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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
e6120ad7-3811-41f6-8609-eec758086aca | a-closer-look-at-invalid-action-masking-in | 2006.14171 | null | https://arxiv.org/abs/2006.14171v3 | https://arxiv.org/pdf/2006.14171v3.pdf | A Closer Look at Invalid Action Masking in Policy Gradient Algorithms | In recent years, Deep Reinforcement Learning (DRL) algorithms have achieved state-of-the-art performance in many challenging strategy games. Because these games have complicated rules, an action sampled from the full discrete action distribution predicted by the learned policy is likely to be invalid according to the game rules (e.g., walking into a wall). The usual approach to deal with this problem in policy gradient algorithms is to "mask out" invalid actions and just sample from the set of valid actions. The implications of this process, however, remain under-investigated. In this paper, we 1) show theoretical justification for such a practice, 2) empirically demonstrate its importance as the space of invalid actions grows, and 3) provide further insights by evaluating different action masking regimes, such as removing masking after an agent has been trained using masking. The source code can be found at https://github.com/vwxyzjn/invalid-action-masking | ['Santiago Ontañón', 'Shengyi Huang'] | 2020-06-25 | null | null | null | null | ['real-time-strategy-games'] | ['playing-games'] | [ 1.52196243e-01 9.65157300e-02 -4.73944843e-01 7.08482563e-02
-7.09089696e-01 -7.04693139e-01 6.55064285e-01 -1.27574176e-01
-8.97973418e-01 1.29087758e+00 2.25313172e-01 -5.83098531e-01
-1.48086667e-01 -7.22742736e-01 -7.27004051e-01 -8.21330845e-01
-1.70203626e-01 4.07640249e-01 2.80789733e-01 -1.90982580e-01
3.94106120e-01 2.37970531e-01 -1.50179362e+00 9.86343846e-02
7.88100064e-01 5.49007893e-01 2.23402575e-01 6.85277283e-01
1.54217646e-01 8.76272440e-01 -7.25649357e-01 -1.77885577e-01
5.32337308e-01 -8.75184059e-01 -8.11313331e-01 -1.86076481e-02
-1.91470966e-01 -8.03569615e-01 -2.61878014e-01 1.35129678e+00
4.54612583e-01 3.16680431e-01 4.56324697e-01 -1.02669406e+00
-2.25867435e-01 7.47025788e-01 -5.18670380e-01 2.49227911e-01
1.67658180e-01 6.77365065e-01 1.06031156e+00 -3.65269214e-01
5.60254216e-01 1.10584140e+00 1.54462263e-01 8.76606047e-01
-1.34553576e+00 -6.97026789e-01 4.72181171e-01 2.33007833e-01
-1.01530015e+00 -2.79340565e-01 6.50013804e-01 -1.88077196e-01
7.60796726e-01 1.38739839e-01 1.01737332e+00 1.32495499e+00
2.39249527e-01 9.90161300e-01 1.50160599e+00 -3.66628170e-01
7.17068076e-01 -2.02519298e-01 -3.94461423e-01 4.36379313e-01
4.45214868e-01 6.57249510e-01 -3.50667626e-01 -2.33183652e-01
8.32341850e-01 -3.62471521e-01 -1.78593218e-01 -4.93036717e-01
-7.44931281e-01 1.13819575e+00 1.19861446e-01 4.98936884e-03
-6.45590365e-01 4.74262744e-01 3.25304270e-01 3.90383363e-01
3.98162544e-01 6.74774468e-01 -4.43535596e-01 -7.43342459e-01
-7.51405060e-01 7.98675537e-01 7.69132316e-01 3.82701457e-01
6.86375082e-01 2.18215093e-01 -2.05273673e-01 7.19542086e-01
9.12038982e-02 2.17143610e-01 4.65791374e-01 -1.40464091e+00
2.31439576e-01 1.41876489e-01 6.93786263e-01 -2.32791767e-01
-2.28244364e-01 -2.38693744e-01 -3.43635857e-01 8.46791208e-01
8.85220706e-01 -6.53869152e-01 -8.39606285e-01 1.96535397e+00
2.95481026e-01 1.23059742e-01 2.85354145e-02 1.03506029e+00
-3.73061299e-02 3.28819811e-01 1.40478417e-01 -3.27369988e-01
8.02804649e-01 -6.78387702e-01 -5.11415124e-01 -5.64702988e-01
5.65421402e-01 -4.19611156e-01 1.29717660e+00 5.49999237e-01
-1.17779744e+00 -1.15420893e-01 -8.69861662e-01 5.04291952e-01
-9.17146504e-02 -2.98674464e-01 7.94082344e-01 5.28271914e-01
-1.00341260e+00 9.62681532e-01 -1.12294233e+00 -2.32169256e-01
4.97319221e-01 3.55266869e-01 1.04607224e-01 9.96047333e-02
-1.25739384e+00 8.67921174e-01 4.43712294e-01 -8.95614401e-02
-1.26207829e+00 -1.93923861e-01 -5.71774662e-01 2.30998192e-02
1.05230224e+00 -3.63808721e-01 1.86004019e+00 -1.21543670e+00
-1.90476775e+00 4.79929298e-01 -4.12410535e-02 -7.11092234e-01
8.78089964e-01 -4.71304178e-01 1.71820864e-01 -5.10894060e-02
5.44007607e-02 5.13475001e-01 8.04344237e-01 -1.16441572e+00
-7.11303115e-01 -1.21661514e-01 4.66874927e-01 4.85829741e-01
7.25255907e-02 -3.49031426e-02 -4.51123677e-02 -5.26941001e-01
-2.56495625e-01 -1.08244920e+00 -5.11635721e-01 -2.29983017e-01
-3.17510813e-01 -1.88354567e-01 2.19282389e-01 -3.28923732e-01
1.20908737e+00 -1.79553711e+00 7.40205571e-02 6.86634928e-02
-7.33767822e-02 3.25966716e-01 -3.97518314e-02 6.50119662e-01
1.66151837e-01 1.41980648e-01 -1.03696324e-01 -5.15049547e-02
2.28147790e-01 4.10066932e-01 -3.74244601e-01 4.83579844e-01
-7.96528459e-02 8.51920545e-01 -1.19008815e+00 -1.27201200e-01
3.09392452e-01 -6.27115518e-02 -8.90220702e-01 8.35539773e-02
-6.36274338e-01 5.74378431e-01 -6.64198458e-01 2.53768355e-01
3.73743802e-01 -6.20547347e-02 5.32151937e-01 5.84023952e-01
-2.34235883e-01 5.53957880e-01 -1.22120237e+00 1.14894533e+00
-1.59999967e-01 3.49862367e-01 1.00024328e-01 -9.96390164e-01
3.97148430e-01 1.36183783e-01 3.69898081e-01 -7.05786109e-01
2.06452802e-01 1.37338355e-01 4.50776100e-01 -3.13919485e-01
3.45352650e-01 -2.01420709e-01 5.89415338e-03 5.73757529e-01
-3.08216751e-01 -1.80552557e-01 3.37780803e-01 -1.44461751e-01
1.23978353e+00 4.93199140e-01 5.20429492e-01 -2.28530169e-01
2.22210705e-01 9.55243707e-02 6.89057231e-01 1.40969777e+00
-4.82830316e-01 7.49092549e-02 9.47511673e-01 -3.30012530e-01
-8.75987530e-01 -1.03494430e+00 2.56991893e-01 1.03979087e+00
5.61833046e-02 -3.20571214e-01 -8.27113926e-01 -6.58447266e-01
1.30223006e-01 9.55495298e-01 -6.77863121e-01 -3.47082287e-01
-6.03449941e-01 -5.65515816e-01 3.10037434e-01 4.43262130e-01
5.81802905e-01 -1.49341333e+00 -1.01665568e+00 3.89857620e-01
1.24556497e-01 -6.60969436e-01 -1.24799751e-01 3.97931963e-01
-7.81074762e-01 -1.11218667e+00 -4.36128855e-01 -1.32036254e-01
3.97297472e-01 1.28385173e-02 1.00213766e+00 1.42793804e-01
1.23051563e-02 2.42959112e-01 -3.27645928e-01 -3.13749909e-01
-5.67034185e-01 -1.34378493e-01 1.79552555e-01 -3.37829173e-01
3.98367763e-01 -4.67900783e-01 -7.56332636e-01 1.74440444e-01
-7.40665257e-01 5.54169267e-02 5.60218632e-01 9.10119236e-01
4.49118018e-01 1.48030175e-02 5.78219175e-01 -8.24444711e-01
1.02546477e+00 -3.97034377e-01 -1.00995171e+00 -1.31337345e-01
-4.10174966e-01 1.95991829e-01 8.18236530e-01 -6.15952373e-01
-8.97591710e-01 -2.01291323e-01 -2.36305356e-01 -2.97406822e-01
-3.38626504e-01 3.80583942e-01 1.67275995e-01 2.92294979e-01
6.36241615e-01 3.10014188e-01 9.67595875e-02 -3.36605310e-01
2.43327186e-01 3.59423131e-01 -8.02859105e-03 -9.07500625e-01
5.47432721e-01 4.03296649e-01 -2.01058552e-01 -5.83399951e-01
-8.10405850e-01 7.26386160e-02 -1.23232409e-01 -4.73651230e-01
7.19174623e-01 -6.71098173e-01 -8.73480976e-01 4.87862736e-01
-6.20921493e-01 -1.28002238e+00 -5.46092331e-01 5.93321085e-01
-9.26432192e-01 2.32053012e-01 -6.98001564e-01 -1.08562434e+00
1.70603052e-01 -1.20355320e+00 6.21631920e-01 5.40072918e-01
-3.34446847e-01 -6.83796287e-01 1.63249359e-01 -3.97788286e-02
1.91683844e-01 -7.43589997e-02 8.16147029e-01 -4.79504913e-01
-5.07088482e-01 8.94316435e-02 3.52262855e-01 3.92860949e-01
2.49984011e-01 -2.23746747e-01 -7.11806655e-01 -5.35401642e-01
-3.48281823e-02 -4.60210383e-01 8.17292929e-01 8.11898649e-01
1.21543026e+00 -3.02122712e-01 -3.72979529e-02 3.13036263e-01
1.18207061e+00 3.53270680e-01 5.37554085e-01 6.46230996e-01
9.09850448e-02 3.74414533e-01 8.48999619e-01 8.44936252e-01
-1.29876677e-02 6.08636200e-01 6.91465795e-01 3.57839733e-01
2.41303563e-01 -5.23965120e-01 5.24021685e-01 8.31811056e-02
-5.31185567e-02 -2.20389813e-01 -6.08008265e-01 3.79761755e-01
-2.04349446e+00 -1.20298243e+00 5.20027995e-01 2.43609548e+00
9.33307707e-01 7.03116059e-01 4.27375108e-01 -2.03614347e-02
5.43903649e-01 3.59696835e-01 -1.07234788e+00 -6.61101997e-01
2.35539749e-01 2.40205899e-01 6.95983589e-01 7.77648687e-01
-9.53724265e-01 1.32580972e+00 6.12392950e+00 9.10358906e-01
-1.05570042e+00 -2.47047357e-02 7.04823911e-01 -4.42651302e-01
-8.86503980e-02 1.73421934e-01 -6.89473033e-01 6.56960011e-01
6.67354703e-01 -3.09043050e-01 9.88372445e-01 8.40733945e-01
6.98546886e-01 -6.56436622e-01 -7.74552941e-01 4.82217878e-01
-5.26916027e-01 -1.12253368e+00 -2.88812041e-01 3.86937678e-01
7.54689693e-01 1.54410928e-01 8.56989175e-02 7.31996179e-01
1.05833125e+00 -9.21438634e-01 7.78821826e-01 1.45040438e-01
5.13334692e-01 -8.85148942e-01 3.93635571e-01 7.00061083e-01
-7.09088027e-01 -3.90889227e-01 -4.53498721e-01 -6.27979636e-01
-5.35326600e-02 2.06452340e-01 -5.78061640e-01 1.72635198e-01
6.10765100e-01 3.98748577e-01 -3.25009197e-01 1.01421285e+00
-7.19246447e-01 9.32719290e-01 -3.23653162e-01 -3.83947968e-01
6.63097501e-01 -4.30845410e-01 5.61046481e-01 6.03489816e-01
1.27466142e-01 7.41427317e-02 2.48125985e-01 8.84836197e-01
1.86707124e-01 -2.49469101e-01 -6.83360398e-01 -2.16632441e-01
4.62380916e-01 7.33656585e-01 -7.97020137e-01 -1.94226980e-01
-3.87580246e-01 7.82453120e-01 5.39766848e-01 6.37278855e-01
-9.55639303e-01 -2.34044529e-02 1.14138103e+00 1.19798062e-02
4.43625301e-01 -2.76050717e-01 -1.44715324e-01 -8.93602014e-01
1.58308651e-02 -1.07055569e+00 2.43383780e-01 -5.08571148e-01
-8.98643255e-01 1.79724276e-01 1.96206972e-01 -1.12892783e+00
-4.68834549e-01 -5.23700297e-01 -6.78251922e-01 6.58774197e-01
-1.19243157e+00 -1.97542831e-01 2.61356890e-01 3.36735040e-01
7.72100985e-01 5.59554324e-02 5.33053219e-01 -1.41667455e-01
-5.25921345e-01 3.15687984e-01 4.03717697e-01 -6.78712800e-02
4.11265939e-01 -1.37589979e+00 3.85905951e-01 8.33080411e-01
-6.54060841e-02 4.02312398e-01 9.21389043e-01 -7.49765575e-01
-1.16678631e+00 -7.08385587e-01 1.53666183e-01 -2.68594891e-01
7.75569081e-01 -2.73834467e-01 -8.15402567e-01 7.51303077e-01
8.92178491e-02 -2.63623863e-01 3.32284629e-01 1.82379596e-02
9.12321657e-02 3.42447788e-01 -9.74088132e-01 1.33276534e+00
1.04325116e+00 -1.93969205e-01 -3.48775804e-01 1.93501517e-01
2.90492475e-01 -4.84029055e-01 -1.95839271e-01 2.34980792e-01
4.79506761e-01 -1.32811666e+00 7.47950315e-01 -9.17147994e-01
3.96329015e-01 -1.05424486e-01 5.28895408e-02 -1.58834839e+00
-2.50883192e-01 -9.69329059e-01 -2.27998123e-01 3.61550212e-01
2.12718263e-01 -7.92184412e-01 1.09373474e+00 3.75386447e-01
9.74083170e-02 -1.26667273e+00 -1.04513693e+00 -1.04527819e+00
4.45859820e-01 -3.72378409e-01 4.46643412e-01 5.48689902e-01
1.43355444e-01 -8.96640345e-02 -5.97253799e-01 -2.58555170e-02
4.66158867e-01 2.92840339e-02 6.60747170e-01 -7.35615671e-01
-7.39399135e-01 -7.75470316e-01 1.04045220e-01 -1.34091401e+00
2.56539226e-01 -4.35532779e-01 1.31918982e-01 -1.44246995e+00
-3.22840782e-03 -3.08233947e-01 -3.13952774e-01 4.52912152e-01
-3.87909412e-01 -1.44343466e-01 4.04683858e-01 -1.24271987e-02
-7.14825869e-01 6.79033995e-01 1.32872474e+00 2.62785286e-01
-3.66597950e-01 2.39803866e-01 -6.18239522e-01 8.36648345e-01
1.30464685e+00 -4.52025294e-01 -5.50961494e-01 -5.86692058e-02
2.08715707e-01 2.35026538e-01 2.15769649e-01 -8.07441711e-01
-2.19608694e-01 -7.85837829e-01 3.34870130e-01 -5.37052527e-02
4.06758279e-01 -5.52860141e-01 -1.19040370e-01 9.50339258e-01
-5.60141385e-01 5.77419661e-02 3.13820302e-01 4.90433246e-01
2.51129806e-01 -4.62490052e-01 8.30816865e-01 -3.74508560e-01
-6.42262340e-01 -2.55425572e-02 -9.88819778e-01 4.06195104e-01
1.04740119e+00 -2.22280890e-01 -2.88369507e-01 -7.86499321e-01
-7.13087738e-01 3.98484349e-01 6.34497881e-01 2.48203084e-01
3.48081440e-01 -1.02478516e+00 -4.18252766e-01 3.47144669e-04
-3.19175810e-01 -2.28599131e-01 9.20242444e-02 6.25916421e-01
-3.29164684e-01 2.28059039e-01 -2.53126264e-01 -8.59382823e-02
-9.50013459e-01 4.55178231e-01 6.13281786e-01 -4.71382797e-01
-6.83153272e-01 7.04868019e-01 1.46264851e-01 -1.45748690e-01
2.04454094e-01 -2.35817954e-01 5.47065400e-02 -2.42865250e-01
3.48088354e-01 2.03317240e-01 -3.96594822e-01 -3.57584772e-03
-1.97412670e-01 1.14800692e-01 -2.44173974e-01 -6.10809684e-01
1.07843423e+00 2.06618384e-01 3.77964407e-01 3.41027141e-01
5.42249799e-01 -7.46382773e-02 -1.98661113e+00 -1.25402376e-01
6.14553988e-02 -6.84442103e-01 -1.64938301e-01 -7.57942915e-01
-7.60794759e-01 5.34329653e-01 3.39338988e-01 4.62013930e-01
9.71056700e-01 -1.45391375e-01 3.69224817e-01 3.43425095e-01
5.88375688e-01 -1.37080085e+00 1.17239170e-01 6.42259777e-01
6.71355963e-01 -1.08658409e+00 5.95362969e-02 2.41975740e-01
-9.02146757e-01 7.75921762e-01 7.90483117e-01 -5.01221418e-01
3.98150444e-01 9.86172482e-02 5.95466085e-02 -5.51984273e-03
-1.02056861e+00 -5.60735643e-01 -4.35588092e-01 4.25568104e-01
1.20100506e-01 3.00518513e-01 -5.67072213e-01 1.59206465e-01
-2.28847295e-01 4.31469530e-02 7.17372835e-01 1.23165107e+00
-6.02275908e-01 -1.26780665e+00 -2.35028118e-01 7.42231905e-01
-5.74721694e-01 -4.82162237e-02 -3.79769057e-01 8.40036929e-01
-2.89047867e-01 9.20945287e-01 -1.28954649e-01 -1.52223155e-01
1.80343747e-01 8.86434317e-02 6.68794572e-01 -6.16705775e-01
-4.31983799e-01 1.56309709e-01 9.68076140e-02 -7.74953127e-01
-4.56277877e-02 -7.06283867e-01 -1.15026009e+00 -2.62863278e-01
4.21242900e-02 9.27779153e-02 2.13340372e-01 9.94367480e-01
2.32072681e-01 4.49216992e-01 3.54331583e-01 -1.02904367e+00
-1.21250594e+00 -8.01470697e-01 -5.99398851e-01 2.69356936e-01
3.91334027e-01 -9.05512035e-01 -5.54156899e-01 -5.97724736e-01] | [3.884488582611084, 1.9050101041793823] |
65e676a6-8d20-447e-a541-8bbcdaa4c4f4 | deep-pipeline-embeddings-for-automl | 2305.14009 | null | https://arxiv.org/abs/2305.14009v2 | https://arxiv.org/pdf/2305.14009v2.pdf | Deep Pipeline Embeddings for AutoML | Automated Machine Learning (AutoML) is a promising direction for democratizing AI by automatically deploying Machine Learning systems with minimal human expertise. The core technical challenge behind AutoML is optimizing the pipelines of Machine Learning systems (e.g. the choice of preprocessing, augmentations, models, optimizers, etc.). Existing Pipeline Optimization techniques fail to explore deep interactions between pipeline stages/components. As a remedy, this paper proposes a novel neural architecture that captures the deep interaction between the components of a Machine Learning pipeline. We propose embedding pipelines into a latent representation through a novel per-component encoder mechanism. To search for optimal pipelines, such pipeline embeddings are used within deep-kernel Gaussian Process surrogates inside a Bayesian Optimization setup. Furthermore, we meta-learn the parameters of the pipeline embedding network using existing evaluations of pipelines on diverse collections of related datasets (a.k.a. meta-datasets). Through extensive experiments on three large-scale meta-datasets, we demonstrate that pipeline embeddings yield state-of-the-art results in Pipeline Optimization. | ['Josif Grabocka', 'Sebastian Pineda Arango'] | 2023-05-23 | null | null | null | null | ['automatic-machine-learning-model-selection', 'automl', 'hyperparameter-optimization', 'bayesian-optimization'] | ['methodology', 'methodology', 'methodology', 'methodology'] | [-2.34097868e-01 1.99266687e-01 -2.20939741e-01 -5.52574217e-01
-8.54514599e-01 -8.67511570e-01 6.47040546e-01 1.72270805e-01
-5.06117046e-01 -2.33880833e-01 3.90514225e-01 -1.79894924e-01
-1.64552420e-01 -3.85195166e-01 -9.04256940e-01 -3.68651330e-01
-6.00220114e-02 9.04804826e-01 -6.25653863e-02 8.08223560e-02
4.28591520e-01 4.60226893e-01 -1.01723027e+00 2.92159021e-01
5.23616195e-01 6.29894435e-01 -1.14617288e-01 9.74371552e-01
-5.80559485e-02 5.56906283e-01 -3.47694218e-01 -6.99590206e-01
3.48043442e-01 2.96266936e-02 -8.76525223e-01 -2.17496231e-01
4.10791397e-01 -1.61197275e-01 -3.47274929e-01 8.26827526e-01
3.27590168e-01 -6.26296699e-02 6.75973892e-01 -1.33430600e+00
-6.84333026e-01 9.18487489e-01 -2.04870179e-01 -6.37192950e-02
-2.82932758e-01 7.22723424e-01 1.38646424e+00 -7.90245414e-01
2.78884023e-01 1.47505772e+00 6.38394833e-01 5.45835495e-01
-1.58251274e+00 -2.01070115e-01 7.30279908e-02 1.38414115e-01
-1.28170335e+00 -6.50711417e-01 6.79281175e-01 -8.33501458e-01
1.35302067e+00 -9.02628303e-02 6.05921328e-01 1.07642126e+00
3.08908671e-01 9.58620369e-01 3.13507915e-01 -2.48692170e-01
4.67942059e-01 2.77723134e-01 3.75426978e-01 1.10552382e+00
3.42255592e-01 -1.54820830e-01 -8.32947731e-01 -3.24253798e-01
3.15212369e-01 -2.73457587e-01 3.16507638e-01 -3.63154173e-01
-1.08757341e+00 7.17165828e-01 3.00469100e-01 7.10462481e-02
-2.44933948e-01 6.22046471e-01 1.81239113e-01 -5.17182983e-02
3.73364687e-01 1.29093039e+00 -8.76901865e-01 -3.32763821e-01
-1.05878496e+00 2.91683376e-01 1.12149239e+00 9.75808918e-01
1.01249480e+00 -2.61960477e-01 -5.37875652e-01 3.75190347e-01
7.35850513e-01 -4.58090194e-02 3.54273617e-01 -1.06824577e+00
5.06815910e-01 9.58003521e-01 3.37370038e-02 -6.51859581e-01
-4.42700028e-01 -3.68430644e-01 -2.45946914e-01 -1.96455061e-01
3.15178692e-01 -5.83247095e-02 -6.93829954e-01 1.47025824e+00
2.88906693e-01 2.39826918e-01 -1.25948638e-01 4.25246119e-01
2.29356855e-01 5.90731382e-01 3.03012371e-01 4.49542344e-01
1.52206349e+00 -1.65646791e+00 -4.96591091e-01 -6.29083753e-01
8.67303014e-01 -7.20030487e-01 1.31818461e+00 4.21934992e-01
-1.19432127e+00 -4.39667732e-01 -1.01648355e+00 -4.83335406e-01
-3.20673674e-01 6.05569065e-01 7.89186180e-01 5.58763981e-01
-1.07381380e+00 9.44714844e-01 -1.49199462e+00 1.18170232e-01
7.05078006e-01 4.51206714e-01 -1.72390282e-01 2.24545226e-01
-5.64204097e-01 9.80218172e-01 4.67756331e-01 2.27241322e-01
-1.30850565e+00 -1.13980126e+00 -9.02655482e-01 2.89914548e-01
1.37658894e-01 -8.86109352e-01 1.44234598e+00 -4.38896656e-01
-1.70649445e+00 6.73388481e-01 -2.68183649e-01 -6.75653458e-01
1.44782320e-01 -9.23210680e-01 1.22512951e-01 -3.47527750e-02
-2.60003120e-01 6.47125125e-01 1.04619336e+00 -4.93023813e-01
-4.34169829e-01 -4.00081962e-01 -1.14166729e-01 -1.69719115e-01
-7.13753283e-01 2.82061845e-01 -7.50247955e-01 -2.07238615e-01
-4.63547558e-01 -9.88637984e-01 -5.65421045e-01 -3.47808897e-02
-6.48045003e-01 -5.04297376e-01 3.48940283e-01 -6.85891747e-01
1.35212266e+00 -2.10067296e+00 6.81919277e-01 -2.18219116e-01
3.56880754e-01 2.19506249e-01 -3.44082415e-01 2.52426744e-01
2.26077542e-01 5.25594711e-01 -2.06684753e-01 -1.09697843e+00
5.14717579e-01 1.56174779e-01 -9.09863859e-02 5.57058394e-01
1.00745118e+00 1.20198739e+00 -8.76312912e-01 -4.78710145e-01
-2.34944075e-02 6.17306232e-01 -1.01076329e+00 7.53664315e-01
-6.73655748e-01 1.44135728e-01 -4.92045283e-01 8.23570311e-01
2.53187884e-02 -6.04279101e-01 -2.22918108e-01 -6.42480314e-01
-2.56594062e-01 5.21208405e-01 -7.84480810e-01 2.19337726e+00
-5.04838169e-01 7.17318177e-01 -1.47450417e-01 -6.62915528e-01
6.70875609e-01 2.00883433e-01 3.47352475e-01 1.76259309e-01
9.76892039e-02 -3.94318439e-03 1.49706751e-01 -4.12332267e-01
5.62364459e-01 6.21075034e-01 -1.97172865e-01 4.01798069e-01
6.31625772e-01 -2.47074917e-01 2.59165943e-01 9.65506360e-02
1.58204961e+00 5.42158306e-01 1.47441149e-01 -1.75504193e-01
1.11608319e-01 2.86289573e-01 3.82219702e-01 7.21982241e-01
-1.18836209e-01 3.74788404e-01 7.23403037e-01 -5.74035406e-01
-1.34927416e+00 -7.20998347e-01 1.35329992e-01 1.23931277e+00
-5.93568563e-01 -9.79591250e-01 -9.80589867e-01 -5.66152096e-01
-1.12245746e-01 8.51364613e-01 -7.13814795e-01 -4.90290225e-01
-6.03454471e-01 -1.10752392e+00 8.04706156e-01 3.79463583e-01
-1.32324114e-01 -6.79245055e-01 -7.49665678e-01 3.03042263e-01
3.50367546e-01 -1.30884826e+00 -4.88676101e-01 2.73489416e-01
-8.07470441e-01 -1.01679015e+00 4.06163791e-03 -2.91098118e-01
6.09957695e-01 -3.01048487e-01 1.47948468e+00 -2.97760576e-01
-5.72036445e-01 3.24366838e-01 -3.61716449e-02 -2.47915179e-01
-4.49314147e-01 6.12225115e-01 2.02045087e-02 -9.01440829e-02
4.23829943e-01 -4.09371018e-01 -6.24600887e-01 -1.62611485e-01
-6.41390145e-01 2.05221683e-01 8.02242041e-01 6.87465727e-01
4.45177287e-01 -5.86040877e-02 -2.32862204e-01 -7.60471404e-01
5.55126429e-01 -5.07321477e-01 -1.14182413e+00 5.12777090e-01
-8.33186030e-01 7.63010561e-01 5.40905893e-01 -4.59599733e-01
-8.00338089e-01 3.36360991e-01 -1.03128338e-02 -5.51394820e-01
-3.97994462e-03 4.36195552e-01 -3.82471502e-01 2.19899848e-01
5.59732258e-01 -3.48932773e-01 -2.44730875e-01 -8.64273727e-01
8.89569938e-01 3.14818084e-01 4.32779074e-01 -8.08030665e-01
9.59712386e-01 3.08224857e-01 3.75503376e-02 -5.55734694e-01
-1.05884695e+00 -3.74790162e-01 -9.15424645e-01 3.74438584e-01
1.10961163e+00 -9.53278840e-01 -8.26532304e-01 1.12270921e-01
-1.57592118e+00 -5.81243694e-01 -1.67400792e-01 2.97838748e-01
-4.80128467e-01 -1.98703691e-01 -5.62388301e-01 -6.30209863e-01
-5.90850294e-01 -1.70837712e+00 1.27965021e+00 4.42753613e-01
-5.08486509e-01 -1.09059834e+00 3.83008599e-01 3.52145046e-01
4.52324271e-01 -1.52986780e-01 9.08179343e-01 -9.34746683e-01
-8.30327213e-01 -3.04775089e-01 5.34705929e-02 5.03945827e-01
-1.67943865e-01 6.86504900e-01 -1.11550105e+00 7.59229586e-02
-1.90080509e-01 -5.34528606e-02 8.06269228e-01 2.32825488e-01
1.41189003e+00 -6.92358136e-01 -3.34100306e-01 1.26321232e+00
1.22318077e+00 -3.17035824e-01 2.31318280e-01 4.82685596e-01
1.18920064e+00 6.50354266e-01 1.95039913e-01 3.68648082e-01
3.33758622e-01 6.17887437e-01 3.49328309e-01 3.49424809e-01
2.22117454e-01 -4.62996274e-01 8.04624557e-01 7.33462870e-01
4.71257448e-01 -8.01167712e-02 -1.19109845e+00 5.99538267e-01
-1.83243108e+00 -4.05501544e-01 9.50405225e-02 1.82526445e+00
1.14795232e+00 1.38619632e-01 -1.48389533e-01 -3.92291069e-01
1.07948810e-01 4.29514386e-02 -7.51770914e-01 -3.85370791e-01
2.82357395e-01 4.81378704e-01 6.91217721e-01 5.16306043e-01
-1.38353717e+00 1.32724297e+00 6.22598934e+00 5.79113781e-01
-6.88614190e-01 1.81834430e-01 5.81991196e-01 -3.58037233e-01
-1.94536820e-01 5.09491622e-01 -1.47469354e+00 3.84304345e-01
1.66939425e+00 -6.00557365e-02 6.52529299e-01 1.10261655e+00
1.44731060e-01 3.29772413e-01 -1.87143445e+00 6.65180624e-01
-1.61900312e-01 -1.66579223e+00 -1.32853627e-01 -7.99589306e-02
4.28544700e-01 2.99119651e-01 8.20215121e-02 2.80122161e-01
8.86065364e-01 -1.31208944e+00 7.06451714e-01 5.97075164e-01
1.73155099e-01 -4.21689630e-01 2.56721884e-01 1.47759944e-01
-6.96937621e-01 -1.26481459e-01 -2.53477782e-01 1.61244437e-01
2.22608387e-01 6.65616751e-01 -1.15728021e+00 1.13655619e-01
6.59749925e-01 7.55531907e-01 -1.12359405e+00 8.27454746e-01
-6.19449556e-01 9.11440194e-01 -3.23126584e-01 5.45598119e-02
2.52973348e-01 -5.62891066e-02 4.87030447e-01 1.44585681e+00
3.72930150e-03 -6.32323861e-01 -1.40157238e-01 1.47414398e+00
-1.79885536e-01 -3.03210109e-01 -2.28358194e-01 -8.67695689e-01
5.80578089e-01 1.70188844e+00 -2.38932386e-01 -6.64595366e-02
-1.55008778e-01 8.30835342e-01 5.93164265e-01 1.32714331e-01
-7.64068186e-01 -1.11900277e-01 1.01985919e+00 -2.86912709e-01
5.83951101e-02 -5.44445992e-01 -5.39182663e-01 -1.18566716e+00
-1.53283939e-01 -8.35547388e-01 4.40260954e-02 -5.25127053e-01
-1.25908053e+00 4.45153505e-01 6.92210793e-02 -5.73835969e-01
-2.99879134e-01 -8.83677185e-01 -6.35698855e-01 1.03443313e+00
-1.29586744e+00 -1.43268323e+00 5.33272028e-02 -1.24891043e-01
4.76721257e-01 -4.59992468e-01 8.42701197e-01 9.64267179e-02
-1.18395174e+00 5.44111609e-01 -1.60573930e-01 1.48920342e-01
7.09420741e-01 -1.44003570e+00 1.11907899e+00 1.07320094e+00
3.49221766e-01 9.78045464e-01 7.71453977e-01 -3.57002586e-01
-2.10501790e+00 -1.20446551e+00 6.93021297e-01 -1.00801349e+00
1.23903000e+00 -6.01720035e-01 -7.42967725e-01 7.85951912e-01
3.08537036e-01 -1.00620046e-01 8.66739094e-01 3.35467488e-01
-5.27439296e-01 -4.57268842e-02 -6.24627650e-01 6.12341464e-01
6.21874154e-01 -6.02123916e-01 -3.96487802e-01 4.10038918e-01
1.26753616e+00 -1.42719880e-01 -1.11806762e+00 1.42957807e-01
6.66669250e-01 -3.71320844e-01 1.09804523e+00 -1.11361408e+00
8.62233579e-01 -3.23872268e-01 -4.87547331e-02 -1.29793274e+00
-2.36521110e-01 -1.04128337e+00 -8.71639073e-01 1.39505637e+00
5.49303591e-01 -8.32088068e-02 7.48422146e-01 1.09267294e+00
-3.01798016e-01 -9.04260933e-01 -4.27134335e-01 -3.92531574e-01
1.99969746e-02 -4.70292389e-01 5.52179039e-01 4.82722044e-01
-3.70735586e-01 4.93505180e-01 -1.83821410e-01 5.02029061e-01
8.73972833e-01 -4.01039481e-01 1.03105938e+00 -1.04066014e+00
-8.02890718e-01 -7.47470915e-01 -1.33864865e-01 -7.42619216e-01
4.40542489e-01 -8.32015097e-01 9.07324255e-02 -1.49860954e+00
8.61555636e-02 -4.76536483e-01 -1.71954080e-01 7.14283884e-01
-3.50033283e-01 -2.94313043e-01 -3.60161960e-02 3.23170006e-01
-5.89299619e-01 6.10696852e-01 6.11967444e-01 -2.71533102e-01
-3.19986910e-01 -1.39297947e-01 -8.61375511e-01 8.74861360e-01
9.45771098e-01 -7.60435462e-01 -1.13686569e-01 -1.09113181e+00
5.91951549e-01 -5.40241301e-01 2.81828195e-01 -7.01413631e-01
6.59082949e-01 -1.20090090e-01 5.34243621e-02 -4.47941810e-01
4.04545903e-01 -5.61466038e-01 1.30569473e-01 2.17764899e-01
-6.82743967e-01 3.42840642e-01 2.59796739e-01 3.62721443e-01
9.38393250e-02 -5.03570795e-01 5.31423330e-01 -9.30472091e-02
-4.78760034e-01 3.87706846e-01 -4.01805378e-02 4.35798205e-02
5.98829985e-01 4.92599845e-01 -4.28821117e-01 4.99061167e-01
-3.65885824e-01 4.94005382e-01 5.24746120e-01 5.49203813e-01
2.70418584e-01 -8.35768938e-01 -6.07536376e-01 4.00972851e-02
1.41891673e-01 3.67567092e-01 -1.95168972e-01 6.82266295e-01
-6.93924785e-01 3.50549638e-01 1.07623570e-01 -4.50132221e-01
-8.34615469e-01 2.62065947e-01 3.69554758e-01 -7.47909188e-01
-1.46320075e-01 1.26153791e+00 -1.91878870e-01 -3.62606972e-01
2.96368539e-01 -3.95297885e-01 1.71916619e-01 -7.25314170e-02
5.98289013e-01 1.98906854e-01 7.00244978e-02 -1.79415315e-01
-3.85396034e-01 1.82105839e-01 -2.52576500e-01 -2.96107978e-01
1.67801762e+00 3.99246961e-01 -4.83627528e-01 6.48567438e-01
1.36227000e+00 -2.43929341e-01 -1.47087026e+00 -1.53109923e-01
5.47891378e-01 -1.55415952e-01 3.25796545e-01 -4.91195947e-01
-7.29074955e-01 1.27301168e+00 3.27399135e-01 -3.47812802e-01
7.04213858e-01 1.52065054e-01 3.27343047e-01 6.85216486e-01
-1.44884005e-01 -1.02968216e+00 2.53283799e-01 5.90553880e-01
5.25178611e-01 -1.12750936e+00 1.53758422e-01 1.27619222e-01
-3.69445741e-01 1.27273631e+00 5.53157210e-01 -1.92945138e-01
8.44436169e-01 4.97589558e-01 -4.46368635e-01 -2.99095184e-01
-1.31184566e+00 1.01469472e-01 4.96444911e-01 1.95348665e-01
5.46191394e-01 9.47501324e-03 3.51864517e-01 9.64150190e-01
-2.52420604e-01 -2.48809159e-01 2.40511224e-01 6.32786393e-01
-1.61558241e-01 -1.31667185e+00 -7.06070587e-02 2.58068472e-01
-5.20831406e-01 -3.68359983e-01 -1.84826091e-01 3.00443649e-01
-7.20603690e-02 6.69324458e-01 1.44299411e-03 -5.33622205e-01
2.36931369e-01 2.46178627e-01 5.05044103e-01 -8.47853422e-01
-9.33461666e-01 -8.51389095e-02 -1.59034654e-01 -8.01798940e-01
-7.44584724e-02 -5.75245678e-01 -7.72980690e-01 8.85803252e-02
-1.14872783e-01 3.80438543e-03 1.15042782e+00 9.00397122e-01
7.39170492e-01 7.58833945e-01 1.64746657e-01 -1.02303302e+00
-8.97781432e-01 -9.42411423e-01 1.05159670e-01 -5.71024120e-02
7.06825852e-02 -4.39347714e-01 -3.62869263e-01 3.66861254e-01] | [8.548941612243652, 4.002139568328857] |
cee8799c-ef4d-4202-be34-9009cb99a871 | ratt-leveraging-unlabeled-data-to-guarantee | 2105.00303 | null | https://arxiv.org/abs/2105.00303v2 | https://arxiv.org/pdf/2105.00303v2.pdf | RATT: Leveraging Unlabeled Data to Guarantee Generalization | To assess generalization, machine learning scientists typically either (i) bound the generalization gap and then (after training) plug in the empirical risk to obtain a bound on the true risk; or (ii) validate empirically on holdout data. However, (i) typically yields vacuous guarantees for overparameterized models. Furthermore, (ii) shrinks the training set and its guarantee erodes with each re-use of the holdout set. In this paper, we introduce a method that leverages unlabeled data to produce generalization bounds. After augmenting our (labeled) training set with randomly labeled fresh examples, we train in the standard fashion. Whenever classifiers achieve low error on clean data and high error on noisy data, our bound provides a tight upper bound on the true risk. We prove that our bound is valid for 0-1 empirical risk minimization and with linear classifiers trained by gradient descent. Our approach is especially useful in conjunction with deep learning due to the early learning phenomenon whereby networks fit true labels before noisy labels but requires one intuitive assumption. Empirically, on canonical computer vision and NLP tasks, our bound provides non-vacuous generalization guarantees that track actual performance closely. This work provides practitioners with an option for certifying the generalization of deep nets even when unseen labeled data is unavailable and provides theoretical insights into the relationship between random label noise and generalization. | ['Zachary C. Lipton', 'J. Zico Kolter', 'Sivaraman Balakrishnan', 'Saurabh Garg'] | 2021-05-01 | null | null | null | null | ['holdout-set'] | ['computer-vision'] | [ 1.90514892e-01 3.59688729e-01 -2.82383263e-01 -4.61536169e-01
-1.23129761e+00 -9.68749106e-01 2.83547014e-01 2.34226346e-01
-7.07253039e-01 8.19716990e-01 -3.51818085e-01 -6.93761051e-01
-1.88165590e-01 -6.71587110e-01 -1.02969742e+00 -1.01287484e+00
-2.32609078e-01 4.15173292e-01 -1.62434742e-01 3.24768215e-01
-1.31779788e-02 5.04721701e-01 -1.21416831e+00 -1.91524640e-01
8.44588280e-01 1.22437871e+00 -3.20334584e-01 5.35327256e-01
1.87636957e-01 4.87556934e-01 -2.82425135e-01 -7.71085978e-01
6.94364250e-01 -2.47137070e-01 -8.23288381e-01 7.38876015e-02
7.48815596e-01 -3.23452085e-01 1.53939798e-02 1.50164509e+00
2.82945484e-01 1.54847145e-01 6.85048699e-01 -1.46295810e+00
-6.88548326e-01 1.04565835e+00 -6.16865814e-01 2.30890155e-01
-3.45442444e-01 1.32473230e-01 1.21291018e+00 -6.16767168e-01
3.31033617e-01 8.11622620e-01 1.08228099e+00 8.79767835e-01
-1.51407552e+00 -8.56731713e-01 3.58477354e-01 -2.57439166e-01
-1.23884249e+00 -3.84025663e-01 2.98680276e-01 -4.44929689e-01
5.23004174e-01 1.32021844e-01 2.54867792e-01 1.29523242e+00
-1.99321315e-01 6.50225043e-01 9.77621198e-01 -3.68288636e-01
6.97259426e-01 3.20638090e-01 5.99496543e-01 5.39460480e-01
4.14941341e-01 5.44045210e-01 -1.00197591e-01 -2.82864064e-01
3.56780142e-01 -3.60353440e-02 -2.74627417e-01 -2.82955021e-01
-6.82906687e-01 8.49651754e-01 4.54359055e-01 2.18805838e-02
-1.05114400e-01 4.16096240e-01 5.15400410e-01 5.32504439e-01
4.97631997e-01 4.47657794e-01 -6.17919207e-01 4.61458862e-02
-8.92295063e-01 1.71158433e-01 9.21281934e-01 8.46466362e-01
6.46623135e-01 -1.11361938e-02 3.62101823e-01 7.62846351e-01
-3.29501592e-02 3.50292355e-01 2.97613919e-01 -1.11767697e+00
3.02500874e-01 9.48377103e-02 2.64916569e-01 -5.11030912e-01
-4.07249719e-01 -8.14022124e-01 -7.91297376e-01 3.83807212e-01
9.17316020e-01 -4.03664559e-01 -9.66660142e-01 2.32697868e+00
1.48366436e-01 -7.09942030e-03 1.30484685e-01 7.41090059e-01
1.55199068e-02 2.78784722e-01 1.79409787e-01 -2.46053323e-01
8.87646139e-01 -7.00804114e-01 -3.07686538e-01 -2.07561329e-01
1.08865917e+00 -1.41812876e-01 1.31637752e+00 7.52479017e-01
-8.89496803e-01 -1.59416497e-01 -1.07791698e+00 3.30231972e-02
-1.37796819e-01 -2.91496366e-01 6.19586825e-01 1.09660017e+00
-1.00722957e+00 8.86553645e-01 -1.06671596e+00 -7.87441730e-02
6.58649445e-01 4.55061764e-01 -4.17116374e-01 -4.90701874e-04
-8.92954528e-01 7.03058958e-01 4.66692448e-01 2.23424375e-01
-1.09290147e+00 -8.79637361e-01 -7.05822110e-01 1.35147989e-01
5.24682701e-01 -4.53491718e-01 1.40068007e+00 -1.16923213e+00
-1.06122613e+00 8.01867902e-01 2.76622653e-01 -8.40711832e-01
8.44762623e-01 -2.11237296e-01 7.18496367e-03 -1.64688870e-01
-3.57881635e-02 4.47125912e-01 6.11963809e-01 -1.36482143e+00
-7.31313467e-01 -6.11817181e-01 1.18161537e-01 -8.37941170e-02
-5.06165087e-01 -3.33999485e-01 1.47719130e-01 -3.60749453e-01
1.25205636e-01 -8.94942462e-01 -3.20530295e-01 8.97014663e-02
-4.75131720e-01 -2.22395435e-01 3.93921107e-01 -3.27655911e-01
8.69395196e-01 -2.38677168e+00 -3.37631226e-01 4.24626231e-01
4.23555791e-01 3.50785889e-02 -6.72500506e-02 1.08232811e-01
-2.28506386e-01 5.69999158e-01 -4.75736201e-01 -4.51553345e-01
2.96971202e-01 3.07260126e-01 -5.83301663e-01 7.95459747e-01
9.42374244e-02 7.16350198e-01 -8.74940395e-01 -1.61035761e-01
-1.86746195e-01 5.19480146e-02 -7.44053066e-01 -2.20115632e-01
-1.20150506e-01 1.72072500e-01 -1.65391937e-01 3.02658737e-01
6.99448109e-01 -4.20128942e-01 9.76398960e-02 -3.98296081e-02
3.43327641e-01 5.91364577e-02 -1.22997749e+00 1.22324979e+00
-5.89707434e-01 4.76234913e-01 1.81849405e-01 -1.20228350e+00
5.40787935e-01 6.77658021e-02 1.82152793e-01 -5.12789786e-02
3.31259131e-01 4.30183947e-01 -7.54124895e-02 -3.07528764e-01
1.69238914e-02 -6.76249623e-01 5.31881526e-02 4.19517159e-01
8.83993953e-02 1.68202356e-01 -1.12296239e-01 7.96451047e-02
1.20510340e+00 -1.22767389e-01 -9.04452614e-03 -4.39917505e-01
5.55877052e-02 -2.77043641e-01 6.01818800e-01 1.07375419e+00
-2.77500182e-01 4.01281029e-01 8.18468034e-01 -3.36596966e-01
-1.15976965e+00 -1.33694232e+00 -5.80919981e-01 1.12845373e+00
-5.32701872e-02 8.03545192e-02 -8.13249648e-01 -1.05280805e+00
1.85676008e-01 8.24899912e-01 -1.00307965e+00 -2.90247262e-01
-2.33642027e-01 -8.77012730e-01 7.10644186e-01 8.67152333e-01
2.34572530e-01 -5.63071907e-01 -1.76216513e-01 -1.20586559e-01
3.06597501e-01 -9.01161969e-01 -3.53336006e-01 8.19101036e-01
-8.19678962e-01 -1.06177866e+00 -3.78692865e-01 -6.97097063e-01
8.27210486e-01 -1.74469084e-01 8.37027788e-01 2.68491715e-01
1.92096364e-02 2.77253836e-01 -1.22096337e-01 -2.65550256e-01
-5.16482294e-01 1.74585402e-01 2.52873808e-01 -2.42010251e-01
2.79017389e-01 -7.34037220e-01 -2.40680233e-01 1.76798865e-01
-8.62327099e-01 -7.14300632e-01 3.83311838e-01 1.14357650e+00
5.56522667e-01 1.67465925e-01 8.50465178e-01 -1.11951804e+00
4.93339509e-01 -6.81251407e-01 -1.11605239e+00 2.44916826e-01
-9.91520762e-01 3.14354897e-01 8.89865696e-01 -7.01420307e-01
-8.40120852e-01 9.80439112e-02 -5.01273982e-02 -5.93474805e-01
-8.84326026e-02 4.51537281e-01 -8.87664929e-02 -4.25360203e-02
1.20413244e+00 -2.37567306e-01 -1.70569181e-01 -6.34134710e-01
5.06608188e-01 5.15451670e-01 7.58787870e-01 -9.68047082e-01
8.39006782e-01 4.78596747e-01 1.22050159e-01 -3.09115469e-01
-1.47103631e+00 -4.09236848e-02 -4.30741191e-01 1.92552313e-01
5.03908157e-01 -6.41779244e-01 -9.77932155e-01 2.26718053e-01
-6.39303505e-01 -6.37160420e-01 -6.34479225e-01 4.70010310e-01
-4.42406505e-01 5.08845329e-01 -6.99740231e-01 -1.23375714e+00
-1.76897749e-01 -9.13240254e-01 8.10903847e-01 -2.42406316e-03
-7.01528862e-02 -1.20232606e+00 -2.82275826e-01 7.12739080e-02
1.22201942e-01 1.64835393e-01 9.75230157e-01 -1.27595901e+00
-2.02438727e-01 -3.89372051e-01 -2.94375330e-01 9.20287848e-01
-8.88183489e-02 -1.39470071e-01 -1.23685253e+00 -4.00914460e-01
1.81669891e-01 -7.95039058e-01 1.02879083e+00 3.22403163e-01
1.48090613e+00 -5.02338588e-01 -6.94344863e-02 7.03342199e-01
1.33729517e+00 -1.85417280e-01 5.32111973e-02 3.51905227e-01
4.84507680e-01 6.02850914e-01 1.67990476e-01 3.44071090e-01
-1.08371615e-01 3.60401310e-02 4.63220179e-01 2.45689198e-01
3.15501988e-01 -3.24603230e-01 2.05579817e-01 2.80216068e-01
3.78376901e-01 -1.98755085e-01 -8.48337829e-01 5.03211260e-01
-1.50670862e+00 -7.72585809e-01 3.71112209e-03 2.58301878e+00
1.19790280e+00 6.46156788e-01 8.57451111e-02 1.91160366e-01
7.23887622e-01 -4.06432092e-01 -9.91841614e-01 -3.56595278e-01
-6.82657361e-02 2.33597025e-01 9.50285733e-01 8.18242431e-01
-1.01733506e+00 6.55506790e-01 6.60263824e+00 9.71315145e-01
-9.58678544e-01 2.52182126e-01 1.01262379e+00 -2.35395238e-01
-3.28594416e-01 2.03489736e-02 -8.66369128e-01 3.35790217e-01
1.04638946e+00 -2.79635698e-01 5.76884985e-01 1.32888401e+00
-3.05348128e-01 -1.35429017e-02 -1.66434169e+00 8.27400208e-01
-3.32353085e-01 -1.01607156e+00 -3.87259066e-01 3.39556783e-01
8.71123195e-01 1.44682154e-01 3.63842398e-01 3.78510565e-01
8.25957537e-01 -1.12482226e+00 7.76820421e-01 -4.54726890e-02
7.07600534e-01 -8.80516112e-01 8.71620417e-01 6.21658742e-01
-5.16545951e-01 -4.28328663e-01 -4.80150938e-01 3.46448831e-02
-1.51017889e-01 9.04139400e-01 -7.20774353e-01 8.16819519e-02
4.21129376e-01 2.74504453e-01 -2.94375479e-01 9.83485579e-01
-2.19630450e-01 8.61455619e-01 -8.66222918e-01 1.50873721e-01
3.42296660e-01 -9.19898301e-02 2.72889286e-01 1.13416886e+00
5.31149991e-02 -8.72118995e-02 1.34184018e-01 9.83505189e-01
-4.87620085e-01 -1.29247651e-01 -4.82924640e-01 -2.95234169e-03
7.14289546e-01 1.09555757e+00 -7.91139901e-01 -2.25322232e-01
-1.03664286e-01 7.07519114e-01 5.75672269e-01 5.00557721e-01
-6.85486257e-01 -1.79653540e-01 4.57813114e-01 -1.53289929e-01
2.45129898e-01 9.44582745e-03 -7.22081006e-01 -9.62546170e-01
2.89512038e-01 -6.77349985e-01 4.81866896e-01 -2.55841434e-01
-1.79776466e+00 4.54016268e-01 -5.78068532e-02 -7.87629366e-01
-2.96644986e-01 -8.63705814e-01 -3.34232092e-01 8.22161973e-01
-1.17481339e+00 -6.93157971e-01 8.82101879e-02 2.62895882e-01
1.55231416e-01 2.18435839e-01 6.79020345e-01 2.95243710e-01
-7.48679876e-01 1.18713772e+00 2.98968017e-01 2.36732677e-01
4.57723469e-01 -1.49989271e+00 2.23107979e-01 8.47641766e-01
1.61155790e-01 6.87575281e-01 8.68098915e-01 -3.94519895e-01
-9.09461021e-01 -9.88297701e-01 3.48658472e-01 -6.46922290e-01
9.69907343e-01 -3.19716334e-01 -1.01560724e+00 9.45118606e-01
-4.04857278e-01 3.66713077e-01 7.14667320e-01 4.98494148e-01
-8.05403590e-01 -1.44385308e-01 -1.41188633e+00 4.24097747e-01
1.04205561e+00 -5.10459423e-01 -4.04555976e-01 5.51653147e-01
7.06608355e-01 -2.76721805e-01 -8.14618170e-01 3.43958080e-01
5.37725329e-01 -8.29667926e-01 4.59001452e-01 -1.12651479e+00
2.80868292e-01 1.91427827e-01 -3.98932815e-01 -1.24494052e+00
1.41450688e-01 -6.94444776e-01 2.69254087e-03 1.17667031e+00
7.86262929e-01 -7.94788957e-01 9.85708535e-01 1.23292720e+00
-8.16443115e-02 -1.01673424e+00 -9.58383322e-01 -1.28591597e+00
6.49798274e-01 -8.05465281e-01 2.22288594e-01 1.09144390e+00
-1.66910857e-01 -1.02331862e-02 -1.50090232e-02 3.89512926e-01
8.98821473e-01 -3.00516635e-01 3.65573823e-01 -1.22703576e+00
-4.19290751e-01 -6.10161006e-01 -4.35753286e-01 -9.08442795e-01
4.52821732e-01 -1.03708124e+00 1.37894407e-01 -8.16274941e-01
2.76566118e-01 -8.43965530e-01 -4.25987452e-01 7.99969852e-01
-1.15796775e-01 6.69333860e-02 -1.32754460e-01 -5.95383905e-03
-3.26366633e-01 2.31738836e-01 7.18063891e-01 1.86572343e-01
-6.02937043e-02 1.61475196e-01 -9.61760700e-01 8.48918378e-01
8.72849882e-01 -6.12621486e-01 -4.66618657e-01 -2.61778921e-01
5.03918350e-01 -3.87870610e-01 6.64106965e-01 -7.63665438e-01
5.31862378e-02 -1.01395091e-02 2.56422877e-01 -9.74512026e-02
-1.01433974e-02 -8.58761311e-01 -2.77171731e-01 4.16618288e-01
-8.92481685e-01 -3.13399315e-01 7.91731104e-02 8.32400262e-01
3.61086011e-01 -7.06919432e-01 1.10655153e+00 1.14179403e-01
-1.07709944e-01 3.18330973e-01 1.23450924e-02 5.64677656e-01
8.04026604e-01 1.21617153e-01 -3.39062095e-01 -3.19936305e-01
-1.07862031e+00 3.57896864e-01 5.53789198e-01 -5.42784519e-02
1.41328707e-01 -1.12351072e+00 -4.85736847e-01 1.09709226e-01
-1.24912567e-01 1.48673356e-01 5.15538640e-02 7.79572725e-01
-1.17751800e-01 -1.43333629e-01 2.83037543e-01 -5.13110697e-01
-6.73741400e-01 9.27817941e-01 6.25816643e-01 -1.68208312e-02
-6.49219871e-01 1.29695404e+00 3.45806658e-01 -2.55872846e-01
7.98750639e-01 -4.90754008e-01 5.22800684e-01 -6.20882474e-02
5.63270450e-01 2.75364876e-01 1.94455564e-01 6.12962879e-02
-1.51328519e-01 8.65061805e-02 -3.37013721e-01 -3.00292194e-01
1.24299049e+00 -6.13117926e-02 1.44037187e-01 5.65045595e-01
1.58390307e+00 -9.54866558e-02 -1.58587348e+00 -2.67088830e-01
2.73455143e-01 -3.18738461e-01 1.46159336e-01 -8.47610533e-01
-1.09358335e+00 9.44016874e-01 6.34501696e-01 5.20037949e-01
1.01958251e+00 1.17731154e-01 6.01903319e-01 8.90262365e-01
2.06806749e-01 -1.05256808e+00 -2.64059126e-01 7.90205225e-02
5.41735351e-01 -1.31713176e+00 -1.50983602e-01 -2.01587856e-01
-4.48172271e-01 9.57349300e-01 4.51867461e-01 -2.35957801e-01
7.24463642e-01 5.66128671e-01 2.33508758e-02 3.07343360e-02
-7.64449120e-01 1.50074258e-01 7.77435023e-03 4.62452739e-01
4.73155566e-02 -8.99478793e-02 -1.20401226e-01 9.86634433e-01
-5.16820252e-01 -6.78996667e-02 3.92753690e-01 5.44929981e-01
-6.15163565e-01 -8.69530916e-01 -1.26836136e-01 5.96480489e-01
-7.19326615e-01 -3.18694860e-01 -1.93284035e-01 8.36867452e-01
3.76225216e-03 8.68330300e-01 -6.68540969e-02 -3.27559412e-01
1.55791417e-02 2.75794387e-01 5.08362830e-01 -4.99598324e-01
-2.67207265e-01 -1.58589616e-01 2.69377176e-02 -3.42675000e-01
3.09449453e-02 -5.82388818e-01 -1.27940392e+00 -4.03801471e-01
-5.94764233e-01 2.35390320e-01 6.63809717e-01 1.13227081e+00
7.02040344e-02 4.40549962e-02 8.10557067e-01 -5.94561517e-01
-1.69096637e+00 -8.39552581e-01 -7.96469927e-01 3.98817688e-01
5.92421353e-01 -5.70187807e-01 -1.18480563e+00 -1.55172916e-02] | [8.125164031982422, 4.067102909088135] |
786aae86-9cd8-4f87-8684-4a2153f8c484 | synthesizing-programs-with-continuous | 2211.00828 | null | https://arxiv.org/abs/2211.00828v2 | https://arxiv.org/pdf/2211.00828v2.pdf | Synthesizing Programs with Continuous Optimization | Automatic software generation based on some specification is known as program synthesis. Most existing approaches formulate program synthesis as a search problem with discrete parameters. In this paper, we present a novel formulation of program synthesis as a continuous optimization problem and use a state-of-the-art evolutionary approach, known as Covariance Matrix Adaptation Evolution Strategy to solve the problem. We then propose a mapping scheme to convert the continuous formulation into actual programs. We compare our system, called GENESYS, with several recent program synthesis techniques (in both discrete and continuous domains) and show that GENESYS synthesizes more programs within a fixed time budget than those existing schemes. For example, for programs of length 10, GENESYS synthesizes 28% more programs than those existing schemes within the same time budget. | ['Abdullah Muzahid', 'Justin Gottschlich', 'Javier Turek', 'Todd A. Anderson', 'Shantanu Mandal'] | 2022-11-02 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 3.71792167e-01 -9.22720358e-02 -2.73556739e-01 -3.56306404e-01
-5.58012128e-01 -6.02924824e-01 6.42482117e-02 -1.03782043e-02
-1.54038489e-01 8.19207668e-01 -5.15963256e-01 -4.58335668e-01
1.25131914e-02 -1.13616920e+00 -8.29236746e-01 -2.52287835e-01
3.44302416e-01 3.42000216e-01 4.46676731e-01 -4.44509983e-01
6.85388684e-01 9.18396860e-02 -1.78259218e+00 -1.14421718e-01
1.37491965e+00 3.53392661e-01 2.32227519e-01 8.96497846e-01
-2.92877644e-01 2.08550677e-01 -8.35444510e-01 -1.99313432e-01
2.62843609e-01 -8.64620507e-01 -4.58454251e-01 -6.83813989e-02
-8.85474756e-02 8.19276720e-02 4.90179844e-02 1.37237465e+00
3.36793184e-01 -6.16188534e-02 3.69556457e-01 -1.53582859e+00
-9.03721750e-01 6.75601006e-01 -5.57170570e-01 -3.25715601e-01
3.61393094e-01 1.19393863e-01 7.73560464e-01 -7.53626645e-01
5.54109395e-01 1.18408108e+00 5.68556786e-01 5.93223214e-01
-1.68363082e+00 -5.90632260e-01 -1.99132804e-02 -3.14827383e-01
-1.66150653e+00 -1.57089457e-01 6.78875268e-01 -6.12862110e-01
1.32261896e+00 4.62519199e-01 7.42661297e-01 5.39173722e-01
4.34812635e-01 4.99342233e-01 1.07942331e+00 -6.99909210e-01
6.69515610e-01 2.89230257e-01 3.61499786e-01 8.05228531e-01
6.22258484e-01 4.82401848e-01 -6.91907853e-02 -7.47362852e-01
5.13304651e-01 -4.90762085e-01 -1.57428622e-01 -4.60536093e-01
-1.06904292e+00 1.16573250e+00 -2.18042701e-01 1.20430388e-01
-2.41525963e-01 4.85166430e-01 3.32459480e-01 4.91674691e-01
1.64812922e-01 7.26569533e-01 -4.33717698e-01 -2.75104731e-01
-1.14704645e+00 6.22936726e-01 1.21781337e+00 1.41150558e+00
8.28490317e-01 5.40757477e-01 -2.05598369e-01 5.51401258e-01
5.02212942e-01 3.96005452e-01 4.89894390e-01 -6.96200013e-01
2.70799696e-01 8.83376777e-01 2.71176100e-02 -8.27516437e-01
3.78722772e-02 -1.69455007e-01 -1.85091764e-01 5.43507159e-01
-2.97198355e-01 -6.89114749e-01 -5.72219372e-01 1.56556475e+00
3.17272812e-01 -8.61337688e-03 9.86507386e-02 3.39375675e-01
3.51870507e-01 9.73887265e-01 -5.25905192e-01 -4.95873004e-01
9.21817124e-01 -1.32534254e+00 -7.35261858e-01 -6.07197657e-02
4.41631973e-01 -7.44644582e-01 9.56623614e-01 4.93393332e-01
-1.02551496e+00 -5.21298587e-01 -1.48239303e+00 5.46263635e-01
-2.50262290e-01 2.46169254e-01 5.89002073e-01 1.44432986e+00
-1.43090665e+00 5.35697281e-01 -8.14747393e-01 -1.23628750e-01
-1.28503054e-01 6.43024266e-01 3.31705540e-01 4.94627267e-01
-6.45898223e-01 6.39638901e-01 5.49852908e-01 -3.43866497e-01
-7.86743402e-01 -7.32508063e-01 -9.27533627e-01 1.77198406e-02
5.37351131e-01 -7.11770952e-01 1.34117353e+00 -9.85215902e-01
-2.11474490e+00 5.59159219e-01 -2.67474651e-01 -2.76064575e-01
2.89941490e-01 2.63600469e-01 -2.88754702e-01 -5.93997598e-01
-2.44399413e-01 1.94854617e-01 8.88249159e-01 -1.18729687e+00
-5.14840603e-01 1.33357793e-01 1.67342156e-01 -2.44707420e-01
-3.89689028e-01 4.05650824e-01 -2.03362882e-01 -7.35373616e-01
-3.36853087e-01 -1.35339606e+00 -5.64758718e-01 -1.23882391e-01
-1.92823470e-01 -1.78516567e-01 7.43156314e-01 -5.51119506e-01
1.74876904e+00 -1.94079435e+00 7.28739560e-01 3.89925122e-01
-1.42402455e-01 1.24319188e-01 -3.05430293e-02 5.45461059e-01
-3.91848832e-02 2.55943239e-01 -6.60481572e-01 9.01163816e-02
4.61927563e-01 1.70155033e-01 -1.69616029e-01 2.28694052e-01
2.31681794e-01 7.56686211e-01 -7.29417562e-01 -4.04177457e-01
-3.54170024e-01 3.09029669e-02 -1.15691113e+00 2.24756718e-01
-7.46917248e-01 -1.30979821e-01 -6.59980118e-01 6.26535892e-01
6.89462900e-01 9.25675705e-02 3.00210178e-01 2.18070790e-01
-4.86586750e-01 -3.43725204e-01 -1.39862549e+00 1.63322198e+00
-5.66507578e-01 5.81029475e-01 3.76597568e-02 -9.18189347e-01
1.43161821e+00 2.35279799e-01 2.41336033e-01 1.05892457e-02
1.57008454e-01 3.72450143e-01 2.31357843e-01 -4.70986336e-01
7.51659214e-01 2.82067031e-01 -3.17354023e-01 6.71625078e-01
-2.19459087e-01 -8.72611642e-01 6.02939427e-01 -2.89880902e-01
1.16365635e+00 3.23880583e-01 4.76551265e-01 -4.62456107e-01
6.73690677e-01 3.88149142e-01 6.99345887e-01 8.29102337e-01
1.47194624e-01 3.44726205e-01 8.24414551e-01 4.48746644e-02
-1.33780503e+00 -4.73851353e-01 5.94863780e-02 7.35784292e-01
-1.42309099e-01 -6.75311983e-01 -9.76442397e-01 -5.70486963e-01
1.35560393e-01 9.93701398e-01 -3.49557161e-01 -2.50725836e-01
-7.86623061e-01 -9.86266494e-01 5.63070357e-01 2.87614822e-01
3.59069347e-01 -9.89954412e-01 -9.43572581e-01 4.32906747e-01
4.81557488e-01 -5.69558501e-01 -7.40265727e-01 1.00538172e-01
-7.91215003e-01 -6.41863525e-01 -6.79926634e-01 -9.13988948e-01
8.30856502e-01 -3.54796767e-01 1.08664703e+00 4.18642700e-01
-5.12659073e-01 2.79004183e-02 -4.49955940e-01 -2.51113832e-01
-1.07334828e+00 -4.10729125e-02 -2.01829925e-01 -3.55363697e-01
-3.29831336e-03 -4.94011372e-01 4.40539159e-02 9.18980688e-02
-9.66201663e-01 5.96618652e-03 4.91409838e-01 1.27486885e+00
5.09886861e-01 6.07710719e-01 5.54098845e-01 -8.92227471e-01
1.13352585e+00 -4.29133862e-01 -1.45849752e+00 5.79800010e-01
-1.08813536e+00 4.92441684e-01 5.44887543e-01 -6.37604773e-01
-1.01366210e+00 4.71127898e-01 -6.21319972e-02 -1.34995610e-01
2.97921658e-01 7.71959841e-01 -1.18472256e-01 -7.05639005e-01
7.53367424e-01 5.17901838e-01 -3.11768278e-02 -6.57161325e-02
2.08393067e-01 7.15925753e-01 9.43497047e-02 -1.10041618e+00
8.95013154e-01 -2.94442832e-01 -1.29113257e-01 -4.07732666e-01
3.40254843e-01 2.14809194e-01 -3.29477191e-01 1.07840680e-01
6.04403853e-01 -3.27879071e-01 -5.74242473e-01 2.42941678e-01
-1.17863894e+00 -3.61675203e-01 -3.29073668e-02 6.15819395e-02
-7.88961232e-01 2.23149657e-01 -1.73266038e-01 -8.71499717e-01
-4.15057480e-01 -1.74341249e+00 1.03891504e+00 2.27745369e-01
-5.68579733e-01 -6.70429826e-01 4.13083285e-01 -2.83020198e-01
5.99146903e-01 5.17211914e-01 1.18603730e+00 -2.47228086e-01
-5.94417155e-01 -1.03949144e-01 1.99778840e-01 2.54114211e-01
9.23312306e-02 7.51411200e-01 -1.65072531e-01 -2.70750433e-01
1.64623797e-01 1.71592817e-01 7.74224699e-02 1.37552559e-01
1.00858128e+00 -2.39785835e-01 -4.09639657e-01 7.75417864e-01
1.76848221e+00 1.08424151e+00 4.97611701e-01 4.03615654e-01
1.23228796e-01 2.69847453e-01 8.41748834e-01 5.94109476e-01
-7.23632723e-02 9.70245779e-01 1.54555291e-01 2.34748080e-01
2.48492211e-01 -7.82734528e-02 5.78009427e-01 8.50434899e-01
2.93963224e-01 -5.07561900e-02 -1.09543395e+00 5.44244707e-01
-2.05032516e+00 -7.36350775e-01 -6.09991699e-02 1.98793459e+00
1.25643468e+00 4.88531403e-02 2.52741575e-01 1.04021020e-01
8.73087823e-01 -3.23085427e-01 -5.56400418e-01 -9.78961408e-01
1.37711897e-01 6.79875791e-01 5.30461013e-01 4.90015239e-01
-6.57408297e-01 8.50509942e-01 7.44847536e+00 7.69642115e-01
-9.83150065e-01 -6.36694506e-02 1.45922378e-01 1.38969406e-01
-6.03961229e-01 3.35674971e-01 -7.82440484e-01 5.65411448e-01
1.10427558e+00 -1.26892233e+00 5.86373806e-01 1.23198116e+00
-8.02717637e-03 -3.99124529e-03 -1.26130188e+00 6.77586317e-01
1.93210930e-01 -1.34051359e+00 -4.33377087e-01 -1.09042004e-01
1.26416695e+00 -7.02314436e-01 -1.15040429e-02 4.51495707e-01
5.38936198e-01 -1.00986445e+00 8.28185856e-01 2.07741469e-01
6.02710426e-01 -1.02623522e+00 4.41756010e-01 3.21491122e-01
-1.12239289e+00 -9.50321779e-02 -1.03343301e-01 9.80191901e-02
1.74229428e-01 4.24144089e-01 -6.09157979e-01 6.14134550e-01
3.62701535e-01 3.13045979e-01 -6.67039216e-01 1.34899175e+00
-1.97672561e-01 5.42319596e-01 -3.20720747e-02 -6.72971606e-01
-8.43122303e-02 -4.60270941e-01 7.03083158e-01 1.05918086e+00
8.62418294e-01 5.92862666e-02 1.58061281e-01 1.71838415e+00
3.40678543e-01 1.15799911e-01 -5.17279327e-01 -4.84091967e-01
5.96962929e-01 7.27621853e-01 -7.05197215e-01 -4.72841531e-01
-3.48665357e-01 8.75051260e-01 -1.90741464e-01 2.80282706e-01
-1.27262139e+00 -1.10532010e+00 3.82822335e-01 -4.67834234e-01
3.66859049e-01 -3.95927370e-01 -5.47193706e-01 -9.82155263e-01
5.71506238e-03 -1.48747456e+00 -2.02668622e-01 -5.34741819e-01
-5.16135216e-01 8.67364705e-01 1.71355680e-01 -1.00556695e+00
-5.85801721e-01 -4.62108493e-01 -6.34904206e-01 1.00345993e+00
-9.81039524e-01 -4.53606039e-01 -1.60544351e-01 4.95067798e-02
6.50847137e-01 -5.75706840e-01 8.74105215e-01 2.66278327e-01
-7.53434122e-01 7.97201276e-01 1.70302257e-01 -5.62387109e-01
3.61024290e-01 -1.09520268e+00 5.70678115e-01 9.04132366e-01
-4.55345213e-01 9.28522348e-01 1.10133767e+00 -8.94671261e-01
-1.97473085e+00 -9.14827704e-01 5.39751649e-01 5.92852570e-03
7.08937526e-01 -2.71075577e-01 -6.70048654e-01 5.31987667e-01
4.17577147e-01 -4.64082181e-01 3.14562112e-01 -4.17288989e-01
7.81519115e-02 1.84321895e-01 -1.36516857e+00 7.23887324e-01
8.96338880e-01 -2.03682408e-01 -5.59582412e-01 2.80094415e-01
1.13775182e+00 -6.89690232e-01 -8.13666046e-01 4.68980789e-01
3.56887132e-01 -6.33840144e-01 6.72620058e-01 -1.57295287e-01
4.41114157e-01 -6.07770562e-01 -1.95045918e-01 -1.36395085e+00
-2.76220255e-02 -1.01806879e+00 2.67051701e-02 1.24919021e+00
5.25000513e-01 -8.68485749e-01 8.23035002e-01 6.21137857e-01
-1.66888446e-01 -6.06635332e-01 -4.98794615e-01 -1.12812638e+00
1.25912324e-01 4.19645570e-02 1.06413317e+00 7.09751546e-01
1.36122838e-01 -8.98723528e-02 -1.88998461e-01 5.45426346e-02
4.93776292e-01 5.23598135e-01 7.22699881e-01 -7.46659040e-01
-9.68221366e-01 -6.26766145e-01 -3.20160121e-01 -5.07730722e-01
5.11315107e-01 -8.36600006e-01 2.36373559e-01 -9.23692465e-01
6.87090158e-02 -4.54635084e-01 2.95458347e-01 3.36378902e-01
-8.95452350e-02 -4.03472394e-01 5.49841300e-02 -4.30761814e-01
2.27228552e-01 5.16233802e-01 8.73719633e-01 -3.09353739e-01
-6.01449072e-01 2.02212945e-01 -6.30097151e-01 2.93008387e-01
7.90309191e-01 -7.32973158e-01 -5.69824874e-01 -2.57976621e-01
4.97994304e-01 6.00495338e-01 -1.34358898e-01 -9.70115244e-01
2.71844596e-01 -6.98832929e-01 -5.30753732e-01 -5.39640963e-01
-1.64889712e-02 -7.06685305e-01 1.06325746e+00 8.70352089e-01
-2.81364948e-01 6.03354871e-01 4.32888597e-01 1.32703573e-01
-3.75909656e-01 -9.61304307e-01 5.47131300e-01 9.18932408e-02
-6.58464074e-01 -1.21065974e-01 -4.21871990e-01 -2.97188133e-01
1.39660370e+00 -1.55938923e-01 -2.52447665e-01 4.96869013e-02
-5.30918419e-01 1.41146272e-01 7.67628849e-01 3.49578917e-01
7.16246545e-01 -1.33929658e+00 -5.76904178e-01 2.56006092e-01
-5.49072251e-02 -4.72675353e-01 -2.70716995e-01 5.40447414e-01
-9.03085113e-01 3.48826021e-01 -2.07682505e-01 -3.79857749e-01
-1.54952550e+00 5.92509866e-01 3.32331508e-01 -4.45055552e-02
-9.89146307e-02 8.02076459e-01 -2.17996314e-01 -7.54769027e-01
-1.58381566e-01 -2.79085457e-01 1.84572205e-01 -3.27193618e-01
1.75154194e-01 2.74487793e-01 -7.55626038e-02 -1.01108573e-01
-3.10392827e-01 7.13458598e-01 4.47379619e-01 -3.15528870e-01
1.34881592e+00 5.29950261e-01 -6.08434379e-01 1.78657070e-01
1.06449234e+00 3.94069739e-02 -6.10956132e-01 2.76648421e-02
1.77930400e-01 -6.57042086e-01 -8.23723599e-02 -5.34986615e-01
-9.40663218e-01 3.47939730e-01 6.30445838e-01 1.78915471e-01
1.21844363e+00 -7.02631891e-01 4.08411294e-01 3.17520887e-01
6.86197042e-01 -9.68772650e-01 2.20258553e-02 5.54575682e-01
9.21382368e-01 -8.41401935e-01 3.10701467e-02 -6.95578396e-01
-3.95693690e-01 1.28411746e+00 8.13242197e-01 -4.51207519e-01
5.25976539e-01 1.00512266e+00 -6.03986502e-01 1.91375345e-01
-9.86731231e-01 1.87126294e-01 2.25578904e-01 5.29490888e-01
4.61856186e-01 1.60740269e-03 -9.91567910e-01 6.80197060e-01
-4.28719193e-01 3.27569425e-01 8.42465878e-01 1.41531765e+00
-4.31428581e-01 -1.74807584e+00 -6.37153566e-01 2.79956609e-01
-8.63452554e-02 -8.87939334e-02 -2.56181538e-01 7.44741857e-01
7.06860274e-02 7.77666330e-01 -4.36512411e-01 -5.69801211e-01
4.26928133e-01 6.89490605e-03 5.72197795e-01 -9.96468127e-01
-7.94606924e-01 -4.28071916e-02 2.93565333e-01 -2.71101832e-01
-1.86968446e-01 -6.98584497e-01 -1.17230260e+00 -1.81330785e-01
-6.12290621e-01 2.27463186e-01 7.30951250e-01 4.50888544e-01
2.95797467e-01 9.81381536e-01 6.95929527e-01 -5.73885500e-01
-9.57020342e-01 -2.20839620e-01 -3.22154433e-01 -3.70252192e-01
1.19770065e-01 -4.96976137e-01 -7.07252026e-02 8.57492983e-02] | [8.048089981079102, 7.302132606506348] |
f5a188b1-92ae-41a2-b239-0d15de717733 | is-chatgpt-a-highly-fluent-grammatical-error | 2304.01746 | null | https://arxiv.org/abs/2304.01746v1 | https://arxiv.org/pdf/2304.01746v1.pdf | Is ChatGPT a Highly Fluent Grammatical Error Correction System? A Comprehensive Evaluation | ChatGPT, a large-scale language model based on the advanced GPT-3.5 architecture, has shown remarkable potential in various Natural Language Processing (NLP) tasks. However, there is currently a dearth of comprehensive study exploring its potential in the area of Grammatical Error Correction (GEC). To showcase its capabilities in GEC, we design zero-shot chain-of-thought (CoT) and few-shot CoT settings using in-context learning for ChatGPT. Our evaluation involves assessing ChatGPT's performance on five official test sets in three different languages, along with three document-level GEC test sets in English. Our experimental results and human evaluations demonstrate that ChatGPT has excellent error detection capabilities and can freely correct errors to make the corrected sentences very fluent, possibly due to its over-correction tendencies and not adhering to the principle of minimal edits. Additionally, its performance in non-English and low-resource settings highlights its potential in multilingual GEC tasks. However, further analysis of various types of errors at the document-level has shown that ChatGPT cannot effectively correct agreement, coreference, tense errors across sentences, and cross-sentence boundary errors. | ['Yue Zhang', 'Lidia S. Chao', 'Jinpeng Hu', 'Derek F. Wong', 'Kaixin Lan', 'Shu Yang', 'Tao Fang'] | 2023-04-04 | null | null | null | null | ['grammatical-error-correction'] | ['natural-language-processing'] | [-5.17379977e-02 2.30058014e-01 4.09719676e-01 -3.97229493e-01
-1.16232955e+00 -3.43588084e-01 3.71934682e-01 6.77986801e-01
-8.30326557e-01 8.43948483e-01 4.14515108e-01 -6.21131599e-01
9.77077335e-02 -3.25281292e-01 -6.19487941e-01 -5.03265932e-02
-1.27850860e-01 7.74813354e-01 1.52719989e-01 -7.37480581e-01
5.93300402e-01 -2.40733162e-01 -8.29742074e-01 6.04789197e-01
1.56169152e+00 -2.92854719e-02 4.02378857e-01 7.54657328e-01
-2.15641484e-01 8.61480296e-01 -8.39796126e-01 -1.02022350e+00
-3.46714914e-01 -1.60148039e-01 -1.35716558e+00 -6.98659360e-01
5.26215672e-01 1.61888540e-01 4.81126159e-01 9.43468332e-01
6.84973001e-01 1.28872007e-01 2.17309698e-01 -7.65143335e-01
-8.71845722e-01 9.89523172e-01 2.17302293e-02 4.66949642e-01
6.00887477e-01 2.71689624e-01 1.19262624e+00 -1.11293805e+00
9.60473001e-01 1.44055569e+00 1.20806277e+00 8.91538978e-01
-1.12002969e+00 -3.68338048e-01 -1.37576088e-02 1.30816996e-01
-1.13482523e+00 -5.62984347e-01 1.34774864e-01 -1.23528816e-01
2.30372214e+00 1.76830873e-01 4.11410719e-01 9.93742943e-01
7.65443325e-01 6.39984429e-01 9.69263434e-01 -8.84539247e-01
-1.04424715e-01 -8.43068883e-02 1.95922494e-01 6.61380410e-01
5.15204221e-02 -1.23698905e-03 -8.10976505e-01 7.66591802e-02
-2.04133078e-01 -7.37320304e-01 -3.23856175e-01 5.27498543e-01
-1.07523406e+00 6.87890947e-01 1.63627744e-01 6.68272018e-01
5.34642600e-02 2.75628775e-01 8.22442591e-01 6.79591656e-01
6.38924718e-01 7.18907118e-01 -8.13722432e-01 -7.14929104e-01
-6.91256940e-01 2.61859328e-01 1.03862751e+00 1.23886108e+00
3.30758661e-01 4.47000377e-02 -2.88804144e-01 1.08454990e+00
3.30376148e-01 5.27547002e-01 4.72581118e-01 -8.49620581e-01
1.29582059e+00 5.59887469e-01 -1.24685854e-01 -9.70007777e-01
-4.44342762e-01 -1.58731639e-01 -4.67276573e-01 -2.55253434e-01
4.28507209e-01 -9.52313989e-02 -3.77309829e-01 1.81066048e+00
-1.54342219e-01 -4.35605943e-01 2.40955994e-01 5.41965961e-01
9.08039927e-01 5.18017113e-01 6.53103530e-01 -1.71743631e-01
1.09523153e+00 -8.05576146e-01 -7.88040578e-01 -7.50268459e-01
1.52905977e+00 -1.08210301e+00 1.29215086e+00 1.28372684e-01
-1.31128454e+00 -3.23036164e-01 -5.68043828e-01 -6.47358358e-01
-3.13635916e-01 3.84388678e-02 3.67940724e-01 5.54920435e-01
-1.25448537e+00 6.70313835e-01 -8.65370929e-01 -7.66646266e-01
-6.33994192e-02 9.44385491e-03 -3.87707740e-01 -2.24044532e-01
-1.52029014e+00 1.48941588e+00 4.71485585e-01 9.09920335e-02
-2.51116723e-01 -7.18578041e-01 -1.00435865e+00 1.63190160e-02
2.42492884e-01 -4.94012386e-01 1.64942837e+00 -5.98549783e-01
-1.15097082e+00 1.05931056e+00 -2.46882513e-01 -5.25827169e-01
2.50967652e-01 -4.78520453e-01 -5.64671040e-01 -2.47836903e-01
3.98624390e-01 5.39615929e-01 1.86093837e-01 -6.65266871e-01
-6.69596195e-01 -2.78875738e-01 -3.17552805e-01 4.67095464e-01
1.09033652e-01 8.24548841e-01 -9.72958356e-02 -6.10041916e-01
-1.01895250e-01 -8.42718124e-01 1.06551290e-01 -5.76945782e-01
-1.64133549e-01 -7.17286885e-01 2.33898461e-01 -1.09593940e+00
1.65190792e+00 -2.01038146e+00 1.31158024e-01 -1.93705246e-01
-1.10113844e-01 6.06501997e-01 -1.89879537e-01 9.13752317e-01
2.02272251e-01 6.52296066e-01 -3.87367785e-01 -7.62236655e-01
-8.96478295e-02 4.49100912e-01 5.63107245e-02 -4.85485606e-02
3.41904610e-01 1.09321213e+00 -1.20096421e+00 -6.50467515e-01
-7.17274174e-02 1.05956040e-01 -8.12952399e-01 -1.33206606e-01
-1.72336116e-01 3.86271328e-01 2.44808495e-02 5.61727524e-01
3.03181380e-01 1.96046963e-01 4.02307630e-01 5.27221680e-01
-4.35052991e-01 1.10881865e+00 -5.79436064e-01 1.84302318e+00
-6.10157728e-01 3.57261539e-01 1.29334390e-01 -6.14181638e-01
5.64436495e-01 5.56677341e-01 -4.39417273e-01 -7.85460591e-01
8.76031667e-02 8.17391992e-01 4.56834435e-01 -5.18640637e-01
9.51828539e-01 -1.15217082e-01 -3.78178954e-01 5.02161801e-01
4.26348448e-01 -3.06203663e-01 4.30882841e-01 6.50398254e-01
1.16374469e+00 -2.16956750e-01 3.38791192e-01 -4.80557859e-01
4.40842181e-01 3.54009598e-01 6.84447885e-01 1.00028265e+00
-3.81919742e-01 3.85707080e-01 2.65600234e-01 3.99938002e-02
-8.07536840e-01 -5.39440215e-01 -2.38384139e-02 1.14269280e+00
-4.25158322e-01 -9.55200911e-01 -8.14496458e-01 -7.89771616e-01
-9.38113630e-02 1.34925342e+00 -1.57305494e-01 -1.46424532e-01
-1.02647769e+00 -5.21207988e-01 9.59362209e-01 3.93784106e-01
4.58909899e-01 -1.50804639e+00 -1.50313288e-01 7.37749934e-01
-7.62577116e-01 -1.03967011e+00 -7.68706858e-01 2.99222227e-02
-8.84558201e-01 -1.12765229e+00 1.77728273e-02 -1.01865411e+00
3.39340687e-01 -2.46499125e-02 1.55990207e+00 6.69700682e-01
1.40973687e-01 4.30872291e-01 -7.02194333e-01 -3.89666378e-01
-1.10143805e+00 2.94827551e-01 1.76186431e-02 -9.26996946e-01
7.62391865e-01 -2.56768078e-01 4.45214473e-02 -1.70998558e-01
-1.67568654e-01 -1.24334686e-01 2.99652576e-01 1.03576303e+00
1.68107703e-01 -3.43634307e-01 6.20335817e-01 -1.26796293e+00
9.15278196e-01 -3.21239322e-01 -8.81606266e-02 4.69168276e-01
-8.71909320e-01 -1.75471440e-01 4.56471443e-01 1.55824140e-01
-1.48537147e+00 -7.29528069e-01 -5.71409285e-01 4.88150507e-01
1.35336682e-01 9.19057608e-01 1.01590753e-01 5.07374760e-03
7.60863543e-01 1.83484703e-02 -1.06442444e-01 -5.68170190e-01
2.23208398e-01 7.35664666e-01 5.31808913e-01 -9.20795918e-01
2.47466475e-01 -4.87281799e-01 -7.64410734e-01 -6.21307135e-01
-9.62262332e-01 -3.45065176e-01 -5.39112926e-01 -1.47412410e-02
7.52320826e-01 -9.42574978e-01 -4.78472054e-01 7.54940093e-01
-1.63600314e+00 -5.45955420e-01 5.02480790e-02 2.16576442e-01
-2.23839939e-01 4.53423291e-01 -1.17981589e+00 -5.23241937e-01
-8.86059761e-01 -8.97304714e-01 8.76195669e-01 -1.01327904e-01
-6.31621718e-01 -1.33148193e+00 1.44732729e-01 4.80417669e-01
5.35806775e-01 -4.06769395e-01 1.27487159e+00 -8.04557383e-01
-2.91502833e-01 2.08164435e-02 8.40580165e-02 4.85412419e-01
-3.52065891e-01 -2.18594689e-02 -3.30485761e-01 -4.41677600e-01
-1.81189463e-01 -3.89431804e-01 5.16030133e-01 -1.14757605e-01
5.18729031e-01 -4.40538764e-01 9.99325886e-03 3.17147344e-01
1.29805422e+00 -2.62823794e-02 6.01080954e-01 4.26070899e-01
5.77255845e-01 5.04043281e-01 5.75134933e-01 -5.87137230e-02
7.74908543e-01 6.33593917e-01 1.24485511e-02 5.57204306e-01
-3.73749614e-01 -4.32761669e-01 5.22382736e-01 1.60764658e+00
1.08242497e-01 -6.56741619e-01 -1.34604347e+00 7.57075548e-01
-1.72047639e+00 -7.94436514e-01 -7.49080539e-01 1.87731314e+00
1.24104786e+00 1.75315142e-01 -6.68485403e-01 -5.20581342e-02
5.93055785e-01 -2.16753885e-01 2.21716374e-01 -1.28158307e+00
-2.40319580e-01 4.17865068e-01 5.44824190e-02 6.89185321e-01
-6.69405103e-01 1.51427782e+00 6.50808287e+00 5.26182353e-01
-6.68597937e-01 5.82313001e-01 1.43822152e-02 2.22991183e-01
-2.02679113e-01 2.60554075e-01 -1.09653318e+00 5.90505004e-01
1.18915129e+00 -1.51921302e-01 2.75916696e-01 4.25898463e-01
2.02525914e-01 -2.91660786e-01 -9.43909883e-01 5.04435062e-01
1.09230153e-01 -1.19285405e+00 8.55581276e-03 -5.26540518e-01
6.46836877e-01 3.97292376e-01 -4.58603919e-01 9.88482654e-01
5.81424832e-01 -7.15441048e-01 8.00504744e-01 2.07507491e-01
7.92121530e-01 -5.91091812e-01 1.08651996e+00 5.72254658e-01
-6.01105094e-01 5.13211451e-02 -3.98985893e-01 -3.66605997e-01
4.60263044e-01 2.51134992e-01 -8.42150688e-01 5.54221213e-01
8.43526840e-01 8.24996233e-01 -9.44202900e-01 7.84827769e-01
-7.84567118e-01 1.01454377e+00 -1.19219817e-01 -1.80768833e-01
1.18470840e-01 1.11865096e-01 8.22076440e-01 1.72790933e+00
3.20477813e-01 3.20981145e-01 1.02200452e-02 5.74050307e-01
-1.90643176e-01 4.49967682e-01 -2.30946854e-01 6.96273670e-02
9.19203877e-01 8.37404191e-01 -1.38928428e-01 -3.38440329e-01
-4.35934961e-01 1.03745520e+00 1.01364553e+00 -3.22247483e-02
-2.75505930e-01 -5.01692116e-01 4.72502202e-01 -9.14184600e-02
-3.11036911e-02 -3.14258605e-01 -3.74710858e-01 -1.18014574e+00
1.84042111e-01 -1.22925508e+00 4.00402516e-01 -8.33799541e-01
-1.32921207e+00 5.73774159e-01 -1.73901111e-01 -6.81103289e-01
-2.68433541e-01 -5.27070820e-01 -9.35337424e-01 1.11168873e+00
-1.34425271e+00 -1.09151924e+00 1.79590642e-01 4.87866312e-01
5.84329545e-01 -8.10456574e-02 1.10527837e+00 4.29779530e-01
-5.96030116e-01 9.78019059e-01 9.61382464e-02 3.29891711e-01
1.02494752e+00 -1.47262168e+00 8.19857478e-01 1.24018192e+00
-1.17399395e-01 9.68082905e-01 6.95147455e-01 -1.02593732e+00
-9.03083742e-01 -1.07720780e+00 2.25812960e+00 -7.86703944e-01
7.38515973e-01 -3.18604201e-01 -1.33662844e+00 9.42707658e-01
3.23353469e-01 -2.99330324e-01 4.49804217e-01 6.65523946e-01
-1.96034402e-01 2.39008769e-01 -1.09569216e+00 4.36029673e-01
1.22359359e+00 -6.61491811e-01 -1.23400235e+00 4.52562809e-01
8.45535755e-01 -9.08157825e-01 -8.96372139e-01 7.66838863e-02
-1.21458806e-01 -6.70965552e-01 3.89454961e-01 -7.46807039e-01
5.55940628e-01 1.79892611e-02 7.52436817e-02 -1.89690137e+00
-4.11151797e-01 -6.86240375e-01 2.24716619e-01 1.71386659e+00
6.96148217e-01 -6.53692961e-01 -4.95867170e-02 6.72780156e-01
-1.06827271e+00 -2.42869511e-01 -1.18846989e+00 -8.11184585e-01
5.45962870e-01 -6.89893246e-01 2.51460969e-01 8.45060170e-01
4.76784527e-01 5.12802958e-01 -1.24844834e-01 -5.21621220e-02
3.09576064e-01 -5.90158880e-01 3.66376489e-01 -1.15386224e+00
-1.67236239e-01 -1.50074467e-01 -1.21998917e-02 -5.53279281e-01
3.72573167e-01 -1.39043796e+00 4.94674683e-01 -1.46882319e+00
2.64539104e-03 -7.92513311e-01 3.18854511e-01 7.30865836e-01
-4.51743484e-01 -1.14088453e-01 2.90207595e-01 2.50761747e-01
-5.83835006e-01 5.06417453e-01 9.41221535e-01 1.60700485e-01
-1.03629373e-01 -3.29970807e-01 -7.53337920e-01 5.13093770e-01
7.99820542e-01 -8.13234389e-01 1.97160468e-01 -1.20585120e+00
5.91305733e-01 -4.68450710e-02 1.12023205e-02 -8.13353419e-01
3.54174674e-01 1.29872829e-01 -2.79405564e-01 -1.48158789e-01
-3.04174155e-01 -1.90916955e-01 -1.79617614e-01 3.39459002e-01
-3.99795890e-01 5.09082735e-01 4.41516101e-01 2.07611203e-01
-2.12100521e-01 -5.84317267e-01 6.56382918e-01 -4.46450740e-01
-8.77723992e-01 -2.48520464e-01 -7.73405790e-01 8.34902585e-01
3.59540582e-01 6.90320879e-02 -7.45158017e-01 -2.16155429e-03
-5.38078010e-01 5.69388747e-01 2.79949754e-01 6.33067489e-01
3.16630304e-01 -9.19898987e-01 -1.05758011e+00 1.56984136e-01
5.24911284e-01 -1.50227845e-01 1.52202979e-01 9.71056998e-01
-4.81613755e-01 7.13114858e-01 9.42490399e-02 -4.91746932e-01
-1.40423727e+00 5.24349622e-02 3.77779812e-01 -5.72428644e-01
-7.04531789e-01 1.14131010e+00 -6.89973712e-01 -1.14165449e+00
-6.31844178e-02 -2.40791559e-01 7.39978179e-02 -3.10705900e-01
4.81891066e-01 2.37863868e-01 7.55805790e-01 -6.44575953e-01
-4.98769283e-01 4.37014587e-02 -5.57176173e-01 7.70698488e-02
1.04770708e+00 -4.87940609e-01 -4.26803231e-01 3.14405829e-01
7.77534664e-01 1.22386582e-01 -4.87013787e-01 -2.80633181e-01
5.21496415e-01 -2.60990351e-01 -1.17761485e-01 -1.46457267e+00
-3.74093920e-01 8.92305613e-01 5.62615879e-02 -2.32530028e-01
5.49146116e-01 -1.83445528e-01 9.32102501e-01 6.23765349e-01
6.72157049e-01 -1.55589497e+00 -1.85589716e-01 1.57282174e+00
1.00514138e+00 -1.30772853e+00 -4.72677827e-01 -3.69826287e-01
-7.38984883e-01 8.38310063e-01 9.62935209e-01 2.12270409e-01
3.61530036e-01 -3.88229154e-02 1.62270829e-01 -3.34407032e-01
-1.08612442e+00 1.46701008e-01 -4.66740541e-02 5.25932074e-01
1.05597138e+00 1.33650854e-01 -1.04538286e+00 5.04438996e-01
-6.05891824e-01 -2.09486768e-01 7.31770456e-01 9.36158836e-01
-3.66231769e-01 -1.25919235e+00 5.57336211e-02 2.52046436e-01
-6.79843605e-01 -7.95415998e-01 -3.26081246e-01 1.04482305e+00
1.06190115e-01 1.16498315e+00 5.63420393e-02 -2.75878515e-03
5.27478635e-01 5.15937269e-01 4.22240049e-01 -1.21426797e+00
-1.40518093e+00 -4.40891445e-01 6.53362691e-01 -6.18836880e-01
-2.64674634e-01 -9.02625799e-01 -1.42926860e+00 -6.50754333e-01
-4.72177118e-01 3.87724370e-01 4.43009615e-01 1.21247494e+00
5.07249594e-01 3.07156473e-01 -1.38948917e-01 -2.46394381e-01
-7.30298340e-01 -1.55101717e+00 -2.70703673e-01 4.87825334e-01
-9.90331843e-02 -2.52695799e-01 -4.39283490e-01 -2.56149620e-01] | [11.085907936096191, 10.716623306274414] |
05e4aa11-fab3-4175-9bbb-cef76cfcad0d | variational-auto-encoding-of-protein | 1712.03346 | null | http://arxiv.org/abs/1712.03346v3 | http://arxiv.org/pdf/1712.03346v3.pdf | Variational auto-encoding of protein sequences | Proteins are responsible for the most diverse set of functions in biology.
The ability to extract information from protein sequences and to predict the
effects of mutations is extremely valuable in many domains of biology and
medicine. However the mapping between protein sequence and function is complex
and poorly understood. Here we present an embedding of natural protein
sequences using a Variational Auto-Encoder and use it to predict how mutations
affect protein function. We use this unsupervised approach to cluster natural
variants and learn interactions between sets of positions within a protein.
This approach generally performs better than baseline methods that consider no
interactions within sequences, and in some cases better than the
state-of-the-art approaches that use the inverse-Potts model. This generative
model can be used to computationally guide exploration of protein sequence
space and to better inform rational and automatic protein design. | ['Eric Kelsic', 'George M. Church', 'Sam Sinai', 'Martin A. Nowak'] | 2017-12-09 | null | null | null | null | ['protein-design'] | ['medical'] | [ 7.08187878e-01 7.54772276e-02 -8.50956589e-02 -4.44753885e-01
-3.26830029e-01 -8.89407992e-01 4.24693495e-01 4.02625471e-01
-4.15241420e-01 1.28961849e+00 2.81787604e-01 -5.36507308e-01
-4.74952208e-03 -4.59671676e-01 -1.02195716e+00 -1.19298065e+00
3.31443138e-02 8.31952274e-01 2.34482720e-01 -3.04729432e-01
2.00775847e-01 5.25498092e-01 -1.38427877e+00 4.73991483e-01
9.94005740e-01 1.19524494e-01 6.42835617e-01 6.46042585e-01
-2.06234977e-01 4.15918410e-01 -3.10058653e-01 -3.35992754e-01
-2.72027314e-01 -9.13499951e-01 -6.49501264e-01 -2.63925493e-01
-2.19137475e-01 3.89737576e-01 2.94913556e-02 8.55509222e-01
4.60713267e-01 1.49385020e-01 9.88757968e-01 -4.54257488e-01
-9.01752174e-01 1.45696029e-01 -3.01914155e-01 1.65894583e-01
3.13019276e-01 3.02973509e-01 1.36572838e+00 -7.95524895e-01
1.14905751e+00 1.17345619e+00 5.10320246e-01 6.09599531e-01
-1.95357537e+00 -8.88351351e-02 -1.51089534e-01 2.25063071e-01
-1.11561179e+00 -1.10102654e-01 3.37018937e-01 -7.47722387e-01
1.54283333e+00 -4.92050461e-02 7.33608961e-01 9.69978511e-01
6.61174119e-01 5.80363214e-01 5.86976111e-01 -3.28151911e-01
2.78700620e-01 -2.70061374e-01 -8.07861388e-02 8.74264300e-01
1.72928479e-02 -1.21001787e-02 -5.76275706e-01 -7.96596229e-01
1.04594171e-01 3.41152668e-01 -2.96692550e-01 -7.03447759e-01
-1.12449503e+00 1.01337397e+00 9.35901552e-02 1.89459682e-01
-7.42442071e-01 9.90704671e-02 1.11998349e-01 -1.14979282e-01
2.99248099e-01 6.75643265e-01 -9.34900701e-01 -3.28401297e-01
-8.57186556e-01 4.70965415e-01 8.05981338e-01 2.48735920e-01
8.81629825e-01 -3.87720972e-01 1.02194346e-01 9.49987113e-01
4.36948836e-01 9.02827457e-02 5.12803793e-01 -6.70666575e-01
-3.67115289e-01 5.35463870e-01 1.28359139e-01 -3.95432830e-01
-1.54353693e-01 6.70487508e-02 -2.48832494e-01 4.34168845e-01
3.63497585e-01 5.69175407e-02 -9.36229408e-01 1.81572783e+00
4.40964103e-01 -2.12978013e-02 1.43437639e-01 5.22925377e-01
3.18785906e-01 7.31945097e-01 3.07658464e-01 -3.34986448e-01
1.13605833e+00 -3.07365924e-01 -6.54294193e-01 1.20322265e-01
6.67271912e-01 -6.51081204e-01 7.55637825e-01 4.57040191e-01
-9.61367428e-01 -2.42142975e-01 -1.10811281e+00 -2.26929709e-01
-5.92959702e-01 -1.85769260e-01 6.51864052e-01 5.24054766e-01
-7.47927129e-01 1.17359018e+00 -1.06332409e+00 -2.12834105e-01
5.79018474e-01 5.78767717e-01 -4.29166645e-01 7.55369104e-03
-1.08841562e+00 1.26035321e+00 5.60764968e-01 -3.45762074e-01
-6.52580798e-01 -9.46577489e-01 -6.63837254e-01 -6.22988082e-02
2.04824582e-01 -5.33335268e-01 9.78742838e-01 -7.09142148e-01
-1.33164871e+00 8.68605256e-01 -7.51009881e-01 -6.30409360e-01
-4.49329615e-02 -1.74571440e-01 1.41968966e-01 -1.12833828e-01
-2.49742910e-01 7.13874102e-01 3.05712342e-01 -7.99525559e-01
-2.31898710e-01 -3.84298384e-01 -4.06653285e-01 1.03325330e-01
2.93989509e-01 -1.76013693e-01 4.82096337e-02 -4.47582066e-01
-2.25591958e-01 -1.12175131e+00 -4.90862250e-01 4.96312939e-02
-3.23383570e-01 -3.39994013e-01 5.83073676e-01 -7.28273809e-01
8.48962724e-01 -1.77228713e+00 9.57371891e-01 7.89297894e-02
2.91638523e-01 4.65731561e-01 2.30136216e-01 9.50518191e-01
-4.09383923e-01 1.76602036e-01 -5.76342404e-01 2.93530583e-01
-1.31568417e-01 4.10367638e-01 -6.62337476e-03 6.18359983e-01
4.81214583e-01 1.05174196e+00 -8.32812905e-01 -7.39428848e-02
2.79475480e-01 9.64787185e-01 -7.84671664e-01 2.37100735e-01
-8.47852767e-01 4.92371529e-01 -5.99509656e-01 3.25460613e-01
2.28117108e-01 -5.40046871e-01 8.34305942e-01 -5.85086979e-02
-3.34198878e-04 3.75055790e-01 -5.19070029e-01 1.53234339e+00
8.33695680e-02 5.43931842e-01 -3.86460006e-01 -1.17169809e+00
7.55166590e-01 2.81584829e-01 7.78877854e-01 1.99559912e-01
-9.36745033e-02 1.52548344e-03 5.53666174e-01 -4.46623534e-01
-8.52126703e-02 -4.33977544e-01 3.15828145e-01 2.84171492e-01
1.82207108e-01 6.36393502e-02 2.65251398e-01 1.44659996e-01
1.27348351e+00 7.01354146e-01 7.45486498e-01 -2.30381936e-01
5.01083076e-01 2.64976501e-01 7.94556379e-01 2.92485386e-01
5.25458902e-02 4.22899395e-01 7.94158936e-01 -4.95667607e-01
-1.38819325e+00 -9.97610509e-01 -1.51265398e-01 1.05826807e+00
-1.63418606e-01 -4.18487370e-01 -6.41806424e-01 -4.06012177e-01
2.46520996e-01 7.15729475e-01 -7.42836654e-01 -5.80545366e-01
-3.48499149e-01 -1.16711056e+00 3.17234397e-01 2.28144601e-01
-3.84032875e-01 -1.25795805e+00 -5.09781241e-01 5.07592916e-01
2.46772245e-02 -5.00659108e-01 -4.30875868e-01 7.17450678e-01
-5.50905168e-01 -1.37655580e+00 -6.42114043e-01 -6.78019106e-01
4.73110050e-01 -3.18703771e-01 1.17647755e+00 -6.70443475e-02
-8.27976704e-01 -2.04938710e-01 -2.76740283e-01 -6.28694415e-01
-8.13811898e-01 -1.52211964e-01 7.18657449e-02 -1.17389768e-01
8.46123278e-01 -7.43529201e-01 -5.59713244e-01 -2.17035655e-02
-9.98096168e-01 -2.33773012e-02 5.38410723e-01 1.18155313e+00
8.80433619e-01 -2.69327134e-01 3.68164003e-01 -1.29629612e+00
6.54157996e-01 -5.08254111e-01 -5.68114936e-01 4.96398121e-01
-5.44207990e-01 8.29561651e-01 6.01528168e-01 -2.88727820e-01
-8.82703245e-01 7.08755314e-01 -3.24527919e-01 -7.76291564e-02
-2.69092828e-01 4.99975264e-01 -3.45025092e-01 1.57145977e-01
6.26061559e-01 5.04234672e-01 2.86382735e-01 -5.24354398e-01
3.91541839e-01 3.66048336e-01 1.22208968e-01 -3.77538651e-01
2.06249774e-01 3.19745213e-01 1.74976051e-01 -1.04885101e+00
-4.70826834e-01 -3.54117334e-01 -8.88936698e-01 2.34181002e-01
1.20799494e+00 -4.20620322e-01 -1.00340772e+00 -1.06959157e-01
-1.08222103e+00 -2.19327837e-01 -3.66080143e-02 2.51156956e-01
-8.62851858e-01 5.35681069e-01 -4.11105007e-01 -6.67670906e-01
4.76025008e-02 -1.52052104e+00 1.13638115e+00 1.33026810e-02
-4.84990567e-01 -9.81394351e-01 4.85781968e-01 2.11915046e-01
-9.91143510e-02 3.91520888e-01 1.34491205e+00 -7.25937843e-01
-4.34909910e-01 2.73996413e-01 4.72605318e-01 9.69087407e-02
3.61254483e-01 2.00263306e-01 -5.00393867e-01 -1.07070701e-02
-4.09603029e-01 -3.47468257e-01 1.20794356e+00 6.74942851e-01
9.41578567e-01 -1.90850813e-02 -6.38387084e-01 4.47463721e-01
1.33459306e+00 4.81127292e-01 7.52617598e-01 -3.20992456e-03
4.56223696e-01 7.73011446e-01 5.45077980e-01 2.48850822e-01
-1.91378757e-01 8.13741922e-01 2.03044638e-01 1.89495087e-01
3.27741027e-01 -3.34822506e-01 3.75359058e-01 6.30673468e-02
-1.14445873e-01 -5.08523405e-01 -9.07101989e-01 3.63260657e-01
-1.98391724e+00 -1.17174423e+00 5.93707748e-02 1.94373751e+00
1.25587785e+00 1.71561874e-02 1.53026551e-01 -3.22175860e-01
6.91057503e-01 -1.54367447e-01 -1.10213506e+00 -4.88223672e-01
-3.00148457e-01 4.75832582e-01 5.98755240e-01 6.04960859e-01
-7.31752992e-01 1.06390202e+00 7.63193130e+00 5.00417471e-01
-5.99847913e-01 -2.47985706e-01 5.19019961e-01 -8.61466005e-02
-3.19262058e-01 1.18135758e-01 -7.96689451e-01 4.92435724e-01
1.23707342e+00 -2.93232091e-02 5.21941006e-01 5.27472079e-01
4.40025270e-01 7.93071836e-02 -1.12402749e+00 6.20628238e-01
-1.48689881e-01 -1.84007657e+00 -2.94822335e-01 3.73319298e-01
5.43188751e-01 2.49527112e-01 -1.28989711e-01 -3.52997184e-01
8.53560746e-01 -1.27783895e+00 -5.05229039e-03 8.36199701e-01
4.61951196e-01 -9.01704252e-01 5.17771959e-01 4.53823358e-01
-6.19269133e-01 2.94931471e-01 -5.53807557e-01 2.11985722e-01
2.20310032e-01 7.22094476e-01 -9.85654771e-01 -1.07003026e-01
3.36879164e-01 8.21559429e-01 -4.37036827e-02 5.27573466e-01
-2.14314699e-01 5.76674044e-01 -1.04909733e-01 -5.57347298e-01
4.48218286e-02 -5.87584078e-01 4.18998986e-01 9.61272359e-01
-1.22447319e-01 2.36192301e-01 2.13576108e-01 1.10398424e+00
8.92908685e-03 9.04543176e-02 -5.36732078e-01 -7.44844854e-01
2.69484252e-01 7.68645108e-01 -5.51631927e-01 -3.33925247e-01
-3.55718851e-01 1.17740357e+00 5.44167936e-01 3.93233180e-01
-7.56205678e-01 -2.71065176e-01 1.47766900e+00 1.06908688e-02
7.40877628e-01 -3.18794936e-01 3.99403214e-01 -9.05678332e-01
-1.86848685e-01 -1.00519538e+00 6.11122064e-02 -6.24930799e-01
-1.22361720e+00 1.18659459e-01 -3.37951303e-01 -4.89603102e-01
-4.42240566e-01 -1.01737142e+00 -1.09346367e-01 1.11236072e+00
-1.06700242e+00 -5.82963169e-01 6.74513876e-01 -9.66261700e-02
6.19458616e-01 -2.25331888e-01 1.17253518e+00 -3.45503181e-01
-3.92158419e-01 1.12982541e-01 9.34888422e-01 -4.30159718e-01
5.53086340e-01 -1.57474017e+00 6.25846863e-01 3.16986918e-01
2.40876272e-01 1.05403697e+00 1.26454437e+00 -9.89518881e-01
-1.42450130e+00 -8.92381907e-01 8.23791146e-01 -6.77772284e-01
3.62741172e-01 -5.60597658e-01 -1.19257522e+00 5.25169492e-01
8.97386745e-02 -1.16556332e-01 1.39049292e+00 1.19058169e-01
-2.93152601e-01 5.77678978e-01 -1.14064622e+00 4.98843580e-01
8.71041536e-01 -5.00119865e-01 -7.90612221e-01 5.92721939e-01
6.42196655e-01 -6.37038145e-03 -9.32544708e-01 2.07625274e-02
7.58657575e-01 -1.00707102e+00 1.11399055e+00 -1.24752212e+00
5.24828851e-01 -4.30435359e-01 -1.74502537e-01 -1.44021511e+00
-6.39934719e-01 -5.31658411e-01 -8.38245302e-02 6.01982296e-01
8.27074349e-01 -5.06412268e-01 1.07225966e+00 2.93408543e-01
-5.15739201e-03 -9.11150575e-01 -5.55855989e-01 -3.50088179e-01
2.07067862e-01 7.83991218e-02 3.93627524e-01 6.84020579e-01
2.39676446e-01 4.95683819e-01 -3.76791358e-01 -1.02606282e-01
3.83552372e-01 1.68963466e-02 3.94670904e-01 -1.20963061e+00
-7.74438083e-01 -2.99299389e-01 -6.77832007e-01 -8.67476046e-01
3.01208138e-01 -9.39277828e-01 1.79131672e-01 -1.41121602e+00
4.01600361e-01 3.86729032e-01 -3.27353805e-01 3.45872492e-01
-3.54917884e-01 8.01226497e-02 -3.79517823e-01 4.90354039e-02
-3.13194633e-01 6.39256239e-01 6.56594276e-01 3.04512251e-02
-4.96186502e-02 -3.03817242e-01 -5.89685380e-01 4.73451287e-01
7.06123114e-01 -5.87347746e-01 -3.75112325e-01 2.69409746e-01
2.79812753e-01 -2.37629712e-01 -8.48403666e-03 -3.93613577e-01
-2.94861674e-01 -4.58283067e-01 6.24641359e-01 -6.60930991e-01
4.32240188e-01 -5.22242785e-01 4.07646328e-01 8.03625405e-01
-6.16747200e-01 -4.89399843e-02 -8.92934203e-03 9.57903981e-01
4.81871292e-02 -1.95648566e-01 8.78523648e-01 -3.70687664e-01
-3.51669669e-01 2.08534569e-01 -8.05954158e-01 -1.27690524e-01
1.16810405e+00 -1.93148300e-01 2.58081675e-01 -1.62739098e-01
-1.04299664e+00 -1.08573310e-01 8.06769967e-01 2.72742867e-01
6.02053761e-01 -9.10290062e-01 -5.76047599e-01 2.12429911e-01
5.30685671e-02 -5.73107302e-01 5.53399771e-02 4.35405910e-01
-8.14062893e-01 6.16724610e-01 -8.70888308e-02 -5.30079901e-01
-1.63969696e+00 6.13803029e-01 4.13005650e-01 -1.87530890e-01
-4.16328430e-01 8.30653667e-01 3.21044564e-01 -4.80965734e-01
-3.28635931e-01 1.95925593e-01 -1.72075748e-01 -2.79976606e-01
5.38369298e-01 -1.00584343e-01 -1.03673376e-01 -7.71288455e-01
-4.60701257e-01 2.92675018e-01 -3.41971010e-01 1.93606168e-01
1.73002720e+00 2.90350229e-01 -2.59643108e-01 5.55904984e-01
1.28109765e+00 -2.32706159e-01 -1.35747063e+00 -4.19713557e-02
2.39751905e-01 -3.12560797e-01 -2.55959570e-01 -8.72668862e-01
-3.94827962e-01 8.73647749e-01 5.12083054e-01 -2.89748818e-01
6.20927274e-01 2.64716595e-01 5.70724189e-01 4.69833732e-01
9.25772488e-02 -8.74702811e-01 -2.52221286e-01 2.74306744e-01
3.57670754e-01 -1.11438513e+00 2.41301265e-02 -2.85533100e-01
-6.98737621e-01 1.04102278e+00 -1.63258407e-02 -1.25870049e-01
5.79762876e-01 3.19128603e-01 -2.06511870e-01 -3.31299633e-01
-1.18623030e+00 -1.83374196e-01 3.61951321e-01 6.52503967e-01
1.25965834e+00 4.10192236e-02 -5.94341993e-01 2.13184908e-01
9.56105590e-02 -8.23865682e-02 1.78575814e-01 1.07799387e+00
-6.93906188e-01 -1.87320387e+00 -6.26045763e-02 5.81135929e-01
-8.86171877e-01 -3.15388709e-01 -1.02906728e+00 2.00946257e-01
1.86728179e-01 6.60331428e-01 -3.64081591e-01 -7.74338096e-02
2.58889496e-02 7.96052337e-01 6.04288876e-01 -8.53880525e-01
-3.36970299e-01 1.80074438e-01 -7.69270360e-02 -4.56207931e-01
-4.45796609e-01 -1.09070075e+00 -1.52952218e+00 -3.58842015e-02
-4.44207162e-01 4.21357930e-01 5.50947309e-01 9.80285287e-01
7.60778725e-01 6.30153656e-01 1.02723032e-01 -5.07916212e-01
-4.92674112e-01 -6.72676980e-01 -5.84739506e-01 5.23610830e-01
1.50291234e-01 -7.24508584e-01 7.75457621e-02 4.61937219e-01] | [4.740192890167236, 5.600770950317383] |
647c09e4-1b74-4ae0-98c4-72e8eb3b54b2 | bbc-oxford-british-sign-language-dataset | 2111.03635 | null | https://arxiv.org/abs/2111.03635v1 | https://arxiv.org/pdf/2111.03635v1.pdf | BBC-Oxford British Sign Language Dataset | In this work, we introduce the BBC-Oxford British Sign Language (BOBSL) dataset, a large-scale video collection of British Sign Language (BSL). BOBSL is an extended and publicly released dataset based on the BSL-1K dataset introduced in previous work. We describe the motivation for the dataset, together with statistics and available annotations. We conduct experiments to provide baselines for the tasks of sign recognition, sign language alignment, and sign language translation. Finally, we describe several strengths and limitations of the data from the perspectives of machine learning and linguistics, note sources of bias present in the dataset, and discuss potential applications of BOBSL in the context of sign language technology. The dataset is available at https://www.robots.ox.ac.uk/~vgg/data/bobsl/. | ['Andrew Zisserman', 'Andrew McParland', 'Rob Cooper', 'Bencie Woll', 'Neil Fox', 'Himel Chowdhury', 'Triantafyllos Afouras', 'Hannah Bull', 'Liliane Momeni', 'Gül Varol', 'Samuel Albanie'] | 2021-11-05 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 3.71287167e-02 -1.16675124e-01 -4.29485559e-01 -4.53283578e-01
-9.92802560e-01 -5.88198483e-01 5.50523579e-01 -7.80500174e-01
-8.94886911e-01 5.89593887e-01 1.14289510e+00 -2.14493960e-01
1.51185066e-01 1.89338550e-01 -5.64952970e-01 -5.06655693e-01
1.72265559e-01 2.75459498e-01 3.69717360e-01 -3.07669520e-01
2.34776139e-01 3.69275272e-01 -1.50896096e+00 4.69500363e-01
4.29926455e-01 4.59950417e-01 1.80801898e-01 9.04055059e-01
1.84721142e-01 9.77143407e-01 -2.16621265e-01 -3.96560490e-01
3.83049458e-01 -6.39811218e-01 -1.15292585e+00 -1.59568876e-01
1.27732289e+00 -5.99141419e-01 -6.38568878e-01 7.80067861e-01
9.56225991e-01 -2.22113311e-01 5.07339656e-01 -1.29434502e+00
-5.97271025e-01 5.24990499e-01 -8.90514329e-02 8.06311667e-02
4.61421728e-01 5.81270695e-01 1.11007357e+00 -7.26348937e-01
1.52940106e+00 1.05088675e+00 6.42103672e-01 1.17618382e+00
-3.99852455e-01 -5.71226418e-01 1.21537507e-01 3.42655063e-01
-9.38317716e-01 -7.84597814e-01 4.10395890e-01 -6.96915984e-01
7.99036741e-01 8.94351229e-02 1.10132587e+00 1.32750988e+00
-5.33378720e-01 1.54915798e+00 1.24464524e+00 -7.42882907e-01
-2.83548266e-01 -5.20584524e-01 2.61823833e-01 8.95314455e-01
2.61536777e-01 2.62225270e-01 -7.83768654e-01 3.96144465e-02
7.59728670e-01 -7.17738032e-01 -4.17130411e-01 -5.46287119e-01
-1.60248566e+00 4.37344909e-01 2.72560567e-01 2.35419825e-01
-7.23114014e-02 3.74628752e-01 6.21141374e-01 3.43464673e-01
-2.52616137e-01 9.49589238e-02 -6.20865345e-01 -6.58095419e-01
-3.15123588e-01 4.38250005e-01 5.31040192e-01 1.30203640e+00
-9.66311693e-02 -1.78137347e-01 6.55142739e-02 9.43267286e-01
5.53328693e-01 7.11405635e-01 6.31582439e-01 -1.05451405e+00
5.91268778e-01 1.39041096e-01 -9.48753133e-02 -9.53644216e-02
-4.32143450e-01 6.59664392e-01 -1.25511661e-01 3.98675919e-01
9.50852334e-01 -2.00845420e-01 -1.30977488e+00 1.65555489e+00
3.58122960e-02 -5.56797944e-02 1.72958508e-01 1.21614265e+00
1.38091004e+00 -1.89747453e-01 2.63164580e-01 3.65592748e-01
1.18284523e+00 -1.06524289e+00 -4.77629274e-01 -1.62275024e-02
9.98351097e-01 -7.70185351e-01 1.14975965e+00 1.63270548e-01
-8.73862267e-01 -9.52880830e-02 -5.85364878e-01 -5.49518466e-01
-3.11881095e-01 6.33509696e-01 7.02003062e-01 3.02987456e-01
-1.11073351e+00 -1.94618806e-01 -1.04571033e+00 -1.10407221e+00
4.91980493e-01 1.68911107e-02 -6.75341904e-01 -1.55245215e-01
-7.26759553e-01 1.14910769e+00 1.52930751e-01 2.02560917e-01
-1.94199502e-01 -1.02990806e-01 -8.99051785e-01 -1.14688528e+00
-4.66509871e-02 -2.99456477e-01 1.80958819e+00 -8.94621909e-01
-1.51201034e+00 1.51247549e+00 -1.78562164e-01 -2.58140951e-01
9.42569673e-01 -3.82415414e-01 -5.29650688e-01 4.73572016e-02
-8.44603498e-03 9.12048340e-01 2.94574052e-01 -6.77445412e-01
-1.01447058e+00 -1.29860699e-01 -9.22166035e-02 1.28445983e-01
4.96778339e-01 6.57228410e-01 -6.62974417e-01 -7.24625647e-01
1.59908935e-01 -1.27733541e+00 -1.03641018e-01 3.44216138e-01
-1.10599883e-01 -1.34039178e-01 3.81585538e-01 -1.09703231e+00
8.42439055e-01 -2.02528071e+00 2.07396820e-01 1.01604097e-01
2.70516742e-02 4.18472290e-01 -4.11998272e-01 5.02039313e-01
8.83716866e-02 -1.48252591e-01 -4.45701391e-01 -7.31223449e-02
-3.49259824e-02 3.95691365e-01 -1.02374695e-01 6.67021990e-01
-1.81462951e-02 1.24629855e+00 -9.83881295e-01 -4.38411772e-01
2.52043039e-01 1.92221344e-01 -4.37474787e-01 -2.54922181e-01
-8.01653638e-02 7.07870722e-01 -8.17718655e-02 1.09944952e+00
1.73946485e-01 3.03229094e-01 9.27721411e-02 -1.04616329e-01
-3.17815900e-01 3.75405550e-01 -8.39362621e-01 1.76787221e+00
-1.05843179e-01 1.41266859e+00 1.28392249e-01 -5.33356309e-01
3.53515506e-01 3.79847556e-01 4.64897364e-01 -4.17405814e-01
2.04463825e-01 8.10486555e-01 3.17543775e-01 -9.51039970e-01
2.68748164e-01 2.19452977e-01 -1.40977919e-01 3.39550585e-01
9.39937755e-02 -2.33837485e-01 5.87061226e-01 -1.28645524e-02
9.71315503e-01 5.62981844e-01 5.46657920e-01 1.04403980e-01
4.05811906e-01 3.46604794e-01 3.48557085e-01 4.83054549e-01
-8.33575606e-01 6.72766685e-01 1.76586956e-01 -3.16660881e-01
-1.01340723e+00 -8.86396289e-01 -1.96826950e-01 7.98151910e-01
-4.66539264e-02 -4.55392569e-01 -2.92237848e-01 -6.79292917e-01
1.86405376e-01 1.74789518e-01 -4.94415015e-01 5.46945333e-01
-1.07984900e+00 -2.28141233e-01 1.22357941e+00 8.76044214e-01
4.29010302e-01 -1.58440876e+00 -9.02996302e-01 -4.01565731e-01
-2.18214989e-01 -1.27803981e+00 -6.02770627e-01 -5.39983273e-01
-5.17639935e-01 -1.53325915e+00 -9.97732043e-01 -1.33875465e+00
6.23442829e-01 -1.71925321e-01 5.57713807e-01 7.25245178e-02
-4.21145618e-01 8.01334321e-01 -7.77873099e-01 -5.49743414e-01
-7.37696588e-01 -1.07936390e-01 7.13467970e-02 -5.24182618e-01
6.32701159e-01 4.31598201e-02 -2.75818944e-01 3.24630857e-01
-5.25448859e-01 3.71937782e-01 5.86381257e-01 9.71006155e-01
3.02121341e-01 -1.16043985e+00 -2.81606521e-02 -3.92595977e-01
3.75682414e-01 2.51087517e-01 -7.02692926e-01 2.72781849e-01
3.30437208e-03 -1.12782247e-01 -3.35121870e-01 -2.47339219e-01
-8.04299533e-01 2.82563180e-01 -3.44812900e-01 2.08215773e-01
-4.86406386e-01 2.71394491e-01 -7.64957592e-02 -3.24167162e-01
4.68359649e-01 1.09056339e-01 2.29460999e-01 -6.12293065e-01
7.23749995e-01 1.15762174e+00 8.30299258e-01 -3.76258731e-01
2.57242203e-01 6.48771048e-01 -1.91469118e-01 -9.83290851e-01
-3.15603167e-01 -7.36512780e-01 -1.29767609e+00 -5.55796385e-01
6.78128600e-01 -8.16446245e-01 -3.50707889e-01 1.07999492e+00
-8.67040098e-01 -9.00691271e-01 -4.44286942e-01 8.60108495e-01
-1.04135919e+00 3.63943368e-01 -4.77411777e-01 -5.10010362e-01
6.30679578e-02 -1.03914034e+00 1.13844025e+00 -9.56981853e-02
-6.54255152e-01 -5.35470068e-01 3.36409867e-01 4.94974554e-01
3.79185788e-02 3.97583634e-01 3.04102629e-01 -2.24427432e-01
-4.16121274e-01 -3.05376858e-01 -1.12164930e-01 4.90957171e-01
1.04299270e-01 2.07933351e-01 -5.73735297e-01 -1.81214139e-01
-1.01105607e+00 -7.16336012e-01 1.03577435e+00 3.90571862e-01
3.82490724e-01 -9.65420231e-02 -9.91040468e-02 5.85490286e-01
1.03343153e+00 -2.07221918e-02 4.93953943e-01 5.98411202e-01
7.61063933e-01 6.87446535e-01 5.65467417e-01 4.07246947e-02
6.82728708e-01 7.68724918e-01 -3.62722874e-01 3.20695713e-02
-9.10911679e-01 -5.10904551e-01 4.76616651e-01 1.00133932e+00
-6.42443240e-01 9.46172103e-02 -1.31729889e+00 9.43883955e-01
-2.04609656e+00 -1.04738772e+00 -3.21131974e-01 1.63725734e+00
7.72206724e-01 -5.82934439e-01 6.31078243e-01 -4.70467992e-02
3.28068197e-01 5.28126322e-02 -3.81239742e-01 -2.32517421e-01
-7.57274508e-01 6.79474846e-02 9.45178092e-01 6.84561014e-01
-1.31481040e+00 1.50772500e+00 6.48364305e+00 1.10035814e-01
-1.17071581e+00 -1.37954205e-02 -4.10775304e-01 -3.56158108e-01
2.40047872e-01 -6.14627302e-02 -8.04819047e-01 2.35596240e-01
4.62635517e-01 5.01384512e-02 2.99700022e-01 6.28880560e-01
2.91761428e-01 -6.01848476e-02 -9.58814740e-01 1.06166995e+00
3.82670224e-01 -1.04484463e+00 -1.65161043e-01 4.91625369e-02
6.01900041e-01 1.12394333e+00 -3.84334445e-01 -5.72818294e-02
6.08038723e-01 -5.00405550e-01 1.18144143e+00 4.70342368e-01
1.39967978e+00 2.13728100e-01 6.77471817e-01 -1.88934878e-02
-1.07516849e+00 -7.14403540e-02 2.01264769e-01 -1.64223582e-01
7.02254832e-01 -6.06854558e-01 -4.80173379e-01 -4.72859740e-02
8.21766973e-01 1.27512145e+00 -6.38170540e-01 1.33869267e+00
-7.48319983e-01 7.64411449e-01 -5.55748284e-01 -2.73596495e-01
1.98605478e-01 -8.19327310e-02 7.82560825e-01 1.36533785e+00
1.28512278e-01 1.33592978e-01 4.15111519e-02 -5.64675574e-05
1.78047985e-01 2.48132855e-01 -6.65067792e-01 -2.56387979e-01
3.90212424e-02 4.58240122e-01 -3.32662344e-01 -3.08552951e-01
-6.63766563e-01 1.01274729e+00 2.19270274e-01 5.36346734e-01
-4.53212798e-01 -2.56731391e-01 9.88461673e-01 -1.35541663e-01
2.69500613e-01 -4.76832598e-01 -1.46914035e-01 -1.43628860e+00
4.45031673e-01 -8.95343959e-01 5.59299648e-01 -8.81200254e-01
-1.20267463e+00 2.18030676e-01 2.88949478e-02 -1.59306419e+00
-4.52077955e-01 -1.22636938e+00 -1.62554067e-03 5.45442581e-01
-1.40805948e+00 -1.64217806e+00 -4.31712449e-01 5.25443077e-01
4.82653081e-01 -3.99548799e-01 7.21156716e-01 1.65394068e-01
-5.32762967e-02 5.70097268e-01 2.17337668e-01 8.53766799e-01
9.91000175e-01 -1.00334084e+00 8.61548245e-01 8.96914482e-01
4.20244008e-01 2.96378970e-01 3.11358154e-01 -7.26601303e-01
-9.77812290e-01 -6.18462205e-01 1.23378336e+00 -9.59728599e-01
8.54478598e-01 -4.50599454e-02 -1.75931528e-02 1.07150280e+00
-8.17445964e-02 -1.07143305e-01 5.40444672e-01 -3.84329498e-01
-4.51655120e-01 4.68874127e-01 -9.73959923e-01 1.00454867e+00
1.90788758e+00 -6.53937876e-01 -8.52422893e-01 4.15946871e-01
1.17398672e-01 -8.52992833e-01 -4.30147678e-01 3.51180345e-01
1.61711860e+00 -5.53657591e-01 5.70557356e-01 -1.03822422e+00
3.19205582e-01 -3.72545689e-01 -3.42583865e-01 -1.00767398e+00
-3.05809788e-02 -5.08296430e-01 1.85782015e-01 6.96660638e-01
3.51164967e-01 -6.37541294e-01 6.75742745e-01 5.47061384e-01
-1.26250267e-01 -4.39239591e-01 -1.18483090e+00 -8.77198219e-01
1.60251439e-01 -8.76307070e-01 2.75615364e-01 6.79895818e-01
3.87606919e-01 -3.32801253e-01 -4.76394027e-01 -2.40910307e-01
5.03113210e-01 -1.66366920e-02 1.29195035e+00 -8.63016784e-01
-1.53400972e-01 -8.05357575e-01 -8.99737179e-01 -1.21933007e+00
1.40481368e-01 -9.21396911e-01 2.68762052e-01 -1.85812008e+00
-3.22809428e-01 -6.53448030e-02 6.04730397e-02 7.79181302e-01
2.82834917e-01 5.47847271e-01 5.65237999e-01 5.16535640e-01
-2.81301588e-01 1.15837213e-02 1.18398213e+00 -5.80508634e-02
-1.07071109e-01 -8.57283324e-02 -2.54929245e-01 9.11217153e-01
1.06521070e+00 4.34018783e-02 2.76297718e-01 -8.55601430e-01
-3.69989812e-01 -6.07637465e-01 4.99665201e-01 -6.32041693e-01
1.83851257e-01 -4.83372509e-02 -2.39763051e-01 -5.76894641e-01
1.38744488e-01 -4.68214869e-01 -3.45843166e-01 5.73082089e-01
-4.65714216e-01 -6.56597540e-02 1.93943292e-01 7.56831542e-02
-1.66126132e-01 1.31689712e-01 7.65150547e-01 -2.59277701e-01
-1.46537447e+00 6.84646703e-03 -5.76610982e-01 2.35858381e-01
6.62893891e-01 -3.90352398e-01 -3.22831392e-01 -3.60188127e-01
-5.29631615e-01 4.89423603e-01 4.91132289e-01 9.40201938e-01
4.01770711e-01 -1.42407870e+00 -8.23303759e-01 4.30520564e-01
8.32886636e-01 -3.86444449e-01 -9.67914090e-02 9.47218776e-01
-1.10361767e+00 7.16145813e-01 -5.27222991e-01 -2.95551956e-01
-1.91074443e+00 -3.46222371e-01 2.31126174e-01 4.85209107e-01
-1.11289597e+00 9.72747624e-01 -4.69556391e-01 -7.23188639e-01
4.75669473e-01 -8.02594721e-01 -1.71452641e-01 -3.22457522e-01
4.63852823e-01 3.73519808e-01 -4.34195101e-01 -1.50016975e+00
-5.59682786e-01 1.01863301e+00 3.07480961e-01 -6.88934028e-01
1.15846169e+00 -3.57919522e-02 -1.22015141e-01 4.70306605e-01
8.95636141e-01 1.53459921e-01 -7.89026201e-01 -3.72426748e-01
4.34217483e-01 -4.37016100e-01 -3.20771098e-01 -1.10296822e+00
-6.54377341e-01 4.63541746e-01 6.40532374e-01 -9.38524544e-01
8.82912815e-01 5.05662441e-01 8.93464267e-01 4.60475266e-01
4.93180394e-01 -1.41136158e+00 -7.07252562e-01 8.52795780e-01
1.33448207e+00 -1.39471531e+00 -1.84920043e-01 -1.91941202e-01
-9.21171248e-01 9.21664953e-01 4.73992288e-01 -5.74985966e-02
7.95265436e-01 1.97190538e-01 1.00906730e+00 -1.24657802e-01
-6.36803955e-02 -8.16611290e-01 4.94555026e-01 7.90440798e-01
4.63664085e-01 2.47555047e-01 -8.64022851e-01 2.51405895e-01
-7.34259069e-01 5.88089883e-01 4.86908495e-01 1.36919153e+00
1.54915163e-02 -1.18530798e+00 -9.32603925e-02 3.68380755e-01
7.97933191e-02 7.72928074e-02 -9.33586597e-01 1.00881851e+00
3.27042282e-01 4.95173723e-01 -2.06065774e-01 -1.49766997e-01
6.53899848e-01 2.82619029e-01 9.21868563e-01 -2.81396598e-01
-4.64827083e-02 -5.31699881e-02 5.26131094e-01 -6.00792408e-01
-9.72682834e-01 -1.33310974e+00 -1.18388188e+00 2.16398403e-01
1.06525077e-02 -4.78242218e-01 5.99903643e-01 8.93772423e-01
5.39857000e-02 -1.52818874e-01 -4.76425380e-01 -7.47577608e-01
-2.17599630e-01 -1.10403919e+00 -4.32729810e-01 5.25667667e-01
5.39851427e-01 -4.93604243e-01 -2.19164357e-01 4.99580294e-01] | [9.151691436767578, -6.473094940185547] |
3a25fc97-e529-4226-bc1b-be0db5f3c764 | feded-federated-learning-via-ensemble | null | null | https://aclanthology.org/2020.emnlp-main.165 | https://aclanthology.org/2020.emnlp-main.165.pdf | FedED: Federated Learning via Ensemble Distillation for Medical Relation Extraction | Unlike other domains, medical texts are inevitably accompanied by private information, so sharing or copying these texts is strictly restricted. However, training a medical relation extraction model requires collecting these privacy-sensitive texts and storing them on one machine, which comes in conflict with privacy protection. In this paper, we propose a privacy-preserving medical relation extraction model based on federated learning, which enables training a central model with no single piece of private local data being shared or exchanged. Though federated learning has distinct advantages in privacy protection, it suffers from the communication bottleneck, which is mainly caused by the need to upload cumbersome local parameters. To overcome this bottleneck, we leverage a strategy based on knowledge distillation. Such a strategy uses the uploaded predictions of ensemble local models to train the central model without requiring uploading local parameters. Experiments on three publicly available medical relation extraction datasets demonstrate the effectiveness of our method. | ['Weijian Sun', 'Yuantao Xie', 'Yantao Jia', 'Jun Zhao', 'Yubo Chen', 'Dianbo Sui'] | null | null | null | null | emnlp-2020-11 | ['medical-relation-extraction'] | ['medical'] | [ 3.51682276e-01 4.98636484e-01 -4.53718692e-01 -4.00217474e-01
-8.95163000e-01 -6.62222385e-01 2.56851703e-01 4.35375243e-01
-6.21296287e-01 9.99485433e-01 2.32309356e-01 -5.86193681e-01
-2.29341567e-01 -9.12334442e-01 -5.66253841e-01 -9.52383876e-01
3.87540236e-02 2.03668207e-01 -1.35417566e-01 2.64714986e-01
-3.77975345e-01 2.07375512e-01 -6.87696934e-01 4.07654107e-01
9.89360809e-01 9.03831244e-01 -3.02767187e-01 2.90757149e-01
-9.06558335e-02 9.21826005e-01 -4.54107672e-01 -9.75610316e-01
4.02015448e-01 -3.30670923e-01 -1.11823118e+00 -4.51061577e-01
-1.65611103e-01 -4.25310463e-01 -3.79049540e-01 1.04508328e+00
4.71266866e-01 -2.50468701e-01 1.54867768e-01 -1.11493218e+00
-3.29279065e-01 7.17878938e-01 -3.35941494e-01 -3.69167507e-01
2.11344570e-01 3.46363224e-02 9.06443715e-01 -5.15212044e-02
7.59895325e-01 4.69093502e-01 7.34360993e-01 5.69076657e-01
-1.24275422e+00 -7.79024720e-01 -2.58806437e-01 -1.62186965e-01
-1.37561238e+00 -4.43270802e-01 5.87351859e-01 -2.83079058e-01
5.48493147e-01 6.74308419e-01 4.57019269e-01 1.07310212e+00
3.33319575e-01 7.56228745e-01 9.77040350e-01 -1.41104385e-01
2.28416070e-01 5.74580550e-01 -7.74236545e-02 5.47474504e-01
4.75816160e-01 -1.02800451e-01 -5.24171889e-01 -9.74217772e-01
8.41347054e-02 3.83159131e-01 -5.50381541e-01 -4.92078900e-01
-1.02175295e+00 6.77809119e-01 3.09280068e-01 1.85279354e-01
-1.70867085e-01 -4.00524974e-01 5.91843843e-01 4.41168398e-01
6.31162584e-01 1.31899819e-01 -8.31211090e-01 2.44701326e-01
-6.67196929e-01 -7.78050125e-02 1.16268086e+00 8.86281550e-01
8.27483952e-01 -1.02359092e+00 -7.32213119e-03 5.19527793e-01
1.99910730e-01 2.12950855e-01 5.23544729e-01 -4.90902036e-01
7.38189280e-01 7.39122152e-01 4.46780547e-02 -1.06610644e+00
-3.53029430e-01 -2.28236362e-01 -1.28957665e+00 -4.50965106e-01
4.77091283e-01 -5.54909348e-01 -2.76406080e-01 1.74027753e+00
8.49632502e-01 5.39442338e-02 4.20576662e-01 6.28745735e-01
5.07227421e-01 1.82650194e-01 2.67886698e-01 -4.01012331e-01
1.41690290e+00 -7.02749014e-01 -9.31744635e-01 2.41388783e-01
1.17835009e+00 -4.20751065e-01 3.76965344e-01 7.59881660e-02
-6.75044954e-01 3.94079447e-01 -8.44391763e-01 -2.55431205e-01
-5.65893352e-01 -2.40256101e-01 8.85588109e-01 9.17244136e-01
-7.86569774e-01 7.92422831e-01 -1.13745534e+00 -4.94449958e-02
9.45920706e-01 5.38459957e-01 -9.15549576e-01 1.69322476e-01
-1.36428535e+00 5.83381295e-01 4.95510310e-01 -5.83961979e-03
-8.82711485e-02 -9.41550136e-01 -8.31221104e-01 1.70514390e-01
5.42289495e-01 -9.01314795e-01 1.01667976e+00 -4.70353961e-01
-1.46240687e+00 8.92028034e-01 8.75209644e-02 -6.25979602e-01
9.47598100e-01 -7.78386369e-02 -4.46647882e-01 -4.31960672e-02
-1.75410181e-01 -1.86449096e-01 5.91005147e-01 -8.49464417e-01
-6.24293029e-01 -6.80134177e-01 -3.06307822e-01 -1.76613957e-01
-6.70251787e-01 -1.33004069e-01 -2.72372037e-01 -3.97176027e-01
7.47247860e-02 -7.31415153e-01 -3.92290860e-01 1.03919670e-01
-8.51698577e-01 -3.19717894e-03 8.76244366e-01 -7.95982957e-01
1.38438368e+00 -2.31221104e+00 -3.49218875e-01 5.59725702e-01
5.55669010e-01 3.26681018e-01 2.91754156e-01 4.87253815e-01
2.86238462e-01 3.54158223e-01 -4.43979681e-01 -5.31247318e-01
-1.34990796e-01 2.32521102e-01 -3.18537176e-01 5.94404578e-01
-1.12267449e-01 9.31406915e-01 -9.14440334e-01 -9.21080709e-01
-1.71899110e-01 5.38023233e-01 -4.71239179e-01 6.60943165e-02
-1.45241991e-01 5.88475525e-01 -9.44579601e-01 3.65588903e-01
1.02249193e+00 -5.26427329e-01 8.52077127e-01 2.19828878e-02
3.59473139e-01 3.80548716e-01 -8.79532635e-01 1.76215017e+00
-3.20815474e-01 -1.18759766e-01 3.22471708e-01 -7.23866522e-01
5.89743495e-01 6.62390172e-01 6.27006710e-01 -1.80245310e-01
3.09662640e-01 1.65664285e-01 -4.94306237e-01 -6.32665694e-01
1.01381943e-01 -1.10579595e-01 -1.64245844e-01 9.96061563e-01
-3.63225639e-02 6.19099677e-01 -5.78232646e-01 3.70435715e-01
1.27357996e+00 -9.55777615e-02 5.04234135e-01 -2.57702731e-02
6.86861038e-01 -3.27687323e-01 9.20983016e-01 5.56979537e-01
-2.18718439e-01 7.26235881e-02 6.98058665e-01 -5.85761786e-01
-4.80720490e-01 -6.86363697e-01 -3.19635153e-01 6.59389853e-01
-3.22539434e-02 -7.76340306e-01 -7.38715947e-01 -1.39336824e+00
2.31769443e-01 3.83451432e-01 -5.60780883e-01 -1.98456481e-01
-4.86873329e-01 -8.29268515e-01 6.59064591e-01 1.06414735e-01
2.84073174e-01 -6.78649843e-01 -7.69303679e-01 1.25524729e-01
-4.40574884e-01 -1.03573251e+00 -5.68921626e-01 1.63861051e-01
-8.45561326e-01 -1.20207822e+00 -3.69425565e-01 -3.57487649e-01
7.97548771e-01 -1.86450668e-02 6.28875971e-01 7.52392262e-02
-2.75616407e-01 -9.15844366e-02 -7.64600188e-02 -5.49043477e-01
-4.97384518e-01 5.76623142e-01 -2.84107953e-01 3.76554787e-01
6.34983242e-01 -6.49722457e-01 -8.07703674e-01 1.95404440e-02
-1.00979650e+00 1.56031981e-01 4.68529880e-01 9.66755569e-01
5.01971126e-01 1.03778914e-02 4.68077689e-01 -1.62294078e+00
4.50095147e-01 -6.85666978e-01 -5.54591298e-01 5.76858819e-01
-6.74222112e-01 7.59885237e-02 7.99178839e-01 -1.58800587e-01
-1.17320859e+00 5.50600529e-01 1.85289383e-01 -5.16030416e-02
-1.06331564e-01 3.97178084e-01 -5.45530379e-01 -8.83909687e-02
3.88961971e-01 1.58623576e-01 4.23831940e-01 -6.46836340e-01
4.29222822e-01 1.10187340e+00 3.88242394e-01 -3.66494387e-01
7.69862711e-01 5.30204833e-01 -2.59224921e-02 -3.48162830e-01
-7.09249556e-01 -3.00162941e-01 -5.96154928e-01 5.24649382e-01
7.47002304e-01 -8.24429214e-01 -1.11747932e+00 2.62384027e-01
-8.56229961e-01 9.50681120e-02 -3.33596289e-01 4.76584047e-01
-4.19288099e-01 4.69784945e-01 -7.02285767e-01 -7.04295099e-01
-8.02261531e-01 -6.47682250e-01 7.33951390e-01 -1.16227262e-01
-2.29193017e-01 -8.00028026e-01 2.37926111e-01 6.10915303e-01
3.69831741e-01 4.50163096e-01 9.31343913e-01 -1.10844111e+00
-7.37784207e-01 -6.35639012e-01 -3.23833637e-02 -7.52507821e-02
5.05073726e-01 -5.47784865e-01 -1.14022398e+00 -3.75605792e-01
2.36783624e-01 -3.19978476e-01 5.01736343e-01 -2.56802261e-01
1.48034656e+00 -9.61754262e-01 -7.74482727e-01 9.78498518e-01
1.33132064e+00 -2.09325358e-01 3.74893636e-01 1.08074218e-01
6.22961104e-01 7.25725114e-01 2.44950891e-01 6.70515776e-01
5.59040368e-01 3.35982293e-01 9.97755453e-02 -1.67764172e-01
7.32236445e-01 -5.16084254e-01 -2.30302542e-01 6.05542004e-01
2.26396292e-01 -4.49092388e-02 -7.11595535e-01 3.46288323e-01
-2.02071786e+00 -5.72643340e-01 3.27059031e-01 2.42072463e+00
1.57836735e+00 -4.80169803e-01 -1.54493824e-01 -8.68892148e-02
2.70825893e-01 1.28480211e-01 -6.10431015e-01 3.46216895e-02
3.69048156e-02 1.89520210e-01 7.72933185e-01 2.50101566e-01
-1.15124714e+00 7.30375051e-01 4.96501446e+00 6.87760949e-01
-1.07376027e+00 4.55772072e-01 7.72144318e-01 -3.00952166e-01
-3.25370431e-01 -1.97603423e-02 -4.28918332e-01 6.97963476e-01
9.41934168e-01 -5.49382746e-01 5.99369965e-02 8.05882514e-01
-1.95861772e-01 5.58826923e-02 -1.15788805e+00 9.28830326e-01
-4.36665237e-01 -1.42389607e+00 -7.26553872e-02 4.94152993e-01
4.21071440e-01 -8.66204277e-02 -1.09688893e-01 -1.89826369e-01
5.92479765e-01 -8.44004333e-01 8.88661947e-03 5.25512516e-01
7.73387313e-01 -8.78856182e-01 9.18019474e-01 5.53336859e-01
-6.12592161e-01 -2.04244051e-02 -3.34613562e-01 4.37996566e-01
3.69453803e-02 7.80242503e-01 -7.47665823e-01 1.06595385e+00
6.34742558e-01 4.67386872e-01 -3.32523823e-01 6.05536938e-01
-9.02068540e-02 3.66603345e-01 -4.74310100e-01 2.20476940e-01
-3.14969480e-01 -1.67666867e-01 1.01046758e-02 1.02054596e+00
1.59907997e-01 4.02403951e-01 5.81359342e-02 6.10176921e-01
-4.48205858e-01 5.41435421e-01 -8.44323695e-01 -4.79839258e-02
5.41647196e-01 1.43666995e+00 -1.57743737e-01 -2.05765188e-01
-4.90582347e-01 9.56053555e-01 4.75565791e-01 9.15213898e-02
-4.35696721e-01 -4.75572348e-01 7.67218411e-01 -9.07996520e-02
5.73647246e-02 2.62912571e-01 -3.15687478e-01 -1.43903852e+00
3.24867338e-01 -8.29656661e-01 9.51527476e-01 1.80211902e-01
-1.51635814e+00 4.49305922e-01 -5.25473773e-01 -1.24767303e+00
-2.21584544e-01 -2.76839584e-02 -2.53050178e-01 9.47604775e-01
-1.45924532e+00 -1.32317734e+00 1.57088526e-02 1.09647465e+00
-8.01436365e-01 3.35049666e-02 1.26944625e+00 2.94326127e-01
-8.27956796e-01 1.31958342e+00 1.83061004e-01 5.03098667e-01
9.34612334e-01 -8.44706059e-01 1.22352220e-01 4.70435143e-01
-1.61661252e-01 1.00776327e+00 1.91088587e-01 -6.76113784e-01
-1.44068742e+00 -1.36870372e+00 1.48184168e+00 -6.02783382e-01
5.48760772e-01 -6.63900852e-01 -1.29610145e+00 9.29015279e-01
-7.75565254e-03 4.69743967e-01 1.56095290e+00 1.36275664e-01
-5.02112985e-01 -2.78181285e-01 -1.60432172e+00 2.96154052e-01
7.70675898e-01 -8.15447927e-01 -2.39973068e-01 6.28367305e-01
9.23762918e-01 -3.14784974e-01 -1.25248408e+00 1.33893430e-01
4.48546290e-01 -6.48804486e-01 7.14911401e-01 -8.87640595e-01
1.02896832e-01 -1.75341487e-01 1.85897335e-01 -8.00538123e-01
-3.50419506e-02 -1.11883640e+00 -3.75585437e-01 1.32807553e+00
6.25119925e-01 -1.37850523e+00 1.07686627e+00 1.25025594e+00
6.13505423e-01 -7.14399755e-01 -1.20640671e+00 -5.18155873e-01
6.74098060e-02 5.76931098e-03 1.20587099e+00 1.49176180e+00
6.83460414e-01 -5.75394481e-02 -5.29555380e-01 3.16940129e-01
7.49360979e-01 1.34346008e-01 8.61189663e-01 -1.18318951e+00
-4.05139685e-01 2.37791479e-01 -2.05839738e-01 -3.67198527e-01
1.97664216e-01 -9.91496861e-01 -2.80201614e-01 -9.90490854e-01
4.06306654e-01 -9.80094433e-01 -5.54160476e-01 9.83597875e-01
-2.96027482e-01 -2.90959656e-01 -7.27065057e-02 3.48323792e-01
-5.33618569e-01 4.25553441e-01 9.25417721e-01 1.09898113e-01
-1.87346026e-01 2.73032963e-01 -1.04356885e+00 3.97818804e-01
8.57804477e-01 -8.42818201e-01 -4.90477860e-01 -3.52498561e-01
2.51902670e-01 3.25722814e-01 2.59206682e-01 -6.21252537e-01
5.74229538e-01 -8.92549008e-02 2.95548618e-01 -3.86746109e-01
-3.16075496e-02 -1.31421602e+00 5.14613509e-01 5.64213574e-01
-3.23864609e-01 -5.22508979e-01 -3.48038897e-02 8.42146158e-01
-2.54608672e-02 1.89209849e-01 4.76610512e-01 5.01966197e-03
1.09763145e-01 5.57738125e-01 6.99251816e-02 -2.19326973e-01
1.27428234e+00 1.49737999e-01 -4.34976906e-01 1.97250620e-02
-8.31230402e-01 5.11258900e-01 4.95968848e-01 4.72485982e-02
7.12000802e-02 -8.75944316e-01 -4.52158689e-01 5.76674104e-01
1.62734032e-01 3.29722732e-01 5.67267478e-01 9.80493784e-01
-3.28452215e-02 4.65064913e-01 2.08432660e-01 -2.42377609e-01
-1.38185084e+00 1.00788033e+00 2.46614471e-01 -6.07658803e-01
-1.01258075e+00 6.12316489e-01 5.98233901e-02 -7.36988425e-01
2.28318885e-01 2.42660921e-02 1.77416801e-01 -1.38965070e-01
7.70052016e-01 5.42199239e-02 3.55852902e-01 -1.79830015e-01
-4.14247751e-01 2.15898864e-02 -5.31234324e-01 2.50610262e-01
1.28317809e+00 -1.37342408e-01 -4.56259191e-01 -1.47859482e-02
1.34769917e+00 1.11385249e-01 -1.00057697e+00 -7.53027797e-01
1.31626979e-01 -6.17265761e-01 -1.77138433e-01 -8.02913666e-01
-1.28754413e+00 4.84273821e-01 3.48423868e-01 -2.04595789e-01
1.21280587e+00 -4.38040718e-02 1.24429762e+00 3.63724798e-01
7.63286829e-01 -7.37234056e-01 -9.28920090e-01 -2.75072698e-02
1.86332285e-01 -1.26095581e+00 2.08449885e-01 -6.49940431e-01
-4.72768009e-01 6.32587075e-01 3.33181560e-01 6.38538241e-01
9.68719542e-01 4.32263166e-01 3.92433137e-01 4.82402109e-02
-8.15534413e-01 4.99883890e-01 -1.75902635e-01 4.96708423e-01
1.63143679e-01 2.41188586e-01 -4.25573170e-01 1.15507865e+00
-3.54731351e-01 3.58695984e-01 6.34494722e-02 1.35537732e+00
3.59753072e-01 -1.57949865e+00 1.29482048e-02 5.55831909e-01
-9.91647899e-01 -2.23634243e-02 -3.63338023e-01 2.79247940e-01
6.40257820e-02 7.59838581e-01 -2.19653800e-01 -2.89979637e-01
3.60839404e-02 3.51793915e-01 -1.06657691e-01 -3.35558981e-01
-9.61346328e-01 -2.39001662e-01 1.96598396e-02 -7.76243567e-01
-6.35426417e-02 -5.77448547e-01 -1.05296588e+00 -4.77455556e-01
-3.63449186e-01 3.94888341e-01 4.75415647e-01 7.85155356e-01
9.70246017e-01 3.88072953e-02 7.45781004e-01 1.94122270e-01
-8.63574207e-01 -4.34728742e-01 -6.93701565e-01 4.55768049e-01
6.50239885e-01 1.50950864e-01 -9.89929661e-02 3.92739475e-02] | [6.098748207092285, 6.518949508666992] |
0bb710fa-1993-40b5-9013-d985455a172a | ih-vit-vision-transformer-based-integrated | 2302.04521 | null | https://arxiv.org/abs/2302.04521v1 | https://arxiv.org/pdf/2302.04521v1.pdf | IH-ViT: Vision Transformer-based Integrated Circuit Appear-ance Defect Detection | For the problems of low recognition rate and slow recognition speed of traditional detection methods in IC appearance defect detection, we propose an IC appearance defect detection algo-rithm IH-ViT. Our proposed model takes advantage of the respective strengths of CNN and ViT to acquire image features from both local and global aspects, and finally fuses the two features for decision making to determine the class of defects, thus obtaining better accuracy of IC defect recognition. To address the problem that IC appearance defects are mainly reflected in the dif-ferences in details, which are difficult to identify by traditional algorithms, we improved the tra-ditional ViT by performing an additional convolution operation inside the batch. For the problem of information imbalance of samples due to diverse sources of data sets, we adopt a dual-channel image segmentation technique to further improve the accuracy of IC appearance defects. Finally, after testing, our proposed hybrid IH-ViT model achieved 72.51% accuracy, which is 2.8% and 6.06% higher than ResNet50 and ViT models alone. The proposed algorithm can quickly and accurately detect the defect status of IC appearance and effectively improve the productivity of IC packaging and testing companies. | ['Chu Wang', 'Jianlan Guo', 'Yuntao Zou', 'Shuang Gao', 'Xiaoibin Wang'] | 2023-02-09 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 1.90411836e-01 -3.72200102e-01 6.49554729e-02 -2.11790845e-01
-3.42892170e-01 -9.78656933e-02 -1.22676946e-01 -5.11553884e-02
-7.98974559e-02 3.79186660e-01 -4.52728540e-01 -2.44271889e-01
2.01665416e-01 -1.00808251e+00 -1.63581520e-01 -5.78181148e-01
4.62447345e-01 3.33290070e-01 3.54110330e-01 1.16579115e-01
5.28224528e-01 5.84935546e-01 -1.46803331e+00 5.26950181e-01
1.00749350e+00 1.39990008e+00 1.27397045e-01 4.88553584e-01
-4.57693547e-01 8.54037106e-01 -9.69717503e-01 -1.68944001e-02
3.40151750e-02 -4.16667491e-01 -5.11440992e-01 3.77683043e-01
2.28963137e-01 -4.69462276e-01 -3.13778430e-01 1.04228246e+00
5.40576279e-01 -3.04429471e-01 6.31612122e-01 -1.02405119e+00
-3.92916083e-01 8.05093274e-02 -8.04368794e-01 5.84565639e-01
1.06930695e-01 -5.11087514e-02 5.34815371e-01 -7.16889560e-01
3.87033463e-01 1.21793568e+00 6.98066235e-01 3.41666996e-01
-6.82790399e-01 -1.01493251e+00 -2.63750087e-02 5.15862703e-01
-1.27928519e+00 -2.63153911e-01 9.18689311e-01 -2.18045712e-01
8.14757943e-01 2.38663197e-01 5.05345345e-01 6.36224985e-01
3.82202953e-01 8.68375838e-01 9.38033581e-01 -4.76640135e-01
-4.50095683e-02 1.22304231e-01 4.02079314e-01 8.72285247e-01
3.51035953e-01 -2.26323187e-01 -1.67851523e-01 1.06983222e-01
1.17783308e+00 3.75169873e-01 5.74857974e-03 3.46717834e-01
-6.53936446e-01 4.91034776e-01 3.89726013e-01 7.45005906e-01
-2.13481694e-01 6.67958111e-02 3.75516623e-01 4.66452569e-01
4.15437669e-01 3.01876605e-01 -3.96627903e-01 -5.56541793e-02
-9.11023498e-01 -4.05861199e-01 2.97273278e-01 4.46789742e-01
6.61885798e-01 3.10904592e-01 2.34567579e-02 1.25964403e+00
3.99202347e-01 4.41448301e-01 7.47914255e-01 -3.63463998e-01
4.01495814e-01 1.02536821e+00 -3.45006943e-01 -1.17988944e+00
-2.50142843e-01 -5.03433287e-01 -9.33854401e-01 2.12978095e-01
1.45292655e-01 5.49350232e-02 -1.51692891e+00 7.68594742e-01
4.65948917e-02 1.99230403e-01 -2.54166424e-01 6.98716700e-01
1.04422998e+00 7.51070440e-01 -1.37623409e-02 7.02392831e-02
1.26825309e+00 -7.87156105e-01 -8.75416994e-01 -1.28659025e-01
5.48525035e-01 -9.64837492e-01 6.61086559e-01 4.92616415e-01
-8.05816889e-01 -8.78596961e-01 -1.45956457e+00 1.70375556e-01
-4.50672001e-01 8.76433969e-01 5.66632748e-01 7.60999382e-01
-9.80179369e-01 5.72621942e-01 -7.37265170e-01 -2.77650416e-01
6.93671882e-01 7.38054514e-01 -1.08420245e-01 -3.72778535e-01
-6.85656905e-01 5.54925025e-01 2.13535931e-02 4.77653027e-01
-7.47409761e-01 -5.60904145e-01 -6.57886922e-01 -9.30543393e-02
1.02678619e-01 6.30780980e-02 7.60902166e-01 -1.20765638e+00
-1.34721923e+00 5.83044112e-01 -9.86014232e-02 1.25303835e-01
1.77775264e-01 2.34570339e-01 -8.24435830e-01 1.85555860e-01
-6.52733147e-02 2.82469541e-01 9.72218454e-01 -1.06231618e+00
-7.87491739e-01 -6.11371160e-01 -1.65667444e-01 -2.62398385e-02
-4.44797963e-01 4.08121943e-02 -7.04826534e-01 -4.61103082e-01
6.29610002e-01 -3.57147634e-01 -2.69506313e-02 -1.29804225e-03
-3.52812678e-01 -5.32411113e-02 1.53371584e+00 -1.13777041e+00
1.05028534e+00 -2.36059475e+00 -3.56983215e-01 4.16482240e-01
4.67858166e-01 6.95075929e-01 -1.35694414e-01 -5.66426441e-02
-1.00598857e-01 1.46532431e-01 -5.94593510e-02 -3.85527790e-01
-4.99288470e-01 1.23088390e-01 3.63202423e-01 5.31131685e-01
3.98594946e-01 6.72820926e-01 -1.90505087e-01 -5.72693467e-01
4.65998381e-01 2.85579056e-01 -1.07943349e-01 1.37180403e-01
1.81364551e-01 2.28703395e-01 -7.22514272e-01 1.26250219e+00
1.02208543e+00 1.50482906e-02 -5.05988114e-02 -3.88657719e-01
4.56630737e-02 7.66038224e-02 -1.18749356e+00 1.21382403e+00
-6.01016164e-01 8.14075351e-01 1.67182133e-01 -1.27763915e+00
1.35446036e+00 3.19481760e-01 5.05054891e-01 -1.12428975e+00
4.92323667e-01 1.79921657e-01 4.81108427e-02 -5.18563747e-01
2.71279067e-01 7.38804787e-02 2.61753529e-01 2.12329894e-01
8.77018192e-06 8.41054246e-02 -1.93700120e-01 -1.12630017e-01
1.21579897e+00 -4.21542346e-01 -3.01782280e-01 -1.23351634e-01
6.05855286e-01 -1.53656483e-01 7.84498513e-01 5.83205879e-01
-2.59221673e-01 7.51696408e-01 4.26135778e-01 -5.54902911e-01
-7.82335341e-01 -7.21475840e-01 -1.85653493e-01 3.18351835e-01
4.99086857e-01 9.22641659e-04 -3.93208385e-01 -9.50958312e-01
-7.59156188e-04 7.05614984e-02 -6.73145592e-01 -1.52313396e-01
-4.02208149e-01 -1.06727183e+00 4.88476396e-01 5.76727092e-01
9.24385309e-01 -1.21818817e+00 -1.74534872e-01 2.60653555e-01
1.22762144e-01 -9.69964862e-01 -1.58411056e-01 1.46362215e-01
-1.18630612e+00 -1.21894085e+00 -5.36280692e-01 -1.10029936e+00
1.12717342e+00 1.62074640e-01 6.87418938e-01 6.31459653e-01
-9.71099198e-01 4.60633077e-02 -4.59346920e-01 -1.80923253e-01
-3.58732998e-01 -2.06219986e-01 -4.43757832e-01 2.16386914e-01
1.85435593e-01 -2.64412969e-01 -7.22799003e-01 5.42486429e-01
-7.83418894e-01 -1.73790202e-01 8.19220185e-01 7.89793909e-01
4.01745081e-01 7.38183260e-01 6.40617013e-01 -7.94958591e-01
3.95536780e-01 -1.09799325e-01 -3.80685210e-01 1.53212860e-01
-5.87830186e-01 -3.34942430e-01 4.49777603e-01 -3.86709392e-01
-1.21801233e+00 -1.13621742e-01 -3.83292347e-01 -3.14486206e-01
-2.27464825e-01 2.00937226e-01 -3.04583102e-01 -5.06480217e-01
7.98243508e-02 5.16925529e-02 1.00058779e-01 -6.79147184e-01
-3.72692376e-01 9.88914192e-01 1.55432358e-01 -2.03098759e-01
3.19004506e-01 2.90022850e-01 -3.89548451e-01 -1.03929496e+00
-9.09998864e-02 -7.09000170e-01 -4.41305280e-01 -3.56108993e-01
7.52160966e-01 -7.21113682e-01 -6.22448206e-01 1.08583808e+00
-1.08875310e+00 -2.27473136e-02 -2.11326703e-02 1.99397191e-01
2.87173063e-01 6.45821810e-01 -1.14073193e+00 -6.56264484e-01
-5.77286661e-01 -1.38743317e+00 9.74429488e-01 3.59258950e-01
2.40097284e-01 -8.65286469e-01 -5.39062738e-01 3.15591812e-01
5.78826785e-01 5.59409223e-02 1.08831787e+00 -2.29859456e-01
-3.88944983e-01 -7.24840641e-01 -6.71353757e-01 6.01678133e-01
4.66863692e-01 1.67548418e-01 -9.91929948e-01 -9.61327851e-02
1.24452129e-01 -3.13952789e-02 8.39205086e-01 3.95005167e-01
1.31159580e+00 2.96658278e-01 -4.69804406e-01 2.67138749e-01
1.75405347e+00 8.88865530e-01 1.14370024e+00 1.33024096e-01
8.41635406e-01 3.24924380e-01 6.81950092e-01 4.00294513e-01
-1.23566471e-01 5.51030159e-01 4.05005783e-01 -6.85434222e-01
-4.97509032e-01 6.18398711e-02 1.78470790e-01 1.04907405e+00
1.61021188e-01 -2.46379778e-01 -5.90725541e-01 6.23887956e-01
-1.47696674e+00 -3.55195999e-01 -3.72084409e-01 1.81309211e+00
3.32711786e-01 2.73971945e-01 -2.14961171e-01 4.19632524e-01
1.00188720e+00 -2.07576603e-01 -5.38554907e-01 -4.94617313e-01
9.88247804e-03 5.38665235e-01 6.29357934e-01 4.30534966e-03
-8.85555983e-01 8.55971992e-01 6.38624287e+00 1.28081334e+00
-1.34602582e+00 1.14427961e-01 1.01972795e+00 1.60660461e-01
7.54205734e-02 -3.71031672e-01 -6.87493742e-01 4.86477941e-01
4.01774734e-01 7.02397227e-01 8.34412724e-02 5.40865898e-01
2.02803351e-02 -3.53073001e-01 -6.60749733e-01 1.16173804e+00
2.50956386e-01 -1.12194085e+00 -1.18616745e-01 -5.69563359e-03
6.69253826e-01 -3.19098532e-01 1.70756489e-01 1.01448685e-01
-4.68753045e-03 -1.01251316e+00 2.78359234e-01 2.23685548e-01
8.49200726e-01 -7.12268412e-01 1.23579764e+00 -1.65064603e-01
-1.25503910e+00 -2.46613637e-01 -6.41628206e-01 1.19367108e-01
-3.45956951e-01 8.84781957e-01 -6.66477740e-01 6.06329441e-01
7.37842739e-01 9.72327292e-01 -6.80086672e-01 1.30064499e+00
2.12080166e-01 8.34809244e-01 -2.91333675e-01 7.57255852e-02
6.91984519e-02 2.17533596e-02 -1.11130560e-02 9.23121870e-01
4.21201378e-01 -1.61362648e-01 -9.86196846e-03 8.29757214e-01
2.84142345e-02 1.60711966e-02 -2.16200009e-01 -5.32157114e-03
1.83568493e-01 1.25400293e+00 -1.18076241e+00 -3.77273411e-01
-3.00061911e-01 1.08974719e+00 -2.00324520e-01 1.67917505e-01
-7.44219303e-01 -6.18913352e-01 4.35993552e-01 -5.11519946e-02
5.98141015e-01 5.48734237e-03 -6.74886167e-01 -7.65005529e-01
1.22823998e-01 -8.50530326e-01 3.03692609e-01 -4.58112627e-01
-9.75234807e-01 4.86605942e-01 -6.83993757e-01 -1.18227255e+00
4.71006811e-01 -8.99124384e-01 -1.01900125e+00 6.65303290e-01
-1.28570735e+00 -1.05704296e+00 -4.27345812e-01 3.60353798e-01
8.84903491e-01 -9.59908888e-02 4.37743366e-01 6.87927723e-01
-1.27026153e+00 8.01179528e-01 1.23615794e-01 9.37353298e-02
5.03696799e-01 -8.74457240e-01 2.79450119e-01 7.86148071e-01
-2.49744326e-01 1.05431907e-01 1.34251133e-01 -8.51502001e-01
-1.29245448e+00 -1.06812215e+00 3.23742002e-01 1.49260193e-01
1.86485037e-01 -1.95893496e-01 -8.61570537e-01 1.78455085e-01
-1.35263830e-01 -9.10008922e-02 2.62259901e-01 -8.45522657e-02
4.91909645e-02 -3.56675476e-01 -1.59036350e+00 2.13930517e-01
5.78024745e-01 -3.38972777e-01 -1.19104877e-01 1.27572447e-01
4.39996332e-01 -1.99243799e-01 -7.86101937e-01 5.24424374e-01
5.28141558e-01 -7.58457661e-01 8.48330081e-01 2.97653377e-01
2.83358514e-01 -2.86900997e-01 5.26473373e-02 -8.11604202e-01
-2.39470720e-01 2.33905077e-01 4.45178688e-01 1.43780971e+00
4.11414295e-01 -7.15498865e-01 1.02915859e+00 1.18436448e-01
-2.93993741e-01 -7.07379639e-01 -8.61011088e-01 -4.70781296e-01
-4.89494950e-01 -4.88427430e-01 4.87222344e-01 7.71707356e-01
-3.29161823e-01 -2.32464373e-01 -2.48687357e-01 2.84908891e-01
2.74321407e-01 -1.01486467e-01 2.15172559e-01 -1.13554513e+00
-2.09384710e-01 -3.04671586e-01 -8.25562835e-01 -8.87255788e-01
-2.96582818e-01 -3.72460335e-01 1.16078153e-01 -1.54529834e+00
3.40198092e-02 -8.45402718e-01 -6.34982526e-01 4.89977747e-01
-5.02193421e-02 8.44406605e-01 -2.20356390e-01 2.06359625e-01
-5.43902278e-01 3.72105002e-01 1.49690950e+00 -5.11011541e-01
-2.42622063e-01 -2.67304182e-02 -4.78329301e-01 4.48638588e-01
7.34750211e-01 -2.73224711e-01 -1.83955461e-01 -4.75728214e-01
-4.14246738e-01 3.76709886e-02 1.79907098e-01 -1.25302553e+00
1.90773189e-01 4.65985507e-01 8.93334687e-01 -7.94574857e-01
3.80783111e-01 -7.97081113e-01 -4.19195704e-02 6.99355364e-01
2.43475989e-01 -1.55063450e-01 5.75387359e-01 3.35388243e-01
-4.58528936e-01 -4.73828465e-01 7.58334517e-01 -8.22784901e-02
-9.90295529e-01 2.59090602e-01 -6.56225801e-01 -7.08665371e-01
8.94111872e-01 -6.34139001e-01 -1.99048683e-01 -1.14034630e-01
-3.94814432e-01 -1.75300129e-02 2.20175967e-01 3.79738897e-01
1.01092589e+00 -9.64747846e-01 -3.41522664e-01 6.45163238e-01
1.66991223e-02 -1.94047242e-01 6.75556958e-01 6.94220841e-01
-9.49076533e-01 1.56245336e-01 -3.51213485e-01 -5.75824201e-01
-1.42438853e+00 7.01219961e-02 4.68802959e-01 -3.24222416e-01
-6.24441147e-01 8.72542977e-01 2.10494727e-01 -1.79714605e-01
1.57322690e-01 -3.34642082e-01 -3.39703828e-01 -2.12298661e-01
4.18999642e-01 5.08804262e-01 5.73378384e-01 -3.95891428e-01
-2.96573818e-01 8.01298320e-01 -5.36234140e-01 3.23218107e-01
9.81417000e-01 -7.87679553e-02 -2.68292129e-01 1.29327923e-01
1.28594434e+00 -3.35642874e-01 -8.08334589e-01 1.68937862e-01
-3.04001898e-01 -5.65526247e-01 4.63859051e-01 -1.06494808e+00
-1.73037601e+00 9.19453800e-01 1.25558972e+00 9.80148613e-02
1.37530720e+00 -1.38056114e-01 1.00900137e+00 -1.74208269e-01
2.10021421e-01 -1.00801528e+00 2.24235788e-01 6.91193938e-02
4.81330574e-01 -1.17659366e+00 -6.91524595e-02 -8.23246539e-01
-2.53129005e-01 1.23618722e+00 9.88806367e-01 -2.78029889e-01
5.13265014e-01 3.47677648e-01 1.26991481e-01 -4.75081444e-01
-2.06131205e-01 6.27594218e-02 2.39944622e-01 5.69766521e-01
1.38829276e-01 -8.29055253e-03 -2.99836278e-01 3.13830376e-01
4.55733210e-01 -2.13710576e-01 3.23362321e-01 9.88819361e-01
-5.22903562e-01 -1.14487243e+00 -5.88024437e-01 9.89410222e-01
-6.75536811e-01 1.57014027e-01 -3.41162443e-01 7.02747703e-01
5.30156076e-01 1.27481246e+00 3.90095204e-01 -7.17682183e-01
2.22220242e-01 -3.73974115e-01 3.91422629e-01 -4.40125853e-01
-4.52512145e-01 1.61072537e-01 -5.82759082e-02 -5.22832811e-01
-6.56025205e-03 -3.72502327e-01 -1.09856570e+00 -1.71423450e-01
-8.18041265e-01 -6.33221790e-02 1.03063691e+00 1.12784088e+00
2.12744310e-01 1.12881744e+00 8.36821854e-01 -5.21831870e-01
8.97734463e-02 -1.07899034e+00 -9.37957466e-01 2.16376603e-01
1.06946163e-01 -6.16075397e-01 -4.04477656e-01 -2.48003423e-01] | [7.423717021942139, 1.748457908630371] |
e5a4034f-ecba-4b38-b022-834248a42a64 | native-language-identification-with | null | null | https://aclanthology.org/J18-3003 | https://aclanthology.org/J18-3003.pdf | Native Language Identification With Classifier Stacking and Ensembles | Ensemble methods using multiple classifiers have proven to be among the most successful approaches for the task of Native Language Identification (NLI), achieving the current state of the art. However, a systematic examination of ensemble methods for NLI has yet to be conducted. Additionally, deeper ensemble architectures such as classifier stacking have not been closely evaluated. We present a set of experiments using three ensemble-based models, testing each with multiple configurations and algorithms. This includes a rigorous application of meta-classification models for NLI, achieving state-of-the-art results on several large data sets, evaluated in both intra-corpus and cross-corpus modes. | ['Shervin Malmasi', 'Mark Dras'] | 2018-09-01 | null | null | null | cl-2018-9 | ['cross-corpus', 'native-language-identification'] | ['computer-vision', 'natural-language-processing'] | [ 2.88405508e-01 -4.84471321e-01 -1.74255982e-01 -5.51383376e-01
-9.64832723e-01 -8.43846679e-01 1.11499441e+00 3.21325064e-02
-5.68998158e-01 7.49218702e-01 1.79981932e-01 -6.29271150e-01
-2.05322474e-01 -5.32424785e-02 -1.75490826e-01 -4.16944772e-01
-1.35462821e-01 8.24779809e-01 -4.13672060e-01 -3.49023670e-01
2.37236381e-01 2.55481768e-02 -1.94825649e+00 5.45582473e-01
1.13820338e+00 7.62353063e-01 -5.23531199e-01 8.52889419e-01
-3.87386680e-01 6.16522789e-01 -9.36773777e-01 -8.14531028e-01
-1.80427030e-01 -7.98380077e-02 -7.71636009e-01 -6.24179304e-01
8.85818899e-01 1.86260343e-01 -1.23618200e-01 8.38350892e-01
8.64231884e-01 -8.35480168e-03 8.39190662e-01 -1.01252854e+00
-5.98147690e-01 1.08122945e+00 -2.05921575e-01 4.28595603e-01
6.99071169e-01 -1.69191718e-01 9.46920156e-01 -9.27461088e-01
4.88571703e-01 1.36983144e+00 9.04078603e-01 7.95405924e-01
-1.43061018e+00 -1.17097306e+00 1.73439294e-01 1.92875177e-01
-1.31358683e+00 -9.38530147e-01 5.96343338e-01 -4.32817370e-01
1.43194234e+00 3.56944680e-01 8.86476189e-02 1.59218740e+00
1.00272344e-02 1.16984975e+00 1.74668956e+00 -8.47258687e-01
-9.59527791e-02 4.27029938e-01 6.81947768e-01 1.57768071e-01
1.88261241e-01 3.38457972e-01 -7.52462983e-01 -1.65249720e-01
-1.75088063e-01 -7.10471630e-01 -5.00686280e-02 1.75006256e-01
-1.37869227e+00 9.72294033e-01 -3.91009569e-01 9.40626502e-01
1.05130516e-01 -2.97384173e-01 7.26227164e-01 5.05838454e-01
1.04896653e+00 5.80042064e-01 -8.25749934e-01 -4.28408831e-01
-1.19633877e+00 3.87559503e-01 1.14060616e+00 7.62002587e-01
1.33376941e-01 1.61368623e-02 -1.68859854e-01 9.98008847e-01
9.93298143e-02 3.59111726e-01 6.37875378e-01 -3.95225883e-01
8.18308532e-01 7.09542871e-01 -3.01334202e-01 -5.96005857e-01
-3.35836202e-01 -4.58752126e-01 -1.00511312e+00 3.44139934e-01
6.16964340e-01 -1.97122484e-01 -4.68093932e-01 1.67214966e+00
-1.51156172e-01 3.42769265e-01 3.58274668e-01 2.39211962e-01
1.20884669e+00 2.96621472e-01 4.88718569e-01 -6.97233602e-02
1.16457593e+00 -7.35934913e-01 -7.04023302e-01 -3.37648839e-01
7.67534077e-01 -7.68755496e-01 4.59455758e-01 5.91167390e-01
-8.10837865e-01 -8.10237646e-01 -8.25262070e-01 2.28326634e-01
-7.54984438e-01 3.77865225e-01 6.65823698e-01 1.22191167e+00
-9.61374581e-01 4.29175705e-01 -4.10946190e-01 -2.61052817e-01
1.83422580e-01 6.39008284e-01 -8.75044167e-01 1.40374959e-01
-1.31291080e+00 1.09246075e+00 6.20974958e-01 -1.03562213e-01
-3.67062628e-01 -8.42004776e-01 -7.23221064e-01 -3.20914567e-01
-3.20226960e-02 -4.35496867e-01 1.19438696e+00 -9.67497408e-01
-1.40109730e+00 1.13612723e+00 -3.86108458e-01 -4.73799706e-01
3.11534047e-01 -2.73418307e-01 -8.96796823e-01 -7.30537057e-01
-1.41667694e-01 3.88968527e-01 5.86800575e-01 -1.23313391e+00
-7.87114143e-01 -7.86721945e-01 -3.05257440e-01 1.91808909e-01
-4.64271486e-01 6.59926295e-01 3.05789113e-01 -8.23048472e-01
-2.45177314e-01 -9.01349187e-01 1.20863482e-01 -1.12014210e+00
-4.34378058e-01 -7.61653960e-01 7.82612085e-01 -9.90494251e-01
1.92283428e+00 -1.95250952e+00 3.79335850e-01 -1.95277900e-01
4.38599512e-02 6.73504174e-01 -1.15007289e-01 3.50695103e-01
-3.33240747e-01 5.96308053e-01 -1.13420047e-01 -9.34841454e-01
-2.90139560e-02 -9.70302299e-02 -1.77138403e-01 -6.20658062e-02
-8.26625247e-03 7.60811508e-01 -7.53671646e-01 -3.82947654e-01
2.72876322e-01 5.90040565e-01 -2.46256545e-01 2.37322062e-01
-4.30026501e-02 7.81068206e-01 -3.67477462e-02 5.92337012e-01
3.24801475e-01 2.24662438e-01 2.76715308e-01 -4.27684747e-02
-3.72973442e-01 4.59409893e-01 -1.29403150e+00 1.62634254e+00
-5.30350387e-01 7.24885821e-01 -2.37405617e-02 -1.10420942e+00
8.71176422e-01 4.60848987e-01 2.72961557e-01 -2.27618903e-01
1.31174847e-01 6.69266224e-01 4.38892454e-01 -1.58629026e-02
4.16080117e-01 3.97166103e-01 -4.21791762e-01 5.17925203e-01
4.38115269e-01 3.07487696e-01 4.64074045e-01 -7.88701847e-02
8.54323387e-01 2.61457562e-01 5.85502386e-01 -3.64419758e-01
1.22228038e+00 -7.41915330e-02 7.53695741e-02 1.17509568e+00
-1.86967298e-01 2.26545244e-01 -1.94224805e-01 -4.08846289e-01
-8.87997746e-01 -6.86505258e-01 -4.52372432e-01 1.49787736e+00
-6.82640314e-01 -3.86589855e-01 -9.90097642e-01 -8.05584311e-01
8.25216621e-02 9.24234569e-01 -3.80196810e-01 2.60144621e-01
-6.53441072e-01 -9.93234396e-01 9.14212883e-01 3.63241076e-01
6.61985397e-01 -1.17451799e+00 -1.01406105e-01 5.12949407e-01
-3.38800639e-01 -1.11353469e+00 -1.88442245e-01 1.14066064e-01
-6.70712650e-01 -9.43309426e-01 -3.34849924e-01 -6.83680177e-01
-1.45134963e-02 -1.71454504e-01 1.49356747e+00 6.19235784e-02
-1.98362082e-01 4.85368341e-01 -4.03225660e-01 -7.29430258e-01
-9.97097552e-01 6.94742024e-01 4.27876353e-01 -1.07824862e-01
9.06322777e-01 -3.60268623e-01 -6.42014295e-02 2.16327906e-01
-2.84885049e-01 -1.70143768e-01 2.19594985e-01 1.10111368e+00
1.39035046e-01 2.87912618e-02 6.10837519e-01 -7.79447138e-01
8.61668527e-01 -2.63647586e-01 -2.62075692e-01 5.71721435e-01
-5.57654142e-01 1.31368846e-01 3.58921081e-01 -7.97968686e-01
-1.22247922e+00 -1.53301820e-01 -3.90536666e-01 1.68978244e-01
-7.68453002e-01 4.33906078e-01 -1.61949918e-01 -3.26845437e-01
3.71047199e-01 2.39228845e-01 -4.40204889e-03 -9.00865138e-01
2.03972995e-01 1.29675245e+00 4.35884565e-01 -6.78550899e-01
2.60904074e-01 -1.82977200e-01 -4.90143239e-01 -7.86847174e-01
-9.36063230e-01 -6.14640713e-01 -1.09319448e+00 -1.93213448e-01
6.75989091e-01 -9.15685236e-01 -6.19245887e-01 8.27865303e-01
-1.23125327e+00 -1.27152190e-01 4.25442398e-01 4.40372139e-01
-2.82095850e-01 2.22658411e-01 -2.08355099e-01 -1.19488168e+00
-7.45490670e-01 -1.36709952e+00 9.02783453e-01 -4.66639400e-02
-7.64808595e-01 -1.12566853e+00 4.05642569e-01 5.83111107e-01
7.31391191e-01 -2.83239949e-02 8.22506249e-01 -1.45485830e+00
1.04561850e-01 -2.86270618e-01 1.07034944e-01 3.25258464e-01
-1.72505870e-01 1.42064571e-01 -1.29928148e+00 -3.97104830e-01
-3.73836458e-01 -1.84933662e-01 8.31910253e-01 1.70378968e-01
1.12248409e+00 -2.29552984e-01 -6.44980252e-01 4.56926525e-01
1.15628779e+00 1.58936039e-01 4.12231743e-01 4.64331359e-01
6.12586856e-01 6.95556521e-01 1.67062879e-01 1.43656254e-01
4.51315373e-01 9.66329873e-01 -1.18229546e-01 5.20737469e-01
-2.69316673e-01 -2.86513660e-02 4.21974719e-01 9.89541888e-01
-2.02466115e-01 -4.22071695e-01 -1.37472034e+00 3.46588016e-01
-1.58901691e+00 -1.29867268e+00 -1.57596231e-01 2.32895517e+00
7.88655102e-01 -8.08002129e-02 2.46629864e-01 6.13183260e-01
5.73500931e-01 -7.73407146e-02 -2.67792404e-01 -5.56495547e-01
-3.25138032e-01 4.07759309e-01 9.48177353e-02 4.94308889e-01
-1.67682564e+00 9.98804152e-01 7.35541773e+00 9.29572344e-01
-8.48365426e-01 3.63561481e-01 8.63093793e-01 1.37489036e-01
-3.50665376e-02 -5.69938235e-02 -1.66033304e+00 6.64727449e-01
1.28843379e+00 -3.45150717e-02 6.16125524e-01 6.59118593e-01
-6.28737688e-01 -1.36231324e-02 -1.19177628e+00 1.02924049e+00
3.34195465e-01 -9.62339222e-01 -1.49773791e-01 1.27446681e-01
9.09895062e-01 3.09049606e-01 6.25844300e-02 8.29876959e-01
3.02155882e-01 -1.28397954e+00 6.62695646e-01 4.48839754e-01
9.83746707e-01 -7.97455430e-01 8.53238523e-01 6.67526543e-01
-1.03725433e+00 -3.76937538e-01 2.80170590e-01 8.87068361e-02
-1.41770855e-01 3.77977014e-01 -3.09957802e-01 6.19417727e-01
8.64042103e-01 5.13271153e-01 -8.63375247e-01 6.02511942e-01
1.84400186e-01 9.13530171e-01 -5.07941484e-01 -3.22380692e-01
-6.96234182e-02 1.37532979e-01 7.92119861e-01 1.56769454e+00
2.26859465e-01 -1.23873316e-01 1.39544621e-01 4.44737464e-01
1.87647671e-01 2.93282837e-01 -7.97919989e-01 -2.50056654e-01
6.12001300e-01 1.27407885e+00 2.37706210e-02 -3.22266996e-01
-4.98133510e-01 7.77525365e-01 6.71974897e-01 9.54481661e-02
-2.86887527e-01 -1.80011138e-01 8.48914385e-01 -3.51192027e-01
-1.93489313e-01 -3.65743697e-01 -5.91384888e-01 -1.19138205e+00
-3.32798988e-01 -1.49957466e+00 7.69511700e-01 -8.69553089e-02
-1.64163077e+00 8.69446278e-01 3.67698431e-01 -9.29846644e-01
-7.67773390e-01 -1.19487488e+00 -6.29378676e-01 1.02973199e+00
-1.11942816e+00 -1.27900600e+00 -1.72356367e-01 2.74975419e-01
5.12213111e-01 -1.17354405e+00 1.43490672e+00 3.04335773e-01
-8.24031055e-01 1.05071199e+00 3.32168221e-01 2.57108629e-01
8.13125193e-01 -1.25349987e+00 5.29236972e-01 5.42429626e-01
4.22127724e-01 7.98760951e-01 5.42098403e-01 -4.82015878e-01
-1.25021529e+00 -6.09842062e-01 1.42882550e+00 -1.10484385e+00
6.38291180e-01 -4.34451997e-01 -7.65197158e-01 6.29927039e-01
4.32396531e-01 -4.43409741e-01 1.04786289e+00 6.83910251e-01
-5.44030070e-01 -9.79564786e-02 -1.10773849e+00 3.64592344e-01
1.13760734e+00 -4.69433963e-01 -5.28740048e-01 1.18685648e-01
2.28906959e-01 -3.69102478e-01 -1.26632380e+00 7.45437801e-01
8.00963759e-01 -1.02504551e+00 1.19035447e+00 -6.14869058e-01
2.34607741e-01 1.40485853e-01 -1.79000393e-01 -1.60718882e+00
-2.71962672e-01 -4.79720056e-01 -2.27219105e-01 1.86216390e+00
7.22425461e-01 -8.80055666e-01 3.65892231e-01 4.80990112e-01
-3.86561325e-04 -6.82324708e-01 -1.23336911e+00 -8.69242966e-01
4.57453847e-01 -8.85678589e-01 9.13767040e-01 8.22781146e-01
7.63929114e-02 4.94640768e-01 -1.23649187e-01 -2.93240935e-01
7.70479858e-01 -2.97618270e-01 8.85369658e-01 -1.45818555e+00
-3.35325971e-02 -1.10341632e+00 -3.86284292e-01 -1.79193825e-01
9.03988063e-01 -1.39978921e+00 -4.05673474e-01 -1.04059255e+00
3.27203721e-01 -4.93075788e-01 -2.43695855e-01 3.50977808e-01
-3.95052016e-01 3.28038245e-01 1.77913666e-01 2.70650208e-01
-4.50902611e-01 1.39651373e-01 4.33749437e-01 -3.81173283e-01
-7.34308809e-02 -5.55588817e-03 -6.99383140e-01 5.59439600e-01
1.00584686e+00 -1.82531059e-01 -8.13128874e-02 -4.28130120e-01
-4.81275208e-02 -2.98315793e-01 2.81419582e-03 -1.27551603e+00
2.01155901e-01 4.92722154e-01 4.28598404e-01 -6.64031565e-01
1.57609642e-01 -4.02654946e-01 2.17333466e-01 3.85794863e-02
-4.92609233e-01 4.84670669e-01 5.55032730e-01 1.84527114e-01
-3.40952069e-01 -3.47749531e-01 5.18913150e-01 -5.16784452e-02
-9.37359273e-01 2.13122100e-01 -2.10658222e-01 2.70311475e-01
6.82833910e-01 -1.81312367e-01 -1.89061817e-02 -7.48388544e-02
-6.25046849e-01 -1.06237113e-01 -7.20790704e-04 8.36113989e-01
2.03566268e-01 -1.19724643e+00 -1.34655380e+00 4.93578047e-01
3.32942486e-01 -7.26791680e-01 1.19283982e-01 6.66607499e-01
7.79789388e-02 7.60373890e-01 1.98661491e-01 -6.66427135e-01
-1.73338997e+00 2.94690073e-01 4.22720462e-01 -9.12639916e-01
-1.12089492e-01 8.43284547e-01 -3.19178015e-01 -1.33963370e+00
3.84931237e-01 5.66866338e-01 -7.15788782e-01 1.42285302e-01
8.73130381e-01 5.51531494e-01 1.49370670e-01 -1.04491246e+00
-5.57736158e-01 4.95021403e-01 -3.02530348e-01 -2.20425501e-01
8.83060396e-01 1.05456337e-02 -3.45842421e-01 6.01600885e-01
1.04877412e+00 -2.97268391e-01 -1.69701442e-01 -1.79909378e-01
4.34246778e-01 -4.62370738e-02 -3.08391452e-02 -1.47601926e+00
-3.81119549e-01 9.31122541e-01 7.36337602e-01 -5.65392710e-02
8.16793263e-01 -4.00296897e-01 3.90668541e-01 2.06575841e-01
6.63973033e-01 -1.09694469e+00 -3.78194630e-01 9.96667266e-01
9.40675735e-01 -1.72857225e+00 -7.75872767e-02 -1.48712009e-01
-4.59391832e-01 1.29027355e+00 5.57124734e-01 3.03110003e-01
7.95755148e-01 5.23402512e-01 9.27303582e-02 4.23786551e-01
-8.01041484e-01 1.70201343e-02 8.83062959e-01 6.64966941e-01
1.28041267e+00 4.40244883e-01 -3.67939651e-01 7.84828305e-01
-4.08028513e-01 -2.30894268e-01 -2.67680019e-01 2.63524622e-01
1.86983466e-01 -1.50597811e+00 -5.81021667e-01 6.41890526e-01
-7.34991312e-01 -3.48516077e-01 -5.62751710e-01 8.82758021e-01
2.32654139e-01 1.32822311e+00 2.76703611e-02 -6.02049708e-01
2.44042829e-01 7.50486255e-01 6.17587984e-01 -4.02285576e-01
-1.35916626e+00 -4.99033839e-01 5.97073495e-01 -1.67805240e-01
-4.98376936e-01 -1.09601212e+00 -4.01852697e-01 -5.43536961e-01
-4.45274860e-01 -2.69662049e-02 1.00644362e+00 1.07802570e+00
3.65224689e-01 2.05128491e-01 3.16732526e-01 -9.76721406e-01
-7.89941788e-01 -1.47869051e+00 -6.18162975e-02 5.05673945e-01
-4.28661220e-02 -6.02133632e-01 -6.12951934e-01 -4.08679783e-01] | [10.38045883178711, 10.560400009155273] |
ad6f480c-bc23-4957-ac6c-0fe33adb4a5f | group-extract-and-aggregate-summarizing-a | 1910.05032 | null | https://arxiv.org/abs/1910.05032v1 | https://arxiv.org/pdf/1910.05032v1.pdf | Group, Extract and Aggregate: Summarizing a Large Amount of Finance News for Forex Movement Prediction | Incorporating related text information has proven successful in stock market prediction. However, it is a huge challenge to utilize texts in the enormous forex (foreign currency exchange) market because the associated texts are too redundant. In this work, we propose a BERT-based Hierarchical Aggregation Model to summarize a large amount of finance news to predict forex movement. We firstly group news from different aspects: time, topic and category. Then we extract the most crucial news in each group by the SOTA extractive summarization method. Finally, we conduct interaction between the news and the trade data with attention to predict the forex movement. The experimental results show that the category based method performs best among three grouping methods and outperforms all the baselines. Besides, we study the influence of essential news attributes (category and region) by statistical analysis and summarize the influence patterns for different currency pairs. | ['Xu sun', 'Shuming Ma', 'Qi Su', 'Deli Chen', 'Ruihan Bao', 'Keiko Harimoto'] | 2019-10-11 | group-extract-and-aggregate-summarizing-a-1 | https://aclanthology.org/D19-5106 | https://aclanthology.org/D19-5106.pdf | ws-2019-11 | ['stock-market-prediction'] | ['time-series'] | [-6.20544016e-01 -2.54359514e-01 -6.61824584e-01 -1.22260310e-01
-9.32042301e-01 -6.79272711e-01 1.07304132e+00 3.98330092e-01
-3.64056438e-01 9.41920757e-01 1.29658258e+00 -2.91318536e-01
-4.20607440e-02 -8.27256739e-01 -4.97994572e-01 -3.44024837e-01
-8.02486464e-02 5.42354226e-01 3.09585243e-01 -3.97791654e-01
6.84420228e-01 -1.11213468e-01 -1.17452300e+00 4.22349513e-01
1.23460603e+00 1.06403553e+00 2.25390494e-03 7.70001337e-02
-6.01830900e-01 1.32390749e+00 -9.85486388e-01 -4.32259083e-01
4.74836260e-01 -8.89791071e-01 -5.79383254e-01 5.17178550e-02
-4.65423912e-02 -3.60482514e-01 -1.11289687e-01 1.00960541e+00
3.04438680e-01 4.90729026e-02 7.51517355e-01 -6.91010296e-01
-4.88275617e-01 1.53355765e+00 -8.96899521e-01 9.47768211e-01
-3.71846594e-02 -4.15004432e-01 1.51454413e+00 -7.69362628e-01
7.83483565e-01 1.01176512e+00 2.70930350e-01 -1.72638953e-01
-6.72423124e-01 -7.13216543e-01 5.06654024e-01 -3.45406607e-02
-6.52925491e-01 -2.49266401e-01 9.12392557e-01 -4.70395684e-01
9.31286275e-01 2.69179523e-01 8.17267537e-01 9.33715224e-01
4.18642759e-01 8.80619705e-01 1.16111195e+00 -7.89604858e-02
1.78365782e-01 -4.92511056e-02 6.26889765e-01 4.43977863e-02
4.48050529e-01 -2.85330713e-01 -8.40178549e-01 -3.44395310e-01
2.66343743e-01 1.12076223e-01 -3.07447582e-01 5.19721866e-01
-1.65922272e+00 1.17441618e+00 2.07644433e-01 4.18546319e-01
-9.21138763e-01 -1.79796144e-01 5.76022625e-01 6.31686330e-01
1.36223507e+00 8.20151389e-01 -9.02432501e-01 -1.47709861e-01
-1.10489476e+00 5.74028313e-01 9.80258286e-01 7.64606595e-01
2.65187174e-01 7.41756856e-02 -3.05052280e-01 9.42588568e-01
1.21843696e-01 3.58767569e-01 9.74166691e-01 -4.93740797e-01
1.01829851e+00 5.69919705e-01 1.92173757e-02 -1.07904732e+00
-4.03795242e-01 -6.49602056e-01 -8.06353748e-01 -1.94796383e-01
2.64066935e-01 -5.55894017e-01 -6.67415082e-01 1.09751606e+00
8.29635859e-02 1.00244492e-01 1.91458389e-01 8.37140262e-01
9.39695001e-01 1.08385336e+00 6.93374425e-02 -9.05508697e-01
1.58868980e+00 -1.19794953e+00 -9.98228073e-01 -1.05470732e-01
4.09629852e-01 -1.06692994e+00 4.21197087e-01 1.70674354e-01
-1.15281284e+00 -1.82795599e-01 -9.03608918e-01 8.75303373e-02
-3.29704702e-01 8.75516161e-02 4.45990026e-01 -3.12984586e-02
-4.99100417e-01 6.63451552e-01 -5.96994281e-01 1.60968527e-01
6.70781732e-02 1.05647035e-02 1.46584362e-01 6.22121215e-01
-1.55112135e+00 8.01742315e-01 6.90263033e-01 -4.27993327e-01
-1.03675120e-01 -8.99158955e-01 -5.76623797e-01 2.11666524e-01
4.65348065e-01 -6.68985069e-01 1.21206295e+00 -5.52591205e-01
-1.43811965e+00 2.34274626e-01 1.49846300e-01 -8.11352968e-01
6.54364526e-01 -3.79082888e-01 -2.86418021e-01 1.07662506e-01
3.11748922e-01 7.88269714e-02 3.97070020e-01 -6.77572191e-01
-1.30810392e+00 -3.16753656e-01 -1.03649966e-01 2.90406197e-01
-2.97158748e-01 3.47714335e-01 -3.98391902e-01 -1.47945678e+00
3.32566053e-01 -5.58196425e-01 -1.73340514e-01 -1.24337173e+00
-5.25360286e-01 -5.81324458e-01 3.84414554e-01 -1.19205928e+00
1.74480557e+00 -1.78946745e+00 1.91574380e-01 1.59271464e-01
2.55702168e-01 -3.85427594e-01 2.56966412e-01 5.76755345e-01
1.84391350e-01 2.56152034e-01 2.17683062e-01 -1.32444233e-01
8.43483135e-02 -1.95551977e-01 -1.00803995e+00 1.29849657e-01
-2.92755157e-01 9.19644058e-01 -6.56215191e-01 -5.41456044e-01
-3.30001086e-01 -1.32542163e-01 -4.67216849e-01 -1.29316390e-01
-4.61889923e-01 2.33818680e-01 -8.54043663e-01 5.93621910e-01
2.99640030e-01 -2.76247978e-01 1.04755335e-01 -8.66030753e-02
-5.45776963e-01 8.05452347e-01 -6.83742046e-01 1.04790890e+00
-1.07816957e-01 3.96235317e-01 -2.65396386e-01 -8.17627728e-01
9.67383862e-01 2.88335055e-01 5.55844963e-01 -7.84960091e-01
2.32618183e-01 5.06201982e-01 -3.55952196e-02 -8.71965364e-02
9.03699100e-01 -1.88367382e-01 -4.86336112e-01 7.27284074e-01
-1.94571257e-01 1.34110093e-01 5.32318830e-01 4.02146608e-01
8.10498893e-01 -3.37893039e-01 6.83892846e-01 -5.73959887e-01
9.43401530e-02 3.45895290e-01 9.19375718e-01 5.78687847e-01
2.30971575e-01 6.41891539e-01 8.32558095e-01 -4.06944811e-01
-1.01369715e+00 -2.70939022e-01 7.11426884e-02 1.04003859e+00
1.63713619e-01 -5.95693409e-01 -4.40353364e-01 -7.91926682e-01
2.14455441e-01 7.81076133e-01 -7.26395249e-01 2.29729831e-01
-6.48410201e-01 -1.28163052e+00 3.91524844e-02 4.46696132e-01
6.89267695e-01 -1.06093597e+00 -2.80837268e-01 4.44065422e-01
-5.52322388e-01 -6.39490545e-01 -6.82249784e-01 2.07673460e-02
-6.77443087e-01 -7.05022931e-01 -1.07146263e+00 -3.71679813e-01
1.13437228e-01 2.14704484e-01 1.06720150e+00 -2.33493894e-01
5.30770481e-01 -1.70389235e-01 -7.40929365e-01 -7.97928452e-01
-2.76901364e-01 5.78324735e-01 -5.17031029e-02 2.30399985e-02
1.98444843e-01 -3.64052117e-01 -5.63674867e-01 1.13988668e-01
-4.68552500e-01 -4.78389068e-03 5.10383546e-01 6.16507828e-01
3.84099543e-01 3.68375927e-01 8.88496280e-01 -1.03554881e+00
1.04805923e+00 -9.34845507e-01 -5.71330547e-01 3.86851817e-01
-7.46054351e-01 1.46986350e-01 4.50732768e-01 -4.16687965e-01
-1.25386977e+00 -6.69812202e-01 2.52984017e-01 1.72887832e-01
4.18725312e-01 1.33760989e+00 1.59977466e-01 8.71775091e-01
8.63701925e-02 3.59141201e-01 -1.73457384e-01 -8.51441145e-01
-8.72273445e-02 6.58041120e-01 2.51658946e-01 -2.17815638e-01
6.12484813e-01 1.64431199e-01 -6.48463368e-01 -7.68195271e-01
-1.16415191e+00 -5.07937610e-01 -2.17445940e-01 -5.70022687e-02
8.61653805e-01 -1.12315869e+00 -1.34031549e-01 6.05678082e-01
-9.92684484e-01 9.74300578e-02 -1.30561039e-01 9.85651314e-01
-1.91277683e-01 3.45351219e-01 -1.06840050e+00 -7.26656377e-01
-5.78494608e-01 -7.38350749e-01 6.87377870e-01 1.10737890e-01
-3.02943379e-01 -8.72278690e-01 2.93539196e-01 4.89015490e-01
2.62477726e-01 3.26574564e-01 7.75545061e-01 -1.41614938e+00
-5.11027217e-01 -1.94966011e-02 -1.08499061e-02 -1.06836252e-01
2.94589132e-01 -2.28112623e-01 -2.31445178e-01 5.16552180e-02
2.71169662e-01 4.58908938e-02 1.50685477e+00 6.83195114e-01
3.40067387e-01 -7.75143683e-01 -8.53853747e-02 3.87267530e-01
9.32579637e-01 3.73821497e-01 1.86541706e-01 8.39573801e-01
4.60215449e-01 7.36476779e-01 8.61207604e-01 6.43596709e-01
4.81966674e-01 3.93131465e-01 -2.00921059e-01 1.50989488e-01
1.94830388e-01 -1.69816151e-01 5.69074810e-01 1.45819581e+00
-4.16984677e-01 -2.56354362e-01 -7.25024879e-01 4.44542050e-01
-2.00541306e+00 -1.28145266e+00 -2.33296067e-01 1.75717998e+00
9.51592505e-01 4.06004906e-01 4.55464214e-01 -9.03395005e-03
6.76965475e-01 5.03861845e-01 -1.68142706e-01 3.06566000e-01
-4.67988014e-01 -4.02281672e-01 6.77990377e-01 3.77311110e-01
-1.25316978e+00 8.64345372e-01 6.13276911e+00 8.91157031e-01
-8.84999275e-01 -1.66623756e-01 9.12899435e-01 -2.00757533e-01
-4.84140366e-01 -1.01865986e-02 -1.14354026e+00 9.64375973e-01
7.87521541e-01 -8.30589950e-01 -1.00674666e-01 7.18409956e-01
3.28493088e-01 -1.12289906e-01 -3.92277479e-01 4.97380942e-01
2.19720021e-01 -1.62308764e+00 3.06937009e-01 4.13070232e-01
9.03319240e-01 1.85702860e-01 -2.28181243e-01 3.01352113e-01
6.03101909e-01 -1.84789255e-01 1.11260271e+00 4.47172254e-01
-3.65543254e-02 -8.81592751e-01 9.95822906e-01 5.26063561e-01
-1.04709888e+00 -2.24505141e-01 -4.64655399e-01 -1.35971248e-01
4.75008547e-01 6.83214605e-01 -2.50904083e-01 8.60404074e-01
6.92973673e-01 1.00984550e+00 -4.71824378e-01 8.44117761e-01
3.32871564e-02 9.04021025e-01 -2.88839698e-01 -2.22639665e-01
6.08202100e-01 -5.77372611e-01 6.72181845e-01 9.03179944e-01
6.76495612e-01 5.80998838e-01 2.23231047e-01 2.48146966e-01
-2.96149284e-01 6.63347900e-01 -2.53164321e-01 -7.52196535e-02
4.66029018e-01 7.64737785e-01 -1.21048009e+00 -8.49144816e-01
-6.87412977e-01 6.72847748e-01 6.57516271e-02 2.14072600e-01
-6.38450563e-01 -3.04473996e-01 2.74120837e-01 4.78221066e-02
5.68272233e-01 1.22545317e-01 -5.10329723e-01 -1.68519318e+00
1.00675069e-01 -7.76477516e-01 8.20911348e-01 -4.18616772e-01
-1.38144588e+00 7.22738981e-01 1.90016285e-01 -1.38437653e+00
-1.64284527e-01 -8.34311098e-02 -7.82095075e-01 4.76327956e-01
-1.48877585e+00 -7.84936368e-01 4.36341435e-01 4.78752181e-02
9.82310951e-01 -7.81688750e-01 1.55649289e-01 6.86087832e-02
-6.47028327e-01 2.90058982e-02 4.56566155e-01 3.53383362e-01
5.55290937e-01 -1.42124379e+00 5.08704484e-01 7.37641573e-01
3.01387727e-01 6.29559577e-01 6.01322830e-01 -1.20610392e+00
-5.62826037e-01 -9.19550598e-01 1.51780844e+00 -3.53645980e-01
1.07607853e+00 -2.96308799e-03 -8.40659201e-01 6.64854586e-01
7.74345338e-01 -9.50918257e-01 6.56945169e-01 3.46632361e-01
-2.76449233e-01 -8.13814849e-02 -3.85379732e-01 3.64724040e-01
4.65380639e-01 9.05089602e-02 -1.35976505e+00 2.54356116e-01
9.87047791e-01 -8.32481608e-02 -8.63209069e-01 1.01058565e-01
1.93283349e-01 -8.07953596e-01 8.56201708e-01 -5.16174018e-01
7.43332326e-01 -1.37586324e-02 1.97769076e-01 -1.67275953e+00
-3.98823380e-01 -6.26688004e-01 -8.13644081e-02 1.66315079e+00
7.94905484e-01 -8.57374609e-01 4.04479891e-01 3.19404244e-01
-2.24556290e-02 -4.54483598e-01 -7.12833405e-01 -6.45790339e-01
2.80551463e-01 1.81013823e-01 7.27689028e-01 1.29819393e+00
4.95403171e-01 7.79951155e-01 -7.67981291e-01 -2.43230075e-01
2.90569067e-01 7.77506292e-01 4.92277712e-01 -1.39868045e+00
-1.88842788e-01 -8.25055480e-01 1.23910502e-01 -1.03436673e+00
9.78566110e-02 -6.84746385e-01 -3.04092884e-01 -1.47731972e+00
4.44259346e-01 -1.32108657e-02 -4.16461587e-01 2.76963925e-03
-5.52929997e-01 -1.53894916e-01 3.64724666e-01 9.85789716e-01
-6.41267002e-01 5.51639557e-01 1.02886140e+00 -1.40923545e-01
-5.16353130e-01 2.74630010e-01 -9.45299923e-01 7.62804210e-01
8.97989392e-01 -4.25638169e-01 -1.21832624e-01 -2.28188142e-01
4.06420201e-01 3.22917014e-01 -2.64096290e-01 -5.53498507e-01
4.38549340e-01 -2.71778643e-01 1.51273951e-01 -1.27059138e+00
-2.80054003e-01 -4.47590232e-01 -1.35300281e-02 3.41551811e-01
-6.18020356e-01 4.35542077e-01 -2.30750948e-01 5.85393965e-01
-7.65669107e-01 9.27699544e-03 3.84680867e-01 -2.91837633e-01
-3.81837428e-01 1.92587003e-01 -4.67930257e-01 4.31336105e-01
8.44212711e-01 3.07394475e-01 -4.14505363e-01 -5.57287931e-01
-4.73415971e-01 4.62244391e-01 -7.95132965e-02 5.40163398e-01
1.69890180e-01 -1.21645999e+00 -1.36792147e+00 -1.96686089e-01
-1.73889905e-01 -2.42014721e-01 9.97142494e-02 8.77024531e-01
-5.73212743e-01 5.46567798e-01 2.14130599e-02 1.83538556e-01
-1.08835995e+00 6.20466352e-01 -2.45558575e-01 -8.11719537e-01
-8.55708003e-01 5.11567950e-01 2.88212627e-01 9.24842060e-02
1.60338804e-02 -6.68275774e-01 -8.84054124e-01 1.06848383e+00
8.73215139e-01 3.89008164e-01 -2.49054134e-01 -6.68044031e-01
5.50881848e-02 4.50960875e-01 -4.59265918e-01 -3.47136885e-01
1.85945630e+00 -3.30285221e-01 -2.91648477e-01 5.23020864e-01
8.82584214e-01 3.87053132e-01 -9.76704776e-01 -4.31400239e-01
6.81936622e-01 -2.03046650e-02 2.09741667e-01 -6.08184457e-01
-1.20062828e+00 3.08712810e-01 -5.80107093e-01 6.97195113e-01
7.84876883e-01 1.69356182e-01 9.62264836e-01 2.03613237e-01
-3.66811268e-03 -1.42864382e+00 -2.35851288e-01 7.10942805e-01
1.07127643e+00 -1.02546859e+00 4.08400446e-01 -2.47457236e-01
-1.04728651e+00 9.16367710e-01 1.85040683e-01 -1.77080899e-01
1.01909423e+00 1.16895340e-01 4.07121569e-01 -3.62763703e-01
-1.02562308e+00 -8.91085789e-02 4.18400019e-01 -3.68051499e-01
3.14982504e-01 -1.82160303e-01 -1.07130742e+00 1.00936389e+00
-6.03654623e-01 -4.39904362e-01 5.89333296e-01 7.05625594e-01
-7.17106998e-01 -8.45414400e-01 -3.61915797e-01 8.41721177e-01
-1.24277246e+00 -2.41076484e-01 -7.03404188e-01 9.66809511e-01
-4.38458562e-01 7.01102197e-01 2.77328283e-01 -2.44941950e-01
3.40013951e-01 1.95106849e-01 -4.76945013e-01 -4.93923873e-01
-9.10514772e-01 8.95286083e-01 2.57813901e-01 1.23820035e-02
-4.94725466e-01 -1.07206964e+00 -9.38411772e-01 -3.67188096e-01
-3.73721838e-01 7.59853542e-01 1.98026210e-01 8.09238434e-01
2.53320605e-01 5.42361081e-01 8.69515836e-01 -4.96685565e-01
-7.45287061e-01 -1.33600521e+00 -9.25156057e-01 3.73581886e-01
3.72056365e-01 -5.09092271e-01 -7.33700097e-01 1.34223893e-01] | [4.40833044052124, 4.286250114440918] |
fb9c9529-71f5-4f99-9e22-8a4a66a72498 | robust-reinforcement-learning-with | 2206.06841 | null | https://arxiv.org/abs/2206.06841v1 | https://arxiv.org/pdf/2206.06841v1.pdf | Robust Reinforcement Learning with Distributional Risk-averse formulation | Robust Reinforcement Learning tries to make predictions more robust to changes in the dynamics or rewards of the system. This problem is particularly important when the dynamics and rewards of the environment are estimated from the data. In this paper, we approximate the Robust Reinforcement Learning constrained with a $\Phi$-divergence using an approximate Risk-Averse formulation. We show that the classical Reinforcement Learning formulation can be robustified using standard deviation penalization of the objective. Two algorithms based on Distributional Reinforcement Learning, one for discrete and one for continuous action spaces are proposed and tested in a classical Gym environment to demonstrate the robustness of the algorithms. | ['Erwan Le Pennec', 'Stéphanie Allassonière', 'Pierre Clavier'] | 2022-06-14 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [ 1.41520062e-02 4.30973291e-01 4.93680350e-02 -2.65986711e-01
-7.78254092e-01 -2.60005176e-01 3.49068105e-01 3.91679257e-01
-1.01874816e+00 1.45041215e+00 -1.02922745e-01 1.95786208e-01
-8.05852175e-01 -6.50906980e-01 -9.30449426e-01 -8.23585033e-01
-5.49497545e-01 3.42461675e-01 1.47218734e-01 -4.59589630e-01
3.86928171e-01 3.56704831e-01 -1.57513499e+00 -3.98754954e-01
9.34549153e-01 8.02187979e-01 2.17995182e-01 7.74935961e-01
4.89605814e-01 6.53147399e-01 -6.48564219e-01 9.83069539e-02
5.47943115e-01 -3.40290695e-01 -3.50758255e-01 -7.20130578e-02
-1.42812384e-02 -2.33051479e-01 -3.37854475e-01 1.25111520e+00
5.45942664e-01 9.83823061e-01 7.89122701e-01 -1.11211407e+00
-1.05520092e-01 5.16741574e-01 -2.01066658e-01 2.30800565e-02
2.45311052e-01 3.78496230e-01 5.91018975e-01 -1.94458127e-01
3.31692815e-01 1.29515588e+00 4.59987462e-01 5.37455976e-01
-1.46103501e+00 -3.24272752e-01 3.16523522e-01 1.02091223e-01
-1.26716185e+00 5.34792319e-02 3.51394475e-01 -4.82635558e-01
7.26310253e-01 -1.66241154e-02 5.40450692e-01 9.59384561e-01
6.66803658e-01 3.76928031e-01 1.48545742e+00 -2.24231288e-01
9.73876834e-01 3.39494497e-02 -3.06814402e-01 4.16877657e-01
1.54782787e-01 1.03187037e+00 -7.42990747e-02 -2.19739363e-01
5.04038274e-01 -2.04529643e-01 8.74379557e-03 -8.02811503e-01
-7.77612448e-01 1.02180755e+00 3.56785923e-01 -8.83326903e-02
-6.23030603e-01 4.16782975e-01 4.32693064e-01 4.62626100e-01
2.21229464e-01 6.41989231e-01 -2.41282552e-01 -4.67731386e-01
-6.21284485e-01 1.00012708e+00 6.14257395e-01 6.42386138e-01
3.93538922e-01 6.10120654e-01 -3.06089342e-01 5.51552355e-01
2.87174851e-01 5.67431152e-01 5.42193830e-01 -1.20516455e+00
3.35465819e-01 2.06697747e-01 5.84195197e-01 -6.47534788e-01
-4.53985900e-01 -1.18167296e-01 -2.62661844e-01 8.98714840e-01
5.71590543e-01 -5.88609636e-01 -7.02155352e-01 1.89275420e+00
4.99574363e-01 1.21590339e-01 2.17231046e-02 6.63969696e-01
-3.48744780e-01 4.50433224e-01 1.27319276e-01 -5.80750883e-01
2.37580225e-01 -3.81694049e-01 -9.68178272e-01 -5.89226410e-02
2.87319630e-01 -3.16475034e-01 1.05067337e+00 5.73979795e-01
-1.23587072e+00 -6.37005448e-01 -1.18998742e+00 6.91013634e-01
-2.85443455e-01 -3.93388689e-01 -1.80191711e-01 5.68797410e-01
-7.49259591e-01 1.32370925e+00 -9.02739763e-01 1.17739007e-01
5.02301427e-03 5.67236185e-01 -8.43045190e-02 4.07214105e-01
-1.21160400e+00 1.15590155e+00 8.37112129e-01 -3.92273962e-02
-1.10355723e+00 -3.18368763e-01 -8.27817142e-01 -2.01926589e-01
5.02431810e-01 -1.40374480e-02 1.33116877e+00 -6.90544963e-01
-2.02901530e+00 1.59535594e-02 7.34568119e-01 -7.04226911e-01
9.52706993e-01 -2.00391561e-01 2.07047965e-02 -1.78862866e-02
-1.05469041e-01 2.26025984e-01 1.16874909e+00 -7.57961154e-01
-5.84609091e-01 -4.69398052e-01 -2.69498937e-02 3.48556757e-01
1.77428588e-01 -4.68904883e-01 5.76766729e-01 -7.89579034e-01
-5.58419704e-01 -1.09544325e+00 -5.80348969e-01 -1.45717442e-01
1.00610335e-03 -2.21836269e-01 3.31952721e-01 -4.98440117e-01
1.03688085e+00 -2.22718835e+00 4.57956135e-01 4.29991454e-01
-5.42873144e-01 1.34876430e-01 -6.18964843e-02 4.40350324e-01
-5.71742952e-02 -2.02490315e-01 -3.91190141e-01 1.50695890e-01
3.69028509e-01 4.76079524e-01 -3.44274461e-01 7.09198594e-01
-3.47488299e-02 3.26321512e-01 -9.41457450e-01 -1.47880524e-01
1.34184942e-01 1.29118741e-01 -4.99684095e-01 5.35778046e-01
-3.58490169e-01 3.14757019e-01 -4.28691745e-01 2.10040864e-02
3.82982075e-01 7.91743517e-01 9.71443206e-02 4.66254145e-01
-1.90200627e-01 -2.79208779e-01 -1.61897790e+00 1.25396478e+00
-2.58282304e-01 -6.56055734e-02 1.23127401e-01 -1.46407413e+00
1.08444393e+00 1.81723967e-01 5.75914502e-01 -6.84435964e-01
1.55135512e-01 2.23921508e-01 1.50952712e-01 -3.69506150e-01
4.50818211e-01 -5.96208513e-01 -2.54133195e-01 2.80715585e-01
2.72876173e-01 -6.79154694e-01 2.16404572e-01 -2.76889026e-01
8.36122513e-01 5.81783772e-01 2.69406170e-01 -5.88394642e-01
2.39815667e-01 -2.10830137e-01 6.59320652e-01 7.93966830e-01
-5.57776034e-01 3.14462185e-03 4.79605407e-01 -1.30464405e-01
-7.24543869e-01 -1.33919668e+00 -1.58723235e-01 7.85768270e-01
-3.29386406e-02 -4.64456826e-02 -7.87868023e-01 -8.70488882e-01
4.53222394e-01 1.02350569e+00 -7.69474745e-01 -6.47928774e-01
-2.92145342e-01 -5.09594202e-01 1.41117156e-01 6.39702857e-01
2.93791872e-02 -9.41062570e-01 -9.56319928e-01 5.07644117e-01
2.92866170e-01 -5.45652926e-01 -3.14341933e-01 6.01875067e-01
-8.47716212e-01 -1.02291191e+00 -5.73596776e-01 -2.95799196e-01
3.68247688e-01 -6.78601921e-01 5.14455557e-01 -4.57170099e-01
-2.51903385e-01 8.53894293e-01 -2.13922992e-01 -7.04487681e-01
-3.92628014e-01 -5.56087017e-01 8.50983024e-01 -1.96085230e-01
-1.71822965e-01 -2.70593017e-01 -3.83293033e-01 3.12844872e-01
-8.43196273e-01 -8.40649664e-01 9.34747607e-02 1.12749016e+00
9.18372989e-01 4.53493655e-01 8.19687009e-01 -5.15755415e-01
9.03106987e-01 -3.38260025e-01 -1.01577413e+00 4.46306951e-02
-8.60740662e-01 5.00433624e-01 6.88030779e-01 -6.30278885e-01
-9.73321915e-01 2.48822242e-01 -4.72630449e-02 -1.95932969e-01
-1.79922506e-01 3.30227017e-01 1.50029033e-01 -2.42235586e-02
8.32735062e-01 4.98690605e-02 1.65763110e-01 -2.77968228e-01
3.12124372e-01 3.93802971e-01 2.19106004e-01 -8.50819349e-01
7.74896681e-01 2.24899556e-02 2.88693696e-01 -7.66629875e-01
-4.90539968e-01 6.34371936e-02 -4.90451962e-01 -4.28360164e-01
7.47563004e-01 -5.61580241e-01 -8.37202489e-01 3.03065717e-01
-3.13790828e-01 -8.65241587e-01 -1.18540180e+00 8.79534245e-01
-1.65801716e+00 1.28839359e-01 -3.03610981e-01 -1.24789846e+00
1.31457239e-01 -1.07271743e+00 4.55101192e-01 2.99383789e-01
5.90718314e-02 -1.02547956e+00 6.24997020e-01 -2.47789040e-01
2.96723455e-01 6.23627722e-01 5.76901615e-01 -6.00124002e-01
1.33713290e-01 -2.30669498e-01 5.13767362e-01 6.75739408e-01
1.14758397e-02 -2.47172877e-01 -6.51352346e-01 -7.12429643e-01
2.40297914e-01 -7.70892203e-01 4.56515461e-01 6.11062467e-01
9.71904635e-01 -4.65561211e-01 2.28489757e-01 9.90154371e-02
1.41225719e+00 5.80749035e-01 2.49292806e-01 4.84244049e-01
1.44523472e-01 7.00253189e-01 1.29360616e+00 9.11404371e-01
-1.39512882e-01 6.01743460e-01 7.18669295e-01 5.05260885e-01
6.72010303e-01 -4.99932438e-01 6.79316223e-01 2.79829323e-01
8.70417878e-02 3.70357275e-01 -5.89409351e-01 3.38207036e-01
-2.14151144e+00 -1.12230241e+00 5.07049203e-01 2.60329986e+00
8.74546349e-01 2.69238025e-01 6.00750268e-01 -5.76976016e-02
5.71149647e-01 -2.63279378e-01 -1.07410228e+00 -9.67908978e-01
2.50317097e-01 2.80804843e-01 5.43094754e-01 7.63627708e-01
-1.00411296e+00 3.51358980e-01 7.20365763e+00 8.55550051e-01
-7.05003440e-01 -3.45378876e-01 5.12396276e-01 -2.85430908e-01
1.82537362e-01 -3.17811221e-01 -6.08149886e-01 6.77713811e-01
1.34123111e+00 -4.12676752e-01 5.66772282e-01 9.86358643e-01
5.33611774e-01 -3.53281975e-01 -1.05805099e+00 5.59332669e-01
-3.94745618e-01 -6.68570220e-01 -5.93965590e-01 4.78866808e-02
7.01111078e-01 -7.28163123e-02 2.20476642e-01 9.19661760e-01
6.10678434e-01 -8.95185292e-01 8.29436660e-01 9.03589427e-01
3.26039672e-01 -1.29329765e+00 7.41605639e-01 7.14141905e-01
-8.89210939e-01 -4.90447909e-01 -4.34719026e-01 -3.37702066e-01
-6.39587194e-02 3.01126093e-02 -4.32781011e-01 2.62912869e-01
4.74233478e-01 3.78096610e-01 -2.98938155e-01 1.19856977e+00
-1.64036125e-01 4.31984931e-01 -4.45491225e-01 -3.16962332e-01
4.43314105e-01 -5.53092480e-01 7.34980285e-01 8.20828319e-01
3.75700682e-01 -5.81774265e-02 6.83289945e-01 7.41806567e-01
5.37712216e-01 1.70062095e-01 -8.42788637e-01 8.16826969e-02
-7.36597255e-02 8.02067518e-01 -4.24868524e-01 8.01905841e-02
2.31549397e-01 5.77226520e-01 5.42700112e-01 2.61569828e-01
-7.39894152e-01 -6.07080519e-01 6.00378096e-01 -6.76585957e-02
3.82239521e-01 -2.38797829e-01 3.13239098e-01 -7.85002887e-01
-1.30303413e-01 -8.05821538e-01 6.04153574e-01 -1.83667779e-01
-1.39400220e+00 2.22787604e-01 3.37859303e-01 -1.14588785e+00
-7.81870246e-01 -6.94303811e-01 -3.87081891e-01 8.14655185e-01
-1.23555899e+00 -1.75899774e-01 4.95106786e-01 8.34205568e-01
4.08677310e-01 -1.18830785e-01 7.57582486e-01 -2.64507592e-01
-5.81092358e-01 4.11636978e-01 8.56323004e-01 -4.25163060e-01
6.66751087e-01 -1.85939932e+00 -2.97069788e-01 4.54855323e-01
-3.79366100e-01 1.37590840e-01 1.24872434e+00 -5.97144961e-01
-1.15562665e+00 -1.06231940e+00 -1.72115676e-02 -1.34701043e-01
9.60327864e-01 -7.13728517e-02 -8.67052019e-01 5.41805804e-01
1.21039189e-01 7.73582757e-02 3.34096700e-01 -2.30150312e-01
2.32048646e-01 -1.14957109e-01 -1.55958617e+00 5.12943983e-01
3.69371384e-01 -2.60926813e-01 -7.54364729e-01 2.60281533e-01
4.68901038e-01 -4.53370601e-01 -1.10903764e+00 3.46662641e-01
3.82451773e-01 -5.39632499e-01 6.62594199e-01 -9.66587067e-01
-5.17785624e-02 -1.16374455e-01 -3.34426701e-01 -2.07188225e+00
-2.07469925e-01 -8.51227522e-01 -3.11463863e-01 7.20347106e-01
8.12062621e-02 -4.06080186e-01 5.04483104e-01 6.90980732e-01
1.33909225e-01 -7.53540516e-01 -1.39061975e+00 -1.36626923e+00
6.52880371e-01 -2.60509074e-01 6.02084473e-02 4.39141542e-01
6.13242328e-01 -2.11892024e-01 -4.79937553e-01 -5.62339909e-02
9.72889483e-01 -1.43949538e-01 4.45712239e-01 -8.58277082e-01
-6.40551984e-01 -3.05846542e-01 -5.15630960e-01 -3.13247681e-01
3.50520462e-01 -4.18718338e-01 6.30968571e-01 -1.03682256e+00
-3.80040020e-01 -9.79552716e-02 -5.84302366e-01 2.11667240e-01
-2.61482805e-01 -2.32882097e-01 1.61185279e-01 -5.90193868e-01
-5.14349639e-01 1.16542113e+00 9.68184710e-01 -1.17030069e-01
-4.06902105e-01 5.16838849e-01 -7.88110644e-02 9.01090443e-01
1.09198558e+00 -5.34344494e-01 -6.16931677e-01 2.12369949e-01
1.17720790e-01 2.68505216e-01 -3.64186876e-02 -1.13708913e+00
-2.05967695e-01 -6.01086855e-01 2.31909811e-01 -2.47797936e-01
2.78887510e-01 -8.58507812e-01 -1.97852939e-01 9.41400826e-01
-8.35814655e-01 2.20528156e-01 3.35553378e-01 1.08676493e+00
-1.86978742e-01 -6.50180936e-01 1.19862962e+00 -2.89248396e-02
-3.87102157e-01 2.44391531e-01 -6.78634226e-01 3.46510798e-01
1.37292182e+00 1.26052722e-01 2.16467395e-01 -5.14267147e-01
-1.10359418e+00 4.26864207e-01 6.43890947e-02 1.67997018e-01
7.42942870e-01 -1.23824036e+00 -5.04660010e-01 -1.65952090e-02
3.81598398e-02 -6.34072781e-01 2.61515260e-01 5.92295587e-01
-1.99506640e-01 -2.72095390e-02 -7.02207148e-01 -1.30245537e-01
-8.57584536e-01 8.03380191e-01 5.90740323e-01 -2.84043133e-01
-3.87622178e-01 3.11775446e-01 -3.89388055e-01 -5.80153942e-01
5.81654310e-01 -3.98010820e-01 -1.33194625e-01 1.23471683e-02
5.63005805e-01 6.46340013e-01 -1.96218148e-01 -2.95232683e-01
5.45095876e-02 2.30179340e-01 6.41309172e-02 -5.02572000e-01
1.39137828e+00 -6.46443218e-02 5.44374704e-01 7.96469212e-01
6.57322049e-01 -4.40192074e-01 -2.05188227e+00 -6.42585605e-02
1.70115188e-01 -5.00776231e-01 1.73710912e-01 -7.71304607e-01
-5.35870194e-01 8.02281439e-01 1.20799625e+00 2.99362600e-01
7.92865813e-01 -7.44990289e-01 3.83374803e-02 4.55764830e-01
5.33407032e-01 -2.02109480e+00 1.77013904e-01 5.62283933e-01
9.75636005e-01 -1.03197420e+00 2.67615281e-02 5.85358262e-01
-9.04519975e-01 1.04885125e+00 7.91481376e-01 -6.84315085e-01
7.59569347e-01 3.49262983e-01 -2.79204756e-01 3.71461809e-01
-6.01518214e-01 -2.69653708e-01 8.88204351e-02 9.29663599e-01
-5.43928370e-02 3.08628380e-01 -4.90844041e-01 3.51706535e-01
1.40690684e-01 -2.12654695e-01 6.62841439e-01 1.07099211e+00
-6.48527563e-01 -1.12229526e+00 -4.49631482e-01 1.63043976e-01
-3.46630335e-01 5.62054813e-01 -4.00010236e-02 7.59095609e-01
-6.07079305e-02 8.15808952e-01 -2.11143702e-01 -1.00539334e-01
7.44514942e-01 1.46647796e-01 6.04771495e-01 -5.17749310e-01
-4.46240038e-01 9.55892429e-02 -2.35765085e-01 -9.21189070e-01
-1.62203699e-01 -9.49171007e-01 -1.30953658e+00 5.65894879e-03
-2.32343465e-01 3.63607347e-01 5.90839505e-01 8.88290286e-01
-3.78268436e-02 6.12823248e-01 8.86908054e-01 -7.52454758e-01
-1.77938378e+00 -8.58041823e-01 -1.22470653e+00 2.86489099e-01
3.35100859e-01 -1.00994563e+00 -4.54579532e-01 -4.19579655e-01] | [4.292448997497559, 2.3087759017944336] |
e7cabd66-d233-4aaf-af28-8e0805ebf33e | multilingual-synthetic-question-and-answer | 2010.12008 | null | https://arxiv.org/abs/2010.12008v3 | https://arxiv.org/pdf/2010.12008v3.pdf | Towards Zero-Shot Multilingual Synthetic Question and Answer Generation for Cross-Lingual Reading Comprehension | We propose a simple method to generate multilingual question and answer pairs on a large scale through the use of a single generative model. These synthetic samples can be used to improve the zero-shot performance of multilingual QA models on target languages. Our proposed multi-task training of the generative model only requires the labeled training samples in English, thus removing the need for such samples in the target languages, making it applicable to far more languages than those with labeled data. Human evaluations indicate the majority of such samples are grammatically correct and sensible. Experimental results show our proposed approach can achieve large gains on the XQuAD dataset, reducing the gap between zero-shot and supervised performance of smaller QA models on various languages. | ['Linting Xue', 'Mihir Sanjay Kale', 'Noah Constant', 'Siamak Shakeri'] | 2020-10-22 | null | https://aclanthology.org/2021.inlg-1.4 | https://aclanthology.org/2021.inlg-1.4.pdf | inlg-acl-2021-8 | ['cross-lingual-question-answering'] | ['natural-language-processing'] | [-2.23078460e-01 5.77121139e-01 5.84864467e-02 -6.05083346e-01
-1.98953199e+00 -7.82891273e-01 6.72385037e-01 -2.25710273e-01
-4.39111412e-01 1.20171344e+00 1.93537652e-01 -5.24647534e-01
4.75002587e-01 -8.90556931e-01 -9.06877100e-01 -4.46309000e-01
6.30047023e-01 1.20562398e+00 1.23393171e-01 -7.53184736e-01
-3.79570603e-01 -4.59530771e-01 -1.41513228e+00 4.68444228e-01
1.74918056e+00 3.50878268e-01 2.41656512e-01 5.43172359e-01
-3.68628025e-01 6.79133534e-01 -8.76838863e-01 -1.01851499e+00
2.64301956e-01 -7.03923821e-01 -8.70434046e-01 -6.58450574e-02
9.20087039e-01 -1.44650996e-01 7.20945671e-02 1.08130682e+00
8.33776355e-01 -6.95191184e-03 6.47004783e-01 -1.03387856e+00
-9.85605955e-01 1.00719202e+00 -1.68355629e-01 -1.95867583e-01
6.82495475e-01 1.49182647e-01 1.21893370e+00 -1.02305150e+00
8.54789257e-01 1.64472175e+00 4.70900059e-01 6.08009875e-01
-1.13943076e+00 -5.86121023e-01 -3.90819967e-01 1.63186625e-01
-1.29315174e+00 -8.29383790e-01 4.40629274e-01 -1.05705000e-01
8.68094146e-01 1.09867536e-01 6.05735630e-02 1.21642315e+00
-1.88030489e-02 8.14924598e-01 1.43399298e+00 -8.35248888e-01
8.91332924e-02 3.92987221e-01 2.41899669e-01 7.98717141e-01
1.43594429e-01 -2.03242674e-01 -2.72295266e-01 -3.91958207e-01
1.46270245e-01 -8.50682080e-01 -3.19049954e-01 -2.15396330e-01
-1.05770981e+00 1.34455502e+00 7.61303380e-02 3.49249202e-03
7.74996057e-02 -2.27937311e-01 5.08170485e-01 7.10051775e-01
5.87033451e-01 4.12703335e-01 -4.65439856e-01 8.06757286e-02
-7.25515068e-01 4.13130313e-01 8.16506445e-01 1.31424463e+00
9.38550591e-01 1.73001930e-01 -3.64908487e-01 1.18598628e+00
2.43232995e-01 1.20525742e+00 4.50434804e-01 -9.72328365e-01
9.11179543e-01 5.72197378e-01 3.62547904e-01 -3.18544656e-01
2.30603516e-02 -2.51088649e-01 -3.82866859e-01 -1.70379847e-01
7.52013922e-01 -3.41186970e-01 -8.81621957e-01 1.86247516e+00
4.15779620e-01 -6.61783814e-01 5.67149639e-01 5.23560286e-01
9.88590479e-01 9.53692257e-01 1.19553626e-01 -3.99062820e-02
1.51821363e+00 -1.20884180e+00 -7.77428389e-01 -3.87406260e-01
1.13538074e+00 -9.66551900e-01 1.77355444e+00 -6.55376166e-02
-1.10794568e+00 -6.73029006e-01 -7.86092758e-01 -2.81507522e-01
-2.06218660e-01 2.47450262e-01 6.37302518e-01 1.10484362e+00
-9.99132156e-01 -2.71741580e-02 -2.89299875e-01 -3.54045898e-01
-3.40146869e-02 -3.72113436e-02 -1.96000457e-01 -5.91737449e-01
-1.65516698e+00 1.13578486e+00 4.47105318e-01 -2.85097301e-01
-1.10429752e+00 -3.56030434e-01 -1.36823511e+00 1.20230310e-01
2.17830718e-01 -6.81748927e-01 1.45599580e+00 -7.81696677e-01
-1.29721451e+00 9.37626362e-01 -3.79669517e-01 -3.71378332e-01
4.55151856e-01 -1.42774895e-01 -4.30247933e-01 2.63547897e-02
5.53429067e-01 9.27660942e-01 6.18045211e-01 -1.25772274e+00
-4.47725207e-01 5.43323951e-03 1.66846156e-01 3.49817395e-01
-1.01931944e-01 1.19005136e-01 -2.90516794e-01 -2.41929695e-01
-2.63918519e-01 -1.01655316e+00 -1.01704672e-01 -6.74794674e-01
-2.34145552e-01 -5.12432516e-01 2.44106367e-01 -1.04801953e+00
8.37747455e-01 -1.58600771e+00 -1.78678021e-01 -1.53508306e-01
-4.26500410e-01 1.52762473e-01 -5.48127353e-01 5.45430779e-01
1.85187116e-01 -4.24706712e-02 -2.30129242e-01 -3.75830710e-01
6.17758892e-02 2.85870969e-01 -5.71719527e-01 1.06319070e-01
1.91813298e-02 1.18100083e+00 -1.17852592e+00 -6.93926215e-01
-5.24654388e-02 1.33419245e-01 -5.82984030e-01 3.26185316e-01
-5.02013326e-01 3.95166457e-01 -1.25147626e-01 6.66797101e-01
5.17778218e-01 -1.83264896e-01 2.21476123e-01 8.38862583e-02
4.26450431e-01 6.62750483e-01 -9.15240526e-01 1.93176210e+00
-8.23003590e-01 3.94082010e-01 -3.03318948e-01 -4.34718370e-01
1.12928677e+00 6.70405447e-01 -2.68385440e-01 -9.60379004e-01
3.96618480e-03 5.75807929e-01 1.75573841e-01 -3.90720189e-01
6.69306815e-01 -1.63869709e-01 -6.78147078e-01 6.51346028e-01
5.05903065e-01 -4.95470375e-01 6.55076027e-01 3.48243237e-01
4.84125316e-01 2.38111854e-01 2.08103538e-01 -4.31889892e-01
5.74831843e-01 3.08566213e-01 5.02586007e-01 8.24760199e-01
-1.36798486e-01 2.65138209e-01 1.71847120e-01 -1.61191955e-01
-1.36085105e+00 -1.13494015e+00 -1.64399728e-01 1.29771781e+00
-1.27907410e-01 -2.72866905e-01 -1.03407073e+00 -1.00823534e+00
-4.05606747e-01 1.25570953e+00 -3.57634306e-01 -1.62187987e-03
-6.58824861e-01 -6.71612680e-01 8.46136332e-01 4.53798436e-02
4.02446091e-01 -1.01234305e+00 -9.48412716e-02 1.77902371e-01
-9.36296582e-01 -1.19798911e+00 -4.72374111e-01 -1.36327893e-01
-7.27076113e-01 -9.00549173e-01 -8.68127584e-01 -1.11953092e+00
7.43987858e-01 6.77552149e-02 1.59748280e+00 -3.76345873e-01
2.95430779e-01 1.38222128e-01 -4.44278091e-01 -3.02229226e-01
-1.12721527e+00 2.41169915e-01 -4.27429192e-02 -3.88608962e-01
5.22701144e-01 -1.43724963e-01 -1.18347570e-01 1.54913619e-01
-6.12280369e-01 1.54043794e-01 2.31267989e-01 1.03115904e+00
3.34904432e-01 -6.36266232e-01 1.11578143e+00 -1.18842494e+00
9.90790308e-01 -3.95233333e-01 -4.64694470e-01 9.34324920e-01
-3.24707448e-01 3.63018692e-01 7.80101418e-01 -4.28916246e-01
-1.35469019e+00 -2.94563502e-01 -2.33015612e-01 5.42364269e-02
-4.92354110e-02 3.79922450e-01 -3.72077733e-01 8.00284371e-02
1.08629215e+00 1.66784465e-01 -2.83966690e-01 -4.37678397e-01
9.74322736e-01 9.59546268e-01 4.20696169e-01 -8.92270684e-01
6.00172222e-01 -2.51652393e-02 -6.36916161e-01 -5.30867577e-01
-9.98461545e-01 -1.71510786e-01 -2.90936410e-01 -1.09339215e-01
6.03677809e-01 -1.31201124e+00 -4.19584289e-03 2.24724516e-01
-1.23712242e+00 -1.69146672e-01 -3.64232033e-01 2.48632744e-01
-5.30072510e-01 4.23225939e-01 -7.99464583e-01 -5.99499702e-01
-7.84039080e-01 -1.23263097e+00 1.28866279e+00 1.35277197e-01
-2.43247747e-01 -1.08853233e+00 5.17082453e-01 8.07664752e-01
1.53722882e-01 -2.60750741e-01 1.22123504e+00 -8.47424090e-01
-5.10004461e-01 -9.35307741e-02 9.71959010e-02 3.39314401e-01
2.84205880e-02 -3.92303556e-01 -1.04201329e+00 -4.93559927e-01
6.40359893e-02 -1.17068505e+00 4.92863834e-01 9.32074860e-02
2.45154127e-01 -2.07042888e-01 7.28482753e-02 1.56445757e-01
1.31155348e+00 -1.37143033e-02 5.65892994e-01 4.19075042e-02
4.79336679e-01 6.65281355e-01 9.47208881e-01 -1.59090698e-01
8.17983866e-01 6.10644877e-01 -1.71468630e-01 -2.01097224e-02
-3.52370322e-01 -5.20788312e-01 7.38370657e-01 1.44515204e+00
3.94012660e-01 -2.15232134e-01 -1.11239779e+00 9.60737228e-01
-1.60260522e+00 -6.92723513e-01 -2.04723224e-01 2.20400739e+00
1.32400811e+00 -6.30367622e-02 -2.68153995e-01 -4.51018959e-01
8.45988095e-01 -5.39092943e-02 -4.73663174e-02 -5.29380083e-01
-4.04273361e-01 5.09922028e-01 2.04088360e-01 9.68757391e-01
-8.12110126e-01 1.41901541e+00 7.05881023e+00 1.09182930e+00
-6.54452562e-01 6.69618666e-01 6.79948509e-01 2.23574057e-01
-8.85021627e-01 2.85753489e-01 -1.17482972e+00 3.18874240e-01
1.35529590e+00 -4.03479874e-01 1.98751509e-01 9.61149871e-01
-8.25853199e-02 -8.47587213e-02 -1.01965785e+00 7.53627539e-01
3.42609376e-01 -1.03750050e+00 5.03657162e-01 -2.43511736e-01
1.17575693e+00 1.36053741e-01 -2.58565664e-01 9.72750604e-01
8.92560601e-01 -9.43484306e-01 5.78957319e-01 1.99689671e-01
9.61953998e-01 -6.72774494e-01 7.02074707e-01 6.32720649e-01
-6.90008819e-01 2.86249965e-01 -7.23011732e-01 1.97978154e-01
5.95812857e-01 4.30334985e-01 -1.13964593e+00 6.08610868e-01
3.33354861e-01 -1.74710408e-01 -8.57342899e-01 7.19098210e-01
-5.41467428e-01 6.84955835e-01 -1.03269272e-01 -8.03047419e-02
1.96068183e-01 -4.08764631e-01 3.48880857e-01 9.74577129e-01
5.00074267e-01 -4.74086106e-01 1.98647097e-01 7.03414083e-01
-2.10006014e-01 6.07327938e-01 -8.02572548e-01 2.95409895e-02
5.01337945e-01 1.07093275e+00 -2.16890812e-01 -8.49147797e-01
-6.08505905e-01 8.65529418e-01 6.24169111e-01 3.84055793e-01
-7.61408806e-01 -3.53737205e-01 3.18903476e-02 -2.73245394e-01
-1.07915783e-02 -7.93051496e-02 -1.24265067e-01 -1.50719512e+00
8.82443115e-02 -1.61683238e+00 4.48176444e-01 -8.55516911e-01
-1.26334774e+00 9.07367647e-01 -1.08250134e-01 -1.23513556e+00
-9.16074932e-01 -3.61152858e-01 -3.28651220e-01 1.29627180e+00
-1.38773286e+00 -1.46147573e+00 -2.91103516e-02 6.18587613e-01
7.89009333e-01 -4.01783526e-01 1.26292896e+00 3.99678916e-01
-1.12597808e-01 8.22296441e-01 1.30871087e-01 1.07751243e-01
1.16084409e+00 -1.30829132e+00 7.09004879e-01 1.02728927e+00
4.95291770e-01 6.25746071e-01 6.86021268e-01 -7.10259795e-01
-8.38469148e-01 -8.46000493e-01 1.62675464e+00 -7.32307911e-01
5.35822153e-01 -5.72649479e-01 -8.47300291e-01 7.57105708e-01
7.87554741e-01 -5.49439728e-01 8.14865887e-01 2.51979858e-01
-3.99546206e-01 6.71010166e-02 -1.13248789e+00 6.93681240e-01
5.78446507e-01 -7.79612422e-01 -9.94390726e-01 7.60634422e-01
8.54667902e-01 -3.28716278e-01 -8.72787654e-01 4.01782185e-01
-5.90327457e-02 -5.35688877e-01 6.93977594e-01 -7.39671707e-01
4.42402750e-01 -1.45069033e-01 -2.10591286e-01 -1.58615518e+00
2.23002821e-01 -4.85806972e-01 4.14066650e-02 1.32269681e+00
8.88800561e-01 -5.65758646e-01 3.18655223e-01 5.89082181e-01
-1.92302451e-01 -3.50782126e-01 -1.09408271e+00 -7.22399652e-01
4.86679941e-01 -1.88555226e-01 5.61882377e-01 1.16012585e+00
-1.49461433e-01 9.39274371e-01 -5.16090035e-01 -8.89552161e-02
6.64290965e-01 3.45627964e-01 1.02185380e+00 -7.64059663e-01
-5.04157245e-01 1.73576087e-01 2.46352911e-01 -1.06900787e+00
3.12145621e-01 -1.19823325e+00 1.06464654e-01 -1.54496932e+00
2.26856813e-01 -6.63326442e-01 3.87625843e-01 3.94943506e-01
-7.01424956e-01 2.58375466e-01 4.78815213e-02 1.07325412e-01
-5.73558390e-01 8.58298123e-01 1.35483468e+00 -1.72968179e-01
7.73726255e-02 -2.35098988e-01 -7.03333795e-01 4.72853273e-01
5.47120929e-01 -6.32165849e-01 -6.33032203e-01 -7.89195776e-01
3.36747020e-01 1.64121106e-01 -2.19014585e-01 -7.11911380e-01
-1.87094569e-01 1.15831658e-01 -6.05047680e-03 -2.95772940e-01
3.20528686e-01 -3.08845729e-01 -1.57059267e-01 2.93711960e-01
-1.54470623e-01 1.36423662e-01 3.01934686e-02 1.12521417e-01
-4.07541007e-01 -6.07794344e-01 6.88045084e-01 -3.43766272e-01
-5.58383584e-01 -8.04684460e-02 1.22076977e-04 6.94686115e-01
7.15217054e-01 1.99529290e-01 -6.61828995e-01 -6.59369171e-01
-5.81883788e-01 4.49124277e-01 5.01236618e-01 6.21032059e-01
1.94622219e-01 -1.52973676e+00 -1.23262548e+00 1.56855083e-03
4.35253024e-01 -2.44066358e-01 1.99009508e-01 3.86990219e-01
-6.43487334e-01 7.86122561e-01 -1.14286542e-01 -5.35254657e-01
-9.86524999e-01 3.51316392e-01 3.88144553e-01 -6.37826025e-01
-2.51383364e-01 6.75306499e-01 1.23208374e-01 -1.26332128e+00
-1.29800662e-01 1.11277707e-01 -1.62742034e-01 8.41476992e-02
2.94954002e-01 9.46072266e-02 8.58914405e-02 -7.38555968e-01
1.25321761e-01 1.15862757e-01 -1.76118895e-01 -5.83827496e-01
8.83411527e-01 -1.43595695e-01 -1.10508360e-01 5.82119763e-01
9.01316822e-01 3.92967105e-01 -7.02270508e-01 -3.84011656e-01
1.76581144e-02 -2.54649609e-01 -5.51172674e-01 -8.65884125e-01
-5.36215425e-01 9.98558223e-01 4.10413772e-01 -1.52500123e-01
6.14802301e-01 2.33385652e-01 1.03651381e+00 6.56030536e-01
8.17958176e-01 -1.09545052e+00 -1.06333159e-01 7.55753040e-01
8.89931083e-01 -1.45947397e+00 -5.39280713e-01 -4.44212526e-01
-8.57795537e-01 6.71315253e-01 7.55582392e-01 1.11697100e-01
-6.88114464e-02 -1.47792786e-01 6.41007841e-01 5.54436296e-02
-8.37640345e-01 -2.11210445e-01 2.36183986e-01 6.32490873e-01
6.83999240e-01 2.80085981e-01 -6.03625655e-01 6.05872512e-01
-6.19499445e-01 -4.39274490e-01 4.82551038e-01 6.20428741e-01
-5.21965265e-01 -1.57698882e+00 -3.61503869e-01 2.83291370e-01
-3.21779609e-01 -6.56709254e-01 -2.46645674e-01 7.58210182e-01
-7.25486353e-02 1.12537706e+00 -1.39444232e-01 1.15544990e-01
2.13668451e-01 7.45821357e-01 6.17037892e-01 -8.53571773e-01
-6.09686852e-01 -7.77213871e-02 7.06673741e-01 -1.79785103e-01
-1.55933022e-01 -4.94865805e-01 -8.85606229e-01 -6.43980652e-02
-3.98397982e-01 5.47195375e-01 3.94590884e-01 1.02431500e+00
2.18614727e-01 1.24804489e-01 5.06569266e-01 -1.27604589e-01
-1.04414165e+00 -1.53735304e+00 -1.70138359e-01 6.67441845e-01
-1.86257929e-01 -3.31926733e-01 -2.86910385e-01 1.36026023e-02] | [11.411441802978516, 8.423086166381836] |
97e8adfc-ef1f-4e52-9a32-fef3df13bc5d | mask2former-for-video-instance-segmentation | 2112.10764 | null | https://arxiv.org/abs/2112.10764v1 | https://arxiv.org/pdf/2112.10764v1.pdf | Mask2Former for Video Instance Segmentation | We find Mask2Former also achieves state-of-the-art performance on video instance segmentation without modifying the architecture, the loss or even the training pipeline. In this report, we show universal image segmentation architectures trivially generalize to video segmentation by directly predicting 3D segmentation volumes. Specifically, Mask2Former sets a new state-of-the-art of 60.4 AP on YouTubeVIS-2019 and 52.6 AP on YouTubeVIS-2021. We believe Mask2Former is also capable of handling video semantic and panoptic segmentation, given its versatility in image segmentation. We hope this will make state-of-the-art video segmentation research more accessible and bring more attention to designing universal image and video segmentation architectures. | ['Alexander G. Schwing', 'Rohit Girdhar', 'Alexander Kirillov', 'Ishan Misra', 'Anwesa Choudhuri', 'Bowen Cheng'] | 2021-12-20 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 4.39240560e-02 1.86741754e-01 -4.40881222e-01 -3.91758323e-01
-7.95634151e-01 -7.69043982e-01 1.46246389e-01 -4.90970224e-01
-3.23914766e-01 3.63359809e-01 -3.14368516e-01 -6.51566267e-01
5.01454949e-01 -6.49738252e-01 -8.94732416e-01 -3.49050283e-01
-1.91513062e-01 3.64600569e-01 7.45225012e-01 1.59580290e-01
1.38897091e-01 5.52896976e-01 -1.18960750e+00 4.67136025e-01
8.13169122e-01 1.48579526e+00 8.84054303e-02 1.07310402e+00
-2.29141429e-01 5.43096840e-01 -2.84450978e-01 -4.29803342e-01
7.64904737e-01 -4.30299520e-01 -1.18277073e+00 4.21337187e-01
1.04297185e+00 -7.61708915e-01 -7.67887712e-01 9.55539107e-01
4.54237163e-02 9.04809311e-02 4.87205505e-01 -1.17365372e+00
-4.69371438e-01 5.95899999e-01 -6.87797427e-01 7.47996449e-01
2.96010822e-02 5.47429144e-01 7.81895876e-01 -3.83685082e-01
6.72696233e-01 9.24742460e-01 7.54678428e-01 7.47282922e-01
-1.01832604e+00 -3.61051500e-01 5.16300082e-01 6.95317909e-02
-1.15301335e+00 -3.42222780e-01 2.35069603e-01 -4.92740840e-01
1.08057106e+00 3.62289190e-01 9.40646946e-01 5.82309306e-01
8.20910335e-02 1.27723098e+00 6.53398752e-01 2.09661305e-01
1.10358573e-01 -3.09473753e-01 1.26076639e-01 8.77371550e-01
-4.10576686e-02 -3.74830693e-01 -1.57203957e-01 5.35327017e-01
1.26010227e+00 -1.70963883e-01 -2.43358970e-01 -7.21255839e-02
-1.12168598e+00 5.34331381e-01 5.00501871e-01 1.71017215e-01
-1.75746530e-01 6.70393407e-01 5.66676974e-01 1.20841853e-01
7.37510383e-01 3.78394127e-01 -6.44020855e-01 -3.27695251e-01
-1.76421678e+00 1.46618843e-01 5.95490992e-01 1.37812269e+00
6.45251513e-01 3.72458845e-01 -4.96081039e-02 4.31254447e-01
1.65265992e-01 4.80893135e-01 2.44186208e-01 -1.63084495e+00
3.95777196e-01 2.83867359e-01 -1.99925229e-01 -3.58819395e-01
-4.77943331e-01 -1.39692187e-01 -3.83693278e-01 -7.60091990e-02
6.10172868e-01 -3.52431029e-01 -1.55914986e+00 1.16654480e+00
1.38510346e-01 4.99715567e-01 -2.79768854e-01 1.05924523e+00
1.08650851e+00 7.99094141e-01 7.09068477e-02 -1.32567674e-01
1.23835468e+00 -1.42040110e+00 -2.54961014e-01 -2.51432419e-01
5.25176823e-01 -6.21950865e-01 8.92583668e-01 4.98738050e-01
-1.63207126e+00 -5.81430018e-01 -8.93937349e-01 -3.91596138e-01
-8.83780792e-02 -2.40590394e-01 9.34990406e-01 7.52305269e-01
-1.24697268e+00 6.38618290e-01 -1.17732382e+00 -3.98213238e-01
8.66174996e-01 4.32779610e-01 -3.01731825e-02 1.61710396e-01
-6.56174958e-01 3.71791095e-01 2.61077017e-01 -1.13549434e-01
-1.00832033e+00 -1.10290551e+00 -5.84592700e-01 -9.77454409e-02
4.96484846e-01 -7.45452881e-01 1.57968855e+00 -1.11040020e+00
-1.38594592e+00 1.20425701e+00 -2.08542999e-02 -7.17118800e-01
7.96013653e-01 -1.56552792e-01 -2.27873623e-01 8.11243594e-01
7.73833096e-02 1.47267687e+00 7.61953890e-01 -1.05458653e+00
-8.70324194e-01 -2.36605838e-01 4.11306210e-02 7.85877705e-02
1.19247404e-03 -1.82426929e-01 -1.30314314e+00 -4.80365545e-01
2.46439576e-01 -8.34785938e-01 -3.02843124e-01 1.73702493e-01
-4.66895759e-01 6.18465804e-03 1.09393406e+00 -6.84766293e-01
1.08648527e+00 -1.89317238e+00 5.19701168e-02 -2.25454327e-02
3.52143198e-01 3.79121840e-01 -2.81474262e-01 -3.91325086e-01
3.27202946e-01 4.84728873e-01 -2.25607172e-01 -2.52846479e-01
-3.89453560e-01 1.31223559e-01 -1.57640710e-01 6.03423893e-01
-3.52270305e-02 1.15034664e+00 -5.23318708e-01 -5.94366491e-01
5.84624350e-01 2.51168251e-01 -8.36696684e-01 7.92216584e-02
-5.69119632e-01 4.01981801e-01 -6.06210411e-01 9.39375222e-01
8.29553127e-01 -4.03047979e-01 -1.02344431e-01 -2.36779362e-01
-2.33201876e-01 -9.82635990e-02 -6.18690014e-01 1.95670080e+00
-1.79421324e-02 1.01344359e+00 4.42785650e-01 -8.60204518e-01
4.23766583e-01 5.51990457e-02 1.11228859e+00 -6.48451149e-01
2.92824030e-01 3.45666744e-02 -1.36812612e-01 -5.46893597e-01
6.74193442e-01 2.01554999e-01 3.03463787e-01 1.87115595e-01
2.11027980e-01 -5.38615942e-01 5.64573765e-01 3.22219640e-01
8.74578953e-01 1.27659991e-01 -4.10850018e-01 -5.68290234e-01
3.42467070e-01 1.72708079e-01 2.62616247e-01 6.43844485e-01
-5.74364901e-01 1.08146954e+00 6.66960657e-01 -4.23947990e-01
-1.29486191e+00 -1.32948530e+00 -3.83150160e-01 9.33596075e-01
3.93023491e-01 -3.67995977e-01 -1.37660861e+00 -9.27847028e-01
-2.62416266e-02 4.85800505e-01 -4.75269109e-01 4.78085071e-01
-7.04142332e-01 -5.21630526e-01 7.26364315e-01 8.07723701e-01
7.45836973e-01 -7.80933797e-01 -6.49908364e-01 7.99715668e-02
-9.08417627e-02 -1.56497669e+00 -7.25975215e-01 -3.73371579e-02
-1.36696970e+00 -1.17877257e+00 -1.00986826e+00 -8.62320364e-01
5.35765469e-01 4.27366763e-01 1.17911649e+00 2.29740471e-01
-3.23793262e-01 6.28125668e-01 -2.88367003e-01 1.80562288e-01
-7.72930160e-02 5.31148732e-01 -3.11266392e-01 -4.82494950e-01
2.31945887e-01 -3.54928374e-01 -1.11862469e+00 4.43385929e-01
-9.73387063e-01 2.22110093e-01 -5.00276163e-02 1.15397781e-01
9.69908535e-01 -3.66044343e-01 8.76293182e-02 -8.96684527e-01
8.96066353e-02 -4.05593365e-01 -6.68544412e-01 3.40318307e-02
-4.57866967e-01 -4.25010592e-01 5.59185326e-01 -2.03660697e-01
-7.51941502e-01 -1.67650327e-01 -5.31822920e-01 -6.16187096e-01
-1.37815520e-01 -1.29676461e-01 1.66065231e-01 -1.89303145e-01
3.74263287e-01 6.95448145e-02 -1.10363387e-01 -2.90274531e-01
5.56880355e-01 4.89879698e-01 6.89267635e-01 -4.25323546e-01
3.18346262e-01 7.12405443e-01 -1.95606038e-01 -9.81880188e-01
-7.05429673e-01 -6.48865819e-01 -6.47346377e-01 -4.51947063e-01
1.57676041e+00 -9.71420705e-01 -5.69635987e-01 5.43961048e-01
-9.07756031e-01 -9.72531378e-01 -2.79819548e-01 1.01753049e-01
-8.34730089e-01 3.47271234e-01 -1.21600580e+00 -2.39247471e-01
-5.35759449e-01 -1.67650449e+00 1.12416029e+00 5.02092063e-01
-1.69113040e-01 -1.01871264e+00 -4.90056902e-01 7.66057551e-01
2.54777908e-01 2.40405276e-02 3.77941072e-01 -4.32780176e-01
-9.97363031e-01 1.89624205e-01 -6.50860965e-01 4.73688513e-01
-1.30657300e-01 3.86818945e-01 -8.51915836e-01 -5.12144081e-02
-1.85096636e-01 -1.50611058e-01 1.07451749e+00 1.10487127e+00
1.76668906e+00 1.14920996e-01 -3.41990054e-01 1.31494689e+00
1.27762866e+00 3.08550596e-01 9.74100113e-01 3.68197352e-01
1.03252685e+00 1.21865548e-01 2.62914091e-01 9.25903097e-02
4.18169051e-01 4.50328320e-01 3.42701972e-01 -3.06996673e-01
-4.74054188e-01 -2.95480434e-02 1.11162044e-01 7.38438606e-01
-1.51372790e-01 -4.61532295e-01 -9.33023930e-01 3.28281641e-01
-1.70936120e+00 -8.99972796e-01 -4.43903357e-01 1.59078121e+00
5.65545857e-01 3.38637352e-01 2.90319055e-01 -4.45117742e-01
5.13004720e-01 3.32576990e-01 -8.39372218e-01 -5.32769203e-01
-6.36722520e-02 9.07041803e-02 1.01509249e+00 5.25015652e-01
-1.42488706e+00 1.48680913e+00 7.56281424e+00 9.46656048e-01
-1.25371063e+00 1.05998099e-01 1.44614339e+00 -3.63335520e-01
-3.19755316e-01 -2.94000894e-01 -6.62482500e-01 4.94403034e-01
8.68465841e-01 1.84757769e-01 5.01493871e-01 8.67241919e-01
1.58848390e-01 -3.52971226e-01 -1.00918996e+00 1.12626421e+00
1.28891943e-02 -1.79294300e+00 9.81507525e-02 1.21213414e-01
1.00222611e+00 4.70712900e-01 1.49458289e-01 1.38074115e-01
-2.17150092e-01 -1.02912176e+00 8.20855141e-01 2.74135560e-01
1.13857114e+00 -5.30146241e-01 3.30014437e-01 -1.62128165e-01
-1.15354443e+00 2.65312344e-01 -1.88099638e-01 2.50959098e-01
6.74686611e-01 3.12725961e-01 -3.70531917e-01 6.49605936e-04
9.22131360e-01 8.70537579e-01 -4.81960505e-01 1.10985792e+00
1.23010546e-01 6.69133961e-01 -4.05613333e-01 3.63152444e-01
7.93429613e-01 -2.40602672e-01 3.02669168e-01 1.50842309e+00
1.25657097e-01 4.66061652e-01 4.28832620e-01 6.84869170e-01
-5.41462064e-01 -3.72132771e-02 -1.95960298e-01 -1.62434578e-01
7.51578733e-02 1.18030548e+00 -1.57192707e+00 -8.33816826e-01
-4.96094048e-01 1.21075451e+00 -1.89251885e-01 5.74991107e-01
-1.17076373e+00 1.17204957e-01 9.17010248e-01 3.75992030e-01
6.75317824e-01 -4.91775423e-01 -7.78257489e-01 -1.31405342e+00
-3.40240210e-01 -5.41813195e-01 1.31377056e-01 -7.24661648e-01
-1.10102344e+00 6.91856265e-01 7.93676004e-02 -6.87952518e-01
-1.73042319e-03 -8.84813428e-01 -4.74595219e-01 2.02627361e-01
-1.30363488e+00 -8.36731672e-01 -2.61908621e-01 5.53332388e-01
1.26283967e+00 1.20809600e-02 1.20676465e-01 3.21133047e-01
-7.41174221e-01 6.45615578e-01 5.41801713e-02 3.39565337e-01
3.37867469e-01 -1.13492107e+00 8.08703959e-01 9.34935629e-01
1.89782575e-01 9.95030552e-02 4.36834127e-01 -6.79481506e-01
-1.46420467e+00 -1.10262823e+00 -1.75457805e-01 -4.23148006e-01
7.64426112e-01 -1.12970136e-01 -6.06406987e-01 8.66678953e-01
4.11356062e-01 1.55541584e-01 4.54091817e-01 -2.35516801e-01
-7.19822496e-02 -2.48890147e-02 -1.35351598e+00 6.88581467e-01
1.33905494e+00 -2.91907668e-01 -1.32482544e-01 3.83789062e-01
1.00507164e+00 -8.93468857e-01 -9.61717129e-01 2.69283086e-01
4.32751745e-01 -1.05018389e+00 9.62126911e-01 -5.10692120e-01
6.67103529e-01 -1.27148656e-02 -1.75058335e-01 -7.49774039e-01
-1.30322546e-01 -5.84601164e-01 -3.74093801e-01 8.53767395e-01
5.08941472e-01 -2.90438086e-01 1.47317731e+00 9.30914819e-01
-5.64446628e-01 -1.02203751e+00 -8.49873364e-01 -5.83398104e-01
3.95611256e-01 -7.46785343e-01 1.94173634e-01 6.98143780e-01
-3.27514172e-01 -1.59519494e-01 2.82833632e-02 -2.01136112e-01
4.83702719e-01 4.56618331e-02 5.81695974e-01 -6.45430088e-01
-7.68715739e-02 -8.55609119e-01 -4.05676842e-01 -1.87941658e+00
1.11054219e-01 -8.24873388e-01 -6.76703975e-02 -1.63458860e+00
-2.33888496e-02 -3.21349800e-01 7.21475333e-02 2.73544401e-01
1.39091834e-01 8.38237286e-01 5.37398577e-01 1.44415021e-01
-9.65171635e-01 1.56878442e-01 1.65544748e+00 -2.83517390e-01
-2.29123712e-01 -1.50300950e-01 -6.10010624e-01 7.79889882e-01
9.64732230e-01 -2.77235415e-02 -4.91357356e-01 -7.85231769e-01
-4.28608745e-01 6.88343421e-02 9.03664082e-02 -9.95126426e-01
4.24009003e-02 -2.30792150e-01 5.33895969e-01 -5.71487248e-01
2.80026197e-01 -6.18500650e-01 -2.59954840e-01 2.51875162e-01
-8.62418488e-02 9.71564651e-02 6.20538354e-01 6.29767850e-02
-9.07677338e-02 -1.34807467e-01 7.74650455e-01 -4.23860252e-01
-1.19663656e+00 7.70849228e-01 -5.54249585e-01 6.61277473e-01
1.26125872e+00 -6.39114857e-01 -3.16479892e-01 -7.71724209e-02
-8.10168326e-01 6.26010299e-01 7.17591763e-01 3.10845912e-01
5.35666943e-01 -8.15965474e-01 -5.14432728e-01 1.26625687e-01
-5.33670545e-01 1.50200650e-01 4.24212694e-01 8.30699921e-01
-1.36251962e+00 4.83703017e-01 -2.19244301e-01 -9.94110405e-01
-9.94342029e-01 3.31622213e-01 5.42182505e-01 2.21534833e-01
-9.36420023e-01 1.19076312e+00 2.73893952e-01 9.19821858e-02
2.87296832e-01 -5.91504574e-01 3.35863203e-01 -4.99870442e-03
4.19432342e-01 5.33738971e-01 -1.45871744e-01 -4.62312251e-01
-3.66609067e-01 6.57349348e-01 -1.59215182e-01 -9.37820598e-03
9.95115578e-01 -2.18187034e-01 1.00513235e-01 1.80762082e-01
1.29695308e+00 -4.02920842e-01 -1.91541851e+00 5.24162889e-01
-3.73666316e-01 -4.44386274e-01 3.52230310e-01 -6.02372110e-01
-1.92996061e+00 8.45478058e-01 4.88421887e-01 2.89448649e-01
1.15533602e+00 2.08061442e-01 1.43653846e+00 6.29842952e-02
3.52253586e-01 -1.28452909e+00 -1.41983435e-01 3.81305844e-01
3.16980571e-01 -1.31140673e+00 1.90844014e-01 -6.17105007e-01
-6.16012692e-01 1.20226347e+00 9.04258728e-01 -4.19394165e-01
6.82133913e-01 4.58248228e-01 9.51415449e-02 -1.91247314e-01
-4.17622060e-01 -1.81148991e-01 4.59018707e-01 3.98431182e-01
6.24803245e-01 1.79961890e-01 -2.17324257e-01 2.29983777e-01
-9.35928300e-02 -8.81681517e-02 5.61992228e-01 4.82188970e-01
-5.26696026e-01 -7.50999272e-01 -1.14290461e-01 7.59511948e-01
-5.74406147e-01 -3.04420382e-01 -7.90486708e-02 8.72838259e-01
1.46875516e-01 7.02139795e-01 6.73313856e-01 -1.94478944e-01
1.64161250e-01 -1.47655174e-01 7.59431124e-01 -3.57531250e-01
-6.70826375e-01 3.47851753e-01 8.29646550e-03 -1.02243078e+00
-3.40013862e-01 -5.85382581e-01 -1.72869074e+00 -7.43054092e-01
4.54346016e-02 -1.34041816e-01 5.78713417e-01 7.46372461e-01
2.93981552e-01 6.10292137e-01 2.80468836e-02 -1.18344569e+00
1.38382718e-01 -5.14687777e-01 -5.78530490e-01 1.15689225e-01
1.66063428e-01 -2.05110595e-01 -7.44017586e-02 2.56082952e-01] | [9.262736320495605, 0.022270046174526215] |
5f1e30be-315f-4aef-994d-0d7275205344 | trigonometric-comparison-measure-a-feature | null | null | https://doi.org/10.1016/j.datak.2018.10.003 | https://doi.org/10.1016/j.datak.2018.10.003 | Trigonometric comparison measure: A feature selection method for text categorization | Text data represented using vector space model is high dimensional data since the number of words can easily grow to tens of thousands for a moderate sized dataset. It may contain lots of redundant or irrelevant features that degrade the performance of a classifier for text categorization. To address this problem, feature selection can be applied for dimensionality reduction and it aims to find a set of highly distinguishing features. Most of filter feature selection methods for text categorization are based on document frequencies in positive and negative classes. Considering only document frequencies favors terms frequently used in a larger class and ignores relative document frequencies in the classes. In this paper, we present a new filter feature selection method, named Trigonometric Comparison Measure (TCM) considering relative document frequencies. The proposed method utilizes true positive rate and false positive rate to determine a better subset of features for text categorization and prefers terms that appear only in documents of one class with high probability. In order to assign a higher rank to terms that are frequently used in one class and rarely appears in another class, TCM calculates off-axis angles of a vector represented as and gives a larger score to terms with a small angle using and functions. The proposed method is compared with eight well-known filter feature selection methods including balanced accuracy measure (ACC2), information gain (IG), chi-squared (CHI), odds ratio (OR), Gini index (Gini), Deviation from a Poisson distribution (DP), distinguishing feature selector (DFS) and normalized difference measure (NDM) on ten datasets using the multinomial naïve Bayes and support vector machines. The experimental results show that TCM achieves significantly better performance for text categorization. | ['See Young Zzang', 'Kyoungok Kim'] | 2019-01-02 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 6.39049162e-04 -5.60764551e-01 -1.84945539e-01 -5.25288463e-01
-2.22775295e-01 -5.84155738e-01 6.10013366e-01 9.15681422e-01
-6.39640093e-01 7.68081427e-01 2.41643950e-01 -3.35571647e-01
-7.41713166e-01 -1.05260229e+00 1.78440511e-01 -6.02978289e-01
-1.17085531e-01 6.43405676e-01 3.11192334e-01 -2.04309911e-01
8.02861214e-01 4.79017198e-01 -2.14650846e+00 3.70497972e-01
7.09581137e-01 9.70314741e-01 6.51085973e-02 5.81844568e-01
-4.35800374e-01 2.13486820e-01 -8.30099225e-01 9.37994346e-02
2.07245216e-01 -4.09965217e-01 -8.52562964e-01 -3.38124722e-01
7.94203952e-03 -5.71068302e-02 6.89207911e-02 9.29651380e-01
2.47709602e-01 5.33343017e-01 1.34639323e+00 -1.29597485e+00
-3.99794489e-01 5.28045475e-01 -7.26969242e-01 5.37410319e-01
4.63214964e-01 -6.67148471e-01 8.76147509e-01 -8.96462023e-01
4.12262142e-01 1.38451695e+00 3.68056566e-01 -1.00499012e-01
-9.76791680e-01 -7.07310557e-01 1.08757704e-01 2.31601357e-01
-1.47883236e+00 2.09859759e-01 4.32829082e-01 -4.21910197e-01
1.08030427e+00 7.28029251e-01 5.39177120e-01 2.96660930e-01
6.43864930e-01 3.08718443e-01 9.52519476e-01 -8.04744124e-01
1.69981107e-01 5.78369796e-01 9.50757086e-01 3.34218800e-01
6.70436740e-01 -9.58198756e-02 -3.52965891e-01 -5.36243320e-01
-1.59467220e-01 3.25876176e-01 1.05142426e-02 1.93243429e-01
-1.02645373e+00 1.19082701e+00 6.20381124e-02 8.17975223e-01
-2.25813583e-01 -4.35137212e-01 6.03920758e-01 5.15893638e-01
1.47613853e-01 2.63839662e-01 -6.19717360e-01 -1.38320461e-01
-7.45362043e-01 3.07549357e-01 8.00615549e-01 6.10934913e-01
7.52220809e-01 -3.55453044e-01 -4.00545835e-01 8.20177615e-01
4.18199360e-01 3.51689875e-01 1.14628291e+00 6.68481439e-02
1.63828284e-01 1.21206605e+00 -1.35496035e-01 -1.34521151e+00
-6.83949769e-01 -1.13858804e-01 -6.93696797e-01 1.10123284e-01
1.13712326e-01 1.38447648e-02 -7.66471088e-01 1.24565279e+00
4.26276773e-01 -7.92380691e-01 -3.25328894e-02 6.16017222e-01
6.79583788e-01 7.86062419e-01 1.91043988e-02 -5.63733339e-01
1.63309014e+00 -1.07576527e-01 -6.06094480e-01 3.56337726e-01
7.08634853e-01 -1.14639461e+00 9.73944128e-01 6.25615120e-01
-2.80723453e-01 -4.51544881e-01 -9.53438580e-01 3.50974143e-01
-8.88292670e-01 8.00526440e-02 5.48479557e-01 7.88961530e-01
-4.83110994e-01 5.26144981e-01 -3.62227201e-01 -6.17880881e-01
1.23888075e-01 5.35400808e-01 -5.01961350e-01 1.13750882e-01
-1.33378768e+00 7.05312490e-01 6.07867241e-01 -6.47076368e-01
-3.79905179e-02 -2.85890877e-01 -5.34489453e-01 1.38752028e-01
-5.09272292e-02 -2.96317726e-01 5.20585001e-01 -8.38395417e-01
-8.85357082e-01 4.65021461e-01 -9.68430191e-02 -1.60607502e-01
2.14741230e-01 1.77837744e-01 -6.69656038e-01 3.29918303e-02
9.70841348e-02 2.32722700e-01 7.30431795e-01 -6.88617229e-01
-1.26963067e+00 -7.96605587e-01 -3.07241410e-01 3.27281088e-01
-6.81335926e-01 1.11045152e-01 3.55976820e-01 -6.86472178e-01
3.80275965e-01 -6.43199384e-01 2.21886367e-01 -4.03317153e-01
-5.04759252e-01 -8.02349746e-01 1.24629927e+00 -2.34617636e-01
1.57973790e+00 -1.92529333e+00 -1.59751862e-01 6.75642014e-01
4.53343131e-02 1.41966626e-01 3.64243299e-01 6.19901717e-01
-1.57901514e-02 3.03236365e-01 1.08721793e-01 6.41164303e-01
-2.47373998e-01 -1.88004121e-01 3.05322874e-02 4.59076852e-01
-1.07103184e-01 -1.14966452e-01 -5.73211372e-01 -6.54132068e-01
3.40113252e-01 3.94377559e-01 -2.67615825e-01 -1.02958106e-01
4.10419852e-01 -2.98939884e-01 -6.66280329e-01 4.83631164e-01
7.31366515e-01 1.97050586e-01 -5.64013235e-02 -4.22725320e-01
-3.79521221e-01 1.41753461e-02 -1.77888703e+00 4.69954014e-01
-3.80546689e-01 5.22642851e-01 -6.44552946e-01 -1.17540145e+00
1.25575018e+00 1.32810712e-01 2.85899907e-01 -3.42311144e-01
6.67458117e-01 1.69963792e-01 1.67368874e-01 -4.57127005e-01
5.10538340e-01 -9.76144522e-02 -2.62519587e-02 5.35924494e-01
5.27829342e-02 2.51197606e-01 5.81144512e-01 1.80340812e-01
8.02787781e-01 -6.92024052e-01 8.06285441e-01 -6.31069481e-01
8.67607236e-01 6.36225939e-02 1.51049808e-01 7.40262926e-01
-3.71508561e-02 2.35038936e-01 5.10823548e-01 -4.45035994e-01
-7.32080698e-01 -5.95402479e-01 -7.39317119e-01 1.18296373e+00
1.10145293e-01 -3.87148201e-01 -2.35546440e-01 -8.18758309e-01
4.13308442e-01 8.11397314e-01 -7.59057879e-01 -3.10197681e-01
-3.08366157e-02 -1.00921142e+00 2.08583012e-01 4.76198122e-02
2.66148150e-01 -6.35794342e-01 -6.69214487e-01 -9.05847475e-02
2.05128342e-01 1.71207811e-03 -2.28103638e-01 5.38515449e-01
-7.76243627e-01 -1.29648304e+00 -4.51255351e-01 -8.09423566e-01
6.96827352e-01 5.17256379e-01 6.35138631e-01 1.13001801e-01
-5.44287264e-01 2.66734255e-03 -9.79643583e-01 -6.48272336e-01
-1.67977840e-01 3.70059237e-02 3.52454424e-01 -1.33852854e-01
1.04562318e+00 -6.13424294e-02 -5.14614046e-01 3.71078998e-01
-9.41587985e-01 -6.69124126e-01 2.83378750e-01 1.09588730e+00
3.26733381e-01 7.84698367e-01 5.50132394e-01 -7.49606371e-01
1.00864410e+00 -6.13998532e-01 -3.39630961e-01 1.40927941e-01
-1.08486640e+00 1.47174641e-01 8.36563349e-01 -5.73518097e-01
-7.28830636e-01 -2.20709398e-01 2.61424720e-01 3.60162050e-01
-9.97444019e-02 5.49502552e-01 6.83680102e-02 9.60289314e-02
9.67480361e-01 3.00472647e-01 -1.39117450e-03 -3.54116261e-01
-1.58957705e-01 1.35266292e+00 -3.54045749e-01 -1.36181563e-01
4.99777943e-01 3.49743366e-01 1.51824579e-01 -8.48443687e-01
-2.86725521e-01 -8.52739096e-01 -6.64192319e-01 2.31084460e-03
3.29775482e-01 -3.26475501e-01 -9.13596034e-01 1.33190095e-01
-6.67238235e-01 8.68948877e-01 1.44051537e-01 1.01284242e+00
1.86085463e-01 2.98572063e-01 6.71382695e-02 -1.11425889e+00
-6.97099924e-01 -9.93931234e-01 6.35941386e-01 4.84904617e-01
-5.60732484e-01 -6.87418520e-01 -1.51906192e-01 -3.23365033e-01
3.08100104e-01 5.89448065e-02 1.37179887e+00 -1.35168672e+00
4.80369091e-01 -8.51526141e-01 -1.46588326e-01 2.73010582e-01
4.56349224e-01 2.61092007e-01 -4.56651658e-01 -2.90165037e-01
-5.89962862e-02 2.28622910e-02 9.05578911e-01 5.35027683e-01
9.13861692e-01 -3.22357625e-01 -7.89763927e-01 1.81093752e-01
1.52738559e+00 8.56755912e-01 2.88554728e-01 5.45969486e-01
3.81295323e-01 7.22845376e-01 9.62879837e-01 8.26746464e-01
-1.34320855e-01 3.78232449e-01 -2.68483497e-02 4.58285481e-01
3.53686005e-01 8.87386724e-02 -1.53705299e-01 4.39688295e-01
1.97893843e-01 -5.55941820e-01 -8.25486541e-01 3.35281819e-01
-1.37349117e+00 -8.74048173e-01 -4.70367491e-01 2.61987424e+00
5.91328382e-01 2.64748573e-01 3.09688509e-01 9.98305678e-01
1.01624835e+00 -4.46988761e-01 -1.46718830e-01 -6.96636438e-01
-8.52927044e-02 1.14579061e-02 4.24618959e-01 5.54257154e-01
-1.01124847e+00 4.29042876e-01 5.15451574e+00 1.03662491e+00
-1.09583068e+00 -3.21319222e-01 6.13010049e-01 8.36635381e-02
-8.55590105e-02 1.30206356e-02 -1.17118895e+00 6.02021813e-01
7.08263338e-01 -7.81812906e-01 7.33207092e-02 8.28434050e-01
-2.25722902e-02 -6.22323990e-01 -8.17974031e-01 1.03518057e+00
4.72600833e-02 -7.51430631e-01 5.68135560e-01 3.73389609e-02
4.02653784e-01 -5.49820244e-01 -2.85199434e-02 2.01664805e-01
-1.29055947e-01 -8.08889747e-01 3.43643039e-01 3.77579272e-01
4.41675782e-01 -1.27313530e+00 1.02259266e+00 2.47222945e-01
-1.11683261e+00 -2.56011009e-01 -7.25206077e-01 -1.67639226e-01
-3.53192270e-01 8.29880893e-01 -8.74431968e-01 3.08476478e-01
7.15882003e-01 2.57993370e-01 -5.49791336e-01 1.01913941e+00
6.80833220e-01 3.76819700e-01 -6.69292212e-01 -8.86945367e-01
2.19980687e-01 -4.99753170e-02 6.68831289e-01 1.15699494e+00
5.12430668e-01 2.60500927e-02 -5.53275682e-02 2.77488679e-01
4.26966161e-01 9.64370966e-01 -5.55421948e-01 1.40985856e-02
8.40194404e-01 1.17241931e+00 -1.16610014e+00 -4.28189605e-01
-2.85857946e-01 5.81000149e-01 -2.63109863e-01 -1.25178233e-01
-3.45216334e-01 -1.21911073e+00 2.36434773e-01 1.70595318e-01
-1.56046161e-02 2.55009562e-01 -3.46915454e-01 -5.88818491e-01
-2.20739916e-01 -6.19128346e-01 9.07047212e-01 -4.56866138e-02
-1.30534363e+00 9.28217590e-01 3.18716466e-01 -1.47745252e+00
-1.09624363e-01 -7.12522149e-01 -3.54017049e-01 1.10148919e+00
-7.91136146e-01 -6.61088943e-01 -3.06966454e-01 6.31912172e-01
5.12826264e-01 -5.86514831e-01 9.55296874e-01 1.44414604e-01
-2.07486808e-01 7.01945424e-01 6.23964310e-01 -1.39405444e-01
7.15978742e-01 -1.28191042e+00 -2.79003799e-01 2.49356791e-01
-2.06111237e-01 7.14595973e-01 8.17158759e-01 -7.95147121e-01
-1.03810191e+00 -7.07986116e-01 1.15800035e+00 -7.18084499e-02
2.74529725e-01 5.20580783e-02 -7.65023291e-01 1.36585951e-01
-4.83556151e-01 -2.38376006e-01 8.95619273e-01 1.03270337e-01
-1.65666431e-01 -1.89131394e-01 -1.60333431e+00 2.95172065e-01
2.55992323e-01 1.16294414e-01 -5.61532676e-01 5.04770458e-01
1.92760661e-01 2.17799857e-01 -8.30285072e-01 5.18537983e-02
8.81515384e-01 -9.28457081e-01 6.81585431e-01 -6.42190814e-01
-3.77062149e-02 -4.10007596e-01 -3.45933467e-01 -1.31483138e+00
-5.12144268e-01 -3.53078730e-02 3.86604011e-01 1.37041271e+00
4.37583864e-01 -7.77255774e-01 4.12244171e-01 1.91136435e-01
4.20101255e-01 -7.38840640e-01 -8.78201127e-01 -7.15348363e-01
1.30716592e-01 2.72064544e-02 6.19149745e-01 1.11375785e+00
4.85496640e-01 4.00267631e-01 8.87422189e-02 -3.18325728e-01
3.16537082e-01 6.69019669e-02 3.42478991e-01 -2.01563573e+00
1.64703205e-01 -5.42958617e-01 -9.85971808e-01 -2.50094205e-01
-3.07569563e-01 -1.02782071e+00 -3.32960188e-01 -1.40753412e+00
3.50766718e-01 -4.50462639e-01 -3.44638020e-01 2.01076582e-01
-2.18509495e-01 -1.59158614e-02 -1.63022414e-01 2.64665008e-01
3.35598551e-02 1.77822262e-01 8.06776285e-01 -1.52978793e-01
-5.26884913e-01 9.76611525e-02 -7.11112559e-01 7.24584877e-01
7.11498559e-01 -7.50882924e-01 -4.09850538e-01 2.64279366e-01
2.30093598e-01 -2.68771708e-01 -4.23661232e-01 -7.91533530e-01
1.24043465e-01 -2.34239921e-01 9.16611612e-01 -7.29478896e-01
-2.67239958e-01 -8.50768387e-01 -2.22602226e-02 7.27906942e-01
-4.38314348e-01 2.66877562e-01 2.54080724e-02 4.97821659e-01
-1.54077083e-01 -6.52683794e-01 7.93483615e-01 1.02756374e-01
-4.64914054e-01 -2.21717417e-01 -6.67853773e-01 -3.17660689e-01
1.27222788e+00 -5.31625986e-01 -3.26449066e-01 -1.05283879e-01
-4.81801778e-01 8.90877321e-02 1.30957022e-01 5.34666240e-01
5.81478357e-01 -1.32546270e+00 -7.33592093e-01 2.05442131e-01
4.66091573e-01 -6.58896625e-01 5.58782816e-02 6.16114914e-01
-4.91184413e-01 5.33069611e-01 -2.42969275e-01 -3.98401111e-01
-1.76259196e+00 3.28317076e-01 1.11862585e-01 -7.03671351e-02
-3.68640184e-01 9.07326281e-01 -1.68283820e-01 -1.42020702e-01
1.18874423e-01 -2.74695218e-01 -9.47727323e-01 5.65217257e-01
8.51560593e-01 8.62277567e-01 3.28266680e-01 -8.33175182e-01
-6.71686471e-01 6.92988396e-01 -5.18768013e-01 6.99938387e-02
8.34442854e-01 2.21576691e-02 -2.93287486e-01 5.50189495e-01
1.50575531e+00 -6.57257810e-02 -1.50248572e-01 3.33071984e-02
-4.49764393e-02 -7.16688812e-01 3.53011549e-01 -8.16388667e-01
-4.36910629e-01 4.44039375e-01 1.10429323e+00 7.41654694e-01
9.44200397e-01 -2.28393137e-01 4.60471034e-01 5.52057683e-01
1.14002161e-01 -1.43222344e+00 -2.70278275e-01 5.60926557e-01
7.31977403e-01 -1.23026311e+00 2.08038419e-01 -2.42805183e-01
-6.34324074e-01 1.61747777e+00 4.49599266e-01 -1.73513904e-01
1.34020758e+00 9.82162654e-02 -2.51458257e-01 -6.21763915e-02
-5.24702191e-01 -1.37383953e-01 4.49925780e-01 5.23874760e-01
8.36538851e-01 1.71847090e-01 -1.38152778e+00 6.34396553e-01
-5.20158827e-01 -2.40369320e-01 3.75838786e-01 1.07314754e+00
-1.00910985e+00 -8.70566547e-01 -6.76489472e-01 1.36016786e+00
-5.63515663e-01 5.98896742e-02 -3.63843828e-01 8.57521474e-01
2.85280019e-01 1.24535596e+00 3.84705424e-01 -7.51431823e-01
3.04992020e-01 2.50188679e-01 2.27159098e-01 -4.16111380e-01
-5.58417857e-01 -1.65103912e-01 -1.16102427e-01 1.31389096e-01
4.43465970e-02 -6.68187797e-01 -1.48755682e+00 -4.12453920e-01
-8.42595160e-01 5.53043306e-01 9.34004426e-01 9.24544394e-01
6.67195171e-02 3.43363613e-01 8.92278016e-01 -4.07184809e-01
-5.47796130e-01 -1.26277506e+00 -8.95887017e-01 4.41527396e-01
8.63277316e-02 -1.05900443e+00 -8.52842569e-01 -1.79034933e-01] | [10.544482231140137, 7.21713924407959] |
15d7ca43-aace-4e4c-8735-f7386563cc59 | achieving-strong-regularization-for-deep | null | null | https://openreview.net/forum?id=Bys_NzbC- | https://openreview.net/pdf?id=Bys_NzbC- | Achieving Strong Regularization for Deep Neural Networks | L1 and L2 regularizers are critical tools in machine learning due to their ability to simplify solutions. However, imposing strong L1 or L2 regularization with gradient descent method easily fails, and this limits the generalization ability of the underlying neural networks. To understand this phenomenon, we investigate how and why training fails for strong regularization. Specifically, we examine how gradients change over time for different regularization strengths and provide an analysis why the gradients diminish so fast. We find that there exists a tolerance level of regularization strength, where the learning completely fails if the regularization strength goes beyond it. We propose a simple but novel method, Delayed Strong Regularization, in order to moderate the tolerance level. Experiment results show that our proposed approach indeed achieves strong regularization for both L1 and L2 regularizers and improves both accuracy and sparsity on public data sets. Our source code is published. | ['Chiu Man Ho', 'Dae Hoon Park', 'Yi Chang'] | 2018-01-01 | null | null | null | iclr-2018-1 | ['l2-regularization'] | ['methodology'] | [-1.23342581e-01 -4.81405482e-02 -5.60707629e-01 -3.38793576e-01
-3.37859303e-01 -4.91283506e-01 1.26007676e-01 -6.62777126e-02
-3.78133833e-01 8.76626611e-01 2.02876478e-01 -4.11291271e-01
-8.44167322e-02 -5.21672130e-01 -7.60498226e-01 -7.75493681e-01
6.17703125e-02 -3.73081326e-01 1.01475202e-01 -1.79018766e-01
2.95715213e-01 4.32686388e-01 -1.03012538e+00 9.60628390e-02
1.03949046e+00 7.26679027e-01 -1.31100621e-02 2.16087356e-01
-2.39347875e-01 9.45497155e-01 -3.61657798e-01 -1.02970190e-01
5.10355592e-01 -3.37018996e-01 -5.00072718e-01 9.76819247e-02
3.71648550e-01 9.82197970e-02 -1.75494611e-01 1.27903593e+00
1.50705904e-01 7.14441985e-02 4.81446832e-01 -8.23717833e-01
-7.54737318e-01 5.79615533e-01 -9.33455527e-01 3.82835418e-01
-9.72690359e-02 1.46652177e-01 8.82656693e-01 -9.88799334e-01
4.17405128e-01 1.05442655e+00 1.06051612e+00 5.90674460e-01
-1.20386302e+00 -6.37924910e-01 6.04314625e-01 -4.61365521e-01
-1.34962952e+00 -2.68751025e-01 9.30456579e-01 -3.36577028e-01
7.15295732e-01 8.87698159e-02 2.54990786e-01 7.01850891e-01
9.76181701e-02 8.01590145e-01 1.08711350e+00 -4.60744351e-01
1.78475797e-01 3.38086516e-01 5.42810738e-01 8.21137607e-01
4.56189811e-01 -1.69534266e-01 -3.05241346e-01 1.48734543e-02
8.83892357e-01 7.49901263e-03 -2.28454024e-01 4.49284650e-02
-6.74380183e-01 9.64970469e-01 5.64863205e-01 5.49542844e-01
-2.60565430e-01 1.75255030e-01 3.31431955e-01 5.22233367e-01
5.59212148e-01 5.81295133e-01 -5.85933626e-01 2.11979553e-01
-8.48268986e-01 -4.31407578e-02 7.07261860e-01 4.50980961e-01
8.66467595e-01 4.11861211e-01 -5.26633635e-02 1.11289644e+00
4.55552638e-02 2.37592697e-01 5.47737479e-01 -9.10830617e-01
4.87956166e-01 7.13341892e-01 -7.58588761e-02 -1.21617544e+00
-5.13243198e-01 -7.57363260e-01 -1.06275153e+00 2.14239016e-01
6.83876097e-01 -4.91359055e-01 -7.97391236e-01 1.94330883e+00
1.08903542e-01 2.97213793e-01 -2.86467723e-03 9.75650489e-01
6.23431861e-01 5.14169872e-01 1.00065760e-01 -5.15864253e-01
7.86036849e-01 -9.24497485e-01 -7.88005948e-01 -5.63417256e-01
9.16554749e-01 -7.38165677e-01 1.46682656e+00 2.47475103e-01
-1.08498180e+00 -4.67340052e-01 -9.53714311e-01 2.14712083e-01
-1.05723657e-01 3.58014584e-01 8.31134498e-01 6.27377450e-01
-7.00181246e-01 7.95353174e-01 -7.77892172e-01 -1.52343705e-01
3.41382384e-01 2.83135921e-01 -1.71202123e-01 -9.56920069e-03
-1.10847604e+00 6.68685079e-01 2.32384294e-01 2.72796333e-01
-2.26519510e-01 -8.44318330e-01 -6.91881001e-01 2.42915973e-02
1.86399311e-01 -2.06292078e-01 8.52894604e-01 -1.20592630e+00
-1.32606828e+00 8.59889507e-01 -1.08313367e-01 -5.55700898e-01
5.65330744e-01 -5.23032904e-01 -1.94156319e-01 -2.64929324e-01
-7.01540709e-02 1.65115088e-01 1.03527188e+00 -1.22510993e+00
-3.21935773e-01 -2.87242144e-01 8.23038965e-02 1.20254178e-02
-6.65593028e-01 -1.93366498e-01 -5.24448633e-01 -6.89938068e-01
3.81941825e-01 -8.61857355e-01 -5.27296901e-01 -1.23785570e-01
-5.17702222e-01 -4.56448384e-02 7.38290608e-01 -3.94366264e-01
1.61181390e+00 -2.27143097e+00 -1.84841588e-01 4.91744459e-01
2.90613770e-01 5.54169655e-01 -1.14947163e-01 1.33880377e-01
-1.39931947e-01 3.85187835e-01 -3.60095829e-01 -6.82772696e-02
-3.63405913e-01 4.24099058e-01 -3.49044979e-01 6.42231226e-01
1.80899680e-01 6.29688978e-01 -7.66707659e-01 -6.84064925e-02
-1.59216747e-02 5.34236491e-01 -8.53678048e-01 -6.58121854e-02
2.05856822e-02 4.80822235e-01 -6.35575414e-01 3.84465188e-01
9.44250941e-01 -4.84842420e-01 3.56176823e-01 -2.35640123e-01
-3.80432099e-01 2.37733468e-01 -1.45646191e+00 1.10959160e+00
-4.99593586e-01 5.65627277e-01 2.82997578e-01 -1.40627563e+00
9.57607746e-01 3.19798663e-02 4.45168585e-01 -5.23411632e-01
8.23228359e-02 4.19320017e-01 -6.27675653e-02 -4.76406395e-01
5.39110322e-03 -1.55377403e-01 2.08923057e-01 2.48017088e-01
-3.72632176e-01 1.75284609e-01 3.87461960e-01 -2.11565625e-02
1.01296318e+00 -2.41821662e-01 2.69350678e-01 -5.61537623e-01
7.83355653e-01 -3.03504288e-01 8.65413606e-01 8.37269723e-01
-1.00952826e-01 4.83144373e-01 8.17167282e-01 -6.73452556e-01
-8.93315971e-01 -6.48880780e-01 -2.61817902e-01 1.08679616e+00
6.22075722e-02 -3.72861534e-01 -6.52503192e-01 -6.77919030e-01
2.20350280e-01 3.96466821e-01 -6.88865662e-01 -3.33310634e-01
-8.74058545e-01 -1.16958022e+00 4.51294065e-01 4.97362643e-01
7.29877830e-01 -8.12198043e-01 -1.81043103e-01 1.12541407e-01
1.60622865e-01 -9.84065413e-01 -4.94880289e-01 3.29844564e-01
-1.27053726e+00 -1.05561936e+00 -5.86668849e-01 -6.93836451e-01
1.01050222e+00 2.65577227e-01 1.03977835e+00 6.10451102e-01
-1.38999289e-02 1.01123964e-02 -1.01178378e-01 3.25933546e-02
-2.36754820e-01 2.13164255e-01 3.05684954e-01 -1.38873175e-01
2.28745475e-01 -8.68508518e-01 -6.24658525e-01 3.02265912e-01
-8.38952422e-01 -3.40031713e-01 6.02229834e-01 8.48500192e-01
4.29846406e-01 -4.08385918e-02 6.81941748e-01 -1.37583065e+00
1.00407267e+00 -4.90672290e-01 -7.31572390e-01 5.59248105e-02
-8.55505168e-01 3.37848783e-01 8.60454500e-01 -5.97342193e-01
-8.19178760e-01 1.45373538e-01 -1.63299829e-01 -4.32769477e-01
2.07794234e-01 8.72916281e-01 3.08298945e-01 -4.20492321e-01
1.00779617e+00 -6.78016106e-03 -1.98644489e-01 -5.95022678e-01
2.52282470e-01 2.13444069e-01 1.89539507e-01 -6.89260542e-01
8.83945167e-01 3.83354902e-01 -5.78687973e-02 -9.69821274e-01
-1.33212829e+00 -3.72221112e-01 -4.62791353e-01 9.84552354e-02
3.06663603e-01 -8.01559627e-01 -5.56301892e-01 1.39837399e-01
-9.91445780e-01 -3.76352102e-01 -3.19610238e-01 5.70274532e-01
-9.79452059e-02 2.55570114e-01 -6.49241388e-01 -6.89863920e-01
-1.89688846e-01 -8.54247332e-01 2.13349193e-01 2.87815541e-01
-5.85221276e-02 -1.29442704e+00 7.67454281e-02 -6.80612549e-02
5.50337076e-01 9.62819979e-02 7.32954443e-01 -3.99592608e-01
-2.25916266e-01 -1.10266387e-01 -3.46664727e-01 7.97862113e-01
2.40822896e-01 2.99189072e-02 -7.49159515e-01 -3.62709999e-01
2.54038453e-01 -3.08507204e-01 1.14863300e+00 7.08954394e-01
1.24303985e+00 -5.01242161e-01 -1.43449351e-01 9.50833559e-01
1.63246143e+00 -2.70153165e-01 6.34653628e-01 2.42669508e-01
8.02457631e-01 2.19205439e-01 4.96288985e-02 4.18247789e-01
-2.39990592e-01 2.63535827e-01 9.76959541e-02 -4.60270852e-01
-1.03967153e-01 -1.94088067e-03 4.25132185e-01 1.07894588e+00
-2.52030224e-01 2.24278197e-01 -7.91363239e-01 4.78090882e-01
-1.85678482e+00 -6.51704907e-01 -2.74610281e-01 2.27173543e+00
1.07337165e+00 4.17122990e-01 -9.11865383e-02 1.97706267e-01
4.35525388e-01 2.63691396e-01 -7.15350807e-01 -4.49634522e-01
-1.94563985e-01 5.62135614e-02 7.54063904e-01 7.50258148e-01
-1.00697219e+00 1.07664800e+00 7.28834248e+00 7.32288003e-01
-1.50140345e+00 6.33797422e-03 5.77576518e-01 -6.90588877e-02
-2.27612481e-01 -9.81068537e-02 -8.89007330e-01 2.97935575e-01
5.74489534e-01 -8.47418755e-02 3.18399757e-01 8.96033883e-01
4.11780775e-01 9.60031077e-02 -8.13062131e-01 8.32747340e-01
-2.05808356e-01 -1.40234554e+00 -6.89989328e-02 -3.09179842e-01
1.00419283e+00 1.10013939e-01 2.43207783e-01 4.34396863e-01
2.47481510e-01 -9.56062376e-01 9.56206918e-02 2.39603788e-01
5.39984941e-01 -6.65396333e-01 4.67383891e-01 4.31026727e-01
-1.00585556e+00 -2.32275322e-01 -6.40358686e-01 -4.36237365e-01
-1.56681061e-01 1.07821834e+00 -3.80353898e-01 -5.59303910e-02
4.54259485e-01 8.16395223e-01 -5.24009645e-01 8.35308909e-01
-5.25767446e-01 8.98338377e-01 -4.59953129e-01 2.16474965e-01
3.41863304e-01 -3.76262456e-01 4.74886924e-01 1.35675418e+00
1.92539781e-01 2.05291137e-01 4.09770519e-01 8.37755203e-01
-3.34705561e-01 2.68294185e-01 -6.26853049e-01 4.75908257e-02
3.64852577e-01 1.07770908e+00 -7.19947815e-01 -1.48975894e-01
-5.46765268e-01 6.02434456e-01 6.40198112e-01 7.63386369e-01
-7.19084799e-01 -1.82605818e-01 6.34971619e-01 2.26089463e-01
1.07815415e-01 -3.98257375e-01 -7.62542188e-01 -1.38394713e+00
1.59176052e-01 -8.30405235e-01 3.87679100e-01 -9.40324664e-02
-1.30543482e+00 4.36143935e-01 -4.19018179e-01 -9.93808508e-01
1.50984362e-01 -6.09078526e-01 -6.48741782e-01 6.98748767e-01
-1.75652468e+00 -6.71679854e-01 -1.11583523e-01 5.97259998e-01
4.56500828e-01 -1.62213758e-01 4.30084497e-01 5.58746934e-01
-8.84725451e-01 7.97146678e-01 2.94318974e-01 1.57882527e-01
5.29968262e-01 -8.43057930e-01 -5.23996353e-02 8.57936919e-01
-1.94729716e-02 9.91820335e-01 8.13036084e-01 -5.27852178e-01
-1.24156940e+00 -1.02370036e+00 7.21799433e-01 -5.46310470e-02
6.87512636e-01 -5.48894592e-02 -1.36233985e+00 6.17268920e-01
-8.95435363e-02 4.22764212e-01 5.42103112e-01 4.20145065e-01
-4.48088825e-01 -1.24313235e-01 -1.10815191e+00 5.22958338e-01
7.95469105e-01 -3.79906625e-01 -2.92615354e-01 5.59054554e-01
5.94694972e-01 -2.47730196e-01 -5.98957300e-01 5.85713089e-01
3.61322612e-01 -9.89057899e-01 1.01297748e+00 -6.72351539e-01
5.30733407e-01 -1.81627259e-01 -2.17876166e-01 -1.08809161e+00
-3.20590913e-01 -6.81831419e-01 -1.63226649e-01 9.40519989e-01
6.66555643e-01 -7.95070767e-01 9.08339381e-01 6.61944985e-01
1.06669232e-01 -9.65211630e-01 -6.19828761e-01 -1.05157590e+00
4.09930706e-01 -4.31914508e-01 -4.91793174e-03 1.26702535e+00
-7.51070008e-02 2.20738307e-01 -5.16182661e-01 1.91611677e-01
5.23155928e-01 4.24483567e-02 4.61673379e-01 -1.15369463e+00
-1.22743450e-01 -5.29751062e-01 -1.15469005e-02 -1.26451612e+00
2.34192923e-01 -9.30299282e-01 -1.92568019e-01 -1.15566313e+00
-5.71989343e-02 -7.59420216e-01 -6.06517017e-01 5.92562973e-01
-4.00891304e-01 4.25231934e-01 2.91494839e-02 5.38071573e-01
-3.25590700e-01 3.86318177e-01 1.20910335e+00 5.92382252e-02
-6.93149447e-01 5.58426753e-02 -8.59157860e-01 1.06715429e+00
9.46946740e-01 -5.65963686e-01 -3.53779972e-01 -5.50138116e-01
5.64801335e-01 -4.75917697e-01 -7.45445043e-02 -9.84150589e-01
1.78030357e-02 -3.26672435e-01 3.09272915e-01 -2.88776285e-03
-9.41785648e-02 -6.02859139e-01 -3.64422679e-01 5.44212043e-01
-5.66782892e-01 -2.63615072e-01 3.41455996e-01 3.21977466e-01
-3.07810217e-01 -3.71819586e-01 1.12509763e+00 -1.31267264e-01
-4.05585289e-01 2.05130547e-01 -2.81817704e-01 3.49469513e-01
6.56187832e-01 1.03640303e-01 1.66997258e-02 -2.84948230e-01
-7.23017573e-01 2.38834605e-01 2.87019283e-01 -4.20783740e-03
4.91567940e-01 -1.36918652e+00 -7.14406967e-01 4.08417106e-01
-3.17739457e-01 -2.79916763e-01 -9.89090800e-02 1.01952708e+00
-3.26061428e-01 1.33228227e-01 -3.33263315e-02 -4.61932302e-01
-1.03174007e+00 5.17331064e-01 4.65782195e-01 -3.62619102e-01
-7.16618180e-01 1.04444146e+00 2.42695123e-01 -4.68752354e-01
3.00192088e-01 -4.95081991e-01 -2.58817255e-01 -8.62767771e-02
4.76850301e-01 2.36817420e-01 -7.06831664e-02 -3.23506117e-01
-3.54744107e-01 8.93153250e-01 -2.05369249e-01 4.58780438e-01
1.41851497e+00 -9.10208002e-02 -1.30995363e-01 4.55838174e-01
1.35566795e+00 2.79340118e-01 -1.28385973e+00 -3.41638058e-01
1.03582643e-01 -4.85705048e-01 1.66039273e-01 -2.30045214e-01
-1.40731227e+00 8.57115626e-01 3.47971648e-01 4.26914424e-01
1.09362459e+00 -3.05044293e-01 7.20058382e-01 6.05932891e-01
-2.16063693e-01 -1.28723359e+00 1.34289250e-01 9.91943181e-01
8.07653844e-01 -1.37146628e+00 3.08491200e-01 -5.64439595e-01
-4.47025955e-01 1.08513796e+00 6.45694971e-01 -6.33904636e-01
7.78260410e-01 3.82355660e-01 2.48378143e-01 -2.77326703e-01
-5.62500894e-01 2.88468748e-02 2.18709946e-01 1.49036407e-01
8.96919668e-01 -2.65351951e-01 -6.91199243e-01 4.97998297e-01
-1.58582851e-02 4.86835614e-02 3.20147276e-01 8.10406029e-01
-5.38990021e-01 -1.08396816e+00 -2.73579746e-01 3.79105061e-01
-6.66944504e-01 -1.50788218e-01 -1.66213438e-01 7.73790896e-01
2.77660545e-02 7.21330702e-01 -2.55532920e-01 -1.60687655e-01
3.47607672e-01 -4.31977548e-02 3.49643797e-01 -4.69237447e-01
-5.74465156e-01 2.55438536e-02 -7.32445195e-02 -5.37273705e-01
-3.98230880e-01 -2.12231785e-01 -1.47030258e+00 -2.58315176e-01
-3.42993647e-01 1.44846007e-01 4.32254225e-01 1.04227293e+00
1.42136738e-01 5.29381990e-01 7.72036314e-01 -4.95689750e-01
-6.80039048e-01 -8.13409746e-01 -5.50548613e-01 5.84644377e-01
6.14924312e-01 -5.19156635e-01 -8.09819043e-01 4.66273837e-02] | [8.434615135192871, 3.607640266418457] |
5fb4009e-7b1d-4e4e-a010-3eafb6fa3d81 | a-markovian-formalism-for-active-querying | 2306.08001 | null | https://arxiv.org/abs/2306.08001v1 | https://arxiv.org/pdf/2306.08001v1.pdf | A Markovian Formalism for Active Querying | Active learning algorithms have been an integral part of recent advances in artificial intelligence. However, the research in the field is widely varying and lacks an overall organizing leans. We outline a Markovian formalism for the field of active learning and survey the literature to demonstrate the organizing capability of our proposed formalism. Our formalism takes a partially observable Markovian system approach to the active learning process as a whole. We specifically outline how querying, dataset augmentation, reward updates, and other aspects of active learning can be viewed as a transition between meta-states in a Markovian system, and give direction into how other aspects of active learning can fit into our formalism. | ['Sid Ijju'] | 2023-06-13 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 4.90438491e-01 7.24023700e-01 -5.04065990e-01 -6.11519694e-01
-6.73905015e-01 -6.26917064e-01 1.00591207e+00 4.88224655e-01
-7.47467995e-01 8.73480797e-01 -2.76766773e-02 -9.18573588e-02
-5.89936256e-01 -9.07601357e-01 -5.58497071e-01 -9.28780437e-01
-1.15151703e-01 7.20774591e-01 5.57379723e-01 1.69725910e-01
4.23416704e-01 6.56604350e-01 -1.33406162e+00 9.02309921e-03
7.25277603e-01 6.42032683e-01 -2.21219316e-01 6.93127930e-01
-7.11284697e-01 1.50787222e+00 -5.66224933e-01 -1.62507534e-01
1.68364346e-01 -3.08904082e-01 -1.50354898e+00 4.00037795e-01
4.25809771e-02 2.78229378e-02 -1.54535130e-01 6.92945719e-01
1.58000559e-01 5.14616072e-01 5.36467791e-01 -1.53437781e+00
-4.06114519e-01 5.88418007e-01 2.84569830e-01 1.32766053e-01
3.30556005e-01 1.72346056e-01 8.74009013e-01 -6.98816597e-01
7.22977281e-01 9.91288781e-01 4.04158756e-02 6.00760221e-01
-1.22170269e+00 -1.28317490e-01 5.74024320e-01 5.53984821e-01
-7.95927107e-01 -7.49679446e-01 8.25332701e-01 -3.48828822e-01
8.67153823e-01 4.21289444e-01 1.29926503e+00 6.85844481e-01
3.05456426e-02 1.68244231e+00 9.43264842e-01 -1.09156251e+00
8.34848702e-01 3.28477889e-01 7.80209899e-01 8.09234381e-01
2.27770030e-01 2.61820763e-01 -8.83508086e-01 -2.55391836e-01
7.04089820e-01 3.55223954e-01 3.88914585e-01 -1.28397167e+00
-8.38947773e-01 9.45627511e-01 2.90605221e-02 1.04040578e-01
-3.17732304e-01 2.23401546e-01 5.09310812e-02 4.98090297e-01
3.27128559e-01 5.80181658e-01 -3.17756236e-01 -2.63909787e-01
-6.20667100e-01 -3.70684899e-02 1.18232286e+00 7.09372580e-01
1.08108401e+00 8.45731720e-02 -9.86959040e-03 4.51695591e-01
8.23060632e-01 1.08672656e-01 1.23346090e-01 -1.29284477e+00
-2.81062815e-02 1.03705370e+00 2.62623489e-01 -1.16427712e-01
-6.55953959e-02 -1.19708136e-01 -6.69404417e-02 5.74647307e-01
7.34773949e-02 -8.57989863e-02 -6.70718372e-01 1.22602749e+00
2.50475198e-01 -1.06925584e-01 2.40775660e-01 2.45490849e-01
6.50804698e-01 6.36654317e-01 3.30468506e-01 -9.87984896e-01
4.47188228e-01 -1.24701560e+00 -8.23055208e-01 -2.29562491e-01
4.58314419e-01 -3.83324176e-01 6.73394799e-01 3.98948491e-01
-1.58350217e+00 -2.95934379e-01 -8.70650768e-01 1.62994683e-01
-5.07827401e-01 -6.20864868e-01 1.11579895e+00 7.24137545e-01
-1.13380301e+00 3.02700132e-01 -1.41308188e+00 -4.65866178e-01
4.34528232e-01 5.13283253e-01 -6.14848323e-02 3.74444097e-01
-9.51344728e-01 1.08386981e+00 7.79571474e-01 -4.94672842e-02
-1.02907038e+00 -1.64152846e-01 -8.20089340e-01 -2.25056723e-01
8.73244822e-01 -5.95657706e-01 1.67745483e+00 -1.21705246e+00
-1.64452469e+00 6.66485727e-01 -8.75772014e-02 -5.98145247e-01
3.59569967e-01 -4.18533623e-01 -4.49805915e-01 6.80115074e-02
-5.30975342e-01 6.80800676e-01 6.68561518e-01 -1.27076483e+00
-8.04825306e-01 -2.59450406e-01 5.18138349e-01 7.36640334e-01
-4.01726216e-01 1.09679019e-02 -1.41452312e-01 -2.77749926e-01
1.22352906e-01 -8.07590008e-01 -6.57300591e-01 2.88182139e-01
1.72870114e-01 -5.08985519e-01 9.43739653e-01 5.43784022e-01
1.49997556e+00 -2.03284502e+00 6.87190145e-02 1.78750485e-01
2.84724712e-01 5.33265918e-02 1.57127857e-01 9.67511952e-01
1.03502713e-01 -1.06018141e-01 -4.01774585e-01 -1.89306378e-01
7.93187544e-02 5.89550912e-01 -2.99455434e-01 3.24513942e-01
1.32817835e-01 8.70245337e-01 -1.14146769e+00 -8.56657028e-01
3.42852265e-01 3.05040777e-02 -2.71117032e-01 3.98821503e-01
-3.23322356e-01 1.76263988e-01 -7.21649289e-01 5.78557789e-01
6.47385314e-04 -2.73922503e-01 2.44028270e-01 5.88879168e-01
-4.76826370e-01 3.36531013e-01 -1.47003353e+00 1.81427038e+00
-1.12411879e-01 5.13425827e-01 -2.26027966e-01 -1.25774980e+00
7.28911161e-01 4.88816679e-01 8.97533655e-01 -3.59553665e-01
-4.08438593e-01 -6.37760460e-02 1.60889640e-01 -5.91462672e-01
6.54712498e-01 4.64206003e-02 9.29651707e-02 1.10617161e+00
3.32786560e-01 -2.24438915e-03 4.07820731e-01 4.07707304e-01
1.16529953e+00 1.79117784e-01 8.86449993e-01 -2.56511569e-02
5.11310101e-01 1.96300268e-01 3.45094293e-01 1.21438920e+00
-5.13074756e-01 -3.50693800e-02 1.05035357e-01 -5.68828523e-01
-5.78435004e-01 -1.19148374e+00 -9.33038741e-02 1.28223670e+00
-1.23384222e-02 -4.73302633e-01 -4.24570501e-01 -9.33307588e-01
-4.33278471e-01 5.53250492e-01 -4.20397550e-01 -1.31913528e-01
-4.58232105e-01 -8.58234406e-01 2.63460547e-01 3.90896350e-01
5.38244605e-01 -1.82760072e+00 -1.21868813e+00 3.83265615e-01
2.67081082e-01 -3.18680525e-01 3.37326705e-01 8.15022409e-01
-1.41739488e+00 -1.05122864e+00 -2.74020344e-01 -6.72481656e-01
6.17189944e-01 -1.22961678e-01 1.28718960e+00 -1.00914769e-01
-8.52889270e-02 1.18703449e+00 -4.32192892e-01 -8.06656659e-01
-4.82500017e-01 1.74791604e-01 -2.67606169e-01 -2.66497899e-02
5.11416614e-01 -2.42830485e-01 -3.79276991e-01 -1.39310539e-01
-1.16829598e+00 -1.37994036e-01 6.50466442e-01 5.25207162e-01
5.81996918e-01 -3.27929622e-03 2.06278831e-01 -1.47073138e+00
4.05597925e-01 -3.66077721e-01 -6.39430344e-01 4.80379790e-01
-1.10730636e+00 1.75156131e-01 -2.10158542e-01 -3.51116508e-01
-1.48050773e+00 6.40500844e-01 1.25595421e-01 2.01408386e-01
-3.79573613e-01 4.49830860e-01 -2.68789083e-01 -1.78928629e-01
6.08955860e-01 2.98113316e-01 -1.40213802e-01 -7.16612414e-02
4.92537051e-01 4.52497393e-01 -1.13558725e-01 -5.13872027e-01
5.84059358e-01 5.60901225e-01 -9.75968763e-02 -6.66469038e-01
-8.14778268e-01 -7.33064234e-01 -9.17739034e-01 -4.43130970e-01
3.79368365e-01 -4.15006310e-01 -6.71762109e-01 4.66071844e-01
-9.17553365e-01 -4.83936280e-01 -1.30658960e+00 4.17768776e-01
-1.09725630e+00 1.61759809e-01 -3.66862178e-01 -1.48690224e+00
-6.36072308e-02 -1.02841115e+00 4.26765978e-01 4.62738812e-01
-3.53320539e-01 -1.53431749e+00 8.42489541e-01 3.49789172e-01
3.28259110e-01 -3.42778236e-01 6.90385461e-01 -1.13002181e+00
-1.03397286e+00 -1.19907983e-01 6.62044704e-01 2.07311232e-02
4.21427451e-02 2.17666864e-01 -1.09421623e+00 -1.27854362e-01
1.50783181e-01 -4.37837541e-01 6.63772345e-01 1.84424981e-01
7.13801324e-01 -4.88546848e-01 -2.85963565e-01 -1.33526802e-01
1.36168051e+00 8.73530984e-01 7.51937270e-01 4.70354646e-01
7.61875585e-02 6.67074442e-01 8.75589788e-01 3.90540570e-01
9.11821052e-02 3.59512120e-01 4.01554018e-01 -2.57681847e-01
3.95696729e-01 1.68217182e-01 6.09274924e-01 1.01700366e+00
-1.06699944e-01 -1.15752272e-01 -1.15958488e+00 3.17870229e-01
-2.21172237e+00 -1.35471034e+00 9.53369867e-03 2.22191143e+00
1.05690515e+00 3.85112137e-01 1.42032102e-01 5.14588118e-01
4.25604701e-01 3.77153426e-01 -7.10124433e-01 -4.72357750e-01
2.22544670e-02 3.20015490e-01 7.26018101e-02 6.18917704e-01
-1.25520706e+00 8.25757861e-01 8.31125927e+00 5.08403480e-01
-6.66785002e-01 1.65446177e-01 6.31942451e-02 -8.56895838e-03
-9.71856788e-02 6.24360800e-01 -6.60918236e-01 -7.40556642e-02
8.42336535e-01 -6.10840246e-02 1.21305525e-01 8.69523406e-01
5.54606505e-02 -3.65671098e-01 -1.44339883e+00 5.77039540e-01
-4.99100722e-02 -1.33897626e+00 1.24719471e-01 1.06278241e-01
7.86634564e-01 -1.82337075e-01 -1.91672832e-01 2.65702277e-01
8.46219838e-01 -5.68009794e-01 7.78484285e-01 1.01427054e+00
1.06818043e-01 -5.12599409e-01 2.47999802e-01 7.66210020e-01
-8.97519410e-01 -3.54745179e-01 -3.60639170e-02 -4.29774642e-01
3.53531763e-02 2.31127515e-01 -7.30617642e-01 4.11265969e-01
3.80207837e-01 7.35187709e-01 -6.14124775e-01 1.39476252e+00
-8.19378793e-02 8.52404177e-01 -1.02297753e-01 -1.65316328e-01
6.65088072e-02 -3.67296398e-01 6.03290379e-01 8.84491563e-01
-5.26315689e-01 1.30348414e-01 3.70256364e-01 5.24859846e-01
1.13351308e-01 3.43780182e-02 -5.67502558e-01 -3.19892526e-01
5.22804737e-01 1.14900815e+00 -9.35751200e-01 -6.31679714e-01
-4.59751546e-01 6.28136754e-01 3.36857766e-01 3.30660254e-01
-2.93675810e-01 -2.11045802e-01 -5.53357117e-02 -1.72815114e-01
-2.62744486e-01 -9.59190130e-02 -1.84601650e-01 -1.08811915e+00
-3.15965146e-01 -8.12915444e-01 7.73162961e-01 -6.12732172e-01
-1.00981534e+00 7.46637359e-02 2.18575716e-01 -1.13700867e+00
-4.03718531e-01 -7.32021257e-02 -5.22314906e-01 2.57733375e-01
-1.06035089e+00 -7.33871341e-01 -1.63052455e-01 7.18051374e-01
9.27236855e-01 -3.86571795e-01 1.27458966e+00 -1.55162424e-01
-5.57613313e-01 6.49945289e-02 6.60837293e-02 -3.49608203e-03
3.50700110e-01 -1.42352545e+00 2.04774216e-01 9.37309921e-01
7.99586117e-01 8.61123502e-01 4.32547867e-01 -5.73696733e-01
-1.26086164e+00 -7.36389518e-01 7.45269954e-01 -7.17083752e-01
6.06251299e-01 -1.51920855e-01 -8.00654113e-01 1.08188915e+00
4.53013599e-01 -2.69094110e-02 9.23693478e-01 2.65046984e-01
1.18231907e-01 -9.71024781e-02 -9.45846260e-01 3.15785885e-01
1.02425969e+00 -5.44087827e-01 -9.76409197e-01 6.07744679e-02
3.85932446e-01 -1.12183236e-01 -6.93091989e-01 3.28309417e-01
3.69149894e-01 -8.91362488e-01 7.94353366e-01 -9.55545187e-01
-2.57831544e-01 -1.13185257e-01 1.45426676e-01 -8.10437977e-01
-3.54369998e-01 -7.23339140e-01 -7.53935456e-01 1.10990381e+00
4.08409655e-01 -6.32583857e-01 1.16077316e+00 7.18655527e-01
-1.14923209e-01 -7.18455255e-01 -7.09432781e-01 -5.23536086e-01
-6.37056902e-02 -5.34532309e-01 1.44648328e-01 7.49742925e-01
2.24764362e-01 3.25410575e-01 1.24734454e-01 -4.29466188e-01
7.74971187e-01 -6.33928850e-02 4.51972127e-01 -1.80665183e+00
-2.02457726e-01 -1.69163913e-01 -2.36139566e-01 -1.00277519e+00
1.30134253e-02 -6.98720276e-01 1.02008462e-01 -1.75888395e+00
5.61422408e-01 -7.48671949e-01 -5.70086837e-01 7.32709825e-01
1.42625064e-01 -4.50336747e-02 3.15664381e-01 7.24032104e-01
-1.33487785e+00 4.80772376e-01 8.23049724e-01 -1.16357312e-01
-4.19808507e-01 4.39889699e-01 -2.61731923e-01 1.02361012e+00
5.68841934e-01 -6.84777498e-01 -8.62528026e-01 -1.31427884e-01
6.65420771e-01 1.42745093e-01 1.38099402e-01 -8.35927904e-01
7.83709407e-01 -6.64879560e-01 2.02192858e-01 -5.81086636e-01
4.86651838e-01 -1.15875268e+00 2.65282422e-01 7.13463604e-01
-1.04092622e+00 -1.18038543e-01 -3.53342354e-01 6.01702332e-01
-3.41595232e-01 -6.78754807e-01 6.85013354e-01 -4.39989150e-01
-9.20660675e-01 3.20176274e-01 -9.31811392e-01 2.12130070e-01
1.20785058e+00 -5.32501161e-01 3.35917994e-02 -5.12471981e-02
-1.14011407e+00 -6.93947151e-02 4.99310076e-01 1.68756962e-01
4.66075897e-01 -9.65271175e-01 -6.82791770e-02 3.03411651e-02
1.40451819e-01 8.65221322e-02 -2.92273760e-01 6.57886565e-01
-2.50500143e-01 5.62018156e-01 -2.68129766e-01 -5.43910801e-01
-1.16190720e+00 4.17906880e-01 3.64029855e-01 -4.80319619e-01
-3.43978703e-01 4.43164229e-01 -1.67510286e-01 -3.85968298e-01
7.89778173e-01 2.53368825e-01 -5.93542159e-01 3.36016715e-01
1.21322557e-01 6.35449052e-01 6.04282059e-02 8.36263523e-02
-3.36679190e-01 3.77095491e-03 -1.42074510e-01 -5.57861447e-01
1.18395829e+00 -2.91850328e-01 -5.70180342e-02 1.45536757e+00
4.06299442e-01 -2.84909993e-01 -1.13049030e+00 -4.39864933e-01
5.00835121e-01 -2.01578978e-02 -2.27305889e-02 -7.91423678e-01
-7.10498214e-01 8.03936541e-01 8.41777146e-01 7.29615688e-01
8.01334321e-01 1.55866981e-01 -4.57617678e-02 9.44478691e-01
5.56819081e-01 -1.44800127e+00 3.10776234e-01 5.00517309e-01
4.64479327e-01 -1.12233698e+00 1.96685016e-01 -1.59104183e-01
-5.76175988e-01 1.15777028e+00 5.18150926e-01 1.79224520e-03
7.74374068e-01 3.82016182e-01 2.35088423e-01 -2.71659106e-01
-1.21782625e+00 -3.76076251e-01 1.05805442e-01 5.94357908e-01
4.36884612e-01 -3.32337201e-01 -2.43122309e-01 -6.39392287e-02
3.94450903e-01 2.59403408e-01 6.04927659e-01 1.81561565e+00
-8.27505171e-01 -1.38049531e+00 -3.02796662e-01 2.92142242e-01
-1.65346533e-01 1.83663204e-01 -5.97899795e-01 6.15879476e-01
7.94864669e-02 8.83788526e-01 3.00133198e-01 3.68908644e-01
1.41676068e-01 3.83430123e-01 8.22873175e-01 -1.14857149e+00
-7.22563982e-01 -2.91350037e-01 -9.15840119e-02 -4.69610006e-01
-1.13766122e+00 -8.57721925e-01 -1.35775101e+00 1.43097684e-01
-4.22426969e-01 5.45817852e-01 4.43351418e-01 1.19429421e+00
-1.04326949e-01 2.32157111e-01 5.68639219e-01 -1.36213452e-01
-4.63022858e-01 -7.58022487e-01 -5.01822114e-01 -8.69950354e-02
2.21206620e-01 -5.23647845e-01 -3.40572417e-01 3.17030251e-01] | [9.508034706115723, 4.260340213775635] |
489a0ed4-1f59-4e59-9bf7-0bcb8be4aaa7 | privacy-preserving-collaborative-chinese-text | 2305.05602 | null | https://arxiv.org/abs/2305.05602v1 | https://arxiv.org/pdf/2305.05602v1.pdf | Privacy-Preserving Collaborative Chinese Text Recognition with Federated Learning | In Chinese text recognition, to compensate for the insufficient local data and improve the performance of local few-shot character recognition, it is often necessary for one organization to collect a large amount of data from similar organizations. However, due to the natural presence of private information in text data, different organizations are unwilling to share private data, such as addresses and phone numbers. Therefore, it becomes increasingly important to design a privacy-preserving collaborative training framework for the Chinese text recognition task. In this paper, we introduce personalized federated learning (pFL) into the Chinese text recognition task and propose the pFedCR algorithm, which significantly improves the model performance of each client (organization) without sharing private data. Specifically, based on CRNN, to handle the non-iid problem of client data, we add several attention layers to the model and design a two-stage training approach for the client. In addition, we fine-tune the output layer of the model using a virtual dataset on the server, mitigating the problem of character imbalance in Chinese documents. The proposed approach is validated on public benchmarks and two self-built real-world industrial scenario datasets. The experimental results show that the pFedCR algorithm can improve the performance of local personalized models while also improving their generalization performance on other client data domains. Compared to local training within an organization, pFedCR improves model performance by about 20%. Compared to other state-of-the-art personalized federated learning methods, pFedCR improves performance by 6%~8%. Moreover, through federated learning, pFedCR can correct erroneous information in the ground truth. | ['xiangyang xue', 'Bin Li', 'Haiyang Yu', 'Shangchao Su'] | 2023-05-09 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [ 9.99824107e-02 -4.86243308e-01 -2.95101911e-01 -7.07601011e-01
-8.90933156e-01 -3.74580741e-01 1.53078958e-01 -9.85770822e-02
-4.03748572e-01 6.65629387e-01 8.55170712e-02 -3.20264071e-01
3.53428870e-02 -7.66356468e-01 -6.14682555e-01 -9.34840977e-01
6.44521952e-01 3.60942781e-01 7.87559245e-03 1.86291456e-01
2.59861469e-01 3.63367438e-01 -1.16162598e+00 7.33996391e-01
1.13340259e+00 1.21844268e+00 7.59758055e-02 4.18247074e-01
-3.69111538e-01 9.24288690e-01 -7.22905874e-01 -8.69031489e-01
2.38033429e-01 -1.19863056e-01 -6.87080860e-01 -6.23046607e-03
2.59584278e-01 -6.55942082e-01 -4.47306514e-01 1.10441065e+00
5.73037505e-01 4.99921888e-02 1.93256214e-01 -1.14504409e+00
-9.19244885e-01 9.12374794e-01 -4.44375724e-01 -1.57831520e-01
-1.29362509e-01 -3.97258699e-02 8.58103871e-01 -8.04750502e-01
3.60635996e-01 7.05196917e-01 5.92040300e-01 7.51184702e-01
-7.27486789e-01 -8.37786674e-01 4.17513400e-01 2.46385500e-01
-1.24431467e+00 -2.99248874e-01 6.78464770e-01 -6.34577572e-02
6.24105394e-01 5.24982154e-01 -6.19907603e-02 1.14509356e+00
2.03330815e-02 1.30355692e+00 8.14587772e-01 -2.69649148e-01
2.30399296e-01 4.44484025e-01 3.07712346e-01 2.34384254e-01
3.24176699e-01 -4.02639717e-01 -5.19573689e-01 -3.21850806e-01
1.72036201e-01 6.95838094e-01 -2.17712149e-01 -2.22631782e-01
-9.72031951e-01 3.77182484e-01 1.79292932e-01 3.09795737e-01
-1.26357064e-01 -1.03348196e-01 6.43807948e-01 1.16839804e-01
5.07557511e-01 -3.77964750e-02 -8.72270525e-01 -1.67346910e-01
-7.54162550e-01 3.12645771e-02 8.09429467e-01 9.83954072e-01
7.53461301e-01 -1.16778769e-01 -2.86686212e-01 8.47205281e-01
-1.57688428e-02 4.68692422e-01 9.31771636e-01 -3.89044613e-01
1.07814884e+00 8.77735496e-01 6.20353818e-02 -1.06091583e+00
9.43831354e-02 -5.80976009e-01 -9.77749169e-01 -4.98925567e-01
3.68199736e-01 -4.31609452e-01 -5.28179884e-01 1.18205583e+00
2.16153771e-01 2.45070249e-01 3.09871823e-01 9.85814929e-01
2.88590521e-01 6.84191048e-01 -1.58823803e-01 -3.59969065e-02
1.00072837e+00 -1.16017938e+00 -7.06840158e-01 -7.62210563e-02
9.56164777e-01 -6.15650594e-01 1.03595150e+00 3.84113073e-01
-2.57386267e-01 -1.30807757e-01 -7.98442185e-01 -1.74374674e-02
-4.54858363e-01 5.74299812e-01 5.66722751e-01 1.10182226e+00
-2.46490479e-01 5.73811412e-01 -6.19228363e-01 -9.64713618e-02
7.32951105e-01 2.87738323e-01 -4.47179615e-01 -6.81191564e-01
-9.67017412e-01 2.09130704e-01 2.94246137e-01 2.17979193e-01
-6.19601488e-01 -7.39673436e-01 -3.04354787e-01 3.53735447e-01
5.35317659e-01 -8.07074234e-02 1.08380914e+00 -9.57612395e-01
-1.46667588e+00 4.17215198e-01 5.98106608e-02 -3.78030121e-01
8.45378995e-01 -3.49857062e-01 -6.08178556e-01 -2.30197817e-01
-2.82978296e-01 -2.30136037e-01 6.16955936e-01 -1.12431026e+00
-8.03193986e-01 -6.70821428e-01 -5.52610517e-01 3.50344814e-02
-9.74910617e-01 4.43017818e-02 -7.24217117e-01 -5.90676785e-01
-2.11884797e-01 -5.74690461e-01 -1.41935185e-01 -2.67846674e-01
-5.12957036e-01 -7.69517049e-02 1.39647114e+00 -9.63405609e-01
1.17664158e+00 -2.32835388e+00 -3.69214028e-01 1.64391905e-01
-1.30288312e-02 5.26043355e-01 -1.42079145e-01 1.97241813e-01
1.81960866e-01 2.10408092e-01 -2.46343985e-01 -5.96956074e-01
1.45696715e-01 2.76122034e-01 -4.50285196e-01 5.05872428e-01
-1.67590782e-01 8.50741148e-01 -4.92289007e-01 -3.06654036e-01
-5.31127229e-02 2.04245970e-01 -4.23855007e-01 1.65992141e-01
-1.34973183e-01 1.51465749e-02 -8.20468247e-01 7.82413781e-01
1.04865289e+00 -4.21604395e-01 5.18245399e-01 4.22737002e-02
4.91331741e-02 -6.98966207e-03 -1.33097994e+00 1.38087440e+00
-2.14507729e-01 2.78202504e-01 -2.43440047e-02 -1.08644676e+00
1.25164723e+00 3.43953907e-01 4.45734113e-01 -7.03302860e-01
3.01417321e-01 3.64700049e-01 -4.33434367e-01 -2.90771484e-01
6.32095754e-01 3.89596373e-01 -2.61139661e-01 7.83916116e-01
-2.39899144e-01 7.16418207e-01 -4.77786005e-01 1.77639157e-01
1.06788671e+00 -2.01543108e-01 -1.78238466e-01 1.79353312e-01
8.23067188e-01 -1.72603533e-01 1.03297162e+00 7.50314951e-01
-3.07424188e-01 5.94713569e-01 2.93168068e-01 -6.10439479e-01
-9.48028922e-01 -3.63606066e-01 1.17476541e-03 1.11862516e+00
1.09930761e-01 -1.97015256e-01 -8.09421182e-01 -1.09312963e+00
2.31808871e-01 6.14397287e-01 -3.23605746e-01 -3.07203799e-01
-5.29416740e-01 -9.28792238e-01 7.22438097e-01 5.86420774e-01
9.24675703e-01 -6.61713243e-01 -1.20506354e-01 1.06203601e-01
-3.62337172e-01 -1.17735648e+00 -8.60213518e-01 1.52985543e-01
-7.41796434e-01 -8.44088674e-01 -7.37949252e-01 -6.66332126e-01
6.79187357e-01 4.74114209e-01 3.85646611e-01 4.22613025e-02
-2.47362316e-01 1.33878589e-01 -6.06066406e-01 -1.60955384e-01
-1.76719666e-01 4.48999643e-01 -1.04273282e-01 7.02384293e-01
5.78332365e-01 -6.61580935e-02 -3.93117964e-01 5.82118869e-01
-7.44905233e-01 -1.52012423e-01 4.08438087e-01 1.11798751e+00
4.36793178e-01 1.57141194e-01 6.64432824e-01 -1.22577906e+00
4.37941194e-01 -5.01087487e-01 -6.63344145e-01 6.74435854e-01
-9.16965365e-01 -8.21275041e-02 1.01212168e+00 -6.05031908e-01
-1.53450346e+00 1.97641209e-01 2.51130432e-01 -5.00449538e-01
2.77653262e-02 4.92526829e-01 -6.77772403e-01 2.39197873e-02
3.82085860e-01 5.17915308e-01 -1.23627745e-01 -9.19440210e-01
3.10153347e-02 1.45507610e+00 5.93263745e-01 -6.56975865e-01
5.36807597e-01 2.38771573e-01 -6.18883014e-01 -2.64617383e-01
-5.14471173e-01 -4.79508936e-01 -3.00553858e-01 2.13647708e-01
4.65283662e-01 -9.76058602e-01 -9.26139653e-01 9.47990716e-01
-8.11763883e-01 -5.34887202e-02 -1.04142174e-01 4.38471198e-01
-1.62317216e-01 6.02547467e-01 -6.31582618e-01 -9.33628201e-01
-6.84847593e-01 -9.41007376e-01 8.29454303e-01 2.60221004e-01
5.63086510e-01 -5.96517980e-01 -3.18839222e-01 7.75213242e-01
4.89949077e-01 -1.26272872e-01 8.50070119e-01 -1.22973776e+00
-7.70503223e-01 -8.39370549e-01 -2.57444233e-01 6.09085083e-01
2.03207150e-01 -1.72089949e-01 -1.15850639e+00 -2.61624664e-01
-5.01182154e-02 -3.27424437e-01 5.59043765e-01 -2.21635312e-01
1.59774709e+00 -6.52718484e-01 -3.30211520e-01 6.81353688e-01
1.34227848e+00 1.51162893e-01 6.22463882e-01 3.05502623e-01
8.80105913e-01 5.67788959e-01 6.99008465e-01 8.14079106e-01
3.62047374e-01 4.39122736e-01 3.60345513e-01 3.33930641e-01
3.39463145e-01 -5.38472354e-01 3.11150610e-01 7.36079514e-01
3.79871219e-01 -4.26252782e-01 -8.14786673e-01 4.50302392e-01
-2.17235613e+00 -8.12372804e-01 -4.20658998e-02 2.31939650e+00
8.10761213e-01 -3.92455399e-01 -3.97131234e-01 2.98303254e-02
9.78186548e-01 -6.58240914e-02 -9.34263468e-01 -2.24211022e-01
-4.07581061e-01 -3.01418036e-01 8.21448326e-01 -1.15351722e-01
-1.00869393e+00 9.08917129e-01 4.97631168e+00 1.02473080e+00
-1.39688408e+00 1.77711770e-01 1.12380946e+00 5.47509901e-02
-2.22280875e-01 -8.09124261e-02 -1.03198600e+00 7.03944325e-01
8.26183975e-01 -4.49626744e-01 6.12534821e-01 1.35779083e+00
-2.37179577e-01 3.40789855e-01 -8.86996388e-01 1.10761702e+00
1.11060418e-01 -1.46751273e+00 -9.57161859e-02 1.29712433e-01
9.68997300e-01 1.92303047e-01 9.48808435e-03 2.57043839e-01
3.04024130e-01 -8.59747171e-01 3.98834705e-01 4.80045915e-01
7.17088580e-01 -8.75531435e-01 1.06311643e+00 7.70577610e-01
-7.81383991e-01 -4.86314446e-01 -5.55194378e-01 2.95371473e-01
-4.00105119e-01 5.74799597e-01 -5.95604122e-01 8.36373568e-01
9.17984188e-01 5.63240230e-01 -6.06193006e-01 9.40418541e-01
2.89652348e-01 6.24922752e-01 -1.65200725e-01 -2.44087085e-01
-9.37900543e-02 -3.78139913e-02 -9.22973529e-02 9.58326399e-01
2.82640517e-01 -4.59179431e-02 -2.86180788e-04 5.38600624e-01
-6.36869729e-01 4.67785388e-01 -3.44518661e-01 -2.32388571e-01
6.48647189e-01 1.31315947e+00 -8.44508037e-02 -1.68058217e-01
-3.85031730e-01 1.24431169e+00 3.73035848e-01 2.59649545e-01
-6.76199794e-01 -6.49725497e-01 6.15503132e-01 -5.01383066e-01
4.70113337e-01 1.78639472e-01 -4.49430615e-01 -1.53468764e+00
3.68462414e-01 -1.26689088e+00 5.44481337e-01 -3.67627293e-01
-1.49942577e+00 4.48664218e-01 -7.44751155e-01 -1.10610318e+00
-8.51824433e-02 -5.21653593e-01 -5.16505480e-01 1.03355181e+00
-1.33099389e+00 -1.25756633e+00 -4.23291683e-01 8.90999377e-01
1.59344271e-01 -4.32318658e-01 7.63332367e-01 6.51523113e-01
-1.05877745e+00 1.37156487e+00 8.10542226e-01 5.68728864e-01
9.18771327e-01 -6.78243101e-01 3.74272346e-01 9.91627395e-01
-1.67565376e-01 6.66816115e-01 -1.72738675e-02 -7.14835048e-01
-1.78441823e+00 -1.38662255e+00 9.37084079e-01 -2.31305733e-01
2.82608628e-01 -5.11987388e-01 -1.18079996e+00 7.49398470e-01
-2.56076634e-01 5.12711823e-01 8.46719503e-01 -1.80440880e-02
-5.71825325e-01 -6.77671194e-01 -1.52767718e+00 2.48522684e-01
6.37558818e-01 -6.25943303e-01 -5.24662882e-02 4.41945344e-01
8.80888700e-01 -3.63496304e-01 -8.32283199e-01 -6.22413158e-02
5.38239658e-01 -5.97822905e-01 4.47872102e-01 -8.22108567e-01
1.40047699e-01 -2.14160830e-01 -5.69726586e-01 -9.09668803e-01
-7.32571781e-02 -5.58186650e-01 -2.04162784e-02 1.71215117e+00
3.00036699e-01 -9.94823635e-01 1.30078506e+00 1.20747364e+00
2.53363070e-03 -4.53588217e-01 -1.08806098e+00 -6.47383571e-01
-1.01590110e-02 -4.64450270e-01 1.24214661e+00 1.22814584e+00
7.18212686e-04 -1.36758491e-01 -7.08796918e-01 3.00359964e-01
6.05047345e-01 1.81331873e-01 9.98324513e-01 -8.94679546e-01
-2.69962341e-01 4.73089181e-02 -1.18122853e-01 -9.68688786e-01
1.61432534e-01 -8.40374947e-01 -1.05334327e-01 -9.92424190e-01
3.57281864e-01 -7.58732557e-01 -4.52543259e-01 7.35825717e-01
-1.47578269e-01 -1.00993283e-01 3.35441113e-01 2.72630334e-01
-6.23028994e-01 6.49565995e-01 1.00272000e+00 -3.04303676e-01
6.05792291e-02 2.18109310e-01 -9.23207164e-01 1.41376734e-01
8.48734021e-01 -4.95427400e-01 1.77133046e-02 -7.73126185e-01
-9.91540775e-02 -8.78430828e-02 1.17564738e-01 -7.51052737e-01
7.50754476e-01 -1.72999784e-01 5.54180503e-01 -7.33595610e-01
-9.46700945e-02 -1.17471910e+00 1.10122412e-02 3.58383030e-01
-2.66907752e-01 -2.43330136e-01 -1.51354581e-01 7.07539916e-01
-1.25117376e-01 -2.74356991e-01 4.46259290e-01 1.43037185e-01
-4.73624617e-01 4.17131275e-01 -4.37224582e-02 -2.91443467e-01
9.69179869e-01 -5.74765094e-02 -7.53668249e-01 -9.47881714e-02
-1.27242789e-01 4.37654555e-01 4.42404211e-01 4.60748106e-01
4.33313608e-01 -1.17239571e+00 -4.70802218e-01 2.91054517e-01
2.33937562e-01 1.58917736e-02 7.57552803e-01 5.53343713e-01
-2.86461294e-01 5.31070471e-01 4.31672409e-02 -3.18561792e-01
-1.20229375e+00 5.89917421e-01 4.01646882e-01 -3.32076192e-01
-5.51806986e-01 5.72666943e-01 -2.15168372e-01 -9.53557074e-01
5.03693044e-01 -6.09642752e-02 1.76008195e-01 -2.07394019e-01
8.19489658e-01 6.36238635e-01 4.21543360e-01 -2.84864187e-01
-3.91333550e-01 1.71119392e-01 -5.51288068e-01 3.78526688e-01
1.38806820e+00 -1.29868716e-01 -1.69740379e-01 1.12052128e-01
1.38581610e+00 1.60897538e-01 -1.30607104e+00 -5.15587211e-01
-1.56682491e-01 -7.79159427e-01 9.21898782e-02 -1.01384473e+00
-1.45714831e+00 7.34732211e-01 7.14235604e-01 -3.06872964e-01
1.15438509e+00 -5.57666063e-01 9.92642820e-01 5.99232256e-01
5.00453830e-01 -1.27164221e+00 -8.92413110e-02 4.03534502e-01
2.90689021e-01 -1.21325314e+00 -9.21424776e-02 -2.19469026e-01
-1.08645201e+00 1.01563907e+00 7.86044061e-01 3.26573730e-01
4.70674455e-01 2.00515792e-01 1.85590982e-01 4.83984590e-01
-7.93947637e-01 6.12184405e-01 -6.55883104e-02 4.64478195e-01
-2.23613605e-02 1.86372057e-01 -1.21537998e-01 1.36153543e+00
2.67750382e-01 2.13443130e-01 4.64493275e-01 1.04268515e+00
-9.88885611e-02 -1.30093932e+00 -3.33426178e-01 5.67138314e-01
-7.01184154e-01 -7.45540718e-03 -3.52151960e-01 2.21009299e-01
-6.28350750e-02 9.28484499e-01 6.21928945e-02 -7.49346912e-01
1.53100997e-01 3.07378292e-01 -2.38676354e-01 -2.10168660e-01
-8.75144601e-01 -1.20720677e-01 -1.10941350e-01 -3.96721572e-01
1.60894573e-01 -8.07088971e-01 -1.24096680e+00 -6.32980704e-01
-5.48920870e-01 2.17030317e-01 8.64438891e-01 7.51324356e-01
8.87089729e-01 3.17185372e-01 1.16223419e+00 -1.76308706e-01
-1.12205076e+00 -7.08131313e-01 -7.34583914e-01 2.46664748e-01
-2.32682340e-02 4.42908816e-02 -3.01422596e-01 -7.86530450e-02] | [5.845946788787842, 6.353536128997803] |
0b59cda7-e6b7-4dd5-a8a1-47af579d0634 | sindiffusion-learning-a-diffusion-model-from | 2211.12445 | null | https://arxiv.org/abs/2211.12445v1 | https://arxiv.org/pdf/2211.12445v1.pdf | SinDiffusion: Learning a Diffusion Model from a Single Natural Image | We present SinDiffusion, leveraging denoising diffusion models to capture internal distribution of patches from a single natural image. SinDiffusion significantly improves the quality and diversity of generated samples compared with existing GAN-based approaches. It is based on two core designs. First, SinDiffusion is trained with a single model at a single scale instead of multiple models with progressive growing of scales which serves as the default setting in prior work. This avoids the accumulation of errors, which cause characteristic artifacts in generated results. Second, we identify that a patch-level receptive field of the diffusion network is crucial and effective for capturing the image's patch statistics, therefore we redesign the network structure of the diffusion model. Coupling these two designs enables us to generate photorealistic and diverse images from a single image. Furthermore, SinDiffusion can be applied to various applications, i.e., text-guided image generation, and image outpainting, due to the inherent capability of diffusion models. Extensive experiments on a wide range of images demonstrate the superiority of our proposed method for modeling the patch distribution. | ['Houqiang Li', 'Lu Yuan', 'Dong Chen', 'Dongdong Chen', 'Wengang Zhou', 'Jianmin Bao', 'Weilun Wang'] | 2022-11-22 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 3.91724676e-01 -5.14997467e-02 9.52034891e-02 2.59270146e-02
-3.75237823e-01 -4.67008144e-01 4.24796999e-01 -5.29821932e-01
1.65690944e-01 8.31035137e-01 3.54914784e-01 1.86144173e-01
1.76453635e-01 -8.99504840e-01 -7.24652350e-01 -1.09134579e+00
4.79288489e-01 2.26663351e-02 2.87133485e-01 -3.20959181e-01
2.35730410e-01 5.08309841e-01 -1.23058724e+00 2.91768730e-01
9.50504005e-01 1.06669724e+00 4.65861231e-01 5.11417568e-01
-2.60247737e-02 7.87896574e-01 -7.70791471e-01 -2.95221418e-01
3.86079818e-01 -8.38762760e-01 -3.44008714e-01 3.94130975e-01
4.11895066e-01 -5.61740279e-01 -4.06359524e-01 1.03775775e+00
5.88440716e-01 -2.91612409e-02 7.23731995e-01 -1.05861259e+00
-1.20250392e+00 1.59184813e-01 -1.04591680e+00 1.08980611e-01
2.00914182e-02 2.90995747e-01 6.90102160e-01 -6.02298677e-01
7.32039452e-01 1.29424632e+00 6.72544301e-01 7.17747331e-01
-1.37026918e+00 -6.29014432e-01 -3.93880308e-02 -2.11582601e-01
-1.16322613e+00 -2.95262218e-01 1.22080708e+00 -2.72007912e-01
4.08702791e-01 1.68581367e-01 8.06134462e-01 1.26187074e+00
3.84459555e-01 8.07008982e-01 1.30855572e+00 -4.04850692e-01
2.16722921e-01 -1.32466033e-01 -5.18678546e-01 4.87326533e-01
1.27404360e-02 1.07190676e-01 -4.11409557e-01 -1.42841563e-01
1.63337433e+00 1.88303947e-01 -4.31704998e-01 -3.61193508e-01
-1.17067134e+00 7.76160598e-01 5.46584725e-01 2.70632744e-01
-6.35382056e-01 3.64033192e-01 1.51711687e-01 2.05209985e-01
5.23307323e-01 2.22729802e-01 -3.85897979e-02 1.54745206e-02
-9.27967966e-01 2.39601359e-01 2.74825990e-01 9.13429677e-01
7.63597250e-01 4.04061466e-01 -4.73964125e-01 1.13491035e+00
1.12689070e-01 5.00660658e-01 5.09713352e-01 -1.16480339e+00
1.26169980e-01 4.62726027e-01 1.28987506e-01 -1.11898232e+00
3.13934177e-01 -3.79956692e-01 -1.41715479e+00 2.76445448e-01
2.80587021e-02 -1.80682838e-01 -1.18773043e+00 1.74240351e+00
3.08746845e-01 1.79594800e-01 -1.67318866e-01 7.63769269e-01
6.00663483e-01 8.03567231e-01 -2.29101688e-01 -9.20128077e-02
9.72703397e-01 -1.15528703e+00 -8.03164005e-01 -8.70313719e-02
-5.62123135e-02 -8.92159045e-01 1.14249909e+00 4.26193625e-01
-1.26002097e+00 -6.63527727e-01 -9.73370969e-01 1.25225291e-01
8.83936137e-02 -2.28151083e-02 5.67204773e-01 4.28330868e-01
-1.30429256e+00 4.32722926e-01 -4.70127136e-01 -1.09483317e-01
7.25453436e-01 1.72320916e-03 -1.60447404e-01 -2.12069288e-01
-8.67361724e-01 3.71286571e-01 -8.75618532e-02 1.05048679e-01
-1.03766322e+00 -5.50430834e-01 -6.20589793e-01 4.51773629e-02
-3.09407301e-02 -9.15963471e-01 8.84032905e-01 -1.23996162e+00
-1.59485233e+00 5.07733285e-01 -1.89299762e-01 -1.48998439e-01
5.89812100e-01 9.20497552e-02 -2.00104788e-01 1.93145052e-01
1.51500031e-01 8.82344186e-01 1.36378288e+00 -1.56907248e+00
-3.11201364e-01 -1.90275088e-01 -2.06848070e-01 1.04261093e-01
-3.29737842e-01 -3.25768948e-01 -4.49703634e-01 -1.27282643e+00
-8.35357141e-03 -7.66950488e-01 -2.42786229e-01 2.49425739e-01
-2.74561584e-01 1.59665659e-01 1.18664062e+00 -6.45722032e-01
1.00641024e+00 -2.20239210e+00 8.60408694e-02 1.18524611e-01
4.46372628e-01 1.88860372e-01 -3.76521766e-01 6.64820492e-01
5.92715442e-02 5.64754978e-02 -4.36075658e-01 -3.50630194e-01
-2.69330084e-01 2.17691198e-01 -4.12471205e-01 2.08819494e-01
1.41540036e-01 1.06735837e+00 -6.75792158e-01 -3.76317978e-01
1.09130740e-01 6.81613505e-01 -5.36068439e-01 3.39912564e-01
-2.32592732e-01 8.20102215e-01 -3.66749495e-01 5.20252347e-01
1.06045842e+00 -1.53762609e-01 -3.26378047e-01 -2.46634364e-01
1.98910579e-01 -1.90815836e-01 -9.71268535e-01 1.85050142e+00
-5.31228781e-01 4.60940212e-01 2.21036628e-01 -6.65712297e-01
1.02065003e+00 1.40863270e-01 3.18942606e-01 -7.46318340e-01
-4.40874981e-04 8.09694082e-02 -1.44314542e-01 -2.70457208e-01
3.09994787e-01 -2.65363365e-01 3.83503169e-01 7.09021568e-01
-1.80577189e-02 -3.12972784e-01 -1.20886276e-02 1.74836174e-01
9.42455769e-01 4.55605686e-02 -1.06055439e-01 -2.79091775e-01
3.35979819e-01 -3.48455936e-01 6.21138215e-01 6.39971852e-01
2.79484950e-02 1.12380922e+00 3.60709578e-01 -3.87022048e-01
-1.36773431e+00 -1.03865480e+00 6.64178580e-02 4.58501399e-01
3.21638763e-01 -1.45834059e-01 -9.08984125e-01 -6.34217680e-01
-2.91455448e-01 4.25808936e-01 -9.57420826e-01 -1.21859029e-01
-4.05684233e-01 -6.50991678e-01 4.91214395e-01 5.35798252e-01
9.83037770e-01 -1.26017237e+00 -3.10882956e-01 1.06663875e-01
-2.33179018e-01 -8.08719575e-01 -8.78116131e-01 -3.49886924e-01
-9.57961917e-01 -7.57000983e-01 -1.25665224e+00 -1.05886471e+00
9.39383507e-01 5.62436521e-01 1.14716744e+00 2.58024633e-01
-2.73359686e-01 9.47843865e-02 -3.01611394e-01 -2.52328627e-02
-5.87013185e-01 -2.52352417e-01 -4.25044596e-01 2.45863691e-01
-2.42133886e-01 -1.00268948e+00 -1.25696647e+00 4.54805642e-01
-1.43402040e+00 2.46367812e-01 8.63531232e-01 1.22231901e+00
6.71424031e-01 3.89592677e-01 6.20220006e-01 -7.94249117e-01
1.02452648e+00 -2.22894624e-01 -2.74233490e-01 1.41951412e-01
-4.49056596e-01 -2.07519844e-01 7.39890635e-01 -7.09077358e-01
-1.31717122e+00 -2.80284256e-01 -1.76723480e-01 -5.34541726e-01
-2.92885583e-02 9.90724750e-03 -1.62186265e-01 -2.92221963e-01
5.41068673e-01 6.99586511e-01 2.73271531e-01 -3.00030321e-01
3.74634594e-01 5.69345593e-01 3.65818113e-01 -5.50956249e-01
6.34386539e-01 6.40554547e-01 -1.83677748e-01 -7.03589857e-01
-3.42532873e-01 -5.57705872e-02 -3.56549621e-01 -1.07793242e-01
6.27467811e-01 -8.57231975e-01 -5.09433687e-01 9.30628955e-01
-1.21776748e+00 -5.15329123e-01 -4.65124756e-01 -5.04582264e-02
-5.70735514e-01 4.40588236e-01 -8.62939060e-01 -5.65187871e-01
-5.57917595e-01 -1.08031046e+00 1.21546650e+00 2.26375505e-01
2.81406641e-02 -1.03728116e+00 1.25974134e-01 1.68703213e-01
7.28321850e-01 3.55997771e-01 1.02922976e+00 3.81884247e-01
-5.28390825e-01 1.21871360e-01 -2.58462757e-01 7.41376519e-01
4.89095867e-01 -5.66078648e-02 -8.53275299e-01 -4.74208325e-01
3.84649873e-01 -3.28730851e-01 8.19631100e-01 5.76755345e-01
1.09134650e+00 -3.81756485e-01 -2.29465917e-01 6.04805112e-01
1.53036273e+00 2.16546029e-01 1.15853250e+00 1.18197061e-01
7.57293284e-01 4.20223087e-01 1.63880467e-01 3.47041368e-01
6.56715482e-02 5.47957957e-01 3.65756929e-01 -6.24131799e-01
-7.18335032e-01 -5.46481788e-01 1.37106180e-01 7.43814111e-01
9.59919244e-02 -3.31070572e-01 -4.11278635e-01 6.84017658e-01
-1.64889967e+00 -9.69621480e-01 1.16122954e-01 1.90418279e+00
8.28665435e-01 -1.15348242e-01 1.93076283e-02 -1.02587320e-01
8.61587822e-01 2.32643351e-01 -7.91426003e-01 -2.43678048e-01
-3.31342667e-01 1.26357049e-01 1.98295072e-01 2.45175436e-01
-4.80257750e-01 8.85097682e-01 6.86672926e+00 1.28024638e+00
-1.22284842e+00 1.39478296e-01 1.05826890e+00 1.26961842e-01
-6.52182698e-01 -6.57038242e-02 -4.16543722e-01 7.22802401e-01
1.35935917e-01 1.07433945e-01 5.99407196e-01 4.44582641e-01
2.65882671e-01 -1.40845224e-01 -6.11012936e-01 1.15440178e+00
2.12081760e-01 -1.48937261e+00 4.38677669e-01 1.71698153e-01
1.24999607e+00 -3.51627648e-01 3.67080480e-01 -2.59279847e-01
3.45196426e-01 -1.10814774e+00 7.22153723e-01 4.08210605e-01
8.86713207e-01 -7.62037635e-01 5.81059933e-01 3.79805535e-01
-8.73456717e-01 8.81759673e-02 -4.64245677e-01 2.69456476e-01
2.47040719e-01 8.69813800e-01 -3.78902167e-01 3.21184188e-01
7.60710955e-01 7.60438800e-01 -3.79483104e-01 8.14487815e-01
-3.36046576e-01 5.88132143e-01 -1.38886556e-01 2.87738055e-01
1.47752732e-01 -5.29672205e-01 3.91487777e-01 7.67800927e-01
5.08117974e-01 -9.48109850e-02 -3.16282094e-01 1.28239191e+00
-1.55631632e-01 -1.05951935e-01 -8.02748799e-01 2.31136188e-01
4.78641152e-01 1.23341346e+00 -7.31056213e-01 -2.59195566e-01
-2.08740577e-01 1.41394544e+00 2.52152264e-01 5.64840913e-01
-8.06039095e-01 -2.82230526e-01 2.98555315e-01 3.41747016e-01
7.51372099e-01 -1.56308606e-01 -3.00043374e-01 -1.05085456e+00
1.47663355e-01 -1.17562544e+00 -1.72537550e-01 -1.05289519e+00
-1.41651332e+00 9.09928739e-01 -2.76824921e-01 -1.30036128e+00
4.30725850e-02 -6.42029494e-02 -8.83039951e-01 9.69835997e-01
-1.27800465e+00 -1.36714005e+00 -5.46629667e-01 6.76767826e-01
6.08299434e-01 -1.41898310e-02 5.82905293e-01 1.49415344e-01
-5.07830203e-01 4.79533881e-01 3.07486594e-01 -2.68367585e-02
7.26032734e-01 -1.00092983e+00 5.69274366e-01 8.64564240e-01
-2.92525142e-02 5.03007174e-01 4.61401433e-01 -7.82485127e-01
-1.08297527e+00 -1.07836032e+00 1.01129115e-01 -8.62379149e-02
2.65116930e-01 -3.14000875e-01 -7.41607785e-01 3.79988104e-01
6.57823443e-01 -1.60715982e-01 3.86543274e-01 -5.57769179e-01
-6.45047203e-02 -2.88474649e-01 -1.33052313e+00 7.48677731e-01
1.04301941e+00 -2.88384914e-01 -4.00883779e-02 4.88803722e-02
4.99426484e-01 -3.21305364e-01 -8.66967022e-01 2.63866574e-01
4.26144570e-01 -1.30120432e+00 8.99778903e-01 1.09324627e-01
7.50029743e-01 -3.89308631e-01 4.88170758e-02 -1.58068240e+00
-5.57089746e-01 -9.30023491e-01 -8.18871334e-02 1.54259551e+00
8.21947977e-02 -6.63532972e-01 8.42638552e-01 1.95834979e-01
1.46894321e-01 -8.30112278e-01 -6.03383124e-01 -6.43163383e-01
6.77991733e-02 2.80258566e-01 8.82650018e-01 7.95170248e-01
-7.98957825e-01 3.79780918e-01 -6.87346876e-01 -1.96621582e-01
7.59994805e-01 2.16406494e-01 8.53666008e-01 -7.35725820e-01
-4.87735063e-01 -3.55735928e-01 -1.91630006e-01 -1.54817986e+00
-2.75878489e-01 -3.88229877e-01 -5.54268248e-03 -1.51171458e+00
2.29372859e-01 -6.69025064e-01 3.14422622e-02 1.74450263e-01
-1.03092469e-01 6.54945374e-01 2.58941352e-02 6.03575826e-01
1.77894849e-02 1.00993717e+00 2.04252386e+00 -2.39278778e-01
-8.38173926e-02 -2.29664728e-01 -9.74154830e-01 4.94045049e-01
7.45461881e-01 -3.32200944e-01 -7.79543817e-01 -7.02663362e-01
1.03009660e-02 4.33236957e-02 3.44491154e-01 -8.86736989e-01
1.82229523e-02 -1.57599762e-01 6.39230847e-01 -3.17623228e-01
2.43384972e-01 -6.61265314e-01 4.06658202e-01 3.18380266e-01
-1.96148559e-01 2.12682515e-01 1.31533667e-01 8.31841111e-01
-5.30151844e-01 -4.16591167e-02 1.05289376e+00 -2.49831006e-01
-4.85534608e-01 4.16453212e-01 -9.72350910e-02 -4.39142212e-02
9.90878284e-01 -5.34718037e-01 -2.46655047e-01 -6.53005421e-01
-2.93616772e-01 -1.46052271e-01 9.17043805e-01 1.79297641e-01
8.51714790e-01 -1.59341967e+00 -6.06932640e-01 5.61556756e-01
-2.21050784e-01 3.54341090e-01 4.87614930e-01 7.32096851e-01
-7.40836501e-01 -2.05074206e-01 -4.19254363e-01 -6.31790936e-01
-8.51250231e-01 4.61342752e-01 1.93304822e-01 -2.17917114e-01
-8.11303556e-01 7.67345488e-01 7.39525020e-01 -6.67731911e-02
-4.29979749e-02 -1.45463347e-01 7.18330666e-02 -4.80060577e-01
5.64481199e-01 1.15182169e-01 -1.45373642e-01 -4.97459471e-01
1.17085934e-01 7.70276308e-01 -2.66158760e-01 -1.42604589e-01
1.34149528e+00 -3.14856589e-01 -3.09331864e-01 3.89362201e-02
9.15908039e-01 1.62652865e-01 -1.78638279e+00 2.28496362e-03
-9.13020253e-01 -7.03059673e-01 1.97567925e-01 -7.83805966e-01
-1.55769026e+00 8.40435505e-01 4.94070023e-01 1.59149066e-01
1.39083862e+00 -3.23526591e-01 1.31382489e+00 -2.54595071e-01
3.08201313e-01 -8.38447034e-01 4.06623840e-01 6.19655894e-03
1.10287309e+00 -9.10225809e-01 -1.33478716e-01 -4.32402521e-01
-7.03256786e-01 8.77594709e-01 6.66574717e-01 -3.93019944e-01
4.66483891e-01 3.56165439e-01 1.57056987e-01 -9.32906419e-02
-5.97853899e-01 2.64775366e-01 1.57916155e-02 8.56257677e-01
2.01750532e-01 -2.64132738e-01 -2.19065323e-01 1.72522902e-01
1.02455951e-01 8.78247768e-02 5.50172508e-01 8.58767688e-01
-1.36257276e-01 -1.16170621e+00 -3.81006420e-01 3.47491324e-01
-2.90295869e-01 -2.54908234e-01 -3.03644150e-01 5.25446951e-01
1.77807823e-01 8.17283928e-01 -1.11721784e-01 -2.49646634e-01
1.56920239e-01 -5.32181203e-01 6.48361742e-01 -3.25509042e-01
-3.06834072e-01 3.89928222e-01 -4.64622706e-01 -3.80412579e-01
-5.03617287e-01 -1.75444067e-01 -7.42950737e-01 -3.99904460e-01
-2.26421446e-01 3.69946323e-02 3.34854275e-01 5.38982034e-01
7.28664219e-01 4.33810085e-01 7.86375999e-01 -8.89628649e-01
-2.94525921e-01 -1.05534446e+00 -9.00490344e-01 6.68143868e-01
3.54802072e-01 -6.20857656e-01 -3.27124566e-01 2.41279319e-01] | [11.455056190490723, -0.8792266249656677] |
2d362ee8-68e2-4a40-84dc-18a28957c12d | scalable-object-detection-for-stylized | 1711.09822 | null | http://arxiv.org/abs/1711.09822v2 | http://arxiv.org/pdf/1711.09822v2.pdf | Scalable Object Detection for Stylized Objects | Following recent breakthroughs in convolutional neural networks and
monolithic model architectures, state-of-the-art object detection models can
reliably and accurately scale into the realm of up to thousands of classes.
Things quickly break down, however, when scaling into the tens of thousands,
or, eventually, to millions or billions of unique objects. Further, bounding
box-trained end-to-end models require extensive training data. Even though -
with some tricks using hierarchies - one can sometimes scale up to thousands of
classes, the labor requirements for clean image annotations quickly get out of
control. In this paper, we present a two-layer object detection method for
brand logos and other stylized objects for which prototypical images exist. It
can scale to large numbers of unique classes. Our first layer is a CNN from the
Single Shot Multibox Detector family of models that learns to propose regions
where some stylized object is likely to appear. The contents of a proposed
bounding box is then run against an image index that is targeted for the
retrieval task at hand. The proposed architecture scales to a large number of
object classes, allows to continously add new classes without retraining, and
exhibits state-of-the-art quality on a stylized object detection task such as
logo recognition. | ['Willi Richert', 'Thilo Will', 'William Darling', 'Clemens Marschner', 'Aayush Garg'] | 2017-11-27 | null | null | null | null | ['logo-recognition'] | ['computer-vision'] | [ 8.85687023e-02 -2.18371585e-01 -1.80698544e-01 -1.25497073e-01
-8.09280872e-01 -7.89326608e-01 6.57171190e-01 2.41279781e-01
-1.55080736e-01 1.32168740e-01 -4.01588261e-01 5.53335063e-03
2.41823703e-01 -9.11920190e-01 -1.08516049e+00 -3.14239740e-01
-2.78966334e-02 8.52936745e-01 8.87148976e-01 -1.42238200e-01
1.65106878e-01 8.11674714e-01 -2.05107713e+00 4.38433111e-01
4.54773843e-01 1.26361656e+00 1.19314186e-01 7.82173991e-01
-4.16174322e-01 5.39334953e-01 -7.51461804e-01 -7.34177828e-01
5.12745082e-01 2.43168324e-01 -5.04770696e-01 4.02497709e-01
1.45502245e+00 -7.36577332e-01 -2.22094253e-01 1.00873101e+00
2.97152311e-01 -2.87203610e-01 7.46532142e-01 -1.06956172e+00
-1.00745678e+00 4.91817832e-01 -7.58912683e-01 2.77055115e-01
-1.72534898e-01 1.99682802e-01 1.25213981e+00 -1.31240404e+00
7.70520806e-01 1.51108646e+00 5.43461204e-01 6.41213357e-01
-1.00970507e+00 -8.37607980e-01 3.63054663e-01 -1.91113845e-01
-1.16731405e+00 -2.97837019e-01 2.36478835e-01 -6.16883278e-01
9.57311094e-01 -5.08654155e-02 5.63138843e-01 9.88747239e-01
6.85735717e-02 1.03516853e+00 6.30605698e-01 -2.50257343e-01
-1.20979780e-02 3.00240159e-01 3.69257718e-01 1.09653926e+00
8.15851510e-01 -9.50726941e-02 -7.05373704e-01 3.35893743e-02
5.94025970e-01 2.28247806e-01 5.52040935e-01 -4.41307038e-01
-8.24577928e-01 7.77327180e-01 5.62503159e-01 1.23732597e-01
1.60721645e-01 4.74171847e-01 5.39413095e-01 1.14908377e-02
5.70480108e-01 5.77081859e-01 -2.03237638e-01 2.40588650e-01
-1.22338820e+00 4.49715346e-01 7.04030514e-01 1.03589571e+00
7.25774765e-01 2.51771957e-02 6.70646206e-02 6.32722855e-01
-5.20362183e-02 5.31290472e-01 3.94254178e-01 -3.22760791e-01
4.17198956e-01 8.99968922e-01 1.36057869e-01 -7.13038981e-01
-1.38471693e-01 -4.78642851e-01 -4.69879746e-01 5.78843892e-01
5.31773746e-01 3.98835868e-01 -1.33753431e+00 1.17073107e+00
4.65143085e-01 -1.00820832e-01 -4.63176906e-01 7.60233998e-01
8.73291373e-01 5.36825836e-01 -7.61032477e-02 4.56899315e-01
1.77703464e+00 -1.02351403e+00 -2.93323338e-01 -7.52047002e-01
3.83540690e-01 -8.20409894e-01 1.31422424e+00 4.37218428e-01
-9.90363717e-01 -7.39578485e-01 -1.27032864e+00 -2.10094184e-01
-9.02287424e-01 2.34145060e-01 7.63250768e-01 7.27809489e-01
-6.63181365e-01 6.82899952e-01 -6.20399833e-01 -6.13267899e-01
9.74870920e-01 4.12127048e-01 -2.08836153e-01 1.28291324e-01
-5.43640971e-01 8.22639644e-01 3.61053735e-01 -1.63928747e-01
-1.31799376e+00 -6.46326661e-01 -6.61940932e-01 3.90443653e-01
6.80932403e-01 -3.30202758e-01 1.27551270e+00 -1.00296557e+00
-8.34860921e-01 1.18900120e+00 -7.02794790e-02 -3.79704475e-01
4.95457232e-01 -7.05528140e-01 -2.26321191e-01 1.07102215e-01
3.72120887e-01 7.19759405e-01 1.45659006e+00 -1.24087203e+00
-9.85418916e-01 -3.57067943e-01 7.18884841e-02 -1.28180817e-01
-7.34264195e-01 4.32515591e-01 -5.18188119e-01 -4.74302709e-01
5.30043319e-02 -5.75707436e-01 7.52627924e-02 4.09615934e-01
-3.54213953e-01 -4.12023693e-01 1.10395980e+00 -1.66565821e-01
1.01241851e+00 -2.06376481e+00 -1.46369830e-01 -1.76289648e-01
4.57131028e-01 3.55980724e-01 -3.04043531e-01 1.59124210e-01
-2.02736054e-02 2.05843702e-01 5.08104041e-02 -5.01912415e-01
1.72187239e-01 -7.54380971e-02 -7.73264587e-01 3.57296646e-01
6.69623673e-01 9.38584268e-01 -7.53020883e-01 -5.17985404e-01
2.07432181e-01 6.86623678e-02 -3.65656585e-01 9.20872316e-02
-4.42191273e-01 -2.84256220e-01 -2.05031261e-01 1.16806412e+00
4.29376543e-01 -6.66999578e-01 -3.87521476e-01 1.42047897e-01
1.90502226e-01 -2.09905505e-02 -1.29975760e+00 1.18417573e+00
-2.50505179e-01 7.18949080e-01 2.12068539e-02 -5.28888822e-01
8.79939675e-01 -1.57317922e-01 2.74772167e-01 -2.42442444e-01
1.76137567e-01 4.81437832e-01 -2.56885707e-01 -1.95251375e-01
7.40356863e-01 -5.69440462e-02 -1.90878466e-01 4.06568974e-01
4.04039532e-01 -2.30206773e-01 4.90990698e-01 3.64958346e-01
1.08033550e+00 3.03448644e-02 4.21041340e-01 8.07848796e-02
1.44108728e-01 -7.89159071e-03 1.30031139e-01 1.06903493e+00
-1.75516516e-01 5.56593776e-01 5.00005424e-01 -9.45966661e-01
-1.34511757e+00 -1.18103409e+00 -2.12135613e-01 1.74568784e+00
2.42002845e-01 -3.88148218e-01 -7.22213387e-01 -7.26344526e-01
4.01316643e-01 1.24685923e-02 -7.09870696e-01 -5.32607920e-02
-6.83757186e-01 -7.24916995e-01 6.15173042e-01 7.08226919e-01
2.77285635e-01 -1.12093627e+00 -7.48191595e-01 2.28210568e-01
5.46668768e-01 -1.15629482e+00 -4.25033212e-01 4.24307674e-01
-9.39055562e-01 -1.08121908e+00 -6.55430734e-01 -7.86768138e-01
5.31770706e-01 4.23616230e-01 1.31529427e+00 2.03032345e-01
-9.72539425e-01 8.48388076e-02 -1.52601182e-01 -8.73399973e-01
-2.47127473e-01 1.15658998e-01 2.34245941e-01 1.32195055e-01
6.23416364e-01 -2.53075749e-01 -3.78834546e-01 1.82071343e-01
-9.82606947e-01 -3.41831505e-01 7.18254328e-01 7.50832200e-01
4.06775147e-01 -3.61161590e-01 4.52769697e-01 -9.24959123e-01
3.18876475e-01 -1.46758646e-01 -9.70507324e-01 3.04088891e-01
-4.36401069e-01 -5.60249537e-02 5.27384818e-01 -8.24009120e-01
-6.71506584e-01 3.49474967e-01 3.04909855e-01 -6.88247263e-01
-2.35694632e-01 -1.73711836e-01 -1.24164587e-02 -9.07941386e-02
9.21777785e-01 -6.15816191e-02 -5.23132801e-01 -5.54149985e-01
8.15018058e-01 7.08668172e-01 5.72568953e-01 -2.93463796e-01
1.13458633e+00 7.38704622e-01 -1.62868887e-01 -8.10752392e-01
-1.28497171e+00 -7.71423280e-01 -9.39922750e-01 -1.43194526e-01
7.61803806e-01 -9.28872108e-01 -6.67285740e-01 5.34246683e-01
-1.15335274e+00 -3.06834131e-01 -4.26584482e-01 -3.17256689e-01
-2.30479360e-01 2.51246959e-01 -6.68671012e-01 -8.64349067e-01
-4.24574107e-01 -8.90856922e-01 1.73649633e+00 3.32131147e-01
-1.11730695e-01 -3.82445067e-01 -2.87146747e-01 1.35188550e-01
1.76476672e-01 1.39341161e-01 9.28651273e-01 -6.64040983e-01
-1.07652640e+00 -9.94499564e-01 -4.64528680e-01 6.92515001e-02
3.34344544e-02 2.36934602e-01 -1.20183349e+00 -2.97029227e-01
-4.54953939e-01 -5.65820575e-01 1.15441358e+00 1.76788002e-01
1.28260660e+00 -1.65963829e-01 -6.10327005e-01 5.20242095e-01
1.31662786e+00 -1.37791559e-01 3.00866306e-01 3.58491451e-01
7.95724869e-01 4.73592311e-01 3.90887260e-01 3.18628848e-01
-1.86696216e-01 5.24634838e-01 7.66379774e-01 1.06414519e-02
-2.93337077e-01 -1.71219394e-01 3.06538254e-01 2.02527270e-01
1.56331405e-01 -4.62338209e-01 -7.13709176e-01 6.66583121e-01
-1.60817707e+00 -9.93659616e-01 1.82198092e-01 2.15260458e+00
5.15984654e-01 6.68504536e-01 4.00647074e-01 -1.11523055e-01
7.59203732e-01 2.60943592e-01 -8.47678125e-01 -3.19257796e-01
-1.21483624e-01 3.11311662e-01 5.39196074e-01 1.10831536e-01
-1.58622503e+00 1.47696447e+00 6.83465338e+00 1.03055561e+00
-1.15072572e+00 1.33537114e-01 5.31226516e-01 -4.85146195e-01
4.79945391e-01 -2.65285283e-01 -1.47467172e+00 1.77157626e-01
6.46145761e-01 1.20178819e-01 2.51321673e-01 1.44404697e+00
-4.17887926e-01 -2.50261992e-01 -1.31127787e+00 9.22789931e-01
3.06179762e-01 -1.54925787e+00 3.08686256e-01 1.23597249e-01
6.96736634e-01 8.19612294e-02 2.64229447e-01 3.98714423e-01
3.27514529e-01 -1.16925514e+00 1.19391251e+00 3.89426015e-02
8.21735322e-01 -5.80853641e-01 3.15230966e-01 2.71689773e-01
-1.43681681e+00 -4.81159806e-01 -6.77158296e-01 1.25342786e-01
-8.23058859e-02 3.75011116e-01 -9.69859540e-01 -2.14977682e-01
9.93797660e-01 4.93845552e-01 -1.16034853e+00 1.14773166e+00
-8.04541185e-02 4.49288785e-01 -3.81642789e-01 -4.67645645e-01
4.46387023e-01 3.09577912e-01 3.65457028e-01 1.25806272e+00
1.41010404e-01 -2.55949706e-01 1.19333871e-01 8.57361436e-01
-4.56223249e-01 -1.60655752e-02 -6.84363723e-01 -2.48415202e-01
3.70269120e-01 1.54030800e+00 -1.17639065e+00 -8.56943429e-01
-3.47593606e-01 9.66912091e-01 7.02553630e-01 -1.65181886e-02
-6.64722562e-01 -6.10674560e-01 5.81166267e-01 5.05081415e-01
6.22543156e-01 -5.09602837e-02 -4.99271333e-01 -9.99476969e-01
1.66926369e-01 -8.34510088e-01 2.85286725e-01 -7.88091958e-01
-1.55233729e+00 4.91524279e-01 -2.32377276e-01 -1.19806731e+00
3.76970880e-02 -1.36874259e+00 -7.28413641e-01 4.93381023e-01
-1.10432053e+00 -1.59176648e+00 -3.61395329e-01 8.33809078e-02
7.30991185e-01 -4.08997416e-01 6.58034682e-01 2.42503420e-01
-4.45364863e-01 4.77340341e-01 2.62454093e-01 5.09536505e-01
8.58932376e-01 -1.36073315e+00 8.73725951e-01 8.67969811e-01
5.34162521e-01 4.79619950e-01 5.66730678e-01 -7.53134191e-01
-1.16816545e+00 -1.14684284e+00 6.70312583e-01 -8.50823402e-01
1.06126785e+00 -1.01935518e+00 -9.91435826e-01 6.91079915e-01
-2.42167488e-01 3.44087958e-01 2.09726781e-01 3.41744959e-01
-1.02895606e+00 -2.63987899e-01 -9.01850522e-01 5.73348880e-01
1.06734622e+00 -5.72870016e-01 -4.27101612e-01 8.01963329e-01
6.23942316e-01 -3.84272188e-01 -3.60668212e-01 2.19851686e-03
7.29411066e-01 -7.94624031e-01 1.09485734e+00 -1.10753644e+00
5.55429876e-01 -1.30870536e-01 7.82038942e-02 -6.05681777e-01
-2.61930496e-01 -6.20390773e-01 -4.78919029e-01 1.06751513e+00
3.11703503e-01 -1.23876601e-01 1.14766753e+00 3.16326767e-01
-2.83556227e-02 -5.89161992e-01 -9.01379704e-01 -1.05577481e+00
1.34220034e-01 -1.89321220e-01 6.20153666e-01 3.70214969e-01
-1.74063087e-01 4.52456385e-01 -3.41205269e-01 9.80798751e-02
7.19873488e-01 3.80843222e-01 1.07620478e+00 -1.56444776e+00
-3.25847328e-01 -7.72574961e-01 -5.95173955e-01 -8.66878450e-01
-1.90698624e-01 -7.45655477e-01 1.77033737e-01 -1.24185419e+00
3.46155524e-01 -2.69178420e-01 2.11257543e-02 4.76208270e-01
-3.36925507e-01 6.66887164e-01 3.32052618e-01 2.75961101e-01
-8.20939839e-01 6.65694624e-02 9.93333459e-01 -4.75125015e-01
-7.05775321e-02 -6.80276006e-02 -4.53464150e-01 9.97429729e-01
4.50380772e-01 -5.67554653e-01 2.15214714e-01 -4.80275899e-01
3.25862855e-01 -5.95273197e-01 6.11252308e-01 -1.16620135e+00
-1.33418422e-02 1.73340719e-02 6.35255218e-01 -8.63340735e-01
6.36588752e-01 -5.85863173e-01 -4.48123872e-01 4.53956336e-01
-1.55982226e-01 -2.63742089e-01 3.17099988e-01 7.70087719e-01
3.91888320e-02 -5.48037171e-01 1.08782303e+00 -3.82032126e-01
-8.21174383e-01 6.13159478e-01 -3.30033690e-01 -1.00739799e-01
1.07879519e+00 -4.03042614e-01 -4.41265106e-01 -9.06364247e-02
-7.02893794e-01 -8.02298263e-02 4.42646593e-01 7.54642546e-01
4.45036203e-01 -1.18971002e+00 -3.87851328e-01 -8.00617933e-02
5.07533610e-01 7.64222369e-02 -1.83896497e-01 1.68673947e-01
-6.80668414e-01 3.08173567e-01 -2.13511437e-02 -7.01941013e-01
-1.39757752e+00 8.48780811e-01 3.73126030e-01 -3.28157634e-01
-6.98337674e-01 1.23914814e+00 4.07230318e-01 -6.23372607e-02
4.79133755e-01 -4.06483293e-01 1.26893625e-01 3.91996861e-01
7.78714120e-01 2.14569092e-01 1.52419418e-01 -4.69104618e-01
-2.64764994e-01 6.50278151e-01 -3.95831883e-01 1.55546486e-01
1.30626690e+00 3.33112210e-01 4.55075782e-03 3.63965899e-01
8.90914559e-01 -1.41963512e-01 -1.30331159e+00 -2.68675178e-01
3.60052228e-01 -4.11359996e-01 -3.94432664e-01 -6.08639896e-01
-6.93544328e-01 1.32037997e+00 8.42467070e-01 2.18995661e-01
5.86014450e-01 4.44417894e-01 7.70681798e-01 9.11924422e-01
3.87339890e-01 -1.05855036e+00 7.59949446e-01 5.91810822e-01
6.10886455e-01 -1.56157386e+00 1.09416701e-01 -3.09971213e-01
-2.07783684e-01 1.33635557e+00 8.53977323e-01 -6.59777939e-01
5.49809635e-01 5.94299376e-01 -1.28224745e-01 -5.12293041e-01
-4.57963526e-01 -3.94924372e-01 4.99354154e-01 4.45742697e-01
1.65316135e-01 2.72818133e-02 4.56364036e-01 3.92146766e-01
-8.51640552e-02 -3.85024667e-01 3.88887554e-01 8.75288129e-01
-1.23221147e+00 -7.41226435e-01 -6.33909523e-01 8.17111433e-01
-3.75552595e-01 -1.70820281e-01 -7.62910426e-01 8.88557673e-01
2.47565925e-01 5.57495058e-01 4.19242084e-01 -9.03378353e-02
2.94766456e-01 1.93802267e-01 4.51175064e-01 -1.02869284e+00
-8.17880213e-01 1.17472544e-01 -2.39731073e-01 -5.68488598e-01
-2.12081242e-02 -5.07751346e-01 -9.39765573e-01 -2.17909276e-01
-6.80112481e-01 -5.11441648e-01 4.94409174e-01 9.05760050e-01
6.32814169e-02 3.05009753e-01 1.89128384e-01 -1.13613582e+00
-6.03892922e-01 -1.09107506e+00 -6.37740970e-01 2.94933319e-01
3.43843073e-01 -6.92026913e-01 -1.39854506e-01 2.88270324e-01] | [9.33842945098877, 1.2319231033325195] |
db4c0c1c-741c-4e59-99c8-70c50be2e9fa | weakly-supervised-action-segmentation-and | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Lu_Weakly-Supervised_Action_Segmentation_and_Alignment_via_Transcript-Aware_Union-of-Subspaces_Learning_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Lu_Weakly-Supervised_Action_Segmentation_and_Alignment_via_Transcript-Aware_Union-of-Subspaces_Learning_ICCV_2021_paper.pdf | Weakly-Supervised Action Segmentation and Alignment via Transcript-Aware Union-of-Subspaces Learning | We address the problem of learning to segment actions from weakly-annotated videos, i.e., videos accompanied by transcripts (ordered list of actions). We propose a framework in which we model actions with a union of low-dimensional subspaces, learn the subspaces using transcripts and refine video features that lend themselves to action subspaces. To do so, we design an architecture consisting of a Union-of-Subspace Network, which is an ensemble of autoencoders, each modeling a low-dimensional action subspace and can capture variations of an action within and across videos. For learning, at each iteration, we generate positive and negative soft alignment matrices using the segmentations from the previous iteration, which we use for discriminative training of our model. To regularize the learning, we introduce a constraint loss that prevents imbalanced segmentations and enforces relatively similar duration of each action across videos. To have a real-time inference, we develop a hierarchical segmentation framework that uses subset selection to find representative transcripts and hierarchically align a test video with increasingly refined representative transcripts. Our experiments on three datasets show that our method improves the state-of-the-art action segmentation and alignment, while speeding up the inference time by a factor of 4 to 13. | ['Ehsan Elhamifar', 'Zijia Lu'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['weakly-supervised-action-segmentation'] | ['computer-vision'] | [ 4.38882679e-01 8.59849826e-02 -4.74493712e-01 -5.16442537e-01
-8.07127953e-01 -7.56477118e-01 3.00305516e-01 -4.99503344e-01
-2.79679000e-01 3.02842140e-01 7.66599000e-01 1.26724958e-01
1.18966587e-01 -2.32771575e-01 -1.12547421e+00 -8.26144874e-01
-1.21734060e-01 6.21948361e-01 2.07012698e-01 3.69154483e-01
-7.60615245e-03 2.35794336e-01 -1.38198102e+00 7.45373428e-01
5.77425063e-01 9.19542253e-01 -8.59116614e-02 6.78022385e-01
2.10577726e-01 8.58934462e-01 -3.97972524e-01 -1.61516264e-01
6.78049803e-01 -8.24992180e-01 -8.91505182e-01 9.21450317e-01
7.29298770e-01 -5.41307509e-01 -5.58949113e-01 7.45184720e-01
5.66994399e-02 4.11664754e-01 7.12767601e-01 -1.22038043e+00
-3.87705714e-01 7.31686652e-01 -7.00439334e-01 -6.97905719e-02
4.05737430e-01 2.74258554e-01 1.25465870e+00 -7.01201379e-01
8.18091750e-01 1.27598596e+00 6.17062151e-01 7.38605976e-01
-1.62822807e+00 -4.57560003e-01 5.45050800e-01 1.95759311e-01
-1.18998480e+00 -5.06615400e-01 6.10251904e-01 -6.70033932e-01
9.25583124e-01 2.29308411e-01 1.06083357e+00 1.32173586e+00
8.84841606e-02 1.28746378e+00 4.94776547e-01 -4.47587036e-02
4.04354662e-01 -4.50134218e-01 -9.84004512e-02 6.19242549e-01
-2.43436471e-01 -5.24907172e-01 -6.47688627e-01 -9.15044919e-02
9.15749252e-01 1.79573059e-01 -2.90017664e-01 -8.41577411e-01
-1.37980127e+00 7.40470946e-01 2.70068329e-02 1.54383600e-01
-5.41941881e-01 1.27951920e-01 4.91288453e-01 1.19150810e-01
2.76110232e-01 2.98124284e-01 -6.25147820e-01 -2.04399347e-01
-1.15689898e+00 1.98170453e-01 7.28736103e-01 8.22192132e-01
7.71729231e-01 -9.74853635e-02 -4.54257011e-01 6.36906981e-01
5.21115810e-02 2.40904540e-01 5.00297010e-01 -1.70817113e+00
4.48461860e-01 6.69172704e-01 3.88593376e-02 -5.36385953e-01
-2.32300013e-01 -5.66224530e-02 -7.89611101e-01 -2.94493705e-01
4.71683681e-01 -6.11092895e-02 -1.15564525e+00 1.99732685e+00
3.52130443e-01 6.54205382e-01 -3.34055163e-02 9.14299786e-01
3.79854776e-02 6.95565939e-01 -8.61919597e-02 -4.55485433e-01
9.79436636e-01 -1.17343414e+00 -5.12511492e-01 -1.08827986e-01
7.47941554e-01 -3.87382865e-01 1.05132771e+00 5.16420841e-01
-1.08910918e+00 -6.34554565e-01 -7.50036299e-01 -1.79141555e-02
2.99150378e-01 2.29225799e-01 3.36658150e-01 1.27166912e-01
-1.02584887e+00 7.43039131e-01 -1.40510082e+00 -2.92301714e-01
4.50194806e-01 5.57514369e-01 -5.62000751e-01 3.17047685e-02
-7.48767138e-01 3.54844540e-01 4.43594962e-01 -1.07493341e-01
-1.17928469e+00 -5.25505424e-01 -1.13957763e+00 2.72913240e-02
5.74496686e-01 -7.36947715e-01 1.23552752e+00 -1.56496084e+00
-1.50900769e+00 6.94391966e-01 -3.25704038e-01 -5.87707400e-01
3.28200310e-01 -5.11281729e-01 -4.01947312e-02 3.96252006e-01
2.49646887e-01 9.62270379e-01 1.02736449e+00 -9.89233911e-01
-5.22438049e-01 -4.25196528e-01 1.05320022e-01 3.72239739e-01
-4.84962583e-01 -1.30450539e-02 -1.05047905e+00 -6.01370692e-01
2.02034056e-01 -1.42080307e+00 -4.23189640e-01 -1.43502802e-01
-4.70645249e-01 -1.07839674e-01 7.18691349e-01 -8.62944961e-01
1.31010377e+00 -2.23659611e+00 9.51502800e-01 1.59058854e-01
9.72286984e-02 -2.71978788e-02 -4.81044292e-01 1.82746217e-01
-1.05622515e-01 -1.10781670e-01 -4.01208103e-01 -6.00515604e-01
-3.79259624e-02 6.44611061e-01 -1.72949120e-01 5.82303822e-01
2.77695864e-01 6.77170813e-01 -7.56156981e-01 -5.61167061e-01
8.52453411e-02 3.68100703e-01 -1.00149572e+00 5.53639293e-01
-4.74980146e-01 7.51108587e-01 -4.92605418e-01 5.31131566e-01
1.95197269e-01 -3.13703299e-01 4.44023252e-01 -2.39821836e-01
9.77959037e-02 4.26630937e-02 -1.40363789e+00 2.15388322e+00
-1.95197061e-01 3.53089273e-01 4.89345416e-02 -1.25006664e+00
4.42256838e-01 3.88457477e-01 9.75973368e-01 -1.47848859e-01
4.14570160e-02 -1.60840541e-01 -1.57692656e-01 -6.62139177e-01
1.31345257e-01 1.50404319e-01 -1.17305860e-01 6.01624966e-01
4.28020090e-01 1.53276533e-01 5.61690688e-01 2.17037916e-01
1.23119318e+00 5.85715234e-01 2.18395114e-01 2.45206758e-01
4.56740439e-01 -2.31945589e-01 9.66146469e-01 4.03852940e-01
-1.63901374e-01 6.08567595e-01 7.77590215e-01 -6.59499228e-01
-1.16413653e+00 -9.45653677e-01 2.15294570e-01 1.28091729e+00
-1.32342964e-01 -6.22754395e-01 -1.05624580e+00 -9.51422751e-01
-1.16641939e-01 3.97006005e-01 -7.19119966e-01 -9.06374231e-02
-8.52019310e-01 -4.08429652e-01 2.36539736e-01 9.86516118e-01
1.80504024e-01 -8.96958709e-01 -6.13757372e-01 6.31283894e-02
-4.11168575e-01 -1.17023945e+00 -8.10077190e-01 1.75732568e-01
-9.26583469e-01 -1.12907493e+00 -6.07724607e-01 -7.73594737e-01
8.21698964e-01 3.65043171e-02 1.05440700e+00 -2.37919554e-01
4.54638228e-02 5.45708537e-01 -3.71891022e-01 1.23178512e-01
-3.38559031e-01 -1.03650525e-01 4.08941984e-01 3.39124262e-01
3.37765306e-01 -6.38711929e-01 -4.70398158e-01 3.20661664e-01
-1.01145244e+00 1.16014011e-01 6.88997984e-01 9.12881374e-01
9.37554181e-01 -2.63763636e-01 1.16696870e-02 -6.33537829e-01
-1.54834967e-02 -3.72534871e-01 -3.69414032e-01 1.74272001e-01
5.12377806e-02 1.61579683e-01 8.96891773e-01 -6.22316778e-01
-8.13223362e-01 8.25919092e-01 1.29769281e-01 -1.11278391e+00
-3.30276906e-01 1.76062867e-01 -3.28401595e-01 2.53290236e-01
4.44721043e-01 2.53411353e-01 1.42629752e-02 -3.72470200e-01
6.25312924e-01 5.31924427e-01 6.31076634e-01 -4.76795405e-01
5.97020030e-01 5.44388175e-01 -2.78912425e-01 -7.10627317e-01
-1.03863740e+00 -6.83847189e-01 -1.24938726e+00 -3.24274689e-01
1.18923759e+00 -1.10578585e+00 -4.87999856e-01 2.10311890e-01
-9.49948013e-01 -7.36377835e-01 -3.23594242e-01 5.43761194e-01
-1.08350205e+00 5.31553447e-01 -8.20286274e-01 -5.11516035e-01
2.30372492e-02 -1.13014913e+00 1.29514694e+00 1.17255021e-02
-5.64589858e-01 -6.09460890e-01 3.66342515e-01 6.21577501e-01
-2.99594671e-01 2.24328756e-01 5.99831522e-01 -7.04751074e-01
-6.58824444e-01 -1.91114144e-03 3.37313026e-01 5.87775409e-01
-3.17450060e-04 7.44255185e-02 -6.15874946e-01 -4.42891538e-01
-1.51684890e-02 -6.24009132e-01 1.07253003e+00 5.65260530e-01
1.48638391e+00 -5.44852912e-01 -3.68136793e-01 6.86864614e-01
9.52390492e-01 1.52935833e-01 4.47983384e-01 1.26822546e-01
9.83645439e-01 5.70745409e-01 7.01436877e-01 5.79803944e-01
1.63797274e-01 7.30860710e-01 2.46305764e-01 1.41949371e-01
1.85373068e-01 -2.67929196e-01 7.55906284e-01 8.15181613e-01
-1.64572045e-01 -7.31643960e-02 -6.23138130e-01 6.43539667e-01
-2.10574913e+00 -1.16409981e+00 4.41308856e-01 2.19764805e+00
8.88970912e-01 5.08447587e-02 6.53156817e-01 -1.64835881e-02
5.49094915e-01 3.39371175e-01 -7.70292819e-01 -8.75323936e-02
2.95797259e-01 2.23158002e-02 2.39120215e-01 3.44478220e-01
-1.44675708e+00 1.06806910e+00 6.48330736e+00 5.07905602e-01
-9.50064003e-01 -2.29741246e-01 6.40414715e-01 -4.09959286e-01
-1.67923793e-01 -1.32175293e-02 -4.11015689e-01 4.08218503e-01
9.26660657e-01 2.92226136e-01 5.47996819e-01 8.97181153e-01
3.64744037e-01 1.03265680e-01 -1.47918248e+00 6.78183615e-01
3.30901802e-01 -1.13277638e+00 2.66153067e-01 -2.29386427e-02
1.08037925e+00 2.52834540e-02 -2.28688478e-01 3.35102409e-01
3.80565733e-01 -8.48564208e-01 6.23666942e-01 3.44716907e-01
4.76223737e-01 -5.98098099e-01 3.51519048e-01 4.08453852e-01
-1.13052714e+00 -4.13516045e-01 -2.83042938e-01 -5.79447858e-02
2.70664126e-01 1.90862209e-01 -4.86955851e-01 2.19942689e-01
5.94175756e-01 1.12817919e+00 -2.37275854e-01 7.15270281e-01
-2.15965614e-01 7.39459157e-01 -2.13123500e-01 4.01318163e-01
4.16188121e-01 -5.18036962e-01 5.04844069e-01 1.19551146e+00
1.96795136e-01 2.44529232e-01 7.85142958e-01 5.51033795e-01
-2.03764588e-01 -2.40094494e-02 -5.51056564e-01 -8.54634121e-02
2.53981650e-01 1.13595045e+00 -6.62383676e-01 -5.72679102e-01
-6.03614986e-01 1.39826155e+00 3.24093372e-01 4.81244236e-01
-9.69295025e-01 1.85585633e-01 7.97054827e-01 -1.03139199e-01
6.48031294e-01 -1.98912963e-01 2.47822925e-02 -1.40965152e+00
2.01208130e-01 -1.20570171e+00 4.96731967e-01 -4.45443183e-01
-9.09477413e-01 2.36272797e-01 -6.16788454e-02 -1.43637156e+00
-5.82451761e-01 -4.00031865e-01 -4.98583436e-01 2.24971533e-01
-5.64273775e-01 -1.13415766e+00 -2.99748443e-02 6.63746238e-01
9.43824291e-01 2.88236933e-03 7.35232413e-01 7.71232098e-02
-8.91120136e-01 3.03581178e-01 2.02439167e-02 4.38018084e-01
7.03924775e-01 -1.30517554e+00 3.17621022e-01 9.49644923e-01
6.93325698e-01 5.21450579e-01 5.81732094e-01 -6.70144260e-01
-1.42625141e+00 -1.23693383e+00 4.63569820e-01 -5.65283060e-01
7.55184710e-01 -4.58992064e-01 -8.50224376e-01 1.39163089e+00
1.39895931e-01 6.15456626e-02 9.05251145e-01 1.22273304e-01
-4.18262064e-01 -6.97905198e-02 -6.69700325e-01 6.61157072e-01
1.30988121e+00 -5.87708235e-01 -6.49575889e-01 3.70660126e-01
6.99395537e-01 -5.07091284e-01 -8.47826958e-01 2.90275604e-01
7.77932644e-01 -1.03429151e+00 7.91708469e-01 -1.00779259e+00
5.99942565e-01 -3.43661815e-01 -1.27404511e-01 -1.27764964e+00
-4.84593213e-01 -5.94959557e-01 -4.77465689e-01 9.08372164e-01
2.08462030e-01 1.86722085e-01 1.16035473e+00 6.40680671e-01
-2.64193982e-01 -9.52059686e-01 -8.08484137e-01 -7.78477669e-01
-2.33614519e-01 -3.37076157e-01 1.26283884e-01 8.10962021e-01
1.99524537e-01 3.50022554e-01 -5.86383224e-01 7.73900300e-02
3.99323851e-01 3.04013193e-01 9.46877301e-01 -8.15110981e-01
-7.03135073e-01 -2.01621771e-01 -5.08096933e-01 -1.47953856e+00
6.49692714e-01 -5.15209675e-01 2.43688941e-01 -1.30247879e+00
6.88672662e-01 3.60296786e-01 -2.49246687e-01 7.21914470e-01
-1.34646773e-01 3.09096634e-01 2.26718694e-01 2.36163512e-01
-1.16390133e+00 6.39333487e-01 1.04412127e+00 -2.27696568e-01
-4.34339136e-01 -7.45549947e-02 -4.15238649e-01 1.00315964e+00
4.71873403e-01 -3.46214861e-01 -4.69594538e-01 -5.20911455e-01
-2.34710023e-01 1.36635870e-01 5.84351160e-02 -1.15266001e+00
7.04777893e-03 -2.85716832e-01 7.03319311e-01 -5.03035963e-01
3.99453342e-01 -7.02549696e-01 8.90624672e-02 3.14006895e-01
-6.22628987e-01 -2.31696978e-01 -6.57967553e-02 6.99647367e-01
-3.12174261e-01 2.22000569e-01 6.92609608e-01 -1.33633822e-01
-6.59456313e-01 4.50940222e-01 -4.34902400e-01 5.78213111e-02
1.18627226e+00 -3.31594765e-01 3.56523335e-01 -4.77344960e-01
-9.23358142e-01 4.75112796e-01 6.88655257e-01 3.75361830e-01
3.05408001e-01 -1.40147173e+00 -4.94793206e-01 1.19273059e-01
4.55495976e-02 1.98862582e-01 3.21618795e-01 9.53072369e-01
-1.35180548e-01 2.30682135e-01 -4.17531550e-01 -8.40079725e-01
-1.37376714e+00 6.19028747e-01 1.08400770e-01 -3.17625314e-01
-5.77734113e-01 8.98604691e-01 4.24026608e-01 -2.51325965e-01
4.50773507e-01 -3.57443988e-01 -2.37950996e-01 1.17629945e-01
4.75919366e-01 3.03151339e-01 -4.31902766e-01 -9.03269529e-01
-2.29485661e-01 7.38093734e-01 -5.04559763e-02 -1.13052048e-01
1.38151765e+00 -2.69639846e-02 -6.47373945e-02 5.68213820e-01
1.31902587e+00 -2.35815316e-01 -1.95640767e+00 -8.39294717e-02
-7.74408430e-02 -4.20604557e-01 -3.80417347e-01 -3.41577649e-01
-1.12423098e+00 5.48640370e-01 1.46114841e-01 -1.23005807e-01
1.29342890e+00 1.81738347e-01 9.30988193e-01 4.66348827e-01
7.63138905e-02 -1.28767228e+00 4.94440019e-01 5.16780734e-01
7.41634011e-01 -8.92842293e-01 -1.86525047e-01 -1.82501957e-01
-8.98424327e-01 1.04009330e+00 7.13061154e-01 -1.57631159e-01
1.35226533e-01 1.69266105e-01 -2.38916799e-02 -2.06236374e-02
-8.00679326e-01 -9.00778174e-02 3.52533281e-01 3.26337427e-01
2.92920142e-01 6.56719357e-02 -1.44468501e-01 5.24777889e-01
1.05419084e-01 2.10487265e-02 3.68678331e-01 7.19846189e-01
-3.69330674e-01 -1.09703302e+00 -3.37674379e-01 4.86801326e-01
-3.44007701e-01 2.89888471e-01 -6.40057504e-01 5.57845712e-01
2.53849715e-01 5.92329323e-01 4.27897900e-01 -6.63641095e-01
2.75579751e-01 3.80116224e-01 4.72698718e-01 -6.78847432e-01
-1.63666472e-01 4.83903140e-01 1.19695123e-02 -1.18343306e+00
-7.21466422e-01 -1.12039793e+00 -1.24808955e+00 1.32168829e-01
-7.95186311e-03 8.35972428e-02 -8.81620776e-03 1.09068835e+00
3.31952333e-01 3.04145008e-01 7.52861261e-01 -1.17299342e+00
-6.52360737e-01 -9.02481496e-01 -6.32996798e-01 7.41626561e-01
2.43728429e-01 -5.45339584e-01 -1.94841191e-01 7.07389534e-01] | [8.558161735534668, 0.6555895805358887] |
40e6b1ac-01a7-47b2-97b4-79a09dca1642 | lexicon-learning-for-few-shot-neural-sequence | 2106.03993 | null | https://arxiv.org/abs/2106.03993v1 | https://arxiv.org/pdf/2106.03993v1.pdf | Lexicon Learning for Few-Shot Neural Sequence Modeling | Sequence-to-sequence transduction is the core problem in language processing applications as diverse as semantic parsing, machine translation, and instruction following. The neural network models that provide the dominant solution to these problems are brittle, especially in low-resource settings: they fail to generalize correctly or systematically from small datasets. Past work has shown that many failures of systematic generalization arise from neural models' inability to disentangle lexical phenomena from syntactic ones. To address this, we augment neural decoders with a lexical translation mechanism that generalizes existing copy mechanisms to incorporate learned, decontextualized, token-level translation rules. We describe how to initialize this mechanism using a variety of lexicon learning algorithms, and show that it improves systematic generalization on a diverse set of sequence modeling tasks drawn from cognitive science, formal semantics, and machine translation. | ['Jacob Andreas', 'Ekin Akyürek'] | 2021-06-07 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [ 5.40877938e-01 2.04344094e-01 -5.13722181e-01 -4.78148848e-01
-7.47905433e-01 -7.51531422e-01 6.03532195e-01 1.54402554e-01
-5.19987702e-01 9.57078397e-01 4.67755139e-01 -1.13473248e+00
2.78362095e-01 -7.67716348e-01 -1.06259215e+00 -4.46583293e-02
3.47707361e-01 6.56543791e-01 1.86326846e-01 -5.41574121e-01
1.99943170e-01 1.13461480e-01 -1.17330515e+00 4.16220784e-01
8.52464795e-01 2.12494999e-01 3.00026387e-01 3.69548798e-01
-4.65698093e-01 7.10483849e-01 -3.45520526e-01 -3.54123205e-01
-7.32257143e-02 -6.94405496e-01 -1.02903008e+00 -4.59820807e-01
2.95413703e-01 -1.94282249e-01 -1.27751112e-01 1.05714059e+00
1.81086093e-01 1.64201185e-02 6.44217491e-01 -7.61317909e-01
-1.07605386e+00 1.06353581e+00 -1.92804039e-01 4.51321155e-01
3.00459445e-01 2.29863688e-01 1.10272598e+00 -7.40930498e-01
7.93987036e-01 1.61356330e+00 6.89170301e-01 1.07261515e+00
-1.50927794e+00 -6.22485220e-01 1.96470290e-01 -5.36864102e-02
-8.77493083e-01 -5.26274920e-01 3.90555590e-01 -3.47466707e-01
1.60999417e+00 -1.78869620e-01 2.50745773e-01 1.42332721e+00
3.56956571e-01 7.92205453e-01 1.09879768e+00 -7.18710959e-01
7.20285550e-02 -1.73036918e-01 3.31271380e-01 7.96140075e-01
6.23961508e-01 1.13525473e-01 -6.52577698e-01 -1.30148649e-01
6.21554434e-01 -4.24714416e-01 -8.46013725e-02 -2.99625657e-03
-9.90470529e-01 9.65460062e-01 -5.73122948e-02 4.08676505e-01
-4.11112383e-02 3.31200600e-01 5.36518395e-01 6.97084427e-01
3.80301774e-01 9.04643953e-01 -9.97093678e-01 -1.80362061e-01
-4.82061177e-01 2.15746209e-01 9.87380266e-01 9.39412773e-01
7.78954446e-01 2.62699068e-01 2.98126489e-01 7.25363374e-01
1.95853725e-01 4.70184773e-01 9.81246591e-01 -8.31364989e-01
4.98523712e-01 4.14533973e-01 -2.98874676e-01 -5.52227318e-01
-4.40420181e-01 -2.08032638e-01 -1.12027980e-01 -3.47283334e-01
4.39284176e-01 -2.70358622e-01 -9.80478048e-01 2.49686432e+00
-1.48654014e-01 -1.49187177e-01 2.19140634e-01 5.57860792e-01
7.00803816e-01 5.51136434e-01 7.18759716e-01 -5.66882379e-02
1.26265693e+00 -5.78016877e-01 -4.10591394e-01 -1.03553939e+00
1.10596180e+00 -5.67298055e-01 1.47608542e+00 8.69785920e-02
-1.27728879e+00 -3.39476496e-01 -1.03008115e+00 -2.91448712e-01
-4.25954282e-01 -3.38619381e-01 8.21604431e-01 5.42577922e-01
-1.22330189e+00 5.33542514e-01 -8.90068352e-01 -7.45689511e-01
2.33458966e-01 2.82298952e-01 -2.76752532e-01 -8.60562623e-02
-1.40929651e+00 1.29127371e+00 6.41620755e-01 -3.65494370e-01
-7.34830976e-01 -5.25143802e-01 -1.00337219e+00 9.80004445e-02
3.95064265e-01 -1.09861016e+00 1.67970216e+00 -1.28862250e+00
-1.46948338e+00 1.03654182e+00 -3.96234125e-01 -4.85425651e-01
-1.70395330e-01 -2.53217727e-01 -1.05445057e-01 -2.73161381e-01
1.31803572e-01 7.42628634e-01 3.80884945e-01 -9.20828938e-01
-3.27423424e-01 -5.10848105e-01 -1.69912353e-01 3.02878231e-01
-1.80758908e-01 5.32098860e-02 8.26261714e-02 -4.43143398e-01
7.37336604e-03 -7.81963825e-01 -7.01474547e-02 -7.42161930e-01
-1.25832245e-01 -2.87036568e-01 2.07165897e-01 -6.28488362e-01
1.04851913e+00 -1.81192768e+00 3.31947654e-01 -2.38568157e-01
-5.92630580e-02 2.44291127e-01 -4.06672180e-01 4.61903632e-01
-2.53567696e-02 4.28744853e-01 -2.23168060e-01 -1.55121401e-01
1.11704215e-01 5.21938205e-01 -5.31686604e-01 -5.97738177e-02
3.84581238e-01 1.31646359e+00 -9.56303060e-01 -1.42848372e-01
-2.65793800e-01 -1.41994925e-02 -7.12012947e-01 -1.21431239e-01
-8.04481328e-01 1.99039757e-01 -4.41592395e-01 3.56263906e-01
-6.73396066e-02 -3.59250724e-01 6.33743107e-01 4.44330335e-01
1.25977546e-01 1.15829206e+00 -3.92820239e-01 2.12677932e+00
-4.15860176e-01 5.82802653e-01 -1.73350707e-01 -1.01778066e+00
6.14123285e-01 2.53590196e-01 -7.72970319e-02 -7.45942354e-01
1.69989035e-01 4.50041622e-01 3.98050517e-01 -5.32064259e-01
5.06460547e-01 -6.89127922e-01 -2.18558580e-01 8.84494662e-01
2.83020765e-01 -6.73039863e-03 6.36513308e-02 1.71026900e-01
1.21471286e+00 3.98021132e-01 3.96913886e-01 -3.48000854e-01
9.34905559e-02 4.94672567e-01 6.31885946e-01 8.42221916e-01
5.30151054e-02 2.38389507e-01 4.06054497e-01 -4.13538516e-01
-1.13837147e+00 -1.07023418e+00 1.25393137e-01 1.73543465e+00
-1.61819085e-01 -2.54563898e-01 -9.84074235e-01 -4.80720252e-01
-1.15008228e-01 1.06225121e+00 -2.66541392e-01 -5.66997766e-01
-8.82395923e-01 -8.82169306e-01 9.34102535e-01 6.45079434e-01
4.96473350e-03 -1.30367506e+00 -4.72060412e-01 5.25327086e-01
-2.19647080e-01 -1.17374837e+00 -2.38833144e-01 4.26876843e-01
-1.05654228e+00 -8.76481771e-01 6.31240457e-02 -1.02259409e+00
3.87373030e-01 2.42798589e-02 1.44953954e+00 3.23186845e-01
-1.87226105e-02 1.79640070e-01 -5.65760098e-02 -4.92676139e-01
-1.19279945e+00 6.08438909e-01 2.11568877e-01 -7.86260784e-01
1.05724740e+00 -4.71534997e-01 8.13239589e-02 -3.25679220e-02
-1.07200682e+00 4.08259369e-02 6.30179763e-01 1.02577019e+00
1.73934504e-01 -5.82978845e-01 1.06027079e+00 -1.12455344e+00
1.17862082e+00 -6.80045962e-01 -4.47051555e-01 3.07862639e-01
-6.98529601e-01 5.82099080e-01 7.48034120e-01 -4.45993274e-01
-9.09752905e-01 -2.03086182e-01 -1.43414870e-01 4.04275469e-02
-4.41066176e-01 6.85189545e-01 -1.67461187e-01 2.16201991e-01
9.02032137e-01 4.44097459e-01 1.59048378e-01 -4.46228027e-01
6.61720455e-01 3.20154548e-01 4.56728697e-01 -1.17382395e+00
5.07966220e-01 -9.08147171e-02 -1.56955391e-01 -7.44543135e-01
-7.30590641e-01 -3.27403247e-02 -4.63827372e-01 5.58800399e-01
9.32595015e-01 -7.62499928e-01 -5.07596314e-01 1.30583182e-01
-1.31862295e+00 -7.87131965e-01 -1.69516981e-01 4.80593592e-01
-7.36057818e-01 2.12505162e-01 -1.11247170e+00 -1.72782570e-01
-2.82694101e-01 -1.09091151e+00 8.28801274e-01 9.41250771e-02
-6.04842424e-01 -1.30317521e+00 1.53263733e-01 4.87767756e-02
6.38113558e-01 -3.87563497e-01 1.78993189e+00 -1.11392200e+00
-5.03218472e-01 3.63746136e-01 3.17456387e-02 2.09633425e-01
7.26572201e-02 -3.45278203e-01 -7.76545167e-01 -3.34646255e-02
1.84491090e-02 -6.41492903e-01 9.19467211e-01 1.52302459e-01
7.18296349e-01 -2.47268632e-01 -3.20030123e-01 4.43616658e-01
1.20015097e+00 1.99528128e-01 2.43982941e-01 3.59242916e-01
4.53854531e-01 5.46126425e-01 -2.10722551e-01 -4.86210912e-01
6.07477844e-01 3.57251883e-01 -8.42335746e-02 1.75097793e-01
-2.77804404e-01 -6.61693275e-01 5.87109745e-01 1.06521022e+00
2.78421730e-01 -1.89220265e-01 -1.09082484e+00 4.57865506e-01
-1.80048549e+00 -8.10050130e-01 3.08890104e-01 1.95954585e+00
1.30895936e+00 3.33324850e-01 -1.50069714e-01 -4.29007292e-01
3.96569312e-01 -4.39458452e-02 -6.08005166e-01 -9.43666875e-01
-1.68715417e-01 4.84147817e-01 4.91443694e-01 7.73159325e-01
-5.52709937e-01 1.72717357e+00 7.43982506e+00 5.08802593e-01
-1.07623637e+00 2.67221242e-01 4.01280552e-01 2.13700309e-01
-7.52897263e-01 1.45484740e-02 -8.52345645e-01 1.82868298e-02
1.41636205e+00 -1.60278291e-01 7.35349476e-01 5.23503065e-01
-8.96730945e-02 1.64892182e-01 -1.57311094e+00 3.80183756e-01
1.72325656e-01 -1.38515556e+00 4.10171181e-01 -2.41760030e-01
5.87592363e-01 3.65024120e-01 3.35823186e-02 5.55158675e-01
8.29977214e-01 -1.28095865e+00 5.56365311e-01 1.30503019e-02
6.90168083e-01 -3.69879752e-01 2.14730874e-01 5.59933603e-01
-5.63396931e-01 -4.35998887e-02 -3.93291712e-01 -4.48624015e-01
1.53057143e-01 -5.69006689e-02 -1.03880739e+00 1.17711872e-02
-3.47350948e-02 4.13778424e-01 -4.49397117e-01 3.87810111e-01
-5.61184883e-01 8.88170183e-01 -1.75526172e-01 -3.37061316e-01
3.31148833e-01 5.43491766e-02 4.83539909e-01 1.32661402e+00
1.55934587e-01 1.28591359e-01 9.61588509e-03 9.85290170e-01
-3.38122904e-01 1.20896578e-01 -9.90408182e-01 -2.37830520e-01
5.92908561e-01 5.17011642e-01 -6.48916066e-01 -3.90992880e-01
-6.72150671e-01 7.47548461e-01 5.79801738e-01 6.57642663e-01
-3.58686328e-01 2.49984325e-03 8.75790119e-01 -2.63371412e-02
8.60553458e-02 -5.54281116e-01 -5.90025187e-01 -1.32824564e+00
-2.20252052e-01 -1.14950979e+00 2.02270508e-01 -7.71586061e-01
-1.21608961e+00 4.80574787e-01 1.28694192e-01 -1.87254667e-01
-9.62500334e-01 -1.03941607e+00 -3.23835939e-01 1.08680022e+00
-1.36686766e+00 -9.39678311e-01 6.00868583e-01 3.53015274e-01
8.74326587e-01 -3.09335500e-01 1.03467739e+00 4.19595018e-02
-4.00427580e-01 5.76814234e-01 -1.34456858e-01 1.66595176e-01
5.26545703e-01 -1.07431567e+00 9.89162922e-01 9.75415707e-01
2.85863638e-01 1.14455557e+00 7.16033936e-01 -6.64012671e-01
-1.58421254e+00 -8.75544012e-01 1.18552935e+00 -8.92006099e-01
8.22056890e-01 -4.88250822e-01 -1.12258518e+00 1.25335395e+00
2.88079441e-01 -4.58429128e-01 7.06227362e-01 3.86145264e-01
-7.21419513e-01 4.78840470e-01 -6.13724768e-01 8.08270752e-01
1.25440264e+00 -9.55003679e-01 -1.19066203e+00 1.36103988e-01
1.07592976e+00 -3.17457676e-01 -4.23668474e-01 1.62629798e-01
4.38314408e-01 -4.10592467e-01 6.40803576e-01 -1.42180753e+00
6.41509354e-01 1.40406072e-01 -3.34376127e-01 -1.53821409e+00
-1.98917419e-01 -6.23447657e-01 1.48560017e-01 7.11148441e-01
9.15946484e-01 -8.26168358e-01 6.36639237e-01 5.34109890e-01
-4.35276449e-01 -5.65883279e-01 -7.75788069e-01 -7.30500877e-01
8.51791739e-01 -5.76579630e-01 7.16110528e-01 1.02713501e+00
2.84078836e-01 1.09267414e+00 1.57156646e-01 -2.57406741e-01
3.50133955e-01 4.48585637e-02 4.90497738e-01 -1.21740651e+00
-4.60278839e-01 -8.12833607e-01 -1.48619294e-01 -1.28115582e+00
7.00868964e-01 -1.33794713e+00 2.89036483e-01 -1.42654729e+00
3.99276882e-01 -3.62379849e-01 -1.12036027e-01 7.25709140e-01
-2.31608674e-01 -1.28187358e-01 -9.57213044e-02 9.23015103e-02
-3.65716666e-01 2.23717958e-01 1.00115967e+00 7.29637146e-02
-2.82222182e-02 -4.88230616e-01 -1.10780430e+00 8.09034765e-01
9.15796280e-01 -6.37341797e-01 -5.57433426e-01 -1.14415145e+00
6.34284496e-01 6.18608817e-02 3.99257708e-03 -4.47297692e-01
3.43430229e-02 -3.71443748e-01 5.70200384e-02 1.53455600e-01
1.71266738e-02 -2.47361466e-01 -3.27440917e-01 5.02996325e-01
-8.07928860e-01 4.70689625e-01 5.75341642e-01 3.51459742e-01
1.00538768e-01 -2.66235352e-01 4.52044606e-01 -5.40523648e-01
-9.28748071e-01 -7.23681822e-02 -6.63761199e-01 5.62275648e-01
4.22282755e-01 -1.60979386e-02 -7.01231778e-01 -1.64369836e-01
-6.47748411e-01 -4.15051877e-02 7.22361028e-01 7.03914940e-01
3.68400216e-01 -1.07281089e+00 -5.85102677e-01 3.05620223e-01
-6.86308071e-02 -1.61796689e-01 -4.26485628e-01 5.23895741e-01
-3.84063631e-01 8.73455167e-01 -2.14307308e-01 -2.97722906e-01
-7.03967571e-01 5.71503222e-01 4.46677208e-01 -1.15559839e-01
-4.07714516e-01 7.20016003e-01 4.56067562e-01 -7.36415982e-01
-4.23377641e-02 -7.66807675e-01 1.84097156e-01 -5.16579032e-01
3.11255604e-01 -1.15251176e-01 1.47496790e-01 -5.64117968e-01
-2.49550372e-01 3.58560890e-01 -1.11004867e-01 -2.92871505e-01
1.08442163e+00 -1.65837348e-01 -6.60992712e-02 6.50227904e-01
9.71184254e-01 -2.06617281e-01 -8.64890754e-01 -5.60425937e-01
3.83340567e-01 2.69263774e-01 -3.97005558e-01 -8.36871982e-01
-3.10375541e-01 1.10301995e+00 -3.76577526e-02 5.35431728e-02
6.66575611e-01 1.06034681e-01 1.00147796e+00 7.59943783e-01
4.81735080e-01 -9.47853208e-01 -1.01450704e-01 1.15766513e+00
2.43138045e-01 -1.11363244e+00 -4.58628684e-01 -2.46543661e-02
-2.88454622e-01 1.02170587e+00 6.88911021e-01 1.39070883e-01
3.03615361e-01 3.69784057e-01 3.87878120e-02 -1.58551589e-01
-1.20886970e+00 -9.86345112e-02 -2.25796383e-02 4.17123437e-01
8.37506831e-01 2.03131020e-01 -4.76311296e-01 6.76519573e-01
-2.99411029e-01 9.91823003e-02 5.63848734e-01 9.54550028e-01
-8.27596664e-01 -1.30298364e+00 -1.06953815e-01 2.43366241e-01
-6.41778648e-01 -6.18781328e-01 -4.89675045e-01 7.34766960e-01
-2.22562447e-01 7.13518441e-01 7.95383900e-02 -3.08953434e-01
2.97074318e-02 8.32909822e-01 7.00785279e-01 -1.11440229e+00
-4.80818242e-01 -1.57042190e-01 3.14348996e-01 -4.96272862e-01
-1.49549559e-01 -6.10033214e-01 -1.57453895e+00 -2.75206923e-01
2.12006923e-02 1.67998940e-01 5.95430493e-01 1.32493210e+00
3.72172147e-01 3.53792757e-01 -9.04725417e-02 -2.52672344e-01
-8.41109395e-01 -9.46613014e-01 1.09693175e-02 4.66595024e-01
2.14601174e-01 -3.16008568e-01 -1.36830047e-01 1.15489341e-01] | [10.715076446533203, 9.037577629089355] |
e4e88cd7-7a41-4886-be93-f039a14e87dc | first-order-motion-model-for-image-animation-1 | 2003.00196 | null | https://arxiv.org/abs/2003.00196v3 | https://arxiv.org/pdf/2003.00196v3.pdf | First Order Motion Model for Image Animation | Image animation consists of generating a video sequence so that an object in a source image is animated according to the motion of a driving video. Our framework addresses this problem without using any annotation or prior information about the specific object to animate. Once trained on a set of videos depicting objects of the same category (e.g. faces, human bodies), our method can be applied to any object of this class. To achieve this, we decouple appearance and motion information using a self-supervised formulation. To support complex motions, we use a representation consisting of a set of learned keypoints along with their local affine transformations. A generator network models occlusions arising during target motions and combines the appearance extracted from the source image and the motion derived from the driving video. Our framework scores best on diverse benchmarks and on a variety of object categories. Our source code is publicly available. | ['Stéphane Lathuilière', 'Elisa Ricci', 'Aliaksandr Siarohin', 'Sergey Tulyakov', 'Nicu Sebe'] | 2020-02-29 | first-order-motion-model-for-image-animation | http://papers.nips.cc/paper/8935-first-order-motion-model-for-image-animation | http://papers.nips.cc/paper/8935-first-order-motion-model-for-image-animation.pdf | neurips-2019-12 | ['video-reconstruction', 'image-animation'] | ['computer-vision', 'computer-vision'] | [ 3.45186710e-01 1.54929250e-01 -2.29381353e-01 -2.90360481e-01
-4.63412076e-01 -7.34330654e-01 7.64283240e-01 -5.13630450e-01
-7.67244324e-02 4.46297169e-01 1.08606070e-01 2.36655727e-01
5.28639495e-01 -6.13898993e-01 -9.79058146e-01 -7.74983525e-01
8.23567063e-02 4.12851095e-01 3.78511906e-01 -3.03603876e-02
6.68728128e-02 5.39855599e-01 -1.51791739e+00 4.20485586e-01
4.02174920e-01 7.81894445e-01 1.29481390e-01 7.29805708e-01
1.51183024e-01 9.09841895e-01 -5.30049741e-01 -2.60991961e-01
4.29413944e-01 -6.51728392e-01 -7.60571361e-01 6.35930419e-01
9.28124785e-01 -4.91652846e-01 -6.33845389e-01 9.39416647e-01
-5.15279435e-02 2.54472375e-01 8.09849501e-01 -1.56078827e+00
-3.52694511e-01 1.28534675e-01 -5.59980512e-01 8.60591233e-02
4.36062187e-01 4.92079616e-01 8.25748622e-01 -8.19106460e-01
1.18626499e+00 1.21826339e+00 1.15124114e-01 7.89708853e-01
-1.40615225e+00 -5.78323007e-01 2.86666363e-01 1.14267841e-01
-1.07246864e+00 -7.33842731e-01 9.57624972e-01 -6.75968468e-01
4.65260118e-01 1.06236739e-02 7.59807587e-01 1.24792171e+00
1.68605596e-01 6.05754137e-01 6.10470593e-01 -1.29527360e-01
3.04543287e-01 -5.30099235e-02 -2.61765033e-01 7.47082651e-01
1.59630959e-03 6.43681288e-02 -3.68589789e-01 -8.36266428e-02
9.98670042e-01 -1.57189444e-01 -1.56151935e-01 -9.60168421e-01
-1.28285205e+00 6.88381612e-01 2.34066010e-01 3.07008065e-02
-3.84148419e-01 4.40374613e-01 1.48983911e-01 -7.03281164e-02
2.12277710e-01 1.61250696e-01 -2.15144619e-01 8.65164921e-02
-8.26281130e-01 4.41016078e-01 6.23630524e-01 9.47323620e-01
9.28820372e-01 2.72529870e-01 -4.84256111e-02 3.61110419e-01
2.27889180e-01 5.18716455e-01 1.97075233e-01 -1.55575132e+00
3.63021344e-01 4.47445780e-01 3.18978071e-01 -1.06213570e+00
-6.43952936e-02 -3.49314064e-02 -5.93331218e-01 5.40952146e-01
6.56911135e-01 -1.01727791e-01 -1.07516420e+00 2.05113697e+00
7.33060360e-01 4.09235090e-01 6.41122237e-02 9.83629704e-01
6.21151626e-01 8.29834998e-01 6.12150915e-02 -1.56897213e-02
1.08963084e+00 -1.18738079e+00 -5.38816512e-01 -5.10170162e-01
3.54369223e-01 -6.85839415e-01 6.80456519e-01 2.38725230e-01
-1.19545567e+00 -6.59664333e-01 -7.87998676e-01 -1.59449071e-01
7.92579576e-02 9.96344388e-02 2.24848062e-01 2.44369164e-01
-9.74710822e-01 4.76570129e-01 -9.15709436e-01 -2.34384850e-01
3.16519082e-01 2.71254897e-01 -6.42942727e-01 -4.05921154e-02
-7.69935668e-01 6.71297073e-01 2.34225690e-01 -1.34219781e-01
-1.30265355e+00 -4.61585611e-01 -1.11082768e+00 -1.88954428e-01
1.80116490e-01 -8.92441571e-01 1.20477688e+00 -1.82686889e+00
-1.44129002e+00 9.94605184e-01 -2.12769687e-01 -2.72849381e-01
7.44713366e-01 -3.16321015e-01 -1.73941478e-01 5.13788939e-01
1.74567297e-01 1.05978394e+00 1.48219240e+00 -1.40040600e+00
-6.48929775e-01 -5.27882725e-02 1.72048345e-01 2.37123579e-01
-8.69484246e-02 1.23822451e-01 -8.30701888e-01 -8.58548105e-01
-2.50771552e-01 -1.45082307e+00 -1.49127752e-01 3.58326972e-01
-3.25779498e-01 1.36055529e-01 1.31608462e+00 -6.90412819e-01
8.20849180e-01 -2.29019904e+00 6.79515660e-01 1.00307271e-01
1.17645033e-01 1.48539171e-01 -4.46687698e-01 1.47526667e-01
-2.93882519e-01 -2.16570616e-01 -3.13006431e-01 -3.14385742e-01
-3.28943431e-01 3.05811524e-01 -2.63691515e-01 7.09414601e-01
4.23251539e-01 8.16337645e-01 -1.11607492e+00 -4.86199021e-01
3.02186757e-01 5.29452384e-01 -6.86772108e-01 5.13943017e-01
-4.44204867e-01 8.66502285e-01 -3.52278262e-01 3.34084362e-01
3.81750435e-01 -3.16541761e-01 1.03722937e-01 -3.40533465e-01
2.16947243e-01 8.29359815e-02 -1.28832924e+00 1.74895418e+00
-2.44185328e-01 6.29756391e-01 2.00309411e-01 -7.63489962e-01
6.48489833e-01 1.85253009e-01 7.21809745e-01 -1.47475436e-01
1.09734453e-01 -1.14276968e-01 1.61088824e-01 -7.12218881e-01
1.23868398e-01 1.65565506e-01 1.37034640e-01 5.60307622e-01
1.08846053e-01 -7.45559335e-02 3.57694328e-01 3.46223772e-01
1.06432641e+00 6.44942403e-01 5.25346100e-02 -9.01328176e-02
3.85562271e-01 5.42090684e-02 5.99071980e-01 2.54193366e-01
-3.71692963e-02 8.35080802e-01 3.90774518e-01 -6.10538185e-01
-1.43185377e+00 -1.03977978e+00 1.67371273e-01 9.09641087e-01
1.72079369e-01 -3.74197453e-01 -7.51828015e-01 -7.22437263e-01
-6.53367117e-02 3.04788053e-01 -8.12163293e-01 -1.20092273e-01
-8.30636024e-01 -1.89233065e-01 4.95645702e-02 4.91216093e-01
2.52312064e-01 -1.06407678e+00 -7.90383637e-01 1.21113002e-01
-3.34035009e-01 -1.39379501e+00 -8.28880429e-01 -2.75779396e-01
-7.04541862e-01 -1.20967436e+00 -7.04016328e-01 -9.41450238e-01
9.41018045e-01 2.71334678e-01 1.14752209e+00 -2.85585760e-03
-2.58106172e-01 5.71481049e-01 -1.43352076e-01 4.61661145e-02
-7.53305793e-01 -1.58611715e-01 -4.54852358e-03 5.87796748e-01
-1.74418062e-01 -4.74883258e-01 -6.78943276e-01 2.63560236e-01
-1.03122258e+00 3.69335234e-01 2.56331772e-01 5.39959133e-01
3.20535570e-01 -2.36785427e-01 1.68080647e-02 -7.38960803e-01
-2.75731504e-01 -5.07240653e-01 -4.98213947e-01 9.88051854e-03
2.33613431e-01 7.35755563e-02 5.42396069e-01 -7.54895687e-01
-8.47730756e-01 8.32780302e-01 2.86525786e-01 -8.79359305e-01
-2.92244494e-01 -1.17249601e-01 -3.59323263e-01 6.24657534e-02
4.72690523e-01 4.73576672e-02 2.19227716e-01 -1.63594708e-01
5.65751195e-01 1.76005706e-01 9.00328636e-01 -5.36002755e-01
1.15053880e+00 7.64539659e-01 9.82697532e-02 -7.90238857e-01
-6.80174530e-01 -2.35888645e-01 -8.57577026e-01 -3.99418831e-01
1.05786026e+00 -9.79409397e-01 -2.81462759e-01 4.86008376e-01
-1.30080974e+00 -5.52815616e-01 -1.57118350e-01 3.05779397e-01
-7.40097463e-01 2.42545143e-01 -3.90512764e-01 -4.49140280e-01
-2.81085540e-02 -1.14909256e+00 1.30123770e+00 1.57422096e-01
-4.44774508e-01 -8.53426695e-01 -5.30699966e-03 1.81859314e-01
1.12920394e-02 6.88091457e-01 7.64420927e-01 -2.35224888e-01
-8.74524355e-01 -1.15141638e-01 1.71657428e-01 3.39223176e-01
2.84510225e-01 5.37767649e-01 -7.64376521e-01 -3.91823024e-01
-1.65834263e-01 -2.60830253e-01 5.79179287e-01 2.98219174e-01
9.08339679e-01 -5.15376449e-01 -4.01667118e-01 7.45814323e-01
1.18871117e+00 1.91437110e-01 6.23433352e-01 2.54081756e-01
1.04493582e+00 9.04131174e-01 4.48085010e-01 1.31857917e-01
2.32902616e-01 1.03053129e+00 5.18410861e-01 -9.34735090e-02
-3.20865959e-01 -2.88073331e-01 5.47451258e-01 2.94637620e-01
-4.19748500e-02 -4.38646376e-02 -8.19285393e-01 5.74652374e-01
-1.81213605e+00 -1.14876390e+00 -3.67040001e-02 2.12231445e+00
6.17960334e-01 -3.50078754e-02 4.18540686e-01 -1.93697587e-01
7.96555698e-01 1.66302755e-01 -8.29032600e-01 -1.17788222e-02
-5.27803861e-02 -7.31212944e-02 1.40930325e-01 4.60592955e-01
-1.24495578e+00 9.32566881e-01 6.30188322e+00 1.50426596e-01
-1.27886331e+00 -5.02805226e-02 7.33660400e-01 -1.07695028e-01
-1.77808106e-01 1.99538723e-01 -5.42105198e-01 4.96835291e-01
6.68309212e-01 -1.93574622e-01 3.91529024e-01 8.59253347e-01
1.67435646e-01 -3.69041860e-02 -1.41026807e+00 8.42261970e-01
2.86370099e-01 -1.38626623e+00 1.37654975e-01 -2.16232780e-02
1.00521028e+00 -2.01805755e-01 1.88473329e-01 -7.67729664e-03
3.43823522e-01 -9.38784480e-01 1.01138854e+00 3.36554468e-01
7.88680255e-01 -4.24686790e-01 5.38532846e-02 2.12262779e-01
-1.09231472e+00 8.77656490e-02 -1.59798618e-02 7.82863274e-02
2.37297401e-01 1.06102452e-02 -5.54704309e-01 1.71232015e-01
5.09802341e-01 1.00434899e+00 -4.83365536e-01 7.44491577e-01
-4.02522564e-01 3.93434227e-01 -2.00616121e-01 5.89226246e-01
2.19425663e-01 -3.55640769e-01 6.25007987e-01 1.06423628e+00
1.45685107e-01 1.11962855e-01 2.80445158e-01 8.15362871e-01
-2.52913665e-02 2.69505903e-02 -8.03068697e-01 1.22883908e-01
2.51732290e-01 1.42862093e+00 -6.19370103e-01 -5.13385773e-01
-5.36740601e-01 1.16919053e+00 1.97321162e-01 5.81936181e-01
-9.35810089e-01 7.21819028e-02 8.69702816e-01 3.88594568e-01
4.70493972e-01 -3.94384980e-01 2.73803622e-01 -1.42184091e+00
2.29771473e-02 -1.01374340e+00 1.83313161e-01 -1.02110171e+00
-8.29329193e-01 5.95995724e-01 1.18838824e-01 -1.45891726e+00
-6.35951698e-01 -4.06939805e-01 -7.69711316e-01 5.97345114e-01
-1.08999920e+00 -1.16499436e+00 -6.35789216e-01 6.23827875e-01
6.64037168e-01 -1.16930122e-03 5.21917164e-01 1.20144047e-01
-5.03880084e-01 1.03734680e-01 -1.78169459e-01 3.61058354e-01
6.69193208e-01 -9.16188180e-01 6.05992138e-01 8.72003078e-01
3.13169688e-01 2.76269764e-01 8.55616212e-01 -5.74172974e-01
-1.41925299e+00 -1.41361475e+00 3.55505913e-01 -6.96758151e-01
5.73545635e-01 -4.09456491e-01 -9.59792554e-01 9.49607313e-01
2.07777292e-01 3.94286633e-01 2.23835722e-01 -6.21429205e-01
-3.77658695e-01 3.10328268e-02 -7.41357744e-01 6.15570009e-01
1.15298057e+00 -3.57238680e-01 -3.52700442e-01 2.84909457e-01
3.33876669e-01 -7.48907506e-01 -5.76001883e-01 2.73492128e-01
5.70192575e-01 -7.19225764e-01 1.00822997e+00 -9.32633400e-01
8.36754739e-01 -5.43202162e-01 3.39749716e-02 -1.18968737e+00
-3.08315247e-01 -7.09767878e-01 -3.52316827e-01 9.41430390e-01
-2.07656529e-02 -3.15585732e-02 9.55803812e-01 7.11905420e-01
1.18968748e-01 -3.79293740e-01 -6.50005817e-01 -6.90044701e-01
-3.77748571e-02 7.93956667e-02 2.63603032e-01 8.57744157e-01
-3.88638377e-01 4.57536370e-01 -5.97338974e-01 2.70815820e-01
7.54162014e-01 2.90268362e-01 1.43498564e+00 -9.60341632e-01
-2.93503493e-01 -1.48369357e-01 -6.34427011e-01 -9.61427987e-01
5.91201723e-01 -8.44862342e-01 -2.74960976e-02 -1.27581239e+00
2.44221970e-01 -2.23808944e-01 2.25582570e-01 5.45963883e-01
-2.69067526e-01 5.27289271e-01 4.20537770e-01 4.60836321e-01
-3.92016947e-01 4.85638678e-01 1.41406190e+00 -1.22328497e-01
-1.74204409e-01 -2.53579736e-01 -2.89535701e-01 9.39088285e-01
6.45916462e-01 -5.29925525e-01 -4.20229644e-01 -4.97370124e-01
-1.87400132e-01 2.86218643e-01 6.03750706e-01 -1.03236246e+00
-1.63351670e-02 -3.99527282e-01 6.21425629e-01 -7.39978328e-02
5.40095806e-01 -8.22746515e-01 7.30155468e-01 4.97918665e-01
-3.42109978e-01 1.73596099e-01 5.18416241e-02 7.30270803e-01
-7.85378590e-02 -5.40005863e-02 1.01898205e+00 -2.11426206e-02
-7.08181143e-01 5.28030813e-01 -2.24321589e-01 4.33672741e-02
1.42694497e+00 -1.49228886e-01 -1.17684975e-01 -7.12833524e-01
-8.81253898e-01 2.78730579e-02 1.05778730e+00 7.36565173e-01
4.85259295e-01 -1.48801005e+00 -6.69089317e-01 2.96097249e-01
1.80848658e-01 1.27086267e-01 2.36499403e-02 5.61734140e-01
-5.49221337e-01 -3.11936205e-03 -5.12789905e-01 -7.94129252e-01
-1.29126310e+00 6.32167876e-01 4.99893457e-01 1.74316481e-01
-8.28875721e-01 3.89813304e-01 8.58956814e-01 1.20055728e-01
1.32705301e-01 -1.71195433e-01 -9.24942791e-02 -2.94405729e-01
6.20793521e-01 9.74611565e-02 -2.89787769e-01 -1.33956766e+00
-2.68248737e-01 7.01972902e-01 4.59205620e-02 -2.16819420e-01
1.18642235e+00 5.92210405e-02 7.07692057e-02 3.92504394e-01
1.56414068e+00 4.79134824e-03 -1.96508181e+00 -6.89007714e-02
-2.91136205e-01 -6.72999442e-01 -3.53803605e-01 -2.32345924e-01
-1.39238071e+00 4.83447552e-01 2.30404988e-01 -3.11939418e-01
1.00757599e+00 4.27045077e-02 6.60487056e-01 2.61314642e-02
3.51602107e-01 -7.61003435e-01 5.22131026e-01 2.61512369e-01
9.29539025e-01 -1.14358056e+00 -1.40523046e-01 -4.50195581e-01
-7.86046863e-01 1.13986313e+00 7.04252005e-01 -5.40991843e-01
4.35599357e-01 2.14491263e-01 1.95616141e-01 -8.76739770e-02
-7.45522797e-01 1.95978605e-03 4.08287078e-01 7.06285894e-01
2.07737625e-01 -2.13071004e-01 -4.26519243e-03 -8.50875452e-02
6.82957396e-02 -1.20354682e-01 6.09898806e-01 9.81470048e-01
-2.37539530e-01 -1.10950732e+00 -4.19584215e-01 1.84063151e-01
-2.60038376e-01 2.70346195e-01 -4.65571493e-01 6.88833714e-01
1.22411139e-01 6.70967221e-01 2.99815297e-01 -1.83443695e-01
9.53691602e-02 2.06537638e-02 6.40855074e-01 -8.66752982e-01
-2.01233253e-01 2.71463573e-01 -1.03683196e-01 -8.66219997e-01
-8.41635168e-01 -9.48391616e-01 -1.13663435e+00 -5.92747750e-03
1.78926051e-01 -6.14375807e-02 4.59336340e-01 6.28685713e-01
2.56051987e-01 2.11417571e-01 7.27206349e-01 -1.26793647e+00
-1.60261989e-01 -5.31664133e-01 -3.69696766e-01 1.02928972e+00
5.60984671e-01 -6.68774664e-01 -3.16630065e-01 8.54803741e-01] | [10.820234298706055, -0.7983798384666443] |
9cf5f293-22f1-4ebe-92c2-c28d76f7d767 | remask-a-robust-information-masking-approach | 2305.02858 | null | https://arxiv.org/abs/2305.02858v1 | https://arxiv.org/pdf/2305.02858v1.pdf | ReMask: A Robust Information-Masking Approach for Domain Counterfactual Generation | Domain shift is a big challenge in NLP, thus, many approaches resort to learning domain-invariant features to mitigate the inference phase domain shift. Such methods, however, fail to leverage the domain-specific nuances relevant to the task at hand. To avoid such drawbacks, domain counterfactual generation aims to transform a text from the source domain to a given target domain. However, due to the limited availability of data, such frequency-based methods often miss and lead to some valid and spurious domain-token associations. Hence, we employ a three-step domain obfuscation approach that involves frequency and attention norm-based masking, to mask domain-specific cues, and unmasking to regain the domain generic context. Our experiments empirically show that the counterfactual samples sourced from our masked text lead to improved domain transfer on 10 out of 12 domain sentiment classification settings, with an average of 2% accuracy improvement over the state-of-the-art for unsupervised domain adaptation (UDA). Further, our model outperforms the state-of-the-art by achieving 1.4% average accuracy improvement in the adversarial domain adaptation (ADA) setting. Moreover, our model also shows its domain adaptation efficacy on a large multi-domain intent classification dataset where it attains state-of-the-art results. We release the codes publicly at \url{https://github.com/declare-lab/remask}. | ['Soujanya Poria', 'Somak Aditya', 'Navonil Majumdar', 'Rishabh Bhardwaj', 'Pengfei Hong'] | 2023-05-04 | null | null | null | null | ['intent-classification'] | ['natural-language-processing'] | [ 4.90207225e-01 1.50146130e-02 -5.19700527e-01 -3.86140466e-01
-1.38615096e+00 -8.68396103e-01 8.46569061e-01 -3.16275544e-02
-3.59087318e-01 1.24103880e+00 3.51605296e-01 -4.20167506e-01
2.40227252e-01 -5.86905777e-01 -8.44696999e-01 -5.96989393e-01
3.62796813e-01 3.08726877e-01 -7.11131990e-02 -2.96627492e-01
-6.60711303e-02 -3.39523666e-02 -1.05909181e+00 5.90841234e-01
1.18747163e+00 9.40409839e-01 -1.58510283e-01 1.51750937e-01
-1.78394541e-01 4.56445128e-01 -7.88293779e-01 -6.54428601e-01
3.31481516e-01 -4.95728344e-01 -7.71600008e-01 -2.77565092e-01
3.66624713e-01 -3.80436748e-01 -1.80344984e-01 1.12861991e+00
6.16507590e-01 5.40210046e-02 8.73869598e-01 -1.11408901e+00
-9.96080101e-01 4.08318758e-01 -5.29501975e-01 2.46286988e-01
3.83251429e-01 1.81646645e-01 8.70198786e-01 -7.95997143e-01
6.49625540e-01 1.00801837e+00 6.06271088e-01 8.57519031e-01
-1.47327316e+00 -1.32650006e+00 2.41833270e-01 -2.25950009e-03
-1.07726264e+00 -5.68210781e-01 1.00005305e+00 -4.12859440e-01
6.93258226e-01 1.13290064e-02 -8.20216089e-02 2.00729251e+00
-9.56914276e-02 6.48787796e-01 1.52281690e+00 -3.84319901e-01
4.50901300e-01 4.56322342e-01 -6.85638562e-02 7.59292021e-02
2.80859441e-01 2.98018932e-01 -5.81339478e-01 -3.79503429e-01
2.83950180e-01 -1.40759468e-01 -2.95772523e-01 -2.05980510e-01
-9.31742311e-01 9.39712822e-01 2.64077455e-01 1.79697260e-01
-4.15608644e-01 -3.55376422e-01 4.66560453e-01 4.14228976e-01
8.40354145e-01 4.59459960e-01 -8.09463739e-01 1.33355139e-02
-8.17382276e-01 5.19765377e-01 6.77916765e-01 8.02096605e-01
6.44190133e-01 -1.62459631e-02 -2.22367108e-01 9.57520366e-01
-1.22734725e-01 5.46998084e-01 7.82964051e-01 -5.52398980e-01
7.97147512e-01 4.36562598e-01 2.32338876e-01 -6.21412337e-01
-3.99677642e-02 -4.05870110e-01 -6.85876548e-01 5.21569550e-02
6.48070037e-01 -3.76153767e-01 -8.74276996e-01 2.13254380e+00
4.31810319e-01 1.92257151e-01 3.60289454e-01 8.63181710e-01
4.83409345e-01 4.14535940e-01 4.31802690e-01 -7.30911866e-02
1.57634103e+00 -5.72142422e-01 -6.12099409e-01 -6.32712841e-01
6.76198244e-01 -7.28170753e-01 1.42143953e+00 2.69842654e-01
-5.13516784e-01 -3.93679529e-01 -1.11209905e+00 1.37177780e-01
-4.82166618e-01 -1.86225027e-01 4.12660182e-01 8.54926884e-01
-3.53927910e-01 3.32340896e-01 -4.51620460e-01 -2.30056092e-01
7.44574964e-01 2.44709980e-02 -4.50852156e-01 -3.10667038e-01
-1.78820515e+00 7.30823934e-01 3.32489997e-01 -7.06928968e-01
-8.47610235e-01 -1.26241803e+00 -7.80922651e-01 -1.87084496e-01
3.06581676e-01 -6.59952104e-01 1.27317464e+00 -1.21228373e+00
-1.39414334e+00 1.14248848e+00 -2.86598831e-01 -8.25804114e-01
7.37336218e-01 -4.50086921e-01 -7.41835237e-01 4.63127606e-02
4.67025310e-01 4.32710886e-01 1.09137142e+00 -1.13948584e+00
-6.74581707e-01 -3.22440177e-01 5.45410812e-03 1.03442118e-01
-6.18561089e-01 -5.52252568e-02 1.65174425e-01 -1.05101943e+00
-4.45273161e-01 -8.29644501e-01 8.02920684e-02 -2.88509786e-01
-2.76990294e-01 1.37691228e-02 8.41896057e-01 -8.83717477e-01
1.11367762e+00 -2.36410475e+00 -3.67338568e-01 -1.73554495e-01
-1.71764061e-01 4.89626288e-01 -4.06092126e-03 1.90150708e-01
-3.57175827e-01 1.83822066e-01 -6.67816579e-01 -3.38585466e-01
6.75608143e-02 -1.01486042e-01 -1.01141858e+00 3.94560963e-01
4.31382030e-01 6.41494930e-01 -9.06904995e-01 -2.32525051e-01
-3.53336101e-03 2.57532120e-01 -7.38712430e-01 1.37541384e-01
-3.92263114e-01 6.74220204e-01 -3.96491766e-01 5.87164462e-01
9.74392891e-01 -1.01992682e-01 3.12063366e-01 1.12419091e-01
3.44763517e-01 7.93552995e-01 -8.14345777e-01 1.75579119e+00
-5.26763022e-01 3.97674114e-01 -5.60586900e-02 -1.14007390e+00
7.94705868e-01 3.63945276e-01 2.06429750e-01 -7.23751247e-01
-1.42095655e-01 3.51333797e-01 -1.60503611e-01 -7.83958063e-02
2.64815658e-01 -6.84346080e-01 -5.58392704e-01 3.54779303e-01
2.92640179e-02 8.37298408e-02 -3.74813199e-01 4.85812947e-02
1.12044358e+00 9.18549374e-02 5.36481380e-01 -1.44810572e-01
4.81676996e-01 2.15728194e-01 8.39135289e-01 6.59657419e-01
-5.39129972e-01 5.37941098e-01 4.57484871e-01 -5.37764914e-02
-8.13782275e-01 -1.22612560e+00 -1.40032858e-01 1.22094727e+00
-9.30570066e-02 1.54368490e-01 -7.83413470e-01 -1.20067775e+00
2.59197563e-01 1.15687752e+00 -6.23039722e-01 -5.02763212e-01
-4.45108593e-01 -8.53852928e-01 8.52777064e-01 2.83762336e-01
7.56241858e-01 -8.66913021e-01 1.06191440e-02 1.23252667e-01
-4.95221853e-01 -1.25765562e+00 -5.41462302e-01 6.79835230e-02
-7.32682526e-01 -7.64611244e-01 -8.09359908e-01 -6.63334966e-01
3.90527904e-01 3.67467217e-02 1.00778759e+00 -6.69015169e-01
2.77778715e-01 -8.76941904e-02 -3.93416852e-01 -5.09464562e-01
-6.36725307e-01 3.10904026e-01 2.70105779e-01 6.41882196e-02
8.22892427e-01 -7.51486480e-01 -6.75717771e-01 2.92518109e-01
-9.96633053e-01 -3.57532024e-01 5.67040026e-01 1.07725883e+00
3.68409574e-01 -1.95147265e-02 1.19068265e+00 -1.23118317e+00
8.51725340e-01 -8.24359655e-01 -3.86292368e-01 -6.93772286e-02
-6.17556691e-01 -3.92447226e-03 1.01212740e+00 -8.08704197e-01
-1.41944969e+00 -1.08569406e-01 7.27605224e-02 -3.62809718e-01
-5.67239106e-01 2.60734171e-01 -5.31261861e-01 3.42394501e-01
1.17700756e+00 3.11854005e-01 -8.60054120e-02 -5.19011199e-01
4.20811653e-01 9.02730048e-01 4.70873028e-01 -7.13912785e-01
9.48262155e-01 6.92793727e-01 -5.60002089e-01 -5.69866002e-01
-1.08024812e+00 -4.29715037e-01 -3.29606861e-01 3.78588825e-01
6.13676548e-01 -1.25064540e+00 -1.10518590e-01 4.26606029e-01
-1.05435407e+00 -4.13251728e-01 -1.32197991e-01 3.28192443e-01
-3.57886523e-01 2.79708713e-01 -1.99297905e-01 -6.60723388e-01
-3.47662330e-01 -7.24695325e-01 7.87599206e-01 -9.42357853e-02
-5.76769233e-01 -9.96980488e-01 1.24695897e-01 6.17058635e-01
2.82140255e-01 3.39230031e-01 1.03502560e+00 -1.24985862e+00
-8.53932798e-02 -1.50455087e-01 -1.28106445e-01 6.64991558e-01
4.20465499e-01 -7.58897901e-01 -1.43285728e+00 -3.02593917e-01
2.20531896e-01 -3.53150129e-01 8.58932078e-01 1.67407140e-01
1.02311480e+00 -5.38478851e-01 -3.63281250e-01 4.91858453e-01
1.19256485e+00 1.13407932e-01 4.64617729e-01 4.43403721e-01
3.66255701e-01 5.89198768e-01 8.58848631e-01 4.17664289e-01
2.30761334e-01 5.91232359e-01 9.10802856e-02 6.85162544e-02
-1.79939359e-01 -6.06164217e-01 6.30573273e-01 1.98149949e-01
5.31703353e-01 -1.43375322e-01 -9.06597555e-01 1.03244185e+00
-1.48552644e+00 -9.94995713e-01 2.99654096e-01 2.37224102e+00
1.36671972e+00 4.27139789e-01 2.69281387e-01 -2.92767137e-02
8.38466048e-01 2.92751908e-01 -8.73486936e-01 -2.78795063e-01
-1.55559883e-01 3.87141377e-01 6.13569140e-01 3.78729850e-01
-1.35985529e+00 1.15160954e+00 5.27889729e+00 1.08181655e+00
-1.14595509e+00 3.35891545e-01 5.68206906e-01 -1.21143825e-01
-4.12008047e-01 -7.98665732e-02 -7.74187922e-01 8.88635397e-01
9.75510776e-01 -3.84274304e-01 4.60742861e-01 1.06339216e+00
-7.96684437e-03 2.64993101e-01 -1.06870019e+00 6.58429444e-01
-1.68319538e-01 -1.14187324e+00 1.00771926e-01 1.80285469e-01
8.71814728e-01 -6.41843379e-02 4.39516723e-01 7.23618865e-01
5.00111818e-01 -8.49093854e-01 6.35889947e-01 -1.51231959e-01
1.25013018e+00 -7.54984140e-01 5.66090703e-01 4.54800099e-01
-6.30998850e-01 -1.32177979e-01 -2.28070155e-01 -2.36697659e-01
-4.61822823e-02 8.83925259e-01 -1.11441576e+00 5.38054049e-01
6.32231355e-01 4.58172023e-01 -4.76917662e-02 2.58161843e-01
-4.01176453e-01 1.06444252e+00 -8.06511119e-02 1.55075505e-01
-3.24266851e-02 1.87285706e-01 7.41014957e-01 1.24598050e+00
1.26631305e-01 5.98261366e-03 -2.59543121e-01 9.87992346e-01
-5.19365251e-01 -1.57961562e-01 -9.20201242e-01 -3.52681801e-02
8.13422263e-01 7.79180229e-01 -1.10345259e-02 -3.06555629e-01
-4.44542915e-01 1.24764323e+00 2.81182885e-01 5.56242108e-01
-1.02560318e+00 -5.58266044e-01 1.33957732e+00 1.36505052e-01
3.85721236e-01 2.59691536e-01 -6.40295982e-01 -1.23673260e+00
3.08628261e-01 -1.26509678e+00 5.70595026e-01 -2.44270429e-01
-1.79962468e+00 2.66340792e-01 4.08508396e-03 -1.44165766e+00
-3.82714391e-01 -4.99735832e-01 -4.25989211e-01 1.11953855e+00
-1.70385993e+00 -1.00781703e+00 4.51409705e-02 6.07353389e-01
4.51083302e-01 -3.05219948e-01 1.06118429e+00 3.88290346e-01
-3.03356051e-01 1.13488591e+00 3.58312875e-01 3.64770710e-01
1.44907022e+00 -1.11945570e+00 6.10987723e-01 8.99812043e-01
-2.78628707e-01 8.19836199e-01 6.97043419e-01 -7.06204057e-01
-8.43656361e-01 -1.31141138e+00 9.56932127e-01 -7.25716829e-01
7.91484892e-01 -6.63429439e-01 -1.18027925e+00 7.03874767e-01
-1.60470866e-02 -6.00578338e-02 8.85828197e-01 1.06427118e-01
-9.41536427e-01 -1.92112833e-01 -1.59835589e+00 6.77928448e-01
1.08268487e+00 -7.87475586e-01 -1.01665056e+00 1.68769524e-01
1.01367033e+00 -3.00718248e-01 -6.66998208e-01 2.68320262e-01
3.67760628e-01 -8.38757277e-01 1.08636081e+00 -8.28394532e-01
7.72666752e-01 -1.03895374e-01 -3.29351872e-01 -1.53793442e+00
-1.37688622e-01 -6.49161577e-01 6.99080601e-02 1.58401871e+00
5.37131071e-01 -1.18774056e+00 7.15342760e-01 5.13323367e-01
2.08620399e-01 -2.33352050e-01 -1.13251555e+00 -1.02376354e+00
6.51828647e-01 -5.10136962e-01 8.58127177e-01 1.45191622e+00
-6.13897620e-03 3.55655193e-01 -2.68544346e-01 2.75621563e-01
6.42269552e-01 1.00489758e-01 7.63314366e-01 -9.06869769e-01
-2.68412948e-01 -2.91793615e-01 8.83634686e-02 -9.18697774e-01
5.81282556e-01 -9.79768872e-01 -2.11438999e-01 -9.93237615e-01
-8.95510837e-02 -3.84414315e-01 -4.75286394e-01 4.88751829e-01
-4.22369570e-01 6.01670295e-02 4.01795693e-02 1.49837226e-01
-2.02442572e-01 5.14937520e-01 8.93222988e-01 -1.95292652e-01
-1.39545590e-01 -1.62736401e-02 -1.37513387e+00 5.33768833e-01
1.03404129e+00 -8.45658600e-01 -3.55439097e-01 -2.10219100e-01
-2.43391976e-01 -1.64678350e-01 4.81207520e-01 -8.09415221e-01
-2.69850463e-01 -2.99619108e-01 3.89912456e-01 -8.51240754e-02
1.91876769e-01 -7.13110685e-01 -2.11099193e-01 4.08524692e-01
-4.23254341e-01 -4.65601325e-01 4.74389255e-01 8.36151302e-01
-2.87801206e-01 5.17656356e-02 8.75985563e-01 -6.49862736e-03
-6.06871724e-01 1.42880098e-03 -1.10287465e-01 6.48316264e-01
8.17856133e-01 -4.09257375e-02 -6.03895843e-01 -2.20850334e-01
-2.50328988e-01 -6.45114928e-02 5.84707379e-01 4.78809476e-01
1.79660380e-01 -1.35500252e+00 -8.29814553e-01 2.43443280e-01
3.64576548e-01 -4.54216562e-02 3.80329579e-01 5.04771590e-01
1.45252436e-01 4.05733109e-01 -6.99463636e-02 -3.00374031e-01
-9.38152015e-01 5.65596282e-01 1.98940441e-01 -4.89990443e-01
-2.52147377e-01 7.99311459e-01 7.16925621e-01 -7.06155479e-01
-1.62905931e-01 -1.54481530e-01 2.48973668e-01 2.43208725e-02
5.03639102e-01 2.59917438e-01 2.27803528e-01 -2.26601601e-01
-6.96326673e-01 1.08877204e-01 -3.16982657e-01 -2.55865455e-01
9.91975725e-01 2.00409964e-02 3.99241179e-01 6.91743642e-02
1.31665349e+00 2.39255592e-01 -1.44930804e+00 -5.29569924e-01
-1.84184648e-02 -4.67862725e-01 -2.64838666e-01 -1.21167564e+00
-4.02090967e-01 7.63996482e-01 4.31312293e-01 2.66398638e-02
1.24496353e+00 -9.30328593e-02 9.23722208e-01 -8.18373635e-03
1.82261989e-01 -1.09768438e+00 -2.43753921e-02 4.90086287e-01
8.34326506e-01 -1.49471617e+00 -1.77672595e-01 -2.34857261e-01
-9.96585727e-01 3.44417036e-01 5.73107004e-01 -1.17304236e-01
5.10507345e-01 -1.72204971e-02 3.02601248e-01 3.56159687e-01
-6.66683495e-01 3.32388841e-02 1.91761151e-01 1.06407845e+00
2.31173009e-01 2.61087805e-01 -1.35285720e-01 1.24275219e+00
-3.98432404e-01 -3.65394540e-02 7.92496875e-02 6.86500430e-01
-3.98379713e-02 -1.11347294e+00 -5.56932926e-01 3.21321905e-01
-8.93171191e-01 -2.71495879e-01 -5.35135031e-01 8.70923221e-01
-8.45618173e-02 9.45883632e-01 -2.33533204e-01 -2.73938447e-01
4.76016968e-01 5.85792422e-01 3.88695076e-02 -5.88056743e-01
-4.93052423e-01 -3.17064583e-01 1.47953346e-01 -3.53626132e-01
-1.10036887e-01 -8.23716938e-01 -1.17162859e+00 -3.35240871e-01
1.66730583e-01 -7.20567778e-02 3.99943680e-01 9.02994037e-01
7.04319596e-01 3.10955226e-01 5.98801076e-01 -3.03641766e-01
-9.42696571e-01 -1.09465444e+00 -3.48431677e-01 9.42371130e-01
5.48019409e-01 -7.01447427e-01 -6.03522003e-01 1.56855062e-01] | [10.32102108001709, 3.1466715335845947] |
b74fdb67-4fd5-45ef-9354-d8a71e870e79 | sciannotate-a-tool-for-integrating-weak | 2208.10241 | null | https://arxiv.org/abs/2208.10241v1 | https://arxiv.org/pdf/2208.10241v1.pdf | SciAnnotate: A Tool for Integrating Weak Labeling Sources for Sequence Labeling | Weak labeling is a popular weak supervision strategy for Named Entity Recognition (NER) tasks, with the goal of reducing the necessity for hand-crafted annotations. Although there are numerous remarkable annotation tools for NER labeling, the subject of integrating weak labeling sources is still unexplored. We introduce a web-based tool for text annotation called SciAnnotate, which stands for scientific annotation tool. Compared to frequently used text annotation tools, our annotation tool allows for the development of weak labels in addition to providing a manual annotation experience. Our tool provides users with multiple user-friendly interfaces for creating weak labels. SciAnnotate additionally allows users to incorporate their own language models and visualize the output of their model for evaluation. In this study, we take multi-source weak label denoising as an example, we utilized a Bertifying Conditional Hidden Markov Model to denoise the weak label generated by our tool. We also evaluate our annotation tool against the dataset provided by Mysore which contains 230 annotated materials synthesis procedures. The results shows that a 53.7% reduction in annotation time obtained AND a 1.6\% increase in recall using weak label denoising. Online demo is available at https://sciannotate.azurewebsites.net/(demo account can be found in README), but we don't host a model server with it, please check the README in supplementary material for model server usage. | ['Le Song', 'Chao Zhang', 'Yinghao Li', 'Leonard Thong', 'Haozheng Luo', 'Mengyang Liu'] | 2022-08-07 | null | null | null | null | ['text-annotation'] | ['natural-language-processing'] | [ 6.83092475e-02 3.72726113e-01 -3.24388631e-02 -2.68675864e-01
-1.06356573e+00 -8.59803915e-01 4.89780486e-01 3.60815734e-01
-7.12729514e-01 8.68328154e-01 2.25170195e-01 -3.94819379e-01
2.50083148e-01 -5.49764216e-01 -5.03078759e-01 -6.30388796e-01
8.12981725e-01 4.80016768e-01 1.98967189e-01 7.68278092e-02
1.15987942e-01 3.91063690e-01 -1.25618577e+00 3.67317915e-01
7.15728104e-01 3.83334845e-01 6.15768015e-01 7.19082773e-01
-5.28103709e-01 7.05797970e-01 -8.52261662e-01 -4.94108588e-01
-8.00674967e-03 -4.37311441e-01 -1.07630920e+00 -4.23073769e-01
1.90665469e-01 2.34143108e-01 2.59522796e-01 1.10520470e+00
8.44284117e-01 3.13383597e-03 4.48245019e-01 -7.85935223e-01
-4.08511013e-01 8.34214985e-01 -9.22017023e-02 7.53319860e-02
3.68885010e-01 1.73352540e-01 8.14016044e-01 -8.15154195e-01
9.05668497e-01 8.12879682e-01 7.26270616e-01 7.69803643e-01
-1.02548146e+00 -9.03196275e-01 -4.77830619e-01 -1.53117910e-01
-1.23661411e+00 -5.50219059e-01 3.37398559e-01 -4.98844326e-01
8.82422686e-01 3.50876421e-01 5.87694943e-02 1.21643198e+00
1.07455254e-01 4.10821438e-01 1.33160651e+00 -6.92954719e-01
2.59499371e-01 4.63914812e-01 4.25661415e-01 7.76785433e-01
1.37812898e-01 -4.92425859e-01 -6.03906035e-01 -1.20308317e-01
5.14486849e-01 -3.43324184e-01 -1.31588891e-01 2.70707339e-01
-1.07825124e+00 4.41276312e-01 -1.17821395e-01 7.66286612e-01
-2.84741759e-01 -7.99307451e-02 5.02399504e-01 1.65724739e-01
6.34585857e-01 4.38797325e-01 -7.28897512e-01 -4.35344040e-01
-9.24046040e-01 -2.87544370e-01 1.10745215e+00 1.05327201e+00
8.42005432e-01 -1.39885217e-01 -1.92264557e-01 8.62721860e-01
3.11319649e-01 2.47232437e-01 3.59558642e-01 -1.01586533e+00
1.40238523e-01 5.60874701e-01 1.50727183e-01 -3.25618088e-01
-5.63394964e-01 -4.31763023e-01 -6.21445060e-01 5.01502194e-02
6.06458843e-01 -3.27411979e-01 -8.94291282e-01 1.30826855e+00
3.71675372e-01 -7.96584040e-02 -6.33894047e-03 4.51941639e-01
1.25908387e+00 4.78751689e-01 6.21312141e-01 -1.62989601e-01
1.65658414e+00 -8.99946570e-01 -1.16467464e+00 2.36999199e-01
1.22660530e+00 -1.03217828e+00 1.09828043e+00 4.97777104e-01
-8.48341167e-01 -4.35385227e-01 -7.60716081e-01 -3.48621815e-01
-8.87323201e-01 3.63958448e-01 4.31695521e-01 7.52586663e-01
-9.48246360e-01 7.01940358e-01 -1.02559614e+00 -8.48981917e-01
1.24131165e-01 1.27091363e-01 -6.04253352e-01 2.42875248e-01
-1.21078694e+00 1.10866225e+00 5.47216356e-01 2.33525205e-02
-7.77726889e-01 -5.59029579e-01 -8.02152991e-01 -5.42449020e-02
5.25153637e-01 -3.61646473e-01 1.41901457e+00 -4.87283081e-01
-1.50263715e+00 1.06685138e+00 -1.76830620e-01 -6.40703961e-02
4.99551684e-01 -3.71691227e-01 -6.03642762e-01 -4.24914360e-02
3.83218229e-01 4.81889695e-01 7.50216022e-02 -1.05131245e+00
-4.40011889e-01 -7.20546618e-02 -2.84433365e-02 -1.45520717e-01
-3.68851095e-01 5.35159528e-01 -5.02074957e-01 -6.11494124e-01
-2.53624976e-01 -8.05458128e-01 -1.81889936e-01 -1.72081605e-01
-4.78469580e-01 -3.56463253e-01 7.41117597e-01 -1.02182245e+00
1.36819196e+00 -2.01447558e+00 -2.87140638e-01 -2.33183667e-01
-3.07319351e-02 3.14263791e-01 1.21654809e-01 5.29634833e-01
-2.43044242e-01 7.68419325e-01 -4.57988083e-01 -6.30748510e-01
1.54429272e-01 1.51874423e-01 2.20575631e-01 3.62180352e-01
-1.26848677e-02 5.06915510e-01 -1.13342285e+00 -6.00140750e-01
2.64466733e-01 6.81191802e-01 2.97976974e-02 2.02648416e-01
-1.90807253e-01 6.17098987e-01 -1.82307005e-01 6.83462858e-01
6.37339890e-01 -2.85315931e-01 3.43887419e-01 -2.34504044e-01
-5.84093869e-01 3.12233806e-01 -1.51590526e+00 1.88762403e+00
-4.30782109e-01 3.79953623e-01 2.36321464e-01 -5.21412432e-01
1.01797760e+00 7.47261047e-01 1.11882478e-01 -1.91726521e-01
1.01288050e-01 3.19087654e-01 -6.24547720e-01 -4.77850795e-01
5.55910170e-01 -1.46763697e-01 -1.07705981e-01 5.10093033e-01
4.40987289e-01 -4.02541757e-02 3.82230401e-01 2.54266769e-01
1.22532153e+00 6.66409016e-01 3.65949303e-01 -2.70289481e-01
4.77434814e-01 1.24451235e-01 7.47605264e-01 6.16735339e-01
-1.65426105e-01 4.75928336e-01 4.41751271e-01 -1.19530864e-01
-9.60321724e-01 -5.56625068e-01 -5.22122562e-01 1.13741803e+00
-5.18183887e-01 -8.78803849e-01 -8.77930939e-01 -7.98621714e-01
-6.60572648e-01 1.01977086e+00 -3.96714419e-01 3.52225512e-01
-1.82782486e-01 -6.29992425e-01 1.09179914e+00 3.12047571e-01
5.30275404e-01 -1.42622721e+00 -3.92567575e-01 -3.16002779e-02
-3.91698003e-01 -7.89101958e-01 -3.79927397e-01 7.53222406e-01
-6.63675606e-01 -8.29217851e-01 -6.84383214e-01 -7.92918503e-01
4.65238959e-01 -2.03515738e-01 9.84041095e-01 7.17731640e-02
-7.56027400e-02 2.64809102e-01 -5.09938776e-01 -5.86869061e-01
-7.32585251e-01 5.05502522e-01 -1.47360161e-01 -6.24282658e-01
6.64016187e-01 -2.38737151e-01 -2.04509526e-01 3.41232777e-01
-1.25790393e+00 1.79650933e-01 4.90506321e-01 5.97651660e-01
6.94998920e-01 -4.48008552e-02 5.27018547e-01 -1.30897045e+00
2.69529223e-01 -5.48005283e-01 -6.45393491e-01 2.19154790e-01
-6.92336798e-01 2.51540363e-01 3.49567056e-01 9.25698280e-02
-1.39627314e+00 3.57744813e-01 -7.54574895e-01 5.26857935e-03
-7.14734137e-01 7.50414252e-01 -4.54176068e-01 4.31037307e-01
6.56127810e-01 -1.12015188e-01 -2.46092096e-01 -1.14273942e+00
2.66173154e-01 1.02245617e+00 5.11673152e-01 -4.42446053e-01
4.38400358e-01 1.64338186e-01 -2.43104026e-01 -7.46182561e-01
-1.02009904e+00 -8.05377066e-01 -8.63780022e-01 -2.74526864e-01
1.22729123e+00 -8.04818928e-01 -5.76872289e-01 3.82133156e-01
-1.15176189e+00 -4.32467639e-01 -1.96191609e-01 4.41461951e-01
-2.17341438e-01 5.03694713e-01 -7.53156424e-01 -7.22738683e-01
-4.68343109e-01 -8.78537834e-01 8.65240872e-01 2.98792422e-01
-4.81833488e-01 -1.07758033e+00 2.54322648e-01 6.82960689e-01
1.41363367e-01 2.58083373e-01 5.20193338e-01 -1.11781490e+00
-6.37357682e-02 -1.49811313e-01 1.01966105e-01 4.40226138e-01
1.33386180e-02 9.37809274e-02 -1.34838712e+00 8.56531560e-02
-4.24617566e-02 -1.90203637e-01 6.32099450e-01 -5.34763075e-02
8.49311471e-01 -6.03312775e-02 -2.81943411e-01 2.61106938e-01
1.26284671e+00 -5.41754477e-02 7.01525331e-01 6.85040653e-01
8.77718925e-01 8.47839296e-01 6.42457545e-01 3.92755210e-01
1.38251707e-02 6.42599761e-01 2.23539114e-01 -2.68795669e-01
-1.05119042e-01 -3.15801293e-01 3.88318032e-01 1.05032921e+00
-1.84241101e-01 -3.15471143e-01 -1.04982150e+00 2.07185671e-01
-1.81030571e+00 -9.54152346e-01 -7.67699480e-01 2.06298089e+00
1.07936525e+00 8.49588960e-02 -1.14909090e-01 -3.72321508e-03
8.86670589e-01 -1.55424446e-01 -1.70632496e-01 -3.91048253e-01
-1.06476419e-01 3.24548453e-01 6.51379466e-01 5.66317558e-01
-1.06876707e+00 9.84628677e-01 5.50871611e+00 9.76730108e-01
-9.25349414e-01 6.33977711e-01 3.67258400e-01 1.95942074e-01
-2.45224312e-01 2.34608367e-01 -1.05382359e+00 4.30224359e-01
1.51639056e+00 7.49867111e-02 -1.64041936e-01 8.18046153e-01
6.57353222e-01 -3.63158397e-02 -8.55488658e-01 5.52647948e-01
-1.56228215e-01 -1.23062992e+00 -2.69422710e-01 1.81990951e-01
2.80567527e-01 1.62915424e-01 -6.34756625e-01 3.64169031e-01
5.88113487e-01 -8.02282095e-01 6.93816900e-01 7.78215885e-01
8.78145278e-01 -6.75041199e-01 1.14682662e+00 3.12268317e-01
-1.06312490e+00 4.96114820e-01 -2.72058219e-01 1.45901546e-01
3.75011444e-01 8.65726769e-01 -8.48786116e-01 7.80914545e-01
8.68847966e-01 5.36680460e-01 -6.91780031e-01 9.60564971e-01
-4.09275055e-01 8.83470058e-01 -2.04045638e-01 1.89268053e-01
-1.43620357e-01 -3.25035185e-01 3.87732029e-01 1.79593217e+00
2.48202190e-01 -4.25316989e-02 5.89034371e-02 5.78553796e-01
-2.65618503e-01 3.51677805e-01 -3.31840545e-01 -2.12780431e-01
3.67939889e-01 1.66985178e+00 -1.10062325e+00 -4.15311188e-01
-3.83650601e-01 1.11158824e+00 4.06695083e-02 6.63185045e-02
-7.80778885e-01 -7.38447130e-01 9.83753894e-03 6.14254475e-02
1.23988278e-01 -3.00225168e-02 -3.73392761e-01 -1.04712915e+00
-2.60930330e-01 -8.58763456e-01 7.03918397e-01 -1.08870518e+00
-9.01577473e-01 6.47815406e-01 -5.73559515e-02 -1.02529097e+00
1.64877266e-01 -5.05122840e-01 -3.17709297e-01 9.69108224e-01
-9.52327549e-01 -1.18643522e+00 -2.78063864e-01 2.79233128e-01
5.48533142e-01 2.63167500e-01 1.22420585e+00 5.69837987e-01
-8.06463718e-01 3.04351807e-01 1.07342042e-01 3.18057060e-01
1.22644210e+00 -1.55452311e+00 9.36285034e-02 7.59498239e-01
1.17245756e-01 7.50034273e-01 8.97670388e-01 -9.98581529e-01
-7.24335134e-01 -1.08845842e+00 1.50987005e+00 -9.58797157e-01
7.43192792e-01 -4.60283875e-01 -9.46085393e-01 7.55692601e-01
5.32216012e-01 -3.60690385e-01 1.19676876e+00 -1.13880588e-02
-1.36764720e-01 2.68070996e-01 -1.23818028e+00 4.54272717e-01
8.61030638e-01 -5.20942271e-01 -5.25753498e-01 5.90444446e-01
6.73022270e-01 -2.02648267e-01 -1.20910144e+00 6.88375384e-02
3.37667555e-01 -6.71939254e-01 3.08426797e-01 -1.48651987e-01
1.44999564e-01 -7.02634990e-01 1.53577700e-01 -8.77097130e-01
-2.26910412e-01 -5.70105672e-01 1.90125555e-01 1.81832516e+00
8.52895319e-01 -4.44777906e-01 6.34634793e-01 6.74392521e-01
-4.55315500e-01 -1.14876762e-01 -7.46593595e-01 -5.70102334e-01
-7.67723173e-02 -6.46665096e-01 2.63159633e-01 1.20430565e+00
1.41173899e-01 5.22404850e-01 -2.72454172e-01 1.06534831e-01
3.74883175e-01 -6.69486105e-01 4.19067532e-01 -1.36233008e+00
-2.69372463e-01 -2.02325117e-02 1.23596638e-01 -4.83339906e-01
1.09407924e-01 -1.19350553e+00 1.02147564e-01 -1.82565844e+00
2.91549563e-01 -2.32727215e-01 -3.16350222e-01 1.10228312e+00
5.65229915e-02 4.38731968e-01 1.47877380e-01 4.28276688e-01
-7.00460374e-01 5.88003211e-02 7.32180893e-01 7.10898712e-02
-1.27029255e-01 -9.46184695e-02 -7.73555696e-01 7.41981626e-01
9.19486105e-01 -1.06391895e+00 4.48328033e-02 -1.68573275e-01
3.90046537e-01 -4.34659809e-01 3.08727063e-02 -9.97320771e-01
2.22040936e-01 1.52156711e-01 2.23854348e-01 -4.52841848e-01
-5.61341606e-02 -7.26341009e-01 7.39616811e-01 4.04720694e-01
-3.35329443e-01 4.42761146e-02 2.82856643e-01 3.42873842e-01
2.98499838e-02 -9.45311427e-01 5.61474383e-01 -5.22407532e-01
-5.69717169e-01 -9.11802873e-02 -7.92364478e-01 -1.70139298e-01
7.77211845e-01 -1.41969502e-01 -4.04613942e-01 -1.03656352e-01
-1.06755316e+00 -1.66419409e-02 6.61361217e-01 6.32713586e-02
-1.66646957e-01 -8.84673238e-01 -5.24945319e-01 -4.08508480e-01
1.01687968e-01 2.96861380e-02 1.49669647e-01 7.72140443e-01
-6.30508900e-01 3.88671041e-01 -6.00412488e-02 -1.78641960e-01
-1.49097252e+00 5.09400368e-01 2.73675490e-02 -4.73852456e-01
-4.46393251e-01 5.76906621e-01 -2.18205735e-01 -8.34792733e-01
2.38260642e-01 -9.92341265e-02 -5.36203325e-01 2.38111436e-01
5.56979179e-01 7.22390532e-01 4.52962816e-01 -5.66644847e-01
-4.04648036e-01 9.95253399e-02 1.39020026e-01 -2.10930750e-01
1.51634383e+00 -1.33716092e-01 -2.01820552e-01 8.25258255e-01
8.39711130e-01 3.78191024e-01 -8.22651505e-01 1.40666872e-01
4.96174067e-01 1.49791777e-01 3.17681767e-02 -1.18527424e+00
-6.60085082e-01 5.07902086e-01 7.33983755e-01 1.54491171e-01
9.37097609e-01 1.64245754e-01 3.02989691e-01 5.14423072e-01
2.13168994e-01 -1.26015973e+00 -3.93085390e-01 6.41781628e-01
6.47014320e-01 -1.12031555e+00 -7.41945356e-02 -5.35685301e-01
-6.60963655e-01 1.01868868e+00 4.83096570e-01 6.44166946e-01
5.31772852e-01 5.06229579e-01 4.98708189e-01 -3.06930393e-01
-5.43538213e-01 -2.21354067e-01 -2.59397805e-01 6.49547756e-01
1.05163729e+00 -5.88642247e-02 -7.95782745e-01 8.01204562e-01
-6.38940334e-02 2.17806756e-01 7.88836658e-01 1.20270109e+00
-3.58108938e-01 -1.55047393e+00 -5.01603484e-01 2.70631075e-01
-9.54506040e-01 -3.81222397e-01 -3.77341062e-01 3.30032855e-01
2.56527096e-01 1.11416566e+00 -2.79424965e-01 -8.17401558e-02
2.82906562e-01 6.08134389e-01 -1.72545582e-01 -1.06063998e+00
-1.03951025e+00 1.84550986e-01 6.52494073e-01 -2.07622528e-01
-6.18671656e-01 -5.05146444e-01 -1.33125854e+00 -2.31063530e-01
-4.71266598e-01 5.92328072e-01 9.39086378e-01 7.96475589e-01
3.89560431e-01 4.39182162e-01 -1.66600287e-01 -6.86679780e-01
-8.42866153e-02 -1.26839721e+00 -5.61820686e-01 2.20800608e-01
-3.29198271e-01 -3.66317987e-01 -5.51258445e-01 5.85517287e-01] | [9.648259162902832, 9.286924362182617] |
e7d750b9-ec48-4528-b387-3335ed5df792 | foreground-background-segmentation-based-on | 1410.6472 | null | http://arxiv.org/abs/1410.6472v1 | http://arxiv.org/pdf/1410.6472v1.pdf | Foreground-Background Segmentation Based on Codebook and Edge Detector | Background modeling techniques are used for moving object detection in video.
Many algorithms exist in the field of object detection with different purposes.
In this paper, we propose an improvement of moving object detection based on
codebook segmentation. We associate the original codebook algorithm with an
edge detection algorithm. Our goal is to prove the efficiency of using an edge
detection algorithm with a background modeling algorithm. Throughout our study,
we compared the quality of the moving object detection when codebook
segmentation algorithm is associated with some standard edge detectors. In each
case, we use frame-based metrics for the evaluation of the detection. The
different results are presented and analyzed. | ['Eugène C. Ezin', 'Mikaël A. Mousse', 'Cina Motamed'] | 2014-10-23 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 1.85575828e-01 -5.45605123e-01 7.65593126e-02 3.35914828e-02
-1.48352146e-01 -3.91243219e-01 3.77637208e-01 3.22050929e-01
-6.76078260e-01 3.37313205e-01 -1.73718676e-01 -2.79388309e-01
1.33037627e-01 -6.63207829e-01 -2.27829069e-01 -6.81286275e-01
-2.25046754e-01 7.18599930e-02 1.00220716e+00 5.45531884e-02
4.68960017e-01 7.66564965e-01 -1.46327853e+00 9.32584777e-02
4.73126799e-01 8.01729918e-01 3.96369368e-01 1.34711266e+00
-2.91685939e-01 7.08912075e-01 -6.64998651e-01 -5.93031794e-02
5.55836618e-01 -7.46932626e-01 -5.06731749e-01 6.17072642e-01
2.13738501e-01 -2.35369012e-01 -1.92750335e-01 1.24656522e+00
3.21059257e-01 2.70046026e-01 7.59832025e-01 -1.35427833e+00
1.83576256e-01 3.21270317e-01 -6.57607317e-01 9.66899574e-01
3.43260050e-01 -2.43078217e-01 3.05026889e-01 -8.70878637e-01
5.89876831e-01 8.92266095e-01 3.79288852e-01 4.38593477e-01
-8.73427987e-01 -3.70375454e-01 6.32817820e-02 6.20801449e-01
-1.86487460e+00 -4.86814052e-01 6.83415294e-01 -6.61248684e-01
4.53403294e-01 4.61229682e-01 6.98636591e-01 1.57218069e-01
1.64267153e-01 8.53781521e-01 7.91574001e-01 -8.52947056e-01
3.16102773e-01 2.16488928e-01 5.32710314e-01 7.17453778e-01
6.51177883e-01 -2.64377464e-02 5.06757498e-02 -4.83975708e-02
7.07383811e-01 -1.01331901e-02 -4.01375800e-01 -3.53203088e-01
-7.38472223e-01 6.28379822e-01 -1.35586396e-01 7.73981929e-01
-1.42992318e-01 1.38466522e-01 2.96186775e-01 -2.48980746e-02
3.34223211e-01 -2.80296266e-01 6.50910214e-02 -5.81386164e-02
-1.45847154e+00 6.18095323e-02 8.21523190e-01 1.23197746e+00
3.63024056e-01 -4.81285097e-04 -1.78489789e-01 9.80489552e-02
3.58733654e-01 3.26336175e-01 3.55023474e-01 -7.62203753e-01
8.72875936e-03 4.06366467e-01 3.48199099e-01 -1.25834394e+00
-4.17169780e-01 -2.12064296e-01 -4.41755861e-01 4.88049001e-01
4.99931425e-01 -7.44011551e-02 -6.33758366e-01 1.08549166e+00
5.85074246e-01 5.96872985e-01 1.20243929e-01 5.45499384e-01
6.13270044e-01 6.27667367e-01 6.26477078e-02 -6.90873921e-01
1.27977073e+00 -1.03442168e+00 -1.07837462e+00 4.16230053e-01
5.09948075e-01 -1.07575023e+00 2.73135364e-01 4.85665709e-01
-1.04257691e+00 -7.53370345e-01 -1.04556513e+00 3.68837297e-01
-2.81580836e-01 2.69688368e-01 1.95969269e-01 1.15119469e+00
-1.07053077e+00 3.59964222e-01 -1.00580907e+00 -6.86093032e-01
9.48105454e-02 3.28069419e-01 3.90035771e-02 3.12592804e-01
-5.56013644e-01 8.74405265e-01 7.69358158e-01 -4.99477163e-02
-8.04737687e-01 1.61951169e-01 -3.71736318e-01 -7.64368922e-02
3.38938534e-01 -4.79908258e-01 1.18221366e+00 -1.13521123e+00
-1.15694988e+00 7.60052979e-01 -3.09832960e-01 -5.72576225e-01
9.12291527e-01 -3.34568545e-02 -7.00034559e-01 4.35853451e-01
-4.47940975e-02 2.64493853e-01 8.95916224e-01 -1.19135082e+00
-1.34127414e+00 -2.39985585e-02 -2.20951275e-03 -1.82543918e-02
-9.33301970e-02 4.79946822e-01 -8.37691307e-01 -6.61753297e-01
1.03644945e-01 -5.33991933e-01 -2.15087831e-01 -1.35381773e-01
-1.07436530e-01 -1.48693085e-01 1.21098983e+00 -5.25601089e-01
1.61591768e+00 -2.39700985e+00 -1.81444705e-01 2.20368683e-01
3.14853579e-01 2.89706826e-01 2.56210893e-01 8.39876980e-02
1.08944841e-01 7.15937018e-02 -2.43938223e-01 -3.51637155e-01
-5.16901612e-01 9.81802866e-03 9.74021107e-02 7.34663725e-01
-1.02375954e-01 2.05782950e-01 -7.07359374e-01 -1.17940056e+00
4.46961880e-01 3.40845764e-01 -3.97871852e-01 9.78820026e-02
2.54473358e-01 1.43417358e-01 -3.65138054e-01 6.51325703e-01
8.37994695e-01 2.36196995e-01 -1.49422036e-02 -1.90628707e-01
-6.05847359e-01 -4.95200127e-01 -1.79463196e+00 1.18331397e+00
7.40699992e-02 9.04048324e-01 3.43544126e-01 -9.93002415e-01
6.40313327e-01 5.07659256e-01 5.58769047e-01 -7.44597316e-02
5.35443485e-01 1.98965743e-01 2.15415031e-01 -5.27469873e-01
7.34376371e-01 4.77824407e-03 5.22344828e-01 -8.84254463e-03
-1.18911915e-01 2.49583766e-01 8.13753903e-01 6.49151281e-02
1.05158043e+00 -8.55998397e-02 5.62134624e-01 -3.97836864e-01
9.18520510e-01 2.41340205e-01 4.09581184e-01 9.00876939e-01
-6.07772946e-01 4.04913217e-01 -3.00069395e-02 -1.47458181e-01
-8.83310199e-01 -9.12703514e-01 -4.10166830e-01 8.17121744e-01
6.16845608e-01 -4.23736572e-01 -1.08174574e+00 -4.83014017e-01
-4.79120493e-01 3.75826299e-01 -4.44330394e-01 1.81897923e-01
-7.21695721e-01 -9.15855110e-01 2.78861403e-01 5.11693478e-01
6.12221777e-01 -6.34437382e-01 -1.13736594e+00 5.30914187e-01
-4.83898958e-03 -1.21526527e+00 -4.42582995e-01 -5.52861542e-02
-1.11939812e+00 -1.16798735e+00 -7.59896815e-01 -8.63058090e-01
8.12666774e-01 7.52533436e-01 8.33495617e-01 4.02159482e-01
-3.92770678e-01 7.77819693e-01 -6.67121649e-01 -4.11001563e-01
-6.77309334e-01 -3.82130265e-01 1.16006047e-01 2.78286964e-01
4.83746290e-01 3.47950938e-03 -5.10530889e-01 4.48893696e-01
-1.12698996e+00 -7.98312053e-02 3.56703103e-01 1.67008147e-01
2.73926824e-01 4.92116690e-01 -1.80035308e-01 -4.81052846e-01
2.46422037e-01 -2.30462775e-01 -1.05749583e+00 1.85717106e-01
-3.06154281e-01 -3.07171047e-01 1.79680929e-01 -3.92286956e-01
-8.67031515e-01 2.89189965e-01 4.69202287e-02 -2.91645020e-01
-2.79479116e-01 -9.96130258e-02 -3.42662305e-01 -3.97912621e-01
3.88389528e-01 1.81391686e-01 -4.78099465e-01 -4.37842786e-01
2.68828928e-01 6.51000977e-01 5.11140645e-01 -9.23597589e-02
6.55928075e-01 6.90617800e-01 5.50804585e-02 -1.19278967e+00
-1.56360134e-01 -1.10389090e+00 -7.11263955e-01 -6.53450310e-01
1.11486471e+00 -4.71635789e-01 -3.52858692e-01 5.24736047e-01
-1.35071361e+00 2.10618675e-02 -1.45469114e-01 6.53556883e-01
-5.61712623e-01 7.50183403e-01 -2.39568785e-01 -1.29443443e+00
-4.63035367e-02 -1.12683427e+00 6.75485313e-01 3.57674509e-01
1.47934690e-01 -1.04892600e+00 1.66930661e-01 -1.52839795e-01
2.06184775e-01 1.14556484e-01 2.50556797e-01 -7.43301988e-01
-7.91891932e-01 -4.47441399e-01 -5.97429127e-02 2.23086402e-01
2.28010282e-01 4.03170228e-01 -6.94376707e-01 -1.09659858e-01
4.32993740e-01 9.41338122e-01 8.04792047e-01 7.55726933e-01
9.48579371e-01 1.63230710e-02 -7.66095281e-01 3.62448961e-01
1.71273088e+00 9.37734663e-01 7.54466116e-01 4.90366459e-01
4.30550158e-01 3.19488823e-01 7.59208858e-01 3.09109300e-01
-2.81681597e-01 8.29249144e-01 -9.17416532e-03 -8.41232315e-02
3.19687426e-02 4.09708738e-01 3.25550407e-01 6.79343879e-01
-2.86356688e-01 -5.62214255e-01 -1.07919264e+00 4.77498740e-01
-1.81079423e+00 -1.16063380e+00 -6.07013166e-01 2.14791656e+00
2.94906735e-01 4.53233808e-01 5.08024395e-01 5.12934029e-01
1.27689898e+00 -3.35902244e-01 1.54968664e-01 -1.29134834e-01
4.94800461e-03 -1.07590027e-01 6.02724850e-01 7.29118168e-01
-1.48806059e+00 7.34476328e-01 7.25079060e+00 8.70849252e-01
-8.76726568e-01 3.34061146e-01 2.59358108e-01 2.54839718e-01
4.76561517e-01 6.69148862e-02 -1.11687016e+00 5.01789927e-01
6.71134710e-01 -4.52927709e-01 -1.78122088e-01 8.36020350e-01
4.17931706e-01 -5.70281982e-01 -9.04163718e-01 1.00662792e+00
2.97327369e-01 -1.05624783e+00 -5.19480184e-03 -2.18003228e-01
5.25709867e-01 -5.91403544e-01 -4.28206742e-01 -1.23110838e-01
-2.93052822e-01 -5.31378031e-01 8.52059960e-01 5.66741467e-01
5.66246621e-02 -6.51043355e-01 7.53376126e-01 2.88928151e-01
-1.62668109e+00 9.15552080e-02 -2.03581363e-01 -5.24153747e-02
3.71154577e-01 2.13368833e-01 -8.93595219e-01 4.12590325e-01
3.44560355e-01 1.46959409e-01 -5.72121263e-01 1.93535841e+00
4.02806222e-01 6.48454130e-01 -3.63470227e-01 -1.90924421e-01
8.30423683e-02 -3.93230200e-01 9.90399182e-01 1.69477403e+00
2.03223124e-01 2.36055627e-01 5.07936478e-01 6.61356330e-01
3.83156270e-01 6.05559230e-01 -5.31552613e-01 1.54151261e-01
3.00467759e-01 1.17933381e+00 -1.67234206e+00 -9.16238904e-01
-6.45903349e-01 9.45615590e-01 -5.49983263e-01 8.03723410e-02
-9.75490332e-01 -4.72065687e-01 1.69366553e-01 3.05206060e-01
3.57786626e-01 -5.54875970e-01 7.91267827e-02 -8.80557954e-01
-2.89049357e-01 -2.53215045e-01 3.43059093e-01 -5.09707212e-01
-5.66263497e-01 5.39932966e-01 4.67558831e-01 -1.41157877e+00
2.23984346e-01 -7.25610733e-01 -8.14312756e-01 4.48728144e-01
-9.51976955e-01 -5.45665562e-01 -4.12104189e-01 5.31477571e-01
8.81833673e-01 -1.38687193e-01 2.26332590e-01 6.22855723e-01
-5.84270597e-01 1.55218691e-01 1.41514719e-01 5.06988347e-01
2.08198309e-01 -1.12902367e+00 -1.29299769e-02 1.58759844e+00
3.49533826e-01 3.73933405e-01 1.09635532e+00 -7.64242053e-01
-8.36780608e-01 -9.68807459e-01 4.88836139e-01 -3.47976983e-01
4.73308474e-01 -4.83648516e-02 -7.72812128e-01 5.57598412e-01
1.98346555e-01 4.02040854e-02 5.82765520e-01 -6.73272192e-01
5.89305282e-01 1.15696549e-01 -1.12790155e+00 5.60920298e-01
8.37937593e-01 3.44899707e-02 -7.50467777e-01 1.60539880e-01
3.30143988e-01 -2.34706789e-01 -3.16098690e-01 3.66651297e-01
3.28084677e-01 -8.25490355e-01 7.91314185e-01 -4.26221013e-01
-4.04327810e-01 -9.41565037e-01 -2.93837577e-01 -4.94140804e-01
-3.90824944e-01 -4.73914415e-01 -1.64366364e-01 1.13359547e+00
-2.45663356e-02 -2.12981746e-01 7.89475560e-01 2.11747199e-01
1.61939055e-01 -4.74295840e-02 -9.57786202e-01 -1.04578102e+00
-3.85936707e-01 -4.61330086e-01 -1.69792905e-01 5.78679919e-01
-2.46471107e-01 8.17839727e-02 -1.29530802e-01 3.71596187e-01
7.87537992e-01 -3.65660079e-02 6.32263184e-01 -1.21876323e+00
-1.94864407e-01 -6.99792624e-01 -1.13462698e+00 -8.50280762e-01
-3.27308655e-01 -4.43466544e-01 1.36443108e-01 -1.42583811e+00
3.34210902e-01 3.48564312e-02 -3.14489007e-01 -1.72119483e-01
-2.79879749e-01 2.45693788e-01 4.22694713e-01 1.02215573e-01
-8.42554688e-01 -1.54749313e-02 6.68353379e-01 -7.86006749e-02
-2.89625257e-01 1.87798336e-01 6.26843795e-02 1.01280212e+00
8.94997358e-01 -6.98190808e-01 -7.24145696e-02 3.43507156e-02
-2.94198662e-01 -5.35016246e-02 3.39440137e-01 -1.53355038e+00
4.29838806e-01 -3.30671333e-02 3.06670547e-01 -8.72791827e-01
-8.92988071e-02 -1.17419159e+00 2.66645938e-01 7.99600601e-01
1.36710823e-01 2.31809214e-01 2.40642667e-01 6.43975735e-01
-1.29471794e-01 -9.28761959e-01 9.54139173e-01 -3.90992351e-02
-1.18745160e+00 -3.72010507e-02 -1.07691848e+00 -1.97846979e-01
1.51279426e+00 -7.20096767e-01 2.43794143e-01 -7.53935874e-02
-1.10002673e+00 -1.04641385e-01 5.13233483e-01 7.93436170e-02
5.47630548e-01 -1.10773218e+00 -5.47019184e-01 -9.00314972e-02
1.02512250e-02 -4.30673867e-01 -1.46510154e-01 1.07749426e+00
-1.00064576e+00 2.30390355e-01 -2.55491585e-01 -6.80584013e-01
-1.82404590e+00 9.80281234e-01 5.06775916e-01 2.52581537e-01
-4.95629132e-01 4.87942010e-01 9.08220485e-02 7.34533072e-01
2.64402211e-01 -4.66784716e-01 -4.86848533e-01 -4.60074246e-02
7.29243159e-01 9.41022992e-01 -1.48584008e-01 -6.96725965e-01
-4.61699963e-01 6.84860110e-01 2.36732095e-01 -3.59628409e-01
4.78918701e-01 -3.42577934e-01 7.13431239e-02 6.62440896e-01
8.17777097e-01 4.54722017e-01 -8.59208465e-01 1.32799909e-01
4.17739511e-01 -5.89210153e-01 1.45636648e-01 -1.13656633e-01
-9.23648119e-01 5.90446293e-01 1.35312617e+00 6.42552435e-01
1.42993557e+00 -2.16548413e-01 2.12918058e-01 1.63597256e-01
5.71927965e-01 -1.08646357e+00 -2.29820013e-01 1.20340690e-01
4.01184440e-01 -1.04523647e+00 5.26467245e-03 -7.22693264e-01
-1.63156688e-01 1.31931448e+00 4.24354464e-01 -1.16711512e-01
8.67683470e-01 5.79627097e-01 1.74188241e-02 -9.50639695e-03
-1.97824612e-01 -7.99524307e-01 1.33868441e-01 5.45798182e-01
2.91084886e-01 -8.69125724e-02 -9.07325923e-01 1.40332058e-01
4.40342426e-01 1.33871064e-01 7.26551175e-01 1.21197677e+00
-9.95700061e-01 -1.09307206e+00 -8.56310070e-01 -3.38747464e-02
-5.84181011e-01 2.17385441e-02 -1.94308445e-01 9.48536873e-01
4.22405928e-01 1.32167828e+00 1.93169922e-01 -1.73968196e-01
1.32612646e-01 3.04652341e-02 5.50807893e-01 -3.97839844e-01
-1.47439912e-01 3.80569458e-01 -1.89429998e-01 -1.97158322e-01
-9.55244780e-01 -7.38129020e-01 -1.31959295e+00 -4.63371798e-02
-7.56479979e-01 4.24164563e-01 6.42609060e-01 7.00578868e-01
-1.63933679e-01 5.58466792e-01 4.08644289e-01 -5.78883708e-01
9.29590464e-02 -7.90635705e-01 -7.69653499e-01 1.51090458e-01
2.08564982e-01 -8.00328434e-01 -3.42864156e-01 5.17252982e-01] | [8.9120512008667, -0.8833074569702148] |
24e63234-2dfb-49d3-9445-753d326123ac | data-free-backbone-fine-tuning-for-pruned | 2306.12881 | null | https://arxiv.org/abs/2306.12881v1 | https://arxiv.org/pdf/2306.12881v1.pdf | Data-Free Backbone Fine-Tuning for Pruned Neural Networks | Model compression techniques reduce the computational load and memory consumption of deep neural networks. After the compression operation, e.g. parameter pruning, the model is normally fine-tuned on the original training dataset to recover from the performance drop caused by compression. However, the training data is not always available due to privacy issues or other factors. In this work, we present a data-free fine-tuning approach for pruning the backbone of deep neural networks. In particular, the pruned network backbone is trained with synthetically generated images, and our proposed intermediate supervision to mimic the unpruned backbone's output feature map. Afterwards, the pruned backbone can be combined with the original network head to make predictions. We generate synthetic images by back-propagating gradients to noise images while relying on L1-pruning for the backbone pruning. In our experiments, we show that our approach is task-independent due to pruning only the backbone. By evaluating our approach on 2D human pose estimation, object detection, and image classification, we demonstrate promising performance compared to the unpruned model. Our code is available at https://github.com/holzbock/dfbf. | ['Vasileios Belagiannis', 'Klaus Dietmayer', 'Achyut Hegde', 'Adrian Holzbock'] | 2023-06-22 | null | null | null | null | ['pose-estimation', '2d-human-pose-estimation', 'model-compression'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 4.83396024e-01 4.22911525e-01 -1.71129480e-01 -3.47073615e-01
-4.38641906e-01 -3.03209782e-01 3.80732045e-02 -3.44097614e-02
-7.81401873e-01 6.64331317e-01 -1.70235738e-01 -1.37469828e-01
2.07618073e-01 -7.48280108e-01 -1.30139339e+00 -5.74707687e-01
6.60593137e-02 5.46819031e-01 2.07028806e-01 3.55209827e-01
-1.25459880e-02 6.09864533e-01 -1.59672987e+00 4.44687843e-01
4.53243673e-01 1.21223521e+00 4.63833392e-01 5.67770898e-01
3.71600658e-01 5.90568542e-01 -7.47349918e-01 -5.36769986e-01
6.57625794e-01 -1.47319824e-01 -7.28070915e-01 7.23915622e-02
6.34776831e-01 -7.97571182e-01 -4.95487988e-01 1.14247823e+00
5.18149018e-01 9.36282799e-02 1.83038056e-01 -1.27534306e+00
-1.02455383e-02 5.55720925e-01 -4.45201129e-01 -1.29220188e-01
-2.71837234e-01 7.30741024e-02 6.14325762e-01 -8.71929526e-01
7.84820437e-01 1.06785464e+00 6.18113339e-01 8.68526757e-01
-1.38949311e+00 -8.87370229e-01 1.33351162e-01 5.06027304e-02
-1.43718922e+00 -6.48336768e-01 7.71985233e-01 -1.60679117e-01
6.84996665e-01 1.72276333e-01 8.38153720e-01 1.25563896e+00
-1.57136858e-01 9.13416028e-01 6.96086586e-01 -4.55221176e-01
3.37822735e-01 1.85516134e-01 3.39416899e-02 8.79299223e-01
3.86199772e-01 1.87126800e-01 -4.38287407e-01 -1.35963693e-01
7.74465322e-01 2.79320050e-02 -4.02746707e-01 -6.15343690e-01
-8.01570296e-01 6.53333664e-01 7.77279437e-01 -2.95245290e-01
-3.82361144e-01 3.53746295e-01 4.50135648e-01 2.21130267e-01
2.70388931e-01 1.72039226e-01 -5.89547515e-01 9.73133892e-02
-1.14834297e+00 4.67001945e-01 8.57205987e-01 1.02572179e+00
6.78273261e-01 -2.24769428e-01 -1.19114956e-02 1.02129865e+00
4.04620022e-02 5.34439869e-02 5.48096478e-01 -1.20454359e+00
4.91175026e-01 3.58872473e-01 -1.79556787e-01 -9.23226535e-01
-2.55547434e-01 -6.13274395e-01 -9.13908124e-01 2.64470756e-01
4.78521526e-01 -2.08604336e-01 -9.93200123e-01 1.93339813e+00
3.62481028e-01 9.64873508e-02 -1.08111538e-01 8.26017261e-01
6.87935412e-01 3.23187739e-01 -1.90160915e-01 1.29179165e-01
1.13653290e+00 -1.25968015e+00 -1.34466484e-01 -3.77397716e-01
6.78556323e-01 -3.50164980e-01 8.66336763e-01 4.19556290e-01
-1.19794798e+00 -5.53174853e-01 -1.15518606e+00 -1.08683452e-01
-8.49856958e-02 4.97711003e-01 1.74690247e-01 3.35848421e-01
-1.03435469e+00 9.65486884e-01 -1.03214538e+00 -5.56969978e-02
8.77747595e-01 5.86399734e-01 -5.17534077e-01 -3.21955591e-01
-9.90753710e-01 5.63171446e-01 7.90853500e-01 4.02342714e-02
-9.33916628e-01 -5.35310566e-01 -6.57203555e-01 1.45606905e-01
4.45033520e-01 -8.20153296e-01 1.50161469e+00 -7.36495554e-01
-1.21538699e+00 8.32421839e-01 -1.14427917e-01 -9.11043704e-01
8.22351098e-01 -1.45959660e-01 2.88570404e-01 3.32012057e-01
-1.23939924e-01 1.19591975e+00 9.49636817e-01 -1.10550869e+00
-5.09011328e-01 -3.46511751e-01 -2.64487602e-02 1.85059771e-01
-2.93425828e-01 -2.40280762e-01 -9.11336958e-01 -6.64437473e-01
9.55413580e-02 -1.15099001e+00 -3.54176790e-01 3.39972824e-01
-6.83563828e-01 3.63190711e-01 8.19413781e-01 -6.40969932e-01
9.96164858e-01 -2.23783112e+00 -1.33724913e-01 3.06115508e-01
4.80043232e-01 4.40255344e-01 -3.05382282e-01 -4.95380759e-02
-9.74521562e-02 -6.40031707e-04 -2.95367926e-01 -7.48970151e-01
-2.37641677e-01 2.22993597e-01 -1.55263707e-01 3.68675858e-01
8.90688747e-02 8.14025581e-01 -4.17597651e-01 -4.74156916e-01
-9.10169184e-02 5.05548537e-01 -8.54926467e-01 -5.46046421e-02
-2.90203840e-01 2.82418251e-01 -2.56880641e-01 4.30787355e-01
7.58847952e-01 -3.99824321e-01 1.00029059e-01 -3.32449764e-01
3.16792876e-01 4.31184679e-01 -9.23979342e-01 1.64829767e+00
-2.18660891e-01 5.82752824e-01 1.93823591e-01 -1.07748389e+00
8.49050343e-01 5.95997833e-02 1.89913973e-01 -5.20241499e-01
3.41389358e-01 1.31430373e-01 -2.67000012e-02 -1.16243213e-01
2.06565052e-01 3.33154023e-01 2.00795010e-01 5.25807798e-01
4.76016924e-02 3.04205388e-01 2.14519814e-01 2.90768500e-02
1.16522825e+00 1.87303126e-01 9.34198685e-03 1.87057748e-01
1.06844045e-01 7.27895647e-02 5.79388022e-01 7.59422481e-01
-2.11864814e-01 7.32916832e-01 4.94810849e-01 -4.56252873e-01
-1.31862485e+00 -7.50776827e-01 -1.41597405e-01 1.06956232e+00
-1.80667669e-01 -4.56331879e-01 -1.27463126e+00 -7.84443796e-01
-1.54065965e-02 4.85663056e-01 -6.84443414e-01 -3.74707341e-01
-7.08310843e-01 -4.56955284e-01 8.02257776e-01 5.41373968e-01
7.28175998e-01 -9.64130819e-01 -1.04443491e+00 1.41685918e-01
-1.46418422e-01 -1.16874969e+00 -3.03629816e-01 3.88559729e-01
-1.21394491e+00 -1.06454694e+00 -7.20302045e-01 -7.86853194e-01
9.97288823e-01 9.79926288e-02 8.09071362e-01 2.74463177e-01
-3.81511509e-01 -2.49666139e-01 -2.93870829e-02 -4.05632466e-01
-3.87446970e-01 2.81451523e-01 -1.00260325e-01 -3.12581629e-01
2.75477290e-01 -6.61642313e-01 -7.48687744e-01 2.65142411e-01
-7.00206459e-01 4.43859428e-01 7.87179947e-01 9.78875756e-01
9.81125593e-01 4.54066023e-02 1.22428797e-01 -8.87589395e-01
4.50278163e-01 -5.02713062e-02 -7.15344250e-01 -5.63958324e-02
-3.54605615e-01 1.88250571e-01 5.09612799e-01 -7.07362473e-01
-7.34314501e-01 4.72053796e-01 -1.85670003e-01 -1.04188311e+00
-3.29111725e-01 3.17118496e-01 -5.15284128e-02 -1.27413437e-01
7.89710462e-01 1.40587315e-01 2.14608923e-01 -7.28419662e-01
3.29558440e-02 5.53801596e-01 6.80727243e-01 -4.09841537e-01
5.43375313e-01 2.92791486e-01 2.22574938e-02 -5.94095170e-01
-8.05128992e-01 1.74710765e-01 -5.67522466e-01 4.42138128e-02
5.28318822e-01 -8.68231893e-01 -4.27666217e-01 4.14934188e-01
-1.31805968e+00 -5.24995029e-01 -4.74461615e-01 4.88100111e-01
-4.98749077e-01 1.41387265e-02 -8.80236387e-01 -4.93009984e-01
-4.77612913e-01 -1.12497437e+00 9.20752466e-01 -5.60970269e-02
-1.40323356e-01 -4.54220414e-01 -2.14061901e-01 3.45086068e-01
1.81517109e-01 3.69124472e-01 7.75129497e-01 -9.07893479e-01
-7.36945093e-01 -6.27993166e-01 -2.48850510e-01 6.11710608e-01
-4.01406229e-01 -3.51352870e-01 -1.14184582e+00 -3.79990965e-01
1.30612746e-01 -5.47541440e-01 9.64657903e-01 3.31460297e-01
1.65276265e+00 -6.77051485e-01 -4.59864676e-01 8.06483924e-01
1.01989305e+00 -5.29249273e-02 4.93726194e-01 3.70838940e-01
6.62750602e-01 5.54062188e-01 4.42736626e-01 3.96817386e-01
6.19627684e-02 4.54404294e-01 6.39775574e-01 1.03455685e-01
-1.52115330e-01 -4.90853518e-01 1.93828776e-01 2.99353212e-01
-1.10991262e-02 -9.72939283e-02 -7.87372947e-01 2.36565873e-01
-1.67922699e+00 -8.41243148e-01 3.23205262e-01 2.44361520e+00
8.34778368e-01 3.54459643e-01 -6.09562397e-02 1.89186290e-01
8.02816391e-01 -4.03809287e-02 -7.80641437e-01 -1.66862741e-01
1.48649663e-01 1.04974262e-01 7.23878622e-01 4.05544460e-01
-1.13270020e+00 9.00253415e-01 5.38933325e+00 7.75771856e-01
-1.23796487e+00 8.39864910e-02 8.31216514e-01 -7.92345703e-01
2.39175692e-01 -2.41378084e-01 -9.67839777e-01 3.71027201e-01
8.82542253e-01 2.19190735e-02 5.30641198e-01 1.18173182e+00
-5.54674193e-02 1.10115625e-01 -1.26718092e+00 9.65393007e-01
-4.77137091e-03 -1.36394644e+00 -6.71481118e-02 2.64970720e-01
4.66482997e-01 3.21976662e-01 -4.80427966e-02 1.59937337e-01
4.23115008e-02 -9.42706168e-01 8.52527320e-01 2.37619653e-01
9.27146316e-01 -8.41401815e-01 5.78106582e-01 7.10985482e-01
-7.45488226e-01 -2.14533135e-01 -5.84638417e-01 6.54772744e-02
-1.91294357e-01 7.13264167e-01 -1.05777860e+00 -5.32307774e-02
9.13887143e-01 2.73395479e-01 -6.15017295e-01 9.89452362e-01
-1.87123507e-01 5.19943953e-01 -6.72911584e-01 1.53977171e-01
-3.24325971e-02 8.25648233e-02 4.35759455e-01 1.01633775e+00
2.44872883e-01 -1.37721285e-01 1.26946368e-03 8.63978982e-01
-4.83482659e-01 -9.65466052e-02 -6.37413800e-01 3.20077576e-02
6.73980713e-01 9.12922978e-01 -6.07144594e-01 -4.02425677e-01
-1.75072983e-01 1.11788404e+00 6.47666037e-01 3.02173138e-01
-8.20067704e-01 -3.26654702e-01 4.91721600e-01 3.10469836e-01
5.46103537e-01 2.51129381e-02 -2.39001691e-01 -9.07005429e-01
3.04129034e-01 -8.89493942e-01 2.32501358e-01 -6.41500533e-01
-6.51566505e-01 7.80407310e-01 1.34130761e-01 -1.04771948e+00
-3.44594508e-01 -5.77818632e-01 -3.33358884e-01 7.33706534e-01
-1.09396029e+00 -8.24614763e-01 -3.48949075e-01 5.06732225e-01
2.45517388e-01 -9.73184556e-02 7.20418632e-01 3.22943270e-01
-7.51574159e-01 1.09126854e+00 3.67883369e-02 3.35855842e-01
4.39219654e-01 -6.79058075e-01 5.07784069e-01 8.48996699e-01
-4.94827218e-02 6.18497431e-01 5.34495175e-01 -6.01544678e-01
-8.54048550e-01 -1.32169163e+00 7.57942975e-01 9.66634601e-02
2.34011814e-01 -6.62145257e-01 -9.76965368e-01 8.30964386e-01
-3.06683183e-01 1.97530955e-01 5.42146266e-01 -2.75131226e-01
-3.79431754e-01 -1.45543188e-01 -1.34108555e+00 7.35139906e-01
1.12558222e+00 -2.04431787e-01 -3.14264625e-01 3.76655906e-01
7.48612523e-01 -6.70813560e-01 -5.82320571e-01 4.46674794e-01
6.60536230e-01 -8.67005467e-01 9.58685160e-01 -4.96216327e-01
6.16213381e-01 -1.42058842e-02 -1.92462057e-01 -9.98212874e-01
-9.18773189e-02 -3.79567683e-01 -5.19148469e-01 6.46756411e-01
5.56112170e-01 -4.29230541e-01 1.41092181e+00 6.87730432e-01
-1.90667547e-02 -1.04505491e+00 -9.02758777e-01 -6.97833359e-01
-6.83769360e-02 -4.35285568e-01 6.00069880e-01 4.99478072e-01
-4.96766359e-01 2.08350182e-01 -3.96767706e-01 1.75461322e-01
8.30268562e-01 5.17310016e-02 9.35037017e-01 -1.04865897e+00
-7.38153696e-01 -3.83884281e-01 -2.46771589e-01 -1.29874110e+00
6.54952787e-03 -8.57000709e-01 7.64406398e-02 -1.06313181e+00
1.53277740e-01 -4.76058841e-01 -7.59453773e-02 7.49081612e-01
2.23455772e-01 5.24503767e-01 3.89041394e-01 3.48233134e-01
-3.08045834e-01 3.73326987e-01 1.35658777e+00 -4.46836129e-02
-1.75375715e-01 2.25263506e-01 -4.99893755e-01 8.48396540e-01
1.24608111e+00 -7.58993685e-01 -4.99157012e-01 -5.21539807e-01
-1.88094914e-01 6.37976900e-02 7.24444687e-01 -1.24842048e+00
3.01376939e-01 2.90788770e-01 6.54578745e-01 -5.80595255e-01
5.75892985e-01 -8.10300052e-01 1.54124841e-01 8.18777442e-01
-6.52547240e-01 -1.12877913e-01 2.79958457e-01 4.09777850e-01
-4.28656898e-02 -5.51349938e-01 9.34484482e-01 -2.09748134e-01
-2.52604544e-01 5.33398032e-01 3.78070516e-03 -1.25935063e-01
8.43842328e-01 -4.09022897e-01 -2.04288825e-01 -2.69731551e-01
-9.10846472e-01 4.48364876e-02 5.99074662e-01 1.07812576e-01
7.18841493e-01 -1.15425622e+00 -4.18515861e-01 5.29119134e-01
-1.25027254e-01 5.21984935e-01 1.96354702e-01 6.45994008e-01
-6.11511230e-01 2.38719121e-01 -2.64452279e-01 -5.41839838e-01
-1.48845160e+00 6.90753996e-01 4.50266659e-01 -2.86066145e-01
-7.68573463e-01 1.05103004e+00 2.87227571e-01 -3.50554407e-01
8.52878809e-01 -2.55894452e-01 2.19021901e-01 -2.31124744e-01
5.50087631e-01 2.52435833e-01 2.02308655e-01 -3.66920352e-01
-1.27985328e-01 1.42709181e-01 -4.79523152e-01 -1.49614766e-01
1.23629653e+00 1.25676632e-01 8.41069371e-02 2.43767519e-02
1.55934393e+00 -3.87481689e-01 -1.57230031e+00 -3.02310735e-01
-4.47350830e-01 -4.48321462e-01 4.62559052e-02 -7.16267169e-01
-1.47794318e+00 9.47563171e-01 5.98722696e-01 -2.94265211e-01
1.26721990e+00 -5.40626086e-02 8.50130856e-01 8.67620230e-01
3.83463442e-01 -1.07336760e+00 -6.11968711e-02 3.70855987e-01
9.34152901e-01 -9.23234701e-01 6.32111495e-03 -3.44095498e-01
-3.44215930e-01 9.44944739e-01 8.05222869e-01 -2.34075651e-01
6.39099419e-01 4.99802023e-01 -9.41753089e-02 3.29795629e-02
-8.31160784e-01 3.41503501e-01 -2.12975871e-02 4.79444385e-01
3.47254379e-03 -1.47044525e-01 4.12800722e-02 4.93497968e-01
-6.22916043e-01 2.67546445e-01 2.06271306e-01 1.11036098e+00
-4.71105695e-01 -1.07081175e+00 -2.76682019e-01 8.00634265e-01
-4.42828506e-01 -1.26569077e-01 -3.25593144e-01 6.86173856e-01
2.35286430e-01 3.42085451e-01 2.01773331e-01 -5.37738144e-01
3.19307625e-01 9.66075733e-02 5.19417107e-01 -4.85056996e-01
-2.86616832e-01 4.48200805e-03 4.36803102e-02 -8.19409966e-01
2.17019916e-02 -5.22195399e-01 -1.08812225e+00 -3.72452587e-01
-1.76042393e-01 -1.43131107e-01 7.10598767e-01 5.72060287e-01
6.12025857e-01 3.63612562e-01 2.97588408e-01 -1.11739683e+00
-8.05417776e-01 -8.79611731e-01 -1.87131315e-01 3.02474022e-01
3.65656763e-01 -5.37749290e-01 -2.12769553e-01 -4.18351311e-03] | [8.586209297180176, 3.2162117958068848] |
d55e20ec-2668-43aa-808d-f22576666979 | trankit-a-light-weight-transformer-based | 2101.03289 | null | https://arxiv.org/abs/2101.03289v5 | https://arxiv.org/pdf/2101.03289v5.pdf | Trankit: A Light-Weight Transformer-based Toolkit for Multilingual Natural Language Processing | We introduce Trankit, a light-weight Transformer-based Toolkit for multilingual Natural Language Processing (NLP). It provides a trainable pipeline for fundamental NLP tasks over 100 languages, and 90 pretrained pipelines for 56 languages. Built on a state-of-the-art pretrained language model, Trankit significantly outperforms prior multilingual NLP pipelines over sentence segmentation, part-of-speech tagging, morphological feature tagging, and dependency parsing while maintaining competitive performance for tokenization, multi-word token expansion, and lemmatization over 90 Universal Dependencies treebanks. Despite the use of a large pretrained transformer, our toolkit is still efficient in memory usage and speed. This is achieved by our novel plug-and-play mechanism with Adapters where a multilingual pretrained transformer is shared across pipelines for different languages. Our toolkit along with pretrained models and code are publicly available at: https://github.com/nlp-uoregon/trankit. A demo website for our toolkit is also available at: http://nlp.uoregon.edu/trankit. Finally, we create a demo video for Trankit at: https://youtu.be/q0KGP3zGjGc. | ['Viet Dac Lai', 'Minh Van Nguyen', 'Thien Huu Nguyen', 'Amir Pouran Ben Veyseh'] | 2021-01-09 | null | https://aclanthology.org/2021.eacl-demos.10 | https://aclanthology.org/2021.eacl-demos.10.pdf | eacl-2021-2 | ['morphological-tagging', 'multilingual-nlp'] | ['natural-language-processing', 'natural-language-processing'] | [-5.21530926e-01 -8.07748362e-02 -2.50290930e-01 -3.78158927e-01
-1.45580387e+00 -1.04605269e+00 2.78748184e-01 2.93282211e-01
-5.46372592e-01 5.90541780e-01 3.85665774e-01 -8.33546519e-01
5.45528054e-01 -5.26755512e-01 -6.98859870e-01 -3.23499054e-01
1.83389395e-01 6.38878703e-01 2.13716701e-01 -3.17085177e-01
-1.63425937e-01 6.38689473e-02 -5.65390050e-01 4.28843349e-01
9.07402992e-01 3.52385789e-01 3.98032397e-01 8.06705713e-01
-3.50454360e-01 4.88579214e-01 -1.91930488e-01 -7.56704867e-01
2.38083363e-01 -6.97019175e-02 -1.12578928e+00 -5.75441360e-01
3.27299714e-01 3.11177634e-02 -2.94245988e-01 1.05925584e+00
5.71051598e-01 -6.08262755e-02 1.27488896e-01 -8.79614115e-01
-7.16775239e-01 1.19560230e+00 -2.54405767e-01 3.82263601e-01
5.13258040e-01 3.18236560e-01 1.44940197e+00 -1.02159941e+00
7.59788394e-01 1.22584438e+00 6.68088853e-01 4.38666463e-01
-1.12130010e+00 -8.31970096e-01 2.06207186e-01 -6.39924454e-03
-1.21170950e+00 -5.90067804e-01 3.34315598e-01 -2.92738587e-01
1.65224504e+00 -8.50464106e-02 4.49467897e-01 9.96858418e-01
2.91856021e-01 1.18363976e+00 9.30737615e-01 -4.68126714e-01
-1.92477167e-01 -1.64791614e-01 3.62873286e-01 9.54935491e-01
-1.25160933e-01 -3.80753547e-01 -3.93761009e-01 -4.98695709e-02
7.01567352e-01 -2.04750910e-01 -2.94751208e-02 1.62995324e-01
-1.28480673e+00 6.35279417e-01 1.01035565e-01 5.61024606e-01
-6.53587505e-02 1.19882062e-01 6.72003627e-01 3.49915564e-01
6.43896461e-01 3.08748007e-01 -1.14083636e+00 -3.08086693e-01
-6.29477143e-01 -2.99738180e-02 8.32130194e-01 1.05590165e+00
8.51971567e-01 4.88414988e-02 1.66970417e-02 1.05360460e+00
1.59459874e-01 6.03089869e-01 4.28950608e-01 -7.22814023e-01
8.24787736e-01 6.14907920e-01 -2.66388863e-01 -2.20610887e-01
-3.74255538e-01 -2.58233342e-02 -2.94920832e-01 -4.40851778e-01
5.56461573e-01 -5.47264695e-01 -9.53964353e-01 1.60660005e+00
3.68760139e-01 -5.18657304e-02 1.64438099e-01 5.21402895e-01
8.99177849e-01 8.63230109e-01 4.39002514e-01 2.32866541e-01
1.77809465e+00 -1.13289928e+00 -5.17526209e-01 -4.12962139e-01
1.08769119e+00 -1.16505373e+00 1.37507355e+00 1.69411719e-01
-1.13422549e+00 -3.03563178e-01 -4.32415575e-01 -7.61753321e-01
-6.17183745e-01 2.14185491e-01 6.97916269e-01 3.55538875e-01
-1.21916366e+00 2.89347202e-01 -1.15731752e+00 -7.39198148e-01
1.61598310e-01 1.61202222e-01 -5.60180902e-01 -2.13548630e-01
-1.25927663e+00 8.97856474e-01 5.73778510e-01 -7.76308253e-02
-8.17021370e-01 -9.24556494e-01 -1.19772005e+00 -1.44474417e-01
1.68286368e-01 -4.35430735e-01 1.68566740e+00 -5.75397313e-01
-1.53132617e+00 1.02498043e+00 -3.05316418e-01 -3.64149243e-01
1.16772488e-01 -5.96375167e-01 -4.71109688e-01 6.59532985e-03
3.58086675e-01 6.53672695e-01 2.18990922e-01 -5.20110846e-01
-7.66900063e-01 -8.16057473e-02 2.36995704e-03 1.96919486e-01
-8.76396596e-02 8.09511185e-01 -7.84891427e-01 -5.99434495e-01
-3.08369935e-01 -7.99910843e-01 -4.64940816e-01 -5.88977873e-01
-5.53694248e-01 -4.82909173e-01 5.21639466e-01 -1.20048189e+00
1.14031160e+00 -2.10031915e+00 -1.48072734e-01 -2.68205225e-01
-2.45001659e-01 4.81163979e-01 -5.20257473e-01 8.25940192e-01
-1.12037234e-01 3.59356105e-01 -3.33762407e-01 -6.10125244e-01
1.61567166e-01 3.04911375e-01 -1.81266800e-01 2.19452292e-01
2.90172875e-01 1.22201824e+00 -1.15075266e+00 -5.39386809e-01
3.11654150e-01 6.34730160e-01 -4.35612738e-01 1.24103948e-01
-3.35129023e-01 5.49273014e-01 -2.97671109e-01 8.96223307e-01
5.45174003e-01 1.98686361e-01 5.44440269e-01 4.94317487e-02
-4.69945192e-01 1.09466553e+00 -6.79718494e-01 2.17119217e+00
-7.08062291e-01 4.43611473e-01 1.86655238e-01 -7.44110525e-01
7.97879815e-01 6.09654367e-01 2.02906981e-01 -4.94981080e-01
1.67703077e-01 5.16758978e-01 -1.27696201e-01 -3.00577372e-01
6.09609485e-01 9.46568400e-02 -5.75865984e-01 4.68061268e-01
5.87339461e-01 -3.56590748e-01 6.00852489e-01 5.43994009e-01
1.14095736e+00 4.07165378e-01 4.24762219e-01 -3.15283149e-01
3.23362082e-01 2.40267590e-01 8.84680867e-01 3.13035160e-01
-1.26150191e-01 2.56248683e-01 5.15956104e-01 -3.36719275e-01
-1.07297015e+00 -1.18622923e+00 -1.90598726e-01 1.55949354e+00
-6.90436363e-01 -7.14539587e-01 -6.27690136e-01 -8.83367896e-01
-1.65286675e-01 6.96355224e-01 -1.79111317e-01 4.59217280e-01
-8.79517972e-01 -6.98300421e-01 1.04748082e+00 5.57139397e-01
2.62111872e-01 -1.30348217e+00 1.31081805e-01 3.24958056e-01
-4.93293732e-01 -1.47674787e+00 -7.30697930e-01 4.49283004e-01
-6.91940129e-01 -1.00541675e+00 -5.08687675e-01 -1.33353245e+00
5.01922309e-01 -1.29910573e-01 1.28370643e+00 -2.04560488e-01
-1.35327607e-01 2.53845721e-01 -4.72820014e-01 -3.00955385e-01
-4.01482910e-01 4.78384674e-01 1.66539997e-02 -5.07484555e-01
6.42657697e-01 -4.72570419e-01 -3.24490249e-01 -1.54174194e-02
-5.76664329e-01 -6.76881569e-03 4.80208427e-01 5.33344984e-01
7.23344684e-01 -3.46709698e-01 1.70613214e-01 -1.04212451e+00
4.57239360e-01 -5.11799335e-01 -7.73378551e-01 2.49800801e-01
-8.96236897e-02 -3.80528122e-02 7.90713489e-01 -2.52434224e-01
-1.17466557e+00 1.89981192e-01 -7.64276564e-01 1.48938760e-01
-3.21008533e-01 6.07937932e-01 -3.68346632e-01 2.53698081e-01
2.95230240e-01 2.32152715e-01 -5.14035702e-01 -9.08833683e-01
7.07884371e-01 6.51696801e-01 5.46076775e-01 -8.11538994e-01
6.68879390e-01 5.13461828e-02 -8.30093563e-01 -6.74637377e-01
-9.02625442e-01 -6.62905395e-01 -1.04890800e+00 2.95182228e-01
9.44443524e-01 -1.20988381e+00 -2.98262209e-01 5.89216352e-01
-1.11726689e+00 -9.40054774e-01 -2.22587898e-01 4.20240521e-01
-6.50253296e-02 4.22395080e-01 -1.33179128e+00 -2.33781978e-01
-7.40868092e-01 -9.83269334e-01 8.77539158e-01 6.09522462e-02
-1.39636248e-01 -1.36379838e+00 3.19944501e-01 4.13904965e-01
2.42073238e-01 -1.17588751e-01 7.72688627e-01 -1.03624451e+00
-2.74918646e-01 -1.09243289e-01 -1.03402935e-01 4.35310781e-01
8.08953196e-02 1.43749744e-01 -6.52909398e-01 -2.33466893e-01
-5.14388263e-01 -4.02359903e-01 8.21855426e-01 2.96455532e-01
4.55355734e-01 -4.10056025e-01 -2.98027664e-01 7.33839929e-01
1.32396972e+00 -6.94112033e-02 3.41923356e-01 4.46723580e-01
8.75642121e-01 4.05844420e-01 5.37044048e-01 6.48769885e-02
8.98635805e-01 2.89190918e-01 -5.11879660e-02 -2.41784230e-01
-1.25231043e-01 -4.98789042e-01 8.37784350e-01 1.46002340e+00
2.85312414e-01 -1.46208420e-01 -1.48987317e+00 1.07231152e+00
-1.73454225e+00 -4.64408040e-01 -1.74263299e-01 1.81853020e+00
1.26573300e+00 -3.10258335e-03 -1.26037905e-02 -3.95085722e-01
5.51364958e-01 1.91137955e-01 -2.50910342e-01 -8.81597638e-01
-1.24922641e-01 5.72564483e-01 6.18096828e-01 9.25533473e-01
-1.11273193e+00 1.85820234e+00 5.25374699e+00 8.22982132e-01
-1.23539412e+00 4.46362734e-01 3.32279056e-01 1.76827554e-02
-3.92259479e-01 3.83408099e-01 -1.20032978e+00 2.25297347e-01
1.14182019e+00 -2.18602583e-01 2.95069218e-01 6.06153369e-01
2.15664402e-01 6.78418726e-02 -7.30397344e-01 5.02820373e-01
-2.63470709e-01 -1.07931674e+00 8.84805340e-03 -1.70039669e-01
4.96336669e-01 1.11321855e+00 -1.03312068e-01 5.37978172e-01
1.21826959e+00 -7.33602762e-01 5.44284523e-01 -2.64201492e-01
9.41128016e-01 -6.61352098e-01 6.63250208e-01 1.88734263e-01
-1.64522755e+00 2.40199372e-01 -4.17421192e-01 9.71854758e-03
5.72697043e-01 7.53967524e-01 -8.51234317e-01 5.60004234e-01
7.96603918e-01 8.86059880e-01 -4.67828780e-01 7.85549164e-01
-1.06661975e+00 9.86650705e-01 -5.28872430e-01 2.44339854e-01
4.04721171e-01 -6.28184974e-02 4.70865458e-01 1.94498181e+00
7.67218247e-02 1.63109109e-01 4.59254473e-01 2.04623446e-01
-2.41647735e-01 6.06346250e-01 -3.22323680e-01 -1.81406468e-01
6.59336865e-01 1.64602971e+00 -6.01017177e-01 -4.92252290e-01
-5.35684228e-01 9.78289127e-01 7.16209471e-01 4.17973489e-01
-7.12512076e-01 -5.31740129e-01 8.75570953e-01 -8.36805552e-02
4.30596471e-01 -6.47372246e-01 1.12627812e-01 -1.61292589e+00
-1.62753418e-01 -9.73408937e-01 7.41902351e-01 -4.52357560e-01
-1.13763773e+00 7.40825593e-01 -2.03651145e-01 -7.03620315e-01
-8.45863298e-02 -7.33328342e-01 -7.15473235e-01 9.61456478e-01
-1.57478738e+00 -1.66654801e+00 4.50221509e-01 7.58070886e-01
7.02322245e-01 2.77328286e-02 1.03826177e+00 4.11508650e-01
-9.12891448e-01 6.36736512e-01 -9.57323238e-03 8.60591888e-01
1.00295985e+00 -1.50359654e+00 1.16262770e+00 1.21809292e+00
2.40250096e-01 7.83108771e-01 3.44730198e-01 -7.54138649e-01
-1.31027114e+00 -1.35973394e+00 1.62000799e+00 -5.30344486e-01
1.33057213e+00 -7.81786084e-01 -8.42046559e-01 1.44599402e+00
6.34780109e-01 3.77076082e-02 8.56741667e-01 4.55579937e-01
-6.21093631e-01 1.88589454e-01 -9.75567877e-01 5.81611931e-01
8.31677556e-01 -7.48783529e-01 -6.25753522e-01 5.67523181e-01
9.06354904e-01 -6.24054372e-01 -1.12533581e+00 -1.89483818e-02
4.91236955e-01 -3.38590235e-01 6.69793010e-01 -4.08343643e-01
2.71756321e-01 -1.94433361e-01 -2.21288316e-02 -1.44385707e+00
-3.74014735e-01 -7.65753984e-01 3.58448952e-01 1.55601096e+00
9.04484808e-01 -9.72141922e-01 3.03750157e-01 2.58497417e-01
-4.24883008e-01 -6.34489715e-01 -8.78852129e-01 -7.87047446e-01
5.88493407e-01 -8.75709713e-01 3.67422879e-01 1.18544066e+00
5.01890481e-01 6.58735871e-01 -5.67211844e-02 2.20121652e-01
5.26517808e-01 -1.09019421e-01 5.72479010e-01 -6.43489182e-01
-2.63929039e-01 -2.66506672e-01 1.36240363e-01 -1.20153046e+00
3.61619085e-01 -1.33707392e+00 8.37295949e-02 -1.81985188e+00
7.90287405e-02 -5.71663439e-01 -2.52532780e-01 1.29994631e+00
-1.81908831e-01 2.83839375e-01 2.87131339e-01 1.46853417e-01
-6.35979474e-01 1.35930136e-01 8.52095306e-01 -1.60977785e-02
-2.74953127e-01 -3.76271367e-01 -6.45396113e-01 7.50698864e-01
1.36348033e+00 -6.80683315e-01 1.55807137e-01 -1.22117341e+00
2.18576014e-01 -2.83913314e-01 -1.29852235e-01 -6.09123230e-01
2.10442320e-01 1.13207810e-01 -1.08250203e-02 -5.74939668e-01
1.92399874e-01 -1.79180816e-01 -3.46059382e-01 4.34622228e-01
5.56807555e-02 4.80270714e-01 6.26636088e-01 -2.02039152e-01
-1.82712957e-01 1.04374224e-02 5.11336148e-01 -5.03406107e-01
-7.52169967e-01 3.92016053e-01 -5.96320927e-01 4.52753842e-01
5.56157291e-01 3.12844157e-01 -5.56241035e-01 7.89893717e-02
-7.49811709e-01 6.00659609e-01 4.12153602e-01 5.33746004e-01
1.46770865e-01 -8.28730762e-01 -9.42315876e-01 6.69318661e-02
-4.66015637e-02 5.86198121e-02 3.18477973e-02 7.96062112e-01
-6.58258379e-01 6.57450557e-01 -2.16277484e-02 -1.51856855e-01
-1.25617445e+00 2.38434449e-01 1.16532482e-01 -7.95967102e-01
-5.36315918e-01 1.17901409e+00 1.18675761e-01 -1.19897723e+00
-2.04066068e-01 -7.04220176e-01 -4.68068160e-02 -9.92059931e-02
4.80322719e-01 -2.62746267e-04 9.77952313e-03 -6.73775911e-01
-6.33829236e-01 4.70648140e-01 -3.31881613e-01 -2.45777369e-01
1.37972939e+00 -1.33352578e-01 -3.94341171e-01 3.99904639e-01
9.84665453e-01 5.17768860e-01 -7.88452387e-01 -3.76756191e-01
2.00867131e-01 5.80209270e-02 -2.74417400e-01 -9.82958615e-01
-8.93301785e-01 9.20246899e-01 -9.10506025e-02 -3.64874601e-01
9.68081474e-01 3.16351533e-01 1.34990811e+00 3.19018662e-01
3.72289240e-01 -1.04763460e+00 -4.19301689e-01 1.31105006e+00
5.20146310e-01 -1.10825324e+00 -2.84944743e-01 -4.52222109e-01
-7.17033207e-01 7.61638463e-01 4.00109380e-01 -1.17193833e-01
7.98027754e-01 6.52959049e-01 6.19556487e-01 1.91250746e-03
-8.95186305e-01 -3.79475474e-01 -7.58304074e-02 5.63123167e-01
1.00469899e+00 3.79509419e-01 -4.48162675e-01 6.49085402e-01
-6.25153065e-01 -2.41915658e-01 3.49070996e-01 8.89486134e-01
-1.50412768e-01 -1.71721971e+00 -9.22307447e-02 5.66304149e-03
-9.92397547e-01 -1.01809788e+00 -1.48140237e-01 6.44782066e-01
2.93622762e-02 7.75262475e-01 3.78834419e-02 -1.08283252e-01
2.45696411e-01 3.10533822e-01 2.35594481e-01 -1.07503867e+00
-1.02287388e+00 1.76691338e-01 4.45783406e-01 -5.53732872e-01
-5.70062920e-02 -8.37137878e-01 -1.46773612e+00 -2.89829999e-01
3.26560698e-02 5.23789883e-01 5.79941511e-01 7.89997816e-01
3.84037286e-01 1.89531267e-01 1.12940364e-01 -6.49252534e-01
5.97961210e-02 -1.15233696e+00 -2.98368245e-01 -1.18465014e-01
3.88190225e-02 6.59992173e-02 -1.06168166e-01 1.46869212e-01] | [10.504111289978027, 9.973572731018066] |
9a46a272-b615-44be-91bf-1c67dcdf9e18 | inference-from-stationary-time-sequences-via | 2006.03258 | null | https://arxiv.org/abs/2006.03258v4 | https://arxiv.org/pdf/2006.03258v4.pdf | Learned Factor Graphs for Inference from Stationary Time Sequences | The design of methods for inference from time sequences has traditionally relied on statistical models that describe the relation between a latent desired sequence and the observed one. A broad family of model-based algorithms have been derived to carry out inference at controllable complexity using recursive computations over the factor graph representing the underlying distribution. An alternative model-agnostic approach utilizes machine learning (ML) methods. Here we propose a framework that combines model-based algorithms and data-driven ML tools for stationary time sequences. In the proposed approach, neural networks are developed to separately learn specific components of a factor graph describing the distribution of the time sequence, rather than the complete inference task. By exploiting stationary properties of this distribution, the resulting approach can be applied to sequences of varying temporal duration. Learned factor graph can be realized using compact neural networks that are trainable using small training sets, or alternatively, be used to improve upon existing deep inference systems. We present an inference algorithm based on learned stationary factor graphs, which learns to implement the sum-product scheme from labeled data, and can be applied to sequences of different lengths. Our experimental results demonstrate the ability of the proposed learned factor graphs to learn to carry out accurate inference from small training sets for sleep stage detection using the Sleep-EDF dataset, as well as for symbol detection in digital communications with unknown channels. | ['Yonina C. Eldar', 'Nariman Farsad', 'Andrea J. Goldsmith', 'Nir Shlezinger'] | 2020-06-05 | null | null | null | null | ['sleep-stage-detection'] | ['medical'] | [ 5.19727588e-01 -3.94213013e-02 -4.74359095e-01 -5.14726460e-01
-5.53445518e-01 -4.40214872e-01 4.40640748e-01 6.65400252e-02
-3.52745384e-01 7.65325785e-01 -3.80560189e-01 -7.13585556e-01
-3.73541504e-01 -5.39187014e-01 -7.45357037e-01 -8.98602068e-01
-6.37740672e-01 5.92384338e-01 -2.56919060e-02 5.07065915e-02
1.89924940e-01 6.45806134e-01 -1.37680495e+00 -1.40134409e-01
4.25162941e-01 1.17958474e+00 2.69969374e-01 1.13207030e+00
-1.67261492e-02 7.79994905e-01 -7.53423333e-01 -1.29093543e-01
7.10827634e-02 -6.11620724e-01 -6.80441409e-02 1.13942340e-01
9.28356349e-02 -3.57409060e-01 -6.27930403e-01 9.00342286e-01
3.60987693e-01 -1.06748827e-01 7.75082588e-01 -1.23812723e+00
-8.55283737e-02 9.22671080e-01 -1.26716554e-01 5.00782609e-01
2.66781598e-02 -1.65386200e-01 8.56816530e-01 -3.38138014e-01
4.10211310e-02 1.15794492e+00 6.62052512e-01 3.31597596e-01
-1.37771547e+00 -8.12227964e-01 -1.12084553e-01 3.65952909e-01
-1.44107449e+00 -5.95124066e-01 7.17390239e-01 -2.33769909e-01
1.08853614e+00 1.05682723e-02 6.28633082e-01 1.26090169e+00
4.23105597e-01 6.80168688e-01 9.85604525e-01 -6.88970089e-01
5.33345163e-01 -1.83977157e-01 2.36280128e-01 7.88912773e-01
1.94140434e-01 3.17676336e-01 -4.23277318e-01 -2.19900861e-01
6.58911824e-01 -3.05605400e-02 -1.94322422e-01 -1.69387490e-01
-9.50810432e-01 8.11137378e-01 -1.89835936e-01 2.12314397e-01
-6.48446083e-02 6.25462294e-01 5.61186075e-01 5.20567536e-01
3.69182050e-01 -2.11896263e-02 -5.75475156e-01 -3.45821410e-01
-1.30892920e+00 2.49972776e-01 1.26197100e+00 9.40535963e-01
8.22872877e-01 3.51091832e-01 -3.17939550e-01 2.35403627e-01
2.70612955e-01 7.55075157e-01 4.01413262e-01 -7.51787841e-01
2.94794828e-01 1.78051535e-02 -8.57064351e-02 -8.77230644e-01
-5.84058821e-01 -8.48677039e-01 -9.31953371e-01 -1.15109578e-01
4.79027003e-01 -1.26496792e-01 -9.53967929e-01 1.97207975e+00
-2.23255187e-01 5.60377240e-01 -1.33405309e-02 2.33548433e-01
-1.22314781e-01 5.49331367e-01 -3.97548914e-01 -5.81257224e-01
1.04075801e+00 -4.63880271e-01 -9.05476093e-01 -1.21817641e-01
4.27846670e-01 -2.22058758e-01 5.82955897e-01 8.14892828e-01
-9.10276115e-01 -6.03233814e-01 -1.44370735e+00 3.04398566e-01
-1.75566882e-01 3.08041751e-01 5.47043860e-01 9.32017088e-01
-1.06293464e+00 8.82163286e-01 -1.19772446e+00 -9.61857790e-04
2.35169679e-01 6.01666451e-01 1.56927601e-01 1.00640334e-01
-1.20941579e+00 6.73357129e-01 7.89200425e-01 3.37148547e-01
-1.31766725e+00 -4.80060548e-01 -8.39980364e-01 3.75571936e-01
3.38350743e-01 -5.22535026e-01 1.26333678e+00 -8.64410460e-01
-1.62441444e+00 3.03885430e-01 -3.23550165e-01 -1.10459602e+00
2.57703453e-01 -1.14321120e-01 -5.46713829e-01 3.25727224e-01
-3.50663006e-01 -2.25891083e-01 1.55958295e+00 -6.83670342e-01
-3.67077172e-01 -1.81493521e-01 -9.15449187e-02 -5.81956863e-01
-3.35515410e-01 -4.43411261e-01 -4.18025851e-01 -5.11496007e-01
-3.26610178e-01 -1.00824451e+00 -3.44307274e-01 -3.20705235e-01
-4.14740056e-01 1.18656680e-02 8.35111380e-01 -6.21864140e-01
1.65169704e+00 -1.99858999e+00 1.76410854e-01 5.49945295e-01
3.63576449e-02 2.09010422e-01 1.11084424e-01 7.11517930e-01
-2.01467853e-02 -3.46918285e-01 -4.18617517e-01 -5.65519691e-01
1.66771993e-01 5.74861825e-01 -5.50455034e-01 6.83716595e-01
1.72942523e-02 6.57637715e-01 -9.78940785e-01 -2.99659193e-01
1.82045236e-01 3.53013337e-01 -4.19892848e-01 3.41524780e-01
-5.42379081e-01 3.97007287e-01 -2.39298582e-01 9.91289914e-02
3.96797538e-01 -3.54474485e-01 4.89707232e-01 -6.83523342e-02
1.13131784e-01 2.28939936e-01 -1.01619923e+00 1.85692215e+00
-8.85308683e-01 8.32248330e-01 -8.82836506e-02 -1.56049895e+00
9.29622769e-01 4.39270020e-01 2.73284674e-01 -3.92867535e-01
3.27189237e-01 1.73483253e-01 7.53927082e-02 -5.03933430e-01
6.64914325e-02 -2.73281306e-01 -2.28889003e-01 6.35193467e-01
3.85011494e-01 2.07460389e-01 4.07005399e-01 1.82140395e-01
1.37080109e+00 3.61309983e-02 4.98328745e-01 -1.70654833e-01
6.27256334e-01 -5.95677733e-01 2.46866152e-01 1.21242833e+00
2.47143999e-01 -5.85849322e-02 5.38721383e-01 -3.12998444e-01
-1.05035305e+00 -1.15187895e+00 -3.32760364e-02 8.50715220e-01
-3.25328201e-01 -5.01006842e-01 -6.40243351e-01 -3.83507997e-01
-2.23041654e-01 7.86650956e-01 -4.77364779e-01 -4.33579385e-01
-5.21606922e-01 -7.02688396e-01 7.29823887e-01 4.91471201e-01
2.73191184e-01 -7.92208195e-01 -6.37768209e-01 5.33974886e-01
7.41776824e-02 -1.37799168e+00 -2.29179382e-01 7.49718964e-01
-9.15445089e-01 -6.82558894e-01 -3.37184489e-01 -3.41737330e-01
4.85599577e-01 -1.80122718e-01 1.01345193e+00 3.50228138e-02
-2.16942757e-01 4.50191379e-01 -1.15577891e-01 -4.21087086e-01
-7.59861887e-01 1.37437448e-01 2.47101635e-01 2.87394196e-01
2.23766550e-01 -1.28578484e+00 -3.11524481e-01 4.52201813e-02
-1.06218922e+00 -1.79680839e-01 7.95721233e-01 8.05355906e-01
3.11819136e-01 3.92048687e-01 6.70908868e-01 -8.13054204e-01
5.02556622e-01 -5.12284517e-01 -1.01448596e+00 2.33117267e-01
-6.48421466e-01 8.12218308e-01 1.06831741e+00 -5.41980982e-01
-6.95909917e-01 3.54511812e-02 -3.01380694e-01 -7.68772423e-01
1.16760381e-01 6.66269064e-01 -1.46838903e-01 1.95633024e-01
4.34695095e-01 7.04521596e-01 1.00397006e-01 -2.26687446e-01
3.31026793e-01 6.31933987e-01 7.37140238e-01 -6.16929412e-01
8.70051563e-01 5.99757493e-01 4.59293783e-01 -9.34688568e-01
-1.02452683e+00 -4.26519841e-01 -5.89197278e-01 -6.19120076e-02
5.34790397e-01 -8.19419444e-01 -8.64880383e-01 3.33105028e-01
-1.05538881e+00 -5.52607954e-01 -1.72551051e-02 6.29194677e-01
-9.49562848e-01 4.66055274e-01 -5.59560716e-01 -1.16362429e+00
-2.62068123e-01 -7.73263574e-01 1.20533442e+00 -6.76243529e-02
-1.90736443e-01 -1.45353329e+00 2.32632980e-01 -3.17307234e-01
3.53053451e-01 1.69091657e-01 1.13356531e+00 -8.46593499e-01
-5.57968020e-01 -5.51822126e-01 1.74705565e-01 5.72395802e-01
1.06652863e-01 -1.49998918e-01 -9.67481971e-01 -5.96289933e-01
2.64267564e-01 -1.02763966e-01 7.46829987e-01 5.47844827e-01
1.31698477e+00 -3.66424888e-01 -5.50217867e-01 6.49704695e-01
1.39423430e+00 1.27214327e-01 5.37113607e-01 -2.24520162e-01
4.16381031e-01 -4.32901829e-02 3.87600273e-01 7.94250190e-01
2.68117321e-04 6.51619554e-01 2.14506432e-01 3.73580813e-01
2.25554287e-01 -2.39984125e-01 5.75596869e-01 8.92018139e-01
3.54948640e-02 -3.51436943e-01 -5.05766094e-01 2.67989427e-01
-1.84490240e+00 -1.26836407e+00 2.24267319e-01 2.41686726e+00
7.10626066e-01 7.05796480e-01 1.42312437e-01 5.43649137e-01
4.55339819e-01 7.08942115e-03 -7.10478961e-01 -3.85270834e-01
3.37653100e-01 6.94700778e-01 6.24074996e-01 5.21641672e-01
-8.59726667e-01 3.69541764e-01 6.73442173e+00 1.10817373e+00
-1.18163514e+00 -1.77566577e-02 2.04058304e-01 1.69038311e-01
-8.37960765e-02 1.04755715e-01 -9.21405375e-01 4.52571094e-01
1.86893523e+00 -2.19374388e-01 7.20957637e-01 6.63515270e-01
5.13915956e-01 2.02160582e-01 -1.57427502e+00 1.12866604e+00
-7.92335048e-02 -1.20009112e+00 -1.67528525e-01 1.73617452e-01
3.12848270e-01 -2.75446951e-01 -2.00075228e-02 5.07840633e-01
8.12252313e-02 -1.08652580e+00 5.89282632e-01 8.74532104e-01
6.12689078e-01 -8.41283202e-01 5.73036015e-01 7.57714868e-01
-1.18265021e+00 -2.59403586e-01 -2.81774163e-01 -3.64970565e-01
1.48452461e-01 9.28536832e-01 -1.23613524e+00 5.44838786e-01
2.34437376e-01 8.17524791e-01 -2.26703092e-01 7.22343624e-01
-1.78953648e-01 9.11688983e-01 -4.22641039e-01 -2.00060502e-01
3.74173671e-02 -1.45326167e-01 5.01873672e-01 1.35521972e+00
5.51989079e-01 -4.50000077e-01 1.23450063e-01 6.83457673e-01
1.85523301e-01 -5.06466150e-01 -6.39119804e-01 -3.64155382e-01
4.64174926e-01 1.12953866e+00 -7.88992524e-01 -4.25530553e-01
-4.01123881e-01 8.61870289e-01 4.13679749e-01 4.82781410e-01
-9.73881066e-01 -3.90761822e-01 3.44806343e-01 -1.00583181e-01
8.63632321e-01 -7.41673827e-01 1.27830341e-01 -1.00502717e+00
-1.80433095e-01 -9.02848125e-01 3.90453756e-01 -3.16974252e-01
-1.07963622e+00 5.02630830e-01 3.77132624e-01 -1.22532380e+00
-9.60920691e-01 -6.49930775e-01 -5.20432532e-01 7.74625361e-01
-1.44078648e+00 -7.61724830e-01 7.44256973e-02 6.92080259e-01
3.77719879e-01 -2.18537658e-01 1.00335050e+00 2.45777890e-01
-3.78157139e-01 6.17682993e-01 2.85043746e-01 7.24332929e-02
9.96213928e-02 -1.39872921e+00 2.70825505e-01 8.89002919e-01
3.97529513e-01 7.54144013e-01 9.71169829e-01 -3.29691231e-01
-1.62452936e+00 -1.05585611e+00 4.39457893e-01 -2.90166289e-01
7.93017209e-01 -6.62923396e-01 -6.64560139e-01 8.26210499e-01
7.22761974e-02 6.73546717e-02 8.51347387e-01 3.78116705e-02
-3.00457865e-01 -2.46408239e-01 -6.18854642e-01 4.01641458e-01
9.02005374e-01 -7.40275323e-01 -2.90376931e-01 1.48082688e-01
4.97149110e-01 -3.33464563e-01 -7.69694924e-01 1.66800171e-01
7.42082834e-01 -9.14028645e-01 7.87728846e-01 -4.35071915e-01
3.92943760e-03 -3.39574397e-01 -2.27818042e-01 -1.19452083e+00
-5.12482151e-02 -1.08470547e+00 -9.04268205e-01 7.69372404e-01
3.36063802e-01 -4.04929131e-01 6.57026529e-01 -1.03116356e-01
-1.62298918e-01 -3.90227199e-01 -9.81966257e-01 -9.59255517e-01
-3.48786563e-01 -8.59380126e-01 3.55656594e-01 2.84113914e-01
-7.74134174e-02 5.94321430e-01 -7.28293002e-01 5.84515214e-01
8.47649872e-01 3.38279098e-01 7.28245616e-01 -9.83408749e-01
-1.14626145e+00 -2.16838285e-01 -6.24219358e-01 -1.34751952e+00
5.41496933e-01 -9.46464777e-01 1.89113811e-01 -9.21325564e-01
-1.40395030e-01 -1.78610191e-01 -2.98350155e-01 1.41188800e-01
1.62181780e-01 -2.36844290e-02 -1.90082848e-01 -1.61721244e-01
-5.39806604e-01 4.30230141e-01 7.88143516e-01 -8.81341100e-02
-1.53449783e-02 5.41830063e-01 -2.00545430e-01 4.91759211e-01
7.84697592e-01 -5.28288543e-01 -7.29404390e-01 1.59190059e-01
4.61458385e-01 5.11742413e-01 3.48288745e-01 -1.25586629e+00
2.17576101e-01 2.46938810e-01 2.84122407e-01 -6.12654984e-01
3.39339852e-01 -9.73056972e-01 2.80397624e-01 8.29007208e-01
-4.57586706e-01 -1.98400199e-01 1.49186671e-01 1.02617836e+00
-6.40277639e-02 -3.89894277e-01 4.74266320e-01 6.33434132e-02
-4.04711992e-01 3.70591074e-01 -4.65624988e-01 -1.29052982e-01
6.74400806e-01 7.15062767e-02 1.90203846e-01 -8.00593197e-01
-9.47488487e-01 -8.60346388e-03 -1.24041967e-01 -5.23828678e-02
5.63374102e-01 -1.25054979e+00 -3.98383111e-01 2.56311029e-01
-1.98574718e-02 -3.32573533e-01 -2.86153350e-02 8.49275529e-01
-2.69256949e-01 6.50249898e-01 -8.86420533e-03 -7.98402607e-01
-8.53332043e-01 8.13275635e-01 2.96736598e-01 -6.96473539e-01
-4.49288696e-01 3.44469666e-01 -1.80186808e-01 -8.80221557e-03
3.49180937e-01 -6.98306143e-01 2.08346620e-01 -2.33049884e-01
7.34750390e-01 1.52175277e-01 1.71805114e-01 -1.35934487e-01
-1.40540957e-01 2.82876462e-01 1.77322142e-02 -8.80012363e-02
1.11134708e+00 -2.05498576e-01 4.03165333e-02 9.55707967e-01
1.58947730e+00 -1.52736217e-01 -1.30983901e+00 -4.22838122e-01
1.77381068e-01 -6.37405142e-02 5.37451766e-02 -3.32395345e-01
-8.81006837e-01 9.75369453e-01 6.44807756e-01 5.24891138e-01
1.33668256e+00 -1.77555054e-01 6.92699313e-01 5.66898406e-01
5.67676306e-01 -6.33087397e-01 2.08822079e-02 3.74335885e-01
3.06040198e-01 -7.45008349e-01 -8.07273835e-02 -2.43104286e-02
1.58505559e-01 1.70103705e+00 -2.07199186e-01 -1.17685132e-01
8.90054584e-01 6.31000638e-01 -3.93389463e-01 -1.30482882e-01
-9.28391397e-01 -2.31275707e-01 1.52336597e-01 5.05535841e-01
1.79568365e-01 5.31416461e-02 -1.17760830e-01 5.55485845e-01
-1.90254048e-01 2.44470149e-01 5.53131461e-01 8.53410363e-01
-3.90630245e-01 -1.15891671e+00 -2.04579547e-01 6.50226057e-01
-4.91499245e-01 -1.79992080e-01 2.95106441e-01 6.79857433e-01
-1.53459191e-01 8.11954319e-01 1.43640503e-01 -3.24387491e-01
-1.27963409e-01 2.10945442e-01 8.20504785e-01 -4.24048483e-01
1.23009399e-01 3.93216424e-02 1.61728993e-01 -4.39921290e-01
-5.22261560e-01 -6.03581429e-01 -1.08631933e+00 -7.36215711e-02
-2.27012038e-01 2.06373036e-01 5.90417564e-01 1.42627895e+00
-1.33742383e-02 7.04797864e-01 6.98816180e-01 -9.66428816e-01
-8.17561507e-01 -9.57847774e-01 -8.24189842e-01 -1.06251873e-01
7.79386342e-01 -4.80637223e-01 -4.84461457e-01 1.99676186e-01] | [6.572680950164795, 1.6391844749450684] |
ddde0dd0-5a41-49da-8d4c-3c066a320860 | transrac-encoding-multi-scale-temporal | 2204.01018 | null | https://arxiv.org/abs/2204.01018v1 | https://arxiv.org/pdf/2204.01018v1.pdf | TransRAC: Encoding Multi-scale Temporal Correlation with Transformers for Repetitive Action Counting | Counting repetitive actions are widely seen in human activities such as physical exercise. Existing methods focus on performing repetitive action counting in short videos, which is tough for dealing with longer videos in more realistic scenarios. In the data-driven era, the degradation of such generalization capability is mainly attributed to the lack of long video datasets. To complement this margin, we introduce a new large-scale repetitive action counting dataset covering a wide variety of video lengths, along with more realistic situations where action interruption or action inconsistencies occur in the video. Besides, we also provide a fine-grained annotation of the action cycles instead of just counting annotation along with a numerical value. Such a dataset contains 1,451 videos with about 20,000 annotations, which is more challenging. For repetitive action counting towards more realistic scenarios, we further propose encoding multi-scale temporal correlation with transformers that can take into account both performance and efficiency. Furthermore, with the help of fine-grained annotation of action cycles, we propose a density map regression-based method to predict the action period, which yields better performance with sufficient interpretability. Our proposed method outperforms state-of-the-art methods on all datasets and also achieves better performance on the unseen dataset without fine-tuning. The dataset and code are available. | ['Shenghua Gao', 'Zhengxin Li', 'Dongze Lian', 'Yiqun Zhao', 'Sixun Dong', 'Huazhang Hu'] | 2022-04-03 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Hu_TransRAC_Encoding_Multi-Scale_Temporal_Correlation_With_Transformers_for_Repetitive_Action_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Hu_TransRAC_Encoding_Multi-Scale_Temporal_Correlation_With_Transformers_for_Repetitive_Action_CVPR_2022_paper.pdf | cvpr-2022-1 | ['repetitive-action-counting'] | ['computer-vision'] | [ 4.62816805e-01 -5.23474634e-01 -5.94486237e-01 -2.56039143e-01
-6.18439436e-01 -4.03222293e-01 3.86672884e-01 -6.60339072e-02
-4.94062185e-01 6.89806938e-01 5.23056805e-01 1.34695098e-01
-1.32103980e-01 -5.57372808e-01 -5.51385701e-01 -6.22219622e-01
-2.54659921e-01 6.32264093e-03 5.48752129e-01 1.56388864e-01
2.03741118e-01 2.34722063e-01 -1.53903770e+00 4.17692840e-01
6.85369730e-01 9.51450109e-01 2.42574990e-01 7.05198348e-01
2.76550889e-01 1.25170958e+00 -7.44889200e-01 -3.84717882e-01
1.07254900e-01 -5.50165832e-01 -6.15666807e-01 4.63397712e-01
4.25687701e-01 -5.55351079e-01 -5.46728194e-01 9.26427364e-01
3.23581636e-01 4.11255956e-01 4.65323448e-01 -1.20400262e+00
-2.84480304e-01 3.78827482e-01 -8.96578193e-01 6.91137612e-01
6.31722212e-01 3.61469239e-01 8.37457001e-01 -3.75601858e-01
3.93663555e-01 1.17976344e+00 5.83805144e-01 3.03016245e-01
-5.92289388e-01 -5.97290397e-01 2.49308884e-01 6.73708439e-01
-1.46642220e+00 -2.52815545e-01 5.60919106e-01 -4.54658806e-01
7.23772287e-01 2.61361539e-01 7.97423899e-01 1.30325234e+00
-1.61925331e-02 8.72323990e-01 6.70742035e-01 -1.80877998e-01
1.65103763e-01 -5.12609839e-01 -2.54495740e-01 4.92080599e-01
2.26759180e-01 -2.46265844e-01 -4.67591524e-01 7.50505626e-02
8.43718708e-01 4.39566821e-01 -2.82793432e-01 7.15205595e-02
-1.58148599e+00 5.81003368e-01 3.47125009e-02 3.25795263e-01
-4.36939210e-01 2.96966374e-01 8.72225881e-01 -1.09904781e-01
3.32114458e-01 1.62806287e-01 -3.80047321e-01 -8.69707644e-01
-9.41944480e-01 1.59887210e-01 4.99646574e-01 1.10897207e+00
3.69812101e-01 2.06254516e-02 -7.19011068e-01 8.32551301e-01
-2.81917006e-01 3.25376749e-01 6.05979979e-01 -1.41948187e+00
7.22993135e-01 6.23718023e-01 1.98803887e-01 -1.31632066e+00
-5.20404875e-01 -2.03318223e-01 -1.03842711e+00 -4.13867325e-01
6.21543825e-01 2.42842995e-02 -4.36198145e-01 1.61048126e+00
3.69882107e-01 6.75653398e-01 -2.89766759e-01 1.09996843e+00
4.25084949e-01 4.70036954e-01 1.31362438e-01 -5.81055164e-01
1.61148500e+00 -1.19526553e+00 -1.00545537e+00 -1.66963443e-01
9.57772791e-01 -3.62465858e-01 1.31464875e+00 5.43487132e-01
-7.34978318e-01 -6.59743905e-01 -7.94491291e-01 -1.08593982e-03
-7.04945177e-02 4.35339153e-01 7.35248268e-01 4.89646375e-01
-3.94148320e-01 6.39265776e-01 -9.58333313e-01 -3.47947091e-01
5.54653406e-01 -1.16103971e-02 -4.07789290e-01 -2.14965016e-01
-1.09597945e+00 5.99468350e-01 5.57679594e-01 1.33085269e-02
-6.26558185e-01 -5.29922485e-01 -1.08121347e+00 2.76836790e-02
8.71731639e-01 -2.99445868e-01 1.19794381e+00 -6.99053645e-01
-1.14515173e+00 4.27420259e-01 -2.06938639e-01 -4.00967538e-01
7.40020216e-01 -4.10796463e-01 -5.15013337e-01 3.90252620e-01
8.77624750e-02 3.71423662e-01 5.88101268e-01 -5.84012926e-01
-6.51333570e-01 -1.87898993e-01 3.99102956e-01 2.17060164e-01
-7.23211586e-01 -6.17076643e-02 -7.87628472e-01 -8.61882150e-01
-3.91094595e-01 -9.10479605e-01 -1.42753094e-01 -1.83855612e-02
-1.79138720e-01 -2.86578685e-01 7.82252789e-01 -5.95895469e-01
1.95918345e+00 -2.29526305e+00 6.64144978e-02 -4.00527537e-01
1.54065892e-01 2.68822134e-01 -3.61423828e-02 3.69159162e-01
9.08977762e-02 -1.47123709e-01 -9.33003947e-02 -2.49821797e-01
-1.57207966e-01 4.51417446e-01 8.63660648e-02 5.59359670e-01
-6.57616463e-03 7.54955649e-01 -1.23215544e+00 -9.53245699e-01
3.89868766e-01 2.84754306e-01 -4.83830124e-01 1.39668155e-02
-3.71376872e-02 7.05235720e-01 -5.54330647e-01 7.95644104e-01
4.29230034e-01 -4.68936294e-01 1.75290897e-01 -3.55080694e-01
5.67427874e-02 -2.63986737e-01 -1.33139217e+00 1.93403816e+00
-4.32725936e-01 5.36088347e-01 -5.53965330e-01 -9.70519543e-01
5.23024261e-01 2.97726065e-01 7.84886897e-01 -5.74141145e-01
3.87235451e-03 -8.21418539e-02 -5.94457202e-02 -9.73237395e-01
5.41076064e-01 2.24990726e-01 -2.47954965e-01 2.95760661e-01
-1.69775918e-01 2.66634136e-01 8.52809608e-01 3.65043469e-02
1.53480566e+00 2.85712421e-01 7.29439676e-01 2.00214356e-01
6.38585091e-01 -2.16087848e-01 8.81227612e-01 6.71896040e-01
-4.97701436e-01 6.76465392e-01 5.61337292e-01 -6.43653572e-01
-7.89635956e-01 -5.40286303e-01 2.21576206e-02 1.05948758e+00
1.92720160e-01 -6.45479858e-01 -7.09435761e-01 -7.79314578e-01
-1.75529286e-01 3.79261285e-01 -6.69489324e-01 -1.22111358e-01
-8.64226818e-01 -7.30431974e-01 6.27434254e-01 9.03752387e-01
1.04811025e+00 -1.09427965e+00 -6.67711139e-01 1.14739448e-01
-7.77462721e-01 -1.66342604e+00 -8.60960603e-01 -3.77972454e-01
-8.76514614e-01 -1.38713801e+00 -7.64019489e-01 -2.37023488e-01
3.44560325e-01 4.11137849e-01 1.18925798e+00 1.59057617e-01
-3.85895342e-01 2.69512981e-01 -6.15893602e-01 -7.19163641e-02
-1.51552498e-01 2.03425344e-02 1.06026806e-01 3.17215268e-03
6.91087067e-01 -5.01825333e-01 -9.04688656e-01 6.70936584e-01
-9.96461868e-01 -5.62707037e-02 4.84304488e-01 6.61134005e-01
5.16202807e-01 2.55839437e-01 4.98807222e-01 -6.28317177e-01
2.85469711e-01 -5.78197479e-01 -3.68292451e-01 1.80649281e-01
-9.09321755e-02 -2.68156767e-01 7.31164634e-01 -9.00754511e-01
-8.47238243e-01 6.58269078e-02 4.66794968e-02 -6.88734472e-01
-1.11418098e-01 1.73597649e-01 -7.19269514e-02 3.35469514e-01
5.89738846e-01 2.08981380e-01 -3.63762498e-01 -3.32880288e-01
4.52928692e-02 6.20220184e-01 6.21645331e-01 -4.80661094e-01
4.23475176e-01 5.03584981e-01 2.01864913e-01 -7.23227918e-01
-1.20149660e+00 -7.94680238e-01 -6.81114256e-01 -3.95172954e-01
9.96099651e-01 -1.03082657e+00 -9.63881552e-01 4.73519176e-01
-1.03363347e+00 -3.17496091e-01 -3.27091604e-01 8.78718555e-01
-8.45651448e-01 8.19270074e-01 -7.34096348e-01 -8.06744099e-01
6.87671974e-02 -1.01671731e+00 1.29324126e+00 8.44035745e-02
-2.93817729e-01 -7.48205245e-01 2.54202256e-04 5.09235561e-01
-7.95261636e-02 4.81842428e-01 4.24620122e-01 -3.78472716e-01
-5.32666802e-01 -2.97934353e-01 -2.86058575e-01 4.73133475e-01
3.74526083e-01 -1.13811389e-01 -7.15012610e-01 -1.30976394e-01
-1.37645066e-01 -2.55080730e-01 6.81625426e-01 4.61299866e-01
1.89636493e+00 -2.37765417e-01 -2.56346881e-01 4.35176998e-01
1.10446739e+00 4.61207554e-02 9.28051114e-01 2.38696277e-01
8.66495550e-01 4.41531330e-01 1.20112514e+00 8.24006915e-01
3.97632271e-01 9.32882547e-01 4.48699892e-01 1.25203013e-01
4.23005819e-02 -1.75722346e-01 1.80727169e-01 8.53134394e-01
-7.50162780e-01 -4.46886808e-01 -5.24990082e-01 5.01630783e-01
-2.07777214e+00 -1.48837543e+00 -4.36210707e-02 2.26301551e+00
6.76300287e-01 8.28709975e-02 5.12348354e-01 4.75391895e-01
7.56786048e-01 4.49143887e-01 -4.85017747e-01 3.04060243e-02
1.66543171e-01 -3.60550016e-01 5.82512081e-01 -4.12084870e-02
-1.28052068e+00 5.99465847e-01 6.09106588e+00 1.30631042e+00
-5.57927847e-01 1.19578280e-01 5.35579145e-01 -2.39671722e-01
3.57960612e-01 -3.94062012e-01 -6.86202407e-01 9.88954842e-01
8.51823986e-01 1.48943231e-01 3.34402710e-01 7.75008559e-01
6.68739259e-01 -3.98083359e-01 -1.02735698e+00 1.28732479e+00
1.20404914e-01 -1.18433416e+00 -1.66975573e-01 -4.71438002e-03
6.01302028e-01 -3.46455663e-01 -5.02951503e-01 4.90356296e-01
-9.42979679e-02 -8.24642956e-01 4.82881814e-01 5.69797635e-01
9.55739260e-01 -4.89116967e-01 7.48007059e-01 5.20033598e-01
-1.47765684e+00 -4.05658066e-01 -3.00946325e-01 -2.96015888e-01
3.64652604e-01 6.12525880e-01 -2.04156473e-01 5.91075540e-01
8.19766164e-01 1.19205618e+00 -5.53513050e-01 1.10431218e+00
-1.25358820e-01 5.11641741e-01 -2.26903841e-01 7.60428086e-02
9.32975411e-02 -1.39065415e-01 2.06530154e-01 1.46859813e+00
5.27440965e-01 2.58559048e-01 2.89237946e-01 1.51006207e-01
-6.92842826e-02 -4.98424396e-02 -6.76879764e-01 -1.77858651e-01
4.32739854e-01 1.26172686e+00 -7.08353460e-01 -5.66070318e-01
-7.45802462e-01 1.09209085e+00 1.71954647e-01 2.30178982e-01
-1.36861002e+00 -8.18158537e-02 6.03754580e-01 2.44123369e-01
2.93864816e-01 -3.22351336e-01 2.22047910e-01 -1.37097216e+00
3.60291600e-01 -8.21224153e-01 7.04702437e-01 -7.65722930e-01
-1.03865147e+00 2.65412599e-01 3.60855967e-01 -1.81439984e+00
-3.97009015e-01 -4.00284767e-01 -4.03951019e-01 2.09931597e-01
-1.24332249e+00 -7.92377889e-01 -9.07398880e-01 5.92357099e-01
1.05681169e+00 1.38194740e-01 5.36918879e-01 8.44671428e-01
-6.14059091e-01 4.68661636e-01 -3.11427623e-01 2.16638163e-01
7.73557901e-01 -9.95519936e-01 8.32020789e-02 6.76595986e-01
4.41217721e-02 1.52264461e-01 5.67166090e-01 -5.52940249e-01
-1.16173947e+00 -1.27830136e+00 5.58545053e-01 -5.63051462e-01
5.84334373e-01 -1.81657717e-01 -7.94499815e-01 6.68241620e-01
-2.26512879e-01 4.58646804e-01 6.69740081e-01 -9.86781493e-02
-9.42559093e-02 -7.43749365e-02 -1.05153430e+00 4.77419347e-01
1.52127767e+00 -3.07702720e-01 -2.26320103e-01 6.47195160e-01
6.36969745e-01 -5.89718699e-01 -1.04262567e+00 4.48824823e-01
6.49794757e-01 -9.81657743e-01 9.89966512e-01 -4.35866475e-01
7.05186844e-01 -4.61806297e-01 -1.80009320e-01 -8.48509133e-01
-2.78554320e-01 -3.86683732e-01 -5.74717641e-01 9.98189986e-01
-2.17570841e-01 -1.49618283e-01 7.60870516e-01 2.18320951e-01
-6.87654987e-02 -8.82746696e-01 -8.73596787e-01 -1.09807253e+00
-5.18953919e-01 -5.24276555e-01 6.21845067e-01 9.00981843e-01
-1.25014521e-02 -5.82977794e-02 -9.01596367e-01 -5.57401851e-02
3.03626984e-01 -4.29988913e-02 8.76183689e-01 -8.60149920e-01
-4.23900932e-01 -4.99024652e-02 -8.88040125e-01 -1.34638870e+00
-3.41055766e-02 -2.44932279e-01 -7.90004581e-02 -1.32292902e+00
5.50573647e-01 5.77466972e-02 -1.53951034e-01 3.15498888e-01
-2.39698544e-01 5.16106904e-01 1.51030391e-01 2.53657222e-01
-1.25228417e+00 5.10091007e-01 1.62976956e+00 -2.66928952e-02
1.08930031e-02 1.38426170e-01 -2.44722709e-01 1.01339722e+00
5.75285375e-01 -4.65415686e-01 -5.53612769e-01 -3.57602835e-01
5.89000359e-02 3.80442649e-01 4.50091809e-01 -1.30441415e+00
2.78710984e-02 -4.66244012e-01 3.40239823e-01 -5.72633743e-01
4.92300987e-01 -8.48874092e-01 1.95096135e-01 3.17604601e-01
-2.54613787e-01 7.82597885e-02 1.41909746e-02 1.01564705e+00
-3.26285154e-01 -2.22799122e-01 4.98674601e-01 -3.05593491e-01
-8.27466846e-01 4.23839837e-01 -2.20982954e-01 2.76188999e-01
1.32939124e+00 -5.85655153e-01 -4.01209265e-01 -4.85574603e-01
-7.55439818e-01 3.40310335e-01 3.37644428e-01 5.31366050e-01
2.48372987e-01 -1.55980384e+00 -6.08599544e-01 -2.03010067e-01
3.63388717e-01 -9.07718837e-02 6.73438609e-01 1.22539008e+00
-4.07396138e-01 3.18896174e-01 -1.43829405e-01 -7.90196657e-01
-1.15779006e+00 6.89425826e-01 7.91334286e-02 -4.48308915e-01
-7.87081957e-01 2.71093607e-01 2.74398565e-01 3.13854963e-02
1.99116781e-01 -5.51675200e-01 -4.04984534e-01 9.67285037e-02
9.16378677e-01 7.37737715e-01 -1.55512497e-01 -6.07808352e-01
-3.68424833e-01 6.22188151e-01 2.95496911e-01 4.58081216e-01
9.20599341e-01 -2.71047264e-01 4.40290958e-01 6.40847385e-01
1.08712292e+00 -7.54172429e-02 -1.55025136e+00 -1.61714301e-01
-1.71050027e-01 -1.03544438e+00 -3.25136900e-01 -3.31174970e-01
-1.23697531e+00 7.28633881e-01 3.87033850e-01 5.63976690e-02
1.36626244e+00 -1.86934099e-01 9.05183554e-01 2.52190113e-01
5.18123865e-01 -1.19084907e+00 4.12807673e-01 2.66265333e-01
6.54630840e-01 -1.54045987e+00 3.37955743e-01 -5.66570163e-01
-8.17514300e-01 1.05908215e+00 6.83606029e-01 1.40090436e-01
2.57551104e-01 1.59858778e-01 -3.98133129e-01 -1.19322844e-01
-5.52716613e-01 -1.55933768e-01 1.59343258e-01 5.21461427e-01
3.48747969e-01 3.16469707e-02 -5.48247337e-01 5.44486165e-01
2.44090065e-01 5.18579900e-01 5.74506402e-01 8.54862392e-01
-2.16470212e-01 -6.43134415e-01 -1.70553759e-01 5.94270706e-01
-6.26084447e-01 2.27186039e-01 3.31000298e-01 8.80175710e-01
2.18944848e-01 9.57164645e-01 1.11420743e-01 -2.48756766e-01
4.08196270e-01 -2.58537173e-01 4.73802358e-01 -4.57919508e-01
-2.20543623e-01 -8.33233669e-02 2.48273924e-01 -9.12275672e-01
-8.56571615e-01 -7.48060226e-01 -1.04781485e+00 -5.81745625e-01
-3.15407276e-01 -2.13299796e-01 7.13072568e-02 9.39094961e-01
1.48085967e-01 7.50788748e-01 4.02500808e-01 -9.18991446e-01
-5.21777153e-01 -1.15206206e+00 -7.59636760e-01 7.05623448e-01
4.63669933e-02 -9.84013617e-01 -3.82743776e-01 3.44231755e-01] | [8.31790542602539, 0.5544586181640625] |
fc5c5d75-d75d-4a01-8620-8d79f5933663 | task3-dcase2021-challenge-sound-event | 2107.14561 | null | https://arxiv.org/abs/2107.14561v1 | https://arxiv.org/pdf/2107.14561v1.pdf | TASK3 DCASE2021 Challenge: Sound event localization and detection using squeeze-excitation residual CNNs | Sound event localisation and detection (SELD) is a problem in the field of automatic listening that aims at the temporal detection and localisation (direction of arrival estimation) of sound events within an audio clip, usually of long duration. Due to the amount of data present in the datasets related to this problem, solutions based on deep learning have positioned themselves at the top of the state of the art. Most solutions are based on 2D representations of the audio (different spectrograms) that are processed by a convolutional-recurrent network. The motivation of this submission is to study the squeeze-excitation technique in the convolutional part of the network and how it improves the performance of the system. This study is based on the one carried out by the same team last year. This year, it has been decided to study how this technique improves each of the datasets (last year only the MIC dataset was studied). This modification shows an improvement in the performance of the system compared to the baseline using MIC dataset. | ['Maximo Cobos', 'Francesc J. Ferri', 'Pedro Zuccarello', 'Sergi Perez-Castanos', 'Javier Naranjo-Alcazar'] | 2021-07-30 | null | null | null | null | ['direction-of-arrival-estimation', 'sound-event-localization-and-detection'] | ['audio', 'audio'] | [ 1.23921484e-01 -1.53490424e-01 7.78850377e-01 -6.14026040e-02
-7.56994963e-01 -3.49404156e-01 3.79408479e-01 2.67732471e-01
-5.84769368e-01 3.90175223e-01 6.23162866e-01 1.60918925e-02
-3.80315006e-01 -5.63882411e-01 -4.23798919e-01 -6.42617166e-01
-3.98213148e-01 -6.54124236e-03 6.95794821e-01 -2.64434785e-01
3.32307220e-01 4.97286230e-01 -2.04758763e+00 8.61565530e-01
-2.79476076e-01 8.42647791e-01 3.46438140e-01 1.05918849e+00
-4.94150594e-02 7.28762984e-01 -8.80410492e-01 1.95170373e-01
9.33655798e-02 -6.55944765e-01 -8.02433312e-01 -4.25594896e-01
3.34817767e-01 1.47052482e-01 -1.49493650e-01 6.21166885e-01
1.15886378e+00 1.81013823e-01 3.72629374e-01 -6.93384767e-01
4.93065268e-01 7.28898227e-01 4.54638079e-02 7.04023302e-01
7.37759471e-01 -5.35761833e-01 6.77350461e-01 -9.77391899e-01
4.95790362e-01 9.19651330e-01 1.07717800e+00 3.17818075e-02
-8.61653030e-01 -4.84246999e-01 -4.29218441e-01 5.84547400e-01
-1.41228724e+00 -5.59873283e-01 8.60280097e-01 -3.82672191e-01
1.31833112e+00 3.05666983e-01 5.10688901e-01 1.02314448e+00
-2.17901971e-02 5.89116454e-01 1.07209468e+00 -7.96900213e-01
4.73365426e-01 1.16527252e-01 -2.32103363e-01 1.24632239e-01
-3.61378878e-01 2.82291353e-01 -6.91187978e-01 1.85515195e-01
4.30982113e-01 -3.52780521e-01 -1.17455341e-01 5.43582253e-02
-9.38808978e-01 7.95575142e-01 2.38384023e-01 1.11081755e+00
-6.14635706e-01 1.10042073e-01 9.22997534e-01 5.09038746e-01
4.11014974e-01 4.34490621e-01 -4.05089021e-01 -7.16803133e-01
-1.68095851e+00 3.24978381e-01 1.06785035e+00 8.99609327e-02
1.22465201e-01 2.27557465e-01 -4.00774963e-02 6.58427477e-01
1.21034093e-01 -9.15418491e-02 7.70537734e-01 -5.56517422e-01
1.60594180e-01 1.19718239e-01 -2.34963924e-01 -1.04711258e+00
-7.82389760e-01 -5.52284598e-01 -6.12620056e-01 3.33022445e-01
4.16880608e-01 -4.36616838e-01 -7.18584836e-01 1.35267031e+00
1.39326096e-01 3.40137839e-01 -2.41692647e-01 7.73074150e-01
9.01873350e-01 9.13234293e-01 -3.25202405e-01 -2.45178297e-01
1.09591246e+00 -6.42665744e-01 -9.70095813e-01 3.25236112e-01
4.42302912e-01 -1.26868999e+00 3.96971405e-01 8.66304696e-01
-1.13466156e+00 -8.12230527e-01 -1.13225019e+00 3.40347528e-01
-7.80826271e-01 1.31168902e-01 2.16659501e-01 7.58248925e-01
-1.27206433e+00 6.63273692e-01 -6.10186636e-01 -4.55262065e-01
-1.77459568e-01 3.05079997e-01 -2.70914555e-01 4.50952381e-01
-1.50369465e+00 7.12676466e-01 4.15126622e-01 1.76119030e-01
-6.37179971e-01 -5.74372292e-01 -4.57079053e-01 2.43665697e-03
1.57771543e-01 -1.66893601e-02 1.31122220e+00 -7.74353921e-01
-1.68433321e+00 6.74540401e-01 1.38148591e-01 -8.52187276e-01
7.21132338e-01 -3.31159979e-01 -5.50758243e-01 1.63165033e-01
-9.37864631e-02 3.93949360e-01 8.11062932e-01 -6.16043448e-01
-8.93597603e-01 4.28523645e-02 -2.16985941e-01 -1.87575310e-01
2.00109929e-01 4.43719536e-01 -5.33385836e-02 -8.25853765e-01
9.49102640e-02 -7.22704232e-01 -1.72506571e-02 -8.43872011e-01
-2.04883918e-01 -3.08186948e-01 7.74113715e-01 -8.25323761e-01
1.40325189e+00 -2.52246904e+00 1.01766922e-01 -3.45037058e-02
-2.42713332e-01 4.63548809e-01 1.79961845e-01 1.08687222e+00
-6.76637113e-01 -1.32695630e-01 1.15829380e-03 -4.24361289e-01
2.83991601e-02 -1.72724053e-01 -5.35561204e-01 4.31164801e-01
-1.66163146e-01 2.25467965e-01 -5.01781404e-01 -3.73208821e-01
3.09574693e-01 6.06830955e-01 -2.63509154e-01 2.70031661e-01
8.56953561e-02 1.99667469e-01 -3.51691954e-02 3.19984779e-02
6.21523619e-01 5.76995134e-01 -1.92249283e-01 -2.57161140e-01
-7.70250499e-01 6.22397304e-01 -1.76549900e+00 1.77051485e+00
-4.62521434e-01 9.67322826e-01 -5.12120612e-02 -1.22920847e+00
1.05131423e+00 1.12550020e+00 6.37608290e-01 -4.93073434e-01
2.56321549e-01 4.42315042e-01 2.04977050e-01 -6.95811450e-01
3.77984077e-01 -2.60039091e-01 1.74211547e-01 5.05521238e-01
3.33435625e-01 -1.95339322e-02 2.05170602e-01 -3.53781700e-01
1.16299200e+00 1.65972590e-01 1.06232539e-01 -2.19390765e-02
5.37708163e-01 -4.22401607e-01 -1.41797429e-02 7.31474936e-01
-1.55284494e-01 9.57661986e-01 3.98560405e-01 -5.81290424e-01
-9.53033268e-01 -6.30374074e-01 -2.39038214e-01 1.05636132e+00
-6.21120512e-01 -5.22161603e-01 -8.14616740e-01 -3.31850111e-01
-5.14050364e-01 3.66106719e-01 -6.74489081e-01 -5.99474460e-02
-7.74325430e-01 -5.84816992e-01 8.48015487e-01 3.54293287e-01
2.72663176e-01 -1.62056375e+00 -1.11619508e+00 7.11725771e-01
-9.30044055e-02 -8.55482399e-01 3.30293119e-01 7.79462934e-01
-6.27206981e-01 -6.01989686e-01 -6.77821815e-01 -7.17926383e-01
-3.68139595e-01 -1.65874615e-01 9.87347782e-01 -4.24611032e-01
-5.51978946e-01 3.28819990e-01 -8.99783969e-01 -9.42617774e-01
-3.23909044e-01 3.38674664e-01 -1.40385419e-01 9.95233431e-02
3.94020855e-01 -9.67772663e-01 -4.87494379e-01 7.84206167e-02
-1.11367834e+00 -5.16027153e-01 4.74913508e-01 2.58896232e-01
1.86393693e-01 3.29146564e-01 6.15394294e-01 -4.96192008e-01
7.88442016e-01 -2.25128859e-01 -3.43715191e-01 -3.90432149e-01
-8.24157670e-02 -2.20038787e-01 4.27994996e-01 -3.45219612e-01
-5.49504519e-01 2.30599314e-01 -7.23272562e-01 -1.98088855e-01
-6.20200515e-01 4.48270649e-01 2.01449752e-01 2.11585104e-03
5.45193136e-01 -1.70854945e-02 -3.40846002e-01 -9.96262074e-01
-7.72676840e-02 8.16026688e-01 4.37600642e-01 1.52067959e-01
2.68192321e-01 2.43755996e-01 1.62398174e-01 -9.54993904e-01
-5.32369614e-01 -7.38913655e-01 -7.17755675e-01 -4.26126629e-01
8.84200215e-01 -6.37951553e-01 -4.79351342e-01 4.62799460e-01
-1.24055767e+00 2.47875620e-02 -5.62919617e-01 8.38513851e-01
-4.05880302e-01 -7.36891031e-02 -3.60831112e-01 -1.10264051e+00
-2.40812331e-01 -9.61285710e-01 9.11080360e-01 1.35737211e-02
-4.09208179e-01 -7.48934805e-01 6.33770943e-01 -9.43333209e-02
6.78545177e-01 3.24403763e-01 4.85307902e-01 -8.12753141e-01
-2.23225262e-02 -6.65029645e-01 3.03708732e-01 5.23182452e-01
-7.93308243e-02 -8.26083794e-02 -1.58502376e+00 2.91755516e-03
4.11450118e-01 5.59107261e-03 1.04835677e+00 6.05422139e-01
9.25619006e-01 3.66318583e-01 -4.58694622e-02 1.64701536e-01
1.20575821e+00 3.87034595e-01 8.43429089e-01 4.85434592e-01
-3.64545025e-02 7.23308444e-01 4.80413467e-01 5.87210357e-01
-3.12460870e-01 1.02260077e+00 4.81914341e-01 -5.61227277e-03
-4.36499834e-01 6.77109808e-02 5.28237402e-01 9.75132406e-01
-6.12123497e-02 -2.19349310e-01 -7.28932619e-01 7.64739811e-01
-1.59347498e+00 -1.11879730e+00 -5.19254267e-01 2.17574191e+00
4.33549047e-01 3.01193476e-01 4.46041316e-01 1.12400591e+00
6.86159849e-01 3.40583950e-01 2.68510342e-01 -1.00013196e+00
6.45809472e-02 6.72763705e-01 -1.56118954e-02 3.43782634e-01
-1.35415685e+00 4.16859686e-01 6.44457531e+00 8.00941527e-01
-1.66078436e+00 1.52350724e-01 -6.39413893e-02 -2.70019770e-01
7.02340126e-01 -3.15646857e-01 -6.38679862e-01 5.34663081e-01
1.58233273e+00 4.49074477e-01 2.71608591e-01 4.65138733e-01
4.02336657e-01 -2.72414267e-01 -1.01900280e+00 9.65632319e-01
2.52285093e-01 -1.11783469e+00 -5.70546389e-01 -1.02664411e-01
3.04340512e-01 9.92647856e-02 1.02453474e-02 3.14088106e-01
-4.43955809e-01 -1.00777793e+00 7.95896292e-01 7.76488781e-01
3.59964788e-01 -9.56800163e-01 1.17992938e+00 2.12716714e-01
-1.09492266e+00 -9.20034051e-02 -2.58965105e-01 -3.11740726e-01
5.16777098e-01 7.27022946e-01 -1.09204757e+00 6.40430510e-01
1.16738892e+00 4.54376787e-01 -4.86054093e-01 1.51393437e+00
-1.36589576e-02 9.88916457e-01 -5.15166342e-01 -8.09277892e-02
2.64942259e-01 3.17286551e-01 9.97130454e-01 1.60964143e+00
4.35509980e-01 -4.54319686e-01 -1.96024969e-01 4.39516991e-01
3.18805069e-01 2.63356388e-01 -5.37930489e-01 1.63325980e-01
-4.41777296e-02 1.13582349e+00 -7.47399569e-01 -1.67368352e-02
-8.66425708e-02 6.22595489e-01 -1.73795670e-01 -7.68272877e-02
-5.93813121e-01 -7.75305212e-01 1.25016615e-01 1.70194834e-01
8.17231357e-01 -2.61767283e-02 -6.91893697e-02 -2.73558587e-01
4.33004647e-02 -6.57731414e-01 3.56407404e-01 -9.51842964e-01
-8.56407106e-01 9.34053063e-01 -5.61925173e-02 -1.26644099e+00
-6.72849417e-01 -3.86354923e-01 -7.34860361e-01 8.85939837e-01
-9.85109866e-01 -4.99022901e-01 2.03383550e-01 3.87876421e-01
6.03107989e-01 -2.11535215e-01 1.10179257e+00 6.57012761e-01
-2.04893097e-01 2.95396477e-01 -9.66191292e-03 -7.01226667e-02
8.47434580e-01 -1.27667892e+00 1.35413796e-01 7.68446565e-01
6.03635848e-01 1.67123541e-01 1.07410169e+00 -2.32452035e-01
-7.64360607e-01 -7.68114805e-01 1.39807081e+00 -1.96406424e-01
5.82282543e-01 -4.52065080e-01 -7.74297655e-01 1.82368889e-01
5.95786750e-01 -1.34570718e-01 6.80074573e-01 -9.02158469e-02
1.17764689e-01 -2.87065566e-01 -7.98093736e-01 -1.11880958e-01
2.41524696e-01 -6.09462559e-01 -8.08056474e-01 6.34048581e-02
5.49719810e-01 -4.21778768e-01 -6.12261236e-01 2.90091306e-01
4.24632341e-01 -1.39490509e+00 8.47675264e-01 -2.03410566e-01
2.71438837e-01 -3.74984354e-01 -1.84074596e-01 -1.08000815e+00
-1.15364799e-02 -7.08942652e-01 -3.92196327e-02 1.33843839e+00
2.55310833e-01 -2.36378238e-01 3.86702955e-01 -3.84694993e-01
-7.75909498e-02 -4.71545875e-01 -1.30140114e+00 -5.26966572e-01
-2.41113871e-01 -8.93947423e-01 2.39142448e-01 4.42634940e-01
-2.33198464e-01 4.70962286e-01 -2.91729659e-01 1.05100252e-01
-8.64768997e-02 4.84471843e-02 3.97551537e-01 -1.34459770e+00
-2.97272116e-01 -4.40028995e-01 -8.82522941e-01 -4.16714281e-01
-3.21785510e-01 -5.94704032e-01 1.15925893e-01 -1.41642582e+00
-4.93478030e-01 4.11734544e-02 -5.21495223e-01 2.84244031e-01
4.01677936e-01 5.03576875e-01 2.21201420e-01 -1.92304641e-01
-4.92204994e-01 3.39127295e-02 6.16184294e-01 2.71805197e-01
-3.16588849e-01 5.35416484e-01 -2.76212376e-02 5.72902143e-01
7.05451488e-01 -7.47052908e-01 -1.58105657e-01 -5.38896546e-02
6.61804259e-01 1.34449750e-01 4.35769439e-01 -1.69473064e+00
3.77110034e-01 8.22607040e-01 2.87301838e-01 -1.01463675e+00
4.92022872e-01 -9.79560912e-01 3.01047862e-01 4.24039155e-01
-4.12482381e-01 1.13217510e-01 4.28317398e-01 2.56798089e-01
-7.42489398e-01 -6.58148289e-01 7.01052964e-01 -2.11915538e-01
-6.11760259e-01 -3.84078741e-01 -7.94742167e-01 -2.26450846e-01
6.90403461e-01 -2.25169986e-01 4.05248433e-01 -5.65943658e-01
-9.88966823e-01 -6.78347886e-01 -4.12683368e-01 3.11080128e-01
3.43110442e-01 -9.88196611e-01 -8.10727537e-01 4.88256514e-02
-2.62835443e-01 -2.08172470e-01 5.10720551e-01 1.11175418e+00
-4.59871560e-01 6.26878083e-01 -2.24387035e-01 -4.93072391e-01
-1.45429015e+00 3.99717391e-01 6.81256950e-01 -1.79814786e-01
-5.70610285e-01 9.90423322e-01 -4.63924468e-01 -3.05682346e-02
6.51341438e-01 -4.00270015e-01 -9.29730058e-01 5.36657274e-01
6.71028972e-01 5.95551968e-01 7.60577917e-01 -5.33614635e-01
-4.21611011e-01 4.89484519e-01 6.59721866e-02 -4.20829892e-01
1.84118497e+00 8.92023370e-02 -1.45538986e-01 9.50328708e-01
1.39912057e+00 3.74549896e-01 -7.72814393e-01 2.18998164e-01
6.72386289e-02 -4.83895019e-02 5.02536476e-01 -9.31738317e-01
-8.61982048e-01 1.37814546e+00 1.26281631e+00 9.62771952e-01
1.28760505e+00 -2.52392322e-01 5.53259373e-01 -2.48426641e-03
1.01744667e-01 -1.14698911e+00 -1.35949105e-01 7.40162909e-01
1.02573442e+00 -7.04257011e-01 -1.64402768e-01 -1.78765133e-02
-1.63539335e-01 1.59081018e+00 -8.79439712e-02 -3.24199259e-01
8.79806519e-01 5.85066378e-01 2.04167753e-01 -2.05142647e-01
-5.74827850e-01 -3.85650337e-01 2.38474935e-01 5.93627453e-01
7.13423967e-01 -7.44910017e-02 -3.82404298e-01 5.54191589e-01
-4.77405399e-01 6.10703975e-02 4.59276527e-01 9.05722678e-01
-3.75773221e-01 -1.22732317e+00 -6.23740852e-01 1.13265645e-02
-9.76741016e-01 -1.81512788e-01 -5.85904539e-01 7.08878696e-01
5.25158763e-01 1.23489487e+00 -1.01284690e-01 -4.53357637e-01
5.62880278e-01 3.70014340e-01 2.72323102e-01 -4.20730054e-01
-1.23717630e+00 3.67097020e-01 -3.94178480e-02 -5.46869695e-01
-5.57299674e-01 -8.95906389e-01 -1.06568360e+00 3.24962318e-01
-1.71319619e-01 4.03414309e-01 1.08582664e+00 8.51722419e-01
2.40047663e-01 1.09534776e+00 7.97117472e-01 -1.03098500e+00
-2.00655028e-01 -1.26512647e+00 -7.60437369e-01 1.22375146e-01
5.10001242e-01 -4.62757140e-01 -4.46077675e-01 1.11193061e-01] | [15.162205696105957, 5.254172325134277] |
62a89d0d-9fdc-4d73-9b1f-2427f0b60d3d | rita-a-study-on-scaling-up-generative-protein | 2205.05789 | null | https://arxiv.org/abs/2205.05789v2 | https://arxiv.org/pdf/2205.05789v2.pdf | RITA: a Study on Scaling Up Generative Protein Sequence Models | In this work we introduce RITA: a suite of autoregressive generative models for protein sequences, with up to 1.2 billion parameters, trained on over 280 million protein sequences belonging to the UniRef-100 database. Such generative models hold the promise of greatly accelerating protein design. We conduct the first systematic study of how capabilities evolve with model size for autoregressive transformers in the protein domain: we evaluate RITA models in next amino acid prediction, zero-shot fitness, and enzyme function prediction, showing benefits from increased scale. We release the RITA models openly, to the benefit of the research community. | ['Debora Marks', 'Iacopo Poli', 'Pascal Notin', 'Niccoló Zanichelli', 'Daniel Hesslow'] | 2022-05-11 | null | null | null | null | ['protein-design'] | ['medical'] | [ 3.31519365e-01 3.63381535e-01 5.27178459e-02 -4.03226435e-01
-6.95155084e-01 -6.53770685e-01 1.37726605e-01 -3.12427014e-01
-1.84701815e-01 1.01856351e+00 3.05515945e-01 -6.13815427e-01
-1.20520927e-01 -4.12358284e-01 -9.95480955e-01 -8.66198897e-01
-1.23011909e-01 1.02310038e+00 5.40582053e-02 -3.45094562e-01
-1.67060588e-02 3.26713890e-01 -1.13922179e+00 2.93729872e-01
8.33281100e-01 3.79114389e-01 4.91425544e-01 6.49797738e-01
2.30195016e-01 6.52899086e-01 -4.88739043e-01 -5.21268845e-01
-9.39541757e-02 -5.57638705e-01 -6.81911647e-01 -3.70796561e-01
-2.31667578e-01 -1.79044809e-02 -1.59082964e-01 4.79025513e-01
8.63199770e-01 1.17778517e-01 9.11558926e-01 -8.09902728e-01
-1.08344126e+00 7.69180775e-01 -1.72792628e-01 1.43258467e-01
1.40908644e-01 7.39906311e-01 1.17337108e+00 -9.03069258e-01
1.18566549e+00 1.13743639e+00 8.62446070e-01 7.37311542e-01
-1.65092397e+00 -3.40348512e-01 -1.71238691e-01 3.16985510e-02
-1.17664492e+00 -4.72771794e-01 2.11187482e-01 -5.58823764e-01
2.16183448e+00 -1.45170182e-01 5.51763475e-01 1.31777930e+00
8.12124550e-01 2.63856769e-01 4.62710112e-01 4.94380668e-02
3.51390660e-01 -5.65543950e-01 2.95389116e-01 6.57299876e-01
1.66773453e-01 -1.43912777e-01 -6.34734929e-01 -9.02030885e-01
3.24923307e-01 1.39117405e-01 -4.22954038e-02 -6.25616372e-01
-8.83898556e-01 9.48295116e-01 -1.11681662e-01 -1.66171938e-01
-1.06021881e+00 2.19628707e-01 3.20612192e-01 2.70439446e-01
5.36714375e-01 7.95681119e-01 -7.95088291e-01 -6.34333551e-01
-5.37198246e-01 3.32313657e-01 8.19355488e-01 9.12500203e-01
3.64720136e-01 1.22590557e-01 2.25565910e-01 8.97103488e-01
3.61092448e-01 3.42196137e-01 6.30236447e-01 -7.07850695e-01
-1.43083319e-01 2.32729867e-01 1.59771487e-01 -1.36035442e-01
-5.27208984e-01 -9.96273458e-02 -4.97461200e-01 -3.68398093e-02
3.17509733e-02 -2.96144783e-01 -1.00529647e+00 1.55913293e+00
1.71691090e-01 -5.72162829e-02 5.34460366e-01 4.86254752e-01
8.10317278e-01 6.53781235e-01 5.50177872e-01 -3.11541826e-01
1.23164606e+00 -7.63494253e-01 -5.05634069e-01 2.01952346e-02
8.42232108e-01 -6.51353896e-01 8.54141653e-01 5.86528182e-01
-1.27352464e+00 -4.24454570e-01 -8.87357950e-01 -3.39336693e-01
-4.31200653e-01 -9.11047757e-02 9.10231292e-01 5.92877924e-01
-9.24703717e-01 8.48862052e-01 -1.16948652e+00 -2.82159984e-01
2.17989743e-01 4.21864182e-01 -3.09444368e-01 2.40860656e-01
-1.16888070e+00 9.18914914e-01 2.78323263e-01 -3.78201187e-01
-8.09256971e-01 -1.02121842e+00 -5.01351714e-01 2.63717949e-01
-5.43297417e-02 -9.90030944e-01 1.01704812e+00 -3.13272178e-01
-1.43376255e+00 9.62892354e-01 -4.21480149e-01 -1.02813470e+00
4.07265387e-02 -4.80655968e-01 -6.40912428e-02 -3.82332474e-01
-2.94960469e-01 5.87178946e-01 1.85987473e-01 -2.94562727e-01
1.63184166e-01 -3.99915874e-01 -4.32169229e-01 1.09682731e-01
3.10831368e-01 1.02580935e-01 1.19176067e-01 -5.51401138e-01
-2.28744656e-01 -1.07414639e+00 -6.18341684e-01 -6.25423253e-01
-5.34440391e-02 -3.97052884e-01 1.65793449e-01 -8.38906825e-01
1.21472704e+00 -1.83883917e+00 6.94099247e-01 -6.04096539e-02
2.52802014e-01 1.03130691e-01 -1.59643978e-01 6.83278739e-01
-5.07417977e-01 1.19413286e-01 -1.59091175e-01 1.32429495e-01
-5.21694683e-02 1.61595553e-01 -4.49063569e-01 2.91582942e-01
3.50731015e-01 1.28249598e+00 -6.13397539e-01 5.43866791e-02
-1.26137853e-01 8.01589489e-01 -8.27983379e-01 2.42714450e-01
-6.58168614e-01 2.05754176e-01 -4.40154791e-01 5.90257823e-01
3.13517243e-01 -8.92700791e-01 6.50353014e-01 3.81337060e-03
9.43459570e-02 4.88619059e-01 -1.43216550e-01 1.58078754e+00
2.82889396e-01 1.39112473e-01 -5.92886865e-01 -5.70236683e-01
1.09664714e+00 1.82799771e-01 8.12384546e-01 -2.66420171e-02
-9.86152142e-02 -6.10156283e-02 5.64898551e-01 -1.80614263e-01
3.35097849e-01 -5.89260161e-02 5.76739646e-02 3.86033744e-01
3.93139184e-01 1.29553944e-01 3.47457439e-01 1.34623528e-01
1.24881184e+00 6.90866172e-01 6.21781588e-01 -2.04890564e-01
-7.34474361e-02 2.85286009e-01 5.71862757e-01 5.93118608e-01
8.93670544e-02 3.20976526e-01 6.47157967e-01 -7.56980419e-01
-1.36590326e+00 -6.86406732e-01 -1.61391050e-01 1.60703754e+00
-6.29770815e-01 -8.90607059e-01 -7.07168818e-01 -3.46995324e-01
2.20020175e-01 7.07690716e-01 -5.13366640e-01 -5.66999793e-01
-5.30680597e-01 -1.34998810e+00 5.71338892e-01 4.79989618e-01
-3.19109082e-01 -1.27767909e+00 -5.79254389e-01 5.04962027e-01
-3.88803845e-03 -5.85071206e-01 -4.83401120e-01 7.69532144e-01
-8.49114776e-01 -8.79498839e-01 -8.91070902e-01 -5.59039831e-01
1.32910237e-01 -9.10232589e-02 1.53411055e+00 -2.16007322e-01
-5.74646115e-01 -4.52849343e-02 -9.61626694e-02 -4.24071938e-01
-7.50562370e-01 2.52167463e-01 1.87297374e-01 -7.85783470e-01
8.79823804e-01 -7.47577429e-01 -4.67622250e-01 3.62223297e-01
-4.74080145e-01 2.60951102e-01 4.90436107e-01 1.05638123e+00
1.12797344e+00 -6.13527834e-01 6.53579473e-01 -1.30214679e+00
7.27560222e-01 -7.19139695e-01 -7.64205337e-01 5.95782757e-01
-9.75079656e-01 5.14807522e-01 4.14508998e-01 -5.66558838e-01
-8.41614664e-01 3.68774205e-01 -5.25657356e-01 -3.57254148e-01
2.85325080e-01 7.21970081e-01 9.23877209e-02 1.34043336e-01
8.64648640e-01 4.50864553e-01 1.28475085e-01 -6.77341700e-01
4.95447665e-01 1.14580102e-01 2.12543905e-01 -3.88274193e-01
1.70179993e-01 -2.59995967e-01 1.47599101e-01 -5.27933836e-01
-2.05265447e-01 -2.71054775e-01 -3.89449716e-01 3.65552843e-01
7.64615119e-01 -1.01726437e+00 -1.05448294e+00 3.33783925e-01
-7.14874387e-01 -5.12288868e-01 -1.03934847e-01 1.17717572e-01
-1.19521677e+00 5.27552426e-01 -8.83462965e-01 -5.75398743e-01
-8.76085937e-01 -1.29759300e+00 1.08254719e+00 -2.91304849e-02
-6.12186611e-01 -6.96958482e-01 4.97157753e-01 1.48971349e-01
2.87089258e-01 -8.73575509e-02 1.15869224e+00 -9.25315082e-01
-3.64717305e-01 3.44988853e-01 4.26442832e-01 -1.18425779e-01
2.31046863e-02 2.71741450e-01 -3.83961648e-01 -1.26990438e-01
-2.57002622e-01 -3.92927378e-01 9.91041601e-01 8.07633817e-01
9.58243489e-01 4.29452024e-03 -5.41751862e-01 7.68477261e-01
9.95041728e-01 3.87981981e-01 7.49172986e-01 8.75003934e-02
4.14792985e-01 7.38039464e-02 7.45537102e-01 5.32373965e-01
-1.84728995e-01 5.52187085e-01 4.99063320e-02 -5.66945709e-02
1.24024093e-01 -3.80759776e-01 3.02251279e-01 4.89139378e-01
-4.99333471e-01 -5.37805021e-01 -1.02815413e+00 5.93824759e-02
-1.86152303e+00 -1.14775014e+00 7.35084638e-02 1.87487710e+00
1.09869123e+00 -5.63550629e-02 3.66361052e-01 -8.47402930e-01
2.77413219e-01 -1.47360399e-01 -1.24684858e+00 -3.74994248e-01
-4.45585310e-01 3.95812124e-01 6.52817607e-01 1.79654554e-01
-7.80380905e-01 1.08040643e+00 8.54138947e+00 6.34847820e-01
-6.54420972e-01 -1.33805662e-01 1.03607810e+00 -2.35435426e-01
-2.68198192e-01 1.43498212e-01 -1.19273806e+00 6.40543282e-01
1.74729013e+00 -4.64950681e-01 5.06560504e-01 1.27429032e+00
2.33812824e-01 4.01653528e-01 -9.07939374e-01 4.89029795e-01
-2.75626779e-01 -1.66114461e+00 -6.07337207e-02 5.90992212e-01
5.53691626e-01 6.52936757e-01 1.59247369e-01 3.87152553e-01
1.12736976e+00 -1.26626909e+00 -6.92852680e-03 9.18314517e-01
6.54808164e-01 -8.19637716e-01 4.74462688e-01 6.58497289e-02
-5.43232203e-01 3.26566249e-01 -8.29983115e-01 4.81578946e-01
2.37087399e-01 3.36500108e-01 -1.29511631e+00 1.32574484e-01
4.35843915e-01 9.19435322e-01 -4.13235635e-01 5.71718335e-01
2.00865269e-01 9.60362673e-01 -2.21539751e-01 -8.55627656e-02
-1.20760024e-01 -4.35874283e-01 2.66404271e-01 1.07119977e+00
1.62200019e-01 2.55334944e-01 -4.77043018e-02 8.93139601e-01
-1.58134088e-01 1.83503851e-01 -4.37697262e-01 -5.22858024e-01
2.80187428e-01 8.20681989e-01 -3.32513750e-01 -5.21628082e-01
-1.32955626e-01 8.98663461e-01 4.41501439e-01 4.83797193e-01
-1.07598853e+00 -9.93243679e-02 1.13367546e+00 5.12906015e-02
3.38546067e-01 -5.85907437e-02 2.72571981e-01 -9.61011052e-01
-4.79703158e-01 -1.27410400e+00 4.54793960e-01 -1.26969469e+00
-1.57906544e+00 3.71168196e-01 -2.55180538e-01 -5.42184591e-01
-8.84123921e-01 -8.81479442e-01 -4.57734913e-02 1.24219608e+00
-7.84387231e-01 -8.71429682e-01 4.77508157e-01 -2.14439645e-01
7.67847717e-01 -3.35117280e-01 1.12629116e+00 -1.04454823e-01
-5.86522996e-01 4.70531970e-01 6.78833842e-01 -6.14132166e-01
8.55145752e-01 -1.12520659e+00 1.43146908e+00 3.23640496e-01
-2.32010514e-01 1.38356268e+00 1.03602433e+00 -1.01326764e+00
-1.51315856e+00 -8.20487559e-01 8.62217724e-01 -6.28096938e-01
6.19287968e-01 -2.72887498e-01 -1.32689893e+00 9.43139791e-01
3.57421152e-02 -5.79679430e-01 1.13167512e+00 4.92832214e-01
-2.82485396e-01 6.28533781e-01 -9.94496524e-01 2.63470501e-01
1.28419733e+00 -3.50898772e-01 -3.43485683e-01 5.94143271e-01
1.10233736e+00 -2.90337205e-01 -1.50599599e+00 3.92039567e-01
8.09630334e-01 -6.23014033e-01 1.40278578e+00 -1.39534104e+00
4.49882120e-01 3.94361541e-02 5.77283353e-02 -1.08451152e+00
-9.20557439e-01 -9.02705491e-01 -3.91404867e-01 5.91899276e-01
7.58107483e-01 -8.22726786e-01 7.16573656e-01 6.17737055e-01
-2.80952960e-01 -9.58622098e-01 -6.29668951e-01 -5.19969642e-01
2.67604649e-01 3.70478034e-01 8.19189727e-01 5.69647014e-01
2.17378527e-01 6.12036943e-01 -6.13560319e-01 -4.67206389e-01
1.29772695e-02 -3.73678878e-02 7.92514205e-01 -1.20011270e+00
-1.05724037e+00 -3.52179825e-01 -1.65597036e-01 -1.11850858e+00
-3.84302773e-02 -7.18557358e-01 -1.55908212e-01 -1.14828098e+00
6.98249102e-01 1.88856304e-01 -3.46665829e-01 5.84384620e-01
-4.82818007e-01 -1.79644898e-01 -4.58077222e-01 2.97234356e-01
-4.87265348e-01 4.89536524e-01 6.87067926e-01 2.43792355e-01
-5.61584486e-03 -8.06381777e-02 -6.07873142e-01 2.94616640e-01
7.86403954e-01 -2.80769944e-01 -4.44295257e-01 8.29437748e-02
4.28255945e-01 -2.24322733e-02 -8.10676515e-02 -4.78130162e-01
-2.70916790e-01 -3.29355985e-01 4.70870465e-01 -7.97668695e-01
5.02672434e-01 -7.95087516e-02 1.01065600e+00 6.08443379e-01
-6.10146701e-01 5.90657473e-01 1.78986549e-01 6.45281732e-01
4.44812447e-01 2.04989966e-02 6.25567257e-01 -2.04809502e-01
-4.36995596e-01 1.36314601e-01 -5.92789590e-01 -2.28276223e-01
7.11647570e-01 -8.07159394e-02 -3.59529912e-01 -1.20081536e-01
-1.04486871e+00 -1.38950229e-01 7.88805187e-01 4.25374389e-01
3.87359202e-01 -6.31412804e-01 -6.70524955e-01 2.24788889e-01
2.30233654e-01 -7.05554545e-01 3.56469184e-01 5.38642287e-01
-4.24686372e-01 8.54204476e-01 -2.95574725e-01 -4.01027501e-01
-1.65512037e+00 9.27160442e-01 4.23664838e-01 -5.61478138e-01
-4.91469771e-01 7.36491501e-01 4.40199137e-01 -2.07569987e-01
-3.31711024e-01 -7.07842037e-02 -1.10090105e-02 -3.70883286e-01
2.99045503e-01 1.04584374e-01 -8.88788328e-02 -5.07726848e-01
-1.39761940e-01 1.03133418e-01 -3.68693143e-01 3.75187218e-01
1.68814099e+00 1.61519632e-01 -4.03830335e-02 3.02046984e-01
7.66597748e-01 -4.94687974e-01 -1.18212295e+00 2.57129893e-02
-1.95933003e-02 2.15807393e-01 -4.07873392e-01 -1.00107074e+00
-3.35524112e-01 4.38298613e-01 6.79914534e-01 -3.92625570e-01
7.74391592e-01 1.19204476e-01 5.74896276e-01 6.44562364e-01
4.06848222e-01 -6.34120345e-01 -2.11451739e-01 5.97993553e-01
4.63782191e-01 -8.83333504e-01 1.10842578e-01 -8.17095414e-02
-7.25217760e-01 6.62737548e-01 2.49174401e-01 3.27981971e-02
2.11274728e-01 2.30242088e-01 -3.43839824e-01 -4.79538381e-01
-1.51512873e+00 5.64960465e-02 1.51160881e-01 6.83183968e-01
1.14635992e+00 1.83253437e-01 -4.10502464e-01 7.36217797e-01
-2.61473358e-01 2.31990125e-02 -1.87740382e-02 7.26353765e-01
-6.87081993e-01 -1.34728336e+00 9.56449211e-02 4.51427370e-01
-6.49578214e-01 -6.43156648e-01 -7.62463331e-01 2.51738638e-01
-5.48254371e-01 5.70051610e-01 -3.54916990e-01 -3.69967846e-03
1.77251294e-01 7.88908243e-01 3.92802894e-01 -3.99036288e-01
-5.94010115e-01 2.97690153e-01 2.14364603e-01 -3.44698071e-01
-4.17087004e-02 -8.13999414e-01 -1.34326398e+00 -3.03329974e-01
-2.98198283e-01 3.72649342e-01 5.19719422e-01 2.45990008e-01
1.19908857e+00 6.16696835e-01 4.14023474e-02 -3.04467857e-01
-7.65818417e-01 -1.27621710e+00 -1.97816253e-01 2.41027951e-01
-3.28741729e-01 -4.58295047e-01 1.21455081e-01 5.07374406e-01] | [4.700778484344482, 5.6249189376831055] |
536d6498-a603-4210-b4ff-60a19758b7bc | context-patch-face-hallucination-based-on | 1809.00665 | null | http://arxiv.org/abs/1809.00665v2 | http://arxiv.org/pdf/1809.00665v2.pdf | Context-Patch Face Hallucination Based on Thresholding Locality-constrained Representation and Reproducing Learning | Face hallucination is a technique that reconstruct high-resolution (HR) faces
from low-resolution (LR) faces, by using the prior knowledge learned from HR/LR
face pairs. Most state-of-the-arts leverage position-patch prior knowledge of
human face to estimate the optimal representation coefficients for each image
patch. However, they focus only the position information and usually ignore the
context information of image patch. In addition, when they are confronted with
misalignment or the Small Sample Size (SSS) problem, the hallucination
performance is very poor. To this end, this study incorporates the contextual
information of image patch and proposes a powerful and efficient context-patch
based face hallucination approach, namely Thresholding Locality-constrained
Representation and Reproducing learning (TLcR-RL). Under the context-patch
based framework, we advance a thresholding based representation method to
enhance the reconstruction accuracy and reduce the computational complexity. To
further improve the performance of the proposed algorithm, we propose a
promotion strategy called reproducing learning. By adding the estimated HR face
to the training set, which can simulates the case that the HR version of the
input LR face is present in the training set, thus iteratively enhancing the
final hallucination result. Experiments demonstrate that the proposed TLcR-RL
method achieves a substantial increase in the hallucinated results, both
subjectively and objectively. Additionally, the proposed framework is more
robust to face misalignment and the SSS problem, and its hallucinated HR face
is still very good when the LR test face is from the real-world. The MATLAB
source code is available at https://github.com/junjun-jiang/TLcR-RL | ['Suhua Tang', 'Yi Yu', 'Junjun Jiang', 'Akiko Aizawa', 'Jiayi Ma', 'Kiyoharu Aizawa'] | 2018-09-03 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 1.78069666e-01 6.33110991e-04 -6.55450672e-02 -8.29030126e-02
-8.05836141e-01 -4.19481285e-03 2.86565632e-01 -6.31824791e-01
1.99025311e-03 6.74206793e-01 3.55229318e-01 3.18939060e-01
5.13323210e-02 -7.09833264e-01 -6.30943477e-01 -9.32183444e-01
4.13384944e-01 -1.82560325e-01 -2.56679446e-01 -3.12863514e-02
1.18224144e-01 5.25177300e-01 -1.78186035e+00 2.52139449e-01
8.94010723e-01 1.07588863e+00 5.24524271e-01 3.02537978e-02
2.50220537e-01 5.86597323e-01 -3.86115640e-01 -1.88205913e-01
4.16575164e-01 -5.12343526e-01 -1.53435722e-01 3.85542959e-01
3.96399796e-01 -3.93905550e-01 -4.08566356e-01 1.16244793e+00
7.54262388e-01 8.49902481e-02 3.80276471e-01 -9.61036146e-01
-6.95568144e-01 -8.86474997e-02 -1.10769081e+00 2.90238461e-03
5.28447628e-01 6.10084534e-02 3.37520868e-01 -1.49228334e+00
3.86350960e-01 1.46404862e+00 5.97430110e-01 3.57203275e-01
-1.07423198e+00 -1.01923609e+00 -1.59535393e-01 2.08259270e-01
-1.85764027e+00 -9.27725255e-01 9.86103117e-01 -2.06294507e-01
3.30401897e-01 9.15340260e-02 3.30261856e-01 9.18742120e-01
-1.30505208e-02 3.95097584e-01 1.46223283e+00 -3.95684481e-01
-5.37296273e-02 4.11327958e-01 -4.04642522e-01 7.50637114e-01
1.12665728e-01 2.95345038e-01 -4.70297068e-01 -1.70424372e-01
1.12274861e+00 2.38485292e-01 -7.42952347e-01 -8.22895989e-02
-8.88573527e-01 5.50328016e-01 3.86418462e-01 3.13026279e-01
-4.52396929e-01 -4.26691353e-01 -3.03916670e-02 2.40832001e-01
4.10575241e-01 5.41000143e-02 7.97938555e-02 4.41812903e-01
-1.10679770e+00 -4.47184555e-02 4.19770926e-01 8.36263478e-01
9.77920294e-01 1.95240885e-01 -1.81481704e-01 1.24438274e+00
3.44166964e-01 5.18027306e-01 4.81310189e-01 -1.05460715e+00
2.47137502e-01 2.16443837e-01 1.72480211e-01 -1.18239987e+00
5.67451157e-02 -5.35418868e-01 -1.08885717e+00 1.84823230e-01
4.54283739e-03 6.81006685e-02 -8.18443239e-01 1.85063052e+00
5.59679687e-01 5.79288542e-01 1.22378558e-01 1.13246715e+00
1.04275155e+00 6.81913376e-01 -3.33040327e-01 -1.01153672e+00
1.31058288e+00 -7.91967571e-01 -1.00225163e+00 -6.84890747e-02
-1.40804306e-01 -1.10974967e+00 1.02541018e+00 4.61469620e-01
-1.08056247e+00 -9.96187031e-01 -9.97931778e-01 7.26974234e-02
2.01117814e-01 5.30964494e-01 1.18244976e-01 4.74172622e-01
-9.02934432e-01 3.17140400e-01 -3.32252979e-01 -2.67816186e-01
5.38325667e-01 1.79933384e-01 -5.99876046e-01 -6.40990555e-01
-1.01289690e+00 7.73746133e-01 6.00906014e-02 3.31799209e-01
-9.25980091e-01 -6.03581905e-01 -7.54979372e-01 -3.33811864e-02
5.22190511e-01 -4.08015132e-01 8.33837330e-01 -9.26324606e-01
-1.44733024e+00 5.97269952e-01 -5.75945795e-01 1.90768436e-01
3.78938794e-01 -3.58136147e-02 -6.74969852e-01 3.46970230e-01
1.53406590e-01 3.60443741e-01 1.38484001e+00 -1.73516417e+00
-1.32543340e-01 -6.28433883e-01 -4.02868062e-01 4.17988777e-01
-1.75597459e-01 -8.86361524e-02 -5.37305951e-01 -6.65767133e-01
2.74140865e-01 -5.87850213e-01 9.86877233e-02 2.19713878e-02
-2.20872581e-01 1.67442441e-01 1.01572847e+00 -8.54535997e-01
1.01626956e+00 -2.58332372e+00 -1.29544973e-01 1.18883826e-01
8.86574313e-02 3.28135461e-01 -2.25513905e-01 2.85911947e-01
-2.84078330e-01 -1.59998968e-01 -2.17562243e-01 -3.67283434e-01
-4.55397427e-01 8.19940343e-02 -3.24370056e-01 7.00103879e-01
1.11902140e-01 5.42304099e-01 -5.34843028e-01 -5.92080653e-01
2.24413857e-01 8.33080649e-01 -3.34391326e-01 5.73072731e-01
2.73551375e-01 8.49840224e-01 -2.78017431e-01 7.71323204e-01
1.23732531e+00 -4.17980216e-02 6.57192916e-02 -6.65989816e-01
-1.08426921e-01 -4.49242294e-01 -1.33873439e+00 1.44842374e+00
-5.08680105e-01 1.80535793e-01 2.94049829e-01 -7.38157690e-01
1.25077295e+00 5.14830709e-01 4.01789635e-01 -7.59857655e-01
-6.96185529e-02 1.74409926e-01 -2.26300672e-01 -6.78201914e-01
1.66696817e-01 -4.73968506e-01 5.40639341e-01 2.11476266e-01
-8.86844844e-02 1.49941295e-01 -2.99301803e-01 -2.07302123e-01
6.70143485e-01 1.01339266e-01 5.75949132e-01 1.30496338e-01
7.24905312e-01 -6.91606820e-01 7.17145860e-01 2.50429839e-01
-7.44828358e-02 9.57782030e-01 1.40357241e-01 1.86794270e-02
-8.92022848e-01 -1.04897070e+00 -4.12996650e-01 6.38382196e-01
2.93847114e-01 -1.55245095e-01 -6.15206599e-01 -2.51343668e-01
-1.82799041e-01 4.68216211e-01 -4.44568038e-01 -2.09617570e-01
-3.99473280e-01 -6.13312662e-01 2.42908970e-01 2.30961785e-01
9.33191419e-01 -1.08994317e+00 -7.84781277e-02 -1.49293616e-01
-3.94599110e-01 -1.06502867e+00 -5.49336314e-01 -6.36167586e-01
-6.21533632e-01 -8.75302374e-01 -9.05503929e-01 -7.81310022e-01
9.46692586e-01 5.36965072e-01 5.66934586e-01 1.57803833e-01
-1.77880272e-01 2.32212052e-01 -2.75167257e-01 1.58663988e-01
-4.94157858e-02 -6.22142136e-01 1.10053316e-01 6.26074016e-01
2.26940755e-02 -7.32890844e-01 -7.17862248e-01 5.54293931e-01
-7.52104640e-01 -4.94996179e-03 9.73710358e-01 1.00976217e+00
7.74770558e-01 4.76468414e-01 7.76468813e-01 -6.25989616e-01
4.77398962e-01 -5.41420877e-01 -2.66288608e-01 1.12262577e-01
-5.92272282e-01 -3.07993472e-01 6.90869153e-01 -5.53925693e-01
-1.48481655e+00 4.86350134e-02 -8.61657858e-02 -9.46458161e-01
-1.46891341e-01 3.98207217e-01 -5.62363148e-01 -3.13055605e-01
5.04023969e-01 4.47922707e-01 3.23526621e-01 -6.21224523e-01
8.97080358e-03 8.50670695e-01 6.49256229e-01 -4.23030823e-01
8.53206635e-01 5.06555855e-01 -5.93535416e-02 -9.32137311e-01
-7.22608387e-01 -4.03976440e-01 -3.13989669e-01 -3.33833277e-01
4.52102244e-01 -1.13670337e+00 -6.45879030e-01 3.73807073e-01
-9.35704827e-01 3.87271866e-02 -2.37969030e-02 5.38129687e-01
-5.19420981e-01 6.07103765e-01 -3.75024617e-01 -1.05462134e+00
-1.51737526e-01 -1.16973186e+00 1.06138623e+00 4.31043446e-01
3.06420058e-01 -4.76913303e-01 -1.85904533e-01 5.81864715e-01
3.27591956e-01 2.30792031e-01 6.57305896e-01 -4.51471061e-02
-5.19964337e-01 -6.07629307e-02 -4.41568494e-01 4.42885667e-01
2.53062308e-01 -4.06803757e-01 -1.15939879e+00 -6.10387444e-01
4.62641001e-01 -2.76433617e-01 6.31983638e-01 2.88763672e-01
1.21079838e+00 -5.04342139e-01 -6.32829070e-02 5.57164967e-01
1.54842603e+00 7.84640387e-02 1.03333783e+00 -1.97926730e-01
6.58277929e-01 7.55081832e-01 9.71170127e-01 6.61163211e-01
1.65029377e-01 9.14579690e-01 2.61586905e-01 -2.85246372e-01
-4.04868424e-01 -5.00457942e-01 5.80740929e-01 5.95866740e-01
-2.62188911e-01 8.59795436e-02 -3.39438677e-01 3.22293073e-01
-1.69338477e+00 -1.13217926e+00 2.86201715e-01 2.57154417e+00
7.47038007e-01 -4.60858524e-01 -1.85583159e-01 1.59705907e-01
1.03986609e+00 1.49226651e-01 -5.64382136e-01 9.70058069e-02
-8.90443623e-02 1.38776317e-01 -6.30610585e-02 4.69142884e-01
-6.46532714e-01 7.68810391e-01 5.10582829e+00 1.07955098e+00
-1.13742387e+00 3.02614748e-01 6.28970742e-01 -3.44138616e-03
-1.11847468e-01 -1.07809253e-01 -6.68930292e-01 4.68870252e-01
4.79019701e-01 -4.86456417e-02 7.44904578e-01 6.02495074e-01
5.40317535e-01 -2.15763018e-01 -8.03150594e-01 1.41759896e+00
4.90309387e-01 -7.67648399e-01 -2.04538498e-02 1.77128434e-01
6.01881981e-01 -5.62721610e-01 4.43879545e-01 2.75989532e-01
-5.20411909e-01 -1.29233146e+00 3.28504235e-01 7.72462368e-01
1.28385854e+00 -9.18386221e-01 7.71144629e-01 3.45237374e-01
-1.28779697e+00 -1.49110198e-01 -6.12056851e-01 2.09157586e-01
3.16313393e-02 6.32415652e-01 -6.41030252e-01 7.39294767e-01
6.73527598e-01 7.49618530e-01 -4.92633045e-01 8.29245746e-01
-2.37695575e-01 3.18664759e-01 -4.22172099e-02 8.44589114e-01
-3.92667145e-01 -3.29390407e-01 7.18645692e-01 7.17545152e-01
6.16297007e-01 4.41426665e-01 2.67841280e-01 9.30698633e-01
-5.89836426e-02 2.49764964e-01 -6.79980993e-01 2.87329882e-01
5.97695768e-01 1.45711422e+00 -2.55994171e-01 -1.61788762e-02
-4.43201214e-01 9.42416131e-01 1.69422224e-01 6.81800187e-01
-7.55783021e-01 -1.16647258e-01 1.95479348e-01 4.47840929e-01
4.45991755e-01 1.15551926e-01 -7.27096274e-02 -1.11273611e+00
1.73922479e-01 -1.16699111e+00 1.12985402e-01 -1.05712569e+00
-1.28448260e+00 5.93245208e-01 -5.79421669e-02 -1.44930446e+00
-2.13871803e-02 -1.23260960e-01 -5.54757595e-01 8.96954715e-01
-1.54802394e+00 -1.23550093e+00 -6.39041543e-01 9.02136445e-01
4.63970155e-01 -3.02162111e-01 5.77866375e-01 3.99113417e-01
-6.20093584e-01 7.66666651e-01 -5.54460287e-02 -1.21472821e-01
9.78073895e-01 -5.37780046e-01 -6.03199363e-01 7.25245655e-01
-1.31577343e-01 7.89629042e-01 6.57197237e-01 -6.95161343e-01
-1.35089171e+00 -1.21521580e+00 4.24160868e-01 1.36434436e-01
-4.18110155e-02 -1.44356042e-01 -1.13662207e+00 4.20304120e-01
-3.15497257e-02 3.47685874e-01 6.35405779e-01 -1.26964629e-01
-3.08052301e-01 -3.70918989e-01 -1.60502851e+00 5.75753510e-01
8.41086924e-01 -6.32645309e-01 -4.75208819e-01 1.20206162e-01
3.92256558e-01 -2.36891612e-01 -1.10010457e+00 6.76998496e-01
6.51689649e-01 -1.24770045e+00 9.33710575e-01 2.67056614e-01
4.01466876e-01 -7.02380240e-01 -3.50200385e-01 -1.02436280e+00
-2.55731881e-01 -4.39558536e-01 -1.83153152e-01 1.37642252e+00
-2.75041405e-02 -7.32143521e-01 4.37428921e-01 1.51342347e-01
-8.64792243e-03 -8.44569921e-01 -9.00215387e-01 -6.75170064e-01
-4.56894785e-01 1.28288716e-01 5.20551801e-01 9.80217814e-01
-1.88401133e-01 2.97466397e-01 -7.22204626e-01 5.05559206e-01
9.03519154e-01 2.05477685e-01 7.64632940e-01 -8.87116671e-01
-4.45535362e-01 1.61947772e-01 -2.04277828e-01 -7.92793989e-01
2.32154101e-01 -5.79104722e-01 3.61395702e-02 -1.24114072e+00
3.75600040e-01 -2.08384171e-01 -2.26574630e-01 4.40719187e-01
-1.61455914e-01 6.01329327e-01 1.01380982e-01 5.93066275e-01
-2.46521026e-01 8.61308575e-01 1.63259041e+00 2.95783043e-01
-1.88856602e-01 -1.71524566e-02 -7.71712244e-01 5.56734085e-01
5.74715972e-01 -2.63817608e-01 -6.02122366e-01 1.59886777e-02
-3.71845901e-01 5.51849186e-01 5.51348150e-01 -9.91789043e-01
1.99308917e-01 -9.10964310e-02 7.66939163e-01 -4.77310956e-01
7.37416089e-01 -7.47480512e-01 5.62945127e-01 1.94286138e-01
-8.45030509e-03 -3.22528511e-01 4.82147969e-02 6.80072486e-01
-4.44474846e-01 -7.92844072e-02 1.28680134e+00 -2.25817576e-01
-4.30191576e-01 4.70060885e-01 1.16755247e-01 -3.29088420e-01
1.02369249e+00 -4.68058795e-01 -2.45343402e-01 -5.78506052e-01
-5.87260008e-01 -1.73023283e-01 6.34962976e-01 1.61042392e-01
1.06130624e+00 -1.50264108e+00 -7.55516171e-01 6.21452153e-01
3.98676656e-02 -2.04291806e-01 5.62659085e-01 1.15835166e+00
4.60686199e-02 5.08243889e-02 -3.74051452e-01 -3.69456708e-01
-1.31672931e+00 6.54296339e-01 2.02358529e-01 6.77768663e-02
-5.30549347e-01 4.82757151e-01 5.08869827e-01 -2.13459879e-01
1.69825908e-02 4.64890331e-01 -4.83141810e-01 -1.22792274e-01
6.82862103e-01 4.02209729e-01 -1.59025356e-01 -1.01363254e+00
-9.89784226e-02 8.26184928e-01 -1.65535137e-01 -1.99746192e-01
1.14482903e+00 -3.37609351e-01 -2.32548833e-01 1.88946187e-01
1.13637209e+00 3.63658726e-01 -1.11476958e+00 -3.43908668e-01
-6.30728900e-01 -1.02830732e+00 2.06774503e-01 -5.92854321e-01
-1.29688454e+00 6.82617426e-01 8.43747437e-01 -3.14796180e-01
1.56293201e+00 -1.28698424e-01 6.34840310e-01 -3.91945131e-02
4.75291550e-01 -7.37412930e-01 4.12997186e-01 -2.31699869e-02
1.26477373e+00 -1.24417484e+00 2.70946324e-01 -7.15336859e-01
-6.95104420e-01 8.88751924e-01 8.00064206e-01 -2.49114167e-02
5.78953326e-01 6.04415275e-02 -1.11872926e-01 -5.52046336e-02
-4.97820407e-01 -1.95084155e-01 1.45151496e-01 6.11710429e-01
3.12200606e-01 -2.01554433e-01 -2.62846917e-01 5.57373166e-01
-9.92474407e-02 6.73265308e-02 3.83860707e-01 4.85588491e-01
-4.64110136e-01 -8.77018690e-01 -7.16196418e-01 2.61457115e-01
-3.99597079e-01 -6.29278198e-02 2.75486335e-02 4.54182714e-01
3.40427011e-01 1.12570822e+00 -2.25501493e-01 -3.81164402e-01
2.37629652e-01 -2.27277562e-01 6.17561519e-01 -6.51392758e-01
-3.80971655e-02 5.50659478e-01 -2.37367585e-01 -5.15688598e-01
-4.03931022e-01 -4.88394767e-01 -1.12705958e+00 -2.91646898e-01
-4.21777755e-01 -3.03104008e-03 2.80425042e-01 7.20472634e-01
4.25169826e-01 5.55012636e-02 1.13398147e+00 -9.48562205e-01
-4.57087815e-01 -1.01003337e+00 -8.49353671e-01 2.74222910e-01
4.23746616e-01 -8.94869924e-01 -5.72359324e-01 -4.98883948e-02] | [12.844440460205078, -0.0016654033679515123] |
b8a112d9-3402-4713-88bf-9da591136132 | response-to-significance-and-stability-of | 2206.04934 | null | https://arxiv.org/abs/2206.04934v1 | https://arxiv.org/pdf/2206.04934v1.pdf | Response to: Significance and stability of deep learning-based identification of subtypes within major psychiatric disorders. Molecular Psychiatry (2022) | Recently, Winter and Hahn [1] commented on our work on identifying subtypes of major psychiatry disorders (MPDs) based on neurobiological features using machine learning [2]. They questioned the generalizability of our methods and the statistical significance, stability, and overfitting of the results, and proposed a pipeline for disease subtyping. We appreciate their earnest consideration of our work, however, we need to point out their misconceptions of basic machine-learning concepts and delineate some key issues involved. | ['Weixiong Zhang', 'Fei Wang', 'Xizhe Zhang'] | 2022-06-10 | null | null | null | null | ['misconceptions'] | ['miscellaneous'] | [ 2.17992872e-01 2.93818712e-01 -3.93895388e-01 -6.05855644e-01
-5.06592095e-01 -4.11794305e-01 3.08653563e-01 5.18840611e-01
-5.40146172e-01 8.66365969e-01 1.22450195e-01 -4.87588584e-01
-5.58673799e-01 -1.57929555e-01 -8.85233805e-02 -3.19370657e-01
-5.17606378e-01 6.72742426e-01 -1.46494687e-01 1.76876299e-02
6.42845809e-01 8.51554334e-01 -9.59360003e-01 3.79502714e-01
8.30413699e-01 4.76764500e-01 -1.25318304e-01 2.54974157e-01
1.57714605e-01 5.41451693e-01 -7.61016250e-01 -8.14244032e-01
-4.10963178e-01 -3.53502259e-02 -1.10088718e+00 -6.46906495e-02
1.60253927e-01 -4.03010786e-01 -1.47266194e-01 8.98243487e-01
5.01566648e-01 -1.49963126e-01 9.37252343e-01 -1.33845758e+00
-6.20671034e-01 2.41239384e-01 -4.71723974e-01 7.45043099e-01
2.77865559e-01 1.24322385e-01 1.06508422e+00 -6.73927844e-01
7.26296425e-01 1.13274872e+00 9.50958550e-01 8.47295403e-01
-1.29179537e+00 -8.98738146e-01 4.22532886e-01 4.56105739e-01
-1.47298491e+00 -5.67354560e-01 1.01739630e-01 -8.92632306e-01
1.19903672e+00 1.44037634e-01 9.70149577e-01 1.03846908e+00
5.48169911e-01 3.90289336e-01 1.09284103e+00 6.85588196e-02
3.80210012e-01 7.34777749e-02 2.73609877e-01 5.13428330e-01
3.57233018e-01 -4.32930350e-01 -7.46202394e-02 -9.77384031e-01
7.62979507e-01 2.34991983e-02 -9.13856830e-03 -1.99963208e-02
-5.99784791e-01 8.81860077e-01 -9.25425347e-03 6.40076280e-01
-1.01752326e-01 -4.08128977e-01 7.38317132e-01 -2.14750543e-01
3.86849433e-01 2.50553846e-01 -4.59517479e-01 -3.28193933e-01
-1.06180203e+00 8.90339240e-02 2.75475025e-01 3.06789428e-01
-1.76892594e-01 -1.71767920e-01 3.40951979e-01 1.19269633e+00
1.77590728e-01 1.91104919e-01 6.69302225e-01 -8.86680543e-01
2.78703012e-02 4.66757357e-01 -1.64951429e-01 -1.02040493e+00
-1.15089869e+00 -2.33341634e-01 -8.36994946e-01 -1.75412357e-01
-1.23065270e-01 -3.54641348e-01 -4.43985999e-01 1.64037180e+00
-3.45522702e-01 -1.17223434e-01 -2.22295836e-01 4.14452463e-01
5.80712855e-01 -2.97450155e-01 6.78506732e-01 -3.45282376e-01
1.17882228e+00 -3.04909855e-01 -5.51759064e-01 -4.67705876e-01
7.68933952e-01 -3.16462934e-01 5.11271834e-01 4.44904804e-01
-1.25287855e+00 -5.75711057e-02 -6.88881695e-01 4.84226132e-03
-1.90702304e-01 1.08485229e-01 1.07964885e+00 1.03622878e+00
-1.18304205e+00 6.88367069e-01 -1.11406982e+00 -9.30498242e-01
6.98914349e-01 7.13744164e-01 -6.56499982e-01 3.20702106e-01
-7.70741701e-01 1.10530043e+00 2.34063655e-01 9.43253860e-02
-3.37895274e-01 -6.55031979e-01 -3.31608444e-01 -4.76184376e-02
-1.40930250e-01 -6.89037204e-01 8.32915246e-01 -5.06298840e-01
-9.27400053e-01 1.36531305e+00 -2.05103844e-01 -1.78079516e-01
-1.27771229e-01 2.94370174e-01 -1.09968579e+00 5.14980972e-01
-1.60131685e-03 4.12921429e-01 1.64481446e-01 -5.87569058e-01
-6.82964802e-01 -7.95112669e-01 -9.33062006e-03 9.28320140e-02
-1.31105483e-01 3.42350692e-01 3.73810619e-01 -3.91032338e-01
2.38921508e-01 -7.30223179e-01 -3.31600666e-01 -6.83127567e-02
-3.24123679e-03 -9.46663618e-02 2.75689751e-01 -5.66848040e-01
1.08029187e+00 -2.07285571e+00 -3.43680650e-01 6.94533437e-02
7.26633132e-01 1.07840449e-01 4.41413149e-02 4.90822554e-01
-5.85449576e-01 6.33674622e-01 -1.24505922e-01 1.90892458e-01
2.15448048e-02 -9.31735113e-02 -1.91393450e-01 8.01661611e-01
6.37031794e-01 9.75563288e-01 -1.07774568e+00 -1.55911729e-01
1.39857396e-01 2.84513712e-01 -7.50576794e-01 -2.97831595e-01
5.46257079e-01 2.38249883e-01 -2.73534775e-01 6.78189099e-01
7.32364058e-01 -3.62745136e-01 7.57122815e-01 -9.95072350e-02
-4.85280231e-02 6.87408030e-01 -6.95536852e-01 1.05014110e+00
4.15739089e-01 3.86613905e-01 2.42777258e-01 -1.11076903e+00
5.88840961e-01 4.74280298e-01 7.92295814e-01 -3.20032477e-01
-1.28014252e-01 3.79549235e-01 5.04106700e-01 -5.14965057e-01
3.84690985e-03 -6.49028242e-01 2.50830919e-01 3.55762601e-01
7.15178996e-03 7.14972615e-02 -2.25592718e-01 2.59727072e-02
1.30459881e+00 -2.71854132e-01 5.96601009e-01 -4.71031159e-01
1.21049225e-01 -2.10600540e-01 5.18947184e-01 5.82957149e-01
-7.54099250e-01 5.89086711e-01 9.63094532e-01 -3.72077882e-01
-5.28228998e-01 -9.17560995e-01 -8.45570564e-01 4.98223692e-01
-6.85879588e-01 -4.53341752e-01 -4.26869422e-01 -7.01877892e-01
-1.07633621e-01 6.74098253e-01 -8.60875309e-01 -3.33537430e-01
-5.28722703e-02 -1.64682794e+00 9.72706378e-01 5.44833958e-01
-3.18082958e-01 -5.36366761e-01 -4.65981752e-01 -5.17116189e-02
1.59552976e-01 -9.84184206e-01 1.93456665e-01 2.19607070e-01
-1.08100283e+00 -1.27997673e+00 -5.66698551e-01 -5.74922264e-01
5.69171607e-01 1.40662640e-01 6.70169950e-01 1.60875246e-01
-5.35645366e-01 3.66625696e-01 -1.59348667e-01 -1.88970432e-01
-1.88107789e-01 2.35047247e-02 4.42288399e-01 -5.73158979e-01
1.03933036e+00 -8.98055315e-01 -6.53806865e-01 -8.92609358e-02
-7.10671186e-01 -5.73327541e-01 5.80716968e-01 4.04184401e-01
5.02263904e-02 -3.07534784e-01 7.71222711e-01 -9.81413603e-01
7.99575388e-01 -9.43586528e-01 9.59083810e-02 -4.10246477e-02
-7.96342254e-01 -5.65083444e-01 -5.33878058e-02 -2.12147743e-01
-4.37424541e-01 -6.65865123e-01 -4.66718614e-01 1.86009213e-01
-4.30040419e-01 2.25198492e-01 2.28562966e-01 -1.43412217e-01
5.15408099e-01 -1.73372790e-01 -1.32286042e-01 -3.51568192e-01
-2.79273331e-01 6.74472272e-01 -5.51503412e-02 -3.33193690e-01
4.21058834e-02 6.80563092e-01 -7.21301064e-02 -9.99719381e-01
-3.19090813e-01 -3.90439123e-01 -6.62441015e-01 2.50941545e-01
7.38329291e-01 -6.71861708e-01 -1.12121093e+00 1.35689378e-01
-9.43560302e-01 6.36543557e-02 2.61629999e-01 6.76217794e-01
-3.78237993e-01 5.26503444e-01 -6.70144737e-01 -8.01162481e-01
-3.22193325e-01 -1.21032143e+00 7.71116853e-01 -2.16167346e-01
-9.29048896e-01 -1.38500428e+00 3.82336348e-01 2.82067209e-01
1.58634797e-01 1.34785131e-01 1.58008826e+00 -9.04347539e-01
6.48181736e-01 -2.19993234e-01 -5.03481090e-01 -2.06501424e-01
4.28100616e-01 2.49844402e-01 -1.14414871e+00 -9.22846943e-02
8.33912343e-02 -3.73184174e-01 3.87147129e-01 5.66258430e-01
1.08620334e+00 -1.36647582e-01 -4.60769475e-01 2.08582595e-01
1.10947442e+00 4.07614112e-01 5.44906914e-01 3.98880512e-01
2.04393804e-01 7.76415169e-01 -7.50350207e-02 4.58860815e-01
4.30693686e-01 4.92783874e-01 1.17745064e-01 -1.13381252e-01
4.09641325e-01 1.05170034e-01 2.88671225e-01 4.76798236e-01
-1.67264014e-01 2.31267169e-01 -9.32826459e-01 3.76327455e-01
-1.47759151e+00 -9.21352565e-01 -2.79835612e-01 1.99614823e+00
5.25591493e-01 2.85694122e-01 6.60737097e-01 -2.24828750e-01
6.79076731e-01 6.23008311e-02 -5.38062334e-01 -7.92411149e-01
-1.66188449e-01 9.51040089e-02 2.33383730e-01 3.00509632e-01
-5.49088001e-01 7.55394638e-01 9.43260002e+00 5.50933301e-01
-1.05683410e+00 1.34281784e-01 8.07291031e-01 -2.51548797e-01
-1.43211722e-01 -1.60263315e-01 -7.63028800e-01 4.15280849e-01
1.20648897e+00 -1.28037974e-01 1.62708610e-01 3.76641184e-01
4.58926976e-01 -1.69438094e-01 -9.37821090e-01 6.97885573e-01
6.85491040e-02 -1.09347737e+00 -1.62837178e-01 1.85811266e-01
1.86879754e-01 2.92669415e-01 1.53655097e-01 8.93812999e-02
1.44676819e-01 -1.26332283e+00 2.08756216e-02 5.54256678e-01
7.04735816e-01 -4.80141819e-01 6.11074388e-01 -1.65228784e-01
-4.23959643e-01 -6.91337064e-02 -5.52707076e-01 -3.76673639e-01
1.59728248e-02 6.49140179e-01 -1.02877128e+00 9.06076506e-02
4.08584654e-01 5.17379284e-01 -4.86684889e-01 9.44796145e-01
1.00314043e-01 6.24611199e-01 -7.92763010e-02 2.83356011e-01
2.55314231e-01 -1.27422318e-01 1.16279297e-01 1.18517208e+00
-3.64830680e-02 3.78416747e-01 -2.96146691e-01 8.53987873e-01
3.87713343e-01 2.57128716e-01 -3.64844561e-01 -8.45847487e-01
2.27551669e-01 1.43285608e+00 -1.04068458e+00 -1.06923781e-01
-6.27986848e-01 7.06049681e-01 3.46274555e-01 7.64283687e-02
-2.95906544e-01 -2.46824905e-01 7.14201987e-01 2.14594707e-01
-8.72775614e-02 7.74213001e-02 -6.70993388e-01 -1.00342858e+00
-9.06193256e-02 -8.13771129e-01 5.90599000e-01 -6.18018985e-01
-1.24841917e+00 3.75955492e-01 -6.00678138e-02 -7.20827162e-01
-4.74374741e-03 -7.37183332e-01 -5.09016275e-01 7.70615041e-01
-8.65533888e-01 -8.98930132e-01 4.91865307e-01 6.75617382e-02
-2.15857644e-02 1.05441324e-02 1.12436187e+00 1.77340582e-01
-7.97690988e-01 4.59807396e-01 2.11052835e-01 -9.79532227e-02
8.81363571e-01 -8.65678787e-01 2.47683793e-01 -1.74666554e-01
-7.16462791e-01 1.04403174e+00 4.43461388e-01 -7.58714736e-01
-8.45945477e-01 -5.81905603e-01 1.20034683e+00 -5.91629744e-01
1.12946808e+00 -2.64523566e-01 -7.15792239e-01 1.08901680e+00
1.05161875e-01 -6.80876315e-01 1.50691628e+00 6.49666548e-01
-1.46949872e-01 2.55141407e-01 -1.58667314e+00 7.70561814e-01
8.99688601e-01 -3.17248672e-01 -7.18171477e-01 6.66760921e-01
-2.37215925e-02 1.47951782e-01 -1.24557972e+00 3.67393851e-01
8.98630798e-01 -1.03798497e+00 9.08753633e-01 -1.01315820e+00
2.72366792e-01 6.31575823e-01 2.01426372e-01 -1.03735113e+00
-7.01211751e-01 -1.78136230e-01 6.39822900e-01 7.84975708e-01
3.23473245e-01 -7.88354516e-01 5.94866455e-01 1.02997470e+00
7.82975331e-02 -9.56107438e-01 -9.16621327e-01 -3.71351928e-01
6.20869219e-01 -4.08871084e-01 3.51344913e-01 1.32804942e+00
1.21270275e+00 2.41492659e-01 5.60134239e-02 2.14063630e-01
2.26413369e-01 -4.07749057e-01 -2.91180816e-02 -1.41575050e+00
-9.58367512e-02 -7.51139224e-01 -8.28116834e-01 -9.89354476e-02
2.63636768e-01 -9.12915826e-01 -8.68874907e-01 -1.45373416e+00
8.18114519e-01 -2.38176987e-01 -3.63380104e-01 5.23810625e-01
-8.67251083e-02 4.10578698e-01 -2.05909759e-01 2.14013129e-01
-4.31149572e-01 7.44330585e-02 7.86310852e-01 1.52634099e-01
-2.65637897e-02 -1.94285944e-01 -1.35534561e+00 9.41137433e-01
9.84580636e-01 -5.02007365e-01 -4.78716284e-01 -2.56253302e-01
1.89265072e-01 -1.99517071e-01 3.52862120e-01 -5.32129824e-01
-2.13412434e-01 -5.73303819e-01 6.96312070e-01 -3.39540064e-01
6.81946397e-01 -2.14549765e-01 -5.04541658e-02 6.82804227e-01
-1.56017661e-01 3.18618655e-01 6.11632288e-01 -1.06846020e-01
4.10714090e-01 -6.76273644e-01 8.57693672e-01 -2.12040380e-01
-3.03004920e-01 2.35980079e-02 -1.06751776e+00 9.48882848e-02
6.68350995e-01 -5.57027347e-02 -3.35280359e-01 -3.30781549e-01
-9.55340385e-01 -4.75913398e-02 4.87412393e-01 -1.09229155e-01
4.55876529e-01 -8.59281778e-01 -4.73543137e-01 2.21073106e-02
-1.09827749e-01 -1.06334388e+00 4.51678962e-01 1.58618438e+00
-2.69602448e-01 9.86781478e-01 -5.26906431e-01 -5.49298599e-02
-1.08829618e+00 4.94352430e-01 3.35782647e-01 2.98723895e-02
-4.67002243e-01 4.40248966e-01 2.11012706e-01 -2.19048634e-01
2.42777895e-02 1.24696858e-01 -3.09871674e-01 2.41220742e-01
7.60741472e-01 6.33888543e-01 4.35848266e-01 -4.53375965e-01
-8.64889443e-01 2.03973711e-01 -6.10218823e-01 -2.34392881e-01
1.51750171e+00 -6.12000842e-03 -4.66904908e-01 5.26365817e-01
1.12096083e+00 1.15708739e-01 -2.67617583e-01 3.40097189e-01
8.53342488e-02 -6.94697499e-02 1.05914101e-01 -7.00109422e-01
-5.93774974e-01 9.06666577e-01 6.43445551e-01 2.57583112e-01
8.23474288e-01 -3.34432931e-03 2.79814005e-01 2.85676062e-01
2.25311607e-01 -1.11651993e+00 -5.58134019e-01 3.76405805e-01
6.41328096e-01 -7.40500450e-01 3.47817332e-01 -3.11478317e-01
-5.81015587e-01 1.15448928e+00 3.05452198e-01 -1.03846461e-01
4.40430820e-01 3.35247703e-02 -9.34535414e-02 -5.12884855e-01
-8.51466715e-01 1.68153811e-02 -1.34503305e-01 1.11829364e+00
7.44337261e-01 -1.64594408e-02 -1.14538181e+00 9.71503854e-01
-3.05873185e-01 1.44918382e-01 4.83351469e-01 7.68142819e-01
-4.83628660e-01 -1.33708704e+00 -3.70389581e-01 9.16329384e-01
-7.01093793e-01 -2.93978959e-01 -8.90612304e-01 7.71649599e-01
1.99212551e-01 8.29919159e-01 8.14355761e-02 -2.80464798e-01
1.19326204e-01 4.22051579e-01 5.57455778e-01 -7.43602633e-01
-4.96209532e-01 2.34774470e-01 1.42244339e-01 -2.68350303e-01
-2.68287480e-01 -1.13950598e+00 -1.07557476e+00 -4.07270491e-01
1.27976641e-01 1.84432581e-01 4.44894046e-01 1.15928364e+00
5.57394028e-01 3.09196949e-01 3.04058969e-01 -1.93753809e-01
-2.98526496e-01 -9.62738097e-01 -1.00011456e+00 -1.45574138e-01
2.40423068e-01 -5.57473600e-01 -3.89166981e-01 -2.16534346e-01] | [8.118496894836426, 5.647578716278076] |
e5f53afa-f10a-4f28-9b56-2fa63b19e102 | llm-assisted-generation-of-hardware | 2306.14027 | null | https://arxiv.org/abs/2306.14027v1 | https://arxiv.org/pdf/2306.14027v1.pdf | LLM-assisted Generation of Hardware Assertions | The security of computer systems typically relies on a hardware root of trust. As vulnerabilities in hardware can have severe implications on a system, there is a need for techniques to support security verification activities. Assertion-based verification is a popular verification technique that involves capturing design intent in a set of assertions that can be used in formal verification or testing-based checking. However, writing security-centric assertions is a challenging task. In this work, we investigate the use of emerging large language models (LLMs) for code generation in hardware assertion generation for security, where primarily natural language prompts, such as those one would see as code comments in assertion files, are used to produce SystemVerilog assertions. We focus our attention on a popular LLM and characterize its ability to write assertions out of the box, given varying levels of detail in the prompt. We design an evaluation framework that generates a variety of prompts, and we create a benchmark suite comprising real-world hardware designs and corresponding golden reference assertions that we want to generate with the LLM. | ['Jeyavijayan Rajendran', 'Ramesh Karri', 'Shailja Thakur', 'Brendan Dolan-Gavitt', 'Benjamin Tan', 'Hammond Pearce', 'Rahul Kande'] | 2023-06-24 | null | null | null | null | ['code-generation'] | ['computer-code'] | [ 2.13666290e-01 7.73983970e-02 -5.82000971e-01 -4.04302299e-01
-7.17867136e-01 -8.61454308e-01 6.30169690e-01 7.06987321e-01
2.91227311e-01 5.34356654e-01 -8.04700553e-02 -1.55165613e+00
5.69647431e-01 -7.63129711e-01 -8.28619301e-01 5.54271460e-01
-1.17928170e-01 -4.02066857e-01 7.85041928e-01 -5.27551949e-01
3.19479376e-01 2.16844708e-01 -1.42570269e+00 1.98138431e-01
5.75864255e-01 8.41166437e-01 -8.72813523e-01 8.32737684e-01
-1.35787334e-02 1.28975737e+00 -9.45551217e-01 -5.18063605e-01
4.63578030e-02 -2.27024361e-01 -7.03891277e-01 -3.94637138e-01
3.00566792e-01 -3.75042468e-01 3.67319971e-01 1.33516943e+00
2.52611679e-03 -7.79230535e-01 1.18292630e-01 -2.14768267e+00
-1.56015055e-02 1.13953757e+00 -4.09909964e-01 -2.24365249e-01
8.51106346e-01 3.97354603e-01 9.57457364e-01 -7.32880771e-01
3.00648719e-01 1.00277853e+00 7.05075800e-01 6.70904815e-01
-1.48569739e+00 -6.53273523e-01 -5.30332923e-02 -3.10724616e-01
-1.58022285e+00 -6.57978833e-01 8.31764460e-01 -6.56606853e-01
1.36796868e+00 6.17297828e-01 3.32093716e-01 1.18973076e+00
1.01919341e+00 2.31395110e-01 1.26141167e+00 -7.44196475e-01
6.58677757e-01 5.41735590e-01 7.74996698e-01 7.53312647e-01
8.38031709e-01 2.62930244e-01 -3.71918827e-01 -1.15626168e+00
2.27375776e-01 -2.76500821e-01 2.11154148e-02 -1.37313485e-01
-9.93998706e-01 6.46079898e-01 -2.95603454e-01 -7.20545202e-02
3.08808953e-01 7.06173301e-01 8.76948833e-01 5.81775367e-01
-1.69390485e-01 5.49703777e-01 -6.85262859e-01 -2.35378325e-01
-1.08671546e+00 3.76767367e-01 1.22239566e+00 9.47023630e-01
4.72892344e-01 3.65922958e-01 4.66629788e-02 -4.61415589e-01
8.35130632e-01 4.22148615e-01 6.91273343e-03 -3.84626389e-01
2.24712137e-02 8.16981792e-01 1.70211092e-01 -9.27820504e-01
9.06542540e-02 -1.39606133e-01 -2.46342823e-01 9.49482918e-01
-1.69035241e-01 -1.09285727e-01 -5.22124469e-01 1.71958864e+00
2.27232069e-01 2.06401497e-01 3.82508606e-01 4.24945951e-01
3.85236889e-01 3.53189647e-01 -1.03797905e-01 -1.23354696e-01
1.39938498e+00 -4.02243286e-01 -4.34540153e-01 -1.84199050e-01
7.41456509e-01 -5.34284234e-01 9.36903954e-01 5.17595768e-01
-8.90817821e-01 -2.92275161e-01 -1.83781219e+00 5.35496473e-01
-2.88905680e-01 -2.30189502e-01 4.70538944e-01 1.14507723e+00
-1.36810219e+00 -4.00117449e-02 -8.63870740e-01 3.96633804e-01
-1.63700745e-01 3.41002554e-01 -4.37458865e-02 4.28771228e-01
-1.22461081e+00 7.57735848e-01 2.59627461e-01 -2.23679915e-01
-1.45291114e+00 -9.55088735e-01 -1.40431845e+00 -9.96369794e-02
6.06414795e-01 -3.06964397e-01 1.66127968e+00 -7.37893164e-01
-1.25936770e+00 3.24884295e-01 2.80296355e-02 -6.94913328e-01
1.72047928e-01 2.01971814e-01 -9.59473372e-01 -3.97487432e-01
-7.52448943e-03 2.42452115e-01 1.12612557e+00 -1.22609580e+00
-2.95669913e-01 1.35703951e-01 7.33618677e-01 -1.30992699e+00
-9.28708389e-02 8.50168467e-01 5.09872735e-01 -3.12357277e-01
-6.61908329e-01 -8.05760562e-01 -3.26477945e-01 -4.12196107e-02
-7.23712862e-01 2.13847101e-01 8.51803780e-01 -4.18362647e-01
1.50648415e+00 -2.12826443e+00 -5.08472323e-01 6.42716885e-01
3.97450745e-01 8.28329101e-02 2.67331809e-01 6.08919561e-01
-2.44501472e-01 8.04857373e-01 -9.76458639e-02 9.72642973e-02
5.57605088e-01 -1.93589121e-01 -1.06604421e+00 3.04710418e-01
5.86480319e-01 8.24438751e-01 -7.65612304e-01 -5.30018270e-01
-6.86693117e-02 1.47652864e-01 -6.97249413e-01 1.12605706e-01
-8.93717051e-01 -1.61299258e-01 -4.15005833e-01 1.04246032e+00
5.20229757e-01 -2.76897371e-01 3.46776634e-01 5.16002849e-02
-3.91016304e-01 4.72515404e-01 -1.21904552e+00 1.28878248e+00
-8.03642869e-01 5.09853482e-01 -1.80537961e-02 1.14242509e-01
8.49066377e-01 6.78540528e-01 -3.03199798e-01 -2.21525654e-01
1.48334384e-01 2.50577569e-01 1.29141524e-01 9.44091231e-02
6.10573411e-01 -2.90386938e-02 -8.86163056e-01 1.07876396e+00
-4.53934371e-01 -3.37436706e-01 8.69433116e-03 4.64091390e-01
1.56468189e+00 -3.21603823e-03 7.97939360e-01 -3.95521194e-01
8.50741863e-01 2.59881407e-01 3.79240572e-01 7.76868582e-01
-1.81911830e-02 -8.62921923e-02 1.05523014e+00 -5.04910171e-01
-1.12243521e+00 -6.91224754e-01 7.30933324e-02 4.07196671e-01
-2.60876030e-01 -1.49531019e+00 -5.14180064e-01 -1.02421665e+00
-2.89295346e-01 9.36805129e-01 -4.97024924e-01 -7.03548193e-01
-1.69028103e-01 1.70590594e-01 1.37917161e+00 7.31773973e-01
1.08357951e-01 -8.27781916e-01 -1.30528176e+00 4.11760807e-02
4.76209402e-01 -1.04559147e+00 -3.95291418e-01 5.95369712e-02
-2.19172448e-01 -1.07375908e+00 4.38119322e-01 -1.92928836e-01
8.04411232e-01 -2.89280772e-01 1.35216999e+00 8.43178511e-01
-3.53284150e-01 1.79076508e-01 -4.85698044e-01 -6.81924224e-01
-1.33908677e+00 -2.87117898e-01 1.52161047e-01 -3.60151410e-01
-1.32719025e-01 -2.41686836e-01 1.56896517e-01 3.36385012e-01
-1.25679266e+00 1.79644898e-02 9.69468951e-02 7.41389155e-01
-2.12557074e-02 2.68618315e-01 2.10072279e-01 -7.94015586e-01
8.34522307e-01 -1.19191304e-01 -1.53494143e+00 3.94470632e-01
-6.99772835e-01 3.88438612e-01 9.85405028e-01 -5.31425416e-01
-3.88964266e-01 -2.01545134e-01 -2.68162429e-01 -2.64838099e-01
1.84189990e-01 1.01830459e+00 -2.47916073e-01 -4.86602098e-01
8.54942739e-01 -3.33195999e-02 -1.03761598e-01 4.40623254e-01
-1.98947620e-02 3.57136995e-01 2.57760614e-01 -1.31480050e+00
1.16430676e+00 -3.31949666e-02 1.25192642e-01 -1.76727265e-01
-3.46501678e-01 4.20599788e-01 1.25880182e-01 -1.42916545e-01
2.97550354e-02 -4.60736811e-01 -1.12001657e+00 1.04187548e-01
-1.14165926e+00 -5.83432198e-01 -4.11990255e-01 -2.87582397e-01
-1.17834195e-01 2.64697462e-01 -3.60958099e-01 -1.23866284e+00
-5.08461654e-01 -1.74399173e+00 1.20397878e+00 -6.64931759e-02
-9.90816057e-01 -7.09265053e-01 3.75792533e-01 -5.49462855e-01
6.13423347e-01 3.70977730e-01 1.19646823e+00 -7.78797626e-01
-4.11244988e-01 -4.74272698e-01 1.69385731e-01 3.47838283e-01
5.60721532e-02 7.33395696e-01 -8.62915695e-01 -3.18249792e-01
-9.06850323e-02 -4.67094928e-01 -4.66746718e-01 -3.53981048e-01
8.58740389e-01 -6.01259291e-01 -7.18657821e-02 -7.90197179e-02
1.47019613e+00 -8.81028771e-02 6.56056285e-01 7.37304166e-02
-2.32290253e-02 -4.05531526e-02 5.98233521e-01 4.61969703e-01
1.41173750e-01 6.59422100e-01 4.27712917e-01 2.27128819e-01
3.65255177e-01 -5.22590697e-01 7.97151327e-01 5.84198162e-02
6.53460979e-01 2.87478566e-01 -1.32268572e+00 2.57714242e-01
-1.51102400e+00 -9.39299047e-01 -1.84303612e-01 2.25513387e+00
1.26163471e+00 7.64503777e-01 -4.25715521e-02 6.78334534e-01
1.02287233e-01 -9.60785151e-02 -8.73776376e-02 -6.38929188e-01
5.74523032e-01 6.27943873e-01 2.50942349e-01 8.76527011e-01
-5.44969022e-01 7.52499878e-01 6.76938963e+00 5.17970324e-01
-1.51224124e+00 -1.10954426e-01 8.95260125e-02 6.81690097e-01
-8.65622461e-01 8.32841098e-01 -1.06159914e+00 1.76553011e-01
1.73155606e+00 -7.48971343e-01 -1.25304058e-01 1.07075703e+00
1.76408932e-01 -9.39865112e-02 -1.48455811e+00 2.42470071e-01
-4.07066345e-02 -1.45295715e+00 -1.13897167e-01 -8.20725188e-02
4.39584404e-01 -6.52647913e-01 2.31996644e-02 5.21178722e-01
6.82439864e-01 -1.06072879e+00 1.44611037e+00 1.36853710e-01
9.51006472e-01 -7.11712301e-01 8.91445637e-01 2.80197978e-01
-1.11620271e+00 1.33935660e-01 2.82149881e-01 -2.86231279e-01
1.32867366e-01 5.66178143e-01 -1.50961292e+00 4.47251350e-01
2.12869033e-01 1.65390372e-01 -9.24553037e-01 7.67600656e-01
-5.67648768e-01 7.85226762e-01 -2.02258825e-01 -3.33659708e-01
-2.86008298e-01 7.05279052e-01 2.00154647e-01 1.01338100e+00
9.70865637e-02 -2.50492096e-01 4.12262678e-01 1.06020570e+00
1.93861797e-01 -4.95413154e-01 -6.39350891e-01 -3.37955713e-01
4.76883352e-01 1.17515826e+00 -6.63229764e-01 -3.91701311e-01
-5.61824441e-01 2.04731241e-01 -3.02492350e-01 -1.25938088e-01
-1.40121305e+00 -1.90721959e-01 6.81448400e-01 4.18278784e-01
-1.80968985e-01 -6.04647875e-01 -2.22177073e-01 -1.11309063e+00
2.25706607e-01 -1.71019900e+00 -1.42818782e-02 -9.23549891e-01
-6.48139358e-01 8.82434070e-01 4.54000711e-01 -1.14409614e+00
-4.43025529e-01 -4.22104090e-01 -7.96273708e-01 8.02191973e-01
-1.19451749e+00 -1.19364953e+00 -8.65132269e-03 2.27722526e-01
1.61018834e-01 2.08864361e-01 9.99886990e-01 3.10675144e-01
-3.52674365e-01 1.00678766e+00 -1.28044391e+00 9.08183157e-02
5.94261944e-01 -8.21517646e-01 9.35682476e-01 1.43983316e+00
-8.15720111e-02 1.39004230e+00 1.25141907e+00 -1.00445914e+00
-2.03558850e+00 -1.00592637e+00 7.11727381e-01 -5.38458705e-01
1.41930664e+00 -6.47553921e-01 -6.42590880e-01 1.06795144e+00
1.45117134e-01 2.52380371e-01 6.89075828e-01 -1.81299180e-01
-1.02649415e+00 1.30352348e-01 -1.07478654e+00 8.16613734e-01
2.13905230e-01 -1.13193738e+00 -2.31254682e-01 -8.33830684e-02
1.10859120e+00 -4.76897508e-01 -8.85349751e-01 2.60335296e-01
5.52051127e-01 -6.61074817e-01 6.60243928e-01 -7.88303852e-01
5.19976735e-01 -1.18619657e+00 -3.83663744e-01 -7.69443035e-01
4.56588954e-01 -9.05484319e-01 -3.83074552e-01 1.44262838e+00
1.02287030e+00 -8.34129930e-01 3.97445828e-01 9.72742379e-01
-9.44108590e-02 -4.71521139e-01 -5.27799308e-01 -6.28845096e-01
-3.54034215e-01 -9.29566681e-01 1.02361071e+00 9.05752063e-01
7.55118489e-01 2.00722009e-01 -1.62117153e-01 4.97756541e-01
3.97343755e-01 1.01719713e-02 8.99406672e-01 -1.02282786e+00
-6.57735884e-01 -7.58241042e-02 -6.11483455e-01 -1.66686431e-01
5.53875923e-01 -6.30019426e-01 1.56268671e-01 -5.70068300e-01
8.04283917e-02 -4.01217490e-01 2.18197584e-01 1.05935383e+00
3.64878923e-01 -6.41772151e-02 -5.30725121e-02 -3.19757104e-01
-4.94766593e-01 1.14923775e-01 1.40645668e-01 -6.35316074e-01
1.46351323e-01 1.02316506e-01 -6.80590928e-01 4.51963902e-01
5.43307304e-01 -5.23860633e-01 -5.18171430e-01 2.97504097e-01
1.14117861e+00 6.31240845e-01 7.84270644e-01 -1.17644966e+00
3.44911069e-01 -5.08419156e-01 -3.97001833e-01 -1.39556825e-01
7.95294940e-02 -9.55584228e-01 6.61380291e-01 7.34346509e-01
-4.86138046e-01 6.63109839e-01 6.78245604e-01 -3.76981571e-02
-2.97695905e-01 -5.50111532e-01 2.13849440e-01 3.83469433e-01
-7.23712444e-01 1.05148360e-01 -5.96267819e-01 -8.26154724e-02
1.16007173e+00 1.97002769e-01 -6.85318232e-01 -1.66083321e-01
-2.53820062e-01 -2.55640782e-02 1.15904212e+00 3.12659651e-01
9.80666399e-01 -1.11393952e+00 -4.18632448e-01 5.16806602e-01
5.94620109e-01 -6.38195097e-01 -3.94449115e-01 7.19289541e-01
-4.35471237e-01 1.80449560e-01 -1.40352234e-01 -4.36586499e-01
-1.47071266e+00 9.14254665e-01 1.88870415e-01 -2.68593311e-01
-2.67029732e-01 2.79299349e-01 -3.16626340e-01 -2.91147768e-01
-3.77064168e-01 -8.57218206e-01 3.80259812e-01 -8.80495608e-01
1.02021873e+00 -5.46402454e-01 2.72012353e-01 -3.00484244e-02
-9.02036369e-01 -1.93399534e-01 -2.96208989e-02 -6.79322243e-01
7.83780396e-01 7.66173184e-01 -5.19497931e-01 4.47005421e-01
7.85266936e-01 6.75834596e-01 -5.39030313e-01 1.94202214e-01
1.28779590e-01 -3.04569542e-01 -9.54819322e-02 -6.81467295e-01
-3.92179340e-01 5.13635218e-01 2.28505861e-02 5.52227736e-01
7.87091553e-01 -8.86165351e-02 3.93157095e-01 7.43589476e-02
1.26944375e+00 -1.95861757e-01 -3.99387740e-02 4.15897310e-01
5.31090081e-01 -8.58802438e-01 1.12997651e-01 -5.74427485e-01
-1.77768990e-01 9.98867512e-01 7.13503838e-01 -4.27333377e-02
5.85555196e-01 1.52010560e+00 -1.40763700e-01 -5.24440855e-02
-1.10733509e+00 6.63811445e-01 -1.74691841e-01 5.34997880e-01
5.56865394e-01 -1.29747793e-01 -2.33579189e-01 8.41697931e-01
-1.77863613e-01 2.75653452e-01 1.30191815e+00 1.89746034e+00
1.07465051e-02 -1.74488747e+00 -7.17850447e-01 6.81962967e-02
-3.50661874e-01 -3.40899736e-01 -5.72568476e-01 7.29675651e-01
-2.73416936e-01 1.02375376e+00 -6.97718620e-01 -8.18820894e-01
4.62192781e-02 -5.90473749e-02 4.10799861e-01 -8.12391043e-01
-9.34481859e-01 -6.31647527e-01 7.69028783e-01 -5.54130614e-01
1.48136467e-01 -3.98909092e-01 -1.35181904e+00 -5.05907178e-01
-3.51908982e-01 5.52988768e-01 6.68229759e-01 8.47187102e-01
2.30768993e-01 6.69860125e-01 4.02760267e-01 -2.41149619e-01
-7.23448932e-01 -2.36565933e-01 -1.09231748e-01 -3.35037887e-01
6.48472190e-01 -5.16134143e-01 -1.90001249e-01 1.89694241e-01] | [7.786944389343262, 7.624784469604492] |
f5684463-858b-4080-9503-d0742e1de83c | learning-sampling-dictionaries-for-efficient | 2306.00851 | null | https://arxiv.org/abs/2306.00851v1 | https://arxiv.org/pdf/2306.00851v1.pdf | Learning Sampling Dictionaries for Efficient and Generalizable Robot Motion Planning with Transformers | Motion planning is integral to robotics applications such as autonomous driving, surgical robots, and industrial manipulators. Existing planning methods lack scalability to higher-dimensional spaces, while recent learning based planners have shown promise in accelerating sampling-based motion planners (SMP) but lack generalizability to out-of-distribution environments. To address this, we present a novel approach, Vector Quantized-Motion Planning Transformers (VQ-MPT) that overcomes the key generalization and scaling drawbacks of previous learning-based methods. VQ-MPT consists of two stages. Stage 1 is a Vector Quantized-Variational AutoEncoder model that learns to represent the planning space using a finite number of sampling distributions, and stage 2 is an Auto-Regressive model that constructs a sampling region for SMPs by selecting from the learned sampling distribution sets. By splitting large planning spaces into discrete sets and selectively choosing the sampling regions, our planner pairs well with out-of-the-box SMPs, generating near-optimal paths faster than without VQ-MPT's aid. It is generalizable in that it can be applied to systems of varying complexities, from 2D planar to 14D bi-manual robots with diverse environment representations, including costmaps and point clouds. Trained VQ-MPT models generalize to environments unseen during training and achieve higher success rates than previous methods. | ['Michael Yip', 'Ahmed H Qureshi', 'Jacob J Johnson'] | 2023-06-01 | null | null | null | null | ['motion-planning'] | ['robots'] | [ 4.61966880e-02 5.90281069e-01 -4.77732122e-01 -1.78536270e-02
-9.21415150e-01 -2.51040727e-01 4.95219231e-01 -1.21537104e-01
-4.00849432e-01 7.25176752e-01 2.06620514e-01 -4.88838434e-01
-4.83164579e-01 -8.70064795e-01 -8.80519092e-01 -6.85798645e-01
-2.75930434e-01 1.27872348e+00 3.80961806e-01 -4.11431909e-01
3.74705613e-01 7.16684401e-01 -1.07266462e+00 -5.48886396e-02
5.67817032e-01 2.98292577e-01 8.45192432e-01 6.46354914e-01
3.56215946e-02 5.27139187e-01 -3.41006696e-01 5.87077618e-01
4.69597697e-01 2.58110315e-02 -8.91405046e-01 -1.55316472e-01
-1.74521625e-01 -5.54944575e-01 -5.38036048e-01 6.62676573e-01
4.72065270e-01 4.39757615e-01 1.01781881e+00 -1.39746165e+00
-3.28059256e-01 2.87545413e-01 -2.67176867e-01 -3.09273452e-01
2.82245398e-01 4.62156624e-01 7.13294625e-01 -7.19395459e-01
1.07589006e+00 1.63711512e+00 7.58536458e-01 7.25870311e-01
-1.21757913e+00 -2.31690541e-01 -1.54109716e-01 3.95793691e-02
-1.13140035e+00 7.88552221e-03 5.22761285e-01 -5.89795470e-01
1.46960461e+00 -3.93350899e-01 6.35314703e-01 9.83872950e-01
1.13119996e+00 8.25050712e-01 3.27156216e-01 -1.08959995e-01
6.15996182e-01 -3.71090949e-01 -5.05485713e-01 8.10118735e-01
-8.06359760e-03 3.72338682e-01 -2.24743068e-01 -3.92289966e-01
1.46438086e+00 -3.93355116e-02 -7.87057355e-02 -1.06502759e+00
-1.65931058e+00 1.26663101e+00 6.77555561e-01 1.94666386e-02
-6.55694902e-01 5.80512047e-01 2.64267266e-01 3.91795747e-02
-7.35444129e-02 9.50330973e-01 -6.40819907e-01 -2.17267409e-01
-6.85066819e-01 7.60377586e-01 6.65804803e-01 1.46023917e+00
9.31970060e-01 1.14335284e-01 9.45727434e-03 6.23037040e-01
3.29054564e-01 3.47230047e-01 5.28924823e-01 -1.66957545e+00
4.28303599e-01 2.62820482e-01 2.47717083e-01 -8.32056522e-01
-1.00051904e+00 7.67573863e-02 -5.86957932e-01 6.14029408e-01
-5.22397161e-02 -3.12675387e-01 -1.12541234e+00 1.45543325e+00
4.71674055e-01 -2.83411503e-01 2.89144218e-01 8.03512931e-01
3.92360181e-01 8.89203191e-01 -1.18716098e-01 5.87703474e-02
8.93248200e-01 -1.22641003e+00 -4.30527270e-01 -4.04782742e-01
6.39092982e-01 -2.76163995e-01 8.51883650e-01 3.60494375e-01
-9.82795835e-01 -3.89039606e-01 -1.04335380e+00 -3.22185874e-01
-9.72111300e-02 -1.47316292e-01 6.48828030e-01 -5.30263409e-02
-1.40820909e+00 1.03279495e+00 -1.45812070e+00 -4.15213674e-01
6.47713363e-01 6.97919309e-01 -3.83997649e-01 -1.41988367e-01
-6.25294030e-01 1.35307837e+00 6.66268110e-01 -3.79272431e-01
-1.37544405e+00 -5.24812460e-01 -1.26411700e+00 -1.14850827e-01
2.80622661e-01 -1.02378225e+00 1.58504772e+00 -7.49009103e-02
-1.84059405e+00 1.51332458e-02 -1.12179533e-01 -4.29065228e-01
2.33935893e-01 -8.12652335e-03 4.14382905e-01 7.41508380e-02
4.73881096e-01 1.38711119e+00 7.46101081e-01 -1.05540097e+00
-5.20321846e-01 -1.02101110e-01 -4.34077978e-02 5.80009460e-01
2.53357649e-01 -7.31908202e-01 -1.62642747e-01 -5.03529720e-02
4.12243634e-01 -1.24962044e+00 -1.08866394e+00 2.29643628e-01
-3.26792777e-01 -5.27872503e-01 6.51467085e-01 -2.07092017e-01
3.59444946e-01 -1.63577676e+00 5.99407017e-01 -3.82613279e-02
4.34583649e-02 -2.04190359e-01 -3.68567139e-01 6.83860898e-01
4.57557499e-01 -3.03829134e-01 -3.28025997e-01 -2.14595079e-01
6.17109872e-02 8.21655929e-01 -1.04852192e-01 5.07907450e-01
5.31995416e-01 9.29453671e-01 -1.46096218e+00 -6.05420172e-01
4.66682583e-01 3.40131581e-01 -8.22380543e-01 -3.99609329e-03
-5.95051765e-01 6.65910244e-01 -7.98525035e-01 5.34671485e-01
2.48056933e-01 -1.65403381e-01 3.44332606e-02 2.72827685e-01
-1.69683829e-01 1.95368290e-01 -8.61977816e-01 2.25229049e+00
-3.68167281e-01 7.13470221e-01 1.60988271e-02 -9.01198208e-01
8.77668023e-01 1.16883293e-01 9.97077405e-01 -3.36590409e-01
1.22247592e-01 2.24064291e-01 -2.15445295e-01 -5.35779059e-01
8.35149705e-01 -2.89069384e-01 -2.23422438e-01 1.46483481e-01
3.14964026e-01 -1.00014794e+00 -1.15371443e-01 -1.34640276e-01
1.27196538e+00 5.61550438e-01 5.11576653e-01 -2.31386825e-01
-1.55267771e-02 7.75618672e-01 5.81707358e-01 6.26695395e-01
-4.10540938e-01 5.07067740e-01 3.62108856e-01 -5.45578837e-01
-1.37253654e+00 -1.21761882e+00 -9.68594775e-02 7.58227110e-01
4.62033451e-01 -8.07521269e-02 -4.80589837e-01 -3.55498761e-01
2.33737141e-01 8.67636740e-01 -4.00106460e-01 5.07213846e-02
-8.66807938e-01 -1.78489789e-01 3.07051182e-01 7.62775958e-01
-1.10683069e-01 -1.31724167e+00 -1.20949340e+00 6.34605706e-01
-1.68912893e-03 -9.25692022e-01 -3.27725500e-01 3.97670686e-01
-1.08462358e+00 -1.12709773e+00 -7.10616767e-01 -1.04495275e+00
5.66486239e-01 2.26548433e-01 8.34532320e-01 -5.62667251e-01
-2.97270000e-01 4.27957267e-01 -8.19742382e-02 -7.39707172e-01
-5.94763935e-01 2.05083370e-01 3.36333692e-01 -9.31856871e-01
-9.73310098e-02 -5.43236673e-01 -4.73434955e-01 2.65595347e-01
-4.42321271e-01 1.32436439e-01 8.68675530e-01 1.10006785e+00
9.33072209e-01 4.54868451e-02 3.80966961e-01 -3.03553194e-01
5.96557617e-01 -7.22255349e-01 -4.94174808e-01 -4.38831389e-01
-4.41538423e-01 4.09547657e-01 5.28146207e-01 -5.63377976e-01
-5.68044484e-01 4.30000454e-01 -1.87650293e-01 -8.18788707e-01
-5.77198081e-02 3.15233737e-01 2.80938804e-01 -8.46812502e-02
9.26494181e-01 1.16783909e-01 5.09026110e-01 3.41759361e-02
6.19182706e-01 2.53392547e-01 6.14290535e-01 -5.28262436e-01
7.18831480e-01 4.44222182e-01 2.99980730e-01 -9.68203425e-01
1.06459945e-01 -4.31877971e-01 -7.71602452e-01 -4.80164075e-03
1.02955234e+00 -6.46244168e-01 -6.51946306e-01 -3.29321586e-02
-1.32659566e+00 -8.74819756e-01 -5.82827210e-01 6.96718633e-01
-1.47451794e+00 -1.10487781e-04 -5.74325621e-01 -6.57166600e-01
-1.76120281e-01 -1.71723258e+00 1.46918488e+00 8.25299695e-02
-3.43711048e-01 -8.38400364e-01 3.37918252e-01 -2.46691912e-01
2.84506351e-01 5.73697984e-01 9.16673362e-01 -3.44949484e-01
-7.59411573e-01 -5.13312183e-02 3.29800516e-01 -2.11065859e-01
-3.54579389e-02 -4.15699571e-01 -4.01824325e-01 -4.03457046e-01
-2.01203629e-01 -3.68413299e-01 4.78641361e-01 9.81604457e-01
8.85817289e-01 -3.35414410e-01 -9.95618403e-01 5.29321373e-01
1.24396563e+00 4.50688392e-01 4.92533654e-01 6.15226805e-01
4.78638142e-01 4.75413114e-01 8.15240681e-01 2.59569168e-01
5.18802583e-01 6.13306284e-01 8.67092431e-01 2.54076898e-01
2.38905191e-01 -4.36044425e-01 2.82757759e-01 4.72110987e-01
1.60297602e-01 6.92537650e-02 -1.04744995e+00 7.62603641e-01
-2.03010869e+00 -7.01462984e-01 2.59918660e-01 1.76448739e+00
5.20986378e-01 -9.17458069e-03 7.73596168e-02 -2.29792282e-01
3.29347849e-01 9.68865305e-02 -9.48306084e-01 -4.92780298e-01
5.93425035e-01 2.47854255e-02 6.86744213e-01 6.89180315e-01
-1.05359530e+00 1.02569175e+00 6.60282516e+00 4.98284012e-01
-9.15211976e-01 4.84616682e-02 1.02895936e-02 -5.16382307e-02
-3.07115465e-01 -7.21465722e-02 -5.63451052e-01 1.16843507e-02
7.90213227e-01 -5.73051609e-02 6.08608425e-01 1.44855595e+00
1.21551514e-01 1.11910962e-01 -1.34961700e+00 7.95585215e-01
-3.05116713e-01 -1.57793856e+00 -2.01178715e-01 2.54358500e-01
6.57222271e-01 6.18949711e-01 -3.53942327e-02 6.62003398e-01
7.12518752e-01 -1.25774944e+00 7.55790532e-01 2.16274083e-01
6.52375460e-01 -5.89761019e-01 6.29949868e-01 8.99382412e-01
-9.88817215e-01 -4.09171313e-01 -8.28476131e-01 -5.14600873e-02
5.68780601e-01 5.24115078e-02 -1.68833244e+00 4.42739725e-01
3.72743696e-01 5.87930143e-01 3.72459084e-01 8.17117095e-01
-1.47541864e-02 -1.79484680e-01 -4.94667530e-01 -3.73933703e-01
6.87066078e-01 2.15921272e-02 8.15197408e-01 7.17058539e-01
4.48684961e-01 2.26124115e-02 4.88209724e-01 1.04764879e+00
4.21949863e-01 -2.98649162e-01 -1.02167356e+00 1.01398975e-01
6.72984779e-01 7.86523402e-01 -6.14436626e-01 -1.53415347e-03
-5.55806123e-02 7.53256619e-01 4.25400257e-01 2.84032851e-01
-6.29311740e-01 -3.19910467e-01 7.67875195e-01 -6.28974289e-02
4.48151290e-01 -7.86137104e-01 -2.79874742e-01 -6.43466115e-01
-3.05387318e-01 -5.26879430e-01 -9.86380354e-02 -7.50730574e-01
-8.36275578e-01 5.27969420e-01 2.97515869e-01 -1.43493640e+00
-7.64979541e-01 -8.49238157e-01 -3.30939502e-01 6.56392515e-01
-1.36350274e+00 -1.00700486e+00 -1.53742507e-01 3.86061072e-01
1.10940492e+00 -2.55780309e-01 1.05272710e+00 -4.30948824e-01
6.72015920e-02 7.57999495e-02 1.67847008e-01 -1.29152849e-01
3.16917241e-01 -1.24430585e+00 6.36334658e-01 2.49786347e-01
-2.13473707e-01 4.25725669e-01 7.18370855e-01 -6.96051836e-01
-1.66656554e+00 -1.27081811e+00 2.48005852e-01 -5.99657476e-01
3.41269642e-01 7.42385313e-02 -5.22742391e-01 9.09040093e-01
-5.84394857e-02 -1.38269678e-01 1.48943113e-02 -1.17187120e-01
3.07794333e-01 4.83188957e-01 -1.31422710e+00 7.65567899e-01
9.65286553e-01 -8.94429907e-02 -7.28577673e-01 6.65365338e-01
1.05297041e+00 -1.10860169e+00 -9.52655315e-01 4.36747968e-01
5.73107660e-01 -4.31959033e-01 1.09430587e+00 -5.34979105e-01
6.20159149e-01 -1.72650337e-01 -2.16123506e-01 -1.67086172e+00
-6.25654757e-01 -6.93365693e-01 -6.54480234e-02 4.96725217e-02
3.93975109e-01 -5.87953329e-01 1.01007831e+00 2.80736566e-01
-8.29116762e-01 -1.23020828e+00 -1.18182456e+00 -7.34033644e-01
6.03966594e-01 -4.04677242e-01 6.52933776e-01 5.18658698e-01
2.48886853e-01 2.23475963e-01 -1.53768053e-02 3.66125762e-01
5.68106472e-01 -2.26601526e-01 1.00864375e+00 -9.54944015e-01
-2.90626615e-01 -4.91032094e-01 -5.45572579e-01 -1.42467058e+00
2.13708699e-01 -8.14214766e-01 8.29891384e-01 -2.17458081e+00
-3.59491765e-01 -6.27969086e-01 2.61129767e-01 6.88914537e-01
2.68002182e-01 -6.27311945e-01 1.98957905e-01 4.14683312e-01
-1.95425823e-01 9.39994454e-01 1.68326437e+00 -3.73621911e-01
-7.14200914e-01 7.44615402e-03 -2.57620931e-01 7.09487796e-01
6.29203618e-01 -4.03869599e-01 -8.46635699e-01 -6.56027138e-01
-3.33525658e-01 7.25806475e-01 1.36782229e-01 -1.17711496e+00
5.14411867e-01 -7.18357563e-01 1.85562894e-01 -7.71519959e-01
5.45405865e-01 -6.01903439e-01 5.26551679e-02 7.75898278e-01
-2.37867847e-01 1.45237967e-01 2.47490719e-01 8.59512806e-01
2.97917217e-01 -1.88271224e-01 6.41855955e-01 -4.19510096e-01
-8.87867570e-01 5.82847953e-01 -7.36408532e-01 -3.05328697e-01
1.10838449e+00 -3.39496911e-01 -9.10034105e-02 -2.71275133e-01
-5.71446300e-01 6.60014391e-01 5.40397346e-01 4.68269169e-01
9.50659335e-01 -1.33293414e+00 -6.13351583e-01 2.39341468e-01
-3.28783356e-02 1.07298577e+00 8.49081874e-02 5.64484656e-01
-1.01091015e+00 6.81562722e-01 -3.89243007e-01 -1.16200411e+00
-4.97677773e-01 6.95024908e-01 3.01877797e-01 -2.03429386e-01
-1.17381358e+00 7.44558871e-01 9.97803509e-02 -1.07484293e+00
8.59938562e-02 -6.80572748e-01 -4.52626795e-02 -6.70997024e-01
1.81372702e-01 4.31492448e-01 -3.27327102e-01 -4.45641696e-01
-3.36780131e-01 5.02628803e-01 3.30587506e-01 -3.62429202e-01
1.47354686e+00 2.01668113e-01 3.56738120e-01 3.22891116e-01
1.09406424e+00 -7.68013358e-01 -1.79001975e+00 1.15894288e-01
-7.48619959e-02 -9.80696082e-02 -1.23196483e-01 -1.50294751e-01
-5.33998609e-01 7.84139872e-01 2.12941468e-01 -2.76773691e-01
5.26924253e-01 4.72102799e-02 9.88859773e-01 7.30002463e-01
1.12096930e+00 -9.73525703e-01 3.25368583e-01 8.39451671e-01
1.05558205e+00 -1.10804152e+00 1.25827324e-02 -1.17654108e-01
-8.51097286e-01 1.19032419e+00 4.70649093e-01 -4.31605637e-01
5.09258509e-01 2.32394770e-01 -2.02318691e-02 -2.27787822e-01
-7.05015779e-01 1.80168390e-01 -5.05868047e-02 1.15977669e+00
-2.16455996e-01 2.49177769e-01 8.13152343e-02 2.61604995e-01
-4.18638587e-01 3.54426950e-02 6.30709410e-01 1.25333321e+00
-7.56263077e-01 -8.14626694e-01 -2.87821054e-01 3.88768911e-01
4.56958562e-01 4.07833487e-01 2.54253447e-01 1.11907125e+00
-3.77407414e-04 7.08926320e-01 2.39892289e-01 -5.40567815e-01
2.91138917e-01 -1.08147904e-01 5.73283017e-01 -7.92970240e-01
6.43286780e-02 -3.24098617e-02 -5.09378500e-02 -1.18302059e+00
-6.78600445e-02 -7.73131430e-01 -1.80038869e+00 6.85569346e-02
-1.62787154e-01 -4.57906723e-02 9.34979081e-01 9.25516903e-01
4.68421966e-01 6.66787326e-01 3.64437014e-01 -1.90424681e+00
-1.08086252e+00 -8.49375427e-01 -2.66165674e-01 -1.37219384e-01
6.73926055e-01 -1.05917978e+00 8.34633559e-02 -2.50051737e-01] | [4.67059850692749, 1.0042400360107422] |
04506bd7-96b2-4e9b-8fad-5708c5dbd69e | improving-analytical-tomographic | 1609.06604 | null | http://arxiv.org/abs/1609.06604v1 | http://arxiv.org/pdf/1609.06604v1.pdf | Improving analytical tomographic reconstructions through consistency conditions | This work introduces and characterizes a fast parameterless filter based on
the Helgason-Ludwig consistency conditions, used to improve the accuracy of
analytical reconstructions of tomographic undersampled datasets. The filter,
acting in the Radon domain, extrapolates intermediate projections between those
existing. The resulting sinogram, doubled in views, is then reconstructed by a
standard analytical method. Experiments with simulated data prove that the
peak-signal-to-noise ratio of the results computed by filtered backprojection
is improved up to 5-6 dB, if the filter is used prior to reconstruction. | ['Marco Stampanoni', 'Filippo Arcadu', 'Jakob Vogel', 'Federica Marone'] | 2016-09-21 | null | null | null | null | ['tomographic-reconstructions'] | ['medical'] | [ 1.23778023e-01 1.75927907e-01 4.17219013e-01 -1.98433772e-01
-3.56805384e-01 6.49601594e-02 6.05707943e-01 -3.22219878e-01
-6.71358824e-01 1.06968939e+00 3.82091939e-01 -2.54234612e-01
-1.21310458e-01 -9.77826893e-01 -5.00396490e-01 -6.12992108e-01
-8.59404448e-03 7.11254776e-01 6.12507880e-01 5.38630784e-02
-5.55284545e-02 6.45612717e-01 -1.09065688e+00 6.10153489e-02
4.13949817e-01 8.18201542e-01 1.15876652e-01 5.16686141e-01
-1.74800865e-03 5.77262402e-01 -1.54928520e-01 -8.62188861e-02
4.11344677e-01 -6.47350550e-01 -6.17088914e-01 8.10748190e-02
-1.45601645e-01 -8.27503145e-01 -4.83498394e-01 1.05312216e+00
1.66941941e-01 -1.94377601e-01 6.75542235e-01 -2.36054540e-01
1.89359978e-01 3.00966769e-01 -1.32149100e-01 1.98412910e-01
4.90148097e-01 -1.24809362e-01 1.34978756e-01 -1.19096541e+00
8.68438482e-01 8.55747342e-01 1.01314270e+00 1.51649848e-01
-1.36112106e+00 -8.06507245e-02 -9.97625649e-01 -2.24499285e-01
-1.00672126e+00 -1.44873604e-01 6.09655619e-01 -3.65857840e-01
7.22680926e-01 5.76090097e-01 9.97043192e-01 3.37138027e-01
4.37745959e-01 2.93734670e-02 1.63488269e+00 -6.75511420e-01
2.53981411e-01 3.74485493e-01 8.84099379e-02 6.50058687e-01
3.84067833e-01 4.78155226e-01 -2.53078371e-01 -3.55854690e-01
1.06540942e+00 -3.70593779e-02 -7.80260026e-01 -2.71598339e-01
-9.98864532e-01 3.84918988e-01 3.29323798e-01 5.48709989e-01
-5.88557243e-01 -1.16431274e-01 2.29332820e-01 1.48627907e-01
7.66802013e-01 2.28976265e-01 1.38226107e-01 1.41413659e-01
-1.21158504e+00 2.62220800e-01 7.84315169e-01 5.26097834e-01
5.69413126e-01 1.08327948e-01 2.08298445e-01 2.77877867e-01
3.80556822e-01 9.67972517e-01 2.13323936e-01 -8.04143012e-01
-5.05001110e-04 5.37229367e-02 2.30553165e-01 -4.85000193e-01
-3.08170974e-01 -3.40546280e-01 -4.94784653e-01 3.86962831e-01
3.48603368e-01 5.79901189e-02 -7.92775095e-01 9.08212245e-01
7.18553543e-01 7.99744576e-03 2.04543218e-01 9.18742180e-01
6.26563311e-01 7.61126578e-01 -4.85893279e-01 -9.37775075e-01
8.60018790e-01 -2.44300216e-01 -1.01807654e+00 7.85662904e-02
-3.88750359e-02 -1.01798034e+00 5.86945772e-01 4.81539547e-01
-1.45944715e+00 -3.42851758e-01 -1.11476231e+00 3.50941181e-01
2.60062307e-01 -9.88684446e-02 -5.68765178e-02 4.29610789e-01
-8.46787691e-01 1.12988758e+00 -9.01247144e-01 -1.43023342e-01
6.83891624e-02 -2.00895876e-01 -3.85841638e-01 1.08332023e-01
-8.31882358e-01 1.10840309e+00 1.94667757e-01 7.75867477e-02
-4.08883959e-01 -7.61268675e-01 -2.38952577e-01 -1.77918017e-01
-2.53517717e-01 -6.34827971e-01 1.14054418e+00 -3.91497761e-01
-1.81252551e+00 6.88010216e-01 7.60508627e-02 -5.64298749e-01
1.03099918e+00 -2.23069474e-01 -3.80465418e-01 7.38913774e-01
4.58160639e-02 -4.55003083e-01 6.79442048e-01 -1.37808788e+00
-4.00954857e-02 -4.27938849e-01 -6.92785144e-01 1.16566449e-01
3.52015823e-01 2.56178109e-03 -4.34751362e-02 -1.65604696e-01
9.65406358e-01 -5.60681462e-01 -4.06587154e-01 1.15075834e-01
-2.71860093e-01 7.18810797e-01 7.96220124e-01 -1.11513495e+00
5.93247116e-01 -2.09264135e+00 -2.52926946e-01 5.08332729e-01
-1.47275440e-03 1.22916758e-01 7.26784289e-01 5.92940569e-01
-2.84950435e-02 -4.41332966e-01 -6.85129404e-01 -6.96341991e-02
-7.01072931e-01 9.63455532e-03 -4.90246922e-01 1.22442722e+00
-6.95711851e-01 2.25805685e-01 -6.87123358e-01 -2.20931903e-01
6.16611183e-01 6.54430687e-01 -1.43508956e-01 2.53964275e-01
2.59342760e-01 5.78169346e-01 -5.05252719e-01 -1.44640077e-02
1.30548310e+00 4.24995944e-02 1.23254545e-01 -2.60932386e-01
-5.43336511e-01 3.73307943e-01 -9.74905074e-01 1.11515009e+00
-4.02379096e-01 5.55785716e-01 4.24037129e-01 -7.57547379e-01
1.05117071e+00 5.82356393e-01 6.53143048e-01 -8.07478786e-01
1.30627573e-01 6.62515640e-01 -4.45112497e-01 -4.70903903e-01
4.13408667e-01 -1.00297034e+00 4.26466018e-01 2.73351192e-01
-1.53112978e-01 -7.31347501e-01 -2.03857794e-01 1.26694888e-01
9.68078613e-01 2.83842385e-01 6.19817495e-01 -8.71915877e-01
5.05927086e-01 2.29607150e-01 1.52124196e-01 2.09089622e-01
2.64502376e-01 8.51370871e-01 5.05976863e-02 -7.36742139e-01
-1.58910382e+00 -1.36230981e+00 -1.04968107e+00 -2.61788398e-01
2.09705070e-01 -1.65215611e-01 -8.64172041e-01 -2.89314479e-01
-9.64460745e-02 9.66976464e-01 -2.85327584e-01 2.32233807e-01
-6.68432355e-01 -8.48066449e-01 2.05668762e-01 2.80503362e-01
8.05227160e-01 -7.60223627e-01 -1.08945620e+00 4.10431892e-01
-6.66986853e-02 -1.02614832e+00 5.01698494e-01 1.36774704e-01
-1.30520511e+00 -1.15137279e+00 -7.74193943e-01 -6.60894066e-03
7.14530885e-01 6.78961072e-03 1.01184440e+00 -4.63638045e-02
-1.09317720e-01 4.51511025e-01 -1.83336467e-01 1.65304989e-01
-8.70041966e-01 -1.08664989e+00 -1.54260462e-02 -3.62294674e-01
-1.06124036e-01 -8.24423075e-01 -3.87193859e-01 3.30206126e-01
-7.03457534e-01 4.86428365e-02 7.72716179e-02 9.62160528e-01
6.81139827e-01 -1.29321804e-02 7.43743479e-02 -9.23701882e-01
2.05958992e-01 5.22894673e-02 -9.87141550e-01 -1.74825519e-01
-6.72448635e-01 1.95278063e-01 8.68839562e-01 8.46789703e-02
-1.38708997e+00 -7.84673691e-02 -4.13636506e-01 -1.13435112e-01
6.16184762e-03 9.97100100e-02 2.54510254e-01 -4.00995493e-01
9.53768671e-01 3.99722755e-01 1.02678053e-01 -5.31691432e-01
-2.47837640e-02 4.46426839e-01 7.84190655e-01 -2.33432114e-01
7.31504679e-01 1.17293990e+00 3.76400918e-01 -1.44399941e+00
-3.55151057e-01 -6.12110615e-01 -3.61412436e-01 -4.61764902e-01
7.50046194e-01 -6.28476202e-01 -1.96727529e-01 2.69293219e-01
-1.16071939e+00 -1.47978410e-01 -8.77952695e-01 1.19285285e+00
-7.72937000e-01 5.60865641e-01 -6.01232171e-01 -9.27122831e-01
-4.04315412e-01 -8.35288525e-01 6.06491029e-01 -7.31647462e-02
1.29228890e-01 -9.27392185e-01 3.84848088e-01 4.04434800e-02
4.62887615e-01 1.80936411e-01 3.74222726e-01 -4.36699949e-02
-4.56753075e-01 -1.78060934e-01 2.26383377e-02 4.45958614e-01
-3.79524976e-01 -2.60362715e-01 -1.09174001e+00 -1.36474088e-01
1.01717627e+00 -1.01017766e-01 8.93696189e-01 6.14965439e-01
5.42741895e-01 -1.14274606e-01 -4.79308397e-01 8.08688939e-01
1.90466988e+00 -1.22471102e-01 1.01422322e+00 1.20673157e-01
-8.82769004e-02 2.71073669e-01 5.07926941e-01 3.77584487e-01
-4.18611109e-01 4.23179239e-01 1.87426463e-01 3.02958731e-02
-3.49710047e-01 -1.81058288e-01 -1.76163726e-02 8.16084921e-01
-4.29754585e-01 1.12646863e-01 -6.20316863e-01 4.12739605e-01
-1.33186817e+00 -1.03024757e+00 -7.10124493e-01 2.55669141e+00
5.21001518e-01 1.92124262e-01 -2.56072074e-01 4.05333340e-01
4.28752154e-01 9.24082398e-02 1.37608260e-01 -2.75039107e-01
-4.16234285e-02 4.38258976e-01 1.11269093e+00 9.39842165e-01
-5.64610660e-01 2.29447514e-01 7.66659164e+00 4.13211375e-01
-1.12121987e+00 3.23567063e-01 6.44242838e-02 3.58190596e-01
-6.08408868e-01 3.47318739e-01 -3.78605425e-01 5.04945040e-01
1.02780366e+00 -2.04886138e-01 3.10285032e-01 6.24648273e-01
2.33021080e-01 -8.62277687e-01 -4.74897414e-01 5.68557322e-01
-1.48821816e-01 -1.31316769e+00 -4.27578598e-01 -2.13916171e-02
6.05465412e-01 -2.11092341e-03 -3.77627641e-01 -3.13087493e-01
-7.57453963e-02 -6.11179590e-01 6.99627221e-01 8.44510615e-01
8.28368425e-01 -5.60466647e-01 7.47281849e-01 5.99526584e-01
-1.01414287e+00 5.77756166e-01 -5.85995138e-01 3.40368412e-02
8.42898071e-01 1.18481123e+00 -1.22097278e+00 8.75572920e-01
6.16248190e-01 1.56623423e-01 1.35228008e-01 1.08278286e+00
-2.11794659e-01 7.17556655e-01 -9.11016881e-01 1.80534065e-01
1.62796006e-02 -6.79878294e-01 8.90887380e-01 9.92421329e-01
6.40001655e-01 4.81788039e-01 -9.09196436e-02 8.85998726e-01
4.91462767e-01 4.32841033e-02 -8.33759785e-01 5.72537959e-01
1.53562620e-01 7.42141008e-01 -9.13256586e-01 -4.94128883e-01
-1.15875520e-01 7.62303948e-01 -6.09427154e-01 3.28818291e-01
-5.93277752e-01 -7.02061132e-02 -3.74466240e-01 8.75793993e-01
1.51665255e-01 -2.80801840e-02 -3.80876809e-01 -7.56596625e-01
2.06896275e-01 -1.51553810e-01 2.45367736e-02 -8.49861801e-01
-8.65442038e-01 6.33117795e-01 4.50719178e-01 -1.27767575e+00
-3.48884255e-01 -5.41731119e-01 -6.67167187e-01 1.14124143e+00
-8.79419804e-01 -8.16127837e-01 -1.98438853e-01 3.72835279e-01
1.64944101e-02 2.19593406e-01 7.40926087e-01 1.60200447e-01
5.02994299e-01 -3.30910057e-01 5.97342968e-01 -2.26188362e-01
1.23547897e-01 -8.37725759e-01 8.56937543e-02 9.88175094e-01
-2.37071455e-01 2.05218852e-01 1.32308269e+00 -9.87821162e-01
-1.11857986e+00 -5.28760672e-01 6.79102898e-01 2.54600376e-01
2.60941923e-01 1.44817680e-01 -1.04449582e+00 6.52406991e-01
2.93167144e-01 5.18851519e-01 -8.29703212e-02 -5.03033161e-01
1.44259542e-01 1.80870742e-02 -1.73862970e+00 -1.78103343e-01
3.66238445e-01 -5.08388817e-01 -9.20473993e-01 3.22710305e-01
1.45106003e-01 -6.13028228e-01 -8.32971692e-01 6.05470717e-01
4.92671609e-01 -1.60269284e+00 1.03727400e+00 7.04302713e-02
3.34182709e-01 -2.35214368e-01 -2.00040802e-01 -1.33495414e+00
-4.28690985e-02 -4.31567490e-01 4.66205835e-01 4.09359008e-01
8.02442357e-02 -8.58644605e-01 7.04358876e-01 -2.87438538e-02
-2.52718806e-01 -4.97942030e-01 -1.35118198e+00 -7.23403394e-01
-1.15485720e-01 -1.69625551e-01 2.77118180e-02 6.51686430e-01
1.64062038e-01 4.90926690e-02 -2.82202512e-01 1.61095455e-01
1.00966740e+00 1.16908085e-02 5.20415664e-01 -1.12110388e+00
-5.45356750e-01 3.26394141e-01 -3.45673025e-01 -9.96582925e-01
-4.31965679e-01 -6.34576142e-01 2.00746015e-01 -1.65197718e+00
-1.23842873e-01 -2.28285238e-01 4.85949725e-01 -4.40216988e-01
3.73672724e-01 1.88027829e-01 -2.08568163e-02 3.35857689e-01
3.94212067e-01 5.82428277e-01 1.31672132e+00 4.51161355e-01
4.53588367e-02 2.60451972e-01 5.58288574e-01 1.30669773e+00
3.77980471e-01 -5.71267784e-01 -7.68326893e-02 7.83556849e-02
5.74547313e-02 6.56369984e-01 4.85279679e-01 -1.49934864e+00
6.23150170e-02 1.20567173e-01 5.58379412e-01 -1.15470433e+00
6.22579694e-01 -1.09416652e+00 9.43343222e-01 8.51749241e-01
1.16964474e-01 -1.60296023e-01 -7.58346766e-02 4.06917214e-01
-2.67090827e-01 -7.79558122e-01 1.28514576e+00 -5.15729725e-01
-1.25965685e-01 -2.60826796e-01 -2.50995785e-01 -1.05072558e-01
7.84795642e-01 -8.38261470e-02 -9.19963643e-02 -2.39799604e-01
-6.93284392e-01 -7.28359699e-01 7.81086564e-01 -1.07115483e+00
7.88802207e-01 -1.19601440e+00 -7.20655024e-01 5.53720236e-01
-5.23757041e-01 -2.90233456e-02 3.14283401e-01 1.16972470e+00
-1.39931810e+00 2.27250874e-01 -2.52574891e-01 -5.95236540e-01
-9.38101530e-01 3.71074140e-01 6.57602787e-01 -3.77592415e-01
-1.44446611e+00 4.38798785e-01 -3.66274178e-01 -3.25357229e-01
-4.40374583e-01 -2.86213011e-01 1.77544117e-01 -4.62456524e-01
5.92016160e-01 6.82722688e-01 3.00704956e-01 -4.36931700e-01
-1.36896759e-01 5.93940556e-01 5.98629773e-01 -6.43372655e-01
1.38573134e+00 8.11289921e-02 -1.58639565e-01 5.16431630e-01
9.96121049e-01 6.10006213e-01 -1.06862032e+00 -1.26362309e-01
-4.31348890e-01 -7.24989414e-01 3.35044712e-01 -5.44943929e-01
-5.14095545e-01 7.06725836e-01 4.49353039e-01 4.71595883e-01
1.00073862e+00 -2.02526152e-01 6.05311334e-01 8.71361047e-02
5.09192884e-01 -9.87152755e-01 -5.00467479e-01 2.12371200e-01
8.88660252e-01 -5.48217058e-01 9.30649281e-01 -5.31612754e-01
-9.41770151e-02 1.52697241e+00 -2.44972646e-01 -6.32094145e-01
9.57099974e-01 7.37691104e-01 -2.36342922e-01 -1.38783813e-01
-9.76291522e-02 2.71244645e-01 -8.39748532e-02 3.48037332e-01
2.68505156e-01 6.13981225e-02 -9.61396694e-01 -8.25791880e-02
-1.56009510e-01 2.93273002e-01 7.50869989e-01 8.56212497e-01
-8.53735089e-01 -6.91906810e-01 -1.07490516e+00 2.51743227e-01
-2.17771530e-01 1.20857611e-01 1.92781925e-01 1.10832310e+00
-2.79431075e-01 2.77522475e-01 2.51657099e-01 2.29277745e-01
2.01642081e-01 -4.82112430e-02 8.11346710e-01 -2.38987122e-04
-3.03777069e-01 5.66187561e-01 2.66729802e-01 -4.74485785e-01
-6.24638259e-01 -6.63580477e-01 -1.56226861e+00 -1.84264496e-01
-1.24669686e-01 4.98501956e-01 7.47879922e-01 8.43847692e-01
-5.37389696e-01 3.50091785e-01 7.07233489e-01 -6.63121521e-01
-9.89018261e-01 -1.04149377e+00 -8.39978337e-01 1.32909402e-01
2.71611869e-01 -4.49143887e-01 -7.28358150e-01 -8.64006132e-02] | [12.810315132141113, -2.756901979446411] |
af0a4b4a-dab9-40ba-bb49-0e02ba28fdf3 | neural-interpretation-of-generic-source-code | 2304.00989 | null | https://arxiv.org/abs/2304.00989v1 | https://arxiv.org/pdf/2304.00989v1.pdf | Neural Interpretation of Generic Source Code | Can a generic (Python) program be executed statement-by-statement by neural networks composed according to the source code? We formulate the Abstract Neural Execution Problem and introduce Neural Interpretation, the first neural model that abstractly executes generic source code, where every variable has a vector encoding, and every function executes a neural network. Neural Interpretation is a model of computers with a compiler architecture, which can assemble neural layers ''programmed'' by partial source code. Neural Interpretation can be trained with flexible learning objectives. We demonstrate white-box execution without concrete inputs for variable misuse localization and repair. | ['Jin Tian', 'Yaojie Hu'] | 2023-03-23 | null | null | null | null | ['variable-misuse'] | ['computer-code'] | [ 2.94099122e-01 5.33806026e-01 -5.89855075e-01 -6.51962340e-01
-1.69444352e-01 -4.58741158e-01 2.14321077e-01 -8.88331160e-02
-7.02837482e-02 6.63602293e-01 1.93364948e-01 -1.12923610e+00
3.79595131e-01 -1.19472599e+00 -1.37112927e+00 -2.96741396e-01
-3.55911762e-01 2.18194544e-01 -3.32436442e-01 -3.28178018e-01
7.27627501e-02 1.33679301e-01 -1.57212293e+00 8.49160314e-01
5.04575372e-01 7.47323155e-01 3.42425257e-01 1.36964893e+00
-8.09159040e-01 1.39089680e+00 -1.04423308e+00 -3.40957761e-01
5.22410609e-02 -8.82269740e-02 -9.29422379e-01 -4.32969719e-01
-1.51843242e-02 -2.95678288e-01 -2.68665887e-02 1.19273341e+00
-8.27416331e-02 -2.16938734e-01 1.76258639e-01 -1.28432691e+00
-1.07576716e+00 1.15883148e+00 1.90537110e-01 -6.20325394e-02
5.03791094e-01 7.36666501e-01 1.10935473e+00 -5.18233597e-01
4.95527476e-01 1.09469080e+00 9.18335021e-01 9.34865475e-01
-1.39123607e+00 -1.61606029e-01 1.46362800e-02 -1.19399145e-01
-8.56746376e-01 1.63511291e-01 4.47372437e-01 -7.35678136e-01
1.96756661e+00 5.40323436e-01 7.21595943e-01 1.41129410e+00
4.69063371e-01 9.01993155e-01 1.51037976e-01 -1.49667189e-01
4.14148360e-01 -2.82340854e-01 7.75680423e-01 1.17153263e+00
-4.54370007e-02 4.78609830e-01 -1.48787498e-02 -3.38022441e-01
5.73503017e-01 9.99172777e-03 -2.82547057e-01 1.68421447e-01
-1.24027026e+00 9.06913042e-01 6.57209098e-01 4.90895212e-02
-3.08031678e-01 9.23405111e-01 1.15336978e+00 5.69312990e-01
-1.04968540e-01 9.88383174e-01 -9.67969894e-01 -2.63434887e-01
-7.40709364e-01 2.58257508e-01 1.10132802e+00 1.22567487e+00
8.77051830e-01 1.11556852e+00 -9.42300484e-02 4.31563050e-01
3.95854712e-02 4.80131388e-01 1.14326382e+00 -8.51048470e-01
4.25252974e-01 8.90437841e-01 -4.23548281e-01 -4.30010378e-01
-5.10836303e-01 -4.94975895e-01 -8.02571774e-01 6.59472048e-01
-5.34082204e-02 -3.26120198e-01 -6.59327328e-01 1.47209370e+00
-3.67278934e-01 -5.40578626e-02 4.32308048e-01 4.76362467e-01
9.98197854e-01 9.03426945e-01 1.29298985e-01 2.83182263e-01
8.83148313e-01 -1.29501784e+00 -5.01034021e-01 -5.16314566e-01
8.54864538e-01 2.33475238e-01 1.29896772e+00 4.27570105e-01
-1.24515891e+00 -7.56723225e-01 -1.49672925e+00 -5.89434803e-02
-7.08446205e-01 5.50519228e-02 1.18982375e+00 5.50724685e-01
-1.34769642e+00 7.94026494e-01 -7.56288528e-01 3.02725405e-01
1.37645960e-01 7.15970039e-01 -4.38621908e-01 7.04099774e-01
-8.07116270e-01 6.00154161e-01 1.27699161e+00 3.50450613e-02
-1.39413357e+00 -4.98150736e-01 -1.31581616e+00 1.97188050e-01
-1.84234187e-01 -1.13568425e+00 1.74300456e+00 -2.09587502e+00
-1.47801042e+00 6.04987085e-01 -3.73026580e-02 -8.79858553e-01
-2.37860426e-01 2.81929020e-02 -4.37731504e-01 -4.80076700e-01
-1.31459549e-01 3.53497237e-01 8.85544479e-01 -1.09053063e+00
-5.90199828e-01 -2.16740265e-01 4.25154030e-01 -5.65413535e-01
4.42705639e-02 8.09388384e-02 -2.15069205e-01 -5.88504493e-01
-5.85991383e-01 -5.18588781e-01 -3.15987408e-01 -1.88024119e-01
-6.40208781e-01 -1.02821290e-01 5.95215142e-01 -9.27409470e-01
1.30932236e+00 -2.03771996e+00 4.72313821e-01 3.14810336e-01
4.55661327e-01 1.36652753e-01 -9.99911353e-02 -1.42199919e-01
-6.44376159e-01 3.65863353e-01 -5.12811661e-01 1.74380168e-01
2.18617573e-01 6.88856840e-01 -5.47818184e-01 1.01482071e-01
2.49361277e-01 1.30227864e+00 -1.07106578e+00 7.73748979e-02
-6.83813766e-02 1.98461771e-01 -8.06076765e-01 9.73158419e-01
-1.15605140e+00 -1.89155728e-01 -4.90566730e-01 9.37446952e-01
1.88445956e-01 -3.89643431e-01 4.03340645e-02 2.76088625e-01
-1.37300864e-01 1.15551122e-01 -5.31011403e-01 1.59278524e+00
-8.60221088e-01 9.36677158e-01 1.45199746e-01 -9.99691606e-01
7.23507166e-01 3.64683390e-01 -4.56022501e-01 -4.10057247e-01
7.17872428e-03 2.55570352e-01 -1.00862801e-01 -8.45077693e-01
6.02277815e-01 1.90889150e-01 -6.43884718e-01 3.38789612e-01
2.34157205e-01 -1.92710327e-03 -5.35104983e-02 -1.38683453e-01
1.26966393e+00 5.54381311e-01 6.39645636e-01 3.11569106e-02
4.89296883e-01 1.24847792e-01 2.64847726e-01 1.18843782e+00
4.10793990e-01 3.01105559e-01 1.01992917e+00 -1.06070042e+00
-1.15985608e+00 -1.27736580e+00 2.37039089e-01 1.75975728e+00
-5.94200730e-01 -5.20760238e-01 -1.08864105e+00 -5.93557537e-01
-2.01862127e-01 1.15507913e+00 -9.03311610e-01 -1.00345172e-01
-8.29314351e-01 -5.51543355e-01 1.28241968e+00 1.01671505e+00
4.06447202e-01 -1.61278915e+00 -1.02201235e+00 4.29704756e-01
4.08591956e-01 -3.39720070e-01 3.49060027e-03 1.09325635e+00
-9.47418392e-01 -9.24711645e-01 -5.92823662e-02 -9.84826624e-01
6.98837101e-01 -7.80216098e-01 1.88811386e+00 6.43224061e-01
2.86553930e-02 3.81181799e-02 7.26050362e-02 -1.94506019e-01
-1.10371029e+00 2.77946144e-01 -3.45788658e-01 -6.41907394e-01
1.44224748e-01 -6.37705803e-01 2.40271315e-02 -4.18805420e-01
-1.13258600e+00 3.67446065e-01 3.74301732e-01 1.23613429e+00
3.68698120e-01 -3.33420694e-01 -3.12242527e-02 -1.22630095e+00
8.81006837e-01 -8.58431637e-01 -9.39680696e-01 1.01186894e-01
-1.88241586e-01 6.58019006e-01 1.53109503e+00 -8.24054554e-02
-8.04101646e-01 1.59996837e-01 -8.86017859e-01 -2.13698149e-01
-3.37536037e-01 8.59745085e-01 -2.30468467e-01 7.86798727e-03
1.31231010e+00 3.96229923e-01 -1.24072641e-01 -2.12772146e-01
4.05617326e-01 5.71120083e-01 1.25679374e+00 -1.12529039e+00
5.38262188e-01 -7.47198313e-02 -2.07199901e-01 -3.16628724e-01
-3.03476781e-01 2.92190671e-01 -3.95255208e-01 8.22657719e-02
8.33034337e-01 -5.07477820e-01 -1.15298700e+00 2.41866738e-01
-1.63581407e+00 -9.66145754e-01 -4.16622847e-01 -3.86605650e-01
-7.77387261e-01 -2.63727099e-01 -6.52776241e-01 -4.09139305e-01
-6.89007998e-01 -1.60384583e+00 8.39948356e-01 9.40768942e-02
-5.20317078e-01 -1.28033185e+00 -9.63870808e-02 -5.53706944e-01
6.50185823e-01 5.38057745e-01 1.42958057e+00 -8.84433627e-01
-7.13553011e-01 -2.18978465e-01 6.07225858e-02 7.58238852e-01
-4.73906070e-01 1.00788131e-01 -8.87253702e-01 1.26853272e-01
-2.31237002e-02 -2.38845706e-01 3.55215520e-01 1.18534327e-01
2.23406935e+00 -9.32771981e-01 -1.47045329e-01 1.34421754e+00
1.62716460e+00 1.07185639e-01 5.48857212e-01 6.87472820e-01
7.45301664e-01 1.41111920e-02 -4.54714477e-01 3.83140236e-01
3.44007909e-02 1.32087007e-01 1.10515547e+00 -1.64843106e-03
1.31756350e-01 -3.53961349e-01 7.20343828e-01 6.29403889e-01
-2.34709792e-02 1.66536167e-01 -1.34926009e+00 4.82677996e-01
-1.60210335e+00 -1.01594913e+00 -1.37187541e-01 1.73422742e+00
8.07762086e-01 1.90749392e-01 -1.14169382e-01 9.67953075e-03
3.30471754e-01 4.50723320e-02 -6.00089788e-01 -1.32545054e+00
-1.30874723e-01 7.78749764e-01 6.69208407e-01 5.57785332e-01
-8.50839317e-01 7.55708992e-01 7.90215111e+00 3.09748709e-01
-1.37899148e+00 1.11117706e-01 4.45705742e-01 1.47560403e-01
-6.66145146e-01 -1.97688967e-01 -5.17481327e-01 4.65338439e-01
1.68790758e+00 -2.44807974e-01 9.16167855e-01 1.67407370e+00
-3.59923482e-01 4.84122902e-01 -1.72074521e+00 5.68833947e-01
-8.31173360e-02 -1.89462113e+00 2.26322964e-01 -5.05110383e-01
4.00010049e-01 3.42178226e-01 2.52785478e-02 1.04270256e+00
6.70560837e-01 -1.56927288e+00 7.87935078e-01 4.61122394e-01
8.07042301e-01 -6.50595427e-01 6.18456006e-01 3.36990923e-01
-9.49757814e-01 -5.80135584e-01 -1.61642089e-01 -3.08192194e-01
-3.70893270e-01 6.22331649e-02 -9.61689770e-01 1.41562939e-01
5.99883914e-01 5.49415231e-01 -7.69352734e-01 5.35533607e-01
-6.68657839e-01 5.37622273e-01 1.69364795e-01 -3.31718773e-01
4.45545286e-01 3.36148083e-01 4.28157240e-01 1.89866853e+00
1.09780528e-01 -4.34461504e-01 -8.21665749e-02 1.40494967e+00
-1.37057737e-01 -2.40495682e-01 -8.29643846e-01 -5.11816777e-02
-3.04912440e-02 7.12327361e-01 -3.13618816e-02 -6.30399108e-01
-5.14750779e-01 1.09190464e+00 5.11611581e-01 7.37909436e-01
-9.48658109e-01 -8.00953984e-01 7.36164808e-01 -2.32202962e-01
9.10256654e-02 1.31055256e-02 -7.92731941e-01 -1.11399424e+00
3.87329757e-02 -1.20073140e+00 1.80392951e-01 -1.21185601e+00
-8.59831214e-01 8.41750920e-01 -7.02190697e-02 -5.37076056e-01
-9.86250162e-01 -1.25463295e+00 -8.12940001e-01 1.00154054e+00
-1.00239396e+00 -8.73185337e-01 -8.27162713e-02 4.37485904e-01
5.38247883e-01 -7.83919394e-01 1.38940883e+00 -8.07934180e-02
-6.28784060e-01 8.06618392e-01 3.09351608e-02 5.16209006e-01
-6.43586218e-01 -1.58829021e+00 9.89729166e-01 8.01394641e-01
-4.37893480e-01 1.23041618e+00 9.09992874e-01 -3.72519463e-01
-2.01476693e+00 -1.59402692e+00 4.92143303e-01 -3.24868530e-01
7.77941108e-01 -2.51987100e-01 -1.04345465e+00 1.47771549e+00
3.26647311e-01 1.21403486e-01 6.45568311e-01 1.40084311e-01
-7.60562837e-01 2.29779765e-01 -1.02331698e+00 7.42298663e-01
8.72879565e-01 -9.07342196e-01 -9.44327414e-01 4.21277672e-01
1.05477703e+00 -8.73770833e-01 -8.57377708e-01 4.25186865e-02
4.27209288e-01 -8.12684655e-01 1.00859904e+00 -1.01793110e+00
9.64792669e-01 -4.49559361e-01 -3.36740136e-01 -1.34407580e+00
-9.94175375e-02 -3.65632564e-01 -7.53723919e-01 6.18880570e-01
6.31336093e-01 -5.58552980e-01 6.80241406e-01 5.24307072e-01
-7.07463801e-01 -6.33594751e-01 -7.87242413e-01 -6.65060818e-01
2.22732574e-01 -9.86080468e-01 1.14400280e+00 6.51328981e-01
9.64153185e-02 2.11937517e-01 -2.09238026e-02 2.47152820e-01
-5.90947503e-03 6.42396510e-02 7.46397555e-01 -7.89907694e-01
-1.15880096e+00 -8.17210853e-01 -4.99445111e-01 -1.09624350e+00
9.91801739e-01 -1.49359322e+00 2.47946396e-01 -1.12804711e+00
-6.51956499e-02 2.29299217e-01 -1.74928866e-02 1.05319095e+00
5.93548454e-02 -4.44887310e-01 -3.16648394e-01 -1.55169323e-01
-3.24075580e-01 1.40164897e-01 4.20428187e-01 -7.06028044e-01
1.48102835e-01 -2.62038678e-01 -5.25469124e-01 7.54671097e-01
8.59884441e-01 -3.97561312e-01 -5.64052053e-02 -1.01155603e+00
7.90177107e-01 4.50919688e-01 6.79892004e-01 -1.07995808e+00
-1.83943135e-03 -2.46115163e-01 3.40515196e-01 1.25110932e-02
-1.48254633e-01 -6.84707463e-01 3.97772014e-01 6.20342195e-01
-6.35034561e-01 8.07391405e-01 4.66094911e-01 6.68864772e-02
-1.99481085e-01 -7.25680590e-01 4.17973220e-01 -6.99558437e-01
-1.20636535e+00 -2.35712063e-02 -7.47749031e-01 -2.19004497e-01
5.24137735e-01 -1.66645259e-01 -3.52782637e-01 1.24332890e-01
-7.40596116e-01 -6.24200627e-02 6.13494992e-01 4.97324854e-01
6.72542393e-01 -1.05916011e+00 -2.74144024e-01 5.82898438e-01
2.40617454e-01 7.12400721e-03 -3.13969165e-01 -1.42383814e-01
-1.09000111e+00 2.69912004e-01 -4.69233066e-01 -5.00042677e-01
-9.05244052e-01 9.55954671e-01 9.10205722e-01 1.67623290e-03
-6.57185197e-01 8.77769947e-01 1.86319500e-01 -1.09604549e+00
2.36668751e-01 -7.97116339e-01 2.11191893e-01 -8.96625340e-01
8.65824938e-01 -9.46533605e-02 5.33704907e-02 -1.33299395e-01
-6.39699167e-03 -3.19072232e-02 5.00112116e-01 1.99935183e-01
1.24309695e+00 8.98826897e-01 -5.75083196e-01 6.92403495e-01
1.47619915e+00 -6.10496700e-01 -8.00947905e-01 2.87476331e-01
1.93211913e-01 1.05895884e-01 -2.08439901e-01 -8.95630836e-01
-1.12648726e+00 8.61873209e-01 3.18910211e-01 1.37574285e-01
1.03028595e+00 -2.80064374e-01 5.00818551e-01 9.44722354e-01
3.77980143e-01 -8.94749045e-01 -1.97291106e-01 1.20646131e+00
9.69841480e-01 -6.78416431e-01 -5.33388197e-01 3.43390077e-01
-1.89450294e-01 1.72142649e+00 7.20869243e-01 -2.48314783e-01
2.08648235e-01 9.68390226e-01 -2.31196508e-01 -1.72813907e-01
-9.56369936e-01 2.32443899e-01 -2.14014769e-01 9.01024997e-01
6.61363721e-01 3.87040079e-01 1.76947519e-01 1.12180185e+00
-5.63476205e-01 3.69753599e-01 5.76415479e-01 8.77219260e-01
-2.16274098e-01 -1.01935232e+00 -5.31072259e-01 7.09683120e-01
-4.61927474e-01 -5.19195676e-01 9.08653513e-02 7.86551416e-01
3.01037997e-01 1.64419055e-01 2.48277366e-01 -6.24427080e-01
2.67170250e-01 3.65238309e-01 1.99778289e-01 -8.89444411e-01
-1.66397452e+00 -1.05994046e+00 1.84556827e-01 -6.84856117e-01
4.45809782e-01 -9.07788947e-02 -1.69687784e+00 -4.07942533e-01
5.65838993e-01 7.39532411e-02 9.52631414e-01 6.58974290e-01
1.89439580e-01 1.14960253e+00 1.19422369e-01 -8.48361850e-01
-5.54242432e-01 -3.62778217e-01 -1.09007426e-01 2.68468171e-01
8.33413601e-01 1.64451376e-01 -2.57771790e-01 3.18296343e-01] | [7.936130046844482, 7.586585521697998] |
1448a455-da47-4d66-ab80-5fc2d85bd148 | leveraging-long-and-short-term-information-in-1 | null | null | http://dx.doi.org/10.1109/tcyb.2019.2896766 | http://dx.doi.org/10.1109/tcyb.2019.2896766 | Leveraging Long and Short-Term Information in Content-Aware Movie Recommendation via Adversarial Training | Movie recommendation systems provide users with ranked lists of movies based on individual’s preferences and constraints. Two types of models are commonly used to generate ranking results: 1) long-term models and 2) session-based models. The long-term-based models represent the interactions between users and movies that are supposed to change slowly across time, while the session-based models encode the information of users’ interests and changing dynamics of movies’ attributes in short terms. In this paper, we propose the LSIC model, leveraging long and short-term information for content-aware movie recommendation using adversarial training. In the adversarial process, we train a generator as an agent of reinforcement learning which recommends the next movie to a user sequentially. We also train a discriminator which attempts to distinguish the generated list of movies from the real records. The poster information of movies is integrated to further improve the performance of movie recommendation, which is specifically essential when few ratings are available. The experiments demonstrate that the proposed model has robust superiority over competitors and achieves the state-of-the-art results. | ['and Ying Shen', 'Xiaojun Chen', 'Zhou Zhao', 'Jianbo Ye', 'Min Yang', 'Benyou Wang', 'Wei Zhao'] | 2020-01-01 | null | null | null | ieee-transactions-on-cybernetics-2020-1 | ['movie-recommendation'] | ['miscellaneous'] | [-6.85567632e-02 -6.38075233e-01 -2.62416989e-01 -7.34150231e-01
-4.63049948e-01 -9.78280842e-01 6.37006998e-01 -3.86081278e-01
-2.86346227e-01 7.07502246e-01 4.66632426e-01 -6.21009618e-02
-3.68102103e-01 -9.43136573e-01 -6.97927833e-01 -6.89028382e-01
-2.13418230e-01 3.77417207e-01 2.71685570e-01 -6.97552204e-01
4.49047297e-01 1.75724998e-01 -1.60360193e+00 7.92727649e-01
7.84209490e-01 1.18158519e+00 2.28964552e-01 7.18309999e-01
3.95250395e-02 1.01401031e+00 -5.86613595e-01 -6.79079533e-01
5.11132479e-01 -5.57716727e-01 -2.42392868e-01 -5.37263528e-02
2.16226995e-01 -6.42126322e-01 -7.31880188e-01 9.01620626e-01
5.03808260e-01 6.88126564e-01 5.98409474e-01 -1.03298974e+00
-1.19354403e+00 8.14174592e-01 1.05154306e-01 3.86496246e-01
4.40442473e-01 5.96443713e-02 1.08257091e+00 -7.40384579e-01
5.30530035e-01 8.16090226e-01 2.19256550e-01 8.69166732e-01
-6.52239203e-01 -6.21614635e-01 7.15963185e-01 2.43682474e-01
-8.11040998e-01 -2.12548211e-01 8.13862741e-01 -5.59759200e-01
3.45658123e-01 4.14718211e-01 5.60547054e-01 1.28768396e+00
3.53418618e-01 5.82271755e-01 8.05553317e-01 1.51528507e-01
2.86753953e-01 4.20663893e-01 -2.25665960e-02 1.37874752e-01
-4.62784588e-01 3.85070741e-01 -3.95736843e-01 -3.35812241e-01
8.90153229e-01 5.52002370e-01 -1.07634075e-01 -1.63117960e-01
-9.58719492e-01 9.43649292e-01 4.03177679e-01 1.60826281e-01
-5.65581262e-01 -1.53732195e-01 2.97339410e-01 9.13319349e-01
4.87122416e-01 6.11490548e-01 -3.95003051e-01 2.09471449e-01
-7.55809188e-01 5.60842335e-01 4.62303638e-01 8.97559345e-01
-7.06760138e-02 6.12406954e-02 -6.30190492e-01 9.57877278e-01
2.45405853e-01 3.18533599e-01 8.23331237e-01 -8.06015134e-01
3.83480430e-01 -6.27218783e-02 7.56525338e-01 -1.02007425e+00
6.96255416e-02 -6.69776618e-01 -7.24349439e-01 -1.50430530e-01
2.07878605e-01 -2.71273911e-01 -6.07483268e-01 1.85671520e+00
6.89266324e-02 4.49133337e-01 2.23894417e-02 1.08451915e+00
9.17777717e-01 1.03249836e+00 -2.36284673e-01 -7.03308403e-01
6.73013985e-01 -1.36629117e+00 -8.57120574e-01 1.64534584e-01
-1.57758430e-01 -6.53384805e-01 1.08710921e+00 6.94157124e-01
-1.38264322e+00 -1.01075673e+00 -8.85804355e-01 4.97174531e-01
-2.56391555e-01 2.46664256e-01 6.31323040e-01 1.81731716e-01
-9.35490251e-01 1.02024639e+00 -1.31426290e-01 9.32443589e-02
-1.62210409e-02 4.78519022e-01 1.63694173e-01 1.52758390e-01
-1.81830943e+00 1.95844471e-01 -2.91340739e-01 1.11016117e-01
-1.21529579e+00 -4.13863748e-01 -1.58175454e-01 1.01205543e-01
3.82455319e-01 -5.85457683e-01 1.31336296e+00 -1.14854312e+00
-1.91401446e+00 2.10735664e-01 3.89660031e-01 -1.51875749e-01
5.17551720e-01 -1.27635881e-01 -1.00547016e+00 -1.12497903e-01
-2.67012209e-01 -2.84524225e-02 1.00663924e+00 -1.17750883e+00
-9.44285095e-01 -1.23430304e-01 6.74498260e-01 4.17584360e-01
-6.40630066e-01 1.54514730e-01 -5.21549821e-01 -1.14877295e+00
-3.42220455e-01 -1.09526086e+00 -3.35869938e-01 -2.82282114e-01
-4.39474024e-02 -2.15326414e-01 1.99254423e-01 -5.69507480e-01
1.45621061e+00 -2.11451030e+00 2.87687182e-01 2.14558139e-01
-2.87355363e-01 2.62922227e-01 -4.22566295e-01 7.00229466e-01
2.46915504e-01 -1.48180231e-01 4.98892128e-01 -1.36778295e-01
-1.75955638e-01 -3.55680026e-02 -8.01400781e-01 -3.97069752e-02
-5.42020857e-01 7.38738716e-01 -1.08547306e+00 1.97619393e-01
-9.21932235e-02 1.67509094e-01 -8.29501987e-01 8.92363369e-01
-4.48554188e-01 7.72480845e-01 -7.51782358e-01 -8.92526843e-03
4.39378917e-01 -3.19028199e-01 3.15462559e-01 -8.83357972e-02
1.12849563e-01 3.86486411e-01 -1.19616163e+00 1.76142085e+00
-5.41853726e-01 -2.70337731e-01 -2.24194169e-01 -5.63922822e-01
8.61956656e-01 4.93032396e-01 4.12468314e-01 -5.54475963e-01
1.07887618e-01 -1.53706402e-01 -4.34958376e-02 -5.94619334e-01
7.04473436e-01 -5.89198619e-02 -1.50446087e-01 6.51524603e-01
5.46718016e-02 1.88684776e-01 1.20550886e-01 5.37906885e-01
8.69990110e-01 1.34368718e-01 -2.84243971e-01 4.65552434e-02
7.74962366e-01 -6.02113783e-01 6.36582673e-01 1.00836825e+00
1.01422377e-01 5.12177527e-01 8.56704712e-02 -4.60564762e-01
-7.02747166e-01 -9.89928365e-01 2.73594797e-01 1.56701827e+00
3.69950563e-01 -2.80124009e-01 -3.50261688e-01 -9.98130202e-01
-5.31808697e-02 7.70822465e-01 -8.03484321e-01 -4.98129904e-01
-2.53415763e-01 -4.12284493e-01 -2.30333939e-01 3.86199445e-01
2.21261545e-03 -1.38856232e+00 2.94691056e-01 4.61582333e-01
-7.26841688e-02 -6.49067879e-01 -1.11890030e+00 -3.31525773e-01
-8.14139783e-01 -5.87225676e-01 -9.30277467e-01 -6.36683822e-01
7.42058754e-01 3.22760910e-01 9.57463920e-01 -3.04211807e-02
5.38140655e-01 1.89153224e-01 -8.10867846e-01 2.16842547e-01
-3.53386611e-01 -3.21631283e-01 4.68463272e-01 5.75849295e-01
-8.43696594e-02 -7.31253803e-01 -1.07166243e+00 6.71971142e-01
-9.36922729e-01 -1.88199043e-01 2.57904619e-01 9.25309777e-01
5.87362409e-01 2.92945445e-01 9.75314915e-01 -1.50426054e+00
9.75622475e-01 -8.13491702e-01 -2.47082144e-01 3.76353592e-01
-7.58796275e-01 -3.17030191e-01 1.57910097e+00 -9.06980217e-01
-1.23967385e+00 -3.34076166e-01 -1.52119726e-01 -3.63449365e-01
3.71147431e-02 3.88259977e-01 -1.34165525e-01 3.80253255e-01
4.51875240e-01 3.14075738e-01 -4.30480897e-01 -8.32753837e-01
5.72364986e-01 8.21647942e-01 5.58452964e-01 -4.51141506e-01
6.49505913e-01 2.10392937e-01 -5.44472873e-01 3.19950521e-01
-1.39074194e+00 -4.12072897e-01 -3.22080702e-01 -4.68693554e-01
5.29294550e-01 -8.23181272e-01 -5.30599535e-01 4.83344764e-01
-6.69097483e-01 -2.39393637e-01 -3.78599226e-01 5.86578488e-01
-5.49908996e-01 -5.35174385e-02 -1.14698648e+00 -5.24125576e-01
-3.79425228e-01 -1.03598249e+00 4.44965839e-01 4.20790702e-01
3.75670999e-01 -9.47708547e-01 2.91701794e-01 3.90344650e-01
5.24559677e-01 -1.19016424e-01 5.47984242e-01 -8.65940094e-01
-3.19786698e-01 -4.18405324e-01 4.41302359e-01 6.84155941e-01
1.47973806e-01 -1.24710687e-02 -5.74540257e-01 -6.45139039e-01
1.41029414e-02 -3.55298132e-01 3.93672794e-01 2.63234019e-01
1.46566677e+00 -6.92078173e-01 1.46731228e-01 5.15048683e-01
1.18130136e+00 7.05565810e-01 5.17771006e-01 -3.65164988e-02
5.67333519e-01 5.37778378e-01 1.13336766e+00 6.45820558e-01
2.76051015e-01 8.55869114e-01 3.30059916e-01 2.83871412e-01
2.46280074e-01 -6.56264842e-01 5.56356490e-01 1.37249482e+00
-2.49861121e-01 -6.51991248e-01 2.23349243e-01 8.12432393e-02
-2.15978599e+00 -1.40983224e+00 1.92291677e-01 2.54544950e+00
7.88241148e-01 3.95023733e-01 3.32383037e-01 -3.13880533e-01
7.66117275e-01 1.87711850e-01 -8.33076119e-01 -4.25899208e-01
2.04986125e-01 -4.33143042e-02 1.27732038e-01 3.89400214e-01
-1.02124608e+00 7.68588543e-01 6.04107237e+00 8.10982883e-01
-1.01530635e+00 1.54310092e-01 5.38033009e-01 -5.00100076e-01
-5.68337739e-01 -2.33951285e-01 -5.80294609e-01 1.05714893e+00
9.31327224e-01 -3.56864035e-01 9.44506407e-01 6.79814219e-01
3.29352111e-01 6.27340615e-01 -1.10093510e+00 5.58158517e-01
2.31761962e-01 -1.16024435e+00 2.68730611e-01 -1.14310250e-01
1.30733943e+00 -3.81390333e-01 4.64374274e-01 5.94239116e-01
5.17808974e-01 -6.87477112e-01 7.55252779e-01 1.19991243e+00
7.16004133e-01 -9.21257973e-01 6.34647369e-01 6.42678201e-01
-9.42816019e-01 -5.38498521e-01 -6.62321389e-01 -1.96977392e-01
3.71422857e-01 3.70100468e-01 5.83330169e-02 6.25700355e-01
5.34673214e-01 1.06455576e+00 -3.25396478e-01 8.93414319e-01
-3.58002752e-01 7.66538978e-01 4.21746343e-01 -9.35754701e-02
4.90834340e-02 -6.70517802e-01 3.60779494e-01 6.54110730e-01
4.23278123e-01 4.53659147e-01 3.87477845e-01 2.43380785e-01
-2.81825632e-01 3.46113503e-01 -2.35180214e-01 1.80733532e-01
4.76265579e-01 1.45578289e+00 -2.08968773e-01 -4.17765886e-01
-4.02131796e-01 1.15796554e+00 2.12357476e-01 3.27517211e-01
-1.03908014e+00 -1.73563763e-01 3.54245067e-01 2.09394485e-01
5.31369686e-01 2.74323761e-01 4.91676480e-01 -1.31478989e+00
-1.35408044e-01 -1.20579648e+00 6.17734849e-01 -9.27363455e-01
-1.85225475e+00 9.00831461e-01 -3.94315898e-01 -1.64551556e+00
-2.43705988e-01 -8.60300437e-02 -8.74917626e-01 6.32650673e-01
-1.12400365e+00 -9.49127197e-01 -2.51968931e-02 9.02170539e-01
7.00109720e-01 -6.18196666e-01 7.30367601e-01 6.25539899e-01
-3.43373656e-01 9.67459500e-01 7.57233202e-01 -2.80863315e-01
1.03313231e+00 -1.12188995e+00 2.05054924e-01 5.55696368e-01
2.65141457e-01 8.80694509e-01 6.45600677e-01 -4.98818219e-01
-1.19315708e+00 -1.22872448e+00 5.46998203e-01 -4.27686542e-01
5.04122853e-01 -3.22956920e-01 -5.64367890e-01 5.35437584e-01
-9.72980186e-02 -6.93576783e-03 9.00520027e-01 -7.98564106e-02
-2.68696755e-01 -4.43360269e-01 -1.14915001e+00 5.10408044e-01
1.04658079e+00 -3.77051741e-01 -5.15147448e-01 6.38125658e-01
8.92010510e-01 -2.82202512e-01 -1.00623429e+00 2.58549750e-01
8.40721488e-01 -8.95989060e-01 9.89624798e-01 -1.22986400e+00
8.06390226e-01 -2.40429997e-01 -4.54413556e-02 -1.63953221e+00
-9.30178404e-01 -7.02592909e-01 -3.95558596e-01 1.40768433e+00
3.58935207e-01 -2.63609320e-01 6.17196679e-01 5.62856019e-01
5.09147309e-02 -8.98416996e-01 -2.81110018e-01 -6.66815758e-01
1.39547391e-02 2.08273113e-01 9.43418443e-01 9.65603411e-01
-5.23995124e-02 4.77885395e-01 -1.22735369e+00 8.70886073e-03
1.94777802e-01 5.02446771e-01 5.61366677e-01 -9.78594720e-01
-1.06325006e+00 -6.34961277e-02 1.83757544e-01 -1.43098867e+00
6.40868116e-03 -8.41749728e-01 -4.86198068e-02 -1.26979756e+00
3.07056010e-01 -5.41144967e-01 -1.19572556e+00 -9.62935165e-02
-2.88458765e-01 2.69120038e-01 1.74945176e-01 3.93322825e-01
-9.64341879e-01 5.89931548e-01 1.62300694e+00 1.70833036e-01
-3.00387830e-01 8.44939351e-01 -8.85489821e-01 4.10053670e-01
6.23814642e-01 -5.54796994e-01 -9.79870260e-01 -2.36048192e-01
4.23295200e-01 5.78764260e-01 -3.08049530e-01 -7.07506239e-01
2.99591064e-01 -5.56764066e-01 2.36215666e-01 -3.79814774e-01
2.77275950e-01 -7.33394682e-01 3.15157861e-01 1.03474408e-01
-1.21491146e+00 1.15119517e-01 -5.54784477e-01 9.61720288e-01
-1.40408918e-01 -3.01405430e-01 5.22176921e-01 -2.16261148e-01
-3.29107046e-01 9.12956059e-01 -2.42630035e-01 -1.69135392e-01
8.66495013e-01 4.40142542e-01 -1.60372823e-01 -1.06848407e+00
-1.07817519e+00 2.40564033e-01 2.38749847e-01 8.38688672e-01
5.44968605e-01 -1.64821053e+00 -6.44085407e-01 -4.53848727e-02
3.69300060e-02 -7.19775319e-01 6.40080512e-01 2.68532149e-02
1.14261299e-01 1.31485403e-01 -2.61395514e-01 2.50132531e-01
-1.08997929e+00 1.13942909e+00 9.59217995e-02 -6.34028494e-01
-8.06293413e-02 1.06219518e+00 1.79305583e-01 -5.93891501e-01
4.48833078e-01 3.58370006e-01 -9.66281950e-01 -3.40033695e-02
6.85623944e-01 2.47446865e-01 -2.46959984e-01 -4.92892385e-01
9.11929905e-02 2.60317236e-01 -5.64048171e-01 -8.54109153e-02
1.39822805e+00 -3.18221211e-01 2.22248271e-01 6.46772444e-01
8.17559421e-01 4.03220922e-01 -1.33170772e+00 -4.16178107e-01
-7.78170288e-01 -8.68098021e-01 -1.85111649e-02 -1.25494134e+00
-1.44919240e+00 5.56237996e-01 5.18791735e-01 5.27172983e-01
1.11874104e+00 -3.58898222e-01 1.04716897e+00 6.02211524e-03
5.96701324e-01 -1.13728392e+00 3.65363091e-01 2.98281193e-01
9.81383562e-01 -9.71791089e-01 -9.84882377e-03 -8.43583420e-02
-9.40426230e-01 7.85806715e-01 7.86862612e-01 -5.53782046e-01
9.41295266e-01 -3.18451911e-01 1.12676635e-01 1.17801763e-01
-1.18351161e+00 1.77349120e-01 6.21012032e-01 2.39782944e-01
3.62443060e-01 1.40030608e-01 -6.83565438e-01 1.49909091e+00
-4.86924350e-02 1.61738858e-01 4.86194521e-01 4.84075367e-01
-1.79075032e-01 -1.46355104e+00 1.96900010e-01 8.62541556e-01
-7.07943738e-01 -3.03676352e-02 -8.45896825e-02 -1.87760428e-01
2.41543785e-01 1.26380265e+00 -3.13115418e-02 -9.52179909e-01
5.45499086e-01 -3.47112566e-01 2.09078386e-01 -6.95766389e-01
-9.81286764e-01 -9.07430146e-03 4.36258391e-02 -4.97620940e-01
-1.93466365e-01 -5.56077421e-01 -7.36450315e-01 -2.38708258e-01
-5.61134636e-01 6.90181196e-01 3.94618779e-01 5.89562416e-01
5.17545760e-01 6.92650378e-01 1.76971436e+00 -6.73674345e-01
-1.20318127e+00 -8.06231558e-01 -9.14441764e-01 9.84746337e-01
-4.98672910e-02 -4.56954360e-01 -4.63109046e-01 1.84812889e-01] | [10.120061874389648, 5.596893787384033] |
03b69d89-e27c-4214-baf9-481e1c324bd7 | interpretable-edge-enhancement-and | 2209.09483 | null | https://arxiv.org/abs/2209.09483v1 | https://arxiv.org/pdf/2209.09483v1.pdf | Interpretable Edge Enhancement and Suppression Learning for 3D Point Cloud Segmentation | 3D point clouds can flexibly represent continuous surfaces and can be used for various applications; however, the lack of structural information makes point cloud recognition challenging. Recent edge-aware methods mainly use edge information as an extra feature that describes local structures to facilitate learning. Although these methods show that incorporating edges into the network design is beneficial, they generally lack interpretability, making users wonder how exactly edges help. To shed light on this issue, in this study, we propose the Diffusion Unit (DU) that handles edges in an interpretable manner while providing decent improvement. Our method is interpretable in three ways. First, we theoretically show that DU learns to perform task-beneficial edge enhancement and suppression. Second, we experimentally observe and verify the edge enhancement and suppression behavior. Third, we empirically demonstrate that this behavior contributes to performance improvement. Extensive experiments performed on challenging benchmarks verify the superiority of DU in terms of both interpretability and performance gain. Specifically, our method achieves state-of-the-art performance in object part segmentation using ShapeNet part and scene segmentation using S3DIS. Our source code will be released at https://github.com/martianxiu/DiffusionUnit. | ['Masashi Matsuoka', 'Qiong Chang', 'Takayuki Shinohara', 'Kyoung-Sook Kim', 'Weimin WANG', 'Xin Liu', 'Haoyi Xiu'] | 2022-09-20 | null | null | null | null | ['scene-segmentation', 'point-cloud-segmentation'] | ['computer-vision', 'computer-vision'] | [-1.49552645e-02 1.15539849e-01 -4.00406688e-01 -3.61830115e-01
-2.33496383e-01 -6.20112658e-01 2.45996252e-01 3.45432684e-02
2.71243244e-01 2.31056958e-01 3.53617892e-02 -4.62908387e-01
-1.19695559e-01 -7.93358564e-01 -1.10396111e+00 -4.63151336e-01
-7.08788112e-02 2.83927232e-01 3.40717852e-01 -7.47794658e-02
2.43373901e-01 8.59920859e-01 -1.42308474e+00 1.85492724e-01
1.09593821e+00 1.09221709e+00 1.59437269e-01 2.24946737e-01
-3.32522333e-01 4.12504405e-01 -2.18866035e-01 -1.81862473e-01
5.77046454e-01 8.26706216e-02 -7.17925906e-01 3.39824915e-01
5.22178829e-01 -2.99214303e-01 -2.18991697e-01 1.04826510e+00
7.54747093e-02 -1.46371974e-02 6.69249415e-01 -1.47548664e+00
-6.93291724e-01 3.73927265e-01 -8.27015758e-01 -3.65424193e-02
4.15876880e-02 2.25557551e-01 1.28293395e+00 -9.86349344e-01
5.45078814e-01 1.22082937e+00 6.67874813e-01 3.49287450e-01
-1.07141185e+00 -6.91543043e-01 4.73040521e-01 1.57101557e-01
-1.31430972e+00 -1.20217495e-01 1.25663507e+00 -1.77826837e-01
7.96076715e-01 4.60226864e-01 7.54065335e-01 5.99533916e-01
-1.86662316e-01 1.33048332e+00 8.71136427e-01 -1.17817596e-01
1.43108934e-01 -2.02545986e-01 3.95686507e-01 7.54124641e-01
4.65291113e-01 -6.59836233e-02 -3.94728392e-01 1.44292163e-02
1.04231203e+00 2.97527760e-01 -5.21341205e-01 -6.52131736e-01
-9.36796725e-01 6.18345916e-01 9.53393638e-01 2.04855613e-02
-4.10140634e-01 3.34880710e-01 2.05380663e-01 1.75002709e-01
6.83066308e-01 3.07261705e-01 -4.83008921e-01 2.38670483e-02
-5.68776071e-01 -5.15217707e-02 6.93827450e-01 1.24443388e+00
9.00188625e-01 -5.02612954e-03 5.44461459e-02 8.59767079e-01
3.85368377e-01 4.40215200e-01 -1.77269474e-01 -1.13186491e+00
2.45334059e-01 1.12360156e+00 -1.79562300e-01 -1.04280078e+00
-2.79973954e-01 -4.51555848e-01 -9.02495861e-01 4.67982054e-01
2.06592411e-01 2.47863397e-01 -1.09939075e+00 1.38658547e+00
5.10968983e-01 4.22141343e-01 -3.77492666e-01 1.03115106e+00
9.00672853e-01 4.80249643e-01 -1.29755199e-01 4.03278202e-01
1.39657795e+00 -1.19870520e+00 -3.30423385e-01 -3.71301740e-01
4.24265504e-01 -6.00318789e-01 1.19232929e+00 3.16376477e-01
-9.99857724e-01 -4.17253643e-01 -1.00882506e+00 -1.48784950e-01
-1.09450869e-01 1.23238638e-02 1.04287398e+00 5.14163494e-01
-9.35106874e-01 6.81127608e-01 -1.08521962e+00 -1.62632335e-02
8.81737113e-01 4.32606608e-01 -2.25194335e-01 -1.24445699e-01
-5.67349672e-01 2.65266806e-01 2.19936445e-01 1.33872023e-02
-3.30919474e-01 -7.59979069e-01 -7.89304316e-01 2.42050707e-01
5.73933959e-01 -8.58791769e-01 1.22401655e+00 -9.93341625e-01
-1.18988597e+00 6.83645010e-01 -3.54813129e-01 -2.85831004e-01
5.54821193e-01 -2.05177128e-01 5.60334362e-02 3.77996981e-01
-1.20882832e-01 8.87504697e-01 7.09179640e-01 -1.72962856e+00
-6.10825539e-01 -5.64081967e-01 3.54108840e-01 1.66414618e-01
-3.62049341e-01 -5.23195207e-01 -9.29325223e-01 -7.39631534e-01
5.09517848e-01 -9.84178603e-01 -7.38466159e-02 5.88728011e-01
-5.29623806e-01 -4.49006319e-01 1.04586482e+00 -3.53812873e-01
9.28371489e-01 -2.17463160e+00 -3.01217616e-01 4.15636837e-01
7.13178396e-01 1.42977372e-01 -4.80151735e-02 6.03432469e-02
-4.89739887e-03 4.52908069e-01 -4.58601683e-01 -4.86270726e-01
7.66329989e-02 3.68425280e-01 -1.18310384e-01 3.41701448e-01
4.18441892e-01 1.26136208e+00 -6.92902505e-01 -2.93574244e-01
3.81657064e-01 6.42025054e-01 -6.91671133e-01 -8.68403986e-02
-2.55434215e-01 3.53605747e-01 -7.20780551e-01 1.07486308e+00
1.03562903e+00 -6.39386058e-01 -1.74406648e-01 -4.03626323e-01
6.04976900e-02 2.85308242e-01 -1.17164838e+00 1.44869840e+00
-3.43852937e-01 7.10250854e-01 1.73452318e-01 -9.48698461e-01
8.71343911e-01 -1.02771949e-02 5.69577217e-01 -4.69786674e-01
1.71393350e-01 1.83790147e-01 -1.13414310e-01 -2.03137800e-01
4.14833575e-01 2.20688745e-01 4.84616011e-01 3.86567235e-01
-4.77904618e-01 -2.09942058e-01 -9.05662477e-02 7.12788999e-02
9.63321209e-01 8.18144977e-02 1.38900042e-01 -3.15391153e-01
2.29656190e-01 6.95877671e-02 5.33468962e-01 6.32375777e-01
-1.12640612e-01 7.82849669e-01 4.44855809e-01 -3.30283105e-01
-8.98889780e-01 -9.67531383e-01 -2.63107419e-01 8.32281232e-01
7.71295011e-01 -3.20520014e-01 -6.42170131e-01 -7.50354946e-01
2.91866302e-01 7.37109303e-01 -5.34259796e-01 -1.43119767e-01
-4.61514682e-01 -3.65206778e-01 2.76308060e-01 9.75079894e-01
6.59126878e-01 -7.93640375e-01 -3.74476999e-01 -2.64812589e-01
-1.09177060e-01 -1.07768345e+00 -6.18424356e-01 -1.41108995e-02
-1.30848360e+00 -1.10827541e+00 -4.89823937e-01 -7.78654397e-01
1.03761697e+00 9.72229183e-01 1.24754620e+00 5.23599029e-01
-2.32936293e-02 5.00098407e-01 -3.81535053e-01 -5.01082420e-01
-1.07395768e-01 1.38721362e-01 -3.10687125e-01 -2.52000779e-01
4.16625232e-01 -8.13977599e-01 -8.38019788e-01 5.16937971e-01
-1.02478933e+00 3.05769444e-01 6.36281610e-01 6.49989128e-01
8.43964338e-01 -5.74844927e-02 1.69545114e-01 -1.01721549e+00
4.94235873e-01 -3.64153743e-01 -5.40979028e-01 1.35197774e-01
-6.68376088e-01 2.10490718e-01 3.09118748e-01 -2.55743295e-01
-8.42391789e-01 1.22993678e-01 -2.07224011e-01 -7.62885690e-01
-2.43402332e-01 3.45915556e-01 -1.99422956e-01 -4.75696415e-01
4.35300231e-01 4.85762581e-02 9.76527296e-03 -5.11351764e-01
3.52007329e-01 6.67054236e-01 3.36419672e-01 -7.36304641e-01
8.30143809e-01 8.32979739e-01 8.96565840e-02 -8.90394270e-01
-6.54308438e-01 -7.71732926e-01 -4.75702643e-01 -2.23520726e-01
5.33586621e-01 -8.73943627e-01 -9.31415737e-01 3.21895421e-01
-1.24633980e+00 -4.27593738e-01 -1.64184883e-01 4.11389470e-02
-3.97734225e-01 4.39392567e-01 -4.86086786e-01 -7.17711508e-01
-4.44652051e-01 -1.16467404e+00 1.27845144e+00 2.61090040e-01
1.11144319e-01 -1.13704073e+00 -5.43552577e-01 3.79048258e-01
1.97140157e-01 3.08089375e-01 7.24187672e-01 -3.44487637e-01
-1.10775614e+00 -2.73006596e-02 -5.93277335e-01 3.75553280e-01
3.55622888e-01 2.48631611e-02 -9.44568217e-01 -1.05114043e-01
-4.60900329e-02 1.29575536e-01 9.37542021e-01 3.92345548e-01
1.66257262e+00 -1.44146398e-01 -4.37129170e-01 8.62960994e-01
1.36321974e+00 -1.36899382e-01 6.09531820e-01 1.84831813e-01
1.08993471e+00 3.04295570e-01 4.51728523e-01 2.36375749e-01
5.00427723e-01 5.32970130e-01 8.28611255e-01 -3.66704881e-01
-3.38212132e-01 -2.27088571e-01 -6.68892339e-02 5.84836185e-01
-3.64271343e-01 -3.14972818e-01 -1.09193397e+00 4.97403353e-01
-1.94391525e+00 -5.59311271e-01 -4.97306705e-01 1.96697915e+00
6.15731597e-01 4.48174238e-01 -1.26362843e-02 3.66558954e-02
6.53222620e-01 2.70025153e-02 -8.50463927e-01 -6.47129607e-04
-1.02791987e-01 1.58172622e-02 7.00029731e-01 3.77031028e-01
-9.32722867e-01 9.45367157e-01 5.30312967e+00 7.93092608e-01
-1.21662104e+00 -9.76616740e-02 9.01026249e-01 1.55642644e-01
-5.66304505e-01 -9.33160111e-02 -5.80559313e-01 2.57138669e-01
1.22960016e-01 -9.84089226e-02 3.23695570e-01 1.12956011e+00
1.49918199e-01 1.15408257e-01 -1.10044432e+00 1.04571164e+00
-6.05740547e-02 -1.54201603e+00 6.39989898e-02 1.70929626e-01
7.99554348e-01 3.17798495e-01 -5.74753545e-02 -1.22782192e-03
4.44133170e-02 -8.32468331e-01 8.22885633e-01 2.45209500e-01
5.48096538e-01 -7.25560725e-01 4.85807389e-01 3.16898048e-01
-1.49096501e+00 1.86225861e-01 -3.65459949e-01 -6.97511435e-02
8.18590596e-02 7.28367150e-01 -8.39336693e-01 6.16146147e-01
6.62591994e-01 8.34097505e-01 -4.89231735e-01 1.31359780e+00
-4.90980148e-01 7.24335313e-01 -6.52873576e-01 4.94787097e-02
2.45683387e-01 -3.34787905e-01 8.13839674e-01 1.05044436e+00
1.27851114e-01 1.65924683e-01 2.47740388e-01 1.19374073e+00
-4.56542909e-01 -6.94828108e-02 -6.73940957e-01 5.46303317e-02
5.20062029e-01 1.19790077e+00 -1.14840019e+00 -1.83681950e-01
-4.56915557e-01 9.73687470e-01 3.19510967e-01 3.54195505e-01
-7.28718638e-01 -2.18429849e-01 9.71331000e-01 2.79677421e-01
5.35002351e-01 -4.04484868e-01 -9.28241313e-01 -1.06547868e+00
3.33048761e-01 -5.62950373e-01 7.74242333e-04 -7.03429818e-01
-1.42277753e+00 3.69029135e-01 1.05505437e-02 -1.31347954e+00
3.18015933e-01 -8.18358779e-01 -6.20209217e-01 4.40126479e-01
-1.72511721e+00 -1.29020536e+00 -6.59010887e-01 2.17652455e-01
7.13871241e-01 4.16958719e-01 2.10107729e-01 3.24313372e-01
-3.85165155e-01 6.25545323e-01 -1.05971478e-01 1.85771912e-01
3.59900981e-01 -1.35471404e+00 8.39232028e-01 6.13011420e-01
3.23690355e-01 6.42000318e-01 3.87063712e-01 -7.25784242e-01
-1.53362572e+00 -1.15828514e+00 2.88308501e-01 -3.94729376e-01
4.00929660e-01 -3.31395507e-01 -1.20742035e+00 6.35278881e-01
-4.79168929e-02 1.52430534e-01 3.19298863e-01 1.14682466e-01
-3.84701848e-01 9.84554738e-02 -1.05046523e+00 7.51647055e-01
1.51015937e+00 -2.65017897e-01 -4.15449798e-01 2.86446601e-01
9.44688499e-01 -6.40533030e-01 -7.48716116e-01 6.48777127e-01
3.45286429e-01 -9.59739447e-01 1.17410028e+00 -3.48457307e-01
5.45311332e-01 -4.29313064e-01 2.70996317e-02 -1.11443770e+00
-3.77243668e-01 -3.98273498e-01 -4.12890583e-01 1.00614882e+00
3.50844830e-01 -9.01682615e-01 1.06778395e+00 6.92262888e-01
-5.28774679e-01 -1.15060580e+00 -6.48726285e-01 -1.02976692e+00
6.87749662e-06 -7.24567950e-01 8.82480204e-01 9.18265820e-01
-4.14150745e-01 1.82213232e-01 2.73275822e-02 2.82578856e-01
6.24392629e-01 4.31346416e-01 8.37052763e-01 -1.37325776e+00
-9.23744589e-02 -8.34200263e-01 -4.46024984e-01 -1.69498003e+00
9.62110907e-02 -9.77134466e-01 5.87070175e-02 -1.80771840e+00
7.44076222e-02 -7.27809846e-01 -1.26055166e-01 7.48631358e-01
-3.78916144e-01 3.57028186e-01 2.49135539e-01 3.43679011e-01
-4.88601297e-01 6.23192430e-01 1.49090362e+00 -2.14330554e-01
-2.54076660e-01 2.17984408e-01 -9.07931745e-01 8.75583947e-01
1.05843759e+00 -2.86856055e-01 -4.51490968e-01 -7.30275631e-01
5.87829389e-02 -4.63678867e-01 5.57035863e-01 -7.79345691e-01
1.70693621e-01 -2.06155270e-01 1.48762003e-01 -6.94463253e-01
4.00840938e-01 -9.80865061e-01 -3.52425426e-02 2.99460113e-01
1.95980102e-01 -1.07127033e-01 4.39742863e-01 6.86392426e-01
-1.44477352e-01 -1.23506933e-01 5.54234266e-01 4.84054834e-02
-8.58532548e-01 6.19967461e-01 2.53441900e-01 3.25567052e-02
9.06392038e-01 -5.83382547e-01 -1.88650981e-01 -3.74421567e-01
-2.59579539e-01 5.22365093e-01 8.70569825e-01 4.20733124e-01
7.39640296e-01 -1.32805741e+00 -5.48564017e-01 1.37712613e-01
1.65397152e-01 5.08300900e-01 7.54745156e-02 8.06361198e-01
-7.63907790e-01 1.96531639e-01 2.10448243e-02 -9.99239028e-01
-1.20639563e+00 2.09785581e-01 1.75354198e-01 1.69069797e-01
-9.44515944e-01 8.85457754e-01 4.91019040e-01 -3.95063996e-01
2.81184524e-01 -6.35091603e-01 2.00180277e-01 -3.83568436e-01
1.05171844e-01 3.83583933e-01 1.26679838e-01 -2.97007293e-01
-3.22004318e-01 6.51627779e-01 -3.15355957e-02 3.67176741e-01
1.50547981e+00 6.24167584e-02 -5.31762615e-02 1.61547557e-01
1.10080075e+00 -2.04960830e-04 -1.57646322e+00 -1.33739978e-01
-1.41454890e-01 -6.40395999e-01 2.31332690e-01 -4.85720903e-01
-1.49894357e+00 7.89580703e-01 4.08129573e-01 4.33650076e-01
1.21161342e+00 3.16092640e-01 9.52748775e-01 4.69067931e-01
2.20766529e-01 -7.87686884e-01 4.74120602e-02 4.31518942e-01
8.39966774e-01 -1.45533848e+00 8.80603418e-02 -1.21282887e+00
-4.88829255e-01 1.01178789e+00 6.45559967e-01 -1.82423502e-01
7.97582150e-01 2.54936934e-01 -1.35147452e-01 -5.47374904e-01
-4.60939527e-01 -2.33875334e-01 6.04449153e-01 4.81348187e-01
2.99505413e-01 1.75897822e-01 -1.56019717e-01 4.23661351e-01
-1.79381579e-01 -1.99048594e-01 1.74800962e-01 9.11762416e-01
-4.20303106e-01 -1.02287245e+00 -2.81520903e-01 5.32156289e-01
6.36170525e-03 1.12405885e-02 -6.10508025e-01 8.54404867e-01
-1.42547265e-01 6.35560572e-01 1.66793108e-01 -2.04459697e-01
4.15134847e-01 -2.57464588e-01 3.37507874e-01 -4.45712984e-01
-2.49023020e-01 -5.24457730e-02 -1.61584079e-01 -6.21183932e-01
-2.60567844e-01 -4.86203879e-01 -1.66685581e+00 -3.59670162e-01
-5.04656672e-01 -1.53922439e-01 7.40097880e-01 8.03908765e-01
7.60147572e-01 5.98250866e-01 6.04508519e-01 -9.14142191e-01
-2.75022924e-01 -5.67810059e-01 -2.91494161e-01 5.19727111e-01
2.58480370e-01 -8.05218935e-01 -4.24753040e-01 -3.86491418e-02] | [7.928774833679199, -3.3069980144500732] |
b9e060ef-0094-4084-94d8-190dc926b141 | color-constancy-by-reweighting-image-feature | 1806.09248 | null | https://arxiv.org/abs/1806.09248v3 | https://arxiv.org/pdf/1806.09248v3.pdf | Color Constancy by Reweighting Image Feature Maps | In this study, a novel illuminant color estimation framework is proposed for computational color constancy, which incorporates the high representational capacity of deep-learning-based models and the great interpretability of assumption-based models. The well-designed building block, feature map reweight unit (ReWU), helps to achieve comparative accuracy on benchmark datasets with respect to prior state-of-the-art deep learning based models while requiring more compact model size and cheaper computational cost. In addition to local color estimation, a confidence estimation branch is also included such that the model is able to simultaneously produce point estimate and its uncertainty estimate, which provides useful clues for local estimates aggregation and multiple illumination estimation. The source code and the dataset have been made available at https://github.com/QiuJueqin/Reweight-CC. | ['Zhengnan Ye', 'Jueqin Qiu', 'Haisong Xu'] | 2018-06-25 | null | null | null | null | ['color-constancy'] | ['computer-vision'] | [-3.99200976e-01 -4.13522333e-01 -1.17878713e-01 -3.78583342e-01
-7.05618739e-01 -2.47394249e-01 4.88884151e-01 -2.21757352e-01
-1.24959543e-01 8.36297333e-01 -1.52605459e-01 -2.06027806e-01
4.84553762e-02 -7.22761333e-01 -6.78423703e-01 -8.67603660e-01
1.44506022e-01 -3.56444158e-02 -6.65687099e-02 1.45988300e-01
2.92981386e-01 5.02291322e-01 -1.59123921e+00 -1.20410752e-02
1.20845568e+00 1.44082940e+00 2.83263862e-01 6.16770923e-01
-1.52249783e-01 4.84385967e-01 -2.87363112e-01 -4.77313817e-01
2.62824953e-01 -4.69070166e-01 -1.92921445e-01 -1.77435488e-01
7.33771741e-01 -5.74645638e-01 -1.47814527e-01 9.83108163e-01
5.84167242e-01 7.48486519e-02 7.99736917e-01 -1.26214170e+00
-9.24910545e-01 -5.61558918e-05 -7.52956450e-01 -1.66227877e-01
-2.51582619e-02 3.75960320e-01 1.06211817e+00 -1.03487885e+00
2.78993607e-01 9.11620021e-01 7.06670821e-01 2.66567200e-01
-1.15750217e+00 -7.86233246e-01 1.79587126e-01 3.85201395e-01
-1.42627597e+00 -3.70252401e-01 8.98645818e-01 -2.14573950e-01
6.09476686e-01 2.71807373e-01 8.81602645e-01 1.00332344e+00
2.40760177e-01 8.49903047e-01 1.32557762e+00 -4.67254549e-01
2.78215617e-01 1.08413852e-03 -3.67690444e-01 8.80985200e-01
4.51035976e-01 3.82977813e-01 -4.90977705e-01 1.86968431e-01
1.24298775e+00 -6.69876784e-02 -3.01909775e-01 -4.75536883e-01
-9.16566014e-01 7.32972503e-01 9.76284444e-01 1.59443617e-02
-3.84311557e-01 6.08335257e-01 6.26812279e-02 -2.33341217e-01
8.73209417e-01 2.61520952e-01 -5.10644317e-01 -5.41503467e-02
-8.68722975e-01 -1.26444414e-01 2.18421534e-01 9.40451503e-01
9.61875021e-01 4.24079359e-01 -1.52497053e-01 8.55152547e-01
4.79817182e-01 7.87049413e-01 5.40554803e-03 -1.02433765e+00
4.46352735e-02 6.35353267e-01 2.14763373e-01 -1.08512127e+00
-4.15951043e-01 -6.79648101e-01 -9.03688490e-01 5.69182217e-01
3.56935680e-01 -1.41552821e-01 -9.24417436e-01 1.56549919e+00
3.52650434e-01 4.23016429e-01 -3.38947415e-01 1.03600514e+00
8.00487459e-01 4.74342823e-01 -4.73619904e-03 2.13018462e-01
1.18240213e+00 -9.64451015e-01 -5.82743526e-01 -6.28058314e-02
1.73924267e-01 -8.57768834e-01 1.10633385e+00 3.79215598e-01
-9.66653347e-01 -8.36218178e-01 -1.10323882e+00 -3.15719485e-01
-4.85315025e-01 6.01505697e-01 1.04066753e+00 6.13875449e-01
-1.08985031e+00 3.15091789e-01 -6.78695440e-01 -1.86933532e-01
6.13909245e-01 -1.82514742e-01 -1.12442248e-01 -2.50145972e-01
-9.86202657e-01 8.96701753e-01 1.23223439e-01 4.81769115e-01
-8.23480368e-01 -7.57578492e-01 -9.58746731e-01 -7.65510350e-02
1.20452136e-01 -6.81861103e-01 9.64091182e-01 -1.00157726e+00
-1.55682909e+00 4.96514827e-01 -3.64936620e-01 4.89353985e-02
7.82415628e-01 -3.50605845e-01 -2.99035549e-01 1.30217955e-01
-2.81643003e-01 6.68342292e-01 9.86286104e-01 -1.71402752e+00
-3.96952957e-01 -2.43348494e-01 -1.53558746e-01 7.65022337e-02
-1.51389286e-01 -2.69416541e-01 -7.46987641e-01 -5.64386904e-01
1.21079952e-01 -6.35669172e-01 2.28269096e-03 6.86698973e-01
-3.13603729e-01 2.02402123e-03 4.28838074e-01 -7.04421997e-01
1.08315384e+00 -1.94332528e+00 -1.55130014e-01 2.28403106e-01
1.71707168e-01 1.23868302e-01 -3.63817513e-01 2.79803693e-01
-2.23719217e-02 -1.35854244e-01 -2.66488820e-01 -5.80610096e-01
1.87934071e-01 -2.09304895e-02 -1.13414750e-02 7.07149267e-01
3.41839403e-01 1.16739917e+00 -7.72856593e-01 -3.37525517e-01
8.17019284e-01 7.38424063e-01 -1.74942911e-01 1.15056835e-01
-1.19713299e-01 1.94081575e-01 -1.99505746e-01 9.18230832e-01
1.32384562e+00 -2.00460076e-01 -5.46102971e-02 -6.03888750e-01
-3.92975062e-01 -1.00396223e-01 -1.36564529e+00 1.76770639e+00
-6.90037012e-01 8.12071264e-01 3.41266654e-02 -3.78126681e-01
9.41852570e-01 -8.90346244e-02 3.08832884e-01 -9.64971781e-01
3.09302121e-01 1.27626434e-01 -2.26361409e-01 -2.10769072e-01
4.36623156e-01 1.67150363e-01 4.78162199e-01 3.23801309e-01
-3.40886265e-02 -3.58430237e-01 -1.34537071e-01 -2.98961475e-02
1.71352670e-01 8.63185883e-01 2.03997940e-01 -4.06821668e-01
2.99605519e-01 -2.88602650e-01 5.43957353e-01 4.68394041e-01
-3.01133901e-01 8.15150321e-01 1.25668615e-01 -3.44137311e-01
-1.05819166e+00 -1.06849825e+00 -5.83214223e-01 8.59216273e-01
3.74650121e-01 -1.79013610e-01 -5.14359355e-01 -2.64790148e-01
2.57948309e-01 8.96439195e-01 -9.43340540e-01 4.16626930e-02
-8.15530345e-02 -6.90773606e-01 2.05051109e-01 8.03637147e-01
7.23534405e-01 -6.08359516e-01 -6.04217887e-01 -1.35351181e-01
-1.52325869e-01 -5.96003354e-01 -2.34796897e-01 1.38696164e-01
-6.70636475e-01 -1.21024489e+00 -8.25919509e-01 -2.63080299e-01
7.29804337e-01 3.24011028e-01 1.21325564e+00 1.49135649e-01
-5.55398226e-01 5.15915155e-01 -3.40865791e-01 -7.02472985e-01
1.46258980e-01 -3.81145477e-01 -2.19216660e-01 -1.13563484e-03
4.26458240e-01 -1.38771087e-01 -1.07068849e+00 2.85281062e-01
-7.01603949e-01 4.99261677e-01 5.65508425e-01 9.80491579e-01
6.65479898e-01 -2.21728519e-01 1.92764863e-01 -3.71709704e-01
2.17173293e-01 -2.87982821e-01 -9.20419931e-01 3.62106442e-01
-6.74710274e-01 -3.77539024e-02 3.80614072e-01 -1.79351214e-02
-1.45343590e+00 -1.32739693e-01 7.70655554e-03 -2.31249571e-01
-1.78335145e-01 2.68716991e-01 3.03672906e-02 -3.61954123e-01
5.49553275e-01 2.97164530e-01 2.43640132e-02 -3.90151888e-01
6.25174880e-01 3.62611473e-01 2.55707294e-01 -6.63578808e-01
7.17041552e-01 5.29451609e-01 1.99326366e-01 -6.34820879e-01
-5.86593807e-01 -3.62443268e-01 -8.88757646e-01 -6.26511693e-01
4.84684706e-01 -1.07893062e+00 -7.46607423e-01 7.79342294e-01
-9.31589901e-01 -6.07209861e-01 5.86453415e-02 3.40962499e-01
-4.14789677e-01 3.32737982e-01 -3.56722206e-01 -1.12849116e+00
-1.69796899e-01 -8.49543273e-01 1.13881242e+00 5.36274016e-01
2.50813365e-01 -1.27484274e+00 -8.56600776e-02 2.96206027e-02
7.47227669e-01 5.41756928e-01 6.27559185e-01 2.96704918e-01
-7.81861246e-01 -2.16317266e-01 -9.01660383e-01 3.82716119e-01
3.05035144e-01 6.26346052e-01 -1.46873212e+00 -2.02319235e-01
-3.88286054e-01 -3.66486847e-01 1.18535471e+00 8.70173812e-01
1.45100832e+00 1.37309805e-01 -4.83716913e-02 9.67655599e-01
1.97409570e+00 -1.03679366e-01 7.35566318e-01 3.86819392e-01
7.55047739e-01 3.81597549e-01 5.73001921e-01 7.06196666e-01
4.39137399e-01 5.18530667e-01 7.40596652e-01 -6.37537479e-01
-5.18899620e-01 -8.71543959e-02 -2.57872283e-01 4.22906488e-01
-4.74002481e-01 -1.66156456e-01 -7.13517904e-01 2.79254258e-01
-1.90547681e+00 -7.49746621e-01 -2.56495476e-01 2.19652653e+00
7.12162554e-01 -2.77997166e-01 -1.92261413e-01 -1.22670621e-01
4.40517604e-01 2.90916562e-01 -7.76873946e-01 -2.09747627e-01
-2.62415439e-01 3.96709889e-02 5.68877995e-01 4.87252265e-01
-9.92479205e-01 7.84179688e-01 6.29478121e+00 8.62031281e-01
-1.19035721e+00 -6.48933509e-03 8.80052805e-01 2.83969883e-02
-4.35302466e-01 -2.87281156e-01 -4.89174604e-01 4.55409169e-01
6.19993627e-01 3.03829521e-01 5.13074696e-01 7.80317426e-01
3.57127309e-01 -5.81485152e-01 -7.28917539e-01 1.07033134e+00
1.32660449e-01 -1.34506989e+00 -8.65048021e-02 -5.91706112e-02
8.72841299e-01 1.64867416e-01 3.37276012e-01 -8.35647061e-02
2.01299489e-01 -9.08866823e-01 7.87841797e-01 8.86061668e-01
1.18974435e+00 -7.85742462e-01 8.20653141e-01 -2.46176854e-01
-1.45400834e+00 -6.26599416e-02 -7.59694099e-01 2.23784260e-02
-5.73020540e-02 7.47425795e-01 -3.79785240e-01 9.41776037e-01
7.32128084e-01 8.24715257e-01 -8.96307647e-01 1.42908669e+00
-5.41020632e-01 4.39662904e-01 -2.49067321e-01 -7.01764822e-02
1.77794784e-01 -5.06967783e-01 -1.71012264e-02 1.34207773e+00
3.00648689e-01 -2.23278239e-01 -1.97863817e-01 1.17230308e+00
1.36332169e-01 -3.79987769e-02 -2.81975627e-01 3.33183765e-01
4.77941602e-01 1.60570967e+00 -6.20831370e-01 -1.83964849e-01
-5.54661870e-01 1.12073505e+00 4.08423990e-01 7.16374218e-01
-1.08028221e+00 -4.10735905e-01 8.14406216e-01 -1.42970353e-01
2.77359933e-01 -5.06491661e-01 -6.54914796e-01 -1.15371168e+00
-2.98558585e-02 -2.73407966e-01 9.92913172e-02 -1.31442261e+00
-1.37460256e+00 5.14420331e-01 -1.92663506e-01 -1.39961946e+00
1.65454865e-01 -1.15174723e+00 -7.75648236e-01 1.17240012e+00
-2.02333236e+00 -1.67445087e+00 -8.72637033e-01 5.30954957e-01
3.19898486e-01 1.78533614e-01 8.27986956e-01 3.78200635e-02
-7.66813159e-01 7.21201241e-01 6.49523020e-01 -2.78331097e-02
9.63760018e-01 -1.51358676e+00 3.12108904e-01 1.08604586e+00
-7.90946856e-02 6.28305852e-01 4.22217876e-01 -3.98488104e-01
-1.35468173e+00 -1.11108708e+00 2.88639367e-01 -5.74122429e-01
6.38443768e-01 -3.99922132e-01 -7.10087597e-01 2.83507228e-01
1.84053242e-01 2.40227982e-01 6.79834068e-01 2.15801820e-01
-5.04786134e-01 -4.34521735e-01 -1.03960311e+00 6.09739900e-01
7.87100613e-01 -4.85390186e-01 1.05821788e-01 8.70532542e-02
3.26663256e-01 -5.11103153e-01 -6.38054669e-01 1.36337653e-01
9.07272756e-01 -1.22616088e+00 1.01995671e+00 -9.33386311e-02
5.72562993e-01 -4.67723668e-01 -1.90524593e-01 -1.31377375e+00
-6.65765345e-01 -1.61768310e-02 -3.25741976e-01 1.18195200e+00
4.18880373e-01 -7.33981371e-01 4.64776874e-01 9.42018628e-01
-1.60552502e-01 -7.90685952e-01 -8.51284802e-01 -6.70806348e-01
-2.18678731e-02 -6.02929771e-01 6.37752891e-01 7.65111148e-01
-5.36497772e-01 -4.55425620e-01 -5.50768197e-01 2.18258709e-01
8.63934278e-01 3.45083803e-01 7.26730943e-01 -1.12534940e+00
2.29226649e-02 -6.91422820e-01 -8.08639601e-02 -7.57389307e-01
-1.38174504e-01 -5.62124312e-01 4.29336242e-02 -1.95645392e+00
1.70800537e-01 -5.47448456e-01 -7.15360522e-01 5.39566696e-01
-4.05034631e-01 7.16973245e-01 6.13772832e-02 8.33163932e-02
-4.78044540e-01 8.97474051e-01 1.43988466e+00 -2.49605533e-02
5.50251789e-02 -1.05284303e-01 -7.47138977e-01 5.09965122e-01
9.99376714e-01 6.01098798e-02 -2.96421051e-01 -6.91088617e-01
6.95817769e-02 -5.83372772e-01 6.71882033e-01 -8.79105151e-01
3.90228108e-02 -2.95412779e-01 1.21436059e+00 -4.65416163e-01
5.64948082e-01 -9.32934403e-01 -1.47089427e-02 3.34827453e-01
2.44048219e-02 -1.28370956e-01 5.44045925e-01 5.07991076e-01
5.21694776e-03 1.09863644e-02 9.11591113e-01 8.02506655e-02
-1.03311765e+00 4.62971181e-01 1.01020493e-01 -4.59889024e-01
8.07633162e-01 -4.33453947e-01 -5.86279452e-01 -3.34609151e-01
-2.63237897e-02 5.30791134e-02 6.75132692e-01 2.19896004e-01
6.13468885e-01 -1.48034203e+00 -7.51630306e-01 2.97900319e-01
4.56832767e-01 -2.11189628e-01 5.31208575e-01 7.79850900e-01
-7.41794407e-01 3.06519121e-01 -4.21528250e-01 -5.47066629e-01
-8.43892455e-01 3.09120387e-01 2.78456718e-01 4.64319855e-01
-5.04329264e-01 9.97869611e-01 1.30724490e-01 -2.43355647e-01
7.41114095e-02 -5.71882010e-01 7.67229348e-02 -6.27996996e-02
5.39730251e-01 5.44719458e-01 -1.18684605e-01 -4.32331711e-01
-3.38334560e-01 7.59316921e-01 3.92633736e-01 2.07478508e-01
1.24452722e+00 -3.93688023e-01 -2.09088191e-01 4.31873679e-01
9.83175397e-01 -1.06242690e-02 -1.88847470e+00 -1.73489168e-01
-5.06294191e-01 -1.04145586e+00 5.24316013e-01 -1.25178790e+00
-1.13817561e+00 1.00461102e+00 8.69476318e-01 -1.36991560e-01
1.28902221e+00 -3.04566681e-01 2.83018231e-01 1.78285003e-01
2.89301068e-01 -1.11080110e+00 2.42611580e-02 1.63235202e-01
1.08613122e+00 -1.76605809e+00 3.99336785e-01 -2.22511381e-01
-6.75029516e-01 1.32809293e+00 9.08639073e-01 -7.00945705e-02
4.83446419e-01 8.85614008e-02 1.87628105e-01 3.15978751e-02
-3.25104237e-01 -2.79486537e-01 7.08926260e-01 7.68880427e-01
6.32528722e-01 2.23326698e-01 8.19517598e-02 2.75551707e-01
1.58330381e-01 -1.14746526e-01 1.84466556e-01 4.03240561e-01
-4.36159790e-01 -6.36142671e-01 -2.90264428e-01 3.29159796e-01
2.03001559e-01 -3.63418311e-01 -4.70070988e-02 8.84780347e-01
1.23053119e-01 8.00101876e-01 2.28708908e-01 -1.27759263e-01
-1.16197877e-01 -1.93523079e-01 5.49423456e-01 -4.70576398e-02
-2.83350777e-02 2.48141691e-01 -1.92051232e-02 -8.48395169e-01
-4.54549223e-01 -3.60979140e-01 -7.76327193e-01 -3.84669363e-01
-5.04236281e-01 -2.60817915e-01 9.93406892e-01 4.67525214e-01
3.90811384e-01 6.33060813e-01 6.21763825e-01 -1.23269796e+00
6.27302155e-02 -1.06463921e+00 -7.26297617e-01 1.19097911e-01
3.66151124e-01 -8.96329999e-01 -2.85708636e-01 -1.62514776e-01] | [10.396772384643555, -2.592088222503662] |
4d6d3fc4-c1f5-4908-a6d7-6a4e57bed9e9 | identification-of-conditional-causal-effects | null | null | http://papers.nips.cc/paper/9327-identification-of-conditional-causal-effects-under-markov-equivalence | http://papers.nips.cc/paper/9327-identification-of-conditional-causal-effects-under-markov-equivalence.pdf | Identification of Conditional Causal Effects under Markov Equivalence | Causal identification is the problem of deciding whether a post-interventional distribution is computable from a combination of qualitative knowledge about the data-generating process, which is encoded in a causal diagram, and an observational distribution. A generalization of this problem restricts the qualitative knowledge to a class of Markov equivalent causal diagrams, which, unlike a single, fully-specified causal diagram, can be inferred from the observational distribution.
Recent work by (Jaber et al., 2019a) devised a complete algorithm for the identification of unconditional causal effects given a Markov equivalence class of causal diagrams. However, there are identifiable conditional causal effects that cannot be handled by that algorithm. In this work, we derive an algorithm to identify conditional effects, which are particularly useful for evaluating conditional plans or policies. | ['Amin Jaber', 'Jiji Zhang', 'Elias Bareinboim'] | 2019-12-01 | null | null | null | neurips-2019-12 | ['causal-identification'] | ['reasoning'] | [ 4.91395324e-01 5.44445872e-01 -9.10477757e-01 -4.72438008e-01
-3.76409441e-01 -7.79941559e-01 9.14141476e-01 6.05020523e-01
1.13637961e-01 1.16685593e+00 6.59044325e-01 -1.15327263e+00
-8.42263103e-01 -9.71412003e-01 -8.34571719e-01 -3.55014235e-01
-4.84115392e-01 5.06332874e-01 1.29943177e-01 4.63761270e-01
2.98735835e-02 5.22558570e-01 -1.25456738e+00 3.41889590e-01
1.03980482e+00 2.74179459e-01 -1.23851724e-01 8.26719701e-01
1.75456092e-01 1.40616894e+00 -2.25324959e-01 1.40271559e-02
-3.11708301e-01 -9.75536525e-01 -1.08580554e+00 -3.91139865e-01
-1.03965625e-01 -6.10778332e-01 -2.53817916e-01 1.04528284e+00
-9.67334509e-02 -2.33407006e-01 1.27557063e+00 -1.60190690e+00
-5.01501143e-01 9.43906903e-01 -1.63901225e-01 1.18151912e-02
8.72805297e-01 7.18279853e-02 1.04299057e+00 -1.62769288e-01
5.74804008e-01 1.76537812e+00 2.89302468e-01 2.14366764e-01
-1.75407279e+00 -7.34491646e-01 1.05884723e-01 -2.36911396e-03
-1.04095089e+00 -2.10494459e-01 3.77022028e-01 -1.01290035e+00
6.10979021e-01 5.10489941e-01 6.98836923e-01 9.30422962e-01
6.30180240e-01 2.37059683e-01 1.57699382e+00 -7.25302637e-01
5.34328640e-01 -2.26421088e-01 3.23150963e-01 6.90379560e-01
4.48337436e-01 1.00377440e+00 -4.94982749e-01 -6.68263853e-01
8.97838891e-01 -2.26901397e-01 9.47946310e-03 -3.63893509e-01
-1.13674605e+00 9.83688653e-01 1.70938611e-01 1.17375441e-01
-3.90985489e-01 5.30719399e-01 2.10394576e-01 3.74043584e-01
6.30872399e-02 3.10775697e-01 -4.28886205e-01 7.64969587e-02
-4.20518607e-01 3.72271419e-01 9.66917574e-01 6.36474013e-01
5.44478595e-01 -5.07001638e-01 -3.19346964e-01 -2.25651726e-01
5.46686471e-01 7.95994699e-01 -3.24540019e-01 -1.01036143e+00
2.29523126e-02 7.27924049e-01 5.50588906e-01 -6.54012382e-01
-4.64823037e-01 4.41503704e-01 -7.61904240e-01 3.08862895e-01
6.65618479e-01 -3.75230402e-01 -8.64095211e-01 1.91935992e+00
3.43990296e-01 -1.40319139e-01 2.78943703e-02 4.91656631e-01
5.94143450e-01 3.28260064e-01 4.78966355e-01 -7.11616576e-01
1.09497142e+00 1.97001740e-01 -1.10468102e+00 3.03926498e-01
6.57124996e-01 -4.34753865e-01 6.97955310e-01 1.67628452e-01
-7.53738105e-01 -5.04677147e-02 -6.75572753e-01 3.35663915e-01
-1.56743899e-01 -1.49702072e-01 9.98932064e-01 7.76379764e-01
-1.02287853e+00 2.73302972e-01 -7.21405387e-01 -3.31885546e-01
1.34391665e-01 4.03385222e-01 -1.70944899e-01 3.13914716e-02
-1.54608047e+00 9.99629974e-01 5.86154342e-01 -1.43233240e-01
-1.45792282e+00 -9.17866945e-01 -7.10347116e-01 7.85811543e-02
6.27731204e-01 -9.47856486e-01 1.35404658e+00 -8.03721547e-01
-1.05690873e+00 4.58077520e-01 -3.64147067e-01 -2.20614031e-01
4.92074788e-01 2.27209032e-01 -6.22259617e-01 -1.02590481e-02
1.77698731e-01 2.04674020e-01 5.49706280e-01 -1.28976011e+00
-7.40197122e-01 -4.24281269e-01 5.68859935e-01 -5.42459190e-02
3.53830755e-01 4.43909287e-01 3.55171561e-01 -1.10557579e-01
-1.77152038e-01 -7.95721650e-01 -6.11206532e-01 -2.93194383e-01
-9.30776656e-01 -3.77386779e-01 3.94624531e-01 -4.43262815e-01
1.47073400e+00 -1.84771061e+00 -8.70328248e-02 5.49247563e-01
2.89055318e-01 -6.87219024e-01 3.91457766e-01 6.02968574e-01
-6.66322529e-01 5.34679949e-01 -5.27533114e-01 6.83102131e-01
1.06750861e-01 2.21991211e-01 -5.06501615e-01 7.19284177e-01
2.69267499e-01 7.93619335e-01 -1.13940728e+00 -6.78380251e-01
3.46876860e-01 -1.91499710e-01 -5.56836367e-01 2.68228829e-01
-4.41870004e-01 5.67910969e-01 -7.59172559e-01 1.73785910e-01
4.17786419e-01 -2.36245450e-02 8.74090552e-01 3.42261046e-01
-5.04810929e-01 5.04470527e-01 -1.26140475e+00 9.66697276e-01
-1.53963834e-01 3.37368727e-01 -3.43892097e-01 -9.30255115e-01
5.35456181e-01 6.97837412e-01 6.34254694e-01 -3.42039108e-01
1.70560941e-01 -1.63656592e-01 1.34541333e-01 -6.37451172e-01
-1.00295983e-01 -7.47602105e-01 -4.58250672e-01 8.02370727e-01
-1.54270222e-02 -1.16400808e-01 1.63504869e-01 3.74173969e-01
1.37223601e+00 -1.83015347e-01 8.60136151e-01 -7.58436382e-01
-8.72841477e-03 2.78287947e-01 6.48587167e-01 1.14678025e+00
-2.47940216e-02 4.53892816e-03 1.30219710e+00 -2.09053606e-01
-7.68704593e-01 -1.53704762e+00 -3.61791611e-01 5.76342702e-01
-1.56207889e-01 -4.87175524e-01 -4.86413419e-01 -7.51473665e-01
1.15426429e-01 9.52806115e-01 -9.80306208e-01 -1.22607358e-01
-1.24452502e-01 -7.65340805e-01 3.71826172e-01 4.36091185e-01
-1.35911614e-01 -8.42837334e-01 -5.80608010e-01 1.40821442e-01
-2.19575480e-01 -4.97389406e-01 -5.28924679e-03 3.02606285e-01
-9.92768526e-01 -1.74684250e+00 1.63587153e-01 -2.03228191e-01
7.70917237e-01 -4.62293655e-01 9.86554921e-01 -2.98969626e-01
1.07674459e-02 4.96644050e-01 -1.26518741e-01 -7.34570622e-01
-9.79709089e-01 -5.87445140e-01 1.43100709e-01 -3.44739586e-01
1.91302583e-01 -4.07921284e-01 -2.19692200e-01 7.84468651e-02
-8.60322833e-01 1.92171410e-02 3.75618190e-01 6.28562570e-01
3.61670315e-01 6.14635527e-01 7.26981699e-01 -1.12399721e+00
6.64104164e-01 -7.04478204e-01 -7.99021721e-01 5.53020656e-01
-6.94522023e-01 3.70925069e-01 4.14111286e-01 -1.22087099e-01
-1.61736465e+00 2.57141441e-01 3.72654974e-01 2.59727776e-01
-3.93257141e-01 1.02137041e+00 -4.37770784e-01 5.10119081e-01
8.07041466e-01 -5.21565378e-01 -3.02999139e-01 -3.38598251e-01
6.93869174e-01 4.29852188e-01 3.47178906e-01 -8.09486210e-01
5.50930858e-01 4.86998737e-01 4.26636249e-01 -1.24431327e-01
-6.03161752e-01 -3.08529854e-01 -9.25618052e-01 -3.35229456e-01
9.54614282e-01 -5.98907888e-01 -1.12144947e+00 -2.12307349e-01
-1.23360717e+00 -5.25856376e-01 -4.73600119e-01 8.91234159e-01
-8.06857586e-01 -1.80062011e-01 -1.08885892e-01 -1.26051474e+00
6.16615534e-01 -8.60335588e-01 4.96357322e-01 -2.11681247e-01
-7.63882160e-01 -1.26494122e+00 5.68310797e-01 -2.11963475e-01
-3.79657745e-01 5.40037751e-01 1.51269925e+00 -1.69536278e-01
-4.95489687e-01 -5.71177341e-02 -2.22351179e-01 -3.77586961e-01
3.69602174e-01 5.65497398e-01 -6.78084433e-01 6.99510872e-02
-1.33143142e-01 -1.17288038e-01 5.28038383e-01 9.75589454e-01
9.78696167e-01 -6.63817763e-01 -6.77094281e-01 -2.30857655e-01
1.34263134e+00 8.77926052e-01 4.83649582e-01 -2.99307376e-01
3.67232382e-01 8.99783731e-01 5.07264197e-01 3.51886719e-01
3.10258031e-01 2.88955450e-01 2.75706142e-01 -1.27740279e-01
5.60884923e-02 -7.33516455e-01 3.23275089e-01 1.81451708e-01
-2.54935324e-01 -5.17669059e-02 -1.06898475e+00 5.63723683e-01
-2.02377415e+00 -1.15741897e+00 -7.28708863e-01 2.24016833e+00
1.21684062e+00 3.06509268e-02 3.68159771e-01 2.16478966e-02
8.26914728e-01 -3.91082555e-01 -4.43985522e-01 -6.07669532e-01
3.89173031e-02 2.05029666e-01 6.61226511e-01 7.44815469e-01
-8.71050835e-01 3.84651989e-01 8.62251472e+00 5.11947811e-01
-2.25618035e-01 9.15055424e-02 5.47003686e-01 3.72659773e-01
-1.02645719e+00 7.72311211e-01 -2.62504011e-01 3.13022196e-01
1.34117353e+00 -5.66808105e-01 7.59301856e-02 2.34592661e-01
8.34125757e-01 -7.03680933e-01 -1.57104349e+00 1.31057560e-01
-7.45142877e-01 -1.28015149e+00 4.86760847e-02 2.17441022e-01
9.29533660e-01 -6.60286784e-01 -2.89615422e-01 -1.75795794e-01
1.60998654e+00 -1.32116091e+00 7.60122120e-01 7.20882893e-01
1.07308149e+00 -6.82400703e-01 3.84973675e-01 3.19371611e-01
-9.46067393e-01 -3.34923804e-01 2.38473006e-02 -5.69697797e-01
1.98075876e-01 1.13626671e+00 -7.95333862e-01 8.34638417e-01
6.69059038e-01 5.74999809e-01 -7.23639280e-02 8.80895376e-01
-8.11953962e-01 1.11737525e+00 -2.95848604e-02 1.17343977e-01
-1.61606595e-01 -9.84022319e-02 3.58231872e-01 1.02115345e+00
3.92951854e-02 5.50512254e-01 -5.44593073e-02 1.13981116e+00
2.56443739e-01 -3.47712159e-01 -9.64573801e-01 -2.53928393e-01
4.49503779e-01 4.05906588e-01 -5.86804688e-01 -3.55959356e-01
-3.58974129e-01 4.52313349e-02 -8.33386928e-02 5.86726427e-01
-6.54515266e-01 2.09701911e-01 3.91002178e-01 6.69889003e-02
-4.43766385e-01 1.43513501e-01 -6.00842893e-01 -7.61705458e-01
-5.02391994e-01 -7.35249102e-01 1.01568210e+00 -7.86654651e-01
-1.18346369e+00 -4.18004036e-01 7.33105600e-01 -7.82049775e-01
-4.20744628e-01 -2.52230883e-01 -5.08338213e-01 1.02391779e+00
-7.14452505e-01 -7.53028631e-01 3.54293913e-01 6.85871899e-01
-2.67351896e-01 3.85643005e-01 1.00550473e+00 -2.15660959e-01
-4.19620454e-01 -1.05337970e-01 -1.78782437e-02 -1.05126828e-01
4.92752224e-01 -1.66755843e+00 -3.55573803e-01 8.88793766e-01
-4.88543600e-01 5.66520393e-01 8.78071070e-01 -1.22413969e+00
-1.31678236e+00 -9.62978065e-01 1.19869244e+00 -6.15974545e-01
1.13005078e+00 -1.99386284e-01 -3.70102972e-01 1.09674037e+00
2.38962412e-01 -6.54919624e-01 6.23150468e-01 7.06832290e-01
-3.46209049e-01 1.14172138e-01 -1.04380071e+00 6.79218411e-01
1.09900606e+00 -7.10245907e-01 -8.55884373e-01 1.25818655e-01
7.54580498e-01 4.91290167e-02 -1.00153303e+00 4.70484197e-01
6.35071695e-01 -6.84530616e-01 6.77369833e-01 -1.13439167e+00
6.65297687e-01 -4.35115486e-01 9.67700183e-02 -1.36932981e+00
-4.60133523e-01 -5.34150660e-01 1.43830210e-01 9.64586735e-01
5.46492517e-01 -6.05666220e-01 2.16778904e-01 8.88072073e-01
2.23306522e-01 2.35729348e-02 -8.40687215e-01 -7.26399958e-01
1.86617523e-01 -5.07726550e-01 6.32827461e-01 1.35543478e+00
5.84583104e-01 3.10067534e-01 -2.25506246e-01 2.27537543e-01
6.73961401e-01 4.59667414e-01 2.80238986e-01 -1.45422769e+00
-1.76158935e-01 -2.65897125e-01 -2.55610704e-01 -3.09813023e-01
2.27264583e-01 -8.42034221e-01 1.90086976e-01 -1.79949915e+00
8.77198040e-01 -6.37447655e-01 -1.21301197e-01 7.42532372e-01
-1.34825677e-01 -7.20424950e-01 -2.05576196e-01 -4.78542894e-02
-5.63175455e-02 2.52241373e-01 1.30037785e+00 -1.61431372e-01
-2.49362677e-01 7.29321241e-02 -7.58025825e-01 6.72797740e-01
7.75611341e-01 -9.49423432e-01 -7.30763793e-01 2.63046831e-01
5.37934899e-01 9.82129991e-01 8.26217532e-01 -1.54748425e-01
1.28718570e-01 -1.03712392e+00 8.36263895e-02 -6.17492795e-01
-6.02882504e-01 -7.56105304e-01 9.48606849e-01 8.91643524e-01
-7.67203033e-01 -6.52344748e-02 1.17369205e-01 7.10036755e-01
-1.71470106e-01 -2.94592321e-01 2.80926526e-01 8.21902156e-02
-5.07965267e-01 -5.25506847e-02 -8.68362904e-01 -8.15481842e-02
1.02337742e+00 1.50528818e-01 -4.53154594e-01 -3.12247872e-01
-9.89144385e-01 1.43724337e-01 1.54843867e-01 -1.16502987e-02
3.97803336e-01 -1.34740019e+00 -6.90807104e-01 -3.43260646e-01
-1.39186447e-02 -1.97998017e-01 2.42880940e-01 1.00797057e+00
-9.92382020e-02 6.98440075e-01 -1.63699940e-01 -2.04444706e-01
-9.07257855e-01 8.90216351e-01 3.40935439e-01 -3.11183721e-01
-2.78119057e-01 1.96757287e-01 7.51512289e-01 -3.41186583e-01
-1.24863215e-01 -5.51578104e-01 -3.69035974e-02 -2.68366113e-02
3.81136149e-01 4.76589710e-01 -4.46041733e-01 -1.09725595e-01
-5.46204031e-01 -1.06131256e-01 6.92446291e-01 -5.73241055e-01
9.48431015e-01 -8.90335217e-02 -5.71464658e-01 8.06344211e-01
7.51951277e-01 -3.87401357e-02 -1.04106247e+00 1.78664789e-01
1.32122532e-01 -3.21924001e-01 -4.71309535e-02 -1.25578713e+00
-3.91603470e-01 4.13724452e-01 2.54777521e-01 5.38143277e-01
1.17293775e+00 3.74141544e-01 -5.33200443e-01 -7.56680742e-02
2.73448884e-01 -6.42441988e-01 -4.88517612e-01 6.19137846e-02
9.88753021e-01 -1.01501644e+00 6.62965104e-02 -6.70716524e-01
-1.88853413e-01 9.06384468e-01 1.00811839e-01 2.34367147e-01
9.90484893e-01 6.11080945e-01 -5.27698755e-01 -5.12301862e-01
-9.44421172e-01 -2.15829864e-01 2.63213962e-01 7.23991036e-01
3.94748420e-01 9.84697104e-01 -6.18802249e-01 4.15382624e-01
-7.91810751e-02 5.12395620e-01 7.58395910e-01 8.48730862e-01
-2.33331621e-01 -8.87540936e-01 -6.17168486e-01 7.76875436e-01
-1.31676629e-01 -6.49805740e-02 -7.66243517e-01 9.20072496e-01
1.03359520e-01 1.43923438e+00 2.13058159e-01 -1.20743118e-01
4.22512233e-01 -1.54829621e-01 5.73883176e-01 -5.16333997e-01
-4.25900817e-02 -2.19606906e-01 5.80563128e-01 -6.26663506e-01
-9.36034739e-01 -9.61267114e-01 -1.23758388e+00 -6.06897652e-01
-1.67750522e-01 2.66563267e-01 9.37964171e-02 1.16988945e+00
-3.25756431e-01 7.26762414e-01 5.94208241e-01 2.49214932e-01
-2.29501262e-01 -6.04736269e-01 -8.55486453e-01 2.11989041e-02
3.73896748e-01 -8.48519623e-01 -2.83693224e-01 4.71473932e-01] | [8.061615943908691, 5.559758186340332] |
4e78ba88-15f5-48af-82ac-1b79ffbf8bcd | audio-captioning-with-composition-of-acoustic | 2105.06355 | null | https://arxiv.org/abs/2105.06355v1 | https://arxiv.org/pdf/2105.06355v1.pdf | Audio Captioning with Composition of Acoustic and Semantic Information | Generating audio captions is a new research area that combines audio and natural language processing to create meaningful textual descriptions for audio clips. To address this problem, previous studies mostly use the encoder-decoder based models without considering semantic information. To fill this gap, we present a novel encoder-decoder architecture using bi-directional Gated Recurrent Units (BiGRU) with audio and semantic embeddings. We extract semantic embedding by obtaining subjects and verbs from the audio clip captions and combine these embedding with audio embedding to feed the BiGRU-based encoder-decoder model. To enable semantic embeddings for the test audios, we introduce a Multilayer Perceptron classifier to predict the semantic embeddings of those clips. We also present exhaustive experiments to show the efficiency of different features and datasets for our proposed model the audio captioning task. To extract audio features, we use the log Mel energy features, VGGish embeddings, and a pretrained audio neural network (PANN) embeddings. Extensive experiments on two audio captioning datasets Clotho and AudioCaps show that our proposed model outperforms state-of-the-art audio captioning models across different evaluation metrics and using the semantic information improves the captioning performance. Keywords: Audio captioning; PANNs; VGGish; GRU; BiGRU. | ['Mustafa Sert', 'Ayşegül Özkaya Eren'] | 2021-05-13 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.09102112e-01 2.52992123e-01 1.27667580e-02 -4.07948971e-01
-1.39025557e+00 -2.61631548e-01 2.22988829e-01 5.85669391e-02
-1.73645634e-02 6.35577738e-01 1.13884616e+00 3.50377917e-01
2.78389513e-01 -4.78988707e-01 -9.96902585e-01 -2.44398922e-01
-1.27846450e-01 1.57217041e-01 -1.62098687e-02 -2.14169040e-01
-3.69757861e-02 -3.40318888e-01 -1.78579903e+00 9.32084203e-01
2.74381220e-01 1.38731587e+00 2.13074476e-01 1.06460285e+00
-8.43621641e-02 8.92154515e-01 -6.90673888e-01 -6.74418285e-02
-2.10461304e-01 -6.79652750e-01 -8.39006960e-01 -2.53869683e-01
8.59884545e-02 -1.57025352e-01 -4.97604728e-01 6.94103479e-01
9.95289803e-01 3.33296150e-01 5.58673859e-01 -1.38191450e+00
-1.14409828e+00 1.26740086e+00 -9.93040502e-02 2.02679828e-01
9.59264576e-01 -2.60987043e-01 1.53690660e+00 -9.39905405e-01
3.01765680e-01 1.23917866e+00 6.93459451e-01 7.53775656e-01
-8.51118743e-01 -6.58367395e-01 -1.88074987e-02 6.95868194e-01
-1.48668969e+00 -3.50522965e-01 1.12339962e+00 -3.59148055e-01
9.26463127e-01 3.32842052e-01 5.82298875e-01 1.56152725e+00
-6.90706298e-02 9.13095236e-01 4.26931679e-01 -3.52527082e-01
2.44768068e-01 8.19242746e-02 1.32530071e-02 -2.03215014e-02
-5.90114474e-01 -4.71281223e-02 -8.30775738e-01 -1.64436013e-01
4.67070252e-01 -2.77077198e-01 -2.74262398e-01 3.24757695e-01
-1.24009454e+00 9.24346626e-01 4.97549802e-01 2.31553838e-01
-5.31477332e-01 6.00472927e-01 8.45328629e-01 2.22488314e-01
2.60980904e-01 5.07978261e-01 -2.39432469e-01 -4.72795337e-01
-7.75422513e-01 2.01412261e-01 5.66149771e-01 9.59596753e-01
3.33765954e-01 4.36120868e-01 -4.36787784e-01 1.18781400e+00
2.98410743e-01 4.60793763e-01 1.05588400e+00 -5.89949846e-01
6.31619871e-01 1.77179754e-01 -1.00396059e-01 -7.95878291e-01
-1.53959557e-01 -2.81457871e-01 -5.50620139e-01 -7.09107041e-01
-3.85428727e-01 -4.41353083e-01 -8.54660511e-01 1.75373840e+00
-1.33526936e-01 6.63831413e-01 4.00676817e-01 1.28602350e+00
1.33838749e+00 1.72303975e+00 3.07516664e-01 9.38762426e-02
1.52160263e+00 -1.19347358e+00 -1.04292345e+00 -2.31966138e-01
3.58618468e-01 -7.71355510e-01 1.22567821e+00 1.11304373e-01
-9.76947606e-01 -9.10265982e-01 -1.14643645e+00 -1.21587560e-01
-2.85182327e-01 2.00949505e-01 2.04640776e-01 2.19030559e-01
-8.72994900e-01 3.45450550e-01 -4.52625245e-01 -1.67318791e-01
-4.06219028e-02 2.22216070e-01 -1.09445900e-01 4.12099361e-01
-1.92527866e+00 2.65184343e-01 7.93802261e-01 -4.73139286e-02
-1.31307709e+00 -8.42826962e-01 -1.12152004e+00 4.96698856e-01
-8.13689232e-02 -4.76051033e-01 1.59979975e+00 -1.10045385e+00
-1.45089674e+00 2.37840772e-01 4.35914472e-02 -8.68862331e-01
-2.86779344e-01 -3.83955270e-01 -7.37407446e-01 5.35017669e-01
4.97206412e-02 1.33187628e+00 8.78399670e-01 -9.61401999e-01
-4.95990902e-01 2.38224611e-01 -2.24739462e-01 4.20647025e-01
-6.93766534e-01 2.32050881e-01 7.21466467e-02 -9.87874031e-01
-3.32392365e-01 -7.85337865e-01 1.11199573e-01 -5.42695999e-01
-4.74903136e-01 -2.15589315e-01 7.78369367e-01 -1.01442039e+00
1.49900794e+00 -2.55300689e+00 1.27589121e-01 -2.09282860e-01
-2.40622833e-01 9.85193998e-03 -4.48732883e-01 6.33719862e-01
-3.33212286e-01 -7.44608268e-02 -1.06543317e-01 -1.40281379e-01
3.52874517e-01 4.84104548e-03 -9.42129731e-01 -2.71318763e-01
3.64399850e-01 9.08805430e-01 -9.03327286e-01 -4.97645140e-01
4.65653911e-02 1.02085364e+00 -7.13090241e-01 5.95654905e-01
-4.39472228e-01 1.68509185e-01 -3.12311560e-01 3.83327544e-01
-3.23980786e-02 1.69959795e-02 -3.80998164e-01 -2.09856197e-01
2.99022347e-01 6.94677055e-01 -1.11508226e+00 1.84604347e+00
-7.25808859e-01 8.08170199e-01 -2.99719959e-01 -8.03157866e-01
1.23627102e+00 1.14860225e+00 4.33512121e-01 -4.40368474e-01
2.99870789e-01 3.13089758e-01 -4.34259802e-01 -6.85380340e-01
5.37815928e-01 -3.44820023e-01 -5.50744116e-01 3.44608068e-01
6.33418679e-01 5.77890463e-02 -2.05585197e-01 -4.53656539e-02
1.03471875e+00 3.47420550e-03 1.24931624e-02 4.50310379e-01
5.67167819e-01 -2.79698670e-01 1.86924830e-01 3.37744415e-01
-4.41087782e-02 1.31501210e+00 2.97475159e-01 -2.69383371e-01
-1.09332621e+00 -1.09016335e+00 1.95324644e-01 1.66248059e+00
-1.79805726e-01 -7.39400148e-01 -8.85940909e-01 -1.98938787e-01
-3.80825341e-01 7.35981882e-01 -6.48957074e-01 -3.73233259e-01
-3.18708301e-01 -3.96902084e-01 8.96115780e-01 7.81854391e-01
7.74467960e-02 -1.64650226e+00 -1.85183600e-01 6.43747091e-01
-7.74587154e-01 -1.24572802e+00 -7.34421372e-01 8.28972310e-02
-4.36898410e-01 -3.52931589e-01 -1.01451194e+00 -1.29796553e+00
6.72989339e-02 -2.45375827e-01 8.23382497e-01 -6.90389514e-01
-8.43104050e-02 4.46486801e-01 -9.49376345e-01 -6.56162381e-01
-4.59633261e-01 3.56131345e-01 4.64637950e-02 4.13557738e-01
4.70130444e-01 -8.91035378e-01 -3.66201729e-01 2.89201457e-02
-9.33257759e-01 1.98589459e-01 5.59054792e-01 6.95702136e-01
6.97276890e-01 -4.78565842e-01 1.22525811e+00 -1.37707517e-01
1.04231620e+00 -8.17607284e-01 6.62146360e-02 -1.70092896e-01
1.28771186e-01 4.50787805e-02 9.41766322e-01 -7.76695192e-01
-7.22428441e-01 1.80435419e-01 -4.86963570e-01 -7.31750011e-01
-1.37221336e-01 6.40289605e-01 -1.21468455e-01 6.17351413e-01
6.48868620e-01 3.92973244e-01 -3.38456243e-01 -6.77008688e-01
4.56237733e-01 1.40248597e+00 7.98326313e-01 -3.49662751e-01
4.99302626e-01 -3.85460779e-02 -6.68279707e-01 -9.02118683e-01
-1.02348411e+00 -4.15486813e-01 -1.30132124e-01 -3.53485107e-01
1.11043262e+00 -1.27701306e+00 -4.37406093e-01 -2.93648690e-01
-1.53237426e+00 2.16626018e-01 -6.42981589e-01 8.66088986e-01
-9.43930984e-01 -9.76430699e-02 -7.26937354e-01 -8.17077994e-01
-7.54616201e-01 -1.04077625e+00 1.42929864e+00 1.19177774e-01
-5.07240772e-01 -6.38823092e-01 3.94036859e-01 3.99361074e-01
3.73005867e-01 3.37275147e-01 6.26823545e-01 -1.08629322e+00
-1.91920638e-01 -2.46918544e-01 -3.52112129e-02 6.05158746e-01
-1.21997900e-01 -5.24517596e-01 -1.44737375e+00 1.32692993e-01
-2.56452709e-01 -6.31768942e-01 7.26006806e-01 3.65543306e-01
1.36971831e+00 -6.76834941e-01 1.33016422e-01 2.68500447e-01
1.21988153e+00 2.82323301e-01 6.05726600e-01 1.09581903e-01
6.65971696e-01 3.25597972e-01 5.10885298e-01 7.52445161e-01
2.41732642e-01 6.90924346e-01 2.73542166e-01 1.40769586e-01
-3.20900261e-01 -9.57747340e-01 1.01066387e+00 1.60234618e+00
2.89774328e-01 -2.05048829e-01 -6.80015385e-01 9.87537324e-01
-1.82812190e+00 -9.79754388e-01 2.55074561e-01 1.63939548e+00
1.01465607e+00 2.34689862e-02 1.53716087e-01 5.39278448e-01
9.74488318e-01 1.56320706e-01 -1.17154598e-01 -8.43388379e-01
1.64200321e-01 3.12931716e-01 -8.87953341e-02 3.31777275e-01
-1.02905893e+00 6.76677346e-01 5.70720005e+00 9.26121354e-01
-1.07775736e+00 3.71867716e-01 2.31254146e-01 -2.64068425e-01
-3.55395019e-01 -3.08963686e-01 -6.68397009e-01 6.34959459e-01
1.79205322e+00 -3.80457938e-01 4.60725904e-01 9.64668453e-01
3.65612596e-01 7.91485846e-01 -1.36837757e+00 1.35174108e+00
3.76948297e-01 -1.23581350e+00 5.65026462e-01 -5.84395409e-01
4.98051405e-01 -4.35602739e-02 6.99632093e-02 5.56085348e-01
-2.10234597e-01 -1.14065850e+00 8.97999167e-01 2.97662526e-01
8.64734054e-01 -7.48620987e-01 9.71620739e-01 -2.83627272e-01
-1.47391427e+00 -1.92057624e-01 -3.66494447e-01 -1.75757244e-01
5.82568288e-01 1.72513634e-01 -1.30053961e+00 4.22691196e-01
6.50681436e-01 7.21348047e-01 -3.97038996e-01 1.14007401e+00
-1.69104427e-01 1.10319483e+00 -2.00273827e-01 -2.75061697e-01
5.25364399e-01 3.88871282e-01 6.66562498e-01 1.52978373e+00
5.73585510e-01 -2.36846820e-01 -1.85555238e-02 7.75790155e-01
-3.47448915e-01 4.52528924e-01 -4.06511039e-01 -4.77151453e-01
5.72454035e-01 8.51512134e-01 -1.65496096e-01 -2.39679649e-01
-2.93458998e-01 8.64844382e-01 -2.71000952e-01 2.06514001e-01
-1.27341008e+00 -1.04277432e+00 5.40135026e-01 1.76221684e-01
2.87123173e-01 2.39100203e-01 1.38262808e-01 -9.72017109e-01
8.05932730e-02 -6.13017082e-01 4.63110626e-01 -1.39770412e+00
-1.23462844e+00 9.61762547e-01 9.86372977e-02 -1.49376678e+00
-6.20583713e-01 -1.22913830e-01 -5.64848304e-01 4.97171134e-01
-1.38879693e+00 -1.17460430e+00 -1.55152008e-01 4.86222357e-01
9.59129274e-01 -3.40580761e-01 9.24889743e-01 6.25934958e-01
-5.06969988e-02 5.04371345e-01 -1.47204593e-01 3.92652154e-01
7.46185482e-01 -9.95064795e-01 4.06686485e-01 2.86658436e-01
5.24439991e-01 2.81546693e-02 8.61703217e-01 -2.74282813e-01
-1.07777596e+00 -1.48727560e+00 1.07702661e+00 -2.12629840e-01
8.41015220e-01 -6.87084854e-01 -9.14508820e-01 8.64549458e-01
4.62221861e-01 -1.82685032e-01 1.14431620e+00 -2.85138756e-01
-1.98329464e-01 -2.01262847e-01 -5.76826930e-01 3.61936837e-01
7.15808034e-01 -7.93617725e-01 -1.20980978e+00 2.74334013e-01
1.67566133e+00 -1.79316983e-01 -9.26464379e-01 1.10234529e-01
4.82374638e-01 -2.04668701e-01 9.53712463e-01 -6.40001237e-01
7.20790505e-01 -3.44234198e-01 -4.48992193e-01 -1.33916581e+00
-4.13427614e-02 -9.18925643e-01 -6.60415068e-02 1.51803613e+00
7.35901833e-01 -1.07793678e-02 4.73710924e-01 -1.35782495e-01
-5.61124206e-01 -4.42016035e-01 -1.03336871e+00 -7.12564588e-01
-3.31781417e-01 -7.72442222e-01 8.09736311e-01 6.43044174e-01
4.24072236e-01 9.30903792e-01 -7.71217167e-01 1.43831387e-01
1.29336059e-01 -1.08269230e-01 3.47380340e-01 -1.02148294e+00
-2.70297498e-01 1.17022157e-01 -7.32015491e-01 -8.43174338e-01
1.99683800e-01 -1.08420682e+00 2.72870243e-01 -1.56340277e+00
-5.51042296e-02 1.15537107e-01 -5.29840052e-01 6.41568482e-01
2.16488898e-01 5.74316025e-01 3.07241887e-01 -1.85526982e-01
-8.39284956e-01 1.04231906e+00 7.77027786e-01 -3.55009258e-01
-5.28988659e-01 -2.84459412e-01 -7.25482941e-01 3.43228191e-01
8.01603854e-01 -6.98501527e-01 -4.44670171e-01 -4.64757085e-01
2.81300604e-01 1.91337407e-01 2.54542768e-01 -1.49392116e+00
5.68026863e-02 2.50439942e-01 2.53271669e-01 -4.30879593e-01
8.32378805e-01 -7.19613373e-01 1.52959168e-01 2.24437881e-02
-9.96122181e-01 -1.17636345e-01 3.46386522e-01 5.39633036e-01
-9.50378954e-01 -2.60717839e-01 4.39133853e-01 2.55453855e-01
-5.36508441e-01 8.24777186e-02 -5.50891221e-01 1.81018353e-01
7.94265687e-01 2.75620166e-03 1.82583928e-01 -8.87425065e-01
-1.14164627e+00 1.38036519e-01 -4.70464826e-01 1.01691461e+00
9.23581898e-01 -2.06121683e+00 -9.96903419e-01 1.09362610e-01
3.88811380e-01 -4.17475462e-01 2.81728119e-01 2.57899582e-01
-2.24166319e-01 6.60949230e-01 -2.98068762e-01 -4.37212348e-01
-9.88048077e-01 4.45906132e-01 9.78043769e-03 9.45873857e-02
-4.55332279e-01 6.73772454e-01 1.27559394e-01 -1.38401359e-01
6.14416778e-01 -5.29484153e-01 -5.82700074e-01 2.83937037e-01
8.22277367e-01 1.46840096e-01 -4.07255068e-02 -7.48214006e-01
-1.73268974e-01 3.38720828e-01 2.52308726e-01 -4.56539840e-01
1.42452586e+00 -6.53313771e-02 3.92690241e-01 7.89835572e-01
1.58559644e+00 -3.85027111e-01 -8.37376177e-01 -1.69994570e-02
-2.23549306e-01 9.49761495e-02 -1.48711400e-02 -5.75414717e-01
-7.64005780e-01 1.25743389e+00 5.33876956e-01 2.65690207e-01
1.14941442e+00 1.90535396e-01 1.45570123e+00 9.41518322e-02
-9.45262611e-02 -1.25158548e+00 3.21138442e-01 4.56647664e-01
1.24371266e+00 -7.90766299e-01 -7.06744492e-01 -8.90740082e-02
-9.88285005e-01 1.15855861e+00 5.18634379e-01 -2.39266321e-01
4.90494996e-01 -1.16327789e-03 -6.53339177e-02 1.89649716e-01
-9.78019178e-01 -1.57469884e-01 3.86002094e-01 5.94107449e-01
4.36189830e-01 1.24370486e-01 -6.74716476e-03 1.51750374e+00
-7.62688339e-01 1.35198623e-01 7.07783461e-01 6.16166651e-01
-6.38179064e-01 -8.59836996e-01 -5.86643636e-01 2.07373768e-01
-5.72446883e-01 -1.67514637e-01 -5.00823379e-01 1.23794802e-01
-1.31371945e-01 8.71391356e-01 2.55557805e-01 -8.61875594e-01
3.48985970e-01 4.90461230e-01 -6.44010678e-02 -8.50005150e-01
-6.61603272e-01 1.87439233e-01 8.64646882e-02 -2.50923783e-01
-2.44487911e-01 -1.27505749e-01 -1.40840328e+00 4.22334939e-01
-1.64801538e-01 8.43128145e-01 7.52836406e-01 5.76903105e-01
5.69775045e-01 9.61039841e-01 8.44663858e-01 -9.77632940e-01
-4.52309012e-01 -1.33756816e+00 -4.37129974e-01 3.86698246e-01
4.65393126e-01 -4.14151013e-01 -4.80700374e-01 5.22678435e-01] | [15.290803909301758, 4.910626411437988] |
c93c6553-4c54-431f-8f83-728749f4c115 | improved-descriptors-for-patch-matching-and | 1701.06854 | null | http://arxiv.org/abs/1701.06854v4 | http://arxiv.org/pdf/1701.06854v4.pdf | Improved Descriptors for Patch Matching and Reconstruction | We propose a convolutional neural network (ConvNet) based approach for
learning local image descriptors which can be used for significantly improved
patch matching and 3D reconstructions. A multi-resolution ConvNet is used for
learning keypoint descriptors. We also propose a new dataset consisting of an
order of magnitude more number of scenes, images, and positive and negative
correspondences compared to the currently available Multi-View Stereo (MVS)
[18] dataset. The new dataset also has better coverage of the overall
viewpoint, scale, and lighting changes in comparison to the MVS dataset. We
evaluate our approach on publicly available datasets, such as Oxford Affine
Covariant Regions Dataset (ACRD) [12], MVS [18], Synthetic [6] and Strecha [15]
datasets to quantify the image descriptor performance. Scenes from the Oxford
ACRD, MVS and Synthetic datasets are used for evaluating the patch matching
performance of the learnt descriptors while the Strecha dataset is used to
evaluate the 3D reconstruction task. Experiments show that the proposed
descriptor outperforms the current state-of-the-art descriptors in both the
evaluation tasks. | ['Sharat Chandran', 'Rahul Mitra', 'Arjun Jain', 'Shuaib Ahmed', 'Sanath Narayan', 'Jiakai Zhang'] | 2017-01-24 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [-8.97324234e-02 -6.59134209e-01 -2.52125598e-02 -4.99077857e-01
-9.61827636e-01 -5.44473946e-01 9.52063799e-01 9.62970331e-02
-4.82379347e-01 3.90084386e-01 3.25625122e-01 4.39295769e-01
-2.58734912e-01 -8.23141456e-01 -6.81562483e-01 -6.13672197e-01
-4.25338484e-02 2.97727466e-01 4.58093137e-01 -4.63985741e-01
5.49761772e-01 1.04500568e+00 -1.76383626e+00 2.46048883e-01
3.05046421e-02 1.48406088e+00 2.72316128e-01 4.21840310e-01
4.15461421e-01 5.38035929e-01 -9.57702324e-02 -8.86384398e-02
7.21642196e-01 1.95304766e-01 -5.93312979e-01 -1.96884558e-01
1.35687459e+00 -3.99727196e-01 -6.08059764e-01 8.73660505e-01
8.18463027e-01 2.26229072e-01 6.71965659e-01 -9.23049867e-01
-6.70322239e-01 -1.77156657e-01 -4.81407613e-01 2.37624884e-01
6.26947820e-01 -1.29425079e-01 1.19797051e+00 -1.24511099e+00
1.13677430e+00 1.33238626e+00 8.92360568e-01 2.72809584e-02
-1.02663374e+00 -4.92126405e-01 -3.70381713e-01 3.83590281e-01
-1.52933729e+00 -2.88607657e-01 8.32686722e-01 -4.09009069e-01
1.23852539e+00 1.51120827e-01 6.42747641e-01 1.15035987e+00
2.69824922e-01 5.92212737e-01 1.10437429e+00 -4.16298628e-01
8.09069648e-02 -1.44107699e-01 -1.22391693e-01 7.03897715e-01
-4.25671823e-02 4.60690141e-01 -4.59937990e-01 -2.44839460e-01
1.14186120e+00 1.53714627e-01 -1.52111650e-01 -9.59967136e-01
-1.52469587e+00 9.64270592e-01 6.79459631e-01 3.79003286e-01
-4.87266541e-01 1.27481014e-01 7.32837915e-01 2.52375692e-01
5.28362155e-01 3.43305230e-01 -4.84979987e-01 -1.08390927e-01
-7.88060188e-01 6.16919756e-01 4.57014769e-01 9.51180160e-01
1.00939727e+00 1.69081066e-03 -5.98941371e-03 1.22075427e+00
1.33161411e-01 5.51244557e-01 2.34249905e-01 -1.26958168e+00
3.32772344e-01 5.89229047e-01 1.44799709e-01 -1.41638124e+00
-5.03742397e-01 -3.16739678e-01 -9.25587177e-01 3.58049542e-01
1.60062656e-01 6.90417707e-01 -8.32039237e-01 1.27864289e+00
1.24130920e-01 -1.12093829e-01 -1.11895472e-01 9.26528275e-01
1.17758417e+00 2.90042251e-01 -4.80570614e-01 3.78401190e-01
1.06305361e+00 -8.79923880e-01 -1.72995210e-01 1.71180502e-01
2.93028951e-01 -1.06521642e+00 7.74832070e-01 2.61255622e-01
-1.00008976e+00 -9.36680794e-01 -1.17805099e+00 -3.41325819e-01
-4.87272412e-01 2.14347199e-01 6.03922784e-01 3.24152075e-02
-1.32100260e+00 7.38668919e-01 -4.90235656e-01 -6.82616830e-01
4.49071050e-01 2.53233492e-01 -9.10823107e-01 -4.96944010e-01
-1.02947831e+00 9.47147667e-01 2.23857597e-01 -2.72836894e-01
-1.09694362e+00 -8.68882596e-01 -1.08765280e+00 -1.35887131e-01
-3.35277379e-01 -6.07126117e-01 5.34264386e-01 -6.53489172e-01
-9.95508790e-01 1.51396108e+00 3.37967217e-01 -1.98217124e-01
5.54401636e-01 9.41857249e-02 -2.37170279e-01 4.41535950e-01
2.79334486e-01 9.57073569e-01 5.65719306e-01 -1.14783144e+00
-5.35142541e-01 -5.05217075e-01 3.41099620e-01 2.09632710e-01
1.99587718e-01 -7.60371760e-02 -2.94031709e-01 -6.71704531e-01
3.40014011e-01 -8.22394729e-01 6.04160391e-02 3.62865180e-01
-1.36446998e-01 -1.10951640e-01 8.49762619e-01 -5.51806211e-01
5.06007731e-01 -2.36432028e+00 3.20440322e-01 6.59742877e-02
8.52621049e-02 1.24554612e-01 -4.36477602e-01 7.24503040e-01
-2.89232403e-01 -3.97021294e-01 1.20160215e-01 -3.64277482e-01
-1.05146579e-01 3.25456381e-01 -3.34937014e-02 9.26050186e-01
-8.09179544e-02 6.58895016e-01 -5.65593481e-01 -2.81330109e-01
8.23328197e-01 7.06576467e-01 -4.48615462e-01 1.81954220e-01
1.91155255e-01 5.02819717e-01 -1.42546594e-01 7.90313065e-01
8.81719112e-01 1.09122261e-01 -6.10134482e-01 -6.82000577e-01
-2.76381016e-01 -1.51310042e-01 -1.27542615e+00 2.12815738e+00
-5.66615582e-01 7.06283391e-01 -2.17373163e-01 -9.10498440e-01
1.28269911e+00 3.89201939e-02 6.58945560e-01 -1.08129811e+00
-6.37821481e-02 3.78085464e-01 -5.05010903e-01 -2.12589860e-01
6.11464858e-01 9.83074903e-02 1.69636533e-01 -1.07074924e-01
4.33614403e-01 -1.30068392e-01 4.02142517e-02 1.32361904e-01
1.05173481e+00 1.60473570e-01 4.08395916e-01 -5.54157913e-01
8.46429229e-01 -2.06916332e-01 3.03450465e-01 5.71421683e-01
-2.49817237e-01 9.89298642e-01 9.86937657e-02 -1.13058078e+00
-1.44824576e+00 -1.16549516e+00 -5.26871741e-01 7.04494655e-01
2.89773911e-01 -3.06720704e-01 -2.25109190e-01 -2.88790196e-01
2.20340535e-01 1.64144590e-01 -8.07860374e-01 2.32938975e-01
-5.43310165e-01 -8.93721506e-02 5.88951170e-01 5.80214798e-01
8.95957053e-01 -9.66469944e-01 -6.14853799e-01 -1.55705869e-01
9.59547460e-02 -1.23564720e+00 -3.99137974e-01 -1.33393660e-01
-6.09929025e-01 -1.27134895e+00 -7.75043070e-01 -8.64989638e-01
3.72042209e-01 3.74645501e-01 1.31048882e+00 -2.18283609e-01
-5.74085832e-01 7.77371526e-01 -3.81253809e-01 7.05749914e-02
-1.40679225e-01 -1.76537424e-01 -6.01521544e-02 -1.41818568e-01
2.30888739e-01 -8.36817086e-01 -9.52636182e-01 4.11077827e-01
-9.40230310e-01 -1.02832183e-01 4.39601511e-01 1.15042746e+00
9.55901980e-01 -4.93263960e-01 9.21956450e-02 -4.48197097e-01
1.63153827e-01 -1.34748101e-01 -7.11204588e-01 3.35383957e-04
-4.36831802e-01 -1.07357569e-01 4.42580968e-01 -1.48836836e-01
-6.19077682e-01 1.16854630e-01 -3.20512742e-01 -7.41016567e-01
-4.52149421e-01 1.55547723e-01 1.66642293e-01 -7.63597548e-01
6.87349975e-01 3.69151115e-01 -1.91237316e-01 -5.78319609e-01
2.44899705e-01 5.16958833e-01 5.85645318e-01 -3.51862043e-01
6.53369188e-01 7.56379128e-01 3.68350416e-01 -8.21674407e-01
-6.14453554e-01 -9.39201772e-01 -1.10168135e+00 -1.18190870e-01
8.64356279e-01 -1.16560400e+00 -4.70935911e-01 5.36563694e-01
-1.22443533e+00 1.67763203e-01 -2.46710673e-01 5.23924232e-01
-1.06462979e+00 3.08933616e-01 -4.45888460e-01 -1.84171021e-01
-2.84033656e-01 -1.42281151e+00 1.79200816e+00 -6.84321001e-02
1.48707137e-01 -1.27432215e+00 3.77065361e-01 2.47041151e-01
6.34077549e-01 8.75323832e-01 8.20174098e-01 -4.01345074e-01
-6.68807328e-01 -3.53176415e-01 -4.93904561e-01 5.07182002e-01
-6.81944415e-02 -1.81076273e-01 -9.40580487e-01 -6.28230631e-01
-3.03887099e-01 -4.95355248e-01 9.33960855e-01 3.08191210e-01
1.00006890e+00 1.86746642e-01 5.29839247e-02 1.11199963e+00
1.90406978e+00 2.76404489e-02 8.43759716e-01 7.07698286e-01
6.97391033e-01 3.46363515e-01 5.85216880e-01 4.73461717e-01
3.86097640e-01 1.21339309e+00 6.43538833e-01 -3.07062536e-01
-3.39750409e-01 -8.96782279e-02 1.27773017e-01 7.55460262e-01
-2.54172087e-01 2.68032670e-01 -8.00931334e-01 5.35290480e-01
-1.66329372e+00 -9.47949648e-01 5.38546266e-03 2.10322952e+00
2.95796990e-01 -3.75293791e-01 -1.03881523e-01 1.06794775e-01
3.79552394e-01 4.93824929e-01 -3.34529817e-01 -3.37273151e-01
-3.14135402e-01 4.42532927e-01 6.80335402e-01 2.86188781e-01
-1.38452303e+00 7.01957762e-01 6.13977671e+00 9.16304588e-01
-1.26561725e+00 5.43771647e-02 3.77786875e-01 2.98833698e-01
3.61062437e-02 -1.19663201e-01 -6.48600459e-01 -2.58743688e-02
5.94411016e-01 2.15309277e-01 4.62822944e-01 9.64368224e-01
-2.51172900e-01 -5.22947311e-02 -1.16529047e+00 1.58061111e+00
5.44389427e-01 -1.69642770e+00 6.85430318e-02 6.09092899e-02
9.36834633e-01 6.83729053e-01 -2.41374001e-02 -1.47861904e-02
-1.52860180e-01 -9.47936893e-01 7.23199189e-01 6.64733768e-01
1.01755178e+00 -8.94629240e-01 8.97753239e-01 -1.94826171e-01
-1.34060287e+00 1.92593217e-01 -7.42386818e-01 2.73718327e-01
-1.59296542e-01 2.36226097e-01 -4.31489408e-01 8.46880972e-01
1.09920609e+00 1.33900952e+00 -1.03761709e+00 9.68760252e-01
2.22858280e-01 -6.73108324e-02 -1.46172538e-01 3.83220702e-01
4.48630154e-01 -2.56615400e-01 6.64513886e-01 1.09861994e+00
1.83272943e-01 -3.47687662e-01 2.65323281e-01 8.09709489e-01
-3.12902555e-02 1.72344804e-01 -1.05386078e+00 4.11834717e-01
2.04964280e-01 1.49781716e+00 -4.39409912e-01 -1.51386499e-01
-5.57940483e-01 9.55525637e-01 2.90753782e-01 2.59533376e-01
-3.30850959e-01 -4.33255285e-01 6.70249283e-01 3.47602181e-02
6.28351629e-01 -4.55895245e-01 3.20219457e-01 -1.14670992e+00
-6.51476905e-02 -6.97908938e-01 3.55523288e-01 -1.06376457e+00
-1.49604487e+00 5.89469612e-01 2.27109954e-01 -1.44850147e+00
-2.56012797e-01 -9.13960516e-01 -2.92019218e-01 8.36102545e-01
-1.63257468e+00 -1.47496927e+00 -7.73788154e-01 8.96331370e-01
5.05876064e-01 -6.48207188e-01 1.00475466e+00 5.35027921e-01
1.73330590e-01 4.03185219e-01 3.51120681e-01 1.93039969e-01
8.09914172e-01 -1.13415003e+00 3.57356220e-01 3.03307891e-01
1.97018668e-01 4.29844230e-01 2.86087722e-01 -7.54571259e-02
-1.60332680e+00 -1.11962271e+00 4.48378265e-01 -5.01936734e-01
4.42980021e-01 -4.19733077e-01 -5.64423084e-01 5.39214611e-01
1.65148169e-01 5.69199622e-01 3.41076761e-01 -2.66793132e-01
-6.85635686e-01 -3.16498339e-01 -1.35028601e+00 5.95747530e-02
1.08595788e+00 -6.96535945e-01 -3.94050688e-01 3.99949044e-01
3.53511542e-01 -6.46081090e-01 -1.40304339e+00 5.25675952e-01
8.00161242e-01 -1.52583718e+00 1.37304413e+00 -1.26551956e-01
4.85712111e-01 -1.57748953e-01 -1.00378680e+00 -1.25136697e+00
-3.83792222e-01 4.83962931e-02 4.54956800e-01 9.71578836e-01
-1.53228998e-01 -5.50440788e-01 3.13051492e-01 -2.01301739e-01
-1.30803093e-01 -7.47046292e-01 -1.31869483e+00 -7.58289158e-01
3.79884355e-02 -1.53847039e-01 4.22875047e-01 8.74905050e-01
-8.43311548e-01 6.98564500e-02 -2.90156245e-01 -1.59283996e-01
7.86601722e-01 1.59924343e-01 8.99498284e-01 -1.22125244e+00
1.06198257e-02 -2.44921878e-01 -1.41751552e+00 -6.09293461e-01
5.15508242e-02 -1.03491020e+00 -3.69522363e-01 -1.47047091e+00
3.81785661e-01 -3.91099721e-01 -4.17406440e-01 1.55986756e-01
5.45213103e-01 6.92524552e-01 1.89661801e-01 2.40110412e-01
-7.26077139e-01 5.70645452e-01 1.33902681e+00 -1.98235184e-01
3.85984689e-01 -4.43154335e-01 -1.24251440e-01 5.22066116e-01
3.67880732e-01 -1.01002365e-01 -4.78466637e-02 -2.96881318e-01
1.42835826e-01 3.41381505e-02 7.45796025e-01 -1.13211048e+00
-9.93324537e-03 2.79696524e-01 7.42309928e-01 -8.94371271e-01
5.29771864e-01 -6.98413074e-01 2.06605256e-01 2.83674598e-01
-3.74528617e-01 4.71312195e-01 1.88130617e-01 5.84800065e-01
-6.11191750e-01 6.45979196e-02 1.14647758e+00 -4.57422495e-01
-9.06458318e-01 6.51828349e-01 4.91336673e-01 -1.92785934e-02
8.45844567e-01 -3.72751743e-01 -2.24698320e-01 -3.91027540e-01
-4.05323148e-01 -1.77176416e-01 7.93394208e-01 6.95109189e-01
9.32816088e-01 -1.87645626e+00 -7.92828798e-01 5.42653322e-01
6.01571202e-01 -1.04011282e-01 2.00897992e-01 8.97636056e-01
-1.06656206e+00 5.45176208e-01 -8.44659507e-01 -9.48979080e-01
-1.26602852e+00 3.56430113e-01 4.24643189e-01 -1.58041045e-01
-8.34061801e-01 4.35041577e-01 3.43905032e-01 -7.69543707e-01
3.10494453e-01 -2.14027807e-01 -2.94422746e-01 -1.16057411e-01
3.87152880e-01 4.93375301e-01 2.17271984e-01 -1.07909143e+00
-4.08043176e-01 1.05561209e+00 2.78970376e-02 1.69317320e-01
1.93285060e+00 -1.26221925e-01 -3.68831933e-01 3.73939812e-01
1.89094973e+00 -2.36022696e-01 -9.17476833e-01 -2.73442417e-01
-1.85320333e-01 -8.30635607e-01 2.23761737e-01 -4.77580100e-01
-1.13227832e+00 8.98769617e-01 1.30485570e+00 -1.48167625e-01
1.05852747e+00 2.07733978e-02 5.91283023e-01 2.91260809e-01
7.55677462e-01 -9.02881384e-01 2.07677886e-01 5.80644250e-01
1.49310923e+00 -1.42288733e+00 1.27282873e-01 6.53270781e-02
-3.83800596e-01 1.43892360e+00 3.29026431e-01 -6.60481930e-01
6.60945952e-01 -1.19538672e-01 -6.04626648e-02 -5.32904387e-01
-3.15542668e-01 -3.24870422e-02 4.67373192e-01 7.78028727e-01
1.86028242e-01 -1.94334611e-01 2.06290819e-02 -2.57581264e-01
-1.13866804e-02 -2.87530720e-01 2.29954883e-01 6.67641878e-01
-1.53498128e-01 -1.10289156e+00 -2.89343983e-01 2.49440685e-01
-3.08170527e-01 2.42430829e-02 -4.42162603e-01 1.05951083e+00
1.83078483e-01 3.86753410e-01 6.60964400e-02 -4.09765661e-01
5.86862266e-01 -3.48999023e-01 7.72524297e-01 -2.08632722e-01
-5.22525430e-01 -8.42516348e-02 -5.71825318e-02 -1.13772058e+00
-8.68418396e-01 -6.72864735e-01 -6.56637371e-01 -2.07865730e-01
-6.57515377e-02 -3.90621156e-01 7.09692955e-01 6.20420218e-01
3.68877441e-01 1.90268219e-01 8.32382083e-01 -1.34538579e+00
-4.03058797e-01 -9.83176410e-01 -7.55490065e-01 9.03804541e-01
2.97336429e-01 -1.12071097e+00 -2.79095143e-01 -3.82724881e-01] | [8.198591232299805, -1.9473379850387573] |
489fae3d-112f-4433-bcf3-f6e361c01b05 | interpretable-clustering-on-dynamic-graphs | 2012.08740 | null | https://arxiv.org/abs/2012.08740v2 | https://arxiv.org/pdf/2012.08740v2.pdf | Interpretable Clustering on Dynamic Graphs with Recurrent Graph Neural Networks | We study the problem of clustering nodes in a dynamic graph, where the connections between nodes and nodes' cluster memberships may change over time, e.g., due to community migration. We first propose a dynamic stochastic block model that captures these changes, and a simple decay-based clustering algorithm that clusters nodes based on weighted connections between them, where the weight decreases at a fixed rate over time. This decay rate can then be interpreted as signifying the importance of including historical connection information in the clustering. However, the optimal decay rate may differ for clusters with different rates of turnover. We characterize the optimal decay rate for each cluster and propose a clustering method that achieves almost exact recovery of the true clusters. We then demonstrate the efficacy of our clustering algorithm with optimized decay rates on simulated graph data. Recurrent neural networks (RNNs), a popular algorithm for sequence learning, use a similar decay-based method, and we use this insight to propose two new RNN-GCN (graph convolutional network) architectures for semi-supervised graph clustering. We finally demonstrate that the proposed architectures perform well on real data compared to state-of-the-art graph clustering algorithms. | ['Carlee Joe-Wong', 'Yuhang Yao'] | 2020-12-16 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [-2.45373935e-01 8.59278664e-02 -2.14384109e-01 -1.22167915e-01
-2.09778883e-02 -6.65967405e-01 4.69281435e-01 4.15895790e-01
-3.02033335e-01 2.24820837e-01 1.49590537e-01 -2.62614399e-01
-1.95359781e-01 -8.83786321e-01 -6.67764425e-01 -9.12727594e-01
-7.08786905e-01 9.88968134e-01 4.74994451e-01 -1.79699913e-01
7.18841851e-02 2.87046909e-01 -1.18700886e+00 5.42046577e-02
4.58088696e-01 1.93599448e-01 2.86670744e-01 8.89847517e-01
-2.06706002e-01 8.57537687e-01 -6.09380066e-01 -8.85075256e-02
7.95969591e-02 -5.12598097e-01 -9.27065253e-01 2.10119575e-01
-1.66787580e-01 2.79460460e-01 -8.72497320e-01 9.94949579e-01
1.77470759e-01 1.69281736e-01 7.38310456e-01 -1.28265977e+00
-6.52002811e-01 1.37183845e+00 -6.96409285e-01 3.74270350e-01
-1.74814537e-01 -1.87632322e-01 1.21618092e+00 -1.91970110e-01
7.84174383e-01 1.10370553e+00 8.11045647e-01 5.89333773e-01
-1.43553805e+00 -5.42951345e-01 6.19888306e-01 2.86920458e-01
-1.54426587e+00 -1.70076445e-01 8.58298063e-01 -5.12089729e-01
8.57585132e-01 -2.35381529e-01 9.29668605e-01 9.82166886e-01
2.60116197e-02 6.10182285e-01 3.44032556e-01 -2.84039378e-01
2.98262894e-01 -3.94518197e-01 4.56827611e-01 7.81431556e-01
2.82632291e-01 -3.65867019e-01 -3.55288804e-01 -1.14833675e-01
5.54151893e-01 2.19293267e-01 -3.65904003e-01 -5.46145022e-01
-1.01665759e+00 8.29246938e-01 6.81935847e-01 6.25420153e-01
-2.81604409e-01 7.30692029e-01 4.22068506e-01 5.41771650e-01
4.48166281e-01 -1.63837940e-01 -2.70385295e-01 -8.86792988e-02
-9.80906725e-01 -2.59210050e-01 9.66496766e-01 9.25056696e-01
8.57383311e-01 -6.21300517e-03 1.29528254e-01 7.17977166e-01
4.18784320e-01 1.14491157e-01 4.44678575e-01 -8.35822344e-01
5.29154800e-02 6.71563506e-01 -3.71496350e-01 -1.15362990e+00
-6.42391622e-01 -5.89602590e-01 -1.34518433e+00 -3.91635269e-01
4.24688548e-01 8.24209824e-02 -8.56737494e-01 2.09059453e+00
2.13081893e-02 6.80339992e-01 -1.61270320e-01 4.34708744e-01
3.86340350e-01 6.60617948e-01 -2.82760352e-01 -4.89931703e-01
8.40701878e-01 -8.52041066e-01 -6.45879328e-01 1.31413952e-01
7.60112286e-01 -6.20766617e-02 7.61660755e-01 2.62007296e-01
-8.48371327e-01 -1.05483525e-01 -7.66311467e-01 4.74029332e-01
-2.02105924e-01 -3.90846759e-01 5.64387441e-01 6.24903679e-01
-1.74937618e+00 1.02648890e+00 -1.36458468e+00 -7.27702260e-01
5.45841977e-02 5.09308398e-01 1.04004398e-01 1.43972203e-01
-1.01619506e+00 2.68645942e-01 3.68706882e-01 2.79473156e-01
-1.05830503e+00 -4.01996613e-01 -5.35977602e-01 2.05951497e-01
1.80260807e-01 -5.96280932e-01 8.54858160e-01 -1.12702501e+00
-1.13223660e+00 7.49627948e-01 -3.16652179e-01 -7.14002430e-01
1.01013936e-01 3.97500336e-01 -4.15109515e-01 2.76573777e-01
-9.25560519e-02 4.50251341e-01 8.69960964e-01 -1.19885015e+00
-2.83741802e-01 -2.04196289e-01 -3.56797963e-01 2.11602841e-02
-6.69012487e-01 -1.92718893e-01 -8.15067887e-01 -5.23076534e-01
1.81710780e-01 -1.24236774e+00 -4.36787695e-01 -4.18947428e-01
-3.97845089e-01 -3.83902907e-01 7.59949327e-01 -2.73978800e-01
1.66494334e+00 -2.01992202e+00 5.99868357e-01 6.31839037e-01
7.36659527e-01 -1.58989146e-01 -3.39359164e-01 6.96169972e-01
-1.07162647e-01 2.72742242e-01 -3.87053698e-01 -3.98303866e-01
-3.22102636e-01 4.15383369e-01 5.75154610e-02 5.50586641e-01
-1.69878453e-01 8.59782994e-01 -1.21461570e+00 -2.31453821e-01
-1.30454212e-01 4.88765210e-01 -4.58603740e-01 -8.26195348e-03
-2.97429506e-03 1.83356851e-01 -4.63002026e-02 1.71855435e-01
2.80295610e-01 -8.21424365e-01 9.37490404e-01 9.04487260e-03
2.34187752e-01 6.88121691e-02 -1.15671349e+00 1.45332205e+00
-6.00554049e-02 8.69867563e-01 3.20992768e-01 -1.37672007e+00
8.25615466e-01 1.35131702e-01 7.54260480e-01 7.88191482e-02
3.98796611e-02 -1.34829506e-01 3.68472219e-01 -5.63514652e-03
4.82206315e-01 1.96454510e-01 1.25342086e-01 8.93537700e-01
-6.54579559e-03 3.81037176e-01 4.64247108e-01 1.02306771e+00
1.78621066e+00 -6.82726026e-01 -2.28192702e-01 -5.02583861e-01
1.88634977e-01 -2.34691292e-01 4.23510998e-01 8.08690965e-01
-2.89398700e-01 3.97069186e-01 7.55996525e-01 -2.51090854e-01
-9.64833796e-01 -1.01922452e+00 5.00216365e-01 1.04792619e+00
7.16710463e-02 -7.80411005e-01 -8.73487592e-01 -4.04425979e-01
-8.53621364e-02 1.49579003e-01 -8.29930604e-01 -4.96572256e-01
-6.14491880e-01 -1.00827444e+00 4.89367962e-01 3.15143824e-01
8.01914260e-02 -1.02347839e+00 -4.72900420e-02 5.25749266e-01
-2.38195360e-01 -8.33460271e-01 -7.10043132e-01 3.41020495e-01
-1.14587319e+00 -1.19270945e+00 -3.66174608e-01 -1.05578160e+00
9.65375066e-01 5.67325294e-01 1.32128930e+00 7.26692677e-01
-1.14413604e-01 8.27764452e-01 -5.00954509e-01 4.18479413e-01
-7.66052723e-01 6.30099177e-01 2.26909295e-01 8.94833878e-02
2.32534602e-01 -1.04442918e+00 -6.78276300e-01 1.50023118e-01
-8.94018650e-01 -2.18858123e-01 1.65276751e-01 6.91892385e-01
3.16863179e-01 3.52559298e-01 4.66872394e-01 -1.05927944e+00
8.97962809e-01 -6.82250440e-01 -4.97330755e-01 2.43346378e-01
-8.18161726e-01 2.97794282e-01 5.99902570e-01 -7.77476966e-01
-3.07540685e-01 -6.93082390e-03 3.59286964e-01 -7.36690223e-01
4.16549593e-01 6.43212736e-01 3.72521639e-01 1.77361608e-01
4.86703336e-01 1.75170407e-01 2.02499002e-01 -1.35939643e-01
7.86569774e-01 4.56962794e-01 4.47298080e-01 -4.49679881e-01
7.88190246e-01 7.51071274e-01 -1.15148216e-01 -7.72486567e-01
-3.62367630e-01 -6.65020049e-01 -8.23508680e-01 -5.66111565e-01
5.21952331e-01 -8.68619084e-01 -9.60859179e-01 5.67790389e-01
-1.05679893e+00 -8.80166352e-01 -1.27723470e-01 2.28898853e-01
-2.86635339e-01 5.85081041e-01 -1.30935514e+00 -7.69951880e-01
-2.66448140e-01 -6.86523318e-01 5.44406056e-01 -1.08992867e-01
-1.72125414e-01 -1.44752216e+00 3.38008046e-01 -3.26561511e-01
2.66070306e-01 1.16816990e-01 1.04168034e+00 -4.73143756e-01
-5.88463545e-01 7.91842118e-02 1.38872927e-02 -2.53466796e-02
2.77453780e-01 4.33884770e-01 -2.93922096e-01 -7.38307416e-01
-4.49217260e-01 1.97168291e-01 1.24675071e+00 7.45606840e-01
9.45943475e-01 -3.66548121e-01 -7.42553711e-01 4.53827202e-01
1.36581910e+00 -1.13505729e-01 4.12754655e-01 6.55738357e-03
1.02192771e+00 5.96607387e-01 -2.80540138e-01 3.55807722e-01
5.95593929e-01 2.86800712e-01 4.95892793e-01 1.63463011e-01
-1.35983810e-01 -2.03393325e-01 4.58966315e-01 1.86130822e+00
-5.80820031e-02 -3.46076071e-01 -1.15763152e+00 8.11240375e-01
-2.28436756e+00 -1.14757812e+00 -5.48568249e-01 2.07652998e+00
6.76335275e-01 2.62107641e-01 5.30387819e-01 5.90062775e-02
1.35716963e+00 1.52179495e-01 -6.49610877e-01 -6.90115914e-02
-1.68125704e-01 -3.82536016e-02 6.79169238e-01 6.04373515e-01
-6.85690403e-01 1.12646496e+00 7.27782059e+00 5.79134643e-01
-7.78633177e-01 8.84410217e-02 5.14466345e-01 -6.46311417e-02
-3.74926925e-01 1.57218128e-01 -4.28429842e-01 4.69378233e-01
1.45083225e+00 -3.26998442e-01 9.62979674e-01 4.63077158e-01
2.85823286e-01 4.25390452e-01 -1.05241334e+00 8.72478664e-01
-5.94239570e-02 -1.45336747e+00 6.02649972e-02 1.56127781e-01
7.55103230e-01 5.00269890e-01 -2.50095874e-01 1.06911190e-01
1.24708664e+00 -7.56864011e-01 4.46641952e-01 4.95500714e-01
3.46479625e-01 -8.42615008e-01 2.30389535e-01 3.26305181e-01
-1.67981398e+00 -1.27441809e-01 -4.62550819e-01 9.06227622e-03
8.59711394e-02 9.77724314e-01 -6.78942919e-01 3.27571332e-01
7.55232155e-01 1.22410142e+00 -6.29563034e-01 7.34365582e-01
-1.57123148e-01 1.06472027e+00 -3.71486753e-01 -1.55843154e-01
1.83850184e-01 -6.22788548e-01 4.43775386e-01 1.30885029e+00
1.52229011e-01 -5.41668050e-02 2.62176828e-03 8.68487477e-01
-4.34098333e-01 -1.82798907e-01 -5.93394399e-01 -1.99116141e-01
8.82687211e-01 1.33080328e+00 -1.39464235e+00 -3.31386149e-01
-1.36110306e-01 1.03211713e+00 8.15985322e-01 6.36297047e-01
-6.96714580e-01 -2.96113282e-01 5.75652182e-01 1.75823539e-01
5.38684070e-01 -6.10741735e-01 4.27182794e-01 -1.09922087e+00
-2.78324127e-01 -5.00588059e-01 6.53159797e-01 -4.06342447e-01
-1.40679514e+00 5.96757829e-01 -3.61189783e-01 -8.61902297e-01
-1.73322991e-01 -7.85568580e-02 -7.67905533e-01 1.45051762e-01
-1.07321179e+00 -8.05630445e-01 -1.42065585e-01 8.05739403e-01
1.53308138e-01 -8.32253024e-02 4.13121045e-01 2.01199651e-01
-7.60656297e-01 4.11975801e-01 4.66454744e-01 4.63627726e-01
4.56114829e-01 -1.34198356e+00 5.83635032e-01 8.53572667e-01
3.54714215e-01 8.25860083e-01 5.64690948e-01 -8.15545440e-01
-1.34324479e+00 -1.41484511e+00 4.95665044e-01 -2.36153215e-01
1.10691702e+00 -7.23940372e-01 -1.14883113e+00 8.78426492e-01
2.03073472e-01 -1.18292145e-01 5.25872886e-01 3.62637460e-01
-5.76482475e-01 -9.49587896e-02 -6.73015416e-01 6.79835677e-01
1.45550251e+00 -5.88226438e-01 -8.99443328e-02 3.47642034e-01
1.07667673e+00 2.02807799e-01 -7.92524576e-01 -8.68260264e-02
2.61139810e-01 -7.92590916e-01 6.72418654e-01 -6.30749047e-01
4.29575099e-03 -4.60290372e-01 2.21858844e-01 -1.58620584e+00
-7.43819416e-01 -1.04237604e+00 -4.70957279e-01 1.09632611e+00
3.58873516e-01 -4.78543371e-01 1.05398273e+00 4.98009846e-02
1.81426972e-01 -3.19232255e-01 -8.15603614e-01 -9.37858582e-01
1.10579602e-01 -3.21507275e-01 2.79090106e-01 1.24093831e+00
1.80710644e-01 5.76109827e-01 -2.51238704e-01 1.89579710e-01
8.15993071e-01 -3.79804261e-02 5.03252923e-01 -1.52962351e+00
-2.64529318e-01 -8.52726281e-01 -4.46915001e-01 -9.52906311e-01
4.95294541e-01 -1.21398354e+00 -1.46142140e-01 -1.51913977e+00
4.52342570e-01 -3.46212417e-01 -3.13509911e-01 2.62759030e-01
-1.40983090e-01 -1.06730081e-01 2.26067286e-02 6.41207039e-01
-9.31623161e-01 3.87112558e-01 5.59957206e-01 -3.03248703e-01
-5.86241841e-01 -1.75224498e-01 -5.28914750e-01 3.12371790e-01
8.16475749e-01 -7.35737562e-01 -5.26276529e-01 -2.69160867e-01
6.00280225e-01 -1.57884985e-01 -3.71494927e-02 -8.07529688e-01
7.43310392e-01 2.85700977e-01 -2.69952089e-01 -5.77530921e-01
-1.72753707e-01 -6.55343115e-01 4.16796863e-01 8.32132995e-01
-4.31804031e-01 2.74576694e-01 -2.77264863e-01 1.26902294e+00
1.07243210e-01 1.22461423e-01 6.81338310e-01 -1.32707343e-01
-2.30530620e-01 5.92758715e-01 -8.44724238e-01 1.26534134e-01
7.65071273e-01 -3.67497653e-02 -1.59251556e-01 -8.38759959e-01
-1.03931952e+00 6.43211961e-01 4.98167366e-01 3.85053873e-01
5.28127909e-01 -1.28719211e+00 -7.35619366e-01 -2.76298046e-01
-4.84353900e-02 -8.50007311e-02 -3.70922647e-02 8.65567803e-01
-4.05900061e-01 -1.29418045e-01 2.79438078e-01 -8.68928373e-01
-1.16926754e+00 8.42326045e-01 5.19922554e-01 -4.76248682e-01
-6.89937532e-01 5.94725013e-01 -1.03308566e-01 -4.33570862e-01
3.72453928e-01 -1.98754877e-01 -1.66189060e-01 2.10677102e-01
2.34482750e-01 3.43318820e-01 -1.13314085e-01 -4.65087503e-01
-2.76226401e-01 5.31024575e-01 -1.48776099e-01 -6.27236441e-02
1.43914366e+00 -4.17987168e-01 -5.48627615e-01 8.48479390e-01
1.15435338e+00 -4.75931525e-01 -9.85101521e-01 -4.27548110e-01
3.83504659e-01 1.39069825e-01 -2.25455135e-01 1.19112745e-01
-1.48932612e+00 6.38723850e-01 3.19434613e-01 7.18650579e-01
8.96834910e-01 3.48638624e-01 5.86189389e-01 5.25415659e-01
2.42559612e-01 -1.11024773e+00 3.21211874e-01 6.57496929e-01
2.06540257e-01 -6.73267663e-01 -2.02894464e-01 -2.42276922e-01
-2.41073415e-01 1.19410002e+00 2.40249038e-01 -2.72161573e-01
1.10640264e+00 2.03891262e-01 -2.48086795e-01 -5.56768119e-01
-1.13007915e+00 -2.89368659e-01 -3.44556957e-01 7.95295894e-01
2.93896228e-01 2.57280618e-01 -1.23925470e-01 5.39651476e-02
-4.32167128e-02 -5.97113609e-01 9.56979036e-01 5.59037745e-01
-5.09824872e-01 -1.06694508e+00 3.48109789e-02 5.79029322e-01
-7.28728995e-02 -1.97841838e-01 -6.44093812e-01 5.33605456e-01
-5.40354967e-01 1.04164016e+00 4.79382336e-01 -6.27673328e-01
-1.32717863e-01 -2.58803312e-02 3.41763288e-01 -6.38996303e-01
-6.47912085e-01 7.28859678e-02 -2.37496942e-01 -4.13203269e-01
-7.59761333e-01 -6.47866964e-01 -1.43623257e+00 -7.48977840e-01
-4.30551499e-01 4.47008848e-01 3.47349405e-01 5.32893062e-01
4.30035025e-01 7.65956581e-01 8.52754176e-01 -5.26463747e-01
8.32594931e-03 -9.23172534e-01 -8.88518631e-01 5.62714517e-01
1.82742134e-01 -2.70992756e-01 -7.06405878e-01 1.05053559e-01] | [7.131778717041016, 5.782120704650879] |
9c8634fd-ccd5-460e-8fe7-e44c256dbb84 | clas-coordinating-multi-robot-manipulation | 2211.15824 | null | https://arxiv.org/abs/2211.15824v1 | https://arxiv.org/pdf/2211.15824v1.pdf | CLAS: Coordinating Multi-Robot Manipulation with Central Latent Action Spaces | Multi-robot manipulation tasks involve various control entities that can be separated into dynamically independent parts. A typical example of such real-world tasks is dual-arm manipulation. Learning to naively solve such tasks with reinforcement learning is often unfeasible due to the sample complexity and exploration requirements growing with the dimensionality of the action and state spaces. Instead, we would like to handle such environments as multi-agent systems and have several agents control parts of the whole. However, decentralizing the generation of actions requires coordination across agents through a channel limited to information central to the task. This paper proposes an approach to coordinating multi-robot manipulation through learned latent action spaces that are shared across different agents. We validate our method in simulated multi-robot manipulation tasks and demonstrate improvement over previous baselines in terms of sample efficiency and learning performance. | ['Patrick van der Smagt', 'Maximilian Karl', 'Elie Aljalbout'] | 2022-11-28 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 5.21924347e-02 1.76523343e-01 -1.66325778e-01 4.84904312e-02
-9.11391675e-01 -8.34116757e-01 6.87765002e-01 3.76215093e-02
-6.87850356e-01 1.07922292e+00 4.90132086e-02 -1.20734744e-01
-4.45411503e-01 -5.70553184e-01 -8.17681372e-01 -6.33197188e-01
-5.22099495e-01 1.06937456e+00 1.86911479e-01 -3.55255991e-01
1.28664494e-01 3.65630031e-01 -1.33643198e+00 1.29966214e-01
5.88484883e-01 4.45220858e-01 4.57578272e-01 9.66397226e-01
4.17124271e-01 7.27879882e-01 -4.75613803e-01 5.41418433e-01
7.44675696e-01 -2.67592430e-01 -9.20012116e-01 3.70537043e-01
-1.56554449e-02 -5.30017436e-01 -3.24909776e-01 8.84444356e-01
3.93599302e-01 6.71214938e-01 6.96452975e-01 -1.71652043e+00
-1.59316048e-01 8.14018548e-01 -7.05944359e-01 -2.10586041e-01
1.94762245e-01 6.96831167e-01 9.42934513e-01 -4.60722297e-02
5.57421327e-01 1.79982913e+00 4.14183661e-02 6.03001952e-01
-1.52151549e+00 -5.40247262e-01 6.62707567e-01 1.62964389e-02
-7.33308077e-01 -3.39715391e-01 1.51515096e-01 -2.73505211e-01
1.16912913e+00 -3.16167384e-01 3.88843775e-01 1.15636027e+00
5.00814497e-01 8.52972746e-01 1.07191563e+00 -1.52484566e-01
4.76264387e-01 -4.00606275e-01 -3.57130438e-01 6.74536347e-01
1.88384071e-01 1.69804797e-01 -2.76231229e-01 -3.36816490e-01
1.02483618e+00 4.04036604e-02 1.99325234e-01 -1.06223238e+00
-1.62215483e+00 8.25811625e-01 4.01805639e-01 -7.49886036e-02
-7.71461785e-01 6.11671150e-01 5.04675329e-01 8.54225278e-01
-3.47271204e-01 1.05087197e+00 -7.61783600e-01 -2.64052123e-01
-1.04699217e-01 8.79941285e-01 9.33872998e-01 1.30806935e+00
5.67742169e-01 1.69442855e-02 1.14151485e-01 5.12407780e-01
1.59579501e-01 4.06671226e-01 3.80366892e-01 -1.62536550e+00
7.86385238e-01 3.27341676e-01 6.42329574e-01 -2.99373418e-01
-7.44090557e-01 1.91035151e-01 -5.79592168e-01 8.99113834e-01
6.48907125e-01 -8.63990784e-01 -8.53127360e-01 1.87759459e+00
5.54694474e-01 -3.67925823e-01 3.35320562e-01 8.54856074e-01
-3.05422395e-01 6.57474995e-01 6.03968278e-03 -2.23032519e-01
1.22432005e+00 -1.33108687e+00 -5.64394236e-01 -7.14372456e-01
4.58283842e-01 -3.82677227e-01 7.89530218e-01 6.36731327e-01
-1.27382576e+00 -3.77009600e-01 -9.51593995e-01 1.70735419e-01
-1.02855302e-01 -1.12887599e-01 7.82005191e-01 -2.26955548e-01
-8.69751513e-01 6.25595093e-01 -1.35410917e+00 -2.27814093e-01
1.83681011e-01 8.42536271e-01 -4.96082455e-01 1.46138296e-02
-7.79956937e-01 1.37912953e+00 5.71780264e-01 9.64127667e-03
-1.53387558e+00 -1.32366702e-01 -6.51150048e-01 1.47542581e-01
9.70153987e-01 -7.80255616e-01 1.93243957e+00 -5.69083810e-01
-1.85002315e+00 -4.84220721e-02 3.43439907e-01 -2.18545750e-01
4.39108938e-01 -3.29074621e-01 4.06150907e-01 -2.03046016e-02
3.32504213e-01 7.30546772e-01 1.06621969e+00 -1.27233303e+00
-9.55356479e-01 -1.54972434e-01 5.58721364e-01 6.46879673e-01
9.01899710e-02 -3.15896183e-01 1.52044505e-01 -1.50125787e-01
-3.93535607e-02 -1.52657747e+00 -8.43352973e-01 -7.52664879e-02
-6.76858649e-02 -2.56370097e-01 6.73552930e-01 -8.54108632e-02
3.19977164e-01 -1.92072487e+00 8.92107964e-01 4.19363678e-02
2.58710623e-01 -3.06886494e-01 -6.40599906e-01 7.52835691e-01
-2.58608256e-03 -2.33837888e-01 1.35630876e-01 -6.93116635e-02
4.49865729e-01 4.03690606e-01 -1.42361820e-01 6.38866425e-01
4.16690148e-02 6.89167321e-01 -1.33350146e+00 -1.87992334e-01
3.93940844e-02 -2.86948979e-01 -6.79288864e-01 2.49167800e-01
-7.13689744e-01 6.17229819e-01 -9.68121290e-01 3.42666864e-01
4.13934141e-02 -1.69169202e-01 6.75264120e-01 3.29316705e-01
-5.60511351e-02 1.30337924e-01 -1.58493423e+00 1.90181875e+00
-6.32584155e-01 1.54183164e-01 6.32980943e-01 -9.32541430e-01
3.48188728e-01 5.19265771e-01 7.93869674e-01 -3.10753226e-01
6.37878254e-02 1.78962201e-01 6.13215446e-01 -4.96602982e-01
5.39400339e-01 -1.28613353e-01 -5.17650485e-01 8.85723531e-01
9.65460613e-02 -7.34304368e-01 4.04999942e-01 4.26675081e-02
1.56477118e+00 3.52955371e-01 4.43307251e-01 1.42834544e-01
-9.38881561e-03 3.82380992e-01 4.03830647e-01 1.12286472e+00
-2.76792645e-01 -1.36758178e-01 7.86361516e-01 -1.80119693e-01
-1.14185894e+00 -1.03277493e+00 4.83199239e-01 1.26679313e+00
1.86344713e-01 -2.79718220e-01 -2.30946645e-01 -5.76357067e-01
3.31754744e-01 4.38156039e-01 -4.50660437e-01 -2.87454855e-02
-8.64726126e-01 -6.66567445e-01 -3.35823707e-02 3.97705078e-01
2.45512307e-01 -1.40672410e+00 -1.26927221e+00 5.47387183e-01
1.33510619e-01 -1.05676889e+00 -2.74119943e-01 5.88679194e-01
-6.80042624e-01 -1.17946637e+00 -2.58728713e-01 -7.01293290e-01
5.73039532e-01 3.40572923e-01 7.19329417e-01 -4.03677195e-01
-4.46017772e-01 7.26441801e-01 -1.15325786e-01 -4.02458131e-01
-5.73668540e-01 2.02508703e-01 4.61468160e-01 -5.87192595e-01
-3.81962091e-01 -5.97447157e-01 -3.73147219e-01 3.70688550e-02
-7.00874984e-01 -1.16850240e-02 1.03851748e+00 1.04157650e+00
1.39855459e-01 2.64609575e-01 6.06315017e-01 -5.05495548e-01
1.04187179e+00 -5.33667386e-01 -9.07786071e-01 1.30889565e-01
-1.97555259e-01 4.17478859e-01 5.53373575e-01 -9.52766895e-01
-9.73899484e-01 4.93683785e-01 7.95502841e-01 -2.26256698e-01
-2.76938677e-01 2.79901594e-01 9.65535119e-02 1.52921051e-01
5.69453895e-01 -1.38970435e-01 3.96555245e-01 -3.87514085e-02
6.23848140e-01 4.24227804e-01 9.65994671e-02 -1.17065823e+00
7.89109766e-01 -5.87314740e-03 3.27777237e-01 -5.76659262e-01
-1.56178460e-01 -1.08836807e-01 -5.22545874e-01 -9.58909020e-02
6.91132903e-01 -9.61019874e-01 -1.41345716e+00 3.36948752e-01
-1.26440251e+00 -1.04834425e+00 -3.54230762e-01 5.77654541e-01
-1.27386153e+00 6.75527006e-02 -7.13676035e-01 -7.78289139e-01
3.80589128e-01 -1.53716803e+00 1.13916600e+00 7.87188783e-02
-2.74864644e-01 -5.50333619e-01 2.04719916e-01 -4.50797230e-02
4.79037076e-01 2.05148026e-01 1.04523349e+00 -5.09561837e-01
-8.53342175e-01 -6.33645579e-02 2.22249269e-01 -1.71590373e-01
2.92882323e-01 -4.97753918e-01 -4.02281463e-01 -7.35208929e-01
-1.44602492e-01 -1.06422901e+00 3.04918557e-01 3.31469238e-01
8.21759880e-01 -4.25341010e-01 -6.17349684e-01 -2.43219763e-01
1.20826423e+00 3.49541366e-01 5.71344085e-02 5.43563366e-01
3.01130235e-01 7.53540158e-01 9.83237743e-01 7.50772536e-01
4.33372945e-01 7.16909587e-01 6.44402266e-01 3.79899532e-01
3.75319332e-01 8.99233222e-02 6.72850013e-01 2.19143420e-01
2.81314328e-02 -3.28743488e-01 -8.43059838e-01 4.46668595e-01
-2.45809650e+00 -9.78827119e-01 4.03462499e-01 1.89862907e+00
8.84418607e-01 8.51296112e-02 3.68606120e-01 -4.48301941e-01
4.01258260e-01 -5.11364639e-03 -1.04682004e+00 -2.49158949e-01
4.95449483e-01 -7.41682351e-02 5.38344562e-01 5.92072964e-01
-1.10912824e+00 9.47890878e-01 6.52490139e+00 3.10420305e-01
-6.63336039e-01 6.41777888e-02 -3.96902896e-02 -5.73200762e-01
5.35173535e-01 7.18077272e-02 -5.48003435e-01 1.06068760e-01
8.64596844e-01 -1.57675460e-01 1.04587996e+00 7.92666256e-01
2.37760454e-01 -6.33749068e-01 -1.74630058e+00 6.80492699e-01
-4.96180981e-01 -7.78934717e-01 -3.13334018e-01 2.41583973e-01
8.17992032e-01 3.43361765e-01 -9.66204479e-02 8.16777766e-01
1.39213562e+00 -9.33374584e-01 5.76200664e-01 1.62082821e-01
2.19072714e-01 -5.06265700e-01 2.00292245e-01 7.87033558e-01
-1.03519332e+00 -8.49104285e-01 -2.15965107e-01 -6.07782841e-01
3.12759012e-01 -3.99352700e-01 -8.55346203e-01 2.66767263e-01
4.21379864e-01 2.35636339e-01 1.90102756e-02 7.88589120e-01
-1.09938309e-01 -2.11245716e-01 -3.78438056e-01 -1.00864731e-01
5.55390954e-01 -3.09892148e-01 4.75726664e-01 4.48030502e-01
1.70765016e-02 1.56734750e-01 1.08846509e+00 5.75159132e-01
2.52374530e-01 -4.98008072e-01 -9.31265235e-01 -2.68216163e-01
5.58110952e-01 1.23645079e+00 -7.42101550e-01 -3.64726573e-01
-4.36655432e-01 7.58883834e-01 6.32174134e-01 4.74604607e-01
-7.53978968e-01 -3.71375471e-01 1.00898910e+00 -4.08941418e-01
3.55782092e-01 -7.97668636e-01 2.09963366e-01 -9.76247609e-01
-5.58128431e-02 -1.50010788e+00 1.14921935e-01 -5.05864441e-01
-1.24176693e+00 1.20729148e-01 3.54683071e-01 -1.15661883e+00
-7.71590054e-01 -6.39561176e-01 -1.32007822e-01 6.05670631e-01
-1.22214222e+00 -9.89467859e-01 9.97236073e-02 4.56623524e-01
8.66262615e-01 -3.48722398e-01 1.00910652e+00 -2.76044279e-01
-4.08890784e-01 -7.90635571e-02 1.65418550e-01 -1.49678156e-01
6.83734596e-01 -1.33008051e+00 1.00780152e-01 3.69376570e-01
-3.26503724e-01 5.13822973e-01 7.05306053e-01 -5.39277971e-01
-1.95360351e+00 -6.46550000e-01 -7.37799928e-02 -4.75288808e-01
1.13948762e+00 -3.88385594e-01 -4.93661344e-01 1.10288453e+00
4.26637590e-01 -1.43969864e-01 1.02473497e-01 2.99474746e-01
-2.43195761e-02 1.96707353e-01 -8.81051838e-01 8.63722384e-01
9.11709607e-01 -2.58354098e-01 -7.10982144e-01 7.13758707e-01
5.05926907e-01 -7.40581155e-01 -8.30288529e-01 1.53961763e-01
5.48789799e-01 -2.92838007e-01 8.71893942e-01 -1.05281591e+00
3.98881674e-01 -3.78495306e-01 3.49114090e-03 -1.84453285e+00
-5.34114778e-01 -7.91949928e-01 -1.77693352e-01 4.61614490e-01
3.19910020e-01 -5.97530842e-01 5.62346578e-01 8.43293905e-01
-1.93049181e-02 -4.17239219e-01 -8.75252426e-01 -8.84717822e-01
2.68531680e-01 1.89742371e-01 3.12835246e-01 6.93767309e-01
5.21413386e-01 6.96969569e-01 -3.51412386e-01 3.48759472e-01
4.31119353e-01 3.09053868e-01 1.15019369e+00 -9.28750634e-01
-7.96402454e-01 -4.34711099e-01 4.16074432e-02 -1.13426006e+00
5.51283062e-01 -6.11517370e-01 6.81702614e-01 -1.48558414e+00
1.40109852e-01 -7.18080759e-01 6.14235662e-02 8.17085564e-01
1.34228215e-01 -6.82122529e-01 5.99563122e-01 1.36512503e-01
-9.89179671e-01 6.66512251e-01 1.53239775e+00 -2.99211204e-01
-4.82791662e-01 -4.84864675e-02 -3.74679714e-01 6.14458501e-01
9.62045491e-01 -3.33506703e-01 -6.72483742e-01 -8.42674077e-01
1.03700824e-01 8.01404536e-01 2.11061224e-01 -9.22196150e-01
4.48227197e-01 -9.00769174e-01 1.66955665e-01 1.20771006e-02
6.58322573e-01 -1.01750326e+00 1.29374983e-02 9.06354904e-01
-6.74295425e-01 5.65554857e-01 1.37289673e-01 8.81915450e-01
4.98316772e-02 -1.88438624e-01 5.89299917e-01 -6.34203434e-01
-5.97267449e-01 2.08694890e-01 -9.15500879e-01 -1.33260190e-01
1.52572501e+00 3.61647099e-01 -4.41210538e-01 -4.21487659e-01
-8.98766935e-01 9.31176901e-01 2.95640588e-01 5.20782351e-01
3.34331959e-01 -8.80905867e-01 -4.52298045e-01 -1.03906669e-01
-1.25337899e-01 3.03422093e-01 -5.03983051e-02 6.40693009e-01
-5.48377596e-02 3.06728691e-01 -7.65008867e-01 -2.47857571e-01
-1.05754948e+00 7.09957004e-01 2.74474084e-01 -4.79166597e-01
-5.80412447e-01 3.60175848e-01 8.68752226e-02 -6.77735806e-01
3.57033700e-01 -4.49565023e-01 8.19682404e-02 5.21250591e-02
1.80309385e-01 3.52968603e-01 -5.54880738e-01 7.68530145e-02
-7.93715939e-02 1.40819073e-01 -3.62130076e-01 -5.20187259e-01
1.54611540e+00 -4.44455259e-02 -1.42468125e-01 4.27037030e-01
5.23605824e-01 -5.83629608e-01 -1.60478687e+00 -1.52003512e-01
2.02381592e-02 -2.62423724e-01 -3.14018428e-01 -7.05651402e-01
-5.30575573e-01 4.63319033e-01 2.41887271e-01 1.35466456e-01
5.13667524e-01 -5.39105237e-02 3.49636167e-01 1.34698606e+00
9.64726448e-01 -1.45743656e+00 5.96042693e-01 8.86568844e-01
9.42971468e-01 -1.38976574e+00 2.22590804e-01 1.28931209e-01
-8.15688431e-01 1.20075452e+00 1.10020733e+00 -5.10298908e-01
2.90484816e-01 4.46957916e-01 -1.71531007e-01 -1.48264915e-01
-1.34730875e+00 -1.32206216e-01 -4.18064892e-01 6.47045851e-01
-1.84605233e-02 8.67844373e-02 1.02141589e-01 2.51561496e-02
9.07259136e-02 -2.10004419e-01 8.25094044e-01 1.55548501e+00
-5.90276301e-01 -1.41549993e+00 -3.45588952e-01 4.93234277e-01
7.09705576e-02 5.92919946e-01 5.44326566e-03 1.00458193e+00
-1.88615426e-01 7.79172897e-01 -1.84763130e-03 9.28068608e-02
2.92138517e-01 8.55209976e-02 9.93105948e-01 -1.10010231e+00
-4.37049955e-01 7.64510855e-02 4.92810495e-02 -9.02642488e-01
-3.84024829e-01 -1.03943455e+00 -1.49661636e+00 -1.72720905e-02
-5.52721582e-02 -2.96722371e-02 5.99048913e-01 8.42337549e-01
3.64730150e-01 8.24064672e-01 5.66618979e-01 -1.43579757e+00
-1.50972962e+00 -8.40594947e-01 -5.04798114e-01 7.37470686e-02
7.67833412e-01 -9.99072313e-01 2.35295705e-02 -2.08967507e-01] | [4.292304515838623, 1.3393892049789429] |
c5597686-3675-48f7-bb88-a3ee90e1f9b5 | evaluating-copy-blend-augmentation-for-low | 2103.05889 | null | https://arxiv.org/abs/2103.05889v1 | https://arxiv.org/pdf/2103.05889v1.pdf | Evaluating COPY-BLEND Augmentation for Low Level Vision Tasks | Region modification-based data augmentation techniques have shown to improve performance for high level vision tasks (object detection, semantic segmentation, image classification, etc.) by encouraging underlying algorithms to focus on multiple discriminative features. However, as these techniques destroy spatial relationship with neighboring regions, performance can be deteriorated when using them to train algorithms designed for low level vision tasks (low light image enhancement, image dehazing, deblurring, etc.) where textural consistency between recovered and its neighboring regions is important to ensure effective performance. In this paper, we examine the efficacy of a simple copy-blend data augmentation technique that copies patches from noisy images and blends onto a clean image and vice versa to ensure that an underlying algorithm localizes and recovers affected regions resulting in increased perceptual quality of a recovered image. To assess performance improvement, we perform extensive experiments alongside different region modification-based augmentation techniques and report observations such as improved performance, reduced requirement for training dataset, and early convergence across tasks such as low light image enhancement, image dehazing and image deblurring without any modification to baseline algorithm. | ['Kyung-Soo Kim', 'Kuk-Jin Yoon', 'Sandeep Singh Sengar', 'Pranjay Shyam'] | 2021-03-10 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 1.08434224e+00 -9.20387730e-02 2.96733111e-01 -2.46491641e-01
-7.44118392e-01 -4.61889118e-01 7.27636755e-01 2.56432444e-01
-6.47735596e-01 5.28761506e-01 1.26463294e-01 -7.31422380e-02
1.41219541e-01 -4.66410786e-01 -8.49494874e-01 -1.00647712e+00
2.06089392e-01 -2.44956240e-01 5.84410369e-01 -7.23588243e-02
5.01055539e-01 6.83618724e-01 -1.53025258e+00 3.24350536e-01
6.37248337e-01 7.37263858e-01 2.00310558e-01 1.00619006e+00
2.16803685e-01 6.87159657e-01 -6.63961709e-01 8.10694471e-02
4.60983753e-01 -3.68761212e-01 -7.44034111e-01 4.47751224e-01
1.00775707e+00 -3.79423201e-01 -1.34849474e-01 1.05418575e+00
4.01267737e-01 4.41285431e-01 6.94022775e-01 -9.77940440e-01
-5.90603650e-01 -1.54976398e-01 -1.14329743e+00 5.26947856e-01
1.65411308e-02 3.30036432e-01 3.22338253e-01 -7.55055845e-01
5.86685598e-01 9.37809289e-01 8.52776885e-01 3.61286312e-01
-1.67846787e+00 -5.69139302e-01 -3.04455012e-01 -1.26672491e-01
-8.44676137e-01 -6.79398119e-01 5.81821263e-01 -1.38280064e-01
8.05681765e-01 3.29304695e-01 1.92689791e-01 5.23594022e-01
2.19068259e-01 4.93554175e-01 1.68307137e+00 -5.37575185e-01
1.01009861e-01 2.89505899e-01 4.94239703e-02 4.11013275e-01
2.62618512e-01 3.70468199e-01 -4.84856576e-01 1.78577751e-01
8.54025424e-01 -2.01139823e-01 -4.31392848e-01 -2.30919391e-01
-1.10335743e+00 5.04045606e-01 7.06309378e-01 3.44988674e-01
-4.30370271e-01 2.25291774e-01 2.12299213e-01 3.84649694e-01
5.07541776e-01 6.38917327e-01 -1.90425083e-01 2.14121461e-01
-1.26386476e+00 1.51782647e-01 1.44449785e-01 4.34765309e-01
9.49360430e-01 1.37064442e-01 -1.57195285e-01 1.00884163e+00
-8.80285501e-02 2.81691879e-01 4.56571430e-01 -1.21848118e+00
1.60837859e-01 4.18058842e-01 -4.20003720e-02 -9.67807174e-01
-3.05487096e-01 -5.05150735e-01 -8.78034890e-01 9.67857718e-01
4.78968829e-01 4.93709818e-02 -1.48324323e+00 1.65048695e+00
4.73685771e-01 2.63678133e-01 3.23744774e-01 9.24682736e-01
5.05252004e-01 7.50760376e-01 1.12442806e-01 -1.10075302e-01
1.27069771e+00 -9.48049188e-01 -6.07267201e-01 -4.86428261e-01
3.54158014e-01 -1.18581033e+00 9.72509384e-01 5.21271586e-01
-1.28690469e+00 -8.36188197e-01 -1.10030246e+00 -2.62580812e-01
-3.90248358e-01 -2.35194847e-01 3.82288724e-01 7.07114697e-01
-1.22776103e+00 5.76013863e-01 -6.01018250e-01 -3.52079898e-01
5.65091968e-01 3.74718517e-01 -7.38386035e-01 -5.20166874e-01
-6.02203071e-01 9.30889785e-01 4.44859356e-01 -2.74279743e-01
-9.12636638e-01 -8.41299474e-01 -8.40312362e-01 -3.43654245e-01
4.76246141e-02 -3.27043504e-01 7.47151554e-01 -1.04048443e+00
-9.19053614e-01 9.60217178e-01 -7.86521435e-02 -5.36132753e-01
3.62465233e-01 -2.11219192e-02 -2.50367999e-01 2.47526407e-01
1.00152127e-01 1.07143712e+00 1.42914093e+00 -1.67420578e+00
-7.04446375e-01 -3.97110313e-01 -2.51689225e-01 3.34367841e-01
-9.95349512e-02 1.04410931e-01 -2.63133824e-01 -1.01240230e+00
2.02575400e-01 -7.84830749e-01 -1.86325520e-01 6.44775778e-02
-1.18489705e-01 3.81143063e-01 1.38803065e+00 -9.73371506e-01
5.46866834e-01 -2.30295730e+00 -2.31096059e-01 1.29368111e-01
2.16691181e-01 4.55813408e-01 -4.13402140e-01 -7.93842599e-02
-3.61425132e-01 -8.39987304e-03 -7.66234457e-01 -3.09465468e-01
-6.77994728e-01 2.66714960e-01 -1.12493262e-01 7.75008798e-01
5.15632093e-01 8.85241091e-01 -6.30643725e-01 -3.69450063e-01
5.78450978e-01 7.68315196e-01 -4.49328363e-01 6.52987808e-02
1.46525756e-01 4.28618491e-01 2.29126528e-01 6.08026922e-01
8.99618685e-01 1.16785474e-01 -4.88487810e-01 -4.28636581e-01
-5.51840253e-02 -2.86979020e-01 -1.14378548e+00 1.30817258e+00
-5.68904757e-01 1.05095899e+00 4.10798341e-01 -9.51348126e-01
7.27255940e-01 1.60700217e-01 3.84378761e-01 -8.55775177e-01
5.55708967e-02 5.13985865e-02 7.10888952e-02 -3.34050119e-01
6.83331907e-01 -1.92043930e-01 4.49456006e-01 5.77483594e-01
-1.82897434e-01 -4.35048133e-01 -4.10942622e-02 2.22582743e-01
1.10454488e+00 -4.40878756e-02 -1.08275667e-01 -1.04801573e-01
3.12104702e-01 1.56348854e-01 -4.08561118e-02 8.63192737e-01
-2.30180904e-01 1.06865382e+00 -1.03558533e-01 -3.89967747e-02
-1.47291243e+00 -9.57921684e-01 -3.50815922e-01 1.09610617e+00
3.23361844e-01 2.61603981e-01 -8.39500308e-01 -4.55342054e-01
-3.19524080e-01 6.56020939e-01 -7.52185285e-01 -4.33437198e-01
-6.17817879e-01 -1.15019703e+00 5.04584491e-01 4.35788095e-01
9.11011040e-01 -1.03946519e+00 -6.79126740e-01 2.04759866e-01
-7.88543075e-02 -1.13361466e+00 -5.33557832e-01 4.71977562e-01
-9.78418052e-01 -7.85632730e-01 -8.82293403e-01 -9.43337619e-01
9.72017705e-01 6.33368075e-01 9.08164144e-01 1.74447134e-01
-5.72197139e-01 2.48040706e-01 -3.03170234e-01 -9.16831493e-02
-6.62363887e-01 -3.82415712e-01 -2.46472254e-01 -2.60348655e-02
-2.80675232e-01 -4.60272014e-01 -7.86276877e-01 4.65014666e-01
-1.59431732e+00 -5.06556891e-02 8.69101465e-01 1.04022670e+00
5.01527011e-01 5.99936366e-01 2.05702916e-01 -7.76397407e-01
6.41078949e-01 2.94046383e-02 -2.26602793e-01 1.02227099e-01
-8.02232623e-01 1.38881907e-01 2.27595553e-01 -6.79040611e-01
-1.09282970e+00 8.76331106e-02 9.21170190e-02 -3.07049602e-01
-4.61107522e-01 9.86636952e-02 2.01907367e-01 -6.43272281e-01
8.76673162e-01 5.34755409e-01 3.60382169e-01 -3.26592237e-01
4.21996534e-01 5.33098221e-01 1.01308286e+00 -1.97031632e-01
9.73650634e-01 6.66229725e-01 -2.91542131e-02 -9.57256794e-01
-5.43179989e-01 -6.64761901e-01 -5.29759228e-01 -1.77524880e-01
8.67041886e-01 -7.49644279e-01 -1.62947461e-01 5.67222238e-01
-9.08614337e-01 -5.42981267e-01 -3.69766295e-01 2.04007655e-01
-3.68474513e-01 4.18887824e-01 -4.48647231e-01 -6.22740805e-01
-2.20401064e-01 -1.07729614e+00 1.08940768e+00 4.19694841e-01
-7.24152848e-02 -9.34584558e-01 -1.64560840e-01 7.01138914e-01
6.52323723e-01 4.23995078e-01 8.37692678e-01 -1.21510163e-01
-3.84454995e-01 -2.23503351e-01 -6.02706909e-01 7.60073245e-01
3.54218751e-01 -2.37111181e-01 -1.18721485e+00 -5.09326398e-01
6.90981224e-02 -2.79569834e-01 1.08685839e+00 5.49671769e-01
9.56172526e-01 -2.30159059e-01 -7.43848905e-02 5.26652396e-01
1.52465117e+00 6.63112924e-02 9.87371564e-01 5.18694401e-01
5.40558159e-01 7.65824020e-01 4.68823910e-01 -2.80456871e-01
-3.51813734e-01 6.36702955e-01 4.32262748e-01 -9.04389381e-01
-8.08529675e-01 2.85460204e-01 3.70714426e-01 -1.25385104e-02
1.47257477e-01 -9.42986161e-02 -5.39033353e-01 8.50802481e-01
-1.24819696e+00 -8.15732121e-01 -2.03901321e-01 2.34409308e+00
9.18509185e-01 7.87479505e-02 1.18112408e-01 4.01100218e-01
6.93270862e-01 3.84931043e-02 -6.82009757e-01 -4.29656357e-01
-2.65719742e-01 3.72321069e-01 9.21397746e-01 6.43649220e-01
-1.10485804e+00 8.57244492e-01 6.45668459e+00 8.93063068e-01
-1.14013457e+00 1.03447177e-01 1.09118760e+00 2.27195080e-02
1.30952194e-01 -2.48501942e-01 -3.27905715e-01 2.38673583e-01
5.45044601e-01 4.04486865e-01 5.26814699e-01 3.60643625e-01
3.35021168e-01 -5.90092003e-01 -6.80728734e-01 8.51765096e-01
2.19895631e-01 -1.22936273e+00 -1.41702592e-01 1.48482427e-01
1.16079903e+00 -7.27608148e-03 3.73013705e-01 -2.47333273e-01
2.15789795e-01 -1.19612861e+00 4.21392620e-01 1.92171171e-01
6.08500004e-01 -6.71418726e-01 6.90737307e-01 -2.41187941e-02
-7.69157231e-01 1.42805979e-01 -2.09893867e-01 3.35240096e-01
-1.70876812e-02 5.64857006e-01 -9.50430274e-01 2.54779160e-01
7.59049952e-01 4.72870618e-01 -9.08737302e-01 1.07852888e+00
4.39583696e-02 5.06139457e-01 -3.45405161e-01 6.15089834e-01
4.16760057e-01 -1.70686647e-01 6.02700353e-01 1.13514483e+00
-1.28974795e-01 8.97569731e-02 -1.62320927e-01 8.35341811e-01
7.61309192e-02 -1.96051151e-01 -5.34590185e-01 2.29109451e-01
3.25081050e-02 1.30886459e+00 -1.19254303e+00 -3.51447374e-01
-2.35139877e-01 1.35274482e+00 -3.17351639e-01 5.31239808e-01
-6.71659052e-01 -4.57904339e-01 5.21104693e-01 4.99275535e-01
1.93970814e-01 -1.35627523e-01 -6.59134388e-01 -5.86581111e-01
-1.13467149e-01 -1.01656151e+00 1.15534492e-01 -1.06898427e+00
-9.61584091e-01 5.04763305e-01 -1.91194251e-01 -9.00198817e-01
1.16894566e-01 -3.63167614e-01 -7.17120826e-01 9.51114357e-01
-1.56959128e+00 -1.12951100e+00 -4.18392032e-01 7.04481423e-01
6.61892831e-01 2.67894566e-02 1.83287248e-01 2.85796911e-01
-2.91381150e-01 4.88516271e-01 1.93756551e-01 7.72369001e-03
8.56210709e-01 -1.33497918e+00 2.20094994e-01 1.35530090e+00
1.14818707e-01 4.00842518e-01 8.92750084e-01 -7.48612404e-01
-9.54853833e-01 -1.25651801e+00 7.66903684e-02 -2.59309977e-01
2.15681449e-01 -3.83057119e-03 -1.14802718e+00 2.55869180e-01
4.32679147e-01 1.34168282e-01 6.16222098e-02 -4.21169221e-01
-1.89269751e-01 -7.18091726e-02 -1.53289223e+00 5.26469409e-01
4.40937549e-01 -3.37012351e-01 -3.16412598e-01 1.83260128e-01
5.00575602e-01 -1.97177961e-01 -7.45821416e-01 4.44816649e-01
3.53514642e-01 -9.17248785e-01 1.31474054e+00 -2.74847358e-01
5.31916738e-01 -6.09769344e-01 -4.82652187e-02 -1.17852402e+00
-7.92652369e-02 -4.14378643e-01 2.83908874e-01 1.33499396e+00
3.87424618e-01 -5.13196230e-01 8.04070890e-01 5.31211495e-01
-1.06948018e-02 -2.62808830e-01 -6.72890663e-01 -5.85874021e-01
-1.23477645e-01 -1.47588238e-01 -8.45843926e-02 9.85623062e-01
-7.14595675e-01 1.29530892e-01 -4.41742420e-01 3.22789520e-01
9.12612021e-01 -1.91368893e-01 7.30015337e-01 -6.58402979e-01
-2.15915218e-01 -4.22015190e-01 -4.33440745e-01 -6.87046587e-01
-3.94274324e-01 -6.09020650e-01 3.12916666e-01 -1.38800216e+00
8.43068436e-02 -6.26713336e-01 -2.21316069e-01 6.45670831e-01
-3.59295905e-01 1.22519994e+00 3.47537659e-02 2.10003212e-01
-2.32238665e-01 1.04967989e-01 1.35269582e+00 -2.45148614e-01
-3.00909817e-01 -1.36361480e-01 -7.37807453e-01 3.97933155e-01
8.85316193e-01 -4.99756396e-01 -3.93478215e-01 -3.37061733e-01
-2.18879461e-01 -2.08073750e-01 7.01012790e-01 -1.23595464e+00
4.65251505e-02 1.36753201e-01 7.14993775e-01 -2.18264744e-01
4.03660208e-01 -8.44854891e-01 -1.12300748e-02 5.31246185e-01
-4.16045576e-01 3.70482281e-02 6.04793370e-01 6.08769119e-01
-2.96269715e-01 -2.64441937e-01 1.45225585e+00 3.59300710e-02
-9.13412988e-01 -1.67664900e-01 -5.31124294e-01 -1.60301104e-01
1.04706931e+00 -7.83876419e-01 -2.10703298e-01 -3.67128521e-01
-7.18969643e-01 -2.92132080e-01 7.44466007e-01 2.51224101e-01
7.45617211e-01 -8.84380639e-01 -7.42239833e-01 3.87964010e-01
-1.11965947e-01 3.11848223e-02 1.60260439e-01 9.25587654e-01
-7.50443041e-01 -1.31901965e-01 -4.50785965e-01 -6.80330753e-01
-1.79241478e+00 4.77869242e-01 4.07709539e-01 -6.73797056e-02
-7.71438599e-01 9.07428443e-01 2.94867784e-01 -4.96445186e-02
1.99368164e-01 -3.79224449e-01 4.54991572e-02 -2.58520752e-01
6.29090011e-01 3.16832155e-01 4.00341749e-01 -5.24865210e-01
-8.66563916e-02 7.04341412e-01 -4.18578148e-01 -2.46724337e-01
1.34801292e+00 -4.09036905e-01 -1.43464059e-01 -7.51372650e-02
1.21321595e+00 -6.71352372e-02 -1.47436500e+00 -2.87367314e-01
-1.72882333e-01 -6.80933952e-01 8.66537631e-01 -1.10087693e+00
-1.29387259e+00 7.27167010e-01 1.26959860e+00 2.14999765e-01
1.56949556e+00 -1.29958734e-01 6.76653981e-01 -1.05396681e-01
-7.78924301e-02 -9.45962250e-01 3.63807410e-01 -7.28029460e-02
7.28854239e-01 -1.47762179e+00 2.72798777e-01 -3.14529836e-01
-6.37993515e-01 7.55520046e-01 5.05424380e-01 -2.14609757e-01
4.43468541e-01 4.39832896e-01 1.93657294e-01 -1.70448497e-01
-3.37710321e-01 -2.08128437e-01 3.09401363e-01 9.61562693e-01
1.86788797e-01 -4.75861073e-01 2.32734606e-01 -6.11591518e-01
1.12511322e-01 -3.63044232e-01 6.95896029e-01 9.02000010e-01
-5.41252255e-01 -9.52268302e-01 -8.80575061e-01 4.74204689e-01
-4.59875584e-01 -2.50458568e-01 -3.82652700e-01 8.03954124e-01
2.13264704e-01 1.02910435e+00 1.69639394e-01 -4.62741181e-02
4.49123681e-02 -3.04738522e-01 6.12050176e-01 -4.03464645e-01
-6.46011889e-01 2.90468633e-01 -1.96749926e-01 -3.85662377e-01
-6.52348578e-01 -6.19088769e-01 -9.83127654e-01 -1.58677220e-01
-4.79086906e-01 -1.89289331e-01 9.52314079e-01 9.05453146e-01
2.69411933e-02 7.48324394e-01 3.52654189e-01 -9.92604077e-01
-8.65406543e-03 -9.51912344e-01 -5.55151105e-01 7.00401604e-01
6.74164772e-01 -4.41960573e-01 -2.45166212e-01 5.58871806e-01] | [10.983824729919434, -1.9942022562026978] |
416c7640-9a9d-4670-9501-66afb82a3511 | enhancing-deep-knowledge-tracing-with | 2302.07942 | null | https://arxiv.org/abs/2302.07942v1 | https://arxiv.org/pdf/2302.07942v1.pdf | Enhancing Deep Knowledge Tracing with Auxiliary Tasks | Knowledge tracing (KT) is the problem of predicting students' future performance based on their historical interactions with intelligent tutoring systems. Recent studies have applied multiple types of deep neural networks to solve the KT problem. However, there are two important factors in real-world educational data that are not well represented. First, most existing works augment input representations with the co-occurrence matrix of questions and knowledge components\footnote{\label{ft:kc}A KC is a generalization of everyday terms like concept, principle, fact, or skill.} (KCs) but fail to explicitly integrate such intrinsic relations into the final response prediction task. Second, the individualized historical performance of students has not been well captured. In this paper, we proposed \emph{AT-DKT} to improve the prediction performance of the original deep knowledge tracing model with two auxiliary learning tasks, i.e., \emph{question tagging (QT) prediction task} and \emph{individualized prior knowledge (IK) prediction task}. Specifically, the QT task helps learn better question representations by predicting whether questions contain specific KCs. The IK task captures students' global historical performance by progressively predicting student-level prior knowledge that is hidden in students' historical learning interactions. We conduct comprehensive experiments on three real-world educational datasets and compare the proposed approach to both deep sequential KT models and non-sequential models. Experimental results show that \emph{AT-DKT} outperforms all sequential models with more than 0.9\% improvements of AUC for all datasets, and is almost the second best compared to non-sequential models. Furthermore, we conduct both ablation studies and quantitative analysis to show the effectiveness of auxiliary tasks and the superior prediction outcomes of \emph{AT-DKT}. | ['Jian Weng', 'Weiqi Luo', 'Boyu Gao', 'Shuyan Huang', 'Jiahao Chen', 'Qiongqiong Liu', 'Zitao Liu'] | 2023-02-14 | null | null | null | null | ['auxiliary-learning', 'knowledge-tracing'] | ['methodology', 'miscellaneous'] | [ 6.14603758e-02 1.06704324e-01 -2.02595428e-01 -4.15224731e-01
-3.53687465e-01 -6.14387453e-01 3.44799012e-01 3.94820273e-01
-3.91765893e-01 8.16967547e-01 1.90621048e-01 -7.12998450e-01
-6.36034489e-01 -9.38704610e-01 -7.15387166e-01 -2.44056568e-01
3.12314957e-01 1.55108392e-01 5.26509166e-01 -4.83853102e-01
1.67400271e-01 2.04348192e-01 -1.70731008e+00 3.89785886e-01
1.47637188e+00 1.09698904e+00 1.70603096e-01 6.02778494e-01
-4.83727604e-01 1.30915165e+00 -8.62356961e-01 -7.32223749e-01
-2.60791898e-01 -5.61192095e-01 -1.20858753e+00 -6.62249446e-01
4.48367357e-01 -3.03067207e-01 -4.87532467e-01 8.38029921e-01
3.08063596e-01 4.08670992e-01 5.37774146e-01 -1.14516628e+00
-1.20010006e+00 8.53445172e-01 -2.25200742e-01 4.41752613e-01
4.68133390e-01 -1.02233008e-01 9.15423095e-01 -6.28372550e-01
2.15204030e-01 6.66927755e-01 1.00531435e+00 4.88993317e-01
-7.95721650e-01 -7.03544378e-01 3.58606100e-01 7.95614183e-01
-9.72702861e-01 1.58869445e-01 7.79680908e-01 -6.56933188e-01
5.01909316e-01 3.41318876e-01 7.73310125e-01 1.01051676e+00
-1.73426270e-02 1.14116335e+00 1.29077923e+00 -4.37831193e-01
-1.63148642e-01 4.06646729e-01 8.94780278e-01 7.81713605e-01
-8.51561427e-02 -5.34536801e-02 -6.75301313e-01 9.41079035e-02
5.69775045e-01 6.65013567e-02 -4.69473720e-01 1.65068641e-01
-9.41179752e-01 6.49816930e-01 3.22232932e-01 6.03649974e-01
-1.03564501e-01 -5.44218421e-02 8.10600743e-02 6.04106605e-01
2.54042804e-01 4.68475372e-01 -9.85736609e-01 -3.91910732e-01
-5.77496588e-01 2.09323585e-01 8.47650945e-01 9.81457770e-01
6.00792050e-01 3.01132984e-02 -4.75854456e-01 9.61620986e-01
-2.23762374e-02 1.99573666e-01 9.62320805e-01 -7.79211581e-01
3.26556176e-01 1.03054774e+00 -1.94192886e-01 -7.21100986e-01
-4.18232083e-01 -8.41798246e-01 -4.90836501e-01 -3.22830081e-01
5.74930012e-01 -3.78905654e-01 -7.90702462e-01 1.91464114e+00
1.86299041e-01 6.84816539e-01 2.24154145e-01 3.62703770e-01
1.70559466e+00 5.14224410e-01 2.77636796e-01 -2.19288498e-01
1.27064288e+00 -9.82350290e-01 -9.32299435e-01 1.16106868e-01
8.01512122e-01 -7.09757149e-01 1.19225800e+00 3.60688508e-01
-9.27190006e-01 -8.74249399e-01 -7.29828238e-01 -1.43461898e-01
-7.23201036e-01 2.32075915e-01 5.34556746e-01 6.30092084e-01
-9.23228085e-01 6.38121068e-01 -2.86161959e-01 6.04700781e-02
2.15824679e-01 3.79863828e-01 5.93433436e-03 2.96940744e-01
-1.62856340e+00 8.15234601e-01 2.19005570e-01 -2.93197073e-02
-5.71620405e-01 -1.31444204e+00 -6.25839412e-01 3.90321463e-01
2.99599856e-01 -4.83090550e-01 1.51736557e+00 -9.19564426e-01
-1.91873336e+00 5.27537823e-01 6.67274743e-02 -1.10856496e-01
4.25102741e-01 -5.02342403e-01 -4.38626707e-01 -1.41663387e-01
-9.42825377e-02 1.55969322e-01 1.57264426e-01 -1.00895667e+00
-7.94946194e-01 -2.19267070e-01 4.06912118e-01 3.18488121e-01
-9.61736619e-01 -4.33387488e-01 -3.58565897e-01 -6.02865040e-01
9.19790342e-02 -6.32907152e-01 1.68999657e-01 -6.56526327e-01
-1.69049446e-02 -1.00952935e+00 6.58995569e-01 -8.33436251e-01
1.81911922e+00 -1.71066797e+00 -1.00860745e-01 -3.38836014e-02
2.89226919e-01 6.88467741e-01 2.65701059e-02 2.96206802e-01
-1.15107410e-01 -9.04190633e-03 2.61708796e-01 1.31654903e-01
9.11095440e-02 2.13144109e-01 -3.93211722e-01 -2.67697006e-01
-1.95685714e-01 1.08778358e+00 -1.05079544e+00 -3.78612131e-01
6.49799258e-02 1.24119751e-01 -3.15026671e-01 4.97297227e-01
-3.13292950e-01 2.91023135e-01 -6.60124779e-01 3.84376794e-01
1.61972731e-01 -3.37575734e-01 7.69637153e-02 2.02017222e-02
-1.02929279e-01 4.19083059e-01 -8.99736047e-01 1.39216852e+00
-4.78181124e-01 6.58740342e-01 -4.60187405e-01 -1.13552201e+00
1.00331950e+00 5.45521140e-01 4.87653404e-01 -8.26997399e-01
7.57792816e-02 1.35925442e-01 2.95037925e-02 -8.43193471e-01
5.96763849e-01 5.54019399e-02 4.18238305e-02 3.71570289e-01
2.45240375e-01 2.85781682e-01 -1.82126820e-01 1.44003108e-01
1.09885967e+00 3.69885832e-01 -8.81291553e-02 -2.52832949e-01
8.43900621e-01 -1.24136381e-01 6.93813026e-01 7.56456614e-01
-3.19670945e-01 1.23431720e-01 5.68239391e-01 -4.13527489e-01
-3.06988865e-01 -8.81898165e-01 -9.54589620e-02 1.59766173e+00
1.19386889e-01 -3.40092272e-01 -7.53968000e-01 -9.46666420e-01
-4.77240272e-02 8.79604220e-01 -8.73940110e-01 -4.31536674e-01
-7.03616977e-01 -4.71969515e-01 7.36382902e-01 7.53871977e-01
6.46684706e-01 -1.08474874e+00 -3.60215902e-01 2.42448017e-01
-3.27503324e-01 -9.72729564e-01 -2.52295166e-01 1.48620456e-01
-6.73924923e-01 -1.09087718e+00 -6.25572383e-01 -9.25872266e-01
4.71049547e-01 1.90722197e-02 1.17353356e+00 3.85491520e-01
7.07118362e-02 7.35095322e-01 -5.10361075e-01 -5.32212079e-01
2.40612533e-02 1.31921828e-01 5.17558306e-02 -1.15555353e-01
6.59235597e-01 -5.75181067e-01 -6.94238603e-01 2.53708482e-01
-7.02553034e-01 1.14201687e-01 5.14839351e-01 8.12247813e-01
3.22306603e-01 1.95433944e-01 1.08273172e+00 -1.12046731e+00
9.10445213e-01 -6.06473863e-01 -2.47836426e-01 6.00510657e-01
-8.33037734e-01 -2.85582002e-02 7.59990215e-01 -7.82264709e-01
-1.36806071e+00 -4.37068373e-01 -2.90114582e-01 -3.58074158e-01
-1.84619620e-01 9.46968019e-01 3.72951590e-02 -5.61095029e-03
5.59303463e-01 5.05457580e-01 -4.58821625e-01 -6.53887451e-01
8.34402069e-02 4.52730149e-01 5.64514995e-01 -1.03736532e+00
4.75771278e-01 -3.69594991e-01 -2.97794044e-01 -4.18982267e-01
-1.46473467e+00 -2.99610794e-01 -6.05170846e-01 -5.17987728e-01
6.21614814e-01 -6.87008798e-01 -1.28160453e+00 5.71817040e-01
-8.16193461e-01 -5.25356174e-01 -3.92731667e-01 5.76330900e-01
-2.19944149e-01 2.78616458e-01 -6.68171167e-01 -7.52365291e-01
-2.76714593e-01 -9.36106801e-01 1.78074747e-01 8.49404514e-01
-4.22049798e-02 -1.35021222e+00 3.09310500e-02 8.99948180e-01
3.28894258e-01 1.03169791e-01 1.30301070e+00 -9.91503954e-01
-2.70301104e-01 -2.22460530e-03 1.57103747e-01 3.74968112e-01
7.98285529e-02 -1.13825805e-01 -1.00654793e+00 3.92757952e-02
4.25115488e-02 -5.70799112e-01 7.32865393e-01 1.10991411e-01
1.49916780e+00 -3.35363030e-01 -1.31781101e-01 2.98578531e-01
1.13906181e+00 4.43502694e-01 3.28859121e-01 3.88873577e-01
8.12125921e-01 7.48745859e-01 4.95967090e-01 4.07400988e-02
8.73771608e-01 4.77219790e-01 -3.91534083e-02 3.49934965e-01
-1.92485556e-01 -3.25654179e-01 4.12162900e-01 1.40750492e+00
-1.62554070e-01 -1.40540093e-01 -1.27082741e+00 8.56608510e-01
-1.91405499e+00 -7.77140498e-01 -5.71764410e-01 1.95249820e+00
1.37271035e+00 3.60919610e-02 -2.46385559e-01 4.79784012e-02
4.47610497e-01 -1.85302943e-01 -5.80147147e-01 -4.53956246e-01
1.79075524e-01 5.57242513e-01 -6.88986033e-02 4.38766688e-01
-6.76007032e-01 9.56421852e-01 5.21476030e+00 9.31902230e-01
-8.64171207e-01 1.65397152e-01 4.63277757e-01 2.73069233e-01
-4.96143818e-01 -1.95093483e-01 -1.09348381e+00 4.39555019e-01
1.18001056e+00 -3.65656346e-01 -5.34196794e-02 6.19434834e-01
-2.48953566e-01 -1.17005073e-01 -1.13313138e+00 4.82388437e-01
-5.43126948e-02 -1.14845037e+00 1.67550862e-01 -1.86385572e-01
1.10551620e+00 -5.69124043e-01 3.95476133e-01 9.93782640e-01
6.58880174e-01 -1.03737974e+00 4.22552288e-01 1.05423427e+00
4.44431007e-01 -6.81107283e-01 9.05325234e-01 6.07641935e-01
-1.08744693e+00 -2.36422911e-01 -7.63617381e-02 -1.83592767e-01
-4.02782470e-01 2.69883245e-01 -8.84076893e-01 8.94232750e-01
8.81692111e-01 3.92116159e-01 -7.22727239e-01 7.50287175e-01
-4.65756923e-01 1.19693005e+00 1.93895027e-02 -3.35461140e-01
2.62594849e-01 -1.18186818e-02 -4.02929150e-02 8.85518432e-01
4.41029668e-01 8.40419292e-01 8.49873200e-02 6.21044517e-01
-2.34637275e-01 2.22273543e-01 -2.01678887e-01 5.30857556e-02
6.42144382e-01 9.81845737e-01 -3.95596892e-01 -4.77998883e-01
-6.05412364e-01 5.62580824e-01 5.73335528e-01 4.66312230e-01
-7.43371010e-01 -3.99888277e-01 5.84858954e-01 1.52440056e-01
2.16527924e-01 8.87903646e-02 -4.07589346e-01 -9.03799951e-01
-4.83095497e-02 -7.29329824e-01 6.51078284e-01 -8.16806912e-01
-1.36744583e+00 4.23992813e-01 -1.40234053e-01 -9.88806903e-01
5.11997528e-02 -6.08951271e-01 -8.76784742e-01 9.87121582e-01
-1.64648128e+00 -1.04047215e+00 -6.33970439e-01 9.98272657e-01
4.27539170e-01 -2.75108784e-01 7.82922804e-01 2.92420775e-01
-7.21215427e-01 1.07508230e+00 3.82424086e-01 3.75720501e-01
6.41683459e-01 -1.47814953e+00 -2.76321143e-01 5.24569631e-01
-6.75203949e-02 6.90572143e-01 4.36092943e-01 -7.18737841e-01
-1.08067834e+00 -9.51346636e-01 1.02661920e+00 -7.79353261e-01
7.71061063e-01 1.81877181e-01 -1.41148520e+00 8.32799494e-01
2.93670297e-01 -2.96256840e-01 1.24613404e+00 3.66938829e-01
-1.08378038e-01 -2.91732490e-01 -7.42622316e-01 4.79079425e-01
7.31164992e-01 -4.46401179e-01 -1.19586432e+00 1.84043869e-01
9.42001164e-01 -5.36772132e-01 -1.48392582e+00 7.07821786e-01
4.83540952e-01 -9.23621714e-01 7.80804217e-01 -7.62229323e-01
5.67389667e-01 4.87078615e-02 3.33518267e-01 -1.12262380e+00
-3.08282107e-01 -2.14901403e-01 -5.09651542e-01 1.57969475e+00
3.02639604e-01 -3.92399400e-01 9.82600629e-01 7.01243818e-01
-4.49873835e-01 -1.12647152e+00 -6.41824543e-01 -5.06807208e-01
4.46805030e-01 -3.00632447e-01 7.99659789e-01 1.59489489e+00
1.11586697e-01 4.14895952e-01 -9.98334959e-02 1.75254524e-01
2.81498551e-01 4.41545725e-01 4.92650181e-01 -1.53315067e+00
-1.36698827e-01 -6.78635657e-01 -1.96597818e-02 -1.23847580e+00
2.45084316e-01 -8.38234365e-01 -2.54274994e-01 -1.56851149e+00
2.75925230e-02 -5.55911720e-01 -6.84862733e-01 6.66883290e-01
-8.54758501e-01 -3.82161826e-01 -2.43961334e-01 1.92983411e-02
-6.43373668e-01 8.93434823e-01 1.48399043e+00 1.45423219e-01
-2.81435102e-01 1.34749681e-01 -6.95856392e-01 7.70225942e-01
6.93178654e-01 -2.00020462e-01 -7.44655252e-01 -5.11100769e-01
2.75041282e-01 3.27661783e-01 2.12752014e-01 -1.02680540e+00
6.31067455e-01 -3.27099025e-01 6.12843275e-01 -5.73258877e-01
1.70430467e-01 -5.48436403e-01 -1.75586358e-01 3.28323007e-01
-4.81807888e-01 -6.67908415e-02 3.61372620e-01 4.73638296e-01
-2.81866610e-01 -4.28413481e-01 4.22219872e-01 -8.39536190e-02
-8.23134005e-01 2.27232724e-01 -3.01054209e-01 1.74429223e-01
9.30416048e-01 -2.24524081e-01 -6.06889784e-01 -3.05946916e-01
-8.65105450e-01 7.24639356e-01 -3.58911067e-01 6.78152382e-01
6.72020197e-01 -1.26278520e+00 -4.72061515e-01 1.73047483e-02
-9.98767912e-02 1.19785525e-01 6.10178888e-01 7.83156514e-01
-3.62303704e-01 5.59851885e-01 -2.71376699e-01 -4.31216508e-01
-1.29260147e+00 2.99622923e-01 3.81478548e-01 -5.95334709e-01
-3.75300467e-01 1.24240923e+00 1.31240621e-01 -9.13740873e-01
5.05313396e-01 -3.74134779e-01 -1.09086871e+00 2.12634698e-01
3.66480380e-01 3.36188763e-01 2.42687631e-02 -2.73799479e-01
1.09871253e-01 4.58199650e-01 -3.65231156e-01 2.49636218e-01
1.39962232e+00 1.00850113e-01 1.29565850e-01 5.89280963e-01
7.04949796e-01 -2.22670048e-01 -1.01888680e+00 -7.23042786e-01
1.82076424e-01 2.52604745e-02 -8.93742144e-02 -1.30979931e+00
-1.03099442e+00 9.70072865e-01 4.68252212e-01 1.27368942e-01
1.10971308e+00 -2.34619975e-01 8.19544137e-01 5.46848774e-01
1.57793224e-01 -1.09352005e+00 4.99696553e-01 9.00813758e-01
6.59487963e-01 -9.78601336e-01 -9.58643034e-02 -2.90232509e-01
-4.11257207e-01 1.09257936e+00 1.51589346e+00 3.69102985e-01
7.59765029e-01 -2.58354515e-01 1.42170191e-01 -2.43150845e-01
-8.48069251e-01 -1.82668582e-01 6.33486629e-01 -2.71178260e-02
6.55441463e-01 4.15400118e-02 -4.44172800e-01 1.45529795e+00
-6.09676242e-01 1.71710234e-02 4.46540415e-01 1.08417368e+00
-5.39238691e-01 -1.06817043e+00 -2.67820477e-01 6.98685110e-01
-4.30373877e-01 -1.30281255e-01 -5.21161556e-01 6.42556608e-01
4.41703975e-01 9.79703605e-01 -3.81209135e-01 -7.79716909e-01
6.13680661e-01 5.37062168e-01 3.83393705e-01 -6.18353367e-01
-1.30891621e+00 -5.86915970e-01 -2.18840271e-01 -1.11427404e-01
-2.42651880e-01 -4.93665248e-01 -1.27745461e+00 -3.31361562e-01
-4.14957196e-01 6.59995317e-01 1.16566777e-01 1.20278740e+00
8.07054061e-03 1.17143726e+00 3.61240655e-01 2.24816069e-01
-4.56430167e-01 -1.17070210e+00 -3.10531199e-01 3.33979368e-01
1.78659186e-01 -7.43280053e-01 -1.50313050e-01 1.48152530e-01] | [10.11495590209961, 7.153869152069092] |
093f747e-2218-409e-b73a-0ef6a8fad3f2 | investigating-lstms-for-joint-extraction-of | null | null | https://aclanthology.org/P16-1087 | https://aclanthology.org/P16-1087.pdf | Investigating LSTMs for Joint Extraction of Opinion Entities and Relations | null | ['Arzoo Katiyar', 'Claire Cardie'] | 2016-08-01 | null | null | null | acl-2016-8 | ['fine-grained-opinion-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.44317626953125, 3.5877878665924072] |
a409b00c-60ec-4452-a177-3bbf6715daae | t-cell-receptor-protein-sequences-and-sparse | 2304.13145 | null | https://arxiv.org/abs/2304.13145v1 | https://arxiv.org/pdf/2304.13145v1.pdf | T Cell Receptor Protein Sequences and Sparse Coding: A Novel Approach to Cancer Classification | Cancer is a complex disease characterized by uncontrolled cell growth and proliferation. T cell receptors (TCRs) are essential proteins for the adaptive immune system, and their specific recognition of antigens plays a crucial role in the immune response against diseases, including cancer. The diversity and specificity of TCRs make them ideal for targeting cancer cells, and recent advancements in sequencing technologies have enabled the comprehensive profiling of TCR repertoires. This has led to the discovery of TCRs with potent anti-cancer activity and the development of TCR-based immunotherapies. In this study, we investigate the use of sparse coding for the multi-class classification of TCR protein sequences with cancer categories as target labels. Sparse coding is a popular technique in machine learning that enables the representation of data with a set of informative features and can capture complex relationships between amino acids and identify subtle patterns in the sequence that might be missed by low-dimensional methods. We first compute the k-mers from the TCR sequences and then apply sparse coding to capture the essential features of the data. To improve the predictive performance of the final embeddings, we integrate domain knowledge regarding different types of cancer properties. We then train different machine learning (linear and non-linear) classifiers on the embeddings of TCR sequences for the purpose of supervised analysis. Our proposed embedding method on a benchmark dataset of TCR sequences significantly outperforms the baselines in terms of predictive performance, achieving an accuracy of 99.8\%. Our study highlights the potential of sparse coding for the analysis of TCR protein sequences in cancer research and other related fields. | ['Murray Patterson', 'Taslim Murad', 'Prakash Chourasia', 'Sarwan Ali', 'Zahra Tayebi'] | 2023-04-25 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [ 5.67733645e-01 -5.15963614e-01 -5.64350307e-01 -1.83490783e-01
-7.78674483e-01 -5.72798193e-01 4.06040579e-01 7.54170179e-01
-4.24272031e-01 7.51750827e-01 4.20546949e-01 -2.79223740e-01
-8.20541903e-02 -6.77357137e-01 -2.78743953e-01 -1.28459346e+00
1.34321228e-01 6.31421804e-01 1.16251968e-02 -1.97120860e-01
2.24153683e-01 6.87858403e-01 -1.26871943e+00 6.53529406e-01
7.81276405e-01 9.31522667e-01 2.13907301e-01 4.55244660e-01
-6.53812468e-01 3.27232987e-01 -3.08660358e-01 1.11641670e-02
-2.10304141e-01 -1.79199561e-01 -2.17866108e-01 -2.29075685e-01
-1.23329528e-01 4.98863429e-01 -2.45305702e-01 8.35463047e-01
6.09960854e-01 -1.76009744e-01 1.00978851e+00 -8.18597853e-01
-5.48736334e-01 -9.03230906e-02 -6.07433498e-01 2.44309768e-01
5.63738525e-01 -3.00108315e-03 1.01503849e+00 -8.34016621e-01
8.36342514e-01 1.14087093e+00 6.71676517e-01 4.65202928e-01
-1.46840644e+00 -4.81886506e-01 -1.89221397e-01 2.56039500e-01
-1.28196931e+00 -1.66007087e-01 7.61652350e-01 -7.10387290e-01
1.21649516e+00 4.68302757e-01 6.31983936e-01 1.09624910e+00
7.57511318e-01 5.63739598e-01 1.12577021e+00 -3.84724408e-01
2.32921928e-01 8.43063965e-02 3.38388562e-01 7.00156510e-01
7.67936334e-02 1.28065899e-01 -5.83387971e-01 -8.58103216e-01
2.29062289e-01 6.90016866e-01 -2.75735080e-01 -5.89385331e-01
-1.10271049e+00 1.30299306e+00 3.00619066e-01 5.71608365e-01
-4.33809996e-01 -1.75855696e-01 6.58080220e-01 1.70358926e-01
2.37135962e-01 1.72482714e-01 -3.66455823e-01 1.57914475e-01
-3.46155047e-01 -1.37065560e-01 4.09725189e-01 2.21116066e-01
6.56509280e-01 -2.13263422e-01 3.34835500e-02 1.14806795e+00
3.52360159e-02 3.22693527e-01 9.85739350e-01 -1.01812063e-02
-1.89619407e-01 9.89504099e-01 -3.54057789e-01 -9.92782176e-01
-5.37336648e-01 -3.91212404e-01 -9.17804182e-01 -3.03341299e-01
2.65471131e-01 3.10140342e-01 -6.75976336e-01 1.43820512e+00
4.72292513e-01 5.29414900e-02 1.98347017e-01 9.11985040e-01
5.35229385e-01 4.84905511e-01 2.03295618e-01 -1.78980857e-01
1.55535972e+00 -3.31823677e-01 -5.44950128e-01 1.85278982e-01
7.90798783e-01 -9.47702050e-01 8.92924249e-01 9.38461870e-02
4.15776335e-02 -1.38231158e-01 -7.23991513e-01 9.77797806e-02
-4.33655471e-01 -1.35143995e-01 1.11425483e+00 4.71571952e-01
-5.40246665e-01 2.31990099e-01 -6.95828378e-01 -4.66798544e-01
4.28362578e-01 5.25803626e-01 -7.58546650e-01 -5.43685794e-01
-1.14502299e+00 7.50932574e-01 1.63180903e-01 -3.09520245e-01
-5.21318436e-01 -7.14385033e-01 -8.43565166e-01 -1.63215593e-01
6.90121800e-02 -4.37084556e-01 3.82879883e-01 -7.56797791e-01
-1.08331490e+00 8.85035515e-01 -2.94950843e-01 -3.66825551e-01
-3.54901016e-01 1.40057385e-01 -4.03665304e-01 1.88272104e-01
-6.12305589e-02 2.39167973e-01 5.32458782e-01 -5.42010009e-01
-4.64648932e-01 -4.64248031e-01 -5.66683650e-01 -1.62502810e-01
-2.21612021e-01 -1.51374221e-01 2.85998639e-02 -6.30542994e-01
-3.97796258e-02 -1.11437893e+00 -4.34570074e-01 -1.29413322e-01
-1.72627583e-01 -2.57968575e-01 6.10945225e-01 -3.45197499e-01
8.36316943e-01 -2.38164949e+00 3.69846851e-01 2.68956661e-01
1.97687417e-01 3.11627090e-01 -2.95100868e-01 7.44603217e-01
-1.67759627e-01 -2.26243630e-01 -2.71037936e-01 4.34727579e-01
-3.65898162e-01 2.44452015e-01 -8.73879716e-02 7.98340380e-01
3.17681760e-01 8.50927591e-01 -8.29172552e-01 -2.73404121e-01
1.92593858e-01 9.49690640e-01 -3.96092415e-01 1.47643074e-01
-3.58011007e-01 3.19558561e-01 -7.53555238e-01 8.56193483e-01
3.92830640e-01 -3.36418569e-01 7.43805528e-01 -3.71163607e-01
1.67990595e-01 1.21674925e-01 -5.39176404e-01 1.43522549e+00
-2.25958630e-01 4.76705134e-01 -3.55672181e-01 -1.03445923e+00
9.53839302e-01 2.65240103e-01 7.06614494e-01 -6.87058210e-01
4.36334945e-02 1.94189057e-01 2.20034242e-01 -2.84674793e-01
-2.34634370e-01 -5.05433440e-01 -2.34657660e-01 4.68086034e-01
-1.80201694e-01 3.34298134e-01 -2.36581430e-01 1.14189491e-01
1.27965558e+00 -4.46126729e-01 5.46049833e-01 -2.13005289e-01
6.39229536e-01 3.00947785e-01 6.75941706e-01 -1.79602467e-02
-5.50853200e-02 2.46250793e-01 5.13124585e-01 -9.18326020e-01
-8.78319085e-01 -7.53247857e-01 -3.63321751e-01 9.46948349e-01
-2.03824893e-01 5.71828373e-02 -9.70680173e-03 -5.53535700e-01
4.78571773e-01 3.19333732e-01 -7.95004308e-01 -3.11710387e-01
-3.32723558e-01 -1.04452860e+00 4.55915362e-01 3.08462292e-01
-2.93820888e-01 -7.19849527e-01 -2.98561364e-01 2.95889795e-01
9.75775626e-03 -8.63352597e-01 -5.13854086e-01 7.21888483e-01
-7.45380402e-01 -1.50872159e+00 -5.76815546e-01 -1.05840659e+00
7.87950099e-01 2.55890995e-01 5.69495857e-01 -1.48286009e-02
-9.03002083e-01 1.59843385e-01 -2.39814997e-01 -1.10252894e-01
-3.35476816e-01 -9.55835357e-03 2.29335412e-01 1.52685538e-01
8.10945272e-01 -3.76462638e-01 -6.22658849e-01 9.38998759e-02
-1.01579392e+00 -2.74545044e-01 9.04551566e-01 1.28662550e+00
1.15523112e+00 -2.88930953e-01 6.79981768e-01 -1.25921953e+00
7.25479841e-01 -7.09180057e-01 -2.72497922e-01 2.56388962e-01
-2.92806566e-01 4.64207262e-01 7.59941101e-01 -6.71087623e-01
-7.18527555e-01 3.52814645e-01 -3.43031853e-01 -2.05905661e-01
-1.01273306e-01 5.54162860e-01 6.63912520e-02 -3.53230864e-01
4.95763958e-01 7.35618889e-01 4.30097491e-01 -4.10053730e-01
1.44865185e-01 7.53525496e-01 -3.95164778e-03 -3.93154413e-01
1.74926043e-01 4.93921041e-01 2.06432849e-01 -1.23742867e+00
-5.15375078e-01 -8.97410929e-01 -2.43668184e-01 3.02199781e-01
7.03975320e-01 -7.38798082e-01 -5.43892086e-01 9.73554403e-02
-6.80761456e-01 1.77758932e-01 1.30494192e-01 5.29359639e-01
-2.97270566e-01 6.16864800e-01 -8.88490319e-01 -4.17741001e-01
-4.61912692e-01 -1.16694832e+00 9.47632730e-01 -9.80600715e-02
-3.71587008e-01 -1.07183492e+00 6.65340900e-01 2.24685520e-01
3.42901707e-01 4.24893856e-01 1.71344721e+00 -1.03326046e+00
-2.41048232e-01 -3.95010889e-01 -9.36378762e-02 -7.13124648e-02
4.43293959e-01 -3.63180280e-01 -7.89891422e-01 -4.17407393e-01
-1.32044807e-01 -3.99804413e-01 1.02088714e+00 1.93175480e-01
8.09960723e-01 -2.51361914e-02 -7.76027679e-01 6.18457317e-01
1.66794062e+00 3.04978937e-01 4.56706822e-01 8.46509039e-02
5.26373148e-01 5.43180823e-01 5.54855883e-01 3.41933250e-01
-1.44355118e-01 6.88719034e-01 1.57493860e-01 -5.13244362e-04
1.97910085e-01 -1.15190111e-01 1.34007856e-01 8.10814023e-01
3.92408967e-01 1.11669995e-01 -7.58733511e-01 6.03894711e-01
-1.31423569e+00 -7.46470392e-01 -5.60073927e-03 2.08861375e+00
9.87601280e-01 -1.75636902e-01 1.20793364e-03 8.72854441e-02
5.16900718e-01 9.49125290e-02 -6.88399017e-01 -7.43353188e-01
-2.74724752e-01 4.87504274e-01 4.02862549e-01 2.98142940e-01
-6.90262914e-01 6.71095669e-01 6.37917948e+00 9.14681613e-01
-1.39110053e+00 -1.98592797e-01 5.25207281e-01 7.10754562e-03
-5.90423584e-01 -2.69263446e-01 -7.84017801e-01 3.69209945e-01
9.21548963e-01 1.03105512e-02 9.38336030e-02 5.84359169e-01
-8.42911005e-02 4.69383560e-02 -9.35568511e-01 9.45178509e-01
1.04475543e-02 -1.66364443e+00 9.69590470e-02 1.43793881e-01
5.90647936e-01 7.90533870e-02 2.31530890e-01 4.36872616e-02
5.78503273e-02 -1.20949411e+00 -3.89029235e-01 3.54735196e-01
1.07168126e+00 -1.00552845e+00 9.75380898e-01 2.62986690e-01
-9.66684759e-01 8.16206709e-02 -5.22715688e-01 2.59961903e-01
-1.04590148e-01 8.48487735e-01 -1.11808348e+00 1.64298937e-01
9.85178053e-02 6.99185967e-01 -1.83690816e-01 6.11417949e-01
1.86380595e-01 3.10577065e-01 -2.25817472e-01 -3.33220124e-01
2.32385233e-01 -1.85413912e-01 2.17693344e-01 1.25311065e+00
4.71251793e-02 1.90775588e-01 3.02822500e-01 2.53848374e-01
1.65837124e-01 4.34935451e-01 -7.02831745e-01 -4.84262675e-01
3.29184175e-01 1.13484943e+00 -5.40025055e-01 5.17443679e-02
-5.84678948e-01 7.99384952e-01 4.82868731e-01 1.28608510e-01
-4.30913359e-01 -2.87163794e-01 1.28668368e+00 7.74220377e-02
4.35053051e-01 1.50531698e-02 7.91661590e-02 -8.69412065e-01
-5.21576464e-01 -1.32632399e+00 5.55604875e-01 1.30730331e-01
-1.51071620e+00 5.97445071e-01 -6.24477983e-01 -9.02250350e-01
-1.66949872e-02 -6.96281612e-01 -1.86924413e-01 8.64583194e-01
-1.68731165e+00 -8.85194242e-01 1.82590365e-01 4.95029926e-01
4.36188042e-01 -3.81448954e-01 1.39471841e+00 1.01296924e-01
-4.64191467e-01 4.66375321e-01 6.81136012e-01 -1.27864152e-01
8.40565264e-01 -8.65514219e-01 -5.96331395e-02 -6.52442202e-02
6.14719689e-02 7.48354495e-01 5.60489178e-01 -6.94574535e-01
-1.81612539e+00 -9.39979613e-01 9.55519438e-01 -1.23484820e-01
6.96192265e-01 -3.65473896e-01 -9.80152845e-01 1.88060656e-01
-2.62972206e-01 3.98345850e-02 1.69836974e+00 1.31529272e-01
-5.37204802e-01 1.55223459e-01 -1.23687899e+00 4.14303631e-01
4.87322152e-01 -7.61580586e-01 -4.35341895e-01 5.02406657e-01
6.11340702e-01 7.83200562e-02 -9.63747740e-01 -2.15152763e-02
9.05882597e-01 -8.04762304e-01 1.10967004e+00 -8.13425839e-01
1.13138802e-01 -2.53339529e-01 -4.49432999e-01 -1.37308967e+00
-5.22944570e-01 -6.63963631e-02 1.28195807e-01 3.76983225e-01
1.48343071e-01 -8.18688154e-01 8.49042535e-01 4.95003127e-02
6.92936853e-02 -1.17919195e+00 -1.03960574e+00 -3.04162741e-01
4.47183400e-02 -4.01430055e-02 3.46007019e-01 8.88849139e-01
1.82499096e-01 2.66921222e-01 -1.41240567e-01 -1.87385589e-01
5.84838033e-01 5.51491618e-01 3.48570853e-01 -1.22239709e+00
-2.30016768e-01 -2.46917233e-01 -8.57518375e-01 -6.05060279e-01
3.10434729e-01 -1.13713443e+00 -4.40054357e-01 -1.13028860e+00
4.26792532e-01 -4.90168750e-01 -6.69461429e-01 3.19346845e-01
-3.73978615e-02 1.53122261e-01 -3.20659369e-01 1.99719191e-01
-3.56467038e-01 4.35847849e-01 8.99935424e-01 -3.69504362e-01
-3.86546217e-02 -2.39158213e-01 -8.12475622e-01 2.86125213e-01
6.94333673e-01 -5.77213407e-01 -1.27146766e-01 5.10264151e-02
7.14442357e-02 7.92976245e-02 -1.64106518e-01 -4.50741321e-01
-2.72822939e-02 -3.27515155e-01 5.44050217e-01 -5.45468688e-01
3.65587562e-01 -9.29999113e-01 4.32179421e-01 1.02033675e+00
-3.69752616e-01 3.23325545e-02 1.47496626e-01 9.71495807e-01
-3.84417355e-01 1.71126291e-01 7.58408546e-01 1.08536765e-01
-5.91392457e-01 1.74638659e-01 -8.29012632e-01 -1.98908985e-01
1.22055244e+00 -1.85843930e-01 -1.60339698e-01 1.69448510e-01
-4.69555646e-01 -1.71238095e-01 7.10987151e-01 2.81344622e-01
9.23323452e-01 -1.35356164e+00 -5.47277510e-01 5.14132023e-01
4.96409178e-01 -5.63830614e-01 3.08999062e-01 8.48779082e-01
-6.41903400e-01 8.45767677e-01 -3.74686807e-01 -6.47690713e-01
-1.56425810e+00 8.39030981e-01 4.41433564e-02 -4.25874114e-01
-6.68695331e-01 7.18108952e-01 3.88823509e-01 -3.33015054e-01
-2.61321869e-02 -9.37587954e-03 -3.86182874e-01 1.10732302e-01
6.28907621e-01 1.12999072e-02 5.09610362e-02 -7.20296741e-01
-8.57844532e-01 8.24979842e-01 -4.01450485e-01 6.37175441e-01
1.56050968e+00 4.31093544e-01 -1.81543663e-01 3.78631532e-01
1.56467307e+00 1.08456038e-01 -6.09791636e-01 -4.59301829e-01
1.84228837e-01 -4.45524663e-01 -1.20291747e-01 -6.57052994e-01
-8.27875614e-01 7.92086959e-01 7.31835783e-01 -4.35554385e-01
9.16362345e-01 2.58336961e-02 1.00374389e+00 2.29598761e-01
4.58787531e-01 -6.65192544e-01 6.47768378e-02 3.37015301e-01
4.22610641e-01 -1.19065130e+00 8.19849409e-03 -3.45236689e-01
-8.26959074e-01 1.00007141e+00 -4.81534079e-02 -8.27940032e-02
4.26543742e-01 3.24162483e-01 1.75067067e-01 -3.03200573e-01
-1.20419264e+00 -2.39773327e-03 2.23796755e-01 8.43150079e-01
8.59519124e-01 3.06242436e-01 -6.49631321e-01 4.65735108e-01
5.02303064e-01 -2.21536264e-01 -8.98187459e-02 8.46430838e-01
-6.91964269e-01 -1.62801886e+00 -3.01280439e-01 6.98778033e-01
-4.94389653e-01 -5.67239225e-02 -6.54960394e-01 3.85673553e-01
-1.03092305e-01 5.54704309e-01 -7.76341259e-02 -3.32930028e-01
7.25325570e-02 3.49609375e-01 3.40882897e-01 -6.25360191e-01
-4.82418269e-01 2.25749537e-01 -1.13791721e-02 -3.66994560e-01
-2.81069428e-02 -7.50881374e-01 -1.29358363e+00 -2.31886059e-01
-2.18632579e-01 3.25553685e-01 6.40341640e-01 6.22221828e-01
6.00385308e-01 4.51900601e-01 7.37032056e-01 -1.64035693e-01
-6.07327104e-01 -4.80212897e-01 -6.52595878e-01 5.11710584e-01
2.58618474e-01 -7.61807859e-01 -2.90045053e-01 -3.41148704e-01] | [4.8468098640441895, 5.58116340637207] |
dab2d799-bf87-41b9-bd0f-c2d5cb317f30 | impact-of-asr-on-alzheimers-disease-detection | 1904.01684 | null | https://arxiv.org/abs/1904.01684v3 | https://arxiv.org/pdf/1904.01684v3.pdf | Impact of ASR on Alzheimer's Disease Detection: All Errors are Equal, but Deletions are More Equal than Others | Automatic Speech Recognition (ASR) is a critical component of any fully-automated speech-based dementia detection model. However, despite years of speech recognition research, little is known about the impact of ASR accuracy on dementia detection. In this paper, we experiment with controlled amounts of artificially generated ASR errors and investigate their influence on dementia detection. We find that deletion errors affect detection performance the most, due to their impact on the features of syntactic complexity and discourse representation in speech. We show the trend to be generalisable across two different datasets for cognitive impairment detection. As a conclusion, we propose optimising the ASR to reflect a higher penalty for deletion errors in order to improve dementia detection performance. | ['Ksenia Shkaruta', 'Jekaterina Novikova', 'Aparna Balagopalan'] | 2019-04-02 | null | null | null | null | ['alzheimer-s-disease-detection'] | ['medical'] | [ 4.77382779e-01 2.15779260e-01 2.79773593e-01 -4.37416762e-01
-8.75736833e-01 -1.57200128e-01 7.67287433e-01 3.65513474e-01
-8.48176241e-01 4.80423868e-01 1.01005137e+00 -5.49296200e-01
-2.28292510e-01 -5.17969489e-01 -1.87195942e-01 -6.68862239e-02
1.23500608e-01 3.43126565e-01 3.57000768e-01 -2.05661699e-01
2.82995462e-01 3.44145238e-01 -1.37397957e+00 5.68885624e-01
9.37888861e-01 3.10941577e-01 3.84708524e-01 8.57977033e-01
1.58311665e-01 1.01141679e+00 -1.10513115e+00 -3.93614113e-01
-1.39065742e-01 -4.65685815e-01 -9.59188104e-01 3.24158847e-01
1.17477387e-01 -4.68061537e-01 -6.78921878e-01 9.07484174e-01
9.23320651e-01 -6.58055115e-03 5.78621149e-01 -4.25291628e-01
-3.40144604e-01 6.01774216e-01 3.22930008e-01 8.17820191e-01
5.74025512e-01 2.39917144e-01 7.37741828e-01 -3.91116560e-01
4.76780385e-01 1.51236534e+00 5.97137332e-01 5.61205566e-01
-1.14406610e+00 -1.52336463e-01 9.47928503e-02 3.84532541e-01
-1.04546964e+00 -1.14490175e+00 3.07191461e-01 -5.09172320e-01
1.26314020e+00 4.49279666e-01 6.18216336e-01 1.10153949e+00
4.60036332e-03 3.23705345e-01 1.12928009e+00 -6.98117316e-01
3.16203624e-01 2.21335679e-01 4.16182280e-01 2.30966002e-01
4.49575454e-01 -7.87596554e-02 -3.87555301e-01 -4.28902179e-01
3.14582497e-01 -4.17651951e-01 -4.32597399e-01 4.76698220e-01
-8.65102470e-01 7.79116571e-01 -5.21270046e-03 4.35018837e-01
-3.73453170e-01 -1.62358776e-01 7.57386386e-01 4.99988794e-01
6.16496265e-01 4.91005093e-01 -2.52994418e-01 -7.14957595e-01
-8.63460064e-01 3.71669412e-01 6.77812934e-01 3.00327063e-01
-3.27906370e-01 -1.95638537e-02 -5.97048342e-01 1.35244560e+00
3.98726135e-01 2.68157214e-01 9.34862792e-01 -7.68625975e-01
5.76479733e-01 7.63925493e-01 -1.38426632e-01 -7.45994925e-01
-4.45003599e-01 -2.73368448e-01 -1.91432923e-01 -3.57327750e-03
4.70443577e-01 -1.31566003e-01 -8.52828741e-01 1.44778788e+00
-2.59101242e-01 -4.78318542e-01 -1.49393650e-02 8.86027277e-01
5.23775995e-01 1.97299138e-01 3.67698878e-01 -1.85948536e-01
1.55446470e+00 -2.67245859e-01 -6.76507056e-01 -7.87890971e-01
1.04509330e+00 -7.91937172e-01 1.06737125e+00 3.50520127e-02
-8.08489382e-01 -1.33715272e-01 -8.35976422e-01 6.98595941e-02
-4.34942506e-02 2.55542845e-01 2.44863763e-01 1.29582977e+00
-1.08168459e+00 3.58455896e-01 -8.38168085e-01 -5.99770308e-01
2.60763943e-01 1.54626042e-01 -2.00199172e-01 -2.01227501e-01
-1.12859190e+00 1.41249645e+00 2.72228450e-01 -1.36307433e-01
-2.44833857e-01 -3.48517090e-01 -7.68201649e-01 -6.73631951e-02
5.29971011e-02 -5.01708448e-01 1.38191271e+00 -9.54085112e-01
-1.08163607e+00 9.66686845e-01 -2.91541398e-01 -7.48752236e-01
7.02691972e-01 -7.31454492e-02 -5.37287951e-01 1.21342704e-01
-1.51345134e-01 1.85491979e-01 4.80288476e-01 -7.66880155e-01
-4.33090955e-01 -7.75584519e-01 -1.24370202e-01 3.22045356e-01
-2.51617581e-01 6.79147303e-01 2.24684119e-01 -9.36197162e-01
-1.82900324e-01 -8.18045735e-01 1.02045573e-01 -2.25568965e-01
2.78304983e-02 -3.07797194e-01 3.35445702e-01 -1.37754667e+00
1.65263712e+00 -2.01201916e+00 -1.98842645e-01 -3.36574800e-02
1.33515522e-01 6.28628016e-01 1.43586114e-01 2.10663274e-01
-5.23170568e-02 2.16476053e-01 -4.07651424e-01 -3.69503647e-01
-2.90872529e-02 3.17092478e-01 5.26576489e-02 4.91132736e-01
3.26997757e-01 6.45432353e-01 -6.11325145e-01 -1.49807245e-01
3.55057031e-01 5.14716089e-01 -6.67349219e-01 7.27154091e-02
2.34379053e-01 5.17856777e-02 5.06010652e-02 2.71614760e-01
2.19028994e-01 1.10715933e-01 2.88658828e-01 3.40576291e-01
-1.46823242e-01 1.23279107e+00 -7.33194172e-01 7.31836140e-01
-2.56697327e-01 9.02032375e-01 1.11315258e-01 -1.08635259e+00
8.38099837e-01 4.93967861e-01 -8.47677737e-02 -9.11989033e-01
1.09928183e-01 2.95711577e-01 8.52364361e-01 -5.89702487e-01
1.84167191e-01 -2.31342018e-01 3.84907693e-01 1.28218010e-01
-4.13067907e-01 3.05193663e-01 2.13524744e-01 7.00646117e-02
1.50676370e+00 -8.09253275e-01 2.56624222e-01 -2.21385270e-01
4.86017674e-01 -1.54315382e-01 2.44614214e-01 8.77424896e-01
-5.43848753e-01 6.86820686e-01 4.05889004e-01 1.53401747e-01
-1.06936073e+00 -6.55113339e-01 -2.60872662e-01 8.66204560e-01
-7.47183800e-01 -6.43962622e-01 -9.15439546e-01 -4.29984778e-01
-1.40314639e-01 1.22877002e+00 -2.33548254e-01 -5.04647017e-01
-7.05539286e-01 -9.77692723e-01 8.01273644e-01 5.68094313e-01
3.04453880e-01 -1.10705566e+00 -6.38385296e-01 1.74434841e-01
-1.56267866e-01 -1.12376642e+00 -4.43627566e-01 4.20895889e-02
-7.43016303e-01 -1.03151381e+00 -1.03788984e+00 -6.36924446e-01
3.35865945e-01 1.16159558e-01 9.27477002e-01 2.73631275e-01
-2.60482669e-01 7.32574701e-01 -6.28246188e-01 -2.83750445e-01
-1.18599594e+00 1.32356156e-02 -1.34767726e-01 -4.87465918e-01
5.03219903e-01 -2.60949999e-01 -3.43097031e-01 7.36512691e-02
-6.12410247e-01 -3.84376198e-01 6.08194768e-01 5.17107725e-01
-1.93999276e-01 -1.77861169e-01 7.05606818e-01 -5.27280807e-01
1.22790182e+00 -2.95890540e-01 6.72623962e-02 1.89522892e-01
-7.43405819e-01 2.56399270e-02 1.94422051e-01 -4.03851002e-01
-9.16156590e-01 -3.75939637e-01 -5.75907230e-01 2.84043789e-01
-5.27478218e-01 5.26398301e-01 -1.43199250e-01 2.10865006e-01
6.51256025e-01 2.73075938e-01 4.31288451e-01 -5.02917171e-01
-3.15569639e-01 1.22337031e+00 -5.24610234e-03 -1.94357082e-01
-3.78791951e-02 -1.20539762e-01 -6.38500094e-01 -1.53170311e+00
-4.73932773e-01 -5.39481223e-01 -1.58093035e-01 -1.55341864e-01
7.15313017e-01 -9.33511794e-01 -1.21396549e-01 5.07996380e-01
-1.12838280e+00 -5.14568508e-01 2.83364151e-02 5.98475337e-01
-3.95980299e-01 5.10486007e-01 -5.02636909e-01 -1.01143706e+00
-3.74734342e-01 -1.25252163e+00 8.12400222e-01 -2.32894599e-01
-9.48036969e-01 -7.91053712e-01 -4.73688692e-02 6.62406564e-01
4.90899891e-01 -4.18738633e-01 1.03108585e+00 -1.05586195e+00
1.25013337e-01 -1.60616919e-01 -2.34573260e-01 6.49263918e-01
2.97078490e-01 -4.17563498e-01 -5.47388434e-01 -7.71167129e-02
1.60356507e-01 1.83851227e-01 7.74940670e-01 3.30753446e-01
6.05837107e-01 -6.29063487e-01 -5.44718020e-02 -3.49374086e-01
7.95287371e-01 3.02696913e-01 9.44640696e-01 5.94334304e-01
2.30663866e-01 6.64194345e-01 -3.14086825e-02 3.41146141e-01
3.43822271e-01 9.84655082e-01 -3.47459614e-01 4.82176870e-01
-5.76390862e-01 1.71914071e-01 7.13488281e-01 6.82826817e-01
1.17074259e-01 -2.60374635e-01 -1.34999144e+00 6.94801152e-01
-1.32526803e+00 -8.65688324e-01 -2.08063915e-01 2.28499913e+00
7.89002419e-01 4.30011809e-01 6.34918749e-01 6.01030350e-01
8.25751722e-01 1.60026886e-02 3.07237520e-03 -4.16775018e-01
-2.02023253e-01 4.94340844e-02 4.73429114e-01 4.90383029e-01
-6.11548960e-01 4.21279043e-01 7.06486225e+00 3.05793375e-01
-9.02498901e-01 7.82332122e-02 6.51921451e-01 -2.09640637e-01
-1.10492446e-01 -4.30958033e-01 -6.41431451e-01 8.14436913e-01
1.43078959e+00 -1.43075272e-01 3.73480201e-01 4.62359816e-01
6.47664726e-01 -2.03312993e-01 -8.39176536e-01 7.21076250e-01
1.88508868e-01 -7.08585858e-01 8.63579065e-02 -2.98237801e-03
-1.58323854e-01 9.13314745e-02 -2.87218720e-01 1.98434189e-01
-1.16794318e-01 -9.69983697e-01 8.11394453e-01 3.24175954e-01
3.30137432e-01 -5.53975582e-01 8.02780271e-01 2.18393698e-01
-6.05085433e-01 -3.00467283e-01 -8.35873885e-04 -3.71554345e-01
2.31339514e-01 4.94372606e-01 -1.28984046e+00 -2.24641785e-01
3.69559467e-01 1.22668751e-01 -8.01501513e-01 1.09275186e+00
-3.81399468e-02 1.25201690e+00 -3.28385204e-01 -1.89141139e-01
-1.57696187e-01 3.34195137e-01 7.64394104e-01 1.58180845e+00
3.23428810e-01 1.67057678e-01 -2.46833354e-01 5.28028786e-01
2.18370348e-01 1.41687572e-01 -2.82470554e-01 -4.16374624e-01
7.12662995e-01 3.57201993e-01 -7.04342067e-01 -3.27867657e-01
-5.56762338e-01 9.16910529e-01 2.00725347e-01 1.53512321e-02
-3.05695355e-01 -9.76745710e-02 7.69237161e-01 7.04551876e-01
7.35958591e-02 -2.57748961e-01 -4.85830396e-01 -7.49143779e-01
2.90242642e-01 -1.25837719e+00 3.71840179e-01 -4.52057421e-01
-9.35708225e-01 1.64332569e-01 -1.71202183e-01 -5.33652008e-01
-1.61710188e-01 -5.18739164e-01 -4.69572157e-01 1.00437057e+00
-1.00387192e+00 -4.01515126e-01 -3.89519846e-03 1.27851859e-01
8.75309706e-01 -1.65968999e-01 7.10914195e-01 4.27798718e-01
-5.11635661e-01 5.56904972e-01 -2.52469063e-01 2.29478851e-01
5.63552618e-01 -9.84191239e-01 6.05432689e-01 7.85439253e-01
-2.65342861e-01 6.98472440e-01 8.70641887e-01 -1.09086347e+00
-7.35367596e-01 -1.01653802e+00 1.17816889e+00 -5.38850248e-01
5.66256940e-01 -6.65758029e-02 -9.96079862e-01 4.27115709e-01
-2.05576092e-01 -6.83934569e-01 5.07925034e-01 1.94538627e-02
-1.48486778e-01 3.10994834e-01 -1.27179909e+00 6.31354928e-01
1.17737913e+00 -7.78543353e-01 -9.81446564e-01 1.85206085e-01
7.51078248e-01 -1.60466079e-02 -9.04580295e-01 4.65498179e-01
3.62899154e-01 -8.16623271e-01 9.70909059e-01 -4.07700658e-01
2.55460203e-01 9.44408998e-02 -2.21545801e-01 -1.33714461e+00
-4.50774014e-01 -2.07986787e-01 5.58023527e-02 1.17145205e+00
3.94650906e-01 -7.83829391e-01 5.01228511e-01 1.02527583e+00
-2.64466077e-01 -3.64026725e-01 -1.20937455e+00 -8.00688088e-01
-4.94691208e-02 -4.87416208e-01 7.40628615e-02 5.42704523e-01
1.56656414e-01 3.99373621e-01 7.34853819e-02 1.77768141e-01
2.16185272e-01 -1.03193927e+00 1.71621963e-02 -1.23871446e+00
-1.36308238e-01 -7.09612668e-01 -7.54645705e-01 -4.85355079e-01
-5.38785234e-02 -7.47081876e-01 -1.03928849e-01 -1.64743257e+00
1.53764784e-01 -9.71848443e-02 1.66385785e-01 2.95335680e-01
-4.12776262e-01 -2.92127401e-01 3.57207030e-01 3.13320421e-02
-2.46356621e-01 4.67042089e-01 5.65456450e-01 -3.92512158e-02
-2.65363038e-01 2.02730626e-01 -8.59299362e-01 4.84778821e-01
1.06425178e+00 -3.86909693e-01 -4.28474188e-01 -3.92471462e-01
-1.70993239e-01 -1.61810413e-01 4.58343923e-01 -1.04594922e+00
-9.47284847e-02 2.33589873e-01 2.11770684e-01 -1.46550804e-01
1.76165074e-01 -2.48527095e-01 -1.92000851e-01 5.94217658e-01
-5.86637497e-01 8.44024345e-02 2.34335512e-01 3.05886865e-01
1.42641932e-01 -5.31801105e-01 7.05192685e-01 -2.26629823e-01
-3.19440842e-01 -5.24449229e-01 -1.10477984e+00 3.28968078e-01
4.30660814e-01 -2.24333599e-01 -3.26547891e-01 -3.56375039e-01
-7.37125874e-01 -2.50989109e-01 3.64247084e-01 5.86695015e-01
4.77683425e-01 -7.89960146e-01 -8.02163243e-01 3.08361650e-02
8.26885179e-02 -4.73075479e-01 -4.90286984e-02 8.78189743e-01
-2.46088624e-01 7.08730876e-01 1.94877058e-01 -5.21280356e-02
-1.68531406e+00 9.09968838e-02 4.64825660e-01 -2.59041544e-02
-7.97474027e-01 5.13658345e-01 -4.21258956e-01 -1.31392792e-01
5.55972993e-01 -3.09868276e-01 -3.11626554e-01 6.11282438e-02
9.40634966e-01 8.89936507e-01 5.47427952e-01 -5.59156299e-01
-5.23015082e-01 -3.27109456e-01 -3.77260327e-01 -3.72234166e-01
1.22866094e+00 -2.08611876e-01 5.18248118e-02 4.44183350e-01
8.86140287e-01 -2.26274669e-01 -7.15036273e-01 1.61414921e-01
5.14391482e-01 -1.67348981e-01 2.14627743e-01 -9.51544404e-01
-3.63943398e-01 3.70766908e-01 9.85706031e-01 3.96052390e-01
6.41109347e-01 2.43688866e-01 5.13511300e-01 4.17831481e-01
5.17725050e-02 -1.24649453e+00 -2.40614355e-01 5.31821430e-01
1.06564975e+00 -9.50389385e-01 -8.22903588e-02 -2.92668700e-01
-5.58091998e-01 7.69989550e-01 3.05369020e-01 1.26695316e-02
3.15464377e-01 2.11712196e-01 -1.56781495e-01 -3.49817909e-02
-5.86678803e-01 -2.87054837e-01 1.69601887e-01 7.45357156e-01
8.48744988e-01 3.27251226e-01 -8.47137332e-01 7.24560380e-01
-4.58946496e-01 6.34623989e-02 5.13693452e-01 8.89405191e-01
-7.79429317e-01 -9.49783623e-01 -6.25112653e-01 9.45760846e-01
-6.30645275e-01 -1.70777544e-01 -8.34787130e-01 4.15718228e-01
-2.33427092e-01 1.10213935e+00 2.63008147e-01 -8.23139250e-02
6.35246038e-01 5.13058901e-01 4.33496058e-01 -8.89728010e-01
-5.13166726e-01 -3.61823142e-01 9.41712677e-01 -1.78316817e-01
-2.03405291e-01 -1.05139327e+00 -9.62276518e-01 -1.91696271e-01
-1.05381094e-01 -2.64714926e-01 3.74811083e-01 1.21322668e+00
4.91851687e-01 7.11025536e-01 -2.05387697e-01 -2.09442899e-01
-8.75620604e-01 -1.43315327e+00 -2.00689763e-01 4.49395865e-01
3.01099211e-01 -5.39299488e-01 -4.51891869e-01 -1.62673339e-01] | [13.963807106018066, 5.457158088684082] |
9db86e7f-bb4b-4d6b-a1a5-c18854b9f5f3 | examining-temporalities-on-stance-detection | 2304.04806 | null | https://arxiv.org/abs/2304.04806v2 | https://arxiv.org/pdf/2304.04806v2.pdf | Examining Temporalities on Stance Detection towards COVID-19 Vaccination | Previous studies have highlighted the importance of vaccination as an effective strategy to control the transmission of the COVID-19 virus. It is crucial for policymakers to have a comprehensive understanding of the public's stance towards vaccination on a large scale. However, attitudes towards COVID-19 vaccination, such as pro-vaccine or vaccine hesitancy, have evolved over time on social media. Thus, it is necessary to account for possible temporal shifts when analysing these stances. This study aims to examine the impact of temporal concept drift on stance detection towards COVID-19 vaccination on Twitter. To this end, we evaluate a range of transformer-based models using chronological (split the training, validation and testing sets in the order of time) and random splits (randomly split these three sets) of social media data. Our findings demonstrate significant discrepancies in model performance when comparing random and chronological splits across all monolingual and multilingual datasets. Chronological splits significantly reduce the accuracy of stance classification. Therefore, real-world stance detection approaches need to be further refined to incorporate temporal factors as a key consideration. | ['Xingyi Song', 'Kalina Bontcheva', 'Mali Jin', 'Yida Mu'] | 2023-04-10 | null | null | null | null | ['stance-detection'] | ['natural-language-processing'] | [ 7.67584816e-02 -5.95881268e-02 -5.50779998e-01 -4.11377549e-01
-4.10309941e-01 -7.89348662e-01 1.07889581e+00 9.50832665e-01
-7.96152890e-01 5.88417590e-01 6.28147840e-01 -7.81937480e-01
9.28633958e-02 -8.71141493e-01 -6.03005290e-01 -4.84332234e-01
-1.10451784e-02 5.85563600e-01 2.99523890e-01 -5.35821259e-01
1.55633703e-01 -1.19702622e-01 -1.31001651e+00 2.82260358e-01
1.00019288e+00 1.32627949e-01 8.02585855e-03 3.69151533e-01
-2.08722930e-02 7.11162925e-01 -7.04809308e-01 -5.43416083e-01
-1.45521313e-01 -4.02759075e-01 -5.72706938e-01 -3.36526215e-01
1.35655984e-01 -5.30508086e-02 4.11551535e-01 9.75708902e-01
4.47322458e-01 -5.45148492e-01 6.01178944e-01 -1.03490937e+00
-1.70937449e-01 8.79676640e-01 -5.26500762e-01 5.74776173e-01
3.84862602e-01 5.71292341e-02 9.31216121e-01 -3.21807563e-01
1.24155498e+00 1.24477077e+00 9.71842885e-01 1.64330646e-01
-1.29072642e+00 -7.65953183e-01 3.99662316e-01 2.02614635e-01
-7.28935719e-01 -2.40840599e-01 6.03344738e-01 -1.12839127e+00
5.59628785e-01 2.13394135e-01 1.07884514e+00 1.21987259e+00
4.35647100e-01 2.64894128e-01 1.75265121e+00 -2.57279992e-01
-1.63078055e-01 3.18568259e-01 3.45694393e-01 2.06812009e-01
6.12735629e-01 -4.46605906e-02 -3.21309954e-01 -5.70368528e-01
-1.74321339e-01 -9.34842229e-02 -1.55955240e-01 1.07589774e-01
-1.10882878e+00 1.33100903e+00 2.52587497e-01 5.03327250e-01
-6.21476889e-01 -5.59402406e-01 6.55313551e-01 4.38882977e-01
1.09139860e+00 2.61820257e-01 -4.58421856e-01 -3.02412629e-01
-9.93722975e-01 4.41615134e-01 6.11694217e-01 2.69969441e-02
2.62193531e-01 -2.73096919e-01 6.46429956e-02 7.49989569e-01
2.12407053e-01 7.23228812e-01 1.45927012e-01 -3.02650869e-01
2.79023647e-01 5.41546285e-01 8.58050957e-02 -1.16739249e+00
-6.20603204e-01 -5.33250690e-01 -3.68571311e-01 -2.27822930e-01
4.06290412e-01 -3.25855017e-01 -6.68980658e-01 2.03529048e+00
6.55258298e-01 -1.57387599e-01 -2.30858371e-01 6.65163875e-01
5.97200274e-01 3.95629972e-01 4.08001423e-01 -5.64421773e-01
1.68910754e+00 -1.01459064e-01 -6.58619761e-01 -1.99599475e-01
6.73027396e-01 -1.14675355e+00 6.38992012e-01 5.39001599e-02
-8.43114853e-01 -1.43433198e-01 -8.69444132e-01 4.53164041e-01
-5.09442449e-01 -5.33998311e-01 2.00458989e-01 7.87529826e-01
-7.81490564e-01 2.35900909e-01 -8.11039388e-01 -6.72338903e-01
7.22326688e-04 -3.20675336e-02 4.78245690e-02 7.27252066e-02
-1.72122610e+00 1.23259079e+00 1.52691290e-01 3.88984494e-02
-3.42287034e-01 -6.57334089e-01 -5.30389845e-01 -7.79402316e-01
6.76302016e-02 -3.85443032e-01 1.08686566e+00 -9.30636704e-01
-6.05685413e-01 1.17483103e+00 -2.29707927e-01 -4.73734409e-01
5.68088233e-01 1.79600105e-01 -7.05962598e-01 -1.31292939e-01
5.16492546e-01 3.98799181e-01 6.46530569e-01 -9.19098973e-01
-7.01746643e-01 -4.89187986e-01 4.81850728e-02 -1.11994922e-01
5.16724959e-02 6.07983053e-01 4.29131985e-01 -5.87085545e-01
-1.77812397e-01 -1.19964266e+00 5.55265546e-02 -7.67461121e-01
2.26442978e-01 -4.26492274e-01 6.24998569e-01 -9.26834822e-01
1.40683341e+00 -1.62261772e+00 -1.17716715e-01 -3.98752540e-02
2.85158515e-01 3.09102088e-01 1.93234861e-01 7.18413472e-01
5.87758161e-02 2.16221094e-01 2.76696980e-01 1.59810722e-01
-3.54732573e-01 2.01643810e-01 -3.01095486e-01 7.93320775e-01
8.83396417e-02 7.59686291e-01 -1.10203123e+00 -4.78227228e-01
-1.11716278e-01 3.97593468e-01 -7.76842475e-01 -2.96458393e-01
-5.17487824e-01 4.38658714e-01 -2.00437218e-01 3.54062825e-01
6.46150768e-01 -2.52871305e-01 7.21228540e-01 -5.87632209e-02
-7.15647399e-01 8.61658931e-01 -1.46191522e-01 6.45167947e-01
-5.25849462e-02 9.05118763e-01 1.46254778e-01 -6.28535926e-01
6.05875373e-01 5.59137285e-01 6.52107716e-01 -8.75452816e-01
3.37591231e-01 2.94679374e-01 6.17534459e-01 -3.23297411e-01
5.08974254e-01 -4.51151937e-01 -6.74857423e-02 7.55643249e-01
-7.06554830e-01 3.49779636e-01 3.13785583e-01 8.24486688e-02
5.78380406e-01 -2.08942175e-01 1.85465872e-01 -5.93264580e-01
3.29987526e-01 5.96755981e-01 7.18042254e-01 3.18985671e-01
-2.26388574e-01 9.59369019e-02 6.80398822e-01 -3.08879256e-01
-9.67543066e-01 -5.69173932e-01 -5.00488400e-01 1.03360224e+00
-3.68178427e-01 -3.16000789e-01 -6.74588025e-01 -4.93617833e-01
7.49145374e-02 7.68541217e-01 -9.11802828e-01 1.99911132e-01
-6.97043180e-01 -1.25741029e+00 3.79901081e-01 -2.42074812e-03
3.07940006e-01 -5.68458676e-01 -8.39763224e-01 5.07018864e-01
-8.66141737e-01 -8.68417025e-01 -9.48167369e-02 -1.25989869e-01
-6.54303074e-01 -1.43876195e+00 -5.17879069e-01 -5.29069602e-01
4.88836080e-01 4.05108750e-01 9.23512101e-01 1.62767500e-01
3.73310953e-01 1.07869722e-01 -3.73334467e-01 -6.78589702e-01
-9.70238149e-01 2.61445284e-01 -1.55863706e-02 -4.10388917e-01
7.39971697e-01 -1.38608918e-01 -7.43953884e-01 3.27160269e-01
-7.83167303e-01 -3.48835369e-03 5.70225380e-02 5.54673672e-01
-1.02916256e-01 -2.00028718e-01 7.54294157e-01 -1.17132294e+00
9.17542577e-01 -1.12624776e+00 -5.08641005e-01 7.37336092e-03
-8.46724689e-01 -3.17276061e-01 9.64884683e-02 -6.16901278e-01
-9.34027016e-01 -1.09960997e+00 -3.20483923e-01 5.28030515e-01
2.71161079e-01 1.12569427e+00 7.06194997e-01 4.39666748e-01
7.04314411e-01 -2.24329412e-01 4.99877393e-01 -3.02918077e-01
-4.52951714e-03 1.05709004e+00 -1.30874157e-01 -3.45325857e-01
5.25639176e-01 6.17651224e-01 -6.36269808e-01 -9.48569298e-01
-7.98648238e-01 -4.86687392e-01 -2.71997184e-01 -5.19803286e-01
9.39159214e-01 -1.16001821e+00 -5.71357191e-01 6.01806462e-01
-9.97067690e-01 -4.10092473e-01 5.64197123e-01 4.62326974e-01
1.73788145e-01 1.46225080e-01 -5.77744067e-01 -4.96001780e-01
-4.57891881e-01 -1.08637655e+00 5.16798615e-01 -2.33723372e-01
-1.07563603e+00 -1.33759522e+00 6.80741370e-01 6.13375604e-01
4.93011326e-01 7.31506824e-01 1.05918813e+00 -7.62073755e-01
-2.23316271e-02 -2.18502097e-02 6.34348318e-02 -5.31099327e-02
4.70840245e-01 2.48649150e-01 -5.66755772e-01 -6.20351434e-01
2.81918734e-01 -1.71615362e-01 6.58024013e-01 4.36321974e-01
-1.81850776e-01 -5.43437362e-01 -3.50133181e-01 -3.63813549e-01
1.18658340e+00 2.63059944e-01 7.74265826e-02 6.57956839e-01
8.76531526e-02 9.30329621e-01 7.34344482e-01 1.92309707e-01
8.20923388e-01 6.53442562e-01 3.07313856e-02 1.33792624e-01
-5.37813036e-03 -2.17804372e-01 5.58405697e-01 1.20677745e+00
3.18735354e-02 1.43115921e-02 -1.32875299e+00 7.56057620e-01
-1.31933975e+00 -1.05818570e+00 -4.96426970e-01 2.16137171e+00
1.10104954e+00 5.75432122e-01 5.83510280e-01 1.18879646e-01
6.71565950e-01 5.36136508e-01 -6.67621568e-02 -5.70851624e-01
-2.30647042e-01 -2.68957913e-01 3.77871156e-01 6.68515325e-01
-8.96832168e-01 4.37920481e-01 6.49195766e+00 2.22313017e-01
-1.64435112e+00 4.97184694e-01 2.76847988e-01 4.77451384e-02
-6.70255840e-01 -8.45190361e-02 -6.89528465e-01 7.07052946e-01
1.18621171e+00 -2.59424835e-01 -6.44441620e-02 4.33678329e-02
6.54571354e-01 -3.89543414e-01 -5.70276499e-01 1.45728420e-02
1.62037015e-01 -1.16431320e+00 -5.48537672e-01 1.09326988e-01
8.30701530e-01 5.65516651e-01 5.17614111e-02 1.48073792e-01
4.40427125e-01 -4.91545975e-01 9.27002966e-01 -3.60952839e-02
5.50614715e-01 -2.43768901e-01 7.85727143e-01 3.76995921e-01
-8.07731271e-01 2.09677309e-01 1.81836233e-01 -3.01400959e-01
3.61858428e-01 5.00804603e-01 -1.18344712e+00 2.31108010e-01
4.12282825e-01 7.23881841e-01 -5.62055469e-01 4.80032355e-01
-3.31444517e-02 1.11577129e+00 -1.95506632e-01 -1.62766859e-01
4.15891945e-01 -9.57687721e-02 5.01910210e-01 9.93611455e-01
-1.41590476e-01 -8.19318593e-02 -1.77700356e-01 2.33487725e-01
4.23123688e-01 -4.31042835e-02 -4.98326898e-01 -4.24398780e-01
4.86234933e-01 6.29771173e-01 -7.95745432e-01 -2.95573771e-01
-7.52039731e-01 -2.07634587e-02 1.13817066e-01 1.82237133e-01
-9.17538524e-01 4.98186409e-01 7.94378042e-01 8.55672002e-01
9.15585309e-02 -1.68273389e-01 -1.07231155e-01 -7.41689861e-01
-2.68478841e-01 -1.27597952e+00 6.56399906e-01 -2.39332438e-01
-9.32156682e-01 2.85916060e-01 2.91915745e-01 -9.02067602e-01
-4.12752241e-01 -1.64883658e-01 -2.72118121e-01 7.45034277e-01
-1.46442974e+00 -1.15697598e+00 7.01921359e-02 5.75234108e-02
4.06657577e-01 4.32553321e-01 4.48618859e-01 2.96504289e-01
-3.03387046e-01 1.13221318e-01 1.36485128e-02 7.67587796e-02
8.12014699e-01 -7.06872761e-01 4.41982895e-01 4.68468875e-01
-4.71928090e-01 8.53651464e-01 1.26834679e+00 -1.23921001e+00
-6.79138601e-01 -9.37241554e-01 1.53105199e+00 -6.15438759e-01
1.02551150e+00 -2.88096279e-01 -8.21155190e-01 8.07519019e-01
5.07945776e-01 -1.23970175e+00 9.16008830e-01 5.58651805e-01
-4.86636639e-01 1.92339435e-01 -1.05566621e+00 8.05692375e-01
7.16329753e-01 -5.93667388e-01 -9.54014897e-01 3.49485308e-01
6.42692208e-01 -2.02081010e-01 -1.04671192e+00 5.07979393e-01
8.03119540e-01 -8.74139845e-01 7.10901260e-01 -4.46346194e-01
3.83500636e-01 -1.20029829e-01 1.66913956e-01 -1.12455118e+00
-2.28952274e-01 -2.42259666e-01 5.65413833e-01 9.38322544e-01
4.84116137e-01 -1.13981223e+00 5.13635159e-01 1.19079158e-01
3.97238314e-01 -7.54242063e-01 -7.31559277e-01 -3.42105955e-01
4.45107996e-01 -2.77334005e-01 5.32181561e-01 1.23028731e+00
3.70378830e-02 3.42262238e-01 -1.49696529e-01 9.03218612e-02
4.00554746e-01 2.77517885e-01 5.28820992e-01 -1.41844976e+00
2.27628529e-01 -5.40652633e-01 -1.66458592e-01 -2.73213297e-01
-5.46475574e-02 -7.06218183e-01 -5.30517220e-01 -1.47723150e+00
3.28025728e-01 -5.99929094e-01 4.83471639e-02 1.18057214e-01
-3.03326607e-01 6.29430264e-02 1.23166591e-01 4.76035386e-01
-2.12301135e-01 1.62354335e-01 9.84527051e-01 -2.42871642e-01
-1.82162479e-01 1.33539155e-01 -6.83254182e-01 5.73890686e-01
9.53068435e-01 -6.82156563e-01 -4.09975916e-01 -2.57520884e-01
9.01172340e-01 -8.26463103e-02 1.49716243e-01 -3.16713154e-01
3.49297039e-02 -3.61739367e-01 -1.83839828e-01 -8.80282700e-01
-2.89338194e-02 -4.09593880e-01 5.84392250e-01 1.09112859e+00
-1.69530377e-01 5.52695811e-01 1.26634300e-01 2.56455660e-01
-1.92819715e-01 1.25817016e-01 5.37440240e-01 1.32725313e-01
-7.62484446e-02 -2.20748201e-01 -8.15168917e-01 4.51571763e-01
8.83597732e-01 -1.11693703e-01 -7.37764776e-01 -2.68657267e-01
-4.81119245e-01 3.13352168e-01 7.60539711e-01 7.68542469e-01
1.09478556e-01 -1.10222483e+00 -1.04589224e+00 -3.73105332e-02
4.38130163e-02 -6.04421616e-01 1.51162744e-01 1.50711071e+00
-5.21297157e-01 7.53546059e-01 -4.56327274e-02 -6.93985939e-01
-1.56488883e+00 4.10460472e-01 1.73277393e-01 -3.21113408e-01
-4.32294644e-02 3.20124298e-01 -2.68651396e-01 -5.06880879e-01
-2.19976917e-01 -1.53380573e-01 -4.88595903e-01 1.03958607e+00
2.30122060e-01 4.95379955e-01 -1.07732869e-03 -9.81425822e-01
-5.40397108e-01 2.51328260e-01 -2.25108311e-01 -2.59588718e-01
1.15724421e+00 -1.31427273e-01 -5.14282405e-01 7.64146566e-01
9.77972925e-01 3.08878720e-01 -7.28622735e-01 -1.41879722e-01
1.95810065e-01 -2.45618358e-01 -4.08819050e-01 -7.14018345e-01
-3.84399205e-01 3.98475856e-01 3.25269222e-01 4.94359046e-01
6.11105263e-01 -5.87607473e-02 6.00285709e-01 -4.32374805e-01
3.95736754e-01 -1.15286398e+00 -2.73091048e-01 3.83057058e-01
8.06918323e-01 -1.19509268e+00 1.51035413e-01 -1.70822859e-01
-3.51829469e-01 3.82222414e-01 6.05560169e-02 4.02241617e-01
7.58300483e-01 -1.15227140e-01 5.25552094e-01 -5.08328319e-01
-9.92083013e-01 -1.36560323e-02 1.49727523e-01 4.29008186e-01
9.58750963e-01 5.14641941e-01 -1.25145245e+00 -5.05300947e-02
-4.29948062e-01 -2.10919872e-01 5.43702364e-01 9.70055103e-01
-3.73702139e-01 -1.30207109e+00 -5.46931565e-01 5.96740186e-01
-7.38894105e-01 -1.53545048e-02 -5.30496061e-01 9.67354596e-01
3.42894137e-01 1.16947651e+00 1.12954594e-01 -2.36900151e-01
1.23666756e-01 2.12259859e-01 2.22661987e-01 -4.70871389e-01
-1.00880265e+00 9.47676823e-02 7.61078835e-01 6.72503412e-02
-1.01564550e+00 -1.32548535e+00 -7.61824369e-01 -8.33094597e-01
-2.09051281e-01 2.71354169e-01 6.13367081e-01 9.97673929e-01
1.61634669e-01 1.79504663e-01 4.52153116e-01 -2.76020411e-02
-4.71238613e-01 -1.01415694e+00 1.21908024e-01 5.00112355e-01
3.86671156e-01 -6.16488576e-01 -2.27636144e-01 -2.75361747e-01] | [8.623737335205078, 9.809857368469238] |
bc074a6e-d578-4b88-815d-ca6ddca1e8ef | enhancement-of-seismic-imaging-an-innovative | 1909.06016 | null | http://arxiv.org/abs/1909.06016v1 | http://arxiv.org/pdf/1909.06016v1.pdf | Enhancement of seismic imaging: An innovative deep learning approach | Enhancing the frequency bandwidth of the seismic data is always the pursuance
at the geophysical community. High resolution of seismic data provides the key
resource to extract detailed stratigraphic knowledge. Here, a novel approach,
based on deep learning model, is introduced by extracting reflections from well
log data to broaden spectrum bandwidth of seismic data through boosting low and
high frequencies. The corresponding improvement is observed from the
enhancement of resolution of seismic data as well as elimination of sidelobe
artifacts from seismic wavelets. During the training stage of deep learning
model, geo-spatial information by taking consideration of multiple wells
simultaneously is fully guaranteed, which assures that laterally and vertically
geological information are constrained by and accurate away from the well
controls during the inversion procedure. Extensive experiments prove that the
enhanced seismic data is consistent with well log information, and honors rock
property relationships defined from the wells at given locations. Uncertainty
analysis could also be quantitatively assessed to determine the possibilities
of a range of seismic responses by leveraging the outputs from the proposed
approach. | [] | 2019-09-13 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [-1.67547569e-01 1.58593226e-02 2.67763406e-01 -2.71848440e-01
-1.09144902e+00 -2.81670153e-01 3.66078973e-01 4.14560474e-02
-3.87384385e-01 9.59106982e-01 4.72304821e-01 -3.09122843e-03
-7.54164577e-01 -1.27104211e+00 -7.16187179e-01 -1.09406447e+00
-6.04836047e-01 1.34035423e-01 1.95540071e-01 -3.26216161e-01
4.00701433e-01 7.13162303e-01 -1.24485505e+00 2.01311529e-01
6.82510138e-01 9.80577528e-01 1.33164018e-01 2.85458058e-01
-6.98292479e-02 4.77493465e-01 -3.21062237e-01 3.00247312e-01
1.50856465e-01 4.36934195e-02 -4.00702089e-01 -7.46417567e-02
-1.09308898e-01 -5.61066031e-01 -1.39130414e-01 8.93136442e-01
6.27130449e-01 1.41025886e-01 6.33335054e-01 -3.76684040e-01
-7.01066921e-04 9.11865056e-01 -8.96286070e-01 1.01212941e-01
-2.08185818e-02 -1.52618542e-01 9.43998992e-01 -1.19911575e+00
-3.32675278e-02 8.19543779e-01 8.17759573e-01 -4.52922851e-01
-1.08329010e+00 -5.28660655e-01 -3.39680105e-01 1.54701561e-01
-1.51055515e+00 -3.60544592e-01 1.26363969e+00 -5.11069298e-01
5.50587416e-01 2.09033415e-01 5.82399666e-01 6.16986811e-01
1.27546221e-01 2.63448834e-01 1.05288506e+00 -5.60957909e-01
2.40596116e-01 -1.89748347e-01 -4.17036898e-02 3.90221596e-01
4.88721937e-01 4.59926516e-01 -9.57476854e-01 -1.39336646e-01
8.87880504e-01 -1.81958750e-01 -5.26418328e-01 1.26832038e-01
-9.19449985e-01 7.93088019e-01 4.47286904e-01 4.91266191e-01
-7.14485228e-01 2.39716053e-01 3.45001072e-01 -2.37084795e-02
4.95952010e-01 2.83557922e-01 -2.73527831e-01 8.97985995e-02
-1.18212831e+00 2.91148156e-01 5.07879376e-01 4.91726756e-01
1.08273780e+00 7.74090469e-01 3.12145472e-01 4.15447712e-01
6.49860561e-01 8.37709010e-01 3.13394755e-01 -7.89427459e-01
3.44671011e-01 1.72856480e-01 1.73957899e-01 -1.07149851e+00
-4.86002415e-01 -6.39363408e-01 -9.29794550e-01 2.56847054e-01
1.94682807e-01 -3.46809238e-01 -4.65348572e-01 1.31932425e+00
2.90798008e-01 2.46573731e-01 2.13697419e-01 6.83271885e-01
5.07373989e-01 7.51075387e-01 -1.41713619e-01 -1.53356805e-01
1.17365468e+00 4.98194210e-02 -6.01554811e-01 -2.73189336e-01
2.58219570e-01 -6.01768434e-01 7.76606798e-01 3.94023508e-01
-6.61434531e-01 -6.17139935e-01 -1.22960699e+00 4.05166179e-01
-2.55555026e-02 -9.79884863e-02 6.49468362e-01 5.71607828e-01
-7.42071390e-01 6.89716637e-01 -1.11964965e+00 5.28830111e-01
-2.00701766e-02 1.58660501e-01 -2.43298128e-01 2.49866486e-01
-1.63309455e+00 6.48776054e-01 5.37514448e-01 8.44993234e-01
-8.13065410e-01 -8.32100153e-01 -7.77154565e-01 3.62544209e-01
-6.18559159e-02 -6.48570508e-02 6.51612401e-01 -5.22341728e-01
-1.12528133e+00 1.34010240e-01 3.26216072e-01 -4.83446926e-01
5.57768881e-01 -4.22950625e-01 -3.71854812e-01 3.01114827e-01
-3.67374420e-02 -1.60928056e-01 7.29120851e-01 -1.28960884e+00
-5.62241077e-01 -3.33670378e-01 -4.13127840e-01 -1.96904793e-01
-4.03330773e-01 -3.95275474e-01 2.25120813e-01 -5.00073969e-01
4.48312581e-01 -2.86018461e-01 -2.58006841e-01 -4.78538483e-01
-2.58989662e-01 3.42817098e-01 4.37546819e-01 -1.13384044e+00
1.08080149e+00 -2.17209625e+00 -2.81675071e-01 7.59368122e-01
-1.38394967e-01 -2.70841330e-01 3.37324649e-01 7.46970475e-01
-4.49249744e-02 -1.19693510e-01 -6.36388063e-01 2.31887385e-01
-1.50750890e-01 8.72755200e-02 -4.14244920e-01 6.36385143e-01
4.77042526e-01 1.61690906e-01 -4.29020911e-01 -3.29580009e-01
1.12856194e-01 5.28321803e-01 -4.80169177e-01 2.69587010e-01
1.24887660e-01 6.23769581e-01 -8.98301840e-01 4.89363641e-01
1.06807995e+00 2.16129735e-01 -1.62745729e-01 -3.71040821e-01
-4.65460837e-01 -2.16470286e-02 -1.53469884e+00 1.29948342e+00
-4.63654459e-01 2.58384883e-01 1.92188472e-01 -1.31714213e+00
1.33164012e+00 4.61422801e-01 5.17275214e-01 -7.94744253e-01
-2.89452463e-01 6.19598806e-01 -4.86742109e-02 -9.35987473e-01
3.16924155e-01 -3.79867375e-01 -1.10312372e-01 8.88619795e-02
-2.03226820e-01 -2.41889477e-01 -2.59842068e-01 -3.35647255e-01
4.63833600e-01 3.98112744e-01 8.99221599e-02 -6.56650186e-01
5.71671367e-01 -2.17535928e-01 4.47396934e-01 6.10902131e-01
5.08985281e-01 3.24071467e-01 2.79974759e-01 -3.10999125e-01
-1.16812205e+00 -9.36577857e-01 -5.34187376e-01 6.42672896e-01
-1.97740227e-01 4.40374643e-01 -4.35025513e-01 1.94473401e-01
4.48111743e-02 4.09077674e-01 -5.35246015e-01 -1.96465537e-01
-8.46729875e-01 -1.15999174e+00 6.66051865e-01 6.28816664e-01
5.67583680e-01 -7.07537115e-01 -9.52628314e-01 4.35869575e-01
-2.32541233e-01 -7.13626981e-01 4.07760441e-01 5.42973638e-01
-1.04074514e+00 -6.73940480e-01 -6.83329165e-01 -4.15995508e-01
1.70520931e-01 -3.49608481e-01 6.74923599e-01 -2.07862809e-01
1.15589060e-01 -1.26272857e-01 -3.16309959e-01 -2.75411308e-01
-3.40119511e-01 -4.62078769e-03 -1.57896787e-01 3.18225175e-01
4.89688180e-02 -8.96485865e-01 -5.20007074e-01 -3.14616114e-02
-7.32226789e-01 -1.62227482e-01 7.94814765e-01 9.56364870e-01
3.49048704e-01 4.66474414e-01 9.16669846e-01 -4.84786570e-01
2.53964365e-01 -5.96125126e-01 -8.48854542e-01 -4.60570715e-02
-2.86530763e-01 3.55282903e-01 3.60529900e-01 5.59249483e-02
-1.48310518e+00 -2.74123996e-01 -6.12430155e-01 2.24691525e-01
-1.55075803e-01 1.11826491e+00 -1.34591639e-01 -3.19149077e-01
6.47196651e-01 4.43078518e-01 -3.00848365e-01 -5.33775985e-01
-1.36162832e-01 6.61206305e-01 6.51673794e-01 -9.62248981e-01
7.23371208e-01 6.07452750e-01 2.24174917e-01 -1.28844309e+00
-4.60032791e-01 -4.51617867e-01 -5.47154665e-01 -4.05890644e-01
4.57969069e-01 -8.60139608e-01 -6.46186709e-01 5.34604847e-01
-9.33955729e-01 1.19719572e-01 1.38629950e-03 9.73392427e-01
-2.50297636e-01 5.10735512e-01 -6.19159698e-01 -1.25219142e+00
-2.52909034e-01 -9.10285830e-01 9.64616299e-01 9.36526656e-02
4.14200965e-03 -1.01346612e+00 -3.58024500e-02 -1.71415862e-02
4.57391590e-01 7.26485670e-01 9.27517593e-01 -5.71431518e-01
-7.20661044e-01 -3.96608919e-01 1.45506440e-02 2.87395656e-01
1.28389597e-01 3.89038175e-02 -1.26420629e+00 -1.48044750e-01
5.66830397e-01 -2.10146934e-01 9.31782842e-01 6.81392968e-01
6.83615625e-01 -1.30811527e-01 1.42622292e-01 6.94865763e-01
1.83021009e+00 9.05786976e-02 6.67836428e-01 5.47500253e-01
3.44777972e-01 8.34643006e-01 4.71499801e-01 1.00094295e+00
-2.06706047e-01 6.73047081e-02 6.37770057e-01 -1.46148920e-01
4.50973362e-01 -1.32921943e-02 6.29776902e-03 7.63107955e-01
-5.21181345e-01 1.47317320e-01 -1.11519814e+00 6.77282274e-01
-1.58003199e+00 -9.85953510e-01 -2.67298102e-01 2.16400075e+00
7.47905314e-01 3.92249644e-01 -3.50504220e-01 5.76656759e-01
6.29886270e-01 3.71350795e-01 -2.38509715e-01 9.68723595e-02
-1.14131160e-01 3.94316077e-01 7.84450710e-01 7.68042505e-01
-8.01396191e-01 3.23401988e-01 5.83878899e+00 6.51754916e-01
-1.23842740e+00 -2.60346711e-01 4.00229394e-01 5.70352137e-01
-7.39043593e-01 -7.95045123e-02 -6.61518633e-01 3.43213916e-01
8.67015064e-01 -1.00444804e-03 1.62912160e-02 6.18152440e-01
7.37230301e-01 -3.02524894e-01 -6.16430879e-01 3.89484584e-01
-4.43500191e-01 -1.49290943e+00 -1.49943933e-01 1.34590968e-01
4.55587745e-01 -1.27445281e-01 -5.73573187e-02 -7.67334457e-03
-5.89892305e-02 -8.39438021e-01 9.05551016e-01 1.04079866e+00
5.36908984e-01 -1.14986789e+00 1.04196334e+00 4.26415384e-01
-1.25451076e+00 -7.63510168e-02 -1.76145643e-01 -9.12153199e-02
2.94920981e-01 8.39599252e-01 -8.06200027e-01 1.07706022e+00
6.69487238e-01 4.46309924e-01 3.53598744e-02 9.87796247e-01
-1.26354024e-01 1.01364028e+00 -5.06402254e-01 1.85170889e-01
3.79272789e-01 -2.85299629e-01 2.88783252e-01 1.20925128e+00
7.45852530e-01 7.28015974e-02 5.80094494e-02 8.73121202e-01
3.94670993e-01 1.37306415e-02 -4.61451352e-01 2.58263409e-01
6.42729819e-01 8.39173555e-01 -3.58125091e-01 9.49930102e-02
-3.57387811e-01 6.08152710e-02 -2.71236539e-01 5.91158867e-01
-7.23567069e-01 -3.35414201e-01 2.18425319e-02 3.21601599e-01
2.40266025e-01 -4.25794542e-01 -5.73964894e-01 -5.47344446e-01
1.10766210e-01 -4.90961641e-01 2.76272565e-01 -5.02473831e-01
-9.90111530e-01 2.72968918e-01 2.23401606e-01 -1.27479827e+00
-3.18670690e-01 -4.32679117e-01 -6.54266655e-01 1.28917825e+00
-1.85256672e+00 -1.12186861e+00 -3.50172102e-01 4.43881214e-01
1.52047455e-01 -7.45586976e-02 6.91761494e-01 4.90384072e-01
4.02552783e-02 1.07599031e-02 4.97756958e-01 2.34810352e-01
4.74189490e-01 -8.59813869e-01 4.83759902e-02 9.09788609e-01
-1.83571532e-01 2.02610508e-01 1.04166842e+00 -7.33422339e-01
-1.12004113e+00 -7.26268828e-01 5.57799637e-01 3.72717261e-01
9.34134126e-01 1.25275537e-01 -1.42327702e+00 4.07316923e-01
-7.68353045e-02 -3.70758498e-05 5.68414569e-01 -4.84972782e-02
8.42496157e-02 -4.90472168e-01 -7.96175182e-01 4.83077392e-02
1.43364236e-01 -5.36487699e-01 -8.12254071e-01 -4.61369678e-02
4.07635242e-01 -3.26832294e-01 -1.03895068e+00 7.37207413e-01
6.00314677e-01 -1.06518185e+00 1.04993296e+00 -2.85736263e-01
6.23903990e-01 -1.58442363e-01 -4.70433235e-01 -1.06500685e+00
-2.05142975e-01 -1.53956011e-01 1.86587483e-01 1.25680089e+00
3.33384961e-01 -5.72553635e-01 8.95667553e-01 2.17749566e-01
-3.55738282e-01 -3.28520149e-01 -1.12849033e+00 -6.07750714e-01
1.70253009e-01 -4.46757376e-01 5.46949267e-01 7.79273570e-01
-2.72101551e-01 -2.83617705e-01 -5.01353681e-01 9.20695961e-01
8.91694307e-01 1.23403504e-01 2.89531827e-01 -1.10767746e+00
-5.20297289e-01 -1.57038808e-01 -1.07607387e-01 -5.65370440e-01
-1.74762741e-01 -2.19143689e-01 2.38383099e-01 -1.26110995e+00
-2.32261941e-01 -5.14416397e-01 -5.55554569e-01 1.05655529e-01
-4.68790196e-02 7.97678679e-02 -4.68042165e-01 2.03139991e-01
6.51214659e-01 6.50265336e-01 1.06952214e+00 4.34104865e-03
-1.84126720e-02 1.07670367e-01 -9.84176397e-02 1.03548396e+00
5.80156207e-01 -4.32095975e-01 -3.35430205e-01 -6.11675262e-01
4.86841887e-01 6.61514759e-01 3.19535583e-01 -1.16665888e+00
1.56487256e-01 -2.76026279e-01 4.36493337e-01 -7.50090361e-01
1.24750666e-01 -9.64079440e-01 3.92738760e-01 5.69065392e-01
-1.27210781e-01 -2.70499825e-01 1.28184468e-01 5.27846098e-01
-7.28474736e-01 -4.40175056e-01 8.32961679e-01 -1.79568768e-01
-1.03662562e+00 -2.04365351e-03 -2.66223937e-01 -2.41011709e-01
5.43161988e-01 -3.17480654e-01 2.35028327e-01 -5.05390503e-02
-6.98190272e-01 1.58416107e-03 -2.60444671e-01 -3.53278399e-01
6.33848310e-01 -1.01063097e+00 -1.09100866e+00 5.10568380e-01
-3.78654599e-02 2.06883296e-01 6.55489981e-01 5.59235275e-01
-8.23328018e-01 1.20887458e-01 -2.58336455e-01 -6.43895924e-01
-5.14390528e-01 -6.68609096e-03 5.67729592e-01 -1.44746855e-01
-6.39878869e-01 7.73788691e-01 -9.99090448e-02 -2.30443361e-03
-4.41178083e-02 -2.55615503e-01 -4.67319310e-01 3.43113035e-01
4.81995016e-01 4.95926768e-01 2.78017402e-01 -4.40249354e-01
-2.29426786e-01 6.26277030e-01 3.88493568e-01 -4.46961790e-01
1.69202232e+00 -4.97494638e-02 -3.12229116e-02 5.65690875e-01
9.50593233e-01 4.25601214e-01 -1.34521937e+00 -2.49860525e-01
4.29169387e-01 -4.24110413e-01 2.83450782e-01 -3.70605320e-01
-8.28425527e-01 1.02445340e+00 4.79828030e-01 6.04874976e-02
1.00657272e+00 -2.82228321e-01 5.86403906e-01 3.80265117e-01
4.19509590e-01 -1.11633492e+00 8.83744359e-02 2.46064365e-01
9.67913747e-01 -1.11924720e+00 -3.17241997e-02 -3.17835547e-02
-1.95341006e-01 1.58960950e+00 1.51466459e-01 -3.37181658e-01
7.59190440e-01 7.42594004e-01 -4.33690473e-02 -1.59862354e-01
-3.71150672e-02 1.57335922e-01 -5.24496995e-02 2.85639614e-01
4.51725543e-01 -1.11248165e-01 -1.57681614e-01 7.06814110e-01
-1.50796339e-01 -2.31798798e-01 4.42135513e-01 9.44186330e-01
-9.91980970e-01 -6.45321846e-01 -9.76586282e-01 8.11263695e-02
-4.28068191e-01 -2.07698181e-01 5.37140727e-01 8.95547807e-01
3.73746790e-02 5.13613641e-01 1.46021526e-02 1.10287890e-01
2.23294824e-01 -4.70122695e-02 9.08801556e-02 -3.57086927e-01
-9.76789296e-02 6.89162374e-01 2.76391178e-01 -5.23349084e-02
-4.48557496e-01 -5.31006157e-01 -1.50532866e+00 -8.72904956e-02
-3.75022262e-01 6.30341947e-01 5.73093653e-01 1.06484687e+00
-3.22073877e-01 7.67892003e-01 8.68791044e-01 -9.62079227e-01
-8.82239521e-01 -9.99840200e-01 -1.03803658e+00 -7.08842650e-02
7.46138096e-01 -5.67037642e-01 -6.87236309e-01 8.11107606e-02] | [6.8802289962768555, 2.5773119926452637] |
1a5875cd-de89-4a07-a0dd-e14e738d422d | a-corpus-of-tables-in-full-text-biomedical | null | null | https://aclanthology.org/W16-5108 | https://aclanthology.org/W16-5108.pdf | A Corpus of Tables in Full-Text Biomedical Research Publications | The development of text mining techniques for biomedical research literature has received increased attention in recent times. However, most of these techniques focus on prose, while much important biomedical data reside in tables. In this paper, we present a corpus created to serve as a gold standard for the development and evaluation of techniques for the automatic extraction of information from biomedical tables. We describe the guidelines used for corpus annotation and the manner in which they were developed. The high inter-annotator agreement achieved on the corpus, and the generic nature of our annotation approach, suggest that the developed guidelines can serve as a general framework for table annotation in biomedical and other scientific domains. The annotated corpus and the guidelines are available at \url{http://www.csse.monash.edu.au/research/umnl/data/index.shtml}. | ['Tatyana Shmanina', 'Ai Lee Cheam', 'Lawrence Cavedon', 'Thomas Bochynek', 'Ingrid Zukerman'] | 2016-12-01 | null | null | null | ws-2016-12 | ['table-annotation', 'table-annotation'] | ['knowledge-base', 'natural-language-processing'] | [ 3.04343551e-01 1.48653418e-01 -4.90695715e-01 -4.99313802e-01
-9.33251083e-01 -4.66062248e-01 1.72120571e-01 9.90784645e-01
-4.89869237e-01 1.08010650e+00 4.65128362e-01 -5.80905259e-01
-1.63743809e-01 -5.29045045e-01 -2.64596224e-01 -4.97790962e-01
4.25724149e-01 5.72543442e-01 4.15190160e-02 -7.39772245e-02
5.92471719e-01 3.83696973e-01 -1.17347419e+00 5.83692431e-01
7.47868180e-01 5.57549715e-01 2.55833000e-01 3.51049751e-01
-4.65929359e-01 6.51367962e-01 -7.01456189e-01 -7.09751666e-01
-3.96986336e-01 -6.27158046e-01 -1.23353887e+00 -2.51720846e-01
-1.00458056e-01 1.74277842e-01 -1.12759806e-01 9.33026731e-01
7.68326759e-01 -2.52272934e-01 5.38161933e-01 -6.89558268e-01
-3.46323371e-01 7.72751451e-01 -1.64165854e-01 3.70424837e-01
5.53809345e-01 -4.71236199e-01 1.06522298e+00 -7.41745353e-01
1.13074684e+00 8.82510245e-01 3.94623697e-01 8.01582575e-01
-9.52168643e-01 -7.46774316e-01 -4.65260804e-01 -1.03867436e-02
-1.32734013e+00 -7.75021970e-01 5.20122349e-01 -6.24901175e-01
8.17803085e-01 5.11458755e-01 5.56381047e-01 8.37706268e-01
4.90590483e-01 4.56936955e-01 1.11004853e+00 -8.29183578e-01
1.98431551e-01 4.99435335e-01 2.63520002e-01 5.21212697e-01
8.64586115e-01 -6.31114960e-01 -5.02570093e-01 -4.27292317e-01
6.50486708e-01 -4.38490331e-01 -1.27205625e-01 -1.42152272e-02
-1.08041942e+00 7.35669196e-01 -2.36644551e-01 8.54865789e-01
-1.84203699e-01 -4.53928113e-01 1.01503050e+00 1.22457512e-01
6.09692454e-01 4.70131218e-01 -5.32309711e-01 -2.32785225e-01
-8.83346200e-01 1.65122420e-01 8.61177027e-01 1.05447209e+00
4.11690250e-02 -4.76185054e-01 1.13269649e-01 1.15514231e+00
3.22279245e-01 -1.18835652e-02 4.23441619e-01 -8.39919329e-01
2.93013066e-01 6.09236240e-01 9.88007933e-02 -1.07217956e+00
-5.41211426e-01 -5.01055084e-02 -5.95768273e-01 -3.91186655e-01
6.57607973e-01 -2.73143947e-01 -5.28850734e-01 1.39429057e+00
4.10465837e-01 -9.42964792e-01 1.48537263e-01 3.90407503e-01
1.15720534e+00 2.45308936e-01 3.73972893e-01 -6.55957162e-01
1.78845990e+00 -1.98563814e-01 -1.40485179e+00 2.30244011e-01
1.11080778e+00 -1.01546204e+00 4.72508252e-01 3.57692122e-01
-1.16934800e+00 -2.24663660e-01 -8.19556117e-01 -3.49454939e-01
-5.75837672e-01 1.33122474e-01 5.00718713e-01 6.52272940e-01
-8.79807413e-01 3.37311804e-01 -8.24859798e-01 -9.08957601e-01
5.86840868e-01 2.21934542e-01 -5.27710199e-01 1.22192815e-01
-1.13967407e+00 1.05423939e+00 6.91109180e-01 2.13298239e-02
3.13370749e-02 -3.02275896e-01 -6.12819433e-01 -5.07782161e-01
3.71021897e-01 -3.45704108e-01 1.17232645e+00 -2.94870704e-01
-7.03264952e-01 1.54597223e+00 -2.84465075e-01 -1.91910788e-01
1.34960562e-01 3.17077674e-02 -6.60357416e-01 4.64820296e-01
4.64866072e-01 3.51022720e-01 -2.70193070e-01 -7.81905353e-01
-7.53731430e-01 -5.59250474e-01 -4.40290719e-01 1.23419128e-02
-2.57261097e-01 5.86270452e-01 -5.59726298e-01 -8.05511057e-01
-1.40737891e-02 -7.69513786e-01 -4.97193262e-02 -1.00431241e-01
-2.65486002e-01 -4.78238404e-01 1.23517320e-01 -1.00178587e+00
1.79794645e+00 -1.91217971e+00 -2.18113050e-01 1.93810746e-01
2.19305575e-01 -6.02490194e-02 5.36001623e-01 9.27555144e-01
-1.25178128e-01 5.40823102e-01 -3.24079126e-01 2.48426676e-01
-1.59000158e-01 3.11414719e-01 1.63701773e-01 6.41976357e-01
-6.46880567e-02 5.56423306e-01 -1.08186400e+00 -1.06841898e+00
-1.77677602e-01 4.04769719e-01 -3.51820216e-02 -5.15406989e-02
7.08905831e-02 5.19364119e-01 -6.62924588e-01 6.95107162e-01
2.24846020e-01 -2.00939327e-01 6.16000056e-01 -8.94941762e-02
-2.88191289e-01 6.78300798e-01 -8.38388145e-01 1.86350667e+00
3.46899748e-01 3.96844685e-01 2.44532794e-01 -9.21864212e-01
1.07819057e+00 7.36750901e-01 9.10485566e-01 -3.40894192e-01
3.45091045e-01 5.65456033e-01 2.87747324e-01 -7.36358166e-01
1.57726005e-01 -1.60662785e-01 -1.78224668e-02 3.13965023e-01
-1.19377626e-03 5.13362736e-02 9.47148323e-01 2.46687680e-01
8.46835613e-01 2.90061146e-01 8.25065136e-01 -5.93579531e-01
7.91783154e-01 5.41534483e-01 7.07767189e-01 3.27792734e-01
-8.71868879e-02 1.52119830e-01 5.87294340e-01 -3.12964201e-01
-1.10555363e+00 -3.35817099e-01 -9.98094916e-01 6.14578843e-01
-4.23516810e-01 -7.98679054e-01 -9.52895343e-01 -4.40262318e-01
-2.06113026e-01 8.02697837e-01 -6.93603933e-01 3.23307633e-01
-4.47504133e-01 -1.06805944e+00 6.18059456e-01 1.75015509e-01
1.71341226e-02 -1.24946952e+00 -7.96306789e-01 5.42272091e-01
-4.95971978e-01 -9.37483668e-01 -1.95525467e-01 4.36562866e-01
-9.92402196e-01 -1.24977338e+00 -5.26763141e-01 -7.88414061e-01
6.88504219e-01 -2.24979594e-01 8.32118928e-01 2.43606001e-01
-4.66821045e-01 1.68036804e-01 -5.39519191e-01 -8.69327068e-01
-8.12511086e-01 2.00924605e-01 -1.22481480e-01 -6.94978833e-01
1.04025972e+00 -1.24137491e-01 -3.63094538e-01 1.69297174e-01
-1.13361835e+00 -3.36851068e-02 4.75767821e-01 7.39285827e-01
7.53932357e-01 -1.51194364e-01 8.00742090e-01 -1.47150350e+00
6.87624574e-01 -4.93357182e-01 -3.20365548e-01 1.86455786e-01
-9.19240177e-01 -1.51267350e-01 -1.02678157e-01 2.02113152e-01
-8.56928051e-01 -1.50672153e-01 -6.85936332e-01 4.76182163e-01
-5.20171285e-01 9.52178419e-01 -8.97210538e-02 1.79909915e-01
6.94821775e-01 -7.16791227e-02 1.88532770e-01 -7.45756805e-01
-2.67078519e-01 1.03776455e+00 1.56870410e-01 -5.57202697e-01
-8.03379714e-02 2.06772298e-01 1.44317495e-02 -9.38658416e-01
-6.49611294e-01 -7.01981843e-01 -8.03698421e-01 -1.27824575e-01
1.03588951e+00 -5.96680999e-01 -3.70976180e-01 -1.83556825e-01
-7.96822786e-01 -5.40204253e-03 -1.27894253e-01 6.62094116e-01
-3.36255521e-01 4.19377059e-01 -6.36622846e-01 -7.10970640e-01
-6.45614028e-01 -1.04019117e+00 7.46532261e-01 -9.09649134e-02
-1.05244732e+00 -1.08031332e+00 2.79347509e-01 6.77701294e-01
-2.69762188e-01 6.58061028e-01 1.05656481e+00 -9.77236331e-01
2.82414466e-01 -3.39578331e-01 1.78161338e-01 -1.85023203e-01
5.02066314e-01 1.18355632e-01 -7.29097486e-01 7.42325410e-02
1.37171745e-01 -1.67955846e-01 2.58695304e-01 2.17233911e-01
9.76686358e-01 -3.99579436e-01 -5.53113639e-01 -5.38111739e-02
1.30374229e+00 6.28512502e-01 5.48608840e-01 6.58953071e-01
3.15816373e-01 1.01399660e+00 7.09616482e-01 3.52848351e-01
3.11135799e-02 6.57233953e-01 -2.71602184e-01 3.89716513e-02
2.61394203e-01 2.43967846e-01 -1.32639587e-01 1.06690013e+00
-1.38183132e-01 -1.01391807e-01 -1.30934834e+00 6.03562653e-01
-1.60351157e+00 -8.45603466e-01 -3.14396501e-01 1.99060023e+00
1.46282387e+00 1.43868968e-01 4.37029272e-01 3.78545433e-01
5.36656141e-01 -4.01299715e-01 -1.06598206e-01 -7.09771156e-01
-2.98046432e-02 2.42003784e-01 2.95012295e-01 1.95144594e-01
-8.26936066e-01 5.30953109e-01 6.25525475e+00 7.60894418e-01
-5.74712873e-01 1.02027871e-01 6.18158162e-01 1.68899854e-03
-4.33913283e-02 -2.73234606e-01 -8.73530746e-01 4.31550920e-01
1.33239520e+00 -4.65368867e-01 -2.57918507e-01 4.74205375e-01
4.27843869e-01 -1.83144152e-01 -9.90285516e-01 6.33974195e-01
-2.56223857e-01 -1.54108334e+00 -1.22272357e-01 2.77089983e-01
9.39412117e-02 -2.08211556e-01 -3.58436018e-01 -2.77862161e-01
-1.50850981e-01 -7.01619864e-01 4.98782516e-01 3.19901437e-01
1.11292815e+00 -4.99680966e-01 9.46255028e-01 -2.89186388e-02
-7.78087378e-01 3.95751208e-01 -2.77881503e-01 1.34256393e-01
4.20308970e-02 6.93242669e-01 -9.00852203e-01 7.37151504e-01
7.76878238e-01 5.34144163e-01 -5.54491103e-01 9.42569613e-01
-9.03660357e-02 5.34862101e-01 1.42308520e-02 -3.00308764e-01
-1.42061561e-02 -2.29599223e-01 3.80139083e-01 1.72972798e+00
4.36499491e-02 2.69686520e-01 -3.63969244e-02 4.75515366e-01
5.18401302e-02 9.57504034e-01 -4.99190927e-01 -5.39578021e-01
5.98502576e-01 1.11790490e+00 -1.12297225e+00 -3.20906281e-01
-7.20428526e-01 2.41552323e-01 1.50846571e-01 -3.70910615e-02
-3.79211247e-01 -7.12624907e-01 4.16473031e-01 1.70080215e-01
5.42969769e-03 1.15243666e-01 -5.76440811e-01 -8.06174040e-01
-1.07586123e-02 -1.15574265e+00 8.56996238e-01 -3.41155976e-01
-9.67827857e-01 6.04222059e-01 3.88762772e-01 -9.16943848e-01
-3.17740083e-01 -6.34039760e-01 2.38470901e-02 9.27474201e-01
-5.93405247e-01 -6.72611773e-01 2.59105355e-01 1.61264315e-01
3.23651403e-01 4.33689691e-02 1.19735503e+00 2.39281937e-01
-7.26829767e-01 4.84361380e-01 3.13284993e-01 4.45030510e-01
8.50337207e-01 -1.22459912e+00 4.24197689e-02 2.50060916e-01
-1.24579370e-01 1.01521182e+00 8.82626474e-01 -1.00131178e+00
-1.06563091e+00 -7.40333974e-01 1.51082969e+00 -4.74187732e-01
7.61322260e-01 -3.64821017e-01 -9.70883071e-01 5.96721590e-01
3.75721365e-01 -7.65549302e-01 1.53340316e+00 3.54975834e-02
1.58604831e-01 1.46216989e-01 -1.31590879e+00 5.43995500e-01
7.29420066e-01 -3.34452987e-01 -6.92660809e-01 6.79173052e-01
1.39816865e-01 -6.25554562e-01 -1.63137078e+00 6.28418475e-02
8.32558334e-01 -4.60238636e-01 5.95260561e-01 -4.94224101e-01
3.97615284e-01 -9.94874611e-02 6.44949600e-02 -5.48799634e-01
-3.42294186e-01 -5.13504624e-01 3.16699088e-01 1.22402275e+00
8.02973926e-01 -5.89068711e-01 6.54954433e-01 6.68227136e-01
-1.68616474e-01 -9.58393455e-01 -9.37069237e-01 -1.65862992e-01
2.36785352e-01 -2.91127980e-01 2.06693307e-01 1.04717755e+00
8.22363079e-01 3.79684061e-01 7.95070753e-02 -4.65419233e-01
5.65529823e-01 -1.77284151e-01 2.77729392e-01 -1.33564472e+00
2.72083610e-01 -3.92342091e-01 -3.12939346e-01 -3.12671810e-02
-2.35714421e-01 -1.01744282e+00 -2.51347333e-01 -1.87657726e+00
1.90428481e-01 -2.72990465e-01 -3.04805577e-01 6.36947811e-01
-1.04716495e-01 1.71721056e-01 -2.89009929e-01 4.39469010e-01
-1.33684710e-01 -3.26136619e-01 8.45131993e-01 2.23833993e-01
-1.62565783e-01 -3.44267309e-01 -1.21000993e+00 3.87311816e-01
1.38370788e+00 -8.55315208e-01 -1.09496213e-01 -4.16507944e-02
3.05920690e-01 -1.19878270e-01 -3.94067585e-01 -6.87782407e-01
1.32307485e-01 -2.10496709e-01 5.38263202e-01 -5.55444062e-01
-6.41055554e-02 -7.39448607e-01 6.07350588e-01 6.13658547e-01
-6.30897164e-01 1.26250058e-01 5.16043842e-01 1.04311042e-01
-1.26864970e-01 -6.60231411e-01 6.03136897e-01 -4.92162108e-01
-1.72817707e-01 -1.66220754e-01 -7.44042873e-01 3.79445888e-02
7.36822307e-01 -3.60210419e-01 -1.26336619e-01 1.35485440e-01
-9.61696804e-01 2.73373902e-01 5.64225554e-01 9.20628458e-02
3.17156702e-01 -1.11882031e+00 -7.31389880e-01 -2.76946932e-01
4.12377954e-01 -1.35405704e-01 1.10855408e-01 1.18021810e+00
-8.27607810e-01 9.47823942e-01 -5.11072397e-01 -1.42356098e-01
-1.48570371e+00 7.23845482e-01 -6.77580759e-02 -2.01566935e-01
-6.90548480e-01 3.78204554e-01 4.07539308e-02 -6.28194073e-03
2.84426749e-01 -1.11180499e-01 -7.31092989e-01 2.05963418e-01
5.77028215e-01 2.75628626e-01 2.70283729e-01 -7.04853117e-01
-5.35836041e-01 1.08165815e-01 -3.39875072e-01 -1.47454143e-01
1.57305622e+00 -1.90901294e-01 -5.46349168e-01 9.93606508e-01
9.43089843e-01 1.58481270e-01 4.08416390e-02 2.92841885e-02
5.66296160e-01 -7.60080963e-02 -1.30005524e-01 -9.08088624e-01
-5.67323089e-01 4.76445943e-01 1.58543006e-01 3.42234939e-01
9.70185399e-01 4.49737124e-02 2.76148438e-01 2.61964977e-01
3.24570686e-02 -1.27930486e+00 -7.51089633e-01 8.26517865e-02
1.00867963e+00 -8.53731215e-01 3.94540638e-01 -8.13824058e-01
-5.46015143e-01 1.38917232e+00 1.81888327e-01 5.57214618e-01
6.15858316e-01 3.75079900e-01 1.46356776e-01 -6.18006289e-01
-6.44861102e-01 5.50132692e-02 1.86681941e-01 5.65084815e-01
1.32517838e+00 -2.61402994e-01 -1.51938939e+00 4.16528612e-01
-2.52656579e-01 4.87947464e-01 5.03268838e-01 1.47152233e+00
-2.71571249e-01 -1.45758235e+00 -4.49228287e-01 8.76570880e-01
-1.36538577e+00 -1.58014059e-01 -7.35316038e-01 9.22729313e-01
9.72910970e-02 9.59378004e-01 -2.05781192e-01 3.82728800e-02
2.65962511e-01 4.80298072e-01 4.71801937e-01 -7.56431282e-01
-6.81201160e-01 7.23832548e-01 7.73290038e-01 -1.61797136e-01
-8.35848570e-01 -1.08355463e+00 -1.22126150e+00 -1.72747120e-01
-1.69432357e-01 9.37094212e-01 4.51266557e-01 6.71220839e-01
1.36078775e-01 5.46124279e-01 -1.25904828e-01 5.82462251e-02
-7.71150738e-02 -1.16437864e+00 -7.47127652e-01 1.49935603e-01
-1.36111572e-01 -5.06679714e-01 -4.76885736e-02 5.52732646e-01] | [8.574207305908203, 8.673831939697266] |
42165d23-834d-4bf7-acd8-3746d80c4b42 | correntropy-based-logistic-regression-with | 2207.09693 | null | https://arxiv.org/abs/2207.09693v1 | https://arxiv.org/pdf/2207.09693v1.pdf | Correntropy-Based Logistic Regression with Automatic Relevance Determination for Robust Sparse Brain Activity Decoding | Recent studies have utilized sparse classifications to predict categorical variables from high-dimensional brain activity signals to expose human's intentions and mental states, selecting the relevant features automatically in the model training process. However, existing sparse classification models will likely be prone to the performance degradation which is caused by noise inherent in the brain recordings. To address this issue, we aim to propose a new robust and sparse classification algorithm in this study. To this end, we introduce the correntropy learning framework into the automatic relevance determination based sparse classification model, proposing a new correntropy-based robust sparse logistic regression algorithm. To demonstrate the superior brain activity decoding performance of the proposed algorithm, we evaluate it on a synthetic dataset, an electroencephalogram (EEG) dataset, and a functional magnetic resonance imaging (fMRI) dataset. The extensive experimental results confirm that not only the proposed method can achieve higher classification accuracy in a noisy and high-dimensional classification task, but also it would select those more informative features for the decoding scenarios. Integrating the correntropy learning approach with the automatic relevance determination technique will significantly improve the robustness with respect to the noise, leading to more adequate robust sparse brain decoding algorithm. It provides a more powerful approach in the real-world brain activity decoding and the brain-computer interfaces. | ['Yasuharu Koike', 'Natsue Yoshimura', 'Yuxi Shi', 'Badong Chen', 'Yuanhao Li'] | 2022-07-20 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 5.17841041e-01 -3.75944555e-01 2.05425352e-01 -4.92727816e-01
-2.10809618e-01 1.32703006e-01 3.90453041e-01 -6.64706826e-02
-2.28712827e-01 8.55376363e-01 4.46095288e-01 3.07235777e-01
-4.81125653e-01 -4.80463415e-01 -3.40499580e-01 -7.93999314e-01
-1.61772609e-01 -1.16845131e-01 -2.82332301e-01 1.86320335e-01
5.52361131e-01 2.98918545e-01 -1.90073657e+00 2.13627458e-01
1.20555437e+00 1.21514690e+00 5.56294084e-01 -1.05160348e-01
5.27285933e-02 6.11164570e-01 -5.12680173e-01 1.31030023e-01
-1.02257624e-01 -6.04605675e-01 -8.32308829e-02 -1.54432012e-02
-9.78502110e-02 1.41324803e-01 -2.67555919e-02 1.38882351e+00
5.87541938e-01 1.96630687e-01 8.52606654e-01 -9.98777866e-01
-4.39786553e-01 5.57523489e-01 -5.69341123e-01 5.77796340e-01
5.80388963e-01 -9.58740413e-02 5.41909218e-01 -1.03717804e+00
1.81154281e-01 1.01839983e+00 4.56104398e-01 2.45448992e-01
-1.12733042e+00 -1.02053535e+00 9.12691467e-03 5.30620098e-01
-1.52109253e+00 -6.38942242e-01 1.25632477e+00 -6.32151723e-01
6.33908153e-01 2.03232557e-01 9.61859345e-01 1.47513533e+00
4.83708799e-01 5.09028971e-01 1.46303129e+00 -1.98090717e-01
4.66082811e-01 2.04428732e-01 3.95168692e-01 5.11258781e-01
5.50012827e-01 3.63150612e-02 -6.46903515e-01 -2.58363247e-01
6.19988561e-01 1.80600360e-01 -6.51492298e-01 -2.38810241e-01
-1.43805552e+00 7.46016979e-01 2.44007140e-01 7.62700021e-01
-8.01169813e-01 -3.48121405e-01 2.89085984e-01 1.01525992e-01
4.25819367e-01 2.82691240e-01 -5.65954112e-02 -1.49047494e-01
-1.04149270e+00 -2.12547570e-01 6.41381860e-01 4.39936548e-01
5.15340805e-01 4.34265494e-01 -1.38046890e-01 1.01230347e+00
3.66267204e-01 4.22634602e-01 1.23462248e+00 -5.30499578e-01
4.80065346e-01 3.98539662e-01 -3.18338573e-01 -1.60163331e+00
-5.93244076e-01 -8.46153796e-01 -1.21824837e+00 -8.75740871e-02
-2.19931185e-01 -1.03688143e-01 -3.37851644e-01 1.69524896e+00
3.78765911e-02 5.61026633e-01 9.01686847e-02 1.13720536e+00
6.31929994e-01 6.65861726e-01 9.32879746e-02 -8.45650315e-01
1.38648176e+00 -2.57513195e-01 -1.05908525e+00 -1.10790782e-01
2.69210994e-01 -5.05607843e-01 9.46681440e-01 6.97448134e-01
-5.74275911e-01 -4.71420109e-01 -1.16346586e+00 5.84570944e-01
6.70422018e-02 4.96946186e-01 8.39649558e-01 5.23353577e-01
-5.72821259e-01 3.08704436e-01 -7.87845016e-01 -2.96748549e-01
5.11689007e-01 3.53145748e-01 -6.09233320e-01 5.34697063e-03
-1.16963840e+00 7.95105219e-01 5.25067031e-01 3.29514980e-01
-4.99833047e-01 -2.07840592e-01 -9.62623715e-01 3.65348279e-01
-4.03756127e-02 -4.61652964e-01 5.92459559e-01 -1.17459393e+00
-1.44683599e+00 2.21258312e-01 -3.27059180e-01 -4.50026572e-01
-1.51304781e-01 -3.29728685e-02 -5.97374082e-01 2.12994486e-01
4.28328663e-02 1.27943680e-01 1.24922371e+00 -7.56811142e-01
-2.41571628e-02 -7.01184750e-01 -8.35783958e-01 3.27014655e-01
-6.75879240e-01 -1.39734536e-01 4.10734892e-01 -9.23498034e-01
6.82239056e-01 -5.47575831e-01 1.29063129e-01 -6.03800535e-01
2.43073385e-02 -9.83889177e-02 6.30671859e-01 -6.11828983e-01
1.18161595e+00 -2.29673386e+00 2.01929152e-01 4.52026874e-01
9.62707847e-02 5.51676601e-02 -4.60210480e-02 -1.46620348e-02
-4.00562137e-01 -3.86188447e-01 -3.59212220e-01 -2.25099139e-02
-3.74638915e-01 -1.98496103e-01 -5.03306240e-02 7.44040370e-01
2.23191887e-01 4.83593285e-01 -5.94353914e-01 -4.41270441e-01
5.69962785e-02 6.02923751e-01 -4.80957329e-01 2.49033958e-01
4.17989969e-01 5.85367441e-01 -6.53202891e-01 6.35379195e-01
4.64616537e-01 -1.40111193e-01 -1.36440516e-01 -3.83502871e-01
1.41223446e-01 -1.13444328e-01 -1.27581930e+00 1.69679952e+00
-2.76685417e-01 5.28520584e-01 1.32725775e-01 -1.44822800e+00
1.36241508e+00 3.91190529e-01 6.66300535e-01 -8.68878603e-01
3.29763681e-01 2.23858088e-01 2.84105003e-01 -8.76781583e-01
-4.89895046e-02 -2.72684038e-01 2.45457590e-01 3.84548366e-01
1.66275263e-01 2.45628774e-01 -3.02654654e-01 -7.44480863e-02
8.89916539e-01 -2.05673754e-01 5.53871334e-01 -4.17662859e-01
9.02034581e-01 -5.40756226e-01 7.02200770e-01 2.05211312e-01
-2.69481957e-01 3.13459396e-01 3.28498989e-01 -2.31648698e-01
-3.94327968e-01 -5.42055666e-01 -4.61183012e-01 6.57475352e-01
1.83546469e-02 -1.25313148e-01 -6.66839421e-01 -2.81582206e-01
-3.01125884e-01 7.25404620e-01 -4.33059216e-01 -5.37112057e-01
-1.82921439e-01 -1.13970280e+00 9.04394016e-02 1.84806913e-01
4.59571302e-01 -1.18944061e+00 -4.82611984e-01 2.89715767e-01
-2.71097183e-01 -7.81699598e-01 -2.23664209e-01 2.24893600e-01
-1.08082891e+00 -7.76210606e-01 -5.58890820e-01 -8.27084005e-01
7.63804257e-01 1.42439529e-01 4.30157095e-01 -3.80844265e-01
-8.32384899e-02 1.82926029e-01 -4.42883909e-01 -1.40557796e-01
6.84383139e-02 -3.28371599e-02 4.03290451e-01 6.22532248e-01
6.16377592e-01 -9.47282791e-01 -6.09029651e-01 1.45960301e-01
-7.66908824e-01 -3.04647796e-02 9.05223966e-01 9.99466896e-01
4.90291357e-01 2.04218715e-01 1.19581103e+00 -5.47371924e-01
1.23905110e+00 -9.95103002e-01 -5.16001284e-01 -1.28095061e-01
-7.39580154e-01 1.46415532e-01 7.60910153e-01 -6.13803625e-01
-1.07932794e+00 3.18658024e-01 -1.13688946e-01 -3.60095918e-01
-2.11204290e-01 8.75299335e-01 -1.64056450e-01 -8.21461827e-02
6.91687703e-01 7.95135260e-01 4.55167592e-02 -5.23610473e-01
-1.60260439e-01 9.22845602e-01 3.03347319e-01 -3.34667385e-01
3.43847632e-01 -9.94038954e-03 -1.79461613e-02 -8.69395614e-01
-4.01199102e-01 -3.54575455e-01 -4.17781651e-01 -1.97872907e-01
6.16840422e-01 -1.04318750e+00 -6.24212384e-01 2.06794351e-01
-1.08431625e+00 4.51035202e-01 2.10353047e-01 1.25072038e+00
-5.93059838e-01 5.76822281e-01 -2.74411827e-01 -1.05974412e+00
-5.37344456e-01 -1.20818865e+00 8.02789927e-01 2.38893583e-01
-2.93915838e-01 -5.91479182e-01 -4.47318777e-02 2.61957645e-01
5.13522506e-01 1.02790520e-01 8.95262897e-01 -8.56239319e-01
-2.85177141e-01 -3.00637841e-01 2.80045252e-03 3.91472280e-01
9.03282464e-02 -5.90912640e-01 -8.21514726e-01 -2.47836232e-01
9.16786551e-01 -1.88614309e-01 7.99236476e-01 2.84043908e-01
1.28648448e+00 -4.14719701e-01 -1.25883818e-01 6.25537276e-01
1.27883148e+00 3.70286167e-01 5.80097377e-01 1.22854505e-02
4.64139044e-01 5.49758196e-01 5.66380978e-01 6.11177146e-01
6.66709915e-02 4.28182781e-01 3.15291286e-01 2.73353636e-01
1.50239170e-01 -1.82119563e-01 5.72362423e-01 1.38538301e+00
4.34948057e-02 2.01416239e-01 -7.00394809e-01 3.08127761e-01
-1.79874516e+00 -1.15605044e+00 -1.46360900e-02 2.24504471e+00
6.69484377e-01 -1.47443399e-01 -2.08678365e-01 4.14648771e-01
7.71089494e-01 3.51006389e-02 -5.30466437e-01 -1.30641758e-01
-1.88172653e-01 1.88096777e-01 -1.56500131e-01 5.52794263e-02
-8.35210204e-01 4.33034271e-01 5.77726793e+00 6.97351336e-01
-1.40352261e+00 1.74099743e-01 4.64813918e-01 -2.12310821e-01
-2.34269887e-01 -3.65233600e-01 -3.96032095e-01 8.50381911e-01
1.01977444e+00 -1.87631369e-01 6.55310094e-01 8.60253930e-01
6.18380189e-01 -1.57385111e-01 -8.45798254e-01 1.69483066e+00
4.56060737e-01 -9.31055486e-01 4.08628881e-02 7.42801279e-03
4.97678995e-01 -1.16999641e-01 -2.50638928e-03 1.87980205e-01
-6.18963480e-01 -1.13497782e+00 3.79010320e-01 9.90963578e-01
4.96139228e-01 -6.23967171e-01 9.00999427e-01 7.06333280e-01
-8.70655477e-01 -4.35251743e-01 -5.14539480e-01 -2.16686726e-01
-2.30197027e-01 9.57951069e-01 -6.04095459e-01 1.61122084e-01
4.07108158e-01 1.09543788e+00 -5.53446651e-01 1.11768985e+00
3.59535888e-02 6.94076419e-01 -1.37867510e-01 -2.87942380e-01
-3.38002264e-01 -4.53951836e-01 6.70564115e-01 1.04997134e+00
7.63574183e-01 3.66390020e-01 -2.26739332e-01 1.11853468e+00
1.74359500e-01 6.55086637e-01 -8.40618551e-01 -3.13285552e-02
3.79537463e-01 1.22128773e+00 -6.50554419e-01 -1.04786552e-01
-3.21673572e-01 8.87197256e-01 1.85021207e-01 4.15059954e-01
-6.81961298e-01 -4.82948184e-01 2.44875640e-01 5.42431921e-02
1.65133700e-01 -2.89090186e-01 -5.11517167e-01 -1.65497708e+00
1.95815101e-01 -9.34185982e-01 5.98370805e-02 -7.55188346e-01
-1.34223938e+00 7.59866297e-01 -9.60773379e-02 -1.40650606e+00
-2.22260296e-01 -2.76884198e-01 -5.08055031e-01 9.99842286e-01
-1.38850713e+00 -4.53218758e-01 -3.68532956e-01 9.29306984e-01
4.97121483e-01 -6.31124735e-01 8.81559670e-01 3.63035351e-01
-6.83729589e-01 2.20165282e-01 9.61840749e-02 -1.75808191e-01
4.84157383e-01 -5.85006237e-01 -6.62422955e-01 7.95619309e-01
1.79462820e-01 8.04279149e-01 4.69748437e-01 -5.88963509e-01
-1.49901712e+00 -8.13675582e-01 4.84125376e-01 9.23026204e-02
5.14840186e-01 -3.16820443e-01 -1.05418169e+00 2.75522470e-01
-5.06651849e-02 -6.83729202e-02 9.52054739e-01 6.94460720e-02
1.27603352e-01 -3.79958153e-01 -1.31960869e+00 3.36126924e-01
7.43968248e-01 -5.05786598e-01 -1.11616504e+00 2.40743488e-01
2.93540657e-01 1.56670704e-01 -9.14059103e-01 3.72261524e-01
6.21317029e-01 -8.34359229e-01 6.16319835e-01 -6.01049662e-02
2.66961187e-01 -3.16295415e-01 -1.63430095e-01 -1.42583454e+00
-5.74302673e-01 -2.44961858e-01 -3.41708452e-01 1.07020044e+00
1.30491674e-01 -9.54857707e-01 5.59482694e-01 4.61229324e-01
1.51935101e-01 -6.96389198e-01 -1.21022820e+00 -4.46093768e-01
-5.16126275e-01 -4.33071971e-01 4.79039252e-01 9.85517800e-01
5.09817123e-01 5.85547805e-01 -4.82628524e-01 7.12407902e-02
6.78062797e-01 -1.76233221e-02 2.43721962e-01 -1.59017730e+00
-9.56414193e-02 -3.10129344e-01 -8.45097005e-01 -6.77232325e-01
4.17288840e-01 -9.96059775e-01 -9.31520313e-02 -1.18315125e+00
3.07733595e-01 -1.42559454e-01 -6.94370627e-01 2.54171342e-01
-1.51289240e-01 -6.38123788e-03 -9.65881050e-02 5.11755764e-01
-3.30173552e-01 1.10524964e+00 9.76626813e-01 -9.38546807e-02
-2.79258579e-01 1.23342708e-01 -9.01557207e-01 5.91149926e-01
8.27270687e-01 -6.14775777e-01 -7.08403945e-01 -1.35798514e-01
-1.37966916e-01 3.80027115e-01 2.44040385e-01 -1.36676085e+00
1.79199353e-01 2.34507769e-02 6.76089764e-01 -2.13775247e-01
4.28724408e-01 -1.13499284e+00 1.37762234e-01 4.69382763e-01
-1.59999430e-01 -1.52397364e-01 -3.22881043e-02 7.23460197e-01
-5.06491840e-01 -2.35312611e-01 6.98775887e-01 3.43704373e-02
-4.89894986e-01 2.03084409e-01 -5.07960141e-01 -4.09087241e-01
1.05141234e+00 -5.04291534e-01 -8.77416879e-02 -3.75848800e-01
-7.46286273e-01 -1.91331133e-01 -1.61696933e-02 3.20667982e-01
1.02120769e+00 -1.41730165e+00 -6.18293464e-01 8.89644325e-01
-2.20239852e-02 -6.46069586e-01 1.95009664e-01 1.31013203e+00
2.55076587e-03 4.36972737e-01 -5.50332189e-01 -6.59190714e-01
-9.57139492e-01 4.87220407e-01 1.29194930e-01 2.55247712e-01
-5.36860824e-01 3.34006310e-01 -6.36803173e-03 1.84019685e-01
2.46588722e-01 -3.74758333e-01 -8.34408045e-01 2.08684772e-01
5.89235842e-01 1.29661843e-01 1.91936865e-01 -9.12317395e-01
-4.57435787e-01 5.17677546e-01 2.87986010e-01 -4.73822057e-02
1.46896148e+00 -3.22919376e-02 -2.09531739e-01 5.43015718e-01
1.20086551e+00 -4.46533486e-02 -8.01635683e-01 -1.74964279e-01
6.51580840e-02 -5.03591418e-01 3.53637666e-01 -6.74865305e-01
-1.04787064e+00 9.12797809e-01 8.72790515e-01 1.69311821e-01
1.33660793e+00 -4.03399020e-01 5.19552112e-01 4.76527721e-01
6.80170357e-01 -8.90617967e-01 -5.96242025e-02 2.23442420e-01
1.06546354e+00 -1.18450308e+00 1.40017748e-01 -1.91474676e-01
-6.66115880e-01 1.22077620e+00 4.07708645e-01 -2.07966119e-01
8.48640442e-01 -1.28524438e-01 -3.83986771e-01 -2.89764971e-01
-7.03420401e-01 1.35757238e-01 3.74412358e-01 6.28279448e-01
5.09593129e-01 1.01897828e-02 -7.99286723e-01 1.21596980e+00
-2.22132936e-01 2.69460768e-01 2.85445452e-01 5.32906771e-01
-6.55619681e-01 -5.93853951e-01 -6.29695296e-01 8.82083237e-01
-3.20177734e-01 -1.64941788e-01 -1.26744229e-02 2.81438790e-03
-5.75980358e-03 9.19132411e-01 -1.88348711e-01 -4.32362854e-01
2.59814143e-01 2.95893908e-01 2.16817498e-01 -6.76676869e-01
-2.81215221e-01 1.18994825e-01 -2.92715132e-01 -5.84004462e-01
-5.49540877e-01 -7.70382524e-01 -1.04056799e+00 2.54907727e-01
-4.97009099e-01 3.13094974e-01 7.67577052e-01 8.72347593e-01
5.89293242e-01 4.74774569e-01 7.61586964e-01 -8.06018829e-01
-5.47367334e-01 -1.18224716e+00 -8.63749444e-01 3.78882051e-01
1.08225517e-01 -9.85687435e-01 -5.21217704e-01 -1.03743009e-01] | [13.023978233337402, 3.461026191711426] |
f6606a03-ba05-4c7f-84c7-676dfcb6cd90 | no-intruder-no-validity-evaluation-criteria | 2103.09263 | null | https://arxiv.org/abs/2103.09263v1 | https://arxiv.org/pdf/2103.09263v1.pdf | No Intruder, no Validity: Evaluation Criteria for Privacy-Preserving Text Anonymization | For sensitive text data to be shared among NLP researchers and practitioners, shared documents need to comply with data protection and privacy laws. There is hence a growing interest in automated approaches for text anonymization. However, measuring such methods' performance is challenging: missing a single identifying attribute can reveal an individual's identity. In this paper, we draw attention to this problem and argue that researchers and practitioners developing automated text anonymization systems should carefully assess whether their evaluation methods truly reflect the system's ability to protect individuals from being re-identified. We then propose TILD, a set of evaluation criteria that comprises an anonymization method's technical performance, the information loss resulting from its anonymization, and the human ability to de-anonymize redacted documents. These criteria may facilitate progress towards a standardized way for measuring anonymization performance. | ['Bennett Kleinberg', 'Maximilian Mozes'] | 2021-03-16 | null | null | null | null | ['text-anonymization'] | ['natural-language-processing'] | [ 2.38688067e-01 1.57967985e-01 -1.18131965e-01 -6.64624333e-01
-8.43331993e-01 -1.09615278e+00 5.96830308e-01 7.95703590e-01
-4.79095727e-01 9.11794066e-01 6.20611966e-01 -1.09957382e-01
-1.36716038e-01 -7.52588689e-01 -2.74955899e-01 -2.45221749e-01
4.01672721e-01 4.59213436e-01 -3.27505767e-01 3.57188940e-01
1.93238661e-01 7.86886632e-01 -9.58077848e-01 3.15579563e-01
1.23015821e+00 3.66074890e-01 -7.45813608e-01 3.83158833e-01
-2.23914921e-01 4.90201056e-01 -7.61316001e-01 -1.07887363e+00
5.95063448e-01 -3.94225031e-01 -1.20352209e+00 -3.61544818e-01
4.22692835e-01 -3.18039536e-01 -1.41729698e-01 1.34063435e+00
4.34310406e-01 -2.67957836e-01 4.17291939e-01 -1.65615821e+00
-7.76114285e-01 7.77440012e-01 -3.73956263e-01 5.05116135e-02
5.87828755e-01 1.15349360e-01 1.08223248e+00 -2.13954493e-01
9.02120829e-01 8.82870734e-01 8.76609683e-01 4.62712973e-01
-1.67935526e+00 -5.36366045e-01 -2.87218750e-01 -2.72136360e-01
-1.47457790e+00 -6.87235355e-01 3.75274569e-01 -6.35154247e-01
3.74231935e-01 7.44691491e-01 4.07192975e-01 8.12665343e-01
5.60557470e-02 5.22864640e-01 9.90838051e-01 -2.62175471e-01
2.58280724e-01 5.38967073e-01 2.45564029e-01 1.05682135e-01
1.03383636e+00 -3.28383923e-01 -3.98183525e-01 -6.02181017e-01
2.91622896e-03 -2.76976287e-01 -3.38305235e-01 -6.16370857e-01
-9.82719004e-01 6.01967394e-01 -3.18891928e-02 5.98683059e-01
-2.30854914e-01 -1.81941107e-01 7.08564103e-01 3.78310353e-01
5.70747972e-01 1.01449454e+00 -2.51479238e-01 -1.77245796e-01
-1.03067279e+00 5.39906859e-01 8.70801032e-01 1.02711833e+00
7.86320806e-01 -5.21424949e-01 -2.03702375e-01 5.85501254e-01
-1.04699554e-02 2.26931155e-01 4.56070364e-01 -7.70891726e-01
5.59136450e-01 1.12176752e+00 5.35565376e-01 -1.25199425e+00
-1.15798555e-01 9.35770124e-02 -6.66154861e-01 -2.80078351e-02
5.03314674e-01 -1.83489010e-01 -2.74261206e-01 1.71994257e+00
1.93344787e-01 -6.75464511e-01 3.54202986e-01 5.64351976e-01
4.11181480e-01 1.54851750e-01 4.38018799e-01 -1.24366082e-01
1.37854028e+00 -8.20777565e-02 -9.68858719e-01 5.62994294e-02
1.07650125e+00 -7.91965365e-01 9.99553680e-01 -5.10490276e-02
-1.08791792e+00 -4.99122553e-02 -8.33712578e-01 -2.07780644e-01
-5.85286617e-01 -1.82774693e-01 5.78665555e-01 1.45379102e+00
-8.18122923e-01 6.69828892e-01 -6.58094764e-01 -7.49960661e-01
7.07704961e-01 5.13959467e-01 -1.02117682e+00 9.93340611e-02
-1.18802452e+00 6.32349968e-01 5.09414613e-01 -1.65953457e-01
-5.42845875e-02 -8.10608447e-01 -5.35505950e-01 1.11624755e-01
3.73280086e-02 -7.91276395e-01 1.02675295e+00 -9.87135828e-01
-7.30772376e-01 1.35140228e+00 -9.17402655e-02 -4.80903953e-01
9.24001932e-01 -2.63990641e-01 -7.48972714e-01 -1.27248928e-01
4.45161998e-01 1.27047122e-01 4.02569830e-01 -1.22936702e+00
-7.21072137e-01 -8.53300095e-01 -2.89231777e-01 5.67559190e-02
-6.55455649e-01 2.89362878e-01 -3.96481622e-03 -6.99389875e-01
7.37670287e-02 -9.64677453e-01 -2.68163346e-03 9.68811940e-03
-7.52580643e-01 1.48605719e-01 8.27318192e-01 -8.46631765e-01
1.36712432e+00 -2.07634163e+00 -3.48735005e-01 3.89643639e-01
3.97496879e-01 6.36964858e-01 7.42567927e-02 7.61163294e-01
-6.21976741e-02 9.91461217e-01 -3.73340189e-01 -4.77520287e-01
1.81923866e-01 -9.39866602e-02 -3.85974735e-01 7.16069579e-01
-8.42526779e-02 8.84731829e-01 -7.48635828e-01 -3.92698944e-01
5.24566602e-03 2.39201561e-01 -3.95831764e-01 1.02748193e-01
7.48203099e-02 2.37671211e-01 -5.95670164e-01 5.47641039e-01
9.45198476e-01 2.18804941e-01 4.43619579e-01 2.33227890e-02
-9.71476883e-02 1.60224617e-01 -1.09032667e+00 1.27557051e+00
5.93422651e-02 6.54207051e-01 9.84329581e-02 -3.11038822e-01
8.91831696e-01 3.69544178e-01 6.64460897e-01 -5.23905039e-01
7.87239671e-02 3.04260492e-01 -2.86024064e-01 -3.45490098e-01
9.41344380e-01 -6.93737566e-02 -2.83218056e-01 9.68720913e-01
-4.56051350e-01 1.58422679e-01 8.60990211e-02 3.60547721e-01
1.02053213e+00 -3.64048660e-01 2.96244293e-01 -4.99838471e-01
5.62101126e-01 1.69783307e-03 6.94435716e-01 7.25424767e-01
-4.84701246e-01 4.77551162e-01 5.49010396e-01 -4.77070421e-01
-1.41537964e+00 -8.38227570e-01 -1.24705486e-01 6.00886285e-01
-1.92875698e-01 -5.64844072e-01 -1.15177023e+00 -8.12760830e-01
5.27249873e-01 8.94191563e-01 -6.25829577e-01 -3.78837913e-01
-3.70424896e-01 -6.98031843e-01 1.18675232e+00 2.12618873e-01
3.94633442e-01 -6.66791320e-01 -3.89994919e-01 6.70178458e-02
-4.31699216e-01 -9.81924593e-01 -6.59338951e-01 -4.64607686e-01
-6.87074482e-01 -1.25760341e+00 -3.60634416e-01 -1.65612832e-01
8.19634676e-01 7.39923865e-02 7.63250053e-01 1.31243557e-01
-2.41131529e-01 4.37639356e-01 -4.21154767e-01 -5.31530619e-01
-8.74291599e-01 4.62467521e-01 1.17923714e-01 1.84701294e-01
8.81140053e-01 -4.58301693e-01 -4.64364737e-01 3.91467452e-01
-1.33696449e+00 -4.62238342e-01 -1.98074547e-03 1.87285706e-01
2.45577410e-01 6.80295154e-02 4.88486439e-01 -1.37435842e+00
1.24477386e+00 -2.29740798e-01 -3.87817740e-01 6.16688550e-01
-9.38897491e-01 1.13748856e-01 6.47746861e-01 4.05109860e-02
-9.79971230e-01 2.46080142e-02 8.80521983e-02 1.15723141e-01
-2.61303514e-01 2.37041473e-01 -6.82229102e-01 -2.29951486e-01
6.02201760e-01 -1.00703225e-01 6.47462904e-02 -6.04223967e-01
5.37326276e-01 9.02733445e-01 5.57288706e-01 -5.96695065e-01
9.27200496e-01 6.27413332e-01 -3.07747215e-01 -6.12113297e-01
-4.44897711e-01 -4.44770277e-01 -7.41240025e-01 1.62503734e-01
5.67045510e-01 -7.15309024e-01 -8.22416425e-01 3.94313842e-01
-1.00743306e+00 4.16531086e-01 -5.86553156e-01 3.20113063e-01
-3.38127077e-01 8.05999219e-01 -1.21359549e-01 -6.59609616e-01
-5.63000679e-01 -8.04100096e-01 6.82957947e-01 -5.57872951e-02
-1.05471933e+00 -8.07178855e-01 2.89104611e-01 6.48590267e-01
3.36678505e-01 4.46735978e-01 8.96882355e-01 -1.09481943e+00
-2.03060061e-01 -7.35974729e-01 -2.18152493e-01 3.12093705e-01
4.74980950e-01 1.14243105e-01 -9.45285499e-01 -5.47174454e-01
8.46647471e-02 3.84164490e-02 2.79635578e-01 -8.59988853e-02
8.09067309e-01 -8.40102673e-01 -3.70602429e-01 7.02299237e-01
1.40571475e+00 4.20147926e-02 8.79749000e-01 4.12552238e-01
7.05961406e-01 9.78892326e-01 9.15339068e-02 4.89352882e-01
1.43316776e-01 5.90353251e-01 -6.06904626e-02 3.78609374e-02
3.43963176e-01 -5.99378526e-01 -1.64243311e-01 4.53858137e-01
2.08027452e-01 -6.02661192e-01 -9.30007219e-01 5.99312127e-01
-1.65547347e+00 -1.07366753e+00 -2.98798770e-01 2.46731448e+00
6.03836894e-01 -2.92884290e-01 2.97834128e-01 2.96789229e-01
8.96571338e-01 -1.93990692e-02 -3.86834264e-01 -6.68911338e-01
-4.57846642e-01 -2.86361158e-01 8.51530373e-01 3.43583822e-01
-8.07628572e-01 8.26546907e-01 6.61927509e+00 2.68153459e-01
-7.25865364e-01 -2.35168427e-01 6.83683574e-01 -1.78136490e-03
-7.79391110e-01 2.85869837e-01 -6.50349200e-01 6.86932564e-01
8.89901221e-01 -1.05297065e+00 2.48416990e-01 5.62555373e-01
2.11764768e-01 1.08015433e-01 -1.29016519e+00 6.73368216e-01
-6.42743409e-02 -1.13830638e+00 3.47254455e-01 4.71924007e-01
7.16375232e-01 -5.14738679e-01 3.78171466e-02 -3.66685688e-01
6.89817429e-01 -1.05832183e+00 4.82232809e-01 5.77854812e-01
8.59610021e-01 -8.06182861e-01 6.96217954e-01 -7.51913637e-02
-8.71026218e-01 7.34785721e-02 -2.72148818e-01 2.35694617e-01
1.54912502e-01 6.49740636e-01 -9.50545251e-01 6.19762778e-01
5.40331304e-01 3.03336382e-01 -7.15304494e-01 7.91172743e-01
-1.53844906e-02 2.56333470e-01 -2.05765560e-01 2.25512624e-01
-2.70219352e-02 -3.64029378e-01 5.93194366e-01 1.05129945e+00
2.41141230e-01 1.56445410e-02 -2.90700644e-01 9.71310794e-01
-5.79777658e-01 4.71368462e-01 -7.93610632e-01 -7.37029970e-01
8.98466408e-01 8.88636410e-01 -6.04276597e-01 -7.43406191e-02
-2.48576283e-01 1.11993349e+00 2.68610179e-01 1.41341597e-01
-2.94777989e-01 -5.22191703e-01 1.12132692e+00 3.72900754e-01
-2.95371264e-01 8.55252892e-02 -6.48789406e-01 -9.86206055e-01
4.95679498e-01 -1.06282270e+00 5.55974543e-01 -4.10518050e-01
-1.30573714e+00 4.71497834e-01 -3.59165341e-01 -9.10191894e-01
1.85219534e-02 3.88876274e-02 -3.02902728e-01 1.11493576e+00
-8.82902443e-01 -1.16090333e+00 -1.63658500e-01 3.34411025e-01
-4.21820372e-01 -1.52934656e-01 9.50983644e-01 2.38633379e-01
-5.80981433e-01 1.04868841e+00 5.36787748e-01 3.32706094e-01
8.35722983e-01 -1.00726354e+00 9.09106016e-01 1.04982090e+00
-2.41945926e-02 1.08981025e+00 7.81148434e-01 -9.61839914e-01
-1.13941145e+00 -1.14213824e+00 1.47597957e+00 -9.16127741e-01
4.55385596e-01 -3.99142534e-01 -1.15630937e+00 1.07755339e+00
-5.99779822e-02 -4.76318896e-01 1.15758812e+00 1.81717157e-01
-5.82773983e-01 -1.29010960e-01 -1.73612046e+00 4.42030817e-01
8.95796895e-01 -9.39410329e-01 -5.79106927e-01 2.11277753e-01
4.43577051e-01 5.55554219e-02 -1.18429494e+00 -7.75565347e-03
6.10279620e-01 -1.01327956e+00 7.82623112e-01 -9.41310346e-01
4.62636439e-04 -2.68440396e-01 -4.06323262e-02 -9.13437545e-01
-2.02796578e-01 -8.85925591e-01 4.68621254e-01 1.73357201e+00
1.46337494e-01 -9.66935039e-01 1.10993981e+00 1.57761645e+00
4.63604271e-01 -2.89921537e-02 -8.85638177e-01 -8.01501930e-01
2.33191341e-01 -7.46778250e-02 1.32687092e+00 1.55151188e+00
2.20112383e-01 -1.00141793e-01 -4.77288961e-01 1.84458151e-01
5.56317925e-01 -2.21897095e-01 9.94193017e-01 -1.59377670e+00
6.41226709e-01 -3.07698935e-01 -5.95925868e-01 -1.72922328e-01
2.34468132e-01 -8.97034407e-01 -5.62716722e-01 -1.37823987e+00
3.11621130e-01 -5.71604967e-01 1.33931220e-01 4.29710567e-01
-6.05026335e-02 -2.77369972e-02 1.44959971e-01 5.07558405e-01
-2.82271564e-01 4.16310012e-01 6.15872920e-01 3.94226387e-02
-2.94581562e-01 -3.07289301e-03 -1.29999399e+00 3.81054401e-01
8.97981346e-01 -9.85059917e-01 -2.69309044e-01 -4.41991329e-01
4.57149833e-01 -3.41628700e-01 3.10976058e-01 -9.93216097e-01
4.21108454e-01 -1.70718789e-01 1.63664505e-01 -1.44252881e-01
-1.75522193e-01 -1.02937746e+00 7.56339073e-01 3.80279362e-01
-5.44324875e-01 3.07661235e-01 2.02014044e-01 4.32260007e-01
-1.56498432e-01 -2.78610319e-01 6.77569330e-01 -9.24529135e-02
-1.87553883e-01 3.85677814e-01 -2.93400705e-01 3.10637921e-01
1.12993109e+00 -5.60429871e-01 -4.14563686e-01 -1.80933028e-01
-5.09435594e-01 2.68029243e-01 1.39412785e+00 3.56079906e-01
1.24545828e-01 -1.14451826e+00 -8.69689584e-01 1.47151932e-01
3.05952966e-01 -5.15704274e-01 2.72545278e-01 1.74167797e-01
-6.08581960e-01 4.23783988e-01 -3.93993050e-01 5.95617220e-02
-1.59436846e+00 6.69623494e-01 1.68920025e-01 -1.40671909e-01
-5.49068749e-01 4.24554139e-01 -1.84353873e-01 -5.97130775e-01
-1.42434137e-02 1.21276438e-01 1.42196000e-01 2.08493426e-01
6.22515082e-01 7.10244238e-01 2.00753614e-01 -9.96458411e-01
-4.60974783e-01 1.06481448e-01 -4.00303483e-01 -5.65070584e-02
1.18039203e+00 -3.70785952e-01 -4.59997863e-01 -4.21528742e-02
1.40566766e+00 5.22424817e-01 -5.82235456e-01 -1.39610633e-01
4.94445622e-01 -9.42672729e-01 -3.08538944e-01 -6.47015333e-01
-7.80773282e-01 3.56938750e-01 4.63113993e-01 5.31735957e-01
9.13718522e-01 -3.87489825e-01 7.02822506e-01 2.49689907e-01
1.39636174e-01 -1.13936532e+00 -6.84051514e-01 -5.30479811e-02
6.15849555e-01 -1.00688612e+00 4.87854540e-01 -5.75228155e-01
-5.25991678e-01 7.58382678e-01 2.41879031e-01 4.58138049e-01
2.03706682e-01 -8.11986625e-03 1.43176556e-01 -1.00397088e-01
-1.21278942e-01 4.80475396e-01 4.38659899e-02 8.26657057e-01
4.17813718e-01 3.19405317e-01 -6.45622671e-01 5.40501952e-01
-6.73206270e-01 4.36351597e-02 7.38968194e-01 9.64269519e-01
-4.13348787e-02 -1.55464113e+00 -3.94463539e-01 6.72839403e-01
-6.06318533e-01 4.92965095e-02 -1.09060574e+00 8.49632502e-01
-6.70390129e-02 8.29235911e-01 2.69451421e-02 -2.45038956e-01
5.69512784e-01 1.89219430e-01 -9.27178655e-03 -4.55150902e-01
-9.73812938e-01 -8.27241004e-01 2.64253736e-01 -4.58128899e-01
-3.84204358e-01 -9.75862563e-01 -8.29223096e-01 -9.98723149e-01
-1.13934621e-01 5.54197788e-01 6.20441973e-01 5.74407935e-01
6.66212261e-01 -1.68765008e-01 4.79235530e-01 1.55615985e-01
-4.79722381e-01 -3.72952491e-01 -8.37663114e-01 9.60379422e-01
1.87058344e-01 1.00585017e-02 -3.32157969e-01 1.54325336e-01] | [6.152437210083008, 6.914242267608643] |
f822b1c7-ae01-4ec6-91be-2a751abfd6fe | an-investigation-for-implicatures-in-chinese | null | null | https://aclanthology.org/W14-2603 | https://aclanthology.org/W14-2603.pdf | An Investigation for Implicatures in Chinese : Implicatures in Chinese and in English are similar ! | null | ['Janyce Wiebe', 'Lingjia Deng'] | 2014-06-01 | null | null | null | ws-2014-6 | ['implicatures'] | ['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] | [-7.469485759735107, 3.5449655055999756] |
c36be268-93c8-44ed-9130-2a5e03a1d5e7 | skit-s2i-an-indian-accented-speech-to-intent | 2212.13015 | null | https://arxiv.org/abs/2212.13015v1 | https://arxiv.org/pdf/2212.13015v1.pdf | Skit-S2I: An Indian Accented Speech to Intent dataset | Conventional conversation assistants extract text transcripts from the speech signal using automatic speech recognition (ASR) and then predict intent from the transcriptions. Using end-to-end spoken language understanding (SLU), the intents of the speaker are predicted directly from the speech signal without requiring intermediate text transcripts. As a result, the model can optimize directly for intent classification and avoid cascading errors from ASR. The end-to-end SLU system also helps in reducing the latency of the intent prediction model. Although many datasets are available publicly for text-to-intent tasks, the availability of labeled speech-to-intent datasets is limited, and there are no datasets available in the Indian accent. In this paper, we release the Skit-S2I dataset, the first publicly available Indian-accented SLU dataset in the banking domain in a conversational tonality. We experiment with multiple baselines, compare different pretrained speech encoder's representations, and find that SSL pretrained representations perform slightly better than ASR pretrained representations lacking prosodic features for speech-to-intent classification. The dataset and baseline code is available at \url{https://github.com/skit-ai/speech-to-intent-dataset} | ['Kumarmanas Nethil', 'Swaraj Dalmia', 'Shangeth Rajaa'] | 2022-12-26 | null | null | null | null | ['spoken-language-understanding', 'intent-classification', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'speech'] | [ 1.96902335e-01 3.91919643e-01 -7.44040012e-02 -1.13093841e+00
-1.32485068e+00 -6.58917844e-01 2.72437215e-01 -4.74904805e-01
-2.74603993e-01 4.09999400e-01 1.10261810e+00 -5.27739644e-01
4.15150642e-01 -4.58996743e-02 -5.07531166e-01 -2.70403355e-01
1.58929333e-01 7.18714833e-01 -4.56629723e-01 -5.12628257e-01
3.20726521e-02 -7.69849569e-02 -1.09927690e+00 7.68200994e-01
4.05898273e-01 9.35549259e-01 5.71480751e-01 1.26457095e+00
-1.67669073e-01 1.32472646e+00 -6.32578790e-01 -1.34989442e-02
9.30832922e-02 -4.29851085e-01 -1.25070846e+00 -5.02436236e-02
7.45272338e-02 -5.19911289e-01 -6.56085849e-01 4.27133024e-01
7.93327153e-01 2.62068391e-01 3.82089883e-01 -8.46956491e-01
-7.56867766e-01 9.87632751e-01 1.40714630e-01 4.31445092e-01
9.26287711e-01 1.42783463e-01 1.44420183e+00 -9.04417157e-01
4.07037258e-01 1.33603072e+00 2.93169349e-01 8.63346100e-01
-1.15637982e+00 -5.44743955e-01 2.26378947e-01 1.98941648e-01
-1.29297495e+00 -1.48714209e+00 7.96915293e-01 -6.92398921e-02
1.61274195e+00 6.02635503e-01 1.15747519e-01 1.68545961e+00
-1.54132277e-01 1.34868419e+00 8.96302283e-01 -4.36043233e-01
1.94735005e-01 1.24862172e-01 5.83602726e-01 2.10010588e-01
-1.03385663e+00 5.23229539e-02 -1.16957462e+00 -5.80413938e-02
1.17732801e-01 -2.38291234e-01 -5.56459546e-01 6.62568688e-01
-1.09147930e+00 8.02614033e-01 9.95632708e-02 1.15243077e-01
-4.60425735e-01 -1.14844523e-01 7.50328720e-01 7.22902298e-01
6.97533786e-01 1.36112377e-01 -8.98864031e-01 -7.63144672e-01
-7.45134532e-01 -2.01712479e-03 1.23096085e+00 1.02832329e+00
5.89208841e-01 2.01454610e-01 7.19977617e-02 1.45620251e+00
3.80021453e-01 4.15658921e-01 9.31783855e-01 -9.04876530e-01
7.52093434e-01 3.54416557e-02 -2.13374987e-01 -3.10324907e-01
-3.45209599e-01 6.08069897e-02 -5.38872898e-01 -3.71693641e-01
1.48125470e-01 -5.85863352e-01 -8.29514205e-01 1.51565385e+00
-2.12660179e-01 1.57228202e-01 6.63695514e-01 1.00853479e+00
1.16578233e+00 1.32756984e+00 -2.47303128e-01 -4.46266353e-01
1.27691960e+00 -1.19039416e+00 -9.28145170e-01 -7.69710720e-01
7.98833787e-01 -9.17432666e-01 1.31996906e+00 3.49862665e-01
-1.01815748e+00 -5.50975561e-01 -6.89176261e-01 -3.50714922e-01
-1.38417050e-01 -3.66755039e-03 4.17286128e-01 5.57924449e-01
-1.51840115e+00 -2.08774775e-01 -6.32008374e-01 -4.30012792e-01
-1.97087482e-01 3.31563681e-01 -3.66563797e-01 7.44896084e-02
-1.44420481e+00 6.86193109e-01 1.36132225e-01 8.92586783e-02
-7.94508994e-01 -4.62896526e-01 -1.10505903e+00 5.71786240e-02
1.96986228e-01 1.26576070e-02 2.09361959e+00 -1.09569490e+00
-2.10468650e+00 7.62222826e-01 -8.25366855e-01 -8.05612385e-01
-4.43677306e-02 -3.70569438e-01 -4.42526877e-01 -1.28720760e-01
-3.81018519e-02 6.93971992e-01 6.11286342e-01 -8.59266400e-01
-5.76314628e-01 -3.31640422e-01 -2.11327001e-01 5.56797445e-01
-3.43761854e-02 4.74498361e-01 -1.83306620e-01 -5.22597194e-01
1.34357074e-02 -9.01153803e-01 2.51300987e-02 -1.01511800e+00
-3.55163515e-01 -4.81172562e-01 8.75574648e-01 -1.24058127e+00
1.21764469e+00 -2.34724808e+00 -1.03124557e-02 -4.14196730e-01
-3.46824467e-01 -8.79664905e-03 -2.84503639e-01 6.39200807e-01
-2.34138578e-01 7.47489706e-02 3.37679721e-02 -7.66495347e-01
4.24883850e-02 3.81009370e-01 -9.03235912e-01 2.60045603e-02
1.40207306e-01 8.22307229e-01 -4.69853789e-01 -1.26432702e-01
3.49032462e-01 2.53656626e-01 -5.48898876e-01 8.69469523e-01
-2.24067673e-01 4.82715577e-01 -1.06801011e-01 4.53367203e-01
1.58238754e-01 1.59202293e-01 2.61879116e-01 1.90315217e-01
-3.39429490e-02 1.59475040e+00 -6.52542353e-01 1.72229218e+00
-1.07911193e+00 8.59921038e-01 4.69926000e-01 -1.06639254e+00
1.08702528e+00 8.75406742e-01 1.06387049e-01 -6.59722984e-01
-4.03668135e-02 2.20529363e-01 1.15958586e-01 -2.67408937e-01
6.76164746e-01 -1.25968084e-01 -4.68758464e-01 6.16940379e-01
2.74219781e-01 -2.18086079e-01 -3.10247540e-01 1.47909984e-01
1.16746557e+00 -4.12785113e-01 4.59973186e-01 -8.05783086e-03
4.27460790e-01 -9.71502736e-02 4.06026036e-01 7.12158084e-01
-3.08826715e-01 7.55601525e-01 2.64371067e-01 -5.25286794e-01
-7.04818606e-01 -8.73103917e-01 -2.73987874e-02 1.82538283e+00
-5.39753854e-01 -5.34969270e-01 -6.79438949e-01 -6.21803701e-01
-3.78158510e-01 1.28816080e+00 -7.90139511e-02 2.86347568e-01
-7.14528024e-01 -2.06235513e-01 6.52522385e-01 5.32845080e-01
1.52926832e-01 -1.37386382e+00 2.35753208e-01 4.27497923e-01
-6.85994148e-01 -1.26463604e+00 -8.66837740e-01 5.24841785e-01
-3.87310654e-01 -2.49614686e-01 -3.22162986e-01 -9.25569475e-01
8.02606717e-03 5.06964251e-02 9.87739027e-01 -3.83900285e-01
1.36722162e-01 4.05912548e-01 -7.10069001e-01 -4.16691303e-01
-9.08006489e-01 3.01168025e-01 4.02351052e-01 -5.15523255e-02
7.66781807e-01 -4.49198872e-01 -7.62937665e-02 2.79251873e-01
-2.00319991e-01 7.78155997e-02 3.47739309e-01 8.54955256e-01
1.69765487e-01 -4.16462868e-01 1.02975774e+00 -5.42925000e-01
7.29511321e-01 -3.75445306e-01 -1.56902760e-01 -8.29823837e-02
-2.68719066e-02 -1.32444561e-01 8.58699441e-01 -2.15527534e-01
-1.20863390e+00 2.12561622e-01 -8.35512042e-01 -2.83400953e-01
-6.60851419e-01 6.18165195e-01 -2.64427364e-01 6.86684787e-01
3.64036083e-01 6.78280711e-01 1.15081370e-01 -5.04685044e-01
3.40889633e-01 1.82241476e+00 5.14039099e-01 -2.91965961e-01
-4.40902710e-02 -2.53241152e-01 -1.25003743e+00 -1.44208527e+00
-9.55606282e-01 -8.74871910e-01 -4.14916158e-01 -1.02040313e-01
7.97287166e-01 -1.29900277e+00 -7.51195550e-01 3.31998885e-01
-1.50483596e+00 -7.00422406e-01 -7.71186575e-02 5.37077308e-01
-9.55018342e-01 2.38330513e-01 -1.05245256e+00 -1.23280871e+00
-6.74733460e-01 -1.34007490e+00 1.09133291e+00 -6.35297000e-02
-9.00648296e-01 -7.81686366e-01 -1.55030578e-01 1.02390647e+00
3.56420994e-01 -9.48203385e-01 4.39094156e-01 -1.26921690e+00
-2.16263473e-01 -5.52817732e-02 1.90082997e-01 7.45472252e-01
3.42790008e-01 -5.00891566e-01 -1.55669081e+00 -1.79614082e-01
3.08080584e-01 -6.93405151e-01 6.23202026e-01 5.29411614e-01
9.57445562e-01 -8.63930881e-01 1.19926378e-01 2.93106377e-01
4.51271027e-01 4.67114657e-01 5.34564972e-01 -1.37353569e-01
4.16997075e-01 7.46793270e-01 4.91698682e-01 2.68834144e-01
6.45868957e-01 6.62954926e-01 -1.68171361e-01 3.04723591e-01
-9.37189311e-02 -2.19744429e-01 1.21055377e+00 1.61730587e+00
4.21589583e-01 -5.48352242e-01 -1.05886304e+00 6.71562970e-01
-1.68943286e+00 -8.85281384e-01 3.52983087e-01 1.80483282e+00
1.10493529e+00 1.82619281e-02 1.50265306e-01 -9.50590521e-02
5.82907259e-01 5.74273825e-01 -2.75156468e-01 -9.54150319e-01
2.29999691e-01 5.67227742e-03 2.74128795e-01 1.10974872e+00
-9.47447836e-01 1.32495308e+00 6.35499954e+00 5.74272633e-01
-1.14223230e+00 2.97485322e-01 9.44559634e-01 -2.30156288e-01
-1.54405579e-01 -1.42907813e-01 -1.07250690e+00 4.33031410e-01
1.78927433e+00 -1.64696485e-01 8.05722535e-01 1.19696307e+00
5.84622264e-01 2.00177521e-01 -1.33575428e+00 1.03572714e+00
2.14224800e-01 -1.08380222e+00 -1.78150564e-01 -2.15897962e-01
4.98867854e-02 5.30352056e-01 -1.65402800e-01 7.32024908e-01
4.89689201e-01 -1.11079705e+00 6.33220613e-01 1.75975397e-01
7.67162323e-01 -7.35836387e-01 8.67852211e-01 5.18323779e-01
-1.00547028e+00 -6.08146470e-03 -1.41848430e-01 -4.25243139e-01
6.23750091e-01 1.90018147e-01 -1.65797055e+00 1.52380064e-01
6.61669493e-01 5.75270772e-01 5.53703308e-03 -6.45898134e-02
-8.71657878e-02 1.41450989e+00 -3.70412946e-01 -3.50434184e-01
1.88921079e-01 -6.58510029e-02 6.22218311e-01 1.45384598e+00
6.52890205e-02 2.51621395e-01 3.77675265e-01 4.89570320e-01
-2.07884133e-01 2.35492975e-01 -8.99723291e-01 -4.94733453e-01
5.30134201e-01 7.69462228e-01 -1.19093224e-01 -2.61787534e-01
-5.77449262e-01 1.31168079e+00 3.97688121e-01 4.31399912e-01
-4.21820998e-01 -2.50931710e-01 1.11682582e+00 -1.36928812e-01
-6.25164583e-02 -5.20188272e-01 -2.61424989e-01 -9.99308765e-01
-3.73831503e-02 -1.22280312e+00 2.32429266e-01 -9.17640269e-01
-1.36187088e+00 8.40684295e-01 -3.07228804e-01 -6.53307557e-01
-9.80305612e-01 -6.40479386e-01 -7.23969698e-01 1.05430055e+00
-1.23116994e+00 -1.01087344e+00 2.32863963e-01 2.65100986e-01
1.70757902e+00 -4.69341129e-01 1.12390566e+00 -3.09209712e-02
-5.16281486e-01 5.76680541e-01 -2.64586598e-01 4.47083116e-01
9.33616519e-01 -1.10873151e+00 8.58700693e-01 5.39538443e-01
2.11050302e-01 6.58526778e-01 6.02894366e-01 -4.48507667e-01
-1.40994954e+00 -8.06484938e-01 1.15273690e+00 -5.99645078e-01
8.31200123e-01 -7.96271086e-01 -8.49366844e-01 1.34834230e+00
7.66774178e-01 -4.38435763e-01 1.03699982e+00 5.51978409e-01
9.16130317e-04 -1.04208022e-01 -5.13304889e-01 6.40509188e-01
8.79307210e-01 -1.03274274e+00 -1.07741547e+00 4.18954790e-01
1.38733006e+00 -3.33202600e-01 -6.90135777e-01 9.67162177e-02
3.60976934e-01 -7.15362728e-01 7.02401459e-01 -6.20744705e-01
2.67022192e-01 2.22593352e-01 -7.27348685e-01 -1.40481734e+00
-4.52009626e-02 -9.15934563e-01 5.12969531e-02 1.38845229e+00
9.08291221e-01 -7.10269809e-01 5.62652111e-01 4.83408093e-01
-5.13902605e-01 -4.31892455e-01 -1.12640011e+00 -4.86792386e-01
-2.41511703e-01 -9.95352805e-01 4.32068676e-01 7.76204109e-01
4.10533428e-01 9.45464492e-01 -5.14212549e-01 3.23156029e-01
5.68969473e-02 -7.85776302e-02 8.18359613e-01 -6.09882414e-01
-3.93566459e-01 -5.19236475e-02 -2.16457039e-01 -1.64585507e+00
7.30637074e-01 -9.47526515e-01 5.84478319e-01 -1.32639837e+00
-3.65088433e-01 -2.14404732e-01 8.77065137e-02 7.73618102e-01
-1.49785811e-02 -3.34655255e-01 8.10994133e-02 6.68674856e-02
-3.66760850e-01 7.52671897e-01 5.51197708e-01 -1.76267430e-01
-5.04000604e-01 2.68629700e-01 -6.12892091e-01 5.73486567e-01
1.08339107e+00 -1.63471356e-01 -4.12405074e-01 -3.16258252e-01
-4.32525784e-01 6.57640398e-01 -1.65809482e-01 -7.55092025e-01
1.84108913e-01 -8.62879902e-02 -1.13776937e-01 -8.21863472e-01
8.19203854e-01 -3.15331101e-01 -4.55729038e-01 -3.24538425e-02
-7.57554352e-01 -2.75712162e-01 1.19648322e-01 1.84164509e-01
-4.68228906e-01 -2.84715652e-01 4.89146471e-01 -1.76229849e-01
-9.15193856e-01 -1.28007814e-01 -1.10241032e+00 2.35850979e-02
4.56883341e-01 1.52687758e-01 -1.92406237e-01 -1.20135093e+00
-8.47997487e-01 2.52477437e-01 2.73394324e-02 8.45478475e-01
6.12483323e-01 -7.98820615e-01 -8.77723515e-01 4.31194663e-01
9.16443169e-02 5.58621027e-02 3.23817074e-01 6.62286162e-01
-1.23266503e-01 7.56614745e-01 3.37758690e-01 -5.79106390e-01
-1.35252607e+00 1.46005407e-01 2.27115944e-01 1.12683244e-01
-4.53946441e-01 1.06875443e+00 3.91117543e-01 -1.16941273e+00
4.82664645e-01 -3.30305576e-01 1.07506914e-02 -1.35890797e-01
7.83246934e-01 5.54431835e-03 5.38245812e-02 -7.77733445e-01
-4.61060762e-01 -2.91214824e-01 -5.60679972e-01 -5.36721170e-01
1.32149231e+00 -5.49677014e-01 1.55408755e-01 8.52450728e-01
1.35836363e+00 5.01398593e-02 -8.21967602e-01 -5.22379130e-02
7.98297301e-02 6.23915382e-02 1.60492271e-01 -6.89020514e-01
-4.41703439e-01 8.70639682e-01 2.02928171e-01 2.79537469e-01
8.79623353e-01 3.38079810e-01 1.12626517e+00 8.96225095e-01
1.43464699e-01 -1.14505243e+00 3.47812213e-02 1.20236850e+00
1.30484223e+00 -1.39517617e+00 -7.06529260e-01 -2.13306457e-01
-1.32320714e+00 9.03762877e-01 5.94754100e-01 2.10676000e-01
4.61178809e-01 6.30499184e-01 7.78629363e-01 2.34181717e-01
-1.16454601e+00 -1.81472197e-01 -6.05085678e-02 5.17581046e-01
8.39333713e-01 3.02818358e-01 1.82198137e-01 9.87226248e-01
-9.57617700e-01 -3.04105997e-01 6.67678118e-01 6.96480691e-01
-4.44545597e-01 -9.59772110e-01 -3.51186752e-01 5.32107949e-01
-4.26975638e-01 -4.57290471e-01 -6.85633123e-01 6.52318448e-02
-7.55882859e-01 1.53034842e+00 2.24005863e-01 -6.53575182e-01
2.83426046e-01 7.99075961e-01 -3.45918298e-01 -9.74014044e-01
-4.21233952e-01 2.23694682e-01 8.45544577e-01 -5.64418137e-01
7.80114904e-03 -6.92136645e-01 -1.33276951e+00 -1.84482798e-01
-1.78891152e-01 2.59036541e-01 6.19590104e-01 9.77763534e-01
4.70642865e-01 4.23227340e-01 1.03239453e+00 -9.69686866e-01
-5.71298182e-01 -1.46526504e+00 -4.19304132e-01 -4.19772230e-02
6.36571825e-01 7.19939619e-02 -6.57473564e-01 2.27841049e-01] | [14.070131301879883, 6.988542556762695] |
502920b0-1814-41ea-8961-713e30705432 | improving-accuracy-and-explainability-of | 2209.09102 | null | https://arxiv.org/abs/2209.09102v1 | https://arxiv.org/pdf/2209.09102v1.pdf | Improving Accuracy and Explainability of Online Handwriting Recognition | Handwriting recognition technology allows recognizing a written text from a given data. The recognition task can target letters, symbols, or words, and the input data can be a digital image or recorded by various sensors. A wide range of applications from signature verification to electronic document processing can be realized by implementing efficient and accurate handwriting recognition algorithms. Over the years, there has been an increasing interest in experimenting with different types of technology to collect handwriting data, create datasets, and develop algorithms to recognize characters and symbols. More recently, the OnHW-chars dataset has been published that contains multivariate time series data of the English alphabet collected using a ballpoint pen fitted with sensors. The authors of OnHW-chars also provided some baseline results through their machine learning (ML) and deep learning (DL) classifiers. In this paper, we develop handwriting recognition models on the OnHW-chars dataset and improve the accuracy of previous models. More specifically, our ML models provide $11.3\%$-$23.56\%$ improvements over the previous ML models, and our optimized DL models with ensemble learning provide $3.08\%$-$7.01\%$ improvements over the previous DL models. In addition to our accuracy improvements over the spectrum, we aim to provide some level of explainability for our models to provide more logic behind chosen methods and why the models make sense for the data type in the dataset. Our results are verifiable and reproducible via the provided public repository. | ['Koray Karabina', 'Jonathan Gold', 'Steven Chang', 'Hilda Azimi'] | 2022-09-14 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 3.35289359e-01 -4.43926424e-01 -3.14161062e-01 -5.57366192e-01
-3.39338928e-01 -6.66194201e-01 4.76771027e-01 -4.01704729e-01
-1.65800944e-01 4.12159294e-01 -8.46336782e-02 -3.68378133e-01
-3.39444906e-01 -9.07268763e-01 -5.44133306e-01 -7.37741709e-01
8.97204354e-02 2.54390121e-01 -1.60183147e-01 -1.68069929e-01
6.26479089e-01 9.75563765e-01 -1.27795947e+00 5.70740163e-01
6.00940347e-01 1.25326681e+00 -8.69017541e-02 8.20068538e-01
-5.62602393e-02 8.31410408e-01 -7.22647727e-01 -3.66232187e-01
4.08503234e-01 -2.60634184e-01 -3.45132560e-01 1.28160864e-01
4.01210755e-01 -4.39333975e-01 -6.87811434e-01 6.19950056e-01
5.49558938e-01 -9.69265476e-02 6.77213550e-01 -1.04803848e+00
-8.77901733e-01 6.17414594e-01 -2.56765753e-01 -1.35454193e-01
2.75125563e-01 2.15917736e-01 6.82054460e-01 -8.70530546e-01
4.85954672e-01 7.71452069e-01 8.45219791e-01 5.57912409e-01
-7.61924565e-01 -7.37302065e-01 -9.55503061e-03 2.83784062e-01
-1.16082382e+00 -4.31029469e-01 7.96636403e-01 -3.78000319e-01
1.03335750e+00 3.90439302e-01 3.45534086e-01 1.39019394e+00
1.17964581e-01 1.12899876e+00 1.28116560e+00 -7.39188433e-01
5.24399728e-02 -1.32899899e-02 6.04042888e-01 7.06440389e-01
1.83833808e-01 4.64282185e-03 -6.41504884e-01 1.80648059e-01
8.53612900e-01 3.57013047e-01 -1.33748531e-01 2.94388503e-01
-9.82056081e-01 3.54248822e-01 -1.26057658e-02 3.83185893e-01
-1.14294656e-01 1.33871928e-01 1.24338098e-01 3.20292324e-01
-7.93682411e-02 4.83616710e-01 -1.98527291e-01 -5.85700512e-01
-9.71968412e-01 1.66381210e-01 9.88798857e-01 1.08737504e+00
2.56033301e-01 1.71269298e-01 -2.68569499e-01 9.59115386e-01
8.08468834e-02 8.68439198e-01 5.64783752e-01 -6.17550910e-01
8.03173006e-01 6.34134352e-01 5.39921112e-02 -9.61507797e-01
-3.97213280e-01 -1.21390179e-01 -8.96724582e-01 1.15141585e-01
5.57563186e-01 -4.71120551e-02 -1.21686411e+00 9.76606190e-01
-7.21722960e-01 -8.97841901e-02 -4.10975590e-02 8.11811388e-01
5.64884484e-01 5.23881674e-01 -5.60219467e-01 1.18853990e-02
9.07103479e-01 -7.59583533e-01 -8.18735182e-01 -3.82610634e-02
2.87011623e-01 -6.91013336e-01 1.08612740e+00 7.61174262e-01
-7.54139006e-01 -6.48104668e-01 -1.16689742e+00 9.70094576e-02
-3.65778059e-01 5.82622886e-01 6.90327704e-01 9.42548692e-01
-4.66552228e-01 7.33647943e-01 -9.30401683e-01 -3.16881806e-01
5.23175597e-01 3.51370782e-01 -1.46989897e-01 -5.40463105e-02
-8.12137485e-01 7.22741961e-01 3.76621075e-02 2.03941599e-01
-3.72187078e-01 -2.59831756e-01 -3.16804916e-01 -3.30529660e-01
-7.59749338e-02 2.88922310e-01 1.01722872e+00 -6.26383126e-01
-1.65221846e+00 6.56395614e-01 -2.83392578e-01 -3.31688821e-01
5.49606502e-01 1.52188689e-01 -9.19216871e-01 -2.04779088e-01
-4.80770409e-01 -8.86106417e-02 9.08534169e-01 -8.59158278e-01
-4.52554643e-01 -4.29996580e-01 -5.71451187e-01 -3.35803539e-01
-4.60099518e-01 2.50965487e-02 -4.81488526e-01 -7.61757314e-01
2.69189090e-01 -9.21926141e-01 3.60862136e-01 -2.71445215e-01
-4.86486703e-01 -5.69124781e-02 1.08765829e+00 -8.74071777e-01
1.24236321e+00 -2.06632948e+00 -2.44107276e-01 4.88824129e-01
-1.50097251e-01 5.30036509e-01 -2.93257851e-02 6.26832664e-01
2.07422331e-01 4.63384204e-02 -1.98593721e-01 -1.91792905e-01
6.14260696e-02 1.97793260e-01 -8.35809290e-01 3.23493928e-01
2.24719211e-01 9.20767546e-01 -4.17504847e-01 1.16606206e-01
1.89810738e-01 1.54805496e-01 8.26121271e-02 1.23488782e-02
-5.28601743e-02 -5.86910881e-02 -4.35598493e-01 1.25829089e+00
5.38385034e-01 -1.10800035e-01 1.05452761e-01 -1.47644758e-01
-1.61934897e-01 8.66298229e-02 -1.21548319e+00 1.07229495e+00
-9.70293507e-02 1.09455907e+00 -2.51662374e-01 -9.73934591e-01
1.49923670e+00 -3.80530320e-02 3.51483166e-01 -8.04377258e-01
1.23762870e-02 5.45353293e-01 1.35536000e-01 -6.54797435e-01
5.00442386e-01 2.09232509e-01 -2.08955612e-02 6.84150696e-01
-1.61958352e-01 -1.90173581e-01 1.65143803e-01 -3.35033476e-01
1.21044505e+00 3.81478816e-02 -2.97206938e-01 3.97133119e-02
1.96931928e-01 -8.31515342e-02 2.83651143e-01 1.06946373e+00
-1.74875304e-01 7.44364381e-01 2.45906085e-01 -5.10450363e-01
-1.03160036e+00 -7.90143371e-01 -5.14317632e-01 7.35923171e-01
1.66485598e-03 -9.85980704e-02 -2.83734798e-01 -2.67772138e-01
5.10154366e-01 6.14691556e-01 -3.75327766e-01 1.18102744e-01
-6.40637875e-01 -7.89211690e-01 1.18824589e+00 1.08983350e+00
8.15777421e-01 -1.09837687e+00 -4.34861511e-01 1.73482850e-01
1.79915398e-01 -1.05893624e+00 -2.34189913e-01 3.45590532e-01
-8.54211509e-01 -9.02034044e-01 -4.50212300e-01 -5.68814874e-01
4.86543030e-01 -1.31294206e-01 4.09459502e-01 -9.89433005e-02
-4.50737000e-01 2.36522570e-01 -5.29818177e-01 -8.29284430e-01
-2.37894773e-01 4.37439829e-02 3.76968592e-01 2.05009431e-01
7.05042660e-01 -3.34093034e-01 -1.43305525e-01 3.97959560e-01
-6.55141354e-01 -6.95137084e-02 6.62643433e-01 8.09371531e-01
2.63526589e-01 -8.05666298e-02 4.01595563e-01 -6.22138441e-01
8.99658203e-01 2.80778315e-02 -5.41064441e-01 4.41576660e-01
-8.18127096e-01 1.22679055e-01 6.94560766e-01 -4.02862728e-01
-6.59755707e-01 -7.74917006e-02 -6.23649620e-02 -3.44937831e-01
-1.02346338e-01 5.35156667e-01 5.69956228e-02 -1.48079261e-01
6.64521694e-01 6.40123665e-01 5.97905107e-02 -5.87617874e-01
-7.14232102e-02 1.50612962e+00 6.93365097e-01 -6.73013985e-01
4.40432847e-01 5.02563298e-01 -9.47375149e-02 -9.23683643e-01
-2.41865948e-01 -1.01310454e-01 -7.97467709e-01 -1.52910456e-01
2.33580694e-01 -4.16926980e-01 -8.63177001e-01 1.23190987e+00
-9.08082247e-01 -4.65342432e-01 -2.50447765e-02 3.77544254e-01
-4.45138574e-01 4.43292677e-01 -7.25274503e-01 -1.25595665e+00
-2.18074396e-01 -9.61193919e-01 9.88651693e-01 3.50426316e-01
-2.98950046e-01 -8.11919928e-01 -3.26859921e-01 3.45463336e-01
4.95596945e-01 2.65461147e-01 7.02053845e-01 -6.36776328e-01
-5.53100824e-01 -7.50969052e-01 -3.40167612e-01 5.82411945e-01
4.06182468e-01 4.57224071e-01 -1.14800131e+00 -2.62599885e-01
-2.94599056e-01 -2.74474859e-01 1.11737561e+00 1.04383148e-01
1.73887515e+00 -1.77492961e-01 -3.14986795e-01 6.76492751e-01
1.15469861e+00 8.60171616e-01 8.72936130e-01 2.77234226e-01
5.42108953e-01 1.35802969e-01 3.24398726e-01 6.06235445e-01
-7.38018304e-02 6.04441106e-01 -8.98886994e-02 2.66222984e-01
-4.63250466e-02 -3.09188753e-01 4.78844583e-01 7.78061271e-01
-5.66112697e-01 -3.41102809e-01 -1.25217414e+00 2.96466947e-01
-1.73697925e+00 -1.14867306e+00 -2.04113185e-01 2.03336167e+00
8.60625625e-01 5.26413172e-02 1.35791987e-01 5.61832666e-01
5.84784746e-01 8.81403238e-02 -7.99633443e-01 -5.50807536e-01
-5.39897978e-01 4.57848936e-01 7.10192859e-01 3.33031952e-01
-9.60119843e-01 7.63156056e-01 6.94817400e+00 5.44509172e-01
-1.78568459e+00 -5.84196031e-01 4.18209583e-01 -1.96226716e-01
7.93238282e-02 -3.58022898e-01 -1.15995824e+00 7.89056897e-01
7.83914566e-01 -7.51815140e-02 5.77667058e-01 6.07518435e-01
1.22841932e-01 2.45897342e-02 -1.37733972e+00 1.33248222e+00
3.82288039e-01 -1.58022082e+00 -1.11180320e-01 1.19504951e-01
5.90455770e-01 -7.34587386e-02 1.75131842e-01 1.79385111e-01
2.15144157e-02 -1.41045082e+00 6.66330338e-01 1.02908170e+00
1.08082962e+00 -3.47719550e-01 7.03798473e-01 4.31368947e-01
-6.24552906e-01 -4.03034478e-01 -1.68042243e-01 -4.37364399e-01
-1.65565938e-01 5.68915308e-01 -8.97481143e-01 4.20968294e-01
4.81546819e-01 8.54512691e-01 -5.32778442e-01 7.18142152e-01
-1.00216135e-01 8.53791475e-01 -2.71830946e-01 -6.38521254e-01
-7.85533637e-02 -1.89130142e-01 2.10866854e-01 1.41617692e+00
4.65627104e-01 2.01041907e-01 -1.67472497e-01 8.90645087e-01
-1.31444931e-01 -3.07662904e-01 -4.87156272e-01 -5.89720607e-01
5.33574402e-01 8.01253319e-01 -4.48321939e-01 -2.39144459e-01
-2.56249487e-01 8.88221443e-01 -2.59627763e-04 3.22612971e-01
-5.91417432e-01 -9.53346968e-01 4.46392804e-01 1.57313213e-01
1.38040915e-01 -5.92139542e-01 -8.74439180e-01 -1.19313717e+00
4.39835995e-01 -9.81664479e-01 5.15532456e-02 -7.20805705e-01
-1.39727569e+00 1.98761329e-01 -3.75755250e-01 -1.04305160e+00
-1.81604415e-01 -1.34967184e+00 -5.06852806e-01 1.12209857e+00
-9.67866182e-01 -9.39603508e-01 -6.10789299e-01 3.49861562e-01
3.75433743e-01 -8.91856611e-01 8.27780962e-01 8.69181380e-02
-7.66815007e-01 8.56030345e-01 5.62790632e-01 8.02076280e-01
7.55341649e-01 -9.00795460e-01 4.04320776e-01 9.12492812e-01
1.88346341e-01 6.54099464e-01 2.87929714e-01 -5.65379798e-01
-1.91584659e+00 -9.12593305e-01 6.66602671e-01 -8.37446749e-01
6.36693299e-01 -3.75690371e-01 -7.77477145e-01 7.68155694e-01
-1.90013409e-01 -1.37149021e-01 5.81629992e-01 -8.27712379e-03
-4.69719023e-01 -3.47470015e-01 -1.11774826e+00 4.19020206e-01
1.07917643e+00 -7.63311982e-01 -4.84062284e-01 1.36142239e-01
-2.11105019e-01 -5.62124372e-01 -8.77289832e-01 3.86996537e-01
1.09920645e+00 -7.64756083e-01 4.89249051e-01 -6.59533620e-01
6.64590836e-01 -2.09404394e-01 -4.65375096e-01 -9.70461667e-01
-3.03017467e-01 -3.30978036e-01 -4.49659288e-01 1.05199909e+00
3.95163000e-01 -7.33130634e-01 8.84371459e-01 1.01982737e+00
-2.02040896e-01 -8.24797690e-01 -6.95927262e-01 -1.12926924e+00
1.02031983e-01 -8.52757812e-01 7.25212693e-01 9.60177839e-01
2.51135677e-01 -3.51183772e-01 -3.85458469e-01 6.76885666e-03
4.82453883e-01 3.62559259e-01 7.80256271e-01 -1.05464792e+00
-4.05766815e-01 -7.31526077e-01 -6.91615522e-01 -1.09769869e+00
-6.43667486e-03 -8.79581571e-01 -2.82802939e-01 -1.36315835e+00
-2.29687523e-02 -5.52880704e-01 -2.39734575e-01 8.98236156e-01
2.40591720e-01 1.18111894e-01 2.68808961e-01 4.84427899e-01
-1.19816817e-01 3.76976058e-02 1.01592028e+00 -5.15456140e-01
-1.28125623e-01 2.36074820e-01 -5.10873795e-01 2.35854164e-01
8.75414193e-01 1.81226060e-01 7.94006977e-03 -6.43382668e-01
-4.01672311e-02 -1.91053376e-01 3.64546448e-01 -8.91061366e-01
4.38779354e-01 -2.54237086e-01 8.58497202e-01 -5.74522913e-01
2.41874501e-01 -4.53198075e-01 -1.66383886e-03 4.22165304e-01
-2.62012929e-01 -3.20996284e-01 2.10032567e-01 2.05404222e-01
-1.71362191e-01 -3.24804664e-01 6.17641270e-01 1.92927271e-01
-8.20767105e-01 9.23901144e-03 -3.95075351e-01 -5.45653105e-01
8.49764645e-01 -6.44652605e-01 -5.71119130e-01 -2.98202813e-01
-5.44752121e-01 -1.23532258e-01 3.42145503e-01 6.84377849e-01
7.46688664e-01 -1.33702874e+00 -6.29278004e-01 7.19046652e-01
1.08953848e-01 -3.12430352e-01 -2.63978392e-01 6.03646636e-01
-6.00163937e-01 8.76291335e-01 -4.62163270e-01 -5.61249614e-01
-1.24661171e+00 6.50309119e-03 4.25905436e-01 2.21037164e-01
-3.88912559e-01 7.83652246e-01 -1.01828730e+00 -3.82684141e-01
4.10275608e-01 -6.18157029e-01 2.48626515e-01 -1.88990384e-01
6.53474331e-01 7.35121369e-01 4.79846716e-01 -1.74555883e-01
-4.78994399e-01 5.78190804e-01 -8.27639848e-02 -8.82646516e-02
1.44364989e+00 5.58455229e-01 -1.54644018e-02 7.17867196e-01
9.39274192e-01 1.12274766e-01 -1.09067464e+00 -7.54376352e-02
1.20147161e-01 -5.89788079e-01 -2.43840545e-01 -1.09277880e+00
-8.13573122e-01 8.90814245e-01 6.88060403e-01 1.76252648e-01
1.07416940e+00 -3.82646501e-01 5.68647623e-01 9.45686638e-01
6.15615964e-01 -1.52755427e+00 -1.24682426e-01 8.11572075e-01
1.01142669e+00 -1.12972867e+00 -6.88329041e-02 4.94237877e-02
-5.40401757e-01 1.66614449e+00 4.71574754e-01 -2.55867839e-02
4.69148993e-01 7.00608730e-01 1.73738897e-01 5.23965806e-02
-5.48132539e-01 2.16190532e-01 1.96685895e-01 6.17850125e-01
6.75580442e-01 2.47018099e-01 -1.85361445e-01 9.40680385e-01
-2.75979191e-01 3.65539223e-01 6.09622002e-01 1.05960429e+00
-1.92037925e-01 -1.42594504e+00 -5.77222824e-01 1.01129830e+00
-8.13299865e-02 2.04459235e-01 -8.74215841e-01 5.42503655e-01
-8.84978008e-03 8.41120720e-01 1.21625915e-01 -7.19929457e-01
5.52743137e-01 4.90118444e-01 7.29925573e-01 -2.14396805e-01
-3.36307645e-01 -5.48086166e-01 -1.52319491e-01 -1.54225573e-01
-3.43556143e-02 -9.34469223e-01 -1.25948477e+00 -5.43548405e-01
-2.09193885e-01 -2.73026466e-01 8.20175588e-01 9.00950253e-01
4.47824270e-01 2.86450386e-01 7.02648878e-01 -6.95400953e-01
-7.33159125e-01 -1.03301883e+00 -8.19150388e-01 3.85874510e-01
7.50375986e-02 -3.37777138e-01 -1.39260262e-01 2.15362772e-01] | [11.869140625, 2.530611991882324] |
77dc80cb-be9a-44ef-ae7e-60b6437e5183 | dynamic-quantized-consensus-under-dos-attacks | 2306.00279 | null | https://arxiv.org/abs/2306.00279v1 | https://arxiv.org/pdf/2306.00279v1.pdf | Dynamic quantized consensus under DoS attacks: Towards a tight zooming-out factor | This paper deals with dynamic quantized consensus of dynamical agents in a general form under packet losses induced by Denial-of-Service (DoS) attacks. The communication channel has limited bandwidth and hence the transmitted signals over the network are subject to quantization. To deal with agent's output, an observer is implemented at each node. The state of the observer is quantized by a finite-level quantizer and then transmitted over the network. To solve the problem of quantizer overflow under malicious packet losses, a zooming-in and out dynamic quantization mechanism is designed. By the new quantized controller proposed in the paper, the zooming-out factor is lower bounded by the spectral radius of the agent's dynamic matrix. A sufficient condition of quantization range is provided under which the finite-level quantizer is free of overflow. A sufficient condition of tolerable DoS attacks for achieving consensus is also provided. At last, we study scalar dynamical agents as a special case and further tighten the zooming-out factor to a value smaller than the agent's dynamic parameter. Under such a zooming-out factor, it is possible to recover the level of tolerable DoS attacks to that of unquantized consensus, and the quantizer is free of overflow. | ['Shengyuan Xu', 'Hideaki Ishii', 'Maopeng Ran', 'Shuai Feng'] | 2023-06-01 | null | null | null | null | ['quantization'] | ['methodology'] | [-1.21184856e-01 2.26287335e-01 -2.29611710e-01 1.62369147e-01
-3.74616951e-01 -6.84273362e-01 3.32492292e-01 1.85174704e-01
-5.65568388e-01 7.23132312e-01 -4.46445763e-01 -3.21362793e-01
1.01797633e-01 -1.06156528e+00 -3.59641612e-01 -1.21699846e+00
-6.13127589e-01 2.41333842e-01 4.99954760e-01 -4.60655749e-01
9.98676419e-02 5.30522048e-01 -1.08053756e+00 -4.88923281e-01
5.36347747e-01 1.05860281e+00 -1.74906746e-01 9.33985651e-01
2.54619390e-01 4.36303586e-01 -1.19900942e+00 2.57560670e-01
6.67812824e-01 -5.41364491e-01 -2.32266515e-01 4.46724594e-01
-1.64185837e-01 -6.06960833e-01 -5.19902170e-01 1.91519082e+00
2.72885978e-01 -2.00918809e-01 3.76790404e-01 -1.86420846e+00
-3.67938966e-01 7.50457287e-01 -3.04886341e-01 2.49010384e-01
1.73577443e-01 2.86684096e-01 6.96479321e-01 3.19674872e-02
6.23740077e-01 1.54078937e+00 4.61981557e-02 5.76039374e-01
-1.39147854e+00 -1.13570654e+00 2.12046221e-01 -7.68818110e-02
-1.66741884e+00 -2.14356020e-01 5.18306136e-01 -7.09909350e-02
4.08598036e-01 3.37097526e-01 5.69669843e-01 3.61293584e-01
8.24942350e-01 -3.69844399e-02 1.05600715e+00 -2.58373052e-01
9.06018436e-01 -1.01596348e-01 -2.84871876e-01 5.57879865e-01
5.56769609e-01 1.70618236e-01 4.67194468e-02 -5.42762756e-01
1.11922896e+00 -5.82056716e-02 -4.48603064e-01 -1.29875019e-01
-1.21085727e+00 9.78374541e-01 3.25991452e-01 2.06735253e-01
-4.98371810e-01 3.88214767e-01 6.19407177e-01 1.12615752e+00
1.93605617e-01 2.06506625e-01 -1.71085790e-01 1.63899079e-01
-3.19951117e-01 1.42049477e-01 1.25994968e+00 9.95431840e-01
7.84899652e-01 2.86704987e-01 3.81648958e-01 -1.46574631e-01
2.69137681e-01 1.15542018e+00 6.39901683e-02 -1.26408434e+00
3.32181782e-01 3.08003932e-01 3.49036932e-01 -1.07990873e+00
-1.95789173e-01 -8.06869939e-02 -1.20973468e+00 6.47877872e-01
5.05145252e-01 -5.85825622e-01 -4.25566226e-01 1.94921458e+00
5.75095832e-01 1.01657704e-01 3.80145848e-01 1.25359547e+00
-2.04679430e-01 1.10999966e+00 -2.71675706e-01 -1.09498763e+00
1.33672762e+00 -1.98246181e-01 -1.20472455e+00 3.08135122e-01
4.44745034e-01 -5.85942864e-01 4.36614007e-01 4.81444657e-01
-6.70049846e-01 1.49218991e-01 -1.42058504e+00 8.75443935e-01
-7.28878006e-02 -7.50197351e-01 6.01958930e-02 7.12483525e-01
-1.01293683e+00 1.68623835e-01 -1.00268972e+00 -2.14356557e-01
-6.49028659e-01 6.31061256e-01 -2.92217851e-01 3.94588619e-01
-1.62674987e+00 6.67416036e-01 5.67425549e-01 6.19027540e-02
-9.33607578e-01 -9.14994627e-02 -6.19578719e-01 -5.65446056e-02
4.80621070e-01 -4.07735854e-01 1.15449846e+00 -1.01698637e+00
-1.82550538e+00 1.36463925e-01 4.21597749e-01 -6.39280021e-01
5.45595884e-01 6.16222560e-01 -4.83806789e-01 5.33814788e-01
5.83766773e-02 9.54047218e-03 9.01494205e-01 -1.11265659e+00
-4.43210304e-01 -3.15376699e-01 3.25598568e-01 1.19833477e-01
-3.72625291e-01 -1.79537162e-02 1.69553518e-01 -9.37600508e-02
2.03123853e-01 -1.05611479e+00 -5.65526187e-01 2.12509483e-01
-2.35271513e-01 -9.41464677e-02 1.18922830e+00 8.01801905e-02
1.14262164e+00 -2.19179320e+00 -1.22444123e-01 4.72540557e-01
2.28218779e-01 2.42257237e-01 -3.16199362e-02 8.41040552e-01
2.68297940e-01 1.60103932e-01 2.09989533e-01 1.18788913e-01
6.00594953e-02 4.88851547e-01 -3.96963805e-01 1.16782677e+00
-2.47137412e-01 -2.05740735e-01 -9.27757263e-01 -2.74077773e-01
1.88093275e-01 3.04081380e-01 -6.97009623e-01 1.53832033e-01
-1.67126536e-01 3.14823329e-01 -9.60890293e-01 2.91816711e-01
9.74706888e-01 1.12959906e-01 3.03140968e-01 5.92973679e-02
-5.24077892e-01 -8.57578367e-02 -1.68265522e+00 7.79123366e-01
-2.01533005e-01 2.50991762e-01 9.30559456e-01 -7.16464579e-01
1.07874823e+00 7.81758130e-01 4.81568784e-01 -4.10549402e-01
6.47357523e-01 2.34409884e-01 1.03704631e-01 -1.88609213e-01
1.66964844e-01 -3.61657321e-01 -3.81461203e-01 3.91733855e-01
-5.11122048e-01 -2.10879073e-01 6.79928735e-02 4.41662610e-01
1.31096578e+00 -1.05687249e+00 4.10233527e-01 -5.31891346e-01
8.52497399e-01 -1.86769411e-01 8.44363272e-01 6.67053342e-01
-6.45512998e-01 -4.97477025e-01 8.26200306e-01 -9.76593345e-02
-1.31474519e+00 -7.46545315e-01 2.83774547e-02 6.04690790e-01
6.19369090e-01 -3.59817535e-01 -8.53180408e-01 -4.98599298e-02
1.35483116e-01 4.88362089e-02 -3.98948491e-01 -3.13491702e-01
-3.04435253e-01 -3.41552347e-01 6.87633038e-01 -2.21068814e-01
8.85471046e-01 -5.78919768e-01 -6.51144981e-01 5.66993833e-01
5.31621948e-02 -1.10313809e+00 -6.89038873e-01 1.70231480e-02
-6.48346305e-01 -1.10593617e+00 -8.74794051e-02 -5.08579195e-01
7.41600037e-01 2.93461829e-01 1.90379560e-01 1.36740908e-01
3.07517529e-01 2.83132493e-01 -2.85096288e-01 -1.05613180e-01
-9.36436057e-01 -2.27554500e-01 5.75140834e-01 5.92255779e-02
-8.07693973e-02 -4.29693282e-01 -6.39989436e-01 4.64571029e-01
-1.30912626e+00 -7.19517171e-01 5.32940589e-02 7.31726170e-01
1.66814774e-01 6.08604550e-01 7.51377404e-01 -6.09527081e-02
7.60818422e-01 -3.45529467e-01 -1.52562463e+00 -4.48387921e-01
-2.89110631e-01 7.09128752e-03 1.10292315e+00 -6.81070507e-01
-4.43503857e-01 -1.40049592e-01 2.53836870e-01 -4.10050482e-01
2.23899081e-01 5.69238737e-02 -4.33190852e-01 -4.31323379e-01
3.77535433e-01 -5.59509695e-02 5.64591765e-01 1.32893309e-01
3.36901248e-01 1.04827011e+00 2.37662047e-01 -1.67097330e-01
9.79693592e-01 7.42430747e-01 3.17220837e-01 -1.02146614e+00
-1.31612539e-03 -8.14870186e-03 -1.83881640e-01 -2.27971479e-01
4.50171679e-01 -7.91695356e-01 -1.73173594e+00 6.18081331e-01
-1.18372667e+00 -1.54629663e-01 3.06091346e-02 6.35447621e-01
-4.91526723e-01 5.34893811e-01 -1.01914024e+00 -1.08752894e+00
-2.46747524e-01 -1.28232682e+00 4.08911675e-01 7.88774937e-02
2.24725515e-01 -8.73824656e-01 3.97744104e-02 -4.34885710e-01
4.32831138e-01 3.12466472e-01 4.06294882e-01 -7.30526865e-01
-8.01441848e-01 -3.63128901e-01 -5.09342318e-03 4.02626365e-01
-5.26579246e-02 1.08018935e-01 -7.68623501e-02 -9.31959391e-01
3.93188715e-01 -7.65682161e-02 9.88532677e-02 1.11958072e-01
3.45093787e-01 -9.29609716e-01 -6.35872111e-02 3.95113200e-01
1.48143589e+00 5.51250160e-01 2.84282595e-01 4.74509001e-01
6.09160401e-03 -5.33535006e-03 5.72832704e-01 9.19340670e-01
1.61416277e-01 4.12948012e-01 1.13653600e+00 4.31418031e-01
6.26387596e-01 1.39134690e-01 6.92939043e-01 9.45716679e-01
5.37990391e-01 -4.58471209e-01 -4.43493932e-01 2.88299382e-01
-1.48384690e+00 -6.84036314e-01 4.29070927e-02 2.33377504e+00
9.78393912e-01 3.53242934e-01 -2.79080607e-02 3.98102641e-01
1.10144603e+00 1.54981151e-01 -4.97456133e-01 -8.79574597e-01
1.55024245e-01 -5.68649769e-01 1.06190038e+00 9.36691761e-01
-6.48937464e-01 7.06206203e-01 5.59689903e+00 6.28722787e-01
-1.39955902e+00 2.43681502e-02 -1.90768749e-01 5.61971605e-01
1.10209510e-01 2.09957749e-01 -7.23744571e-01 6.21973574e-01
1.23875475e+00 -6.27741635e-01 5.61710775e-01 7.43011177e-01
6.54886127e-01 -2.22482279e-01 -7.42289960e-01 5.57906926e-01
-3.29184115e-01 -8.48187387e-01 -1.50680348e-01 4.94296759e-01
5.27798295e-01 -3.07320178e-01 -3.55728902e-02 -2.95195654e-02
4.67364371e-01 -1.90617561e-01 3.47641289e-01 -7.48705938e-02
8.10288370e-01 -9.88439262e-01 8.70625436e-01 6.00933313e-01
-1.21472800e+00 -2.46157989e-01 -6.30088925e-01 -4.86619741e-01
2.13427678e-01 5.54923952e-01 -6.69997752e-01 1.54065832e-01
-6.60961941e-02 2.11799279e-01 2.97643572e-01 8.48119497e-01
3.47844884e-02 3.92865330e-01 -7.46889710e-01 -1.76308066e-01
3.99905741e-01 -4.54963028e-01 9.52874541e-01 8.56776655e-01
1.51897758e-01 6.80861056e-01 8.56345534e-01 3.65116209e-01
-1.05657838e-01 -2.15652376e-01 -5.33409953e-01 1.06179513e-01
1.28840089e+00 7.34051645e-01 -6.37835681e-01 -6.91438198e-01
1.26601234e-01 7.74633527e-01 -3.62831175e-01 3.84450525e-01
-5.83061457e-01 -9.17440116e-01 9.85203385e-01 -1.43265203e-01
-1.56601202e-02 -4.63711977e-01 2.02999130e-01 -1.02885473e+00
-4.37869728e-01 -9.00176942e-01 2.63551146e-01 -2.68669337e-01
-1.09919882e+00 4.90520328e-01 -1.35631308e-01 -1.49098527e+00
-3.18475842e-01 6.07694872e-02 -4.46684152e-01 4.85699236e-01
-1.03575277e+00 -3.04269046e-01 6.35393336e-02 1.09807360e+00
6.45215660e-02 8.70711282e-02 9.32557225e-01 2.55273014e-01
-4.97148752e-01 3.36052239e-01 4.25660610e-01 1.71627104e-01
5.80280423e-01 -8.61782312e-01 -3.01118642e-01 8.34745526e-01
-9.04745519e-01 5.96705496e-01 1.37946427e+00 -5.95292926e-01
-1.78806639e+00 -9.39309716e-01 6.39168859e-01 4.01756942e-01
1.31936443e+00 -2.34570429e-01 -7.56415784e-01 8.59932959e-01
3.07869136e-01 5.80768138e-02 3.75292934e-02 -9.39804316e-01
-3.66352528e-01 -3.57052982e-01 -1.59227037e+00 4.96828228e-01
9.04535726e-02 -4.38442051e-01 -4.15787250e-01 4.17301536e-01
9.67475712e-01 -3.53187859e-01 -1.20869732e+00 -1.97786212e-01
1.74403578e-01 -6.15740716e-01 3.87618929e-01 -2.53698051e-01
-5.75210571e-01 -6.29052222e-01 -1.44747630e-01 -1.47973442e+00
-6.19310364e-02 -1.32309592e+00 3.81308883e-01 9.70277727e-01
-2.67055154e-01 -1.10886419e+00 4.22036201e-01 9.57082734e-02
2.46895716e-01 5.14877513e-02 -1.48775256e+00 -9.46057081e-01
7.05964118e-02 1.54471487e-01 3.30306023e-01 8.28811765e-01
8.00020695e-01 3.87705639e-02 -3.43292058e-01 9.42509472e-01
1.04625237e+00 -5.05599380e-01 7.94377804e-01 -9.68819380e-01
2.15242714e-01 -6.46931380e-02 -7.28937089e-01 -1.03902781e+00
2.59266555e-01 -2.82334208e-01 1.08050399e-01 -9.90723133e-01
-3.44079822e-01 -3.87176216e-01 -1.07395776e-01 3.42599973e-02
4.80391622e-01 1.31935984e-01 5.87376595e-01 4.50331926e-01
-5.95985949e-01 3.60921085e-01 1.21348584e+00 -7.30378646e-03
-4.31790054e-01 1.51920229e-01 1.42917007e-01 6.03168368e-01
9.16527390e-01 -5.57367742e-01 -3.50274354e-01 2.33213454e-01
-1.55150056e-01 1.06451428e+00 1.82961419e-01 -7.02011347e-01
4.98027861e-01 -5.16275704e-01 -5.82533181e-01 -4.00790662e-01
2.59772420e-01 -1.23833644e+00 2.41167590e-01 1.27345514e+00
-4.39139873e-01 1.22010425e-01 -3.82321864e-01 8.45374465e-01
-5.25636375e-02 1.33839160e-01 1.12846959e+00 1.65042534e-01
-1.30276993e-01 1.76935285e-01 -1.20645308e+00 -2.34263197e-01
1.09949362e+00 4.09322008e-02 -4.55500096e-01 -7.43251026e-01
-8.46668720e-01 5.77730358e-01 6.39502287e-01 9.80233848e-02
4.21227068e-01 -1.20192754e+00 -3.80932629e-01 3.73882324e-01
-3.03805441e-01 -4.18686748e-01 7.09098950e-02 7.63161361e-01
-5.78220606e-01 3.48334074e-01 -2.85465211e-01 -5.55359364e-01
-1.20613253e+00 9.98699784e-01 6.60442412e-01 1.31671920e-01
-4.44070369e-01 1.28315955e-01 -3.18500578e-01 1.44761980e-01
4.38776314e-01 -5.16324461e-01 2.43645355e-01 -1.60442322e-01
7.60740757e-01 4.50788945e-01 -2.54182786e-01 -6.71528637e-01
-2.57043868e-01 3.66950154e-01 1.99106157e-01 -4.17497367e-01
7.10002124e-01 -7.07773089e-01 -6.41417623e-01 1.94308028e-01
1.20980799e+00 -8.51054192e-02 -1.22779512e+00 -2.54753172e-01
-4.75095987e-01 -5.98219156e-01 9.16295052e-02 2.49625500e-02
-1.09278536e+00 4.66127485e-01 4.78995770e-01 1.00261164e+00
9.14785206e-01 -4.86391246e-01 6.07829869e-01 5.34975290e-01
8.17027569e-01 -8.77710402e-01 -6.16192631e-03 5.54077685e-01
3.75947267e-01 -6.44980669e-01 -2.82531738e-01 -5.57327092e-01
-2.85813600e-01 1.14480984e+00 4.26691562e-01 -7.00482965e-01
7.34161735e-01 7.06604958e-01 1.05961962e-02 1.82734653e-01
-1.06176543e+00 1.43599063e-01 -8.43715727e-01 7.07056880e-01
-3.16027075e-01 2.84991533e-01 -6.87911034e-01 -2.43731096e-01
-3.56485806e-02 -4.80461493e-02 1.53280544e+00 8.24953616e-01
-1.25203502e+00 -7.92876244e-01 -9.68761444e-01 -3.76648843e-01
-4.31155711e-01 4.33942139e-01 5.24563715e-02 8.13120604e-01
-2.15692475e-01 1.43308294e+00 2.94234753e-01 -1.05768032e-01
2.80144393e-01 -4.10148233e-01 -1.24133177e-01 -2.66879320e-01
-3.11996102e-01 4.87689495e-01 -1.62239432e-01 -5.37617266e-01
-1.95779115e-01 -9.56075117e-02 -1.62582290e+00 -9.49098110e-01
-5.92505038e-01 7.68557966e-01 4.80062574e-01 5.85286856e-01
-1.25787795e-01 3.01771909e-01 1.20788133e+00 -4.39989358e-01
-1.32452810e+00 -6.85961425e-01 -1.17726362e+00 -1.06796911e-02
9.80936348e-01 -4.30988550e-01 -1.06756592e+00 -1.84256926e-01] | [5.218894958496094, 2.6680984497070312] |
943b3a8e-518b-427b-ad2b-385a22465111 | magic123-one-image-to-high-quality-3d-object | 2306.17843 | null | https://arxiv.org/abs/2306.17843v1 | https://arxiv.org/pdf/2306.17843v1.pdf | Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priors | We present Magic123, a two-stage coarse-to-fine approach for high-quality, textured 3D meshes generation from a single unposed image in the wild using both2D and 3D priors. In the first stage, we optimize a neural radiance field to produce a coarse geometry. In the second stage, we adopt a memory-efficient differentiable mesh representation to yield a high-resolution mesh with a visually appealing texture. In both stages, the 3D content is learned through reference view supervision and novel views guided by a combination of 2D and 3D diffusion priors. We introduce a single trade-off parameter between the 2D and 3D priors to control exploration (more imaginative) and exploitation (more precise) of the generated geometry. Additionally, we employ textual inversion and monocular depth regularization to encourage consistent appearances across views and to prevent degenerate solutions, respectively. Magic123 demonstrates a significant improvement over previous image-to-3D techniques, as validated through extensive experiments on synthetic benchmarks and diverse real-world images. Our code, models, and generated 3D assets are available at https://github.com/guochengqian/Magic123. | ['Bernard Ghanem', 'Sergey Tulyakov', 'Peter Wonka', 'Ivan Skorokhodov', 'Hsin-Ying Lee', 'Bing Li', 'Aliaksandr Siarohin', 'Jian Ren', 'Abdullah Hamdi', 'Jinjie Mai', 'Guocheng Qian'] | 2023-06-30 | null | null | null | null | ['image-to-3d'] | ['computer-vision'] | [ 2.28162020e-01 1.10774867e-01 2.16134638e-01 -2.38279641e-01
-8.02912414e-01 -5.35662651e-01 6.38211191e-01 -2.58508086e-01
1.67186305e-01 6.61245465e-01 9.90123823e-02 1.78538635e-03
4.16525491e-02 -9.71410453e-01 -1.02011776e+00 -6.13301814e-01
1.84663162e-01 3.84151608e-01 6.15215562e-02 -1.15073472e-01
3.14415842e-01 6.54806972e-01 -1.63277376e+00 2.14593470e-01
1.02292192e+00 1.15126097e+00 2.89393514e-01 5.43197036e-01
2.14669108e-01 4.32751447e-01 -9.40653682e-02 -3.34751993e-01
6.56695247e-01 -3.26238245e-01 -6.97841406e-01 5.20795047e-01
6.14784181e-01 -3.60426307e-01 -7.75885489e-03 1.15272987e+00
3.63436103e-01 1.08871244e-01 6.14254355e-01 -8.33374918e-01
-7.10349977e-01 -1.30238712e-01 -1.22718179e+00 -2.42014691e-01
4.22414839e-01 5.47286987e-01 7.45791197e-01 -1.26045787e+00
9.29393649e-01 1.52047527e+00 4.79221851e-01 2.36600846e-01
-1.73383009e+00 -5.45806408e-01 1.13134526e-01 -4.24177229e-01
-1.46948600e+00 -4.75288898e-01 1.03692305e+00 -6.32185519e-01
5.29525876e-01 2.13017225e-01 7.07447469e-01 9.21589434e-01
2.46546209e-01 3.86929005e-01 1.29113114e+00 -2.99607933e-01
1.88820034e-01 -4.22569923e-02 -5.42036593e-01 1.03483784e+00
-1.45696141e-02 4.33281988e-01 -5.88215768e-01 -2.93639779e-01
1.49031508e+00 -1.75133288e-01 -4.92231488e-01 -6.97592258e-01
-1.16333997e+00 7.43355393e-01 5.04476964e-01 -2.66061068e-01
-7.13923156e-01 1.80973202e-01 -1.52778894e-01 2.75595915e-02
9.81685936e-01 4.72413182e-01 -2.41367564e-01 2.44782493e-02
-7.33752012e-01 4.30957377e-01 4.23128009e-01 7.56941319e-01
1.06840873e+00 8.84395093e-02 -5.63703664e-02 9.33661878e-01
4.43690836e-01 7.51310289e-01 -1.86928615e-01 -1.62315631e+00
2.18300521e-01 5.76981127e-01 3.99815619e-01 -1.06308532e+00
5.32510430e-02 -3.43499303e-01 -9.17755425e-01 7.57207155e-01
3.43136340e-01 -1.45767555e-01 -1.13179433e+00 1.47954917e+00
6.87653184e-01 2.96229552e-02 -3.18320930e-01 1.12318158e+00
6.31100714e-01 7.17471480e-01 -4.56593573e-01 1.61225572e-01
1.25727355e+00 -1.01518917e+00 -3.88030440e-01 -2.29123935e-01
2.92899203e-03 -8.54234695e-01 1.16915417e+00 5.05034387e-01
-1.62628841e+00 -4.52904999e-01 -8.05398703e-01 -1.85730264e-01
-7.19495444e-03 -1.05285505e-02 4.14724559e-01 2.74121970e-01
-1.08037794e+00 6.47296429e-01 -7.03497410e-01 1.00708447e-01
6.33726776e-01 -3.00684404e-02 -2.74726659e-01 -2.00740591e-01
-9.49274123e-01 4.73554581e-01 6.45270050e-02 1.35144815e-01
-8.96382630e-01 -9.10946965e-01 -9.35135007e-01 -2.31391311e-01
3.70963037e-01 -1.05841362e+00 9.02898669e-01 -1.03586102e+00
-1.65549123e+00 1.24638927e+00 -7.96728209e-02 4.33999486e-02
7.24933922e-01 -1.50271252e-01 1.35969028e-01 2.57829994e-01
1.56464666e-01 9.01477396e-01 1.14646387e+00 -1.80104852e+00
-2.93905973e-01 -2.93048650e-01 1.04186557e-01 4.96036410e-01
4.20516640e-01 -4.62977469e-01 -8.70868146e-01 -9.29342985e-01
1.51601091e-01 -8.16616535e-01 -2.81511396e-01 5.66878140e-01
-5.61319828e-01 3.42385381e-01 4.26629841e-01 -6.61905587e-01
7.33639598e-01 -2.20416856e+00 6.05239987e-01 4.46773708e-01
3.60714138e-01 -3.01330268e-01 -1.11282073e-01 7.98638270e-04
7.08851889e-02 1.81721732e-01 -4.40847576e-01 -6.32343948e-01
-2.18847305e-01 1.68197993e-02 -1.30412459e-01 5.04437923e-01
4.31929976e-01 7.92334437e-01 -1.02449322e+00 -2.55008191e-01
4.05087322e-01 8.45382452e-01 -8.88182461e-01 3.08814853e-01
-4.13685828e-01 9.46150303e-01 -5.79939663e-01 6.61963224e-01
9.41278934e-01 -5.91066420e-01 -8.71845782e-02 -4.14663732e-01
-2.55364090e-01 -3.74216139e-02 -1.22211087e+00 1.95604479e+00
-7.32172430e-01 3.05901766e-01 5.32762587e-01 -3.75102192e-01
9.20700550e-01 -4.47166488e-02 3.47384393e-01 -8.05623412e-01
1.93222001e-01 2.47771248e-01 -6.28020108e-01 -3.30785006e-01
5.19903064e-01 -1.20607212e-01 2.20368177e-01 3.26737493e-01
-3.87342364e-01 -6.82146013e-01 -3.75444256e-02 -1.04765240e-02
5.76894045e-01 5.70446491e-01 -4.90069985e-02 -3.42738122e-01
4.31603283e-01 -7.81448334e-02 4.17464972e-01 4.97913539e-01
4.84566510e-01 1.20694864e+00 4.75402862e-01 -4.71127868e-01
-1.22955894e+00 -1.21913970e+00 -2.59803921e-01 4.99013990e-01
4.90743816e-01 -3.73818539e-02 -6.21652484e-01 -3.67178798e-01
8.05101320e-02 6.07558310e-01 -8.58350277e-01 7.73559287e-02
-4.78266388e-01 -4.53320980e-01 -8.21596608e-02 1.70413509e-01
6.71586096e-01 -8.98536861e-01 -7.28602648e-01 2.75298487e-02
-2.73473650e-01 -8.23600113e-01 -6.72005534e-01 -4.94877249e-02
-8.59192371e-01 -9.60743487e-01 -1.08056283e+00 -5.69575965e-01
9.43006814e-01 1.78139657e-01 1.45437944e+00 2.17621326e-01
-2.98140973e-01 2.44780704e-01 -1.08747497e-01 5.97226247e-02
-3.84950131e-01 -2.12554812e-01 -3.08694184e-01 2.88290262e-01
-5.08731544e-01 -7.55377293e-01 -8.89256060e-01 3.83644164e-01
-8.36636961e-01 6.95731819e-01 3.89357716e-01 9.16559279e-01
9.81727779e-01 -1.53180465e-01 -1.28768682e-01 -7.32423604e-01
3.19966555e-01 -4.01487589e-01 -9.26625311e-01 1.77658405e-02
-2.83791453e-01 1.62811041e-01 3.07494670e-01 -3.30028445e-01
-1.23924065e+00 7.52805248e-02 -1.25477716e-01 -8.46406817e-01
-3.45388129e-02 2.35336825e-01 -1.25694364e-01 -1.78238511e-01
5.71913958e-01 8.32906440e-02 1.81759223e-01 -6.31721795e-01
4.82704073e-01 2.07091317e-01 3.58459234e-01 -8.49676728e-01
8.50601077e-01 7.47861445e-01 -6.03708625e-02 -6.55333519e-01
-9.62970197e-01 1.03110939e-01 -6.46752298e-01 -3.06850910e-01
9.58481133e-01 -9.93102551e-01 -5.86065412e-01 6.82454586e-01
-1.05335510e+00 -8.30363929e-01 -4.43575412e-01 1.00644246e-01
-7.42491186e-01 2.07460806e-01 -4.47019488e-01 -6.34547293e-01
-3.74191821e-01 -1.35452294e+00 1.61306643e+00 2.31108189e-01
-9.68065411e-02 -8.52616966e-01 -4.05798964e-02 3.77485037e-01
4.07822847e-01 6.72647476e-01 9.24489319e-01 5.46513498e-01
-9.99381900e-01 4.14563388e-01 -4.24292862e-01 2.69384503e-01
1.54013932e-01 7.47428536e-02 -7.97113955e-01 -2.13413179e-01
-6.85672909e-02 -5.01202106e-01 6.35135412e-01 5.48425317e-01
1.13933277e+00 -3.22265029e-01 -1.11863241e-01 1.11137259e+00
1.42790902e+00 -9.82838571e-02 6.09368563e-01 1.60304576e-01
8.27161789e-01 5.27177393e-01 5.19563437e-01 5.59090078e-01
3.69019121e-01 7.78364360e-01 6.18652225e-01 -3.82283837e-01
-3.09531212e-01 -4.27974254e-01 -5.02245277e-02 3.02399606e-01
-2.75443703e-01 -1.82126626e-01 -7.47532964e-01 3.79490167e-01
-1.50833750e+00 -6.64335787e-01 4.46389206e-02 2.35262966e+00
9.14172053e-01 1.37813747e-01 -4.87465039e-02 -3.50008845e-01
5.89526653e-01 3.26886833e-01 -7.67200232e-01 -7.85589144e-02
-1.73928976e-01 1.75263837e-01 3.08948040e-01 9.17231977e-01
-7.93896019e-01 8.57484639e-01 5.29938412e+00 7.68215895e-01
-1.26732981e+00 -7.27109164e-02 1.17074692e+00 -3.38093549e-01
-8.86046410e-01 -5.18760756e-02 -4.36020494e-01 4.48838621e-01
2.76588947e-01 1.47030964e-01 6.50647163e-01 5.19712806e-01
3.06224287e-01 -2.45012552e-01 -7.66535997e-01 1.19493413e+00
-1.23926409e-01 -1.80242729e+00 1.65389493e-01 2.09737912e-01
1.11998844e+00 5.45771606e-02 2.10737273e-01 -2.48525605e-01
3.25774074e-01 -9.91612017e-01 1.11564767e+00 7.93704093e-01
1.18985128e+00 -8.19482028e-01 1.31342962e-01 1.34978026e-01
-1.12688386e+00 3.43776554e-01 -7.89878294e-02 2.70487547e-01
4.91999686e-01 7.79429495e-01 6.44180272e-03 6.07713759e-01
9.03814971e-01 8.53251278e-01 -3.38077486e-01 7.14608252e-01
-2.40094617e-01 1.74828786e-02 -3.04328203e-01 4.88719642e-01
2.02755988e-01 -6.43370032e-01 7.16655552e-01 6.20483160e-01
2.68206835e-01 3.03979695e-01 1.90672308e-01 1.55585659e+00
-1.67155817e-01 -1.33815467e-01 -5.88330984e-01 4.01321352e-01
5.02595782e-01 1.15436316e+00 -6.26385510e-01 -1.86780870e-01
-7.80402869e-02 1.15442765e+00 3.96419644e-01 5.19859612e-01
-8.18180382e-01 -4.18892689e-02 6.84759378e-01 4.92708713e-01
3.37835997e-01 -2.12094605e-01 -5.22174895e-01 -1.20681441e+00
1.37367725e-01 -9.22538817e-01 -8.86371061e-02 -1.06033623e+00
-1.25308621e+00 7.50158846e-01 4.42356281e-02 -1.13556135e+00
-5.01111988e-03 -2.71081746e-01 -5.10999501e-01 1.27974987e+00
-1.54184437e+00 -1.01149321e+00 -5.67169785e-01 4.09222394e-01
5.50475836e-01 3.35844964e-01 5.44215739e-01 1.20768189e-01
-4.26751792e-01 1.43594950e-01 1.21846898e-02 -2.47825965e-01
5.27866125e-01 -1.09340882e+00 7.00726032e-01 5.72240174e-01
-2.78674543e-01 3.08696717e-01 4.75473285e-01 -7.94156313e-01
-1.46198332e+00 -1.05966508e+00 1.85334891e-01 -4.05927688e-01
1.35387078e-01 -3.45205367e-01 -9.36773062e-01 3.51008892e-01
1.94379419e-01 8.24112967e-02 2.56227758e-02 -2.68049031e-01
-1.52578712e-01 1.83834165e-01 -1.34876335e+00 7.83984542e-01
1.14395690e+00 -2.54681498e-01 -1.30429119e-01 1.47731557e-01
5.31529844e-01 -8.44383776e-01 -1.04253948e+00 4.20801371e-01
5.43584585e-01 -1.24833274e+00 1.18340194e+00 4.74237204e-02
7.30605006e-01 -4.43263471e-01 -5.28521873e-02 -1.35339487e+00
-3.63522470e-01 -9.69409347e-01 -1.38062790e-01 8.61721754e-01
3.49431545e-01 -4.69942242e-01 7.56178439e-01 4.92303044e-01
-5.21122143e-02 -1.15447390e+00 -6.45174503e-01 -5.05128145e-01
3.94960828e-02 -1.34875864e-01 6.60298288e-01 7.86163926e-01
-6.73125327e-01 2.01309502e-01 -3.97408217e-01 1.56151131e-01
1.01475239e+00 5.34218848e-01 8.11731637e-01 -1.12566495e+00
-3.33772570e-01 -4.66383696e-01 1.66017696e-01 -1.50819016e+00
-2.29220375e-01 -6.47854447e-01 -1.38020981e-02 -1.37972844e+00
-1.14914579e-02 -7.13363290e-01 3.13600093e-01 2.32614353e-01
-9.84308198e-02 6.13468051e-01 8.36729445e-03 1.82558015e-01
-7.00335652e-02 8.71979713e-01 1.81762683e+00 8.75835270e-02
-3.85100633e-01 -2.95567513e-01 -6.85124874e-01 8.18737626e-01
5.62561691e-01 -2.42209211e-01 -3.94813418e-01 -8.56108189e-01
2.87728399e-01 2.17827275e-01 6.48498297e-01 -6.23889804e-01
-2.60113120e-01 -3.59687865e-01 6.35753572e-01 -7.33319700e-01
7.83287585e-01 -5.33211946e-01 4.54045802e-01 2.38313600e-01
-1.16480224e-01 -4.46972810e-03 1.78568915e-01 5.29639542e-01
2.59858053e-02 1.85783148e-01 1.19339013e+00 -3.58267576e-01
-4.17046785e-01 5.38081586e-01 8.20784196e-02 4.16473031e-01
8.14952075e-01 -3.08420479e-01 -4.86276986e-05 -3.58474433e-01
-6.80423141e-01 3.26450348e-01 1.20843232e+00 4.81601447e-01
8.35697353e-01 -1.44925749e+00 -7.77512610e-01 6.40211046e-01
-6.69721067e-02 6.04177833e-01 5.16335130e-01 6.66705489e-01
-8.31841707e-01 -2.26037472e-01 -2.48081133e-01 -8.56171489e-01
-9.08298135e-01 2.95878321e-01 5.67940116e-01 -3.39919887e-02
-9.47303593e-01 9.16050971e-01 5.28341413e-01 -4.79403824e-01
2.13789627e-01 -1.60486981e-01 2.72754580e-01 -2.90278107e-01
4.23169106e-01 2.42672518e-01 -1.96634829e-01 -5.70499301e-01
-1.72494754e-01 9.27261949e-01 -1.88249722e-02 -2.68989474e-01
1.44961250e+00 -3.45390946e-01 -1.42933792e-02 2.10904092e-01
1.06445539e+00 1.97132096e-01 -1.99344242e+00 -1.36674434e-01
-4.85973507e-01 -9.15621102e-01 2.71833688e-01 -7.19205558e-01
-1.44679368e+00 6.12326503e-01 4.16757822e-01 -8.68295580e-02
1.12727988e+00 -1.38823772e-02 6.84577525e-01 -2.41574660e-01
3.95443231e-01 -6.51511908e-01 2.15756699e-01 4.48737979e-01
1.30995095e+00 -1.15627313e+00 1.42080754e-01 -6.04887068e-01
-6.03978217e-01 9.01314855e-01 6.09212995e-01 -2.37491757e-01
7.92962372e-01 2.74390459e-01 7.68158361e-02 -5.63386738e-01
-6.81895018e-01 2.02068925e-01 4.95777667e-01 3.96828562e-01
2.21819833e-01 -2.24781215e-01 2.97477603e-01 3.47145051e-02
-9.64461768e-04 -1.24682732e-01 3.91174138e-01 7.76834488e-01
-9.05230641e-02 -8.88137817e-01 -5.62448263e-01 2.75011033e-01
-1.83686137e-01 -1.10744305e-01 -1.21393099e-01 4.96552318e-01
1.91974491e-01 6.14861250e-01 1.57700673e-01 -9.71137434e-02
4.34197783e-01 -3.88110995e-01 6.52534246e-01 -6.14369512e-01
-3.60754251e-01 4.08750206e-01 1.16932467e-01 -9.35696602e-01
-3.06678325e-01 -4.44157511e-01 -9.32028890e-01 -3.96283150e-01
-1.14732042e-01 -1.87111631e-01 4.34254795e-01 4.85076547e-01
7.19337940e-01 3.87358338e-01 8.43441248e-01 -1.54443431e+00
-1.67786330e-01 -5.23878038e-01 -4.97401208e-01 4.43040043e-01
3.07012230e-01 -8.05688918e-01 -3.95100236e-01 1.09625898e-01] | [9.28408432006836, -3.1358799934387207] |
9231585c-fe6b-4fe4-9095-eb599b46b225 | story-cloze-ending-selection-baselines-and | 1703.04330 | null | http://arxiv.org/abs/1703.04330v1 | http://arxiv.org/pdf/1703.04330v1.pdf | Story Cloze Ending Selection Baselines and Data Examination | This paper describes two supervised baseline systems for the Story Cloze Test
Shared Task (Mostafazadeh et al., 2016a). We first build a classifier using
features based on word embeddings and semantic similarity computation. We
further implement a neural LSTM system with different encoding strategies that
try to model the relation between the story and the provided endings. Our
experiments show that a model using representation features based on average
word embedding vectors over the given story words and the candidate ending
sentences words, joint with similarity features between the story and candidate
ending representations performed better than the neural models. Our best model
achieves an accuracy of 72.42, ranking 3rd in the official evaluation. | ['Anette Frank', 'Todor Mihaylov'] | 2017-03-13 | story-cloze-ending-selection-baselines-and-1 | https://aclanthology.org/W17-0913 | https://aclanthology.org/W17-0913.pdf | ws-2017-4 | ['cloze-test'] | ['natural-language-processing'] | [-1.52577534e-01 -5.52315190e-02 -3.02671313e-01 -4.96335566e-01
-9.85430300e-01 -3.35719287e-01 6.74281895e-01 4.99314696e-01
-6.16847277e-01 2.90794760e-01 9.30098295e-01 3.91803496e-02
-1.75562158e-01 -7.75983751e-01 -3.74339104e-01 -4.18499112e-01
-1.97964489e-01 4.05712992e-01 -1.95622653e-01 -4.38009739e-01
5.39798558e-01 -2.05177918e-01 -1.50279164e+00 1.08021545e+00
4.40077245e-01 7.95850992e-01 1.01003744e-01 5.90605736e-01
1.17501050e-01 9.94222343e-01 -6.86646640e-01 -4.24841702e-01
-2.28723332e-01 -4.44719881e-01 -1.06360710e+00 -4.43217486e-01
6.26534820e-01 -5.12930691e-01 -6.99505329e-01 7.15290010e-01
6.17551684e-01 5.16408741e-01 9.23900843e-01 -7.03888834e-01
-1.43118060e+00 1.15783930e+00 -3.10770601e-01 4.42085475e-01
7.27885246e-01 2.09399555e-02 1.64977705e+00 -1.16171312e+00
7.18565881e-01 9.73659158e-01 7.84024537e-01 7.90645719e-01
-9.69114900e-01 -5.62215507e-01 8.49228352e-02 1.01331079e+00
-1.40441966e+00 -3.47748369e-01 6.12181485e-01 -4.49267358e-01
1.63992667e+00 1.33006558e-01 5.26005387e-01 1.71969783e+00
-8.30261558e-02 8.20896924e-01 8.55495274e-01 -6.17125928e-01
6.10054806e-02 7.96969756e-02 9.50850308e-01 3.71740043e-01
-1.98267661e-02 -1.62238821e-01 -7.89448440e-01 -3.13812971e-01
9.75134596e-02 -2.89078951e-01 -3.23830247e-01 1.60348192e-01
-7.07209468e-01 1.46482384e+00 5.26943386e-01 7.13738143e-01
-2.92846501e-01 2.21780777e-01 7.69720197e-01 2.80466616e-01
7.13915825e-01 6.43686652e-01 -3.01795691e-01 -1.02533668e-01
-9.64264691e-01 6.54092908e-01 4.66933310e-01 4.01286870e-01
1.27394691e-01 -9.94861200e-02 -7.46427357e-01 1.28117943e+00
3.65627587e-01 -5.74448891e-02 1.17852628e+00 -3.15462112e-01
4.86878037e-01 1.66635931e-01 -2.39625454e-01 -1.09272647e+00
-2.90145040e-01 -2.71671861e-01 -3.11851948e-01 -3.17302823e-01
1.25819132e-01 1.37260988e-01 -6.91444635e-01 1.80518675e+00
-4.72649962e-01 2.23098233e-01 1.78673178e-01 7.38730252e-01
1.49787474e+00 6.95496738e-01 4.15576756e-01 -1.12781800e-01
1.46008813e+00 -1.10742605e+00 -9.49440897e-01 -4.62239921e-01
1.17594337e+00 -4.61457312e-01 1.42868459e+00 4.51545477e-01
-1.12481844e+00 -4.24350649e-01 -1.47628856e+00 -3.55845839e-01
-6.77964389e-01 3.76333147e-01 4.52900916e-01 5.05842805e-01
-1.11683905e+00 7.40950406e-01 -2.42120132e-01 -5.97274423e-01
4.30478603e-01 6.46963418e-02 -3.30869496e-01 -1.67598113e-01
-1.66838467e+00 1.29674101e+00 5.93073845e-01 -4.50743854e-01
-1.00838971e+00 -4.04249370e-01 -1.00384271e+00 2.18706131e-01
-3.72357517e-01 -2.84238517e-01 1.16432786e+00 -7.02155948e-01
-1.18239820e+00 1.22373557e+00 -2.06102282e-01 -6.06177628e-01
-1.79054543e-01 -4.80480015e-01 -5.23217380e-01 -4.99754027e-02
1.36323795e-01 3.46991062e-01 5.05875707e-01 -7.25147963e-01
-6.87296763e-02 -2.26547599e-01 -2.22977828e-02 2.39688098e-01
-6.34502769e-01 5.50364137e-01 4.55762535e-01 -7.09330559e-01
6.93411380e-02 -3.18662852e-01 2.32746705e-01 -7.43817866e-01
-1.53947145e-01 -9.18871701e-01 6.01825893e-01 -9.51848924e-01
1.54818952e+00 -2.20718813e+00 -1.12379104e-01 -2.34918565e-01
1.65262833e-01 3.99961889e-01 -4.48286533e-01 7.56209016e-01
-6.11628354e-01 3.13592494e-01 6.45333678e-02 -5.35110831e-01
6.33656606e-02 -1.88197985e-01 -8.29464257e-01 4.66016680e-01
8.60439688e-02 7.73814321e-01 -9.53452289e-01 3.73151079e-02
-1.81190580e-01 4.15267885e-01 -4.62125689e-01 1.87972933e-01
-7.21563399e-02 -3.46159488e-01 -7.15068076e-03 1.43826991e-01
3.67097288e-01 5.03566265e-02 -1.52853072e-01 2.19524875e-01
3.02979171e-01 1.23868525e+00 -5.01370966e-01 1.64544713e+00
-4.72763658e-01 1.02103198e+00 -7.52041578e-01 -5.67427874e-01
1.06701267e+00 8.46465111e-01 -1.54786780e-01 -5.21687627e-01
5.29920280e-01 1.49153769e-01 -5.48308380e-02 -8.07381392e-01
5.82764447e-01 -2.08418652e-01 -1.21835873e-01 6.27848983e-01
5.25106132e-01 -4.16813791e-02 1.48567617e-01 2.56672621e-01
1.13225555e+00 2.99207997e-02 3.39739799e-01 -6.16307184e-02
2.61116028e-01 -3.28249335e-01 -8.54914412e-02 7.20229626e-01
-2.31485039e-01 9.71615136e-01 4.88998622e-01 -4.64589328e-01
-1.01063919e+00 -7.53265977e-01 -2.44842038e-01 1.35256338e+00
-4.58413631e-01 -1.01421034e+00 -7.42215693e-01 -6.54267550e-01
-4.43860769e-01 1.59973967e+00 -1.00666559e+00 -5.51699102e-01
-5.16711831e-01 -5.91725349e-01 8.15239727e-01 7.14765370e-01
2.38754958e-01 -1.12072372e+00 -4.15847003e-01 2.36105710e-01
-4.86274004e-01 -7.69752741e-01 -1.17722072e-01 4.13166910e-01
-5.53084970e-01 -7.44246125e-01 -3.65389049e-01 -1.08239198e+00
-1.54871002e-01 1.44638777e-01 1.03716588e+00 2.22367227e-01
6.11370467e-02 -2.61471897e-01 -7.86031961e-01 -2.09247828e-01
2.74212807e-02 -9.65420678e-02 5.95518388e-02 -2.51549780e-01
1.17082369e+00 -4.93710786e-01 -6.32943511e-02 -4.41308409e-01
-6.25543892e-01 -1.39024958e-01 -6.86187744e-02 1.13966572e+00
9.86803174e-02 -3.77259910e-01 6.79041445e-01 -5.21142483e-01
1.16028869e+00 -9.18645084e-01 3.91216487e-01 8.19434375e-02
-2.89516330e-01 -1.23723209e-01 4.42592472e-01 -5.31719089e-01
-6.70965493e-01 -3.77813488e-01 -4.82857078e-01 -1.95772663e-01
-2.95094132e-01 6.12663805e-01 2.37839893e-02 6.40187919e-01
1.07395732e+00 3.87028940e-02 -6.23041987e-01 -5.92116296e-01
5.02605259e-01 9.09699202e-01 2.44746223e-01 -2.93845147e-01
1.72280356e-01 1.44916922e-02 -6.96392894e-01 -5.70952356e-01
-1.30486369e+00 -5.93997240e-01 -5.12960911e-01 -6.06771074e-02
1.13866842e+00 -8.79881620e-01 -3.51557523e-01 1.60055265e-01
-1.85573351e+00 6.14733212e-02 -1.58269688e-01 7.52703190e-01
-6.54265702e-01 -1.07254824e-02 -9.68141198e-01 -7.19369590e-01
-2.92618781e-01 -7.31260240e-01 8.27158749e-01 -1.84067577e-01
-9.38598990e-01 -9.17982221e-01 4.95684355e-01 2.61948198e-01
3.43523294e-01 1.96547136e-01 1.26066661e+00 -1.36231446e+00
2.50918120e-01 -4.52532500e-01 -1.82247639e-01 2.71334559e-01
-3.01321119e-01 -1.58239916e-01 -1.46356750e+00 -1.36226594e-01
4.35865581e-01 -9.17337596e-01 1.47078860e+00 4.02757257e-01
1.04823720e+00 -3.69307399e-01 4.86199558e-03 4.21742618e-01
1.43209171e+00 3.61775532e-02 7.72961020e-01 5.39617777e-01
5.50186455e-01 5.86268187e-01 2.89335281e-01 3.07314932e-01
2.96878368e-01 6.29605472e-01 1.95030123e-01 5.99894643e-01
-2.96493918e-01 -6.04520679e-01 6.73346698e-01 9.46435988e-01
1.75048336e-01 -4.20548856e-01 -8.77001762e-01 8.78442645e-01
-1.78772104e+00 -1.29166913e+00 -4.96099055e-01 1.79935288e+00
6.45579934e-01 -4.60408628e-02 -8.70967358e-02 4.14708704e-01
7.04841673e-01 5.02136648e-01 1.13019475e-03 -1.12174833e+00
-4.42697316e-01 5.95119238e-01 1.07604645e-01 5.84355533e-01
-1.23965967e+00 1.16099477e+00 7.12818480e+00 1.08579040e+00
-6.83613420e-01 6.25240028e-01 5.16577899e-01 -5.17539442e-01
-2.06137002e-01 -2.39380240e-01 -5.93451083e-01 1.60013482e-01
1.29771113e+00 -2.74722606e-01 3.54586154e-01 8.79347622e-01
-5.62952226e-03 -3.41392979e-02 -1.36591196e+00 1.06155121e+00
9.61994231e-01 -1.15413952e+00 2.15994850e-01 -3.66152495e-01
6.79694712e-01 1.00726195e-01 3.67886186e-01 5.42555273e-01
2.26841331e-01 -1.47668839e+00 9.09121871e-01 4.11098301e-01
4.74229664e-01 -8.43142390e-01 1.02814090e+00 2.42054477e-01
-7.23558187e-01 -6.86297789e-02 -5.88682055e-01 -7.38324523e-01
-1.05426081e-01 1.26724735e-01 -8.80710602e-01 1.27319451e-02
7.08245277e-01 9.43538308e-01 -8.45646918e-01 8.79624844e-01
-8.20197284e-01 8.85239542e-01 2.19646081e-01 -4.73741382e-01
5.19707739e-01 3.77987087e-01 4.96162295e-01 1.36325777e+00
2.14861512e-01 -3.44621949e-04 -2.15769231e-01 8.65587771e-01
-1.27281740e-01 6.13104284e-01 -1.02907348e+00 -1.11883409e-01
3.33269298e-01 8.25509846e-01 -4.24475700e-01 -5.23574591e-01
-1.31978154e-01 8.92432511e-01 8.76729965e-01 2.67105460e-01
-8.22335303e-01 -5.67230701e-01 6.46443069e-01 -1.32558346e-01
2.14899853e-01 -1.97998621e-02 -7.36873209e-01 -1.00237310e+00
5.34086116e-03 -5.60735345e-01 4.63638127e-01 -1.04267108e+00
-1.62517929e+00 1.02033997e+00 2.78657582e-02 -9.82776940e-01
-3.93410802e-01 -6.20854557e-01 -1.02202225e+00 9.54751968e-01
-9.04608428e-01 -1.11211014e+00 1.88696846e-01 4.16481972e-01
8.05179298e-01 -4.11917150e-01 1.33119607e+00 7.78734833e-02
-3.22111607e-01 7.27744699e-01 7.04594776e-02 2.52405554e-01
6.95525050e-01 -9.46738839e-01 5.11842728e-01 5.84585011e-01
4.21055883e-01 6.57562792e-01 7.33914316e-01 -5.75791955e-01
-3.80454034e-01 -9.05472398e-01 1.82752836e+00 -8.00355434e-01
7.31290102e-01 -4.15544093e-01 -1.07652617e+00 7.36085892e-01
8.33522677e-01 -5.55544794e-01 9.93542194e-01 4.14648294e-01
-8.89409065e-01 4.72817123e-01 -1.01803648e+00 5.51887870e-01
9.42445934e-01 -9.82049227e-01 -1.21352398e+00 7.68473804e-01
8.52432311e-01 -5.13660209e-03 -2.21312135e-01 8.33057612e-02
2.27669477e-01 -8.36292565e-01 7.75622725e-01 -1.22534859e+00
1.01419568e+00 2.11241931e-01 -5.07844090e-01 -1.59304881e+00
-6.49704337e-01 1.48456350e-01 -8.00621882e-02 1.06450534e+00
4.92342025e-01 -1.53929800e-01 3.33965868e-01 3.76858592e-01
-1.50953144e-01 -7.19471097e-01 -1.02975917e+00 -7.41932690e-01
5.05716085e-01 -7.40563869e-01 4.31758881e-01 1.05031025e+00
6.07053518e-01 7.50693977e-01 -3.04626524e-01 -2.10697591e-01
2.45207120e-02 -5.14765084e-01 -1.16902284e-01 -8.42888415e-01
-3.61211479e-01 -7.37888753e-01 -5.47895253e-01 -5.77998817e-01
5.68037391e-01 -1.32039058e+00 -5.40033430e-02 -1.77307928e+00
4.51778203e-01 2.67530859e-01 -4.53187525e-01 6.44858539e-01
-2.10711509e-01 4.86165464e-01 1.15067691e-01 -1.71824515e-01
-5.87974846e-01 5.09760737e-01 5.82601368e-01 -2.57351816e-01
8.48513097e-02 -5.16551137e-01 -6.48981869e-01 9.60738003e-01
1.08961391e+00 -6.67516291e-01 -3.00330043e-01 -9.00641263e-01
3.46236259e-01 -3.37095231e-01 3.52472663e-01 -1.05105233e+00
-5.70730753e-02 2.32924372e-01 3.31865132e-01 -5.72028995e-01
4.74535525e-01 -4.67874482e-02 -3.85339588e-01 2.60511041e-01
-1.00947082e+00 9.57146287e-02 1.24781124e-01 1.11652538e-01
-3.09264153e-01 -9.14480448e-01 6.49642766e-01 -4.42646258e-02
-3.69530112e-01 -2.33782098e-01 -6.63499177e-01 4.91202585e-02
7.33479798e-01 -1.54325545e-01 -3.28184098e-01 -7.13068962e-01
-9.98642504e-01 -2.13129401e-01 -4.25634608e-02 7.25537479e-01
9.86716092e-01 -1.72700679e+00 -1.13880551e+00 9.13069323e-02
4.58256364e-01 -8.39193165e-01 2.49052629e-01 3.40222955e-01
-1.98476240e-01 6.13220334e-01 -3.72847263e-03 1.62517875e-01
-1.36490262e+00 6.06147647e-01 1.26644686e-01 -2.88923115e-01
-3.88821602e-01 1.37571609e+00 -7.56986886e-02 -4.47162837e-02
2.54308820e-01 -2.15513781e-01 -7.92739332e-01 3.97152841e-01
1.18515849e+00 4.32992399e-01 4.05796647e-01 -9.93752539e-01
-1.27347589e-01 2.81232864e-01 -3.85143012e-01 -2.63599515e-01
1.51783490e+00 1.34563744e-01 -1.25088975e-01 8.14175248e-01
1.58997691e+00 -3.49995822e-01 -1.85433328e-01 -2.87887752e-01
4.02587742e-01 -6.02099478e-01 3.31773669e-01 -9.19500351e-01
-6.85561240e-01 1.17538738e+00 6.07540727e-01 2.33093977e-01
7.26838946e-01 1.87959597e-01 9.30958629e-01 4.02550578e-01
1.53116137e-01 -1.13649702e+00 1.21165179e-01 1.05394423e+00
1.25011826e+00 -9.35080826e-01 -3.55035126e-01 1.04416963e-02
-7.75751948e-01 1.22673547e+00 6.53342187e-01 -6.81614757e-01
4.26476449e-01 -3.52693796e-01 -8.52409825e-02 -2.78233916e-01
-1.10640991e+00 -3.01283449e-01 4.22253877e-01 6.24735296e-01
9.81321335e-01 1.60928145e-01 -1.02175951e+00 1.31116855e+00
-4.98307616e-01 -4.42051262e-01 5.87160051e-01 3.26438844e-01
-5.49501419e-01 -8.87209117e-01 -2.16750935e-01 5.41515589e-01
-5.14311969e-01 -5.88481843e-01 -6.36460900e-01 3.78267556e-01
3.46186817e-01 1.22960699e+00 2.81187534e-01 -7.23848104e-01
2.71588117e-01 6.66602314e-01 3.05513203e-01 -1.08059001e+00
-9.53436673e-01 -3.52247298e-01 5.16727090e-01 -4.51691449e-01
-4.89979377e-03 -6.67352974e-01 -1.10415125e+00 -1.24803111e-01
-4.20505971e-01 -6.54763132e-02 6.28301799e-01 1.09515333e+00
-6.27714023e-02 3.97466421e-01 4.83680904e-01 -5.17853916e-01
-5.06007731e-01 -1.62191129e+00 -7.64271200e-01 5.30997455e-01
1.82988971e-01 -7.84139276e-01 -3.95729959e-01 -1.67259187e-01] | [11.279988288879395, 8.909173011779785] |
2749e795-e271-47e5-aec7-f1225fe9b9b0 | class-aware-visual-prompt-tuning-for-vision | 2208.08340 | null | https://arxiv.org/abs/2208.08340v4 | https://arxiv.org/pdf/2208.08340v4.pdf | Dual Modality Prompt Tuning for Vision-Language Pre-Trained Model | With the emergence of large pre-trained vison-language model like CLIP, transferable representations can be adapted to a wide range of downstream tasks via prompt tuning. Prompt tuning tries to probe the beneficial information for downstream tasks from the general knowledge stored in the pre-trained model. A recently proposed method named Context Optimization (CoOp) introduces a set of learnable vectors as text prompt from the language side. However, tuning the text prompt alone can only adjust the synthesized "classifier", while the computed visual features of the image encoder can not be affected , thus leading to sub-optimal solutions. In this paper, we propose a novel Dual-modality Prompt Tuning (DPT) paradigm through learning text and visual prompts simultaneously. To make the final image feature concentrate more on the target visual concept, a Class-Aware Visual Prompt Tuning (CAVPT) scheme is further proposed in our DPT, where the class-aware visual prompt is generated dynamically by performing the cross attention between text prompts features and image patch token embeddings to encode both the downstream task-related information and visual instance information. Extensive experimental results on 11 datasets demonstrate the effectiveness and generalization ability of the proposed method. Our code is available in https://github.com/fanrena/DPT. | ['Yanning Zhang', 'Peng Wang', 'Guoqiang Liang', 'Shizhou Zhang', 'De Cheng', 'Qirui Wu', 'Yinghui Xing'] | 2022-08-17 | null | null | null | null | ['general-knowledge'] | ['miscellaneous'] | [ 1.65061489e-01 -8.93212408e-02 -2.54732013e-01 -4.72446591e-01
-7.66721785e-01 -5.59047401e-01 8.20236742e-01 1.12352304e-01
-3.34362328e-01 4.56401646e-01 5.18278539e-01 -5.89939989e-02
-4.49724495e-03 -5.62729776e-01 -8.26291025e-01 -9.65617776e-01
4.40621048e-01 -8.10492039e-02 1.01306975e-01 -1.64663240e-01
4.25782591e-01 2.02560887e-01 -1.45380032e+00 6.97167397e-01
7.39306748e-01 1.09162986e+00 9.21116531e-01 4.03473020e-01
-3.94023925e-01 6.25122666e-01 -3.47283095e-01 -8.80941674e-02
2.62878239e-01 -2.90885031e-01 -4.66553509e-01 1.23846322e-01
4.04817730e-01 -1.89870358e-01 -2.86688566e-01 7.75728762e-01
5.94326317e-01 3.68442267e-01 5.64124048e-01 -1.24333572e+00
-1.20579767e+00 5.44587016e-01 -5.71912289e-01 3.89941931e-01
1.30569004e-02 6.29380822e-01 1.10027516e+00 -1.32713413e+00
5.22420347e-01 1.35371208e+00 8.11099820e-03 6.94396198e-01
-1.17593300e+00 -7.44536996e-01 8.40357900e-01 5.58769941e-01
-1.27599490e+00 -4.51113939e-01 9.25466478e-01 -4.53263551e-01
8.88971567e-01 1.51898101e-01 5.22970498e-01 1.20607507e+00
8.31632912e-02 7.96171010e-01 1.19426298e+00 -4.32344973e-01
2.80812513e-02 4.82624143e-01 -2.75030971e-01 7.47023106e-01
-2.33538851e-01 2.95248836e-01 -8.73098671e-01 8.71882513e-02
8.05759311e-01 2.25131065e-01 -4.43152726e-01 -3.98656458e-01
-1.35258090e+00 7.07445979e-01 7.59432971e-01 2.76547164e-01
-4.78675932e-01 1.82917133e-01 3.09828669e-01 1.52468801e-01
4.24277633e-01 3.55637252e-01 -7.08997488e-01 2.16996238e-01
-5.00083745e-01 -3.78517509e-02 1.13440417e-01 1.03170419e+00
9.43554401e-01 -9.50933173e-02 -1.14527857e+00 7.99000263e-01
4.67094123e-01 3.31579834e-01 5.60406864e-01 -6.36440933e-01
8.02262723e-01 6.73358917e-01 -9.02411342e-02 -8.03629935e-01
-9.06134248e-02 -4.04685199e-01 -5.32528579e-01 7.21850544e-02
6.67267889e-02 -8.49639699e-02 -1.07828522e+00 1.93677771e+00
5.18701911e-01 2.30950519e-01 -1.51669085e-01 1.01261842e+00
9.52773750e-01 1.01569557e+00 4.83810186e-01 -2.26673052e-01
1.37466097e+00 -1.30348241e+00 -6.22979283e-01 -4.65313941e-01
4.35389906e-01 -8.20235848e-01 1.48564601e+00 3.57433781e-02
-8.23902249e-01 -8.00341249e-01 -8.95011485e-01 -3.94982606e-01
-3.11064541e-01 2.73812354e-01 3.02852541e-01 -1.12767465e-01
-1.09926963e+00 1.49822071e-01 -4.09237266e-01 -2.47124508e-01
6.43383145e-01 2.74180949e-01 -1.74774125e-01 -2.55503088e-01
-1.04368091e+00 6.88671827e-01 4.98853534e-01 1.44184068e-01
-1.10188746e+00 -8.23646724e-01 -7.15424180e-01 3.07176203e-01
6.72617257e-01 -7.49203324e-01 1.09212947e+00 -1.27526224e+00
-1.65474975e+00 6.78469479e-01 -3.16249937e-01 2.47722670e-01
2.58886337e-01 -1.76164195e-01 -8.26200992e-02 2.35733807e-01
7.16198161e-02 9.73537445e-01 1.40242946e+00 -1.46669376e+00
-7.02732742e-01 -2.23943621e-01 2.40574151e-01 5.22295535e-01
-5.68410575e-01 -2.15796992e-01 -7.08035290e-01 -7.13038623e-01
-3.77989322e-01 -7.14155316e-01 -2.39012063e-01 3.22205752e-01
-2.54566193e-01 -4.20470089e-01 7.23854423e-01 -4.94751036e-01
1.13088620e+00 -2.41375518e+00 2.92557806e-01 -5.09246206e-03
1.43158421e-01 3.50449324e-01 -7.20419109e-01 5.25247514e-01
-2.24772811e-01 -1.85934920e-02 -6.90905750e-02 -2.52528399e-01
-2.31366381e-01 1.37209222e-01 -3.02810222e-01 1.98679730e-01
5.43193579e-01 1.15424359e+00 -9.45374250e-01 -5.87514937e-01
3.36609393e-01 5.89795828e-01 -7.60131240e-01 5.79991817e-01
-4.18234229e-01 7.29226351e-01 -7.49846101e-01 4.83619630e-01
4.32375848e-01 -3.83388311e-01 -2.01020073e-02 -4.16686535e-01
-2.21036404e-01 -1.40811373e-02 -7.28338361e-01 1.88016748e+00
-8.16064537e-01 4.18110102e-01 -5.07998988e-02 -1.13828599e+00
9.44668114e-01 2.20321342e-01 1.22763887e-01 -1.12595391e+00
1.87821060e-01 -2.66666561e-01 -2.35648930e-01 -8.96256626e-01
3.23435128e-01 2.11419817e-02 1.53511569e-01 1.90456614e-01
1.60484195e-01 3.00102443e-01 -1.83572054e-01 2.08946750e-01
6.47324204e-01 3.75269324e-01 3.56995046e-01 -1.79246813e-01
7.03614712e-01 -4.54473943e-02 4.51197058e-01 4.48097348e-01
-3.43018442e-01 3.33339483e-01 3.87546241e-01 -1.52616829e-01
-8.22071671e-01 -9.49812829e-01 6.25892729e-02 1.73052394e+00
3.17604631e-01 -3.73732537e-01 -5.07101953e-01 -6.94528162e-01
-1.82675514e-02 5.81307888e-01 -8.84529889e-01 -3.90903831e-01
-4.80460256e-01 -3.46187204e-01 -1.78463519e-01 6.13736033e-01
3.48112613e-01 -1.38384259e+00 -2.55679935e-01 -9.18720849e-04
-1.42264232e-01 -1.02658880e+00 -8.31028879e-01 1.45729154e-01
-7.79468238e-01 -7.20514238e-01 -6.17502868e-01 -1.00299585e+00
8.67614448e-01 5.15903294e-01 7.09969580e-01 2.01304063e-01
-1.89916953e-01 5.63143611e-01 -4.77403969e-01 -9.20483768e-02
-2.82544699e-02 -3.89759801e-02 -3.66765916e-01 4.45582956e-01
2.32669815e-01 -3.63669395e-01 -9.70495284e-01 3.08075789e-02
-8.48063648e-01 3.29563588e-01 8.64686608e-01 1.09113622e+00
8.97749662e-01 -6.60526335e-01 6.63334727e-01 -7.70458698e-01
5.57082951e-01 -6.72042310e-01 -4.19624954e-01 3.49868894e-01
-6.75083876e-01 2.53470421e-01 7.25402832e-01 -7.55827844e-01
-1.41636312e+00 9.78412777e-02 1.64728969e-01 -7.18856394e-01
-1.01908281e-01 4.75694150e-01 -3.72101188e-01 1.40777856e-01
4.87247735e-01 3.21156591e-01 -2.57383555e-01 -3.95790875e-01
8.73076439e-01 5.86199522e-01 3.68131042e-01 -7.59959400e-01
7.22465992e-01 3.47207516e-01 -2.39464298e-01 -5.55940151e-01
-9.86698091e-01 -4.99915004e-01 -4.45224822e-01 -2.40586728e-01
9.97191668e-01 -9.81433868e-01 -5.86292028e-01 4.70670201e-02
-1.11455524e+00 -5.36207616e-01 -1.79896235e-01 1.48004308e-01
-5.46560109e-01 5.43476343e-02 -2.01932698e-01 -4.63372558e-01
-3.03796619e-01 -1.21137738e+00 1.22570741e+00 4.48520064e-01
1.81705281e-01 -9.44548726e-01 -1.00404121e-01 3.39965820e-01
3.11004281e-01 -7.08026215e-02 1.04445302e+00 -4.36899543e-01
-9.40023720e-01 1.50314271e-01 -7.31649935e-01 3.30768108e-01
8.98545831e-02 -2.51011461e-01 -1.20802999e+00 -3.50277603e-01
-1.98985443e-01 -3.11426342e-01 9.67014849e-01 3.78171772e-01
1.65527952e+00 -4.63770717e-01 -3.93837094e-01 7.96103299e-01
1.67047358e+00 1.79697588e-01 4.05957520e-01 3.08412373e-01
1.02131391e+00 6.67929709e-01 9.13867712e-01 4.24283057e-01
4.65232641e-01 6.83505535e-01 5.25315881e-01 2.15985738e-02
-3.44453603e-01 -6.45138919e-01 3.43502462e-01 5.85292637e-01
2.07356922e-03 -2.94507474e-01 -5.98216772e-01 6.10270798e-01
-1.91194570e+00 -7.75159121e-01 1.99040323e-01 1.98773026e+00
9.50277865e-01 -2.38130122e-01 -3.48161608e-01 -4.52064633e-01
7.63258755e-01 2.78338611e-01 -7.30800688e-01 -3.10662210e-01
2.43099540e-01 1.23108000e-01 1.74806178e-01 4.31040555e-01
-8.27507019e-01 1.13365304e+00 4.61788225e+00 9.26522732e-01
-1.36595750e+00 4.26013231e-01 4.58295614e-01 -3.77447903e-01
-5.38749158e-01 1.40538275e-01 -7.61638641e-01 4.74544883e-01
5.05744934e-01 -2.38592550e-01 5.06440282e-01 6.55445993e-01
4.91652906e-01 1.27107203e-01 -1.19629920e+00 1.12335312e+00
7.44103640e-02 -1.28639281e+00 3.74792159e-01 -8.10778290e-02
7.23409951e-01 -8.95786062e-02 3.49501580e-01 5.95439136e-01
-2.68129617e-01 -7.84415543e-01 7.38369644e-01 6.01012290e-01
1.04574275e+00 -6.41249418e-01 5.45578077e-02 1.67362973e-01
-1.33792615e+00 -4.40419227e-01 -2.78255135e-01 1.45995602e-01
-9.68756229e-02 1.79552481e-01 -7.53348172e-01 2.42608607e-01
7.79980302e-01 8.31654429e-01 -7.37352490e-01 7.74507701e-01
-4.75351602e-01 4.61981475e-01 2.22622871e-01 -1.16239250e-01
2.95885056e-01 6.93958178e-02 4.98319834e-01 1.17197168e+00
1.63350448e-01 3.30712736e-01 2.26460978e-01 9.01943862e-01
-1.40178008e-02 3.31153840e-01 -3.48721564e-01 -1.14319429e-01
5.35359800e-01 1.46741009e+00 -3.76286983e-01 -3.10175300e-01
-5.00074744e-01 9.70660865e-01 7.52087474e-01 7.88896799e-01
-6.95139587e-01 -2.10091159e-01 6.44152761e-01 -3.55914645e-02
6.37029707e-01 4.02830765e-02 -2.10734084e-01 -1.16250134e+00
1.58250540e-01 -6.28788769e-01 3.32112998e-01 -8.67068529e-01
-1.34847200e+00 4.64412659e-01 8.52368865e-03 -1.26107872e+00
1.36839852e-01 -6.92063928e-01 -6.65507436e-01 9.21204925e-01
-1.85264373e+00 -1.40914059e+00 -4.71232355e-01 8.88996303e-01
9.17488515e-01 -3.91675495e-02 4.99836415e-01 1.58514053e-01
-5.92403591e-01 7.33856201e-01 -8.55047032e-02 -2.92874336e-01
9.97082472e-01 -1.02265453e+00 -1.26570523e-01 6.65573657e-01
-1.79216266e-01 4.88187671e-01 5.59856892e-01 -4.38385755e-01
-1.32872343e+00 -1.36284876e+00 7.11127341e-01 -3.41283858e-01
5.54358661e-01 -4.30699766e-01 -1.10417986e+00 7.28870630e-01
4.87100124e-01 1.83487460e-01 5.43764770e-01 -2.82526389e-03
-5.22620559e-01 -3.03746581e-01 -7.49336362e-01 7.04985440e-01
1.12133062e+00 -5.62339604e-01 -5.63895404e-01 2.85346985e-01
1.04593086e+00 -2.64451951e-01 -5.45495927e-01 9.06283408e-02
2.45425254e-01 -6.87920213e-01 9.81194019e-01 -5.95073104e-01
7.77686715e-01 -2.72184163e-01 -1.66730568e-01 -1.40458012e+00
-5.87470770e-01 -3.30228001e-01 -8.75690952e-02 1.56098580e+00
3.13714325e-01 -5.98296762e-01 1.97803929e-01 3.45312029e-01
-3.46632332e-01 -1.00469279e+00 -6.83077037e-01 -3.85440469e-01
6.72525167e-02 -2.47008458e-01 6.48244679e-01 9.04133499e-01
-1.13630258e-01 5.93834877e-01 -2.14791313e-01 1.63426027e-01
3.44437778e-01 2.76535243e-01 5.33627391e-01 -7.46197402e-01
-4.00855154e-01 -4.86182690e-01 -2.07704306e-02 -1.11536467e+00
1.73887059e-01 -1.17407870e+00 2.56517798e-01 -1.51812482e+00
4.53902155e-01 -4.45745081e-01 -6.98827028e-01 7.84135997e-01
-6.25545204e-01 -1.45289227e-01 4.40731227e-01 4.07518782e-02
-5.08605897e-01 7.98283696e-01 1.79978061e+00 -2.38650039e-01
-2.35949501e-01 -2.72863001e-01 -9.78593409e-01 3.69110167e-01
6.89801335e-01 -4.88767326e-01 -7.71382570e-01 -7.61522353e-01
2.02383325e-01 3.36079709e-02 5.50248742e-01 -2.98137277e-01
1.50395602e-01 -5.28160453e-01 4.33080554e-01 -3.13719213e-01
3.56877714e-01 -7.51971543e-01 -5.30244350e-01 1.28942668e-01
-6.84101641e-01 -5.29610291e-02 2.84099877e-01 8.34952235e-01
-1.22621924e-01 2.33129114e-02 8.10506523e-01 -1.07428744e-01
-1.08280921e+00 6.25025928e-01 -7.96732828e-02 1.55905103e-02
1.10103452e+00 4.39200550e-02 -5.62215149e-01 -9.79608819e-02
-5.71491361e-01 4.85996217e-01 2.28235587e-01 8.88570964e-01
8.06634426e-01 -1.51240516e+00 -6.19154155e-01 2.08619297e-01
6.72718287e-01 -9.85911191e-02 6.08736634e-01 7.27426946e-01
7.71684945e-02 4.54547763e-01 -1.76659271e-01 -6.97498441e-01
-1.00681067e+00 1.04242110e+00 1.31991446e-01 1.69945974e-02
-6.00118935e-01 1.04359901e+00 9.55926418e-01 -1.17427267e-01
1.30893871e-01 -2.51694381e-01 -4.26376224e-01 8.62210765e-02
6.95986092e-01 -1.26748949e-01 -2.97761470e-01 -5.03296852e-01
-2.35078529e-01 6.23924613e-01 -1.78644076e-01 -5.68415262e-02
1.29343832e+00 -5.47492206e-01 5.01756221e-02 3.70617449e-01
1.37316287e+00 -1.70731112e-01 -1.62919760e+00 -4.99846369e-01
-3.70569348e-01 -5.94211876e-01 2.16187447e-01 -8.81258428e-01
-1.28540134e+00 1.09802687e+00 7.77865469e-01 -2.27469191e-01
1.53614199e+00 3.00199896e-01 6.12355649e-01 1.74576595e-01
-1.67265888e-02 -9.07513618e-01 4.73143876e-01 2.00970873e-01
1.21965945e+00 -1.34401333e+00 -2.75682777e-01 -2.10162938e-01
-9.18634057e-01 8.72518003e-01 1.09795272e+00 -1.42048940e-01
4.64049250e-01 -1.41655222e-01 1.52812764e-01 -1.08435035e-01
-1.12881696e+00 -2.00776130e-01 4.80203688e-01 7.07473099e-01
2.70885646e-01 1.65482722e-02 -2.92139798e-01 7.79827654e-01
2.82784253e-01 -6.42978102e-02 1.33939356e-01 7.50001311e-01
-4.38886255e-01 -1.03098524e+00 -2.15336829e-01 3.30618382e-01
-1.98244769e-02 -2.66815305e-01 -1.75080776e-01 5.01745880e-01
4.14454460e-01 7.93630183e-01 7.05856015e-04 -2.72815317e-01
2.73773968e-01 -2.56031733e-02 5.33773005e-01 -8.73169005e-01
-4.78976101e-01 5.73700555e-02 -2.87667632e-01 -8.14914763e-01
-3.58026356e-01 -5.68583190e-01 -1.16977310e+00 2.50264883e-01
-1.45195186e-01 -1.44231364e-01 5.54335952e-01 1.05610764e+00
6.09414458e-01 7.39101708e-01 8.20439696e-01 -9.49422836e-01
-4.40935522e-01 -7.88361847e-01 -1.67409033e-01 3.38967055e-01
5.35917640e-01 -7.68737376e-01 -2.18686014e-01 2.68374860e-01] | [10.192989349365234, 1.8232231140136719] |
f896ed91-bc0e-4890-86f2-d4e9d8a72e00 | multi-level-and-multi-scale-feature | 1703.01793 | null | http://arxiv.org/abs/1703.01793v2 | http://arxiv.org/pdf/1703.01793v2.pdf | Multi-Level and Multi-Scale Feature Aggregation Using Pre-trained Convolutional Neural Networks for Music Auto-tagging | Music auto-tagging is often handled in a similar manner to image
classification by regarding the 2D audio spectrogram as image data. However,
music auto-tagging is distinguished from image classification in that the tags
are highly diverse and have different levels of abstractions. Considering this
issue, we propose a convolutional neural networks (CNN)-based architecture that
embraces multi-level and multi-scaled features. The architecture is trained in
three steps. First, we conduct supervised feature learning to capture local
audio features using a set of CNNs with different input sizes. Second, we
extract audio features from each layer of the pre-trained convolutional
networks separately and aggregate them altogether given a long audio clip.
Finally, we put them into fully-connected networks and make final predictions
of the tags. Our experiments show that using the combination of multi-level and
multi-scale features is highly effective in music auto-tagging and the proposed
method outperforms previous state-of-the-arts on the MagnaTagATune dataset and
the Million Song Dataset. We further show that the proposed architecture is
useful in transfer learning. | ['Juhan Nam', 'Jongpil Lee'] | 2017-03-06 | null | null | null | null | ['music-auto-tagging'] | ['music'] | [ 2.30117619e-01 -3.79888892e-01 7.94343278e-02 -2.96342790e-01
-8.43176126e-01 -6.76588655e-01 4.02998894e-01 -3.11324373e-03
-5.42034626e-01 3.31744105e-01 3.64960790e-01 3.00769240e-01
-8.09169337e-02 -7.71849573e-01 -8.17069232e-01 -4.40903604e-01
-2.37837538e-01 1.23288296e-01 1.85845390e-01 1.62498370e-01
8.21706578e-02 2.65462935e-01 -1.90606546e+00 6.26796007e-01
3.56464028e-01 1.61742198e+00 1.57636628e-01 6.77612782e-01
-1.77806437e-01 8.60723674e-01 -5.06542563e-01 -3.09948653e-01
2.52551317e-01 -3.46425414e-01 -9.74396586e-01 1.71808034e-01
7.57016063e-01 -1.45058244e-01 -3.49136651e-01 1.09750545e+00
6.41248405e-01 5.41161261e-02 5.19950569e-01 -1.06887031e+00
-6.50516570e-01 9.04462337e-01 -2.72268534e-01 1.23536892e-01
1.15035042e-01 -2.53730953e-01 1.29263830e+00 -9.30579364e-01
6.61425069e-02 1.04717660e+00 9.86102939e-01 1.93494186e-01
-1.04674983e+00 -9.70689237e-01 -4.98666056e-02 1.54629752e-01
-1.62509680e+00 -2.58889675e-01 8.53655457e-01 -7.06328392e-01
6.99942172e-01 2.45713051e-02 9.12838101e-01 8.72100115e-01
-9.33573917e-02 7.56594777e-01 1.04108238e+00 -4.43182290e-01
3.14395241e-02 -2.49214098e-01 -1.54175565e-01 6.16051137e-01
-3.65598649e-01 -1.67751402e-01 -8.77256095e-01 -1.11737579e-01
8.37134182e-01 1.37705415e-01 -2.74703335e-02 -2.11982980e-01
-1.61522877e+00 7.32044935e-01 6.39423549e-01 7.04230905e-01
-2.34177396e-01 5.84469318e-01 6.94121182e-01 2.38475233e-01
5.48047364e-01 5.78044772e-01 -4.09325689e-01 -2.80614376e-01
-1.16892803e+00 1.50765091e-01 4.39472973e-01 7.82649636e-01
7.40808368e-01 2.56937176e-01 -1.87626481e-01 1.13469362e+00
-1.74208418e-01 1.25855953e-01 7.24996388e-01 -8.60961199e-01
1.44694045e-01 3.05033863e-01 -2.22888604e-01 -1.00271440e+00
-4.72357213e-01 -9.24539745e-01 -1.06903255e+00 2.61812620e-02
1.75935507e-01 7.87156522e-02 -7.93731987e-01 1.72555876e+00
-6.11081049e-02 6.86270535e-01 -2.15872064e-01 9.36940432e-01
9.45629597e-01 4.54904526e-01 1.75602794e-01 2.47838125e-01
1.53499365e+00 -1.22155571e+00 -5.79132020e-01 1.21904045e-01
2.64014304e-01 -9.80259955e-01 1.22933257e+00 4.25776184e-01
-8.68767619e-01 -1.24239969e+00 -1.07077646e+00 -5.05623892e-02
-4.92378026e-01 6.07537270e-01 5.28730035e-01 2.04689577e-01
-1.10115767e+00 9.02525902e-01 -5.66702127e-01 -2.36951381e-01
6.11457407e-01 3.86859059e-01 -2.07157478e-01 4.46443290e-01
-1.16178548e+00 1.44961745e-01 6.96923018e-01 -1.43905997e-01
-8.89051557e-01 -7.25069761e-01 -6.86347425e-01 3.19212347e-01
1.08506404e-01 -6.60037398e-01 1.30503500e+00 -1.23594606e+00
-1.56285405e+00 8.30044568e-01 2.73126215e-01 -4.92125571e-01
1.72742844e-01 -3.56710851e-01 -4.93263513e-01 6.61311820e-02
1.69029906e-01 9.77077603e-01 1.01277232e+00 -9.72319603e-01
-9.44198668e-01 -1.91144735e-01 2.96740919e-01 2.00966880e-01
-8.44764531e-01 -3.03425398e-02 -4.56271738e-01 -1.10652137e+00
2.51457263e-02 -1.06872845e+00 6.44523352e-02 -1.29149586e-01
-4.33660269e-01 -5.77349998e-02 4.80480820e-01 -1.93503708e-01
1.19010866e+00 -2.55871797e+00 4.79609668e-02 -4.04127687e-02
2.31213480e-01 3.83264907e-02 -4.10261184e-01 2.21830085e-01
-1.15388855e-01 4.23930772e-03 7.98024759e-02 -3.63027066e-01
1.84069291e-01 -1.70429319e-01 -3.92470270e-01 -3.34873497e-02
8.15457627e-02 8.72554183e-01 -9.01153386e-01 -5.46057343e-01
6.11422434e-02 4.63464648e-01 -6.23326600e-01 1.79159969e-01
-5.96338809e-02 4.74927515e-01 -2.56988943e-01 6.62509978e-01
2.64290094e-01 -2.55898178e-01 6.72070216e-03 -5.87690055e-01
-6.57430142e-02 3.80165577e-01 -1.08364081e+00 2.25031853e+00
-6.24294817e-01 6.11055970e-01 -2.42571399e-01 -1.03161168e+00
8.91738474e-01 4.72264946e-01 8.66518736e-01 -4.67235506e-01
1.63723335e-01 2.48720124e-01 -1.20768011e-01 -2.79814273e-01
6.13471508e-01 -1.43071085e-01 -4.10396129e-01 5.05688608e-01
6.72163486e-01 1.00515850e-01 2.67023873e-02 -1.83976129e-01
9.49422538e-01 9.82030407e-02 1.06654286e-01 -1.07963927e-01
4.49174792e-01 -2.17798874e-01 5.81614256e-01 7.36293316e-01
-4.19558659e-02 6.46757901e-01 7.77543858e-02 -7.29711235e-01
-1.12786508e+00 -8.69189024e-01 -1.68071613e-01 1.44127679e+00
-1.58545077e-01 -7.90632963e-01 -6.84099495e-01 -6.56276166e-01
1.40915647e-01 -1.38874084e-01 -6.91894650e-01 -2.05721874e-02
-1.66679502e-01 -2.88897246e-01 8.80994499e-01 8.10934186e-01
6.70554817e-01 -1.33305109e+00 -4.23926800e-01 2.72108883e-01
-1.74587533e-01 -1.17601967e+00 -5.54438531e-01 3.44931483e-01
-7.04601645e-01 -8.66168141e-01 -6.99695110e-01 -1.03326678e+00
8.57378319e-02 2.17450440e-01 1.32298803e+00 -1.33994818e-01
-1.62898093e-01 3.78546685e-01 -5.75919390e-01 -5.45391619e-01
-3.08546927e-02 5.08196473e-01 2.36153036e-01 3.92567366e-01
4.30664092e-01 -1.07691991e+00 -5.69564998e-01 1.68581858e-01
-8.97384644e-01 3.22765969e-02 6.10556841e-01 6.87798917e-01
9.50710416e-01 1.72312990e-01 6.37322843e-01 -6.45870030e-01
6.00742400e-01 -6.51294366e-02 -2.55837291e-01 -4.47731800e-02
-9.92264301e-02 -1.27801239e-01 7.12117970e-01 -8.49417388e-01
-4.00282919e-01 5.84620178e-01 -1.66538268e-01 -7.87013710e-01
-4.76137102e-01 4.96925294e-01 3.35506955e-03 -6.69098645e-02
4.99310762e-01 1.21632047e-01 -3.40032667e-01 -8.21214437e-01
4.01926309e-01 9.06818867e-01 7.89792120e-01 -4.78626907e-01
8.15342128e-01 4.24367696e-01 -6.53762892e-02 -5.10026336e-01
-1.41135263e+00 -5.39539158e-01 -9.17454362e-01 -4.26633179e-01
1.12868023e+00 -1.29141140e+00 -8.18444192e-01 5.48356473e-01
-8.06726217e-01 -1.45150512e-01 -5.65429568e-01 6.98967874e-01
-8.00369978e-01 2.18664035e-02 -6.35234296e-01 -5.10345161e-01
-3.56199354e-01 -9.59040821e-01 1.37331152e+00 9.71202180e-02
-2.61056662e-01 -6.94090486e-01 3.34835887e-01 1.64513543e-01
3.67952645e-01 7.39059821e-02 8.73015821e-01 -8.50945830e-01
-3.50123674e-01 -8.24723318e-02 -1.78660139e-01 4.22495395e-01
1.75426841e-01 -3.95508587e-01 -1.50391448e+00 -2.58445263e-01
-4.48816359e-01 -7.65669405e-01 1.13511324e+00 3.64604026e-01
1.65984523e+00 -9.35599506e-02 3.08475532e-02 7.74426103e-01
1.31307757e+00 -1.38484314e-02 2.50644952e-01 5.20273268e-01
8.12540770e-01 1.64561167e-01 4.91272658e-01 4.80180234e-01
2.34191388e-01 8.56648326e-01 5.18611372e-01 -2.62121111e-01
-2.77032048e-01 -3.52118105e-01 2.71063983e-01 1.17673504e+00
-1.36385009e-01 3.59442741e-01 -5.63272178e-01 6.27016544e-01
-1.89119482e+00 -9.17859018e-01 2.90100068e-01 1.93052816e+00
8.57982159e-01 -2.80984268e-02 4.81072277e-01 4.72560853e-01
7.01874316e-01 2.90086746e-01 -2.46943384e-01 -8.07571188e-02
-3.00613251e-02 5.29621065e-01 8.97195488e-02 -3.12417420e-03
-1.70196974e+00 1.07272446e+00 6.33626366e+00 1.01910079e+00
-1.25446975e+00 2.69509345e-01 2.57115811e-01 -2.00469479e-01
1.32619172e-01 -2.10019782e-01 -3.71396571e-01 3.23800057e-01
8.64750504e-01 2.65906185e-01 5.64947367e-01 8.03857684e-01
-2.65639395e-01 4.54903692e-01 -1.06679952e+00 1.30923975e+00
1.26299769e-01 -1.34278929e+00 2.10668162e-01 4.98065166e-02
6.90992296e-01 1.55020347e-02 1.71593964e-01 3.20275694e-01
1.75975934e-02 -8.88903558e-01 1.16455245e+00 6.92040741e-01
8.57266009e-01 -9.83488262e-01 6.89190090e-01 -3.36734168e-02
-1.75872242e+00 -2.80809015e-01 -5.23856759e-01 -1.44411981e-01
-1.63160190e-01 6.09827220e-01 -5.97126186e-01 6.00357771e-01
1.07684994e+00 1.02786410e+00 -5.42268455e-01 1.10885608e+00
1.73911497e-01 6.05660200e-01 1.28177297e-03 2.79096454e-01
2.15656847e-01 1.09415263e-01 2.14683086e-01 1.33759534e+00
6.18039489e-01 -3.11476469e-01 4.25733149e-01 7.78037429e-01
-3.41454864e-01 3.00294250e-01 -3.20942551e-01 -3.14823717e-01
3.82337064e-01 1.57124674e+00 -6.45233691e-01 -4.48914677e-01
-3.99893731e-01 9.49570954e-01 3.19653958e-01 9.45941061e-02
-7.13123381e-01 -4.78298157e-01 7.49135792e-01 -2.35450249e-02
6.19184434e-01 -1.28676757e-01 2.91221291e-02 -1.15155184e+00
-1.96510777e-01 -7.56888151e-01 4.07171428e-01 -1.01413321e+00
-1.27952373e+00 1.03065312e+00 -5.33530354e-01 -1.81236923e+00
-1.04406305e-01 -5.04712462e-01 -3.59474868e-01 4.45713311e-01
-1.56845927e+00 -1.52158856e+00 -3.75854015e-01 8.52231860e-01
5.55284202e-01 -5.75875223e-01 1.21277595e+00 7.29137242e-01
-1.25436991e-01 6.82524800e-01 -2.16012634e-02 6.74077213e-01
8.56808901e-01 -1.28810954e+00 1.52101591e-01 2.61419594e-01
1.03358078e+00 2.16906473e-01 5.95819913e-02 -2.00937048e-01
-8.00703824e-01 -1.44425261e+00 8.27258170e-01 -1.17351644e-01
8.18732083e-01 -4.47188228e-01 -6.68562889e-01 4.97084349e-01
1.84743121e-01 1.83129385e-01 1.07245350e+00 3.15682471e-01
-6.05067909e-01 -3.86177689e-01 -3.99479806e-01 4.56787311e-02
1.12499976e+00 -1.02799010e+00 -4.48005080e-01 1.65477648e-01
6.56201601e-01 -2.51065731e-01 -1.22334456e+00 4.01075184e-01
8.11741412e-01 -7.35363781e-01 1.03298509e+00 -6.92938924e-01
5.04313827e-01 -4.07134324e-01 -3.58967692e-01 -1.28167415e+00
-7.63437986e-01 -3.56383920e-01 1.76283538e-01 1.31997335e+00
2.55372494e-01 1.16000324e-01 5.74590445e-01 -5.34730732e-01
-3.08195978e-01 -6.02828443e-01 -8.03929806e-01 -7.97931612e-01
-1.54639885e-01 -6.14460111e-01 7.65782177e-01 1.12501287e+00
-3.60866450e-03 4.52352077e-01 -6.43199623e-01 2.97775008e-02
3.49723339e-01 5.27424574e-01 8.83739769e-01 -1.72609866e+00
-3.27861726e-01 -3.35296899e-01 -8.05351794e-01 -8.53965878e-01
3.57775956e-01 -1.11972702e+00 4.35671881e-02 -1.12732875e+00
3.52523595e-01 -3.89457226e-01 -8.90405476e-01 7.27569580e-01
1.02076113e-01 1.02175152e+00 3.64266664e-01 3.92243564e-01
-9.11479056e-01 5.32525659e-01 1.13604963e+00 -4.11295027e-01
-1.04890145e-01 -1.65124498e-02 -4.29245263e-01 8.22326660e-01
7.37955987e-01 -4.63723272e-01 -1.64709851e-01 -4.01475459e-01
2.62281805e-01 -2.88684100e-01 5.08309364e-01 -1.56924701e+00
1.21550895e-01 3.13256443e-01 5.41120350e-01 -6.34180963e-01
6.00050628e-01 -8.56522739e-01 1.55073270e-01 9.41667706e-02
-7.27967083e-01 -4.47612144e-02 2.31215715e-01 5.21937311e-01
-7.38550246e-01 5.28044254e-03 6.00093067e-01 -1.48053944e-01
-5.65139472e-01 4.49701339e-01 -2.68733978e-01 -8.96767080e-02
4.46174413e-01 -1.43375725e-01 2.36561373e-01 -4.74782020e-01
-1.08919621e+00 -3.80055875e-01 1.99918419e-01 6.37325466e-01
2.82970995e-01 -1.98837960e+00 -4.89613086e-01 6.81612641e-02
3.50914091e-01 -3.69639993e-01 2.82470465e-01 7.19579577e-01
-8.68589729e-02 4.23617333e-01 -6.16108775e-01 -8.78348470e-01
-1.21944857e+00 5.12664855e-01 2.56020814e-01 -1.45263225e-01
-5.17535567e-01 7.62460828e-01 1.84188738e-01 -3.65511209e-01
5.57523727e-01 -4.55024630e-01 -4.17302638e-01 3.53612930e-01
4.08165962e-01 -1.43005416e-01 5.86272813e-02 -8.89299095e-01
-1.10625282e-01 9.65353251e-01 1.88494280e-01 -9.14355814e-02
1.64670825e+00 7.60285109e-02 -4.64139096e-02 7.73785412e-01
1.17630696e+00 4.79282025e-04 -1.16910744e+00 -5.59423566e-01
-2.00061232e-01 -3.78804147e-01 1.18814878e-01 -4.79905576e-01
-1.47910881e+00 1.10741925e+00 8.49928558e-01 4.40532923e-01
1.32230663e+00 9.41556096e-02 7.74197757e-01 3.19824040e-01
1.15562990e-01 -1.23495042e+00 4.62527871e-01 5.38597286e-01
7.14427531e-01 -8.06942880e-01 -2.72849560e-01 -3.39623317e-02
-4.94368225e-01 1.16415155e+00 4.89880413e-01 -3.46186697e-01
8.33182752e-01 3.07956845e-01 2.54488885e-01 -2.11724579e-01
-3.28225225e-01 -5.91961324e-01 7.49553561e-01 4.41944808e-01
5.99892497e-01 2.74034124e-02 1.09623663e-01 9.33852971e-01
-6.71413481e-01 1.83946639e-01 1.20819353e-01 6.93558872e-01
-4.29446548e-01 -1.11779916e+00 -2.75859565e-01 2.78420925e-01
-6.65171206e-01 -1.64172918e-01 -5.86421132e-01 5.77684760e-01
6.00014925e-01 6.83285654e-01 4.04422253e-01 -1.02710044e+00
1.86703891e-01 3.08051497e-01 4.06288922e-01 -6.16961837e-01
-8.28044116e-01 4.93110925e-01 -1.51635394e-01 -5.16032994e-01
-9.53199685e-01 -1.76981911e-01 -9.74999607e-01 9.32244360e-02
-3.08690012e-01 2.70193189e-01 6.10270441e-01 8.83460879e-01
4.99647409e-01 9.48953927e-01 6.48638248e-01 -1.18457901e+00
-3.33025038e-01 -1.16038907e+00 -9.68310416e-01 5.20750999e-01
3.26695591e-01 -7.48196840e-01 6.67401217e-03 3.46163660e-01] | [15.722698211669922, 5.206240177154541] |
4a343228-9587-4284-858a-0da1390f9d69 | random-access-neural-compression-of-material | 2305.17105 | null | https://arxiv.org/abs/2305.17105v1 | https://arxiv.org/pdf/2305.17105v1.pdf | Random-Access Neural Compression of Material Textures | The continuous advancement of photorealism in rendering is accompanied by a growth in texture data and, consequently, increasing storage and memory demands. To address this issue, we propose a novel neural compression technique specifically designed for material textures. We unlock two more levels of detail, i.e., 16x more texels, using low bitrate compression, with image quality that is better than advanced image compression techniques, such as AVIF and JPEG XL. At the same time, our method allows on-demand, real-time decompression with random access similar to block texture compression on GPUs, enabling compression on disk and memory. The key idea behind our approach is compressing multiple material textures and their mipmap chains together, and using a small neural network, that is optimized for each material, to decompress them. Finally, we use a custom training implementation to achieve practical compression speeds, whose performance surpasses that of general frameworks, like PyTorch, by an order of magnitude. | ['Aaron Lefohn', 'Pontus Ebelin', 'Tomas Akenine-Möller', 'Bartlomiej Wronski', 'Marco Salvi', 'Karthik Vaidyanathan'] | 2023-05-26 | null | null | null | null | ['image-compression'] | ['computer-vision'] | [ 5.16497850e-01 -6.75910041e-02 1.08459052e-02 1.41790286e-01
-3.68743747e-01 -2.59938955e-01 4.77802813e-01 2.42644578e-01
-2.78017521e-01 4.16096509e-01 7.63662830e-02 -5.34428477e-01
2.15588942e-01 -1.25651848e+00 -9.19688344e-01 -6.77317202e-01
-4.21335287e-02 4.18340951e-01 5.26316822e-01 -2.08306387e-01
4.36064303e-01 6.09397769e-01 -2.07117701e+00 5.78194141e-01
6.09033704e-01 1.30341303e+00 2.73463339e-01 9.59875941e-01
-1.51392460e-01 8.27208459e-01 -2.64925689e-01 -4.61888582e-01
4.04791296e-01 3.29821631e-02 -6.49780631e-01 -5.94181456e-02
7.96479464e-01 -1.01705110e+00 -6.26923382e-01 6.23367906e-01
2.43332401e-01 3.74758691e-02 3.73940319e-01 -5.60384274e-01
-3.07074338e-01 4.90257144e-01 -9.77932632e-01 -8.62267911e-02
6.41390011e-02 1.58869103e-01 7.36396432e-01 -4.24002886e-01
6.56420648e-01 1.12236679e+00 6.58474267e-01 2.66016603e-01
-1.27140558e+00 -4.15817380e-01 -4.39816535e-01 1.20988630e-01
-1.17649817e+00 -3.95146906e-01 4.06302154e-01 -1.24587722e-01
1.16405606e+00 8.11680257e-01 1.00964224e+00 7.42644489e-01
3.47718716e-01 4.76550519e-01 1.02811646e+00 -3.59147400e-01
2.37106130e-01 -3.28337073e-01 -4.49563473e-01 4.30652231e-01
2.05076814e-01 6.71110377e-02 -5.63679457e-01 -6.25740290e-02
1.31585157e+00 3.85860093e-02 -2.61311442e-01 -1.75282270e-01
-1.12496793e+00 5.06612420e-01 2.14043364e-01 5.52774407e-02
-4.94078636e-01 6.25816286e-01 7.57713497e-01 3.81719261e-01
5.77481270e-01 1.93589434e-01 -1.78020880e-01 -4.88131464e-01
-1.42509270e+00 3.63797247e-01 9.08276320e-01 6.99191570e-01
7.17980504e-01 3.29798125e-02 -1.55719398e-02 8.41691077e-01
-1.33435041e-01 5.14560401e-01 3.16617817e-01 -1.21087682e+00
6.07517898e-01 3.16509277e-01 -2.74877161e-01 -1.31121552e+00
-7.66678303e-02 -8.44754651e-02 -1.43766963e+00 5.61381102e-01
8.62967297e-02 5.27690053e-01 -7.32963622e-01 1.21678567e+00
4.22700644e-01 1.06801532e-01 -2.31059611e-01 6.32245958e-01
2.49846578e-01 1.13020205e+00 -1.72135994e-01 1.09465368e-01
1.63585353e+00 -1.00422919e+00 -3.11417341e-01 1.37260213e-01
4.03096288e-01 -9.92267191e-01 1.37794006e+00 8.19733799e-01
-1.42132640e+00 -4.24231887e-01 -1.24595356e+00 -6.67972088e-01
-1.52265936e-01 -1.31355926e-01 8.51135552e-01 4.61838275e-01
-1.14649975e+00 1.07234454e+00 -1.18532002e+00 1.47397676e-02
3.77409220e-01 3.74540210e-01 -3.05944473e-01 -1.78651839e-01
-5.28924763e-01 3.59179825e-01 5.27366638e-01 -2.29200512e-01
-6.20197058e-01 -9.75565791e-01 -5.23487389e-01 5.85783958e-01
1.22202985e-01 -5.50993443e-01 1.05275118e+00 -4.69437331e-01
-1.76846719e+00 7.90589511e-01 1.56952545e-01 -7.27798760e-01
6.11039877e-01 -3.71110380e-01 -5.14567504e-03 4.69180286e-01
-5.26234388e-01 9.56244946e-01 1.07660258e+00 -1.12657320e+00
-5.27502656e-01 -1.05663888e-01 4.73442189e-02 2.44977161e-01
-7.14684784e-01 -2.31776401e-01 -7.88352728e-01 -7.56944358e-01
-4.22668830e-02 -6.43067598e-01 -2.20963821e-01 3.25727999e-01
-1.92651287e-01 1.53172389e-01 1.05766320e+00 -1.01612806e+00
1.32993317e+00 -2.26589036e+00 2.96295941e-01 2.05193728e-01
6.11799121e-01 3.69318187e-01 -6.28088564e-02 4.40549552e-01
1.53880805e-01 -6.24117218e-02 -3.46967429e-01 -5.46160817e-01
-1.14705339e-01 3.97051603e-01 -5.99708021e-01 4.56333570e-02
-4.81154695e-02 6.40386462e-01 -4.27162915e-01 -3.02452892e-01
3.38973463e-01 8.40619922e-01 -1.03300238e+00 1.78270251e-01
-4.46461231e-01 2.31299266e-01 -6.70892075e-02 3.99449944e-01
1.07221484e+00 -3.81486475e-01 2.90038377e-01 -3.88410419e-01
-3.95786881e-01 3.88072431e-01 -8.90505493e-01 1.63611186e+00
-7.67062366e-01 7.85941184e-01 9.68673080e-02 -7.69279897e-01
7.63023496e-01 1.09704658e-01 3.43454927e-01 -9.94918466e-01
-1.38061652e-02 3.61971945e-01 -3.32384944e-01 -2.33421490e-01
1.15526605e+00 2.66532809e-01 2.69887686e-01 6.25074387e-01
-5.01852751e-01 -4.90837425e-01 2.85655797e-01 1.21647649e-01
1.13116384e+00 2.07836360e-01 -8.42081234e-02 -1.00135975e-01
1.99184299e-01 -1.76278546e-01 -1.69870406e-01 5.91541648e-01
6.23178005e-01 6.19728327e-01 6.47974849e-01 -7.24164963e-01
-1.90886116e+00 -6.85544729e-01 -1.67646006e-01 9.97106910e-01
-5.01318462e-02 -8.51674318e-01 -9.71695304e-01 2.49762639e-01
-4.71319221e-02 3.87398094e-01 -4.84643966e-01 1.84113324e-01
-1.04632199e+00 -5.24517298e-01 4.70468849e-01 3.92297894e-01
7.69553244e-01 -1.02036250e+00 -1.25285852e+00 2.64200777e-01
8.16398561e-02 -9.89978611e-01 -8.49819258e-02 2.04515100e-01
-1.31558192e+00 -6.71863914e-01 -7.69723594e-01 -5.16436577e-01
3.15523177e-01 3.50856394e-01 1.28508484e+00 4.25725222e-01
-3.68266851e-01 -1.61467716e-01 -1.45451546e-01 2.14176059e-01
-5.27018487e-01 3.28891695e-01 -4.70017999e-01 -2.70201296e-01
-2.03146860e-01 -1.06366801e+00 -8.98394287e-01 1.12202428e-01
-1.45563030e+00 8.51527750e-01 6.15017891e-01 6.79964900e-01
9.01315868e-01 3.18637520e-01 -4.91062105e-01 -6.94937944e-01
5.19562185e-01 -1.18990988e-01 -7.92067468e-01 -5.34055643e-02
-4.40219700e-01 -7.75873736e-02 8.97904336e-01 -4.90427464e-01
-7.92219162e-01 -3.39959562e-01 -3.31695527e-01 -4.57221568e-01
9.01086703e-02 4.28948432e-01 3.23378921e-01 -3.05391222e-01
5.27725875e-01 3.08936268e-01 2.21076250e-01 -7.20012307e-01
3.02870482e-01 5.01429558e-01 7.87548900e-01 -6.42443359e-01
6.00656033e-01 5.95001101e-01 2.56297946e-01 -9.14306998e-01
-2.47177660e-01 3.10733244e-02 -2.87331581e-01 5.64212613e-02
5.96221149e-01 -7.13599741e-01 -1.00951362e+00 6.92360878e-01
-1.05967033e+00 -8.16117525e-01 -4.86627787e-01 8.68932456e-02
-7.80776620e-01 4.76020843e-01 -1.19580281e+00 -2.89421678e-01
-5.04657924e-01 -1.21016300e+00 1.28846788e+00 6.84925318e-02
-2.90555861e-02 -5.95523059e-01 -8.80565569e-02 1.97796136e-01
8.61401320e-01 2.61373460e-01 1.00029290e+00 3.23659271e-01
-1.11942959e+00 -1.94211558e-01 -7.40576744e-01 2.95857906e-01
-1.80721790e-01 5.36525026e-02 -5.52634776e-01 -3.55012745e-01
1.35057956e-01 -4.36478823e-01 9.04747128e-01 1.46338791e-01
1.83023334e+00 -5.01321673e-01 -1.16020344e-01 1.08894885e+00
1.63687980e+00 5.33439815e-02 1.28027678e+00 5.50546646e-01
8.07081938e-01 2.73769468e-01 2.27161035e-01 6.81143701e-01
2.82340407e-01 8.73578429e-01 4.50637072e-01 -2.82702327e-01
-4.85530972e-01 -1.90185472e-01 -1.54522628e-01 1.28703189e+00
-4.94233698e-01 -2.57248491e-01 -8.14985037e-01 -2.22173426e-02
-1.44016111e+00 -7.60437787e-01 -6.22290671e-02 2.37001300e+00
1.01855707e+00 6.20907806e-02 -1.51355073e-01 3.03991824e-01
2.89367288e-01 3.20274174e-01 -3.57389003e-01 -7.30669618e-01
-1.03194810e-01 6.84057295e-01 7.72940814e-01 3.92247736e-01
-8.92960727e-01 9.66647148e-01 6.61001635e+00 1.37063265e+00
-1.44573152e+00 -9.48371217e-02 1.01616085e+00 -1.68009818e-01
-3.35058302e-01 -9.75808129e-02 -4.97773349e-01 4.78408962e-01
1.10247564e+00 1.06515378e-01 1.09786284e+00 7.02571273e-01
9.80818793e-02 -2.65995890e-01 -7.09356964e-01 1.02832329e+00
5.33919074e-02 -1.70643163e+00 3.04719985e-01 4.46188599e-01
5.25779486e-01 -7.96311498e-02 2.38884643e-01 -1.47606006e-05
3.99129320e-04 -9.86290932e-01 6.36089087e-01 1.61953732e-01
1.31828761e+00 -7.02009499e-01 3.55108321e-01 8.55473354e-02
-1.05041492e+00 1.08536713e-01 -7.20883369e-01 -2.71141250e-02
2.43466750e-01 5.87455571e-01 -2.08594859e-01 2.96616346e-01
9.39422429e-01 5.18586755e-01 -4.42068309e-01 9.52208042e-01
9.15343836e-02 5.56942284e-01 -6.62502825e-01 2.95327157e-01
2.15220019e-01 -2.97724664e-01 1.39810994e-01 1.04888785e+00
5.95971227e-01 1.56467497e-01 -2.64266759e-01 3.96748096e-01
-1.83656052e-01 1.75340459e-01 -4.48897153e-01 -3.26801166e-02
2.60437787e-01 1.14290702e+00 -9.14898336e-01 -7.42854118e-01
-2.74199843e-01 1.27667797e+00 4.02348608e-01 3.89842596e-03
-8.54501963e-01 -3.99095535e-01 3.16340864e-01 4.12495315e-01
5.97418964e-01 -5.59682429e-01 -4.07557845e-01 -1.01032817e+00
1.73196107e-01 -1.16808629e+00 -1.90251783e-01 -6.59806490e-01
-6.87762082e-01 7.30997324e-01 -8.09467435e-02 -1.17988789e+00
-6.33721724e-02 -6.17999315e-01 -3.22111279e-01 6.60407126e-01
-1.43376982e+00 -1.02475286e+00 -5.26255488e-01 3.16205502e-01
3.30900282e-01 8.63695592e-02 8.28589082e-01 6.24424279e-01
-1.79614127e-01 4.63828743e-01 3.86373729e-01 -3.88404727e-01
3.37269336e-01 -8.41006517e-01 9.47757304e-01 3.97158027e-01
-1.52757972e-01 4.04315859e-01 5.88765979e-01 -7.03834474e-01
-1.72466040e+00 -6.90213919e-01 4.63334113e-01 2.95923680e-01
5.95956802e-01 -4.60987598e-01 -1.16879559e+00 3.92708272e-01
3.41267675e-01 -3.23224723e-01 4.12160397e-01 -1.27887249e-01
-5.27599096e-01 -1.29909322e-01 -1.02160156e+00 9.08070147e-01
9.11163986e-01 -3.24590355e-01 5.67624308e-02 4.10401076e-01
8.11861396e-01 -9.37699199e-01 -9.32201385e-01 2.97153592e-01
8.11615467e-01 -1.34372449e+00 1.23380923e+00 -7.71761164e-02
1.29941201e+00 -5.63595518e-02 -4.33835685e-01 -8.28148365e-01
-3.95051211e-01 -5.94991326e-01 -5.59705913e-01 6.98085785e-01
-1.01993188e-01 -4.56056446e-01 1.06987214e+00 3.10848415e-01
-1.68699510e-02 -9.84089553e-01 -9.91001785e-01 -5.39689064e-01
-1.01306088e-01 -2.59931892e-01 8.49419892e-01 6.39857113e-01
-3.98555428e-01 -1.42686591e-01 -7.94121563e-01 -1.88738316e-01
5.33006012e-01 1.47759810e-01 9.30248916e-01 -8.68718624e-01
-7.93605685e-01 -5.98714471e-01 -4.44967806e-01 -1.45279062e+00
-5.28231621e-01 -6.47716641e-01 -2.23964438e-01 -1.21423161e+00
4.69991118e-02 -6.96515441e-01 2.03063920e-01 4.37150121e-01
2.17071876e-01 9.45113659e-01 2.72836149e-01 3.94157708e-01
-1.33642673e-01 5.14551103e-01 1.44963622e+00 -1.40482277e-01
-1.00028656e-01 -6.11315966e-01 -3.68158191e-01 5.41571319e-01
7.12310791e-01 -1.13164589e-01 -2.29900494e-01 -1.00906169e+00
3.47331643e-01 2.15658963e-01 3.34397763e-01 -1.43223917e+00
1.32431701e-01 9.42600295e-02 3.50581735e-01 -7.68096566e-01
6.78142488e-01 -8.18618417e-01 5.00997305e-01 4.55758780e-01
-2.16816306e-01 2.44333431e-01 4.37354922e-01 1.90300763e-01
-2.37594277e-01 -1.54187098e-01 7.56553471e-01 -2.05243602e-02
-2.18441576e-01 4.10290956e-01 -2.56239086e-01 -5.21899641e-01
6.54725850e-01 -3.52437913e-01 -3.34081113e-01 -1.90339431e-01
-3.31842184e-01 -2.86070973e-01 8.35889995e-01 9.44001526e-02
5.38093507e-01 -1.29027140e+00 -6.57895982e-01 4.39692825e-01
-4.48674738e-01 2.01713845e-01 6.48802936e-01 6.70074940e-01
-1.58491659e+00 2.29575053e-01 -4.23792660e-01 -6.68185890e-01
-1.24062490e+00 5.52117884e-01 -1.24151997e-01 -5.47150552e-01
-1.21078777e+00 5.92589736e-01 1.48473576e-01 1.77370936e-01
7.58306459e-02 -3.32602620e-01 1.86246604e-01 -2.55579114e-01
9.37608719e-01 4.47283775e-01 2.15408131e-01 -1.02357112e-01
3.41488838e-01 8.31161261e-01 -2.61991233e-01 -8.05742368e-02
1.49280000e+00 6.33351579e-02 -5.73881686e-01 4.17024419e-02
1.18491256e+00 2.52670974e-01 -1.44506240e+00 1.05451614e-01
-5.71770012e-01 -8.36104035e-01 1.84004903e-01 -4.14628297e-01
-1.24155450e+00 9.38850224e-01 4.53466237e-01 5.35718799e-01
1.37856889e+00 -3.45563352e-01 1.43765664e+00 2.09447756e-01
4.04082149e-01 -9.52407300e-01 -1.28760889e-01 3.33199114e-01
9.17359054e-01 -6.47900462e-01 2.49111265e-01 -7.01356649e-01
-1.13310173e-01 1.35016215e+00 2.74356365e-01 -3.62823248e-01
3.45358551e-01 7.65628159e-01 -3.71484101e-01 -1.11031167e-01
-8.59695613e-01 5.23726106e-01 -6.03693724e-02 2.85055697e-01
3.41029584e-01 2.47621596e-01 -2.55193859e-01 -3.79442811e-01
-6.02005243e-01 1.27454326e-01 3.94565105e-01 8.94116282e-01
-3.83841395e-01 -1.33168876e+00 -4.05720621e-01 6.83814049e-01
-3.25294226e-01 -4.02066201e-01 4.24374133e-01 5.16401947e-01
-1.33397475e-01 2.69969285e-01 6.17800176e-01 -3.58337939e-01
1.88666140e-03 -4.51400161e-01 6.31769359e-01 -4.76097129e-02
-5.21774173e-01 1.65753782e-01 -1.68329000e-01 -7.25115597e-01
4.91154641e-02 -3.56641412e-01 -6.35269463e-01 -1.08017755e+00
2.14938536e-01 -2.35181212e-01 8.81869793e-01 2.95780987e-01
4.67411757e-01 5.67676425e-01 4.43320513e-01 -1.41898394e+00
-2.26793244e-01 -5.94504774e-01 -4.15555537e-01 3.25049192e-01
-1.11099714e-02 -2.52900571e-01 -1.25288561e-01 6.93940595e-02] | [11.259970664978027, -1.3943321704864502] |
c4a7aeae-511b-4c9d-8ac5-b956792ba283 | towards-reducing-aleatoric-uncertainty-for | 2110.11012 | null | https://arxiv.org/abs/2110.11012v2 | https://arxiv.org/pdf/2110.11012v2.pdf | Towards Reducing Aleatoric Uncertainty for Medical Imaging Tasks | In safety-critical applications like medical diagnosis, certainty associated with a model's prediction is just as important as its accuracy. Consequently, uncertainty estimation and reduction play a crucial role. Uncertainty in predictions can be attributed to noise or randomness in data (aleatoric) and incorrect model inferences (epistemic). While model uncertainty can be reduced with more data or bigger models, aleatoric uncertainty is more intricate. This work proposes a novel approach that interprets data uncertainty estimated from a self-supervised task as noise inherent to the data and utilizes it to reduce aleatoric uncertainty in another task related to the same dataset via data augmentation. The proposed method was evaluated on a benchmark medical imaging dataset with image reconstruction as the self-supervised task and segmentation as the image analysis task. Our findings demonstrate the effectiveness of the proposed approach in significantly reducing the aleatoric uncertainty in the image segmentation task while achieving better or on-par performance compared to the standard augmentation techniques. | ['Deepti R. Bathula', 'Narayanan C. Krishnan', 'Abhishek Singh Sambyal'] | 2021-10-21 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 6.30678356e-01 7.83877671e-01 1.63297839e-02 -8.30921948e-01
-1.10126901e+00 -2.35155508e-01 5.36499977e-01 6.55809045e-01
-6.11428440e-01 9.91097331e-01 1.18493229e-01 -3.46953005e-01
-2.70465255e-01 -5.64602196e-01 -8.49239588e-01 -7.38806963e-01
4.12357807e-01 7.39308119e-01 2.02866495e-01 5.07724404e-01
2.00369582e-01 3.30691226e-02 -1.24179661e+00 1.88798979e-01
1.09466934e+00 1.40625191e+00 -1.00296125e-01 5.21319449e-01
-1.34828046e-01 9.88085806e-01 -5.90220511e-01 -2.96394318e-01
2.00839743e-01 -1.24121346e-01 -7.65818119e-01 4.80373986e-02
1.20732076e-01 -5.72663583e-02 2.30197102e-01 1.26010418e+00
2.21379682e-01 -6.02421686e-02 9.26271200e-01 -1.12681496e+00
-3.11421622e-02 9.32832241e-01 -4.79680389e-01 1.53475001e-01
-1.71023324e-01 3.63024208e-03 5.02814531e-01 -5.39148986e-01
1.94649130e-01 1.11356437e+00 7.12122977e-01 3.86861324e-01
-1.43277359e+00 -5.66126347e-01 -1.66401770e-02 -8.39520469e-02
-1.38205767e+00 -2.91387081e-01 6.01036906e-01 -5.78912079e-01
3.96822810e-01 2.86555707e-01 2.48413205e-01 7.77178705e-01
6.58226967e-01 5.03665030e-01 1.18519282e+00 -2.27229118e-01
7.61890590e-01 6.16350770e-01 3.62232298e-01 4.07067388e-01
5.12936771e-01 2.69826531e-01 -2.37628058e-01 -3.54957104e-01
3.93938512e-01 -2.28419960e-01 -2.13700667e-01 1.34165091e-02
-9.92552519e-01 6.79776371e-01 2.60481775e-01 -4.56464477e-02
-5.25300860e-01 2.66108096e-01 4.85359699e-01 8.06844458e-02
5.78744531e-01 6.67225540e-01 -7.03176796e-01 -2.88354158e-01
-1.01141679e+00 -9.21069980e-02 5.91922045e-01 7.77004004e-01
3.25594455e-01 8.28413069e-02 -3.77009541e-01 5.05199134e-01
4.01483387e-01 2.96196640e-01 5.21999478e-01 -6.95664763e-01
1.65086895e-01 4.75192040e-01 2.18306988e-01 -8.20583880e-01
-5.54553509e-01 -7.61141002e-01 -1.02333760e+00 4.72864836e-01
3.45555067e-01 -2.23536953e-01 -1.33110011e+00 1.86842203e+00
4.35302138e-01 3.79782796e-01 1.68379992e-01 6.40918493e-01
8.11283529e-01 5.09071171e-01 5.04863620e-01 -3.50175411e-01
1.35851192e+00 -6.17301583e-01 -1.06020582e+00 -2.91119218e-01
4.01479274e-01 -7.05444396e-01 6.34273529e-01 6.15955114e-01
-7.56616652e-01 -4.03368592e-01 -1.14675844e+00 2.45806292e-01
6.20673709e-02 4.69881948e-03 4.12300766e-01 7.41944313e-01
-3.20259452e-01 6.31801784e-01 -9.42069173e-01 2.82745034e-01
6.89009368e-01 3.14335763e-01 -2.77898908e-01 1.63861275e-01
-1.21954739e+00 1.08441103e+00 7.14392424e-01 9.13441554e-02
-9.13863599e-01 -1.18683159e+00 -9.76366878e-01 -7.76339769e-02
6.43333077e-01 -5.76938927e-01 1.30219543e+00 -5.84112942e-01
-1.10701132e+00 3.89843136e-01 1.12813644e-01 -8.56575966e-01
9.65063214e-01 -3.79282296e-01 -3.08600008e-01 -2.30960727e-01
-5.23065776e-02 5.72246015e-01 1.08623528e+00 -1.35517609e+00
-6.15455151e-01 -3.77982497e-01 -4.24569339e-01 -7.30951428e-02
2.49464110e-01 -4.79055166e-01 -2.83901505e-02 -7.45608449e-01
5.41478455e-01 -9.89208102e-01 -5.18429339e-01 -1.09454148e-01
-7.40902185e-01 1.59544218e-02 6.88617945e-01 -6.59458756e-01
1.29938579e+00 -1.91715837e+00 -2.75824130e-01 3.91227216e-01
3.13881874e-01 1.37593672e-01 3.26170743e-01 -2.09322184e-01
-2.46835247e-01 4.00700271e-01 -8.47085178e-01 -3.43708396e-01
-3.36689413e-01 3.60548198e-01 2.53911428e-02 3.23557973e-01
5.05637288e-01 6.83271825e-01 -6.79195940e-01 -6.81072116e-01
3.87948453e-01 4.31147009e-01 -1.69364110e-01 2.06917167e-01
-3.55763167e-01 8.60655248e-01 -7.12965906e-01 3.58409554e-01
7.70280361e-01 -2.23139882e-01 -2.53009737e-01 -4.72531199e-01
3.06419373e-01 1.18915420e-02 -1.30561495e+00 1.35168695e+00
-5.90702951e-01 2.11682990e-01 -2.52236813e-01 -8.64686012e-01
8.85301471e-01 3.63093019e-01 3.16389710e-01 -8.80319029e-02
4.12084967e-01 1.14114672e-01 1.74382374e-01 -5.48465669e-01
1.87421322e-01 -4.86317486e-01 -1.14843585e-01 3.01279545e-01
-8.20770934e-02 -5.39375067e-01 -3.22784573e-01 1.89409018e-01
1.02397692e+00 -3.22093144e-02 6.43434882e-01 -4.55340683e-01
5.44890106e-01 3.66537571e-02 8.01071107e-01 8.44885468e-01
-2.68952161e-01 6.83371127e-01 6.90826356e-01 -3.97633404e-01
-9.33342576e-01 -9.83366966e-01 -6.76627219e-01 9.48124975e-02
-2.09109709e-01 -3.95812467e-02 -6.79844797e-01 -9.91812766e-01
5.82738519e-02 1.16335142e+00 -1.02026224e+00 -3.55521321e-01
2.57962476e-02 -8.65119040e-01 3.48010033e-01 6.38651848e-01
3.38987857e-01 -7.24185169e-01 -6.12715662e-01 1.48215592e-01
3.73528451e-02 -1.21122956e+00 -2.14551479e-01 4.03695375e-01
-1.02761650e+00 -1.17856991e+00 -2.94867873e-01 1.28727285e-02
9.46895838e-01 -5.77927649e-01 1.14908028e+00 -2.19511494e-01
-3.83530051e-01 8.49553421e-02 -9.41575468e-02 -9.90198195e-01
-8.32561791e-01 -3.73799324e-01 -9.92377475e-03 -1.10268109e-01
1.87212780e-01 -2.73400456e-01 -4.59003538e-01 1.15119331e-01
-1.16820955e+00 -5.96951805e-02 4.92700279e-01 1.05536652e+00
9.47638571e-01 4.46348310e-01 8.90458941e-01 -1.67859328e+00
6.24416232e-01 -5.40172935e-01 -6.59044564e-01 1.26860261e-01
-1.13152587e+00 4.90428507e-01 3.42651218e-01 -2.02928707e-01
-1.43769097e+00 4.12827641e-01 -9.93166417e-02 -3.91206175e-01
-1.46056235e-01 7.20014215e-01 1.20194323e-01 1.40603542e-01
7.58862317e-01 -3.24561924e-01 2.33030498e-01 -2.20163360e-01
4.93620783e-02 6.29480600e-01 4.41376835e-01 -5.18696606e-01
4.35825199e-01 2.70519108e-01 4.17910069e-01 -4.06681210e-01
-1.19239497e+00 -1.07975535e-01 -3.66234720e-01 -1.20516956e-01
8.92774045e-01 -7.01883197e-01 -5.33204138e-01 2.90899038e-01
-9.17309463e-01 2.51294166e-01 -6.68124914e-01 7.71641374e-01
-2.97777981e-01 4.28127736e-01 -2.67196327e-01 -1.20652878e+00
-5.78065515e-01 -1.32265437e+00 9.51010585e-01 2.61370838e-01
-3.81524742e-01 -1.01490915e+00 -1.48472682e-01 4.76088971e-01
3.81052941e-01 6.03130698e-01 9.98800099e-01 -1.05002558e+00
-2.98714250e-01 -5.63201070e-01 -1.68243051e-01 6.46411359e-01
2.21343935e-01 -1.81588709e-01 -1.16303122e+00 2.66371906e-01
5.29797137e-01 -3.72952074e-01 8.83508384e-01 6.79174542e-01
1.45693958e+00 -2.01737702e-01 4.38057026e-03 -9.75206960e-03
1.50837374e+00 2.13593781e-01 5.30417085e-01 -4.14890796e-02
3.78031105e-01 5.94890237e-01 8.71527255e-01 6.42129302e-01
3.53719760e-03 3.25371534e-01 5.80255628e-01 -4.95924102e-03
1.05353996e-01 -6.60414249e-02 -3.62325698e-01 3.28822792e-01
2.23445207e-01 -8.75934213e-03 -1.12573576e+00 3.68279010e-01
-1.99109542e+00 -4.12003040e-01 -2.91461915e-01 2.28776360e+00
1.04260993e+00 4.58937258e-01 -4.88356024e-01 3.77309233e-01
6.86433613e-01 -2.58279085e-01 -8.32875907e-01 -2.79361635e-01
1.63803235e-01 3.16402689e-02 5.38684487e-01 6.39862895e-01
-1.05794251e+00 5.05629063e-01 6.13586998e+00 9.56921160e-01
-6.90886736e-01 2.22029969e-01 1.48091340e+00 1.27131790e-01
-3.54652673e-01 -2.42807001e-01 -3.52280408e-01 4.96223092e-01
9.24549580e-01 -8.05452764e-02 -2.03743294e-01 1.04434907e+00
2.55805552e-01 -7.61740923e-01 -1.39349520e+00 8.61433148e-01
-6.90443739e-02 -1.22266042e+00 -1.54666096e-01 -2.18963012e-01
8.21242452e-01 -4.64916319e-01 1.98436573e-01 1.73060074e-02
8.85430723e-02 -1.25274885e+00 4.91278231e-01 7.43970215e-01
6.21913135e-01 -9.18072999e-01 1.47613907e+00 4.65234488e-01
-4.64813501e-01 1.62530035e-01 -1.86334595e-01 1.94093913e-01
2.25114435e-01 1.23434150e+00 -1.10114467e+00 3.99021924e-01
5.24386346e-01 1.35320753e-01 -5.03543019e-01 8.03582191e-01
-2.28468210e-01 6.93363190e-01 -3.88902336e-01 1.74483269e-01
-2.00683344e-02 -1.35882989e-01 5.16179979e-01 9.51347768e-01
4.90361229e-02 3.49491447e-01 -1.47164702e-01 1.05198038e+00
7.34661222e-02 -1.85507357e-01 -2.99464196e-01 -2.54168380e-02
3.93428534e-01 1.13268709e+00 -6.77226841e-01 -4.73290324e-01
4.56758663e-02 5.37711024e-01 -1.49834439e-01 -4.66999114e-02
-8.81092429e-01 2.55592708e-02 2.64272988e-01 -8.14971626e-02
-3.16990972e-01 4.50356275e-01 -1.00621414e+00 -6.15508974e-01
-2.61827353e-02 -5.87216794e-01 4.96808052e-01 -6.19790316e-01
-1.36660254e+00 9.94445026e-01 9.12698209e-02 -1.17119658e+00
-3.43564004e-01 -3.78333837e-01 -3.15041363e-01 9.23560381e-01
-1.25442994e+00 -8.04044545e-01 -2.91814923e-01 1.31151110e-01
5.48531532e-01 -7.33644664e-02 8.23397934e-01 7.50335827e-02
-5.01801968e-01 4.71235305e-01 -1.46140695e-01 -3.01021874e-01
6.76085651e-01 -1.24197769e+00 -5.14266714e-02 7.57550061e-01
-3.45470279e-01 3.32848579e-01 1.06486392e+00 -9.85035717e-01
-7.85679102e-01 -1.23163331e+00 5.25138557e-01 -6.15370035e-01
4.95813996e-01 5.56271076e-02 -9.70012069e-01 4.66488063e-01
-1.46143764e-01 3.25329512e-01 7.08789706e-01 -7.68151581e-02
2.92799752e-02 -4.94854432e-03 -1.88822663e+00 1.64519757e-01
3.54816139e-01 -2.65243948e-01 -6.23316228e-01 2.57301748e-01
9.33800161e-01 -7.11150050e-01 -1.14580595e+00 7.77026296e-01
4.19581592e-01 -7.33351648e-01 5.82726955e-01 -6.26226962e-01
6.80067420e-01 -2.93539852e-01 -8.07559192e-02 -1.34777021e+00
2.41286345e-02 -1.87566087e-01 -1.10944524e-01 1.14530635e+00
8.87965143e-01 -5.88087559e-01 9.35836256e-01 1.39346445e+00
-1.27254948e-01 -8.85166526e-01 -1.16412938e+00 -4.68266189e-01
-1.18878230e-01 -9.41555560e-01 4.61852044e-01 7.77824342e-01
-3.51706326e-01 -7.14607490e-03 -4.34795588e-01 4.19630796e-01
9.20368195e-01 -5.53020775e-01 1.94788530e-01 -1.41026330e+00
-1.48223758e-01 1.79024994e-01 -6.06845975e-01 -3.36635858e-02
-1.20041184e-02 -4.73565549e-01 3.47260147e-01 -1.33197331e+00
1.98757902e-01 -6.20971739e-01 -4.52156484e-01 2.47173265e-01
-3.65347713e-01 -7.97139108e-03 -1.98686719e-02 -1.77850127e-02
-1.64704174e-01 4.93039787e-01 1.07781017e+00 -1.25549585e-01
-1.37796879e-01 4.06405717e-01 -6.38412893e-01 9.30780172e-01
7.58933187e-01 -8.12770605e-01 -7.47940242e-01 -9.27845314e-02
3.02452594e-01 2.69645900e-01 2.02664301e-01 -1.12247515e+00
1.80683628e-01 4.47636321e-02 4.71504927e-01 -5.72775960e-01
1.34748131e-01 -1.27288508e+00 3.84253234e-01 5.84791183e-01
-6.71518028e-01 -2.57491469e-01 4.47007388e-01 1.03709662e+00
-2.55201876e-01 -5.76454103e-01 1.10163820e+00 -8.19825307e-02
-3.78159046e-01 5.15888818e-02 -2.10086331e-01 -6.09536795e-03
1.15116704e+00 3.78495976e-02 -1.04983605e-01 -3.12080055e-01
-1.30413902e+00 1.38113037e-01 -1.08639389e-01 3.06629598e-01
7.12962151e-01 -9.98291850e-01 -7.74002790e-01 4.64554280e-02
2.95449644e-01 3.58312756e-01 5.66212416e-01 7.55884945e-01
-3.03781450e-01 4.04450059e-01 1.56347558e-01 -9.29177463e-01
-1.07697594e+00 3.45647812e-01 2.85061210e-01 -4.61311489e-01
-3.21453154e-01 9.32248950e-01 9.52421352e-02 -2.54889488e-01
3.74699354e-01 -5.73191941e-01 -2.10069612e-01 -1.53928980e-01
5.15022814e-01 4.11747545e-01 3.29907000e-01 -2.17903495e-01
-1.96675465e-01 1.96493849e-01 -1.98792487e-01 -1.65382430e-01
1.16600668e+00 -5.80347888e-02 -2.11628694e-02 6.52389467e-01
6.19356990e-01 -3.68817985e-01 -1.21659899e+00 -3.06755662e-01
8.03026557e-02 -4.03786212e-01 6.31879926e-01 -1.40509534e+00
-8.79777491e-01 6.60211325e-01 9.78905022e-01 -3.64673287e-02
1.05963469e+00 -8.75473693e-02 2.95771986e-01 2.48203486e-01
2.32591987e-01 -1.24895144e+00 -2.01789334e-01 4.75058593e-02
1.02990866e+00 -1.90943182e+00 1.40853837e-01 -7.15578854e-01
-1.15753555e+00 8.60935926e-01 6.03228509e-01 1.51591763e-01
1.16411972e+00 5.15238523e-01 1.19617367e-02 -2.74700508e-03
-6.14153385e-01 3.18501830e-01 5.54304063e-01 4.07130390e-01
4.16696727e-01 6.54134825e-02 -3.52380216e-01 1.11683786e+00
1.10935934e-01 2.38338992e-01 5.20673394e-01 7.65022993e-01
-2.59626210e-01 -7.06164062e-01 -5.39778173e-01 9.64475572e-01
-6.30437434e-01 -1.41117290e-01 1.17244013e-01 5.59784234e-01
1.79508895e-01 1.18362796e+00 1.10576581e-03 -3.74849051e-01
1.98633716e-01 1.30152717e-01 4.31196541e-02 -6.81077540e-01
-4.80343074e-01 -1.58086475e-02 3.39564294e-01 -5.56701005e-01
-2.67383695e-01 -5.30577838e-01 -1.46674752e+00 2.04233497e-01
-4.66341764e-01 2.81030983e-01 1.01618183e+00 1.12806690e+00
2.55395472e-01 1.13764596e+00 3.86662096e-01 -1.21278085e-01
-8.74091387e-01 -1.17959261e+00 -5.27909040e-01 2.48482004e-01
3.39444906e-01 -6.61863029e-01 -4.63854134e-01 -3.29746753e-02] | [14.380502700805664, -2.0491151809692383] |
d6faf967-0e76-4a96-ac0d-f5d9367f16a2 | sparse-range-constrained-learning-and-its | 1807.10571 | null | http://arxiv.org/abs/1807.10571v1 | http://arxiv.org/pdf/1807.10571v1.pdf | Sparse Range-constrained Learning and Its Application for Medical Image Grading | Sparse learning has been shown to be effective in solving many real-world
problems. Finding sparse representations is a fundamentally important topic in
many fields of science including signal processing, computer vision, genome
study and medical imaging. One important issue in applying sparse
representation is to find the basis to represent the data,especially in
computer vision and medical imaging where the data is not necessary incoherent.
In medical imaging, clinicians often grade the severity or measure the risk
score of a disease based on images. This process is referred to as medical
image grading. Manual grading of the disease severity or risk score is often
used. However, it is tedious, subjective and expensive. Sparse learning has
been used for automatic grading of medical images for different diseases. In
the grading, we usually begin with one step to find a sparse representation of
the testing image using a set of reference images or atoms from the dictionary.
Then in the second step, the selected atoms are used as references to compute
the grades of the testing images. Since the two steps are conducted
sequentially, the objective function in the first step is not necessarily
optimized for the second step. In this paper, we propose a novel sparse
range-constrained learning(SRCL)algorithm for medical image grading.Different
from most of existing sparse learning algorithms, SRCL integrates the objective
of finding a sparse representation and that of grading the image into one
function. It aims to find a sparse representation of the testing image based on
atoms that are most similar in both the data or feature representation and the
medical grading scores. We apply the new proposed SRCL to CDR computation and
cataract grading. Experimental results show that the proposed method is able to
improve the accuracy in cup-to-disc ratio computation and cataract grading. | ['Cheng Jun'] | 2018-07-11 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 2.92559117e-01 -2.98622489e-01 -1.34817094e-01 -3.16301674e-01
-6.93243980e-01 -1.75714791e-01 -2.25154217e-02 2.76188016e-01
-1.40267879e-01 6.03759527e-01 3.76853317e-01 1.45858064e-01
-5.40531576e-01 -7.51130283e-01 -1.38421580e-01 -8.51877213e-01
1.52377069e-01 4.67245609e-01 1.88255683e-01 5.28833903e-02
4.06233847e-01 4.10336524e-01 -1.67753839e+00 2.52020359e-01
1.19655406e+00 1.02862179e+00 5.24198294e-01 4.28116471e-01
-1.47387356e-01 7.34061182e-01 -6.11018956e-01 -4.49070521e-02
3.10932308e-01 -9.19673145e-01 -4.31242406e-01 4.25785840e-01
3.74961674e-01 2.49763932e-02 9.76103321e-02 1.50201201e+00
6.43188059e-01 8.02999735e-02 8.51594448e-01 -7.98265755e-01
-4.51294571e-01 1.22772306e-01 -8.37536752e-01 1.92671806e-01
3.39988053e-01 -3.76365960e-01 5.29683769e-01 -8.39097738e-01
4.25282031e-01 9.09581006e-01 3.70495230e-01 2.10472479e-01
-6.85214162e-01 -5.25462568e-01 -1.83703691e-01 3.54012340e-01
-1.48862171e+00 -2.30903611e-01 8.25977504e-01 -5.77680349e-01
9.47987884e-02 3.71438473e-01 8.35997760e-01 4.61052284e-02
1.90194517e-01 3.04596782e-01 1.33119559e+00 -5.12069523e-01
3.31168234e-01 9.42358673e-02 1.29079938e-01 1.17203724e+00
4.56379086e-01 -2.04701066e-01 -2.38487989e-01 -1.92714125e-01
7.14666843e-01 3.13643932e-01 -5.26864767e-01 -2.68439174e-01
-1.02777565e+00 9.51019824e-01 4.63004082e-01 6.06582582e-01
-5.46562850e-01 -3.28427374e-01 9.54575092e-02 8.83978382e-02
1.69003919e-01 2.68742085e-01 1.84011385e-02 2.25739956e-01
-1.29563701e+00 3.32923286e-04 5.61979175e-01 2.50400931e-01
7.26402342e-01 4.24856693e-03 -1.70501113e-01 1.14666271e+00
2.99365014e-01 4.04537976e-01 9.62107599e-01 -9.00038183e-01
4.43746969e-02 8.50661278e-01 -2.51878668e-02 -1.27316320e+00
-1.91465735e-01 -4.94457006e-01 -1.11591089e+00 2.89401948e-01
3.22932541e-01 -1.12800300e-01 -1.09147978e+00 1.23010993e+00
3.18694770e-01 5.24380505e-01 -5.86461313e-02 1.27411413e+00
9.72833693e-01 6.62003875e-01 -2.42907643e-01 -6.24402404e-01
1.34269917e+00 -5.93701959e-01 -5.06389737e-01 -5.34471944e-02
4.72629339e-01 -1.07176077e+00 7.23070204e-01 6.00624502e-01
-8.77628922e-01 -4.63858008e-01 -8.68197501e-01 1.46280393e-01
1.56172663e-01 6.76239014e-01 6.63873494e-01 3.33903521e-01
-8.44003677e-01 2.88228065e-01 -5.92956305e-01 -4.02485222e-01
3.53597611e-01 4.18825895e-01 -4.17643696e-01 -4.90619451e-01
-7.27542043e-01 7.87949324e-01 4.15859856e-02 6.39506429e-02
-6.15273058e-01 -5.31814218e-01 -8.31964850e-01 7.70795494e-02
1.24713346e-01 -7.25982130e-01 7.58121848e-01 -9.62746859e-01
-1.16121423e+00 9.27487791e-01 -4.26638305e-01 -2.44642884e-01
8.97049904e-02 1.64338589e-01 -3.16006154e-01 2.26069793e-01
3.30641180e-01 2.38750264e-01 8.13432574e-01 -9.87301230e-01
-6.39384866e-01 -4.70963538e-01 -2.56219089e-01 3.91565144e-01
-1.74714550e-01 -1.36144124e-02 -6.05630159e-01 -6.98788226e-01
5.41772664e-01 -7.84197569e-01 -3.88643473e-01 1.37906335e-02
-1.51090980e-01 -1.66993871e-01 7.60922134e-01 -8.26848030e-01
1.14475226e+00 -2.22208977e+00 3.76078963e-01 5.76628804e-01
2.10796133e-01 2.40059778e-01 4.68212478e-02 -1.85265124e-01
-2.17159376e-01 -1.97128832e-01 -4.02576238e-01 1.32539282e-02
-6.84223175e-01 6.98114187e-02 3.35109860e-01 5.82278669e-01
-6.83655888e-02 2.94477284e-01 -7.47392714e-01 -8.67147744e-01
1.94442093e-01 4.47499186e-01 -6.26664817e-01 3.23787570e-01
1.90014243e-01 6.17114902e-01 -5.16414404e-01 8.24865639e-01
4.25461084e-01 -4.64003891e-01 -5.22872014e-03 -5.84264040e-01
-7.55159259e-02 -2.45124951e-01 -1.45926499e+00 1.66318405e+00
-3.03596556e-01 3.76238614e-01 -8.96007344e-02 -1.33990252e+00
1.06435704e+00 4.07936990e-01 8.18993568e-01 -5.00917256e-01
2.41231158e-01 3.55170280e-01 2.88199782e-01 -7.43890464e-01
-5.03962412e-02 -3.55808735e-01 2.97886282e-01 2.10242480e-01
-7.56694451e-02 -3.58522147e-01 5.00408173e-01 3.48765291e-02
8.69885325e-01 -4.06242341e-01 6.19079828e-01 3.43758352e-02
9.54568326e-01 5.18497415e-02 8.90663266e-01 1.76433519e-01
4.73538227e-02 7.70268679e-01 3.85002047e-01 -4.29349571e-01
-7.85381496e-01 -7.39424229e-01 -1.95471466e-01 2.85582513e-01
1.59212753e-01 -2.43536577e-01 -4.93104964e-01 -4.58102643e-01
-1.28599882e-01 5.22198938e-02 -4.20159847e-01 -1.53961316e-01
-3.85500610e-01 -7.92390406e-01 -2.10813761e-01 8.20357725e-02
4.59285200e-01 -8.19783807e-01 -3.43569756e-01 -4.12536599e-02
-1.09682113e-01 -5.87023795e-01 -4.97070849e-01 -2.54220605e-01
-1.06510770e+00 -1.27827561e+00 -1.18208706e+00 -1.17683625e+00
1.25163054e+00 4.00309920e-01 6.75745249e-01 4.29026425e-01
-5.47949672e-01 -2.90905591e-02 -3.56953621e-01 -2.36066848e-01
-8.85004029e-02 -3.92007738e-01 -1.04724288e-01 3.49146664e-01
3.63709927e-02 -4.78466630e-01 -8.34473968e-01 2.92345524e-01
-7.13816941e-01 -2.15479150e-01 9.59505200e-01 8.50060344e-01
1.08996630e+00 4.14661646e-01 2.76190609e-01 -1.01025212e+00
6.21685207e-01 -4.75889117e-01 -6.51877761e-01 2.16655269e-01
-4.70287263e-01 -4.06978466e-02 5.86025238e-01 -2.73161024e-01
-7.54381299e-01 2.62708247e-01 -9.72819701e-02 -5.27609825e-01
4.90923524e-02 9.67181981e-01 4.28729318e-02 -3.63397449e-01
5.01414895e-01 3.80095243e-01 2.16503337e-01 -4.07839686e-01
-1.28459826e-01 5.58287144e-01 4.84115481e-01 -2.97946334e-01
5.50046384e-01 2.93069869e-01 2.76221752e-01 -8.66600633e-01
-9.54663038e-01 -7.78228521e-01 -2.32488021e-01 -1.96945250e-01
8.92667890e-01 -7.53952801e-01 -4.39372480e-01 1.73274085e-01
-7.44222581e-01 4.08585817e-01 -2.67817229e-01 8.79133523e-01
-3.08467537e-01 5.95848858e-01 -1.76537216e-01 -5.40748715e-01
-4.15806770e-01 -1.44288719e+00 9.35817659e-01 6.16107523e-01
-1.06596380e-01 -7.12763131e-01 1.26259148e-01 4.90761012e-01
2.42726147e-01 2.19882011e-01 1.13257551e+00 -2.64102340e-01
-3.93325150e-01 -4.71368909e-01 -1.81567624e-01 4.27632183e-01
5.74044526e-01 -3.39777470e-01 -3.77269983e-01 -2.92125076e-01
2.76094437e-01 -1.50793940e-01 7.42375910e-01 7.84299791e-01
1.25261593e+00 -2.42908165e-01 -1.39653593e-01 7.87201464e-01
1.57951140e+00 5.10848761e-01 6.59735858e-01 -1.32276028e-01
6.29055202e-01 5.03405631e-01 7.37047255e-01 3.71769637e-01
4.29060869e-02 5.79702675e-01 2.40636975e-01 -3.65668416e-01
-3.46617877e-01 1.16030619e-01 7.72402361e-02 1.16568720e+00
-3.18221688e-01 2.02612862e-01 -7.64105380e-01 6.33245647e-01
-1.67059684e+00 -9.94339705e-01 -3.61094400e-02 2.47393799e+00
7.76660740e-01 -2.40826949e-01 -1.08653478e-01 3.97503674e-01
7.71008074e-01 -2.42170498e-01 -4.55720216e-01 -1.93560869e-01
1.41331315e-01 4.72896904e-01 1.11103915e-01 4.66020882e-01
-7.44682431e-01 4.84255970e-01 5.49055624e+00 6.63433969e-01
-1.32181454e+00 1.33112580e-01 6.12726331e-01 -2.68808212e-02
-6.98594004e-02 3.89016084e-02 -6.11345649e-01 6.08726084e-01
3.94325823e-01 -2.23268494e-01 3.37265253e-01 6.89518988e-01
3.78896892e-01 -5.26015818e-01 -7.06086814e-01 1.24446487e+00
3.89455676e-01 -1.18268275e+00 2.94797063e-01 -1.28769651e-01
9.54602480e-01 -4.56724763e-01 5.69911040e-02 -2.07349285e-01
-1.02253973e-01 -1.09134316e+00 7.13976547e-02 7.92030632e-01
9.39178109e-01 -7.69615412e-01 1.07686949e+00 2.48170048e-01
-1.12482965e+00 -1.82207394e-02 -5.00981510e-01 1.73478201e-01
-1.65626518e-02 9.39848244e-01 -8.01613152e-01 3.59756261e-01
5.44157684e-01 8.62493992e-01 -4.27463144e-01 1.60380745e+00
-2.18964070e-01 5.40634394e-01 -6.62924647e-02 1.71649992e-01
2.89073735e-02 -6.33564115e-01 5.35165131e-01 6.93602622e-01
8.91391456e-01 2.91016966e-01 4.21258301e-01 5.75338125e-01
4.55680415e-02 5.01235723e-01 -6.50861263e-01 6.87191710e-02
2.02220976e-01 1.22260439e+00 -8.28243732e-01 -3.19005102e-01
-4.91855681e-01 8.23585749e-01 -1.10318027e-01 1.95933923e-01
-3.41075689e-01 -3.82127613e-01 3.34481299e-01 3.52427363e-01
1.27034485e-01 -2.94772591e-02 -1.20240919e-01 -1.04175794e+00
-1.61472097e-01 -1.12185681e+00 4.97923315e-01 -7.63274252e-01
-9.95463133e-01 4.66347724e-01 -2.82263041e-01 -1.50960016e+00
-1.73589230e-01 -2.59160578e-01 -7.26128042e-01 1.11819243e+00
-1.52314043e+00 -8.92602801e-01 -6.75902247e-01 9.84961510e-01
5.84172785e-01 -4.33138072e-01 6.84923410e-01 4.12028939e-01
-4.46708798e-01 2.51250207e-01 -7.17674866e-02 8.77221301e-02
7.87052214e-01 -1.12817323e+00 -8.83217275e-01 9.44522738e-01
1.45020396e-01 6.74476683e-01 5.38164973e-01 -6.83280110e-01
-9.82316732e-01 -9.28467512e-01 8.12517107e-01 3.61194402e-01
2.15601042e-01 4.76220906e-01 -7.36880124e-01 1.78406790e-01
-6.38288483e-02 2.91437238e-01 8.95454347e-01 -1.12502180e-01
2.85545051e-01 -3.70551139e-01 -1.25158608e+00 2.75608301e-01
5.18921971e-01 -2.02512488e-01 -5.32319188e-01 5.06700575e-01
1.06852546e-01 -4.46401775e-01 -9.32198584e-01 4.10957694e-01
4.07628685e-01 -8.58777523e-01 8.32727134e-01 -3.47668648e-01
4.66494441e-01 -7.58564055e-01 -1.25336602e-01 -1.37281144e+00
-4.93395418e-01 -1.22754797e-01 1.65075108e-01 8.79778624e-01
1.36672065e-01 -4.07742530e-01 9.02886987e-01 1.74790306e-03
-1.41860545e-01 -9.12787199e-01 -7.42357194e-01 -2.39545077e-01
-4.46459025e-01 1.87562600e-01 1.98172823e-01 9.15158689e-01
-3.20775390e-01 2.04115942e-01 -3.57261956e-01 4.44828182e-01
7.86282063e-01 5.87218106e-01 4.15382355e-01 -1.52886224e+00
-3.93048406e-01 -2.41360590e-01 -7.80098975e-01 -4.34069067e-01
-2.30395123e-01 -9.56173956e-01 -1.86612368e-01 -1.95130944e+00
4.38433379e-01 -5.50559938e-01 -4.40053433e-01 3.17994595e-01
-2.62626946e-01 4.16416973e-01 9.13321041e-03 4.81327295e-01
-2.52526909e-01 1.96940646e-01 1.51169050e+00 -2.70312726e-01
-3.43175173e-01 1.63736358e-01 -9.22870994e-01 8.20780694e-01
7.03452468e-01 -5.26514590e-01 -5.97368121e-01 -1.96330518e-01
2.03191563e-01 3.03087145e-01 1.34792015e-01 -1.21956289e+00
3.42290103e-01 -1.52483732e-01 4.85631257e-01 -3.85949999e-01
1.39202431e-01 -8.43859732e-01 1.34014666e-01 7.22165704e-01
8.15979168e-02 -1.13618895e-01 -2.34453857e-01 3.60629916e-01
-6.67698920e-01 -6.09499037e-01 1.07358086e+00 -3.77416670e-01
-6.88242495e-01 3.85838866e-01 -2.16911398e-02 -1.50476694e-01
1.07102823e+00 -3.27017128e-01 -2.52675787e-02 -3.89672995e-01
-9.11716163e-01 8.37827101e-02 3.15865278e-01 -3.03782858e-02
9.33092356e-01 -1.22093654e+00 -9.18687165e-01 1.95636734e-01
8.41589570e-02 9.35160592e-02 2.48461202e-01 1.15857041e+00
-6.24575973e-01 2.81126857e-01 -2.47871265e-01 -6.83518350e-01
-1.62742341e+00 3.30480933e-01 2.69574642e-01 -3.55956763e-01
-3.64483923e-01 7.88656235e-01 1.12868570e-01 2.03474492e-01
8.41020122e-02 -3.15325081e-01 -7.38230586e-01 1.22948863e-01
5.52214265e-01 2.90700316e-01 1.20812446e-01 -7.24877477e-01
-1.49222136e-01 1.26160121e+00 -6.76639006e-02 2.70584196e-01
1.43484724e+00 1.81348994e-01 -5.75193882e-01 2.51018763e-01
1.06164694e+00 3.10331076e-01 -4.32914436e-01 -1.76322401e-01
-3.29138339e-01 -5.41411042e-01 4.26937044e-01 -7.50800073e-01
-1.50051177e+00 6.08661294e-01 1.04794538e+00 -1.15654454e-01
1.50221348e+00 -1.37449071e-01 6.94466054e-01 7.99127594e-02
4.70171899e-01 -7.16972828e-01 2.89656669e-02 1.07847832e-01
9.32017088e-01 -1.32654762e+00 3.07895154e-01 -5.82528770e-01
-5.07657290e-01 8.68394256e-01 3.38976234e-01 -2.85185397e-01
7.68851042e-01 1.40708178e-01 5.39754741e-02 -3.39800715e-01
-1.27227932e-01 -2.12010115e-01 6.68860555e-01 4.77669567e-01
6.13586664e-01 -3.63159962e-02 -8.94735873e-01 4.57623005e-01
-1.36209726e-01 2.98989922e-01 3.07580769e-01 7.27506936e-01
-6.97817624e-01 -1.17956114e+00 -7.20571518e-01 8.90855014e-01
-4.86085385e-01 -1.30194817e-02 -1.87582627e-01 1.28108963e-01
5.06660402e-01 7.89111376e-01 -1.90230533e-01 -9.59148705e-02
1.90824300e-01 -8.51333141e-02 4.54948515e-01 -9.84782517e-01
-1.91587567e-01 1.77609056e-01 -2.74273962e-01 -5.66739738e-01
-6.15653992e-01 -5.62643766e-01 -1.33017004e+00 2.18306392e-01
-2.96847045e-01 4.64086533e-01 6.02282226e-01 8.35801482e-01
1.18939243e-01 4.10553277e-01 6.80435896e-01 -4.65289533e-01
-1.30655795e-01 -7.51435101e-01 -7.79461443e-01 4.59752083e-01
2.45992720e-01 -7.71294892e-01 -2.64934927e-01 2.38724858e-01] | [12.465533256530762, 0.35658663511276245] |
2648aab0-4c54-4188-a79b-72bdda4c9ef0 | generalizability-of-deep-adult-lung | 2211.02475 | null | https://arxiv.org/abs/2211.02475v2 | https://arxiv.org/pdf/2211.02475v2.pdf | Generalizability of Deep Adult Lung Segmentation Models to the Pediatric Population: A Retrospective Study | Lung segmentation in chest X-rays (CXRs) is an important prerequisite for improving the specificity of diagnoses of cardiopulmonary diseases in a clinical decision support system. Current deep learning models for lung segmentation are trained and evaluated on CXR datasets in which the radiographic projections are captured predominantly from the adult population. However, the shape of the lungs is reported to be significantly different across the developmental stages from infancy to adulthood. This might result in age-related data domain shifts that would adversely impact lung segmentation performance when the models trained on the adult population are deployed for pediatric lung segmentation. In this work, our goal is to (i) analyze the generalizability of deep adult lung segmentation models to the pediatric population and (ii) improve performance through a stage-wise, systematic approach consisting of CXR modality-specific weight initializations, stacked ensembles, and an ensemble of stacked ensembles. To evaluate segmentation performance and generalizability, novel evaluation metrics consisting of mean lung contour distance (MLCD) and average hash score (AHS) are proposed in addition to the multi-scale structural similarity index measure (MS-SSIM), the intersection of union (IoU), Dice score, 95% Hausdorff distance (HD95), and average symmetric surface distance (ASSD). Our results showed a significant improvement (p < 0.05) in cross-domain generalization through our approach. This study could serve as a paradigm to analyze the cross-domain generalizability of deep segmentation models for other medical imaging modalities and applications. | ['Sameer Antani', 'Zhiyun Xue', 'Ghada Zamzmi', 'Feng Yang', 'Sivaramakrishnan Rajaraman'] | 2022-11-04 | null | null | null | null | ['ms-ssim'] | ['computer-vision'] | [ 9.17315111e-02 -3.88728678e-02 -1.38134003e-01 -3.66274953e-01
-7.40402222e-01 -6.38161480e-01 3.09773266e-01 3.56491774e-01
-4.33667094e-01 5.45318127e-01 1.40938535e-01 -6.36303067e-01
-3.24261427e-01 -6.94292247e-01 -3.43680203e-01 -7.02967346e-01
-1.63901551e-03 7.80291557e-01 4.97276247e-01 2.15764359e-01
-1.09003454e-01 7.98497200e-01 -9.83969271e-01 2.14435205e-01
1.10491741e+00 8.38860691e-01 1.93960980e-01 6.67698026e-01
-1.30610213e-01 5.07477820e-01 -3.89282465e-01 -3.19526017e-01
2.11619809e-01 -5.24878860e-01 -8.37935984e-01 -1.81706741e-01
6.89712405e-01 -4.44760233e-01 -2.42120430e-01 7.33220816e-01
8.60836923e-01 7.91575983e-02 8.57558787e-01 -6.26326501e-01
-3.40107352e-01 6.15188777e-01 -3.63068879e-01 5.81089854e-01
-1.19796976e-01 3.44773114e-01 6.47442877e-01 -7.27147341e-01
2.31775150e-01 6.90948009e-01 1.02803504e+00 7.42364347e-01
-7.81076670e-01 -4.93913233e-01 -3.15013647e-01 -1.94300458e-01
-1.09627616e+00 2.77626753e-01 2.65189111e-01 -7.19216347e-01
6.21233463e-01 4.67125386e-01 7.53458560e-01 6.18405521e-01
3.74574155e-01 4.30505395e-01 9.43635643e-01 -6.71908036e-02
7.20288157e-02 -1.16716750e-01 6.59149140e-02 7.23032892e-01
6.34977877e-01 -9.69534926e-03 1.74374655e-01 -6.56733941e-03
8.88656139e-01 1.31364390e-01 -3.87955964e-01 -2.78934568e-01
-1.31675792e+00 7.97186434e-01 6.23090267e-01 6.81918561e-01
-4.43327755e-01 5.43461740e-03 4.25637633e-01 -2.60060579e-01
3.84817094e-01 5.42639792e-01 -3.42938334e-01 4.68643866e-02
-1.21568632e+00 6.24069758e-02 5.68675041e-01 4.10180777e-01
2.65085492e-02 -3.97680849e-02 -6.22561991e-01 9.87745404e-01
1.53492942e-01 4.80979830e-01 9.52838361e-01 -8.00838649e-01
5.34691513e-01 5.56631625e-01 -5.52252054e-01 -5.17493844e-01
-8.43674779e-01 -5.19912124e-01 -8.08921516e-01 -2.73223370e-01
5.34371138e-01 -2.14113176e-01 -1.15405655e+00 1.50080419e+00
4.45068717e-01 4.85527329e-02 -2.50891238e-01 9.95164454e-01
1.16779387e+00 1.93468899e-01 4.09009516e-01 -7.55139738e-02
1.22327423e+00 -7.12716103e-01 -8.68167058e-02 1.22409582e-01
8.69254708e-01 -4.79033232e-01 1.17775548e+00 -5.76837212e-02
-1.28884733e+00 -6.11644447e-01 -8.69508207e-01 1.26545221e-01
-1.36256754e-01 7.23562837e-02 3.65285575e-01 9.23614502e-01
-1.03238475e+00 8.44678283e-01 -1.25572181e+00 -4.21814322e-01
7.64015913e-01 4.87974018e-01 -1.56011239e-01 9.07465145e-02
-9.75883722e-01 8.81656945e-01 3.26492101e-01 -2.63486266e-01
-6.22520506e-01 -1.23128152e+00 -4.05974954e-01 9.49113667e-02
4.83183526e-02 -1.06238437e+00 9.88875628e-01 -4.19092655e-01
-1.17514515e+00 1.11800253e+00 3.42767715e-01 -4.83932287e-01
7.37768769e-01 2.01001018e-02 -1.36521468e-02 4.63063955e-01
3.66754420e-02 7.22928524e-01 3.51321042e-01 -1.01145422e+00
-5.37088752e-01 -5.83958328e-01 -4.69152302e-01 3.10003161e-01
-1.95876583e-01 -1.92031085e-01 -3.75819862e-01 -7.73125470e-01
2.38258839e-01 -1.06787550e+00 -2.63120055e-01 -8.48790258e-02
-2.24984914e-01 -1.68456972e-01 5.59079111e-01 -9.61124182e-01
1.35889065e+00 -1.84831667e+00 -1.66879743e-01 2.49662817e-01
5.01444101e-01 5.46661854e-01 2.03039601e-01 -1.22854099e-01
-2.06330851e-01 4.49962437e-01 -5.61349213e-01 -3.41145955e-02
-5.64435959e-01 -3.66480388e-02 1.32314399e-01 5.41472256e-01
-3.78637761e-02 9.09721732e-01 -6.87846005e-01 -9.06397581e-01
2.99261540e-01 2.93997675e-01 -5.96602201e-01 3.89477611e-01
7.95937553e-02 8.33433926e-01 -4.39855844e-01 4.73584443e-01
5.47895133e-01 -4.80864495e-01 -9.18703601e-02 -1.98425099e-01
2.13289969e-02 1.77604824e-01 -5.31291306e-01 1.46086752e+00
-5.15803635e-01 3.27411085e-01 -3.56427222e-01 -7.14593351e-01
6.64405406e-01 5.35193741e-01 1.10183287e+00 -4.19501156e-01
2.14794278e-01 3.21761876e-01 5.43175220e-01 -4.76204067e-01
-1.40152782e-01 -5.66077530e-01 3.26413721e-01 4.62129682e-01
-1.42957821e-01 -3.89277875e-01 2.69053187e-02 -4.48301993e-02
1.05638337e+00 -4.30210888e-01 1.70053422e-01 -4.88145113e-01
6.17323816e-01 7.36670708e-03 1.55682340e-01 6.52489245e-01
-3.74258101e-01 1.29984176e+00 2.74370372e-01 -2.77681142e-01
-1.10034275e+00 -1.59765482e+00 -6.77481771e-01 6.88532591e-01
-6.86185285e-02 1.92999974e-01 -8.92178774e-01 -9.96085048e-01
1.11006171e-01 6.57287836e-01 -4.37904596e-01 -2.05865204e-01
-8.23966742e-01 -9.70332325e-01 9.06917274e-01 1.01527119e+00
3.92559290e-01 -9.72876251e-01 -9.54555333e-01 1.96775552e-02
-1.05766607e-02 -9.60600436e-01 -5.82739532e-01 5.58023639e-02
-1.31777918e+00 -1.12748790e+00 -1.37684608e+00 -6.52010024e-01
5.77839017e-01 -1.39618129e-01 1.03750217e+00 2.91703194e-01
-3.62820864e-01 6.08178020e-01 -8.64862278e-02 -3.15073699e-01
-5.95613062e-01 2.88401663e-01 -3.03589664e-02 -5.52039564e-01
-4.07446921e-02 -2.61223674e-01 -1.06655359e+00 3.39562923e-01
-9.49700356e-01 -6.93565980e-02 4.80155408e-01 6.86229885e-01
7.77040660e-01 -2.48644784e-01 4.77874070e-01 -1.07578480e+00
6.32183611e-01 -5.32273233e-01 -2.26031989e-01 5.49623191e-01
-8.42740119e-01 -3.32242250e-01 6.80145621e-01 -4.16354656e-01
-9.00635779e-01 -3.64042819e-01 -3.55025858e-01 -6.37377799e-01
-1.72178090e-01 3.34627241e-01 4.23278868e-01 -1.55980447e-02
6.32796168e-01 6.69652894e-02 -2.89438851e-02 -3.43000978e-01
-4.34834734e-02 5.34633338e-01 5.27518868e-01 -6.36547208e-01
6.24257326e-01 5.13219953e-01 4.00564194e-01 -6.77521706e-01
-6.14369631e-01 -6.35764420e-01 -9.77767646e-01 -1.77444026e-01
1.32760179e+00 -6.66617453e-01 -1.95128515e-01 3.75453711e-01
-6.77496493e-01 -4.25821304e-01 -4.43114132e-01 8.62539053e-01
-3.04643452e-01 5.00553370e-01 -5.87794304e-01 -2.20652655e-01
-8.17958176e-01 -1.34311748e+00 8.15289378e-01 2.95161366e-01
-3.49672914e-01 -1.41227567e+00 3.34549755e-01 5.39263070e-01
4.74395007e-01 4.35691476e-01 1.31109095e+00 -1.29725647e+00
-1.05841095e-02 -4.81518283e-02 -2.48062834e-01 6.48553073e-01
2.94045597e-01 1.04954466e-01 -6.32690132e-01 -3.58096302e-01
-1.90640688e-02 -4.36372943e-02 8.80382538e-01 8.38275313e-01
1.66511655e+00 1.52028129e-01 -2.71942496e-01 9.05555189e-01
1.25805688e+00 5.70309341e-01 2.61085451e-01 8.28322489e-03
9.26496148e-01 3.82627964e-01 2.00283363e-01 3.46238077e-01
2.40837902e-01 1.62469655e-01 2.33074889e-01 -3.38258624e-01
-4.14764076e-01 -1.40416741e-01 -1.43334389e-01 9.64312315e-01
-2.10413858e-01 -1.82684958e-01 -1.54125834e+00 5.82308888e-01
-1.00329232e+00 -5.78772724e-01 -2.47318342e-01 2.26599598e+00
6.19043648e-01 2.77162008e-02 3.77651215e-01 -2.01935560e-01
8.23993325e-01 -5.59206121e-02 -6.89018607e-01 -3.12760592e-01
1.69658840e-01 6.11392558e-01 6.93701863e-01 1.76651459e-02
-9.90260601e-01 4.24645424e-01 6.34726048e+00 6.28067791e-01
-1.63535035e+00 1.47668958e-01 9.78016376e-01 -4.94023301e-02
-1.69721797e-01 -2.14491010e-01 -4.97225791e-01 6.21099234e-01
8.14140499e-01 1.55564519e-02 6.21772744e-03 6.56060636e-01
-8.64373967e-02 -1.92462131e-02 -1.17133725e+00 6.19075894e-01
-9.13833231e-02 -1.24083328e+00 3.65644172e-02 6.06891289e-02
7.82025397e-01 4.18659896e-01 4.60462272e-01 1.85929075e-01
-1.57646850e-01 -1.14605629e+00 2.31587604e-01 3.33252013e-01
1.30330801e+00 -2.91721255e-01 7.94283390e-01 1.17142871e-01
-1.00537086e+00 9.43627283e-02 -2.08624229e-01 7.42204845e-01
-4.55407538e-02 3.66468370e-01 -1.38049483e+00 4.71873313e-01
7.74540007e-01 2.97151685e-01 -7.54982412e-01 1.22968006e+00
3.53104472e-01 8.75381052e-01 -4.61592883e-01 1.46604195e-01
2.49946699e-01 -3.20424944e-01 4.44392592e-01 1.24829137e+00
5.41021883e-01 3.97185594e-01 -1.51946217e-01 8.50020051e-01
-9.39813480e-02 3.82037222e-01 -3.30859303e-01 7.08064884e-02
3.08926016e-01 1.10627353e+00 -1.14652145e+00 -3.28206480e-01
-5.20610929e-01 3.76422167e-01 -1.89987525e-01 2.33756378e-02
-1.11230862e+00 9.70579758e-02 1.26664773e-01 6.28167093e-01
1.37704894e-01 3.21945131e-01 -7.51588881e-01 -9.16064084e-01
-2.67821014e-01 -7.43277133e-01 7.85747170e-01 -4.55808550e-01
-1.30182481e+00 4.68733430e-01 2.53703296e-01 -1.16137218e+00
9.54807922e-02 -4.86851811e-01 -1.13811076e+00 7.99825847e-01
-1.05692518e+00 -7.51864076e-01 -4.12894487e-01 3.11701447e-01
4.03369248e-01 -1.82278514e-01 3.03231567e-01 3.42583150e-01
-6.16333842e-01 7.86805272e-01 8.44694749e-02 3.25636923e-01
4.86956686e-01 -1.50936091e+00 1.05573393e-01 5.30287981e-01
-1.48868144e-01 6.66779339e-01 2.19641045e-01 -7.41667211e-01
-6.75328016e-01 -1.30042219e+00 2.59090662e-01 -5.41511357e-01
3.33355188e-01 4.75576073e-01 -1.08148718e+00 5.19295096e-01
-3.05137098e-01 1.24882266e-01 9.07703280e-01 -3.21986109e-01
5.24977073e-02 -4.38295268e-02 -1.41047621e+00 4.09508437e-01
7.03521550e-01 -1.88073024e-01 -5.99726140e-01 1.40327156e-01
5.15158534e-01 -6.28146887e-01 -1.56238055e+00 1.07902861e+00
6.50617003e-01 -1.12110388e+00 1.18893981e+00 -5.38807511e-01
4.82101887e-01 6.75336982e-04 6.31082729e-02 -9.80072200e-01
-2.59126335e-01 1.94893569e-01 1.38088718e-01 9.33621049e-01
4.68759775e-01 -5.65940380e-01 9.01215911e-01 7.49022365e-01
-3.33685577e-01 -1.48644495e+00 -1.05957186e+00 -4.57620472e-01
8.22716177e-01 -2.07485065e-01 6.63869739e-01 7.49859333e-01
-4.73926723e-01 -1.05007224e-01 4.91112500e-01 3.65458839e-02
2.82381117e-01 -3.09609063e-02 2.14864850e-01 -1.14316559e+00
-2.31946111e-01 -8.10572088e-01 -9.71560702e-02 -5.84371209e-01
-1.93421930e-01 -1.21094608e+00 -2.27209151e-01 -1.84452307e+00
4.07716334e-01 -6.58063829e-01 -6.46367967e-01 7.44026527e-02
-4.53542262e-01 -5.14754876e-02 2.64272064e-01 3.68729949e-01
-2.90018111e-01 2.17453107e-01 1.66074193e+00 1.47244066e-01
-1.93891019e-01 5.24618566e-01 -3.83436292e-01 7.75634766e-01
8.45186234e-01 -5.03808916e-01 -5.00559390e-01 -3.20672959e-01
-4.17897344e-01 3.66418689e-01 2.21638918e-01 -1.31606829e+00
3.79590802e-02 6.41453639e-02 7.78242350e-01 -7.29341328e-01
-1.17538899e-01 -7.62807727e-01 4.10339870e-02 9.60982800e-01
-3.15544337e-01 2.65902966e-01 1.57448828e-01 1.96031034e-01
8.22248217e-03 -2.52256602e-01 1.24163282e+00 -2.35444307e-01
-4.69913669e-02 6.60922527e-01 -1.53064653e-01 4.60634112e-01
1.08158851e+00 -4.27452922e-01 -8.52443874e-02 -1.45113124e-02
-6.18226111e-01 1.04244284e-01 4.69765812e-01 1.02818049e-01
5.40225565e-01 -1.00146949e+00 -7.04117715e-01 1.78776309e-02
-6.86811656e-02 4.65310097e-01 3.59167933e-01 1.23521578e+00
-1.14918935e+00 6.87045932e-01 -4.64122742e-02 -9.48151648e-01
-1.17329109e+00 2.59680331e-01 8.62310648e-01 -7.39854813e-01
-5.73095977e-01 1.09505951e+00 5.72445393e-01 -4.06564295e-01
1.29058644e-01 -8.57688725e-01 -1.16948023e-01 -1.34345680e-01
-1.47414744e-01 5.80839038e-01 2.17381626e-01 -5.90414226e-01
-5.14384627e-01 8.22831213e-01 -2.81838817e-03 2.65647352e-01
1.04605198e+00 1.80124372e-01 2.60985438e-02 2.45616496e-01
1.19391048e+00 -1.46600455e-01 -1.07684016e+00 2.01096274e-02
1.28985122e-02 -7.61324447e-03 -2.41599791e-02 -7.81425297e-01
-1.35057282e+00 1.11388719e+00 1.03723538e+00 1.61072351e-02
1.17657137e+00 3.10333550e-01 1.11076987e+00 -1.08323164e-01
-1.59293726e-01 -8.89591217e-01 7.66363963e-02 4.17969264e-02
5.40246010e-01 -1.24920034e+00 1.79275364e-01 -1.69589937e-01
-8.83426666e-01 1.15487456e+00 8.65802050e-01 3.30284499e-02
7.78223634e-01 7.74623007e-02 1.51830360e-01 -2.26240292e-01
-4.03241724e-01 -6.89217672e-02 6.51583076e-01 3.94288421e-01
8.68339479e-01 1.54617652e-01 -4.55665499e-01 6.99804604e-01
-1.93267867e-01 -2.84399718e-01 8.73832330e-02 5.42179644e-01
-5.05124629e-01 -7.76135623e-01 -3.93770754e-01 9.15577590e-01
-9.04641032e-01 1.23988022e-03 -1.75627723e-01 9.36270535e-01
3.84535730e-01 3.73480499e-01 1.49602801e-01 -2.63048321e-01
2.88202375e-01 8.01852122e-02 5.69965243e-01 -8.53053868e-01
-9.13562357e-01 -1.87824503e-01 -4.49456245e-01 -7.96347857e-02
-1.95135996e-01 -7.20592797e-01 -1.55649543e+00 -9.26269963e-02
-2.09441617e-01 -7.52040446e-02 4.93003935e-01 8.66525292e-01
-7.11281076e-02 7.40322292e-01 4.93296504e-01 -2.78375089e-01
-6.86349452e-01 -8.85514021e-01 -2.86435813e-01 5.16513228e-01
1.80346370e-01 -4.61674064e-01 -4.41328764e-01 -1.65934786e-01] | [15.158475875854492, -2.086452007293701] |
a0d61da0-eb62-4414-b032-30c774729015 | hybrid-quantum-neural-network-for-drug | 2211.05777 | null | https://arxiv.org/abs/2211.05777v2 | https://arxiv.org/pdf/2211.05777v2.pdf | Hybrid quantum neural network for drug response prediction | Cancer is one of the leading causes of death worldwide. It is caused by a variety of genetic mutations, which makes every instance of the disease unique. Since chemotherapy can have extremely severe side effects, each patient requires a personalized treatment plan. Finding the dosages that maximize the beneficial effects of the drugs and minimize their adverse side effects is vital. Deep neural networks automate and improve drug selection. However, they require a lot of data to be trained on. Therefore, there is a need for machine-learning approaches that require less data. Hybrid quantum neural networks were shown to provide a potential advantage in problems where training data availability is limited. We propose a novel hybrid quantum neural network for drug response prediction, based on a combination of convolutional, graph convolutional, and deep quantum neural layers of 8 qubits with 363 layers. We test our model on the reduced Genomics of Drug Sensitivity in Cancer dataset and show that the hybrid quantum model outperforms its classical analog by 15% in predicting IC50 drug effectiveness values. The proposed hybrid quantum machine learning model is a step towards deep quantum data-efficient algorithms with thousands of quantum gates for solving problems in personalized medicine, where data collection is a challenge. | ['Alexey Melnikov', 'Tatiana Tomashuk', 'Daria Kosichkina', 'Nurbolat Kenbayev', 'Mohammad Kordzanganeh', 'Asel Sagingalieva'] | 2022-11-10 | null | null | null | null | ['drug-response-prediction'] | ['medical'] | [ 3.80943954e-01 -2.13778481e-01 -4.48375195e-01 -2.28889957e-01
-8.15735519e-01 -4.22152728e-01 3.28869708e-02 7.93225408e-01
-4.64381903e-01 9.67835426e-01 -1.85887948e-01 -5.13303041e-01
-1.61862209e-01 -1.38851142e+00 -6.09898031e-01 -1.10412335e+00
1.00342825e-01 6.29318297e-01 -7.79657289e-02 -5.58602393e-01
3.43774796e-01 4.85089600e-01 -9.24478769e-01 4.01215971e-01
1.05297828e+00 8.17344844e-01 9.13051143e-02 7.94078529e-01
2.58172899e-01 6.46909118e-01 -3.13541710e-01 -2.88818240e-01
-1.89849343e-02 -6.91440940e-01 -7.86229968e-01 -6.28542244e-01
-1.10275885e-02 -3.49922866e-01 -5.79631031e-01 1.22215927e+00
1.07632256e+00 -6.86210692e-02 6.90260530e-01 -7.20321476e-01
-6.07532501e-01 5.33323705e-01 5.72660565e-02 9.86638367e-02
1.00580335e-01 3.66772532e-01 1.06964481e+00 7.77537003e-02
3.77225518e-01 8.21437180e-01 5.64548254e-01 8.81474495e-01
-1.28463042e+00 -5.10449648e-01 -1.10962141e+00 5.42893052e-01
-1.33206761e+00 -9.58644375e-02 1.84420288e-01 -2.15807602e-01
1.32944119e+00 1.06868230e-01 5.50389647e-01 8.81655097e-01
9.55677927e-01 1.26121640e-01 7.81906426e-01 -8.66360068e-02
5.28290093e-01 -1.57019287e-01 -2.94040889e-02 9.77247059e-01
3.34524900e-01 3.87636244e-01 -4.12516117e-01 -2.94094324e-01
4.21790779e-01 4.74790365e-01 -2.56302029e-01 1.87772378e-01
-8.75651240e-01 1.01799929e+00 9.36837614e-01 3.28733683e-01
-3.86768609e-01 6.53507948e-01 3.17008227e-01 1.29955947e-01
-1.43113166e-01 7.04521298e-01 -5.08754790e-01 1.18751684e-02
-4.00154203e-01 1.50385007e-01 5.54119825e-01 4.40033406e-01
8.06386530e-01 -4.14539635e-01 -2.17387512e-01 2.99003482e-01
2.56344993e-02 6.28201008e-01 4.99932945e-01 -4.99282897e-01
-1.12496436e-01 7.37842321e-01 -1.09157585e-01 -7.28106260e-01
-1.05392456e+00 -4.55148697e-01 -1.19021022e+00 -2.60144353e-01
1.28067911e-01 -1.12688832e-01 -7.58729815e-01 1.64553618e+00
3.09583127e-01 1.79686010e-01 2.74615198e-01 7.19596386e-01
9.78359818e-01 6.24927878e-01 1.84249073e-01 -1.21231891e-01
1.31680667e+00 -3.94171894e-01 -6.68064117e-01 3.94578815e-01
1.53175962e+00 -4.92099494e-01 4.64476436e-01 3.00649613e-01
-5.95067024e-01 -1.92267857e-02 -1.17816365e+00 -2.29665413e-01
-6.72377706e-01 -3.29082638e-01 1.17656410e+00 8.71597171e-01
-8.00186753e-01 1.46602631e+00 -7.84847677e-01 -2.86902845e-01
6.50107741e-01 1.11094487e+00 -3.85808885e-01 -1.13091558e-01
-1.68106198e+00 9.40690041e-01 6.81993484e-01 -5.08742929e-02
-8.12038779e-01 -6.19397640e-01 -5.44543028e-01 1.56402513e-01
-7.69875273e-02 -9.55960751e-01 1.13794851e+00 -1.43894643e-01
-1.68848610e+00 3.20483565e-01 2.96519578e-01 -5.45284927e-01
-1.57199368e-01 3.69265139e-01 -1.55766904e-01 1.40585529e-03
-2.66644955e-01 4.71602321e-01 1.57242984e-01 1.44516751e-01
-4.56783116e-01 -6.73451126e-01 -3.07562146e-02 -2.79776733e-02
-4.43409145e-01 -4.60631847e-01 1.21993355e-01 1.92300633e-01
-3.13888431e-01 -1.09944248e+00 -5.30572057e-01 -5.63086748e-01
-3.88311058e-01 -3.58534932e-01 3.23432952e-01 -1.62144542e-01
1.14872754e+00 -1.76080966e+00 2.44292766e-01 5.54731116e-02
3.61131459e-01 2.97034979e-01 -1.81481361e-01 7.14498758e-01
1.52967954e-02 1.86054796e-01 -1.94169223e-01 5.43811500e-01
-1.92007199e-01 8.91713053e-02 2.11942136e-01 7.33908892e-01
7.39207342e-02 1.13200438e+00 -1.19675076e+00 -2.24724680e-01
1.22998998e-01 5.72795689e-01 -8.12095344e-01 -4.90879864e-02
-5.31824350e-01 6.37031674e-01 -7.67842829e-01 5.66735387e-01
6.01247251e-01 -7.90121496e-01 3.69979411e-01 -2.69162267e-01
1.52371377e-01 3.93474668e-01 -5.73161542e-01 1.72023928e+00
-2.17978463e-01 1.86138570e-01 -8.39153290e-01 -9.17941809e-01
4.08148497e-01 3.87780756e-01 7.46851027e-01 -1.03688633e+00
6.40008926e-01 4.68690008e-01 5.60784280e-01 -4.03511673e-01
3.94942276e-02 -4.89942074e-01 -8.66024196e-02 2.94104189e-01
1.01619504e-01 -2.94372439e-01 2.64129043e-01 7.56352097e-02
1.65153742e+00 -3.99433345e-01 2.95905501e-01 2.26864610e-02
4.08485830e-01 7.64632504e-03 5.98842978e-01 6.11706436e-01
-4.39817607e-01 2.45179251e-01 6.16484940e-01 -7.51813650e-01
-9.86965120e-01 -6.07774675e-01 -6.11550868e-01 6.55529559e-01
1.24612220e-01 -4.30164307e-01 -7.85923481e-01 -2.85216808e-01
-1.04351260e-01 4.18823421e-01 -6.53311253e-01 -5.38996756e-01
-1.56987697e-01 -1.54107928e+00 7.05381632e-01 1.58093974e-01
3.96617144e-01 -7.00538337e-01 -4.20859188e-01 4.58976597e-01
-5.77179641e-02 -9.10893619e-01 -1.64301142e-01 7.16829717e-01
-8.92785132e-01 -1.35188425e+00 -3.62951010e-01 -5.11258066e-01
3.67396176e-01 -2.83342842e-02 7.20669270e-01 2.14847907e-01
-4.99183357e-01 -4.69188005e-01 -2.70583779e-01 -3.69190723e-01
-7.30679095e-01 2.83272952e-01 2.14741021e-01 -3.12709123e-01
5.92202127e-01 -3.01753283e-01 -9.51374829e-01 -1.64518505e-01
-9.36369538e-01 -2.15367042e-02 8.15538466e-01 1.30673194e+00
8.35666418e-01 2.72654206e-01 5.68136215e-01 -1.10246849e+00
5.32007039e-01 -3.59601170e-01 -7.60710239e-01 2.50232488e-01
-5.57957113e-01 4.70527798e-01 1.13506711e+00 1.15507133e-01
-4.21477973e-01 3.92708749e-01 -3.56292337e-01 2.70780355e-01
4.19632085e-02 6.40435398e-01 1.22262947e-01 -5.42811394e-01
1.09066832e+00 -7.02038929e-02 -2.17499912e-01 9.48363692e-02
2.36218765e-01 8.11647058e-01 -1.12278871e-01 1.40825687e-02
3.60245593e-02 2.26123542e-01 9.34893370e-01 -1.01086044e+00
-7.69129634e-01 -1.94994256e-01 -3.74181181e-01 3.49285662e-01
1.22019196e+00 -6.16738379e-01 -1.43003988e+00 4.84490067e-01
-1.11215031e+00 -3.04595500e-01 2.49083206e-01 7.08989978e-01
-5.60671508e-01 3.24719399e-01 -9.99973357e-01 -1.95561215e-01
-8.33574116e-01 -1.49269509e+00 8.06553304e-01 4.80922401e-01
5.21202683e-02 -9.63667452e-01 4.15116221e-01 2.85406560e-01
3.26702535e-01 3.07306617e-01 1.19518328e+00 -6.31493866e-01
-7.03337729e-01 -4.53473568e-01 -1.34061068e-01 7.13285282e-02
3.73137653e-01 -1.49780333e-01 -8.83316994e-01 -3.81941170e-01
-4.04549420e-01 -6.22246623e-01 8.09209645e-01 6.61408961e-01
1.19373500e+00 -1.12459511e-02 -3.05639297e-01 8.97358954e-01
1.58085251e+00 3.78660768e-01 6.98842227e-01 -7.05596358e-02
6.64237499e-01 -1.97710708e-01 6.82935491e-02 5.26472986e-01
1.78400964e-01 5.67915618e-01 7.69354701e-01 7.38251656e-02
2.56510973e-01 2.71961652e-03 -6.56929053e-03 5.43093145e-01
-1.00634992e-01 -4.02947724e-01 -8.94580364e-01 -8.29079840e-03
-1.48941731e+00 -9.19928372e-01 -4.28608119e-01 2.47198987e+00
1.05877984e+00 -1.00247860e-01 -1.21126898e-01 -4.14941907e-02
3.76164824e-01 -4.61783797e-01 -7.82281280e-01 -6.29963458e-01
-8.17320794e-02 8.18527281e-01 9.91485715e-01 2.91828752e-01
-1.20550764e+00 9.65679824e-01 5.70711994e+00 8.18619370e-01
-1.29861581e+00 -1.05306707e-01 7.72407472e-01 1.97264373e-01
-8.19382593e-02 -9.04648453e-02 -6.72018170e-01 2.99116760e-01
1.60721672e+00 -2.03008667e-01 6.75128460e-01 2.24665835e-01
2.33664438e-01 -1.32599279e-01 -1.42284560e+00 1.04087543e+00
-4.38856035e-01 -1.71664107e+00 3.71267423e-02 2.25600362e-01
8.30201864e-01 4.76990074e-01 1.80270985e-01 2.31484361e-02
2.51021594e-01 -1.45628393e+00 -3.71682972e-01 4.55904365e-01
7.78104722e-01 -1.04626024e+00 1.06563270e+00 3.74111325e-01
-5.80787539e-01 -1.84748054e-01 -8.01566958e-01 -6.90097213e-02
-3.33235860e-01 4.47767615e-01 -1.11816537e+00 5.23833334e-01
3.67340922e-01 7.38855004e-01 -4.22193706e-01 1.03110921e+00
-1.08657561e-01 2.49157220e-01 -3.12999994e-01 -7.03672945e-01
2.24635959e-01 -3.18955898e-01 -2.64424503e-01 7.35475183e-01
3.78034115e-01 4.24654961e-01 -1.87066168e-01 7.43019462e-01
-3.50804448e-01 2.41325572e-01 -7.20007718e-01 -7.37855136e-01
1.69012278e-01 1.14250541e+00 -5.95609248e-01 -1.02478549e-01
-2.15594366e-01 9.72678781e-01 3.10049981e-01 -1.85658231e-01
-8.09818685e-01 -6.37100399e-01 3.51593167e-01 -3.55480224e-01
-1.81381837e-01 2.44440198e-01 -2.53284603e-01 -8.85904312e-01
-9.13019001e-01 -7.77747154e-01 4.72940236e-01 -2.57161409e-01
-9.57582831e-01 3.10374290e-01 -8.64068925e-01 -1.15777600e+00
-2.04785421e-01 -1.03682637e+00 -2.17513233e-01 9.76360440e-01
-1.37977695e+00 -6.46559656e-01 4.43457104e-02 4.48774785e-01
-2.46772856e-01 2.08945602e-01 1.49573195e+00 5.43135285e-01
-8.09301436e-01 6.15358651e-01 6.19504154e-01 -7.29252473e-02
6.62981212e-01 -1.25162637e+00 4.92609404e-02 1.66742489e-01
-2.01243758e-01 5.04703701e-01 6.22006238e-01 -5.34375787e-01
-1.98949039e+00 -9.99165058e-01 9.04658973e-01 -2.79700369e-01
6.56045794e-01 6.40405938e-02 -6.09654963e-01 1.31084532e-01
-2.76185083e-03 -3.79763432e-02 1.28998065e+00 -6.32606894e-02
8.10781419e-02 -1.78406656e-01 -1.15054035e+00 4.37387615e-01
3.25816482e-01 -6.15758359e-01 1.25316530e-01 1.11429000e+00
6.05354011e-01 -4.63445336e-01 -1.04978311e+00 3.34229797e-01
4.65631723e-01 -8.68371844e-01 6.78335845e-01 -8.89653206e-01
5.22290766e-01 -2.12641075e-01 -9.03169960e-02 -1.45405746e+00
-5.02092183e-01 -5.23062587e-01 1.25517026e-01 9.51357372e-03
6.65789366e-01 -5.75556636e-01 9.57868755e-01 6.00011826e-01
-2.92443577e-02 -9.04888868e-01 -1.04972708e+00 -4.07415658e-01
3.72046620e-01 -1.85045749e-01 5.92659116e-01 7.85181761e-01
3.75899941e-01 7.07353115e-01 -2.83149630e-01 6.43904209e-02
5.62511086e-01 -2.70652268e-02 4.34669197e-01 -1.02399600e+00
-4.23282713e-01 -4.90247458e-01 -8.48681092e-01 -4.95926410e-01
-2.97102541e-01 -1.24304450e+00 -3.13238263e-01 -1.51230049e+00
4.85865027e-01 -2.23539814e-01 -7.93145895e-01 5.11351347e-01
-5.82277887e-02 9.72955823e-02 -3.82343739e-01 -2.43300393e-01
-4.39763188e-01 5.82521081e-01 1.45466340e+00 -5.70937335e-01
-2.13451236e-01 -4.04896215e-02 -5.03043234e-01 1.88975111e-01
8.82351577e-01 -6.15568161e-01 -2.43182242e-01 -2.06278160e-01
7.65444994e-01 3.97284657e-01 1.51311897e-03 -1.04656005e+00
3.26986253e-01 -3.17355394e-01 3.72034758e-01 -4.01663750e-01
2.01201424e-01 -5.70253968e-01 1.95454821e-01 1.26657891e+00
-1.42173976e-01 -2.38497928e-01 8.14298913e-02 6.30089045e-01
1.15033485e-01 -1.39250219e-01 1.00077760e+00 -7.66438991e-02
-2.74673820e-01 8.32618654e-01 -1.67295575e-01 -3.44782889e-01
9.89060760e-01 2.17088848e-01 -5.11025369e-01 -8.01682994e-02
-4.23745841e-01 1.06637433e-01 3.72935951e-01 -1.15837693e-01
5.61620355e-01 -1.13298714e+00 -5.63300550e-01 -3.00754812e-02
2.50049740e-01 -2.70742804e-01 7.32033491e-01 8.45908403e-01
-1.13068628e+00 9.76899505e-01 -2.32747942e-01 -4.62870270e-01
-8.37797940e-01 7.50323653e-01 7.57131755e-01 -3.71153623e-01
-1.82618663e-01 9.06397462e-01 -1.29563082e-02 -3.46342385e-01
-1.75845847e-01 -2.77920872e-01 -2.44961858e-01 -3.77014041e-01
5.44497192e-01 2.83580154e-01 4.72400457e-01 -5.17970860e-01
-4.43741709e-01 3.89418900e-01 -1.88027009e-01 5.10238469e-01
1.39487064e+00 5.94194114e-01 -3.98708940e-01 -2.54751500e-02
1.35915041e+00 -5.67149699e-01 -4.99765247e-01 3.39217708e-02
-2.20492914e-01 4.81869429e-02 4.31202054e-01 -8.88549268e-01
-7.82532454e-01 1.21399724e+00 1.02775109e+00 6.77788407e-02
1.20578349e+00 -2.65248001e-01 1.15401042e+00 1.02219057e+00
5.83742023e-01 -1.06818986e+00 -2.55270332e-01 6.03221834e-01
1.43959880e-01 -1.43102717e+00 1.15097754e-01 -5.74961789e-02
-2.93307770e-02 1.37271023e+00 1.74267858e-01 5.85004315e-02
5.36795139e-01 -2.52258480e-01 -2.01433599e-01 -3.88579249e-01
-7.77977884e-01 -1.20982252e-01 7.00972006e-02 9.83750075e-02
8.53359580e-01 5.85777938e-01 -6.50988102e-01 3.77877742e-01
-3.90591361e-02 2.86536306e-01 6.03885114e-01 8.47604275e-01
-4.47462946e-01 -1.65873623e+00 -1.52751552e-02 6.70735478e-01
-5.51132917e-01 -3.41605157e-01 -3.29591721e-01 1.73502624e-01
1.69296354e-01 8.99006903e-01 -4.80220228e-01 -6.82658672e-01
6.18263753e-03 -1.01858668e-01 8.49447608e-01 -5.70451975e-01
-4.58474964e-01 -2.63768762e-01 -2.88860649e-01 -5.62290609e-01
-2.22840279e-01 -4.72891741e-02 -1.81107354e+00 -7.22113729e-01
-5.51998854e-01 4.82002124e-02 7.84223616e-01 9.34112906e-01
6.51927948e-01 7.89489508e-01 6.94414735e-01 -2.23881453e-01
-8.41213584e-01 -7.36840010e-01 -8.21226239e-01 2.07231149e-01
3.28557640e-01 -4.85706121e-01 8.61247033e-02 -3.48011971e-01] | [5.342650413513184, 5.396031379699707] |
a452f749-191c-451c-931c-2b30cab23f82 | multi-temporal-sentinel-1-and-2-data-fusion | 1807.09954 | null | http://arxiv.org/abs/1807.09954v1 | http://arxiv.org/pdf/1807.09954v1.pdf | Multi-temporal Sentinel-1 and -2 Data Fusion for Optical Image Simulation | In this paper, we present the optical image simulation from a synthetic
aperture radar (SAR) data using deep learning based methods. Two models, i.e.,
optical image simulation directly from the SAR data and from multi-temporal
SARoptical data, are proposed to testify the possibilities. The deep learning
based methods that we chose to achieve the models are a convolutional neural
network (CNN) with a residual architecture and a conditional generative
adversarial network (cGAN). We validate our models using the Sentinel-1 and -2
datasets. The experiments demonstrate that the model with multi-temporal
SAR-optical data can successfully simulate the optical image, meanwhile, the
model with simple SAR data as input failed. The optical image simulation
results indicate the possibility of SARoptical information blending for the
subsequent applications such as large-scale cloud removal, and optical data
temporal superresolution. We also investigate the sensitivity of the proposed
models against the training samples, and reveal possible future directions. | ['Naoto Yokoya', 'Wei He'] | 2018-07-26 | null | null | null | null | ['cloud-removal'] | ['computer-vision'] | [ 3.10029328e-01 -3.73084724e-01 5.26840448e-01 -3.19559753e-01
-6.78829193e-01 -5.77193081e-01 6.75204694e-01 -8.63803148e-01
-3.01567852e-01 9.67033386e-01 -7.49529451e-02 -4.33702976e-01
-1.78456366e-01 -1.08338141e+00 -6.03881001e-01 -1.10007870e+00
-1.63814351e-01 3.14038754e-01 1.70911476e-01 -3.34426910e-01
-1.66699782e-01 1.16984522e+00 -1.41444969e+00 4.20048773e-01
1.09336817e+00 1.18173480e+00 2.71399647e-01 5.21329105e-01
3.47440183e-01 9.60283101e-01 -6.51247978e-01 2.20952611e-02
8.39136839e-01 -1.03830136e-01 -2.31832504e-01 3.64735305e-01
5.70022881e-01 -7.20029175e-01 -6.74957871e-01 1.16420412e+00
4.91882622e-01 -4.77443375e-02 2.78238297e-01 -8.77881527e-01
-7.84295380e-01 6.86683431e-02 -4.16257113e-01 1.90420732e-01
-2.12118104e-01 6.57569110e-01 1.47563130e-01 -5.68117023e-01
6.15643978e-01 1.10350227e+00 7.83168733e-01 2.95733213e-01
-1.07004666e+00 -7.60942578e-01 -2.61736810e-01 -1.15887932e-01
-1.50964808e+00 -2.92572409e-01 7.14544654e-01 -7.79047906e-01
4.66002226e-01 3.14104348e-01 7.72284150e-01 9.29380894e-01
5.17686903e-01 3.57650906e-01 1.73909235e+00 -1.11398406e-01
-1.02662817e-02 2.33562477e-02 -2.51368675e-02 4.61449146e-01
3.37460220e-01 1.00132751e+00 6.07351363e-02 -1.19876161e-01
1.01242065e+00 1.22568756e-01 -5.19423425e-01 2.33049989e-01
-9.29251432e-01 9.35601532e-01 8.56563628e-01 8.90096053e-02
-5.69925129e-01 -1.33098722e-01 -1.80344686e-01 3.33559036e-01
6.90702915e-01 6.47305608e-01 -2.05492705e-01 1.00356889e+00
-1.40211570e+00 4.79115695e-01 1.07726157e-01 7.32975423e-01
6.94412231e-01 9.24039543e-01 -2.30257303e-01 2.66122848e-01
3.25253069e-01 1.07549798e+00 2.92566240e-01 -6.15312517e-01
8.22769105e-02 7.72455260e-02 6.44226134e-01 -7.97187865e-01
-3.10168564e-01 -6.55387402e-01 -1.07676589e+00 5.75618148e-01
-7.80000314e-02 -4.52606350e-01 -1.42899108e+00 1.21216226e+00
-1.59333311e-02 4.23211664e-01 7.08611608e-01 1.35882723e+00
1.03203344e+00 9.04526353e-01 -2.68443406e-01 -4.48946923e-01
9.41055000e-01 -7.28595793e-01 -8.70143354e-01 -5.08435249e-01
2.52512307e-03 -8.70205700e-01 6.51143789e-01 1.93631768e-01
-8.32587421e-01 -8.45245838e-01 -1.30669045e+00 4.38132226e-01
-2.95681059e-01 3.22848290e-01 7.45718420e-01 1.95371866e-01
-1.01445174e+00 4.18665022e-01 -8.31830561e-01 -1.18674181e-01
4.10092562e-01 -4.64820042e-02 -4.92345467e-02 -1.22102641e-01
-1.37568355e+00 5.43825269e-01 4.36844379e-01 6.64444447e-01
-1.34819388e+00 -6.61707640e-01 -5.05292475e-01 -1.50154576e-01
-1.68307170e-01 -6.83426380e-01 7.09503651e-01 -1.14354444e+00
-1.18601978e+00 7.71582603e-01 4.08226550e-01 -8.31440032e-01
4.74001139e-01 -2.73935586e-01 -9.96596456e-01 3.14412028e-01
-1.45427078e-01 4.86654699e-01 1.03830349e+00 -1.69386506e+00
-2.20330834e-01 -3.47343743e-01 -7.54795298e-02 -1.88338220e-01
5.12515604e-01 1.10150598e-01 3.44530076e-01 -7.13552058e-01
2.07007229e-01 -9.47762728e-01 -3.42532247e-01 -1.48329154e-01
-5.25170803e-01 8.22629631e-01 9.51432288e-01 -6.85266852e-01
6.05183125e-01 -2.15198588e+00 -1.96911663e-01 -1.20650701e-01
-9.95025858e-02 5.56038320e-01 -3.88867021e-01 4.04367030e-01
-1.84263036e-01 1.02028124e-01 -6.12474799e-01 2.05184728e-01
-6.95981801e-01 -8.31874162e-02 -9.73886967e-01 4.34798717e-01
4.52393115e-01 1.02443480e+00 -3.97275180e-01 3.37828882e-02
-5.75492233e-02 4.15492237e-01 -1.20182969e-01 7.08752453e-01
-5.14225245e-01 8.74600530e-01 -5.91537893e-01 9.16997969e-01
1.45390129e+00 -6.61593527e-02 -5.64492121e-03 -7.50965714e-01
-3.92680645e-01 -4.36355203e-01 -7.38388240e-01 9.11463737e-01
-4.93728481e-02 5.58839917e-01 1.71172708e-01 -6.45421803e-01
1.13358569e+00 -9.13889781e-02 2.03064144e-01 -9.54575121e-01
-3.08465082e-02 5.64543232e-02 2.17338294e-01 -7.11746037e-01
6.19134903e-01 -7.37217247e-01 2.94488221e-01 2.67336041e-01
-3.35851073e-01 -6.19289041e-01 -4.82517034e-01 -9.84043702e-02
5.36635220e-01 7.09758177e-02 -2.15281337e-01 -2.30558798e-01
3.33732486e-01 6.87205732e-01 6.13320470e-01 6.66655064e-01
5.69932126e-02 9.43435252e-01 -2.08058298e-01 -9.40846503e-01
-1.09589708e+00 -1.08128035e+00 -3.00872415e-01 -1.57679617e-01
2.18065262e-01 4.33602124e-01 -2.96509415e-01 -4.23189074e-01
-1.73173156e-02 6.37897372e-01 -3.01131457e-01 -3.04857977e-02
-3.00356477e-01 -1.39845681e+00 6.44352019e-01 2.11357951e-01
1.03594673e+00 -1.07829273e+00 -5.39564610e-01 5.90550639e-02
-4.23848396e-03 -1.61505532e+00 4.15333807e-01 -2.85186768e-01
-9.80071843e-01 -1.07013130e+00 -2.98990220e-01 -3.32546115e-01
4.34353024e-01 5.82303166e-01 8.87330294e-01 -2.18856800e-02
-4.33415383e-01 -2.28068344e-02 -3.49911749e-01 -2.88453668e-01
-5.03537595e-01 -6.81022704e-01 2.36811399e-01 4.08178329e-01
-5.81826530e-02 -6.83310449e-01 -6.72920704e-01 7.36498907e-02
-1.44646132e+00 2.25602403e-01 8.74082625e-01 8.66216362e-01
6.66616023e-01 2.56924212e-01 1.00538582e-01 -7.91608632e-01
2.75314569e-01 -2.92705387e-01 -1.09777153e+00 1.58344835e-01
-2.46036351e-01 -1.14223361e-01 6.02905869e-01 -2.40556598e-01
-1.23556447e+00 -2.94420030e-02 -3.55158821e-02 -8.16343009e-01
-1.88099995e-01 9.34428871e-01 1.26297483e-02 -3.82944614e-01
8.03491235e-01 7.33429253e-01 -2.35729907e-02 -4.29411203e-01
-4.32714038e-02 6.99256718e-01 6.20380104e-01 -5.31827956e-02
1.42216849e+00 1.04099059e+00 -7.69598261e-02 -1.07575822e+00
-1.03924203e+00 1.14611685e-01 -5.21505415e-01 -2.35837936e-01
1.02322173e+00 -1.48240268e+00 -1.04762621e-01 8.65250170e-01
-1.25128794e+00 -5.91380775e-01 8.38767271e-03 8.55285883e-01
-3.32592398e-01 1.94054782e-01 -5.16536415e-01 -8.99675310e-01
-4.92211163e-01 -1.10254657e+00 1.12455833e+00 2.18639970e-01
7.44917333e-01 -6.29006028e-01 3.10406629e-02 5.08241236e-01
5.80235958e-01 7.48724163e-01 8.09169590e-01 -2.46847287e-01
-1.39806616e+00 -1.09315842e-01 -5.63266039e-01 5.47863841e-01
1.77511631e-03 1.49746254e-01 -1.11914158e+00 -8.00486565e-01
4.34644669e-01 -4.67402160e-01 1.07472074e+00 5.48521757e-01
1.12003434e+00 -4.49862570e-01 -4.58031595e-02 1.21906543e+00
2.01759672e+00 4.24416363e-01 1.35885525e+00 1.89716697e-01
5.49223006e-01 3.28100115e-01 7.80842602e-01 2.75712490e-01
-3.02441359e-01 4.42197204e-01 8.13165188e-01 -5.03466606e-01
-9.54820663e-02 6.13581948e-02 3.92756939e-01 4.77755606e-01
-2.74415433e-01 -4.21325982e-01 -7.81554997e-01 1.97773725e-01
-1.50286138e+00 -1.14147723e+00 -3.72726828e-01 2.12374282e+00
1.96978331e-01 -1.87023461e-01 -5.73187292e-01 -4.19328779e-01
4.12972599e-01 7.12594032e-01 -5.38276076e-01 2.05871031e-01
-6.72817647e-01 4.67834860e-01 7.73773432e-01 6.57972932e-01
-1.15405214e+00 1.31311738e+00 6.62272549e+00 4.55599725e-01
-1.61161458e+00 1.38931721e-01 4.52530801e-01 5.89550808e-02
-4.06894654e-01 1.31673500e-01 -7.29091585e-01 4.39115107e-01
6.99441731e-01 2.31038973e-01 4.11683768e-01 1.91349223e-01
4.47735935e-01 -4.97152060e-02 -3.49662423e-01 6.09887958e-01
-2.91613326e-03 -1.58479309e+00 6.55413270e-01 4.76750173e-02
9.17061031e-01 4.40579474e-01 1.13849239e-02 1.54260322e-01
5.00918090e-01 -1.14429486e+00 3.67539585e-01 1.15624654e+00
1.06870937e+00 -5.83321273e-01 8.79418850e-01 3.96130472e-01
-6.64639056e-01 -8.33370611e-02 -5.26142657e-01 1.19601879e-02
6.27639368e-02 8.48668456e-01 -5.46359301e-01 1.14550900e+00
6.48601592e-01 5.50491810e-01 -5.69202423e-01 8.61547112e-01
-1.21769547e-01 5.47902763e-01 -3.49590778e-02 6.01661682e-01
3.84413511e-01 -6.11915112e-01 7.27337718e-01 8.10136616e-01
7.08961308e-01 5.16348243e-01 3.29394937e-01 1.26055062e+00
4.10222352e-01 -5.03766358e-01 -8.37065101e-01 -3.10663879e-01
2.09431589e-01 1.35260367e+00 -1.33654162e-01 -1.35655940e-01
-1.76733643e-01 7.20743299e-01 -1.02977112e-01 8.01594198e-01
-9.11758363e-01 6.62476048e-02 7.92643785e-01 3.50415885e-01
2.90061802e-01 -5.62967360e-01 7.47750103e-02 -1.31864810e+00
-2.34122157e-01 -9.20689225e-01 -6.17751814e-02 -1.75510252e+00
-1.33058083e+00 1.23489618e+00 -2.18278542e-02 -1.75419152e+00
1.50092706e-01 -7.42129743e-01 -7.58136451e-01 1.14152801e+00
-2.17476511e+00 -1.55951142e+00 -8.69223595e-01 4.87178087e-01
7.62966871e-02 -6.59521580e-01 5.98229408e-01 -3.25184092e-02
-3.17376971e-01 -7.73325190e-02 1.79590106e-01 3.05431783e-02
4.41721857e-01 -4.35952097e-01 3.07806432e-01 1.28427982e+00
-3.30740780e-01 1.09238952e-01 7.94864118e-01 -8.61409247e-01
-1.38622928e+00 -1.66940212e+00 2.68595099e-01 6.03472665e-02
6.64797723e-01 -3.17896418e-02 -9.01995897e-01 8.23647141e-01
3.75947297e-01 3.74793410e-01 3.19057375e-01 -7.99496412e-01
-9.32538584e-02 -4.22394097e-01 -1.30899203e+00 4.17643636e-01
6.40728474e-01 -4.32847470e-01 -5.21098733e-01 6.74083769e-01
1.03886271e+00 -2.88138926e-01 -6.18739903e-01 1.23092055e+00
2.21454695e-01 -1.25646472e+00 1.03321671e+00 -7.90232778e-01
7.53568947e-01 -7.27923632e-01 -6.18633747e-01 -1.53787184e+00
-4.73988891e-01 -3.33549023e-01 2.94104517e-01 6.61230505e-01
2.11789250e-01 -6.27418101e-01 4.16939020e-01 -1.14792846e-01
-3.07488352e-01 -1.68307245e-01 -5.95466435e-01 -1.05607951e+00
1.58102922e-02 -7.69264027e-02 7.97132075e-01 1.24165106e+00
-1.28865254e+00 1.50108263e-01 -5.75323701e-01 1.16917944e+00
7.70194590e-01 6.48640096e-01 7.53834009e-01 -1.28035307e+00
-2.34613985e-01 1.79758757e-01 -2.71838397e-01 -7.61533141e-01
5.88373132e-02 -6.69327736e-01 -9.37679633e-02 -1.19505298e+00
5.73955365e-02 -4.16954279e-01 -6.84486255e-02 1.78407669e-01
1.53483123e-01 4.64459449e-01 1.13705754e-01 5.04186630e-01
3.61506790e-01 7.86306798e-01 1.62021494e+00 -4.70835000e-01
1.07029758e-01 2.03957111e-02 -2.27080911e-01 4.15122896e-01
8.32756519e-01 -3.23379487e-01 -8.47600922e-02 -7.16313183e-01
6.36796933e-03 4.24460620e-01 9.85490918e-01 -1.32514751e+00
1.04651935e-01 -4.50684994e-01 5.49058437e-01 -8.58556330e-01
4.52443212e-01 -8.46064925e-01 5.49966156e-01 4.34664398e-01
2.37752393e-01 -9.86147001e-02 5.01356721e-01 3.28646362e-01
-5.51123857e-01 7.15220124e-02 1.15760970e+00 -3.55428100e-01
-5.74706495e-01 6.38138115e-01 -3.04874212e-01 -3.43815893e-01
7.92559564e-01 -4.34236713e-02 -6.56371534e-01 -3.33294928e-01
-9.15690422e-01 6.50930330e-02 3.33434433e-01 -6.26820400e-02
7.59015739e-01 -1.29805696e+00 -1.19652700e+00 7.49106109e-01
5.43943048e-02 -6.77031651e-02 5.21552742e-01 5.62407732e-01
-8.94208312e-01 3.27147663e-01 -4.83033925e-01 -6.41755641e-01
-9.80992973e-01 5.91953099e-01 8.78368437e-01 -2.91623235e-01
-4.60306048e-01 4.29347068e-01 3.44973803e-01 -6.28801286e-01
-5.48854887e-01 -7.08505288e-02 8.88952985e-02 -1.89860091e-01
4.53838140e-01 -2.21667305e-01 3.18256542e-02 -4.85634774e-01
-5.49626462e-02 4.38858122e-01 1.27513662e-01 -2.20729008e-01
1.64102733e+00 1.63999110e-01 -2.03716010e-01 2.79965177e-02
6.11481547e-01 1.23328269e-01 -1.39380181e+00 -3.53839576e-01
-5.63034356e-01 -6.00590110e-01 5.23057461e-01 -1.09856653e+00
-1.53555441e+00 9.13817406e-01 8.82701159e-01 2.53006488e-01
1.36425686e+00 -3.57168943e-01 6.67663872e-01 4.97558981e-01
3.01675290e-01 -4.45760489e-01 -1.53798148e-01 5.52446842e-01
1.34295344e+00 -1.32575345e+00 2.85143763e-01 -2.89860189e-01
-6.63994908e-01 1.15325153e+00 5.87904453e-01 -3.87272120e-01
6.48663521e-01 4.47333097e-01 4.12466794e-01 -4.94748294e-01
-5.24012625e-01 -4.36433822e-01 8.45855325e-02 5.26021063e-01
-1.04726762e-01 1.98404104e-01 1.73021391e-01 4.34874117e-01
-1.32311583e-01 1.73677564e-01 8.55668843e-01 5.76609492e-01
-3.41031611e-01 -6.39666438e-01 -6.05243564e-01 2.79770285e-01
-6.99314615e-03 -4.55284029e-01 -4.31895435e-01 1.05524600e+00
4.26965579e-02 6.77232862e-01 3.93679529e-01 -5.58147013e-01
1.09308645e-01 -1.60312355e-01 3.83150846e-01 -4.68143851e-01
-3.93138736e-01 7.40182549e-02 -5.32443002e-02 -3.20266008e-01
-7.14190483e-01 -3.18063438e-01 -6.34419560e-01 -2.42227539e-01
-1.73134699e-01 2.17949655e-02 3.54196697e-01 8.25372100e-01
4.70808446e-01 4.55756247e-01 1.13137448e+00 -9.15475667e-01
-3.67286772e-01 -1.09634411e+00 -9.06154573e-01 2.62232404e-02
6.53619885e-01 -3.95819694e-01 -5.63212514e-01 -5.15656173e-02] | [10.053825378417969, -2.0603463649749756] |
50bd7629-0f7f-4232-a59c-e082d68a777c | the-economics-of-recommender-systems-evidence | 2211.14219 | null | https://arxiv.org/abs/2211.14219v1 | https://arxiv.org/pdf/2211.14219v1.pdf | The Economics of Recommender Systems: Evidence from a Field Experiment on MovieLens | We conduct a field experiment on a movie-recommendation platform to identify if and how recommendations affect consumption. We use within-consumer randomization at the good level and elicit beliefs about unconsumed goods to disentangle exposure from informational effects. We find recommendations increase consumption beyond its role in exposing goods to consumers. We provide support for an informational mechanism: recommendations affect consumers' beliefs, which in turn explain consumption. Recommendations reduce uncertainty about goods consumers are most uncertain about and induce information acquisition. Our results highlight the importance of recommender systems' informational role when considering policies targeting these systems in online marketplaces. | ['Joseph Konstan', 'Ruoyan Kong', 'Daniel Kluver', 'Duarte Goncalves', 'Guy Aridor'] | 2022-11-25 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [-3.83171707e-01 3.09230268e-01 -1.16517448e+00 -3.45964342e-01
-2.17781335e-01 -1.09738290e+00 5.13034940e-01 3.89735818e-01
-5.43624997e-01 9.17361304e-02 1.04176676e+00 -9.92380381e-01
-1.37400748e-02 -1.11665380e+00 -8.97502482e-01 -5.61022460e-02
2.19233617e-01 -4.24389988e-01 -3.19104105e-01 -8.62049088e-02
4.97738421e-01 -1.84275672e-01 -1.18654239e+00 3.76988173e-01
7.09191561e-01 6.58785880e-01 -5.31141199e-02 1.47603124e-01
1.22904748e-01 8.09480309e-01 -1.33196309e-01 -1.09900582e+00
8.58856320e-01 -2.12283745e-01 -2.02215627e-01 4.59468186e-01
2.45515049e-01 -1.05983484e+00 -3.95589620e-01 1.19550252e+00
9.15438309e-02 2.93022215e-01 5.74625790e-01 -9.49491024e-01
-1.84269202e+00 1.50328422e+00 -1.95134252e-01 3.48008722e-01
6.42921150e-01 1.86711788e-01 1.46609271e+00 -2.91278124e-01
6.33685648e-01 1.34447360e+00 2.81647414e-01 7.31565952e-02
-1.64731467e+00 -5.04599810e-01 7.86592007e-01 -1.90010637e-01
-6.78495586e-01 -5.62496960e-01 4.26597297e-01 -8.50483775e-01
5.80778718e-01 6.10691547e-01 7.63114274e-01 9.90234017e-01
2.57765025e-01 8.51451397e-01 1.43146038e+00 -3.72862488e-01
6.13730371e-01 9.46030498e-01 6.47012234e-01 -1.84066728e-01
1.23778164e+00 6.70972347e-01 -2.01429665e-01 -2.93806046e-01
7.63274789e-01 1.70197293e-01 -9.40214694e-02 1.19988009e-01
-3.62985849e-01 1.24416006e+00 2.14779064e-01 -7.98275024e-02
-1.03169632e+00 -8.32346752e-02 -7.20107257e-02 8.51514757e-01
4.90620524e-01 1.04901099e+00 -5.14030397e-01 -2.26975251e-02
3.52366984e-01 2.47419551e-01 1.05136621e+00 8.33939970e-01
3.61718744e-01 -2.46468604e-01 -1.75564140e-01 4.74369675e-01
7.97212064e-01 5.10513663e-01 3.30834955e-01 -9.18525279e-01
2.03441381e-01 2.20371559e-01 7.96932280e-01 -1.18805218e+00
-1.13316730e-01 -4.88708377e-01 1.32875159e-01 -1.71960108e-02
6.36707246e-01 -7.19506741e-01 -2.58933771e-02 1.54764724e+00
1.27008855e-01 -3.82514387e-01 -2.12061673e-01 9.82480466e-01
3.13362718e-01 5.53238750e-01 3.75331849e-01 -3.08505982e-01
1.54156625e+00 -5.25221586e-01 -6.70654953e-01 -1.51713029e-01
6.78880453e-01 -5.23249805e-01 1.13174653e+00 5.75927734e-01
-9.82215405e-01 -1.57053754e-01 -6.12670898e-01 3.07341695e-01
-2.67828107e-01 -3.23037177e-01 1.29306638e+00 1.31546533e+00
-5.92006862e-01 5.33245564e-01 -3.94642413e-01 -1.76675826e-01
1.47304490e-01 3.30665596e-02 6.08710349e-01 1.51773274e-01
-1.25759411e+00 6.51675224e-01 -4.38965380e-01 -5.08909583e-01
-8.32004905e-01 -8.03262770e-01 -6.19941592e-01 2.40275398e-01
3.39628607e-01 -5.89224339e-01 1.43431699e+00 -1.45949972e+00
-1.82477915e+00 -8.95241052e-02 5.37601948e-01 -4.35171336e-01
1.44747302e-01 -2.88764328e-01 -6.95127308e-01 -3.95070195e-01
2.29524061e-01 5.34322672e-02 5.01274705e-01 -1.09632301e+00
-8.96003604e-01 -4.44623262e-01 8.56206656e-01 3.71424764e-01
-3.73566419e-01 1.78964272e-01 3.94789547e-01 -3.57711822e-01
-2.06021786e-01 -9.03848112e-01 -3.30845416e-01 -5.55833042e-01
-3.87607664e-01 -2.09946334e-01 -3.19211155e-01 -5.67812562e-01
9.12880242e-01 -2.01049256e+00 -8.93983722e-01 2.42246196e-01
-2.70183273e-02 -4.23130870e-01 -1.69178128e-01 5.83687127e-01
2.54898518e-01 8.10097456e-01 1.01495409e+00 8.85118321e-02
8.11708927e-01 -3.72620374e-01 -2.36514315e-01 8.34365964e-01
-5.96692920e-01 8.15973639e-01 -5.22403002e-01 4.31018919e-01
3.05215027e-02 2.05643624e-01 -9.76582229e-01 -1.35916650e-01
-4.10140693e-01 1.11176863e-01 -7.43268251e-01 1.94739714e-01
7.52772808e-01 -1.05770141e-01 3.50801259e-01 1.86073065e-01
-6.70624971e-01 1.30877197e+00 -9.41620588e-01 8.74482453e-01
-4.60813016e-01 -2.99940072e-03 1.68889984e-01 -1.43525779e-01
1.76978722e-01 2.43623070e-02 7.26619661e-02 -1.05504465e+00
3.65745157e-01 -4.08860475e-01 4.21266079e-01 -6.14266872e-01
4.57458317e-01 -2.06298083e-02 1.75455902e-02 9.76389885e-01
-6.54128909e-01 2.81748325e-01 -1.59837022e-01 6.46882653e-01
6.87648714e-01 -2.59989738e-01 2.43833065e-01 -9.95662868e-01
-3.84379089e-01 -1.57210588e-01 5.74920774e-01 1.13307858e+00
1.56623513e-01 -4.04405475e-01 4.81918275e-01 -4.07194272e-02
-7.27124929e-01 -6.48246884e-01 -7.17284679e-02 1.59473956e+00
4.32858557e-01 -2.23904252e-01 -3.57886255e-01 -7.61400521e-01
7.47544467e-01 1.85230184e+00 -9.62095141e-01 -1.73227731e-02
6.35658085e-01 -6.03650808e-01 -4.21952814e-01 3.08609128e-01
6.81621730e-02 -3.62916648e-01 -3.39806646e-01 -4.68990766e-02
3.42527628e-02 -3.61817122e-01 -1.08606029e+00 -5.27672410e-01
-6.42658710e-01 -9.46073532e-01 -1.17399819e-01 2.44878754e-02
4.92772311e-01 8.14697683e-01 8.58668923e-01 9.20945853e-02
5.36613166e-01 7.37692654e-01 -5.77528894e-01 -5.44315279e-01
-3.53936940e-01 -4.03300822e-01 2.46792845e-03 1.13922805e-01
7.81141818e-01 -3.99985641e-01 -1.09600973e+00 6.62326515e-01
-6.37399197e-01 -3.39087009e-01 4.04798657e-01 -1.63101986e-01
-1.95276335e-01 2.54293054e-01 8.79079640e-01 -1.38980186e+00
1.17300487e+00 -1.27711082e+00 -1.16791737e+00 -1.01887800e-01
-1.18692625e+00 -2.92107463e-01 4.79537278e-01 -7.64405370e-01
-1.62145293e+00 -3.28274727e-01 1.60566755e-02 8.23309124e-01
-2.13313460e-01 6.00883186e-01 -2.69188613e-01 1.45062789e-01
9.22009230e-01 -5.19573450e-01 -1.74793184e-01 -7.18274653e-01
7.74361968e-01 5.14472723e-01 -3.79195333e-01 -4.70642060e-01
4.94802564e-01 4.86272067e-01 -7.97978878e-01 -4.19697911e-01
-6.13266706e-01 -2.34650746e-01 2.55265713e-01 -7.57556558e-02
6.15980148e-01 -1.33424389e+00 -1.41508365e+00 -1.58629075e-01
-4.05769110e-01 -5.65203369e-01 -2.11273804e-01 1.16612196e+00
-1.47857741e-01 1.12786427e-01 -9.84691679e-01 -1.14465749e+00
1.70775771e-01 -8.78032088e-01 -9.25581455e-02 2.46601611e-01
-2.04056382e-01 -8.63230646e-01 -1.61574811e-01 5.00217974e-01
7.91766226e-01 -5.61422467e-01 5.70130885e-01 -9.13164198e-01
-8.75290871e-01 -1.87975436e-01 1.03980333e-01 -2.79255986e-01
1.85331970e-01 -6.50252476e-02 -6.93151057e-01 -2.95229644e-01
1.15861133e-01 -1.46502241e-01 4.61808264e-01 8.82512152e-01
3.61884952e-01 -1.08438492e+00 -1.97765261e-01 -3.52537371e-02
1.39914012e+00 5.85292578e-01 4.65727299e-01 4.02938336e-01
-1.37817800e-01 6.84981167e-01 5.80518663e-01 1.01471245e+00
6.49451494e-01 5.79943299e-01 8.66540000e-02 2.80537933e-01
2.39150375e-01 -1.03110695e+00 6.61237955e-01 2.51893117e-03
4.52756256e-01 -5.30752838e-01 3.10437411e-01 4.40915167e-01
-1.47175562e+00 -7.55994856e-01 -2.45168135e-01 2.24991226e+00
2.28119716e-01 2.86050200e-01 8.13940883e-01 -3.60495508e-01
6.72771752e-01 -3.22392851e-01 -7.68048108e-01 -5.15513957e-01
2.08619609e-01 -7.55351722e-01 1.01184821e+00 9.45646405e-01
-3.89984071e-01 6.64847195e-01 7.04111052e+00 -9.84195694e-02
-5.42963326e-01 2.86477536e-01 1.13347888e+00 -3.59812416e-02
-1.47698319e+00 9.08403024e-02 -8.12537789e-01 5.61883986e-01
1.10027719e+00 -6.66142464e-01 6.96932197e-01 9.83588040e-01
6.43828750e-01 -3.41836601e-01 -1.26927865e+00 1.40144825e-01
-4.10531759e-01 -1.10137916e+00 -3.53282064e-01 7.14394033e-01
1.03867531e+00 -3.64754051e-01 7.13573873e-01 2.43526742e-01
1.40466738e+00 -4.56932515e-01 1.06961513e+00 1.65512204e-01
-9.02992785e-02 -5.68126023e-01 2.22519159e-01 4.57037359e-01
-3.70670438e-01 -5.48445880e-01 -6.89749718e-01 -9.90574241e-01
2.56313652e-01 6.05796516e-01 -5.80785930e-01 -2.44297758e-01
8.10463503e-02 3.58913541e-01 -3.60406578e-01 9.54009652e-01
-2.41802469e-01 8.33343863e-01 -2.67568231e-02 -3.27750862e-01
-6.54591694e-02 -7.72449613e-01 1.86276976e-02 6.46478713e-01
1.79578722e-01 8.35052669e-01 -1.43789500e-01 1.20484710e+00
-1.90263867e-01 5.25171101e-01 -5.83098471e-01 -3.26250762e-01
7.53480792e-01 8.89951587e-01 -7.00252056e-01 -2.83784240e-01
-8.48420441e-01 5.01053274e-01 -3.59090090e-01 4.61027861e-01
-5.65046430e-01 3.52700680e-01 1.04951513e+00 2.42450073e-01
-5.10687847e-03 2.60064781e-01 -5.65387547e-01 -1.32741690e+00
-6.18908584e-01 -9.49595630e-01 3.19357246e-01 -3.89899254e-01
-1.40262699e+00 -4.98543292e-01 -2.57307559e-01 -3.71264309e-01
2.02570736e-01 -2.62425452e-01 -3.89212668e-01 7.23244548e-01
-1.20358121e+00 -1.38370141e-01 4.67102021e-01 2.51971334e-01
5.06174028e-01 5.25734901e-01 5.53561151e-01 2.09012359e-01
-3.74118418e-01 6.86835289e-01 3.55721235e-01 -3.21000606e-01
4.79138434e-01 -1.06596065e+00 7.39256084e-01 7.65879095e-01
4.59497422e-02 1.20922923e+00 8.69307756e-01 -1.05856526e+00
-2.01490927e+00 -6.98961079e-01 4.34450805e-01 -6.46101415e-01
5.57295799e-01 -3.33718926e-01 -2.09932521e-01 1.07807004e+00
5.03448963e-01 -1.02369928e+00 1.39254355e+00 8.21310520e-01
-5.47278404e-01 3.53570394e-02 -1.56359613e+00 7.94970036e-01
1.12203813e+00 -4.63168949e-01 -1.70903355e-01 5.95795572e-01
1.10658467e+00 4.16939527e-01 -1.04371631e+00 -5.66343307e-01
8.04035008e-01 -1.06083596e+00 7.38947213e-01 -6.23793781e-01
3.96143466e-01 3.53493363e-01 -4.32875067e-01 -1.63950157e+00
-1.05577242e+00 -7.41065919e-01 5.93354821e-01 1.21243203e+00
1.04105055e+00 -1.12458110e+00 7.75877237e-01 1.81856966e+00
1.56849608e-01 3.44990075e-01 1.46964669e-01 -5.57342172e-01
1.53033078e-01 -2.97895312e-01 9.49707210e-01 1.15028667e+00
7.95362234e-01 3.18238139e-01 -1.99421272e-01 4.91553843e-01
7.13170111e-01 2.56751806e-01 4.52968776e-01 -9.00232852e-01
-8.10412407e-01 -2.95470566e-01 5.01023710e-01 -1.51918399e+00
-1.61114827e-01 -4.94032294e-01 -3.26579630e-01 -1.26795280e+00
3.12260956e-01 -4.09381479e-01 -1.42739356e-01 -1.70729995e-01
3.35058540e-01 -2.16663003e-01 4.17588413e-01 2.61510145e-02
-4.61624265e-01 1.54646588e-02 9.80176806e-01 2.00745612e-01
-4.33023691e-01 4.84005690e-01 -2.11259961e+00 4.86968219e-01
7.94438779e-01 -9.29802507e-02 -7.50116408e-01 -3.50648254e-01
9.98100996e-01 3.44185412e-01 -2.55979747e-01 8.03080499e-02
-1.80660188e-01 -7.72821784e-01 2.48661458e-01 -7.39687309e-02
8.16629268e-03 -9.68232989e-01 6.76847160e-01 4.03255045e-01
-1.18686724e+00 9.36955884e-02 -6.27884194e-02 7.49842644e-01
8.08536232e-01 -1.96939573e-01 2.76316732e-01 -1.45908207e-01
-1.02247067e-01 -7.73627730e-03 -1.23308337e+00 -4.67218608e-01
7.45639980e-01 2.92940199e-01 -6.22000754e-01 -1.24354553e+00
-7.39390075e-01 -3.48391905e-02 9.06099558e-01 5.04076183e-01
7.52162114e-02 -1.18479264e+00 -3.45012158e-01 -8.24876726e-02
4.13628332e-02 -1.35395622e+00 1.82948336e-01 2.41272017e-01
2.17214733e-01 6.59922421e-01 1.00352801e-01 5.19037902e-01
-7.06011474e-01 1.18310046e+00 1.61268726e-01 5.09620726e-01
-5.93085438e-02 9.73304033e-01 5.39241016e-01 8.59303102e-02
9.47828293e-02 -4.75551695e-01 -2.25682035e-01 2.27351505e-02
9.58821535e-01 5.64518034e-01 -3.65773290e-01 -1.17357999e-01
9.96657610e-02 -4.34431791e-01 -2.33851597e-01 -4.41914886e-01
1.12610877e+00 -9.55753505e-01 2.16240302e-01 3.27591091e-01
9.43511248e-01 6.32510126e-01 -1.40363777e+00 -1.08038552e-01
-1.58051297e-01 -1.54819357e+00 3.62222165e-01 -1.33707142e+00
-1.10492671e+00 -1.24040902e-01 3.19460481e-01 1.16165149e+00
5.61375260e-01 -6.73688799e-02 2.59217829e-01 1.17524311e-01
5.96902192e-01 -1.08330917e+00 -1.41140208e-01 -6.37532294e-01
4.59940106e-01 -1.06862688e+00 -2.63443422e-02 -7.92815030e-01
-7.36745775e-01 2.85289139e-01 4.23467755e-01 -2.28086501e-01
1.13790309e+00 -2.16442063e-01 -8.06717202e-02 -1.04987867e-01
-9.40424025e-01 -9.91181433e-02 -2.78631095e-02 4.90308672e-01
7.53189206e-01 6.36483073e-01 -1.08437550e+00 1.26447487e+00
-3.35505903e-01 -1.03695579e-01 1.20260274e+00 2.29269862e-01
-5.62495053e-01 -1.08753848e+00 -4.10307080e-01 8.96946311e-01
-9.27466750e-01 -2.74963260e-01 -6.75123751e-01 5.46658814e-01
-2.21032575e-01 1.93045974e+00 1.44493803e-01 -4.23227280e-01
3.94818634e-01 -3.77593070e-01 1.79634318e-01 -7.09732831e-01
-7.45011568e-01 4.95578468e-01 7.86315680e-01 -5.95514596e-01
-9.08871740e-02 -1.16470063e+00 -3.79839182e-01 -6.47081494e-01
-8.25932503e-01 4.77179557e-01 6.03592277e-01 5.43136418e-01
7.64684677e-01 -1.29818961e-01 1.10907841e+00 -2.37184484e-02
-1.04680872e+00 -8.78658712e-01 -1.26746011e+00 8.65928352e-01
1.59121618e-01 -4.13487464e-01 -6.56965256e-01 -3.58968318e-01] | [9.54659366607666, 5.619514465332031] |
7e40f644-9fc6-45d6-96b8-80b6c1e83f22 | vision-transformer-based-video-hashing | 2112.08117 | null | https://arxiv.org/abs/2112.08117v2 | https://arxiv.org/pdf/2112.08117v2.pdf | Vision Transformer Based Video Hashing Retrieval for Tracing the Source of Fake Videos | In recent years, the spread of fake videos has brought great influence on individuals and even countries. It is important to provide robust and reliable results for fake videos. The results of conventional detection methods are not reliable and not robust for unseen videos. Another alternative and more effective way is to find the original video of the fake video. For example, fake videos from the Russia-Ukraine war and the Hong Kong law revision storm are refuted by finding the original video. We use an improved retrieval method to find the original video, named ViTHash. Specifically, tracing the source of fake videos requires finding the unique one, which is difficult when there are only small differences in the original videos. To solve the above problems, we designed a novel loss Hash Triplet Loss. In addition, we designed a tool called Localizator to compare the difference between the original traced video and the fake video. We have done extensive experiments on FaceForensics++, Celeb-DF and DeepFakeDetection, and we also have done additional experiments on our built three datasets: DAVIS2016-TL (video inpainting), VSTL (video splicing) and DFTL (similar videos). Experiments have shown that our performance is better than state-of-the-art methods, especially in cross-dataset mode. Experiments also demonstrated that ViTHash is effective in various forgery detection: video inpainting, video splicing and deepfakes. Our code and datasets have been released on GitHub: \url{https://github.com/lajlksdf/vtl}. | ['Xuyuan Lai', 'Jinchuan Li', 'Yun Cao', 'Xianfeng Zhao', 'Pengfei Pei'] | 2021-12-15 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [-1.44351050e-01 -4.87557441e-01 -2.72777323e-02 -1.42930165e-01
-5.61409414e-01 -5.88040471e-01 3.85686129e-01 -2.52490103e-01
-2.08078131e-01 9.19256985e-01 -2.87262835e-02 -2.30563179e-01
3.10426444e-01 -6.29827142e-01 -8.61785591e-01 -4.79977667e-01
-9.10018384e-03 7.14100450e-02 3.23601812e-01 -2.10674524e-01
5.52402735e-01 4.74885762e-01 -1.34430230e+00 5.47702074e-01
6.97002232e-01 7.04361558e-01 -8.98937806e-02 4.11904216e-01
1.07710816e-01 6.03746474e-01 -8.43911171e-01 -9.63060975e-01
4.75172907e-01 -6.86111033e-01 -6.57790363e-01 1.22759655e-01
5.86051881e-01 -8.78670454e-01 -8.05019915e-01 1.29101574e+00
4.97761756e-01 -2.19698623e-01 1.95182100e-01 -1.71617365e+00
-6.29053235e-01 2.18481585e-01 -9.64436591e-01 4.35399622e-01
5.80646098e-01 1.41280726e-01 2.21824437e-01 -9.95551288e-01
8.90389562e-01 1.47684431e+00 6.70469344e-01 5.85871458e-01
-6.28194332e-01 -1.23047876e+00 -5.00961900e-01 6.61812603e-01
-1.55251455e+00 -5.72418749e-01 6.66486979e-01 -3.13996583e-01
4.18398499e-01 2.61053711e-01 5.96824706e-01 1.26949346e+00
2.70568818e-01 6.80798471e-01 1.16036057e+00 -2.02712521e-01
-3.14156353e-01 2.38040239e-01 -3.57415587e-01 9.64525342e-01
4.87548381e-01 2.54561305e-01 -7.53373325e-01 -4.15471882e-01
7.07583487e-01 1.01640739e-01 -6.06996238e-01 2.68371403e-01
-9.90458667e-01 9.66521561e-01 1.09120741e-01 3.54919523e-01
-4.71153893e-02 1.96491301e-01 4.85648036e-01 6.30840659e-01
3.17553878e-01 7.41124228e-02 5.63835353e-02 -1.49326712e-01
-1.25056839e+00 3.88562083e-01 5.85651398e-01 7.55540490e-01
7.08749056e-01 -1.28124312e-01 5.85095547e-02 6.31142616e-01
2.00770840e-01 4.86324042e-01 4.90659088e-01 -9.81051922e-01
6.01153195e-01 9.91890877e-02 1.92469016e-01 -1.73743260e+00
3.73017550e-01 2.04019994e-01 -6.34840667e-01 -1.07010365e-01
4.41638559e-01 -7.37531930e-02 -8.00723195e-01 1.30248892e+00
2.70931363e-01 6.15781128e-01 -2.96662748e-01 1.10678792e+00
7.59870946e-01 7.35690594e-01 -4.22556579e-01 -2.67038643e-01
1.16145742e+00 -7.67716885e-01 -9.08640087e-01 1.11194171e-01
5.93349814e-01 -1.26217258e+00 7.36631691e-01 6.73078120e-01
-8.42047334e-01 -3.15754056e-01 -8.77727032e-01 1.15061603e-01
-2.76992261e-01 1.53457001e-01 3.47225308e-01 8.95618320e-01
-8.63816738e-01 7.93920875e-01 -5.22315145e-01 -3.47723961e-01
7.04096079e-01 1.08545713e-01 -7.89853692e-01 -3.48548800e-01
-1.34630048e+00 5.97445428e-01 3.82293433e-01 2.51750499e-01
-1.04363704e+00 -3.18448812e-01 -4.20390040e-01 -3.17421824e-01
4.33345497e-01 -5.39781302e-02 7.81914711e-01 -1.31579888e+00
-1.00278485e+00 9.50322986e-01 2.07962431e-02 -3.33421379e-01
8.67881536e-01 -2.06268579e-01 -7.65277803e-01 5.56585133e-01
2.55856633e-01 4.86533016e-01 1.33089435e+00 -1.09539390e+00
-4.54957962e-01 -4.10741597e-01 -2.85642266e-01 -2.16170579e-01
-2.62621641e-01 3.61848056e-01 -4.83307838e-01 -9.40484464e-01
-4.17957716e-02 -9.23506260e-01 4.79098797e-01 1.45207822e-01
-5.11179030e-01 8.78961682e-02 1.52827179e+00 -1.25407958e+00
1.25597644e+00 -2.35048270e+00 -2.22392961e-01 1.67072892e-01
5.83398081e-02 6.78311765e-01 -1.49752945e-01 6.45264328e-01
-1.81907997e-01 4.12742287e-01 -8.51220414e-02 -8.38037282e-02
-3.35702449e-01 3.96857858e-02 -5.35868645e-01 9.74174619e-01
-5.39690368e-02 5.31653404e-01 -8.80491376e-01 -6.91829801e-01
-6.13976158e-02 4.43500370e-01 -4.17603076e-01 5.78775369e-02
1.82645485e-01 3.83596808e-01 -2.38239840e-01 1.04215026e+00
1.14439702e+00 2.66679138e-01 5.84490411e-02 -1.65074512e-01
4.36844826e-02 -7.66002834e-02 -1.16244674e+00 1.45379472e+00
1.74065068e-01 9.81566489e-01 3.37037854e-02 -7.28752732e-01
7.01279223e-01 4.08561081e-01 2.39040583e-01 -4.84781235e-01
2.51262814e-01 4.70098287e-01 -3.19200903e-01 -8.50949168e-01
6.11799002e-01 2.99996929e-03 1.38466224e-01 4.68263716e-01
1.94897521e-02 1.29135102e-01 1.20863825e-01 5.23666620e-01
1.04269922e+00 1.24272920e-01 -1.15089990e-01 1.74746320e-01
3.12588155e-01 -2.77470462e-02 6.25749350e-01 5.75729668e-01
-4.75391597e-01 5.79942226e-01 6.45773888e-01 -3.32209229e-01
-9.08291340e-01 -8.67137969e-01 1.00543208e-01 3.46895903e-01
3.89388442e-01 -7.14356482e-01 -7.67260909e-01 -9.37303066e-01
1.86958671e-01 3.14882994e-01 -3.88741791e-01 -2.11841390e-01
-6.15841508e-01 -4.04053807e-01 1.13879454e+00 -2.06012309e-01
1.10011303e+00 -1.02414382e+00 -2.49231145e-01 -8.36893842e-02
-6.15868688e-01 -1.11899900e+00 -7.00627446e-01 -8.50862324e-01
-7.10941911e-01 -1.35735512e+00 -6.90612078e-01 -6.37025654e-01
6.11917615e-01 6.72801197e-01 5.33861518e-01 7.86713719e-01
-5.28657734e-01 1.93207916e-02 -5.36460638e-01 -4.67912666e-02
-4.67092216e-01 -6.10152602e-01 1.59935206e-01 1.77995682e-01
2.18028337e-01 -1.97429150e-01 -5.50426424e-01 5.27900636e-01
-1.11757588e+00 -3.67931962e-01 1.57929197e-01 7.95185149e-01
1.58841982e-01 4.24799353e-01 3.33302677e-01 -7.46445358e-01
4.97047931e-01 -4.92124736e-01 -5.21573603e-01 2.29907572e-01
-2.14671463e-01 -3.28670591e-01 4.71577346e-01 -3.72886062e-01
-8.90828192e-01 -2.33253002e-01 -5.07006533e-02 -9.06857729e-01
-3.16718705e-02 1.71461955e-01 -1.70368906e-02 -3.62624854e-01
3.54987860e-01 2.04238251e-01 1.81473926e-01 -4.38732833e-01
5.82537754e-03 1.00160241e+00 4.66843516e-01 -2.70841956e-01
1.01619911e+00 6.37624681e-01 -1.75280362e-01 -1.02557886e+00
-1.41757339e-01 -2.37627968e-01 -1.77968398e-01 -5.00220478e-01
4.43039089e-01 -8.63847613e-01 -6.55592084e-01 9.97835159e-01
-1.40295267e+00 1.78455979e-01 4.40809309e-01 4.39831734e-01
-7.28661194e-02 1.23990762e+00 -6.65939271e-01 -6.35736823e-01
-1.30742550e-01 -1.14715052e+00 7.92378783e-01 6.38130456e-02
8.51011276e-02 -5.84751010e-01 -7.49844536e-02 7.07557619e-01
8.22870806e-02 4.74911839e-01 4.77679163e-01 -2.62165427e-01
-6.82297051e-01 -2.71617383e-01 -3.48527193e-01 4.83114362e-01
1.97089002e-01 3.58756036e-01 -6.79073453e-01 -6.38870358e-01
9.48488340e-02 -3.35603803e-01 9.00395155e-01 -4.12038937e-02
1.12580037e+00 -5.25794148e-01 -3.88723344e-01 5.89084208e-01
1.40895522e+00 1.88751951e-01 1.22718966e+00 2.53023952e-01
5.38552284e-01 4.87423539e-01 9.29729044e-01 4.97119337e-01
-9.10353288e-02 7.56523490e-01 3.23331058e-01 1.86782479e-01
-1.65532902e-01 -3.64073128e-01 7.71228254e-01 4.81755733e-01
-8.61198679e-02 -4.73292917e-01 -4.63285178e-01 5.30847132e-01
-1.70702493e+00 -1.54507422e+00 -4.01789665e-01 2.34923816e+00
6.16199732e-01 -1.50288403e-01 1.56293198e-01 1.32618934e-01
9.50995564e-01 6.28881454e-02 -1.33314177e-01 -4.32409555e-01
-1.66400045e-01 1.27579823e-01 5.92931032e-01 3.13154906e-01
-1.02921379e+00 1.18559647e+00 5.34387445e+00 1.17847383e+00
-1.23731172e+00 5.04757524e-01 4.39324886e-01 -1.31378651e-01
-3.41088995e-02 1.81684405e-01 -5.59063554e-01 1.02206028e+00
7.12144971e-01 2.47522965e-02 5.66826761e-01 4.27688956e-01
4.52738225e-01 -3.70589882e-01 -4.87039000e-01 1.33108103e+00
4.02389377e-01 -1.47868609e+00 1.14508308e-02 1.36571184e-01
4.37265277e-01 -3.38227868e-01 -3.85541990e-02 4.02984098e-02
-2.80788451e-01 -8.74428749e-01 7.26210594e-01 2.33333096e-01
9.45529282e-01 -8.03300619e-01 7.80982137e-01 1.20098434e-01
-8.53001893e-01 1.07583046e-01 -5.95230997e-01 2.49672145e-01
8.64696354e-02 6.17803335e-01 -7.92942822e-01 5.81840813e-01
9.77371633e-01 8.12240124e-01 -4.85527188e-01 1.07053459e+00
-2.99493641e-01 6.40436709e-01 -1.98022410e-01 2.58268237e-01
9.63185579e-02 -2.23061308e-01 7.90702403e-01 1.07002449e+00
6.61263585e-01 3.98863554e-02 -2.42906272e-01 6.75919592e-01
-3.73538494e-01 -9.24265850e-03 -8.86004806e-01 -3.01195204e-01
5.02149522e-01 9.91977930e-01 -7.52179623e-01 -1.77329406e-01
-2.19735071e-01 1.42939842e+00 -6.12397343e-02 1.04069650e-01
-1.34874165e+00 -6.02742910e-01 6.06671870e-01 3.52616817e-01
4.16545123e-01 -1.01019137e-01 5.09966731e-01 -1.52476799e+00
1.38462797e-01 -1.22476053e+00 3.52534264e-01 -7.18469679e-01
-1.05249345e+00 4.92360115e-01 -9.52747390e-02 -1.24187815e+00
8.25577602e-02 -2.34884173e-01 -3.98715496e-01 4.35236275e-01
-1.24152648e+00 -8.39185953e-01 -3.19907516e-01 9.03361082e-01
4.82594371e-01 -3.63440633e-01 3.14649552e-01 6.61350667e-01
-6.01589561e-01 6.56266451e-01 8.82895812e-02 2.87178278e-01
1.11603498e+00 -4.47730273e-01 1.17234081e-01 1.13492072e+00
1.09406307e-01 5.06981790e-01 7.21312821e-01 -1.10423505e+00
-1.49204421e+00 -9.43156898e-01 8.51230323e-01 -8.98177847e-02
5.17526805e-01 -2.40251243e-01 -8.14235449e-01 5.53363621e-01
3.27264220e-01 -1.41472727e-01 3.75308335e-01 -8.95841956e-01
-4.28948730e-01 1.32803367e-02 -1.57284033e+00 2.51495630e-01
8.90942156e-01 -5.18023014e-01 -3.78732264e-01 5.71287334e-01
3.72700006e-01 -2.32423231e-01 -4.39272016e-01 -2.54173446e-02
7.34678447e-01 -1.41184533e+00 7.88112044e-01 -4.02699471e-01
4.76381570e-01 -4.27508891e-01 -9.53320488e-02 -9.83801663e-01
1.65495992e-01 -9.39866066e-01 6.05491176e-02 1.30212188e+00
-6.03600554e-02 -6.88953996e-01 7.52792835e-01 -2.82303169e-02
2.74349362e-01 -3.38842154e-01 -1.03140426e+00 -9.13711488e-01
-4.01207447e-01 -1.02328233e-01 5.79750001e-01 1.28095293e+00
-2.88353443e-01 -3.46913218e-01 -1.09573007e+00 1.67689204e-01
7.60647655e-01 -5.52182756e-02 8.65357876e-01 -6.89413726e-01
-2.11947113e-01 3.43671143e-02 -6.83662415e-01 -6.21325374e-01
2.32405644e-02 -5.73897958e-01 -3.63853991e-01 -9.75513697e-01
2.67692149e-01 -7.91378766e-02 2.13227719e-01 5.47486901e-01
1.48783684e-01 7.19998360e-01 3.43756258e-01 5.07651031e-01
-1.57448292e-01 3.66013527e-01 1.30808866e+00 -1.68663859e-02
1.26296267e-01 -1.89163804e-01 -2.41857961e-01 4.54762995e-01
8.50728095e-01 -9.67524946e-01 -6.28261715e-02 -3.36475909e-01
1.85899215e-03 2.15583012e-01 7.55405605e-01 -9.23397958e-01
-2.56650392e-02 -8.22718590e-02 3.12195927e-01 -6.01878881e-01
3.98929805e-01 -6.69397712e-01 3.74470979e-01 7.94228852e-01
2.31210843e-01 2.94709414e-01 1.15434617e-01 6.16506159e-01
-3.10948819e-01 -3.48781854e-01 8.79867435e-01 -2.42823660e-01
-7.03647673e-01 1.99908063e-01 -2.92099595e-01 -7.45104924e-02
1.17576611e+00 -4.47295815e-01 -7.47923255e-01 -6.19958043e-01
-3.14478248e-01 -1.50866255e-01 6.79807544e-01 4.64620799e-01
1.17096364e+00 -1.21058369e+00 -7.71319926e-01 3.03047240e-01
-2.43364915e-01 -8.67773056e-01 1.97657466e-01 8.71318340e-01
-1.09708416e+00 5.09005226e-02 -5.94213843e-01 -2.30075732e-01
-1.75791955e+00 5.09597838e-01 1.68031782e-01 2.67921388e-01
-3.80053788e-01 6.86047316e-01 -3.97577107e-01 -8.00445583e-03
4.40437458e-02 1.87777385e-01 2.92162150e-01 5.55486046e-02
6.62663102e-01 6.18084431e-01 4.84783165e-02 -9.74827588e-01
-5.52242219e-01 3.31247151e-01 -2.36810312e-01 4.65172790e-02
1.17557549e+00 -7.41678104e-02 -4.57166076e-01 -2.67138958e-01
1.43408895e+00 2.67864436e-01 -7.68545985e-01 2.61879474e-01
-2.86684841e-01 -1.37634325e+00 9.62921605e-02 -4.91383851e-01
-1.45827591e+00 6.17047608e-01 6.36904001e-01 1.08259909e-01
1.09041154e+00 -3.83869916e-01 1.34544718e+00 3.54225673e-02
6.15499914e-01 -8.59647214e-01 2.77364999e-01 2.54052859e-02
9.01900768e-01 -1.13069379e+00 2.65449315e-01 -4.98302966e-01
-5.64357221e-01 1.23204541e+00 3.98730308e-01 -1.56408027e-01
5.16578913e-01 -1.02961943e-01 -9.88169387e-02 -1.66885540e-01
-2.92505264e-01 3.68524045e-01 -1.49660200e-01 3.58452708e-01
6.09513447e-02 -8.78179148e-02 -5.83801210e-01 6.32247329e-02
-6.39976710e-02 2.99686164e-01 8.74140620e-01 1.08558202e+00
-2.60498554e-01 -1.40740609e+00 -8.11610162e-01 2.53741801e-01
-7.10926116e-01 5.66136166e-02 -6.87098444e-01 9.71632063e-01
2.83203274e-01 1.11798978e+00 -2.33661249e-01 -6.63455784e-01
-1.17641494e-01 -2.78134912e-01 6.22989535e-01 -2.10161611e-01
-6.04041517e-01 5.86956479e-02 1.89933367e-02 -8.28917444e-01
-3.96055400e-01 -5.82423389e-01 -9.13241863e-01 -1.16552758e+00
-5.42020917e-01 2.49130771e-01 5.80496609e-01 5.83914578e-01
3.81260008e-01 -2.99934000e-01 7.88985789e-01 -5.77873290e-01
-3.09271008e-01 -6.66182995e-01 -7.18289614e-01 5.51867127e-01
2.25897908e-01 -6.80551529e-01 -6.11709476e-01 1.15251467e-01] | [12.521074295043945, 1.0634490251541138] |
5208e290-4c74-4820-b155-8edafcfbe4a2 | deep-attention-fusion-feature-for-speech | 2003.07544 | null | https://arxiv.org/abs/2003.07544v1 | https://arxiv.org/pdf/2003.07544v1.pdf | Deep Attention Fusion Feature for Speech Separation with End-to-End Post-filter Method | In this paper, we propose an end-to-end post-filter method with deep attention fusion features for monaural speaker-independent speech separation. At first, a time-frequency domain speech separation method is applied as the pre-separation stage. The aim of pre-separation stage is to separate the mixture preliminarily. Although this stage can separate the mixture, it still contains the residual interference. In order to enhance the pre-separated speech and improve the separation performance further, the end-to-end post-filter (E2EPF) with deep attention fusion features is proposed. The E2EPF can make full use of the prior knowledge of the pre-separated speech, which contributes to speech separation. It is a fully convolutional speech separation network and uses the waveform as the input features. Firstly, the 1-D convolutional layer is utilized to extract the deep representation features for the mixture and pre-separated signals in the time domain. Secondly, to pay more attention to the outputs of the pre-separation stage, an attention module is applied to acquire deep attention fusion features, which are extracted by computing the similarity between the mixture and the pre-separated speech. These deep attention fusion features are conducive to reduce the interference and enhance the pre-separated speech. Finally, these features are sent to the post-filter to estimate each target signals. Experimental results on the WSJ0-2mix dataset show that the proposed method outperforms the state-of-the-art speech separation method. Compared with the pre-separation method, our proposed method can acquire 64.1%, 60.2%, 25.6% and 7.5% relative improvements in scale-invariant source-to-noise ratio (SI-SNR), the signal-to-distortion ratio (SDR), the perceptual evaluation of speech quality (PESQ) and the short-time objective intelligibility (STOI) measures, respectively. | ['Jian-Hua Tao', 'Cunhang Fan', 'Bin Liu', 'Zhengqi Wen', 'Xuefei Liu', 'Jiangyan Yi'] | 2020-03-17 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-3.42408232e-02 -4.13219035e-01 4.79670197e-01 -1.28541216e-01
-1.04078424e+00 -1.44983843e-01 4.78414185e-02 -1.06827691e-01
-3.80211115e-01 3.79147828e-01 4.46006149e-01 -6.24125935e-02
-3.01415801e-01 -3.41955841e-01 -2.25315556e-01 -1.07643747e+00
7.43346065e-02 -3.75467032e-01 8.87742937e-02 -2.62330890e-01
-4.63338085e-02 3.36719006e-01 -1.79623091e+00 1.47716060e-01
1.26626492e+00 1.29864013e+00 5.72883129e-01 7.89615393e-01
9.40352380e-02 3.39332014e-01 -8.78269136e-01 5.22228479e-02
7.65375793e-04 -5.79233944e-01 -1.65029392e-01 5.74788190e-02
3.84272030e-03 -2.22548798e-01 -5.59908390e-01 1.46131980e+00
1.19866037e+00 5.09504974e-01 4.05781597e-01 -9.32305813e-01
-4.64277357e-01 6.32247090e-01 -4.70150858e-01 6.45832479e-01
5.07499650e-02 1.93382502e-01 5.52675605e-01 -1.06592917e+00
-4.23263043e-01 1.15396357e+00 3.08338732e-01 3.01866919e-01
-7.72370696e-01 -9.20555651e-01 -3.69118638e-02 4.43586498e-01
-1.30562496e+00 -9.67028141e-01 9.85176682e-01 -1.96188420e-01
9.82007205e-01 4.54564720e-01 4.67360526e-01 4.38846886e-01
-3.36742252e-02 7.56572247e-01 9.33727026e-01 -4.76863295e-01
-6.35457411e-02 -2.15550229e-01 3.89845669e-01 2.36665294e-01
-2.87029207e-01 3.96809250e-01 -3.44036758e-01 3.61987084e-01
3.51936162e-01 -1.81653500e-01 -8.32618237e-01 6.71111584e-01
-8.15761447e-01 3.33296359e-01 4.04190183e-01 7.07900524e-01
-5.33253729e-01 -4.02873158e-01 2.84335196e-01 2.33002722e-01
6.00727499e-01 1.07729167e-01 -4.42919016e-01 -2.09265202e-01
-1.09886932e+00 -1.54979691e-01 4.43851858e-01 5.54026067e-01
4.18148220e-01 5.49633920e-01 -4.07540202e-01 1.20729196e+00
4.24513727e-01 7.41897106e-01 8.43184531e-01 -6.99138939e-01
5.06962776e-01 -8.68829265e-02 -1.23343900e-01 -7.42760062e-01
-1.85501099e-01 -8.88524652e-01 -9.09874082e-01 1.33875221e-01
-2.23617703e-02 -3.10538650e-01 -1.08436882e+00 1.70366597e+00
2.09006742e-01 3.40238720e-01 3.51852655e-01 1.29542768e+00
1.05526745e+00 1.08023822e+00 -2.38472819e-01 -8.00945461e-01
1.34150386e+00 -1.18477249e+00 -1.21715641e+00 -2.00242132e-01
-2.43648112e-01 -1.19909549e+00 6.84827030e-01 4.61982101e-01
-1.40894616e+00 -1.18052721e+00 -1.14153707e+00 1.50038436e-01
2.76704971e-02 4.75842595e-01 -2.29795292e-01 7.05710053e-01
-8.98511946e-01 4.22068268e-01 -7.60867536e-01 2.44174734e-01
1.90534845e-01 2.85964966e-01 -1.24554895e-01 7.05511868e-02
-1.40182686e+00 6.34892166e-01 1.82347104e-01 3.00721198e-01
-9.41699147e-01 -5.98374128e-01 -7.66868353e-01 6.40364349e-01
1.75646916e-01 -4.30791944e-01 1.25044739e+00 -1.05379975e+00
-1.81622982e+00 2.37405717e-01 -3.94362599e-01 -2.95492262e-01
-1.27143100e-01 -2.86625028e-01 -1.09988081e+00 2.18258828e-01
-1.15771279e-01 1.20292023e-01 1.06067884e+00 -8.04640532e-01
-8.20645452e-01 -2.76421636e-01 -4.28706676e-01 5.82879543e-01
-2.41257176e-01 2.95179367e-01 -4.36850965e-01 -9.11319792e-01
2.53541201e-01 -4.90317822e-01 7.47991726e-02 -6.74900115e-01
-3.20172131e-01 -1.18993938e-01 7.95618713e-01 -1.29235518e+00
1.30411184e+00 -2.85460591e+00 1.71389207e-01 -1.09683201e-01
1.60680190e-01 9.25741971e-01 -3.26497436e-01 -8.73305723e-02
-3.38963509e-01 -2.62409598e-01 -1.75625131e-01 -4.73880589e-01
-1.19443953e-01 -3.99831921e-01 -1.17700219e-01 4.66025114e-01
1.46962389e-01 2.92620957e-01 -6.86408103e-01 -2.89258450e-01
5.25812089e-01 8.39111924e-01 -3.27253819e-01 5.32720149e-01
5.51486075e-01 3.88661474e-01 1.01257697e-01 1.13309965e-01
1.02302313e+00 5.51756978e-01 -5.06516278e-01 -3.67861301e-01
-3.37370753e-01 6.52718127e-01 -1.23376584e+00 1.53217626e+00
-6.10028148e-01 7.72183120e-01 6.95392311e-01 -9.12305117e-01
1.00400960e+00 6.82792783e-01 1.99003771e-01 -7.62492299e-01
4.90419745e-01 3.81302774e-01 5.75629115e-01 -5.57559788e-01
-1.28453197e-02 -3.35114539e-01 4.01314318e-01 -3.51065025e-02
3.50462466e-01 -1.28466636e-01 -4.22012731e-02 -1.63817510e-01
5.28189182e-01 -4.76003349e-01 9.38190892e-02 -1.24865368e-01
1.05428612e+00 -9.25883055e-01 7.23408580e-01 2.77279057e-02
-4.25414681e-01 7.13163257e-01 -2.43010595e-01 2.71019608e-01
-5.20680964e-01 -1.18345022e+00 -1.17804892e-02 8.11469257e-01
2.47064233e-01 -2.41111130e-01 -8.98180366e-01 -1.13944694e-01
-3.80915016e-01 8.91552508e-01 -4.23439145e-02 -5.88592887e-01
-4.04386878e-01 -4.05445486e-01 4.27925050e-01 4.89974707e-01
6.36008143e-01 -9.92954373e-01 -1.73455313e-01 3.50464910e-01
-5.52608132e-01 -8.32699120e-01 -7.73248196e-01 2.31661692e-01
-4.44100887e-01 -5.36007822e-01 -9.47215438e-01 -9.00609732e-01
3.95352244e-01 5.49043119e-01 2.96928108e-01 -1.71921790e-01
1.16028532e-01 3.87168713e-02 -2.03401402e-01 -5.93965888e-01
-3.83396417e-01 -3.72030944e-01 3.12690645e-01 3.94851834e-01
3.50796908e-01 -7.40329742e-01 -6.66819930e-01 2.52120972e-01
-6.94305241e-01 -8.37991536e-02 6.40641868e-01 7.07056284e-01
4.36345488e-01 8.47236872e-01 7.38915682e-01 3.44166756e-01
9.20075059e-01 -1.54424638e-01 -4.65450764e-01 -1.84405163e-01
-2.31612369e-01 -3.23579282e-01 9.57554758e-01 -7.03399062e-01
-1.39999056e+00 -2.98759699e-01 -4.88353074e-01 -6.67574584e-01
-3.15169603e-01 5.36330283e-01 -7.81481564e-01 3.46052498e-01
1.78466350e-01 4.83267814e-01 -2.89204437e-02 -6.99677706e-01
-2.21749605e-03 1.41142011e+00 7.19258726e-01 3.36876847e-02
6.66624129e-01 -4.38683107e-02 -5.55327415e-01 -1.01730156e+00
-6.05428219e-01 -6.63824737e-01 -1.22020811e-01 -1.54917929e-02
8.63753557e-01 -1.12885511e+00 -5.41089833e-01 8.97911310e-01
-1.22561741e+00 2.78287157e-02 -1.09457254e-01 1.24607217e+00
-2.40883902e-01 4.36314702e-01 -6.81323171e-01 -1.23515844e+00
-7.46354520e-01 -1.49448049e+00 7.71628201e-01 7.05208242e-01
1.86371386e-01 -3.53734255e-01 -2.21705094e-01 3.09944779e-01
6.72320962e-01 -5.11066794e-01 5.00375509e-01 -7.61942685e-01
-2.79077142e-03 1.55637292e-02 1.05759114e-01 9.62473214e-01
5.06563723e-01 -2.35770345e-01 -1.40861309e+00 -3.65962863e-01
6.81651890e-01 2.95542985e-01 7.89288640e-01 8.18021417e-01
9.27521527e-01 -2.74699926e-01 5.05788438e-02 6.66978657e-01
9.63856578e-01 8.68274629e-01 8.18337619e-01 -2.38771662e-01
4.53224778e-01 4.21667695e-01 5.75180948e-01 2.13375434e-01
7.17245862e-02 5.36070168e-01 8.52685869e-02 -4.64927852e-01
-7.34610260e-01 1.27344191e-01 5.55448711e-01 1.63925517e+00
1.28546312e-01 -2.61983544e-01 -4.88634169e-01 7.31837153e-01
-1.30412221e+00 -1.01374424e+00 -1.09932691e-01 2.26336694e+00
9.15452957e-01 2.34528989e-01 -5.69311045e-02 8.26767623e-01
1.10514355e+00 6.34478033e-02 -4.54103619e-01 -3.51139486e-01
-1.52758300e-01 5.70852935e-01 3.89294028e-02 7.52957046e-01
-1.07824004e+00 4.38733906e-01 4.19592667e+00 1.24825287e+00
-1.43436611e+00 1.71274781e-01 3.81372392e-01 -2.20046595e-01
1.33298352e-01 -3.87575001e-01 -4.03177708e-01 7.41054714e-01
1.19017005e+00 -3.74747187e-01 8.70052636e-01 4.37448323e-01
3.60311806e-01 -4.37792242e-02 -6.52071357e-01 1.30186319e+00
1.44906506e-01 -4.63604361e-01 -4.24661368e-01 -2.27021813e-01
3.21831197e-01 -1.01020224e-01 2.28886470e-01 3.93450230e-01
-3.20227712e-01 -9.07241344e-01 7.67125070e-01 5.92866600e-01
8.77046943e-01 -1.16692257e+00 7.95519054e-01 4.68770027e-01
-1.44706643e+00 -2.29404941e-01 -2.35944837e-01 4.83691171e-02
1.72567740e-01 9.66006994e-01 -4.40208256e-01 7.75789678e-01
7.84545898e-01 2.94211864e-01 2.96662580e-02 1.28790081e+00
-4.68037695e-01 8.87417018e-01 -2.64757991e-01 2.49501005e-01
-6.96828067e-02 -1.32645192e-02 9.72461045e-01 1.28682685e+00
5.08476138e-01 4.62531000e-01 -2.26044416e-01 6.05282128e-01
8.44871998e-02 -2.94557959e-03 7.01363757e-02 -1.86741836e-02
4.98856574e-01 1.19740760e+00 -2.55039990e-01 -3.88339132e-01
-2.58602411e-01 8.70563209e-01 -2.43642598e-01 6.68160856e-01
-1.00284636e+00 -1.39930332e+00 8.55770946e-01 -2.86787748e-01
5.64980328e-01 -7.65129626e-02 1.05508298e-01 -8.93090963e-01
6.66848645e-02 -1.09319127e+00 -4.52774875e-02 -1.01511586e+00
-1.04435086e+00 8.44809055e-01 -4.66757596e-01 -1.28172457e+00
1.00659102e-01 -3.28685492e-01 -8.01146150e-01 1.71508777e+00
-1.61900640e+00 -6.94925725e-01 -1.63459286e-01 5.89593649e-01
8.76023054e-01 -2.03487128e-01 5.74734032e-01 6.85051739e-01
-7.43907690e-01 7.04642951e-01 1.10714450e-01 1.07493386e-01
6.71631098e-01 -9.55020130e-01 2.03879461e-01 1.33636725e+00
-2.04896614e-01 5.15166938e-01 5.12711048e-01 -3.91238928e-01
-8.96170139e-01 -1.01783776e+00 8.08743298e-01 4.16698426e-01
2.09917620e-01 -2.05363855e-01 -9.58631694e-01 -4.57030647e-02
3.31183195e-01 -4.00546715e-02 6.36911750e-01 -5.24946749e-01
6.16414845e-02 -5.66468775e-01 -9.44707572e-01 4.28811848e-01
5.53744435e-01 -7.66218126e-01 -9.84862328e-01 -2.57238984e-01
1.07721353e+00 -4.25352961e-01 -5.85186064e-01 4.23584908e-01
3.45147043e-01 -9.22451675e-01 1.00443816e+00 -1.09020203e-01
1.52451545e-01 -7.11677849e-01 -2.34405711e-01 -1.94689465e+00
-5.43878317e-01 -8.73174250e-01 2.56696325e-02 1.66588831e+00
2.25610048e-01 -6.76770151e-01 -1.68543339e-01 1.20649457e-01
-6.90208614e-01 -3.90278310e-01 -9.02950048e-01 -8.14546764e-01
-1.85638562e-01 -4.56361353e-01 7.27410853e-01 6.68645501e-01
3.93449180e-02 4.55302864e-01 -1.46254316e-01 6.73073828e-01
3.87679398e-01 -8.34568590e-02 2.76308835e-01 -9.25582290e-01
-3.22061956e-01 -7.04744995e-01 -1.46714866e-01 -1.10099149e+00
3.32116522e-02 -6.68754578e-01 3.70749980e-01 -1.70009315e+00
-3.81186604e-01 1.96588844e-01 -8.19087267e-01 8.14754516e-02
-7.02153862e-01 -2.78898835e-01 2.83242524e-01 -1.59452423e-01
-1.40190005e-01 9.48203027e-01 1.26928723e+00 -3.49001557e-01
-4.66690809e-01 3.54368091e-01 -6.64623916e-01 6.10546291e-01
6.88996732e-01 -3.51524174e-01 -3.89478385e-01 -3.80209357e-01
-6.86361134e-01 3.79920006e-01 5.00493832e-02 -1.39972782e+00
4.17733312e-01 1.68288469e-01 3.87031913e-01 -7.99828351e-01
5.62525809e-01 -8.91815960e-01 -1.11918978e-01 3.54508638e-01
-1.54875860e-01 -6.07396305e-01 5.25958598e-01 2.97079504e-01
-6.47026181e-01 -4.02041450e-02 1.01857817e+00 2.69886553e-01
-1.87307149e-01 4.19797488e-02 -4.66352940e-01 -7.40979537e-02
6.89237475e-01 -3.79848713e-03 -1.77035242e-01 -4.64933634e-01
-6.96987927e-01 1.67501196e-02 -3.03757370e-01 2.94201493e-01
7.80423462e-01 -1.37533736e+00 -1.00819099e+00 5.56461453e-01
-3.73045117e-01 -2.89062075e-02 7.88878918e-01 1.17872751e+00
7.54109174e-02 2.87057608e-01 -1.54100418e-01 -3.68722856e-01
-1.36290407e+00 6.94771409e-01 6.18152440e-01 2.31866509e-01
-2.37895787e-01 1.09482801e+00 5.06844342e-01 1.75039679e-01
4.18028146e-01 -5.21804750e-01 -4.57559317e-01 -3.02833728e-02
8.99781287e-01 6.62117541e-01 3.10575455e-01 -8.65379453e-01
-4.26103592e-01 4.98733103e-01 1.86158821e-01 -2.77941823e-01
1.30958688e+00 -4.44389910e-01 -8.08371753e-02 2.48476416e-01
1.43275511e+00 5.93576431e-01 -1.05760503e+00 -2.00242162e-01
-6.79532826e-01 -4.13585395e-01 6.48340046e-01 -1.04421496e+00
-1.47307944e+00 1.44417942e+00 9.43070054e-01 3.98263335e-01
1.87159169e+00 -3.82268697e-01 1.25608873e+00 -2.69046158e-01
-2.77266502e-01 -8.57417703e-01 -6.90268725e-02 5.22191942e-01
1.02066803e+00 -7.77665257e-01 -4.16429102e-01 -2.03910530e-01
-3.47708970e-01 1.04020405e+00 5.36125362e-01 1.20057493e-01
6.42715633e-01 4.50511098e-01 2.22348511e-01 2.86089391e-01
-2.45123491e-01 -5.38255692e-01 6.41761065e-01 5.73877454e-01
3.95622045e-01 6.39001653e-02 -1.55055314e-01 1.28560698e+00
-4.15960640e-01 -4.20714319e-01 1.15426004e-01 3.34216684e-01
-6.45285964e-01 -6.20226860e-01 -7.93691993e-01 1.46855339e-01
-6.56620026e-01 -4.06183839e-01 -1.02576703e-01 7.77718937e-03
2.43193179e-01 1.80157161e+00 9.97455865e-02 -6.40798926e-01
6.63000762e-01 6.49290085e-02 1.56194672e-01 -2.87320137e-01
-6.42142057e-01 8.93786311e-01 -3.29789221e-02 -1.72265112e-01
-2.39405781e-01 -2.43579641e-01 -1.37332141e+00 9.89512801e-02
-7.96901107e-01 4.14167106e-01 6.79301620e-01 8.35372984e-01
3.99143338e-01 1.23861802e+00 9.88326967e-01 -8.67367983e-01
-3.05022836e-01 -1.28915882e+00 -6.20162249e-01 3.17421734e-01
9.92185771e-01 -4.78542328e-01 -8.23164523e-01 -8.24212749e-03] | [14.905877113342285, 5.8133392333984375] |
c117e2cf-deee-4141-b778-91d3f2984a45 | an-artificial-life-simulation-library-based | 2304.13520 | null | https://arxiv.org/abs/2304.13520v1 | https://arxiv.org/pdf/2304.13520v1.pdf | An Artificial Life Simulation Library Based on Genetic Algorithm, 3-Character Genetic Code and Biological Hierarchy | Genetic algorithm (GA) is inspired by biological evolution of genetic organisms by optimizing the genotypic combinations encoded within each individual with the help of evolutionary operators, suggesting that GA may be a suitable model for studying real-life evolutionary processes. This paper describes the design of a Python library for artificial life simulation, Digital Organism Simulation Environment (DOSE), based on GA and biological hierarchy starting from genetic sequence to population. A 3-character instruction set that does not take any operand is introduced as genetic code for digital organism. This mimics the 3-nucleotide codon structure in naturally occurring DNA. In addition, the context of a 3-dimensional world composing of ecological cells is introduced to simulate a physical ecosystem. Using DOSE, an experiment to examine the changes in genetic sequences with respect to mutation rates is presented. | ['Maurice HT Ling'] | 2023-02-19 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 4.62616161e-02 -3.56540173e-01 3.98417920e-01 2.90860441e-02
1.08561397e+00 -5.39372981e-01 4.64070380e-01 1.61102369e-01
-4.67836797e-01 9.50007379e-01 -3.23478520e-01 -4.58593428e-01
2.20917702e-01 -1.14031172e+00 -6.54276848e-01 -8.98098290e-01
-4.30217564e-01 1.26416549e-01 2.32633606e-01 -5.21224678e-01
6.07066453e-01 5.47510326e-01 -2.12811327e+00 -7.64509961e-02
1.08522308e+00 -3.90365124e-02 4.31491345e-01 1.06967521e+00
-4.12225991e-01 1.30584553e-01 -9.44894075e-01 -6.31611003e-03
2.77362198e-01 -1.10648799e+00 -2.34188676e-01 -1.61891237e-01
-5.96628785e-01 3.06768626e-01 5.68745248e-02 8.36233854e-01
5.88644743e-01 6.48966655e-02 6.99998558e-01 -1.00729573e+00
-8.43793094e-01 4.25798833e-01 -1.76799625e-01 4.33122963e-02
4.24224466e-01 6.32798672e-01 1.36196807e-01 -3.67545664e-01
5.19087732e-01 1.31822169e+00 5.98540008e-01 5.44059873e-01
-8.20270419e-01 -1.20957345e-01 -5.07385492e-01 -1.63087294e-01
-1.62561417e+00 3.51135910e-01 1.01786599e-01 -3.87433439e-01
1.30078900e+00 5.97550452e-01 1.46226513e+00 6.86727226e-01
1.30103528e+00 3.33096795e-02 9.25953865e-01 -7.21843123e-01
7.60080218e-01 -2.42900610e-01 -1.25947475e-01 7.64000773e-01
8.84464145e-01 4.92547631e-01 -4.19117659e-01 -5.68991601e-01
7.39460528e-01 -4.10127640e-01 -7.37882257e-02 -8.16960037e-02
-8.44045222e-01 3.67633849e-01 -1.13750445e-02 3.17753315e-01
-4.66882348e-01 2.66896397e-01 1.18194796e-01 3.78684402e-02
-3.71287674e-01 5.48836589e-01 -4.66832042e-01 -2.74591625e-01
-1.43843085e-01 3.22037905e-01 1.05108821e+00 7.05817163e-01
5.06327391e-01 4.15279418e-01 7.17776269e-02 5.09046316e-01
6.69271290e-01 2.52575219e-01 9.90770817e-01 -6.91124499e-01
-7.26790249e-01 1.01229429e+00 -8.55033621e-02 -6.59877539e-01
-3.73501807e-01 -4.76744324e-01 -5.96596360e-01 4.47620511e-01
5.79283647e-02 -2.22987249e-01 -9.30686533e-01 1.38006711e+00
5.14139891e-01 1.50914073e-01 2.80988604e-01 4.79321778e-01
3.99988025e-01 9.13531542e-01 2.40674034e-01 -3.95108938e-01
1.35116744e+00 -6.66885972e-01 -4.37605947e-01 1.98140696e-01
5.03898978e-01 -3.52622539e-01 8.78165007e-01 1.56091675e-01
-8.65957916e-01 -3.16059649e-01 -1.26923764e+00 4.01450008e-01
-8.17900658e-01 -5.21945179e-01 6.33181155e-01 1.41037130e+00
-1.13243818e+00 6.54843509e-01 -8.22937608e-01 -7.59205639e-01
-2.50692934e-01 2.23966137e-01 2.03858197e-01 4.32465225e-01
-1.10434115e+00 8.12380016e-01 7.34930813e-01 -8.07613730e-02
-7.89908648e-01 -4.34935480e-01 -5.17640769e-01 7.07932785e-02
-1.48666352e-01 -1.09112930e+00 5.80274522e-01 -9.38630164e-01
-2.01093292e+00 8.27794790e-01 3.12683824e-03 -3.06004167e-01
3.64185959e-01 5.00220239e-01 -3.42501372e-01 -4.13495362e-01
-5.11550903e-01 4.73699003e-01 2.28810996e-01 -1.18318546e+00
-2.75406837e-01 -2.45104983e-01 -2.88610458e-01 2.31609836e-01
4.51166444e-02 1.61404848e-01 -1.11012325e-01 -7.04436481e-01
-2.57334858e-01 -1.09480393e+00 -4.82757390e-01 -3.43148828e-01
1.57573953e-01 8.75360072e-02 2.70695865e-01 -5.04642189e-01
1.42638373e+00 -2.02304888e+00 1.33007616e-01 2.46297464e-01
-5.26342750e-01 3.72925311e-01 -2.26359833e-02 8.19115877e-01
9.79135409e-02 4.09333020e-01 -8.41112435e-01 6.11797929e-01
-1.49540454e-01 4.01045203e-01 5.35967231e-01 2.43139237e-01
-8.12868401e-02 6.03279650e-01 -9.37828660e-01 -1.75374553e-01
-2.35939071e-01 4.21030074e-01 -4.89562750e-01 2.21895967e-02
-3.55369329e-01 3.83520991e-01 -3.50813746e-01 7.82035649e-01
7.07882404e-01 1.25621244e-01 4.31574851e-01 7.21930444e-01
-5.88483155e-01 -3.25624496e-01 -8.11573565e-01 1.32652724e+00
-4.65693511e-02 1.33889094e-01 -1.08611681e-01 -3.74674112e-01
1.24364436e+00 1.69616327e-01 1.17889404e-01 -4.57776994e-01
2.79444307e-01 3.63625556e-01 7.95099676e-01 -5.37645996e-01
4.68997538e-01 1.05048969e-01 1.07421353e-01 4.85648006e-01
-3.35152358e-01 -4.78863508e-01 4.66016412e-01 -1.82167426e-01
1.05093324e+00 5.40890992e-01 6.93495691e-01 -8.42522085e-01
7.80451357e-01 2.45247379e-01 7.78790772e-01 7.12543964e-01
-3.51481616e-01 1.67641640e-01 2.86643684e-01 -6.56276286e-01
-1.34020162e+00 -8.88021529e-01 -1.86208099e-01 7.14795887e-01
2.75880963e-01 -3.37486207e-01 -1.17850709e+00 3.39274555e-01
3.21314414e-03 1.02272153e+00 -5.20498514e-01 -4.24711108e-01
-6.42446697e-01 -1.67404819e+00 7.37340152e-01 -3.45634758e-01
5.79492867e-01 -1.34628880e+00 -1.26881003e+00 3.95292401e-01
5.31104684e-01 -2.73378134e-01 3.94843630e-02 1.69897288e-01
-9.44371939e-01 -7.39824355e-01 -1.69615492e-01 -6.34795427e-01
6.25444114e-01 -1.10163778e-01 1.03587866e+00 8.87115777e-01
-8.82990837e-01 3.92948054e-02 -4.07628983e-01 -7.33856499e-01
-1.04415321e+00 -2.76547581e-01 4.20192443e-02 -4.03645009e-01
3.40988368e-01 -6.93896234e-01 -4.46355641e-01 1.91295475e-01
-1.19585526e+00 -1.53796956e-01 1.19488701e-01 7.76767373e-01
3.34741861e-01 3.27623963e-01 2.71051109e-01 -2.05244601e-01
7.77938843e-01 -4.95587379e-01 -7.11469233e-01 5.62071741e-01
-3.52538109e-01 2.23557323e-01 7.09435761e-01 -1.57795548e-01
-9.05908227e-01 1.47734843e-02 1.59658045e-02 5.55451512e-01
-7.60299936e-02 4.05230939e-01 -1.56880140e-01 -1.79094285e-01
6.43571198e-01 7.44936764e-01 4.29630458e-01 -1.68161497e-01
-2.81859457e-01 7.82646239e-01 3.06484699e-01 -6.73634529e-01
2.01641768e-01 1.61523804e-01 3.37802947e-01 -1.07675719e+00
5.94961584e-01 4.46466923e-01 -4.71831143e-01 -4.37250316e-01
7.21083462e-01 -2.76406050e-01 -9.01859045e-01 9.79533315e-01
-8.33357513e-01 -2.67491817e-01 -9.41968411e-02 3.32806796e-01
-5.26940644e-01 3.23777199e-01 -5.58485627e-01 -1.00634062e+00
-4.31847721e-01 -9.65762675e-01 3.84108573e-01 7.43442416e-01
-1.22291230e-01 -8.78676355e-01 4.93987739e-01 -2.97596306e-01
5.00343204e-01 4.03026700e-01 1.22903562e+00 -2.01535046e-01
-3.93363148e-01 1.60310000e-01 7.48109996e-01 -2.05081776e-01
1.65529847e-01 1.04268610e+00 -3.59678566e-01 -2.80142073e-02
1.61974654e-01 5.50512195e-01 2.10411549e-01 2.76329368e-01
7.80478239e-01 1.38320372e-01 -3.67219388e-01 6.18269861e-01
1.67281294e+00 1.08959937e+00 1.14413524e+00 9.57530141e-01
1.14312500e-01 6.20516002e-01 3.91965896e-01 7.49734879e-01
1.16892019e-02 3.12432051e-01 4.06664640e-01 3.76222372e-01
2.11583480e-01 7.77835399e-02 4.44458872e-01 6.96653485e-01
-3.72277051e-01 -6.59873664e-01 -1.28708732e+00 1.10368453e-01
-1.39866376e+00 -8.85444582e-01 -3.89229715e-01 2.45088983e+00
8.14270020e-01 -2.96966553e-01 2.26102605e-01 -1.07887529e-01
1.03552186e+00 -7.50960290e-01 -5.68094969e-01 -1.06707728e+00
-5.42376637e-01 5.84422760e-02 5.80093682e-01 2.88954735e-01
-3.39310557e-01 7.51099110e-01 7.67548609e+00 3.89094740e-01
-9.94503319e-01 -3.59793365e-01 5.23409188e-01 1.18295848e-01
-2.33014449e-01 1.04614161e-01 -6.04777157e-01 9.94750977e-01
1.39444208e+00 -8.00548315e-01 6.75838947e-01 2.50503659e-01
4.59774464e-01 -4.96379435e-01 -4.10857916e-01 3.62665415e-01
-1.66910395e-01 -1.02737594e+00 3.05749744e-01 2.05913693e-01
6.95342600e-01 -3.54067773e-01 -2.71423519e-01 -8.85823891e-02
4.66124833e-01 -1.12003171e+00 6.91961050e-01 5.35609484e-01
9.11630690e-02 -8.66523147e-01 6.20348454e-01 4.99225199e-01
-7.52289951e-01 -2.07178101e-01 -5.57194293e-01 -3.07267219e-01
2.09363714e-01 2.01710433e-01 -7.88202047e-01 6.00163817e-01
7.71460593e-01 -4.52792421e-02 -6.57009125e-01 1.68335795e+00
9.02724192e-02 4.95254904e-01 -3.78813505e-01 -6.08153462e-01
9.15044323e-02 -7.83640325e-01 7.67032504e-01 1.17162871e+00
9.74045396e-01 4.02562261e-01 -4.98559028e-01 8.55894268e-01
3.82016331e-01 2.77171016e-01 -5.11577368e-01 -2.80548304e-01
6.66782260e-01 6.96645856e-01 -1.09467518e+00 -3.23206604e-01
1.04760095e-01 8.70696545e-01 -3.63663018e-01 1.81985840e-01
-8.94053042e-01 -4.54532266e-01 7.62981534e-01 -1.15606233e-01
1.13299796e-02 -5.03060639e-01 -3.81094992e-01 -3.74661595e-01
-5.02905309e-01 -9.85665858e-01 6.24339432e-02 -7.94327617e-01
-8.66475344e-01 4.99077946e-01 -1.37171641e-01 -8.17872047e-01
-1.04979321e-01 -6.33548260e-01 -6.78862214e-01 9.75602031e-01
-7.28790462e-01 -1.85801208e-01 -2.95828342e-01 -5.13863713e-02
3.03272069e-01 -3.24253172e-01 8.39314342e-01 -2.07121328e-01
-9.75400448e-01 2.67066866e-01 6.35961831e-01 -5.61460137e-01
2.82968078e-02 -9.22804534e-01 7.43309259e-01 7.37532079e-01
-7.47276068e-01 9.11884189e-01 1.05815971e+00 -1.20201921e+00
-1.76506662e+00 -7.44285703e-01 4.98429477e-01 5.50141558e-02
2.45638996e-01 -2.63578415e-01 -9.75618541e-01 2.62373775e-01
5.39809167e-01 -4.30976063e-01 8.21685672e-01 -8.50608587e-01
3.30119342e-01 4.94547635e-01 -1.54688120e+00 1.13627696e+00
1.25623035e+00 2.64536589e-01 -1.58172980e-01 1.24332272e-01
7.17284024e-01 -2.35466957e-01 -6.92567647e-01 2.23888189e-01
7.18061745e-01 -9.78833318e-01 9.13144112e-01 -4.08082277e-01
5.91683611e-02 -8.32566321e-01 2.83234045e-02 -1.23469079e+00
-5.22892833e-01 -8.39562058e-01 2.82727629e-01 8.99039805e-01
2.18848661e-01 -1.11192286e+00 4.23779517e-01 3.60832125e-01
-2.29975685e-01 -4.07916158e-01 -7.07716227e-01 -1.17418063e+00
1.28554285e-01 5.66788793e-01 1.17795122e+00 7.47945964e-01
-1.02802463e-01 -1.48598969e-01 3.43068354e-02 1.96968749e-01
4.75068361e-01 -4.77999955e-01 4.46298927e-01 -1.02980518e+00
-4.97234762e-01 -6.22220457e-01 -8.12182128e-01 6.32772734e-03
-2.69682080e-01 -5.68703592e-01 1.14419401e-01 -1.07164943e+00
1.10215679e-01 -2.76742131e-01 -6.29936680e-02 -4.03128192e-02
-3.21184993e-01 -1.13525495e-01 -7.89589882e-02 -1.19752802e-01
3.21359396e-01 5.29232323e-01 1.01018775e+00 2.69352168e-01
-3.59564424e-01 -5.00460625e-01 -2.58940935e-01 4.22171175e-01
1.08619440e+00 -5.37697852e-01 -1.75790176e-01 -9.23565328e-02
3.60784292e-01 -1.46828026e-01 -1.60110965e-01 -1.07460618e+00
-1.84258018e-02 -5.32960415e-01 3.76402438e-02 -1.63617611e-01
-1.22701868e-01 -7.35532284e-01 1.21180058e+00 1.35372376e+00
3.33764441e-02 7.75888741e-01 4.15896446e-01 4.35746580e-01
1.21496640e-01 -9.30845439e-01 7.36509025e-01 -4.30084050e-01
-6.77158058e-01 -3.40897799e-01 -1.21501088e+00 -7.14266241e-01
1.61248326e+00 -1.05783856e+00 -4.10135120e-01 2.78630525e-01
-2.97010034e-01 2.85145901e-02 1.27371383e+00 1.36025250e-01
2.82723039e-01 -7.98206091e-01 -7.34398782e-01 2.17697516e-01
-1.53853014e-01 -7.01723576e-01 1.00668333e-02 2.49660611e-01
-1.83428514e+00 1.66041210e-01 -9.44576740e-01 -2.78816134e-01
-1.39838541e+00 6.93410933e-01 8.24953616e-01 3.15408498e-01
-2.01343819e-01 8.43704522e-01 -8.65497887e-02 -6.47638202e-01
-3.11490119e-01 4.83054817e-02 -2.83700645e-01 -5.60943305e-01
5.71492136e-01 5.08312941e-01 -5.55937029e-02 -3.74360889e-01
-4.01457280e-01 6.04902089e-01 6.64716363e-01 -1.15004733e-01
1.28220463e+00 -1.94972813e-01 -8.96153212e-01 6.12787306e-01
5.44776797e-01 -1.12471685e-01 -6.63626075e-01 7.08614886e-01
1.08540483e-01 -5.20396709e-01 -5.48385024e-01 -9.02984023e-01
-4.40434933e-01 2.31904402e-01 9.09568727e-01 5.48784360e-02
1.12293458e+00 -7.78379440e-01 2.42827520e-01 3.25969100e-01
7.36688435e-01 -9.39947486e-01 -3.56732279e-01 7.68544614e-01
6.73528552e-01 -4.42500800e-01 -4.00327109e-02 -4.36473310e-01
-2.03150019e-01 1.27991366e+00 5.14641225e-01 -9.56228077e-02
3.89484584e-01 6.02527678e-01 -2.88473278e-01 2.61739343e-02
-9.77162063e-01 2.38181546e-01 -5.60184360e-01 6.66340709e-01
6.40262663e-01 1.27626553e-01 -1.27986813e+00 -3.04679777e-02
-3.02799851e-01 1.43103406e-01 1.02230263e+00 1.55333352e+00
-7.17321634e-01 -1.31351924e+00 -8.51918280e-01 -6.86018541e-02
-2.93675154e-01 -2.01876715e-01 -7.61328638e-01 4.57768500e-01
3.87527496e-01 6.65260732e-01 2.57430166e-01 -1.80275708e-01
-1.30653024e-01 1.73207983e-01 5.79195678e-01 -4.21062648e-01
-1.21973908e+00 -4.87860739e-01 -4.16308455e-02 3.43211144e-02
5.20469137e-02 -9.12522197e-01 -1.78073180e+00 -6.41721189e-01
1.03262410e-01 5.58909237e-01 1.12201738e+00 3.13837707e-01
6.76582456e-01 7.06662655e-01 3.07913959e-01 -4.54147488e-01
1.18852228e-01 -5.13039947e-01 -7.26440668e-01 -1.10920601e-01
-4.15981084e-01 -3.99238527e-01 -2.40869820e-01 2.03208610e-01] | [5.625804901123047, 4.136998653411865] |
87cf1212-7a0f-4fcb-b6a9-40fd8ed5d1d6 | ghrs-graph-based-hybrid-recommendation-system | 2111.11293 | null | https://arxiv.org/abs/2111.11293v2 | https://arxiv.org/pdf/2111.11293v2.pdf | GHRS: Graph-based Hybrid Recommendation System with Application to Movie Recommendation | Research about recommender systems emerges over the last decade and comprises valuable services to increase different companies' revenue. Several approaches exist in handling paper recommender systems. While most existing recommender systems rely either on a content-based approach or a collaborative approach, there are hybrid approaches that can improve recommendation accuracy using a combination of both approaches. Even though many algorithms are proposed using such methods, it is still necessary for further improvement. In this paper, we propose a recommender system method using a graph-based model associated with the similarity of users' ratings, in combination with users' demographic and location information. By utilizing the advantages of Autoencoder feature extraction, we extract new features based on all combined attributes. Using the new set of features for clustering users, our proposed approach (GHRS) has gained a significant improvement, which dominates other methods' performance in the cold-start problem. The experimental results on the MovieLens dataset show that the proposed algorithm outperforms many existing recommendation algorithms on recommendation accuracy. | ['Mohammad Hadi Valipour', 'Zahra Zamanzadeh Darban'] | 2021-11-06 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [-2.52546221e-01 -3.53621036e-01 -1.29933104e-01 -5.00726163e-01
-1.51813537e-01 -3.73159707e-01 5.46595335e-01 2.59404510e-01
-3.00816417e-01 4.13062930e-01 4.30165946e-01 -6.24991655e-02
-7.79653966e-01 -1.10812485e+00 3.65577787e-02 -7.25803971e-01
2.92500168e-01 4.73530084e-01 2.55233735e-01 -6.67327702e-01
9.64348376e-01 3.45109880e-01 -1.78033817e+00 3.34460109e-01
8.94212484e-01 8.98926616e-01 2.60365158e-01 3.13105971e-01
-5.48509598e-01 6.15163088e-01 -2.61964232e-01 -4.47494268e-01
4.16774988e-01 -2.33617112e-01 -4.87676024e-01 -2.50831902e-01
-4.92176674e-02 -2.66084373e-01 -3.48607033e-01 7.72929549e-01
5.98795116e-01 7.90678680e-01 6.81361794e-01 -9.67337847e-01
-1.03355241e+00 8.47427189e-01 -3.87824357e-01 1.74858242e-01
5.28948128e-01 -6.19404137e-01 1.11302888e+00 -7.96599030e-01
4.04114783e-01 8.76543939e-01 6.98490024e-01 2.23331377e-01
-6.05668724e-01 -5.66630363e-01 3.37733954e-01 4.92999375e-01
-1.33440697e+00 5.28424084e-02 8.87012064e-01 -4.59452391e-01
9.13363993e-01 2.78994411e-01 7.51060009e-01 6.48705065e-01
-7.27303699e-02 4.59285080e-01 1.13096106e+00 -4.65678990e-01
2.40599483e-01 6.90057158e-01 7.59814560e-01 2.93069243e-01
1.99372530e-01 -1.37534691e-02 -1.10710800e-01 -3.01860124e-01
3.30250055e-01 7.58067071e-01 -1.33077428e-01 -1.14897296e-01
-8.04494381e-01 1.04560757e+00 3.26827317e-01 8.04222405e-01
-4.62687641e-01 -6.35645032e-01 1.35178193e-01 3.67156982e-01
4.95655477e-01 5.39789736e-01 -1.94452107e-01 -1.13801174e-02
-1.01954818e+00 3.10525090e-01 9.89280939e-01 8.51416826e-01
4.98269588e-01 -5.14036901e-02 8.30373690e-02 8.73927355e-01
5.35969615e-01 2.93601900e-01 9.49652493e-01 -3.81345570e-01
-7.45930569e-03 9.19184506e-01 1.06708884e-01 -1.61850441e+00
-5.26475430e-01 -6.45227730e-01 -5.97304165e-01 -9.06583220e-02
5.75427711e-02 -1.55552462e-01 -5.22404492e-01 1.10438693e+00
3.84499818e-01 2.27004305e-01 2.14960381e-01 1.09132445e+00
1.31215334e+00 7.70062506e-01 -2.37826601e-01 -4.06749755e-01
9.34256375e-01 -1.10171473e+00 -8.62598479e-01 7.66004443e-01
5.06230950e-01 -9.25449252e-01 6.57688797e-01 8.62989306e-01
-6.11903131e-01 -7.16957688e-01 -1.04634738e+00 4.31058824e-01
-9.06847000e-01 1.88754201e-01 7.64645517e-01 9.47080076e-01
-1.05367219e+00 9.64140177e-01 -1.43553048e-01 -8.04274440e-01
-2.20170632e-01 6.15049362e-01 7.65165407e-03 2.14052666e-02
-1.09250057e+00 7.03224480e-01 1.71988070e-01 1.91735700e-02
-1.91571444e-01 -1.45598292e-01 -9.13744196e-02 1.45128429e-01
2.65264988e-01 -4.44123089e-01 7.15951920e-01 -1.07358372e+00
-1.56671989e+00 -1.68643534e-01 3.22419763e-01 -2.35167921e-01
-9.35159102e-02 -2.76944190e-01 -1.03250086e+00 -1.96932212e-01
-4.56758797e-01 -8.97223577e-02 4.45219517e-01 -1.02514112e+00
-8.64982545e-01 -4.41036135e-01 2.06851453e-01 4.11516607e-01
-1.02752936e+00 1.91601709e-01 -3.37175459e-01 -4.15101826e-01
5.56003861e-02 -1.00688791e+00 -4.86955673e-01 -8.80614281e-01
-7.31287077e-02 -5.47669113e-01 6.19418800e-01 -5.59099734e-01
1.64191103e+00 -1.77697849e+00 -6.38431869e-03 6.20088935e-01
8.89170840e-02 5.64903736e-01 -4.52050194e-02 9.33736145e-01
2.27814600e-01 -6.32934645e-02 3.02567869e-01 1.85053319e-01
-6.47157878e-02 -8.05542320e-02 -1.06215596e-01 8.61523747e-02
-6.83499813e-01 2.09144130e-01 -8.07217538e-01 -3.78512919e-01
2.92075664e-01 5.94334841e-01 -6.00343406e-01 3.05765897e-01
2.11521655e-01 -1.59772336e-02 -8.09689820e-01 3.33448857e-01
5.90735853e-01 6.20863633e-04 4.42252874e-01 -2.85730720e-01
-5.98286726e-02 -2.14019697e-02 -1.61909235e+00 1.43776727e+00
-3.22793990e-01 2.38932744e-02 -3.42183203e-01 -1.23792350e+00
1.43003774e+00 3.86230290e-01 7.99471319e-01 -3.34438294e-01
4.56945091e-01 9.97745469e-02 2.46284902e-01 -6.80417776e-01
9.05526698e-01 2.08717391e-01 4.16106552e-01 6.32484794e-01
-1.09166745e-03 6.55386269e-01 2.74267405e-01 4.39557523e-01
1.07149863e+00 2.99294144e-01 1.95157364e-01 -2.24925607e-01
7.71405995e-01 -4.10698317e-02 5.11158943e-01 7.48464942e-01
1.96421653e-01 3.55975211e-01 -2.97736049e-01 -5.59404910e-01
-6.96789622e-01 -4.53612059e-01 1.49330154e-01 1.00363493e+00
1.40315428e-01 -8.58571351e-01 -6.18702650e-01 -9.42035854e-01
8.86622518e-02 4.95016932e-01 -5.38228214e-01 -2.26265825e-02
-1.97686106e-01 -8.03841352e-01 1.06961928e-01 4.79228467e-01
1.90602258e-01 -1.00735116e+00 3.93249951e-02 4.39821035e-01
5.00115380e-02 -5.65244675e-01 -1.66813865e-01 -1.90530121e-01
-1.10185170e+00 -8.64747763e-01 -6.33050501e-01 -5.34510612e-01
5.85408628e-01 8.58040392e-01 7.62705803e-01 5.10481834e-01
1.76243082e-01 5.08603513e-01 -1.19829726e+00 -2.70007491e-01
-8.00417922e-03 3.86410505e-01 2.74244606e-01 3.65842730e-01
8.24175358e-01 -7.81049788e-01 -5.77419162e-01 2.81804860e-01
-5.59695899e-01 -4.37032342e-01 6.31330907e-01 5.59182048e-01
2.57984847e-01 3.95263672e-01 1.00413036e+00 -1.43497717e+00
1.02359760e+00 -9.09620404e-01 -2.43615523e-01 1.96085453e-01
-1.49631727e+00 -8.97419751e-02 1.05162144e+00 -3.39275122e-01
-1.38809168e+00 -2.39239976e-01 -3.24337363e-01 -1.48177505e-01
-3.66558582e-01 7.16220260e-01 1.45789711e-02 -1.59154311e-01
5.89994550e-01 1.54064223e-01 -2.33362734e-01 -8.84851456e-01
4.36570287e-01 1.23595071e+00 -1.02831930e-01 -3.20450038e-01
6.33504927e-01 1.42047420e-01 -2.58831501e-01 -5.22804916e-01
-6.02573872e-01 -1.01373565e+00 -6.12871349e-01 -4.09769028e-01
6.18201852e-01 -5.69169939e-01 -8.34837914e-01 -1.15840457e-01
-5.81394017e-01 6.75171077e-01 5.58927990e-02 1.02442718e+00
-8.45561400e-02 5.18240333e-01 -4.97736722e-01 -9.61981595e-01
-6.12098217e-01 -9.21884716e-01 2.27944076e-01 6.17681801e-01
8.45006332e-02 -9.11097527e-01 4.06473488e-01 5.07968009e-01
7.58639097e-01 -1.26727328e-01 6.75035119e-01 -1.31079626e+00
-2.66501546e-01 -4.94100928e-01 3.94247770e-02 2.86088914e-01
2.16928080e-01 6.34324178e-02 -5.59434354e-01 -2.85755575e-01
-5.25659919e-02 2.50554174e-01 6.27694666e-01 1.09104425e-01
9.88929987e-01 -2.95299679e-01 -5.49419105e-01 3.37009400e-01
1.78761244e+00 7.58051455e-01 5.52622020e-01 5.11251569e-01
8.13658416e-01 4.06901240e-01 8.56510878e-01 6.12640142e-01
3.91011178e-01 4.93825078e-01 2.25185543e-01 6.31704554e-02
2.80547202e-01 6.99094161e-02 8.33705366e-02 1.49444330e+00
-7.98915029e-01 -3.58677179e-01 -5.00619769e-01 2.01690912e-01
-2.18052649e+00 -1.18958044e+00 -3.52848351e-01 2.21763277e+00
1.27088934e-01 -1.22644687e-02 5.51873267e-01 3.84649068e-01
6.24991894e-01 -1.58786595e-01 -1.38982609e-01 -7.96091557e-01
1.30075932e-01 1.73506066e-01 3.72511208e-01 1.20769195e-01
-8.81133556e-01 6.84473693e-01 5.78914261e+00 5.73919952e-01
-9.27258253e-01 1.39458412e-02 -1.09635748e-01 -4.33076732e-03
-5.55095971e-01 1.19435221e-01 -8.66896033e-01 6.25432551e-01
1.08179355e+00 -1.85708717e-01 5.75338244e-01 9.80705202e-01
1.16928004e-01 4.74598296e-02 -4.46035355e-01 7.45575547e-01
6.40275061e-01 -1.13352680e+00 1.48433357e-01 1.72517687e-01
8.00938785e-01 -2.39473671e-01 -9.09349173e-02 3.41465235e-01
4.61556315e-01 -5.16005278e-01 -6.55826777e-02 1.06310105e+00
-3.78887296e-01 -1.08622575e+00 1.25903738e+00 4.28397954e-01
-9.72437680e-01 -3.03767174e-01 -6.16960824e-01 -2.95270592e-01
-2.39000708e-01 6.05042577e-01 -7.12181926e-01 1.11193681e+00
7.94702828e-01 1.02030468e+00 -6.19877338e-01 1.36340547e+00
2.14785621e-01 8.97491395e-01 -1.60399172e-02 -7.17975318e-01
2.45587882e-02 -9.54971671e-01 3.45469862e-01 1.12949145e+00
5.16182423e-01 4.62179542e-01 1.61043048e-01 2.22703725e-01
3.25898439e-01 1.05043519e+00 -6.95952773e-01 -6.23843307e-03
4.37374771e-01 1.81144071e+00 -9.00775433e-01 -4.59553003e-01
-7.20636308e-01 6.20598018e-01 1.01701200e-01 2.06387952e-01
-6.91439688e-01 -7.03056395e-01 3.16181034e-02 1.77030087e-01
4.00443494e-01 -2.72848867e-02 -6.89993985e-03 -9.43260550e-01
-2.96369553e-01 -9.60109651e-01 5.49329817e-01 -5.92255771e-01
-1.32880163e+00 8.20789397e-01 -2.37658605e-01 -1.60334635e+00
-8.21856856e-02 -3.30737561e-01 -4.56309766e-01 5.51149368e-01
-9.17668521e-01 -1.15555072e+00 -3.48793566e-01 7.03455210e-01
3.58433783e-01 -7.96356022e-01 9.99404132e-01 6.27882659e-01
-5.55472612e-01 5.25979280e-01 8.45690012e-01 -2.57647157e-01
8.92766416e-01 -1.21289098e+00 -3.23904634e-01 6.83595896e-01
4.07041967e-01 9.96650934e-01 4.49673980e-01 -5.25661647e-01
-1.35677993e+00 -6.11729860e-01 7.97460914e-01 -3.14600945e-01
3.17748666e-01 2.34127253e-01 -7.27943122e-01 4.62797374e-01
4.99253541e-01 -5.34490228e-01 1.33995414e+00 8.25248778e-01
-1.19427554e-01 -4.47840035e-01 -1.08940983e+00 3.44790787e-01
7.91336834e-01 3.46725644e-03 -8.17948937e-01 2.90305197e-01
2.59972036e-01 3.15261602e-01 -1.33886814e+00 1.26114994e-01
7.67278552e-01 -1.16243541e+00 7.71563888e-01 -7.40944386e-01
3.02905560e-01 -5.95957875e-01 -2.39029735e-01 -1.43639922e+00
-9.36375499e-01 -1.31485134e-01 1.80881955e-02 1.58503139e+00
2.64405608e-01 -5.81318974e-01 7.48166800e-01 4.03408349e-01
-2.78343856e-02 -7.23901212e-01 -1.24273784e-01 -5.29271424e-01
-2.69978642e-01 1.67697947e-02 8.22652817e-01 1.19082391e+00
3.83622527e-01 5.11559546e-01 -6.74373746e-01 -2.46934012e-01
2.91760981e-01 4.95027959e-01 7.09068239e-01 -1.72659981e+00
-4.68799710e-01 -3.36623609e-01 -3.62933606e-01 -5.53392291e-01
-1.55695513e-01 -1.08085835e+00 -2.82556683e-01 -1.74771237e+00
3.02777141e-01 -5.59315026e-01 -1.00776160e+00 7.30425790e-02
-1.73447311e-01 1.59934253e-01 1.21253081e-01 3.28509063e-01
-7.47202396e-01 1.69527307e-01 9.10523057e-01 7.23256320e-02
-4.17002589e-01 3.12994570e-01 -1.01837325e+00 7.91600406e-01
8.50943923e-01 -4.97880578e-01 -6.65893555e-01 -1.46198466e-01
4.07195419e-01 -1.23473525e-01 -5.15826225e-01 -1.18206310e+00
4.12028700e-01 -2.69624949e-01 4.67789114e-01 -8.01051199e-01
9.80986953e-02 -9.92266178e-01 3.79841805e-01 1.55878402e-02
-3.43619734e-01 1.99866846e-01 -4.10742372e-01 7.72788763e-01
-1.27984256e-01 -4.42784935e-01 2.87357301e-01 -1.49187297e-01
-6.88583970e-01 1.66958436e-01 -5.05514801e-01 -6.85126424e-01
8.22811306e-01 -3.37640256e-01 -4.23296094e-02 -4.30252552e-01
-8.42495382e-01 -1.50414094e-01 2.50933319e-01 7.34186172e-01
6.18333578e-01 -1.37730575e+00 -4.31192160e-01 -1.13779090e-01
8.70675817e-02 -9.82178330e-01 3.50942969e-01 7.04760909e-01
-1.68413162e-01 3.60500067e-01 -4.91070628e-01 1.22519009e-01
-1.26486123e+00 9.20672953e-01 -1.54058084e-01 -2.94563860e-01
-3.85933280e-01 5.81600428e-01 -4.16953027e-01 -4.95765060e-01
2.08037749e-01 2.09382296e-01 -1.44893479e+00 3.11504215e-01
5.06195188e-01 6.08989894e-01 2.13962957e-01 -7.58594096e-01
-1.67643577e-01 6.07144654e-01 -3.89810205e-01 2.22053211e-02
1.50097466e+00 -1.37564898e-01 6.99144229e-02 3.62892151e-01
7.68201709e-01 4.26242352e-01 -1.42886311e-01 -4.24848855e-01
1.69971827e-02 -6.29408300e-01 2.98892140e-01 -8.52970243e-01
-1.41061771e+00 5.48548460e-01 8.61594319e-01 6.17581069e-01
1.17752063e+00 -5.95653892e-01 8.16029072e-01 6.89557076e-01
3.91328901e-01 -1.42907989e+00 -2.78687328e-01 2.22669408e-01
4.65203226e-01 -1.12451112e+00 2.66670734e-01 -2.99328685e-01
-5.05509675e-01 1.30562949e+00 6.30532622e-01 -4.77994114e-01
1.24273026e+00 -1.85534935e-02 8.86133239e-02 4.15101014e-02
-6.96346998e-01 -4.27743673e-01 4.12136167e-01 6.24953866e-01
9.23495054e-01 1.01343870e-01 -1.32652974e+00 1.23212183e+00
-1.02194004e-01 1.60481438e-01 6.40162051e-01 7.71772742e-01
-7.71821380e-01 -1.57889903e+00 -1.82171836e-01 1.01342905e+00
-3.29479635e-01 -1.46331668e-01 -4.36876863e-01 5.56134641e-01
3.87517422e-01 1.49229133e+00 -3.48498791e-01 -1.16574168e+00
3.04813176e-01 -1.02505647e-01 1.97962284e-01 -5.07366002e-01
-1.08544612e+00 1.39445394e-01 1.73225477e-01 -3.23331922e-01
-7.36608624e-01 -6.71106815e-01 -9.34059203e-01 -3.98480773e-01
-9.30517793e-01 7.76799142e-01 8.78247976e-01 9.34929192e-01
3.92095625e-01 4.41196799e-01 9.66114402e-01 -4.12422448e-01
-4.21372801e-01 -1.11753511e+00 -8.99160743e-01 4.87431228e-01
-4.99726504e-01 -8.19663465e-01 -2.53665686e-01 -1.92156509e-01] | [10.06755256652832, 5.80987548828125] |
ab11203a-16e1-41da-b5bc-04f8eb4be72b | n2dnot-too-deep-clustering-via-clustering-the | 1908.05968 | null | https://arxiv.org/abs/1908.05968v6 | https://arxiv.org/pdf/1908.05968v6.pdf | N2D: (Not Too) Deep Clustering via Clustering the Local Manifold of an Autoencoded Embedding | Deep clustering has increasingly been demonstrating superiority over conventional shallow clustering algorithms. Deep clustering algorithms usually combine representation learning with deep neural networks to achieve this performance, typically optimizing a clustering and non-clustering loss. In such cases, an autoencoder is typically connected with a clustering network, and the final clustering is jointly learned by both the autoencoder and clustering network. Instead, we propose to learn an autoencoded embedding and then search this further for the underlying manifold. For simplicity, we then cluster this with a shallow clustering algorithm, rather than a deeper network. We study a number of local and global manifold learning methods on both the raw data and autoencoded embedding, concluding that UMAP in our framework is best able to find the most clusterable manifold in the embedding, suggesting local manifold learning on an autoencoded embedding is effective for discovering higher quality discovering clusters. We quantitatively show across a range of image and time-series datasets that our method has competitive performance against the latest deep clustering algorithms, including out-performing current state-of-the-art on several. We postulate that these results show a promising research direction for deep clustering. The code can be found at https://github.com/rymc/n2d | ['Robert J. Piechocki', 'Raul Santos-Rodriguez', 'Ryan McConville', 'Ian Craddock'] | 2019-08-16 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [-6.69938505e-01 -1.71119288e-01 1.02249056e-01 -3.65065366e-01
-7.95967340e-01 -5.53484499e-01 5.34590721e-01 8.39790553e-02
-3.61072928e-01 -1.69553041e-01 2.95319349e-01 5.09238690e-02
-3.94834965e-01 -6.29481673e-01 -6.37547553e-01 -1.10419607e+00
-5.36485136e-01 7.45789468e-01 -3.97257805e-01 2.72266418e-01
1.33930415e-01 2.08423167e-01 -1.46488011e+00 2.52864510e-01
5.53633034e-01 7.49563277e-01 2.52724111e-01 5.06604075e-01
8.60189497e-02 5.54118991e-01 -4.40365404e-01 -7.85559863e-02
2.89082944e-01 -3.64539444e-01 -8.62818956e-01 2.57504076e-01
4.33408320e-02 -3.24889310e-02 -5.63332975e-01 9.94789183e-01
4.03936237e-01 3.61144871e-01 7.57253051e-01 -1.31761706e+00
-8.03324342e-01 8.45045447e-01 -6.02864265e-01 1.00042270e-02
-2.09872633e-01 1.31391808e-01 1.22160184e+00 -8.18025291e-01
4.59566951e-01 1.14919412e+00 8.10798824e-01 3.50923210e-01
-1.35709393e+00 -5.39621472e-01 -1.17690876e-01 3.09591979e-01
-1.67065048e+00 -3.04657280e-01 8.42772067e-01 -6.44884944e-01
8.11403930e-01 -1.70703694e-01 5.90378821e-01 9.64249671e-01
-4.06781249e-02 7.35113084e-01 5.86769760e-01 -9.90039557e-02
5.31476676e-01 -1.99765116e-01 1.26256719e-01 6.17940784e-01
-1.89699024e-01 -4.62717190e-02 -6.55110851e-02 -5.70417419e-02
6.78903222e-01 4.74041879e-01 -7.92163685e-02 -6.02483749e-01
-1.35051072e+00 1.19723797e+00 9.75637674e-01 6.75865531e-01
-4.45844173e-01 4.50537890e-01 3.43288809e-01 3.53280514e-01
2.38161236e-01 6.86988294e-01 -3.90440494e-01 -1.64621532e-01
-1.22346818e+00 -2.31604785e-01 7.15343714e-01 4.88710374e-01
1.12617552e+00 -2.61319429e-01 3.45264047e-01 8.42389047e-01
2.99448729e-01 -5.20805269e-02 6.52625501e-01 -1.47713673e+00
-1.86970290e-02 8.18873882e-01 -3.37321520e-01 -1.21761298e+00
-6.01902366e-01 -4.33691055e-01 -1.29421520e+00 9.40583460e-03
1.91447392e-01 -2.62912661e-01 -7.45185196e-01 1.62752545e+00
2.32479259e-01 3.90452564e-01 -4.26006056e-02 1.00567210e+00
3.13024759e-01 9.12065148e-01 -2.02058226e-01 9.51790512e-02
1.07038283e+00 -8.90310764e-01 -5.58172822e-01 2.41360709e-01
8.22016895e-01 -5.30959010e-01 9.08958018e-01 4.00453895e-01
-9.88444090e-01 -5.65131724e-01 -9.61621761e-01 -3.77666056e-02
-5.51055551e-01 1.57403097e-01 7.16387749e-01 7.92687237e-02
-1.60186815e+00 9.25212562e-01 -1.50702536e+00 -7.04898775e-01
5.53300917e-01 6.29431009e-01 -3.46916467e-01 -1.26929462e-01
-9.34113443e-01 3.70354295e-01 3.82453382e-01 4.87596355e-03
-9.88801658e-01 -6.38180196e-01 -6.05928838e-01 1.85871333e-01
-2.39132829e-02 -6.52489007e-01 7.14735448e-01 -8.23363841e-01
-1.30238056e+00 7.20384061e-01 7.13874772e-02 -4.95452493e-01
-1.19801499e-02 -6.72695935e-02 -4.01311219e-01 5.31113744e-01
1.47763729e-01 1.14796102e+00 8.84022892e-01 -1.31753027e+00
-3.03464949e-01 -3.90775383e-01 -2.97722369e-01 1.69690400e-01
-7.34019041e-01 -1.85049534e-01 -6.53829813e-01 -3.36414427e-01
3.35637182e-01 -9.66139555e-01 -4.37851429e-01 -2.53256887e-01
-4.39410001e-01 -6.17711902e-01 1.13657498e+00 -4.83314097e-01
1.22374427e+00 -2.39280009e+00 4.69196886e-01 3.38439703e-01
4.90781873e-01 -1.71686873e-01 -1.92874342e-01 5.89149415e-01
-2.25745663e-01 2.85104513e-01 -3.90385985e-01 -6.35736942e-01
2.10419148e-01 9.08762366e-02 7.47232661e-02 8.65926981e-01
1.88030332e-01 1.05196702e+00 -7.71789551e-01 -3.45019519e-01
4.94137406e-01 8.54697585e-01 -6.53997242e-01 2.13834479e-01
-3.95089015e-03 3.62154961e-01 -5.35092838e-02 3.65207851e-01
3.59728336e-01 -8.74143362e-01 2.31231496e-01 -2.11997554e-01
-4.14576866e-02 -4.79234792e-02 -1.11846912e+00 2.17238331e+00
-1.86002508e-01 8.06077540e-01 2.89401501e-01 -1.50722897e+00
6.40078723e-01 3.32701653e-01 1.05021858e+00 -2.59479821e-01
2.25642979e-01 -9.50664431e-02 1.68363616e-01 -3.19799751e-01
1.54061139e-01 9.78270099e-02 -2.08675619e-02 7.84318566e-01
3.64375293e-01 2.66694397e-01 1.48568787e-02 4.43966359e-01
1.36834300e+00 -3.22673023e-01 -3.64544421e-01 -4.94143039e-01
1.04449660e-01 1.67016774e-01 2.38412783e-01 4.97869402e-01
-2.61261165e-01 6.80722296e-01 3.89658153e-01 -3.51001531e-01
-1.29117191e+00 -1.06898177e+00 -1.36158869e-01 1.04094291e+00
-1.12940118e-01 -5.75199485e-01 -1.05952609e+00 -3.73063058e-01
-9.60909426e-02 2.50618488e-01 -7.83084273e-01 -3.39269727e-01
-4.47806031e-01 -8.21649253e-01 5.25468886e-01 3.99546087e-01
3.17175895e-01 -9.84690189e-01 -2.48148888e-01 2.70591855e-01
-8.85176882e-02 -7.01051593e-01 -3.28392476e-01 5.06436050e-01
-1.05839479e+00 -1.00239944e+00 -4.02858347e-01 -1.09473515e+00
6.26331806e-01 2.73271054e-01 1.08697701e+00 3.81354600e-01
-3.79841596e-01 5.11343002e-01 -3.47675592e-01 3.91028553e-01
-2.14787021e-01 3.87914926e-01 1.53281763e-01 2.54262779e-02
8.04649472e-01 -9.46721554e-01 -9.66416419e-01 1.40853405e-01
-9.60775971e-01 -4.57527846e-01 5.65910161e-01 6.74675763e-01
4.52051759e-01 6.60002708e-01 3.64286751e-01 -3.64509851e-01
4.68316466e-01 -9.40845013e-01 -3.33076119e-01 -2.48763889e-01
-4.38417166e-01 4.56701256e-02 7.37602413e-01 -2.50181317e-01
-3.27272207e-01 2.72452921e-01 6.09465875e-02 -1.17726946e+00
-5.25351524e-01 4.72138673e-01 -8.67662486e-03 4.61537361e-01
5.93004286e-01 1.16085842e-01 3.15692574e-01 -5.27008414e-01
7.67260253e-01 5.59335232e-01 5.83205223e-01 -5.33982694e-01
8.48077834e-01 9.69410956e-01 -3.33657295e-01 -8.27737927e-01
-4.58494902e-01 -8.91744375e-01 -1.16756296e+00 -9.07643512e-02
1.37465966e+00 -1.05655110e+00 -8.25586557e-01 1.65949658e-01
-8.18518996e-01 -5.32597184e-01 -1.82126895e-01 5.31569064e-01
-5.21588147e-01 2.51575440e-01 -9.61525142e-01 -3.89531016e-01
-3.18260342e-02 -1.16312742e+00 1.21062315e+00 -1.74670398e-01
-2.73580283e-01 -1.49728394e+00 2.39115968e-01 1.47423863e-01
2.71256208e-01 3.00903052e-01 8.62357438e-01 -6.51056468e-01
-4.83552635e-01 -1.27252921e-01 -6.45020381e-02 2.77057618e-01
1.30265743e-01 2.68091738e-01 -8.68647516e-01 -6.70594454e-01
6.57277852e-02 -3.65825266e-01 1.14691627e+00 7.30354071e-01
1.58624244e+00 -2.55221009e-01 -4.50063944e-01 8.24164450e-01
1.60080814e+00 -8.73575807e-02 5.66029966e-01 1.90620571e-01
1.07840645e+00 7.09397495e-01 2.08491739e-02 4.23212767e-01
5.65278947e-01 1.89390317e-01 5.08889079e-01 -2.13003784e-01
8.28236714e-02 3.22695710e-02 4.06908244e-01 1.51844215e+00
1.66923672e-01 4.00966266e-03 -1.30706739e+00 8.51547658e-01
-2.01725292e+00 -1.05645728e+00 7.30347037e-02 1.76112032e+00
6.14862442e-01 -2.41647184e-01 3.41733813e-01 2.12398127e-01
9.43102658e-01 7.25934133e-02 -5.88339329e-01 4.72937264e-02
-2.94187665e-02 6.60921112e-02 4.92150001e-02 4.70838726e-01
-1.46710300e+00 9.55032170e-01 6.13134909e+00 6.36249721e-01
-1.16793871e+00 1.99232906e-01 7.37582684e-01 -2.56433874e-01
-4.39229272e-02 2.28350181e-02 -1.96807712e-01 4.83238995e-01
1.42807949e+00 2.90274888e-01 7.24502802e-01 8.04706037e-01
1.88536912e-01 4.80814219e-01 -1.42161798e+00 9.72193956e-01
-1.80915803e-01 -1.53079891e+00 -5.30760959e-02 5.13921559e-01
8.60611796e-01 4.53006059e-01 2.75809318e-01 3.29655588e-01
7.18331158e-01 -1.41010427e+00 2.46124730e-01 4.89075929e-01
3.51100653e-01 -9.27112937e-01 5.35188675e-01 1.46506861e-01
-1.18697429e+00 -1.73110992e-01 -5.47087133e-01 2.32733339e-02
-3.33793283e-01 6.98011160e-01 -5.77674925e-01 3.94334584e-01
1.16772878e+00 1.41157830e+00 -6.21828377e-01 8.36808562e-01
2.19201043e-01 7.61664033e-01 -3.58701736e-01 2.57178932e-01
5.95706403e-01 -6.13804400e-01 2.36374378e-01 1.24955845e+00
2.36154422e-01 1.12534203e-01 3.64734501e-01 1.16704798e+00
-2.11104318e-01 -2.01617196e-01 -5.07681012e-01 -3.21130186e-01
6.88317299e-01 1.52191496e+00 -9.86861289e-01 -3.73339742e-01
-2.90994197e-01 1.03690743e+00 5.53415477e-01 4.44413841e-01
-6.85984015e-01 -3.57295781e-01 9.39506829e-01 -1.32395938e-01
5.80346704e-01 -4.44410384e-01 -1.86953858e-01 -1.08498704e+00
-3.46168518e-01 -7.30272949e-01 6.31508172e-01 -5.93282700e-01
-1.63559425e+00 4.39704835e-01 -3.47458363e-01 -1.18266141e+00
-1.03485063e-01 -4.42441404e-01 -7.82314956e-01 2.61487722e-01
-9.16025996e-01 -8.76415908e-01 -1.59949884e-01 9.99757230e-01
3.34480494e-01 -1.89070925e-01 7.35649407e-01 4.45342958e-01
-8.18806529e-01 2.91391701e-01 6.76373899e-01 5.27472317e-01
5.79030216e-01 -1.61239314e+00 7.65680671e-02 7.03889549e-01
4.07949477e-01 9.51799393e-01 3.42721641e-01 -2.29864031e-01
-1.63698804e+00 -1.42037284e+00 2.18483031e-01 -7.11538434e-01
9.70544755e-01 -6.37108266e-01 -1.06973815e+00 6.59076929e-01
6.59515858e-01 -1.43957958e-01 8.37450325e-01 3.04430604e-01
-3.72238964e-01 3.40238810e-02 -7.41058707e-01 4.39132869e-01
7.58323669e-01 -7.43327141e-01 -4.26658511e-01 3.91943246e-01
8.73895526e-01 2.85996765e-01 -1.53747296e+00 7.20377564e-02
8.37565586e-02 -1.04200292e+00 9.76568580e-01 -3.84943306e-01
6.08550310e-01 -5.22545040e-01 -3.51522654e-01 -1.47730696e+00
-7.29529321e-01 -5.13522446e-01 -2.73297310e-01 1.06664205e+00
2.27665707e-01 -1.30442426e-01 9.03008819e-01 1.75616995e-01
-3.52263480e-01 -7.34328032e-01 -6.97023869e-01 -6.56717777e-01
5.34030259e-01 -2.56894529e-01 5.75613916e-01 1.60190165e+00
1.34120509e-01 4.74348933e-01 8.09460133e-02 3.93150210e-01
8.69943142e-01 1.75887793e-01 6.21802866e-01 -1.34390473e+00
-5.74109517e-02 -7.74307430e-01 -4.04333591e-01 -8.11939955e-01
3.41553211e-01 -1.27358496e+00 -3.89634669e-02 -1.57952309e+00
1.97750464e-01 -3.13497066e-01 -5.33711612e-01 4.00607198e-01
7.69416317e-02 2.31806770e-01 -2.73832176e-02 7.09814548e-01
-1.08603096e+00 6.89354360e-01 8.54537368e-01 -1.69595063e-01
-3.08998317e-01 -5.42046130e-01 -6.88342631e-01 5.47064364e-01
9.15736139e-01 -4.66520876e-01 -2.69203722e-01 -5.72557211e-01
5.88968489e-03 -2.46825293e-01 3.99360418e-01 -1.23287964e+00
7.01957583e-01 3.60155523e-01 5.92352629e-01 -6.93805635e-01
3.01823735e-01 -9.20992672e-01 9.88416448e-02 4.89857674e-01
-2.94515252e-01 1.62024617e-01 -7.23162070e-02 6.66595519e-01
-4.50890154e-01 3.74645025e-01 7.83124506e-01 -1.69976965e-01
-5.86159587e-01 6.11955881e-01 -3.86378139e-01 -1.29468381e-01
8.43827903e-01 -1.42876133e-01 -1.84256479e-01 -4.19646353e-01
-1.04117262e+00 4.99926805e-01 7.47923136e-01 3.06020141e-01
4.69003916e-01 -1.45671761e+00 -5.76403022e-01 3.98225933e-02
-1.94429204e-01 1.59599379e-01 1.14785217e-01 1.02728963e+00
-3.93752754e-01 2.37091213e-01 -3.09125874e-02 -1.24968815e+00
-5.49874663e-01 9.14310277e-01 4.32206601e-01 7.50590637e-02
-7.15735614e-01 7.86582172e-01 1.65543407e-01 -7.99959242e-01
4.59656090e-01 -8.26867074e-02 -4.01897952e-02 1.16764084e-01
1.84155479e-01 4.52399343e-01 -3.00357435e-02 -5.30019939e-01
-4.00839955e-01 6.25873625e-01 6.20818213e-02 -2.94067878e-02
1.69481838e+00 -2.82450974e-01 -6.34551644e-01 6.45760953e-01
1.89004731e+00 -6.06014609e-01 -1.35296750e+00 -2.19236352e-02
2.18333349e-01 8.60948116e-02 1.46159053e-01 -2.74409980e-01
-1.42090940e+00 1.15631568e+00 6.60719991e-01 5.04677415e-01
1.09153056e+00 4.40930188e-01 6.84079826e-01 6.00403905e-01
-1.00598969e-01 -1.21149027e+00 5.56851447e-01 2.62495428e-01
5.73553205e-01 -1.45160937e+00 -1.16656221e-01 3.12943250e-01
-6.00106776e-01 1.02128196e+00 4.55846071e-01 -7.15077639e-01
1.22602236e+00 8.30587521e-02 1.95720792e-01 -8.60041976e-01
-9.04400945e-01 -7.88396895e-02 1.48978949e-01 3.81376773e-01
4.65674609e-01 1.89930946e-01 5.04088640e-01 3.23469907e-01
-4.07731563e-01 -7.73484349e-01 3.02036136e-01 4.28116441e-01
-4.22475338e-01 -8.97704184e-01 -3.43691349e-01 5.04874945e-01
-2.47569039e-01 3.42483409e-02 -5.16390145e-01 8.31048489e-01
7.17169931e-03 1.09315062e+00 5.93879640e-01 -6.73301876e-01
-3.39111567e-01 1.33982956e-01 8.91986210e-03 -5.16103983e-01
-5.33253610e-01 4.17447090e-01 -6.44975066e-01 -8.78254533e-01
-7.06863999e-01 -8.21057737e-01 -1.41602027e+00 -5.62196434e-01
-1.23027101e-01 4.22133297e-01 6.85993433e-01 6.39897704e-01
6.92157984e-01 4.80968952e-01 8.78972530e-01 -1.02701700e+00
-1.92472935e-01 -8.58468473e-01 -6.71572685e-01 6.82233512e-01
4.32337940e-01 -3.99048448e-01 -6.57751083e-01 1.58785447e-01] | [9.106363296508789, 3.2656302452087402] |
812e8389-3224-41a9-bcdf-1b59f0f1fdc1 | predict-to-detect-prediction-guided-3d-object | 2306.08528 | null | https://arxiv.org/abs/2306.08528v1 | https://arxiv.org/pdf/2306.08528v1.pdf | Predict to Detect: Prediction-guided 3D Object Detection using Sequential Images | Recent camera-based 3D object detection methods have introduced sequential frames to improve the detection performance hoping that multiple frames would mitigate the large depth estimation error. Despite improved detection performance, prior works rely on naive fusion methods (e.g., concatenation) or are limited to static scenes (e.g., temporal stereo), neglecting the importance of the motion cue of objects. These approaches do not fully exploit the potential of sequential images and show limited performance improvements. To address this limitation, we propose a novel 3D object detection model, P2D (Predict to Detect), that integrates a prediction scheme into a detection framework to explicitly extract and leverage motion features. P2D predicts object information in the current frame using solely past frames to learn temporal motion features. We then introduce a novel temporal feature aggregation method that attentively exploits Bird's-Eye-View (BEV) features based on predicted object information, resulting in accurate 3D object detection. Experimental results demonstrate that P2D improves mAP and NDS by 3.0% and 3.7% compared to the sequential image-based baseline, illustrating that incorporating a prediction scheme can significantly improve detection accuracy. | ['Dongsuk Kum', 'In-Jae Lee', 'Youngseok Kim', 'Sanmin Kim'] | 2023-06-14 | null | null | null | null | ['3d-object-detection', 'depth-estimation'] | ['computer-vision', 'computer-vision'] | [ 3.43251824e-01 -2.55994290e-01 -1.57873794e-01 -2.14754492e-01
-5.12612522e-01 -4.04808134e-01 7.00445712e-01 -4.57213931e-02
-3.81824821e-01 1.81717232e-01 1.23105645e-01 -1.08493445e-02
1.72230810e-01 -5.53868413e-01 -5.88762224e-01 -5.40045559e-01
-1.34853914e-01 -2.68213600e-01 1.18870509e+00 5.03327101e-02
4.34267312e-01 5.95501006e-01 -1.90382218e+00 4.09287572e-01
4.62268561e-01 1.19234192e+00 6.84802175e-01 8.58381331e-01
2.18172714e-01 8.83600235e-01 -4.66061056e-01 2.57650584e-01
7.22381353e-01 -2.16376752e-01 -1.51094601e-01 3.21291924e-01
7.68590808e-01 -1.30465341e+00 -4.88224775e-01 6.63952231e-01
4.97545630e-01 1.14970230e-01 3.25146288e-01 -1.15690351e+00
-2.04316229e-02 -2.33509578e-02 -1.01634800e+00 5.17478585e-01
6.05172813e-01 4.40738976e-01 8.22814405e-01 -1.07725692e+00
5.62934518e-01 1.32523584e+00 6.09576344e-01 5.55934608e-01
-1.17555165e+00 -5.47229707e-01 3.43477130e-01 9.86772627e-02
-1.22933424e+00 -5.55821657e-01 5.96248984e-01 -5.32163382e-01
1.28639030e+00 6.20567501e-02 7.53230989e-01 7.46543586e-01
1.13456041e-01 1.05971003e+00 9.29604888e-01 -3.57593626e-01
-1.40137197e-02 -6.40328974e-02 6.67782575e-02 6.97572947e-01
3.97268564e-01 6.38490915e-01 -8.76013041e-01 1.08654320e-01
8.46719742e-01 8.48974586e-02 -2.18083516e-01 -5.50124168e-01
-1.34490728e+00 5.97629845e-01 5.70059955e-01 -1.62684172e-01
-3.48137051e-01 1.55818969e-01 6.57314360e-02 -8.54147449e-02
3.48835200e-01 1.64882332e-01 -2.08506659e-01 9.99939255e-03
-1.05496001e+00 3.65826070e-01 2.36588955e-01 9.66999114e-01
8.70914340e-01 2.25073267e-02 -3.38163882e-01 3.72451186e-01
4.65811670e-01 6.95824802e-01 7.50664622e-02 -1.35453451e+00
3.00024182e-01 7.00189829e-01 5.01543880e-01 -9.39584672e-01
-4.75949079e-01 -1.63689271e-01 -1.79665416e-01 5.32982409e-01
1.84738219e-01 -4.41959128e-03 -9.77045178e-01 1.44630671e+00
6.41122460e-01 3.66690278e-01 5.32818027e-02 1.26289272e+00
8.45665276e-01 4.30553257e-01 -2.35462517e-01 -1.55901596e-01
9.87495840e-01 -9.56986964e-01 -3.15867454e-01 -3.41670275e-01
5.96719265e-01 -7.23407686e-01 4.94774640e-01 3.19197953e-01
-1.11129844e+00 -6.88757837e-01 -1.20533168e+00 4.53888327e-02
3.44958976e-02 9.33242887e-02 5.58502734e-01 6.70584679e-01
-1.04180896e+00 3.20617944e-01 -1.09403574e+00 -5.89377940e-01
4.16419148e-01 4.43040252e-01 -2.37327605e-01 -2.70866483e-01
-6.34559810e-01 8.82447004e-01 6.17030382e-01 -2.24117905e-01
-1.10938716e+00 -6.91767216e-01 -8.55171025e-01 -2.40708262e-01
5.69137156e-01 -9.54695821e-01 1.35635757e+00 -5.64173877e-01
-1.21148968e+00 5.94544172e-01 -4.73586142e-01 -8.77673626e-01
5.36297679e-01 -6.14978254e-01 4.07488756e-02 4.92518693e-01
1.86835766e-01 1.16021359e+00 8.47616792e-01 -1.23481536e+00
-1.31826091e+00 -1.92142308e-01 3.14869404e-01 4.19479996e-01
-2.07184449e-01 2.27342593e-04 -7.00571716e-01 -3.18509877e-01
4.29904759e-01 -8.29879940e-01 -2.02577114e-01 3.44587952e-01
-7.89346322e-02 -2.40051020e-02 1.19341338e+00 -1.88421875e-01
1.14035940e+00 -2.13932085e+00 -2.13595361e-01 -2.82791466e-01
3.51160556e-01 3.37592781e-01 -1.15315750e-01 2.36667812e-01
5.47847450e-01 -3.65484446e-01 1.58364195e-02 -5.00568688e-01
-3.45330089e-01 1.80316240e-01 -1.54624462e-01 4.85789984e-01
3.77531588e-01 6.84320867e-01 -1.09224737e+00 -4.72675711e-01
8.40490878e-01 4.75074023e-01 -9.14655447e-01 1.71231538e-01
-2.47873500e-01 4.03854072e-01 -4.57664698e-01 8.13306093e-01
7.88794637e-01 -1.90761521e-01 -2.27796271e-01 -2.05153748e-01
-4.86303926e-01 3.19850266e-01 -1.15147030e+00 1.63052225e+00
-1.58436894e-02 7.47039735e-01 -8.72434750e-02 -3.79370600e-01
8.05418789e-01 -3.98563966e-02 6.53359890e-01 -5.72072864e-01
7.80103682e-03 2.37011507e-01 -6.12417832e-02 -3.10740501e-01
7.87771046e-01 2.58345991e-01 2.69247144e-01 5.00031784e-02
-1.04896300e-01 -1.38995409e-01 9.71598774e-02 3.19515139e-01
1.31179726e+00 3.90023887e-01 4.23474371e-01 1.34515360e-01
3.18811595e-01 9.50985029e-02 7.13360131e-01 1.10815001e+00
-5.36903918e-01 7.72643149e-01 -5.07662930e-02 -2.91829169e-01
-8.86975527e-01 -1.02020919e+00 -9.33236256e-02 7.34467208e-01
7.82881856e-01 -4.49845642e-01 -2.80899227e-01 -6.37706041e-01
2.68676937e-01 5.43465316e-01 -4.58440274e-01 -5.54519482e-02
-4.86603230e-01 -6.13334417e-01 1.28047615e-01 6.62324309e-01
6.70274436e-01 -3.60832751e-01 -1.51914358e+00 3.58331174e-01
-1.03715315e-01 -1.41659689e+00 -2.73215503e-01 1.05960995e-01
-9.43404913e-01 -9.27177310e-01 -5.00995815e-01 -2.73606420e-01
3.21519315e-01 1.28608263e+00 7.23726392e-01 -1.23086445e-01
-3.25957090e-01 7.39371955e-01 -5.81995010e-01 -4.17755783e-01
-1.13173388e-01 -4.32894409e-01 2.28208795e-01 -9.86132100e-02
6.25548184e-01 -3.50052893e-01 -1.05385149e+00 4.34043139e-01
-5.27392924e-01 3.44677985e-01 7.62625992e-01 8.44805360e-01
3.73129845e-01 -3.05172384e-01 1.30751267e-01 -1.25391200e-01
-2.60182440e-01 -3.02610427e-01 -7.68879533e-01 -2.11023748e-01
-5.74673533e-01 -1.78071946e-01 -2.08801299e-01 -4.10560071e-01
-1.13083827e+00 3.83656472e-01 2.82701015e-01 -8.73182237e-01
-1.61466584e-01 -6.11536689e-02 2.32897297e-01 -3.40647936e-01
5.62933385e-01 3.15902591e-01 -4.91616912e-02 -2.88712382e-01
9.73549038e-02 4.80935991e-01 5.34691393e-01 -1.93747163e-01
6.25198483e-01 9.38238382e-01 1.86245903e-01 -7.15397596e-01
-9.39722717e-01 -1.00740993e+00 -7.04754949e-01 -5.36286592e-01
8.66464674e-01 -1.41528857e+00 -6.28367484e-01 3.45194399e-01
-1.29066765e+00 -1.16125055e-01 -7.46065155e-02 9.36235189e-01
-5.65325201e-01 4.23151016e-01 -4.83934611e-01 -1.16502190e+00
-2.50266679e-02 -9.08335626e-01 1.46634769e+00 1.43455982e-01
-3.28947455e-02 -3.45636606e-01 -2.54060388e-01 2.45468214e-01
2.78862476e-01 3.21440160e-01 1.57226205e-01 -1.92273930e-01
-1.29810667e+00 -2.28721485e-01 -6.12547159e-01 1.21027648e-01
-2.25034244e-02 -1.66166842e-01 -1.11072600e+00 -2.12252185e-01
1.58621352e-02 4.44001965e-02 1.22930324e+00 5.77112198e-01
5.12053668e-01 1.04511678e-01 -5.81108153e-01 4.40283090e-01
1.35883534e+00 2.97016978e-01 3.19673091e-01 4.41425592e-01
6.09642088e-01 5.43757200e-01 9.38775659e-01 7.73696780e-01
4.54006493e-01 9.86518741e-01 7.29434431e-01 1.25601709e-01
-5.97145796e-01 -2.46690616e-01 6.58441007e-01 2.65992820e-01
-7.87041634e-02 -2.66984761e-01 -8.19244266e-01 5.99778652e-01
-1.85667288e+00 -1.03129172e+00 -2.47288376e-01 2.11445761e+00
1.18841611e-01 3.66796046e-01 1.62583381e-01 1.06419832e-01
6.74266636e-01 2.08132952e-01 -7.04010904e-01 2.68730521e-01
-2.80088276e-01 -4.50938523e-01 8.10456872e-01 3.82350296e-01
-1.15131843e+00 8.72586668e-01 5.94902277e+00 4.78189647e-01
-9.36973691e-01 1.90061349e-02 2.01366632e-03 -4.71764565e-01
1.19694673e-01 1.44079521e-01 -1.33110428e+00 2.47883692e-01
4.38031435e-01 1.55694829e-02 -8.55604187e-02 7.29013741e-01
5.01017570e-01 -6.72884226e-01 -1.08694327e+00 1.08787870e+00
2.10069269e-01 -1.15731406e+00 3.91447097e-02 2.00027063e-01
7.93177724e-01 3.10724229e-01 3.97881446e-03 7.49756694e-02
1.68766379e-01 -4.27523792e-01 9.58054364e-01 3.79697144e-01
2.40023598e-01 -3.79211605e-01 5.44434369e-01 4.95933235e-01
-1.39436269e+00 -3.27934414e-01 -3.09552252e-01 -4.52568829e-01
4.57299680e-01 5.22729397e-01 -1.11916947e+00 4.29159462e-01
1.03028464e+00 1.17680025e+00 -6.63960338e-01 1.26560104e+00
5.25584444e-02 3.87099057e-01 -6.36496425e-01 6.26448765e-02
4.24955606e-01 3.11313421e-01 1.10639083e+00 9.93932664e-01
5.32445312e-01 2.52161771e-01 3.90092999e-01 6.55231833e-01
4.72258240e-01 -3.19035679e-01 -6.54223561e-01 3.67795944e-01
4.92093712e-01 8.47437501e-01 -6.41455233e-01 -3.16387564e-01
-9.20734346e-01 1.00621045e+00 -1.07482046e-01 1.77655876e-01
-7.35359371e-01 -6.30095899e-02 7.24919140e-01 3.76529515e-01
9.24043536e-01 -4.54305619e-01 -5.16964570e-02 -1.05492377e+00
1.26641899e-01 -2.92452574e-01 2.83335835e-01 -8.33607018e-01
-9.71380770e-01 2.85354733e-01 1.07780762e-01 -1.64012527e+00
-1.95973217e-01 -5.82613230e-01 -2.42211163e-01 4.65851724e-01
-1.69336188e+00 -1.16559958e+00 -5.52068412e-01 5.23410618e-01
6.73944831e-01 1.25650272e-01 1.76925510e-01 2.02824235e-01
-2.61221886e-01 2.24175006e-01 -1.20516755e-01 -2.09073097e-01
7.37608850e-01 -9.52180743e-01 2.95627624e-01 1.19884288e+00
-1.21199593e-01 3.52790534e-01 5.28361380e-01 -6.76786661e-01
-1.52273858e+00 -1.06583297e+00 6.00492179e-01 -7.42262840e-01
3.59405249e-01 -2.87668854e-01 -6.08381867e-01 4.19979572e-01
-1.85501322e-01 2.89444417e-01 2.89326429e-01 -4.00265753e-01
-2.14339063e-01 -5.74275814e-02 -9.46179509e-01 5.30917585e-01
1.37175369e+00 -2.52024859e-01 -5.32155037e-01 -1.35249659e-01
9.13524866e-01 -5.65663993e-01 -4.87436324e-01 8.59778345e-01
7.64074743e-01 -1.48016143e+00 1.29201400e+00 3.65843847e-02
9.93875489e-02 -8.82394373e-01 -5.84974647e-01 -4.55129772e-01
-2.08483741e-01 -4.39089686e-01 -6.97230756e-01 8.93943548e-01
-1.02436095e-01 -3.91315550e-01 7.91666090e-01 4.29351449e-01
-2.26428658e-01 -4.43693757e-01 -9.34487402e-01 -9.43033397e-01
-6.55205488e-01 -5.79508126e-01 2.87839442e-01 4.72408056e-01
-3.09587419e-01 1.72498912e-01 -6.03232145e-01 5.05845964e-01
7.33981311e-01 3.58605236e-01 1.14613318e+00 -1.09681010e+00
-4.04340357e-01 -3.80651444e-01 -9.33298290e-01 -1.86888552e+00
-4.57173020e-01 -4.10148412e-01 2.35352620e-01 -1.43280876e+00
2.22829431e-01 -1.86593220e-01 -2.79115826e-01 2.45867640e-01
-4.35225904e-01 3.88676316e-01 4.14957345e-01 3.84806037e-01
-9.31538999e-01 3.90778095e-01 1.08463299e+00 9.41872224e-02
-3.61092657e-01 -1.56106159e-01 -2.10050836e-01 6.97813809e-01
4.26432520e-01 -2.89522678e-01 -2.96364844e-01 -5.43367922e-01
-2.93364018e-01 1.87680483e-01 8.76607835e-01 -1.35146725e+00
4.83180523e-01 4.59710918e-02 6.62151754e-01 -1.18726838e+00
7.28058994e-01 -7.30797529e-01 -2.09903494e-01 6.49807751e-01
4.37049083e-02 -2.59687483e-01 4.34403539e-01 1.07862282e+00
-1.68313071e-01 1.02211848e-01 6.83780968e-01 -7.13012218e-02
-1.38051867e+00 2.41592050e-01 -3.79378825e-01 -4.46279794e-01
1.26588416e+00 -7.09432364e-01 -2.32687339e-01 -1.57331750e-01
-3.09729546e-01 3.36257309e-01 5.94328165e-01 4.13388997e-01
7.94257641e-01 -1.14605319e+00 -6.72709346e-01 2.39195466e-01
3.76612991e-01 2.95251422e-02 2.60512173e-01 1.10697210e+00
-2.93069810e-01 5.54521561e-01 -5.22434339e-02 -1.41812348e+00
-1.37589300e+00 6.22501373e-01 1.88272074e-01 1.99312791e-01
-8.08083117e-01 9.01033044e-01 5.70284545e-01 1.85440451e-01
1.72100127e-01 -3.69752645e-01 2.34313663e-02 -9.90147740e-02
7.37845659e-01 4.12095368e-01 -2.39369661e-01 -5.89805841e-01
-4.62460250e-01 7.47067273e-01 -2.28399917e-01 -6.13604523e-02
1.13242519e+00 -6.48245335e-01 4.67062652e-01 3.60989898e-01
8.98717701e-01 -1.13580473e-01 -2.04004097e+00 -5.16403496e-01
-1.48666173e-01 -9.74343777e-01 3.82326037e-01 -4.99781400e-01
-6.54247522e-01 8.18331897e-01 8.40329111e-01 -1.76083788e-01
1.27671134e+00 6.93052113e-02 7.15993047e-01 3.95825922e-01
7.55898833e-01 -6.37338758e-01 3.32040429e-01 4.43610609e-01
4.14010882e-01 -1.60680902e+00 1.80871829e-01 -5.07663608e-01
-3.76162410e-01 1.11691391e+00 8.20944250e-01 1.07844725e-01
2.89197475e-01 2.19084591e-01 -2.13367403e-01 -3.31800506e-02
-8.47462952e-01 -8.26509595e-01 2.42637828e-01 6.43314898e-01
-1.10217854e-01 -3.28999937e-01 5.28917834e-02 1.28891200e-01
3.94311428e-01 -6.90774545e-02 3.85002285e-01 1.18923903e+00
-8.44409049e-01 -6.85179412e-01 -4.33843434e-01 3.97305608e-01
-9.98803228e-02 -1.38947014e-02 -1.35506555e-01 7.80010343e-01
3.09622109e-01 1.13123810e+00 2.14982927e-01 -6.03629410e-01
2.53389627e-01 -3.18238825e-01 6.77139103e-01 -6.23050988e-01
-1.15756460e-01 3.55333954e-01 -1.47166014e-01 -9.55284238e-01
-9.07416582e-01 -9.48740661e-01 -9.82295334e-01 -9.56616923e-02
-6.79830372e-01 -4.94033664e-01 4.48775738e-01 6.33425772e-01
5.48290730e-01 3.21010083e-01 5.12526989e-01 -1.32576871e+00
-1.81516260e-01 -6.71928346e-01 -3.39822352e-01 7.14163259e-02
7.60875225e-01 -1.03740168e+00 -4.41595823e-01 5.22006042e-02] | [7.982998371124268, -2.1086273193359375] |
aa65a3ec-87a6-4af1-a474-2cc186e388f6 | vatlm-visual-audio-text-pre-training-with | 2211.11275 | null | https://arxiv.org/abs/2211.11275v2 | https://arxiv.org/pdf/2211.11275v2.pdf | VATLM: Visual-Audio-Text Pre-Training with Unified Masked Prediction for Speech Representation Learning | Although speech is a simple and effective way for humans to communicate with the outside world, a more realistic speech interaction contains multimodal information, e.g., vision, text. How to design a unified framework to integrate different modal information and leverage different resources (e.g., visual-audio pairs, audio-text pairs, unlabeled speech, and unlabeled text) to facilitate speech representation learning was not well explored. In this paper, we propose a unified cross-modal representation learning framework VATLM (Visual-Audio-Text Language Model). The proposed VATLM employs a unified backbone network to model the modality-independent information and utilizes three simple modality-dependent modules to preprocess visual, speech, and text inputs. In order to integrate these three modalities into one shared semantic space, VATLM is optimized with a masked prediction task of unified tokens, given by our proposed unified tokenizer. We evaluate the pre-trained VATLM on audio-visual related downstream tasks, including audio-visual speech recognition (AVSR), visual speech recognition (VSR) tasks. Results show that the proposed VATLM outperforms previous the state-of-the-art models, such as audio-visual pre-trained AV-HuBERT model, and analysis also demonstrates that VATLM is capable of aligning different modalities into the same space. To facilitate future research, we release the code and pre-trained models at https://aka.ms/vatlm. | ['Furu Wei', 'Jinyu Li', 'Daxin Jiang', 'LiRong Dai', 'Jie Zhang', 'Binxing Jiao', 'Shujie Liu', 'Ziqiang Zhang', 'Long Zhou', 'Qiushi Zhu'] | 2022-11-21 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 2.00127661e-01 -1.08887956e-01 -2.68366843e-01 -4.51298475e-01
-1.01224124e+00 -5.14224112e-01 8.65064561e-01 -1.48113117e-01
-3.17935497e-01 2.12014526e-01 6.01221979e-01 -3.70647460e-01
4.54268664e-01 -2.07531795e-01 -6.42115831e-01 -4.38406974e-01
4.92761701e-01 3.17492872e-01 5.26022725e-02 -2.27360707e-02
2.27462556e-02 1.45380810e-01 -1.61633921e+00 9.78496671e-01
5.02959371e-01 9.76681173e-01 6.29979789e-01 7.54188180e-01
-6.54044867e-01 8.04360330e-01 -4.40517575e-01 -1.64823502e-01
-1.69812202e-01 -3.59571040e-01 -9.34581697e-01 8.02205279e-02
1.98027223e-01 -2.59556800e-01 -4.76585269e-01 6.64140284e-01
6.53994143e-01 3.58755320e-01 8.61813962e-01 -1.50049436e+00
-1.00801730e+00 7.62847602e-01 -5.16126454e-01 -1.46869466e-01
5.29401422e-01 3.37916613e-01 1.03452075e+00 -1.16620159e+00
5.40671587e-01 1.68298507e+00 6.09386042e-02 8.05938125e-01
-1.03170073e+00 -5.98281205e-01 3.42041373e-01 3.91495854e-01
-1.43899345e+00 -8.30169439e-01 6.09225035e-01 -4.02677417e-01
1.09921682e+00 3.17456603e-01 3.35419476e-01 1.70499671e+00
-1.01980835e-01 1.32013392e+00 8.99905562e-01 -4.37421441e-01
-1.05327368e-01 3.91434222e-01 3.40567231e-02 3.91712010e-01
-4.51450348e-01 -3.72555405e-02 -9.77922022e-01 1.45732835e-01
5.30369639e-01 7.16012269e-02 -2.22416997e-01 -3.28883052e-01
-1.35982215e+00 5.58766246e-01 3.29042047e-01 3.71330380e-01
-5.52205108e-02 1.12505564e-02 7.49533355e-01 2.89831489e-01
2.01175734e-01 -1.28332883e-01 -4.34551686e-01 -1.29966006e-01
-8.09019506e-01 -3.73058468e-01 3.83539677e-01 9.27527010e-01
5.78299046e-01 2.45539799e-01 -5.49087882e-01 1.22531879e+00
1.04620135e+00 7.04238176e-01 7.47106194e-01 -5.77237010e-01
7.34207094e-01 4.13014084e-01 -4.42812771e-01 -2.88946778e-01
-1.34009644e-01 -5.07694930e-02 -7.26217031e-01 4.36382778e-02
-1.36844488e-02 5.18567786e-02 -1.34169292e+00 1.77177954e+00
3.19720022e-02 1.68790087e-01 6.01812005e-01 9.77862477e-01
1.66061497e+00 1.02487457e+00 4.39482272e-01 7.05147535e-02
1.50716901e+00 -1.34679210e+00 -7.65758932e-01 -3.76722991e-01
4.50948745e-01 -9.04808521e-01 1.26511097e+00 4.94665280e-02
-9.95510161e-01 -6.48376048e-01 -8.40616047e-01 -4.68772769e-01
-5.28568745e-01 5.37890136e-01 2.12166816e-01 4.28014874e-01
-1.15389264e+00 -2.28259653e-01 -6.85056806e-01 -6.92310691e-01
2.47255683e-01 -1.47820711e-02 -7.14968264e-01 -6.91939220e-02
-1.18440700e+00 8.48965764e-01 3.64219457e-01 7.79381171e-02
-1.51597404e+00 -3.54477644e-01 -1.25407350e+00 1.17345855e-01
3.12518895e-01 -9.65879023e-01 1.41848481e+00 -1.17286265e+00
-1.75052130e+00 9.44455385e-01 -6.32526934e-01 -2.15098009e-01
1.11851908e-01 1.06233902e-01 -5.19201696e-01 1.93272889e-01
-1.10687232e-02 9.93800223e-01 1.15709281e+00 -1.50023150e+00
-3.72173846e-01 -2.17142493e-01 -1.58425808e-01 5.12638688e-01
-3.03115934e-01 2.56248385e-01 -7.37121940e-01 -6.39911890e-01
-6.73193261e-02 -5.87900937e-01 2.16511741e-01 -4.58701551e-02
-4.98871565e-01 -2.51602173e-01 7.16690838e-01 -7.69374609e-01
9.22304690e-01 -2.53524637e+00 2.76868701e-01 2.59454716e-02
6.82540983e-02 3.55871469e-01 -6.46832764e-01 5.99640548e-01
-1.12357691e-01 1.10514536e-01 -3.15972082e-02 -1.08081651e+00
1.99075118e-01 2.44425505e-01 -3.19655359e-01 1.50460497e-01
6.88074678e-02 9.86108184e-01 -6.34527206e-01 -6.94946408e-01
6.29680097e-01 8.04982364e-01 -1.67370290e-01 4.17331457e-01
-1.46601349e-01 6.07506692e-01 -2.06831366e-01 9.04179811e-01
4.73052680e-01 -2.29206353e-01 2.16400810e-03 -3.71583492e-01
5.15736379e-02 4.55066442e-01 -9.53218281e-01 2.11800933e+00
-7.42812335e-01 7.87818432e-01 2.37974972e-01 -8.21065187e-01
7.61636257e-01 7.82118618e-01 1.98390216e-01 -8.35310638e-01
3.22069943e-01 -3.22054396e-03 -3.52619678e-01 -7.13048875e-01
3.96783382e-01 -5.42826839e-02 -2.32778955e-02 3.37919831e-01
6.87381923e-01 2.12084547e-01 -2.65776038e-01 4.78655577e-01
7.63715506e-01 -1.09158270e-01 2.26750240e-01 5.77531338e-01
6.27844572e-01 -2.80989051e-01 1.49520501e-01 5.37395597e-01
-3.19060653e-01 8.03630054e-01 1.16752110e-01 2.81425238e-01
-6.76925182e-01 -1.46387446e+00 2.09875926e-02 1.53608835e+00
1.22658037e-01 -6.77426755e-01 -3.83802950e-01 -4.99760479e-01
-1.02489419e-01 6.02018774e-01 -4.16791201e-01 -1.39657944e-01
1.08136445e-01 -4.02370915e-02 7.51328468e-01 5.89960635e-01
4.02271092e-01 -1.19059551e+00 5.52196391e-02 -2.31690332e-01
-4.85900074e-01 -1.22516191e+00 -5.49950004e-01 -1.13384284e-01
-4.00296032e-01 -5.98822236e-01 -9.00833964e-01 -9.07362223e-01
3.53837609e-01 6.05694175e-01 7.68768489e-01 -3.55465263e-01
-4.47053090e-02 8.99660647e-01 -5.96831858e-01 -2.28883192e-01
-4.51096028e-01 -9.75067094e-02 -1.74414009e-01 3.76969188e-01
2.47737795e-01 -3.39171350e-01 -3.09127569e-01 1.35481477e-01
-8.40468526e-01 3.57040823e-01 7.58619368e-01 8.73933673e-01
5.32883823e-01 -6.16246462e-01 4.33591992e-01 -2.16376528e-01
5.24692774e-01 -5.64878702e-01 1.86476354e-02 4.80989814e-01
-1.27856031e-01 4.20162780e-03 2.77341545e-01 -6.41300142e-01
-1.26451492e+00 1.66819602e-01 -2.46025011e-01 -9.44856644e-01
-5.01580417e-01 7.08142877e-01 -5.33783555e-01 1.69286206e-01
2.22089708e-01 5.08133054e-01 1.57234341e-01 -6.03816390e-01
8.65998209e-01 1.19525838e+00 6.44248843e-01 -4.00225103e-01
4.23925221e-01 2.83243448e-01 -4.48240489e-01 -1.00833941e+00
-3.00439894e-01 -7.39417136e-01 -2.69322902e-01 -2.64700919e-01
1.09969878e+00 -1.43761063e+00 -6.58527434e-01 3.29165667e-01
-1.44355154e+00 -2.70761311e-01 -7.39035681e-02 7.28510618e-01
-4.91947800e-01 4.51065212e-01 -5.40724754e-01 -1.02170455e+00
-2.69748509e-01 -1.34934366e+00 1.39330697e+00 8.83543342e-02
-1.60375163e-01 -8.45856607e-01 -4.43742983e-02 8.75287831e-01
2.63721436e-01 -4.47971344e-01 5.00633299e-01 -9.08300817e-01
-4.98166978e-01 2.74894893e-01 -4.11814988e-01 4.58064884e-01
1.27811685e-01 1.31983280e-01 -1.45189619e+00 -1.99110955e-01
-4.47724283e-01 -5.27494311e-01 1.23811102e+00 3.81320506e-01
9.67721462e-01 -1.23212049e-02 -3.40122700e-01 5.58148265e-01
8.86311471e-01 2.23447338e-01 5.96597254e-01 1.17129438e-01
9.73388970e-01 6.42540634e-01 5.02056599e-01 3.80839497e-01
8.08754146e-01 6.47213697e-01 4.33623403e-01 -2.27373913e-01
-4.11021471e-01 -3.70019585e-01 9.08554733e-01 1.02326047e+00
3.07821631e-01 -4.94412601e-01 -7.84439504e-01 6.54164553e-01
-1.96741605e+00 -9.52324271e-01 2.06766516e-01 1.88557458e+00
6.44439459e-01 -3.26249212e-01 9.66913998e-02 -2.08022296e-01
6.63162708e-01 2.95346320e-01 -3.25840116e-01 -3.98801357e-01
-2.38599822e-01 7.64680952e-02 -2.38234624e-02 5.77709913e-01
-1.09036565e+00 1.16354620e+00 5.86768627e+00 9.14113283e-01
-1.34810960e+00 5.13585746e-01 1.95594281e-01 -2.16190532e-01
-5.42550683e-01 -4.45458889e-02 -5.22596180e-01 3.40732634e-01
9.82332766e-01 3.22388597e-02 4.78594482e-01 6.54015481e-01
3.85830998e-01 2.61333317e-01 -1.15962541e+00 1.53381169e+00
3.90559644e-01 -1.16949439e+00 5.64742506e-01 -2.71470487e-01
1.50179505e-01 2.76370347e-01 3.16345632e-01 5.85905790e-01
1.70325533e-01 -1.18576193e+00 8.32201242e-01 4.73510593e-01
9.24756467e-01 -2.86298633e-01 4.41245496e-01 1.60039857e-01
-1.57399726e+00 -6.62527326e-03 -1.31522333e-02 4.11896139e-01
3.64301622e-01 2.99379183e-03 -9.18522596e-01 7.48922288e-01
7.34039307e-01 8.68391871e-01 -4.61296707e-01 8.38194430e-01
-2.02891111e-01 4.51520860e-01 -3.82990427e-02 2.48826176e-01
1.60864875e-01 2.17117891e-01 5.22390008e-01 1.34914732e+00
3.39341283e-01 -3.20600867e-01 1.84696645e-01 8.28305185e-01
-7.25541562e-02 3.99438232e-01 -6.94836378e-01 -3.99634570e-01
6.28193557e-01 1.25670457e+00 -2.13456139e-01 -4.16757613e-01
-8.26083958e-01 1.09058547e+00 1.31338567e-01 6.81955278e-01
-7.07306027e-01 -1.56621441e-01 7.50173688e-01 -3.49363297e-01
1.30777091e-01 -1.18886240e-01 -1.80159464e-01 -1.39473891e+00
-1.32633438e-02 -8.53305101e-01 4.63669240e-01 -1.35351038e+00
-1.34292734e+00 7.44736731e-01 -7.65211657e-02 -1.44052112e+00
-3.30640733e-01 -7.17109919e-01 -5.59158742e-01 1.13417542e+00
-1.52602613e+00 -1.69718909e+00 -1.75835907e-01 1.11814749e+00
9.12780941e-01 -6.18113935e-01 6.81761861e-01 3.53150368e-01
-5.48610449e-01 7.62649775e-01 -2.69634247e-01 2.57579625e-01
1.16851711e+00 -8.09006631e-01 1.39536709e-01 7.09187925e-01
4.04110670e-01 7.20670521e-01 3.46360058e-01 -4.89490360e-01
-1.49148273e+00 -1.09572399e+00 6.74850285e-01 -3.82666171e-01
6.03214800e-01 -4.42550153e-01 -8.89800310e-01 8.76339972e-01
6.88046813e-01 -7.59078041e-02 9.35883999e-01 4.79428284e-02
-7.82329202e-01 -4.92105372e-02 -8.30154657e-01 5.84949076e-01
7.49301314e-01 -1.25328279e+00 -7.38760948e-01 9.33338981e-03
1.12162673e+00 -2.38988593e-01 -6.44683421e-01 2.23794460e-01
5.52472174e-01 -5.04487991e-01 1.08064795e+00 -5.77127695e-01
2.13340715e-01 -4.55274940e-01 -5.77563405e-01 -1.33642197e+00
6.66776597e-02 -3.80420506e-01 -1.60566214e-02 1.57750058e+00
5.20654857e-01 -6.27870440e-01 1.66613355e-01 1.03768587e-01
-4.44953233e-01 -3.19666356e-01 -1.11354399e+00 -5.91332257e-01
-2.73390859e-01 -7.52712131e-01 3.92913669e-01 9.99390841e-01
2.22085729e-01 7.66335666e-01 -4.98317808e-01 3.48680794e-01
2.38845319e-01 -1.71056017e-01 8.03332388e-01 -9.39577997e-01
-3.64947617e-01 -4.41454142e-01 -3.32361311e-01 -1.21082771e+00
4.88841355e-01 -1.28431284e+00 5.94160110e-02 -1.89461446e+00
3.01228553e-01 -1.04463026e-01 -5.06865859e-01 8.47459137e-01
1.47058785e-01 -6.60663098e-02 5.24203002e-01 1.82431310e-01
-7.80171812e-01 9.39227819e-01 1.19925928e+00 -5.55164814e-01
-2.71931261e-01 -2.22698048e-01 -6.33405149e-01 4.32331651e-01
5.60396314e-01 -9.53973457e-02 -6.41514242e-01 -6.59785330e-01
-2.62427539e-01 3.04194450e-01 5.11971653e-01 -6.80409253e-01
3.74332786e-01 -1.49820819e-01 2.57977426e-01 -7.50384390e-01
9.45033729e-01 -6.75341189e-01 -8.34494606e-02 -2.01421604e-01
-5.75439751e-01 -1.94438770e-01 2.88171321e-01 3.70345861e-01
-6.52102351e-01 -2.54675690e-02 4.40556586e-01 1.94081699e-03
-9.24246311e-01 1.95247799e-01 -6.71822429e-01 -3.39927197e-01
7.98250377e-01 -1.31830409e-01 -6.51063859e-01 -6.45951509e-01
-1.01466918e+00 4.40776557e-01 1.68056682e-01 8.93506765e-01
1.03398526e+00 -1.44150662e+00 -6.10198855e-01 2.02338889e-01
5.54579437e-01 -3.24391633e-01 6.91632092e-01 8.38523030e-01
1.41919434e-01 5.35589755e-01 -1.85258985e-01 -8.07355225e-01
-1.65346575e+00 6.21407092e-01 2.13056937e-01 1.98885396e-01
-3.55066419e-01 7.20042765e-01 5.15407979e-01 -7.18298972e-01
6.53778851e-01 -4.06220883e-01 -4.19325203e-01 1.00509576e-01
6.08616471e-01 1.90785136e-02 -1.78938881e-01 -1.07953238e+00
-5.78618705e-01 4.44752753e-01 3.20459865e-02 -5.61389863e-01
9.12529588e-01 -5.09796798e-01 1.32488059e-02 8.44397545e-01
1.32457602e+00 -1.95178464e-01 -8.46175075e-01 -4.70477670e-01
-5.32117605e-01 -1.53812006e-01 7.80827776e-02 -7.94344842e-01
-1.14111292e+00 1.38893187e+00 7.85021305e-01 -1.65522680e-01
1.15883040e+00 4.28007066e-01 5.65804064e-01 2.45700300e-01
6.04694113e-02 -8.85959208e-01 2.93294579e-01 6.06505513e-01
1.03856730e+00 -1.40989149e+00 -4.93282616e-01 -3.73322636e-01
-1.22637618e+00 8.95931005e-01 7.49795318e-01 6.58243239e-01
5.66143572e-01 6.17260002e-02 3.94140244e-01 1.04514547e-01
-1.10112464e+00 -6.92093730e-01 5.62420309e-01 8.02482784e-01
6.12356842e-01 1.06527559e-01 2.41039723e-01 8.69514883e-01
2.07263529e-01 -2.57219285e-01 1.34988382e-01 8.52443755e-01
-2.70956755e-01 -1.23942304e+00 -5.23036420e-01 6.93259314e-02
-7.87214264e-02 -2.70230949e-01 -4.68581408e-01 3.59803200e-01
1.63126439e-02 1.24730897e+00 8.19695070e-02 -6.77345872e-01
1.94325849e-01 2.91592896e-01 2.91064084e-01 -8.02584410e-01
-5.11049330e-01 3.81275207e-01 5.56389391e-02 -4.27831411e-01
-5.53416431e-01 -4.82457519e-01 -1.30410945e+00 4.93018478e-02
1.50705818e-02 -2.46675983e-01 7.45018363e-01 1.03551376e+00
5.89510620e-01 7.09751427e-01 3.69322181e-01 -9.52046216e-01
-1.32933959e-01 -1.02850127e+00 -3.90779138e-01 3.06256890e-01
5.41316926e-01 -7.59710133e-01 -3.48355174e-01 1.12823159e-01] | [14.154187202453613, 5.058468341827393] |
74514ebe-8a15-4472-8d99-43e5a27b05a3 | membership-inference-attacks-and-defenses-in-1 | 2202.03335 | null | https://arxiv.org/abs/2202.03335v2 | https://arxiv.org/pdf/2202.03335v2.pdf | Membership Inference Attacks and Defenses in Neural Network Pruning | Neural network pruning has been an essential technique to reduce the computation and memory requirements for using deep neural networks for resource-constrained devices. Most existing research focuses primarily on balancing the sparsity and accuracy of a pruned neural network by strategically removing insignificant parameters and retraining the pruned model. Such efforts on reusing training samples pose serious privacy risks due to increased memorization, which, however, has not been investigated yet. In this paper, we conduct the first analysis of privacy risks in neural network pruning. Specifically, we investigate the impacts of neural network pruning on training data privacy, i.e., membership inference attacks. We first explore the impact of neural network pruning on prediction divergence, where the pruning process disproportionately affects the pruned model's behavior for members and non-members. Meanwhile, the influence of divergence even varies among different classes in a fine-grained manner. Enlighten by such divergence, we proposed a self-attention membership inference attack against the pruned neural networks. Extensive experiments are conducted to rigorously evaluate the privacy impacts of different pruning approaches, sparsity levels, and adversary knowledge. The proposed attack shows the higher attack performance on the pruned models when compared with eight existing membership inference attacks. In addition, we propose a new defense mechanism to protect the pruning process by mitigating the prediction divergence based on KL-divergence distance, whose effectiveness has been experimentally demonstrated to effectively mitigate the privacy risks while maintaining the sparsity and accuracy of the pruned models. | ['Lan Zhang', 'Xiaoyong Yuan'] | 2022-02-07 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 3.92233521e-01 9.61650610e-02 -2.72314698e-01 -4.71866578e-01
1.38876811e-02 -4.69502687e-01 -1.66314647e-01 -6.30074963e-02
-5.23881733e-01 6.91238582e-01 -1.89519092e-01 -6.59831405e-01
-2.79970407e-01 -8.00913155e-01 -8.03513467e-01 -8.54191184e-01
-4.34253365e-03 -3.53923887e-01 3.85804400e-02 1.25393555e-01
2.44513825e-01 6.58694327e-01 -1.25718760e+00 1.52101323e-01
8.94846797e-01 1.23422635e+00 -4.51564521e-01 1.49787873e-01
3.30244958e-01 5.15610993e-01 -9.05310690e-01 -9.01344657e-01
7.39878356e-01 -1.58180118e-01 -2.99256444e-01 -5.78123569e-01
5.08828282e-01 -7.63216078e-01 -4.67812806e-01 1.50382292e+00
3.41461092e-01 3.16712037e-02 2.82958329e-01 -1.36927891e+00
-4.65619296e-01 9.26696777e-01 -6.23118758e-01 2.93655545e-01
-3.87940377e-01 -1.12948999e-01 6.37569368e-01 -1.48349717e-01
-8.71090963e-02 9.09655154e-01 8.02199543e-01 8.00172925e-01
-1.08930373e+00 -1.56297350e+00 2.52813756e-01 1.68054067e-02
-1.61746383e+00 -5.03903508e-01 7.00421154e-01 -1.12826608e-01
5.21272600e-01 6.07606232e-01 3.08629751e-01 1.04712546e+00
2.33634129e-01 3.60664159e-01 7.81169653e-01 -1.83495790e-01
3.45419765e-01 4.72849488e-01 5.04987895e-01 6.49904311e-01
9.92998898e-01 2.00279906e-01 -2.70313382e-01 -6.52759850e-01
4.93708044e-01 2.91403711e-01 -4.03165668e-01 -2.38363132e-01
-2.10110500e-01 7.16270864e-01 2.63385028e-01 -7.58179054e-02
-6.56564757e-02 6.48241118e-02 5.12985349e-01 2.14594573e-01
2.74719089e-01 1.88469887e-01 -6.04163945e-01 2.63363272e-01
-7.80001700e-01 3.80919129e-03 8.21940064e-01 9.63388026e-01
5.10499239e-01 1.81975454e-01 -7.02926815e-02 4.25944686e-01
1.33952916e-01 1.55382380e-01 3.93123627e-01 -7.54724443e-01
4.68197554e-01 4.53087717e-01 -3.58394682e-01 -1.37011373e+00
1.19645506e-01 -4.98691678e-01 -1.11003530e+00 1.23997338e-01
1.57498762e-01 -7.93624341e-01 -5.72198749e-01 2.14546275e+00
2.51826614e-01 4.01991129e-01 1.36456817e-01 5.27390420e-01
5.00592947e-01 3.80922318e-01 2.04889491e-01 -2.30181754e-01
1.05866754e+00 -5.58414519e-01 -5.21450460e-01 1.28053069e-01
4.73465681e-01 -1.39325678e-01 6.43186927e-01 2.69956827e-01
-8.84738386e-01 -2.10610077e-01 -1.46484303e+00 2.43973956e-01
-1.94053248e-01 7.80808106e-02 6.16102099e-01 1.50518870e+00
-6.64644599e-01 5.78218699e-01 -8.57461095e-01 1.07875466e-01
9.52161551e-01 8.59028041e-01 -1.73945174e-01 2.53882587e-01
-1.27056277e+00 5.01976907e-01 4.00186479e-01 1.54542267e-01
-4.39013273e-01 -1.01153302e+00 -5.42687476e-01 3.95967245e-01
2.04629213e-01 -4.81481373e-01 8.09806466e-01 -7.73996234e-01
-1.30195796e+00 3.92391324e-01 1.25628740e-01 -1.12476206e+00
4.44355488e-01 -1.54062495e-01 -5.88018417e-01 -1.56048775e-01
-4.65939462e-01 3.63645554e-01 7.82233775e-01 -9.84819233e-01
-6.80776656e-01 -6.57942832e-01 9.98147279e-02 -4.52320352e-02
-1.04808772e+00 -1.51274145e-01 -1.03542723e-01 -7.63806939e-01
-1.68714643e-01 -7.72651732e-01 -3.42826724e-01 1.86422206e-02
-6.60137177e-01 3.19433749e-01 1.19477522e+00 -3.78101587e-01
1.52658927e+00 -2.58727121e+00 -5.65484941e-01 6.59672856e-01
4.13111836e-01 6.62949800e-01 2.47622803e-01 -1.70560494e-01
1.10403471e-01 4.41889554e-01 -1.45901233e-01 -2.12743670e-01
-2.06218511e-01 2.16874495e-01 -6.88461304e-01 5.70658922e-01
-1.30464241e-01 3.73329133e-01 -1.90661311e-01 -1.34342298e-01
-1.71990126e-01 5.00607610e-01 -8.33898365e-01 -3.64602245e-02
1.20814562e-01 -2.53970884e-02 -5.88379443e-01 4.50918674e-01
1.04699111e+00 8.72338042e-02 3.14571619e-01 -2.60651052e-01
1.84667110e-01 4.10753973e-02 -8.85245383e-01 7.24503458e-01
-5.33159450e-02 3.31254065e-01 1.41895935e-01 -9.63524938e-01
1.07770836e+00 4.51422073e-02 2.22104430e-01 -4.19323981e-01
5.59109807e-01 -5.05236052e-02 2.57889807e-01 1.44798763e-03
3.85325789e-01 2.89004929e-02 -8.00920352e-02 4.17088419e-01
-4.27493542e-01 8.24591100e-01 -4.82748568e-01 -8.36221948e-02
9.24161792e-01 -5.78853607e-01 1.92136511e-01 -4.45743829e-01
4.13561404e-01 -3.49648654e-01 9.17820275e-01 9.31528807e-01
-4.62460399e-01 1.62117243e-01 5.02556205e-01 -3.48311514e-01
-5.73001862e-01 -8.45490336e-01 -3.10129404e-01 8.63747001e-01
3.85664821e-01 -3.10070932e-01 -9.74389613e-01 -8.74996841e-01
3.18197161e-01 7.98729599e-01 -6.82991922e-01 -7.48460114e-01
-3.95538598e-01 -7.98472285e-01 1.31696582e+00 3.31791282e-01
8.65003765e-01 -5.07363081e-01 -9.22973037e-01 -4.56080854e-01
3.69274169e-01 -9.60161030e-01 -5.19447684e-01 3.07944089e-01
-1.01826775e+00 -9.30649340e-01 -2.86459439e-02 -5.69665968e-01
8.47073555e-01 2.14049071e-01 2.40358695e-01 1.21573061e-01
-6.20508529e-02 -1.41295075e-01 -3.46311275e-03 -8.24244440e-01
-1.16245195e-01 1.85387343e-01 3.66779953e-01 1.55012250e-01
5.10212064e-01 -9.26537216e-01 -5.21263599e-01 3.00461739e-01
-8.62039447e-01 -3.44637334e-01 5.98156691e-01 5.33576727e-01
5.37179232e-01 5.53127587e-01 7.03594744e-01 -1.37137830e+00
8.74722660e-01 -5.64142942e-01 -6.04665041e-01 3.64721656e-01
-7.17692196e-01 2.63947342e-02 9.39027488e-01 -9.80274379e-01
-1.00802159e+00 -7.15438947e-02 -8.83065630e-03 -7.21210480e-01
9.32900980e-02 8.64461511e-02 -5.75998604e-01 -4.77147162e-01
4.46655363e-01 6.93755820e-02 8.78573284e-02 -3.67427737e-01
8.95485133e-02 6.24548972e-01 4.65682715e-01 -5.63159943e-01
5.64087451e-01 5.21213710e-01 1.43444657e-01 -7.88029194e-01
-5.09258628e-01 1.29946724e-01 -1.54996719e-02 3.13124418e-01
4.44538742e-01 -6.10245347e-01 -9.85742271e-01 4.04579788e-01
-7.89422929e-01 2.01652735e-01 -1.77014887e-01 2.48402774e-01
1.55735806e-01 5.49476624e-01 -4.73564357e-01 -8.97818148e-01
-8.97229314e-01 -9.16634142e-01 1.21149562e-01 3.71676594e-01
-1.29773304e-01 -6.16835415e-01 -3.86103064e-01 2.42555603e-01
4.42525983e-01 3.73707443e-01 1.14653683e+00 -1.26019979e+00
-3.75969589e-01 -2.97904789e-01 -2.36186758e-01 5.33822060e-01
1.28936216e-01 -1.49734750e-01 -9.75706875e-01 -3.82005155e-01
5.13072371e-01 4.04729769e-02 7.00320303e-01 2.98504978e-01
1.68749833e+00 -9.90771174e-01 -3.49041849e-01 1.01339614e+00
1.13413751e+00 4.99623388e-01 6.44304812e-01 8.69645402e-02
6.70696259e-01 4.10043716e-01 3.11508954e-01 6.33901119e-01
-2.05216810e-01 1.43271998e-01 5.00332236e-01 2.27518633e-01
5.57172418e-01 -2.89330870e-01 1.94050550e-01 2.83058286e-01
3.15423727e-01 -1.71485692e-01 -4.03380871e-01 9.57978219e-02
-1.62311065e+00 -8.98416638e-01 2.80163348e-01 2.44719172e+00
9.17640328e-01 3.69929969e-01 -1.86448693e-01 7.59565905e-02
7.56450772e-01 3.58808115e-02 -9.03989434e-01 -5.71749508e-01
-3.76362018e-02 3.30951422e-01 8.79220307e-01 1.03865243e-01
-1.05686343e+00 6.62953734e-01 5.92851925e+00 9.33838725e-01
-1.24051821e+00 2.51981877e-02 1.07177508e+00 -5.33124268e-01
-1.17472567e-01 -3.12571079e-02 -1.17688203e+00 6.83732927e-01
8.94874513e-01 -3.00429672e-01 2.01359496e-01 1.20171237e+00
-1.95136428e-01 2.84256130e-01 -9.73240912e-01 8.16470563e-01
4.52764817e-02 -1.27940261e+00 2.14877218e-01 3.09226304e-01
4.55148220e-01 -2.29838192e-01 4.04665887e-01 2.81530261e-01
2.27499887e-01 -9.65844095e-01 5.52685916e-01 2.21056536e-01
6.05677843e-01 -1.29914165e+00 6.44616902e-01 5.04181325e-01
-6.39191628e-01 -4.44727182e-01 -5.40000618e-01 5.50080184e-03
-2.84247011e-01 5.16115248e-01 -5.35490811e-01 1.30793065e-01
8.66790473e-01 7.39995763e-03 -3.20137292e-01 6.58848822e-01
1.19163409e-01 7.11187661e-01 -4.44660366e-01 -1.71849295e-01
-1.45363539e-01 -1.73514098e-01 4.33955312e-01 8.38595688e-01
1.04904495e-01 3.69043350e-01 -1.43341377e-01 8.74605119e-01
-3.72130483e-01 -7.79685229e-02 -6.31124496e-01 1.72120720e-01
9.52962279e-01 8.74695659e-01 -1.82984620e-01 5.33145256e-02
-1.49807572e-01 5.70095897e-01 3.98234069e-01 1.71568468e-01
-9.70174909e-01 -5.37939847e-01 1.08915019e+00 1.45352453e-01
3.47051114e-01 1.24757245e-01 -7.41658568e-01 -8.01008523e-01
6.16554618e-02 -8.17420304e-01 5.76284826e-01 1.18370475e-02
-1.21889138e+00 5.44637501e-01 8.66894126e-02 -7.05113351e-01
2.30814606e-01 -2.45556518e-01 -7.71565557e-01 7.94240654e-01
-9.88990545e-01 -6.72904074e-01 1.41649380e-01 5.94278753e-01
-1.51470035e-01 -3.90186191e-01 6.41407728e-01 3.79404753e-01
-1.22501981e+00 1.73958230e+00 8.46271291e-02 2.28303432e-01
2.60580242e-01 -2.16564149e-01 3.60723175e-02 1.16883016e+00
-9.79093537e-02 1.34992731e+00 4.39621598e-01 -7.60990202e-01
-1.16751504e+00 -1.31664956e+00 3.71402472e-01 4.24740426e-02
3.32665175e-01 -4.01356816e-01 -1.11242938e+00 5.70504725e-01
-1.73091561e-01 -5.13467863e-02 1.12480187e+00 3.42913158e-02
-6.12922907e-01 -3.52981150e-01 -1.75375879e+00 8.87918234e-01
8.60348582e-01 -2.89393634e-01 -1.24992758e-01 -2.99982816e-01
9.85888183e-01 -1.80460006e-01 -5.84946573e-01 8.13772738e-01
7.82006919e-01 -9.58063126e-01 9.46809411e-01 -7.77069509e-01
4.49212790e-02 7.93103948e-02 -3.55663776e-01 -4.25000489e-01
-1.80430755e-01 -6.95907414e-01 -6.66265607e-01 1.31103778e+00
4.66666102e-01 -9.03603792e-01 1.42239940e+00 1.10680246e+00
3.04445386e-01 -1.05867016e+00 -9.10654485e-01 -6.50112808e-01
5.61108207e-03 -1.35207802e-01 9.79967892e-01 9.84342754e-01
-2.14466467e-01 -1.15317721e-02 -5.65483630e-01 6.29268348e-01
5.96882164e-01 -1.36033028e-01 5.62963009e-01 -9.75013256e-01
-3.59571517e-01 -4.47843164e-01 -4.19493407e-01 -8.08146954e-01
2.35310927e-01 -5.49759746e-01 -4.07905787e-01 -3.84024084e-01
1.30378991e-01 -5.93327582e-01 -7.08089650e-01 6.51771188e-01
-4.71231975e-02 2.60964278e-02 6.39040545e-02 1.71960101e-01
-1.99591622e-01 3.48122448e-01 6.23293161e-01 1.83812566e-02
-3.33061069e-01 4.70022023e-01 -1.18540108e+00 6.74326301e-01
1.23785794e+00 -6.99663937e-01 -9.01265383e-01 -2.30755448e-01
-1.53202087e-01 -2.01341197e-01 2.13062599e-01 -1.06896996e+00
5.13760030e-01 -1.23997301e-01 2.53129900e-01 -4.45623964e-01
2.07296178e-01 -1.20693636e+00 3.84692520e-01 6.88097119e-01
-4.78659958e-01 -3.17229778e-02 4.36642796e-01 8.02986801e-01
2.22125828e-01 -2.84413517e-01 8.65825355e-01 2.85078019e-01
-1.17429987e-01 5.47042191e-01 -2.22815260e-01 -1.12433411e-01
1.27475524e+00 -6.68298841e-01 -3.86649340e-01 8.88599455e-03
-2.93650836e-01 1.08922534e-01 3.60736817e-01 1.95598006e-01
6.62347198e-01 -9.44975913e-01 -2.22993433e-01 6.63995028e-01
-2.72511274e-01 -9.29904506e-02 4.22974974e-01 4.64286506e-01
-3.44988972e-01 2.08315328e-01 -3.71612966e-01 2.79889945e-02
-1.71103311e+00 6.29816711e-01 3.37739050e-01 -1.14783317e-01
-5.02577722e-01 9.90679622e-01 3.16013366e-01 -1.86956674e-01
8.36841643e-01 -2.75672793e-01 5.62858675e-03 -5.21606624e-01
4.92917478e-01 4.75604892e-01 -6.86383322e-02 -1.82365656e-01
-3.58767927e-01 1.63593013e-02 -6.33318663e-01 3.12149495e-01
8.66165698e-01 1.45626247e-01 -7.68708661e-02 -2.58137226e-01
1.26922703e+00 1.99822158e-01 -1.16067815e+00 -1.84712708e-01
-3.01972449e-01 -5.45422673e-01 4.51463200e-02 -5.22249699e-01
-1.41987085e+00 6.06641889e-01 5.30418515e-01 6.90052658e-02
1.31860042e+00 -6.04904771e-01 1.02970564e+00 5.19129276e-01
3.29163224e-01 -7.13032842e-01 -4.07065272e-01 2.44584844e-01
1.08298667e-01 -7.22622275e-01 2.88168579e-01 -5.85441768e-01
-4.50643867e-01 5.75800240e-01 9.79442954e-01 -7.10794553e-02
1.05606008e+00 5.15582919e-01 -3.67333651e-01 1.67280510e-01
-5.05104244e-01 7.35606134e-01 1.00110255e-01 5.47905564e-01
-2.09681034e-01 8.95164907e-02 -2.21191406e-01 1.42741716e+00
-3.45794946e-01 -2.39780679e-01 1.96769208e-01 1.01568520e+00
-3.02402943e-01 -1.00061464e+00 -3.68780158e-02 7.46620655e-01
-9.56770301e-01 -1.57849312e-01 -5.44488490e-01 5.52423179e-01
3.24187517e-01 6.76843345e-01 -2.25623157e-02 -8.80038381e-01
2.00103402e-01 -3.30937833e-01 1.02176286e-01 -3.38024944e-01
-8.76302421e-01 -5.55176616e-01 3.00599113e-02 -2.70265132e-01
2.15646520e-01 -3.31413060e-01 -1.09312880e+00 -7.61816144e-01
-5.76904476e-01 2.54635185e-01 3.51523995e-01 6.20246589e-01
8.60226810e-01 2.74645567e-01 6.61592126e-01 -8.52038190e-02
-1.09579432e+00 -3.61687779e-01 -5.89371622e-01 3.13303471e-02
1.39611319e-01 -3.51633608e-01 -4.22435313e-01 -5.85367501e-01] | [5.892332553863525, 7.106832504272461] |
41829039-27dc-4526-9450-77054a4747a0 | revision-for-concision-a-constrained | 2210.14257 | null | https://arxiv.org/abs/2210.14257v1 | https://arxiv.org/pdf/2210.14257v1.pdf | Revision for Concision: A Constrained Paraphrase Generation Task | Academic writing should be concise as concise sentences better keep the readers' attention and convey meaning clearly. Writing concisely is challenging, for writers often struggle to revise their drafts. We introduce and formulate revising for concision as a natural language processing task at the sentence level. Revising for concision requires algorithms to use only necessary words to rewrite a sentence while preserving its meaning. The revised sentence should be evaluated according to its word choice, sentence structure, and organization. The revised sentence also needs to fulfil semantic retention and syntactic soundness. To aide these efforts, we curate and make available a benchmark parallel dataset that can depict revising for concision. The dataset contains 536 pairs of sentences before and after revising, and all pairs are collected from college writing centres. We also present and evaluate the approaches to this problem, which may assist researchers in this area. | ['Kwan Hui Lim', 'Wenchuan Mu'] | 2022-10-25 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 4.85698432e-01 2.02219576e-01 -2.41244689e-01 -5.97731113e-01
-5.90260625e-01 -8.22743058e-01 4.18563932e-01 8.09004903e-01
-6.42659903e-01 1.16248775e+00 5.97926021e-01 -5.99682868e-01
-3.02240670e-01 -5.11152446e-01 -3.90439630e-01 1.92519143e-01
6.01679683e-01 8.44867751e-02 1.05065115e-01 -4.28784579e-01
1.08044386e+00 5.33854067e-01 -1.05711913e+00 5.85971951e-01
1.07725585e+00 5.16717471e-02 7.35335648e-01 9.02928174e-01
-4.74769384e-01 5.81985891e-01 -1.13352513e+00 -9.83447015e-01
-1.98675722e-01 -5.28681219e-01 -1.48207986e+00 1.68120444e-01
6.80069506e-01 7.19270483e-02 1.56264558e-01 1.07281578e+00
2.82867014e-01 2.52211630e-01 3.74017119e-01 -4.95258182e-01
-1.03995538e+00 8.92026484e-01 -2.31155261e-01 6.46048069e-01
7.46405721e-01 -3.30726147e-01 1.22681010e+00 -1.00277102e+00
8.50680053e-01 1.06247067e+00 5.96312940e-01 5.78324020e-01
-9.29323316e-01 -2.42901698e-01 2.99762338e-01 3.41867924e-01
-1.10156178e+00 -8.32420826e-01 6.12016201e-01 -1.91036806e-01
1.41038704e+00 8.22485209e-01 9.22086775e-01 7.81711578e-01
6.74773395e-01 4.59295869e-01 8.10845971e-01 -8.74996066e-01
-5.72512746e-02 1.64622828e-01 7.66079187e-01 6.15384996e-01
6.21070147e-01 -7.33610511e-01 -7.15203047e-01 5.09667575e-01
-6.27636770e-03 -2.49920428e-01 -3.87072772e-01 4.12807703e-01
-1.25186050e+00 3.60810131e-01 -1.67877480e-01 6.67371213e-01
-6.49808273e-02 -1.69216156e-01 4.19497371e-01 4.98994410e-01
1.36702359e-01 1.14844310e+00 -2.82599509e-01 -5.43921173e-01
-1.23372757e+00 3.77808034e-01 7.71072984e-01 1.20581222e+00
2.85259485e-01 -4.77244467e-01 -4.29020077e-01 9.68042910e-01
1.35825008e-01 4.00155425e-01 6.17688239e-01 -9.43208337e-01
8.32964838e-01 4.99185920e-01 -1.94327325e-01 -1.26096296e+00
-1.96304813e-01 -5.54442644e-01 -7.70211816e-01 -5.91557860e-01
-1.47177158e-02 3.24917018e-01 -1.72008231e-01 1.24736977e+00
-3.41803700e-01 -5.90000153e-01 2.62526572e-01 4.07458603e-01
1.35769534e+00 7.01142788e-01 -2.12041289e-01 -6.88912392e-01
1.42575061e+00 -1.24307454e+00 -1.15752804e+00 -2.39022404e-01
8.47568631e-01 -1.40447390e+00 1.33903885e+00 4.11097169e-01
-1.60623014e+00 -5.65630198e-01 -1.36394393e+00 -6.48782074e-01
-2.92159647e-01 3.46123368e-01 -8.40003267e-02 5.26965320e-01
-1.06140244e+00 9.07654464e-01 -4.34490949e-01 -5.58173239e-01
2.64273405e-01 -1.54967457e-01 -3.31454009e-01 7.87162781e-02
-1.03139937e+00 1.13065970e+00 2.82839268e-01 2.29536191e-01
3.61883253e-01 -9.24093604e-01 -8.01768899e-01 1.56366140e-01
1.70123130e-01 -8.16774845e-01 1.64049232e+00 -5.29240608e-01
-1.50714588e+00 1.06298268e+00 -6.23648524e-01 -2.61806160e-01
3.58924180e-01 -1.66302577e-01 -4.78275061e-01 2.09397212e-01
5.17278731e-01 2.27903232e-01 5.19068480e-01 -6.89905107e-01
-4.83635634e-01 -1.18938230e-01 -2.03540951e-01 8.07106197e-02
-4.87109601e-01 2.99544722e-01 -4.26379025e-01 -1.15458095e+00
9.75419357e-02 -5.65779567e-01 2.39870086e-01 -1.17348164e-01
-7.47604787e-01 -4.09594268e-01 4.37637955e-01 -9.36836720e-01
1.90618527e+00 -1.80030632e+00 1.56872943e-01 -5.22320345e-02
3.92562777e-01 3.71975005e-01 -2.51571566e-01 7.77925432e-01
-8.96951333e-02 7.26547778e-01 -4.25050437e-01 -6.12138093e-01
-2.14888286e-02 1.90839544e-01 -4.45144534e-01 1.33829623e-01
2.84159690e-01 1.19220197e+00 -9.24836218e-01 -7.24009573e-01
-1.13586642e-01 -1.97464883e-01 -3.34990740e-01 6.48076981e-02
2.38907099e-01 -1.43966660e-01 -3.22298706e-01 3.48438233e-01
4.47124302e-01 -1.30325094e-01 4.83425148e-02 1.08461808e-02
-3.92165363e-01 8.14650297e-01 -8.42824757e-01 1.58000553e+00
-5.23283720e-01 1.10092890e+00 -4.00978476e-01 -8.39963257e-01
1.01708031e+00 -9.14586857e-02 -7.85291120e-02 -1.03929865e+00
-3.73692125e-01 2.52476692e-01 -1.48250237e-01 -6.36204064e-01
1.29584455e+00 -3.11166435e-01 -2.40071774e-01 7.29714513e-01
-5.30292690e-01 -8.35510850e-01 1.01838279e+00 6.39876127e-01
8.02496374e-01 -2.70902783e-01 5.98380268e-01 -5.91304481e-01
8.39354634e-01 -5.18543608e-02 1.82765961e-01 9.21680629e-01
-5.06716669e-02 4.65163559e-01 6.08377695e-01 -3.55568051e-01
-1.08744740e+00 -9.24323678e-01 -1.60799712e-01 8.11005235e-01
-1.09406402e-02 -1.17687654e+00 -5.53602278e-01 -4.03664023e-01
-2.19566599e-01 1.22114217e+00 -1.77059710e-01 -2.59944081e-01
-6.81372344e-01 -9.95893031e-03 4.99344647e-01 3.31270218e-01
4.23297733e-01 -9.18912470e-01 -7.15308309e-01 3.72705221e-01
-3.27974588e-01 -1.00769937e+00 -8.04253340e-01 -4.32366580e-02
-6.65248930e-01 -1.04497504e+00 -6.79326594e-01 -9.95138168e-01
7.85935283e-01 2.58524120e-01 1.22422361e+00 5.48702002e-01
5.65468967e-02 3.00032884e-01 -5.17869055e-01 -5.40140867e-01
-8.20662320e-01 4.45543766e-01 -1.88427851e-01 -8.39419723e-01
1.93233024e-02 -2.29299918e-01 1.41855419e-01 -3.09412181e-01
-9.27902997e-01 5.67300320e-01 3.44591826e-01 7.63135195e-01
3.23950261e-01 -4.33308594e-02 5.39629519e-01 -1.00669992e+00
1.58641517e+00 1.36060447e-01 -8.36571082e-02 9.20312524e-01
-7.55922258e-01 2.72626787e-01 8.80812407e-01 1.77930430e-01
-9.69824255e-01 -6.35451317e-01 -2.47525856e-01 5.82908869e-01
2.96888381e-01 7.31596172e-01 2.42902175e-01 2.90086925e-01
4.96452957e-01 4.37055051e-01 -1.17967255e-01 -4.59102094e-01
1.77657083e-01 6.77935183e-01 7.93976188e-01 -7.89522886e-01
5.53831041e-01 -2.62466818e-01 1.95465535e-02 -1.17389822e+00
-1.09538400e+00 -2.20924571e-01 -9.49021161e-01 -1.28070295e-01
3.04901272e-01 -3.39872658e-01 -6.34016752e-01 -4.04317528e-02
-1.70850468e+00 1.87791497e-01 -4.51959163e-01 1.70682758e-01
-1.70981765e-01 1.00976717e+00 -5.41858256e-01 -3.90177816e-01
-5.11749208e-01 -1.02499974e+00 1.01271749e+00 3.09945434e-01
-1.04826832e+00 -1.12711883e+00 -6.71877339e-03 3.80192935e-01
3.15152854e-01 -2.17561305e-01 1.19955730e+00 -4.67817932e-01
4.17110771e-02 5.93941025e-02 -1.58675954e-01 4.37273324e-01
4.16075468e-01 4.20198917e-01 -3.39742064e-01 -1.85195625e-01
-8.80169719e-02 -1.00905649e-01 6.94444895e-01 -7.65990168e-02
1.25923574e+00 -6.99288189e-01 1.36650547e-01 1.97916895e-01
8.35376263e-01 -2.35622320e-02 4.79612648e-01 5.66657007e-01
4.25289512e-01 7.02958047e-01 3.90116006e-01 4.69177514e-01
5.00554323e-01 4.11790162e-01 -4.31891114e-01 3.70672643e-01
-3.95818621e-01 -1.62670031e-01 2.03098744e-01 1.87708545e+00
2.62820780e-01 -3.16729456e-01 -1.02038884e+00 3.11397910e-01
-1.57016993e+00 -1.09544468e+00 -5.50494850e-01 1.65172386e+00
1.43446326e+00 2.85941452e-01 -4.05243307e-01 3.52044433e-01
5.36429465e-01 2.42425203e-01 5.60490526e-02 -1.24508536e+00
-4.45044249e-01 4.98435378e-01 -1.27176449e-01 1.03552401e+00
-4.75111991e-01 1.15413976e+00 6.27945375e+00 7.07796931e-01
-1.05897522e+00 -3.95103276e-01 4.97049630e-01 -3.08621973e-02
-7.80969799e-01 -7.36390576e-02 -9.50956523e-01 4.20309305e-01
8.00181448e-01 -9.22617495e-01 2.41215497e-01 1.63571686e-01
5.86578608e-01 -3.22785258e-01 -1.34938657e+00 1.03664565e+00
3.69211733e-01 -1.78629458e+00 5.30802786e-01 -7.24658787e-01
7.39828467e-01 -7.97890186e-01 -8.48858505e-02 1.79265141e-01
-4.33157384e-01 -1.19372320e+00 9.45768416e-01 1.02883434e+00
4.38239247e-01 -6.16759717e-01 6.30845428e-01 4.46779400e-01
-9.87802863e-01 4.95616168e-01 -3.53054523e-01 -7.22463071e-01
1.90640185e-02 5.63103557e-01 -8.26467335e-01 6.23596191e-01
3.32772553e-01 1.29372346e+00 -1.25460696e+00 7.02324867e-01
-6.40322626e-01 3.44904721e-01 7.84807876e-02 -8.36255193e-01
1.40441284e-01 -1.69634447e-01 7.00612307e-01 1.55456614e+00
3.77753645e-01 1.55774310e-01 -9.82276648e-02 7.45612264e-01
-2.14369789e-01 5.92500389e-01 -2.20573366e-01 -5.84760308e-01
6.27678216e-01 8.67467046e-01 -6.97483063e-01 -5.12602568e-01
-7.93584511e-02 1.13446999e+00 4.45507377e-01 2.56666869e-01
-2.66675055e-01 -9.43603694e-01 2.82297403e-01 -4.90047522e-02
-1.13167658e-01 -4.39831197e-01 -8.68047476e-01 -1.26883924e+00
6.20150805e-01 -8.68731678e-01 6.26703501e-02 -9.68952179e-01
-1.18616545e+00 5.16444743e-01 -1.60567045e-01 -7.06321061e-01
1.57050323e-02 -4.79939461e-01 -6.92080438e-01 1.05663359e+00
-1.52036595e+00 -5.48502386e-01 -7.22688884e-02 -2.90817041e-02
1.08580196e+00 -7.86565710e-03 6.98983073e-01 5.87724000e-02
-8.48637581e-01 6.23844266e-01 -3.75668779e-02 -2.00048760e-01
8.85640085e-01 -1.13756692e+00 3.71278763e-01 1.21672606e+00
2.93622404e-01 1.19351363e+00 8.34014237e-01 -7.59611845e-01
-1.20586908e+00 -9.94374692e-01 2.28262115e+00 -4.93690878e-01
8.12120318e-01 -1.81968324e-02 -9.40545380e-01 4.04756039e-01
5.97402096e-01 -7.68122315e-01 7.21710563e-01 2.44300067e-02
1.21848680e-01 -3.33691798e-02 -5.53979814e-01 1.01974058e+00
1.01569581e+00 -6.46561503e-01 -1.59641922e+00 5.82822680e-01
6.72425866e-01 -3.91435295e-01 -6.08347476e-01 5.90183362e-02
3.17921758e-01 -6.64024055e-01 5.87255657e-01 -9.63638365e-01
1.07082283e+00 -5.42334542e-02 2.16522768e-01 -1.40718889e+00
-5.23297250e-01 -8.67421091e-01 1.80296823e-01 1.33349776e+00
5.70370197e-01 -3.38220835e-01 3.44851226e-01 7.09430099e-01
-5.16053677e-01 -9.01714146e-01 -7.51140833e-01 -8.50336850e-01
2.81831115e-01 -5.14476955e-01 6.41212821e-01 7.68661737e-01
5.89633048e-01 7.77656257e-01 4.24118564e-02 -5.13696492e-01
9.58500206e-02 2.54699498e-01 4.36697602e-01 -1.02705967e+00
9.72530022e-02 -1.00297177e+00 9.65013802e-02 -1.12927926e+00
5.77156067e-01 -1.27670777e+00 -1.84522897e-01 -1.93409264e+00
4.58449513e-01 1.40106633e-01 3.77059966e-01 5.36807835e-01
-3.34466726e-01 2.78798379e-02 4.40283656e-01 1.17225148e-01
-7.73689151e-01 6.24841809e-01 1.61916721e+00 -3.62867385e-01
-1.19904034e-01 -9.43153873e-02 -1.12109649e+00 5.66114902e-01
7.73252010e-01 -2.58943021e-01 -3.24659050e-01 -6.11554384e-01
6.12992167e-01 -1.39069155e-01 4.92708646e-02 -4.84330148e-01
6.08695388e-01 -4.76265073e-01 3.75924945e-01 -6.53297126e-01
-1.40492707e-01 -3.73094022e-01 -2.08397597e-01 6.71722949e-01
-8.81892860e-01 8.83275747e-01 2.35919505e-01 -1.70531824e-01
-1.57910004e-01 -8.07136416e-01 7.07379580e-01 -4.82322052e-02
-5.69313228e-01 -4.02286887e-01 -7.29155302e-01 5.34724832e-01
7.63390422e-01 -3.04076105e-01 -4.88764197e-01 -2.09485084e-01
-4.70511109e-01 1.60894305e-01 3.29605818e-01 5.79931855e-01
9.56474423e-01 -1.12817943e+00 -9.11408126e-01 4.41608280e-02
8.29740912e-02 -2.67588854e-01 -2.96294298e-02 5.12180626e-01
-1.00689936e+00 8.86959851e-01 -1.13137409e-01 -1.43630102e-01
-1.59445894e+00 2.67418832e-01 -1.35548741e-01 -2.42532119e-01
-6.74615264e-01 8.78572285e-01 -5.46933115e-01 -2.32104078e-01
2.28238311e-02 -8.77396584e-01 -3.90287429e-01 1.38517022e-01
7.64687717e-01 4.98609066e-01 3.70165318e-01 -5.61887801e-01
-3.59517097e-01 5.30620277e-01 -2.83706158e-01 -1.08167063e-02
1.31683028e+00 -5.38679481e-01 -7.48265445e-01 6.67951047e-01
1.27362537e+00 5.96639395e-01 -2.46008679e-01 -2.29707986e-01
5.43201566e-01 -5.26588440e-01 -1.43122599e-01 -8.51301134e-01
-1.87480375e-01 7.06910729e-01 -5.22990406e-01 2.23520637e-01
8.39757502e-01 -1.33160591e-01 8.38386416e-01 1.15095270e+00
-2.17484951e-01 -1.42291868e+00 1.66901916e-01 1.15922379e+00
1.35524631e+00 -8.66203547e-01 5.39126694e-01 -5.21720827e-01
-3.44249010e-01 1.78721035e+00 3.44903320e-01 2.99143195e-01
2.92343765e-01 -3.65801156e-02 -2.66637951e-01 -1.05176017e-01
-7.23768771e-01 4.34231967e-01 6.32556975e-01 1.07981972e-01
8.80142033e-01 -1.83713622e-02 -1.04544377e+00 9.11120713e-01
-1.07909214e+00 -2.40257427e-01 1.05320370e+00 9.44042504e-01
-4.73003805e-01 -1.33777404e+00 -2.70142257e-01 5.57501674e-01
-1.98283553e-01 -5.27629137e-01 -9.26039696e-01 4.22216803e-01
-2.49798775e-01 1.08995426e+00 -1.82332639e-02 1.41143158e-01
5.52239895e-01 -8.12726393e-02 6.66870892e-01 -1.03149188e+00
-8.05086672e-01 -5.90726912e-01 3.78001988e-01 -3.44082713e-02
-2.28188857e-01 -5.35590112e-01 -1.37373900e+00 -8.15623343e-01
1.34951934e-01 5.91986477e-01 3.69644076e-01 1.26631939e+00
2.62438864e-01 8.39580715e-01 2.73250908e-01 -7.73228556e-02
-7.43289709e-01 -7.61688232e-01 -4.51249182e-01 2.44581163e-01
2.01571018e-01 6.93756109e-03 1.18626133e-01 3.97926956e-01] | [12.033157348632812, 9.462803840637207] |
0073446e-7d9e-4a5a-a1a3-40ccdc6ddb9d | ehrsql-a-practical-text-to-sql-benchmark-for-1 | 2301.07695 | null | https://arxiv.org/abs/2301.07695v4 | https://arxiv.org/pdf/2301.07695v4.pdf | EHRSQL: A Practical Text-to-SQL Benchmark for Electronic Health Records | We present a new text-to-SQL dataset for electronic health records (EHRs). The utterances were collected from 222 hospital staff members, including physicians, nurses, and insurance review and health records teams. To construct the QA dataset on structured EHR data, we conducted a poll at a university hospital and used the responses to create seed questions. We then manually linked these questions to two open-source EHR databases, MIMIC-III and eICU, and included various time expressions and held-out unanswerable questions in the dataset, which were also collected from the poll. Our dataset poses a unique set of challenges: the model needs to 1) generate SQL queries that reflect a wide range of needs in the hospital, including simple retrieval and complex operations such as calculating survival rate, 2) understand various time expressions to answer time-sensitive questions in healthcare, and 3) distinguish whether a given question is answerable or unanswerable. We believe our dataset, EHRSQL, can serve as a practical benchmark for developing and assessing QA models on structured EHR data and take a step further towards bridging the gap between text-to-SQL research and its real-life deployment in healthcare. EHRSQL is available at https://github.com/glee4810/EHRSQL. | ['Edward Choi', 'Jong-Yeup Kim', 'Minjoon Seo', 'Seongjun Yang', 'Woncheol Shin', 'Yeonsu Kwon', 'Seongsu Bae', 'Hyeonji Hwang', 'Gyubok Lee'] | 2023-01-16 | ehrsql-a-practical-text-to-sql-benchmark-for | https://openreview.net/forum?id=B2W8Vy0rarw | https://openreview.net/pdf?id=B2W8Vy0rarw | neurips-2022-datasets-and-benchmarks-2022-12 | ['text-to-sql'] | ['computer-code'] | [-2.31020465e-01 2.78779775e-01 5.42836823e-02 -9.57520843e-01
-1.60577023e+00 -7.35633910e-01 -2.91323364e-01 8.86325836e-01
-1.53152287e-01 7.09359288e-01 8.61532390e-01 -8.67847502e-01
-2.21268594e-01 -8.99325967e-01 -3.57439071e-01 -9.64618921e-02
-1.89638995e-02 8.86124134e-01 -3.71163875e-01 -1.85012355e-01
-2.29324102e-01 6.62613958e-02 -9.00087714e-01 6.80563450e-01
9.53442752e-01 9.95725036e-01 -2.30577737e-01 8.88732135e-01
-3.87686980e-03 1.38243711e+00 -5.36812127e-01 -4.63058919e-01
6.26623556e-02 -3.46220642e-01 -1.06448615e+00 -4.87330437e-01
-1.40424252e-01 -4.67443377e-01 -2.29910150e-01 5.67920566e-01
9.13355291e-01 -4.94679719e-01 9.27704349e-02 -1.01521063e+00
-4.28182960e-01 5.01921892e-01 4.55398053e-01 5.60989976e-03
1.13593328e+00 5.40396571e-01 7.91586518e-01 -3.53220135e-01
7.67054498e-01 8.57227683e-01 9.27866817e-01 3.62203866e-01
-8.56700301e-01 -3.29170853e-01 -7.26575494e-01 -1.02517582e-01
-1.23605359e+00 -5.54836929e-01 3.16139087e-02 -4.16546017e-01
7.54311442e-01 7.92349935e-01 5.83467066e-01 9.37745690e-01
3.95510852e-01 6.57633603e-01 1.04178131e+00 -1.96938403e-03
3.33599001e-01 1.00447759e-01 3.51111442e-01 5.76374710e-01
-1.64763376e-01 2.27380246e-02 -1.37886256e-01 -9.59885657e-01
5.66864848e-01 4.83184934e-01 -4.65460628e-01 4.16265249e-01
-1.45162451e+00 8.01987290e-01 3.05528492e-02 -1.32693246e-01
-7.52840638e-01 -3.00213039e-01 5.09318709e-01 6.31023407e-01
-7.67614320e-02 6.33862376e-01 -1.27869332e+00 -6.08230472e-01
-4.80339319e-01 3.76243562e-01 1.43045735e+00 1.37915587e+00
5.48900962e-01 -7.40777969e-01 -7.42359817e-01 5.42150021e-01
1.16792329e-01 6.29467905e-01 1.98317736e-01 -1.16235256e+00
3.84200841e-01 8.15242112e-01 6.15536034e-01 -7.70303488e-01
-7.19876766e-01 1.39534652e-01 -6.68370664e-01 -9.40400958e-01
4.27147388e-01 -6.70561135e-01 -7.13094950e-01 1.48051918e+00
4.40701753e-01 -3.57291996e-01 2.67068952e-01 6.06356144e-01
1.31659698e+00 4.94222075e-01 2.67662436e-01 -3.46935004e-01
1.98684072e+00 -2.01792851e-01 -1.07967758e+00 1.90097347e-01
1.23281109e+00 -7.83421338e-01 1.20341027e+00 9.10182744e-02
-1.36487806e+00 -1.16128027e-01 -9.05341879e-02 -8.20089504e-02
-1.63951010e-01 -3.02328885e-01 3.12266052e-01 5.87348521e-01
-1.00329089e+00 1.38064861e-01 -1.07725680e+00 -3.67383689e-01
2.06701830e-01 5.18536977e-02 -1.07779399e-01 -3.58656168e-01
-1.68126941e+00 6.34129524e-01 -6.98877200e-02 -2.20516801e-01
-7.63924301e-01 -1.02000618e+00 -1.06741357e+00 6.02118000e-02
4.46488053e-01 -1.20243752e+00 1.75299776e+00 -2.98890844e-02
-8.34001899e-01 8.20117235e-01 -3.59802932e-01 -7.78610930e-02
4.60368961e-01 -6.32770360e-02 -6.82907760e-01 1.19157352e-01
2.73811728e-01 9.73448753e-02 -1.76672772e-01 -6.77859068e-01
-3.79665196e-01 -5.96595407e-01 -1.73988760e-01 -2.21030802e-01
1.37497500e-01 2.98726290e-01 -2.94688851e-01 -4.40137774e-01
-2.19432473e-01 -5.90768039e-01 -5.20313561e-01 -3.58490497e-01
-4.88126308e-01 -8.10589492e-02 2.47526858e-02 -1.16269648e+00
1.85962200e+00 -1.93524563e+00 -7.03853071e-01 2.01808244e-01
4.01065856e-01 -2.81667024e-01 -7.41227940e-02 1.01469231e+00
-9.44178700e-02 2.73061216e-01 -1.62149772e-01 4.48152095e-01
-2.19642177e-01 1.61220923e-01 -1.52352124e-01 7.72329718e-02
2.94994920e-01 1.38178134e+00 -1.07621181e+00 -8.53285909e-01
-2.32690617e-01 1.98786139e-01 -7.74599791e-01 8.82051170e-01
-2.45956421e-01 6.09182298e-01 -7.00516999e-01 1.02400744e+00
3.18449229e-01 -7.08519757e-01 1.82057381e-01 -2.34880283e-01
4.46883179e-02 6.55226767e-01 -8.78174543e-01 1.40163517e+00
-2.45743632e-01 -8.51040855e-02 -1.37999617e-02 -5.97965300e-01
9.06229854e-01 7.64667809e-01 9.92442787e-01 -6.39759839e-01
8.97671282e-02 1.57982007e-01 -3.20104957e-01 -1.37760496e+00
1.29202679e-01 -5.45402756e-03 -5.53626239e-01 4.16454047e-01
-4.37680244e-01 -2.31700435e-01 -5.16019128e-02 1.13479100e-01
1.85779881e+00 -4.81564730e-01 5.96094668e-01 -1.47762209e-01
1.60151437e-01 3.47046673e-01 7.26649821e-01 7.12017059e-01
-2.86585271e-01 6.78280592e-01 6.71096921e-01 -7.09202588e-01
-7.26295650e-01 -1.11279607e+00 -3.94837976e-01 7.00277865e-01
-5.48182905e-01 -7.33006537e-01 -5.64694881e-01 -5.00126183e-01
-6.89477995e-02 7.53143549e-01 -4.95162606e-01 1.09765746e-01
-3.71980011e-01 -4.86918747e-01 7.99421489e-01 4.06970173e-01
1.07480228e-01 -9.98147845e-01 -1.06075871e+00 6.80325985e-01
-1.00566638e+00 -1.01591086e+00 -8.75341535e-01 2.31098831e-02
-8.62929165e-01 -1.47226918e+00 -3.80219519e-01 -5.58607578e-01
3.55095506e-01 -4.52137828e-01 1.62210429e+00 2.88756136e-02
-6.43056333e-01 7.65722275e-01 -3.92404705e-01 -7.63865173e-01
-7.35155165e-01 -1.02529682e-01 -5.13777554e-01 -4.80961233e-01
8.13630879e-01 9.17511582e-02 -1.04164040e+00 5.83076656e-01
-1.19490206e+00 -3.02372634e-01 9.93832871e-02 8.03755581e-01
5.42112470e-01 -4.35075164e-01 9.78358448e-01 -1.22567177e+00
1.07046604e+00 -9.06406581e-01 -3.29806715e-01 5.92887521e-01
-6.82347596e-01 -1.62696224e-02 4.94735479e-01 1.64117366e-01
-6.07595086e-01 8.10349286e-02 -6.93272650e-01 3.18290219e-02
-3.58784556e-01 9.66630042e-01 1.19973496e-02 8.77505243e-01
8.53567243e-01 7.58980736e-02 3.12853754e-01 -2.27259159e-01
1.75002083e-01 1.10231578e+00 3.90568316e-01 -5.84980130e-01
1.30686536e-01 1.33755282e-01 -5.65042913e-01 -2.59880334e-01
-7.98112094e-01 -7.28769779e-01 5.62489741e-02 1.42180279e-01
9.21356678e-01 -1.04445279e+00 -1.25773084e+00 4.72678691e-02
-1.08260655e+00 -4.20019507e-01 -7.42671609e-01 2.52256155e-01
-5.11997283e-01 2.12895781e-01 -9.95704710e-01 -7.47270286e-01
-6.69605017e-01 -9.66887951e-01 1.18085778e+00 -8.14624876e-02
-8.31209838e-01 -1.00709426e+00 2.39970163e-01 5.20500243e-01
6.24980628e-01 4.51283365e-01 1.15446639e+00 -9.16781366e-01
-3.43433976e-01 -3.63486797e-01 -8.88897255e-02 -1.63996190e-01
5.71442425e-01 -1.35629922e-01 -5.62342942e-01 -2.30631068e-01
4.94383514e-01 -6.07937813e-01 -5.24926111e-02 4.42172199e-01
1.35139716e+00 -7.18337774e-01 -1.48858801e-01 5.16014636e-01
1.17499840e+00 4.54639047e-01 7.80475378e-01 -1.06144659e-01
-2.72039045e-02 6.41612947e-01 8.60933244e-01 1.13747776e+00
1.22198761e+00 -4.76629753e-03 1.14191681e-01 -2.71540791e-01
7.69510806e-01 -2.22171456e-01 -4.95974533e-03 1.17034793e+00
3.33542734e-01 -7.53563195e-02 -1.54681480e+00 6.30253673e-01
-1.66441584e+00 -5.52604139e-01 -3.59997272e-01 2.05195570e+00
1.50265825e+00 -3.78534883e-01 1.06331751e-01 -3.39337111e-01
1.76373124e-01 -4.44861412e-01 -5.96166730e-01 -2.37774342e-01
1.90125868e-01 4.13176298e-01 3.16058546e-01 2.18192965e-01
-6.72408938e-01 2.81257570e-01 6.71227551e+00 -3.64579968e-02
-8.45912397e-01 6.12111203e-02 1.02503538e+00 1.18861578e-01
-5.97947896e-01 -2.87928939e-01 -4.26310837e-01 5.06394207e-01
1.61386347e+00 -3.41586888e-01 1.71133339e-01 6.78968310e-01
5.43147564e-01 9.04443040e-02 -1.55136645e+00 9.48765516e-01
-4.49256212e-01 -1.36529362e+00 -3.12243134e-01 -1.72222972e-01
5.55092275e-01 1.71992809e-01 -3.04314226e-01 3.45490873e-01
5.97518027e-01 -1.25815868e+00 1.19829506e-01 1.00665140e+00
1.14777946e+00 -7.11711794e-02 1.15796363e+00 3.22806984e-01
-9.97784376e-01 -1.00008398e-01 3.03968694e-02 2.12815106e-01
3.06251347e-01 6.69156730e-01 -1.35262847e+00 7.88688123e-01
8.78093004e-01 2.86186129e-01 -4.06297445e-01 9.20511782e-01
4.00111496e-01 6.82568550e-01 -3.12279135e-01 7.03798383e-02
-2.38853246e-01 1.12210356e-01 -1.67249113e-01 1.29927897e+00
5.61320066e-01 6.65202856e-01 5.44014312e-02 7.34586835e-01
3.71966884e-02 3.16948891e-01 -6.14994407e-01 -3.31169337e-01
7.29016602e-01 1.01277840e+00 -5.42263277e-02 -6.06338024e-01
-5.14683306e-01 3.71572554e-01 -8.89240205e-02 5.96303821e-01
-8.46925676e-01 -4.39955831e-01 6.91591322e-01 4.88182455e-01
-2.25032836e-01 2.09065259e-01 -3.13167959e-01 -1.13440442e+00
-1.41161531e-01 -1.60683668e+00 1.08709562e+00 -7.87822187e-01
-1.53213429e+00 6.69633210e-01 -1.64552554e-01 -1.07395172e+00
-6.53338075e-01 -2.28012830e-01 -8.57120156e-02 8.94045353e-01
-1.32519650e+00 -5.00925004e-01 -7.84040272e-01 1.03180051e+00
1.58646837e-01 3.07033271e-01 1.33273399e+00 7.58028090e-01
-2.72106826e-01 5.52835226e-01 -2.78627872e-01 3.99124771e-01
9.98942912e-01 -1.01871490e+00 3.07281882e-01 -1.32163048e-01
-7.17110217e-01 1.14730954e+00 5.07910132e-01 -7.85579920e-01
-1.73405647e+00 -1.33681488e+00 1.20179689e+00 -1.17949677e+00
4.02580380e-01 -7.87566602e-02 -1.05323029e+00 9.04002488e-01
-6.32534549e-02 -1.39703661e-01 1.28553367e+00 -2.81481534e-01
6.80044666e-02 -8.57489482e-02 -1.36438441e+00 2.76110679e-01
7.02633560e-01 -6.80834174e-01 -6.99902892e-01 4.99073625e-01
1.05517161e+00 -7.03302801e-01 -1.70226812e+00 6.80826724e-01
4.17789161e-01 -6.50646091e-01 7.56558180e-01 -1.23942852e+00
5.83507538e-01 -4.02615778e-02 -2.71280557e-01 -1.03230095e+00
4.20854315e-02 -8.24957371e-01 1.54388007e-02 6.63190484e-01
5.29107094e-01 -1.00114667e+00 6.06221616e-01 1.37984264e+00
-3.17785218e-02 -1.27010798e+00 -5.67809045e-01 -2.40476102e-01
-2.59853005e-01 -3.96657199e-01 1.09740984e+00 1.28367305e+00
2.30790704e-01 3.23886573e-02 8.20497200e-02 1.38851076e-01
1.66784644e-01 7.82265589e-02 7.55570352e-01 -1.03432333e+00
-1.63649365e-01 2.75304884e-01 1.77621648e-01 -1.03558314e+00
-7.91967750e-01 -5.89566588e-01 4.49138805e-02 -1.90107191e+00
1.92449808e-01 -7.10732818e-01 -1.08290613e-01 6.81322515e-01
-3.06308359e-01 -4.96524900e-01 -2.70526350e-01 7.75897503e-02
-4.78770256e-01 1.11333504e-01 1.15939081e+00 9.84526649e-02
-2.26097465e-01 1.46891594e-01 -9.73430097e-01 2.34616622e-01
6.08482957e-01 -5.13907731e-01 -2.29368329e-01 -3.25516433e-01
5.78608632e-01 1.24024916e+00 2.10003272e-01 -3.50755960e-01
3.90992582e-01 -4.32645708e-01 -3.91594470e-02 -6.34985209e-01
-1.04948670e-01 -9.17733550e-01 6.70355201e-01 7.68428147e-01
-4.31527674e-01 5.99778712e-01 6.86628819e-02 8.83059278e-02
-4.77494091e-01 1.08294256e-01 2.25729376e-01 -4.49639231e-01
1.04511574e-01 5.03680587e-01 -3.36938441e-01 7.82890439e-01
8.97397399e-01 1.63832337e-01 -5.11476099e-01 -8.31739962e-01
-8.26038480e-01 9.71781313e-01 3.28525841e-01 1.80405766e-01
8.11701775e-01 -9.73345518e-01 -9.50218558e-01 2.49282956e-01
4.79094386e-01 1.68227464e-01 3.97573948e-01 9.39844251e-01
-7.63325512e-01 5.56973100e-01 2.12923959e-01 -5.69404185e-01
-9.40918803e-01 5.88662386e-01 1.86599597e-01 -4.37304914e-01
-3.17421526e-01 3.61440718e-01 -7.02352598e-02 -8.74468923e-01
2.64139175e-01 -1.01798594e+00 2.99068153e-01 -8.32611620e-02
3.86310428e-01 1.35832846e-01 1.38091460e-01 1.22794934e-01
-5.17309725e-01 1.51962236e-01 3.08498800e-01 2.29589999e-01
1.34414089e+00 -9.69703048e-02 -2.96448767e-01 4.68103498e-01
1.27897322e+00 -8.63356367e-02 -4.82221156e-01 -1.52317226e-01
2.19366610e-01 -1.55781224e-01 -6.65156901e-01 -9.39252615e-01
-5.84377050e-01 4.13823724e-01 3.93149793e-01 5.04567027e-01
1.31485701e+00 2.63811320e-01 1.11769986e+00 4.99915212e-01
2.42686391e-01 -7.96602845e-01 -5.85655943e-02 3.48974437e-01
8.81370485e-01 -1.40154445e+00 -3.90197396e-01 -2.35761628e-01
-8.11013818e-01 9.81925547e-01 3.84023130e-01 7.13299155e-01
8.91583681e-01 5.63566983e-01 7.49537051e-01 -6.76710069e-01
-1.29842448e+00 2.43103147e-01 -2.45182812e-01 2.69260883e-01
7.95534670e-01 2.66586453e-01 -2.10613206e-01 9.03072238e-01
-3.93439591e-01 7.51843810e-01 5.47992408e-01 1.14810622e+00
9.78367999e-02 -1.03747487e+00 -4.50216025e-01 1.11898232e+00
-7.69047558e-01 -3.11250418e-01 1.01272382e-01 2.87443936e-01
-1.84945256e-01 1.14051259e+00 -1.50750667e-01 -2.13404313e-01
9.16100264e-01 3.61715674e-01 -1.61931813e-01 -7.17772543e-01
-1.01185524e+00 -2.36244932e-01 5.02054870e-01 -7.70663857e-01
1.05072536e-01 -6.53460503e-01 -1.48642635e+00 -8.32046419e-02
2.03071639e-01 4.60980296e-01 4.89230037e-01 4.21316445e-01
8.43174696e-01 5.67463338e-01 5.58130264e-01 5.25513470e-01
-1.08409834e+00 -6.97752357e-01 -1.09399483e-01 7.41101384e-01
5.86641908e-01 2.46708378e-01 2.90325247e-02 2.28171930e-01] | [8.719667434692383, 8.488037109375] |
e5bb5b1e-263e-4209-96e9-d032a4529238 | task-discrepancy-maximization-for-fine-1 | 2207.01376 | null | https://arxiv.org/abs/2207.01376v1 | https://arxiv.org/pdf/2207.01376v1.pdf | Task Discrepancy Maximization for Fine-grained Few-Shot Classification | Recognizing discriminative details such as eyes and beaks is important for distinguishing fine-grained classes since they have similar overall appearances. In this regard, we introduce Task Discrepancy Maximization (TDM), a simple module for fine-grained few-shot classification. Our objective is to localize the class-wise discriminative regions by highlighting channels encoding distinct information of the class. Specifically, TDM learns task-specific channel weights based on two novel components: Support Attention Module (SAM) and Query Attention Module (QAM). SAM produces a support weight to represent channel-wise discriminative power for each class. Still, since the SAM is basically only based on the labeled support sets, it can be vulnerable to bias toward such support set. Therefore, we propose QAM which complements SAM by yielding a query weight that grants more weight to object-relevant channels for a given query image. By combining these two weights, a class-wise task-specific channel weight is defined. The weights are then applied to produce task-adaptive feature maps more focusing on the discriminative details. Our experiments validate the effectiveness of TDM and its complementary benefits with prior methods in fine-grained few-shot classification. | ['Jae-Pil Heo', 'WonJun Moon', 'SuBeen Lee'] | 2022-07-04 | task-discrepancy-maximization-for-fine | http://openaccess.thecvf.com//content/CVPR2022/html/Lee_Task_Discrepancy_Maximization_for_Fine-Grained_Few-Shot_Classification_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Lee_Task_Discrepancy_Maximization_for_Fine-Grained_Few-Shot_Classification_CVPR_2022_paper.pdf | cvpr-2022-1 | ['few-shot-image-classification'] | ['computer-vision'] | [ 2.65097201e-01 -2.18516991e-01 -3.81033778e-01 -5.56007206e-01
-8.16842139e-01 -1.95875227e-01 6.62262201e-01 1.07732028e-01
-2.21847519e-01 6.72874749e-01 3.52510303e-01 3.19859952e-01
-1.36566922e-01 -7.18096197e-01 -4.57877934e-01 -7.76989162e-01
7.18178749e-02 -7.25587830e-02 5.29503584e-01 6.77312315e-02
5.01315534e-01 3.74096811e-01 -1.80003917e+00 6.01192236e-01
7.60609329e-01 1.43091214e+00 6.00789189e-01 3.20241630e-01
-3.48200470e-01 5.58117688e-01 -4.42316115e-01 -1.10424802e-01
2.44166851e-01 -5.57142079e-01 -5.23192108e-01 2.64155477e-01
6.58490241e-01 -2.45783135e-01 -1.84767663e-01 1.24103594e+00
4.90751386e-01 3.07132155e-01 9.37946558e-01 -1.23508191e+00
-6.18510306e-01 2.45497808e-01 -7.90143311e-01 5.84999204e-01
6.97715953e-02 1.96989343e-01 1.30841196e+00 -1.03877997e+00
4.29807276e-01 1.16557837e+00 2.79332280e-01 4.68112409e-01
-1.20493698e+00 -6.04885221e-01 4.29941595e-01 3.13903421e-01
-1.52434599e+00 -5.89228332e-01 1.00019312e+00 -5.18974423e-01
6.17137790e-01 2.59553283e-01 5.69601059e-01 7.85218060e-01
2.24088609e-01 8.04822922e-01 1.18246865e+00 -2.45945364e-01
5.31159222e-01 2.40294740e-01 2.17628494e-01 7.35101819e-01
2.16374114e-01 -1.23559758e-01 -7.82429099e-01 1.78040657e-02
8.02946448e-01 2.21125841e-01 -2.82126635e-01 -5.41356802e-01
-1.05636299e+00 8.60208094e-01 5.53950608e-01 3.28505635e-01
-3.82301956e-01 1.13784760e-01 2.61546254e-01 4.21374440e-02
4.55894738e-01 3.45466107e-01 -3.09120327e-01 1.34488136e-01
-1.05436277e+00 2.88385153e-02 3.44626635e-01 8.13616037e-01
1.24369383e+00 -3.75498384e-02 -1.09743464e+00 1.05740392e+00
3.91685337e-01 1.44579187e-01 5.14863312e-01 -6.13035858e-01
1.77010596e-01 6.78309679e-01 -5.86450361e-02 -9.86610651e-01
-1.53556734e-01 -6.38579428e-01 -5.89606225e-01 1.65580854e-01
2.18180656e-01 1.66582912e-01 -1.05691385e+00 1.80513537e+00
2.58674771e-01 2.52306879e-01 -2.72183806e-01 1.19286227e+00
1.02113509e+00 4.32130039e-01 3.26960921e-01 -7.60658979e-02
1.57277119e+00 -9.55870807e-01 -4.20303971e-01 -3.79574001e-01
2.72291273e-01 -6.29439592e-01 1.22384369e+00 -2.10007891e-01
-7.30351746e-01 -7.50148475e-01 -1.14210236e+00 6.34085760e-02
-3.01101059e-01 8.88204724e-02 6.03617311e-01 4.54524994e-01
-8.21703136e-01 4.30402070e-01 -1.83054626e-01 -4.35961396e-01
7.30648160e-01 -8.82815272e-02 -1.26488328e-01 -7.54989907e-02
-1.06100857e+00 6.71571732e-01 2.14978606e-01 -2.49032870e-01
-1.13800430e+00 -6.27786994e-01 -1.00767028e+00 2.45168209e-01
2.29971141e-01 -6.79775536e-01 9.31546032e-01 -8.41888011e-01
-1.23482978e+00 1.03138661e+00 -3.40159565e-01 -1.70953467e-01
1.49151087e-01 2.32188985e-01 -3.24877918e-01 2.74159521e-01
4.78386819e-01 8.29173028e-01 1.49207997e+00 -1.18005764e+00
-8.20991933e-01 -4.28915501e-01 1.32382050e-01 2.40757629e-01
-4.36248153e-01 -8.90450552e-02 -5.74769199e-01 -8.10155213e-01
-9.74432006e-03 -3.42537075e-01 -3.35071236e-02 2.66542494e-01
-4.41254944e-01 -2.85422266e-01 7.69175947e-01 -2.84377545e-01
1.36404634e+00 -2.33859301e+00 -1.45431861e-01 8.54182914e-02
3.93617183e-01 4.78014387e-02 -4.13506329e-01 1.72323287e-01
1.69120148e-01 -2.13683769e-01 -1.76041827e-01 -1.49372652e-01
-1.02818036e-03 -2.52083372e-02 -8.79751220e-02 3.87755960e-01
6.79365516e-01 1.12753808e+00 -7.79721916e-01 -6.70462728e-01
3.11363161e-01 1.82526916e-01 -3.01818311e-01 1.55336514e-01
-1.56161413e-01 2.07142651e-01 -6.82838976e-01 8.90632629e-01
7.11588979e-01 -3.15373302e-01 -2.90467620e-01 -5.50868094e-01
-1.11065909e-01 -1.46631464e-01 -1.13339770e+00 1.66166556e+00
-4.07197267e-01 5.45160949e-01 6.70224428e-03 -1.04542494e+00
1.16197026e+00 -1.34880856e-01 4.45637144e-02 -8.80138278e-01
1.73758313e-01 -1.00218272e-02 -7.94188753e-02 -4.88656431e-01
4.32822675e-01 -2.87571818e-01 -1.51807591e-01 3.65046412e-01
2.42045105e-01 8.32376406e-02 -2.82487907e-02 2.63600290e-01
8.73391032e-01 -7.89730027e-02 5.53826928e-01 -4.80281413e-01
3.95158261e-01 -4.04492497e-01 6.94176316e-01 8.37903976e-01
-5.95376134e-01 6.98938429e-01 1.99274376e-01 -2.32045546e-01
-8.42733026e-01 -1.06201947e+00 -2.82784820e-01 1.48285043e+00
6.30275369e-01 -2.58232832e-01 -5.20122349e-01 -8.21136415e-01
9.89509150e-02 3.75258267e-01 -1.03637588e+00 -3.59558731e-01
-1.06741367e-02 -4.10194427e-01 8.54697153e-02 4.86565441e-01
7.01209247e-01 -1.10793281e+00 -8.28904867e-01 3.32452431e-02
1.42250195e-01 -7.45589197e-01 -7.92334497e-01 4.08559799e-01
-5.89706063e-01 -9.78159189e-01 -1.20588672e+00 -8.25620294e-01
6.92869782e-01 8.66113782e-01 9.63619351e-01 -2.10442543e-01
-5.66055655e-01 4.26699072e-01 -4.08898741e-01 -2.55494297e-01
2.47028425e-01 -2.02498734e-01 -1.90414026e-01 5.91705203e-01
4.95413095e-01 -4.72839266e-01 -9.59840477e-01 3.39725405e-01
-7.23541141e-01 2.40305066e-02 9.37485576e-01 8.43120217e-01
7.25573838e-01 -6.06378168e-02 7.14775205e-01 -8.16957831e-01
5.88993609e-01 -4.69849616e-01 -1.89380467e-01 2.87085235e-01
-3.20688099e-01 1.29494473e-01 4.32764202e-01 -3.78011584e-01
-1.06370544e+00 4.79311273e-02 2.34784439e-01 -5.47432840e-01
-3.70949000e-01 1.94137663e-01 -3.26854736e-01 -1.36851773e-01
6.12539530e-01 3.65378320e-01 -2.41616979e-01 -4.74418759e-01
4.00811464e-01 7.07681298e-01 2.11851791e-01 -5.04669964e-01
5.95337152e-01 4.69236672e-01 -1.44811183e-01 -8.59412789e-01
-1.10054171e+00 -8.14552844e-01 -4.92528170e-01 -3.20164859e-01
9.97796237e-01 -9.05863881e-01 -1.75503209e-01 3.61309230e-01
-8.28470409e-01 -1.37636378e-01 -5.22878051e-01 3.63307387e-01
-4.58655328e-01 2.90438771e-01 -2.64774054e-01 -7.46223509e-01
-2.06915632e-01 -9.39476371e-01 1.40811849e+00 5.15343785e-01
-1.75819933e-01 -7.80665338e-01 -7.37372115e-02 6.09529437e-03
4.47546959e-01 -1.20489402e-02 8.59517038e-01 -4.31959718e-01
-4.77194786e-01 -1.46608829e-01 -7.33544648e-01 2.45420173e-01
1.07748300e-01 -2.61128128e-01 -1.41591179e+00 -2.78629810e-01
-2.05695525e-01 -3.72673303e-01 1.36894608e+00 5.94758987e-01
1.27071965e+00 -3.34913768e-02 -3.54927838e-01 6.03026569e-01
1.48658967e+00 3.70007865e-02 5.02581775e-01 2.11795226e-01
6.22994840e-01 5.49151719e-01 9.24781263e-01 5.64348876e-01
6.77769631e-02 7.33621538e-01 3.00105542e-01 -1.61467046e-01
-4.71616238e-01 -2.89499730e-01 2.30806127e-01 2.18147248e-01
6.42885342e-02 1.82439893e-01 -2.43825763e-01 5.86711884e-01
-1.70850706e+00 -1.01443815e+00 2.31162265e-01 2.18854618e+00
6.34092987e-01 1.54902861e-01 1.20866425e-01 -1.16439104e-01
9.23533738e-01 5.49571335e-01 -7.82323480e-01 -1.86775208e-01
-1.69961169e-01 1.94252804e-01 1.26221865e-01 5.54215834e-02
-1.17482364e+00 8.57771993e-01 5.53873920e+00 1.47587228e+00
-1.16487443e+00 1.13590427e-01 6.35786414e-01 -1.11726823e-03
-5.48927367e-01 -5.99404201e-02 -9.63248968e-01 6.47056162e-01
2.94050395e-01 -8.11999813e-02 1.72364548e-01 9.20467496e-01
1.12213276e-01 -2.98489839e-01 -9.52917874e-01 1.08528471e+00
1.82311162e-01 -1.33004177e+00 2.89468259e-01 -3.47517878e-02
7.90221095e-01 -2.78162777e-01 1.53916687e-01 4.07254606e-01
-1.95768490e-01 -8.09546709e-01 8.43673825e-01 6.16615951e-01
1.03819835e+00 -5.99051416e-01 4.94900137e-01 2.16505215e-01
-1.49729025e+00 -1.80793285e-01 -8.24103594e-01 -1.18066244e-01
-9.35162231e-02 7.40713656e-01 -5.20371377e-01 2.01535568e-01
6.03885412e-01 8.91538501e-01 -6.85562253e-01 1.18845201e+00
-1.89105600e-01 3.40560794e-01 2.38814890e-01 -6.69580176e-02
2.29120597e-01 -3.79275419e-02 4.62886631e-01 1.40408254e+00
1.99009091e-01 8.80778879e-02 2.08421394e-01 9.91240680e-01
4.46304567e-02 7.09946677e-02 -5.58101296e-01 1.13234363e-01
6.54329419e-01 1.64663696e+00 -8.68765116e-01 -3.29206586e-01
-5.42724967e-01 1.31433904e+00 4.29242820e-01 4.77031201e-01
-7.02550709e-01 -5.84151983e-01 8.08989108e-01 1.12433195e-01
5.92092693e-01 1.78962395e-01 -2.10948244e-01 -1.09831572e+00
-1.41083404e-01 -3.79965454e-01 4.71205771e-01 -6.90084040e-01
-1.60700023e+00 4.85507816e-01 -2.92680740e-01 -1.51594937e+00
3.26575786e-01 -4.08493221e-01 -9.22877491e-01 9.42300797e-01
-1.80940795e+00 -1.30770350e+00 -6.82814419e-01 7.89420545e-01
8.51383209e-01 -1.74460784e-01 7.84533203e-01 2.16911092e-01
-6.48686528e-01 8.27323854e-01 -1.83351099e-01 -1.35664627e-01
8.56208026e-01 -1.15935481e+00 2.24642046e-02 7.07437396e-01
1.14299089e-01 5.50162375e-01 4.27949458e-01 -5.42028308e-01
-9.89690721e-01 -1.31850028e+00 5.61823010e-01 -1.02370165e-01
4.99627233e-01 -2.60158747e-01 -8.71337235e-01 9.24224854e-02
-2.53076226e-01 2.44013041e-01 6.68138266e-01 1.48696736e-01
-5.52822828e-01 -2.77064711e-01 -1.07656026e+00 2.05788478e-01
1.04568863e+00 -8.22115958e-01 -5.17938673e-01 -3.96941155e-02
5.61775446e-01 1.93066105e-01 -6.00174069e-01 1.99630931e-01
6.11854613e-01 -1.06663644e+00 8.46676290e-01 -4.49385583e-01
5.17037809e-01 -5.22871673e-01 -4.08702672e-01 -1.25607145e+00
-9.23324227e-01 -2.99489312e-02 -1.45511970e-01 1.35793483e+00
1.76228181e-01 -2.83296943e-01 6.30632401e-01 1.80503622e-01
-2.48800144e-01 -8.34377825e-01 -8.66436899e-01 -7.43789434e-01
-3.91064793e-01 -3.25419366e-01 4.58068281e-01 1.03207290e+00
-2.50713795e-01 5.18948078e-01 -4.81811643e-01 7.21620815e-03
5.99625468e-01 6.57581329e-01 5.52592218e-01 -1.22946954e+00
-3.16285133e-01 -5.04032254e-01 -6.56690478e-01 -1.12675858e+00
1.90107420e-03 -8.91671777e-01 1.79252312e-01 -1.57781470e+00
8.41323614e-01 -4.43842620e-01 -7.91790664e-01 3.70399147e-01
-4.72650856e-01 7.21570790e-01 1.65955752e-01 3.34670752e-01
-1.01635790e+00 6.43696845e-01 1.40788400e+00 -3.44954431e-01
-6.95922077e-02 1.17865568e-02 -9.79801178e-01 5.82595050e-01
5.65937281e-01 -1.65725172e-01 -3.18902194e-01 -2.58242916e-02
-2.89268434e-01 -3.63855809e-01 5.21263897e-01 -9.84788477e-01
6.03279360e-02 -2.63156116e-01 7.04038978e-01 -4.77534413e-01
3.81735742e-01 -4.67042863e-01 -3.42770189e-01 3.39992225e-01
-3.46438229e-01 -8.45924675e-01 4.10263017e-02 7.60112107e-01
-3.26903820e-01 -1.97875351e-01 1.12033784e+00 -1.94891989e-01
-1.56414020e+00 5.04531324e-01 -1.65388033e-01 9.03823376e-02
1.22901964e+00 -6.23328507e-01 -2.28648379e-01 -2.72228748e-01
-6.53475463e-01 1.62963450e-01 3.37118149e-01 4.21925604e-01
8.52009714e-01 -1.42508817e+00 -6.58573270e-01 5.24603844e-01
8.34857047e-01 -3.91482741e-01 6.84238672e-01 6.66312099e-01
2.08588868e-01 3.70463997e-01 -5.33893228e-01 -5.66745877e-01
-1.15609515e+00 5.38594604e-01 1.50388971e-01 1.86367452e-01
-3.95253807e-01 1.23828375e+00 9.34902549e-01 1.86835945e-01
2.04168379e-01 -2.13982612e-01 -4.61678505e-01 3.57860386e-01
7.37908065e-01 3.39785606e-01 -2.44664699e-01 -5.70979714e-01
-3.77078623e-01 7.96695769e-01 -1.11712120e-01 1.82712644e-01
1.15888107e+00 -3.81784856e-01 1.56373188e-01 5.53812265e-01
1.13147914e+00 2.08395552e-02 -1.66433656e+00 -4.29764479e-01
-2.69556046e-01 -7.62280285e-01 2.79792786e-01 -7.45516598e-01
-9.01346743e-01 1.05469787e+00 7.94088125e-01 1.80871990e-02
1.24084902e+00 4.56822664e-01 5.16959071e-01 -1.77109495e-01
3.81255418e-01 -1.06181228e+00 3.80873680e-01 3.36789668e-01
7.41479814e-01 -1.37008870e+00 -2.11147964e-01 -3.52556348e-01
-6.41217291e-01 8.75771701e-01 7.54835546e-01 -5.15433997e-02
4.78213161e-01 -1.29334822e-01 -1.59353867e-01 -3.82973492e-01
-4.25145715e-01 -7.11400867e-01 8.19884658e-01 7.70974100e-01
2.90978789e-01 1.62863195e-01 -3.34886879e-01 8.85671079e-01
3.15987408e-01 -1.53774321e-01 -8.46932754e-02 6.62770867e-01
-1.03724778e+00 -6.93432868e-01 -1.48918822e-01 8.40927124e-01
-1.10673308e-01 -2.39422008e-01 -3.61041248e-01 3.52490216e-01
4.06609416e-01 7.16413736e-01 3.07863533e-01 -4.95171279e-01
8.21533650e-02 -1.20241202e-01 5.19181788e-01 -9.40246820e-01
-8.42742622e-02 1.17811278e-01 -7.15370476e-02 -6.81689441e-01
-2.51194745e-01 -3.06010276e-01 -1.00616837e+00 4.79868166e-02
-3.50103438e-01 9.32499766e-03 2.94016451e-01 8.07048261e-01
3.68629277e-01 6.37640655e-01 7.31725454e-01 -9.72141266e-01
-3.36620599e-01 -9.14697886e-01 -1.07491684e+00 5.04858196e-01
2.77200967e-01 -9.57690835e-01 -2.44894415e-01 -2.48953179e-01] | [9.77492618560791, 2.0416064262390137] |
9aff1aac-1e88-4e11-a1c0-e66a8413008b | video-killed-the-hd-map-predicting-driving | 2305.11856 | null | https://arxiv.org/abs/2305.11856v1 | https://arxiv.org/pdf/2305.11856v1.pdf | Video Killed the HD-Map: Predicting Driving Behavior Directly From Drone Images | The development of algorithms that learn behavioral driving models using human demonstrations has led to increasingly realistic simulations. In general, such models learn to jointly predict trajectories for all controlled agents by exploiting road context information such as drivable lanes obtained from manually annotated high-definition (HD) maps. Recent studies show that these models can greatly benefit from increasing the amount of human data available for training. However, the manual annotation of HD maps which is necessary for every new location puts a bottleneck on efficiently scaling up human traffic datasets. We propose a drone birdview image-based map (DBM) representation that requires minimal annotation and provides rich road context information. We evaluate multi-agent trajectory prediction using the DBM by incorporating it into a differentiable driving simulator as an image-texture-based differentiable rendering module. Our results demonstrate competitive multi-agent trajectory prediction performance when using our DBM representation as compared to models trained with rasterized HD maps. | ['Frank Wood', 'Adam Ścibior', 'Berend Zwartsenberg', 'Xiaoxuan Liang', 'Dylan Green', 'Setareh Dabiri', 'Justice Sefas', 'Matthew Niedoba', 'Jonathan Wilder Lavington', 'Vasileios Lioutas', 'Yunpeng Liu'] | 2023-05-19 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [-2.59490401e-01 1.44319190e-02 4.43127975e-02 -7.22241402e-01
-5.76850593e-01 -6.39297068e-01 9.43764031e-01 8.37443769e-02
-5.33075094e-01 8.93532574e-01 8.69014859e-02 -5.20321906e-01
1.04714386e-01 -1.17404819e+00 -9.66299295e-01 -1.99985147e-01
-3.83248210e-01 9.34696913e-01 7.25886345e-01 -5.92180431e-01
1.75692737e-02 7.88594246e-01 -2.12747312e+00 1.08739987e-01
1.14716852e+00 3.42977375e-01 5.87345004e-01 1.01313043e+00
2.78386474e-01 9.34485018e-01 -3.71031642e-01 4.45980579e-02
4.64428216e-01 2.10443400e-02 -3.19742829e-01 -2.71116287e-01
5.83417892e-01 -7.67947137e-01 -6.97162509e-01 4.33739483e-01
2.42835909e-01 5.45506716e-01 4.79174137e-01 -1.82932985e+00
-9.15613864e-03 -3.01443618e-02 -3.01598422e-02 2.15428799e-01
1.91289127e-01 6.32491887e-01 4.30893838e-01 -4.08946604e-01
8.82912934e-01 9.17109966e-01 6.58741713e-01 2.86902726e-01
-1.34559774e+00 -5.90973377e-01 1.77128017e-01 4.20934558e-01
-1.38646603e+00 -1.87916815e-01 4.24691558e-01 -6.96782649e-01
1.25717318e+00 3.61869097e-01 9.26574290e-01 7.99630165e-01
2.65868783e-01 5.94613433e-01 1.03279257e+00 2.41317913e-01
1.37053221e-01 1.93000570e-01 -7.65364692e-02 1.14799678e+00
-1.09206744e-01 6.75872147e-01 -2.99606323e-01 -1.55426590e-02
9.50373888e-01 -2.10568398e-01 1.28824115e-01 -6.39940560e-01
-1.35595357e+00 5.12082577e-01 5.64212024e-01 -5.59139013e-01
-5.64135611e-01 5.11515081e-01 3.27482432e-01 2.04004407e-01
3.50949436e-01 2.49417692e-01 -1.24602623e-01 -6.10500872e-01
-8.22072804e-01 1.07223821e+00 7.12811947e-01 1.37570691e+00
1.30714643e+00 1.21090926e-01 1.18646197e-01 4.44320679e-01
-1.15175053e-01 9.47295547e-01 -2.33190134e-01 -1.38498914e+00
4.81934160e-01 5.68986714e-01 5.98892272e-01 -1.04586971e+00
-6.21137559e-01 -7.56241828e-02 -3.55819583e-01 8.32419097e-01
2.93265790e-01 -3.52828205e-01 -9.12661850e-01 1.58251691e+00
4.58475024e-01 4.90964144e-01 -1.96366962e-02 1.03698945e+00
5.32610357e-01 9.33097899e-01 3.76616687e-01 6.50152862e-01
6.27726555e-01 -1.20770109e+00 -2.79882252e-01 -1.71444193e-01
9.98262048e-01 -1.52474895e-01 1.06492043e+00 1.02072619e-01
-7.32037663e-01 -8.51306498e-01 -9.55071688e-01 -1.15005232e-01
-7.77657688e-01 7.75128528e-02 6.39419615e-01 5.57251871e-01
-1.28778827e+00 3.51436794e-01 -1.03514194e+00 -2.58908212e-01
2.15778723e-01 5.35910666e-01 -3.71933967e-01 -8.40523466e-02
-9.05104458e-01 1.39112580e+00 1.19751781e-01 -2.08362594e-01
-1.46048713e+00 -9.95124102e-01 -9.24082339e-01 -2.45970860e-01
1.88534379e-01 -5.76081574e-01 1.21121359e+00 -4.73710746e-01
-1.40525496e+00 4.91612107e-01 -2.72902727e-01 -6.39791369e-01
8.84012401e-01 -5.39828949e-02 -5.61172962e-01 -5.04261442e-02
1.46874934e-01 1.16134572e+00 3.52397293e-01 -1.34717786e+00
-1.35960269e+00 1.53456712e-02 4.14008111e-01 3.54659051e-01
4.74083394e-01 -3.89942557e-01 -2.12214634e-01 -1.03166156e-01
-6.17868900e-01 -1.18739998e+00 -5.71963072e-01 1.00761071e-01
-6.19411878e-02 -1.89842526e-02 1.31202257e+00 -8.42045248e-01
9.03489590e-01 -1.75279272e+00 -1.20752409e-01 4.40511584e-01
8.24023560e-02 1.35399833e-01 -1.89326927e-01 6.31986856e-01
5.50406158e-01 -3.23521614e-01 -9.46072936e-02 -2.64641672e-01
3.44211668e-01 5.67204952e-01 -4.80244458e-01 3.23762357e-01
-7.26261586e-02 9.91944671e-01 -1.04171801e+00 -3.36912155e-01
7.23053992e-01 4.99183446e-01 -6.54182911e-01 3.25468332e-01
-4.68882531e-01 7.51291811e-01 -3.66232038e-01 1.89580768e-01
6.29088342e-01 1.48571014e-01 2.33895071e-02 1.74405932e-01
-8.20962608e-01 3.70472461e-01 -8.17611635e-01 1.73379564e+00
-6.38332725e-01 1.02045822e+00 -8.72893035e-02 -5.80165207e-01
9.46911216e-01 -1.04308464e-02 4.12108809e-01 -8.95846128e-01
-3.99440259e-01 2.22664904e-02 -2.17283487e-01 -5.11188328e-01
9.57455873e-01 3.28144789e-01 -2.41351902e-01 4.17113483e-01
-4.15895581e-01 -5.70288301e-01 2.00281113e-01 1.28882602e-01
1.07792211e+00 6.25219703e-01 -2.34856516e-01 -3.07233363e-01
2.84556121e-01 1.16945410e+00 4.41129118e-01 6.56439483e-01
-2.80592173e-01 2.13114604e-01 1.46213055e-01 -9.97032881e-01
-1.62934566e+00 -1.21112978e+00 2.38172993e-01 1.15639341e+00
4.42390531e-01 -4.13603157e-01 -8.30274761e-01 -5.29857516e-01
2.66808063e-01 1.13160670e+00 -4.54982251e-01 1.64328143e-01
-1.12457204e+00 -4.30130899e-01 6.15979850e-01 6.33274317e-01
6.48607075e-01 -7.43816495e-01 -1.26251888e+00 3.68285149e-01
-7.43492879e-03 -9.84695494e-01 -1.48805916e-01 -1.42597154e-01
-4.36270624e-01 -8.82262766e-01 -2.91732043e-01 -6.86537921e-01
5.76322675e-01 3.52059394e-01 9.83584166e-01 2.38769408e-02
-1.43258378e-01 6.40512407e-01 1.95541158e-02 -5.82426071e-01
-5.35194218e-01 1.57377422e-01 -2.46962830e-02 -5.41194916e-01
2.52052993e-01 -6.51116431e-01 -5.25062680e-01 7.33238816e-01
-3.24273884e-01 9.11353528e-01 2.26204067e-01 3.06467265e-01
6.16150081e-01 -9.85534638e-02 5.57984829e-01 -6.00711048e-01
4.32556182e-01 -5.34495294e-01 -1.12654662e+00 2.69988570e-02
-3.69935572e-01 -4.27152254e-02 5.92903435e-01 -1.40684918e-01
-1.20426631e+00 3.45375538e-01 -4.35009003e-02 -1.39006019e-01
-5.80423832e-01 3.35793853e-01 1.64449930e-01 -2.39450768e-01
5.35302877e-01 2.31805876e-01 1.90465316e-01 -2.60688774e-02
5.85281968e-01 5.01906693e-01 7.44595170e-01 -6.12907290e-01
8.44314337e-01 5.76109231e-01 1.44861713e-01 -1.07586014e+00
-1.40877953e-02 -2.17911303e-01 -8.26743484e-01 -7.32499301e-01
8.55188668e-01 -1.05632222e+00 -1.06091344e+00 2.93344766e-01
-8.70168984e-01 -1.33087528e+00 -8.27537552e-02 5.12634277e-01
-1.09553289e+00 -1.20867372e-01 -1.81502804e-01 -6.12417698e-01
3.88882399e-01 -1.38268232e+00 9.64675188e-01 5.61239123e-02
-2.07782254e-01 -1.00277185e+00 3.63225460e-01 1.51589006e-01
5.83829463e-01 5.54684699e-01 9.00717318e-01 -1.31854236e-01
-1.31703949e+00 -2.09007680e-01 -1.91293418e-01 -2.66450435e-01
-1.59902856e-01 -2.64561534e-01 -7.44179308e-01 -2.13823058e-02
-8.30252588e-01 -1.59393445e-01 3.79781693e-01 2.09309071e-01
9.59823370e-01 1.07734771e-02 -7.28945851e-01 4.59186375e-01
1.23805964e+00 2.44112626e-01 5.62572896e-01 5.46711087e-01
8.32769334e-01 7.53705323e-01 9.86199617e-01 2.60227591e-01
1.29285753e+00 9.26180065e-01 4.20705974e-01 -3.44446927e-01
-1.29567087e-01 -6.16773307e-01 1.49029329e-01 2.74595976e-01
-3.69265199e-01 -2.06435367e-01 -1.39333391e+00 6.50084555e-01
-2.18120766e+00 -1.15604293e+00 -3.61158878e-01 2.21700621e+00
1.95725977e-01 -8.68446529e-02 5.05626678e-01 -5.37008524e-01
4.59957898e-01 7.92628601e-02 -7.19447613e-01 -3.25372368e-01
6.53725713e-02 -2.17522159e-01 9.58640575e-01 8.72066498e-01
-1.00703585e+00 1.19990659e+00 6.93874836e+00 3.82373840e-01
-8.49059165e-01 7.93779790e-02 2.05253974e-01 -1.25994667e-01
-2.52948135e-01 -8.81357938e-02 -8.41830850e-01 2.37484619e-01
1.41044521e+00 -4.09764230e-01 5.95977962e-01 9.45703447e-01
8.38932514e-01 -3.50115180e-01 -1.06183577e+00 7.58445144e-01
-2.61497676e-01 -1.64740062e+00 -8.18330944e-02 3.86023521e-01
7.56667018e-01 4.72783387e-01 1.42143667e-01 6.33926868e-01
9.09496188e-01 -1.05627668e+00 8.62631798e-01 7.81737626e-01
6.22880042e-01 -9.76647377e-01 2.75948465e-01 7.48666286e-01
-1.37037241e+00 3.81515399e-02 -2.77126014e-01 -1.66475683e-01
5.39151013e-01 -3.32713246e-01 -1.34517622e+00 3.65766197e-01
4.99653757e-01 8.00253570e-01 -4.32943314e-01 1.12287962e+00
3.44099365e-02 3.68331343e-01 -5.14050782e-01 8.14287364e-03
6.19089007e-01 -4.24514264e-01 5.48792422e-01 1.18470240e+00
2.72806674e-01 1.45026833e-01 5.54856479e-01 9.16119874e-01
4.85252589e-01 -2.68804163e-01 -1.06974757e+00 3.44677359e-01
3.15582126e-01 8.66572380e-01 -3.54683012e-01 -5.81061482e-01
-4.36169356e-01 8.27202797e-01 4.86664981e-01 4.63136733e-01
-1.19898796e+00 -1.13700740e-01 1.26820016e+00 4.83166337e-01
6.71233907e-02 -9.65273321e-01 -1.69035923e-02 -5.81338823e-01
-3.46006781e-01 -4.70039815e-01 -1.61669955e-01 -9.39277768e-01
-5.53596199e-01 5.62463224e-01 3.85272205e-01 -1.33744395e+00
-5.81273079e-01 -3.89478683e-01 -4.21896398e-01 1.01400852e+00
-1.68557489e+00 -1.50763333e+00 -7.66421914e-01 4.04001027e-01
5.43385863e-01 -1.59884274e-01 8.06825042e-01 3.67839247e-01
-1.96834043e-01 3.54507305e-02 3.29557359e-02 -2.24007204e-01
3.58912319e-01 -1.34966338e+00 1.01629066e+00 5.88950813e-01
-2.55256414e-01 1.26252636e-01 7.69208014e-01 -9.27257121e-01
-1.32502508e+00 -1.45510232e+00 4.61462647e-01 -8.28612924e-01
3.59281063e-01 -4.64689642e-01 -8.24829757e-01 1.11589086e+00
5.43747796e-03 -1.56967357e-01 3.14396054e-01 -2.31311575e-01
1.07904136e-01 -8.76365229e-02 -9.58656907e-01 8.51850033e-01
1.35588646e+00 -5.02472162e-01 -1.70357615e-01 2.43313864e-01
4.57895309e-01 -5.99008977e-01 -6.98920488e-01 7.37116039e-02
6.16783023e-01 -1.02841413e+00 1.10239017e+00 -4.93737787e-01
1.25229418e-01 -7.06196249e-01 -8.86135101e-02 -1.69791818e+00
-3.02282542e-01 -3.52564871e-01 2.32584149e-01 4.39800948e-01
4.79057878e-01 -6.20159566e-01 1.00653970e+00 1.05358613e+00
-7.16450930e-01 -2.82799810e-01 -6.83431506e-01 -7.81213284e-01
-7.37124607e-02 -4.73526001e-01 1.11016917e+00 5.89355707e-01
-8.43087304e-03 -2.70462781e-01 -4.74344999e-01 6.31011188e-01
5.31080663e-01 6.67507574e-02 1.63544607e+00 -1.08540440e+00
3.05819422e-01 -1.92447543e-01 -6.83174193e-01 -1.34457493e+00
3.92138988e-01 -8.67039502e-01 2.23158658e-01 -1.75500619e+00
-3.59673858e-01 -1.09792113e+00 3.57373059e-01 2.95669913e-01
1.90853924e-01 -2.86033172e-02 2.86749788e-02 9.94178876e-02
-5.22916913e-01 5.37728608e-01 1.13955140e+00 -2.93989033e-02
-3.23884547e-01 -1.14609651e-01 3.51898521e-01 5.67760706e-01
9.26012993e-01 -1.43565536e-01 -7.54034758e-01 -6.96122468e-01
1.88124888e-02 1.79561287e-01 8.58776689e-01 -1.51224363e+00
4.97134060e-01 -4.65382099e-01 1.95195109e-01 -1.01722944e+00
7.92894840e-01 -9.81422901e-01 5.60705960e-01 3.63614291e-01
-2.74894893e-01 5.17752469e-01 5.69724441e-01 6.16385281e-01
2.73098797e-02 3.70798469e-01 3.59190166e-01 -1.31148458e-01
-1.40959907e+00 4.48967516e-01 -8.81570518e-01 -2.71092951e-01
1.36098766e+00 -4.20334965e-01 -5.71474075e-01 -5.60339034e-01
-5.82280695e-01 6.10172510e-01 8.66094828e-01 5.29668391e-01
5.65846682e-01 -1.12229609e+00 -6.45268023e-01 4.21630889e-01
1.44221127e-01 -9.16205943e-02 4.06872958e-01 5.16317368e-01
-1.07125819e+00 5.86597800e-01 -6.43075466e-01 -6.01506233e-01
-1.09771550e+00 4.10869658e-01 4.40731078e-01 5.61204739e-02
-1.10417712e+00 2.05508471e-01 1.30377576e-01 -8.84723902e-01
-1.82365045e-01 -4.32012171e-01 3.53845730e-02 -5.65545499e-01
5.08540630e-01 8.86866450e-01 -1.17687762e-01 -9.67233181e-01
-3.30820717e-02 2.24397272e-01 3.25257480e-01 -5.10424376e-01
1.42277241e+00 -2.82872885e-01 4.82877910e-01 2.98657984e-01
7.25450039e-01 -2.86143601e-01 -1.91661870e+00 1.50435954e-01
-6.81665838e-02 -6.66365445e-01 6.81904852e-02 -9.05261219e-01
-5.11072755e-01 7.92282820e-01 6.51257515e-01 -1.08242840e-01
3.87351036e-01 -4.48999017e-01 9.20442164e-01 6.29738033e-01
9.94255841e-01 -1.19110405e+00 -4.57308680e-01 4.62961137e-01
8.14913034e-01 -1.35240424e+00 -3.29825103e-01 -4.26042438e-01
-9.78169203e-01 8.13905120e-01 8.92336309e-01 -2.73012608e-01
5.48895121e-01 3.67083073e-01 9.88980383e-02 -2.75260806e-01
-8.37461889e-01 -2.97186822e-01 9.23554525e-02 1.27956593e+00
-3.40309381e-01 4.55466330e-01 4.63171482e-01 -1.33051634e-01
-5.55225492e-01 7.39034638e-02 6.74989641e-01 9.76778686e-01
-6.67006195e-01 -9.12253082e-01 -5.88433035e-02 4.39363182e-01
5.45303941e-01 2.83299208e-01 2.17276379e-01 1.08517015e+00
1.40004382e-01 7.75703251e-01 5.18127620e-01 -4.69778329e-01
5.28057396e-01 -1.39362916e-01 3.37135971e-01 -4.61665183e-01
-5.20405233e-01 -7.02382684e-01 5.03358901e-01 -8.18783343e-01
-1.40343994e-01 -7.62773037e-01 -1.45777750e+00 -7.95020103e-01
4.70416218e-01 -6.46295100e-02 8.64167750e-01 6.37030661e-01
6.50125146e-01 3.76499683e-01 1.99844077e-01 -1.32135403e+00
1.17251329e-01 -4.64278668e-01 -2.84728646e-01 1.98278382e-01
5.19316256e-01 -8.50284874e-01 2.67295748e-01 1.67765290e-01] | [5.553201675415039, 0.849898099899292] |
a0dca224-70af-43ac-b698-7b4f9d3b9b4a | universal-domain-adaptive-object-detector | 2207.01756 | null | https://arxiv.org/abs/2207.01756v1 | https://arxiv.org/pdf/2207.01756v1.pdf | Universal Domain Adaptive Object Detector | Universal domain adaptive object detection (UniDAOD)is more challenging than domain adaptive object detection (DAOD) since the label space of the source domain may not be the same as that of the target and the scale of objects in the universal scenarios can vary dramatically (i.e, category shift and scale shift). To this end, we propose US-DAF, namely Universal Scale-Aware Domain Adaptive Faster RCNN with Multi-Label Learning, to reduce the negative transfer effect during training while maximizing transferability as well as discriminability in both domains under a variety of scales. Specifically, our method is implemented by two modules: 1) We facilitate the feature alignment of common classes and suppress the interference of private classes by designing a Filter Mechanism module to overcome the negative transfer caused by category shift. 2) We fill the blank of scale-aware adaptation in object detection by introducing a new Multi-Label Scale-Aware Adapter to perform individual alignment between the corresponding scale for two domains. Experiments show that US-DAF achieves state-of-the-art results on three scenarios (i.e, Open-Set, Partial-Set, and Closed-Set) and yields 7.1% and 5.9% relative improvement on benchmark datasets Clipart1k and Watercolor in particular. | ['ShiLiang Pu', 'WeiJie Chen', 'Lei Zhang', 'Wenxu Shi'] | 2022-07-05 | null | null | null | null | ['multi-label-learning'] | ['methodology'] | [ 2.19039336e-01 -3.28957111e-01 -2.12030392e-02 -5.70117533e-01
-6.97419822e-01 -7.31878757e-01 3.32541734e-01 -1.03499405e-01
-6.16239488e-01 4.54631120e-01 -2.58443922e-01 1.91137195e-01
1.03457406e-01 -6.34235680e-01 -6.24716759e-01 -8.03314447e-01
1.84234068e-01 4.68460321e-01 1.01736546e+00 -2.22334601e-02
-1.20658232e-02 6.34434640e-01 -1.45999968e+00 4.12711680e-01
9.55157638e-01 1.14107835e+00 4.69485521e-01 4.28647786e-01
-2.06997678e-01 3.59748930e-01 -6.30647302e-01 -8.71743113e-02
5.83364069e-01 -2.59509325e-01 -5.09762406e-01 3.41308504e-01
8.07140410e-01 -4.59696770e-01 -2.58315384e-01 1.32383239e+00
7.62339652e-01 3.50156687e-02 7.55624175e-01 -1.49518979e+00
-7.69242108e-01 3.27685595e-01 -1.07499993e+00 4.07144219e-01
-3.69933933e-01 1.34729534e-01 7.03881145e-01 -9.44241405e-01
5.18135071e-01 1.54000270e+00 5.33949196e-01 7.89861321e-01
-1.33987355e+00 -1.11724627e+00 6.12797916e-01 1.82648912e-01
-1.42286146e+00 -1.17548928e-01 5.52969098e-01 -4.72615868e-01
3.78135234e-01 -3.34726349e-02 1.80856869e-01 9.14225399e-01
-8.93235132e-02 7.79381335e-01 1.13625407e+00 -3.45557988e-01
3.33916277e-01 3.74364585e-01 1.62024066e-01 3.30712706e-01
3.71184021e-01 -1.05774492e-01 -2.28862345e-01 -1.72685285e-03
8.55957687e-01 2.19637707e-01 5.61990403e-02 -9.07736421e-01
-1.12832427e+00 6.86708629e-01 6.74230278e-01 2.90904511e-02
-9.21722949e-02 -1.68446109e-01 5.99986255e-01 2.88563937e-01
3.39905679e-01 2.83277541e-01 -7.41038442e-01 5.29817641e-01
-4.88558948e-01 2.92324722e-01 2.97768176e-01 1.35198581e+00
7.67591178e-01 -1.70238972e-01 -4.73569900e-01 1.14952767e+00
1.31599188e-01 6.70368850e-01 5.70340693e-01 -7.11004615e-01
4.74484563e-01 8.28120291e-01 1.68305486e-01 -5.99734306e-01
-6.30077779e-01 -5.30296922e-01 -6.99200094e-01 4.13087040e-01
6.75719619e-01 -2.31728360e-01 -1.13418543e+00 2.00223351e+00
7.32287586e-01 2.54527062e-01 4.16185521e-02 1.03689265e+00
8.82343650e-01 4.95734572e-01 3.51757646e-01 -5.41156121e-02
1.56794524e+00 -1.13194251e+00 -4.43481892e-01 -5.46116590e-01
5.44617474e-01 -7.11119175e-01 1.36318016e+00 1.29069060e-01
-6.12194300e-01 -7.87785232e-01 -1.17104745e+00 -1.55352959e-02
-5.00666201e-01 3.43844384e-01 3.42968076e-01 4.78176653e-01
-3.66906554e-01 1.36880949e-01 -5.43494463e-01 -4.86101359e-01
6.66750133e-01 2.63604432e-01 -2.33453274e-01 -2.23108470e-01
-1.01401997e+00 6.07087851e-01 7.22789764e-01 -4.50392455e-01
-8.80332470e-01 -7.40715146e-01 -4.69875902e-01 -8.45654383e-02
5.82436800e-01 -3.47326517e-01 1.16416168e+00 -1.32454479e+00
-1.10194373e+00 8.97615790e-01 3.38894844e-01 -1.50183752e-01
5.44119954e-01 -3.25873107e-01 -6.05306327e-01 1.55863361e-02
2.97691464e-01 9.11854923e-01 8.10492992e-01 -1.13706231e+00
-1.17628765e+00 -5.28068244e-01 8.06441307e-02 4.07628745e-01
-6.20596766e-01 1.30815461e-01 -5.31785369e-01 -8.28727484e-01
1.68222070e-01 -9.93109703e-01 3.10107018e-03 6.31508470e-01
-9.38183144e-02 -3.14609349e-01 1.19138908e+00 -3.79076719e-01
1.07067263e+00 -2.38222003e+00 -8.34671780e-02 -2.61138469e-01
1.30474910e-01 4.43724871e-01 -4.45874035e-01 -2.70705134e-01
-1.50998861e-01 -3.41448426e-01 -2.00934082e-01 -6.92488030e-02
-1.00858025e-01 1.68537408e-01 -1.09651260e-01 4.10128415e-01
4.33303863e-01 3.97749901e-01 -8.84142578e-01 -4.82655287e-01
3.37945856e-03 8.63572210e-02 -4.80338454e-01 2.08895832e-01
-2.05901295e-01 1.18325323e-01 -4.67432112e-01 7.83191383e-01
1.25663781e+00 -2.64663011e-01 7.64145330e-02 -2.13877797e-01
-5.31594343e-02 -8.40511471e-02 -1.85683358e+00 1.53157496e+00
-2.01293185e-01 2.80218035e-01 7.46419728e-02 -6.89695358e-01
9.10727203e-01 -1.21772617e-01 2.14322746e-01 -6.81248665e-01
-3.95599306e-02 2.11391330e-01 1.09384127e-01 -2.52286166e-01
2.97659785e-01 1.69759579e-02 -1.51878163e-01 1.95545554e-01
2.13404432e-01 1.50832042e-01 6.33686483e-02 1.55085489e-01
8.90756071e-01 -4.51910533e-02 4.31699067e-01 -5.34111142e-01
5.11649489e-01 -1.70191437e-01 9.74173129e-01 8.11836123e-01
-6.61118507e-01 6.73915088e-01 3.41241449e-01 -2.48594612e-01
-9.55383420e-01 -1.21520364e+00 -2.71995187e-01 1.85073102e+00
6.67142570e-01 2.94861645e-01 -7.01581895e-01 -1.10719025e+00
4.04518783e-01 4.25332695e-01 -7.29556620e-01 -2.72324443e-01
-5.02484918e-01 -9.07360017e-01 3.57566148e-01 8.34391534e-01
7.53149807e-01 -8.58143926e-01 -3.64542663e-01 1.00263394e-01
1.17980503e-01 -1.28850913e+00 -9.35857177e-01 3.93616378e-01
-6.20359659e-01 -1.07144475e+00 -7.66639352e-01 -1.04176497e+00
6.47402227e-01 7.45591938e-01 8.59739304e-01 -4.70963925e-01
-2.14227632e-01 2.87808161e-02 -2.94939846e-01 -6.10933125e-01
-1.79270998e-01 5.26298210e-03 2.22523525e-01 2.53213674e-01
4.68282610e-01 -2.95071691e-01 -7.53976405e-01 8.78646910e-01
-9.07100856e-01 -1.52675465e-01 6.56035781e-01 8.11060488e-01
8.12148213e-01 -5.11446297e-02 8.57648075e-01 -9.68927383e-01
1.86481416e-01 -4.85453397e-01 -8.51362824e-01 4.48038578e-01
-5.69983542e-01 -2.16433719e-01 5.73446095e-01 -1.15595531e+00
-1.14702451e+00 2.64848471e-01 3.71551514e-01 -3.83914500e-01
-2.23041952e-01 -3.92992407e-01 -6.09827399e-01 -1.58347860e-01
1.03920805e+00 -4.76277769e-02 -2.57168561e-01 -6.92310333e-01
3.60214561e-01 8.68186414e-01 5.96781135e-01 -4.05394405e-01
9.68590677e-01 5.80361605e-01 -2.25879461e-01 -3.58700752e-01
-9.57114637e-01 -7.69676566e-01 -8.17908287e-01 -9.13204476e-02
7.70427763e-01 -1.34894788e+00 -6.14129081e-02 8.91189337e-01
-9.30491269e-01 -3.08798343e-01 -4.17422414e-01 3.78008902e-01
-1.25023931e-01 2.29035690e-01 -3.20820838e-01 -3.34592670e-01
-4.28363502e-01 -1.03813004e+00 1.06302607e+00 6.17947936e-01
3.73414606e-01 -4.60441232e-01 -2.74163812e-01 -1.69167537e-02
3.16046506e-01 8.23086053e-02 7.36456156e-01 -9.45546925e-01
-3.24090391e-01 -3.71177122e-02 -8.42898369e-01 5.77689707e-01
1.92013815e-01 -2.28970841e-01 -9.80544388e-01 -6.01514816e-01
-3.10229570e-01 -4.14384842e-01 7.77818143e-01 1.92424029e-01
1.15227878e+00 1.83032564e-04 -4.48833555e-01 5.61704099e-01
1.42454195e+00 1.52497560e-01 2.45842233e-01 4.66222197e-01
5.98208368e-01 4.59225595e-01 9.90176499e-01 4.99761969e-01
2.51792669e-01 9.72297430e-01 3.93529981e-01 -8.73366445e-02
-6.16061985e-01 1.54018141e-02 3.03407192e-01 7.59247690e-02
5.57611406e-01 -1.19092084e-01 -7.11302161e-01 5.30362427e-01
-1.89754450e+00 -4.80990589e-01 -9.54982489e-02 2.13128185e+00
8.36912155e-01 3.82420301e-01 5.24856150e-01 -2.84377575e-01
1.20560277e+00 -2.71961689e-01 -1.09686637e+00 1.04933605e-01
-3.76566082e-01 -3.16342086e-01 7.09622264e-01 1.45078506e-02
-1.55656683e+00 8.93073559e-01 5.25746870e+00 1.09651518e+00
-1.09430313e+00 4.21022087e-01 3.77143979e-01 -2.27824390e-01
3.83451939e-01 -3.33981127e-01 -1.21091616e+00 5.81476033e-01
3.75370800e-01 -3.99040654e-02 2.22258940e-01 1.30154288e+00
-2.10304722e-01 8.86130333e-02 -9.78033781e-01 8.33514512e-01
-1.85505271e-01 -8.14630568e-01 -3.60144563e-02 -3.22716355e-01
8.13983321e-01 2.02083752e-01 -1.08302608e-02 5.70634842e-01
2.81365424e-01 -1.72835276e-01 9.17550981e-01 -1.10539660e-01
1.04509664e+00 -5.75643301e-01 5.59421182e-01 3.80069792e-01
-1.25461984e+00 -4.79155838e-01 -6.85162604e-01 3.40251625e-01
-3.13035756e-01 5.78256011e-01 -5.97970724e-01 1.23063482e-01
9.65099633e-01 4.07761514e-01 -8.21946859e-01 1.19877625e+00
-2.46404577e-02 4.14403915e-01 -4.69331115e-01 8.70419741e-02
6.06053360e-02 2.58113414e-01 6.18605256e-01 1.47000873e+00
8.33938196e-02 -7.50771165e-02 3.06786925e-01 5.99989176e-01
-2.45796800e-01 6.50945008e-02 -1.79927513e-01 4.94923532e-01
7.90662348e-01 1.34075081e+00 -8.32262337e-01 -3.36156994e-01
-6.14203930e-01 9.79973316e-01 2.88748205e-01 2.49144092e-01
-9.49611247e-01 -4.51653540e-01 7.43241787e-01 2.20827878e-01
5.97740531e-01 7.44579211e-02 -2.87333190e-01 -1.02120221e+00
-4.79935296e-02 -7.23489523e-01 9.17869449e-01 -4.91801858e-01
-1.49701798e+00 3.27793419e-01 9.36321467e-02 -1.40972710e+00
4.38661337e-01 -7.32083321e-01 -2.87936866e-01 7.01820254e-01
-1.66133308e+00 -1.23798037e+00 -5.98333955e-01 6.60711586e-01
6.68764293e-01 -3.24407250e-01 5.00274837e-01 7.10594058e-01
-6.92757428e-01 1.02356923e+00 4.21518981e-01 1.29113764e-01
1.34178865e+00 -1.32604456e+00 2.42350996e-01 9.13458228e-01
-3.14020663e-01 2.34920159e-01 4.03251678e-01 -3.90505344e-01
-9.50391233e-01 -1.58232141e+00 3.20851594e-01 -1.81513995e-01
5.26595354e-01 -6.72494769e-01 -1.11703539e+00 4.41130877e-01
-4.14953500e-01 5.52600920e-01 3.94319743e-01 -3.79546434e-02
-8.44293118e-01 -5.60895741e-01 -1.49005520e+00 3.79736036e-01
1.23261082e+00 -2.25702405e-01 -3.22909027e-01 4.91162121e-01
9.51624870e-01 -4.60203499e-01 -6.77629650e-01 5.68126917e-01
4.65309709e-01 -6.83029890e-01 1.01424873e+00 -4.69783902e-01
5.86727932e-02 -7.15263009e-01 -3.31376225e-01 -1.01945031e+00
-7.73552179e-01 -4.21169586e-02 -6.25298768e-02 1.52201152e+00
1.29225239e-01 -7.17481613e-01 5.74990153e-01 2.59378731e-01
-6.65489212e-02 -4.55040962e-01 -9.91526544e-01 -1.05743921e+00
7.40490258e-02 -9.08174366e-03 7.41353989e-01 9.04711664e-01
-6.63184345e-01 4.30549741e-01 -1.35220408e-01 5.55909932e-01
6.42972052e-01 2.72888064e-01 8.08489859e-01 -1.25269961e+00
-2.17973500e-01 -3.46650749e-01 -3.62188786e-01 -1.26495731e+00
-2.50402063e-01 -6.94501996e-01 7.12695345e-02 -1.01786685e+00
5.64920306e-01 -6.85091317e-01 -7.38449395e-01 7.32545912e-01
-4.00882781e-01 3.24446261e-01 3.07830513e-01 3.71839255e-01
-9.64137197e-01 3.20251763e-01 1.34302759e+00 -9.04258266e-02
-2.57119328e-01 8.90633464e-02 -9.03389692e-01 6.86020851e-01
6.11064613e-01 -6.46333337e-01 -2.96809435e-01 -3.93619895e-01
-2.91091383e-01 -5.45233905e-01 3.06295902e-01 -1.17129970e+00
6.51709512e-02 -3.41552258e-01 4.81964439e-01 -5.19681633e-01
-3.59854214e-02 -9.28989887e-01 -3.49404842e-01 4.18327481e-01
-3.31021070e-01 -3.55561823e-01 4.82396394e-01 7.55725861e-01
6.70709535e-02 -4.92507443e-02 1.51851714e+00 1.86525702e-01
-1.22684467e+00 3.18681240e-01 1.49101645e-01 2.21206725e-01
1.29938471e+00 -1.50832430e-01 -5.88246167e-01 2.39915311e-01
-4.49026555e-01 5.52649438e-01 3.93249601e-01 7.25882530e-01
3.22070152e-01 -1.43020868e+00 -7.75814354e-01 1.83873102e-01
6.04667306e-01 2.86720157e-01 5.09176016e-01 3.82473946e-01
-1.81624785e-01 -2.83010118e-02 -3.53375703e-01 -7.06444860e-01
-1.42244697e+00 6.94159746e-01 6.13000691e-01 -2.04596072e-01
-4.40509379e-01 1.20064187e+00 8.95850837e-01 -7.63567746e-01
4.58250433e-01 -1.79214954e-01 -1.83863670e-01 2.04969287e-01
7.26547718e-01 4.76119399e-01 -2.56184135e-02 -4.57578957e-01
-5.65186143e-01 5.03856838e-01 -4.57694829e-01 6.02275133e-01
1.18939364e+00 -2.42637753e-01 2.49450922e-01 2.62588680e-01
1.06157744e+00 -3.59566957e-01 -1.85986352e+00 -7.89920151e-01
-2.65417129e-01 -4.01671916e-01 -1.35240316e-01 -1.20426726e+00
-9.57574129e-01 6.47700727e-01 1.40768218e+00 2.60322336e-02
1.33261740e+00 1.25753820e-01 6.82650864e-01 2.30324954e-01
6.88707083e-02 -1.42242312e+00 3.34590405e-01 3.53501260e-01
7.24748015e-01 -1.43962049e+00 6.82821572e-02 -5.79689145e-01
-6.51476443e-01 9.19360995e-01 1.21739709e+00 3.03833093e-02
4.65109318e-01 2.09024429e-01 5.83133996e-02 8.38288590e-02
-3.93397868e-01 -2.50782251e-01 3.73990268e-01 7.07760692e-01
-5.48370257e-02 2.72596091e-01 -1.02802143e-01 6.99502587e-01
5.47436416e-01 -1.81368187e-01 2.06048489e-01 8.19645703e-01
-7.94342875e-01 -7.82566488e-01 -5.74957728e-01 4.18241352e-01
-2.77575940e-01 1.84934586e-01 -2.38658115e-01 8.50520551e-01
6.13768578e-01 6.95376635e-01 1.83213279e-01 -3.09302270e-01
7.30169773e-01 -1.51622295e-01 3.12869370e-01 -7.76487529e-01
-3.57631147e-01 4.62708101e-02 -3.01985025e-01 -2.57275999e-01
-2.28904396e-01 -8.14060688e-01 -1.34274769e+00 8.70719030e-02
-6.80784583e-01 -4.10726637e-01 3.65383089e-01 5.02442479e-01
5.02904117e-01 5.82928002e-01 6.88222229e-01 -5.93699872e-01
-8.64364445e-01 -1.05136204e+00 -8.91443431e-01 5.73025882e-01
2.94287980e-01 -1.01517546e+00 -1.97139010e-01 -1.08203918e-01] | [9.363673210144043, 1.488131046295166] |
537ec6c5-afde-4709-bad3-e111225f2021 | image-processing-based-scene-text-detection | 2004.08079 | null | https://arxiv.org/abs/2004.08079v1 | https://arxiv.org/pdf/2004.08079v1.pdf | Image Processing Based Scene-Text Detection and Recognition with Tesseract | Text Recognition is one of the challenging tasks of computer vision with considerable practical interest. Optical character recognition (OCR) enables different applications for automation. This project focuses on word detection and recognition in natural images. In comparison to reading text in scanned documents, the targeted problem is significantly more challenging. The use case in focus facilitates the possibility to detect the text area in natural scenes with greater accuracy because of the availability of images under constraints. This is achieved using a camera mounted on a truck capturing likewise images round-the-clock. The detected text area is then recognized using Tesseract OCR engine. Even though it benefits low computational power requirements, the model is limited to only specific use cases. This paper discusses a critical false positive case scenario occurred while testing and elaborates the strategy used to alleviate the problem. The project achieved a correct character recognition rate of more than 80\%. This paper outlines the stages of development, the major challenges and some of the interesting findings of the project. | ['Bénédicte Bernier', 'Ebin Zacharias', 'Martin Teuchler'] | 2020-04-17 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 8.62511694e-01 -2.83617288e-01 2.30136916e-01 -2.03764141e-01
-2.25399777e-01 -6.22177660e-01 7.19821632e-01 -8.20162520e-02
-6.32636487e-01 5.30171096e-01 -5.17255545e-01 -4.18637693e-01
1.18662551e-01 -4.12259072e-01 -3.15116078e-01 -6.59042060e-01
3.48083645e-01 5.29293776e-01 4.12297428e-01 1.09411068e-01
9.55450058e-01 9.67082083e-01 -1.49151790e+00 2.18445361e-01
5.76569974e-01 5.13301849e-01 7.85361409e-01 1.27285242e+00
-3.81481051e-01 7.18066037e-01 -8.55551600e-01 -9.70462039e-02
2.08858445e-01 -3.38452995e-01 -6.55327737e-01 9.92789388e-01
4.43913728e-01 -5.44472516e-01 3.77675220e-02 1.07580030e+00
2.21236661e-01 -1.40388474e-01 6.68492913e-01 -9.12015438e-01
-3.55297357e-01 -1.15444347e-01 -7.05588818e-01 3.51784348e-01
6.14250183e-01 6.69079125e-02 4.58259732e-01 -7.26829171e-01
5.14557481e-01 8.16508055e-01 3.73172641e-01 2.56131560e-01
-1.05049694e+00 -1.58433959e-01 -1.87637061e-01 2.06614491e-02
-1.26038051e+00 -2.86489755e-01 3.47510695e-01 -3.51788640e-01
1.33001804e+00 5.02972424e-01 4.59038287e-01 5.85603416e-01
2.86641121e-01 8.43672335e-01 1.27226019e+00 -1.08822477e+00
3.06493863e-02 7.16371000e-01 4.04112250e-01 5.57434380e-01
4.15376514e-01 -3.22376698e-01 -1.04045630e-01 2.61443377e-01
6.89200521e-01 8.14024732e-02 -1.50844827e-01 -2.11620182e-02
-1.04706752e+00 5.43384433e-01 -3.42386097e-01 7.03259289e-01
-1.89166874e-01 -2.23126337e-01 2.23509669e-01 1.01830391e-02
4.45465855e-02 4.53421861e-01 -7.84383267e-02 -5.68655193e-01
-1.13807356e+00 -8.92011449e-02 8.01115215e-01 1.08849287e+00
3.28871310e-01 3.42173845e-01 2.91142493e-01 8.84785414e-01
3.04653645e-01 7.82149374e-01 3.41585845e-01 -2.63678312e-01
3.58432114e-01 6.99328184e-01 1.20618612e-01 -7.92362809e-01
-1.87587723e-01 -3.32751237e-02 -1.55427754e-01 5.38737535e-01
6.17392838e-01 -1.96001947e-01 -1.13644743e+00 3.42798322e-01
3.94083858e-02 -5.73676288e-01 4.44070809e-03 8.04769158e-01
1.46203876e-01 9.58553553e-01 -4.04859811e-01 -2.25693345e-01
1.54376698e+00 -6.53822005e-01 -8.98289502e-01 -3.86399299e-01
4.76389021e-01 -1.33887112e+00 8.03731978e-01 7.45587409e-01
-5.97824275e-01 -2.89181113e-01 -1.16482151e+00 1.91389546e-01
-5.37777722e-01 7.41104901e-01 2.76822448e-01 1.08236885e+00
-7.73373425e-01 7.76454210e-02 -5.70645869e-01 -1.01621699e+00
-1.68423578e-01 4.11421865e-01 -3.69713962e-01 -1.92255616e-01
-3.38981599e-01 1.22407103e+00 2.91527092e-01 2.91836917e-01
-3.41105551e-01 1.57625586e-01 -6.58970952e-01 -2.63970017e-01
3.92508596e-01 3.77744853e-01 1.08281279e+00 -9.80180383e-01
-1.46240878e+00 1.01006758e+00 -2.62427986e-01 -2.74997324e-01
8.53847325e-01 -3.71999860e-01 -5.24063826e-01 4.87649679e-01
-5.81236370e-03 2.72893995e-01 9.95677352e-01 -8.50227654e-01
-9.61459875e-01 -2.51131982e-01 -4.08115089e-01 9.29483771e-02
-7.77096748e-02 5.14089286e-01 -5.83039880e-01 -3.23874831e-01
-8.36007297e-02 -7.93734610e-01 9.13398564e-02 -1.58045948e-01
-3.49192202e-01 -1.09932177e-01 1.41959560e+00 -6.81528687e-01
8.76722395e-01 -1.95110095e+00 -6.03244007e-01 1.31835952e-01
-3.56800467e-01 6.07496440e-01 1.89485013e-01 6.67554677e-01
6.86289892e-02 -1.94991380e-01 -1.48515359e-01 5.55001870e-02
-2.17468336e-01 1.07284397e-01 -3.41906965e-01 6.71644926e-01
4.14977610e-01 2.95457959e-01 -4.09393460e-01 -5.13758242e-01
6.92655802e-01 2.91040510e-01 3.72196883e-01 -5.76556250e-02
5.72730638e-02 -3.91465455e-01 -4.44131553e-01 1.02388370e+00
6.97039664e-01 2.53078103e-01 9.31715295e-02 2.18467250e-01
-6.68758214e-01 -2.71181047e-01 -1.52193606e+00 7.02192605e-01
-8.91901404e-02 1.30730891e+00 7.41369203e-02 -1.02528083e+00
1.36415339e+00 2.18496323e-01 1.63826961e-02 -5.41918874e-01
8.16834420e-02 3.41827303e-01 -1.85946167e-01 -9.53277349e-01
1.11982501e+00 5.08458167e-02 4.81629401e-01 3.14677149e-01
-4.36673880e-01 -2.60325402e-01 3.92121583e-01 1.26922103e-02
7.40903139e-01 1.67622998e-01 2.85527229e-01 -3.49905699e-01
5.59536457e-01 6.07192934e-01 -1.13112517e-02 7.40998626e-01
-2.68572867e-01 6.39115274e-01 2.38880351e-01 -3.10582727e-01
-1.18189549e+00 -3.62762153e-01 -3.64765406e-01 5.25533855e-01
1.81522369e-01 7.16535971e-02 -8.10887396e-01 -2.37312049e-01
-8.95548984e-02 8.44580829e-01 -3.91883731e-01 5.62624574e-01
-5.54193258e-01 -6.99313283e-01 4.83645678e-01 3.79994482e-01
5.00773907e-01 -1.02813864e+00 -1.20473623e+00 7.53095448e-02
2.30366930e-01 -1.22657979e+00 -1.40285164e-01 1.78679287e-01
-1.01936305e+00 -1.01910138e+00 -9.48916554e-01 -9.25688505e-01
1.00839174e+00 7.47357488e-01 4.94496077e-01 8.97189155e-02
-1.05531585e+00 7.67019212e-01 -5.02111077e-01 -6.58053458e-01
-5.15245080e-01 -2.08319277e-01 -2.20385626e-01 -2.06978084e-03
9.05656576e-01 4.69701648e-01 -1.38262093e-01 4.92290795e-01
-9.07600462e-01 -1.98201224e-01 7.31095493e-01 5.10201693e-01
2.76079595e-01 3.06449920e-01 6.21548072e-02 -5.48118591e-01
5.93084931e-01 3.05263638e-01 -1.05988741e+00 3.41125786e-01
-6.23224854e-01 -2.77446747e-01 4.35131907e-01 -3.94593924e-01
-1.00276935e+00 3.30908060e-01 3.53961021e-01 2.54704446e-01
-6.40886128e-01 3.62853520e-02 1.04001559e-01 -2.01507047e-01
4.89034772e-01 5.79551637e-01 2.33126417e-01 -2.75910556e-01
-4.31174159e-01 1.26374674e+00 3.50547075e-01 -9.20683295e-02
6.35656714e-01 4.95701611e-01 -2.37812568e-02 -1.91516769e+00
8.38660449e-02 -8.86932671e-01 -5.97802341e-01 -5.55873334e-01
8.37428987e-01 -5.69946647e-01 -4.65937436e-01 7.10124373e-01
-1.18722415e+00 -4.92450148e-02 1.52845874e-01 7.77760684e-01
-1.65930867e-01 7.44259179e-01 -3.41384351e-01 -1.48702419e+00
-1.70811638e-01 -1.15079546e+00 9.76856411e-01 3.44732970e-01
-9.95432660e-02 -8.00045013e-01 -2.94467688e-01 5.02264738e-01
2.09313333e-01 3.75733823e-02 4.21382904e-01 -4.60211247e-01
-6.39745653e-01 -9.30660248e-01 -3.56098562e-01 3.76566917e-01
2.65832096e-01 6.11368954e-01 -8.97345185e-01 -1.50248826e-01
1.68145467e-02 -4.64071780e-02 5.66728950e-01 2.52946347e-01
5.10712743e-01 4.15566593e-01 -4.59746331e-01 -1.86373636e-01
1.77848709e+00 7.32775450e-01 1.00653720e+00 7.64011323e-01
4.70382154e-01 7.09765911e-01 9.56151247e-01 1.51718810e-01
-3.86833638e-01 6.06631100e-01 1.93788037e-01 7.12796301e-02
1.45145595e-01 3.34849268e-01 5.07391870e-01 1.85243905e-01
-3.44660394e-02 -3.79338175e-01 -1.19344103e+00 6.12875223e-01
-1.32295084e+00 -9.09669161e-01 -7.82728195e-01 2.11690927e+00
2.99609393e-01 2.10591510e-01 3.44207883e-02 5.13108969e-01
1.24430394e+00 -1.97265625e-01 -4.95249406e-02 -8.88973832e-01
8.08910578e-02 -2.06019208e-01 8.84136677e-01 6.96532130e-01
-9.59370732e-01 7.41158724e-01 6.63664484e+00 5.46592712e-01
-1.30553424e+00 -5.65624416e-01 2.14093700e-01 2.07473293e-01
5.14747858e-01 -8.08888599e-02 -1.28587413e+00 3.86746407e-01
6.15051627e-01 2.65938669e-01 1.17717065e-01 7.77614594e-01
2.77145475e-01 -1.01735878e+00 -8.49643767e-01 8.96995544e-01
5.83848596e-01 -9.93294656e-01 1.13245202e-02 3.20059121e-01
2.29823813e-01 -2.53985167e-01 -2.58294672e-01 -3.36144030e-01
-2.56522357e-01 -7.64318407e-01 5.89809060e-01 4.22523946e-01
8.47430646e-01 -7.44305491e-01 7.51011968e-01 3.63839865e-01
-8.01031947e-01 5.81989773e-02 -6.60459340e-01 -3.31350923e-01
9.05400440e-02 4.73437339e-01 -1.46625030e+00 4.18988615e-02
4.14409488e-01 1.98082849e-01 -6.57731831e-01 1.11311376e+00
7.50696426e-03 4.87132519e-01 -4.21914339e-01 -7.70979285e-01
4.18008894e-01 -4.92852360e-01 5.76464832e-01 1.84957516e+00
3.18065733e-01 -7.15677291e-02 -1.86865628e-01 4.63481516e-01
5.90727627e-01 1.45705640e-01 -8.43945444e-01 -3.17095190e-01
1.24103062e-01 1.28904629e+00 -1.31731594e+00 -3.98567021e-01
-4.32767630e-01 1.29187191e+00 -3.10499102e-01 9.41910967e-02
-4.98346984e-01 -9.87431049e-01 8.92151799e-03 2.52212018e-01
4.95196939e-01 -3.98988187e-01 -3.41769487e-01 -7.87643254e-01
2.09093899e-01 -8.17559183e-01 -1.25640482e-01 -8.92371595e-01
-5.32751918e-01 3.19803476e-01 -1.53027549e-01 -1.17828417e+00
-2.07855836e-01 -1.28598905e+00 -6.09887600e-01 1.07149088e+00
-1.12956357e+00 -8.93406987e-01 -3.17210913e-01 2.80132413e-01
1.13730133e+00 -4.32166368e-01 7.22035766e-01 -9.40886512e-02
-7.13202357e-01 4.28481966e-01 4.64540929e-01 3.07652414e-01
5.69860697e-01 -1.32662535e+00 2.13573128e-01 1.22171760e+00
-1.90705791e-01 4.76417571e-01 9.05328929e-01 -7.98531175e-01
-1.64890468e+00 -5.28746426e-01 1.14220190e+00 -3.60945672e-01
6.40418231e-01 -3.63630205e-01 -6.58073723e-01 3.11947644e-01
3.68639767e-01 -5.92036724e-01 4.33182538e-01 -4.66576159e-01
1.29311547e-01 5.89540675e-02 -1.16843486e+00 4.99040961e-01
8.31510350e-02 -5.27761936e-01 -7.08735704e-01 4.35011894e-01
-1.66482523e-01 -2.62842774e-01 -2.66805351e-01 -3.97656560e-01
8.14882576e-01 -8.74645531e-01 4.05426472e-01 -1.06768414e-01
3.23719323e-01 -6.02108240e-01 -1.83960162e-02 -6.06114030e-01
1.53359801e-01 -4.50361162e-01 6.08884871e-01 1.22781610e+00
5.09154499e-01 -6.14555418e-01 8.24924707e-01 7.06136763e-01
1.32676661e-01 -2.95406915e-02 -5.64770579e-01 -7.37178802e-01
-5.94450891e-01 -2.86601633e-01 -3.03961784e-01 7.73093283e-01
1.18753292e-01 1.12972111e-01 -3.59070361e-01 2.83170462e-01
5.52447557e-01 -2.41368054e-03 7.13900149e-01 -1.04995477e+00
-1.55797780e-01 -9.92608815e-02 -8.12671363e-01 -7.98048556e-01
-4.69603419e-01 -2.54642099e-01 1.70886040e-01 -1.61288548e+00
9.60222036e-02 2.98862636e-01 4.64818716e-01 6.32481873e-02
6.98059052e-02 3.41836393e-01 4.33934301e-01 1.87337607e-01
-2.00600669e-01 -5.16458035e-01 8.58558416e-01 -2.05412775e-01
-7.44596124e-04 -6.71614753e-03 -2.58306473e-01 5.81638575e-01
9.92264986e-01 -1.79724872e-01 -1.97170809e-01 -5.83818674e-01
1.80830769e-02 -2.90410548e-01 1.45725116e-01 -8.87137175e-01
4.23914075e-01 -9.24829841e-02 7.39193797e-01 -1.08125305e+00
2.41769105e-01 -1.19718945e+00 -1.00922413e-01 7.05981731e-01
1.15256123e-02 3.06980699e-01 2.11279184e-01 5.54596841e-01
-8.83405954e-02 -1.05333853e+00 8.41152370e-01 -2.20907554e-01
-1.02080393e+00 -6.72373354e-01 -1.16370738e+00 -7.32086360e-01
1.35325694e+00 -1.27961886e+00 -2.77958006e-01 -1.13340527e-01
-3.58629465e-01 -1.05823889e-01 5.43500721e-01 1.96418688e-01
6.51671767e-01 -5.41092813e-01 -5.29806376e-01 2.85590678e-01
1.43764302e-01 -2.78452009e-01 -1.31970882e-01 7.49743283e-01
-1.40069687e+00 7.41358519e-01 -3.04264218e-01 -8.87433410e-01
-1.92582047e+00 4.07343835e-01 1.83599606e-01 2.33545899e-01
-7.30481386e-01 3.94870609e-01 -3.87922108e-01 3.09825748e-01
2.37947255e-01 -1.78459898e-01 -4.05594796e-01 1.66457668e-02
9.48677242e-01 7.15052128e-01 1.89415708e-01 -5.24379671e-01
-3.16055447e-01 7.63601363e-01 -5.66518009e-01 -3.05682421e-01
1.03480017e+00 -1.87644467e-01 -1.61783144e-01 3.66356969e-01
8.43964219e-01 2.14132443e-01 -1.08873212e+00 1.37386784e-01
1.68415502e-01 -6.79328740e-01 -3.42651717e-02 -7.88266599e-01
-4.45713460e-01 8.74818623e-01 8.66349101e-01 4.10325021e-01
8.47386539e-01 -4.56983060e-01 1.49934322e-01 8.74615788e-01
1.75090328e-01 -1.77583408e+00 -1.69468597e-01 5.17724454e-01
6.95903778e-01 -1.17439961e+00 3.78119051e-01 -3.86765838e-01
-7.29192734e-01 1.83095098e+00 3.55514556e-01 -1.83288585e-02
-2.65444685e-02 5.51889241e-01 2.58024424e-01 -5.48381619e-02
-2.36882776e-01 -6.07758090e-02 -2.77172178e-01 8.56250405e-01
6.19622290e-01 4.25201692e-02 -5.70072889e-01 -4.76989746e-01
3.07667881e-01 -2.70694979e-02 1.23587692e+00 1.47008610e+00
-9.99210358e-01 -8.81181479e-01 -1.12588346e+00 6.08492494e-01
-6.55461907e-01 1.14323139e-01 -7.81217933e-01 9.84946609e-01
-2.50997394e-01 1.16601443e+00 1.24356918e-01 -3.87419648e-02
2.72468984e-01 1.88160032e-01 3.35262120e-01 -3.22743922e-01
-2.24416271e-01 3.18316251e-01 2.71879017e-01 -2.09157132e-02
-3.45819056e-01 -8.77997279e-01 -1.14692342e+00 -1.99657023e-01
-7.57677674e-01 1.17691800e-01 1.26667857e+00 9.38453078e-01
-1.78184509e-01 1.77493274e-01 4.35481906e-01 -6.96677864e-01
-5.28871894e-01 -9.61948693e-01 -9.37481403e-01 -1.38030201e-01
2.64944643e-01 -2.50525355e-01 -2.69578844e-01 4.53667223e-01] | [11.80321979522705, 2.603389024734497] |
f0bad5c4-3f1c-4d3f-abb7-f43790173192 | authnet-a-deep-learning-based-authentication | 2012.02515 | null | https://arxiv.org/abs/2012.02515v2 | https://arxiv.org/pdf/2012.02515v2.pdf | AuthNet: A Deep Learning based Authentication Mechanism using Temporal Facial Feature Movements | Biometric systems based on Machine learning and Deep learning are being extensively used as authentication mechanisms in resource-constrained environments like smartphones and other small computing devices. These AI-powered facial recognition mechanisms have gained enormous popularity in recent years due to their transparent, contact-less and non-invasive nature. While they are effective to a large extent, there are ways to gain unauthorized access using photographs, masks, glasses, etc. In this paper, we propose an alternative authentication mechanism that uses both facial recognition and the unique movements of that particular face while uttering a password, that is, the temporal facial feature movements. The proposed model is not inhibited by language barriers because a user can set a password in any language. When evaluated on the standard MIRACL-VC1 dataset, the proposed model achieved an accuracy of 98.1%, underscoring its effectiveness as an effective and robust system. The proposed method is also data-efficient since the model gave good results even when trained with only 10 positive video samples. The competence of the training of the network is also demonstrated by benchmarking the proposed system against various compounded Facial recognition and Lip reading models. | ['Sowmya Kamath', 'B R Mukesh', 'Pravan Omprakash', 'Mohit Raghavendra'] | 2020-12-04 | null | null | null | null | ['lip-password-classification'] | ['time-series'] | [ 6.14912324e-02 -1.48250490e-01 -1.77062958e-01 -2.42799237e-01
-2.11486787e-01 -3.36618632e-01 5.61863184e-01 -5.12152493e-01
-7.79172003e-01 6.12105906e-01 -3.00447017e-01 -1.95171893e-01
5.83210737e-02 -3.87763470e-01 -2.95057923e-01 -7.84571826e-01
1.55089587e-01 -1.56052476e-02 -8.29007775e-02 -2.65421197e-02
1.95353076e-01 8.46308351e-01 -1.90476179e+00 -7.29935244e-02
5.60124695e-01 1.22206950e+00 -2.59958327e-01 5.28094947e-01
1.88560531e-01 1.26274779e-01 -6.30776048e-01 -6.62536860e-01
2.57891893e-01 -9.07545909e-02 -4.45478082e-01 7.67108798e-02
5.30969441e-01 -6.43220425e-01 -1.61068693e-01 8.08392346e-01
8.08989167e-01 -1.61397353e-01 4.04325396e-01 -1.23311388e+00
-3.40628058e-01 -2.37493366e-02 -7.75081873e-01 -1.45983502e-01
6.78371191e-01 2.53808498e-01 4.30971563e-01 -1.11850452e+00
2.26658702e-01 9.14806843e-01 7.55269587e-01 1.07587361e+00
-8.12864304e-01 -1.15037572e+00 -2.45714635e-01 2.08603084e-01
-1.65209174e+00 -1.00318539e+00 6.26976490e-01 -2.83715338e-01
8.67546082e-01 9.93796661e-02 6.24533594e-01 1.17145097e+00
9.56107751e-02 5.65318882e-01 1.11649215e+00 -6.05607331e-01
-5.83304949e-02 5.50702155e-01 -8.28705803e-02 8.60505104e-01
4.49180543e-01 5.15468754e-02 -6.09452724e-01 -2.30619714e-01
4.92217034e-01 -1.29673667e-02 -2.34084442e-01 -1.61837354e-01
-6.93095446e-01 5.14648676e-01 -1.58377007e-01 5.13546109e-01
-4.11594808e-01 -2.43768305e-01 3.32051575e-01 6.93014171e-03
1.76145718e-01 -9.88623220e-03 -4.32689428e-01 -3.75023544e-01
-9.90878165e-01 -2.18116850e-01 7.85495698e-01 4.21214998e-01
3.21646482e-01 3.09436828e-01 3.77345204e-01 8.69423926e-01
5.85975170e-01 7.99005926e-01 7.38837779e-01 -2.57871330e-01
1.31442592e-01 5.04458129e-01 2.12421522e-01 -9.99879241e-01
-4.75798696e-01 1.87349722e-01 -9.88938451e-01 3.93694997e-01
7.29976773e-01 -1.68838054e-01 -9.18757915e-01 1.71669686e+00
3.37973714e-01 2.44840771e-01 6.86317682e-02 5.08536696e-01
7.73545265e-01 3.15845519e-01 1.31349593e-01 -2.40157500e-01
1.46872032e+00 -3.21674615e-01 -1.01118100e+00 8.95778462e-02
2.06061408e-01 -8.23794544e-01 1.05960238e+00 6.83655739e-01
-8.30144346e-01 -6.69321299e-01 -1.09062064e+00 3.38172346e-01
-4.93300706e-01 5.29827237e-01 6.30348325e-01 1.69814670e+00
-1.18051386e+00 4.39523786e-01 -6.59168720e-01 -7.22306728e-01
6.31860375e-01 1.10513401e+00 -8.27742815e-01 1.95426121e-01
-9.96100128e-01 6.75565362e-01 -1.11792319e-01 3.44023943e-01
-3.89466792e-01 -1.11841105e-01 -7.60026276e-01 -5.11266105e-02
1.00184558e-02 -1.94277633e-02 8.14448357e-01 -9.91371334e-01
-2.01941490e+00 1.10483968e+00 -3.79860222e-01 -2.89555758e-01
4.10361201e-01 -2.71911800e-01 -8.60006988e-01 2.36028060e-01
-5.49779594e-01 5.12746155e-01 1.33674729e+00 -6.91377699e-01
-3.08906317e-01 -5.58131754e-01 -2.70879060e-01 -2.35976815e-01
-7.57334828e-01 3.28938663e-01 -1.71183571e-01 -1.82022154e-01
-1.15163818e-01 -1.00328660e+00 4.47484642e-01 7.98297748e-02
-3.38274479e-01 -1.78411692e-01 1.24770987e+00 -8.34469855e-01
9.97822225e-01 -2.17886138e+00 -4.67133522e-01 4.10801709e-01
-8.81343335e-02 1.02937245e+00 4.92537245e-02 1.87879771e-01
-6.68339208e-02 2.14258432e-01 1.12116687e-01 -3.83381784e-01
-2.50925094e-01 -1.58812508e-01 1.10318109e-01 7.21344948e-01
1.78961784e-01 6.31683886e-01 -2.70394892e-01 -3.07089478e-01
3.56708020e-01 1.03320789e+00 -2.57541120e-01 1.30947292e-01
5.49365044e-01 3.38373989e-01 -1.02122992e-01 1.02142406e+00
8.43751669e-01 1.77869484e-01 1.69738039e-01 -1.46206722e-01
1.36378393e-01 -2.56759137e-01 -1.31084573e+00 1.19562376e+00
-3.54393661e-01 6.55187130e-01 1.78002417e-01 -6.75810695e-01
1.02317882e+00 7.47948229e-01 4.25209850e-01 -5.74453592e-01
3.99923474e-01 3.33578050e-01 1.15020797e-01 -7.11326182e-01
2.32171416e-01 2.98590884e-02 4.72831488e-01 4.45580453e-01
6.69013038e-02 3.85093749e-01 -3.06625128e-01 -2.95036703e-01
4.94774997e-01 1.71080157e-01 4.60242689e-01 -7.80794993e-02
1.17031765e+00 -7.98685372e-01 2.63942122e-01 4.41453755e-01
-5.88399887e-01 1.65878624e-01 -1.18666179e-01 -4.02607530e-01
-5.44499815e-01 -5.29682338e-01 -2.55485862e-01 6.34590805e-01
3.44338864e-02 -1.94279805e-01 -1.08050883e+00 -6.77897334e-01
-1.08961619e-01 -1.62593275e-02 -5.10918975e-01 -1.44280210e-01
-4.51445341e-01 -4.63301122e-01 9.95912910e-01 2.38931075e-01
8.23998809e-01 -1.11658525e+00 -6.68426633e-01 2.49545071e-02
4.70245183e-02 -1.21643817e+00 -3.01744998e-01 -4.12015140e-01
-7.63298750e-01 -1.17948627e+00 -8.68905962e-01 -6.89324498e-01
5.62567055e-01 6.22269139e-02 4.56342548e-01 4.86124396e-01
-3.84004354e-01 6.11946344e-01 -1.89226754e-02 -6.13355696e-01
-2.05142736e-01 4.11991999e-02 7.85029590e-01 7.22097099e-01
9.10833061e-01 -2.53962874e-01 -6.54643834e-01 5.57285964e-01
-5.92980385e-01 -4.87837613e-01 4.19367582e-01 1.01082122e+00
8.46478343e-02 -1.90940008e-01 7.82134652e-01 -4.93327677e-01
6.62494838e-01 -4.13495079e-02 -5.57982624e-01 2.16756672e-01
-6.07648671e-01 -3.06774974e-01 4.71023798e-01 -7.03984976e-01
-1.06396878e+00 2.89514035e-01 -6.30725622e-01 5.04132137e-02
-5.19013882e-01 -1.61295757e-02 -5.21848738e-01 -6.90407991e-01
4.50167120e-01 2.24232405e-01 4.16480184e-01 -4.44696575e-01
-1.48400843e-01 1.18978107e+00 3.58772844e-01 -2.41314858e-01
8.48130345e-01 4.75342929e-01 -9.61147472e-02 -1.47637248e+00
-4.65244912e-02 -4.36104417e-01 -7.48745084e-01 -3.96983266e-01
7.25054026e-01 -5.62455654e-01 -1.56492543e+00 1.29556441e+00
-9.74913359e-01 2.12602749e-01 3.11607420e-01 5.05349636e-01
-1.10860191e-01 6.18746459e-01 -4.11123127e-01 -1.49987423e+00
-6.27140522e-01 -1.14451742e+00 8.27869952e-01 6.17284894e-01
-1.27126426e-01 -7.93106019e-01 -2.78988123e-01 3.74212056e-01
7.04762757e-01 5.98864630e-02 3.91118526e-01 -6.69216633e-01
-1.22698873e-01 -7.96739817e-01 -5.57277761e-02 4.56703991e-01
6.03733718e-01 2.97955990e-01 -1.59164965e+00 -4.04784888e-01
3.16522755e-02 -4.46547002e-01 3.12885523e-01 2.68461436e-01
9.84642625e-01 -2.93938458e-01 -2.76494890e-01 4.68274385e-01
1.20616746e+00 5.14970720e-01 8.14810991e-01 1.21459015e-01
4.64699030e-01 6.67252183e-01 2.04263970e-01 3.65862519e-01
7.12255063e-03 8.34025800e-01 3.37164611e-01 -3.10380608e-01
1.37774065e-01 -1.27057051e-02 3.87853056e-01 4.13892835e-01
-3.54927331e-01 -3.53460535e-02 -8.05089951e-01 1.56597331e-01
-1.30326009e+00 -1.18867695e+00 1.97745338e-01 2.51402855e+00
5.79299688e-01 -8.55482519e-02 3.29470932e-01 6.09574795e-01
6.04333162e-01 -2.27050155e-01 -4.96850431e-01 -5.39905906e-01
-1.85991257e-01 6.42358661e-01 2.90344745e-01 4.28408623e-01
-1.08024728e+00 9.53394830e-01 6.00113153e+00 4.30656642e-01
-1.65193295e+00 -2.04026885e-02 6.02448463e-01 1.15057953e-01
4.84000325e-01 -6.69202566e-01 -9.83597755e-01 4.17656273e-01
1.07463992e+00 2.98435479e-01 2.13165104e-01 6.91796124e-01
2.97754973e-01 -3.40332448e-01 -8.79176795e-01 1.33684218e+00
2.44464546e-01 -1.03827775e+00 -1.25501603e-01 3.52565080e-01
1.78469479e-01 -5.34098268e-01 3.81426305e-01 -1.69338975e-02
-5.83519459e-01 -1.36154246e+00 2.73046196e-01 4.33290362e-01
1.15878391e+00 -7.64483392e-01 9.95900452e-01 2.32574627e-01
-1.10025299e+00 -1.06432922e-01 -1.32564485e-01 -2.75417138e-02
-1.13108054e-01 -1.30387336e-01 -8.67288530e-01 2.85365105e-01
6.47817254e-01 2.61370897e-01 -5.75476050e-01 9.19626296e-01
3.79167236e-02 6.71325922e-01 -4.85265195e-01 -2.33951703e-01
-8.73144567e-02 -9.80433356e-03 1.00156583e-01 1.05766690e+00
3.75329256e-01 1.97405238e-02 -4.15666938e-01 1.90202221e-01
-1.23911379e-02 3.88434291e-01 -8.39602292e-01 -6.62563816e-02
2.79204369e-01 1.32653487e+00 -5.42900860e-01 8.07622299e-02
-5.74956656e-01 9.35177088e-01 -1.96179748e-01 2.34712645e-01
-6.36355042e-01 -5.16684771e-01 8.65207314e-01 2.43378907e-01
7.41439983e-02 -2.28006855e-01 -1.55424455e-03 -9.83271539e-01
1.88843235e-02 -1.15535414e+00 1.08375646e-01 -4.75638360e-01
-8.16221833e-01 8.33572447e-01 -3.87242466e-01 -1.16768324e+00
-4.14533198e-01 -8.64691913e-01 -2.81395614e-01 1.04745221e+00
-1.41490185e+00 -1.23931074e+00 -4.75846201e-01 1.10776412e+00
1.89941406e-01 -6.75566614e-01 1.23667812e+00 4.93446767e-01
-6.27084613e-01 1.35687995e+00 -2.37415850e-01 4.44252670e-01
8.08126926e-01 -6.69353306e-01 1.08728796e-01 8.40589285e-01
2.66291887e-01 9.74584877e-01 3.69969189e-01 -3.05197030e-01
-1.47779799e+00 -2.21590385e-01 8.43641877e-01 -1.67794898e-01
1.38620600e-01 -5.42609334e-01 -8.78456533e-01 1.80351555e-01
3.36504906e-01 -8.03648494e-03 9.15485799e-01 -2.09663585e-01
-2.87169605e-01 -3.80659074e-01 -1.61081648e+00 4.42179203e-01
6.75012350e-01 -6.38705730e-01 -2.19930306e-01 1.39646083e-01
-2.06166238e-01 -1.41889900e-01 -6.25711083e-01 2.70123303e-01
1.19044685e+00 -1.09403741e+00 7.45987356e-01 -4.99037296e-01
-4.04397041e-01 -1.48112759e-01 8.93375184e-03 -6.08175814e-01
2.67552137e-01 -9.72914815e-01 -1.71887517e-01 1.44230866e+00
3.88574541e-01 -1.03432727e+00 1.14876068e+00 8.59743476e-01
7.81444252e-01 -6.38113022e-01 -1.06899047e+00 -7.04850554e-01
-3.57371926e-01 -3.48721176e-01 6.26870930e-01 7.71709740e-01
-3.19480933e-02 -6.01734268e-03 -8.53397131e-01 1.62920311e-01
5.22174299e-01 -6.19386673e-01 8.99215579e-01 -1.45389962e+00
8.84330943e-02 -1.90298975e-01 -7.19660997e-01 -6.27284467e-01
2.03615278e-01 -2.68176764e-01 -3.46559376e-01 -7.52581000e-01
-1.41860709e-01 -3.83341849e-01 -2.60651588e-01 6.77854300e-01
9.30372477e-02 6.44339383e-01 2.59284731e-02 1.54367700e-01
1.64177760e-01 2.33448863e-01 7.49855816e-01 -6.65144026e-02
-3.25294554e-01 3.94180834e-01 -4.73107874e-01 7.36058235e-01
9.37460303e-01 -1.01840556e-01 -2.73465663e-01 1.35504037e-01
-1.98885962e-01 -2.67560035e-01 5.19032851e-02 -1.09881866e+00
3.29061776e-01 2.91106641e-01 6.80835903e-01 -2.06546143e-01
7.62365580e-01 -1.13128781e+00 4.47763428e-02 5.38644254e-01
2.03158274e-01 1.26598492e-01 5.20390034e-01 2.17452988e-01
-1.09884560e-01 -4.85862270e-02 1.04816425e+00 1.91665590e-01
-6.26643836e-01 3.17554355e-01 -4.66024816e-01 -5.00442982e-01
1.06057084e+00 -8.88878584e-01 -6.76355585e-02 -5.33570170e-01
-6.18801057e-01 -5.34875154e-01 3.93294454e-01 5.66319585e-01
7.92887926e-01 -1.04941893e+00 -3.62436831e-01 1.00346220e+00
-8.76565129e-02 -8.43267262e-01 1.43697694e-01 7.98219085e-01
-3.17224383e-01 5.10985553e-01 -5.69857240e-01 -6.52515829e-01
-2.00588703e+00 2.12568864e-01 5.36360562e-01 2.30104566e-01
-3.34708184e-01 7.19998240e-01 -3.84450972e-01 -1.64808914e-01
5.81167936e-01 8.94079283e-02 -4.97887790e-01 -2.56197825e-02
9.16308939e-01 9.72473621e-02 2.03154743e-01 -1.00614846e+00
-7.35211849e-01 9.38420713e-01 -8.50438476e-02 -1.41452253e-01
9.59519744e-01 1.14322908e-01 7.38078728e-02 2.60683261e-02
9.51921821e-01 3.77622366e-01 -9.72334266e-01 1.62372235e-02
1.02302969e-01 -6.58605814e-01 -2.94915736e-01 -7.30026305e-01
-1.13152719e+00 1.01000142e+00 1.20861959e+00 9.66209918e-02
1.13980532e+00 -5.40364563e-01 6.54983938e-01 3.14381778e-01
4.32803929e-01 -9.24972951e-01 -4.33977284e-02 2.26397410e-01
5.19359410e-01 -1.45547605e+00 -3.42339985e-02 -1.30377293e-01
-4.61682886e-01 1.26602399e+00 5.56511402e-01 4.05730218e-01
8.90978456e-01 2.69812673e-01 3.41707587e-01 1.44149318e-01
-1.81917340e-01 5.49039468e-02 4.38647449e-01 1.10852242e+00
5.93264878e-01 -2.01778263e-01 -2.61996090e-01 4.63532329e-01
1.70912929e-02 2.28057235e-01 3.94982904e-01 8.28696251e-01
-9.73603949e-02 -1.27794969e+00 -4.19076145e-01 3.10722232e-01
-9.54487562e-01 1.98132977e-01 -4.63384062e-01 1.09078610e+00
-1.04026794e-02 1.22275805e+00 -2.97468752e-01 -5.98967373e-01
1.66415140e-01 4.02611107e-01 4.40712035e-01 -1.44603863e-01
-6.71542823e-01 -1.03198111e-01 -9.90946144e-02 -6.24000311e-01
-7.16898620e-01 -6.52556241e-01 -7.81077802e-01 -4.69928563e-01
-5.37275493e-01 -6.17340393e-02 9.89915609e-01 1.00225735e+00
3.85356903e-01 -2.87328452e-01 7.12184429e-01 -8.40516746e-01
-4.37383175e-01 -1.01646709e+00 -5.60763419e-01 1.74187794e-01
3.36395234e-01 -7.26823866e-01 -2.57084936e-01 -1.70146849e-03] | [13.303565979003906, 1.1823457479476929] |
3be86768-7c4d-4e28-a335-5e1d9439b5ac | nlatool-an-application-for-enhanced-deep-text | null | null | https://aclanthology.org/C18-2026 | https://aclanthology.org/C18-2026.pdf | NLATool: an Application for Enhanced Deep Text Understanding | Today, we see an ever growing number of tools supporting text annotation. Each of these tools is optimized for specific use-cases such as named entity recognition. However, we see large growing knowledge bases such as Wikipedia or the Google Knowledge Graph. In this paper, we introduce NLATool, a web application developed using a human-centered design process. The application combines supporting text annotation and enriching the text with additional information from a number of sources directly within the application. The tool assists users to efficiently recognize named entities, annotate text, and automatically provide users additional information while solving deep text understanding tasks. | ['Eric H{\\"a}mmerle', 'Markus G{\\"a}rtner', 'Valentin Schwind', 'Sven Mayer', 'Jonas Kuhn', 'Lars Lischke', 'Florin Rheinwald', 'Emine Turcan', 'Gustav Murawski'] | 2018-08-01 | nlatool-an-application-for-enhanced-deep-text-1 | https://aclanthology.org/C18-2026 | https://aclanthology.org/C18-2026.pdf | coling-2018-8 | ['text-annotation'] | ['natural-language-processing'] | [-4.81593639e-01 4.95514899e-01 -4.03116763e-01 -2.41116405e-01
-3.41210812e-01 -8.77701938e-01 4.62376446e-01 7.62629867e-01
-6.04920149e-01 7.48223662e-01 5.41057110e-01 -2.38428757e-01
-1.63313225e-01 -7.98514605e-01 -2.23138295e-02 3.48355561e-01
4.94458258e-01 7.93358207e-01 2.32344747e-01 -2.43068367e-01
2.96792358e-01 2.79585630e-01 -1.40408540e+00 4.33359534e-01
1.44443262e+00 5.65729797e-01 2.03762874e-01 3.49903733e-01
-1.45094407e+00 9.18382704e-01 -5.47847509e-01 -6.56852961e-01
-5.48133627e-02 9.33309346e-02 -1.17026889e+00 -3.66315782e-01
5.23135662e-01 -8.49055573e-02 -1.61456451e-01 8.35929155e-01
3.94856244e-01 1.42138109e-01 5.27892262e-02 -1.20449662e+00
-7.39490390e-01 9.84843314e-01 1.56149361e-03 -8.31386000e-02
7.00875044e-01 -4.77528602e-01 8.80857587e-01 -1.12226987e+00
9.60625708e-01 6.86054230e-01 7.82039881e-01 4.81977522e-01
-5.12831509e-01 -3.89319628e-01 -8.06378499e-02 -9.60001547e-04
-1.20685375e+00 -3.78523201e-01 4.10266429e-01 -7.00286925e-01
1.36901903e+00 1.11198612e-01 4.13245529e-01 6.31741047e-01
-2.81571716e-01 7.27647364e-01 4.35701936e-01 -6.77617729e-01
9.41412374e-02 6.34009123e-01 7.66521752e-01 1.05034566e+00
6.43944144e-01 -7.16566861e-01 -9.13075030e-01 -2.48887911e-01
3.97963673e-01 -9.51126888e-02 -1.74531683e-01 -3.66160184e-01
-1.02747118e+00 2.91231215e-01 -1.93191227e-02 6.46369457e-01
-6.24245286e-01 -1.13334350e-01 7.39752054e-01 -1.10312030e-01
6.44959509e-01 8.62378001e-01 -8.46125245e-01 -3.70101869e-01
-8.23586702e-01 -8.45083967e-02 1.54189074e+00 1.32512867e+00
9.25672948e-01 -3.71893555e-01 -1.92710012e-01 8.34693551e-01
5.38270950e-01 1.62268594e-01 7.77889311e-01 -6.87232375e-01
7.22960949e-01 1.64233637e+00 4.64733064e-01 -6.99186146e-01
-6.16413116e-01 -2.97039926e-01 -2.11080045e-01 -2.24687974e-03
5.50603271e-01 -4.00554150e-01 -7.42046952e-01 1.10365379e+00
4.55553353e-01 -3.91033918e-01 3.80164199e-02 3.24423403e-01
1.39204812e+00 2.43427724e-01 7.59691298e-01 2.03768805e-01
1.53004432e+00 -1.03599441e+00 -1.37675691e+00 -4.63123739e-01
1.34891927e+00 -7.51759171e-01 1.16568339e+00 5.94577491e-02
-7.76244998e-01 -3.14168304e-01 -5.43305218e-01 -6.65844858e-01
-1.50511467e+00 5.14514685e-01 6.32940352e-01 7.50045836e-01
-1.09070826e+00 4.23227102e-01 -5.84068120e-01 -9.46762562e-01
4.22702432e-01 1.28595605e-01 -6.00876987e-01 1.25163123e-01
-1.25181210e+00 1.14985144e+00 9.32614088e-01 -3.62107605e-01
-6.78698346e-02 -9.94831562e-01 -9.43458915e-01 2.66865134e-01
5.30922294e-01 -9.70815837e-01 1.29494524e+00 -8.05136383e-01
-1.26508772e+00 1.15077102e+00 -1.57312766e-01 -5.09723537e-02
2.45829791e-01 -8.93856525e-01 -5.02775133e-01 -2.68078029e-01
3.28105092e-01 2.52432138e-01 9.09919292e-02 -8.12701166e-01
-7.39797235e-01 -7.37285972e-01 2.63297595e-02 4.40441549e-01
-1.02773452e+00 4.08150434e-01 -6.10971630e-01 -5.32260418e-01
-4.98436153e-01 -2.93976784e-01 -1.96087494e-01 3.12725715e-02
-4.76829976e-01 -5.43394387e-01 1.09792423e+00 -1.14489090e+00
1.88039911e+00 -1.79594147e+00 -1.43618941e-01 8.98606107e-02
5.25667548e-01 3.26847166e-01 2.07145855e-01 1.02952743e+00
5.93803637e-02 6.77794993e-01 1.95309427e-02 -3.88994843e-01
1.76585883e-01 1.46889966e-02 -1.85999889e-02 -3.48936528e-01
-4.06827301e-01 9.91414964e-01 -9.53152180e-01 -6.58528507e-01
2.33995225e-02 4.97508675e-01 -1.49880558e-01 -2.50297934e-02
-4.77309674e-01 2.02349294e-02 -8.33954632e-01 4.91544902e-01
7.50069991e-02 -6.07250929e-01 1.97384581e-01 -1.15355618e-01
-4.49396223e-01 5.31741977e-01 -1.22781706e+00 2.15150380e+00
-6.64242387e-01 8.14209223e-01 9.85016674e-02 -4.01739448e-01
6.57867193e-01 6.33338869e-01 4.13454682e-01 -1.93465054e-01
-5.13005555e-02 2.68585354e-01 -7.67246485e-01 -9.63393927e-01
9.38557208e-01 5.92179179e-01 8.12633708e-02 5.89641154e-01
3.71281564e-01 1.37735069e-01 5.13773263e-01 6.79302514e-01
1.25996935e+00 4.38881487e-01 9.44601238e-01 -1.40070915e-01
4.24150914e-01 4.37894762e-01 -4.35453281e-03 5.93557239e-01
1.48320168e-01 -3.20397347e-01 -7.89774880e-02 -5.24361968e-01
-1.01741374e+00 -4.88794476e-01 -2.48891823e-02 1.50151002e+00
-2.28892475e-01 -1.13273108e+00 -8.25003028e-01 -9.91255283e-01
6.16826266e-02 1.02118707e+00 -2.04017818e-01 2.46877164e-01
-1.89100742e-01 7.18187019e-02 7.01470613e-01 6.82303250e-01
6.60009444e-01 -1.14343655e+00 -4.55477506e-01 1.36570394e-01
-2.79292047e-01 -1.13633072e+00 -3.24403495e-01 -2.70155054e-02
-8.69525433e-01 -9.00684237e-01 -3.20595771e-01 -8.90001237e-01
7.61395216e-01 -6.03458807e-02 1.52069807e+00 1.45545781e-01
-2.78554559e-01 9.14110720e-01 -4.96049225e-01 -6.23511970e-01
-1.15853526e-01 4.57258999e-01 -3.69725823e-01 -5.47624290e-01
8.29696715e-01 -2.32309893e-01 -1.52664810e-01 5.78896282e-03
-9.94368792e-01 2.52159655e-01 2.19654024e-01 1.91913858e-01
4.18388955e-02 1.08364984e-01 5.25744319e-01 -1.20787728e+00
8.44893813e-01 -6.90199196e-01 -2.92036146e-01 5.53758562e-01
-7.18958676e-01 9.79798362e-02 4.59491670e-01 -2.48541683e-02
-1.51163626e+00 4.24679577e-01 -6.15113862e-02 2.62970746e-01
-5.63709259e-01 1.28080678e+00 -4.12759632e-01 8.01896155e-02
9.57369268e-01 -2.09843352e-01 -6.11222148e-01 -8.61576200e-01
8.37316096e-01 1.01125050e+00 3.61208320e-01 -5.13441861e-01
6.42740369e-01 3.72523367e-02 -5.59148729e-01 -8.95666838e-01
-1.04478204e+00 -8.95882070e-01 -9.19001698e-01 -3.24428141e-01
7.72111952e-01 -8.09484422e-01 -5.51904023e-01 8.41868445e-02
-1.26000690e+00 -3.70010138e-01 -6.11557782e-01 2.42457777e-01
-6.36635199e-02 3.31233591e-01 -2.63458550e-01 -7.31280506e-01
-6.99040294e-01 -2.70393103e-01 8.74761760e-01 5.12798846e-01
-5.20991087e-01 -1.42243814e+00 1.51935518e-01 6.22054696e-01
3.91967714e-01 4.19775583e-02 9.43797886e-01 -1.60892463e+00
-2.03489915e-01 -5.53281248e-01 -2.04846203e-01 1.92242842e-02
3.00473362e-01 2.82391515e-02 -8.13817620e-01 4.59535301e-01
-6.34207666e-01 -3.05958483e-02 1.07128248e-01 -3.68665308e-01
1.15192056e+00 -5.11418402e-01 -7.67680287e-01 1.43750653e-01
1.26179636e+00 2.18259856e-01 5.17410874e-01 6.01893246e-01
1.00890315e+00 9.39703643e-01 3.58110607e-01 4.41150576e-01
5.51143169e-01 5.05904555e-01 1.97124586e-01 1.48025051e-01
4.93904166e-02 -3.87402087e-01 -4.05284017e-02 5.48986673e-01
1.49773598e-01 -5.35197735e-01 -1.56279135e+00 5.91877818e-01
-2.16357517e+00 -7.73049295e-01 -4.21417326e-01 1.91934991e+00
1.03793490e+00 -2.85384268e-01 -1.68880343e-01 -3.01445723e-01
7.03938782e-01 -3.83006364e-01 -5.61401784e-01 1.65503412e-01
1.65443301e-01 3.22817624e-01 4.27705348e-01 3.33864123e-01
-8.06714237e-01 1.23888469e+00 6.15891790e+00 5.60247540e-01
-6.43892467e-01 8.82377177e-02 5.45465723e-02 1.65981695e-01
-1.97746530e-01 2.04218775e-02 -1.02654314e+00 1.78192288e-01
8.54999304e-01 -8.82513225e-01 1.63753971e-01 1.33471024e+00
1.65350437e-01 3.85123007e-02 -1.00271869e+00 8.47215891e-01
1.48621187e-01 -1.49360943e+00 1.20640785e-01 -1.59443438e-01
4.12580371e-01 3.18861790e-02 -5.29949427e-01 4.20933694e-01
5.68757594e-01 -7.85724819e-01 3.05171043e-01 7.58266985e-01
6.82598352e-01 -2.39946485e-01 5.83732188e-01 4.90182877e-01
-1.07666731e+00 1.32733062e-01 3.32407624e-01 1.80011690e-01
2.04152241e-01 8.03944886e-01 -9.19363201e-01 5.18956542e-01
8.01390469e-01 5.99598706e-01 -9.03617740e-01 1.41441643e+00
-3.42990935e-01 1.55886307e-01 -4.94229168e-01 -1.39619917e-01
-3.66251141e-01 -2.62143724e-02 5.55142343e-01 1.45228803e+00
4.16486502e-01 1.59965292e-01 2.45024532e-01 4.54572678e-01
-5.20210981e-01 8.23177338e-01 -9.23321664e-01 -7.79811323e-01
6.52887642e-01 1.70199096e+00 -7.54482806e-01 -6.14721417e-01
-6.67100608e-01 1.15902174e+00 3.84055763e-01 2.99230963e-01
-3.49114567e-01 -1.08765411e+00 4.46689039e-01 4.42643315e-01
-1.84450492e-01 -4.62883174e-01 -5.93140006e-01 -1.35472989e+00
5.33016622e-02 -6.87609136e-01 5.37790000e-01 -1.30353236e+00
-1.10494602e+00 4.01013643e-01 -4.24903370e-02 -6.51538014e-01
-1.74481183e-01 -7.86497176e-01 -4.26365793e-01 7.75844693e-01
-9.67337787e-01 -1.11772096e+00 -8.08770537e-01 5.44985592e-01
5.60167968e-01 -1.72544941e-02 1.04388821e+00 7.44100869e-01
-5.46881616e-01 -3.86010520e-02 5.73917069e-02 4.49196547e-01
9.78994489e-01 -1.32875597e+00 5.16617417e-01 5.86905003e-01
2.59035021e-01 9.37341630e-01 3.22208136e-01 -1.31786096e+00
-1.06541610e+00 -9.23484087e-01 1.50753546e+00 -9.66126323e-01
9.51404393e-01 -2.52437711e-01 -1.10008359e+00 1.02434814e+00
5.41062593e-01 -1.42206177e-01 1.27749753e+00 2.97759145e-01
-4.01339322e-01 3.94194752e-01 -1.04055285e+00 5.14920831e-01
1.23297215e+00 -6.84297860e-01 -8.91721487e-01 5.80877721e-01
7.10603416e-01 -6.47052705e-01 -9.39480305e-01 2.06785556e-02
3.22493285e-01 -2.68427700e-01 6.95445299e-01 -1.01445174e+00
9.52425152e-02 -3.59946460e-01 3.19551587e-01 -1.19807756e+00
6.00976497e-02 -6.48118913e-01 -6.06067777e-01 1.56578195e+00
8.23693573e-01 -4.54602957e-01 8.08952212e-01 1.46968281e+00
-3.03308755e-01 4.04361971e-02 -4.70949948e-01 -4.07466590e-01
-6.16793513e-01 -3.99523497e-01 4.61004883e-01 1.57992470e+00
7.94530809e-01 5.90027153e-01 2.46753380e-01 3.51253636e-02
1.83679610e-01 -5.11499286e-01 6.81356728e-01 -1.85526812e+00
2.55360693e-01 -3.37227613e-01 -3.25298076e-03 -7.32665539e-01
1.03978999e-01 -1.10020208e+00 -3.44465435e-01 -2.48229790e+00
5.33451959e-02 -2.32045978e-01 2.04643041e-01 1.08819747e+00
-1.19098723e-01 -4.45780307e-01 -1.60499200e-01 1.65568963e-01
-9.27305341e-01 1.38000175e-01 5.91646254e-01 2.17961743e-01
-5.01208782e-01 -3.37653756e-01 -7.88351357e-01 1.04973233e+00
9.12144065e-01 -4.63366956e-01 -2.10253358e-01 -6.68219864e-01
1.13448107e+00 -4.07786399e-01 5.04084751e-02 -1.02771211e+00
8.18496466e-01 -8.34844932e-02 4.00203645e-01 -5.09377718e-01
-2.43822128e-01 -1.20285118e+00 -1.25391437e-02 -8.66636187e-02
-5.53202093e-01 -3.52962269e-03 5.58273911e-01 2.61785626e-01
-1.80927366e-01 -6.57923520e-01 2.22913161e-01 -3.05313826e-01
-1.07153594e+00 8.62458423e-02 -5.39118230e-01 2.90585309e-01
1.02810085e+00 -1.35865137e-01 -7.64102876e-01 -2.83030301e-01
-9.46904242e-01 4.42989647e-01 2.75695622e-01 6.93215251e-01
1.05805308e-01 -1.13679874e+00 -1.65857285e-01 -3.22956473e-01
4.37570274e-01 3.02628092e-02 -1.04895430e-02 3.40695292e-01
-5.81005573e-01 5.31285048e-01 -9.17088091e-02 2.03177050e-01
-1.34115112e+00 3.56632829e-01 1.90130129e-01 -3.39520901e-01
-6.16759419e-01 4.41684306e-01 -3.11946332e-01 -7.39344895e-01
4.02354509e-01 -2.51015816e-02 -7.86813319e-01 3.69119227e-01
7.32605338e-01 7.12334752e-01 3.17023247e-01 -2.05961049e-01
-3.30638319e-01 2.65999347e-01 8.23861510e-02 -7.79074952e-02
1.31134546e+00 -3.44641834e-01 -3.75089675e-01 3.47160399e-01
6.66135430e-01 3.74606043e-01 -3.64465088e-01 -3.09194505e-01
8.58335853e-01 -5.99692240e-02 2.14835837e-01 -1.37981427e+00
-7.24723935e-01 5.41820288e-01 2.36063406e-01 6.24301672e-01
8.60613883e-01 1.46928906e-01 5.79505503e-01 1.22532940e+00
2.01381773e-01 -1.34176600e+00 -2.13911027e-01 8.53381634e-01
7.89871931e-01 -1.06587982e+00 2.67035533e-02 -4.83527094e-01
-5.65815866e-01 1.38260043e+00 9.60332930e-01 7.25404322e-01
6.75475061e-01 4.35114145e-01 7.33375251e-02 -6.13867462e-01
-4.76973116e-01 -4.94366914e-01 4.50344175e-01 7.97889113e-01
1.05484891e+00 -5.70410967e-01 -2.67685831e-01 7.95110345e-01
-1.06183179e-01 4.80336487e-01 2.87269861e-01 1.00413597e+00
-5.86676240e-01 -1.29518855e+00 -1.46377474e-01 5.86309910e-01
-6.56941116e-01 -5.55463731e-01 -9.86088157e-01 7.49077439e-01
-6.19855076e-02 7.71994710e-01 1.37400135e-01 1.42789185e-01
5.38676381e-01 6.89374864e-01 -8.36902186e-02 -1.29704833e+00
-1.14125526e+00 -5.58036387e-01 7.36094832e-01 -4.65918660e-01
-2.25780189e-01 -1.97472051e-01 -1.74227679e+00 -2.93928385e-02
-3.71058553e-01 5.22551239e-01 1.09163892e+00 8.50202024e-01
8.67286444e-01 2.27429867e-01 -4.32376176e-01 -2.68978119e-01
3.47350180e-01 -8.45809340e-01 -2.14208513e-01 1.52701989e-01
-4.07752693e-01 -1.29296303e-01 -2.02626530e-02 5.03784299e-01] | [9.304187774658203, 8.757075309753418] |
1816bbb8-b87e-4cdf-949e-6b9756176e65 | better-modeling-the-programming-world-with | 2201.03346 | null | https://arxiv.org/abs/2201.03346v2 | https://arxiv.org/pdf/2201.03346v2.pdf | Better Modeling the Programming World with Code Concept Graphs-augmented Multi-modal Learning | The progress made in code modeling has been tremendous in recent years thanks to the design of natural language processing learning approaches based on state-of-the-art model architectures. Nevertheless, we believe that the current state-of-the-art does not focus enough on the full potential that data may bring to a learning process in software engineering. Our vision articulates on the idea of leveraging multi-modal learning approaches to modeling the programming world. In this paper, we investigate one of the underlying idea of our vision whose objective based on concept graphs of identifiers aims at leveraging high-level relationships between domain concepts manipulated through particular language constructs. In particular, we propose to enhance an existing pretrained language model of code by joint-learning it with a graph neural network based on our concept graphs. We conducted a preliminary evaluation that shows gain of effectiveness of the models for code search using a simple joint-learning method and prompts us to further investigate our research vision. | ['Bang Liu', 'Houari Sahraoui', 'Martin Weyssow'] | 2022-01-10 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-8.20106417e-02 4.64907318e-01 -2.47973070e-01 -4.18475956e-01
-4.18220729e-01 -2.69868433e-01 6.89211667e-01 5.44066966e-01
-3.70750166e-02 -1.27109230e-01 3.88808668e-01 -8.62105310e-01
-3.04097801e-01 -7.98859715e-01 -7.75372446e-01 2.20984578e-01
-4.03445989e-01 3.02728534e-01 9.54978764e-02 -3.87354076e-01
5.69899142e-01 1.07161887e-01 -1.54710412e+00 6.97581470e-01
4.63983417e-01 4.44739550e-01 1.25663817e-01 4.31122869e-01
-7.55797386e-01 1.51920319e+00 -2.96635568e-01 -4.28861111e-01
-8.96512717e-02 -7.29787201e-02 -1.00286055e+00 -8.74178931e-02
3.98603231e-01 -8.98149759e-02 -1.41796440e-01 8.86746764e-01
-1.79833725e-01 -2.45512977e-01 3.26669663e-01 -1.29667652e+00
-1.22256804e+00 9.90470171e-01 -4.00324672e-01 7.50301555e-02
5.52042067e-01 1.04933254e-01 1.29859400e+00 -7.46884167e-01
6.86594903e-01 1.11301255e+00 9.86291885e-01 6.90393865e-01
-1.22730851e+00 -3.54651630e-01 4.05169368e-01 1.21971510e-01
-1.01903594e+00 -1.97824493e-01 1.04286635e+00 -8.67480755e-01
1.57701421e+00 -2.48075530e-01 6.15204036e-01 6.75687909e-01
3.01453233e-01 8.93843889e-01 7.60918379e-01 -9.67057586e-01
-5.53006977e-02 3.28101516e-01 4.28243548e-01 1.33994174e+00
4.30053622e-02 1.09733962e-01 -3.92502159e-01 -3.08416665e-01
7.02088356e-01 -3.25674489e-02 2.95673817e-01 -1.00328434e+00
-8.49895477e-01 9.43300426e-01 4.50138420e-01 8.08493614e-01
8.40117969e-03 6.17576659e-01 5.76563954e-01 4.70571548e-01
3.87428552e-01 6.54921412e-01 -4.65898901e-01 -3.11400414e-01
-1.07090020e+00 1.54214785e-01 1.02688503e+00 9.95969772e-01
9.29408133e-01 -4.60125357e-02 3.24372321e-01 5.60439706e-01
9.29948092e-01 -3.33307713e-01 3.77787352e-01 -8.05731654e-01
3.21318746e-01 1.16491485e+00 -4.19282794e-01 -9.87426817e-01
-3.50717008e-01 -2.17792720e-01 1.60029069e-01 3.16108555e-01
1.80354804e-01 3.43489610e-02 -7.17095256e-01 1.56130815e+00
-3.27456892e-01 1.66943014e-01 -3.54216397e-02 1.93452343e-01
6.39280260e-01 3.46098363e-01 2.75904030e-01 2.96214908e-01
1.08517253e+00 -9.46947753e-01 -3.13871384e-01 -4.89729494e-01
1.13759184e+00 -5.37684083e-01 1.17364705e+00 3.50556761e-01
-1.09127796e+00 -6.20499969e-01 -1.14703929e+00 -6.33034557e-02
-7.16183245e-01 2.16694605e-02 1.31851995e+00 8.77805591e-01
-1.36427689e+00 2.51543999e-01 -1.01361644e+00 -6.63694024e-01
4.73529607e-01 2.48124421e-01 -2.54625082e-01 -2.72731692e-01
-7.95724154e-01 1.05503452e+00 4.30465937e-01 -3.02717805e-01
-1.07451379e+00 -8.89664114e-01 -1.15564919e+00 1.54655814e-01
5.56940615e-01 -7.87785292e-01 1.24282169e+00 -1.07183695e+00
-1.04105306e+00 1.12959313e+00 1.71735868e-01 -4.72302824e-01
-8.58697072e-02 -1.19672202e-01 -4.02707845e-01 -1.56615585e-01
-9.78320912e-02 4.16034788e-01 5.61217844e-01 -1.64076424e+00
-3.56398672e-01 -2.14465991e-01 7.49353468e-01 -3.77324373e-01
-7.95057714e-01 5.17310023e-01 -2.23258168e-01 -4.15155977e-01
-3.55216026e-01 -7.30840206e-01 -2.19748512e-01 1.38592690e-01
2.29671046e-01 -4.86377925e-01 7.55106986e-01 -3.78188133e-01
1.65191555e+00 -2.16103601e+00 1.54699340e-01 3.29246372e-02
3.50111425e-01 2.32328400e-01 -3.40992332e-01 1.00566077e+00
-3.95539671e-01 4.23952043e-01 -1.87474683e-01 -1.73131391e-01
2.89699465e-01 6.61059096e-02 -3.41933697e-01 1.53974637e-01
4.33986872e-01 1.02419627e+00 -1.12460279e+00 -4.20019954e-01
1.58260852e-01 3.40662569e-01 -9.56989944e-01 2.73758024e-01
-5.94779909e-01 -3.62089097e-01 -5.03972769e-01 6.54142857e-01
1.88351005e-01 -2.54960269e-01 4.82933789e-01 1.22896612e-01
-2.37120524e-01 3.59361231e-01 -8.72012258e-01 2.27460790e+00
-7.36743450e-01 5.88604689e-01 -2.35071629e-02 -1.43884718e+00
9.16425169e-01 1.08580977e-01 4.82490569e-01 -3.43615264e-01
-1.95435062e-01 6.78022057e-02 1.15830742e-01 -8.78425896e-01
3.83585989e-01 -1.42272815e-01 -1.46625087e-01 7.06260324e-01
2.44288921e-01 -2.82085568e-01 9.62952748e-02 3.88920784e-01
1.21197128e+00 6.52927518e-01 2.64256239e-01 -3.93396616e-01
6.43093228e-01 7.13408962e-02 -1.74784258e-01 6.20044470e-01
5.48396371e-02 2.95708533e-02 7.25280046e-01 -6.68929994e-01
-7.27331400e-01 -7.73153126e-01 1.46339595e-01 1.30488372e+00
-4.76069719e-01 -1.10347009e+00 -6.48823500e-01 -9.19615388e-01
3.56232673e-02 9.23587739e-01 -6.38341606e-01 -1.95841387e-01
-6.49317324e-01 -4.26559836e-01 6.74882650e-01 7.57061124e-01
-6.78632632e-02 -9.99887884e-01 -8.91884565e-01 1.82568431e-01
4.94149387e-01 -6.72645152e-01 1.40118405e-01 1.98630899e-01
-9.27953959e-01 -1.17855191e+00 -1.49214447e-01 -1.08579743e+00
5.79067409e-01 1.83655366e-01 1.69134176e+00 6.00974143e-01
-3.16799253e-01 1.07512534e+00 -4.62236702e-01 -6.77613318e-01
-8.25296164e-01 9.11199488e-03 -5.03496349e-01 -5.60289800e-01
8.46851051e-01 -5.74137866e-01 1.07891555e-03 -3.55911046e-01
-1.17017162e+00 -1.61650822e-01 3.82304370e-01 6.03201210e-01
-1.77208945e-01 1.55401438e-01 2.91083276e-01 -1.15904176e+00
9.44345653e-01 -8.24141264e-01 -5.68975627e-01 5.29189527e-01
-7.09659219e-01 3.43195975e-01 3.65120977e-01 -5.77181019e-02
-1.04122639e+00 -7.17367902e-02 1.09543920e-01 -1.49473488e-01
-1.33455202e-01 1.29975569e+00 1.91847518e-01 -4.14598405e-01
6.76065326e-01 6.30224571e-02 -1.06636599e-01 -2.44799361e-01
7.97690392e-01 5.06378293e-01 1.37072444e-01 -1.16960466e+00
7.94198215e-01 3.02052587e-01 -1.11492336e-01 -7.41329432e-01
-3.16136569e-01 -4.01574850e-01 -6.37237072e-01 -1.08821630e-01
6.81032479e-01 -8.29811633e-01 -3.96237195e-01 3.07243824e-01
-1.13698804e+00 -3.74055743e-01 -3.86350900e-02 6.39519095e-02
-6.82015002e-01 5.43837368e-01 -5.12698889e-01 -8.82858217e-01
1.82763189e-01 -1.06592357e+00 8.31085682e-01 -1.35308400e-01
-2.28254780e-01 -1.54237401e+00 2.39374802e-01 3.61686915e-01
6.54806793e-01 1.51037619e-01 1.64097369e+00 -6.09585583e-01
-8.69311690e-01 -2.91867077e-01 -1.56044483e-01 1.96108773e-01
7.53247216e-02 2.56807536e-01 -8.07938039e-01 -1.37174428e-01
-2.39764880e-02 -7.12965488e-01 6.14294648e-01 5.06370142e-02
1.23340094e+00 1.04194820e-01 -4.32681322e-01 3.35506618e-01
1.83854532e+00 6.92196786e-02 4.03881967e-01 5.62694907e-01
8.41019869e-01 6.78934395e-01 3.32804859e-01 3.35856736e-01
8.45935702e-01 4.66559857e-01 7.14725494e-01 7.63348863e-02
-1.64363965e-01 -3.36999267e-01 3.34568918e-01 8.12483728e-01
1.60017852e-02 1.17994204e-01 -1.62927055e+00 8.50434422e-01
-1.87797332e+00 -7.09645748e-01 1.68868318e-01 1.83095729e+00
7.73511767e-01 2.85298318e-01 -2.33801864e-02 -3.15355539e-01
1.43038660e-01 2.81970829e-01 -1.62854165e-01 -5.75402677e-01
5.04553556e-01 2.45673150e-01 -5.88940047e-02 3.25766712e-01
-1.03603411e+00 8.74126911e-01 6.81283426e+00 2.03785807e-01
-9.36601043e-01 -1.87959328e-01 -2.84913983e-02 4.00017321e-01
-6.44585371e-01 4.59688127e-01 -5.18530011e-01 -3.97110172e-02
1.19068730e+00 -3.63141030e-01 5.82276821e-01 1.14314699e+00
-1.69356689e-01 1.45327970e-01 -1.64570379e+00 6.50918305e-01
4.88728613e-01 -1.50290751e+00 1.55302480e-01 1.45223169e-02
6.01526797e-01 2.16853783e-01 -7.15409145e-02 7.89828002e-01
6.45648777e-01 -1.06468928e+00 5.73342800e-01 6.40824139e-01
1.66009769e-01 -4.80054617e-01 4.43359137e-01 4.66976643e-01
-1.21748888e+00 -4.77515906e-01 -2.49041885e-01 -4.26014900e-01
-2.77478963e-01 1.23935729e-01 -8.47863436e-01 6.64514422e-01
3.32387120e-01 1.04511106e+00 -8.84770811e-01 8.65485489e-01
-1.27051607e-01 3.82658333e-01 6.02186061e-02 -8.57827961e-02
3.25194895e-01 1.98387936e-01 1.09898686e-01 1.65348351e+00
2.27254499e-02 -3.32653970e-01 5.49357474e-01 1.31978703e+00
-1.05364263e-01 7.99353328e-03 -1.02679205e+00 -6.06340051e-01
1.93426788e-01 9.89618540e-01 -5.02591670e-01 -3.09706807e-01
-1.34809852e+00 9.37177837e-02 6.62736475e-01 1.57010585e-01
-6.58568025e-01 -2.76152581e-01 1.90890953e-01 1.13733575e-01
3.35791647e-01 -5.69970012e-01 -2.95995593e-01 -1.23163545e+00
5.03435917e-02 -1.21179283e+00 4.38181311e-01 -8.12027991e-01
-1.13712108e+00 3.65860373e-01 2.65033096e-01 -7.51968801e-01
-2.92600304e-01 -9.02637959e-01 -6.62982643e-01 7.92075157e-01
-1.63041127e+00 -1.66070843e+00 2.39105932e-02 3.55527788e-01
5.82471848e-01 -4.44944322e-01 1.17495382e+00 1.61241353e-01
-7.34409988e-02 4.50284362e-01 -4.44131106e-01 2.29785025e-01
4.38301116e-01 -1.27688098e+00 5.63154042e-01 9.65460658e-01
5.26342332e-01 1.37978292e+00 5.59113979e-01 -5.05280852e-01
-1.64452159e+00 -7.06310391e-01 9.21584606e-01 -8.64164770e-01
1.29585326e+00 -2.78000832e-01 -1.13055933e+00 1.07059276e+00
4.25303221e-01 5.09501733e-02 8.04848909e-01 4.47401434e-01
-8.19703817e-01 1.65839165e-01 -6.85104787e-01 3.13398719e-01
8.70253325e-01 -1.15162969e+00 -1.18892479e+00 1.71557590e-01
5.94506085e-01 -1.33915499e-01 -9.67423797e-01 2.13840812e-01
4.30042565e-01 -9.58377242e-01 1.01245952e+00 -1.17011285e+00
8.80523503e-01 -1.78749025e-01 -3.07130069e-01 -1.05964756e+00
-2.83193260e-01 -3.57761323e-01 -2.55549490e-01 1.29721260e+00
3.23783964e-01 -1.50287047e-01 8.75333607e-01 7.93215096e-01
-2.99986780e-01 -6.45356178e-01 -3.90687615e-01 -5.96220136e-01
4.46080863e-01 -8.68786156e-01 4.07502472e-01 1.24493098e+00
6.50609255e-01 2.32810348e-01 5.90059049e-02 -5.54949557e-03
4.33430612e-01 1.09719038e-01 7.38899887e-01 -1.36272109e+00
-6.66376114e-01 -5.07413328e-01 -4.56742346e-01 -8.04307699e-01
6.39199138e-01 -1.19687152e+00 -2.61705905e-01 -1.73458278e+00
3.87917548e-01 -2.74074823e-01 -4.65820402e-01 7.69294202e-01
4.46709953e-02 -4.50167328e-01 3.46773386e-01 -5.30460738e-02
-6.03119850e-01 9.34626833e-02 6.37671471e-01 -3.85058582e-01
9.29658413e-02 -3.21745634e-01 -8.81889224e-01 7.96776891e-01
5.60279727e-01 -3.95870805e-01 -8.23549569e-01 -9.53768075e-01
6.97586894e-01 5.43909818e-02 3.53676498e-01 -9.91467774e-01
4.01193678e-01 -2.16972781e-03 -2.53511667e-01 3.73946205e-02
-1.04097769e-01 -9.37166035e-01 -2.33845398e-01 5.06818354e-01
-6.21728539e-01 9.78977531e-02 5.91438770e-01 5.58210850e-01
-3.24729443e-01 -6.17991865e-01 3.13088685e-01 -6.35606170e-01
-1.26983690e+00 -1.80409357e-01 -4.12036926e-01 3.26741375e-02
7.76445687e-01 -1.39100835e-01 -3.94802004e-01 -6.31080493e-02
-4.40106511e-01 1.19411908e-01 6.14290118e-01 9.58438575e-01
4.89922613e-01 -8.22589278e-01 -4.24263835e-01 2.55656868e-01
6.83906913e-01 -5.97803652e-01 -2.98731297e-01 2.43670598e-01
-4.19689238e-01 5.56949854e-01 -3.15576673e-01 -3.05330455e-01
-1.20870316e+00 1.06663203e+00 3.01017761e-01 -4.36895132e-01
-3.06216836e-01 8.36160958e-01 -1.24601705e-03 -5.49570799e-01
2.99634427e-01 -4.80839103e-01 -3.92326772e-01 -2.01917797e-01
3.14787716e-01 4.17677574e-02 9.03422236e-02 -2.26317734e-01
-4.75737363e-01 6.33271933e-01 -2.28895649e-01 2.23472238e-01
1.65566659e+00 3.27414811e-01 -5.23816884e-01 6.45551860e-01
1.04572451e+00 -1.36572933e-02 -8.50235760e-01 -2.65849710e-01
7.81092465e-01 -2.98913270e-01 1.80849984e-01 -9.04183567e-01
-8.55705380e-01 1.09109414e+00 4.56642807e-01 4.04826313e-01
8.47331643e-01 1.65407225e-01 1.42392844e-01 6.55800164e-01
6.93620861e-01 -7.98230767e-01 3.66223991e-01 6.56542480e-01
7.22032428e-01 -1.23797047e+00 2.20750511e-01 -3.15152824e-01
-2.26140171e-01 1.76800072e+00 6.02394581e-01 -2.95685768e-01
9.16041911e-01 5.50849557e-01 1.55248091e-01 -8.06531549e-01
-9.41296518e-01 -2.28144079e-02 6.15323186e-02 8.81216049e-01
1.18148768e+00 -1.55097470e-01 -1.55092567e-01 3.15497220e-01
3.33924204e-01 4.23711091e-01 7.86148131e-01 1.38240480e+00
-3.72979730e-01 -1.67293429e+00 -5.44686131e-02 2.67826498e-01
-4.07472730e-01 -3.87786388e-01 -4.76352006e-01 9.75302696e-01
1.83744043e-01 8.28559935e-01 -2.53804713e-01 -3.39932591e-01
3.55666101e-01 2.78646201e-01 6.16314471e-01 -1.29535365e+00
-8.00384820e-01 -3.84498924e-01 1.70728907e-01 -6.11827910e-01
-8.10546160e-01 -6.49482191e-01 -1.18024385e+00 1.95822477e-01
-1.62001207e-01 -4.56616431e-02 6.85403764e-01 8.67151022e-01
2.24921793e-01 7.51174092e-01 2.63842382e-02 -5.98871350e-01
-5.66640556e-01 -6.23238027e-01 -2.64199406e-01 2.42239550e-01
2.37548321e-01 -4.20007408e-01 -1.98461916e-02 3.49157661e-01] | [7.6586737632751465, 7.787206172943115] |
bac28a73-c144-4a47-a4ff-ff12b014011f | semi-supervised-confidence-level-based | 2211.15066 | null | https://arxiv.org/abs/2211.15066v1 | https://arxiv.org/pdf/2211.15066v1.pdf | Semi-Supervised Confidence-Level-based Contrastive Discrimination for Class-Imbalanced Semantic Segmentation | To overcome the data-hungry challenge, we have proposed a semi-supervised contrastive learning framework for the task of class-imbalanced semantic segmentation. First and foremost, to make the model operate in a semi-supervised manner, we proposed the confidence-level-based contrastive learning to achieve instance discrimination in an explicit manner, and make the low-confidence low-quality features align with the high-confidence counterparts. Moreover, to tackle the problem of class imbalance in crack segmentation and road components extraction, we proposed the data imbalance loss to replace the traditional cross entropy loss in pixel-level semantic segmentation. Finally, we have also proposed an effective multi-stage fusion network architecture to improve semantic segmentation performance. Extensive experiments on the real industrial crack segmentation and the road segmentation demonstrate the superior effectiveness of the proposed framework. Our proposed method can provide satisfactory segmentation results with even merely 3.5% labeled data. | ['Kangcheng Liu'] | 2022-11-28 | null | null | null | null | ['road-segementation'] | ['computer-vision'] | [ 5.22056103e-01 3.01265806e-01 -3.59541118e-01 -7.83007264e-01
-1.11006343e+00 -2.22184919e-02 -3.03815864e-02 2.04693094e-01
-3.79890263e-01 6.73094034e-01 -3.95414859e-01 -1.07415058e-01
-2.15513110e-01 -9.17127192e-01 -5.42165518e-01 -8.62433910e-01
3.74587029e-01 3.27569991e-01 3.64757866e-01 2.80826867e-01
6.02504790e-01 2.22022340e-01 -1.75397062e+00 2.26362929e-01
1.53556359e+00 1.23846149e+00 1.33308530e-01 2.55791664e-01
-2.06161991e-01 6.25748754e-01 -4.62669820e-01 -1.42817885e-01
2.06417784e-01 -1.89602807e-01 -9.04477239e-01 4.79940504e-01
3.03074270e-01 -1.69513434e-01 1.37583852e-01 1.21499455e+00
4.09561157e-01 -1.13276109e-01 8.01153183e-01 -1.32840550e+00
-4.14773449e-02 5.97120404e-01 -1.13961077e+00 9.54729095e-02
-1.39576927e-01 3.19859684e-02 9.37661886e-01 -7.84666300e-01
1.23078160e-01 1.13094723e+00 5.92552602e-01 8.86338502e-02
-7.76423752e-01 -7.87569880e-01 1.67146876e-01 1.96146175e-01
-1.22590053e+00 -1.42480284e-01 1.17497659e+00 -3.95031214e-01
4.53119487e-01 7.18939453e-02 5.56894898e-01 3.23136419e-01
-3.79665196e-02 1.02115285e+00 1.46607721e+00 -4.81924683e-01
8.13436359e-02 -6.33038729e-02 5.67919493e-01 9.60829198e-01
4.16826755e-01 -1.89997002e-01 -9.09890383e-02 1.53805166e-01
5.84257782e-01 -1.41287521e-01 1.17053855e-02 -2.30899110e-01
-9.79529023e-01 7.37143576e-01 5.27949154e-01 2.33667567e-01
-1.27744302e-01 -1.27338022e-01 4.87403691e-01 -1.87074170e-01
7.48572469e-01 1.88805580e-01 -5.70730090e-01 2.93452442e-01
-1.19787633e+00 -7.03042522e-02 3.52319390e-01 5.69616914e-01
1.01667714e+00 1.07570402e-01 -8.11276883e-02 9.25311208e-01
6.09387696e-01 3.40347588e-01 3.19515526e-01 -9.61281598e-01
7.05895483e-01 7.39774644e-01 -3.21019888e-01 -1.11846447e+00
-3.92999470e-01 -6.50145471e-01 -9.12548244e-01 4.08336252e-01
3.75981420e-01 1.06290141e-02 -1.21209896e+00 1.34807467e+00
4.73173112e-01 1.36756212e-01 -2.32207160e-02 8.41957450e-01
5.37093639e-01 5.15906632e-01 1.63695499e-01 -4.75854784e-01
1.26698673e+00 -1.00825441e+00 -8.41003597e-01 -5.47911189e-02
4.70091432e-01 -7.37575114e-01 1.03694403e+00 4.77130562e-01
-9.64628637e-01 -7.57184744e-01 -1.47448254e+00 1.48648843e-01
-1.66218609e-01 3.60679626e-01 5.71928442e-01 7.49605775e-01
-3.42181593e-01 6.82776392e-01 -7.46506751e-01 2.78920680e-01
8.68822038e-01 3.56566548e-01 -9.10872072e-02 4.88067307e-02
-1.16957271e+00 4.47181612e-01 9.30472970e-01 5.12207210e-01
-3.73372376e-01 -4.20335948e-01 -7.79953361e-01 -2.25293972e-02
5.11080921e-01 -3.04623514e-01 9.70555484e-01 -1.02496731e+00
-1.24200916e+00 9.16727126e-01 1.80838257e-01 -2.09081814e-01
4.26265717e-01 -2.17971101e-01 -3.81859243e-01 3.78956169e-01
2.31057644e-01 5.53325176e-01 8.00640166e-01 -1.52269208e+00
-8.00387979e-01 -6.10099852e-01 -1.22621618e-01 4.52540189e-01
-1.80866733e-01 -4.69332665e-01 -2.50354439e-01 -6.79266691e-01
7.15732932e-01 -5.29414177e-01 -2.90231973e-01 -9.85507965e-02
-6.50766790e-01 -1.28267810e-01 9.81707692e-01 -9.14906561e-01
1.13197505e+00 -2.04967594e+00 -3.08661848e-01 4.93299574e-01
1.94230869e-01 2.69451469e-01 3.83822471e-01 -4.79841202e-01
-1.47873119e-01 2.21099332e-01 -1.08120334e+00 -9.12374705e-02
-2.22426027e-01 8.37722868e-02 2.00419426e-01 3.75575811e-01
3.87548447e-01 4.60091978e-01 -6.64149404e-01 -1.14748323e+00
2.47127697e-01 1.90664485e-01 -4.19668764e-01 8.35313052e-02
-6.99284822e-02 2.67887026e-01 -6.51621461e-01 9.00135696e-01
1.07771015e+00 -4.12164293e-02 -1.99844643e-01 -6.77092016e-01
9.76501852e-02 -1.11631252e-01 -1.52433622e+00 1.66605759e+00
-2.65891552e-01 6.60848394e-02 1.57672763e-01 -1.31949353e+00
1.15495932e+00 -1.26477424e-03 7.85360932e-01 -8.30940604e-01
2.02603087e-01 3.92036617e-01 -2.80892789e-01 -5.02301395e-01
6.14528239e-01 -1.65794417e-01 -1.77433982e-01 1.23064622e-01
-1.62205443e-01 -4.03948247e-01 2.77806986e-02 -1.30075559e-01
5.00359178e-01 2.52869606e-01 1.21655114e-01 -4.22864169e-01
8.62441838e-01 2.68528223e-01 1.05742776e+00 2.86626697e-01
-3.25579971e-01 6.43394411e-01 2.64884770e-01 -1.05190333e-02
-7.48956859e-01 -9.77209091e-01 -3.00423384e-01 4.54943866e-01
6.10930741e-01 8.07889923e-02 -1.19546437e+00 -9.73638952e-01
-2.02067420e-01 4.41867322e-01 -1.25587896e-01 -1.96931511e-01
-6.62415326e-01 -1.34699893e+00 5.82345068e-01 5.72249115e-01
1.25210857e+00 -8.09881628e-01 -4.13308859e-01 9.33708549e-02
-3.82134050e-01 -1.00678217e+00 -1.81016505e-01 3.49355072e-01
-1.01068783e+00 -1.29963219e+00 -4.49396223e-01 -1.12268138e+00
7.82172024e-01 -3.03945486e-02 9.01122451e-01 1.81399778e-01
-4.45598692e-01 -7.41985440e-02 -2.08741143e-01 -1.46750733e-01
-2.21785665e-01 1.20075539e-01 -3.44431579e-01 -6.07665628e-02
1.78855471e-02 -3.11611265e-01 -6.86672032e-01 3.79566193e-01
-7.65521228e-01 1.45357978e-02 6.08943820e-01 9.63424325e-01
9.88417685e-01 6.99306130e-01 9.71873581e-01 -1.07027459e+00
4.31802213e-01 -6.90580010e-02 -4.06614631e-01 2.11128905e-01
-1.17934966e+00 -1.15053862e-01 3.32395017e-01 -6.01643175e-02
-1.49538720e+00 2.60668784e-01 -3.81478548e-01 1.62513182e-02
-2.00353161e-01 2.81255662e-01 -4.92698997e-01 1.56379104e-01
1.18633687e-01 -1.32285982e-01 -4.02968042e-02 -3.47742438e-01
3.22602898e-01 9.66100514e-01 7.11022973e-01 -7.35173285e-01
6.73928797e-01 3.43878299e-01 -2.84543615e-02 -4.84767884e-01
-9.11272109e-01 -4.90091413e-01 -7.91347146e-01 -2.56687254e-01
1.23547161e+00 -1.00905728e+00 -5.93331814e-01 1.08796132e+00
-8.33736420e-01 -1.25913739e-01 -4.14475203e-01 4.06634033e-01
-5.90391338e-01 7.78954864e-01 -7.16668308e-01 -9.35809672e-01
-4.82543945e-01 -1.21510720e+00 1.07425845e+00 5.57377279e-01
3.93917054e-01 -7.62861192e-01 -3.26869518e-01 9.30106759e-01
-1.35149090e-02 4.90750313e-01 1.14397025e+00 -5.39285302e-01
-4.63902920e-01 -7.55986795e-02 -5.38717210e-01 7.32649505e-01
1.56388909e-01 1.92602515e-01 -1.16859400e+00 5.81843555e-02
1.29636824e-01 -4.63746578e-01 9.68318939e-01 2.97478497e-01
1.46132445e+00 1.93633914e-01 -3.06766719e-01 2.99753219e-01
1.48033559e+00 2.77613878e-01 5.51240981e-01 1.32070705e-01
9.76678252e-01 6.59704268e-01 1.35372829e+00 2.17257500e-01
3.54447544e-01 4.47933316e-01 5.30098915e-01 -5.48209012e-01
-1.85577407e-01 -8.99535343e-02 -2.28599831e-01 1.13707316e+00
2.03388602e-01 -1.35351568e-01 -6.17097199e-01 4.36682701e-01
-1.68908226e+00 -6.00628197e-01 -5.90071380e-01 1.94492185e+00
9.05737281e-01 6.68489814e-01 -6.61376789e-02 7.96173275e-01
1.06471241e+00 2.20019341e-01 -5.24803400e-01 -1.90813929e-01
-1.03247240e-01 1.44647315e-01 5.14905930e-01 3.78848910e-01
-1.46270502e+00 7.52256811e-01 5.68372202e+00 1.42866743e+00
-8.99891257e-01 8.84646848e-02 1.12537539e+00 5.60056686e-01
-2.43609384e-01 -1.83080323e-02 -6.54452741e-01 6.83157265e-01
4.10757601e-01 3.72091532e-01 -1.98551223e-01 7.30447710e-01
1.41794271e-05 -4.12903279e-01 -5.41060567e-01 9.16505039e-01
-3.89710553e-02 -9.12083745e-01 -3.29515815e-01 -2.18645513e-01
7.24409521e-01 -5.69494188e-01 2.59017460e-02 1.80640638e-01
-4.29799268e-03 -8.13068032e-01 4.76443470e-01 3.33508521e-01
7.26612568e-01 -9.79921937e-01 8.67796719e-01 3.09905499e-01
-1.28064191e+00 -2.55949795e-01 -1.43614858e-01 3.50221172e-02
2.38418296e-01 1.26547229e+00 -6.20697498e-01 9.97577012e-01
6.24788761e-01 6.56957507e-01 -3.81568819e-01 8.78708065e-01
-4.21682447e-02 6.66167736e-01 -2.14468017e-01 4.11013991e-01
1.33300781e-01 -2.19446138e-01 1.82942227e-01 9.70516682e-01
2.79196855e-02 3.91806252e-02 3.94917399e-01 6.94939017e-01
-1.19256619e-02 3.35274816e-01 -1.19343974e-01 2.90570438e-01
3.35589916e-01 1.31642795e+00 -1.21353650e+00 -3.71080071e-01
-2.10104156e-02 7.58428812e-01 -8.78837034e-02 -1.74223687e-02
-8.67486537e-01 -4.99571919e-01 6.80437535e-02 -2.79834457e-02
9.83425826e-02 1.19123504e-01 -8.96866500e-01 -8.57561588e-01
2.78877050e-01 -5.89506626e-01 4.27919567e-01 -5.95634699e-01
-1.27287889e+00 3.46370250e-01 -1.48205217e-02 -1.19012678e+00
1.32201791e-01 -5.12351215e-01 -5.75709939e-01 5.74436963e-01
-1.70698774e+00 -1.17266202e+00 -4.32860196e-01 3.24620456e-01
8.17653656e-01 1.50753990e-01 3.79823267e-01 6.63798809e-01
-8.05950940e-01 4.82548743e-01 -7.50624835e-02 7.64913261e-02
6.95224822e-01 -1.22548199e+00 -1.11274056e-01 8.67681086e-01
-2.77996600e-01 -6.17896626e-03 4.18026716e-01 -7.72603929e-01
-4.90638733e-01 -1.07731938e+00 2.61376262e-01 1.22856803e-01
1.04606554e-01 5.77688701e-02 -1.11191428e+00 6.01794012e-02
-3.76370177e-02 -9.69493389e-02 3.24216187e-01 -3.29626471e-01
-1.21067926e-01 -3.78647774e-01 -1.73813558e+00 4.85652164e-02
8.74985218e-01 -3.22386026e-01 -8.04616272e-01 2.70326912e-01
8.58584583e-01 -3.91481102e-01 -9.92956758e-01 1.11054265e+00
3.56728345e-01 -9.18379724e-01 1.00368249e+00 1.29534379e-01
5.08214176e-01 -5.43932140e-01 -8.36183056e-02 -9.00344610e-01
-2.32534874e-02 1.22564629e-01 2.63188660e-01 1.59365332e+00
3.80004793e-01 -5.09272933e-01 1.03226137e+00 4.26304430e-01
-3.84089977e-01 -8.72386873e-01 -8.93846691e-01 -4.19111669e-01
7.83110559e-02 -3.36401075e-01 4.48941022e-01 9.53192830e-01
-3.11405569e-01 1.29626423e-01 -2.00580478e-01 2.14195535e-01
7.89691985e-01 2.04329863e-01 2.29451552e-01 -1.50841486e+00
-1.87147394e-01 -1.86293483e-01 -3.43476743e-01 -7.93194830e-01
2.97849208e-01 -7.78847575e-01 4.42238897e-01 -1.45110095e+00
2.32859775e-01 -8.64177704e-01 -5.54479241e-01 3.36295664e-01
-4.45848644e-01 4.15432513e-01 -2.62563229e-01 7.85806403e-02
-4.19110179e-01 4.87400383e-01 1.46680117e+00 -4.64357197e-01
-4.65731062e-02 1.40392825e-01 -6.77505732e-01 9.00854111e-01
8.45130563e-01 -5.58029413e-01 -6.24212801e-01 -9.19528753e-02
-1.15899280e-01 5.86554855e-02 2.79773682e-01 -1.10740185e+00
4.43715900e-02 -2.42939722e-02 3.48477930e-01 -9.91288483e-01
1.50944903e-01 -1.08366656e+00 -3.13670695e-01 5.01180232e-01
-2.17322886e-01 -3.01239789e-01 -3.72910053e-02 5.34644842e-01
-4.90666270e-01 -3.90486002e-01 9.98295486e-01 -8.64995457e-03
-7.70860672e-01 4.99480702e-02 3.17657106e-02 8.87686908e-02
1.14581001e+00 -4.52546626e-01 -3.05136025e-01 2.14997619e-01
-6.56251311e-01 3.64351094e-01 3.78656805e-01 -7.40424171e-02
5.83885431e-01 -1.03741837e+00 -3.99921507e-01 2.58627862e-01
-5.65907732e-03 4.93185282e-01 6.43268943e-01 6.94007397e-01
-4.24333543e-01 -1.59244895e-01 -3.00508678e-01 -7.70513713e-01
-1.14068258e+00 3.87881160e-01 2.47823715e-01 -4.71331388e-01
-3.93060118e-01 6.17371738e-01 -2.08955593e-02 -6.41727388e-01
3.11576054e-02 -3.82963181e-01 -3.42030078e-01 1.31995112e-01
-4.13565934e-02 7.04063237e-01 2.37365350e-01 -5.98859131e-01
-3.39526147e-01 9.16813254e-01 4.95758988e-02 1.06760480e-01
1.04932165e+00 -2.83932149e-01 -8.42911527e-02 3.84284407e-01
1.09076893e+00 -2.28361920e-01 -1.32877386e+00 5.78372441e-02
5.80009259e-02 -2.53621936e-01 2.97275573e-01 -8.62746358e-01
-1.45617878e+00 8.96588027e-01 9.72040176e-01 4.17286828e-02
1.40625513e+00 -4.38890964e-01 1.04511297e+00 1.82200596e-01
2.59286195e-01 -1.55730569e+00 1.28732458e-01 -4.90296818e-02
7.45901465e-02 -1.46272242e+00 3.17280322e-01 -9.38185811e-01
-5.50146520e-01 1.05715406e+00 8.46413732e-01 -8.00487101e-02
6.74938500e-01 3.70182663e-01 1.13038654e-02 -2.73858160e-01
-1.02542721e-01 -8.26110616e-02 1.71801955e-01 4.07988995e-01
1.61446720e-01 8.40481743e-02 -5.48992753e-01 6.37540221e-01
2.88357288e-01 -3.87828723e-02 3.89091045e-01 8.16581070e-01
-6.27086997e-01 -9.18667495e-01 -3.85656238e-01 6.72434211e-01
-5.85956097e-01 2.08391443e-01 1.20212510e-01 7.30788350e-01
4.56685513e-01 1.09968042e+00 1.56100571e-01 -5.56230187e-01
2.12603420e-01 5.30686341e-02 2.40800291e-01 -3.44209641e-01
-4.46780324e-01 2.64600635e-01 2.07191184e-02 -3.92458826e-01
-7.91091442e-01 -4.54220682e-01 -1.75607765e+00 3.50299656e-01
-8.21521580e-01 1.79440349e-01 6.81738138e-01 1.09963000e+00
-7.46857002e-02 8.70056093e-01 9.45224404e-01 -5.62576652e-01
-4.91239369e-01 -7.42583394e-01 -7.15586126e-01 3.56890380e-01
6.14959151e-02 -7.03159809e-01 -5.23518741e-01 1.04396485e-01] | [14.49334716796875, -2.0312771797180176] |
856b036a-45d0-45ff-9e8c-96e8aa437da1 | h-vfi-hierarchical-frame-interpolation-for | 2211.11309 | null | https://arxiv.org/abs/2211.11309v1 | https://arxiv.org/pdf/2211.11309v1.pdf | H-VFI: Hierarchical Frame Interpolation for Videos with Large Motions | Capitalizing on the rapid development of neural networks, recent video frame interpolation (VFI) methods have achieved notable improvements. However, they still fall short for real-world videos containing large motions. Complex deformation and/or occlusion caused by large motions make it an extremely difficult problem in video frame interpolation. In this paper, we propose a simple yet effective solution, H-VFI, to deal with large motions in video frame interpolation. H-VFI contributes a hierarchical video interpolation transformer (HVIT) to learn a deformable kernel in a coarse-to-fine strategy in multiple scales. The learnt deformable kernel is then utilized in convolving the input frames for predicting the interpolated frame. Starting from the smallest scale, H-VFI updates the deformable kernel by a residual in succession based on former predicted kernels, intermediate interpolated results and hierarchical features from transformer. Bias and masks to refine the final outputs are then predicted by a transformer block based on interpolated results. The advantage of such a progressive approximation is that the large motion frame interpolation problem can be decomposed into several relatively simpler sub-tasks, which enables a very accurate prediction in the final results. Another noteworthy contribution of our paper consists of a large-scale high-quality dataset, YouTube200K, which contains videos depicting a great variety of scenarios captured at high resolution and high frame rate. Extensive experiments on multiple frame interpolation benchmarks validate that H-VFI outperforms existing state-of-the-art methods especially for videos with large motions. | ['Yu-Wing Tai', 'Chi-Keung Tang', 'Xin Tao', 'Yanan sun', 'Guangyang Wu', 'Changlin Li'] | 2022-11-21 | null | null | null | null | ['video-frame-interpolation'] | ['computer-vision'] | [ 1.30599057e-02 -4.44184244e-01 -2.30921909e-01 -1.65451020e-01
-6.58860624e-01 -9.51379463e-02 3.10053378e-01 -4.52325553e-01
-2.46612042e-01 8.87457848e-01 1.88317761e-01 1.16007030e-01
6.33922368e-02 -7.01618254e-01 -9.95872200e-01 -7.16270983e-01
-4.00553793e-02 -4.54717614e-02 6.23948812e-01 -1.50025517e-01
1.20295718e-01 3.40864599e-01 -1.42660367e+00 5.74932635e-01
9.35333967e-01 1.23626447e+00 3.75231922e-01 5.41384339e-01
5.95882423e-02 9.58437979e-01 -2.05049455e-01 -1.31991595e-01
4.18593347e-01 -4.00837988e-01 -6.69530571e-01 2.35904045e-02
7.17447221e-01 -8.22007477e-01 -3.88940424e-01 7.79080868e-01
3.50909024e-01 3.91127259e-01 4.08266515e-01 -1.02733433e+00
-6.21430933e-01 3.11912447e-01 -6.56961381e-01 2.65860260e-01
5.82842827e-01 1.82289049e-01 5.36088288e-01 -1.29675996e+00
7.23268449e-01 1.44739592e+00 1.01027572e+00 5.79997778e-01
-1.31558156e+00 -6.66880488e-01 1.29603326e-01 3.64125580e-01
-1.42970669e+00 -3.26346338e-01 7.80088127e-01 -6.04739428e-01
7.09607959e-01 3.48542899e-01 7.88310707e-01 7.59203792e-01
2.10125953e-01 6.73547208e-01 7.92104840e-01 -1.19753703e-01
7.85943344e-02 -4.29036170e-01 -2.50812829e-01 6.07645154e-01
-2.33791307e-01 9.90225449e-02 -3.58452171e-01 6.44689128e-02
1.45288837e+00 1.37991533e-01 -7.15719640e-01 1.02420002e-02
-1.57519245e+00 5.80826223e-01 5.79552233e-01 2.19844311e-01
-4.11578655e-01 2.25746930e-01 4.46741909e-01 2.07962707e-01
5.85756421e-01 -1.55210137e-01 -3.47764075e-01 1.05693154e-01
-1.26819587e+00 5.37458837e-01 4.37709033e-01 8.42272639e-01
1.01591814e+00 3.53700995e-01 -5.51821768e-01 7.58860886e-01
2.46292189e-01 2.68423200e-01 3.97523761e-01 -1.23486972e+00
6.37826085e-01 2.08764046e-01 5.31396985e-01 -1.15481806e+00
-5.30815013e-02 4.25657667e-02 -1.39034665e+00 3.35295856e-01
5.40423274e-01 4.10586186e-02 -7.00432539e-01 1.36906219e+00
5.02661645e-01 9.15910482e-01 -3.39196414e-01 1.17018366e+00
8.51380110e-01 1.08341026e+00 3.11809212e-01 -3.71720165e-01
1.10312700e+00 -1.12461174e+00 -7.66175330e-01 2.18767732e-01
2.51229197e-01 -8.46624851e-01 8.95299315e-01 3.62484783e-01
-1.33967531e+00 -1.33121812e+00 -7.86867917e-01 -3.88169825e-01
2.18774766e-01 2.74780482e-01 3.31097603e-01 -1.37685786e-03
-1.21028709e+00 8.25836062e-01 -6.85046792e-01 1.26881108e-01
4.34822202e-01 4.51889127e-01 -2.67317891e-01 9.89523008e-02
-1.30100429e+00 4.82080638e-01 5.44820309e-01 4.33019221e-01
-6.95741534e-01 -1.00185931e+00 -9.64325666e-01 1.18596479e-02
1.37065589e-01 -7.64330864e-01 9.39565778e-01 -1.36884022e+00
-1.56444430e+00 3.62915993e-01 -4.64451939e-01 -4.23117578e-01
8.87818396e-01 -3.03773910e-01 -3.54072452e-01 7.41625950e-02
-1.88724492e-02 8.61109793e-01 1.36289763e+00 -1.11487532e+00
-8.99643302e-01 8.91862214e-02 -6.80074990e-02 1.61814377e-01
-2.13743210e-01 -2.93296445e-02 -5.77242315e-01 -1.16758120e+00
-2.40477771e-01 -7.77686954e-01 -2.16095164e-01 2.59168267e-01
5.73796183e-02 -3.62797886e-01 1.15566206e+00 -9.92577553e-01
1.64573455e+00 -2.13888597e+00 3.95694733e-01 -2.66278535e-01
1.80252478e-01 5.88755429e-01 -5.34691475e-02 -1.77692287e-02
-1.82733864e-01 -1.22034632e-01 -1.39134601e-01 -3.00136268e-01
-4.37230855e-01 1.13081999e-01 -4.20058340e-01 1.82480067e-01
3.21387678e-01 8.49312305e-01 -9.11758959e-01 -5.88118970e-01
5.18388569e-01 8.37408006e-01 -7.10402966e-01 4.04468268e-01
-1.01898983e-01 8.35063517e-01 -3.64526927e-01 4.10682321e-01
9.07914162e-01 -3.01796466e-01 -3.35780919e-01 -7.75311649e-01
-4.06230330e-01 -3.08257729e-01 -1.33702028e+00 1.84013510e+00
-2.72231102e-01 6.20531917e-01 4.51534837e-02 -9.84501302e-01
7.82252371e-01 4.39712733e-01 5.67998648e-01 -2.05151439e-01
-6.39266223e-02 2.19811112e-01 -5.06503880e-01 -6.14412427e-01
4.72996801e-01 1.75450429e-01 4.54659939e-01 -3.01727593e-01
-7.71507695e-02 1.97945088e-01 2.55123138e-01 -8.80480707e-02
5.01082718e-01 7.25672662e-01 2.20755711e-01 -3.32059503e-01
1.04228234e+00 -3.15253943e-01 8.71621370e-01 4.16600198e-01
-9.28292722e-02 9.91673708e-01 3.16497274e-02 -1.10175860e+00
-1.25720954e+00 -9.22188282e-01 -1.30532607e-01 8.03145826e-01
3.63996774e-01 -2.60772765e-01 -8.68248820e-01 -3.79310876e-01
-5.42855673e-02 3.51518244e-02 -6.45214200e-01 2.28968933e-01
-1.23941028e+00 -4.43510056e-01 9.15983170e-02 7.17480540e-01
8.60203683e-01 -1.17035651e+00 -3.20432454e-01 5.23163021e-01
-4.40026045e-01 -1.21493196e+00 -9.09346700e-01 -5.12067437e-01
-9.44348574e-01 -7.11990714e-01 -1.21600807e+00 -1.06600916e+00
6.31108344e-01 3.71649057e-01 9.72420335e-01 2.85180628e-01
-1.33615583e-01 -2.04686284e-01 -1.31358027e-01 2.39875540e-01
-4.16429490e-01 -3.17991525e-01 -1.13927526e-02 4.36031044e-01
-1.21833399e-01 -4.43544418e-01 -9.78698850e-01 5.57931960e-01
-9.82585192e-01 5.18664360e-01 1.08062521e-01 9.73409593e-01
8.58409643e-01 -1.79924011e-01 4.93433863e-01 -5.39477170e-01
1.90705553e-01 -4.00792062e-01 -7.31410921e-01 1.72385812e-01
-1.30856976e-01 -1.40651837e-01 1.00767422e+00 -7.01019228e-01
-1.31656909e+00 1.15834296e-01 -4.09492999e-01 -8.50171328e-01
-2.60912664e-02 8.67829174e-02 1.99856702e-02 -3.40531498e-01
4.33633685e-01 2.01063707e-01 -1.99903801e-01 -3.87153298e-01
1.08904175e-01 3.58823240e-01 6.57015562e-01 -5.27424872e-01
9.50781465e-01 3.78816992e-01 5.06647453e-02 -7.39800870e-01
-5.99496722e-01 -1.85368165e-01 -7.36384690e-01 -5.14806151e-01
1.20864046e+00 -1.29120719e+00 -7.99063444e-01 7.36449778e-01
-1.26393378e+00 -5.83965540e-01 -2.70820230e-01 4.67967391e-01
-5.32009125e-01 5.57488739e-01 -1.04228950e+00 -3.22652817e-01
-4.06569481e-01 -1.44885719e+00 1.05868244e+00 3.73639941e-01
7.85598904e-03 -9.64288414e-01 -1.27108008e-01 2.95025706e-01
5.22145629e-01 4.00598973e-01 5.70175350e-01 3.15699697e-01
-9.81492162e-01 1.70788899e-01 -3.20388138e-01 4.88666266e-01
3.15316468e-01 2.80000508e-01 -7.10698783e-01 -4.82113451e-01
-1.45457759e-01 -1.00351416e-01 8.45610499e-01 6.34157658e-01
1.57454872e+00 -3.65424126e-01 -1.27103791e-01 1.08732069e+00
1.62162626e+00 1.89894244e-01 9.30444121e-01 2.31164351e-01
1.21400285e+00 2.96430178e-02 7.58825362e-01 5.26921213e-01
3.15718293e-01 9.53334212e-01 1.98181182e-01 -1.21495359e-01
-3.36900145e-01 -1.79324850e-01 3.86883944e-01 7.85220087e-01
-9.28439796e-01 1.26854032e-02 -4.16180253e-01 3.53615344e-01
-1.97375369e+00 -1.20364475e+00 -2.44429290e-01 2.31416655e+00
9.89385784e-01 -1.23479359e-01 1.98264077e-01 1.43553168e-01
8.75810742e-01 2.34981552e-01 -3.39576334e-01 -1.51588628e-02
-9.16735157e-02 9.90256518e-02 8.49000365e-02 6.10089183e-01
-1.25779903e+00 8.17297637e-01 5.46697521e+00 1.21678507e+00
-1.23130679e+00 -1.26627749e-02 9.72714186e-01 2.48708606e-01
-1.80917814e-01 -4.97301184e-02 -9.04658079e-01 8.80699277e-01
6.23745203e-01 1.37031332e-01 5.15349746e-01 6.93299770e-01
6.12687588e-01 3.82610895e-02 -9.71265972e-01 1.08402503e+00
-2.49194860e-01 -1.94721782e+00 3.77244949e-01 -3.50635201e-01
9.31203544e-01 -2.25899324e-01 -4.34085019e-02 1.71101898e-01
-2.02973858e-01 -8.12273979e-01 8.13388169e-01 6.04318202e-01
1.12453532e+00 -6.65620387e-01 5.17582119e-01 2.04024419e-01
-1.80846012e+00 -4.48384956e-02 -5.34969568e-01 -1.10477813e-01
3.55466187e-01 3.82394820e-01 -1.06134623e-01 6.32321537e-01
9.47273731e-01 1.13032043e+00 -2.13396773e-01 9.63142872e-01
2.41399147e-02 5.44562876e-01 -4.14127260e-02 5.41180849e-01
1.19521968e-01 -4.76109803e-01 3.30353320e-01 1.30533588e+00
5.85054755e-01 3.75590056e-01 3.88364762e-01 6.28721535e-01
-3.53924967e-02 1.04120001e-01 -2.54708886e-01 6.43588960e-01
1.69778690e-01 1.23883808e+00 -4.29668695e-01 -7.14021802e-01
-7.09507287e-01 1.19696081e+00 2.53151327e-01 6.83446705e-01
-1.15326488e+00 -4.58533913e-02 6.17588103e-01 3.54748249e-01
4.84254420e-01 -2.01940089e-01 8.64832550e-02 -1.49418390e+00
1.21418752e-01 -7.25831628e-01 2.20359176e-01 -8.54635894e-01
-9.60720360e-01 7.48615026e-01 -7.71822333e-02 -1.83199728e+00
-2.75272816e-01 -3.11570823e-01 -5.05094647e-01 1.04719162e+00
-1.62564421e+00 -1.19515026e+00 -8.27707648e-01 9.79974687e-01
1.12916028e+00 9.50184017e-02 5.33891201e-01 6.44257665e-01
-5.08219719e-01 4.25958663e-01 1.16423599e-01 3.00659508e-01
7.99369216e-01 -8.19682300e-01 4.18532491e-01 7.64513731e-01
-2.77073681e-01 2.91945368e-01 4.39991742e-01 -7.42006123e-01
-1.27082705e+00 -1.47126460e+00 6.91964865e-01 -4.81688343e-02
4.26543891e-01 4.31784317e-02 -1.25763822e+00 6.87036037e-01
2.29219258e-01 8.30260456e-01 -1.34916911e-02 -8.91745865e-01
-2.44705621e-02 -4.11492556e-01 -1.04449427e+00 5.01968861e-01
8.98496926e-01 -2.10595906e-01 -2.15210065e-01 1.06180958e-01
7.04109848e-01 -7.70449936e-01 -1.28885150e+00 5.88827670e-01
5.48290312e-01 -1.07952940e+00 1.40499961e+00 -3.20086539e-01
8.67644966e-01 -6.60698473e-01 1.42726392e-01 -9.81482327e-01
-7.37667203e-01 -9.03921545e-01 -2.77705461e-01 1.21209610e+00
-2.30151147e-01 -2.90487349e-01 5.68441629e-01 6.44002318e-01
-1.00662529e-01 -8.99451196e-01 -9.48907673e-01 -5.26985824e-01
-2.49526009e-01 -1.06900772e-02 4.60574299e-01 9.85549092e-01
-5.03958523e-01 -1.33709297e-01 -9.22580898e-01 -5.65023087e-02
8.00361514e-01 9.85742584e-02 7.24724650e-01 -9.75167096e-01
-2.14133620e-01 -2.32668240e-02 -4.40241516e-01 -1.49977124e+00
1.29528284e-01 -4.48755771e-01 -5.96142560e-02 -1.23361230e+00
-5.68951517e-02 -4.16048139e-01 -1.34401366e-01 1.96761236e-01
-6.66824102e-01 6.60308421e-01 2.79395521e-01 4.97553498e-01
-3.30251127e-01 5.26930094e-01 1.63072753e+00 -1.22975662e-01
-3.43214780e-01 1.07186204e-02 9.41920951e-02 9.82425094e-01
4.43895578e-01 2.20805798e-02 -2.81202227e-01 -6.83839023e-01
-2.42049873e-01 5.58597982e-01 5.28568208e-01 -1.27118826e+00
7.99310207e-02 -3.78062665e-01 8.69591057e-01 -5.37325621e-01
2.99353421e-01 -8.13750148e-01 6.01765394e-01 4.01776642e-01
-2.74441123e-01 1.49098203e-01 7.92239085e-02 4.53408062e-01
-4.96968865e-01 1.33307919e-01 1.00537944e+00 -6.45636246e-02
-1.17445064e+00 8.06426108e-01 -8.95357355e-02 -8.65955744e-03
1.05918622e+00 -5.49706459e-01 1.93106055e-01 -1.42523780e-01
-7.26671636e-01 -8.49804729e-02 4.95000035e-01 3.80560040e-01
7.82731414e-01 -1.71200776e+00 -8.76654446e-01 4.05223757e-01
-3.74733061e-01 3.26123625e-01 5.64719200e-01 8.44360530e-01
-8.54932845e-01 -6.16339454e-03 -4.07354087e-01 -7.27493346e-01
-1.15220034e+00 6.71812356e-01 1.49994344e-01 -1.47717655e-01
-8.74585986e-01 7.72654653e-01 4.94463623e-01 3.78529847e-01
1.13735825e-01 -7.00739145e-01 -2.84777820e-01 -2.69315969e-02
8.48253369e-01 4.30538148e-01 -2.42733419e-01 -9.73398089e-01
6.41609356e-03 9.86544669e-01 1.08479686e-01 4.14392054e-01
1.20986331e+00 -2.71555930e-01 -1.25487387e-01 6.68062568e-02
1.18152571e+00 -1.06584840e-01 -1.94695544e+00 -1.47170991e-01
-5.25875390e-01 -8.99504185e-01 -2.57289290e-01 -5.54035939e-02
-1.36356080e+00 6.67944849e-01 5.04422605e-01 -4.50028926e-02
1.43813598e+00 -5.62520325e-01 1.23176575e+00 -8.14311951e-02
4.91713941e-01 -7.64104545e-01 -1.56009600e-01 3.23329300e-01
9.37582850e-01 -1.27323973e+00 5.61935678e-02 -6.23782575e-01
-3.90188068e-01 1.43232417e+00 7.44177818e-01 -4.29076761e-01
4.20033783e-01 2.34811231e-01 -1.46437153e-01 6.81060553e-01
-3.79520953e-01 3.69700402e-01 4.82245862e-01 3.57493669e-01
5.54295361e-01 -3.09685290e-01 -4.10711795e-01 2.02673420e-01
3.85585397e-01 6.93388104e-01 2.64989763e-01 4.30329472e-01
-2.89208710e-01 -9.41451788e-01 -5.64022481e-01 2.52933860e-01
-4.37017411e-01 -7.74256513e-02 6.07069671e-01 6.21514320e-01
3.37572783e-01 5.03211915e-01 4.35018316e-02 -2.74866015e-01
9.81969833e-02 -3.39141697e-01 3.75678152e-01 1.96583476e-02
-6.39056146e-01 3.26152503e-01 -3.12492281e-01 -9.11769986e-01
-7.62169361e-01 -4.51317698e-01 -1.30789280e+00 -2.37983525e-01
-3.54797430e-02 1.17754489e-02 -7.89486617e-02 7.64173925e-01
3.41805518e-01 5.50732493e-01 5.30357599e-01 -1.63880193e+00
-3.07759017e-01 -6.93432391e-01 -3.01629901e-01 8.83697212e-01
7.02762902e-01 -5.55549264e-01 -2.33421251e-01 7.54252613e-01] | [10.749916076660156, -1.4522819519042969] |
315377bc-3ed6-40fe-a226-9499796ad7de | backpack-language-models | 2305.16765 | null | https://arxiv.org/abs/2305.16765v1 | https://arxiv.org/pdf/2305.16765v1.pdf | Backpack Language Models | We present Backpacks: a new neural architecture that marries strong modeling performance with an interface for interpretability and control. Backpacks learn multiple non-contextual sense vectors for each word in a vocabulary, and represent a word in a sequence as a context-dependent, non-negative linear combination of sense vectors in this sequence. We find that, after training, sense vectors specialize, each encoding a different aspect of a word. We can interpret a sense vector by inspecting its (non-contextual, linear) projection onto the output space, and intervene on these interpretable hooks to change the model's behavior in predictable ways. We train a 170M-parameter Backpack language model on OpenWebText, matching the loss of a GPT-2 small (124Mparameter) Transformer. On lexical similarity evaluations, we find that Backpack sense vectors outperform even a 6B-parameter Transformer LM's word embeddings. Finally, we present simple algorithms that intervene on sense vectors to perform controllable text generation and debiasing. For example, we can edit the sense vocabulary to tend more towards a topic, or localize a source of gender bias to a sense vector and globally suppress that sense. | ['Percy Liang', 'Christopher D. Manning', 'John Thickstun', 'John Hewitt'] | 2023-05-26 | null | null | null | null | ['word-embeddings'] | ['methodology'] | [ 4.28261608e-01 4.23608720e-01 -2.99116224e-01 -5.42709649e-01
-5.62578738e-01 -1.10566986e+00 6.34268820e-01 1.95768729e-01
-6.38349235e-01 3.85165811e-01 6.38785839e-01 -8.11840236e-01
1.84445128e-01 -8.12283576e-01 -7.80095279e-01 -4.16961402e-01
4.48523074e-01 7.65148759e-01 -1.86281487e-01 -9.00736451e-01
3.64265561e-01 -4.80136089e-02 -1.30408537e+00 5.19877136e-01
6.71692312e-01 3.91588330e-01 2.53660321e-01 6.40319109e-01
-4.78153378e-01 2.15319052e-01 -9.68270063e-01 -6.34203076e-01
9.29575711e-02 -1.55064330e-01 -9.43959773e-01 -3.61608863e-01
6.94383204e-01 1.72571212e-01 3.10747400e-02 9.09869134e-01
3.94030184e-01 1.79600745e-01 7.93678761e-01 -1.00764382e+00
-1.52749705e+00 1.32197392e+00 -4.08538640e-01 2.98377782e-01
4.47724134e-01 3.28226089e-01 1.39232218e+00 -8.05159807e-01
6.83657885e-01 1.66200495e+00 7.28487551e-01 5.49936950e-01
-1.73993206e+00 -7.51055956e-01 3.05537432e-01 -7.42966756e-02
-1.18753123e+00 -2.74327546e-01 4.46995169e-01 -2.29324773e-01
1.32032275e+00 5.85824788e-01 1.13285959e+00 1.79319549e+00
5.44088900e-01 4.48552072e-01 5.87714434e-01 -4.86221820e-01
3.22508663e-01 9.61119011e-02 6.67821884e-01 5.73765576e-01
1.87551245e-01 -2.38608513e-02 -7.55841970e-01 -4.72765118e-01
5.17152905e-01 -1.73294276e-01 -9.02111977e-02 -3.04452479e-01
-1.05772221e+00 1.05536532e+00 3.77646267e-01 3.67604911e-01
-1.52412325e-01 3.86987329e-01 3.75315070e-01 7.20566511e-01
5.37151873e-01 1.25191486e+00 -7.01095581e-01 -9.90890041e-02
-6.02684259e-01 4.49549615e-01 8.30799580e-01 8.09826732e-01
7.56194055e-01 8.37490335e-02 -4.41984922e-01 8.70115519e-01
3.25669080e-01 7.32180834e-01 1.07417130e+00 -7.26116776e-01
2.87534207e-01 3.67263049e-01 -8.44826773e-02 -9.82810140e-01
-3.56940031e-01 -3.63635123e-01 -6.67358339e-01 -1.06926180e-01
1.02311596e-01 -2.82718331e-01 -1.00590980e+00 2.01853871e+00
-5.06462194e-02 1.52587548e-01 -3.71439457e-02 6.60533190e-01
4.44670528e-01 7.01916933e-01 4.46791261e-01 1.18895702e-01
1.56590414e+00 -4.49688643e-01 -4.33516532e-01 -8.43382299e-01
7.85975814e-01 -7.90048182e-01 1.98402202e+00 2.94027776e-01
-9.52128649e-01 -3.84486198e-01 -1.31836224e+00 -3.02605659e-01
-6.66511536e-01 -3.35530996e-01 8.39777410e-01 6.30016744e-01
-1.20103502e+00 5.73437750e-01 -5.16265988e-01 -5.16035795e-01
-1.77103102e-01 3.34048390e-01 -3.44873875e-01 3.87096792e-01
-1.80055141e+00 9.51910913e-01 6.10689282e-01 -4.40824658e-01
-3.21489841e-01 -1.12809360e+00 -9.91341650e-01 7.10757896e-02
4.05447111e-02 -1.19316375e+00 1.23432875e+00 -1.17505944e+00
-1.17832303e+00 9.70674515e-01 -2.51064897e-01 -5.31027734e-01
-1.70287594e-01 -1.58525214e-01 -5.88845909e-01 -4.82926250e-01
1.88288048e-01 7.30418086e-01 1.11474955e+00 -1.05818927e+00
-3.34642053e-01 -3.74164850e-01 7.55295157e-02 3.67167264e-01
-7.39912212e-01 7.09470036e-03 -1.83626726e-01 -1.11927330e+00
-9.60117280e-02 -9.49499726e-01 -2.95694351e-01 -1.64697722e-01
-5.61082304e-01 -3.32321316e-01 5.15236139e-01 -2.42143169e-01
1.52660251e+00 -2.02379513e+00 3.16672921e-01 2.72394985e-01
4.83015835e-01 2.12597191e-01 -4.92686242e-01 2.91346341e-01
-5.93450010e-01 8.20909917e-01 -7.59513453e-02 -4.49794263e-01
2.77578920e-01 4.67404395e-01 -7.62534976e-01 1.08802401e-01
1.39418885e-01 9.12497103e-01 -1.02614021e+00 -4.11715135e-02
4.74904664e-02 4.41369414e-01 -1.11245096e+00 -3.13232243e-02
-3.52283835e-01 -4.82534587e-01 -2.58312345e-01 -1.81720644e-01
3.87470603e-01 -2.33621716e-01 3.80292773e-01 -3.18102866e-01
3.02828014e-01 5.95227599e-01 -1.07347667e+00 1.72425890e+00
-9.02577817e-01 5.85429311e-01 -2.19362721e-01 -5.72764039e-01
7.84750164e-01 3.42453308e-02 -8.95080417e-02 -5.12330890e-01
-1.59408242e-01 -4.08652797e-02 -7.09084421e-02 -3.75346392e-01
1.20333886e+00 -4.20306832e-01 -4.70934212e-01 8.86329710e-01
1.40759647e-01 -3.33475471e-01 -1.61810741e-01 3.62918347e-01
1.06286335e+00 -4.18934435e-01 3.25085849e-01 -5.28186500e-01
-4.42350619e-02 -3.32371205e-01 1.73448086e-01 1.07495129e+00
2.02943921e-01 5.79560220e-01 5.88356733e-01 -5.75693667e-01
-1.12420583e+00 -1.32984006e+00 4.54957783e-02 1.91412103e+00
-8.57560858e-02 -9.42915022e-01 -7.20296085e-01 -4.73030597e-01
2.50415623e-01 1.64873064e+00 -9.31823194e-01 -6.96599185e-01
-6.09968305e-01 -7.07021296e-01 8.77521038e-01 4.40404475e-01
-1.62220642e-01 -1.05330122e+00 -1.71734139e-01 2.94648260e-01
-4.91920784e-02 -2.36376539e-01 -1.01977456e+00 6.81981862e-01
-3.97575647e-01 -6.14932835e-01 -5.98135293e-02 -6.05877817e-01
4.06871617e-01 -1.53681645e-02 1.62702584e+00 1.14449188e-01
1.62714608e-02 9.55585688e-02 -1.93017110e-01 -3.86149406e-01
-8.00761163e-01 7.89583564e-01 2.90064722e-01 -2.90058911e-01
6.40765011e-01 -7.30673611e-01 -2.53728628e-01 6.29184721e-03
-1.19476545e+00 -9.67275128e-02 1.07553467e-01 1.15581632e+00
3.06131989e-01 -6.03294253e-01 1.08207949e-01 -1.13541186e+00
1.44238186e+00 -6.48890555e-01 -1.00770518e-01 3.01062524e-01
-7.62967229e-01 6.70754492e-01 5.30815721e-01 -7.75272012e-01
-7.65323281e-01 -2.68761367e-01 -2.37229258e-01 -2.92871177e-01
9.57333818e-02 3.83387119e-01 -1.85496137e-02 6.07578695e-01
1.25155497e+00 1.76775485e-01 -4.15415615e-02 -2.58387029e-01
1.11089635e+00 6.64334178e-01 4.59374160e-01 -7.59027421e-01
9.60685015e-01 -4.99669276e-02 -8.49377275e-01 -6.16869032e-01
-9.15971577e-01 -4.98351604e-02 -2.93301374e-01 5.23539722e-01
9.96604085e-01 -6.09362960e-01 -5.59913814e-01 1.29181007e-03
-1.39592052e+00 -3.75930369e-01 -4.14420903e-01 -1.52303696e-01
-3.18680555e-01 -1.83242500e-01 -5.21775544e-01 -3.05374861e-01
-6.56823218e-01 -1.02678168e+00 1.16351783e+00 1.12550072e-01
-1.35796750e+00 -1.25841582e+00 6.13728240e-02 -2.39680689e-02
6.21627331e-01 -6.01857364e-01 1.54667068e+00 -1.16519845e+00
7.51450583e-02 2.46644855e-01 2.23922312e-01 1.88654035e-01
5.52568659e-02 4.07358520e-02 -7.93675840e-01 -2.70185471e-01
-1.31711379e-01 -3.11731279e-01 7.18058109e-01 -2.97086071e-02
1.17187202e+00 -7.55725384e-01 -5.45597136e-01 8.74600828e-01
8.86339486e-01 -1.62982583e-01 3.18792224e-01 3.17333221e-01
6.93251848e-01 3.30000296e-02 8.80365148e-02 2.80278862e-01
2.39242807e-01 5.54779410e-01 1.72393188e-01 -2.67623723e-01
1.75351664e-01 -6.70909166e-01 3.92485321e-01 7.04106390e-01
5.77455282e-01 -5.70112407e-01 -8.17340672e-01 4.36627686e-01
-1.44672370e+00 -1.02903140e+00 1.71985567e-01 1.98966312e+00
1.35984480e+00 4.60136980e-01 -3.92315865e-01 -3.85187268e-01
5.01046717e-01 4.94164973e-01 -6.59117162e-01 -9.70498621e-01
-3.07244658e-01 5.28439701e-01 7.01336324e-01 9.74019885e-01
-7.27829695e-01 1.44443786e+00 7.33076763e+00 9.81025815e-01
-1.31877172e+00 6.47781268e-02 5.41525662e-01 -3.77405286e-01
-1.44846046e+00 -1.11205161e-01 -6.39801443e-01 4.70721513e-01
9.82123911e-01 -6.11084819e-01 7.35799372e-01 7.35378027e-01
1.03911109e-01 4.61778343e-01 -1.32700610e+00 9.28056061e-01
2.48794869e-01 -1.53135347e+00 6.20055020e-01 -1.29858822e-01
3.67765486e-01 1.64411515e-01 3.79822671e-01 6.57165349e-01
8.76518369e-01 -1.34692156e+00 8.00697505e-01 4.51054662e-01
7.53977001e-01 -7.91761577e-01 1.55081168e-01 8.72569680e-02
-6.92215979e-01 4.59422730e-02 -4.62399691e-01 -4.25716899e-02
2.73061674e-02 2.49604329e-01 -9.91736233e-01 -2.87185490e-01
4.19809222e-01 4.67736453e-01 -9.14049089e-01 1.73193414e-03
-3.18691373e-01 7.38610506e-01 -3.54286462e-01 -1.99606895e-01
1.98160801e-02 -1.19158588e-01 6.54606581e-01 1.21625662e+00
2.54029751e-01 6.24865405e-02 6.36981241e-03 1.25323737e+00
-3.98776650e-01 7.66740739e-02 -8.79666686e-01 -8.58198926e-02
9.72256839e-01 9.58049655e-01 -1.12480685e-01 -6.70861065e-01
1.20772712e-01 1.29352593e+00 4.14013475e-01 4.75907832e-01
-7.03995585e-01 -2.83018291e-01 1.65311956e+00 2.53188342e-01
1.34334743e-01 5.20549268e-02 -6.00331485e-01 -1.27474868e+00
-1.83669969e-01 -1.25865543e+00 3.03324431e-01 -1.24750161e+00
-1.36859167e+00 6.29890859e-01 -6.01420589e-02 -4.42744851e-01
-6.51629388e-01 -6.19236112e-01 -7.43917942e-01 1.08232677e+00
-5.06341517e-01 -9.88126218e-01 4.53853339e-01 3.44652921e-01
6.45435870e-01 -3.17880452e-01 1.31615937e+00 -1.90935507e-01
-3.98960412e-01 1.21372080e+00 -1.39116809e-01 -4.59220931e-02
9.43306983e-01 -1.54583836e+00 1.27083659e+00 3.70115966e-01
2.88967758e-01 1.60079610e+00 1.29109454e+00 -5.23572624e-01
-1.26703250e+00 -9.76111114e-01 1.11849463e+00 -9.66227233e-01
1.10910773e+00 -6.91917419e-01 -1.03407848e+00 1.15031421e+00
6.20053828e-01 -4.32879478e-01 9.73201036e-01 5.97576261e-01
-8.67217362e-01 4.50886004e-02 -8.76592338e-01 1.38901532e+00
1.31239522e+00 -8.30081940e-01 -1.26039386e+00 2.49037921e-01
1.45038366e+00 -4.31669742e-01 -6.21757388e-01 -2.91057229e-01
4.91646588e-01 -7.48241484e-01 1.15581238e+00 -1.30016065e+00
3.74843836e-01 -2.08449475e-02 -1.18167616e-01 -2.17058682e+00
-4.69170958e-01 -6.74472392e-01 3.86852413e-01 9.87808526e-01
7.93105066e-01 -6.91491067e-01 7.30210364e-01 6.80999398e-01
1.59730352e-02 -4.79711711e-01 -7.41571963e-01 -2.62712777e-01
6.50235057e-01 -8.01329315e-01 1.04969418e+00 1.20813501e+00
3.03836137e-01 1.07882690e+00 -1.41870201e-01 -4.55837995e-02
1.26685277e-01 -3.13455641e-01 4.89987463e-01 -9.92447078e-01
-5.42465687e-01 -6.65345252e-01 -1.07590847e-01 -1.22516727e+00
4.55285281e-01 -1.25596869e+00 -2.27639750e-01 -1.02914321e+00
-7.13992864e-02 -2.13415176e-01 -2.49444872e-01 7.92625904e-01
-4.63759631e-01 -3.29287723e-02 2.86238790e-01 -1.82757989e-01
-2.21229598e-01 4.86381084e-01 7.56077945e-01 -5.25668204e-01
-5.39785959e-02 -4.44591403e-01 -1.43147004e+00 6.81671441e-01
7.55003452e-01 -5.77443659e-01 -7.99756825e-01 -8.17324698e-01
7.26445735e-01 -8.01332593e-02 1.87675416e-01 -4.56692040e-01
3.87113988e-02 -3.99724036e-01 4.36099887e-01 -1.30974665e-01
3.17802608e-01 -3.54953468e-01 1.90065559e-02 2.40637675e-01
-9.39277053e-01 7.07098663e-01 1.90355554e-01 2.29906365e-01
2.49286100e-01 -1.48312241e-01 4.47390467e-01 -7.92215541e-02
-5.62596500e-01 -1.53975025e-01 -6.63752258e-01 4.02263016e-01
3.29559356e-01 -4.56138216e-02 -2.81121582e-01 -3.93796146e-01
-6.16505027e-01 6.54993653e-02 6.91872478e-01 1.06632793e+00
3.15148503e-01 -1.34236491e+00 -4.83151823e-01 6.56392634e-01
3.80516857e-01 -4.13121432e-01 -1.50141671e-01 -1.50455564e-01
-4.58987877e-02 7.06953257e-02 -4.70794039e-03 -3.74599695e-01
-1.07683587e+00 5.81605375e-01 5.07162452e-01 -1.33883655e-01
-2.21006930e-01 1.14906311e+00 2.97028601e-01 -9.51996207e-01
-1.23646103e-01 -6.38278902e-01 4.42962721e-02 1.93659142e-01
6.46492243e-01 -2.35048145e-01 8.16120431e-02 -2.16771066e-01
-2.85244167e-01 3.18337291e-01 -3.98506105e-01 -3.51049930e-01
9.92393970e-01 -4.29981090e-02 -1.90036371e-01 6.14370227e-01
1.51296651e+00 5.06186746e-02 -6.85314894e-01 -8.62203445e-03
-4.97554690e-02 -2.33152229e-03 -1.64977536e-01 -8.50712240e-01
-7.33700931e-01 4.90710527e-01 4.79528993e-01 2.66495496e-01
6.53408587e-01 9.10665467e-02 6.53313100e-01 7.15395629e-01
9.32677835e-02 -1.11315334e+00 1.42031014e-01 9.73764837e-01
1.13187957e+00 -8.58053744e-01 -3.72341365e-01 -3.72225568e-02
-7.68282712e-01 9.24948514e-01 5.88868439e-01 -1.91710312e-02
5.80800235e-01 3.96239311e-01 4.30067480e-01 -2.00978920e-01
-1.25692415e+00 3.56565744e-01 2.05074623e-02 7.40204871e-01
7.30073631e-01 3.01134080e-01 -2.60617167e-01 7.07982481e-01
-1.08348262e+00 -3.36403191e-01 5.29399693e-01 3.03239435e-01
-4.30651218e-01 -1.32983577e+00 -4.41323876e-01 6.96236908e-01
-5.93238845e-02 -8.37537706e-01 -3.93281877e-01 5.77301264e-01
7.83570483e-02 7.67646194e-01 5.82592487e-01 -7.56585300e-01
4.23812062e-01 3.95391762e-01 -7.33477399e-02 -1.03031433e+00
-8.30075741e-01 -3.13260019e-01 2.01930374e-01 -6.84443235e-01
5.95810175e-01 -4.60072428e-01 -1.37087917e+00 -5.70924520e-01
1.63318709e-01 9.49511230e-02 4.22850758e-01 9.85093057e-01
4.61463541e-01 3.73124838e-01 2.64494598e-01 -5.33632815e-01
-9.68901813e-01 -1.11316514e+00 -4.21099335e-01 8.45728338e-01
1.84493765e-01 -2.14511499e-01 -3.03077698e-01 8.12208503e-02] | [10.589326858520508, 8.744301795959473] |
b458df83-a112-4661-99e9-f7496884943d | l-seqsleepnet-whole-cycle-long-sequence | 2301.03441 | null | https://arxiv.org/abs/2301.03441v2 | https://arxiv.org/pdf/2301.03441v2.pdf | L-SeqSleepNet: Whole-cycle Long Sequence Modelling for Automatic Sleep Staging | Human sleep is cyclical with a period of approximately 90 minutes, implying long temporal dependency in the sleep data. Yet, exploring this long-term dependency when developing sleep staging models has remained untouched. In this work, we show that while encoding the logic of a whole sleep cycle is crucial to improve sleep staging performance, the sequential modelling approach in existing state-of-the-art deep learning models are inefficient for that purpose. We thus introduce a method for efficient long sequence modelling and propose a new deep learning model, L-SeqSleepNet, which takes into account whole-cycle sleep information for sleep staging. Evaluating L-SeqSleepNet on four distinct databases of various sizes, we demonstrate state-of-the-art performance obtained by the model over three different EEG setups, including scalp EEG in conventional Polysomnography (PSG), in-ear EEG, and around-the-ear EEG (cEEGrid), even with a single EEG channel input. Our analyses also show that L-SeqSleepNet is able to alleviate the predominance of N2 sleep (the major class in terms of classification) to bring down errors in other sleep stages. Moreover the network becomes much more robust, meaning that for all subjects where the baseline method had exceptionally poor performance, their performance are improved significantly. Finally, the computation time only grows at a sub-linear rate when the sequence length increases. | ['Maarten De Vos', 'Kaare Mikkelsen', 'Mathias Baumert', 'Alfred Mertins', 'Philipp Koch', 'Minh C. Tran', 'Oliver Y. Chén', 'Elisabeth Heremans', 'Kristian P. Lorenzen', 'Huy Phan'] | 2023-01-09 | null | null | null | null | ['sleep-staging'] | ['medical'] | [ 7.02752396e-02 -9.93761346e-02 9.18762907e-02 -3.25323224e-01
-2.17656150e-01 -2.85923094e-01 1.41594425e-01 -1.66945934e-01
-6.92593873e-01 9.01052952e-01 1.48013353e-01 -3.37306857e-01
-3.85060370e-01 -1.84549615e-01 -3.23108286e-01 -7.43961751e-01
-2.27240041e-01 2.00918183e-01 2.10877523e-01 -3.50047916e-01
6.46023080e-02 1.85596347e-01 -1.60513270e+00 1.97477847e-01
7.39994943e-01 1.05521405e+00 3.41889650e-01 6.78626478e-01
2.05720067e-01 5.46822071e-01 -8.39183867e-01 -2.74561286e-01
1.70950219e-02 -6.32654727e-01 -6.95767879e-01 -3.15062046e-01
1.88235164e-01 2.63696194e-01 -1.81840003e-01 9.52226520e-01
8.88114333e-01 -4.67998236e-02 6.37480393e-02 -1.02174067e+00
-1.12799950e-01 4.77010369e-01 -1.37385145e-01 9.29079473e-01
3.92097354e-01 2.24945173e-01 7.78473556e-01 -3.83907884e-01
2.02354297e-01 4.02443707e-01 1.05998695e+00 7.75154233e-01
-1.36971986e+00 -8.89561117e-01 -2.29899436e-01 6.14756107e-01
-1.44170105e+00 -6.68191850e-01 5.25693357e-01 -1.67703807e-01
1.46971309e+00 4.21108902e-01 1.29599655e+00 1.41604912e+00
8.68089795e-01 3.62035453e-01 1.38931620e+00 -2.65771002e-01
2.98547298e-01 -7.98450336e-02 4.24540460e-01 4.10053819e-01
1.30455837e-01 3.71142216e-02 -1.00983870e+00 4.19674337e-01
1.67741477e-01 1.01147830e-01 -2.03799039e-01 2.25095868e-01
-9.77395713e-01 3.12269658e-01 2.29720518e-01 7.93660164e-01
-3.27759534e-01 -3.33468914e-02 5.28539956e-01 5.10643840e-01
6.91145301e-01 6.24310851e-01 -7.69744992e-01 -7.84644425e-01
-1.87230122e+00 -4.93252128e-02 9.84151900e-01 7.02730238e-01
6.09925032e-01 1.28538604e-03 -2.40064159e-01 7.33002365e-01
-1.20923549e-01 1.50804818e-01 1.00023913e+00 -5.77127099e-01
1.06619827e-01 6.45402789e-01 -1.84006929e-01 -4.95170444e-01
-1.31632352e+00 -8.26429844e-01 -1.23957872e+00 -1.89074442e-01
1.97392330e-01 -1.82912543e-01 -7.96477616e-01 1.73878264e+00
-4.74414915e-01 4.21133846e-01 -1.95665196e-01 6.53085649e-01
7.33291924e-01 3.16872001e-01 -3.71751972e-02 -4.23899680e-01
1.65824163e+00 -8.51982594e-01 -8.12715888e-01 -8.05065930e-01
3.43819261e-01 -3.70000392e-01 1.05091882e+00 8.47189784e-01
-1.26293445e+00 -7.45874524e-01 -1.28581727e+00 -1.51629761e-01
-3.14078987e-01 1.00755781e-01 4.52220529e-01 8.63011360e-01
-1.66803992e+00 7.27749944e-01 -1.22573960e+00 -5.37331522e-01
3.61644030e-01 1.01709127e+00 -3.36326420e-01 3.79049301e-01
-1.22380352e+00 1.14561307e+00 2.10776508e-01 2.54717141e-01
-7.30449021e-01 -7.65841961e-01 -5.36941230e-01 5.08117914e-01
7.63289258e-02 -7.49912262e-01 9.99463022e-01 -7.17619359e-01
-1.35552788e+00 9.94302869e-01 -5.93641281e-01 -7.58466959e-01
1.98608592e-01 -1.62269384e-01 -7.18565762e-01 -8.11026022e-02
-5.88265918e-02 6.01757586e-01 8.67592692e-01 -5.02451003e-01
-2.54802018e-01 -4.41407025e-01 -7.07679540e-02 -2.92172849e-01
-4.04368699e-01 8.08481053e-02 -3.07279229e-01 -5.01927257e-01
-2.67142415e-01 -1.00501537e+00 -5.38250916e-02 -6.10397756e-01
-2.69556254e-01 -8.16418976e-02 1.91726074e-01 -5.42386830e-01
1.71696782e+00 -2.17729330e+00 2.85333395e-01 -1.57361209e-01
4.22818244e-01 2.85931945e-01 3.93946171e-02 4.06030685e-01
-4.54192579e-01 2.18259115e-02 -3.12383503e-01 -8.54334056e-01
2.73315720e-02 4.26798642e-01 1.39049113e-01 5.06835222e-01
-1.68353170e-01 1.05908048e+00 -6.67483151e-01 -1.39074564e-01
1.18987523e-02 3.19115222e-01 -5.24907589e-01 1.43282980e-01
4.07450289e-01 4.11276758e-01 4.30000842e-01 2.64492422e-01
4.51315105e-01 -4.44256067e-01 1.16162159e-01 -2.81831250e-02
-2.19230011e-01 6.38933480e-01 -6.08019412e-01 2.13244367e+00
-7.07481980e-01 9.29062128e-01 -1.73379838e-01 -8.73005748e-01
4.83603209e-01 3.01406413e-01 4.49713767e-01 -1.03259587e+00
1.90950528e-01 2.88489312e-02 4.91746217e-01 -4.60198134e-01
4.11206365e-01 -4.88484442e-01 1.17463410e-01 5.53462923e-01
5.12269199e-01 2.25306809e-01 5.98417401e-01 -1.08933821e-01
1.45085287e+00 -3.22074085e-01 4.11822349e-01 -7.15607822e-01
5.19021690e-01 -5.81318855e-01 5.63104928e-01 7.21884131e-01
-2.49223590e-01 5.44769049e-01 7.31374085e-01 -4.63402748e-01
-6.78614616e-01 -7.49818921e-01 -1.98215798e-01 1.07658267e+00
-4.03400287e-02 -9.60945904e-01 -1.03603995e+00 -3.87761742e-01
-5.17517030e-01 6.75474823e-01 -8.05327654e-01 -4.58503991e-01
-4.18333173e-01 -1.25280762e+00 6.89221919e-01 4.64077115e-01
3.59220058e-01 -1.30656362e+00 -1.02876127e+00 2.35052332e-01
-6.46782964e-02 -1.07173812e+00 -2.75891602e-01 8.65865171e-01
-1.01893473e+00 -9.47943330e-01 -5.72673500e-01 -4.02900875e-01
2.71558672e-01 -2.46824831e-01 1.28708041e+00 1.94831192e-01
-3.40055853e-01 -8.23719800e-02 -3.40257019e-01 -3.34151864e-01
-6.11553527e-02 5.63969910e-01 3.31016660e-01 -2.35538110e-01
7.28018701e-01 -1.14540982e+00 -8.77097726e-01 2.44677991e-01
-7.66224563e-01 8.06698203e-02 5.47900319e-01 7.22017407e-01
1.38036579e-01 5.21417223e-02 6.33539677e-01 -6.19867861e-01
6.33207023e-01 -4.36102271e-01 -3.27486664e-01 -1.22120194e-01
-9.76571262e-01 7.56041035e-02 7.25706041e-01 -2.15413883e-01
-4.65557963e-01 -2.72313505e-01 -6.92540646e-01 -1.27457559e-01
-1.62231445e-01 2.66807884e-01 1.31573990e-01 -3.34847867e-02
5.80299079e-01 6.48054898e-01 -1.28954932e-01 -5.02920449e-01
-2.53130496e-01 4.28511351e-01 3.61507475e-01 8.48501921e-02
2.84462988e-01 2.75398105e-01 -1.00992816e-02 -1.10631359e+00
-9.01837766e-01 -5.77449322e-01 -8.17468941e-01 8.26206896e-03
1.00440049e+00 -8.22479784e-01 -9.33108628e-01 5.02746999e-01
-9.60544229e-01 -6.55193567e-01 -2.91610211e-01 3.39377403e-01
-4.01770383e-01 1.29486710e-01 -5.46665490e-01 -6.03343368e-01
-5.88387787e-01 -1.28837657e+00 1.00635421e+00 3.90996672e-02
-6.41356051e-01 -1.01431406e+00 1.81798339e-01 3.80772315e-02
4.74380493e-01 -3.77014577e-01 7.91303515e-01 -6.43506467e-01
-1.56887785e-01 1.44724965e-01 6.37595505e-02 5.80075026e-01
8.22484270e-02 -6.18946910e-01 -1.26758063e+00 -3.83099586e-01
4.82928842e-01 9.86848474e-02 7.71661818e-01 4.14285183e-01
1.19562399e+00 -1.47472456e-01 -9.28603336e-02 1.01569688e+00
1.24715662e+00 1.75519332e-01 1.00547945e+00 3.46988082e-01
3.16022187e-01 1.98213667e-01 -1.86030388e-01 3.92962277e-01
3.22113037e-01 5.07191181e-01 2.17847914e-01 1.95794832e-02
-2.41662234e-01 3.55711877e-01 5.11241615e-01 1.23155141e+00
-1.49282545e-01 -3.63400519e-01 -8.61861229e-01 5.47419786e-01
-1.37640607e+00 -8.43456507e-01 -1.13569677e-01 1.81486988e+00
8.58960092e-01 4.08096433e-01 1.85955495e-01 4.99667138e-01
6.83625937e-02 2.95467824e-01 -4.83689815e-01 -5.47683597e-01
-3.25698107e-02 1.00730574e+00 3.55218410e-01 6.98363706e-02
-6.25471294e-01 6.39056802e-01 7.06399679e+00 6.66303039e-01
-1.34402168e+00 6.81410849e-01 2.84803241e-01 -6.39695168e-01
2.89858431e-01 -3.48507375e-01 -7.47179031e-01 1.08139920e+00
1.81460547e+00 4.87265997e-02 8.36569607e-01 5.29258907e-01
4.83882457e-01 -4.35225964e-01 -1.26741338e+00 1.43890429e+00
1.75858364e-01 -1.27721584e+00 -6.51641607e-01 3.09771933e-02
5.05670965e-01 5.20632923e-01 -2.38643855e-01 3.83035958e-01
-5.32567501e-01 -1.00789595e+00 6.98517382e-01 6.77680969e-01
1.02521789e+00 -6.50303364e-01 9.84591961e-01 4.59248513e-01
-9.89166379e-01 -2.32807443e-01 -3.28315735e-01 -3.99147153e-01
1.29985005e-01 7.21803308e-01 -4.99183863e-01 4.91180629e-01
9.80442226e-01 1.00212455e+00 -1.19610214e+00 1.14544535e+00
-4.12723005e-01 9.88969445e-01 -1.94023088e-01 4.11884822e-02
7.77579173e-02 1.92857776e-02 2.74688631e-01 1.39528406e+00
4.52763796e-01 -2.86262743e-02 -5.98872423e-01 7.47686148e-01
-6.00508340e-02 -5.27505517e-01 -2.75037766e-01 1.17102802e-01
3.64397429e-02 1.28037810e+00 -9.12022889e-01 -1.18593179e-01
-3.28727156e-01 1.20500028e+00 1.54705301e-01 2.31206253e-01
-5.99530637e-01 -3.87742609e-01 5.65253317e-01 8.66376385e-02
1.66221172e-01 -1.29580840e-01 -6.40768409e-01 -1.24880683e+00
3.19892652e-02 -7.44169950e-01 1.10597968e-01 -9.37160075e-01
-1.02148104e+00 9.45752382e-01 -9.97598916e-02 -9.68588769e-01
-2.83129781e-01 -5.05111575e-01 -6.42890155e-01 7.88767457e-01
-1.34771109e+00 -6.88226581e-01 -3.01146895e-01 6.17432415e-01
6.00228846e-01 2.05018166e-02 7.91946292e-01 8.05365920e-01
-6.61616564e-01 6.92975163e-01 -1.71077847e-01 -3.37571144e-01
5.47297299e-01 -1.39646435e+00 5.29697597e-01 9.57669377e-01
2.23430663e-01 9.91716027e-01 7.09311366e-01 -2.36661658e-01
-1.20429647e+00 -7.56870151e-01 1.22165096e+00 -5.58486104e-01
6.96447670e-01 -7.90175140e-01 -7.34255433e-01 5.47732174e-01
5.43777466e-01 -2.13151202e-01 8.50960612e-01 2.68852621e-01
2.93112308e-01 -5.28998017e-01 -7.22914636e-01 2.26647347e-01
1.17711639e+00 -7.28618562e-01 -9.74308431e-01 1.15569070e-01
4.22319621e-01 -2.27710858e-01 -6.53392911e-01 1.19428940e-01
6.85076356e-01 -1.54276347e+00 5.61891019e-01 -1.16074264e-01
2.51651168e-01 5.61640523e-02 3.44549954e-01 -1.46892095e+00
-2.55408555e-01 -1.05107868e+00 -2.20083132e-01 7.41148353e-01
2.55073428e-01 -6.06247485e-01 8.57262850e-01 2.10512623e-01
-6.57211959e-01 -8.58437240e-01 -1.26633203e+00 -8.71764004e-01
-4.08174336e-01 -7.32465327e-01 5.32709897e-01 3.80166769e-01
9.73205864e-02 6.37558639e-01 -3.84547681e-01 -1.22817405e-01
4.72163744e-02 -2.28696629e-01 3.34394217e-01 -1.29683363e+00
-3.55849043e-02 -3.88234109e-01 -5.44037819e-01 -9.07365441e-01
2.48213515e-01 -1.01211333e+00 1.10803947e-01 -1.36492765e+00
2.68290132e-01 -9.50202942e-02 -7.64269412e-01 5.76336026e-01
-5.53440340e-02 5.82327366e-01 1.08063340e-01 -2.37221867e-02
-7.33571887e-01 2.68368810e-01 8.20411682e-01 2.00892419e-01
-1.23914130e-01 9.84995142e-02 -7.21539199e-01 6.95073307e-01
7.00235605e-01 -7.08010435e-01 -4.36552167e-01 -2.69451678e-01
6.83141172e-01 -4.58850525e-02 3.16875041e-01 -1.60882986e+00
4.39842254e-01 7.07307994e-01 4.19705629e-01 -5.01641333e-01
4.75443661e-01 -7.13032484e-01 2.67684013e-01 5.55843472e-01
-1.02686800e-01 5.01342237e-01 5.35366535e-01 2.47559384e-01
-7.84948841e-02 -1.33917481e-01 6.31683111e-01 -1.30866274e-01
-6.16318703e-01 1.11817501e-01 -6.74890101e-01 1.98927239e-01
4.43561316e-01 -3.54125410e-01 -9.80754867e-02 -7.36627504e-02
-1.02390623e+00 -4.83181253e-02 2.67101020e-01 2.80394465e-01
3.44193071e-01 -7.85901546e-01 -1.66916192e-01 8.02873552e-01
-6.50796741e-02 -5.21657467e-01 5.05976498e-01 1.49682117e+00
-2.79398382e-01 7.20827520e-01 -4.63314027e-01 -6.66133642e-01
-1.19314110e+00 4.12935257e-01 3.76571596e-01 -4.77976739e-01
-5.63093781e-01 1.03557682e+00 -1.73634030e-02 -6.33684034e-03
1.89859465e-01 -8.68061841e-01 -2.51107305e-01 4.93692130e-01
5.43104112e-01 5.43221772e-01 9.18479323e-01 -3.52747858e-01
-6.30715311e-01 4.88506764e-01 2.21517161e-01 2.43924946e-01
1.51912761e+00 -2.37918586e-01 -5.68658888e-01 7.23939240e-01
1.03477156e+00 -8.71835798e-02 -9.09827590e-01 5.59614778e-01
9.89921540e-02 1.12238184e-01 1.04764678e-01 -9.89998043e-01
-1.06702864e+00 1.14217401e+00 9.54499125e-01 3.69660825e-01
1.55584490e+00 -3.01163971e-01 9.50781882e-01 3.55249107e-01
4.98336017e-01 -7.45606303e-01 -2.42045850e-01 5.68487704e-01
5.12032568e-01 -8.93262327e-01 8.76143873e-02 3.73365074e-01
-4.32764947e-01 1.07125723e+00 2.76412576e-01 -7.00939745e-02
7.99651861e-01 3.63139540e-01 -2.70123541e-01 -4.83740747e-01
-8.79621923e-01 -2.17313915e-01 3.61282706e-01 3.21909636e-01
3.02193254e-01 -1.10010229e-01 -4.65754956e-01 1.02924657e+00
-7.18504608e-01 4.07259405e-01 5.08125246e-01 6.31732941e-01
-1.51509538e-01 -1.10796714e+00 9.55103040e-02 7.47269034e-01
-9.46422279e-01 -5.62859356e-01 -7.52375880e-03 7.80640185e-01
5.22455156e-01 1.01016212e+00 1.46367744e-01 -6.91434026e-01
2.70470649e-01 3.78212512e-01 5.58847249e-01 -8.57566297e-01
-1.09744751e+00 -1.34232000e-01 -6.11501150e-02 -9.42243040e-01
-5.96356511e-01 -3.93688411e-01 -8.77197981e-01 -2.09181085e-01
-7.86382481e-02 2.08742127e-01 5.11994898e-01 1.21698141e+00
5.11107743e-01 9.32322979e-01 -3.64974886e-02 -1.08252060e+00
-7.96960294e-03 -1.27961648e+00 -9.03082252e-01 -3.42143737e-02
7.54483461e-01 -5.59683204e-01 -4.10026282e-01 4.86569814e-02] | [13.375988006591797, 3.6271185874938965] |
a4e21d6a-8900-4830-9e17-55081baab275 | the-elements-of-end-to-end-deep-face | 2009.13290 | null | https://arxiv.org/abs/2009.13290v4 | https://arxiv.org/pdf/2009.13290v4.pdf | The Elements of End-to-end Deep Face Recognition: A Survey of Recent Advances | Face recognition is one of the most popular and long-standing topics in computer vision. With the recent development of deep learning techniques and large-scale datasets, deep face recognition has made remarkable progress and been widely used in many real-world applications. Given a natural image or video frame as input, an end-to-end deep face recognition system outputs the face feature for recognition. To achieve this, a typical end-to-end system is built with three key elements: face detection, face alignment, and face representation. The face detection locates faces in the image or frame. Then, the face alignment is proceeded to calibrate the faces to the canonical view and crop them with a normalized pixel size. Finally, in the stage of face representation, the discriminative features are extracted from the aligned face for recognition. Nowadays, all of the three elements are fulfilled by the technique of deep convolutional neural network. In this survey article, we present a comprehensive review about the recent advance of each element. To start with, we present an overview of the end-to-end deep face recognition. Then, we review the advance of each element, respectively, covering many aspects such as the to-date algorithm designs, evaluation metrics, datasets, performance comparison, existing challenges, and promising directions for future research. Also, we provide a detailed discussion about the effect of each element on its subsequent elements and the holistic system. Through this survey, we wish to bring contributions in two aspects: first, readers can conveniently identify the methods which are quite strong-baseline style in the subcategory for further exploration; second, one can also employ suitable methods for establishing a state-of-the-art end-to-end face recognition system from scratch. | ['Xiao-Ping Zhang', 'Hailin Shi', 'Hang Du', 'Dan Zeng', 'Tao Mei'] | 2020-09-28 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [ 2.26956427e-01 -4.07039076e-01 -1.18516229e-01 -8.64336491e-01
-5.17696083e-01 -2.25692332e-01 3.82607788e-01 -8.43806326e-01
-2.37397075e-01 2.14827791e-01 -1.44820184e-01 2.37460002e-01
-3.32799554e-02 -6.15889668e-01 -4.96755183e-01 -9.50822890e-01
-2.30845045e-02 2.15671927e-01 -5.36428273e-01 -1.02447197e-02
1.06643461e-01 9.94624734e-01 -1.91775870e+00 2.82794982e-01
8.99738893e-02 1.46418238e+00 -1.35312378e-01 4.37219799e-01
-4.32437584e-02 4.36805427e-01 -5.04606783e-01 -7.04983234e-01
3.83054942e-01 -3.91033262e-01 -5.59804142e-01 5.09734213e-01
8.72291565e-01 -6.01875722e-01 -6.16865337e-01 1.02482867e+00
9.91232336e-01 4.56350259e-02 5.31819046e-01 -1.33630204e+00
-5.81629336e-01 3.54573280e-02 -6.49806440e-01 9.23911482e-02
2.67141104e-01 3.31457146e-02 4.54678476e-01 -1.57929325e+00
3.69341046e-01 1.42479050e+00 6.35455191e-01 8.94659102e-01
-7.48197615e-01 -8.02899659e-01 9.42979902e-02 3.85553300e-01
-1.55640352e+00 -1.22338414e+00 8.48859310e-01 -3.06582332e-01
8.74955595e-01 7.33214095e-02 3.19323927e-01 9.93696213e-01
-2.73413926e-01 8.27340961e-01 7.50552356e-01 -5.45120299e-01
-1.29634693e-01 -3.39964241e-01 1.14286534e-01 9.73964572e-01
-6.10587001e-02 2.78709948e-01 -5.82071304e-01 1.16040140e-01
7.09275842e-01 2.44641587e-01 -1.06320664e-01 -1.52207181e-01
-6.79548204e-01 6.68692946e-01 3.56701255e-01 3.47663671e-01
-3.50069880e-01 -1.32291734e-01 4.76547211e-01 2.74372488e-01
3.68103564e-01 -3.84215623e-01 -1.81615442e-01 1.61803603e-01
-1.14346385e+00 1.21282497e-02 6.24598265e-01 7.78665721e-01
6.74432874e-01 3.60957891e-01 -2.73629874e-01 1.17943954e+00
4.03141320e-01 4.75613922e-01 3.44830722e-01 -7.14963317e-01
1.38149530e-01 3.41150224e-01 -3.41580600e-01 -8.99075925e-01
-3.78802329e-01 -2.65413880e-01 -1.07650173e+00 3.07481140e-01
2.45500848e-01 -1.33448452e-01 -9.06988382e-01 1.66444707e+00
4.36909229e-01 3.33903044e-01 -6.91989064e-02 8.96892548e-01
1.30416358e+00 4.60092008e-01 -1.38885915e-01 -4.59700227e-01
1.81241155e+00 -1.03939974e+00 -7.03101814e-01 -2.18305156e-01
2.21051108e-02 -9.04696167e-01 5.57717681e-01 9.94690806e-02
-1.03760803e+00 -9.94245231e-01 -8.83404374e-01 -1.89069793e-01
-2.50560522e-01 7.15074956e-01 5.00595152e-01 9.68526959e-01
-1.21342611e+00 2.75238901e-01 -7.30890334e-01 -5.73449492e-01
7.44599342e-01 6.63397133e-01 -6.81624651e-01 -2.71366298e-01
-8.00599515e-01 6.11845016e-01 -1.32028744e-01 6.31332517e-01
-1.07279289e+00 -3.36963832e-01 -7.75990129e-01 4.17029746e-02
2.62893796e-01 -5.30398548e-01 1.26971972e+00 -1.27715397e+00
-1.64697647e+00 1.37119639e+00 -5.54298222e-01 9.62167829e-02
3.05750668e-01 -1.30860358e-01 -7.00781405e-01 8.28089565e-02
-2.04335868e-01 6.61483347e-01 1.24516296e+00 -9.03286636e-01
-5.77363551e-01 -8.66340339e-01 -2.09243461e-01 6.95172250e-02
-5.07455230e-01 8.41775775e-01 -1.01575744e+00 -5.46972215e-01
-9.54254419e-02 -6.62377715e-01 2.23497555e-01 3.41647297e-01
-9.80511457e-02 -4.72669184e-01 9.79448140e-01 -4.79962468e-01
1.05875742e+00 -2.55856705e+00 3.60080041e-02 -9.68391374e-02
6.50837719e-02 6.04453444e-01 -4.22140032e-01 2.68647522e-01
-5.04802406e-01 -3.92674327e-01 -2.21144259e-01 -6.81643486e-01
-8.00977647e-02 -1.58419147e-01 -2.97895998e-01 7.39293814e-01
3.21357459e-01 9.22753096e-01 -4.61486369e-01 -2.10519344e-01
3.01656604e-01 8.93416107e-01 -2.55012065e-01 4.71886545e-01
2.58203536e-01 3.62887233e-01 -2.11391553e-01 9.23669875e-01
9.99222875e-01 1.35618255e-01 1.01832218e-01 -4.13039416e-01
-4.79780929e-03 -2.76651651e-01 -1.16072798e+00 1.26218426e+00
-2.68766135e-01 6.44643903e-01 5.11652708e-01 -1.16097856e+00
1.11284339e+00 4.33937222e-01 4.40324306e-01 -5.06402791e-01
1.40694052e-01 3.08953315e-01 -1.59129322e-01 -5.97952366e-01
2.37684790e-02 -1.01514324e-03 3.97880495e-01 2.82785267e-01
2.64619291e-01 5.03082335e-01 5.66481501e-02 -4.16217297e-01
5.02959907e-01 -6.28680065e-02 4.78541344e-01 1.26190353e-02
1.11670566e+00 -6.86230838e-01 3.79178733e-01 2.72588044e-01
-5.31856775e-01 6.35428131e-01 2.86817133e-01 -9.86323953e-01
-6.49361134e-01 -6.65973365e-01 -2.74326473e-01 1.21137059e+00
-3.76093775e-01 -1.79820746e-01 -1.13160264e+00 -7.24469662e-01
-2.13951692e-02 -1.83385596e-01 -8.18043768e-01 5.30862138e-02
-7.13558614e-01 -8.58081818e-01 6.67119682e-01 7.37813950e-01
9.38678384e-01 -1.32703805e+00 -2.38504410e-01 -1.65916994e-01
6.24477398e-03 -1.11080849e+00 -5.12078285e-01 -2.71424532e-01
-7.06110835e-01 -1.09307373e+00 -9.96776938e-01 -1.38000536e+00
9.35368121e-01 5.58100939e-01 1.00763988e+00 4.72859323e-01
-4.70629573e-01 4.09527481e-01 -9.68797132e-02 -8.38406980e-02
1.68133974e-01 -2.20672235e-01 3.65961313e-01 5.08469522e-01
9.50960279e-01 -3.05670589e-01 -6.81164086e-01 4.16262507e-01
-6.60424650e-01 -3.98805231e-01 5.23471773e-01 9.33153629e-01
4.36036378e-01 -1.94607288e-01 5.92376471e-01 -4.90299225e-01
2.17965439e-01 -1.04034074e-01 -6.65324152e-01 3.20280492e-01
-2.31179789e-01 -4.25761968e-01 5.56255698e-01 -7.04321191e-02
-9.85099614e-01 4.98830587e-01 -6.62480891e-01 -4.66328919e-01
-3.39522392e-01 2.02188879e-01 -5.96489608e-01 -3.34475398e-01
4.08294857e-01 4.79369164e-01 3.21776599e-01 -5.40544987e-01
3.42994481e-01 1.07843840e+00 6.11330807e-01 -4.21623170e-01
6.08018577e-01 4.82338637e-01 -1.25565633e-01 -1.18533242e+00
-7.07330287e-01 -2.62661010e-01 -8.02550018e-01 -3.65640044e-01
6.63055718e-01 -8.97970736e-01 -9.83014584e-01 1.05204654e+00
-1.25609243e+00 1.76067222e-02 1.38893127e-01 1.96710497e-01
-4.55868393e-01 3.55521172e-01 -6.78358078e-01 -8.68465543e-01
-5.71677387e-01 -1.47295356e+00 1.21291316e+00 6.04783893e-01
3.17810595e-01 -7.21053600e-01 -2.18723074e-01 4.18943733e-01
3.84953707e-01 -1.28894776e-01 5.44503629e-01 -5.64109147e-01
-3.18163663e-01 -3.20842147e-01 -3.67692083e-01 4.94612992e-01
2.57566214e-01 2.70755321e-01 -1.50645185e+00 -7.11798608e-01
1.83359236e-01 -4.35024321e-01 7.91522861e-01 4.44268942e-01
1.44946933e+00 -9.03617293e-02 -4.04474378e-01 1.00946259e+00
1.19536269e+00 3.22230726e-01 7.71332383e-01 -7.42841735e-02
5.28387189e-01 8.28504324e-01 3.34099084e-01 2.66704619e-01
1.00825019e-01 9.45334256e-01 2.24904537e-01 -1.26221538e-01
-3.36330533e-01 4.69754189e-02 5.64903259e-01 4.90570754e-01
-1.90330088e-01 -2.01506130e-02 -5.15750289e-01 9.71413851e-02
-1.36178589e+00 -1.08558488e+00 3.38115454e-01 2.17587113e+00
3.99874151e-01 -5.45612693e-01 3.91137540e-01 9.83074829e-02
1.11248028e+00 2.63279349e-01 -5.74189961e-01 -1.05371155e-01
-6.09347485e-02 2.67577112e-01 -8.79883394e-02 4.52904910e-01
-1.25623274e+00 1.08586085e+00 6.67226791e+00 8.38077724e-01
-1.41197014e+00 -6.99016303e-02 1.03780520e+00 3.77041586e-02
6.37296796e-01 -5.25942147e-01 -1.29792523e+00 3.15615535e-01
6.79619491e-01 3.10343534e-01 6.46695316e-01 9.88989949e-01
2.58209370e-03 2.64160126e-01 -1.34508741e+00 1.57599521e+00
5.73574603e-01 -1.11512661e+00 6.48659319e-02 4.83783893e-02
4.86221015e-01 -1.91566184e-01 2.93635368e-01 2.71974415e-01
-4.40738261e-01 -1.29560685e+00 4.09985214e-01 3.31966460e-01
1.26504457e+00 -8.71578038e-01 6.68576479e-01 -6.07222319e-02
-1.46484685e+00 -2.29351729e-01 -5.17306030e-01 3.90972290e-03
-7.48950243e-02 4.90355015e-01 -2.03095838e-01 4.42536533e-01
6.45524859e-01 6.21970475e-01 -2.92973220e-01 8.02331090e-01
-1.16524816e-01 2.73230582e-01 -1.83850855e-01 2.28960484e-01
6.79985993e-03 -1.99793339e-01 7.63990283e-02 1.29303408e+00
3.18011343e-01 2.00931400e-01 7.88295791e-02 5.08525550e-01
-4.48307067e-01 1.28513090e-02 -5.90274692e-01 -1.10850958e-02
4.86055821e-01 1.61425960e+00 -6.44888937e-01 -2.54343063e-01
-8.73203933e-01 9.93653238e-01 3.83071423e-01 2.87202090e-01
-6.40486240e-01 -4.04420644e-01 1.00111210e+00 -3.22035313e-01
3.30099255e-01 -6.06727488e-02 -8.28101560e-02 -8.20939362e-01
2.12509751e-01 -1.22861290e+00 4.25443560e-01 -2.94610113e-01
-1.27236891e+00 9.93759751e-01 -2.23948270e-01 -9.89879429e-01
-1.76360160e-01 -1.07692683e+00 -6.69400215e-01 9.23002839e-01
-1.50817537e+00 -1.11161280e+00 -6.18449748e-01 7.68303037e-01
7.89039612e-01 -7.33114481e-01 9.61613417e-01 6.66362643e-01
-1.17747068e+00 1.07246208e+00 1.20055869e-01 7.35371768e-01
7.47776508e-01 -4.01581287e-01 6.25739753e-01 9.73969579e-01
2.53113896e-01 7.42429972e-01 9.55812261e-02 -2.75625676e-01
-1.65525627e+00 -1.02256119e+00 8.03524673e-01 -1.84853718e-01
5.96024245e-02 -4.49209303e-01 -7.18324721e-01 5.88249385e-01
1.35143116e-01 5.82420707e-01 6.78703308e-01 -8.85960609e-02
-4.20066237e-01 -6.78213537e-01 -1.26477122e+00 4.63671625e-01
1.15551543e+00 -6.42186642e-01 -1.91186950e-01 2.90023506e-01
-5.57101667e-02 -3.29758644e-01 -5.37973046e-01 3.88205200e-01
9.46337044e-01 -1.09627759e+00 1.03166509e+00 -5.71667492e-01
2.42822424e-01 -2.73290396e-01 -2.72533149e-01 -9.26928341e-01
-5.18666506e-01 -7.25885928e-01 -1.59715071e-01 1.45350051e+00
-1.05527535e-01 -5.21313190e-01 8.80079746e-01 4.92306501e-01
2.00438071e-02 -8.75548065e-01 -9.97086048e-01 -5.92128932e-01
-1.55675337e-01 -1.96919650e-01 6.48667634e-01 5.60004115e-01
-3.76079708e-01 2.93195695e-01 -3.31823260e-01 1.07879035e-01
6.24102414e-01 1.02670640e-01 8.21356595e-01 -1.03690004e+00
1.82052881e-01 -5.13636708e-01 -5.41147768e-01 -1.30578303e+00
4.00770813e-01 -7.51729071e-01 -3.18736583e-02 -1.07124567e+00
4.29993540e-01 9.80413482e-02 -1.51782483e-01 4.96617347e-01
-1.08580463e-01 7.62716889e-01 2.07756713e-01 1.77720264e-01
-2.75371999e-01 5.19856930e-01 9.78207469e-01 -1.42225819e-02
1.23670094e-01 1.29891053e-01 -7.68277228e-01 6.27620816e-01
5.04853308e-01 -1.60078555e-02 -1.22783780e-01 -6.24905229e-01
-4.86833751e-01 -1.29531145e-01 1.81928858e-01 -9.46888685e-01
3.50384265e-01 1.40718207e-01 9.35068727e-01 -4.20266926e-01
5.59533656e-01 -8.76402736e-01 -4.50154059e-02 4.24748629e-01
-2.26964131e-02 1.89054191e-01 2.39590406e-01 5.75980581e-02
-4.36666489e-01 -1.92094028e-01 1.29362702e+00 -4.60212759e-04
-8.81622255e-01 7.62388587e-01 -1.98341638e-01 -4.26979840e-01
1.18859375e+00 -4.42691386e-01 -2.47890502e-02 -2.61295557e-01
-5.73302627e-01 -1.21135237e-02 1.52757511e-01 5.53476334e-01
7.63034225e-01 -1.44437087e+00 -9.60926652e-01 9.05166924e-01
-6.56460226e-02 -3.62289786e-01 5.34376323e-01 7.27533281e-01
-3.32315385e-01 5.15792847e-01 -4.35183078e-01 -6.08123362e-01
-1.68655801e+00 6.71409190e-01 5.60717583e-01 3.05065900e-01
-3.27737063e-01 1.12973976e+00 3.66821676e-01 -1.86362237e-01
6.08382463e-01 3.48254830e-01 -3.90320599e-01 1.27957553e-01
1.20553434e+00 2.54027337e-01 3.40558589e-01 -1.12363243e+00
-7.74825633e-01 9.91209447e-01 -1.34317011e-01 2.26374581e-01
1.37653947e+00 -5.10072336e-02 -3.95925522e-01 -2.21990019e-01
1.39667237e+00 -2.15312570e-01 -1.29896677e+00 -2.04353005e-01
-3.12051952e-01 -5.75540960e-01 -4.18554805e-02 -5.51351368e-01
-1.83844149e+00 1.11238527e+00 9.89030480e-01 -2.28293434e-01
1.55671763e+00 -1.93112746e-01 5.27202070e-01 3.20313931e-01
2.77806461e-01 -6.57984674e-01 6.27105534e-02 6.43143713e-01
1.11872602e+00 -1.24629939e+00 -1.60346463e-01 -4.60881919e-01
-2.85500944e-01 1.46742666e+00 7.65359521e-01 1.85415596e-02
8.05656731e-01 3.56406599e-01 2.48601913e-01 -1.27512559e-01
-3.07899594e-01 -2.06236795e-01 1.75241441e-01 7.61937559e-01
7.82692969e-01 -2.49429762e-01 1.11062638e-01 4.86153841e-01
1.17894160e-02 1.01463884e-01 -2.13456377e-01 5.48208177e-01
-2.94237554e-01 -1.22070241e+00 -6.19847536e-01 2.71915168e-01
-5.74892163e-01 4.97530699e-02 -4.31185305e-01 4.20515954e-01
8.59604105e-02 9.51941192e-01 2.41712734e-01 -4.04918015e-01
3.09203416e-01 2.03585804e-01 7.19423771e-01 -4.31089312e-01
-3.67883295e-01 -2.48207208e-02 -2.63525277e-01 -5.39878964e-01
-4.34070766e-01 -5.45568049e-01 -7.96951056e-01 -5.85112810e-01
-1.30975068e-01 -1.09225191e-01 6.64136350e-01 9.38601434e-01
4.70848680e-01 7.81868845e-02 9.55694079e-01 -1.28437817e+00
-5.45143783e-01 -9.99889970e-01 -4.49802339e-01 1.36496708e-01
3.36873531e-01 -7.97208309e-01 -1.20632529e-01 1.17285907e-01] | [13.285872459411621, 0.7376279830932617] |
a38be23d-cfd5-4591-828c-8711a9021872 | a-comparative-study-of-semi-and-self | 2011.08076 | null | https://arxiv.org/abs/2011.08076v2 | https://arxiv.org/pdf/2011.08076v2.pdf | A comparative study of semi- and self-supervised semantic segmentation of biomedical microscopy data | In recent years, Convolutional Neural Networks (CNNs) have become the state-of-the-art method for biomedical image analysis. However, these networks are usually trained in a supervised manner, requiring large amounts of labelled training data. These labelled data sets are often difficult to acquire in the biomedical domain. In this work, we validate alternative ways to train CNNs with fewer labels for biomedical image segmentation using. We adapt two semi- and self-supervised image classification methods and analyse their performance for semantic segmentation of biomedical microscopy images. | ['Nico Scherf', 'Ingo Roeder', 'Sebastian Wagner', 'Sebastian Niehaus', 'Alisa Mironenko', 'Nastassya Horlava'] | 2020-11-11 | null | null | null | null | ['self-supervised-image-classification'] | ['computer-vision'] | [ 5.42489111e-01 2.86334038e-01 4.45765555e-02 -7.01943457e-01
-3.31720173e-01 -4.12987322e-01 2.54711956e-01 4.88042891e-01
-1.19017112e+00 9.63090420e-01 -6.04914606e-01 -4.57777202e-01
2.33259946e-01 -7.71427393e-01 -5.87190390e-01 -7.88322926e-01
3.20705980e-01 8.90655398e-01 3.01219523e-01 1.05918065e-01
3.69764306e-02 7.98926294e-01 -1.38306201e+00 3.44603300e-01
7.96469808e-01 1.04127800e+00 2.48765573e-01 6.05201304e-01
-4.02580470e-01 4.76968020e-01 -6.54458523e-01 5.10623977e-02
2.53794957e-02 -4.99626011e-01 -1.20952547e+00 1.58092812e-01
-3.04540861e-02 1.93897393e-02 1.19210683e-01 1.12849557e+00
5.78477442e-01 -2.62403637e-01 7.44237661e-01 -7.92578101e-01
-1.66069523e-01 2.73854136e-01 -9.41708684e-02 4.00968611e-01
-1.48097113e-01 2.69901101e-03 3.02203685e-01 -4.03262377e-01
7.78769910e-01 7.66078115e-01 7.39090800e-01 7.87673116e-01
-1.47445297e+00 -5.29789031e-01 -4.22182649e-01 3.73018421e-02
-1.19986129e+00 -4.33275402e-01 7.05598354e-01 -6.13720298e-01
6.75661325e-01 9.87217575e-03 5.83436251e-01 9.78982389e-01
5.35042062e-02 6.41237617e-01 1.42975700e+00 -5.35817802e-01
4.58775073e-01 2.51630515e-01 1.72289297e-01 6.40414774e-01
3.06519836e-01 -1.02286570e-01 4.55794507e-04 7.06670433e-02
9.54705954e-01 5.28675616e-02 -1.05068974e-01 -2.09998697e-01
-1.25750780e+00 9.37283456e-01 4.48256254e-01 9.51843917e-01
-2.51855820e-01 -8.83851424e-02 6.35994792e-01 1.62135020e-01
8.27880442e-01 6.99715853e-01 -7.16761827e-01 8.44692662e-02
-1.18654680e+00 -1.32566422e-01 6.99386179e-01 5.78999162e-01
6.80806041e-01 -1.55603319e-01 -4.46720570e-02 9.55734670e-01
2.35434324e-01 4.09566425e-02 7.55620897e-01 -5.74821413e-01
-2.72027671e-01 8.27241838e-01 -2.27745444e-01 -6.50893092e-01
-9.31858957e-01 -4.25679892e-01 -1.24449134e+00 2.00331897e-01
9.55909014e-01 -1.04029529e-01 -1.11606371e+00 1.23705149e+00
1.99134961e-01 -2.85628468e-01 1.24383382e-01 7.90154040e-01
1.01912856e+00 3.27175498e-01 2.66866326e-01 -2.45186523e-01
1.07290936e+00 -8.54959548e-01 -8.40167880e-01 -2.43437905e-02
1.03491879e+00 -7.03378737e-01 5.07480204e-01 4.55550820e-01
-7.34224379e-01 -5.57706952e-01 -7.65903056e-01 3.08537874e-02
-8.02873969e-01 2.13230044e-01 5.55112720e-01 6.88308716e-01
-1.05351663e+00 9.35469449e-01 -8.84367228e-01 -5.68932772e-01
1.16344702e+00 7.51572371e-01 -6.23978198e-01 9.64697897e-02
-9.59709704e-01 7.62899458e-01 7.42873609e-01 2.59874612e-01
-7.96931386e-01 -3.21845025e-01 -6.88337028e-01 -1.16290301e-01
-3.55848819e-02 -3.22243661e-01 1.12253022e+00 -1.09228468e+00
-1.47680354e+00 1.73959947e+00 1.63673386e-01 -5.37335634e-01
5.80650687e-01 2.17966214e-01 -4.77933325e-02 4.89229828e-01
-1.15207106e-01 9.58195090e-01 6.41425252e-01 -1.22992718e+00
-2.14216620e-01 -3.76422822e-01 -2.94863105e-01 -4.57073480e-01
-1.92848429e-01 1.04425766e-01 1.31489158e-01 -4.88776088e-01
1.08181655e-01 -7.24427402e-01 -5.66632152e-01 7.33165666e-02
-6.27468228e-01 -3.32049608e-01 8.49031508e-01 -4.30820882e-01
5.05173087e-01 -2.04670691e+00 9.47667733e-02 1.40809089e-01
3.90000135e-01 7.86451578e-01 1.33109644e-01 1.64957978e-02
-1.53700367e-01 2.30520844e-01 -5.16852140e-01 -4.15300608e-01
-3.00941050e-01 4.84983653e-01 4.99790609e-01 6.23547316e-01
1.96143687e-01 9.69152927e-01 -9.31594491e-01 -1.03073931e+00
4.46097821e-01 4.10442770e-01 -1.74816683e-01 4.17453259e-01
-2.80057847e-01 1.26469874e+00 -3.55306417e-01 5.59337318e-01
4.81222093e-01 -5.54498017e-01 3.02536219e-01 -2.53252059e-01
1.09626316e-02 -2.29328349e-01 -4.59568352e-01 1.80836415e+00
-8.72950554e-02 7.27195740e-01 4.51193117e-02 -1.89178944e+00
9.04961288e-01 4.41617787e-01 7.36334741e-01 -5.02200603e-01
6.60349548e-01 4.11453068e-01 7.83024635e-03 -6.95262611e-01
-1.96552053e-01 -5.04828453e-01 2.50417084e-01 1.33941948e-01
6.01871908e-01 -9.92916003e-02 4.54004347e-01 -4.25654233e-01
1.01787329e+00 7.16345832e-02 2.09242508e-01 -5.09708822e-01
8.32074285e-01 3.05466294e-01 3.94523621e-01 4.83690768e-01
-3.82217467e-01 7.80123770e-01 4.06376243e-01 -8.62679541e-01
-1.11802757e+00 -5.86447537e-01 -5.92772663e-01 6.16045654e-01
-2.47641742e-01 1.31387115e-01 -1.26969087e+00 -8.58091116e-01
-3.45358878e-01 -7.67812580e-02 -6.61822259e-01 1.88453004e-01
-4.33409661e-01 -9.57427263e-01 6.83306932e-01 3.57017308e-01
5.62911332e-01 -1.44147778e+00 -7.89870024e-01 4.87782598e-01
-1.51578961e-02 -1.32517576e+00 2.64189601e-01 7.53928661e-01
-1.10296679e+00 -1.30938160e+00 -1.16418171e+00 -1.15993440e+00
1.25487196e+00 -2.45848626e-01 1.11894822e+00 4.41353977e-01
-6.78374887e-01 -6.82893768e-02 -4.16684389e-01 -5.65926015e-01
-6.18543565e-01 3.78587663e-01 -2.22038627e-01 6.97299019e-02
4.12089586e-01 -4.00347173e-01 -6.79009199e-01 3.34574372e-01
-1.19234264e+00 4.10443135e-02 7.89068282e-01 1.01573086e+00
8.75143588e-01 6.84333295e-02 5.24222791e-01 -1.51521361e+00
3.50484014e-01 -1.07278980e-01 -4.06359166e-01 -1.22832274e-02
-4.23813224e-01 -1.57060415e-01 8.24092984e-01 -3.67780060e-01
-6.56943202e-01 3.81647468e-01 -6.99857056e-01 -1.45922661e-01
-1.17492831e+00 5.86804867e-01 1.56868830e-01 -5.22731841e-01
7.12898672e-01 -1.33641940e-02 3.32261980e-01 -5.90403914e-01
-1.05415076e-01 7.93938518e-01 3.46459270e-01 -3.12097400e-01
4.28988874e-01 6.23268962e-01 2.47450218e-01 -9.09339666e-01
-1.00669944e+00 -7.35169232e-01 -1.17266846e+00 -2.22212389e-01
1.41364229e+00 -3.84883195e-01 -5.13256907e-01 8.85968328e-01
-1.14812505e+00 -6.83643639e-01 -2.67702341e-01 2.89215922e-01
-4.44532752e-01 3.84246320e-01 -7.06204891e-01 -4.50524092e-01
-3.60174507e-01 -1.19648373e+00 1.02286589e+00 3.44971120e-01
-2.68215567e-01 -1.22966528e+00 7.83282220e-02 4.01876718e-01
4.15895969e-01 5.91377676e-01 8.99687111e-01 -1.05026388e+00
5.76108508e-02 -2.84961104e-01 -4.02295560e-01 5.96587300e-01
2.38781616e-01 -2.91795880e-01 -1.16520500e+00 -3.09190691e-01
-8.54809210e-02 -4.78916436e-01 8.37460279e-01 5.18578470e-01
1.58525312e+00 7.64682889e-02 -5.66019118e-01 5.57146728e-01
1.39465964e+00 3.24306071e-01 6.87567115e-01 4.19722944e-01
5.86462855e-01 7.89480925e-01 1.01478115e-01 -1.41767502e-01
-1.71419263e-01 3.11303943e-01 2.84032524e-01 -7.63349473e-01
1.68785885e-01 3.92531723e-01 -5.79441369e-01 8.20156932e-01
-1.45895734e-01 -7.22192824e-02 -1.18617320e+00 6.16008222e-01
-1.73386157e+00 -2.72073388e-01 -2.88497090e-01 1.73047578e+00
1.14352524e+00 2.91566908e-01 5.75430095e-02 5.67910433e-01
6.92775428e-01 -3.88537198e-01 -2.65454471e-01 -1.08720303e-01
-4.22315188e-02 6.95002675e-01 5.44278920e-01 -5.85360602e-02
-1.40210438e+00 8.70832205e-01 6.84439182e+00 6.41311347e-01
-1.25834858e+00 2.51810580e-01 1.07756436e+00 5.09409964e-01
5.27018964e-01 -4.36934084e-01 -3.26649129e-01 5.29567540e-01
1.03419828e+00 7.54333436e-01 -7.92295560e-02 6.62875772e-01
-5.03737619e-03 -4.04021651e-01 -1.16822100e+00 9.48015571e-01
-2.74834484e-01 -1.41738522e+00 -2.13645682e-01 1.28019243e-01
7.45095372e-01 -1.42821178e-01 -2.52453834e-01 -4.41441052e-02
1.45046934e-02 -1.33567071e+00 1.21184282e-01 4.30101991e-01
8.84851098e-01 -5.94708741e-01 1.42899978e+00 3.24058950e-01
-4.94093895e-01 3.24562788e-01 -4.37858313e-01 1.08114563e-01
7.95511156e-03 8.70604217e-01 -4.89521295e-01 3.31137627e-01
9.21111643e-01 4.34519053e-01 -6.52181625e-01 1.16473413e+00
2.61953920e-02 7.76536405e-01 -1.75476611e-01 -2.02361848e-02
4.93111014e-01 -1.06947213e-01 1.89444534e-02 1.34168351e+00
-1.22205250e-01 -1.33163676e-01 2.55169809e-01 7.64056444e-01
-1.80135250e-01 1.06112443e-01 -3.61262053e-01 -5.34239352e-01
-2.33866245e-01 1.52073109e+00 -1.68756759e+00 -3.94982785e-01
-1.21869698e-01 7.27741122e-01 3.26253623e-01 1.83023483e-01
-3.99703830e-01 -5.04405260e-01 2.66519524e-02 3.25182267e-02
3.11977509e-02 -4.34418172e-02 -3.32996100e-01 -7.78248370e-01
-5.54405510e-01 -5.49996257e-01 1.92824110e-01 -4.91185158e-01
-1.43430865e+00 6.10515237e-01 -1.76542968e-01 -9.55570698e-01
5.41098863e-02 -1.16836882e+00 -3.64840120e-01 5.37322283e-01
-1.60785687e+00 -9.49663699e-01 -3.83109808e-01 3.82554084e-01
2.64004141e-01 -2.66217142e-01 1.21532643e+00 3.60368460e-01
-4.07512635e-01 7.11075813e-02 1.92923963e-01 5.81431627e-01
5.61052620e-01 -1.32365215e+00 -1.30074471e-01 1.67953849e-01
-5.01567200e-02 5.00673950e-01 5.04697025e-01 -1.69444904e-01
-6.93582177e-01 -9.69789565e-01 8.35717678e-01 1.63736448e-01
3.53813857e-01 -2.80933440e-01 -9.71234500e-01 4.11265790e-01
2.69885987e-01 3.90668869e-01 1.10321558e+00 -1.48678496e-01
1.39251858e-01 5.46620153e-02 -1.54288650e+00 1.87437594e-01
5.29652834e-01 -3.85027379e-01 -3.28946501e-01 7.61350870e-01
2.70517141e-01 -3.55988353e-01 -1.10903192e+00 4.88180995e-01
2.23615438e-01 -8.73648822e-01 8.68641913e-01 -6.05920017e-01
5.01557589e-01 -1.36143938e-01 4.11691397e-01 -1.22198284e+00
-3.34501565e-02 -3.29903752e-01 2.87376463e-01 8.90181959e-01
3.05849254e-01 -8.01098287e-01 1.13249254e+00 5.66941239e-02
-1.04344457e-01 -9.06801283e-01 -9.98977959e-01 -7.22538888e-01
3.30853879e-01 9.78266150e-02 1.75222501e-01 1.06619573e+00
-1.18788272e-01 2.24075124e-01 2.56289274e-01 -4.83595788e-01
6.43432200e-01 -1.64921865e-01 3.42236757e-01 -1.70554864e+00
1.68223694e-01 -6.49151385e-01 -7.92269289e-01 -4.17016566e-01
3.46663028e-01 -8.73085022e-01 2.34697655e-01 -1.68445659e+00
1.31015601e-02 -4.61393416e-01 -5.29570520e-01 5.97344220e-01
2.42809474e-01 8.08649600e-01 -3.68861943e-01 3.35871011e-01
-6.60899997e-01 -1.05980247e-01 1.19640887e+00 -3.40520293e-01
7.19237477e-02 -1.33128002e-01 -1.83241025e-01 8.05660665e-01
1.36221707e+00 -6.77149951e-01 -4.96164151e-02 -1.53285444e-01
-6.57715872e-02 -3.91807050e-01 4.29096848e-01 -1.15326333e+00
1.92139938e-01 1.39448375e-01 7.54927337e-01 -4.94911969e-01
-1.87913433e-01 -9.06897604e-01 3.97856161e-03 5.67664623e-01
-4.54996526e-01 -4.54311013e-01 2.06833705e-01 2.43791118e-01
-3.66239280e-01 -6.30475640e-01 1.29370105e+00 -7.59116650e-01
-2.47752130e-01 1.99985400e-01 -7.65188694e-01 -9.62538347e-02
9.76682186e-01 -3.56607020e-01 -6.90447763e-02 1.00516833e-01
-1.08769047e+00 -1.19432837e-01 4.61135477e-01 -1.58144638e-01
4.41235691e-01 -1.05158973e+00 -2.81046212e-01 1.57784577e-02
1.58631891e-01 5.61667204e-01 1.94519073e-01 1.09945858e+00
-1.12957180e+00 6.68075562e-01 -6.39013946e-01 -8.61525953e-01
-1.14569747e+00 6.01139724e-01 4.05716360e-01 -5.85343659e-01
-4.34958071e-01 7.26074934e-01 -4.72966470e-02 -7.59831190e-01
7.39606768e-02 -2.83475518e-01 -6.68052554e-01 -5.76250535e-03
1.90251455e-01 3.33689526e-02 2.53576726e-01 -6.95062876e-01
-2.49680340e-01 3.44057977e-01 -5.92646338e-02 2.59021550e-01
1.53424072e+00 1.37583286e-01 -5.43236494e-01 4.82893050e-01
1.41748714e+00 -8.52474570e-01 -8.63592446e-01 -5.28742820e-02
3.19215834e-01 4.95672375e-02 1.12815626e-01 -6.59931719e-01
-1.25442588e+00 1.09675324e+00 7.03699172e-01 5.70216835e-01
9.44185019e-01 -1.61864236e-02 7.76919782e-01 4.99832481e-01
4.78145093e-01 -1.30097365e+00 -1.70876607e-01 1.95481211e-01
3.54993463e-01 -1.45602620e+00 -1.56213015e-01 -5.09468555e-01
-3.06240488e-02 1.37304640e+00 3.91023070e-01 -1.28587484e-02
7.43264079e-01 1.71103463e-01 3.92657667e-01 -4.64272290e-01
3.83460782e-02 -4.47871238e-01 1.29556358e-01 8.27554941e-01
7.56961584e-01 -1.15705051e-01 -3.18273902e-01 2.81485885e-01
1.31296247e-01 2.67606795e-01 2.72724569e-01 1.21501863e+00
-2.58224368e-01 -1.22390389e+00 -2.86388457e-01 6.74948990e-01
-1.03486907e+00 9.85933691e-02 -4.80355144e-01 6.22025669e-01
3.59191895e-01 1.02456653e+00 -9.79238525e-02 5.18488474e-02
2.90575996e-02 1.67126730e-01 6.06834829e-01 -8.19369435e-01
-7.36653924e-01 1.26709506e-01 -1.46372035e-01 -2.07957461e-01
-1.01224947e+00 -2.74660468e-01 -1.33821547e+00 1.27725065e-01
-2.36491516e-01 1.96438670e-01 7.82848418e-01 1.17847192e+00
3.43958102e-02 8.90245676e-01 2.85584867e-01 -1.12005103e+00
1.22090779e-01 -1.21193337e+00 -6.54599309e-01 6.55741394e-01
1.10690385e-01 -5.81158578e-01 -2.53761023e-01 5.46879709e-01] | [14.461690902709961, -2.874837875366211] |
5320f2c6-6d75-4cb3-b679-ff66c7688545 | machine-translation-with-weakly-paired | null | null | https://openreview.net/forum?id=ryza73R9tQ | https://openreview.net/pdf?id=ryza73R9tQ | Machine Translation With Weakly Paired Bilingual Documents | Neural machine translation, which achieves near human-level performance in some languages, strongly relies on the availability of large amounts of parallel sentences, which hinders its applicability to low-resource language pairs. Recent works explore the possibility of unsupervised machine translation with monolingual data only, leading to much lower accuracy compared with the supervised one. Observing that weakly paired bilingual documents are much easier to collect than bilingual sentences, e.g., from Wikipedia, news websites or books, in this paper, we investigate the training of translation models with weakly paired bilingual documents. Our approach contains two components/steps. First, we provide a simple approach to mine implicitly bilingual sentence pairs from document pairs which can then be used as supervised signals for training. Second, we leverage the topic consistency of two weakly paired documents and learn the sentence-to-sentence translation by constraining the word distribution-level alignments. We evaluate our proposed method on weakly paired documents from Wikipedia on four tasks, the widely used WMT16 German$\leftrightarrow$English and WMT13 Spanish$\leftrightarrow$English tasks, and obtain $24.1$/$30.3$ and $28.0$/$27.6$ BLEU points separately, outperforming
state-of-the-art unsupervised results by more than 5 BLEU points and reducing the gap between unsupervised translation and supervised translation up to 50\%. | ['Tie-Yan Liu', 'Jinhua Zhu', 'Fei Gao', 'Di He', 'Xu Tan', 'Tao Qin', 'Lijun Wu'] | null | null | null | null | iclr-2019-5 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 9.65443179e-02 -1.70088708e-01 -5.71160853e-01 -3.99633825e-01
-1.44527876e+00 -7.79954672e-01 8.25355291e-01 1.01032704e-01
-6.69767678e-01 1.25503588e+00 1.31648421e-01 -5.93153059e-01
2.18859389e-01 -6.70679748e-01 -9.89358664e-01 -7.31566608e-01
3.41602087e-01 7.30500460e-01 -1.64356694e-01 -5.00540376e-01
1.23875909e-01 -1.25094354e-01 -9.79404271e-01 3.08666587e-01
1.25038505e+00 5.00079036e-01 3.98883134e-01 5.57175465e-02
-4.27512735e-01 1.16867870e-01 -3.10745329e-01 -7.83798099e-01
3.82563412e-01 -7.47502506e-01 -7.02948749e-01 -1.53181806e-01
3.34325373e-01 8.91152248e-02 1.07911170e-01 1.24562919e+00
4.47579890e-01 -2.41397545e-01 5.86718619e-01 -6.82633579e-01
-5.86632252e-01 1.07914245e+00 -7.75837600e-01 1.20024845e-01
3.09242129e-01 1.79565381e-02 1.17135823e+00 -1.20448232e+00
8.22292209e-01 8.96425784e-01 3.57775718e-01 4.10925359e-01
-1.28472877e+00 -8.71477604e-01 1.99922342e-02 -1.12145878e-01
-1.38362515e+00 -5.50914884e-01 6.31907463e-01 -3.80584061e-01
1.17837501e+00 1.40388653e-01 3.35014045e-01 1.31348586e+00
1.38550609e-01 5.56909800e-01 1.58718097e+00 -7.76459813e-01
-1.58825070e-01 3.71980846e-01 2.85051372e-02 5.02463996e-01
2.10485086e-01 4.42901216e-02 -7.86956787e-01 -8.61008000e-03
3.47484857e-01 -3.13279897e-01 -2.32628599e-01 -9.84470025e-02
-1.65313447e+00 6.71602368e-01 2.23190621e-01 6.67850912e-01
-1.58019617e-01 -3.36473852e-01 3.99026036e-01 5.78501165e-01
7.15528548e-01 2.88580179e-01 -5.41757464e-01 -1.46316603e-01
-1.01277459e+00 7.65268365e-03 6.36302948e-01 1.27507961e+00
1.10254657e+00 -2.26263478e-01 -6.02181070e-03 1.10426104e+00
1.25358254e-01 9.86138821e-01 5.89742243e-01 -3.03827882e-01
1.17799270e+00 3.25894266e-01 1.18450366e-01 -7.98782885e-01
2.08808202e-02 -5.23478866e-01 -1.04435802e+00 -6.36730373e-01
4.55873251e-01 -6.07084036e-02 -5.97287059e-01 1.94236314e+00
1.58668086e-02 -3.89900714e-01 3.42842281e-01 7.85425484e-01
4.05227780e-01 7.69206583e-01 -2.05604389e-01 -6.17302179e-01
1.30803978e+00 -1.17702889e+00 -5.15724301e-01 -5.55702925e-01
7.94504046e-01 -1.36701441e+00 1.31600237e+00 1.58993676e-01
-1.08384621e+00 -5.43192148e-01 -8.65607560e-01 4.69483212e-02
-3.00275952e-01 4.39537525e-01 4.03029710e-01 4.28564399e-01
-8.75002742e-01 4.46091324e-01 -7.93759286e-01 -6.01882577e-01
1.15671419e-01 3.53210777e-01 -3.90390605e-01 -4.21096683e-01
-1.34541810e+00 1.02864921e+00 3.09389174e-01 1.17702268e-01
-6.36194825e-01 -3.41702461e-01 -7.10741520e-01 -2.73745358e-01
3.06631207e-01 -5.37631750e-01 9.58929658e-01 -9.83629286e-01
-1.51597631e+00 1.09723437e+00 -4.21697348e-01 -3.35073262e-01
5.92020333e-01 -1.69538990e-01 -3.56127083e-01 -1.43212661e-01
4.41144288e-01 6.06784880e-01 5.36912441e-01 -1.01719761e+00
-5.26997983e-01 -4.17667389e-01 -2.63874322e-01 3.18685800e-01
-5.20401061e-01 1.94116250e-01 -6.23439074e-01 -7.21625149e-01
2.58141249e-01 -1.14977765e+00 -1.12910196e-01 -6.92735672e-01
-4.58959430e-01 -1.44808561e-01 1.57846376e-01 -8.59055161e-01
9.77105379e-01 -1.90756214e+00 3.37484032e-01 8.00057575e-02
-1.90247789e-01 -2.47679055e-02 -3.58517081e-01 6.16198599e-01
2.64289021e-01 6.32047281e-02 -5.24856389e-01 -3.76729161e-01
4.60671932e-02 2.44043767e-01 -2.51584321e-01 3.39818358e-01
1.20507009e-01 8.68013501e-01 -8.88726771e-01 -4.83173907e-01
-3.16195250e-01 7.80868679e-02 -1.98887989e-01 1.86240114e-02
-1.04552738e-01 6.96526647e-01 -2.18282357e-01 5.80991924e-01
6.57688498e-01 2.39843391e-02 6.56907022e-01 6.61886781e-02
-2.41211787e-01 8.56121421e-01 -6.15853846e-01 2.08418965e+00
-8.01983416e-01 5.29968083e-01 -7.12089939e-03 -1.15125585e+00
1.12196636e+00 3.17037493e-01 2.23662481e-01 -8.96553636e-01
4.30301614e-02 8.11710119e-01 2.02955723e-01 -2.52740771e-01
3.36624950e-01 -4.59866375e-01 -3.23176235e-01 6.77643239e-01
2.13605165e-01 -1.78121597e-01 6.46972060e-01 -8.85362998e-02
7.16692924e-01 2.38460630e-01 2.87202030e-01 -6.83092237e-01
5.40733159e-01 6.82794163e-03 6.84498906e-01 5.18561304e-01
2.16885522e-01 4.70062196e-01 2.81072259e-01 -9.19642672e-02
-1.18818557e+00 -1.13208055e+00 -1.58245236e-01 9.74831223e-01
6.50555789e-02 -4.27956939e-01 -8.34935248e-01 -8.32991540e-01
-3.38820279e-01 6.37404799e-01 -1.42402753e-01 1.18035711e-01
-8.89724791e-01 -1.08397925e+00 5.38059652e-01 3.24620634e-01
4.98851627e-01 -8.31358790e-01 1.40563712e-01 8.24125484e-02
-7.21741557e-01 -1.21125555e+00 -6.38949454e-01 3.24784845e-01
-1.03405499e+00 -6.22471571e-01 -8.95589471e-01 -1.02660763e+00
8.71245682e-01 2.29870498e-01 1.39195907e+00 -1.64411396e-01
3.50684583e-01 -3.04820240e-01 -3.08746636e-01 -6.20942228e-02
-5.63422263e-01 5.96157670e-01 4.41866249e-01 -1.98677704e-01
7.89655209e-01 -6.36402011e-01 -2.47672126e-01 4.41694558e-01
-5.98749220e-01 2.28187054e-01 1.06055713e+00 1.20622742e+00
7.09777594e-01 -3.37984592e-01 4.87300396e-01 -8.51904392e-01
5.17901838e-01 -3.73949677e-01 -5.11230946e-01 3.94211888e-01
-8.00348759e-01 2.60697514e-01 8.10894787e-01 -3.30767006e-01
-8.74311388e-01 -2.00479776e-01 -1.06459931e-01 9.11357403e-02
-1.95847219e-03 6.89250410e-01 -2.39238724e-01 3.40130985e-01
7.21436024e-01 4.89803165e-01 -1.36722520e-01 -5.37475586e-01
1.92649782e-01 9.09696639e-01 2.57894784e-01 -9.14927900e-01
1.00948679e+00 1.08586051e-01 -4.55875874e-01 -5.45843184e-01
-6.29319727e-01 -3.74612212e-01 -8.43159556e-01 2.41644531e-01
5.58530569e-01 -1.03805017e+00 7.63834571e-04 1.56657875e-01
-1.21017647e+00 -3.17741632e-01 4.04530503e-02 8.18112552e-01
-4.71591562e-01 2.75840729e-01 -7.29689121e-01 -3.22351664e-01
-4.34978276e-01 -1.25836360e+00 9.80279028e-01 -2.13383168e-01
-2.58084118e-01 -8.25690150e-01 2.05705181e-01 5.49995303e-01
2.79124081e-01 -2.79761702e-01 1.11984909e+00 -7.08641231e-01
-4.64385122e-01 1.65436394e-03 -1.86374173e-01 5.13588786e-01
2.62287319e-01 -4.74974185e-01 -6.17428839e-01 -4.86244857e-01
9.56130959e-03 -3.57362479e-01 7.25879848e-01 -6.75093755e-02
4.00852412e-01 -2.82889724e-01 -3.00571710e-01 3.38812470e-01
1.35842288e+00 -7.46835694e-02 3.90663654e-01 2.30039150e-01
4.78002608e-01 7.62034118e-01 6.59964561e-01 -4.06366251e-02
3.30888808e-01 8.46363485e-01 -2.54614770e-01 -2.14598724e-03
-6.82251081e-02 -3.87248099e-01 7.52932072e-01 1.84922385e+00
-1.29772946e-01 -1.37502164e-01 -1.09976339e+00 7.21112072e-01
-1.72374272e+00 -5.73918819e-01 -3.93393412e-02 2.36958408e+00
1.34985888e+00 3.19508344e-01 -8.79608393e-02 -2.62806028e-01
7.21984863e-01 -7.31537957e-03 -8.94502103e-02 -3.80594313e-01
-3.63253564e-01 4.45758730e-01 4.55235630e-01 6.46789014e-01
-7.33951926e-01 1.22200871e+00 5.08061981e+00 1.02191472e+00
-1.10873330e+00 4.26481843e-01 5.39419532e-01 -2.32718483e-01
-5.57533205e-01 1.58168778e-01 -8.94364119e-01 5.97372830e-01
9.70805705e-01 -1.69709504e-01 5.54836035e-01 4.42563593e-01
2.52812237e-01 -3.06272879e-02 -1.27445984e+00 9.00613904e-01
1.68614611e-01 -1.08241773e+00 9.90779474e-02 1.29235297e-01
1.12164259e+00 3.04747015e-01 8.62844139e-02 3.41887027e-01
2.94820249e-01 -8.54003549e-01 6.34493589e-01 7.07490072e-02
1.02721810e+00 -6.64042592e-01 9.25896466e-01 6.43648088e-01
-8.77890468e-01 3.84626031e-01 -4.37063605e-01 -5.90749122e-02
1.34239301e-01 8.41251135e-01 -6.64369583e-01 8.68096471e-01
6.78556204e-01 7.67706275e-01 -4.46133882e-01 2.61695743e-01
-5.04426956e-01 6.39692783e-01 -4.61001009e-01 -1.80712268e-01
3.36587697e-01 -6.78778172e-01 4.08357590e-01 1.37056530e+00
7.09639966e-01 -3.14711481e-01 9.91758406e-02 7.24995136e-01
-4.66442972e-01 6.13255441e-01 -7.81184673e-01 -1.32003322e-01
3.37480128e-01 9.32067454e-01 -6.35167897e-01 -3.04519773e-01
-6.21182561e-01 1.22413313e+00 5.30434906e-01 4.14241433e-01
-6.74947917e-01 -2.42933601e-01 3.67698461e-01 -4.26809257e-03
7.95684382e-02 -4.28069115e-01 -3.54568064e-01 -1.53405976e+00
5.61390758e-01 -1.16918266e+00 -7.19625950e-02 -3.00919920e-01
-1.43767059e+00 9.42047536e-01 -1.27248392e-01 -1.41959333e+00
-5.40325582e-01 -4.11972433e-01 -2.83482671e-01 1.23942399e+00
-1.50605190e+00 -1.18525589e+00 1.94493726e-01 3.34574282e-01
6.17663443e-01 -3.99846733e-01 8.92681658e-01 4.67700362e-01
-4.79308695e-01 8.07826340e-01 5.17793119e-01 1.63607016e-01
1.16596830e+00 -1.02223253e+00 4.49441791e-01 9.52978313e-01
6.54921830e-01 9.88213122e-01 5.75992048e-01 -5.45425177e-01
-1.39085007e+00 -8.40728879e-01 1.69629204e+00 -5.50995648e-01
6.99529529e-01 -8.05164576e-01 -5.88275313e-01 5.22181392e-01
6.40689015e-01 -3.33453536e-01 6.93251789e-01 3.02632660e-01
-4.04840142e-01 -3.08518767e-01 -7.73030043e-01 8.78212512e-01
1.21241128e+00 -7.78714657e-01 -6.01171672e-01 4.83163089e-01
5.09808123e-01 -2.51289487e-01 -7.28236079e-01 3.84387791e-01
4.19237882e-01 -7.19920278e-01 7.28206277e-01 -4.32139218e-01
7.08816767e-01 -1.47162661e-01 -4.16529000e-01 -1.52936399e+00
5.07542603e-02 -5.68536937e-01 3.50524127e-01 1.35930753e+00
1.04568672e+00 -6.17523432e-01 5.93067586e-01 8.66573527e-02
-1.60165459e-01 -6.70308352e-01 -1.05625665e+00 -9.75746870e-01
7.06937432e-01 -2.26309121e-01 4.14329976e-01 1.06615341e+00
1.54501393e-01 7.10010827e-01 -2.95212775e-01 -2.18374893e-01
4.90174264e-01 4.89195585e-01 6.98718965e-01 -7.48894155e-01
-4.54294413e-01 -4.25737023e-01 1.79238468e-01 -1.40240431e+00
4.52894837e-01 -1.31420350e+00 3.02177757e-01 -1.22015643e+00
4.45021778e-01 -4.93185908e-01 -1.57867163e-01 3.30160648e-01
-1.29020497e-01 6.25883698e-01 -6.33768067e-02 4.45074826e-01
-2.41187975e-01 5.72197735e-01 1.04688632e+00 -3.93194407e-01
-9.68359262e-02 -2.29713142e-01 -6.47080362e-01 4.79804546e-01
8.50423574e-01 -6.53585434e-01 -1.16425201e-01 -9.57881033e-01
4.30595189e-01 -1.34165972e-01 -1.55390635e-01 -5.92267931e-01
6.74149618e-02 -1.34939656e-01 -2.65485235e-02 -3.43198538e-01
-4.07040007e-02 -6.26582146e-01 -1.37042984e-01 3.04019809e-01
-3.01944196e-01 4.05513197e-01 1.27004296e-01 3.30525190e-01
-5.97879410e-01 -2.27622375e-01 5.57758451e-01 -2.22404718e-01
-1.73025236e-01 3.08424961e-02 -2.22909704e-01 1.64106965e-01
6.01696908e-01 1.45402208e-01 -1.90734014e-01 -1.05208077e-01
-3.93265128e-01 2.44242083e-02 5.88714063e-01 5.25401890e-01
2.27321312e-01 -1.44969881e+00 -9.91906524e-01 2.54387349e-01
3.17148775e-01 -1.36267871e-01 -1.24569476e-01 1.28646374e+00
-2.53010154e-01 6.78460479e-01 -3.77122909e-02 -7.67047107e-01
-9.55288112e-01 3.82747352e-01 -1.42859280e-01 -5.10442972e-01
-2.64028847e-01 6.30709887e-01 -9.86035168e-03 -8.22851658e-01
-1.55482665e-01 -1.11988887e-01 3.24846983e-01 -1.04404703e-01
7.54189342e-02 1.88825075e-02 3.60083640e-01 -7.90594399e-01
-3.41129631e-01 7.37439692e-01 -3.18028122e-01 -4.97560203e-01
1.14009202e+00 -2.79663950e-01 -5.53703845e-01 4.66376543e-01
1.30783546e+00 3.89251292e-01 -5.91071963e-01 -5.07917464e-01
2.70507067e-01 -4.18741345e-01 -4.47657496e-01 -8.33718479e-01
-7.10856676e-01 1.02754748e+00 2.71755755e-01 -2.29306892e-01
9.09321547e-01 1.14167064e-01 1.02639985e+00 5.24900496e-01
6.88225150e-01 -1.21594512e+00 -2.09060535e-01 6.68791413e-01
6.26152515e-01 -1.50413704e+00 -1.83404297e-01 -4.01884764e-01
-3.90158474e-01 8.18760157e-01 3.31582546e-01 1.87145278e-01
3.09608698e-01 1.78419784e-01 2.40110055e-01 2.12457776e-01
-5.83754778e-01 -7.25700632e-02 3.70754093e-01 3.56566191e-01
8.62299681e-01 1.33551806e-01 -1.01051581e+00 5.95700562e-01
-4.08428788e-01 -3.00835103e-01 3.14828455e-02 6.82372749e-01
-3.72607298e-02 -1.79874074e+00 -2.28320211e-01 2.91524023e-01
-4.80673641e-01 -6.26912236e-01 -4.77830946e-01 7.00834155e-01
8.33807811e-02 9.47993577e-01 -8.43524933e-02 -2.62083679e-01
1.47396838e-02 3.69218796e-01 6.34002924e-01 -6.71181321e-01
-5.78748047e-01 3.13453257e-01 1.92138776e-01 -1.76317811e-01
-5.53791642e-01 -7.80760586e-01 -8.90739799e-01 -3.24420154e-01
-2.21042603e-01 5.00589371e-01 6.07910573e-01 1.12781656e+00
2.81753659e-01 -5.27100563e-02 6.22103393e-01 -5.53408682e-01
-6.34995878e-01 -1.32576942e+00 -2.30275840e-01 3.14347118e-01
-8.09269920e-02 -3.01607311e-01 -2.74785250e-01 3.67579684e-02] | [11.596829414367676, 10.359140396118164] |
bac1693b-578b-4676-81b6-3298669e839d | understanding-a-class-of-decentralized-and | 2204.12663 | null | https://arxiv.org/abs/2204.12663v2 | https://arxiv.org/pdf/2204.12663v2.pdf | Understanding A Class of Decentralized and Federated Optimization Algorithms: A Multi-Rate Feedback Control Perspective | Distributed algorithms have been playing an increasingly important role in many applications such as machine learning, signal processing, and control. Significant research efforts have been devoted to developing and analyzing new algorithms for various applications. In this work, we provide a fresh perspective to understand, analyze, and design distributed optimization algorithms. Through the lens of multi-rate feedback control, we show that a wide class of distributed algorithms, including popular decentralized/federated schemes, can be viewed as discretizing a certain continuous-time feedback control system, possibly with multiple sampling rates, such as decentralized gradient descent, gradient tracking, and federated averaging. This key observation not only allows us to develop a generic framework to analyze the convergence of the entire algorithm class. More importantly, it also leads to an interesting way of designing new distributed algorithms. We develop the theory behind our framework and provide examples to highlight how the framework can be used in practice. | ['Nicola Elia', 'Mingyi Hong', 'Xinwei Zhang'] | 2022-04-27 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-1.30111977e-01 -4.48775589e-01 -1.72643095e-01 -1.78377435e-01
-6.42311633e-01 -7.51224697e-01 3.34030420e-01 1.84319466e-01
-2.15651676e-01 8.97455871e-01 1.65966317e-01 -3.29110056e-01
-3.97106469e-01 -6.28690183e-01 -5.82714498e-01 -8.89599919e-01
-2.54280478e-01 -1.09014295e-01 -1.21439286e-01 -2.70651072e-01
2.00148672e-02 4.88922536e-01 -1.05148399e+00 -3.13771278e-01
7.08736062e-01 1.07195675e+00 9.86510888e-04 9.14736390e-01
4.05703068e-01 7.27835894e-01 -4.11590815e-01 -1.41112372e-01
5.41965902e-01 -7.49983013e-01 -5.23002267e-01 3.66104618e-02
3.95555258e-01 -5.05913973e-01 -3.16006064e-01 1.17596483e+00
7.65107036e-01 4.19201136e-01 1.51084289e-01 -1.14694452e+00
-2.82886028e-01 6.27591491e-01 -4.24243867e-01 1.10618234e-01
3.70326862e-02 -3.23467441e-02 9.44600344e-01 -7.39983201e-01
4.98831868e-01 1.07827747e+00 6.75035298e-01 3.91090512e-01
-1.17002976e+00 -3.12960505e-01 2.71835506e-01 5.77427596e-02
-1.05228341e+00 -4.66099381e-01 6.97635770e-01 -3.34724694e-01
2.21483529e-01 6.09156191e-01 7.01858461e-01 4.01632428e-01
1.95744604e-01 1.06275916e+00 9.71002936e-01 -4.86734688e-01
6.15135670e-01 -5.07889725e-02 4.62429449e-02 7.74851501e-01
3.67377520e-01 -8.81797671e-02 -5.95594168e-01 -7.03925550e-01
8.47126186e-01 2.64031053e-01 -4.50608850e-01 -5.96668124e-01
-1.24697506e+00 1.02919269e+00 3.71302038e-01 2.54830867e-01
-4.07555163e-01 3.80416691e-01 5.15178382e-01 8.54830205e-01
7.95799375e-01 2.51111418e-01 -3.90572131e-01 -3.04766800e-02
-6.90404654e-01 5.29292107e-01 1.03333056e+00 4.80999261e-01
8.30294430e-01 2.27311477e-01 -2.00324491e-01 6.38747156e-01
1.03202760e-01 6.63777530e-01 2.19341502e-01 -1.25155604e+00
3.21363777e-01 1.52123168e-01 5.35335660e-01 -1.18017673e+00
-2.59212226e-01 -4.56314176e-01 -1.12777448e+00 -2.04736181e-02
3.84833992e-01 -6.90667987e-01 1.48673937e-01 1.75938511e+00
7.55547583e-01 9.73135233e-02 -2.71555722e-01 1.12402916e+00
-4.79030877e-01 7.76105344e-01 -5.47824860e-01 -5.99705160e-01
7.05853641e-01 -8.12987626e-01 -7.13952959e-01 2.73161322e-01
5.91608167e-01 -5.23798108e-01 5.67313194e-01 5.25281489e-01
-1.07991099e+00 -3.54945920e-02 -7.94495046e-01 1.85838923e-01
1.48431912e-01 2.24464983e-01 8.18002343e-01 5.85593283e-01
-1.31624401e+00 8.45503867e-01 -1.10892630e+00 -5.59154093e-01
6.71219965e-03 1.57712117e-01 2.58638173e-01 3.62977721e-02
-8.68627608e-01 4.74270105e-01 -2.83100098e-01 3.46883804e-01
-8.62303197e-01 -6.40990853e-01 -3.01936746e-01 5.15182083e-03
3.89273942e-01 -1.14032185e+00 1.50226486e+00 -9.59540606e-01
-1.79779935e+00 1.62405387e-01 -2.72326708e-01 -7.41577327e-01
8.02476227e-01 -4.31221873e-01 5.36446385e-02 3.93603742e-02
-1.10328779e-01 -4.38409358e-01 1.08477712e+00 -5.25719285e-01
-6.94641471e-01 -6.29867673e-01 -4.07188609e-02 2.52662569e-01
-8.39697897e-01 2.84002036e-01 3.03768724e-01 -6.50527000e-01
-1.93162471e-01 -8.43480170e-01 -6.92189515e-01 4.19140995e-01
-1.49180740e-01 -1.72510788e-01 7.14498460e-01 -1.64984599e-01
1.36244655e+00 -2.03125763e+00 5.04085243e-01 1.97712258e-01
4.79306996e-01 4.00805511e-02 2.54052162e-01 8.64403307e-01
4.68543977e-01 9.90998074e-02 -1.36989787e-01 -3.85606349e-01
1.39946938e-01 -1.01550110e-01 -7.01316893e-01 1.08881164e+00
-4.03726220e-01 6.35422468e-01 -9.94965613e-01 4.33004573e-02
1.76040769e-01 2.08802730e-01 -5.50722361e-01 -7.35589340e-02
-1.29689336e-01 7.51508594e-01 -1.06405509e+00 2.55955577e-01
2.30846211e-01 -4.10240054e-01 2.67567247e-01 1.80224270e-01
-4.47599828e-01 -1.36019319e-01 -1.49946105e+00 1.53553843e+00
-5.55194199e-01 6.99550450e-01 1.16804993e+00 -1.34491098e+00
5.88411689e-01 4.14211988e-01 7.71612287e-01 -9.48999524e-02
1.96365476e-01 3.18895727e-01 -4.91556138e-01 -2.45215163e-01
5.71712911e-01 6.11711517e-02 -4.93282042e-02 8.50580096e-01
-2.22336322e-01 -4.25360277e-02 3.75779048e-02 3.76567155e-01
1.13658202e+00 -5.78917146e-01 3.66864055e-01 -4.35680449e-01
5.19906938e-01 -2.19938397e-01 4.81225848e-01 1.01649177e+00
-4.36354548e-01 -7.28564858e-02 1.56370655e-01 -4.62918937e-01
-1.03672004e+00 -7.22478509e-01 1.77423090e-01 1.29875171e+00
-3.74919991e-03 -4.71260935e-01 -5.07600605e-01 -3.38395536e-02
2.80068070e-01 -2.14254037e-01 -3.27739984e-01 -9.76246521e-02
-5.13884068e-01 -5.48360944e-01 4.15040225e-01 2.86039472e-01
5.19175708e-01 -2.81829119e-01 -6.02856576e-01 3.71323019e-01
-6.98531047e-02 -6.85208321e-01 -8.36506248e-01 -2.62653053e-01
-1.28778362e+00 -7.98045933e-01 -8.79368782e-01 -3.83744568e-01
4.37218040e-01 5.94219983e-01 7.35458791e-01 -1.91586595e-02
-2.30161697e-01 1.04850590e+00 -4.71223630e-02 -4.00668621e-01
-1.57123789e-01 2.82717682e-02 4.88087207e-01 6.37688160e-01
-4.41656232e-01 -6.29613400e-01 -7.97023892e-01 3.56562287e-01
-8.15061629e-01 -2.45245889e-01 1.53901562e-01 7.64210463e-01
4.54190493e-01 -3.56160492e-01 8.27058554e-01 -7.77047515e-01
1.14630568e+00 -6.14931345e-01 -1.08482993e+00 2.73591161e-01
-5.76702952e-01 9.48069245e-02 1.06324756e+00 -3.25463355e-01
-8.25455129e-01 1.62958473e-01 2.75179744e-01 -3.48249495e-01
3.68766159e-01 6.20943785e-01 2.32460320e-01 -4.91075993e-01
7.73294568e-01 2.50357836e-01 3.63548994e-01 -4.86859500e-01
8.21261406e-01 7.52820015e-01 1.16766572e-01 -9.21006978e-01
5.03528595e-01 9.55827832e-01 1.23494573e-01 -9.84044313e-01
-5.61064839e-01 -5.95984459e-01 -4.62431982e-02 -2.30582684e-01
1.49240449e-01 -8.38324130e-01 -9.30146933e-01 6.14131331e-01
-9.42992151e-01 -6.22899890e-01 -4.21417534e-01 4.97616500e-01
-8.46482217e-01 2.72165477e-01 -8.83234382e-01 -1.08128977e+00
-4.59779829e-01 -7.73312688e-01 7.97133565e-01 3.38288248e-01
1.48208156e-01 -1.29889929e+00 4.89099175e-01 -1.53060868e-01
9.79737818e-01 2.70441979e-01 2.46706307e-01 -2.35232279e-01
-6.45111978e-01 -1.55776009e-01 1.14590190e-01 2.49333486e-01
1.40130714e-01 4.07651030e-02 -5.28727710e-01 -7.48146236e-01
3.85635048e-01 -4.01914328e-01 6.84551656e-01 5.34417212e-01
9.21958029e-01 -5.72252095e-01 -3.83327544e-01 6.48096621e-01
1.58434677e+00 -4.53104585e-01 -1.65203601e-01 -1.92909405e-01
4.63907391e-01 2.68307745e-01 3.35860401e-01 1.02139664e+00
1.86219722e-01 4.50197071e-01 1.57050997e-01 2.59569734e-01
4.11646515e-01 3.18545438e-02 4.22816724e-01 9.30505276e-01
-3.91671546e-02 1.60483211e-01 -4.96429771e-01 4.61659491e-01
-2.28583860e+00 -9.95128453e-01 5.84757030e-02 2.42478824e+00
9.24321830e-01 -7.71273732e-01 4.60402608e-01 -1.04599051e-01
1.03286314e+00 1.81245238e-01 -8.27681422e-01 -3.31382304e-01
2.49761995e-02 1.72089681e-01 7.39754498e-01 3.50456417e-01
-9.88517404e-01 5.10572493e-01 6.94181442e+00 6.24423087e-01
-1.41285086e+00 2.32798442e-01 3.82747620e-01 -1.36170000e-01
8.41376856e-02 1.43653732e-02 -5.49732387e-01 2.69771188e-01
1.03441489e+00 -9.14596140e-01 9.28183496e-01 9.99499738e-01
7.80899823e-01 4.50334977e-03 -8.42987835e-01 9.44900930e-01
-4.05895412e-01 -1.33483052e+00 -3.53754610e-01 1.19169004e-01
1.02477586e+00 2.32579753e-01 5.47321923e-02 -1.11223035e-01
6.03245556e-01 -4.63001162e-01 3.49477500e-01 3.31665009e-01
4.71522510e-01 -5.47150433e-01 1.61196515e-02 5.69840968e-01
-1.28685749e+00 -4.51232404e-01 -4.37896430e-01 -7.18971968e-01
2.15997919e-01 8.84735942e-01 -1.12695493e-01 7.90038049e-01
5.52187860e-01 8.38481128e-01 2.66102850e-02 1.41120148e+00
2.92937383e-02 7.70323038e-01 -7.60001957e-01 -5.00395536e-01
9.41864997e-02 -4.03018802e-01 7.55154788e-01 6.73072338e-01
2.62308538e-01 2.06569918e-02 4.23786253e-01 5.28044403e-01
-2.00331479e-01 2.67069846e-01 -6.87288940e-01 9.31804255e-02
4.65982407e-01 1.34924400e+00 -4.99979168e-01 -4.24412429e-01
-5.42678237e-01 8.52183044e-01 3.67693037e-01 5.62335312e-01
-5.86847007e-01 -4.57791299e-01 9.80574787e-01 -2.73346066e-01
1.79849550e-01 -7.49735773e-01 -1.51358947e-01 -1.51572108e+00
1.01826265e-01 -7.98370123e-01 3.08363408e-01 -8.02241117e-02
-1.30643344e+00 -6.08097799e-02 -1.48853347e-01 -9.51186299e-01
-2.99987525e-01 -1.82450324e-01 -7.66030252e-01 6.03789032e-01
-1.09252131e+00 -4.79610741e-01 -1.00733452e-01 9.97435212e-01
2.86493689e-01 2.87746668e-01 6.04888737e-01 3.79307896e-01
-6.54670596e-01 3.19838136e-01 1.00586748e+00 2.66998447e-02
7.40402997e-01 -1.10916436e+00 -5.91087453e-02 1.00685906e+00
9.54289809e-02 7.09481001e-01 7.63537526e-01 -1.88733310e-01
-2.14016175e+00 -1.21658313e+00 3.79452199e-01 -1.14898317e-01
1.15042603e+00 -3.05447340e-01 -4.35681373e-01 6.01691782e-01
1.17360413e-01 2.69508511e-01 3.85088265e-01 1.62915036e-01
1.84499785e-01 -6.33172095e-01 -9.85285521e-01 5.38901150e-01
7.73147225e-01 -4.22956109e-01 -1.90927424e-02 6.71239078e-01
4.83133793e-01 -4.05320853e-01 -7.79044151e-01 -1.72624603e-01
4.83490527e-01 -7.33081102e-01 6.26672804e-01 -6.49463654e-01
-1.97494090e-01 -1.21446289e-01 -1.18863754e-01 -1.59304261e+00
1.42660467e-02 -1.51806939e+00 -4.72725958e-01 9.04512942e-01
-9.97085869e-02 -1.18700719e+00 5.95906913e-01 6.37546122e-01
1.01035677e-01 -7.21762300e-01 -9.53545213e-01 -7.86108077e-01
1.96317390e-01 -2.25107625e-01 2.56798178e-01 8.11862290e-01
2.09633514e-01 1.15391202e-01 -5.56834400e-01 1.41964108e-01
8.10822845e-01 2.91604161e-01 8.54509473e-01 -9.48833108e-01
-4.02934104e-01 -4.23614383e-01 -2.65732825e-01 -1.51766729e+00
-1.46878511e-01 -6.88974977e-01 -1.11750953e-01 -1.10893226e+00
-1.42992526e-01 -4.85579193e-01 -2.58604437e-01 1.20642841e-01
7.47159421e-02 1.53486524e-02 2.36879528e-01 5.45332432e-01
-8.06338966e-01 6.36966825e-01 1.10794151e+00 4.79573458e-02
-3.43464345e-01 3.37915361e-01 -5.82030594e-01 4.61933434e-01
8.93585622e-01 -3.50900531e-01 -3.06262463e-01 -6.49834216e-01
4.46146637e-01 3.20827216e-01 4.82368648e-01 -8.36222887e-01
7.16436565e-01 -3.79591405e-01 -2.13976502e-01 3.62441614e-02
-2.77756080e-02 -6.48170292e-01 -1.87655300e-01 7.10549533e-01
-4.10600096e-01 1.62431732e-01 -2.69503683e-01 9.50219572e-01
-2.59574264e-01 2.13199869e-01 6.44926667e-01 -1.58769742e-01
-2.31289968e-01 3.89540881e-01 -5.35842538e-01 5.96210435e-02
1.00925136e+00 4.78062034e-01 -1.09181345e-01 -8.12165141e-01
-7.66373456e-01 6.65655673e-01 3.31064790e-01 -1.06607251e-01
2.50251800e-01 -1.11681676e+00 -6.18516862e-01 5.18452153e-02
-6.15649045e-01 -2.39745885e-01 1.93790957e-01 1.15435171e+00
-2.82359749e-01 4.00625497e-01 2.86503017e-01 -6.67556643e-01
-8.79525959e-01 2.05366328e-01 4.45736825e-01 -1.68483019e-01
-4.83661741e-01 4.98769462e-01 -1.78333059e-01 -5.50688878e-02
1.34309858e-01 -4.22241300e-01 6.00329936e-01 -2.59740442e-01
7.47816086e-01 8.14625144e-01 2.51055006e-02 1.39312834e-01
-1.90278515e-01 4.55274582e-01 1.93619624e-01 -3.59531969e-01
1.23976624e+00 -6.17941856e-01 -4.40818340e-01 5.85540771e-01
1.12945867e+00 1.54217824e-01 -1.22623968e+00 -3.68619740e-01
-1.32750154e-01 -2.79577196e-01 -1.25100380e-02 -2.15040430e-01
-1.12750423e+00 7.40521550e-01 3.48038763e-01 6.75285757e-01
1.17660236e+00 -2.95497715e-01 7.39402890e-01 7.00025856e-01
8.15753102e-01 -9.63059545e-01 -2.40963772e-01 6.03211343e-01
5.90457618e-01 -8.67241025e-01 1.54646235e-02 1.61725702e-03
-3.83888446e-02 1.17275906e+00 -7.28548095e-02 -7.54239440e-01
7.74176657e-01 1.81329861e-01 -3.24646771e-01 2.49945775e-01
-9.47180927e-01 1.64349899e-01 -3.06815624e-01 6.66810945e-02
4.05278146e-01 6.78116977e-02 -6.95471525e-01 3.23459119e-01
2.57858485e-01 3.36059868e-01 6.54144466e-01 1.09788597e+00
-6.48112297e-01 -1.12473011e+00 -3.03412795e-01 4.57846642e-01
-5.94052970e-01 2.42514640e-01 -3.07373464e-01 2.90302068e-01
-7.73263156e-01 8.93581092e-01 -2.99231887e-01 8.87824513e-04
-1.21286921e-01 -4.26913388e-02 4.33327377e-01 -4.22290355e-01
-5.91078520e-01 -2.88474112e-04 -3.47124815e-01 -6.88580513e-01
-5.16869307e-01 -5.18080771e-01 -8.16944897e-01 -6.34367287e-01
-4.17169034e-01 5.21507204e-01 7.79170036e-01 8.61490488e-01
7.69675434e-01 1.16754040e-01 1.11220098e+00 -7.68261194e-01
-1.35498714e+00 -7.10427880e-01 -8.14020395e-01 -1.77787617e-01
7.07828879e-01 -2.95217425e-01 -5.81378222e-01 -1.27873525e-01] | [6.2996697425842285, 4.88827657699585] |
d5945e76-5979-4b93-a19d-bb646fc3f316 | stochastic-model-predictive-control-with-1 | 2305.19262 | null | https://arxiv.org/abs/2305.19262v1 | https://arxiv.org/pdf/2305.19262v1.pdf | Stochastic Model Predictive Control with Dynamic Chance Constraints | In this work, we introduce a stochastic model predictive control scheme for dynamic chance constraints. We consider linear discrete-time systems affected by unbounded additive stochastic disturbance and subject to chance constraints that are defined by time-varying probabilities with a common, fixed lower bound. By utilizing probabilistic reachable tubes with dynamic cross-sections, we are reformulating the stochastic optimization problem into a deterministic tube-based MPC problem with time-varying tightened constraints. We show that the resulting deterministic MPC formulation with dynamic tightened constraints is recursively feasible and that the closed-loop stochastic system will satisfy the corresponding dynamic chance constraints. In addition, we will also introduce a novel implementation using zonotopes to describe the tightening analytically. Finally, we will end with an example to illustrate the benefits of the developed approach to stochastic MPC with dynamic chance constraints. | ['Mircea Lazar', 'Sofie Haesaert', 'Maico Hendrikus Wilhelmus Engelaar'] | 2023-05-30 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [ 1.93071336e-01 4.35839325e-01 -4.17901754e-01 3.35054904e-01
-8.23032260e-01 -1.01675820e+00 5.64071119e-01 -6.04051054e-02
9.44411159e-02 1.40984631e+00 -6.19560704e-02 -6.67012155e-01
-7.70096123e-01 -8.81839991e-01 -6.63420677e-01 -1.05962706e+00
-2.96782553e-01 4.63724673e-01 9.80147421e-02 -2.43938323e-02
-9.04736295e-02 5.31790733e-01 -1.23349130e+00 -3.50379169e-01
1.02252793e+00 8.08671713e-01 1.72708973e-01 8.55652213e-01
3.74332905e-01 5.03816187e-01 -5.70354939e-01 1.28922343e-01
4.67218846e-01 -1.55638263e-01 -3.08079630e-01 2.88084239e-01
-5.14952123e-01 -3.15434076e-02 -7.25594833e-02 8.82216215e-01
2.17080683e-01 7.82334864e-01 6.99677646e-01 -1.26811624e+00
-2.81543378e-02 3.91447127e-01 -3.46441060e-01 -5.26886694e-02
2.39511058e-01 4.32291240e-01 7.17087746e-01 -1.50006264e-01
4.44952726e-01 1.07778931e+00 3.22953612e-01 5.94640672e-01
-1.50888169e+00 4.00263444e-02 5.82569838e-01 -3.44248116e-01
-1.25452185e+00 1.98171228e-01 1.74363554e-01 -5.69969356e-01
9.68701005e-01 8.62688720e-01 5.30073941e-01 5.49883187e-01
6.44349813e-01 3.89557123e-01 1.01348948e+00 -2.63079405e-01
6.42456353e-01 4.65143695e-02 -5.96459769e-02 2.24581212e-01
4.80962574e-01 7.15387166e-01 3.55160564e-01 -3.57675135e-01
4.67972785e-01 9.11522508e-02 -3.67746741e-01 -5.75790405e-01
-8.72572482e-01 1.00218225e+00 2.30193920e-02 -3.50104235e-02
-3.23376268e-01 2.92715222e-01 3.75744799e-04 -1.14814691e-01
3.77473861e-01 4.48067635e-01 -5.00072896e-01 1.14569865e-01
-5.88050544e-01 6.08554840e-01 1.17200828e+00 1.33429384e+00
-1.37685947e-02 4.59256440e-01 -5.56386530e-01 6.05531856e-02
3.59991401e-01 8.76045942e-01 -1.23621359e-01 -9.78757799e-01
5.78063190e-01 9.62637365e-03 8.28082860e-01 -3.98922622e-01
-2.43541881e-01 -6.69915318e-01 -5.33263266e-01 4.85835135e-01
3.34765464e-01 -8.28737736e-01 -8.07263374e-01 1.60394454e+00
2.14318991e-01 1.91748440e-01 2.16689229e-01 6.59661770e-01
-4.43701714e-01 1.29349077e+00 -2.93167859e-01 -9.37082946e-01
6.12894416e-01 -8.13429058e-01 -8.52916896e-01 3.39300007e-01
1.98641881e-01 -4.76146728e-01 6.87522769e-01 5.34086406e-01
-9.94053304e-01 2.43886352e-01 -8.46964121e-01 7.49676585e-01
-3.10480893e-01 8.47990215e-02 1.99039757e-01 9.68798637e-01
-7.42922962e-01 6.51788354e-01 -7.48538673e-01 -1.29145365e-02
-3.21098804e-01 4.66562271e-01 4.09600705e-01 4.72778678e-01
-9.20270324e-01 1.07078707e+00 3.72652054e-01 4.52790707e-01
-1.14709747e+00 -6.38604224e-01 -7.38583922e-01 1.33851305e-01
7.60290861e-01 -6.90919936e-01 1.45109081e+00 -2.39774227e-01
-1.99009311e+00 -6.01304136e-02 -5.44082597e-02 -5.92768550e-01
7.67634749e-01 -5.22325747e-03 -5.55828325e-02 -1.71656147e-01
-1.55514851e-01 -6.11851811e-01 3.12355042e-01 -1.43459511e+00
-6.58342063e-01 2.05972195e-01 5.62075794e-01 3.00603330e-01
2.24743694e-01 -3.15377600e-02 -3.32030095e-03 -3.88260454e-01
-3.97095293e-01 -1.21024501e+00 -1.00175130e+00 -6.40589416e-01
-8.88021946e-01 3.18956792e-01 5.28344750e-01 -1.59959868e-01
1.50201166e+00 -1.66714811e+00 6.08839810e-01 4.00237799e-01
-3.62255752e-01 -6.65523782e-02 3.85731161e-01 7.56333411e-01
-1.51847936e-02 2.96793580e-01 -6.35201633e-01 -2.17267260e-01
2.34171823e-01 3.95179331e-01 -8.10936391e-01 5.63995004e-01
3.17558274e-02 1.94714814e-01 -9.15541053e-01 5.10573238e-02
7.26335406e-01 -1.53741762e-01 -4.57647026e-01 1.03437133e-01
-5.64063311e-01 3.69134218e-01 -7.40212202e-01 3.81600201e-01
5.96903920e-01 6.50169075e-01 -6.67777937e-03 5.79488695e-01
-5.99561214e-01 -3.77737314e-01 -1.52983797e+00 6.82054877e-01
-8.71524155e-01 2.90232837e-01 6.01564229e-01 -6.44384384e-01
2.87932724e-01 2.58342892e-01 4.66109484e-01 5.94836712e-01
3.75103980e-01 1.10617414e-01 -2.29510218e-01 -6.80646747e-02
6.55247569e-01 -6.36564672e-01 -3.74925047e-01 3.07794157e-02
-2.94182986e-01 -9.79948580e-01 9.76581573e-02 4.44550402e-02
7.16580570e-01 -9.18440707e-03 5.23346782e-01 -5.87452650e-01
7.29191124e-01 2.91782260e-01 6.65076077e-01 5.41634321e-01
-2.43676439e-01 3.09146434e-01 8.54477286e-01 2.07012773e-01
-6.50853872e-01 -1.19189131e+00 -8.14017579e-02 4.60100949e-01
2.13236213e-01 1.45624414e-01 -4.99589294e-01 -1.53344154e-01
3.74418736e-01 1.43022966e+00 -4.70502138e-01 3.25942761e-04
-2.20522702e-01 -1.01993763e+00 -2.05787614e-01 -3.84756215e-02
-1.85375184e-01 -4.63464148e-02 -2.91145325e-01 5.56851447e-01
5.09742141e-01 -8.28027904e-01 -4.67426270e-01 5.54617286e-01
-7.39902020e-01 -9.24606860e-01 -7.26323724e-01 -2.53961205e-01
6.64975107e-01 -2.20793653e-02 5.14018655e-01 -6.35175467e-01
2.55968496e-02 5.45475006e-01 1.47149071e-01 -4.98582870e-01
-3.01409930e-01 -2.70974606e-01 2.31878340e-01 -1.48285091e-01
-8.21612895e-01 -1.69667348e-01 -1.30379587e-01 3.80613983e-01
-9.18965101e-01 -7.07255602e-02 -2.60665417e-01 8.12637448e-01
9.16440070e-01 4.82444525e-01 3.56183380e-01 -5.64073563e-01
7.74935007e-01 -4.71775591e-01 -1.46999264e+00 2.00649753e-01
-3.32493395e-01 -8.51084106e-03 1.02966332e+00 -4.31363016e-01
-1.31853998e+00 5.39823532e-01 4.39894527e-01 -3.82333517e-01
2.04036593e-01 4.66372877e-01 -3.71752739e-01 -2.34430537e-01
1.38397753e-01 2.58652158e-02 -3.49971086e-01 -2.47557625e-01
5.91797590e-01 2.26815954e-01 3.53815854e-01 -9.31789875e-01
9.66283739e-01 3.70142490e-01 5.24858117e-01 -3.58516425e-01
-5.69766760e-01 -2.32213542e-01 -2.75448918e-01 -4.67107058e-01
7.03603864e-01 -5.42211711e-01 -9.67932105e-01 3.06854397e-01
-8.65509212e-01 -6.32282436e-01 -8.12803984e-01 6.36793196e-01
-1.12667632e+00 -2.75543681e-03 -3.55673403e-01 -1.74870789e+00
3.11620355e-01 -1.22876775e+00 7.35002279e-01 2.93335736e-01
2.88624614e-01 -1.35254502e+00 3.60724360e-01 -2.88393676e-01
3.31029296e-01 9.77102876e-01 5.14280677e-01 -2.03299224e-01
-8.42707038e-01 -3.74508440e-01 4.06786978e-01 1.66706711e-01
-6.98014498e-02 4.70196337e-01 -5.06516755e-01 -2.74760991e-01
3.38002592e-01 4.48613584e-01 4.30505127e-01 8.20203304e-01
9.17868674e-01 -5.93349159e-01 -6.88080668e-01 4.74083811e-01
1.98020351e+00 4.43366438e-01 -1.14766032e-01 3.20500076e-01
3.12428832e-01 5.47971487e-01 9.11566377e-01 7.82811224e-01
8.16394538e-02 4.94793326e-01 8.91788304e-01 2.17581034e-01
6.85519278e-01 -4.59646322e-02 3.37459892e-01 2.17434883e-01
-1.87713370e-01 -6.35937214e-01 -8.22293758e-01 7.44227529e-01
-1.89991868e+00 -1.02609050e+00 -5.24554372e-01 2.54838419e+00
7.30658233e-01 7.91515782e-02 -1.37434036e-01 -2.25390896e-01
1.03792965e+00 -5.10799997e-02 -2.94595927e-01 -1.05289543e+00
3.75907943e-02 -2.50591844e-01 1.45494092e+00 1.13086343e+00
-1.20397067e+00 4.68964487e-01 6.88431120e+00 7.93797731e-01
-8.19549918e-01 -1.42489463e-01 6.33179128e-01 -5.69959342e-01
-3.93768042e-01 1.28515244e-01 -8.85363698e-01 6.43523097e-01
1.17797089e+00 -1.16029215e+00 5.76029181e-01 9.47510064e-01
1.07443976e+00 -4.87536520e-01 -8.19164872e-01 2.44918957e-01
-6.28706574e-01 -1.15899050e+00 -3.20595354e-01 2.79782742e-01
1.47982919e+00 -2.91339219e-01 3.00471783e-01 2.39923239e-01
9.90301430e-01 -8.26365530e-01 7.54732370e-01 6.69387817e-01
4.98473793e-01 -1.09090960e+00 7.94739962e-01 6.61516368e-01
-1.20986629e+00 -4.03074831e-01 -3.83385643e-02 -3.44257563e-01
8.06790829e-01 7.32426465e-01 -3.33371103e-01 9.55522835e-01
-3.24895591e-01 1.58526778e-01 4.81507450e-01 1.53047693e+00
-3.99428576e-01 4.55681026e-01 -7.17063904e-01 -3.74450952e-01
4.03599232e-01 -5.01472771e-01 9.74374175e-01 9.82097030e-01
5.13622880e-01 3.56598765e-01 4.61551100e-01 8.83568048e-01
6.96088731e-01 -3.58458877e-01 -8.85709465e-01 3.56342256e-01
9.32423472e-02 7.46992052e-01 -4.20185208e-01 -2.03572273e-01
-2.71475226e-01 5.98625422e-01 -5.16243637e-01 7.18838096e-01
-1.22388375e+00 -4.08886075e-01 7.76928127e-01 -3.23953360e-01
8.07375610e-02 -5.03159583e-01 -5.74919462e-01 -9.55069065e-01
-1.73769966e-01 -1.11628376e-01 2.54534274e-01 -5.23421764e-01
-1.01105034e+00 2.26550460e-01 4.18246657e-01 -1.41165090e+00
-4.83554810e-01 -5.88547468e-01 -1.02788293e+00 1.42313766e+00
-1.19439566e+00 -6.38402641e-01 5.92892110e-01 3.39511216e-01
5.07080317e-01 1.38742164e-01 5.50718844e-01 -2.67271668e-01
-9.86932456e-01 -2.03710541e-01 1.06709969e+00 -8.25801015e-01
-2.63752528e-02 -1.64457345e+00 -3.19742620e-01 1.26423526e+00
-9.49564457e-01 5.31072080e-01 1.36754370e+00 -7.33333766e-01
-1.51626527e+00 -1.48442900e+00 5.19943058e-01 -2.95936227e-01
1.32952189e+00 -1.33471936e-01 -2.90209919e-01 6.22972667e-01
1.44213364e-01 -1.68313026e-01 1.97439685e-01 -5.49547255e-01
6.93032801e-01 1.45457730e-01 -1.54091012e+00 5.50752282e-01
5.25602937e-01 -9.88408178e-02 -4.95021105e-01 7.94496596e-01
1.03430688e+00 -7.05922902e-01 -6.74374998e-01 3.61930341e-01
1.17938682e-01 -1.25385925e-01 7.17863321e-01 -4.51765120e-01
5.63763715e-02 -4.53541696e-01 -1.37653917e-01 -1.72288024e+00
-3.31341736e-02 -1.33705842e+00 -5.61957844e-02 9.94852126e-01
6.82789445e-01 -9.35466766e-01 5.95494330e-01 1.01022410e+00
-2.97483653e-01 -1.05757451e+00 -1.40580642e+00 -1.39372909e+00
3.74087930e-01 -5.63889742e-01 3.49636078e-01 3.51662844e-01
4.63329047e-01 -5.18708646e-01 -3.64860564e-01 8.10733318e-01
5.37580073e-01 2.75982291e-01 2.46297449e-01 -4.18560237e-01
-5.66247642e-01 -3.74460876e-01 1.46461681e-01 -6.39849007e-01
8.54199603e-02 -3.88224661e-01 5.64939439e-01 -1.62380469e+00
1.11928536e-02 -1.34829164e-01 -1.53141081e-01 -2.29858160e-01
-2.14882106e-01 -6.38698578e-01 3.86006802e-01 -2.80340254e-01
-2.31674984e-01 9.00658906e-01 1.16157484e+00 7.05394074e-02
-6.26796424e-01 8.54054034e-01 -7.95812011e-02 6.82759881e-01
9.70711827e-01 -2.72900343e-01 -6.40518129e-01 6.59461990e-02
-1.03380390e-01 6.69805527e-01 9.45705622e-02 -8.18145275e-01
2.45018117e-02 -1.15521002e+00 -4.42492366e-01 -8.76056075e-01
2.38381594e-01 -1.11087036e+00 3.98658335e-01 8.19481194e-01
-1.53424412e-01 -7.94060230e-02 4.40289497e-01 1.10006642e+00
-7.26104975e-02 -4.54667330e-01 7.23797023e-01 1.35827377e-01
-1.76129699e-01 2.30168521e-01 -1.16515982e+00 -3.78314048e-01
1.84043229e+00 -9.59266871e-02 -5.55140385e-03 -4.97262388e-01
-1.40349710e+00 8.31927776e-01 4.14638370e-01 -2.17954606e-01
6.48376718e-02 -8.23374689e-01 -9.98334363e-02 -4.54223901e-01
-6.23967171e-01 -1.99553475e-01 1.78840309e-01 5.45355439e-01
-3.18147331e-01 8.94410312e-01 3.42136800e-01 -2.98736513e-01
-8.42343390e-01 9.73715484e-01 7.56985068e-01 -5.05541921e-01
2.21424267e-01 7.17849791e-01 7.93829039e-02 -4.71851200e-01
9.23327804e-02 -9.83542144e-01 1.31333232e-01 -1.45622492e-01
1.54013736e-02 6.05217397e-01 -3.33379954e-01 -1.67055309e-01
-2.33061954e-01 4.46122408e-01 9.66591954e-01 -6.39415503e-01
1.08990538e+00 -5.34640908e-01 1.27451643e-01 3.66366863e-01
6.11058176e-01 2.82830894e-01 -1.57923365e+00 3.33570302e-01
-9.29602310e-02 -2.50820011e-01 4.34482982e-03 -7.79415250e-01
-7.68277526e-01 4.04776424e-01 2.30705097e-01 4.36290473e-01
9.74397421e-01 -4.72245187e-01 2.99996525e-01 2.97680765e-01
6.76437259e-01 -9.25099611e-01 -9.11375999e-01 8.51530015e-01
9.93021548e-01 -3.84216130e-01 4.31465656e-02 -9.85036016e-01
-5.51376879e-01 1.07420182e+00 3.30994219e-01 -5.40524840e-01
5.77459276e-01 5.99395216e-01 -8.39575768e-01 5.62553942e-01
-1.17492568e+00 -4.76812750e-01 -4.43237536e-02 6.95003510e-01
-9.82619673e-02 6.56000197e-01 -8.79107654e-01 6.32897377e-01
6.41217530e-02 -1.02957904e-01 1.04668736e+00 7.62785375e-01
-5.08290231e-01 -8.69032860e-01 -8.54735374e-01 2.97064990e-01
-2.88070321e-01 2.74733659e-02 3.27281505e-02 9.04858589e-01
-7.87759945e-02 1.12608004e+00 -2.76702553e-01 2.17404049e-02
5.78395605e-01 -1.22192167e-01 1.61380395e-01 -1.06119347e+00
-5.74115157e-01 1.42135918e-01 5.05939424e-01 -4.34970051e-01
2.39810139e-01 -8.00174594e-01 -1.23437774e+00 -2.77109563e-01
-6.16966546e-01 1.83851108e-01 4.57726598e-01 7.42154419e-01
-5.38797192e-02 9.17514682e-01 9.57506120e-01 -7.84535527e-01
-1.13577437e+00 -5.39420843e-01 -6.98636234e-01 -8.68881941e-01
3.38935614e-01 -7.44224370e-01 -7.13993013e-01 -7.95133337e-02] | [4.771667003631592, 2.414045810699463] |
aa172d2b-090d-44c4-892f-c7f6932ced69 | tesla-test-time-self-learning-with-automatic | 2303.09870 | null | https://arxiv.org/abs/2303.09870v1 | https://arxiv.org/pdf/2303.09870v1.pdf | TeSLA: Test-Time Self-Learning With Automatic Adversarial Augmentation | Most recent test-time adaptation methods focus on only classification tasks, use specialized network architectures, destroy model calibration or rely on lightweight information from the source domain. To tackle these issues, this paper proposes a novel Test-time Self-Learning method with automatic Adversarial augmentation dubbed TeSLA for adapting a pre-trained source model to the unlabeled streaming test data. In contrast to conventional self-learning methods based on cross-entropy, we introduce a new test-time loss function through an implicitly tight connection with the mutual information and online knowledge distillation. Furthermore, we propose a learnable efficient adversarial augmentation module that further enhances online knowledge distillation by simulating high entropy augmented images. Our method achieves state-of-the-art classification and segmentation results on several benchmarks and types of domain shifts, particularly on challenging measurement shifts of medical images. TeSLA also benefits from several desirable properties compared to competing methods in terms of calibration, uncertainty metrics, insensitivity to model architectures, and source training strategies, all supported by extensive ablations. Our code and models are available on GitHub. | ['Jean-Philippe Thiran', 'Behzad Bozorgtabar', 'Guillaume Vray', 'Devavrat Tomar'] | 2023-03-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tomar_TeSLA_Test-Time_Self-Learning_With_Automatic_Adversarial_Augmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tomar_TeSLA_Test-Time_Self-Learning_With_Automatic_Adversarial_Augmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['self-learning'] | ['natural-language-processing'] | [ 4.21168685e-01 3.29965740e-01 -3.47006768e-01 -5.01823902e-01
-1.35817492e+00 -7.11323321e-01 3.02935064e-01 1.71999156e-01
-5.77279568e-01 9.55708265e-01 -3.10327441e-01 -2.12126359e-01
-1.54327318e-01 -5.50278842e-01 -9.52650666e-01 -6.94872081e-01
-1.03494644e-01 6.60433769e-01 4.11711901e-01 -1.64051931e-02
-2.43069127e-01 1.63097799e-01 -9.40319717e-01 1.91678051e-02
1.24512005e+00 1.20122361e+00 -2.96972394e-01 5.66152334e-01
4.52069454e-02 5.90671599e-01 -5.22149861e-01 -7.86411166e-01
4.58309025e-01 -6.31180227e-01 -8.34531903e-01 -1.62429944e-01
3.43798995e-01 -2.56427050e-01 -3.16429973e-01 1.18042493e+00
8.23010683e-01 -7.63849691e-02 5.92806697e-01 -1.12009513e+00
-6.24156654e-01 7.39351213e-01 -2.71903574e-01 6.84881881e-02
-6.50386736e-02 2.03252673e-01 4.57810044e-01 -5.77432573e-01
5.59512854e-01 5.45730114e-01 1.00813794e+00 6.95484698e-01
-1.37049389e+00 -7.89383769e-01 -2.67625581e-02 2.54448384e-01
-1.23219085e+00 -1.76250905e-01 9.35236454e-01 -2.85527170e-01
5.40193856e-01 1.71669722e-01 4.50842410e-01 1.46464527e+00
2.22463280e-01 9.18708324e-01 1.40896797e+00 -3.69945258e-01
5.04726648e-01 5.17399967e-01 -2.35528037e-01 7.61193097e-01
1.19343884e-01 2.76933521e-01 -3.16737413e-01 -2.42898002e-01
6.95427716e-01 -2.99424797e-01 -3.43149036e-01 -9.49942291e-01
-1.26436400e+00 6.11121953e-01 3.65716755e-01 8.17017332e-02
3.72701213e-02 -1.52619541e-01 8.11725378e-01 4.95456100e-01
6.32334828e-01 5.22450209e-01 -8.73142600e-01 -5.69390878e-02
-1.03883064e+00 -1.74225405e-01 9.14218247e-01 9.99605238e-01
6.76102579e-01 9.69333276e-02 -3.91337186e-01 7.97298789e-01
-6.91311806e-02 6.61602378e-01 8.47526073e-01 -8.29451740e-01
3.69428933e-01 3.63891870e-01 -2.83640951e-01 -3.57542515e-01
-3.76894504e-01 -8.74336541e-01 -9.47923362e-01 1.54299557e-01
1.87535554e-01 -2.43087843e-01 -1.08000958e+00 1.91868246e+00
5.74882448e-01 5.51508963e-01 2.36516982e-01 4.78712052e-01
7.29591489e-01 -6.39502928e-02 4.67110351e-02 -2.42765203e-01
9.64673698e-01 -9.69588280e-01 -6.90186620e-01 -2.35909820e-01
5.90196192e-01 -4.88428503e-01 1.03724456e+00 2.15962216e-01
-1.16141665e+00 -5.44431269e-01 -1.26081061e+00 1.91940963e-01
-4.54189509e-01 -2.14610472e-01 2.85878032e-01 9.57429230e-01
-8.95124018e-01 8.24029028e-01 -1.10647428e+00 4.51702848e-02
6.85852349e-01 4.39061970e-01 -3.05344045e-01 3.20124626e-03
-1.29733741e+00 8.05341661e-01 6.15864813e-01 -3.45035821e-01
-8.18178415e-01 -1.11976647e+00 -1.07047939e+00 -2.37428963e-01
4.94983405e-01 -8.50346744e-01 1.27889085e+00 -1.15117788e+00
-1.93235207e+00 9.57824171e-01 4.28214073e-01 -8.00651371e-01
1.05193710e+00 -1.19661421e-01 -4.62763220e-01 1.69247016e-01
3.26188616e-02 6.24753237e-01 8.94000649e-01 -1.04914176e+00
-1.43842429e-01 -2.36446112e-01 -2.47590572e-01 -2.18950007e-02
-5.37559450e-01 -5.17480433e-01 -5.04217327e-01 -1.02402568e+00
-1.75100133e-01 -9.46047127e-01 -3.30839127e-01 1.99306328e-02
-4.46011186e-01 3.34389836e-01 6.76037908e-01 -6.26656532e-01
1.04542530e+00 -2.06675339e+00 5.67988353e-03 3.04735214e-01
-6.81479797e-02 3.99424016e-01 -1.70052886e-01 -8.49499330e-02
-1.86975628e-01 3.76816690e-02 -1.03531039e+00 -2.53234416e-01
3.04117892e-02 2.40294993e-01 -9.58033800e-02 5.25753498e-01
2.63512194e-01 1.13814116e+00 -1.07704091e+00 -8.01637411e-01
2.61232495e-01 3.88991922e-01 -6.34324014e-01 1.67250618e-01
-1.80452496e-01 8.25173855e-01 -2.73625284e-01 6.07486129e-01
8.34174156e-01 -3.67023975e-01 3.38749811e-02 -8.75781104e-02
5.93724966e-01 -4.34924029e-02 -9.18663919e-01 2.03162098e+00
-5.68613708e-01 2.54996210e-01 -2.08992347e-01 -1.15795302e+00
7.07844496e-01 3.65052372e-01 6.95983469e-01 -6.45921469e-01
9.46156234e-02 4.06516790e-01 -2.06979454e-01 -3.77082318e-01
-8.94156992e-02 -8.10214877e-02 -1.58132881e-01 2.73403972e-01
5.22541761e-01 -2.98763871e-01 -5.38955666e-02 4.61819805e-02
1.17763638e+00 5.05058169e-01 3.93314183e-01 -7.20045045e-02
5.71697712e-01 -2.10058853e-01 7.36129880e-01 8.77311349e-01
-4.01878417e-01 6.28891587e-01 2.54054308e-01 -6.69246539e-02
-1.06111622e+00 -1.39512014e+00 -5.30447900e-01 7.18354642e-01
1.05798014e-01 -1.44444853e-02 -9.72685933e-01 -1.28023064e+00
4.28121053e-02 7.15043306e-01 -9.07073557e-01 -5.54885268e-01
-4.17063683e-01 -6.47262692e-01 6.79451942e-01 7.15922475e-01
8.13422978e-01 -7.66254127e-01 -2.73042113e-01 2.40094990e-01
-1.48395538e-01 -1.33768225e+00 -6.68581963e-01 3.24912071e-01
-1.13560820e+00 -1.04984426e+00 -9.99169111e-01 -6.52851701e-01
6.94960237e-01 -4.53201711e-01 1.26849604e+00 -5.50985634e-01
-2.83934921e-01 7.03532577e-01 -3.19501579e-01 -5.88989258e-01
-7.09639966e-01 3.14200759e-01 -6.57336786e-02 2.91194227e-02
1.75797388e-01 -6.37991965e-01 -7.92862773e-01 5.05100310e-01
-9.91193175e-01 -1.51690647e-01 7.70437241e-01 1.19358504e+00
1.03244853e+00 -2.86781490e-01 7.24928796e-01 -1.30453122e+00
1.77064329e-01 -5.61450541e-01 -7.74827898e-01 4.89621520e-01
-1.06652546e+00 4.43540454e-01 6.67516053e-01 -6.51430368e-01
-1.18740273e+00 1.74024671e-01 -2.23561257e-01 -6.58433378e-01
-1.08473994e-01 2.61461198e-01 -1.15933921e-02 -4.38534766e-01
1.05078852e+00 3.40913147e-01 1.51503623e-01 -8.38944912e-02
2.71590561e-01 3.48491699e-01 7.98339784e-01 -5.84215343e-01
9.97678936e-01 6.25283480e-01 -6.40942752e-02 -1.49226591e-01
-1.01737952e+00 -3.15612942e-01 -7.93346763e-01 7.52793401e-02
7.25572288e-01 -9.21155155e-01 -2.96598226e-01 5.31871974e-01
-5.69325984e-01 -6.65649235e-01 -8.69225442e-01 6.12161100e-01
-9.83441174e-01 5.47353923e-01 -4.80433285e-01 -3.19006115e-01
-6.60195827e-01 -9.63217437e-01 8.99379432e-01 7.14863762e-02
1.66650444e-01 -1.36224723e+00 2.68462688e-01 3.20251971e-01
6.25439346e-01 6.18649185e-01 5.98412752e-01 -9.01653230e-01
-3.08864176e-01 -3.64378899e-01 1.15606427e-01 9.64437425e-01
1.28736407e-01 -5.61320484e-01 -1.07434094e+00 -5.16319394e-01
2.88544178e-01 -5.47845304e-01 9.49623883e-01 4.64112014e-01
1.54169869e+00 -1.96934894e-01 -2.93546468e-01 1.11473930e+00
1.32068169e+00 -6.65003061e-02 5.69249272e-01 4.31050301e-01
4.51796114e-01 2.93803304e-01 5.56695342e-01 3.30388397e-01
1.23752408e-01 3.73463809e-01 2.99748749e-01 -2.28014037e-01
-1.56620815e-01 -1.48434684e-01 3.68434727e-01 9.88455951e-01
3.77423912e-01 8.74148961e-03 -8.34273219e-01 5.99928856e-01
-1.69630158e+00 -6.03881896e-01 4.65795010e-01 2.47743344e+00
1.28526640e+00 1.94648176e-01 -3.90778743e-02 -9.06183645e-02
6.21805787e-01 -6.64703995e-02 -1.20900428e+00 -3.26491706e-02
-2.63907194e-01 5.58733344e-01 9.25266027e-01 2.34317094e-01
-1.29699647e+00 5.84460557e-01 6.24811840e+00 1.00191343e+00
-9.40272331e-01 6.37711823e-01 9.70412374e-01 -1.30388722e-01
-1.39100119e-01 -3.15697968e-01 -3.47846717e-01 4.00623500e-01
1.13629806e+00 -1.06641591e-01 2.50002205e-01 9.90793645e-01
-4.43366528e-01 1.57695055e-01 -1.32428849e+00 8.45057309e-01
2.59983659e-01 -1.20405376e+00 -3.51233184e-01 -2.25434691e-01
1.08526325e+00 3.66797298e-01 3.55084270e-01 5.33996046e-01
4.77606982e-01 -6.69620872e-01 3.28895509e-01 3.47302884e-01
1.26690042e+00 -7.31781006e-01 9.13066447e-01 6.58918321e-02
-7.38575816e-01 3.97971123e-02 -6.21632934e-02 7.42838323e-01
-2.85379887e-02 5.71890056e-01 -8.75241101e-01 8.19638908e-01
6.49578750e-01 4.91267979e-01 -5.94573438e-01 1.15945804e+00
-2.89555848e-01 8.76827598e-01 -2.80973345e-01 3.91764194e-01
-7.02633187e-02 2.06089258e-01 3.99741381e-01 1.29851270e+00
2.51788199e-01 -2.05233529e-01 6.83004558e-02 7.60234416e-01
-2.80264020e-01 1.51773274e-01 -4.30638492e-01 2.46558428e-01
3.38841140e-01 1.09974551e+00 -4.94985640e-01 -3.36284876e-01
-3.03508818e-01 1.31440496e+00 9.25693437e-02 2.79995620e-01
-1.11501873e+00 -4.69764709e-01 2.53666222e-01 8.76580086e-03
3.71389449e-01 1.53251752e-01 -2.11755037e-01 -1.26095927e+00
3.10730875e-01 -8.82878482e-01 5.99359035e-01 -3.15739840e-01
-1.62096512e+00 5.96889734e-01 6.75931871e-02 -1.58364689e+00
-4.26093370e-01 -3.98854107e-01 -3.89285654e-01 5.77035069e-01
-1.88368785e+00 -1.08136284e+00 -3.03162664e-01 8.18644047e-01
3.25468600e-01 -3.81311893e-01 1.06340015e+00 3.26615483e-01
-3.15664768e-01 1.31939483e+00 5.71302533e-01 3.10150623e-01
1.24372625e+00 -1.40750289e+00 4.24330294e-01 6.48162007e-01
-4.34582075e-03 -1.94216929e-02 3.35744977e-01 -6.06790602e-01
-9.10702288e-01 -1.31491232e+00 1.46035507e-01 -6.50742173e-01
6.82486653e-01 -3.26289237e-01 -1.19634759e+00 7.19464481e-01
1.04980491e-01 6.04247749e-01 9.67268646e-01 -1.22750066e-01
-4.36401367e-01 -3.46928775e-01 -1.55921376e+00 2.90276617e-01
9.10981059e-01 -5.54355085e-01 -4.12407428e-01 5.90205967e-01
9.00278151e-01 -8.64911616e-01 -1.25336826e+00 7.13829219e-01
2.24402800e-01 -6.97921574e-01 9.66601133e-01 -4.79459941e-01
2.09020764e-01 8.27277917e-03 2.51537740e-01 -1.35740125e+00
-1.62691250e-01 -8.14041913e-01 -3.25697601e-01 1.14353096e+00
5.45179665e-01 -9.26190078e-01 1.00627506e+00 5.41305065e-01
-1.91580698e-01 -8.43830466e-01 -1.20770502e+00 -1.22379458e+00
3.46777171e-01 -4.50719476e-01 4.65574384e-01 1.09939051e+00
-9.42994878e-02 -1.14157446e-01 -2.38694787e-01 1.71421364e-01
1.01161218e+00 -1.90153703e-01 5.11820138e-01 -1.04856157e+00
-7.32622564e-01 -1.79159865e-01 -4.28775132e-01 -5.93250155e-01
2.29874939e-01 -1.13893592e+00 -2.59960978e-03 -9.40343380e-01
1.92014292e-01 -5.41673005e-01 -8.47295702e-01 5.63123345e-01
-2.56861091e-01 2.15300933e-01 -2.22240075e-01 7.65904337e-02
-7.15350509e-01 5.98469198e-01 1.28056633e+00 -2.81621218e-01
-2.79972732e-01 9.15987566e-02 -5.39997637e-01 5.18975198e-01
8.53450239e-01 -6.64872289e-01 -6.27535462e-01 -3.71991813e-01
-6.54590270e-03 5.54506108e-02 3.23269755e-01 -1.27871621e+00
3.17578435e-01 5.35016172e-02 4.39164460e-01 -2.72300959e-01
1.58773661e-01 -7.67955899e-01 -5.16750850e-02 5.86305857e-01
-3.58988792e-01 -3.52669477e-01 6.43061459e-01 7.98629463e-01
-2.75329083e-01 -2.66315788e-01 1.12186348e+00 -5.15646897e-02
-4.18238193e-01 4.80694920e-01 3.60643893e-01 4.80521113e-01
1.14506340e+00 -1.60208523e-01 -1.47965595e-01 -1.61892369e-01
-9.91421402e-01 3.49498808e-01 2.99971849e-01 2.08245113e-01
4.25217718e-01 -1.34660709e+00 -8.75195503e-01 3.45920384e-01
2.07951441e-01 2.23109886e-01 5.24329185e-01 9.07496154e-01
-3.06972593e-01 1.61560133e-01 -1.97720394e-01 -6.86738312e-01
-7.74569452e-01 7.85882473e-01 5.50725281e-01 -7.45374978e-01
-4.02133852e-01 9.00818050e-01 2.08807454e-01 -7.74456263e-01
1.90752357e-01 -2.50341445e-01 3.95419091e-01 -2.88183838e-01
2.25386590e-01 1.98531687e-01 2.41017744e-01 5.01454715e-03
-2.88642734e-01 3.97977293e-01 -2.44351566e-01 -5.93387075e-02
1.17054021e+00 1.24674596e-01 3.18465918e-01 3.40116143e-01
1.23020101e+00 -2.62189597e-01 -1.61966467e+00 -7.09057927e-01
-1.37446702e-01 -1.70607284e-01 -2.94297636e-02 -1.26166272e+00
-1.08685482e+00 6.57660544e-01 1.15045953e+00 -2.07089186e-01
1.41065252e+00 4.56488840e-02 8.74501407e-01 4.29232150e-01
2.57772803e-01 -1.51698673e+00 2.77125925e-01 3.07135940e-01
7.29735792e-01 -1.74626446e+00 -8.27278793e-02 -3.12288374e-01
-6.03315175e-01 8.50272357e-01 6.26339495e-01 1.04127079e-02
8.13835621e-01 4.87813681e-01 1.62188280e-02 2.86289245e-01
-4.24173772e-01 4.16556299e-02 3.02957475e-01 8.51316035e-01
1.71373069e-01 -3.67624424e-02 -2.36660615e-03 6.16315663e-01
-2.85410918e-02 1.89032599e-01 1.02058724e-01 7.90162742e-01
8.66964385e-02 -1.20851839e+00 -2.09235236e-01 4.64854509e-01
-6.37670040e-01 -1.25136137e-01 -9.65743139e-03 8.42445850e-01
5.46221621e-03 2.94401139e-01 -1.77458636e-02 -3.43605280e-01
2.26845995e-01 1.85202003e-01 6.02430284e-01 -4.87535834e-01
-5.21838546e-01 -1.51749015e-01 -3.16950053e-01 -4.73517954e-01
-3.11697990e-01 -7.42050946e-01 -1.09566832e+00 1.49951041e-01
-4.81364399e-01 7.74350250e-03 5.73012888e-01 8.15059066e-01
4.87816006e-01 6.96527958e-01 6.69409096e-01 -4.98855054e-01
-8.90279830e-01 -8.93550098e-01 -3.35332304e-01 5.60833693e-01
2.34102234e-01 -5.02205551e-01 -2.08708793e-01 7.91596621e-02] | [14.514430046081543, -1.9173448085784912] |
b2b21fd1-871b-4abb-895b-aefd1e684da2 | jiff-jointly-aligned-implicit-face-function | 2204.10549 | null | https://arxiv.org/abs/2204.10549v1 | https://arxiv.org/pdf/2204.10549v1.pdf | JIFF: Jointly-aligned Implicit Face Function for High Quality Single View Clothed Human Reconstruction | This paper addresses the problem of single view 3D human reconstruction. Recent implicit function based methods have shown impressive results, but they fail to recover fine face details in their reconstructions. This largely degrades user experience in applications like 3D telepresence. In this paper, we focus on improving the quality of face in the reconstruction and propose a novel Jointly-aligned Implicit Face Function (JIFF) that combines the merits of the implicit function based approach and model based approach. We employ a 3D morphable face model as our shape prior and compute space-aligned 3D features that capture detailed face geometry information. Such space-aligned 3D features are combined with pixel-aligned 2D features to jointly predict an implicit face function for high quality face reconstruction. We further extend our pipeline and introduce a coarse-to-fine architecture to predict high quality texture for our detailed face model. Extensive evaluations have been carried out on public datasets and our proposed JIFF has demonstrates superior performance (both quantitatively and qualitatively) over existing state-of-the-arts. | ['Kwan-Yee K. Wong', 'Wenqi Yang', 'Kai Han', 'GuanYing Chen', 'Yukang Cao'] | 2022-04-22 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Cao_JIFF_Jointly-Aligned_Implicit_Face_Function_for_High_Quality_Single_View_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Cao_JIFF_Jointly-Aligned_Implicit_Face_Function_for_High_Quality_Single_View_CVPR_2022_paper.pdf | cvpr-2022-1 | ['face-model', '3d-human-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.44630790e-01 8.39185342e-02 1.31338909e-01 -6.59623146e-01
-8.20514143e-01 -2.11903647e-01 5.41034818e-01 -7.29618967e-01
3.07726324e-01 3.68852288e-01 6.06678545e-01 3.13958704e-01
5.22441082e-02 -7.37042546e-01 -6.96725726e-01 -2.42277816e-01
1.87019125e-01 7.49661982e-01 -5.75983524e-02 -3.06248337e-01
2.38493457e-01 1.09332490e+00 -1.90718055e+00 5.06774127e-01
5.26804209e-01 1.02376139e+00 -7.41344988e-02 4.87512887e-01
3.10393777e-02 2.77357966e-01 4.25168201e-02 -4.62182790e-01
6.18534923e-01 -2.99055666e-01 -6.43217981e-01 5.83183050e-01
1.03166437e+00 -6.92616403e-01 -3.19518894e-01 6.81733549e-01
7.44846880e-01 -1.43223077e-01 6.96242332e-01 -9.48412001e-01
-2.81885982e-01 -5.31041384e-01 -8.21257949e-01 -3.93145531e-01
8.33678305e-01 7.84870908e-02 6.59559548e-01 -1.41632843e+00
9.72000897e-01 1.76881874e+00 9.92477298e-01 8.26389551e-01
-1.31456017e+00 -5.80228090e-01 -6.86114430e-02 -2.78000921e-01
-1.46064973e+00 -1.07958126e+00 8.65857005e-01 -5.36688745e-01
1.09748638e+00 2.22228855e-01 7.53649116e-01 8.13707054e-01
1.20368950e-01 6.12510204e-01 1.37492943e+00 -4.24857467e-01
-1.43411919e-01 -2.72855431e-01 -5.88450670e-01 1.28743017e+00
-2.30028167e-01 3.35753888e-01 -8.59692752e-01 -2.25242928e-01
1.50390244e+00 7.83776641e-02 -3.37654352e-01 -4.94995445e-01
-8.01975489e-01 3.61655802e-01 2.88930386e-01 -2.36183107e-01
-3.57330620e-01 1.02550566e-01 -8.11160058e-02 1.78621367e-01
9.60343421e-01 -1.68917611e-01 -4.83268619e-01 1.66044328e-02
-1.04537833e+00 5.30687511e-01 6.90388918e-01 9.30099785e-01
9.33266997e-01 -1.37244537e-01 -9.08999518e-02 9.63495195e-01
4.70296353e-01 6.82103753e-01 -3.85009646e-01 -1.48748553e+00
5.43526784e-02 5.89622676e-01 1.30273074e-01 -1.07073343e+00
-3.00807506e-01 -3.28225829e-02 -4.38875675e-01 6.24835432e-01
1.22465886e-01 3.68872344e-01 -1.08602560e+00 1.30291259e+00
7.92235553e-01 3.04840535e-01 -5.93959093e-01 1.12532628e+00
9.90836918e-01 2.94695199e-01 -4.12908047e-01 -4.89350595e-02
1.13203013e+00 -8.17858577e-01 -5.68843663e-01 7.25990385e-02
1.75820775e-02 -9.51933503e-01 1.02597427e+00 4.38331425e-01
-1.40358496e+00 -4.42487150e-01 -6.16990447e-01 -2.66340941e-01
2.81388134e-01 4.36867625e-02 4.64210540e-01 6.54657722e-01
-1.43475771e+00 8.36956859e-01 -8.17548990e-01 -3.71428192e-01
7.36381948e-01 6.49343908e-01 -8.60677838e-01 -4.06271398e-01
-3.64220381e-01 8.47197890e-01 -4.92831826e-01 -4.92212921e-02
-8.93162489e-01 -8.56922626e-01 -9.53811288e-01 -1.34237364e-01
1.69303879e-01 -1.12501407e+00 1.19792318e+00 -7.09716678e-01
-1.69604373e+00 1.46837223e+00 -4.92846847e-01 3.36639702e-01
5.43075562e-01 -2.04216331e-01 4.00442779e-02 3.64460081e-01
-1.58791840e-02 8.11933637e-01 1.08988035e+00 -1.66825914e+00
-2.92861670e-01 -7.97809601e-01 -4.56754193e-02 4.14140880e-01
-4.19214033e-02 4.32036482e-02 -7.96087623e-01 -6.42526627e-01
3.01072031e-01 -7.28581667e-01 1.29309684e-01 8.74802291e-01
-1.24818303e-01 1.39343739e-01 8.20294261e-01 -9.66828823e-01
7.29733229e-01 -1.85981798e+00 3.71787786e-01 2.39226446e-01
2.39502057e-01 -1.38851982e-02 -2.67526388e-01 1.49488240e-01
2.63925821e-01 -1.04290508e-01 -2.14218065e-01 -1.10318410e+00
-2.98790857e-02 1.53004572e-01 -2.36495864e-03 5.92282951e-01
3.43732715e-01 8.37356150e-01 -4.28151786e-01 -4.55297977e-01
3.56266111e-01 1.16848266e+00 -1.06294429e+00 3.94821495e-01
-1.68592513e-01 8.15324903e-01 -4.43519711e-01 1.11883616e+00
1.10190797e+00 -2.21176416e-01 1.11355945e-01 -3.80245686e-01
2.22790707e-02 1.28546670e-01 -8.09860468e-01 2.03280759e+00
-5.11643648e-01 1.23666033e-01 6.28453612e-01 -3.44732195e-01
9.52202916e-01 3.49665016e-01 7.06272483e-01 -6.76661074e-01
1.09804690e-01 2.69184530e-01 -6.87512457e-01 -3.62925828e-01
2.55204886e-01 -4.31695282e-01 4.53642756e-01 3.49269450e-01
1.14282310e-01 -4.85831410e-01 -7.46045589e-01 -1.54665112e-01
8.21062624e-01 8.58325124e-01 -3.68535109e-02 -3.31264049e-01
5.08761346e-01 -2.84315944e-01 4.91384268e-01 -6.90380782e-02
-2.68712379e-02 1.29052007e+00 2.27615863e-01 -5.97049952e-01
-1.24939597e+00 -9.21283007e-01 -2.91019231e-01 6.29168749e-01
1.44812306e-02 -5.65647960e-01 -9.38086927e-01 -6.94824338e-01
2.91086167e-01 -8.92342627e-02 -7.49147654e-01 3.31631660e-01
-6.56144083e-01 -2.58488953e-01 1.92331508e-01 4.60209191e-01
4.59683895e-01 -8.34330142e-01 -2.12888062e-01 -9.11752582e-02
-1.64721102e-01 -1.09502852e+00 -7.15925872e-01 -6.04796469e-01
-9.58954573e-01 -9.66629803e-01 -7.46141434e-01 -6.07751846e-01
9.10285056e-01 1.85043231e-01 1.33459151e+00 5.42507231e-01
-4.62381363e-01 7.15749860e-01 -1.10010810e-01 3.73711921e-02
-1.31959990e-01 -4.98510361e-01 2.40963802e-01 1.62834272e-01
-5.63704856e-02 -8.08311403e-01 -9.00734007e-01 6.29531741e-01
-4.90862221e-01 3.11816901e-01 2.32359678e-01 8.15870941e-01
9.79576945e-01 -3.27564836e-01 2.75535733e-01 -9.18541431e-01
-5.37847448e-03 -8.70249271e-02 -4.17150587e-01 8.91354308e-02
-3.80275339e-01 -8.44632387e-02 9.47701707e-02 -5.80142513e-02
-1.23721170e+00 3.17415833e-01 -5.91706336e-01 -7.79122829e-01
-2.14712764e-03 -6.81992248e-02 -4.54391122e-01 -6.12389028e-01
3.02850127e-01 2.31541600e-02 3.72953624e-01 -9.81396079e-01
-6.24806946e-03 4.86388475e-01 3.26810777e-01 -7.51092851e-01
7.85333991e-01 9.37281847e-01 2.38516152e-01 -8.72043431e-01
-7.58876622e-01 -1.97668985e-01 -8.82754922e-01 -4.44172710e-01
6.68513298e-01 -1.15968752e+00 -7.77827561e-01 4.89930868e-01
-1.15653872e+00 -3.83547872e-01 -1.15061432e-01 8.42778534e-02
-1.07872677e+00 2.60820091e-01 -6.18577421e-01 -8.78780663e-01
-4.40201133e-01 -1.13885999e+00 1.99025691e+00 2.33284645e-02
-8.78344178e-02 -8.26326907e-01 2.61331797e-02 7.31950104e-01
3.81308764e-01 5.72332561e-01 6.15348041e-01 4.56813216e-01
-7.86727667e-01 1.35079756e-01 -2.42550641e-01 1.45069435e-01
1.58003464e-01 -7.77611732e-02 -1.31471002e+00 -4.53320324e-01
-1.30075350e-01 -2.86350131e-01 5.63390434e-01 4.47845608e-01
1.11191654e+00 -1.87916070e-01 -2.54600644e-01 1.11267340e+00
1.35704899e+00 -5.83595157e-01 8.67258787e-01 -1.83477074e-01
8.83775890e-01 7.34628618e-01 6.08665705e-01 6.23243153e-01
5.73444307e-01 9.78255153e-01 4.57006454e-01 -1.41693935e-01
-5.82335830e-01 -5.37925243e-01 4.46028769e-01 5.98910928e-01
-6.77274287e-01 1.93180680e-01 -5.53214788e-01 2.76697487e-01
-1.41464221e+00 -7.09276795e-01 7.71210864e-02 2.12812471e+00
6.21041417e-01 -2.58512050e-01 1.74968660e-01 -5.97510785e-02
4.17859077e-01 -1.28071710e-01 -3.35344583e-01 -1.27694964e-01
5.15685044e-02 6.56931102e-01 -6.32643402e-02 7.83438385e-01
-8.18790019e-01 9.73673582e-01 6.40986824e+00 7.97522008e-01
-9.89861071e-01 1.90067738e-01 6.01734161e-01 -1.28445178e-01
-6.61051154e-01 -6.64333552e-02 -8.25346231e-01 7.06438124e-02
5.00032842e-01 3.48415613e-01 5.11961818e-01 5.59227109e-01
2.15748027e-01 -5.73157556e-02 -1.10911143e+00 1.41070509e+00
3.02073210e-01 -1.45606530e+00 1.69785738e-01 5.84909141e-01
9.00794089e-01 -2.02463940e-01 1.46235034e-01 -1.35390803e-01
-1.84607610e-01 -1.42379284e+00 8.48733306e-01 9.60200906e-01
1.52116632e+00 -7.99093068e-01 2.66629130e-01 2.18522668e-01
-1.32467747e+00 3.19744080e-01 -1.60664901e-01 6.24444410e-02
2.21507400e-01 4.06591922e-01 -6.49820209e-01 5.39194465e-01
9.36476886e-01 8.90262425e-01 -2.53817946e-01 7.56991565e-01
1.41483918e-01 5.61680458e-02 -3.03042263e-01 7.26957440e-01
-4.22361016e-01 -1.21029705e-01 5.87135792e-01 8.47024739e-01
4.47697788e-01 5.15590429e-01 1.60464153e-01 9.13675308e-01
-3.07251692e-01 4.71962020e-02 -7.10334837e-01 1.52634010e-01
1.08576104e-01 1.19609809e+00 -5.56391656e-01 -8.06622803e-02
-4.95443791e-01 1.35527861e+00 4.67214227e-01 -4.47194604e-03
-3.15906763e-01 3.42256308e-01 9.81134355e-01 9.87444997e-01
3.40445727e-01 -3.18639517e-01 -2.00054839e-01 -1.24910498e+00
1.46238521e-01 -7.59157598e-01 3.76889594e-02 -8.64396155e-01
-1.51073694e+00 6.57020509e-01 -2.00812027e-01 -1.24027085e+00
-6.40996769e-02 -7.26893246e-01 -2.21980959e-01 1.01858294e+00
-1.61516297e+00 -1.60266805e+00 -4.39250350e-01 7.15555787e-01
6.39239252e-01 -5.77801205e-02 1.10855389e+00 4.00854617e-01
-1.07684597e-01 5.89693248e-01 -3.17100197e-01 -8.41870084e-02
7.73169339e-01 -8.31165314e-01 5.50548553e-01 3.18394572e-01
-1.39074460e-01 6.13943517e-01 3.28242987e-01 -9.14383411e-01
-1.92629266e+00 -9.00968850e-01 6.92053616e-01 -9.30043638e-01
-2.51511693e-01 -4.27980214e-01 -6.23961627e-01 6.35772347e-01
-2.21632391e-01 6.00497365e-01 4.39878047e-01 1.02221165e-02
-4.08569902e-01 1.09213879e-02 -1.75337982e+00 2.47432694e-01
1.66095567e+00 -6.82071269e-01 -2.37396747e-01 2.43783612e-02
2.65593439e-01 -6.45713031e-01 -1.12852991e+00 7.13807225e-01
1.07578397e+00 -1.34812534e+00 1.10295892e+00 -3.22644502e-01
4.73671138e-01 -2.77791232e-01 -4.83706445e-01 -1.05126810e+00
-3.04465175e-01 -7.45785236e-01 -4.29072738e-01 1.02819073e+00
-1.70485288e-01 -2.92544872e-01 1.01167786e+00 6.47224724e-01
-1.55338407e-01 -1.06688321e+00 -8.29936385e-01 -5.49625397e-01
-1.76346511e-01 -7.64346346e-02 7.89737225e-01 7.49399066e-01
-3.50810766e-01 5.57720661e-02 -4.84864861e-01 2.73986328e-02
8.80292416e-01 3.32595170e-01 9.36647952e-01 -1.40441203e+00
-1.27326757e-01 -1.06977828e-01 -5.20004392e-01 -1.28987098e+00
3.04261774e-01 -9.03360546e-01 -9.88328010e-02 -1.34738958e+00
3.68571252e-01 -7.01611578e-01 3.76075178e-01 4.67765480e-01
1.22626491e-01 9.05979097e-01 -9.32670534e-02 9.22954455e-02
-2.64451951e-01 8.31453383e-01 1.69440866e+00 3.28988492e-01
5.42933447e-03 -2.69868612e-01 -4.85062093e-01 7.41364658e-01
1.80939153e-01 -1.26973063e-01 -1.87044874e-01 -6.91969156e-01
-4.61097844e-02 3.67760241e-01 4.90050375e-01 -7.78850675e-01
-2.13659644e-01 -1.55171826e-01 9.03404772e-01 -5.93750119e-01
9.44772124e-01 -9.29018021e-01 3.64463121e-01 -1.33122317e-02
2.44705439e-01 -9.71703753e-02 1.73890769e-01 4.83757764e-01
6.08635955e-02 4.29135233e-01 1.00606799e+00 -3.08255345e-01
-4.69805330e-01 9.60583389e-01 3.86022985e-01 -1.05275683e-01
6.42804503e-01 -4.02272373e-01 2.17835769e-01 -4.47455734e-01
-8.65983009e-01 -9.03043970e-02 1.31931078e+00 4.25647259e-01
1.21599543e+00 -1.53554606e+00 -9.79450524e-01 7.04187930e-01
-7.48898759e-02 -8.03367514e-03 3.90276045e-01 7.66232073e-01
-8.37209880e-01 7.61067821e-03 -4.92332220e-01 -8.76857519e-01
-1.52293336e+00 4.85356450e-02 5.48771203e-01 1.13847815e-01
-9.84544516e-01 8.64340901e-01 3.01056743e-01 -8.72933984e-01
5.68545870e-02 6.84803054e-02 1.42021328e-01 -3.64896089e-01
6.56945825e-01 2.09664896e-01 3.40714365e-01 -1.22840083e+00
-4.31961536e-01 1.31478393e+00 1.86672121e-01 -1.37747303e-01
1.76835465e+00 -2.06021130e-01 -2.04801947e-01 -1.06088564e-01
1.22925735e+00 1.98116615e-01 -1.71459877e+00 -2.52855986e-01
-5.33499777e-01 -1.24864507e+00 1.17843345e-01 -6.98973298e-01
-1.24481606e+00 9.11404848e-01 5.87459862e-01 -6.67650461e-01
1.16017985e+00 6.85513914e-02 8.99628401e-01 -3.06762666e-01
9.13704097e-01 -7.55331159e-01 1.25752136e-01 4.45775509e-01
1.19822538e+00 -1.20361829e+00 2.92006224e-01 -9.70281005e-01
-3.69811952e-01 1.10297167e+00 6.16866350e-01 -2.89378911e-01
9.25069451e-01 4.01942968e-01 -1.82579905e-01 -5.77858806e-01
-6.28955007e-01 -9.37139839e-02 6.98798418e-01 6.52542591e-01
5.48848212e-01 -3.00217718e-02 1.63585916e-01 3.39375079e-01
-1.91699475e-01 1.40421525e-01 2.66742595e-02 9.49390650e-01
-2.53101856e-01 -1.33623135e+00 -5.53484023e-01 2.95010239e-01
-2.66652048e-01 1.54760182e-01 -3.21623862e-01 4.58989859e-01
5.65560907e-02 6.25483930e-01 9.00049582e-02 -4.03220147e-01
5.48188627e-01 -3.48611437e-02 1.37007463e+00 -8.35544705e-01
-3.85311604e-01 3.17338705e-01 1.58051878e-01 -1.00171113e+00
-4.28865582e-01 -6.45440400e-01 -1.18337214e+00 -7.30403781e-01
-4.59982641e-02 -4.59734410e-01 6.30370855e-01 6.72145307e-01
7.43870020e-01 -1.25136822e-01 7.17180789e-01 -1.71627867e+00
-1.44145012e-01 -6.55126691e-01 -6.26921237e-01 4.86615986e-01
5.04264057e-01 -1.10202086e+00 -1.47876218e-01 8.48557204e-02] | [13.142935752868652, -0.03410336747765541] |
8819b0a4-865b-47b5-9ea4-77d1af258089 | in-defense-of-pseudo-labeling-an-uncertainty-1 | 2101.06329 | null | https://arxiv.org/abs/2101.06329v3 | https://arxiv.org/pdf/2101.06329v3.pdf | In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learning | The recent research in semi-supervised learning (SSL) is mostly dominated by consistency regularization based methods which achieve strong performance. However, they heavily rely on domain-specific data augmentations, which are not easy to generate for all data modalities. Pseudo-labeling (PL) is a general SSL approach that does not have this constraint but performs relatively poorly in its original formulation. We argue that PL underperforms due to the erroneous high confidence predictions from poorly calibrated models; these predictions generate many incorrect pseudo-labels, leading to noisy training. We propose an uncertainty-aware pseudo-label selection (UPS) framework which improves pseudo labeling accuracy by drastically reducing the amount of noise encountered in the training process. Furthermore, UPS generalizes the pseudo-labeling process, allowing for the creation of negative pseudo-labels; these negative pseudo-labels can be used for multi-label classification as well as negative learning to improve the single-label classification. We achieve strong performance when compared to recent SSL methods on the CIFAR-10 and CIFAR-100 datasets. Also, we demonstrate the versatility of our method on the video dataset UCF-101 and the multi-label dataset Pascal VOC. | ['Mubarak Shah', 'Yogesh S Rawat', 'Kevin Duarte', 'Mamshad Nayeem Rizve'] | 2021-01-15 | in-defense-of-pseudo-labeling-an-uncertainty | https://openreview.net/forum?id=-ODN6SbiUU | https://openreview.net/pdf?id=-ODN6SbiUU | iclr-2021-1 | ['semi-supervised-video-classification', 'semi-supervised-medical-image-classification'] | ['computer-vision', 'medical'] | [ 4.84923691e-01 2.36863717e-01 -5.26738405e-01 -8.88580382e-01
-1.24027240e+00 -5.32840967e-01 6.89926505e-01 2.19547957e-01
-5.12419701e-01 1.12504041e+00 -1.50157094e-01 -5.89069165e-02
2.41699532e-01 -4.52714205e-01 -9.23850000e-01 -6.83881938e-01
5.56679010e-01 6.44509733e-01 3.62802036e-02 2.39717424e-01
-1.69382498e-01 -7.03220665e-02 -1.64596498e+00 4.92808104e-01
7.65170753e-01 9.84301984e-01 -1.22974351e-01 2.02089086e-01
-5.82306497e-02 1.04757380e+00 -3.25872809e-01 -4.63387340e-01
1.56922698e-01 -3.45670968e-01 -8.30042064e-01 3.11368555e-01
7.22885907e-01 -1.66313071e-02 3.29366595e-01 1.21605670e+00
1.72191992e-01 1.53976530e-01 8.49926293e-01 -1.56181705e+00
-5.36059141e-01 5.50641596e-01 -6.40569866e-01 -4.85788375e-01
8.82234573e-02 -2.98263103e-01 1.04084980e+00 -1.15685213e+00
6.92599297e-01 1.35631502e+00 8.26834500e-01 9.48685944e-01
-1.34132755e+00 -7.52605736e-01 1.50863081e-01 5.45648858e-03
-1.25247371e+00 -3.16447765e-01 6.00164473e-01 -4.49914724e-01
4.11998510e-01 2.35622525e-01 2.17352688e-01 1.44374084e+00
-4.17789429e-01 1.08676493e+00 1.65941370e+00 -8.05428684e-01
3.38365138e-01 4.99080569e-01 4.10724938e-01 6.12923026e-01
1.76714659e-01 2.63276339e-01 -5.90063155e-01 -2.59655148e-01
3.93781453e-01 -1.41743541e-01 -2.25189522e-01 -5.34573317e-01
-1.08922374e+00 8.08131278e-01 1.23699076e-01 1.80722307e-02
6.43730164e-06 5.44288708e-03 4.73695874e-01 2.15093926e-01
7.76459515e-01 5.34211516e-01 -7.47510552e-01 2.10244790e-01
-9.52918649e-01 1.42149433e-01 5.56285977e-01 1.00004053e+00
8.47719908e-01 -7.21968040e-02 -3.65681827e-01 1.16560340e+00
5.67538261e-01 4.21844274e-01 4.04947251e-01 -1.10418379e+00
1.65843129e-01 5.31218231e-01 2.25253239e-01 -3.69757831e-01
-5.25284231e-01 -7.27671981e-01 -9.04291213e-01 2.14760542e-01
3.51547658e-01 -2.67877243e-02 -1.19075477e+00 2.10440636e+00
3.91655475e-01 3.30880642e-01 1.12797394e-01 6.30972743e-01
9.26219523e-01 3.83287072e-01 5.26237071e-01 -5.56154609e-01
8.71008813e-01 -1.44623494e+00 -9.98703480e-01 -3.70735973e-01
1.18798625e+00 -5.51470041e-01 1.17106724e+00 4.74377394e-01
-5.92460752e-01 -5.64684272e-01 -9.57500756e-01 1.85935661e-01
-3.62235546e-01 3.17038029e-01 6.84577823e-01 6.59800291e-01
-7.90346742e-01 5.47138155e-01 -4.71746951e-01 -6.85288012e-02
6.08342052e-01 2.37200528e-01 -5.96356630e-01 -3.60795915e-01
-1.18163443e+00 9.19921637e-01 6.66092753e-01 -2.29008436e-01
-7.42772758e-01 -7.11909831e-01 -1.01032329e+00 -3.40203077e-01
6.80934608e-01 -3.73677760e-01 1.24207342e+00 -1.35531569e+00
-1.19238555e+00 1.17964649e+00 -8.35206881e-02 -4.19770539e-01
6.66595817e-01 -1.88970909e-01 -5.02278566e-01 -8.02618265e-02
2.70932257e-01 1.14762354e+00 9.77295637e-01 -1.82982552e+00
-5.64253867e-01 -1.16131559e-01 -1.36260793e-01 2.85189211e-01
-2.97287852e-01 -3.36802006e-01 -2.19666690e-01 -7.64181316e-01
2.56641597e-01 -1.32899940e+00 -2.52982140e-01 -1.53896451e-01
-5.44037998e-01 -3.46915632e-01 7.45981038e-01 -3.72220948e-02
9.24712658e-01 -2.15782046e+00 -1.80976942e-01 1.02207042e-01
-4.85172644e-02 5.43455184e-01 -2.76958227e-01 -1.55415326e-01
-3.91622394e-01 3.09931964e-01 -3.54924440e-01 -8.20918798e-01
-5.42013673e-03 4.53368127e-01 -2.75886744e-01 2.90151000e-01
3.30300122e-01 8.05910051e-01 -1.06510496e+00 -7.22954154e-01
3.00990760e-01 2.83657670e-01 -3.11820596e-01 9.96120945e-02
-3.41842026e-01 7.00732410e-01 -6.23963773e-02 7.99713135e-01
7.02686250e-01 -5.18573046e-01 1.46039501e-01 -2.77686030e-01
3.08417708e-01 -8.69615451e-02 -1.19526100e+00 1.51534033e+00
-3.60874116e-01 5.12442030e-02 -2.71794081e-01 -9.68679965e-01
7.00797737e-01 4.05463040e-01 3.87533188e-01 -3.46758634e-01
1.56231239e-01 4.26596224e-01 -5.56620479e-01 -2.98639834e-01
2.75085151e-01 -4.32324529e-01 -1.58678126e-02 4.72441196e-01
4.10876513e-01 -2.05783784e-01 3.41166764e-01 2.18084738e-01
6.46741807e-01 6.09218538e-01 2.59234101e-01 -1.16019934e-01
6.32805347e-01 4.00454877e-03 8.17161620e-01 8.16742301e-01
-3.55521470e-01 8.32004607e-01 1.49460569e-01 -1.34021401e-01
-8.31217587e-01 -8.96129847e-01 -5.47531903e-01 1.15801954e+00
8.95745978e-02 -4.48250830e-01 -5.82732320e-01 -1.31851339e+00
-1.32341191e-01 9.59889770e-01 -6.55759871e-01 -2.89503753e-01
-3.73023972e-02 -1.25652122e+00 4.57694978e-01 5.03702044e-01
3.87207061e-01 -8.27832639e-01 1.66054830e-01 4.88561057e-02
-4.66033697e-01 -1.36669910e+00 -9.57293436e-02 6.09764695e-01
-7.29325056e-01 -1.11712718e+00 -5.06390810e-01 -7.45955467e-01
9.61346924e-01 -3.27100568e-02 1.17961049e+00 -1.34269699e-01
9.95699465e-02 2.99778879e-01 -5.92080951e-01 -4.88621682e-01
-7.85688281e-01 -2.02311859e-01 3.22471619e-01 2.34766543e-01
5.19824803e-01 -1.23022176e-01 7.49330968e-02 4.22146916e-01
-8.14294636e-01 2.43371189e-01 4.08199698e-01 1.37610316e+00
1.08792150e+00 -9.60220844e-02 1.00098979e+00 -1.61486471e+00
1.56091318e-01 -4.04714048e-01 -3.73414338e-01 5.92773259e-01
-1.05251122e+00 1.25932544e-01 5.49070776e-01 -6.79289639e-01
-1.30992806e+00 4.00541663e-01 -1.06013790e-01 -5.14973640e-01
-2.72778988e-01 2.41381928e-01 3.04001551e-02 -2.49701142e-01
8.71101618e-01 -2.36157969e-01 -2.65801907e-01 -5.11079133e-01
4.15472567e-01 6.75330997e-01 4.69128102e-01 -5.95795512e-01
6.72580183e-01 5.00501156e-01 1.74040243e-01 -3.75333130e-01
-1.87003434e+00 -5.69390833e-01 -7.39672005e-01 -2.38086194e-01
6.71905041e-01 -1.16836917e+00 -1.25186592e-01 4.96914029e-01
-7.17058718e-01 -2.73019135e-01 -5.09666383e-01 6.11268520e-01
-5.58860004e-01 4.55808282e-01 -5.22863925e-01 -8.56077373e-01
-8.50451961e-02 -1.01847482e+00 1.12082112e+00 9.06343162e-02
-2.47674540e-01 -9.74109471e-01 -1.22775007e-02 6.24056339e-01
6.79808781e-02 2.06993192e-01 8.37699533e-01 -8.62222373e-01
-5.39460033e-02 -2.77113676e-01 -2.12237209e-01 8.91784549e-01
-1.91041291e-01 -1.97266877e-01 -1.39211679e+00 -3.22869182e-01
-1.36935368e-01 -1.17460287e+00 9.71218348e-01 2.27200747e-01
1.17495954e+00 1.90824658e-01 -2.34338537e-01 3.53810430e-01
1.42882180e+00 -1.64220124e-01 2.05191255e-01 2.60669500e-01
8.71749461e-01 7.38151073e-01 1.18091142e+00 2.11667895e-01
2.79167444e-01 5.77360928e-01 5.66440821e-01 -2.66831756e-01
-2.70096928e-01 -2.61248201e-01 8.08625668e-02 6.38152003e-01
2.02892378e-01 -1.89708829e-01 -7.38857806e-01 3.31496477e-01
-2.08336473e+00 -5.32736838e-01 -5.96074402e-01 2.33194804e+00
1.05962563e+00 1.51684145e-02 -3.39417130e-01 3.78532112e-01
7.23387599e-01 -9.31318998e-02 -4.28998351e-01 8.01259801e-02
-3.63268584e-01 1.65366501e-01 5.03141582e-01 4.25561816e-01
-1.61965394e+00 9.58733380e-01 6.27815008e+00 1.04359126e+00
-6.82862461e-01 5.29850304e-01 9.11422253e-01 3.27411108e-02
-1.73135564e-01 -1.97165906e-02 -1.06679094e+00 1.48435980e-01
7.83868849e-01 4.83670652e-01 3.58037949e-02 1.14548135e+00
-3.74578476e-01 -4.11850989e-01 -1.28080738e+00 1.19269800e+00
2.90760964e-01 -9.56912935e-01 4.03859764e-02 -3.13551486e-01
1.20642626e+00 7.34048802e-03 -3.77702410e-04 5.88680089e-01
3.61664563e-01 -9.09476459e-01 8.68260860e-01 2.71722585e-01
1.04517937e+00 -5.57516277e-01 9.42142308e-01 5.16428292e-01
-6.18173063e-01 -1.34526953e-01 -3.35724592e-01 1.51294783e-01
1.85180768e-01 9.07104969e-01 -6.60583079e-01 4.88337785e-01
5.80438197e-01 6.86677396e-01 -7.95550406e-01 7.52556086e-01
-5.62343001e-01 7.49224067e-01 -2.78567582e-01 4.28500056e-01
1.02066815e-01 -4.26987000e-02 1.13112248e-01 1.08022130e+00
-3.75464521e-02 -1.84518471e-01 3.83971930e-01 5.92688382e-01
-3.43157351e-01 2.11938545e-01 -3.98685902e-01 1.24418452e-01
3.88585269e-01 1.28720522e+00 -5.19054711e-01 -6.44449115e-01
-4.23677534e-01 7.04998314e-01 5.55873334e-01 2.69931048e-01
-8.26658547e-01 2.61693865e-01 -4.99766134e-02 -2.23141938e-01
-3.28597158e-01 3.29861313e-01 -3.34644228e-01 -1.40304673e+00
-3.21105570e-02 -7.09402323e-01 4.13530737e-01 -7.80320942e-01
-1.55910659e+00 4.97516006e-01 -4.05819453e-02 -1.33988202e+00
-2.54055887e-01 -3.74593824e-01 2.49444276e-01 5.85930765e-01
-1.77331233e+00 -1.40367329e+00 -1.91505522e-01 3.99113774e-01
4.56456870e-01 1.11533850e-02 1.15074956e+00 5.73663652e-01
-5.49386799e-01 6.73115075e-01 1.43229991e-01 -1.13328077e-01
1.31950486e+00 -1.34434938e+00 -1.64100394e-01 7.04429865e-01
3.21360171e-01 7.34655485e-02 4.77751195e-01 -6.52373135e-01
-5.07806063e-01 -1.50588310e+00 9.33767796e-01 -6.54829443e-01
3.51067603e-01 -3.09967726e-01 -9.20587003e-01 7.84135640e-01
-3.38030338e-01 4.53284621e-01 1.00233054e+00 9.45596844e-02
-6.44400001e-01 9.98580605e-02 -1.33844399e+00 2.65941501e-01
8.99418414e-01 -4.35438007e-01 -4.37664330e-01 8.52858603e-01
5.44626534e-01 -3.76724690e-01 -6.40233576e-01 8.82591307e-01
2.15573832e-01 -8.42083037e-01 7.77538836e-01 -4.85948414e-01
3.58268827e-01 -4.02920574e-01 -3.22424531e-01 -1.17654717e+00
-2.11163536e-01 2.41763908e-02 -1.27877921e-01 1.53500593e+00
6.15347445e-01 -4.51210141e-01 7.01731682e-01 6.68442547e-01
-5.30243777e-02 -6.18799567e-01 -6.86910033e-01 -8.81545067e-01
-1.11430012e-01 -7.10313737e-01 2.62150198e-01 1.44092846e+00
-3.03783506e-01 4.89634782e-01 -6.82229936e-01 -3.18753943e-02
9.80270505e-01 -4.64342795e-02 4.65868860e-01 -1.42527103e+00
-2.90341437e-01 1.98193491e-01 3.62542607e-02 -6.65581644e-01
5.61004162e-01 -1.02647448e+00 2.32654169e-01 -1.10344100e+00
3.17233533e-01 -1.01755762e+00 -4.09499109e-01 9.23231721e-01
-2.25754440e-01 9.36427951e-01 3.53094302e-02 3.73569459e-01
-9.25875306e-01 5.29351532e-01 1.17867196e+00 4.47425572e-03
7.43263140e-02 6.76882046e-04 -4.74394202e-01 1.07496738e+00
5.44842839e-01 -8.02722752e-01 -4.72918779e-01 -1.00475408e-01
2.95318991e-01 -2.40412220e-01 2.07967937e-01 -8.78815591e-01
-2.51657695e-01 -4.02350910e-02 3.23170185e-01 -4.57670122e-01
3.79451543e-01 -1.00411499e+00 6.79679662e-02 1.99350387e-01
-6.64404333e-01 -6.60655797e-01 -1.74541891e-01 8.28357279e-01
-3.33947361e-01 -5.64549208e-01 1.19926882e+00 -1.68147296e-01
-6.92112267e-01 1.63299367e-01 6.97375685e-02 2.60453433e-01
1.19663930e+00 1.20323621e-01 -2.20277131e-01 -9.53456759e-02
-1.07801580e+00 3.16286981e-01 4.67665702e-01 4.07614738e-01
5.47657311e-02 -1.56971443e+00 -4.83616263e-01 1.29069418e-01
4.08277631e-01 2.25294888e-01 1.87224209e-01 8.18702757e-01
-9.02579725e-02 1.86406359e-01 1.44568518e-01 -7.68742561e-01
-1.30266261e+00 6.20572031e-01 2.21205637e-01 -4.97816652e-01
-1.42546311e-01 1.12889373e+00 2.38726661e-01 -8.34669769e-01
4.59844798e-01 1.63581312e-01 -3.70928317e-01 1.78252786e-01
4.82827991e-01 1.55015752e-01 1.80477560e-01 -7.53198445e-01
-2.65827894e-01 4.73892957e-01 -1.22477122e-01 5.94968162e-02
1.14759004e+00 -1.17601246e-01 -9.56171006e-02 7.31315136e-01
1.03291667e+00 -2.97308981e-01 -1.23241389e+00 -4.36797172e-01
1.81447774e-01 -3.50989103e-01 1.73584968e-02 -1.15099943e+00
-8.97430182e-01 7.23125517e-01 6.35766089e-01 -1.87931895e-01
9.17859256e-01 -5.88181913e-02 3.97156864e-01 3.95876080e-01
6.20290995e-01 -1.45731366e+00 2.75251232e-02 3.42412710e-01
4.24683988e-01 -1.89216065e+00 8.05930048e-02 -9.13028359e-01
-8.98167491e-01 8.70470822e-01 8.16884398e-01 2.44124547e-01
5.87372303e-01 2.06074476e-01 2.82167643e-01 6.33636191e-02
-7.42517889e-01 -1.80310354e-01 2.58861035e-01 5.90734363e-01
3.92490685e-01 1.69714347e-01 -3.71797532e-01 5.97338438e-01
3.28686178e-01 1.69584557e-01 4.80165958e-01 9.39025939e-01
-1.27751350e-01 -1.44455945e+00 -3.79410654e-01 6.51230454e-01
-4.58709031e-01 -4.40060981e-02 -2.39857227e-01 5.73626935e-01
4.19094145e-01 1.03746676e+00 -3.82386118e-01 -3.15290242e-01
-2.70435810e-02 5.17816365e-01 4.00133908e-01 -9.83716369e-01
-3.25785428e-01 2.43547913e-02 3.15455526e-01 -3.36568564e-01
-1.09982538e+00 -5.36240995e-01 -1.23053229e+00 2.61083901e-01
-8.48626792e-01 1.33534253e-01 6.51958406e-01 1.11227870e+00
4.29676846e-02 3.46503019e-01 4.65482175e-01 -6.12397373e-01
-6.77955329e-01 -1.05850399e+00 -7.20618606e-01 8.83895636e-01
8.78295377e-02 -1.05507886e+00 -6.52212679e-01 1.65254667e-01] | [9.50585651397705, 3.924499273300171] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.