191918e-01
-6.82698488e-01 -9.01282072e-01 1.73703060e-01 3.74491841e-01
6.76997542e-01 -1.14132249e+00 2.20975369e-01 2.57005006e-01
1.42621532e-01 -1.00010240e+00 -5.41002333e-01 1.11319207e-01
-1.03892374e+00 -1.15681303e+00 -3.06130379e-01 -1.14409471e+00
1.16502404e+00 2.77356952e-01 1.04040504e+00 6.78027809e-01
2.25898877e-01 -2.37688199e-02 -3.12305391e-01 -1.14968549e-02
-3.73676568e-01 5.30357003e-01 -1.16173789e-01 2.74657041e-01
-1.25570849e-01 -1.03448772e+00 -4.90610808e-01 2.80629128e-01
-9.56910849e-01 2.90917784e-01 8.53522122e-01 1.05230117e+00
5.44600606e-01 3.43714505e-01 8.93945992e-01 -1.29624379e+00
8.40120554e-01 -5.89118481e-01 -8.04693758e-01 3.55710506e-01
-1.04027963e+00 3.62830609e-01 1.15339541e+00 -4.02976424e-01
-7.31076777e-01 -2.47515291e-02 6.44077137e-02 -4.54035908e-01
1.44213140e-01 1.04298377e+00 -6.27503157e-01 -1.85230136e-01
4.42040801e-01 2.73386955e-01 2.17079833e-01 -4.08034623e-01
1.90412760e-01 2.97851294e-01 3.52512866e-01 -6.47345603e-01
1.08561647e+00 7.36136883e-02 1.68198347e-01 -4.13262427e-01
-9.01104152e-01 -6.97607696e-02 -4.44316208e-01 -1.13894485e-01
3.26458603e-01 -6.10860229e-01 -5.26811182e-01 4.59638864e-01
-8.38627577e-01 -4.55198616e-01 -4.70378846e-02 2.96467930e-01
-1.70734733e-01 6.41530693e-01 -4.34284627e-01 -5.07375836e-01
-3.67274851e-01 -1.01931369e+00 5.81765711e-01 4.09193933e-01
3.73724401e-01 -1.14960921e+00 -5.33762500e-02 -9.09712464e-02
2.44877905e-01 3.94456178e-01 1.21671152e+00 -7.43532479e-01
-6.83226347e-01 -8.72012898e-02 -5.27427912e-01 4.71217215e-01
4.20208722e-01 -1.34853140e-01 -5.02895236e-01 -5.56463659e-01
-2.01127306e-01 -4.34731185e-01 6.84171379e-01 1.52400792e-01
1.31021714e+00 -3.65287423e-01 -4.34667915e-01 7.93921232e-01
1.42938340e+00 1.54056624e-01 3.28173250e-01 1.58050656e-01
1.19482422e+00 5.88469088e-01 3.41632277e-01 1.59914941e-01
3.15742970e-01 3.35848689e-01 6.14633262e-01 -2.42744744e-01
3.13545465e-02 -7.43891478e-01 8.38283524e-02 1.19133365e+00
8.16196725e-02 -4.16349739e-01 -8.61966550e-01 2.49704078e-01
-1.96999121e+00 -3.89838576e-01 1.76246852e-01 2.19589019e+00
6.60770297e-01 6.30694807e-01 4.85235313e-03 -8.93457886e-03
1.03491139e+00 4.29432392e-01 -8.98603141e-01 2.18096197e-01
7.55803753e-03 3.76703292e-02 2.32415602e-01 4.27395344e-01
-9.68994617e-01 1.06838250e+00 5.11930752e+00 8.07322621e-01
-1.26416111e+00 -2.24395514e-01 6.95860445e-01 5.14567852e-01
-4.57367420e-01 2.92430222e-01 -4.85051185e-01 6.15363240e-01
4.80772287e-01 -1.43555611e-01 5.34762681e-01 8.94247830e-01
2.50165313e-01 4.43777353e-01 -9.71993208e-01 1.06416690e+00
-1.30404070e-01 -1.11091542e+00 2.63546228e-01 8.34824110e-04
6.67091489e-01 -1.50861844e-01 -1.61306411e-01 5.71504533e-01
2.83061802e-01 -9.76443052e-01 2.21772686e-01 3.00634146e-01
6.75083995e-01 -7.25810826e-01 7.28156686e-01 5.51434755e-01
-1.46209979e+00 -1.24841526e-01 -3.05028766e-01 7.12309405e-02
2.43346766e-02 6.84778512e-01 -8.00549746e-01 9.21655893e-01
3.88202310e-01 1.02000308e+00 -7.74470150e-01 1.01059687e+00
-4.58306342e-01 8.10544372e-01 -2.85089433e-01 -1.66737899e-01
2.99434423e-01 -5.27397454e-01 5.03397524e-01 7.26022243e-01
3.18235196e-02 -3.73944007e-02 6.64841712e-01 9.57039535e-01
-3.67981642e-01 4.04999703e-01 -6.91238523e-01 -2.80296177e-01
5.51559627e-01 1.40228069e+00 -9.27990615e-01 -7.48785883e-02
-3.18029344e-01 6.19930923e-01 8.59440863e-01 4.79519904e-01
-6.57451093e-01 -4.78290558e-01 -4.41740081e-02 2.80428082e-01
9.58153009e-02 -1.24502972e-01 -5.32031693e-02 -1.13085079e+00
2.79551238e-01 -9.61887240e-01 4.75539982e-01 -3.60064298e-01
-1.56616211e+00 6.46955371e-01 -1.18465543e-01 -1.22952032e+00
1.36977494e-01 -4.17806268e-01 -1.00146163e+00 5.20030677e-01
-1.49951911e+00 -1.08823037e+00 -6.13673985e-01 5.65721154e-01
1.12271003e-01 -1.98064432e-01 3.18171442e-01 2.90498883e-01
-8.21937919e-01 6.77867353e-01 -6.69955537e-02 4.15993661e-01
5.16072035e-01 -1.24150109e+00 4.06035036e-01 7.87292361e-01
9.16257575e-02 5.28425872e-01 3.75004500e-01 -7.86089361e-01
-1.27828026e+00 -1.37556410e+00 2.28573084e-01 1.50743827e-01
7.10765064e-01 -5.24433553e-01 -1.06682289e+00 6.93089724e-01
-1.64735168e-01 4.60333109e-01 1.83601871e-01 2.41028622e-01
-1.73534229e-01 -5.24938464e-01 -9.75590944e-01 6.78143322e-01
1.31225884e+00 -2.46232077e-01 -2.33550400e-01 3.70445997e-01
9.24101830e-01 -3.48351508e-01 -6.25946164e-01 7.61539757e-01
1.46491960e-01 -7.87605464e-01 6.30481303e-01 -5.78367531e-01
1.79115325e-01 -4.63051945e-01 3.52300107e-01 -1.47751987e+00
-3.51879478e-01 -7.42806435e-01 -1.63375929e-01 1.33757889e+00
3.67667645e-01 -9.66583431e-01 1.08324540e+00 2.29185566e-01
-2.02068731e-01 -1.10973287e+00 -5.78940570e-01 -6.57570064e-01
-2.24343881e-01 3.62838246e-02 7.03352630e-01 1.06196201e+00
-2.44269326e-01 8.29589307e-01 -3.27458680e-01 2.23055169e-01
5.72970152e-01 1.04871586e-01 9.45973635e-01 -1.44230199e+00
-4.39033985e-01 -3.81547451e-01 -4.86332208e-01 -1.18619728e+00
3.59802485e-01 -1.24438214e+00 5.42806126e-02 -1.52075303e+00
1.21937603e-01 -8.34562421e-01 -5.57717562e-01 5.09501040e-01
-4.83621925e-01 -3.82520825e-01 -1.22438492e-02 1.84839383e-01
-6.57528877e-01 8.26827526e-01 1.42087328e+00 -2.50251979e-01
-4.08694953e-01 4.98418100e-02 -7.48382926e-01 5.85516751e-01
7.92089581e-01 -5.74162364e-01 -9.38352346e-01 -2.47887954e-01
1.99509740e-01 1.10595776e-02 2.58205473e-01 -9.37513769e-01
2.42624313e-01 -9.25915837e-02 4.31475639e-02 -3.32970262e-01
-2.54000664e-01 -7.52488971e-01 1.35505749e-02 4.60514754e-01
-2.32010022e-01 8.37770775e-02 -1.42659605e-01 8.71365964e-01
-2.02078909e-01 -1.79100931e-01 7.65432656e-01 -1.17013104e-01
-5.76743543e-01 8.24759305e-01 3.42922568e-01 3.04530799e-01
8.04740906e-01 -1.61634788e-01 -2.98792899e-01 -3.16264451e-01
-6.13532543e-01 6.41339421e-01 3.63241881e-01 2.68219590e-01
5.70078433e-01 -1.31712449e+00 -5.40701747e-01 2.98880726e-01
1.92075118e-01 5.71774900e-01 9.54400748e-02 7.18467712e-01
-4.31418478e-01 -9.23746526e-02 7.32389987e-02 -6.47232652e-01
-8.30890357e-01 6.74913287e-01 3.46459270e-01 -6.64313853e-01
-7.81695366e-01 5.96801221e-01 4.41257715e-01 -8.09360683e-01
2.89227486e-01 -1.89413950e-01 -2.56231576e-01 -4.08814371e-01
2.29264367e-02 -4.37196307e-02 -3.51657905e-02 -2.40543544e-01
-8.26213956e-02 4.11855936e-01 -3.06327134e-01 4.56148475e-01
1.38762939e+00 -1.01964124e-01 -1.51158616e-01 3.20543706e-01
1.04648662e+00 -2.77819373e-02 -1.35496294e+00 -4.28528428e-01
2.27063686e-01 -1.60322949e-01 -9.85880345e-02 -5.08971334e-01
-1.45288181e+00 5.98967791e-01 4.01385754e-01 3.76477659e-01
1.14475715e+00 -1.58196270e-01 7.69300461e-01 5.19380450e-01
2.29354352e-01 -8.58822584e-01 3.21323395e-01 2.60062248e-01
5.74582100e-01 -1.31464124e+00 -6.08331412e-02 -6.18398070e-01
-2.94907093e-01 9.85799730e-01 9.78623390e-01 -3.29175144e-01
8.01096559e-01 -3.60490680e-02 -1.11771494e-01 -3.55198979e-01
-4.59465444e-01 -8.67779404e-02 1.37313902e-01 6.06463194e-01
8.90004784e-02 -2.38218959e-02 -3.10448647e-01 6.50765419e-01
2.87161712e-02 -1.91148147e-01 2.42927685e-01 7.71085024e-01
-2.95480371e-01 -1.17630529e+00 9.43649337e-02 6.59645438e-01
-2.20916439e-02 -8.24096501e-02 -4.27670091e-01 9.35119271e-01
2.64426172e-02 7.22509444e-01 -3.76890540e-01 -6.24145389e-01
3.29728901e-01 -3.24676394e-01 2.06261426e-01 -7.67646253e-01
-3.40439647e-01 -3.83758955e-02 -1.00582868e-01 -2.94367462e-01
-3.33919674e-01 -5.80104031e-02 -1.24560201e+00 -8.27367082e-02
-6.00539804e-01 4.68973070e-01 1.13732360e-01 9.74278629e-01
4.11226094e-01 4.91121352e-01 9.52667356e-01 -5.09000719e-01
-6.64515495e-01 -9.39697087e-01 -6.87133610e-01 4.90059733e-01
1.71641320e-01 -6.70877576e-01 -5.57518363e-01 -3.48465860e-01] | [7.332619667053223, 6.247817516326904] |
af8917d0-588e-4b7a-ad36-9d0cd9527c41 | dnabert-2-efficient-foundation-model-and | 2306.15006 | null | https://arxiv.org/abs/2306.15006v1 | https://arxiv.org/pdf/2306.15006v1.pdf | DNABERT-2: Efficient Foundation Model and Benchmark For Multi-Species Genome | Decoding the linguistic intricacies of the genome is a crucial problem in biology, and pre-trained foundational models such as DNABERT and Nucleotide Transformer have made significant strides in this area. Existing works have largely hinged on k-mer, fixed-length permutations of A, T, C, and G, as the token of the genome language due to its simplicity. However, we argue that the computation and sample inefficiencies introduced by k-mer tokenization are primary obstacles in developing large genome foundational models. We provide conceptual and empirical insights into genome tokenization, building on which we propose to replace k-mer tokenization with Byte Pair Encoding (BPE), a statistics-based data compression algorithm that constructs tokens by iteratively merging the most frequent co-occurring genome segment in the corpus. We demonstrate that BPE not only overcomes the limitations of k-mer tokenization but also benefits from the computational efficiency of non-overlapping tokenization. Based on these insights, we introduce DNABERT-2, a refined genome foundation model that adapts an efficient tokenizer and employs multiple strategies to overcome input length constraints, reduce time and memory expenditure, and enhance model capability. Furthermore, we identify the absence of a comprehensive and standardized benchmark for genome understanding as another significant impediment to fair comparative analysis. In response, we propose the Genome Understanding Evaluation (GUE), a comprehensive multi-species genome classification dataset that amalgamates $28$ distinct datasets across $7$ tasks, with input lengths ranging from $70$ to $1000$. Through comprehensive experiments on the GUE benchmark, we demonstrate that DNABERT-2 achieves comparable performance to the state-of-the-art model with $21 \times$ fewer parameters and approximately $56 \times$ less GPU time in pre-training. | ['Han Liu', 'Ramana Davuluri', 'Pratik Dutta', 'Weijian Li', 'Yanrong Ji', 'Zhihan Zhou'] | 2023-06-26 | null | null | null | null | ['splice-site-prediction', 'covid-variant-prediction', 'promoter-detection', 'transcription-factor-binding-site-prediction', 'dna-analysis', 'transcription-factor-binding-site-prediction-1', 'transcription-factor-binding-site-prediction-2', 'genome-understanding', 'core-promoter-detection', 'epigenetic-marks-prediction', 'data-compression'] | ['medical', 'medical', 'medical', 'medical', 'medical', 'medical', 'medical', 'medical', 'medical', 'medical', 'time-series'] | [ 6.24985456e-01 -1.48777992e-01 -2.66542524e-01 -1.41331956e-01
-8.14928651e-01 -9.90903795e-01 2.11175252e-02 5.34351707e-01
-6.87723219e-01 7.53404975e-01 -1.10351332e-01 -7.80439854e-01
-6.08625039e-02 -7.95609236e-01 -1.05353284e+00 -7.03195274e-01
-8.93581361e-02 5.38042903e-01 -1.58037201e-01 2.91753896e-02
5.85151970e-01 2.02566143e-02 -1.56042004e+00 3.18334818e-01
9.31246936e-01 4.97953802e-01 4.50662434e-01 9.61293459e-01
-3.54697704e-01 3.57332416e-02 -6.04613185e-01 -6.82922304e-01
5.97142279e-02 -5.46297789e-01 -8.92769814e-01 -3.62775117e-01
3.08432821e-02 -6.63366541e-02 6.69545457e-02 6.27476990e-01
6.10344112e-01 -2.61288553e-01 6.39942884e-01 -9.17585313e-01
-5.08173585e-01 9.91495788e-01 -4.23608452e-01 1.32772565e-01
1.04290538e-01 4.55314159e-01 1.13420677e+00 -5.48326492e-01
8.02286804e-01 9.17333066e-01 1.10186398e+00 3.15780818e-01
-1.08167601e+00 -3.20580781e-01 -1.77939638e-01 -1.41723650e-02
-1.47378206e+00 -1.72140628e-01 5.76562397e-02 -3.72840941e-01
1.92974675e+00 4.68605697e-01 6.96538329e-01 6.89306557e-01
3.68293792e-01 4.69049066e-01 5.73439062e-01 -5.23470104e-01
9.20673162e-02 -4.78864104e-01 2.50843108e-01 1.01089871e+00
4.78559494e-01 -2.10562155e-01 -4.19891864e-01 -4.85017270e-01
2.55587369e-01 -3.87440503e-01 -1.36535823e-01 1.11801913e-02
-1.19887531e+00 6.81441426e-01 -1.67901933e-01 -2.96653602e-02
-1.65406302e-01 3.76890808e-01 7.68063784e-01 -6.46015257e-02
2.68904537e-01 6.56656563e-01 -6.33506656e-01 -8.04044425e-01
-5.48067987e-01 1.36261210e-01 8.69737089e-01 9.46453035e-01
6.62767053e-01 -3.77644636e-02 2.62393564e-01 7.91900218e-01
-9.86689031e-02 3.33149821e-01 6.21183097e-01 -7.01231003e-01
6.97929487e-02 5.02756059e-01 -1.14245452e-01 -5.53695440e-01
-4.00206119e-01 -4.72971469e-01 -6.57241046e-01 -6.21442318e-01
4.51765120e-01 -2.49981787e-02 -8.44473004e-01 1.94863713e+00
4.27367657e-01 1.77345008e-01 2.06313819e-01 2.36149892e-01
4.55395430e-01 5.34281075e-01 1.83952600e-01 -3.03734113e-02
1.75358117e+00 -6.49344385e-01 -3.05455118e-01 -1.69082701e-01
1.19088805e+00 -9.23763454e-01 1.03016877e+00 4.22024786e-01
-9.10627902e-01 -3.68848503e-01 -1.06952608e+00 -3.53944689e-01
-5.94224751e-01 -1.42515361e-01 1.07961166e+00 1.16603184e+00
-7.86085010e-01 5.72522283e-01 -8.25541019e-01 -5.52672207e-01
1.38857692e-01 1.91127867e-01 -2.24095687e-01 -3.10380220e-01
-8.68523180e-01 8.58190775e-01 8.31847489e-01 -1.66539758e-01
-7.14109361e-01 -9.08412874e-01 -8.50947201e-01 1.48566112e-01
3.87733757e-01 -8.46576214e-01 8.01845908e-01 -1.65073484e-01
-1.12367678e+00 9.53647494e-01 -4.29144621e-01 -5.78275800e-01
-9.75403190e-02 -6.75894171e-02 9.38318744e-02 -9.38986838e-02
-2.55245537e-01 6.89998090e-01 5.48164025e-02 -7.17676103e-01
-6.19262934e-01 -2.61221249e-02 -1.65233478e-01 -3.85495611e-02
-4.46530059e-02 -8.65265876e-02 -5.60200512e-01 -6.14362538e-01
-7.40411133e-02 -9.05132234e-01 -9.19917971e-02 -5.68118155e-01
-3.47324222e-01 -2.17793003e-01 2.87470549e-01 -7.58386612e-01
1.03025031e+00 -1.94996154e+00 4.60887291e-02 2.03495789e-02
3.29170376e-02 4.48305696e-01 -5.06025016e-01 8.02532673e-01
-1.07867427e-01 5.26285768e-01 -4.60101604e-01 -1.87190354e-01
1.40471965e-01 3.43281537e-01 -2.51665056e-01 3.03585649e-01
2.02162832e-01 1.07032061e+00 -9.65852082e-01 -1.54062644e-01
-1.45029664e-01 5.08717477e-01 -9.35607493e-01 -2.67547481e-02
-5.59924185e-01 -1.58635587e-01 -1.09113917e-01 7.62686551e-01
6.62741780e-01 -3.99477750e-01 5.84627509e-01 -2.14015767e-01
-2.44983286e-01 4.78567153e-01 -6.63963139e-01 2.03882837e+00
-6.91134408e-02 2.57226944e-01 -2.24966526e-01 -1.11176109e+00
7.48873651e-01 1.31177045e-02 3.37415338e-01 -2.65052319e-01
1.17908455e-01 4.71968085e-01 2.16196865e-01 -4.36628878e-01
7.74159670e-01 3.71530205e-02 -1.13098472e-01 6.72753632e-01
-1.06743023e-01 -3.74606073e-01 4.94149357e-01 1.58059224e-01
1.27567327e+00 5.25176227e-01 4.76033598e-01 -2.59400904e-01
4.29190025e-02 3.38126808e-01 8.94673645e-01 8.91695321e-01
-6.17777407e-02 3.36045295e-01 4.77935374e-01 -5.22332251e-01
-1.34609163e+00 -7.11202204e-01 -6.23236150e-02 1.03951800e+00
-2.05617040e-01 -8.46516371e-01 -1.11128855e+00 -3.81767422e-01
9.04132053e-02 7.11281896e-01 -3.92712772e-01 -1.63469315e-01
-8.22852731e-01 -1.55597353e+00 1.43898106e+00 3.09718817e-01
2.30595931e-01 -6.16221428e-01 -8.85143399e-01 4.09976989e-01
-6.41414046e-01 -9.20829654e-01 -6.34028912e-01 6.06819272e-01
-4.78675902e-01 -1.24258840e+00 -3.49197507e-01 -8.71352375e-01
6.17445588e-01 2.42782101e-01 1.05424929e+00 2.31995374e-01
-8.23121130e-01 5.35847992e-02 -4.60833579e-01 -6.06116295e-01
-5.21073699e-01 2.61238605e-01 2.18217950e-02 -8.84804726e-01
6.47821665e-01 -3.26592803e-01 -5.20464599e-01 1.61331281e-01
-1.06811011e+00 2.84852326e-01 5.61305821e-01 1.19570088e+00
8.04840982e-01 -6.00987347e-03 8.76115203e-01 -1.04677153e+00
3.26086581e-01 -4.57099885e-01 -4.97758061e-01 6.22461200e-01
-5.79257965e-01 1.08622573e-01 6.31279826e-01 -2.14582548e-01
-6.93588912e-01 -1.68264762e-01 -5.60366988e-01 3.90648872e-01
1.58135563e-01 1.02106893e+00 -6.78093964e-03 -4.57353778e-02
4.34648812e-01 8.70370150e-01 5.08632958e-02 -4.17312652e-01
5.38041651e-01 6.74251437e-01 6.15178645e-01 -1.06923020e+00
2.90357560e-01 2.75424749e-01 -7.52572641e-02 -9.70670223e-01
-2.98715264e-01 -3.24954331e-01 -3.91924202e-01 4.86727238e-01
5.65573156e-01 -7.25281894e-01 -9.99011278e-01 7.22875893e-01
-9.88103449e-01 -3.95608962e-01 -1.53994903e-01 3.62030119e-01
-6.92846298e-01 9.83936667e-01 -9.46260452e-01 -6.50534272e-01
-6.45955443e-01 -1.11585951e+00 9.56834733e-01 -1.41282633e-01
-3.48937958e-01 -5.66203952e-01 -8.63181520e-03 4.93849665e-01
3.70230585e-01 1.27132535e-01 1.77108586e+00 -6.93146884e-01
-6.10806763e-01 -9.27456766e-02 -2.10911781e-01 6.15278296e-02
2.89525062e-01 1.63870864e-02 -7.47947633e-01 -4.74117130e-01
-2.85137206e-01 -5.60814083e-01 9.29431379e-01 1.90154195e-01
1.24572229e+00 -1.98975727e-02 -3.63842726e-01 1.06269574e+00
1.48201394e+00 3.36188912e-01 6.80980146e-01 4.08443630e-01
4.97351050e-01 4.29222375e-01 5.00439942e-01 4.66672003e-01
5.48901260e-01 3.05503130e-01 2.71968722e-01 1.46795318e-01
-5.10228984e-03 -2.99237043e-01 8.60882998e-02 1.20240271e+00
2.22595707e-02 -7.26779759e-01 -9.58058178e-01 6.63962483e-01
-1.35538459e+00 -7.56768942e-01 1.93935260e-01 2.31108904e+00
1.18314886e+00 1.23884464e-02 -1.12549886e-01 1.02214262e-01
3.55112642e-01 -1.08472332e-01 -6.66380584e-01 -6.17681980e-01
-4.37246442e-01 4.13588524e-01 7.30274737e-01 2.53933787e-01
-7.82205164e-01 1.11009336e+00 6.82730865e+00 1.22822773e+00
-8.72512221e-01 -2.74695396e-01 8.25436473e-01 4.26595584e-02
-4.07149076e-01 -2.74707545e-02 -9.89210904e-01 4.72320467e-01
1.15396988e+00 -3.47725272e-01 6.42899394e-01 6.66463017e-01
-3.86281945e-02 -1.35710880e-01 -1.22786903e+00 7.39738464e-01
9.50126424e-02 -1.68455243e+00 1.51463240e-01 8.03255960e-02
3.53771627e-01 2.14773744e-01 2.51772418e-03 2.41163865e-01
4.22263116e-01 -1.31773841e+00 5.09072363e-01 -2.55885976e-03
1.09069812e+00 -7.59029984e-01 7.92522490e-01 3.98546010e-01
-1.29448450e+00 3.31151754e-01 -7.77933002e-01 -3.78345251e-02
1.66847587e-01 6.75981462e-01 -1.27535129e+00 8.29156041e-01
3.70674312e-01 2.55295038e-01 -2.56241281e-02 8.22130799e-01
2.81195611e-01 7.45225191e-01 -5.22887170e-01 -6.49399459e-02
1.67254642e-01 -1.01648457e-01 2.38174021e-01 1.81659245e+00
5.74481368e-01 2.51255840e-01 4.40049693e-02 5.43930352e-01
-1.19788237e-01 6.83337301e-02 -3.61891866e-01 -5.44009686e-01
6.98855698e-01 8.26385558e-01 -9.69938338e-01 -4.24390584e-01
-2.21488401e-01 8.48506629e-01 4.42400336e-01 4.56864573e-02
-9.14061487e-01 -6.76405787e-01 9.87485409e-01 -4.50120211e-01
5.09590685e-01 -2.77423382e-01 -1.62188321e-01 -8.12547445e-01
-2.99570441e-01 -1.23480785e+00 5.04950225e-01 -3.93170267e-01
-1.03249621e+00 3.33744079e-01 -2.56314814e-01 -4.93161798e-01
-2.62962162e-01 -6.06155038e-01 -1.18634336e-01 1.04789555e+00
-1.52588570e+00 -1.10543776e+00 6.24466874e-02 -2.40941405e-01
5.20812571e-01 2.77757406e-01 1.25130892e+00 2.53260911e-01
-7.64461040e-01 1.01204002e+00 4.59319502e-01 -7.26415440e-02
5.58473885e-01 -1.12353003e+00 1.29049134e+00 7.05872715e-01
-3.00280273e-01 1.22139883e+00 6.03046596e-01 -8.16833138e-01
-2.05982804e+00 -1.11177194e+00 1.07713687e+00 -4.07321393e-01
3.22623104e-01 -5.25885463e-01 -8.60117674e-01 7.93577433e-01
-8.36434662e-02 -6.30735636e-01 1.21670926e+00 7.77550787e-02
-6.55360937e-01 3.76708478e-01 -1.16052413e+00 5.72178841e-01
1.33172774e+00 -5.05076647e-01 -3.77462178e-01 2.58434713e-01
1.08104670e+00 -5.58195710e-01 -9.84117448e-01 3.07394773e-01
8.67034554e-01 -6.63455725e-01 1.21292567e+00 -6.29311144e-01
3.47351640e-01 -1.10200860e-01 -2.65927404e-01 -1.30757380e+00
-2.91870594e-01 -6.19126618e-01 3.44086111e-01 1.25240993e+00
3.74631017e-01 -7.89940417e-01 9.27818418e-01 1.09839484e-01
-6.19471908e-01 -7.98024893e-01 -9.01891291e-01 -8.15692604e-01
2.29630426e-01 -4.43927974e-01 1.00565112e+00 1.05961502e+00
3.97477120e-01 -1.54371694e-01 -4.60464776e-01 -2.03769609e-01
5.07762194e-01 6.51278645e-02 8.37951660e-01 -5.71643472e-01
-5.63173532e-01 -3.93622756e-01 -1.22039050e-01 -1.24098384e+00
-2.19177678e-01 -1.15966463e+00 2.45635286e-01 -1.20913601e+00
5.49508333e-01 -5.46104848e-01 -2.64312536e-01 5.64698339e-01
-4.29281384e-01 1.48695782e-01 -2.03697681e-01 -8.57175887e-02
-2.08734363e-01 1.83829993e-01 6.79203629e-01 -4.86681834e-02
2.32427195e-01 -9.57539976e-01 -7.91252971e-01 5.14471769e-01
8.06422710e-01 -3.90862852e-01 -3.73303950e-01 -7.25510359e-01
4.72194850e-01 -1.62911534e-01 -2.44082272e-01 -6.24145627e-01
7.43801147e-02 -2.76525319e-01 3.50359753e-02 -7.15211689e-01
3.19213718e-01 -6.48509711e-02 5.45549750e-01 9.36451137e-01
-2.42200822e-01 3.39756399e-01 5.34200311e-01 4.64819819e-01
2.78584868e-01 -3.67984504e-01 4.70847845e-01 -3.91530395e-01
-5.66472054e-01 2.64545735e-02 -4.37793255e-01 7.13603124e-02
6.91172242e-01 -4.58231598e-01 -9.71100092e-01 4.50350791e-01
1.38530374e-01 -7.90319219e-03 7.83021986e-01 2.48949621e-02
3.53716999e-01 -4.23089087e-01 -8.19144785e-01 3.10898006e-01
2.46706590e-01 -2.80969143e-01 4.10060078e-01 3.72221649e-01
-1.09426761e+00 8.77989471e-01 -1.72125384e-01 -3.77489001e-01
-1.42784417e+00 5.70533633e-01 -2.12818990e-03 -3.11429024e-01
-3.19375724e-01 1.10606623e+00 1.34086907e-01 -6.40310884e-01
2.64749452e-02 -5.38110435e-01 3.68715614e-01 -2.80265331e-01
4.90883797e-01 3.50482613e-01 2.66475707e-01 -3.25308770e-01
-3.86827290e-01 6.39679372e-01 -2.20031694e-01 4.05795097e-01
1.19507873e+00 2.53935512e-02 -3.84375632e-01 3.69463526e-02
1.04237258e+00 -7.71666765e-02 -6.83951020e-01 -4.35698479e-02
4.56022210e-02 -4.08617556e-01 -5.04093468e-01 -9.09622908e-01
-4.73664045e-01 6.08433366e-01 2.27209494e-01 -1.98878467e-01
8.70233774e-01 -3.69624406e-01 1.17524016e+00 5.96930802e-01
6.15676343e-01 -9.44487631e-01 -1.66099936e-01 7.48355269e-01
1.94328889e-01 -7.98147321e-01 -5.00542074e-02 -7.46700048e-01
-2.67883241e-02 7.84812212e-01 3.15851063e-01 5.62908173e-01
-1.73977786e-03 5.56159198e-01 -2.13738412e-01 -2.44597113e-03
-1.01449490e+00 3.58784832e-02 -4.29969728e-01 6.87724710e-01
7.13646054e-01 9.98250023e-02 -8.12824547e-01 6.84912443e-01
-4.98541355e-01 1.61279112e-01 4.14972991e-01 1.21383178e+00
-5.42925477e-01 -1.50764775e+00 -1.37895226e-01 5.89580894e-01
-6.98351443e-01 -5.54627478e-01 -8.88943151e-02 5.96652627e-01
3.75698954e-01 8.50470662e-01 9.66963097e-02 -4.85820234e-01
-1.65720195e-01 4.14374202e-01 5.51448762e-01 -5.44751525e-01
-6.95841968e-01 2.60306820e-02 3.83460820e-01 -1.96023479e-01
9.12777036e-02 -5.24520993e-01 -1.37626719e+00 -7.75111914e-01
-2.67651737e-01 4.19806510e-01 7.89599240e-01 6.99524164e-01
7.68338025e-01 5.73806405e-01 3.21440995e-02 -4.24688458e-01
-5.85845649e-01 -7.47434378e-01 -4.18255895e-01 2.38950282e-01
-2.72616655e-01 -2.52099216e-01 -1.99701145e-01 2.78398395e-01] | [10.789790153503418, 7.466174602508545] |
1f421e53-facb-4d78-8789-b53901328f5a | saliency-aware-spatio-temporal-artifact | 2301.01069 | null | https://arxiv.org/abs/2301.01069v1 | https://arxiv.org/pdf/2301.01069v1.pdf | Saliency-Aware Spatio-Temporal Artifact Detection for Compressed Video Quality Assessment | Compressed videos often exhibit visually annoying artifacts, known as Perceivable Encoding Artifacts (PEAs), which dramatically degrade video visual quality. Subjective and objective measures capable of identifying and quantifying various types of PEAs are critical in improving visual quality. In this paper, we investigate the influence of four spatial PEAs (i.e. blurring, blocking, bleeding, and ringing) and two temporal PEAs (i.e. flickering and floating) on video quality. For spatial artifacts, we propose a visual saliency model with a low computational cost and higher consistency with human visual perception. In terms of temporal artifacts, self-attention based TimeSFormer is improved to detect temporal artifacts. Based on the six types of PEAs, a quality metric called Saliency-Aware Spatio-Temporal Artifacts Measurement (SSTAM) is proposed. Experimental results demonstrate that the proposed method outperforms state-of-the-art metrics. We believe that SSTAM will be beneficial for optimizing video coding techniques. | ['Tiesong Zhao', 'Chengdong Lan', 'Weiling Chen', 'Yang Zheng', 'Liqun Lin'] | 2023-01-03 | null | null | null | null | ['blocking'] | ['natural-language-processing'] | [ 1.99036330e-01 -7.38057256e-01 -6.22765394e-03 -9.86221731e-02
-4.74812627e-01 -3.11411053e-01 1.97302788e-01 3.50493252e-01
2.94798752e-03 5.47178686e-01 3.93099368e-01 -5.68583980e-02
-2.46999115e-02 -2.87652194e-01 -6.23727083e-01 -5.13716042e-01
-3.84995192e-01 -9.58378732e-01 7.27445662e-01 -5.74665740e-02
8.43366027e-01 2.08421811e-01 -1.74643099e+00 3.75454813e-01
1.24815583e+00 1.37741411e+00 5.57210207e-01 5.40457249e-01
1.85387120e-01 1.15374291e+00 -8.05356443e-01 -7.90223703e-02
6.56037703e-02 -5.96161962e-01 -3.83657128e-01 2.14979067e-01
2.66881049e-01 -5.28851032e-01 -2.05799475e-01 1.35915542e+00
2.89483249e-01 1.24430850e-01 3.63559127e-01 -1.38137555e+00
-7.87179649e-01 1.35424376e-01 -8.88487637e-01 8.69841754e-01
6.44236684e-01 3.19775850e-01 6.47442698e-01 -1.05942070e+00
2.91681170e-01 1.20359373e+00 5.49461842e-01 1.82155408e-02
-7.48208106e-01 -5.52485585e-01 -1.03627563e-01 9.36544716e-01
-1.31998420e+00 -5.53455651e-01 1.01615071e+00 -5.47783971e-01
5.19373596e-01 6.46572948e-01 5.50021112e-01 6.79315209e-01
6.84464037e-01 6.56635523e-01 1.15629232e+00 -2.70464838e-01
4.64518070e-01 1.16469234e-01 -1.40536457e-01 5.04807949e-01
2.23818511e-01 4.96319346e-02 -6.44849181e-01 -8.69670510e-02
8.66659284e-01 1.77364890e-02 -7.37535298e-01 -4.75949526e-01
-1.15745997e+00 3.03339630e-01 3.32174420e-01 1.67692780e-01
-3.67960691e-01 5.10698743e-03 3.16963106e-01 -1.30636245e-01
2.87853926e-01 4.56143647e-01 -1.78795889e-01 -3.33974898e-01
-1.11272275e+00 -8.17127228e-02 -6.82121664e-02 9.60809231e-01
5.58408260e-01 2.05872834e-01 -6.92986608e-01 9.68583465e-01
2.57916331e-01 6.21090591e-01 5.08106649e-01 -1.05026078e+00
2.98968911e-01 4.92000282e-01 5.79586983e-01 -1.51461327e+00
-1.63376153e-01 -3.21391314e-01 -6.66879237e-01 3.95999998e-01
6.86301887e-02 2.86430657e-01 -8.08853686e-01 1.30299282e+00
-1.02610767e-01 4.91312593e-01 -3.24554980e-01 1.37613928e+00
7.48093784e-01 7.47224748e-01 1.14276700e-01 -6.43124878e-01
1.37357354e+00 -9.40429688e-01 -1.24912548e+00 -3.85387354e-02
-3.79654653e-02 -1.08544552e+00 1.19630802e+00 5.17248869e-01
-1.58904243e+00 -8.04200590e-01 -1.41180813e+00 1.00985892e-01
1.02706477e-02 7.46495351e-02 2.25901991e-01 5.51238596e-01
-1.00070715e+00 4.11784381e-01 -7.52638578e-01 -1.47566274e-01
1.72851920e-01 -1.74411982e-01 2.10853100e-01 3.69136259e-02
-1.11493528e+00 8.41610909e-01 6.78799599e-02 3.14147770e-02
-1.08255911e+00 -6.97387218e-01 -7.15613246e-01 1.38356730e-01
4.42115515e-01 -2.73960680e-01 1.09731328e+00 -1.32289279e+00
-1.27496374e+00 2.84268558e-01 -4.76521671e-01 -1.63651466e-01
3.65468353e-01 -2.67438143e-01 -8.58357668e-01 4.17877078e-01
3.32515128e-02 2.56715655e-01 1.21247292e+00 -1.57530093e+00
-5.79396129e-01 5.59247052e-03 -8.88544694e-02 1.95792049e-01
-3.59758973e-01 4.12970096e-01 -4.59941208e-01 -8.98191452e-01
1.07957125e-01 -5.53827286e-01 1.79308385e-01 2.70480722e-01
-1.56382382e-01 1.35750383e-01 1.00409317e+00 -9.28617358e-01
1.83885682e+00 -2.25324917e+00 -4.11810093e-02 -1.69500336e-02
3.55762780e-01 5.94357371e-01 1.47537459e-02 1.18108459e-01
7.06144869e-02 1.14581294e-01 -1.08069927e-01 1.09925173e-01
-3.49918962e-01 -3.02815706e-01 -3.18175778e-02 3.95978630e-01
2.78609842e-01 7.20954955e-01 -1.19390333e+00 -5.35880148e-01
5.76534927e-01 3.79545838e-01 -3.20563942e-01 3.14424425e-01
1.27266884e-01 2.90448040e-01 -2.87949890e-01 8.38513732e-01
1.06896675e+00 -2.13213637e-01 -2.75828600e-01 -5.59738636e-01
-5.16677380e-01 6.00506142e-02 -1.01700163e+00 1.46116889e+00
-2.53271848e-01 8.44646871e-01 -1.31563902e-01 -1.42282933e-01
7.92967439e-01 2.24585310e-01 1.74318880e-01 -1.04055190e+00
1.68086872e-01 2.01105788e-01 -1.26487300e-01 -8.41416061e-01
7.70836890e-01 2.74634868e-01 4.41515654e-01 -1.29965231e-01
-1.57453805e-01 3.61412585e-01 2.39308953e-01 2.44403362e-01
1.10572577e+00 -3.84614393e-02 3.55575323e-01 -5.54342031e-01
6.81237042e-01 -5.44653535e-01 7.28971541e-01 4.02757287e-01
-7.66278148e-01 8.23330104e-01 4.26533490e-01 -1.95461437e-01
-1.04445565e+00 -1.28062117e+00 1.33393526e-01 8.08208227e-01
9.49412644e-01 -4.55287427e-01 -7.10996568e-01 -8.37722421e-02
-2.64814138e-01 6.74987435e-01 -4.32989776e-01 -2.83293277e-01
-2.82717019e-01 -2.90449083e-01 1.41771007e-02 4.39885765e-01
7.02135324e-01 -1.04262197e+00 -1.03730559e+00 1.54686287e-01
-6.03143811e-01 -1.01602674e+00 -8.58812332e-01 -4.12894398e-01
-7.37298608e-01 -1.04607260e+00 -9.58656132e-01 -5.88845372e-01
5.29065132e-01 1.05111563e+00 1.01516342e+00 1.81126535e-01
-2.58102924e-01 9.69514251e-02 -6.79126263e-01 -9.36994255e-02
-1.34176731e-01 -7.91637957e-01 -1.75422534e-01 3.47292632e-01
7.01681897e-02 -4.04924124e-01 -1.20734596e+00 5.55014610e-01
-1.06421053e+00 1.85453683e-01 5.20537436e-01 3.81008089e-01
3.54923129e-01 2.68095523e-01 8.87635797e-02 -1.46675944e-01
7.02352822e-01 -4.62372363e-01 -4.10365164e-01 2.13426143e-01
-3.69919896e-01 -2.26600870e-01 6.30532563e-01 -5.04238844e-01
-1.29023337e+00 -6.67236924e-01 5.41394055e-01 -7.48317003e-01
-1.50922388e-01 2.29362428e-01 -1.33059889e-01 -2.69640684e-01
5.11888385e-01 3.25015783e-01 -4.38500404e-01 -4.03275520e-01
-1.91096380e-01 6.84125781e-01 5.18470764e-01 -4.79078665e-02
5.50267816e-01 3.13980520e-01 -4.18219939e-02 -8.08474422e-01
-5.55548012e-01 -5.86954236e-01 -1.03744656e-01 -6.80139661e-01
9.06422853e-01 -9.71916974e-01 -5.55739820e-01 6.35879993e-01
-1.16322160e+00 2.79981196e-01 2.69275576e-01 3.92966092e-01
-3.87131274e-01 8.35719943e-01 -6.93607211e-01 -9.96770978e-01
-1.06992997e-01 -1.35868990e+00 8.57408106e-01 5.14881015e-01
7.11478759e-03 -5.81957102e-01 -2.80540973e-01 2.85055876e-01
7.08214164e-01 2.43905708e-01 7.03992009e-01 3.87285233e-01
-7.96781003e-01 4.48575258e-01 -5.46384037e-01 4.00936753e-01
3.62317413e-01 2.05761120e-02 -6.46416426e-01 -2.50241041e-01
2.93568164e-01 1.07477017e-01 4.77788240e-01 5.94949305e-01
1.32190812e+00 -3.07855308e-01 -5.40128313e-02 5.81373036e-01
1.55363977e+00 8.49857390e-01 1.19946694e+00 5.08957148e-01
4.93804783e-01 1.27107456e-01 1.00326204e+00 6.98178649e-01
5.26425093e-02 8.55041742e-01 6.13633633e-01 -2.91785896e-01
-3.49925101e-01 6.11730665e-03 4.95974779e-01 8.67862940e-01
-1.97205096e-01 -3.70503962e-01 -7.55548000e-01 6.19029462e-01
-1.69971132e+00 -8.68102849e-01 -4.10073072e-01 2.09259939e+00
6.60674155e-01 1.30491614e-01 -1.51891440e-01 2.87147701e-01
9.85695541e-01 2.59952605e-01 -4.38350856e-01 -4.30443496e-01
-2.52014875e-01 -1.66687697e-01 4.66291904e-01 3.02363753e-01
-9.99636531e-01 5.48325539e-01 5.90896034e+00 8.21169138e-01
-1.10737419e+00 1.12352677e-01 6.43245399e-01 -2.92147323e-02
-3.16864043e-01 -6.64071813e-02 -1.03516005e-01 1.06954718e+00
6.13029301e-01 -3.60766016e-02 6.11386895e-01 5.07436097e-01
9.38036501e-01 -5.16122222e-01 -6.98182225e-01 1.21738791e+00
4.02653813e-01 -8.90794873e-01 6.38040602e-02 -3.53250533e-01
6.89722598e-01 -4.76386338e-01 3.92211527e-01 -3.76579463e-01
-2.99717337e-01 -7.28411794e-01 1.13449931e+00 6.54154181e-01
7.03813851e-01 -6.34964108e-01 7.43313134e-01 -2.65738130e-01
-1.34927809e+00 -1.73242137e-01 -2.32762769e-01 7.54364952e-02
1.87076166e-01 6.51671708e-01 -7.53504485e-02 3.88985395e-01
8.56184900e-01 9.25468028e-01 -8.48913610e-01 1.65238941e+00
-1.53884351e-01 5.03770590e-01 2.10812807e-01 8.58043805e-02
-3.10721085e-03 -1.15579024e-01 7.89615631e-01 1.10234487e+00
6.28921807e-01 2.33418629e-01 -2.44546011e-01 9.55256701e-01
3.79016161e-01 8.85897651e-02 -4.68344465e-02 2.23649129e-01
5.35901308e-01 9.82165217e-01 -6.24132812e-01 -2.41849720e-01
-4.87726241e-01 1.20116186e+00 -3.19625765e-01 5.54689586e-01
-1.27776670e+00 -4.03853148e-01 7.20256150e-01 2.09593534e-01
4.10858542e-01 -1.10975832e-01 -3.50425392e-01 -1.20400786e+00
3.37676406e-01 -8.15542817e-01 -1.58516034e-01 -1.37856066e+00
-8.90868247e-01 4.25197154e-01 -2.04439819e-01 -1.84204829e+00
4.63135004e-01 -1.80410475e-01 -6.37897611e-01 6.50499940e-01
-1.53891885e+00 -5.94440341e-01 -7.87105203e-01 5.88099658e-01
8.72864425e-01 2.56892174e-01 1.27709627e-01 4.29190099e-01
-4.55241531e-01 4.17435497e-01 -8.71638954e-02 -3.25137079e-01
6.23496234e-01 -8.78907859e-01 1.04941549e-02 1.26652932e+00
-2.39628166e-01 4.19797868e-01 1.08917916e+00 -6.08033419e-01
-1.40228844e+00 -1.04721272e+00 5.92501760e-01 4.21273522e-03
5.41717708e-01 -1.93057470e-02 -9.79620814e-01 1.39845535e-01
2.25777850e-01 -1.38717711e-01 3.13956320e-01 -4.84243900e-01
-2.32465714e-01 -2.37413287e-01 -1.06510115e+00 6.11584663e-01
9.62012470e-01 -3.33723307e-01 -4.09572095e-01 -1.95493296e-01
5.68320870e-01 -1.12203017e-01 -6.61019444e-01 5.60099006e-01
4.15212154e-01 -1.59525204e+00 9.17036116e-01 1.97568282e-01
6.38407111e-01 -7.58929670e-01 1.39452331e-02 -1.16036153e+00
-8.20893705e-01 -5.28570473e-01 -2.71432042e-01 1.08795178e+00
-5.60494997e-02 -1.70288965e-01 1.71158299e-01 2.06789240e-01
-3.19078207e-01 -3.89587700e-01 -6.94267750e-01 -8.27440858e-01
-8.16640496e-01 -1.52484268e-01 4.59517688e-01 8.21712494e-01
2.61312425e-01 -3.15978713e-02 -7.28532791e-01 3.22522968e-01
6.43891215e-01 -1.00028262e-01 3.09632331e-01 -7.97059953e-01
-5.38698770e-02 -4.04640496e-01 -7.03137577e-01 -1.03712440e+00
-6.49919331e-01 -7.47163892e-02 9.64439064e-02 -1.28172791e+00
5.21678567e-01 4.88324389e-02 -8.58050585e-01 -4.87904325e-02
-5.70419848e-01 3.02531421e-01 3.22481900e-01 2.89165974e-01
-1.06502557e+00 6.71931326e-01 1.27260947e+00 2.13345909e-03
-9.74155888e-02 -5.17685235e-01 -4.27882552e-01 7.87644923e-01
8.13333392e-01 -1.51438549e-01 -3.67199481e-01 -4.57627177e-01
-6.06537350e-02 3.21548313e-01 4.64751273e-01 -1.44840312e+00
8.02930743e-02 -3.13731819e-01 4.24837023e-01 -6.41704857e-01
3.04573387e-01 -7.57212341e-01 1.14641272e-01 5.03820479e-01
-1.05990231e-01 9.83434096e-02 2.74023831e-01 5.69029331e-01
-5.08650959e-01 3.90257575e-02 1.05538476e+00 7.15959296e-02
-9.90339279e-01 -1.76805943e-01 -4.89020646e-01 -6.34259433e-02
1.24932754e+00 -4.57015157e-01 -5.03448963e-01 -6.03259504e-01
-2.85494566e-01 3.74368057e-02 7.35149384e-01 7.03718483e-01
1.11941326e+00 -1.31636691e+00 -5.37741065e-01 2.53598183e-01
2.56400615e-01 -9.65041876e-01 5.80983102e-01 9.22828496e-01
-6.49933398e-01 4.02851194e-01 -4.95516032e-01 -6.84268951e-01
-1.52713060e+00 8.25390399e-01 3.39083001e-02 1.07595623e-01
-8.91834199e-02 7.14587927e-01 3.55112791e-01 8.64368796e-01
2.80605227e-01 -6.43432021e-01 -2.78912842e-01 -4.00187641e-01
8.22219014e-01 8.13187420e-01 -1.17762133e-01 -7.29296386e-01
-4.36332434e-01 5.15766799e-01 1.47077858e-01 1.69149652e-01
7.71712422e-01 -6.03923559e-01 -2.68220231e-02 5.19097090e-01
9.58436430e-01 4.91801463e-02 -1.46552002e+00 9.26334634e-02
-5.07229008e-02 -1.14145219e+00 1.77237511e-01 -9.82528925e-01
-1.08370090e+00 7.92046487e-01 1.01890922e+00 2.96807528e-01
1.71569085e+00 -3.14521998e-01 1.03262043e+00 -7.07535028e-01
6.15155935e-01 -1.00282073e+00 4.05904144e-01 8.64119232e-02
1.02185690e+00 -1.14205170e+00 1.04567371e-01 -7.04998553e-01
-6.35525823e-01 8.26991677e-01 6.01487637e-01 1.97730418e-02
3.08973312e-01 9.52356085e-02 3.36321676e-03 1.23901501e-01
-7.02477992e-01 -9.01497826e-02 6.16105556e-01 4.08322573e-01
3.15257847e-01 -3.77792008e-02 -5.54610133e-01 3.30204368e-01
2.65111208e-01 1.74462005e-01 7.40166664e-01 9.35640574e-01
-6.85251117e-01 -4.69092906e-01 -6.39065742e-01 2.92547315e-01
-5.93357921e-01 -2.53958434e-01 -1.47376314e-01 1.99043944e-01
5.49455643e-01 1.51811373e+00 5.22037223e-02 -5.43244839e-01
3.68751943e-01 -5.37744641e-01 3.82825911e-01 -4.52830270e-02
-3.83305520e-01 2.20483273e-01 -1.96499899e-01 -9.80688930e-01
-5.10965466e-01 -4.06953841e-01 -8.96007180e-01 -2.14259699e-01
-3.77283961e-01 2.61873174e-02 5.72221994e-01 5.42837083e-01
3.93453270e-01 6.86093688e-01 6.23762906e-01 -7.66951680e-01
-8.40277225e-03 -1.01745391e+00 -7.19794571e-01 8.52809608e-01
6.19935393e-01 -9.13945735e-01 -6.06632531e-01 2.79693455e-01] | [11.719966888427734, -1.9278206825256348] |
e88c56ee-a322-4d04-b3f1-f9187716abb6 | environmental-sound-classification-on | null | null | https://github.com/jonnor/ESC-CNN-microcontroller/blob/master/README.md#abstract | https://github.com/jonnor/ESC-CNN-microcontroller/releases/download/print1/report-print1.pdf | Environmental Sound Classification on Microcontrollers using Convolutional Neural Networks | Noise is a growing problem in urban areas, and according to the WHO is the second environmental cause of health problems in Europe. Noise monitoring using Wireless Sensor Networks are being applied in order to understand and help mitigate these noise problems. It is desirable that these sensor systems, in addition to logging the sound level, can indicate what the likely sound source is. However, transmitting audio to a cloud system for classification is energy-intensive and may cause privacy issues. It is also critical for widespread adoption and dense sensor coverage that individual sensor nodes are low-cost. Therefore we propose to perform the noise classification on the sensor node, using a low-cost microcontroller.
Several Convolutional Neural Networks were designed for the STM32L476 low-power microcontroller using the Keras deep-learning framework, and deployed using the vendor-provided X-CUBE-AI inference engine. The resource budget for the model was set at maximum 50% utilization of CPU, RAM, and FLASH. 10 model variations were evaluated on the Environmental Sound Classification task using the standard Urbansound8k dataset.
The best models used Depthwise-Separable convolutions with striding for downsampling, and were able to reach 70.9% mean 10-fold accuracy while consuming only 20% CPU. To our knowledge, this is the highest reported performance on Urbansound8k using a microcontroller. One of the models was also tested on a microcontroller development device, demonstrating the classification of environmental sounds in real-time.
These results indicate that it is computationally feasible to classify environmental sound on low-power microcontrollers. Further development should make it possible to create wireless sensor-networks for noise monitoring with on-edge noise source classification. | ['Jon Nordby'] | 2019-05-15 | null | null | null | n-a-2019-5 | ['environmental-sound-classification', 'sound-classification'] | ['audio', 'audio'] | [ 1.57378048e-01 -2.62244344e-01 5.43702960e-01 -3.30143839e-01
-4.80757356e-01 -1.29453853e-01 -1.95245430e-01 2.27022871e-01
-7.82415271e-01 3.37701321e-01 -2.72157103e-01 -4.97275710e-01
-1.51926070e-01 -1.24741411e+00 -3.17727834e-01 -7.97726333e-01
-3.19226831e-01 -6.11662157e-02 3.36230665e-01 9.51611772e-02
-3.00665289e-01 5.19289732e-01 -2.05708456e+00 1.64221391e-01
1.55596614e-01 1.50011253e+00 3.92801195e-01 1.11700869e+00
2.70047933e-01 5.56254864e-01 -1.19256699e+00 1.89541727e-01
2.32085466e-01 -2.72941142e-02 -3.66871715e-01 -9.29544568e-01
2.70116568e-01 -5.59209883e-01 1.65760610e-02 8.68132293e-01
1.08238697e+00 -9.19505879e-02 1.52613211e-04 -1.40378821e+00
4.69694525e-01 7.40720212e-01 3.74858864e-02 3.30876946e-01
1.26905024e-01 2.33336747e-01 4.99810219e-01 -1.38110384e-01
-2.58811355e-01 8.27812195e-01 1.07535183e+00 3.26635450e-01
-9.06041265e-01 -1.33850384e+00 -5.43170691e-01 2.75055468e-01
-1.53741741e+00 -4.99463737e-01 7.91450143e-01 -5.70424274e-02
1.50969541e+00 6.40347183e-01 9.40546155e-01 9.83805716e-01
2.70173788e-01 -1.88211903e-01 1.23975289e+00 -5.02064824e-01
9.28301692e-01 -1.08437449e-01 -2.50551570e-02 2.56325632e-01
6.71689808e-01 1.02181964e-01 -5.41806757e-01 -4.26055014e-01
5.79403341e-02 -1.28130034e-01 -6.95726424e-02 5.44921339e-01
-5.75700223e-01 6.09806418e-01 4.23337400e-01 4.14810985e-01
-3.86277556e-01 1.06682265e+00 5.70313990e-01 3.46658766e-01
5.59374034e-01 3.60533506e-01 -7.40492225e-01 -7.65338719e-01
-9.45762634e-01 -1.26626091e-02 1.10506189e+00 5.34706295e-01
6.84289873e-01 3.62034351e-01 7.39111781e-01 5.32121956e-01
5.84643006e-01 9.15063202e-01 4.69915956e-01 -9.07631934e-01
8.20082575e-02 1.24735855e-01 -2.84872085e-01 -1.02538896e+00
-9.28029656e-01 -3.04918021e-01 -8.36938620e-01 6.70736670e-01
2.85123587e-01 -6.53162479e-01 -7.67113090e-01 1.37140226e+00
1.31972760e-01 4.05012161e-01 1.30291373e-01 5.33127606e-01
1.02982438e+00 6.35273099e-01 3.01980257e-01 1.75964043e-01
1.37351370e+00 -1.78655803e-01 -6.27034307e-01 -6.88788295e-02
4.71657962e-01 -5.37298799e-01 8.80463898e-01 5.94817281e-01
-4.50304329e-01 -5.76851249e-01 -1.33831012e+00 3.55250657e-01
-7.14022636e-01 -2.14996487e-02 3.80946875e-01 1.61525488e+00
-1.10371029e+00 5.23970425e-01 -1.34958327e+00 -3.87973934e-01
4.33573842e-01 6.82489395e-01 8.91632140e-02 2.12053075e-01
-1.37728167e+00 7.42128193e-01 5.36777684e-03 2.34813541e-01
-6.25532925e-01 -9.54488635e-01 -4.96677727e-01 1.92078710e-01
-1.00656740e-01 -2.49919787e-01 1.34780014e+00 -5.61590552e-01
-1.65280616e+00 4.37005237e-02 3.75494719e-01 -6.28060460e-01
2.02021420e-01 -1.52418688e-01 -8.71156514e-01 1.14397146e-01
-1.45634338e-01 3.99055779e-01 6.96834207e-01 -7.99552917e-01
-8.02949905e-01 -1.36325911e-01 -1.09963648e-01 -4.31104630e-01
-6.88930392e-01 -7.82485008e-02 4.26500082e-01 -8.33270103e-02
-3.43553007e-01 -7.93934047e-01 -3.42253774e-01 4.55893837e-02
3.67965400e-02 -5.69371358e-02 1.25185442e+00 -6.77387178e-01
1.16365767e+00 -1.99824798e+00 -1.05679321e+00 7.46535420e-01
-1.25834346e-02 5.40673792e-01 1.52719930e-01 7.30661824e-02
-1.51945695e-01 1.73498005e-01 -1.74607001e-02 -1.46569267e-01
-1.75281450e-01 3.37847292e-01 2.19350636e-01 3.55999202e-01
-7.00081885e-02 1.47115573e-01 -6.02228820e-01 1.11013837e-02
4.64704454e-01 8.14147770e-01 -2.79439718e-01 -3.30487229e-02
-2.51444280e-02 -1.80358246e-01 -2.32011259e-01 5.07572114e-01
7.79667974e-01 5.58938563e-01 -1.80351093e-01 -2.83825099e-01
-3.72635841e-01 4.02030528e-01 -1.61620307e+00 1.33083606e+00
-1.04336584e+00 8.86204541e-01 6.12549365e-01 -8.24013114e-01
1.00815868e+00 5.33304751e-01 4.33918446e-01 -6.75579071e-01
3.81672591e-01 5.83571792e-01 5.47705218e-02 -8.67180645e-01
1.21803418e-01 9.60976034e-02 -3.45916189e-02 2.90069640e-01
-3.69109780e-01 -4.28697258e-01 -1.91890925e-01 -3.87067854e-01
1.57201469e+00 -2.85954922e-01 6.80600703e-02 -3.82044882e-01
7.03373849e-02 -2.41689421e-02 1.31647393e-01 6.33621752e-01
-1.75917521e-01 2.65441954e-01 6.04785094e-03 -4.44128901e-01
-6.09181702e-01 -4.92999047e-01 -3.06471199e-01 8.30941439e-01
-3.29547584e-01 -3.18088681e-01 -1.01980388e+00 4.78771748e-03
-1.48217857e-01 7.68693745e-01 -4.21364829e-02 -9.80781093e-02
-4.05740082e-01 -7.69173145e-01 1.18915892e+00 4.70278144e-01
7.41681576e-01 -9.83807921e-01 -1.83325732e+00 6.16129279e-01
2.69507408e-01 -8.56999397e-01 4.02357161e-01 1.11202037e+00
-3.90338004e-01 -9.01992679e-01 -5.63365184e-02 -5.39350748e-01
4.34794985e-02 1.17767729e-01 9.30029154e-01 1.39839470e-01
-6.28403008e-01 8.63206148e-01 -2.90995538e-01 -1.24878681e+00
-3.33579540e-01 1.27383664e-01 1.96310133e-01 -5.35858512e-01
6.80959165e-01 -1.00468516e+00 -6.38625562e-01 -1.61579214e-02
-7.53995538e-01 -5.38497746e-01 2.36402035e-01 2.36370340e-01
3.38943630e-01 8.27241540e-01 6.78259850e-01 -1.85980558e-01
5.37662625e-01 -5.59985161e-01 -7.70421386e-01 -2.74106324e-01
-5.09443104e-01 -3.66449893e-01 8.69498551e-01 -3.46129119e-01
-7.76043892e-01 4.73879576e-01 -7.13357270e-01 2.02709083e-02
-6.42517090e-01 2.68082380e-01 -4.29883927e-01 -2.42257312e-01
7.26745248e-01 -3.61455679e-01 -8.66100006e-03 -3.51059884e-01
-2.34250516e-01 1.37350798e+00 2.90149838e-01 -1.85910955e-01
4.98345107e-01 4.26125139e-01 3.84181321e-01 -1.74214399e+00
-8.73117819e-02 -3.00180137e-01 -1.26604944e-01 -3.37653339e-01
9.41235542e-01 -1.21887922e+00 -1.08014250e+00 7.29723811e-01
-1.01436305e+00 -5.45247734e-01 -2.29356095e-01 4.58951712e-01
2.14198735e-02 -1.66895509e-01 -4.79380973e-02 -1.20943272e+00
-8.04778934e-01 -9.74718690e-01 6.28445745e-01 3.72137964e-01
-4.99112189e-01 -5.88689566e-01 -7.00454563e-02 -2.12178696e-02
9.79241908e-01 1.47317976e-01 4.74417359e-01 -4.21661466e-01
-1.93529248e-01 -2.90072948e-01 1.60000101e-01 6.48459911e-01
5.43959737e-02 -5.05231991e-02 -1.57867491e+00 4.72699963e-02
2.76097268e-01 -2.11359441e-01 7.44363010e-01 4.85237509e-01
1.45794618e+00 2.32796394e-03 -2.47908190e-01 5.08669972e-01
1.53440654e+00 5.07818162e-01 6.15104258e-01 3.90172869e-01
4.78185415e-01 2.00392962e-01 1.59796447e-01 2.81784654e-01
1.64390355e-01 3.74987841e-01 9.12507832e-01 -8.95332098e-02
-2.47538432e-01 2.81100154e-01 4.03078526e-01 4.58017230e-01
9.00197625e-02 -2.54698694e-01 -9.34476793e-01 4.20378685e-01
-1.07059133e+00 -8.39418054e-01 -5.00349820e-01 1.88552320e+00
5.99315941e-01 1.30908072e-01 1.01747215e-02 8.55442822e-01
3.03911716e-01 -1.33413494e-01 -3.53773057e-01 -7.01525867e-01
3.87376457e-01 1.07169557e+00 8.03732276e-01 4.17503953e-01
-1.20969737e+00 1.31622940e-01 5.32400179e+00 6.60456538e-01
-1.58943594e+00 2.81842023e-01 4.84811604e-01 -4.65623617e-01
1.64035872e-01 -5.19600928e-01 -7.07128704e-01 7.12345600e-01
1.81707084e+00 3.21902156e-01 2.84591496e-01 1.18553460e+00
5.29838026e-01 -5.45598865e-01 -6.90979481e-01 8.69597137e-01
-2.98689574e-01 -9.15463984e-01 -8.51796210e-01 -1.32174075e-01
7.49355108e-02 3.34044635e-01 -2.37691537e-01 -2.60054097e-02
1.67375237e-01 -1.01375365e+00 6.03236675e-01 3.08293670e-01
7.34919608e-01 -1.15936303e+00 9.96783853e-01 1.54365510e-01
-1.42391908e+00 -2.29196876e-01 -4.27992821e-01 -6.22560143e-01
-2.24099398e-01 9.78218377e-01 -8.52439880e-01 1.05396517e-01
1.66630208e+00 3.46810780e-02 -4.91235703e-01 8.86499107e-01
6.31660223e-02 1.28724372e+00 -1.23367965e+00 -7.96082556e-01
-3.86003144e-02 2.38479227e-01 1.71122998e-01 1.25522530e+00
8.71066928e-01 6.93596601e-02 -1.45987704e-01 2.53668576e-01
3.03247422e-01 -2.35823780e-01 -4.47375089e-01 5.12774646e-01
5.81838667e-01 1.35151160e+00 -7.03567028e-01 1.07489705e-01
-4.13163304e-01 3.18117768e-01 -5.02372682e-01 -5.25086932e-02
-6.24913573e-01 -8.04751158e-01 1.19031358e+00 2.70496279e-01
3.01722407e-01 -9.11566466e-02 -6.10936284e-01 -2.79825423e-02
2.32765470e-02 -6.31766737e-01 -6.61137700e-02 -7.47722328e-01
-8.20021451e-01 6.41189694e-01 2.25509256e-02 -9.39355552e-01
-1.60627868e-02 -7.17843831e-01 -7.59874165e-01 7.96104789e-01
-1.48004222e+00 -9.91939783e-01 -7.74117827e-01 4.03679460e-01
1.16896100e-01 -9.74220876e-03 1.40532529e+00 4.97262150e-01
-3.17737997e-01 4.29905087e-01 1.12837240e-01 -6.14515282e-02
1.84718162e-01 -1.06892645e+00 2.03160420e-01 6.39358342e-01
-1.67183265e-01 6.45878986e-02 7.09308743e-01 -5.51214457e-01
-1.31793630e+00 -1.40483475e+00 6.41766548e-01 8.22796971e-02
6.41538858e-01 -5.76425910e-01 -6.68489993e-01 4.23327237e-02
3.53283614e-01 9.08985063e-02 8.60495985e-01 -3.00960481e-01
1.17269620e-01 -8.16010952e-01 -1.44473743e+00 3.36059123e-01
6.17417514e-01 -2.44013652e-01 4.11382020e-02 1.59491643e-01
6.00175440e-01 -1.38537288e-01 -8.16516459e-01 -9.50364545e-02
4.36313242e-01 -8.59140575e-01 7.88289726e-01 3.32331866e-01
1.14443280e-01 -4.77095068e-01 -3.14013898e-01 -1.35964978e+00
1.24945343e-02 -5.58322608e-01 3.94193418e-02 1.51492918e+00
3.46665919e-01 -7.26684868e-01 8.39195669e-01 6.66466296e-01
-1.36690468e-01 -3.66103560e-01 -1.66344273e+00 -6.29098296e-01
-1.98514104e-01 -1.43153334e+00 8.43689799e-01 2.35302567e-01
-2.15052694e-01 6.29012361e-02 -1.84881836e-02 5.81086993e-01
4.91147101e-01 -7.33975768e-01 5.84087670e-01 -1.37470186e+00
2.41457820e-02 -2.36310616e-01 -8.39307606e-01 -1.82701871e-01
-6.04723468e-02 -5.28191388e-01 2.13643461e-01 -1.27958226e+00
-6.05700195e-01 -6.60774171e-01 -1.63669705e-01 8.90876889e-01
3.26917857e-01 4.50207591e-01 -2.42005154e-01 -5.58421731e-01
6.41686767e-02 4.69589531e-02 2.63292670e-01 -3.03888381e-01
-1.04676485e-01 3.56311798e-01 -4.35596377e-01 8.57941508e-01
1.21876526e+00 -8.70530367e-01 -4.30929810e-01 -5.27747512e-01
3.00891757e-01 -4.60980713e-01 6.81403995e-01 -1.99775922e+00
5.83927870e-01 2.49575362e-01 3.30296695e-01 -5.18228829e-01
4.90839630e-01 -1.70495605e+00 5.49264848e-01 1.05611324e+00
9.64852199e-02 -1.26156107e-01 5.20421386e-01 4.15272862e-01
5.59797660e-02 -2.91701823e-01 8.02635014e-01 4.61675785e-02
-7.42798686e-01 -2.95441896e-01 -1.01295853e+00 -5.21855891e-01
1.02216065e+00 -1.31746635e-01 -1.51416197e-01 -4.69974339e-01
-3.14757049e-01 -1.55802920e-01 -1.19478039e-01 3.69031280e-02
4.82323855e-01 -9.20299053e-01 -3.46438974e-01 4.50267047e-01
-2.16766492e-01 1.65523842e-01 3.51078123e-01 2.56250769e-01
-9.42363918e-01 1.06962234e-01 -2.31255069e-02 -6.81772530e-01
-1.68928313e+00 -9.77593958e-02 7.37230241e-01 2.93327272e-01
-6.26848042e-01 8.76024663e-01 -1.06861484e+00 -2.72660077e-01
6.25894129e-01 -9.50513661e-01 -3.95065546e-02 1.00148745e-01
7.28262365e-01 8.99263561e-01 7.16991127e-01 4.20391411e-02
-7.42804766e-01 5.27834535e-01 7.39758611e-01 -1.25249708e-02
1.59073961e+00 1.64285704e-01 -1.31562352e-01 3.07200104e-01
1.28416073e+00 -6.00293800e-02 -8.91324103e-01 3.12103122e-01
-1.04603298e-01 -3.59166116e-02 6.16876066e-01 -8.33474755e-01
-1.33173788e+00 9.14622605e-01 1.44673443e+00 5.10118902e-01
1.63461065e+00 -3.68196964e-01 6.88541532e-01 7.19071984e-01
4.56823289e-01 -1.27340424e+00 -2.76172549e-01 3.35852623e-01
4.77570653e-01 -8.87122154e-01 4.92599383e-02 -1.46038085e-01
9.37882885e-02 1.40665245e+00 5.39498270e-01 3.32905240e-02
1.43791115e+00 1.39708257e+00 5.24082482e-01 -1.46117643e-01
-5.42760253e-01 -3.19747813e-02 -5.76876044e-01 1.17378271e+00
2.16958016e-01 4.12839234e-01 2.21054956e-01 8.91895056e-01
-6.93882108e-01 1.98642790e-01 5.63834488e-01 1.09825945e+00
-6.29826665e-01 -8.10194910e-01 -7.29690492e-01 7.07015514e-01
-6.77580714e-01 -3.47579867e-02 -2.10220534e-02 5.75399697e-01
7.44860053e-01 1.63870251e+00 4.56848949e-01 -6.66627705e-01
5.09430289e-01 -2.85807997e-02 -2.23936215e-01 -2.67766684e-01
-1.15136480e+00 -5.76713681e-02 2.94772357e-01 -4.91783291e-01
-3.74399275e-01 -6.34911120e-01 -1.17573476e+00 -4.18328524e-01
-2.53038436e-01 9.33636650e-02 1.35329080e+00 5.62171161e-01
2.33975753e-01 1.07194364e+00 4.98426825e-01 -8.10866296e-01
-3.31045330e-01 -1.09157813e+00 -6.90234005e-01 -3.51651698e-01
3.15070331e-01 -2.85178900e-01 -7.22482502e-01 -1.67238832e-01] | [14.620237350463867, 5.474834442138672] |
53bdb159-a1b8-498d-aab0-5b9896b30d6c | relational-reasoning-over-spatial-temporal | null | null | https://ieeexplore.ieee.org/abstract/document/9750933 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9750933 | Relational Reasoning Over Spatial-Temporal Graphs for Video Summarization | In this paper, we propose a dynamic graph modeling approach to learn spatial-temporal representations for video summarization. Most existing video summarization methods extract image-level features with ImageNet pre-trained deep models. Differently, our method exploits object-level and relation-level information to capture spatial-temporal dependencies. Specifically, our method builds spatial graphs on the detected object proposals. Then, we construct a temporal graph by using the aggregated representations of spatial graphs. Afterward, we perform relational reasoning over spatial and temporal graphs with graph convolutional networks and extract spatial-temporal representations for importance score prediction and key shot selection. To eliminate relation clutters caused by densely connected nodes, we further design a self-attention edge pooling module, which disregards meaningless relations of graphs. We conduct extensive experiments on two popular benchmarks, including the SumMe and TVSum datasets. Experimental results demonstrate that the proposed method achieves superior performance against state-of-the-art video summarization methods. | ['Jie zhou', 'Jiwen Lu', 'Yucheng Han', 'Wencheng Zhu'] | 2022-04-06 | null | null | null | ieee-transactions-on-image-processing-2022-4 | ['supervised-video-summarization', 'relational-reasoning'] | ['computer-vision', 'natural-language-processing'] | [ 1.39163032e-01 1.76026151e-01 -4.56202596e-01 -2.70099103e-01
-4.63513047e-01 -1.44880980e-01 5.17143488e-01 4.52542663e-01
-2.19490439e-01 3.50253910e-01 7.98605144e-01 1.06676131e-01
-3.68316919e-02 -9.63024855e-01 -8.33499849e-01 -3.57095301e-01
-3.41830581e-01 -7.51844049e-02 6.23708010e-01 -3.01660281e-02
3.46804410e-01 2.41188243e-01 -1.27311718e+00 5.16583681e-01
8.96856964e-01 9.37334001e-01 1.80901766e-01 7.02619910e-01
-1.28799334e-01 1.54979539e+00 -6.25137508e-01 -3.65647525e-01
-2.23384336e-01 -6.80060089e-01 -8.04183483e-01 3.88188660e-01
6.19919360e-01 -6.21426821e-01 -1.23406029e+00 1.11898768e+00
3.24663967e-01 6.32603407e-01 4.65815485e-01 -9.33743536e-01
-1.11981440e+00 7.15214789e-01 -9.61009264e-01 6.90914869e-01
5.32546043e-01 3.01448971e-01 1.38487923e+00 -7.79642522e-01
7.04908550e-01 1.26403570e+00 4.31170434e-01 9.48340967e-02
-9.39351678e-01 -4.27810907e-01 7.07345068e-01 6.39263570e-01
-1.26943398e+00 -2.52868623e-01 1.01507688e+00 -1.63186625e-01
1.34722853e+00 1.18000261e-01 1.03106463e+00 7.24457204e-01
2.43599832e-01 1.17081487e+00 1.98774382e-01 6.01564348e-02
9.44289416e-02 -7.75313139e-01 3.15192372e-01 1.11933172e+00
2.06968859e-01 -6.60445333e-01 -7.81756043e-01 4.29325998e-02
8.33990157e-01 2.96380252e-01 -3.73844147e-01 -3.84490043e-01
-1.19564748e+00 6.02751732e-01 9.92867351e-01 2.69034386e-01
-8.46046388e-01 5.25824666e-01 6.54647529e-01 4.07765359e-02
7.01925755e-01 8.28234777e-02 1.10673577e-01 3.19158703e-01
-1.19161117e+00 1.56864971e-01 4.38457459e-01 1.25320232e+00
7.26582289e-01 -6.01107925e-02 -8.48737121e-01 7.25959778e-01
6.00119159e-02 -1.02868758e-01 3.41463238e-01 -8.75741780e-01
7.13936031e-01 1.03293765e+00 -3.52436066e-01 -1.36184227e+00
-2.97812909e-01 -4.93894160e-01 -1.01789939e+00 -5.56082249e-01
-3.32337737e-01 2.11123526e-01 -1.04191625e+00 1.25655937e+00
7.72794411e-02 6.69868290e-01 -4.69017699e-02 1.00638604e+00
1.52861619e+00 9.08313572e-01 1.95423275e-01 -4.60757107e-01
1.15245259e+00 -1.52438068e+00 -8.58487785e-01 -1.57026097e-01
5.66847801e-01 -2.04633474e-01 8.99092197e-01 -1.71851516e-01
-1.34936666e+00 -4.28547531e-01 -1.00918591e+00 -4.17078733e-01
-2.82266773e-02 -3.89949530e-02 6.10222101e-01 -2.16930509e-01
-1.31811583e+00 7.29620039e-01 -8.28435838e-01 -5.70844114e-01
8.72588813e-01 1.69690385e-01 -2.50519603e-01 -2.28281155e-01
-1.02045071e+00 4.40085441e-01 5.97659647e-01 3.52947861e-02
-8.52770269e-01 -4.59162414e-01 -1.27662766e+00 5.34257829e-01
6.50066674e-01 -1.04570353e+00 1.07939243e+00 -7.95164287e-01
-1.14182472e+00 6.68770373e-01 -3.74622107e-01 -5.68372071e-01
8.91470388e-02 -2.83549070e-01 -1.47054300e-01 6.65230274e-01
2.28302553e-01 6.05006576e-01 5.82440197e-01 -9.32590425e-01
-5.58552623e-01 -2.44294882e-01 3.66381288e-01 4.32187617e-01
-5.03832340e-01 -7.37166405e-03 -1.03786182e+00 -8.67257595e-01
1.58680201e-01 -3.21293175e-01 -3.56168091e-01 -3.64029795e-01
-5.93258142e-01 -3.64519000e-01 8.31945360e-01 -9.58308339e-01
1.68630695e+00 -2.00127554e+00 5.50353885e-01 -4.61218990e-02
6.62404060e-01 1.11381494e-01 -4.10671622e-01 4.43292230e-01
5.15372306e-03 1.51491627e-01 -1.14905141e-01 -3.09846252e-01
-2.73283064e-01 2.53424495e-02 -1.77942976e-01 2.88768023e-01
2.97802001e-01 1.36333311e+00 -1.15816653e+00 -7.81060338e-01
2.26829171e-01 3.29282522e-01 -7.59429455e-01 1.97504506e-01
-3.44566107e-01 2.44249590e-02 -7.68426776e-01 4.97883081e-01
4.26397085e-01 -5.18786073e-01 1.73992097e-01 -4.43250626e-01
1.19440235e-01 3.90424758e-01 -5.36527812e-01 2.10583663e+00
-1.50697753e-01 6.26174510e-01 -4.98870224e-01 -1.17305350e+00
6.12297893e-01 -6.51583821e-02 6.04547620e-01 -8.00529182e-01
1.57702118e-01 -3.72725904e-01 -3.78053665e-01 -5.80853879e-01
7.82222211e-01 4.11379069e-01 -4.55457978e-02 1.54433951e-01
4.54966009e-01 1.82474837e-01 3.07287753e-01 8.78888488e-01
1.46244562e+00 2.15664297e-01 4.71026272e-01 -5.18754795e-02
5.30846894e-01 -8.31817836e-02 5.63383520e-01 5.33768773e-01
-1.13668703e-01 7.68850386e-01 9.95062232e-01 -5.51464498e-01
-7.02815831e-01 -1.03592598e+00 8.74107540e-01 1.12654817e+00
6.38648510e-01 -1.01008630e+00 -6.54268861e-01 -7.91532397e-01
-3.04680854e-01 7.02544391e-01 -7.12200105e-01 -4.92218494e-01
-8.60522270e-01 -4.61224377e-01 8.28563347e-02 7.39760816e-01
7.80782580e-01 -1.18908584e+00 -3.92865390e-01 3.81613642e-01
-3.75226378e-01 -1.14028454e+00 -8.32885683e-01 -5.17515242e-01
-9.94534731e-01 -1.12856793e+00 -7.62567639e-01 -1.05097413e+00
6.02455080e-01 7.56509900e-01 1.24584532e+00 3.72163653e-01
-1.61898732e-01 4.88307327e-01 -5.73824406e-01 7.44582638e-02
1.19541332e-01 1.86406061e-01 -5.17843544e-01 -1.24966854e-03
1.61805585e-01 -7.63706267e-01 -9.69608724e-01 -5.14130890e-02
-7.67729282e-01 2.83846378e-01 6.93524420e-01 5.65931797e-01
7.74623394e-01 5.25507033e-02 5.63812971e-01 -6.73100948e-01
7.05316365e-01 -4.80892092e-01 -2.58253068e-01 3.81813139e-01
5.32444147e-03 5.54125942e-02 3.68545026e-01 -2.74885803e-01
-1.02528906e+00 -1.92020342e-01 2.76469946e-01 -7.86116600e-01
2.87076801e-01 8.58951271e-01 -1.40795574e-01 3.62006605e-01
3.83705646e-01 4.48092937e-01 -3.23235273e-01 -1.17427282e-01
5.09806752e-01 1.18109450e-01 6.28552556e-01 -1.69144049e-01
5.63401461e-01 4.46665823e-01 -1.43708527e-01 -8.25680017e-01
-1.08902967e+00 -5.02134740e-01 -5.84294975e-01 -4.31075513e-01
1.29059660e+00 -1.00535345e+00 -4.67804193e-01 2.66080618e-01
-1.30456388e+00 -3.21881950e-01 -3.13652575e-01 3.55003864e-01
-6.24250472e-01 6.23273432e-01 -7.81586289e-01 -4.45730239e-01
-5.02004445e-01 -8.84679735e-01 1.27590704e+00 5.04496574e-01
-5.33409864e-02 -8.22967708e-01 -1.51979685e-01 2.71394551e-01
7.06441924e-02 3.69218022e-01 8.19848835e-01 -3.42657298e-01
-1.07614100e+00 -6.52354509e-02 -5.23498833e-01 1.24967098e-02
8.79258886e-02 7.15250820e-02 -4.89679098e-01 -1.19851135e-01
-4.21069384e-01 -1.50497139e-01 1.64180720e+00 7.50210524e-01
1.50380361e+00 -6.10402882e-01 -4.61896777e-01 6.27180696e-01
1.04674745e+00 -1.55968860e-01 7.43342161e-01 1.06289260e-01
1.29825354e+00 4.99705881e-01 6.00618839e-01 4.13956493e-01
7.57915556e-01 3.70842993e-01 5.34527540e-01 -4.07468937e-02
-3.63538504e-01 -5.15277803e-01 2.81136662e-01 8.84858131e-01
-4.01362658e-01 -4.77308780e-01 -5.69258392e-01 6.92232430e-01
-2.45111489e+00 -1.29892254e+00 -1.07263438e-01 1.74245894e+00
5.02309322e-01 1.75797790e-01 1.87121093e-01 -4.07386869e-01
1.03292930e+00 9.31372523e-01 -4.80424464e-01 -2.33181901e-02
-1.16733767e-01 -5.75799122e-02 3.94543707e-01 2.09772691e-01
-1.25038600e+00 1.33487904e+00 5.27397013e+00 7.63712347e-01
-6.89040840e-01 5.49416393e-02 8.24859142e-01 -4.19218183e-01
-4.54507828e-01 -1.17617279e-01 -1.76758263e-02 1.34984896e-01
4.13161486e-01 -4.96480823e-01 2.65153676e-01 6.05824709e-01
1.44760624e-01 -1.69132371e-02 -1.06879508e+00 1.11739612e+00
4.69584703e-01 -1.80314732e+00 3.87853831e-01 -2.21891552e-01
8.64549756e-01 4.34802845e-03 -1.69451416e-01 5.06369770e-01
2.93766111e-01 -8.33850443e-01 6.61720455e-01 7.80687451e-01
4.82431322e-01 -9.06226337e-01 6.24338329e-01 4.14272174e-02
-1.77313018e+00 4.09057289e-02 -3.92481446e-01 -3.70360389e-02
2.23729849e-01 3.31596524e-01 -3.06443632e-01 9.40361500e-01
6.91289186e-01 1.41071725e+00 -6.25102818e-01 1.00735009e+00
-4.34702665e-01 5.00999391e-01 1.23591840e-01 6.71668909e-03
3.35883707e-01 -2.01387763e-01 6.29902065e-01 1.22212541e+00
2.78171659e-01 4.64487970e-01 2.68703699e-01 7.50676215e-01
-5.33517241e-01 1.05282724e-01 -7.01058984e-01 -1.80266440e-01
2.59810984e-01 1.16666985e+00 -8.42688084e-01 -6.97062194e-01
-6.18282735e-01 1.36234641e+00 6.86462700e-01 6.56902850e-01
-1.09208226e+00 -3.09700310e-01 3.41054797e-01 1.35781288e-01
4.97937173e-01 -1.74793735e-01 6.52482882e-02 -1.42186940e+00
1.20153897e-01 -3.19283903e-01 7.74310350e-01 -8.79506230e-01
-1.04880559e+00 4.47131217e-01 1.37992784e-01 -9.29318428e-01
5.12727909e-03 9.43838432e-02 -1.16167510e+00 4.28680569e-01
-1.31563270e+00 -1.30490768e+00 -5.57367980e-01 5.00592113e-01
9.18428242e-01 -1.08297989e-01 2.32216150e-01 -6.55921875e-03
-6.91592336e-01 2.61215508e-01 -5.05502462e-01 3.11473578e-01
2.91986555e-01 -1.00013173e+00 6.70395970e-01 1.00222111e+00
2.11558402e-01 4.42786694e-01 4.57440555e-01 -9.45161760e-01
-1.33972073e+00 -1.60498798e+00 7.45748401e-01 1.85893290e-02
6.91642165e-01 -7.37883449e-02 -1.09207821e+00 9.22457099e-01
5.48860133e-01 2.91935295e-01 4.15157616e-01 -9.02093388e-03
-3.32458884e-01 5.09920791e-02 -6.59910262e-01 8.70880425e-01
1.71935320e+00 -3.85038882e-01 -8.17947686e-01 4.49440867e-01
1.24306440e+00 -2.88043678e-01 -5.95477104e-01 3.86395276e-01
2.79737204e-01 -9.72960353e-01 9.27642882e-01 -6.27499878e-01
7.13338435e-01 -1.91928983e-01 1.18291356e-01 -1.29131579e+00
-7.30379164e-01 -5.64319134e-01 -5.46047509e-01 1.24409676e+00
9.37964022e-02 -1.85483158e-01 7.63964832e-01 1.05782449e-01
-4.83946353e-01 -8.96655023e-01 -5.17887890e-01 -5.32337725e-01
-5.51244855e-01 8.65158346e-03 4.80268836e-01 8.92570555e-01
1.13475330e-01 7.18367279e-01 -2.41089180e-01 1.71470657e-01
6.15965545e-01 3.92300159e-01 7.48245835e-01 -8.94756854e-01
-1.71381146e-01 -6.70388579e-01 -7.38855302e-01 -1.18116295e+00
5.31904399e-01 -9.15092945e-01 -2.17668772e-01 -2.35851216e+00
6.51638091e-01 4.91703302e-01 -4.81756955e-01 2.91066378e-01
-6.06142938e-01 6.50257468e-02 2.31543407e-01 1.25461370e-01
-1.51357627e+00 9.30461347e-01 1.43042958e+00 -5.27566254e-01
-3.25250089e-01 -4.53559965e-01 -5.81131458e-01 7.91202307e-01
7.69408107e-01 -2.56664246e-01 -7.34981596e-01 -5.45311987e-01
2.26377279e-01 2.07177296e-01 5.22923708e-01 -9.46270227e-01
3.76770139e-01 -2.88946539e-01 3.92543882e-01 -9.17114258e-01
1.04084872e-01 -2.41214007e-01 -2.24208370e-01 1.94079995e-01
-4.63526398e-01 -5.49618565e-02 -1.95550686e-03 9.47595417e-01
-4.32145655e-01 2.78355867e-01 4.36551422e-01 -1.82443321e-01
-9.53155756e-01 8.03530335e-01 -1.56608626e-01 5.46457469e-02
1.12422800e+00 -1.65201887e-01 -4.19252783e-01 -5.86351335e-01
-7.13373482e-01 5.95872045e-01 4.58186388e-01 4.92092848e-01
1.03811240e+00 -1.41236508e+00 -6.13908708e-01 -2.47959718e-01
2.51402080e-01 3.25177610e-01 6.71869338e-01 8.20695996e-01
-6.19238555e-01 9.96332690e-02 -1.84681579e-01 -4.70427930e-01
-1.27720499e+00 8.22682619e-01 1.37494519e-01 -2.56610811e-01
-1.01934826e+00 9.65177417e-01 7.33666241e-01 2.37149730e-01
2.12118804e-01 -5.25055349e-01 -5.40110290e-01 7.74186403e-02
5.23857832e-01 2.30698600e-01 -3.24419618e-01 -7.06321418e-01
-4.59947288e-01 3.48386556e-01 -2.79046416e-01 2.19103128e-01
1.45635426e+00 -1.30000383e-01 -2.60871619e-01 2.00514808e-01
1.21749020e+00 -2.97988027e-01 -1.25309086e+00 -3.99134278e-01
-5.61426841e-02 -3.88680905e-01 1.06404908e-01 9.15706158e-03
-1.40314543e+00 7.35624790e-01 -2.02404454e-01 1.57899618e-01
1.38394022e+00 3.61594766e-01 9.42444444e-01 2.63299465e-01
6.66977931e-03 -9.90524769e-01 6.45619273e-01 3.63821238e-01
1.20914268e+00 -9.28273559e-01 4.09317255e-01 -4.89000827e-01
-7.60220587e-01 8.75943303e-01 8.16646039e-01 -5.60344994e-01
3.03386837e-01 -3.41554374e-01 -7.34784484e-01 -5.61210215e-01
-9.08705950e-01 -4.58658338e-01 5.14142275e-01 3.86000991e-01
4.28616047e-01 -1.56635791e-02 -3.98795664e-01 6.42801762e-01
2.53077716e-01 -7.54588246e-02 5.10739028e-01 8.78107369e-01
-6.90981627e-01 -4.46767837e-01 1.23962618e-01 6.78296149e-01
-2.60521233e-01 -2.75863767e-01 -6.25846446e-01 5.92230082e-01
-2.51493812e-01 7.63595819e-01 3.18040699e-01 -5.41996837e-01
5.25794566e-01 -3.60647947e-01 4.90717918e-01 -8.40208650e-01
-3.81866723e-01 1.51303649e-01 1.52477711e-01 -9.00727570e-01
-6.12830281e-01 -5.01299024e-01 -1.60931754e+00 -1.67645395e-01
1.25490939e-02 -6.06611036e-02 -2.62815263e-02 6.41838610e-01
5.35586178e-01 1.28319764e+00 4.02923346e-01 -1.14836335e+00
1.87686272e-02 -9.54421699e-01 -3.75120431e-01 4.54015851e-01
3.04224551e-01 -5.62408864e-01 5.04586734e-02 8.75706747e-02] | [9.844778060913086, 0.7471922636032104] |
8a8d89bd-e963-4f44-8341-bb9d669ba192 | d2c-diffusion-denoising-models-for-few-shot | 2106.06819 | null | https://arxiv.org/abs/2106.06819v1 | https://arxiv.org/pdf/2106.06819v1.pdf | D2C: Diffusion-Denoising Models for Few-shot Conditional Generation | Conditional generative models of high-dimensional images have many applications, but supervision signals from conditions to images can be expensive to acquire. This paper describes Diffusion-Decoding models with Contrastive representations (D2C), a paradigm for training unconditional variational autoencoders (VAEs) for few-shot conditional image generation. D2C uses a learned diffusion-based prior over the latent representations to improve generation and contrastive self-supervised learning to improve representation quality. D2C can adapt to novel generation tasks conditioned on labels or manipulation constraints, by learning from as few as 100 labeled examples. On conditional generation from new labels, D2C achieves superior performance over state-of-the-art VAEs and diffusion models. On conditional image manipulation, D2C generations are two orders of magnitude faster to produce over StyleGAN2 ones and are preferred by 50% - 60% of the human evaluators in a double-blind study. | ['Stefano Ermon', 'Chenlin Meng', 'Jiaming Song', 'Abhishek Sinha'] | 2021-06-12 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 4.36821550e-01 5.01183629e-01 -3.26634884e-01 -1.56618834e-01
-9.09155130e-01 -4.21249717e-01 9.97985542e-01 -5.34179389e-01
-1.56526342e-01 8.34153354e-01 3.61140072e-01 1.00615732e-01
4.32563394e-01 -7.97381282e-01 -8.62689376e-01 -7.97627330e-01
2.66411901e-01 5.71405530e-01 -1.81386307e-01 -5.86381927e-02
-4.99725007e-02 1.66932955e-01 -1.60557151e+00 3.95212889e-01
7.33696759e-01 5.79219759e-01 5.78468502e-01 1.02698040e+00
-1.35851696e-01 6.89118385e-01 -7.52724111e-01 -4.93347168e-01
2.13265583e-01 -1.09268820e+00 -4.70196933e-01 5.11407852e-01
2.54563063e-01 -5.21021545e-01 -2.46136993e-01 1.11846519e+00
6.80498540e-01 9.85488072e-02 1.33768535e+00 -9.30503309e-01
-1.68599868e+00 4.86841261e-01 -3.33802104e-01 -7.86127672e-02
1.72845706e-01 5.61120629e-01 7.02135503e-01 -1.07564676e+00
1.13975739e+00 1.36275792e+00 1.11729570e-01 1.41847837e+00
-1.72250271e+00 -5.97208261e-01 -1.20913722e-02 -1.34605840e-01
-1.27711403e+00 -5.74156463e-01 6.95905685e-01 -7.38388360e-01
9.13009405e-01 -1.58726033e-02 6.07984960e-01 1.84740663e+00
2.38223728e-02 1.01146042e+00 1.03712201e+00 -4.72391725e-01
4.65749800e-01 5.06330609e-01 -5.82767844e-01 8.01889181e-01
3.01260705e-04 4.78216022e-01 -5.71769536e-01 2.06391633e-01
1.12089348e+00 -2.49929890e-01 -2.65930772e-01 -2.90921271e-01
-1.22632611e+00 1.31952083e+00 4.21671420e-01 2.10599959e-01
-6.36130154e-01 3.92288595e-01 1.09554203e-02 3.40851396e-01
7.66137064e-01 6.82426214e-01 -3.05221714e-02 -1.74796075e-01
-9.06745911e-01 1.88114330e-01 4.01672512e-01 1.34796226e+00
6.14233851e-01 7.59676993e-01 -6.08507991e-01 8.23708773e-01
2.12038323e-01 9.11764681e-01 8.49753261e-01 -1.22370362e+00
-1.57044064e-02 5.08097187e-02 1.58453435e-01 -5.52059352e-01
3.06245089e-01 -3.52858454e-01 -9.67650950e-01 4.86973822e-01
-1.61912233e-01 -4.24105406e-01 -1.56243169e+00 1.79127419e+00
-9.15849879e-02 -9.53931436e-02 2.17763275e-01 6.76916242e-01
7.59260297e-01 1.02614093e+00 9.30498838e-02 -4.35847789e-01
6.82712138e-01 -9.88858461e-01 -8.70845437e-01 -2.51589328e-01
4.88300860e-01 -6.33599222e-01 1.05616474e+00 3.51825118e-01
-1.27233827e+00 -7.83667147e-01 -9.76214945e-01 1.07527852e-01
-2.99212128e-01 1.20123841e-01 4.29953516e-01 9.23576593e-01
-1.38006234e+00 5.03728926e-01 -8.04030597e-01 -1.69404358e-01
6.16792738e-01 6.89620599e-02 -9.84979495e-02 -4.03828263e-01
-9.79092062e-01 7.86340237e-01 2.87713170e-01 -4.54672247e-01
-1.88033736e+00 -5.78635991e-01 -1.07934070e+00 -1.46454513e-01
-8.00754689e-03 -8.22138906e-01 1.13107932e+00 -1.09038627e+00
-1.95273769e+00 9.38292801e-01 -1.26171708e-01 -4.30361986e-01
4.79194432e-01 -5.33326492e-02 -3.96006495e-01 2.55622596e-01
2.67057955e-01 1.43277943e+00 1.44434142e+00 -1.53731740e+00
-2.05186188e-01 6.97196350e-02 -2.58563340e-01 1.80789411e-01
-3.46888423e-01 -3.80649239e-01 -3.95356953e-01 -9.50635731e-01
-3.38137120e-01 -1.05053449e+00 -2.90074408e-01 3.77844647e-02
-4.80908126e-01 -2.08445221e-01 8.19696903e-01 -6.43812478e-01
7.59440184e-01 -2.05151272e+00 7.98432112e-01 -3.91350836e-01
2.06340432e-01 2.91230619e-01 -4.98379469e-01 9.29452404e-02
6.92071300e-03 2.70142674e-01 -2.09199607e-01 -6.83779597e-01
1.31888896e-01 2.41868392e-01 -3.05195689e-01 1.36668131e-01
5.76042712e-01 1.07213831e+00 -8.86667490e-01 -2.80694991e-01
2.86977738e-01 7.02197731e-01 -6.40232623e-01 3.63358021e-01
-6.74500167e-01 5.93533158e-01 -8.05435404e-02 5.07340431e-01
4.65254426e-01 -4.99930203e-01 -4.92855124e-02 1.46704838e-01
3.72931182e-01 -3.40881497e-01 -6.77008688e-01 1.99211299e+00
-5.69062412e-01 8.39353442e-01 -3.20643604e-01 -8.06457043e-01
8.56685519e-01 5.02618611e-01 -1.46428850e-02 -6.08626366e-01
1.62959889e-01 6.77128062e-02 -3.49680662e-01 -4.12833482e-01
4.80106801e-01 -5.24366498e-01 3.70028466e-02 4.85423207e-01
7.61056364e-01 -6.30963624e-01 3.69017184e-01 5.34019053e-01
5.84788501e-01 3.07687789e-01 -4.94082943e-02 -2.57614590e-02
-8.18395764e-02 -1.99960232e-01 4.55782682e-01 8.29653025e-01
1.96597889e-01 8.33764017e-01 2.74410248e-01 -1.63658217e-01
-1.27608776e+00 -1.21492064e+00 2.29837835e-01 8.79009366e-01
-1.12543605e-01 -2.46763095e-01 -8.93852770e-01 -6.22896314e-01
-3.96120459e-01 1.35998023e+00 -9.09203112e-01 -2.94218361e-01
8.59901235e-02 -5.86913943e-01 2.52077848e-01 5.59288681e-01
2.92895317e-01 -1.14329052e+00 -3.60481024e-01 2.31150702e-01
-6.15652986e-02 -8.83044302e-01 -4.45199996e-01 -3.04116905e-02
-8.63336205e-01 -4.81302440e-01 -1.66458714e+00 -8.89052808e-01
9.10636842e-01 -2.34447680e-02 1.11784601e+00 -3.75491738e-01
-4.00206149e-01 5.05575478e-01 -4.67359275e-01 -3.98749650e-01
-8.76397848e-01 -2.08297923e-01 1.47107780e-01 1.79055762e-02
1.43972319e-02 -5.05487323e-01 -4.73719716e-01 6.16282299e-02
-9.47616756e-01 3.94210011e-01 8.44979048e-01 1.02871656e+00
7.09614813e-01 -1.44163027e-01 4.72997934e-01 -1.04536605e+00
7.46053100e-01 -4.19476867e-01 -5.73804379e-01 1.94241270e-01
-8.66989136e-01 3.96672249e-01 3.67305309e-01 -9.14878845e-01
-1.39216197e+00 -3.12354807e-02 1.58098012e-01 -9.19343412e-01
-9.35286954e-02 2.63540089e-01 1.96164057e-01 2.13377222e-01
1.21219647e+00 4.76027489e-01 -7.95669630e-02 -2.65113384e-01
1.07305288e+00 4.04459298e-01 3.71328086e-01 -4.90907162e-01
7.98526824e-01 2.17482552e-01 -3.32102656e-01 -7.96246469e-01
-5.98146319e-01 2.29155734e-01 -4.42654222e-01 -2.85184413e-01
1.39401889e+00 -1.06746995e+00 6.28381548e-03 4.99191523e-01
-1.10457921e+00 -7.52083004e-01 -7.24716961e-01 5.46465635e-01
-7.53239989e-01 -9.46326032e-02 -7.45095015e-01 -7.90645480e-01
-2.09982395e-01 -1.28158915e+00 1.01432621e+00 3.42476010e-01
-1.02649748e-01 -9.97479260e-01 1.34853691e-01 2.23717704e-01
4.99326259e-01 3.00169140e-01 8.81329656e-01 -2.21496411e-02
-7.01656282e-01 -2.14508876e-01 8.41051936e-02 6.86632276e-01
9.18789282e-02 -1.39447972e-01 -9.20207620e-01 -5.19801974e-01
-2.18678087e-01 -8.56248856e-01 9.79200184e-01 6.13661289e-01
8.79719973e-01 -3.10683101e-01 -2.94304460e-01 5.66287041e-01
1.31314647e+00 2.58575916e-01 7.80586541e-01 -3.50443065e-01
5.90285718e-01 3.91011804e-01 1.74929366e-01 4.91251379e-01
-6.18736520e-02 3.59951049e-01 2.66546071e-01 -2.95064356e-02
-8.16874802e-01 -6.12854183e-01 6.04742289e-01 9.21869218e-01
-1.55177906e-01 -4.64113504e-01 -4.04350996e-01 8.28492820e-01
-1.26078761e+00 -9.03883934e-01 2.31918782e-01 1.92703652e+00
1.08739507e+00 5.38931601e-02 -2.11377591e-01 -4.67777491e-01
9.09455538e-01 1.48333460e-01 -8.02482665e-01 -2.09401503e-01
-2.20969349e-01 3.81133139e-01 1.61028355e-01 4.76527452e-01
-6.62079275e-01 1.19515657e+00 7.20764160e+00 1.05957353e+00
-9.96825516e-01 3.39348406e-01 8.42053533e-01 -3.54754925e-01
-6.74716055e-01 -8.94440934e-02 -6.96048558e-01 3.68264854e-01
1.00778842e+00 -6.39237314e-02 5.07637799e-01 9.43962514e-01
-3.01018745e-01 8.28462094e-02 -1.06141078e+00 1.17581093e+00
5.43312132e-01 -1.63605094e+00 3.72187167e-01 1.89618081e-01
1.60425067e+00 -2.24344172e-02 7.91672885e-01 5.41548014e-01
9.27243173e-01 -1.19682026e+00 6.61926925e-01 5.55987358e-01
1.51898229e+00 -4.96867925e-01 2.61990577e-01 3.48953605e-01
-5.65769136e-01 1.47455052e-01 -6.29990578e-01 2.96710759e-01
4.07226115e-01 3.82858604e-01 -6.88320994e-01 4.37363703e-03
4.17991728e-01 7.21224010e-01 -4.53857541e-01 3.32373410e-01
-7.75729358e-01 7.50634015e-01 1.01968527e-01 -1.38644233e-01
2.60356218e-01 -1.14967413e-02 4.74113077e-01 9.94032025e-01
6.53967977e-01 1.03425235e-01 -2.46195182e-01 1.35686743e+00
-3.41478616e-01 -1.08231716e-01 -8.51314723e-01 -5.25487244e-01
5.69756441e-02 8.97191644e-01 -4.68809754e-01 -7.90399015e-01
-1.08485214e-01 1.60395217e+00 1.67584106e-01 7.36367702e-01
-6.42705917e-01 -2.79773831e-01 2.48772830e-01 -1.60698816e-01
6.69526160e-01 -7.48809278e-02 8.34654868e-02 -1.34335637e+00
-5.93301773e-01 -8.84308815e-01 8.39747861e-03 -1.30117953e+00
-1.22212851e+00 8.09365630e-01 -6.70994446e-02 -1.21803355e+00
-7.92608142e-01 -5.83057702e-01 -3.08725148e-01 8.65871072e-01
-1.29089081e+00 -1.15125406e+00 -7.66482726e-02 6.62572742e-01
9.89698529e-01 -6.91420317e-01 1.24158037e+00 -9.42898393e-02
-2.31866822e-01 4.84859586e-01 3.10851455e-01 -7.67686963e-02
5.86819470e-01 -1.14626324e+00 4.28675950e-01 9.19553280e-01
4.27590132e-01 3.94264787e-01 7.03161299e-01 -7.53481030e-01
-9.69763041e-01 -1.16088068e+00 5.46829283e-01 -4.02956039e-01
3.46074104e-02 -3.45296115e-01 -5.19275188e-01 8.10817599e-01
6.04111671e-01 1.43253906e-02 7.54591107e-01 -2.02803403e-01
-4.41421986e-01 4.80152518e-01 -1.08419776e+00 6.92882001e-01
1.04886329e+00 -6.98472679e-01 -2.98342526e-01 5.34900546e-01
9.62811589e-01 -1.91256613e-01 -8.08457971e-01 -4.57614101e-02
9.76016596e-02 -7.21909106e-01 8.20232332e-01 -5.89730442e-01
8.38000476e-01 5.10359257e-02 -2.57425696e-01 -1.71968257e+00
-5.04241705e-01 -7.85970569e-01 -3.61749262e-01 1.07241559e+00
6.90930068e-01 -1.09553650e-01 7.53598332e-01 3.70894998e-01
1.07423216e-02 -2.45505139e-01 -4.29603249e-01 -8.86106610e-01
2.53795743e-01 -2.51691550e-01 2.12627172e-01 9.29432750e-01
-3.73905092e-01 6.44916475e-01 -8.55296552e-01 -2.39807650e-01
8.45365167e-01 -9.86781791e-02 7.02976763e-01 -8.69173169e-01
-6.12719238e-01 -4.03071791e-01 -3.21940690e-01 -1.29394114e+00
1.65502355e-01 -9.65851724e-01 2.44186044e-01 -1.64260542e+00
1.36555299e-01 5.16669676e-02 -5.78821488e-02 2.19570756e-01
-1.98854581e-01 4.63180006e-01 2.59493470e-01 3.09095591e-01
-4.65312749e-01 1.11393809e+00 1.56789231e+00 -6.67785943e-01
-2.12280780e-01 -1.91550151e-01 -6.44295394e-01 3.86997819e-01
5.58853328e-01 -5.85411489e-01 -8.86242807e-01 -6.41416848e-01
1.21540003e-01 2.27409229e-01 2.72001438e-02 -9.19736147e-01
-5.81998415e-02 -1.84865683e-01 7.24532664e-01 -2.64538050e-01
5.87197721e-01 -1.41573817e-01 2.43894354e-01 5.19113719e-01
-6.76356435e-01 -3.41357499e-01 -5.92787676e-02 8.41349006e-01
-7.46453553e-02 -4.47308183e-01 8.57638717e-01 -5.27065635e-01
-8.23086202e-01 3.50554734e-01 -7.26558506e-01 4.75891046e-02
1.00313485e+00 4.36117919e-03 -1.80277854e-01 -8.15729558e-01
-1.07905269e+00 -2.13603750e-01 4.45617765e-01 4.73774493e-01
1.18260658e+00 -1.62689471e+00 -1.00195837e+00 2.78400183e-01
2.55221993e-01 -6.20380156e-02 3.22856516e-01 2.97591269e-01
-3.66092116e-01 1.39306113e-01 -1.75747082e-01 -5.83885014e-01
-8.83418620e-01 8.67137551e-01 2.68144514e-02 -1.08415075e-01
-4.31128740e-01 1.55179799e+00 4.85186845e-01 -3.16248149e-01
1.61231589e-02 1.44895270e-01 2.61243861e-02 -7.40860775e-02
5.52303135e-01 4.00265940e-02 -6.45033836e-01 -5.04681170e-01
3.81509990e-01 4.01103824e-01 -2.88535357e-01 -7.35935330e-01
1.17963111e+00 3.68295982e-02 3.81401658e-01 5.09708822e-01
1.22307599e+00 -4.76472646e-01 -1.80716717e+00 -3.17443758e-01
-6.33223176e-01 -4.48835880e-01 2.23826334e-01 -8.91780496e-01
-1.20091426e+00 1.18077111e+00 7.93903053e-01 -1.02795795e-01
8.63927960e-01 2.18097359e-01 6.19207263e-01 2.18610421e-01
1.87827781e-01 -1.16472292e+00 8.72459352e-01 3.17038186e-02
1.02508211e+00 -1.22883093e+00 -1.66991949e-01 3.59847881e-02
-1.14686942e+00 6.95234120e-01 3.95468384e-01 -1.31062061e-01
5.35572469e-01 -1.23310946e-01 8.57787952e-02 -1.79684222e-01
-8.31911445e-01 -2.06380382e-01 3.23298097e-01 1.24114096e+00
2.14624584e-01 2.37864450e-01 9.41685140e-02 9.02294889e-02
-4.22748812e-02 2.71838848e-02 4.92784828e-01 6.72275960e-01
-1.93761751e-01 -1.03466785e+00 -4.52168807e-02 3.93401295e-01
-8.91476870e-02 -2.49573037e-01 -1.57049716e-01 4.13248003e-01
1.12646595e-01 9.66537714e-01 1.25771210e-01 -3.69452119e-01
-1.59239456e-01 2.05814555e-01 7.74029672e-01 -8.24793279e-01
1.21467873e-01 2.75124401e-01 -9.36959684e-02 -2.69559592e-01
-5.24908185e-01 -6.99703395e-01 -8.20407808e-01 -1.16421491e-01
-4.43577647e-01 -4.73089702e-02 8.17886293e-01 6.28352046e-01
5.59216738e-01 6.58975244e-01 5.93664408e-01 -9.27302003e-01
-4.66454268e-01 -1.22846842e+00 -6.97347879e-01 6.38421774e-01
1.20443746e-01 -6.59611344e-01 -3.15282971e-01 7.41416991e-01] | [11.385379791259766, -0.21780355274677277] |
b507784c-4be9-412a-bf68-182eac50dace | segmentation-of-vhr-eo-images-using | 2108.04222 | null | https://arxiv.org/abs/2108.04222v2 | https://arxiv.org/pdf/2108.04222v2.pdf | Segmentation of VHR EO Images using Unsupervised Learning | Semantic segmentation is a crucial step in many Earth observation tasks. Large quantity of pixel-level annotation is required to train deep networks for semantic segmentation. Earth observation techniques are applied to varieties of applications and since classes vary widely depending on the applications, therefore, domain knowledge is often required to label Earth observation images, impeding availability of labeled training data in many Earth observation applications. To tackle these challenges, in this paper we propose an unsupervised semantic segmentation method that can be trained using just a single unlabeled scene. Remote sensing scenes are generally large. The proposed method exploits this property to sample smaller patches from the larger scene and uses deep clustering and contrastive learning to refine the weights of a lightweight deep model composed of a series of the convolution layers along with an embedded channel attention. After unsupervised training on the target image/scene, the model automatically segregates the major classes present in the scene and produces the segmentation map. Experimental results on the Vaihingen dataset demonstrate the efficacy of the proposed method. | ['Xiao Xiang Zhu', 'Muhammad Shahzad', 'Lichao Mou', 'Sudipan Saha'] | 2021-07-09 | null | null | null | null | ['unsupervised-semantic-segmentation'] | ['computer-vision'] | [ 5.99934995e-01 2.77671311e-02 2.11662516e-01 -6.83633566e-01
-2.00128183e-01 -5.80264270e-01 1.90239504e-01 3.36891949e-01
-6.33758128e-01 4.21417803e-01 -2.56226629e-01 -2.81339020e-01
-7.92429820e-02 -1.03656363e+00 -4.31578368e-01 -9.11058426e-01
1.16923138e-01 4.30800557e-01 4.77858841e-01 3.43320966e-02
3.33302587e-01 3.44922692e-01 -1.42883456e+00 -6.25183573e-03
1.12184811e+00 1.07256186e+00 9.30965245e-01 6.16319716e-01
-5.34231722e-01 4.98044193e-01 -3.43914360e-01 3.04635674e-01
5.36392450e-01 -4.44156200e-01 -9.66910362e-01 6.44182146e-01
2.94421971e-01 -1.30480096e-01 1.92181572e-01 1.46673834e+00
3.59910965e-01 5.95737398e-01 6.82803214e-01 -6.74559176e-01
-1.48561642e-01 5.24329722e-01 -6.62917256e-01 5.62857687e-01
-5.00190377e-01 -1.70098647e-01 9.26506162e-01 -5.44229805e-01
1.51367813e-01 8.91744316e-01 3.58444661e-01 1.64551973e-01
-9.20470119e-01 -4.49517339e-01 3.47121358e-01 -1.57766059e-01
-1.31872833e+00 -2.89126873e-01 8.54830503e-01 -5.83088994e-01
5.64281583e-01 1.20151431e-04 6.02644563e-01 -3.49396914e-02
-6.75440952e-02 4.29312319e-01 1.22495437e+00 -2.67838895e-01
3.48641962e-01 2.70198882e-01 3.87891889e-01 4.60016131e-01
1.98557407e-01 -3.74041498e-01 2.57417470e-01 2.22546086e-01
7.57837296e-01 3.31842810e-01 -1.73506215e-01 -2.11935282e-01
-8.57071459e-01 8.42175364e-01 8.24209571e-01 1.34877279e-01
-4.98561084e-01 2.59512458e-02 2.03711033e-01 8.60787034e-02
6.85850263e-01 2.68106997e-01 -4.57113683e-01 5.53568721e-01
-1.20689869e+00 1.52882608e-02 4.44874227e-01 5.50034344e-01
1.17108285e+00 4.64646295e-02 3.55395913e-01 1.06329393e+00
4.31142688e-01 3.93788815e-01 3.50996703e-01 -5.57964087e-01
2.83560663e-01 9.38348234e-01 2.44642962e-02 -8.03570330e-01
-6.21596038e-01 -6.45112514e-01 -8.41711760e-01 3.05363864e-01
2.37366542e-01 -2.56171107e-01 -1.42375469e+00 1.20623147e+00
6.78696394e-01 1.32497445e-01 1.93572342e-01 1.06004930e+00
7.32748032e-01 8.21600497e-01 4.05418038e-01 2.73484796e-01
1.28258693e+00 -9.41981971e-01 -1.56493202e-01 -5.51678061e-01
4.89739180e-01 -6.22412026e-01 9.22189236e-01 1.75708458e-02
-5.21949589e-01 -6.30056441e-01 -1.03206682e+00 9.18183997e-02
-5.45299947e-01 2.80056179e-01 7.81034589e-01 5.88854313e-01
-8.59856665e-01 3.72918636e-01 -9.17446494e-01 -5.49045384e-01
7.83712566e-01 4.01927084e-01 -1.08846754e-01 4.39128727e-02
-9.24169242e-01 5.05546451e-01 1.02690887e+00 4.59748775e-01
-1.06800663e+00 -1.48431987e-01 -7.58188069e-01 2.99334943e-01
3.03710282e-01 -2.44723797e-01 8.43295932e-01 -1.40254116e+00
-1.30277240e+00 9.21291649e-01 2.93205947e-01 -4.23684627e-01
2.42548689e-01 -1.82159841e-01 -1.50313675e-01 2.68464208e-01
2.81238407e-01 8.47316206e-01 8.77892256e-01 -1.04414487e+00
-1.08449435e+00 -3.61077338e-01 1.22621410e-01 6.18554056e-01
-4.54090722e-03 -2.29594529e-01 -2.49266043e-01 -4.33151245e-01
5.62698066e-01 -7.02608049e-01 -5.63049138e-01 -3.19609433e-01
-3.06310594e-01 -4.75120507e-02 9.96266901e-01 -6.83038592e-01
6.64457679e-01 -2.19420671e+00 -6.81806430e-02 2.99081475e-01
1.27660736e-01 7.52620548e-02 2.12351501e-01 3.50068919e-02
-3.04673854e-02 -3.30336168e-02 -8.47485423e-01 1.23046048e-01
-2.37540931e-01 1.01093329e-01 -2.00987577e-01 6.46703005e-01
2.13679507e-01 5.50157428e-01 -7.32835948e-01 -4.90432441e-01
4.21041250e-01 2.37395644e-01 -4.85986590e-01 3.11261207e-01
-5.62077284e-01 1.02129662e+00 -6.70473337e-01 5.23864746e-01
8.66235316e-01 -3.16346377e-01 -3.66449729e-02 7.70712122e-02
-2.90876538e-01 7.30397599e-03 -1.27070856e+00 1.66931057e+00
-3.44176441e-01 4.69058782e-01 5.67098521e-02 -1.72448361e+00
9.40444052e-01 1.35600656e-01 2.44127095e-01 -4.30548161e-01
4.43912059e-01 2.71496266e-01 1.47832885e-01 -7.39968240e-01
4.45533872e-01 -3.61651093e-01 -9.72719339e-04 3.02442163e-01
4.40274961e-02 -3.43234271e-01 1.11923732e-01 -8.96258131e-02
4.21997637e-01 1.49743631e-01 1.71924382e-01 -6.15501225e-01
5.64152539e-01 3.24373722e-01 5.15908599e-01 6.28843606e-01
-1.38617173e-01 3.80371720e-01 -2.69510262e-02 -6.89687073e-01
-1.02229965e+00 -7.11177826e-01 -2.90048748e-01 1.31042421e+00
5.16953886e-01 5.32061040e-01 -9.30293322e-01 -5.20352125e-01
-3.93947124e-01 3.49912375e-01 -5.93908608e-01 2.05968156e-01
-1.34673983e-01 -1.17568099e+00 2.99168855e-01 2.82400876e-01
1.07949877e+00 -1.24587703e+00 -9.88197803e-01 2.69697189e-01
-7.03987926e-02 -1.13449085e+00 -4.15341370e-02 3.37585628e-01
-1.09934616e+00 -1.05882835e+00 -5.31953216e-01 -1.01465607e+00
9.83828187e-01 5.20228505e-01 8.03140759e-01 1.08252347e-01
-2.47443810e-01 -1.85434297e-01 -4.63054746e-01 -3.81553471e-01
5.05931415e-02 3.27234864e-01 -4.59378660e-01 3.03178579e-01
3.14465582e-01 -6.24773324e-01 -8.77561510e-01 8.61896500e-02
-1.17472434e+00 1.85541525e-01 7.92783678e-01 5.85493863e-01
4.55098987e-01 6.38698280e-01 5.62124670e-01 -1.30407393e+00
-1.72511768e-02 -5.80743909e-01 -9.46683347e-01 6.39107451e-02
-2.14587986e-01 -5.42178079e-02 3.73519212e-01 -1.84339568e-01
-1.22218597e+00 4.34689939e-01 -9.34631303e-02 1.48699462e-01
-6.53259754e-01 8.45478594e-01 -2.63198853e-01 -6.39896616e-02
5.16524732e-01 3.78782421e-01 -4.10293669e-01 -4.07397270e-01
2.51116872e-01 8.77583921e-01 5.00300169e-01 -3.52644086e-01
6.00092888e-01 6.66386843e-01 -3.43420297e-01 -1.13500619e+00
-1.12482643e+00 -7.06400514e-01 -7.34113693e-01 -2.40310635e-02
1.34118450e+00 -1.08234584e+00 -3.96183543e-02 4.73505765e-01
-6.78277731e-01 -5.78662455e-01 -3.33734266e-02 6.14651084e-01
-1.40339509e-01 2.09323481e-01 -1.41860083e-01 -8.46344590e-01
-4.90339130e-01 -1.04967487e+00 9.05189991e-01 6.04711711e-01
2.56517261e-01 -9.48278010e-01 -1.08262084e-01 2.89357007e-01
2.86604524e-01 1.51682392e-01 7.06942081e-01 -6.98838592e-01
-6.12247467e-01 -1.30188107e-01 -5.65865159e-01 4.90723968e-01
3.35442513e-01 -2.21648857e-01 -1.02492332e+00 -8.61972272e-02
1.32575601e-01 -3.60375494e-01 1.14380813e+00 6.79194748e-01
1.22242367e+00 1.25605334e-02 -2.09582224e-01 7.91757345e-01
1.74864006e+00 1.83158323e-01 5.54688931e-01 3.37556809e-01
9.66827095e-01 7.88966656e-01 4.46643859e-01 3.13234955e-01
2.02734172e-01 -1.84889287e-02 4.31806505e-01 -3.93504888e-01
2.41251603e-01 1.91024557e-01 -1.50722727e-01 5.02204955e-01
-8.93256161e-03 -8.44228268e-02 -9.07009661e-01 8.78804266e-01
-1.47209466e+00 -7.51709282e-01 -1.01834320e-01 2.07245302e+00
6.59991622e-01 -3.94996889e-02 -2.64036566e-01 1.77186176e-01
8.68330121e-01 2.34480351e-01 -5.80170155e-01 3.59707251e-02
7.11235330e-02 3.33694011e-01 7.84314394e-01 5.25761187e-01
-1.53807402e+00 1.31729114e+00 5.08004618e+00 6.93182886e-01
-1.51523578e+00 1.15240999e-01 7.66703248e-01 4.58444923e-01
8.20506513e-02 1.62435040e-01 -6.57827079e-01 3.55703533e-01
5.18457532e-01 4.58131075e-01 1.38791189e-01 7.58497059e-01
3.74522060e-01 -4.95947570e-01 -5.16662121e-01 7.88944542e-01
-3.42701912e-01 -9.25853312e-01 2.35117022e-02 -2.04833254e-01
9.84536350e-01 3.80075872e-01 -2.15698019e-01 -3.42374593e-02
4.24143881e-01 -1.02410853e+00 5.58964074e-01 1.17176570e-01
4.81767178e-01 -6.67670786e-01 7.87033141e-01 4.90162998e-01
-1.25875425e+00 -1.02866836e-01 -8.02300513e-01 -1.97847724e-01
-6.63543120e-02 6.54460073e-01 -6.20423138e-01 4.01213437e-01
7.64631331e-01 6.08350337e-01 -4.14612919e-01 1.19590294e+00
-1.98183313e-01 7.62604296e-01 -5.00840724e-01 3.13613594e-01
6.00795209e-01 -5.79228282e-01 1.32615820e-01 1.06530762e+00
1.89311624e-01 3.44442487e-01 5.20117939e-01 7.02241600e-01
-1.29317101e-02 3.36934298e-01 -3.47607404e-01 -2.85171539e-01
2.25004569e-01 1.53756917e+00 -1.55533922e+00 -4.60372746e-01
-3.34927499e-01 1.03042293e+00 2.15609610e-01 3.96633178e-01
-5.46523929e-01 -5.24458706e-01 1.32450715e-01 -4.39434918e-03
4.44793403e-01 -1.75621256e-01 -3.94666821e-01 -1.03140903e+00
-3.29191804e-01 -4.37059224e-01 4.52522397e-01 -5.14019847e-01
-9.06060934e-01 6.55112863e-01 -1.22817397e-01 -1.03789461e+00
4.44893062e-01 -2.50944942e-01 -7.43872225e-01 8.85071158e-01
-1.81390107e+00 -1.20986438e+00 -8.11223030e-01 4.40709710e-01
8.26975524e-01 1.37440979e-01 5.53142011e-01 3.51131231e-01
-5.85242629e-01 -2.21348971e-01 1.34701714e-01 4.50509727e-01
2.30092034e-01 -1.22647691e+00 1.17034353e-01 1.08117557e+00
1.78540945e-02 2.51756251e-01 5.78100085e-01 -6.15304172e-01
-6.50614083e-01 -1.49824226e+00 3.59983504e-01 1.68575644e-01
4.03805375e-01 -1.05702959e-01 -9.33299899e-01 7.45729744e-01
1.00163735e-01 2.01416239e-01 6.48554206e-01 -3.12929988e-01
1.22358099e-01 -6.65663928e-02 -1.12863410e+00 1.25606507e-01
4.87139672e-01 -3.17075491e-01 -5.87299585e-01 3.68667990e-01
4.68377739e-01 -3.01452816e-01 -5.26732147e-01 2.92524099e-01
1.75034910e-01 -8.01696777e-01 6.64169848e-01 -3.06713223e-01
3.82905215e-01 -6.61464036e-01 -1.98571637e-01 -1.15673339e+00
-7.55134672e-02 7.97112063e-02 7.83605814e-01 9.82225597e-01
3.88296664e-01 -6.52649641e-01 8.19516599e-01 3.47962350e-01
-3.10589314e-01 -1.29283130e-01 -3.62045050e-01 -2.54517943e-01
-2.96021327e-02 -5.40108494e-02 4.01404232e-01 1.17844081e+00
-6.02957487e-01 4.84687150e-01 -3.62089043e-03 7.25851476e-01
8.25805962e-01 4.13236827e-01 6.43869698e-01 -1.61765003e+00
-1.09322987e-01 -2.73903191e-01 -3.73931259e-01 -1.12653291e+00
-1.97150707e-02 -8.81524384e-01 5.09076893e-01 -1.70923030e+00
2.65967816e-01 -8.68831336e-01 -2.44853780e-01 4.02550250e-01
-4.55062330e-01 4.37960058e-01 -3.04888576e-01 3.05021167e-01
-5.43495536e-01 5.23552418e-01 9.03379977e-01 -3.43220711e-01
-4.20835346e-01 1.32554676e-02 -5.18197596e-01 7.67588019e-01
9.96322393e-01 -5.58668792e-01 -6.26524925e-01 -6.90321743e-01
1.23255976e-01 -3.41843158e-01 4.24751639e-01 -1.12441158e+00
1.82970464e-01 -2.09870145e-01 4.19643104e-01 -5.71413517e-01
-2.19054818e-01 -9.44470763e-01 1.51247814e-01 3.28863621e-01
-8.72905105e-02 -6.16290927e-01 6.94287419e-02 5.47558486e-01
-3.42717499e-01 -4.46998894e-01 1.16200042e+00 -4.79123056e-01
-1.21622097e+00 3.90131623e-01 -2.59349644e-01 -1.61798876e-02
9.58976686e-01 -3.71948898e-01 1.44981861e-01 1.30268976e-01
-8.43917549e-01 3.10077488e-01 3.30120593e-01 -9.31150764e-02
4.13243443e-01 -6.81693196e-01 -7.41440535e-01 2.30326250e-01
1.07477888e-01 8.28136325e-01 5.23951948e-01 7.47079194e-01
-1.08543491e+00 1.24593899e-01 -2.89896131e-01 -7.97893524e-01
-7.89661586e-01 2.61277199e-01 5.16883254e-01 7.91850872e-03
-8.12863886e-01 9.37231183e-01 6.88357174e-01 -6.28549755e-01
-6.12104461e-02 -4.77092534e-01 -5.02197206e-01 4.28485982e-02
2.76987731e-01 -5.24180569e-02 -1.24393729e-02 -6.88258350e-01
-7.63218403e-02 5.45675278e-01 1.02507532e-01 -5.17517775e-02
1.55527329e+00 -5.19210279e-01 -2.26189882e-01 3.35930914e-01
9.27895010e-01 -3.57094467e-01 -1.49249232e+00 -2.60541171e-01
2.34407783e-02 -2.56195635e-01 3.97291452e-01 -5.00555754e-01
-1.44306707e+00 1.13018620e+00 7.65302956e-01 2.61661589e-01
1.26052976e+00 -1.05275497e-01 6.22705877e-01 3.53303462e-01
8.45126063e-03 -1.19970000e+00 -3.40432316e-01 3.45700741e-01
1.81519285e-01 -1.58602357e+00 -8.88737217e-02 -4.55479175e-01
-4.66234893e-01 8.63409817e-01 4.39286500e-01 -3.81129503e-01
8.05754483e-01 -1.34324014e-01 4.21123475e-01 -5.57853997e-01
9.24978629e-02 -4.98006225e-01 2.35571042e-02 3.85282993e-01
3.33666980e-01 1.38080448e-01 -2.07562119e-01 3.82880241e-01
-7.79720768e-03 -2.00810462e-01 2.78766602e-01 9.18698549e-01
-1.19437921e+00 -6.70645714e-01 -3.77719462e-01 5.43155253e-01
-7.47100055e-01 -3.25118750e-01 1.39038172e-03 3.33658576e-01
2.99814969e-01 9.24014926e-01 2.27216572e-01 -8.22761655e-03
-2.06463039e-01 8.51770956e-03 1.20262437e-01 -8.46780121e-01
-4.10713911e-01 2.75885671e-01 -3.10119271e-01 -1.82832312e-02
-8.68165135e-01 -3.46344769e-01 -1.50785911e+00 1.95531666e-01
-4.31814462e-01 3.81780893e-01 7.79355764e-01 1.23434699e+00
-4.34993282e-02 5.78627467e-01 6.56388938e-01 -9.78076696e-01
-9.52634960e-02 -1.12529695e+00 -9.16617930e-01 3.27482879e-01
2.37886727e-01 -6.32096052e-01 -2.41924435e-01 4.83290106e-01] | [9.498333930969238, -1.0298399925231934] |
1a6dc22a-696b-4aa4-8f43-55a8d5374ecf | multi-task-end-to-end-training-improves-1 | 2305.06218 | null | https://arxiv.org/abs/2305.06218v1 | https://arxiv.org/pdf/2305.06218v1.pdf | Multi-Task End-to-End Training Improves Conversational Recommendation | In this paper, we analyze the performance of a multitask end-to-end transformer model on the task of conversational recommendations, which aim to provide recommendations based on a user's explicit preferences expressed in dialogue. While previous works in this area adopt complex multi-component approaches where the dialogue management and entity recommendation tasks are handled by separate components, we show that a unified transformer model, based on the T5 text-to-text transformer model, can perform competitively in both recommending relevant items and generating conversation dialogue. We fine-tune our model on the ReDIAL conversational movie recommendation dataset, and create additional training tasks derived from MovieLens (such as the prediction of movie attributes and related movies based on an input movie), in a multitask learning setting. Using a series of probe studies, we demonstrate that the learned knowledge in the additional tasks is transferred to the conversational setting, where each task leads to a 9%-52% increase in its related probe score. | ['Judith Yue Li', 'Ambarish Jash', 'Santiago Ontanon', 'Moustafa Farid Alzantot', 'Ellie Ka In Chio', 'Dima Kuzmin', 'Naveen Ram'] | 2023-05-08 | null | null | null | null | ['movie-recommendation', 'dialogue-management'] | ['miscellaneous', 'natural-language-processing'] | [ 3.65387380e-01 2.24919751e-01 -4.78689037e-02 -8.20601761e-01
-1.02283108e+00 -7.86995292e-01 7.47184694e-01 -9.09755081e-02
-3.80268931e-01 6.25495374e-01 8.77791703e-01 -3.40604037e-01
-1.74657464e-01 -4.28905219e-01 -3.44738156e-01 -3.37715745e-01
1.38857052e-01 1.00945544e+00 2.18281582e-01 -6.10526860e-01
3.85523885e-01 -1.92815945e-01 -1.23416400e+00 1.14124155e+00
6.22199953e-01 1.00381052e+00 1.83784187e-01 9.07752216e-01
-1.34638622e-01 9.40626919e-01 -3.93408746e-01 -9.03946161e-01
4.59502339e-02 -3.25520217e-01 -1.18027782e+00 1.55451298e-01
4.12246108e-01 -5.80907762e-01 8.79632030e-03 2.81206220e-01
6.54244065e-01 8.70688558e-01 9.53365505e-01 -7.76933134e-01
-5.40271282e-01 8.35282564e-01 1.79730311e-01 9.67102796e-02
5.81721783e-01 -4.58020508e-01 1.63807392e+00 -1.14483702e+00
3.28006357e-01 1.31173813e+00 4.85435247e-01 7.55229831e-01
-1.34314024e+00 -2.31692821e-01 3.55849087e-01 -1.44560695e-01
-4.01930630e-01 -6.48358941e-01 4.36082482e-01 -4.07992363e-01
1.15136814e+00 5.08473933e-01 5.59589565e-02 1.39364338e+00
7.61916637e-02 9.33934450e-01 8.45833182e-01 -2.09886953e-01
6.48237094e-02 6.05155945e-01 3.12027872e-01 2.20565647e-01
-5.26251316e-01 -3.24996680e-01 -7.46425450e-01 -4.38573718e-01
3.04342180e-01 4.48676012e-02 -1.03927091e-01 -1.67662472e-01
-1.01198828e+00 1.19729781e+00 -5.39872758e-02 1.06521890e-01
-2.87013173e-01 -4.37971503e-01 4.19106483e-01 7.55461514e-01
1.06304860e+00 8.46038520e-01 -8.35756302e-01 -3.35503668e-01
-4.52778190e-01 5.14445901e-01 1.29238284e+00 8.84495020e-01
1.99852511e-01 -4.32002455e-01 -6.73997939e-01 1.46013618e+00
3.59307021e-01 2.95606226e-01 4.98397291e-01 -1.16388166e+00
6.37400925e-01 7.60879368e-02 5.76209664e-01 -6.45459771e-01
-5.62580824e-01 -3.69578749e-01 -2.72884458e-01 -5.21959722e-01
5.95404088e-01 -6.66731179e-01 -5.66816218e-02 1.76271129e+00
4.15791035e-01 -1.75324768e-01 2.07937405e-01 7.33189225e-01
1.08931386e+00 6.05693698e-01 -1.02923743e-01 -4.30167258e-01
1.40099490e+00 -1.50788200e+00 -5.75077236e-01 -1.31836310e-01
7.09895849e-01 -9.73255575e-01 1.24661326e+00 6.36072218e-01
-1.20241725e+00 -5.56083083e-01 -5.08104622e-01 -2.39824235e-01
-2.18824460e-03 1.64798290e-01 5.31956911e-01 3.63179713e-01
-1.01457763e+00 5.83480477e-01 -1.85909986e-01 -4.98088121e-01
-2.28884801e-01 2.46036828e-01 -3.55847478e-02 1.00415960e-01
-1.28533125e+00 9.43073392e-01 -4.59937215e-01 -7.68073350e-02
-7.88529754e-01 -4.41239119e-01 -4.91079450e-01 1.87428012e-01
4.72152442e-01 -7.48323321e-01 2.01747322e+00 -7.43764460e-01
-2.05748820e+00 3.70711237e-01 -1.52966693e-01 -2.11977765e-01
3.17282945e-01 -4.04638976e-01 -2.09022611e-01 -1.62644297e-01
-8.60167667e-02 3.13788503e-01 6.34694517e-01 -8.31221700e-01
-1.06220436e+00 -7.97494650e-02 5.95282853e-01 6.77966952e-01
-7.00042546e-01 2.91577816e-01 -2.62563795e-01 -5.71370780e-01
-4.27557945e-01 -1.13831508e+00 -2.69360483e-01 -6.57615721e-01
-2.18461841e-01 -6.27180040e-01 6.13514818e-02 -4.92864728e-01
1.07204294e+00 -1.78562665e+00 3.99841547e-01 3.95151861e-02
1.52434081e-01 -1.07574396e-01 -3.52908134e-01 7.45320559e-01
4.27738845e-01 -2.46386081e-01 3.98825288e-01 -9.28218424e-01
2.17692330e-01 -9.65937972e-02 -6.15923703e-01 -6.96079358e-02
-2.38409162e-01 6.62267804e-01 -7.87774622e-01 -6.35631159e-02
-2.44619161e-01 1.50835246e-01 -1.04024041e+00 7.75250733e-01
-4.07111853e-01 4.92954135e-01 -6.13449037e-01 -1.16107173e-01
-2.54070070e-02 -4.73944068e-01 4.61076349e-01 -1.37970060e-01
3.05837452e-01 1.02342141e+00 -6.32800579e-01 1.60509169e+00
-8.81604970e-01 2.53527731e-01 1.42264739e-01 -6.72725081e-01
8.52753282e-01 6.50036633e-01 4.28367019e-01 -5.08501112e-01
1.33955181e-02 -2.35911265e-01 1.00395672e-01 -5.61980963e-01
9.45881844e-01 -2.00732812e-01 -2.45061681e-01 1.22667503e+00
2.93976575e-01 1.00074507e-01 2.55718734e-02 6.04235649e-01
1.01867974e+00 7.39545599e-02 -1.06308997e-01 -1.51968449e-01
3.79597366e-01 -2.18883961e-01 3.07256356e-02 9.54188704e-01
3.47315729e-01 2.51337439e-01 3.46146047e-01 -2.83839852e-01
-7.65023232e-01 -7.03395963e-01 -5.07768281e-02 2.31316304e+00
-3.10298324e-01 -6.11063182e-01 -5.34219265e-01 -1.05551493e+00
-2.76658386e-02 1.04934847e+00 -5.90757668e-01 -5.55062070e-02
-2.87066698e-01 -7.89528966e-01 2.99766082e-02 3.79182637e-01
-1.22604296e-01 -1.09239781e+00 -3.51364240e-02 4.74346876e-01
-6.88870430e-01 -1.16528058e+00 -9.87192810e-01 -4.51842099e-02
-7.94171572e-01 -6.44868374e-01 -5.26734591e-01 -5.95215559e-01
4.29248780e-01 4.64822233e-01 1.38855135e+00 -2.15755045e-01
5.07302523e-01 7.09947169e-01 -6.70194328e-01 -2.98991837e-02
-5.24866819e-01 2.68056750e-01 1.17231920e-01 2.90798426e-01
2.78876364e-01 -7.05922067e-01 -5.06208539e-01 9.35805202e-01
-4.11824822e-01 3.38119954e-01 2.04555005e-01 9.75533485e-01
-7.26271644e-02 -6.14365816e-01 1.22908199e+00 -1.41427732e+00
1.32061040e+00 -6.67594373e-01 3.02183088e-02 2.35887766e-01
-8.44433308e-01 -6.78303465e-02 6.21578932e-01 -5.24238467e-01
-1.51136959e+00 -1.79195359e-01 -3.73478651e-01 2.65490741e-01
2.85768881e-02 5.48798740e-01 1.88952714e-01 2.29213268e-01
8.04598510e-01 8.62704441e-02 -1.03856549e-01 -7.96539903e-01
6.92614555e-01 1.05493045e+00 1.48022115e-01 -9.34673548e-01
2.19125316e-01 -6.15672655e-02 -6.46606266e-01 -2.34781593e-01
-1.46462667e+00 -5.88845909e-01 -3.08053046e-01 -3.83236885e-01
6.42952442e-01 -8.71820033e-01 -8.89025569e-01 -7.36367330e-02
-1.04385555e+00 -6.02359414e-01 -6.47801384e-02 5.98053634e-01
-6.99469626e-01 2.46546030e-01 -1.13922739e+00 -8.36177886e-01
-6.21047556e-01 -1.03765035e+00 9.95373726e-01 -1.39711335e-01
-4.36193943e-01 -1.18630636e+00 2.96393394e-01 9.75470841e-01
6.84626222e-01 -8.19617093e-01 8.41393709e-01 -1.57993734e+00
7.26923496e-02 -1.45899475e-01 1.53829798e-01 2.57486254e-01
5.86335286e-02 -4.91449654e-01 -1.03721535e+00 -3.99354696e-01
1.25259250e-01 -8.17665696e-01 6.62630916e-01 6.79353178e-02
7.56408691e-01 -5.44877470e-01 -7.42220730e-02 -7.24355653e-02
3.96151245e-01 1.02408864e-01 1.92748234e-01 -1.20479465e-01
3.86670679e-01 1.02920854e+00 8.85812283e-01 5.76858461e-01
8.63503635e-01 1.10959053e+00 4.89741974e-02 2.38811567e-01
2.43713856e-01 -2.84495056e-01 5.87238491e-01 1.11063075e+00
-9.43197981e-02 -6.19147182e-01 -1.82086140e-01 2.40837112e-01
-2.19057250e+00 -8.62277269e-01 1.28029212e-01 2.21915960e+00
1.08688259e+00 1.47647485e-01 5.91060042e-01 -4.86064583e-01
4.95732367e-01 -7.48656839e-02 -3.82448405e-01 -6.31852806e-01
3.88131917e-01 -7.06446543e-02 -1.35888323e-01 7.43965626e-01
-9.50802207e-01 8.80480826e-01 6.57247353e+00 6.36319160e-01
-8.62248063e-01 3.29003125e-01 7.46090889e-01 -3.95585716e-01
-5.06954432e-01 -3.31095934e-01 -9.49637175e-01 3.08078110e-01
1.18211854e+00 -1.68452501e-01 7.64308155e-01 6.95481300e-01
2.86489189e-01 2.70467103e-01 -1.48934639e+00 4.70877498e-01
2.81512260e-01 -1.05887341e+00 8.70215148e-02 1.04413554e-01
6.19330406e-01 1.19930711e-02 6.23252429e-02 7.90470302e-01
6.95130289e-01 -7.36695707e-01 3.55513394e-01 3.91694635e-01
4.42655474e-01 -3.86936277e-01 6.61556304e-01 5.90909064e-01
-6.50104463e-01 -1.52060866e-01 -4.22539204e-01 -1.63879246e-01
3.43270391e-01 2.30845839e-01 -1.23108172e+00 3.38908970e-01
4.42328870e-01 7.73903251e-01 -1.74000561e-01 4.96425360e-01
-1.29292412e-02 8.61460030e-01 1.16221324e-01 -2.66158521e-01
3.10717616e-02 -3.43640178e-01 4.52445298e-01 1.36304045e+00
2.38433510e-01 2.74649978e-01 3.74351770e-01 3.64006758e-01
-3.84794623e-01 6.47764444e-01 -2.82419801e-01 8.21598172e-02
2.88553685e-01 1.85310328e+00 -2.03717187e-01 -3.45436394e-01
-6.05657697e-01 8.37516844e-01 5.81032157e-01 3.35916758e-01
-4.72086400e-01 4.24272902e-02 4.90698397e-01 -1.06106624e-01
6.21426284e-01 1.57551289e-01 1.31094098e-01 -1.19487774e+00
-3.43334675e-01 -1.13657200e+00 3.91319185e-01 -7.11107314e-01
-1.80535257e+00 7.66977727e-01 -2.33773038e-01 -1.07707179e+00
-8.08667541e-01 -3.89524519e-01 -6.78211331e-01 7.93668211e-01
-1.06318319e+00 -1.00845349e+00 1.21281162e-01 6.89510405e-01
9.02250886e-01 -4.21256006e-01 1.23245907e+00 4.74812806e-01
-3.12833607e-01 8.22518229e-01 3.53546858e-01 -1.95297480e-01
1.45411122e+00 -1.35969210e+00 3.20935518e-01 1.52790651e-01
2.57655501e-01 6.60694063e-01 7.72476971e-01 -2.88222313e-01
-1.25775611e+00 -9.85416055e-01 1.08540142e+00 -8.60013783e-01
6.67118192e-01 -6.91268384e-01 -5.37785232e-01 9.04577851e-01
3.36284518e-01 -4.70008552e-01 1.23096848e+00 1.01828182e+00
-3.65861684e-01 -1.13998344e-02 -1.05677307e+00 4.55682099e-01
1.04471743e+00 -5.78021586e-01 -6.28986537e-01 7.05750406e-01
8.79228115e-01 -4.18294996e-01 -1.23273039e+00 1.58075064e-01
7.66175210e-01 -6.82108760e-01 8.38941813e-01 -1.44131088e+00
7.96483397e-01 4.10411716e-01 -2.89520144e-01 -1.63369298e+00
-5.80012441e-01 -7.84550309e-01 -3.39421853e-02 1.28345811e+00
9.23453689e-01 -4.03913826e-01 4.52078402e-01 8.59373927e-01
-3.56432855e-01 -7.48162389e-01 -6.08867586e-01 -1.02949575e-01
-1.11679416e-02 -3.08740705e-01 2.55212814e-01 8.10760140e-01
6.08691990e-01 1.41168451e+00 -1.01894855e+00 -3.77690822e-01
3.47415321e-02 4.48528141e-01 8.92564833e-01 -1.33952999e+00
-9.38942134e-01 -2.00975344e-01 7.44538426e-01 -1.77527964e+00
1.46598861e-01 -9.00765896e-01 3.48583937e-01 -1.39122140e+00
3.59347850e-01 -4.94878560e-01 -3.55829448e-01 1.19287163e-01
-2.50523508e-01 1.03400677e-01 1.39885873e-01 1.58770740e-01
-1.23933816e+00 5.29908419e-01 1.30961776e+00 1.20765112e-01
-1.78443581e-01 7.69145548e-01 -1.13696432e+00 4.40265000e-01
3.82864088e-01 -4.28408414e-01 -7.27940261e-01 -3.19253981e-01
5.29220700e-01 5.60545564e-01 -4.20077384e-01 -2.14549303e-01
2.24098787e-01 1.07766036e-02 -1.13292374e-01 -9.07539800e-02
7.75123537e-01 -5.49392819e-01 -2.75752336e-01 -3.47327501e-01
-1.36830688e+00 -1.41673191e-02 -6.32119253e-02 7.67804146e-01
1.77634180e-01 -2.52355188e-01 1.56363785e-01 2.49583796e-02
9.76419151e-02 1.92610502e-01 -7.30142117e-01 1.98995739e-01
4.15606230e-01 3.87393087e-01 -5.36366463e-01 -1.09742975e+00
-1.10575187e+00 2.82853901e-01 -1.66720703e-01 6.42798245e-01
2.93962181e-01 -1.15410531e+00 -9.43405151e-01 -2.45337531e-01
1.52698204e-01 -5.04816234e-01 3.82033646e-01 9.13618028e-01
6.53225303e-01 5.37443936e-01 -5.76012302e-03 -2.52114505e-01
-1.43743336e+00 4.07287151e-01 2.63323277e-01 -6.62177384e-01
-2.43710220e-01 1.02460492e+00 4.66360122e-01 -9.14512038e-01
4.87050802e-01 -1.24397486e-01 -9.21232641e-01 2.20047683e-01
6.72493279e-01 2.09809735e-01 9.83135328e-02 -2.65861213e-01
1.46861479e-01 2.18369230e-03 -5.83129525e-01 -3.78114253e-01
1.39014816e+00 -3.98964018e-01 5.16114198e-02 5.69431245e-01
7.89380729e-01 4.18694228e-01 -9.85177398e-01 -7.74816513e-01
-1.44687921e-01 -3.04124653e-01 -1.48400187e-01 -1.25946105e+00
-7.75777221e-01 5.51225781e-01 1.05995134e-01 6.19901001e-01
6.85845256e-01 1.02141343e-01 8.46945643e-01 8.58102918e-01
2.74769694e-01 -9.48043823e-01 3.57038945e-01 7.25669384e-01
9.50718820e-01 -1.23791277e+00 -3.45172614e-01 -2.77483493e-01
-1.19268155e+00 9.88810003e-01 6.55694425e-01 1.89063951e-01
6.55991495e-01 -7.51051754e-02 3.54002416e-02 -7.79808611e-02
-1.59334695e+00 2.19123259e-01 5.81991911e-01 9.03515741e-02
9.14178014e-01 -6.46229088e-02 -3.17920119e-01 1.26686144e+00
-1.58180714e-01 -6.75533116e-02 4.74596649e-01 3.26458663e-01
-5.21514356e-01 -1.27200806e+00 2.61149138e-01 9.33188498e-01
-5.30529559e-01 -3.44864279e-01 -4.40146238e-01 -1.77554473e-01
-5.39084435e-01 1.55778551e+00 -4.61359136e-03 -8.37427258e-01
5.04427433e-01 1.85633332e-01 1.60954326e-01 -9.26571965e-01
-1.19324839e+00 2.61519849e-01 1.16727793e+00 -3.83203894e-01
-3.90821695e-01 -6.66573942e-01 -7.22724259e-01 -2.19678044e-01
-5.04263818e-01 7.92048216e-01 3.74641120e-01 1.19884694e+00
5.28934240e-01 4.48252976e-01 1.02719021e+00 -8.32629502e-01
-1.11284459e+00 -1.50436187e+00 -5.02058208e-01 5.29438794e-01
-8.66341684e-03 -5.65051019e-01 -3.02410364e-01 -1.99800551e-01] | [12.38705825805664, 7.570472717285156] |
2f03a2d7-eb37-46cb-8b58-a7d0cdddee02 | are-pre-trained-language-models-useful-for | 2305.15183 | null | https://arxiv.org/abs/2305.15183v1 | https://arxiv.org/pdf/2305.15183v1.pdf | Are Pre-trained Language Models Useful for Model Ensemble in Chinese Grammatical Error Correction? | Model ensemble has been in widespread use for Grammatical Error Correction (GEC), boosting model performance. We hypothesize that model ensemble based on the perplexity (PPL) computed by pre-trained language models (PLMs) should benefit the GEC system. To this end, we explore several ensemble strategies based on strong PLMs with four sophisticated single models. However, the performance does not improve but even gets worse after the PLM-based ensemble. This surprising result sets us doing a detailed analysis on the data and coming up with some insights on GEC. The human references of correct sentences is far from sufficient in the test data, and the gap between a correct sentence and an idiomatic one is worth our attention. Moreover, the PLM-based ensemble strategies provide an effective way to extend and improve GEC benchmark data. Our source code is available at https://github.com/JamyDon/PLM-based-CGEC-Model-Ensemble. | ['Yunfang Wu', 'Xiuyu Wu', 'Chenming Tang'] | 2023-05-24 | null | null | null | null | ['grammatical-error-correction'] | ['natural-language-processing'] | [-2.32126638e-01 -4.42886120e-03 1.46760553e-01 -5.36810935e-01
-8.45427394e-01 -3.05029392e-01 6.13358200e-01 3.51977736e-01
-4.51256484e-01 8.86129677e-01 2.13529661e-01 -5.85438013e-01
1.67697862e-01 -5.99409699e-01 -5.75979352e-01 -6.00520611e-01
5.37507832e-02 4.32398885e-01 -8.93316716e-02 -6.82732165e-01
5.16440570e-01 1.77249815e-02 -1.42865765e+00 4.12874430e-01
1.55028152e+00 2.89576292e-01 4.10363287e-01 6.88828111e-01
-2.18736097e-01 6.86445117e-01 -5.44384420e-01 -8.46197307e-01
-2.81563610e-01 -6.56520486e-01 -8.42143536e-01 -7.07469165e-01
3.22963446e-01 1.54663906e-01 2.55203068e-01 1.24343824e+00
6.20180786e-01 -5.12196915e-03 6.19149983e-01 -8.92306924e-01
-6.22572243e-01 9.58647609e-01 -2.67569214e-01 8.93569216e-02
1.52561948e-01 8.62035155e-03 9.94243383e-01 -1.18123233e+00
7.21880853e-01 1.31312239e+00 7.79936373e-01 8.76699686e-01
-9.10511732e-01 -4.95427132e-01 2.77721852e-01 3.85928720e-01
-1.09056377e+00 -3.35159391e-01 5.65667927e-01 -2.17026815e-01
1.21621752e+00 3.52112114e-01 4.40199703e-01 1.30337965e+00
4.50433493e-01 7.23230362e-01 1.36963165e+00 -8.51472378e-01
-1.04434438e-01 2.71168768e-01 5.52492380e-01 6.55995905e-01
3.71242493e-01 1.79900333e-01 -3.46011311e-01 1.44180069e-02
-1.50138885e-02 -3.11133862e-01 -4.94241685e-01 4.48145449e-01
-8.91993999e-01 8.60465705e-01 1.39854863e-01 6.41389251e-01
-2.97779769e-01 1.26139596e-01 3.28524888e-01 3.54394138e-01
8.44331503e-01 5.95763206e-01 -7.56898046e-01 -4.48953599e-01
-8.89761686e-01 3.93116862e-01 8.14565539e-01 6.43399477e-01
4.58116144e-01 1.63099453e-01 -3.00200619e-02 1.14570367e+00
2.82758623e-01 6.59733415e-01 3.78702670e-01 -7.62345433e-01
5.89150250e-01 6.21532083e-01 -1.92198128e-01 -9.28419352e-01
-2.97711015e-01 -6.94634795e-01 -8.54970515e-01 2.85973456e-02
3.78627509e-01 -3.00722510e-01 -6.45413220e-01 1.80038404e+00
-6.15921579e-02 -2.00876649e-02 -1.35431722e-01 6.99874103e-01
7.76515722e-01 6.21851504e-01 4.75464821e-01 -2.20415339e-01
1.16912150e+00 -9.72195864e-01 -8.76367629e-01 -2.16470838e-01
1.23870814e+00 -8.67930174e-01 9.42063749e-01 4.70439255e-01
-9.31329489e-01 -5.21778584e-01 -9.32774484e-01 1.69723555e-02
-4.04581040e-01 3.83862436e-01 5.13054132e-01 5.55465460e-01
-9.03593838e-01 8.69233668e-01 -8.49019706e-01 -4.37818408e-01
5.61867431e-02 1.22919464e-02 -3.94106239e-01 -2.11704239e-01
-1.21321476e+00 1.41715598e+00 5.51079869e-01 1.87868029e-01
-4.54495877e-01 -4.05510426e-01 -6.34446561e-01 -1.59408420e-01
1.10718243e-01 -5.48029661e-01 1.16625345e+00 -8.87237549e-01
-1.34196544e+00 8.03843081e-01 -4.37430054e-01 -5.12211561e-01
3.71527404e-01 -3.15511405e-01 -5.35758436e-01 -4.50368822e-01
-1.50028750e-01 4.76143718e-01 5.12843132e-01 -1.33970237e+00
-5.43893695e-01 -3.99253011e-01 -3.48520249e-01 -3.59706059e-02
8.79323632e-02 4.49327439e-01 -5.67147508e-03 -6.24032915e-01
5.95853813e-02 -9.56871271e-01 -3.96191292e-02 -8.48400354e-01
-3.71293157e-01 -5.17095745e-01 3.34940910e-01 -1.16381109e+00
1.84642601e+00 -1.76212847e+00 1.30396113e-01 -7.25565478e-02
-3.49640287e-02 7.00489938e-01 -2.88902223e-01 6.62894905e-01
-2.44100213e-01 5.03532529e-01 -3.41406912e-01 -7.58224785e-01
-1.45208657e-01 1.63565859e-01 -1.32507026e-01 -1.43079951e-01
3.48561794e-01 8.94031703e-01 -8.74788642e-01 -3.30180675e-01
-1.27895445e-01 3.74625117e-01 -6.37944639e-01 -8.57609957e-02
-2.01347083e-01 6.58236086e-01 -1.78959444e-02 6.20766163e-01
8.29403698e-01 -1.59563106e-02 3.73126984e-01 2.65201896e-01
-1.50216833e-01 7.46353626e-01 -8.33275974e-01 1.48264515e+00
-3.92857313e-01 4.06956553e-01 -3.37891042e-01 -8.51215720e-01
1.05524576e+00 8.29258412e-02 -1.73041776e-01 -6.76344514e-01
1.55495912e-01 8.10822606e-01 4.78040487e-01 -2.41873294e-01
7.66349435e-01 -7.63241872e-02 3.92953083e-02 4.85298306e-01
4.07705843e-01 1.14281893e-01 4.94431287e-01 2.39433512e-01
8.65488470e-01 2.01715529e-01 3.42647225e-01 -4.79932874e-01
6.17275000e-01 1.09462701e-02 8.57081652e-01 7.39083707e-01
-1.21221468e-01 5.30539036e-01 3.92536581e-01 -1.82841539e-01
-9.08895969e-01 -6.35508776e-01 -1.44525260e-01 8.31550539e-01
-2.87389964e-01 -9.51269031e-01 -9.40107465e-01 -1.04120255e+00
-2.99763680e-01 1.31852198e+00 -3.78969312e-01 -1.65684178e-01
-7.82692313e-01 -1.04284739e+00 5.56074262e-01 4.74383891e-01
4.43471044e-01 -1.07787287e+00 -6.76870123e-02 3.78718376e-01
-6.71353877e-01 -7.06489801e-01 -1.37263358e-01 1.30179331e-01
-1.04524875e+00 -9.21007991e-01 -3.78319800e-01 -5.21622419e-01
5.81332684e-01 -1.77601054e-01 1.45611811e+00 6.34973526e-01
3.38694215e-01 7.46622160e-02 -8.56896579e-01 -8.46164644e-01
-9.64100599e-01 2.65628755e-01 4.66476344e-02 -4.04937863e-01
5.31782448e-01 -4.14657593e-01 -2.18414679e-01 9.32993251e-04
-2.58126169e-01 2.58687168e-01 2.23185658e-01 1.03090346e+00
2.34792084e-01 -3.15689147e-01 7.07552969e-01 -1.04634643e+00
6.31227374e-01 -3.91260684e-01 -2.85742760e-01 5.19928455e-01
-9.43474472e-01 8.24931115e-02 3.53106201e-01 -5.18742874e-02
-1.25858128e+00 -5.65821290e-01 -7.63864040e-01 2.96243936e-01
5.14653474e-02 7.03509033e-01 3.04913614e-02 2.21064761e-01
6.51803195e-01 9.54377949e-02 -9.22134295e-02 -8.72801542e-01
1.34077609e-01 8.01559448e-01 1.68607235e-01 -6.93220556e-01
3.04062009e-01 -1.54133841e-01 -3.99091661e-01 -3.88564318e-01
-8.62831354e-01 -3.00722779e-03 -6.26154363e-01 -2.10308567e-01
5.46154201e-01 -8.40684474e-01 -3.30523551e-01 7.02751637e-01
-1.74367440e+00 -3.83414388e-01 2.75475472e-01 6.21239483e-01
-2.34020680e-01 3.90445977e-01 -7.68070936e-01 -9.90945041e-01
-6.63867772e-01 -1.03404248e+00 7.04602063e-01 1.79591656e-01
-3.42575520e-01 -1.18584263e+00 2.22274989e-01 3.86341631e-01
5.79852581e-01 2.34853216e-02 1.14337230e+00 -9.22045648e-01
-3.24769378e-01 -1.94103926e-01 3.40419672e-02 7.22489595e-01
-2.18274742e-01 1.99842736e-01 -8.78706396e-01 -2.53086835e-01
-6.58838004e-02 9.28058922e-02 1.13704419e+00 2.06514373e-01
1.04422235e+00 -1.43131271e-01 -1.91225603e-01 2.51193136e-01
1.51253164e+00 6.80126995e-02 7.71184325e-01 2.65547752e-01
5.01684368e-01 4.44863766e-01 6.01012945e-01 1.58079147e-01
6.26185596e-01 5.62598705e-01 3.13227355e-01 3.79634559e-01
-2.77085871e-01 -3.63483429e-01 6.00112259e-01 1.48147917e+00
-5.97907364e-01 -6.22040808e-01 -1.13496757e+00 4.27907050e-01
-1.82505894e+00 -8.25271845e-01 -8.43464792e-01 2.14182281e+00
8.62612724e-01 1.35462210e-02 -2.82081693e-01 8.93342346e-02
6.86760485e-01 -9.37882885e-02 1.87725022e-01 -7.79994607e-01
-4.68807101e-01 3.07094693e-01 1.13788888e-01 6.92261398e-01
-8.11305881e-01 9.97964919e-01 5.85269547e+00 9.71867919e-01
-1.00098228e+00 5.16412735e-01 5.43203175e-01 2.53510654e-01
-4.51806098e-01 3.16291392e-01 -1.24108005e+00 8.63293767e-01
1.40597725e+00 -4.08940278e-02 2.34835744e-01 5.89869380e-01
1.59765333e-01 -3.00955117e-01 -7.15774536e-01 7.27485895e-01
1.95475355e-01 -1.19496286e+00 -5.29892556e-02 6.20152391e-02
8.76537919e-01 3.44835371e-01 -3.00027162e-01 8.03419650e-01
3.37114394e-01 -8.27868342e-01 5.58153510e-01 7.99877822e-01
5.21035552e-01 -5.12076139e-01 1.26012063e+00 7.47471273e-01
-5.14192700e-01 1.71041880e-02 -4.34730381e-01 -1.86192423e-01
2.37096548e-01 7.45020330e-01 -5.37473500e-01 9.37450171e-01
7.35764742e-01 7.12861478e-01 -9.99370933e-01 8.65906358e-01
-5.78658104e-01 1.14233828e+00 -1.18244134e-01 -1.41919449e-01
-7.66192600e-02 -2.39487603e-01 7.34190524e-01 1.44300675e+00
6.54098213e-01 2.10319888e-02 -3.83737653e-01 7.96024919e-01
2.46101227e-02 3.45487624e-01 -4.40752834e-01 8.91103595e-03
4.72628117e-01 1.27146006e+00 -2.89651394e-01 -4.48251635e-01
-2.90523440e-01 9.93807673e-01 7.27391541e-01 1.07309081e-01
-9.04560566e-01 -2.18768835e-01 3.74575198e-01 -1.00138240e-01
-1.18639298e-01 -1.69668898e-01 -5.77827334e-01 -1.39828825e+00
2.40687677e-03 -1.12253487e+00 2.47631595e-01 -6.36264920e-01
-1.32974851e+00 8.06486964e-01 -1.87391564e-01 -1.10392869e+00
-2.84366310e-01 -7.40708590e-01 -9.29229975e-01 1.10809565e+00
-1.42839539e+00 -1.05844498e+00 -1.68848723e-01 3.84422578e-02
4.89704102e-01 -2.08648741e-01 1.11949372e+00 4.53765541e-01
-7.73047924e-01 8.06493163e-01 2.61040688e-01 -3.14783379e-02
9.48158443e-01 -1.27347803e+00 4.43230659e-01 1.10541904e+00
1.91179782e-01 8.61733913e-01 6.84422135e-01 -7.74407566e-01
-6.17948711e-01 -9.69802916e-01 1.89385986e+00 -1.00924170e+00
3.10680419e-01 -1.16399236e-01 -1.27617633e+00 6.24227703e-01
3.42803657e-01 -4.60568905e-01 5.28437912e-01 5.11173189e-01
-2.42419213e-01 9.72280949e-02 -8.75547945e-01 4.34569150e-01
1.03121459e+00 -2.86827743e-01 -6.07518792e-01 1.04401939e-01
4.65653300e-01 -3.85974556e-01 -7.54582465e-01 6.28828883e-01
2.67060697e-01 -1.08101249e+00 3.73515993e-01 -7.93475211e-01
7.85213649e-01 -1.97743744e-01 -2.57566929e-01 -1.74940431e+00
-2.46858656e-01 -1.08715981e-01 2.68796701e-02 1.64233911e+00
7.68020689e-01 -8.74443054e-01 1.00081168e-01 4.28165346e-01
-4.73719746e-01 -7.82739043e-01 -6.47651732e-01 -9.26403284e-01
5.62146187e-01 -7.95937121e-01 5.13900340e-01 9.92515922e-01
7.30071217e-02 3.55473787e-01 -4.15675938e-01 -8.94196779e-02
3.24403405e-01 -2.34074086e-01 5.93524992e-01 -1.25296986e+00
-2.62717724e-01 -3.85575056e-01 3.19292955e-02 -6.33045137e-01
2.87473708e-01 -1.29076755e+00 -5.84299155e-02 -1.36537135e+00
3.00914764e-01 -4.46674019e-01 -1.95298746e-01 5.11689425e-01
-6.50920808e-01 9.61426366e-03 5.33963978e-01 1.22736655e-01
-5.42144895e-01 5.83609462e-01 9.27375138e-01 8.81512314e-02
8.48032013e-02 -1.00601688e-01 -7.75565207e-01 6.90307140e-01
1.18712425e+00 -7.27578521e-01 2.17886165e-01 -5.94022632e-01
3.73593479e-01 -1.56741902e-01 3.35938245e-01 -8.09176385e-01
-5.67732528e-02 1.78846568e-01 2.53218897e-02 -5.16523778e-01
-1.09036863e-01 -2.06543535e-01 2.09682554e-01 5.44887424e-01
-6.85637072e-02 2.55496681e-01 1.88424215e-01 1.35732383e-01
-3.44907939e-01 -6.65879369e-01 7.48647809e-01 -2.00766265e-01
-6.42682374e-01 -1.24680720e-01 -2.69348204e-01 -9.81714204e-02
2.70951569e-01 1.85827881e-01 -5.68337142e-01 -1.65682539e-01
-6.59043849e-01 2.53592402e-01 3.43355358e-01 4.75175679e-01
2.88448781e-01 -1.27367890e+00 -1.22183704e+00 1.09967873e-01
2.73518860e-01 -4.78205204e-01 4.00763929e-01 1.26183593e+00
-4.20518279e-01 5.40800691e-01 4.21826206e-02 -3.38410974e-01
-1.51535523e+00 1.61180776e-02 3.93527120e-01 -6.86395943e-01
-1.94973499e-01 9.30350602e-01 -3.53612453e-01 -9.34076190e-01
-2.16963962e-01 -1.49338618e-01 -1.88569099e-01 -1.26136184e-01
4.91553098e-01 4.88248080e-01 3.60691786e-01 -5.47228038e-01
-3.59582424e-01 3.51874620e-01 -2.10793972e-01 1.27786160e-01
1.33667731e+00 -1.99184984e-01 -4.36955303e-01 5.67918897e-01
8.04636300e-01 8.90418589e-02 -5.91826320e-01 -1.38129935e-01
3.29400778e-01 -1.83483690e-01 -1.33078024e-01 -1.38993669e+00
-6.44992054e-01 1.07810903e+00 4.03538167e-01 -1.35554776e-01
1.00996006e+00 -2.15717360e-01 5.64874768e-01 1.60316125e-01
6.05474830e-01 -1.20617008e+00 -2.45735884e-01 1.15164089e+00
1.14696968e+00 -1.55666268e+00 -2.92157888e-01 -2.77041167e-01
-7.99597740e-01 1.07433522e+00 7.35374689e-01 -1.56919524e-01
6.21248305e-01 -9.53465626e-02 6.78734556e-02 -9.64480117e-02
-1.12413538e+00 -1.73810259e-01 2.30469882e-01 3.76078814e-01
9.67483580e-01 3.43848586e-01 -1.18213129e+00 1.03483856e+00
-3.47788513e-01 -2.42168307e-01 4.67080832e-01 4.03167546e-01
-3.57764810e-01 -1.62287104e+00 -1.77330285e-01 3.42402995e-01
-5.63023090e-01 -4.75893974e-01 -3.74684095e-01 8.14368844e-01
3.71924758e-01 1.03313863e+00 -1.17877580e-01 -5.36510229e-01
2.07680717e-01 6.19338095e-01 5.92775226e-01 -5.83883464e-01
-8.01287830e-01 -2.63513923e-01 4.66526210e-01 -3.52457166e-01
-2.74902642e-01 -7.62188911e-01 -1.10375416e+00 -6.32892132e-01
-6.55355752e-01 2.32543901e-01 7.59764671e-01 9.33673680e-01
4.42123175e-01 3.21242988e-01 3.88979316e-01 -4.87729669e-01
-5.87283909e-01 -1.59740007e+00 -3.12958658e-01 3.10854852e-01
-1.00409344e-01 -4.54004586e-01 -5.00456929e-01 -2.50658005e-01] | [11.090837478637695, 10.703198432922363] |
de8653fb-751e-48ba-ab16-6a9a3d998c4b | rgb-d-salient-object-detection-a-survey | 2008.00230 | null | https://arxiv.org/abs/2008.00230v4 | https://arxiv.org/pdf/2008.00230v4.pdf | RGB-D Salient Object Detection: A Survey | Salient object detection (SOD), which simulates the human visual perception system to locate the most attractive object(s) in a scene, has been widely applied to various computer vision tasks. Now, with the advent of depth sensors, depth maps with affluent spatial information that can be beneficial in boosting the performance of SOD, can easily be captured. Although various RGB-D based SOD models with promising performance have been proposed over the past several years, an in-depth understanding of these models and challenges in this topic remains lacking. In this paper, we provide a comprehensive survey of RGB-D based SOD models from various perspectives, and review related benchmark datasets in detail. Further, considering that the light field can also provide depth maps, we review SOD models and popular benchmark datasets from this domain as well. Moreover, to investigate the SOD ability of existing models, we carry out a comprehensive evaluation, as well as attribute-based evaluation of several representative RGB-D based SOD models. Finally, we discuss several challenges and open directions of RGB-D based SOD for future research. All collected models, benchmark datasets, source code links, datasets constructed for attribute-based evaluation, and codes for evaluation will be made publicly available at https://github.com/taozh2017/RGBDSODsurvey | ['Ming-Ming Cheng', 'Deng-Ping Fan', 'Jianbing Shen', 'Tao Zhou', 'Ling Shao'] | 2020-08-01 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 8.23387969e-03 -1.40143633e-01 -3.18030924e-01 -3.74017984e-01
-4.87069070e-01 -3.14092785e-01 3.02020758e-01 9.47051346e-02
-1.47546962e-01 5.22390425e-01 1.00973167e-01 -1.13386428e-02
-6.45980388e-02 -7.53376722e-01 -3.85602146e-01 -9.39703286e-01
3.51809636e-02 -1.35811031e-01 6.53083503e-01 -3.03312093e-01
2.77581811e-01 7.52294242e-01 -2.03862715e+00 1.55850440e-01
9.03666437e-01 1.61270821e+00 6.41760588e-01 3.09958339e-01
8.03070962e-02 6.01720273e-01 -3.62711221e-01 -3.50147396e-01
5.12938380e-01 -1.49321362e-01 -4.86508578e-01 1.99802697e-01
5.81246912e-01 -4.26561952e-01 -8.14317226e-01 1.13286746e+00
8.23955417e-01 -9.54971984e-02 5.18969715e-01 -1.57135427e+00
-8.85084867e-01 -6.08621128e-02 -6.59089446e-01 4.47974890e-01
4.18097109e-01 2.43634552e-01 9.11704957e-01 -1.13973439e+00
5.28954148e-01 1.03128159e+00 3.24000776e-01 4.14505959e-01
-6.41587019e-01 -5.65044463e-01 1.36854380e-01 5.37364125e-01
-1.17091918e+00 -1.95744693e-01 1.33363807e+00 -1.24260552e-01
6.35873795e-01 4.51387912e-01 9.84204769e-01 8.76299143e-01
5.34052216e-02 1.42426622e+00 1.24388182e+00 -2.79918879e-01
1.15334243e-01 3.11427176e-01 -3.89966294e-02 6.02179766e-01
2.02929065e-01 2.32726246e-01 -9.24856305e-01 1.28592148e-01
6.84909523e-01 1.47654235e-01 -2.16779023e-01 -8.84981453e-01
-1.17086256e+00 6.87603951e-01 1.20483816e+00 -2.02071249e-01
-3.50579321e-01 -1.37711465e-01 1.18360080e-01 -1.41088828e-01
4.74511147e-01 1.12722419e-01 -2.19505861e-01 2.49135554e-01
-3.95253897e-01 3.49254131e-01 6.20242991e-02 1.27445042e+00
9.09470379e-01 -9.80173051e-02 -1.34907275e-01 8.27080965e-01
3.57484132e-01 7.82672465e-01 2.91019857e-01 -9.27570820e-01
4.50805157e-01 9.65834618e-01 8.15857649e-02 -1.03213382e+00
-5.86922944e-01 -2.62921840e-01 -8.23479116e-01 3.70828599e-01
9.74358022e-02 3.46396029e-01 -7.68379986e-01 1.04091287e+00
5.72194099e-01 -1.12996414e-01 -7.07690418e-02 1.38126838e+00
1.59919000e+00 4.29158926e-01 -2.53928527e-02 5.90981655e-02
1.37911952e+00 -8.91394258e-01 -5.15759349e-01 -4.73117471e-01
1.97779477e-01 -7.41944909e-01 1.23631752e+00 3.94305825e-01
-1.06554949e+00 -5.98968625e-01 -9.91682172e-01 -5.68058431e-01
-5.00554323e-01 2.99207181e-01 1.08269846e+00 5.07374465e-01
-1.01966989e+00 1.06146485e-01 -7.08522379e-01 -6.60557151e-01
7.94506013e-01 7.91375339e-02 -2.76111931e-01 -3.79034758e-01
-1.20415974e+00 7.79939473e-01 1.88531578e-01 3.30724359e-01
-9.68257785e-01 -3.78871262e-01 -9.16508615e-01 -4.77741241e-01
2.63747931e-01 -5.07876098e-01 1.11169720e+00 -4.29503500e-01
-9.94455099e-01 1.15741873e+00 -2.49498129e-01 -2.10567638e-01
5.06007612e-01 -1.86705180e-02 -2.06520662e-01 3.07775527e-01
1.68689415e-01 9.57016766e-01 3.87261242e-01 -1.43115067e+00
-9.61561561e-01 -5.94010115e-01 3.44836295e-01 5.75725257e-01
-3.84614021e-01 4.76041399e-02 -7.38131821e-01 -4.77072448e-01
5.63438356e-01 -6.40999734e-01 -2.47530788e-01 7.19575584e-01
-5.11934042e-01 -2.81142682e-01 7.00475514e-01 -1.23642325e-01
1.06234419e+00 -2.14778423e+00 -7.96615854e-02 -3.51401448e-01
3.54741424e-01 1.46470517e-01 1.76001266e-01 2.77292699e-01
2.90305614e-01 -1.86665192e-01 -3.08643520e-01 -5.06672382e-01
-1.68801010e-01 1.83168650e-01 -2.44037524e-01 7.25338280e-01
1.45070255e-01 9.49197114e-01 -8.23596299e-01 -6.74235880e-01
7.04545736e-01 3.95998031e-01 -1.20626301e-01 2.16348216e-01
7.75068179e-02 3.71977657e-01 -7.77869225e-01 1.35472620e+00
9.72393811e-01 -1.16973914e-01 -5.66375017e-01 -3.76229078e-01
-3.03798914e-01 7.04979897e-02 -1.08114409e+00 1.59880733e+00
-2.64859460e-02 8.25568855e-01 -2.03259751e-01 -8.30049098e-01
1.12776637e+00 -1.98119804e-01 5.82736969e-01 -1.22534454e+00
-5.17735370e-02 2.49643981e-01 -3.97083968e-01 -5.94602585e-01
5.63414633e-01 1.82464942e-01 6.20018020e-02 6.65342808e-02
-3.75570685e-01 -3.88616651e-01 1.01415977e-01 1.33297279e-01
5.54697216e-01 1.29385665e-01 4.08263743e-01 7.12091848e-02
4.65600818e-01 3.44931990e-01 5.58817029e-01 6.41418755e-01
-8.26661408e-01 9.29696023e-01 2.12409168e-01 -6.95726991e-01
-7.91021824e-01 -1.26562166e+00 -5.19896328e-01 8.02440584e-01
1.06830478e+00 -6.70237243e-02 -2.24500656e-01 -4.03123617e-01
3.29961061e-01 3.06556463e-01 -8.46631706e-01 -1.50166750e-01
-2.72378981e-01 -9.35354233e-01 2.62015402e-01 6.84678137e-01
1.04608655e+00 -1.17842531e+00 -9.28690970e-01 -1.08542331e-01
-3.57561737e-01 -1.17174625e+00 1.44227162e-01 1.27491966e-01
-1.13899684e+00 -1.09973145e+00 -8.62338901e-01 -7.72235036e-01
5.26358187e-01 1.05058324e+00 1.09712505e+00 2.76848841e-02
-5.66860855e-01 3.36176276e-01 -4.97989744e-01 -1.02625823e+00
3.40359330e-01 -1.68826491e-01 1.12588562e-01 -3.19578767e-01
5.40453613e-01 -2.01336145e-01 -1.19750392e+00 6.35729432e-01
-8.15275490e-01 3.00019741e-01 7.27336466e-01 3.10924321e-01
8.43159676e-01 -2.28586644e-01 2.76125282e-01 -1.69492751e-01
6.47357181e-02 -3.84089589e-01 -5.60138047e-01 1.68571156e-02
-5.90669811e-01 -5.00637949e-01 2.02612672e-02 1.11902677e-01
-9.26025867e-01 2.13855848e-01 -1.58636212e-01 -2.80294240e-01
-2.79547989e-01 6.99779578e-03 -2.39380628e-01 -2.94323683e-01
7.14242339e-01 4.43898827e-01 -9.56134945e-02 -7.34217703e-01
2.90839821e-01 9.04593229e-01 3.93493831e-01 -2.71834046e-01
6.00911140e-01 9.82744575e-01 7.21911117e-02 -8.38084757e-01
-1.14354956e+00 -8.03945422e-01 -6.46190345e-01 -4.66742069e-01
6.29033983e-01 -1.08996141e+00 -6.64138377e-01 7.89572179e-01
-9.71765578e-01 -2.35169962e-01 -1.30093724e-01 4.34128910e-01
-5.79853356e-01 2.65100330e-01 -3.66405159e-01 -7.80442715e-01
-1.30198210e-01 -1.22950125e+00 1.33919489e+00 6.69569850e-01
3.83697689e-01 -7.59226918e-01 -2.39677802e-01 5.42990744e-01
2.89427340e-01 2.54361898e-01 5.99517465e-01 2.43144594e-02
-9.86815929e-01 -6.50620311e-02 -6.57988012e-01 1.79684624e-01
2.06830382e-01 -2.17006922e-01 -1.30756283e+00 -7.16465935e-02
-5.69469901e-03 -3.58178854e-01 9.89462554e-01 7.33254492e-01
1.26408780e+00 1.36339873e-01 -6.03731155e-01 9.44752455e-01
1.50484562e+00 1.97352748e-02 6.84666276e-01 8.72715831e-01
6.91685736e-01 6.73504293e-01 1.16984010e+00 5.54228067e-01
6.93219423e-01 7.78676331e-01 1.05511534e+00 -5.16528726e-01
-4.24611092e-01 3.69998440e-02 -6.06970526e-02 3.42851102e-01
-2.97093481e-01 -2.94162959e-01 -9.79968727e-01 4.91363078e-01
-1.59675407e+00 -6.60320461e-01 -3.85007381e-01 1.95045924e+00
5.73538482e-01 7.88896158e-02 3.09481353e-01 2.85676271e-01
5.63020527e-01 2.44426027e-01 -9.22484875e-01 1.57484397e-01
-7.75459945e-01 -3.28507096e-01 6.47080600e-01 3.87406647e-02
-1.31324661e+00 8.45544040e-01 6.27075386e+00 5.68280220e-01
-1.10200405e+00 -8.88867825e-02 7.25263178e-01 -2.13988632e-01
-6.94844499e-02 -2.21028790e-01 -9.13236141e-01 3.23227972e-01
1.06719555e-02 -1.82734653e-01 -4.07485664e-02 1.21387899e+00
3.72085601e-01 -6.93373680e-01 -7.90332496e-01 1.37711179e+00
2.13186055e-01 -1.20717990e+00 -3.69690941e-04 5.47796721e-03
6.87849402e-01 3.32195461e-01 3.83883208e-01 -3.10274381e-02
-7.82201216e-02 -7.15344965e-01 7.78925717e-01 2.75673240e-01
6.23503268e-01 -5.84968448e-01 8.20634484e-01 1.68951169e-01
-1.33208334e+00 -2.40183741e-01 -9.40727234e-01 -2.19803676e-01
-1.15800790e-01 6.17356181e-01 -3.95555705e-01 7.51029968e-01
1.40777755e+00 1.24735463e+00 -1.06045580e+00 1.61148226e+00
-3.92365366e-01 1.51705191e-01 -1.97917223e-01 -1.81509808e-01
1.76532209e-01 -6.73784986e-02 3.88925523e-01 9.09787834e-01
3.43513489e-01 2.33331487e-01 -1.51178502e-02 7.19159842e-01
2.40110382e-01 -1.00982161e-02 -5.87897718e-01 4.58093017e-01
4.25266176e-01 1.27891457e+00 -6.94145620e-01 -5.87324016e-02
-6.20438576e-01 9.02097285e-01 1.96101904e-01 4.01082218e-01
-7.73758888e-01 -3.24116528e-01 9.78461385e-01 2.12259918e-01
2.51426734e-02 -1.90835461e-01 -5.37126958e-01 -8.89940917e-01
1.87023208e-01 -5.37410855e-01 4.46418852e-01 -1.38329160e+00
-1.23328745e+00 4.53030795e-01 2.23477669e-02 -1.88236642e+00
4.70995754e-01 -8.56846035e-01 -3.05988640e-01 7.71737814e-01
-2.10248828e+00 -1.14765489e+00 -1.04624641e+00 5.93306065e-01
5.15793920e-01 6.95498511e-02 2.87624300e-01 2.02066556e-01
-5.57265282e-01 1.46008074e-01 1.51338443e-01 1.48135379e-01
7.20530570e-01 -1.09519184e+00 3.48117858e-01 6.24740660e-01
-4.47910205e-02 3.77016157e-01 6.73594654e-01 -4.74423051e-01
-1.59036362e+00 -1.01543033e+00 4.30610299e-01 -6.35050058e-01
2.87101775e-01 -3.37361008e-01 -7.23154902e-01 3.95015895e-01
-3.06033622e-02 4.69921142e-01 2.89196253e-01 -3.23191822e-01
3.10874004e-02 -4.69514132e-01 -1.00189328e+00 4.63927478e-01
1.41254306e+00 -2.53867716e-01 -3.77330333e-01 3.59329551e-01
5.88683903e-01 -7.04441011e-01 -6.12470627e-01 5.35123527e-01
4.27337587e-01 -1.60290956e+00 1.36622846e+00 -1.36489153e-01
4.30397451e-01 -5.26935756e-01 -3.21568161e-01 -9.47510600e-01
-2.38306731e-01 -1.79732349e-02 -1.40140951e-01 9.44732666e-01
-8.60546604e-02 -4.94490355e-01 1.04783082e+00 4.26129550e-01
-3.95209014e-01 -1.11600125e+00 -8.48926067e-01 -7.29079366e-01
-2.99538046e-01 -6.61571503e-01 5.60269654e-01 4.93211269e-01
-2.68940002e-01 -1.41092241e-01 -1.81243464e-01 3.45948040e-01
9.04261172e-01 3.87062103e-01 8.00204277e-01 -1.21976352e+00
3.47535878e-01 -5.16631007e-01 -8.21963787e-01 -1.27487898e+00
-4.46334809e-01 -7.30771184e-01 -7.41986036e-02 -2.17313957e+00
3.33214074e-01 -6.50304973e-01 -4.33394015e-01 4.12763029e-01
-3.09627086e-01 6.87830925e-01 1.02773324e-01 4.77820158e-01
-6.53050661e-01 8.37416768e-01 1.57216907e+00 -1.97138637e-01
-5.58937714e-02 9.13555324e-02 -8.10670733e-01 7.03021646e-01
6.48890793e-01 -3.16821724e-01 -4.75692004e-01 -4.14266616e-01
1.15818426e-01 -2.96161979e-01 6.75815344e-01 -1.00580871e+00
3.61552447e-01 -3.21006447e-01 7.98372805e-01 -1.11206889e+00
6.89857006e-01 -7.07617283e-01 -3.95881116e-01 3.92240375e-01
1.06379338e-01 -8.75304639e-02 2.55042762e-01 4.58720565e-01
-4.44206297e-01 1.61555275e-01 8.69443178e-01 -1.41326711e-01
-1.45166588e+00 6.71988010e-01 6.31894320e-02 -9.68595445e-02
1.36414433e+00 -8.32825243e-01 -4.55075264e-01 -2.78382480e-01
-3.34020466e-01 3.43512803e-01 6.13325775e-01 4.68127221e-01
1.12880468e+00 -1.52135241e+00 -6.01395130e-01 2.55296350e-01
8.02575946e-01 2.84769207e-01 4.49783415e-01 8.78669500e-01
-5.72801471e-01 4.93908346e-01 -5.50799489e-01 -8.91239941e-01
-1.10586190e+00 6.76154852e-01 2.94109195e-01 5.85609257e-01
-6.51700675e-01 1.07242501e+00 5.46146750e-01 -1.77974090e-01
4.04941112e-01 -3.93151999e-01 -2.62208581e-01 -6.58243597e-02
4.96692419e-01 4.38497424e-01 2.52608806e-01 -6.66755080e-01
-7.29202807e-01 8.59269619e-01 2.71705866e-01 3.76221895e-01
1.29246593e+00 -7.19182909e-01 1.66055746e-02 3.56797993e-01
9.25755978e-01 -2.94219464e-01 -1.54913080e+00 -4.47952837e-01
-3.85003597e-01 -9.48012471e-01 1.13722809e-01 -5.41138411e-01
-1.21115839e+00 1.04063952e+00 8.56223226e-01 1.55076593e-01
1.55927050e+00 4.31413412e-01 5.55145025e-01 2.08940864e-01
6.44560993e-01 -9.88597274e-01 3.77489090e-01 8.16523433e-02
1.03927302e+00 -1.88369441e+00 4.65744585e-01 -6.55185699e-01
-8.51831257e-01 8.94005895e-01 8.90423656e-01 1.00284979e-01
6.27584636e-01 2.05315296e-02 2.50044227e-01 -2.05998778e-01
-3.10086936e-01 -5.98293185e-01 3.27872515e-01 1.01355207e+00
2.12318286e-01 -1.33005396e-01 1.97833270e-01 3.50426108e-01
-1.66759476e-01 -1.59241781e-01 3.95914465e-01 9.42579329e-01
-7.28251874e-01 -6.79903746e-01 -4.15396929e-01 3.87245148e-01
2.14304030e-02 1.42488480e-02 -4.08834606e-01 8.93456459e-01
1.58658370e-01 1.01132560e+00 1.82034392e-02 -3.30736279e-01
5.88559330e-01 -5.97143531e-01 4.43030596e-01 -5.03070295e-01
1.50745427e-02 -2.31009603e-01 -2.50106990e-01 -7.79041350e-01
-6.95127189e-01 -6.57005548e-01 -9.95569289e-01 -3.52704793e-01
-1.89367577e-01 -4.44326609e-01 6.86528385e-01 5.21024644e-01
2.61118442e-01 1.81468546e-01 6.74087822e-01 -1.03293788e+00
-5.40927693e-04 -7.33259976e-01 -8.01271856e-01 2.85010785e-01
3.71717095e-01 -1.08131242e+00 -3.79636317e-01 -7.31760934e-02] | [9.637410163879395, -0.7927422523498535] |
b4e18d5a-2f6f-49dd-8824-5287b2c4538d | incremental-few-shot-text-classification-with | 2104.11882 | null | https://arxiv.org/abs/2104.11882v1 | https://arxiv.org/pdf/2104.11882v1.pdf | Incremental Few-shot Text Classification with Multi-round New Classes: Formulation, Dataset and System | Text classification is usually studied by labeling natural language texts with relevant categories from a predefined set. In the real world, new classes might keep challenging the existing system with limited labeled data. The system should be intelligent enough to recognize upcoming new classes with a few examples. In this work, we define a new task in the NLP domain, incremental few-shot text classification, where the system incrementally handles multiple rounds of new classes. For each round, there is a batch of new classes with a few labeled examples per class. Two major challenges exist in this new task: (i) For the learning process, the system should incrementally learn new classes round by round without re-training on the examples of preceding classes; (ii) For the performance, the system should perform well on new classes without much loss on preceding classes. In addition to formulating the new task, we also release two benchmark datasets in the incremental few-shot setting: intent classification and relation classification. Moreover, we propose two entailment approaches, ENTAILMENT and HYBRID, which show promise for solving this novel problem. | ['Philip Yu', 'Yihao Feng', 'Wenpeng Yin', 'Congying Xia'] | 2021-04-24 | null | https://aclanthology.org/2021.naacl-main.106 | https://aclanthology.org/2021.naacl-main.106.pdf | naacl-2021-4 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 6.05687261e-01 3.31689477e-01 -4.57101256e-01 -7.29799092e-01
-7.48002231e-01 -3.57815206e-01 5.88210583e-01 6.80570781e-01
-6.00460649e-01 8.91377628e-01 6.84741586e-02 -1.78814143e-01
1.55015633e-01 -9.71557677e-01 -2.58411974e-01 -4.65368271e-01
2.44093686e-01 8.82975519e-01 4.88251328e-01 -3.39039862e-01
1.17834702e-01 5.86248562e-02 -1.75591660e+00 7.04714954e-01
7.05208778e-01 1.15716219e+00 -9.12215374e-03 5.97777903e-01
-3.22198004e-01 1.07039022e+00 -2.68871784e-01 -5.93143642e-01
1.06879227e-01 -4.86785203e-01 -1.47309184e+00 2.71654397e-01
1.74942553e-01 -2.94349700e-01 -2.02022381e-02 7.86805034e-01
3.30172986e-01 6.78859174e-01 7.00207472e-01 -1.17340755e+00
-4.71094012e-01 1.07175493e+00 -3.58415812e-01 2.71907717e-01
4.74269152e-01 -2.11354166e-01 1.33702564e+00 -1.21010482e+00
6.99962974e-01 8.66923392e-01 6.13871813e-01 5.86056232e-01
-9.39339221e-01 -3.81667495e-01 6.41080678e-01 5.75490832e-01
-1.14553249e+00 -5.54214060e-01 6.76375806e-01 -3.27437550e-01
8.92359555e-01 3.52233320e-01 3.16549599e-01 9.75655973e-01
-3.24333459e-01 9.66493189e-01 6.11903131e-01 -6.47372901e-01
6.31322503e-01 4.48373646e-01 1.06679142e+00 2.21720114e-01
1.32560283e-01 -3.19217563e-01 -2.12136045e-01 -1.49685055e-01
-5.06891847e-01 4.40792620e-01 -2.37329766e-01 -2.16447026e-01
-9.76476729e-01 9.74781573e-01 1.37034059e-01 5.47058105e-01
-3.56609598e-02 -3.15885365e-01 6.64992988e-01 4.99715358e-01
8.80638182e-01 4.04090017e-01 -8.43639791e-01 -1.02090023e-01
-6.70844972e-01 1.88837126e-01 1.13982737e+00 1.20303619e+00
8.96649778e-01 -5.90535760e-01 -2.86425620e-01 9.54938352e-01
-1.15428165e-01 1.17797712e-02 9.23737824e-01 -3.41997176e-01
6.87694073e-01 8.86973500e-01 -2.66880710e-02 -6.06240094e-01
-5.98895013e-01 -3.40301782e-01 -8.84368062e-01 -5.09517252e-01
2.68709481e-01 -2.10819989e-01 -7.98393905e-01 1.39674532e+00
6.14994109e-01 4.42698076e-02 4.21537101e-01 2.91213572e-01
9.09679472e-01 9.33908343e-01 -7.11496770e-02 -8.34756017e-01
1.42805314e+00 -1.28670228e+00 -7.49373198e-01 -6.59278572e-01
1.14229715e+00 -4.92365360e-01 1.32717407e+00 2.13032544e-01
-5.87038279e-01 -5.39070606e-01 -8.94841790e-01 -1.02791816e-01
-6.45358264e-01 -1.49064407e-01 4.97818142e-01 4.42237258e-01
-4.12938297e-01 5.22606373e-01 -3.74016643e-01 -6.12841904e-01
3.27199101e-01 8.38644430e-02 -8.58408585e-02 -4.45362478e-01
-1.42585909e+00 7.26971269e-01 7.51698792e-01 -2.70306110e-01
-4.25658733e-01 -6.48784280e-01 -9.05024588e-01 2.72395402e-01
8.97008121e-01 -4.04076278e-01 1.72428322e+00 -9.64628875e-01
-1.28139579e+00 7.61750937e-01 -3.25728774e-01 -4.13474441e-01
5.11777639e-01 -6.56556040e-02 -3.00552785e-01 -9.14583877e-02
2.79308498e-01 1.48224548e-01 7.14051545e-01 -1.05129135e+00
-1.17199802e+00 -5.28698623e-01 7.22973272e-02 5.16638458e-01
-6.74869776e-01 -2.38601238e-01 1.25941932e-02 -5.06882191e-01
2.41319556e-02 -7.85380900e-01 -1.57136142e-01 -3.46054286e-01
-3.68877083e-01 -7.36178458e-01 9.94312823e-01 -2.83938497e-02
1.19524705e+00 -2.17402744e+00 -1.48251683e-01 -3.45702857e-01
1.63110331e-01 3.56854677e-01 5.56518659e-02 2.89375961e-01
-3.23137604e-02 -9.39403698e-02 -2.17067167e-01 -3.50282550e-01
-8.71318802e-02 2.01763153e-01 -6.39633536e-01 -2.36534812e-02
-1.39595881e-01 9.40387309e-01 -1.19666231e+00 -3.00639212e-01
-1.67160928e-01 -3.05391967e-01 -3.66686940e-01 2.28362143e-01
-4.11277503e-01 9.92589593e-02 -3.01166654e-01 3.81403834e-01
4.65662748e-01 -3.92506272e-01 1.44674018e-01 -2.98180617e-03
3.31464410e-01 1.61069915e-01 -1.16431701e+00 1.33828771e+00
-5.03320456e-01 4.54399705e-01 -6.46132827e-01 -1.59617400e+00
7.81209290e-01 4.11143512e-01 1.49151206e-01 -1.53753296e-01
2.85095960e-01 1.07540358e-02 -1.22644119e-01 -7.45536566e-01
5.23600221e-01 -5.67066908e-01 -2.86991686e-01 9.13552105e-01
2.71012872e-01 -4.26886342e-02 5.63426137e-01 1.77280560e-01
1.00145996e+00 -4.34461504e-01 8.00975144e-01 8.74575526e-02
4.51686233e-01 1.95179686e-01 5.02642572e-01 9.13372338e-01
-2.46375427e-01 3.96628827e-01 4.68068391e-01 -9.05077815e-01
-7.50520170e-01 -4.17577475e-01 -1.71187341e-01 1.71548033e+00
2.29334623e-01 -4.14577991e-01 -3.59209418e-01 -1.32530797e+00
-1.86460257e-01 9.79085982e-01 -7.66985178e-01 -3.93728465e-01
-2.33174577e-01 -8.95985186e-01 1.09863110e-01 4.34921890e-01
3.51066232e-01 -1.02455580e+00 -4.88147110e-01 3.46770376e-01
-3.87592793e-01 -1.26777864e+00 -5.85827112e-01 7.07611859e-01
-8.38261485e-01 -1.18227005e+00 -5.69905162e-01 -1.23467267e+00
8.23336482e-01 3.74449760e-01 8.94183099e-01 1.90940350e-02
-1.57652915e-01 1.53731674e-01 -1.07553124e+00 -4.28604692e-01
-4.33418036e-01 5.30295014e-01 1.04915621e-02 2.22671375e-01
7.47452378e-01 -2.96392083e-01 -1.37798429e-01 2.52568394e-01
-9.01674867e-01 1.98760152e-01 2.72954077e-01 1.04488146e+00
4.22425687e-01 3.24612170e-01 9.16356683e-01 -1.58815229e+00
6.33444786e-01 -6.46227002e-01 -3.56449112e-02 5.53342819e-01
-7.48054326e-01 -4.39680368e-02 1.09058630e+00 -8.40797305e-01
-1.14970386e+00 8.50780681e-02 -2.35233530e-02 2.58771926e-01
-2.29444578e-01 7.87996829e-01 -4.03468311e-03 5.06980121e-01
1.05981803e+00 1.95388392e-01 -4.05981839e-01 -3.85171086e-01
6.00586653e-01 1.03329825e+00 9.52031463e-02 -3.31176728e-01
6.89245164e-01 3.12173665e-01 -4.81445372e-01 -7.71578670e-01
-1.70986736e+00 -7.58134663e-01 -1.01611602e+00 1.35370865e-01
3.96335781e-01 -5.63403130e-01 -2.43591592e-01 4.32129174e-01
-1.00285506e+00 -4.23395276e-01 -8.88796687e-01 3.31043363e-01
-3.42648983e-01 4.51785982e-01 -5.08147061e-01 -6.48100853e-01
-5.81915319e-01 -7.76687682e-01 7.42087960e-01 2.13500336e-01
-2.96630383e-01 -1.03334117e+00 2.06498533e-01 3.02979380e-01
1.32596031e-01 -2.60535508e-01 1.04036009e+00 -1.54812658e+00
3.21811646e-01 -7.57516503e-01 3.82267348e-02 2.65225708e-01
3.64326954e-01 -5.21511972e-01 -1.11840737e+00 -3.37675005e-01
2.50612468e-01 -8.00118983e-01 8.97796094e-01 -2.18229756e-01
1.01235771e+00 -5.57113290e-01 -4.34137374e-01 1.66553944e-01
8.96731555e-01 5.31511068e-01 8.08772370e-02 9.40677822e-02
4.07013565e-01 7.09285915e-01 9.18149650e-01 5.83177507e-01
4.82861757e-01 6.03792548e-01 -1.92974970e-01 1.87786713e-01
1.55189484e-01 -1.21438093e-01 1.26434593e-02 7.24513412e-01
2.47134492e-01 -2.40691334e-01 -1.07069957e+00 3.68790030e-01
-2.16472745e+00 -8.96867275e-01 3.01404417e-01 2.14363503e+00
1.29957700e+00 2.37768844e-01 -1.14918500e-01 4.11980838e-01
8.25858533e-01 -1.19383475e-02 -7.71339655e-01 -2.46257428e-02
2.81849384e-01 4.48522344e-02 -2.56173223e-01 5.09591162e-01
-1.30851233e+00 9.75369692e-01 5.31714535e+00 7.50004530e-01
-8.40262711e-01 2.62651354e-01 9.51807618e-01 -1.23643829e-02
5.07981144e-02 2.37868652e-01 -1.34329021e+00 3.08576494e-01
8.82208347e-01 -6.35985315e-01 1.38977692e-01 1.10235643e+00
-4.79463011e-01 -1.45878717e-01 -1.58400416e+00 7.76387513e-01
4.84396815e-01 -1.10707414e+00 1.84891745e-01 -4.59243119e-01
8.05713713e-01 -2.00544417e-01 -2.98352182e-01 1.06554091e+00
2.95885533e-01 -5.33440650e-01 3.70477468e-01 1.01629056e-01
7.14071810e-01 -6.53309226e-01 1.01738524e+00 8.93917024e-01
-1.09923935e+00 -3.34308326e-01 -5.64401567e-01 -3.50076884e-01
1.95357967e-02 6.88501894e-01 -9.15937960e-01 3.36600035e-01
4.10064310e-01 8.93138707e-01 -6.22115195e-01 7.75220871e-01
-2.89736658e-01 4.41530913e-01 -2.43325666e-01 -3.33138794e-01
3.92351411e-02 3.26730102e-01 1.06296957e-01 1.01629734e+00
4.64494377e-02 4.71767694e-01 4.98052746e-01 1.92543343e-01
-3.45938414e-01 4.47887719e-01 -5.32881916e-01 2.03220114e-01
4.28697824e-01 1.28474104e+00 -9.55549002e-01 -9.11435723e-01
-5.23064673e-01 9.95219946e-01 5.02660871e-01 2.61156589e-01
-3.93013209e-01 -7.22658515e-01 -1.34251386e-01 -1.23456977e-01
1.15266986e-01 4.84885931e-01 -1.82027936e-01 -1.61186635e+00
4.96699698e-02 -5.90116739e-01 8.80639791e-01 -4.99219656e-01
-1.58680332e+00 8.39069605e-01 -4.65035401e-02 -1.19215262e+00
-4.04737383e-01 -3.95322710e-01 -4.98082846e-01 1.51968673e-01
-1.48351455e+00 -9.92316604e-01 -3.90859962e-01 4.56028670e-01
1.21609485e+00 -1.30702347e-01 9.44957495e-01 2.31186092e-01
-4.67861801e-01 6.59415245e-01 2.74170369e-01 2.28868470e-01
9.49871838e-01 -1.18695259e+00 3.67148697e-01 6.60624325e-01
2.76349127e-01 1.91727161e-01 5.25699258e-01 -4.90909696e-01
-8.01545918e-01 -1.28247714e+00 1.31722784e+00 -6.37635767e-01
6.38099432e-01 -5.13253927e-01 -1.11908841e+00 1.02172840e+00
-1.70148343e-01 3.84647638e-01 1.05114460e+00 4.67128187e-01
-3.18323344e-01 -6.21841475e-02 -1.02337670e+00 3.98913443e-01
8.03812504e-01 -2.96354949e-01 -1.02304697e+00 8.37528825e-01
1.00112128e+00 -2.81591892e-01 -4.52719361e-01 4.68873233e-01
3.27291846e-01 -2.79922336e-01 5.51935077e-01 -1.02766788e+00
3.97671461e-01 -2.12955326e-02 -1.01914793e-01 -1.49486268e+00
-2.02402830e-01 -3.12576592e-01 -3.71766895e-01 1.24974871e+00
7.31376946e-01 -6.11134231e-01 7.57345259e-01 7.91237652e-01
2.40023285e-02 -9.60529029e-01 -8.76865625e-01 -8.49839151e-01
1.72548890e-01 -4.60366756e-01 3.87673885e-01 1.29979968e+00
7.04745233e-01 1.22449684e+00 -4.11619633e-01 -3.32291752e-01
2.12099254e-01 4.95686859e-01 7.36496150e-01 -1.53517115e+00
-2.00104654e-01 -7.65415579e-02 -2.17835858e-01 -1.05333245e+00
6.85040234e-03 -1.16567183e+00 3.20992172e-01 -1.41603446e+00
7.94973016e-01 -4.68058407e-01 -1.11379877e-01 8.00910115e-01
-6.76842988e-01 -1.80075660e-01 9.64996219e-02 3.28265101e-01
-1.12357497e+00 6.55678332e-01 8.78528774e-01 -3.69987190e-01
-4.16188031e-01 5.19457519e-01 -8.45561326e-01 7.80253768e-01
6.28402770e-01 -6.97284222e-01 -7.02121437e-01 -3.10552195e-02
3.60777438e-01 -1.08144857e-01 -3.84767622e-01 -6.89741135e-01
4.73133683e-01 -2.77685791e-01 2.33235732e-02 -5.88033020e-01
-3.16610709e-02 -8.50605011e-01 -4.37172592e-01 3.96145791e-01
-8.74116838e-01 -5.64765096e-01 -3.00673068e-01 6.83587313e-01
-1.54842094e-01 -8.05433750e-01 9.09055114e-01 -4.86239083e-02
-8.18130136e-01 4.67047423e-01 -3.96292835e-01 5.94372511e-01
1.38194036e+00 -4.29450534e-03 -3.84880573e-01 -2.94193029e-01
-1.21073234e+00 4.72664267e-01 1.02256276e-01 4.70188618e-01
3.89173090e-01 -1.18993342e+00 -5.77552915e-01 3.74626666e-02
7.38503575e-01 4.77941811e-01 2.60697126e-01 5.92516303e-01
1.32513267e-03 2.33360335e-01 4.34738576e-01 -2.22465068e-01
-1.23482931e+00 1.09373939e+00 1.15034739e-02 -6.51514649e-01
-6.20562911e-01 8.29659522e-01 1.67993620e-01 -7.75723636e-01
5.21617234e-01 -3.72684658e-01 -6.64996088e-01 6.95145249e-01
9.07906115e-01 1.73078239e-01 2.92901486e-01 -2.99975008e-01
3.88962179e-02 1.94778621e-01 -6.62201226e-01 3.30791563e-01
1.34748125e+00 -3.01007718e-01 1.70218453e-01 1.10124063e+00
1.25284767e+00 -4.24865246e-01 -7.52983272e-01 -9.80007052e-01
3.31600755e-01 -2.51762837e-01 -2.12154433e-01 -7.70935774e-01
-6.75680697e-01 9.00375962e-01 3.71545702e-01 4.83706236e-01
8.53828132e-01 2.41782516e-01 9.28253174e-01 1.08741474e+00
2.71787435e-01 -1.22000408e+00 2.91941136e-01 9.72310483e-01
4.69276130e-01 -1.44664657e+00 1.04593918e-01 -4.05293763e-01
-7.94236422e-01 1.23923683e+00 6.57090962e-01 4.70914960e-01
9.67130125e-01 1.84174301e-03 -1.24708906e-01 -1.06396824e-01
-1.03877330e+00 -2.96496123e-01 2.28180870e-01 3.55837882e-01
2.72343695e-01 -1.32998720e-01 -2.68824339e-01 9.90140915e-01
-1.35268390e-01 5.16641187e-03 5.46939790e-01 1.22302687e+00
-9.45985675e-01 -1.02348351e+00 1.31731167e-01 9.07225132e-01
-2.20678806e-01 -1.57018259e-01 -3.39350462e-01 3.92757505e-01
2.08212882e-02 9.24071252e-01 5.38857505e-02 -3.04030746e-01
4.42563295e-01 5.84387243e-01 -1.22125462e-01 -1.41924822e+00
-2.04809725e-01 -3.99653822e-01 6.94051236e-02 1.38808087e-01
-1.34138525e-01 -5.62817454e-01 -1.22041285e+00 3.60570513e-02
-7.34986067e-01 5.80020905e-01 1.58997715e-01 1.39391232e+00
1.45835787e-01 3.67251307e-01 1.01014721e+00 -4.78808433e-01
-7.59548128e-01 -1.25234532e+00 -6.50819063e-01 5.43422580e-01
1.66804492e-01 -4.43037212e-01 -6.07299626e-01 3.43550742e-01] | [10.23849105834961, 3.7335102558135986] |
1d5354c7-d3af-4db1-9a39-5c762b2d27dd | attention-transfer-network-for-nature-image | null | null | https://www.semanticscholar.org/paper/Attention-Transfer-Network-for-Nature-Image-Matting-Zhou-Tian/426480d271a49efd97b4586b918188f8e44d20e4 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9197694&tag=1 | Attention Transfer Network for Nature Image Matting | Natural image matting is an important problem that widely applied in computer vision and graphics. Recent deep learning matting approaches have made an impressive process in both accuracy and efficiency. However, there are still two fundamental problems remain largely unsolved: 1) accurately separating an object from the image with similar foreground and background color or lots of details; 2) exactly extracting an object with fine structures from complex background. In this paper, we propose an attention transfer network (ATNet) to overcome these challenges. Specifically, we firstly design a feature attention block to effectively distinguish the foreground object from the color-similar regions by activating foreground-related features as well as suppressing others. Then, we introduce a scale transfer block to magnify the feature maps without adding extra information. By integrating the above blocks into an attention transfer module, we effectively reduce the artificial content in results and decrease the computational complexity. Besides, we use a perceptual loss to measure the difference between the feature representations of the predictions and the ground-truths. It can further capture the high-frequency details of the image, and consequently, optimize the fine structures of the object. Extensive experiments on two publicly common datasets (i.e., Composition-1k matting dataset, and www.alphamatting.com dataset) show that the proposed ATNet obtains significant improvements over the previous methods. The source code and compiled models have been made publicly available at https://github.com/ailsaim/ATNet. | ['and Zhiquan Qi', 'IEEE', 'Member', 'Yingjie Tian', 'Fenfen Zhou'] | 2020-09-15 | null | null | null | ieee-transactions-on-circuits-and-systems-for-8 | ['image-matting'] | ['computer-vision'] | [ 2.56392986e-01 -2.17808709e-01 1.12221219e-01 -3.09270054e-01
-5.17198980e-01 -1.13805354e-01 2.58041441e-01 -2.75958031e-01
-1.10226154e-01 3.96610469e-01 4.50084358e-02 4.32409309e-02
1.38089225e-01 -8.13684046e-01 -9.61850226e-01 -8.95004988e-01
4.42928553e-01 1.18655227e-01 5.13309717e-01 3.78697701e-02
1.63583711e-01 2.89490461e-01 -1.35042584e+00 3.99409771e-01
1.21289933e+00 1.32073188e+00 5.32823145e-01 1.65795878e-01
-4.18802321e-01 8.20542872e-01 -3.14941794e-01 -3.43851596e-01
3.33422810e-01 -5.00840008e-01 -5.05466700e-01 3.66817892e-01
6.53422177e-01 -4.73168582e-01 -5.64550817e-01 1.44828939e+00
3.03548396e-01 -5.26521504e-02 3.98387343e-01 -1.15838635e+00
-8.57757688e-01 4.54035521e-01 -1.06285250e+00 1.25193849e-01
-1.36525586e-01 4.15769905e-01 6.55697703e-01 -9.84390080e-01
1.44860849e-01 1.45142853e+00 4.22919989e-01 3.00203502e-01
-1.15801835e+00 -8.77927244e-01 3.12158644e-01 4.26299334e-01
-1.37455702e+00 -3.53113830e-01 1.00176597e+00 -4.18761045e-01
2.05980778e-01 2.90107429e-01 7.85477102e-01 8.10732961e-01
2.71625280e-01 1.04276502e+00 1.12452114e+00 -4.70361598e-02
-9.24872980e-02 7.65916482e-02 -4.60913889e-02 7.98411012e-01
3.05339068e-01 -8.79729912e-02 -2.01994926e-01 2.25185171e-01
1.04056191e+00 3.91549706e-01 -6.29164457e-01 -3.62753898e-01
-1.23859286e+00 5.20912647e-01 8.33729208e-01 2.05079988e-01
-4.59875613e-01 2.20335454e-01 1.68064266e-01 2.04330515e-02
5.77708066e-01 -1.08063556e-02 -2.94890940e-01 2.73762286e-01
-7.79485762e-01 7.71028325e-02 2.14932606e-01 1.03416741e+00
1.07581842e+00 2.02160865e-01 -2.39306316e-01 8.59997153e-01
3.19483310e-01 6.38933420e-01 4.04494733e-01 -8.35184693e-01
5.93157351e-01 9.81982112e-01 4.43126298e-02 -1.33076072e+00
-3.67231411e-03 -3.76003087e-01 -1.00995374e+00 3.04925919e-01
2.88117021e-01 -9.97045729e-03 -1.16264737e+00 1.52424085e+00
5.02231419e-01 4.64205712e-01 -3.24100524e-01 1.22862136e+00
9.79441762e-01 9.64869797e-01 -1.60187706e-01 -3.24519947e-02
1.34708273e+00 -1.26446438e+00 -6.79962754e-01 -5.75761735e-01
-5.50629431e-03 -9.56525266e-01 1.23232150e+00 2.27090999e-01
-1.17442882e+00 -8.28094304e-01 -1.04902899e+00 -2.66356826e-01
-5.74986674e-02 3.82650316e-01 6.01510048e-01 2.48491853e-01
-6.61975801e-01 5.50511122e-01 -9.30313289e-01 2.89162099e-02
7.90199459e-01 2.68596113e-01 -1.37457356e-01 -2.53372163e-01
-9.03376281e-01 5.46789110e-01 4.69214529e-01 4.07552361e-01
-9.38839138e-01 -7.14185774e-01 -8.46375942e-01 1.45220026e-01
5.47866762e-01 -6.02877378e-01 9.82230663e-01 -1.57017684e+00
-1.27862895e+00 7.02407062e-01 2.19051875e-02 -5.51664717e-02
5.79422474e-01 -4.94514227e-01 -2.86616266e-01 -2.72676516e-02
4.01997752e-02 6.13387048e-01 9.90547836e-01 -1.28492892e+00
-5.30702591e-01 -3.11120033e-01 -1.27339363e-01 2.86155522e-01
-4.12721962e-01 -1.41641989e-01 -8.49876761e-01 -9.63746548e-01
2.69610524e-01 -6.34954274e-01 -8.02852064e-02 3.89176399e-01
-4.95843351e-01 1.93575293e-01 8.32362175e-01 -9.13241327e-01
9.00582016e-01 -2.22927165e+00 3.44582617e-01 -1.53185964e-01
4.84986275e-01 4.64429528e-01 -2.55872428e-01 -9.07471403e-02
-6.31096661e-02 -1.38333574e-01 -4.40235823e-01 -6.47241548e-02
-3.03083926e-01 4.84838299e-02 -3.47409487e-01 4.48205650e-01
4.57623452e-01 1.04739881e+00 -6.81504607e-01 -5.26708126e-01
3.92126918e-01 4.68276262e-01 -4.26158160e-01 3.69520932e-01
-3.03030431e-01 4.25694644e-01 -5.99597991e-01 6.06177330e-01
1.12359250e+00 -2.54142135e-01 -2.18871310e-01 -5.89888155e-01
-8.84790197e-02 9.26663652e-02 -1.33794785e+00 1.56076634e+00
-1.91564843e-01 4.98631954e-01 2.91282564e-01 -9.81060684e-01
8.81238818e-01 -1.28191218e-01 2.91897207e-01 -6.94087803e-01
4.27502364e-01 9.50743854e-02 1.21507920e-01 -4.09898937e-01
1.54300079e-01 1.56590808e-02 3.40979964e-01 2.47528911e-01
-8.03962275e-02 5.45288473e-02 -1.83337182e-01 4.72456142e-02
6.07040524e-01 7.12207630e-02 -3.15616690e-02 -2.74984300e-01
6.10957503e-01 -1.03178166e-01 9.53958988e-01 3.70115936e-01
-1.53783202e-01 8.90897930e-01 2.80034035e-01 -4.94520485e-01
-8.42722476e-01 -1.06493139e+00 5.70163205e-02 7.71376014e-01
8.50694895e-01 -1.92435861e-01 -8.65699828e-01 -5.43677330e-01
6.01471961e-02 4.17457491e-01 -8.02859366e-01 -3.30394685e-01
-7.51274288e-01 -6.35519147e-01 1.44538194e-01 6.29267335e-01
1.05031681e+00 -1.23797870e+00 -2.83898503e-01 5.56215458e-03
-3.65575314e-01 -1.00413752e+00 -7.99472749e-01 -1.32093742e-01
-7.79639661e-01 -9.21850443e-01 -8.18840623e-01 -9.22130346e-01
8.03606629e-01 6.96055889e-01 1.01712799e+00 4.64461923e-01
-3.86796087e-01 -1.24341495e-01 -1.65181220e-01 -3.70216519e-01
-9.08637233e-03 -1.02568924e-01 -3.77696902e-01 4.06651795e-01
2.26905778e-01 -4.09704536e-01 -7.85364032e-01 3.70629191e-01
-1.01315522e+00 6.81664646e-01 9.49926615e-01 8.74971509e-01
7.64113188e-01 1.37204930e-01 3.00316185e-01 -8.54504287e-01
1.43987611e-01 -3.85363281e-01 -6.63564980e-01 2.04387888e-01
-3.58417094e-01 -2.99500786e-02 5.73067248e-01 -4.67872977e-01
-1.18613696e+00 5.89151941e-02 1.04498994e-02 -7.27451026e-01
-6.92917791e-04 2.25398585e-01 -6.51443660e-01 -1.57665133e-01
8.09159279e-02 5.25429428e-01 1.12383254e-03 -5.97563982e-01
1.29276007e-01 3.75473827e-01 4.58384931e-01 -5.77684402e-01
1.08931863e+00 5.27660251e-01 -2.59763598e-01 -4.77959931e-01
-1.03778386e+00 -2.27423862e-01 -4.72555488e-01 -1.67291448e-01
9.22672987e-01 -9.45824742e-01 -4.52004999e-01 7.79103816e-01
-9.85542536e-01 -4.59243357e-01 -1.76538229e-01 2.68078923e-01
-3.62861484e-01 4.51502711e-01 -6.86710536e-01 -4.69988495e-01
-5.58776855e-01 -1.36072278e+00 1.09647346e+00 5.94770432e-01
4.46753502e-01 -6.50358438e-01 -3.24645907e-01 4.07250911e-01
3.99626881e-01 1.92970335e-01 8.75836074e-01 6.15569502e-02
-1.09442806e+00 1.28887713e-01 -7.43903458e-01 4.49935108e-01
3.37610096e-01 4.01003771e-02 -9.64344561e-01 -2.06167668e-01
1.25441417e-01 -2.80591756e-01 1.15549326e+00 4.50182140e-01
1.55766749e+00 -2.92043775e-01 -2.47250333e-01 1.01318264e+00
1.36572897e+00 1.39319584e-01 7.32245266e-01 2.10885763e-01
1.05057275e+00 4.73316461e-01 7.78287292e-01 2.06057876e-01
8.02073553e-02 5.67929924e-01 5.67299485e-01 -4.70213234e-01
-3.89847010e-01 -3.26982409e-01 3.11763644e-01 8.47984850e-01
4.51592468e-02 3.69275198e-03 -5.95280349e-01 3.25161219e-01
-1.93771791e+00 -7.89529383e-01 -1.80607319e-01 2.02182078e+00
8.38012636e-01 2.45385006e-01 -1.42872721e-01 -1.11383915e-01
9.45127547e-01 2.11509436e-01 -8.87768090e-01 4.96156774e-02
-2.46183738e-01 6.46673292e-02 3.29464167e-01 3.76284838e-01
-1.23575389e+00 1.02675235e+00 4.78318453e+00 1.07933700e+00
-1.26239896e+00 -8.89764130e-02 1.00575483e+00 1.70707610e-02
-2.95083046e-01 -1.32566160e-02 -5.54266870e-01 7.79342175e-01
2.53092408e-01 -2.05515221e-01 5.46476245e-01 8.04525256e-01
8.70809108e-02 8.84517506e-02 -8.45295191e-01 1.03073299e+00
1.20334044e-01 -1.28912723e+00 3.75330061e-01 -1.00363292e-01
6.80486083e-01 -1.73645571e-01 1.64418757e-01 2.74718851e-01
-6.65533617e-02 -1.01074946e+00 9.50845301e-01 5.66485584e-01
7.53710568e-01 -6.86852336e-01 7.40444362e-01 2.54200250e-01
-1.45025504e+00 6.76628873e-02 -6.01027310e-01 1.06218316e-01
-1.39088079e-01 7.73911536e-01 -2.33886763e-01 5.89475274e-01
8.27882588e-01 8.05960953e-01 -6.80145383e-01 1.20477319e+00
-1.79375783e-01 5.47524631e-01 -5.30175082e-02 2.60846794e-01
1.63656995e-01 -5.30515671e-01 3.40676039e-01 1.03065050e+00
1.93885624e-01 8.92153159e-02 3.49127561e-01 1.27640462e+00
-1.89622477e-01 8.05215985e-02 -1.14498697e-01 8.60184282e-02
3.19715530e-01 1.47321403e+00 -7.71058857e-01 -4.04625893e-01
-4.34372187e-01 1.18722761e+00 3.61447424e-01 3.53950709e-01
-1.03457546e+00 -4.71235663e-01 6.71033442e-01 2.74428070e-01
3.67082745e-01 -1.69373676e-02 -4.24062014e-01 -1.33467603e+00
1.99179336e-01 -9.72593069e-01 1.66495815e-01 -9.36085105e-01
-1.22268343e+00 4.96528327e-01 -2.82235920e-01 -1.25649059e+00
5.41607201e-01 -7.00430334e-01 -8.59799206e-01 9.90054190e-01
-1.43634200e+00 -1.17683220e+00 -9.21792448e-01 4.67221797e-01
6.62946701e-01 1.92520291e-01 2.50229329e-01 5.64190090e-01
-8.47276926e-01 4.65959460e-01 6.67182207e-02 2.86302716e-01
7.71783650e-01 -1.20162225e+00 3.72704864e-01 9.35114145e-01
-1.73808202e-01 5.58723927e-01 3.27668905e-01 -6.64942384e-01
-1.40918148e+00 -1.40127051e+00 1.59302875e-01 -1.18658088e-01
4.21065837e-01 -5.43719232e-01 -1.31074023e+00 6.24441087e-01
2.58079022e-01 1.24633811e-01 4.69738208e-02 -3.19392145e-01
-3.09864312e-01 -3.71332228e-01 -8.83151650e-01 6.78027093e-01
9.96613026e-01 -1.45750016e-01 -3.83006066e-01 2.21783876e-01
7.91739464e-01 -4.72668678e-01 -4.48027700e-01 5.80803931e-01
4.07143205e-01 -1.04656017e+00 9.88849163e-01 -2.42119491e-01
6.96030676e-01 -6.41331017e-01 5.82201630e-02 -1.20165551e+00
-6.32303715e-01 -2.94222057e-01 -4.40198742e-02 1.44841230e+00
-4.90367599e-02 -5.67133009e-01 7.51813293e-01 3.20891261e-01
-2.28151590e-01 -1.03036845e+00 -5.97419262e-01 -4.38524753e-01
2.78372895e-02 1.06739402e-02 5.65573990e-01 8.88894677e-01
-7.54754663e-01 3.63710701e-01 -5.47592759e-01 1.45682499e-01
6.87447250e-01 6.74936056e-01 7.91919589e-01 -1.03416049e+00
-2.76314348e-01 -5.31054258e-01 -3.98312479e-01 -1.20487475e+00
-4.21850570e-02 -7.48317361e-01 1.64523199e-01 -1.55556834e+00
6.34387195e-01 -4.36281711e-01 -4.17670906e-01 4.92588967e-01
-6.86993480e-01 2.14777768e-01 8.80788937e-02 2.15381876e-01
-5.58649063e-01 1.00305235e+00 1.62949681e+00 -4.16708529e-01
1.31267175e-01 -1.93155408e-01 -7.79245853e-01 8.92370045e-01
8.67234707e-01 -3.51922840e-01 -2.77667165e-01 -7.35948920e-01
-2.81164318e-01 -2.35314026e-01 5.21535099e-01 -9.84279692e-01
-1.64886303e-02 -4.45181847e-01 9.04219747e-01 -5.97575545e-01
3.63460809e-01 -8.05889189e-01 8.31149369e-02 5.89490175e-01
-1.01303227e-01 -6.28690422e-02 3.39322627e-01 5.72346747e-01
-1.79769233e-01 -1.07713103e-01 1.11772585e+00 -1.03781454e-01
-7.45481908e-01 6.51424289e-01 -9.07108653e-03 7.26154298e-02
1.06172776e+00 -1.39360176e-02 -4.78378713e-01 -2.29474500e-01
-2.83875465e-01 3.24546903e-01 6.74487829e-01 3.98592472e-01
8.36212754e-01 -1.47267187e+00 -7.79939771e-01 4.34144437e-01
-7.31516480e-02 2.70461231e-01 5.19564629e-01 8.19345832e-01
-7.06857204e-01 -9.51342285e-02 -4.70212221e-01 -5.11564374e-01
-1.17995918e+00 6.19500756e-01 4.39599246e-01 7.07921386e-02
-8.59769225e-01 8.96242440e-01 1.01701915e+00 -2.14218214e-01
2.03398541e-01 -4.20256466e-01 3.26425545e-02 -3.23478937e-01
7.30046749e-01 1.89045236e-01 -1.11366168e-01 -6.30910635e-01
-2.71095425e-01 7.10283637e-01 -3.11507523e-01 3.77217263e-01
1.24591088e+00 -1.50742769e-01 -2.23141074e-01 3.25676084e-01
1.04785669e+00 7.89881567e-04 -1.60496020e+00 -3.91027629e-01
-3.00468087e-01 -8.55234742e-01 1.95299506e-01 -5.53800642e-01
-1.65184414e+00 1.14258313e+00 7.64618814e-01 -6.67138025e-02
1.34433353e+00 -1.01152837e-01 1.03861308e+00 -4.88495305e-02
-4.68668453e-02 -7.19887912e-01 3.48390847e-01 1.76413551e-01
1.05344450e+00 -1.29649186e+00 1.22242413e-01 -6.08453810e-01
-5.80415487e-01 9.03046489e-01 1.18658268e+00 -4.12367046e-01
4.83857810e-01 2.24402621e-01 1.79086566e-01 -1.78472638e-01
-5.10030746e-01 -2.05476388e-01 4.33180809e-01 3.31234634e-01
2.58278310e-01 -1.08242035e-03 7.33055826e-03 7.71171629e-01
1.25946417e-01 -2.34367520e-01 2.59435654e-01 7.04912901e-01
-6.82186007e-01 -7.21188366e-01 -4.84634638e-01 5.84602296e-01
-4.12564725e-01 -2.46755078e-01 -4.58480418e-01 4.50628608e-01
3.31068665e-01 6.45985663e-01 8.36064816e-02 -5.25429249e-01
2.72110224e-01 -3.57286274e-01 4.09751385e-01 -5.37734866e-01
-3.77094209e-01 2.59979695e-01 -4.27675068e-01 -4.98361796e-01
-1.57080188e-01 -3.16646993e-01 -1.14032423e+00 -2.98960835e-01
-5.13057351e-01 -2.86745615e-02 2.49304205e-01 6.39772832e-01
3.47430915e-01 8.27872872e-01 5.87822795e-01 -1.03448057e+00
-3.58992666e-01 -8.68858516e-01 -4.66387868e-01 5.16775727e-01
2.31441781e-01 -7.51410782e-01 -2.23090798e-01 1.40486151e-01] | [10.591136932373047, -0.9438309669494629] |
30865d03-4f8c-475a-abb8-7a807add246f | how-helpful-is-inverse-reinforcement-learning | null | null | https://aclanthology.org/2021.acl-short.11 | https://aclanthology.org/2021.acl-short.11.pdf | How Helpful is Inverse Reinforcement Learning for Table-to-Text Generation? | Existing approaches for the Table-to-Text task suffer from issues such as missing information, hallucination and repetition. Many approaches to this problem use Reinforcement Learning (RL), which maximizes a single manually defined reward, such as BLEU. In this work, we instead pose the Table-to-Text task as Inverse Reinforcement Learning (IRL) problem. We explore using multiple interpretable unsupervised reward components that are combined linearly to form a composite reward function. The composite reward function and the description generator are learned jointly. We find that IRL outperforms strong RL baselines marginally. We further study the generalization of learned IRL rewards in scenarios involving domain adaptation. Our experiments reveal significant challenges in using IRL for this task. | ['Shashank Srivastava', 'Snigdha Chaturvedi', 'Zheng Qi', 'Sayan Ghosh'] | 2021-08-01 | null | null | null | acl-2021-5 | ['table-to-text-generation'] | ['natural-language-processing'] | [ 2.95964867e-01 5.09678423e-01 -6.32250607e-01 -4.38314319e-01
-1.56049371e+00 -7.35870302e-01 9.18676257e-01 1.49777606e-01
-4.41013873e-01 1.18898177e+00 7.75376916e-01 -1.60398081e-01
1.50226131e-01 -3.43614548e-01 -8.57977808e-01 -3.23397249e-01
1.23818070e-01 8.39113235e-01 -3.56260180e-01 -4.93778765e-01
4.83680367e-01 -6.16966896e-02 -1.03069007e+00 3.26828301e-01
8.33443105e-01 4.48699951e-01 2.60273695e-01 6.14217103e-01
-1.06135681e-01 1.44470572e+00 -8.06395710e-01 -3.93989265e-01
1.66049182e-01 -7.07879484e-01 -1.09757972e+00 9.19362828e-02
-9.05261841e-03 -8.39868307e-01 -2.15091676e-01 7.42403090e-01
4.90747839e-01 4.21123952e-01 8.81656170e-01 -1.46135449e+00
-1.05634832e+00 1.14987230e+00 -4.94119763e-01 -5.09520173e-02
6.16007566e-01 2.12925494e-01 1.46014154e+00 -7.74315119e-01
7.37384498e-01 1.56319821e+00 1.71722874e-01 7.78402746e-01
-1.39333940e+00 -5.34043372e-01 3.35622975e-03 6.11467361e-02
-1.04718459e+00 -5.14397144e-01 4.19982940e-01 -2.42560610e-01
1.24772680e+00 -1.26387089e-01 -9.81909186e-02 1.40555990e+00
2.14314759e-02 1.17886066e+00 1.09885263e+00 -4.43362892e-01
1.49781227e-01 2.14040712e-01 -2.88528383e-01 3.00403088e-01
2.10248545e-01 1.75928935e-01 -7.23725080e-01 -6.61600754e-02
7.80296147e-01 -2.93937832e-01 1.83557212e-01 -7.66175911e-02
-1.22926295e+00 1.19132626e+00 2.63328999e-01 -2.20910475e-01
-2.22778678e-01 5.41972399e-01 3.24121773e-01 4.28234905e-01
2.90124089e-01 8.78055692e-01 -5.03182232e-01 -3.86534095e-01
-5.12096405e-01 5.90093911e-01 8.64383757e-01 1.29298949e+00
7.53015518e-01 3.87711413e-02 -5.93718767e-01 1.03898799e+00
3.21524173e-01 4.54262972e-01 4.67526793e-01 -1.36654043e+00
7.08583176e-01 1.09350294e-01 4.68646705e-01 -1.18035778e-01
-4.40471619e-01 -1.93873510e-01 -3.68028760e-01 7.93995857e-02
5.11414170e-01 -3.92507613e-01 -7.17904150e-01 2.01695228e+00
-7.38854036e-02 -2.89529562e-01 5.09558320e-01 6.97616100e-01
6.97192490e-01 6.42369926e-01 2.65881866e-01 -1.42590746e-01
1.04283214e+00 -1.28912652e+00 -9.48168516e-01 -6.46897376e-01
8.11180830e-01 -6.46655679e-01 1.28890979e+00 4.07396376e-01
-1.30867505e+00 -1.21725731e-01 -8.44768941e-01 -3.36543143e-01
-8.95225480e-02 2.92798668e-01 4.17467266e-01 1.48660123e-01
-9.47439194e-01 6.50348902e-01 -3.25001508e-01 -1.74691081e-01
1.95632339e-01 3.26716423e-01 -7.30096176e-02 -1.96671382e-01
-1.34035349e+00 1.16075957e+00 5.06604314e-01 -3.45981717e-01
-9.60968435e-01 -9.17481408e-02 -9.12620306e-01 5.52129820e-02
7.87608385e-01 -4.90263343e-01 1.96028578e+00 -8.52462649e-01
-1.62319410e+00 7.56688356e-01 -1.86056998e-02 -4.78805810e-01
3.47227097e-01 -3.98861378e-01 -1.07059315e-01 -1.66641444e-01
4.61326450e-01 8.83461237e-01 7.56710768e-01 -1.36633241e+00
-3.59921485e-01 -6.69157207e-02 1.21984117e-01 6.03899598e-01
8.68419185e-02 -1.03011099e-03 -3.51503104e-01 -8.52635503e-01
-4.36296433e-01 -9.91715729e-01 -2.23823905e-01 -5.96413672e-01
-4.45982277e-01 -7.20867932e-01 1.17596902e-01 -6.28168285e-01
1.22936583e+00 -1.86740303e+00 1.27429813e-01 -1.48338273e-01
6.27276003e-02 -3.24741185e-01 -6.40467763e-01 7.18270838e-01
7.40905926e-02 3.41264009e-01 -4.64015901e-02 -3.55763555e-01
8.28779191e-02 3.78138512e-01 -3.13447028e-01 -8.15132260e-02
6.28708005e-01 1.11331260e+00 -1.31861389e+00 -4.55942065e-01
-3.21695060e-01 -2.86243781e-02 -7.76193798e-01 4.25853819e-01
-8.31790149e-01 4.30960685e-01 -4.69589859e-01 5.19663215e-01
2.30557293e-01 -5.26202202e-01 3.98177773e-01 3.88450623e-01
2.31758893e-01 7.30110645e-01 -9.26913738e-01 1.86345983e+00
-3.96801710e-01 3.02749008e-01 -3.46936971e-01 -6.82831645e-01
1.01181185e+00 2.91959792e-01 3.73590887e-01 -8.25077057e-01
-5.11240512e-02 1.82747662e-01 3.48854847e-02 -3.05032134e-01
9.03897047e-01 -2.66272962e-01 -4.61243391e-01 1.00864875e+00
3.71029198e-01 -2.48037711e-01 2.46115088e-01 6.92167521e-01
1.16586673e+00 5.47778964e-01 5.93971908e-01 4.18460071e-02
-4.49891128e-02 1.26114324e-01 3.60732943e-01 1.12372112e+00
-1.16918445e-01 4.33693647e-01 8.75567794e-01 -2.19185036e-02
-1.19281435e+00 -1.07063925e+00 1.45225137e-01 1.42092812e+00
-5.68414815e-02 -6.04497075e-01 -5.33471584e-01 -1.03629386e+00
3.49058472e-02 1.29865324e+00 -4.81866121e-01 -1.92141265e-01
-3.42448950e-01 -5.06054997e-01 4.60337013e-01 5.66571057e-01
9.84493643e-02 -1.50910532e+00 -3.32068145e-01 4.51100826e-01
-5.41319549e-01 -1.19578063e+00 -6.91390574e-01 5.76581240e-01
-7.95026839e-01 -5.98998308e-01 -5.54342806e-01 -4.24651206e-01
4.04827207e-01 1.82549655e-01 1.63719666e+00 -6.08499311e-02
7.95880258e-02 4.49832022e-01 -5.69275439e-01 -4.71287638e-01
-8.80995989e-01 4.61198855e-04 -6.42670617e-02 -5.52672803e-01
3.05630624e-01 -2.39141658e-01 -4.10774142e-01 1.48117945e-01
-8.01840723e-01 -6.32230612e-03 9.37750340e-01 1.11516595e+00
4.81458277e-01 -4.86879349e-01 1.02713585e+00 -1.06129491e+00
1.24081254e+00 -7.36799479e-01 -3.17846656e-01 3.19897801e-01
-9.12215650e-01 8.24570596e-01 6.58932209e-01 -4.43174720e-01
-1.09957552e+00 2.11213022e-01 8.57517496e-02 -2.28208266e-02
2.35034823e-02 4.04186577e-01 8.57084841e-02 5.64941704e-01
9.34839964e-01 2.22792953e-01 -7.90899172e-02 -2.22029552e-01
8.60962451e-01 7.69795537e-01 4.29645270e-01 -7.61652112e-01
6.92884922e-01 -1.40635222e-01 -3.46389890e-01 -2.45499849e-01
-1.36040533e+00 -3.12601566e-01 -3.25346470e-01 -1.45318076e-01
6.93530083e-01 -1.25596666e+00 -8.04697812e-01 -3.63493592e-01
-1.23383820e+00 -7.38423467e-01 -5.84193230e-01 4.26889032e-01
-1.09250462e+00 2.38150507e-01 -6.92253888e-01 -9.86633003e-01
-2.56279737e-01 -9.63363945e-01 1.18149483e+00 1.70854837e-01
-4.97412920e-01 -7.82539964e-01 1.43004254e-01 4.17114854e-01
2.60491878e-01 -5.83129153e-02 8.04252505e-01 -8.81026089e-01
-6.41591847e-01 2.15796664e-01 -1.90652788e-01 1.58442393e-01
-1.20834000e-01 -4.63660628e-01 -7.73033023e-01 -2.02105418e-01
-2.12062776e-01 -1.34784782e+00 8.85693371e-01 1.83489382e-01
1.01000166e+00 -7.12087214e-01 2.06472754e-01 1.74838215e-01
1.29300201e+00 2.61383116e-01 6.38821244e-01 4.35974389e-01
4.06834811e-01 6.44599020e-01 9.09027517e-01 8.60467911e-01
7.25264549e-01 5.59380889e-01 2.82459497e-01 5.80952764e-02
2.42922381e-02 -7.57180691e-01 7.09717631e-01 4.79383707e-01
9.50659886e-02 -5.21548748e-01 -6.92561984e-01 2.89344937e-01
-2.02594948e+00 -9.85818267e-01 2.39825219e-01 2.03505611e+00
1.22857547e+00 2.40441129e-01 2.15817645e-01 -3.54784399e-01
4.24274445e-01 5.08017056e-02 -7.84013271e-01 -6.74048781e-01
-4.17131111e-02 -1.89719293e-02 6.17448270e-01 5.91802359e-01
-7.05470681e-01 1.31475770e+00 7.57243824e+00 6.89312398e-01
-4.48963284e-01 -4.52328660e-02 6.39543414e-01 -5.68700396e-02
-5.94018877e-01 2.49780402e-01 -8.19271982e-01 5.65010570e-02
1.10869741e+00 -3.39748472e-01 1.15516686e+00 8.17852139e-01
2.04138592e-01 -1.50354505e-01 -1.38394284e+00 8.26716721e-01
1.12548441e-01 -8.62830162e-01 2.08384201e-01 2.54523419e-02
7.80604482e-01 1.51594311e-01 2.62087700e-03 8.73594940e-01
1.27833033e+00 -1.30215943e+00 5.70366204e-01 3.04033190e-01
1.04432380e+00 -7.22908616e-01 2.18881100e-01 4.28275198e-01
-6.18063748e-01 -2.05288455e-01 -2.84625590e-01 -6.60817325e-02
-2.69059576e-02 -9.98687092e-03 -1.40097213e+00 1.74473122e-01
2.24921346e-01 8.42712820e-01 -4.84876215e-01 4.10663694e-01
-6.82537973e-01 4.57581550e-01 6.28579855e-02 -2.52452672e-01
2.79216141e-01 -1.29120603e-01 3.34284723e-01 1.09060752e+00
1.22742921e-01 -3.06665264e-02 3.25819165e-01 1.26940560e+00
-6.10940158e-01 1.17840864e-01 -8.37008476e-01 -3.37874085e-01
4.05889601e-01 1.12382698e+00 -4.02588338e-01 -3.66977096e-01
-4.11228865e-01 1.14519203e+00 6.40862703e-01 4.60341156e-01
-5.54345131e-01 -7.81548545e-02 3.86832148e-01 -2.48802543e-01
6.64091855e-02 -2.04490826e-01 -3.41220021e-01 -1.15960217e+00
-1.15384810e-01 -1.13794589e+00 4.01913762e-01 -1.02633667e+00
-1.38334727e+00 3.22844952e-01 5.38306348e-02 -1.29239917e+00
-1.02534068e+00 -2.37055048e-01 -2.67920047e-01 7.01489687e-01
-1.74489045e+00 -7.85348415e-01 1.62131116e-01 5.40415943e-01
9.07295704e-01 -2.24956483e-01 9.00040209e-01 -3.06189775e-01
-3.73838097e-01 6.30623639e-01 3.13471675e-01 5.74772581e-02
1.14683819e+00 -1.68797016e+00 5.97909272e-01 4.37974691e-01
2.17699975e-01 3.76433313e-01 6.86990857e-01 -8.21771324e-01
-1.31337786e+00 -8.92885983e-01 1.03595519e+00 -7.70186663e-01
6.50646448e-01 -3.30386579e-01 -5.88645279e-01 8.80512357e-01
3.93051356e-01 -3.67529899e-01 7.39795625e-01 2.31391281e-01
-4.80149239e-01 2.68315375e-01 -9.55001533e-01 6.33652270e-01
7.24538028e-01 -5.30830383e-01 -5.77656448e-01 5.89834213e-01
9.82797503e-01 -3.53470981e-01 -9.06275094e-01 -1.62933797e-01
3.24042588e-01 -5.88699162e-01 8.41018319e-01 -8.71926129e-01
1.22735035e+00 9.11538024e-03 -1.77811012e-02 -1.57082057e+00
-3.03860843e-01 -7.32130706e-01 -1.80940434e-01 1.05846810e+00
6.78618014e-01 -1.26448497e-01 5.55117309e-01 8.72154593e-01
1.20703623e-01 -3.70712548e-01 -5.70731103e-01 -1.00890315e+00
2.66951233e-01 -1.48897126e-01 3.21187377e-01 8.39885712e-01
3.70543987e-01 1.07433319e+00 -9.22934175e-01 -4.76075351e-01
6.20744050e-01 -2.35273302e-01 7.91756272e-01 -1.04495525e+00
-7.00971723e-01 -1.75521597e-01 5.31184077e-01 -1.21100664e+00
1.96049169e-01 -1.17179298e+00 4.81538981e-01 -1.66510129e+00
4.64318424e-01 -2.56245136e-01 -2.49717668e-01 6.23616934e-01
-2.89051443e-01 -2.94577360e-01 5.28866708e-01 3.69982481e-01
-1.12396824e+00 6.29057229e-01 1.51382899e+00 -3.45184766e-02
-3.78974795e-01 -1.80613458e-01 -1.17672122e+00 2.92339683e-01
9.91035819e-01 -8.85195971e-01 -4.18181241e-01 -4.43656772e-01
4.45550680e-01 8.22031319e-01 -4.77366745e-02 -3.72136384e-01
5.78971626e-03 -4.56889868e-01 3.24380875e-01 -4.69025970e-01
2.73375988e-01 -3.60355616e-01 -5.20342648e-01 1.55912459e-01
-1.11022294e+00 1.99813485e-01 -2.59392988e-02 7.57397115e-01
-1.19624078e-01 -4.66785252e-01 5.05696774e-01 -3.46433640e-01
-3.32264423e-01 -5.01686856e-02 -5.40756226e-01 6.26495183e-01
7.34151840e-01 4.41880107e-01 -4.17045116e-01 -1.11930466e+00
-6.07904613e-01 5.34074545e-01 1.92362487e-01 6.32221580e-01
5.85035682e-01 -1.37011242e+00 -1.06936586e+00 -2.89465398e-01
3.52460504e-01 -4.66262475e-02 -2.87991673e-01 3.22799683e-01
1.53506652e-01 4.59476709e-01 -2.11293921e-01 -8.41432437e-02
-6.47175252e-01 5.98989367e-01 6.97335741e-03 -8.84888947e-01
-3.74969572e-01 4.84241098e-01 6.06937408e-02 -5.41430652e-01
3.68791938e-01 -2.06922926e-02 -2.90869594e-01 -1.25528663e-01
4.40320551e-01 9.37814787e-02 -3.81679446e-01 -1.76718712e-01
-3.50784138e-02 -5.55081926e-02 -3.86742115e-01 -6.36451244e-01
1.34674776e+00 -2.69903779e-01 3.51015627e-01 3.84258509e-01
1.01021421e+00 -1.24830335e-01 -1.42645872e+00 -4.81734931e-01
4.08221066e-01 -1.88587323e-01 -1.53327480e-01 -1.35130215e+00
-3.98709089e-01 6.32534504e-01 7.31712803e-02 -4.67353277e-02
8.94568920e-01 1.57622844e-01 6.14985406e-01 8.27624798e-01
1.30404711e-01 -1.61951530e+00 6.42224371e-01 7.48242319e-01
1.06201971e+00 -1.60767853e+00 7.38210753e-02 2.25696847e-01
-1.50777590e+00 1.19523096e+00 7.71918476e-01 -9.05515254e-02
-1.84921045e-02 3.39222610e-01 -1.16013132e-01 9.51904356e-02
-1.29061782e+00 -4.50266421e-01 1.19328313e-01 5.70968509e-01
8.27147543e-01 4.69692647e-02 -4.24280286e-01 6.34832561e-01
-1.51599944e-01 8.91342387e-02 9.69773889e-01 9.08912897e-01
-3.78428131e-01 -1.56219792e+00 -2.93268174e-01 5.63683212e-01
-4.61020112e-01 -3.36810410e-01 -7.30715454e-01 6.43751502e-01
-5.27893960e-01 1.09412146e+00 -1.50327414e-01 -3.72288346e-01
1.70108140e-01 1.26709595e-01 4.54691410e-01 -7.78484106e-01
-7.08773136e-01 3.01119059e-01 2.56634414e-01 -5.45493484e-01
-1.04767896e-01 -7.01575100e-01 -1.47481298e+00 2.54068542e-02
-1.50221661e-02 3.80172208e-02 6.79233551e-01 8.34382832e-01
1.40527666e-01 4.47342336e-01 7.62295544e-01 -5.27634263e-01
-1.24737072e+00 -9.96984780e-01 -5.91375709e-01 6.61495388e-01
4.02049571e-01 -4.11801577e-01 -2.31656283e-01 -1.71335302e-02] | [11.702709197998047, 8.910517692565918] |
8b5e6563-0027-4c77-a843-37f580814cf2 | robust-proxy-improving-adversarial-robustness | 2306.15457 | null | https://arxiv.org/abs/2306.15457v1 | https://arxiv.org/pdf/2306.15457v1.pdf | Robust Proxy: Improving Adversarial Robustness by Robust Proxy Learning | Recently, it has been widely known that deep neural networks are highly vulnerable and easily broken by adversarial attacks. To mitigate the adversarial vulnerability, many defense algorithms have been proposed. Recently, to improve adversarial robustness, many works try to enhance feature representation by imposing more direct supervision on the discriminative feature. However, existing approaches lack an understanding of learning adversarially robust feature representation. In this paper, we propose a novel training framework called Robust Proxy Learning. In the proposed method, the model explicitly learns robust feature representations with robust proxies. To this end, firstly, we demonstrate that we can generate class-representative robust features by adding class-wise robust perturbations. Then, we use the class representative features as robust proxies. With the class-wise robust features, the model explicitly learns adversarially robust features through the proposed robust proxy learning framework. Through extensive experiments, we verify that we can manually generate robust features, and our proposed learning framework could increase the robustness of the DNNs. | ['Yong Man Ro', 'Hong Joo Lee'] | 2023-06-27 | null | null | null | null | ['adversarial-robustness'] | ['adversarial'] | [ 1.22695580e-01 -7.24360645e-02 -1.24484068e-03 -3.23445082e-01
-9.54980791e-01 -1.06401920e+00 6.95421159e-01 -2.91485548e-01
-2.42557108e-01 7.68621981e-01 1.81411892e-01 -9.21723098e-02
-2.24370256e-01 -9.41564500e-01 -1.09288394e+00 -9.21375871e-01
6.14189729e-02 -2.96160996e-01 1.98276713e-01 -1.80533707e-01
1.88030541e-01 8.26515436e-01 -1.14948368e+00 -4.54150662e-02
9.44918156e-01 9.45380092e-01 -3.96092683e-01 3.61383349e-01
5.20218372e-01 6.79161012e-01 -8.64133060e-01 -3.48103940e-01
6.38021708e-01 -1.44715175e-01 -6.24983311e-01 -1.33312494e-01
3.93986464e-01 -7.10101485e-01 -9.57487285e-01 1.41795087e+00
5.27816534e-01 1.77431867e-01 4.81468886e-01 -1.68260801e+00
-8.63401055e-01 7.63081729e-01 -1.84964061e-01 4.11337465e-02
1.72635496e-01 4.02510464e-01 6.64827108e-01 -5.97033978e-01
3.17775249e-01 1.39272296e+00 3.57571632e-01 8.65187228e-01
-8.07562590e-01 -1.07500160e+00 4.73204643e-01 1.77005157e-01
-1.18646932e+00 -2.48159751e-01 1.18327379e+00 -1.86188817e-01
2.58954406e-01 2.65103221e-01 7.97078833e-02 1.41049504e+00
1.34229949e-02 7.85353422e-01 1.09452915e+00 -1.16617329e-01
1.60771787e-01 8.01302418e-02 -1.09300114e-01 5.75993538e-01
3.05340946e-01 5.55392444e-01 3.65945622e-02 -2.90777177e-01
6.71131909e-01 4.03122991e-01 -5.25659323e-01 -2.14588389e-01
-9.64699686e-01 9.24373090e-01 8.59429419e-01 1.89708903e-01
2.64638657e-05 3.21422815e-01 5.35154581e-01 3.53103399e-01
1.55649453e-01 3.82992148e-01 -4.59690005e-01 3.00001591e-01
-3.98721457e-01 2.69735336e-01 5.16896188e-01 9.53640759e-01
6.75906956e-01 4.52752799e-01 -1.25994772e-01 6.46390080e-01
4.55764383e-01 6.43088818e-01 5.01405537e-01 -8.74947786e-01
5.71196139e-01 4.23243672e-01 -1.43247724e-01 -1.17788863e+00
-1.11770324e-01 -2.95605719e-01 -9.54249442e-01 3.67214471e-01
1.87845275e-01 -4.70606506e-01 -9.07356739e-01 2.12047243e+00
2.81720072e-01 5.22598803e-01 4.83347178e-01 6.22739196e-01
5.31875312e-01 6.04410589e-01 -1.66411668e-01 -6.65586293e-02
7.72894681e-01 -8.77040327e-01 -4.88509983e-01 1.78994928e-02
1.63737372e-01 -6.14668131e-01 7.86079407e-01 2.67052203e-01
-6.90597713e-01 -4.77726281e-01 -1.33297682e+00 3.11301798e-01
-3.33983868e-01 -3.59341651e-01 4.76125717e-01 9.38068211e-01
-5.68926990e-01 8.23142409e-01 -7.76456594e-01 4.47033793e-02
5.93324959e-01 4.43545997e-01 -6.39349818e-01 -1.79335311e-01
-1.44744658e+00 7.22731411e-01 4.69471484e-01 -4.43456182e-03
-1.48573530e+00 -5.34170926e-01 -8.89268279e-01 -7.24252984e-02
2.57717311e-01 -4.22567904e-01 1.07505512e+00 -8.82409036e-01
-1.54951441e+00 3.20894390e-01 5.43951929e-01 -4.39050674e-01
6.63077235e-01 -3.96453768e-01 -4.72712338e-01 3.13063264e-01
-3.15692395e-01 1.45456076e-01 1.24240243e+00 -1.54243374e+00
-2.02008337e-01 -2.40264907e-01 4.83916879e-01 -2.45936990e-01
-8.53683770e-01 2.87383795e-01 1.47993788e-01 -1.15088868e+00
-1.86506003e-01 -9.59943235e-01 -3.70378196e-01 1.89713091e-01
-5.64163566e-01 2.14869782e-01 1.02683794e+00 -3.68099660e-01
8.58619452e-01 -2.33111358e+00 1.55814085e-02 4.48551178e-01
1.64456248e-01 6.54857874e-01 -3.47258449e-01 3.07157695e-01
-2.76350468e-01 5.47497869e-01 -4.66939390e-01 1.45918638e-01
1.82600677e-01 4.62079555e-01 -8.91347826e-01 6.45198882e-01
4.06636268e-01 5.63305378e-01 -8.93388689e-01 -1.66724682e-01
8.46009478e-02 5.99577725e-01 -7.38708138e-01 4.48487729e-01
7.70307556e-02 4.10595387e-01 -8.41949940e-01 7.32904196e-01
1.05539131e+00 4.74942356e-01 -2.26474226e-01 -1.92905858e-01
3.02170873e-01 -2.42119536e-01 -1.13540363e+00 1.20384502e+00
-2.99482316e-01 2.89463878e-01 -3.68217342e-02 -1.10021746e+00
1.01955938e+00 3.53872538e-01 1.70680612e-01 -1.63019955e-01
3.32769632e-01 2.49902725e-01 -4.16403040e-02 -2.34380260e-01
9.96759310e-02 1.73894186e-02 -2.29437634e-01 3.26090544e-01
8.32482427e-02 -8.74761567e-02 -4.17886615e-01 9.60615799e-02
1.25725794e+00 6.81861653e-04 2.03919858e-02 1.40750960e-01
8.46323788e-01 -5.63955247e-01 9.35805917e-01 5.91371834e-01
-3.57455283e-01 5.93153715e-01 3.95456821e-01 -3.07169825e-01
-8.41865718e-01 -1.26083577e+00 -1.03597902e-01 7.79174805e-01
6.11003302e-02 -1.87060177e-01 -9.16381359e-01 -1.18664455e+00
1.09786779e-01 3.12748671e-01 -5.33839822e-01 -9.16062117e-01
-6.30075693e-01 -5.47827601e-01 1.12201548e+00 7.94950128e-01
7.10259378e-01 -9.74393129e-01 4.35958430e-02 -1.46180233e-02
2.25066379e-01 -9.75079954e-01 -3.97865832e-01 1.52478755e-01
-7.84222722e-01 -9.64601874e-01 -6.06922925e-01 -6.83568120e-01
9.95171964e-01 2.74398208e-01 4.00660694e-01 3.78081590e-01
-1.46905199e-01 2.08021015e-01 -5.38383663e-01 -2.29399964e-01
-4.84570473e-01 -1.90930352e-01 7.09665656e-01 4.13259082e-02
-2.86561459e-01 -9.04612303e-01 -3.71119142e-01 3.08387250e-01
-1.35074282e+00 -6.89748466e-01 6.85383022e-01 8.38735223e-01
4.62782502e-01 1.88310325e-01 8.16471398e-01 -7.72813737e-01
5.60922444e-01 -5.28866053e-01 -4.75343615e-01 1.32105172e-01
-2.96443075e-01 2.63619512e-01 1.36254895e+00 -6.74615264e-01
-7.52076089e-01 -2.09588581e-03 -4.03307140e-01 -9.25676048e-01
-2.96881557e-01 2.97674239e-01 -9.30745244e-01 -7.17615128e-01
7.86643088e-01 1.39436603e-01 -2.61556506e-01 -3.68860781e-01
6.18745804e-01 5.76130569e-01 6.20404005e-01 -1.03448415e+00
1.83362973e+00 2.99313486e-01 -1.18298056e-02 -2.18628839e-01
-8.40901434e-01 2.26419792e-01 -4.98014599e-01 -1.87912229e-02
4.35153246e-01 -7.37950921e-01 -4.76016045e-01 8.11091840e-01
-9.17355299e-01 -1.77214265e-01 -8.03681761e-02 4.59699184e-01
-7.03806460e-01 4.86522645e-01 -4.71137762e-01 -4.99053001e-01
-3.45668644e-01 -1.23323941e+00 4.99690861e-01 4.02807921e-01
3.38880986e-01 -8.67306828e-01 -2.57104635e-02 4.06432338e-02
3.78854036e-01 7.86419928e-01 6.70979738e-01 -9.82117176e-01
-5.87044537e-01 -4.78031397e-01 -1.27719000e-01 8.91764164e-01
4.06653374e-01 3.55789989e-01 -1.14773452e+00 -5.12084186e-01
6.86545596e-02 -5.50104499e-01 8.44016731e-01 -2.60040402e-01
1.55200195e+00 -7.06271648e-01 -1.12478346e-01 1.10535502e+00
1.25249863e+00 6.06114492e-02 6.50749683e-01 4.73315805e-01
8.27006102e-01 2.29785115e-01 5.64323366e-01 4.30574954e-01
-6.45107999e-02 3.10315847e-01 9.28678036e-01 1.36169180e-01
2.81914830e-01 -3.42122436e-01 7.01550066e-01 5.86959660e-01
-2.03238726e-02 -7.30361417e-02 -5.32390118e-01 3.38544279e-01
-1.67147219e+00 -1.08830929e+00 5.05084455e-01 1.88147509e+00
9.99017715e-01 3.93049628e-01 -8.40386599e-02 4.31021512e-01
7.80735195e-01 3.59135449e-01 -7.40244508e-01 -3.26253504e-01
-8.31454098e-02 7.29135647e-02 4.44203615e-01 2.50964850e-01
-1.36278856e+00 9.68294978e-01 5.76133871e+00 6.97713912e-01
-1.13726270e+00 -2.05187410e-01 2.37364963e-01 -2.01668888e-02
-4.44455504e-01 -5.34976497e-02 -6.17695272e-01 5.37880242e-01
6.94530487e-01 -5.50533652e-01 4.16872621e-01 1.11574197e+00
-9.85067785e-02 9.14283276e-01 -1.01611042e+00 6.02306664e-01
2.11396784e-01 -1.06432939e+00 3.37877452e-01 -5.23723848e-02
8.17904651e-01 -3.97623748e-01 4.59018946e-01 4.27587748e-01
7.50851870e-01 -9.19127643e-01 5.69408417e-01 6.42060995e-01
5.22921205e-01 -1.30473173e+00 6.77272737e-01 3.11678320e-01
-9.60528731e-01 -3.50100100e-01 -6.18817210e-01 2.85861433e-01
-1.22675255e-01 4.20489877e-01 -2.90080369e-01 7.36372232e-01
5.15646875e-01 6.91720665e-01 -7.15107501e-01 1.08059323e+00
-7.62734234e-01 6.54789448e-01 -4.96731885e-02 3.89024466e-01
2.39349440e-01 1.72175318e-01 5.32615185e-01 8.10796976e-01
3.12210053e-01 8.31343327e-03 3.23397368e-01 6.41479254e-01
-5.30060709e-01 -1.88887507e-01 -8.99498999e-01 -6.22135736e-02
6.68341815e-01 1.30858338e+00 -3.16600353e-01 5.60626760e-02
-1.57347232e-01 9.28690255e-01 3.63037497e-01 3.11994940e-01
-1.11861265e+00 -6.79680645e-01 1.00797713e+00 -4.35444564e-01
9.27930921e-02 -1.90059051e-01 8.11761543e-02 -1.19744229e+00
7.65158832e-02 -1.16280615e+00 2.51448393e-01 -4.90689367e-01
-1.73016965e+00 6.69692636e-01 -2.15055317e-01 -1.42335510e+00
-1.32063210e-01 -6.23355567e-01 -1.00758517e+00 6.62672877e-01
-1.58503079e+00 -1.34908175e+00 -7.12571442e-02 1.12679899e+00
2.96482686e-02 -5.34684062e-01 8.37081671e-01 2.62594402e-01
-8.13612163e-01 1.27858925e+00 2.61210263e-01 5.81907868e-01
9.34607387e-01 -1.01894271e+00 3.61340553e-01 1.20712399e+00
2.47639939e-02 9.95048881e-01 6.39425337e-01 -4.41924959e-01
-1.32280397e+00 -1.54841113e+00 -2.73137912e-02 -2.83104807e-01
9.00094450e-01 -2.64740020e-01 -1.08029032e+00 8.18020940e-01
-4.38713841e-02 6.94614708e-01 8.07020664e-01 -5.10030389e-01
-9.41748440e-01 -3.84565502e-01 -1.62797785e+00 6.42265141e-01
9.51696038e-01 -7.53234863e-01 -7.45720685e-01 4.34531391e-01
1.24220455e+00 -2.39579409e-01 -9.54502881e-01 5.08736014e-01
3.73354942e-01 -4.86187220e-01 1.17772615e+00 -8.08385491e-01
3.37681800e-01 -4.80112761e-01 -3.76217276e-01 -1.52246428e+00
-3.05258691e-01 -6.82701290e-01 -1.96704134e-01 1.64084470e+00
3.59927453e-02 -8.00018907e-01 4.68400002e-01 4.09850478e-01
-3.25008005e-01 -5.30568719e-01 -8.84546101e-01 -1.19674635e+00
5.25261521e-01 -2.25154221e-01 9.17382002e-01 1.03645468e+00
-2.90470958e-01 -5.58012366e-01 -5.09820223e-01 7.49198914e-01
5.11163533e-01 -1.11190259e-01 7.96297610e-01 -1.06869733e+00
-2.25646585e-01 -2.50932395e-01 -7.32497394e-01 -4.97973531e-01
6.15401745e-01 -8.87582839e-01 9.31306630e-02 -1.03760958e+00
6.65287971e-02 -3.73636246e-01 -7.04269409e-01 8.55049431e-01
-5.25112748e-01 1.69490933e-01 2.11480752e-01 -1.08077072e-01
-4.03550595e-01 6.65391564e-01 1.11581302e+00 -2.29151741e-01
2.30131641e-01 -1.52649696e-03 -1.06732631e+00 8.86100173e-01
1.16224670e+00 -6.88731015e-01 -3.82657021e-01 -3.97758514e-01
-9.67343077e-02 -4.49974358e-01 5.23748159e-01 -1.07933080e+00
3.36809270e-02 -4.89822030e-01 3.73161465e-01 -1.66988984e-01
6.21138848e-02 -1.05539179e+00 -2.86065668e-01 4.43623513e-01
-2.32155576e-01 -1.07033633e-01 1.63257718e-01 6.43404484e-01
-2.20713362e-01 -3.46514106e-01 1.01248991e+00 1.07461013e-01
-3.85569930e-01 7.43984818e-01 -6.64592758e-02 1.25220165e-01
1.30074525e+00 2.22055733e-01 -5.44002771e-01 -1.34753257e-01
-4.96051252e-01 2.15594471e-01 4.95644987e-01 5.25141954e-01
8.99086833e-01 -1.71232665e+00 -5.43797433e-01 3.61387700e-01
9.97314379e-02 -1.00496218e-01 1.92059726e-01 8.42759535e-02
-2.45224208e-01 -1.68374613e-01 -4.92462277e-01 -2.56106276e-02
-1.09735286e+00 1.20470989e+00 2.98075795e-01 3.01691629e-02
-4.01567847e-01 8.35146487e-01 -4.69630621e-02 -6.27668262e-01
3.30616355e-01 -4.64448072e-02 -1.37034029e-01 -1.39157653e-01
6.78863883e-01 1.93109527e-01 -2.09568366e-01 -6.28814518e-01
-3.53083968e-01 6.67447865e-01 -4.26188052e-01 6.28329515e-02
1.39296329e+00 1.65490612e-01 3.42229381e-02 -1.23773187e-01
1.34208441e+00 2.64863253e-01 -1.53229129e+00 -2.37759739e-01
-2.04202428e-01 -6.66976154e-01 -2.32551962e-01 -4.55320179e-01
-1.52159703e+00 9.49637532e-01 4.82001841e-01 5.59465401e-02
1.26342380e+00 -3.67635876e-01 8.49855781e-01 6.17315650e-01
2.96201766e-01 -7.35561848e-01 2.94457167e-01 5.38833559e-01
9.87040401e-01 -1.04319739e+00 -8.79014954e-02 -2.97396064e-01
-3.88991535e-01 1.04474914e+00 8.76759768e-01 -6.58663690e-01
6.64790273e-01 2.71690816e-01 1.33163571e-01 3.68061662e-01
-5.12632251e-01 1.34395897e-01 8.13638121e-02 9.18656766e-01
-1.54888272e-01 -1.67083889e-01 -6.28686026e-02 9.58239734e-01
-4.01352316e-01 -4.37424392e-01 5.71975946e-01 9.18745220e-01
-4.20370489e-01 -1.40132177e+00 -5.35167754e-01 7.23000541e-02
-6.28496170e-01 7.39722401e-02 -4.40032542e-01 5.37186980e-01
1.28039688e-01 9.83872533e-01 -5.65312564e-01 -7.86178648e-01
4.20151472e-01 -1.96154594e-01 3.15296680e-01 -4.50738519e-01
-5.52047908e-01 -3.91930312e-01 -4.01964307e-01 -5.13732851e-01
-2.23884329e-01 -4.43044901e-01 -1.33139908e+00 -3.32980037e-01
-3.32663506e-01 1.03757784e-01 3.55636686e-01 9.50068414e-01
-4.16377261e-02 6.39644802e-01 1.38361883e+00 -7.56638765e-01
-1.17037165e+00 -6.39503062e-01 -5.15906572e-01 4.85751748e-01
5.56542039e-01 -7.71732628e-01 -7.53024042e-01 -1.29505411e-01] | [5.564391136169434, 7.924444198608398] |
55b5c39f-991e-4838-99e9-5c040b21cb14 | sample-efficient-social-navigation-using | 2106.10318 | null | https://arxiv.org/abs/2106.10318v1 | https://arxiv.org/pdf/2106.10318v1.pdf | Sample Efficient Social Navigation Using Inverse Reinforcement Learning | In this paper, we present an algorithm to efficiently learn socially-compliant navigation policies from observations of human trajectories. As mobile robots come to inhabit and traffic social spaces, they must account for social cues and behave in a socially compliant manner. We focus on learning such cues from examples. We describe an inverse reinforcement learning based algorithm which learns from human trajectory observations without knowing their specific actions. We increase the sample-efficiency of our approach over alternative methods by leveraging the notion of a replay buffer (found in many off-policy reinforcement learning methods) to eliminate the additional sample complexity associated with inverse reinforcement learning. We evaluate our method by training agents using publicly available pedestrian motion data sets and compare it to related methods. We show that our approach yields better performance while also decreasing training time and sample complexity. | ['Gregory Dudek', 'Bobak H. Baghi'] | 2021-06-18 | null | null | null | null | ['social-navigation'] | ['robots'] | [-5.23906909e-02 3.12380403e-01 -2.59592682e-01 -2.92817533e-01
-5.96608043e-01 -5.17436206e-01 8.85986507e-01 -1.00133223e-02
-1.15336943e+00 1.25643277e+00 1.76719695e-01 -4.22566533e-01
-9.48418751e-02 -8.95873368e-01 -9.42356825e-01 -4.87493277e-01
-3.65782470e-01 6.89107776e-01 7.13169098e-01 -4.23981845e-01
4.04113203e-01 4.28498745e-01 -1.59517884e+00 -5.46028651e-02
7.11614549e-01 2.10377946e-01 2.96182990e-01 1.18180430e+00
4.16871905e-01 9.25858200e-01 -1.77277803e-01 -6.68046251e-02
3.65586877e-01 -5.32244325e-01 -8.34459364e-01 3.35155785e-01
-8.62233862e-02 -6.81976020e-01 -4.96074706e-01 7.11310983e-01
3.85888755e-01 5.83235562e-01 8.91123116e-01 -1.46494150e+00
-1.44436315e-01 5.04365385e-01 -1.12429082e-01 1.53654411e-01
4.63278830e-01 5.91317475e-01 7.45678723e-01 -2.42564514e-01
8.72687101e-01 1.35203183e+00 6.77033722e-01 7.14483321e-01
-1.19649732e+00 -3.22791159e-01 3.02492172e-01 4.80195999e-01
-6.53756022e-01 -5.94997704e-01 3.49676132e-01 -3.75153035e-01
1.13556135e+00 -2.41057873e-01 9.32544887e-01 1.19844365e+00
-2.29594782e-01 1.14628303e+00 1.00205374e+00 -3.40654552e-01
6.16525233e-01 -1.04730569e-01 -5.57325743e-02 9.78344440e-01
2.14746088e-01 5.87119043e-01 -1.83539867e-01 -2.56679565e-01
7.95943737e-01 -1.19166851e-01 7.46072680e-02 -9.20169830e-01
-1.11423314e+00 7.98063278e-01 4.28701282e-01 -1.80443108e-01
-4.30659711e-01 7.28471041e-01 3.09775740e-01 5.26046038e-01
-1.82752192e-01 4.66974676e-02 -3.57585132e-01 -7.05838859e-01
-1.80278718e-01 7.31079400e-01 1.01174307e+00 1.07678437e+00
1.03639328e+00 -1.14460006e-01 2.75698274e-01 5.16478419e-01
4.11893487e-01 7.55113780e-01 2.79185712e-01 -1.68334341e+00
4.39958274e-01 2.82217950e-01 7.97597468e-01 -6.61004961e-01
-5.27625680e-01 1.63172573e-01 -3.87565531e-02 5.62601686e-01
9.74160314e-01 -4.05075312e-01 -6.24122858e-01 1.76927710e+00
5.06927848e-01 2.79720366e-01 4.05624062e-01 7.25099146e-01
-5.85170947e-02 3.47674251e-01 1.59982428e-01 2.40800288e-02
7.82558560e-01 -1.08550406e+00 -2.23366022e-01 -1.64722711e-01
1.12319267e+00 -2.05802277e-01 1.02317226e+00 2.68383950e-01
-8.42572212e-01 -1.50935814e-01 -8.65831912e-01 1.88715890e-01
-1.89500615e-01 -1.41927049e-01 5.03886044e-01 7.29981005e-01
-1.06646729e+00 1.11375833e+00 -1.12494051e+00 -7.98959494e-01
3.42542112e-01 5.92843652e-01 -2.64065534e-01 2.19810158e-01
-7.70191193e-01 8.54900658e-01 2.29148895e-01 -4.91696179e-01
-9.62022960e-01 -6.11366890e-02 -8.95966053e-01 -2.99069941e-01
6.07917845e-01 -4.79978621e-01 1.80440187e+00 -7.35464931e-01
-2.00471687e+00 3.84624034e-01 -9.69254449e-02 -9.02015746e-01
8.54830623e-01 -3.37150991e-01 1.27771497e-01 2.29123279e-01
1.80708304e-01 8.21933925e-01 5.59633553e-01 -1.30105984e+00
-9.84105110e-01 -2.39411891e-02 3.87992322e-01 2.72412121e-01
5.59137156e-03 -4.31669295e-01 -9.91970524e-02 -5.95598407e-02
-4.60173905e-01 -1.47361851e+00 -8.41891110e-01 9.22533199e-02
3.02509852e-02 -1.61219165e-01 6.68371737e-01 -8.33531097e-02
6.32355511e-01 -1.79006195e+00 -7.28871375e-02 3.40335310e-01
-1.18456043e-01 2.09674284e-01 -2.73236722e-01 6.83678269e-01
6.24476910e-01 -4.14039850e-01 -3.54585707e-01 -3.72036278e-01
2.99809009e-01 7.70879686e-01 -2.08275333e-01 6.55463398e-01
-1.48138359e-01 9.32196558e-01 -1.53281677e+00 -3.80661577e-01
3.58301908e-01 -3.58165950e-02 -1.07745874e+00 1.60461381e-01
-2.95651942e-01 5.67464054e-01 -5.65256476e-01 1.90220833e-01
1.21285304e-01 -6.81572184e-02 3.28875273e-01 8.27440262e-01
-1.11881658e-01 2.49716014e-01 -1.16697359e+00 1.27139688e+00
-3.51566344e-01 5.34698009e-01 9.19550378e-03 -1.02461386e+00
6.17628634e-01 1.23169824e-01 4.13924783e-01 -5.87203503e-01
3.57212089e-02 3.49543422e-01 4.60892804e-02 -7.06718981e-01
5.47037363e-01 2.19689533e-02 -2.15103433e-01 7.02272236e-01
-2.05137640e-01 1.84778437e-01 1.66525036e-01 -2.73841899e-02
1.25728571e+00 5.82565427e-01 6.15861654e-01 -2.00120151e-01
5.28864026e-01 4.49264050e-01 3.39139998e-01 1.19075203e+00
-6.62277341e-01 -1.59449086e-01 3.82467270e-01 -4.36458796e-01
-1.14930940e+00 -1.13289988e+00 5.68248987e-01 1.15818119e+00
2.05039784e-01 -2.18210280e-01 -9.20776010e-01 -8.38004529e-01
2.06767112e-01 6.50076330e-01 -5.46587169e-01 6.56138361e-02
-1.15285599e+00 -3.74165595e-01 4.92308229e-01 5.77469468e-01
3.30893487e-01 -1.22555959e+00 -1.12199008e+00 4.73254561e-01
4.24745083e-02 -1.10867882e+00 -3.07093829e-01 2.52734069e-02
-7.28703260e-01 -1.30126607e+00 -5.76211631e-01 -5.09233654e-01
6.37857556e-01 3.78444731e-01 6.91571712e-01 1.74973294e-01
3.05056721e-02 1.00432050e+00 -4.12647069e-01 -2.51520157e-01
-7.16061711e-01 7.62000307e-02 4.07000095e-01 -2.90675104e-01
3.75729054e-01 -6.48178101e-01 -6.25167251e-01 2.87849903e-01
-3.41474265e-01 -1.04245491e-01 3.62477660e-01 9.34932709e-01
1.12963021e-01 -4.77562010e-01 5.91599822e-01 -8.12131882e-01
6.57043159e-01 -6.30146861e-01 -9.45579469e-01 -3.24127942e-01
-6.16793990e-01 3.44262153e-01 8.22297633e-01 -4.96805131e-01
-8.02714646e-01 4.18609142e-01 7.08754435e-02 5.81794307e-02
-5.57715297e-01 -1.42763183e-01 8.19492489e-02 -1.53034404e-01
7.52377033e-01 2.80966252e-01 4.31597888e-01 -1.75736710e-01
6.95343316e-01 6.25807464e-01 3.60157639e-01 -8.10619771e-01
7.04303086e-01 6.95271373e-01 8.51309374e-02 -9.73568916e-01
-2.26348594e-01 -4.82255876e-01 -5.30782998e-01 -4.97862905e-01
6.47361219e-01 -6.95581853e-01 -1.41676629e+00 2.97877848e-01
-8.74019921e-01 -1.01057887e+00 -4.15369928e-01 6.71761155e-01
-1.63717210e+00 5.45273900e-01 -5.88631511e-01 -1.15858567e+00
5.04206479e-01 -1.16727638e+00 8.44324350e-01 1.56766400e-01
-2.50163049e-01 -1.01897383e+00 4.66524869e-01 4.33740625e-03
2.16822699e-01 -4.28807139e-02 4.46780801e-01 -7.12761462e-01
-9.60202336e-01 2.49310911e-01 2.74734143e-02 -2.87864000e-01
6.13299664e-03 -4.53286260e-01 -6.45479262e-01 -2.80387580e-01
-5.55207133e-01 -3.75616908e-01 6.87367976e-01 3.26482683e-01
4.92823720e-01 -5.95833480e-01 -5.50141931e-01 9.82296541e-02
1.18559909e+00 3.69945109e-01 4.19352174e-01 9.41787839e-01
2.77956724e-01 9.84194756e-01 8.95655215e-01 6.44494116e-01
7.63415158e-01 6.54973567e-01 3.80959541e-01 3.14255506e-01
1.95077747e-01 -5.57004869e-01 6.70495450e-01 2.64187783e-01
-1.60316616e-01 -1.24482289e-01 -8.22461247e-01 8.26293647e-01
-2.67299247e+00 -1.22933900e+00 -5.30599989e-02 1.95244753e+00
5.22290528e-01 2.75164843e-02 9.06958282e-01 5.98413274e-02
3.98274869e-01 -2.69893587e-01 -7.32029200e-01 -3.07633966e-01
3.44737947e-01 -2.76780397e-01 7.92770088e-01 1.11502194e+00
-1.05409443e+00 1.25601530e+00 6.89120865e+00 3.38683218e-01
-3.96885335e-01 -1.16527779e-02 5.58982268e-02 4.70607774e-03
7.45746046e-02 1.31574675e-01 -7.61124611e-01 1.66755453e-01
1.13257527e+00 -4.33417968e-02 7.40535676e-01 9.39550817e-01
4.43455487e-01 -4.52594787e-01 -1.27221704e+00 6.78525269e-01
-4.53110307e-01 -1.21885180e+00 -3.61461312e-01 3.83055180e-01
5.34291983e-01 1.52824134e-01 -1.95813000e-01 4.87252474e-01
1.15816081e+00 -7.69579768e-01 6.34326994e-01 4.82628822e-01
7.64218941e-02 -9.17473555e-01 2.49285325e-01 7.35449314e-01
-1.08485365e+00 -4.37309921e-01 -2.60956198e-01 -6.40761852e-01
4.33048278e-01 -3.03226739e-01 -1.30717850e+00 2.11038277e-01
4.93789375e-01 8.64522815e-01 -1.30384222e-01 1.06806183e+00
-1.54697657e-01 5.39481401e-01 -4.71937180e-01 -7.28997171e-01
6.61468685e-01 -4.97080743e-01 7.13433981e-01 1.20211494e+00
2.42174804e-01 1.80333883e-01 4.90885437e-01 3.83183390e-01
4.40913141e-01 -1.08446740e-01 -1.19471562e+00 1.64901450e-01
4.02253985e-01 6.62608743e-01 -8.45054507e-01 -4.53889310e-01
-3.13208550e-01 9.41123724e-01 5.29717982e-01 3.92137229e-01
-6.40374303e-01 -1.41576171e-01 9.48192537e-01 1.54660828e-02
5.65635800e-01 -6.48255825e-01 2.19854608e-01 -7.28108406e-01
-1.37040347e-01 -6.22645795e-01 1.46950483e-01 -2.82913297e-01
-1.00806963e+00 1.72818646e-01 1.51505202e-01 -1.46895051e+00
-1.02808464e+00 -6.47462904e-01 -1.52496174e-01 1.51664451e-01
-1.56213582e+00 -6.34054005e-01 1.33381084e-01 4.78096515e-01
4.59635198e-01 -3.52342963e-01 6.15528107e-01 -5.15527325e-03
9.42418166e-03 2.54692078e-01 2.28312358e-01 8.00367892e-02
4.27585334e-01 -1.41045022e+00 6.59340918e-01 5.34210563e-01
-1.27460241e-01 3.56433302e-01 9.78147149e-01 -5.86791158e-01
-1.18788445e+00 -8.60268891e-01 4.65733498e-01 -6.03770733e-01
8.29724252e-01 -1.61510840e-01 -4.29990679e-01 9.20540571e-01
1.34870946e-01 -1.67822674e-01 4.80884612e-01 -9.82164368e-02
1.10938743e-01 2.50106156e-01 -1.08303928e+00 1.24431610e+00
1.43324006e+00 -2.24475861e-01 -6.19794309e-01 8.00573304e-02
6.02172256e-01 -2.13402599e-01 -3.25778008e-01 -9.73302405e-03
5.85329235e-01 -1.06784379e+00 1.01854658e+00 -8.88115048e-01
6.24239296e-02 -4.94935274e-01 -2.10267603e-01 -1.29764986e+00
-6.41687438e-02 -8.92866075e-01 -1.72048539e-01 5.70100486e-01
3.12301159e-01 -8.38385046e-01 1.16518712e+00 5.62727511e-01
1.01628408e-01 -5.16350806e-01 -1.02721786e+00 -1.12740958e+00
2.06813514e-02 -4.93832797e-01 4.51825917e-01 3.71617734e-01
3.49353969e-01 7.92019367e-02 -5.98494530e-01 1.25203222e-01
9.06969070e-01 -1.83489889e-01 1.34545600e+00 -8.18324327e-01
-5.07132113e-01 -3.66754055e-01 -2.65369892e-01 -1.53477323e+00
4.01856512e-01 -7.14090943e-01 4.37688500e-01 -1.31582689e+00
-9.73532274e-02 -5.62362194e-01 1.47103831e-01 3.15082222e-01
1.51770875e-01 -2.22159013e-01 2.93868244e-01 4.14981544e-02
-1.02442515e+00 7.82346427e-01 1.11748016e+00 3.21722329e-01
-3.33448052e-01 3.29550177e-01 -1.84751824e-01 1.03176284e+00
9.83740747e-01 -6.58150911e-01 -5.68021476e-01 -5.29254787e-02
-1.18603297e-01 3.24306607e-01 5.29902220e-01 -1.08392727e+00
4.90693897e-01 -4.34266567e-01 -3.57218869e-02 -2.86723226e-01
4.97928888e-01 -9.03619230e-01 -3.03180754e-01 1.15171671e+00
-4.22673672e-01 2.06690907e-01 -1.28144816e-01 1.23884046e+00
4.20258909e-01 -3.93963635e-01 6.40316248e-01 -3.05787235e-01
-8.86705697e-01 7.64194876e-02 -1.29177713e+00 -8.53028074e-02
1.24946201e+00 -3.78850162e-01 -2.20562577e-01 -7.93938100e-01
-7.26810634e-01 5.24988413e-01 8.62594783e-01 2.42205381e-01
6.24424636e-01 -1.05408895e+00 -3.22008640e-01 4.01489921e-02
2.42083333e-04 -5.64714491e-01 -1.95222482e-01 6.94372356e-01
-5.45979023e-01 3.29013437e-01 -4.32956755e-01 -4.98092949e-01
-1.01498187e+00 5.66986382e-01 4.42718327e-01 -1.96527749e-01
-9.32754815e-01 2.11340085e-01 -2.09774211e-01 -9.18932617e-01
2.30534717e-01 -2.86192626e-01 -3.46779048e-01 -4.32936281e-01
4.13399458e-01 7.54464746e-01 -6.55651748e-01 -4.67325896e-01
-3.00342023e-01 3.19896281e-01 1.50915205e-01 -8.43579710e-01
1.37045264e+00 -3.85612309e-01 4.73488599e-01 3.78323466e-01
7.87759602e-01 -5.93444519e-02 -1.97231770e+00 -2.70195872e-01
5.83399534e-01 -3.78880769e-01 -7.04617321e-01 -2.45813712e-01
-3.57259661e-01 5.08136988e-01 5.11305511e-01 8.54953751e-02
5.04846871e-01 -2.23733187e-01 9.03130770e-01 1.19993615e+00
9.43408251e-01 -1.34512699e+00 2.12071672e-01 7.94029117e-01
2.41558149e-01 -1.41851771e+00 -2.37813517e-01 -4.95428815e-02
-6.84235334e-01 1.09135413e+00 5.56439757e-01 -6.84175313e-01
6.00587189e-01 2.58229882e-01 -7.69066298e-03 2.16525942e-01
-8.35767567e-01 -6.79835737e-01 -4.59340811e-01 1.19550598e+00
-2.45190382e-01 6.23723976e-02 -2.31638789e-01 2.48274565e-01
-2.75292009e-01 1.30082980e-01 8.90484750e-01 1.26490259e+00
-1.16931605e+00 -1.42158937e+00 -2.29674131e-01 7.78539702e-02
7.51086548e-02 4.75664377e-01 -2.27044970e-01 1.02530873e+00
-3.65177363e-01 1.16258335e+00 -8.00224468e-02 -2.88682044e-01
2.56045431e-01 -1.90310970e-01 5.81363916e-01 -2.62802184e-01
-1.86182678e-01 -2.99547791e-01 4.15791571e-01 -9.46134031e-01
-6.69990420e-01 -1.04301977e+00 -1.67663229e+00 -3.84130627e-01
4.59462225e-01 1.12195529e-01 3.19587499e-01 1.14153135e+00
2.26967663e-01 3.23378593e-02 5.60535371e-01 -1.24919236e+00
-7.17599928e-01 -4.03010249e-01 -1.46197721e-01 2.19128385e-01
7.36672699e-01 -9.14580524e-01 -8.35644007e-02 -5.03072105e-02] | [4.671984672546387, 1.1172970533370972] |
88dd2717-aceb-4a09-9c74-8b06068cdfad | smartbrush-text-and-shape-guided-object | 2212.05034 | null | https://arxiv.org/abs/2212.05034v1 | https://arxiv.org/pdf/2212.05034v1.pdf | SmartBrush: Text and Shape Guided Object Inpainting with Diffusion Model | Generic image inpainting aims to complete a corrupted image by borrowing surrounding information, which barely generates novel content. By contrast, multi-modal inpainting provides more flexible and useful controls on the inpainted content, \eg, a text prompt can be used to describe an object with richer attributes, and a mask can be used to constrain the shape of the inpainted object rather than being only considered as a missing area. We propose a new diffusion-based model named SmartBrush for completing a missing region with an object using both text and shape-guidance. While previous work such as DALLE-2 and Stable Diffusion can do text-guided inapinting they do not support shape guidance and tend to modify background texture surrounding the generated object. Our model incorporates both text and shape guidance with precision control. To preserve the background better, we propose a novel training and sampling strategy by augmenting the diffusion U-net with object-mask prediction. Lastly, we introduce a multi-task training strategy by jointly training inpainting with text-to-image generation to leverage more training data. We conduct extensive experiments showing that our model outperforms all baselines in terms of visual quality, mask controllability, and background preservation. | ['Kun Zhang', 'Tobias Hinz', 'Zhe Lin', 'Zhifei Zhang', 'Shaoan Xie'] | 2022-12-09 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xie_SmartBrush_Text_and_Shape_Guided_Object_Inpainting_With_Diffusion_Model_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xie_SmartBrush_Text_and_Shape_Guided_Object_Inpainting_With_Diffusion_Model_CVPR_2023_paper.pdf | cvpr-2023-1 | ['image-inpainting'] | ['computer-vision'] | [ 3.54425758e-01 3.46132278e-01 -2.90528268e-01 -9.35118198e-02
-7.04769313e-01 -5.45170248e-01 4.99279559e-01 -2.29883492e-01
2.80097201e-02 7.75952697e-01 3.33806545e-01 6.68668468e-03
3.36105406e-01 -6.75768137e-01 -1.01399279e+00 -6.81343555e-01
4.98216450e-01 2.76332259e-01 3.15561235e-01 -8.43652040e-02
9.55330133e-02 5.87922275e-01 -1.17376077e+00 6.12082183e-01
1.11365294e+00 1.17669868e+00 7.34004021e-01 5.25258243e-01
-8.12725872e-02 1.01972854e+00 -4.85421330e-01 -8.80011767e-02
7.19045401e-01 -6.26847506e-01 -3.85706842e-01 7.66327083e-01
8.05077672e-01 -8.42854500e-01 -4.04079318e-01 7.74818540e-01
4.58739311e-01 3.27430874e-01 6.92892075e-01 -8.46721590e-01
-9.26116228e-01 2.85329491e-01 -9.27682161e-01 -7.44590983e-02
1.76528290e-01 5.42110443e-01 6.98218226e-01 -9.00098026e-01
9.47000623e-01 1.31469679e+00 4.13055837e-01 8.01361263e-01
-1.59361804e+00 -4.70901191e-01 4.02711540e-01 -3.54576796e-01
-9.23954368e-01 -6.48197412e-01 1.17467129e+00 -3.29623461e-01
4.08926040e-01 4.41212922e-01 5.75566709e-01 1.04901648e+00
2.71658123e-01 1.03845215e+00 1.10397911e+00 -4.46463406e-01
1.61004975e-01 1.33597001e-01 -5.58206797e-01 7.41074145e-01
-1.05763830e-01 1.86596379e-01 -4.26768184e-01 -4.75563742e-02
1.33604920e+00 2.63600081e-01 -4.76794630e-01 -5.14102697e-01
-1.22714806e+00 7.25236535e-01 4.39611912e-01 -8.65725577e-02
-4.80176359e-01 3.36785704e-01 7.73459077e-02 2.95597255e-01
1.01141107e+00 5.57187021e-01 -3.22684944e-01 3.41317415e-01
-1.24855602e+00 4.66999471e-01 4.13126826e-01 1.16562724e+00
8.74862850e-01 2.90433943e-01 -6.09078705e-01 9.46965992e-01
3.15633044e-02 5.25252938e-01 1.46996066e-01 -1.74953783e+00
3.86717349e-01 4.66216236e-01 4.96548414e-01 -8.05489719e-01
1.55077219e-01 -2.42941320e-01 -8.57439637e-01 7.97915697e-01
1.97105318e-01 -3.35651219e-01 -1.16121221e+00 1.56876063e+00
4.03274953e-01 4.87218425e-02 -4.61583823e-01 9.54364896e-01
4.80727762e-01 9.69247580e-01 -3.40707690e-01 -3.43216300e-01
8.50016057e-01 -1.47294605e+00 -8.18030357e-01 -4.13583100e-01
2.91877598e-01 -8.06470335e-01 9.40873682e-01 4.97887224e-01
-1.67795324e+00 -5.48809648e-01 -7.26338446e-01 -3.46475869e-01
8.45412463e-02 6.09271694e-03 3.45759064e-01 2.09308341e-01
-1.21057022e+00 6.85311794e-01 -6.91641927e-01 4.72122319e-02
7.52359629e-01 -9.67605859e-02 -1.66091979e-01 -2.87454724e-01
-5.82291067e-01 7.69510686e-01 1.51051983e-01 -3.62091124e-01
-1.02799690e+00 -1.04071116e+00 -8.88042629e-01 -5.03727235e-02
5.37027001e-01 -1.02983344e+00 9.07079101e-01 -1.23661113e+00
-1.48933923e+00 6.59519434e-01 -7.92506784e-02 -3.60363424e-01
1.04844391e+00 -9.83189866e-02 1.19119085e-01 4.57793027e-01
3.16296011e-01 1.41691601e+00 1.36412871e+00 -1.82668710e+00
-5.72820127e-01 -9.42265149e-03 -1.25439651e-02 3.45656067e-01
-1.57646298e-01 -1.37909740e-01 -7.33263373e-01 -1.37053537e+00
-2.77422871e-02 -7.08572686e-01 -4.54744458e-01 1.02029562e+00
-5.05230606e-01 3.16096097e-01 1.31714189e+00 -8.67339551e-01
1.12154007e+00 -2.30823898e+00 1.77955538e-01 -1.85895994e-01
3.68017435e-01 1.22689359e-01 -4.11919206e-01 1.96532369e-01
1.40459120e-01 1.32205501e-01 -5.39109170e-01 -9.08134878e-01
-3.62395436e-01 4.25593972e-01 -6.30555332e-01 3.00823510e-01
4.13246542e-01 9.40976083e-01 -8.34815562e-01 -6.17435217e-01
3.74449015e-01 5.25253415e-01 -8.25289130e-01 2.59941697e-01
-8.01288366e-01 6.11782372e-01 -4.00718778e-01 8.39781106e-01
8.28629315e-01 -2.05043599e-01 -3.09924752e-01 -1.49825469e-01
-1.34275869e-01 -2.79701263e-01 -1.10178137e+00 1.87031794e+00
-6.41179144e-01 6.68441713e-01 6.50756121e-01 -4.86814976e-01
6.90264583e-01 1.67265050e-02 5.59707880e-01 -6.08597398e-01
-2.01927364e-01 -2.26819376e-03 -4.10827011e-01 -3.56314778e-01
6.17153406e-01 8.34631920e-02 4.56650764e-01 5.61955690e-01
-3.84982228e-01 -6.07160926e-01 5.04721813e-02 3.84339482e-01
9.05447245e-01 4.31609422e-01 -1.78074330e-01 -1.95394561e-01
2.07524169e-02 -2.63088405e-01 5.26613057e-01 8.21329534e-01
-9.34972335e-03 1.25670969e+00 2.80465424e-01 -3.60654831e-01
-1.23790550e+00 -9.39538658e-01 -1.14157572e-02 8.68628621e-01
2.45239809e-01 -2.03694217e-02 -7.09796607e-01 -7.01255202e-01
1.43988384e-02 7.82020509e-01 -7.42268622e-01 1.30828917e-01
-6.90983176e-01 -1.84169605e-01 9.35958400e-02 3.87230337e-01
4.85955656e-01 -1.27101874e+00 -5.99970698e-01 3.34346533e-01
-4.40768003e-01 -1.02131462e+00 -1.32619369e+00 1.09421156e-01
-1.01909435e+00 -6.72311902e-01 -1.11537480e+00 -7.90980995e-01
9.36873615e-01 4.21646684e-01 1.01765203e+00 2.34007075e-01
-2.50258565e-01 3.06600213e-01 -2.79978365e-01 -1.59008175e-01
-6.80183768e-01 -2.33619928e-01 -2.58287609e-01 2.48017326e-01
-7.91563332e-01 -6.98608518e-01 -9.73208368e-01 3.27626556e-01
-1.41835213e+00 5.16998410e-01 7.38264143e-01 8.77141118e-01
7.87562013e-01 -1.25308514e-01 3.00531894e-01 -6.67394102e-01
4.60144013e-01 -3.08675677e-01 -2.42384583e-01 1.18146442e-01
-4.78470922e-01 8.51054955e-03 5.83994508e-01 -9.35987771e-01
-1.14089692e+00 1.35246739e-01 1.80072770e-01 -1.06520760e+00
1.23103723e-01 -1.01790145e-01 -1.64154202e-01 -1.82063684e-01
6.35352194e-01 4.20174420e-01 3.00453305e-01 -4.87320840e-01
6.08623803e-01 1.69802710e-01 4.06471580e-01 -7.59489715e-01
6.54084444e-01 9.24324453e-01 -7.41013214e-02 -6.61314964e-01
-7.88548708e-01 -1.47409558e-01 -2.74711579e-01 -1.37807488e-01
7.81608105e-01 -8.83296311e-01 -1.13401920e-01 5.02021313e-01
-1.31218767e+00 -9.47912037e-01 -7.04231024e-01 -1.79205775e-01
-7.82597899e-01 5.27401924e-01 -7.91160166e-01 -7.15327740e-01
-3.30648720e-01 -1.08105278e+00 1.35335171e+00 -1.40635982e-01
1.70799401e-02 -9.05365109e-01 -3.44475865e-01 2.60119468e-01
6.19338572e-01 5.10740459e-01 5.87598681e-01 1.04981266e-01
-9.99243498e-01 1.53215438e-01 -3.79963964e-01 5.95604241e-01
2.86718100e-01 -1.37053356e-01 -7.26118207e-01 -3.60400170e-01
1.50251120e-01 -4.78752106e-01 1.24711788e+00 6.76490605e-01
1.37710130e+00 -7.54868567e-01 -4.73726124e-01 7.94514418e-01
1.25535059e+00 9.45106894e-02 7.74815679e-01 1.37531623e-01
8.45372796e-01 6.97909355e-01 5.91764688e-01 5.17545104e-01
1.35346845e-01 7.31449902e-01 5.50112367e-01 -4.15302962e-01
-7.50022411e-01 -3.83077413e-01 4.32027102e-01 1.54633880e-01
1.94378808e-01 -5.46189249e-01 -2.34849811e-01 5.59243798e-01
-1.73719811e+00 -9.66122448e-01 1.33382842e-01 1.83834803e+00
1.20487249e+00 -4.62421626e-02 2.00771051e-03 -2.78571129e-01
6.43462420e-01 4.74876165e-01 -7.94545233e-01 8.11236277e-02
-2.90904224e-01 -1.84511453e-01 4.16451603e-01 9.71680939e-01
-7.87063003e-01 9.90170479e-01 6.11759520e+00 1.19977546e+00
-1.05305624e+00 1.82473242e-01 1.25983250e+00 -4.12297636e-01
-6.38414145e-01 -2.54376773e-02 -4.95995015e-01 6.76742256e-01
9.04216543e-02 2.76763856e-01 4.85931247e-01 6.72665715e-01
5.51936030e-01 -3.44209224e-01 -9.31598485e-01 8.98922622e-01
2.08354309e-01 -1.80476665e+00 4.28277254e-01 1.92129642e-01
1.19894075e+00 -1.60510182e-01 2.43877590e-01 -1.38576463e-01
2.21003741e-01 -5.95377028e-01 1.15040052e+00 7.96870589e-01
9.96447921e-01 -4.27417397e-01 -8.99116173e-02 4.20516253e-01
-7.89044142e-01 -7.99556077e-02 -2.62477756e-01 1.11035757e-01
3.09023827e-01 7.94062495e-01 -2.95649886e-01 2.63078958e-01
5.00852823e-01 6.49357855e-01 -4.03349102e-01 7.18583584e-01
-3.07366922e-02 2.51118600e-01 -1.41295433e-01 4.78617966e-01
8.98619592e-02 -3.48583251e-01 7.96777904e-01 1.03659356e+00
3.47042531e-01 2.02403799e-01 5.64488053e-01 1.46047926e+00
-1.84136480e-02 -1.25485867e-01 -6.23341858e-01 3.23731601e-01
2.73259878e-01 1.08722484e+00 -7.32190788e-01 -5.45163631e-01
-2.47933134e-01 1.48536289e+00 1.25309750e-01 5.91136277e-01
-6.89466476e-01 -4.28733081e-02 4.69171226e-01 6.79434061e-01
5.76851010e-01 -1.85621709e-01 -5.45560718e-01 -1.25993276e+00
1.50603414e-01 -7.41481423e-01 -9.96549055e-02 -1.21621299e+00
-1.25692856e+00 4.27720815e-01 -8.79492089e-02 -1.28322184e+00
-8.12088698e-02 -2.99589038e-01 -6.69504404e-01 8.87165487e-01
-1.53541613e+00 -1.36105466e+00 -2.04830185e-01 5.49929261e-01
1.03034759e+00 5.41634709e-02 2.31707469e-01 8.45520422e-02
-2.72522092e-01 3.17796648e-01 4.11545299e-03 -1.49378568e-01
1.01914036e+00 -1.10714662e+00 1.99176237e-01 9.24688160e-01
-2.58089751e-01 3.90179992e-01 7.85138965e-01 -9.75186288e-01
-1.19136167e+00 -1.48586094e+00 3.33196700e-01 -5.08946300e-01
2.90115327e-01 -3.27488631e-01 -8.43825996e-01 5.31489372e-01
4.37644929e-01 3.35536987e-01 -9.97336060e-02 -6.97427630e-01
-1.97149217e-01 -1.12369783e-01 -1.31819689e+00 8.38371754e-01
1.02992857e+00 -1.41808480e-01 -1.29192800e-03 4.35975224e-01
1.01635015e+00 -6.46943450e-01 -4.50653464e-01 6.21012226e-02
2.07484886e-01 -1.12877464e+00 1.02070904e+00 -8.81381929e-02
8.53433847e-01 -3.78407955e-01 -9.66987386e-02 -1.43555570e+00
-3.97019267e-01 -1.05042315e+00 -5.15991390e-01 1.07631588e+00
1.61711007e-01 -3.53781104e-01 8.31140280e-01 6.16023302e-01
-5.30355453e-01 -1.01138389e+00 -5.85821211e-01 -7.64364719e-01
4.50265668e-02 -2.57607669e-01 2.82618999e-01 9.19446945e-01
-5.47388315e-01 -6.19099922e-02 -7.99686432e-01 -1.84511572e-01
5.87379813e-01 1.45448908e-01 7.90396512e-01 -7.07567811e-01
-4.63302970e-01 -5.11266232e-01 2.06705719e-01 -1.58897507e+00
-1.38743892e-02 -4.95133132e-01 1.51314825e-01 -1.61703444e+00
1.06773630e-01 -4.34924453e-01 2.40016714e-01 6.25058651e-01
-1.49208099e-01 4.45746928e-01 3.69971126e-01 3.55078042e-01
-3.91443521e-01 7.06629515e-01 2.07839417e+00 -2.97645599e-01
-4.52795178e-01 -1.72607213e-01 -7.84721255e-01 5.07187188e-01
4.44143713e-01 -2.70495385e-01 -4.81218755e-01 -6.26570761e-01
2.27416437e-02 5.07400155e-01 5.08206367e-01 -5.88895738e-01
6.78154603e-02 -4.14561152e-01 6.78957760e-01 -5.13890445e-01
5.38981378e-01 -7.81729877e-01 -2.81120893e-02 3.05234909e-01
-3.84582996e-01 -1.27696201e-01 1.31776690e-01 7.83915818e-01
4.58994284e-02 3.36355269e-02 9.93281424e-01 -2.59463936e-01
-3.05416316e-01 6.38325512e-01 -2.29174390e-01 -1.22166250e-03
1.01113749e+00 -4.79103655e-01 -1.67391241e-01 -6.85810030e-01
-7.38677502e-01 2.79199868e-01 9.93280113e-01 3.71222317e-01
8.45849872e-01 -1.32706046e+00 -8.47802520e-01 2.89643139e-01
-1.60577834e-01 3.00741673e-01 3.30344111e-01 6.12178683e-01
-4.88046914e-01 -1.51721016e-01 -2.66044568e-02 -5.71730375e-01
-8.75529110e-01 7.27240622e-01 2.89400697e-01 -1.47778302e-01
-9.03557599e-01 6.74848020e-01 7.70729661e-01 -2.65720463e-03
2.62486100e-01 -4.86100972e-01 4.51315880e-01 -8.90400261e-02
6.33019269e-01 2.62180597e-01 -3.64181608e-01 -1.96685851e-01
2.20351920e-01 3.09835076e-01 -3.59101862e-01 -3.66069734e-01
1.26631534e+00 -4.85872686e-01 -1.62024274e-01 1.49237409e-01
7.84019947e-01 3.11400652e-01 -2.24326158e+00 -2.61475652e-01
-7.30572701e-01 -7.25592434e-01 2.44278952e-01 -9.76659179e-01
-1.39439499e+00 4.89036232e-01 3.59689236e-01 1.05711080e-01
1.21183956e+00 -5.20457290e-02 9.93144512e-01 4.25058184e-04
1.29023463e-01 -1.06380332e+00 7.47959793e-01 1.49406314e-01
1.42736304e+00 -1.25601232e+00 2.61261705e-02 -5.06932318e-01
-7.88609147e-01 9.38131154e-01 8.52149189e-01 -2.07171425e-01
5.31096280e-01 4.32360917e-01 3.94285023e-02 1.04420893e-01
-7.57358551e-01 1.31219164e-01 3.22263896e-01 7.17363119e-01
2.53232680e-02 -2.97968298e-01 -2.41205040e-02 1.28577337e-01
2.89559960e-01 -5.71178347e-02 6.44620001e-01 8.19710255e-01
-5.06780744e-01 -9.55078125e-01 -5.33077240e-01 5.53655505e-01
-2.93238163e-01 -2.51175553e-01 -2.47855008e-01 4.98941034e-01
3.06561977e-01 8.35915685e-01 -3.18249641e-03 7.84388483e-02
-2.18735859e-02 -3.93549711e-01 5.35340548e-01 -7.79587448e-01
-2.99998432e-01 5.06052434e-01 -2.19266355e-01 -7.11550057e-01
-3.61616820e-01 -5.75078249e-01 -8.86811256e-01 -8.29691291e-02
-2.65311778e-01 -3.78451556e-01 2.91299075e-01 7.58887589e-01
5.64968467e-01 5.20380020e-01 7.01446891e-01 -1.37148285e+00
-3.03890377e-01 -8.30183506e-01 -5.51666319e-01 2.72134751e-01
7.25665092e-01 -3.46374512e-01 -4.03054088e-01 3.63356799e-01] | [11.423259735107422, -0.7251989841461182] |
2d27f67f-105f-46a8-8f29-c8cdc7b00fd6 | jpeg-compressed-images-can-bypass-protections | 2304.02234 | null | https://arxiv.org/abs/2304.02234v2 | https://arxiv.org/pdf/2304.02234v2.pdf | JPEG Compressed Images Can Bypass Protections Against AI Editing | Recently developed text-to-image diffusion models make it easy to edit or create high-quality images. Their ease of use has raised concerns about the potential for malicious editing or deepfake creation. Imperceptible perturbations have been proposed as a means of protecting images from malicious editing by preventing diffusion models from generating realistic images. However, we find that the aforementioned perturbations are not robust to JPEG compression, which poses a major weakness because of the common usage and availability of JPEG. We discuss the importance of robustness for additive imperceptible perturbations and encourage alternative approaches to protect images against editing. | ['Tom Goldstein', 'Jonas Geiping', 'Pedro Sandoval-Segura'] | 2023-04-05 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 5.63183665e-01 -1.53674558e-01 1.55233830e-01 1.29344931e-03
-4.33488041e-01 -8.65476847e-01 1.03799188e+00 3.90145555e-02
-4.16395098e-01 5.80929220e-01 2.55853355e-01 -5.24397850e-01
1.54687837e-01 -6.97032571e-01 -7.62012720e-01 -5.03244758e-01
-8.05599168e-02 -4.62127149e-01 1.45217359e-01 -2.65735388e-01
7.08153963e-01 7.06939459e-01 -1.17355478e+00 1.18595779e-01
7.88904905e-01 4.83781427e-01 6.05180413e-02 1.15591848e+00
1.87517509e-01 1.12801683e+00 -1.07613409e+00 -7.13587165e-01
6.33223593e-01 -5.49275100e-01 -4.25498903e-01 4.11784768e-01
4.99354511e-01 -8.98564219e-01 -7.21431792e-01 1.21377647e+00
5.79228520e-01 -2.29003802e-01 5.14782012e-01 -1.33636558e+00
-1.38602018e+00 2.63186783e-01 -7.50588953e-01 2.78265446e-01
4.68544245e-01 4.88259435e-01 2.23630890e-01 -3.31801742e-01
7.01687038e-01 1.24037993e+00 4.96441692e-01 6.13654137e-01
-1.11588836e+00 -5.79826951e-01 -3.70262355e-01 -9.57634896e-02
-1.43590319e+00 -8.94905746e-01 5.10025024e-01 -2.85740614e-01
9.03891385e-01 6.29221082e-01 2.85124809e-01 1.19083333e+00
8.50433588e-01 3.49079341e-01 1.30835462e+00 -5.34218431e-01
5.01512773e-02 2.64419705e-01 -5.53523004e-01 4.57455665e-01
3.51186037e-01 1.71717420e-01 -3.24572504e-01 -6.17389977e-01
9.23977196e-01 -3.91057581e-01 -4.12702769e-01 -8.30336101e-03
-1.11031103e+00 7.85553217e-01 1.86808482e-01 1.34915203e-01
-2.07817122e-01 4.03843820e-01 4.39407974e-01 4.80104953e-01
3.59269112e-01 5.42942286e-01 1.98925391e-01 -1.41559571e-01
-9.13260043e-01 -1.60223339e-02 5.58936119e-01 7.92589843e-01
2.38635808e-01 5.21691799e-01 2.59604573e-01 6.23407841e-01
2.76659399e-01 6.34656727e-01 3.81229162e-01 -1.34938133e+00
4.26297843e-01 -1.76121548e-01 1.61884412e-01 -1.59699583e+00
4.05634165e-01 1.50636166e-01 -8.67736042e-01 6.77083075e-01
1.81258559e-01 -1.29155695e-01 -8.45583260e-01 1.34193099e+00
-1.57852113e-01 -2.04838380e-01 5.04887179e-02 4.33962971e-01
1.22144923e-01 8.84077668e-01 2.29378566e-02 -2.02881932e-01
8.46047819e-01 -6.13514900e-01 -8.06265116e-01 -2.70766467e-01
1.49396077e-01 -1.23121166e+00 8.14839363e-01 3.88164997e-01
-1.24098289e+00 -2.87472337e-01 -1.28240550e+00 1.35790892e-02
-3.29946220e-01 -5.35568476e-01 2.04688638e-01 1.43031442e+00
-1.41862917e+00 5.39002240e-01 -4.69253659e-01 -3.12688082e-01
5.36579907e-01 2.55434871e-01 -5.22301793e-01 -8.82622004e-02
-1.19504046e+00 1.17857742e+00 3.37886810e-02 -1.65927470e-01
-9.03549194e-01 -3.41597140e-01 -6.64843857e-01 -3.41721267e-01
-1.42572090e-01 -2.77618527e-01 6.83178365e-01 -1.25299072e+00
-1.26796591e+00 7.98130453e-01 2.20692344e-02 -6.34126723e-01
1.10722470e+00 -2.60510924e-03 -8.34354103e-01 4.13983464e-01
-1.86393857e-01 7.11332262e-01 1.53669906e+00 -1.47864759e+00
-1.75197460e-02 8.81252214e-02 2.47953683e-02 1.13679297e-01
-6.68775558e-01 2.82708287e-01 -2.67519772e-01 -9.95946407e-01
-5.26211023e-01 -8.15782964e-01 -2.65462995e-01 4.09329742e-01
-5.24025619e-01 7.08845794e-01 1.14793265e+00 -8.46302152e-01
1.05256891e+00 -2.44595790e+00 -4.81824249e-01 3.11401993e-01
3.41715366e-01 4.86124933e-01 -3.92235249e-01 6.79643154e-01
5.01737036e-02 9.31196153e-01 -1.40177161e-01 2.21938873e-03
-2.79743165e-01 2.48646066e-01 -4.02957022e-01 7.50310540e-01
-4.25731614e-02 7.18696594e-01 -5.36588073e-01 -4.11362410e-01
3.60538691e-01 9.64214683e-01 -3.10576290e-01 -1.27931014e-01
1.69734254e-01 2.13803217e-01 -1.20463399e-02 4.59124237e-01
9.59230602e-01 1.09316669e-01 -4.14666720e-02 1.05713867e-01
-4.72981445e-02 -6.19567782e-02 -7.80632734e-01 8.83245766e-01
-1.42493382e-01 1.14477038e+00 1.91431850e-01 -2.09545195e-01
6.13254964e-01 2.78608352e-01 1.42971069e-01 -5.08479774e-01
1.01294413e-01 9.54281241e-02 1.10938866e-02 -2.50166684e-01
8.46285462e-01 6.95044696e-02 2.39638925e-01 6.50563776e-01
-6.36694431e-01 -5.02679825e-01 3.48852612e-02 7.29090631e-01
1.03368986e+00 -5.01299500e-01 5.96920028e-02 -2.38050073e-01
2.40110159e-01 -2.69770324e-01 -2.97752526e-02 9.80347633e-01
-3.07468712e-01 7.46722639e-01 2.57549942e-01 -2.37262696e-01
-1.49035132e+00 -9.83029962e-01 1.12009287e-01 4.26048815e-01
2.83162206e-01 -4.68069315e-01 -1.03314507e+00 -4.35828626e-01
4.45178524e-02 7.77265966e-01 -6.25582516e-01 -3.84619147e-01
-2.80975670e-01 -6.36978149e-01 1.11661041e+00 1.15632489e-01
7.59852767e-01 -7.22420871e-01 -5.17682910e-01 8.16854984e-02
-1.60412103e-01 -1.02817452e+00 -8.99053991e-01 -4.26581353e-01
-7.68833637e-01 -7.95944095e-01 -1.04765332e+00 -4.19915617e-01
9.41207051e-01 6.44510150e-01 7.09291220e-01 5.48009992e-01
-4.05716598e-01 7.24340260e-01 -5.18936105e-02 -1.98953703e-01
-1.31521523e+00 -5.03067911e-01 8.71047452e-02 -1.60138998e-02
-1.40867040e-01 -2.05387831e-01 -7.10992694e-01 4.23377544e-01
-1.44674993e+00 -3.03761423e-01 2.65415311e-01 3.08614075e-01
4.15949151e-02 5.96494377e-01 7.21721947e-02 -7.30873942e-01
1.22893059e+00 -2.02525795e-01 -2.96972007e-01 2.51112044e-01
-8.22237313e-01 -1.94292143e-01 6.34623826e-01 -5.34570396e-01
-1.07939029e+00 -3.35021853e-01 -1.77891888e-02 -2.31418815e-02
-6.62615672e-02 1.25170216e-01 1.16334006e-03 -9.67237175e-01
8.41574490e-01 4.00397062e-01 3.72055382e-01 -8.90230164e-02
5.03418505e-01 6.69506192e-01 4.35000300e-01 -2.02203229e-01
1.19714785e+00 7.00609624e-01 -4.09229040e-01 -1.34262455e+00
3.03911000e-01 4.00756806e-01 -1.65480196e-01 -2.89226472e-01
7.32406080e-01 -7.94054627e-01 -3.12327057e-01 1.10234582e+00
-1.05639136e+00 -2.29584098e-01 -7.94483870e-02 -3.04306634e-02
-3.28362942e-01 1.27904463e+00 -1.11247551e+00 -5.20558298e-01
-9.99276415e-02 -1.20146918e+00 3.31498682e-01 1.90357212e-02
-4.70718503e-01 -9.19162333e-01 -1.42256707e-01 4.73368973e-01
9.44548011e-01 2.89829701e-01 8.44247937e-01 1.08766750e-01
-7.33616054e-01 -4.27838564e-01 -2.32543230e-01 6.60974801e-01
4.95920628e-01 6.47731662e-01 -5.70961952e-01 -5.73109150e-01
2.08379462e-01 -2.33237028e-01 5.60511291e-01 2.09692121e-01
9.00638044e-01 -6.71864212e-01 -5.47491536e-02 4.85022902e-01
1.46998084e+00 2.59528309e-01 1.32110703e+00 5.64267933e-01
4.29922938e-01 2.59821475e-01 8.88782367e-02 4.27358329e-01
1.78269669e-02 1.17263168e-01 3.24695110e-01 -2.79358268e-01
-7.26121068e-02 -2.42147267e-01 5.00797033e-01 6.35260940e-01
3.52289751e-02 -6.41588807e-01 -6.83911204e-01 3.48244160e-01
-9.69770372e-01 -1.12340522e+00 -1.98344082e-01 2.15396214e+00
8.09319913e-01 3.59854758e-01 -1.22729771e-01 9.53261033e-02
9.71225441e-01 5.61608315e-01 -3.59072864e-01 -7.91411459e-01
-4.36648190e-01 -2.56404549e-01 1.19403541e+00 6.60155177e-01
-9.57577288e-01 7.59961665e-01 8.20814514e+00 7.26234913e-01
-1.18087304e+00 8.03891867e-02 8.22694004e-01 4.28877361e-02
-4.83444422e-01 -5.09170368e-02 -2.15254545e-01 6.98406279e-01
9.76160526e-01 -5.26186466e-01 5.44494510e-01 2.89598316e-01
3.97098392e-01 -2.95223385e-01 -4.34125185e-01 7.02806950e-01
3.69355828e-01 -1.29157066e+00 4.47582185e-01 3.58701110e-01
8.43318164e-01 -1.16428725e-01 6.27327919e-01 -5.54198027e-01
4.53497440e-01 -1.12473524e+00 7.67994881e-01 7.37462342e-02
9.90936816e-01 -7.64278293e-01 1.66694999e-01 8.31930488e-02
-7.10721135e-01 9.95114222e-02 -6.21691048e-01 5.96018415e-03
3.56839746e-01 3.67573470e-01 -3.49101722e-01 -1.14493296e-01
3.87827367e-01 3.42390031e-01 -8.00899029e-01 8.16834867e-01
-7.88441077e-02 4.23577726e-01 -2.28759080e-01 3.30242455e-01
1.69436842e-01 1.46705687e-01 4.99018490e-01 1.15623021e+00
4.88824308e-01 -1.71759099e-01 -2.68647343e-01 5.80197811e-01
-1.10184811e-01 -2.09875181e-01 -1.03740299e+00 -5.24273992e-01
4.55610365e-01 6.34154856e-01 -8.36042762e-01 -2.63060749e-01
-3.08053225e-01 1.62129962e+00 -4.20900464e-01 5.64908028e-01
-8.17937076e-01 -4.22427803e-01 7.15597451e-01 2.26455301e-01
6.31253421e-02 -6.69727266e-01 -2.99074292e-01 -1.11032510e+00
-6.90160394e-02 -1.48328078e+00 -5.52080683e-02 -7.98504293e-01
-1.10235381e+00 5.90340614e-01 -2.50283957e-01 -9.74063694e-01
8.94159675e-02 -2.14515150e-01 -4.69511688e-01 8.25425088e-01
-1.19724166e+00 -8.91331196e-01 4.59152423e-02 7.90907264e-01
3.66647482e-01 -2.06954360e-01 7.22092509e-01 -3.36242095e-02
-6.43711761e-02 7.24999011e-01 4.91894722e-01 1.19540676e-01
8.61161232e-01 -7.80312181e-01 1.05964482e+00 1.25461042e+00
-3.78545448e-02 8.22585642e-01 8.62322450e-01 -9.16591525e-01
-1.39484036e+00 -7.58927047e-01 6.62483871e-01 -4.40728694e-01
5.01718938e-01 -2.68331945e-01 -8.69225025e-01 5.46766639e-01
6.78984821e-01 -4.09113199e-01 5.41673005e-01 -9.25026655e-01
-5.68388700e-01 2.30903402e-01 -1.58456516e+00 9.80920136e-01
6.12148762e-01 -1.00727725e+00 -1.91036686e-01 1.70401260e-01
4.90993738e-01 -7.09041879e-02 -6.78686440e-01 -2.66574234e-01
4.21620220e-01 -1.11732125e+00 1.25608397e+00 7.33055323e-02
6.45509660e-01 -2.02261940e-01 -1.76820345e-02 -1.21829593e+00
-3.47953588e-01 -1.12650216e+00 1.33167967e-01 1.28091002e+00
2.53433615e-01 -9.46762919e-01 5.04252732e-01 1.07074237e+00
8.31021011e-01 1.50533512e-01 -6.85549617e-01 -9.32509959e-01
1.84824720e-01 -1.23854302e-01 3.06315333e-01 1.28418148e+00
-1.15712516e-01 -4.10484523e-01 -9.79030252e-01 2.14273959e-01
8.65587592e-01 -8.08394849e-01 6.59196377e-01 -4.73628789e-01
-1.72016472e-01 -2.73481071e-01 -5.65127254e-01 -9.13089216e-01
-2.59961426e-01 -2.59855717e-01 -2.57121205e-01 -1.29967761e+00
-1.14204884e-02 -2.29437247e-01 5.46970293e-02 -6.42044768e-02
-8.92738923e-02 7.97508419e-01 5.70801675e-01 5.61785698e-01
4.43097949e-02 9.08676684e-02 1.15437698e+00 -2.94546306e-01
2.59515435e-01 -4.48493242e-01 -9.38604712e-01 5.14486194e-01
1.10112059e+00 -5.39519012e-01 -7.30361044e-01 -7.38155007e-01
2.61057556e-01 -2.01052427e-01 3.50197464e-01 -9.43128228e-01
1.23338155e-01 -2.23363578e-01 4.43632245e-01 7.74652809e-02
2.31217369e-01 -8.52816522e-01 4.44147050e-01 7.66016245e-01
-4.58668172e-01 4.40024078e-01 3.69424015e-01 6.98297858e-01
-1.10350437e-01 -2.85587043e-01 1.08657420e+00 -3.61860037e-01
-5.09597838e-01 -7.75453001e-02 -1.14183974e+00 -8.58743489e-02
1.21651137e+00 -6.52035415e-01 -5.65151870e-01 -1.05738437e+00
-1.96783170e-01 -3.79224509e-01 1.33483338e+00 5.67993641e-01
8.38092625e-01 -9.53867197e-01 -7.44338453e-01 3.37846667e-01
-3.67765903e-01 -1.00481296e+00 2.36355767e-01 2.57900357e-01
-1.27828705e+00 -3.52671444e-02 -4.76749569e-01 -1.56260401e-01
-1.63574994e+00 8.10778558e-01 2.56412029e-01 3.14405710e-01
-5.38667262e-01 7.78985202e-01 -8.11648220e-02 2.79040694e-01
4.86669689e-02 3.70756298e-01 3.32539618e-01 -4.80489701e-01
8.83924603e-01 4.27098513e-01 -2.43150026e-01 -7.73020685e-01
-3.50131363e-01 3.92655343e-01 -5.65663397e-01 -2.98619390e-01
8.80191445e-01 -5.91317773e-01 -3.63403022e-01 -1.87807381e-01
1.22639859e+00 4.89554256e-01 -1.28364432e+00 3.89498055e-01
-4.04150546e-01 -1.12202001e+00 -9.03171971e-02 -7.55146980e-01
-1.09026253e+00 8.20651948e-01 5.56287646e-01 4.76107419e-01
9.17964816e-01 -7.22580731e-01 1.00901616e+00 9.90101919e-02
2.95719117e-01 -1.05781412e+00 -6.17477149e-02 2.30818868e-01
8.56675565e-01 -9.42082763e-01 2.87857175e-01 -4.19941932e-01
-7.03507841e-01 9.45996583e-01 1.97145864e-01 -1.62491463e-02
5.69231749e-01 7.15215445e-01 6.23097479e-01 1.72858641e-01
-4.17559683e-01 6.60975277e-01 -2.14558631e-01 1.12166727e+00
2.37673700e-01 -1.28537416e-01 -3.90722871e-01 -6.53954685e-01
5.06470837e-02 3.37896077e-03 1.24182880e+00 1.31023180e+00
-2.23431900e-01 -1.30091023e+00 -9.35285509e-01 2.05332398e-01
-8.55672836e-01 -2.66138911e-01 -8.53624284e-01 5.32198548e-01
-2.50485897e-01 1.28673935e+00 -3.00681353e-01 -3.53909224e-01
-2.18807325e-01 -2.33004898e-01 4.20502037e-01 7.31153190e-02
-6.82156503e-01 -1.20036930e-01 -6.89669773e-02 -3.37410659e-01
-3.13229561e-01 -2.79613018e-01 -5.54327786e-01 -1.18713415e+00
-2.80248582e-01 -1.25853479e-01 6.04456425e-01 4.76232082e-01
5.72130501e-01 -3.17606181e-02 6.64823532e-01 -5.41006207e-01
-6.76961005e-01 -2.41081551e-01 -6.13016367e-01 6.78271413e-01
3.97030443e-01 7.49196410e-02 -6.23744547e-01 4.26672071e-01] | [12.419756889343262, 1.1325336694717407] |
ec9769b5-904a-4e7b-924b-fea87bc0dd9b | efficient-keyword-spotting-using-dilated | 1811.07684 | null | http://arxiv.org/abs/1811.07684v2 | http://arxiv.org/pdf/1811.07684v2.pdf | Efficient keyword spotting using dilated convolutions and gating | We explore the application of end-to-end stateless temporal modeling to
small-footprint keyword spotting as opposed to recurrent networks that model
long-term temporal dependencies using internal states. We propose a model
inspired by the recent success of dilated convolutions in sequence modeling
applications, allowing to train deeper architectures in resource-constrained
configurations. Gated activations and residual connections are also added,
following a similar configuration to WaveNet. In addition, we apply a custom
target labeling that back-propagates loss from specific frames of interest,
therefore yielding higher accuracy and only requiring to detect the end of the
keyword. Our experimental results show that our model outperforms a max-pooling
loss trained recurrent neural network using LSTM cells, with a significant
decrease in false rejection rate. The underlying dataset - "Hey Snips"
utterances recorded by over 2.2K different speakers - has been made publicly
available to establish an open reference for wake-word detection. | ['Mathieu Poumeyrol', 'Thibault Gisselbrecht', 'David Leroy', 'Mohammed Chlieh', 'Alice Coucke', 'Thibaut Lavril'] | 2018-11-19 | null | null | null | null | ['small-footprint-keyword-spotting'] | ['speech'] | [ 3.82096380e-01 1.08683988e-01 -1.12372786e-01 -3.64412338e-01
-7.74502099e-01 -3.76573533e-01 5.99663854e-01 -1.87379792e-01
-9.25085783e-01 4.73633170e-01 4.71405119e-01 -3.12276989e-01
3.24922532e-01 -1.91265166e-01 -6.22068524e-01 -5.48937857e-01
-4.47353601e-01 -5.12068942e-02 1.79065436e-01 -8.20536315e-02
1.13386177e-01 4.83536422e-01 -1.32211721e+00 5.56739032e-01
5.38977832e-02 1.01384032e+00 2.71423101e-01 1.00013554e+00
6.78447410e-02 1.07880318e+00 -8.06456029e-01 -1.03783913e-01
-1.05057858e-01 -3.72289121e-01 -8.67835045e-01 -5.64835183e-02
3.10119689e-01 -3.19580495e-01 -7.30160475e-01 4.06608433e-01
9.29661453e-01 4.91280377e-01 1.61446892e-02 -5.25754929e-01
-3.33623528e-01 7.33040333e-01 1.21946288e-02 7.24715054e-01
3.84387732e-01 2.10506037e-01 9.50746596e-01 -9.62821901e-01
5.15428543e-01 1.13233340e+00 8.52365375e-01 8.11166942e-01
-1.29876387e+00 -7.11760104e-01 3.75116318e-01 4.06960070e-01
-1.42470586e+00 -1.03711641e+00 4.84662265e-01 -5.69946598e-03
1.92206895e+00 3.73681784e-01 5.01606226e-01 1.76900804e+00
3.74691039e-02 8.42274129e-01 5.71567237e-01 -5.79862237e-01
-7.04459026e-02 -4.46344130e-02 2.63997111e-02 5.64955771e-01
-7.68950462e-01 2.85326652e-02 -1.06174815e+00 -8.64742994e-02
4.54805017e-01 -1.09636337e-01 -2.97847748e-01 2.67846912e-01
-1.26094604e+00 6.12565935e-01 1.82441831e-01 4.89002883e-01
-3.18866283e-01 5.20130813e-01 6.95946336e-01 5.17055213e-01
7.60581017e-01 1.33070767e-01 -6.17689967e-01 -4.80237871e-01
-1.39023471e+00 -8.07330385e-02 7.31555223e-01 7.01354742e-01
3.88998598e-01 3.81444484e-01 -5.14556348e-01 1.05664921e+00
7.24783316e-02 1.42001361e-01 9.02972579e-01 -7.35179067e-01
1.90396041e-01 -1.09913714e-01 -6.91377223e-02 -5.10900378e-01
-5.46801269e-01 -6.88580930e-01 -5.06445944e-01 -2.40978763e-01
-9.60244760e-02 -2.52393067e-01 -1.33912086e+00 1.98053277e+00
-2.84057409e-01 6.17815077e-01 -8.35327432e-02 7.24734187e-01
6.20174468e-01 8.35757017e-01 2.42577530e-02 -3.12651664e-01
1.27922046e+00 -1.01858437e+00 -9.11383748e-01 -2.01859146e-01
7.00233459e-01 -8.69235933e-01 1.02466547e+00 4.03516322e-01
-1.02956915e+00 -4.02265102e-01 -7.79557586e-01 -1.98688805e-01
-3.60299766e-01 2.51714170e-01 2.54013628e-01 5.39446175e-01
-1.39366245e+00 6.52702689e-01 -8.74868214e-01 -5.38795531e-01
4.51173820e-02 4.01074678e-01 -8.24980736e-02 4.64469045e-01
-1.45686662e+00 9.93164062e-01 2.96139121e-01 3.55132073e-01
-1.20910621e+00 -6.24817073e-01 -7.43537605e-01 9.79776587e-03
2.90026605e-01 -2.09148645e-01 1.55793440e+00 -8.50357533e-01
-1.89123011e+00 7.65927017e-01 -6.12758875e-01 -9.91834760e-01
4.72074807e-01 -5.06612003e-01 -6.52329981e-01 1.07317410e-01
-3.11036825e-01 7.68497407e-01 8.89673591e-01 -4.76085007e-01
-2.60248989e-01 2.01937690e-01 -3.01343590e-01 1.18357301e-01
-6.18834674e-01 4.63056266e-01 -6.68906093e-01 -9.85584021e-01
-2.02899650e-01 -8.88466179e-01 -2.26080433e-01 -4.44699943e-01
-5.20767689e-01 -2.45313570e-01 1.05795097e+00 -9.74722207e-01
1.45446563e+00 -2.09225702e+00 -3.52962688e-02 -4.77054268e-02
-1.09511986e-01 3.75170350e-01 -3.99738163e-01 5.21113694e-01
-2.06703261e-01 1.72399685e-01 -8.48408267e-02 -9.16817546e-01
-2.18254477e-01 1.44827798e-01 -6.85563266e-01 3.92417938e-01
3.47105473e-01 8.58951628e-01 -6.26288235e-01 -1.95507735e-01
1.65238008e-01 8.01338434e-01 -2.79195368e-01 1.87173322e-01
-3.30641240e-01 2.63547927e-01 1.07901050e-02 3.61289233e-01
9.95702669e-02 -1.15829587e-01 9.88670513e-02 1.41383246e-01
-3.20386469e-01 1.03020358e+00 -6.28145099e-01 2.08363986e+00
-7.03472316e-01 1.03835666e+00 -2.13504061e-01 -8.67866695e-01
9.18910921e-01 8.60185206e-01 1.83286667e-01 -9.36672866e-01
2.29925606e-02 2.01024190e-02 -2.11935148e-01 -3.96687418e-01
7.33940959e-01 -1.16746306e-01 6.89748442e-03 4.24572200e-01
3.47535640e-01 3.65415603e-01 -1.87665801e-02 1.91040665e-01
1.29649353e+00 2.79138237e-01 -2.94554353e-01 6.39909208e-02
2.56069511e-01 -4.42133158e-01 4.25088108e-01 8.91750932e-01
-3.25875776e-03 8.06079268e-01 2.75568038e-01 -4.32060003e-01
-9.43151891e-01 -6.88612401e-01 2.99181882e-02 1.38040686e+00
-4.13730949e-01 -4.86687064e-01 -5.64928472e-01 -3.39667588e-01
-5.29275298e-01 7.86039889e-01 -6.31954372e-01 -2.24923715e-01
-9.05856907e-01 -6.28295064e-01 1.11515129e+00 3.19233596e-01
1.79184809e-01 -1.57731259e+00 -7.79174984e-01 5.37620127e-01
-3.78729641e-01 -1.23828602e+00 -6.10693514e-01 7.69560635e-01
-6.86796308e-01 -2.61387408e-01 -1.03179955e+00 -5.75935841e-01
4.53585610e-02 -1.00936644e-01 1.12996471e+00 -8.35742727e-02
-3.78005207e-01 2.34052166e-01 -3.90659332e-01 -1.10159509e-01
-2.41966695e-01 3.71736079e-01 1.24869771e-01 -6.04406111e-02
4.79227841e-01 -5.67077219e-01 -5.60174525e-01 8.63345563e-02
-6.93354189e-01 -9.94277745e-02 4.04340595e-01 6.86920643e-01
1.14438318e-01 -4.70464021e-01 5.31146049e-01 -3.65840495e-01
6.56113267e-01 -2.96368450e-01 -3.74361336e-01 9.29714739e-02
-4.56229746e-01 2.13772759e-01 3.79339427e-01 -6.97724760e-01
-9.49220538e-01 -4.00848836e-02 -3.51511776e-01 -7.70793319e-01
-1.10407867e-01 3.88384879e-01 4.01685178e-01 1.47863075e-01
4.28188741e-01 5.57994902e-01 -5.34213111e-02 -7.33332515e-01
2.45393470e-01 4.98417944e-01 6.19864702e-01 -1.02713659e-01
2.84169346e-01 2.58444279e-01 -4.52953577e-01 -9.44558740e-01
-6.82008386e-01 -5.91169775e-01 -2.04411447e-01 -1.30905151e-01
6.75015509e-01 -9.59934652e-01 -8.65390539e-01 4.99808520e-01
-1.36676192e+00 -5.72922587e-01 -1.88501969e-01 5.66996753e-01
-4.73942608e-01 3.24065000e-01 -1.03459179e+00 -1.03046238e+00
-5.47895551e-01 -8.23305488e-01 9.77449000e-01 -6.34614304e-02
-4.88259822e-01 -7.56948054e-01 1.68487042e-01 -1.16025560e-01
7.26274192e-01 -2.08707541e-01 4.79953229e-01 -8.35302770e-01
-3.19580555e-01 -1.06392547e-01 1.54741153e-01 4.56489414e-01
-2.21784025e-01 -1.70914635e-01 -1.52796304e+00 -3.17791849e-01
-1.66572239e-02 -3.28282624e-01 1.34753919e+00 4.82875824e-01
9.77038264e-01 -3.39897513e-01 -2.95421630e-01 4.66007382e-01
9.69157100e-01 1.30585521e-01 6.62900686e-01 2.92314410e-01
4.48550850e-01 5.84344923e-01 1.33089289e-01 5.77529550e-01
3.14076096e-02 8.87993991e-01 1.47328928e-01 -2.03408122e-01
-3.18110377e-01 -4.53007780e-02 7.74078846e-01 8.56229186e-01
1.85130060e-01 -5.15426040e-01 -9.08449888e-01 7.73664415e-01
-1.77537262e+00 -1.18614113e+00 2.25605130e-01 2.20512843e+00
9.53449965e-01 2.23629102e-01 2.17903823e-01 3.06700729e-02
6.12103343e-01 5.19100428e-01 -3.43528181e-01 -7.37273276e-01
-1.46436125e-01 4.59923834e-01 4.99125451e-01 6.51567519e-01
-1.09994316e+00 1.20805192e+00 6.84959745e+00 7.76403844e-01
-1.51344538e+00 3.87875497e-01 6.81962192e-01 -7.60735393e-01
-4.71655354e-02 -1.27501145e-01 -9.23952639e-01 4.43657845e-01
1.77524912e+00 2.69663036e-01 4.63858336e-01 3.63075674e-01
6.92649662e-01 1.32218301e-01 -9.04062629e-01 7.49049544e-01
9.04264450e-02 -1.32570052e+00 -2.30591059e-01 -1.11950420e-01
2.30892137e-01 5.57925999e-01 1.99277103e-01 3.80530983e-01
9.11526680e-02 -1.19227970e+00 1.03376210e+00 6.38326228e-01
8.81562769e-01 -7.72914648e-01 4.02236849e-01 2.92022854e-01
-1.19189906e+00 4.88822758e-02 -6.48696795e-02 -1.20790623e-01
3.41295928e-01 6.36689126e-01 -1.25397444e+00 1.36429489e-01
9.04785812e-01 5.89272439e-01 -2.99077302e-01 8.62487137e-01
1.22843273e-02 9.39242125e-01 -5.33835173e-01 -5.84940473e-03
4.17269439e-01 4.33085650e-01 5.51400840e-01 1.99454594e+00
2.21251383e-01 -3.89578462e-01 -8.25126246e-02 6.70327365e-01
-9.16254148e-02 -1.04726896e-01 -4.55024958e-01 -1.69303268e-01
3.47752035e-01 1.09264874e+00 -5.41481435e-01 -2.27243304e-01
-3.21572363e-01 1.47003233e+00 1.97771192e-01 7.26962864e-01
-8.61158729e-01 -4.02972192e-01 6.96360469e-01 -1.15567036e-01
6.94915593e-01 -2.76337206e-01 1.59204423e-01 -1.09292078e+00
4.82645892e-02 -5.88477790e-01 1.68707266e-01 -5.81085324e-01
-7.47067094e-01 9.54031587e-01 -3.01998913e-01 -9.49418068e-01
-7.19753027e-01 -3.11693996e-01 -5.33264816e-01 1.21534991e+00
-1.54361534e+00 -1.06169903e+00 2.10441977e-01 4.37544316e-01
7.39579737e-01 3.91547531e-02 1.10955727e+00 4.55852479e-01
-6.73369527e-01 7.46418476e-01 -1.27413079e-01 8.92168209e-02
6.72087908e-01 -8.64873052e-01 8.66407931e-01 1.09661603e+00
2.39976928e-01 7.11923659e-01 7.91919410e-01 -3.54244798e-01
-8.08695018e-01 -1.06454885e+00 1.30540323e+00 -3.73764187e-01
5.72376847e-01 -8.91014755e-01 -9.56968665e-01 8.53760064e-01
4.36221093e-01 1.13749810e-01 4.57610428e-01 1.67718366e-01
-4.33266699e-01 2.42170215e-01 -5.48166871e-01 5.54730058e-01
1.00865364e+00 -9.97602403e-01 -4.25712019e-01 2.31464431e-01
1.04317319e+00 -2.54381210e-01 -3.94561112e-01 4.00532991e-01
5.58248162e-01 -9.08149660e-01 9.71731782e-01 -4.60330486e-01
2.49510277e-02 1.22061064e-02 8.04751832e-03 -9.43950832e-01
-1.43487006e-01 -1.14618468e+00 -3.81365597e-01 1.08317041e+00
6.10607862e-01 -3.34622800e-01 8.32502961e-01 1.81460395e-01
-3.41781944e-01 -6.95419610e-01 -1.24320185e+00 -6.68304861e-01
-2.04593867e-01 -7.71576226e-01 7.26946592e-02 7.33867228e-01
-7.41779581e-02 3.01376849e-01 -8.68454278e-01 4.60571088e-02
4.14771549e-02 -3.24124813e-01 1.24080829e-01 -6.70350969e-01
-4.71146941e-01 -4.95256454e-01 -1.71651959e-01 -1.35259414e+00
3.73506576e-01 -6.93251431e-01 2.42601722e-01 -9.69924688e-01
-2.35632993e-02 -1.19703315e-01 -6.57867074e-01 7.71254897e-01
1.60701245e-01 5.39971113e-01 1.33674562e-01 1.20610707e-01
-5.94694078e-01 5.76868415e-01 4.82343376e-01 -2.60817576e-02
-2.62120456e-01 -5.06947711e-02 -4.00130488e-02 4.38891232e-01
6.91282392e-01 -5.36251187e-01 -2.09110826e-01 -3.17642152e-01
2.74189245e-02 -3.48753259e-02 4.53550875e-01 -9.84457135e-01
1.43825412e-01 3.06760907e-01 3.46683145e-01 -6.93240881e-01
7.63217270e-01 -4.09099221e-01 -6.86526299e-02 5.15318096e-01
-8.01589727e-01 -8.73412713e-02 5.26938975e-01 4.85914081e-01
-1.47265106e-01 -1.86763242e-01 4.74289984e-01 -1.36784002e-01
-8.61330867e-01 -4.65361364e-02 -8.76786590e-01 -2.90145904e-01
5.50938785e-01 -1.46309808e-01 -7.43590146e-02 -7.55049706e-01
-1.01081038e+00 -1.04800671e-01 -3.57957333e-02 6.36847794e-01
5.99185348e-01 -8.90638232e-01 -6.16605282e-01 1.38064057e-01
-6.82818191e-03 -5.93457460e-01 1.33532673e-01 8.58051062e-01
-5.82817160e-02 8.13606679e-01 3.48422348e-01 -5.06251633e-01
-1.30625558e+00 3.55619788e-01 5.60442328e-01 -2.43825763e-01
-9.09086645e-01 1.29705405e+00 -6.59394935e-02 -3.67222875e-01
7.19770730e-01 -4.72577155e-01 -6.38470277e-02 1.55177280e-01
6.27210855e-01 8.23110044e-02 4.82167095e-01 -4.74754333e-01
-5.68804622e-01 -1.58278104e-02 -3.57866824e-01 -5.48625052e-01
1.19876850e+00 -2.64988869e-01 2.10381910e-01 7.37852633e-01
1.36490464e+00 -1.45595632e-02 -1.26576364e+00 -5.16626775e-01
2.27777183e-01 9.16548520e-02 2.49419093e-01 -9.77919698e-01
-7.34751105e-01 9.82509553e-01 8.87760758e-01 1.54375330e-01
1.04901743e+00 -5.78580201e-02 9.02958393e-01 4.66467023e-01
-5.35256304e-02 -9.71410215e-01 1.76640853e-01 9.22786057e-01
7.62808919e-01 -8.48506033e-01 -4.23912942e-01 2.70914435e-01
-4.20178771e-01 1.11816180e+00 1.88228667e-01 1.11945257e-01
3.44133258e-01 5.40386677e-01 2.47790381e-01 4.84157279e-02
-1.23418653e+00 -1.67517155e-01 1.05424762e-01 2.07069278e-01
7.96752751e-01 -2.13912010e-01 3.57443355e-02 1.32642224e-01
-4.31135818e-02 9.04642567e-02 3.48141044e-01 8.38366568e-01
-1.95875183e-01 -1.08411396e+00 -1.61856443e-01 2.35744148e-01
-8.68207574e-01 -7.52084553e-01 -2.33971745e-01 2.89120525e-01
-2.98893064e-01 9.09208119e-01 3.61245126e-01 -4.46631134e-01
-8.03539122e-04 5.69247305e-01 1.47112682e-01 -6.64508879e-01
-8.44128251e-01 4.33570892e-01 3.45176041e-01 -6.96348965e-01
-4.04669821e-01 -6.04618609e-01 -1.20728481e+00 -1.00895874e-01
-3.29938889e-01 -1.62160117e-02 6.91183865e-01 9.57849920e-01
5.28230250e-01 7.96053112e-01 3.77835453e-01 -1.04233456e+00
-3.41080189e-01 -1.36887169e+00 -1.88937023e-01 2.28009429e-02
8.20955336e-01 -1.92002282e-01 -3.73400331e-01 1.91814139e-01] | [14.319982528686523, 6.227598667144775] |
fa56a2e3-3277-4499-91ed-6540adbf5f14 | effective-data-augmentation-for-sentence | null | null | https://aclanthology.org/2022.coling-1.305 | https://aclanthology.org/2022.coling-1.305.pdf | Effective Data Augmentation for Sentence Classification Using One VAE per Class | In recent years, data augmentation has become an important field of machine learning. While images can use simple techniques such as cropping or rotating, textual data augmentation needs more complex manipulations to ensure that the generated examples are useful. Variational auto-encoders (VAE) and its conditional variant the Conditional-VAE (CVAE) are often used to generate new textual data, both relying on a good enough training of the generator so that it doesn’t create examples of the wrong class. In this paper, we explore a simpler way to use VAE for data augmentation: the training of one VAE per class. We show on several dataset sizes, as well as on four different binary classification tasks, that it systematically outperforms other generative data augmentation techniques. | ['Philippe Langlais', 'Frédéric Piedboeuf'] | null | null | null | null | coling-2022-10 | ['sentence-classification'] | ['natural-language-processing'] | [ 5.59202254e-01 4.29538906e-01 -1.74653515e-01 -1.94133967e-01
-3.69213372e-01 -5.35222888e-01 1.14709973e+00 1.11193106e-01
-4.72859204e-01 1.07350647e+00 -3.38637270e-02 -4.56935406e-01
4.71919745e-01 -9.53235388e-01 -7.99913406e-01 -7.43817925e-01
2.94811726e-01 7.21475422e-01 5.37199751e-02 -2.90277779e-01
1.81783754e-02 4.61290509e-01 -1.61438024e+00 1.30810574e-01
8.03534210e-01 5.63504100e-01 4.18864045e-04 6.78641796e-01
-3.23049277e-01 6.07097805e-01 -1.07291043e+00 -5.46047628e-01
2.48234674e-01 -7.11649776e-01 -6.24896467e-01 3.98766190e-01
1.59091532e-01 -2.20384285e-01 -4.93139364e-02 7.15947866e-01
1.90789938e-01 6.69007748e-02 1.05000794e+00 -1.60835159e+00
-8.36896181e-01 6.59055531e-01 -5.75189829e-01 -2.54599191e-02
2.66445875e-02 7.73962587e-02 6.79662943e-01 -9.28468168e-01
8.53530288e-01 1.01175797e+00 5.27728319e-01 9.08462524e-01
-1.57348382e+00 -6.36556745e-01 -2.81358123e-01 -6.13992102e-02
-1.02199495e+00 -2.56398857e-01 8.36002946e-01 -4.57621843e-01
4.82204050e-01 3.65006387e-01 7.44673610e-01 1.49460542e+00
-2.32302129e-01 8.54119778e-01 1.31250787e+00 -8.39036345e-01
3.41656446e-01 5.32108665e-01 -2.59484440e-01 3.76738399e-01
4.23668712e-01 4.11941744e-02 -1.20329164e-01 -1.32339686e-01
9.21438456e-01 -8.99324119e-02 -1.39170229e-01 -5.76674521e-01
-1.16026807e+00 1.31794524e+00 3.21020901e-01 4.81652111e-01
-4.67518419e-01 2.57907778e-01 3.04467589e-01 1.81535512e-01
5.50554037e-01 7.02784002e-01 -2.74949074e-01 1.61254574e-02
-8.88940036e-01 3.70361596e-01 3.67033631e-01 8.56350541e-01
6.53817892e-01 4.33665067e-01 -3.92079234e-01 9.01260495e-01
1.11344144e-01 3.37626189e-01 8.07229459e-01 -5.63227713e-01
2.59261787e-01 5.65171778e-01 3.06856874e-02 -3.41959685e-01
7.04029948e-02 -4.20281410e-01 -1.15538371e+00 6.88792944e-01
4.46720034e-01 -2.52731115e-01 -1.54772389e+00 1.88616264e+00
2.88120538e-01 -8.08571205e-02 7.38344863e-02 3.30512464e-01
6.27349079e-01 7.93243289e-01 1.83277786e-01 -2.19517872e-01
9.98244941e-01 -6.88143313e-01 -1.08097506e+00 -2.63862133e-01
7.22820282e-01 -7.04344988e-01 1.15071201e+00 5.26766598e-01
-1.04050303e+00 -7.05442131e-01 -1.09256244e+00 -6.60297275e-02
-7.19096899e-01 2.72471696e-01 7.60524154e-01 1.06858671e+00
-8.22284162e-01 5.00784457e-01 -7.18023837e-01 -4.53680977e-02
7.32496500e-01 1.88433409e-01 -5.46570480e-01 7.12859482e-02
-1.12179089e+00 9.95825291e-01 4.14559424e-01 -1.99054077e-01
-9.56647515e-01 -6.50721848e-01 -1.14094579e+00 -1.29416347e-01
1.43455327e-01 -6.28979564e-01 1.13748693e+00 -9.91234779e-01
-1.38179147e+00 7.87862539e-01 -5.97771741e-02 -6.95965350e-01
7.63466716e-01 -2.32651412e-01 -5.75937331e-02 -2.31367171e-01
-1.37351185e-01 1.08371174e+00 1.25573993e+00 -1.51799166e+00
-2.21324652e-01 -7.62342364e-02 -2.77252376e-01 -1.75979689e-01
-4.45642173e-01 -3.21748167e-01 -7.96534959e-03 -1.04399359e+00
-2.21594617e-01 -9.31598961e-01 -3.91202956e-01 -1.53825596e-01
-6.49177492e-01 -2.87479579e-01 1.27673423e+00 -6.09976828e-01
9.42607284e-01 -2.03795385e+00 2.39723906e-01 1.29816070e-01
1.64102197e-01 5.95839858e-01 1.60493627e-02 1.99337766e-01
-5.81837654e-01 5.66439390e-01 -6.01329386e-01 -7.42519438e-01
-1.80221885e-01 5.48632026e-01 -4.62148041e-01 1.16423026e-01
7.47091472e-01 1.14101875e+00 -6.54220700e-01 -4.69915748e-01
4.35919166e-01 6.49986506e-01 -6.01791561e-01 3.80100086e-02
-4.36194122e-01 5.09910643e-01 7.60232359e-02 1.11791022e-01
5.09346485e-01 -1.96420938e-01 -1.10540040e-01 1.14632934e-01
1.52867034e-01 -2.34074499e-02 -1.14731860e+00 1.34698653e+00
-4.85057443e-01 1.01414537e+00 -4.75776762e-01 -9.97147381e-01
9.91747677e-01 4.02066886e-01 1.06424019e-01 -2.84448206e-01
1.97105914e-01 1.26245514e-01 1.47256911e-01 -3.12399834e-01
6.35797560e-01 -3.38141561e-01 4.52474132e-02 4.82101738e-01
1.66128203e-01 -4.87873793e-01 5.18137932e-01 4.01587307e-01
6.91229284e-01 2.55178154e-01 3.24591607e-01 1.03882745e-01
3.19863915e-01 5.43866456e-02 1.54146150e-01 7.15126514e-01
4.06713039e-01 7.89062858e-01 5.49656212e-01 -1.96262807e-01
-1.81646693e+00 -7.48500645e-01 -2.54974723e-01 5.36372185e-01
-5.31707704e-01 -2.83447653e-01 -9.22740459e-01 -8.15326929e-01
-3.26608360e-01 1.17845607e+00 -9.92980957e-01 -2.61079729e-01
-3.78178477e-01 -1.00175107e+00 3.45672995e-01 6.00602865e-01
4.97209132e-01 -1.37447178e+00 -4.71212596e-01 -2.24910141e-03
-1.05336979e-01 -9.41170096e-01 1.02513994e-03 2.74668813e-01
-9.46259916e-01 -6.61621094e-01 -1.02729797e+00 -5.78182161e-01
9.36502695e-01 -1.26351580e-01 9.91888285e-01 1.67616710e-01
-2.93491989e-01 7.64052421e-02 -7.03053474e-01 -7.86420524e-01
-1.11713600e+00 1.09622903e-01 -8.23394954e-02 8.74264240e-02
4.21760157e-02 -5.86013973e-01 -6.99341074e-02 -7.42827579e-02
-1.28213656e+00 2.84262896e-01 7.89224982e-01 9.56548572e-01
5.55894315e-01 -5.37670627e-02 5.06925523e-01 -1.19803989e+00
7.28424549e-01 -2.52487391e-01 -4.76050287e-01 -1.62293818e-02
-5.27385890e-01 2.48765185e-01 6.73947275e-01 -7.31628299e-01
-8.16914916e-01 1.11587353e-01 -3.23488563e-01 -6.50457263e-01
-3.05844396e-01 3.15536261e-01 4.99457642e-02 2.54330963e-01
1.00053358e+00 3.37009192e-01 1.91623464e-01 -4.14247751e-01
5.43925047e-01 5.51430464e-01 4.32071865e-01 -1.63831040e-01
1.17016482e+00 4.03252035e-01 1.40685737e-01 -9.91778195e-01
-4.96210635e-01 2.58093268e-01 -6.93143427e-01 -1.01401284e-01
1.02067053e+00 -4.50784981e-01 -1.15548126e-01 3.62904370e-01
-1.15702486e+00 -4.63461369e-01 -8.92651498e-01 2.62366027e-01
-3.29983979e-01 1.22982427e-01 -2.69523114e-01 -9.94639456e-01
-1.17296048e-01 -1.14334619e+00 8.92746985e-01 1.44012749e-01
-2.85740882e-01 -8.02480757e-01 4.61386964e-02 1.03147663e-01
3.39795142e-01 5.98608792e-01 9.65334237e-01 -8.56325746e-01
-3.46053511e-01 -5.43811321e-01 1.79905534e-01 6.71293914e-01
2.34725356e-01 4.18636799e-01 -1.04596424e+00 -1.35769367e-01
-1.46767646e-01 -3.92593443e-01 8.87628675e-01 3.60860676e-01
1.42059326e+00 -2.53945231e-01 -3.65845799e-01 2.37179980e-01
1.11033559e+00 3.17362130e-01 1.22713029e+00 1.07249826e-01
6.70169532e-01 4.61045355e-01 4.43906933e-01 1.73954144e-01
-2.00380102e-01 6.18052781e-01 4.50520635e-01 -3.94561410e-01
-2.99325675e-01 -3.15273494e-01 5.70139848e-03 3.60718489e-01
-2.25569338e-01 -4.15461510e-01 -6.17640674e-01 5.21819293e-01
-1.33396780e+00 -8.57779682e-01 -5.01546264e-01 2.28082299e+00
8.93265784e-01 1.98810250e-01 1.24953948e-01 7.99696982e-01
6.06266677e-01 -1.59518290e-02 -1.98950961e-01 -4.16135758e-01
-2.02538118e-01 7.88254082e-01 2.27269545e-01 4.17746395e-01
-1.21346843e+00 7.95824051e-01 6.88514709e+00 7.43258119e-01
-9.47089076e-01 1.70847103e-01 8.88517916e-01 1.98111445e-01
-4.72076416e-01 -1.68551385e-01 -5.71954191e-01 5.01332819e-01
7.37378418e-01 2.60966718e-01 6.49260581e-02 9.70811725e-01
-1.59727678e-01 -2.45418578e-01 -9.92186725e-01 8.93395305e-01
1.07265497e-02 -1.44747472e+00 2.85133839e-01 2.89818823e-01
9.05916512e-01 -5.21240532e-01 2.38109559e-01 4.62258190e-01
4.62656319e-01 -1.25566471e+00 5.32926261e-01 2.54500657e-01
9.83711600e-01 -6.63294554e-01 8.68983626e-01 3.23740691e-01
-4.90034014e-01 1.74679697e-01 -1.62317693e-01 1.12731881e-01
1.95316121e-01 6.97287738e-01 -1.03101230e+00 3.15463185e-01
3.71369004e-01 2.51574159e-01 -8.39649975e-01 7.70922065e-01
-4.76984203e-01 6.68133140e-01 -2.42431983e-01 -2.38244254e-02
2.15242598e-02 -2.14164257e-01 4.45183456e-01 9.34102476e-01
3.57362658e-01 -1.17287979e-01 -3.52358609e-01 9.65595841e-01
-1.98924363e-01 -1.52703170e-02 -8.84740114e-01 -2.97749311e-01
1.66666076e-01 1.26581573e+00 -7.08295047e-01 -6.35702312e-01
-2.29934171e-01 1.12621033e+00 2.53183078e-02 2.11283818e-01
-7.66693473e-01 -3.37595671e-01 2.72719026e-01 7.35436454e-02
3.91399711e-01 -2.14486018e-01 -3.82062435e-01 -7.97960818e-01
-8.42299983e-02 -8.69258523e-01 9.79045928e-02 -1.11053431e+00
-1.00546384e+00 7.58203924e-01 1.22207440e-01 -1.18181503e+00
-5.89332759e-01 -6.18400633e-01 -5.08839250e-01 8.77593815e-01
-1.15377522e+00 -1.02215719e+00 -4.35441166e-01 4.49852526e-01
4.92278755e-01 -2.24141866e-01 9.58115339e-01 2.74306595e-01
-4.50851142e-01 5.60114741e-01 -2.94950783e-01 3.04300010e-01
4.39687848e-01 -1.39308357e+00 5.63389540e-01 9.25580680e-01
5.74356377e-01 3.81745607e-01 1.11655855e+00 -6.76733315e-01
-6.73546493e-01 -8.20906579e-01 8.39692473e-01 -6.22482479e-01
2.37355173e-01 -5.85131049e-01 -9.87873554e-01 9.16079879e-01
4.45320219e-01 -8.68333578e-02 5.27333796e-01 -2.58669913e-01
-1.32975787e-01 3.03754956e-01 -1.16199446e+00 6.62765622e-01
6.62612140e-01 -1.00470535e-01 -4.69710916e-01 4.16260153e-01
6.89048946e-01 -3.10137868e-01 -5.38982332e-01 4.50227618e-01
7.33577162e-02 -8.53391647e-01 8.54414999e-01 -7.95009255e-01
8.81805837e-01 -1.63506657e-01 1.69622213e-01 -1.68878567e+00
9.96007118e-03 -4.92632091e-01 -7.97840357e-02 1.30217731e+00
6.11801982e-01 -5.86908162e-01 9.44258869e-01 3.56330395e-01
-1.93220582e-02 -6.35523677e-01 -7.02337801e-01 -6.84722126e-01
1.60251826e-01 -4.51038450e-01 4.31118041e-01 1.17038548e+00
-4.51089442e-01 5.19913495e-01 -6.39489770e-01 -2.73653179e-01
4.12987441e-01 -3.51464421e-01 1.16118896e+00 -1.49816060e+00
-1.01948246e-01 -4.07652825e-01 -4.31441069e-01 -6.19800806e-01
-1.17321111e-01 -8.04877281e-01 -1.27875984e-01 -1.57032824e+00
4.89810258e-02 -5.38042843e-01 3.38396460e-01 6.62987292e-01
-2.01435760e-01 6.05856895e-01 1.65745616e-02 -8.08274299e-02
1.84607580e-01 7.41092503e-01 1.39521980e+00 -1.97393596e-02
-2.83193231e-01 2.87439656e-02 -4.45127547e-01 4.78773415e-01
8.38483930e-01 -5.49228132e-01 -4.97293800e-01 2.92594135e-02
3.58971149e-01 -3.54795069e-01 4.93271470e-01 -9.30780053e-01
-3.58574778e-01 1.95829317e-01 7.98143923e-01 -6.24213874e-01
5.42635739e-01 -8.13869119e-01 2.06193209e-01 5.87703466e-01
-3.71586561e-01 5.12359440e-02 3.68480712e-01 3.28022093e-01
-1.43028632e-01 -7.46699035e-01 8.45907092e-01 -1.00859165e-01
-2.85214603e-01 1.66409656e-01 -5.24285436e-01 -1.50474697e-01
1.17883730e+00 -1.70999184e-01 -9.94218588e-02 -6.32524431e-01
-8.66928399e-01 -2.09960982e-01 3.97756457e-01 3.89226645e-01
4.88827467e-01 -1.42886198e+00 -7.58959532e-01 5.15213430e-01
-7.19864741e-02 1.12168707e-01 4.73481044e-03 6.47482276e-01
-3.39674175e-01 2.28929982e-01 -2.24530786e-01 -5.11926293e-01
-1.30490100e+00 8.45718682e-01 1.96629822e-01 -2.83079714e-01
-5.29804826e-01 7.88177371e-01 8.71146843e-02 -3.25847954e-01
-7.45832026e-02 -2.62742549e-01 -3.88240129e-01 2.37826526e-01
4.80466843e-01 -1.83018167e-02 4.77571785e-02 -3.58427465e-01
1.73736081e-01 -2.84304097e-02 -9.66952443e-02 -5.05526304e-01
1.35114479e+00 3.75819743e-01 1.61818676e-02 5.58108389e-01
8.65148723e-01 1.05177946e-02 -1.06147552e+00 -7.89964497e-02
-3.02571356e-01 -5.43593407e-01 -9.28124487e-02 -5.91622472e-01
-1.02289522e+00 1.06211770e+00 6.02568746e-01 7.57075667e-01
9.53827918e-01 3.06941010e-02 2.05893040e-01 -2.46883668e-02
5.78080453e-02 -8.84245217e-01 3.05591673e-01 1.12071492e-01
1.07556176e+00 -1.24466181e+00 1.05542317e-01 -4.74601835e-01
-9.31074858e-01 8.55681598e-01 5.34938574e-01 -2.23709028e-02
4.38103765e-01 9.03006196e-02 -1.57099009e-01 -4.69265180e-03
-5.34916580e-01 -2.33283222e-01 2.53668249e-01 8.15927625e-01
4.35430378e-01 -1.19452894e-01 -2.62783974e-01 9.36634615e-02
-5.44206619e-01 -6.47807121e-02 7.38816679e-01 9.00518775e-01
-4.97085117e-02 -1.51572740e+00 -5.28881311e-01 7.56248951e-01
-3.44048858e-01 -1.41209558e-01 -5.93830168e-01 1.32430816e+00
1.05014041e-01 6.24114871e-01 3.83978277e-01 -2.65452296e-01
-4.84719127e-02 5.45432389e-01 6.18324101e-01 -6.19709671e-01
-3.78991216e-01 -1.88954398e-02 2.97881272e-02 4.04863283e-02
-5.97073019e-01 -7.48349130e-01 -9.23814952e-01 -8.21856633e-02
-5.05656123e-01 7.35901371e-02 9.34675038e-01 9.52325881e-01
1.23120695e-02 8.23038936e-01 4.84717935e-01 -8.99324119e-01
-2.61216462e-01 -1.46136236e+00 -3.58330727e-01 6.69562817e-01
2.12268859e-01 -8.17549109e-01 -4.24353212e-01 2.65519947e-01] | [11.60489559173584, -0.18049122393131256] |
62e8483f-7976-473a-899f-198e9f55817f | a-review-of-benchmarks-for-visual-defect | 2305.13261 | null | https://arxiv.org/abs/2305.13261v1 | https://arxiv.org/pdf/2305.13261v1.pdf | A Review of Benchmarks for Visual Defect Detection in the Manufacturing Industry | The field of industrial defect detection using machine learning and deep learning is a subject of active research. Datasets, also called benchmarks, are used to compare and assess research results. There is a number of datasets in industrial visual inspection, of varying quality. Thus, it is a difficult task to determine which dataset to use. Generally speaking, datasets which include a testing set, with precise labeling and made in real-world conditions should be preferred. We propose a study of existing benchmarks to compare and expose their characteristics and their use-cases. A study of industrial metrics requirements, as well as testing procedures, will be presented and applied to the studied benchmarks. We discuss our findings by examining the current state of benchmarks for industrial visual inspection, and by exposing guidelines on the usage of benchmarks. | ['Yves GRANDVALET', 'Alexandre Durupt', 'Philippe Carvalho'] | 2023-05-05 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 2.17207119e-01 -2.98412174e-01 -1.04402892e-01 -6.79408073e-01
-2.10286111e-01 -3.28033775e-01 3.42065096e-02 5.14708698e-01
1.71058446e-01 5.03598928e-01 -4.97230530e-01 -5.03253520e-01
-3.51453602e-01 -1.01745844e+00 -4.89601463e-01 -4.03848767e-01
-3.58085968e-02 2.77370274e-01 3.26349400e-02 -1.19454719e-01
4.70641613e-01 2.86088556e-01 -1.95665765e+00 7.38689721e-01
6.21842504e-01 1.32308042e+00 1.63856670e-01 3.98591757e-01
-2.16004457e-02 8.72569263e-01 -1.41091454e+00 -3.11453372e-01
3.92422467e-01 -3.00740451e-01 -9.80162621e-01 6.07493043e-01
3.86106223e-01 -2.05956861e-01 4.51978594e-01 1.11952746e+00
2.75047600e-01 -3.18739712e-01 5.24257243e-01 -1.57199991e+00
-8.65116179e-01 5.74406028e-01 -4.48767483e-01 2.60394644e-02
3.56650919e-01 1.56901702e-01 9.63041246e-01 -3.86659503e-01
4.04852808e-01 9.82301652e-01 4.91778344e-01 1.99126542e-01
-9.58553135e-01 -5.75889885e-01 2.94692051e-02 4.05147225e-01
-9.47809279e-01 2.72442698e-01 1.04776490e+00 -8.94465566e-01
9.80200231e-01 2.83913493e-01 5.55819869e-01 1.07530785e+00
5.57034552e-01 4.51103657e-01 1.21059000e+00 -7.67600000e-01
3.59989882e-01 3.82604510e-01 5.73375523e-01 7.07690954e-01
6.37708485e-01 1.28837779e-01 -7.39040747e-02 4.37335670e-02
5.77980518e-01 2.78129280e-02 2.64144149e-02 -6.25216186e-01
-8.99614453e-01 9.18732941e-01 3.82642224e-02 5.18653214e-01
-2.39787936e-01 -1.12270661e-01 9.65276659e-01 8.60559285e-01
1.47571489e-01 6.95067048e-01 -8.99596095e-01 -8.77903253e-02
-5.03891408e-01 1.46958813e-01 8.46696794e-01 1.03076088e+00
9.04981554e-01 1.27700970e-01 -3.07864286e-02 7.94701040e-01
4.00228918e-01 -8.76582190e-02 3.71779799e-01 -9.76828873e-01
3.61730844e-01 1.03359020e+00 5.95635958e-02 -1.12467635e+00
-2.47329414e-01 -2.19251350e-01 -7.06803560e-01 9.90972936e-01
2.14226156e-01 -1.47368863e-01 -5.94692349e-01 9.31067228e-01
-1.65594518e-01 -3.39506686e-01 -1.41758064e-03 5.62750697e-01
8.35102439e-01 3.52095664e-01 -1.75971210e-01 -6.93953484e-02
1.14692068e+00 -6.42154753e-01 -6.20146632e-01 -9.96880904e-02
8.36641729e-01 -7.24695504e-01 1.39930224e+00 1.09927201e+00
-7.86930621e-01 -1.06743550e+00 -1.44479215e+00 3.26663643e-01
-7.20618784e-01 3.98256302e-01 5.16277552e-01 9.91617858e-01
-7.12722301e-01 7.79877901e-01 -7.55294025e-01 -2.63368368e-01
2.94900030e-01 2.23306865e-01 -2.18487352e-01 -5.82716018e-02
-9.53121126e-01 8.37752223e-01 4.41960245e-01 1.19155966e-01
-1.26984048e+00 -4.57987010e-01 -8.84025931e-01 -1.14260584e-01
2.83348560e-01 -1.09675646e-01 1.40272224e+00 -8.68089736e-01
-1.01770318e+00 7.36675203e-01 7.88684070e-01 -3.81048083e-01
3.89498979e-01 -3.22768390e-01 -5.92602372e-01 -3.91999990e-01
6.79787025e-02 -3.33746523e-02 5.63316464e-01 -1.57523513e+00
-8.26361299e-01 -1.85923412e-01 7.71235704e-01 -6.99759722e-01
-4.26270306e-01 2.91206688e-01 9.36590880e-02 -2.44398281e-01
-3.80107939e-01 -4.53509301e-01 -1.46721244e-01 -3.36573333e-01
-5.57242811e-01 -3.60217422e-01 9.77384746e-01 -5.32098591e-01
1.23273885e+00 -1.99055123e+00 -4.41627890e-01 9.24750194e-02
2.79872328e-01 1.44093335e-01 -7.98177645e-02 4.85072643e-01
-3.84313911e-01 2.81510532e-01 -2.78169930e-01 1.30138127e-02
1.86208561e-01 6.58214241e-02 1.31133139e-01 4.62685078e-01
3.98212463e-01 2.26820946e-01 -8.32305491e-01 -5.28714240e-01
5.92434824e-01 -7.91379809e-02 -2.50847101e-01 3.31385225e-01
-2.98308790e-01 2.39210621e-01 -5.22868276e-01 1.07088339e+00
6.00076973e-01 -1.51498497e-01 1.57200962e-01 -5.42173743e-01
-1.40076756e-01 9.37682092e-02 -1.23491573e+00 1.40673637e+00
-8.10989976e-01 6.51039541e-01 -2.44081333e-01 -1.44921184e+00
1.39632940e+00 4.06913280e-01 5.28597355e-01 -7.27366388e-01
4.65743542e-01 1.34624839e-01 2.39455551e-01 -1.12912726e+00
3.33263502e-02 2.84368396e-01 -1.75111204e-01 5.12322545e-01
7.12929368e-02 -1.09494619e-01 9.94775295e-01 -5.47207713e-01
1.46550846e+00 1.11307085e-01 2.57225484e-01 -1.74170449e-01
6.74878299e-01 7.28055835e-02 6.96369648e-01 4.05109137e-01
-1.25534594e-01 6.19020402e-01 9.41892982e-01 -9.69778657e-01
-1.15843546e+00 -7.88894951e-01 -1.91077694e-01 6.78579986e-01
-8.12503919e-02 -2.91658938e-01 -7.02717364e-01 -1.20911431e+00
7.72733986e-02 6.52869880e-01 -7.07970083e-01 -2.56304443e-01
-4.88881975e-01 -7.04069376e-01 9.43458006e-02 6.02965355e-01
4.40462917e-01 -1.63673234e+00 -1.38096392e+00 9.92707685e-02
4.19119000e-01 -9.16017652e-01 4.44273055e-01 4.71645296e-01
-8.80297244e-01 -1.73903215e+00 -9.55693722e-02 -8.32294822e-01
6.99629486e-01 -1.31919101e-01 1.90812397e+00 4.11825627e-01
-6.15852594e-01 1.28115550e-01 -8.89794827e-01 -7.50003874e-01
-9.48233843e-01 -1.48990443e-02 -3.97642493e-01 -3.48978758e-01
5.28693020e-01 -3.02734047e-01 -1.59871832e-01 4.11380261e-01
-1.01638806e+00 -4.35852528e-01 9.35311854e-01 8.19288790e-01
3.02282423e-01 4.36859071e-01 5.63482702e-01 -1.05039656e+00
9.55991983e-01 -3.26757342e-01 -9.22587514e-01 5.09239256e-01
-8.99309695e-01 -1.61587298e-01 5.84285975e-01 -6.53753951e-02
-7.92099297e-01 -2.93832630e-01 -9.15981010e-02 -3.26672912e-01
-7.12188482e-01 6.62222028e-01 -4.64115560e-01 1.67992208e-02
7.72079468e-01 -3.62778097e-01 -1.63801879e-01 -6.28633559e-01
-1.89957932e-01 9.16101635e-01 4.40867245e-02 -8.17691803e-01
5.16640604e-01 -1.57207195e-02 -2.63622552e-01 -6.01166785e-01
-5.74303508e-01 -1.82948753e-01 -8.01005840e-01 -4.22752857e-01
8.41892123e-01 -1.52961597e-01 -5.70079029e-01 3.18242371e-01
-1.28553486e+00 -2.65839249e-01 -5.48182189e-01 3.75892371e-01
-6.31488383e-01 2.66842991e-01 -6.14410222e-01 -6.39342189e-01
-1.05394579e-01 -1.67262065e+00 9.36557233e-01 -1.98005110e-01
-3.50100875e-01 -9.29494977e-01 8.34817663e-02 4.48782951e-01
1.60377920e-01 7.72413552e-01 1.39714599e+00 -6.93742216e-01
-3.33585262e-01 -3.12951267e-01 -9.92751792e-02 1.24574411e+00
6.39143229e-01 4.88981634e-01 -8.75975311e-01 -3.00434500e-01
1.74753636e-01 -5.23999512e-01 2.77319163e-01 2.48319790e-01
1.55948591e+00 1.21689282e-01 -1.09824635e-01 6.61841184e-02
1.51505327e+00 7.91501760e-01 6.47825539e-01 7.46161997e-01
4.05779928e-01 8.41992795e-01 1.21663177e+00 6.26074672e-01
-1.32159039e-01 3.72598678e-01 1.09426546e+00 -2.48202845e-01
1.10221937e-01 4.82477218e-01 1.07268021e-01 9.37219679e-01
-1.94089323e-01 -2.01280639e-01 -1.07143807e+00 4.12559301e-01
-1.41283512e+00 -4.63624030e-01 -1.63539425e-01 2.00455165e+00
5.39356709e-01 4.68947291e-01 2.51889434e-02 1.02563286e+00
7.12818027e-01 -1.45654589e-01 -2.60803193e-01 -6.97830975e-01
4.08822417e-01 3.26482058e-01 2.61479095e-02 -1.87951609e-01
-1.24373758e+00 2.40959048e-01 6.84406233e+00 1.18572563e-01
-9.86779869e-01 1.89542875e-01 5.56876719e-01 2.48575956e-01
1.00855336e-01 -2.00551614e-01 -4.01100874e-01 3.82753640e-01
8.13024402e-01 5.26693538e-02 -2.10387670e-02 1.13327217e+00
6.62968829e-02 2.34050721e-01 -1.52585912e+00 6.87181234e-01
1.27679378e-01 -1.10046399e+00 -1.82758838e-01 -1.34646088e-01
7.88476765e-01 -3.67767304e-01 -3.03048063e-02 3.40681553e-01
5.09702086e-01 -9.09144402e-01 7.49091864e-01 1.24703243e-01
4.75506634e-01 -8.76075327e-01 1.43889034e+00 -9.96265560e-02
-7.73159027e-01 -6.09413505e-01 -5.21010876e-01 -1.24556668e-01
-2.24633917e-01 8.34942102e-01 -5.95705628e-01 9.21468735e-01
1.31066084e+00 8.48027229e-01 -8.03333223e-01 1.12479115e+00
-1.69636667e-01 5.06562531e-01 3.98810416e-01 -1.09256335e-01
-2.17007566e-02 -1.24196924e-01 -1.34647727e-01 1.17236364e+00
4.23817813e-01 -9.73237813e-01 4.42041487e-01 1.20924926e+00
1.55020982e-01 1.70609772e-01 -7.65440941e-01 3.74298058e-02
6.06662259e-02 1.27348804e+00 -6.94429994e-01 -8.53704363e-02
-8.76113594e-01 2.74208546e-01 -1.06172509e-01 8.61322209e-02
-9.82763231e-01 -6.63387060e-01 7.40385532e-01 9.64302346e-02
9.69689712e-03 -1.11662686e-01 -3.72111678e-01 -3.64584625e-01
2.79202878e-01 -1.34767497e+00 3.51722479e-01 -7.22213924e-01
-1.45731044e+00 7.97493875e-01 3.36403280e-01 -1.78534520e+00
2.26978231e-02 -1.11464965e+00 -5.65201998e-01 3.52240384e-01
-1.29743707e+00 -9.23775613e-01 -8.18628013e-01 9.23725665e-02
8.35241318e-01 -5.48565149e-01 5.94988942e-01 6.54179931e-01
-6.09627664e-01 2.75430053e-01 -2.30341583e-01 1.95944309e-01
4.51268286e-01 -1.39190710e+00 2.49399662e-01 5.47937810e-01
9.77425054e-02 3.92917663e-01 7.90589511e-01 -5.35804331e-01
-1.10264730e+00 -1.14175642e+00 1.22886121e-01 -3.25444490e-01
5.24015903e-01 -1.93518430e-01 -9.02623355e-01 7.03671217e-01
5.04293084e-01 -2.61415653e-02 5.91311455e-01 9.82968807e-02
1.32342065e-02 -5.76599598e-01 -1.21831357e+00 4.87010181e-02
6.64485812e-01 -1.81403503e-01 -5.14495075e-01 4.52539533e-01
5.12969971e-01 -1.11262634e-01 -1.19798779e+00 7.44599998e-01
2.92760730e-01 -1.44694698e+00 5.85917294e-01 -4.93919373e-01
6.10587656e-01 -3.83315563e-01 2.79296637e-02 -1.39842618e+00
-1.99139208e-01 2.92174648e-02 2.66495615e-01 1.49473119e+00
4.80670184e-01 -3.93322438e-01 7.52372384e-01 6.12487048e-02
-4.03932959e-01 -7.04688191e-01 -3.45973223e-01 -9.23247039e-01
-1.44167706e-01 -5.51683068e-01 9.08543944e-01 9.61332560e-01
-1.99730515e-01 -4.78364388e-03 1.01553418e-01 -4.43625003e-02
2.57174999e-01 2.24125609e-01 7.61254072e-01 -1.74947786e+00
-1.02838483e-02 -2.72387952e-01 -8.76577675e-01 -1.51778877e-01
8.44465271e-02 -4.97400939e-01 1.74959093e-01 -1.77674949e+00
-1.36121571e-01 -3.73698324e-01 -6.92896247e-01 5.68685830e-01
3.93382818e-01 1.26740605e-01 -8.61848220e-02 -1.73479825e-01
-4.22581822e-01 6.94075674e-02 1.17417085e+00 -6.90233111e-01
7.60861263e-02 1.59889728e-01 -4.07317489e-01 7.65761495e-01
1.06362545e+00 -3.52469832e-01 -5.57951689e-01 -4.88547951e-01
4.24589276e-01 -4.61612582e-01 1.34425402e-01 -1.55341005e+00
-3.23463202e-01 -1.22435287e-01 2.21767738e-01 -6.00479424e-01
-2.43981823e-01 -1.15429509e+00 8.60802531e-02 6.95250988e-01
-4.76771653e-01 5.89713037e-01 -8.80570859e-02 3.80013809e-02
-6.71563685e-01 -8.91188323e-01 7.81977832e-01 -4.10556197e-01
-7.87371814e-01 -2.25595962e-02 -1.90010697e-01 -2.68685251e-01
1.43519568e+00 -2.88239837e-01 -1.85974285e-01 9.45861265e-02
-4.56596702e-01 -1.24513760e-01 4.43257242e-01 6.96237862e-01
5.75845420e-01 -1.22818267e+00 -6.27003968e-01 4.66478914e-01
4.57305908e-01 -7.48861730e-02 -5.07141054e-02 3.95277500e-01
-8.84976149e-01 2.65553534e-01 -7.74143994e-01 -6.42532647e-01
-1.21131206e+00 9.84131455e-01 2.56693006e-01 -4.15404975e-01
-3.18648309e-01 2.90734559e-01 -4.92150187e-02 -4.20045733e-01
3.27167153e-01 -9.33824182e-01 -5.96766591e-01 -9.41705704e-02
2.90043682e-01 6.64688826e-01 8.03701401e-01 -4.69617955e-02
-6.97593018e-02 4.44108963e-01 2.03387320e-01 6.33042037e-01
1.32978213e+00 3.08685631e-01 -3.01067561e-01 7.21963704e-01
1.04765189e+00 -3.68204296e-01 -9.25902724e-01 4.72274184e-01
5.72140753e-01 -4.80730593e-01 -1.78754315e-01 -8.30335081e-01
-1.66411304e+00 1.08314061e+00 1.09691143e+00 1.10123122e+00
1.18281269e+00 -1.23474196e-01 3.45731497e-01 4.24131662e-01
6.38457298e-01 -1.21029460e+00 4.37513322e-01 2.30514139e-01
1.05183387e+00 -1.43493843e+00 -6.54157251e-02 -4.90348190e-01
-4.06964332e-01 1.35538459e+00 1.05625808e+00 -1.65491208e-01
6.84580743e-01 7.73042023e-01 3.59793812e-01 -5.24185598e-01
-6.23350501e-01 2.08959103e-01 -9.92776453e-02 9.84645963e-01
7.89121330e-01 -2.10955814e-01 -3.91399175e-01 4.14128870e-01
-7.01546967e-02 1.89313337e-01 4.46726710e-01 1.38621581e+00
-3.21460217e-01 -1.47243154e+00 -5.31164467e-01 7.00402200e-01
-6.70300305e-01 4.87081647e-01 -3.72302264e-01 1.14090335e+00
6.94160342e-01 1.33246720e+00 4.73829433e-02 -6.98242247e-01
7.90886343e-01 -2.52932727e-01 4.17843044e-01 -8.89381528e-01
-8.18911016e-01 -5.53416967e-01 2.91123062e-01 -4.61301327e-01
-3.90421361e-01 -4.42181617e-01 -8.68404508e-01 1.53476834e-01
-4.85483557e-01 1.39123440e-01 7.31407106e-01 7.90547252e-01
1.11159272e-02 1.20107234e+00 7.29732752e-01 -4.27417725e-01
-6.61339164e-01 -1.28873909e+00 -8.65039229e-01 5.69354117e-01
1.52003080e-01 -9.32385504e-01 -3.20774019e-01 4.51153398e-01] | [7.347334384918213, 1.923999309539795] |
49ac7cc0-0c98-4d53-b9ea-5f3d6195bf32 | hopular-modern-hopfield-networks-for-tabular-1 | 2206.00664 | null | https://arxiv.org/abs/2206.00664v1 | https://arxiv.org/pdf/2206.00664v1.pdf | Hopular: Modern Hopfield Networks for Tabular Data | While Deep Learning excels in structured data as encountered in vision and natural language processing, it failed to meet its expectations on tabular data. For tabular data, Support Vector Machines (SVMs), Random Forests, and Gradient Boosting are the best performing techniques with Gradient Boosting in the lead. Recently, we saw a surge of Deep Learning methods that were tailored to tabular data but still underperform compared to Gradient Boosting on small-sized datasets. We suggest "Hopular", a novel Deep Learning architecture for medium- and small-sized datasets, where each layer is equipped with continuous modern Hopfield networks. The modern Hopfield networks use stored data to identify feature-feature, feature-target, and sample-sample dependencies. Hopular's novelty is that every layer can directly access the original input as well as the whole training set via stored data in the Hopfield networks. Therefore, Hopular can step-wise update its current model and the resulting prediction at every layer like standard iterative learning algorithms. In experiments on small-sized tabular datasets with less than 1,000 samples, Hopular surpasses Gradient Boosting, Random Forests, SVMs, and in particular several Deep Learning methods. In experiments on medium-sized tabular data with about 10,000 samples, Hopular outperforms XGBoost, CatBoost, LightGBM and a state-of-the art Deep Learning method designed for tabular data. Thus, Hopular is a strong alternative to these methods on tabular data. | ['Sepp Hochreiter', 'Angela Bitto-Nemling', 'Lukas Gruber', 'Bernhard Schäfl'] | 2022-06-01 | hopular-modern-hopfield-networks-for-tabular | https://openreview.net/forum?id=3zJVXU311-Q | https://openreview.net/pdf?id=3zJVXU311-Q | null | ['classification'] | ['methodology'] | [-5.70051372e-01 -1.90126166e-01 -3.26538920e-01 -5.24319530e-01
-1.62590057e-01 -3.43268126e-01 7.21888542e-01 1.78008020e-01
-5.62203586e-01 1.10357511e+00 -1.70299020e-02 -4.62666899e-01
-2.54424423e-01 -1.12129545e+00 -7.31268466e-01 -7.71568894e-01
-3.67679089e-01 9.30266440e-01 1.42946988e-01 -2.94382066e-01
3.26044738e-01 4.67054039e-01 -1.55080485e+00 7.37553596e-01
4.57087427e-01 1.64513958e+00 -9.47963446e-02 3.09412420e-01
-5.91698408e-01 1.05761969e+00 -4.02313143e-01 -5.34904540e-01
2.05731362e-01 4.25396822e-02 -8.75671506e-01 -5.58507383e-01
8.32148850e-01 -1.71422929e-01 -1.29731610e-01 5.23455441e-01
4.50056553e-01 -1.45584643e-01 4.22185034e-01 -1.36514080e+00
-5.26862204e-01 7.67855704e-01 -3.33022833e-01 2.24438891e-01
2.42922753e-02 8.02658424e-02 1.02946496e+00 -1.25598860e+00
6.23187065e-01 1.37169194e+00 1.24415934e+00 3.63062769e-01
-1.14972401e+00 -7.06737399e-01 3.66040707e-01 5.92206538e-01
-5.97778797e-01 -1.13992088e-01 5.34253061e-01 -4.96484131e-01
1.26950514e+00 1.52172312e-01 1.12326860e+00 1.11599708e+00
6.97428823e-01 1.05571830e+00 1.35644090e+00 -3.79440725e-01
4.62734759e-01 1.30548641e-01 5.41259170e-01 1.01963186e+00
9.83773395e-02 3.21972668e-01 -1.01312089e+00 -1.13180801e-01
3.39761257e-01 8.96739885e-02 7.22260773e-02 -6.53090596e-01
-1.19236767e+00 1.06068230e+00 9.72503185e-01 1.56939507e-01
-4.87041205e-01 2.51592964e-01 7.12325931e-01 6.35851562e-01
4.66159344e-01 5.32829762e-01 -8.44116211e-01 8.17131400e-02
-1.00300646e+00 4.52970237e-01 8.40210974e-01 5.56991935e-01
8.13092053e-01 2.03746140e-01 -8.20634291e-02 7.98766017e-01
2.02728644e-01 3.42014283e-01 8.28744471e-01 -4.90074873e-01
4.29715425e-01 7.24234939e-01 -1.40908301e-01 -9.03957725e-01
-9.59576428e-01 -7.74581134e-01 -1.09276783e+00 6.30024910e-01
5.23310542e-01 -3.67223881e-02 -1.40060186e+00 1.20329559e+00
1.72455281e-01 -6.20232403e-01 -2.61791855e-01 8.15603316e-01
9.99205351e-01 6.40500069e-01 3.02175302e-02 2.26003565e-02
1.09235692e+00 -1.27465999e+00 -2.83338368e-01 -6.69119060e-01
3.20065618e-01 -3.36391449e-01 9.33518231e-01 8.61530900e-01
-7.66709805e-01 -5.32658696e-01 -1.03397727e+00 1.16615310e-01
-9.23520446e-01 -2.98881292e-01 1.18531370e+00 6.89044178e-01
-1.19291162e+00 7.78529167e-01 -5.15255272e-01 -1.35630772e-01
8.65494907e-01 6.17059052e-01 -4.33402479e-01 -7.84449875e-02
-1.36102653e+00 1.23064685e+00 4.19804960e-01 2.44723707e-01
-8.98085713e-01 -5.30122757e-01 -4.61534142e-01 1.16181634e-01
1.73116326e-01 -8.35192382e-01 1.20140970e+00 -1.02740765e+00
-1.48558021e+00 7.10126519e-01 9.73607153e-02 -1.09028316e+00
5.44513404e-01 -1.70506462e-01 -7.49849826e-02 -3.48992169e-01
-1.04005791e-01 8.60379815e-01 1.09393013e+00 -8.38710129e-01
-8.05223107e-01 -8.57904971e-01 -2.43272662e-01 1.46752074e-01
-4.44678873e-01 -3.95864218e-01 1.68166772e-01 -2.35301733e-01
2.75404066e-01 -7.80044794e-01 -2.87805974e-01 -2.55207866e-02
-1.90661520e-01 -3.08883965e-01 6.96673870e-01 -4.08999711e-01
7.78217793e-01 -1.69101048e+00 -1.13189198e-01 1.58177659e-01
2.69043416e-01 2.07908824e-01 8.05515721e-02 3.11749697e-01
-1.89969361e-01 -1.90628573e-01 7.05639496e-02 1.25511080e-01
-6.65187463e-02 4.21079174e-02 -3.26954454e-01 2.93305397e-01
-3.51918489e-02 1.24946916e+00 -7.77749002e-01 -2.64895767e-01
7.43676051e-02 1.07949175e-01 -3.76071393e-01 -2.05916569e-01
-3.34244370e-01 -3.58582795e-01 -6.37530684e-02 8.61868620e-01
5.64296901e-01 -3.66214603e-01 -2.53849365e-02 -1.00608834e-03
-1.90188855e-01 4.48178977e-01 -8.70588958e-01 1.33902919e+00
-5.92232406e-01 9.60226536e-01 -1.44600108e-01 -1.23168802e+00
1.20307159e+00 2.50347722e-02 2.76823670e-01 -1.05431437e+00
-5.55246510e-02 4.06931549e-01 -6.06448017e-02 5.01834368e-03
1.94431737e-01 -3.82922322e-01 1.26407325e-01 1.88005596e-01
3.85269850e-01 4.48638946e-02 2.01083079e-01 -1.15684956e-01
1.07120872e+00 3.28491107e-02 3.80552411e-01 -2.37370223e-01
2.64654458e-01 5.01352489e-01 2.20489025e-01 1.02901351e+00
5.13526984e-02 2.49760032e-01 5.89448571e-01 -1.45108044e+00
-8.21093798e-01 -9.47237015e-01 -3.65538239e-01 1.58845139e+00
-2.77005285e-01 -2.35647440e-01 -5.09988844e-01 -9.21117723e-01
3.63597572e-01 6.29905581e-01 -9.72060382e-01 1.39397485e-02
-2.94272095e-01 -1.17949247e+00 9.02369469e-02 5.60236037e-01
1.00910389e+00 -1.62671316e+00 -7.42925763e-01 4.53063220e-01
3.21286172e-01 -3.42953473e-01 4.14317578e-01 1.25243366e+00
-1.37997067e+00 -8.70395005e-01 -4.24838692e-01 -8.24272811e-01
3.29612494e-01 -2.41609409e-01 1.46016824e+00 -3.09483320e-01
-1.37441590e-01 -4.14049327e-01 -7.59272128e-02 -6.12277567e-01
-2.03050934e-02 4.65281844e-01 -2.02416182e-01 -3.07658643e-01
5.31596005e-01 -5.47711790e-01 -4.66959447e-01 1.71121553e-01
-3.84604603e-01 9.04288441e-02 7.07276165e-01 1.61710405e+00
2.42978171e-01 -2.65987851e-02 7.24335730e-01 -1.13933575e+00
6.63792491e-01 -4.06070709e-01 -7.08793700e-01 6.34282976e-02
-1.21442711e+00 4.19295013e-01 8.80595624e-01 -1.73879504e-01
-7.00753450e-01 2.47673932e-02 -1.44933864e-01 9.63078905e-03
4.52304631e-02 9.60038900e-01 3.23782653e-01 7.06227198e-02
1.13172817e+00 1.69818923e-01 3.30082178e-02 -5.05602896e-01
2.02019110e-01 6.19784713e-01 3.41675371e-01 -1.82743803e-01
3.74226898e-01 4.98148561e-01 5.93828224e-03 -4.61483933e-02
-9.04864669e-01 -1.06733911e-01 -7.06280053e-01 -2.51803309e-01
4.42385256e-01 -6.22266769e-01 -7.62275100e-01 9.31061506e-01
-8.56793046e-01 -5.07550657e-01 -1.08542837e-01 1.31185979e-01
-5.23719907e-01 -4.34068799e-01 -6.52008355e-01 -6.70403004e-01
-8.69738877e-01 -7.23632038e-01 7.20452130e-01 3.63914259e-02
-1.28897615e-02 -9.78370309e-01 -6.98824273e-03 1.48805514e-01
5.42285860e-01 -7.56156221e-02 1.17430413e+00 -7.88634777e-01
-2.68669516e-01 -5.09081423e-01 -1.59290880e-01 1.71203643e-01
-1.25506446e-01 -3.91230881e-01 -1.00681007e+00 -1.87965006e-01
-2.14088559e-01 -6.61472976e-01 1.42488742e+00 5.50732732e-01
1.21401393e+00 -3.18650186e-01 -3.22073311e-01 6.35872900e-01
1.19807124e+00 4.11808074e-01 5.11597455e-01 1.23476362e+00
5.17072439e-01 5.15673041e-01 6.11691415e-01 1.99811190e-01
2.46477112e-01 3.12316298e-01 8.46678615e-01 -6.92835972e-02
1.47959203e-01 -2.23948419e-01 2.96697646e-01 3.51701111e-01
9.58583653e-02 1.00331582e-01 -1.30253184e+00 3.38418186e-01
-1.93603039e+00 -1.17919743e+00 -6.68518469e-02 2.18494534e+00
8.19637537e-01 6.67095840e-01 2.29382768e-01 3.49110156e-01
3.13787162e-01 1.13192992e-03 -9.25258994e-01 -6.81157470e-01
-3.97116579e-02 2.62166411e-01 5.62528729e-01 1.19765095e-01
-1.39982009e+00 1.04300797e+00 6.75311518e+00 6.94812834e-01
-1.50919342e+00 -1.59249395e-01 1.11079669e+00 -2.55772203e-01
4.54169996e-02 -2.37478748e-01 -1.05691731e+00 3.55018526e-01
1.09596300e+00 2.82551497e-01 7.27486849e-01 1.46258521e+00
-3.74301106e-01 -2.31880069e-01 -9.88807082e-01 9.99492049e-01
-1.18549421e-01 -1.78156054e+00 -8.21252912e-02 -8.76455382e-02
4.74624276e-01 5.27731359e-01 1.72912151e-01 6.94436312e-01
6.25347912e-01 -1.42594516e+00 9.34279442e-01 1.99990630e-01
3.72367859e-01 -7.80194819e-01 9.24727738e-01 3.60924751e-01
-4.52870667e-01 -6.40873611e-01 -5.23832858e-01 -2.90894836e-01
-4.75146949e-01 9.41195905e-01 -1.07768059e+00 1.17718510e-01
1.25739741e+00 8.08436155e-01 -8.56708527e-01 1.04161823e+00
-1.06631845e-01 6.21322155e-01 -3.23092282e-01 -5.75705111e-01
6.43733621e-01 1.65860638e-01 1.03258668e-02 1.12192333e+00
3.88655835e-03 -5.64398468e-01 -8.68321508e-02 3.10416192e-01
8.54982287e-02 6.40619099e-02 -5.68120480e-01 2.99482673e-01
2.28957862e-01 1.24808955e+00 -6.90746069e-01 -3.97700608e-01
-3.42990220e-01 5.30529559e-01 6.83281362e-01 1.42569855e-01
-3.96125555e-01 -2.39549026e-01 2.35435516e-01 2.41163418e-01
6.18920863e-01 1.39111742e-01 -8.07526350e-01 -8.53798509e-01
-7.85321817e-02 -1.17412364e+00 5.19981146e-01 -9.22968566e-01
-1.32842302e+00 9.68466818e-01 -2.67841250e-01 -8.09742391e-01
-6.20693743e-01 -1.27375317e+00 -1.57973751e-01 7.59536564e-01
-1.38617468e+00 -1.03418875e+00 -2.93999612e-01 7.27289200e-01
3.36746007e-01 -6.60325170e-01 9.29467797e-01 -1.07093684e-01
-2.58134186e-01 2.64375031e-01 5.20459414e-01 2.23819792e-01
4.26658809e-01 -1.33468723e+00 6.73619986e-01 1.48358926e-01
4.38589096e-01 5.02886713e-01 6.13065422e-01 -4.03204471e-01
-1.27911508e+00 -7.48392820e-01 7.43486047e-01 -1.63744420e-01
8.53354990e-01 -5.94570816e-01 -7.91242540e-01 5.86210787e-01
1.63963050e-01 2.78888494e-01 2.59863973e-01 7.72641659e-01
-5.83300710e-01 -6.98274374e-01 -1.09926879e+00 3.81880730e-01
9.67357337e-01 -1.55882686e-01 -5.01949668e-01 2.88018227e-01
-8.85956176e-03 -4.09094930e-01 -4.85637069e-01 3.75006557e-01
1.01063180e+00 -1.29073822e+00 7.12701619e-01 -8.16253960e-01
6.44951880e-01 1.06531188e-01 6.97518364e-02 -1.86223149e+00
-4.81215745e-01 -1.48942500e-01 -2.81537354e-01 4.91900712e-01
6.25606954e-01 -8.23426664e-01 1.18940508e+00 2.20095709e-01
-1.02269016e-02 -1.19514811e+00 -1.04231322e+00 -6.81734502e-01
4.40196633e-01 -3.30373853e-01 5.84780633e-01 8.08219969e-01
1.01005867e-01 6.19610667e-01 -2.49446154e-01 -5.53892493e-01
5.20591199e-01 6.45766139e-01 5.61634898e-01 -1.68615210e+00
-2.02138156e-01 -5.79945445e-01 -5.78457534e-01 -5.53242028e-01
3.12856510e-02 -1.00500810e+00 6.27402514e-02 -1.69037855e+00
1.02436900e-01 -6.38225853e-01 -5.21185517e-01 1.01690733e+00
-1.86838266e-02 2.36651868e-01 1.23954497e-01 -3.35717853e-03
-3.87505114e-01 2.81463474e-01 9.52051997e-01 -5.73290110e-01
-8.62430334e-02 2.07699761e-01 -6.34384871e-01 6.39870524e-01
9.35841739e-01 -6.13761187e-01 -1.55190080e-01 -3.24696600e-01
7.58389831e-01 3.64994002e-03 3.35368454e-01 -1.20052600e+00
2.30340719e-01 2.54919771e-02 1.13180196e+00 -7.86789715e-01
1.30575702e-01 -7.70682871e-01 -6.78742900e-02 6.13026202e-01
-4.40228134e-01 1.61116406e-01 4.55234081e-01 2.32523367e-01
-2.72035360e-01 -3.01343262e-01 5.94910562e-01 -3.66322130e-01
-7.91931510e-01 3.01688373e-01 -4.73766863e-01 -2.08141550e-01
6.37176573e-01 -2.67387211e-01 -5.96193194e-01 -2.72215515e-01
-8.58506024e-01 2.44608700e-01 -3.32808234e-02 4.90421593e-01
5.42138636e-01 -1.23905730e+00 -6.90355361e-01 3.53887439e-01
-1.01028837e-01 3.91785428e-03 -2.51044512e-01 6.26095235e-01
-4.68296111e-01 6.85484648e-01 -7.20787048e-01 -5.62956274e-01
-1.14157569e+00 7.20443785e-01 5.47802746e-01 -6.70391798e-01
-3.71078938e-01 1.08164859e+00 -1.61084205e-01 -7.14200854e-01
3.42858434e-01 -2.25437894e-01 -2.33844027e-01 5.87212324e-01
3.76464456e-01 2.41867259e-01 7.26751447e-01 1.66324705e-01
-4.99347210e-01 7.55274668e-02 -5.41925311e-01 -1.31577998e-01
1.74002838e+00 4.02162671e-01 -2.54169673e-01 7.69444585e-01
8.66016924e-01 -5.27513206e-01 -1.25511944e+00 -1.67392995e-02
3.13310653e-01 -2.71148145e-01 4.02690321e-01 -1.42952371e+00
-1.14983082e+00 9.52717721e-01 6.68830335e-01 3.34925056e-01
1.10601270e+00 -3.89458686e-01 3.83478016e-01 9.35658395e-01
7.43900359e-01 -1.02808070e+00 -1.13995157e-01 9.14634705e-01
1.00588953e+00 -1.30471730e+00 2.66306043e-01 3.25744390e-01
-4.05851901e-01 1.42726564e+00 6.72840118e-01 -1.67281941e-01
6.67545795e-01 4.65121776e-01 2.40918532e-01 -2.41074905e-01
-1.37317622e+00 1.21158324e-01 1.03166535e-01 6.31774485e-01
4.81589586e-01 -6.47864044e-02 3.06381881e-02 2.69685537e-01
-6.53906524e-01 2.79198378e-01 -2.26062480e-02 7.00036466e-01
-6.94883883e-01 -8.49241793e-01 -5.60195744e-01 1.10081387e+00
-2.20903724e-01 -5.72436452e-01 -4.25717741e-01 8.97563636e-01
-9.46514383e-02 5.31390548e-01 1.53227940e-01 -5.94494343e-01
3.27639543e-02 4.50477362e-01 3.80115956e-01 -1.20063297e-01
-1.12522364e+00 -4.84574318e-01 1.36274114e-01 -5.16779423e-01
1.58977166e-01 -6.70241535e-01 -1.08698654e+00 -5.11166573e-01
-1.87257975e-01 1.10399075e-01 9.86975431e-01 8.19046855e-01
3.80834080e-02 2.64613390e-01 4.62339908e-01 -8.88496697e-01
-7.00516522e-01 -1.10930181e+00 -4.50322151e-01 -1.01480605e-02
4.68964309e-01 -7.53955185e-01 -2.40133658e-01 -3.31534564e-01] | [8.608287811279297, 4.019047260284424] |
143842b4-36e8-4495-94c2-eda44297d6bf | estimating-optimal-policy-value-in-general | 2302.09451 | null | https://arxiv.org/abs/2302.09451v1 | https://arxiv.org/pdf/2302.09451v1.pdf | Estimating Optimal Policy Value in General Linear Contextual Bandits | In many bandit problems, the maximal reward achievable by a policy is often unknown in advance. We consider the problem of estimating the optimal policy value in the sublinear data regime before the optimal policy is even learnable. We refer to this as $V^*$ estimation. It was recently shown that fast $V^*$ estimation is possible but only in disjoint linear bandits with Gaussian covariates. Whether this is possible for more realistic context distributions has remained an open and important question for tasks such as model selection. In this paper, we first provide lower bounds showing that this general problem is hard. However, under stronger assumptions, we give an algorithm and analysis proving that $\widetilde{\mathcal{O}}(\sqrt{d})$ sublinear estimation of $V^*$ is indeed information-theoretically possible, where $d$ is the dimension. We then present a more practical, computationally efficient algorithm that estimates a problem-dependent upper bound on $V^*$ that holds for general distributions and is tight when the context distribution is Gaussian. We prove our algorithm requires only $\widetilde{\mathcal{O}}(\sqrt{d})$ samples to estimate the upper bound. We use this upper bound and the estimator to obtain novel and improved guarantees for several applications in bandit model selection and testing for treatment effects. | ['Emma Brunskill', 'Vidya Muthukumar', 'Aldo Pacchiano', 'Weihao Kong', 'Jonathan N. Lee'] | 2023-02-19 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [ 3.07781816e-01 2.65184373e-01 -8.80197167e-01 -1.24584667e-01
-1.31581259e+00 -7.38921821e-01 -9.28917080e-02 1.84777588e-01
-6.57709837e-01 1.30640101e+00 -2.89928496e-01 -8.26322794e-01
-7.72571146e-01 -6.95264101e-01 -1.17469442e+00 -1.03037667e+00
-3.71183336e-01 8.25984359e-01 -8.08662921e-02 3.29194933e-01
1.83326989e-01 4.21240062e-01 -1.30275452e+00 -1.06597401e-01
1.00652111e+00 1.32833159e+00 5.60588948e-02 9.03810203e-01
8.50724876e-02 3.79648894e-01 -4.74106908e-01 -3.97006363e-01
5.66756248e-01 -7.29511440e-01 -7.38034546e-01 2.75115091e-02
3.47361684e-01 -4.45775509e-01 -8.18070695e-02 1.37156367e+00
4.95121360e-01 8.81412774e-02 5.60006142e-01 -1.04314089e+00
-1.62585080e-01 9.90624785e-01 -8.38156283e-01 4.71446306e-01
-1.22171424e-01 -2.07105443e-01 1.23272729e+00 -5.28114326e-02
3.16514522e-01 1.13043272e+00 3.23993474e-01 3.41036707e-01
-1.57418251e+00 -9.86224055e-01 2.36954063e-01 -1.24389112e-01
-1.14435935e+00 -2.99583763e-01 3.95056814e-01 -4.10651863e-01
3.99954319e-01 6.02062881e-01 4.18441147e-01 6.42870188e-01
-2.58516103e-01 1.00804162e+00 1.34822929e+00 -6.41298413e-01
4.30057615e-01 2.30924025e-01 3.50501388e-01 6.22285545e-01
7.09715188e-01 3.91492546e-01 -2.22993985e-01 -3.63104045e-01
1.10368049e+00 -1.04157887e-01 -3.61700922e-01 -2.30872199e-01
-8.59174430e-01 1.19243765e+00 1.26415908e-01 4.80950437e-02
-3.22991461e-01 5.91281712e-01 1.23478301e-01 4.93372232e-01
5.81995785e-01 3.76031816e-01 -5.82634509e-01 -3.49648684e-01
-9.93614018e-01 4.46761966e-01 7.74695992e-01 1.14444172e+00
4.18844938e-01 -2.15397794e-02 -4.20703620e-01 7.87719905e-01
-2.58278608e-01 9.28241730e-01 -3.43890429e-01 -1.17151582e+00
8.33255291e-01 -1.03092290e-01 8.01867545e-01 -4.61713761e-01
-2.26377621e-01 -7.41823375e-01 -6.44961119e-01 -9.61264372e-02
9.73171115e-01 -4.89885390e-01 -6.11985266e-01 1.91972697e+00
2.54221231e-01 -1.40203446e-01 -3.57704937e-01 5.48149705e-01
-1.57153770e-01 5.64092398e-01 -1.84002504e-01 -9.23023641e-01
1.06610215e+00 -3.54188740e-01 -5.21777272e-01 -2.81701505e-01
6.33101463e-01 -4.77924377e-01 9.24372196e-01 5.39859533e-01
-1.38856721e+00 2.46337682e-01 -6.75100207e-01 4.37460393e-01
3.17672342e-01 -1.32369086e-01 9.77248669e-01 1.05504453e+00
-6.66929126e-01 6.12691879e-01 -6.09185874e-01 4.48447727e-02
7.28456616e-01 6.89578354e-01 4.12109531e-02 -2.41845861e-01
-8.36644053e-01 4.18350965e-01 3.01179916e-01 -2.13271499e-01
-7.84451544e-01 -6.32298887e-01 -5.05880713e-01 2.25532770e-01
8.13064158e-01 -4.50444043e-01 1.35821688e+00 -6.99137390e-01
-1.21733522e+00 6.19763017e-01 -2.54482538e-01 -8.04949820e-01
6.64176762e-01 9.71941501e-02 2.82986283e-01 4.80602533e-02
-6.16075695e-02 -4.38990146e-02 6.61692798e-01 -9.25090671e-01
-9.37856495e-01 -7.29867995e-01 4.52330351e-01 3.62328365e-02
-9.93030146e-02 6.96550906e-02 -7.88121670e-02 -4.83655602e-01
3.29493657e-02 -1.06347871e+00 -4.19798404e-01 -2.73108721e-01
-4.47090536e-01 -1.47610411e-01 -1.86029524e-01 -3.53774369e-01
1.40460086e+00 -1.86965334e+00 -1.30646959e-01 5.30399382e-01
5.06153256e-02 -5.03182858e-02 2.28284329e-01 2.87899643e-01
2.32249320e-01 4.16370392e-01 -1.77225664e-01 -9.62814316e-02
2.10021287e-01 7.27907047e-02 -3.29119325e-01 7.29815722e-01
-5.49143195e-01 5.44648945e-01 -5.80086231e-01 -3.65153581e-01
-1.32756144e-01 -2.02545911e-01 -9.36526418e-01 -1.34222776e-01
-4.48995501e-01 3.56309533e-01 -7.22160995e-01 5.45671105e-01
8.98098111e-01 -3.58715057e-01 2.53359765e-01 3.44126612e-01
5.96916191e-02 1.01802796e-01 -1.37562025e+00 1.02091670e+00
-2.83878863e-01 2.84269780e-01 4.16130185e-01 -1.68899584e+00
4.00926530e-01 5.39831892e-02 6.40700281e-01 -5.45085490e-01
4.05314386e-01 3.72771442e-01 -1.29489809e-01 -2.50543743e-01
-9.05654356e-02 -7.46012986e-01 -2.02337354e-01 4.44752514e-01
-3.66185874e-01 1.01620361e-01 2.27932662e-01 -1.74672633e-01
1.18041503e+00 -3.38718146e-01 1.92912757e-01 -4.92495149e-01
-4.88364659e-02 -1.80940434e-01 5.53015053e-01 1.47746754e+00
-8.02571476e-02 -3.36409942e-03 1.12578809e+00 -3.20779048e-02
-9.95074451e-01 -9.93639588e-01 -5.00869513e-01 1.18782258e+00
-5.45163751e-02 1.09645516e-01 -7.54813313e-01 -5.39411485e-01
4.06250536e-01 9.15686667e-01 -9.53311741e-01 2.06674322e-01
-2.94358999e-01 -1.05052018e+00 1.64011121e-01 4.02134478e-01
1.96044639e-01 -5.71405947e-01 -3.77991468e-01 2.30235204e-01
1.22018032e-01 -9.87955511e-01 -4.82549936e-01 5.35677612e-01
-1.03583407e+00 -8.95292401e-01 -8.45558345e-01 -2.12507203e-01
4.92398441e-01 1.10914595e-02 8.69832814e-01 -2.90059716e-01
1.59213722e-01 2.19230205e-01 -1.20079778e-01 -7.66824305e-01
-1.57464832e-01 8.09462592e-02 -1.50183246e-01 -2.21545324e-01
1.56236023e-01 -4.79818672e-01 -7.22822428e-01 2.41531000e-01
-7.99667120e-01 -1.86236054e-01 5.70231020e-01 8.98857057e-01
8.56589437e-01 1.69292912e-02 5.04994154e-01 -1.23695397e+00
4.62354332e-01 -4.65384871e-01 -1.36370707e+00 2.65843183e-01
-5.11871874e-01 5.22370696e-01 5.19590020e-01 -4.96065587e-01
-5.30553222e-01 -7.74379820e-02 1.92094203e-02 -3.90049726e-01
3.00955772e-01 6.23263836e-01 6.52444316e-03 1.23181500e-01
6.07001066e-01 2.10819885e-01 -5.85875995e-02 -6.61121666e-01
8.49733353e-02 6.16663098e-01 6.69311434e-02 -1.01025116e+00
4.34448391e-01 4.10538733e-01 3.87835681e-01 -6.79964066e-01
-1.30375457e+00 -1.46495625e-01 1.54622748e-01 2.80331641e-01
2.77170420e-01 -6.11850560e-01 -1.45598781e+00 -3.44685584e-01
-6.10778034e-01 -7.48678684e-01 -4.88631338e-01 7.83617556e-01
-1.09714758e+00 6.64968323e-03 -2.33551398e-01 -1.55814576e+00
-9.96797904e-02 -1.04503989e+00 7.31654525e-01 -4.04210687e-02
1.81599870e-01 -7.47552693e-01 -1.48870781e-01 4.38554257e-01
1.67264551e-01 1.49760649e-01 1.00827098e+00 -5.92173874e-01
-7.68818498e-01 -3.51663917e-01 -3.16081136e-01 2.26185098e-01
-2.30389625e-01 -5.05218923e-01 -4.87169981e-01 -4.39963341e-01
-8.09576213e-02 -1.46852031e-01 1.02414834e+00 1.10650802e+00
1.76519406e+00 -8.44334781e-01 -4.19844717e-01 5.80282748e-01
1.37606657e+00 4.52731907e-01 4.65720624e-01 1.52093351e-01
7.49863237e-02 1.90833628e-01 6.84600055e-01 8.61241162e-01
3.52910114e-03 6.53075933e-01 4.22152609e-01 1.78118840e-01
5.96848965e-01 -5.73303364e-02 7.55134672e-02 -3.00659556e-02
-2.82736737e-02 -3.00296992e-01 -5.11352360e-01 6.40688300e-01
-1.73889852e+00 -1.02385557e+00 1.10304952e-01 2.90849686e+00
1.25423479e+00 3.79902154e-01 4.91697073e-01 9.05926451e-02
7.19913960e-01 -2.91172445e-01 -8.49984050e-01 -5.83219051e-01
2.66239364e-02 7.55393386e-01 1.29975641e+00 7.38110662e-01
-8.80277395e-01 6.45251036e-01 6.01352072e+00 1.28821874e+00
-8.03798556e-01 2.68273324e-01 9.15810227e-01 -5.67377985e-01
-3.82138163e-01 4.39312309e-02 -9.65801358e-01 5.84137917e-01
1.00879526e+00 -3.30312252e-01 5.49183488e-01 1.01523483e+00
2.73896694e-01 -5.44371605e-01 -1.13499129e+00 9.51620758e-01
-4.46528763e-01 -1.32964826e+00 -5.70382118e-01 6.28183544e-01
8.89547706e-01 -2.72391826e-01 2.67591298e-01 1.74071804e-01
7.69923389e-01 -1.02561677e+00 4.88015056e-01 3.58594544e-02
1.14121830e+00 -1.06142259e+00 4.77245718e-01 6.78722918e-01
-6.62639618e-01 -3.98336500e-01 -5.52479982e-01 -2.81995200e-02
-9.83302668e-02 9.85205948e-01 -7.01358080e-01 3.18935633e-01
4.97417241e-01 1.06033824e-01 2.35132158e-01 1.25495601e+00
5.70899099e-02 7.72283494e-01 -8.71712267e-01 -1.82932600e-01
1.80105805e-01 -1.51428550e-01 3.70766878e-01 9.48002458e-01
5.92419088e-01 4.14066374e-01 2.09852338e-01 5.75881720e-01
-2.50126988e-01 1.48371056e-01 -3.66583943e-01 -2.17079282e-01
5.41417837e-01 4.70122606e-01 -5.55593371e-01 -2.70723939e-01
1.21940393e-03 5.65353274e-01 3.23366582e-01 3.24568212e-01
-9.43603098e-01 -3.27399299e-02 7.61884749e-01 2.46795058e-01
6.67488873e-01 -6.16803505e-02 -4.10835057e-01 -9.88790333e-01
1.16949998e-01 -5.43141603e-01 7.57056832e-01 -6.64463341e-02
-1.03132939e+00 -2.11841911e-02 4.28876758e-01 -8.79460156e-01
-2.36216411e-01 -5.02954125e-01 1.28348008e-01 7.93023825e-01
-1.22119629e+00 -4.22893584e-01 4.97136980e-01 4.93090719e-01
1.67959064e-01 1.09761432e-01 6.15788758e-01 1.05972197e-02
-5.18764615e-01 9.60009873e-01 8.41108203e-01 -2.36065403e-01
2.77923435e-01 -1.34333313e+00 -5.65038562e-01 5.37103176e-01
-9.85124186e-02 3.74175757e-01 1.10399151e+00 -4.35254008e-01
-1.35294724e+00 -7.78016388e-01 5.48769772e-01 -6.09715991e-02
6.53389513e-01 -3.20053399e-01 -3.11926514e-01 8.89994979e-01
-2.26075828e-01 1.23764023e-01 7.70195782e-01 2.92212218e-01
-1.36692464e-01 -4.92344052e-01 -1.29267669e+00 3.88934374e-01
1.10815322e+00 -8.69497210e-02 1.13262296e-01 6.67678356e-01
5.94655812e-01 -4.56983596e-01 -1.05059969e+00 3.48417848e-01
6.17669523e-01 -8.85664523e-01 8.00591290e-01 -8.49983633e-01
7.28189126e-02 3.80820096e-01 -4.03511703e-01 -9.38185513e-01
5.58370315e-02 -1.06339562e+00 -2.88520277e-01 6.80642784e-01
6.94255829e-01 -8.08585405e-01 1.02579272e+00 7.51600087e-01
3.60332936e-01 -9.17894423e-01 -1.35520089e+00 -1.04533088e+00
6.57313645e-01 -6.83034301e-01 4.03789431e-01 6.82186723e-01
-5.24559349e-04 -1.06906489e-01 -6.70979738e-01 8.19237679e-02
8.75907481e-01 3.47400725e-01 5.40477514e-01 -1.10239196e+00
-7.96487570e-01 -4.90965396e-01 1.64596252e-02 -1.53750813e+00
-2.56608594e-02 -5.46162367e-01 -9.73223671e-02 -1.20843518e+00
5.47805011e-01 -1.07264030e+00 -4.90902007e-01 2.57294029e-01
-1.03411391e-01 -3.78847986e-01 1.46335021e-01 -5.87645806e-02
-4.49876100e-01 2.66953707e-01 1.35295606e+00 6.56670555e-02
-1.18261896e-01 6.80202782e-01 -1.06258929e+00 3.70160550e-01
7.50020564e-01 -6.78715885e-01 -4.27317351e-01 -1.81471705e-01
4.41382229e-01 8.30201983e-01 1.59227625e-01 -5.15490115e-01
-1.19471088e-01 -7.60805368e-01 5.94067425e-02 -5.73640108e-01
2.74639368e-01 -7.36893058e-01 1.45877851e-02 5.48196554e-01
-7.11644590e-01 -4.22471017e-01 3.20178755e-02 6.66027427e-01
4.46244508e-01 -4.34295267e-01 9.40762460e-01 3.08974576e-03
2.62945086e-01 5.78022718e-01 -7.23899622e-03 2.55646735e-01
1.15008271e+00 1.32977106e-02 -1.84761733e-01 -8.58010828e-01
-8.23065341e-01 2.38749340e-01 8.68466794e-02 -4.32429016e-01
1.64771527e-01 -1.10041749e+00 -8.15222263e-01 -8.93053189e-02
-2.34265894e-01 -1.58116054e-02 2.82313555e-01 1.09652555e+00
-1.33800149e-01 7.10990012e-01 4.39205289e-01 -5.50529182e-01
-1.05467343e+00 7.37365961e-01 3.16992015e-01 -5.83611786e-01
-1.93897076e-02 1.03044415e+00 2.09766701e-01 3.05923343e-01
2.63698816e-01 -4.28211749e-01 4.19892102e-01 -2.40126997e-02
6.14789903e-01 3.53874773e-01 -2.28812441e-01 3.88960267e-04
-7.30641037e-02 1.48403421e-01 -2.12102354e-01 -4.26472038e-01
1.36365891e+00 -2.03403875e-01 4.26411591e-02 2.87196428e-01
1.24514663e+00 8.99466872e-02 -1.45223475e+00 -4.59506005e-01
-1.03984445e-01 -7.43134379e-01 -8.98159575e-03 -6.51904523e-01
-1.03157723e+00 7.06157029e-01 6.24133170e-01 6.86012805e-01
1.09420753e+00 3.01442087e-01 3.58631402e-01 2.99751461e-01
6.41537547e-01 -1.10523248e+00 -3.45916301e-01 2.68768013e-01
7.35522509e-01 -1.07857752e+00 7.65475333e-02 -4.17591184e-01
-3.53520304e-01 7.64331818e-01 2.62627322e-02 -6.24862127e-02
7.49292970e-01 1.49051443e-01 -8.29520106e-01 8.38110819e-02
-6.44887686e-01 -3.80755574e-01 1.06412686e-01 2.14886427e-01
2.61148483e-01 4.71495837e-01 -6.29380643e-01 6.91370070e-01
-3.44656020e-01 1.76648572e-01 3.36949885e-01 8.92411470e-01
-7.05224991e-01 -1.26549137e+00 -5.18395245e-01 9.49094534e-01
-1.00592816e+00 -1.22505024e-01 1.05766200e-01 7.90444553e-01
-2.04718634e-01 8.72867286e-01 -8.36672261e-02 2.34743968e-01
1.28867686e-01 -3.46717983e-02 9.59869683e-01 -3.05313259e-01
2.72925328e-02 4.99922097e-01 2.30523840e-01 -3.03115696e-01
-2.77357191e-01 -7.09132373e-01 -8.19571078e-01 -6.88259542e-01
-4.59790349e-01 4.01192307e-01 5.04614949e-01 1.05548704e+00
6.64562061e-02 6.04505688e-02 8.58836651e-01 -3.46016705e-01
-1.08515453e+00 -8.80080521e-01 -1.01039159e+00 -6.65716976e-02
5.08676112e-01 -7.40363538e-01 -3.80862415e-01 -3.67834181e-01] | [4.641083717346191, 3.4116344451904297] |
fb317051-dcf0-423f-a5b0-0d39b1263e48 | tabgsl-graph-structure-learning-for-tabular | 2305.15843 | null | https://arxiv.org/abs/2305.15843v1 | https://arxiv.org/pdf/2305.15843v1.pdf | TabGSL: Graph Structure Learning for Tabular Data Prediction | This work presents a novel approach to tabular data prediction leveraging graph structure learning and graph neural networks. Despite the prevalence of tabular data in real-world applications, traditional deep learning methods often overlook the potentially valuable associations between data instances. Such associations can offer beneficial insights for classification tasks, as instances may exhibit similar patterns of correlations among features and target labels. This information can be exploited by graph neural networks, necessitating robust graph structures. However, existing studies primarily focus on improving graph structure from noisy data, largely neglecting the possibility of deriving graph structures from tabular data. We present a novel solution, Tabular Graph Structure Learning (TabGSL), to enhance tabular data prediction by simultaneously learning instance correlation and feature interaction within a unified framework. This is achieved through a proposed graph contrastive learning module, along with transformer-based feature extractor and graph neural network. Comprehensive experiments conducted on 30 benchmark tabular datasets demonstrate that TabGSL markedly outperforms both tree-based models and recent deep learning-based tabular models. Visualizations of the learned instance embeddings further substantiate the effectiveness of TabGSL. | ['Cheng-Te Li', 'Jay Chiehen Liao'] | 2023-05-25 | null | null | null | null | ['graph-structure-learning'] | ['graphs'] | [-1.30668983e-01 3.85345995e-01 -4.76276487e-01 -3.03629100e-01
-2.88587600e-01 -5.40067136e-01 5.07286549e-01 7.28598416e-01
5.43682337e-01 6.93447888e-01 3.26706439e-01 -3.50587875e-01
-6.59360528e-01 -1.21262503e+00 -6.79933369e-01 -6.65852129e-01
-4.22224879e-01 7.18198657e-01 -2.92180359e-01 -1.48088649e-01
1.37404427e-01 4.67314810e-01 -1.31606960e+00 5.47329903e-01
9.21309173e-01 1.27484238e+00 -3.54832113e-01 1.15635924e-01
-6.77298725e-01 1.20590174e+00 -5.76907516e-01 -8.31995785e-01
1.83723018e-01 -1.45617545e-01 -5.16728044e-01 2.59705275e-01
8.00092161e-01 1.26844183e-01 -6.77245677e-01 9.53888416e-01
2.34530076e-01 -1.00488082e-01 5.09501755e-01 -1.55614221e+00
-1.08364666e+00 9.49538112e-01 -4.92029697e-01 -9.73414928e-02
2.89201558e-01 2.50624008e-02 1.76600599e+00 -6.01327479e-01
6.71787620e-01 1.47326064e+00 9.07721043e-01 7.53421709e-02
-1.48926032e+00 -6.05046928e-01 3.59793603e-01 4.78363395e-01
-9.65670347e-01 8.70774090e-02 1.29585707e+00 -5.44418991e-01
8.42039406e-01 2.33648255e-01 9.23557818e-01 9.73617077e-01
5.14663041e-01 8.92770112e-01 8.91184688e-01 2.29440387e-02
-6.52206913e-02 -6.61503104e-03 4.00279790e-01 1.09033823e+00
6.21472061e-01 7.83167034e-02 -6.49100006e-01 -3.60759976e-03
4.13629025e-01 3.03951442e-01 -2.24194214e-01 -1.11835015e+00
-9.82029021e-01 8.59684169e-01 1.02926755e+00 7.88279176e-02
-2.94424683e-01 1.44277722e-01 7.72106528e-01 4.42618161e-01
6.30998671e-01 7.10932136e-01 -2.82022834e-01 3.05299997e-01
-4.95119184e-01 4.09542285e-02 5.70767045e-01 8.65396500e-01
8.53868127e-01 2.86710024e-01 -3.89409631e-01 9.41274643e-01
2.04799578e-01 3.09524864e-01 1.60279736e-01 -2.62944788e-01
6.46213055e-01 1.51876760e+00 -6.27323091e-01 -1.55709851e+00
-6.99311912e-01 -7.58316040e-01 -9.35771525e-01 -1.89685673e-01
2.02877969e-01 3.85784179e-01 -8.43638897e-01 1.30370677e+00
3.33561033e-01 -7.59635195e-02 -2.55338669e-01 6.70407593e-01
1.28671122e+00 4.82352406e-01 -9.82446745e-02 1.00995466e-01
1.03226840e+00 -8.89742494e-01 -9.25241232e-01 1.02826450e-02
1.04983354e+00 -1.86122492e-01 1.15430498e+00 3.71344209e-01
-5.76967895e-01 -3.00728291e-01 -9.29659545e-01 -1.17555760e-01
-8.42047274e-01 -3.07956904e-01 1.10382271e+00 6.06519938e-01
-9.41520631e-01 7.55385458e-01 -3.60407680e-01 -1.85300320e-01
7.88819790e-01 3.51950228e-01 -5.99038959e-01 -2.30678901e-01
-1.12231290e+00 6.08968198e-01 5.73698044e-01 3.28724831e-01
-5.14809370e-01 -7.44638741e-01 -1.17934251e+00 5.11010110e-01
6.96590126e-01 -5.45932710e-01 5.65831125e-01 -5.69851220e-01
-7.79647112e-01 4.59925443e-01 1.64848149e-01 -5.74937701e-01
7.44271949e-02 1.29042715e-01 -4.71440792e-01 -6.10640086e-02
-8.52377247e-03 1.91369116e-01 7.55680501e-01 -1.46737146e+00
-2.68638700e-01 -5.71027040e-01 4.92867790e-02 1.68552652e-01
-6.67059422e-01 -8.85617495e-01 -2.82841653e-01 -5.06224751e-01
1.98862940e-01 -4.78994280e-01 -1.81548875e-02 -1.65175319e-01
-7.81782269e-01 -4.77635771e-01 1.00199556e+00 -4.54629511e-01
1.40152276e+00 -1.81106234e+00 3.06018498e-02 3.84397507e-01
9.52446640e-01 4.44005542e-02 -3.55086863e-01 8.00126672e-01
-3.65405172e-01 1.74759194e-01 4.48848456e-02 -1.26670375e-01
2.37734377e-01 1.69754103e-01 -3.70424777e-01 3.81066322e-01
3.50014508e-01 1.49580956e+00 -1.02740073e+00 -3.68160039e-01
4.80975240e-01 3.36166054e-01 -5.87914765e-01 1.38689265e-01
-3.73924404e-01 -7.33351335e-02 -2.75901288e-01 1.00461161e+00
6.87932432e-01 -6.92553699e-01 6.30145490e-01 -4.91024196e-01
4.15986121e-01 6.12287104e-01 -7.56907046e-01 1.19375670e+00
-3.86287808e-01 5.36263585e-01 -3.68950367e-01 -1.48312283e+00
1.18422139e+00 -3.10621802e-02 6.47962272e-01 -1.17335463e+00
-1.36329727e-02 -9.36115310e-02 -5.19091859e-02 -2.15707570e-01
4.60310608e-01 5.74259833e-02 9.35707707e-03 2.47164279e-01
2.35106707e-01 5.81481913e-03 3.35089475e-01 4.98650342e-01
1.10493350e+00 -4.54731323e-02 2.76377082e-01 -2.01068431e-01
2.97472000e-01 1.00897014e-01 2.85478085e-01 5.19763708e-01
1.53269365e-01 2.67284691e-01 1.32394326e+00 -7.95701623e-01
-7.22842693e-01 -9.75241244e-01 -7.60884359e-02 7.66004801e-01
2.15723485e-01 -9.52744961e-01 -2.49101415e-01 -1.22254622e+00
5.08171797e-01 5.01636446e-01 -1.07257020e+00 -3.67129922e-01
-3.17213297e-01 -6.79838896e-01 5.70160635e-02 5.08415580e-01
2.03078657e-01 -1.16008139e+00 3.08549017e-01 1.17539391e-01
1.55663058e-01 -9.30208623e-01 -2.50746876e-01 5.67986906e-01
-1.10741007e+00 -1.45844769e+00 2.45647714e-01 -6.15762711e-01
6.86754286e-01 4.47372139e-01 1.56060815e+00 3.87512833e-01
-2.32487112e-01 2.68356621e-01 -2.94661433e-01 1.03398669e-03
-1.81669801e-01 2.09396973e-01 -4.42554533e-01 -4.96001206e-02
3.54383349e-01 -6.01824880e-01 -3.51106495e-01 6.53386563e-02
-6.36445045e-01 8.73627365e-02 5.74573159e-01 1.42699957e+00
5.01453042e-01 1.64596066e-01 7.04051137e-01 -1.35109520e+00
8.37270439e-01 -7.21022666e-01 -8.49641144e-01 3.99804145e-01
-1.03133321e+00 3.61653924e-01 9.99189496e-01 1.67442277e-01
-4.12984818e-01 -2.86148369e-01 1.71598658e-01 -6.39903605e-01
1.80111572e-01 1.13316011e+00 -3.53827566e-01 1.47795752e-01
4.13250953e-01 2.24716485e-01 1.45787522e-01 -2.22217605e-01
4.53139156e-01 1.93687007e-01 2.27536753e-01 -4.31376040e-01
9.33577478e-01 3.26225311e-01 3.67294222e-01 -3.94795418e-01
-1.04311621e+00 -4.82937843e-01 -7.55247116e-01 -3.63872111e-01
5.24001479e-01 -6.41017318e-01 -1.00092292e+00 -6.22833073e-02
-6.82909608e-01 -2.10445166e-01 -1.65849984e-01 -1.11933269e-01
-2.62840092e-01 3.35218579e-01 -3.07127357e-01 -5.10610700e-01
-3.57927024e-01 -8.96244943e-01 1.13712752e+00 -2.53855854e-01
-3.08506773e-03 -1.66649234e+00 -1.25812348e-02 5.33607543e-01
6.17467836e-02 5.06499052e-01 1.47502887e+00 -8.24515581e-01
-9.05168056e-01 -3.80100310e-01 -5.99425316e-01 -4.80682738e-02
5.11308849e-01 -1.62180841e-01 -7.48098791e-01 -4.05412257e-01
-7.12155759e-01 -4.29562420e-01 9.66210306e-01 3.14753890e-01
1.55421317e+00 -4.79243845e-01 -3.71688098e-01 8.18022430e-01
1.36985826e+00 1.86010957e-01 4.50745046e-01 3.75642151e-01
1.43206489e+00 6.77107096e-01 4.34046745e-01 2.52986193e-01
4.23921764e-01 4.40207809e-01 1.04279387e+00 -2.79173762e-01
-1.24442838e-01 -6.80733979e-01 -4.64406461e-02 8.55596960e-01
4.25994605e-01 -4.49603438e-01 -1.03331757e+00 4.28703159e-01
-2.05663323e+00 -8.53007793e-01 -4.12097752e-01 1.95665300e+00
3.50892872e-01 3.76664013e-01 2.20239405e-02 3.13595951e-01
5.02277374e-01 5.24574757e-01 -6.15552306e-01 -3.59662086e-01
-1.56648338e-01 1.64208099e-01 3.80051970e-01 2.01854452e-01
-1.05557847e+00 8.92716825e-01 5.52392483e+00 6.71453059e-01
-1.10924268e+00 -4.02380794e-01 7.37134516e-01 1.85046792e-01
-7.56841421e-01 -1.78145885e-01 -5.05670309e-01 1.47864357e-01
6.90583527e-01 -3.23874682e-01 5.23923516e-01 9.53601122e-01
-3.20746452e-02 4.90249693e-01 -1.26776505e+00 1.06329334e+00
7.17505217e-02 -1.84034562e+00 6.13946855e-01 1.28691003e-01
4.87000763e-01 -3.18785667e-01 2.56311983e-01 6.79689646e-01
4.67799723e-01 -1.42884624e+00 3.45720798e-01 2.23951891e-01
5.19434810e-01 -9.61191297e-01 8.37327719e-01 -2.24330381e-01
-1.51887071e+00 -1.89228997e-01 -4.26212668e-01 3.28100212e-02
-3.37080896e-01 7.37244606e-01 -1.25196278e+00 1.18305695e+00
6.98720396e-01 1.64432001e+00 -1.04596114e+00 9.19076145e-01
-1.81115404e-01 6.48122013e-01 2.28329018e-01 -2.19779044e-01
4.02789891e-01 -4.23710197e-01 2.43635297e-01 9.75633681e-01
3.30141746e-02 -4.60432947e-01 1.16080143e-01 1.07736063e+00
-4.47599381e-01 2.95913547e-01 -1.03966939e+00 -5.38762748e-01
3.97446275e-01 1.59207439e+00 -7.15481818e-01 -2.05896944e-01
-6.32358253e-01 2.60152251e-01 8.79595995e-01 2.01615661e-01
-6.35317981e-01 -1.68730766e-01 4.51986849e-01 3.57884407e-01
2.99161553e-01 -9.08146948e-02 -6.38389289e-01 -1.14871991e+00
3.95652167e-02 -1.13174462e+00 8.49058449e-01 -6.36631370e-01
-1.60220003e+00 6.04257643e-01 -2.36497506e-01 -1.24700546e+00
9.62873623e-02 -8.53138268e-01 -3.94536883e-01 5.62119186e-01
-1.41539586e+00 -1.36771047e+00 -4.90752369e-01 6.05089903e-01
1.52166665e-01 -3.13927978e-01 6.16854906e-01 1.44156635e-01
-6.64148271e-01 7.37469912e-01 1.98164716e-01 2.13935748e-01
3.94991189e-01 -1.77697480e+00 5.15226066e-01 5.52745521e-01
5.13933003e-01 6.87747657e-01 4.26196605e-01 -8.23750257e-01
-1.84592593e+00 -1.45467329e+00 6.86879337e-01 -5.21378160e-01
1.15398359e+00 -8.60249937e-01 -1.13910186e+00 6.80993736e-01
2.06225410e-01 2.45853677e-01 5.39224863e-01 7.74439991e-01
-6.99301243e-01 -4.24473673e-01 -6.88345909e-01 7.64476478e-01
1.18338954e+00 -5.95505714e-01 -3.10376823e-01 3.95395577e-01
6.88740730e-01 -1.26181915e-01 -1.02295983e+00 4.45505023e-01
3.45481962e-01 -8.92169237e-01 7.88050175e-01 -7.77653992e-01
6.45425856e-01 -1.30856514e-01 5.86180165e-02 -1.69467092e+00
-3.77596080e-01 -2.80525774e-01 -4.65552300e-01 9.87091124e-01
4.36019897e-01 -7.79899299e-01 9.53380585e-01 1.95721358e-01
-2.83290386e-01 -8.94694746e-01 -7.23242998e-01 -7.66998410e-01
-2.95049269e-02 -2.17410639e-01 8.57811630e-01 1.40253305e+00
3.33145708e-01 6.26636922e-01 -4.29557264e-01 7.90551007e-02
7.00776339e-01 5.81333160e-01 9.70543683e-01 -1.48195350e+00
2.65792594e-03 -6.43279850e-01 -7.25734651e-01 -6.29238427e-01
3.91607016e-01 -1.59514058e+00 -4.19474900e-01 -2.08901811e+00
2.24877208e-01 -4.81024414e-01 -5.52260101e-01 8.43075931e-01
-3.69287908e-01 2.52612904e-02 7.08281249e-03 -2.62035549e-01
-7.96467006e-01 9.32575762e-01 1.57991409e+00 -7.51545370e-01
5.66874929e-02 -4.87970024e-01 -7.82806098e-01 7.95237496e-02
5.48833311e-01 -4.10115033e-01 -7.52381384e-01 -1.52050003e-01
5.98096550e-01 2.17011601e-01 2.96798438e-01 -6.94997549e-01
3.77903320e-02 -1.13008153e-02 2.14676380e-01 -9.54158962e-01
4.08744030e-02 -1.01574576e+00 1.47932023e-01 2.82853991e-01
-3.45953643e-01 1.28662363e-01 4.59318668e-01 6.75682664e-01
-5.09981573e-01 3.32683861e-01 2.31456727e-01 2.23874003e-01
-6.18542910e-01 7.00899780e-01 1.70304358e-01 -9.85420644e-02
6.69308960e-01 -2.87158787e-01 -8.52911890e-01 -3.64431530e-01
-7.15460241e-01 4.92499202e-01 1.78256631e-01 6.90275013e-01
7.46832609e-01 -1.64010906e+00 -4.15520072e-01 3.40523630e-01
5.53529084e-01 -7.54693523e-02 8.20517018e-02 7.07109094e-01
-3.76116514e-01 5.17023027e-01 -2.27251291e-01 -7.03424394e-01
-1.20115757e+00 1.06597686e+00 3.56623262e-01 -6.50552332e-01
-7.19624162e-01 5.52471101e-01 4.72582221e-01 -8.13487470e-01
1.10665627e-01 -2.61798233e-01 -3.61221611e-01 3.53405595e-01
-1.96897253e-01 6.83559105e-02 4.37924683e-01 -1.54070631e-01
-2.30067462e-01 1.59998700e-01 -3.56171519e-01 7.79598236e-01
1.51072049e+00 2.14750141e-01 -3.59520376e-01 4.90205526e-01
1.08873916e+00 -4.24785092e-02 -9.56795096e-01 -1.63772851e-01
3.51615310e-01 -4.83007222e-01 -7.01757148e-03 -6.99569821e-01
-1.51419997e+00 1.03123665e+00 4.20430079e-02 7.21887946e-01
8.78251612e-01 -6.27516881e-02 4.36826050e-01 5.23921788e-01
2.46197835e-01 -5.80380738e-01 4.50779885e-01 4.49135661e-01
9.30123508e-01 -1.37763643e+00 1.97370514e-01 -5.26850879e-01
-3.39886934e-01 1.23962235e+00 7.70662546e-01 -1.97842270e-01
6.10870063e-01 3.81224118e-02 -1.05928563e-01 -8.39712441e-01
-1.16288781e+00 -1.69992894e-01 6.86853945e-01 6.10940814e-01
5.85209846e-01 8.71663839e-02 -4.61217389e-03 4.36901748e-01
-4.17775691e-01 -5.82432687e-01 2.56018460e-01 4.88988459e-01
-1.15194693e-01 -9.46349382e-01 -1.61327407e-01 1.04171026e+00
4.39529382e-02 -4.80604231e-01 -7.49666631e-01 1.18655109e+00
-4.24450487e-01 7.41615891e-01 2.13212948e-02 -6.49290383e-01
2.90637344e-01 -7.33373985e-02 2.11378634e-01 -5.80291450e-01
-6.62276268e-01 -2.55151421e-01 2.17524663e-01 -7.86068976e-01
1.88536607e-02 -3.95241529e-01 -8.78768981e-01 -5.45497417e-01
-3.65470380e-01 1.80243373e-01 1.23965994e-01 6.24855995e-01
3.63382459e-01 1.06487787e+00 6.04743779e-01 -3.61309111e-01
-1.45256922e-01 -7.62565136e-01 -7.93686628e-01 6.67678773e-01
2.50405729e-01 -9.44408238e-01 -2.97138095e-01 -3.14453900e-01] | [7.102985382080078, 6.286863803863525] |
fd5a3b81-97c3-4c5b-a92d-24a28a576241 | increasingly-packing-multiple-facial | null | null | https://dl.acm.org/doi/10.1145/3323873.3325053 | https://dl.acm.org/doi/pdf/10.1145/3323873.3325053 | Increasingly Packing Multiple Facial-Informatics Modules in A Unified Deep-Learning Model via Lifelong Learning | Simultaneously running multiple modules is a key requirement for a smart multimedia system for facial applications including face recognition, facial expression understanding, and gender identification. To effectively integrate them, a continual learning approach to learn new tasks without forgetting is introduced. Unlike previous methods growing monotonically in size, our approach maintains the compactness in continual learning. The proposed packing-and-expanding method is effective and easy to implement, which can iteratively shrink and enlarge the model to integrate new functions. Our integrated multitask model can achieve similar accuracy with only 39.9% of the original size. | ['Chu-Song Chen', 'Yi-Ming Chan', 'Chein-Hung Chen', 'Jia-Hong Lee', 'Timmy S. T. Wan', 'Steven C. Y. Hung'] | 2019-06-10 | null | null | null | proceedings-of-the-2019-on-international | ['age-and-gender-classification', 'gender-prediction'] | ['computer-vision', 'computer-vision'] | [-2.41753105e-02 -1.24304138e-01 -2.80871391e-01 -5.28238714e-01
-5.65855801e-01 -2.69900560e-01 -1.28695220e-01 -1.58534095e-01
-5.95462382e-01 1.00852752e+00 -5.29916584e-01 -1.27139941e-01
-1.49178401e-01 -5.38067937e-01 -6.85806811e-01 -8.13845932e-01
-8.63914341e-02 4.63545710e-01 3.33143860e-01 7.80562758e-02
-3.26091573e-02 3.38128120e-01 -1.86343122e+00 2.16881290e-01
8.15575182e-01 8.55266631e-01 5.07742465e-01 5.83356678e-01
-3.74040067e-01 4.65475410e-01 -5.17010987e-01 -4.15564358e-01
1.05036460e-01 2.54894406e-01 -7.22338438e-01 4.01772350e-01
3.36656630e-01 -3.04893821e-01 -2.20729306e-01 8.37383449e-01
5.11650503e-01 2.48500377e-01 2.28749990e-01 -1.42644119e+00
-1.73139855e-01 2.93112040e-01 -1.08225620e+00 4.88090143e-02
6.03149496e-02 -2.38016635e-01 5.05532444e-01 -1.09223664e+00
2.41840407e-01 1.24153233e+00 6.35434031e-01 7.22384393e-01
-8.02924514e-01 -1.03207552e+00 5.00555277e-01 2.57154256e-01
-1.65369070e+00 -7.96977043e-01 5.76330483e-01 8.45611375e-03
8.62693012e-01 3.98776472e-01 6.06436253e-01 5.12448907e-01
4.15407009e-02 7.86402404e-01 8.12051952e-01 -2.54563123e-01
8.62868428e-02 3.41769844e-01 -1.44115239e-01 1.01948440e+00
2.67052472e-01 -6.01082027e-01 -8.09554577e-01 -1.77093983e-01
8.39620471e-01 3.95437658e-01 2.48531789e-01 -2.13650212e-01
-5.71425736e-01 5.81017256e-01 -3.26960713e-01 3.06016225e-02
2.60320276e-01 2.97381461e-01 4.92669195e-01 5.82488120e-01
7.65828192e-01 -4.08355035e-02 -9.29548800e-01 -4.07836884e-01
-9.54263747e-01 1.58465747e-02 6.56980574e-01 1.05563104e+00
1.06657958e+00 1.22887380e-01 2.74481773e-01 1.16557217e+00
1.38225362e-01 6.76923573e-01 5.50055265e-01 -9.91078436e-01
1.35144457e-01 4.74997759e-01 -3.03098559e-01 -7.09273517e-01
-6.13081992e-01 -3.83869261e-01 -7.73575246e-01 -1.57151431e-01
1.07812978e-01 -3.27931583e-01 -7.22083151e-01 1.56946039e+00
6.48449779e-01 5.72427452e-01 -3.84951085e-01 2.26806328e-01
7.46652722e-01 6.14554107e-01 9.74217430e-02 -7.37240255e-01
1.09800780e+00 -1.06538534e+00 -6.88719451e-01 -1.64033338e-01
6.21206939e-01 -7.59379804e-01 9.61147964e-01 5.72195411e-01
-1.19871902e+00 -6.46759033e-01 -9.61130261e-01 1.20361492e-01
-1.90204650e-01 9.10772979e-02 1.10130489e+00 1.07103848e+00
-1.15405488e+00 3.36506665e-01 -7.84675121e-01 -9.22512487e-02
7.61856616e-01 8.94187331e-01 -3.75353426e-01 -1.30924806e-01
-6.33454025e-01 5.61712086e-01 4.25733507e-01 -1.94960266e-01
-5.51110506e-01 -7.28884518e-01 -8.56714725e-01 1.38129383e-01
6.22959256e-01 -4.83292609e-01 1.20567727e+00 -9.10680771e-01
-1.77357185e+00 7.54778981e-01 -7.33406126e-01 -9.61879194e-02
1.64736435e-01 -2.02495411e-01 -4.34916973e-01 1.22399703e-02
-2.24748284e-01 7.54614711e-01 1.45068932e+00 -7.84634233e-01
-7.17241287e-01 -5.24658561e-01 -1.49396777e-01 3.85776252e-01
-1.10479140e+00 -1.56116143e-01 -7.71373630e-01 -4.39233005e-01
1.43503994e-01 -7.21103132e-01 -5.79793155e-02 1.89087495e-01
3.84101063e-01 -3.61008108e-01 1.20979095e+00 -3.68742853e-01
1.53692257e+00 -2.23261023e+00 -1.44925844e-02 5.12519740e-02
5.21817744e-01 2.75839835e-01 -2.96347648e-01 -1.89644933e-01
1.22369654e-01 -4.67592292e-02 1.94803700e-01 -6.88532114e-01
-3.01658481e-01 2.55255073e-01 6.68760315e-02 3.09228033e-01
4.69382741e-02 8.47459316e-01 -5.99277794e-01 -6.96225107e-01
-2.28705481e-02 1.68561369e-01 -6.76320791e-01 -9.45791136e-03
-8.47308189e-02 2.49204021e-02 -1.63370207e-01 9.81253088e-01
1.13436306e+00 -2.95712203e-01 2.16704145e-01 6.73234984e-02
2.25288212e-01 -2.14679301e-01 -1.35334873e+00 1.79046810e+00
-6.61017656e-01 3.73362213e-01 2.42896840e-01 -1.14398921e+00
9.57929909e-01 1.76657230e-01 7.77049243e-01 -7.38014519e-01
-8.23201984e-02 1.81402177e-01 -3.93530637e-01 -5.45213997e-01
3.20903510e-01 -1.74885735e-01 1.71255041e-02 5.38405418e-01
4.11729246e-01 1.25061542e-01 9.48463157e-02 -3.74117941e-02
9.18075144e-01 -2.13325560e-01 3.63413811e-01 -1.69141233e-01
6.42950058e-01 -5.57911038e-01 7.11193502e-01 3.86977196e-01
-2.05131263e-01 1.19585268e-01 2.16652572e-01 -8.40480208e-01
-7.15672791e-01 -8.17885041e-01 -1.00486934e-01 1.80205524e+00
5.33258393e-02 -4.38724399e-01 -5.78772545e-01 -7.45801866e-01
1.76947311e-01 6.14538342e-02 -4.18262869e-01 -1.45609379e-01
-5.68179011e-01 -8.04777205e-01 2.74519891e-01 3.30246627e-01
6.51889622e-01 -6.51729941e-01 -4.04012471e-01 1.51605114e-01
-1.45842448e-01 -9.13390875e-01 -7.43178785e-01 1.65356383e-01
-1.22673023e+00 -9.44884360e-01 -6.31237328e-01 -1.09684360e+00
7.24910915e-01 5.71093321e-01 7.69845665e-01 2.59038270e-01
-4.32101130e-01 4.42960888e-01 9.89192501e-02 -3.30223918e-01
1.25283837e-01 4.57653522e-01 2.29032353e-01 6.62488192e-02
3.11953098e-01 -6.02294743e-01 -3.19201678e-01 3.49849403e-01
-8.99133325e-01 -6.23588357e-03 5.37735224e-01 9.83220816e-01
3.91674489e-01 2.09822029e-01 9.32489753e-01 -1.05059314e+00
3.74075115e-01 -4.27873462e-01 -5.08870959e-01 5.26993275e-01
-7.16424644e-01 -7.45542720e-02 4.40427661e-01 -9.03611362e-01
-1.16764331e+00 3.52875859e-01 -1.62962321e-02 -4.25790757e-01
1.73004076e-01 2.35569537e-01 -2.26889014e-01 -5.44874251e-01
1.92379653e-01 2.24912629e-01 2.29644626e-01 -3.82929921e-01
2.98549682e-01 6.15823269e-01 1.28771544e-01 -5.26926279e-01
5.50319612e-01 3.78307104e-01 1.10919893e-01 -9.85843122e-01
-5.36836922e-01 -4.75033492e-01 -6.43904150e-01 -4.78083879e-01
-1.86132528e-02 -1.01566374e+00 -1.23367393e+00 7.12456703e-01
-1.04174292e+00 -2.31437996e-01 -6.42227829e-02 1.30609006e-01
-3.23881805e-01 3.69253427e-01 -4.53973472e-01 -8.45715404e-01
-2.86680490e-01 -8.71577561e-01 1.08471656e+00 4.78020310e-01
8.18252042e-02 -1.14228487e+00 -1.66191131e-01 4.53185320e-01
5.63020706e-01 -4.79580671e-01 8.37714553e-01 -2.72314548e-01
-4.94426072e-01 -4.34916541e-02 -1.38859376e-01 1.31776318e-01
2.99367577e-01 -3.15269023e-01 -9.93778288e-01 -6.46229804e-01
1.67170316e-01 -5.74482441e-01 9.90162373e-01 2.71313548e-01
1.71472800e+00 -3.03219140e-01 -4.95982409e-01 7.50333607e-01
1.19451427e+00 3.45944852e-01 5.70719779e-01 6.63911924e-02
4.27295595e-01 3.12311798e-01 5.95809758e-01 8.44516098e-01
3.29324484e-01 3.28395516e-01 4.63595130e-02 -1.63866580e-01
1.01821169e-01 4.54815254e-02 3.84493500e-01 1.04356980e+00
1.44668579e-01 2.08000898e-01 -6.57191396e-01 2.47672319e-01
-1.95902479e+00 -7.67962456e-01 5.67330658e-01 2.19352150e+00
1.14545357e+00 9.19697881e-02 3.68743837e-02 -7.51203522e-02
6.50276482e-01 -1.94394104e-02 -8.17347527e-01 -3.48630071e-01
1.09999478e-01 5.78632176e-01 4.29178089e-01 4.92045313e-01
-1.20033288e+00 1.15744746e+00 7.43816042e+00 1.41161680e+00
-1.25887811e+00 4.91467595e-01 9.17638838e-01 -4.97386694e-01
-7.08188489e-02 -3.68967354e-01 -1.18195415e+00 2.73899525e-01
7.69007504e-01 -4.35848504e-01 5.01937628e-01 1.02570069e+00
-3.51987071e-02 -3.49682987e-01 -7.95089602e-01 1.46921825e+00
4.34419483e-01 -1.24486041e+00 1.06685624e-01 -2.91219324e-01
7.52023339e-01 -6.37300909e-01 3.72703224e-01 6.52426720e-01
-3.58556658e-01 -1.03332710e+00 4.07961421e-02 5.41159928e-01
1.19521260e+00 -9.46705103e-01 5.49816430e-01 2.84900427e-01
-1.50464714e+00 -1.82322189e-01 -5.21994650e-01 -1.51677936e-01
-1.20179690e-01 5.23904383e-01 -7.10262179e-01 8.20035040e-02
7.35227942e-01 4.19022262e-01 -7.06956625e-01 1.13332450e+00
5.35295606e-01 4.37255055e-01 -2.91992217e-01 -1.84305564e-01
-3.67291242e-01 2.27678660e-02 1.98218256e-01 9.06105995e-01
4.21683192e-01 -7.34484270e-02 4.45321172e-01 6.57150671e-02
-3.14673752e-01 3.91833991e-01 -4.50300008e-01 2.17164680e-01
5.69464386e-01 1.30459416e+00 -7.11066544e-01 -3.01799536e-01
-6.53155744e-01 9.96102273e-01 4.95101333e-01 2.72451639e-01
-1.00275040e+00 -4.22638386e-01 5.74159145e-01 1.05807900e-01
4.01038975e-01 -3.75032604e-01 -3.28005910e-01 -9.50468481e-01
1.51474506e-01 -7.41500974e-01 3.84761900e-01 -3.23390871e-01
-9.51522946e-01 3.44713628e-01 -1.66530889e-02 -9.68756258e-01
1.33120283e-01 -5.22925079e-01 -3.41976136e-01 2.91558146e-01
-1.53873944e+00 -1.03684354e+00 -3.25108737e-01 8.57345700e-01
9.70666230e-01 -5.79790592e-01 8.57133806e-01 6.49380684e-01
-7.08077192e-01 1.08855414e+00 1.29556447e-01 -5.13629436e-01
1.07056546e+00 -7.15978205e-01 -1.90389976e-01 2.63545722e-01
5.19280620e-02 5.77762842e-01 2.49595836e-01 -4.98979121e-01
-1.56182921e+00 -1.24832451e+00 6.13247573e-01 1.46814585e-01
5.24042070e-01 -4.03558224e-01 -8.67838264e-01 4.18586522e-01
-8.08990970e-02 -6.04400784e-03 1.04732490e+00 4.29882497e-01
-3.34411412e-01 -8.97844255e-01 -1.05361199e+00 3.30687076e-01
1.11004114e+00 -2.40309820e-01 -1.69104803e-02 4.72503930e-01
8.04330528e-01 -3.71239811e-01 -5.33332288e-01 4.82307494e-01
8.33446383e-01 -5.52144349e-01 7.52166092e-01 -4.29674655e-01
-5.31119630e-02 9.48008671e-02 1.44795641e-01 -7.95437455e-01
-2.17156202e-01 -1.02921677e+00 -6.40691519e-01 9.48909402e-01
4.37684476e-01 -7.92324305e-01 1.10567534e+00 5.55544376e-01
1.22650631e-01 -1.08107460e+00 -1.25081694e+00 -8.40405583e-01
-1.84279710e-01 -3.70767623e-01 5.53257227e-01 7.16367841e-01
1.39457747e-01 2.65751541e-01 -5.83061755e-01 -1.37776420e-01
5.34672797e-01 -2.48466115e-02 7.35300422e-01 -1.34602249e+00
-3.92501444e-01 -3.81312042e-01 -2.22202584e-01 -1.36876142e+00
1.32247463e-01 -6.93500221e-01 -3.38141739e-01 -8.86103451e-01
6.71669841e-01 -6.05724990e-01 -3.89025509e-01 7.27002800e-01
-2.25226760e-01 3.82253677e-01 -7.10878074e-02 1.02291629e-01
-1.28689396e+00 7.62741327e-01 1.16245759e+00 -2.63399333e-01
-3.97347897e-01 3.52661967e-01 -6.87042475e-01 8.12102020e-01
9.08901036e-01 -4.98779714e-01 -5.26778519e-01 -4.29595619e-01
3.05740416e-01 -9.56537798e-02 -2.59757906e-01 -9.35333014e-01
5.82099795e-01 -1.84816346e-01 3.17557126e-01 -4.35715705e-01
4.80299592e-01 -7.68519461e-01 -1.19563520e-01 3.45097274e-01
5.98126613e-02 6.55385926e-02 6.57240391e-01 5.64606130e-01
1.07013108e-02 -2.04167530e-01 6.17633879e-01 -1.43463099e-02
-7.86821008e-01 6.24267459e-01 -4.17714447e-01 -7.31837898e-02
1.43285143e+00 -2.09731936e-01 -1.52959347e-01 -2.98720658e-01
-1.04920208e+00 5.16068816e-01 9.84132662e-03 4.58727300e-01
9.84974802e-01 -1.32245314e+00 -3.81528407e-01 3.84829700e-01
-2.12916151e-01 -1.17059357e-01 5.96848786e-01 7.72423506e-01
-2.74840057e-01 3.37531716e-01 -1.88289091e-01 -5.29963672e-01
-1.94506025e+00 3.31361294e-01 2.52195746e-02 -2.54666150e-01
-1.31278429e-02 1.02739179e+00 -7.58982599e-02 -4.01902258e-01
6.58946812e-01 9.62897241e-02 -1.94704011e-01 1.29085094e-01
6.18978024e-01 5.95651865e-01 7.74673223e-02 4.74453494e-02
-2.14648813e-01 4.91157532e-01 -4.19285625e-01 1.12745084e-01
1.35835075e+00 -2.78054357e-01 -4.34107721e-01 6.14317000e-01
1.16003704e+00 -1.22147702e-01 -1.29224873e+00 -4.24672186e-01
-1.87309265e-01 -5.73830426e-01 5.13462350e-03 -3.13499063e-01
-1.13911974e+00 8.47728789e-01 6.36698723e-01 -1.67348713e-01
1.38507628e+00 -1.42565474e-01 8.20797741e-01 8.28432381e-01
6.75492227e-01 -1.44128466e+00 6.09424055e-01 5.31854212e-01
4.73943412e-01 -1.26477969e+00 3.60801101e-01 -5.29326379e-01
-4.05415386e-01 1.31888831e+00 1.02811658e+00 4.52770531e-01
1.05165637e+00 5.40076375e-01 -4.51195478e-01 1.15184635e-01
-1.02305579e+00 -8.90411362e-02 4.78268228e-02 5.80629766e-01
3.16981077e-01 -1.68578833e-01 -3.30886930e-01 7.57762790e-01
1.28931463e-01 2.89339777e-02 2.48150676e-01 9.46695685e-01
-7.75086164e-01 -1.21689367e+00 -1.22142971e-01 6.66351378e-01
-1.70422181e-01 -5.76485805e-02 2.62383252e-01 5.68493962e-01
3.48478764e-01 7.59071827e-01 2.51113832e-01 -4.90940064e-01
-1.88601702e-01 3.79940361e-01 7.58361042e-01 -7.51468003e-01
2.65840138e-03 1.38160080e-01 -3.13940316e-01 -3.78342718e-01
-4.72043544e-01 -6.60927653e-01 -1.06826591e+00 -5.22890270e-01
-5.12902319e-01 -9.64964926e-02 4.97354686e-01 9.53675985e-01
4.94130254e-01 3.72056484e-01 1.07437074e+00 -4.45701331e-01
-3.95912975e-01 -8.26832592e-01 -6.02278590e-01 -1.80399776e-01
7.10069239e-02 -7.07378209e-01 -5.37579954e-02 1.66930303e-01] | [9.711394309997559, 3.339000940322876] |
1bfe0e6c-528a-4239-bbfc-f7421d21e7b0 | probing-deep-speaker-embeddings-for-speaker | 2212.07068 | null | https://arxiv.org/abs/2212.07068v1 | https://arxiv.org/pdf/2212.07068v1.pdf | Probing Deep Speaker Embeddings for Speaker-related Tasks | Deep speaker embeddings have shown promising results in speaker recognition, as well as in other speaker-related tasks. However, some issues are still under explored, for instance, the information encoded in these representations and their influence on downstream tasks. Four deep speaker embeddings are studied in this paper, namely, d-vector, x-vector, ResNetSE-34 and ECAPA-TDNN. Inspired by human voice mechanisms, we explored possibly encoded information from perspectives of identity, contents and channels; Based on this, experiments were conducted on three categories of speaker-related tasks to further explore impacts of different deep embeddings, including discriminative tasks (speaker verification and diarization), guiding tasks (target speaker detection and extraction) and regulating tasks (multi-speaker text-to-speech). Results show that all deep embeddings encoded channel and content information in addition to speaker identity, but the extent could vary and their performance on speaker-related tasks can be tremendously different: ECAPA-TDNN is dominant in discriminative tasks, and d-vector leads the guiding tasks, while regulating task is less sensitive to the choice of speaker representations. These may benefit future research utilizing speaker embeddings. | ['Rongzhi Gu', 'Junyi Peng', 'Ding Pan', 'Zifeng Zhao'] | 2022-12-14 | null | null | null | null | ['speaker-recognition', 'speaker-verification'] | ['speech', 'speech'] | [-2.12222219e-01 -1.42175183e-01 -6.57957643e-02 -6.80224776e-01
-7.10389197e-01 -5.40678501e-01 8.59735131e-01 7.01548485e-03
-3.04664314e-01 1.11290492e-01 8.77009571e-01 -2.64916956e-01
2.43925117e-02 -3.17159146e-01 -1.95358291e-01 -9.23089147e-01
-1.71739668e-01 5.51326992e-03 -3.18200067e-02 -2.87113279e-01
2.78553069e-01 6.44873440e-01 -1.66006613e+00 2.46906728e-01
2.97756672e-01 9.68663931e-01 -1.64769635e-01 5.45430601e-01
-4.45364237e-01 4.84862894e-01 -7.38583326e-01 -3.87697428e-01
-1.67321563e-02 -2.30878294e-01 -6.09357595e-01 -8.54402408e-02
1.97296500e-01 -1.76196247e-01 -3.97205234e-01 8.94181669e-01
1.07137942e+00 9.94103029e-02 9.17893767e-01 -1.17842865e+00
-1.04360127e+00 9.09735084e-01 -3.65319163e-01 5.33708811e-01
2.84489930e-01 7.67652914e-02 1.00299668e+00 -1.21920240e+00
1.44520491e-01 1.71298516e+00 4.70824867e-01 7.25172341e-01
-1.14522803e+00 -8.94816399e-01 3.41227353e-01 5.28384387e-01
-1.57613683e+00 -1.08661413e+00 9.56673920e-01 -5.58472514e-01
7.13038683e-01 3.64217132e-01 1.60143211e-01 1.51929069e+00
-5.13449125e-02 8.20746422e-01 1.11886632e+00 -3.39274108e-01
1.72223374e-01 6.44525588e-01 3.32451493e-01 -1.74146406e-02
-2.83647120e-01 3.69556755e-01 -7.71640420e-01 -1.57723054e-01
3.73326838e-01 -2.17894241e-01 -4.74369615e-01 3.23294625e-02
-1.11614776e+00 1.15449858e+00 2.68969983e-01 6.14635348e-01
-2.43411183e-01 -3.01198423e-01 6.83270931e-01 4.81865823e-01
2.88873136e-01 1.48585320e-01 -4.41963673e-01 -8.32232237e-02
-8.11225593e-01 1.04861423e-01 6.20088995e-01 6.52577341e-01
7.17350066e-01 4.66037810e-01 -5.64232945e-01 1.29800534e+00
6.48221791e-01 2.08724976e-01 9.50433671e-01 -2.47433186e-01
4.55379426e-01 1.61136463e-01 -3.28521192e-01 -7.78847933e-01
-4.13059860e-01 -3.43819916e-01 -6.35131299e-01 -8.83778408e-02
2.85333037e-01 -2.10106418e-01 -7.95781910e-01 1.82884157e+00
3.35339934e-01 9.31037888e-02 1.10205188e-01 1.18242168e+00
1.29962027e+00 6.84998572e-01 9.61664245e-02 9.66805890e-02
1.78666735e+00 -8.01232040e-01 -8.50149274e-01 -2.79034376e-01
3.31305295e-01 -9.07965422e-01 9.54625130e-01 -6.61818087e-02
-7.20767975e-01 -7.40373313e-01 -9.08537745e-01 -2.19782498e-02
-5.65716922e-01 -7.37513974e-03 2.41957739e-01 1.10098410e+00
-1.26292396e+00 5.30504510e-02 -3.61449152e-01 -3.83120447e-01
2.28498936e-01 3.36315185e-01 -2.92522937e-01 2.27922127e-01
-1.60343564e+00 9.34007466e-01 -8.31516087e-02 1.67333648e-01
-1.10033071e+00 -6.26118481e-01 -9.54793870e-01 2.05720276e-01
-2.43114397e-01 -2.07395464e-01 1.11424828e+00 -7.76144147e-01
-1.74436343e+00 8.26258123e-01 -3.31718981e-01 -3.11442643e-01
1.92085072e-01 2.24835321e-01 -9.29927528e-01 -1.22488752e-01
1.80273771e-03 8.34732652e-01 1.08029699e+00 -1.02322578e+00
-5.67296267e-01 -7.96820283e-01 -1.80685207e-01 8.58637467e-02
-6.92982495e-01 4.18026000e-01 -7.70801753e-02 -7.10143030e-01
7.68162385e-02 -7.45210350e-01 2.69622028e-01 -1.31088167e-01
-4.82312083e-01 -6.44111753e-01 1.02440357e+00 -8.28615189e-01
1.29828703e+00 -2.64424920e+00 3.61964256e-02 -2.73312349e-02
-1.10501572e-01 2.67006457e-01 -3.84869963e-01 4.94874984e-01
-1.72854587e-01 4.10587579e-01 4.98674670e-03 -3.65580857e-01
2.37469092e-01 -4.95683551e-02 -3.01751494e-01 4.96966779e-01
4.36491758e-01 6.12504482e-01 -2.70202696e-01 -3.94440860e-01
9.48869213e-02 8.89259934e-01 -3.33101749e-01 1.30954936e-01
3.50050658e-01 3.99476826e-01 -3.67108047e-01 8.39289784e-01
7.56262898e-01 6.19337499e-01 -2.01901495e-01 -1.49639860e-01
-2.90598512e-01 6.93561494e-01 -1.10335159e+00 1.31804931e+00
-5.94206333e-01 1.11846828e+00 5.68098009e-01 -8.96186173e-01
1.02704823e+00 6.17541075e-01 6.32614419e-02 -7.02928364e-01
1.23816155e-01 -5.16104996e-02 4.91866499e-01 -5.33698857e-01
5.77592671e-01 -3.46341014e-01 9.93310139e-02 3.83496970e-01
5.76301329e-02 3.26355636e-01 -2.52323359e-01 -1.92286994e-03
6.63637340e-01 -5.60007513e-01 -7.31487200e-02 -3.61462861e-01
6.53084934e-01 -5.01298964e-01 4.20261234e-01 3.95675659e-01
-8.02350283e-01 4.84754741e-01 5.46119332e-01 9.17104259e-02
-5.76622307e-01 -9.35646296e-01 -6.59176111e-01 1.64906359e+00
1.72860902e-02 -5.48159033e-02 -5.87613046e-01 -5.00150025e-01
2.40232572e-01 8.83201718e-01 -7.59676397e-01 -2.91498244e-01
-3.67259532e-01 -4.62901056e-01 8.78072739e-01 5.68509042e-01
-3.58877108e-02 -1.13907123e+00 1.07713178e-01 2.00670391e-01
6.18498549e-02 -9.54752624e-01 -9.21008766e-01 2.79216409e-01
-6.28764749e-01 -6.36031151e-01 -9.16399717e-01 -1.02877390e+00
2.73635536e-01 3.41058671e-01 7.59633660e-01 -5.51579893e-01
1.77652705e-02 3.98661345e-01 -4.32715386e-01 -5.48274457e-01
-5.22661150e-01 -1.08294927e-01 3.14777732e-01 4.56262201e-01
6.70825660e-01 -4.25411522e-01 -6.35825217e-01 5.81737936e-01
-6.58191562e-01 -7.47842789e-01 6.21108830e-01 8.54595184e-01
4.77746502e-02 -1.39934823e-01 9.58538055e-01 -3.11647296e-01
9.36086893e-01 -4.76740271e-01 5.92646413e-02 -1.06707796e-01
-2.78466851e-01 -1.91174485e-02 3.29492271e-01 -6.13513470e-01
-1.01182616e+00 -3.87360841e-01 -5.57523429e-01 -3.95082355e-01
-3.44071537e-01 3.68712723e-01 -5.09045124e-01 1.51739597e-01
5.34458697e-01 5.47601640e-01 3.45414668e-01 -5.79097688e-01
3.09310585e-01 1.45123470e+00 -1.42006213e-02 -3.62905443e-01
4.92964864e-01 1.43146858e-01 -9.15228069e-01 -1.26584709e+00
-2.90187269e-01 -6.05014741e-01 -3.80796075e-01 6.77524582e-02
8.82059455e-01 -1.07523656e+00 -4.72245693e-01 4.28431183e-01
-1.20038199e+00 1.44052461e-01 -4.34868746e-02 7.46273398e-01
1.04196660e-01 2.25798160e-01 -6.82434142e-01 -9.54154611e-01
-2.91319072e-01 -1.66292477e+00 1.11478782e+00 4.54326831e-02
-3.48471910e-01 -1.01107049e+00 -1.44815028e-01 4.84862715e-01
8.03829134e-01 -5.03501177e-01 1.08506835e+00 -1.15484405e+00
-2.38451101e-02 -7.86244869e-02 -2.85533667e-01 5.73014915e-01
2.64820069e-01 2.93075833e-02 -1.64456809e+00 -2.24603549e-01
1.53750852e-01 -7.17671961e-02 8.84825289e-01 5.45111537e-01
1.01089275e+00 -2.90483654e-01 -2.00632066e-01 3.39734852e-01
7.83478677e-01 3.03583980e-01 5.71140349e-01 1.81147039e-01
5.49515724e-01 9.62565303e-01 1.04279853e-01 2.68550634e-01
3.99592787e-01 9.30848658e-01 3.53340656e-01 1.42255768e-01
-3.63868475e-01 -1.37989461e-01 9.17308688e-01 9.23776627e-01
3.57401013e-01 -1.11773334e-01 -5.75278223e-01 7.07242191e-01
-1.08270848e+00 -9.50048029e-01 8.52025002e-02 2.06770015e+00
7.40495384e-01 -1.51398823e-01 4.72579509e-01 4.34363455e-01
1.11621583e+00 5.16445160e-01 -5.85987508e-01 -8.25096369e-01
-1.05306797e-01 -2.85273371e-03 4.80242670e-02 2.61106163e-01
-9.29598927e-01 6.79584563e-01 6.48541451e+00 7.33328342e-01
-1.62618744e+00 3.74739289e-01 4.70225334e-01 8.06882698e-03
-4.15774256e-01 -3.82776111e-01 -1.08920383e+00 5.60257912e-01
1.18228948e+00 -1.23774800e-02 1.86183035e-01 8.52320313e-01
3.84710312e-01 5.59736907e-01 -1.34286833e+00 1.06270361e+00
2.39210650e-01 -8.96955788e-01 8.53887126e-02 2.84157217e-01
2.29599461e-01 1.29287198e-01 2.94625938e-01 7.80016541e-01
-2.15926155e-01 -9.24592197e-01 8.41840327e-01 -1.23958208e-01
4.78475571e-01 -6.14332438e-01 6.97427094e-01 9.45038907e-03
-1.16435289e+00 -1.83329925e-01 -2.93826014e-01 1.59949869e-01
7.43990242e-02 4.53589469e-01 -9.20599937e-01 1.36824325e-01
6.24512374e-01 4.85827744e-01 -3.90074193e-01 6.45140707e-01
-6.20482713e-02 9.53101754e-01 3.41384746e-02 -2.71664619e-01
1.50734782e-01 1.03534676e-01 6.66853249e-01 1.45713544e+00
2.85919756e-01 -3.26128662e-01 -1.69463500e-01 9.22900856e-01
-3.75646390e-02 1.79597467e-01 -4.87239420e-01 -2.00355396e-01
8.79829109e-01 1.24971271e+00 -1.92999095e-01 -4.57042009e-02
-4.90901291e-01 6.95760846e-01 3.73523571e-02 5.41239917e-01
-9.36610162e-01 -2.42440224e-01 1.40189826e+00 3.06862354e-01
4.69906449e-01 -7.45276213e-02 -1.40974119e-01 -8.36919606e-01
-7.58159235e-02 -8.69509697e-01 2.23141640e-01 -2.99409866e-01
-1.55381942e+00 6.49412930e-01 -2.76728004e-01 -1.12687719e+00
-8.42322856e-02 -6.63528502e-01 -1.05438745e+00 1.02065265e+00
-1.63653767e+00 -9.65626657e-01 1.26137704e-01 6.88541472e-01
8.07739258e-01 -5.37461996e-01 7.81205535e-01 6.32559597e-01
-9.85179603e-01 1.14046323e+00 1.48230031e-01 4.02067870e-01
9.15342808e-01 -9.38696504e-01 2.85829246e-01 3.96806449e-01
3.13015074e-01 7.64500082e-01 3.72102201e-01 5.17248958e-02
-1.51583791e+00 -1.05372798e+00 7.65372515e-01 -1.69141382e-01
6.32897735e-01 -4.25277442e-01 -9.34072375e-01 4.35163289e-01
3.22913319e-01 -1.21277228e-01 1.09909272e+00 4.10602868e-01
-5.51317573e-01 -1.77936360e-01 -1.05330741e+00 3.72430861e-01
7.79850066e-01 -9.68732893e-01 -4.61636454e-01 -7.53641352e-02
6.92221582e-01 -1.46501902e-02 -8.17318857e-01 2.52256691e-02
5.11183560e-01 -1.02403176e+00 1.02765453e+00 -5.12497783e-01
1.86429203e-01 -1.62022322e-01 -4.09687370e-01 -1.49322248e+00
-5.80772460e-01 -7.99155906e-02 9.80546977e-03 1.90356660e+00
6.43912435e-01 -9.49031055e-01 3.57474953e-01 4.75842834e-01
-3.81835252e-01 -6.66787684e-01 -1.19524753e+00 -8.31679285e-01
2.50141114e-01 -5.00972211e-01 8.91964376e-01 1.10209668e+00
-1.32515281e-01 5.49473882e-01 -2.15564549e-01 3.89020741e-01
2.29546756e-01 -2.07568467e-01 5.73043108e-01 -1.01730120e+00
5.39879091e-02 -8.50660622e-01 -6.69250131e-01 -9.74116445e-01
4.89161074e-01 -1.13271189e+00 -7.71007910e-02 -1.34612107e+00
-1.46666601e-01 -2.91964591e-01 -5.57337403e-01 2.38257021e-01
-9.37338620e-02 -5.27319200e-02 5.02894931e-02 1.28924951e-01
-2.68305019e-02 7.88698435e-01 9.83243406e-01 -4.73852992e-01
-3.07731718e-01 9.48922262e-02 -1.03691649e+00 4.21843767e-01
6.21741533e-01 -2.31704980e-01 -1.69388667e-01 -4.65248644e-01
-7.93260098e-01 7.94539005e-02 1.03224322e-01 -5.15368760e-01
1.07337467e-01 2.21899644e-01 2.96298951e-01 -4.49694753e-01
5.79128623e-01 -5.52363873e-01 -3.32562029e-01 2.25594506e-01
-5.63731074e-01 -1.26111954e-01 3.05523962e-01 4.41449225e-01
-6.09865725e-01 -1.85653925e-01 6.98766828e-01 3.17967027e-01
-8.15647364e-01 1.70837492e-01 -6.64890647e-01 -4.14845236e-02
8.43232810e-01 -4.95190322e-01 -3.23219337e-02 -4.29497510e-01
-7.76589811e-01 -2.37276703e-02 -1.36688769e-01 8.82293820e-01
5.28879523e-01 -1.48401082e+00 -9.48342800e-01 4.60545570e-01
2.89264470e-01 -5.17812848e-01 4.25280601e-01 7.86760390e-01
2.13148862e-01 5.60703814e-01 1.69007555e-02 -8.22353780e-01
-1.21493804e+00 3.74530733e-01 2.60126323e-01 3.39003026e-01
-1.11431293e-01 1.09175658e+00 4.49986488e-01 -5.78374445e-01
6.45090222e-01 -3.08442384e-01 -5.09124100e-01 7.30742276e-01
7.05008924e-01 4.31479573e-01 2.68546611e-01 -1.08532703e+00
-8.62670541e-01 4.52932507e-01 -3.83899987e-01 -1.01467326e-01
1.04491687e+00 -2.70798177e-01 1.06771573e-01 4.22603697e-01
1.53679204e+00 1.00748681e-01 -8.65736961e-01 -2.87389368e-01
-1.26836360e-01 -3.20620447e-01 3.72270942e-01 -5.88324428e-01
-1.30486298e+00 1.31051505e+00 1.00758553e+00 4.34403628e-01
7.19966054e-01 1.26181066e-01 7.19656527e-01 -3.37558538e-01
-8.47185478e-02 -9.90982175e-01 -4.11304608e-02 3.06571633e-01
1.13439655e+00 -1.45839047e+00 -4.91021663e-01 -9.04271677e-02
-8.83112967e-01 9.67020929e-01 5.10305583e-01 4.71247256e-01
1.01134229e+00 9.73858237e-02 4.94528800e-01 -9.64344814e-02
-5.99896073e-01 -2.30529666e-01 3.12653422e-01 7.45478868e-01
1.04311621e+00 2.71377802e-01 -1.33306207e-02 6.47485733e-01
-3.41429681e-01 -7.44939089e-01 1.69656903e-01 4.68439102e-01
-2.16963530e-01 -1.12312651e+00 -8.19114387e-01 4.26653266e-01
-4.54593241e-01 -1.29677117e-01 -6.29376054e-01 5.27936697e-01
-4.58597578e-02 1.33563399e+00 2.07344651e-01 -6.23408794e-01
4.26216990e-01 2.74216354e-01 -3.42158675e-02 -7.56578386e-01
-7.14859843e-01 1.39279142e-01 1.39451712e-01 -3.32174152e-02
-2.11573809e-01 -8.22265387e-01 -1.09546077e+00 -3.03177595e-01
-7.26575553e-01 2.34086424e-01 1.08039081e+00 7.66849160e-01
5.23788929e-01 5.98738790e-01 8.71761560e-01 -1.06720114e+00
-9.20702457e-01 -1.32466090e+00 -7.61829853e-01 3.35035175e-01
5.12856960e-01 -6.90143466e-01 -7.67446637e-01 -3.62789631e-01] | [14.279175758361816, 6.0721001625061035] |
c60400ee-87ba-49c5-9852-edec6c67274a | adversarial-multi-task-deep-learning-for | 2207.01691 | null | https://arxiv.org/abs/2207.01691v1 | https://arxiv.org/pdf/2207.01691v1.pdf | Adversarial Multi-Task Deep Learning for Noise-Robust Voice Activity Detection with Low Algorithmic Delay | Voice Activity Detection (VAD) is an important pre-processing step in a wide variety of speech processing systems. VAD should in a practical application be able to detect speech in both noisy and noise-free environments, while not introducing significant latency. In this work we propose using an adversarial multi-task learning method when training a supervised VAD. The method has been applied to the state-of-the-art VAD Waveform-based Voice Activity Detection. Additionally the performance of the VADis investigated under different algorithmic delays, which is an important factor in latency. Introducing adversarial multi-task learning to the model is observed to increase performance in terms of Area Under Curve (AUC), particularly in noisy environments, while the performance is not degraded at higher SNR levels. The adversarial multi-task learning is only applied in the training phase and thus introduces no additional cost in testing. Furthermore the correlation between performance and algorithmic delays is investigated, and it is observed that the VAD performance degradation is only moderate when lowering the algorithmic delay from 398 ms to 23 ms. | ['Zheng-Hua Tan', 'Peter Koch', 'Claus Meyer Larsen'] | 2022-07-04 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [ 1.59378037e-01 -2.86270946e-01 3.66839349e-01 3.24323088e-01
-1.01134098e+00 -5.92276454e-01 5.75842738e-01 4.69767600e-01
-6.47642672e-01 5.57937503e-01 -6.52045012e-02 -4.00579274e-01
-2.56400257e-01 -2.94065982e-01 -2.82483190e-01 -8.67384017e-01
-2.43141696e-01 2.43993789e-01 5.22336543e-01 -5.29778711e-02
-1.52838439e-01 8.18676054e-01 -1.75711477e+00 1.71655282e-01
5.57690203e-01 7.53399253e-01 3.19166511e-01 1.10950828e+00
-4.91390675e-02 2.44145766e-01 -1.37874854e+00 1.10778287e-01
9.26598012e-02 -3.38057846e-01 -3.70388836e-01 -1.30269483e-01
1.58848196e-01 4.02325168e-02 9.22264829e-02 7.73170471e-01
1.15000403e+00 1.45838857e-01 6.36204898e-01 -1.02311718e+00
6.08219504e-01 2.44279429e-01 -1.04634590e-01 4.68612909e-01
4.52449203e-01 2.38914669e-01 4.77806717e-01 -6.73386991e-01
1.86915725e-01 1.03867519e+00 3.47542077e-01 3.15123558e-01
-1.25656199e+00 -3.93631548e-01 -3.63770813e-01 2.73142129e-01
-1.16798365e+00 -6.19808435e-01 8.66993606e-01 -4.03903812e-01
1.16941440e+00 3.19128752e-01 2.74312198e-01 1.07884526e+00
2.62235880e-01 5.33149838e-01 1.29704916e+00 -6.90100431e-01
4.08129901e-01 1.45129636e-01 -2.23523557e-01 1.54444113e-01
-4.03824197e-05 1.86391458e-01 -4.52696353e-01 -1.18459456e-01
2.88697511e-01 -6.91294074e-01 -6.83460310e-02 1.34759471e-01
-9.20714974e-01 5.22097886e-01 -2.67108262e-01 7.83559740e-01
-5.15929222e-01 -9.91162360e-02 9.29771483e-01 8.04098368e-01
5.35337865e-01 3.54221225e-01 -4.28364038e-01 -6.04451478e-01
-1.11105239e+00 -2.67947074e-02 8.25128257e-01 3.77350301e-01
2.72703230e-01 7.80014932e-01 -1.88162535e-01 9.76266026e-01
1.47229269e-01 6.91933155e-01 5.71424365e-01 -5.39023221e-01
3.45608562e-01 -4.59369943e-02 -1.07388631e-01 -2.78522462e-01
-6.28215134e-01 -7.20885992e-01 -4.41942841e-01 7.09787190e-01
7.24054098e-01 -5.32738864e-01 -8.44896972e-01 1.40423596e+00
2.12378085e-01 1.54546961e-01 1.98644668e-01 6.12105608e-01
3.48171622e-01 7.08752275e-01 -2.94426568e-02 -8.52168500e-01
1.19410920e+00 -5.98994613e-01 -1.06524205e+00 -2.30147973e-01
3.26421320e-01 -1.40318620e+00 1.05430830e+00 8.33641112e-01
-1.14718175e+00 -7.26931632e-01 -1.22597206e+00 5.87597549e-01
-2.25385502e-01 -1.32176846e-01 -1.56548455e-01 1.45246816e+00
-9.13328409e-01 3.69368881e-01 -5.99106789e-01 -2.44000420e-01
-2.40746066e-01 4.03518260e-01 -5.79947159e-02 2.52121717e-01
-1.14926004e+00 8.93662453e-01 -4.64578904e-02 -1.70514852e-01
-1.04986143e+00 -5.23205519e-01 -5.65311551e-01 -6.13851622e-02
2.80618370e-01 -5.97161464e-02 1.37309217e+00 -8.03742647e-01
-1.87585425e+00 4.37839389e-01 -2.34930366e-01 -6.18145645e-01
9.00817215e-01 -2.51094460e-01 -9.63701844e-01 1.86613873e-01
-3.53480518e-01 -3.58814225e-02 1.34543574e+00 -1.17027175e+00
-3.92024249e-01 -3.39860320e-02 -2.95385927e-01 4.95786630e-02
-1.79371923e-01 1.86312750e-01 -1.21482342e-01 -6.99532330e-01
-2.37067282e-01 -8.46517146e-01 1.12395220e-01 -4.88180935e-01
-5.00154570e-02 -5.78654185e-02 1.10626543e+00 -7.21895397e-01
1.24423754e+00 -2.15904117e+00 -4.78776284e-02 2.29645014e-01
-3.49985152e-01 9.68688846e-01 -3.61141413e-02 4.96410698e-01
-3.98943983e-02 -3.52926105e-02 2.28934991e-03 -4.30677235e-01
-2.53092110e-01 9.84113943e-03 5.18885069e-02 4.90047902e-01
1.39508873e-01 2.02547193e-01 -5.94756663e-01 -2.63632685e-01
5.95368803e-01 4.98487443e-01 -9.53488126e-02 4.83841091e-01
2.09360998e-02 4.38990474e-01 1.09613508e-01 3.20994705e-01
5.20749688e-01 7.49225020e-01 -8.70109648e-02 -4.31002490e-02
-2.42583111e-01 3.25083762e-01 -1.36967576e+00 1.35767019e+00
-1.06120300e+00 9.78484988e-01 2.53859401e-01 -8.95647943e-01
1.13565040e+00 9.48784947e-01 3.81627858e-01 -9.89315093e-01
1.73116267e-01 4.48989034e-01 5.94327152e-01 -4.37257111e-01
1.73007593e-01 -1.03507815e-02 2.72166073e-01 4.82224673e-01
1.06698219e-02 -3.20608228e-01 -5.45600988e-02 -3.82804036e-01
1.03667724e+00 -3.22178334e-01 2.95690268e-01 -1.83483437e-01
8.86006236e-01 -4.95300025e-01 4.66951393e-02 4.33721364e-01
-4.12596375e-01 2.32417449e-01 5.29855371e-01 2.81806022e-01
-9.96727884e-01 -1.06423473e+00 -4.32859249e-02 9.40237045e-01
-3.68728071e-01 -7.94148631e-03 -9.39889252e-01 -2.40227655e-01
-3.04583579e-01 8.20327878e-01 -1.05864011e-01 -1.48589373e-01
-6.53103352e-01 -3.06897491e-01 9.44186151e-01 -3.13110165e-02
1.68259859e-01 -1.20858049e+00 -6.50243342e-01 5.33247709e-01
8.94973502e-02 -1.28979659e+00 -2.35796615e-01 5.55395663e-01
-7.87206650e-01 -5.77352822e-01 -7.24348903e-01 -5.54859340e-01
-6.35115057e-02 7.52964616e-02 7.46930540e-01 -2.82999009e-01
-1.52157322e-01 5.09319842e-01 -4.95300859e-01 -7.83101797e-01
-1.24327540e+00 8.72131530e-03 3.30231279e-01 7.90100098e-02
1.57405928e-01 -6.04167759e-01 -2.63744593e-01 4.86159652e-01
-7.85020590e-01 -7.15118170e-01 5.56129098e-01 6.34974539e-01
2.27698863e-01 1.75871417e-01 1.00781310e+00 -5.09412050e-01
8.43561411e-01 -2.08207890e-01 -4.77072626e-01 -1.64778367e-01
-6.24901175e-01 -7.62560731e-03 8.07728052e-01 -7.16072977e-01
-9.20117080e-01 -1.06279023e-01 -6.69078767e-01 -4.44395512e-01
-4.05432284e-01 1.60970911e-01 -4.17743266e-01 -1.18499272e-01
8.02017331e-01 1.91512741e-02 2.35594615e-01 -6.34496689e-01
-4.88662301e-03 9.07500684e-01 2.45308265e-01 -1.23495288e-01
7.77927220e-01 4.65377085e-02 1.07937358e-01 -1.59093761e+00
-1.86520275e-02 -7.17654586e-01 -4.54831451e-01 -4.75897759e-01
5.13509512e-01 -7.11447120e-01 -6.05530024e-01 7.46258795e-01
-1.23571157e+00 -4.80974108e-01 -1.51968345e-01 6.41233325e-01
-5.66526473e-01 3.28041226e-01 -2.63817370e-01 -1.30412424e+00
-4.59112823e-01 -1.29552448e+00 5.69928527e-01 -1.04200013e-01
-2.70718038e-01 -9.18649852e-01 -1.88670769e-01 2.34616339e-01
3.91451865e-01 7.68382475e-02 9.15112257e-01 -9.51109767e-01
-2.32189056e-02 -1.80439934e-01 4.14116830e-01 7.87010491e-01
2.03546792e-01 -4.83991392e-02 -1.53718245e+00 -4.05800074e-01
2.41211504e-01 8.99684653e-02 3.39127243e-01 6.19961083e-01
5.93483210e-01 1.42027408e-01 1.81635782e-01 -2.29923390e-02
1.18223202e+00 7.09488988e-01 6.98060215e-01 2.25356087e-01
1.87657520e-01 6.15690410e-01 9.34345603e-01 3.48600447e-01
-6.27778232e-01 9.61280465e-01 4.41029608e-01 -1.49782062e-01
-5.84243238e-01 2.60270804e-01 6.87412202e-01 8.17533374e-01
1.62042245e-01 -5.84252775e-01 -8.14467430e-01 6.09156668e-01
-1.16189277e+00 -8.70571494e-01 -3.15216482e-01 2.62962198e+00
5.97822070e-01 7.39098370e-01 4.54202712e-01 1.14404309e+00
7.44529366e-01 3.40337336e-01 -2.73311645e-01 -1.23364651e+00
8.99581332e-03 7.47396231e-01 4.25008655e-01 7.71617055e-01
-8.43307137e-01 5.96541941e-01 5.72691631e+00 9.73542452e-01
-1.54003429e+00 4.87582356e-01 2.32681688e-02 -1.72570676e-01
1.62551761e-01 -5.47073603e-01 -4.31728750e-01 4.03118789e-01
1.55610728e+00 -1.74873427e-01 2.81588584e-01 6.28749013e-01
8.13476264e-01 -3.69032741e-01 -7.92998374e-01 9.96229053e-01
6.95494050e-03 -5.90519190e-01 -4.18515116e-01 9.11052525e-02
2.73289591e-01 -2.12162197e-01 5.76258898e-02 8.32590386e-02
-5.95791936e-01 -8.66943479e-01 7.44820774e-01 1.70405328e-01
9.44699228e-01 -9.63429570e-01 9.02212501e-01 3.77736509e-01
-1.19394875e+00 -4.22202125e-02 1.02118343e-01 7.64185339e-02
3.59333992e-01 6.48540974e-01 -1.38541579e+00 3.00768435e-01
2.90259928e-01 -3.12007546e-01 -2.30263561e-01 1.26905978e+00
-1.50167808e-01 1.05115199e+00 -2.63079315e-01 -2.49304950e-01
1.90171584e-01 2.25515798e-01 1.02055144e+00 1.54829109e+00
3.66752625e-01 -5.69684923e-01 -2.01732352e-01 2.73100529e-02
2.33266965e-01 3.39826018e-01 -6.12252235e-01 1.76747635e-01
5.67324042e-01 9.05901611e-01 -4.59712863e-01 1.16295610e-02
-2.27095723e-01 1.08225393e+00 -5.19243717e-01 2.09433675e-01
-7.09353685e-01 -6.47081852e-01 8.37268651e-01 2.49989718e-01
1.13871284e-01 -5.27761638e-01 -1.01331748e-01 -5.49733676e-02
-5.92097901e-02 -9.13172662e-01 -1.14477791e-01 -3.80493194e-01
-5.45198619e-01 5.89958429e-01 -9.15585607e-02 -1.45507908e+00
-5.74039102e-01 -5.99126399e-01 -7.62205899e-01 1.16316855e+00
-1.33018827e+00 -6.75917625e-01 -7.03847483e-02 5.24065912e-01
1.07663751e+00 -4.11497205e-01 8.44136953e-01 6.26572192e-01
-3.40509057e-01 8.62098932e-01 1.72386065e-01 -2.45769888e-01
9.32309091e-01 -1.30499840e+00 3.99321109e-01 1.09532034e+00
1.82954282e-01 1.21562272e-01 1.24891794e+00 -3.56518865e-01
-1.15375805e+00 -9.60561633e-01 6.91122591e-01 9.51861292e-02
4.69986916e-01 -4.65580791e-01 -9.80217218e-01 -2.22351491e-01
3.67753893e-01 -2.65541822e-01 6.77658319e-01 -1.63093999e-01
-1.20077327e-01 -3.18231493e-01 -1.18205082e+00 3.22744071e-01
4.60818470e-01 -7.97475159e-01 -5.20454824e-01 2.16754675e-01
7.56235242e-01 -2.38617569e-01 -8.27939630e-01 -4.49474156e-02
5.06473005e-01 -9.10866499e-01 8.22164416e-01 -1.14845939e-01
-3.16792697e-01 -2.97023892e-01 9.02454630e-02 -1.61935413e+00
2.23360583e-01 -9.78174567e-01 -2.25313932e-01 1.40733027e+00
7.76862860e-01 -6.76060975e-01 4.30515945e-01 -3.29633296e-01
-2.41360754e-01 -2.75485188e-01 -1.40243840e+00 -1.25632608e+00
-1.73411444e-01 -9.01723206e-01 1.48457244e-01 3.12271595e-01
-2.93728054e-01 2.17343077e-01 -4.71131742e-01 4.38423902e-01
3.64013106e-01 -7.73721099e-01 6.16940677e-01 -1.04735434e+00
-4.26931262e-01 -2.54641920e-01 -4.47500646e-01 -4.75423187e-01
-1.87782392e-01 -3.44108999e-01 1.08122885e-01 -1.11572909e+00
-8.00459981e-01 -2.71545529e-01 -3.91946286e-01 -8.54779184e-02
5.55029437e-02 -9.21105891e-02 2.19182640e-01 -1.17007524e-01
3.67686711e-02 1.19689330e-01 9.23095763e-01 -8.94458145e-02
-6.53840363e-01 7.56929994e-01 2.35557497e-01 4.61937010e-01
8.86901319e-01 -5.66523492e-01 -6.25110090e-01 -3.00339796e-02
-3.53340715e-01 2.11004063e-01 4.38627377e-02 -1.43944454e+00
2.65340865e-01 3.46005917e-01 -8.41775723e-03 -4.99993920e-01
6.61911011e-01 -1.03444719e+00 8.80178362e-02 8.03003728e-01
4.43363264e-02 7.12805018e-02 7.59336710e-01 3.90270025e-01
-2.11553201e-01 -3.82204473e-01 1.03469718e+00 3.56873602e-01
-5.47553241e-01 -2.12523490e-01 -8.66715312e-01 -9.93120223e-02
1.05222166e+00 -3.38497490e-01 -2.06030700e-02 -5.21629393e-01
-7.68662870e-01 -3.60729605e-01 -3.93386893e-02 3.93570811e-01
3.61822188e-01 -7.51120865e-01 -5.95560551e-01 2.65821129e-01
-9.82449874e-02 -5.28986037e-01 4.21297997e-01 8.32406580e-01
-5.05234301e-01 3.89305800e-01 -1.71292081e-01 -6.59096539e-01
-1.92726648e+00 4.60344076e-01 4.60451424e-01 -8.05760399e-02
-1.11683393e-02 4.13396299e-01 -7.01159775e-01 2.55048007e-01
5.57556629e-01 -2.34840214e-01 -3.60243976e-01 4.15916711e-01
4.14960682e-01 7.29913294e-01 7.56297946e-01 -5.44188917e-01
-2.99160480e-01 4.54719275e-01 2.02521652e-01 -6.13938510e-01
9.36282218e-01 -3.29885297e-02 3.48536462e-01 7.37645566e-01
1.18130636e+00 6.49908006e-01 -8.78547728e-01 -2.16863193e-02
1.79854825e-01 -2.37084150e-01 2.97403306e-01 -8.55719745e-01
-5.85605383e-01 1.27544034e+00 1.32938659e+00 4.81658101e-01
1.30965602e+00 -3.76733959e-01 6.20592058e-01 5.59296608e-02
2.94573724e-01 -1.16705191e+00 2.63320327e-01 5.32614529e-01
8.44104946e-01 -8.57597053e-01 -1.80322215e-01 -2.80929714e-01
-6.01200819e-01 1.15285242e+00 2.09641725e-01 3.16444755e-01
6.62736535e-01 5.74905753e-01 3.44849467e-01 3.17571163e-01
-3.31866771e-01 -4.78172094e-01 1.11461453e-01 8.97043765e-01
3.69828492e-01 8.54836404e-02 -5.59666097e-01 2.02820286e-01
3.35646830e-02 -4.20263231e-01 5.26592314e-01 6.90105796e-01
-5.44095278e-01 -1.42237806e+00 -6.33750021e-01 2.56514490e-01
-7.60340154e-01 -1.09010890e-01 -1.73581377e-01 7.32104659e-01
1.35155261e-01 1.47338450e+00 -1.36964312e-02 -4.85552996e-01
8.14244688e-01 4.27644491e-01 2.25019023e-01 -4.44820553e-01
-1.00982130e+00 4.77675200e-01 3.95219833e-01 -2.00639457e-01
-2.30745688e-01 -8.52197766e-01 -1.08507609e+00 -1.28691658e-01
-3.76178622e-01 1.84651062e-01 1.18415785e+00 9.62433279e-01
9.49866027e-02 1.01964843e+00 1.00505114e+00 -5.50590456e-01
-3.78531635e-01 -1.18297553e+00 -5.66558361e-01 -1.42259290e-02
6.52521253e-01 -5.63832760e-01 -6.67250633e-01 -1.84827894e-01] | [14.933513641357422, 5.792105674743652] |
315eae2b-27e5-49e3-84da-c49088274ac4 | applications-of-deep-learning-in-stock-market | 2003.01859 | null | https://arxiv.org/abs/2003.01859v1 | https://arxiv.org/pdf/2003.01859v1.pdf | Applications of deep learning in stock market prediction: recent progress | Stock market prediction has been a classical yet challenging problem, with the attention from both economists and computer scientists. With the purpose of building an effective prediction model, both linear and machine learning tools have been explored for the past couple of decades. Lately, deep learning models have been introduced as new frontiers for this topic and the rapid development is too fast to catch up. Hence, our motivation for this survey is to give a latest review of recent works on deep learning models for stock market prediction. We not only category the different data sources, various neural network structures, and common used evaluation metrics, but also the implementation and reproducibility. Our goal is to help the interested researchers to synchronize with the latest progress and also help them to easily reproduce the previous studies as baselines. Base on the summary, we also highlight some future research directions in this topic. | ['Weiwei Jiang'] | 2020-02-29 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-7.70798802e-01 -4.87860709e-01 -5.55173814e-01 -2.48876661e-01
-2.89703608e-01 -4.80267465e-01 6.11852765e-01 7.40473345e-02
-4.99570012e-01 7.11588323e-01 -9.60936397e-03 -2.55431861e-01
1.71879074e-03 -9.64931726e-01 -3.59182239e-01 -5.27376652e-01
-2.66219527e-01 2.42860883e-01 1.86597720e-01 -5.22507846e-01
6.87263429e-01 3.92901838e-01 -1.46768224e+00 -3.14078294e-02
5.50173581e-01 1.49778497e+00 -1.55473039e-01 -1.50104510e-02
-4.81161565e-01 1.31618536e+00 -5.59573531e-01 -9.88692164e-01
6.09426618e-01 -3.52699250e-01 -3.99523675e-01 -2.82427281e-01
3.07995193e-02 -4.65443105e-01 -4.27809924e-01 1.00348854e+00
5.15682220e-01 -1.23007409e-01 4.37968940e-01 -1.37352002e+00
-8.54237378e-01 8.75682592e-01 -5.90554893e-01 6.77877367e-01
-4.47007686e-01 2.76473016e-02 1.41869760e+00 -9.73788261e-01
3.05885315e-01 6.33117497e-01 6.17083728e-01 1.81875229e-01
-6.64868534e-01 -1.10070693e+00 4.75542545e-01 7.42431104e-01
-9.69542563e-01 -1.77390605e-01 1.04616547e+00 -5.10168135e-01
8.86900783e-01 3.91133539e-02 9.97054696e-01 8.81914556e-01
4.21877027e-01 1.22070312e+00 1.16764867e+00 -1.66341528e-01
2.86842436e-01 1.65526092e-01 3.79429847e-01 2.12616503e-01
4.75894481e-01 4.22599792e-01 -5.43130100e-01 1.33564442e-01
8.29932451e-01 1.35203734e-01 7.02804998e-02 -2.01229990e-01
-1.05368674e+00 1.44530845e+00 3.98627251e-01 8.54230285e-01
-4.51718092e-01 -1.58845894e-02 4.30881500e-01 7.19760597e-01
8.01385939e-01 5.41301906e-01 -7.64779270e-01 -3.72959226e-01
-1.19311714e+00 7.26450861e-01 9.95909214e-01 5.60759604e-01
2.85673767e-01 5.79621792e-01 1.50067016e-01 5.27739823e-01
1.77850693e-01 2.73224801e-01 8.85786176e-01 -5.27195632e-01
2.35112727e-01 6.15930438e-01 1.50015965e-01 -1.09874296e+00
-5.81990361e-01 -1.00012195e+00 -1.08138394e+00 4.28665608e-01
5.31345546e-01 -4.62584883e-01 -3.43274325e-01 1.14749348e+00
-2.07190245e-01 2.18221486e-01 -1.46353215e-01 6.71520829e-01
6.44689322e-01 6.75890863e-01 -2.20033407e-01 -2.95814723e-01
1.17034483e+00 -1.12507021e+00 -6.61746740e-01 -1.41914725e-01
4.23086554e-01 -7.36546755e-01 6.46792412e-01 2.40212545e-01
-1.13346541e+00 -4.30795699e-01 -1.01974022e+00 2.59448439e-01
-6.82719469e-01 -1.46953389e-01 7.25460291e-01 5.19813716e-01
-9.01632249e-01 7.44594753e-01 -8.39294672e-01 9.58192497e-02
2.68753678e-01 1.36634022e-01 2.50319779e-01 5.98104060e-01
-1.61087286e+00 1.24666524e+00 3.11059028e-01 1.53074861e-01
-2.03045860e-01 -7.16365397e-01 -3.12994301e-01 5.97077198e-02
1.88208029e-01 -3.71999860e-01 1.42934060e+00 -8.89295518e-01
-1.70079994e+00 7.05731571e-01 3.56041521e-01 -9.25825238e-01
6.03558898e-01 -1.77577704e-01 -4.48521048e-01 -5.23282230e-01
-1.76519156e-01 2.62217581e-01 5.89980006e-01 -4.93861139e-01
-9.18414116e-01 -1.55404583e-01 -9.03231353e-02 5.48695214e-03
-3.89925033e-01 4.73590821e-01 -1.97510198e-02 -1.13967848e+00
-9.46972668e-02 -6.76842868e-01 -2.16712803e-01 -2.33395860e-01
3.25474627e-02 -5.61168849e-01 3.13875318e-01 -5.46817422e-01
1.43770504e+00 -1.84471977e+00 -7.43277147e-02 1.11639109e-02
1.18976302e-01 2.76089162e-01 1.07888453e-01 6.15095079e-01
-2.57209033e-01 3.71309146e-02 -1.50923222e-01 -3.36856574e-01
3.28990042e-01 -2.05838978e-01 -8.67604971e-01 3.89620364e-01
1.49847135e-01 1.23297346e+00 -5.60226262e-01 -7.59568252e-03
1.93034038e-01 1.71163410e-01 -1.21642202e-01 3.49622890e-02
-1.38959989e-01 6.57571033e-02 -3.68013412e-01 5.93787789e-01
6.19886398e-01 -3.76659036e-01 -1.96174562e-01 3.66513520e-01
-5.92585444e-01 5.44645250e-01 -1.28641462e+00 8.50761235e-01
6.40083030e-02 7.65890896e-01 -2.79817224e-01 -1.38014352e+00
1.14005291e+00 3.31518292e-01 7.60894358e-01 -9.31046247e-01
2.69712299e-01 6.95424974e-01 2.07430303e-01 5.93094043e-02
4.92255002e-01 -4.84427422e-01 7.06966668e-02 6.49039984e-01
-3.17537218e-01 2.44465008e-01 3.73885214e-01 -3.95467550e-01
4.48881269e-01 -1.02498874e-01 4.99898612e-01 -1.83648825e-01
4.11752820e-01 3.54853384e-02 5.15568852e-01 3.00884455e-01
-3.55169505e-01 1.91393703e-01 6.28388584e-01 -1.00583327e+00
-8.26894343e-01 -7.23500550e-01 -2.94297278e-01 1.12011242e+00
-2.04761505e-01 -6.12458810e-02 -3.53527457e-01 -3.84835005e-01
2.95956731e-01 5.15129030e-01 -6.89160466e-01 2.19875336e-01
-6.17775023e-01 -1.24341249e+00 3.08252126e-01 8.23012233e-01
7.63388872e-01 -1.40179324e+00 -5.23417652e-01 3.30816746e-01
1.85047373e-01 -7.09598780e-01 -1.36399269e-01 1.80221274e-01
-1.11601293e+00 -8.99362326e-01 -1.18650460e+00 -7.52860308e-01
-1.80214927e-01 1.41839206e-01 1.50199747e+00 1.74966723e-01
2.66597658e-01 -1.55759960e-01 -4.77762133e-01 -1.02825117e+00
-1.40723735e-01 4.92998093e-01 1.24154678e-02 -1.01240166e-01
7.25821078e-01 -4.31957960e-01 -6.16952538e-01 2.60836989e-01
-8.68598044e-01 -4.15607318e-02 6.65294051e-01 5.56119740e-01
1.42001614e-01 -4.57792729e-02 1.02653193e+00 -5.16395986e-01
8.54942143e-01 -5.46067417e-01 -1.21920562e+00 1.13049215e-02
-1.15087771e+00 3.02099567e-02 3.61450642e-01 -1.33701175e-01
-5.91846108e-01 -5.49824536e-01 -3.52864474e-01 6.16821982e-02
2.36772180e-01 1.11200523e+00 5.59977174e-01 3.01073007e-02
2.07558736e-01 4.22141582e-01 -7.47124404e-02 -5.93806088e-01
8.77239630e-02 4.16861236e-01 -1.39935622e-02 -5.17717786e-02
8.76688421e-01 1.54839709e-01 -1.60654083e-01 -6.62550211e-01
-6.92090452e-01 -8.56885165e-02 -6.50440812e-01 -1.29196346e-01
5.21754503e-01 -8.46035779e-01 -5.85970283e-01 1.06407452e+00
-8.48991394e-01 -3.63804281e-01 -1.19025521e-01 4.87632096e-01
-2.38100633e-01 1.40545890e-01 -1.04975080e+00 -9.85576868e-01
-4.62614506e-01 -9.88622963e-01 3.01425189e-01 3.16322625e-01
2.44789827e-03 -1.32295120e+00 2.72939771e-01 -6.19462021e-02
8.82318914e-01 8.54017586e-02 6.59833968e-01 -9.65193450e-01
-6.87164426e-01 -4.35488850e-01 -2.42981717e-01 4.67316657e-01
-6.74349070e-02 1.82662476e-02 -8.09601367e-01 -2.72219688e-01
1.46227986e-01 -2.39854619e-01 1.01682174e+00 7.16285110e-01
6.38306916e-01 1.39710517e-03 2.04959586e-02 2.98200339e-01
1.22272992e+00 4.38868135e-01 4.54154968e-01 9.71409380e-01
8.30989331e-02 6.51956379e-01 3.66179228e-01 6.10632181e-01
6.12560987e-01 4.71419036e-01 2.85843462e-01 -1.37515604e-01
5.19397795e-01 -9.29726586e-02 3.97375226e-01 1.08855939e+00
-4.17732537e-01 -2.71755960e-02 -7.41602182e-01 5.74593432e-02
-1.90946746e+00 -1.12675619e+00 1.98976994e-02 1.93862236e+00
6.83499455e-01 3.43152434e-01 6.43626988e-01 1.32819325e-01
4.00848687e-01 5.03026068e-01 -8.16388130e-01 8.87512416e-02
-3.02631855e-01 1.46635041e-01 6.25097930e-01 1.20337598e-01
-1.28345776e+00 8.71871889e-01 6.98058367e+00 4.49992210e-01
-1.48366773e+00 -3.40307742e-01 9.09881413e-01 -1.14663742e-01
-7.05757663e-02 -2.25916490e-01 -9.87388253e-01 6.64170980e-01
9.42210019e-01 -6.80639327e-01 1.93140075e-01 9.79353189e-01
2.37742707e-01 2.75219977e-01 -8.51464391e-01 9.94356036e-01
-1.18637219e-01 -1.70617723e+00 -2.09654346e-01 2.44294882e-01
7.02313006e-01 4.52066243e-01 4.63397622e-01 7.30742157e-01
1.72379866e-01 -8.02095652e-01 7.82657027e-01 6.01627588e-01
-6.85335044e-03 -7.86047995e-01 1.18989110e+00 5.09584010e-01
-1.25709724e+00 -3.23561043e-01 -6.25597656e-01 -8.36732924e-01
4.28869687e-02 4.74839598e-01 -9.96196270e-03 4.58905727e-01
7.00882137e-01 1.02485251e+00 -3.60553920e-01 1.25005984e+00
-9.02068019e-02 7.10106313e-01 -1.69071957e-01 -4.84575570e-01
4.86478925e-01 -4.52820122e-01 3.03877182e-02 8.46676767e-01
3.90383750e-01 -2.00366080e-01 3.95346992e-02 8.99423480e-01
-9.13860500e-02 5.42093873e-01 -3.08704883e-01 -3.57511908e-01
1.23108529e-01 8.73372734e-01 -7.63155937e-01 -3.17248225e-01
-1.08328903e+00 3.83715957e-01 2.82435179e-01 1.73390791e-01
-7.90138483e-01 -2.66043305e-01 9.95624125e-01 7.93110728e-02
1.64325550e-01 -3.23589981e-01 -7.37734258e-01 -1.40763319e+00
3.96539904e-02 -7.75041521e-01 4.56610382e-01 -3.32494646e-01
-1.51786089e+00 5.41045785e-01 4.45877053e-02 -1.21329677e+00
-3.80969733e-01 -9.58707750e-01 -6.99132681e-01 8.35788369e-01
-2.04827595e+00 -4.66674179e-01 2.18760118e-01 -5.85637195e-03
6.97276652e-01 -7.41629362e-01 5.97390354e-01 3.30065429e-01
-8.17201495e-01 2.85873115e-01 5.26471972e-01 6.78586841e-01
4.41922128e-01 -1.16174901e+00 1.01950908e+00 7.66918242e-01
4.56160516e-01 3.60443681e-01 5.33885062e-01 -4.69460815e-01
-7.96927273e-01 -7.39682794e-01 1.08190775e+00 -3.10373873e-01
1.26608169e+00 -1.08436309e-01 -9.41671968e-01 7.06176579e-01
5.91135204e-01 -4.52335894e-01 6.27951801e-01 5.08067124e-02
-1.69997156e-01 -2.95287192e-01 -5.33312857e-01 5.36591649e-01
2.87692845e-01 -2.94675957e-02 -5.99047899e-01 -6.32269904e-02
2.85885870e-01 -2.24222228e-01 -7.19685614e-01 8.05181265e-02
6.61815405e-01 -1.46625555e+00 6.99130774e-01 -5.00097334e-01
2.53544807e-01 1.60950705e-01 2.75512815e-01 -1.24024189e+00
-5.36687732e-01 -4.54779506e-01 -8.16304684e-02 8.59125614e-01
5.31926274e-01 -1.01049399e+00 1.08155262e+00 7.51725674e-01
1.87379513e-02 -1.04078197e+00 -7.67383337e-01 -8.26887786e-01
8.62983882e-01 -5.59501112e-01 7.50935674e-01 8.87096465e-01
2.34403014e-01 2.85414457e-01 -5.29632628e-01 -3.98819983e-01
4.04128999e-01 6.22832179e-01 5.26264429e-01 -1.72910857e+00
2.92631355e-03 -1.27444386e+00 -2.67607987e-01 -1.20559847e+00
1.13570847e-01 -7.81496286e-01 -6.52956188e-01 -1.37327504e+00
1.19687334e-01 -1.31950945e-01 -7.88374543e-01 2.17159137e-01
-3.32878008e-02 3.01877886e-01 3.86504918e-01 4.95463163e-01
-1.31499201e-01 5.46564221e-01 1.14937496e+00 -9.26538035e-02
-1.95590571e-01 5.77284873e-01 -8.86914611e-01 6.97638631e-01
1.20911682e+00 -2.94564098e-01 -1.22868516e-01 -1.72962204e-01
7.02667654e-01 -2.07698956e-01 7.24495426e-02 -7.39045620e-01
1.76080987e-01 -2.30837137e-01 4.27210510e-01 -8.16410422e-01
2.78102374e-03 -6.07810915e-01 -1.52076110e-01 7.07595289e-01
-3.26613814e-01 6.65225565e-01 1.79643437e-01 2.41431415e-01
-5.81504166e-01 -3.09544533e-01 8.09321046e-01 -2.31440082e-01
-8.68488312e-01 5.70760250e-01 -4.03416306e-01 4.64293920e-02
1.14753234e+00 -2.69223377e-03 -5.55987023e-02 -7.70893276e-01
-1.90718919e-01 3.29458594e-01 5.45552261e-02 7.17020094e-01
4.10973310e-01 -1.29583752e+00 -1.01452291e+00 1.31838039e-01
-1.70756340e-01 -5.60555279e-01 -6.22262284e-02 7.20446885e-01
-6.32595778e-01 9.47485209e-01 -3.79477382e-01 1.40890598e-01
-6.96197093e-01 7.38763094e-01 4.51372623e-01 -4.82974172e-01
-5.56344867e-01 6.78233683e-01 -3.17258164e-02 1.45858899e-01
4.95254427e-01 -9.02972758e-01 -5.83554447e-01 4.29217577e-01
7.62037396e-01 6.66711152e-01 5.97747043e-02 -4.42275733e-01
-2.13134870e-01 5.32148302e-01 -2.78628208e-02 1.01092905e-01
1.82956624e+00 1.45755097e-01 2.02181134e-02 9.29284930e-01
9.09299612e-01 -2.72320628e-01 -1.03627658e+00 -2.84897774e-01
5.68919003e-01 -7.41214156e-02 1.25106871e-01 -5.04817605e-01
-1.51525986e+00 8.46562862e-01 5.47605276e-01 8.50296378e-01
8.64954591e-01 -2.03484431e-01 9.56348360e-01 3.96711916e-01
3.36535633e-01 -1.22637177e+00 3.39576812e-03 7.57338762e-01
9.23478425e-01 -1.47938263e+00 2.31324732e-02 1.37422055e-01
-6.15073681e-01 1.36115670e+00 1.78874731e-01 -4.49257642e-01
1.10591900e+00 3.81388456e-01 4.52173263e-01 -5.18674180e-02
-7.84806907e-01 -2.53827512e-01 1.96953490e-01 2.14859873e-01
1.01595163e+00 -1.30812049e-01 -4.37733084e-01 7.44975805e-01
-4.59397107e-01 1.75761998e-01 3.14371854e-01 5.81782401e-01
-5.92341959e-01 -1.38970685e+00 -1.63009867e-01 7.03110337e-01
-6.28277361e-01 -1.83777899e-01 -3.91146898e-01 1.15788305e+00
-3.59938949e-01 7.64007747e-01 3.50225687e-01 -2.93619901e-01
4.71694022e-01 1.54564843e-01 -3.73406075e-02 -2.30924651e-01
-6.50990129e-01 -2.32602917e-02 -3.92402649e-01 -7.33681098e-02
-4.01540726e-01 -7.94352889e-01 -7.91580200e-01 -7.20114768e-01
-1.92573056e-01 2.25975350e-01 4.56401795e-01 9.42923963e-01
9.36343521e-02 2.33724192e-01 6.94663286e-01 -7.09844351e-01
-1.13963926e+00 -1.16788912e+00 -1.10198796e+00 -1.68800890e-01
2.38251090e-01 -8.09216678e-01 -3.43135625e-01 -2.36115962e-01] | [4.414448261260986, 4.265755653381348] |
a6d9ae3b-da5f-408e-b76f-000e0a5776b7 | re-think-and-re-design-graph-neural-networks | 2307.00222 | null | https://arxiv.org/abs/2307.00222v1 | https://arxiv.org/pdf/2307.00222v1.pdf | Re-Think and Re-Design Graph Neural Networks in Spaces of Continuous Graph Diffusion Functionals | Graph neural networks (GNNs) are widely used in domains like social networks and biological systems. However, the locality assumption of GNNs, which limits information exchange to neighboring nodes, hampers their ability to capture long-range dependencies and global patterns in graphs. To address this, we propose a new inductive bias based on variational analysis, drawing inspiration from the Brachistochrone problem. Our framework establishes a mapping between discrete GNN models and continuous diffusion functionals. This enables the design of application-specific objective functions in the continuous domain and the construction of discrete deep models with mathematical guarantees. To tackle over-smoothing in GNNs, we analyze the existing layer-by-layer graph embedding models and identify that they are equivalent to l2-norm integral functionals of graph gradients, which cause over-smoothing. Similar to edge-preserving filters in image denoising, we introduce total variation (TV) to align the graph diffusion pattern with global community topologies. Additionally, we devise a selective mechanism to address the trade-off between model depth and over-smoothing, which can be easily integrated into existing GNNs. Furthermore, we propose a novel generative adversarial network (GAN) that predicts spreading flows in graphs through a neural transport equation. To mitigate vanishing flows, we customize the objective function to minimize transportation within each community while maximizing inter-community flows. Our GNN models achieve state-of-the-art (SOTA) performance on popular graph learning benchmarks such as Cora, Citeseer, and Pubmed. | ['Guorong Wu', 'Won Hwa Kim', 'Minjeong Kim', 'Shahar Z Kovalsky', 'Ziquan Wei', 'Jiaqi Ding', 'Tingting Dan'] | 2023-07-01 | null | null | null | null | ['image-denoising', 'graph-embedding', 'graph-learning'] | ['computer-vision', 'graphs', 'graphs'] | [ 4.30877618e-02 3.01520944e-01 1.54362721e-02 -6.24561422e-02
-6.51415437e-02 -6.46243393e-01 6.83832526e-01 3.73162851e-02
-2.97806412e-01 6.14314198e-01 1.67713240e-01 -3.49595904e-01
-2.02125460e-01 -1.31142724e+00 -9.10620868e-01 -7.77762234e-01
-2.52889633e-01 1.50030002e-01 1.34857133e-01 -3.36731583e-01
-1.01806492e-01 6.20330453e-01 -5.82587361e-01 -4.22511343e-03
1.14128828e+00 5.62340081e-01 -9.65298340e-02 6.22571528e-01
-2.40204751e-01 8.51109624e-01 -2.60969818e-01 -6.20470285e-01
3.61868441e-01 -7.26231098e-01 -6.99139237e-01 -3.33768427e-02
4.54184115e-01 6.42254599e-04 -9.60850000e-01 1.35917282e+00
4.48559254e-01 -1.20133394e-02 8.08258474e-01 -1.32504213e+00
-1.30573845e+00 6.29336596e-01 -6.18288934e-01 2.79103875e-01
-3.09523642e-01 3.57752025e-01 1.14211690e+00 -4.85644460e-01
8.56701970e-01 1.28597140e+00 1.04264462e+00 6.29012406e-01
-1.68584454e+00 -4.79409695e-01 3.35776120e-01 -3.24506372e-01
-1.28504753e+00 -5.39535992e-02 9.41062093e-01 -5.59151828e-01
5.62549591e-01 4.28542756e-02 8.19010377e-01 1.23744249e+00
4.45697695e-01 4.80847299e-01 7.81717122e-01 6.65146634e-02
1.83244824e-01 -3.37174058e-01 -2.41502166e-01 9.30096626e-01
3.53858113e-01 -1.84388250e-01 -2.95159310e-01 -1.18457519e-01
1.31371880e+00 2.70673305e-01 -3.78839940e-01 -5.59597552e-01
-9.36733603e-01 1.11396050e+00 1.07761014e+00 2.93202341e-01
-3.73812407e-01 4.81961727e-01 2.29300439e-01 3.83517385e-01
7.43573427e-01 3.16941351e-01 -5.78647256e-02 4.35074389e-01
-8.08015525e-01 1.34013936e-01 8.36357594e-01 6.44681633e-01
7.82284439e-01 4.13221791e-02 -1.81859568e-01 5.99310338e-01
4.39297765e-01 4.00576442e-01 -2.06313794e-03 -8.40322912e-01
1.54099822e-01 7.96384335e-01 -4.37378824e-01 -1.43435025e+00
-2.89916039e-01 -6.18719041e-01 -1.43114817e+00 -4.21194285e-02
4.56381828e-01 -4.39127833e-02 -8.88695240e-01 1.98877943e+00
2.37525150e-01 4.49238420e-01 -2.92516798e-01 8.76453102e-01
5.68692863e-01 4.03626174e-01 7.79309943e-02 3.71908955e-02
8.27866137e-01 -9.00480092e-01 -5.69609821e-01 -1.73909709e-01
5.23140848e-01 -1.00067072e-01 9.74436402e-01 -1.28210619e-01
-1.10599160e+00 -1.02025606e-01 -7.67385304e-01 -2.27478072e-01
-5.43175042e-01 -6.12083197e-01 7.06081510e-01 4.33619201e-01
-1.32480311e+00 8.95973027e-01 -1.14641261e+00 -2.65978724e-01
8.70405376e-01 3.03033590e-01 -1.62459075e-01 7.81996176e-03
-1.12870848e+00 5.22841454e-01 -2.73788840e-01 2.70947546e-01
-9.18110132e-01 -1.05957890e+00 -9.05105770e-01 9.68679488e-02
1.99983448e-01 -9.38269973e-01 2.94580132e-01 -9.57108796e-01
-1.34189594e+00 7.99910665e-01 1.77444547e-01 -5.97668946e-01
7.55311966e-01 2.82827973e-01 -1.75773278e-01 1.25934735e-01
-3.62068936e-02 6.07325852e-01 7.42312133e-01 -8.92374516e-01
1.78277358e-01 -2.68109888e-01 2.38130301e-01 -1.75662532e-01
-5.77704251e-01 -4.75673169e-01 -3.76487911e-01 -9.36599970e-01
-1.21020325e-01 -8.07209671e-01 -4.57972825e-01 4.57023531e-01
-4.69869316e-01 9.18800682e-02 6.86849058e-01 -6.91276848e-01
1.16360641e+00 -2.03391171e+00 7.07536995e-01 4.06716466e-01
8.88531089e-01 4.69794162e-02 -4.64786798e-01 3.28456163e-01
2.44438127e-01 4.97362614e-01 -6.89222515e-01 -3.49691480e-01
1.51727982e-02 3.57257992e-01 3.03896088e-02 6.87095046e-01
4.05652374e-01 1.39393234e+00 -1.02245891e+00 -3.24099243e-01
-9.32181925e-02 9.81656849e-01 -8.92558634e-01 -5.50282076e-02
-2.82427698e-01 4.45726335e-01 -5.45109153e-01 2.13156581e-01
7.53342152e-01 -7.35809147e-01 3.29284400e-01 -1.05285868e-01
1.87209472e-01 7.10423961e-02 -7.62425303e-01 1.76400089e+00
-4.13852811e-01 4.78617519e-01 4.93691206e-01 -1.35005903e+00
6.82389855e-01 -1.47381708e-01 5.74207425e-01 -4.12937164e-01
1.05321929e-01 -3.15773264e-02 9.22581274e-03 -4.44723628e-02
-4.42135856e-02 -5.88988811e-02 2.37740427e-01 2.56809682e-01
6.68737590e-02 1.79000814e-02 1.45611152e-01 5.74994147e-01
1.42642486e+00 -1.58006221e-01 -2.08123997e-01 -6.75626755e-01
3.07860583e-01 -3.61128181e-01 5.21072686e-01 7.48545170e-01
-1.33880422e-01 6.35231733e-01 9.31320667e-01 -4.16145355e-01
-1.00803304e+00 -1.28690195e+00 1.35570735e-01 8.26770723e-01
1.28599912e-01 -2.28806034e-01 -8.47980142e-01 -9.82978940e-01
2.22330540e-01 6.12982847e-02 -9.91389453e-01 -3.11361700e-01
-5.17794549e-01 -8.47492814e-01 6.34542286e-01 3.47140521e-01
4.35079724e-01 -8.04078758e-01 1.28337637e-01 3.28298151e-01
8.29813704e-02 -9.80271935e-01 -9.52109158e-01 -8.88404623e-02
-7.23264456e-01 -1.11020064e+00 -8.51169467e-01 -6.73275352e-01
9.38582897e-01 9.70940068e-02 1.25998974e+00 3.55452091e-01
-3.41545671e-01 4.26612675e-01 4.91837002e-02 -4.31880029e-03
-4.56850886e-01 3.71101409e-01 -2.00465366e-01 1.76720738e-01
4.21958901e-02 -1.00630569e+00 -8.87369871e-01 1.74779221e-01
-1.16630793e+00 -5.11538796e-02 3.99876237e-01 8.27023804e-01
5.78889370e-01 -6.68731332e-02 6.12246037e-01 -1.04185677e+00
8.95071208e-01 -6.70952737e-01 -7.14130759e-01 2.32551500e-01
-6.55637085e-01 2.17963055e-01 7.67527997e-01 -6.29431188e-01
-5.60874104e-01 -3.02236587e-01 -8.57619569e-04 -5.09729624e-01
4.41131592e-01 6.72482193e-01 -7.07860366e-02 -4.96583819e-01
5.09270906e-01 6.73389956e-02 4.61102724e-01 -2.98204809e-01
6.30305052e-01 4.91221100e-02 2.98670769e-01 -3.23834151e-01
8.46172631e-01 7.94707954e-01 2.93017924e-01 -9.11723852e-01
-7.41382122e-01 -2.00624336e-02 -4.81342435e-01 -8.71460736e-02
9.19540942e-01 -7.34040976e-01 -7.79160202e-01 5.29775918e-01
-1.07844627e+00 -7.57620096e-01 -4.13487375e-01 4.18620110e-02
-2.38509387e-01 4.19396520e-01 -1.03052354e+00 -3.77334982e-01
-4.07871723e-01 -9.75149035e-01 8.44256878e-01 1.34919122e-01
9.86482948e-02 -1.71752453e+00 1.42398477e-01 -5.78032769e-02
7.76605368e-01 5.97907901e-01 8.20350349e-01 -2.71775693e-01
-7.23928034e-01 1.87521473e-01 -4.63554353e-01 4.84633803e-01
1.14354208e-01 -7.91847035e-02 -4.50265884e-01 -4.53566909e-01
-2.85256207e-01 6.68716133e-02 1.27760422e+00 7.30245650e-01
1.12436450e+00 -5.93548417e-01 -4.06293213e-01 1.08451974e+00
1.36237001e+00 -5.07797718e-01 6.72929287e-01 -1.19500794e-01
1.29846787e+00 4.79684353e-01 -5.49559176e-01 1.09380884e-02
4.51621383e-01 2.10719496e-01 5.14727890e-01 -2.54473537e-01
-3.43705624e-01 -4.78884757e-01 2.89826423e-01 9.38883901e-01
-2.28045210e-02 -4.72445309e-01 -7.80950069e-01 6.07257962e-01
-1.86621308e+00 -7.48694777e-01 -2.05870211e-01 1.79431427e+00
7.62247205e-01 1.48461252e-01 1.62540495e-01 -4.92000431e-01
7.55537391e-01 5.21865606e-01 -7.73167431e-01 -7.73768779e-03
-3.77537072e-01 2.07454368e-01 7.90980339e-01 8.59533250e-01
-9.49805915e-01 9.27974224e-01 5.85482359e+00 6.81872368e-01
-1.30016649e+00 2.64503390e-01 7.45173454e-01 -4.59367670e-02
-8.02524209e-01 -1.14751838e-01 -3.43228757e-01 5.95896244e-01
7.16650724e-01 -1.71406806e-01 8.50228548e-01 3.72102290e-01
1.26883969e-01 7.01341748e-01 -8.42128992e-01 5.90963006e-01
-1.57139465e-01 -1.86026692e+00 1.61889911e-01 5.46386123e-01
9.96347249e-01 3.77978534e-01 6.40308484e-02 3.21614789e-03
6.58991992e-01 -1.23175681e+00 3.63989115e-01 5.87526739e-01
6.71248734e-01 -6.17479563e-01 3.31104785e-01 1.27932027e-01
-1.18311417e+00 2.15364158e-01 -4.76675719e-01 1.52752489e-01
1.92623436e-01 9.26638901e-01 -2.21976712e-01 5.11276722e-01
3.86214733e-01 1.10538352e+00 -3.81243318e-01 6.86545730e-01
-1.29156336e-01 6.56630814e-01 -4.17760193e-01 3.81624885e-02
3.14465284e-01 -6.69018567e-01 6.96396947e-01 1.16912746e+00
1.21493042e-01 -2.82611370e-01 1.20789766e-01 1.55123377e+00
-6.62209988e-01 -6.95705861e-02 -7.71964908e-01 -4.07708764e-01
2.13117152e-01 1.12991023e+00 -8.57673585e-01 1.03134155e-01
-5.01748979e-01 1.14075339e+00 6.61772609e-01 7.59110451e-01
-8.00556242e-01 -2.69699931e-01 7.63133883e-01 5.20072937e-01
3.34070981e-01 -3.44312131e-01 -1.59031257e-01 -1.32026601e+00
3.02693807e-02 -6.09753549e-01 2.28247404e-01 -2.11699396e-01
-1.79926825e+00 4.03746128e-01 -4.42151994e-01 -5.90132654e-01
4.54277217e-01 -5.97269237e-01 -8.75932634e-01 8.61388147e-01
-1.59368992e+00 -1.31615078e+00 -2.11502969e-01 5.25574028e-01
-6.25293106e-02 1.26934543e-01 3.36344481e-01 5.24404407e-01
-5.83528340e-01 4.85423326e-01 1.00721605e-01 3.99664313e-01
3.82103324e-01 -1.29631162e+00 8.15222383e-01 9.28209662e-01
7.62831122e-02 7.92732835e-01 4.17584985e-01 -8.26724112e-01
-1.48551261e+00 -1.47645259e+00 5.71790099e-01 -3.82686585e-01
1.09960365e+00 -6.69510722e-01 -1.26035905e+00 6.73187673e-01
-3.14750429e-03 6.49970293e-01 1.96549162e-01 -5.84683008e-02
-4.70855922e-01 2.86748055e-02 -1.08645141e+00 6.27129257e-01
1.49211657e+00 -6.93058074e-01 2.23766536e-01 4.17306125e-01
8.90016258e-01 -6.44561276e-02 -9.34819221e-01 2.06436709e-01
3.72217983e-01 -7.56104887e-01 1.14691842e+00 -8.09352100e-01
6.46426678e-01 -8.86804163e-02 2.39830792e-01 -1.57630205e+00
-4.65015918e-01 -9.98280466e-01 -2.01390564e-01 1.07807970e+00
4.14792895e-01 -8.93970132e-01 9.46105540e-01 3.91244859e-01
6.69292212e-02 -7.80103922e-01 -7.01335132e-01 -7.52890706e-01
5.84488988e-01 8.19926411e-02 4.89654452e-01 1.25798488e+00
-4.02165055e-01 2.13800684e-01 -1.81246310e-01 6.65770322e-02
8.99640918e-01 -3.70407432e-01 5.21049976e-01 -1.26671135e+00
-3.62856477e-01 -8.93030047e-01 -4.31301951e-01 -1.22856069e+00
4.15046155e-01 -1.23124588e+00 -3.18610072e-01 -1.62637579e+00
9.39986110e-02 -4.45620745e-01 -3.30129772e-01 2.73765564e-01
-1.39120102e-01 2.89834142e-01 3.60083804e-02 -1.59222614e-02
-3.79318774e-01 7.01084375e-01 1.60708559e+00 -4.18025613e-01
-2.16725633e-01 -3.39771092e-01 -7.64453292e-01 6.32859349e-01
6.66130662e-01 -5.64514160e-01 -4.39888537e-01 -6.05264664e-01
6.30173266e-01 -1.66299298e-01 7.61163294e-01 -5.61040998e-01
3.74797255e-01 -9.20926183e-02 3.15495655e-02 6.24227934e-02
-1.29553199e-01 -5.45619726e-01 1.06029823e-01 4.44008797e-01
-3.92829806e-01 -1.32410154e-01 -6.10483103e-02 9.40585136e-01
3.47828097e-03 3.09076577e-01 9.29310381e-01 -8.34798887e-02
-7.44411945e-02 8.21477056e-01 -1.46471098e-01 4.56731647e-01
6.45689905e-01 1.13405310e-01 -4.98669952e-01 -4.84026790e-01
-6.31009996e-01 3.21873307e-01 6.81304455e-01 2.04230264e-01
4.44865882e-01 -1.29644215e+00 -8.49182248e-01 1.29564717e-01
-2.53673285e-01 2.24860281e-01 3.55943859e-01 1.01848733e+00
-6.86698020e-01 -2.10033551e-01 -7.83936530e-02 -4.16642666e-01
-6.03342235e-01 4.12944943e-01 7.23681569e-01 -4.51983631e-01
-8.13588202e-01 9.89739954e-01 5.38242280e-01 -5.82206607e-01
3.06205228e-02 -2.95803398e-01 1.75458625e-01 -2.43849561e-01
1.66401848e-01 2.57062942e-01 -2.02364117e-01 -3.86255831e-01
-2.98049718e-01 3.14264208e-01 6.19849786e-02 1.83241099e-01
1.50172567e+00 -1.12782493e-01 -5.05637646e-01 1.28160715e-01
1.47151840e+00 1.20695211e-01 -1.62790537e+00 -1.74038678e-01
-2.81624466e-01 -6.91995546e-02 2.90140063e-01 -4.28876251e-01
-1.53631365e+00 9.10060048e-01 1.83678240e-01 5.95785797e-01
7.96013117e-01 5.68867549e-02 9.43808675e-01 -2.37203538e-02
-7.23029226e-02 -7.26611912e-01 1.40853882e-01 5.41097581e-01
7.45232224e-01 -1.03716707e+00 -7.78483301e-02 -4.11629081e-01
-1.67848170e-01 7.59516835e-01 3.43968183e-01 -5.84908128e-01
1.03847694e+00 4.26172018e-01 -3.02710563e-01 -5.33264101e-01
-4.43061143e-01 3.07256524e-02 4.23072636e-01 5.34340024e-01
3.45838308e-01 4.86555770e-02 -1.69001728e-01 3.53665739e-01
1.82591498e-01 -2.03412965e-01 4.47572887e-01 5.41756153e-01
5.31373732e-03 -9.35872138e-01 3.28043550e-01 5.23226440e-01
-5.37013233e-01 -2.91681647e-01 -6.29910171e-01 7.46512473e-01
-2.33958021e-01 4.95331407e-01 2.42670491e-01 -1.41133964e-01
7.01396987e-02 -2.95807898e-01 4.16186363e-01 -3.93894762e-01
-4.33961898e-01 -1.43033592e-02 -3.66815388e-01 -5.67209482e-01
-4.79895979e-01 -3.76381159e-01 -1.03169405e+00 -7.52483130e-01
-1.24992169e-01 3.59603576e-02 3.79311621e-01 6.84122503e-01
6.49630368e-01 8.48099828e-01 3.64099622e-01 -6.51337206e-01
-2.79657334e-01 -7.44498909e-01 -6.90666795e-01 6.67466283e-01
4.34544265e-01 -4.14114445e-01 -6.96218491e-01 -3.59219275e-02] | [6.928215503692627, 6.110955715179443] |
b99fcdea-50f3-4a90-aba2-571bf67121c5 | phishsim-aiding-phishing-website-detection | 2207.10801 | null | https://arxiv.org/abs/2207.10801v1 | https://arxiv.org/pdf/2207.10801v1.pdf | PhishSim: Aiding Phishing Website Detection with a Feature-Free Tool | In this paper, we propose a feature-free method for detecting phishing websites using the Normalized Compression Distance (NCD), a parameter-free similarity measure which computes the similarity of two websites by compressing them, thus eliminating the need to perform any feature extraction. It also removes any dependence on a specific set of website features. This method examines the HTML of webpages and computes their similarity with known phishing websites, in order to classify them. We use the Furthest Point First algorithm to perform phishing prototype extractions, in order to select instances that are representative of a cluster of phishing webpages. We also introduce the use of an incremental learning algorithm as a framework for continuous and adaptive detection without extracting new features when concept drift occurs. On a large dataset, our proposed method significantly outperforms previous methods in detecting phishing websites, with an AUC score of 98.68%, a high true positive rate (TPR) of around 90%, while maintaining a low false positive rate (FPR) of 0.58%. Our approach uses prototypes, eliminating the need to retain long term data in the future, and is feasible to deploy in real systems with a processing time of roughly 0.3 seconds. | ['Sanjay Jha', 'Alan Blair', 'Arindam Pal', 'Rizka Purwanto'] | 2022-07-13 | null | null | null | null | ['phishing-website-detection'] | ['adversarial'] | [ 7.09486976e-02 -3.92927825e-01 -1.51292786e-01 -1.32248282e-01
-6.49619520e-01 -9.65475321e-01 5.33778429e-01 6.15187705e-01
-3.08697760e-01 4.65912044e-01 -5.04867375e-01 -4.92786050e-01
-1.47359341e-01 -9.44028795e-01 -2.95361280e-01 -8.64711881e-01
-3.39062393e-01 3.17851841e-01 8.88805151e-01 1.64652959e-01
8.86974812e-01 2.86631733e-01 -1.73777974e+00 2.01478660e-01
1.14104056e+00 9.42763865e-01 2.31867552e-01 8.08790684e-01
-4.24033552e-02 2.65757203e-01 -7.92012572e-01 -5.96325874e-01
5.69921136e-01 -5.65559492e-02 -4.82539207e-01 4.31329198e-03
5.48378944e-01 -4.06973004e-01 -1.60226867e-01 1.21725357e+00
1.72662899e-01 -2.04929247e-01 6.52146876e-01 -1.20603704e+00
-3.97290289e-01 -9.40076262e-02 -7.31696725e-01 3.28489423e-01
3.88605982e-01 1.91957265e-01 1.05959821e+00 -5.59003472e-01
6.06110096e-01 6.86509669e-01 8.50553632e-01 -5.27354181e-02
-1.14487898e+00 -4.50385690e-01 -2.19935387e-01 5.60299516e-01
-1.52883852e+00 -7.61742517e-02 4.77915287e-01 -2.89742053e-01
7.64313400e-01 3.50239575e-01 6.01360321e-01 7.27201879e-01
2.73742735e-01 5.50068498e-01 1.09782445e+00 -5.10433197e-01
6.83478117e-01 5.07213354e-01 7.03010082e-01 3.84881914e-01
1.06065547e+00 -1.30300745e-01 4.82822955e-02 -7.39200950e-01
4.15839463e-01 3.30198407e-01 -3.46235223e-02 -4.18110013e-01
-8.65782976e-01 1.09300137e+00 6.26924708e-02 3.39340091e-01
-6.33345962e-01 -4.30368066e-01 6.86910272e-01 4.16521966e-01
1.63908631e-01 3.00687790e-01 -5.62899947e-01 -2.79837251e-02
-9.89342093e-01 1.19558372e-01 1.13702440e+00 6.86986089e-01
7.56619692e-01 -5.28536379e-01 2.98645407e-01 6.61409676e-01
1.55621931e-01 4.49034750e-01 6.17501020e-01 -6.49372697e-01
3.81259024e-01 7.60022640e-01 7.18544647e-02 -1.09802210e+00
-3.51680294e-02 -2.94444505e-02 -3.57370913e-01 9.68572795e-02
3.63858163e-01 1.96454272e-01 -6.65741920e-01 1.06156993e+00
3.54362011e-01 2.55270243e-01 -1.41320974e-01 3.61613035e-01
1.49288196e-02 6.13919973e-01 -6.89995335e-03 -3.89330775e-01
1.59084225e+00 -5.56855917e-01 -2.47581095e-01 6.05431981e-02
5.33804297e-01 -7.19951749e-01 1.15865707e+00 6.02257848e-01
-5.18180609e-01 -1.45863175e-01 -1.15160167e+00 6.95764959e-01
-6.68695569e-01 -3.67235541e-01 5.76366544e-01 1.20461726e+00
-9.77243066e-01 6.04237318e-01 -4.74037707e-01 -8.02663743e-01
2.41835654e-01 2.78578997e-01 -1.26852989e-01 -2.84612685e-01
-1.17462623e+00 7.92002797e-01 6.24645591e-01 -6.62898779e-01
-4.50747520e-01 -5.53098202e-01 -3.67306322e-01 3.31303418e-01
4.03403223e-01 -2.18859315e-01 7.62228489e-01 -7.11293519e-01
-9.90893960e-01 5.23609579e-01 -1.02049597e-02 -4.73746598e-01
2.84245074e-01 1.86838023e-02 -7.63681889e-01 4.35823143e-01
1.36321047e-02 -7.22111464e-02 7.51137137e-01 -1.16245759e+00
-8.60227287e-01 -5.20958066e-01 -1.34701177e-01 -2.01804578e-01
-7.66543865e-01 -1.36432365e-01 -3.30327868e-01 -2.85841316e-01
1.73851997e-01 -1.02608848e+00 -1.54833272e-01 -5.61273754e-01
-1.18904226e-01 -3.39668930e-01 9.44761097e-01 -1.02476239e+00
1.51630712e+00 -1.79076171e+00 -7.75965869e-01 7.87385464e-01
1.69839919e-01 8.64241481e-01 -5.83480746e-02 6.11923516e-01
4.52679358e-02 2.26888403e-01 -2.15788990e-01 4.99236763e-01
-2.93432415e-01 -2.07030982e-01 -9.72115397e-02 5.10470748e-01
1.47901289e-02 3.74129355e-01 -1.01129258e+00 -4.65435028e-01
2.33903587e-01 4.75436240e-01 -2.62835801e-01 9.27305147e-02
1.63550541e-01 -5.19617260e-01 -2.56378174e-01 7.72473454e-01
9.53722656e-01 -3.35240841e-01 5.07218242e-01 -4.67268601e-02
-1.37033597e-01 3.15940201e-01 -1.40208721e+00 8.85745049e-01
-3.07278812e-01 4.46364641e-01 -2.65021205e-01 -9.35528576e-01
1.20966613e+00 1.92403287e-01 4.85183299e-01 -6.62531972e-01
-4.89510261e-02 3.98822516e-01 -1.95878193e-01 -7.18245983e-01
2.27808148e-01 4.25740004e-01 1.32294059e-01 6.46569192e-01
-1.28079504e-01 6.13775253e-01 7.00924456e-01 4.36916918e-01
1.54511690e+00 -4.38491821e-01 7.57585108e-01 -2.97205061e-01
8.41255844e-01 7.87445717e-03 2.28868425e-01 6.48793340e-01
-4.03097808e-01 3.40478450e-01 5.01302958e-01 -4.48481798e-01
-1.29309881e+00 -1.12763333e+00 -1.99494094e-01 8.88852775e-01
1.76620096e-01 -4.39156741e-01 -7.97965884e-01 -1.06178033e+00
1.71677291e-01 8.14723670e-01 -4.74753737e-01 -1.00433901e-01
-4.97332394e-01 -1.00274062e+00 2.48832941e-01 5.26827201e-02
4.83446449e-01 -5.60437262e-01 -5.69200218e-01 2.26686150e-01
-2.21870244e-01 -5.51585078e-01 -2.48945713e-01 7.63351843e-02
-1.00288725e+00 -1.56396019e+00 -5.60742736e-01 -8.05014372e-01
6.38959765e-01 7.62619734e-01 8.13727260e-01 -2.00620573e-03
-5.11890411e-01 1.92557842e-01 -4.54157114e-01 -2.09494121e-03
-4.95798290e-01 7.11437166e-02 -1.39788806e-01 -1.67380318e-01
9.80806589e-01 -4.49593991e-01 -7.46076941e-01 4.36078280e-01
-8.37546587e-01 -8.59741449e-01 8.28659534e-01 7.19547689e-01
2.54921615e-01 5.17431378e-01 7.92385578e-01 -9.62084889e-01
7.72398710e-01 -8.80167246e-01 -7.42765665e-01 5.53820014e-01
-1.36684895e+00 -1.12802103e-01 6.67016506e-01 -5.75248361e-01
-9.94093537e-01 1.85435608e-01 2.13025421e-01 5.58046810e-02
-2.29099810e-01 3.01462114e-01 8.52570385e-02 -1.23814903e-02
8.62569988e-01 4.82502341e-01 1.48544401e-01 -5.40874541e-01
1.67312130e-01 1.04912889e+00 3.51542741e-01 6.71466440e-02
9.18262124e-01 3.01394939e-01 -1.30153239e-01 -7.87819445e-01
-2.34520391e-01 -1.29630613e+00 -8.86521220e-01 -3.72215845e-02
2.26799011e-01 -5.29062808e-01 -9.53944027e-01 2.84828961e-01
-8.94068420e-01 5.06039381e-01 2.54928648e-01 2.62316763e-01
-1.72223598e-01 1.12921655e+00 -6.06634021e-01 -8.72920513e-01
-6.63074493e-01 -4.75887895e-01 4.71375704e-01 1.50627106e-01
-2.28899658e-01 -9.74429190e-01 3.76784980e-01 2.80833483e-01
4.75004911e-01 1.47572801e-01 6.89776897e-01 -1.28711450e+00
-3.62226635e-01 -1.06038117e+00 -3.37966949e-01 3.91314894e-01
3.19975525e-01 7.93747231e-02 -7.41292596e-01 -6.07256532e-01
2.57673293e-01 1.50884509e-01 7.70605922e-01 1.23522818e-01
7.27521062e-01 -4.72098678e-01 -5.16564608e-01 1.31803125e-01
1.81513619e+00 6.41238332e-01 7.92883098e-01 8.32616091e-01
2.93542862e-01 6.42742276e-01 8.13710153e-01 5.25767326e-01
1.24399476e-01 3.39969248e-01 4.52409565e-01 2.66418308e-01
4.86349948e-02 -8.37705731e-02 4.62159693e-01 6.77153587e-01
1.39710099e-01 -1.70595527e-01 -9.21045482e-01 5.64339995e-01
-1.56516767e+00 -1.21607804e+00 -4.29797113e-01 2.87984443e+00
5.47356129e-01 1.87631935e-01 5.16536117e-01 4.41755474e-01
1.32902622e+00 -1.84168458e-01 -4.73380864e-01 -2.96214670e-01
3.14927012e-01 3.63813229e-02 8.62263620e-01 2.90604055e-01
-1.26196039e+00 4.66388375e-01 5.79694080e+00 6.13320231e-01
-1.01107872e+00 -6.84455559e-02 4.07783389e-01 1.59725010e-01
2.26610869e-01 1.14004284e-01 -8.09887290e-01 9.80967641e-01
1.45611417e+00 -4.40592587e-01 3.96977514e-01 1.23585200e+00
6.32545575e-02 -2.95082629e-01 -5.54949284e-01 6.61515474e-01
2.75810838e-01 -9.66394722e-01 -1.93093300e-01 2.57365406e-01
5.99035382e-01 -7.99592584e-02 -1.31939650e-01 -1.64601672e-03
1.97288170e-01 -3.94424886e-01 8.40668306e-02 3.09747644e-02
4.55668032e-01 -8.54133427e-01 1.07268858e+00 1.83485791e-01
-1.12550163e+00 -4.05936152e-01 -5.35090923e-01 4.04251665e-01
-5.81994839e-03 1.00935781e+00 -1.45210481e+00 2.20799476e-01
7.42955625e-01 3.12723339e-01 -8.76654863e-01 1.54867601e+00
1.95817232e-01 6.74675286e-01 -3.68434966e-01 -3.91645432e-01
1.03326619e-01 -1.84211172e-02 4.14862454e-01 1.38922477e+00
4.66478795e-01 -2.92178303e-01 -4.98242453e-02 1.99831754e-01
4.40505385e-01 3.07786465e-01 -4.40275669e-01 7.82382786e-02
9.08166528e-01 1.27703142e+00 -1.00818968e+00 -6.54742837e-01
-3.83300841e-01 9.77681398e-01 -1.85282230e-01 1.03157341e-01
-8.71596932e-01 -1.15932775e+00 2.64062226e-01 4.52888072e-01
6.22967541e-01 -8.08373988e-02 -3.11697870e-01 -1.08149552e+00
1.43492222e-01 -7.84063935e-01 7.66974211e-01 -3.01549658e-02
-1.54379320e+00 3.76062006e-01 -2.47569293e-01 -1.49550474e+00
-2.88016319e-01 -5.38222849e-01 -6.58885479e-01 6.75090313e-01
-1.45779705e+00 -6.07418180e-01 -4.63490158e-01 5.69266915e-01
3.90975356e-01 -1.12394996e-01 5.98618507e-01 1.73940063e-01
-2.67601460e-01 4.16345209e-01 6.11826301e-01 -7.99893215e-02
1.00116873e+00 -1.24234486e+00 5.25750279e-01 8.31956267e-01
-4.06528682e-01 1.02636826e+00 6.25284195e-01 -9.91431773e-01
-1.25281668e+00 -9.32819843e-01 1.09550905e+00 -3.11112940e-01
6.66343331e-01 -2.11560994e-01 -1.24337804e+00 2.07633972e-01
-1.35771334e-01 -3.46297801e-01 9.03137088e-01 1.26916364e-01
-8.87135267e-01 -2.13595122e-01 -1.71303678e+00 2.59079158e-01
6.65601969e-01 -9.03748274e-02 -5.69854081e-01 4.47202444e-01
3.60766113e-01 5.43344975e-01 -9.98915911e-01 -1.25115559e-01
7.66627610e-01 -1.14439070e+00 1.09982991e+00 -1.37660667e-01
-1.33580521e-01 -3.99102777e-01 2.34875828e-02 -9.00460958e-01
-7.74582922e-01 -3.06512326e-01 -2.77303517e-01 1.25690365e+00
1.70817465e-01 -1.01979399e+00 9.32440162e-01 4.23176408e-01
5.22693694e-01 -2.57482678e-01 -6.66121483e-01 -1.22856629e+00
-3.38602930e-01 2.36188039e-01 5.11040509e-01 1.15039480e+00
4.17073190e-01 1.41693890e-01 -2.11711675e-01 1.48350954e-01
1.06994998e+00 2.46812627e-02 4.21317875e-01 -1.71400511e+00
-2.20264897e-01 -3.48565698e-01 -6.06902480e-01 -4.91322428e-01
-4.53018308e-01 -6.49541318e-01 -1.77748859e-01 -1.24252963e+00
4.72435713e-01 -4.68134314e-01 -4.92805839e-01 2.58350968e-01
-2.95503736e-01 1.66810125e-01 -2.53233109e-02 6.69362485e-01
-5.66721618e-01 -2.16435611e-01 4.42839086e-01 3.83327529e-02
-4.21891302e-01 2.15880632e-01 -6.39306843e-01 4.17409837e-01
1.23216617e+00 -5.57860434e-01 -3.99220884e-01 4.97044593e-01
-1.60285592e-01 -4.59633261e-01 6.82597980e-04 -9.34812367e-01
2.50893414e-01 -2.01032072e-01 4.94322062e-01 -6.06699228e-01
-2.40054175e-01 -7.43941665e-01 -1.27099399e-02 1.17906237e+00
-1.06517568e-01 1.69653624e-01 -2.41433084e-01 7.58344531e-01
7.55974352e-02 -7.07917809e-01 9.13232386e-01 -1.71083882e-01
-7.91628003e-01 1.33943856e-02 -4.94170547e-01 -5.04703403e-01
1.50492358e+00 -3.20415020e-01 -5.10960340e-01 -4.81920317e-02
-1.84363835e-02 -1.88104138e-01 7.28424430e-01 3.00791621e-01
3.47062647e-01 -9.68856573e-01 -5.24779558e-01 1.60574526e-01
2.74200499e-01 -8.19396198e-01 1.50574639e-01 4.47616488e-01
-8.03016186e-01 7.20694184e-01 -3.10477138e-01 -5.43443203e-01
-1.55167818e+00 1.02782130e+00 -4.05426830e-01 -3.22677135e-01
-5.27723193e-01 1.33059621e-01 -2.62690872e-01 -9.23478156e-02
4.30170149e-02 4.65985864e-01 -3.62540036e-01 1.28551722e-01
9.63819861e-01 9.32565570e-01 4.01136935e-01 -1.59734160e-01
-4.43463176e-01 2.69529223e-01 -5.56044459e-01 1.69879317e-01
1.28031969e+00 -8.55042487e-02 -1.42307699e-01 8.78007114e-02
1.34328127e+00 -4.61422987e-02 -8.79910409e-01 -2.71655656e-02
5.97619593e-01 -9.80748594e-01 -1.71269104e-01 -9.67293382e-01
-6.61094606e-01 4.62274790e-01 9.33248162e-01 7.66613126e-01
1.05901587e+00 -1.54219508e-01 1.03949261e+00 3.90945673e-01
3.69760871e-01 -1.17246079e+00 1.80718765e-01 2.89526224e-01
2.41509870e-01 -1.11228514e+00 1.17790654e-01 -6.59582496e-01
-3.77822429e-01 1.32854915e+00 3.77415717e-01 -3.49348605e-01
6.61060214e-01 4.98695672e-02 -2.10130259e-01 9.47021879e-04
-6.82957590e-01 3.18771228e-02 -1.52251512e-01 1.02569914e+00
2.04620615e-01 1.46634668e-01 -8.92406464e-01 1.11018442e-01
2.51759797e-01 -1.98752135e-01 5.18122733e-01 1.04278946e+00
-1.04957032e+00 -1.23201776e+00 -5.90734601e-01 7.55363345e-01
-4.07719374e-01 -1.89391226e-01 -4.23056781e-01 6.72164559e-01
-1.18081972e-01 9.49893117e-01 1.11885689e-01 -5.71025133e-01
1.88024983e-01 2.75773942e-01 8.56634229e-02 -2.29740694e-01
-5.10398149e-01 -6.91232234e-02 -4.12075184e-02 -3.69400144e-01
1.06297895e-01 -1.04640555e+00 -7.40516365e-01 -7.06722021e-01
-5.56004107e-01 1.66995734e-01 1.01072419e+00 3.42481703e-01
6.75002992e-01 -2.31274799e-01 1.22379398e+00 -3.49243760e-01
-9.11815822e-01 -7.04863787e-01 -5.66310465e-01 5.95546484e-01
-2.25083362e-02 -3.80008101e-01 -6.36024356e-01 1.40628278e-01] | [7.8212385177612305, 9.981996536254883] |
0945f26e-73af-4199-87d8-39664ed22fa6 | open-information-extraction-from-question | 1903.00172 | null | http://arxiv.org/abs/1903.00172v2 | http://arxiv.org/pdf/1903.00172v2.pdf | Open Information Extraction from Question-Answer Pairs | Open Information Extraction (OpenIE) extracts meaningful structured tuples
from free-form text. Most previous work on OpenIE considers extracting data
from one sentence at a time. We describe NeurON, a system for extracting tuples
from question-answer pairs. Since real questions and answers often contain
precisely the information that users care about, such information is
particularly desirable to extend a knowledge base with.
NeurON addresses several challenges. First, an answer text is often hard to
understand without knowing the question, and second, relevant information can
span multiple sentences. To address these, NeurON formulates extraction as a
multi-source sequence-to-sequence learning task, wherein it combines
distributed representations of a question and an answer to generate knowledge
facts. We describe experiments on two real-world datasets that demonstrate that
NeurON can find a significant number of new and interesting facts to extend a
knowledge base compared to state-of-the-art OpenIE methods. | ['Wang-Chiew Tan', 'Nikita Bhutani', 'Alon Halevy', 'Yoshihiko Suhara', 'H. V. Jagadish'] | 2019-03-01 | open-information-extraction-from-question-1 | https://aclanthology.org/N19-1239 | https://aclanthology.org/N19-1239.pdf | naacl-2019-6 | ['open-information-extraction'] | ['natural-language-processing'] | [ 6.84831068e-02 5.66060662e-01 -3.07996958e-01 -4.58939016e-01
-1.47428930e+00 -8.63334298e-01 1.11383528e-01 6.56410754e-01
-2.14816198e-01 1.44780719e+00 4.90007967e-01 -3.19994837e-01
-1.86203659e-01 -1.22018123e+00 -8.86888146e-01 1.89185351e-01
1.54082462e-01 8.45904469e-01 4.05945897e-01 -5.10325730e-01
1.79266751e-01 -4.62373979e-02 -1.62337971e+00 1.02530158e+00
1.22228956e+00 9.52609479e-01 -9.48050022e-02 7.25492954e-01
-1.20089495e+00 1.37081409e+00 -9.21450734e-01 -9.01455760e-01
1.08648211e-01 -2.87069291e-01 -1.56837988e+00 -2.88475424e-01
4.05901551e-01 -9.33773741e-02 -5.71049824e-02 8.16456735e-01
2.70639807e-01 4.72170748e-02 3.97856385e-01 -1.22399366e+00
-8.34361494e-01 1.02570808e+00 -7.77517911e-03 4.21661079e-01
1.05353391e+00 -8.95588025e-02 1.42319226e+00 -8.06673765e-01
9.97132361e-01 1.06592548e+00 4.73938823e-01 3.48594844e-01
-8.04523230e-01 -4.79300261e-01 3.74090411e-02 4.97253954e-01
-1.05811536e+00 -3.70151669e-01 3.96049678e-01 -5.79756722e-02
1.29787219e+00 5.00805557e-01 3.18409145e-01 7.57371664e-01
2.29519367e-01 1.19942045e+00 5.55697203e-01 -4.40341443e-01
2.26359293e-01 2.95312077e-01 6.27854407e-01 4.30162251e-01
5.17782509e-01 -4.75721002e-01 -6.16221964e-01 -4.52891171e-01
3.98526527e-02 -1.78738028e-01 -2.27771267e-01 1.18248872e-01
-9.78243291e-01 8.27365339e-01 3.69699478e-01 2.48178557e-01
-4.90817994e-01 -2.73060411e-01 4.31919813e-01 5.00927091e-01
3.47707838e-01 9.72645223e-01 -1.06292653e+00 -3.43472689e-01
-5.03520012e-01 8.62205327e-01 1.68131530e+00 1.29648030e+00
1.17113733e+00 -8.16660523e-01 -2.80758947e-01 5.14207184e-01
-1.37667847e-03 2.59469718e-01 4.26863104e-01 -1.06660366e+00
1.02555561e+00 1.12602592e+00 3.25749844e-01 -9.32463169e-01
-2.18301222e-01 1.06711769e-02 -3.55626315e-01 -6.39845133e-01
3.87886286e-01 -5.95775187e-01 -4.23026621e-01 1.30458117e+00
5.39610267e-01 -2.78720379e-01 6.00274563e-01 4.46167886e-01
1.66827869e+00 7.61675715e-01 -1.50049359e-01 -2.41168335e-01
1.53736007e+00 -6.79462552e-01 -1.05926478e+00 -4.81128216e-01
6.54153168e-01 -6.06285334e-01 7.90038228e-01 8.95390660e-02
-8.84276986e-01 -2.41098642e-01 -8.85709465e-01 -4.86984074e-01
-7.54493892e-01 -3.12047899e-01 5.34140706e-01 2.22869873e-01
-4.94585723e-01 3.40467721e-01 -1.69375181e-01 -1.76573187e-01
2.33948842e-01 1.35777473e-01 -3.03550988e-01 -2.87312031e-01
-1.60129821e+00 7.67808080e-01 6.47236645e-01 -3.53470176e-01
7.62214437e-02 -9.78118420e-01 -1.07251775e+00 3.55677098e-01
1.06342340e+00 -1.15174317e+00 1.63149762e+00 -4.24107909e-01
-8.40241313e-01 5.37049413e-01 -8.15570235e-01 -6.75802290e-01
-1.60067409e-01 -4.38468218e-01 -6.49094760e-01 3.14313352e-01
5.27439177e-01 5.62499762e-01 3.60873491e-01 -1.11150169e+00
-7.90028691e-01 -5.29614270e-01 2.93069839e-01 -4.55838330e-02
-1.90899491e-01 2.26663873e-01 -3.15094024e-01 -2.88005471e-01
4.49780300e-02 -4.32436854e-01 -1.91023484e-01 -5.55304885e-01
-7.75220037e-01 -6.15991473e-01 7.47448742e-01 -7.49758244e-01
1.60990751e+00 -1.36224639e+00 -4.36141849e-01 5.88342510e-02
6.81776166e-01 3.47737402e-01 -7.79613182e-02 7.57187545e-01
-6.16658330e-02 3.36923808e-01 -2.95313954e-01 3.63930821e-01
-1.52907800e-02 4.88525301e-01 -7.75132120e-01 -6.13294184e-01
2.43800029e-01 1.45047998e+00 -1.04296494e+00 -8.14236224e-01
-6.38954937e-01 -2.52279371e-01 -4.91481066e-01 2.94939876e-01
-1.09580743e+00 -6.94966465e-02 -6.86159790e-01 6.02716625e-01
5.62417805e-01 -5.08313417e-01 9.56793949e-02 -7.89402612e-03
2.78518438e-01 7.26766884e-01 -1.12661350e+00 1.30876255e+00
-3.25269282e-01 4.62342143e-01 -1.91658989e-01 -6.49515450e-01
1.00268781e+00 4.84937191e-01 4.44035798e-01 -4.64420885e-01
-1.60792265e-02 2.29277700e-01 -3.73799592e-01 -1.22617781e+00
9.06185389e-01 -1.42211542e-01 -3.95714998e-01 7.87838161e-01
2.20314786e-01 -1.91581041e-01 7.71244586e-01 7.18180537e-01
1.56828213e+00 -3.30332667e-01 6.09739244e-01 1.14498399e-01
4.14648175e-01 3.91583592e-01 8.67031753e-01 7.60838866e-01
1.30336806e-01 3.00922304e-01 8.02176952e-01 -5.55854797e-01
-6.45082355e-01 -1.01082814e+00 9.66558680e-02 5.71683168e-01
-6.77941814e-02 -9.61336553e-01 -7.73641706e-01 -1.19081235e+00
3.30555409e-01 8.64582479e-01 -4.39859122e-01 1.32521793e-01
-4.75174844e-01 -1.11548007e-01 5.41102290e-01 5.08291543e-01
4.35069650e-01 -1.26052678e+00 -4.23396260e-01 4.09231246e-01
-1.06599307e+00 -1.39985514e+00 -3.18000585e-01 9.21332985e-02
-6.30782485e-01 -1.44930816e+00 -7.32959360e-02 -6.93783998e-01
4.21696693e-01 6.37820214e-02 1.86286652e+00 3.19736600e-02
-1.96498498e-01 2.13701651e-01 -7.64879525e-01 -7.53491223e-01
-4.76539522e-01 5.15885353e-01 -3.15103799e-01 -2.15917349e-01
1.32275033e+00 -3.84554267e-01 -1.11321375e-01 6.38989210e-02
-1.04536128e+00 -2.59292513e-01 3.51220608e-01 5.88283002e-01
5.28238893e-01 7.16100037e-02 1.25201750e+00 -1.38050318e+00
1.18969727e+00 -8.82941186e-01 -2.63413429e-01 8.69310200e-01
-3.41494322e-01 5.94137311e-01 6.04061306e-01 1.28944144e-01
-1.25762129e+00 -1.23170748e-01 -3.31641138e-01 1.38626009e-01
-3.40647012e-01 9.00113761e-01 -4.85436916e-01 6.38679981e-01
9.60406482e-01 -3.61423679e-02 -1.28677696e-01 -4.00548249e-01
6.86041355e-01 9.90448058e-01 7.67491758e-01 -7.78318882e-01
6.29252791e-01 3.75385918e-02 -4.47054386e-01 -6.61158442e-01
-1.52666807e+00 -7.45891452e-01 -6.69060290e-01 2.78258592e-01
5.84326982e-01 -5.25065601e-01 -8.11656237e-01 -1.85802191e-01
-1.50772882e+00 2.06973270e-01 -8.06309044e-01 -1.59028396e-01
-3.10226291e-01 2.25057647e-01 -4.83846784e-01 -5.29581130e-01
-7.20262051e-01 -5.57409942e-01 7.63316214e-01 4.04419899e-01
-7.63082981e-01 -8.75847399e-01 2.94052988e-01 7.72586942e-01
-5.63210770e-02 1.80379421e-01 1.11883104e+00 -1.12170887e+00
-8.39951456e-01 -2.37422585e-01 -2.03348577e-01 9.07348618e-02
3.29490334e-01 -2.38744989e-01 -6.41918302e-01 3.74078631e-01
1.85529515e-01 -6.41777754e-01 8.07174444e-01 -2.09023595e-01
9.89136398e-01 -9.43762839e-01 -2.54371464e-01 7.68199712e-02
1.13246572e+00 6.64909277e-03 6.28147840e-01 2.11498544e-01
3.67932200e-01 9.18140531e-01 6.96358263e-01 5.65989614e-01
8.62385869e-01 1.27869785e-01 -1.14512973e-01 5.27530909e-01
1.43185079e-01 -5.82855701e-01 -4.71195802e-02 6.54855251e-01
7.10572839e-01 -2.37406194e-01 -8.30688000e-01 8.61505628e-01
-1.92954946e+00 -1.21440494e+00 -3.14588219e-01 1.65901303e+00
1.49260664e+00 -5.54861948e-02 -7.68343359e-02 1.10623807e-01
4.10633713e-01 -1.60960659e-01 -8.11681032e-01 -3.73243004e-01
-2.35557646e-01 3.58219445e-01 1.21057622e-01 4.48671013e-01
-7.66503572e-01 9.15849745e-01 6.34082031e+00 7.02365279e-01
-4.01586324e-01 -1.34610623e-01 3.61833572e-01 -7.38724992e-02
-9.42157626e-01 3.27850789e-01 -1.32411468e+00 2.30058312e-01
1.13823533e+00 -6.95167482e-01 5.04355691e-03 7.23826826e-01
-3.75065237e-01 -4.16232795e-01 -1.35066128e+00 7.62246311e-01
2.02012137e-01 -1.76778793e+00 2.69570291e-01 -2.25120857e-01
7.80166984e-01 -2.97242701e-01 -5.54410219e-01 5.19763887e-01
6.46087170e-01 -9.28330898e-01 2.80767709e-01 7.54346371e-01
3.76957625e-01 -8.52774858e-01 8.50855589e-01 8.21696162e-01
-1.15847862e+00 -2.76732117e-01 -5.11096179e-01 -1.59667283e-01
3.65126818e-01 9.86671090e-01 -1.14040077e+00 7.31804907e-01
6.64893508e-01 6.18703663e-01 -8.01224232e-01 9.20788467e-01
-6.44249558e-01 4.84044760e-01 -3.11900496e-01 -4.46765453e-01
4.97151650e-02 2.13468194e-01 2.46629253e-01 1.05208755e+00
1.15685351e-01 5.28995574e-01 1.26038387e-01 1.06550753e+00
-5.24517715e-01 1.20211810e-01 -8.79631639e-01 -1.29037499e-01
6.25891924e-01 1.17550313e+00 -2.75947928e-01 -6.13841116e-01
-5.78445792e-01 6.08617961e-01 6.01253688e-01 3.65219682e-01
-1.66165084e-01 -9.53342438e-01 6.02471113e-01 -4.10943627e-02
3.96916240e-01 1.59812346e-01 -6.76728845e-01 -1.41182923e+00
5.51121593e-01 -1.15700221e+00 9.55059052e-01 -8.53874683e-01
-1.46904659e+00 5.00996947e-01 -2.35938635e-02 -9.12166595e-01
-8.32089543e-01 -2.60418475e-01 -3.98243248e-01 8.46695364e-01
-1.35991681e+00 -6.68077230e-01 -1.01647951e-01 5.02058029e-01
5.10130942e-01 -2.84720082e-02 8.80366623e-01 1.23159148e-01
-3.06017190e-01 5.07942975e-01 -1.59017071e-01 5.72813272e-01
5.05188882e-01 -1.21858561e+00 8.20088148e-01 7.73547590e-01
4.12250370e-01 9.43466187e-01 5.88374138e-01 -1.02987599e+00
-1.45449686e+00 -9.67406869e-01 1.75206399e+00 -1.00917339e+00
5.86339355e-01 -3.29082519e-01 -1.30500770e+00 8.05486381e-01
3.22410405e-01 -8.47327858e-02 1.13439727e+00 4.70009059e-01
-4.07802880e-01 -4.66920398e-02 -1.13274312e+00 4.08421725e-01
6.72580481e-01 -7.34663427e-01 -1.36125863e+00 3.08167011e-01
1.32084155e+00 -3.63238275e-01 -8.91346753e-01 3.54353577e-01
2.08825395e-01 -9.00147796e-01 7.95534551e-01 -1.06085694e+00
7.70895302e-01 -1.51190370e-01 -6.11222051e-02 -1.26640248e+00
1.75110355e-01 -7.04019189e-01 -8.65235686e-01 1.41797662e+00
1.00537157e+00 -3.79197478e-01 9.66779530e-01 1.20828891e+00
1.41277760e-01 -1.12602568e+00 -7.95214832e-01 -7.45485723e-01
1.24459542e-01 -5.03698885e-01 1.14945138e+00 6.04373574e-01
7.58657753e-01 9.91000712e-01 1.79320842e-01 -7.95803070e-02
2.45358527e-01 6.64027810e-01 9.18189526e-01 -1.32756710e+00
-9.50715765e-02 1.39018789e-01 7.25758970e-02 -1.16229582e+00
3.68076235e-01 -8.07980776e-01 8.61989856e-02 -1.94155872e+00
1.44596636e-01 -3.08876544e-01 2.56138980e-01 4.62311834e-01
-6.22080266e-01 -5.33533931e-01 2.44350433e-02 -1.09884150e-01
-9.75588918e-01 5.81736565e-01 1.12406301e+00 -2.44336843e-01
-2.23579347e-01 7.95931965e-02 -1.41408908e+00 6.05225742e-01
6.51939154e-01 -5.59063971e-01 -6.06776416e-01 -2.14521438e-01
8.90215874e-01 4.15186048e-01 -7.95361176e-02 -7.66912282e-01
5.48072100e-01 -2.48507217e-01 1.47338018e-01 -9.73699808e-01
7.18927309e-02 -5.48174322e-01 -4.91078466e-01 2.97610909e-02
-5.27595818e-01 5.43208197e-02 1.51125088e-01 4.00015414e-01
-6.99912906e-01 -6.26971245e-01 5.90487458e-02 -5.11033535e-01
-7.27752328e-01 2.83451855e-01 -2.68008888e-01 1.02920616e+00
7.16351330e-01 1.18939631e-01 -6.83884740e-01 -6.75946593e-01
-3.21383148e-01 7.41219997e-01 -1.68043137e-01 6.07461154e-01
9.77490783e-01 -1.11710382e+00 -8.29467297e-01 6.88036084e-02
5.02781808e-01 2.77530968e-01 1.19791135e-01 1.81980550e-01
-2.57006556e-01 7.83015728e-01 2.59474814e-01 6.62775040e-02
-1.21170771e+00 8.18719625e-01 -5.15832864e-02 -7.65067518e-01
-4.79037404e-01 8.63473952e-01 -2.91504830e-01 -8.56241763e-01
1.08402997e-01 -5.10948896e-01 -4.71829802e-01 3.30120265e-01
9.37189698e-01 1.64823443e-01 3.27903330e-01 -1.24795035e-01
-1.52893364e-01 -1.92833152e-02 -3.93539011e-01 -8.01815912e-02
1.15668678e+00 -9.93848667e-02 -4.26369995e-01 3.50555778e-01
1.28773463e+00 9.84664075e-03 -6.47672355e-01 -8.15205812e-01
4.50762153e-01 -4.32649225e-01 -5.91756523e-01 -8.74859631e-01
-3.89139116e-01 5.69483757e-01 -4.33764160e-01 3.74759942e-01
8.01602066e-01 3.52552176e-01 1.62407851e+00 1.21034849e+00
3.27015489e-01 -1.05365777e+00 5.22597767e-02 1.14083040e+00
9.61974621e-01 -1.37183392e+00 -1.14764489e-01 -6.38203442e-01
-5.74544072e-01 1.08545053e+00 8.54231954e-01 3.85610223e-01
6.95818424e-01 4.05718088e-01 2.62824059e-01 -4.31784391e-01
-1.19066405e+00 -3.72238487e-01 2.85090238e-01 5.32669008e-01
2.68981725e-01 -2.22773582e-01 -2.26610620e-02 1.22771049e+00
-7.60835469e-01 2.56186664e-01 5.91071188e-01 1.28192365e+00
-7.31711388e-01 -1.34599268e+00 -3.40064377e-01 9.88673925e-01
-4.83952612e-01 -2.22373337e-01 -8.69375706e-01 1.79644316e-01
5.44630215e-02 1.46874642e+00 -1.72423959e-01 -4.44884747e-01
3.66003335e-01 4.82460439e-01 1.06258549e-01 -1.05629253e+00
-7.62912214e-01 -8.48756373e-01 6.02177441e-01 -4.46768135e-01
-7.18076974e-02 -5.89062929e-01 -1.68532205e+00 -2.47894138e-01
-1.40519321e-01 8.16119194e-01 2.88932979e-01 1.42422271e+00
6.96770072e-01 3.40374500e-01 3.20326954e-01 2.76196092e-01
-4.97815102e-01 -7.57011712e-01 -4.27426368e-01 3.51564944e-01
3.74111623e-01 -1.53971493e-01 1.70208633e-01 6.30193800e-02] | [9.867986679077148, 8.481878280639648] |
51e2ff4e-bfd2-4101-9bfe-2d6e8c766550 | influence-of-segmentation-on-deep-iris | 1901.10431 | null | https://arxiv.org/abs/1901.10431v2 | https://arxiv.org/pdf/1901.10431v2.pdf | Influence of segmentation on deep iris recognition performance | Despite the rise of deep learning in numerous areas of computer vision and image processing, iris recognition has not benefited considerably from these trends so far. Most of the existing research on deep iris recognition is focused on new models for generating discriminative and robust iris representations and relies on methodologies akin to traditional iris recognition pipelines. Hence, the proposed models do not approach iris recognition in an end-to-end manner, but rather use standard heuristic iris segmentation (and unwrapping) techniques to produce normalized inputs for the deep learning models. However, because deep learning is able to model very complex data distributions and nonlinear data changes, an obvious question arises. How important is the use of traditional segmentation methods in a deep learning setting? To answer this question, we present in this paper an empirical analysis of the impact of iris segmentation on the performance of deep learning models using a simple two stage pipeline consisting of a segmentation and a recognition step. We evaluate how the accuracy of segmentation influences recognition performance but also examine if segmentation is needed at all. We use the CASIA Thousand and SBVPI datasets for the experiments and report several interesting findings. | ['Vitomir Štruc', 'Dejan Štepec', 'Juš Lozej', 'Peter Peer'] | 2019-01-29 | null | null | null | null | ['iris-segmentation'] | ['medical'] | [ 2.68218338e-01 -5.43905124e-02 -1.07918225e-01 -6.41073167e-01
-2.09689423e-01 -3.45585108e-01 7.12785244e-01 4.98109236e-02
-5.45880556e-01 1.92876488e-01 1.13676369e-01 -4.43444908e-01
-3.48755062e-01 -4.32495058e-01 -3.90643865e-01 -7.16188371e-01
2.19635412e-01 4.98840392e-01 -1.95415959e-01 3.45364492e-03
3.96230608e-01 7.23047018e-01 -1.80426371e+00 8.80010501e-02
1.07641292e+00 6.31740332e-01 -7.56125748e-01 8.74682963e-01
4.04760242e-02 6.42150760e-01 -6.90670311e-01 -4.25653517e-01
5.91371536e-01 -7.41341770e-01 -8.95920098e-01 5.49047291e-01
8.50631714e-01 -3.26044917e-01 -1.38261035e-01 1.02702737e+00
6.57685041e-01 1.66072607e-01 5.20142913e-01 -6.00979567e-01
-5.64654708e-01 3.65013063e-01 -8.24537277e-01 5.23414612e-01
-2.93024909e-02 5.89889169e-01 5.02793550e-01 -3.71900648e-01
4.14109230e-01 9.24665630e-01 6.06938064e-01 3.45152795e-01
-1.39446306e+00 -2.68286556e-01 -1.86611280e-01 -1.49825156e-01
-1.18925464e+00 -5.12405872e-01 3.77323717e-01 -5.92771590e-01
7.20585644e-01 3.64076823e-01 5.39976716e-01 8.10032129e-01
-1.02752872e-01 7.12942421e-01 1.57243860e+00 -6.22521996e-01
-9.71019045e-02 2.06374392e-01 4.56808925e-01 5.96589506e-01
3.74271989e-01 6.22907877e-01 -1.58451736e-01 1.60767645e-01
7.09279180e-01 -1.51419565e-01 8.76328126e-02 -2.27972418e-01
-9.36192751e-01 6.85111821e-01 3.39048415e-01 5.90939999e-01
-4.52480614e-01 -1.79985985e-01 2.39370093e-01 4.18683708e-01
3.36124867e-01 7.21018970e-01 -2.17892095e-01 -2.90005296e-01
-1.46616387e+00 1.25882447e-01 5.56091130e-01 1.55549511e-01
5.69512546e-01 4.42498885e-02 -4.44599688e-01 7.51311183e-01
1.18071616e-01 1.00604668e-01 5.45515895e-01 -4.08690721e-01
-1.66524902e-01 9.93598759e-01 -1.12896658e-01 -8.27927053e-01
-7.05160975e-01 -8.47868025e-01 -7.18043625e-01 4.56678599e-01
9.30621922e-01 -4.39031929e-01 -1.66717732e+00 1.19619524e+00
3.74434501e-01 3.02670836e-01 -1.04669467e-01 1.07219231e+00
6.91227853e-01 2.59507485e-02 -1.66614074e-02 2.25592181e-01
1.00033855e+00 -8.54083598e-01 -3.66677076e-01 -1.31066486e-01
5.98362446e-01 -1.17361987e+00 9.77944732e-01 4.65823203e-01
-9.20943499e-01 -5.63437104e-01 -7.65426636e-01 -2.74655819e-01
-3.86549830e-01 4.56233531e-01 7.20296860e-01 1.01674509e+00
-1.07859612e+00 6.30832493e-01 -8.76877904e-01 -7.12786794e-01
4.39814806e-01 8.00753355e-01 -1.06454827e-01 1.40212364e-02
-5.91776192e-01 8.81040394e-01 2.79529244e-01 2.70364940e-01
-3.66024762e-01 -3.32293153e-01 -6.45887673e-01 -9.99520719e-02
7.16110766e-02 -6.18644893e-01 1.12014401e+00 -1.51679766e+00
-1.61230934e+00 1.18131924e+00 -2.06867829e-01 -6.18804097e-01
5.14205992e-01 -1.36007443e-01 -1.30658001e-01 -1.59442499e-01
-4.92072523e-01 4.64327931e-01 9.85349596e-01 -8.87204349e-01
-7.06624508e-01 -5.33449233e-01 9.32293683e-02 -2.50547677e-02
1.87558793e-02 4.03016061e-01 -4.82728541e-01 -6.51617289e-01
-9.19068903e-02 -1.09427476e+00 -3.36921215e-01 -3.52044016e-01
-4.27368969e-01 -1.40178546e-01 2.67041534e-01 -6.75421059e-01
1.08026135e+00 -2.19737935e+00 4.63302173e-02 4.17295545e-01
2.23302692e-01 8.35613847e-01 -1.64166361e-01 -2.52546519e-02
-2.05039337e-01 1.29453942e-01 -3.47405851e-01 -4.25807685e-01
-1.39550582e-01 2.36022294e-01 -6.17918298e-02 6.37554765e-01
2.59372383e-01 8.68831456e-01 -6.52508199e-01 -1.59646496e-01
6.59587920e-01 6.83289766e-01 -4.86928076e-01 1.17644630e-01
-1.43316239e-02 5.54364443e-01 -8.98639858e-02 8.97998214e-01
6.34107590e-01 -2.15612650e-01 -1.41900659e-01 5.97136244e-02
-2.67157525e-01 2.46504936e-02 -1.02157629e+00 1.39897573e+00
-3.80736999e-02 7.33789444e-01 -1.81512445e-01 -1.08368409e+00
7.77882516e-01 1.62896186e-01 4.32401061e-01 -8.64797950e-01
4.00879920e-01 2.55424321e-01 5.30469835e-01 -3.32018763e-01
4.32157576e-01 -1.56288236e-01 4.39385146e-01 4.91599709e-01
-2.17696559e-02 2.89743811e-01 3.08866203e-01 -3.16371083e-01
7.78129637e-01 1.82885736e-01 1.62725925e-01 -1.22459941e-01
4.89630848e-01 2.28181198e-01 4.27026272e-01 7.98258662e-01
-3.79007459e-01 7.15409219e-01 5.25118530e-01 -7.91478813e-01
-7.44278073e-01 -6.46137297e-01 -4.40169126e-01 8.34121704e-01
-1.33031756e-01 -1.03740692e-01 -9.37139034e-01 -6.87660694e-01
-1.02976803e-02 3.66545618e-01 -8.81358981e-01 -3.52181047e-02
-3.14981997e-01 -1.48039353e+00 6.24371886e-01 2.10515752e-01
3.29232544e-01 -1.09046721e+00 -7.43556321e-01 -6.82041422e-02
3.48379344e-01 -6.94824517e-01 -1.79241851e-01 -1.50694638e-01
-9.78410661e-01 -1.43569839e+00 -7.05913723e-01 -5.04516006e-01
8.95337999e-01 -1.45487282e-02 1.01230741e+00 3.94021839e-01
-6.52005672e-01 1.47782564e-01 -2.41375551e-01 -5.77871025e-01
-3.76075685e-01 2.65293181e-01 -2.74022901e-03 5.26201546e-01
1.09574592e+00 -7.02019259e-02 -6.21501029e-01 1.07465744e-01
-1.04271257e+00 -8.65043700e-02 1.08312011e+00 8.35079253e-01
5.13369381e-01 -5.42572979e-03 1.48383200e-01 -9.91799772e-01
6.62034512e-01 -8.61047879e-02 -8.22176635e-01 2.03203633e-01
-8.19666862e-01 1.08694687e-01 3.03842515e-01 -3.38419825e-01
-8.06231856e-01 2.27015093e-01 -2.69125104e-01 -3.06490809e-01
-5.97236812e-01 6.67088807e-01 5.11286199e-01 -4.14196432e-01
8.06614041e-01 1.27482519e-01 2.57524550e-01 -6.69706821e-01
2.00499341e-01 7.83131778e-01 4.60031152e-01 -2.46254951e-01
8.16274405e-01 3.65211248e-01 -9.17638391e-02 -9.28703368e-01
-6.91518188e-01 -5.32999337e-01 -8.42657983e-01 -6.54066866e-03
7.53775835e-01 -5.07619321e-01 -5.51307857e-01 8.17770243e-01
-6.23007298e-01 -4.20939893e-01 -5.06781161e-01 4.84693080e-01
-2.51621068e-01 3.17693919e-01 -4.36369985e-01 -5.64361870e-01
-1.99081048e-01 -1.59737861e+00 6.93427861e-01 8.74179423e-01
-2.36062214e-01 -8.87812138e-01 3.23867589e-01 6.23161674e-01
5.98700821e-01 4.37637001e-01 6.86187029e-01 -6.85944974e-01
-5.28828442e-01 -3.47605944e-01 -4.31637049e-01 4.44652617e-01
3.44616205e-01 3.57083738e-01 -1.25376213e+00 -4.09967363e-01
-1.90412864e-01 -2.35329479e-01 9.46802318e-01 5.67867994e-01
1.12613440e+00 -1.20343819e-01 3.69199887e-02 1.12040257e+00
1.41615927e+00 4.25090045e-02 7.80174434e-01 4.93668556e-01
5.24672031e-01 7.13401914e-01 2.01799154e-01 1.05570881e-02
5.40647879e-02 5.44586658e-01 2.59863436e-01 -7.53314555e-01
-2.58650124e-01 4.76320498e-02 1.58852970e-04 2.41121769e-01
-3.63322556e-01 1.75575554e-01 -1.27855122e+00 8.84241581e-01
-1.61796105e+00 -7.45693982e-01 -2.35303435e-02 2.46446157e+00
9.31038380e-01 -2.62611527e-02 2.73251027e-01 8.04898590e-02
1.98504388e-01 -3.85813378e-02 -6.17822886e-01 -7.26566494e-01
-2.42741546e-03 6.27431810e-01 4.41278517e-01 5.19975960e-01
-1.25939763e+00 9.33684826e-01 6.57974195e+00 3.48862231e-01
-1.61674666e+00 -3.13735902e-01 9.45961177e-01 -1.86921135e-01
2.68753648e-01 -6.01829961e-02 -6.31142557e-01 3.25527459e-01
1.06649232e+00 3.81038576e-01 4.04503345e-01 4.35298741e-01
2.02305645e-01 -2.28283957e-01 -9.19495642e-01 9.17911649e-01
-5.72277904e-02 -1.14236259e+00 -1.73551179e-02 2.95210361e-01
7.47658908e-01 2.09902197e-01 5.13617337e-01 1.19769983e-01
2.43883923e-01 -1.45035684e+00 -1.05343558e-01 7.40803421e-01
5.46864510e-01 -7.34181941e-01 1.03140283e+00 4.43232358e-02
-4.27423120e-01 -1.13065869e-01 -1.51295677e-01 -1.44738019e-01
-3.82864833e-01 4.90659177e-01 -9.72097754e-01 3.15394610e-01
5.11802733e-01 5.79592705e-01 -1.16752040e+00 1.55619931e+00
-1.42118856e-02 9.20202851e-01 -3.68746042e-01 3.79944801e-01
4.74094570e-01 -3.02663743e-01 4.05586243e-01 1.25189710e+00
-3.89610641e-02 -1.28658876e-01 -1.07118152e-01 7.33311594e-01
2.28447199e-04 2.23376006e-01 -3.45705122e-01 -3.25046510e-01
-1.73909113e-01 1.06534588e+00 -7.88378954e-01 -1.05778381e-01
-6.65676951e-01 6.48944139e-01 8.87397975e-02 4.88268197e-01
-3.53100181e-01 -3.03921849e-01 8.62291873e-01 3.53980869e-01
1.63022235e-01 -1.69216599e-02 -7.63829887e-01 -1.05818641e+00
-1.72643512e-01 -1.40611875e+00 2.79091358e-01 -1.54166400e-01
-1.03905129e+00 4.74613845e-01 -3.80172044e-01 -8.37876737e-01
-2.90129662e-01 -5.40808320e-01 -6.23618186e-01 1.13991821e+00
-1.80723906e+00 -1.20034957e+00 -2.34937578e-01 5.04549742e-01
2.43484661e-01 -2.53631771e-01 5.91704369e-01 1.96701452e-01
-1.06278992e+00 8.11122954e-01 2.13698730e-01 5.87202907e-01
8.28228474e-01 -1.41407883e+00 5.56812167e-01 1.43285942e+00
5.04196465e-01 8.39187741e-01 6.82287097e-01 -3.84593129e-01
-9.51385081e-01 -8.41757357e-01 9.16943014e-01 -6.05587840e-01
3.06816936e-01 1.83711067e-01 -9.33936179e-01 6.38886034e-01
3.58383328e-01 9.98032466e-02 8.25013280e-01 4.95640606e-01
-6.68347627e-02 -5.57531416e-02 -1.06394815e+00 4.70573217e-01
4.49947357e-01 -2.92694896e-01 -6.27438247e-01 8.50523114e-02
-3.44460122e-02 -6.85916185e-01 -7.53014743e-01 6.05850637e-01
4.54173744e-01 -1.34068584e+00 8.45322073e-01 -8.63707840e-01
3.23298812e-01 -3.25390428e-01 4.55434889e-01 -1.06113410e+00
-1.17710248e-01 -5.94957590e-01 4.32099886e-02 8.15432012e-01
4.22629863e-01 -6.13191307e-01 9.17294502e-01 8.78410280e-01
1.59911633e-01 -9.06299829e-01 -7.07536995e-01 -2.61654168e-01
1.44275561e-01 5.22906296e-02 6.16787434e-01 9.89550591e-01
-4.87924814e-01 -1.00118723e-02 -1.60965234e-01 2.75419056e-01
6.11187279e-01 4.18974280e-01 1.00531673e+00 -1.34849524e+00
-3.55682850e-01 -7.86744535e-01 -6.69534743e-01 -6.33244038e-01
-2.84601152e-01 -5.74328184e-01 -2.29697704e-01 -1.23150468e+00
-8.91823694e-02 -4.37115997e-01 -4.35997546e-01 6.44535303e-01
-3.68171394e-01 5.00146031e-01 6.06293306e-02 2.27945998e-01
-1.64189935e-01 -1.07921802e-01 9.82560813e-01 -1.63479969e-01
-5.18996596e-01 3.86651754e-01 -8.75589311e-01 6.26686156e-01
7.42680073e-01 -7.52745718e-02 -2.56160170e-01 -4.90798742e-01
-2.49874946e-02 -4.83544469e-01 3.47578883e-01 -9.81830716e-01
3.49371046e-01 1.55143510e-03 5.24380684e-01 -1.38541713e-01
-1.06082045e-01 -5.47214866e-01 -2.96791345e-01 3.27329695e-01
-1.83979973e-01 -2.11405054e-01 5.23565710e-01 8.45875815e-02
-4.58323002e-01 -5.52306846e-02 1.09949517e+00 4.07585651e-02
-6.39487624e-01 2.69694805e-01 -2.24172041e-01 -1.68629542e-01
7.73248434e-01 -5.40374935e-01 -1.52940616e-01 1.24024130e-01
-8.58252585e-01 3.25681083e-02 8.48123670e-01 5.56033731e-01
3.46114397e-01 -5.82141221e-01 -7.67034590e-01 7.95527220e-01
1.04385391e-01 4.77814972e-02 -5.10283858e-02 1.10153592e+00
-7.40691125e-01 5.65790296e-01 -1.83587685e-01 -7.37904072e-01
-1.38977647e+00 4.10577685e-01 9.14250076e-01 -3.03798199e-01
-5.74046433e-01 9.15657699e-01 -1.65197596e-01 -4.50678945e-01
4.12308216e-01 -4.99537706e-01 -4.32083875e-01 3.41378078e-02
6.22493565e-01 6.01408295e-02 3.20330918e-01 -6.83768332e-01
3.98298800e-02 5.71522415e-01 -4.83842373e-01 3.40780407e-01
1.15513873e+00 1.45702869e-01 -1.94441453e-01 8.59261677e-03
5.85262418e-01 -1.98181123e-01 -1.00916481e+00 -2.80169696e-01
2.84706384e-01 -8.00917387e-01 4.40755785e-01 -1.22577131e+00
-1.21566057e+00 9.10028219e-01 1.12971318e+00 7.70205539e-03
1.31733334e+00 -4.00865078e-01 5.84256828e-01 2.80918106e-02
-1.10476471e-01 -1.00837970e+00 -4.96400744e-01 3.21099162e-01
4.53108937e-01 -1.50895298e+00 3.09289116e-02 1.36193871e-01
-4.22448277e-01 9.60493445e-01 4.85136479e-01 2.78548561e-02
5.09980381e-01 3.10870651e-02 5.58101952e-01 -3.08794409e-01
-2.08788514e-01 -6.54610157e-01 6.80879712e-01 5.16475320e-01
7.35245049e-01 2.54365448e-02 -2.58606285e-01 -6.16516694e-02
-1.07470639e-01 3.55880946e-01 5.98459780e-01 6.63760364e-01
-2.27731526e-01 -1.42859948e+00 -4.28959399e-01 8.26627433e-01
-7.54180372e-01 3.07063647e-02 -6.86931133e-01 6.87187791e-01
3.23329955e-01 8.50423574e-01 7.40343928e-02 -3.90313357e-01
2.82893687e-01 6.70866370e-02 3.00341994e-01 -7.86116481e-01
-1.00947368e+00 4.63323267e-05 -3.07747632e-01 -5.07030368e-01
-6.64713919e-01 -7.85242736e-01 -7.73417771e-01 -3.69202077e-01
-1.75564200e-01 -1.50104538e-01 7.12374926e-01 1.15164351e+00
4.00899827e-01 4.10226673e-01 2.81103790e-01 -5.25368750e-01
-6.01826549e-01 -9.49975610e-01 -3.71875167e-01 4.53157663e-01
7.25651622e-01 -4.07317907e-01 -3.06422502e-01 -8.51146728e-02] | [3.743502378463745, -3.6320319175720215] |
ceb769ff-c60d-4514-a619-946fd4e146d8 | self-pair-synthesizing-changes-from-single | 2212.10236 | null | https://arxiv.org/abs/2212.10236v1 | https://arxiv.org/pdf/2212.10236v1.pdf | Self-Pair: Synthesizing Changes from Single Source for Object Change Detection in Remote Sensing Imagery | For change detection in remote sensing, constructing a training dataset for deep learning models is difficult due to the requirements of bi-temporal supervision. To overcome this issue, single-temporal supervision which treats change labels as the difference of two semantic masks has been proposed. This novel method trains a change detector using two spatially unrelated images with corresponding semantic labels such as building. However, training on unpaired datasets could confuse the change detector in the case of pixels that are labeled unchanged but are visually significantly different. In order to maintain the visual similarity in unchanged area, in this paper, we emphasize that the change originates from the source image and show that manipulating the source image as an after-image is crucial to the performance of change detection. Extensive experiments demonstrate the importance of maintaining visual information between pre- and post-event images, and our method outperforms existing methods based on single-temporal supervision. code is available at https://github.com/seominseok0429/Self-Pair-for-Change-Detection. | ['Junghoon Seo', 'Yongjin Jeon', 'Hakjin Lee', 'Minseok Seo'] | 2022-12-20 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 5.13092101e-01 -2.25968152e-01 3.14013243e-01 -6.49222612e-01
-4.36309040e-01 -6.81147099e-01 5.65543234e-01 1.72166303e-01
-4.30854678e-01 6.09676242e-01 -7.98552334e-02 -2.81591326e-01
-2.23717000e-02 -8.19602251e-01 -8.44468415e-01 -8.55377614e-01
9.26619917e-02 -2.55387694e-01 5.32123923e-01 -7.88279101e-02
-6.28377274e-02 4.15123373e-01 -1.63110173e+00 1.91285059e-01
7.39326358e-01 6.90657198e-01 5.52112579e-01 3.47211182e-01
8.28669593e-02 3.21283877e-01 -3.35453421e-01 3.15997675e-02
7.05997527e-01 -7.05171824e-01 -6.68913603e-01 1.29483625e-01
6.89659894e-01 -4.31107521e-01 -2.09040552e-01 1.37796128e+00
4.15589839e-01 2.01370224e-01 4.20771390e-01 -1.21961319e+00
-5.58396220e-01 2.94252425e-01 -7.59099782e-01 6.02611780e-01
-1.75693735e-01 9.02850330e-02 9.49296117e-01 -8.55672240e-01
6.09332085e-01 8.40742290e-01 6.35140300e-01 1.15708984e-01
-1.24281597e+00 -7.72381008e-01 5.90750635e-01 3.36266428e-01
-1.35638928e+00 -5.70250750e-01 9.94775057e-01 -5.07819533e-01
6.36526167e-01 2.92874634e-01 6.97609305e-01 7.79265642e-01
4.44174707e-02 5.31188309e-01 1.19978917e+00 -4.11039859e-01
3.32964659e-01 -6.30378872e-02 -9.11859702e-03 5.50923049e-01
2.52260745e-01 1.71083987e-01 -2.38046631e-01 2.92176366e-01
7.47647941e-01 4.43482965e-01 -5.79607487e-01 -3.80133748e-01
-1.22500956e+00 6.11088514e-01 9.69167769e-01 5.78270912e-01
-2.64933050e-01 8.13049078e-02 2.31155101e-02 4.34901148e-01
6.29641712e-01 1.33485571e-01 -4.82957065e-01 3.07773232e-01
-1.03651619e+00 -2.39675730e-01 1.44452259e-01 4.34712708e-01
1.06184411e+00 -4.34848666e-02 8.59871283e-02 5.44334173e-01
2.58535683e-01 6.50573313e-01 2.80211598e-01 -6.43166721e-01
3.14255893e-01 4.06185865e-01 1.98694661e-01 -1.10332859e+00
-2.98023522e-01 -4.50781941e-01 -8.47203612e-01 5.50758481e-01
3.93171698e-01 7.39523321e-02 -1.23752356e+00 1.71821213e+00
5.67113936e-01 1.79938942e-01 -8.85907486e-02 9.23608780e-01
7.48151362e-01 4.97961134e-01 -1.16884515e-01 -5.26888482e-02
9.87503409e-01 -8.88355732e-01 -6.26551449e-01 -3.96635801e-01
4.08792377e-01 -4.87978131e-01 1.15921926e+00 -1.91581920e-01
-5.88147581e-01 -6.20312870e-01 -1.06031048e+00 -4.77268873e-03
-6.07716203e-01 1.01584367e-01 2.81564265e-01 1.46994069e-01
-1.04154456e+00 5.19701600e-01 -9.96362686e-01 -6.14190638e-01
3.80322039e-01 -9.25303996e-02 -4.59933519e-01 2.25626063e-02
-1.23935723e+00 7.47792542e-01 2.71748841e-01 3.79750937e-01
-9.31334257e-01 -5.06502688e-01 -8.06895852e-01 -3.57035697e-02
1.22162767e-01 -3.41728061e-01 1.01015437e+00 -1.52667296e+00
-1.01288450e+00 1.04160821e+00 -2.46555462e-01 -1.23055227e-01
7.33922362e-01 -7.54822232e-03 -4.49431688e-01 1.33607462e-01
3.68952423e-01 7.70669281e-01 1.00920534e+00 -1.54536366e+00
-7.99152255e-01 -4.46961164e-01 1.69237524e-01 3.04696649e-01
1.83251090e-02 -1.53950796e-01 -2.77044266e-01 -7.01363683e-01
5.21870911e-01 -9.27475214e-01 -8.53967816e-02 3.61205399e-01
-2.17683703e-01 2.72267967e-01 1.08896089e+00 -8.99478018e-01
9.66485441e-01 -2.38913369e+00 -4.03869659e-01 1.15484394e-01
-4.03619073e-02 2.30668202e-01 -2.35667467e-01 1.93881646e-01
-3.52761894e-01 1.86008513e-01 -7.83063114e-01 -1.00329831e-01
-3.10410738e-01 1.21349864e-01 -2.81240851e-01 6.72452331e-01
3.99931908e-01 7.23279536e-01 -1.05075014e+00 -3.10155839e-01
3.40514511e-01 3.27845663e-01 -1.27854884e-01 1.74327232e-02
-5.23591368e-03 8.47506166e-01 -2.64388770e-01 5.54082870e-01
8.65354538e-01 -1.58368483e-01 1.74319148e-02 -1.97363615e-01
-2.76244402e-01 2.80063510e-01 -1.27477586e+00 1.48486042e+00
-1.75558835e-01 8.32626581e-01 1.28445670e-01 -9.57421422e-01
7.57588565e-01 8.04768279e-02 4.14090395e-01 -9.48511541e-01
-1.40513152e-01 7.49674588e-02 7.48109538e-03 -4.34062839e-01
2.44420052e-01 -2.01234594e-01 2.33121186e-01 3.98814887e-01
-5.68693757e-01 -2.25542173e-01 4.59420606e-02 -1.30207673e-01
1.04143715e+00 2.05943167e-01 2.83672035e-01 -1.30964980e-01
2.06470802e-01 9.01875645e-03 8.23779702e-01 7.90008903e-01
-6.37944520e-01 8.24881673e-01 -8.08273405e-02 -3.78556967e-01
-7.28609800e-01 -1.18283927e+00 -4.16874379e-01 8.86499941e-01
5.59659064e-01 1.13573223e-01 -3.64018857e-01 -6.96723998e-01
9.19329301e-02 6.46616101e-01 -6.73132956e-01 -6.88661933e-02
-4.09471005e-01 -6.43011093e-01 2.81401962e-01 4.74426866e-01
8.59100461e-01 -9.49155807e-01 -8.37138832e-01 1.14854731e-01
-3.14150423e-01 -7.81309605e-01 -3.09106857e-01 4.08325672e-01
-8.13750029e-01 -9.85029936e-01 -3.80288094e-01 -9.41608906e-01
9.72745180e-01 9.37954962e-01 7.58746564e-01 1.51133701e-01
-1.49702713e-01 1.67394385e-01 -4.01292711e-01 -1.68871760e-01
-1.58840746e-01 -2.89368778e-01 -3.33836168e-01 7.90326446e-02
1.73405603e-01 -7.47186661e-01 -7.95753598e-01 2.97699094e-01
-1.10601568e+00 2.65719354e-01 3.52859855e-01 7.48439431e-01
5.85999012e-01 4.64065075e-01 3.73628348e-01 -5.00590384e-01
-1.38263851e-01 -3.38941157e-01 -5.51184595e-01 3.56946111e-01
-5.47376812e-01 -7.75458068e-02 2.02191919e-01 -1.61853522e-01
-1.34288573e+00 3.00599098e-01 1.54260159e-01 -2.21804589e-01
-3.44449967e-01 4.17315871e-01 -1.78774461e-01 7.26250708e-02
5.76049745e-01 2.09700510e-01 -1.19979329e-01 -4.38288808e-01
3.41696411e-01 6.53647184e-01 4.43430334e-01 6.29370734e-02
1.02028430e+00 1.12322199e+00 -3.58952284e-01 -6.34134412e-01
-8.14372063e-01 -5.98766804e-01 -9.83331144e-01 -3.03013712e-01
7.28176117e-01 -1.19976366e+00 2.22236142e-01 7.61154711e-01
-9.42108691e-01 -5.32155991e-01 -2.73254097e-01 4.16061759e-01
-7.53466114e-02 3.31620425e-01 -3.20710629e-01 -5.10871947e-01
-1.03414856e-01 -7.71066546e-01 1.00254023e+00 2.90118396e-01
1.45182177e-01 -9.40787792e-01 1.18948914e-01 1.01131022e-01
4.24420565e-01 3.52267385e-01 6.59110129e-01 -2.32791752e-01
-5.53113461e-01 1.33914411e-01 -3.18443835e-01 3.75232875e-01
8.69634151e-01 1.53337494e-01 -1.10992646e+00 -3.73349488e-01
1.17614396e-01 4.06357944e-02 1.11988950e+00 4.08099055e-01
8.36561203e-01 1.03178043e-02 -4.33778018e-01 5.81940472e-01
1.61328208e+00 3.14656317e-01 6.19815052e-01 5.93061805e-01
9.16824698e-01 6.36244178e-01 5.42930186e-01 3.51710707e-01
4.56994206e-01 5.60206234e-01 5.16470015e-01 -7.09603846e-01
-3.05216193e-01 -5.96986152e-02 4.15503383e-01 2.59392977e-01
2.22338170e-01 -1.07466199e-01 -9.35474455e-01 8.71888936e-01
-1.84911442e+00 -1.16686666e+00 -4.13963020e-01 2.27874422e+00
1.01206207e+00 -9.20105055e-02 -3.44714940e-01 1.47480279e-01
1.09581769e+00 4.56573933e-01 -7.80720592e-01 2.17607558e-01
-3.36430430e-01 -1.23259820e-01 4.98701602e-01 6.57974184e-01
-1.40118873e+00 9.76912498e-01 5.46382856e+00 3.58518183e-01
-1.49705100e+00 3.93781602e-01 5.28447628e-01 -1.42488793e-01
-3.89085263e-01 2.39702508e-01 -5.32099605e-01 4.73853052e-01
2.90276587e-01 1.41052932e-01 1.63702026e-01 4.66198057e-01
6.30195677e-01 -4.85897243e-01 -9.18924510e-01 7.23247886e-01
4.35531922e-02 -7.09552169e-01 -1.50713965e-01 -2.72054791e-01
1.11138856e+00 3.32851499e-01 -1.16479337e-01 -9.04357657e-02
2.35288337e-01 -5.06125569e-01 8.59286427e-01 3.94921929e-01
6.24179244e-01 -1.49089202e-01 5.14330864e-01 8.02769884e-03
-1.30957758e+00 1.44548237e-01 -2.09364980e-01 -3.17097932e-01
1.18785843e-01 9.04081047e-01 -6.31946921e-01 4.95853245e-01
1.10855603e+00 1.15135372e+00 -8.54391217e-01 9.04758632e-01
-5.60792625e-01 5.49871147e-01 -4.89632249e-01 6.12736166e-01
9.96818915e-02 -3.87929678e-01 4.67500240e-01 9.27192569e-01
3.43384743e-01 -5.22701144e-02 1.92222849e-01 9.06850040e-01
2.50445306e-01 -2.89214551e-01 -7.03728259e-01 1.94291726e-01
5.12139440e-01 1.08531404e+00 -9.97397006e-01 -2.99562901e-01
-4.79766577e-01 1.28185046e+00 1.00630984e-01 5.30228436e-01
-8.80250871e-01 -3.06704372e-01 6.64121330e-01 8.66160318e-02
6.39048755e-01 -2.50611931e-01 -3.82510364e-01 -1.23291707e+00
5.17231524e-01 -4.75705624e-01 4.09878552e-01 -9.58940923e-01
-1.06125319e+00 3.77719045e-01 3.48642319e-02 -1.49502647e+00
2.57137030e-01 -8.92476141e-02 -8.32045436e-01 7.00243056e-01
-2.01402473e+00 -1.20822287e+00 -7.97297716e-01 5.44691980e-01
3.31043005e-01 5.43284893e-01 4.08053666e-01 3.01516384e-01
-6.34032667e-01 1.78101674e-01 4.69588429e-01 1.50244758e-01
8.60962629e-01 -1.09482706e+00 4.05316740e-01 1.53442538e+00
1.97917774e-01 2.43637517e-01 6.20240271e-01 -7.55308032e-01
-5.57002664e-01 -1.34121811e+00 7.33272254e-01 -2.21495017e-01
6.04422748e-01 -1.34024382e-01 -1.21137989e+00 6.78030431e-01
6.96554631e-02 1.18534341e-01 3.76253515e-01 -2.49392092e-01
-3.65718722e-01 -3.81599456e-01 -9.65378344e-01 5.10134995e-01
1.23930955e+00 -7.02078402e-01 -6.79578722e-01 2.66307771e-01
5.26213586e-01 -2.10853666e-01 -5.08235872e-01 4.90878344e-01
2.06968173e-01 -1.08098757e+00 6.61113441e-01 -1.07475229e-01
2.61885732e-01 -1.01964772e+00 -1.05030663e-01 -1.35215008e+00
-4.46192414e-01 -5.36263101e-02 5.36218762e-01 1.26023114e+00
2.12237597e-01 -7.34187961e-01 3.06860238e-01 3.63672644e-01
-1.14718065e-01 -5.29913120e-02 -9.12193060e-01 -9.31030393e-01
-5.62140718e-02 -9.61686000e-02 5.29096901e-01 1.30855620e+00
-4.88529831e-01 1.47359222e-01 -7.77423233e-02 7.86012650e-01
2.40681440e-01 4.38915044e-01 5.57515860e-01 -1.20877028e+00
-5.33962920e-02 -3.83593559e-01 -4.35302377e-01 -6.70528412e-01
5.18461168e-02 -8.64286423e-01 4.11652207e-01 -1.74065292e+00
3.13181460e-01 -4.11703885e-01 -5.00623882e-01 1.07627630e+00
-4.40044135e-01 3.22025508e-01 5.42390160e-03 4.28570211e-01
-3.54799271e-01 7.59102881e-01 1.03091180e+00 -3.82729262e-01
-2.68794984e-01 -3.05622816e-01 -3.78073722e-01 5.23558855e-01
1.00473309e+00 -7.18648314e-01 -3.63440305e-01 -5.96112251e-01
-1.85975451e-02 -5.96953988e-01 7.07780480e-01 -9.90236580e-01
8.65362957e-02 -2.50163823e-01 3.23953152e-01 -5.73210478e-01
-6.59826025e-03 -1.02953649e+00 5.70649743e-01 7.12744534e-01
-9.35108215e-02 6.65595308e-02 1.88591436e-01 6.19312942e-01
-4.21173602e-01 -2.34345213e-01 9.07942533e-01 -1.70651361e-01
-1.17956543e+00 8.44397545e-02 -2.25537568e-01 -9.25823003e-02
9.93274033e-01 -3.75764221e-01 -4.50732291e-01 -2.90008575e-01
-4.97100055e-01 1.07391737e-01 8.61656249e-01 4.76175785e-01
4.35847044e-01 -1.14845479e+00 -6.91843927e-01 2.19092354e-01
4.09673631e-01 1.12035759e-01 3.96000862e-01 8.32456589e-01
-3.94874901e-01 -8.35617632e-02 -3.56882632e-01 -6.93273067e-01
-1.42390728e+00 3.52894187e-01 6.08832777e-01 1.30956337e-01
-8.00851941e-01 6.61786556e-01 4.90745813e-01 -3.72328043e-01
-2.86925882e-02 -6.44257665e-01 1.59934014e-01 2.16091320e-01
4.04427916e-01 -1.69902649e-02 2.35755667e-01 -5.47489703e-01
-5.91034710e-01 5.50947845e-01 -1.50184976e-02 -1.31656557e-01
1.40720582e+00 -5.27213812e-01 -1.36924461e-01 6.91666186e-01
1.12183714e+00 -1.73480436e-01 -1.64470065e+00 -6.19586110e-01
-2.27360338e-01 -7.01405942e-01 4.03972059e-01 -7.92709410e-01
-1.26294672e+00 6.99322581e-01 1.16729009e+00 4.18274924e-02
1.31051123e+00 -6.19189441e-02 3.71552616e-01 4.46877003e-01
2.17076883e-01 -1.04917073e+00 1.89830512e-02 4.56268609e-01
1.01087463e+00 -1.86199450e+00 -5.11060283e-02 -3.02709997e-01
-4.62664008e-01 7.40954101e-01 5.96721470e-01 2.22024873e-01
8.77004325e-01 -8.34551454e-02 4.85741258e-01 -2.46266007e-01
-2.36581936e-01 -6.82333231e-01 1.52100950e-01 5.22359073e-01
1.59450918e-01 2.02201843e-01 -1.20282553e-01 -1.16579756e-01
6.79949969e-02 -2.86343396e-01 4.19753730e-01 1.30623293e+00
-3.92298371e-01 -7.86794245e-01 -4.03730780e-01 2.61126429e-01
-1.18930489e-02 -2.91750789e-01 -4.22320276e-01 7.50553131e-01
2.93240458e-01 1.10383201e+00 3.70898962e-01 -2.20712841e-01
1.63721025e-01 -1.52670324e-01 2.09254026e-01 -5.83372176e-01
-4.67464536e-01 1.73880719e-02 -1.24901362e-01 -3.77661407e-01
-9.10516679e-01 -8.80583644e-01 -1.35659504e+00 -1.65034477e-02
-3.83301407e-01 -1.04818471e-01 6.81178510e-01 8.61576855e-01
4.62014765e-01 4.01038796e-01 9.12698984e-01 -7.85409033e-01
-2.23829791e-01 -9.51734900e-01 -7.30060995e-01 7.68928170e-01
6.75845802e-01 -7.27165699e-01 -7.92514980e-01 4.99471515e-01] | [9.681890487670898, -1.2945058345794678] |
a8b539fb-f964-49fa-a24b-b4f7e19494a3 | how-large-language-models-are-transforming | 2210.03568 | null | https://arxiv.org/abs/2210.03568v3 | https://arxiv.org/pdf/2210.03568v3.pdf | How Large Language Models are Transforming Machine-Paraphrased Plagiarism | The recent success of large language models for text generation poses a severe threat to academic integrity, as plagiarists can generate realistic paraphrases indistinguishable from original work. However, the role of large autoregressive transformers in generating machine-paraphrased plagiarism and their detection is still developing in the literature. This work explores T5 and GPT-3 for machine-paraphrase generation on scientific articles from arXiv, student theses, and Wikipedia. We evaluate the detection performance of six automated solutions and one commercial plagiarism detection software and perform a human study with 105 participants regarding their detection performance and the quality of generated examples. Our results suggest that large models can rewrite text humans have difficulty identifying as machine-paraphrased (53% mean acc.). Human experts rate the quality of paraphrases generated by GPT-3 as high as original texts (clarity 4.0/5, fluency 4.2/5, coherence 3.8/5). The best-performing detection model (GPT-3) achieves a 66% F1-score in detecting paraphrases. | ['Bela Gipp', 'Frederic Kirstein', 'Terry Ruas', 'Jan Philip Wahle'] | 2022-10-07 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [-1.04429662e-01 3.29471827e-01 8.09875131e-02 2.02998579e-01
-1.27410865e+00 -1.01679516e+00 8.34580839e-01 3.86288851e-01
-1.77236378e-01 6.91338837e-01 5.55082321e-01 -7.49400139e-01
3.71445231e-02 -7.00981855e-01 -8.51204395e-01 8.68235249e-03
5.90704620e-01 3.82968307e-01 -2.72965312e-01 -1.38339937e-01
1.13208115e+00 3.21277648e-01 -8.94578218e-01 4.85532224e-01
1.44838846e+00 -4.92577702e-02 6.95977733e-02 1.15147018e+00
-2.93787241e-01 1.13394380e+00 -1.48810625e+00 -8.55051279e-01
-7.13051409e-02 -6.04895175e-01 -9.87730980e-01 -3.13795805e-01
1.15098548e+00 -2.14308560e-01 -5.15234828e-01 1.14648223e+00
5.83961070e-01 -1.14601620e-01 8.73184681e-01 -1.02906299e+00
-1.29469156e+00 6.49169147e-01 -3.52090180e-01 7.10810781e-01
9.90375817e-01 4.26769823e-01 8.11273336e-01 -1.05173624e+00
7.62954056e-01 1.53479218e+00 6.27471745e-01 3.35361481e-01
-1.37241793e+00 -8.15571249e-01 -6.98358953e-01 5.65427653e-02
-9.94389415e-01 -6.89468145e-01 5.07229924e-01 -7.62233317e-01
1.02958226e+00 3.22479844e-01 4.77471828e-01 1.50704956e+00
6.80183887e-01 3.69549990e-01 1.17787302e+00 -5.05672395e-01
1.26109362e-01 4.56713855e-01 5.44193506e-01 6.04277611e-01
6.60917222e-01 -2.55196720e-01 -6.40591025e-01 -6.27325416e-01
5.80220580e-01 -5.92838168e-01 -2.20748439e-01 4.62828130e-01
-1.09545553e+00 1.00961983e+00 1.54629230e-01 8.14541057e-02
-6.99081942e-02 -2.13644072e-01 3.05033982e-01 7.29742110e-01
4.37836230e-01 1.35152256e+00 3.55111957e-01 -4.58113790e-01
-1.17700326e+00 6.78321123e-01 1.02723002e+00 1.14551723e+00
1.47837967e-01 1.61879733e-01 -5.53371370e-01 8.40126455e-01
-1.19866639e-01 6.97219670e-01 1.00171471e+00 -1.22197080e+00
8.50695908e-01 5.47470868e-01 2.67713577e-01 -1.38093984e+00
1.15028515e-01 -6.54131949e-01 -6.04129374e-01 1.31560378e-02
6.07163310e-01 -6.03876039e-02 -2.40348458e-01 1.13226366e+00
-3.65533113e-01 -3.34727824e-01 4.14975211e-02 4.62257057e-01
8.96801829e-01 7.42091954e-01 6.77294135e-02 1.25178844e-01
1.42931998e+00 -8.09448659e-01 -5.26927888e-01 -5.75771451e-01
8.49196732e-01 -1.47179472e+00 1.34927821e+00 4.37638193e-01
-1.55482733e+00 -6.46568120e-01 -1.04830790e+00 -3.25197637e-01
-3.68488692e-02 3.06367189e-01 6.22823201e-02 8.44826579e-01
-9.73481596e-01 8.08400333e-01 -1.06521904e-01 -2.37028867e-01
5.33776879e-01 -3.87239367e-01 -2.47741371e-01 1.07145116e-01
-1.10945904e+00 1.20179629e+00 -5.55908531e-02 -5.86082816e-01
-6.72161222e-01 -9.76308167e-01 -6.11173868e-01 2.36073673e-01
-2.21337751e-01 -8.49394143e-01 1.41068947e+00 -5.77485204e-01
-1.11184216e+00 1.21134162e+00 -2.64501482e-01 -6.31519556e-01
7.73117721e-01 -4.13631976e-01 -2.23879412e-01 2.60270894e-01
4.79823500e-01 5.99937513e-02 8.48510504e-01 -6.44632936e-01
6.17824961e-04 -2.70073801e-01 -2.75631487e-01 3.88178043e-02
-5.78180730e-01 3.80459040e-01 6.43659115e-01 -8.29533994e-01
-2.92168230e-01 -8.04571807e-01 -1.07939960e-02 -2.47618139e-01
-4.56362277e-01 -3.52817297e-01 4.06570613e-01 -1.17825449e+00
1.31730986e+00 -1.62434483e+00 -2.05017149e-01 -1.31952092e-01
5.53713083e-01 4.68107730e-01 -1.96601868e-01 5.19198954e-01
-9.24553722e-02 7.20391929e-01 2.41418526e-01 -2.30788544e-01
-5.94247580e-02 -5.92349112e-01 -8.65078092e-01 4.34915364e-01
-5.98759216e-04 1.09567332e+00 -1.12967479e+00 -4.41734195e-01
-7.59064406e-02 5.51314652e-02 -3.10173839e-01 2.57673144e-01
1.99843645e-01 -2.87148327e-01 -3.26510608e-01 3.01596522e-01
3.97467852e-01 -4.13402528e-01 -4.14146632e-01 5.44199288e-01
-1.85143128e-01 7.37870097e-01 -6.11991704e-01 1.29738581e+00
-4.74278450e-01 1.36792946e+00 -3.09761405e-01 -3.10998231e-01
1.01511776e+00 2.18145758e-01 -4.54544246e-01 -4.91138756e-01
-2.34240189e-01 2.74859101e-01 -3.88548113e-02 -4.62166429e-01
1.15792525e+00 6.78959563e-02 -1.31830096e-01 9.61065292e-01
-1.02753781e-01 -4.86046702e-01 1.60572842e-01 9.87450421e-01
1.29563618e+00 -4.58908737e-01 9.72521380e-02 -4.28520441e-01
4.08997148e-01 2.12636337e-01 -1.57847360e-01 1.51578295e+00
-1.59736827e-01 7.05368698e-01 8.62840712e-01 -9.23111513e-02
-1.64167917e+00 -9.71811473e-01 2.21385807e-01 7.58619666e-01
-3.46173227e-01 -6.67116225e-01 -9.69902515e-01 -3.73578489e-01
3.55610214e-02 1.41603112e+00 -7.17982054e-02 -6.45224392e-01
-5.02818644e-01 -3.85985732e-01 1.17645693e+00 7.25851953e-02
1.75757632e-01 -1.17321241e+00 -3.43725711e-01 2.00225171e-02
-5.93872130e-01 -8.95639956e-01 -5.88512123e-01 -7.96144664e-01
-8.95851374e-01 -9.10519123e-01 -7.78767943e-01 -6.34285569e-01
5.67686915e-01 6.77014947e-01 1.40949571e+00 6.34636581e-02
-4.24324155e-01 1.49653956e-01 -5.20785302e-02 -3.79502147e-01
-1.38380992e+00 7.32462183e-02 -1.35103434e-01 -8.98398757e-01
4.04371083e-01 -2.33831242e-01 -2.55812794e-01 -4.69617508e-02
-6.01469934e-01 2.16155350e-02 5.85208654e-01 6.43338323e-01
-3.82279843e-01 -5.21040022e-01 6.52009368e-01 -8.56453538e-01
1.63502645e+00 -5.18210649e-01 -2.94153571e-01 2.03328535e-01
-9.11584079e-01 -1.20376252e-01 9.20908928e-01 -4.66215372e-01
-9.97808278e-01 -8.33965063e-01 1.91964373e-01 -3.77607167e-01
-1.21949136e-01 2.96265066e-01 5.42263210e-01 -1.60685450e-01
1.44943690e+00 3.83243918e-01 8.80814791e-02 -3.51403862e-01
2.68376648e-01 8.67867827e-01 6.07960641e-01 -3.97544235e-01
9.01393712e-01 -1.98232323e-01 -3.83671612e-01 -1.01990187e+00
-7.26805627e-01 -2.88024396e-01 -1.03128947e-01 -1.81067437e-01
2.30697796e-01 -1.11666024e+00 -5.36730409e-01 3.78908038e-01
-1.58813894e+00 7.97667056e-02 -2.38072932e-01 2.22954333e-01
-2.59022981e-01 1.07268846e+00 -1.05373514e+00 -4.39910352e-01
-1.00895333e+00 -8.95107508e-01 7.25561500e-01 3.45252216e-01
-1.10011446e+00 -6.85886204e-01 1.36557892e-01 9.70068574e-01
5.08133888e-01 -8.82609412e-02 1.06365049e+00 -9.97956514e-01
-1.92468300e-01 -5.05984724e-01 -3.00063103e-01 1.86610535e-01
-3.45742077e-01 -9.33321938e-02 -7.86746502e-01 -4.23789054e-01
2.65577376e-01 -3.66496354e-01 5.14907658e-01 1.35114402e-01
7.89953053e-01 -7.88797677e-01 3.18401083e-02 3.43010910e-02
8.05101871e-01 -4.14198965e-01 8.31655025e-01 5.03119230e-01
4.93953526e-01 4.84617978e-01 9.87816676e-02 3.61146212e-01
-8.01633596e-02 2.08469421e-01 -1.93142861e-01 5.66171169e-01
-3.62187266e-01 -7.76987910e-01 8.73489261e-01 8.36482644e-01
4.32206482e-01 -4.39789295e-01 -8.72030914e-01 5.05823255e-01
-1.34123921e+00 -1.42000759e+00 -8.88465166e-01 2.35307693e+00
7.89329410e-01 4.78551328e-01 1.34662464e-02 -1.82244048e-01
9.37000334e-01 1.01658896e-01 -2.21740991e-01 -7.96977222e-01
-2.25289211e-01 2.06359684e-01 2.54663944e-01 5.99718034e-01
-4.66201574e-01 9.38735723e-01 6.78564453e+00 9.18066859e-01
-4.60340291e-01 -1.03151925e-01 6.03810787e-01 -2.02979490e-01
-5.17649889e-01 2.51194268e-01 -8.18078279e-01 8.35949421e-01
1.44089937e+00 -1.11046851e+00 2.10066825e-01 9.16451097e-01
5.47358096e-01 1.10989148e-02 -8.42427790e-01 9.44268942e-01
4.77769196e-01 -1.43597186e+00 5.09521902e-01 1.51238322e-01
8.34742546e-01 -4.05720651e-01 9.25674513e-02 5.83857358e-01
5.11147499e-01 -1.18283105e+00 5.80442190e-01 6.37079418e-01
4.81911629e-01 -5.11031449e-01 4.69362944e-01 5.89851856e-01
-1.91971958e-01 -1.18603520e-02 -8.37622583e-01 -2.94505239e-01
-2.93114837e-02 9.45183516e-01 -1.27897024e+00 1.22201540e-01
6.57146797e-02 6.42541647e-01 -1.31253755e+00 1.00492990e+00
-5.74741125e-01 9.47990239e-01 8.66973698e-02 -3.70028347e-01
-1.42781958e-01 -2.10836202e-01 1.13473237e+00 1.47179699e+00
5.49970210e-01 -3.07735950e-01 -3.64293188e-01 1.51019156e+00
-5.84846556e-01 8.95641595e-02 -7.67195225e-01 -3.88824821e-01
7.70668209e-01 1.24334681e+00 -8.93786177e-02 -7.32478261e-01
1.18478671e-01 1.07678711e+00 3.37313920e-01 1.54880628e-01
-5.95454335e-01 -8.36298168e-01 1.60155714e-01 2.59286761e-01
-5.14621794e-01 -8.05075020e-02 -6.79252148e-01 -1.35839009e+00
1.42212838e-01 -1.27011335e+00 2.30917260e-01 -1.36240482e+00
-1.40749645e+00 3.13320965e-01 -3.76532704e-01 -9.53139603e-01
-5.05611300e-01 -1.82688117e-01 -1.03945136e+00 1.32331944e+00
-7.99623430e-01 -6.61345363e-01 -5.02571762e-01 4.46624607e-02
9.40360308e-01 -4.61910546e-01 6.19420052e-01 -1.07363909e-01
-3.26844990e-01 7.71515012e-01 2.05082834e-01 2.25118086e-01
1.14077306e+00 -1.29158998e+00 1.10469913e+00 1.02839899e+00
-1.70883834e-02 1.12843585e+00 9.71525490e-01 -9.57644761e-01
-1.28374207e+00 -9.36915278e-01 1.57514167e+00 -8.55394959e-01
9.15200651e-01 -3.79004739e-02 -1.17989552e+00 5.63661098e-01
3.67380083e-01 -9.57637906e-01 4.29336071e-01 -8.09887797e-02
-5.48634410e-01 6.00874424e-01 -1.00962830e+00 8.47254097e-01
7.17340410e-01 -9.79860723e-01 -1.20735371e+00 8.09078872e-01
4.72106278e-01 -2.05214515e-01 -8.66368413e-01 -3.99284095e-01
5.41250467e-01 -8.58031452e-01 1.08496809e+00 -9.44012702e-01
1.34994519e+00 3.18392217e-01 6.82466328e-01 -1.38805485e+00
-6.30679309e-01 -9.37076449e-01 -1.09304413e-01 1.24493432e+00
2.25046977e-01 -3.80391628e-01 7.70330489e-01 9.42934930e-01
-2.22003460e-01 1.67632818e-01 -6.67768419e-01 -7.39631593e-01
8.23693156e-01 1.36915073e-01 -3.53292637e-02 1.11493635e+00
3.98505092e-01 6.91782951e-01 -1.52236447e-01 -2.81108052e-01
8.07414651e-01 6.33125007e-02 1.00639415e+00 -1.17875099e+00
-1.87600762e-01 -8.85804951e-01 -1.45358115e-01 -9.19563234e-01
3.38211387e-01 -1.23185039e+00 -4.55578983e-01 -1.33948100e+00
4.75982338e-01 3.15798283e-01 5.10729492e-01 1.27734199e-01
-4.39869881e-01 -1.76082119e-01 2.83751637e-01 7.03070462e-01
-2.41500959e-01 3.46049964e-01 9.47013915e-01 -1.50667548e-01
-1.73750371e-01 -4.12272029e-02 -1.02789927e+00 4.77673918e-01
9.33045149e-01 -6.33800447e-01 -2.56524812e-02 -2.29577139e-01
5.52885473e-01 1.55291066e-01 6.89766765e-01 -1.07523072e+00
3.67111683e-01 7.65431672e-02 6.09091640e-01 -1.47956908e-01
-1.47674158e-01 2.29219764e-01 -2.47608055e-03 5.82610428e-01
-1.01294065e+00 3.77880335e-01 2.21235529e-01 3.66535306e-01
2.19874874e-01 -8.82880688e-01 7.52933264e-01 -5.25123477e-01
3.09274271e-02 -6.12227023e-01 -1.00370562e+00 5.26281178e-01
5.46852946e-01 -4.05659787e-02 -1.00817299e+00 -6.41641498e-01
-2.75661826e-01 -8.06911141e-02 6.01769149e-01 5.89197457e-01
6.97292149e-01 -8.77756953e-01 -1.17611265e+00 -2.05999594e-02
-9.62976739e-02 -6.26999319e-01 6.51907846e-02 3.02484393e-01
-6.68395936e-01 5.60288727e-01 -3.16965371e-01 -6.51860535e-02
-1.37426519e+00 3.33383143e-01 8.38084146e-02 -3.06410521e-01
-4.73708510e-01 6.49424791e-01 -2.54832268e-01 -1.29926428e-01
-2.66524285e-01 2.69237965e-01 -4.48387377e-02 -8.97687823e-02
7.79890001e-01 1.09330893e+00 1.05901480e-01 -4.23347414e-01
2.98905998e-01 -6.43479289e-04 -5.72347105e-01 -2.16400743e-01
7.44602025e-01 1.11650929e-01 -2.25242615e-01 2.70045370e-01
1.17811871e+00 3.23464721e-01 -8.71808082e-02 7.30485320e-02
-4.68735397e-02 -6.48561299e-01 -8.87220800e-02 -7.32809603e-01
-6.46790722e-03 8.74997318e-01 -1.40858097e-02 3.21711540e-01
9.94690359e-02 -2.99849927e-01 8.54645193e-01 6.07187569e-01
-2.65963133e-02 -9.88583386e-01 4.12333012e-01 6.71033442e-01
1.14768434e+00 -8.75678658e-01 3.46645117e-01 -1.13474667e-01
-5.63864589e-01 1.25364959e+00 6.15379155e-01 -3.23425174e-01
-2.10519791e-01 -3.66139948e-01 -3.95286292e-01 -2.18971148e-01
-8.69585931e-01 9.96926486e-01 2.69300401e-01 2.37567462e-02
7.38430083e-01 -1.27723798e-01 -6.30457222e-01 3.87136668e-01
-1.02756011e+00 -1.00298539e-01 1.53057754e+00 6.09807432e-01
-6.27188921e-01 -6.08420789e-01 -7.69000649e-01 8.60023856e-01
-6.59803391e-01 -3.14934969e-01 -1.02635098e+00 2.69241989e-01
-9.64576602e-01 1.26447964e+00 -7.07278848e-02 -1.03551060e-01
2.63903946e-01 3.35251689e-01 1.91855863e-01 -6.47629857e-01
-1.18206251e+00 -4.40499604e-01 1.66088924e-01 -1.19929999e-01
5.12220263e-01 -6.08977914e-01 -6.33384883e-01 -1.00499678e+00
-2.14314207e-01 4.54211771e-01 5.03023326e-01 5.17599821e-01
8.78874958e-01 1.15501471e-01 4.70287085e-01 -2.98937708e-01
-1.19100463e+00 -1.33735037e+00 -1.36716202e-01 7.32114434e-01
1.78209394e-01 1.16738901e-01 -7.93929279e-01 1.36691079e-01] | [8.610077857971191, 9.998034477233887] |
090a7d0f-f149-4bdc-beb3-a8e8dbf8ae3b | a-convex-optimal-control-framework-for | 2203.16870 | null | https://arxiv.org/abs/2203.16870v3 | https://arxiv.org/pdf/2203.16870v3.pdf | A Convex Optimal Control Framework for Autonomous Vehicle Intersection Crossing | Cooperative vehicle management emerges as a promising solution to improve road traffic safety and efficiency. This paper addresses the speed planning problem for connected and autonomous vehicles (CAVs) at an unsignalized intersection with consideration of turning maneuvers. The problem is approached by a hierarchical centralized coordination scheme that successively optimizes the crossing order and velocity trajectories of a group of vehicles so as to minimize their total energy consumption and travel time required to pass the intersection. For an accurate estimate of the energy consumption of each CAV, the vehicle modeling framework in this paper captures 1) friction losses that affect longitudinal vehicle dynamics, and 2) the powertrain of each CAV in line with a battery-electric architecture. It is shown that the underlying optimization problem subject to safety constraints for powertrain operation, cornering and collision avoidance, after convexification and relaxation in some aspects can be formulated as two second-order cone programs, which ensures a rapid solution search and a unique global optimum. Simulation case studies are provided showing the tightness of the convex relaxation bounds, the overall effectiveness of the proposed approach, and its advantages over a benchmark solution invoking the widely used first-in-first-out policy. The investigation of Pareto optimal solutions for the two objectives (travel time and energy consumption) highlights the importance of optimizing their trade-off, as small compromises in travel time could produce significant energy savings. | ['Simos A. Evangelou', 'Stelios Timotheou', 'Boli Chen', 'Xiao Pan'] | 2022-03-31 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [-1.63144290e-01 3.79603505e-01 -4.00445461e-01 -5.93377789e-03
-4.34187025e-01 -6.98463857e-01 2.36469910e-01 2.71615952e-01
-6.19935989e-01 8.18384409e-01 -5.57720482e-01 -5.66940784e-01
-1.01635981e+00 -6.92780972e-01 -6.50298834e-01 -1.01016963e+00
-1.36319965e-01 5.27780414e-01 2.64542247e-03 -3.64214659e-01
1.60600305e-01 9.83604968e-01 -1.51617742e+00 -8.36603165e-01
1.21667409e+00 1.05226636e+00 3.87008488e-01 5.05670547e-01
1.81299075e-01 1.86522856e-01 -2.23886415e-01 -3.38733286e-01
1.86325029e-01 1.66815981e-01 -5.68837285e-01 2.87863404e-01
-3.43859673e-01 -2.46093586e-01 1.60073847e-01 8.99633646e-01
3.21383253e-02 6.08343124e-01 7.01844811e-01 -2.05356860e+00
4.89818782e-01 -1.43895030e-01 -4.78940994e-01 2.54897345e-02
-4.20485079e-01 2.52208054e-01 8.19309354e-01 -3.37511301e-01
3.74487966e-01 8.34706485e-01 9.10407081e-02 1.47389919e-01
-1.22767866e+00 -1.82457656e-01 2.17610449e-01 4.12294239e-01
-1.61234713e+00 -2.85119951e-01 5.48390567e-01 -2.77194411e-01
9.95238841e-01 6.14120483e-01 8.97074819e-01 5.01631163e-02
4.85227942e-01 3.78616363e-01 4.32798594e-01 -1.29649103e-01
5.12495339e-01 2.31937736e-01 7.55358785e-02 3.70454818e-01
8.33847582e-01 2.99746543e-02 3.36706966e-01 1.62542284e-01
-1.44258350e-01 -2.28328541e-01 6.44809101e-03 -7.44794190e-01
-5.78434467e-01 7.13213861e-01 1.17398679e-01 4.40195352e-02
-5.63354611e-01 1.25821218e-01 3.10181648e-01 -5.41979261e-02
5.59243895e-02 1.67264625e-01 -3.39505613e-01 -1.23222403e-01
-7.75077403e-01 5.69269955e-01 7.11305380e-01 1.23263264e+00
6.97998822e-01 2.32452720e-01 3.24915978e-03 3.57837349e-01
3.86154056e-01 6.47136331e-01 -5.89388847e-01 -1.20375419e+00
5.50788701e-01 5.96001804e-01 4.71001118e-01 -8.54667246e-01
-6.46134079e-01 -3.71278495e-01 -4.07738119e-01 4.49570686e-01
2.68931121e-01 -3.46902937e-01 -3.13493699e-01 1.45502961e+00
5.82710981e-01 -4.17840540e-01 1.20687418e-01 8.40891421e-01
-1.91733316e-01 7.98812866e-01 2.53519565e-01 -7.64076769e-01
1.18000948e+00 -6.20706856e-01 -9.10225272e-01 -2.14422438e-02
6.11097813e-01 -4.18556482e-01 2.20177755e-01 2.15871423e-01
-1.47251034e+00 -1.55776680e-01 -1.25009203e+00 2.16093779e-01
-4.30290431e-01 1.29101381e-01 -9.48658511e-02 6.80242538e-01
-9.88336086e-01 2.97048390e-01 -7.94071496e-01 -9.06849578e-02
1.58584997e-01 8.03016543e-01 4.69854325e-02 2.34492067e-02
-6.88834786e-01 1.12731731e+00 2.07063183e-01 4.98696476e-01
-6.60237074e-01 -8.75017762e-01 -7.84148037e-01 1.67860746e-01
7.65033364e-01 -4.00427729e-01 1.11710894e+00 -4.39114571e-01
-1.34826350e+00 2.55719423e-01 -1.75218210e-01 -2.54464746e-01
7.15125322e-01 2.13765383e-01 -2.53839344e-01 1.73306003e-01
-2.65665371e-02 2.74804324e-01 2.72148132e-01 -1.29943490e+00
-1.10173571e+00 -1.59072280e-01 1.18730076e-01 3.58215213e-01
-9.48223695e-02 -2.94640034e-01 -4.44332838e-01 2.13973597e-01
-5.76711476e-01 -1.23605144e+00 -5.79845548e-01 -4.19552416e-01
-4.52493638e-01 -5.20729780e-01 1.02766812e+00 -4.45727080e-01
1.14043438e+00 -1.73791552e+00 3.89789760e-01 6.51378512e-01
-2.63467968e-01 1.41620159e-01 1.21183014e-02 6.21974587e-01
2.34561116e-01 1.08686142e-01 -1.24789365e-01 -3.48481208e-01
2.43376121e-02 5.05466342e-01 1.72972068e-01 8.49058807e-01
2.83110887e-01 5.19817173e-01 -7.23609805e-01 -3.83472264e-01
5.89383364e-01 4.78266388e-01 -4.42260355e-01 1.88370571e-01
1.63652040e-02 2.95241531e-02 -7.03277946e-01 3.67439926e-01
8.07640493e-01 6.19555831e-01 2.92657971e-01 -3.22676122e-01
-8.22177827e-01 -4.04431611e-01 -1.28965068e+00 8.16878915e-01
-5.93766689e-01 7.41141498e-01 8.50047708e-01 -1.34483171e+00
6.96823657e-01 1.02071226e-01 9.63634133e-01 -9.22508478e-01
3.95544529e-01 2.90205181e-01 -2.29130223e-01 -5.81982970e-01
8.39150190e-01 6.85665235e-02 -6.23501539e-02 -9.32998657e-02
-2.68052489e-01 -4.45553474e-02 5.57421803e-01 1.49577349e-01
5.03576100e-01 -2.45098561e-01 -6.99496120e-02 -8.56509686e-01
8.45919490e-01 4.02560204e-01 6.10564351e-01 -9.15513933e-02
-9.86172929e-02 -4.28073198e-01 8.24204683e-01 2.45719656e-01
-1.20986962e+00 -6.19673252e-01 -9.20028519e-03 4.64235723e-01
7.67870009e-01 1.38025835e-01 -8.22626650e-01 -1.72607929e-01
7.61354864e-02 1.25685358e+00 -2.82236516e-01 -2.32693329e-01
-8.44433784e-01 -4.68343496e-01 -2.57498860e-01 1.74812600e-01
1.03698067e-01 -2.90257409e-02 -1.11171675e+00 1.68094575e-01
-5.52930683e-02 -1.16232610e+00 -4.17933464e-01 1.99273258e-01
-4.97344762e-01 -1.41766953e+00 -2.32466847e-01 -5.25015235e-01
1.01544130e+00 4.91583884e-01 5.14598668e-01 2.38169596e-01
-4.72998805e-02 6.70139492e-01 3.50858793e-02 -6.02478445e-01
-2.69244254e-01 4.98069786e-02 7.53060877e-02 1.93861172e-01
7.19174277e-03 1.49306804e-02 -5.93453467e-01 6.94537878e-01
-6.17141843e-01 -1.29197016e-01 1.82196304e-01 2.07413241e-01
8.14764082e-01 8.47228944e-01 5.26259184e-01 -9.41561684e-02
4.50599611e-01 -6.04827523e-01 -1.33674622e+00 1.97814852e-01
-8.23714495e-01 -1.46832466e-01 6.48916960e-01 7.47469738e-02
-1.03291476e+00 2.13587001e-01 2.23075554e-01 -2.42117733e-01
7.78784603e-02 8.77976343e-02 -6.53495848e-01 -3.41938853e-01
-4.69419181e-01 5.26688285e-02 4.37574148e-01 -1.51825234e-01
2.02773914e-01 3.12099487e-01 4.61663306e-01 -5.03542781e-01
1.00598264e+00 3.54407668e-01 8.08545411e-01 -9.92280364e-01
-1.44281853e-02 -5.29291749e-01 -5.38319767e-01 -8.02164733e-01
8.16454411e-01 -7.11669683e-01 -1.68806791e+00 4.61153649e-02
-1.00084722e+00 -1.19376443e-01 -1.97576880e-01 4.14197296e-01
-7.89602637e-01 1.79743916e-01 3.27893883e-01 -1.43664420e+00
5.11553735e-02 -1.35241520e+00 5.85407853e-01 4.55470860e-01
1.33497998e-01 -1.04215765e+00 -2.28879020e-01 3.37119997e-01
3.88846606e-01 6.91560209e-01 8.32695305e-01 -7.90994465e-02
-8.16914678e-01 -1.49452120e-01 1.30589500e-01 1.48077145e-01
-3.45897496e-01 2.94715017e-01 -2.13065580e-01 -6.27941370e-01
-2.64449000e-01 3.44198227e-01 1.47564873e-01 4.69418526e-01
7.28140056e-01 -5.69793344e-01 -7.20881164e-01 3.22588652e-01
2.15592623e+00 6.36883259e-01 3.88812244e-01 4.02449131e-01
3.93797904e-01 1.18621397e+00 1.06013882e+00 3.87114495e-01
6.51415825e-01 1.03396106e+00 1.19029343e+00 -1.63127348e-01
5.16191363e-01 2.40597054e-01 4.78056818e-02 4.09049809e-01
-2.90856779e-01 -3.84465575e-01 -6.30482554e-01 9.40181255e-01
-1.75534558e+00 -7.78514802e-01 -6.17642939e-01 2.33737040e+00
-1.99863948e-02 -8.84709284e-02 3.14792186e-01 3.57258409e-01
7.02607334e-01 -3.62779081e-01 -3.49812090e-01 -1.10601234e+00
1.01226695e-01 -3.90417278e-01 1.36162758e+00 6.90084696e-01
-5.60315907e-01 3.97023261e-02 5.75908709e+00 8.81768525e-01
-8.01392794e-01 -7.72779211e-02 5.78763962e-01 -4.17003840e-01
-4.79427904e-01 2.90995948e-02 -8.51468623e-01 4.74427819e-01
1.22076237e+00 -7.11719871e-01 6.23404503e-01 6.07825100e-01
1.06651306e+00 -4.75338697e-01 -7.75075734e-01 3.61611426e-01
-3.95466357e-01 -1.14577341e+00 -4.24601167e-01 5.24039388e-01
6.65922582e-01 -1.85767055e-01 -2.03969955e-01 -3.51380035e-02
-1.17627516e-01 -6.13410652e-01 1.08602262e+00 4.79877025e-01
4.84187186e-01 -1.64982283e+00 5.81481278e-01 3.86636645e-01
-1.46478343e+00 -5.07529974e-01 1.09305479e-01 2.33192772e-01
9.21631932e-01 2.31152505e-01 -4.05925035e-01 9.13701892e-01
2.44839594e-01 6.48241565e-02 3.69721912e-02 1.19706821e+00
2.96894908e-01 5.47696836e-02 -5.45168698e-01 -5.12919247e-01
6.15287423e-01 -8.29043627e-01 7.99302518e-01 9.55335736e-01
2.78437734e-01 1.75773054e-01 7.56405443e-02 7.96047151e-01
4.59542632e-01 1.91071421e-01 -3.84709120e-01 2.23175913e-01
5.00940800e-01 1.45217323e+00 -8.54541361e-01 1.47531286e-01
-2.94546872e-01 1.81065798e-01 -1.59777299e-01 6.26417994e-01
-1.26391423e+00 -6.50852799e-01 1.06269538e+00 3.17316204e-01
3.77276808e-01 -3.21774244e-01 -3.10523123e-01 -1.45013720e-01
2.04094797e-01 1.24164417e-01 5.33213664e-04 -2.20261276e-01
-2.69716352e-01 2.19027013e-01 6.23809397e-01 -1.14521670e+00
-1.89767838e-01 -5.83802164e-01 -8.03648233e-01 7.31818020e-01
-1.81007957e+00 -8.66130054e-01 -3.56777087e-02 3.65028709e-01
4.48754013e-01 2.58200416e-05 -5.87696210e-02 6.59247935e-01
-1.08289886e+00 2.75560200e-01 5.43281972e-01 -7.94271588e-01
-4.39887673e-01 -7.95352340e-01 -4.55625415e-01 9.11470115e-01
-1.09337795e+00 1.13780640e-01 1.15317917e+00 -4.24639761e-01
-2.08003569e+00 -1.19409466e+00 9.03242886e-01 1.77521527e-01
4.88185048e-01 2.25263629e-02 -4.59574938e-01 2.19504237e-01
3.88259500e-01 -4.08671439e-01 2.38433722e-02 -5.65829515e-01
7.93745816e-01 -4.34381574e-01 -1.16430128e+00 5.78518391e-01
4.89161730e-01 1.68801978e-01 2.88743049e-01 1.03478655e-01
4.45032418e-01 -1.03537932e-01 -6.56086266e-01 3.31386834e-01
3.83227140e-01 -3.83510917e-01 6.34347975e-01 -3.47646892e-01
-3.24709952e-01 -4.68263716e-01 -6.00428022e-02 -1.17942131e+00
-8.89835805e-02 -8.27774525e-01 8.89642462e-02 1.45815825e+00
3.66523504e-01 -4.99545455e-01 5.73855639e-01 9.38923061e-01
-4.08464789e-01 -9.70488966e-01 -1.51014268e+00 -9.97444153e-01
-7.57899415e-03 -5.11208296e-01 4.23393935e-01 2.02107802e-01
-7.59657146e-03 -1.12180947e-03 -3.92055027e-02 5.11615157e-01
9.48980451e-01 -3.17310393e-01 6.20565414e-01 -9.81629431e-01
3.84889305e-01 -6.00511074e-01 -1.36316985e-01 -4.92929220e-01
2.71512419e-01 -5.63602507e-01 1.90339461e-01 -1.67877889e+00
-1.87277675e-01 -4.79611725e-01 1.74832597e-01 -6.41690791e-02
3.67934138e-01 -3.37729216e-01 2.50644714e-01 -3.18103552e-01
-4.90365952e-01 5.60322642e-01 9.21775222e-01 -3.18461567e-01
-3.16507578e-01 3.21024090e-01 -4.14970130e-01 4.27159250e-01
7.67396152e-01 -2.88528949e-01 -7.73034036e-01 -2.16427073e-01
1.93165347e-01 3.83704931e-01 2.96036661e-01 -6.64154947e-01
4.97167677e-01 -7.44507849e-01 -6.82931125e-01 -9.65680540e-01
4.81793284e-01 -1.69468462e+00 7.94958353e-01 7.77329683e-01
1.10511199e-01 1.23377554e-01 3.49327147e-01 7.20138550e-01
-3.94717641e-02 -4.09471035e-01 9.94093060e-01 6.19238615e-01
-7.30631113e-01 9.35648456e-02 -8.95779014e-01 -4.18059886e-01
2.04722404e+00 -5.16764224e-01 -1.02377005e-01 -1.28438979e-01
-3.91249388e-01 1.13973963e+00 1.95616096e-01 4.90594298e-01
6.82649165e-02 -1.17929912e+00 -5.45265138e-01 -2.12910503e-01
-2.63035178e-01 -9.27526206e-02 6.18228436e-01 1.16012776e+00
-5.14816821e-01 9.25181627e-01 -2.76818484e-01 -4.91997868e-01
-1.34931540e+00 6.19743109e-01 5.73269784e-01 -9.74416211e-02
-9.63333324e-02 1.62311643e-01 -2.29357988e-01 2.00276613e-01
1.02449432e-01 -1.60423219e-01 -2.74574012e-01 3.10524881e-01
7.80749768e-02 1.41037560e+00 1.22275144e-01 -1.09092665e+00
-5.71244240e-01 8.56007993e-01 6.34252012e-01 7.48836249e-02
1.02945971e+00 -7.33945906e-01 5.02431393e-03 -2.34181672e-01
1.34659612e+00 -1.58797845e-01 -1.32372165e+00 5.52935183e-01
-8.02740902e-02 -1.90138802e-01 4.11717445e-01 -4.00550008e-01
-1.33292043e+00 2.68376619e-01 3.01632047e-01 3.52244437e-01
1.17091727e+00 -3.60031694e-01 6.41008794e-01 3.84395872e-03
3.96448731e-01 -1.58410060e+00 -5.99296033e-01 1.53785050e-01
7.00211406e-01 -6.32566690e-01 -1.53117562e-02 -6.83901787e-01
-4.74561870e-01 1.10449934e+00 6.00861251e-01 5.81984594e-02
5.45642436e-01 2.77758598e-01 -6.27108455e-01 1.41037241e-01
-7.71266222e-01 -2.29259551e-01 1.01852223e-01 3.58703405e-01
-3.00293297e-01 3.24522585e-01 -1.03653932e+00 3.71440262e-01
2.58648902e-01 -2.44018853e-01 7.46590018e-01 7.79593825e-01
-5.82593381e-01 -7.68785477e-01 -4.15343463e-01 1.29312187e-01
-1.12535380e-01 8.68307889e-01 2.74583638e-01 1.17081344e+00
4.14422661e-01 1.32543731e+00 3.71301502e-01 -3.54995281e-02
9.53472197e-01 -3.86200517e-01 4.34746966e-02 1.27915531e-01
-2.08352074e-01 -2.93717664e-02 5.29774427e-01 -5.63530326e-01
-3.58027250e-01 -7.91935980e-01 -1.50259936e+00 -5.79090416e-01
-4.25925732e-01 6.89981997e-01 1.26835120e+00 1.00464964e+00
2.59765208e-01 6.59173489e-01 1.00109792e+00 -7.55111277e-01
-3.04530114e-01 -1.49668589e-01 -7.44710743e-01 -1.85814083e-01
4.00852025e-01 -6.85950875e-01 -3.72968107e-01 -3.24946970e-01] | [5.550437927246094, 1.9085668325424194] |
acc305cb-fa2f-42df-88f1-ea7cd421ebd9 | detecting-edit-failures-in-large-language | 2305.17553 | null | https://arxiv.org/abs/2305.17553v2 | https://arxiv.org/pdf/2305.17553v2.pdf | Detecting Edit Failures In Large Language Models: An Improved Specificity Benchmark | Recent model editing techniques promise to mitigate the problem of memorizing false or outdated associations during LLM training. However, we show that these techniques can introduce large unwanted side effects which are not detected by existing specificity benchmarks. We extend the existing CounterFact benchmark to include a dynamic component and dub our benchmark CounterFact+. Additionally, we extend the metrics used for measuring specificity by a principled KL divergence-based metric. We use this improved benchmark to evaluate recent model editing techniques and find that they suffer from low specificity. Our findings highlight the need for improved specificity benchmarks that identify and prevent unwanted side effects. | ['Fazl Barez', 'Ioannis Konstas', 'Esben Kran', 'Julia Persson', 'Jason Hoelscher-Obermaier'] | 2023-05-27 | null | null | null | null | ['specificity', 'model-editing'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.61272991e-01 2.56620854e-01 -6.59118772e-01 -3.53466302e-01
-8.55334520e-01 -6.03687167e-01 1.00320578e+00 4.72153313e-02
-7.16377020e-01 1.53208971e+00 1.40918583e-01 -4.10201967e-01
-8.39416981e-02 -4.44542170e-01 -1.05087399e+00 -2.86720276e-01
-3.10376465e-01 1.08512379e-01 7.70267025e-02 -3.81230898e-02
6.61483288e-01 3.04982543e-01 -1.43925154e+00 3.71700406e-01
9.29175735e-01 3.97365898e-01 -2.83195496e-01 3.80818903e-01
1.62050366e-01 8.08863223e-01 -9.98271406e-01 -6.93100333e-01
3.27185005e-01 -4.15703446e-01 -7.87955165e-01 -6.25524223e-01
7.39567518e-01 -1.54503018e-01 1.31105110e-01 9.42560852e-01
6.59202814e-01 1.90512225e-01 6.21669531e-01 -1.30351257e+00
-6.67396307e-01 7.58991480e-01 -4.72028434e-01 6.71615779e-01
2.86070436e-01 1.17165640e-01 7.56359458e-01 -8.14194143e-01
6.78211927e-01 1.28875196e+00 1.02871573e+00 9.98994529e-01
-1.52979064e+00 -1.17359877e+00 4.52721000e-01 -4.28162329e-02
-1.11331201e+00 -6.52984977e-01 5.60128450e-01 -2.39459544e-01
1.24409163e+00 6.61400378e-01 3.62992615e-01 1.70001161e+00
4.06645536e-01 5.28062284e-01 1.62854540e+00 -4.76155311e-01
2.12421939e-01 2.19427720e-01 1.80843249e-01 4.18955117e-01
8.07249069e-01 8.13179851e-01 -6.67133868e-01 -3.60558748e-01
5.40503323e-01 -3.54373693e-01 -2.62287110e-01 -3.93696636e-01
-9.42727327e-01 8.36853266e-01 -2.91558690e-02 1.79083273e-01
-1.88491449e-01 2.47165963e-01 4.08936173e-01 5.67221105e-01
6.23563826e-01 1.14995778e+00 -9.34305191e-01 -1.91947252e-01
-1.01903379e+00 5.35162628e-01 8.18028092e-01 6.12166524e-01
2.44373009e-01 6.28275126e-02 -5.09707034e-01 7.35458851e-01
-4.28614095e-02 2.36490905e-01 6.37635767e-01 -1.00515997e+00
5.32880843e-01 2.32058465e-01 2.57189929e-01 -7.70897925e-01
-3.38670850e-01 -6.64621055e-01 -3.77800018e-01 1.54553488e-01
9.48731378e-02 -1.04651190e-01 -8.55503321e-01 2.29690146e+00
6.54843301e-02 3.20267648e-01 -1.46618396e-01 3.79144430e-01
2.72587687e-01 -8.84306580e-02 5.16740620e-01 -6.24201477e-01
5.26652813e-01 -8.10205996e-01 -7.89624393e-01 -3.88632238e-01
1.05010915e+00 -4.50163454e-01 1.27094364e+00 3.55908900e-01
-1.27864420e+00 -4.09978628e-03 -1.17820323e+00 4.64476854e-01
-4.91759688e-01 -5.15164495e-01 1.16986680e+00 1.04630232e+00
-9.04319286e-01 8.73372614e-01 -4.19496268e-01 4.53622490e-02
5.13123572e-01 5.04611909e-01 -1.01545796e-01 1.96075842e-01
-1.54252744e+00 1.46986032e+00 4.03150201e-01 -3.83398801e-01
-8.63959014e-01 -1.35090733e+00 -7.91952610e-01 -1.42164066e-01
4.16191101e-01 -8.46255898e-01 1.26328874e+00 -1.02964151e+00
-1.18229139e+00 6.53010845e-01 4.19598520e-02 -7.62724638e-01
7.00401187e-01 -2.59628862e-01 -6.17984474e-01 -4.01707798e-01
6.48522526e-02 4.86291140e-01 5.52426994e-01 -1.43102634e+00
-4.54029888e-01 -9.49909613e-02 2.39181921e-01 2.00059265e-01
-3.00854474e-01 -1.04604185e-01 6.86280951e-02 -1.12696862e+00
-5.64031482e-01 -7.95605123e-01 -6.91820309e-02 -4.23724830e-01
-2.82208979e-01 1.87947541e-01 5.28393626e-01 -2.79535055e-01
1.53313696e+00 -1.78609002e+00 -3.88946533e-01 1.79153189e-01
1.01347573e-01 5.13520956e-01 -3.44755590e-01 1.69083312e-01
-4.97373641e-01 6.15803003e-01 -3.81494015e-01 -2.58994609e-01
-7.30950311e-02 1.42084464e-01 -4.07639891e-01 8.18342566e-02
2.71344334e-02 9.50368345e-01 -8.88062775e-01 -4.80549395e-01
-3.76889631e-02 1.45840243e-01 -6.95628583e-01 -2.55272269e-01
-1.77690938e-01 6.98013157e-02 -8.84071644e-03 4.21912402e-01
7.48459935e-01 7.34007284e-02 2.65485883e-01 2.03854218e-01
2.88397111e-02 6.99852228e-01 -9.38130736e-01 1.36136711e+00
-4.94696587e-01 3.25488806e-01 -2.97463506e-01 -8.34047794e-01
5.05097032e-01 1.32599533e-01 4.84278351e-02 -8.34924042e-01
-1.74804136e-01 2.24427134e-01 1.99342087e-01 -7.12767690e-02
4.31036741e-01 -5.59498966e-01 2.09615584e-02 3.86321515e-01
-1.12106800e-01 4.84725647e-02 -2.73065157e-02 8.37678909e-02
1.15754080e+00 -1.09646931e-01 4.14267510e-01 -4.77427453e-01
2.12180272e-01 -1.67746559e-01 7.34354794e-01 1.41751444e+00
-2.47825250e-01 3.56026977e-01 5.46278536e-01 -1.58510089e-01
-6.66412890e-01 -1.17954695e+00 -3.31458122e-01 1.01025295e+00
-2.65686691e-01 -2.92277038e-01 -4.83743846e-01 -1.20300734e+00
4.00437117e-01 1.24895906e+00 -9.13607538e-01 -7.48563766e-01
-1.74241379e-01 -1.27067542e+00 9.88338947e-01 5.94796717e-01
2.07048267e-01 -7.47841477e-01 -3.05576563e-01 1.75526794e-02
-6.24332018e-03 -3.81580740e-01 -4.11626875e-01 2.33904004e-01
-1.18335760e+00 -1.18534756e+00 -5.80728829e-01 -3.15539390e-01
4.68168586e-01 2.32833829e-02 1.48300350e+00 2.27418542e-01
1.58529319e-02 3.93910140e-01 1.11049684e-02 -7.55531907e-01
-4.06237215e-01 -1.57817528e-01 4.24506038e-01 -7.39000320e-01
3.78618956e-01 -5.40765345e-01 -5.04425347e-01 1.79608867e-01
-8.18654299e-01 -9.42838714e-02 6.98931158e-01 8.84845376e-01
3.52967590e-01 -3.11327815e-01 1.25319445e+00 -1.47390568e+00
1.06211782e+00 -5.72780430e-01 -3.13691735e-01 4.00545239e-01
-1.42567432e+00 2.26033702e-01 2.88702518e-01 -8.03439856e-01
-1.27409065e+00 -4.54636335e-01 2.13356420e-01 -7.25965500e-02
2.53380954e-01 6.13513529e-01 2.40721386e-02 -3.45647126e-01
9.03280318e-01 -9.68844518e-02 -2.05949917e-01 -5.07954955e-01
1.97901875e-01 2.22242206e-01 1.98066890e-01 -6.73777103e-01
5.05444348e-01 2.02980220e-01 -5.06003164e-02 -1.38579562e-01
-1.08747482e+00 5.61711490e-02 -1.45633042e-01 2.43837386e-01
2.30929013e-02 -9.12958264e-01 -5.42764246e-01 1.98574275e-01
-9.17115867e-01 -5.92770517e-01 -2.08104566e-01 5.33981800e-01
-5.20830631e-01 3.07092726e-01 -5.10683358e-01 -7.18134761e-01
-2.22318798e-01 -5.86063087e-01 4.43975806e-01 -3.64897251e-02
-6.54665470e-01 -1.07310116e+00 4.14665014e-01 7.44063631e-02
5.65374494e-01 2.34420747e-01 9.22954798e-01 -9.18474495e-01
-2.92307362e-02 -1.57250285e-01 1.44700184e-02 3.28613222e-01
1.12822577e-02 -2.00317860e-01 -9.04830813e-01 -2.34288201e-01
6.99898973e-02 -3.72441500e-01 1.33278668e+00 4.88443434e-01
1.24919009e+00 -7.02629209e-01 -4.23467427e-01 4.05438274e-01
1.23311818e+00 2.67217636e-01 6.46307826e-01 5.49110949e-01
4.03655022e-01 3.94572526e-01 7.00081110e-01 1.14981480e-01
2.81243712e-01 6.25893652e-01 5.49665466e-02 1.59143051e-03
-5.10525480e-02 -3.32527846e-01 4.82782364e-01 2.72605747e-01
-1.20442081e-02 -1.43308230e-02 -6.69619679e-01 6.62604451e-01
-1.78624928e+00 -1.10916531e+00 2.11541411e-02 2.68712258e+00
1.47127068e+00 6.99286819e-01 -3.34464312e-02 1.67702466e-01
4.78034735e-01 3.53723429e-02 -4.46056366e-01 -9.23085868e-01
-1.66649729e-01 5.69085181e-01 7.91700423e-01 6.80668533e-01
-8.76931250e-01 8.82979333e-01 8.34169483e+00 7.56745994e-01
-8.14274430e-01 2.11596698e-01 5.37524819e-01 -6.02678359e-01
-6.51813507e-01 1.15042597e-01 -6.80366933e-01 4.31139588e-01
1.12863755e+00 -4.74693805e-01 1.51911080e-01 7.15197682e-01
1.32147789e-01 -3.64984930e-01 -1.42117453e+00 4.05563980e-01
1.67073563e-01 -1.42414761e+00 2.43745461e-01 -6.14594333e-02
1.15972400e+00 -2.30681509e-01 3.78394693e-01 6.80652142e-01
4.01301563e-01 -1.12935865e+00 3.64574164e-01 7.39315391e-01
6.49673581e-01 -9.35338676e-01 5.31991303e-01 1.15368083e-01
-3.81767809e-01 -5.09349070e-02 -1.70811370e-01 -4.79724854e-01
-3.27908188e-01 8.00601244e-01 -7.69340634e-01 1.78805843e-01
2.93815404e-01 3.65906328e-01 -9.10281301e-01 9.92872179e-01
-2.08773494e-01 7.14498103e-01 -3.50734182e-02 1.47520646e-01
-7.93088824e-02 4.27848250e-01 5.73889792e-01 1.27453768e+00
3.77744921e-02 -3.31617773e-01 -4.52737302e-01 1.06858206e+00
-2.77001560e-01 -1.89638674e-01 -9.71138179e-01 -4.59790230e-04
7.53558576e-01 5.30529201e-01 -1.80084839e-01 -5.10405779e-01
-3.41160238e-01 8.17285776e-01 3.13790083e-01 3.31309646e-01
-1.00098634e+00 -2.19819441e-01 8.04408848e-01 -2.09968776e-01
-1.34741262e-01 1.05434678e-01 -6.33178055e-01 -1.15355599e+00
3.70501615e-02 -1.14214325e+00 4.64183927e-01 -2.97383636e-01
-1.20378506e+00 -2.92096764e-01 3.87792319e-01 -7.32102334e-01
-2.68933594e-01 -2.15265751e-01 -6.12958610e-01 7.64196336e-01
-1.34675181e+00 -7.48880267e-01 4.93450671e-01 1.89032122e-01
2.72032022e-01 1.86072171e-01 9.71914232e-01 2.37189248e-01
-6.84230924e-01 1.10816860e+00 1.16665967e-01 -3.97338033e-01
1.08511615e+00 -1.30826414e+00 4.06381428e-01 7.83386946e-01
-2.87612900e-02 1.13422942e+00 7.91105092e-01 -1.00349283e+00
-7.41118908e-01 -1.15910542e+00 1.10325921e+00 -9.96044278e-01
3.56190234e-01 -3.11948031e-01 -7.83360362e-01 1.05303609e+00
-1.98077217e-01 -3.25472295e-01 7.95248032e-01 6.46561503e-01
-6.45126164e-01 1.03298575e-01 -1.38346446e+00 7.82861650e-01
1.34118366e+00 -3.41488123e-01 -9.50816929e-01 1.06305167e-01
7.48193622e-01 -2.72339046e-01 -6.63575113e-01 6.90669298e-01
7.52320170e-01 -1.01194859e+00 1.09566402e+00 -1.02280533e+00
4.02354211e-01 2.63564169e-01 1.32158116e-01 -1.51498449e+00
-2.12183774e-01 -3.75168949e-01 -3.35088104e-01 1.33243001e+00
9.52964425e-01 -8.72200847e-01 6.48746133e-01 1.03899264e+00
1.04587741e-01 -5.93759418e-01 -8.83204758e-01 -1.24652886e+00
4.17419255e-01 -5.28125882e-01 4.88578379e-01 1.43317449e+00
2.39896834e-01 4.99401754e-03 -5.34595311e-01 -2.29849204e-01
7.51706243e-01 -2.82391489e-01 3.76238465e-01 -1.00362289e+00
-2.87118167e-01 -5.35495877e-01 -8.73093531e-02 -3.75963539e-01
2.65989721e-01 -7.66526401e-01 -2.74248540e-01 -8.84826660e-01
6.48311257e-01 -2.28536054e-01 -7.71907151e-01 5.54058909e-01
-4.23048556e-01 2.06488326e-01 3.16888392e-02 -4.09995541e-02
-5.92987418e-01 4.73480403e-01 9.62177515e-01 -9.46025271e-03
-2.14737669e-01 -1.38287455e-01 -1.09516287e+00 7.95591772e-01
1.03299749e+00 -8.32976818e-01 -5.56369603e-01 -2.50197768e-01
4.58587050e-01 -3.08595955e-01 4.99042571e-01 -8.32783222e-01
-1.90503180e-01 -5.78050137e-01 5.17467678e-01 -1.45436719e-01
2.43273582e-02 -4.28405941e-01 5.31659424e-02 6.54532373e-01
-7.72280276e-01 6.76516891e-02 6.34644389e-01 6.46262884e-01
1.20999314e-01 -6.80576414e-02 6.91676497e-01 -1.22630060e-01
-4.78889555e-01 -1.99900210e-01 -3.12801242e-01 3.89914542e-01
1.01527286e+00 -4.50774543e-02 -6.43949091e-01 -2.83722103e-01
-4.45065886e-01 1.55666262e-01 5.42346776e-01 5.24958849e-01
5.38381875e-01 -1.28540432e+00 -3.78603399e-01 -9.27897245e-02
6.47912100e-02 -7.20515788e-01 5.89983612e-02 1.00701141e+00
6.85974658e-02 6.06641829e-01 -1.73660398e-01 2.22605303e-01
-1.28653049e+00 5.67493141e-01 4.63638365e-01 -4.75513279e-01
-1.19354390e-01 9.75254476e-01 1.45067036e-01 -4.41360354e-01
2.57265389e-01 -2.62043506e-01 4.30924585e-03 -5.89245036e-02
5.53779244e-01 4.12431598e-01 1.36844844e-01 -3.04251313e-02
-6.52866483e-01 -6.17812611e-02 -3.72419953e-01 -1.75695941e-01
1.18456340e+00 1.54370265e-02 1.26574710e-01 4.71775621e-01
1.00272787e+00 1.77022323e-01 -8.84646654e-01 1.83093742e-01
3.68107766e-01 -5.76092005e-01 -1.05304115e-01 -1.46108711e+00
-6.45021737e-01 3.46001893e-01 6.04223371e-01 -1.40348092e-01
9.66274023e-01 -4.91613209e-01 3.28293651e-01 4.87548977e-01
3.29714566e-01 -1.28274834e+00 -2.37974137e-01 5.02631605e-01
9.04440999e-01 -1.16477060e+00 2.73890048e-01 -1.10835299e-01
-6.85486972e-01 2.31710956e-01 7.80913770e-01 -3.68268341e-02
4.60299700e-01 3.38310927e-01 -1.34478286e-01 -6.78640902e-02
-1.20963883e+00 1.69636816e-01 2.54839927e-01 8.51471722e-01
6.43545866e-01 1.37131274e-01 -1.18334091e+00 9.08578694e-01
-2.54527748e-01 2.61233777e-01 5.89685380e-01 9.02742207e-01
6.79827482e-02 -1.34438884e+00 -1.54149085e-01 9.01541889e-01
-8.34399641e-01 -4.95030075e-01 -7.99973786e-01 1.11764371e+00
-4.82457764e-02 6.89547241e-01 -2.25541756e-01 -5.15510440e-01
3.86966437e-01 4.86390769e-01 7.74961472e-01 -7.04182506e-01
-7.90135264e-01 -5.60245693e-01 5.30106306e-01 -8.03085327e-01
-5.00564933e-01 -5.35182774e-01 -8.68994653e-01 -3.79496902e-01
-5.42390883e-01 4.61449325e-02 4.57768232e-01 8.19385707e-01
4.68791991e-01 3.55186343e-01 4.02780116e-01 -3.87755364e-01
-8.71153057e-01 -9.52994704e-01 -3.83417994e-01 4.78098840e-01
4.10775840e-01 -9.98687863e-01 -6.21897221e-01 -3.30147296e-01] | [10.365972518920898, 8.17609691619873] |
56213e73-5757-459a-987c-9aa0f07cb57f | modeling-the-distributional-uncertainty-for | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tian_Modeling_the_Distributional_Uncertainty_for_Salient_Object_Detection_Models_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tian_Modeling_the_Distributional_Uncertainty_for_Salient_Object_Detection_Models_CVPR_2023_paper.pdf | Modeling the Distributional Uncertainty for Salient Object Detection Models | Most of the existing salient object detection (SOD) models focus on improving the overall model performance, without explicitly explaining the discrepancy between the training and testing distributions. In this paper, we investigate a particular type of epistemic uncertainty, namely distributional uncertainty, for salient object detection. Specifically, for the first time, we explore the existing class-aware distribution gap exploration techniques, i.e. long-tail learning, single-model uncertainty modeling and test-time strategies, and adapt them to model the distributional uncertainty for our class-agnostic task. We define test sample that is dissimilar to the training dataset as being "out-of-distribution" (OOD) samples. Different from the conventional OOD definition, where OOD samples are those not belonging to the closed-world training categories, OOD samples for SOD are those break the basic priors of saliency, i.e. center prior, color contrast prior, compactness prior and etc., indicating OOD as being "continuous" instead of being discrete for our task. We've carried out extensive experimental results to verify effectiveness of existing distribution gap modeling techniques for SOD, and conclude that both train-time single-model uncertainty estimation techniques and weight-regularization solutions that preventing model activation from drifting too much are promising directions for modeling distributional uncertainty for SOD. | ['Yuchao Dai', 'Mochu Xiang', 'Jing Zhang', 'Xinyu Tian'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['salient-object-detection-1'] | ['computer-vision'] | [-2.70945784e-02 4.34081048e-01 -3.22668582e-01 -5.00526667e-01
-7.34632432e-01 -4.02665257e-01 5.51254034e-01 2.31753334e-01
-3.81837845e-01 7.59642482e-01 4.08667773e-02 -3.24365526e-01
-1.44229516e-01 -4.81842786e-01 -8.57436121e-01 -7.02286720e-01
1.57448173e-01 5.61292648e-01 6.70759916e-01 3.46581548e-01
4.02437240e-01 1.15512297e-01 -1.82708120e+00 1.39227256e-01
1.26213217e+00 1.25665665e+00 3.57470453e-01 2.68949747e-01
-2.34211579e-01 6.00579143e-01 -9.24771130e-01 -4.00987327e-01
3.24907526e-02 -3.88762057e-02 -4.52519268e-01 -1.05622299e-01
1.76814139e-01 -3.44114602e-01 4.64741103e-02 1.59469712e+00
6.06711745e-01 2.51312822e-01 1.11202252e+00 -1.60881639e+00
-1.07921207e+00 8.70949745e-01 -8.80328774e-01 4.06740010e-01
-1.90085039e-01 -8.14403072e-02 6.79717183e-01 -1.06392038e+00
2.88966298e-01 1.42981029e+00 3.69643062e-01 6.55143499e-01
-1.07613516e+00 -6.30985796e-01 7.11410940e-01 3.09685796e-01
-1.40674341e+00 -2.89522205e-03 9.91402805e-01 -6.61080778e-01
5.92995405e-01 2.72806525e-01 4.73587900e-01 1.28160310e+00
2.44041532e-01 1.33426511e+00 1.07484090e+00 -5.00730157e-01
8.16446126e-01 5.53056419e-01 2.98927128e-01 1.24638788e-01
6.27443075e-01 1.98854625e-01 -4.04481679e-01 -1.75901800e-01
6.50361478e-01 -2.90632665e-01 -2.17096612e-01 -7.88165033e-01
-9.63294744e-01 8.52306724e-01 2.57471740e-01 -1.60094962e-01
7.10193738e-02 1.11015439e-01 1.81481272e-01 -1.64827287e-01
8.84983599e-01 2.84393668e-01 -3.97436142e-01 -7.77505012e-03
-7.77253866e-01 4.01458114e-01 4.18534070e-01 1.26154208e+00
7.34375060e-01 1.26254261e-01 -7.73326755e-01 8.20242882e-01
6.32842422e-01 3.75108361e-01 5.90151727e-01 -6.09130025e-01
2.43547320e-01 4.77558523e-01 5.02094269e-01 -5.40642738e-01
-2.84670502e-01 -5.38216650e-01 -3.06263268e-01 3.32014561e-01
4.27767485e-01 -3.70344311e-01 -1.22153842e+00 1.98017466e+00
4.86705035e-01 3.77513021e-01 -3.61452587e-02 9.89959478e-01
9.66061354e-01 3.88409704e-01 3.02959055e-01 -8.32786858e-02
1.31481564e+00 -7.26332903e-01 -6.13845110e-01 -4.98655021e-01
5.11680305e-01 -4.47895229e-01 1.51502597e+00 3.26946050e-01
-8.55062008e-01 -4.55208391e-01 -1.21149123e+00 1.77704975e-01
-4.62961107e-01 4.30153832e-02 4.83554602e-01 5.98651886e-01
-5.55908322e-01 2.31608227e-01 -6.47633255e-01 9.26239863e-02
6.36147380e-01 -5.66004328e-02 2.62111455e-01 1.03161216e-01
-1.27871501e+00 9.34809148e-01 5.54562628e-01 -5.14011607e-02
-1.28900838e+00 -9.88271534e-01 -9.36880350e-01 7.02111349e-02
5.18217385e-01 -4.23513561e-01 1.40627861e+00 -8.29022169e-01
-1.07129335e+00 8.39383185e-01 1.82921253e-02 -4.48567331e-01
5.21697462e-01 -4.99965161e-01 -2.92064726e-01 -3.82108927e-01
3.59411314e-02 8.53024125e-01 1.18772686e+00 -1.63432848e+00
-7.29283750e-01 -1.73447207e-01 -5.44925779e-02 2.37963542e-01
-3.19713831e-01 1.21965222e-01 -2.62714624e-01 -7.11864591e-01
1.32713005e-01 -5.88569164e-01 -4.50575426e-02 3.27017196e-02
-6.81355000e-01 -4.90957141e-01 8.98549080e-01 -1.16116770e-01
1.34353471e+00 -2.27500033e+00 -2.41244465e-01 1.05224177e-01
2.32039511e-01 6.73789158e-02 2.28035256e-01 -3.27720076e-01
3.48172970e-02 2.01181650e-01 -3.50628823e-01 -4.41438854e-01
4.96315569e-01 1.65645808e-01 -5.20530105e-01 3.23052496e-01
4.94113743e-01 6.26894295e-01 -1.06987953e+00 -7.49438763e-01
1.15246214e-01 1.78649783e-01 -2.10055441e-01 3.20250452e-01
-7.03272521e-01 1.35753185e-01 -4.25325871e-01 8.07334304e-01
1.07926226e+00 -1.29715636e-01 -5.60808897e-01 -4.19765040e-02
-2.02393960e-02 3.40897106e-02 -1.48251045e+00 1.17140841e+00
5.94779588e-02 5.37842989e-01 -1.56534404e-01 -5.68723619e-01
9.62223709e-01 -1.52899697e-01 -1.19545065e-01 -2.99237967e-01
3.13778281e-01 1.28205463e-01 8.39630365e-02 -3.11506093e-01
6.74497128e-01 -1.13409333e-01 -1.08675912e-01 1.12466782e-01
7.00322390e-02 -2.36664906e-01 1.19338050e-01 7.22092167e-02
5.29462636e-01 3.29821706e-01 1.25675481e-02 -6.99613094e-01
-1.03679478e-01 -3.52008402e-01 8.54920208e-01 1.04805458e+00
-6.32201552e-01 9.26774740e-01 8.19298327e-01 2.63456166e-01
-6.98571801e-01 -1.55721259e+00 -6.73739076e-01 1.06268334e+00
6.31633759e-01 -3.79257761e-02 -7.92568684e-01 -8.25742900e-01
1.55066729e-01 1.33450878e+00 -8.52045715e-01 -5.59353828e-01
1.02922976e-01 -7.81228483e-01 2.27703482e-01 7.53054440e-01
4.08180445e-01 -9.93569374e-01 -7.17613339e-01 9.74019095e-02
-2.50384659e-02 -6.69390798e-01 -4.53291357e-01 3.91785890e-01
-6.12132847e-01 -9.19250250e-01 -8.54136050e-01 -7.26472020e-01
6.45276725e-01 4.68717925e-02 1.24153018e+00 -3.23348761e-01
-1.58445403e-01 3.54479611e-01 -2.58450747e-01 -1.38391256e+00
-1.24824271e-02 -6.01916373e-01 8.31440985e-02 -1.74984094e-02
6.99414670e-01 -2.53624260e-01 -5.56804419e-01 2.88746864e-01
-8.55601430e-01 -1.03513837e-01 4.08767700e-01 7.20416129e-01
9.10804868e-01 1.39054343e-01 8.22291076e-01 -7.48937368e-01
6.94675684e-01 -7.84307539e-01 -5.32672405e-01 3.65858704e-01
-6.11335993e-01 -2.02967431e-02 -1.32417142e-01 -1.04946601e+00
-1.41728365e+00 -3.81296337e-01 2.83154517e-01 -8.09959531e-01
-1.56711742e-01 5.18482089e-01 -4.75289464e-01 3.64572465e-01
7.63543904e-01 -1.66546032e-02 -2.73272425e-01 -2.62086391e-01
2.10700393e-01 6.25344276e-01 4.38127071e-01 -9.74809825e-01
4.34517831e-01 3.12897623e-01 -4.25077230e-01 -5.62149346e-01
-1.15313768e+00 -3.24134797e-01 -1.60699263e-01 -1.97621316e-01
7.18063593e-01 -9.96021748e-01 -9.76671651e-02 6.19952083e-01
-8.17695498e-01 -4.34257030e-01 -7.63597012e-01 7.90780962e-01
-5.90329587e-01 3.07128131e-01 -2.02297240e-01 -1.33187330e+00
6.68035597e-02 -1.04625702e+00 1.13990891e+00 5.51346660e-01
-6.15110621e-02 -1.02483833e+00 -9.05842707e-02 -1.08622201e-01
2.44515672e-01 2.50174165e-01 8.86412740e-01 -8.69717658e-01
-3.85138988e-01 6.25463575e-03 -3.62179309e-01 4.92964834e-01
-1.21618658e-01 1.49180934e-01 -1.23014808e+00 -6.19674586e-02
2.83327907e-01 -4.34190989e-01 9.73862529e-01 8.19600880e-01
1.49282265e+00 -8.25993065e-03 -4.38445210e-01 2.82683402e-01
1.19732809e+00 1.54654756e-01 6.27750039e-01 3.25647026e-01
4.30383623e-01 6.31709695e-01 1.06286049e+00 7.51623154e-01
3.91574740e-01 2.83253312e-01 8.58684659e-01 6.80797920e-02
-6.20800443e-02 -4.59614098e-01 2.41134539e-01 -6.86442927e-02
5.56930304e-01 -3.72259349e-01 -8.52285326e-01 9.57947552e-01
-2.09456062e+00 -7.05264628e-01 -1.20754972e-01 2.33159256e+00
1.07733941e+00 6.10432148e-01 9.83026251e-03 -2.59860307e-01
1.08050799e+00 -4.85126674e-02 -8.74872208e-01 -2.04457492e-01
-2.45180011e-01 -2.74955660e-01 1.68837398e-01 3.99815947e-01
-1.23636830e+00 7.81655014e-01 5.93455362e+00 1.23822987e+00
-6.46093190e-01 2.94236615e-02 7.00144231e-01 -2.41049677e-01
-8.36453080e-01 1.63275346e-01 -1.08574760e+00 7.97386944e-01
5.09199679e-01 -1.73019260e-01 -2.40409404e-01 1.22530985e+00
-1.22030243e-01 -5.38458228e-01 -1.24934471e+00 9.17794108e-01
-3.63101810e-02 -9.32926416e-01 -2.85970662e-02 -2.28811428e-01
9.19809401e-01 -2.12896362e-01 3.84656489e-01 7.23702490e-01
6.03318214e-01 -8.94533575e-01 1.33046389e+00 6.35199964e-01
3.34939569e-01 -7.84057558e-01 8.82255077e-01 5.75771570e-01
-7.58798420e-01 -1.69713110e-01 -7.75958896e-01 1.42977580e-01
-1.35323424e-02 1.12661707e+00 -4.77076441e-01 8.84370059e-02
1.10350943e+00 3.98933947e-01 -7.86940217e-01 1.44210279e+00
-3.16774368e-01 7.43839204e-01 -4.75622684e-01 -2.25278854e-01
9.11739245e-02 1.55887589e-01 7.45780230e-01 1.12801635e+00
4.84699935e-01 -3.31040204e-01 1.02799334e-01 1.54831386e+00
1.94772169e-01 -3.27599943e-01 -3.99015725e-01 3.38945419e-01
7.70422876e-01 9.17847037e-01 -6.45063519e-01 -3.80076885e-01
-1.97908923e-01 2.22872123e-01 3.01722467e-01 4.12041128e-01
-1.29646254e+00 -3.46440583e-01 4.68501151e-01 -1.35313869e-01
4.53003466e-01 2.53273070e-01 -6.06512964e-01 -1.11133134e+00
6.61324859e-02 -3.37392330e-01 4.48829770e-01 -1.07190132e+00
-1.63296354e+00 1.78911790e-01 7.48202145e-01 -1.33220923e+00
3.22135836e-02 -5.22700250e-01 -8.69921148e-01 8.99123788e-01
-1.78219712e+00 -9.85868037e-01 -6.25714958e-02 5.35121381e-01
5.12087047e-01 2.54117399e-02 2.67533153e-01 -1.49248332e-01
-4.80785340e-01 7.13359237e-01 1.75801307e-01 -2.94876456e-01
7.72319973e-01 -1.58749461e+00 -1.26348078e-01 8.98254752e-01
-9.00563523e-02 5.76242864e-01 1.04674435e+00 -8.08835447e-01
-5.53325951e-01 -1.06390607e+00 4.24365491e-01 -6.59256101e-01
7.29821622e-01 -4.90237951e-01 -1.07200658e+00 6.18516743e-01
-3.18254903e-02 8.39077905e-02 4.75554854e-01 2.32485652e-01
-3.20035756e-01 8.86077732e-02 -1.32791233e+00 7.21626163e-01
8.37044477e-01 -2.67477036e-01 -9.70985174e-01 2.54631937e-01
9.83397305e-01 -5.47235727e-01 -5.21117330e-01 6.18433475e-01
9.14909020e-02 -9.49401438e-01 8.15898597e-01 -3.96917552e-01
5.62032796e-02 -5.85045338e-01 -2.35342443e-01 -1.36445236e+00
-2.86547989e-01 -2.75229663e-01 -7.72510827e-01 1.62179399e+00
3.98734629e-01 -4.67570364e-01 8.04982901e-01 8.03581595e-01
-6.53395891e-01 -7.00159490e-01 -1.14198875e+00 -9.38815415e-01
1.72086284e-01 -6.80131853e-01 7.23538816e-01 5.58816791e-01
-2.56024562e-02 5.32850809e-02 -4.93696593e-02 3.73383373e-01
8.76320541e-01 1.12410858e-01 2.70000130e-01 -1.36221945e+00
-2.99011320e-01 -5.14709532e-01 -2.44437888e-01 -9.57365334e-01
1.61733910e-01 -3.28426838e-01 5.17016232e-01 -1.37076473e+00
4.10913408e-01 -4.95713562e-01 -5.87965071e-01 2.80970097e-01
-4.36451226e-01 -2.58621037e-01 -1.95093319e-01 9.87799093e-03
-6.41950250e-01 8.96382630e-01 1.29785228e+00 -2.94708252e-01
-1.70621917e-01 2.28730455e-01 -1.03245628e+00 9.37997878e-01
6.73264623e-01 -5.22011042e-01 -1.05375230e+00 -2.26540118e-01
1.17377974e-01 -4.38384652e-01 5.87106824e-01 -8.62407327e-01
-1.95663311e-02 -5.52952766e-01 4.44202065e-01 -7.81687438e-01
1.93178892e-01 -8.17672491e-01 -3.53267223e-01 7.63046816e-02
-4.25255686e-01 -6.38736904e-01 4.35942322e-01 8.30165088e-01
-2.66268570e-02 -7.31929541e-01 8.34470212e-01 5.86488023e-02
-9.90242243e-01 1.33326009e-01 -1.76635027e-01 4.97278959e-01
1.22561157e+00 -4.78375286e-01 -7.77070343e-01 -1.63761348e-01
-8.60555530e-01 4.24954951e-01 3.13499808e-01 5.12025058e-01
7.54707098e-01 -1.34893608e+00 -6.80796325e-01 7.82938451e-02
5.15633106e-01 5.67629337e-01 4.11222607e-01 5.74957430e-01
2.40748033e-01 1.90141071e-02 3.86051312e-02 -8.89136553e-01
-8.21708262e-01 7.68022895e-01 3.04476082e-01 1.84264854e-02
-2.62184620e-01 1.25062704e+00 7.41080403e-01 -2.94132143e-01
6.99257314e-01 -6.80593669e-01 -2.52024621e-01 5.19671291e-02
4.96300548e-01 2.02042326e-01 -1.57712400e-01 -7.36331493e-02
-3.22820991e-01 4.43449095e-02 -9.16477218e-02 -7.08552748e-02
8.80085111e-01 -1.65674642e-01 2.07286268e-01 9.56782043e-01
6.80747986e-01 -1.05778337e-01 -1.97202420e+00 -1.33747295e-01
4.22876216e-02 -5.49862385e-01 1.16671324e-01 -9.72325861e-01
-5.16021609e-01 9.84529197e-01 8.80955160e-01 2.63895899e-01
8.01738143e-01 4.84715492e-01 6.86481670e-02 1.18098937e-01
4.01679248e-01 -1.51676250e+00 2.47236595e-01 4.43755597e-01
1.02305031e+00 -1.44784880e+00 -1.25889197e-01 -2.33579636e-01
-9.28703547e-01 5.90065360e-01 1.34376657e+00 -6.57999143e-02
9.03000534e-01 2.99309969e-01 -2.42928356e-01 1.40013276e-02
-7.11819887e-01 -2.51382023e-01 5.19337833e-01 1.09033608e+00
2.68111490e-02 6.80892495e-03 2.57636607e-02 1.15776253e+00
6.03276901e-02 -2.24708080e-01 5.19676507e-01 9.65345562e-01
-8.66882980e-01 -4.49326426e-01 -3.86433095e-01 6.68906569e-01
-1.14138573e-01 -1.09800048e-01 -1.75070494e-01 9.04400647e-01
2.89822727e-01 8.38398099e-01 3.43610823e-01 -1.88472733e-01
2.83799797e-01 8.26931670e-02 4.04624879e-01 -8.00425649e-01
6.57582656e-02 -8.58022273e-03 -5.15178330e-02 3.65050845e-02
-1.59503654e-01 -7.26967037e-01 -1.41902614e+00 1.68784544e-01
-8.30205023e-01 8.73906091e-02 4.19268787e-01 8.10204864e-01
2.77040362e-01 5.68560302e-01 2.49236777e-01 -7.15325177e-01
-1.02440894e+00 -1.22053719e+00 -9.87316608e-01 2.39980772e-01
3.66221040e-01 -1.33084035e+00 -7.85978496e-01 -2.83872664e-01] | [9.632970809936523, 1.931174874305725] |
6fabcd94-576f-40b7-ab88-9c7307b247a4 | coherent-hierarchical-multi-label | 2010.10151 | null | https://arxiv.org/abs/2010.10151v1 | https://arxiv.org/pdf/2010.10151v1.pdf | Coherent Hierarchical Multi-Label Classification Networks | Hierarchical multi-label classification (HMC) is a challenging classification task extending standard multi-label classification problems by imposing a hierarchy constraint on the classes. In this paper, we propose C-HMCNN(h), a novel approach for HMC problems, which, given a network h for the underlying multi-label classification problem, exploits the hierarchy information in order to produce predictions coherent with the constraint and improve performance. We conduct an extensive experimental analysis showing the superior performance of C-HMCNN(h) when compared to state-of-the-art models. | ['Thomas Lukasiewicz', 'Eleonora Giunchiglia'] | 2020-10-20 | null | http://proceedings.neurips.cc/paper/2020/hash/6dd4e10e3296fa63738371ec0d5df818-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/6dd4e10e3296fa63738371ec0d5df818-Paper.pdf | neurips-2020-12 | ['protein-function-prediction'] | ['medical'] | [ 6.78208590e-01 4.45523858e-01 -4.31529075e-01 -5.83683074e-01
-7.15027928e-01 -3.80913377e-01 3.86374652e-01 4.86482948e-01
-1.94112480e-01 6.56841278e-01 -1.66285373e-02 -3.29808056e-01
-3.14655691e-01 -3.32105935e-01 -1.14347816e-01 -7.18125939e-01
1.86497360e-01 7.18274653e-01 3.75930011e-01 6.39790595e-02
8.68797377e-02 3.54349464e-01 -1.72764349e+00 1.05890346e+00
3.15415561e-01 1.10822582e+00 -1.77836627e-01 5.52031994e-01
2.28747874e-01 1.24036753e+00 -1.90028757e-01 -5.02627790e-01
-2.79538687e-02 -5.53961754e-01 -1.45482159e+00 1.20938785e-01
4.94378299e-01 4.26309735e-01 2.40683213e-01 1.02026105e+00
2.75608182e-01 2.17075716e-03 1.09089518e+00 -1.49110544e+00
-1.53365076e-01 7.51407146e-01 -5.97168088e-01 -2.66285390e-01
1.80807412e-01 -6.05289936e-01 1.57928550e+00 -6.44008100e-01
8.57390165e-01 1.48325098e+00 1.00912499e+00 9.10346329e-01
-1.57130539e+00 -8.74501288e-01 2.49070376e-01 4.65601504e-01
-1.50581026e+00 -6.37944490e-02 5.78325570e-01 -4.44301456e-01
8.61568749e-01 3.60682517e-01 1.91154823e-01 1.04449248e+00
1.80562273e-01 8.29613924e-01 1.52092934e+00 -7.38844395e-01
2.39298224e-01 8.09638947e-02 6.07032001e-01 8.30802619e-01
-2.06700921e-01 -1.06165387e-01 -9.40931663e-02 -5.59608579e-01
-3.63956578e-02 -1.09016307e-01 -4.93590310e-02 -3.34306866e-01
-7.20523775e-01 1.12890613e+00 4.91636366e-01 3.33650231e-01
-5.97122014e-02 -5.42831160e-02 5.30719638e-01 1.09148040e-01
7.16695309e-01 3.84770811e-01 -6.68752789e-01 5.14037371e-01
-8.95795226e-01 1.11484803e-01 9.13242042e-01 1.11344707e+00
5.24169087e-01 -6.26263976e-01 -2.42665455e-01 1.14281619e+00
3.48797768e-01 -3.38081330e-01 3.20890725e-01 -1.08286965e+00
2.74120718e-01 7.77086139e-01 -5.91227233e-01 -6.37861371e-01
-1.14606130e+00 -6.29215717e-01 -1.22309566e+00 -4.04429100e-02
9.04447213e-02 1.43366262e-01 -5.29803634e-01 1.64320827e+00
3.98516387e-01 3.39233607e-01 3.66598330e-02 3.01099300e-01
6.28611922e-01 3.55731308e-01 5.69103837e-01 -5.93145072e-01
1.25569963e+00 -1.39677286e+00 -6.69897974e-01 1.45250812e-01
1.17746675e+00 -4.62020338e-01 4.20406789e-01 4.22761053e-01
-4.58757728e-01 -5.50372481e-01 -9.35362279e-01 6.66860193e-02
-5.70689440e-01 -2.60414541e-01 4.94150400e-01 5.95764697e-01
-1.06002939e+00 6.78719580e-01 -8.59093294e-03 -3.19510192e-01
3.61786276e-01 4.43471164e-01 -4.80331957e-01 -4.26934749e-01
-1.23064864e+00 9.96853828e-01 9.95891392e-01 -1.79517150e-01
-6.91738605e-01 -4.37105030e-01 -8.40040088e-01 2.16223542e-02
5.47072351e-01 -3.41776758e-01 1.33211422e+00 -5.21025956e-01
-1.02556086e+00 1.09433341e+00 2.20392615e-01 -3.94834936e-01
3.38229001e-01 3.99490088e-01 -4.77532655e-01 2.65221298e-01
1.69680282e-01 1.05292642e+00 4.48999703e-01 -1.78489733e+00
-1.24538267e+00 -4.12530541e-01 -2.92406953e-03 1.54189784e-02
-2.95352340e-01 2.12059319e-01 -1.03071265e-01 -4.84140545e-01
1.36513934e-01 -1.37272584e+00 -4.90341246e-01 -5.97939253e-01
-8.43462169e-01 -7.68877685e-01 9.29661095e-01 -4.98447008e-02
1.44664824e+00 -1.82130134e+00 5.29823124e-01 4.22016501e-01
5.18515348e-01 -1.19703084e-01 -2.58921683e-01 2.82047272e-01
-2.79811025e-01 3.75418067e-01 -2.44183779e-01 -8.59331608e-01
1.24499775e-01 4.43298757e-01 2.21670702e-01 5.90549171e-01
-1.00387692e-01 6.04230940e-01 -7.06993759e-01 -1.05478096e+00
-1.42976061e-01 1.93823695e-01 -4.37910974e-01 5.00835106e-02
-3.46525520e-01 4.22686428e-01 1.50446063e-02 8.43928814e-01
1.94778249e-01 -8.14634383e-01 9.89334047e-01 -1.31132543e-01
1.45903409e-01 -2.52502769e-01 -1.04417706e+00 1.12918794e+00
-3.38203639e-01 4.68985848e-02 -2.15296060e-01 -9.52331960e-01
5.49950540e-01 7.02929974e-01 6.94466293e-01 -2.94887692e-01
3.19623709e-01 2.81138957e-01 -9.38500986e-02 2.37749144e-02
-4.06254269e-02 -4.12874192e-01 -2.06229806e-01 4.85483527e-01
3.37115735e-01 2.71051824e-01 2.67687827e-01 7.91541412e-02
9.68633592e-01 -2.40534857e-01 8.86582971e-01 -5.11196613e-01
7.43252039e-01 -2.89415210e-01 8.37965846e-01 9.67566431e-01
-2.82518804e-01 3.38217229e-01 6.19966388e-01 -4.78539407e-01
-9.60500658e-01 -9.81924310e-02 -4.37251002e-01 1.67521644e+00
-1.21263370e-01 -6.92022026e-01 -7.83148527e-01 -1.41562819e+00
-5.09303389e-03 4.77158636e-01 -9.00862634e-01 1.16216287e-01
-4.26537097e-01 -8.95937860e-01 6.82339907e-01 3.72322530e-01
3.32645714e-01 -9.20684695e-01 -1.04047671e-01 4.36672598e-01
-3.15801680e-01 -1.44115186e+00 -1.70655683e-01 8.56243849e-01
-6.69397414e-01 -1.40056551e+00 -4.28556919e-01 -1.11869192e+00
4.60188627e-01 1.66754499e-02 1.18719685e+00 4.08351451e-01
-2.12332606e-01 -7.95679614e-02 -6.76332235e-01 -1.08227476e-01
-7.27439821e-01 6.92062497e-01 -9.54992473e-02 2.45493487e-01
9.67177376e-02 -3.58429462e-01 -5.59107587e-02 6.72331810e-01
-8.37046564e-01 3.12678993e-01 4.44572896e-01 9.80635822e-01
7.98161805e-01 6.16546810e-01 6.08613551e-01 -1.81654310e+00
1.94036275e-01 -6.74922049e-01 -3.95854652e-01 7.44975030e-01
-1.16073263e+00 -7.48548433e-02 5.98706305e-01 -5.00987232e-01
-6.81140780e-01 5.40036082e-01 -5.06312586e-02 -2.37651184e-01
-4.31892186e-01 4.57114369e-01 -1.93086296e-01 -3.93190384e-01
4.12604511e-01 -3.43747675e-01 -5.00836790e-01 -7.87634194e-01
4.31720644e-01 8.05940986e-01 2.36393794e-01 -5.09208977e-01
2.59874701e-01 1.51944339e-01 7.46851385e-01 -3.21608126e-01
-1.94805264e+00 -1.05441964e+00 -1.26642823e+00 -4.18724507e-01
1.15076435e+00 -7.04554141e-01 -9.62963760e-01 2.53089905e-01
-1.17765653e+00 -3.12327355e-01 4.42376792e-01 1.88269988e-02
-5.56515336e-01 1.90267012e-01 -1.12480676e+00 -8.42369199e-01
-1.29997328e-01 -1.07328129e+00 1.09155428e+00 -1.80514514e-01
-1.68412700e-01 -1.29739273e+00 2.79449485e-02 7.88977981e-01
-1.13694191e-01 3.71524930e-01 1.61961544e+00 -1.10642672e+00
-4.75635044e-02 -1.01795487e-01 -3.79777879e-01 2.43057057e-01
-3.01399857e-01 -5.20688474e-01 -1.08412647e+00 -5.12586653e-01
-2.55844146e-01 -6.92929149e-01 9.21840906e-01 -1.29241303e-01
1.50963402e+00 -4.16092247e-01 -6.24187887e-01 2.28943571e-01
1.85177147e+00 5.87826148e-02 2.79080868e-01 3.96746933e-01
8.94145906e-01 8.90476644e-01 4.85528916e-01 3.07288140e-01
5.58387280e-01 9.75403428e-01 5.54336846e-01 -2.19582021e-01
-5.61105125e-02 -1.41553478e-02 -3.61904055e-01 1.00435293e+00
7.82469660e-02 -5.51868737e-01 -1.12035978e+00 -6.08714372e-02
-2.26340079e+00 -6.04227245e-01 -5.24545074e-01 1.42360008e+00
9.90922451e-01 2.11557612e-01 1.51368394e-01 4.70664084e-01
1.02498162e+00 6.03965819e-02 -3.19957405e-01 -1.88811332e-01
-2.59294808e-01 1.43329099e-01 4.37466443e-01 5.50660610e-01
-1.59756160e+00 9.10317242e-01 7.23948431e+00 1.18145609e+00
-3.06828082e-01 6.03306293e-01 5.99071681e-01 3.69357407e-01
3.28211129e-01 -3.44557129e-02 -1.50387597e+00 9.86436754e-02
1.22785974e+00 5.11529863e-01 1.01558186e-01 7.46645987e-01
-6.64780676e-01 2.87886351e-01 -1.48246264e+00 6.46471679e-01
3.69503677e-01 -1.11199737e+00 1.57680154e-01 2.11500749e-01
1.03127432e+00 -2.93573916e-01 -1.49242043e-01 6.43127024e-01
5.57896137e-01 -8.41487408e-01 5.86474776e-01 1.61098614e-01
6.51938200e-01 -8.21915090e-01 1.00979209e+00 7.16351449e-01
-1.33512866e+00 -6.27556860e-01 -2.35647440e-01 2.02055469e-01
3.78139652e-02 4.81346905e-01 -5.12547135e-01 7.97020912e-01
3.17953229e-01 8.01213086e-01 -1.04243398e+00 7.32230425e-01
-1.47532314e-01 5.72311938e-01 2.72214532e-01 1.98912799e-01
5.36837578e-01 3.79347324e-01 -1.50112376e-01 1.59327173e+00
-1.99722230e-01 2.50555128e-01 9.42220032e-01 1.48331210e-01
-2.97602296e-01 5.30639529e-01 -5.14750242e-01 4.21476096e-01
1.28935620e-01 1.47593260e+00 -9.93734062e-01 -3.73319358e-01
-6.09587431e-01 9.47387159e-01 8.12129796e-01 -3.60342972e-02
-8.08756709e-01 -9.36927721e-02 -1.66073635e-01 -3.34987938e-01
9.00691077e-02 4.21662867e-01 -1.41931653e-01 -7.66769409e-01
-6.33007705e-01 -8.77014577e-01 1.06327820e+00 -3.65903765e-01
-1.53801322e+00 6.90335989e-01 -1.74351186e-01 -1.25629056e+00
-4.68546413e-02 -7.33759463e-01 2.39566997e-01 3.31547320e-01
-1.71185279e+00 -1.66199505e+00 4.17626984e-02 5.29563665e-01
4.86928612e-01 -2.42803574e-01 1.43924713e+00 5.52146912e-01
-5.99618673e-01 7.54612505e-01 1.90947548e-01 -1.46825567e-01
7.04917192e-01 -1.49355102e+00 -2.61802554e-01 9.34339166e-02
1.31694317e-01 -1.41927563e-02 3.30084592e-01 -5.42001426e-01
-2.38625839e-01 -1.48078096e+00 1.27082956e+00 -5.30098379e-01
6.96071446e-01 -4.96771663e-01 -8.10934126e-01 6.15532339e-01
5.21755777e-02 1.67664602e-01 1.60125995e+00 4.95646000e-01
-9.17558134e-01 2.53100455e-01 -1.07964993e+00 1.75218835e-01
1.06080997e+00 -8.08772743e-01 -1.95533723e-01 7.67267644e-01
9.26299334e-01 2.92272717e-02 -1.48478544e+00 7.16097176e-01
6.49244905e-01 -7.05304682e-01 7.78007269e-01 -1.02172399e+00
4.78682160e-01 6.49063885e-02 -5.79929173e-01 -1.09840941e+00
-8.07907581e-01 -1.08120842e-02 -2.58120984e-01 1.12513828e+00
8.14879894e-01 -2.78917223e-01 5.06991863e-01 2.48786584e-01
-3.96795571e-02 -8.48793030e-01 -8.67209136e-01 -8.33009064e-01
2.95421839e-01 -1.69844508e-01 3.52460891e-01 1.40657377e+00
2.21954137e-01 7.33640015e-01 -7.32369900e-01 2.57429272e-01
1.12183893e+00 3.59551728e-01 4.54721227e-03 -2.04614091e+00
-3.42831105e-01 -3.32208544e-01 -3.03577751e-01 -3.82841468e-01
1.09776926e+00 -1.49899268e+00 8.55868012e-02 -1.15199387e+00
7.09104300e-01 -4.65046376e-01 -8.82027328e-01 9.17852044e-01
-1.88566431e-01 5.80062091e-01 5.64641893e-01 4.92007226e-01
-1.59062791e+00 4.62482087e-02 1.06776834e+00 -2.93336451e-01
3.34876955e-01 1.00662164e-01 -5.70546508e-01 6.98462665e-01
6.09533906e-01 -9.35590208e-01 -4.00765240e-02 3.47993761e-01
1.55162349e-01 3.54550898e-01 -2.41154674e-02 -9.95238543e-01
4.59003389e-01 -2.78790016e-02 2.40729406e-01 -5.19595504e-01
1.06719151e-01 -1.01324117e+00 3.36457759e-01 5.85292637e-01
-1.14566529e+00 -2.03653976e-01 -2.67918527e-01 7.67996967e-01
-1.47723511e-01 -5.62715173e-01 1.11765766e+00 -2.31590122e-01
-4.89003599e-01 2.30603784e-01 -3.04490119e-01 -2.40483537e-01
1.02100086e+00 4.69740123e-01 -5.71280241e-01 1.14403546e-01
-1.30423832e+00 5.08130789e-01 1.35129377e-01 3.73444438e-01
1.82278186e-01 -1.58410203e+00 -5.40673316e-01 -1.48374528e-01
6.48765147e-01 -3.78774166e-01 9.75769311e-02 6.90461099e-01
5.25116511e-02 7.90857553e-01 1.96211925e-03 -3.38984966e-01
-1.70788407e+00 9.23338413e-01 3.08210343e-01 -9.41567481e-01
-4.13349986e-01 5.37441909e-01 5.24735376e-02 -8.18382800e-01
6.65771604e-01 2.52072245e-01 -7.62017667e-01 4.46434081e-01
4.17691499e-01 3.76824498e-01 3.11241467e-02 -7.44912207e-01
-1.94874749e-01 5.26299596e-01 -3.91862392e-01 2.54644662e-01
1.10379517e+00 -1.11219049e-01 -4.29908425e-01 6.87554896e-01
1.54507053e+00 -9.37331498e-01 -6.41004622e-01 -4.82430160e-01
7.50255764e-01 6.65238723e-02 3.43717188e-02 -8.96911740e-01
-8.68149340e-01 4.87725288e-01 5.13110161e-01 5.24949908e-01
8.89458477e-01 2.01477468e-01 3.40355754e-01 5.04468858e-01
7.07265317e-01 -1.24623454e+00 2.46673599e-02 5.75729787e-01
4.54464555e-01 -1.28976083e+00 -9.15510207e-02 -9.83417094e-01
-5.64270020e-01 1.10335064e+00 7.05210805e-01 5.58138072e-01
1.10922575e+00 -4.01935652e-02 -6.69565424e-02 -4.44696844e-01
-1.18021262e+00 -5.40087402e-01 2.96572924e-01 2.45151654e-01
4.12790388e-01 2.87245721e-01 -3.29570323e-01 3.85112971e-01
3.53566468e-01 -2.94164747e-01 1.85504928e-01 8.87906492e-01
-5.15384972e-01 -1.35837877e+00 -1.29503578e-01 3.65388185e-01
-7.28132844e-01 6.07116967e-02 -6.77414000e-01 5.77869654e-01
5.83064735e-01 1.21963894e+00 -5.47102928e-01 -6.38307512e-01
2.71087319e-01 5.33641517e-01 2.26311639e-01 -9.68070745e-01
-7.83066034e-01 2.10049704e-01 3.28129798e-01 -4.74635631e-01
-8.90989959e-01 -5.66399038e-01 -1.10698593e+00 -5.21253496e-02
-7.44202971e-01 1.50924072e-01 4.73974705e-01 1.09464824e+00
-1.71718493e-01 6.20378971e-01 9.25062358e-01 -6.09903634e-01
-4.96084362e-01 -1.06795180e+00 -1.09168994e+00 6.12709761e-01
5.06200306e-02 -8.65697205e-01 -4.96680409e-01 -4.89180200e-02] | [9.60743522644043, 4.357751846313477] |
1e2fd338-b3f0-4ea7-9beb-2f84e9cc2fea | simulating-analogue-film-damage-to-analyse | 2302.10004 | null | https://arxiv.org/abs/2302.10004v1 | https://arxiv.org/pdf/2302.10004v1.pdf | Simulating analogue film damage to analyse and improve artefact restoration on high-resolution scans | Digital scans of analogue photographic film typically contain artefacts such as dust and scratches. Automated removal of these is an important part of preservation and dissemination of photographs of historical and cultural importance. While state-of-the-art deep learning models have shown impressive results in general image inpainting and denoising, film artefact removal is an understudied problem. It has particularly challenging requirements, due to the complex nature of analogue damage, the high resolution of film scans, and potential ambiguities in the restoration. There are no publicly available high-quality datasets of real-world analogue film damage for training and evaluation, making quantitative studies impossible. We address the lack of ground-truth data for evaluation by collecting a dataset of 4K damaged analogue film scans paired with manually-restored versions produced by a human expert, allowing quantitative evaluation of restoration performance. We construct a larger synthetic dataset of damaged images with paired clean versions using a statistical model of artefact shape and occurrence learnt from real, heavily-damaged images. We carefully validate the realism of the simulated damage via a human perceptual study, showing that even expert users find our synthetic damage indistinguishable from real. In addition, we demonstrate that training with our synthetically damaged dataset leads to improved artefact segmentation performance when compared to previously proposed synthetic analogue damage. Finally, we use these datasets to train and analyse the performance of eight state-of-the-art image restoration methods on high-resolution scans. We compare both methods which directly perform the restoration task on scans with artefacts, and methods which require a damage mask to be provided for the inpainting of artefacts. | ['Paul Henderson', 'John Williamson', 'Daniela Ivanova'] | 2023-02-20 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 8.52504790e-01 -7.94609189e-02 5.90994060e-01 -1.69092894e-01
-1.14421296e+00 -4.82090354e-01 6.37583971e-01 -2.85496432e-02
-3.17729264e-01 6.65369093e-01 3.32005829e-01 7.39638731e-02
-1.56193852e-01 -6.97944403e-01 -1.10992622e+00 -6.42319262e-01
3.78785208e-02 3.60974878e-01 2.12500334e-01 -2.51830667e-01
2.40072638e-01 6.94568336e-01 -1.55953419e+00 5.33361077e-01
8.02700281e-01 7.20740855e-01 2.06274033e-01 8.29576135e-01
3.35232556e-01 4.77114975e-01 -1.08737385e+00 -4.78464037e-01
5.19750953e-01 -5.70676327e-01 -7.49373794e-01 5.78190506e-01
9.60122526e-01 -6.03119016e-01 -4.02494729e-01 8.39022100e-01
5.18117428e-01 2.04121228e-02 7.61888683e-01 -6.86016738e-01
-8.93829226e-01 -7.46359453e-02 -7.56400406e-01 2.98079193e-01
5.12334287e-01 5.87512612e-01 5.35105646e-01 -5.13593197e-01
8.73570442e-01 1.01557064e+00 9.35815454e-01 3.60584021e-01
-1.66338253e+00 -3.44871581e-01 -5.81692100e-01 1.92142189e-01
-9.72067893e-01 -4.84285563e-01 8.94290745e-01 -4.99175847e-01
7.62817264e-01 4.36088264e-01 6.93535447e-01 1.33476532e+00
2.18367115e-01 4.17616069e-01 1.46224177e+00 -4.89331901e-01
2.74695069e-01 -1.50088474e-01 -5.66144586e-01 2.89908528e-01
-7.76128471e-02 3.99263412e-01 -2.67166078e-01 -2.69410852e-02
1.00795686e+00 -3.17714542e-01 -7.70664692e-01 -8.18901211e-02
-8.87099981e-01 5.18598437e-01 4.66458350e-01 2.24892795e-01
-6.31313026e-01 1.88322961e-01 2.52369851e-01 4.19316888e-01
6.33250296e-01 6.86365187e-01 3.39087956e-02 -1.78689696e-02
-1.43208349e+00 2.77816534e-01 5.15622914e-01 3.50250572e-01
3.63739014e-01 1.97774008e-01 1.88890785e-01 1.18578970e+00
-2.35954151e-01 2.71187365e-01 1.66712806e-01 -1.17384291e+00
2.55614035e-02 1.06545784e-01 1.95234090e-01 -1.00453675e+00
-3.29695605e-02 -2.73682952e-01 -9.90376055e-01 8.21257114e-01
5.35170972e-01 1.23575486e-01 -1.31529677e+00 1.08638740e+00
1.20175749e-01 1.82199284e-01 -8.64885524e-02 1.19051945e+00
4.39533085e-01 6.13659680e-01 -1.22128904e-01 -2.29769379e-01
9.77618396e-01 -5.37346661e-01 -7.18251407e-01 -2.76955813e-01
-8.06260929e-02 -1.07865751e+00 1.15825510e+00 9.98986542e-01
-1.48679698e+00 -4.21994716e-01 -1.18738651e+00 -1.24516882e-01
1.05046108e-01 -2.70790309e-01 1.87629670e-01 4.79605854e-01
-9.53641474e-01 1.14122188e+00 -5.84086835e-01 -2.02352718e-01
7.68342137e-01 5.01625165e-02 -6.29804432e-01 -4.36615735e-01
-9.76173401e-01 1.15314472e+00 -1.55692294e-01 7.83772171e-02
-1.17127419e+00 -9.90403414e-01 -7.97138274e-01 -2.78421789e-01
1.38107553e-01 -4.63891596e-01 1.19327998e+00 -1.22178602e+00
-1.14687264e+00 1.15372121e+00 3.80983084e-01 -4.53793943e-01
9.88999963e-01 -1.92713395e-01 -4.98873621e-01 5.14405072e-01
-6.61004558e-02 2.65958339e-01 1.17594707e+00 -1.97555399e+00
-1.41800448e-01 3.01639270e-03 -1.50678381e-01 -1.26161292e-01
2.24951640e-01 1.68831125e-01 -2.04439268e-01 -1.02490568e+00
-6.65673092e-02 -3.38307530e-01 -6.38661534e-02 3.12995136e-01
-2.61879265e-01 6.49665594e-01 6.01836622e-01 -1.33778942e+00
7.39976645e-01 -2.07814050e+00 9.30400565e-03 3.22658531e-02
-1.11695953e-01 2.97284693e-01 -2.99846232e-01 5.85762739e-01
-3.68096560e-01 2.96512246e-01 -1.00651503e+00 -5.89178979e-01
-2.71133810e-01 4.05546963e-01 -2.82130510e-01 8.42289090e-01
4.22402531e-01 5.76156020e-01 -7.52976716e-01 -2.53803343e-01
4.31635916e-01 6.82713687e-01 -1.26175314e-01 1.90093994e-01
-2.04991177e-01 4.49248880e-01 4.09803987e-01 7.46481657e-01
7.72249043e-01 3.78050953e-01 -3.14272314e-01 -4.88292843e-01
9.77240428e-02 -1.88861743e-01 -1.07385528e+00 1.73163044e+00
-5.25896311e-01 8.89628351e-01 3.88708413e-01 -5.94477177e-01
6.35197401e-01 3.33837748e-01 2.29396671e-01 -1.10816455e+00
3.31056677e-02 2.05381587e-01 -1.97436824e-01 -9.01002109e-01
5.73702276e-01 -7.96238542e-01 3.70799303e-01 5.27718723e-01
-1.48536265e-01 -8.66604865e-01 -6.98160031e-04 5.38968369e-02
1.43640256e+00 5.17752357e-02 -2.81930447e-01 3.96862030e-02
-9.59179364e-03 5.08435220e-02 1.53599590e-01 5.17494082e-01
1.69357046e-01 1.64720726e+00 2.97129929e-01 -4.46701199e-01
-1.64327729e+00 -1.20194924e+00 -2.83740252e-01 2.26724952e-01
1.99191086e-02 1.72631532e-01 -9.89278853e-01 -3.07972640e-01
-9.91457924e-02 9.11123514e-01 -8.72358561e-01 -8.47305804e-02
-7.67477691e-01 -9.45559382e-01 5.67417085e-01 2.17181027e-01
3.29592377e-01 -1.33155775e+00 -5.80457509e-01 1.95088238e-01
-6.71530887e-02 -9.63781357e-01 -2.64555246e-01 -1.36578128e-01
-5.06009161e-01 -1.37943184e+00 -8.52740288e-01 -5.82978070e-01
6.29669845e-01 9.09513012e-02 1.30976677e+00 3.97652119e-01
-8.96098077e-01 4.13743228e-01 -3.67145211e-01 -1.15668967e-01
-1.02282345e+00 -7.37834454e-01 -3.31492811e-01 7.77915642e-02
-5.56108057e-01 -9.42033589e-01 -8.55912983e-01 3.28659952e-01
-1.70586085e+00 -4.46440168e-02 7.38930643e-01 8.84942532e-01
6.52739167e-01 3.91021878e-01 3.03115368e-01 -6.99703574e-01
8.53132725e-01 -3.00856829e-01 -1.48698762e-01 6.35968009e-03
-1.88759655e-01 -3.22035134e-01 4.72122431e-01 -4.77881372e-01
-1.14814317e+00 -2.39805713e-01 -3.01110417e-01 -5.89680076e-01
-4.31849271e-01 2.22738266e-01 -3.69337685e-02 -2.69414812e-01
1.11842752e+00 1.81308612e-01 1.15114041e-01 -7.08121836e-01
2.49219120e-01 5.43467879e-01 1.10542178e+00 -3.95680040e-01
8.91035259e-01 7.34192729e-01 -5.23833185e-02 -1.11661923e+00
-5.13268769e-01 -7.62092918e-02 -6.35050118e-01 -3.31353068e-01
5.82067847e-01 -5.55732310e-01 -6.81385472e-02 7.86260962e-01
-1.17398524e+00 -7.70334721e-01 -6.55626118e-01 -7.23008290e-02
-7.30924904e-01 8.46062779e-01 -7.80881226e-01 -6.38348877e-01
-1.97672397e-01 -1.01768219e+00 1.13241959e+00 -1.64575249e-01
-2.67289668e-01 -8.00002277e-01 8.44602734e-02 6.30436242e-01
2.84583777e-01 9.55605567e-01 1.01821542e+00 2.29245484e-01
-3.14323306e-01 -3.43933493e-01 -1.74201474e-01 1.00470626e+00
2.58209378e-01 7.83088729e-02 -1.15401769e+00 -1.91739574e-01
3.00404757e-01 -3.10865581e-01 9.23856080e-01 3.58177304e-01
1.02683043e+00 -1.58415124e-01 1.19812414e-01 5.38979530e-01
1.59964323e+00 -6.89628497e-02 1.36812270e+00 4.65903133e-01
5.36397576e-01 7.35352337e-01 4.07142341e-01 -1.09275784e-02
-3.00391614e-01 6.59210861e-01 6.83686733e-01 -5.52105486e-01
-7.44051397e-01 3.14315548e-03 8.35099816e-02 2.95969129e-01
-1.83171958e-01 -4.35437113e-01 -7.98314095e-01 9.07132924e-01
-1.22667468e+00 -9.89756048e-01 -5.97986996e-01 2.29967213e+00
1.09690523e+00 2.24620149e-01 -1.42964674e-02 6.06430948e-01
5.61218619e-01 1.41744077e-01 -2.98049122e-01 -4.76811528e-01
-4.54911560e-01 4.97934490e-01 5.06261647e-01 5.86104929e-01
-9.99356806e-01 4.56104726e-01 6.54072809e+00 9.51767743e-01
-8.64113390e-01 2.59522289e-01 6.14466488e-01 -3.17820400e-01
-2.58726627e-01 -3.10356736e-01 2.95403033e-01 6.61076128e-01
7.96564460e-01 3.78451198e-01 5.77613235e-01 1.77247792e-01
6.03080451e-01 -5.33736110e-01 -9.89851356e-01 9.11664307e-01
3.60377520e-01 -1.42994821e+00 -1.93467457e-02 3.56910639e-02
9.55379009e-01 -1.51173711e-01 -1.02042332e-01 -4.31313843e-01
9.82967764e-02 -1.33828402e+00 9.48092699e-01 8.57372642e-01
9.64277744e-01 -7.15657234e-01 7.00497568e-01 1.67945147e-01
-2.58789450e-01 7.50237629e-02 -2.39860281e-01 6.19085915e-02
6.80650890e-01 8.35355341e-01 -5.24293184e-01 6.13920212e-01
6.77618623e-01 3.97375911e-01 -7.23747671e-01 1.38708186e+00
-4.63375151e-01 6.04866445e-01 -2.53210694e-01 1.02561724e+00
4.66093048e-02 -1.19441085e-01 5.93536258e-01 1.27001727e+00
1.88880473e-01 -2.55193491e-03 -3.46769124e-01 1.03425348e+00
-8.33085850e-02 -3.19790274e-01 -4.56913352e-01 3.04562062e-01
8.99098720e-03 1.03193295e+00 -7.01098800e-01 -1.41336456e-01
-1.42887175e-01 1.40996695e+00 -1.33223027e-01 3.45079243e-01
-5.74441731e-01 -9.49447230e-02 5.51577389e-01 7.38497257e-01
1.63011074e-01 -1.78438380e-01 -4.64569747e-01 -6.15934610e-01
3.33218157e-01 -1.14393246e+00 -3.02462410e-02 -1.28562391e+00
-1.73854303e+00 5.87168276e-01 -8.43398347e-02 -1.13035405e+00
2.54121069e-02 -3.12512577e-01 -8.12484980e-01 1.05295277e+00
-1.35811210e+00 -1.20255995e+00 -4.94597733e-01 2.63697565e-01
7.65335441e-01 3.99897695e-01 6.40511692e-01 3.73969197e-01
-2.54622370e-01 2.85228729e-01 1.78858757e-01 -5.91011457e-02
6.97940588e-01 -1.22626841e+00 7.10397959e-01 1.01784503e+00
3.46734449e-02 -2.37106606e-02 1.01795125e+00 -8.44554424e-01
-1.02101994e+00 -9.92584348e-01 8.28443468e-02 -4.62749362e-01
4.66982514e-01 -1.39708936e-01 -1.46726286e+00 3.83463562e-01
4.22836155e-01 -8.91473144e-02 2.48099312e-01 -5.62406480e-01
-1.29019499e-01 2.46125251e-01 -1.56126797e+00 3.64694506e-01
8.67933929e-01 -4.53785121e-01 -6.89779341e-01 4.30264890e-01
3.52626771e-01 -3.95495027e-01 -1.00799012e+00 3.24629575e-01
5.09687006e-01 -1.33686399e+00 1.18942773e+00 -1.16296962e-01
1.01957321e+00 -2.48030320e-01 9.50284228e-02 -1.66269159e+00
-4.01538908e-02 -5.07370412e-01 2.13544741e-01 1.24788642e+00
7.07422793e-02 -1.83056563e-01 6.15812719e-01 5.11032224e-01
-4.14644122e-01 -6.73535347e-01 -9.37531054e-01 -7.43743062e-01
1.13213405e-01 -5.28073192e-01 2.93168843e-01 9.63014483e-01
-7.41982400e-01 -1.79276660e-01 -4.33881074e-01 2.13336274e-01
7.73696303e-01 -2.37897202e-01 6.27983510e-01 -1.03409231e+00
-4.50030744e-01 -4.12724763e-01 -4.82801527e-01 -2.80654520e-01
-1.82488725e-01 -4.06609595e-01 2.16317967e-01 -1.86148942e+00
-5.31370789e-02 -2.61755466e-01 2.48820633e-01 2.26728097e-01
5.22665912e-03 8.04458380e-01 -8.51933733e-02 3.66280317e-01
1.13915943e-01 4.25888240e-01 1.51695824e+00 -2.17191771e-01
1.13518186e-01 -3.08596283e-01 -3.22052747e-01 8.23188841e-01
6.57710135e-01 -5.31520426e-01 -2.05781177e-01 -7.08151817e-01
1.40820011e-01 -3.48190181e-02 1.03334928e+00 -1.15987337e+00
-2.10704908e-01 8.40970725e-02 6.89522445e-01 -1.15436330e-01
5.87749958e-01 -7.11009026e-01 5.95180690e-01 1.62155569e-01
-1.36102453e-01 -2.51867443e-01 4.42210048e-01 6.05894566e-01
-1.71890184e-01 -4.35983896e-01 1.07165587e+00 -4.10551816e-01
-4.35856640e-01 -1.86118901e-01 -2.81250477e-01 -8.67359713e-02
7.52445459e-01 -5.97702205e-01 -2.58556038e-01 -5.54602802e-01
-7.70616949e-01 -3.61621082e-01 1.22651088e+00 1.64068371e-01
8.15201700e-01 -9.59268153e-01 -8.98116052e-01 2.09644422e-01
-3.68222505e-01 2.57113963e-01 6.17645919e-01 6.75294101e-01
-1.02455437e+00 -6.23174727e-01 -4.39830780e-01 -3.50247055e-01
-1.20218790e+00 4.86627400e-01 4.44995224e-01 -4.24308423e-03
-9.48921084e-01 6.86773896e-01 -4.52129217e-03 8.72547925e-03
-1.76152978e-02 -2.16004923e-01 2.39221677e-01 -8.27112943e-02
5.91794312e-01 4.36800718e-01 5.09131074e-01 -5.97464859e-01
1.75584942e-01 6.54002488e-01 6.13813661e-02 -2.14033902e-01
1.63932312e+00 2.11926363e-02 -8.49203989e-02 4.83542234e-02
9.39806104e-01 -6.82129115e-02 -1.61922264e+00 1.48136795e-01
-2.48541400e-01 -8.91777158e-01 2.85618663e-01 -1.24844694e+00
-1.22759902e+00 8.18499446e-01 7.42660880e-01 3.44219655e-01
1.34908903e+00 -2.19808131e-01 9.05911565e-01 -3.49512011e-01
2.86463559e-01 -8.97347152e-01 1.82861060e-01 -2.47369692e-01
1.60023284e+00 -1.02912772e+00 4.73860234e-01 -4.56952661e-01
-5.05933464e-01 8.18344176e-01 1.54573217e-01 -4.55022901e-01
8.51913542e-02 3.39727461e-01 2.24656016e-01 -4.79350388e-01
-2.17443734e-01 1.78545550e-01 1.52295768e-01 9.41901922e-01
-2.93481182e-02 -2.72959203e-01 -6.49405345e-02 2.67048091e-01
-3.18966746e-01 -1.32080927e-01 9.77206826e-01 1.02664757e+00
-6.21337891e-02 -1.04333699e+00 -8.23672652e-01 4.52171445e-01
-4.98349905e-01 1.26936799e-03 -5.54804325e-01 9.49622154e-01
2.98179507e-01 8.72374952e-01 6.47529513e-02 -1.00471251e-01
6.75168693e-01 -1.89730793e-01 8.92371476e-01 -4.40208793e-01
-7.51330018e-01 -4.05218191e-02 3.09877604e-01 -4.80482310e-01
-3.41983557e-01 -7.66257882e-01 -7.82717168e-01 -1.66492820e-01
-4.71661314e-02 -3.61142069e-01 8.87887061e-01 8.82481754e-01
6.00888431e-02 7.95451581e-01 3.17621887e-01 -1.45873988e+00
-2.91758657e-01 -1.07548237e+00 -7.14951098e-01 1.01194346e+00
5.18846512e-01 -5.24702787e-01 -6.05404973e-01 5.83232343e-01] | [11.279410362243652, -2.029863119125366] |
a07328f3-62d9-44ac-b760-b10a57dcf3e9 | viewformer-view-set-attention-for-multi-view | 2305.00161 | null | https://arxiv.org/abs/2305.00161v1 | https://arxiv.org/pdf/2305.00161v1.pdf | ViewFormer: View Set Attention for Multi-view 3D Shape Understanding | This paper presents ViewFormer, a simple yet effective model for multi-view 3d shape recognition and retrieval. We systematically investigate the existing methods for aggregating multi-view information and propose a novel ``view set" perspective, which minimizes the relation assumption about the views and releases the representation flexibility. We devise an adaptive attention model to capture pairwise and higher-order correlations of the elements in the view set. The learned multi-view correlations are aggregated into an expressive view set descriptor for recognition and retrieval. Experiments show the proposed method unleashes surprising capabilities across different tasks and datasets. For instance, with only 2 attention blocks and 4.8M learnable parameters, ViewFormer reaches 98.8% recognition accuracy on ModelNet40 for the first time, exceeding previous best method by 1.1% . On the challenging RGBD dataset, our method achieves 98.4% recognition accuracy, which is a 4.1% absolute improvement over the strongest baseline. ViewFormer also sets new records in several evaluation dimensions of 3D shape retrieval defined on the SHREC'17 benchmark. | ['Deying Li', 'Xudong Cai', 'Peng Wang', 'Yongcai Wang', 'Hongyu Sun'] | 2023-04-29 | null | null | null | null | ['3d-shape-recognition'] | ['computer-vision'] | [-1.76688299e-01 -1.62444055e-01 -1.12817183e-01 -5.59529841e-01
-1.06126010e+00 -9.28858817e-01 1.02200782e+00 -2.57490724e-01
-1.07145108e-01 -1.49161398e-01 4.40514058e-01 2.51080632e-01
-1.33128330e-01 -2.93528765e-01 -5.57243049e-01 -8.37799549e-01
8.89177620e-02 7.71435618e-01 3.93700413e-02 -1.48630545e-01
2.43873850e-01 8.38609397e-01 -1.70003068e+00 6.99922979e-01
1.43446013e-01 1.49428308e+00 1.25104219e-01 4.88378733e-01
5.03179766e-02 2.79100299e-01 -4.78907257e-01 -4.45815891e-01
4.98898476e-01 3.89285564e-01 -4.81051147e-01 7.26486295e-02
1.08644247e+00 -6.44214630e-01 -3.47683311e-01 5.93044341e-01
1.02011871e+00 -2.12087184e-01 5.45791507e-01 -1.00075531e+00
-1.09989655e+00 3.05838943e-01 -7.64650226e-01 2.65526265e-01
4.95739520e-01 -6.07156195e-03 1.33454406e+00 -1.48181570e+00
7.44046330e-01 1.31932127e+00 3.83634120e-01 5.53797364e-01
-1.04818320e+00 -4.90503401e-01 3.32777679e-01 1.09485403e-01
-1.30706000e+00 -6.43480003e-01 6.87796772e-01 -2.67812699e-01
1.55741978e+00 4.60951954e-01 7.22116470e-01 1.12795460e+00
8.81349295e-02 9.18363810e-01 1.13949859e+00 -6.18066713e-02
-1.28454387e-01 -1.14221618e-01 7.61416331e-02 4.60178494e-01
2.40307629e-01 1.29536748e-01 -8.62514496e-01 -1.33815423e-01
5.72432816e-01 1.73234597e-01 -2.65208155e-01 -9.08508778e-01
-1.30266190e+00 4.26457465e-01 4.67265397e-01 1.32602170e-01
-2.72913545e-01 9.79249366e-03 3.11504513e-01 2.54772127e-01
6.59907043e-01 1.40488714e-01 -4.21505541e-01 3.00391298e-02
-4.79248613e-01 -1.83828326e-03 6.13905847e-01 1.18660879e+00
1.74659699e-01 7.90800806e-03 -2.12297961e-02 1.03674638e+00
3.19628894e-01 1.14743447e+00 1.21420383e-01 -6.68570697e-01
6.21733189e-01 7.15754330e-01 -3.39406356e-02 -1.12315309e+00
-4.04284865e-01 -5.48989832e-01 -8.89735699e-01 8.15254524e-02
2.71383263e-02 6.51302636e-01 -9.84421253e-01 1.42409492e+00
2.80457556e-01 -3.32067817e-01 -9.91429836e-02 1.11360979e+00
1.33835709e+00 3.48835409e-01 -4.75041032e-01 4.13949676e-02
1.31227124e+00 -9.77508008e-01 -2.89809287e-01 1.35601684e-01
2.84188896e-01 -7.45439649e-01 1.00029981e+00 5.75493693e-01
-1.33774149e+00 -5.38140059e-01 -1.09933221e+00 -1.75144523e-01
-3.97558302e-01 1.96084887e-01 6.02177143e-01 5.17482221e-01
-1.19264090e+00 2.84342289e-01 -6.84914887e-01 -3.24953437e-01
3.89824927e-01 3.88306260e-01 -9.01376367e-01 -3.49634230e-01
-4.83748287e-01 6.87166810e-01 -2.39587605e-01 2.59652566e-02
-7.84620762e-01 -8.27473283e-01 -5.67179084e-01 8.21577907e-02
4.20897335e-01 -8.11506093e-01 8.14639807e-01 -2.01146483e-01
-1.23052561e+00 1.39497626e+00 -1.73012912e-01 -7.15456679e-02
2.19296962e-01 -4.29678977e-01 -4.77416158e-01 2.38211796e-01
-1.25128433e-01 5.60708225e-01 8.58680785e-01 -1.54618371e+00
-1.14666045e-01 -8.40906680e-01 2.43602633e-01 4.20296639e-01
-2.39818603e-01 -1.14417262e-01 -9.82623756e-01 -3.72289181e-01
6.92458332e-01 -1.11546206e+00 2.72280723e-01 -1.29839266e-02
-2.21735582e-01 -2.39521578e-01 5.62454164e-01 -2.16252536e-01
8.16080332e-01 -2.15113735e+00 5.12525678e-01 1.74547121e-01
5.26083767e-01 9.54116359e-02 -2.73021191e-01 5.50202310e-01
-1.09767675e-01 -3.06884516e-02 1.54857844e-01 -6.20403588e-01
2.72388071e-01 2.36400396e-01 -3.75591844e-01 5.15702426e-01
-7.81999305e-02 1.07287836e+00 -4.74673986e-01 3.78391109e-02
1.11756742e-01 5.75938761e-01 -7.47206450e-01 2.75168896e-01
1.70791924e-01 2.45651230e-01 -2.51280814e-01 9.09666836e-01
9.05021548e-01 -5.89904666e-01 7.32014030e-02 -5.97176313e-01
1.07620381e-01 2.47993842e-01 -1.03008008e+00 2.07361841e+00
-4.32965010e-01 2.78066099e-01 -1.09995052e-01 -5.87529480e-01
9.90508854e-01 3.03832311e-02 8.00422609e-01 -8.20501387e-01
-8.31432417e-02 8.22384655e-02 -2.55007803e-01 -3.66203725e-01
3.82456779e-01 5.79668134e-02 -9.62889120e-02 5.67835391e-01
2.72446394e-01 -2.24046752e-01 -3.91093791e-01 1.46011576e-01
9.27338541e-01 1.90943599e-01 2.15478554e-01 -2.71774322e-01
4.99653161e-01 -6.56406224e-01 1.43567830e-01 7.47705400e-01
-6.53050467e-02 1.01454461e+00 2.54195899e-01 -8.29804838e-01
-9.33959484e-01 -1.13378000e+00 -5.42146154e-02 1.16544998e+00
1.42057210e-01 -7.13072598e-01 -9.05500799e-02 -8.21814477e-01
3.81646603e-01 3.34300756e-01 -6.83106244e-01 -6.74817115e-02
-3.97464722e-01 -5.16659260e-01 4.01738495e-01 7.36071348e-01
4.28481489e-01 -5.04237890e-01 -5.96826792e-01 -3.06966364e-01
7.64933601e-02 -1.36692476e+00 -6.72421396e-01 -3.66958082e-02
-7.28314161e-01 -1.00074148e+00 -7.16341436e-01 -3.16375464e-01
4.79751170e-01 7.74462521e-01 1.41701984e+00 -5.83557449e-02
-7.33260214e-02 9.97765541e-01 -3.26657891e-01 -2.31990531e-01
1.35144755e-01 7.28304833e-02 2.33259514e-01 8.57530460e-02
3.06739390e-01 -7.38708198e-01 -7.68923461e-01 3.61871332e-01
-6.45084977e-01 -4.85324301e-02 6.67957842e-01 9.81805861e-01
9.40028012e-01 -9.51795697e-01 2.79064715e-01 -7.22294748e-01
1.16222665e-01 -1.56641454e-01 -5.61914980e-01 4.47123975e-01
-7.13384986e-01 1.33036554e-01 2.62727171e-01 -2.03618839e-01
-7.37905383e-01 2.03427896e-02 -6.84283003e-02 -9.83784556e-01
-1.66365713e-01 1.44680098e-01 -5.25601506e-01 -2.44116560e-01
3.17927748e-01 4.74579901e-01 -4.03383225e-02 -7.73581207e-01
5.69562256e-01 4.37611908e-01 2.13715941e-01 -4.20705557e-01
5.54603279e-01 8.43864620e-01 8.87630582e-02 -7.69375086e-01
-8.27761710e-01 -5.63621283e-01 -8.63244355e-01 -1.26348749e-01
6.58339262e-01 -1.14365637e+00 -1.08421564e+00 4.05604720e-01
-1.09508729e+00 2.40842223e-01 -7.57204443e-02 2.40620837e-01
-6.80773318e-01 3.58508557e-01 -2.56121963e-01 -5.96666873e-01
-5.50878108e-01 -1.18552339e+00 1.73660290e+00 -3.78784806e-01
1.56417593e-01 -5.23494303e-01 -5.00729345e-02 5.01004338e-01
4.36205864e-01 1.93219110e-01 8.20881367e-01 -8.58799398e-01
-7.59881914e-01 -1.71429083e-01 -3.58760893e-01 1.53675005e-01
4.00198027e-02 -3.42365533e-01 -1.32908773e+00 -6.38726652e-01
1.56057119e-01 -3.60569924e-01 1.11862457e+00 -4.69877943e-02
1.26445043e+00 -1.46731120e-02 -2.41789669e-01 9.16436315e-01
1.32915378e+00 2.83432424e-01 3.42454821e-01 -4.70345430e-02
9.26333308e-01 2.78292984e-01 1.51396215e-01 5.03260434e-01
6.21970594e-01 9.97564077e-01 6.50039494e-01 1.59187838e-01
-2.30888322e-01 -1.43840864e-01 1.35068297e-01 1.35459244e+00
-3.08272809e-01 -3.19390833e-01 -9.91076231e-01 2.22981900e-01
-1.45319462e+00 -1.02400792e+00 1.99492723e-01 2.10461187e+00
4.11367193e-02 6.10411689e-02 2.50147022e-02 -2.49420926e-01
1.49700254e-01 5.88555098e-01 -7.72285819e-01 -2.74658114e-01
-4.27626431e-01 1.42471358e-01 3.06668013e-01 2.64288127e-01
-1.12077856e+00 6.83674395e-01 6.25839567e+00 4.83679920e-01
-1.06646609e+00 -2.23471364e-03 3.22422534e-01 -5.85029483e-01
-5.63946366e-01 -4.04485166e-01 -9.19016302e-01 -9.77894887e-02
5.61076105e-01 1.43241584e-01 3.06883961e-01 7.02280819e-01
-4.50463951e-01 3.31090569e-01 -1.32046747e+00 1.51914859e+00
8.22830677e-01 -1.12615943e+00 5.54930806e-01 3.35756212e-01
6.52251363e-01 3.70496660e-01 3.88893783e-01 1.65288765e-02
-2.29123205e-01 -8.80063891e-01 6.36508167e-01 7.35591292e-01
1.02110720e+00 -5.98391712e-01 4.25866932e-01 1.91559359e-01
-1.19597590e+00 -4.33008783e-02 -3.08941513e-01 4.10465181e-01
-8.12356472e-02 2.33225554e-01 -5.03070831e-01 8.98010552e-01
8.40200782e-01 1.14938676e+00 -7.10686564e-01 7.45913684e-01
3.16549093e-01 -6.04628325e-02 -4.66789067e-01 8.00208896e-02
1.41404152e-01 3.92415412e-02 8.72400045e-01 8.08807909e-01
4.43384260e-01 2.91536838e-01 -3.41639407e-02 5.56849539e-01
-6.95650801e-02 -7.55012929e-02 -1.10131538e+00 1.53079942e-01
2.69652963e-01 1.10386693e+00 -4.27517921e-01 -2.90843338e-01
-5.08837223e-01 1.10250759e+00 4.78199095e-01 2.07342058e-01
-6.89056814e-01 3.18812162e-01 6.38846159e-01 -8.04755464e-02
7.55865097e-01 -3.51125181e-01 -2.32956484e-01 -1.45559514e+00
3.46250862e-01 -1.01692057e+00 4.53146756e-01 -9.44974720e-01
-1.48953485e+00 9.33146000e-01 1.08615130e-01 -1.42553389e+00
-5.02377413e-02 -8.53173375e-01 1.06059819e-01 6.84000552e-01
-1.21642208e+00 -1.53774691e+00 -3.98977906e-01 3.90765667e-01
4.81706023e-01 -4.85608637e-01 1.09090722e+00 3.66198301e-01
-1.11445844e-01 7.98768044e-01 8.55114684e-02 -3.25368464e-01
8.30047548e-01 -1.26583242e+00 6.35904789e-01 2.11434141e-01
6.32819116e-01 6.41624331e-01 1.82093516e-01 -2.33049303e-01
-2.16887069e+00 -4.64935035e-01 8.42566311e-01 -1.05575359e+00
3.03638816e-01 -7.35309958e-01 -8.56794953e-01 5.42821288e-01
2.90580124e-01 9.51939300e-02 8.65962803e-01 1.72353238e-01
-1.04699421e+00 -3.74099553e-01 -8.38899374e-01 2.72083879e-01
1.62008762e+00 -7.81566978e-01 -5.03633797e-01 1.37844339e-01
6.31782472e-01 -7.57289648e-01 -1.10912919e+00 7.65754163e-01
1.04548597e+00 -1.33515394e+00 1.44198799e+00 -8.27952087e-01
3.37220848e-01 7.89093785e-03 -8.67121816e-01 -1.12527275e+00
-5.21996796e-01 -1.66354463e-01 -3.31732988e-01 9.96734560e-01
2.49124318e-01 -6.21938288e-01 5.82076550e-01 5.09881496e-01
-1.78466663e-01 -1.30005074e+00 -1.16127861e+00 -7.85332561e-01
8.13058391e-02 -4.12667453e-01 8.61864865e-01 7.71730661e-01
-5.61896682e-01 4.16668296e-01 -4.07501072e-01 2.77204514e-01
7.24624276e-01 8.28324378e-01 8.45344484e-01 -1.17535937e+00
-4.27506328e-01 -4.74835873e-01 -4.96063918e-01 -1.47055125e+00
6.58316165e-02 -1.16165721e+00 -5.46778440e-01 -1.39690101e+00
4.99453485e-01 -3.09778184e-01 -4.74430621e-01 3.25370431e-01
2.80891597e-01 2.81305879e-01 6.09953165e-01 3.08884323e-01
-8.10981095e-01 6.84716702e-01 1.46877182e+00 -2.73587108e-01
1.82607725e-01 -5.97818121e-02 -7.41999269e-01 5.42632759e-01
4.42521363e-01 -2.97763548e-03 -3.88258606e-01 -8.60032022e-01
3.18428963e-01 1.13524340e-01 5.10024965e-01 -7.38520861e-01
3.34711194e-01 1.06492780e-01 5.23535311e-01 -1.11252129e+00
9.06067729e-01 -9.54367876e-01 2.80713290e-01 4.14692275e-02
-2.08438098e-01 5.26132762e-01 2.90113449e-01 8.35848987e-01
-6.52128309e-02 4.42769051e-01 4.21101332e-01 -1.05391458e-01
-6.12440169e-01 5.03324389e-01 3.67512971e-01 4.45984192e-02
5.84366024e-01 -1.71249256e-01 -4.67904717e-01 -3.86267722e-01
-8.97254407e-01 9.02452320e-02 3.85843813e-01 8.57609153e-01
9.34846103e-01 -1.79447615e+00 -5.60303926e-01 6.20491385e-01
5.27087331e-01 -3.07109714e-01 4.89230990e-01 6.43979490e-01
3.92518193e-02 6.06978476e-01 -2.13614762e-01 -1.00078046e+00
-1.42028391e+00 5.89527965e-01 2.49565929e-01 -2.25236416e-01
-7.71497488e-01 9.57458496e-01 3.34642410e-01 -3.23401749e-01
1.22319706e-01 -2.73679495e-01 -1.43399149e-01 2.81266749e-01
3.23016822e-01 1.43236727e-01 3.54811370e-01 -9.69193816e-01
-6.61477983e-01 1.08423007e+00 -2.53319204e-01 2.99490262e-02
1.67309153e+00 -2.06829697e-01 -2.35726647e-02 6.92244172e-01
1.46724331e+00 -2.15724409e-02 -1.00759280e+00 -2.84225702e-01
-1.77048713e-01 -7.02972472e-01 -1.33529812e-01 -9.32159424e-01
-1.20289385e+00 1.09421110e+00 7.56901681e-01 1.80105597e-01
1.07719314e+00 3.00792843e-01 6.64155304e-01 5.78685403e-01
6.09868944e-01 -6.26349747e-01 2.13401243e-01 7.17528999e-01
1.54118574e+00 -1.35433507e+00 1.59565821e-01 -2.23390982e-01
-6.06665969e-01 7.86439240e-01 6.92577124e-01 -1.28965631e-01
7.40303278e-01 1.12509497e-01 -6.60798699e-02 -6.50955319e-01
-1.11669898e+00 -1.02031454e-01 9.46711481e-01 3.83705497e-01
2.57093847e-01 -1.10833459e-02 1.22681491e-01 6.62122309e-01
-4.43426035e-02 -7.04996467e-01 -7.84140378e-02 4.61968601e-01
-1.66169599e-01 -9.33476090e-01 -1.15647420e-01 4.09121007e-01
-3.09102386e-01 -2.24910979e-03 -8.43954802e-01 7.52021492e-01
-2.09787250e-01 3.82351428e-01 6.31840453e-02 -5.25466740e-01
8.07306468e-01 2.46473774e-01 7.32112169e-01 -3.58608335e-01
-6.38406992e-01 2.96692163e-01 -7.58949593e-02 -9.81078506e-01
-6.26739085e-01 -5.85434437e-01 -6.08070731e-01 -8.42512622e-02
-2.49851316e-01 -5.13849437e-01 4.64141488e-01 6.35362864e-01
1.02665639e+00 1.28472775e-01 7.42818356e-01 -9.00405943e-01
-7.76828825e-01 -5.95375180e-01 -4.33782995e-01 6.54196978e-01
2.48939171e-01 -9.50783968e-01 -2.65096575e-01 -3.19015235e-01] | [8.230355262756348, -3.703024387359619] |
22ded831-7616-4091-a9ec-69b359158ed9 | yes-but-can-chatgpt-identify-entities-in | 2303.17322 | null | https://arxiv.org/abs/2303.17322v1 | https://arxiv.org/pdf/2303.17322v1.pdf | Yes but.. Can ChatGPT Identify Entities in Historical Documents? | Large language models (LLMs) have been leveraged for several years now, obtaining state-of-the-art performance in recognizing entities from modern documents. For the last few months, the conversational agent ChatGPT has "prompted" a lot of interest in the scientific community and public due to its capacity of generating plausible-sounding answers. In this paper, we explore this ability by probing it in the named entity recognition and classification (NERC) task in primary sources (e.g., historical newspapers and classical commentaries) in a zero-shot manner and by comparing it with state-of-the-art LM-based systems. Our findings indicate several shortcomings in identifying entities in historical text that range from the consistency of entity annotation guidelines, entity complexity, and code-switching, to the specificity of prompting. Moreover, as expected, the inaccessibility of historical archives to the public (and thus on the Internet) also impacts its performance. | ['Antoine Doucet', 'Jose G. Moreno', 'Ahmed Hamdi', 'Nancy Girdhar', 'Emanuela Boros', 'Carlos-Emiliano González-Gallardo'] | 2023-03-30 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [-2.70913333e-01 3.48749459e-01 -1.14997938e-01 -1.01473115e-01
-1.22673678e+00 -9.29119945e-01 9.01781380e-01 4.84216243e-01
-7.03228831e-01 1.00519180e+00 7.90258348e-01 -5.36488712e-01
1.04690224e-01 -4.51116264e-01 -3.64560395e-01 -1.04069427e-01
1.50525570e-01 5.51604390e-01 2.28594482e-01 -3.10933053e-01
5.62058806e-01 3.52348626e-01 -9.14304197e-01 4.81935203e-01
9.00117517e-01 3.50163221e-01 2.15211213e-01 5.58646262e-01
-5.89811444e-01 1.22617733e+00 -7.41979063e-01 -8.60133827e-01
-3.41027379e-01 -3.26598704e-01 -1.29844844e+00 -4.04390097e-01
4.64957729e-02 3.77668999e-03 -3.26828539e-01 6.57672346e-01
4.96534884e-01 1.13611631e-01 4.97448772e-01 -9.43004906e-01
-6.73038125e-01 7.60103762e-01 -8.82289112e-02 3.35933983e-01
6.03160381e-01 -7.79858604e-02 1.24256074e+00 -8.52417588e-01
1.25741649e+00 9.41282809e-01 7.42404938e-01 5.17370164e-01
-8.52032185e-01 -3.78421515e-01 -5.79508245e-02 1.86720312e-01
-1.42514944e+00 -7.40346551e-01 4.04765010e-01 -6.20149910e-01
1.14798462e+00 3.39305311e-01 1.07387066e-01 1.16263187e+00
1.04381062e-01 5.78880966e-01 8.28803420e-01 -7.42282689e-01
2.44656205e-01 5.65219522e-01 -6.62980974e-02 4.05388862e-01
9.31932032e-02 -5.77537358e-01 -5.80974758e-01 -4.66990292e-01
2.16554180e-01 -6.58668816e-01 -3.64394158e-01 3.29681247e-01
-1.10884523e+00 7.71461427e-01 -1.19809240e-01 7.57577658e-01
-2.53860474e-01 -2.44145006e-01 6.43871248e-01 2.71035098e-02
6.03504896e-01 8.48548591e-01 -5.64424336e-01 -5.83234608e-01
-8.45536351e-01 7.06078671e-03 1.50157773e+00 7.98940957e-01
3.91213447e-01 -3.38884175e-01 -2.34342343e-03 7.80467927e-01
1.05556354e-01 7.81622902e-02 6.99524224e-01 -6.81491792e-01
8.58890295e-01 5.10766625e-01 3.12479407e-01 -1.00932407e+00
-3.71775448e-01 -3.40000749e-01 -3.67095709e-01 -5.06733477e-01
6.03732169e-01 -4.99449372e-01 -2.75241345e-01 1.70229661e+00
1.94797382e-01 -1.68712288e-01 1.65167600e-01 5.30295551e-01
7.14382946e-01 7.20118284e-01 3.48798513e-01 -7.22947717e-02
1.36746657e+00 -6.31485224e-01 -5.09544492e-01 -3.32735628e-01
9.46010113e-01 -9.19391692e-01 7.73779154e-01 2.19118968e-02
-8.19018424e-01 -1.44576833e-01 -6.33075237e-01 -1.17868364e-01
-5.47334671e-01 2.43506417e-01 4.33196217e-01 4.72062916e-01
-7.20649004e-01 4.30307478e-01 -8.07318628e-01 -8.81698966e-01
-1.34259030e-01 -1.15355775e-01 -5.40308475e-01 1.84675336e-01
-1.32574701e+00 1.35667038e+00 2.95060307e-01 -4.25619334e-02
-1.84305519e-01 -5.61446011e-01 -5.61870039e-01 1.21761143e-01
4.81841862e-01 -2.76796311e-01 1.20353222e+00 -5.82691729e-01
-1.13803148e+00 8.82428586e-01 -2.16236383e-01 -3.58372688e-01
3.69244546e-01 -2.53504783e-01 -8.03090572e-01 1.25480816e-01
2.63083816e-01 2.29119003e-01 5.64208515e-02 -8.23884130e-01
-5.94587266e-01 -2.37855613e-01 1.50082946e-01 -1.00880384e-03
-3.63205880e-01 7.19835818e-01 -4.18206334e-01 -5.61694384e-01
-1.44523397e-01 -1.08620799e+00 -1.33572802e-01 -5.32381415e-01
-4.01398838e-01 -3.98216337e-01 3.77708822e-01 -1.08988345e+00
1.64558172e+00 -2.06239176e+00 -7.71396086e-02 -5.90614527e-02
-5.74671663e-02 6.33737445e-02 1.67826697e-01 1.06383300e+00
1.38610885e-01 5.84929049e-01 6.28701523e-02 -2.74112105e-01
3.02591687e-03 -4.09794673e-02 -4.17929292e-01 4.34435993e-01
1.28735557e-01 7.84749448e-01 -9.08169389e-01 -8.08386683e-01
-2.80997813e-01 3.69393528e-01 -2.21645683e-01 -8.02330964e-04
-1.88265026e-01 2.92961150e-01 -4.99607921e-01 5.65879762e-01
-1.30785540e-01 -4.80084270e-01 5.44869959e-01 1.94656223e-01
-7.49358952e-01 9.34028387e-01 -1.00142789e+00 1.57226872e+00
-5.07791460e-01 1.11861622e+00 3.31464335e-02 -4.26548004e-01
6.13402605e-01 7.32364297e-01 2.15335086e-01 -4.35638845e-01
-8.20214450e-02 3.85885894e-01 -5.50689222e-03 -5.89557409e-01
9.74101126e-01 1.19829595e-01 -3.13337892e-01 4.88708615e-01
1.06097609e-01 2.68866718e-01 3.56351167e-01 4.11047757e-01
1.06284714e+00 1.38151854e-01 3.69435400e-01 -2.81333834e-01
4.64402914e-01 4.31423873e-01 4.87021804e-01 6.10058844e-01
1.06470965e-01 4.74977493e-01 4.16413218e-01 -2.10121293e-02
-1.14578700e+00 -4.67741936e-01 -2.10728928e-01 9.94478166e-01
-4.03407633e-01 -6.81719184e-01 -7.55793631e-01 -4.80785072e-01
-2.18728960e-01 1.06185460e+00 -3.94015729e-01 1.79182753e-01
-8.34227145e-01 -4.38509524e-01 9.03817952e-01 3.69205475e-01
2.39849076e-01 -9.81879115e-01 -4.91630584e-01 5.29926360e-01
-6.28976643e-01 -1.21163845e+00 -2.56181240e-01 -3.81436828e-03
-3.78909290e-01 -8.06083202e-01 -6.92206204e-01 -5.62687695e-01
3.95618588e-01 -1.77365035e-01 1.19605947e+00 -3.48814763e-02
-9.36782826e-03 5.08314431e-01 -5.10248780e-01 -3.93304288e-01
-7.86657810e-01 6.24538660e-01 -1.25849247e-01 -1.95242614e-01
3.11132908e-01 -3.40895653e-01 -1.43455371e-01 7.43668154e-02
-7.51797915e-01 -2.36810535e-01 5.90294182e-01 3.83148164e-01
-9.44828764e-02 -4.81254935e-01 8.98225784e-01 -1.15697300e+00
7.43514121e-01 -8.64952862e-01 -4.05465692e-01 6.69545054e-01
-5.95129728e-01 1.16033256e-01 4.36301798e-01 -2.36637607e-01
-1.34459531e+00 -3.89393777e-01 -2.43175358e-01 4.26533967e-01
-2.21061423e-01 8.39638531e-01 1.17707364e-01 4.95608151e-02
7.58187413e-01 1.73143335e-02 -4.42731857e-01 -6.59056723e-01
3.60990435e-01 1.04257429e+00 8.05754423e-01 -6.58897221e-01
5.59104145e-01 2.26098578e-02 -5.48339069e-01 -1.01623392e+00
-8.30155492e-01 -6.01570249e-01 -4.55928385e-01 -3.80409837e-01
9.51804817e-01 -9.31024611e-01 -5.28385520e-01 2.14847684e-01
-1.39703608e+00 3.56773683e-03 -6.55740034e-03 5.29298127e-01
3.60201709e-02 4.43865687e-01 -9.83927488e-01 -8.35272968e-01
-3.34981948e-01 -6.67769372e-01 5.95123231e-01 3.28497350e-01
-8.27802420e-01 -1.04742324e+00 2.63197541e-01 4.73579168e-01
4.48318869e-01 2.23668590e-01 1.08338869e+00 -1.19795346e+00
-3.59541476e-01 -2.60162890e-01 -4.91348514e-03 -1.30700976e-01
-1.55735120e-01 8.86357948e-02 -8.17486107e-01 1.27162039e-01
-1.68766081e-01 -1.91321552e-01 2.87833214e-01 -2.36176223e-01
4.19534922e-01 -4.85470742e-01 -3.87553334e-01 -3.01371366e-01
1.32735515e+00 2.03098282e-01 5.33144593e-01 7.93225527e-01
3.32483798e-01 7.99488068e-01 2.49679640e-01 5.11159003e-01
6.96431577e-01 5.05266130e-01 -1.21688806e-01 1.63871646e-01
1.11683309e-01 -3.68432641e-01 9.93641689e-02 9.99215841e-01
6.71589449e-02 -4.14459497e-01 -1.25321436e+00 7.98218369e-01
-1.82092404e+00 -1.00443006e+00 -2.21806943e-01 2.07538319e+00
1.00893545e+00 2.85219610e-01 -1.49208874e-01 -3.97057146e-01
7.69140601e-01 5.69680780e-02 -1.01872586e-01 -2.51729876e-01
-2.48634189e-01 -1.69153288e-01 4.56871271e-01 3.42120290e-01
-6.43676579e-01 7.00659335e-01 5.84111452e+00 6.68095708e-01
-9.92829502e-01 1.15001790e-01 6.52537167e-01 2.87377894e-01
-2.98832893e-01 4.01672542e-01 -1.02404976e+00 5.81111312e-01
1.26018524e+00 -4.55688715e-01 1.69928700e-01 8.53801906e-01
7.20273852e-02 -2.02736989e-01 -9.79376912e-01 5.69955647e-01
2.32091561e-01 -1.56535494e+00 -4.00629580e-01 -5.09621426e-02
5.94482660e-01 3.78704935e-01 -4.93024439e-01 4.66309041e-01
3.09295833e-01 -7.50182629e-01 9.79647517e-01 4.02024686e-01
4.75698173e-01 -3.54654998e-01 7.80404747e-01 6.04264438e-01
-9.62302685e-01 3.50459926e-02 -4.22054715e-02 4.39585093e-03
5.87244272e-01 4.36766446e-01 -8.19499731e-01 5.69616914e-01
4.39343959e-01 -3.91557179e-02 -6.20994031e-01 9.40789998e-01
-2.26887241e-01 8.95084679e-01 -3.79188389e-01 -3.74691993e-01
4.09794688e-01 1.01963654e-01 4.94918734e-01 1.68565929e+00
1.04029886e-01 2.87154227e-01 -3.61395180e-02 5.55208683e-01
-3.66757989e-01 3.66748512e-01 -2.73753077e-01 -4.12092566e-01
8.97795856e-01 1.31541109e+00 -7.77043283e-01 -2.86555290e-01
-7.82962203e-01 7.56316245e-01 5.16211987e-01 1.90039933e-01
-4.76386249e-01 -5.66199422e-01 2.14951709e-01 4.74325046e-02
8.37988257e-02 -4.43917543e-01 -9.33293551e-02 -1.20699191e+00
1.25420481e-01 -8.93785119e-01 4.14954752e-01 -7.66840696e-01
-1.21034729e+00 7.17454076e-01 -2.20294818e-02 -8.62975180e-01
-4.05055434e-01 -2.68175751e-01 -5.30404985e-01 8.08247387e-01
-1.18337941e+00 -9.18528318e-01 2.04618037e-01 9.12220031e-02
4.79875267e-01 2.69487947e-02 8.69423687e-01 5.85007787e-01
-5.70886612e-01 3.82466346e-01 2.49446213e-01 5.61606348e-01
9.09471512e-01 -9.87266243e-01 6.01455450e-01 8.93685818e-01
4.14842039e-01 9.05449688e-01 9.09259975e-01 -5.65882921e-01
-1.29134929e+00 -6.50924683e-01 1.83205140e+00 -8.19239795e-01
1.05826414e+00 -3.63161534e-01 -9.56401885e-01 6.26905501e-01
3.48545581e-01 -6.06441021e-01 7.93619096e-01 3.62287521e-01
-2.51137584e-01 3.94043118e-01 -8.31122816e-01 7.00450718e-01
5.34047723e-01 -1.02074528e+00 -8.50982130e-01 4.26871032e-01
5.15705407e-01 -3.73443335e-01 -8.87812197e-01 -6.47224672e-03
5.39609194e-01 -5.56087852e-01 4.10999745e-01 -7.81920254e-01
4.25221950e-01 -3.95506322e-02 -1.64968781e-02 -9.91158009e-01
-1.91177979e-01 -7.64208555e-01 2.10122257e-01 1.91937470e+00
9.82632577e-01 -4.83422846e-01 5.97500443e-01 1.32218778e+00
-2.59431124e-01 -5.66992342e-01 -9.41137373e-01 -5.08638024e-01
5.45564443e-02 -4.33386624e-01 2.62869924e-01 1.17090189e+00
5.74173391e-01 6.76317930e-01 -2.22741544e-01 1.54298887e-01
-1.59302801e-01 -3.35715525e-03 5.48566937e-01 -1.25606060e+00
-3.95991445e-01 -1.58308968e-01 -2.43746825e-02 -8.00671339e-01
3.00044984e-01 -7.57514119e-01 7.66709074e-02 -1.50675690e+00
1.73252091e-01 -4.93515193e-01 1.20571844e-01 3.92203033e-01
-1.77305356e-01 -2.16575757e-01 2.60760903e-01 5.46266258e-01
-6.50286734e-01 4.75165509e-02 3.32683265e-01 2.82082707e-01
-1.82045445e-01 -2.54418641e-01 -8.01995695e-01 7.19207406e-01
6.57341897e-01 -7.31914341e-01 1.61962733e-01 -3.73432934e-01
7.14922428e-01 3.24922800e-01 2.31550876e-02 -8.94127071e-01
5.53421557e-01 2.33920291e-02 6.72043860e-02 -2.60355771e-01
5.11941426e-02 -4.62266237e-01 3.56088936e-01 1.42990097e-01
-5.99187195e-01 2.74008065e-01 1.75768763e-01 3.94526213e-01
-2.98437238e-01 -7.44654775e-01 3.08972448e-01 -3.90440643e-01
-7.09948897e-01 -3.00545573e-01 -7.36468077e-01 5.08653402e-01
5.52874327e-01 1.12098597e-01 -6.82912529e-01 -4.90991622e-01
-2.96220928e-01 2.15397421e-02 4.17966068e-01 4.10198063e-01
-2.09232524e-01 -7.29835749e-01 -8.23642373e-01 -5.05305231e-01
1.04722887e-01 -4.70512480e-01 6.63305400e-03 8.47554028e-01
-4.10084575e-01 7.47713029e-01 1.29000247e-01 8.11608285e-02
-1.00697911e+00 2.05080315e-01 -5.16688079e-02 -4.72842395e-01
-5.09695947e-01 6.18344247e-01 -3.83903503e-01 -2.01481059e-01
1.02153853e-01 2.00837501e-03 -2.79971004e-01 3.25633556e-01
3.80143434e-01 5.37212968e-01 1.88358903e-01 -5.63872814e-01
-3.50363851e-01 -8.75261948e-02 -9.26812291e-02 -4.69464004e-01
1.28223848e+00 -2.26876006e-01 -4.72903624e-02 6.37073338e-01
9.02555764e-01 6.17319345e-01 -5.97303331e-01 -1.48665905e-01
6.62091017e-01 -1.35940267e-02 -3.57401133e-01 -7.73702264e-01
-3.29588175e-01 4.33060050e-01 -1.65028006e-01 5.41962147e-01
4.88200277e-01 2.91917413e-01 9.32701111e-01 6.91096723e-01
3.51566434e-01 -1.38457346e+00 -3.90576929e-01 7.74896681e-01
5.74612379e-01 -1.09235001e+00 -2.87361760e-02 -1.45745873e-01
-7.64489412e-01 1.04258358e+00 2.57475287e-01 5.46277463e-01
3.44021499e-01 2.13191837e-01 6.50959983e-02 -1.79320753e-01
-8.68459463e-01 3.89700085e-02 1.24546681e-02 1.03337854e-01
9.21236753e-01 -1.91791475e-01 -6.55240834e-01 6.34300709e-01
-2.04693019e-01 -2.66555697e-01 7.05708981e-01 1.03217328e+00
-4.61821407e-01 -1.04262221e+00 -1.06077462e-01 2.16346696e-01
-1.09234738e+00 -4.47516710e-01 -6.40333951e-01 8.71975303e-01
-2.53213286e-01 1.30067265e+00 -1.19548105e-01 -7.08943978e-02
1.85509637e-01 5.25685310e-01 -2.13462219e-01 -7.21039593e-01
-9.42522228e-01 -2.90358692e-01 9.76582348e-01 -1.48143144e-02
-4.12623584e-01 -8.35045159e-01 -1.21949565e+00 -2.88237482e-01
-5.80792010e-01 7.64985800e-01 9.68100131e-01 1.19898057e+00
5.45948744e-01 3.96066316e-04 2.65880764e-01 -4.00894314e-01
-4.17280585e-01 -1.08539248e+00 -2.12829247e-01 1.31321907e-01
-2.05687910e-01 -1.78829759e-01 -3.60883862e-01 1.81374401e-01] | [9.798818588256836, 9.655640602111816] |
bfac7274-2344-4a62-9462-5c8fd6c4cbc2 | learning-to-selectively-learn-for-weakly | 2109.12457 | null | https://arxiv.org/abs/2109.12457v1 | https://arxiv.org/pdf/2109.12457v1.pdf | Learning to Selectively Learn for Weakly-supervised Paraphrase Generation | Paraphrase generation is a longstanding NLP task that has diverse applications for downstream NLP tasks. However, the effectiveness of existing efforts predominantly relies on large amounts of golden labeled data. Though unsupervised endeavors have been proposed to address this issue, they may fail to generate meaningful paraphrases due to the lack of supervision signals. In this work, we go beyond the existing paradigms and propose a novel approach to generate high-quality paraphrases with weak supervision data. Specifically, we tackle the weakly-supervised paraphrase generation problem by: (1) obtaining abundant weakly-labeled parallel sentences via retrieval-based pseudo paraphrase expansion; and (2) developing a meta-learning framework to progressively select valuable samples for fine-tuning a pre-trained language model, i.e., BART, on the sentential paraphrasing task. We demonstrate that our approach achieves significant improvements over existing unsupervised approaches, and is even comparable in performance with supervised state-of-the-arts. | ['Huan Liu', 'Yang Liu', 'Chenlei Guo', 'Xing Fan', 'Alexander Hanbo Li', 'Dingcheng Li', 'Kaize Ding'] | 2021-09-25 | null | https://aclanthology.org/2021.emnlp-main.480 | https://aclanthology.org/2021.emnlp-main.480.pdf | emnlp-2021-11 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 6.14257872e-01 2.52707452e-02 -5.82288206e-01 -4.67656791e-01
-1.38139999e+00 -5.87219834e-01 7.28803098e-01 7.93863907e-02
-1.88420027e-01 9.59198058e-01 6.81929350e-01 -3.72444063e-01
1.02524832e-01 -5.35821617e-01 -8.89621079e-01 -3.37674320e-01
7.18587339e-01 6.45259798e-01 -1.37972385e-01 -4.18545067e-01
5.83972394e-01 2.05466032e-01 -1.28853726e+00 5.68687677e-01
1.25657892e+00 3.53965223e-01 1.27878398e-01 3.15837890e-01
-2.52576470e-01 9.54902947e-01 -5.12980223e-01 -5.41549385e-01
1.21136837e-01 -9.80091035e-01 -1.12870538e+00 4.71122041e-02
2.67468631e-01 -6.73342571e-02 -2.57851213e-01 9.62432623e-01
3.49765033e-01 1.37744114e-01 6.96060836e-01 -9.74740565e-01
-1.02221310e+00 8.47539783e-01 -5.09147942e-01 2.66560644e-01
5.85822463e-01 1.36866197e-01 1.49278378e+00 -1.22931945e+00
5.61303020e-01 9.96644795e-01 4.64004785e-01 6.74498022e-01
-1.44923234e+00 -5.40441096e-01 -1.47167668e-01 1.84932992e-01
-1.09360921e+00 -6.59021556e-01 1.05511808e+00 -1.89863697e-01
9.54927206e-01 4.70772088e-02 4.76936758e-01 1.40065038e+00
-1.68113634e-01 1.07056034e+00 1.04599631e+00 -8.39683175e-01
2.20353037e-01 1.81900188e-01 1.00089557e-01 5.21741271e-01
9.74765420e-02 -1.00545563e-01 -7.99569368e-01 -2.84575403e-01
3.93940061e-01 3.71687114e-02 -2.43752480e-01 -3.90877306e-01
-1.31235313e+00 9.91448224e-01 3.46586406e-01 3.99751782e-01
-3.18243802e-01 -6.94088191e-02 4.39373285e-01 6.87071085e-01
7.05178857e-01 9.91922855e-01 -3.94506872e-01 -1.64531544e-01
-1.36543453e+00 2.57161975e-01 9.00408328e-01 1.12939465e+00
7.58223832e-01 -2.80728132e-01 -3.56591612e-01 1.27879405e+00
-6.57851845e-02 2.86020994e-01 7.74118781e-01 -9.14336681e-01
9.65301692e-01 5.49039483e-01 2.37254515e-01 -4.60190892e-01
1.67598426e-01 -3.61827165e-01 -6.79656088e-01 -5.18029034e-01
3.44477564e-01 7.32166246e-02 -7.73172498e-01 1.71161902e+00
-7.00137168e-02 7.87293762e-02 2.93714464e-01 6.74056411e-01
5.91346979e-01 7.37060726e-01 -3.99693847e-02 -2.39739656e-01
1.02756131e+00 -1.61505723e+00 -3.43226194e-01 -5.98126471e-01
5.09263158e-01 -8.64539623e-01 1.59917223e+00 2.32094482e-01
-1.36965346e+00 -5.68935931e-01 -9.11282778e-01 -2.52432406e-01
4.01319703e-03 2.05122933e-01 4.89755332e-01 2.95185387e-01
-8.30054283e-01 8.16486895e-01 -4.54228669e-01 -4.63819087e-01
5.46221197e-01 2.56581251e-02 -1.32454738e-01 -3.38304549e-01
-1.21463764e+00 9.61782515e-01 2.55801380e-01 -5.61379828e-02
-7.86169589e-01 -8.00317347e-01 -8.08571458e-01 2.43030921e-01
2.68796653e-01 -1.18805206e+00 1.52033234e+00 -1.02205014e+00
-1.72882676e+00 9.71323371e-01 -4.51267511e-01 -4.63040739e-01
3.80992591e-01 -4.47759867e-01 9.69338641e-02 1.25874072e-01
3.94746959e-01 6.15974367e-01 8.67862761e-01 -1.06563723e+00
-3.75497639e-01 -2.06044212e-01 1.10247303e-02 4.37701643e-01
-5.47679365e-01 1.47793487e-01 -3.60404521e-01 -9.15463150e-01
3.58224995e-02 -9.73938286e-01 -4.54083979e-01 -3.54443580e-01
-4.58220452e-01 -2.56396860e-01 2.09820613e-01 -5.26913762e-01
1.09198892e+00 -1.78125668e+00 4.91807908e-01 -1.34206206e-01
-1.23776458e-01 4.67515409e-01 -3.47455800e-01 8.09371173e-01
5.50806336e-03 8.32630321e-02 -4.37731504e-01 -5.55282056e-01
-8.26109350e-02 1.92533266e-02 -8.77060473e-01 -2.54709441e-02
3.38879108e-01 1.33347249e+00 -1.36876190e+00 -5.18751681e-01
-4.07589413e-03 -1.29113689e-01 -5.87912083e-01 4.03561682e-01
-2.71589637e-01 4.65242386e-01 -5.49964368e-01 5.46756089e-01
1.54168874e-01 -5.57561040e-01 1.11757545e-03 6.19480498e-02
2.75134623e-01 7.04421401e-01 -3.62407655e-01 2.23783445e+00
-9.33935940e-01 4.58075345e-01 -3.52970362e-01 -1.29740369e+00
8.50994706e-01 1.63650215e-01 2.52224147e-01 -4.75982785e-01
-9.97198448e-02 5.47039390e-01 -2.68277496e-01 -5.59421599e-01
6.81370914e-01 -6.07844830e-01 -2.16588259e-01 7.62368023e-01
1.29923657e-01 -5.55579901e-01 3.23070854e-01 4.03041244e-01
1.26222503e+00 2.43108988e-01 4.29302603e-01 -7.21512139e-02
7.23713458e-01 2.70685643e-01 3.34521055e-01 8.78316462e-01
-3.38873640e-02 8.39307725e-01 3.44310135e-01 -2.09734906e-02
-1.36491561e+00 -1.24061668e+00 2.37274289e-01 9.88613963e-01
9.66314152e-02 -3.32005382e-01 -5.71822286e-01 -8.10260832e-01
-3.14081386e-02 8.37881207e-01 -2.66676664e-01 -3.46663356e-01
-7.33328700e-01 -6.25032365e-01 6.97565615e-01 5.94810009e-01
3.07225853e-01 -1.39564347e+00 -1.11759617e-03 2.93988883e-01
-5.86057782e-01 -1.03296006e+00 -5.84445596e-01 4.86333966e-02
-1.16100764e+00 -7.68790603e-01 -9.78599727e-01 -1.07528496e+00
7.48670995e-01 4.86856580e-01 1.40745199e+00 -1.31132215e-01
2.09417138e-02 -9.21533182e-02 -5.38305998e-01 -1.16543412e-01
-7.69357443e-01 6.09139860e-01 -1.70732841e-01 -2.93413699e-01
5.35122156e-01 -8.55655313e-01 -5.71392417e-01 -3.38902809e-02
-6.90022409e-01 2.34884024e-01 9.95602787e-01 1.25978208e+00
5.25069952e-01 -5.95160663e-01 1.27771628e+00 -1.22265756e+00
1.12583292e+00 -5.33060193e-01 -9.80098844e-02 4.27223295e-01
-6.88454747e-01 3.54137987e-01 1.09142947e+00 -2.95143515e-01
-1.02429509e+00 -4.08158526e-02 -1.51463181e-01 -4.06961024e-01
-2.93414500e-02 5.07896602e-01 1.06351607e-01 1.99139282e-01
7.97485054e-01 5.85528791e-01 -8.90678391e-02 -4.98474866e-01
6.69903874e-01 9.12516415e-01 4.48001325e-01 -7.56115794e-01
9.77945089e-01 3.03070396e-01 -3.36633861e-01 -6.01136684e-01
-1.45504701e+00 -7.17845976e-01 -5.95654666e-01 2.37112090e-01
3.57618153e-01 -1.10661614e+00 1.52075827e-01 1.02764241e-01
-1.13131833e+00 -2.80078083e-01 -5.99891365e-01 2.00691342e-01
-8.03556859e-01 6.31103992e-01 -9.11420643e-01 -3.07000965e-01
-7.70562053e-01 -8.97375643e-01 1.21938074e+00 4.74630706e-02
-5.38998544e-01 -6.45278156e-01 4.52824354e-01 9.97184455e-01
3.54093820e-01 -3.22610229e-01 1.00101316e+00 -7.32134998e-01
-5.56905210e-01 -1.46679163e-01 -2.87398845e-01 5.33578396e-01
3.40582758e-01 -4.56587017e-01 -7.95329571e-01 -1.62154779e-01
-1.62076741e-03 -1.00497568e+00 9.89251375e-01 -7.45404838e-03
1.08700430e+00 -3.57262492e-01 -2.75592119e-01 4.24981296e-01
1.00918758e+00 -2.53426075e-01 4.55066442e-01 2.11770520e-01
5.24091125e-01 7.21480548e-01 7.11633742e-01 1.93503007e-01
1.72686130e-01 5.91425776e-01 -2.67290384e-01 7.09485784e-02
-1.42523766e-01 -8.49525273e-01 3.77895743e-01 1.06218863e+00
3.19184393e-01 -1.00161962e-01 -5.67083538e-01 9.08788204e-01
-1.99417758e+00 -1.14592981e+00 1.67734683e-01 1.85976875e+00
1.36813116e+00 8.44648033e-02 4.09717439e-03 1.89105943e-02
6.12466633e-01 3.29507440e-01 -6.87612474e-01 -3.62783670e-01
3.59776169e-02 5.67093551e-01 -1.74078364e-02 2.56094128e-01
-7.62995243e-01 1.32680809e+00 6.09897089e+00 8.78199995e-01
-7.70256996e-01 9.22377557e-02 4.92290497e-01 -3.04213285e-01
-6.66896164e-01 2.43679419e-01 -6.36106968e-01 5.50791442e-01
7.89782047e-01 -4.10064995e-01 5.82987189e-01 9.18559670e-01
3.39432597e-01 2.34462962e-01 -1.33543491e+00 8.23944926e-01
4.76686567e-01 -1.30200231e+00 3.68609071e-01 -2.78359622e-01
1.06486642e+00 3.06322016e-02 -1.62449747e-01 5.13188779e-01
4.84122992e-01 -9.87871051e-01 4.74570394e-01 1.49614647e-01
6.90128386e-01 -6.08241796e-01 5.35461366e-01 6.14493072e-01
-8.43227029e-01 -8.60490501e-02 -5.80362201e-01 -2.74295285e-02
4.59890485e-01 7.69979715e-01 -6.97679043e-01 5.66291630e-01
2.13308856e-01 9.19346094e-01 -6.00355864e-01 8.65086436e-01
-8.46068859e-01 8.99170160e-01 -1.17522767e-02 -2.36947894e-01
3.09424818e-01 -2.76961923e-01 3.62396330e-01 1.17132270e+00
2.61959881e-01 -7.10656568e-02 2.44414918e-02 1.14052331e+00
-5.38227916e-01 3.11129659e-01 -7.60355711e-01 -2.23297611e-01
5.79624116e-01 1.02682626e+00 -2.31100217e-01 -4.65665191e-01
-4.82869804e-01 1.33895600e+00 7.76832759e-01 3.43952209e-01
-6.36279166e-01 -5.69401085e-01 2.65483230e-01 -9.56376195e-02
2.27176383e-01 9.74418297e-02 -4.24778879e-01 -1.73983991e+00
2.69413352e-01 -1.03797960e+00 2.56767720e-01 -8.30775142e-01
-1.75532770e+00 4.10354763e-01 -1.99138075e-01 -1.23307145e+00
-6.40294313e-01 -1.19806916e-01 -8.78614366e-01 1.00004971e+00
-1.66367269e+00 -1.33292496e+00 -1.14305116e-01 3.04996550e-01
1.07294750e+00 -2.63537973e-01 6.48068964e-01 1.08374752e-01
-4.45514083e-01 6.67657137e-01 1.94326818e-01 -8.92295390e-02
9.78684306e-01 -1.10842061e+00 7.10897863e-01 9.35678244e-01
5.80709755e-01 9.09766197e-01 6.62089348e-01 -5.04583478e-01
-1.24758947e+00 -1.14694560e+00 1.44001830e+00 -4.72969741e-01
8.72609913e-01 -3.94257009e-01 -8.25911939e-01 6.01248622e-01
2.42146060e-01 -3.58024746e-01 6.71836853e-01 3.54776502e-01
-4.88665372e-01 4.95636323e-03 -8.75283360e-01 8.46597135e-01
1.35703850e+00 -8.35039914e-01 -1.28765333e+00 5.56781590e-01
7.34546483e-01 -9.97487083e-02 -4.74034995e-01 2.88256705e-01
3.93477947e-01 -7.51409054e-01 1.10826659e+00 -9.05148804e-01
1.18601537e+00 5.87598942e-02 1.56498268e-01 -1.50899541e+00
-2.65376925e-01 -7.82435298e-01 -9.94851291e-02 1.36174846e+00
5.98826587e-01 -2.86209822e-01 1.20080149e+00 3.03340048e-01
-1.72822237e-01 -9.57663953e-01 -6.90074444e-01 -8.30572009e-01
3.33796710e-01 -6.58196816e-03 3.26809078e-01 8.50901484e-01
5.17051101e-01 1.11669362e+00 -4.83494878e-01 -3.19243103e-01
5.49996376e-01 7.74268210e-01 9.11137283e-01 -7.91100979e-01
-6.38002574e-01 -3.64212930e-01 2.63585150e-01 -1.56635416e+00
6.54837787e-01 -1.32179916e+00 3.19678724e-01 -1.62013209e+00
6.18759394e-01 -3.50188464e-01 -1.48479670e-01 3.58009636e-01
-5.34235656e-01 2.36232340e-01 2.54015550e-02 6.52744651e-01
-5.48277557e-01 8.57535362e-01 1.12840903e+00 -1.66303441e-01
-2.92106301e-01 1.80918410e-01 -1.05223489e+00 4.56872255e-01
8.29174995e-01 -5.59336960e-01 -7.19530582e-01 -5.30919075e-01
2.61415452e-01 1.56029344e-01 2.78121829e-01 -6.09981358e-01
2.11223289e-01 -2.52962261e-01 -5.37702292e-02 -5.52344918e-01
2.00744435e-01 -2.53574938e-01 -3.30253512e-01 2.44778946e-01
-9.56514835e-01 1.09572083e-01 -2.33084917e-01 7.40299463e-01
-4.81504023e-01 -7.18568385e-01 7.25310206e-01 -3.12862784e-01
-2.78766066e-01 -3.42025422e-02 -2.12114111e-01 6.63297892e-01
5.95801115e-01 3.23278308e-02 -2.34162211e-01 -4.19307321e-01
-3.07911247e-01 1.95863262e-01 6.97816610e-01 5.03131330e-01
5.51535845e-01 -1.21055806e+00 -8.87177408e-01 5.71618453e-02
3.71674210e-01 1.78164133e-04 -1.10840939e-01 3.25425535e-01
-2.45166019e-01 5.77848673e-01 8.82169679e-02 -2.07153171e-01
-9.56072211e-01 6.06393874e-01 -1.20953538e-01 -8.72808397e-01
-5.66996038e-01 8.60723317e-01 2.80747972e-02 -4.85496938e-01
-8.15127194e-02 -8.09432790e-02 -4.30162810e-02 -6.49161786e-02
3.38658720e-01 5.66772744e-02 6.53572679e-02 -2.84952343e-01
-2.87972242e-02 3.83046955e-01 -4.32634532e-01 -2.38192573e-01
1.45231581e+00 -6.73511848e-02 -1.53904837e-02 2.82825142e-01
1.17653322e+00 2.94221547e-02 -7.40917861e-01 -5.22913396e-01
2.55996794e-01 -4.39938188e-01 -4.49369609e-01 -6.95542455e-01
-6.70987844e-01 8.63055587e-01 -4.18596953e-01 -2.00575948e-01
9.84247565e-01 1.42137051e-01 1.29974723e+00 8.13987434e-01
5.14819920e-01 -9.54539716e-01 5.77464402e-01 5.44435918e-01
8.79572928e-01 -1.21580899e+00 5.02131023e-02 -3.84033650e-01
-7.90273190e-01 7.56636679e-01 4.63252187e-01 -3.76437545e-01
9.12458375e-02 -1.86217889e-01 -8.58322755e-02 -8.74430500e-03
-9.55915332e-01 -1.28672167e-03 2.00481102e-01 2.99034357e-01
5.30201852e-01 -4.09383237e-01 -6.28664374e-01 5.78638136e-01
-3.78732979e-01 3.90536189e-01 4.11448091e-01 9.37012911e-01
-3.32314849e-01 -1.50617504e+00 1.06757559e-01 6.49735689e-01
-3.49773169e-01 -5.89832246e-01 -6.32470727e-01 3.01487565e-01
-4.19344991e-01 9.85757470e-01 -4.17491317e-01 -1.50765941e-01
4.17561322e-01 2.72657126e-01 5.44004500e-01 -1.14156532e+00
-7.64199376e-01 -3.48092288e-01 2.13403240e-01 -2.30742440e-01
-3.05314869e-01 -6.20400727e-01 -7.31482685e-01 9.72248521e-03
-3.35103959e-01 4.72020328e-01 2.24841103e-01 1.08345640e+00
5.68142414e-01 7.41502717e-02 8.58253181e-01 -8.20380092e-01
-1.24356711e+00 -1.09643173e+00 -1.93267748e-01 8.32141042e-01
7.37415403e-02 -2.23972380e-01 -5.45729160e-01 1.52590841e-01] | [11.654313087463379, 9.24806022644043] |
abebf61f-a9e8-4aa9-ba3f-0d9bd82dc6b7 | leveraging-a-bilingual-dictionary-to-learn | null | null | https://aclanthology.org/2022.lrec-1.124 | https://aclanthology.org/2022.lrec-1.124.pdf | Leveraging a Bilingual Dictionary to Learn Wolastoqey Word Representations | Word embeddings (Mikolov et al., 2013; Pennington et al., 2014) have been used to bolster the performance of natural language processing systems in a wide variety of tasks, including information retrieval (Roy et al., 2018) and machine translation (Qi et al., 2018). However, approaches to learning word embeddings typically require large corpora of running text to learn high quality representations. For many languages, such resources are unavailable. This is the case for Wolastoqey, also known as Passamaquoddy-Maliseet, an endangered low-resource Indigenous language. As there exist no large corpora of running text for Wolastoqey, in this paper, we leverage a bilingual dictionary to learn Wolastoqey word embeddings by encoding their corresponding English definitions into vector representations using pretrained English word and sequence representation models. Specifically, we consider representations based on pretrained word2vec (Mikolov et al., 2013), RoBERTa (Liu et al., 2019) and sentence-BERT (Reimers and Gurevych, 2019) models. We evaluate these embeddings in word prediction tasks focused on part-of-speech, animacy, and transitivity; semantic clustering; and reverse dictionary search. In all evaluations we demonstrate that approaches using these embeddings outperform task-specific baselines, without requiring any language-specific training or fine-tuning. | ['Paul Cook', 'Diego Bear'] | null | null | null | null | lrec-2022-6 | ['learning-word-embeddings', 'reverse-dictionary'] | ['methodology', 'natural-language-processing'] | [-9.85074341e-02 -1.42733872e-01 -5.72990298e-01 -1.28037512e-01
-4.51136351e-01 -7.84416318e-01 8.65889847e-01 4.77682143e-01
-8.79083514e-01 6.24291837e-01 4.88823146e-01 -7.24776328e-01
-5.35123870e-02 -9.55114484e-01 -4.54830498e-01 -2.87295133e-01
3.70970145e-02 3.90450627e-01 -3.11457843e-01 -6.63443506e-01
2.78005779e-01 3.74280602e-01 -1.19244444e+00 -4.33024950e-02
8.54447603e-01 5.24928570e-01 3.88041556e-01 5.79111338e-01
-2.79573858e-01 3.27958256e-01 -3.42437863e-01 -4.78140324e-01
1.59597576e-01 -1.04162440e-01 -8.36685717e-01 -3.97598326e-01
2.65372604e-01 -1.52591005e-01 -6.01048946e-01 9.23979878e-01
5.16150773e-01 2.65335470e-01 8.40817630e-01 -8.89354885e-01
-1.50188065e+00 8.25372517e-01 -3.56821865e-01 3.46368581e-01
2.48952731e-01 3.42446536e-01 1.60961723e+00 -1.21478748e+00
8.11962783e-01 1.24866438e+00 7.13541508e-01 3.81633371e-01
-1.17004228e+00 -7.03650653e-01 1.04949223e-02 4.74605709e-01
-1.48069298e+00 -2.21746504e-01 3.96713525e-01 -5.06500542e-01
1.45513999e+00 2.16935799e-02 7.32779443e-01 1.28327250e+00
1.49338454e-01 6.37368679e-01 8.61565292e-01 -6.36506677e-01
-1.22926138e-01 -3.14984135e-02 1.80782288e-01 6.47504210e-01
2.34205738e-01 5.60578331e-02 -5.37402868e-01 -1.19299382e-01
6.17100120e-01 1.91148773e-01 -4.17038463e-02 7.76825473e-02
-1.42020214e+00 1.21499383e+00 4.51385617e-01 5.64446032e-01
-4.12100434e-01 1.88581705e-01 6.29149854e-01 3.86405408e-01
5.66074431e-01 8.09333086e-01 -6.13941908e-01 -3.79501879e-01
-6.31078362e-01 1.60062447e-01 5.24375260e-01 9.32230890e-01
7.49179482e-01 1.65004134e-01 5.73113337e-02 1.31486797e+00
2.31790930e-01 5.55304289e-01 9.48064744e-01 -3.21572334e-01
3.18196505e-01 4.15907383e-01 -2.11033002e-01 -9.92563546e-01
-3.29789042e-01 -9.74137932e-02 -5.75670958e-01 -4.89679635e-01
1.79099962e-01 -1.26082927e-03 -9.78109479e-01 1.73814642e+00
1.35870546e-01 -1.23891141e-02 3.01882356e-01 8.94042730e-01
8.51832628e-01 1.09814978e+00 1.78943053e-01 2.18837932e-01
1.51928568e+00 -1.04416656e+00 -5.30531347e-01 -3.92968833e-01
7.33466864e-01 -8.64905655e-01 1.44145167e+00 2.31447190e-01
-6.98770761e-01 -4.52676028e-01 -8.64420831e-01 -4.65713918e-01
-7.79543161e-01 -1.97807439e-02 8.18458378e-01 4.95132923e-01
-7.46046901e-01 4.51958805e-01 -6.84521377e-01 -9.75796759e-01
1.79096609e-01 5.20230457e-02 -5.41654885e-01 -3.14599484e-01
-1.50941491e+00 1.39667892e+00 4.21604216e-01 -1.54162094e-01
-7.98855484e-01 -7.87458956e-01 -1.11393750e+00 -9.09190476e-02
1.18800230e-01 -2.99702734e-01 8.14559042e-01 -7.05328584e-01
-1.18632340e+00 1.23841882e+00 1.46498650e-01 -4.90270913e-01
-4.43466119e-02 -4.13275212e-01 -5.62964082e-01 2.68333554e-02
1.77830607e-01 6.89986110e-01 5.89774430e-01 -6.99814379e-01
-2.48477027e-01 -2.89508909e-01 3.01460207e-01 1.32490888e-01
-7.69054770e-01 3.73380840e-01 -1.35566860e-01 -1.12400115e+00
-2.93494076e-01 -8.78115535e-01 -7.91579410e-02 4.93895598e-02
-5.30578420e-02 -5.36818027e-01 3.14626008e-01 -1.01749909e+00
1.35368323e+00 -2.22496939e+00 2.71383822e-01 -7.91319311e-02
6.58272132e-02 4.72727209e-01 -6.92855537e-01 1.05457544e+00
-3.61096896e-02 5.22881210e-01 -4.46513981e-01 6.53925911e-02
1.84164509e-01 6.56952322e-01 -3.61698985e-01 5.73820114e-01
4.45176214e-01 1.05938303e+00 -1.05602753e+00 -2.82590538e-01
2.21045956e-01 5.19064784e-01 -6.54783487e-01 9.62020084e-02
-8.69489312e-02 7.09079728e-02 -2.89479028e-02 6.43477023e-01
2.32138544e-01 1.75769240e-01 2.91168928e-01 -1.20876729e-01
-3.32835823e-01 7.01038241e-01 -6.78642213e-01 1.85036147e+00
-1.09637964e+00 8.04571688e-01 -3.03703189e-01 -1.09303153e+00
9.63616967e-01 2.60242045e-01 3.30778152e-01 -6.59086585e-01
5.86611480e-02 3.86027873e-01 1.69738278e-01 -5.81420660e-01
8.64428580e-01 -3.40726584e-01 -3.85487020e-01 5.32384157e-01
3.54226977e-01 -3.71356696e-01 3.60881597e-01 1.62728980e-01
1.00951481e+00 2.07917932e-02 6.35684311e-01 -3.91127944e-01
3.36265296e-01 2.22332209e-01 4.67527270e-01 3.30021858e-01
-1.47274047e-01 6.49930358e-01 2.56245315e-01 -3.46233875e-01
-1.26555920e+00 -1.18292177e+00 -5.37362099e-01 1.43255889e+00
-5.78738451e-02 -6.65689051e-01 -2.54626840e-01 -2.45836198e-01
1.99531123e-01 7.88754225e-01 -7.26045191e-01 -3.37269992e-01
-5.90600967e-01 -5.00969470e-01 9.05742407e-01 3.31953287e-01
-4.02044319e-02 -1.20464003e+00 -4.71977025e-01 3.89165908e-01
-9.23508331e-02 -9.70421553e-01 -6.66972876e-01 1.89731151e-01
-4.77394104e-01 -9.15644169e-01 -6.81987286e-01 -8.96213770e-01
3.13865453e-01 1.01660579e-01 1.19605803e+00 9.73760113e-02
-5.60753822e-01 3.66095304e-01 -8.08537304e-01 -3.42816174e-01
-3.45667541e-01 1.19240813e-01 4.69843715e-01 -2.74341941e-01
5.99223614e-01 -3.67543459e-01 -2.49787003e-01 -8.59475508e-02
-1.11745894e+00 -3.20608228e-01 4.07641947e-01 1.25578260e+00
4.79555517e-01 -5.43801725e-01 6.62708223e-01 -6.88350499e-01
7.50126541e-01 -8.80186856e-01 -3.05784732e-01 1.93329722e-01
-4.88544196e-01 -1.62123293e-02 7.92344391e-01 -6.94442987e-01
-3.48660350e-01 -6.27743661e-01 -4.14079875e-01 -3.00599813e-01
8.94598961e-02 8.27814698e-01 1.34595707e-01 2.35879138e-01
5.92612147e-01 3.28854620e-01 -3.40954334e-01 -5.20138443e-01
8.96634221e-01 1.06358278e+00 3.38226259e-01 -7.15867579e-01
7.82343447e-01 6.75797388e-02 -6.74979806e-01 -1.15456223e+00
-6.92223251e-01 -6.57230139e-01 -6.93348169e-01 1.97443739e-01
1.09268081e+00 -9.95297372e-01 -8.21553022e-02 1.65502802e-02
-1.29391813e+00 -2.90453553e-01 -3.22273523e-01 6.99538648e-01
-2.72488952e-01 2.70727038e-01 -6.29910529e-01 -4.87892687e-01
-2.82084286e-01 -8.20521832e-01 8.30743551e-01 -6.77015334e-02
-4.88225549e-01 -1.22206008e+00 3.29261214e-01 1.32030964e-01
4.52241570e-01 8.54722410e-02 1.31152475e+00 -8.69219482e-01
3.46026346e-02 6.68607354e-02 -1.98491454e-01 4.77815807e-01
2.78548241e-01 1.00092702e-01 -5.50581038e-01 -2.49379039e-01
-3.73096675e-01 -5.66762745e-01 7.41436481e-01 -9.22453701e-02
8.28849196e-01 -4.01616395e-01 1.49179578e-01 7.42184877e-01
1.44013846e+00 5.93374204e-03 3.44372630e-01 5.91995120e-01
6.41643167e-01 6.46858633e-01 4.69254524e-01 3.72158796e-01
5.48802555e-01 4.71792668e-01 6.27667904e-02 1.47584543e-01
-2.03827605e-01 -4.87423837e-01 6.89281404e-01 1.39774084e+00
1.41450122e-01 -2.33840674e-01 -1.23014963e+00 1.23604882e+00
-1.45948493e+00 -7.45628595e-01 -4.09688381e-03 1.81314385e+00
1.06659150e+00 -3.58472228e-01 -2.00551748e-01 -2.51336724e-01
4.27175432e-01 4.92202014e-01 -3.81259799e-01 -9.86544251e-01
-2.46638834e-01 6.90330386e-01 4.40534353e-01 3.86749268e-01
-8.36893082e-01 1.46161103e+00 5.19865751e+00 7.51563311e-01
-1.04505277e+00 4.53762889e-01 7.40372017e-02 3.88122797e-02
-6.51941717e-01 9.49904770e-02 -4.93112952e-01 3.68281156e-01
1.10926294e+00 -4.33584422e-01 6.72762334e-01 4.83290464e-01
-4.86537628e-02 2.09948450e-01 -9.68408048e-01 1.07500648e+00
3.93943042e-01 -1.17027676e+00 -3.32518155e-03 -9.71800089e-02
6.59247160e-01 5.22111773e-01 -2.76966095e-02 6.23618841e-01
4.66765612e-01 -1.38292682e+00 6.64898276e-01 5.77097535e-02
9.31903064e-01 -6.83993697e-01 4.75396782e-01 2.25484267e-01
-1.06728899e+00 9.64679345e-02 -7.41074860e-01 -6.05334714e-02
2.40940854e-01 4.38557714e-01 -3.72435778e-01 4.68882620e-01
6.32129967e-01 1.07419205e+00 -3.73696536e-01 4.29305196e-01
-6.65620446e-01 7.87410438e-01 -2.94754356e-01 -2.64736891e-01
5.75982213e-01 -4.61915225e-01 4.68091458e-01 1.43780565e+00
2.77434409e-01 8.77673998e-02 1.43519342e-01 7.36455321e-01
-3.11467111e-01 5.26716173e-01 -7.94073701e-01 -8.01888049e-01
4.88659859e-01 1.18724895e+00 -3.49409342e-01 -1.95377424e-01
-6.48636460e-01 9.54921126e-01 3.83805394e-01 3.13749790e-01
-6.74819708e-01 -6.14559412e-01 1.22014630e+00 -1.05467029e-01
3.24763983e-01 -7.29664922e-01 -1.71368033e-01 -1.39591897e+00
-2.76199818e-01 -9.56467748e-01 2.38526776e-01 -5.07851303e-01
-1.65159738e+00 5.66840231e-01 6.32726699e-02 -1.03515530e+00
-1.92488134e-01 -9.17851806e-01 -4.22837466e-01 9.87949967e-01
-1.53923309e+00 -1.42757261e+00 2.43280351e-01 3.58159512e-01
6.16093338e-01 -4.12674695e-01 1.03458834e+00 3.97207081e-01
-4.92062807e-01 4.96277481e-01 4.02021110e-01 2.94794887e-01
7.48747289e-01 -1.09969318e+00 6.64936483e-01 6.32172108e-01
6.55752540e-01 9.82989252e-01 4.25045997e-01 -4.87915307e-01
-1.80466163e+00 -1.08060551e+00 1.21068883e+00 -4.07773256e-01
1.39198780e+00 -6.69359386e-01 -1.01742852e+00 7.91136861e-01
5.27837753e-01 -9.14160758e-02 9.78430748e-01 3.20891172e-01
-7.20762610e-01 2.77747542e-01 -7.65070558e-01 7.65075684e-01
1.07646036e+00 -9.10557806e-01 -8.46617460e-01 5.75114191e-01
9.58839953e-01 3.75712067e-02 -1.08957148e+00 1.24920104e-02
5.84880054e-01 -3.27969342e-01 8.99400711e-01 -1.08291507e+00
6.13979101e-01 3.66562158e-02 -4.55297381e-01 -1.55772090e+00
-1.82616845e-01 -2.11216882e-01 2.85215974e-01 1.20080340e+00
4.78000849e-01 -6.77716315e-01 1.38755201e-03 2.22972203e-02
-2.18640655e-01 -7.77519703e-01 -8.34406734e-01 -1.05453205e+00
6.84465826e-01 -6.17156982e-01 5.39433181e-01 1.38408113e+00
6.48864955e-02 4.66426998e-01 -3.40369433e-01 -2.07263604e-02
6.02758229e-02 -3.25882435e-02 7.10533381e-01 -8.99049103e-01
-1.22938022e-01 -4.66221064e-01 -5.46249509e-01 -8.87209535e-01
6.30465388e-01 -1.56024969e+00 -7.15739876e-02 -1.82323253e+00
8.64698458e-03 -4.80381340e-01 -3.50400180e-01 5.93345404e-01
-1.31268978e-01 3.57381552e-01 2.66234040e-01 9.01698247e-02
-1.21420823e-01 8.79846156e-01 7.56063402e-01 -2.47835234e-01
1.79492831e-01 -8.83599281e-01 -6.24819279e-01 4.59140629e-01
8.58406842e-01 -5.39311171e-01 -1.39591977e-01 -9.38486278e-01
3.91816378e-01 -5.56796551e-01 1.83380514e-01 -3.77836108e-01
-5.76650724e-02 -4.51056689e-01 -2.27823094e-01 -8.96624103e-02
2.41630271e-01 -4.86033440e-01 -2.83622354e-01 4.95783925e-01
-3.11975002e-01 5.68822861e-01 2.99330801e-01 3.55600744e-01
-3.76425862e-01 -5.06747961e-01 6.96482122e-01 1.87372901e-02
-9.20950294e-01 4.61734772e-01 -5.43147266e-01 3.90020937e-01
8.91586185e-01 1.09239127e-02 -2.21660554e-01 -1.94271296e-01
-2.18090788e-01 2.96065181e-01 4.93180901e-01 9.02442932e-01
7.38605499e-01 -1.43830287e+00 -9.18087900e-01 2.20718861e-01
5.19923270e-01 -2.96003520e-01 -1.31892815e-01 6.05480969e-01
-7.57230163e-01 2.86000282e-01 -2.90348947e-01 -1.68563455e-01
-7.65417218e-01 5.60798943e-01 -1.28186241e-01 -1.60649002e-01
-5.75458348e-01 7.16237605e-01 5.99734038e-02 -1.05547118e+00
-2.56715864e-01 -3.21563542e-01 -2.54001230e-01 3.13553065e-01
3.25287759e-01 5.00961691e-02 -1.79175705e-01 -9.17421103e-01
-4.97766793e-01 4.71838742e-01 -5.22425696e-02 -1.19555019e-01
1.59611464e+00 1.45179396e-02 -3.50986212e-01 6.54561937e-01
1.43575788e+00 1.27508432e-01 -3.37306887e-01 -3.01234990e-01
1.74367979e-01 -4.76616174e-01 -1.17010169e-01 -5.10054469e-01
-7.54652023e-01 1.30468428e+00 3.66893053e-01 -8.12501609e-02
7.35859573e-01 5.53359389e-02 1.17165303e+00 4.37261045e-01
3.41884106e-01 -1.18437612e+00 -9.43079963e-02 1.06419790e+00
9.74137902e-01 -1.19301271e+00 -1.26050666e-01 1.65659547e-01
-6.26640797e-01 1.16900373e+00 3.39889526e-01 -1.70745611e-01
6.66540086e-01 -1.45462483e-01 4.62472811e-02 -1.36889005e-02
-7.98149765e-01 -4.09640431e-01 1.75728157e-01 5.97925544e-01
6.62478089e-01 3.34796906e-01 -8.96832526e-01 5.94829321e-01
-4.02022690e-01 -5.09283304e-01 4.71030504e-01 8.01228762e-01
-2.71526396e-01 -1.31378424e+00 -1.99747905e-01 6.22393370e-01
-2.88389325e-01 -8.30081165e-01 -2.93251365e-01 8.01457405e-01
1.32343680e-01 8.28073978e-01 2.16018587e-01 -2.56638914e-01
3.28489870e-01 3.26429643e-02 2.76779652e-01 -1.17232263e+00
-8.28222811e-01 -2.69206464e-01 1.50847182e-01 -1.57282248e-01
-2.01162219e-01 -4.74280477e-01 -1.25778723e+00 -3.41486871e-01
-2.12986887e-01 1.53420150e-01 7.20243335e-01 8.10850561e-01
2.38836840e-01 2.69909173e-01 3.47291619e-01 -5.33138096e-01
-5.14925897e-01 -1.26057351e+00 -6.24145210e-01 4.88090873e-01
-5.31854555e-02 -6.91613197e-01 -2.37220347e-01 -6.83187619e-02] | [10.835331916809082, 9.866747856140137] |
5e9b6ce2-4c99-4566-a544-0ec6c7c805ba | task-focused-few-shot-object-detection-for | 2201.12437 | null | https://arxiv.org/abs/2201.12437v2 | https://arxiv.org/pdf/2201.12437v2.pdf | Mobile Robot Manipulation using Pure Object Detection | This paper addresses the problem of mobile robot manipulation using object detection. Our approach uses detection and control as complimentary functions that learn from real-world interactions. We develop an end-to-end manipulation method based solely on detection and introduce Task-focused Few-shot Object Detection (TFOD) to learn new objects and settings. Our robot collects its own training data and automatically determines when to retrain detection to improve performance across various subtasks (e.g., grasping). Notably, detection training is low-cost, and our robot learns to manipulate new objects using as few as four clicks of annotation. In physical experiments, our robot learns visual control from a single click of annotation and a novel update formulation, manipulates new objects in clutter and other mobile settings, and achieves state-of-the-art results on an existing visual servo control and depth estimation benchmark. Finally, we develop a TFOD Benchmark to support future object detection research for robotics: https://github.com/griffbr/tfod. | ['Brent Griffin'] | 2022-01-28 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 3.95238884e-02 9.90785360e-02 -1.90570325e-01 -2.37874806e-01
-5.64782977e-01 -7.05258250e-01 1.68755159e-01 -9.26571339e-03
-6.48364723e-01 3.89147311e-01 -4.02256370e-01 -6.54592365e-02
5.09062819e-02 -1.65071458e-01 -1.25456142e+00 -3.19026768e-01
-2.33406350e-01 7.38861442e-01 9.31193531e-01 -2.97996938e-01
4.12991911e-01 5.74064314e-01 -1.67039299e+00 6.87856227e-02
5.23324013e-01 7.83567190e-01 1.09881330e+00 1.22177339e+00
5.64506829e-01 6.17809653e-01 -2.91056901e-01 3.80046099e-01
5.91365755e-01 3.15865934e-01 -8.05167139e-01 1.50899738e-01
5.22954702e-01 -9.76320803e-01 -4.60174263e-01 8.71395409e-01
6.53663695e-01 3.76604229e-01 5.51623821e-01 -1.54012454e+00
-4.22906071e-01 5.44988930e-01 -6.06262326e-01 2.04757880e-02
3.76773775e-01 9.84151721e-01 6.78030491e-01 -1.07875109e+00
8.08755934e-01 1.53535032e+00 4.58332598e-01 7.65339255e-01
-1.21869564e+00 -5.08323073e-01 4.95110989e-01 6.70861453e-02
-7.08096385e-01 -4.93800461e-01 1.80897772e-01 -7.18234181e-01
1.07133710e+00 -1.97630182e-01 5.94600201e-01 1.08330572e+00
8.88713598e-02 1.14584255e+00 4.12517190e-01 -3.06305915e-01
1.70417145e-01 -2.85133690e-01 3.16483416e-02 1.05844831e+00
4.77603197e-01 2.72327989e-01 -3.33980620e-01 2.37516299e-01
9.58173275e-01 7.88389146e-02 -3.02449077e-01 -1.15736341e+00
-1.66730833e+00 3.52704138e-01 5.74672937e-01 -3.50625753e-01
-2.71915793e-01 8.44673276e-01 3.40089113e-01 2.15648323e-01
-3.32822055e-01 6.21482134e-01 -7.90506303e-01 -4.07016009e-01
2.38112453e-02 6.23656452e-01 7.62975752e-01 1.82314146e+00
7.36556053e-01 -2.35606968e-01 -2.28750378e-01 6.98323429e-01
1.55459598e-01 7.50375986e-01 3.88899036e-02 -1.76068783e+00
5.54690957e-01 3.12476188e-01 6.79150879e-01 -5.31079948e-01
-6.61771715e-01 2.43619591e-01 1.56488106e-01 7.42893875e-01
4.84558314e-01 -2.77187228e-01 -1.14478302e+00 1.47589600e+00
6.89856410e-01 -2.32063442e-01 -1.19389996e-01 1.17127001e+00
6.04097903e-01 2.82468647e-01 -1.00915678e-01 2.83047050e-01
1.11700308e+00 -1.41220212e+00 -4.79570538e-01 -7.33854234e-01
5.54593861e-01 -5.24577618e-01 1.31910503e+00 6.58247769e-01
-1.07896030e+00 -5.39919138e-01 -9.28593099e-01 -3.64826083e-01
-1.27795711e-01 5.01707911e-01 6.84849620e-01 -2.77962953e-01
-7.05243826e-01 5.88661671e-01 -1.57626319e+00 -5.62595963e-01
4.17175591e-01 4.71650958e-01 -2.32442528e-01 -1.86187238e-01
-2.51425087e-01 1.05577135e+00 5.25023818e-01 -5.03799208e-02
-1.64210534e+00 -4.22244668e-01 -8.32003117e-01 -3.19947809e-01
1.15333045e+00 -5.33523798e-01 2.10508847e+00 -1.88016057e-01
-1.53449464e+00 7.18729138e-01 1.59081370e-01 -2.22315773e-01
6.30778134e-01 -8.36562335e-01 5.29578328e-01 3.03317666e-01
4.26563472e-01 1.15722251e+00 9.81596112e-01 -1.64742720e+00
-1.06759143e+00 -2.42480278e-01 3.43986362e-01 1.36784881e-01
9.01048779e-02 -2.13487163e-01 -8.24069619e-01 -3.73218685e-01
1.99798375e-01 -1.37943673e+00 -1.45222113e-01 8.42062116e-01
-3.94332886e-01 -3.12894195e-01 1.11926103e+00 -1.71723053e-01
3.74065191e-01 -2.03652477e+00 6.10075653e-01 -3.31321597e-01
2.40481257e-01 -1.02720059e-01 -1.87460020e-01 2.58297592e-01
5.15448570e-01 -3.80419463e-01 6.93656132e-02 -2.79280812e-01
1.33656919e-01 1.88411370e-01 -1.48249775e-01 4.06354576e-01
1.75364926e-01 9.20346558e-01 -1.30731201e+00 -3.12653869e-01
3.14069808e-01 -3.23530063e-02 -9.16792214e-01 3.46655369e-01
-7.51485407e-01 5.31153321e-01 -3.74617606e-01 1.07571185e+00
2.70679355e-01 -2.41687834e-01 5.44702075e-02 -1.67130232e-01
-2.71248758e-01 -3.19160745e-02 -9.94431257e-01 2.12535357e+00
-3.68269503e-01 7.50719309e-01 4.77903575e-01 -7.95507729e-01
5.71451485e-01 -1.79854810e-01 3.42461169e-01 6.46773959e-03
3.81478578e-01 3.03788155e-01 1.04406051e-01 -9.53042388e-01
6.09020293e-01 7.40069270e-01 -4.69667427e-02 9.76325721e-02
2.19886154e-01 -7.35805035e-01 2.14093447e-01 1.36675000e-01
1.38140106e+00 6.62749410e-01 1.62398711e-01 2.06014812e-02
-4.52525795e-01 5.48166752e-01 2.70519495e-01 1.13124287e+00
-4.61194009e-01 4.57441092e-01 1.15452893e-01 -4.05015051e-02
-1.06302667e+00 -1.09152985e+00 2.16869086e-01 1.58431363e+00
7.70054102e-01 -6.61851540e-02 -5.01884937e-01 -4.22226250e-01
5.86970329e-01 4.47764188e-01 -4.53049630e-01 -9.29133072e-02
-8.01168203e-01 -2.81209443e-02 1.61613613e-01 8.12810361e-01
1.93223432e-01 -1.35372186e+00 -1.49307752e+00 2.57258624e-01
3.36994091e-03 -1.24269617e+00 -4.88042444e-01 6.14004910e-01
-7.89863169e-01 -1.37904096e+00 -5.95407605e-01 -1.24675429e+00
7.24645078e-01 7.06492424e-01 6.25664592e-01 -4.92849983e-02
-8.65362823e-01 8.49490643e-01 -4.53014731e-01 -6.86522663e-01
-3.14802915e-01 5.34131303e-02 2.44774446e-01 -7.99097002e-01
-4.48428541e-02 -2.42119551e-01 -5.48038721e-01 4.54569519e-01
-2.47250289e-01 -7.65853748e-02 7.93149352e-01 7.25129008e-01
2.97090024e-01 -5.39334893e-01 1.86284378e-01 -1.99322000e-01
3.57483834e-01 -1.27075657e-01 -1.06195319e+00 -4.49911430e-02
-3.12074512e-01 1.15806766e-01 1.86019149e-02 -1.04050076e+00
-8.44486594e-01 5.08335352e-01 6.70938969e-01 -7.84165144e-01
-1.26044676e-01 5.14125228e-02 3.32072735e-01 -2.11486220e-01
9.74336028e-01 -1.60841316e-01 1.73604280e-01 -4.54939276e-01
5.80693781e-01 5.70024967e-01 9.66399848e-01 -7.84596145e-01
7.70037770e-01 4.46636021e-01 -2.70494282e-01 -5.24990261e-01
-5.86777687e-01 -6.37915134e-01 -8.89899433e-01 -3.22889000e-01
7.69005239e-01 -9.84812021e-01 -1.47425640e+00 6.07597470e-01
-1.32629561e+00 -1.14983797e+00 -1.53652772e-01 6.50454044e-01
-1.01023865e+00 9.05657709e-02 -6.40353739e-01 -7.65643477e-01
-6.38952777e-02 -1.29396725e+00 1.55531287e+00 5.24865910e-02
-2.25743636e-01 -9.74626690e-02 -3.17968398e-01 4.05512229e-02
9.55189019e-02 1.51424050e-01 4.60842073e-01 -1.11454524e-01
-1.03864038e+00 -1.98904932e-01 -1.53086752e-01 -4.09985706e-02
1.00173272e-01 -5.55776320e-02 -5.56177020e-01 -7.40692139e-01
-6.01538002e-01 -8.79570127e-01 8.14296007e-01 2.00395182e-01
1.19115949e+00 -1.57086745e-01 -7.95908451e-01 4.11028713e-01
9.48706448e-01 3.22140992e-01 1.60521865e-01 5.55625260e-01
7.20356226e-01 4.22898591e-01 1.38674176e+00 6.51696265e-01
3.14127266e-01 9.03024912e-01 9.14775789e-01 5.44743001e-01
-1.78610176e-01 -2.68036742e-02 4.62119848e-01 3.57756466e-02
-5.03490604e-02 -1.42357543e-01 -1.03164995e+00 5.39115191e-01
-2.23145032e+00 -6.66538298e-01 9.10939649e-02 1.89655972e+00
9.21620369e-01 5.13661385e-01 1.95916086e-01 -3.18813682e-01
6.32426560e-01 -3.62039417e-01 -1.28375673e+00 2.33665362e-01
4.85347450e-01 -2.05690220e-01 8.25718701e-01 5.41244209e-01
-1.14178908e+00 1.39183307e+00 5.79076910e+00 1.45060033e-01
-1.01112783e+00 -9.26212296e-02 -4.67812151e-01 -3.89010757e-01
5.85144699e-01 -4.80182096e-02 -9.60182905e-01 7.08249509e-02
-5.96636208e-03 1.28448769e-01 6.83794558e-01 1.54048800e+00
1.90855965e-01 -5.20156920e-01 -1.57266331e+00 9.96190548e-01
-2.56878585e-02 -8.34015608e-01 -3.77489775e-01 -2.89192080e-01
3.84077728e-01 5.21026313e-01 -7.66344070e-02 5.37149310e-01
7.06645906e-01 -5.57079732e-01 1.18712461e+00 3.34190667e-01
7.96822190e-01 -5.74969575e-02 1.59334540e-01 6.19928658e-01
-1.10245395e+00 -7.66042888e-01 -2.98640639e-01 -1.60871178e-01
1.41064823e-01 -2.56111205e-01 -1.29563439e+00 -2.86941200e-01
1.09790885e+00 7.77809024e-01 -3.09143424e-01 1.11921012e+00
-3.80226344e-01 4.04581390e-02 -4.77506608e-01 -3.48657459e-01
3.71445119e-02 4.64589655e-01 8.95383418e-01 8.51928294e-01
5.95680811e-02 1.95951745e-01 7.78965473e-01 7.65900314e-01
6.08289391e-02 -6.03860676e-01 -4.50924754e-01 1.14016064e-01
8.00795794e-01 1.15664268e+00 -5.98080933e-01 -3.23697418e-01
-7.95076638e-02 1.08560753e+00 5.67268550e-01 2.65360147e-01
-8.71315300e-01 -6.95175767e-01 6.09690905e-01 6.58075139e-02
7.20111787e-01 -6.98830783e-01 1.95222870e-02 -1.03265405e+00
3.37574482e-01 -7.51019776e-01 -1.14651501e-01 -1.17690778e+00
-7.37908542e-01 -8.31855759e-02 3.57404351e-01 -1.39765847e+00
-1.10629335e-01 -1.05465364e+00 -2.31491417e-01 1.07967213e-01
-1.30000448e+00 -8.21950257e-01 -1.08536160e+00 2.00158238e-01
1.06741416e+00 3.85554917e-02 5.13760865e-01 -5.16007952e-02
-2.38981679e-01 3.49890471e-01 -2.18500674e-01 1.01666138e-01
9.52039063e-01 -1.20112193e+00 3.92782897e-01 4.33317870e-01
-5.18064916e-01 4.31282669e-01 8.33195746e-01 -7.87090361e-01
-1.96946454e+00 -8.71472180e-01 -2.44785309e-01 -7.75261641e-01
5.40820837e-01 -6.16246521e-01 -5.97337842e-01 1.15576887e+00
-2.90407389e-01 2.38348261e-01 -4.59545821e-01 -2.87240148e-01
-9.05384049e-02 7.07921833e-02 -9.78992701e-01 8.37122202e-01
1.66400123e+00 -1.97333656e-02 -7.55022347e-01 6.32304907e-01
1.07970476e+00 -1.03334332e+00 -4.28072035e-01 5.65184116e-01
8.47870827e-01 -4.35192257e-01 9.56674814e-01 -6.96424961e-01
2.91873723e-01 -5.56178331e-01 -1.45694822e-01 -1.18798053e+00
-4.45457041e-01 -7.40929067e-01 -3.73526514e-01 3.85415852e-01
3.50198090e-01 -2.46628881e-01 7.09266961e-01 4.12774503e-01
-5.16525388e-01 -5.64650655e-01 -5.33423483e-01 -1.07431912e+00
-4.53742087e-01 -2.74264574e-01 -1.23261452e-01 4.80767578e-01
2.85104692e-01 2.21825406e-01 -2.76950091e-01 4.47879791e-01
6.89854085e-01 8.23566169e-02 1.56299949e+00 -1.10594273e+00
-4.36955869e-01 -2.73124039e-01 -3.12171370e-01 -1.67271471e+00
4.19901572e-02 -6.36506081e-01 1.04841828e+00 -1.52697837e+00
2.17989728e-01 -5.05326688e-01 2.97923476e-01 9.35260475e-01
-7.10944831e-02 -1.56381205e-01 4.39629436e-01 3.33298594e-01
-9.14871097e-01 4.72244978e-01 1.41357625e+00 -3.43138784e-01
-4.79964346e-01 -5.43054938e-02 -7.63352886e-02 6.95127368e-01
7.32789755e-01 -4.19193894e-01 -1.41874626e-01 -7.21315205e-01
-1.26403883e-01 9.88533795e-02 6.16344452e-01 -1.13972855e+00
4.29145873e-01 -4.37944442e-01 2.51191229e-01 -6.36155784e-01
6.45964265e-01 -8.12204897e-01 -5.52185357e-01 9.85894382e-01
-4.73253399e-01 -1.36620536e-01 3.81260961e-01 7.64289916e-01
5.40280282e-01 -2.96901852e-01 8.14345479e-01 -3.33154261e-01
-1.19020402e+00 1.47927225e-01 -4.07567739e-01 9.39050913e-02
1.32571840e+00 -1.95715234e-01 -5.30348718e-01 -2.01294348e-01
-1.02419829e+00 8.59296262e-01 6.22591257e-01 8.32219839e-01
6.52653098e-01 -9.26444709e-01 -3.35579932e-01 -2.55955398e-01
4.75616097e-01 6.88827634e-01 -3.04551035e-01 7.82474756e-01
-7.23258376e-01 7.51300305e-02 -1.47121817e-01 -1.06062806e+00
-1.30212951e+00 6.62917078e-01 2.46800408e-01 6.37500525e-01
-8.11392725e-01 1.08218586e+00 1.33679003e-01 -6.38291538e-01
8.81575704e-01 -8.32035720e-01 2.58476734e-01 -4.96630371e-01
1.21686198e-01 5.68098128e-01 -3.90817642e-01 1.76521733e-01
-2.66746670e-01 6.30393744e-01 -1.34502068e-01 -7.55512118e-02
1.22985148e+00 -9.78511497e-02 1.37544751e-01 5.96790433e-01
8.80307257e-01 -4.39740747e-01 -2.08797550e+00 -1.98314205e-01
5.88012263e-02 -5.08874834e-01 -3.79799098e-01 -9.34600234e-01
-4.15826440e-01 6.26726031e-01 4.90744323e-01 -4.58244950e-01
3.66045833e-01 4.25512701e-01 5.36431432e-01 1.56814110e+00
7.61268616e-01 -1.47610068e+00 8.92186046e-01 8.71766269e-01
1.28226936e+00 -1.64678955e+00 6.60661682e-02 -5.10122001e-01
-4.48635995e-01 1.16531205e+00 1.32978249e+00 -2.32653797e-01
2.98906356e-01 5.82235575e-01 -1.40427172e-01 -2.55293995e-01
-7.29576826e-01 -2.35716507e-01 -8.38969946e-02 6.74039721e-01
-4.00942743e-01 -9.96963605e-02 4.01690573e-01 1.33948237e-01
-1.01104416e-01 5.87666072e-02 5.94181180e-01 1.77083635e+00
-1.14396727e+00 -3.16955417e-01 -1.91801473e-01 3.49733144e-01
1.85696259e-01 3.08116257e-01 -3.14541340e-01 9.36117947e-01
-1.85064465e-01 7.01230407e-01 -6.33931905e-02 -2.07843885e-01
6.39245927e-01 -2.76827663e-01 9.26388800e-01 -1.10149503e+00
4.32095043e-02 -6.94676712e-02 -7.81396255e-02 -9.67922509e-01
-2.05992728e-01 -7.71419287e-01 -1.80612874e+00 2.97528207e-01
-6.81119859e-01 -4.90974277e-01 8.62484038e-01 6.58634126e-01
3.85294199e-01 6.08477771e-01 2.26164505e-01 -1.67740333e+00
-9.85335052e-01 -1.17252171e+00 -1.82880312e-01 1.86871678e-01
6.56292737e-01 -1.27918053e+00 -1.94780171e-01 9.69680101e-02] | [4.749980449676514, 0.45206817984580994] |
be0d497f-2555-4166-86ed-75031cf22141 | agent-environment-network-for-temporal-action | 2107.08323 | null | https://arxiv.org/abs/2107.08323v3 | https://arxiv.org/pdf/2107.08323v3.pdf | Agent-Environment Network for Temporal Action Proposal Generation | Temporal action proposal generation is an essential and challenging task that aims at localizing temporal intervals containing human actions in untrimmed videos. Most of existing approaches are unable to follow the human cognitive process of understanding the video context due to lack of attention mechanism to express the concept of an action or an agent who performs the action or the interaction between the agent and the environment. Based on the action definition that a human, known as an agent, interacts with the environment and performs an action that affects the environment, we propose a contextual Agent-Environment Network. Our proposed contextual AEN involves (i) agent pathway, operating at a local level to tell about which humans/agents are acting and (ii) environment pathway operating at a global level to tell about how the agents interact with the environment. Comprehensive evaluations on 20-action THUMOS-14 and 200-action ActivityNet-1.3 datasets with different backbone networks, i.e C3D and SlowFast, show that our method robustly exhibits outperformance against state-of-the-art methods regardless of the employed backbone network. | ['Minh-Triet Tran', 'Akihiro Sugimoto', 'Kashu Yamazaki', 'Ngan Le', 'Viet-Khoa Vo-Ho'] | 2021-07-17 | null | null | null | null | ['temporal-action-proposal-generation'] | ['computer-vision'] | [ 4.28225696e-01 -2.59130057e-02 -6.35524914e-02 -7.53391609e-02
1.72422871e-01 -5.43119192e-01 1.12118435e+00 2.80519165e-02
-5.00849605e-01 8.10181797e-01 6.89587891e-01 1.38575304e-02
-8.39868113e-02 -6.95442021e-01 -6.09441757e-01 -7.87272990e-01
-3.25276583e-01 3.00381958e-01 6.19467318e-01 -1.64654538e-01
1.38249755e-01 2.90014714e-01 -1.46425247e+00 3.24208111e-01
2.68998016e-02 8.72179687e-01 2.62749344e-01 8.65052640e-01
3.00110340e-01 1.82937694e+00 -7.75888443e-01 2.32137486e-01
1.57143742e-01 -8.92325580e-01 -9.85836089e-01 3.12902957e-01
1.49272359e-03 -6.49736226e-01 -5.32551467e-01 9.28489029e-01
2.37237439e-01 3.46439540e-01 4.92354542e-01 -1.62349415e+00
-2.25497425e-01 6.59916461e-01 -4.09249179e-02 5.72305858e-01
8.26644301e-01 6.53610408e-01 6.38517082e-01 -3.19098681e-01
1.08323026e+00 1.39640880e+00 1.66466936e-01 5.46321154e-01
-7.16240108e-01 -3.84071380e-01 6.61012888e-01 7.43676364e-01
-1.13389444e+00 -3.07574630e-01 7.24561155e-01 -2.79481560e-01
9.34789598e-01 -3.41759110e-03 8.95729899e-01 1.55142963e+00
4.16096479e-01 7.11184621e-01 8.91144335e-01 -2.35112980e-02
6.29405081e-01 -3.34916323e-01 -1.39939830e-01 4.58796322e-01
-1.78338230e-01 2.52162665e-01 -8.66865873e-01 -7.33082592e-02
9.11574543e-01 -1.00564860e-01 -3.15317869e-01 1.13594308e-02
-1.86169076e+00 2.92141467e-01 2.42642090e-01 2.27603212e-01
-1.06429386e+00 6.20496452e-01 5.89861691e-01 2.12435797e-01
-7.60357827e-04 3.25711727e-01 -2.87259966e-01 -5.50420105e-01
-4.55233276e-01 4.29549694e-01 7.19482720e-01 7.91487277e-01
4.30344880e-01 2.03820258e-01 -3.60035211e-01 -2.28547111e-01
1.40046597e-01 1.97660819e-01 1.59306154e-01 -1.38620722e+00
2.08155036e-01 7.40022004e-01 5.09565234e-01 -9.97935593e-01
-3.00657243e-01 -1.61907718e-01 -5.43015242e-01 4.23291832e-01
2.60005653e-01 -2.98488498e-01 -7.86876678e-01 1.79486990e+00
7.38281608e-01 8.45126331e-01 3.60866010e-01 1.01247871e+00
7.22259223e-01 9.24057066e-01 4.62223768e-01 -3.56316626e-01
1.65552700e+00 -1.12650585e+00 -8.99058342e-01 -2.44295970e-01
3.11976343e-01 -3.90379578e-01 6.60578549e-01 3.59055668e-01
-1.11772823e+00 -7.67525494e-01 -1.00693262e+00 3.20334375e-01
-2.45658189e-01 -2.46976569e-01 3.85882974e-01 -1.20624810e-01
-9.77634788e-01 2.55919397e-01 -7.91222095e-01 -6.48264766e-01
1.01769619e-01 7.80959129e-02 -5.02249181e-01 1.07824437e-01
-1.27898276e+00 7.56327510e-01 6.19300306e-01 1.94249257e-01
-1.92378116e+00 -2.31257409e-01 -7.16836393e-01 -1.36162704e-02
8.53487313e-01 -7.39954174e-01 1.25170052e+00 -1.32344770e+00
-1.55438757e+00 2.76175588e-01 4.80025522e-02 -8.01117480e-01
6.07878029e-01 -1.59900814e-01 -5.38019359e-01 7.14449227e-01
-2.44082580e-03 8.18867445e-01 6.78180277e-01 -9.39198911e-01
-1.08676660e+00 -2.58951336e-01 7.47566819e-01 7.22177625e-01
8.87569860e-02 2.22938642e-01 -6.84283495e-01 -4.31044966e-01
-9.42741334e-02 -1.08505225e+00 -2.65959829e-01 4.27396446e-02
-3.21332574e-01 -4.59008604e-01 1.18567157e+00 -3.88427585e-01
1.12933946e+00 -1.98056400e+00 2.00937241e-01 -3.15077528e-02
2.49359027e-01 2.10987017e-01 -1.89357460e-01 6.96714997e-01
-6.91252053e-02 -2.03192726e-01 2.77834564e-01 -1.44279242e-01
3.28085124e-02 2.49201566e-01 -1.43327579e-01 5.70756733e-01
-1.17645428e-01 4.45202827e-01 -1.30695164e+00 -6.12623632e-01
3.09088558e-01 6.80917859e-01 -1.75188854e-01 4.44107413e-01
-5.08946121e-01 7.34836519e-01 -8.41802239e-01 5.01854420e-01
1.26193687e-01 -8.79000872e-02 4.19213682e-01 -1.89827800e-01
-1.26800984e-01 3.01250607e-01 -1.36622202e+00 1.60164440e+00
4.88987826e-02 4.93735343e-01 1.64130837e-01 -7.15692759e-01
3.29907805e-01 9.86213803e-01 5.46833754e-01 -6.43469572e-01
-1.29704461e-01 -2.77245343e-01 2.42868289e-01 -7.11490691e-01
2.61867225e-01 2.66192406e-01 -4.49146517e-02 7.44689226e-01
-4.92149927e-02 4.78037983e-01 5.13958216e-01 5.35668373e-01
1.67392802e+00 7.74894834e-01 5.89553475e-01 3.19779962e-02
6.35310233e-01 -1.44934189e-02 9.25035417e-01 7.74005055e-01
-7.12018430e-01 -1.38273582e-01 6.95207119e-01 -9.04168963e-01
-6.03172898e-01 -8.64179730e-01 7.13126361e-01 1.35128844e+00
5.89006603e-01 -4.92692977e-01 -8.13093364e-01 -7.84977019e-01
-6.30385160e-01 7.16432154e-01 -9.64929700e-01 -1.02930302e-02
-8.28643501e-01 -8.75556394e-02 4.14323837e-01 6.38726115e-01
1.07492042e+00 -1.82993019e+00 -1.39331686e+00 4.79574591e-01
-4.74905014e-01 -1.50245738e+00 -5.69262564e-01 -1.70626715e-01
-4.99643981e-01 -1.23518491e+00 5.00243343e-02 -4.00371432e-01
6.94116831e-01 6.89331740e-02 1.02207267e+00 1.16973951e-01
1.96198728e-02 5.72565436e-01 -4.09601331e-01 -2.87999928e-01
-4.96109486e-01 -5.02945185e-01 1.16308913e-01 2.19281048e-01
2.62619317e-01 -5.21193326e-01 -8.53529990e-01 6.15564525e-01
-9.02904928e-01 3.94411385e-01 3.19255292e-01 8.57425705e-02
4.79706526e-01 5.48240602e-01 3.10335249e-01 -5.64639032e-01
4.56700861e-01 -5.73518932e-01 -2.18497053e-01 2.39350140e-01
-9.21041891e-02 -8.94572586e-02 4.70276117e-01 -5.93798339e-01
-1.46253908e+00 1.72609299e-01 4.69497383e-01 -2.88194925e-01
-6.07000411e-01 2.99849778e-01 -3.50935429e-01 6.19286418e-01
5.46837270e-01 3.75316292e-01 -3.85289460e-01 1.62300035e-01
1.87291354e-01 1.14244290e-01 7.17136145e-01 -6.49709523e-01
5.57961702e-01 8.69668424e-01 1.81077465e-01 -5.73422730e-01
-3.38958919e-01 -3.39093447e-01 -4.73526239e-01 -1.01631129e+00
1.44980288e+00 -8.41273367e-01 -1.04842353e+00 5.65773726e-01
-1.37184441e+00 -5.88441432e-01 -1.00112900e-01 3.98324102e-01
-8.01440418e-01 2.43845973e-02 -5.39592087e-01 -7.28151202e-01
-1.28674388e-01 -1.15256619e+00 7.71206677e-01 4.45264168e-02
-3.80354255e-01 -8.82641196e-01 1.09771006e-01 3.54578882e-01
2.74329334e-01 5.33707678e-01 5.20216644e-01 -5.67567170e-01
-9.93739069e-01 -1.13090780e-02 1.47598892e-01 4.28731330e-02
3.21620047e-01 -1.16692518e-03 -5.06833792e-01 -2.75709722e-02
7.46146291e-02 -1.14011757e-01 1.84314221e-01 2.87469864e-01
8.79326820e-01 -4.59488124e-01 -4.20746565e-01 6.70006275e-02
9.80711222e-01 8.36425602e-01 1.01401722e+00 2.45861158e-01
4.00064021e-01 6.36159062e-01 8.54627669e-01 5.40266812e-01
4.88385737e-01 7.54940331e-01 7.68089056e-01 -2.83377897e-02
-2.06802398e-01 -2.39355758e-01 8.34433913e-01 9.19073895e-02
-6.03873849e-01 -6.03014648e-01 -6.71511710e-01 4.27849084e-01
-2.32126689e+00 -1.70278895e+00 1.77591354e-01 1.69448042e+00
4.93372232e-01 2.13366061e-01 2.17822984e-01 -1.60833091e-01
7.55882740e-01 4.40986723e-01 -8.16511035e-01 -6.65736496e-02
6.93040341e-02 -4.44945186e-01 -4.37683985e-02 2.37801969e-01
-9.73340750e-01 9.88651693e-01 5.68424320e+00 3.65750104e-01
-7.77645707e-01 1.84544027e-01 4.28101629e-01 -2.41592512e-01
3.18468183e-01 9.89022478e-02 -5.23929954e-01 4.41920519e-01
9.10938501e-01 -6.84166998e-02 5.57429910e-01 4.60848182e-01
8.44980419e-01 -4.11227494e-01 -1.55453420e+00 6.06377482e-01
1.33933881e-02 -1.14700389e+00 -4.16421471e-03 3.09462305e-02
4.89568293e-01 -1.97412968e-01 -3.41705918e-01 2.09440544e-01
4.48113054e-01 -8.10312748e-01 1.05780280e+00 8.07738543e-01
1.67920470e-01 -4.80787724e-01 5.39755583e-01 4.82826203e-01
-1.78328013e+00 -1.01734541e-01 3.28791559e-01 -3.09858024e-01
4.35122848e-01 -2.24931374e-01 -8.90302122e-01 2.61943936e-01
7.55778134e-01 6.96065366e-01 -1.42447308e-01 5.61634958e-01
-6.61795080e-01 5.94919622e-01 -4.04604673e-02 7.10130483e-02
5.89863062e-01 -9.27671939e-02 8.12427104e-01 9.68709528e-01
6.77909702e-02 4.98621345e-01 5.51957846e-01 7.19540298e-01
1.87098756e-01 -4.20111120e-01 -6.38688803e-01 -1.18691087e-01
4.79082763e-01 1.08888817e+00 -8.50349605e-01 -8.20239723e-01
-2.91838020e-01 9.03555214e-01 -1.46063820e-01 7.58828402e-01
-1.22405100e+00 6.67386800e-02 8.27244699e-01 1.76245332e-01
2.55500287e-01 -2.87883043e-01 6.99311972e-01 -6.74791694e-01
-1.77457929e-01 -1.11053383e+00 4.51441973e-01 -1.27791953e+00
-6.13816261e-01 7.25778043e-01 2.78714389e-01 -1.40144753e+00
-4.00144607e-01 -2.11095169e-01 -8.40230584e-01 2.99100369e-01
-9.70145643e-01 -1.07448721e+00 -6.37682080e-01 9.29894030e-01
8.91234815e-01 -6.76002875e-02 5.22556603e-01 2.42935377e-03
-4.39057976e-01 -3.71312201e-01 -9.31978106e-01 4.12543900e-02
4.34832901e-01 -8.99833024e-01 8.95647891e-03 1.12230563e+00
-4.89663100e-03 4.34877932e-01 9.35365617e-01 -9.12448227e-01
-1.38725686e+00 -1.12508750e+00 5.73967457e-01 -3.19004446e-01
6.17206156e-01 -1.48604244e-01 -5.30051887e-01 1.01412868e+00
7.51404345e-01 1.14140034e-01 3.92024100e-01 -5.83419383e-01
7.28516513e-03 -7.72850960e-02 -1.13121676e+00 9.21392381e-01
1.54376411e+00 -4.25942481e-01 -6.57431960e-01 3.61346960e-01
8.43068182e-01 -2.35339612e-01 -5.50800264e-01 1.23648524e-01
4.71463412e-01 -1.15678048e+00 9.22937870e-01 -7.65706062e-01
4.31709021e-01 -8.47987235e-01 -1.99831411e-01 -1.03616154e+00
-2.79106945e-01 -7.92077661e-01 -3.95141900e-01 8.75142872e-01
-3.22971530e-02 -4.68936831e-01 6.12728775e-01 5.33722222e-01
-1.63413763e-01 -2.89393365e-01 -8.55922937e-01 -5.47839224e-01
-8.28734100e-01 -4.68774259e-01 7.34421074e-01 8.05168390e-01
1.30641041e-02 2.83393443e-01 -5.57148993e-01 3.89023840e-01
4.81605947e-01 -3.72762501e-01 9.01514947e-01 -7.33945847e-01
-2.89149731e-01 -1.20205484e-01 -5.72593629e-01 -9.42101657e-01
2.16776460e-01 -2.06217617e-01 2.19256684e-01 -1.76397336e+00
2.38850862e-01 1.37283787e-01 -3.96564603e-01 6.17673278e-01
1.49478137e-01 -1.08487360e-01 2.78836012e-01 2.07692683e-01
-1.12387979e+00 4.42513913e-01 1.32913244e+00 -7.93015957e-03
-1.71345919e-01 -2.55358577e-01 -2.03908876e-01 9.87719238e-01
5.25807023e-01 -2.96407282e-01 -9.26527441e-01 -2.79704899e-01
2.24266231e-01 4.81185466e-01 6.57650888e-01 -1.38660431e+00
6.07347190e-01 -5.57682037e-01 6.02482148e-02 -5.31660676e-01
5.14954805e-01 -1.05301309e+00 6.13504648e-01 6.78899825e-01
-4.28804427e-01 4.26399022e-01 -1.03870928e-01 8.59338701e-01
-2.54265279e-01 1.37008876e-01 3.03615808e-01 -4.75489765e-01
-1.34698999e+00 3.82870853e-01 -8.24614346e-01 -2.45464027e-01
1.50686848e+00 -3.07572812e-01 -6.48613334e-01 -6.86944723e-01
-7.91680574e-01 3.33942920e-01 1.83376744e-01 3.47862482e-01
7.15072572e-01 -1.29580390e+00 -6.34287953e-01 -2.27023780e-01
6.19787257e-03 -3.87457669e-01 3.47983658e-01 8.56600046e-01
-5.03382325e-01 2.98507124e-01 -3.35106969e-01 -4.23566341e-01
-1.42264307e+00 4.93605047e-01 5.51057398e-01 -2.77204752e-01
-6.53710186e-01 5.05543172e-01 5.47251940e-01 4.76052649e-02
3.74739915e-01 -2.17169434e-01 -4.48433310e-01 -6.18488602e-02
8.53051484e-01 6.40534401e-01 -5.51225007e-01 -9.72091317e-01
-4.35541779e-01 1.35720372e-01 1.80929422e-01 -2.27405205e-01
1.13030386e+00 -1.40709981e-01 -1.53983280e-01 2.92690575e-01
6.73927426e-01 -6.93861067e-01 -1.74785960e+00 -2.36311644e-01
-3.57798487e-01 -2.21660867e-01 2.21005920e-03 -9.88304019e-01
-8.73607516e-01 5.05815208e-01 5.63244462e-01 2.35320315e-01
1.27462614e+00 -2.38508917e-02 5.31891644e-01 3.43124598e-01
7.32588470e-01 -1.37610662e+00 6.75177276e-01 3.77097487e-01
1.16792631e+00 -8.67001355e-01 4.38519605e-02 -1.00044064e-01
-8.66191030e-01 1.01047254e+00 9.74690139e-01 1.79360881e-01
4.68893468e-01 1.03067823e-01 -5.67852743e-02 -5.75913906e-01
-1.32433331e+00 3.54792748e-04 -5.57031259e-02 5.16792476e-01
-6.18725382e-02 7.38658535e-04 4.00865301e-02 1.55191347e-01
3.67369711e-01 1.99831322e-01 6.46401227e-01 1.04093766e+00
-2.91230828e-01 -6.22258067e-01 -4.10177559e-01 -1.14120282e-01
-1.29956678e-01 3.22292358e-01 -4.61141557e-01 9.49412048e-01
6.14661515e-01 1.23695219e+00 3.62019926e-01 -2.05286697e-01
3.17389041e-01 -9.34293047e-02 3.03763628e-01 -5.89216709e-01
-6.55670583e-01 7.02930614e-02 3.90888363e-01 -1.14972270e+00
-1.05479360e+00 -8.75012696e-01 -1.61154461e+00 -2.76442945e-01
3.25681180e-01 -4.43184003e-02 4.33168620e-01 1.12782764e+00
3.08529258e-01 7.58554041e-01 2.61675268e-01 -9.06247199e-01
2.53583398e-02 -9.77075815e-01 -3.44490498e-01 5.48485339e-01
2.83892006e-01 -6.09967887e-01 -3.79493296e-01 4.05444950e-01] | [8.452577590942383, 0.6816368699073792] |
15bd42ee-d84c-4825-a56e-681447ebb4d3 | kpi-edgar-a-novel-dataset-and-accompanying | 2210.09163 | null | https://arxiv.org/abs/2210.09163v1 | https://arxiv.org/pdf/2210.09163v1.pdf | KPI-EDGAR: A Novel Dataset and Accompanying Metric for Relation Extraction from Financial Documents | We introduce KPI-EDGAR, a novel dataset for Joint Named Entity Recognition and Relation Extraction building on financial reports uploaded to the Electronic Data Gathering, Analysis, and Retrieval (EDGAR) system, where the main objective is to extract Key Performance Indicators (KPIs) from financial documents and link them to their numerical values and other attributes. We further provide four accompanying baselines for benchmarking potential future research. Additionally, we propose a new way of measuring the success of said extraction process by incorporating a word-level weighting scheme into the conventional F1 score to better model the inherently fuzzy borders of the entity pairs of a relation in this domain. | ['Rafet Sifa', 'Christian Bauckhage', 'Basil Jacob', 'Desiana Nurchalifah', 'Lars Hillebrand', 'Syed Musharraf Ali', 'Tobias Deußer'] | 2022-10-17 | null | null | null | null | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [-4.47015494e-01 3.34258735e-01 -4.88952398e-01 -3.75732064e-01
-8.62697244e-01 -9.18644547e-01 7.06013620e-01 8.77359748e-01
-3.12547833e-01 7.78948784e-01 6.85610473e-01 -4.80547309e-01
-5.40013254e-01 -1.10381544e+00 -1.62282318e-01 -5.00402376e-02
-2.32214943e-01 6.35217845e-01 -3.04081086e-02 -4.14367765e-02
5.81061721e-01 6.11058116e-01 -8.49868536e-01 4.47767735e-01
5.24976909e-01 1.43194342e+00 -6.28510773e-01 4.78495836e-01
-2.91450560e-01 1.27876401e+00 -9.37735200e-01 -1.03286600e+00
1.95765510e-01 2.50896215e-01 -1.24412358e+00 -2.31553450e-01
-7.57600069e-02 -1.08755283e-01 -3.85998100e-01 8.06188703e-01
1.31943166e-01 -5.11266068e-02 9.72184837e-01 -1.14519680e+00
-9.31380510e-01 8.98797393e-01 -2.60509402e-01 4.55180317e-01
5.90885937e-01 -3.64180833e-01 1.58327568e+00 -9.89985824e-01
9.22105014e-01 7.18862653e-01 6.58621192e-01 -3.28632385e-01
-7.78461099e-01 -5.78843057e-01 -1.77303016e-01 3.89353663e-01
-1.50757694e+00 -4.35103744e-01 2.94336200e-01 -5.11105001e-01
1.56267071e+00 2.25573882e-01 2.92140573e-01 3.12707007e-01
-1.70069262e-01 6.50604010e-01 8.50832462e-01 -4.06223863e-01
9.20530260e-02 2.50444710e-01 6.29029393e-01 9.92921144e-02
7.30645955e-01 -1.91228941e-01 -5.32854140e-01 -3.57032180e-01
5.42276740e-01 -2.45470762e-01 -2.91812941e-02 1.97815672e-01
-1.19226921e+00 6.25410616e-01 1.06001839e-01 5.84427714e-01
-7.43797123e-01 -4.97186899e-01 3.31747830e-01 4.31476563e-01
4.12901878e-01 1.15552950e+00 -1.02152085e+00 -1.77910551e-01
-9.63650763e-01 4.00375605e-01 1.29799247e+00 1.03489208e+00
5.15679181e-01 -5.79739690e-01 -4.80502456e-01 4.22892720e-01
5.50597794e-02 2.92896237e-02 2.80604303e-01 -7.92272806e-01
1.01982296e+00 1.22392797e+00 4.69256163e-01 -1.15424025e+00
-5.23299277e-01 -3.99401367e-01 -3.52780014e-01 -3.46389890e-01
3.84445876e-01 -1.65325165e-01 -4.78848249e-01 8.78696263e-01
1.40481442e-01 -6.16356507e-02 6.65217996e-01 4.01318282e-01
1.32096541e+00 4.45185125e-01 3.57083641e-02 -3.69573861e-01
1.47009599e+00 -4.83065337e-01 -9.85652328e-01 1.50659814e-01
9.92210090e-01 -7.74906099e-01 1.97186217e-01 2.25548550e-01
-1.01306999e+00 -1.19618084e-02 -8.67193520e-01 8.34316686e-02
-8.68226469e-01 1.54685825e-01 9.09969509e-01 5.13583839e-01
-7.34650910e-01 6.43934429e-01 -4.93644536e-01 1.85277894e-01
3.81874561e-01 2.11904541e-01 -5.74234843e-01 2.52592951e-01
-1.44852245e+00 1.18342304e+00 6.83696449e-01 -2.92763740e-01
1.55613169e-01 -9.31041420e-01 -8.88207018e-01 1.95631310e-01
4.87257361e-01 -3.23625743e-01 9.75839794e-01 2.28078336e-01
-8.88081908e-01 9.24649060e-01 6.81758076e-02 -6.47226572e-01
-6.41863048e-02 -2.42044464e-01 -9.53160942e-01 7.44365603e-02
1.47459507e-01 -4.91280071e-02 -1.45382136e-01 -7.19042480e-01
-9.34063554e-01 -3.60264540e-01 -8.41573030e-02 -2.70064455e-02
-5.70552528e-01 7.46737182e-01 -3.04444045e-01 -8.30459714e-01
9.87801030e-02 -2.72478580e-01 -5.91415055e-02 -1.04497540e+00
-3.56184155e-01 -6.46742284e-01 2.90719062e-01 -1.19657445e+00
1.94822407e+00 -1.73100281e+00 -1.85342163e-01 5.94690442e-01
4.70149964e-01 3.38779777e-01 3.07650238e-01 6.00619733e-01
-3.62251818e-01 5.85231781e-01 2.17919350e-01 2.15825081e-01
-1.28060868e-02 -1.51213050e-01 -2.99702585e-01 -2.24674270e-02
6.12046659e-01 1.17547119e+00 -7.35390127e-01 -5.83839834e-01
-2.98202574e-01 1.43313199e-01 -1.13197602e-01 7.16569647e-02
3.06610644e-01 -9.26937163e-02 -3.63154471e-01 1.00855196e+00
4.84150708e-01 -1.19227365e-01 2.06204712e-01 -4.30385202e-01
-1.90482065e-01 1.00498402e+00 -1.64446568e+00 7.31923342e-01
-1.70375884e-01 3.87736142e-01 -4.24862683e-01 -1.16142011e+00
1.24463940e+00 5.88458478e-01 6.63668811e-01 -5.85855305e-01
-2.11931989e-02 2.72194952e-01 -3.82449031e-01 -3.65854561e-01
8.64558458e-01 1.95204750e-01 -4.89199579e-01 3.36411685e-01
5.70541099e-02 3.93190421e-02 6.94568932e-01 4.00366813e-01
1.53544652e+00 -3.28707933e-01 7.12434947e-01 2.70560179e-02
4.16884094e-01 1.91693291e-01 7.97910333e-01 3.57451141e-01
3.08918338e-02 3.66163522e-01 7.83741415e-01 -3.29147637e-01
-9.32363570e-01 -6.75968051e-01 -3.91278803e-01 6.35639071e-01
-6.10608220e-01 -7.56600857e-01 -2.43873343e-01 -7.89375126e-01
3.90845776e-01 6.96538806e-01 -4.77924138e-01 7.35156834e-02
-4.27830935e-01 -9.37763631e-01 7.95828998e-01 7.74442852e-01
4.00639266e-01 -8.26142550e-01 -3.36826056e-01 3.67494673e-01
-1.81585088e-01 -1.34635055e+00 2.76677497e-02 3.58483642e-01
-6.23155177e-01 -1.29358840e+00 -5.50073087e-01 -6.77538872e-01
2.76122153e-01 -5.14879882e-01 1.63663673e+00 -2.23632142e-01
1.97845697e-01 4.66076657e-02 -5.31095803e-01 -3.28287005e-01
-2.07983777e-01 3.12120110e-01 -1.59466103e-01 -3.84220243e-01
8.25474381e-01 -1.51732519e-01 -2.33668715e-01 2.86915839e-01
-7.47104943e-01 -4.46304947e-01 5.41319489e-01 4.09079850e-01
2.77449012e-01 4.50430840e-01 8.62721920e-01 -8.80645931e-01
9.94418323e-01 -6.40039921e-01 -4.08769041e-01 7.03782737e-01
-1.07430327e+00 3.16081531e-02 7.93672279e-02 1.88896805e-02
-9.66784656e-01 -2.76685953e-01 9.38387141e-02 3.46624523e-01
-1.07082121e-01 1.28956199e+00 -2.31138274e-01 3.15230250e-01
3.03291440e-01 -3.55950683e-01 -6.11886501e-01 -6.03267193e-01
4.34139580e-01 1.04120338e+00 9.47719991e-01 -4.26010668e-01
6.39273882e-01 -1.07558863e-02 -1.89360417e-02 -2.37050995e-01
-1.05824542e+00 -8.65542471e-01 -8.62457752e-01 2.43303195e-01
3.67770314e-01 -9.11052823e-01 -7.33905971e-01 4.12349612e-01
-1.00631666e+00 3.54212165e-01 -4.66867089e-01 5.53492546e-01
2.88838949e-02 3.24417725e-02 -8.93649757e-01 -7.44742930e-01
-4.38346922e-01 -3.75228941e-01 4.72299993e-01 3.51612538e-01
-7.07485974e-01 -1.09534180e+00 9.60792229e-02 5.44504046e-01
2.15540349e-01 3.88052195e-01 8.72562408e-01 -1.41356444e+00
9.42586958e-02 -7.58535683e-01 -7.10663199e-01 2.84495145e-01
2.13632628e-01 2.35764682e-01 -3.58835787e-01 3.31954628e-01
-5.46309762e-02 -6.80594286e-03 8.47510636e-01 -9.14139077e-02
4.90001708e-01 -6.64751947e-01 -2.98631161e-01 4.15505052e-01
1.18989706e+00 3.84534478e-01 7.07063138e-01 9.29671288e-01
5.19370139e-01 8.00220132e-01 6.56947315e-01 5.63705981e-01
9.09889162e-01 5.58227956e-01 -2.82637805e-01 9.45417657e-02
1.32108822e-01 1.42162681e-01 8.90322104e-02 9.06921029e-01
-3.58445972e-01 1.23216193e-02 -1.45558047e+00 8.43431652e-01
-1.60006475e+00 -8.96493018e-01 -3.74222398e-01 1.96390581e+00
1.25114381e+00 4.53532100e-01 8.74892846e-02 5.64039290e-01
5.29188991e-01 -2.84935325e-01 -3.79215069e-02 -2.31972024e-01
-5.10771215e-01 6.06102824e-01 1.00564110e+00 1.19001269e-01
-1.21939290e+00 6.48983002e-01 6.91409779e+00 3.97820801e-01
-4.77241337e-01 -3.12440783e-01 7.83824503e-01 1.78684056e-01
-1.06177337e-01 1.58703193e-01 -1.22591352e+00 2.13939741e-01
1.41529119e+00 -6.65532827e-01 -1.31986171e-01 5.40071726e-01
-1.40934199e-01 -1.27263188e-01 -1.02232516e+00 4.17115033e-01
-3.15180540e-01 -1.61176288e+00 -1.05623566e-01 1.58178747e-01
7.46536016e-01 -2.83597946e-01 -1.04245864e-01 1.67457461e-01
6.86275959e-01 -9.43312049e-01 5.19034564e-01 9.08665955e-01
5.69732189e-01 -9.88461792e-01 1.31630576e+00 -1.88767910e-01
-1.20871878e+00 -2.48799860e-01 -3.47995050e-02 -1.79570615e-01
2.90536508e-03 7.60181665e-01 -9.88809347e-01 1.11687434e+00
6.34269476e-01 7.82012343e-01 -6.98040307e-01 1.04005873e+00
-1.31049350e-01 8.22564900e-01 -2.25730687e-01 2.14548647e-01
-3.68155204e-02 -3.77845317e-02 1.91332176e-01 1.50905752e+00
2.98871219e-01 3.76200378e-01 -2.80597746e-01 5.96865237e-01
-3.69820714e-01 2.77518749e-01 -1.76058352e-01 -5.25765777e-01
7.50593901e-01 1.16432035e+00 -7.48480499e-01 -7.91720569e-01
-5.49297214e-01 2.29890198e-01 2.77248263e-01 2.49643892e-01
-2.08646551e-01 -9.68024075e-01 4.04481173e-01 -1.48137240e-02
4.61669207e-01 -3.01024646e-01 -7.75790274e-01 -1.17977118e+00
1.38434023e-01 -6.75803661e-01 9.50674236e-01 -4.28398758e-01
-1.31734145e+00 3.21834654e-01 5.29113188e-02 -1.01017749e+00
-4.85519081e-01 -4.65777546e-01 -3.70893776e-01 9.65581834e-01
-1.37390208e+00 -8.80534291e-01 1.07183032e-01 2.02685952e-01
-3.50723326e-01 -3.25164646e-01 9.08150256e-01 5.74716032e-01
-8.35873127e-01 5.81181169e-01 1.57298744e-01 9.88698006e-01
8.29061508e-01 -1.45209289e+00 8.02297175e-01 6.95220172e-01
2.87615061e-01 6.36672616e-01 5.36904812e-01 -9.21425641e-01
-7.93522179e-01 -9.23469484e-01 1.80872679e+00 -8.81884038e-01
1.10152996e+00 1.29414216e-01 -1.24065363e+00 5.99676788e-01
-5.15632816e-02 -1.68686599e-01 8.00358474e-01 3.80816638e-01
-4.32667017e-01 -1.28047958e-01 -1.22270155e+00 -1.29687697e-01
6.35372281e-01 -4.49528098e-01 -1.17487991e+00 -2.47116163e-02
6.29992604e-01 -2.18194187e-01 -1.80779731e+00 5.28567255e-01
2.52575994e-01 -3.72157604e-01 1.03815448e+00 -8.62717330e-01
4.57518995e-01 -1.95219114e-01 -1.31233945e-01 -1.11930764e+00
-3.72143060e-01 -5.16969860e-01 -5.57103693e-01 2.04303074e+00
7.90625989e-01 -6.39513016e-01 5.87375879e-01 1.13432479e+00
3.56315196e-01 -5.84969223e-01 -8.15220654e-01 -7.81335592e-01
9.38289464e-02 -4.81108248e-01 1.23794472e+00 1.28163743e+00
6.26195073e-01 3.25277835e-01 9.62919518e-02 1.41267449e-01
3.76553506e-01 2.18577147e-01 3.58311534e-01 -1.36337912e+00
6.83429390e-02 -5.05852461e-01 -6.79724574e-01 -3.06113154e-01
-5.26359119e-02 -7.72090137e-01 -6.14645243e-01 -1.80434930e+00
9.46816802e-02 -4.61054891e-01 -7.94222236e-01 5.75167537e-01
-2.19112188e-01 7.97848478e-02 -1.49492091e-02 5.79298735e-01
-6.53042138e-01 -6.33411705e-02 5.35658538e-01 -1.87713299e-02
-4.32353497e-01 8.95553753e-02 -1.14455080e+00 6.50335431e-01
5.45517921e-01 -5.48509121e-01 3.79350126e-01 4.99253944e-02
7.44238913e-01 2.10350707e-01 -7.53237754e-02 -7.67841816e-01
3.97398651e-01 -3.74472708e-01 3.43369156e-01 -8.18334818e-01
-4.03680354e-01 -5.06630182e-01 -5.77589544e-03 -9.53833386e-03
-4.41640735e-01 3.23443472e-01 -1.64801031e-01 -3.09558101e-02
-6.42411709e-01 -2.59753883e-01 8.40737447e-02 -7.91181158e-03
-5.78080356e-01 -1.24783128e-01 -1.44505128e-01 2.47837037e-01
9.12148178e-01 2.29568824e-01 -5.74473381e-01 -3.01197171e-02
-6.49101198e-01 2.90692747e-01 1.43951029e-01 2.48543277e-01
4.56598192e-01 -1.37157071e+00 -1.06137645e+00 -1.59636617e-01
3.05053651e-01 -8.91714022e-02 -4.99039859e-01 6.27127290e-01
-3.89192700e-01 8.67549300e-01 3.54781710e-02 3.89657438e-01
-1.10774541e+00 3.72025818e-01 -8.16015452e-02 -1.10655105e+00
-3.03255230e-01 6.50028229e-01 -5.68144202e-01 -1.96137547e-01
7.63249993e-02 -6.30225182e-01 -9.21815097e-01 6.76170111e-01
6.29011214e-01 7.01192796e-01 6.90453708e-01 -7.24312246e-01
-6.05196655e-01 -7.96822011e-02 -2.11804390e-01 -1.24646254e-01
1.75365829e+00 -7.32267126e-02 -3.49744618e-01 2.15434507e-01
8.55011761e-01 1.30487129e-01 -4.63532776e-01 -6.61871552e-01
1.24992681e+00 -2.45469719e-01 9.36792418e-02 -9.35524762e-01
-9.41962659e-01 9.01872292e-02 -1.89016044e-01 5.14147639e-01
8.59496236e-01 2.85087585e-01 8.71186554e-01 5.17473578e-01
-5.82547896e-02 -1.40899765e+00 -3.62901688e-01 7.35685766e-01
6.58868730e-01 -9.27708030e-01 3.00874203e-01 -3.46761465e-01
-7.01332033e-01 1.18135524e+00 -2.61955485e-02 8.75293612e-02
7.71520078e-01 6.32692277e-01 2.03311250e-01 -4.20586705e-01
-7.42986917e-01 -2.22326443e-01 7.50982225e-01 5.61274529e-01
7.04273105e-01 2.11962029e-01 -5.03081083e-01 1.30344367e+00
-5.18461466e-01 6.80072457e-02 4.87964004e-01 1.02823365e+00
-2.02499792e-01 -1.03738332e+00 -4.19614732e-01 8.24136555e-01
-1.24330807e+00 -3.00191164e-01 -7.21794963e-01 5.34332395e-01
-1.55816719e-01 1.08236599e+00 1.17860325e-01 -4.33468431e-01
7.86538184e-01 2.30096653e-01 -9.49499682e-02 -6.32206917e-01
-9.39523220e-01 -3.83816421e-01 6.32448375e-01 -1.82507947e-01
-4.12104309e-01 -1.00503755e+00 -1.31558478e+00 -2.34165385e-01
-3.34825426e-01 4.63210702e-01 5.10029495e-01 1.02803719e+00
4.06363785e-01 5.35200119e-01 6.12529635e-01 9.35858041e-02
-3.24647278e-01 -1.04198182e+00 -6.21033192e-01 5.88517487e-01
-1.65403858e-02 -5.29552221e-01 -2.01482996e-01 1.70994475e-01] | [9.357437133789062, 8.704753875732422] |
d5d53b06-917a-48c6-9a60-3e5c533776b9 | meta-learning-for-low-resource-unsupervised | 2010.09046 | null | https://arxiv.org/abs/2010.09046v2 | https://arxiv.org/pdf/2010.09046v2.pdf | Unsupervised Neural Machine Translation for Low-Resource Domains via Meta-Learning | Unsupervised machine translation, which utilizes unpaired monolingual corpora as training data, has achieved comparable performance against supervised machine translation. However, it still suffers from data-scarce domains. To address this issue, this paper presents a novel meta-learning algorithm for unsupervised neural machine translation (UNMT) that trains the model to adapt to another domain by utilizing only a small amount of training data. We assume that domain-general knowledge is a significant factor in handling data-scarce domains. Hence, we extend the meta-learning algorithm, which utilizes knowledge learned from high-resource domains, to boost the performance of low-resource UNMT. Our model surpasses a transfer learning-based approach by up to 2-4 BLEU scores. Extensive experimental results show that our proposed algorithm is pertinent for fast adaptation and consistently outperforms other baseline models. | ['Cheonbok Park', 'Jaegul Choo', 'Eunjeong Park', 'Mohammad Azam Khan', 'Soyoung Yang', 'Taehee Kim', 'Yunwon Tae'] | 2020-10-18 | null | https://aclanthology.org/2021.acl-long.225 | https://aclanthology.org/2021.acl-long.225.pdf | acl-2021-5 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 3.92911553e-01 -1.13362037e-01 -8.74387264e-01 -5.23470759e-01
-1.36203289e+00 -6.56033397e-01 7.81086504e-01 -1.96468577e-01
-5.61706185e-01 1.41471469e+00 8.17484856e-02 -6.31644666e-01
4.54384893e-01 -4.26196963e-01 -9.99521077e-01 -5.00835061e-01
6.73219979e-01 9.04732585e-01 9.24048796e-02 -3.52622628e-01
6.30990118e-02 -2.55793594e-02 -8.48478436e-01 5.07800758e-01
1.46459246e+00 2.65109003e-01 4.20137346e-01 2.07269285e-02
-3.11088890e-01 3.97723377e-01 -3.02312672e-01 -5.85281909e-01
1.80336028e-01 -8.69375229e-01 -9.40723419e-01 -1.45703211e-01
2.71990627e-01 -3.48280579e-01 -3.55356485e-02 1.02407455e+00
5.35377562e-01 -6.61892295e-02 7.94293344e-01 -7.32755780e-01
-1.02541566e+00 6.82586670e-01 -3.19156408e-01 2.47426257e-01
4.72685844e-02 -1.30467027e-01 7.03855813e-01 -1.33259130e+00
8.05855274e-01 8.34886372e-01 5.01960337e-01 7.41587520e-01
-1.13755357e+00 -7.21067131e-01 -9.78278220e-02 1.90283850e-01
-1.09418845e+00 -6.59529865e-01 7.38664329e-01 -1.69452831e-01
1.21395350e+00 -2.67650545e-01 3.93908806e-02 1.30453920e+00
1.76790074e-01 8.28733563e-01 1.55550969e+00 -9.91702139e-01
1.60404429e-01 3.93968582e-01 -3.46635789e-01 3.94097954e-01
1.89236224e-01 1.94165260e-01 -6.13256931e-01 -1.15850531e-01
7.64385760e-01 -3.42601359e-01 7.40163699e-02 -2.47370973e-01
-1.48798394e+00 7.79665470e-01 3.97681445e-02 5.29296696e-01
-4.11669105e-01 -3.97013724e-01 4.15900081e-01 8.14144671e-01
9.00388837e-01 4.16732520e-01 -9.46874738e-01 -3.64592046e-01
-8.56046498e-01 -2.77901560e-01 6.41561687e-01 1.29314208e+00
9.33131278e-01 1.67329028e-01 1.32401660e-01 1.23815715e+00
-1.60219207e-01 7.33089566e-01 8.81821215e-01 -4.86833066e-01
9.22223926e-01 5.55421710e-01 1.87774777e-01 -2.33521596e-01
2.22461186e-02 -3.70442361e-01 -6.90702140e-01 -2.27654323e-01
2.39241377e-01 -4.14677531e-01 -9.70622540e-01 1.76237929e+00
2.75963634e-01 -7.49584064e-02 4.69566524e-01 7.47202039e-01
3.40314150e-01 7.31924951e-01 2.50869505e-02 -7.21037090e-01
8.31489742e-01 -1.38661611e+00 -7.52205491e-01 -3.59686673e-01
7.80724287e-01 -1.02460313e+00 1.20903301e+00 1.53048694e-01
-1.03188002e+00 -5.68514705e-01 -9.66857374e-01 5.01770526e-02
-3.47738117e-01 2.63382584e-01 5.38899601e-01 5.88749230e-01
-9.17561114e-01 5.03923595e-01 -7.46626616e-01 -8.54012609e-01
1.65199474e-01 6.12012327e-01 -3.89856786e-01 -1.98463961e-01
-1.30929494e+00 1.14081120e+00 6.34494305e-01 -2.55527347e-01
-6.38037980e-01 -3.26554358e-01 -5.86366475e-01 -3.97828162e-01
2.24483088e-01 -6.62946522e-01 1.45494843e+00 -1.61555481e+00
-2.08818340e+00 7.03134060e-01 -2.78251827e-01 -3.46758217e-01
4.63791698e-01 -3.42805952e-01 -5.44018507e-01 3.26255783e-02
7.03276917e-02 4.58258897e-01 8.40736210e-01 -9.91385043e-01
-5.96564889e-01 -1.85497582e-01 -3.17119181e-01 5.09552836e-01
-7.49998212e-01 3.23692322e-01 -4.35363948e-01 -8.04281533e-01
-1.52027264e-01 -1.05679381e+00 -1.05489269e-01 -7.55664945e-01
1.41501874e-01 -1.88422605e-01 5.28481245e-01 -7.97255456e-01
1.04969335e+00 -1.70392931e+00 3.29052031e-01 -2.51048714e-01
-4.14080560e-01 5.37627876e-01 -4.54932898e-01 6.74860835e-01
2.63195693e-01 -1.41334504e-01 -4.34118629e-01 -9.93862823e-02
-3.57295662e-01 3.91954690e-01 -2.99323112e-01 1.26496673e-01
3.00767958e-01 1.09019220e+00 -1.02745199e+00 -5.01091003e-01
-1.67572036e-01 1.08767599e-01 -3.86598945e-01 2.65353382e-01
-3.06634843e-01 8.80162895e-01 -5.22840798e-01 7.99989402e-01
5.98890901e-01 -1.88384950e-01 6.46292388e-01 1.57994792e-01
-1.36332482e-01 4.86006647e-01 -2.77086318e-01 2.17726660e+00
-6.41188800e-01 4.21768099e-01 -4.66285169e-01 -9.92964983e-01
1.17784631e+00 5.01140296e-01 3.42283189e-01 -8.99990261e-01
1.87892765e-02 8.48476291e-01 -2.80446149e-02 -4.74296033e-01
3.62867415e-01 -3.15911591e-01 -6.89562634e-02 6.29679799e-01
3.41901124e-01 2.17040613e-01 2.57557910e-02 -2.36317933e-01
9.44593906e-01 6.27922595e-01 3.72389734e-01 -2.08651975e-01
4.47872072e-01 5.10645747e-01 7.73169935e-01 4.62338090e-01
-2.75779486e-01 3.17898571e-01 -2.33746260e-01 -4.49990422e-01
-1.31861091e+00 -1.03924549e+00 5.71724102e-02 1.42851126e+00
2.08012182e-02 -1.82961971e-02 -8.36219072e-01 -1.15318346e+00
-2.39793032e-01 6.24106050e-01 -2.13508904e-01 -1.79245204e-01
-9.68392491e-01 -9.64894772e-01 4.70171988e-01 5.74113190e-01
5.92965722e-01 -9.65759277e-01 8.87908563e-02 4.87920821e-01
-6.33332014e-01 -1.16053140e+00 -5.24316430e-01 1.83040142e-01
-1.51623321e+00 -4.13978755e-01 -9.33353066e-01 -1.29345834e+00
7.84121394e-01 3.15219790e-01 1.13347983e+00 -4.75254059e-01
6.10237598e-01 -1.38709009e-01 -5.76196134e-01 -2.40860075e-01
-8.24670672e-01 7.24369466e-01 5.89858353e-01 -1.42785400e-01
9.07621384e-01 -7.36264467e-01 -2.13316560e-01 3.80724579e-01
-6.24341488e-01 2.04882964e-01 1.19215596e+00 1.15467668e+00
5.91575921e-01 -3.92194241e-01 1.20616782e+00 -1.08954692e+00
7.05349505e-01 -5.57799041e-01 -4.88861501e-01 5.21885574e-01
-9.89314735e-01 1.30047068e-01 9.62329149e-01 -8.59923959e-01
-1.37045109e+00 8.57575312e-02 2.93078005e-01 -1.75691903e-01
8.61335360e-03 7.65049696e-01 -1.49907708e-01 -9.56893563e-02
8.30116808e-01 5.21420002e-01 -1.30310968e-01 -8.23898613e-01
3.86024147e-01 1.22086632e+00 4.05018955e-01 -8.97875309e-01
9.47315037e-01 -2.54619457e-02 -5.38900614e-01 -3.47908407e-01
-6.57319605e-01 -3.41634601e-01 -1.11398149e+00 1.44487411e-01
3.95543814e-01 -1.13679683e+00 3.68873239e-01 2.91174173e-01
-1.11581445e+00 -4.84859139e-01 8.90713632e-02 8.72274697e-01
-7.56152570e-01 1.95819497e-01 -7.31321216e-01 -4.04996514e-01
-5.28551459e-01 -8.73245895e-01 6.78341210e-01 -7.39804422e-03
4.91193775e-03 -1.08177876e+00 5.64447284e-01 5.31804860e-01
4.16488737e-01 -1.84608445e-01 1.04647923e+00 -9.76768017e-01
-4.08030778e-01 9.22667384e-02 -8.85271206e-02 5.26442170e-01
5.08163035e-01 -5.25110722e-01 -5.98287940e-01 -5.40482640e-01
-4.37354222e-02 -6.73571289e-01 4.56202447e-01 -8.09017047e-02
4.86557871e-01 -3.45045507e-01 -2.27263048e-01 4.30989414e-01
1.40496016e+00 2.45937005e-01 3.69184077e-01 6.61283672e-01
5.03214538e-01 3.52876365e-01 9.30938005e-01 -4.83891517e-02
4.58052009e-01 6.85043037e-01 -2.98182249e-01 -1.69486478e-01
-1.12087049e-01 -4.81178194e-01 8.54109704e-01 1.80341375e+00
-2.71598369e-01 2.53242292e-02 -1.04103470e+00 7.60903776e-01
-1.96941400e+00 -5.59229136e-01 1.91307381e-01 2.17438459e+00
1.45951593e+00 -8.49367026e-03 1.66403562e-01 -5.19568920e-01
8.12105656e-01 -2.92856127e-01 -7.09339261e-01 -6.27792537e-01
-1.58544004e-01 3.68246704e-01 5.04048407e-01 3.36488783e-01
-8.47388029e-01 1.61456048e+00 6.66152668e+00 8.32271695e-01
-1.21869528e+00 6.30466342e-01 1.89703748e-01 1.92787871e-01
-1.26070663e-01 2.02857945e-02 -7.42643714e-01 4.73596394e-01
1.31438732e+00 -3.50835979e-01 5.23246408e-01 7.65338600e-01
6.68469295e-02 4.10175860e-01 -1.16168594e+00 5.56487560e-01
1.89392224e-01 -1.06668735e+00 4.13295299e-01 6.32510334e-02
1.24758673e+00 4.32723373e-01 1.89676523e-01 6.67431414e-01
6.16462529e-01 -5.31735837e-01 1.61141738e-01 9.35599208e-03
1.20891273e+00 -7.09899187e-01 7.47403800e-01 6.95507526e-01
-6.33118331e-01 1.95462063e-01 -7.04769611e-01 -4.07234356e-02
-6.98301941e-02 1.45613611e-01 -1.20692062e+00 7.82015204e-01
2.24033445e-01 6.57930613e-01 -2.95352161e-01 6.33102655e-01
-3.22330415e-01 9.67269778e-01 -7.35938698e-02 7.47379512e-02
2.98401654e-01 -3.06132078e-01 1.26289800e-01 1.15882838e+00
6.43900514e-01 -1.18398845e-01 2.12956876e-01 2.84090221e-01
-4.50238079e-01 7.00706124e-01 -7.31955469e-01 -2.45011836e-01
4.71281201e-01 7.62596488e-01 -1.42557845e-01 -5.00302017e-01
-7.54198074e-01 1.53287828e+00 7.09026337e-01 6.53038144e-01
-5.74127793e-01 -1.84885576e-01 2.50470281e-01 -1.94439575e-01
2.18874574e-01 -3.34148079e-01 -2.40568891e-01 -1.60220397e+00
1.41171381e-01 -1.21903336e+00 2.96884298e-01 -4.82942581e-01
-1.38580191e+00 8.16951632e-01 -1.26618356e-01 -1.62487853e+00
-6.22404754e-01 -4.58611578e-01 -1.26482740e-01 9.58039165e-01
-1.82369685e+00 -1.60346913e+00 3.95238608e-01 5.88805676e-01
9.62554932e-01 -7.36707389e-01 1.05573535e+00 3.47499043e-01
-4.95310009e-01 9.26570535e-01 9.05671418e-01 1.27243191e-01
1.40362871e+00 -9.76966202e-01 5.17752349e-01 8.24090302e-01
2.35532284e-01 7.64074802e-01 4.13979083e-01 -8.87436211e-01
-1.44288981e+00 -1.29054213e+00 1.50055146e+00 -6.35999143e-01
6.14264965e-01 -1.64092287e-01 -8.84183049e-01 9.13078010e-01
5.07230282e-01 -3.07031542e-01 1.01827931e+00 2.32678071e-01
-4.91430521e-01 -1.07598796e-01 -1.07227397e+00 6.02966428e-01
1.02587748e+00 -6.75441742e-01 -1.09523797e+00 4.65742797e-01
7.58864284e-01 -2.64208257e-01 -8.82141292e-01 5.98086417e-01
4.06711847e-01 -3.01006466e-01 6.78860366e-01 -9.85535562e-01
5.42234361e-01 5.00368699e-03 -2.17784226e-01 -1.63820720e+00
-4.14487749e-01 -7.65264392e-01 -2.36864269e-01 1.26546884e+00
7.66030729e-01 -6.91586018e-01 6.17578745e-01 1.69745415e-01
-6.16919138e-02 -4.32180911e-01 -9.90325809e-01 -1.20641232e+00
7.67908752e-01 1.18496396e-01 6.02436125e-01 1.45255446e+00
2.73568302e-01 6.69377208e-01 -7.10926771e-01 -2.29736883e-03
5.56204498e-01 3.72198761e-01 8.23196709e-01 -9.44946468e-01
-3.29903841e-01 3.34509909e-02 1.81441754e-01 -1.22902703e+00
4.13715988e-01 -1.19647574e+00 -2.70079114e-02 -1.28862870e+00
3.62418801e-01 -3.51604789e-01 -8.01746726e-01 5.11411309e-01
-2.93447793e-01 4.07939017e-01 -1.66677058e-01 6.96089149e-01
-5.23434341e-01 5.43001831e-01 1.22331822e+00 -1.56972587e-01
-4.45465237e-01 -1.37511775e-01 -6.39441967e-01 5.12127221e-01
1.10093403e+00 -6.84533417e-01 -3.73038262e-01 -1.01238203e+00
-3.50763202e-02 -6.24061413e-02 -6.08156741e-01 -6.79122746e-01
2.13894203e-01 -5.61574578e-01 2.32842132e-01 -3.20307225e-01
7.82777555e-03 -7.84810483e-01 -1.35808125e-01 3.31768394e-01
-3.85959625e-01 2.99653769e-01 2.06018493e-01 4.14233625e-01
-2.75295436e-01 -8.46532732e-02 5.69179654e-01 -1.72046915e-01
-7.56336749e-01 1.93703309e-01 -4.39470887e-01 1.68482457e-05
6.68166161e-01 1.10185742e-02 -1.91774845e-01 -1.53696656e-01
-4.38274533e-01 2.14881869e-03 7.73228407e-01 6.90879941e-01
4.14390117e-01 -1.70324528e+00 -1.02608287e+00 1.37779444e-01
4.63337839e-01 -5.16057491e-01 -2.56456435e-01 7.00887918e-01
-2.73506284e-01 4.44250911e-01 -4.43215191e-01 -5.12518823e-01
-9.36625183e-01 5.94189167e-01 -7.16014951e-02 -3.55471104e-01
-3.14031065e-01 4.28831369e-01 -1.12300701e-01 -8.85103524e-01
-1.72285676e-01 1.09659560e-01 1.27180606e-01 -2.98837423e-01
3.19056362e-01 6.81465492e-02 6.73831403e-02 -6.34226620e-01
-2.14347675e-01 5.05757987e-01 -5.84516048e-01 -4.43259597e-01
1.33538365e+00 -3.82909238e-01 -8.88893008e-02 5.89675546e-01
1.00711310e+00 6.57740757e-02 -9.85412717e-01 -9.74497139e-01
2.46289894e-01 -4.62388724e-01 -1.70245960e-01 -1.22506011e+00
-4.58723515e-01 8.16268265e-01 4.64305431e-01 -6.98736131e-01
1.26129329e+00 -1.92022756e-01 1.07428002e+00 8.16747010e-01
8.42088878e-01 -1.51659298e+00 9.58786160e-02 7.53311455e-01
4.83304828e-01 -1.56925058e+00 -1.79053381e-01 -1.10625282e-01
-5.33823371e-01 1.03204560e+00 8.17676902e-01 2.26226851e-01
6.88267797e-02 9.07459296e-03 5.84273756e-01 5.68906605e-01
-8.40703130e-01 -1.00049384e-01 2.43337125e-01 7.29557753e-01
5.12516677e-01 1.22508593e-01 -6.24287546e-01 5.01879811e-01
1.04502052e-01 4.09522921e-01 1.77084029e-01 1.02861297e+00
-4.84419763e-01 -1.92595470e+00 -2.16585323e-01 6.12825677e-02
-5.84201872e-01 -4.30980951e-01 -6.83326244e-01 6.13579810e-01
1.22058205e-02 9.44874704e-01 -3.42721045e-01 -4.28985149e-01
9.09673944e-02 4.79644686e-01 6.60851061e-01 -7.62011051e-01
-6.10854983e-01 2.80733109e-01 1.39336705e-01 -9.18940231e-02
-7.07803905e-01 -5.33183694e-01 -7.28866041e-01 -1.46113083e-01
-3.35448116e-01 3.94426405e-01 6.16894841e-01 1.04479229e+00
5.25545061e-01 -6.13769218e-02 8.73832345e-01 -4.02455330e-01
-8.96689296e-01 -1.29215753e+00 7.34770447e-02 2.36394808e-01
1.00399598e-01 -4.38240230e-01 5.43289967e-02 4.09968644e-01] | [11.620697021484375, 10.307747840881348] |
599b8002-09da-4c5c-93d0-512af0aaeef6 | adaptive-weighted-discriminator-for-training | 2012.03149 | null | https://arxiv.org/abs/2012.03149v2 | https://arxiv.org/pdf/2012.03149v2.pdf | Adaptive Weighted Discriminator for Training Generative Adversarial Networks | Generative adversarial network (GAN) has become one of the most important neural network models for classical unsupervised machine learning. A variety of discriminator loss functions have been developed to train GAN's discriminators and they all have a common structure: a sum of real and fake losses that only depends on the actual and generated data respectively. One challenge associated with an equally weighted sum of two losses is that the training may benefit one loss but harm the other, which we show causes instability and mode collapse. In this paper, we introduce a new family of discriminator loss functions that adopts a weighted sum of real and fake parts, which we call adaptive weighted loss functions or aw-loss functions. Using the gradients of the real and fake parts of the loss, we can adaptively choose weights to train a discriminator in the direction that benefits the GAN's stability. Our method can be potentially applied to any discriminator model with a loss that is a sum of the real and fake parts. Experiments validated the effectiveness of our loss functions on an unconditional image generation task, improving the baseline results by a significant margin on CIFAR-10, STL-10, and CIFAR-100 datasets in Inception Scores and FID. | ['Qiang Ye', 'Qiang Cheng', 'Vasily Zadorozhnyy'] | 2020-12-05 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zadorozhnyy_Adaptive_Weighted_Discriminator_for_Training_Generative_Adversarial_Networks_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zadorozhnyy_Adaptive_Weighted_Discriminator_for_Training_Generative_Adversarial_Networks_CVPR_2021_paper.pdf | cvpr-2021-1 | ['conditional-image-generation'] | ['computer-vision'] | [ 2.27393836e-01 1.92401856e-01 1.01404920e-01 -3.76203150e-01
-6.59188867e-01 -5.63897252e-01 6.92971110e-01 -3.90267819e-01
-3.64098340e-01 9.35828328e-01 -1.66761562e-01 -1.06168598e-01
1.57870129e-01 -9.29191828e-01 -7.54384637e-01 -1.03118479e+00
-3.64425476e-03 3.70118022e-01 2.07587525e-01 -5.67849815e-01
-1.06370382e-01 3.14723372e-01 -1.10664308e+00 1.13126531e-01
1.06921792e+00 1.19369626e+00 -3.96691471e-01 5.67754328e-01
2.01690242e-01 9.31648195e-01 -1.13162422e+00 -7.83441782e-01
5.70172250e-01 -8.80951524e-01 -5.02950430e-01 -3.80730867e-01
2.82643080e-01 -2.96973079e-01 -2.61690944e-01 1.16138208e+00
6.26780748e-01 -1.09351441e-01 9.45411623e-01 -1.52588212e+00
-7.99209893e-01 5.67658305e-01 -3.68874043e-01 9.47937891e-02
-1.41185150e-01 2.22049490e-01 6.66908324e-01 -4.96668875e-01
3.48925531e-01 1.13557053e+00 7.84761369e-01 8.52409422e-01
-1.20523560e+00 -9.24062371e-01 -6.57674223e-02 -1.79064602e-01
-1.14302564e+00 -1.15088284e-01 9.06719744e-01 -5.54706395e-01
5.76832831e-01 2.38287404e-01 2.14367434e-01 1.30429423e+00
2.96529889e-01 4.89850968e-01 1.06831348e+00 -3.75432491e-01
1.11271232e-01 4.11727101e-01 -3.33728135e-01 5.23965120e-01
-4.03034985e-02 2.96255171e-01 -3.98949161e-02 -3.49347629e-02
7.99973011e-01 -2.20832422e-01 -4.73776311e-01 -2.50180006e-01
-7.09877908e-01 1.07134652e+00 7.03891695e-01 3.19690049e-01
-4.69072871e-02 3.85043800e-01 3.10483783e-01 5.95868528e-01
5.90531528e-01 6.12772703e-01 -1.09561421e-01 2.37170421e-02
-6.68262720e-01 1.02031805e-01 4.32965249e-01 6.39960468e-01
6.24994397e-01 5.01002431e-01 -4.58700448e-01 9.74042118e-01
-5.29564023e-02 3.36090028e-01 9.08059180e-01 -5.34334838e-01
4.88829821e-01 4.02911544e-01 7.66735375e-02 -9.65331912e-01
7.18102828e-02 -6.48296118e-01 -1.18013668e+00 6.80399179e-01
5.20748675e-01 -3.69458765e-01 -1.06942046e+00 2.06616974e+00
-2.95938514e-02 2.37974331e-01 -9.66681447e-03 7.84713566e-01
5.14237046e-01 5.45399070e-01 -5.49596697e-02 1.91591397e-01
7.09939778e-01 -9.68157709e-01 -5.32838404e-01 -1.82129458e-01
4.63120759e-01 -6.65692925e-01 1.07328594e+00 2.49607220e-01
-1.27970564e+00 -5.50936937e-01 -1.24104202e+00 1.68451458e-01
-4.64355677e-01 1.46175116e-01 3.79750371e-01 8.56579781e-01
-1.00553083e+00 1.07132280e+00 -5.40392339e-01 2.54990071e-01
5.37939191e-01 3.30918342e-01 -1.67330593e-01 3.21164548e-01
-1.46424198e+00 8.72062981e-01 3.08632821e-01 5.02594784e-02
-1.04144824e+00 -6.40782058e-01 -5.07284105e-01 1.26960918e-01
-1.11911163e-01 -5.12442410e-01 9.25056636e-01 -1.60905564e+00
-1.60929632e+00 7.38684595e-01 6.04618669e-01 -7.44438708e-01
1.17336011e+00 -1.31518930e-01 -3.70863914e-01 -1.78985789e-01
7.20717534e-02 5.95496118e-01 1.12417579e+00 -1.17987573e+00
-2.41413742e-01 -6.39573932e-02 -6.68924525e-02 1.87935270e-02
-4.34469283e-01 -3.46696526e-01 2.61632621e-01 -1.06477499e+00
-2.81892627e-01 -8.52702498e-01 7.21733570e-02 1.93989109e-02
-5.38349330e-01 -3.63174938e-02 7.93347776e-01 -4.90006596e-01
9.99540150e-01 -2.28644276e+00 1.17756754e-01 1.09089464e-01
7.01090023e-02 6.02272689e-01 -1.89498380e-01 3.33650649e-01
-2.57864803e-01 2.66088843e-01 -6.28262341e-01 -4.11327392e-01
-9.20276344e-02 1.21057123e-01 -6.97600424e-01 3.74404609e-01
6.04999185e-01 7.18799055e-01 -1.03170311e+00 -2.92049292e-02
-2.36815605e-02 6.32286727e-01 -3.07402462e-01 4.00266558e-01
-1.93236396e-02 4.16160643e-01 -1.08399823e-01 1.04893900e-01
7.36955822e-01 1.73497051e-01 -2.29066715e-01 9.81895179e-02
3.68490309e-01 1.79616734e-01 -8.74571323e-01 1.11634433e+00
-3.29538524e-01 4.54409599e-01 -2.92824686e-01 -1.05563939e+00
1.07044160e+00 2.42503718e-01 1.60224107e-03 -5.63172221e-01
1.65446103e-01 3.37030947e-01 1.05461977e-01 4.33318727e-02
2.39997625e-01 -6.12497211e-01 -5.93415909e-02 3.82437587e-01
3.08110952e-01 -1.25963956e-01 -1.96761683e-01 1.06832579e-01
9.52605665e-01 -6.45339936e-02 -2.39936709e-02 -8.54984447e-02
5.21056652e-01 -4.70823675e-01 5.18918753e-01 7.63966322e-01
-2.31570572e-01 8.73642683e-01 7.73131967e-01 -2.82645762e-01
-1.04109180e+00 -1.22400475e+00 1.42768575e-02 6.98688149e-01
-4.85790186e-02 2.87475824e-01 -7.16698468e-01 -9.97339249e-01
-1.33809909e-01 8.20384383e-01 -7.40342855e-01 -8.00769925e-01
-5.88536024e-01 -9.16225135e-01 1.06934857e+00 5.48349261e-01
8.38395774e-01 -1.07071543e+00 -3.22426796e-01 -4.81582945e-03
2.62183845e-02 -8.02900434e-01 -5.95132768e-01 2.64944494e-01
-6.89824879e-01 -8.37074697e-01 -8.72808278e-01 -7.15641916e-01
7.12018073e-01 -2.05634162e-01 1.13081574e+00 6.43560216e-02
-6.20848453e-03 -6.08036630e-02 -2.74678349e-01 -5.01014292e-01
-6.96545541e-01 -9.70736146e-03 -1.85662836e-01 3.08504999e-01
-2.98313230e-01 -6.48539305e-01 -4.95261610e-01 3.23381484e-01
-1.09508133e+00 -1.34901375e-01 2.60661930e-01 1.12508488e+00
3.87199223e-01 2.85657346e-01 8.91458631e-01 -1.03595603e+00
7.19291747e-01 -5.26841760e-01 -5.02891541e-01 1.47524640e-01
-7.81879365e-01 2.97999531e-01 1.14494014e+00 -6.76572978e-01
-1.08348274e+00 -3.40501994e-01 -2.10450590e-01 -6.53219998e-01
1.97151244e-01 8.82794484e-02 -2.33036473e-01 -9.38005820e-02
6.30412698e-01 2.70010501e-01 -8.38382542e-02 -3.03734243e-01
3.95883471e-01 4.41419482e-01 6.15417480e-01 -4.04270679e-01
9.37954605e-01 3.69492620e-01 8.25067461e-02 -2.98491538e-01
-5.10891318e-01 2.05641448e-01 -1.46436632e-01 -1.63331941e-01
6.55323505e-01 -6.59023821e-01 -2.84300387e-01 1.09224331e+00
-1.01582015e+00 -4.85524327e-01 -7.14126527e-01 2.62763977e-01
-3.39235693e-01 1.57721192e-01 -6.58543468e-01 -6.73193872e-01
-4.78439391e-01 -1.07707202e+00 6.36628568e-01 1.59715250e-01
1.63332537e-01 -1.15533960e+00 1.77677497e-01 -4.72936295e-02
7.34394014e-01 8.00744712e-01 1.01686919e+00 -6.34523094e-01
-2.19004050e-01 -4.28491026e-01 -9.71779972e-03 1.11739099e+00
1.60210371e-01 -2.16570105e-02 -9.34762359e-01 -5.11018276e-01
2.71572232e-01 -5.90840936e-01 1.11635280e+00 3.77501771e-02
1.15393114e+00 -6.60670936e-01 1.18465364e-01 7.15800881e-01
1.57421505e+00 3.83638233e-01 9.82565939e-01 1.74036622e-01
6.00836635e-01 2.34887153e-01 2.35066898e-02 8.23785365e-02
-1.64345548e-01 6.27948046e-01 6.93516731e-01 -1.30932838e-01
-1.94915369e-01 -4.31006372e-01 4.46880698e-01 6.57419562e-01
-8.83943513e-02 -3.75054419e-01 -4.45660830e-01 4.15432245e-01
-1.50927973e+00 -1.06548774e+00 1.10690042e-01 2.45941043e+00
9.45218146e-01 4.09256756e-01 1.59164831e-01 2.86890984e-01
7.31829345e-01 1.75081477e-01 -5.97060084e-01 -6.19803488e-01
-2.88481086e-01 5.75753570e-01 4.62773710e-01 4.83804166e-01
-1.02464747e+00 6.96159303e-01 5.90532827e+00 1.11851656e+00
-1.37245011e+00 2.28312001e-01 9.57783222e-01 8.40877276e-03
-4.65003431e-01 -2.23099515e-01 -3.93334478e-01 8.96185517e-01
8.00331175e-01 4.81626280e-02 4.07553524e-01 8.94193232e-01
-2.34508246e-01 2.81158268e-01 -9.55733180e-01 6.53214872e-01
-3.08881495e-02 -1.06591356e+00 2.15798125e-01 -1.09652884e-01
9.45995212e-01 -9.67026651e-02 4.53471810e-01 4.14212734e-01
4.75952685e-01 -1.30884182e+00 8.07018936e-01 3.40599626e-01
8.48274171e-01 -8.43043149e-01 8.57467949e-01 2.93860048e-01
-7.54497111e-01 1.09867543e-01 -2.81797975e-01 -3.95146683e-02
-2.06167120e-02 6.07930064e-01 -7.52445638e-01 4.77302074e-01
2.75444955e-01 4.53767002e-01 -4.13027018e-01 8.02930176e-01
-4.41170961e-01 6.74846351e-01 -9.79064479e-02 2.62704879e-01
2.53689319e-01 -3.52425784e-01 5.95341086e-01 9.64491069e-01
2.97197312e-01 -4.01729435e-01 -3.34121794e-01 1.10276997e+00
-5.52290261e-01 -3.05582106e-01 -5.17644644e-01 6.38559833e-02
4.21801716e-01 9.70621705e-01 -4.86305475e-01 -1.51014835e-01
6.39939867e-03 1.21297610e+00 4.02748913e-01 3.15136909e-01
-1.14048803e+00 -7.15754032e-01 4.97162223e-01 8.53056312e-02
2.86413372e-01 1.86239108e-01 -3.12806129e-01 -1.06583691e+00
3.55113000e-02 -7.87688732e-01 3.16160530e-01 -6.38303339e-01
-1.56636548e+00 1.06557190e+00 -1.36145294e-01 -1.31969297e+00
-3.20127696e-01 -4.55907315e-01 -1.00743389e+00 1.05423832e+00
-1.53796458e+00 -9.42720652e-01 -2.72718251e-01 6.69391215e-01
1.30205989e-01 -3.85747194e-01 7.84062922e-01 4.53685552e-01
-5.85236073e-01 1.08312643e+00 1.56295881e-01 4.29023564e-01
5.23463249e-01 -1.33817303e+00 4.09499913e-01 8.47332180e-01
7.61936679e-02 1.80659682e-01 6.64030969e-01 -2.88763791e-01
-5.26501656e-01 -1.21031713e+00 7.61884034e-01 -2.48142734e-01
4.78751212e-01 -3.39184582e-01 -8.08084965e-01 8.02284300e-01
1.87573820e-01 8.19548070e-02 3.60288441e-01 -5.79670429e-01
-5.69622457e-01 -2.24546373e-01 -1.50562119e+00 4.15468812e-01
7.15310931e-01 -3.05969596e-01 -2.75085509e-01 2.34011784e-01
6.71458006e-01 -2.97443718e-01 -6.57374680e-01 5.76652467e-01
4.46043640e-01 -1.28988683e+00 9.71124291e-01 -6.28758907e-01
5.95625758e-01 -5.68891987e-02 8.52030516e-02 -1.67462432e+00
-2.32239932e-01 -5.10399401e-01 1.08402602e-01 1.39801073e+00
2.14444548e-01 -9.84529376e-01 6.23815060e-01 2.45048091e-01
-4.55568582e-02 -1.07964921e+00 -9.99210596e-01 -1.09335446e+00
5.51803231e-01 -2.63222419e-02 6.59731090e-01 8.88648629e-01
-5.09196997e-01 2.98242033e-01 -5.66120505e-01 -1.54366881e-01
5.63432336e-01 -2.66316473e-01 6.13872588e-01 -1.02517045e+00
-5.00500321e-01 -6.87888443e-01 -4.52930063e-01 -7.75243163e-01
2.75505912e-02 -9.11356211e-01 -1.94676414e-01 -9.27052796e-01
-2.22341925e-01 -7.17651129e-01 -4.02112067e-01 4.47807550e-01
-8.76481459e-02 2.68668979e-01 2.05001175e-01 3.41898918e-01
9.34694037e-02 8.76353025e-01 1.16155696e+00 -2.25071684e-01
-1.54712901e-01 1.38495773e-01 -5.25835991e-01 6.27874851e-01
9.15274918e-01 -6.84343278e-01 -5.60926735e-01 -3.75361949e-01
-3.50804511e-03 -3.12796980e-01 5.47993541e-01 -1.18608451e+00
-2.26690590e-01 2.43163612e-02 3.18242610e-01 3.64562944e-02
3.83315742e-01 -6.36976063e-01 2.90992141e-01 5.74944496e-01
-5.33042848e-01 -1.50889039e-01 -7.57753924e-02 5.14359474e-01
-3.39526117e-01 -2.43347585e-01 1.33599949e+00 -1.46010458e-01
-4.42882851e-02 2.30335981e-01 1.78467050e-01 3.84456128e-01
1.03075111e+00 -9.61430743e-03 -4.17875946e-01 -6.22453511e-01
-3.44666064e-01 -1.01365931e-01 3.10220271e-01 3.38615984e-01
5.65391541e-01 -1.55022216e+00 -8.28105807e-01 4.40552592e-01
-1.64521694e-01 -1.07557349e-01 3.56286317e-02 4.97994334e-01
-4.61905926e-01 1.19472049e-01 -4.87019271e-01 -1.62812009e-01
-8.53775203e-01 4.23846185e-01 8.25457871e-01 -7.18516946e-01
-2.48165742e-01 1.17499876e+00 3.75968635e-01 -2.69881725e-01
1.75838932e-01 -1.37319103e-01 4.82059456e-02 -2.18598738e-01
2.62925625e-01 4.16866183e-01 1.50089324e-01 -6.70164526e-01
-2.09202528e-01 2.56304204e-01 7.87814185e-02 -1.14002777e-02
1.11263657e+00 2.99243689e-01 3.78372446e-02 3.23395759e-01
1.33245242e+00 -1.89259544e-01 -1.36040282e+00 -4.86913882e-03
-4.92706060e-01 -4.10205930e-01 -1.71920806e-01 -1.04702806e+00
-1.47528446e+00 9.04641509e-01 6.63308322e-01 4.70508516e-01
1.37165844e+00 -2.68536836e-01 9.24906850e-01 -3.37725759e-01
2.72141486e-01 -8.50265622e-01 3.06448907e-01 2.28962883e-01
9.77952421e-01 -1.05431199e+00 -4.96347219e-01 -1.26795799e-01
-7.63396740e-01 8.40860486e-01 5.55600107e-01 -5.99114418e-01
5.46451688e-01 4.08339240e-02 -2.51289401e-02 1.42126098e-01
-4.43461567e-01 1.98700398e-01 2.69604921e-01 5.29739618e-01
2.13388890e-01 1.00536868e-01 -4.37759578e-01 6.36108398e-01
-2.72074878e-01 -1.45373836e-01 4.30280566e-01 6.13328397e-01
-8.60517919e-02 -1.26905596e+00 -9.84220281e-02 4.46712166e-01
-6.13673866e-01 -1.11772276e-01 -4.08471614e-01 7.06074178e-01
5.41546881e-01 7.41023898e-01 7.35638989e-03 -5.65265834e-01
2.82730043e-01 1.09941795e-01 3.20750624e-01 -2.98288524e-01
-7.46048689e-01 -3.64194036e-01 -1.72589689e-01 -2.73113847e-01
-1.65892735e-01 -1.72213495e-01 -1.00903690e+00 -2.66704947e-01
-3.86139363e-01 2.09693268e-01 4.65104878e-01 8.38150561e-01
1.72314271e-01 4.68461573e-01 9.72350001e-01 -6.00242853e-01
-1.10362196e+00 -1.01184869e+00 -6.46599650e-01 7.83900857e-01
4.95632708e-01 -5.97220778e-01 -7.98502088e-01 -6.07152171e-02] | [11.63056468963623, -0.2122114896774292] |
86c8d9bb-ed4d-4d6b-b0bb-6cad150d3063 | cnn-based-posture-free-hand-detection | 1809.10432 | null | http://arxiv.org/abs/1809.10432v1 | http://arxiv.org/pdf/1809.10432v1.pdf | CNN Based Posture-Free Hand Detection | Although many studies suggest high performance hand detection methods, those
methods are likely to be overfitting. Fortunately, the Convolution Neural
Network (CNN) based approach provides a better way that is less sensitive to
translation and hand poses. However the CNN approach is complex and can
increase computational time, which at the end reduce its effectiveness on a
system where the speed is essential.In this study we propose a shallow CNN
network which is fast, and insensitive to translation and hand poses. It is
tested on two different domains of hand datasets, and performs in relatively
comparable performance and faster than the other state-of-the-art hand
CNN-based hand detection method. Our evaluation shows that the proposed shallow
CNN network performs at 93.9% accuracy and reaches much faster speed than its
competitors. | ['Richard Adiguna', 'Yustinus Eko Soelistio'] | 2018-09-27 | null | null | null | null | ['hand-detection'] | ['computer-vision'] | [-4.71670896e-01 -6.68686688e-01 -3.40547591e-01 9.59148258e-02
-1.40486851e-01 -4.08818543e-01 2.14016259e-01 -5.37469149e-01
-5.84685981e-01 4.85517383e-01 2.25149870e-01 -1.45017728e-01
2.40939051e-01 -7.00675130e-01 -2.72911638e-01 -6.43769681e-01
1.78953618e-01 4.15308803e-01 5.35157144e-01 -2.58460164e-01
1.49277866e-01 8.02635312e-01 -1.32149148e+00 2.02799365e-01
1.76936373e-01 6.50211573e-01 1.56183615e-01 7.18857110e-01
1.26958162e-01 8.00920248e-01 -7.97567785e-01 -3.83091301e-01
3.16710263e-01 -4.58982959e-02 -1.07141078e+00 -4.39851791e-01
5.51973403e-01 -9.23060715e-01 -7.27896631e-01 9.33318913e-01
1.30129719e+00 -1.02778673e-01 4.48341697e-01 -8.18065643e-01
-6.96651876e-01 3.90782654e-01 -6.94734037e-01 4.50510502e-01
3.41286957e-01 3.05940449e-01 6.92456603e-01 -9.58689570e-01
5.68592846e-01 1.31388319e+00 9.47528839e-01 7.40090609e-01
-6.53191328e-01 -8.36618423e-01 -1.51205063e-02 -4.43949737e-02
-1.55342388e+00 3.96329165e-02 5.21603823e-01 -3.34017992e-01
1.24370062e+00 -3.08333803e-02 6.86705887e-01 1.44748533e+00
2.87151393e-02 1.25377345e+00 8.47899973e-01 -5.10200679e-01
-2.80449748e-01 -1.31493723e-02 1.40870348e-01 9.13437486e-01
3.78217906e-01 7.40912408e-02 -3.80425900e-01 2.89808866e-03
1.29722798e+00 2.24364087e-01 -1.88922733e-01 1.71226799e-01
-1.24072981e+00 6.16022408e-01 7.77712822e-01 7.19525397e-01
-3.96865934e-01 1.81397185e-01 6.10124111e-01 8.07042718e-02
9.67847928e-02 1.66726619e-01 -5.41391432e-01 -1.23005033e-01
-1.00608337e+00 5.63359857e-01 6.97599173e-01 1.02042079e+00
-2.12355882e-01 1.47839829e-01 -4.57055360e-01 7.54585207e-01
7.15973228e-02 4.13927615e-01 3.44680607e-01 -2.13827819e-01
7.20242620e-01 5.14994264e-01 2.28014849e-02 -9.93991435e-01
-7.91342437e-01 -6.42330766e-01 -9.57457721e-01 4.41066027e-01
8.71631503e-01 -1.35819020e-03 -1.13769913e+00 1.38067496e+00
-7.72402659e-02 -6.02178574e-01 -3.90058577e-01 1.26949513e+00
1.13880563e+00 2.90649921e-01 2.37267047e-01 2.66591102e-01
1.51255476e+00 -1.11907494e+00 -8.58907521e-01 -1.45427272e-01
3.81601006e-01 -1.26168299e+00 1.30996180e+00 6.44298017e-01
-7.91820586e-01 -7.10042655e-01 -9.69948053e-01 -1.72993079e-01
-4.47416455e-01 6.15820408e-01 7.60771215e-01 9.50401843e-01
-8.61909926e-01 6.18853331e-01 -6.82903826e-01 -8.45936358e-01
5.57419658e-01 5.16929746e-01 -2.32227474e-01 1.98102966e-01
-9.23352182e-01 1.10383928e+00 4.72082913e-01 2.75245607e-01
-6.04697287e-01 -1.37121499e-01 -2.51672566e-01 -9.06181857e-02
2.40312025e-01 -5.41106582e-01 1.32838023e+00 -6.37759387e-01
-1.40958571e+00 5.56177616e-01 3.03859264e-02 2.92053223e-02
8.52867186e-01 -4.42210704e-01 -2.13162884e-01 -2.04016760e-01
-2.40815759e-01 5.98019421e-01 9.18628871e-01 -7.64158607e-01
-5.42153656e-01 -5.42044282e-01 2.89750732e-02 -1.62836999e-01
-5.21043718e-01 5.73498785e-01 -5.48696041e-01 -9.02484119e-01
6.37442023e-02 -9.44964290e-01 2.32186913e-01 1.23027958e-01
-5.80338955e-01 -6.51456714e-01 9.99160290e-01 -9.19489384e-01
1.25998318e+00 -1.67773354e+00 -1.46385118e-01 4.87611145e-02
1.61422238e-01 7.80800760e-01 -1.36298835e-01 3.63905728e-01
8.67729336e-02 5.53629212e-02 2.71863073e-01 -2.95207471e-01
-2.68544734e-01 -5.04866987e-02 7.32069239e-02 3.63668382e-01
-8.24158788e-02 8.73391807e-01 -6.27124548e-01 -6.89939916e-01
2.70244747e-01 9.19979095e-01 -4.50449109e-01 2.90876657e-01
1.23141259e-01 2.86007583e-01 -4.20842946e-01 1.17965245e+00
5.25890589e-01 -1.00510351e-01 1.01916224e-01 -4.99616593e-01
-5.81108853e-02 3.42740528e-02 -1.22140324e+00 1.41853750e+00
-2.49469638e-01 1.00650012e+00 -3.52766931e-01 -4.35811996e-01
7.15852201e-01 6.62070155e-01 -1.04838632e-01 -3.46336782e-01
8.38567853e-01 2.67693549e-01 2.69739419e-01 -8.46363902e-01
3.14456969e-01 2.53354698e-01 5.65237045e-01 5.72305679e-01
1.72113985e-01 3.18331689e-01 -1.23238802e-01 -1.53642625e-01
7.88512111e-01 1.18749544e-01 3.35864514e-01 -2.80973077e-01
3.60628515e-01 -3.84859294e-01 2.44589016e-01 6.45385921e-01
-5.22009254e-01 6.62691712e-01 2.25690410e-01 -1.08790350e+00
-1.20657003e+00 -6.83573902e-01 1.70764532e-02 1.13758981e+00
-3.42702329e-01 -2.79665411e-01 -9.75208521e-01 -6.46943986e-01
-1.22766629e-01 -2.31018990e-01 -7.03049958e-01 3.17947060e-01
-9.29649472e-01 -6.49149239e-01 9.00546730e-01 1.42768073e+00
9.90625679e-01 -1.46722221e+00 -6.47437632e-01 2.31174931e-01
-2.99784243e-01 -1.07758176e+00 -5.40118873e-01 -1.86387226e-01
-9.65939105e-01 -1.19358253e+00 -1.41845763e+00 -1.18611777e+00
4.56442565e-01 1.72484383e-01 9.50411558e-01 3.16923350e-01
-6.39351308e-01 -1.39014438e-01 -4.52233583e-01 -6.77813709e-01
7.98225477e-02 7.42502153e-01 3.61168295e-01 -4.26242471e-01
5.88326693e-01 -2.51256287e-01 -9.74988580e-01 2.39901692e-01
-2.57361978e-01 -4.07777488e-01 7.43275821e-01 9.32739079e-01
1.03694841e-01 -6.45171478e-02 1.74497187e-01 -3.33855212e-01
8.62913489e-01 2.48379841e-01 -3.60827416e-01 2.26725906e-01
-6.92004323e-01 -8.21776092e-02 4.19245243e-01 -8.59960854e-01
-1.01493859e+00 3.79050672e-01 -2.04675242e-01 -2.24183008e-01
-4.31361124e-02 7.57161677e-02 8.34753662e-02 -4.80234265e-01
8.45199704e-01 1.38696238e-01 -2.44822502e-01 -8.89128983e-01
-5.94340917e-03 9.96197343e-01 4.62760717e-01 -3.87641042e-01
6.37222111e-01 3.17665279e-01 -9.54469964e-02 -7.77981400e-01
-5.44394016e-01 -3.13115656e-01 -1.01612461e+00 -3.27613443e-01
8.16626370e-01 -5.55482924e-01 -9.50618386e-01 9.60333645e-01
-1.60984087e+00 -1.89448357e-01 5.14443636e-01 5.02479017e-01
-1.72498837e-01 4.21608448e-01 -9.41136718e-01 -9.51104105e-01
-8.61069024e-01 -1.17677045e+00 8.82479250e-01 2.19933137e-01
-4.31674451e-01 -5.77086926e-01 -2.47470349e-01 1.70455933e-01
5.89689612e-01 3.57552879e-02 6.21300399e-01 -2.42375895e-01
-3.99698824e-01 -7.76595175e-01 -5.60982943e-01 4.41304207e-01
1.66432396e-01 2.99822897e-01 -1.11671340e+00 -5.51651895e-01
-4.84698683e-01 -1.82928681e-01 6.98901057e-01 5.05750060e-01
1.22816658e+00 -2.22234398e-01 -3.98131639e-01 3.46844405e-01
1.22788250e+00 5.65165728e-02 7.43622899e-01 6.02121711e-01
8.74917388e-01 4.59041148e-01 2.34800056e-01 3.92597318e-01
2.84717456e-02 7.31593728e-01 2.05035657e-01 -3.97953302e-01
-4.19692308e-01 -9.86663625e-02 -4.04597223e-02 4.15373206e-01
-1.25374591e+00 -2.35417768e-01 -9.83390093e-01 5.85307181e-01
-1.65836930e+00 -1.08806670e+00 -2.01740101e-01 1.78585637e+00
6.99294925e-01 5.82593009e-02 9.29860115e-01 6.85687900e-01
7.13032842e-01 8.12025275e-03 -8.99269655e-02 -2.88306296e-01
5.30593060e-02 6.76709533e-01 5.12545526e-01 -4.66331802e-02
-1.28641391e+00 1.25117517e+00 7.37337875e+00 4.25511658e-01
-1.14394331e+00 3.10985595e-01 -5.29597960e-02 -1.66661099e-01
7.90980995e-01 -7.35762358e-01 -7.90335834e-01 1.64604858e-01
-2.43895426e-02 6.36034310e-01 3.91657352e-01 1.20683527e+00
-1.54917657e-01 3.50550890e-01 -9.88803923e-01 1.24081230e+00
1.12363778e-01 -9.44320023e-01 1.34176224e-01 -9.97499973e-02
4.28361773e-01 3.48574045e-04 7.85468593e-02 5.72858676e-02
-1.95484594e-01 -1.13104594e+00 8.36431324e-01 1.81539748e-02
8.90206099e-01 -8.60901058e-01 1.23987103e+00 2.60025471e-01
-1.41560125e+00 -1.82286114e-01 -3.55033815e-01 -3.28931779e-01
-1.29002050e-01 -7.21359402e-02 -6.51334941e-01 -2.94126242e-01
1.14960647e+00 9.29144323e-02 -4.48964208e-01 1.04309630e+00
-4.09824938e-01 3.36992294e-01 -9.46849063e-02 -5.97155631e-01
1.82896033e-01 6.40446007e-01 3.13280880e-01 1.57886088e+00
3.37296948e-02 3.07591945e-01 -9.03345719e-02 7.09599972e-01
7.57747795e-04 2.42765591e-01 -4.57336038e-01 -7.31001869e-02
3.60290468e-01 1.09619188e+00 -9.35012937e-01 -2.57983416e-01
-5.16544282e-01 1.30736256e+00 2.05901876e-01 2.74574786e-01
-5.27436495e-01 -7.60871351e-01 5.79063058e-01 8.37097317e-02
2.11050674e-01 -3.50546628e-01 -3.79085481e-01 -1.00298786e+00
2.96417177e-01 -8.76582146e-01 3.65020305e-01 -5.97953677e-01
-1.19919407e+00 9.79769349e-01 -5.01948774e-01 -1.16069674e+00
9.56645410e-04 -1.21635890e+00 -5.49810767e-01 9.09619749e-01
-1.19973624e+00 -1.43753076e+00 -7.24349380e-01 9.29819703e-01
6.77417994e-01 -4.15331364e-01 1.05791998e+00 5.15145898e-01
-6.94857657e-01 1.25617480e+00 -4.36202139e-01 9.41387773e-01
6.55738890e-01 -9.96673167e-01 6.99708760e-01 7.41970837e-01
-1.96603417e-01 1.06005776e+00 5.27675092e-01 -5.75957596e-01
-1.20135152e+00 -6.99656904e-01 9.34566140e-01 -5.63765645e-01
1.18030064e-01 -1.59321010e-01 -6.75013661e-01 6.67000234e-01
1.44289002e-01 9.20400098e-02 2.74846524e-01 4.54346687e-01
-8.44618082e-01 1.41582295e-01 -1.38262546e+00 5.23125470e-01
1.28592598e+00 -7.28857517e-01 -6.09518588e-01 4.62773323e-01
1.52707547e-01 -5.17572820e-01 -7.13056505e-01 3.11130047e-01
1.37945437e+00 -8.62438142e-01 1.10833335e+00 -7.39470303e-01
2.64456719e-01 -3.28652672e-02 -9.37811157e-04 -8.55773151e-01
-7.33047009e-01 -3.40374738e-01 -2.31976822e-01 1.02288413e+00
2.85018265e-01 -6.04235709e-01 8.95952165e-01 2.24413440e-01
5.41874886e-01 -7.34260798e-01 -6.95956409e-01 -1.29050338e+00
9.11144540e-02 -2.08944038e-01 8.09440911e-01 7.93138027e-01
-2.60333195e-02 3.03658396e-01 -5.16837418e-01 8.80263001e-02
6.60544932e-01 -1.45586044e-01 7.31666327e-01 -1.25103080e+00
2.52656732e-02 -7.03155100e-01 -5.54396391e-01 -7.22355604e-01
-2.11487055e-01 -3.73026818e-01 -6.90577850e-02 -1.50817418e+00
5.02075732e-01 -1.82819098e-01 -2.35063344e-01 7.47049510e-01
-3.08125727e-02 6.85543895e-01 2.10237443e-01 4.22585547e-01
1.44373760e-01 -1.45282894e-02 1.38753629e+00 -3.16231698e-01
-3.01847428e-01 1.60001308e-01 -1.80244088e-01 8.49480510e-01
1.04796910e+00 -3.59037548e-01 1.04292922e-01 -5.78482211e-01
-1.19012967e-01 -4.45070684e-01 4.18567628e-01 -1.13297927e+00
2.14748383e-01 1.89985633e-01 7.46984124e-01 -8.68590891e-01
1.43141989e-02 -6.02478206e-01 -4.23194110e-01 8.74583423e-01
-5.61651075e-03 1.00717835e-01 4.95916419e-02 -1.10382296e-01
8.65174010e-02 -1.13471299e-02 8.60710859e-01 -3.24077547e-01
-5.60968816e-01 5.22551060e-01 -2.81588137e-01 -4.73380029e-01
4.31771159e-01 -4.75509048e-01 -5.35372905e-02 -2.69767821e-01
-4.28206325e-01 -4.45372999e-01 -6.67396858e-02 6.69822037e-01
4.39707071e-01 -1.33868992e+00 -6.33896470e-01 1.66098714e-01
3.74428891e-02 -3.21055770e-01 -1.09961115e-01 6.96290374e-01
-8.23219657e-01 8.44765365e-01 -4.95498180e-01 -4.03492987e-01
-1.67903948e+00 5.62707484e-01 4.28767979e-01 1.33837357e-01
-8.39234352e-01 1.14187896e+00 -2.62980133e-01 -3.33127499e-01
9.86389458e-01 -4.86171663e-01 -2.33521745e-01 -1.32144973e-01
8.75031292e-01 8.16373706e-01 1.52165771e-01 -4.98332411e-01
-6.25519872e-01 9.62483764e-01 -1.35392606e-01 2.45690435e-01
1.02007711e+00 5.69695294e-01 -2.72569284e-02 -1.81015715e-01
9.19628382e-01 -3.18200678e-01 -7.05512881e-01 -1.97044432e-01
-2.36167759e-01 -6.45578980e-01 1.59853548e-01 -1.12843478e+00
-1.29977775e+00 1.18572783e+00 1.20705462e+00 -4.93490212e-02
9.95209694e-01 5.00352532e-02 1.12032950e+00 6.86382234e-01
5.47092199e-01 -1.21886706e+00 2.33014390e-01 6.08460903e-01
1.09953701e+00 -1.35540104e+00 1.43082201e-01 -5.06825030e-01
-1.64611131e-01 1.56854331e+00 9.11868334e-01 -5.37912883e-02
5.93649805e-01 4.40573245e-01 2.91017056e-01 -3.14714849e-01
2.25216866e-01 -4.41902876e-01 3.48725200e-01 7.01990604e-01
8.31257820e-01 3.12003613e-01 -4.76102442e-01 5.37935674e-01
-2.16208786e-01 4.05139744e-01 -1.71593517e-01 9.99436677e-01
-2.30923399e-01 -9.18736279e-01 -4.87532258e-01 2.18853503e-01
-7.27299988e-01 -5.04229777e-02 -6.39406145e-01 1.18633485e+00
2.24630132e-01 6.80627346e-01 -3.05828959e-01 -7.99267292e-01
6.82929933e-01 -5.70131727e-02 1.03607213e+00 -2.72936285e-01
-8.10062349e-01 -1.11722007e-01 -1.81676015e-01 -6.17603660e-01
-3.03885877e-01 -3.54687691e-01 -9.41235960e-01 -8.45551789e-01
-6.91208482e-01 -6.72259867e-01 5.17615676e-01 9.33291614e-01
-4.33020368e-02 3.41989636e-01 1.05347008e-01 -1.05519891e+00
-6.92215383e-01 -1.52054143e+00 -6.00512147e-01 4.06079553e-02
2.56220192e-01 -9.73014116e-01 6.22010939e-02 -2.65748203e-01] | [6.566018581390381, -0.6256058216094971] |
531aaaf0-fb89-46af-9605-a6fa5b85f60f | were-we-there-already-applying-minimal | null | null | https://aclanthology.org/2021.sigmorphon-1.29 | https://aclanthology.org/2021.sigmorphon-1.29.pdf | Were We There Already? Applying Minimal Generalization to the SIGMORPHON-UniMorph Shared Task on Cognitively Plausible Morphological Inflection | Morphological rules with various levels of specificity can be learned from example lexemes by recursive application of minimal generalization (Albright and Hayes, 2002, 2003).
A model that learns rules solely through minimal generalization was used to predict average human wug-test ratings from German, English, and Dutch in the SIGMORPHON-UniMorph 2021 Shared Task, with competitive results. Some formal properties of the minimal generalization operation were proved. An automatic method was developed to create wug-test stimuli for future experiments that investigate whether the model’s morphological generalizations are too minimal. | ['Jane S.Y. Li', 'Colin Wilson'] | null | null | null | null | acl-sigmorphon-2021-8 | ['morphological-inflection'] | ['natural-language-processing'] | [ 2.09199697e-01 3.16026449e-01 -1.81713596e-01 -1.00916660e+00
-3.22791129e-01 -8.88084531e-01 5.56011558e-01 2.72858202e-01
-8.15119267e-01 9.05106902e-01 1.05435438e-01 -5.51835299e-01
-4.39760536e-01 -6.74359322e-01 -3.63204122e-01 -1.30792081e-01
-4.35706638e-02 4.68760192e-01 4.19795036e-01 -4.97068673e-01
4.03652817e-01 6.68130279e-01 -1.56497693e+00 9.00263727e-01
1.07558823e+00 6.39599979e-01 3.39493334e-01 5.58989406e-01
1.69164374e-01 3.30118299e-01 -5.87952435e-01 -8.72405410e-01
4.65429127e-01 -6.42847478e-01 -7.39951789e-01 -2.25752115e-01
9.91357982e-01 -1.47195943e-02 1.26744986e-01 1.03478992e+00
3.20776135e-01 5.31933069e-01 1.00856757e+00 -5.18741310e-01
-1.16354787e+00 1.19974566e+00 2.36971200e-01 3.00879300e-01
4.85370010e-01 2.64778174e-02 1.39160466e+00 -1.04695344e+00
7.39998996e-01 1.25148618e+00 4.29860622e-01 1.00729239e+00
-1.35738540e+00 -6.47343278e-01 2.63987601e-01 4.10722643e-01
-1.25052536e+00 -1.64425317e-02 2.72416800e-01 -1.71085402e-01
1.50849402e+00 4.88364726e-01 5.98987699e-01 1.04457605e+00
1.89674541e-01 4.99493778e-01 1.57564187e+00 -7.09628522e-01
-6.58359155e-02 3.74757171e-01 3.57069641e-01 7.40091622e-01
6.17063880e-01 4.50241894e-01 -7.34225392e-01 9.56335291e-02
5.29512107e-01 -7.02497602e-01 -1.44806907e-01 -8.42412710e-02
-8.56222391e-01 7.24734426e-01 4.48796123e-01 4.75638598e-01
-1.42966345e-01 -4.56977993e-01 3.23582381e-01 8.85999322e-01
-1.04449587e-02 1.11565089e+00 -9.55699742e-01 2.64173418e-01
-5.63572168e-01 2.66297817e-01 8.43562722e-01 1.02092540e+00
6.36487544e-01 1.33814618e-01 -3.21076587e-02 1.07603502e+00
-2.61282548e-02 4.25173044e-01 1.09199584e+00 -8.01823556e-01
1.96455300e-01 5.98318219e-01 -2.95648307e-01 -6.47558868e-01
-6.03260279e-01 -1.42125383e-01 -1.83670789e-01 1.46204263e-01
5.06223738e-01 1.20170750e-02 -1.00063217e+00 1.95721984e+00
-1.70982495e-01 -4.94207084e-01 1.22522965e-01 4.89030808e-01
8.39772344e-01 4.29683715e-01 4.11971241e-01 -4.02406991e-01
9.10319805e-01 -3.21090251e-01 -3.87149334e-01 -3.16645235e-01
8.77141058e-01 -3.99534166e-01 1.70502543e+00 8.14332843e-01
-1.03752208e+00 -9.46894348e-01 -1.13173854e+00 3.92129347e-02
-7.27643669e-01 4.13343869e-02 9.07037437e-01 1.06423676e+00
-7.79696465e-01 8.76437008e-01 -4.22956616e-01 -6.24710619e-01
-7.10658282e-02 7.14887202e-01 -5.41914701e-01 2.09636688e-02
-1.23265445e+00 1.40036666e+00 1.09356976e+00 -1.69853680e-02
-5.65616965e-01 -5.32385886e-01 -1.04456663e+00 -1.41367599e-01
8.24210346e-02 -3.05821121e-01 1.08732080e+00 -9.68946040e-01
-1.34860146e+00 1.17866659e+00 2.96525478e-01 -3.03757310e-01
9.67888981e-02 -7.43209496e-02 -8.15921128e-01 -1.84526473e-01
2.50038095e-02 6.63425982e-01 4.17583376e-01 -9.93753135e-01
-8.60504389e-01 -3.05855513e-01 3.18015404e-02 3.02795351e-01
-5.84495246e-01 4.49916301e-03 4.51557487e-01 -8.29928994e-01
6.10116869e-02 -8.07022333e-01 9.74538103e-02 -8.13069820e-01
-1.49105877e-01 -7.08508790e-01 -2.80271955e-02 -5.83071530e-01
1.22051740e+00 -2.03767300e+00 3.78918438e-03 4.92176622e-01
-3.48771989e-01 2.77761877e-01 -4.38140303e-01 1.87986299e-01
-2.60546952e-01 2.36421898e-01 -2.99064964e-01 5.17159462e-01
2.65627146e-01 3.27169508e-01 -2.78840214e-01 1.20493947e-02
1.11370496e-01 8.96040082e-01 -9.03198004e-01 -1.84425771e-01
3.81595902e-02 -1.90305993e-01 -8.17941308e-01 -1.37093008e-01
-1.94149554e-01 -3.05023432e-01 1.87167749e-01 5.32674372e-01
2.23050579e-01 4.99362528e-01 4.68877554e-01 -3.95325422e-02
-9.77521762e-02 7.35414028e-01 -1.08774102e+00 1.35443664e+00
-4.19337302e-01 3.97279590e-01 -4.86084312e-01 -6.70763433e-01
1.00781798e+00 1.99892491e-01 -3.04764003e-01 -7.09481776e-01
1.85280427e-01 6.65393054e-01 8.09762836e-01 -3.70076537e-01
3.82432342e-01 -6.73712909e-01 -2.60479569e-01 2.05859944e-01
3.74526143e-01 -6.21422768e-01 8.17500830e-01 -1.49570107e-01
8.54361951e-01 1.62883639e-01 6.74733460e-01 -7.93172836e-01
6.11466885e-01 1.39814973e-01 6.85928285e-01 9.46876049e-01
-6.96074814e-02 9.44836661e-02 2.96627022e-02 -5.91812670e-01
-6.54493034e-01 -1.44546461e+00 -4.82434213e-01 1.59151721e+00
-3.24562907e-01 -6.20134950e-01 -5.20091236e-01 -8.34327817e-01
1.33975297e-01 1.43641019e+00 -7.37936258e-01 -3.38206768e-01
-7.12721646e-01 -6.64912820e-01 5.33912480e-01 6.02328002e-01
4.86778319e-02 -1.71049619e+00 -5.90708673e-01 2.15765938e-01
3.22267830e-01 -8.42099190e-01 -2.43259832e-01 6.77317142e-01
-1.04053271e+00 -1.03396618e+00 -5.60318343e-02 -1.17088223e+00
6.62369967e-01 -2.85671651e-01 1.08566391e+00 1.36468321e-01
-1.97745979e-01 6.30740523e-02 -4.87509906e-01 -6.07426941e-01
-5.01465917e-01 1.26074478e-01 5.04862726e-01 -5.90294302e-01
8.49084198e-01 -3.69939834e-01 1.85472108e-02 3.19941401e-01
-9.04995024e-01 -3.95193219e-01 8.47508013e-01 9.68886137e-01
4.96564299e-01 -1.56722143e-02 8.79477203e-01 -1.33420384e+00
9.93099093e-01 4.36257105e-04 -4.42151278e-01 2.75324702e-01
-7.65153646e-01 4.42890339e-02 7.95133173e-01 -7.10503042e-01
-1.23865354e+00 2.11759545e-02 -8.99025500e-02 3.40638101e-01
-3.61729205e-01 6.49205804e-01 -3.77056241e-01 -6.04432672e-02
1.27211893e+00 3.90746705e-02 -5.64676046e-01 -4.41752374e-01
3.61065060e-01 3.25499833e-01 4.93597835e-01 -9.84364688e-01
8.06985140e-01 -1.93912879e-01 -1.19120620e-01 -9.36251402e-01
-1.07109904e+00 -4.40629348e-02 -8.53415728e-01 2.40491122e-01
5.31747401e-01 -4.81005341e-01 -3.80527228e-01 -9.77384113e-03
-9.36772048e-01 -6.55729055e-01 -5.74915826e-01 7.96847343e-01
-7.00435758e-01 1.25730664e-01 -7.50254333e-01 -1.73538074e-01
5.67185618e-02 -4.74125654e-01 2.53925413e-01 2.07759347e-02
-8.68746936e-01 -9.42638516e-01 1.35432035e-01 -1.14632495e-01
5.18745147e-02 -1.99167699e-01 1.61706960e+00 -1.26679933e+00
4.49014045e-02 -1.51919097e-01 1.87452167e-01 9.00575221e-01
2.95926601e-01 -1.28080279e-01 -5.87378919e-01 -1.75103411e-01
1.50176316e-01 -6.80111289e-01 8.37731004e-01 6.54546469e-02
9.81712401e-01 -2.50156105e-01 1.71046317e-01 7.37435877e-01
1.24191940e+00 3.90877336e-01 3.39974761e-01 2.91348457e-01
3.81821007e-01 7.49277174e-01 5.42533696e-01 -1.62889197e-01
-1.00998513e-01 3.29946935e-01 -2.91051030e-01 4.45186764e-01
-2.90577024e-01 -3.73837203e-01 6.76226556e-01 1.10176802e+00
-2.88072914e-01 3.83632928e-02 -7.12189138e-01 4.55831617e-01
-1.28812194e+00 -1.05866504e+00 -1.03412598e-01 2.19948554e+00
1.09458804e+00 5.61049998e-01 -1.05043771e-02 2.46587902e-01
5.98151565e-01 -2.50184745e-01 -2.54616052e-01 -1.25906682e+00
-6.42226219e-01 7.54261792e-01 2.32719287e-01 6.84034109e-01
-6.19514287e-01 1.50140715e+00 7.59961700e+00 7.92374730e-01
-5.39486170e-01 -2.54168928e-01 1.00736976e-01 1.24523230e-02
-4.25754577e-01 -1.55298993e-01 -9.87225831e-01 -5.53627759e-02
9.99525547e-01 -4.71296102e-01 3.77094239e-01 6.43584073e-01
-3.53325933e-01 1.87406301e-01 -1.35905623e+00 5.26979923e-01
2.91888267e-01 -7.41929650e-01 5.12205780e-01 -8.68992135e-02
1.05979311e+00 -2.08115175e-01 5.32207862e-02 7.83275843e-01
4.24869478e-01 -9.83961940e-01 5.54715037e-01 1.51239604e-01
7.65874207e-01 -7.85643339e-01 4.96309191e-01 2.90011942e-01
-8.39399934e-01 -2.12211892e-01 -9.17512655e-01 -4.56430465e-01
-2.14696869e-01 6.80021867e-02 -1.01334476e+00 2.85658032e-01
3.54296565e-01 2.75307953e-01 -1.13041902e+00 7.55228043e-01
-9.80345130e-01 8.40326428e-01 -2.63479054e-01 -3.70286793e-01
7.59570673e-02 1.25720054e-02 2.43686184e-01 1.54898953e+00
2.75572628e-01 2.26064637e-01 -1.08551960e-02 6.27973795e-01
-1.40816728e-02 7.57048428e-01 -7.96371222e-01 6.62641227e-02
2.18623295e-01 9.87320721e-01 -8.61857414e-01 -2.77663141e-01
-4.72551525e-01 8.92305851e-01 4.77093041e-01 2.99715936e-01
-2.73164719e-01 -6.50151789e-01 3.26728851e-01 8.96861106e-02
2.59381264e-01 -7.07498193e-02 -5.70015788e-01 -9.18744743e-01
-7.52594471e-02 -1.11151564e+00 8.36758733e-01 -5.99230111e-01
-1.77276456e+00 6.32066369e-01 3.69043320e-01 -8.13919723e-01
-3.81875128e-01 -1.62139356e+00 -4.69035357e-01 8.66342723e-01
-7.87995696e-01 -8.38817656e-01 1.35723338e-01 6.78369105e-01
4.93384987e-01 -5.58294594e-01 1.07687557e+00 -1.32816210e-01
1.40165836e-02 8.90607536e-01 -3.34320188e-01 1.49146572e-01
8.89419317e-01 -1.69256806e+00 2.05600053e-01 8.52531612e-01
5.67585468e-01 9.73980129e-01 5.43762326e-01 -7.66508043e-01
-6.52450025e-01 -6.33482456e-01 1.31990218e+00 -8.05580616e-01
6.83458507e-01 -2.32661679e-01 -8.81266117e-01 8.79333079e-01
2.81637460e-01 -4.57813501e-01 1.06807458e+00 4.24043834e-01
-6.49662375e-01 -1.96556941e-01 -9.75941002e-01 8.57474327e-01
1.33383155e+00 -4.64539617e-01 -1.56301224e+00 2.00527593e-01
5.12847066e-01 4.46677506e-02 -7.38965034e-01 8.11278164e-01
5.72237074e-01 -9.07488585e-01 5.83163679e-01 -1.15313447e+00
2.07110316e-01 -7.75830671e-02 -4.15578336e-01 -1.78717208e+00
-5.96753120e-01 -4.23402667e-01 3.20414543e-01 9.36671853e-01
1.02267933e+00 -5.22544622e-01 4.84602809e-01 3.93486798e-01
-4.42724764e-01 -3.58347565e-01 -7.56523073e-01 -1.22750175e+00
4.85221475e-01 -4.83376473e-01 4.71559256e-01 8.65916193e-01
5.96767068e-01 6.52618408e-01 2.03979805e-01 -2.05861852e-01
2.57641613e-01 -9.94706303e-02 2.64004648e-01 -1.40385652e+00
-2.64109224e-01 -8.10990870e-01 -5.64191520e-01 -6.62408412e-01
6.09536469e-01 -1.60241711e+00 1.05579183e-01 -1.25558519e+00
-7.10298773e-03 1.47746697e-01 -7.86069930e-01 5.46924353e-01
-3.05043429e-01 1.16641492e-01 3.48908633e-01 -3.11453283e-01
-1.84630662e-01 1.13685414e-01 1.15144467e+00 8.38845670e-02
-2.39654854e-01 5.23282943e-05 -9.47280109e-01 9.07482207e-01
1.00228858e+00 -5.24830699e-01 -6.52267396e-01 -3.80325764e-01
4.23915714e-01 -6.27133727e-01 -9.30755213e-02 -9.43644345e-01
1.80008560e-01 -4.43579257e-01 6.61302626e-01 -1.19459599e-01
1.83254227e-01 -4.89710331e-01 -3.23417425e-01 5.93646169e-01
-7.35941172e-01 5.42437553e-01 5.37952721e-01 4.16834056e-02
-5.85495681e-02 -8.00532043e-01 8.53886783e-01 -2.92144001e-01
-9.02827203e-01 -8.64942372e-02 -4.90226865e-01 2.87671328e-01
7.95745790e-01 -2.39132836e-01 -1.48799613e-01 5.63107505e-02
-9.98366535e-01 -2.48897552e-01 2.77908772e-01 4.25830543e-01
7.60382116e-01 -1.39710367e+00 -9.76102769e-01 4.43422258e-01
2.75418133e-01 -7.64418185e-01 -3.09772551e-01 3.09241503e-01
-4.07492936e-01 5.44235289e-01 -6.77161992e-01 4.17618155e-02
-9.55034614e-01 6.55456781e-01 1.95070371e-01 -8.07038024e-02
-2.81287909e-01 1.12403572e+00 2.40983129e-01 -7.96346784e-01
-4.01681103e-03 -4.52759862e-01 -7.34927431e-02 7.82881901e-02
4.67842877e-01 2.54325390e-01 3.01689118e-01 -4.58838701e-01
-1.73143432e-01 2.41749823e-01 -2.52430201e-01 -4.26033646e-01
1.19345844e+00 3.71784687e-01 -5.60731031e-02 8.64882231e-01
7.50840127e-01 3.19379032e-01 -4.04864639e-01 -4.30275649e-02
5.37183642e-01 -3.82934839e-01 -5.61438024e-01 -1.31566834e+00
-3.96985441e-01 6.90820038e-01 3.07532668e-01 2.08366975e-01
9.06494915e-01 -4.47689146e-02 1.86277807e-01 1.17530346e+00
4.24984694e-01 -1.59381068e+00 -2.91672200e-01 8.35155547e-01
9.87592638e-01 -9.02757645e-01 -6.58412278e-02 -3.47769916e-01
-7.01171696e-01 1.20952654e+00 9.15335357e-01 -2.62901396e-01
5.42781532e-01 4.13772464e-02 1.34864137e-01 -4.64983322e-02
-9.81239617e-01 -3.03362280e-01 6.32630229e-01 8.47295880e-01
8.82145405e-01 3.82376254e-01 -1.20048976e+00 9.57835257e-01
-7.35499799e-01 -3.84505510e-01 4.29239124e-01 6.36684299e-01
-7.11746454e-01 -1.28938031e+00 1.75882652e-02 8.12149167e-01
-1.89852923e-01 -4.73794609e-01 -1.01695406e+00 1.28276074e+00
3.98376405e-01 7.35146403e-01 -2.40075171e-01 -4.51864123e-01
7.69279957e-01 6.20764315e-01 1.01298499e+00 -1.39916360e+00
-9.64627504e-01 -3.31248790e-01 2.31024384e-01 -2.27718040e-01
-1.58774704e-01 -8.04153204e-01 -1.40497804e+00 4.81769666e-02
-1.68156520e-01 4.16797996e-01 2.54902273e-01 8.95525694e-01
-3.48113179e-01 9.72575024e-02 9.75474566e-02 -2.96049178e-01
-9.89348948e-01 -1.35734260e+00 -8.57774436e-01 8.98610711e-01
-5.12668371e-01 -4.89973515e-01 -6.23317599e-01 6.13757223e-02] | [10.648760795593262, 9.642779350280762] |
0da68c68-fff9-437b-97cd-d838deedebd3 | neural-text-classification-by-jointly-1 | 2011.12184 | null | https://arxiv.org/abs/2011.12184v1 | https://arxiv.org/pdf/2011.12184v1.pdf | Neural Text Classification by Jointly Learning to Cluster and Align | Distributional text clustering delivers semantically informative representations and captures the relevance between each word and semantic clustering centroids. We extend the neural text clustering approach to text classification tasks by inducing cluster centers via a latent variable model and interacting with distributional word embeddings, to enrich the representation of tokens and measure the relatedness between tokens and each learnable cluster centroid. The proposed method jointly learns word clustering centroids and clustering-token alignments, achieving the state of the art results on multiple benchmark datasets and proving that the proposed cluster-token alignment mechanism is indeed favorable to text classification. Notably, our qualitative analysis has conspicuously illustrated that text representations learned by the proposed model are in accord well with our intuition. | ['Shuo Jin', 'Haidong Zhang', 'Yekun Chai'] | 2020-11-24 | neural-text-classification-by-jointly | null | null | null | ['text-clustering'] | ['natural-language-processing'] | [-1.81507081e-01 2.16114432e-01 -3.61522526e-01 -4.39247787e-01
-7.44920850e-01 -6.99590445e-01 1.10079813e+00 9.17645514e-01
-4.72619504e-01 -1.00834310e-01 7.30484426e-01 2.49859355e-02
-3.73446554e-01 -6.34991765e-01 -2.07583517e-01 -1.00739276e+00
-2.21193973e-02 7.38327861e-01 -5.59220791e-01 2.61089534e-01
4.85342443e-01 1.61367238e-01 -1.61355937e+00 2.16377020e-01
7.40657806e-01 6.32472157e-01 1.33816063e-01 2.75116593e-01
-5.77129602e-01 5.26984990e-01 -4.79121625e-01 -2.30249718e-01
-7.45932236e-02 1.80763006e-02 -1.02034080e+00 1.19755432e-01
1.78246826e-01 1.95489958e-01 -2.37470359e-01 9.91672277e-01
1.72895193e-01 5.58580279e-01 1.28478718e+00 -1.12578583e+00
-1.26401865e+00 1.19338191e+00 -6.43336535e-01 -3.69770676e-02
1.77552566e-01 -2.96383828e-01 1.81794286e+00 -1.04264176e+00
4.86213803e-01 1.28229439e+00 4.36019242e-01 4.17466938e-01
-1.11365461e+00 -3.29200059e-01 3.27365607e-01 8.43763426e-02
-1.62804925e+00 -3.41995180e-01 8.40811729e-01 -7.36892164e-01
1.00997722e+00 1.26035037e-02 1.99585557e-01 1.05534887e+00
-1.96965575e-01 7.78297722e-01 2.58496612e-01 -6.74651325e-01
4.55450535e-01 5.99473976e-02 6.49169564e-01 4.18343037e-01
2.22031772e-01 -5.09664357e-01 -3.01631391e-01 -2.53693730e-01
2.04292849e-01 4.81151730e-01 1.53385907e-01 -5.70367873e-01
-1.21118701e+00 1.14798462e+00 4.86993939e-01 8.30274284e-01
-4.65812147e-01 5.78854740e-01 6.71423912e-01 -3.19296360e-01
6.82542145e-01 3.22478622e-01 -1.17196217e-01 -4.87616174e-02
-9.22963798e-01 -3.84409167e-02 3.27127844e-01 1.05181456e+00
8.75681639e-01 1.91013124e-02 -2.77927071e-01 9.31562126e-01
7.25165486e-01 2.62441397e-01 9.04968321e-01 -6.97685182e-01
3.74850094e-01 9.32011843e-01 -2.28092432e-01 -1.21600676e+00
-3.87155354e-01 -1.48766756e-01 -8.40180576e-01 -3.34561765e-01
1.92654535e-01 2.17726380e-02 -4.27490860e-01 1.71705639e+00
8.97212103e-02 -6.97750924e-03 6.33997023e-02 6.30773306e-01
6.49315476e-01 5.06354034e-01 5.93965948e-01 -8.44495464e-03
1.38749576e+00 -6.48379982e-01 -7.06174552e-01 1.13146983e-01
1.06243753e+00 -6.49835885e-01 1.21043861e+00 -1.91833019e-01
-6.51888907e-01 -6.02125943e-01 -7.19380796e-01 -2.12484509e-01
-7.43345618e-01 1.42772362e-01 6.39064550e-01 5.55540979e-01
-1.02180922e+00 2.61748016e-01 -7.94847906e-01 -5.60959458e-01
5.07318735e-01 3.27170342e-01 -2.28153542e-01 -7.73092872e-03
-1.06518567e+00 4.56717491e-01 6.27111912e-01 -1.56606376e-01
-5.66008031e-01 -4.77889061e-01 -1.09717524e+00 4.22834188e-01
-1.64976463e-01 -6.29522383e-01 6.57080531e-01 -1.03125691e+00
-9.65132654e-01 1.02913451e+00 -4.78734702e-01 -2.34855741e-01
5.94348758e-02 3.41948010e-02 -1.78642869e-01 2.63607264e-01
3.85919720e-01 8.98820817e-01 7.50670493e-01 -1.46095073e+00
-4.11981076e-01 -5.91436803e-01 -4.87896055e-01 2.57280499e-01
-9.55056071e-01 -1.22279994e-01 -2.56575912e-01 -7.13893890e-01
5.40090986e-02 -4.19903725e-01 -2.87102461e-01 -4.51950461e-01
-4.84277874e-01 -1.03806031e+00 6.70486033e-01 -2.02346072e-01
1.19095564e+00 -2.29279923e+00 1.40172243e-01 4.73206788e-01
5.88047981e-01 -4.45008159e-01 -1.33178145e-01 7.60442913e-01
-3.05539340e-01 2.34358400e-01 -1.19553559e-01 -9.42728639e-01
6.95242882e-01 1.66757494e-01 -5.18520355e-01 7.72450686e-01
-8.62332433e-02 1.13336420e+00 -9.15115356e-01 -6.20435894e-01
4.28257436e-01 6.30754590e-01 -5.20761847e-01 -9.80170816e-03
-2.70077467e-01 -2.24076375e-01 -5.01438975e-01 3.23710501e-01
3.09247971e-01 -3.04102808e-01 5.68014562e-01 1.53421447e-01
1.65121406e-02 5.04511967e-02 -9.37123120e-01 1.72797287e+00
-2.11655721e-01 8.94393027e-01 -3.30796719e-01 -1.39737797e+00
1.06577253e+00 3.70356530e-01 7.40310133e-01 -3.47462744e-01
2.77659237e-01 -3.44221354e-01 -4.70376909e-01 -5.96082449e-01
9.08187211e-01 -6.14585616e-02 -3.36189121e-01 8.07728946e-01
3.10632437e-01 4.04760480e-01 -1.53958261e-01 6.54637933e-01
6.48594439e-01 -1.94836020e-01 3.45597476e-01 -6.49525940e-01
2.21641049e-01 -3.97424847e-02 -1.45207420e-01 6.06098175e-01
-1.77266821e-01 3.86874735e-01 5.48169136e-01 -1.65390581e-01
-1.03916597e+00 -1.06623161e+00 -2.43080810e-01 1.70381439e+00
6.05591834e-02 -7.17955291e-01 -9.00202036e-01 -6.09677196e-01
1.52782664e-01 8.14784467e-01 -1.13161802e+00 -8.41040611e-02
-1.85495839e-01 -5.78279853e-01 7.21185803e-01 8.74588668e-01
-2.62839109e-01 -1.00159454e+00 -8.35882425e-02 -5.12780994e-03
-2.09734172e-01 -7.52427995e-01 -4.72878665e-01 4.63520467e-01
-5.15553057e-01 -1.05863988e+00 -4.38492715e-01 -1.36569834e+00
9.89565194e-01 4.26140010e-01 9.82011914e-01 1.72063291e-01
-3.73682857e-01 7.73120761e-01 -6.18076026e-01 9.87444371e-02
-1.70521095e-01 3.26367319e-01 1.62398547e-01 5.98476157e-02
9.96182323e-01 -2.71127731e-01 -5.03880024e-01 -1.42987713e-01
-1.23336327e+00 -4.93162841e-01 1.18471190e-01 7.48139024e-01
3.26262683e-01 2.08754405e-01 7.23569989e-01 -7.92470455e-01
6.96277261e-01 -9.89774406e-01 -1.77130118e-01 4.23434973e-02
-7.30064213e-01 1.67709991e-01 7.27343261e-01 -2.29370713e-01
-8.12323689e-01 6.84244782e-02 1.60755113e-01 -3.45688015e-01
-6.18389726e-01 5.31096876e-01 -2.21550301e-01 7.94244468e-01
6.09966576e-01 1.96005940e-01 -2.89098650e-01 -2.91443020e-01
1.21897793e+00 9.22209442e-01 5.54781258e-01 -8.62385452e-01
6.83863580e-01 7.80662239e-01 -3.18398029e-01 -8.14122975e-01
-6.05977416e-01 -1.08756542e+00 -1.34004736e+00 8.51733759e-02
1.30116749e+00 -9.20518637e-01 -9.24769282e-01 -5.03836684e-02
-1.07567847e+00 -8.43936354e-02 -4.21191812e-01 3.06263447e-01
-4.72091675e-01 7.01870441e-01 -5.18359959e-01 -6.30941451e-01
-2.54776508e-01 -9.11849022e-01 1.18510330e+00 -2.07898349e-01
-5.77322841e-01 -1.75520456e+00 2.37686887e-01 2.26439655e-01
6.90389648e-02 1.39797017e-01 1.35771370e+00 -1.12133634e+00
-5.70684597e-02 -3.57306600e-01 -3.56467366e-01 -1.65266186e-01
3.85065615e-01 1.62421793e-01 -1.12537956e+00 -2.73053586e-01
-3.33049983e-01 -1.29790738e-01 1.13713324e+00 4.37039167e-01
1.22032154e+00 -3.72158676e-01 -5.29532075e-01 2.35742614e-01
1.54487288e+00 -2.33518243e-01 3.20552588e-01 2.87578225e-01
1.03939295e+00 8.21244001e-01 1.30649522e-01 7.38268852e-01
5.36594450e-01 5.07197142e-01 3.26099545e-01 -4.20511439e-02
3.44049126e-01 -2.52748251e-01 2.57528633e-01 9.86667156e-01
5.09095132e-01 -3.03591341e-01 -1.22635841e+00 8.43728721e-01
-2.15210629e+00 -1.03880703e+00 -2.93753058e-01 1.53485537e+00
6.27601624e-01 -2.21900076e-01 1.06703892e-01 6.82616979e-02
9.95648265e-01 1.53659686e-01 -1.35094225e-01 -3.28981996e-01
-2.52871327e-02 1.38009906e-01 2.51418918e-01 5.38129568e-01
-1.11952817e+00 1.25344574e+00 6.77897501e+00 9.46648359e-01
-5.27582586e-01 1.16090052e-01 5.04232824e-01 1.35573789e-01
-6.45801067e-01 -8.78457874e-02 -5.57035685e-01 3.29771161e-01
7.55166709e-01 -2.93133855e-01 2.18967572e-01 9.33432162e-01
2.90918887e-01 3.70654941e-01 -1.36753225e+00 8.37425053e-01
3.22430730e-01 -1.48785961e+00 5.97813725e-01 1.56694978e-01
8.10579181e-01 -2.46015504e-01 3.65125060e-01 2.55970746e-01
7.12080717e-01 -1.28382230e+00 5.85757852e-01 3.55308086e-01
6.25326574e-01 -1.09135139e+00 6.57786369e-01 -2.40128245e-02
-1.36645746e+00 -5.61931431e-02 -5.30334711e-01 -1.13523051e-01
-4.29623514e-01 3.25432539e-01 -9.08861995e-01 4.25178409e-01
3.99640620e-01 9.78527784e-01 -7.96021402e-01 4.17985409e-01
-8.51897895e-02 6.09833002e-01 2.53201395e-01 -1.60342306e-01
5.92211008e-01 -2.25635305e-01 2.11290047e-01 1.77142382e+00
-1.55966699e-01 -3.37357134e-01 1.56174034e-01 1.26915860e+00
-2.39101216e-01 4.21088010e-01 -6.58498287e-01 -1.85276702e-01
9.02207315e-01 1.42255723e+00 -1.15295303e+00 -5.59329987e-01
-1.43971995e-01 7.66881704e-01 5.31225383e-01 4.78965580e-01
-6.01774991e-01 -5.55276155e-01 8.43976200e-01 -3.07939947e-01
3.91434193e-01 -3.62382978e-01 -7.21304893e-01 -1.02809751e+00
-3.02457064e-01 -1.94542855e-01 4.96882200e-01 -5.11649489e-01
-1.63418937e+00 3.03619295e-01 -5.59104718e-02 -9.39438641e-01
-3.16161752e-01 -5.14547765e-01 -8.47669065e-01 5.11618614e-01
-1.13844097e+00 -1.27920747e+00 -1.72078788e-01 8.21736336e-01
5.03132403e-01 -1.89428613e-01 1.15865827e+00 2.85438653e-02
-6.17759287e-01 9.34179187e-01 7.74118602e-01 6.54420137e-01
5.89045942e-01 -1.64452302e+00 2.99652278e-01 5.53705513e-01
4.54955548e-01 1.07770669e+00 4.61932659e-01 -3.56565654e-01
-1.24424064e+00 -1.18968320e+00 8.19671631e-01 -7.75601327e-01
9.73342001e-01 -7.38533437e-01 -8.18741143e-01 7.55939364e-01
5.24026573e-01 -3.82043839e-01 1.28406096e+00 4.22362059e-01
-7.30645835e-01 3.88151258e-01 -8.46453786e-01 5.33223152e-01
6.73796892e-01 -8.62702727e-01 -9.26641881e-01 2.67747700e-01
1.00587380e+00 5.19680023e-01 -8.48579466e-01 -5.04870594e-01
2.05315277e-01 -3.29020917e-01 8.57077479e-01 -7.00842321e-01
5.75010955e-01 -9.36486647e-02 -5.15544832e-01 -1.21761167e+00
-5.20733237e-01 -1.43117532e-01 2.55990267e-01 1.86649585e+00
1.36237681e-01 -3.95751089e-01 7.62395442e-01 4.75636959e-01
-1.53066561e-01 -2.88133621e-01 -8.99575770e-01 -3.90362114e-01
6.78818166e-01 -4.99013186e-01 4.70715821e-01 1.78521872e+00
6.75327480e-01 3.42651755e-01 2.64105529e-01 6.11222312e-02
8.99976730e-01 9.86398906e-02 4.58399445e-01 -1.34662032e+00
1.13945983e-01 -1.05446184e+00 -4.28656340e-01 -9.02529001e-01
1.03366113e+00 -1.60439432e+00 5.94471432e-02 -1.55799544e+00
4.77337867e-01 -4.29559618e-01 -4.37354118e-01 4.54606652e-01
-3.29484135e-01 3.82000655e-02 -6.10439405e-02 4.11978990e-01
-1.04162323e+00 6.21204555e-01 5.15937090e-01 -3.76302242e-01
-6.06322810e-02 -6.03504956e-01 -1.02756751e+00 5.74235737e-01
6.28032625e-01 -5.51086187e-01 -1.44186363e-01 -4.77718711e-01
3.80125403e-01 -6.13638580e-01 2.38751575e-01 -3.39336395e-01
4.35351193e-01 8.26049149e-02 5.68040311e-01 -4.49956089e-01
-3.41232195e-02 -7.96676457e-01 -6.22964740e-01 6.05278909e-02
-1.04633224e+00 -1.51387870e-01 -1.02438703e-01 8.84891033e-01
-1.60914987e-01 -2.42217213e-01 6.73122406e-01 7.95224234e-02
-6.55972004e-01 8.63806903e-02 -8.81377220e-01 2.23540097e-01
9.62493896e-01 -3.55817139e-01 -1.74090058e-01 -4.12427224e-02
-7.98548818e-01 3.10704142e-01 5.15930176e-01 5.25330782e-01
5.77965260e-01 -1.56483674e+00 -5.06755769e-01 1.51221529e-01
4.62154835e-01 -6.94577545e-02 -1.74625944e-02 2.70236224e-01
1.40017802e-02 6.05338871e-01 2.28978351e-01 -7.45686948e-01
-9.30030227e-01 8.77224982e-01 7.02375397e-02 -2.42607854e-02
-7.47157395e-01 5.01761436e-01 3.53570312e-01 -5.73171794e-01
4.88221169e-01 -2.60810673e-01 -5.08797050e-01 6.57797396e-01
1.92728058e-01 3.44100833e-01 -1.34462789e-01 -9.01477814e-01
-3.50310981e-01 5.79467535e-01 -1.88218113e-02 -1.21296169e-02
1.31505811e+00 -3.74348879e-01 -2.59936810e-01 6.29265249e-01
1.60205996e+00 -2.01969758e-01 -8.54233742e-01 -4.44157779e-01
6.21114254e-01 -1.10431612e-01 -4.58690003e-02 -1.22416139e-01
-8.30930471e-01 1.06175363e+00 3.35291296e-01 3.00592989e-01
4.38072413e-01 4.85836953e-01 4.99759525e-01 4.75731403e-01
-2.19695017e-01 -1.16513228e+00 5.23522198e-01 4.79926348e-01
2.40822315e-01 -1.01582861e+00 -3.14199299e-01 2.08818227e-01
-7.84441650e-01 1.19625056e+00 2.88524330e-01 -3.48205060e-01
7.64419198e-01 -4.47399020e-02 5.44494912e-02 -5.86395621e-01
-7.87057042e-01 -2.90418327e-01 2.87998110e-01 7.40154624e-01
7.15117157e-01 3.18632960e-01 1.36890203e-01 7.21760690e-01
-2.03951091e-01 -8.88217330e-01 2.48313516e-01 4.91834372e-01
-6.82544708e-01 -9.32737827e-01 -3.32680464e-01 1.89015359e-01
-3.49150062e-01 -2.49282613e-01 -6.18490636e-01 5.99957168e-01
-4.67185006e-02 1.09442985e+00 7.28416264e-01 -5.04684150e-02
-2.23225027e-01 3.83944213e-01 -3.80301252e-02 -8.54499698e-01
-6.56586170e-01 8.99982452e-02 -6.56557739e-01 -9.83919576e-02
-3.99272889e-01 -7.22568274e-01 -1.87267220e+00 -2.79627740e-01
-4.16722149e-01 4.70686972e-01 6.51656091e-01 1.03886020e+00
3.11632335e-01 4.05798048e-01 8.16728830e-01 -8.05580139e-01
-2.88784295e-01 -9.19416070e-01 -7.24062026e-01 8.91467214e-01
1.48171857e-01 -4.80066121e-01 -5.50428808e-01 3.22875112e-01] | [10.415709495544434, 6.796196937561035] |
68222bfb-b245-495d-b338-a83cad2653e2 | direction-of-arrival-estimation-for-a-vector | 2004.05671 | null | https://arxiv.org/abs/2004.05671v1 | https://arxiv.org/pdf/2004.05671v1.pdf | Direction of Arrival Estimation for a Vector Sensor Using Deep Neural Networks | A vector sensor, a type of sensor array with six collocated antennas to measure all electromagnetic field components of incident waves, has been shown to be advantageous in estimating the angle of arrival and polarization of the incident sources. While angle estimation with machine learning for linear arrays has been well studied, there has not been a similar solution for the vector sensor. In this paper, we propose neural networks to determine the number of the sources and estimate the angle of arrival of each source, based on the covariance matrix extracted from received data. Also, we provide a solution for matching output angles to corresponding sources and examine the error distributions with this method. The results show that neural networks can achieve reasonably accurate estimation with up to 5 sources, especially if the field-of-view is limited. | ['Jianyuan Yu', 'Daniel Tait', 'R. Michael Buehrer', 'William W. Howard'] | 2020-04-12 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [ 2.77698636e-01 -2.86307156e-01 1.99139029e-01 -4.15623069e-01
-5.65881789e-01 -5.12928843e-01 -2.32404470e-02 -2.11926788e-01
-1.57409906e-01 4.68111902e-01 9.37312469e-03 -4.04358238e-01
-6.89919412e-01 -7.64948845e-01 -6.65795982e-01 -9.79486525e-01
-3.51341188e-01 2.12785810e-01 -4.84561652e-01 3.15171108e-02
1.74698234e-01 8.44473958e-01 -1.14047956e+00 -1.97060257e-01
3.18799168e-01 1.60176218e+00 7.04983063e-03 6.91732347e-01
2.62379676e-01 5.33617377e-01 -8.89223993e-01 3.91458161e-02
2.64794648e-01 -1.80482537e-01 6.56530485e-02 -1.78814948e-01
5.12694240e-01 -3.25931907e-01 -2.38558680e-01 9.29144025e-01
7.11659014e-01 -3.54622751e-01 1.00180113e+00 -1.29051590e+00
-1.28158346e-01 5.30147910e-01 -4.45040077e-01 1.55274779e-01
1.29473820e-01 -5.31812608e-01 3.84764493e-01 -8.09359133e-01
1.76358371e-04 6.71036482e-01 9.45827782e-01 -6.40371442e-02
-5.34630060e-01 -6.65957987e-01 -6.11845374e-01 2.30851099e-01
-1.15485072e+00 -4.70619977e-01 1.33019221e+00 -4.46151435e-01
7.51139343e-01 2.32276484e-01 5.39477885e-01 7.10028112e-01
4.89361227e-01 1.52796730e-01 8.37398648e-01 -8.88241887e-01
3.88751805e-01 1.37560323e-01 9.95575115e-02 4.72285628e-01
6.17121875e-01 2.43616268e-01 -3.62127423e-01 -2.21878082e-01
7.71681011e-01 -6.13747500e-02 -4.03929681e-01 -7.40824580e-01
-1.06630027e+00 9.72690463e-01 4.89072561e-01 5.46269596e-01
-6.95113838e-01 1.34228408e-01 -2.50498682e-01 7.78083950e-02
4.17917132e-01 9.28236663e-01 -5.57604671e-01 3.80215138e-01
-6.65637791e-01 -1.74412001e-02 1.02103186e+00 7.30323851e-01
5.64837813e-01 8.15076470e-01 3.40445340e-01 8.17193508e-01
5.18841326e-01 1.39804816e+00 -9.85281989e-02 -9.98799682e-01
4.12462473e-01 2.22381428e-01 3.30561996e-01 -1.72801137e+00
-8.88355494e-01 -8.36970806e-01 -1.11313379e+00 3.76455098e-01
5.60990453e-01 -1.27999938e+00 -6.32928550e-01 1.52105749e+00
9.93759185e-02 9.63855237e-02 3.38394105e-01 9.01917696e-01
6.19246066e-01 8.62433195e-01 -5.91708541e-01 -4.64645058e-01
8.94356906e-01 -4.31319177e-01 -8.68002534e-01 -6.58963144e-01
1.76601112e-01 -7.56888449e-01 -4.12243992e-01 5.99361122e-01
-8.35005045e-01 -2.24854589e-01 -1.34801483e+00 9.68044579e-01
-2.10688859e-01 3.91785800e-01 7.37616301e-01 6.01409495e-01
-6.32318854e-01 1.36657134e-01 -6.53333426e-01 1.26865461e-01
-5.79038486e-02 1.19734220e-01 -2.60795265e-01 -7.36776413e-03
-1.07589233e+00 1.10349524e+00 1.11261137e-01 4.32633430e-01
-4.19342630e-02 -6.79227829e-01 -7.88943350e-01 -3.39034246e-04
-1.90646112e-01 -3.00632894e-01 9.23977971e-01 -8.05852592e-01
-1.24758303e+00 -1.18651338e-01 -2.52208740e-01 -1.98216155e-01
-3.90713841e-01 5.68865389e-02 -9.32354569e-01 3.24315965e-01
9.20463204e-02 1.03190830e-02 8.76691699e-01 -1.23494709e+00
-7.02512145e-01 -4.72490489e-01 -4.24844682e-01 8.34903717e-02
-4.17589188e-01 -2.14298323e-01 7.63492808e-02 -1.78505704e-01
7.61955798e-01 -7.32070982e-01 -5.21764517e-01 -1.79603830e-01
-1.12641282e-01 6.96833879e-02 6.59918070e-01 -6.48764670e-01
9.65435266e-01 -2.00865269e+00 -2.01118171e-01 7.16740727e-01
1.57795087e-01 6.67676330e-02 6.11970713e-03 5.41447759e-01
-3.14312845e-01 -6.93039596e-01 7.66490623e-02 3.60095620e-01
-4.94119287e-01 -2.47024387e-01 -1.72212243e-01 6.81217670e-01
-9.22157317e-02 4.31313097e-01 -4.43297267e-01 2.59524465e-01
3.13577682e-01 5.68392336e-01 -2.05166027e-01 4.56723511e-01
3.58813107e-01 1.42886728e-01 -4.81573552e-01 5.74284434e-01
9.16235864e-01 -4.16575938e-01 1.07249498e-01 -8.38757336e-01
-3.54243852e-02 -3.27884436e-01 -1.82183874e+00 9.24419403e-01
-6.88166678e-01 1.02737999e+00 4.70736176e-01 -1.31980300e+00
1.24963033e+00 6.26824319e-01 8.52300942e-01 -7.39139140e-01
2.51630038e-01 1.21134058e-01 2.53357682e-02 -8.38059723e-01
-2.01211303e-01 -5.02863573e-03 5.55287255e-03 3.41706842e-01
1.04315713e-01 -7.27520958e-02 4.50314581e-02 -2.65422672e-01
9.90569770e-01 -5.64832985e-01 5.15712380e-01 1.00605212e-01
2.70550046e-03 -3.43819568e-03 3.94788206e-01 7.94965684e-01
3.68533969e-01 3.91326696e-01 -1.62994221e-01 -5.53593755e-01
-8.29486609e-01 -9.07848001e-01 -2.61296064e-01 4.42635953e-01
2.60294467e-01 2.38470644e-01 -4.12575603e-01 8.29635635e-02
9.22789425e-02 8.38318288e-01 -4.92320620e-02 2.05368862e-01
-5.94447017e-01 -9.85590577e-01 1.48349538e-01 6.71636641e-01
1.03670828e-01 -5.53951263e-01 -9.17787611e-01 3.36109221e-01
-1.05989918e-01 -1.13948190e+00 3.36401105e-01 6.58294559e-01
-5.60283482e-01 -1.14475298e+00 -4.95925307e-01 -6.48469329e-01
8.38273764e-01 3.13031912e-01 7.79263139e-01 -7.44477987e-01
-1.49220582e-02 7.03065753e-01 -1.26072153e-01 -1.09890556e+00
6.27642199e-02 -5.67279875e-01 3.34322959e-01 1.03577375e-01
1.80865675e-01 -7.39727616e-01 -3.96071792e-01 3.49154383e-01
-4.66237694e-01 -2.69669652e-01 8.16932261e-01 4.41541106e-01
-1.25383381e-02 2.82375067e-01 6.17381692e-01 -3.57584029e-01
4.85038698e-01 -6.32450998e-01 -9.43706512e-01 -6.59731328e-02
-5.52314341e-01 -7.84559697e-02 5.57197690e-01 -3.40017915e-01
-8.85943055e-01 9.54452604e-02 -2.73409456e-01 -2.06857622e-01
-4.13796484e-01 7.29290724e-01 -3.03394962e-02 -6.52803183e-01
7.13796258e-01 -2.02441677e-01 -5.01066670e-02 -2.35265017e-01
-3.83718614e-03 7.75598049e-01 3.73493135e-01 1.20396331e-01
9.73356783e-01 2.27027699e-01 5.10645688e-01 -1.22275996e+00
-8.03117752e-01 -7.34475434e-01 -2.12162063e-01 -3.59916657e-01
5.76767504e-01 -9.62510586e-01 -6.20466769e-01 5.90132952e-01
-1.46609473e+00 3.49699169e-01 2.52793163e-01 1.27198625e+00
-2.10438073e-01 -1.76268950e-01 -1.36136055e-01 -8.76976073e-01
-3.25872332e-01 -9.61275160e-01 5.67866504e-01 3.38971674e-01
-9.12013352e-02 -1.06473732e+00 1.49368092e-01 -7.21565038e-02
8.17439556e-01 2.32926041e-01 6.30059600e-01 -5.92689455e-01
-3.96187961e-01 -6.61071360e-01 1.22874714e-02 3.75895321e-01
1.39752716e-01 -5.49115896e-01 -6.76956773e-01 -2.86362410e-01
4.48564947e-01 9.71443057e-02 1.16884172e-01 1.14336586e+00
6.62905276e-01 -4.79075670e-01 -7.32651234e-01 1.00989211e+00
2.03145576e+00 8.33714366e-01 3.22823495e-01 1.08621880e-01
6.22198105e-01 3.37968707e-01 1.66874513e-01 3.74267429e-01
-1.71220392e-01 4.20097739e-01 5.68049788e-01 -2.66225278e-01
3.39052618e-01 2.77513504e-01 -3.13984185e-01 8.57233882e-01
1.13477297e-01 -8.01096737e-01 -7.55521178e-01 3.52826446e-01
-1.22763717e+00 -8.43878865e-01 -3.50504637e-01 1.74657774e+00
6.20931759e-02 -8.42433274e-02 -4.58938897e-01 4.86891836e-01
4.54973161e-01 2.29438856e-01 -4.14745390e-01 -1.72534406e-01
-8.88210312e-02 1.19973853e-01 1.05545413e+00 6.64569378e-01
-9.85714674e-01 -5.76166771e-02 7.38050938e+00 3.47670406e-01
-1.55093384e+00 -2.66522735e-01 1.51187509e-01 1.61608964e-01
-2.92690635e-01 -2.40426585e-01 -8.88555229e-01 3.72341871e-01
7.07693815e-01 2.04566732e-01 2.25393325e-01 7.60553956e-01
-8.23204294e-02 -2.46996939e-01 -9.66430068e-01 1.21177721e+00
5.05484104e-01 -1.18784809e+00 -3.84542137e-01 -2.90048212e-01
6.20129824e-01 -1.41648158e-01 -7.79197644e-03 -2.44018197e-01
-8.41944143e-02 -8.22724462e-01 2.59068578e-01 7.71371484e-01
4.03591663e-01 -6.69772565e-01 8.46445084e-01 6.54416442e-01
-7.19033420e-01 -2.61476904e-01 -2.64300168e-01 -4.00297910e-01
4.14044797e-01 1.24104786e+00 -1.14522219e+00 4.23439503e-01
6.93017960e-01 2.38571897e-01 -1.19617425e-01 1.19334483e+00
-1.22910179e-01 1.01982498e+00 -8.57713044e-01 -6.01918995e-01
-5.84882218e-03 -2.51732469e-01 7.94095278e-01 8.34779203e-01
1.04947841e+00 3.13701093e-01 -4.63717170e-02 3.61588120e-01
4.36781764e-01 -1.97128892e-01 -8.67255330e-01 3.20321381e-01
6.47910655e-01 1.30903447e+00 -3.39896262e-01 1.67687073e-01
-4.53830689e-01 1.73067391e-01 -9.95760933e-02 8.86566758e-01
-2.47394204e-01 -7.98589110e-01 3.27994972e-01 2.20098794e-01
5.89835107e-01 -3.34568083e-01 -3.72884363e-01 -4.30210382e-01
1.99218407e-01 -6.03962481e-01 -1.33984998e-01 -1.13886428e+00
-1.27388668e+00 4.59167093e-01 1.37689769e-01 -1.42934370e+00
-9.03967202e-01 -1.03419197e+00 -6.25835836e-01 9.49246049e-01
-1.10649621e+00 -7.51619756e-01 -1.81611955e-01 2.08779112e-01
-2.57146835e-01 -5.40154099e-01 8.94485593e-01 2.78943628e-01
-5.03372811e-02 2.24155068e-01 5.48161209e-01 4.83102888e-01
4.49732691e-01 -8.88925076e-01 -2.71822780e-01 8.93364310e-01
1.12593487e-01 4.01093394e-01 1.23407638e+00 -2.61943996e-01
-1.78368735e+00 -7.62187719e-01 5.97675085e-01 -2.76253641e-01
4.36532825e-01 -4.23869714e-02 -2.73532540e-01 5.41096866e-01
4.89357710e-01 1.18596047e-01 6.11089528e-01 3.81145887e-02
-2.62463111e-02 -6.54331505e-01 -7.97031820e-01 1.26024842e-01
4.91841018e-01 1.54182792e-01 -3.72705430e-01 3.51432979e-01
1.36761129e-01 -4.83799249e-01 -8.27344358e-01 9.09358501e-01
5.91994464e-01 -9.83504176e-01 1.18610275e+00 -5.13071977e-02
2.06037581e-01 -1.95902154e-01 -4.21069384e-01 -1.89970350e+00
-7.38544345e-01 1.00024275e-01 -8.57745707e-02 8.05277228e-01
4.68563884e-01 -9.53468859e-01 8.01540136e-01 -1.07995041e-01
1.47840776e-03 -8.60995710e-01 -9.37639058e-01 -4.52894419e-01
-3.79300535e-01 -5.48384428e-01 3.06011051e-01 8.15149963e-01
-1.69369340e-01 7.09423125e-01 -5.26121676e-01 1.11308753e+00
1.04725266e+00 2.94946462e-01 4.96469378e-01 -1.25352824e+00
-3.66869479e-01 -7.16520622e-02 -3.58960301e-01 -1.04266083e+00
8.10297430e-02 -5.17399967e-01 2.37278998e-01 -1.87434685e+00
-5.42053998e-01 -6.22911334e-01 -1.06799997e-01 2.58301422e-02
8.05696025e-02 1.97116196e-01 -2.32344955e-01 -4.08660114e-01
3.01783215e-02 -1.31104914e-02 7.63980329e-01 -2.36874580e-01
1.46226242e-01 6.61737502e-01 -6.18337452e-01 1.09528923e+00
7.34807014e-01 -5.69188595e-01 -3.38537961e-01 -8.88920903e-01
6.59078419e-01 5.43118000e-01 1.43229201e-01 -1.60381377e+00
7.12221980e-01 -2.14180369e-02 1.14380062e+00 -8.97643149e-01
4.29739207e-01 -1.35002148e+00 5.23208559e-01 2.40064591e-01
-9.52889845e-02 3.02888416e-02 6.97902357e-03 4.34883446e-01
-3.48467678e-01 -5.68574488e-01 6.61617279e-01 1.30083680e-01
-5.63809574e-01 9.23916250e-02 -5.51873982e-01 -2.01388180e-01
8.42967987e-01 -1.80893824e-01 1.76365655e-02 -8.07128489e-01
-1.26316488e-01 -6.62451088e-02 -4.55297381e-01 7.84463622e-03
7.01773167e-01 -1.47068441e+00 -7.67370164e-01 6.08391702e-01
2.79722735e-02 -2.76279122e-01 1.59639567e-01 2.66978770e-01
-3.30160916e-01 6.77889109e-01 -1.81804538e-01 -6.85620904e-01
-1.06820798e+00 2.65336931e-01 5.60455978e-01 1.64229780e-01
1.49840161e-01 9.43431079e-01 -3.25887024e-01 -4.86668587e-01
2.13673338e-01 1.69945285e-02 -7.37391114e-01 6.25874922e-02
5.13314664e-01 4.47571248e-01 1.83567613e-01 -5.08305430e-01
-3.51285845e-01 1.14694118e+00 6.89430118e-01 -1.11437112e-01
1.47978401e+00 1.34713173e-01 -1.02574438e-01 3.91187280e-01
1.47082341e+00 4.42265511e-01 -8.03746998e-01 -2.12472320e-01
-5.11554182e-01 -1.51491627e-01 3.36042792e-01 -8.00358772e-01
-1.14074886e+00 7.40650713e-01 9.84199107e-01 3.45810145e-01
1.29080319e+00 -1.65753603e-01 3.84729713e-01 8.84042442e-01
5.09407580e-01 -6.50804758e-01 -1.25720367e-01 3.91214222e-01
5.88075101e-01 -1.13047075e+00 -1.17531382e-02 -3.50089699e-01
9.89044458e-02 1.19918323e+00 4.04778451e-01 -1.65476903e-01
1.06581104e+00 1.03693533e+00 4.96246338e-01 -1.84445009e-01
-2.54628479e-01 5.08529127e-01 4.00089860e-01 8.52091432e-01
2.15938732e-01 1.13631397e-01 1.07461162e-01 2.16073588e-01
-7.98582062e-02 -2.28448704e-01 4.80598778e-01 8.47397149e-01
-7.45732427e-01 -6.44910693e-01 -1.07040393e+00 7.36989200e-01
-5.24310291e-01 6.87604100e-02 3.06776702e-01 2.89047688e-01
1.31207705e-01 1.20806670e+00 2.97954470e-01 -1.95139036e-01
5.70555866e-01 -2.98915446e-01 5.38015664e-01 -9.13578551e-03
2.09232301e-01 3.48848142e-02 1.50869325e-01 -3.19668680e-01
-4.25480068e-01 -2.98892707e-01 -7.66548276e-01 3.42419505e-01
-5.62239766e-01 5.04859090e-01 1.15053701e+00 8.84108067e-01
1.28300205e-01 5.83547294e-01 9.82740641e-01 -9.05314445e-01
-5.77838361e-01 -8.89239073e-01 -7.95090914e-01 -1.02022454e-01
6.57390714e-01 -5.36273181e-01 -8.19382966e-01 -5.24914488e-02] | [6.517349720001221, 1.3100148439407349] |
c4d59f3d-6d89-435e-a946-4bb32ac9e5f5 | a-sequential-algorithm-for-training-text | cmp-lg/9407020 | null | https://arxiv.org/abs/cmp-lg/9407020v2 | https://arxiv.org/pdf/cmp-lg/9407020v2.pdf | A Sequential Algorithm for Training Text Classifiers | The ability to cheaply train text classifiers is critical to their use in information retrieval, content analysis, natural language processing, and other tasks involving data which is partly or fully textual. An algorithm for sequential sampling during machine learning of statistical classifiers was developed and tested on a newswire text categorization task. This method, which we call uncertainty sampling, reduced by as much as 500-fold the amount of training data that would have to be manually classified to achieve a given level of effectiveness. | ['William A. Gale', 'David D. Lewis'] | 1994-07-24 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 5.78295887e-01 3.53578866e-01 -5.61330616e-01 -9.79542613e-01
-1.11938727e+00 -5.17994285e-01 1.04749846e+00 8.12069416e-01
-9.60653722e-01 9.75382686e-01 6.73269853e-02 -6.53068244e-01
1.43913832e-02 -6.26571476e-01 -4.71613556e-01 -2.77688622e-01
2.44116232e-01 7.22617269e-01 1.31738096e-01 1.59819514e-01
5.38847387e-01 1.90343738e-01 -1.56158257e+00 5.62136531e-01
5.86525798e-01 1.22383058e+00 -9.10904910e-03 4.60030526e-01
-5.40027797e-01 6.69248939e-01 -6.51268065e-01 -3.97052497e-01
-1.73661292e-01 -1.50049806e-01 -1.16433227e+00 8.58424082e-02
2.69239366e-01 -5.02685197e-02 1.44113138e-01 9.94412422e-01
1.25957251e-01 4.63148922e-01 1.15076101e+00 -9.12424386e-01
7.52427801e-02 1.19293070e+00 -3.53514135e-01 4.51818973e-01
3.66286308e-01 -4.75246280e-01 1.13002670e+00 -6.82729781e-01
4.06016439e-01 1.25117862e+00 3.48699838e-01 4.05271798e-01
-1.33364522e+00 -5.47172606e-01 1.91095233e-01 -2.05152199e-01
-1.09724569e+00 -5.07362962e-01 4.47219998e-01 -5.20277560e-01
8.05360973e-01 2.07013905e-01 3.32913995e-01 9.16836560e-01
3.67258340e-01 7.13466167e-01 1.04694474e+00 -7.65239418e-01
5.89799285e-01 6.78047180e-01 5.44152081e-01 5.30443251e-01
3.18665802e-01 -3.52773368e-01 -3.88137907e-01 -3.93315375e-01
-9.37698036e-02 -2.70888746e-01 6.33956566e-02 1.40098140e-01
-8.53419721e-01 1.46481228e+00 -7.89920613e-03 3.83841813e-01
-1.52433679e-01 -3.74907814e-02 6.75677836e-01 5.90068996e-01
1.09949684e+00 8.08629572e-01 -8.58260036e-01 -2.00342983e-01
-1.01403236e+00 2.12776959e-01 9.00855422e-01 1.05313206e+00
4.03316975e-01 -2.05025271e-01 -3.66975591e-02 1.10675514e+00
8.68430659e-02 2.08721057e-01 8.05463314e-01 -7.18950689e-01
4.52387363e-01 7.62660563e-01 1.23143375e-01 -5.32228291e-01
-4.92993385e-01 3.84938493e-02 -6.54965758e-01 -1.64255530e-01
3.58007073e-01 -2.16287091e-01 -8.80260527e-01 1.26441491e+00
2.30851278e-01 -7.66221523e-01 -1.71961725e-01 3.14744294e-01
6.42147660e-01 6.25586748e-01 3.70847099e-02 -3.91035348e-01
1.37761605e+00 -4.34834331e-01 -5.12462556e-01 -3.90666038e-01
7.88379252e-01 -7.63182104e-01 9.90862310e-01 7.26788163e-01
-8.98425996e-01 -2.59791970e-01 -1.01515222e+00 8.04359615e-02
-5.91254473e-01 -1.14641249e-01 6.91383541e-01 6.39741123e-01
-3.58210355e-01 6.27701223e-01 -5.09687722e-01 2.02630628e-02
6.49276674e-01 2.91265815e-01 -3.04363728e-01 -1.93451956e-01
-1.36325419e+00 1.04009748e+00 6.80284023e-01 -3.38317662e-01
-4.59858090e-01 -5.60442567e-01 -8.41972530e-01 1.40825152e-01
4.18457955e-01 -1.52670756e-01 1.69913137e+00 -9.47428346e-01
-1.19678640e+00 8.35718095e-01 -1.18518300e-01 -6.35481954e-01
6.41196907e-01 -1.43807113e-01 -1.32370457e-01 -1.21512979e-01
2.87059303e-02 5.99124491e-01 1.06958485e+00 -6.80573106e-01
-8.63385439e-01 -4.28020090e-01 -3.28907907e-01 1.38200879e-01
-3.30545902e-01 3.84125680e-01 -6.20588511e-02 -7.38996565e-01
8.55172798e-02 -8.52264345e-01 -2.01900050e-01 -2.65966475e-01
-3.07906747e-01 -6.74675047e-01 7.28966475e-01 -4.41212565e-01
8.15452754e-01 -1.78143251e+00 -1.82656962e-02 5.61521888e-01
8.63384306e-02 -1.11177191e-01 3.07132989e-01 1.66127697e-01
2.02649832e-01 4.64793295e-01 -9.74257886e-02 -1.41680511e-02
-1.27717659e-01 1.49024846e-02 -4.49842960e-01 2.94061303e-01
1.48730323e-01 3.60541821e-01 -7.59468377e-01 -6.07561588e-01
6.09585233e-02 6.03101100e-04 -4.03038740e-01 5.44465904e-04
-6.01463735e-01 -1.21956065e-01 -4.74495858e-01 3.02369028e-01
9.22433734e-02 -1.25237092e-01 -6.39826432e-03 1.47285923e-01
2.53009647e-01 7.05527425e-01 -1.09935343e+00 1.17318618e+00
-6.90133393e-01 9.17522371e-01 -1.37816861e-01 -1.11160696e+00
6.88172340e-01 2.81752735e-01 4.86818045e-01 -2.82784492e-01
6.24911308e-01 3.94838639e-02 -1.89548388e-01 -2.27080286e-01
3.69034052e-01 -2.93801755e-01 -3.51968795e-01 7.33033240e-01
5.94854876e-02 -4.83595252e-01 4.24342006e-01 3.58081549e-01
9.60355461e-01 -6.06495678e-01 4.18517530e-01 -6.82291925e-01
2.96690643e-01 3.43812913e-01 1.22595087e-01 9.93327260e-01
4.26502042e-02 1.03966311e-01 5.81568956e-01 -4.90743250e-01
-1.00262928e+00 -2.85357505e-01 -6.85434937e-01 1.53088748e+00
-3.21054995e-01 -3.34125787e-01 -7.45428264e-01 -8.42900932e-01
7.02158660e-02 1.11933541e+00 -6.16443813e-01 -1.17417842e-01
4.78530228e-02 -6.59719348e-01 2.28466555e-01 2.79432058e-01
9.62395072e-02 -9.79679346e-01 -5.82515180e-01 4.03138191e-01
-1.77892298e-01 -1.08624709e+00 -4.20710504e-01 6.93331182e-01
-9.77437079e-01 -1.08863056e+00 -3.31323832e-01 -5.08971453e-01
5.43132842e-01 -5.74645102e-02 1.05358446e+00 -1.06127463e-01
-4.39632118e-01 1.82779238e-01 -6.68200850e-01 -1.10763431e+00
-7.46673048e-01 4.60987329e-01 1.77769840e-01 -3.35517138e-01
9.35315490e-01 2.23582685e-02 -1.14446416e-01 1.10073440e-01
-1.06747365e+00 -8.61918107e-02 4.72424328e-01 1.35707688e+00
1.68681532e-01 5.56631684e-01 6.86875463e-01 -1.39805388e+00
9.49843943e-01 -3.62533182e-01 -5.17153263e-01 1.67642340e-01
-9.53420699e-01 5.38720846e-01 4.99082774e-01 -5.98183036e-01
-1.05067909e+00 1.70525596e-01 -7.97239598e-03 2.78268784e-01
1.94382423e-03 8.69945049e-01 3.47426116e-01 1.64472789e-01
9.85117555e-01 -3.99820693e-02 2.91772150e-02 -3.44528109e-01
2.51269877e-01 1.24745166e+00 -2.13012144e-01 -6.43172503e-01
2.50267178e-01 1.14565872e-01 -3.41880918e-01 -9.09184694e-01
-1.32685637e+00 -6.57892048e-01 -6.85681641e-01 -1.80823654e-02
2.98899412e-01 -5.31527758e-01 -2.76948571e-01 1.05186515e-01
-9.06303942e-01 -3.16203654e-01 -1.68925658e-01 5.68445086e-01
-2.93889701e-01 -1.03842668e-01 -2.91640460e-01 -8.22860241e-01
-4.55012769e-01 -6.50319993e-01 8.98871124e-01 -4.12813798e-02
-5.67549169e-01 -8.42677236e-01 -3.63255799e-01 3.04753065e-01
3.54867578e-01 -2.80122161e-01 1.21163666e+00 -1.29174864e+00
2.49031335e-01 -8.91821504e-01 -1.61245257e-01 3.09177548e-01
2.44892333e-02 1.07482530e-01 -8.45049858e-01 -2.89386272e-01
1.61072731e-01 -8.91589046e-01 1.00087893e+00 4.80744809e-01
1.52967596e+00 -5.62527180e-01 -3.82840872e-01 -3.70390445e-01
1.07848763e+00 3.30650717e-01 -3.61421704e-02 9.70595330e-02
1.28630400e-01 9.56345856e-01 5.58335602e-01 5.52024662e-01
-2.22485483e-01 4.90679651e-01 -3.99618030e-01 3.14327955e-01
4.68610644e-01 -4.72181179e-02 -1.41271085e-01 4.26510215e-01
5.56959152e-01 -5.03497124e-02 -9.94270504e-01 3.75467658e-01
-1.54440451e+00 -9.73736048e-01 3.44845116e-01 2.09090877e+00
1.19881237e+00 9.19700801e-01 -1.96049903e-02 4.81456906e-01
6.86908007e-01 -1.99328497e-01 -5.24833798e-01 -8.08631063e-01
5.41704178e-01 1.24450184e-01 5.67790449e-01 5.73866963e-01
-1.31201077e+00 7.21326530e-01 7.34071684e+00 8.74948919e-01
-9.40690637e-01 -6.13137856e-02 1.04690039e+00 -1.04151100e-01
-1.51910633e-01 -2.97452271e-01 -9.87690687e-01 4.71841305e-01
1.25422466e+00 -3.19105148e-01 2.88392872e-01 1.28895485e+00
9.34089124e-02 -5.76294661e-01 -1.36165106e+00 7.25147665e-01
1.68398246e-02 -1.32619488e+00 1.33891061e-01 -2.15354711e-01
5.52148938e-01 3.23198102e-02 -3.86532366e-01 4.91229713e-01
5.74437022e-01 -1.04858279e+00 6.63518190e-01 -4.74551655e-02
9.86589253e-01 -6.98679030e-01 9.89345014e-01 7.45422304e-01
-5.51851213e-01 -2.66984016e-01 -2.67265111e-01 6.89534005e-03
-2.00332284e-01 9.81079102e-01 -1.07218122e+00 -1.75786391e-01
5.39589703e-01 3.13268378e-02 -5.29129207e-01 7.44738042e-01
1.36209562e-01 8.55615735e-01 -5.97516358e-01 -6.65701628e-01
2.95511901e-01 4.02207822e-01 1.40847147e-01 1.25314784e+00
-2.03955755e-01 1.23197928e-01 2.30906472e-01 2.53851742e-01
-2.99784124e-01 3.38465214e-01 -5.17168641e-01 -7.75276124e-02
7.78016090e-01 8.23607862e-01 -9.38481987e-01 -6.67239904e-01
-2.75132686e-01 4.57898110e-01 2.65605628e-01 -8.48971084e-02
-3.44691873e-01 -7.55523980e-01 -1.40700132e-01 1.45103469e-01
4.13143635e-03 8.32418054e-02 -4.92826760e-01 -1.07618284e+00
-8.50433558e-02 -8.60013008e-01 5.83586216e-01 -3.61204445e-01
-1.59959042e+00 7.68361628e-01 5.70518315e-01 -8.14517558e-01
-6.83156133e-01 -8.50126743e-01 -3.53930503e-01 7.47601628e-01
-9.26968038e-01 -4.19827610e-01 -9.67859030e-02 2.33372614e-01
9.38911259e-01 -5.41792750e-01 8.56610298e-01 -5.28717637e-02
-3.36370438e-01 4.84876633e-01 3.00515026e-01 1.41687766e-01
6.01172805e-01 -1.36021388e+00 1.61198124e-01 3.04024935e-01
2.08462775e-01 3.02754611e-01 9.68087494e-01 -3.55482966e-01
-8.69723082e-01 -9.60319340e-01 1.04212964e+00 -4.67599511e-01
5.23749411e-01 -4.72374618e-01 -6.62205160e-01 5.64964056e-01
-1.45040274e-01 -2.52231240e-01 7.39907265e-01 5.79657793e-01
-5.58864772e-01 -2.47479960e-01 -1.33323729e+00 3.98992002e-01
2.70520687e-01 -4.99362439e-01 -9.75772440e-01 6.21627331e-01
4.87710297e-01 -2.24679455e-01 -7.12901235e-01 1.02795385e-01
6.55763924e-01 -1.66562602e-01 3.26108634e-01 -9.03257072e-01
4.63556081e-01 4.33764160e-01 -2.60717999e-02 -1.54263461e+00
5.46076149e-02 -3.53642374e-01 1.60113856e-01 9.17115033e-01
8.85739565e-01 -5.63576341e-01 7.78594017e-01 1.15500331e+00
3.28990012e-01 -5.73746443e-01 -8.83459806e-01 -4.74575341e-01
2.71173120e-01 -5.97392142e-01 1.73657104e-01 9.60602760e-01
5.50857484e-01 8.93438697e-01 1.65107146e-01 -6.86867177e-01
6.08561754e-01 4.26622760e-03 2.99558431e-01 -1.71723008e+00
5.00832237e-02 -7.38362849e-01 -3.04393739e-01 -6.33817971e-01
1.97266638e-01 -7.90500700e-01 3.97706687e-01 -1.17830575e+00
2.71692395e-01 -5.82632720e-01 2.84897834e-02 4.94944155e-01
-1.36995211e-01 -1.35072440e-01 -3.85348022e-01 9.96706635e-02
-2.34777763e-01 3.16854209e-01 4.86762613e-01 -5.29092073e-01
-2.69281477e-01 3.05369020e-01 -8.74964356e-01 6.99352324e-01
8.22175503e-01 -9.22293365e-01 -5.32921135e-01 4.93590534e-02
3.47103208e-01 -2.54282169e-02 -3.66969436e-01 -5.40787280e-01
1.40888050e-01 -2.94934213e-01 5.88615477e-01 -4.52340603e-01
1.19683579e-01 -8.46463442e-01 -2.95589358e-01 3.76245469e-01
-1.39000082e+00 -1.28145024e-01 2.72691190e-01 6.13839090e-01
-2.97310352e-01 -8.53507698e-01 9.10915911e-01 -1.95685178e-01
-2.58438468e-01 -2.13327277e-02 -9.09692585e-01 2.05139965e-01
1.13800061e+00 3.76671374e-01 -2.17337027e-01 -3.52702945e-01
-4.81694400e-01 1.39665768e-01 -2.14577205e-02 3.79543304e-01
3.05845857e-01 -8.80967915e-01 -8.24631810e-01 1.75196350e-01
3.21379602e-01 3.86423841e-02 -1.80882558e-01 2.63839215e-01
-4.34424341e-01 7.45264411e-01 1.10344030e-01 -3.60872418e-01
-1.26468301e+00 5.62145233e-01 6.12213984e-02 -2.55105168e-01
-6.06471241e-01 8.15572917e-01 -5.64746797e-01 -1.94645479e-01
4.97180402e-01 -3.45254481e-01 -3.35524082e-01 2.61038870e-01
8.34484875e-01 1.59308642e-01 5.67598701e-01 -1.08633548e-01
-2.30761528e-01 -1.70811862e-01 -6.71641350e-01 -5.28174162e-01
9.65596259e-01 2.51275212e-01 8.55478197e-02 7.82292902e-01
1.19103491e+00 -3.25392544e-01 -7.19057500e-01 -5.39785504e-01
5.17088056e-01 -4.63079244e-01 6.84932470e-01 -8.30637276e-01
-3.11031103e-01 7.76149690e-01 1.87276572e-01 8.07887256e-01
6.80644751e-01 6.93580136e-02 2.94098705e-01 1.18899679e+00
2.99779266e-01 -1.63653278e+00 -2.28933737e-01 3.71458024e-01
8.78659904e-01 -1.64113939e+00 4.30991501e-01 -1.84566483e-01
-6.72678053e-01 1.12172842e+00 2.68287897e-01 1.45069078e-01
1.15786541e+00 3.05808544e-01 -2.88501322e-01 -1.71172529e-01
-1.22911334e+00 2.08341107e-01 5.11395693e-01 1.27735183e-01
5.81990540e-01 6.69570491e-02 -3.57344806e-01 3.85007590e-01
-1.20865598e-01 -7.24529326e-02 3.91923308e-01 1.00574613e+00
-8.21941674e-01 -6.87400579e-01 -1.85731649e-01 1.42679656e+00
-7.86151707e-01 -1.52183533e-01 -5.41769326e-01 4.01391745e-01
-3.92521322e-01 1.17968869e+00 2.95108229e-01 -2.74439782e-01
-6.69089332e-02 5.47375202e-01 2.35418126e-01 -8.49378169e-01
-3.42027783e-01 -5.46856597e-02 5.03768623e-01 -3.25622447e-02
-3.19063932e-01 -7.20535874e-01 -8.21278870e-01 -2.67743081e-01
-7.00836003e-01 5.80639839e-01 1.02079737e+00 1.36469126e+00
-1.32890239e-01 1.99256063e-01 7.41290510e-01 -6.96337402e-01
-9.73102629e-01 -1.50225949e+00 -4.31066930e-01 2.53414541e-01
6.33764639e-02 -6.05506122e-01 -6.82377875e-01 1.58176988e-01] | [10.557442665100098, 8.013365745544434] |
3ed19591-ce79-44f7-a90d-03ac9f310ed1 | attention-mechanisms-for-object-recognition | 1807.09480 | null | http://arxiv.org/abs/1807.09480v2 | http://arxiv.org/pdf/1807.09480v2.pdf | Attention Mechanisms for Object Recognition with Event-Based Cameras | Event-based cameras are neuromorphic sensors capable of efficiently encoding
visual information in the form of sparse sequences of events. Being
biologically inspired, they are commonly used to exploit some of the
computational and power consumption benefits of biological vision. In this
paper we focus on a specific feature of vision: visual attention. We propose
two attentive models for event based vision: an algorithm that tracks events
activity within the field of view to locate regions of interest and a
fully-differentiable attention procedure based on DRAW neural model. We
highlight the strengths and weaknesses of the proposed methods on four
datasets, the Shifted N-MNIST, Shifted MNIST-DVS, CIFAR10-DVS and N-Caltech101
collections, using the Phased LSTM recognition network as a baseline reference
model obtaining improvements in terms of both translation and scale invariance. | ['Marco Cannici', 'Matteo Matteucci', 'Andrea Romanoni', 'Marco Ciccone'] | 2018-07-25 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 6.69659317e-01 -2.82874823e-01 3.19144517e-01 -3.35854292e-01
-1.96765825e-01 -4.23024356e-01 1.10023272e+00 1.09468186e-02
-8.66136849e-01 6.51121736e-01 3.27037215e-01 3.14511359e-01
-2.75903523e-01 -4.89731163e-01 -1.03433955e+00 -9.28147793e-01
-1.04287947e-02 -2.91365534e-02 6.67451560e-01 2.99109638e-01
7.23895550e-01 9.13127363e-01 -1.70182252e+00 6.21178448e-01
9.73366797e-02 1.24886644e+00 5.02790630e-01 7.75719762e-01
2.20710114e-01 1.20378602e+00 -5.57851791e-01 3.27387750e-02
9.23039615e-02 -5.98974884e-01 -3.82619172e-01 -1.62395611e-01
6.00009084e-01 -1.19414695e-01 -9.95463610e-01 9.78714108e-01
5.07798672e-01 3.80918831e-01 6.94774747e-01 -1.06897032e+00
-7.83661366e-01 6.49655908e-02 -2.35414982e-01 1.15543199e+00
2.11121470e-01 3.81408870e-01 6.45292580e-01 -8.92699718e-01
7.68207014e-01 1.08657348e+00 4.68743086e-01 6.35620773e-01
-1.07946682e+00 -1.82955533e-01 6.84518814e-02 7.18631685e-01
-9.32662606e-01 -6.37219548e-01 6.51383340e-01 -3.45192730e-01
1.85484636e+00 -2.38059998e-01 9.95454669e-01 1.67024028e+00
8.75593841e-01 6.54497683e-01 1.08557832e+00 -2.38351092e-01
7.67990232e-01 -3.78936857e-01 2.15503559e-01 5.15531600e-01
1.27171546e-01 3.05395424e-01 -1.28799200e+00 2.40519606e-02
9.85164464e-01 5.88679969e-01 -3.11609983e-01 -1.85554117e-01
-1.44538438e+00 5.84896564e-01 6.88813448e-01 2.66062886e-01
-8.87197971e-01 6.20159388e-01 1.90367475e-01 -1.01888835e-01
-1.56234384e-01 2.68020540e-01 -3.14172447e-01 8.47673267e-02
-8.65253448e-01 -1.64227337e-01 4.16620433e-01 7.74954379e-01
6.04172587e-01 4.46789593e-01 -4.78536189e-01 4.21802163e-01
4.66775477e-01 5.21106720e-01 9.37627971e-01 -9.78725314e-01
-2.16097385e-01 5.91903865e-01 -8.31784382e-02 -6.50629938e-01
-3.30364764e-01 -8.34923163e-02 -6.44695580e-01 4.08876687e-01
1.82500601e-01 2.29291663e-01 -1.31445074e+00 1.65550852e+00
-2.99975753e-01 5.42000532e-01 2.79603809e-01 7.66133070e-01
7.97976792e-01 7.07015097e-01 2.36819476e-01 -8.94835070e-02
1.48589647e+00 -8.29234898e-01 -4.48073566e-01 -7.00681031e-01
-3.44933748e-01 -2.21418649e-01 6.87944293e-01 3.52360755e-01
-1.14572847e+00 -6.15589678e-01 -1.02945805e+00 -2.80155599e-01
-6.62852466e-01 -3.36612575e-02 5.92309356e-01 6.21088929e-02
-1.36438787e+00 6.22479439e-01 -1.13364518e+00 -8.14316511e-01
8.58042240e-01 5.01164854e-01 -2.54986614e-01 2.45552838e-01
-6.66289449e-01 8.17843020e-01 4.40185368e-01 -2.04072222e-01
-1.63110411e+00 -5.55390835e-01 -6.97576404e-01 4.39853996e-01
-4.10347670e-01 -7.97148049e-01 1.05413222e+00 -1.06223130e+00
-1.18026292e+00 1.05069196e+00 -4.39467937e-01 -1.04247975e+00
-1.71796843e-01 -5.19705899e-02 -2.11800054e-01 6.79275990e-01
-2.01843396e-01 1.18960977e+00 8.46532285e-01 -5.48663259e-01
-6.92808628e-01 -5.63434124e-01 -1.64251372e-01 4.62306198e-03
-3.39959800e-01 1.64778665e-01 -1.44109756e-01 -5.76726019e-01
3.30760591e-02 -6.88601911e-01 -1.01520017e-01 2.11060017e-01
5.66690192e-02 -2.58903682e-01 7.47625470e-01 -3.38861138e-01
3.14108163e-01 -2.09898448e+00 1.82093740e-01 -4.60932583e-01
-5.38685769e-02 2.33662963e-01 -1.84790611e-01 2.76293218e-01
-9.05896798e-02 -4.03508514e-01 -2.80847400e-01 -1.11241817e-01
-3.73315215e-01 4.07359183e-01 -7.19526768e-01 5.20226955e-01
6.64671242e-01 1.16957152e+00 -7.83945441e-01 -2.64372408e-01
4.74399984e-01 7.99908757e-01 -2.25051701e-01 -1.32983983e-01
-2.73777872e-01 3.55872095e-01 -6.99031772e-03 8.22466195e-01
1.32031124e-02 -2.76406020e-01 -2.19877377e-01 -3.63119900e-01
-3.01737428e-01 2.39137203e-01 -5.68423748e-01 2.08924150e+00
4.37016152e-02 1.10337007e+00 -5.55637479e-01 -1.04935360e+00
7.27888942e-01 3.89161527e-01 3.29730093e-01 -1.15489769e+00
4.41087544e-01 -1.78330109e-01 -5.71079133e-03 -3.85379970e-01
8.13385099e-02 1.42255142e-01 2.65722603e-01 -1.48360077e-02
8.29548955e-01 3.32932979e-01 1.24140307e-01 3.67777273e-02
1.37512684e+00 1.41748115e-01 3.28535289e-01 -2.09505513e-01
3.09049994e-01 -1.14858426e-01 4.04239714e-01 8.43225837e-01
-5.04970074e-01 7.03162313e-01 1.49119720e-01 -5.03141522e-01
-8.41529489e-01 -1.39740276e+00 4.43495624e-02 9.72270787e-01
3.77719514e-02 2.43314385e-01 -4.68467176e-01 -3.32575649e-01
-2.28776664e-01 6.93053365e-01 -8.56400788e-01 -4.60705996e-01
-5.52119195e-01 -8.47943604e-01 4.74915147e-01 8.00906837e-01
5.39358616e-01 -1.76716292e+00 -1.63698721e+00 2.11487949e-01
2.54334956e-01 -1.28751278e+00 -3.50504965e-02 8.48551452e-01
-9.69028652e-01 -1.05403829e+00 -6.43774688e-01 -1.02648830e+00
5.81540227e-01 2.78098166e-01 9.38975036e-01 -7.67947972e-01
-7.02965736e-01 8.34363043e-01 7.73192048e-02 -8.70154679e-01
2.89316446e-01 -4.71613258e-01 -7.55930543e-02 2.57035673e-01
6.86691463e-01 -1.04590189e+00 -9.06704307e-01 -1.98613897e-01
-8.15596163e-01 -2.07138270e-01 7.26707160e-01 5.98697603e-01
9.85012412e-01 -6.51112616e-01 5.07582426e-01 -5.28533876e-01
1.71952248e-01 -4.35969204e-01 -5.50684631e-01 1.33847803e-01
-3.07711959e-01 -4.13866574e-03 6.27291203e-01 -6.66028321e-01
-9.69312131e-01 5.26032031e-01 2.45154828e-01 -8.20108891e-01
-3.93203914e-01 8.28367993e-02 1.71498999e-01 -3.77470732e-01
8.22651625e-01 8.03702772e-01 -6.54535174e-01 -1.06348835e-01
9.34662446e-02 1.02093473e-01 1.02612996e+00 -5.81376348e-03
1.05200544e-01 1.00965202e+00 1.03598163e-01 -8.91792476e-01
-4.88894045e-01 -5.22214413e-01 -5.67644715e-01 -3.34910333e-01
1.01937962e+00 -8.73331368e-01 -6.73605442e-01 7.64334202e-01
-1.36588526e+00 -3.24645370e-01 -7.23692775e-01 5.39453387e-01
-7.76706219e-01 -3.07539664e-02 -6.90668225e-01 -7.65922368e-01
-4.20328766e-01 -6.88317239e-01 1.03183889e+00 9.36917424e-01
-9.24747612e-04 -9.76548433e-01 2.69272178e-01 -3.58070701e-01
3.60537171e-01 3.89376551e-01 6.84494853e-01 -6.17497027e-01
-6.63899899e-01 1.14468575e-01 -2.66856611e-01 5.64838424e-02
-1.30123779e-01 -1.76784143e-01 -1.49654984e+00 -1.00954384e-01
3.89580488e-01 -3.44941080e-01 1.41860831e+00 8.03024828e-01
1.15490794e+00 -4.10126522e-02 -3.28970402e-01 7.19132364e-01
1.79316437e+00 6.76286399e-01 1.00256872e+00 1.41582623e-01
2.56960720e-01 2.26629168e-01 -1.54050902e-01 4.07925934e-01
2.62073189e-01 2.02917367e-01 8.20475638e-01 2.56093889e-01
-4.64798838e-01 -2.21185863e-01 6.36281550e-01 3.24046373e-01
-1.10734031e-01 -1.97849795e-01 -7.42400348e-01 1.17504942e+00
-1.92857778e+00 -1.14401138e+00 1.89721614e-01 2.07021809e+00
5.00634789e-01 3.40490267e-02 -2.59890318e-01 -7.54028000e-03
4.83449489e-01 1.60527676e-01 -1.08215845e+00 -4.20129508e-01
-5.66815019e-01 5.47952414e-01 4.77774948e-01 -1.76647693e-01
-8.67740273e-01 7.91272402e-01 6.28496504e+00 5.88135272e-02
-1.30503058e+00 2.30925828e-01 2.96914518e-01 -5.76119721e-01
3.74007642e-01 -2.23349795e-01 -8.32085907e-01 5.50816178e-01
1.58463669e+00 -1.75988879e-02 6.23432159e-01 5.14700055e-01
5.07669663e-03 -3.00195873e-01 -1.43882787e+00 1.32343900e+00
4.47098643e-01 -1.60138047e+00 6.72367588e-02 -1.14518389e-01
6.47352397e-01 7.25869775e-01 1.46607161e-01 9.37640220e-02
1.20840751e-01 -9.86723065e-01 8.73814702e-01 1.02893245e+00
4.93182451e-01 -3.46911013e-01 1.76257253e-01 9.93635058e-02
-9.12210107e-01 -2.84682184e-01 -5.85855007e-01 -9.77815017e-02
1.63019933e-02 2.11644620e-01 -3.42289984e-01 -9.80393291e-02
1.25609994e+00 1.11403728e+00 -7.27028310e-01 1.43450963e+00
-2.30357617e-01 6.02500916e-01 -2.15734735e-01 -1.08977161e-01
2.15468675e-01 2.18313038e-01 5.16043425e-01 1.26641357e+00
4.56254691e-01 8.97178352e-02 -3.13786566e-01 1.00793660e+00
-1.61205322e-01 -7.14361310e-01 -7.97189832e-01 -1.66107878e-01
3.08196932e-01 1.34225118e+00 -1.03787780e+00 -3.00643831e-01
-4.74878043e-01 1.29474115e+00 2.83003539e-01 4.56292570e-01
-8.09649706e-01 -4.37331885e-01 4.20298815e-01 -3.02925348e-01
9.60648835e-01 -2.07724392e-01 -1.42383039e-01 -1.05297232e+00
-3.03957820e-01 -4.79443967e-01 3.22094977e-01 -1.23060513e+00
-1.04281306e+00 5.15560627e-01 -3.37201536e-01 -7.13814318e-01
-2.60715246e-01 -9.79951382e-01 -7.26620555e-01 5.88605762e-01
-1.71962881e+00 -9.98725772e-01 -4.38858390e-01 9.71273959e-01
7.23579705e-01 -1.98156312e-01 8.64758193e-01 8.68759975e-02
-4.01786536e-01 6.06184565e-02 -4.94557060e-02 3.56283337e-02
5.71229100e-01 -1.03986120e+00 4.77916658e-01 1.13186359e+00
7.90893376e-01 5.12526095e-01 4.95304972e-01 -2.33161658e-01
-1.71616089e+00 -1.35399377e+00 7.60969102e-01 -4.97423977e-01
4.27851260e-01 -4.25470710e-01 -8.38758349e-01 9.20483708e-01
5.02868950e-01 3.64912361e-01 3.01787049e-01 -6.46970093e-01
-3.74145269e-01 -2.01865017e-01 -1.35645854e+00 4.42467630e-01
1.12013113e+00 -7.30577230e-01 -1.03938746e+00 3.25585961e-01
2.88251698e-01 -4.16042693e-02 -5.12411416e-01 6.17882349e-02
4.45163012e-01 -1.08924258e+00 1.11995459e+00 -4.83864844e-01
2.19243005e-01 -2.06259996e-01 -4.30026919e-01 -9.68525469e-01
-5.87360919e-01 -3.28570753e-01 -6.14455104e-01 1.06188583e+00
-1.08493567e-01 -6.41849995e-01 6.17991924e-01 1.34349279e-02
-2.64062583e-01 -4.35449272e-01 -1.14705944e+00 -6.59129262e-01
-4.89376009e-01 -1.08918943e-01 -1.66394025e-01 2.79363453e-01
-3.68667006e-01 5.06358624e-01 3.07935476e-02 3.27900767e-01
7.19160736e-01 -7.32674599e-02 -1.05864562e-01 -1.38482845e+00
-1.32536098e-01 -3.62025350e-01 -7.70723999e-01 -7.36359715e-01
2.95854248e-02 -7.41758168e-01 2.19846338e-01 -1.64444077e+00
4.16245192e-01 6.72560155e-01 -9.74638999e-01 6.51181757e-01
4.88261551e-01 5.89406490e-01 4.06871230e-04 1.95825502e-01
-7.56074846e-01 4.08241332e-01 4.46203291e-01 -8.67816508e-02
1.38896629e-01 -5.67673564e-01 -5.41375577e-01 8.64351571e-01
6.91639602e-01 -5.67713618e-01 -4.11751062e-01 -6.86130881e-01
6.36015832e-02 -1.15694411e-01 1.10592234e+00 -1.66915512e+00
8.58153999e-01 -2.00625341e-02 9.81310427e-01 -4.84751016e-01
5.69738567e-01 -7.20798850e-01 -9.70350876e-02 7.34507918e-01
-3.55466515e-01 2.69289076e-01 4.58741188e-01 1.03986764e+00
-6.72536120e-02 -1.56324744e-01 1.08465028e+00 -3.61759186e-01
-1.23879802e+00 3.68970752e-01 -7.56377876e-01 2.20423657e-03
1.06881464e+00 -4.91986722e-01 -5.88666439e-01 1.64841384e-01
-3.93174261e-01 -2.21341535e-01 2.71640331e-01 4.62463915e-01
1.04409933e+00 -1.01272225e+00 -2.77358383e-01 3.51720452e-01
6.92944080e-02 -4.47302669e-01 1.30556270e-01 6.77833021e-01
-2.23975226e-01 8.64605069e-01 -1.00161397e+00 -8.05463612e-01
-9.16008294e-01 7.83187032e-01 5.30834377e-01 2.49892935e-01
-6.87655628e-01 1.10882437e+00 3.81245434e-01 2.59260237e-01
3.55739087e-01 -6.36728585e-01 -3.98505718e-01 -9.43981856e-02
6.87432945e-01 2.81985521e-01 1.30780622e-01 -4.47498143e-01
-7.35728383e-01 4.06111360e-01 1.52055129e-01 -2.86970258e-01
1.57362044e+00 -6.18868768e-02 -9.75669995e-02 7.39924550e-01
7.77768612e-01 -5.92693985e-01 -1.78592753e+00 -2.07311764e-01
7.00283423e-02 1.32486984e-01 1.26461908e-01 -9.77779031e-01
-1.08438599e+00 1.23223448e+00 1.01852965e+00 -1.06398106e-01
1.40952051e+00 -3.21496390e-02 6.47001386e-01 6.26829088e-01
3.05589706e-01 -7.88406193e-01 3.84122074e-01 6.38606489e-01
8.83740902e-01 -1.00922668e+00 -4.13690209e-01 3.94548774e-01
-3.76491487e-01 1.09333909e+00 4.64213699e-01 -7.93135703e-01
4.07491177e-01 2.09141761e-01 -4.44734216e-01 -3.14781666e-01
-1.09078085e+00 -2.32074469e-01 2.02491105e-01 8.48048985e-01
5.77121228e-02 -5.31804919e-01 1.96720898e-01 6.52610660e-01
3.10611725e-01 3.54646951e-01 4.81729776e-01 9.02586997e-01
-6.12702250e-01 -3.06517810e-01 -6.94283620e-02 5.06159246e-01
-5.32267332e-01 -2.95318782e-01 -5.18765748e-01 2.98243880e-01
1.37933522e-01 6.99638069e-01 3.25435311e-01 -3.94073278e-02
3.50815535e-01 3.36399287e-01 1.00977814e+00 -4.93069798e-01
-7.08080292e-01 -2.68968821e-01 -5.03697634e-01 -6.80938542e-01
-6.98360145e-01 -7.75760770e-01 -1.45895517e+00 3.00764114e-01
3.27869147e-01 -5.77786922e-01 6.71276033e-01 8.73851657e-01
7.52879441e-01 9.65779006e-01 1.23573281e-01 -1.02427578e+00
-2.20020339e-01 -8.15015316e-01 -3.00515741e-01 3.34634811e-01
3.50686669e-01 -5.63006580e-01 -2.12026775e-01 6.02272213e-01] | [8.252243041992188, 2.2564008235931396] |
86450c69-f797-46fe-9a1c-313b5198129c | saarland-vector-based-models-of-semantic | null | null | https://aclanthology.org/S12-1089 | https://aclanthology.org/S12-1089.pdf | Saarland: Vector-based models of semantic textual similarity | null | ['Georgiana Dinu', 'Stefan Thater'] | 2012-07-01 | null | null | null | semeval-2012-7 | ['video-description'] | ['computer-vision'] | [-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] | [-7.3042192459106445, 3.6937756538391113] |
b9ed9431-226b-4239-b991-74f96f42f0e9 | ordered-tree-decomposition-for-hrg-rule | null | null | https://aclanthology.org/J19-2005 | https://aclanthology.org/J19-2005.pdf | Ordered Tree Decomposition for HRG Rule Extraction | We present algorithms for extracting Hyperedge Replacement Grammar (HRG) rules from a graph along with a vertex order. Our algorithms are based on finding a tree decomposition of smallest width, relative to the vertex order, and then extracting one rule for each node in this structure. The assumption of a fixed order for the vertices of the input graph makes it possible to solve the problem in polynomial time, in contrast to the fact that the problem of finding optimal tree decompositions for a graph is NP-hard. We also present polynomial-time algorithms for parsing based on our HRGs, where the input is a vertex sequence and the output is a graph structure. The intended application of our algorithms is grammar extraction and parsing for semantic representation of natural language. We apply our algorithms to data annotated with Abstract Meaning Representations and report on the characteristics of the resulting grammars. | ['Xiaochang Peng', 'Giorgio Satta', 'Daniel Gildea'] | 2019-06-01 | null | null | null | cl-2019-6 | ['tree-decomposition'] | ['graphs'] | [ 8.87693524e-01 1.07339799e+00 -3.99789326e-02 -2.31383339e-01
-4.21782196e-01 -8.80654335e-01 6.78051859e-02 4.16664094e-01
-2.27150340e-02 5.26392579e-01 -9.26198736e-02 -8.76892090e-01
-1.99896380e-01 -1.58551323e+00 -7.78090715e-01 -2.51591057e-01
-2.67046303e-01 6.48554683e-01 5.74291110e-01 -1.29263207e-01
6.54648989e-02 5.20197272e-01 -1.58534217e+00 1.44019574e-01
5.24883568e-01 3.18680912e-01 3.67461950e-01 1.12065256e+00
-8.56240988e-01 3.46172690e-01 -2.86079168e-01 -5.53160965e-01
3.66765559e-01 -6.63140118e-01 -1.42784214e+00 5.28489411e-01
2.03464493e-01 2.80993223e-01 -1.11246407e-01 1.20446897e+00
-2.09337741e-01 -1.68021306e-01 3.81572336e-01 -1.52104056e+00
-1.28002807e-01 9.74850655e-01 -3.04003745e-01 -1.06035113e-01
7.74248600e-01 -4.47874457e-01 1.60528243e+00 -9.41584408e-02
9.62489903e-01 1.28996897e+00 2.36493021e-01 5.93581438e-01
-1.17288244e+00 1.68869086e-03 3.69347900e-01 -2.82470196e-01
-1.00875497e+00 -1.91955432e-01 4.87590373e-01 -2.13698223e-01
1.37988901e+00 4.82674360e-01 5.79520166e-01 2.54092336e-01
9.47302207e-02 4.04051274e-01 5.87350428e-01 -1.21976745e+00
2.96834528e-01 -3.08446556e-01 7.16068327e-01 1.26793849e+00
7.09350288e-01 7.82010611e-03 4.39463444e-02 -3.58950555e-01
7.23657072e-01 -4.47596550e-01 1.97554901e-02 -4.52652335e-01
-7.49130666e-01 9.17128026e-01 -1.39012694e-01 2.91443646e-01
-1.32243738e-01 3.43400508e-01 4.76436198e-01 6.18243575e-01
-4.34710830e-02 1.40122175e-01 -5.79282820e-01 2.68907458e-01
-2.97067970e-01 2.57766634e-01 1.23045254e+00 1.43554211e+00
8.62995505e-01 -1.57376200e-01 2.60925859e-01 3.92240465e-01
1.96363881e-01 4.13910300e-01 -2.01621816e-01 -8.59087229e-01
6.53713048e-01 9.50997889e-01 -9.29274112e-02 -8.52594733e-01
-3.67264450e-01 1.26942128e-01 -1.46240443e-01 -3.60118225e-02
6.38545692e-01 -1.33717731e-02 -9.56760347e-01 2.13062191e+00
4.82538760e-01 -4.15275186e-01 2.31267676e-01 2.13820115e-01
6.38592303e-01 7.01298833e-01 4.22380567e-02 -5.70849240e-01
1.72066331e+00 -6.31217480e-01 -5.00472724e-01 -2.66196907e-01
1.22181165e+00 -1.93501115e-01 7.61993408e-01 3.10601443e-01
-1.12609780e+00 1.06774853e-03 -8.64636898e-01 -2.36877158e-01
-3.02759349e-01 -3.34318817e-01 8.55661035e-01 9.36251879e-01
-1.23991466e+00 4.78459865e-01 -6.98328853e-01 -4.78851229e-01
-1.97243556e-01 5.07928491e-01 -5.72680175e-01 -2.22535521e-01
-9.18468475e-01 4.52845752e-01 8.37146938e-01 -2.81855613e-01
-1.46319032e-01 2.05921605e-02 -1.25375485e+00 1.74549833e-01
7.27544963e-01 -8.03131104e-01 1.27531815e+00 -1.10524213e+00
-8.97748649e-01 1.19978964e+00 -5.39255261e-01 -3.89518619e-01
-1.45964772e-01 5.97229600e-01 -4.57407594e-01 3.82054150e-01
1.25806957e-01 3.42531621e-01 5.09581566e-01 -1.14348745e+00
-1.02029228e+00 -6.38218760e-01 5.53916693e-01 -3.28744911e-02
2.17264503e-01 2.62049288e-01 -5.88285863e-01 -1.44486204e-01
6.19210243e-01 -9.77734029e-01 -4.03579414e-01 -6.39387190e-01
-5.38359880e-01 -4.89616096e-01 2.75462538e-01 -6.73611999e-01
1.61757016e+00 -1.99153972e+00 1.80856287e-01 6.47667170e-01
5.18084764e-01 -2.96991795e-01 -2.16745630e-01 7.80830145e-01
-2.88592428e-01 4.06408697e-01 -5.17712057e-01 3.47432941e-01
1.01311453e-01 7.34305799e-01 -3.70915323e-01 1.46690488e-01
-5.89799173e-02 7.70458639e-01 -9.28550065e-01 -5.98513484e-01
-2.51936346e-01 -4.05691594e-01 -7.00363219e-01 8.12797472e-02
-6.18775249e-01 -4.18683231e-01 -6.75690293e-01 3.16306293e-01
6.84125006e-01 -9.76913273e-02 1.03101802e+00 3.20979506e-01
2.06218399e-02 5.55973649e-01 -1.45508862e+00 1.19794083e+00
-2.86738485e-01 9.57559720e-02 2.23572552e-01 -7.55086243e-01
8.50923598e-01 2.37955526e-01 1.77221730e-01 -3.05088192e-01
-6.36352971e-02 1.80636361e-01 7.04757646e-02 -3.43036145e-01
4.54301655e-01 -2.68529564e-01 -4.61425066e-01 5.99474192e-01
-5.19890301e-02 2.24458486e-01 7.52412140e-01 7.81650126e-01
1.46463406e+00 1.92262128e-01 6.63649976e-01 -3.02660882e-01
5.36129355e-01 1.51893243e-01 6.64293766e-01 6.27967179e-01
4.29100156e-01 1.61208361e-01 1.30519915e+00 -7.44647622e-01
-9.73558545e-01 -6.75154865e-01 3.47336113e-01 1.01228893e+00
-6.16638213e-02 -1.00013220e+00 -1.16498315e+00 -9.29308593e-01
-2.64112562e-01 7.07125843e-01 -5.71824849e-01 2.09318101e-01
-7.69671321e-01 -2.99512744e-01 3.67004782e-01 4.30914491e-01
-6.70973659e-02 -1.20009685e+00 -7.90270865e-01 5.01272380e-01
-9.35767516e-02 -1.57424283e+00 -3.17908406e-01 3.67693037e-01
-1.23140132e+00 -1.56587660e+00 5.38455606e-01 -1.16180336e+00
1.19228292e+00 1.78869992e-01 1.32407403e+00 6.45573914e-01
-2.77039707e-02 4.44219679e-01 -4.07644898e-01 -2.23988160e-01
-8.24602962e-01 3.06636393e-01 -5.25718451e-01 -4.67250437e-01
2.30411291e-01 -5.37061453e-01 2.67780364e-01 -1.44550458e-01
-1.25592291e+00 3.03965420e-01 1.25987962e-01 4.86530453e-01
7.85724938e-01 4.12041157e-01 1.20484628e-01 -1.80218184e+00
3.79014760e-01 -2.57107526e-01 -9.88218963e-01 6.51512146e-01
-4.39090431e-01 5.59570432e-01 7.86932707e-01 3.58235598e-01
-7.81142056e-01 5.85977256e-01 -8.13648701e-02 3.66014481e-01
-2.19287962e-01 5.06500125e-01 -7.03179538e-01 1.75466195e-01
3.53667378e-01 -5.84923364e-02 -1.12405300e-01 -2.62156308e-01
5.00133812e-01 2.02605933e-01 5.03777206e-01 -9.25359070e-01
6.47590458e-01 1.82844132e-01 7.33388782e-01 -7.88564026e-01
-3.93718779e-01 -2.84088701e-01 -5.92493355e-01 3.04964423e-01
7.10427999e-01 -2.46852010e-01 -4.60807145e-01 -9.50056836e-02
-1.33602440e+00 -1.45051077e-01 -2.07895294e-01 -4.48468365e-02
-6.92186058e-01 7.30803132e-01 -4.27773714e-01 -8.93923104e-01
-2.73944348e-01 -8.42977583e-01 8.51408303e-01 -2.96764731e-01
-2.37409756e-01 -1.02557600e+00 1.27880245e-01 2.71418169e-02
-4.39102560e-01 4.09063160e-01 1.78741062e+00 -8.20660174e-01
-6.61411524e-01 6.17894679e-02 -8.40159953e-02 -2.61725724e-01
1.26760632e-01 -7.30068609e-02 -4.33299094e-01 -7.07394704e-02
-2.91376919e-01 2.13776127e-01 4.85892326e-01 1.45874307e-01
8.17617416e-01 -3.81637663e-01 -7.29006767e-01 2.85393059e-01
1.71108997e+00 3.69688183e-01 6.40728891e-01 1.94606379e-01
5.19742668e-01 8.66104841e-01 4.31973100e-01 -5.81694432e-02
3.89767349e-01 2.97595829e-01 2.34571263e-01 1.35119170e-01
9.24800336e-02 -5.88828981e-01 1.11435562e-01 6.51456535e-01
-7.05531463e-02 -5.00439227e-01 -9.07455266e-01 5.72850347e-01
-1.74831605e+00 -6.14947140e-01 -6.84631646e-01 2.32000256e+00
3.82935464e-01 9.85428467e-02 2.99171329e-01 3.81236017e-01
9.00713086e-01 -1.50341645e-01 3.33945677e-02 -1.27853572e+00
7.52945915e-02 5.48698366e-01 8.22676003e-01 1.04810321e+00
-7.43905246e-01 1.25499594e+00 6.90024233e+00 1.91989094e-01
-2.50429869e-01 -1.35290906e-01 1.12073377e-01 5.57074368e-01
-8.32181990e-01 7.49629080e-01 -7.49881566e-01 -1.78148866e-01
1.19398117e+00 -1.96870372e-01 7.61160314e-01 5.63302755e-01
-3.55024725e-01 -1.23933472e-01 -1.05044413e+00 4.45559621e-01
-3.48316789e-01 -9.76239860e-01 1.98857754e-01 2.66988754e-01
3.08340073e-01 -5.22359550e-01 -7.12090611e-01 1.33096769e-01
6.83990419e-01 -7.90810943e-01 3.79295647e-01 -1.20625548e-01
7.92956948e-01 -1.01501858e+00 3.36953223e-01 3.86012375e-01
-1.47393298e+00 -3.73871885e-02 -4.53340948e-01 -2.12885737e-01
-4.63544065e-03 2.24160135e-01 -6.47561848e-01 9.47947562e-01
1.57419801e-01 -2.96274908e-02 -2.67167330e-01 5.77890456e-01
-5.32126904e-01 4.95733976e-01 -3.91030014e-01 1.00464046e-01
1.35564938e-01 -4.84095931e-01 5.72846055e-01 1.34000969e+00
3.00111502e-01 7.56604373e-01 4.04377282e-01 4.90552545e-01
-3.92044149e-02 2.87849694e-01 -1.03036714e+00 -4.03389245e-01
4.38289791e-01 1.12060404e+00 -1.46126068e+00 -4.06690419e-01
-7.08126366e-01 8.44450474e-01 4.86454338e-01 2.03162462e-01
-3.78194273e-01 -8.56905997e-01 4.57185805e-01 6.77291006e-02
4.43495899e-01 -3.27860057e-01 -9.54836607e-02 -8.92348111e-01
1.77340105e-01 -9.40649033e-01 1.12461972e+00 -6.41302466e-01
-4.73736286e-01 8.09532046e-01 2.91508973e-01 -3.39974254e-01
-3.84951860e-01 -6.79030597e-01 -4.68879312e-01 7.67834067e-01
-9.95914400e-01 -8.89975190e-01 2.27891505e-01 3.59805763e-01
3.01925421e-01 2.98239410e-01 1.13478279e+00 -5.45229793e-01
-1.65091977e-01 2.56459504e-01 -7.35166609e-01 8.61072466e-02
-3.84869844e-01 -1.47925627e+00 9.77112889e-01 1.28018200e+00
4.35226083e-01 6.63594544e-01 8.05627227e-01 -8.18954051e-01
-1.86545277e+00 -8.44475269e-01 1.44197643e+00 -2.91457117e-01
4.57621515e-01 -5.16576946e-01 -8.14037204e-01 1.34693205e+00
2.19118983e-01 -2.00341567e-01 4.44547147e-01 1.10658616e-01
-3.17690790e-01 2.24755108e-01 -1.20876455e+00 4.76479501e-01
1.76941562e+00 -2.24753529e-01 -7.00670362e-01 3.12579572e-01
9.88486469e-01 -5.54277241e-01 -6.37404680e-01 5.24936989e-02
2.51570821e-01 -4.94367272e-01 4.83509839e-01 -1.04205501e+00
8.29005241e-02 -5.20709991e-01 -1.30677804e-01 -1.01739693e+00
-5.20838678e-01 -9.08404052e-01 1.43859312e-01 1.01319039e+00
7.00276554e-01 -7.33330905e-01 8.20209444e-01 6.36469245e-01
-7.92427734e-02 -4.65249836e-01 -6.76343322e-01 -6.92439735e-01
-1.27921999e-01 -6.15538061e-01 9.11736488e-01 6.15602195e-01
2.90596128e-01 8.53313804e-01 5.33728562e-02 6.47072077e-01
7.22481251e-01 7.88315773e-01 6.60212457e-01 -1.51583326e+00
-3.13128829e-01 -4.98755611e-02 -4.40628886e-01 -7.15844035e-01
4.75375086e-01 -1.31541395e+00 -1.54258609e-01 -2.06723285e+00
1.26648217e-01 -2.59761155e-01 2.62287110e-01 8.52140725e-01
2.08475813e-01 -4.58906293e-01 6.79165721e-02 -3.91599417e-01
-4.02206063e-01 -2.18696773e-01 1.06820381e+00 1.38251483e-01
-3.01420301e-01 -1.33106962e-01 -8.72134566e-01 6.12105072e-01
7.22294569e-01 -7.63841033e-01 -7.82848179e-01 -2.89295495e-01
7.24743485e-01 6.09033585e-01 -1.52667724e-02 -2.96493053e-01
-4.39341925e-02 -4.50407147e-01 -4.45382506e-01 -2.55615383e-01
-4.66932029e-01 -8.89782548e-01 5.63604712e-01 5.23854375e-01
-2.82889098e-01 4.53221589e-01 1.08829610e-01 4.47955966e-01
5.00581972e-02 -7.93654442e-01 4.57133770e-01 -3.94400984e-01
-7.78892279e-01 1.61241725e-01 -3.74649495e-01 1.49203718e-01
7.60033131e-01 -5.77894449e-01 -1.80707052e-01 -1.51179478e-01
-1.00825763e+00 1.35313973e-01 6.72706544e-01 1.49315864e-01
4.93697226e-01 -8.67950618e-01 -3.39124769e-01 3.18472803e-01
1.71663791e-01 -9.86717716e-02 -2.34396636e-01 1.87504932e-01
-7.34324813e-01 3.16526622e-01 6.30733669e-02 8.66273716e-02
-1.76790118e+00 1.01120472e+00 6.55924827e-02 -3.94866765e-01
-8.71898830e-01 3.07383716e-01 4.39061314e-01 -3.58065546e-01
-4.22700606e-02 -6.87056184e-01 -1.91284314e-01 -6.23839736e-01
3.11751485e-01 7.77472481e-02 1.52213469e-01 -4.84440714e-01
-3.39875996e-01 4.75024045e-01 -8.19175562e-04 -1.79514363e-01
1.15046966e+00 -9.18267444e-02 -6.47053123e-01 -3.03743817e-02
8.53749752e-01 3.74422669e-01 -4.04211164e-01 -1.80830210e-01
5.41909039e-01 -2.83446819e-01 -3.41277897e-01 -2.89291173e-01
-8.97231162e-01 2.94681132e-01 -2.25696146e-01 7.43197680e-01
1.24023497e+00 2.96477497e-01 8.54769170e-01 6.83199227e-01
7.15461612e-01 -8.68011475e-01 -8.05017471e-01 6.81494594e-01
4.61685389e-01 -3.93448800e-01 -1.83184683e-01 -1.30919290e+00
-3.25674236e-01 1.38524020e+00 2.02931002e-01 -9.68661383e-02
1.77476153e-01 5.76880038e-01 -3.61830741e-01 -2.80870825e-01
-1.07463157e+00 -5.64528525e-01 -3.08538198e-01 6.56757891e-01
1.09920077e-01 3.22983146e-01 -8.37860525e-01 6.07649803e-01
-3.68360877e-01 -1.71756104e-01 7.34635532e-01 1.33137774e+00
-8.20856392e-01 -1.80513060e+00 -2.24916920e-01 1.37168989e-01
-6.19894087e-01 -1.70154184e-01 -8.22504163e-01 9.80411828e-01
1.86010264e-02 9.96483326e-01 -1.33988872e-01 -2.70719558e-01
4.25872415e-01 3.28641176e-01 1.04894936e+00 -1.06679428e+00
-3.77653658e-01 -4.54838723e-02 8.98739994e-01 -3.60866606e-01
-6.82970285e-02 -5.22615910e-01 -2.04096174e+00 -2.66153157e-01
-2.20877692e-01 6.07407033e-01 3.60549331e-01 7.93001056e-01
1.38201222e-01 2.28354335e-01 5.50392747e-01 3.05208117e-01
-1.01445541e-01 -3.81252974e-01 -8.03194523e-01 3.37394416e-01
-1.72446251e-01 -2.01025248e-01 -2.11323112e-01 1.67171896e-01] | [10.287538528442383, 9.556445121765137] |
fa04ddc6-ad1c-4c73-bf3f-8d1e96b36790 | phrase-aware-unsupervised-constituency | null | null | https://openreview.net/forum?id=c9pFDJXSGa5 | https://openreview.net/pdf?id=c9pFDJXSGa5 | Phrase-aware Unsupervised Constituency Parsing | Recent studies have achieved inspiring success in unsupervised grammar induction using masked language modeling (MLM) as the proxy task. Despite their high accuracy in identifying low-level structures, prior arts tend to struggle in capturing high-level structures like clauses, since the MLM task usually only requires information from local context. In this work, we revisit LM-based constituency parsing from a phrase-centered perspective. Inspired by the natural reading process of human, we propose to regularize the parser with phrases extracted by an unsupervised phrase tagger to help the LM model quickly manage low-level structures. For a better understanding of high-level structures, we propose a phrase-guided masking strategy for LM to emphasize more on reconstructing non-phrase words. We show that the initial phrase regularization serves as an effective bootstrap, and phrase-guided masking improves the identification of high-level structures. Experiments on the public benchmark with two different backbone models demonstrate the effectiveness and generality of our method. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['constituency-parsing'] | ['natural-language-processing'] | [ 3.59460503e-01 4.95124429e-01 -5.61533272e-01 -4.54832554e-01
-1.04532349e+00 -6.28401101e-01 2.59902447e-01 2.10883364e-01
-2.29169145e-01 4.97718811e-01 5.19712389e-01 -8.36682022e-01
3.47786754e-01 -7.69633293e-01 -7.44679391e-01 -4.95177418e-01
1.56411409e-01 4.58922863e-01 1.59239545e-01 -2.04268564e-02
2.12439552e-01 2.90248580e-02 -1.08086383e+00 6.49908602e-01
9.66421783e-01 1.00824490e-01 6.19373202e-01 2.40902945e-01
-6.55262649e-01 8.69916022e-01 -4.67208624e-01 -3.90152544e-01
-1.41812093e-03 -5.83329082e-01 -1.04084337e+00 2.05645666e-01
2.75212109e-01 2.39770748e-02 3.62276971e-01 9.48872447e-01
-1.58071537e-02 -2.48931810e-01 3.30407679e-01 -6.80845678e-01
-2.42787212e-01 1.31106865e+00 -5.60411751e-01 6.18245006e-02
2.40372092e-01 -1.08628340e-01 1.85373390e+00 -8.80433142e-01
7.26966023e-01 1.40222800e+00 5.04228711e-01 6.48997426e-01
-1.61426604e+00 -4.85177249e-01 6.69479012e-01 -8.44235271e-02
-1.12247431e+00 -2.39508420e-01 9.18832660e-01 -2.82440037e-01
1.22444975e+00 6.11682124e-02 1.97620809e-01 8.62426579e-01
1.71214789e-01 1.02463806e+00 1.08637702e+00 -9.95489895e-01
-7.93097988e-02 5.82730360e-02 5.90456486e-01 8.54012668e-01
2.55647510e-01 -1.40263870e-01 -5.91627419e-01 -6.66120872e-02
4.92945343e-01 -5.06430268e-01 -1.07133072e-02 -6.09138273e-02
-1.01327026e+00 1.16998172e+00 -8.43222439e-02 5.96085966e-01
-2.24617153e-01 -1.27076298e-01 1.37468129e-01 4.40959558e-02
4.50136095e-01 4.93908852e-01 -6.52777612e-01 3.90173867e-02
-9.54637468e-01 -5.09196036e-02 7.49487281e-01 9.81531084e-01
8.59778762e-01 -1.88272342e-01 -2.07824379e-01 8.90810132e-01
3.50503832e-01 1.40477419e-01 2.54651964e-01 -7.95785010e-01
7.42427111e-01 8.02849412e-01 -3.40843707e-01 -5.99708021e-01
-5.01632154e-01 -5.51251054e-01 -5.30720770e-01 -1.87187731e-01
4.16412890e-01 -1.90555640e-02 -9.31942999e-01 2.13333225e+00
1.10661611e-01 -1.65113539e-01 9.48848128e-02 3.96542996e-01
5.58062792e-01 8.63771975e-01 4.10589546e-01 -3.85420948e-01
1.56220484e+00 -1.08197427e+00 -6.38682008e-01 -6.12719357e-01
1.08673513e+00 -5.01529872e-01 1.38632321e+00 2.23574206e-01
-1.24546862e+00 -4.34304059e-01 -8.60376596e-01 -1.59238443e-01
4.73691486e-02 1.24123134e-01 6.60482824e-01 5.65294027e-01
-8.26455474e-01 3.04186344e-01 -9.63441432e-01 -1.47882313e-01
1.14576459e-01 3.71395677e-01 -2.74526477e-01 6.36064336e-02
-9.94997919e-01 7.01874375e-01 5.77848434e-01 -1.06021017e-01
-4.88969117e-01 -4.58046049e-01 -9.54940617e-01 1.81695849e-01
5.98160446e-01 -4.10324603e-01 1.17199588e+00 -7.37536430e-01
-1.26749182e+00 1.16598356e+00 -7.38501489e-01 -4.62711692e-01
-1.83499455e-01 -1.81157976e-01 1.07248500e-01 6.40230179e-02
2.55719334e-01 6.75560057e-01 6.53339207e-01 -1.29056716e+00
-8.66252720e-01 -2.15881094e-01 1.42488003e-01 6.40271232e-02
-1.25425428e-01 3.38161469e-01 -4.33153629e-01 -6.68291867e-01
6.55834317e-01 -8.52750123e-01 -2.48421431e-01 -1.05435538e+00
-5.46769142e-01 -4.34900910e-01 2.35997602e-01 -7.21291959e-01
1.65410733e+00 -1.97510016e+00 3.14050972e-01 1.65521413e-01
2.23672122e-01 9.39760655e-02 -1.48892194e-01 4.67868537e-01
-1.16736680e-01 4.40316260e-01 -4.59457427e-01 -7.32476652e-01
-2.48379633e-02 6.16394937e-01 -4.41948056e-01 -1.92476287e-01
5.49415767e-01 1.11716998e+00 -7.12907732e-01 -6.63999915e-01
-1.10364154e-01 -4.41457182e-02 -8.05837870e-01 1.83667675e-01
-4.70908493e-01 6.11764669e-01 -3.71654540e-01 7.01378703e-01
4.81881857e-01 -3.36356938e-01 7.63065994e-01 2.16784328e-01
-2.01141864e-01 1.10195076e+00 -9.03845608e-01 1.59628260e+00
-6.11426890e-01 1.52643189e-01 2.28774413e-01 -9.36730683e-01
5.80372810e-01 2.40897790e-01 -2.87959576e-02 -4.33562309e-01
-2.05031112e-01 1.39559448e-01 4.35101390e-01 -1.95064217e-01
2.61119485e-01 -3.06715012e-01 -3.32256556e-01 4.58859712e-01
1.26859352e-01 8.74066502e-02 2.73761064e-01 2.71151274e-01
1.13127887e+00 1.73354909e-01 4.42775339e-01 -4.07882601e-01
5.77891469e-01 1.88786954e-01 9.89303350e-01 8.22804630e-01
2.81928122e-01 4.97269988e-01 5.95577300e-01 -2.63236582e-01
-6.64136946e-01 -9.37508643e-01 -7.93804675e-02 1.58189094e+00
-2.53130019e-01 -1.07579052e+00 -1.00587559e+00 -9.11773860e-01
-4.31474268e-01 1.06717110e+00 -3.28871191e-01 7.50399977e-02
-1.21954393e+00 -1.10429943e+00 3.58059049e-01 6.24676585e-01
1.27507195e-01 -1.37251544e+00 -2.24116579e-01 6.06317937e-01
-4.90770400e-01 -1.28633821e+00 -2.31594428e-01 6.95290744e-01
-9.88117576e-01 -7.36547351e-01 -9.28303413e-03 -1.27929544e+00
9.22588408e-01 -4.39837202e-02 1.53347039e+00 3.98733288e-01
3.01393926e-01 -1.23159923e-01 -3.96913558e-01 -3.55005831e-01
-8.53081763e-01 4.55807656e-01 -1.94617957e-01 -2.63613760e-01
6.03955805e-01 -6.74416184e-01 1.18309408e-01 1.10025369e-01
-7.25371361e-01 4.38866943e-01 6.91024601e-01 9.09947872e-01
6.91557288e-01 -5.32629080e-02 2.77988613e-01 -1.47863460e+00
4.25734818e-01 -9.17774662e-02 -6.54358447e-01 2.69391805e-01
-5.92781246e-01 4.41386878e-01 6.32400036e-01 -2.98641741e-01
-1.06578612e+00 1.40447333e-01 -3.85152102e-01 3.25028002e-01
-1.73018053e-01 6.78133309e-01 -6.71308756e-01 3.19095641e-01
1.76931694e-01 2.69537121e-01 -4.88307923e-01 -7.15177298e-01
3.55152756e-01 2.89342403e-01 3.30603153e-01 -1.11745965e+00
9.39709723e-01 1.40442133e-01 -9.15398151e-02 -6.50503457e-01
-1.23819089e+00 -3.56170267e-01 -8.72567832e-01 4.60785806e-01
9.59803045e-01 -8.66153002e-01 -3.55137527e-01 3.46307643e-02
-1.42759407e+00 -3.83688867e-01 -5.97149916e-02 1.51787296e-01
-4.10541922e-01 4.96533453e-01 -9.01517212e-01 -7.93941855e-01
-2.53107697e-01 -1.12671471e+00 1.17348349e+00 -1.70699298e-01
-5.27132034e-01 -1.07827830e+00 1.20972335e-01 5.40251374e-01
-3.01897787e-02 -2.84237057e-01 1.66858292e+00 -6.67545438e-01
-8.51627290e-01 2.25359574e-01 2.66870409e-02 2.90975422e-01
8.75854120e-02 -3.78731281e-01 -7.91017950e-01 -4.72960807e-02
1.31190538e-01 -1.06673427e-01 9.42754090e-01 3.10543448e-01
8.03152442e-01 -2.67627746e-01 -3.25337231e-01 6.49065554e-01
1.15688813e+00 1.66027285e-02 4.62424189e-01 4.41484541e-01
8.72609019e-01 8.02692533e-01 6.19736493e-01 -6.40874803e-02
5.01093149e-01 3.54530692e-01 2.49396399e-01 -1.12762056e-01
6.49262220e-02 -6.23676598e-01 5.28428793e-01 1.20304275e+00
1.57728940e-01 -1.44380391e-01 -1.06190181e+00 4.90956545e-01
-1.78141522e+00 -5.02096295e-01 -2.29062110e-01 1.92941058e+00
1.21729565e+00 5.89542031e-01 -1.34870093e-02 2.36146227e-02
6.88741982e-01 1.68098599e-01 1.00783259e-01 -4.30220455e-01
-1.96349859e-01 4.90020573e-01 2.03731999e-01 8.79819751e-01
-1.03821313e+00 1.64329779e+00 6.31460190e+00 7.76125610e-01
-8.04206491e-01 5.93222268e-02 5.04018962e-01 3.24576855e-01
-5.16987503e-01 5.26462853e-01 -1.53939569e+00 3.07305217e-01
9.39983547e-01 2.33926907e-01 1.20394751e-01 7.47815013e-01
1.74796253e-01 -6.31290898e-02 -1.19429481e+00 6.40100896e-01
-2.48188555e-01 -1.27039969e+00 2.26786032e-01 1.53390542e-01
4.86497730e-01 -5.72190993e-02 -1.98855564e-01 4.66292650e-01
2.50491053e-01 -9.98508632e-01 6.84825838e-01 -5.57925589e-02
4.08131152e-01 -6.23020768e-01 5.61280429e-01 8.50245416e-01
-1.23163259e+00 9.80148390e-02 -4.28424060e-01 -3.30002218e-01
2.88595438e-01 5.05293667e-01 -8.69495809e-01 3.66298318e-01
1.85516044e-01 3.27547252e-01 -5.43874621e-01 3.01028937e-01
-9.88166511e-01 1.36695445e+00 -2.71125525e-01 6.46233708e-02
3.91256690e-01 -4.66389179e-01 6.32285237e-01 1.51581764e+00
-2.55298018e-01 3.43453616e-01 5.14778256e-01 9.14736271e-01
-8.31309259e-02 4.23242509e-01 -3.42450917e-01 -2.26706415e-01
3.10412198e-01 1.31049180e+00 -9.58966792e-01 -2.43568301e-01
-6.05724812e-01 6.96200728e-01 7.01484621e-01 1.25760525e-01
-6.24429405e-01 1.54003233e-01 3.14634949e-01 2.53045529e-01
3.25163096e-01 -4.98522758e-01 -4.85744089e-01 -1.24506056e+00
2.92499125e-01 -1.07834244e+00 3.81277025e-01 -2.63035923e-01
-1.04063606e+00 5.56688190e-01 -2.36890768e-03 -7.55821466e-01
-4.66023654e-01 -6.55314505e-01 -6.86564744e-01 9.39400256e-01
-1.47280729e+00 -1.41359520e+00 4.05882031e-01 1.17597945e-01
7.28670061e-01 8.35693553e-02 1.07088292e+00 -8.79114717e-02
-6.69495702e-01 6.24438584e-01 -2.89810687e-01 4.28558648e-01
4.20739442e-01 -1.45537293e+00 6.67494714e-01 1.13980949e+00
7.08470345e-01 1.08640409e+00 5.31436145e-01 -7.48863161e-01
-1.23556721e+00 -9.75898206e-01 1.40923488e+00 -7.85925508e-01
6.23687744e-01 -8.63763213e-01 -1.19232106e+00 1.03341186e+00
1.18125714e-01 -5.95825732e-01 8.08109462e-01 6.84849143e-01
-3.25629890e-01 2.49835327e-01 -6.67525709e-01 6.12562180e-01
1.05057800e+00 -4.81782407e-01 -1.11587763e+00 3.11198145e-01
8.94876599e-01 -1.98483780e-01 -4.14805323e-01 5.12030423e-01
1.72489703e-01 -6.54906869e-01 6.17830753e-01 -7.58526862e-01
4.44934338e-01 -1.38200685e-01 -2.70315349e-01 -9.49633896e-01
-5.50077975e-01 -7.68949926e-01 -1.75179094e-02 1.57396626e+00
8.47883642e-01 -4.62987930e-01 1.03289628e+00 3.68318409e-01
-1.67141169e-01 -6.89193964e-01 -6.42808139e-01 -6.85870171e-01
2.93543011e-01 -6.40975058e-01 2.03890905e-01 6.97457492e-01
1.66600659e-01 9.65803385e-01 -2.12579638e-01 4.33804512e-01
5.81499755e-01 5.17957151e-01 6.02248669e-01 -1.22401774e+00
-7.85836279e-01 -2.17098698e-01 2.18500607e-02 -1.31595623e+00
6.92874551e-01 -1.12012982e+00 3.06066543e-01 -1.36568129e+00
5.00794888e-01 -3.75254393e-01 -3.00125182e-01 7.52548218e-01
-4.91700292e-01 -4.24989164e-02 1.84709877e-01 9.39679518e-02
-5.15229523e-01 1.50170609e-01 7.13580489e-01 1.45287029e-02
-3.23653489e-01 4.85267751e-02 -9.22951043e-01 1.18881166e+00
7.50240445e-01 -8.23167443e-01 -1.98654965e-01 -6.11546040e-01
2.73728192e-01 -4.91317995e-02 -1.33458734e-01 -5.40170133e-01
-1.51153440e-02 -1.15714908e-01 -8.17946419e-02 -8.20721686e-01
-1.93231534e-02 -4.71602648e-01 -2.91843772e-01 3.83532912e-01
-3.19167823e-01 3.57812494e-02 1.91378713e-01 1.41464248e-01
-3.46258521e-01 -6.48074627e-01 5.80940127e-01 -3.05450410e-01
-6.51495576e-01 -2.25662868e-02 -3.33705127e-01 2.90109724e-01
3.00815672e-01 -4.33343910e-02 8.04619342e-02 -1.41820192e-01
-8.73673737e-01 1.94481239e-01 4.68747616e-01 2.07128152e-01
3.20272923e-01 -8.24087977e-01 -6.45290971e-01 3.50643307e-01
7.95642138e-02 1.57433063e-01 -3.78051609e-01 6.94616079e-01
-2.31493130e-01 5.27762830e-01 3.74986321e-01 -6.44797146e-01
-1.39058328e+00 4.69020993e-01 -7.07612559e-02 -8.93609762e-01
-5.32374501e-01 1.00588787e+00 7.38816679e-01 -5.32407165e-01
2.69644082e-01 -6.53408170e-01 -3.15218419e-02 -1.86169177e-01
4.38329071e-01 -2.35953018e-01 1.13527268e-01 -5.19032121e-01
-4.46771324e-01 6.57714486e-01 -3.96859825e-01 -1.76613137e-01
1.40469158e+00 -1.65290713e-01 -5.51049471e-01 5.03021538e-01
8.41948926e-01 6.48095667e-01 -9.48760927e-01 -3.43307257e-01
9.65383947e-01 2.06736535e-01 -1.09804444e-01 -6.75374806e-01
-4.95615214e-01 9.36974704e-01 -3.14691722e-01 6.32774830e-03
1.10804367e+00 2.83986688e-01 8.39372098e-01 4.89331752e-01
5.85714698e-01 -8.77636254e-01 -2.16592073e-01 9.90387022e-01
4.11115855e-01 -1.23856688e+00 -2.63547242e-01 -9.33069229e-01
-4.22743052e-01 8.64816248e-01 4.28193629e-01 1.33250300e-02
3.57079476e-01 6.01370931e-01 1.06773496e-01 -3.71701010e-02
-9.22953367e-01 -4.13213491e-01 2.63094574e-01 2.18108267e-01
7.81582296e-01 2.66330130e-02 -6.58710659e-01 1.02754807e+00
-3.84713829e-01 -5.77924252e-01 2.48333901e-01 1.13177156e+00
-6.82354629e-01 -1.90940428e+00 -3.35736215e-01 5.93605824e-02
-9.20496643e-01 -7.90418923e-01 -4.78510380e-01 7.76435852e-01
1.21135771e-01 1.01664734e+00 -5.99650741e-02 -1.81307420e-01
9.91147757e-02 5.90523839e-01 6.07799888e-01 -1.34094608e+00
-5.32719016e-01 4.79359239e-01 2.59390682e-01 -5.38944185e-01
-3.09081852e-01 -6.25178814e-01 -1.61997783e+00 2.46921033e-01
-3.13055307e-01 4.75325435e-01 4.98120785e-01 1.16406906e+00
1.03073239e-01 2.71797210e-01 4.70712334e-01 -3.52812648e-01
-5.11337042e-01 -9.65395153e-01 -3.89789939e-01 2.67634213e-01
2.99605485e-02 -3.12806934e-01 -3.62249672e-01 2.06517175e-01] | [10.40538215637207, 9.596328735351562] |
3283281b-f968-4d97-868a-8788efcc9389 | instance-segmentation-gnns-for-one-shot | 2103.06509 | null | https://arxiv.org/abs/2103.06509v1 | https://arxiv.org/pdf/2103.06509v1.pdf | Instance Segmentation GNNs for One-Shot Conformal Tracking at the LHC | 3D instance segmentation remains a challenging problem in computer vision. Particle tracking at colliders like the LHC can be conceptualized as an instance segmentation task: beginning from a point cloud of hits in a particle detector, an algorithm must identify which hits belong to individual particle trajectories and extract track properties. Graph Neural Networks (GNNs) have shown promising performance on standard instance segmentation tasks. In this work we demonstrate the applicability of instance segmentation GNN architectures to particle tracking; moreover, we re-imagine the traditional Cartesian space approach to track-finding and instead work in a conformal geometry that allows the GNN to identify tracks and extract parameters in a single shot. | ['Gage DeZoort', 'Savannah Thais'] | 2021-03-11 | null | null | null | null | ['3d-instance-segmentation-1'] | ['computer-vision'] | [-5.18022329e-02 3.00685763e-01 -1.20716967e-01 -2.26871207e-01
-6.65772557e-01 -7.68617868e-01 8.16038191e-01 4.51108187e-01
-5.26053786e-01 4.68014687e-01 -5.59717953e-01 -5.11753440e-01
-3.89624983e-01 -1.06733823e+00 -8.57390285e-01 -6.22989714e-01
-9.12158713e-02 1.62604010e+00 5.28872728e-01 -5.99937402e-02
3.93874310e-02 1.11965549e+00 -9.14785862e-01 9.96577591e-02
1.84720278e-01 8.88613820e-01 -8.29153359e-02 8.24111223e-01
-3.28622878e-01 3.99852514e-01 -3.27202260e-01 -3.73587608e-01
3.89106184e-01 -5.61894059e-01 -8.14826488e-01 2.72959583e-02
1.48612887e-01 3.92042518e-01 -4.99629617e-01 1.20322919e+00
1.11059956e-01 2.75501341e-01 6.58833325e-01 -1.28664541e+00
-3.54008153e-02 8.72876883e-01 -1.05717190e-01 4.92245227e-01
-2.15958580e-01 -2.33352054e-02 7.51654565e-01 -5.24330020e-01
9.09888983e-01 9.76280808e-01 1.02548957e+00 2.68280536e-01
-1.21821392e+00 -1.40915215e-01 -4.99042384e-02 1.63151532e-01
-9.73304391e-01 5.75713217e-02 6.09189689e-01 -6.18723094e-01
1.00819731e+00 3.28260660e-01 1.00751734e+00 5.88538408e-01
3.48018885e-01 5.66826224e-01 7.28874683e-01 -4.59453106e-01
3.59427184e-01 -1.44322515e-01 6.67201877e-01 5.41190982e-01
2.97908932e-01 4.98775423e-01 -7.78440014e-02 1.83505982e-01
8.61404181e-01 -5.87466769e-02 1.02995910e-01 -8.00402880e-01
-1.45089710e+00 1.02417111e+00 7.24459231e-01 3.99465084e-01
-2.06229597e-01 3.99757892e-01 5.58780730e-01 2.13727251e-01
5.15784144e-01 7.07919300e-01 -2.52359539e-01 2.22156435e-01
-1.11800754e+00 6.27518177e-01 7.88697898e-01 9.20098484e-01
4.32231724e-01 -6.02196306e-02 -4.17066008e-01 1.31943189e-02
-2.08494738e-02 4.13877845e-01 -1.56403050e-01 -6.20618165e-01
2.16706321e-01 5.08767962e-01 1.62755102e-01 -5.57351172e-01
-1.09924519e+00 -7.53210366e-01 -5.70682585e-01 4.95928973e-01
6.63675487e-01 -3.33607420e-02 -1.11107254e+00 1.03612423e+00
8.64019513e-01 3.51656258e-01 -2.85615176e-01 1.04088151e+00
7.55589724e-01 4.30411935e-01 -7.57621648e-03 -1.66169420e-01
1.43123150e+00 -9.03443456e-01 -5.43301642e-01 1.24297217e-01
5.62419295e-01 -3.99355918e-01 4.29967605e-02 4.58157182e-01
-1.39609337e+00 -7.53517330e-01 -1.15825462e+00 7.03308508e-02
-6.00766063e-01 -5.59169240e-02 6.92432165e-01 5.96192241e-01
-8.59482527e-01 1.17951977e+00 -1.23439217e+00 -2.88260132e-01
4.12035555e-01 7.06092238e-01 2.76466683e-02 4.98009533e-01
-8.20142567e-01 9.28784609e-01 9.68245387e-01 1.18146308e-01
-8.06309998e-01 -8.12534094e-01 -6.31882787e-01 -2.41731793e-01
4.47887778e-01 -8.49310398e-01 1.49298763e+00 -3.63588631e-01
-1.17502105e+00 7.73708999e-01 9.53449458e-02 -1.11588478e+00
4.92754042e-01 1.18882991e-01 -4.93481845e-01 1.26122579e-01
-6.24088608e-02 3.25276077e-01 6.69840813e-01 -1.11267734e+00
-7.63876319e-01 -2.03796759e-01 1.58568993e-01 -4.52219509e-02
7.27237344e-01 3.14958632e-01 -5.67899287e-01 -1.26140624e-01
6.17837727e-01 -1.04525065e+00 -4.75167364e-01 -6.71195745e-01
-7.89218903e-01 -1.13959312e-01 8.19196880e-01 -4.13986921e-01
7.48261273e-01 -1.55081105e+00 4.05712992e-01 3.51310253e-01
6.03336573e-01 2.08430141e-01 5.50825536e-01 2.73705453e-01
-2.42336750e-01 -2.52856582e-01 -2.18757749e-01 -1.55438185e-01
2.93363214e-01 -1.02208138e-01 -3.68686207e-02 6.18690610e-01
-1.35770589e-01 1.30299175e+00 -9.53228354e-01 -2.93804467e-01
3.78597498e-01 2.10512340e-01 -2.61447966e-01 -4.73466575e-01
-8.73737454e-01 9.43001211e-01 -6.38876200e-01 3.44624043e-01
7.67394304e-01 -3.65643054e-01 8.39673132e-02 -2.81923413e-01
-4.86811310e-01 2.91167628e-02 -9.87113833e-01 1.57647264e+00
4.99664657e-02 6.47194326e-01 1.38174435e-02 -1.24779546e+00
6.27093554e-01 2.22116128e-01 9.82188582e-01 -8.74403238e-01
3.78339350e-01 -9.33915079e-02 1.24760687e-01 -1.75293982e-01
7.08174944e-01 -2.69352317e-01 -3.27393621e-01 6.13159895e-01
1.85058583e-02 -1.95709661e-01 2.65166730e-01 2.08222553e-01
1.27693582e+00 1.04140408e-01 -1.47301191e-02 -1.63740724e-01
3.11176836e-01 9.86753881e-01 2.34977424e-01 9.90631819e-01
6.70562387e-02 5.24844646e-01 3.18996459e-01 -8.07467103e-01
-1.47547400e+00 -1.32501388e+00 -3.33225220e-01 5.42134345e-01
2.73475200e-01 -5.30875564e-01 -9.12634313e-01 -9.13116992e-01
-1.10715732e-01 7.26086676e-01 -3.59753221e-01 2.47406870e-01
-8.50279808e-01 -8.35568249e-01 2.66182989e-01 4.87862796e-01
1.06061198e-01 -1.21454871e+00 -5.48491001e-01 4.12522525e-01
3.64222914e-01 -1.19767463e+00 2.38987301e-02 2.95878708e-01
-7.01640666e-01 -1.13068438e+00 -2.82943428e-01 -5.02833426e-01
7.08066404e-01 -5.99168353e-02 1.48785841e+00 1.02200940e-01
-5.81370831e-01 2.60659635e-01 -3.14096034e-01 -8.00061524e-01
-6.12002194e-01 8.63518659e-03 2.06401497e-02 -2.12626353e-01
3.07930470e-01 -1.26136452e-01 -5.21435380e-01 1.14911802e-01
-4.31443214e-01 -2.48213969e-02 9.58576351e-02 3.36863875e-01
1.13125038e+00 4.80673820e-01 -4.85601202e-02 -1.18344069e+00
2.18933970e-01 -4.69161332e-01 -1.18069291e+00 6.58788234e-02
-1.95409074e-01 -1.78408831e-01 3.22620302e-01 -1.80905834e-02
-6.16759896e-01 3.35644454e-01 -2.73366243e-01 -5.33150196e-01
-2.38785267e-01 1.51881993e-01 3.11424658e-02 -4.41276073e-01
4.08865333e-01 -4.57982346e-02 -2.27941230e-01 -3.82421851e-01
6.51709259e-01 -1.23549782e-01 6.18124485e-01 -3.66190076e-01
1.05757391e+00 6.72478855e-01 5.20178020e-01 -6.67947292e-01
-9.53441381e-01 -6.27672017e-01 -1.31582868e+00 -5.33387721e-01
1.49001551e+00 -4.19366151e-01 -1.15782416e+00 -1.60142630e-02
-1.17603898e+00 -2.78816849e-01 -6.69528127e-01 6.39183760e-01
-7.07171679e-01 1.49872184e-01 -3.62226248e-01 -7.30927527e-01
-1.26734793e-01 -1.10155189e+00 1.22915328e+00 3.11753780e-01
7.67099485e-03 -1.21159017e+00 2.88728088e-01 3.09493318e-02
8.46890211e-02 4.90132421e-01 1.00229645e+00 -1.00847483e+00
-9.33913410e-01 -4.53924298e-01 -2.30483040e-01 -3.35080951e-01
-4.51142341e-01 -3.18674594e-01 -5.08965969e-01 -3.44449997e-01
2.28494346e-01 2.55694240e-01 8.25009048e-01 8.06301475e-01
1.09600604e+00 3.32519233e-01 -1.10460341e+00 7.05505133e-01
1.53521776e+00 5.46071589e-01 1.41698137e-01 4.19076890e-01
9.34830427e-01 3.23870838e-01 3.48335028e-01 -7.32731670e-02
-7.25046769e-02 1.08191919e+00 3.39391917e-01 -1.49181351e-01
-2.47505158e-01 6.26707543e-03 -2.34375462e-01 5.60576379e-01
-1.73221007e-02 -3.09589118e-01 -1.00161469e+00 2.35635787e-01
-2.00031757e+00 -1.09207547e+00 -8.89472425e-01 1.97329569e+00
-1.74556505e-02 5.48604429e-01 4.36126590e-01 1.66581243e-01
6.96377814e-01 -2.38135189e-01 -4.84807760e-01 -4.09951210e-01
3.23925525e-01 5.59163630e-01 8.36515427e-01 5.51413119e-01
-1.30907834e+00 8.40728343e-01 6.64174652e+00 6.79666579e-01
-6.00873828e-01 4.95197803e-01 2.65616983e-01 -3.42834145e-01
-8.38803649e-02 2.08298355e-01 -1.10806715e+00 4.59243238e-01
9.56142843e-01 3.56030278e-02 2.68742263e-01 4.93199319e-01
2.67636217e-02 -1.04295593e-02 -1.32262838e+00 6.62889719e-01
-3.38387460e-01 -1.78264463e+00 -3.40791851e-01 1.20549634e-01
6.18519843e-01 3.23598981e-01 -3.54447961e-01 4.66767550e-01
5.92622936e-01 -9.64558721e-01 8.24790657e-01 8.87107968e-01
1.99895576e-01 -7.47927725e-01 3.58373493e-01 4.67870325e-01
-1.26146376e+00 3.39754343e-01 -4.61352170e-01 2.63207525e-01
9.71878290e-01 7.96948195e-01 -1.22078824e+00 1.13907373e+00
4.06058818e-01 2.81413168e-01 -3.45711380e-01 1.75382090e+00
2.84368515e-01 5.48174500e-01 -5.44205368e-01 8.84280726e-02
5.36127329e-01 -7.58231938e-01 8.85089993e-01 1.11320531e+00
3.46696436e-01 -5.74962720e-02 5.74893594e-01 9.23412919e-01
1.29431739e-01 -4.92680579e-01 -6.92878008e-01 4.27341834e-02
-4.75395983e-03 1.32911515e+00 -1.77018976e+00 -1.76755980e-01
-1.74136102e-01 7.09708095e-01 3.50192398e-01 1.56155825e-01
-1.02001464e+00 -4.27279696e-02 2.64083624e-01 2.05826029e-01
6.49799824e-01 -5.04509151e-01 -2.11220846e-01 -4.48877960e-01
-2.82555491e-01 -2.19353557e-01 3.86412531e-01 -6.16540134e-01
-1.02418137e+00 7.01492906e-01 1.77153543e-01 -1.10779655e+00
-1.59877256e-01 -9.58519459e-01 -8.04904640e-01 6.45444572e-01
-1.00339508e+00 -1.20170856e+00 -5.31514622e-02 2.92605251e-01
5.81851244e-01 -4.91859950e-02 3.49141687e-01 1.19374849e-01
-4.55051631e-01 -8.83101225e-02 2.93530554e-01 2.39984676e-01
-3.55155140e-01 -1.67934752e+00 1.17975307e+00 6.72224998e-01
9.65098321e-01 2.08623171e-01 9.42376137e-01 -1.19204748e+00
-1.55468976e+00 -1.18176055e+00 6.83327675e-01 -1.01044786e+00
8.64003062e-01 -6.69093609e-01 -6.68386936e-01 9.61967051e-01
7.33643025e-02 3.29088181e-01 -5.21194525e-02 1.33424297e-01
4.25161779e-01 3.47202331e-01 -1.01953578e+00 3.21238756e-01
1.33503234e+00 -4.36836302e-01 -2.48043090e-01 1.17957604e+00
6.17245555e-01 -9.41939414e-01 -7.59385467e-01 4.54805523e-01
-1.36358216e-01 -7.33952522e-01 1.36560881e+00 -8.82102489e-01
-3.61303240e-01 -4.76627201e-01 4.13076907e-01 -1.19272101e+00
-4.81442273e-01 -6.44199371e-01 -1.01563878e-01 7.59557545e-01
3.04472536e-01 -1.79319277e-01 1.28756070e+00 6.13672808e-02
-3.68320018e-01 -4.12272424e-01 -1.07500315e+00 -9.71325040e-01
1.12540454e-01 -9.10198331e-01 6.61376119e-01 5.09892702e-01
-3.77560675e-01 1.91148698e-01 1.60382986e-01 7.25172043e-01
1.00142181e+00 2.55917341e-01 3.84948164e-01 -1.74508882e+00
-4.58647817e-01 -5.60833633e-01 -5.09875000e-01 -7.38342524e-01
6.99697807e-02 -1.68042922e+00 -2.67963056e-02 -1.74031901e+00
-1.29959017e-01 -7.95702994e-01 -2.20934495e-01 -1.64447919e-01
4.53843445e-01 1.80458114e-01 2.35251456e-01 1.33261725e-01
-1.00371373e+00 1.27122262e-02 1.09233522e+00 -3.59991819e-01
1.08120359e-01 5.14649808e-01 1.28434896e-01 6.65362179e-01
6.39316380e-01 -7.39033580e-01 -2.08526641e-01 -2.91534573e-01
4.94033188e-01 5.79970598e-01 7.39494085e-01 -1.42257810e+00
6.29516721e-01 3.33644986e-01 5.58343709e-01 -1.22596860e+00
4.63482291e-01 -7.72961080e-01 6.36532545e-01 2.83464611e-01
-2.70744525e-02 3.36700797e-01 3.95753175e-01 6.82787538e-01
5.71134035e-03 -7.75612831e-01 4.11868542e-01 -5.69079936e-01
-5.66119254e-01 5.61777413e-01 -1.31707340e-01 -2.07193241e-01
1.34576464e+00 -3.67060244e-01 -3.03074960e-02 3.52982551e-01
-1.35339749e+00 2.13029623e-01 3.26620013e-01 1.90222278e-01
2.01561958e-01 -1.42225266e+00 -3.32976997e-01 -3.11858952e-02
-1.40960455e-01 1.68054402e-02 2.60897785e-01 9.42592263e-01
-8.22510242e-01 7.89940476e-01 2.99230800e-03 -8.94270241e-01
-1.02610242e+00 9.03453588e-01 6.35663629e-01 -4.54584926e-01
-1.26608849e+00 1.00169206e+00 5.96040301e-02 -3.41322631e-01
-5.92888296e-02 -2.86806196e-01 -2.11002171e-01 7.84574002e-02
3.00598949e-01 2.09196612e-01 6.40847504e-01 -4.89375085e-01
-6.25117049e-02 5.34472048e-01 -1.84330434e-01 -1.88340649e-01
1.02153122e+00 3.26917112e-01 6.40357211e-02 3.98783088e-01
6.96968913e-01 -2.72313714e-01 -1.04627848e+00 -2.24027522e-02
4.24689353e-01 1.37992740e-01 1.52915433e-01 -7.53562212e-01
-1.09470654e+00 5.43098211e-01 6.10609591e-01 9.00623202e-01
3.73715878e-01 3.50040495e-01 9.42967057e-01 3.02895010e-01
6.21200800e-01 -9.14526165e-01 -5.08556604e-01 6.86783314e-01
3.19957733e-01 -9.25308824e-01 -9.46410522e-02 -3.30299050e-01
-2.47474149e-01 1.27250648e+00 1.93024516e-01 -2.69302249e-01
7.31077075e-01 4.23456937e-01 -2.79830366e-01 -1.06226420e+00
-2.53312588e-01 -5.54588437e-01 5.80051959e-01 5.25391698e-01
-3.80207561e-02 1.46478340e-01 -1.63138062e-02 2.55405992e-01
-4.41854537e-01 -2.80576169e-01 2.81369418e-01 7.71379173e-01
-6.43651843e-01 -1.26722884e+00 -4.80266273e-01 5.35533428e-01
-2.04411045e-01 2.21883222e-01 -2.69892544e-01 1.12110519e+00
4.09307361e-01 5.68334520e-01 4.41598207e-01 -1.06093794e-01
4.69440937e-01 2.42425799e-01 7.78081119e-01 -7.26249158e-01
-1.05768514e+00 -1.82167545e-01 1.10007301e-01 -5.07595301e-01
-3.07262540e-01 -8.34451020e-01 -1.84023345e+00 -3.07884142e-02
-2.20065892e-01 5.77739418e-01 1.11224759e+00 1.32622015e+00
-1.07564434e-01 9.72362757e-01 1.01619825e-01 -1.14428675e+00
-7.19356313e-02 -4.66346264e-01 -5.99590898e-01 3.40154380e-01
1.51234418e-01 -8.13725531e-01 1.59973368e-01 -2.41058797e-01] | [15.695473670959473, 2.9151132106781006] |
a5dda608-9bbe-4aee-9279-bdfc4561d947 | gnpm-geometric-aware-neural-parametric-models | 2209.10621 | null | https://arxiv.org/abs/2209.10621v1 | https://arxiv.org/pdf/2209.10621v1.pdf | GNPM: Geometric-Aware Neural Parametric Models | We propose Geometric Neural Parametric Models (GNPM), a learned parametric model that takes into account the local structure of data to learn disentangled shape and pose latent spaces of 4D dynamics, using a geometric-aware architecture on point clouds. Temporally consistent 3D deformations are estimated without the need for dense correspondences at training time, by exploiting cycle consistency. Besides its ability to learn dense correspondences, GNPMs also enable latent-space manipulations such as interpolation and shape/pose transfer. We evaluate GNPMs on various datasets of clothed humans, and show that it achieves comparable performance to state-of-the-art methods that require dense correspondences during training. | ['Lourdes Agapito', 'Mirgahney Mohamed'] | 2022-09-21 | null | null | null | null | ['pose-transfer'] | ['computer-vision'] | [-2.14169651e-01 1.80501446e-01 -2.68039465e-01 -1.91547215e-01
-4.11909580e-01 -6.80638194e-01 7.99448192e-01 -2.30478808e-01
-2.71793343e-02 5.03450572e-01 3.82413745e-01 3.33172828e-01
3.95650268e-02 -6.59136117e-01 -1.19850469e+00 -5.26012957e-01
-3.06602716e-01 9.81729329e-01 -2.91523952e-02 -2.33414266e-02
-1.14976130e-01 6.61840558e-01 -8.06414783e-01 -1.13385901e-01
6.24886334e-01 5.42863607e-01 -1.07184328e-01 6.10727906e-01
3.45559180e-01 1.75654098e-01 1.43219307e-02 -4.62639444e-02
5.13824046e-01 3.65223698e-02 -5.99492908e-01 3.24654937e-01
6.75847173e-01 -6.09301150e-01 -6.02479637e-01 5.88432074e-01
1.91856667e-01 2.85075724e-01 8.08802605e-01 -1.08798671e+00
-1.16678977e+00 1.84276074e-01 -3.95218819e-01 -4.32238132e-01
3.00199151e-01 4.89603788e-01 9.01960313e-01 -1.15859997e+00
9.88237381e-01 1.50171220e+00 9.67626512e-01 7.10881412e-01
-2.05051494e+00 -1.96535856e-01 2.47700110e-01 -3.23932022e-01
-1.32019222e+00 -2.66004920e-01 1.20637190e+00 -7.10146189e-01
9.00432885e-01 1.69684794e-02 1.04982436e+00 1.52756286e+00
2.70674706e-01 6.52028203e-01 7.16525972e-01 -7.54209757e-02
1.19026937e-01 -4.77491826e-01 -1.65493250e-01 9.01191890e-01
1.87280253e-01 4.14068937e-01 -6.47779524e-01 -4.00117278e-01
1.71225047e+00 4.04751822e-02 -2.75304109e-01 -1.21794021e+00
-1.64608681e+00 6.37603104e-01 6.30498707e-01 -2.26295009e-01
-5.15714765e-01 7.90102839e-01 1.46018758e-01 -4.07330645e-03
6.18916214e-01 4.96294260e-01 -6.54928446e-01 8.84489790e-02
-7.79597521e-01 6.79419398e-01 7.41656840e-01 1.28787649e+00
7.08952606e-01 6.93114847e-02 -8.70466530e-02 1.79822475e-01
5.08700013e-01 5.92999518e-01 -1.90624103e-01 -1.44920349e+00
4.27931130e-01 5.67627609e-01 3.70803356e-01 -1.16046953e+00
-2.24700242e-01 -7.65577406e-02 -1.08230686e+00 1.46822572e-01
4.97126728e-01 1.78581357e-01 -1.06224036e+00 1.92335427e+00
3.56003582e-01 5.10754526e-01 -4.22064841e-01 1.01900184e+00
4.05393362e-01 4.69613731e-01 7.86819905e-02 2.31494978e-01
9.43431616e-01 -7.67637551e-01 -3.52664083e-01 -9.78506505e-02
4.11582440e-02 -4.16185230e-01 1.06019914e+00 -1.03299180e-03
-1.32778299e+00 -6.87916517e-01 -8.78363192e-01 -4.52563643e-01
2.51661334e-02 -3.43096070e-02 6.95206225e-01 -2.38241978e-06
-1.16107082e+00 1.15540099e+00 -1.55899811e+00 -3.31860393e-01
3.20035875e-01 4.09812719e-01 -6.01067424e-01 1.44434750e-01
-7.35793412e-01 7.64512300e-01 5.33211827e-02 3.78719538e-01
-1.21390963e+00 -1.18087494e+00 -1.13288617e+00 -2.18156755e-01
1.41599625e-01 -1.43738163e+00 9.94487524e-01 -3.41437548e-01
-1.76185477e+00 1.02065432e+00 -1.61885947e-01 -3.10996443e-01
8.75970721e-01 -4.26683128e-01 3.60624701e-01 5.84059209e-02
-3.72048289e-01 9.32096899e-01 9.50686634e-01 -1.37696493e+00
5.10580480e-01 -4.19387370e-01 1.00167632e-01 9.56192091e-02
2.68231392e-01 -5.08225024e-01 -4.07491684e-01 -8.37485969e-01
4.18634415e-01 -1.35642791e+00 -4.85162854e-01 7.38611341e-01
-5.19221604e-01 -6.82509243e-02 9.69112396e-01 -8.76564026e-01
3.68158817e-01 -1.76357841e+00 1.07664728e+00 2.35304311e-01
4.20945615e-01 -2.05005199e-01 -1.86826259e-01 3.19735885e-01
2.45182574e-01 8.22204128e-02 -2.53273100e-01 -9.81053233e-01
4.65771526e-01 6.27981603e-01 -4.02784586e-01 8.20727110e-01
4.66369092e-01 1.53643847e+00 -8.12540114e-01 -2.65941709e-01
4.20481920e-01 7.92254150e-01 -9.01197910e-01 3.38889480e-01
-7.78806746e-01 1.28000736e+00 -4.25677449e-01 4.73766804e-01
5.21669626e-01 -5.11369407e-01 2.35105813e-01 -5.03759086e-01
6.12488016e-02 2.17235848e-01 -9.20769393e-01 2.37840199e+00
-1.20090708e-01 2.20407084e-01 1.01610176e-01 -7.96143949e-01
8.13080072e-01 4.87963349e-01 6.10780716e-01 7.73672089e-02
5.91642149e-02 -2.31907349e-02 -6.31278694e-01 -1.51052371e-01
2.85248697e-01 7.39087164e-02 -1.79802731e-01 2.42053732e-01
2.19703242e-01 -4.81530637e-01 -5.18845856e-01 2.78775189e-02
7.38642812e-01 1.17065454e+00 -1.16852566e-03 -4.50742722e-01
1.40220582e-01 1.90903302e-02 5.42933047e-01 2.59958446e-01
1.27073541e-01 6.82510853e-01 2.25006789e-01 -8.49949777e-01
-1.64794517e+00 -1.60911453e+00 1.54274330e-01 6.81196451e-01
9.94849131e-02 -2.45544121e-01 -5.26553571e-01 -3.72805208e-01
4.78192359e-01 3.29898238e-01 -8.54654610e-01 4.70389128e-02
-1.17730618e+00 -1.86429515e-01 3.40172470e-01 7.84726739e-01
2.65774101e-01 -8.56654048e-01 -4.78088647e-01 1.55214012e-01
-5.49564473e-02 -1.20844376e+00 -8.18833411e-01 -3.32412153e-01
-1.38148773e+00 -9.18649852e-01 -7.05174029e-01 -6.18787587e-01
1.09772420e+00 -1.32585570e-01 1.39107621e+00 -7.13268295e-02
-2.86904007e-01 6.41958654e-01 1.32384270e-01 1.24739699e-01
-3.52483362e-01 1.30092902e-02 3.25558871e-01 -3.76969963e-01
-9.25168172e-02 -1.33497548e+00 -6.60751343e-01 3.72637898e-01
-3.19660097e-01 4.07030195e-01 2.25106135e-01 8.13055098e-01
9.90761101e-01 -8.91822398e-01 -1.90537423e-03 -5.44662118e-01
1.77297041e-01 -3.02686766e-02 -7.92038620e-01 8.09151381e-02
-2.06512585e-01 4.17412609e-01 2.22550645e-01 -7.12040424e-01
-6.77273870e-01 3.71776551e-01 1.55615136e-01 -1.08377731e+00
-1.57764599e-01 1.06664635e-01 -1.07250601e-01 -2.66748965e-01
6.57364368e-01 1.48475096e-02 2.09189415e-01 -8.55054379e-01
8.24851453e-01 -4.68296528e-01 8.68211627e-01 -1.16381145e+00
1.18297541e+00 6.46961451e-01 3.97742659e-01 -5.57696760e-01
-4.71171051e-01 -7.65222386e-02 -1.46989822e+00 1.36819974e-01
1.05012548e+00 -9.80595887e-01 -8.48554611e-01 4.94369328e-01
-1.41554761e+00 -7.26344824e-01 -5.64146578e-01 3.58050078e-01
-1.06710780e+00 3.67968142e-01 -9.83194053e-01 -3.72860223e-01
-2.80938715e-01 -9.16078389e-01 1.36324036e+00 -4.56112057e-01
-5.80440879e-01 -1.31181574e+00 1.31585672e-01 -1.82803228e-01
1.91822127e-01 1.16064525e+00 8.69694591e-01 -1.38775967e-02
-1.28497005e+00 -3.60938758e-02 1.61877215e-01 9.65820178e-02
1.94907770e-01 -1.00385351e-03 -6.62656546e-01 -5.76377392e-01
-1.87159523e-01 -1.11748740e-01 5.24222255e-01 4.89006430e-01
1.08739901e+00 -6.18736029e-01 -4.26020801e-01 1.07827711e+00
1.11168015e+00 -6.69666290e-01 3.26842964e-01 -2.90722311e-01
1.20958006e+00 4.02332097e-01 6.75913841e-02 1.88700020e-01
5.54814875e-01 7.88723886e-01 4.77594107e-01 3.23449746e-02
-1.70340106e-01 -8.04312348e-01 1.92797735e-01 9.67468917e-01
-6.56436026e-01 4.02424514e-01 -7.68855989e-01 4.97348368e-01
-1.93392467e+00 -6.63455307e-01 5.18702529e-02 1.98113513e+00
8.45623314e-01 -3.27396095e-02 9.64761376e-02 -3.36646020e-01
3.60349119e-01 3.60574454e-01 -8.67572069e-01 1.25730067e-01
5.39750122e-02 2.81182081e-01 3.66075158e-01 7.28470981e-01
-9.55480278e-01 9.50642765e-01 6.62854624e+00 -1.17860530e-02
-7.82987714e-01 1.76362813e-01 7.26660565e-02 -1.70521401e-02
-6.13333344e-01 3.72212939e-02 -3.66409332e-01 1.61238909e-02
4.01509047e-01 2.35972535e-02 5.85865974e-01 6.69964314e-01
1.77853867e-01 5.69686115e-01 -1.66345561e+00 7.30907559e-01
-2.17539236e-01 -1.59706950e+00 2.38362938e-01 2.54951596e-01
9.39942539e-01 -2.97668874e-02 1.11001402e-01 -6.10666201e-02
5.77541649e-01 -9.37530160e-01 9.60200548e-01 8.70246589e-01
9.17946041e-01 -2.85155356e-01 -6.38112277e-02 4.80263144e-01
-1.21302533e+00 4.50542867e-01 -3.67410898e-01 -1.49061130e-02
3.42567533e-01 2.87552327e-01 -5.88636875e-01 5.43167353e-01
4.15396303e-01 9.66230631e-01 -1.14939585e-01 6.29481733e-01
-5.00959933e-01 3.52015108e-01 -4.67725843e-01 4.39852864e-01
-7.59579763e-02 -3.61449242e-01 9.56378400e-01 6.34118259e-01
2.44327173e-01 1.21512517e-01 4.72679138e-01 1.58771789e+00
-9.63275656e-02 -5.07510006e-01 -5.88601768e-01 7.68836495e-03
3.36607814e-01 7.69574940e-01 -2.84969032e-01 -2.79538631e-02
5.00294566e-02 1.27792716e+00 3.25947911e-01 5.60098410e-01
-6.97936893e-01 4.35320258e-01 9.13385689e-01 3.66643071e-01
2.56045580e-01 -1.34917724e+00 -3.53418887e-01 -1.42357993e+00
1.10851198e-01 -3.42286497e-01 -6.43264055e-02 -6.83867395e-01
-1.50633621e+00 2.36538976e-01 3.50451440e-01 -9.75080609e-01
-4.99788970e-01 -5.88675857e-01 -4.48151261e-01 1.03247845e+00
-1.03057957e+00 -1.89515316e+00 -4.37725186e-01 5.88638604e-01
2.73581654e-01 4.62679446e-01 1.07270360e+00 -2.35482708e-01
-3.64380050e-03 4.10115242e-01 -3.00336570e-01 2.06800133e-01
5.63243985e-01 -1.39767468e+00 1.28844810e+00 5.11996210e-01
1.83812350e-01 9.29322124e-01 7.06707180e-01 -8.73641491e-01
-1.80352950e+00 -1.14726579e+00 5.02404332e-01 -1.09473002e+00
3.31590056e-01 -7.31609821e-01 -9.69941974e-01 1.22677290e+00
-1.65563092e-01 3.95930588e-01 2.46733740e-01 1.84183508e-01
-6.84628844e-01 2.81427562e-01 -1.07899261e+00 5.97021580e-01
1.63889086e+00 -7.12863386e-01 -6.87236071e-01 2.59960204e-01
1.29083383e+00 -1.02592158e+00 -1.14680970e+00 4.45001274e-01
6.75214589e-01 -4.33572829e-01 1.43622482e+00 -7.23506987e-01
3.14440876e-01 -2.11705893e-01 -1.72858998e-01 -1.28470457e+00
-5.46042442e-01 -1.02686894e+00 -7.30882764e-01 7.29622483e-01
7.00157806e-02 -4.04927492e-01 8.78714800e-01 9.94101048e-01
-1.28098294e-01 -6.82950139e-01 -8.66959631e-01 -7.22610891e-01
2.70758808e-01 -1.30968362e-01 7.97107577e-01 1.10674644e+00
-4.07127500e-01 -1.50584295e-01 -5.81539631e-01 4.15626407e-01
1.06288433e+00 1.92472473e-01 1.09566987e+00 -1.37605774e+00
-4.55443084e-01 -1.52900591e-02 -3.62890720e-01 -1.37580597e+00
3.76735121e-01 -8.08848917e-01 -3.30706500e-02 -1.38238728e+00
-1.02636375e-01 -3.71725589e-01 1.46401316e-01 5.24845958e-01
9.08470750e-02 3.41631979e-01 2.65527785e-01 4.46870983e-01
-2.92973638e-01 9.72160757e-01 1.62813127e+00 1.33107323e-02
-3.17092061e-01 -2.84632266e-01 -3.37970629e-02 8.49957705e-01
3.14245522e-01 -1.52651682e-01 -2.97277212e-01 -8.30185354e-01
4.87237237e-02 1.82269469e-01 1.14481580e+00 -7.57628143e-01
1.53414890e-01 -4.71140265e-01 7.37427056e-01 -5.97887278e-01
7.67576635e-01 -7.43492663e-01 7.36366034e-01 3.88594270e-01
-3.00976902e-01 2.54063934e-01 2.14891732e-01 8.54533195e-01
3.54490012e-01 6.06420398e-01 6.01142943e-01 -3.20927858e-01
-4.07285780e-01 9.77013052e-01 3.32599461e-01 -1.31456822e-01
7.59470761e-01 -1.10350385e-01 1.10016130e-02 -8.37869197e-02
-1.19307232e+00 8.79196897e-02 1.02316308e+00 4.27040517e-01
5.99742055e-01 -1.64990258e+00 -5.90402484e-01 3.72534096e-01
-3.61537486e-02 6.81075215e-01 2.24166006e-01 6.75249159e-01
-5.01969159e-01 1.75651148e-01 -3.50483179e-01 -1.02271950e+00
-9.17065680e-01 5.05232990e-01 3.94410491e-01 -2.16119304e-01
-1.15056169e+00 6.66060805e-01 1.94099665e-01 -8.93926919e-01
3.51504944e-02 -7.85101473e-01 5.57901442e-01 -6.53317809e-01
-1.05867907e-01 2.99677074e-01 -3.38467836e-01 -5.62497020e-01
-1.29740968e-01 9.41412926e-01 4.64495748e-01 -1.95692495e-01
1.52056134e+00 4.89859618e-02 -2.79958308e-01 5.84699869e-01
1.06780100e+00 -2.19913065e-01 -1.99856186e+00 -3.99521083e-01
-2.49530137e-01 -5.70776165e-01 -3.40906143e-01 -3.97145689e-01
-6.90965414e-01 9.80904818e-01 6.26793653e-02 -3.23733747e-01
5.54908812e-01 2.87707709e-02 8.73753726e-01 3.46998066e-01
7.52250731e-01 -5.35368204e-01 3.16341937e-01 5.25446236e-01
1.37489748e+00 -8.44439387e-01 9.17196497e-02 -5.38313806e-01
-2.89577574e-01 9.79463816e-01 4.73311990e-01 -9.74207580e-01
8.74533296e-01 4.12980616e-02 -4.42064583e-01 -2.31580332e-01
-6.34741426e-01 3.45824152e-01 8.09456170e-01 8.09821248e-01
-8.84923264e-02 2.28078857e-01 2.71227419e-01 3.01259428e-01
-1.58087432e-01 7.45697320e-02 -5.21859713e-02 7.82794058e-01
8.93740952e-02 -1.27601933e+00 -1.74851254e-01 -1.41694978e-01
3.82195152e-02 3.01317960e-01 -1.96189672e-01 8.49471569e-01
-9.82180163e-02 3.76721770e-02 1.95143610e-01 -3.16961825e-01
4.41111922e-01 1.61278680e-01 1.12609124e+00 -5.03315508e-01
-2.62384087e-01 7.53325447e-02 -2.34318018e-01 -1.02905607e+00
-5.33860147e-01 -6.67389154e-01 -1.05663836e+00 -5.77216089e-01
1.22273833e-01 -1.47690654e-01 5.15175283e-01 7.55829573e-01
7.00218439e-01 2.93800533e-01 1.20235302e-01 -1.69599748e+00
-8.13212991e-01 -8.02531540e-01 -2.54479915e-01 7.62449563e-01
5.70233047e-01 -8.04053366e-01 -8.36558267e-02 4.11086082e-01] | [6.9492926597595215, -1.2362289428710938] |
206b317d-342f-44f8-87b9-af1dd266ec54 | semg-gesture-recognition-with-a-simple-model | 2006.03645 | null | https://arxiv.org/abs/2006.03645v2 | https://arxiv.org/pdf/2006.03645v2.pdf | sEMG Gesture Recognition with a Simple Model of Attention | Myoelectric control is one of the leading areas of research in the field of robotic prosthetics. We present our research in surface electromyography (sEMG) signal classification, where our simple and novel attention-based approach now leads the industry, universally beating more complex, state-of-the-art models. Our novel attention-based model achieves benchmark leading results on multiple industry-standard datasets including 53 finger, wrist, and grasping motions, improving over both sophisticated signal processing and CNN-based approaches. Our strong results with a straightforward model also indicate that sEMG represents a promising avenue for future machine learning research, with applications not only in prosthetics, but also in other important areas, such as diagnosis and prognostication of neurodegenerative diseases, computationally mediated surgeries, and advanced robotic control. We reinforce this suggestion with extensive ablative studies, demonstrating that a neural network can easily extract higher order spatiotemporal features from noisy sEMG data collected by affordable, consumer-grade sensors. | ['Carson Drake', 'David Josephs', 'Andrew Heroy', 'John Santerre'] | 2020-06-05 | null | null | null | null | ['emg-gesture-recognition', 'electromyography-emg'] | ['medical', 'medical'] | [ 4.50039268e-01 7.18892738e-02 -4.31843042e-01 2.29654670e-01
-6.83639824e-01 -1.47869229e-01 7.35870451e-02 -7.40201354e-01
-4.18947846e-01 6.68542504e-01 4.73312765e-01 4.31676209e-02
-3.38496685e-01 -5.06974049e-02 -7.12765336e-01 -6.87487245e-01
-3.35704774e-01 2.23010376e-01 -2.05437958e-01 -3.54588360e-01
1.59992024e-01 3.51505697e-01 -1.29003108e+00 1.85144678e-01
4.89595175e-01 1.39919198e+00 4.25119907e-01 3.76411647e-01
5.46689272e-01 2.15132207e-01 -6.11052871e-01 1.07942618e-01
4.30931151e-02 -2.63769418e-01 -5.92725396e-01 -5.28653026e-01
8.40335265e-02 -1.07129328e-01 -5.54686725e-01 9.22871172e-01
1.11886489e+00 -3.62913609e-02 4.67046082e-01 -9.33527052e-01
-8.25273335e-01 5.71720719e-01 -1.58515617e-01 1.41552150e-01
2.26747692e-01 5.87746322e-01 8.09804499e-01 -5.01849711e-01
8.44661653e-01 8.67648542e-01 8.43445897e-01 9.51074898e-01
-1.11733770e+00 -5.90103924e-01 1.85450818e-02 4.79237884e-01
-6.08739436e-01 -2.39025831e-01 1.03586018e+00 -3.47230703e-01
1.02955604e+00 2.21579418e-01 8.89872670e-01 2.08988786e+00
7.72166729e-01 1.00848627e+00 8.35209429e-01 -1.01808488e-01
2.35142902e-01 -5.77406943e-01 -1.93305001e-01 -2.31867912e-03
1.33283049e-01 3.22753429e-01 -7.31142223e-01 -1.43860169e-02
1.04996133e+00 -5.70681766e-02 -7.73584306e-01 -8.13082904e-02
-1.83926678e+00 3.02642792e-01 4.69486624e-01 6.52316630e-01
-9.84995186e-01 6.77733362e-01 5.14295697e-01 3.96335512e-01
1.99585870e-01 1.07843459e+00 -7.17444122e-01 -8.78016591e-01
-5.18536150e-01 2.47309744e-01 6.31472647e-01 7.05397964e-01
-4.15011883e-01 3.04354876e-01 -3.34843397e-01 7.46116042e-01
-1.49019966e-02 2.60250241e-01 7.95221746e-01 -1.49738073e+00
3.53659719e-01 2.62808144e-01 -7.03864396e-02 -7.63192415e-01
-7.28976369e-01 -6.64411664e-01 -1.04260659e+00 4.56293225e-01
2.97672987e-01 -3.20728183e-01 -5.84732354e-01 1.78177643e+00
-3.34592164e-01 -2.46667732e-02 -3.65730762e-01 1.39000404e+00
4.28025186e-01 -2.06450596e-02 5.15137687e-02 -1.13969639e-01
1.14074266e+00 -5.26851714e-01 -7.63367772e-01 -2.58197904e-01
7.07745180e-02 -2.11930260e-01 1.08600032e+00 9.31728423e-01
-1.04254973e+00 -4.18265045e-01 -9.17271674e-01 1.08336218e-01
1.63363032e-02 2.98086345e-01 9.86478209e-01 8.67456123e-02
-7.55315661e-01 1.42098749e+00 -1.21477675e+00 -4.19641197e-01
6.52418613e-01 7.80819952e-01 -4.68417346e-01 2.84388512e-01
-9.60901856e-01 1.15149784e+00 -4.26725894e-02 4.03086126e-01
-6.82551861e-01 -7.10073054e-01 -2.17508569e-01 3.38730216e-02
1.48576871e-01 -7.48255670e-01 9.35328186e-01 -4.92225438e-01
-1.79110861e+00 6.36294007e-01 2.01306984e-01 -2.42365524e-01
1.58223897e-01 -6.24928534e-01 -2.98824668e-01 -1.17003754e-01
-1.89074218e-01 4.41208035e-01 8.45245063e-01 -5.41716337e-01
4.73003909e-02 -7.24596202e-01 -3.50229681e-01 -1.92209303e-01
-5.23098469e-01 5.01966290e-02 7.51590803e-02 -1.11949170e+00
3.04360926e-01 -9.49593365e-01 -3.01794797e-01 3.31170052e-01
-2.95493245e-01 -3.35414797e-01 4.94763672e-01 -8.33686531e-01
7.65250683e-01 -2.15529442e+00 1.04895508e+00 -2.31202934e-02
1.96874574e-01 1.28582448e-01 -3.72300833e-01 2.49676839e-01
-1.88670546e-01 -2.19756767e-01 -1.98413804e-01 2.03167543e-01
5.90493865e-02 -1.98968172e-01 -3.71202976e-02 3.59877914e-01
5.54297805e-01 1.24465680e+00 -9.69194055e-01 2.78355241e-01
1.93152383e-01 3.66898060e-01 -4.74051505e-01 4.57150675e-02
-2.43587494e-01 9.10424471e-01 -4.03711468e-01 1.05354810e+00
-1.44878238e-01 -2.28327751e-01 3.80184442e-01 -7.50869274e-01
2.92761117e-01 3.23934823e-01 -7.17665315e-01 2.41757107e+00
-3.57022882e-01 1.03661287e+00 2.89761007e-01 -1.41706288e+00
6.72115505e-01 5.74293077e-01 1.06388056e+00 -5.39631724e-01
7.05018044e-01 4.41066653e-01 4.23834264e-01 -9.09714580e-01
-8.04973766e-02 1.02275573e-01 -5.26119880e-02 2.33183801e-01
4.34185445e-01 1.56547591e-01 -2.78505087e-01 -5.04418850e-01
1.41565704e+00 6.80574834e-01 -1.40542522e-01 -3.25157285e-01
-9.93940383e-02 -6.81787506e-02 4.60779101e-01 5.16770840e-01
-4.49498415e-01 5.87522447e-01 3.12172800e-01 -2.82444172e-02
-7.18022346e-01 -9.13461864e-01 -1.08336024e-01 8.87750149e-01
-9.45962220e-02 1.12945784e-03 -5.41016400e-01 -2.52988841e-02
4.21138167e-01 -1.75298795e-01 -4.83694106e-01 -5.26520789e-01
-8.42441022e-01 -5.73490858e-01 4.82097715e-01 1.11956680e+00
1.36740550e-01 -1.64212656e+00 -6.29510880e-01 6.26182854e-01
-3.53634536e-01 -1.04227626e+00 -1.87037334e-01 3.58330369e-01
-1.25725758e+00 -1.09865582e+00 -1.17157912e+00 -1.07798433e+00
1.71463013e-01 -5.93507946e-01 5.47458351e-01 -4.44792837e-01
-6.48805320e-01 4.11630750e-01 -2.71067619e-01 -4.63967472e-01
-2.48085745e-02 1.99428484e-01 5.03142595e-01 -4.15592164e-01
4.28297400e-01 -1.16424048e+00 -6.28496647e-01 -1.13142026e-03
-2.23289058e-01 -4.61745322e-01 9.65189576e-01 1.11974216e+00
5.15004039e-01 -5.91619849e-01 1.05185187e+00 7.22479168e-03
1.16756403e+00 -4.60215420e-01 9.18563753e-02 2.27830745e-02
-3.39993566e-01 9.50907320e-02 4.66826618e-01 -9.41835999e-01
-3.85114402e-01 -9.94114801e-02 -2.12110952e-01 -5.97935796e-01
-5.79803251e-02 4.70962167e-01 -7.33756796e-02 -2.37831637e-01
8.19363832e-01 3.03931981e-01 5.30883729e-01 -7.30902314e-01
7.13022798e-02 9.35436189e-01 1.11790001e+00 -5.77752173e-01
2.13843271e-01 1.52390793e-01 1.27263173e-01 -7.79490232e-01
8.12472925e-02 -1.14019200e-01 -2.70104825e-01 -3.55461836e-01
6.10680044e-01 -5.38811922e-01 -1.26577854e+00 7.87678480e-01
-1.20941603e+00 -4.92357910e-01 -1.78082913e-01 9.06734169e-01
-1.00307071e+00 1.85724124e-01 -8.82425129e-01 -7.41625249e-01
-7.05139339e-01 -8.87129188e-01 1.20754302e+00 -8.45672935e-02
-8.02785218e-01 -4.94156659e-01 -1.40391096e-01 1.61830753e-01
6.26045406e-01 4.79486197e-01 8.68798912e-01 -1.89708158e-01
-3.77091199e-01 -3.91428024e-01 2.13257268e-01 6.08278930e-01
1.83558449e-01 -6.43200457e-01 -7.94311821e-01 -3.24882418e-01
7.67809749e-02 -5.21820962e-01 7.02975929e-01 7.24157214e-01
1.51790214e+00 1.86936751e-01 -4.33357209e-01 5.46508610e-01
9.95707393e-01 2.72427261e-01 7.60274768e-01 1.59767583e-01
4.63271052e-01 3.75743240e-01 2.95100570e-01 -3.47096706e-03
-2.93639660e-01 8.22937489e-01 3.55911344e-01 1.83657825e-01
-2.37711340e-01 1.46093488e-01 3.75553161e-01 8.94588709e-01
-8.68969738e-01 1.52491197e-01 -4.99484390e-01 6.17855906e-01
-1.75506556e+00 -8.67204249e-01 2.43267879e-01 1.71611691e+00
7.48999596e-01 -1.02222480e-01 -3.63888703e-02 4.98537481e-01
3.57133120e-01 -1.55283123e-01 -1.10737133e+00 -8.43234546e-03
-4.69296426e-02 8.83199513e-01 2.89094448e-01 -4.01915252e-01
-9.26787555e-01 4.49571490e-01 6.65291739e+00 3.65113914e-01
-1.50847375e+00 -7.61748757e-03 -1.08289897e-01 -4.53584641e-01
1.33048430e-01 -8.26142669e-01 -7.91991130e-02 8.70538056e-01
5.84096551e-01 5.15237339e-02 8.89867961e-01 8.61682296e-01
2.52292871e-01 2.13599414e-01 -1.38713634e+00 1.21859097e+00
-2.21369222e-01 -1.49808025e+00 -6.32195592e-01 8.05817097e-02
3.76687109e-01 4.85587955e-01 -1.56018093e-01 1.47909313e-01
-4.07347977e-01 -1.01501107e+00 5.50430596e-01 8.23567510e-01
9.18196380e-01 4.42482680e-02 6.87216759e-01 4.42950614e-02
-8.42144489e-01 -4.31933999e-01 9.77162719e-02 -1.43206388e-01
2.62918413e-01 1.87080741e-01 9.27980989e-02 3.30834478e-01
1.02910006e+00 1.12201393e+00 2.45080680e-01 9.75653410e-01
3.10597830e-02 5.48340440e-01 -2.76635081e-01 -5.10314643e-01
-2.96621114e-01 2.13578910e-01 9.87645030e-01 7.35722780e-01
3.22195262e-01 3.21008265e-02 -4.30796266e-01 1.20239651e+00
-1.09518528e-01 -5.13358176e-01 -4.79682058e-01 -6.31445110e-01
2.08187565e-01 8.11074018e-01 -2.18861625e-01 1.48616403e-01
-5.58710583e-02 1.15540612e+00 -7.55931213e-02 5.09687901e-01
-5.24056792e-01 -8.01526904e-01 1.01651883e+00 -1.53736293e-01
5.28962091e-02 -6.20292127e-01 -7.15502858e-01 -1.19108689e+00
8.51128697e-01 -8.49438131e-01 -2.46087208e-01 -8.17512095e-01
-1.37802029e+00 2.94918418e-01 -2.95602947e-01 -1.44986987e+00
-4.85018134e-01 -1.35527098e+00 -5.73790193e-01 6.98137581e-01
-1.09672594e+00 -5.12549996e-01 -3.28779966e-01 3.68729800e-01
4.62016761e-01 -3.04202467e-01 1.15032041e+00 5.65908611e-01
-3.81977081e-01 3.76043797e-01 2.06188902e-01 2.57959664e-01
5.56385994e-01 -8.33985806e-01 4.63573545e-01 3.58140349e-01
-1.18033133e-01 8.02438796e-01 3.93271148e-01 -4.32750225e-01
-1.92021096e+00 -5.59324920e-01 4.47278708e-01 -3.36511195e-01
5.98456204e-01 -1.82304636e-01 -4.77146924e-01 4.09932733e-01
-1.24845371e-01 -2.24837512e-01 2.00388595e-01 1.07799709e-01
2.95868039e-01 -1.97067454e-01 -8.67442548e-01 6.02773666e-01
1.59247541e+00 -4.18546468e-01 -6.78978264e-01 4.20818716e-01
2.80852973e-01 -3.39210421e-01 -1.31781006e+00 7.64026642e-01
1.33524513e+00 -2.16059357e-01 1.02469671e+00 -9.43298519e-01
5.81965446e-01 2.60932535e-01 -4.61293980e-02 -1.56092799e+00
-4.85733271e-01 -7.63136387e-01 -5.45518696e-01 3.71892214e-01
1.51879296e-01 -6.24186754e-01 8.53766382e-01 4.58696097e-01
-6.61856055e-01 -1.13278246e+00 -1.11466348e+00 -1.15060461e+00
7.90251791e-02 -5.27419209e-01 2.03407139e-01 5.01237512e-01
7.87216485e-01 -6.13635406e-02 -4.40690398e-01 -4.42277402e-01
4.14973021e-01 -1.66126877e-01 2.63636142e-01 -1.42026067e+00
-3.73689502e-01 -7.06112444e-01 -8.90083313e-01 -9.86551225e-01
1.24233715e-01 -8.31457078e-01 3.11698526e-01 -1.64292133e+00
-2.56444454e-01 -8.96811187e-02 -5.22890389e-01 5.70235372e-01
1.58799812e-01 2.70642698e-01 1.31670162e-01 5.58586717e-02
3.82298119e-02 6.31004512e-01 1.45071113e+00 -5.15070856e-01
-1.61602408e-01 3.21768299e-02 -6.65323973e-01 3.60798299e-01
8.61190081e-01 -1.90064490e-01 1.13732710e-01 -4.07490313e-01
-2.62143731e-01 4.38910015e-02 6.82971358e-01 -1.40017259e+00
1.08581513e-01 5.93279414e-02 3.98787349e-01 4.75892238e-02
5.45935631e-01 -7.64569879e-01 -4.90915403e-02 6.97521269e-01
-2.94133216e-01 -4.09704238e-01 5.90046570e-02 5.08174181e-01
-9.39737856e-02 5.43048799e-01 3.49522293e-01 -1.21497847e-01
-7.58029401e-01 1.06647395e-01 -4.89224285e-01 -3.05944066e-02
6.51954830e-01 -4.05901849e-01 -2.44747654e-01 -2.02457547e-01
-1.19464326e+00 -1.20008597e-02 -2.10726440e-01 1.02964771e+00
5.96243262e-01 -1.51224566e+00 -6.03077233e-01 3.98792565e-01
-3.16497274e-02 -4.71788436e-01 2.92026456e-02 1.21975899e+00
4.97886986e-02 5.26836216e-01 -7.36608863e-01 -8.03617358e-01
-6.56638086e-01 -2.53029168e-02 2.99012899e-01 3.03090662e-01
-9.42351282e-01 7.60814607e-01 -6.20431721e-01 -1.18234456e-01
7.33425200e-01 -8.35674286e-01 1.00375295e-01 -4.70993161e-01
8.00600722e-02 5.52435338e-01 2.54972816e-01 1.00943558e-02
-5.46238780e-01 6.15263939e-01 5.87005556e-01 1.38880238e-01
1.52679992e+00 5.47933638e-01 3.81049067e-02 4.66395795e-01
9.45020318e-01 -7.80903101e-01 -1.17062140e+00 2.90107459e-01
1.42268509e-01 9.72541124e-02 4.24101241e-02 -1.14748168e+00
-1.15835202e+00 8.43218148e-01 8.79790843e-01 -3.39308590e-01
1.19774866e+00 6.96104094e-02 1.12506902e+00 6.67316973e-01
9.98198509e-01 -1.28059304e+00 5.58608323e-02 1.77408874e-01
1.47770452e+00 -1.06337750e+00 -2.63250113e-01 3.47969383e-02
-2.79889822e-01 1.11194038e+00 5.14085650e-01 -5.41720331e-01
8.43618870e-01 3.26671004e-01 -1.17262065e-01 -1.83571935e-01
-4.82148856e-01 9.90587920e-02 4.87687290e-01 8.74105096e-01
5.29754043e-01 1.30406767e-01 -7.00973988e-01 1.32651412e+00
-2.57547349e-02 8.74093592e-01 -1.01975769e-01 1.02515888e+00
-8.38272199e-02 -9.72463012e-01 6.36810362e-02 9.92376208e-01
-4.76895154e-01 1.27643570e-02 -2.40908891e-01 6.97606623e-01
-6.56141490e-02 5.82456350e-01 1.82933826e-02 -8.37360919e-01
6.13606691e-01 2.32603341e-01 9.21028912e-01 -3.56428176e-01
-6.30205810e-01 -1.13124907e-01 1.13425702e-01 -1.22850049e+00
-4.12559271e-01 -5.51966548e-01 -1.04404259e+00 2.48311177e-01
-3.35544795e-01 -3.43889505e-01 1.05759323e+00 8.51810694e-01
7.73466051e-01 9.15885389e-01 2.33091414e-01 -1.62856531e+00
-9.07779932e-01 -1.46262312e+00 -5.76241612e-01 2.48445272e-01
2.62911469e-01 -9.84892368e-01 -2.27961928e-01 -1.33642465e-01] | [6.8502302169799805, 0.18503722548484802] |
788f3f92-62a2-4052-afff-d0fcbd7b5d71 | cross-modality-deep-feature-learning-for | 2201.02356 | null | https://arxiv.org/abs/2201.02356v1 | https://arxiv.org/pdf/2201.02356v1.pdf | Cross-Modality Deep Feature Learning for Brain Tumor Segmentation | Recent advances in machine learning and prevalence of digital medical images have opened up an opportunity to address the challenging brain tumor segmentation (BTS) task by using deep convolutional neural networks. However, different from the RGB image data that are very widespread, the medical image data used in brain tumor segmentation are relatively scarce in terms of the data scale but contain the richer information in terms of the modality property. To this end, this paper proposes a novel cross-modality deep feature learning framework to segment brain tumors from the multi-modality MRI data. The core idea is to mine rich patterns across the multi-modality data to make up for the insufficient data scale. The proposed cross-modality deep feature learning framework consists of two learning processes: the cross-modality feature transition (CMFT) process and the cross-modality feature fusion (CMFF) process, which aims at learning rich feature representations by transiting knowledge across different modality data and fusing knowledge from different modality data, respectively. Comprehensive experiments are conducted on the BraTS benchmarks, which show that the proposed cross-modality deep feature learning framework can effectively improve the brain tumor segmentation performance when compared with the baseline methods and state-of-the-art methods. | ['Yizhou Yu', 'Junwei Han', 'Jungong Han', 'Qiang Zhang', 'Guohai Huang', 'Dingwen Zhang'] | 2022-01-07 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [ 3.84362668e-01 -1.91575229e-01 -2.83323210e-02 -3.90476376e-01
-9.90941584e-01 -2.39048786e-02 5.62700629e-01 1.25127539e-01
-5.83798885e-01 5.42578697e-01 2.02203900e-01 -2.82919705e-01
-2.44351581e-01 -7.08033919e-01 -5.14326692e-01 -1.09952819e+00
2.77205825e-01 2.24251106e-01 3.28603178e-01 -1.34377733e-01
-8.31995159e-02 3.29007685e-01 -1.30681300e+00 4.02776152e-01
9.25056040e-01 1.35663676e+00 3.36879015e-01 2.25526467e-01
-6.05129778e-01 7.78621912e-01 -1.67750299e-01 8.62335414e-02
9.44418013e-02 -2.95724660e-01 -1.03790915e+00 3.73120338e-01
-4.34851944e-02 -7.67046139e-02 -3.25628906e-01 1.26473224e+00
5.69922566e-01 -1.18233345e-01 5.42380333e-01 -1.26675653e+00
-3.98321271e-01 3.92626494e-01 -9.33169305e-01 4.02994335e-01
-1.90301407e-02 2.65725311e-02 5.47603488e-01 -7.73565829e-01
4.14948076e-01 9.14340794e-01 5.16756952e-01 4.33312029e-01
-7.24372566e-01 -5.72667778e-01 1.54912919e-01 4.04118806e-01
-1.27951503e+00 -1.76245853e-01 8.07017326e-01 -5.62218428e-01
4.67107981e-01 1.35260940e-01 8.36290240e-01 1.03354776e+00
3.65885407e-01 1.06542778e+00 1.49762487e+00 -2.07755834e-01
-1.54231638e-01 -1.62136674e-01 3.37489009e-01 1.00523889e+00
-3.78991477e-02 1.13765053e-01 -3.45930547e-01 -4.41586114e-02
7.23022163e-01 6.45722926e-01 -4.16465610e-01 -3.12700450e-01
-1.53802550e+00 8.10866296e-01 7.62129009e-01 7.62876928e-01
-5.56473553e-01 -2.22558647e-01 6.71037018e-01 1.62016392e-01
3.34463716e-01 -1.72603101e-01 -5.67280650e-01 1.27874091e-01
-7.95216024e-01 -1.04604535e-01 4.26995069e-01 6.52445316e-01
7.79673040e-01 -2.73274750e-01 -2.97342896e-01 9.43102598e-01
5.08887231e-01 3.68489236e-01 1.16739368e+00 -1.75412983e-01
3.69167030e-01 8.69123101e-01 -4.22289073e-01 -6.05456710e-01
-8.01015437e-01 -4.64870811e-01 -1.23734248e+00 -8.95416643e-03
3.06485832e-01 -1.85430914e-01 -1.51690602e+00 1.47600281e+00
5.35554886e-01 2.90961593e-01 2.02628255e-01 8.68980110e-01
1.27260745e+00 4.32617784e-01 1.95495248e-01 -2.76648372e-01
1.56167126e+00 -1.11017513e+00 -7.25690603e-01 -1.99873131e-02
6.80853248e-01 -6.70961261e-01 6.91177130e-01 1.38574496e-01
-6.63183630e-01 -4.33154345e-01 -8.67844522e-01 -3.82936979e-03
-6.20493948e-01 -1.26039028e-01 1.04323411e+00 2.85997987e-01
-7.83561945e-01 -1.93469413e-02 -9.75919247e-01 -2.28061616e-01
9.24067974e-01 3.42921346e-01 -6.99316204e-01 -3.65570843e-01
-1.09615755e+00 8.51138949e-01 5.93188763e-01 2.50676930e-01
-8.76287401e-01 -8.74379992e-01 -7.47873187e-01 -1.90767393e-01
5.17761767e-01 -7.45532811e-01 1.07368171e+00 -1.21662188e+00
-1.18476713e+00 7.91664302e-01 9.25737247e-03 -9.06844661e-02
5.52350879e-01 1.22787736e-01 -5.60602427e-01 1.73227683e-01
1.29634142e-01 6.29064620e-01 8.28555286e-01 -1.07966447e+00
-9.75813925e-01 -5.74258626e-01 -3.01008105e-01 2.71195590e-01
-2.20876142e-01 -1.88659325e-01 -4.15612012e-01 -4.66513008e-01
3.42138439e-01 -7.37899005e-01 -2.96541452e-01 -2.49383926e-01
-5.01423120e-01 -2.16881141e-01 1.14870501e+00 -6.19954109e-01
7.62460530e-01 -1.96368217e+00 3.96314204e-01 2.69684851e-01
4.96414185e-01 7.11574554e-02 1.58044137e-02 -4.47062217e-02
-2.41528183e-01 -1.14914224e-01 -4.40079033e-01 -1.25628978e-01
-1.88339680e-01 3.36555302e-01 2.45431229e-01 3.73295546e-01
7.37612247e-02 9.57338095e-01 -6.56102180e-01 -8.50721240e-01
3.02122205e-01 4.38768238e-01 -3.78875174e-02 1.35022819e-01
1.41497329e-01 8.63635719e-01 -7.54967749e-01 1.18476355e+00
6.99575365e-01 -3.16428959e-01 -4.72423762e-01 -4.80118364e-01
-4.19906266e-02 -5.30347168e-01 -7.81018615e-01 2.17790484e+00
-3.67489398e-01 9.35352594e-02 -3.14963087e-02 -1.22043872e+00
5.76297522e-01 3.67780924e-01 1.11507285e+00 -8.00655603e-01
5.35205066e-01 3.91141415e-01 1.31421015e-01 -8.19957614e-01
6.20209984e-02 -5.57672143e-01 -4.33396436e-02 1.39312461e-01
2.95431525e-01 -1.01028280e-02 8.81407708e-02 -1.15795635e-01
1.12049389e+00 -2.94566512e-01 2.05991030e-01 -1.34360895e-01
8.20309460e-01 -2.88610458e-02 7.39979684e-01 2.71163702e-01
-4.48106885e-01 3.09070438e-01 2.52246559e-01 -5.58348775e-01
-5.44527411e-01 -1.00830889e+00 -4.22493637e-01 6.79298580e-01
2.18350947e-01 1.00758515e-01 -5.53515971e-01 -1.00074935e+00
-3.45038660e-02 -3.73137146e-02 -8.76456201e-01 -3.38259131e-01
-2.57256180e-01 -1.04456210e+00 2.31120974e-01 6.33749127e-01
9.45766330e-01 -1.03312802e+00 -6.54720485e-01 3.75649333e-02
-2.55887926e-01 -1.18324244e+00 -3.84304583e-01 3.40745717e-01
-8.43981028e-01 -1.22120500e+00 -1.09432471e+00 -9.78357077e-01
7.66032934e-01 2.03278884e-01 7.23078728e-01 1.13561533e-01
-6.38049424e-01 3.64207447e-01 -5.69302499e-01 -4.28965777e-01
-1.41785368e-01 2.63748318e-01 -3.75279576e-01 2.36063436e-01
3.78912032e-01 -4.34314430e-01 -6.16077960e-01 9.85808223e-02
-1.31644011e+00 5.60933411e-01 1.27439725e+00 1.14503562e+00
9.70504105e-01 1.33413061e-01 5.13089240e-01 -8.48639727e-01
3.08967531e-01 -7.25877881e-01 -1.78347558e-01 3.89648825e-01
-2.70375609e-01 9.11793951e-03 3.38311166e-01 -2.44698390e-01
-1.00073946e+00 1.77337691e-01 -2.86673814e-01 -3.55339885e-01
-3.38667005e-01 9.49182153e-01 -2.51285136e-01 -3.23835135e-01
1.52064517e-01 4.92420912e-01 1.23749658e-01 -2.46499121e-01
1.60378069e-01 6.31255627e-01 4.66205746e-01 -4.26436067e-01
5.20202756e-01 4.12077755e-01 1.60356179e-01 -5.04141867e-01
-7.75898814e-01 -5.16363204e-01 -7.68444598e-01 -2.37338111e-01
1.22524154e+00 -7.74841666e-01 -4.77047056e-01 9.63120997e-01
-7.30948627e-01 -9.49985255e-03 -2.01876462e-01 5.59996068e-01
-5.52962959e-01 2.76370853e-01 -4.99999672e-01 -1.42966941e-01
-3.34305406e-01 -1.70838439e+00 1.16719341e+00 7.30006576e-01
6.97352648e-01 -1.07313681e+00 -2.71135825e-03 4.85997677e-01
4.29336518e-01 5.30542195e-01 9.52212870e-01 -7.68734753e-01
-3.69496107e-01 -2.19257861e-01 -5.04685342e-01 2.34709650e-01
5.55780649e-01 -3.41585964e-01 -8.79130840e-01 -3.46411467e-01
-3.21971066e-02 -4.54907149e-01 1.06155634e+00 3.85956585e-01
1.39785898e+00 2.71331280e-01 -6.07881904e-01 7.35463560e-01
1.45562470e+00 1.97024867e-01 2.97990143e-01 3.77312869e-01
1.06257379e+00 2.71241456e-01 3.42920601e-01 3.88004214e-01
5.51062524e-01 3.57100338e-01 6.00925803e-01 -5.09724557e-01
-2.40916740e-02 2.13633016e-01 -1.55804858e-01 1.00267589e+00
1.44051448e-01 3.14062268e-01 -1.24061191e+00 5.71847558e-01
-1.85663891e+00 -5.93048394e-01 2.78106689e-01 1.75307047e+00
7.85062432e-01 -1.20176859e-02 9.66703985e-03 2.23885149e-01
7.28174984e-01 1.01282567e-01 -7.08467066e-01 3.19166891e-02
-5.72074950e-02 1.60197895e-02 5.36761045e-01 -5.55532016e-02
-1.39441252e+00 5.03787756e-01 5.27184772e+00 9.67518210e-01
-1.44275486e+00 4.01830107e-01 6.74489021e-01 1.79003477e-01
-9.61166546e-02 -3.65156800e-01 -3.66384923e-01 4.75943178e-01
6.63694620e-01 2.54085157e-02 7.86696002e-03 5.01911640e-01
-1.31326452e-01 -2.36432999e-01 -8.24593425e-01 1.28187847e+00
6.38386980e-02 -1.13352191e+00 1.12612367e-01 2.09453225e-01
6.62839293e-01 3.11520666e-01 8.87930244e-02 4.03780222e-01
3.51164937e-02 -9.62038100e-01 4.51460809e-01 9.35491025e-01
4.70388860e-01 -8.29750240e-01 9.30872619e-01 1.84896141e-01
-1.40869522e+00 -3.06002975e-01 -1.49667948e-01 5.84023178e-01
6.95273355e-02 7.87383616e-01 -4.22488242e-01 1.15910935e+00
8.03754091e-01 1.03805232e+00 -7.90985107e-01 1.24026239e+00
2.32633784e-01 3.63006592e-01 -1.85002565e-01 3.25267375e-01
4.30158824e-01 -3.32344025e-02 2.01849997e-01 9.77826178e-01
3.87606084e-01 7.47068971e-03 6.10989213e-01 5.89896202e-01
-4.24804837e-02 1.64176956e-01 -3.44873577e-01 -9.24704149e-02
1.48652690e-02 1.62037349e+00 -8.32067072e-01 -2.99876571e-01
-7.56643474e-01 7.75265276e-01 1.23129494e-01 1.94452211e-01
-8.49054456e-01 -1.87659562e-01 2.68170923e-01 -2.97020048e-01
2.24551082e-01 1.06651731e-01 -3.35885227e-01 -1.26023197e+00
-1.15231521e-01 -6.76662087e-01 5.77444792e-01 -4.29418355e-01
-1.50514507e+00 7.63717294e-01 -1.98118668e-02 -1.19982851e+00
5.07574901e-02 -5.66495478e-01 -6.39910638e-01 8.11061382e-01
-2.00537777e+00 -1.81255138e+00 -5.90869963e-01 1.16167438e+00
5.12990236e-01 -2.70193934e-01 6.36870623e-01 4.23883021e-01
-6.65400624e-01 4.58077610e-01 1.45390764e-01 2.08344802e-01
3.69049549e-01 -1.09613776e+00 -4.39773381e-01 6.05323493e-01
-2.38389358e-01 1.16006680e-01 1.47909835e-01 -4.40649539e-01
-1.46362400e+00 -1.10424376e+00 1.30390495e-01 1.35935724e-01
6.89969599e-01 2.21331641e-01 -9.11850631e-01 4.37580794e-01
2.27957577e-01 7.90386856e-01 9.33936059e-01 -6.35794699e-02
-1.54989824e-01 -1.36944577e-01 -1.27755582e+00 8.50118324e-02
6.82839274e-01 -4.44483191e-01 -6.78183079e-01 3.41431946e-01
6.37010694e-01 -5.99741280e-01 -1.27048838e+00 7.35499859e-01
3.81111324e-01 -6.47334456e-01 8.57536256e-01 -5.05032420e-01
4.29324508e-01 -3.30029309e-01 -1.94114640e-01 -1.42717171e+00
-2.65536040e-01 7.88258463e-02 1.97834298e-01 1.10039568e+00
2.57659167e-01 -6.44531131e-01 3.84019881e-01 4.00023222e-01
-5.38354278e-01 -1.12833250e+00 -1.33272219e+00 -4.39568698e-01
3.08394164e-01 -1.35387644e-01 7.32886314e-01 9.65673149e-01
-1.45246595e-01 -5.91979213e-02 -1.59562062e-02 -2.09449120e-02
4.69782501e-01 4.02147532e-01 1.41194746e-01 -1.23030710e+00
-7.43590668e-02 -6.58003271e-01 -6.55070007e-01 -3.63227725e-01
1.22357838e-01 -1.12316275e+00 9.02228281e-02 -1.76296389e+00
5.96087158e-01 -4.11529034e-01 -8.93624246e-01 6.94024622e-01
-2.40017101e-01 1.48729468e-02 8.54517519e-02 1.64220229e-01
-6.73579454e-01 7.31973827e-01 1.68867052e+00 -4.51591820e-01
9.66165215e-02 -1.87717661e-01 -7.42015302e-01 7.50279129e-01
5.64968050e-01 -3.06993902e-01 -2.98685312e-01 -5.00631869e-01
-3.83179098e-01 2.52125025e-01 4.21945870e-01 -1.20921588e+00
4.23040718e-01 -1.60878271e-01 6.34697199e-01 -6.53573811e-01
1.27702698e-01 -1.10891771e+00 -1.52758852e-01 5.02292752e-01
-8.17602500e-02 -8.27317014e-02 2.86959559e-01 6.03188157e-01
-5.36705017e-01 7.04221800e-02 8.53592634e-01 -3.28806132e-01
-9.60988939e-01 8.68747830e-01 -2.30189368e-01 2.13274043e-02
1.18602777e+00 -2.08625555e-01 -3.67856950e-01 3.81446064e-01
-8.84159029e-01 3.14185649e-01 3.47954296e-02 5.26772439e-01
7.04828143e-01 -1.65120602e+00 -5.70295215e-01 2.57805169e-01
3.94393504e-01 3.37720513e-01 6.51686907e-01 1.44585228e+00
-1.34649351e-01 2.60590434e-01 -6.08851612e-01 -9.77330625e-01
-9.64562058e-01 4.88967061e-01 4.69003797e-01 -4.59106445e-01
-6.02234542e-01 7.76947677e-01 2.70040274e-01 -3.77632767e-01
3.74858342e-02 -4.65746075e-01 -4.97707993e-01 1.68044046e-01
4.82631594e-01 -8.26460496e-02 2.17070594e-01 -9.35050905e-01
-4.04967517e-01 5.98949194e-01 -4.16833609e-01 1.23474531e-01
1.43457234e+00 -1.18218929e-01 -2.68403172e-01 4.42308605e-01
1.49491441e+00 -6.62524700e-01 -1.04017174e+00 -6.78415120e-01
-8.04186016e-02 -2.82926202e-01 4.58461076e-01 -1.02129877e+00
-1.73468673e+00 8.85773957e-01 1.10819721e+00 -1.63196924e-03
1.51215506e+00 -3.58475349e-03 1.22729325e+00 1.46526277e-01
2.66411394e-01 -1.02843344e+00 -6.58875108e-02 3.86807501e-01
6.63389444e-01 -1.69704986e+00 -2.50808597e-01 -2.71409601e-01
-6.81634307e-01 1.17909646e+00 6.89360023e-01 3.66703495e-02
1.10403967e+00 2.21254185e-01 1.14161201e-01 -4.38771904e-01
-4.45355475e-01 -4.81381327e-01 3.76557171e-01 4.70551163e-01
2.23677665e-01 1.32298306e-01 -2.39667967e-01 9.74569261e-01
1.78376704e-01 3.40677559e-01 5.33620939e-02 1.10300946e+00
-5.77681720e-01 -1.07168150e+00 -4.36296731e-01 8.69690418e-01
-3.94121736e-01 1.44948721e-01 -4.13392484e-02 7.47112930e-01
5.45912921e-01 7.47313619e-01 -1.84350297e-01 -4.69599396e-01
2.19534785e-01 1.24211848e-01 5.49717546e-01 -3.03956211e-01
-5.35670638e-01 1.81718886e-01 -5.88577032e-01 -4.52870756e-01
-7.25479782e-01 -7.08534062e-01 -1.43817925e+00 1.38399124e-01
-3.32052946e-01 2.53756382e-02 6.93959177e-01 1.48569798e+00
1.38195371e-02 1.13636315e+00 7.47133195e-01 -8.10215473e-01
-2.68593162e-01 -8.75389934e-01 -5.25824010e-01 4.51708704e-01
3.98900151e-01 -1.03240073e+00 9.78107303e-02 -7.15218186e-02] | [14.484533309936523, -2.411526918411255] |
216533a8-e6a7-4b0e-971f-382168bc1df4 | globalmind-global-multi-head-interactive-self | 2304.08687 | null | https://arxiv.org/abs/2304.08687v1 | https://arxiv.org/pdf/2304.08687v1.pdf | GlobalMind: Global Multi-head Interactive Self-attention Network for Hyperspectral Change Detection | High spectral resolution imagery of the Earth's surface enables users to monitor changes over time in fine-grained scale, playing an increasingly important role in agriculture, defense, and emergency response. However, most current algorithms are still confined to describing local features and fail to incorporate a global perspective, which limits their ability to capture interactions between global features, thus usually resulting in incomplete change regions. In this paper, we propose a Global Multi-head INteractive self-attention change Detection network (GlobalMind) to explore the implicit correlation between different surface objects and variant land cover transformations, acquiring a comprehensive understanding of the data and accurate change detection result. Firstly, a simple but effective Global Axial Segmentation (GAS) strategy is designed to expand the self-attention computation along the row space or column space of hyperspectral images, allowing the global connection with high efficiency. Secondly, with GAS, the global spatial multi-head interactive self-attention (Global-M) module is crafted to mine the abundant spatial-spectral feature involving potential correlations between the ground objects from the entire rich and complex hyperspectral space. Moreover, to acquire the accurate and complete cross-temporal changes, we devise a global temporal interactive multi-head self-attention (GlobalD) module which incorporates the relevance and variation of bi-temporal spatial-spectral features, deriving the integrate potential same kind of changes in the local and global range with the combination of GAS. We perform extensive experiments on five mostly used hyperspectral datasets, and our method outperforms the state-of-the-art algorithms with high accuracy and efficiency. | ['Liangpei Zhang', 'Chen Wu', 'Meiqi Hu'] | 2023-04-18 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 4.10993606e-01 -6.87351406e-01 1.26742125e-01 -3.66605520e-01
-4.30504203e-01 -5.19873917e-01 4.11483347e-01 5.53255603e-02
-2.23559886e-01 4.84560579e-01 -1.14830710e-01 -9.52507257e-02
-6.32352293e-01 -1.14525354e+00 -5.21585584e-01 -1.00399017e+00
-2.28249580e-01 4.37585190e-02 1.23513117e-01 -5.44313431e-01
-5.57113327e-02 6.09014452e-01 -1.68984687e+00 -5.05993292e-02
1.31527233e+00 1.00071120e+00 5.54721832e-01 3.02296788e-01
-1.43782854e-01 6.19356334e-02 3.67546491e-02 3.10188949e-01
3.82433355e-01 -4.74404901e-01 -6.31602585e-01 4.17381495e-01
3.52209449e-01 -2.23128036e-01 -2.23786272e-02 1.59298003e+00
5.67039132e-01 3.35839331e-01 3.30135584e-01 -8.84535074e-01
-4.57962900e-01 3.31321031e-01 -1.06491482e+00 4.44058895e-01
-2.24922255e-01 3.66236299e-01 9.56399679e-01 -7.28081226e-01
3.65320414e-01 1.20717049e+00 5.48783064e-01 -2.99485028e-01
-9.53441083e-01 -6.54898465e-01 7.57105410e-01 5.15192986e-01
-1.43384111e+00 5.44071496e-02 1.00057900e+00 -4.41980720e-01
7.12417901e-01 5.77465594e-01 9.16327417e-01 3.97525519e-01
-4.46552085e-03 6.22071624e-01 1.02885044e+00 -8.51598382e-02
-1.94926083e-01 -3.29104483e-01 3.67337614e-01 3.89393240e-01
1.60703927e-01 8.84358138e-02 4.23779562e-02 3.80169958e-01
7.07011938e-01 3.49158704e-01 -6.28840506e-01 -1.13799870e-01
-1.19529724e+00 6.29483998e-01 1.08577549e+00 7.14677751e-01
-8.29907000e-01 -2.96442837e-01 1.35743797e-01 2.15774581e-01
8.01045954e-01 5.53740561e-01 -7.40552306e-01 3.00136209e-01
-7.81395614e-01 -6.08071731e-03 3.01586650e-02 5.25136888e-01
1.33832848e+00 -8.26115832e-02 -2.18889028e-01 8.67728770e-01
1.27973974e-01 6.05662465e-01 6.38522446e-01 -3.93632382e-01
3.17477435e-01 9.31346834e-01 1.35269493e-01 -1.35383797e+00
-6.76506221e-01 -1.05869937e+00 -1.21630168e+00 7.64048323e-02
-2.84240156e-01 -1.28332019e-01 -9.11519289e-01 1.65830183e+00
6.86029494e-01 2.24326581e-01 -1.99821636e-01 1.02412093e+00
5.72741210e-01 9.21232045e-01 3.01096030e-02 -5.63318014e-01
1.26469088e+00 -8.44640195e-01 -7.55017579e-01 -3.46232474e-01
3.87101024e-01 -1.80518433e-01 1.10034847e+00 -4.85563129e-02
-5.19321918e-01 -7.79358089e-01 -1.01569891e+00 1.41525656e-01
-8.13075125e-01 1.80243549e-03 7.75459111e-01 6.13460951e-02
-7.37605929e-01 6.88983619e-01 -7.23732889e-01 -5.26624858e-01
4.94660676e-01 2.90776730e-01 -2.77184695e-01 -1.22828610e-01
-1.39330328e+00 5.19822419e-01 7.00893402e-01 7.47547984e-01
-5.07722855e-01 -7.65769005e-01 -7.46348977e-01 1.80190757e-01
5.61386824e-01 -4.05270129e-01 6.12663627e-01 -1.32240582e+00
-1.26968300e+00 5.85799277e-01 -1.29782725e-02 -1.50699792e-02
2.13832036e-01 -2.14811891e-01 -5.62829077e-01 -2.05203984e-02
1.73813522e-01 5.12560189e-01 7.03505576e-01 -1.22290015e+00
-8.33640814e-01 -8.60842705e-01 -8.93270373e-02 8.07077467e-01
-4.49619770e-01 -2.93380201e-01 -3.79282206e-01 -6.16560876e-01
5.23649573e-01 -6.85120463e-01 -2.59517789e-01 -2.08943427e-01
-1.40800327e-01 1.56993970e-01 1.09146166e+00 -8.11230958e-01
1.19923615e+00 -2.42446303e+00 2.83299655e-01 1.20163493e-01
-4.66736257e-02 5.49453676e-01 -4.16159987e-01 1.10502154e-01
-5.21320999e-01 2.13417798e-01 -8.17764699e-01 2.58131802e-01
-1.93105876e-01 5.13074882e-02 -1.40421793e-01 3.53433520e-01
3.09521675e-01 9.65323627e-01 -9.31429863e-01 -9.40854996e-02
4.02570009e-01 2.43032664e-01 -4.20126587e-01 1.77638024e-01
-2.65942037e-01 6.91870034e-01 -6.96837008e-01 6.20541632e-01
1.11885417e+00 -1.68223023e-01 -2.25109667e-01 -5.28834343e-01
-6.24844849e-01 -2.83219367e-01 -1.19984841e+00 1.63265276e+00
-2.01930791e-01 2.77648121e-01 1.89709947e-01 -1.02551711e+00
8.31408262e-01 -6.93927407e-02 7.07727432e-01 -6.94271863e-01
-9.81224924e-02 1.55337468e-01 -6.15880303e-02 -6.50899351e-01
2.51325756e-01 -6.83899298e-02 3.67199659e-01 -3.21372300e-02
-2.40252033e-01 -2.18465135e-01 -1.09676123e-01 -3.65317255e-01
5.81995845e-01 1.25295833e-01 4.87211972e-01 -3.80030215e-01
6.91173077e-01 4.10401225e-02 4.51325089e-01 4.89261031e-01
-1.98594987e-01 3.55066985e-01 3.13083418e-02 -3.80522013e-01
-5.68650782e-01 -4.65290755e-01 -2.95500219e-01 9.78747606e-01
5.94350994e-01 2.58985281e-01 -3.69751602e-01 -5.03248930e-01
5.57447076e-02 5.46560645e-01 -7.55211055e-01 -2.33166397e-01
-1.39881715e-01 -1.61638165e+00 -9.23310742e-02 2.85809398e-01
1.24221599e+00 -1.14423180e+00 -4.69408244e-01 2.92330116e-01
-2.47339472e-01 -6.53419197e-01 -2.66863793e-01 1.99402168e-01
-8.93595517e-01 -9.20235097e-01 -5.42527795e-01 -3.94832343e-01
3.18141222e-01 6.31339788e-01 5.39346814e-01 -3.03443134e-01
-4.33690846e-01 4.56076078e-02 -5.19596279e-01 -2.16523916e-01
3.48715395e-01 1.64458349e-01 -4.39066857e-01 5.07227421e-01
1.64174214e-01 -7.13773251e-01 -5.79308450e-01 1.10448882e-01
-1.17261970e+00 1.41658694e-01 6.85417414e-01 9.10930574e-01
7.53309369e-01 6.03091061e-01 4.27694112e-01 -7.76075959e-01
1.72053233e-01 -6.57051563e-01 -5.67952096e-01 3.11853647e-01
-6.66441381e-01 -1.87851697e-01 3.51598084e-01 -1.44731477e-01
-1.17650676e+00 -9.49816555e-02 -2.05860958e-01 -2.73991644e-01
-4.07520741e-01 1.09012139e+00 -6.47953331e-01 -8.60151574e-02
4.94854629e-01 5.35747349e-01 -3.17433983e-01 -4.44895387e-01
2.94762224e-01 6.36816740e-01 5.75351954e-01 -6.72606006e-02
8.97735417e-01 5.13646662e-01 -1.65496796e-01 -9.13037896e-01
-9.83503938e-01 -7.98476219e-01 -8.50991488e-01 -6.69034645e-02
7.89543808e-01 -9.52297747e-01 -2.51133621e-01 1.06604290e+00
-7.27271378e-01 -4.03418094e-01 -3.70164603e-01 4.14811760e-01
-1.66895822e-01 3.87034595e-01 -1.65736109e-01 -6.10936761e-01
-4.63534087e-01 -9.08350289e-01 1.20063722e+00 4.50903535e-01
5.45818269e-01 -8.54343116e-01 -2.12029535e-02 3.70379235e-03
5.30930102e-01 4.69824135e-01 9.87886608e-01 -2.60156155e-01
-6.11226499e-01 5.56169301e-02 -7.09391952e-01 1.93639472e-01
7.35599458e-01 -2.42638253e-02 -8.95192385e-01 -3.69462937e-01
1.04445577e-01 1.72035620e-01 1.07074726e+00 7.21851945e-01
1.44845760e+00 -1.54722378e-01 -3.65153641e-01 1.10288858e+00
1.82977772e+00 3.55543405e-01 5.59805572e-01 4.24483061e-01
9.58714962e-01 6.01163447e-01 8.77104402e-01 5.58832943e-01
1.96865663e-01 3.74059469e-01 9.14242446e-01 -5.34993410e-01
2.17967927e-01 2.18574479e-01 5.60186803e-02 5.39145947e-01
-1.39721990e-01 -2.02023730e-01 -7.02579737e-01 6.97696149e-01
-1.87837362e+00 -1.08986044e+00 -2.66454518e-01 1.99309099e+00
7.70561695e-01 -3.72401774e-01 -3.63288224e-01 3.53557877e-02
9.90526736e-01 6.83583021e-01 -1.07249272e+00 4.06756759e-01
-4.66919124e-01 1.48862600e-01 5.38695455e-01 4.46318716e-01
-1.60157824e+00 1.07506442e+00 4.95920515e+00 8.99910390e-01
-1.43499696e+00 1.37312174e-01 5.15989780e-01 2.37889037e-01
-3.89978170e-01 -1.78981841e-01 -5.23237944e-01 2.93535978e-01
3.16737384e-01 8.58597457e-02 5.44817388e-01 4.18071687e-01
3.37922871e-01 1.54917594e-02 -3.46304566e-01 8.88130784e-01
-2.59795904e-01 -8.07902753e-01 1.94601864e-01 -8.05151016e-02
9.53387916e-01 2.78287381e-01 4.14781906e-02 2.11356595e-01
1.84694648e-01 -6.53218567e-01 3.51701468e-01 7.46579707e-01
9.31549430e-01 -6.95431054e-01 7.52892911e-01 4.11825627e-01
-1.45829916e+00 -3.95493716e-01 -2.81862885e-01 1.97589695e-01
-1.89718410e-01 7.93923199e-01 -1.92710131e-01 1.21282005e+00
7.88911939e-01 1.35499895e+00 -7.06586659e-01 1.01275897e+00
7.08851442e-02 4.81181711e-01 -3.60248864e-01 3.86954963e-01
6.05459750e-01 -5.85966647e-01 7.13359416e-01 1.02564168e+00
5.51550627e-01 5.96993089e-01 1.73350796e-01 8.05471957e-01
2.86143899e-01 2.35624716e-01 -3.36409897e-01 -1.74283221e-01
1.44635037e-01 1.42926824e+00 -5.50130188e-01 -1.72102660e-01
-2.01135620e-01 1.08407271e+00 -3.83114303e-03 6.76838100e-01
-6.73287332e-01 -4.95116323e-01 8.18107128e-01 -9.93496254e-02
4.42177236e-01 -1.34792000e-01 -1.50402144e-01 -1.24822545e+00
-8.81443098e-02 -9.27279413e-01 4.81413990e-01 -1.03884268e+00
-1.30723155e+00 6.35547578e-01 -1.31126612e-01 -1.20888793e+00
3.02731305e-01 -3.48120242e-01 -6.43399298e-01 1.12606275e+00
-2.00231123e+00 -1.35753512e+00 -1.05422735e+00 7.61117041e-01
5.93865275e-01 1.45283058e-01 5.61824620e-01 2.35784948e-01
-9.53574240e-01 1.00347646e-01 2.57982016e-01 -1.98215291e-01
5.62450528e-01 -1.02475739e+00 1.38099402e-01 1.17097282e+00
-2.79777467e-01 1.49759591e-01 4.31711555e-01 -7.03958750e-01
-1.06492472e+00 -1.66149867e+00 1.65785745e-01 3.25102150e-01
6.71532512e-01 3.36924791e-01 -1.20718181e+00 5.00309169e-01
-4.65801433e-02 1.37912095e-01 2.39040822e-01 -1.91998836e-02
8.22956711e-02 -5.90799391e-01 -8.89055908e-01 4.55247730e-01
1.19849586e+00 -5.16622841e-01 -1.38742626e-01 4.96852100e-01
9.69462693e-01 -1.69318572e-01 -6.98124528e-01 1.05153024e+00
1.06924109e-01 -8.45488906e-01 7.66071439e-01 -3.25017095e-01
2.75960475e-01 -7.05382705e-01 -2.73950785e-01 -1.47809410e+00
-9.44471955e-01 -2.89784461e-01 2.67982841e-01 1.21708488e+00
9.88214649e-03 -9.47801232e-01 6.65844083e-02 -1.98912881e-02
-4.10967886e-01 -3.83424461e-01 -5.20628691e-01 -4.27721560e-01
-2.24019334e-01 -3.25218230e-01 1.05424011e+00 1.33254004e+00
-5.83570540e-01 3.89452428e-02 -7.13905543e-02 8.14803839e-01
2.81771302e-01 6.83597684e-01 3.09421182e-01 -1.44085491e+00
-1.61058791e-02 -7.61922061e-01 -3.98128390e-01 -9.74236548e-01
7.61916563e-02 -8.21513295e-01 -4.29310231e-03 -1.50657654e+00
2.22579554e-01 -3.19169611e-01 -6.21960461e-01 6.62057936e-01
-7.55209327e-01 5.66815473e-02 -1.59450904e-01 3.45718324e-01
-1.87482819e-01 1.00903261e+00 1.47901404e+00 -4.60163236e-01
-4.18631911e-01 -5.19010983e-02 -7.68809557e-01 4.45170045e-01
5.70069671e-01 -1.00914262e-01 -3.89767498e-01 -4.13310081e-01
1.08931705e-01 -2.40088001e-01 5.15264750e-01 -1.10824370e+00
7.72522613e-02 -5.84134758e-01 3.82798135e-01 -8.65863860e-01
-1.28939256e-01 -8.68995965e-01 3.07519376e-01 2.66618758e-01
-3.37096155e-02 -3.72738689e-01 3.65424573e-01 5.19644976e-01
-5.56591928e-01 3.80260348e-02 9.22284186e-01 -2.66096115e-01
-1.21496069e+00 1.01294696e+00 4.66151275e-02 -3.35277796e-01
8.96735191e-01 -1.38988391e-01 -9.27496850e-02 -5.07819243e-02
-7.29282796e-01 4.86387670e-01 3.82518500e-01 3.82627994e-01
1.97729915e-01 -1.20763814e+00 -7.16544092e-01 6.69946730e-01
2.76530117e-01 3.65271509e-01 9.04667377e-01 8.16423059e-01
-4.71115828e-01 2.06330165e-01 -3.40135098e-01 -7.44512260e-01
-8.47911477e-01 5.22220790e-01 8.24966967e-01 -3.18301111e-01
-4.65547383e-01 8.17157686e-01 5.62252283e-01 -5.62022924e-01
-5.49621224e-01 -4.31091130e-01 -6.38633013e-01 5.12109518e-01
4.09248799e-01 1.11466266e-01 1.15952998e-01 -6.34680510e-01
-2.11931229e-01 9.14499700e-01 3.72166097e-01 9.56933275e-02
1.58699405e+00 -5.29590845e-01 -4.83080715e-01 4.83956546e-01
1.06096900e+00 -3.31886262e-01 -1.59206092e+00 -5.21910429e-01
-3.79795194e-01 -5.08181095e-01 4.94711548e-01 -7.69644737e-01
-1.32051075e+00 9.28492069e-01 1.09870613e+00 3.02028030e-01
1.80152214e+00 -3.99431974e-01 5.06107032e-01 2.20126733e-01
2.49627471e-01 -8.06022882e-01 -3.52000952e-01 6.91537619e-01
9.80869234e-01 -1.43567741e+00 -6.98193535e-02 -4.15102780e-01
-4.23810273e-01 9.31872785e-01 7.70438075e-01 2.24995419e-01
8.12148988e-01 -2.41512954e-01 -1.64849147e-01 -3.56674165e-01
-2.87743479e-01 -6.88054264e-01 4.12611008e-01 3.93622696e-01
5.05515076e-02 3.28387737e-01 -7.56781548e-03 5.21677136e-01
1.49728149e-01 -2.11180881e-01 -2.07993723e-02 5.64440429e-01
-6.81533813e-01 -6.87922299e-01 -3.45329940e-01 4.75926697e-01
5.44020124e-02 -3.64935607e-01 -7.37700239e-02 6.36283815e-01
5.23983300e-01 6.97905719e-01 1.86430976e-01 -3.32188189e-01
3.05974483e-01 -3.25806602e-03 6.06284067e-02 -5.66190839e-01
-3.85100663e-01 2.81133682e-01 -4.88969654e-01 -5.72548926e-01
-5.95275998e-01 -9.37852502e-01 -9.87050414e-01 6.74444810e-02
-5.26888669e-01 -1.08802937e-01 5.18345833e-01 1.07850420e+00
4.64737892e-01 8.25883627e-01 1.06026435e+00 -9.47935939e-01
-3.28298599e-01 -1.33712459e+00 -1.03678894e+00 2.97851682e-01
5.46127737e-01 -5.51267147e-01 -3.02619874e-01 -8.79604891e-02] | [9.863794326782227, -1.4348989725112915] |
bd6fddf6-f6ea-48e5-b2f3-ef7fc700a301 | progressive-adversarial-networks-for-fine | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Wang_Progressive_Adversarial_Networks_for_Fine-Grained_Domain_Adaptation_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_Progressive_Adversarial_Networks_for_Fine-Grained_Domain_Adaptation_CVPR_2020_paper.pdf | Progressive Adversarial Networks for Fine-Grained Domain Adaptation | Fine-grained visual categorization has long been considered as an important problem, however, its real application is still restricted, since precisely annotating a large fine-grained image dataset is a laborious task and requires expert-level human knowledge. A solution to this problem is applying domain adaptation approaches to fine-grained scenarios, where the key idea is to discover the commonality between existing fine-grained image datasets and massive unlabeled data in the wild. The main technical bottleneck lies in that the large inter-domain variation will deteriorate the subtle boundaries of small inter-class variation during domain alignment. This paper presents the Progressive Adversarial Networks (PAN) to align fine-grained categories across domains with a curriculum-based adversarial learning framework. In particular, throughout the learning process, domain adaptation is carried out through all multi-grained features, progressively exploiting the label hierarchy from coarse to fine. The progressive learning is applied upon both category classification and domain alignment, boosting both the discriminability and the transferability of the fine-grained features. Our method is evaluated on three benchmarks, two of which are proposed by us, and it outperforms the state-of-the-art domain adaptation methods.
| [' Jianmin Wang', ' Mingsheng Long', ' Yunbo Wang', ' Xinyang Chen', 'Sinan Wang'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['fine-grained-visual-categorization'] | ['computer-vision'] | [ 3.08323890e-01 -2.89602041e-01 -2.34570473e-01 -3.84690523e-01
-6.36827528e-01 -1.07806373e+00 6.29328609e-01 1.28166750e-01
-4.41696882e-01 8.30268383e-01 3.88879366e-02 -2.50281077e-02
-2.05576003e-01 -7.48264253e-01 -5.65732658e-01 -9.21796918e-01
3.63285363e-01 6.13079250e-01 5.12098193e-01 -2.86193669e-01
1.59157202e-01 4.73931760e-01 -1.40915382e+00 2.48353332e-01
1.00036442e+00 1.31451261e+00 -1.94103234e-02 3.76365006e-01
-5.00727236e-01 5.09446561e-01 -6.68151677e-01 -4.29670185e-01
3.20450574e-01 -2.74014980e-01 -9.13614333e-01 1.45348653e-01
7.19864845e-01 -1.90770045e-01 -6.20598085e-02 1.39956009e+00
3.16907287e-01 5.58093786e-02 9.38189566e-01 -1.43549013e+00
-8.56119573e-01 1.89678937e-01 -7.43104041e-01 2.08725542e-01
-9.28357095e-02 8.20976719e-02 7.72740185e-01 -5.70431411e-01
3.83642495e-01 1.36997688e+00 5.68419158e-01 6.51792824e-01
-1.30609834e+00 -9.02236044e-01 5.10543525e-01 3.51799577e-01
-1.52340484e+00 -1.09730810e-02 8.53897035e-01 -8.80722821e-01
5.02848625e-01 -1.80863336e-01 2.42018878e-01 1.14456844e+00
2.17549011e-01 2.21155003e-01 1.42807567e+00 -4.77636784e-01
3.40444684e-01 1.93134114e-01 1.52095988e-01 2.80010432e-01
1.02680534e-01 3.45455468e-01 -1.04512975e-01 -6.89598694e-02
8.56681705e-01 1.59382716e-01 2.00574640e-02 -8.71349096e-01
-1.08725071e+00 9.72279012e-01 6.27540708e-01 2.79249609e-01
-9.74653438e-02 -3.08141291e-01 4.85824287e-01 5.55670202e-01
4.73880053e-01 4.13233370e-01 -6.04753494e-01 1.22384824e-01
-6.84017360e-01 2.87936360e-01 4.69405383e-01 7.89597213e-01
9.12678123e-01 -2.14569703e-01 -3.27839464e-01 9.69383359e-01
3.37714665e-02 2.97939688e-01 7.35438883e-01 -6.21556640e-01
3.04010272e-01 8.24069798e-01 1.67909376e-02 -7.37009346e-01
-9.09138992e-02 -4.28275794e-01 -1.16053474e+00 4.89781290e-01
9.01978850e-01 -1.52260363e-01 -1.16011357e+00 1.82434964e+00
4.54070121e-01 8.44806209e-02 -1.91705793e-01 9.14692044e-01
6.57068491e-01 2.85424292e-01 7.30726361e-01 1.92137763e-01
1.32156456e+00 -9.99725938e-01 -3.30881476e-01 -4.50528949e-01
-6.16781563e-02 -7.23380923e-01 1.16112876e+00 7.02559426e-02
-5.17499149e-01 -1.07568645e+00 -1.09624076e+00 -3.79201933e-03
-7.28436828e-01 -3.25294733e-01 3.03709745e-01 5.77492476e-01
-6.36959255e-01 3.75940740e-01 -4.30531412e-01 -3.77553195e-01
6.15106344e-01 3.06714863e-01 -4.79469359e-01 -2.58165389e-01
-1.42123675e+00 7.01317728e-01 7.34804153e-01 -3.52658987e-01
-7.59519100e-01 -8.23722601e-01 -6.90234363e-01 1.55183613e-01
2.79951215e-01 -7.45344102e-01 1.02593803e+00 -1.20958233e+00
-1.27209556e+00 1.13509119e+00 2.40237057e-01 -2.72038966e-01
5.83262026e-01 3.04096732e-02 -5.23097157e-01 -7.90819451e-02
2.82236546e-01 6.95915043e-01 1.32509065e+00 -1.26128614e+00
-9.24164712e-01 -5.43380380e-01 2.48309031e-01 1.81378603e-01
-4.02673453e-01 -3.17777514e-01 -1.61159173e-01 -9.77109551e-01
-2.10354537e-01 -7.61928856e-01 -1.52778342e-01 6.67967647e-02
1.41718816e-02 -3.95950884e-01 8.69591653e-01 -3.34175736e-01
9.50293601e-01 -2.33640790e+00 1.99423939e-01 8.84958804e-02
2.80156463e-01 2.71865457e-01 -2.00826094e-01 1.00941837e-01
-1.58982664e-01 -9.03908908e-02 -1.74457148e-01 2.26412654e-01
1.96921363e-01 1.66237071e-01 -5.32190442e-01 2.77031869e-01
3.86653036e-01 9.13556755e-01 -9.42308187e-01 -5.55807829e-01
3.78085315e-01 1.49068221e-01 -4.94412422e-01 5.53739727e-01
-2.08964691e-01 8.34374607e-01 -5.71148217e-01 5.19866228e-01
8.96951854e-01 -4.41639930e-01 -3.14553306e-02 -2.18487486e-01
2.23902643e-01 -4.61439222e-01 -1.16752458e+00 1.67225182e+00
-3.71746689e-01 1.14254698e-01 -9.81254056e-02 -1.13942945e+00
1.00136745e+00 3.82346101e-02 1.54416606e-01 -9.25294578e-01
-2.68880278e-02 9.50224772e-02 3.70264947e-02 -1.30685344e-01
2.67960727e-01 -6.41824365e-01 -6.16549432e-01 3.26790571e-01
3.21110368e-01 -2.90838867e-01 -1.03419460e-02 -6.62246943e-02
7.46115565e-01 8.79951492e-02 6.70062363e-01 -2.38002881e-01
6.82446539e-01 2.23126292e-01 5.35827518e-01 7.33245313e-01
-6.48269176e-01 4.23534811e-01 2.70758629e-01 -4.74920750e-01
-1.11695457e+00 -1.30999482e+00 -1.66778237e-01 1.62135577e+00
4.16248977e-01 2.74892569e-01 -8.70975435e-01 -1.14552379e+00
3.20981383e-01 1.86037108e-01 -1.02448666e+00 -4.56054300e-01
-2.87943512e-01 -4.95870590e-01 3.86816293e-01 6.95581317e-01
7.02505112e-01 -1.12830794e+00 -3.14581007e-01 1.46217272e-01
-1.47393703e-01 -1.20507288e+00 -5.60611248e-01 4.46802825e-01
-6.68810010e-01 -1.01831734e+00 -8.70834708e-01 -8.74525070e-01
5.30365407e-01 2.01011464e-01 1.44233203e+00 -2.45410487e-01
-3.15058291e-01 -1.74927637e-02 -4.28307533e-01 -2.91207552e-01
-3.91392559e-01 1.29002184e-01 7.94218970e-04 1.04390062e-01
7.21377730e-01 -5.08502841e-01 -6.10695899e-01 6.25713587e-01
-8.58053684e-01 -3.86534661e-01 7.14567602e-01 1.36629879e+00
7.23029733e-01 3.51136237e-01 7.67338872e-01 -1.08991921e+00
5.10803521e-01 -4.22371238e-01 -5.81747234e-01 3.02694559e-01
-4.37036514e-01 1.06742300e-01 9.69291866e-01 -6.98905349e-01
-1.18413281e+00 1.34456754e-02 -7.80823380e-02 -4.90397751e-01
-8.13204288e-01 -2.49279998e-02 -5.32523155e-01 -2.45163679e-01
9.41883802e-01 8.90857279e-02 -3.84281635e-01 -4.66370851e-01
6.13257885e-01 6.05963886e-01 7.63947368e-01 -7.50348151e-01
1.08803666e+00 4.26068246e-01 -2.44866386e-01 -4.51513708e-01
-1.04195440e+00 -6.23471200e-01 -1.15975761e+00 1.75913915e-01
9.16898787e-01 -9.13932443e-01 -2.89646387e-01 6.85521901e-01
-7.18371093e-01 -4.58212405e-01 -5.23531675e-01 -2.03804150e-02
-6.55491948e-01 2.66779244e-01 -2.44999647e-01 -1.74102057e-02
-6.34000152e-02 -1.03222013e+00 1.09665251e+00 3.13043356e-01
-1.82455420e-01 -1.17537737e+00 1.49523795e-01 2.60796666e-01
4.07556981e-01 2.79643953e-01 1.27266693e+00 -6.15329325e-01
-2.11845189e-01 -5.11642843e-02 -6.66319013e-01 3.71096462e-01
4.37807053e-01 -4.71294314e-01 -1.01842058e+00 -5.06210387e-01
-2.55063981e-01 -6.68699920e-01 6.83778048e-01 2.00739190e-01
1.39454067e+00 4.86355200e-02 -3.75501603e-01 5.70611835e-01
1.47965741e+00 9.09352824e-02 2.84549952e-01 5.51650345e-01
7.92957306e-01 5.20047486e-01 8.70083809e-01 2.98187196e-01
1.44083127e-01 8.12590182e-01 3.70252967e-01 -1.88657895e-01
-4.59969848e-01 -3.47596437e-01 -3.23499084e-01 2.42844224e-01
1.53600916e-01 3.00278217e-02 -7.36144364e-01 6.15423679e-01
-1.61277676e+00 -8.67532730e-01 4.19456810e-01 1.98193133e+00
1.00639796e+00 2.38982901e-01 2.36118376e-01 2.51240551e-01
9.49265003e-01 9.73519608e-02 -8.43084812e-01 -3.81837636e-01
1.14695035e-01 3.97679299e-01 4.57938045e-01 1.81388810e-01
-1.53601611e+00 1.13673651e+00 5.63865709e+00 1.10575020e+00
-1.25937748e+00 8.53581801e-02 5.11451423e-01 5.06081522e-01
2.54020870e-01 -5.25033534e-01 -6.84472740e-01 5.39442778e-01
5.40373385e-01 -1.38017923e-01 3.40083718e-01 1.08967757e+00
-3.56014073e-01 2.31555894e-01 -1.08710992e+00 8.57323229e-01
-2.34083131e-01 -8.86793196e-01 1.06474385e-01 -1.50875449e-01
9.69735265e-01 -2.70215601e-01 1.74762368e-01 5.90983808e-01
7.24986672e-01 -1.10031736e+00 6.07972264e-01 4.40265574e-02
1.27678990e+00 -8.16593885e-01 6.48408651e-01 3.70187551e-01
-1.35706115e+00 -2.16661870e-01 -5.36006510e-01 4.02618088e-02
-1.95332542e-01 3.85942996e-01 -6.14734828e-01 3.58446449e-01
1.00424135e+00 5.16967535e-01 -6.23509884e-01 8.24426055e-01
-3.05216700e-01 2.36525759e-01 2.80674517e-01 4.13600117e-01
2.53239959e-01 1.24217801e-01 7.06355795e-02 1.22441781e+00
-1.17215598e-02 -4.26995568e-02 4.97076571e-01 5.34656227e-01
-1.58837020e-01 -2.97232062e-01 -4.01503652e-01 1.74360961e-01
4.56468284e-01 1.29011083e+00 -6.88319743e-01 -2.70963579e-01
-4.64164615e-01 1.15866661e+00 5.95180631e-01 2.57704586e-01
-7.40435779e-01 -4.37800765e-01 9.73814726e-01 5.07779196e-02
5.40637851e-01 7.58219957e-02 -4.41015005e-01 -1.09100080e+00
-2.47264773e-01 -1.07329524e+00 8.56187820e-01 -2.95815617e-01
-1.99628210e+00 6.81306243e-01 -1.01186000e-01 -1.35041428e+00
-3.29356283e-01 -6.98259175e-01 -2.69683033e-01 1.16884148e+00
-1.87877691e+00 -1.38776040e+00 -7.81559289e-01 1.15223730e+00
6.24085248e-01 -1.33478642e-01 1.00823033e+00 4.60334271e-01
-1.02976754e-01 9.80522573e-01 2.10676223e-01 3.91257465e-01
1.19564819e+00 -1.49991143e+00 4.68657762e-01 6.78199708e-01
-2.00752616e-01 3.45965445e-01 4.85825479e-01 -3.99284214e-01
-7.92928874e-01 -1.53194880e+00 5.41287303e-01 -5.65247476e-01
8.53833020e-01 -3.88935894e-01 -1.17493618e+00 4.81998563e-01
-2.92019863e-02 5.08364737e-01 6.03768706e-01 -2.28506699e-02
-8.62430871e-01 -2.92951256e-01 -1.34519148e+00 2.81751186e-01
8.77484918e-01 -6.44504130e-01 -9.62071419e-01 -2.17524357e-02
6.53649509e-01 -4.57358718e-01 -9.39769447e-01 4.26113218e-01
5.60261965e-01 -7.50555277e-01 1.22985220e+00 -6.71692789e-01
3.73329997e-01 -4.32058394e-01 -1.62823498e-01 -1.75273561e+00
-7.57270932e-01 5.39162531e-02 1.09896116e-01 1.31846237e+00
-1.96680680e-01 -5.68407118e-01 7.55868137e-01 2.57152200e-01
1.37366474e-01 -1.61414564e-01 -7.65911877e-01 -9.04353321e-01
7.05693364e-01 1.71962067e-01 7.78532803e-01 1.09249246e+00
-4.22701538e-01 4.17051286e-01 -1.45728961e-01 1.36543334e-01
8.44402730e-01 5.79805195e-01 8.12981367e-01 -1.61104321e+00
-2.48013437e-01 -4.84518588e-01 -4.98021096e-01 -8.31365526e-01
2.98548549e-01 -5.35284936e-01 1.79506108e-01 -1.07305956e+00
3.23078573e-01 -7.15569973e-01 -5.92916667e-01 3.74760240e-01
-4.13551211e-01 6.92158580e-01 1.94124684e-01 1.03639394e-01
-5.73847234e-01 3.35628182e-01 1.56478965e+00 -6.17803156e-01
4.95213233e-02 -1.39778763e-01 -9.45738137e-01 6.95890367e-01
6.26889169e-01 -4.70138311e-01 -4.76619810e-01 -1.88720003e-01
-3.68650079e-01 -2.53394067e-01 4.92166817e-01 -1.03014421e+00
1.86301962e-01 -4.99305010e-01 7.42655337e-01 -2.75339991e-01
-4.58186027e-04 -1.08468950e+00 -2.56598536e-02 2.23948732e-01
-2.45460078e-01 -3.82716924e-01 4.46749806e-01 5.97509503e-01
-5.60729384e-01 9.42672566e-02 1.45453238e+00 -2.32316315e-01
-1.46167779e+00 5.52841723e-01 2.48937637e-01 5.62659860e-01
1.22495246e+00 -2.67326206e-01 -3.65578085e-01 1.58161610e-01
-8.34468544e-01 1.53976111e-02 6.09855175e-01 6.55100942e-01
1.67813361e-01 -1.38205123e+00 -6.94688380e-01 3.65481049e-01
6.85549855e-01 7.58346394e-02 6.01075113e-01 1.21631570e-01
3.78269069e-02 3.51734310e-01 -8.76816750e-01 -7.77941287e-01
-1.11280167e+00 1.08787918e+00 3.81045043e-01 -5.93442440e-01
-2.74896711e-01 9.57614958e-01 9.18532789e-01 -5.51611900e-01
1.85937256e-01 2.48649400e-02 -4.42937136e-01 1.46121159e-01
5.25977790e-01 9.16894302e-02 2.54812449e-01 -5.71101725e-01
-3.18043739e-01 1.13356185e+00 -2.65488982e-01 4.55395192e-01
1.06011212e+00 -4.40178186e-01 1.60696417e-01 9.75053832e-02
1.08733356e+00 -2.69911319e-01 -1.65640461e+00 -5.31284153e-01
-4.65399623e-02 -5.76750815e-01 -2.32541531e-01 -1.20948946e+00
-8.39985013e-01 1.15464413e+00 9.38033521e-01 2.72892326e-01
1.43462276e+00 1.70827746e-01 5.20893097e-01 -7.29355812e-02
5.23212016e-01 -9.86015618e-01 2.62943745e-01 6.50650561e-01
6.65967584e-01 -1.58590317e+00 -2.32230067e-01 -4.24954534e-01
-5.79183042e-01 9.18375731e-01 8.37271810e-01 -2.70721436e-01
6.09418571e-01 9.51874331e-02 3.29656631e-01 1.79802850e-01
-3.24887842e-01 -2.10130677e-01 5.14759302e-01 9.92047846e-01
2.28206247e-01 2.01529995e-01 4.70676497e-02 7.68868029e-01
-6.40427619e-02 4.53010872e-02 -3.02314144e-02 5.82775593e-01
-3.26335996e-01 -1.36531663e+00 -5.38898766e-01 2.48889461e-01
-4.82762694e-01 6.06649369e-02 -3.16869378e-01 8.36774111e-01
5.66713274e-01 8.16967905e-01 1.84483513e-01 -1.85681984e-01
5.28488815e-01 -3.11685055e-02 4.95507926e-01 -6.53394818e-01
-4.82734352e-01 -2.53822595e-01 -4.17080700e-01 -4.32772338e-01
-3.34213495e-01 -3.81548911e-01 -8.90996337e-01 -3.86867613e-01
1.92708686e-01 9.59038213e-02 2.37038463e-01 9.68800187e-01
2.15647191e-01 7.19020069e-01 6.82611406e-01 -7.73265719e-01
-8.45272362e-01 -8.36910486e-01 -7.44215012e-01 9.49388564e-01
5.25049567e-01 -1.10363269e+00 -2.40143552e-01 2.53663808e-01] | [9.947298049926758, 2.4785122871398926] |
608ad736-a58d-4514-ab89-b33a8e374fac | aesop-abstract-encoding-of-stories-objects | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Ravi_AESOP_Abstract_Encoding_of_Stories_Objects_and_Pictures_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Ravi_AESOP_Abstract_Encoding_of_Stories_Objects_and_Pictures_ICCV_2021_paper.pdf | AESOP: Abstract Encoding of Stories, Objects, and Pictures | Visual storytelling and story comprehension are uniquely human skills that play a central role in how we learn about and experience the world. Despite remarkable progress in recent years in synthesis of visual and textual content in isolation and learning effective joint visual-linguistic representations, existing systems still operate only at a superficial, factual level. With the goal of developing systems that are able to comprehend rich human-generated narratives, and co-create new stories, we introduce AESOP: a new dataset that captures the creative process associated with visual storytelling. Visual panels are composed of clip-art objects with specific attributes enabling a broad range of creative expression. Using AESOP, we propose foundational storytelling tasks that are generative variants of story cloze tests, to better measure the creative and causal reasoning ability required for visual storytelling. We further develop a generalized story completion framework that models stories as the co-evolution of visual and textual concepts. We benchmark the proposed approach with human baselines and evaluate using comprehensive qualitative and quantitative metrics. Our results highlight key insights related to the dataset, modelling and evaluation of visual storytelling for future research in this promising field of study. | ['Mubbasir Kapadia', 'Jonathan Brandt', 'Scott Cohen', 'Kushal Kafle', 'Hareesh Ravi'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['visual-storytelling', 'story-completion'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.05918133e-01 3.72586548e-02 5.85810766e-02 -1.23470113e-01
-4.04411048e-01 -9.86214578e-01 1.45539773e+00 3.66196007e-01
2.11290002e-01 4.24640387e-01 1.14209890e+00 -9.84504074e-02
-1.20351970e-01 -7.64308870e-01 -6.25533462e-01 -1.09152727e-01
8.33854005e-02 3.48234326e-01 -1.32738411e-01 -3.92709762e-01
4.46354121e-01 1.50018111e-01 -1.72796476e+00 1.07053757e+00
8.14870656e-01 5.18009484e-01 3.20203751e-01 9.60679531e-01
-3.68475497e-01 1.85990989e+00 -7.27240503e-01 -5.26115000e-01
-2.19911158e-01 -8.81761432e-01 -9.07352746e-01 1.45416781e-01
5.65293431e-01 -1.61963090e-01 -3.66687000e-01 5.97168207e-01
3.35643709e-01 2.47755200e-01 8.85986149e-01 -1.30880368e+00
-1.14181566e+00 1.04840326e+00 -2.68500626e-01 1.22575201e-02
1.07298064e+00 5.41859388e-01 1.23196936e+00 -9.68273222e-01
1.20435607e+00 1.30452061e+00 5.94725490e-01 4.39993978e-01
-1.38664496e+00 -4.22721207e-01 6.86993673e-02 3.97320151e-01
-9.25575256e-01 -2.63134271e-01 8.83372843e-01 -1.09358680e+00
1.00974667e+00 4.55358595e-01 1.23269141e+00 1.63395119e+00
-3.07611495e-01 1.15978217e+00 1.27478504e+00 -2.82984018e-01
1.55607676e-02 1.63268089e-01 -2.39620700e-01 6.34752810e-01
-2.49327257e-01 1.17664367e-01 -1.23048866e+00 3.51619720e-01
9.79350746e-01 -2.75402457e-01 -2.27230623e-01 -4.52691376e-01
-1.75120378e+00 8.46582472e-01 4.43241626e-01 5.16687870e-01
-3.93786699e-01 2.84646779e-01 5.74573994e-01 2.03253001e-01
1.42009035e-01 9.08740044e-01 5.67410171e-01 -5.26226103e-01
-1.05837452e+00 7.33218074e-01 5.51350296e-01 1.03812826e+00
6.66849762e-02 1.40499055e-01 -8.65881979e-01 7.48379230e-01
2.81047914e-02 2.62845904e-01 2.03414708e-01 -7.69220769e-01
4.64186370e-01 9.25479591e-01 -1.47999063e-01 -9.55551744e-01
-1.70413032e-01 -4.38002825e-01 -5.03209949e-01 4.65595126e-01
2.40558013e-01 2.91958630e-01 -5.64666927e-01 1.50414026e+00
-1.08376712e-01 1.64107047e-02 -5.27163781e-02 8.57582510e-01
1.46060050e+00 8.56021166e-01 2.76720703e-01 1.14595398e-01
1.47644246e+00 -1.00410378e+00 -7.66100883e-01 -3.06946784e-01
3.08264971e-01 -6.99336469e-01 1.85633934e+00 4.22119141e-01
-1.50515294e+00 -5.21631718e-01 -1.34544563e+00 -5.05965948e-01
-4.83449906e-01 -1.30591597e-02 6.23564959e-01 2.18194962e-01
-9.50241506e-01 4.14580584e-01 -3.61388892e-01 -4.01228368e-01
7.60277033e-01 -7.09035814e-01 -4.18087244e-01 -5.77567257e-02
-9.15260196e-01 1.03952026e+00 5.53805470e-01 -4.83116418e-01
-1.25574064e+00 -1.23776412e+00 -8.96758914e-01 -1.07775696e-01
3.46900195e-01 -9.14298356e-01 1.32091010e+00 -8.41202259e-01
-1.10417783e+00 9.95064020e-01 2.59414017e-01 -2.36493140e-01
7.74524987e-01 -2.84584284e-01 -2.13937685e-01 1.24371417e-01
2.19425574e-01 5.90253770e-01 5.93084395e-01 -1.64142597e+00
-2.91391224e-01 2.25201368e-01 2.42880121e-01 2.33032867e-01
-2.89201498e-01 8.08461159e-02 -1.78330734e-01 -1.10248303e+00
-5.33069551e-01 -2.68288165e-01 2.25384638e-01 -9.33552906e-02
-5.29123425e-01 6.57296330e-02 6.71934247e-01 -7.90638447e-01
1.27801394e+00 -2.04513121e+00 5.01970708e-01 -5.44455610e-02
7.17960179e-01 -2.27595091e-01 -1.57755926e-01 9.98936176e-01
-3.87836806e-02 3.02047461e-01 -1.49962410e-01 -6.27603412e-01
2.72252232e-01 -3.33637297e-01 -4.83218640e-01 -1.90828905e-01
1.73110127e-01 1.30393875e+00 -1.22790170e+00 -6.18532360e-01
4.10667837e-01 6.22674644e-01 -4.72363174e-01 3.04878294e-01
-5.90078592e-01 4.64350373e-01 -5.55222780e-02 4.72056061e-01
4.15837578e-03 -3.59082490e-01 -1.37863541e-02 5.39979041e-02
-3.12030673e-01 2.95855254e-01 -8.04004610e-01 2.06310749e+00
-5.55539846e-01 1.39313388e+00 -6.68047130e-01 -3.44870687e-01
7.65630424e-01 3.54422599e-01 -1.05453774e-01 -8.23459148e-01
7.59522840e-02 -3.94023687e-01 -4.06018347e-01 -8.30357552e-01
7.25069702e-01 -5.17976582e-01 -2.18182102e-01 6.86248004e-01
8.28526393e-02 -5.13939202e-01 3.96363080e-01 6.79933071e-01
1.15863192e+00 4.80564058e-01 5.60454905e-01 1.44182369e-01
-8.34160969e-02 4.27502483e-01 -3.04254562e-01 8.47592175e-01
3.17667902e-01 7.89725244e-01 8.49929690e-01 -2.26248249e-01
-1.32054949e+00 -1.47067702e+00 2.47713253e-01 1.23305047e+00
-1.06290936e-01 -7.89436340e-01 -4.72523004e-01 -2.35138744e-01
-1.77884713e-01 1.33962131e+00 -9.19004917e-01 1.67390019e-01
-4.02747244e-01 -5.72330877e-02 6.09174967e-01 6.65891111e-01
3.78688812e-01 -1.51634765e+00 -1.11489141e+00 5.16372658e-02
-3.50815564e-01 -1.13338435e+00 -3.34421732e-02 -3.25895727e-01
-3.25541347e-01 -9.80779350e-01 -6.52108192e-01 -6.71012402e-01
2.39787981e-01 2.94837683e-01 1.51063049e+00 5.77947833e-02
-4.25482303e-01 5.60164034e-01 -6.49652660e-01 -3.45097631e-01
-8.13947082e-01 -4.81946468e-01 -5.72649121e-01 -2.55632132e-01
-3.68219286e-01 -8.41323793e-01 -4.20898050e-01 -3.52737039e-01
-1.06930649e+00 1.20281315e+00 3.98847491e-01 6.08814061e-01
1.34797141e-01 -1.29295707e-01 1.23946093e-01 -8.90871406e-01
1.05642366e+00 -6.54577613e-01 4.05701771e-02 3.53682190e-01
-1.07046925e-01 -4.69921716e-02 4.97332096e-01 -5.88439286e-01
-1.26395154e+00 -4.25411224e-01 2.74873048e-01 -4.35403109e-01
1.26312226e-02 7.76526392e-01 1.17942190e-03 4.23014820e-01
8.75557363e-01 8.89192373e-02 -3.75208110e-01 -1.76991522e-01
1.04432034e+00 -6.64846227e-02 1.05657232e+00 -6.20636523e-01
1.00565076e+00 4.01468635e-01 -1.47399962e-01 -6.94687665e-01
-6.06772065e-01 -2.21617743e-01 -3.74803871e-01 -1.09906089e+00
8.98831189e-01 -9.64771748e-01 -8.12533021e-01 8.68279114e-02
-1.29112482e+00 -5.26044190e-01 -5.88540316e-01 8.24216902e-02
-8.37749183e-01 9.58985612e-02 -3.17725241e-01 -6.84423208e-01
-1.53562045e-02 -5.36046088e-01 9.83278096e-01 8.69320333e-02
-1.05516875e+00 -1.07089055e+00 3.27213913e-01 3.68063301e-01
2.00364158e-01 1.29218602e+00 1.34490085e+00 -2.78183997e-01
-5.48096359e-01 -4.36955616e-02 -3.72800678e-01 -2.42154911e-01
-4.26363736e-01 8.16281065e-02 -9.54646230e-01 2.03726113e-01
-3.59247804e-01 -8.30064893e-01 9.01659548e-01 -7.30216280e-02
6.95772350e-01 -4.14008826e-01 -1.53765932e-01 3.12600285e-01
1.44036222e+00 9.15151387e-02 7.84783006e-01 3.96986842e-01
9.11806703e-01 8.17176759e-01 4.05512094e-01 7.81067312e-01
5.56338966e-01 5.15169084e-01 3.33648354e-01 -2.42874548e-02
-7.07244992e-01 -1.04687333e+00 4.09478784e-01 5.85626245e-01
-3.34623873e-01 -3.85640830e-01 -1.29978728e+00 7.57982016e-01
-1.82014298e+00 -1.57789624e+00 -3.01775783e-01 1.59920931e+00
8.33048820e-01 1.06422462e-01 3.08529615e-01 2.88860559e-01
2.93845534e-01 5.81879854e-01 -1.78765327e-01 -2.93448985e-01
-4.50607896e-01 2.74475604e-01 -3.08412999e-01 4.35721189e-01
-6.55332685e-01 1.06562054e+00 6.44414139e+00 6.97616577e-01
-8.63378465e-01 1.01343267e-01 5.94736561e-02 -5.35164297e-01
-7.58629143e-01 -1.74784623e-02 1.02093890e-01 -5.83181344e-02
3.12217712e-01 -2.92449266e-01 6.21039927e-01 6.45491838e-01
1.68029100e-01 -6.69144690e-02 -1.48032165e+00 1.14610600e+00
5.04132748e-01 -1.85123193e+00 3.91198486e-01 -4.61177140e-01
9.15821135e-01 -7.09807336e-01 3.40530246e-01 3.42254996e-01
5.38283229e-01 -1.59399211e+00 1.52549148e+00 9.24281359e-01
1.01118577e+00 -5.16788602e-01 4.77261841e-02 1.34201109e-01
-1.12874544e+00 8.96682143e-02 3.32640827e-01 -4.83111829e-01
2.72142828e-01 1.03951663e-01 -9.46999669e-01 2.52247155e-01
4.74975079e-01 1.00703371e+00 -8.76811683e-01 8.91342402e-01
-6.53163731e-01 4.01636362e-01 4.76711214e-01 -4.05340403e-01
5.86237423e-02 4.13037777e-01 8.41089129e-01 1.47416127e+00
3.46768707e-01 5.86645566e-02 -2.17175111e-01 1.50568736e+00
3.33392359e-02 2.09092066e-01 -7.57046938e-01 -8.27442110e-01
4.84630078e-01 1.02359903e+00 -8.08003545e-01 -3.65594655e-01
-2.01562107e-01 8.82256925e-01 5.14335573e-01 3.69279832e-01
-7.89667845e-01 -7.11856782e-02 4.04627204e-01 3.27427626e-01
-3.65899615e-02 -3.01724911e-01 -7.68606544e-01 -9.43175018e-01
2.89586023e-03 -1.00145853e+00 9.05563012e-02 -1.54290771e+00
-1.17276621e+00 4.43159401e-01 3.05688053e-01 -1.06866813e+00
-3.43249291e-01 -3.26650560e-01 -1.02385604e+00 4.19069380e-01
-8.03394258e-01 -1.59047806e+00 -7.99517512e-01 4.60676998e-01
8.20860028e-01 -9.37409699e-02 6.67890728e-01 -3.23755980e-01
-2.97409240e-02 1.80424064e-01 -3.31034213e-01 -8.79975036e-02
4.03983623e-01 -1.41983938e+00 6.85748458e-01 8.49605620e-01
5.15763700e-01 4.01209474e-01 1.22647846e+00 -8.38075459e-01
-1.12087774e+00 -7.54417658e-01 8.24089587e-01 -8.18543911e-01
8.61916959e-01 -8.23479116e-01 -6.48894429e-01 6.13003969e-01
6.68917000e-01 -8.35163295e-01 7.76224256e-01 1.71759069e-01
-7.88882613e-01 7.02576697e-01 -7.25009203e-01 1.03788471e+00
1.57110786e+00 -8.22012126e-01 -1.01761925e+00 2.34886885e-01
8.04217339e-01 -2.37529308e-01 -5.86039841e-01 -1.02347732e-01
7.74426222e-01 -1.16799283e+00 1.06355238e+00 -6.52969062e-01
1.60113955e+00 -2.08073571e-01 1.26043335e-01 -1.49437034e+00
-4.05943453e-01 -7.72960603e-01 5.02400137e-02 1.52671921e+00
2.51963407e-01 1.89251438e-01 4.09203082e-01 8.35764185e-02
-1.97622329e-01 -2.13328674e-01 -4.87510353e-01 -5.45525610e-01
-7.97187611e-02 -8.15353394e-01 5.82481444e-01 1.22172308e+00
5.78648448e-01 4.66987401e-01 -4.34952199e-01 -4.19913709e-01
2.58484006e-01 2.32818201e-02 9.71056104e-01 -1.15108705e+00
-3.19690168e-01 -1.00885975e+00 -3.08219522e-01 -3.62238467e-01
1.49032660e-02 -1.20921504e+00 -3.04021128e-02 -2.10931087e+00
6.77687109e-01 2.93766856e-01 2.41477638e-01 4.00980860e-01
2.38101520e-02 2.62314647e-01 6.59363687e-01 1.03663221e-01
-7.11239398e-01 4.09699857e-01 1.59277439e+00 -1.52711615e-01
-1.64121181e-01 -8.89585614e-01 -7.22744167e-01 6.92361474e-01
3.87027085e-01 -8.50633681e-02 -7.13817894e-01 -3.37249666e-01
1.00473523e+00 2.43069053e-01 9.72323835e-01 -1.15006018e+00
5.45538887e-02 -2.99489975e-01 5.09716749e-01 -5.17307341e-01
3.65809143e-01 -3.61571282e-01 4.89003003e-01 3.80929410e-02
-7.79776216e-01 1.77098975e-01 2.00029194e-01 3.16266596e-01
-1.72819450e-01 1.73301145e-01 3.60522509e-01 -9.26311091e-02
-8.19878459e-01 -2.97418952e-01 -5.42878270e-01 3.20464492e-01
1.14893997e+00 -5.17898679e-01 -7.09788561e-01 -8.44821870e-01
-6.57117665e-01 8.40115622e-02 6.17222846e-01 7.94577837e-01
9.20969427e-01 -1.82352126e+00 -1.01148391e+00 -2.64958471e-01
6.01528347e-01 -3.93055469e-01 3.61173213e-01 2.94273257e-01
-7.89673090e-01 4.55182604e-02 -6.64079726e-01 -1.79276869e-01
-1.27798152e+00 6.24924779e-01 -1.65781081e-01 -2.05524430e-01
-8.89262140e-01 9.49373662e-01 3.54167491e-01 2.56180406e-01
1.36367112e-01 -1.91807136e-01 -5.44624627e-01 3.33556592e-01
7.79986560e-01 4.52956468e-01 -5.86501539e-01 -6.14648759e-01
1.01772614e-01 3.03816974e-01 3.10909927e-01 -6.03278935e-01
1.31736672e+00 9.54704061e-02 -1.20007908e-02 9.94077802e-01
7.88540959e-01 6.31566644e-02 -1.18440795e+00 1.40963579e-02
1.91038176e-02 -5.56939185e-01 -2.52016038e-01 -1.34462225e+00
-3.36155444e-01 1.11264968e+00 7.47496635e-02 2.96558380e-01
8.70480120e-01 4.66623425e-01 4.27530050e-01 -4.80724759e-02
-2.68751718e-02 -7.79640138e-01 7.92000055e-01 4.02583838e-01
1.83804607e+00 -8.41182232e-01 6.93673342e-02 -2.67521054e-01
-1.10243809e+00 9.90739584e-01 4.70174909e-01 -2.05397233e-02
-1.97492704e-01 2.72508413e-01 -2.12551922e-01 -6.14324808e-01
-9.77304041e-01 -3.02280217e-01 7.53474176e-01 7.14506984e-01
5.76897323e-01 3.32633853e-01 9.35463831e-02 8.10152292e-01
-9.12783980e-01 -9.41645876e-02 4.37769234e-01 6.77725494e-01
-1.97175682e-01 -4.68584687e-01 -4.24653918e-01 7.58126974e-02
1.87699363e-01 -8.48674253e-02 -1.06618869e+00 1.07044590e+00
1.64524525e-01 8.00491750e-01 1.06693685e-01 -4.81411576e-01
3.44273776e-01 1.83295324e-01 8.47899735e-01 -5.04319191e-01
-7.89767504e-01 -4.13060546e-01 4.55659717e-01 -4.68747795e-01
-3.33511978e-01 -7.19911039e-01 -1.14345384e+00 -6.53454363e-01
5.42267203e-01 -4.57877010e-01 4.35823172e-01 8.48404169e-01
-1.34681955e-01 9.57138240e-01 -1.70139253e-01 -8.00845325e-01
4.03867513e-01 -9.25960302e-01 -2.55563438e-01 8.47150087e-01
1.39368892e-01 -7.44618833e-01 3.12926471e-02 4.46632892e-01] | [11.190916061401367, 0.9023920893669128] |
855fd7cc-7c40-410b-9cd0-aaa80e01a21a | change-detection-under-global-viewpoint | 1703.00552 | null | http://arxiv.org/abs/1703.00552v1 | http://arxiv.org/pdf/1703.00552v1.pdf | Change Detection under Global Viewpoint Uncertainty | This paper addresses the problem of change detection from a novel perspective
of long-term map learning. We are particularly interested in designing an
approach that can scale to large maps and that can function under global
uncertainty in the viewpoint (i.e., GPS-denied situations). Our approach, which
utilizes a compact bag-of-words (BoW) scene model, makes several contributions
to the problem:
1) Two kinds of prior information are extracted from the view sequence map
and used for change detection. Further, we propose a novel type of prior,
called motion prior, to predict the relative motions of stationary objects and
anomaly ego-motion detection. The proposed prior is also useful for
distinguishing stationary from non-stationary objects.
2) A small set of good reference images (e.g., 10) are efficiently retrieved
from the view sequence map by employing the recently developed
Bag-of-Local-Convolutional-Features (BoLCF) scene model.
3) Change detection is reformulated as a scene retrieval over these reference
images to find changed objects using a novel spatial Bag-of-Words (SBoW) scene
model. Evaluations conducted of individual techniques and also their
combinations on a challenging dataset of highly dynamic scenes in the publicly
available Malaga dataset verify their efficacy. | ['Tanaka Kanji', 'Murase Tomoya'] | 2017-03-01 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 2.77389407e-01 -6.17407620e-01 -1.38540808e-02 -3.50962192e-01
-7.77243316e-01 -7.26620615e-01 9.96528566e-01 -3.72582711e-02
-2.16041535e-01 5.25530636e-01 2.71307200e-01 1.20060019e-01
-2.17803866e-01 -8.10975015e-01 -1.02490687e+00 -9.20423090e-01
-2.14336544e-01 1.83778182e-01 8.17731500e-01 -3.99537683e-01
4.77231205e-01 6.76023126e-01 -1.87484145e+00 7.49634206e-02
5.09821296e-01 9.06011462e-01 6.99516952e-01 9.59954381e-01
3.14658582e-01 9.79375124e-01 -1.41647398e-01 1.52437374e-01
4.20973748e-01 -1.10363197e-02 -6.65701926e-01 2.85307199e-01
8.34608138e-01 -6.13428116e-01 -5.08978248e-01 9.59029853e-01
3.82875621e-01 4.02017355e-01 6.41121924e-01 -1.05735338e+00
-3.46188486e-01 -2.22149432e-01 -4.16110039e-01 9.75730062e-01
4.14000303e-01 9.74895712e-03 7.65361965e-01 -1.11879432e+00
8.19077432e-01 1.10923421e+00 5.11594296e-01 -1.27529785e-01
-9.44705784e-01 -1.79913908e-01 4.41784114e-01 7.86278248e-01
-1.47212684e+00 -7.36835301e-01 9.04967189e-01 -8.60831320e-01
1.04637003e+00 4.45802808e-01 5.02854526e-01 7.13794589e-01
5.60406506e-01 7.05051720e-01 8.28332186e-01 -4.29023057e-01
3.42982620e-01 -3.63083631e-02 -5.18020689e-02 5.17870367e-01
1.38330564e-01 3.38147581e-02 -5.69311023e-01 -1.60137508e-02
7.07448304e-01 2.32054725e-01 -4.82650757e-01 -1.00055397e+00
-1.41227746e+00 6.73847258e-01 4.28463161e-01 3.59428138e-01
-4.64522004e-01 2.39097774e-01 1.38416588e-01 1.75858945e-01
8.17451239e-01 -1.41838100e-02 -3.57640356e-01 -6.89014420e-02
-9.29813862e-01 3.14961463e-01 4.26246256e-01 1.09333611e+00
9.77876306e-01 -5.27045913e-02 -1.36376888e-01 5.83423197e-01
2.33910859e-01 7.94931889e-01 4.22666967e-01 -5.53052008e-01
4.90721941e-01 8.55785236e-02 5.99217892e-01 -1.63858163e+00
-2.76814520e-01 -4.01040375e-01 -6.70153558e-01 1.13086798e-03
-4.49327752e-02 5.79637825e-01 -9.46970522e-01 1.61250794e+00
5.85374773e-01 6.42309427e-01 -7.69169033e-02 6.22987688e-01
7.11169779e-01 8.74357700e-01 -5.21549821e-01 -2.40915999e-01
1.10674512e+00 -1.05395031e+00 -6.80329204e-01 -4.50831205e-01
4.21632886e-01 -5.02255380e-01 5.92967093e-01 6.09373003e-02
-8.71353209e-01 -8.67745459e-01 -1.07453012e+00 1.72724172e-01
-9.30430532e-01 -8.86273235e-02 3.65220994e-01 3.28355491e-01
-1.63139677e+00 2.18035355e-01 -9.96470451e-01 -8.11246455e-01
1.15382448e-01 2.31464401e-01 -4.41913545e-01 -4.11520272e-01
-9.27962840e-01 1.10510874e+00 4.67976063e-01 1.41113520e-01
-1.20287240e+00 -3.51592362e-01 -1.05080438e+00 -8.17024559e-02
3.53467166e-01 -6.07219756e-01 9.36364770e-01 -8.75958860e-01
-1.20961082e+00 8.12561333e-01 -4.63025838e-01 -2.80447870e-01
4.86340046e-01 -3.40228200e-01 -5.26180148e-01 3.37007821e-01
3.44921589e-01 6.01733804e-01 1.20948553e+00 -1.40268290e+00
-9.67649937e-01 -3.30790609e-01 1.11482404e-01 5.12740314e-01
1.64479434e-01 -7.15094954e-02 -7.07273602e-01 -6.75686777e-01
4.76804376e-01 -1.01168168e+00 -1.31394729e-01 -2.49564350e-01
1.51019037e-01 1.05767526e-01 9.62463081e-01 -9.77984965e-01
1.06085014e+00 -2.13518882e+00 1.09013543e-01 1.08590275e-01
-1.41416444e-02 9.02980492e-02 -7.93665275e-02 5.34374058e-01
-6.78208247e-02 -2.88527071e-01 -2.24105135e-01 -2.20332786e-01
-2.89951891e-01 1.94225088e-01 -5.05785465e-01 8.14479172e-01
1.03102615e-02 8.39433312e-01 -1.10875058e+00 -1.86359182e-01
7.07064509e-01 2.05079615e-01 -4.06631410e-01 6.71263561e-02
6.57741129e-02 7.55959332e-01 -3.85865092e-01 5.83259463e-01
1.03931069e+00 -1.37790799e-01 -4.15798545e-01 -8.89064092e-03
-3.59865248e-01 -5.37627302e-02 -1.36074030e+00 1.82482719e+00
-1.74604475e-01 6.90477729e-01 -3.29224885e-01 -8.80489528e-01
7.35762358e-01 -7.29917064e-02 6.60840273e-01 -6.52253211e-01
-2.51167744e-01 2.41514277e-02 -5.54051936e-01 -6.04090691e-01
9.10946608e-01 4.30590034e-01 -3.95674780e-02 -1.41391173e-01
-4.21554707e-02 -9.52593908e-02 6.95040822e-02 1.83947012e-01
1.24256158e+00 3.33564192e-01 6.07801497e-01 -2.91575879e-01
7.72324800e-01 8.32587332e-02 5.29803693e-01 1.03479695e+00
-4.57090437e-01 6.69197321e-01 -2.52133429e-01 -6.03686035e-01
-9.74974751e-01 -1.08180165e+00 -1.04507789e-01 1.04511917e+00
5.24828553e-01 -1.75769866e-01 -1.23717681e-01 -4.49849397e-01
-1.17543014e-02 5.79625010e-01 -4.78855312e-01 -4.48201671e-02
-6.10456884e-01 -7.27020502e-01 -1.47661671e-01 4.16524082e-01
5.98168254e-01 -6.79886758e-01 -8.62160742e-01 2.79075056e-01
-4.96050119e-01 -1.19027030e+00 -3.06839049e-01 -5.07643782e-02
-7.58418083e-01 -8.69280994e-01 -5.61177552e-01 -7.53913164e-01
4.64940697e-01 1.02894104e+00 1.02865303e+00 -6.20212317e-01
-5.69086522e-02 1.14567924e+00 -5.49544215e-01 -8.70328546e-02
-4.68688644e-03 -4.82126981e-01 1.68570548e-01 4.12241787e-01
1.46777764e-01 -5.53576589e-01 -6.66052341e-01 3.36557716e-01
-9.05751169e-01 5.97550683e-02 4.43300545e-01 6.84323013e-01
6.89921677e-01 3.30350637e-01 2.74165988e-01 -3.80157351e-01
-1.97630167e-01 -7.60078311e-01 -7.85596967e-01 2.95791745e-01
-4.76039827e-01 -2.86049545e-01 9.17569101e-02 -1.74949378e-01
-1.21116352e+00 1.21484138e-01 1.84938416e-01 -4.98270661e-01
-2.69455850e-01 3.76495600e-01 -1.94481850e-01 -3.59075785e-01
4.26228642e-01 7.73153722e-01 -4.70241904e-01 -2.93268889e-01
5.74609399e-01 5.08971095e-01 8.22790861e-01 5.87410331e-02
8.87944579e-01 1.10717714e+00 2.34007463e-03 -9.52524185e-01
-5.47440588e-01 -1.30672300e+00 -1.09306479e+00 -5.40219903e-01
8.10020864e-01 -1.42746150e+00 -5.25152907e-02 7.24912405e-01
-1.03524411e+00 -2.48432849e-02 -2.24002395e-02 5.31040668e-01
-9.41862702e-01 5.45048356e-01 -2.58671641e-01 -7.74400413e-01
8.90372321e-04 -6.96200490e-01 1.36862791e+00 6.89023212e-02
2.65868068e-01 -1.03781211e+00 4.70807493e-01 1.61705494e-01
4.69346583e-01 4.60388422e-01 4.29439217e-01 -3.89426500e-01
-1.21632624e+00 -2.67659545e-01 -6.06530793e-02 1.10719413e-01
2.66208261e-01 -3.68454754e-01 -1.11714685e+00 -7.79457748e-01
3.53007019e-01 1.87939718e-01 1.06915450e+00 6.75567150e-01
6.08610630e-01 -1.88090965e-01 -6.53694868e-01 8.27513576e-01
1.80431032e+00 4.01696295e-01 7.18030214e-01 6.58021092e-01
8.37996185e-01 1.84550583e-01 9.31090474e-01 5.21829784e-01
6.04491949e-01 9.05863941e-01 5.57389557e-01 1.69056356e-01
-1.62415095e-02 -1.06954373e-01 5.37394345e-01 8.10732961e-01
6.65522292e-02 -4.01208133e-01 -1.01494050e+00 9.43201423e-01
-2.17043114e+00 -1.04993188e+00 -1.71458691e-01 2.21410942e+00
2.25840017e-01 -2.61041820e-02 -2.60497451e-01 -6.59069046e-02
8.49313974e-01 7.23468602e-01 -4.95503396e-01 3.39016765e-01
-2.27925286e-01 -3.23406279e-01 8.03323925e-01 4.87534046e-01
-1.58592987e+00 9.94502246e-01 5.77846479e+00 6.01837277e-01
-9.73231077e-01 4.33097750e-01 1.07328109e-01 2.45548919e-01
9.79115367e-02 7.58715644e-02 -9.26013410e-01 4.96372163e-01
7.64413953e-01 5.00871129e-02 2.74244756e-01 8.66491556e-01
2.17231005e-01 -6.37098968e-01 -9.98881757e-01 1.32985473e+00
6.81077302e-01 -1.29994977e+00 -2.66231783e-03 -2.17469901e-01
9.74244475e-01 4.03268009e-01 -3.39608751e-02 2.55127847e-01
1.23121412e-02 -3.16404402e-01 1.01321948e+00 1.01577938e+00
5.22111118e-01 -5.20472765e-01 8.10474336e-01 4.18495506e-01
-1.31184411e+00 -9.36681107e-02 -4.81364220e-01 -2.15764288e-02
1.08904339e-01 4.56109881e-01 -9.91761565e-01 1.06130826e+00
1.04079366e+00 1.26047647e+00 -1.02275634e+00 1.33520484e+00
2.11887151e-01 3.43027979e-01 -2.48627707e-01 4.07864213e-01
3.59274447e-01 2.04808246e-02 9.51506436e-01 1.27938688e+00
4.96099055e-01 -1.10184841e-01 2.85468906e-01 5.12387276e-01
5.37888348e-01 -9.68993157e-02 -9.18514669e-01 4.54824686e-01
2.21229464e-01 1.00913858e+00 -8.53481650e-01 -3.14422518e-01
-4.56664383e-01 1.24700153e+00 4.35705110e-02 4.81990010e-01
-5.44340134e-01 -8.59145671e-02 3.67698669e-01 -5.53391427e-02
9.42277253e-01 -4.76659805e-01 3.51645052e-01 -1.38234246e+00
2.76326090e-01 -3.64410520e-01 3.87614042e-01 -1.14622521e+00
-1.01497149e+00 2.77141631e-01 3.61905038e-01 -1.56878829e+00
-3.05155218e-01 -2.97660738e-01 -4.77159977e-01 8.10740411e-01
-1.90416992e+00 -1.38558555e+00 -7.64567018e-01 8.04211795e-01
7.66081691e-01 -1.46286130e-01 5.03851473e-01 2.25056201e-01
-5.27013317e-02 -9.51319709e-02 8.77815127e-01 -2.11732358e-01
7.10126281e-01 -1.17150879e+00 6.26722872e-01 1.38903785e+00
1.33357778e-01 -5.38202003e-02 7.11194694e-01 -7.98614919e-01
-1.33531892e+00 -1.49741924e+00 7.67436624e-01 -8.83246660e-01
4.85326111e-01 -4.68620062e-01 -8.19412053e-01 9.32934642e-01
-1.07496470e-01 5.70513643e-02 2.54216284e-01 -2.54858464e-01
-1.93220135e-02 -2.45010912e-01 -8.99889886e-01 3.45809519e-01
1.07150090e+00 -5.95604599e-01 -7.26284146e-01 4.15770501e-01
6.39001846e-01 -5.06739318e-01 -3.84604365e-01 5.44391215e-01
2.72354126e-01 -8.87525678e-01 1.18782926e+00 -3.39975148e-01
-8.88898149e-02 -7.30116427e-01 -7.08710492e-01 -1.31640637e+00
-9.22085881e-01 -3.46610099e-01 -3.42322826e-01 1.10878789e+00
-4.47327435e-01 -6.25892103e-01 3.56505930e-01 -3.98838520e-02
-2.98629522e-01 -9.32487473e-02 -1.04386842e+00 -9.68202829e-01
-5.27856886e-01 -4.48852897e-01 2.96788961e-01 9.53123808e-01
-5.59745371e-01 1.30994558e-01 -6.83044612e-01 7.05319941e-01
4.66023237e-01 1.57497525e-02 9.84542072e-01 -1.05550659e+00
-6.76266477e-02 4.80354838e-02 -1.12006557e+00 -1.22604954e+00
-8.42795074e-02 -7.89235234e-01 3.53146851e-01 -1.60113406e+00
2.15284243e-01 -2.12283999e-01 -5.24160504e-01 -3.88900079e-02
-1.31547540e-01 -4.73050885e-02 7.12565258e-02 6.36454344e-01
-9.46677923e-01 6.42598867e-01 8.52758408e-01 -2.10299090e-01
-1.28479213e-01 1.19111262e-01 -1.26828209e-01 6.39458239e-01
4.71087307e-01 -4.33736980e-01 -4.45065916e-01 -3.67658705e-01
3.07304651e-01 1.14917673e-01 7.26612806e-01 -1.24920809e+00
4.13829267e-01 -1.70122549e-01 3.30015659e-01 -1.51322711e+00
4.68903452e-01 -8.63303959e-01 2.85483271e-01 2.35925809e-01
1.41198933e-01 2.07438514e-01 8.75046626e-02 1.16188407e+00
-2.58051783e-01 -1.13113299e-01 6.75651670e-01 -1.61992654e-01
-1.57899094e+00 3.04014653e-01 -3.36873382e-01 -3.08785230e-01
1.26061106e+00 -3.33125561e-01 -2.47545317e-01 -6.10602498e-01
-6.76116943e-01 1.97572470e-01 4.97734040e-01 6.70679331e-01
5.66961050e-01 -1.43621278e+00 -6.77913487e-01 3.42581481e-01
6.46595180e-01 -2.18112376e-02 5.76124609e-01 8.78976166e-01
-6.79842472e-01 7.95276105e-01 -2.72180855e-01 -1.19882321e+00
-1.10203588e+00 9.21152294e-01 1.99548304e-01 -1.03207015e-01
-7.88201094e-01 7.16166556e-01 5.29556036e-01 -1.84964761e-01
9.83981192e-02 -4.23662692e-01 -2.92615354e-01 9.29834247e-02
7.13347495e-01 4.33865547e-01 1.89220279e-01 -1.07657540e+00
-6.67424500e-01 7.25080669e-01 -1.97333880e-02 -4.83773425e-02
1.47040701e+00 -8.52066815e-01 -8.01765323e-02 8.14583838e-01
1.09985709e+00 -9.19421576e-03 -1.46994817e+00 -7.26978540e-01
5.58640882e-02 -9.77294266e-01 2.39518166e-01 -4.80178952e-01
-5.04243374e-01 5.60526788e-01 1.22112203e+00 9.83889177e-02
1.13855076e+00 6.21576197e-02 2.60962039e-01 6.23386860e-01
7.34090567e-01 -1.03345990e+00 8.93980116e-02 7.75217474e-01
9.71659005e-01 -1.52392459e+00 -3.98069955e-02 -2.63958871e-01
-4.60951954e-01 8.48535061e-01 3.11904252e-01 -1.56045109e-01
8.42552960e-01 -2.49065176e-01 -3.31513345e-01 -2.03362420e-01
-5.26203692e-01 -5.36354601e-01 3.81756991e-01 7.26981103e-01
-5.12770951e-01 -1.62535980e-02 2.48441190e-01 1.94135457e-02
1.39370322e-01 -1.54517114e-01 5.85588098e-01 1.34989846e+00
-7.60152459e-01 -2.61834741e-01 -6.41736567e-01 1.76561460e-01
-6.45242110e-02 -4.96804863e-02 2.83235516e-02 6.50204301e-01
2.92585194e-01 1.02623940e+00 2.46379614e-01 -3.54872912e-01
1.83197856e-01 1.16502168e-03 3.66573989e-01 -5.22131503e-01
1.47154659e-01 1.14796504e-01 -1.62545636e-01 -7.56984115e-01
-7.65446365e-01 -9.18246865e-01 -4.74747002e-01 -1.33754713e-02
-2.08184719e-01 -2.61645675e-01 6.46180987e-01 9.18087602e-01
3.70619595e-01 4.68099803e-01 9.04031813e-01 -1.26350629e+00
-1.57638460e-01 -9.02913034e-01 -7.55132079e-01 2.86280423e-01
7.94213951e-01 -9.21695054e-01 -4.74669516e-01 3.78916770e-01] | [7.730288505554199, -1.9069499969482422] |
4de678d0-f33d-425c-b57c-2da490a5a9f9 | a-shallow-neural-network-for-native-language | null | null | https://aclanthology.org/W17-5027 | https://aclanthology.org/W17-5027.pdf | A Shallow Neural Network for Native Language Identification with Character N-grams | This paper describes the systems submitted by GadjahMada team to the Native Language Identification (NLI) Shared Task 2017. Our models used a continuous representation of character n-grams which are learned jointly with feed-forward neural network classifier. Character n-grams have been proved to be effective for style-based identification tasks including NLI. Results on the test set demonstrate that the proposed model performs very well on essay and fusion tracks by obtaining more than 0.8 on both F-macro score and accuracy. | ['Meisyarah Dwiastuti', 'Yunita Sari', 'Muhammad Rifqi Fatchurrahman'] | 2017-09-01 | null | null | null | ws-2017-9 | ['native-language-identification'] | ['natural-language-processing'] | [-2.20292658e-01 -2.68378377e-01 -5.36446691e-01 -5.71374953e-01
-7.32044339e-01 -7.98144102e-01 8.40274572e-01 7.47690275e-02
-5.96453726e-01 6.26812339e-01 2.68669665e-01 -4.35916781e-01
2.00960636e-01 -3.80906641e-01 -1.17503040e-01 -1.31991088e-01
2.39750847e-01 6.82359040e-01 -4.74195838e-01 -1.49725303e-01
5.43378592e-01 5.31248331e-01 -1.12810063e+00 3.51991594e-01
8.53355050e-01 8.22023034e-01 -2.85201252e-01 1.21368325e+00
-3.92280310e-01 9.35549796e-01 -8.58626604e-01 -7.02906072e-01
6.97517321e-02 -2.37342730e-01 -9.17580068e-01 -4.72684503e-01
8.22183490e-01 -5.05869746e-01 -4.96469170e-01 1.02919507e+00
3.38905752e-01 1.58006623e-02 1.01432681e+00 -6.01276338e-01
-1.07655835e+00 1.10660386e+00 -2.46339396e-01 9.28215981e-02
5.50517857e-01 -1.12262897e-01 1.14229584e+00 -1.15381694e+00
4.83949214e-01 1.35786355e+00 9.53516722e-01 6.86645508e-01
-1.06785214e+00 -8.20793748e-01 -3.12714249e-01 1.79121703e-01
-1.07890773e+00 -3.67970914e-01 6.52751923e-01 -5.28363466e-01
1.35471153e+00 3.13797951e-01 -1.37413833e-02 1.38270676e+00
1.33855462e-01 1.28589249e+00 1.31275856e+00 -9.64584351e-01
-1.85231507e-01 6.96034670e-01 1.27639151e+00 8.03500712e-01
1.44607753e-01 -8.24279897e-03 -8.33078802e-01 -1.48202270e-01
3.60996515e-01 -2.90213734e-01 1.94504082e-01 6.21436954e-01
-1.05522275e+00 1.08596063e+00 -1.12218112e-01 6.17909372e-01
-4.49403748e-02 -2.34874547e-01 7.05068827e-01 5.64423621e-01
4.87936527e-01 6.08891368e-01 -3.84673417e-01 -4.01580960e-01
-1.14312065e+00 8.83715823e-02 1.04841614e+00 9.26990509e-01
1.36803240e-01 2.14360267e-01 -4.84038383e-01 1.33918977e+00
1.43850669e-01 3.99575293e-01 1.18590748e+00 -4.86459702e-01
4.16529655e-01 5.75929761e-01 -1.10720154e-02 -7.40853369e-01
-2.44431555e-01 -5.39078712e-01 -1.01968253e+00 9.44806784e-02
6.83335364e-01 -1.49701223e-01 -5.06896853e-01 1.33276570e+00
-5.06531000e-01 -2.41090074e-01 9.87910181e-02 4.14426804e-01
9.84595954e-01 8.53591561e-01 2.42510483e-01 2.94046164e-01
1.48042154e+00 -1.17596233e+00 -8.18826556e-01 -1.00821018e-01
6.84916258e-01 -1.06526566e+00 1.05401444e+00 5.99391043e-01
-1.00751328e+00 -1.04426336e+00 -1.15790606e+00 6.21797098e-03
-6.30005002e-01 1.21314001e+00 5.14950871e-01 1.13274455e+00
-9.42363977e-01 6.71138704e-01 -1.22570790e-01 -5.39782166e-01
2.85756856e-01 2.78776586e-01 -4.47440594e-01 3.14044029e-01
-1.12459743e+00 7.91892588e-01 4.69434083e-01 -1.87037095e-01
-4.63200510e-01 -5.58692753e-01 -3.33641052e-01 1.63823813e-01
-4.04490173e-01 -7.94602334e-02 1.31905103e+00 -9.64142442e-01
-2.01800323e+00 1.34093451e+00 -2.67350465e-01 -7.76346087e-01
3.84630144e-01 -3.93003047e-01 -6.20844781e-01 -1.28389269e-01
-3.34902457e-03 5.03230453e-01 5.49471736e-01 -6.05145872e-01
-6.29408479e-01 -2.83898324e-01 -2.17910394e-01 2.63469741e-02
-1.20467508e+00 5.43371439e-01 1.61893204e-01 -7.21604347e-01
-2.50663370e-01 -9.34189260e-01 4.95949060e-01 -5.38087130e-01
-5.41755080e-01 -8.50809693e-01 6.66861296e-01 -1.34786212e+00
1.49231458e+00 -1.82507241e+00 -1.12337358e-01 1.75177678e-02
-8.38501379e-02 6.58553421e-01 -8.88424963e-02 2.40206659e-01
-2.93387398e-02 8.82613659e-02 2.76424527e-01 -6.69448495e-01
1.54798478e-01 -7.37818629e-02 -4.32323337e-01 1.53621271e-01
-4.54293527e-02 9.97552335e-01 -6.08933032e-01 -2.02734083e-01
1.32398745e-02 -2.23654229e-03 4.67259698e-02 7.38637626e-01
5.12230434e-02 -1.02157585e-01 1.27709106e-01 7.23889291e-01
3.17208141e-01 5.35001308e-02 -8.72513801e-02 1.25608385e-01
-1.29562914e-01 4.24243808e-01 -6.59409285e-01 1.46498430e+00
-4.70494151e-01 8.57062757e-01 -1.66367531e-01 -6.52539253e-01
1.33664560e+00 3.75259042e-01 -3.33179116e-01 -1.69131353e-01
2.18070135e-01 4.26776946e-01 8.60579759e-02 7.34006660e-03
6.69236243e-01 1.13695771e-01 -2.35123128e-01 5.73209524e-01
6.19678915e-01 3.47240657e-01 2.01118618e-01 3.26847225e-01
5.40911078e-01 -7.27750733e-02 3.35387409e-01 -5.71562350e-01
1.14636564e+00 -5.33361509e-02 1.67908058e-01 1.03626621e+00
-2.60856658e-01 4.76998687e-01 1.73509806e-01 -3.49846959e-01
-1.13998222e+00 -7.15060174e-01 5.59778847e-02 1.44058990e+00
-7.63643980e-01 -1.13572136e-01 -8.96761894e-01 -8.11981082e-01
3.95963388e-03 9.80354726e-01 -6.55992389e-01 -1.23521179e-01
-5.14560282e-01 -4.10395950e-01 1.25152719e+00 7.47153044e-01
4.89555955e-01 -9.91047740e-01 1.35784477e-01 1.68551713e-01
8.08673352e-02 -9.42466199e-01 -6.83988810e-01 2.53519952e-01
-6.87258661e-01 -6.68955863e-01 -1.00222600e+00 -1.11883879e+00
1.78929880e-01 -4.22187410e-02 1.09416127e+00 -1.68519020e-01
-3.57153952e-01 2.46973813e-01 -4.32776451e-01 -5.96184671e-01
-7.16681719e-01 7.68228829e-01 1.02373928e-01 1.15370348e-01
1.06789100e+00 5.55372611e-02 -8.07869807e-03 -3.14088687e-02
-2.17424363e-01 -6.86005875e-02 4.02425319e-01 1.08304405e+00
-8.64680186e-02 -4.42869633e-01 7.51838803e-01 -9.71340239e-01
1.09295726e+00 -1.53568238e-01 -4.96219039e-01 6.65188134e-01
-7.44300961e-01 -2.38879755e-01 8.50788713e-01 -6.90157056e-01
-1.18044984e+00 -1.85503531e-02 -2.19663158e-01 -9.07204971e-02
-3.53275329e-01 2.71748066e-01 -1.16855145e-01 -2.73983181e-01
5.33347964e-01 4.25224185e-01 2.96321604e-02 -7.94528663e-01
2.04720825e-01 1.27461243e+00 7.53588796e-01 -5.14321744e-01
5.02771437e-01 -3.39335620e-01 -5.10485291e-01 -9.73578393e-01
-8.58876288e-01 -7.36541808e-01 -9.47034001e-01 -1.15404017e-01
7.95939982e-01 -9.38730836e-01 -7.52159655e-01 1.05791962e+00
-1.23295271e+00 -1.20945210e-02 1.42929733e-01 5.06751895e-01
-1.44938841e-01 5.41157424e-01 -1.12260544e+00 -1.00346661e+00
-1.12674642e+00 -8.24263811e-01 3.56993556e-01 5.27831316e-01
-9.62761343e-01 -1.33114743e+00 2.89653838e-01 4.97503281e-01
6.68355405e-01 -4.71591502e-01 1.18791389e+00 -1.41953254e+00
3.90609562e-01 -7.33771682e-01 -2.49868020e-01 6.51669502e-01
-1.12537503e-01 1.74411550e-01 -1.36994731e+00 -2.19577163e-01
-1.97318152e-01 -5.39013028e-01 6.81450009e-01 1.47720337e-01
1.21428692e+00 -3.26405078e-01 1.20109797e-01 5.45767248e-01
1.19187367e+00 -9.33050290e-02 2.21045762e-01 3.70495617e-01
7.58436263e-01 5.80356538e-01 8.22786763e-02 3.01904082e-01
1.04925945e-01 5.42845547e-01 -2.91991711e-01 3.24749887e-01
-4.14320588e-01 -2.93376088e-01 4.42834496e-01 1.07440233e+00
3.27596009e-01 -1.53266355e-01 -1.05893636e+00 4.67260361e-01
-1.61927664e+00 -8.63523483e-01 -3.71694237e-01 1.91555238e+00
9.76933122e-01 2.37209320e-01 3.79147738e-01 2.26132900e-01
5.79830408e-01 -2.22317219e-01 -2.96665132e-01 -1.18471146e+00
-4.86243099e-01 4.04620439e-01 6.57595038e-01 6.37587845e-01
-1.27698958e+00 1.21040905e+00 7.18746519e+00 1.12834215e+00
-9.50876951e-01 4.70701456e-02 9.03350413e-01 4.94973250e-02
5.75584546e-02 -4.60775554e-01 -1.49673843e+00 6.65645719e-01
1.57472968e+00 -3.27462316e-01 3.31222892e-01 1.02120662e+00
-3.09154212e-01 4.90183644e-02 -9.96255875e-01 1.05513334e+00
3.65692317e-01 -1.24656034e+00 2.27665246e-01 -1.28475413e-01
8.03535104e-01 -1.51223034e-01 2.18265921e-01 6.97039306e-01
4.61669415e-01 -1.40542662e+00 5.31684875e-01 4.69987541e-01
9.20466065e-01 -8.47082555e-01 8.36690485e-01 3.99190068e-01
-5.50576925e-01 -1.42376363e-01 -5.66040158e-01 -1.14082225e-01
-4.30409938e-01 3.65694106e-01 -9.43751991e-01 2.83660173e-01
2.57362872e-01 8.21889222e-01 -8.07632864e-01 7.05474973e-01
-1.71883032e-01 1.30425370e+00 -6.90009966e-02 -5.70937216e-01
5.44536233e-01 -1.85190856e-01 3.30656350e-01 1.77060258e+00
8.96814540e-02 -3.13197315e-01 4.25571017e-02 7.70254433e-01
-1.83149874e-01 4.04499948e-01 -5.27087867e-01 -3.26404780e-01
4.14270908e-01 1.48045325e+00 -2.29752943e-01 -7.56423354e-01
-4.05694753e-01 1.27002692e+00 4.76886690e-01 -3.83135863e-02
-2.75739759e-01 -7.56603122e-01 5.78520536e-01 -2.49880284e-01
-1.11848943e-01 -2.36616343e-01 -9.83086050e-01 -1.08360744e+00
-4.44116145e-01 -1.07275414e+00 4.69161391e-01 -3.64630729e-01
-1.63988936e+00 7.97883093e-01 -3.50010455e-01 -9.25103128e-01
-5.53996086e-01 -1.37996590e+00 -8.67469132e-01 1.47987866e+00
-1.30242920e+00 -1.44825947e+00 -8.47689435e-02 2.38721371e-01
8.58123422e-01 -1.32645309e+00 1.53207922e+00 1.61341429e-01
-9.25499499e-01 1.43538606e+00 7.71780610e-01 8.11308861e-01
1.01374435e+00 -1.45558715e+00 5.30202270e-01 5.90796232e-01
4.34306800e-01 8.00811112e-01 3.15169990e-01 -5.14312387e-01
-1.30540609e+00 -9.24233198e-01 1.48405504e+00 -4.12860274e-01
8.45018506e-01 -2.54967332e-01 -8.53514552e-01 5.81964850e-01
5.17613828e-01 -5.68670213e-01 1.02722526e+00 2.50316203e-01
-7.57225871e-01 1.70713246e-01 -1.10570133e+00 1.36753783e-01
3.19388449e-01 -9.09232676e-01 -6.74402118e-01 4.67129856e-01
6.41486943e-02 -3.62494960e-02 -9.11211431e-01 -1.51345581e-01
8.12144399e-01 -6.69216156e-01 8.99681985e-01 -9.66713071e-01
5.69798112e-01 3.91931027e-01 -9.33014508e-03 -1.21904075e+00
-6.47906542e-01 -3.98781806e-01 -3.34289402e-01 1.66238606e+00
5.70713639e-01 -5.02306700e-01 6.49627686e-01 5.43314815e-01
3.63983363e-01 -5.47340333e-01 -6.68414176e-01 -9.85758543e-01
6.54630542e-01 -6.20997585e-02 1.67981058e-01 1.06540668e+00
2.59150982e-01 7.82436073e-01 -3.38223487e-01 -5.01839578e-01
5.54381311e-01 -2.16701657e-01 5.95232368e-01 -1.42682445e+00
-3.49091828e-01 -1.07607138e+00 -1.69830725e-01 -7.39051402e-01
8.14300299e-01 -1.15863168e+00 -3.82322907e-01 -8.80620360e-01
5.04040360e-01 -5.55918030e-02 -3.03681195e-01 4.96359974e-01
-3.18916112e-01 4.54362273e-01 1.16289176e-01 4.18423861e-01
-3.25493157e-01 -1.59201957e-02 1.74663067e-01 -4.38030392e-01
2.42237389e-01 1.55284241e-01 -5.99814296e-01 6.45458519e-01
1.26089871e+00 -4.70837802e-02 -3.58503982e-02 -3.96650523e-01
-7.41319507e-02 -2.49453694e-01 -1.87202275e-01 -1.06471586e+00
5.67348786e-02 2.80348182e-01 7.18908429e-01 -7.97522247e-01
1.57762393e-01 -2.64218152e-01 -4.89815593e-01 5.34559846e-01
-9.92938578e-01 5.31139560e-02 2.46465147e-01 3.15590836e-02
-1.45621076e-01 -9.05576825e-01 7.84106791e-01 -1.53759688e-01
-5.86965203e-01 7.16433860e-03 -6.23274803e-01 -2.50843734e-01
5.88850021e-01 -2.94848710e-01 -4.89439890e-02 -3.23093772e-01
-4.14163440e-01 -1.93428218e-01 2.82188386e-01 6.76278353e-01
2.73682475e-01 -1.26593411e+00 -1.23611629e+00 4.96622086e-01
2.09732980e-01 -1.27147853e+00 -3.24635863e-01 3.16156119e-01
-5.96745372e-01 9.30539370e-01 -4.10696566e-01 -1.78942904e-01
-1.80040419e+00 5.33378087e-02 1.98637664e-01 -6.37426615e-01
-3.26079309e-01 1.30596745e+00 -4.53389525e-01 -8.44517171e-01
6.35591626e-01 2.24673375e-01 -6.83711290e-01 1.19884767e-01
1.12649846e+00 6.97212100e-01 4.60765734e-02 -6.97155714e-01
-2.16911688e-01 2.37754330e-01 -6.02443695e-01 -1.98916256e-01
1.06649792e+00 6.98926300e-02 -3.12067449e-01 8.55996490e-01
1.54595768e+00 -6.76507652e-02 -3.75083983e-01 -5.00107348e-01
3.62061977e-01 -1.41494215e-01 8.44846591e-02 -1.37536502e+00
-4.13005173e-01 1.20651019e+00 7.24395752e-01 3.13960761e-02
3.18821281e-01 -7.44802058e-01 1.03409123e+00 5.60524821e-01
-3.21841687e-02 -1.42954767e+00 -1.76547050e-01 1.13859701e+00
6.07290387e-01 -1.40639436e+00 -1.77783549e-01 1.49497673e-01
-5.64440310e-01 1.73031092e+00 5.46858728e-01 -2.77031422e-01
5.59887886e-01 1.68785349e-01 2.62252927e-01 5.23502290e-01
-4.87300128e-01 3.22626859e-01 6.21381104e-01 3.65317851e-01
1.15464604e+00 5.37511706e-01 -6.01927340e-01 9.18323517e-01
-3.69765908e-01 -2.98837662e-01 3.35534543e-01 3.25939476e-01
-4.39318150e-01 -1.28905058e+00 -5.82121611e-01 6.36496961e-01
-7.43865132e-01 -1.31661862e-01 -9.46269095e-01 2.26053461e-01
-4.90121424e-01 9.68294740e-01 1.87451858e-02 -4.90047663e-01
-2.13206068e-01 8.60852242e-01 3.88283432e-01 -4.71205115e-01
-1.35986352e+00 -4.65936303e-01 2.38070413e-01 -5.58642345e-03
2.64662236e-01 -5.66939235e-01 -6.56177759e-01 -6.05292261e-01
-1.21653885e-01 2.49829948e-01 8.38798404e-01 6.77790701e-01
1.04232542e-01 2.84027696e-01 6.69430196e-01 -5.45004249e-01
-1.37177014e+00 -1.60183537e+00 -6.61093652e-01 3.85888398e-01
-1.20212242e-01 7.49684777e-03 -2.72402287e-01 2.92709563e-02] | [10.378216743469238, 10.510923385620117] |
62d8b7ce-b00f-4d80-9008-cbf610cd1c6c | disentangled-makeup-transfer-with-generative | 1907.01144 | null | https://arxiv.org/abs/1907.01144v1 | https://arxiv.org/pdf/1907.01144v1.pdf | Disentangled Makeup Transfer with Generative Adversarial Network | Facial makeup transfer is a widely-used technology that aims to transfer the makeup style from a reference face image to a non-makeup face. Existing literature leverage the adversarial loss so that the generated faces are of high quality and realistic as real ones, but are only able to produce fixed outputs. Inspired by recent advances in disentangled representation, in this paper we propose DMT (Disentangled Makeup Transfer), a unified generative adversarial network to achieve different scenarios of makeup transfer. Our model contains an identity encoder as well as a makeup encoder to disentangle the personal identity and the makeup style for arbitrary face images. Based on the outputs of the two encoders, a decoder is employed to reconstruct the original faces. We also apply a discriminator to distinguish real faces from fake ones. As a result, our model can not only transfer the makeup styles from one or more reference face images to a non-makeup face with controllable strength, but also produce various outputs with styles sampled from a prior distribution. Extensive experiments demonstrate that our model is superior to existing literature by generating high-quality results for different scenarios of makeup transfer. | ['Yaohui Jin', 'Wenqing Chen', 'Honglun Zhang', 'Hao He'] | 2019-07-02 | null | null | null | null | ['facial-makeup-transfer'] | ['computer-vision'] | [ 4.04486388e-01 4.07520801e-01 -5.98558895e-02 -4.28401381e-01
-7.04835236e-01 -9.93363976e-01 5.76327026e-01 -1.19019973e+00
3.16819400e-01 8.36321533e-01 4.93953452e-02 2.10774750e-01
5.15653491e-01 -9.87296641e-01 -9.65818822e-01 -8.82032216e-01
4.54985321e-01 2.46238485e-01 -6.31333768e-01 -2.43001774e-01
-2.47160032e-01 4.50132489e-01 -1.18030298e+00 3.85374725e-01
7.11319804e-01 8.16457391e-01 -3.36080670e-01 6.13615155e-01
3.30977827e-01 6.88292444e-01 -7.76629627e-01 -1.27872157e+00
6.83480322e-01 -8.08860004e-01 -2.44872764e-01 7.32047409e-02
6.27989352e-01 -8.47939730e-01 -7.19504952e-01 1.17240572e+00
5.49453616e-01 -5.58305860e-01 8.30315173e-01 -1.69820821e+00
-1.73498189e+00 4.37529951e-01 -6.21141434e-01 -5.35051703e-01
3.75989139e-01 3.81882995e-01 6.03615463e-01 -9.27797854e-01
4.99411464e-01 1.52549314e+00 4.11700100e-01 1.20225728e+00
-1.33181190e+00 -1.65807700e+00 -1.76512569e-01 -3.09164375e-01
-1.47157025e+00 -8.68248820e-01 8.44075024e-01 -3.43069166e-01
-1.03099950e-01 3.33126247e-01 3.56473833e-01 1.78154576e+00
2.11376101e-01 4.36933190e-01 1.42324030e+00 -1.13631263e-01
-1.28239363e-01 3.55502993e-01 -5.79576194e-01 8.17746997e-01
4.35762227e-01 5.13506293e-01 -4.95695680e-01 -1.94527134e-01
1.15609610e+00 1.22974508e-01 -4.93048608e-01 -3.96709114e-01
-1.06181610e+00 8.03698838e-01 4.97665823e-01 -2.05303490e-01
-2.18580335e-01 2.80143350e-01 -1.24896675e-01 5.17047465e-01
3.24250281e-01 4.59287584e-01 1.17509224e-01 4.32419688e-01
-6.37085617e-01 1.96552768e-01 7.72279084e-01 1.11769485e+00
8.83613050e-01 1.91883698e-01 -4.25545812e-01 6.27313554e-01
2.52889484e-01 1.05188751e+00 2.34314173e-01 -9.15851176e-01
4.81633365e-01 2.28463322e-01 3.77311796e-01 -1.19715083e+00
7.81965435e-01 3.82679813e-02 -9.85864282e-01 5.93701482e-01
1.41139969e-01 -3.89703810e-01 -9.12994444e-01 2.09161544e+00
2.56669167e-02 4.27351445e-01 8.62442702e-02 9.71325278e-01
7.27356255e-01 5.68118274e-01 -2.43628979e-01 1.64271653e-01
1.17887878e+00 -7.62630343e-01 -8.13881695e-01 -1.43014580e-01
-2.75806427e-01 -8.24652791e-01 8.31574976e-01 1.94498062e-01
-1.14282405e+00 -4.94204640e-01 -1.38496315e+00 -1.54360235e-01
-3.68450098e-02 3.58127654e-01 2.69255638e-01 9.75695193e-01
-9.71520126e-01 4.83656466e-01 -3.95608842e-01 2.76320398e-01
8.17799866e-01 4.52718079e-01 -9.22820449e-01 -1.42701760e-01
-1.28208721e+00 5.84679902e-01 -1.07175358e-01 1.99677259e-01
-1.37485957e+00 -7.89163351e-01 -7.61397779e-01 -8.89348909e-02
7.43955895e-02 -1.14707685e+00 9.67933118e-01 -1.62927175e+00
-1.88619697e+00 1.14688504e+00 8.79706368e-02 6.00297228e-02
7.64664710e-01 3.80825959e-02 -4.63828772e-01 1.04779445e-01
2.42785774e-02 5.97356439e-01 1.62856293e+00 -1.66902590e+00
7.94192217e-03 -3.69636118e-01 1.51852921e-01 9.39110219e-02
-3.35477412e-01 -2.85821185e-02 -1.35679737e-01 -9.64894176e-01
-2.73417354e-01 -1.08206367e+00 4.55172569e-01 4.98686135e-01
-9.10298645e-01 6.29201531e-01 8.78793061e-01 -7.10859656e-01
5.82152665e-01 -2.19591141e+00 4.08656955e-01 -5.03684916e-02
7.29145050e-01 2.10023969e-01 -5.06510377e-01 4.29694831e-01
-3.51537257e-01 4.48376358e-01 -8.55464861e-02 -5.06132364e-01
1.38174921e-01 3.19735229e-01 -7.40113556e-01 5.11013925e-01
4.63380307e-01 1.09240079e+00 -1.01832724e+00 -3.11256684e-02
-1.58635929e-01 7.88470447e-01 -5.48613429e-01 6.08171642e-01
1.14710130e-01 8.12338769e-01 -4.24431145e-01 3.67243439e-01
1.03934610e+00 9.80161875e-02 -4.07220162e-02 -1.28161833e-01
5.13960600e-01 -1.12340450e-01 -6.85802698e-01 1.31648123e+00
-5.80912471e-01 5.30930340e-01 2.60459870e-01 -3.59743655e-01
9.95932221e-01 5.87171674e-01 -6.20878823e-02 -2.33388230e-01
3.24855626e-01 1.40298754e-01 -3.78303342e-02 -3.20862889e-01
1.64446458e-01 -6.88549697e-01 -4.36273217e-03 5.43775320e-01
1.95416063e-01 -3.15906793e-01 -5.52665591e-01 8.82203504e-03
8.77913177e-01 8.68419707e-02 -8.81303027e-02 -4.08412237e-03
3.78239036e-01 -8.71866763e-01 6.36914134e-01 2.26936996e-01
-1.27793536e-01 9.93872225e-01 6.79088891e-01 -1.30007371e-01
-1.24523842e+00 -1.48654807e+00 2.41575554e-01 6.75525427e-01
3.36621404e-01 1.67040497e-01 -9.29307997e-01 -7.52387345e-01
3.06446642e-01 4.15973932e-01 -1.11039197e+00 -6.08709633e-01
-4.25045907e-01 -2.36784458e-01 1.01600873e+00 2.99562812e-01
7.92740107e-01 -8.41677964e-01 2.44161844e-01 -3.69325578e-01
-7.66609535e-02 -1.01094270e+00 -1.03620243e+00 -7.94521868e-01
-2.45660126e-01 -1.14207351e+00 -9.90723908e-01 -6.51004255e-01
1.00080407e+00 2.44228527e-01 9.16532874e-01 -7.49489143e-02
8.63920674e-02 1.48941100e-01 -1.40832141e-01 -4.10445333e-01
-9.10270154e-01 -3.78616512e-01 2.34826311e-01 8.81136358e-01
-7.37049384e-03 -7.48045027e-01 -7.94530571e-01 5.09352446e-01
-1.08203900e+00 3.72777432e-01 4.59770709e-01 8.50363195e-01
1.88938290e-01 -2.71147102e-01 8.46802413e-01 -1.01577163e+00
6.89060092e-01 -5.60746431e-01 -3.17572415e-01 2.60262311e-01
-3.67067516e-01 7.35546425e-02 7.73720324e-01 -7.98613966e-01
-1.03201532e+00 -3.70502502e-01 1.13735750e-01 -1.00143504e+00
1.22222818e-01 -4.72859293e-01 -8.61577213e-01 -3.31386536e-01
6.11063361e-01 1.89344123e-01 3.49676758e-01 -1.08383499e-01
8.43738317e-01 7.98444092e-01 7.25335717e-01 -6.71765924e-01
1.51799059e+00 6.61418140e-01 -2.35017866e-01 -1.54100746e-01
-6.55112028e-01 5.51368415e-01 -4.15351361e-01 -5.89910671e-02
6.76906228e-01 -1.10749161e+00 -7.33505964e-01 7.47368991e-01
-1.09731829e+00 -8.08659866e-02 -2.95848757e-01 4.27265167e-02
-5.94662607e-01 -2.61550434e-02 -6.35578275e-01 -5.11346400e-01
-2.69912720e-01 -1.18250513e+00 1.27335632e+00 2.85644472e-01
7.00608790e-02 -6.30417883e-01 -1.14852473e-01 4.21129525e-01
3.80140930e-01 7.66051531e-01 5.94144166e-01 -3.31859514e-02
-8.62728894e-01 -3.84507060e-01 -2.42838994e-01 7.33646512e-01
6.75721407e-01 1.12452850e-01 -1.11144865e+00 -4.40796554e-01
8.87867138e-02 -4.15071309e-01 5.92763126e-01 -3.44264507e-01
8.83939505e-01 -8.73902023e-01 -1.41235009e-01 1.00842798e+00
1.14965641e+00 7.25346282e-02 8.50862205e-01 -3.66828471e-01
1.01055300e+00 4.83980715e-01 -4.28834371e-02 1.60525411e-01
2.25497872e-01 5.19847155e-01 5.47657192e-01 -4.63294052e-02
-4.16148275e-01 -9.47528303e-01 6.86972499e-01 5.33123493e-01
-3.30764428e-02 -2.68241107e-01 -1.77615777e-01 3.10420722e-01
-1.34436214e+00 -1.12113404e+00 3.99139881e-01 2.22995400e+00
9.39631581e-01 -3.92700702e-01 -1.17483921e-01 -1.28510341e-01
9.29198861e-01 1.60915613e-01 -8.73692811e-01 -2.81677306e-01
-1.18334457e-01 3.35747808e-01 3.84889394e-01 2.74378598e-01
-7.79530466e-01 7.63411105e-01 5.86025238e+00 5.71429968e-01
-1.35015976e+00 3.05762291e-01 7.12961912e-01 -4.04762596e-01
-7.07888067e-01 -1.80117950e-01 -3.36456865e-01 8.69577527e-01
5.68570852e-01 -5.30218422e-01 8.65874887e-01 4.76035893e-01
-1.19515069e-01 6.96909130e-01 -1.33655465e+00 1.19291961e+00
2.75390774e-01 -1.17620361e+00 4.06170517e-01 2.56015688e-01
1.01910388e+00 -6.64992750e-01 7.77357221e-01 1.29962906e-01
5.74822724e-01 -1.57968068e+00 8.37214649e-01 6.52099371e-01
1.58963978e+00 -8.07036698e-01 4.31535602e-01 2.07448706e-01
-7.40739524e-01 8.44422802e-02 -2.13225111e-01 3.53661329e-02
-1.21053919e-01 3.45030487e-01 -5.72239280e-01 6.48558080e-01
1.36270747e-01 5.53436816e-01 -1.54877335e-01 1.79994106e-01
-8.49851549e-01 3.92471701e-01 1.27733067e-01 3.80040199e-01
-2.73704916e-01 -3.81521344e-01 5.05628228e-01 5.02145588e-01
4.45456773e-01 1.24800697e-01 -3.02380115e-01 1.45448089e+00
-1.00992787e+00 -2.36925632e-01 -1.08653080e+00 -5.64348847e-02
6.55049622e-01 1.24860394e+00 1.11748233e-01 -1.62810877e-01
-1.44276517e-02 1.61537063e+00 3.60779107e-01 5.39592624e-01
-9.93228853e-01 -3.10991943e-01 1.27057314e+00 8.85967463e-02
2.82813739e-02 1.32861167e-01 -1.87338039e-01 -1.47413564e+00
5.84614761e-02 -1.03208148e+00 -2.77702004e-01 -1.10619545e+00
-1.68174255e+00 8.06794882e-01 -3.18449467e-01 -1.23626709e+00
-1.04437895e-01 -3.86571616e-01 -7.01048017e-01 1.45756900e+00
-1.37667716e+00 -1.66199815e+00 -3.70956779e-01 5.69524109e-01
4.42890078e-02 -3.35509866e-01 9.70459163e-01 1.76887989e-01
-6.60615027e-01 1.11803949e+00 -4.99230362e-02 5.91648877e-01
8.70310962e-01 -9.46776330e-01 5.60031950e-01 7.00829804e-01
8.15234631e-02 7.91237772e-01 4.39729393e-01 -4.14423764e-01
-1.54422581e+00 -1.28753769e+00 2.95853823e-01 -7.80488253e-01
3.49434018e-01 -6.86393619e-01 -7.57988632e-01 9.52784956e-01
4.82241362e-01 2.71019101e-01 8.61573398e-01 -5.68026721e-01
-1.00635290e+00 -8.17498490e-02 -1.62680686e+00 6.98626101e-01
1.10482275e+00 -9.15295839e-01 -4.86007839e-01 -3.89978401e-02
7.10996568e-01 -5.19318879e-01 -6.77099228e-01 1.39700353e-01
8.91385317e-01 -1.10139191e+00 1.02193117e+00 -5.99611998e-01
9.33686316e-01 -2.30488643e-01 -6.06134161e-02 -1.57551193e+00
-1.84399799e-01 -1.09959972e+00 -1.29160210e-01 1.43593109e+00
4.00846787e-02 -9.66671228e-01 6.73857629e-01 5.87793708e-01
3.68007571e-01 -6.22550428e-01 -7.69149780e-01 -7.56300569e-01
4.07228172e-01 2.44270459e-01 1.23883796e+00 1.02054083e+00
-4.80745316e-01 3.86548609e-01 -9.75253820e-01 4.79723066e-01
8.52241695e-01 1.92326203e-01 9.88452196e-01 -7.39566863e-01
-3.63703549e-01 -2.30500147e-01 -4.01639313e-01 -8.99780810e-01
5.42424381e-01 -1.06394923e+00 -1.58879533e-01 -8.03267539e-01
2.76069134e-01 -3.77796978e-01 -8.05537999e-02 4.38229829e-01
-2.32846960e-01 7.92591512e-01 3.86465162e-01 2.92670667e-01
3.89813691e-01 8.79220963e-01 1.94232130e+00 -2.15590537e-01
3.74469161e-01 -5.02111763e-02 -1.32145619e+00 5.34095109e-01
6.31904960e-01 -5.46786249e-01 -6.20832443e-01 -5.83456457e-01
6.41368851e-02 2.53845185e-01 6.97099447e-01 -6.48227453e-01
-2.92545766e-01 -1.23419121e-01 4.43373799e-01 3.35089177e-01
5.05329609e-01 -6.39528155e-01 5.43795168e-01 1.04578562e-01
-2.26110399e-01 -2.04088077e-01 -1.40576273e-01 5.32685995e-01
-3.18114385e-02 6.89369664e-02 9.72768426e-01 9.74612907e-02
-7.21662045e-02 7.72832215e-01 3.17509770e-01 -1.18848279e-01
1.25363815e+00 -1.53493404e-01 -5.03456771e-01 -6.67925179e-01
-7.17002213e-01 -1.29384741e-01 8.29946220e-01 7.13500619e-01
7.37095594e-01 -1.80160379e+00 -1.06649137e+00 7.55241036e-01
4.86474447e-02 -2.50283778e-01 1.70573443e-01 1.43307179e-01
-4.23469156e-01 -1.68144539e-01 -5.66075742e-01 -4.70905565e-02
-1.14203107e+00 7.56538928e-01 6.28384411e-01 2.67217994e-01
-3.36175382e-01 8.71214509e-01 7.43322253e-01 -5.01326561e-01
-3.92061383e-01 2.36303657e-01 2.52960950e-01 -3.23357224e-01
6.59682572e-01 5.23519143e-02 -3.72783124e-01 -9.60853994e-01
1.70645993e-02 6.09230399e-01 2.04572752e-01 -1.87754348e-01
9.73498642e-01 -4.23886068e-02 -1.29052356e-01 1.04971223e-01
1.56031728e+00 4.21634734e-01 -1.68337941e+00 1.10002737e-02
-1.00019336e+00 -1.03957403e+00 -3.20882231e-01 -7.01561511e-01
-1.45021296e+00 8.45275879e-01 3.63006502e-01 -2.11755233e-03
1.18499207e+00 -2.05642894e-01 8.91661108e-01 -2.89191246e-01
7.47727394e-01 -9.54069793e-02 2.65983373e-01 -1.66983560e-01
1.31849217e+00 -1.07246089e+00 -4.62077081e-01 -6.97040319e-01
-7.36198783e-01 7.83968806e-01 5.48428953e-01 -5.33741593e-01
5.43710053e-01 3.14990819e-01 2.06535250e-01 1.54288203e-01
-5.69359183e-01 2.48342425e-01 2.89873987e-01 9.02936935e-01
4.10252549e-02 3.44299912e-01 2.84869552e-01 7.07644165e-01
-6.03709519e-01 1.82181165e-01 6.10939860e-01 3.72861981e-01
2.72818238e-01 -1.38286293e+00 -4.89740342e-01 2.57276118e-01
-4.02603239e-01 1.89798325e-01 -5.58487296e-01 5.22014499e-01
4.18954581e-01 8.59493852e-01 -5.09744398e-02 -7.08909214e-01
2.98395425e-01 -1.03675567e-01 8.43795002e-01 -6.86860979e-01
-4.59514529e-01 -5.92006981e-01 -2.69038975e-01 -5.12775481e-01
-7.80080929e-02 -2.70208210e-01 -7.06699491e-01 -8.03373992e-01
-5.82995266e-02 -6.74941912e-02 2.87040025e-01 5.16239703e-01
6.19078696e-01 3.35785925e-01 1.35536981e+00 -8.93754423e-01
-7.77028441e-01 -7.06960917e-01 -7.19805419e-01 7.97625840e-01
6.00045502e-01 -6.65242136e-01 -3.00354987e-01 3.38506758e-01] | [12.693376541137695, 0.0876995176076889] |
c75a71d2-cd7f-4a6f-9485-5d7710b4d3ee | improving-sequential-recommendation | 2106.14031 | null | https://arxiv.org/abs/2106.14031v2 | https://arxiv.org/pdf/2106.14031v2.pdf | Improving Sequential Recommendation Consistency with Self-Supervised Imitation | Most sequential recommendation models capture the features of consecutive items in a user-item interaction history. Though effective, their representation expressiveness is still hindered by the sparse learning signals. As a result, the sequential recommender is prone to make inconsistent predictions. In this paper, we propose a model, SSI, to improve sequential recommendation consistency with Self-Supervised Imitation. Precisely, we extract the consistency knowledge by utilizing three self-supervised pre-training tasks, where temporal consistency and persona consistency capture user-interaction dynamics in terms of the chronological order and persona sensitivities, respectively. Furthermore, to provide the model with a global perspective, global session consistency is introduced by maximizing the mutual information among global and local interaction sequences. Finally, to comprehensively take advantage of all three independent aspects of consistency-enhanced knowledge, we establish an integrated imitation learning framework. The consistency knowledge is effectively internalized and transferred to the student model by imitating the conventional prediction logit as well as the consistency-enhanced item representations. In addition, the flexible self-supervised imitation framework can also benefit other student recommenders. Experiments on four real-world datasets show that SSI effectively outperforms the state-of-the-art sequential recommendation methods. | ['Bo Long', 'Zhen He', 'Zhuoye Ding', 'Xiaofang Zhao', 'Yonghao Song', 'Hongshen Chen', 'Xu Yuan'] | 2021-06-26 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [-2.46031344e-01 -3.29637587e-01 -5.93589246e-01 -5.38059711e-01
-1.85181662e-01 -5.12747586e-01 5.27640760e-01 -2.99899966e-01
8.25579762e-02 5.42139530e-01 4.10568953e-01 1.84296623e-01
-7.32602358e-01 -5.22584140e-01 -6.82566047e-01 -5.03773689e-01
-3.33790593e-02 2.55674154e-01 -6.12934912e-03 -3.66246551e-01
1.07297406e-01 -1.20070711e-01 -1.60093188e+00 2.33179480e-01
1.33752072e+00 9.44142640e-01 3.62325549e-01 2.32003722e-02
-4.88795945e-03 7.99682736e-01 -6.57875910e-02 -3.74834090e-01
8.90702605e-02 -7.50456274e-01 -4.47971553e-01 9.38435271e-02
9.09128413e-02 -5.27458489e-01 -4.94867504e-01 8.54376376e-01
1.61304295e-01 6.49486959e-01 5.53885102e-01 -1.20651805e+00
-1.04406917e+00 1.00037193e+00 -1.61227435e-01 -1.73748154e-02
6.77974522e-01 5.36414869e-02 1.18860638e+00 -8.80769193e-01
5.27071178e-01 1.04576993e+00 7.18495965e-01 3.13683629e-01
-9.88820672e-01 -7.78052032e-01 9.06139135e-01 2.92129546e-01
-1.00535727e+00 -9.91919264e-02 8.36932540e-01 -3.29843581e-01
5.13561845e-01 2.28743225e-01 8.74106944e-01 1.49006808e+00
-8.25675949e-02 1.21494770e+00 8.80945444e-01 1.40555516e-01
-4.26850431e-02 2.82758117e-01 5.20508111e-01 5.54706216e-01
1.37539208e-01 4.34051692e-01 -6.99909151e-01 3.08238808e-02
8.99153948e-01 8.79305720e-01 -2.35085607e-01 -4.51559007e-01
-1.29364312e+00 6.71662986e-01 3.61613214e-01 4.57434386e-01
-4.70213652e-01 -2.58381069e-01 2.88183719e-01 7.11668670e-01
2.63711721e-01 3.72416109e-01 -4.25441772e-01 -2.43493184e-01
-7.63791919e-01 1.47926018e-01 8.10263813e-01 1.29690409e+00
5.81589580e-01 8.81358236e-02 -4.43769932e-01 9.72821414e-01
6.03112102e-01 4.08665597e-01 9.44351017e-01 -9.35000777e-01
2.11905554e-01 7.25362480e-01 3.08888793e-01 -1.05409896e+00
-3.28802854e-01 -8.92668784e-01 -1.00630105e+00 -5.95361054e-01
3.68650407e-01 1.76956188e-02 -1.77318498e-01 1.95913291e+00
3.55805188e-01 7.04925895e-01 -6.88091293e-02 9.32585180e-01
7.80857742e-01 5.15470982e-01 -2.32511476e-01 -7.30973125e-01
9.72533166e-01 -1.38247597e+00 -8.01610470e-01 2.44862288e-01
4.47498709e-01 -5.55582404e-01 1.01693296e+00 4.32252795e-01
-1.02768576e+00 -1.05197370e+00 -9.55436826e-01 3.55921298e-01
2.41312180e-02 3.09342146e-01 6.39624178e-01 2.34940395e-01
-4.64036107e-01 9.25147414e-01 -6.70866907e-01 -2.97417790e-01
-3.89202721e-02 3.96679878e-01 1.47686498e-02 1.34597823e-01
-1.42983973e+00 4.62650925e-01 1.85439247e-03 1.66311786e-01
-6.07979178e-01 -6.92995906e-01 -6.48357570e-01 1.18353859e-01
6.24878764e-01 -7.44565308e-01 1.39516366e+00 -1.18230569e+00
-2.14155793e+00 -1.73840728e-02 -7.38732144e-02 -1.28366023e-01
4.75130439e-01 -3.28970373e-01 -7.67503202e-01 -3.87123704e-01
-4.40558568e-02 -8.87344033e-02 8.72011602e-01 -1.06050766e+00
-8.22891712e-01 -2.34389409e-01 4.96383235e-02 4.70423430e-01
-6.87135696e-01 -2.50648856e-01 -6.68544948e-01 -8.44945788e-01
3.36386450e-03 -1.19833291e+00 -2.48209275e-02 -3.80617768e-01
-1.00352190e-01 -5.39431095e-01 3.83793801e-01 -5.61440527e-01
1.63282728e+00 -2.21445513e+00 5.21771073e-01 5.28979301e-01
6.10198490e-02 3.65307480e-02 -3.95932764e-01 4.87784237e-01
2.89723337e-01 -3.37009758e-01 2.99705297e-01 -3.83211672e-01
1.71705052e-01 2.99314052e-01 -4.07286584e-01 2.14241877e-01
-5.17975152e-01 9.84778345e-01 -1.13136899e+00 -1.54006302e-01
8.07561651e-02 2.12525412e-01 -7.77748883e-01 5.91711342e-01
-1.15077436e-01 8.11909497e-01 -7.61251748e-01 3.64624232e-01
3.49482119e-01 -5.84197640e-01 5.93234837e-01 -2.24800184e-01
-6.02872483e-03 4.16494370e-01 -1.46629477e+00 1.90488112e+00
-4.60545212e-01 -2.49768794e-01 -2.70786047e-01 -9.71655309e-01
8.61973643e-01 2.68500298e-01 7.34945714e-01 -8.67885649e-01
-2.94960797e-01 8.18528980e-02 -8.91952515e-02 -3.44604373e-01
6.73713326e-01 3.81875366e-01 -8.87931064e-02 8.63731563e-01
2.15935215e-01 5.29331684e-01 -7.00615123e-02 3.73948574e-01
4.58187193e-01 4.98203784e-01 2.96957761e-01 -1.16931677e-01
5.53237081e-01 -6.07351542e-01 8.33615899e-01 9.94725883e-01
-5.44838384e-02 2.35565305e-01 6.83924705e-02 -2.13394955e-01
-5.52323401e-01 -8.78434360e-01 1.69117510e-01 1.38530266e+00
6.36283040e-01 -6.65550470e-01 -2.76319832e-01 -8.59637022e-01
2.57083416e-01 5.33558249e-01 -6.32508874e-01 -4.99326140e-01
-4.89869416e-01 -1.98431283e-01 -6.67217001e-02 6.49545193e-01
3.19371641e-01 -1.02293205e+00 3.29846919e-01 5.07031024e-01
-2.40020037e-01 -6.53885305e-01 -1.16460729e+00 -3.03265870e-01
-9.91898894e-01 -7.74362206e-01 -5.42631447e-01 -6.91349983e-01
4.21545535e-01 8.27878237e-01 9.23485458e-01 1.93421498e-01
5.01911104e-01 6.61633134e-01 -6.89860582e-01 1.08786844e-01
-1.04631633e-01 3.52926217e-02 6.46940887e-01 2.44457036e-01
2.01963991e-01 -8.43558490e-01 -9.41029847e-01 8.97520363e-01
-4.64010388e-01 4.87298667e-02 5.33099294e-01 1.18845725e+00
5.61343968e-01 -1.83713645e-01 9.46859479e-01 -9.61737335e-01
7.04042137e-01 -8.90113711e-01 -1.64767042e-01 4.86661792e-01
-1.23862791e+00 -1.40684396e-01 9.89992380e-01 -9.79351759e-01
-1.43050361e+00 -2.78051108e-01 2.49072224e-01 -4.11974281e-01
8.63009319e-02 8.01031291e-01 -1.41784549e-02 3.17029834e-01
3.30825061e-01 6.41827762e-01 1.79623395e-01 -6.84898913e-01
5.25483549e-01 6.68196440e-01 3.22490573e-01 -6.96502566e-01
6.18239939e-01 9.85537693e-02 -6.43255293e-01 -3.30308080e-01
-1.09105027e+00 -7.49369562e-01 -4.23725218e-01 -2.91131824e-01
7.22424537e-02 -9.78806734e-01 -9.54618335e-01 5.40991008e-01
-6.36495650e-01 -3.38481516e-01 -2.73376584e-01 7.90016592e-01
-5.40893793e-01 7.17884183e-01 -9.15685177e-01 -6.54042363e-01
-2.55188137e-01 -8.83514702e-01 5.77026665e-01 4.63899344e-01
-2.51360029e-01 -9.56483424e-01 1.89586148e-01 2.87121207e-01
3.88076305e-01 -6.31695747e-01 3.96548152e-01 -9.05514479e-01
-4.39223379e-01 -8.59599859e-02 4.85232696e-02 2.57533789e-01
3.98599714e-01 -1.25424817e-01 -3.28146338e-01 -4.85001117e-01
-9.87276621e-03 -2.64323473e-01 4.63434547e-01 2.03493476e-01
1.12354124e+00 -5.39296746e-01 -2.33910516e-01 4.04834509e-01
8.21886539e-01 1.90913349e-01 1.64487228e-01 1.05256855e-01
9.19788480e-01 4.82239902e-01 1.03395188e+00 7.34949827e-01
7.56791115e-01 9.41286206e-01 -1.68674752e-01 4.46558446e-01
9.99456868e-02 -9.44314480e-01 5.60574830e-01 1.45285881e+00
-3.27516913e-01 -2.17544157e-02 1.48612112e-02 6.83392063e-02
-2.58127236e+00 -1.35708725e+00 4.61173207e-02 2.23740840e+00
7.69842088e-01 -2.11911306e-01 5.49976349e-01 -3.35455179e-01
4.59217638e-01 -4.12612408e-02 -8.32223117e-01 9.40144211e-02
2.04085961e-01 -5.56882918e-01 -5.55188656e-02 2.20501646e-01
-7.46517181e-01 7.15309560e-01 5.80178595e+00 8.49754393e-01
-7.43632674e-01 1.09884381e-01 1.08008631e-01 -2.03975379e-01
-5.51845312e-01 -2.14649335e-01 -7.18010187e-01 1.08338892e+00
7.46310532e-01 -1.82379797e-01 8.80813479e-01 8.60532641e-01
3.40446293e-01 3.09680969e-01 -1.33885920e+00 9.36717153e-01
8.17229971e-02 -1.02701175e+00 8.51642638e-02 3.72922271e-02
1.06840432e+00 -4.47125673e-01 2.38244832e-01 8.72083008e-01
5.23920536e-01 -6.13762498e-01 5.96356928e-01 1.08358932e+00
3.74710828e-01 -6.13900125e-01 5.98898530e-01 6.99682236e-01
-1.39457917e+00 -3.74633759e-01 -2.91785717e-01 -1.62706509e-01
1.59350246e-01 3.22492123e-01 -2.01509427e-03 1.01472938e+00
6.24906778e-01 1.61519825e+00 -2.41899118e-01 7.69021094e-01
-2.14503184e-01 7.52583086e-01 -1.55029476e-01 -1.99880496e-01
7.79927224e-02 -7.80500591e-01 2.81730056e-01 8.66304815e-01
4.73704636e-01 3.20680916e-01 4.16933984e-01 6.69316530e-01
-1.28113264e-02 2.12614492e-01 -4.12932664e-01 7.40112290e-02
8.04118156e-01 1.11579442e+00 -9.52201262e-02 -4.31289792e-01
-7.06182897e-01 1.23957646e+00 5.30927718e-01 4.93994683e-01
-7.57433474e-01 1.48352146e-01 6.49668097e-01 -2.99812347e-01
4.73469108e-01 -1.43152684e-01 -8.08778554e-02 -1.60404086e+00
-6.07943758e-02 -1.17444646e+00 4.62579370e-01 -3.51468265e-01
-1.85066676e+00 4.64494936e-02 -8.28733370e-02 -1.72098255e+00
-3.37556809e-01 -1.19514138e-01 -8.33532512e-01 5.30914485e-01
-1.16468191e+00 -1.20913279e+00 -2.87302285e-01 7.89713860e-01
5.93491793e-01 -4.84881252e-01 6.47567689e-01 3.49931896e-01
-9.53609884e-01 1.17031729e+00 7.00168550e-01 -3.00995946e-01
9.46451724e-01 -1.00980330e+00 -5.61944395e-02 4.34012085e-01
3.54027152e-01 1.36889720e+00 4.09074992e-01 -6.79914713e-01
-1.72711980e+00 -8.87831092e-01 5.57405293e-01 -4.58671987e-01
7.99227655e-01 2.02366300e-02 -1.07756352e+00 8.57472062e-01
-1.97810858e-01 -3.01056266e-01 9.69264209e-01 7.24521697e-01
-4.22291815e-01 -3.25345516e-01 -6.37470245e-01 6.15198731e-01
1.56271029e+00 -5.71623981e-01 -7.05152512e-01 3.06193650e-01
6.74807847e-01 -2.79200375e-01 -1.13944924e+00 2.11997554e-01
9.79288340e-01 -7.82078862e-01 1.02767301e+00 -6.60858035e-01
3.29914570e-01 -2.12659270e-01 1.65760815e-01 -1.39974117e+00
-9.46013212e-01 -7.55737960e-01 -7.70119846e-01 1.35266793e+00
1.86017394e-01 -5.75006604e-01 4.67865765e-01 5.94166696e-01
-9.08823311e-02 -7.47389078e-01 -5.06842434e-01 -1.03491366e+00
-2.75796473e-01 -1.54119357e-01 8.48945975e-01 1.14321673e+00
6.00616455e-01 4.21487480e-01 -1.20121372e+00 -1.20735295e-01
4.56635594e-01 7.76797831e-01 9.02027667e-01 -1.38865089e+00
-8.54503632e-01 -4.31685537e-01 1.87354922e-01 -1.74188614e+00
2.11315930e-01 -8.92013371e-01 -1.59432843e-01 -1.20829403e+00
4.38503236e-01 -4.08116370e-01 -9.10131752e-01 9.09346938e-02
-4.24311936e-01 -1.99717298e-01 6.19558617e-03 8.49845588e-01
-1.26168334e+00 8.26695919e-01 1.46245730e+00 1.77931458e-01
-6.36105299e-01 4.23753977e-01 -8.20887923e-01 5.54910541e-01
6.00125194e-01 -1.78210735e-01 -7.76827097e-01 -4.10453081e-02
1.77380145e-01 2.53259391e-01 1.39978668e-03 -4.91085589e-01
5.10115862e-01 -4.27909315e-01 2.31112823e-01 -3.45888853e-01
1.19076639e-01 -8.29850078e-01 3.29679638e-01 2.97006786e-01
-6.51333809e-01 -1.44328997e-01 -4.16495055e-01 1.27527130e+00
-7.07485378e-02 2.56780349e-02 2.90595442e-01 7.95672685e-02
-6.52812839e-01 6.27600908e-01 -2.54010707e-01 -2.95513213e-01
8.02937329e-01 -1.75876111e-01 3.95317562e-02 -5.43152153e-01
-7.07628310e-01 5.82190752e-01 3.11947018e-01 7.27620304e-01
4.33217734e-01 -1.70224917e+00 -3.22340727e-01 2.99791068e-01
2.44095251e-01 -4.92417872e-01 8.95132065e-01 1.05251873e+00
5.15787303e-01 2.25734591e-01 -1.53547302e-01 -4.52008575e-01
-9.99929190e-01 8.21899056e-01 2.85521317e-02 -4.12675977e-01
-6.86022043e-01 5.31510413e-01 4.21195440e-02 -6.43902957e-01
4.74022925e-01 -2.65229434e-01 -5.71477771e-01 8.42501745e-02
4.82700258e-01 5.55305898e-01 -5.29203713e-01 -3.28748196e-01
-1.21400021e-01 4.56778944e-01 -3.83372188e-01 1.19433895e-01
1.18378031e+00 -5.56996703e-01 2.21106410e-01 6.62799358e-01
8.92289877e-01 4.01529633e-02 -1.39776385e+00 -8.83699417e-01
-1.33331880e-01 -4.04479414e-01 -2.78628170e-01 -8.27075899e-01
-8.98838520e-01 2.97921330e-01 2.86335289e-01 2.02869177e-01
6.76882267e-01 -1.69210300e-01 9.63732481e-01 7.04427123e-01
5.13530433e-01 -1.32102585e+00 3.89848113e-01 4.78481621e-01
7.28815913e-01 -1.44663966e+00 -4.76274826e-02 -1.61612913e-01
-1.01034236e+00 8.13840091e-01 8.91009867e-01 -1.91421732e-01
8.81917298e-01 -3.62847865e-01 -3.11177135e-01 1.21557906e-01
-7.56718934e-01 9.11152363e-02 7.63727248e-01 2.84694523e-01
5.46033561e-01 2.81534314e-01 -6.35890484e-01 1.56841135e+00
7.61151463e-02 2.82370329e-01 7.82512203e-02 4.45835531e-01
-1.44543588e-01 -1.13441873e+00 2.64541388e-01 6.50620937e-01
1.52990267e-01 2.35631555e-01 -1.30863622e-01 3.49261284e-01
-1.67570323e-01 1.00390220e+00 -2.44203165e-01 -8.14368367e-01
4.68549490e-01 -1.44469678e-01 2.47486472e-01 -5.35185874e-01
-7.55272627e-01 8.47709700e-02 -5.90583198e-02 -8.13970804e-01
-3.62400770e-01 -7.07349598e-01 -1.06577563e+00 -4.10478473e-01
-4.29813594e-01 3.44346225e-01 1.77256837e-01 1.12554991e+00
6.26169086e-01 4.69525397e-01 1.11259651e+00 -7.83104539e-01
-1.15218878e+00 -1.02013743e+00 -8.98795664e-01 6.98541939e-01
7.10342964e-03 -8.32631707e-01 -3.25985849e-01 -2.97589660e-01] | [10.105229377746582, 5.584603309631348] |
afa1b163-7fb2-4640-9d35-aa1aabe5d45a | 3d-human-pose-estimation-in-video-with | 1811.11742 | null | http://arxiv.org/abs/1811.11742v2 | http://arxiv.org/pdf/1811.11742v2.pdf | 3D human pose estimation in video with temporal convolutions and semi-supervised training | In this work, we demonstrate that 3D poses in video can be effectively
estimated with a fully convolutional model based on dilated temporal
convolutions over 2D keypoints. We also introduce back-projection, a simple and
effective semi-supervised training method that leverages unlabeled video data.
We start with predicted 2D keypoints for unlabeled video, then estimate 3D
poses and finally back-project to the input 2D keypoints. In the supervised
setting, our fully-convolutional model outperforms the previous best result
from the literature by 6 mm mean per-joint position error on Human3.6M,
corresponding to an error reduction of 11%, and the model also shows
significant improvements on HumanEva-I. Moreover, experiments with
back-projection show that it comfortably outperforms previous state-of-the-art
results in semi-supervised settings where labeled data is scarce. Code and
models are available at https://github.com/facebookresearch/VideoPose3D | ['Michael Auli', 'David Grangier', 'Dario Pavllo', 'Christoph Feichtenhofer'] | 2018-11-28 | 3d-human-pose-estimation-in-video-with-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Pavllo_3D_Human_Pose_Estimation_in_Video_With_Temporal_Convolutions_and_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Pavllo_3D_Human_Pose_Estimation_in_Video_With_Temporal_Convolutions_and_CVPR_2019_paper.pdf | cvpr-2019-6 | ['monocular-3d-human-pose-estimation', 'weakly-supervised-3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-3.26649040e-01 2.00264245e-01 -2.93885022e-01 -4.11318868e-01
-8.95853162e-01 -7.19462633e-01 4.52322662e-01 -5.51048458e-01
-6.50014222e-01 3.60715330e-01 3.83534431e-01 -9.95673425e-03
4.01584715e-01 -3.36465128e-02 -1.29592776e+00 -3.15876991e-01
-2.90046483e-01 5.02599895e-01 1.86323851e-01 1.80540353e-01
-8.89281780e-02 4.50130403e-01 -1.18361366e+00 7.03124143e-03
2.39031166e-01 9.55105245e-01 5.47918715e-02 8.22414339e-01
5.04904687e-01 6.23942137e-01 -5.35275042e-02 -3.04831237e-01
6.84583604e-01 -8.90011154e-03 -8.62046123e-01 3.70204955e-01
1.14004517e+00 -9.04581249e-01 -8.33073854e-01 8.08606565e-01
5.47595561e-01 4.62271273e-02 4.34276074e-01 -1.12541652e+00
-2.82735437e-01 5.84506057e-02 -6.11401677e-01 1.14211805e-01
5.47530770e-01 4.97753657e-02 5.97952306e-01 -1.29832971e+00
9.07565415e-01 1.09367299e+00 1.06456983e+00 6.26391411e-01
-9.38624144e-01 -7.33383417e-01 5.00001423e-02 9.24453959e-02
-1.51153517e+00 -3.74364674e-01 4.84650403e-01 -5.17885506e-01
1.01565790e+00 -1.56182647e-01 7.65960932e-01 1.28558981e+00
4.23839167e-02 1.07782304e+00 7.38536119e-01 -2.60775954e-01
-1.82545751e-01 -2.14462548e-01 -2.00869814e-01 1.02016616e+00
-7.07377791e-02 3.27914596e-01 -7.11418390e-01 1.80789202e-01
1.09651339e+00 2.72732079e-01 -2.51729488e-01 -8.60146880e-01
-1.50123239e+00 4.90858465e-01 5.71551621e-01 -2.84521967e-01
-3.29288602e-01 5.12433112e-01 2.23836422e-01 8.40640217e-02
6.66365921e-01 2.14290153e-02 -7.22194433e-01 -4.11767155e-01
-9.60352302e-01 3.70496690e-01 4.39235330e-01 1.31407809e+00
6.09660864e-01 -2.12147996e-01 1.87653810e-01 3.27334732e-01
4.17775005e-01 8.04968715e-01 1.50686488e-01 -1.54773521e+00
4.76708293e-01 3.03301960e-01 4.30683017e-01 -5.85174859e-01
-5.20174205e-01 -2.50802785e-01 -2.88951635e-01 3.30136895e-01
4.82433438e-01 -4.45486426e-01 -1.11532712e+00 1.48980916e+00
5.50484955e-01 3.81379157e-01 -1.98609665e-01 1.18600368e+00
9.32768762e-01 3.75519395e-01 -2.69780129e-01 1.88042164e-01
9.20835078e-01 -1.41683090e+00 -4.65143681e-01 -2.43083254e-01
6.60403490e-01 -6.79995716e-01 6.43030286e-01 3.47245753e-01
-1.06900477e+00 -6.89551234e-01 -9.04706419e-01 -1.57302320e-01
-4.94261310e-02 4.76846367e-01 5.90699852e-01 2.48356909e-01
-1.24151993e+00 6.85171545e-01 -1.20606089e+00 -6.43286347e-01
4.65556651e-01 3.61584872e-01 -9.15525436e-01 -1.18484989e-01
-7.44118929e-01 8.24421585e-01 1.59198880e-01 1.71902627e-01
-1.12277997e+00 -8.46987367e-01 -1.01648068e+00 -4.27351534e-01
6.86018527e-01 -6.61436021e-01 1.61960995e+00 -5.45694292e-01
-1.41898453e+00 1.05669665e+00 -1.03484213e-01 -5.64969182e-01
9.03303623e-01 -1.10460305e+00 1.32179521e-02 5.58945894e-01
1.27718553e-01 1.29415941e+00 7.53719568e-01 -1.05670619e+00
-4.65676904e-01 -3.12893718e-01 1.32260054e-01 5.23597836e-01
-7.53775612e-02 -1.03611544e-01 -1.17338443e+00 -5.01026273e-01
1.16888918e-01 -1.45008087e+00 -2.69685179e-01 4.17258620e-01
-4.47452843e-01 6.34109899e-02 7.87170887e-01 -7.15803862e-01
6.10539079e-01 -1.96360803e+00 2.28186384e-01 -1.12634763e-01
2.81164259e-01 2.73131222e-01 4.98701073e-02 1.78255141e-01
-2.45886967e-01 -2.15520903e-01 1.33587003e-01 -8.22732806e-01
-1.56508490e-01 -6.08549267e-02 -5.29636927e-02 8.03598881e-01
-1.92648619e-02 1.27498734e+00 -8.77281845e-01 -3.68332237e-01
6.46446407e-01 7.46855021e-01 -5.88224173e-01 2.19726369e-01
-2.43967995e-02 7.04494476e-01 -2.49645904e-01 6.82153344e-01
5.58694243e-01 -4.96619195e-01 -1.82330143e-02 -3.01834285e-01
-3.53036337e-02 2.99744438e-02 -1.03075075e+00 2.31801009e+00
-5.04953563e-02 8.15613031e-01 -3.69357206e-02 -5.42922854e-01
4.46301848e-01 3.70506853e-01 7.12637305e-01 -1.53277263e-01
2.33982936e-01 6.45719394e-02 -5.59495747e-01 -4.04054523e-01
3.92377436e-01 2.55453169e-01 1.60061002e-01 2.76049703e-01
4.64908510e-01 -5.35094701e-02 -1.15241148e-01 3.24030876e-01
9.71185803e-01 9.74537313e-01 4.17620726e-02 -1.90481041e-02
3.29822004e-02 -4.87938942e-03 3.28232348e-01 4.22269523e-01
-3.21915925e-01 1.07260692e+00 1.69202745e-01 -6.26471162e-01
-1.19195437e+00 -1.07402074e+00 1.57216758e-01 6.76976442e-01
1.82123825e-01 -5.22791326e-01 -8.65721583e-01 -9.69769537e-01
1.26398593e-01 8.91300067e-02 -6.91318750e-01 2.28084967e-01
-5.82799792e-01 -1.52941672e-02 5.39852023e-01 1.02859426e+00
5.63672245e-01 -6.41281366e-01 -6.11625433e-01 -1.50811985e-01
-2.86932290e-01 -1.56344044e+00 -6.70761704e-01 1.39986545e-01
-9.23229337e-01 -1.25641465e+00 -1.21483505e+00 -7.34540403e-01
8.40716302e-01 4.75983411e-01 9.44140315e-01 -1.45778045e-01
-1.10840753e-01 6.87201023e-01 -3.87908280e-01 -2.14860648e-01
8.97156224e-02 -9.19018239e-02 5.08546591e-01 -3.15242052e-01
4.65103984e-01 -3.96244705e-01 -8.28099370e-01 4.46490616e-01
-4.26486880e-01 2.18838364e-01 4.95867878e-01 5.65806806e-01
7.44300902e-01 -5.41550398e-01 -1.54028565e-01 -5.76675415e-01
-3.01579386e-01 -1.71999276e-01 -6.08461738e-01 -1.51096210e-02
-3.46276730e-01 -2.54904479e-01 7.09039643e-02 -4.07012939e-01
-9.15678740e-01 6.14902258e-01 -9.12301764e-02 -1.28448987e+00
-4.25796747e-01 1.16902553e-01 2.08318755e-01 -2.77801275e-01
7.23916411e-01 -2.67978804e-03 2.28736833e-01 -4.68602180e-01
5.24996340e-01 3.40358824e-01 7.49159932e-01 -3.32943827e-01
9.50816989e-01 8.71815681e-01 1.72935743e-02 -5.42125463e-01
-1.09185851e+00 -7.10832000e-01 -1.18722749e+00 -4.42935079e-01
1.05384803e+00 -1.55338633e+00 -6.57478571e-01 5.45968473e-01
-1.05525744e+00 -7.62572408e-01 -5.17909639e-02 9.60723400e-01
-9.64650154e-01 5.03773689e-01 -7.82551050e-01 -4.13014054e-01
-2.43601292e-01 -1.07286334e+00 1.44540632e+00 3.51606897e-04
-2.98563272e-01 -9.06228423e-01 4.01769616e-02 4.85944003e-01
-1.00665256e-01 2.72529691e-01 -8.28386396e-02 -5.30154109e-01
-5.68432093e-01 -4.22910541e-01 -2.61940155e-02 4.73757982e-01
-1.25905201e-01 -1.87187076e-01 -9.03893828e-01 -4.41149354e-01
-3.47452462e-01 -6.68573499e-01 7.45972157e-01 6.89726114e-01
8.76807213e-01 1.00580186e-01 -5.07664680e-01 8.95954370e-01
8.52117598e-01 -3.81485909e-01 3.50596875e-01 2.73174375e-01
9.86874104e-01 4.06313688e-01 9.02615726e-01 2.69692838e-01
4.72355098e-01 7.24087179e-01 4.04835522e-01 -1.87264889e-01
-2.07174554e-01 -5.97731531e-01 2.89574742e-01 5.09150624e-01
-4.26353335e-01 1.05198532e-01 -8.77520025e-01 3.95893186e-01
-2.04925942e+00 -7.42381811e-01 6.86542094e-02 2.11778760e+00
4.98661339e-01 2.59697467e-01 2.89436579e-01 -2.76632488e-01
5.49807668e-01 7.99293146e-02 -6.20722175e-01 4.24638391e-01
1.83553547e-01 -2.43429430e-02 8.37521374e-01 4.84191477e-01
-1.51952612e+00 1.08796978e+00 7.08149719e+00 3.91940713e-01
-9.77851987e-01 8.78348053e-02 3.96289229e-01 -6.29090905e-01
3.14288408e-01 -1.56328842e-01 -8.56352329e-01 1.93573192e-01
7.23333240e-01 2.57436126e-01 1.57295108e-01 1.11484659e+00
2.35980630e-01 4.53633554e-02 -1.35659921e+00 1.23901069e+00
2.40848973e-01 -1.36562276e+00 -3.37602139e-01 2.53194898e-01
9.56109047e-01 5.99380672e-01 9.65285599e-02 1.70764759e-01
1.05891153e-01 -8.04317355e-01 9.05362308e-01 4.08266425e-01
1.03484213e+00 -6.71484351e-01 6.23784781e-01 3.72103989e-01
-1.12042987e+00 2.05247760e-01 -2.73799360e-01 -1.64765880e-01
3.65912050e-01 1.70866132e-01 -7.81413376e-01 3.21630597e-01
1.22573519e+00 1.17124200e+00 -4.89443362e-01 1.00134051e+00
-5.05196929e-01 3.88573200e-01 -4.98901278e-01 2.96596497e-01
3.49840313e-01 7.02181086e-02 4.81848747e-01 1.17241204e+00
3.17780703e-01 1.14118889e-01 2.95797527e-01 5.03605008e-01
-2.52115101e-01 -3.07778656e-01 -6.82184339e-01 1.48157686e-01
2.90334910e-01 1.29314578e+00 -6.10630512e-01 -4.11673397e-01
-4.74224567e-01 1.34986329e+00 2.93929964e-01 5.07556856e-01
-9.45636988e-01 -2.57475227e-02 6.44312084e-01 1.21016964e-01
5.84727943e-01 -6.73421085e-01 2.57544249e-01 -1.31561911e+00
2.03411460e-01 -5.50242245e-01 1.09412797e-01 -1.16140568e+00
-9.14964497e-01 3.61416042e-01 3.01173091e-01 -1.60846186e+00
-5.44515371e-01 -9.22860801e-01 -3.23543042e-01 5.13463557e-01
-1.17238808e+00 -1.33937371e+00 -4.04968917e-01 6.80278778e-01
6.27984345e-01 -3.76505703e-02 7.20834613e-01 1.33016050e-01
-2.11487919e-01 5.62087119e-01 5.32159284e-02 4.84930873e-01
9.39824343e-01 -1.21265888e+00 9.24150467e-01 8.35990489e-01
3.66910875e-01 6.29136622e-01 5.51024020e-01 -6.73341691e-01
-1.50969696e+00 -1.00746858e+00 5.56638777e-01 -1.04613054e+00
4.99037027e-01 -4.67726737e-01 -3.93015742e-01 1.26253092e+00
6.42855912e-02 5.97922683e-01 6.02437377e-01 1.18469268e-01
-4.34764266e-01 3.48077506e-01 -9.52239811e-01 6.32528365e-01
1.47664893e+00 -5.76896310e-01 -5.21893322e-01 6.60960615e-01
9.09225404e-01 -1.06473017e+00 -8.11654449e-01 4.45202112e-01
8.09862554e-01 -9.97692585e-01 1.11043417e+00 -4.96791095e-01
3.76538336e-01 -3.07322323e-01 -1.84422940e-01 -1.17472553e+00
-2.65747666e-01 -7.32305288e-01 -5.63308895e-01 4.14885938e-01
2.52830476e-01 -7.71685168e-02 1.32168210e+00 5.29830158e-01
-6.88060448e-02 -9.61506307e-01 -7.98342943e-01 -8.68224025e-01
-5.80412932e-02 -6.48654401e-01 -4.40020524e-02 6.41866505e-01
-6.67009316e-03 1.15569822e-01 -6.40779316e-01 2.62327254e-01
8.22871447e-01 -2.39002511e-01 1.17900324e+00 -7.78235972e-01
-2.78371453e-01 1.91756878e-02 -6.27330720e-01 -1.69961751e+00
2.10644543e-01 -5.87719262e-01 2.78928820e-02 -1.24140668e+00
1.43317878e-01 1.19297959e-01 6.70817420e-02 5.23308694e-01
8.70566536e-03 7.41312385e-01 2.35692382e-01 2.73034662e-01
-8.95633042e-01 3.86968672e-01 1.15923989e+00 1.70957103e-01
1.66196970e-03 -1.05299421e-01 -2.60221869e-01 1.05106080e+00
5.30247152e-01 -1.55567110e-01 -3.80341828e-01 -6.05905890e-01
-1.04592986e-01 4.07829583e-02 5.87168396e-01 -1.23981869e+00
1.66363090e-01 1.35460377e-01 8.86019349e-01 -9.47400630e-01
7.01585889e-01 -7.75692105e-01 3.21357064e-02 3.71453434e-01
-1.94590747e-01 7.25540295e-02 2.38598719e-01 6.36868775e-01
2.09740214e-02 1.47954926e-01 3.68557364e-01 -3.44829798e-01
-9.89264190e-01 7.43846178e-01 -9.55393612e-02 -3.29753608e-02
1.12673461e+00 -3.84448230e-01 1.08302064e-01 -6.56955421e-01
-8.79126072e-01 3.14793825e-01 7.12909162e-01 5.98347425e-01
7.59360015e-01 -1.50073957e+00 -5.09764791e-01 8.83366391e-02
1.82295308e-01 2.58324176e-01 3.80529672e-01 9.68620658e-01
-7.71827757e-01 5.48018515e-01 -5.02375737e-02 -1.04198682e+00
-1.38569784e+00 5.50353229e-01 3.05212677e-01 2.76001275e-01
-8.65805984e-01 1.07436872e+00 8.43485817e-02 -7.19909072e-01
5.82067907e-01 -2.83878684e-01 2.59495020e-01 -3.14121455e-01
6.07511282e-01 3.42091739e-01 -1.01526670e-01 -9.08278763e-01
-4.76758897e-01 8.76603663e-01 -1.39795363e-01 -2.09914878e-01
1.36573613e+00 -1.85312167e-01 4.84464377e-01 3.17876905e-01
1.56426120e+00 -1.40492618e-01 -2.04962015e+00 -4.39323157e-01
-4.36734706e-01 -6.14252090e-01 -1.02618508e-01 -5.32222211e-01
-9.34680521e-01 7.87620127e-01 6.08949661e-01 -5.53110242e-01
6.68960810e-01 2.70664871e-01 6.80901587e-01 7.40485072e-01
4.89821851e-01 -9.97444868e-01 3.99874419e-01 5.39130986e-01
7.84928560e-01 -1.48702765e+00 2.13226259e-01 -3.67962748e-01
-8.21241915e-01 1.06037807e+00 7.76904047e-01 -4.06894654e-01
8.74631286e-01 2.65885085e-01 4.06766683e-01 -2.49653280e-01
-4.23479408e-01 -9.06751081e-02 4.80525315e-01 6.17074311e-01
4.44032282e-01 -9.60405096e-02 3.47737372e-01 1.60016418e-01
-1.81238204e-01 2.88935661e-01 2.35907882e-01 1.06837308e+00
-2.55442977e-01 -6.30195796e-01 -3.09713751e-01 3.53293866e-02
-3.70889515e-01 1.91960596e-02 -2.33710870e-01 9.99403596e-01
-5.07998094e-02 5.42346478e-01 1.29758328e-01 -6.96884215e-01
3.13265473e-01 2.59657949e-02 7.91058004e-01 -6.01117551e-01
-1.08630545e-01 3.57678115e-01 3.36620468e-03 -1.14813328e+00
-7.46925712e-01 -6.89172924e-01 -1.41419911e+00 -3.41428220e-01
-1.43736660e-01 -2.99224615e-01 5.73153734e-01 8.65738571e-01
5.08965015e-01 1.28037944e-01 3.12283337e-01 -1.69817042e+00
-4.85417604e-01 -1.01027691e+00 -3.05157334e-01 4.07439619e-01
4.30496454e-01 -9.37322140e-01 -2.49026343e-01 2.21618801e-01] | [7.1279096603393555, -1.0015801191329956] |
c03993d8-928e-42af-b3da-935a7eb0b308 | cloud-removal-using-atmosphere-model | 2210.01981 | null | https://arxiv.org/abs/2210.01981v1 | https://arxiv.org/pdf/2210.01981v1.pdf | Cloud removal Using Atmosphere Model | Cloud removal is an essential task in remote sensing data analysis. As the image sensors are distant from the earth ground, it is likely that part of the area of interests is covered by cloud. Moreover, the atmosphere in between creates a constant haze layer upon the acquired images. To recover the ground image, we propose to use scattering model for temporal sequence of images of any scene in the framework of low rank and sparse models. We further develop its variant, which is much faster and yet more accurate. To measure the performance of different methods {\em objectively}, we develop a semi-realistic simulation method to produce cloud cover so that various methods can be quantitatively analysed, which enables detailed study of many aspects of cloud removal algorithms, including verifying the effectiveness of proposed models in comparison with the state-of-the-arts, including deep learning models, and addressing the long standing problem of the determination of regularisation parameters. The latter is companioned with theoretic analysis on the range of the sparsity regularisation parameter and verified numerically. | ['Zhuo Wang', 'Feng Li', 'Yi Guo'] | 2022-10-05 | null | null | null | null | ['cloud-removal'] | ['computer-vision'] | [ 6.00681007e-01 -3.53282839e-01 4.78467613e-01 3.95218283e-02
-5.85652113e-01 -4.46102232e-01 4.64192957e-01 -1.67353392e-01
-2.55141526e-01 5.14918447e-01 -1.19397685e-01 -1.53994057e-02
-4.65612143e-01 -8.13163936e-01 -4.98888731e-01 -1.14182281e+00
-1.55814946e-01 4.94443893e-01 -5.69591597e-02 -3.18203777e-01
-4.64030821e-03 7.21560717e-01 -1.65661073e+00 -5.16083725e-02
9.12635326e-01 7.73524284e-01 5.32503963e-01 4.54562753e-01
-2.43525431e-02 5.46009719e-01 -2.53199041e-01 -4.79272865e-02
6.20797992e-01 -2.55942017e-01 -3.98686826e-01 4.62866962e-01
5.97599983e-01 -2.79448964e-02 -1.92952320e-01 1.38013530e+00
2.46077821e-01 1.06145017e-01 7.38832951e-01 -7.27238894e-01
-4.10474427e-02 -2.95431390e-02 -7.40306497e-01 1.68650821e-01
-1.63450658e-01 -1.98633283e-01 6.23106062e-01 -9.25381780e-01
4.89774168e-01 7.33372211e-01 5.47618926e-01 -8.83109048e-02
-9.24419880e-01 -4.45134848e-01 1.43792421e-01 -4.09282465e-03
-1.58826578e+00 -4.31790560e-01 7.21626520e-01 -5.78803778e-01
2.96987265e-01 4.72465485e-01 7.10599244e-01 3.22452307e-01
8.51521790e-02 4.83633816e-01 1.30899930e+00 -6.62858188e-01
1.80283844e-01 6.50826544e-02 1.84886470e-01 4.30334628e-01
7.91900516e-01 2.00783178e-01 -2.03771800e-01 -6.43125251e-02
5.06379068e-01 2.91846484e-01 -5.84546745e-01 -4.07958567e-01
-6.88007951e-01 7.25357056e-01 4.19433922e-01 4.00599599e-01
-6.31625533e-01 -7.67134726e-02 -1.98042050e-01 1.95883363e-01
5.98576546e-01 2.80171663e-01 -7.59285763e-02 6.07268572e-01
-1.47450352e+00 4.91571337e-01 6.43947005e-01 7.24756777e-01
8.54188323e-01 3.72528285e-01 1.01593651e-01 4.91569579e-01
3.13018292e-01 1.01844299e+00 6.44570664e-02 -1.03081238e+00
2.89661616e-01 1.78405464e-01 4.08111900e-01 -1.02661121e+00
-7.85932317e-02 -1.03696334e+00 -1.37498903e+00 3.11978132e-01
8.10791999e-02 1.45495227e-02 -8.24489295e-01 1.09690964e+00
2.41501331e-01 4.84768033e-01 1.48086697e-01 9.95974898e-01
4.78540510e-01 7.79456854e-01 -3.16903591e-01 -6.14722371e-01
1.03242326e+00 -5.18586576e-01 -7.34733701e-01 -2.52237290e-01
1.95460886e-01 -7.25252748e-01 4.28653657e-01 5.49425900e-01
-7.28538454e-01 -3.85791123e-01 -8.60223949e-01 4.23596233e-01
-1.84426904e-01 6.08368292e-02 7.02630460e-01 4.59169954e-01
-9.08611774e-01 5.94717860e-01 -7.73321271e-01 -2.50968933e-01
2.08722353e-01 -1.77562963e-02 -1.41783878e-01 -5.18213630e-01
-8.96149993e-01 8.46420050e-01 3.48541513e-02 7.37516761e-01
-8.41166556e-01 -5.47952056e-01 -5.96622288e-01 1.52060047e-01
6.26038387e-02 -7.00983584e-01 6.74769998e-01 -1.11101198e+00
-8.24187338e-01 7.80896008e-01 -3.20767015e-01 -4.34124112e-01
5.16074181e-01 -4.24806982e-01 -3.76909345e-01 2.23675892e-01
4.40612715e-03 7.80736655e-02 1.32772648e+00 -1.61264789e+00
-3.26410055e-01 -4.73836035e-01 -1.69668987e-01 2.59898692e-01
1.05108947e-01 -2.28373796e-01 -2.98239201e-01 -4.81998622e-01
3.53940547e-01 -1.14079821e+00 -5.75091958e-01 -1.10113226e-01
-9.49098393e-02 6.36526883e-01 6.68579698e-01 -7.29450345e-01
8.61840129e-01 -2.29958820e+00 2.57436279e-02 4.85974371e-01
3.29157889e-01 4.52162504e-01 -2.19820470e-01 6.38219714e-01
-2.34857589e-01 -1.61753803e-01 -7.45414376e-01 -3.36225092e-01
-4.97892678e-01 3.01282525e-01 -4.83851224e-01 9.19051826e-01
1.15832493e-01 4.07674968e-01 -5.83918691e-01 -1.19431928e-01
3.31591278e-01 7.15135276e-01 -2.63767511e-01 1.73771575e-01
2.33600964e-03 5.74859381e-01 -6.23713851e-01 3.70429426e-01
1.18862057e+00 -1.19286507e-01 6.75510615e-02 -1.01709470e-01
-4.93978560e-01 -1.84039652e-01 -1.38928521e+00 1.22043276e+00
-4.19778675e-01 6.15560591e-01 6.02159202e-01 -1.09185839e+00
6.66054845e-01 1.98529124e-01 5.03943503e-01 -5.75842500e-01
8.11541304e-02 5.27804419e-02 -1.24981113e-01 -5.97793043e-01
5.17527103e-01 -3.60543281e-01 5.63531935e-01 2.10966691e-02
-5.25188327e-01 -4.29997385e-01 1.75937377e-02 1.81248710e-01
7.73151934e-01 -7.01940954e-02 1.80447698e-01 -4.27851945e-01
4.92351741e-01 1.29341274e-01 1.60926968e-01 8.57415378e-01
2.37272643e-02 6.79707110e-01 -1.01108000e-01 -3.94738048e-01
-8.79524469e-01 -6.93238318e-01 -3.39112133e-01 3.18057746e-01
2.21293002e-01 6.49220720e-02 -5.79083204e-01 2.26808041e-01
-2.98914164e-02 4.22671229e-01 -4.36800539e-01 1.22740299e-01
-3.83578092e-01 -1.21713197e+00 2.65882444e-03 -2.55977601e-01
5.99784553e-01 -7.18806565e-01 -4.75196272e-01 -3.12024932e-02
-3.35861415e-01 -1.27489996e+00 3.30965310e-01 1.25212401e-01
-9.49535728e-01 -1.15838647e+00 -5.74788392e-01 -3.59602034e-01
5.99194169e-01 9.37883139e-01 1.23845696e+00 3.01964432e-01
-1.96683511e-01 3.40734661e-01 -4.45216775e-01 -5.07354677e-01
-2.22445518e-01 -2.63808221e-01 -1.79304555e-01 3.27992082e-01
-6.92214817e-02 -7.70208836e-01 -5.76046109e-01 -1.17173955e-01
-1.20047009e+00 -3.55029069e-02 7.83875704e-01 5.00397146e-01
6.78428590e-01 5.44879496e-01 -1.34361953e-01 -1.10572577e+00
1.08684488e-01 -5.06980419e-01 -9.38231587e-01 7.39353076e-02
-5.44441640e-01 -5.90935498e-02 1.98561400e-01 1.71701863e-01
-8.70651484e-01 2.83200890e-01 5.22758886e-02 -6.11626267e-01
-6.77378997e-02 7.07498908e-01 -6.38796538e-02 -4.67342019e-01
4.01695669e-01 4.81844664e-01 -6.95969537e-02 -4.76667762e-01
2.39006117e-01 3.60188872e-01 3.95213753e-01 -1.93409666e-01
1.48146260e+00 1.00290048e+00 4.24424112e-01 -1.59284425e+00
-9.26313043e-01 -8.94540071e-01 -6.45568430e-01 -3.23899597e-01
5.21486878e-01 -1.09488797e+00 -1.43109843e-01 5.35914779e-01
-1.03678811e+00 -4.72815633e-02 -1.17298566e-01 4.19471949e-01
-1.75584286e-01 5.46837211e-01 -2.03845248e-01 -1.22760761e+00
-3.62510651e-01 -7.96948433e-01 1.02126741e+00 -2.22625181e-01
3.81005287e-01 -7.82629848e-01 2.39166856e-01 4.11286533e-01
4.32689250e-01 4.49158818e-01 7.13053226e-01 4.70518917e-02
-1.04569960e+00 -1.10095926e-01 -2.03850642e-01 5.27780890e-01
-5.44793680e-02 4.95919250e-02 -1.12340641e+00 -5.83038747e-01
4.98455405e-01 2.73992449e-01 1.12768698e+00 7.14739799e-01
8.33952844e-01 -1.42793447e-01 -2.36973494e-01 9.05170381e-01
2.02780056e+00 -2.86791593e-01 8.85989487e-01 1.63470671e-01
6.63704574e-01 7.09761739e-01 6.26698732e-01 3.57560545e-01
-2.79941540e-02 4.72047299e-01 9.42823231e-01 -4.28740710e-01
5.51708043e-02 4.78206456e-01 3.72597650e-02 8.81004274e-01
-3.91308635e-01 -3.17246437e-01 -7.26578295e-01 6.51894391e-01
-1.59413981e+00 -1.05838561e+00 -5.74482858e-01 2.32915020e+00
1.33544385e-01 -1.96329683e-01 -4.02066499e-01 2.94828087e-01
5.58205843e-01 4.98582691e-01 -1.02162443e-01 1.76714763e-01
-2.85202593e-01 4.48784918e-01 8.38313162e-01 6.89415097e-01
-1.06472874e+00 6.04419589e-01 6.14460564e+00 5.22925496e-01
-1.24404120e+00 8.82173032e-02 1.93477616e-01 1.08939387e-01
-3.47673744e-01 9.92163867e-02 -5.86029649e-01 3.49865288e-01
6.87240660e-01 1.46232784e-01 4.93526429e-01 3.84876788e-01
7.21482933e-01 -1.94591284e-01 -5.43317139e-01 9.26465333e-01
1.18771724e-01 -1.17459357e+00 1.68356746e-01 1.30416304e-01
9.86313164e-01 4.44778740e-01 5.17224148e-02 -4.62694503e-02
1.30677700e-01 -8.45535517e-01 5.12209415e-01 8.24228048e-01
5.66533923e-01 -4.49680120e-01 7.75041938e-01 5.50428867e-01
-1.08216441e+00 9.86270085e-02 -4.78530079e-01 -3.31837356e-01
-3.51596326e-02 1.23310256e+00 -3.35186392e-01 9.67402816e-01
5.96887887e-01 7.36133695e-01 -3.30420792e-01 1.23840153e+00
-7.74823278e-02 6.69578671e-01 -4.62282419e-01 5.31453073e-01
2.76793748e-01 -7.53766179e-01 7.00915575e-01 1.00962257e+00
4.67289180e-01 4.97254372e-01 2.12895319e-01 6.75733387e-01
1.43528178e-01 -6.43374920e-02 -8.39354277e-01 9.27793905e-02
1.21351711e-01 1.19088829e+00 -4.65869337e-01 -2.22782698e-02
-3.63210112e-01 6.60707474e-01 -1.09501123e-01 6.59414709e-01
-3.55975211e-01 1.73982993e-01 7.43383944e-01 4.18674022e-01
4.95027065e-01 -4.49380845e-01 4.96369340e-02 -1.27751243e+00
1.26420543e-01 -8.71613741e-01 -1.30209491e-01 -8.20114255e-01
-1.12872434e+00 5.05960941e-01 -1.48812495e-02 -1.51010084e+00
9.90153402e-02 -4.96920377e-01 -5.85600495e-01 9.21548426e-01
-2.12476683e+00 -1.05541420e+00 -6.94432378e-01 5.71477711e-01
2.89241135e-01 1.17719963e-01 6.82238221e-01 3.63619536e-01
-2.52797186e-01 -3.55597168e-01 6.70195222e-01 -7.13320896e-02
1.43136457e-01 -6.16856396e-01 1.06889695e-01 1.38376153e+00
2.34108180e-01 4.66950059e-01 1.05251372e+00 -5.36175132e-01
-1.41777885e+00 -1.46206415e+00 6.17736816e-01 -2.16486722e-01
4.98268187e-01 -2.01067552e-01 -1.00296772e+00 3.52179557e-01
1.09414780e-03 1.92676052e-01 3.65773380e-01 -1.74189597e-01
-4.57274728e-02 -2.80205697e-01 -8.33412051e-01 1.18891813e-01
7.13976264e-01 -3.97722691e-01 -4.06109363e-01 6.19221330e-01
2.90812075e-01 -1.76400483e-01 -2.80028999e-01 5.46595454e-01
3.34794134e-01 -1.16639888e+00 9.85304832e-01 -2.87196040e-01
3.72610927e-01 -4.52703863e-01 -4.49042708e-01 -1.14752555e+00
-4.88342285e-01 -3.34526688e-01 1.42572224e-01 8.63648474e-01
9.98301953e-02 -4.04755324e-01 7.50616908e-01 8.83521289e-02
7.08359927e-02 -2.62719750e-01 -7.60701358e-01 -6.78745151e-01
1.23381410e-02 -2.81582534e-01 2.31261611e-01 1.15374112e+00
-9.96141613e-01 8.32551196e-02 -5.34981012e-01 1.03802216e+00
1.00683141e+00 5.70416808e-01 6.57313168e-01 -1.53951073e+00
-1.77628934e-01 3.60251404e-02 -2.37789169e-01 -6.98067367e-01
1.00794330e-01 -6.53376043e-01 4.31404859e-02 -1.39176846e+00
1.83099210e-01 -5.33561826e-01 -2.76517868e-01 -1.19748786e-01
-3.88955027e-02 2.68738389e-01 2.73366481e-01 6.11537635e-01
-1.85862243e-01 5.55391133e-01 1.21035159e+00 -2.90721297e-01
6.36953562e-02 2.52511919e-01 -3.52518499e-01 7.73814917e-01
4.58904833e-01 -7.50422835e-01 -3.53713632e-01 -6.60909593e-01
4.20755953e-01 1.62277520e-01 6.30460739e-01 -1.17587042e+00
1.34182815e-02 -1.98172659e-01 2.12476388e-01 -5.84290206e-01
4.54390019e-01 -1.22417891e+00 4.62168425e-01 2.72812456e-01
5.15928306e-02 -1.43286481e-01 6.02592975e-02 6.61660075e-01
-4.64493096e-01 -3.80651534e-01 8.85954201e-01 -4.57672477e-01
-5.19960165e-01 4.44639653e-01 -3.87707561e-01 -1.53481066e-01
4.89067197e-01 -1.21421397e-01 7.56191686e-02 -5.49676001e-01
-7.40549922e-01 -3.37318704e-02 4.27757382e-01 -1.79739177e-01
3.87434512e-01 -6.66328669e-01 -9.67699289e-01 3.23243767e-01
-2.17538383e-02 1.38087377e-01 3.03546548e-01 7.84044921e-01
-7.58794785e-01 1.13934264e-01 1.33956283e-01 -5.67592144e-01
-1.14436960e+00 3.69215965e-01 5.41860878e-01 -2.96979874e-01
-6.34040892e-01 5.12481034e-01 3.71181607e-01 -1.83502570e-01
-1.23786308e-01 -2.26714686e-01 -2.88603812e-01 -1.08973458e-01
4.36302751e-01 1.70967549e-01 4.14998174e-01 -6.39233351e-01
-1.39324635e-01 7.84344792e-01 3.23028415e-01 1.05009630e-01
1.58712041e+00 -1.82712540e-01 -4.99799639e-01 2.33909130e-01
7.54796624e-01 2.75630116e-01 -1.05261457e+00 -3.32071155e-01
-3.03920925e-01 -6.17115378e-01 4.77191538e-01 -3.67511809e-01
-1.26547313e+00 7.86932588e-01 8.14614534e-01 2.26874635e-01
1.19567013e+00 -4.58955705e-01 3.10643435e-01 5.39650142e-01
3.67164344e-01 -5.01288176e-01 -4.62560236e-01 5.09085238e-01
8.76450837e-01 -1.23994875e+00 5.23932397e-01 -7.73302436e-01
-1.59594923e-01 6.47326469e-01 -1.19323693e-01 -4.88863051e-01
9.44141090e-01 2.76461184e-01 1.37034848e-01 -5.02471268e-01
-3.72650892e-01 -4.54585940e-01 4.00708802e-02 5.20051181e-01
8.76000524e-02 8.53222087e-02 -1.97222792e-02 -2.32082441e-01
-6.51706234e-02 -1.15133878e-02 6.17783248e-01 7.48603404e-01
-4.95139599e-01 -8.17380905e-01 -6.35214388e-01 4.04802173e-01
-3.59740376e-01 -4.70678151e-01 -3.41997184e-02 7.44844794e-01
1.96176469e-01 9.36601996e-01 -1.00815393e-01 2.71824002e-01
4.51202951e-02 -1.69382900e-01 4.13677245e-01 -5.96652269e-01
-1.60734832e-01 1.69999674e-01 -1.13246620e-01 -2.33559057e-01
-9.44020450e-01 -8.86777401e-01 -6.83569551e-01 -2.35040292e-01
-4.32304502e-01 3.98427069e-01 6.46415770e-01 1.05265784e+00
7.88827762e-02 3.34037930e-01 7.65498340e-01 -1.09855354e+00
-3.50378841e-01 -8.25983524e-01 -1.13738060e+00 2.22609773e-01
7.36659825e-01 -5.43193996e-01 -6.83167994e-01 1.01072617e-01] | [10.030427932739258, -2.007946252822876] |
572fba95-c71c-4e50-87b8-5099fc0b8d58 | point-process-modelling-of-rumour-dynamics-in | null | null | https://aclanthology.org/P15-2085 | https://aclanthology.org/P15-2085.pdf | Point Process Modelling of Rumour Dynamics in Social Media | null | ['Kalina Bontcheva', 'Michal Lukasik', 'Trevor Cohn'] | 2015-07-01 | point-process-modelling-of-rumour-dynamics-in-1 | https://aclanthology.org/P15-2085 | https://aclanthology.org/P15-2085.pdf | ijcnlp-2015-7 | ['rumour-detection'] | ['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] | [-7.397706508636475, 3.655409574508667] |
7b5ca255-6c94-4403-a381-5ee767e4e11e | theres-a-time-and-place-for-reasoning-beyond | null | null | https://openreview.net/forum?id=2Po_v-AG9u | https://openreview.net/pdf?id=2Po_v-AG9u | There’s a Time and Place for Reasoning Beyond the Image | Images are often more significant than only the pixels to human eyes, as we can infer, associate, and reason with contextual information from other sources to establish a more complete picture. For example, in Figure 1, we can find a way to identify the news articles related to the picture through segment-wise understandings on the signs, the buildings, the crowds, and more. This tells us the time when and the location where the image is taken, which will help us in subsequent tasks, such as evidence retrieval for criminal activities, automatic storyline construction, and upper-stream processing such as image clustering.
In this work, we formulate this problem and introduce TARA: a dataset with 16k images with their associated news, time and location automatically extracted from New York Times (NYT), and an additional 61k examples as distant supervision from WIT. On top of the extractions, we present a crowdsourced subset in which images are believed to be feasible to find their spatio-temporal information for evaluation purpose. We show that there exists a $70\%$ gap between a state-of-the-art joint model and human performance, which is slightly filled by our proposed model that uses segment-wise reasoning, motivating higher-level vision-language joint models that can conduct open-ended reasoning with world knowledge. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['image-clustering'] | ['computer-vision'] | [-1.89350978e-01 2.61923611e-01 -2.45678544e-01 -5.05472183e-01
-8.94105971e-01 -4.67512131e-01 9.39542055e-01 3.03511858e-01
-5.85097492e-01 7.37214744e-01 5.83753645e-01 -1.08892046e-01
-1.82118803e-01 -7.45111465e-01 -8.94820511e-01 -4.29842204e-01
1.87566146e-01 7.51869977e-01 4.44950968e-01 -1.21573120e-01
3.53859454e-01 1.94733083e-01 -1.32342732e+00 7.49086678e-01
5.02993464e-01 9.28362727e-01 3.62114251e-01 5.50520122e-01
1.39048342e-02 1.42567265e+00 -6.05921388e-01 -1.09154344e+00
5.20287566e-02 6.57485574e-02 -8.84745300e-01 3.02336782e-01
5.94966173e-01 -7.10976064e-01 -4.86688137e-01 7.84758389e-01
1.11831345e-01 -1.10824190e-01 8.47783744e-01 -1.33902895e+00
-8.26139748e-01 7.93101013e-01 -8.15181851e-01 2.84356326e-01
4.51496124e-01 2.81645805e-01 1.07144117e+00 -8.46629679e-01
1.02085495e+00 1.15882850e+00 6.88275099e-01 -1.21031009e-01
-6.00901842e-01 -4.03809786e-01 2.87349313e-01 7.46599674e-01
-1.38189745e+00 -2.47326985e-01 4.51904058e-01 -5.62564552e-01
9.08633113e-01 2.04130352e-01 7.31971145e-01 1.16595840e+00
-2.15188846e-01 1.30714607e+00 1.11900377e+00 -4.33056176e-01
-3.73125933e-02 2.35451505e-01 1.98384240e-01 8.39035451e-01
4.23908606e-02 -4.52345997e-01 -8.25916231e-01 -2.14865315e-03
6.95349514e-01 9.78992805e-02 -1.78737134e-01 1.35605196e-02
-1.50031209e+00 5.87650716e-01 7.37987459e-01 2.06534758e-01
-4.36107755e-01 1.47427410e-01 3.85984033e-02 -5.22950411e-01
4.19145644e-01 3.42201978e-01 -2.04392925e-01 -7.05494359e-02
-1.31332016e+00 3.52024227e-01 6.40623629e-01 9.45723176e-01
9.17023063e-01 -5.45017779e-01 -4.43447232e-01 6.27344608e-01
3.29324931e-01 7.11057305e-01 4.54387888e-02 -1.19083810e+00
1.05469930e+00 6.94882989e-01 4.02344257e-01 -1.26957524e+00
-2.30929449e-01 6.23470210e-02 -4.19151515e-01 -2.14557618e-01
7.27140486e-01 -1.30435437e-01 -8.53040934e-01 1.26070940e+00
3.16888601e-01 4.73962277e-01 -9.32033882e-02 9.64355588e-01
8.17957103e-01 7.12981105e-01 3.95654961e-02 1.44854575e-01
1.96201468e+00 -1.00570929e+00 -7.56332278e-01 -5.52160501e-01
3.89131665e-01 -5.82472861e-01 6.71328485e-01 2.39509612e-01
-7.81177700e-01 -4.27883327e-01 -8.63415480e-01 -3.27926069e-01
-7.45755434e-01 5.78177214e-01 5.13820469e-01 9.70746502e-02
-7.64952838e-01 2.79962212e-01 -4.49503720e-01 -6.91673219e-01
6.23675227e-01 -3.81142110e-01 -3.73374492e-01 -4.64785427e-01
-1.32876027e+00 1.23928571e+00 5.30306101e-01 4.84280176e-02
-9.01622415e-01 -4.15707529e-01 -8.79609406e-01 -2.52466291e-01
6.45534217e-01 -5.24189413e-01 1.05068159e+00 -3.84344131e-01
-5.00678778e-01 1.17990673e+00 -1.59678072e-01 -8.54830444e-01
6.24571383e-01 -3.23812157e-01 -3.20269406e-01 4.56040025e-01
8.39680374e-01 9.72437739e-01 8.76658678e-01 -1.34874129e+00
-1.06949282e+00 -4.94226038e-01 4.44699347e-01 2.21826985e-01
-1.68682069e-01 2.51537174e-01 -1.09987700e+00 -4.95738924e-01
-1.71015501e-01 -8.38422298e-01 -1.74609572e-01 9.11612734e-02
-7.60271013e-01 -2.90231556e-01 6.19957387e-01 -1.20645523e+00
1.04746830e+00 -2.00239897e+00 -1.19783588e-01 1.03771098e-01
3.21620792e-01 5.41916639e-02 1.26737908e-01 4.20497894e-01
2.71098197e-01 5.70362154e-03 -3.99705060e-02 -3.31936806e-01
3.85468118e-02 3.58628273e-01 -5.78158617e-01 4.77945805e-01
1.81048498e-01 1.06875539e+00 -1.10961986e+00 -1.00070381e+00
2.07003579e-01 1.99155763e-01 6.64068907e-02 -1.25002891e-01
-3.69474620e-01 1.22565769e-01 -5.57585895e-01 7.06222534e-01
3.22043359e-01 -5.19050062e-01 -3.83710675e-02 -4.03490484e-01
-6.17602989e-02 -1.32431686e-01 -1.11447489e+00 1.54456949e+00
5.49842604e-03 1.09311163e+00 -1.67394832e-01 -8.33130121e-01
5.42305768e-01 1.78842142e-01 2.09084377e-01 -7.75636852e-01
-5.65290377e-02 -2.01635346e-01 -6.24699712e-01 -9.95318532e-01
7.85969734e-01 3.43001664e-01 -1.32282376e-01 3.89331788e-01
-2.69177616e-01 -1.71046764e-01 4.51218307e-01 6.33540750e-01
1.09843040e+00 2.25124016e-01 1.00130379e-01 3.25675815e-01
8.78707692e-02 7.06500232e-01 3.29575688e-01 8.85853350e-01
-1.54535294e-01 7.23885715e-01 5.10088205e-01 -5.43393195e-01
-1.26940739e+00 -7.53920376e-01 -7.17109591e-02 8.36139977e-01
3.34987640e-01 -4.28828746e-01 -8.33607078e-01 -6.01545215e-01
-1.10417455e-01 9.99887049e-01 -9.19481874e-01 6.15229428e-01
-5.63358247e-01 -3.95407230e-01 8.36990416e-01 6.35104597e-01
8.92014384e-01 -9.51107323e-01 -6.56972170e-01 -5.35654612e-02
-8.71204674e-01 -1.65630448e+00 -2.18897238e-01 -1.95661396e-01
-2.82975942e-01 -1.34685707e+00 -9.46109116e-01 -4.75115955e-01
7.41392016e-01 3.60066891e-01 1.04596651e+00 -1.44233719e-01
-3.48162115e-01 6.39365435e-01 -3.81846011e-01 -5.37054121e-01
8.31118450e-02 -4.93713260e-01 -2.60002524e-01 1.43302605e-01
4.84718025e-01 -7.74482489e-02 -6.24001205e-01 3.03036273e-01
-6.57110631e-01 3.02131236e-01 5.84666371e-01 3.07807714e-01
5.26298702e-01 2.55424380e-01 -9.05986205e-02 -7.13274956e-01
4.05796349e-01 -7.60070562e-01 -1.84549585e-01 6.97803617e-01
-2.56704427e-02 -4.79254611e-02 7.24174306e-02 -2.31200323e-01
-1.26251531e+00 7.21963570e-02 4.02472705e-01 -4.37761337e-01
-3.70408416e-01 6.48366332e-01 2.23725989e-01 7.08689034e-01
8.47367764e-01 2.69692898e-01 -3.43330413e-01 -1.80850476e-01
8.57810676e-01 7.68626869e-01 8.73813272e-01 -6.25982583e-01
7.34612167e-01 1.04812133e+00 -3.84275198e-01 -8.41690719e-01
-1.50103986e+00 -8.65525663e-01 -8.42439890e-01 -5.06137490e-01
1.31010950e+00 -1.17256296e+00 -4.74369258e-01 2.67493337e-01
-1.56375504e+00 -1.99432150e-01 -8.28069672e-02 3.61734927e-01
-4.88793224e-01 4.32339013e-01 -6.18881524e-01 -9.54549313e-01
8.32660422e-02 -9.62022722e-01 1.29893148e+00 8.28022286e-02
-2.76774734e-01 -7.06615984e-01 -2.21942902e-01 1.28527462e+00
-1.72149912e-01 2.30590537e-01 5.96497357e-01 -6.09808028e-01
-8.69563460e-01 -3.01212013e-01 -8.14726889e-01 6.28879527e-03
-4.21265960e-01 -4.37286422e-02 -1.04663467e+00 3.57917309e-01
-3.12980771e-01 -2.77554095e-01 9.40036058e-01 3.96937311e-01
8.69605362e-01 -3.30833375e-01 -6.41721606e-01 7.96232745e-02
1.13437426e+00 -6.98402673e-02 8.80105913e-01 6.14233077e-01
7.90582120e-01 9.81324732e-01 9.59944308e-01 4.71039951e-01
8.99017572e-01 6.62583768e-01 5.74749529e-01 -1.15561023e-01
-2.46515244e-01 -4.46358651e-01 3.09483021e-01 2.51717836e-01
-4.77382243e-01 -2.46190995e-01 -1.15785849e+00 9.88595545e-01
-2.40705109e+00 -1.41468132e+00 -4.49962616e-01 1.58111203e+00
6.62690103e-01 2.69906446e-02 -2.48881779e-03 -1.44396067e-01
8.34754348e-01 1.88148245e-01 -4.11401123e-01 2.95658976e-01
-2.73133069e-01 -4.64679062e-01 7.38565862e-01 1.93055451e-01
-1.18221271e+00 9.26224828e-01 5.95436811e+00 1.24649811e+00
-4.13321167e-01 1.85264021e-01 8.97286475e-01 4.31548990e-02
-8.02875236e-02 2.57500142e-01 -9.84331608e-01 3.46536040e-01
4.82465386e-01 4.35530633e-01 2.44560301e-01 8.10936809e-01
3.60101312e-01 -6.41975284e-01 -1.11456585e+00 1.12265313e+00
5.34124672e-01 -1.72993934e+00 -1.03967898e-01 1.32226735e-01
7.52347708e-01 9.69198719e-02 -2.22393557e-01 1.20138697e-01
4.91952747e-01 -1.03604150e+00 1.20282078e+00 8.68434012e-01
4.42873955e-01 -3.30800921e-01 7.17273057e-01 7.77170300e-01
-1.05787408e+00 -1.93581730e-01 -1.63474500e-01 -1.04776084e-01
3.59627098e-01 7.27658629e-01 -1.17111325e+00 5.39207697e-01
8.71421218e-01 8.96777451e-01 -7.82808483e-01 9.40032661e-01
-8.63958120e-01 4.91887063e-01 -3.13611627e-01 -1.93041414e-01
5.71259201e-01 -2.31853742e-02 3.50881547e-01 1.30323732e+00
4.01526958e-01 3.67592663e-01 1.88979656e-01 9.73156810e-01
-2.97023654e-02 -2.35171840e-01 -6.99940562e-01 3.27700794e-01
4.45522249e-01 1.21361828e+00 -1.08382022e+00 -6.55917585e-01
-4.43688840e-01 9.19672489e-01 5.76484442e-01 5.99784195e-01
-1.12973154e+00 -5.81326224e-02 1.94451421e-01 4.03928399e-01
3.58637244e-01 -2.60777891e-01 -2.17172742e-01 -1.08522630e+00
1.63127735e-01 -4.44142491e-01 4.22796458e-01 -1.70194781e+00
-1.26454616e+00 1.83684587e-01 3.15953255e-01 -1.10566342e+00
-4.91875708e-02 -6.20737791e-01 -3.64741266e-01 5.52703142e-01
-1.62698936e+00 -1.52451038e+00 -3.34009707e-01 6.59110665e-01
6.85067713e-01 4.49343175e-02 2.53074914e-01 9.05625895e-02
-4.27612692e-01 -1.70584500e-01 -2.17038050e-01 8.67273092e-01
7.04874992e-01 -1.15456998e+00 1.75166517e-01 1.08158600e+00
6.68484449e-01 3.73168439e-01 6.77046359e-01 -8.79337013e-01
-9.58990276e-01 -1.01478851e+00 1.09436655e+00 -1.10770547e+00
9.77440476e-01 -8.56091529e-02 -4.41475779e-01 7.33875692e-01
2.90000528e-01 -2.18009800e-01 4.26810145e-01 7.09401295e-02
-4.43465054e-01 1.24233134e-01 -9.35152888e-01 8.07431817e-01
8.52846742e-01 -6.88869715e-01 -9.88661468e-01 7.61530042e-01
5.43944359e-01 -3.71060818e-01 -5.35777986e-01 -1.50145769e-01
3.29395533e-01 -8.71807873e-01 1.18901348e+00 -3.20908725e-01
8.88542235e-01 -4.33410883e-01 -1.79465309e-01 -8.88134599e-01
-1.39706597e-01 -4.46020402e-02 7.83526748e-02 1.33324492e+00
6.62239790e-01 3.65314707e-02 6.59772456e-01 1.09537315e+00
1.00039929e-01 -4.45201635e-01 -9.22850072e-01 -4.12757307e-01
-4.95705843e-01 -1.03884113e+00 4.03856874e-01 1.04152954e+00
1.25702024e-01 3.16543907e-01 -5.85596800e-01 5.37928104e-01
6.10724151e-01 1.72825053e-01 7.72862554e-01 -8.43491077e-01
-1.01567112e-01 -1.47400990e-01 -3.63273799e-01 -1.26310968e+00
-1.11196443e-01 -4.56072420e-01 1.39309272e-01 -2.19371009e+00
5.87970555e-01 -4.52256411e-01 1.82526782e-01 8.38531494e-01
-1.24023676e-01 3.35942239e-01 3.30854446e-01 5.10169625e-01
-1.14124918e+00 9.42081586e-02 1.29658806e+00 -5.12165129e-01
3.22535366e-01 -3.24879080e-01 -5.52011073e-01 1.23385572e+00
2.43933007e-01 -4.40912038e-01 -2.17635363e-01 -6.80291772e-01
5.71645737e-01 2.81518608e-01 7.94413090e-01 -9.10991848e-01
7.19090343e-01 -3.36906374e-01 6.36821330e-01 -9.40356314e-01
6.96259260e-01 -9.60916519e-01 -9.55560207e-02 1.10655837e-01
-3.27184051e-01 -1.56008169e-01 -1.94763124e-01 9.46908057e-01
-1.77166417e-01 -5.37498415e-01 3.45209381e-03 -4.93271440e-01
-1.16510415e+00 1.55145749e-01 -2.66117007e-01 3.68069927e-03
1.22337162e+00 -4.11983609e-01 -7.67953336e-01 -5.79619110e-01
-7.36844480e-01 5.57731271e-01 1.73046097e-01 2.15725288e-01
6.51995420e-01 -1.13314140e+00 -8.73440981e-01 -5.20073712e-01
4.02897954e-01 9.13107693e-02 3.47122073e-01 8.75097275e-01
-5.43570459e-01 3.53474498e-01 1.55176878e-01 -6.23853445e-01
-1.11175990e+00 6.59707606e-01 -2.24263236e-01 -4.01594162e-01
-6.49565518e-01 7.18364120e-01 -8.65582079e-02 -1.25645846e-02
2.57284731e-01 -4.75130677e-01 -6.06533229e-01 4.87996846e-01
6.83183372e-01 2.22597644e-01 -3.23460311e-01 -8.91939282e-01
-3.36609960e-01 5.38246751e-01 1.78402126e-01 -3.62250477e-01
1.32952893e+00 -2.29337037e-01 -6.02820441e-02 2.90361911e-01
6.41822100e-01 -6.76588416e-02 -1.36161113e+00 -5.67676663e-01
-1.07994698e-01 -4.56188530e-01 -2.23543063e-01 -8.44913721e-01
-8.27499032e-01 7.69563675e-01 1.29791483e-01 3.86795044e-01
6.45352960e-01 4.81782258e-01 7.27643847e-01 5.93410313e-01
5.17505169e-01 -1.51336694e+00 3.42894793e-01 4.52007741e-01
9.17509139e-01 -1.56402338e+00 3.02784264e-01 -2.71835387e-01
-1.16679108e+00 1.01969218e+00 4.15021420e-01 1.15540646e-01
5.77632725e-01 2.98865233e-02 -1.93480238e-01 -6.44750237e-01
-4.18345660e-01 -6.65296853e-01 2.54110932e-01 8.00033033e-01
-2.20888168e-01 1.68901294e-01 3.09030473e-01 5.62495887e-01
-1.20552659e-01 -2.20937952e-01 3.55104536e-01 6.64231241e-01
-5.99561274e-01 -5.02026081e-01 -7.40204513e-01 4.06629920e-01
-1.62080243e-01 -1.35336041e-01 -4.13956434e-01 7.57412910e-01
5.67643702e-01 1.15150619e+00 -1.62974391e-02 -5.58581911e-02
2.54637182e-01 -1.08428977e-01 2.26052448e-01 -4.71019447e-01
-1.23195730e-01 -2.81467199e-01 6.25798881e-01 -3.92797947e-01
-8.91374826e-01 -6.98809624e-01 -1.19545603e+00 -7.00921938e-02
-1.43435225e-01 -1.47455841e-01 6.50087714e-01 1.51866794e+00
8.38379040e-02 3.00468773e-01 1.47863841e-02 -9.84590888e-01
2.32785389e-01 -8.66895199e-01 -4.15918976e-01 6.91179216e-01
-1.86927561e-02 -5.13177097e-01 -9.30756778e-02 6.44121587e-01] | [10.697948455810547, 1.3101729154586792] |
355f82cf-342b-487f-bb8a-a9be785b209f | learning-decoupling-features-through | 2203.16772 | null | https://arxiv.org/abs/2203.16772v1 | https://arxiv.org/pdf/2203.16772v1.pdf | Learning Decoupling Features Through Orthogonality Regularization | Keyword spotting (KWS) and speaker verification (SV) are two important tasks in speech applications. Research shows that the state-of-art KWS and SV models are trained independently using different datasets since they expect to learn distinctive acoustic features. However, humans can distinguish language content and the speaker identity simultaneously. Motivated by this, we believe it is important to explore a method that can effectively extract common features while decoupling task-specific features. Bearing this in mind, a two-branch deep network (KWS branch and SV branch) with the same network structure is developed and a novel decoupling feature learning method is proposed to push up the performance of KWS and SV simultaneously where speaker-invariant keyword representations and keyword-invariant speaker representations are expected respectively. Experiments are conducted on Google Speech Commands Dataset (GSCD). The results demonstrate that the orthogonality regularization helps the network to achieve SOTA EER of 1.31% and 1.87% on KWS and SV, respectively. | ['Yuexian Zou', 'Yujun Wang', 'Peng Gao', 'Weiji Zhuang', 'Rongzhi Gu', 'Li Wang'] | 2022-03-31 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 3.01430970e-02 -7.48800412e-02 -2.34073848e-01 -6.57883346e-01
-9.19627190e-01 -3.12578142e-01 4.48290706e-01 -5.54890454e-01
-1.54550627e-01 2.71150649e-01 4.32434380e-01 -5.09164691e-01
1.14013635e-01 -9.28414706e-03 -3.22728246e-01 -7.76125610e-01
1.71097010e-01 -2.75217086e-01 -9.25618857e-02 -1.05375312e-01
1.12001806e-01 4.18831289e-01 -1.45215178e+00 1.46548962e-02
7.08199859e-01 1.08499539e+00 3.57206047e-01 6.36566520e-01
-1.07993096e-01 5.28686702e-01 -5.56504548e-01 -1.87191352e-01
3.86355340e-01 -3.31430048e-01 -6.06776357e-01 -1.81779459e-01
4.47668552e-01 -3.94780338e-01 -6.83006048e-01 1.03010011e+00
9.13285613e-01 1.76751986e-01 4.63923901e-01 -1.49255431e+00
-9.52354908e-01 1.04668450e+00 -4.01114911e-01 3.12096238e-01
1.63729899e-02 1.06775045e-01 1.31593692e+00 -1.00459671e+00
-4.16426286e-02 1.21740198e+00 5.24239302e-01 7.05729187e-01
-1.03191531e+00 -1.12485206e+00 3.81767154e-01 3.90771896e-01
-1.61670160e+00 -1.00875759e+00 1.05283260e+00 -2.07066208e-01
8.98851216e-01 5.98364353e-01 2.19882563e-01 1.29212403e+00
-2.24776864e-01 1.27742958e+00 1.00480628e+00 -4.29907829e-01
-1.11679435e-01 5.34634054e-01 5.26192904e-01 4.03195977e-01
-2.12695941e-01 9.17540565e-02 -9.47180212e-01 -3.39628421e-02
2.80292809e-01 -2.78709173e-01 -7.79337764e-01 -1.07678905e-01
-1.24395096e+00 8.26311886e-01 -1.26628473e-01 4.58950937e-01
-9.57849026e-02 -9.55739394e-02 5.94383597e-01 4.16686088e-01
1.29641339e-01 6.83683679e-02 -5.95209718e-01 -1.39598370e-01
-1.04041314e+00 -4.98561375e-02 6.68078959e-01 1.11375892e+00
4.46777940e-01 8.60360384e-01 -3.23694408e-01 1.02699530e+00
5.51189303e-01 7.66680002e-01 1.00444400e+00 -3.39571834e-01
3.76584172e-01 2.36824915e-01 -4.26077634e-01 -7.12431431e-01
-1.27392054e-01 -6.63333893e-01 -6.89654469e-01 -2.79591858e-01
-1.76094607e-01 -3.92102115e-02 -9.65734243e-01 1.97156537e+00
1.02130413e-01 5.05113125e-01 4.52881306e-01 8.07612002e-01
1.31480873e+00 7.41194546e-01 -4.12900932e-02 -1.55862272e-01
1.54530597e+00 -9.52462852e-01 -9.44595397e-01 -1.72497705e-01
4.66513991e-01 -8.15596640e-01 1.35172856e+00 3.23308222e-02
-7.14672029e-01 -6.59522772e-01 -1.22031653e+00 -1.07153319e-03
-2.87655056e-01 5.37947178e-01 3.92203629e-01 9.68820930e-01
-1.04367268e+00 1.10755585e-01 -4.55069214e-01 -8.11323375e-02
2.36302223e-02 4.10583228e-01 -3.18520665e-01 3.75001371e-01
-1.60364437e+00 6.24898732e-01 2.96672434e-01 1.63205206e-01
-1.00801003e+00 -6.22923911e-01 -8.73281002e-01 2.91042775e-01
1.26545653e-01 -2.40799487e-01 1.37575519e+00 -8.72895956e-01
-2.00708914e+00 7.24321961e-01 -3.13120246e-01 -4.13150072e-01
1.10036537e-01 -1.27357930e-01 -1.02549040e+00 3.46371233e-02
-2.18441477e-03 2.94517487e-01 1.27382159e+00 -1.01823008e+00
-6.91060543e-01 -4.06549931e-01 -4.01464671e-01 2.76913226e-01
-7.66597033e-01 4.09730375e-01 -2.81738490e-01 -7.60040879e-01
1.90330729e-01 -6.22815907e-01 4.41467047e-01 -5.11847913e-01
-8.44278693e-01 -5.23688078e-01 1.14673603e+00 -1.16460419e+00
1.33062470e+00 -2.50473380e+00 3.12681422e-02 2.13155761e-01
1.00927420e-01 4.88422513e-01 -1.46898136e-01 3.02821219e-01
-1.71558127e-01 -8.08858946e-02 1.12880841e-02 -6.16761625e-01
1.88548774e-01 -7.33236745e-02 -5.02783895e-01 5.74642897e-01
1.36834785e-01 7.60352492e-01 -3.17007005e-01 -2.34423071e-01
3.01103909e-02 5.38526535e-01 -2.90955603e-01 3.09453279e-01
4.68636930e-01 2.27303922e-01 -3.26539218e-01 7.21800864e-01
8.98620248e-01 1.99285746e-01 1.23871900e-02 -3.28034967e-01
-1.68800086e-01 7.15233207e-01 -1.46865082e+00 1.35175431e+00
-3.61449063e-01 7.26225793e-01 4.19089258e-01 -1.37325954e+00
1.13927019e+00 6.94319665e-01 2.26468537e-02 -5.60122192e-01
1.71759397e-01 2.85397738e-01 1.67479232e-01 -3.13996911e-01
4.05428350e-01 -1.97568804e-01 9.72140282e-02 2.79718041e-01
2.48629481e-01 1.80119142e-01 -6.09137475e-01 6.89873844e-02
3.81802529e-01 -4.48557019e-01 9.30172205e-02 -5.29429138e-01
8.44978571e-01 -7.40117073e-01 8.76545906e-01 5.83076835e-01
-3.92420679e-01 3.66530508e-01 -7.49679701e-03 -7.08207339e-02
-6.02919042e-01 -1.03767931e+00 -2.03688517e-01 1.16252589e+00
1.39653593e-01 -2.64686912e-01 -5.48012674e-01 -4.16929394e-01
-2.92869899e-02 9.07141447e-01 -2.35397220e-01 -4.32285219e-01
-4.92821902e-01 -1.99672490e-01 1.00604463e+00 5.15518844e-01
4.91085768e-01 -7.70320177e-01 -1.37103990e-01 -5.84508851e-02
-7.76509792e-02 -1.27787137e+00 -7.89864063e-01 2.45663807e-01
-3.37000638e-01 -5.94807506e-01 -8.33013892e-01 -1.02856565e+00
2.80331522e-01 6.67970598e-01 6.06148362e-01 -2.31002167e-01
1.85849816e-02 3.46921563e-01 -3.51978600e-01 -4.51492161e-01
-5.10464787e-01 1.79213837e-01 7.10779011e-01 4.23302859e-01
5.46900034e-01 -5.70471227e-01 -2.65011579e-01 2.01909348e-01
-6.09029651e-01 -2.34378576e-02 5.11520386e-01 1.03718686e+00
2.17718966e-02 -2.19587803e-01 9.31122303e-01 -1.63054168e-01
7.75955260e-01 -1.70546353e-01 -6.28281713e-01 3.59698713e-01
-5.84835649e-01 3.72206531e-02 6.50821686e-01 -6.58853412e-01
-9.55350220e-01 -1.06623396e-01 -2.72328883e-01 -5.67710400e-01
-1.66975886e-01 3.10682297e-01 -6.02643132e-01 3.15366691e-04
8.87783095e-02 9.60604072e-01 9.28298235e-02 -5.30453742e-01
2.40199372e-01 1.44436157e+00 3.84981513e-01 -1.92683831e-01
8.67035508e-01 3.30556184e-02 -8.12545598e-01 -1.30104446e+00
-3.63409698e-01 -8.56109917e-01 -2.12821871e-01 -3.05552594e-02
6.86298251e-01 -1.33075738e+00 -7.96827614e-01 6.04955852e-01
-1.05171394e+00 2.08971351e-01 -5.69791393e-03 9.29892898e-01
-2.07181826e-01 5.64065576e-01 -2.48845801e-01 -1.07703698e+00
-6.22058868e-01 -1.42640531e+00 1.09865224e+00 3.05827618e-01
-6.10654652e-02 -6.20077252e-01 -2.69514650e-01 3.51821452e-01
6.76791191e-01 -6.79251730e-01 6.67311966e-01 -1.31261504e+00
-2.92054981e-01 -1.63870919e-02 -9.96676683e-02 8.58931959e-01
3.61275733e-01 -9.26024094e-02 -1.58052540e+00 -3.45924407e-01
3.98848802e-01 -1.01187140e-01 7.24470139e-01 3.20420831e-01
9.64102507e-01 -3.31546426e-01 -1.68039650e-01 7.50444114e-01
8.34830642e-01 4.34301168e-01 3.42635512e-01 6.93200082e-02
5.66022158e-01 4.40318614e-01 1.77252684e-02 2.25832731e-01
3.61907572e-01 8.70217025e-01 -1.77109446e-02 -8.99235159e-02
-1.61133036e-01 -2.59162456e-01 7.33147562e-01 1.40039194e+00
5.57626545e-01 5.68971299e-02 -6.88492060e-01 6.25499904e-01
-1.40495646e+00 -8.69718611e-01 1.33796796e-01 2.18028307e+00
8.27344835e-01 -9.85636562e-02 2.16199774e-02 4.30485994e-01
8.99690688e-01 3.26090813e-01 -5.22547662e-01 -4.12619710e-01
-3.14877778e-01 -8.02610815e-02 2.33322576e-01 5.66108346e-01
-1.07654929e+00 9.64822114e-01 5.82856607e+00 1.00379634e+00
-1.54347408e+00 -8.32156315e-02 5.35238862e-01 1.69494882e-01
-4.19290274e-01 -2.22187057e-01 -1.25360882e+00 4.06611949e-01
1.05108941e+00 -3.13954890e-01 2.80713975e-01 1.08755767e+00
1.61214843e-01 4.79647398e-01 -1.03954327e+00 1.39268172e+00
3.11246246e-01 -1.01178718e+00 5.04684299e-02 -6.15835339e-02
2.24072471e-01 3.10230851e-02 3.07592422e-01 6.55579507e-01
4.71675508e-02 -9.52685893e-01 8.64728272e-01 2.64060311e-03
7.24485457e-01 -6.82745397e-01 5.84724903e-01 4.40467656e-01
-1.35889006e+00 -2.49169990e-02 -2.51937360e-01 1.67304382e-01
9.62353200e-02 2.21511707e-01 -1.13852394e+00 5.85912168e-01
6.87771320e-01 4.20033872e-01 -2.00539514e-01 6.24406278e-01
-6.43642843e-02 9.39298391e-01 -2.07838908e-01 -1.02743641e-01
2.04381764e-01 -7.72856250e-02 7.21354246e-01 1.36600566e+00
2.68971264e-01 -1.74527034e-01 5.73566668e-02 9.12387550e-01
-1.59890205e-01 2.71136880e-01 -5.92949390e-01 -2.60108411e-01
7.96550095e-01 1.15317583e+00 -1.76402688e-01 -2.94920594e-01
-4.52769160e-01 9.28437889e-01 4.11246419e-02 4.57461089e-01
-7.91027188e-01 -5.22433400e-01 1.15873432e+00 -3.31366211e-01
4.22886848e-01 -2.34235108e-01 -6.01677150e-02 -1.27739716e+00
9.98212770e-02 -1.01935267e+00 -7.59502314e-03 -2.59258360e-01
-1.25569868e+00 6.79548264e-01 -2.88646102e-01 -1.12251711e+00
-1.15173556e-01 -4.71860528e-01 -7.33274758e-01 1.26078463e+00
-1.99254036e+00 -1.31178188e+00 3.76096293e-02 7.58910239e-01
7.89968371e-01 -6.82606399e-01 8.81640553e-01 2.28225857e-01
-8.40985835e-01 1.19919765e+00 3.48685116e-01 4.23190206e-01
5.54742992e-01 -9.01529670e-01 6.31477654e-01 9.77793097e-01
3.63361746e-01 9.27340627e-01 4.76198882e-01 -2.06415296e-01
-1.65406120e+00 -6.81360483e-01 9.72572923e-01 1.17139511e-01
5.56859374e-01 -6.23572946e-01 -9.55418050e-01 5.61741889e-01
3.08935583e-01 -4.27812636e-02 9.78186011e-01 1.50527388e-01
-5.95071316e-01 -2.75729597e-01 -7.98523188e-01 6.40664935e-01
7.39456534e-01 -1.05117118e+00 -6.85270250e-01 1.64812043e-01
1.02112138e+00 -1.80249721e-01 -4.80488688e-01 3.41178715e-01
5.50276637e-01 -9.11921561e-01 9.80891466e-01 -6.64130807e-01
-2.79552788e-01 -2.69430429e-01 -5.70523560e-01 -1.21904600e+00
-7.30753243e-02 -7.32146561e-01 -6.29746392e-02 1.64265645e+00
5.29659688e-01 -8.63486826e-01 3.88839900e-01 5.01680017e-01
-3.17976773e-01 -5.33322275e-01 -1.29673421e+00 -1.27057266e+00
-2.29504794e-01 -5.84881842e-01 7.52674222e-01 1.16682851e+00
-1.63631029e-02 5.99921703e-01 -7.43002534e-01 5.83504081e-01
2.98624068e-01 1.38465941e-01 7.30210185e-01 -1.14615428e+00
-1.10031985e-01 -4.92571473e-01 -4.12320048e-01 -1.29907441e+00
5.05948246e-01 -8.53435755e-01 -2.24403925e-02 -1.11409569e+00
-7.03885313e-03 -1.54308826e-01 -5.10080338e-01 3.51276696e-01
-3.04740995e-01 -3.26320678e-01 1.95366696e-01 1.47503912e-01
-6.88016489e-02 1.02735794e+00 6.55153036e-01 -2.56592304e-01
-2.05551103e-01 2.48426393e-01 -9.03369129e-01 5.55399239e-01
9.33821440e-01 -9.63759720e-02 -5.85768521e-01 -2.64175981e-01
-5.49554646e-01 -2.13680312e-01 3.95276211e-02 -7.60872543e-01
3.05483282e-01 5.64281233e-02 -7.42938221e-02 -4.56592321e-01
5.57809770e-01 -6.72543228e-01 -3.04413229e-01 3.19731146e-01
-4.00102228e-01 -3.05679083e-01 2.26610973e-01 5.00528157e-01
-4.96050507e-01 -1.47509977e-01 7.84218431e-01 1.87045515e-01
-8.08032870e-01 1.31735787e-01 -3.66114765e-01 -7.26454258e-02
8.45795572e-01 -3.36106300e-01 4.14420180e-02 -6.93937719e-01
-3.06741267e-01 1.88859478e-01 -3.38250548e-01 8.18906486e-01
8.61380100e-01 -1.48334277e+00 -8.04436982e-01 6.51749015e-01
1.82616130e-01 -4.26474601e-01 3.86351317e-01 8.38096380e-01
7.69566000e-02 7.43735313e-01 2.15862319e-01 -6.48483217e-01
-1.66063142e+00 2.64137924e-01 2.83211559e-01 2.19483852e-01
-5.97584069e-01 1.26030028e+00 3.71851742e-01 -7.11021245e-01
6.55440509e-01 -4.63425428e-01 -2.23356888e-01 -6.91635208e-03
7.44658589e-01 1.15872115e-01 1.87081575e-01 -9.75210249e-01
-6.32396698e-01 3.82790148e-01 -3.52041215e-01 -6.53441921e-02
1.26738238e+00 -2.82365650e-01 2.63229787e-01 3.96483541e-01
1.46803999e+00 2.89382666e-01 -8.72604668e-01 -6.75996602e-01
6.80241287e-02 -3.87636244e-01 3.66466016e-01 -5.37310302e-01
-1.01944649e+00 1.12182319e+00 7.28160143e-01 2.69135267e-01
1.02499521e+00 -1.18182570e-01 9.28086996e-01 1.82521418e-01
-1.15871504e-01 -1.23018324e+00 -1.77989602e-01 5.01420915e-01
8.95390928e-01 -1.29039752e+00 -3.95023137e-01 -3.59455973e-01
-8.70914459e-01 1.05348408e+00 4.37897027e-01 3.71824116e-01
6.19630814e-01 3.30215901e-01 2.83012390e-01 5.60069382e-02
-4.14412469e-01 -8.30567325e-04 4.75933731e-01 5.46813607e-01
4.17544574e-01 1.93902493e-01 5.24002314e-02 9.23026562e-01
-3.91175389e-01 -5.12482941e-01 1.73704207e-01 6.95831299e-01
-4.06747699e-01 -8.86624336e-01 -3.75167102e-01 1.72818959e-01
-4.71134633e-01 -4.75705206e-01 -3.66285712e-01 5.74196279e-01
-4.53167140e-01 1.00268531e+00 -3.67706329e-01 -6.75474405e-01
2.28782386e-01 4.34992790e-01 -2.09536999e-01 -3.73541415e-01
-4.33998972e-01 1.28573090e-01 3.93205620e-02 -3.01878601e-01
-1.61342055e-01 -3.80060613e-01 -1.07237256e+00 -1.20679019e-02
-8.50154579e-01 3.22613388e-01 9.66224253e-01 9.46823478e-01
4.62383449e-01 5.70771873e-01 9.76212204e-01 -3.89083624e-01
-1.13708389e+00 -1.11813045e+00 -9.19912755e-01 4.43638489e-02
6.20643854e-01 -6.16608679e-01 -6.42035782e-01 -1.10666305e-01] | [14.389182090759277, 6.1090264320373535] |
c0f396e4-52a3-476c-90a9-12ea7f9b97fe | multiresolution-fully-convolutional-networks | 2201.02350 | null | https://arxiv.org/abs/2201.02350v1 | https://arxiv.org/pdf/2201.02350v1.pdf | Multiresolution Fully Convolutional Networks to detect Clouds and Snow through Optical Satellite Images | Clouds and snow have similar spectral features in the visible and near-infrared (VNIR) range and are thus difficult to distinguish from each other in high resolution VNIR images. We address this issue by introducing a shortwave-infrared (SWIR) band where clouds are highly reflective, and snow is absorptive. As SWIR is typically of a lower resolution compared to VNIR, this study proposes a multiresolution fully convolutional neural network (FCN) that can effectively detect clouds and snow in VNIR images. We fuse the multiresolution bands within a deep FCN and perform semantic segmentation at the higher, VNIR resolution. Such a fusion-based classifier, trained in an end-to-end manner, achieved 94.31% overall accuracy and an F1 score of 97.67% for clouds on Resourcesat-2 data captured over the state of Uttarakhand, India. These scores were found to be 30% higher than a Random Forest classifier, and 10% higher than a standalone single-resolution FCN. Apart from being useful for cloud detection purposes, the study also highlights the potential of convolutional neural networks for multi-sensor fusion problems. | ['Bhaskar Ramachandra Nikam', 'Prasun Kumar Gupta', 'Claudio Persello', 'Debvrat Varshney'] | 2022-01-07 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [ 6.05502188e-01 -6.94943070e-01 1.03234962e-01 -3.39311302e-01
-8.65981996e-01 -7.33987570e-01 3.49460810e-01 -2.22935677e-01
-1.70908287e-01 8.50129664e-01 -2.22806677e-01 -3.97282213e-01
-3.29540282e-01 -1.18435884e+00 -4.58618492e-01 -9.90431011e-01
-6.98210821e-02 2.64606655e-01 -1.57510698e-01 -5.38314164e-01
-3.48246545e-01 8.17108929e-01 -2.18251395e+00 5.49715340e-01
1.30579698e+00 1.07989919e+00 4.14781719e-01 8.52907956e-01
-4.20787781e-02 6.07850552e-01 -3.13128054e-01 3.40920746e-01
6.63530290e-01 -1.17073156e-01 -5.69548130e-01 1.10457100e-01
1.06491387e+00 -3.45305532e-01 1.92840949e-01 9.61512029e-01
3.30629200e-01 1.06555276e-01 4.02610034e-01 -7.66474783e-01
-4.07028109e-01 3.28953356e-01 -9.17268515e-01 4.96871561e-01
-2.10506395e-01 -1.61816061e-01 9.57350850e-01 -5.47622859e-01
7.13443607e-02 7.87799835e-01 8.07374954e-01 7.34284148e-02
-9.61458147e-01 -1.04503846e+00 1.16677023e-01 -1.74937353e-01
-1.53683686e+00 -5.26710693e-03 2.94849753e-01 -7.54523337e-01
8.69123936e-01 6.87616706e-01 1.09297395e+00 2.60978729e-01
2.40246177e-01 3.59731108e-01 1.62102616e+00 -3.00632358e-01
-3.31921466e-02 -2.90689230e-01 2.08492443e-01 1.40422478e-01
7.92704523e-01 5.42302549e-01 -3.82335216e-01 1.86079778e-02
6.63844764e-01 4.65610623e-01 -2.96690613e-01 2.83434510e-01
-6.37563825e-01 1.07869494e+00 1.09726763e+00 3.66159141e-01
-6.94338799e-01 -4.49639335e-02 -1.43399462e-01 3.47518623e-01
1.02485597e+00 2.13469878e-01 -5.51084518e-01 6.35664940e-01
-1.27732503e+00 5.20601928e-01 -8.31354526e-04 2.80871332e-01
8.65821421e-01 3.36589932e-01 1.05725124e-01 7.63879597e-01
4.04859155e-01 1.16253185e+00 -1.96536675e-01 -7.60393500e-01
3.76594774e-02 2.30376065e-01 2.72987127e-01 -4.11031246e-01
-3.19678634e-01 -1.02861619e+00 -8.84215951e-01 7.64381647e-01
-1.87721029e-01 -1.66742235e-01 -1.36609495e+00 9.16652501e-01
6.28099963e-02 1.83233693e-01 4.66752470e-01 1.40288079e+00
1.25930667e+00 6.87886953e-01 4.01369669e-02 -1.81336731e-01
1.45291936e+00 -5.72154760e-01 -5.99886954e-01 -6.08023703e-01
8.71771947e-02 -6.63763285e-01 5.30902505e-01 9.07061622e-02
-4.25295353e-01 -5.65590501e-01 -1.10077870e+00 4.37148154e-01
-6.39474094e-01 4.16688398e-02 1.06288636e+00 5.69458127e-01
-1.13537097e+00 4.47726041e-01 -5.62638581e-01 -2.18038201e-01
3.23686630e-01 5.86957252e-03 -1.05005935e-01 -7.87052140e-02
-1.29992485e+00 6.76918864e-01 3.53612810e-01 8.25338066e-01
-4.77979988e-01 -6.04977131e-01 -7.34986007e-01 -4.52369265e-02
-1.95849940e-01 -5.17114460e-01 9.15392637e-01 -1.32152355e+00
-7.35585570e-01 8.83628428e-01 -3.24886262e-01 -4.68099236e-01
2.29532033e-01 -1.98867857e-01 -6.73985720e-01 5.19432649e-02
1.75912172e-01 1.70436710e-01 7.79689014e-01 -1.61602557e+00
-1.12831652e+00 -8.14516842e-01 9.99463424e-02 3.87256384e-01
3.36608946e-01 2.76485175e-01 3.53036433e-01 -2.53501356e-01
6.93706512e-01 -1.07911777e+00 -2.24573582e-01 -2.64484406e-01
1.84696883e-01 1.20212846e-01 1.03932023e+00 -7.83945441e-01
4.64846402e-01 -1.81664526e+00 -6.87062085e-01 1.29575461e-01
3.40666682e-01 5.84449112e-01 -5.99167086e-02 -7.61414990e-02
-2.87924856e-01 1.48577794e-01 -4.31517065e-01 4.01777595e-01
-7.21522748e-01 2.24723414e-01 -2.73628145e-01 5.86409867e-01
2.17653915e-01 8.80431354e-01 -7.57588744e-01 9.08091888e-02
4.50251162e-01 8.83692980e-01 2.95183659e-01 9.13248584e-03
-5.34080900e-02 4.55307662e-01 -1.47289202e-01 8.59369814e-01
1.57046700e+00 6.36480227e-02 -6.01760522e-02 -4.19718251e-02
-5.33921599e-01 5.65601848e-02 -9.70683932e-01 9.11930561e-01
-4.02738184e-01 9.68472362e-01 4.29184854e-01 -5.80220878e-01
1.17126596e+00 2.21931949e-01 1.90201387e-01 -9.98237729e-01
-9.44566056e-02 3.21905583e-01 5.44945374e-02 -4.54184830e-01
6.67287767e-01 -6.36460245e-01 4.90230113e-01 -8.00356790e-02
-6.86692595e-01 -2.65311360e-01 -3.95751297e-01 -3.30870390e-01
3.42338055e-01 7.93801174e-02 -1.46699011e-01 -2.04407960e-01
3.32368672e-01 3.87509108e-01 4.66255993e-01 9.74413514e-01
-1.68823916e-02 7.75784194e-01 -4.88924623e-01 -8.73781383e-01
-7.29633391e-01 -1.17639375e+00 -5.19140005e-01 1.13430202e+00
1.02663457e-01 4.13091600e-01 -8.37654248e-02 2.61323620e-02
3.11813742e-01 4.27219272e-01 -6.34631991e-01 4.01331186e-01
-1.16397120e-01 -1.21139216e+00 2.68826246e-01 6.98545039e-01
9.72577810e-01 -8.82680058e-01 -6.15612507e-01 3.68211679e-02
-3.73309791e-01 -1.43734348e+00 6.28657579e-01 3.50829780e-01
-1.20231998e+00 -8.07076931e-01 -3.83161277e-01 -4.10919577e-01
2.80959788e-03 1.04624581e+00 1.32436872e+00 8.71655494e-02
-3.95768434e-01 -3.54550540e-01 -5.35186172e-01 -9.67614293e-01
-4.05927561e-02 -2.90460587e-01 -3.85139197e-01 -2.14222401e-01
7.14593470e-01 -3.12161088e-01 -4.13668931e-01 -1.22482017e-01
-7.84378350e-01 1.27178147e-01 5.43317437e-01 6.04342580e-01
6.26550138e-01 3.13007534e-01 2.02204779e-01 -7.45378196e-01
1.27538726e-01 -2.52146155e-01 -9.32548583e-01 -1.20456763e-01
-4.28293824e-01 -5.62564075e-01 1.97175313e-02 4.25283253e-01
-1.08540261e+00 2.30105579e-01 -1.10943533e-01 -3.63875896e-01
-5.80611587e-01 7.36605883e-01 3.65372479e-01 -4.23146755e-01
8.55282426e-01 3.02773714e-01 -1.35121182e-01 -3.55602115e-01
-1.16417758e-01 1.27963758e+00 4.11537707e-01 1.91780571e-02
1.09432399e+00 9.64873910e-01 -9.00875777e-02 -1.28074288e+00
-1.26062167e+00 -9.58079517e-01 -6.50361836e-01 -3.02154660e-01
1.15710473e+00 -1.73839331e+00 -3.40371817e-01 6.28681839e-01
-6.85544491e-01 -1.48561131e-02 7.22862408e-02 4.80449080e-01
6.49394020e-02 -2.77775936e-02 -2.09071353e-01 -1.56184006e+00
-8.99605513e-01 -7.01506257e-01 1.15743101e+00 5.05178869e-01
4.35377419e-01 -5.64856052e-01 -1.23535655e-01 9.27296937e-01
6.91825390e-01 8.03975463e-01 2.37083852e-01 4.37726360e-03
-6.32920027e-01 -3.04164261e-01 -9.17611182e-01 4.56209719e-01
4.16583955e-01 3.06281298e-01 -1.72800708e+00 -3.99512380e-01
-2.91628093e-01 1.20839514e-02 1.56718755e+00 6.68866575e-01
7.13973463e-01 1.93185732e-01 -5.08130118e-02 9.06265795e-01
2.02297044e+00 1.69747666e-01 5.19168615e-01 6.57455504e-01
7.93984056e-01 4.12272960e-01 8.30996513e-01 1.89765766e-01
1.48029014e-01 1.88615277e-01 9.70550895e-01 -6.98929012e-01
5.39180003e-02 5.14285803e-01 -4.58498672e-02 -7.52934515e-02
-8.42451632e-01 1.72581926e-01 -1.01565170e+00 7.08071232e-01
-1.65932715e+00 -1.26493275e+00 -7.15408802e-01 2.12424874e+00
9.80987176e-02 -9.47225839e-02 -4.42948416e-02 1.21033810e-01
6.95114315e-01 4.54104602e-01 -3.70165467e-01 -3.64878744e-01
-5.67933261e-01 5.89279175e-01 1.20777547e+00 4.26720738e-01
-1.53418767e+00 8.71054888e-01 5.96279240e+00 1.81503490e-01
-1.53201509e+00 1.66868359e-01 1.24041706e-01 -4.72406298e-03
-2.99793094e-01 -1.53340042e-01 -5.43602169e-01 1.24600887e-01
1.11206675e+00 3.91730577e-01 4.88001436e-01 5.01410425e-01
7.25582063e-01 -2.58227468e-01 1.35562465e-01 8.42024148e-01
-3.57181400e-01 -1.45590734e+00 -9.49431509e-02 1.55005962e-01
8.86308134e-01 1.06532240e+00 -1.16269998e-01 -1.80536862e-02
5.58021665e-01 -1.34993029e+00 4.56032664e-01 3.66689652e-01
8.63023460e-01 -1.08044291e+00 1.21409428e+00 2.80151188e-01
-1.55789351e+00 -2.87706137e-01 -6.73859537e-01 -6.39835835e-01
-3.92156154e-01 8.50785017e-01 -5.54028988e-01 1.02711546e+00
1.27995288e+00 5.20102084e-01 -2.38938347e-01 9.64974046e-01
-1.86512843e-01 6.03822708e-01 -4.03091669e-01 3.75930905e-01
5.40498912e-01 -4.64041352e-01 2.15725437e-01 1.25136745e+00
3.65840018e-01 2.68711209e-01 2.14023679e-01 6.25351012e-01
2.49342352e-01 -2.76406497e-01 -7.95517504e-01 5.79442829e-02
2.41005734e-01 1.41126847e+00 -4.47813034e-01 -2.54927844e-01
-3.76017213e-01 5.65547645e-01 -3.96600276e-01 4.43245620e-01
-6.74777389e-01 -1.46982148e-01 1.35002077e+00 1.33076295e-01
4.51189697e-01 -2.41756827e-01 -4.88126189e-01 -9.42936778e-01
2.21150927e-03 -5.68992496e-01 2.37505093e-01 -1.22138119e+00
-1.12483704e+00 8.17583144e-01 -2.99432278e-01 -1.41857362e+00
7.57868737e-02 -6.50270939e-01 -5.01850784e-01 1.60518730e+00
-2.49707389e+00 -1.52304757e+00 -1.31699705e+00 3.86914849e-01
2.55723000e-01 2.30565190e-01 9.50170696e-01 -1.47039965e-01
-3.89694050e-02 -1.65829077e-01 5.85116684e-01 2.36863345e-01
4.64272462e-02 -1.21924663e+00 4.53133583e-01 1.04452753e+00
-2.02657074e-01 1.60913378e-01 7.85057306e-01 -6.47665083e-01
-1.09971905e+00 -1.71797299e+00 7.10025549e-01 1.15732603e-01
2.49004945e-01 4.03682180e-02 -9.78136837e-01 5.50251365e-01
-3.24810483e-02 3.81530315e-01 7.09132493e-01 1.91075668e-01
-5.68268359e-01 -3.43690127e-01 -1.48403990e+00 -2.31775139e-02
7.98101425e-01 -7.34313726e-01 -2.84022063e-01 6.67047322e-01
5.56773603e-01 -5.19908369e-01 -7.66642153e-01 8.84132206e-01
9.21961427e-01 -1.39879370e+00 1.00709546e+00 -2.91604877e-01
4.83502209e-01 -7.04089820e-01 -5.62838912e-01 -1.28445947e+00
-7.50508606e-01 3.16825956e-01 3.85858268e-01 6.14435971e-01
3.04479539e-01 -8.08869123e-01 8.10575008e-01 -9.06461701e-02
1.79202692e-03 6.02933438e-03 -7.58373737e-01 -1.00902474e+00
2.90183183e-02 -4.96156842e-01 7.28622913e-01 1.02367806e+00
-1.13942730e+00 5.54210842e-02 -2.66896989e-02 1.14862728e+00
9.38592315e-01 8.77383232e-01 4.11441058e-01 -2.07053685e+00
2.36821398e-01 -1.15039550e-01 -3.33190024e-01 -2.51437351e-02
-3.35736619e-03 -7.86154866e-01 8.66500884e-02 -2.07956576e+00
1.73028156e-01 -4.93894935e-01 -5.05810261e-01 5.36665738e-01
-1.68005154e-01 7.78940201e-01 2.04778105e-01 3.17851543e-01
3.29123467e-01 8.25364962e-02 1.11461449e+00 -3.47038507e-01
-2.96330266e-02 2.18095005e-01 -6.38378441e-01 6.14307284e-01
9.67397690e-01 -2.01396674e-01 2.04155054e-02 -6.21303499e-01
-5.07163219e-02 -4.10637669e-02 4.33632463e-01 -1.26169860e+00
-4.60184872e-01 -4.60144728e-01 8.99867892e-01 -1.04724455e+00
4.47179794e-01 -9.30424273e-01 4.27870154e-01 3.40722501e-01
2.94394702e-01 -4.17517036e-01 3.74446183e-01 7.50674829e-02
-2.59615868e-01 9.16653574e-02 1.11641502e+00 -3.70848894e-01
-1.05697191e+00 3.37059289e-01 -4.51543868e-01 -7.05616891e-01
7.53589690e-01 -6.31700575e-01 -5.14139831e-01 -3.12810123e-01
-3.96680802e-01 7.56523162e-02 4.99819666e-01 4.01368797e-01
6.27451181e-01 -6.62936509e-01 -1.13126230e+00 5.18241167e-01
5.31356037e-01 2.11777210e-01 4.49315280e-01 4.46456343e-01
-8.37414324e-01 2.59056658e-01 -2.21554235e-01 -9.69034970e-01
-1.56876826e+00 -7.54533783e-02 8.37314308e-01 -1.79924816e-02
-6.82205498e-01 9.20647621e-01 -1.80976525e-01 -7.45203137e-01
-5.18304408e-01 -2.40178242e-01 -3.67116809e-01 2.49960318e-01
6.57783747e-01 -3.16464230e-02 6.14689171e-01 -7.60636866e-01
-5.28080940e-01 7.66227484e-01 2.82234073e-01 1.90593287e-01
1.53481960e+00 -1.39191478e-01 -2.77589798e-01 4.74020153e-01
6.55143201e-01 -2.30987325e-01 -1.02682948e+00 -4.01034951e-01
-4.51226383e-01 -8.01574171e-01 8.15460086e-01 -1.00866473e+00
-1.39409173e+00 7.06086278e-01 1.10237753e+00 3.75682056e-01
1.40050280e+00 -2.61254132e-01 5.23165822e-01 2.80069917e-01
1.15080826e-01 -5.55198610e-01 -1.02832520e+00 8.08552146e-01
5.97360730e-01 -1.58208585e+00 1.01795428e-01 -5.16474783e-01
-4.42800522e-01 1.20245290e+00 4.74368572e-01 -1.54377669e-01
4.39245611e-01 4.05196875e-01 8.02381277e-01 -3.44693869e-01
-3.74148101e-01 -8.88136566e-01 1.13014415e-01 8.23876858e-01
7.83986032e-01 7.33952761e-01 1.70490861e-01 -6.91822693e-02
-3.34016442e-01 3.10477287e-01 6.05526865e-01 1.03576088e+00
-9.81403649e-01 -1.93217859e-01 -9.24408793e-01 9.13450480e-01
-3.08073252e-01 -2.64158219e-01 -7.30948746e-02 4.83139694e-01
2.79889971e-01 1.35556257e+00 6.04822576e-01 -3.77862930e-01
4.86589260e-02 -1.80361986e-01 1.22058161e-01 -3.75695437e-01
-6.70110166e-01 2.19727308e-01 6.96635991e-02 -2.55493075e-01
-9.34045374e-01 -5.58736980e-01 -9.47929144e-01 -4.96285170e-01
-4.44602966e-01 2.74410963e-01 1.10709882e+00 1.00820458e+00
-2.82306559e-02 6.80166900e-01 7.72314787e-01 -1.07512963e+00
-4.24692594e-03 -1.08084691e+00 -1.19584715e+00 -2.61655241e-01
1.09913909e+00 -5.88125706e-01 -4.04315591e-01 -4.91791040e-01] | [9.767284393310547, -1.7205301523208618] |
3e7e520c-49e0-4a16-b10c-3888dfac04da | a-hierarchical-approach-for-generating | 1611.06607 | null | http://arxiv.org/abs/1611.06607v2 | http://arxiv.org/pdf/1611.06607v2.pdf | A Hierarchical Approach for Generating Descriptive Image Paragraphs | Recent progress on image captioning has made it possible to generate novel
sentences describing images in natural language, but compressing an image into
a single sentence can describe visual content in only coarse detail. While one
new captioning approach, dense captioning, can potentially describe images in
finer levels of detail by captioning many regions within an image, it in turn
is unable to produce a coherent story for an image. In this paper we overcome
these limitations by generating entire paragraphs for describing images, which
can tell detailed, unified stories. We develop a model that decomposes both
images and paragraphs into their constituent parts, detecting semantic regions
in images and using a hierarchical recurrent neural network to reason about
language. Linguistic analysis confirms the complexity of the paragraph
generation task, and thorough experiments on a new dataset of image and
paragraph pairs demonstrate the effectiveness of our approach. | ['Li Fei-Fei', 'Justin Johnson', 'Ranjay Krishna', 'Jonathan Krause'] | 2016-11-20 | a-hierarchical-approach-for-generating-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Krause_A_Hierarchical_Approach_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Krause_A_Hierarchical_Approach_CVPR_2017_paper.pdf | cvpr-2017-7 | ['dense-captioning'] | ['computer-vision'] | [ 6.35568261e-01 6.56099558e-01 -1.99225634e-01 -4.47176456e-01
-9.71973121e-01 -5.43496192e-01 7.88751662e-01 1.02228165e-01
1.98417991e-01 1.00982141e+00 8.48239779e-01 -1.88544124e-01
6.02672219e-01 -7.44424284e-01 -1.01922214e+00 -3.12104583e-01
3.44668716e-01 1.98611066e-01 1.05112463e-01 -8.45658034e-02
1.62605926e-01 3.35252285e-01 -1.48745441e+00 1.00121176e+00
3.05679977e-01 4.44742143e-01 6.26128376e-01 8.37339163e-01
-6.26137078e-01 1.25497997e+00 -8.06698918e-01 -3.58080506e-01
-1.86429679e-01 -9.85719085e-01 -1.03243577e+00 8.19650650e-01
7.14570999e-01 -4.94493604e-01 -3.44841689e-01 7.75783658e-01
1.22762332e-02 -2.59202480e-01 7.11937964e-01 -1.02971447e+00
-1.19459474e+00 8.73617351e-01 -6.29425347e-01 -6.13655411e-02
6.61723554e-01 3.02308537e-02 8.96165252e-01 -6.03400767e-01
1.03109300e+00 1.26585865e+00 5.13322175e-01 9.04589534e-01
-1.61573601e+00 -2.01534331e-01 1.72494635e-01 -3.66369754e-01
-1.09258389e+00 -3.80904168e-01 7.51577914e-01 -3.78817618e-01
9.79633152e-01 2.84883082e-01 6.91127896e-01 1.33033669e+00
2.01391324e-01 8.58611703e-01 1.10795355e+00 -3.42009038e-01
-1.49943596e-02 4.25985605e-01 -1.67906404e-01 6.48882985e-01
2.57421583e-01 -2.65449375e-01 -3.21527511e-01 1.35712162e-01
9.99095500e-01 -2.26486698e-01 -2.13811055e-01 -2.27682307e-01
-1.53529882e+00 8.47819865e-01 5.98212183e-01 3.72575432e-01
-4.94872153e-01 4.58708733e-01 2.82171428e-01 -9.05326530e-02
5.59110820e-01 6.79819584e-01 3.36856067e-01 2.67810971e-01
-1.21606839e+00 3.70812297e-01 6.80245459e-01 1.08977377e+00
7.55661428e-01 8.92144665e-02 -5.20588577e-01 5.55095851e-01
-1.90009102e-01 4.66109335e-01 3.42839777e-01 -1.15268958e+00
3.65199149e-01 2.49172151e-01 2.48783842e-01 -1.18095636e+00
-3.37844566e-02 -2.84919113e-01 -8.88108790e-01 6.31557629e-02
2.90519577e-02 4.48552221e-02 -1.07357216e+00 1.73277783e+00
-3.47034574e-01 -2.36055300e-01 2.99616992e-01 8.66552651e-01
1.00390887e+00 1.30883503e+00 2.50141442e-01 -1.67170659e-01
1.49907708e+00 -1.10693955e+00 -8.11976075e-01 -5.90658665e-01
1.72977000e-01 -6.70729041e-01 9.79850650e-01 -2.40060799e-02
-1.58567786e+00 -5.94020545e-01 -9.56030607e-01 -4.21700537e-01
-2.62294352e-01 -2.12299854e-01 2.80872405e-01 3.81399542e-02
-1.50200379e+00 5.92003092e-02 -2.32323006e-01 -5.27085364e-01
4.43855286e-01 -1.28958791e-01 -5.11939943e-01 -1.40340164e-01
-9.45691347e-01 9.35177147e-01 5.39659619e-01 -4.09179300e-01
-8.54233205e-01 -5.22984147e-01 -1.13462079e+00 2.36386120e-01
-4.89690490e-02 -1.26997185e+00 1.35292184e+00 -1.17789769e+00
-1.04368997e+00 1.11259174e+00 -5.94663918e-01 -6.67133272e-01
2.15855882e-01 3.87720972e-01 -1.17790081e-01 7.04143822e-01
3.60609591e-01 1.65391231e+00 9.33696389e-01 -1.86673522e+00
-3.91196847e-01 1.69198588e-01 2.15863988e-01 2.86320329e-01
-1.56237185e-01 -2.46849693e-02 -4.05011386e-01 -6.86305821e-01
-2.41504475e-01 -4.83865917e-01 -3.65758777e-01 -6.10299595e-03
-5.42897582e-01 1.32369429e-01 6.36779547e-01 -6.52082860e-01
8.41747820e-01 -1.92055404e+00 9.55877006e-02 -2.30743632e-01
2.91739881e-01 -9.79345441e-02 -3.67673606e-01 5.55023551e-01
-6.23371378e-02 5.44819057e-01 -4.99230713e-01 -5.87800026e-01
-1.46152049e-01 4.40360039e-01 -8.54738772e-01 -1.13985799e-01
7.49059439e-01 1.33452928e+00 -9.65905666e-01 -8.64125073e-01
2.53038108e-01 7.10112631e-01 -4.18471158e-01 1.62414491e-01
-5.81911206e-01 2.37750679e-01 -2.21055925e-01 2.07621321e-01
3.88559520e-01 -6.37411654e-01 -4.62431498e-02 -1.78449363e-01
-6.29700199e-02 9.04713869e-02 -3.28761101e-01 1.78427207e+00
-6.47767186e-01 1.13555586e+00 -1.90770239e-01 -8.46619785e-01
8.00170302e-01 4.48350191e-01 1.97177157e-01 -6.72398746e-01
-2.57538319e-01 -1.00129873e-01 -7.73576975e-01 -6.59376800e-01
7.54243553e-01 -5.94812632e-01 -4.45196867e-01 7.24007666e-01
-2.40482390e-01 -6.82368279e-01 3.12909663e-01 4.43268269e-01
7.88694680e-01 1.13492049e-02 1.94895193e-01 1.15193538e-01
2.79619396e-01 5.09955347e-01 -2.12715715e-01 9.34179246e-01
1.46924481e-02 1.44318533e+00 5.57427466e-01 -6.12604678e-01
-1.78661120e+00 -1.26819217e+00 1.60001859e-01 5.52304327e-01
1.74284101e-01 -2.92314649e-01 -8.96392643e-01 -3.68755102e-01
-3.14417630e-01 9.72806096e-01 -6.93633676e-01 4.80359420e-02
-4.65205938e-01 -3.34657967e-01 3.96817237e-01 3.45183522e-01
5.35921395e-01 -1.39448631e+00 -7.91377068e-01 9.68660787e-02
-7.67295003e-01 -1.63776302e+00 -5.81669092e-01 -2.21102118e-01
-5.49899340e-01 -5.01183629e-01 -1.30301952e+00 -1.35841024e+00
1.06085098e+00 5.55670559e-01 1.57658577e+00 -7.93892220e-02
-3.80571395e-01 3.24421048e-01 -2.58760959e-01 -2.11329758e-01
-8.89980495e-01 -9.29772705e-02 -5.74043572e-01 -2.35614672e-01
-8.59269127e-02 -3.10371488e-01 -3.58432800e-01 -2.75422484e-01
-1.31497753e+00 9.70494986e-01 9.00738239e-01 7.75924265e-01
5.25655568e-01 -2.39071459e-01 4.23459232e-01 -9.04213428e-01
8.90435338e-01 -5.10712147e-01 -5.04307076e-02 2.60651380e-01
-1.17539071e-01 4.07211661e-01 5.93079746e-01 -3.00668776e-01
-1.07358956e+00 1.93481937e-01 9.33650658e-02 -1.88572660e-01
-3.39603722e-01 4.88043815e-01 4.24897701e-01 3.35369647e-01
6.69609666e-01 7.57948339e-01 2.55993396e-01 -2.44362019e-02
7.26551175e-01 4.66535985e-01 1.03523135e+00 -2.32670575e-01
8.08437467e-01 6.51506484e-01 -1.64101020e-01 -8.22756171e-01
-1.09246016e+00 -1.35670170e-01 -4.19414461e-01 -2.05083504e-01
1.21197963e+00 -1.03925765e+00 -1.48406982e-01 -2.01581135e-01
-1.69142401e+00 9.29558128e-02 -5.05316913e-01 4.32168972e-03
-9.45405900e-01 3.87465417e-01 -5.79933584e-01 -4.82191712e-01
-2.36159578e-01 -9.75746453e-01 1.36688221e+00 1.00564383e-01
-5.57851017e-01 -8.62068653e-01 -1.08267359e-01 3.78219038e-01
4.66934115e-01 7.37626612e-01 7.90849030e-01 3.94967981e-02
-7.18435764e-01 -4.94333841e-02 -6.11136377e-01 2.69978613e-01
1.56850368e-02 -2.25137964e-01 -8.47346842e-01 7.16258660e-02
-5.43778539e-02 -5.00576615e-01 1.03786528e+00 4.68931019e-01
1.14809477e+00 -6.19144022e-01 -1.93080202e-01 2.38009498e-01
1.63020420e+00 -7.18751773e-02 9.44219708e-01 3.48623693e-01
5.68702281e-01 7.55644321e-01 8.30275118e-02 9.19518098e-02
4.58022177e-01 4.29250985e-01 3.08501661e-01 -6.55508637e-01
-6.86120093e-01 -5.92505693e-01 3.89542758e-01 4.81648207e-01
3.60448003e-01 -4.81656730e-01 -7.62045264e-01 9.12050605e-01
-1.65411818e+00 -1.34734106e+00 2.61065811e-02 1.49475586e+00
8.38794351e-01 -2.70294044e-02 -3.50768492e-02 -3.10457051e-01
8.28169823e-01 3.09319317e-01 -3.54292631e-01 -8.49601984e-01
-4.61057007e-01 -1.88941404e-01 2.63785303e-01 6.59874141e-01
-7.07611144e-01 9.82524693e-01 7.55584621e+00 3.91052157e-01
-9.92759407e-01 -2.17951030e-01 1.09819388e+00 1.54497046e-02
-7.18094230e-01 -2.64662560e-02 -4.13673431e-01 3.30992520e-01
9.15622771e-01 -3.93585324e-01 2.45269045e-01 5.98540008e-01
3.73937786e-01 -1.05901137e-01 -9.23406363e-01 1.16818190e+00
7.29684353e-01 -1.90802991e+00 1.04625511e+00 -2.46039420e-01
1.03224421e+00 -2.66936123e-01 2.18480706e-01 -1.74369022e-01
5.76601401e-02 -1.33880758e+00 9.09056127e-01 6.26309454e-01
9.45421696e-01 -6.40565097e-01 5.64157724e-01 2.79498339e-01
-7.72784650e-01 2.39822403e-01 -3.74046683e-01 -1.89434364e-01
4.53235418e-01 4.03493851e-01 -1.09037232e+00 3.29445273e-01
4.36637998e-01 6.77345872e-01 -7.99553216e-01 6.51226044e-01
-2.00598702e-01 1.17123917e-01 1.39937252e-01 -9.09394771e-02
5.36218345e-01 1.27841085e-01 4.42765862e-01 1.36048651e+00
4.30613935e-01 1.92337275e-01 -1.61325373e-02 1.30532193e+00
-1.05099730e-01 -1.01888135e-01 -1.08762670e+00 -3.30178440e-01
4.15778421e-02 9.83477890e-01 -8.48049104e-01 -7.41349757e-01
-3.88161451e-01 1.31188262e+00 2.17557788e-01 5.25177538e-01
-8.30236793e-01 -3.71712565e-01 1.32777512e-01 3.25240910e-01
3.14817965e-01 -1.28760129e-01 -4.32589322e-01 -1.31686842e+00
1.10102244e-01 -7.47240186e-01 -9.21840817e-02 -1.72783935e+00
-1.01059949e+00 1.02366257e+00 1.81470975e-01 -1.00136864e+00
-7.31168211e-01 -2.14711979e-01 -4.38602597e-01 8.34230483e-01
-1.49191177e+00 -1.45067048e+00 -3.43507707e-01 2.02722281e-01
8.82068992e-01 2.37755865e-01 8.49721313e-01 -2.07755119e-01
-1.27284393e-01 8.46071169e-02 -2.24232733e-01 9.97351715e-04
5.87506354e-01 -1.16883886e+00 7.14371264e-01 6.94702804e-01
3.08343887e-01 3.78321826e-01 1.24204743e+00 -5.09048223e-01
-8.46464097e-01 -1.16057348e+00 1.37930477e+00 -4.12235111e-01
3.26899976e-01 -5.62829256e-01 -8.78333271e-01 6.95494711e-01
8.13634515e-01 -4.02270049e-01 3.69200170e-01 -5.21284282e-01
-5.48153281e-01 2.96644986e-01 -9.64390934e-01 8.03661346e-01
7.49599755e-01 -6.87515855e-01 -8.16325128e-01 5.38861811e-01
1.03220332e+00 -4.41384427e-02 -4.82293904e-01 4.16749381e-02
3.63496363e-01 -1.04619122e+00 1.09766114e+00 -3.77508789e-01
1.26995373e+00 -3.68108541e-01 1.72056898e-01 -1.21237361e+00
-2.94433981e-01 -5.31106293e-01 2.77256131e-01 1.13512409e+00
6.58356965e-01 -1.28655687e-01 8.05493772e-01 4.55375642e-01
-9.57545936e-02 -4.85523462e-01 -3.57238263e-01 -4.71248150e-01
7.38101602e-02 -1.60591781e-01 5.49963176e-01 6.85044110e-01
8.09718966e-02 6.88520551e-01 -5.95847607e-01 -1.21781707e-01
5.67132413e-01 4.00682569e-01 6.01080239e-01 -6.33282721e-01
-5.23472615e-02 -5.14759600e-01 -3.86527270e-01 -1.03014672e+00
3.37908745e-01 -8.58530402e-01 2.96552360e-01 -2.30373764e+00
8.01052213e-01 3.11248899e-01 3.41053128e-01 4.65707511e-01
2.70457268e-02 8.40144873e-01 3.14900935e-01 3.40802222e-01
-7.15092063e-01 2.72283435e-01 1.48979759e+00 -5.07875979e-01
1.80560514e-01 -7.74868190e-01 -1.06590545e+00 4.91648018e-01
6.75844848e-01 -2.72093982e-01 -4.89094377e-01 -7.86645770e-01
2.23991528e-01 2.21683398e-01 5.34249961e-01 -9.61855173e-01
4.42625545e-02 -4.14775796e-02 7.85080731e-01 -6.35033429e-01
3.64678770e-01 -4.30928409e-01 2.55877495e-01 4.67314601e-01
-9.64450181e-01 1.85599029e-01 3.93827945e-01 4.80684221e-01
-5.60100019e-01 -5.52185252e-02 7.87893116e-01 -5.68837047e-01
-6.38896763e-01 -6.55141324e-02 -6.70019448e-01 -1.58623323e-01
1.07043183e+00 -4.18051362e-01 -4.11519408e-01 -9.32053328e-01
-7.91721940e-01 1.52442589e-01 8.81593823e-01 5.89569986e-01
1.00620437e+00 -1.41067684e+00 -1.15585446e+00 6.14433223e-03
2.25391805e-01 -1.36120200e-01 2.37082064e-01 7.39489645e-02
-9.07868147e-01 7.60344744e-01 -4.14930314e-01 -6.80935979e-01
-1.11277044e+00 8.62621605e-01 5.46900406e-02 -1.62613705e-01
-7.83023894e-01 5.20105064e-01 7.28586197e-01 3.16604942e-01
-8.01970959e-02 -3.90484393e-01 -2.33819425e-01 -1.67965367e-01
8.14745426e-01 -5.66596746e-01 -6.81768537e-01 -8.71658385e-01
6.87175840e-02 6.92861676e-01 -1.48444280e-01 -4.66170281e-01
1.26590967e+00 -6.61647618e-01 -2.57477999e-01 5.52250564e-01
1.36962187e+00 -3.79218519e-01 -1.14805186e+00 1.64997786e-01
-2.20138863e-01 -2.15630963e-01 -1.43828690e-01 -7.09585488e-01
-5.69340765e-01 8.19882333e-01 -5.46404012e-02 4.78858024e-01
1.21264052e+00 4.85103756e-01 1.11406863e+00 1.40877426e-01
-1.07553070e-02 -5.03507376e-01 4.22880501e-01 2.65418828e-01
1.20732129e+00 -1.33704674e+00 -2.77917646e-02 -3.35161179e-01
-8.87096941e-01 1.24120426e+00 3.30629528e-01 -1.15627766e-01
-8.39458853e-02 8.65702257e-02 3.49283442e-02 -2.94784188e-01
-9.68324065e-01 -8.80995095e-02 2.50955045e-01 6.46345139e-01
4.41595078e-01 -1.16677843e-01 -1.83385229e-04 1.23108558e-01
-3.29159290e-01 -6.14425577e-02 1.14293194e+00 7.66544044e-01
-5.37835002e-01 -7.79884279e-01 -5.64951658e-01 1.71456799e-01
-4.02725935e-01 -1.95682243e-01 -7.29571998e-01 6.85303330e-01
-9.13865119e-02 8.13819289e-01 4.41674471e-01 -3.49380858e-02
-1.83854718e-02 -2.21691234e-03 5.10502398e-01 -8.97477269e-01
-2.64943302e-01 -1.34328783e-01 9.85804573e-02 -3.58640283e-01
-6.66388273e-01 -3.41447204e-01 -1.20372498e+00 -6.38863444e-02
3.79205793e-01 2.46075764e-01 7.30577528e-01 7.29145348e-01
4.55458015e-01 6.02638483e-01 4.67515886e-01 -9.17923212e-01
-2.63422597e-02 -6.12585366e-01 -2.46385261e-01 6.83754086e-01
7.61114240e-01 6.36064634e-02 -4.05851781e-01 8.30055535e-01] | [11.076837539672852, 0.9444981813430786] |
54284512-53d6-4fd0-9850-5865c994605d | joint-monocular-3d-vehicle-detection-and | 1811.10742 | null | https://arxiv.org/abs/1811.10742v3 | https://arxiv.org/pdf/1811.10742v3.pdf | Joint Monocular 3D Vehicle Detection and Tracking | Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a novel online framework for 3D vehicle detection and tracking from monocular videos. The framework can not only associate detections of vehicles in motion over time, but also estimate their complete 3D bounding box information from a sequence of 2D images captured on a moving platform. Our method leverages 3D box depth-ordering matching for robust instance association and utilizes 3D trajectory prediction for re-identification of occluded vehicles. We also design a motion learning module based on an LSTM for more accurate long-term motion extrapolation. Our experiments on simulation, KITTI, and Argoverse datasets show that our 3D tracking pipeline offers robust data association and tracking. On Argoverse, our image-based method is significantly better for tracking 3D vehicles within 30 meters than the LiDAR-centric baseline methods. | ['Philipp Krähenbühl', 'Ji Lin', 'Dequan Wang', 'Qi-Zhi Cai', 'Hou-Ning Hu', 'Min Sun', 'Fisher Yu', 'Trevor Darrell'] | 2018-11-26 | joint-monocular-3d-vehicle-detection-and-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Hu_Joint_Monocular_3D_Vehicle_Detection_and_Tracking_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Hu_Joint_Monocular_3D_Vehicle_Detection_and_Tracking_ICCV_2019_paper.pdf | iccv-2019-10 | ['online-multi-object-tracking'] | ['computer-vision'] | [-5.84704459e-01 -4.33868885e-01 -4.72907513e-01 -4.72894669e-01
-7.70935237e-01 -9.29611444e-01 8.12718868e-01 -1.19322330e-01
-4.66839790e-01 4.19401735e-01 -7.59946778e-02 -5.55190742e-01
3.18083048e-01 -6.40253127e-01 -8.98833454e-01 -4.55003351e-01
-6.20273829e-01 5.39129138e-01 7.09445834e-01 3.66057694e-01
1.19573988e-01 1.22876728e+00 -1.49469709e+00 9.60826315e-03
2.56042004e-01 9.56406772e-01 3.82849067e-01 8.78408372e-01
-6.98891059e-02 5.57429850e-01 -1.27374172e-01 4.91248667e-02
5.32269716e-01 6.75367475e-01 -3.17203216e-02 1.68304592e-01
9.71374452e-01 -1.11676526e+00 -9.12957668e-01 6.74419761e-01
6.63044229e-02 7.51425251e-02 5.62106431e-01 -1.74562943e+00
-4.03935276e-03 2.96094338e-03 -6.35257065e-01 5.74974179e-01
4.64230925e-01 5.79362094e-01 3.81616026e-01 -1.10257292e+00
8.14916134e-01 1.41633940e+00 1.06455398e+00 3.49592477e-01
-9.12558615e-01 -9.68274891e-01 4.94516134e-01 4.76548642e-01
-1.71360874e+00 -7.82996535e-01 4.39931750e-01 -8.70183647e-01
1.16815412e+00 -2.32484788e-01 6.12628162e-01 8.73554647e-01
3.21115226e-01 7.17509031e-01 3.67235869e-01 2.86948800e-01
-1.19408686e-02 2.37226253e-03 -1.02466941e-01 8.37409079e-01
3.82742941e-01 6.14855945e-01 -4.18343902e-01 -6.77097738e-02
5.90685785e-01 2.79675126e-01 2.77444124e-01 -8.02739322e-01
-1.24184763e+00 4.93640542e-01 2.17108399e-01 -2.99389899e-01
-3.49668056e-01 5.18477201e-01 1.22265629e-01 1.63176246e-02
6.63964987e-01 -4.81272638e-01 -4.56014037e-01 6.47297651e-02
-1.11958170e+00 5.99499404e-01 3.35451543e-01 1.62960052e+00
1.04288876e+00 2.71802247e-01 -2.16770157e-01 -1.63098931e-01
5.60788631e-01 1.24794412e+00 -3.16076547e-01 -1.43099809e+00
6.90232456e-01 5.22526383e-01 7.07623303e-01 -9.30875599e-01
-5.62530160e-01 -7.10484385e-02 -1.42378718e-01 5.86449802e-01
3.26781482e-01 -2.79751211e-01 -9.12329912e-01 1.57534432e+00
7.70005882e-01 6.90496743e-01 2.00413074e-03 8.40414762e-01
6.46879137e-01 6.80925548e-01 4.09444198e-02 2.71801651e-03
7.62780130e-01 -5.82747042e-01 -4.40574914e-01 -4.64720637e-01
8.03655922e-01 -3.57149690e-01 -1.74651131e-01 -3.64439726e-01
-8.42403352e-01 -6.21885598e-01 -7.82361984e-01 1.66951090e-01
-3.49843383e-01 1.64740726e-01 2.98101693e-01 4.86507654e-01
-1.27291775e+00 3.36863786e-01 -1.17759323e+00 -4.73784089e-01
5.12042820e-01 3.34432662e-01 -4.56582069e-01 -8.28531310e-02
-6.66774929e-01 1.18580878e+00 1.25269219e-01 8.84802826e-03
-1.45617354e+00 -8.37172925e-01 -1.09266424e+00 -3.62154126e-01
1.55805126e-01 -5.76577485e-01 1.22807407e+00 8.39174166e-02
-7.74076045e-01 8.89435649e-01 -8.12673986e-01 -9.56168234e-01
7.58431971e-01 -2.50734389e-01 -3.41108292e-01 4.96879779e-02
3.74368668e-01 1.12316883e+00 7.03324497e-01 -1.22078824e+00
-1.39067948e+00 -5.38120151e-01 -2.69697934e-01 2.34580845e-01
3.11067492e-01 -3.15998942e-01 -5.76903224e-01 5.54654039e-02
3.02509069e-01 -9.57387030e-01 -1.91547409e-01 4.55675900e-01
-6.82354197e-02 -1.42146677e-01 1.32964623e+00 -4.97940451e-01
6.33140385e-01 -1.90716493e+00 -5.68841100e-01 2.60008797e-02
2.38793984e-01 1.46207079e-01 -5.20330220e-02 1.98700145e-01
5.14057279e-01 -2.17772126e-01 5.21402478e-01 -6.25126660e-01
6.92486018e-02 5.08655154e-04 -6.74591064e-01 8.48383188e-01
1.02867104e-01 1.02850688e+00 -1.00251174e+00 -5.02146304e-01
7.42229581e-01 6.42901123e-01 -2.89310008e-01 7.86034316e-02
-1.20656453e-01 4.49488968e-01 -5.43135941e-01 9.32121634e-01
9.65868711e-01 5.05212694e-02 -2.91885257e-01 -3.27487029e-02
-7.93906868e-01 1.42443657e-01 -8.41430962e-01 1.34504724e+00
-1.25051677e-01 1.14782476e+00 7.31046945e-02 -3.50536138e-01
9.95821655e-01 1.13450676e-01 7.64565408e-01 -5.58356941e-01
-1.02012113e-01 6.70111030e-02 -6.52157843e-01 -5.59941947e-01
8.40762973e-01 5.16110003e-01 5.67559004e-02 3.81298512e-02
-4.74862188e-01 1.19555168e-01 1.42171100e-01 2.08763078e-01
1.26202226e+00 4.52201992e-01 -9.14561599e-02 2.12661937e-01
3.58139068e-01 5.95971406e-01 6.51387393e-01 7.83030868e-01
-6.67325735e-01 1.16960041e-01 -2.39031851e-01 -7.94754028e-01
-1.35871542e+00 -1.28915036e+00 -1.27477646e-01 5.48666060e-01
5.65523624e-01 6.22381642e-02 -2.07573861e-01 -6.09875858e-01
6.63893342e-01 6.38356686e-01 -3.58531982e-01 2.62801468e-01
-8.81349564e-01 -6.03202544e-02 4.09489512e-01 7.03881741e-01
2.89780170e-01 -2.41287544e-01 -8.74278009e-01 3.39150250e-01
3.09292413e-02 -1.63047338e+00 -6.12874985e-01 -3.47530395e-01
-8.20487320e-01 -1.00073147e+00 -3.16044211e-01 -5.43226004e-01
5.34658313e-01 1.17313898e+00 6.21950269e-01 -1.48312122e-01
5.26309051e-02 6.11374140e-01 7.70901665e-02 -4.06653315e-01
-4.32928473e-01 -3.33697975e-01 5.51988184e-01 -2.74187744e-01
7.42956698e-01 -2.99165756e-01 -7.24071860e-01 4.57507670e-01
2.37420171e-01 2.87487153e-02 2.58533895e-01 -4.96869162e-02
5.45402288e-01 -2.65747815e-01 3.29273969e-01 -2.81400484e-04
-3.32612872e-01 -7.62452006e-01 -1.38691115e+00 -2.27743074e-01
-4.53844488e-01 -3.96951228e-01 7.27009103e-02 -3.77744645e-01
-8.13489497e-01 6.48205996e-01 2.30350032e-01 -1.09964645e+00
-3.23316306e-01 -1.63535029e-01 1.54244617e-01 -8.65407437e-02
2.65281320e-01 1.73126221e-01 -4.77954224e-02 -2.92789280e-01
3.14995736e-01 4.61098671e-01 7.65582860e-01 8.66812319e-02
1.30159283e+00 1.20429552e+00 1.26640499e-01 -6.80861712e-01
-3.99019510e-01 -6.61933720e-01 -9.04631615e-01 -7.06009090e-01
9.49226499e-01 -1.79775894e+00 -1.02493286e+00 2.28537261e-01
-1.50799930e+00 -3.21350276e-01 2.72804171e-01 7.60189772e-01
-4.37279254e-01 1.25797555e-01 -4.21808898e-01 -1.24798357e+00
1.74071658e-02 -1.08011794e+00 1.42690277e+00 1.69144258e-01
7.67524960e-03 -8.96994948e-01 1.25864908e-01 8.42599645e-02
1.01606414e-01 3.06141317e-01 1.04942359e-01 -2.82160252e-01
-1.51962340e+00 -5.58275700e-01 -4.07703608e-01 -4.65979785e-01
-1.98536545e-01 5.77582121e-02 -9.53598380e-01 -5.19914746e-01
-4.67269748e-01 4.32940573e-01 9.82675552e-01 7.55165398e-01
4.19172555e-01 -2.00288802e-01 -1.32109106e+00 7.62795448e-01
1.12519121e+00 3.59355479e-01 6.79612011e-02 4.23455983e-01
7.85219431e-01 5.10844052e-01 1.14705968e+00 3.71828496e-01
1.07702887e+00 8.51750433e-01 8.08875084e-01 1.94762260e-01
-1.76390901e-01 -5.25744259e-01 6.34783804e-01 1.51506886e-01
2.86061674e-01 -7.31773153e-02 -9.51854289e-01 9.60806370e-01
-1.96276855e+00 -1.37651002e+00 -3.79270613e-01 2.15000868e+00
-7.91911110e-02 2.75980651e-01 3.08662832e-01 -6.09588861e-01
1.02677226e+00 9.08180177e-02 -8.42106879e-01 2.85142124e-01
1.09479979e-01 -9.40387249e-01 1.32175994e+00 8.64511847e-01
-1.40014029e+00 1.12770140e+00 6.46512461e+00 1.69096664e-01
-8.23261380e-01 1.37109995e-01 1.69078633e-01 -3.91146123e-01
-1.29562005e-01 1.10821374e-01 -1.74412501e+00 4.45177913e-01
8.31368446e-01 -3.31216544e-01 -2.88937520e-02 9.62277472e-01
7.92979121e-01 -1.47060931e-01 -1.23901427e+00 8.82754624e-01
-1.77894875e-01 -1.88332391e+00 -3.11879903e-01 4.30265754e-01
7.08782196e-01 8.20815861e-01 6.32397309e-02 1.82843044e-01
5.45305252e-01 -6.32500589e-01 1.11325324e+00 5.01167476e-01
6.17577553e-01 -6.72227263e-01 4.03774977e-01 6.53902948e-01
-1.75534713e+00 -2.13153452e-01 -3.16084087e-01 -4.94236425e-02
7.41984725e-01 1.45812213e-01 -1.40330112e+00 9.58040804e-02
7.34159470e-01 1.00982738e+00 -2.88807452e-01 1.27808869e+00
1.64102718e-01 1.13661774e-01 -5.22938669e-01 2.27911890e-01
3.94822091e-01 5.56892008e-02 1.09835768e+00 1.12258494e+00
6.24968350e-01 -3.51083092e-03 5.44709265e-01 8.91311109e-01
2.63332456e-01 -7.00750291e-01 -1.32028425e+00 4.76181895e-01
1.22626138e+00 1.12835574e+00 -3.94068539e-01 -4.14298207e-01
-5.28497279e-01 1.83924913e-01 1.52279168e-01 5.18395543e-01
-1.18832278e+00 3.43017161e-01 1.20123041e+00 4.25184369e-01
8.21942031e-01 -8.80772948e-01 1.16042286e-01 -8.04904282e-01
-3.26465182e-02 1.68224320e-01 4.18520207e-03 -8.54717612e-01
-8.18498015e-01 3.77950102e-01 2.01051742e-01 -1.59917748e+00
-5.74099600e-01 -5.23888707e-01 -5.73737144e-01 5.59359968e-01
-1.80401659e+00 -1.28403807e+00 -3.73506725e-01 3.49629521e-01
6.03733718e-01 -1.74949527e-01 4.58140559e-02 4.92926747e-01
-3.24066848e-01 6.27373457e-02 1.06249601e-01 3.36357728e-02
4.17428881e-01 -6.96367741e-01 1.14555395e+00 1.05017769e+00
-7.20472028e-03 5.75383902e-02 7.92102158e-01 -1.19719958e+00
-1.76083052e+00 -1.72925889e+00 7.85798490e-01 -9.83280241e-01
6.51416540e-01 -3.08083057e-01 -6.69452786e-01 1.12679029e+00
-3.20739239e-01 1.39345840e-01 -9.41808745e-02 -5.71939409e-01
-1.63322136e-01 -2.73149163e-01 -1.07366598e+00 5.13573110e-01
1.39653969e+00 -3.64902079e-01 -1.63927123e-01 2.56679267e-01
7.69109726e-01 -7.34698236e-01 -4.58618701e-01 5.74754953e-01
7.54078329e-01 -7.65261114e-01 1.34399760e+00 -2.63312399e-01
-5.18911362e-01 -7.34072626e-01 -3.28350097e-01 -7.37501502e-01
-2.17653677e-01 -5.51964700e-01 -5.11082888e-01 8.74685585e-01
1.89834923e-01 -3.66560042e-01 1.21663845e+00 5.96946061e-01
-2.97802448e-01 -2.23050699e-01 -1.31657994e+00 -1.03307152e+00
-2.48228580e-01 -8.47672045e-01 6.84377730e-01 3.63859296e-01
-5.25558650e-01 -1.09234534e-01 -4.04148430e-01 9.56562400e-01
1.20400774e+00 2.12331116e-01 1.33770788e+00 -1.07629383e+00
6.41409874e-01 -1.85326904e-01 -7.10787416e-01 -1.58465326e+00
4.85311598e-01 -7.84658849e-01 2.08486065e-01 -1.42771876e+00
3.48770022e-02 -4.43032742e-01 3.82855475e-01 1.41654477e-01
3.08214128e-01 5.58116660e-02 1.61528721e-01 3.65290672e-01
-7.99667418e-01 4.36435133e-01 6.50836408e-01 -1.83166042e-01
-1.96777225e-01 2.30457738e-01 2.63443798e-01 9.22249138e-01
5.66099703e-01 -5.27985573e-01 -1.32538304e-01 -8.43564212e-01
-1.89389452e-01 3.29166383e-01 9.56435382e-01 -1.04726744e+00
7.29951262e-01 -2.59761453e-01 6.78545177e-01 -1.95419419e+00
8.19291651e-01 -9.11737919e-01 2.20983744e-01 5.14877498e-01
7.56899565e-02 4.27064687e-01 5.36634088e-01 8.44867587e-01
4.16361064e-01 2.58694530e-01 4.31675434e-01 -8.83931071e-02
-1.35146523e+00 9.32043135e-01 -7.54399657e-01 -2.52201796e-01
1.32149351e+00 -6.82953954e-01 -3.53020847e-01 -4.42962259e-01
-4.16203290e-01 8.70958984e-01 7.25907385e-01 7.20056117e-01
8.81135762e-01 -1.51945531e+00 -7.35706747e-01 1.39715672e-01
-2.78914440e-02 -8.92088413e-02 2.28432283e-01 7.15496004e-01
-3.95995617e-01 8.91266227e-01 4.99625318e-02 -1.30574584e+00
-1.49908185e+00 5.62908947e-01 3.61145943e-01 3.37978303e-01
-7.99039662e-01 5.37195683e-01 2.94388235e-01 -3.35032493e-01
3.94592851e-01 -2.40406096e-01 4.18807380e-02 -1.25010058e-01
7.54563093e-01 6.30549729e-01 -3.11924428e-01 -1.25019145e+00
-7.24563897e-01 7.21461236e-01 3.03666014e-02 -1.97508678e-01
1.15648162e+00 -8.40566874e-01 5.55808425e-01 2.58920461e-01
1.04277825e+00 -1.85069755e-01 -2.27852774e+00 -2.01089546e-01
2.20290676e-01 -7.55694449e-01 1.16149247e-01 -5.40866740e-02
-8.59001517e-01 6.20871544e-01 8.00957561e-01 -1.81417510e-01
1.40793428e-01 1.75240919e-01 8.72585714e-01 5.22510111e-01
7.07054853e-01 -7.11299837e-01 -3.22870046e-01 6.44334376e-01
3.84121895e-01 -1.38935208e+00 3.27675268e-02 -1.87772840e-01
-1.81700960e-01 9.71608222e-01 8.53721023e-01 3.84118333e-02
3.61234874e-01 4.18611705e-01 6.49057925e-02 -1.49197370e-01
-9.94670868e-01 -4.33432370e-01 8.21401626e-02 8.54961097e-01
-4.35941428e-01 -1.33782268e-01 6.43796146e-01 -1.84207976e-01
1.21943496e-01 -2.89224386e-01 4.40249801e-01 6.96583688e-01
-9.05413151e-01 -4.16035414e-01 -6.62180483e-01 7.55917141e-03
1.00075275e-01 2.54167229e-01 8.92634913e-02 9.21626151e-01
1.40377134e-01 1.02258182e+00 6.61700547e-01 -4.30049062e-01
1.81936473e-01 -1.45911738e-01 4.60830033e-01 -3.45057219e-01
2.33423308e-01 -3.19184810e-02 1.82855561e-01 -7.63184190e-01
-2.67860204e-01 -1.14979684e+00 -1.37875104e+00 -6.41573131e-01
-1.97375104e-01 -3.47172320e-01 9.38126862e-01 9.89112198e-01
7.29797542e-01 -7.92690516e-02 8.04273963e-01 -1.52147079e+00
-2.82572240e-01 -4.56126720e-01 -7.78712407e-02 -4.46442105e-02
1.04578888e+00 -8.54531467e-01 -1.94330752e-01 -4.00583027e-03] | [6.764579772949219, -2.1914148330688477] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.